WORKS TRANSLATED BY WILLIAM T. HALL PUBLISHED BY JOHN WILEY & SONS, Inc. P. P. TREADWELL'S ANAL%HCAL CHEMISTRY IN Two VOLUMES. Vol. I. Qualitative Analysis. 8vo, xv +534 pages. Cloth, $3.00. Vol. II. Quantitative Analysis. 8vo, Jri +908 pages, 126 figures. Cloth, $4.00. (With G. W. ROLFE.) H. Claassen's Beet-Sugar Manufacture. 8vo, xiv + 287 pages. Cloth, $3.00. (With A. A. BLANCHARD.) H. and W. Blitz's Laboratory Methods of Inorganic Chemistry. 8vo, xv + 258 pages, 26 figures. Cloth, $3.00. (With J. W. PHELAN.) H. Biltz's Introduction to Experimental Inorganic Chemistry. 12mo, vi + 185 pages. Cloth, $1.25. (With GEORGE DEFREN.) Emil Abderhalden's Text-book of Physiological Chemistry. 8vo, xiii + 722 pages. Cloth, $5.00. (With R. S. WILLIAMS.) W. Ostwald's Introduction to Chemistry. Large 12mo, v + 368 pages, 74 figures. Cloth, $1.50 net. (With C. R. HAYWARD.) W. Borchers's Metallurgy. 8vo, v + 271 pages, 218 figures. Cloth, $3.00 net. (In preparation.) Ready in 1913. A. Classen's Quantitative Analysis by Electrol- ysis. ANALYTICAL CHEMISTRY, BY F. P. TEE AD WELL, -Pn.D., Professor of Analytical Chemistry in the Polytechnic Institute of Zurich, AUTHORIZED TRANSLATION FROM THE GERMAN BY WILLIAM T. HALL, S.B., Assistant Professor of Chemistry, Massachusetts Institute of Technology, VOLUME II. QUANTITATIVE ANALYSIS. THIRD EDITION, REVISED AND ENLARGED TOTAL ISSUE, TEX THOUSAND NEvV YOEK JOHN WILEY & SONS, INC. LONDON: CHAPMAN & HALL, LIMITED 1914 Copyright, 1904, 1910, 1911, BY WILLIAM T. HALL. First and Second Editions entered at Stationers' HalL THE SCIENTIFIC PRESS OiERT ORUMMONO AND COMPANY BROOKLYN. N. Y. TRANSLATOR'S NOTE. THIS translation has been made from the second German edition, but Professor Treadwell has kindly indicated quite a number of changes which he intends to make in the third edition. Since it has been my aim not so much to prepare an exact literal translation as to publish a book which will be useful to English- speaking students, I am under great 'obligations to several of my friends and colleagues for suggesting certain other changes. That part of the proof relating to Gravimetric Analysis has been care- fully read and criticised by Professor Henry Fay, that relating to Volumetric Analysis by Professor F. Jewett Moore, and Pro- fessor Augustus H. Gill has twice read the chapter on Gas Analysis. I have also received valuable assistance in reading the proof from Messrs. R. S. Williams, F. R. Kneeland and J. R. Odell, all of the Massachusetts Institute of Technology. I am indebted to Mr. A. R. Jackson, of Winthrop, for several drawings. WILLIAM T. HALL. MASSACHUSETTS INSTITUTE OF TECHNOLOGY, April, 1904. In preparing the second edition, the text has been compared with the fourth German edition, and certain additions have been made which are not found in the German text. W. T. H. March, 1910. The third edition has been compared with the fifth German edition. In the preparation of dilute acids the volume of con- centrated acid is always referred to the volume of water added. Thus sulphuric acid (1:4) signifies one volume of concentrated acid diluted with four volumes of water. With solutions of solids, the first number signifies the weight in grams of the solid and the second the volume of the solvent in cubic centi- meters. W. T. H. June, 1911. iii 299024 TABLE OF CONTENTS. INTRODUCTION. PAGE Gravimetric and Volumetric Analysis 1 Direct and Indirect Analyses 2 Weighing 6 Reduction of Weighing to Vacuo 13 Testing of Weights 15 Filtration and Washing of Precipitates 18 Drying and Igniting of Precipitates 21 Evaporation of Liquids 30 Drying Substances in Currents of Gases 33 Preparation of the Substance for Analysis 35 Recrystallization 35 PART I. GRAVIMETRIC DETERMINATION OF THE METALS. GROUP V (ALKALIES). Potassium 38 Sodium 43 Separation of Potassium from Sodium 43, 50 Lithium ^53 Determination of Lithium, Potassium, and Sodium 53 Ammonium 57 Magnesium 65 Separation of Magnesium from the Alkalies 68 GROUP IV (ALKALINE EARTHS). Calcium 70 Strontium 72 Barium 74 Separation of Calcium from Magnesium 76 v vi TABLE OF CONTENTS. PAGE Separation of Strontium from Magnesium 78 Separation of Barium from Magnesium 79 Separation of the Alkaline Earths from One Another 79 GROUP III. Aluminium 82 Iron 87 Titanium 100 Chromium 102 Uranium 106 Separation of Group III from Group IV 107, 147 Separation of Iron from Aluminium 107 Separation of Iron, Aluminium, and Phosphoric Acid Ill Separation of Iron from Chromium 113 Separation of Aluminium from Chromium 114 Separation of Iron from Titanium 114 Separation of Aluminium from Titanium 116 Separation of Uranium from Iron and Aluminium 119 Manganese 120 Nickel 129 Cobalt 138 Zinc 140 Separation of Manganese, Nickel, Cobalt, and Zinc from the Alkaline Earths 147 Separation of the Bivalent from the Other Metals of the Ammonium Sulphide Group 149 Separation of Zinc from Nickel, Cobalt, and Manganese 156 Separation of Manganese from Nickel and Cobalt 161 Separation of Cobalt from Nickel 161 Separation of Nickel from Zinc 165 Separation of Nickel from Manganese 165 Separation of Nickel from Iron 166 Removal of Ferric Chloride by Ether 167 GROUP II. (a) Sulpho-Bases. Mercury 168 Lead 174 Bismuth 179 Copper .182 Cadmium 189 Separation of the Sulpho-Bases from the Preceding Groups 192 TABLE OF CONTENTS. vu PAGE Analysis of Brass 193 Separation of the Sulpho-Bases from One Another 194 (b) Sulpha-Acids. Arsenic 205 Antimony 218 Tin 228 Separation of Arsenic, Antimony, and Tin from Members of the Ammo- nium Sulphide Group 235 Separation of Arsenic, Antimony, and Tin from Mercury, Lead, Copper, Cadmium, and Bismuth 235 Analysis of Bronzes 236 Separation of the Sulpho- Acids from One Another 241 Analysis of Bearing Metal 252 Gold 257 Platinum 268 Separation of Gold from Platinum 271 Analysis of Commercial Platinum 272 Selenium 277 Tellurium 279 Separation of Selenium and Tellurium from the Metals of Groups III, IV, and V 279 Separation of Selenium and Tellurium from Metals of Group II 280 Separation of Selenium from Tellurium 282 Molybdenum 284 Tungsten 288 Separation of Molybdenum from Tungsten 293 Analysis of Wolframite 296 Analysis of Tungsten Bronzes 298 Separation of Tungsten from Tin 300 Separation of Tungstic Acid from Silica 302 Vanadium 303 Separation of Vanadium from Arsenic Acid 306 Separation of Vanadium from Phosphoric Acid 307 Separation of Vanadium from Molybdenum 308 Analysis of Vanadinite 308 Determination of Vanadium and Chromium in Iron Ores and Rocks. ... 310 Determination of Vanadium and Chromium in Pig Iron 312 Determination of Vanadium, Molybdenum, Chromium, and Nickel in Steel , , 313 GROUP I. Silver.. ..,. 317 viii TABLE OF CONTENTS. GRAVIMETRIC DETERMINATION OF THE METALLOIDS. GROUP I.* PAGE Hydrochloric Acid 320 Analysis of an Insoluble Chloride 323 Free Chlorine 324 Chlorine in Organic Compounds 325 Hydrobromic Acid 329 Hydriodic Acid " 330 Separation of the Halogens from One Another 331 Hydrocyanic Acid 337 Determination of Hydrocyanic Acid in the Presence of Halogen Hydride 339 Sulphocyanic Acid 339 Determination of Sulphocyanic and Hydrocyanic Acids 342 Determination of Sulphocyanic Acid and Halogen Hydrides 342 Hydroferrocyanic Acid 342 Hydroferricyanic Acid 344 Hypochlorous Acid 344 GROUP II. Nitrous Acid 344 Hydrosulphuric Acid, H 2 S 347 Analysis of Tetrahedrite 359 Acetic Acid 371 Cyanic Acid 371 Determination of Cyanic, Hydrocyanic, and Carbonic Acids 371 Hypophosphorous Acid 372 GROUP III. Sulphurous Acid 373 Selenous and Tellurous Acids 374 Phosphorous Acid 374 Carbonic Acid 375 Determination of Carbon 398 Determination of Carbon and Hydrogen in Organic Substances 414 Dumas Method for Determining Nitrogen 422 Oxalic Acid 427 Boric Acid 428 Molybdic Acid 433 Tartaric Acid 433 Meta- and Pyrophosphoric Acids 433 lodic Acid .433 * For the Division of Acids into Groups cf. Vol. I. TABLE OF CONTENTS. ix GROUP IV. PAGE Phosphoric Acid 434 Determination of Phosphorus and Silicon in Iron and Steel 440 Separation of Phosphoric Acid from the Metals . . . 448 Thiosulphuric Acid 450 GROUP V. Nitric Acid 451 Chloric Acid 460 Perchloric Acid 462 GROUP VI. Sulphuric Acid . 464 Hydrofluoric Acid 471 Separation of Phosphoric and Hydrofluoric Acids 474 Separation of Fluorine from the Metals 481 Separation of Fluorine from the Acids 482 Hydrofluosilicic Acid 483 GROUP VII. Silicic Acid 485 Analysis of Silicates 491 Determination of Zirconium and Sulphur in Rocks 505 Analysis of Chromite 509 Determination of Thorium in Monazite 510 Determination of Water in Silicates 512 PART II. VOLUMETRIC ANALYSIS. Measuring Instruments 514 Normal Volume and Normal Temperature 516 Calibration of Measuring Flasks 522 Calibration of Pipettes 524 Calibration of Burettes 527 Normal Solutions 530 I. ACIDIMETRY AND ALKALIMETRY. Indicators 538 Alkalimetry 558 Acidimetry 571 X TABLE OF CONTENTS. II. OXIDATION AND REDUCTION METHODS. PAGE Permanganate Methods 596 Potassium Dichromate Methods 641 lodimetry 644 Reduction Methods 697 III. PRECIPITATION ANALYSES. Determination of Silver 702 Determination of Halogens 707 Determination of Cyanogen 710 Determination of Sulphocyanic Acid 712 Determination of Sulphuric Acid 714 Determination of Phosphoric Acid 718 Determination of Nickel 720 Determination of Copper 724 Determination of Lead 726 PART III. GAS ANALYSIS. The Collection and Confinement of Gas Samples 730 Calibrating Gas Measuring Instruments 743 Purification of Mercury 747 Determination of Carbon Dioxide 750 Ethylene - 751 Benzene 752 Acetylene 754 Separation of the Heavy Hydrocarbons 756 Oxygen 757 Carbon Monoxide 762 Hydrogen 770 Methane 774 Analysis of Illuminating and Producer Gas 775 Technical Gas Analysis 786 Method of Hempel 768 Method of Winkler-Dennis 794 Orsat's Apparatus 797 Bunte's Apparatus 798 Analysis of Gases which are Absorbed by Water 800 Nitrous Oxide 800 Nitric Oxide . . .802 TABLE OF CONTENTS. XI PAGE Nitrogen 806 Analysis of Gases by Titration 808 Chlorine 808 Hydrochloric Acid 814 Sulphur Dioxide 815 Hydrogen Sulphide 816 Ethylene 818 Gas-Volumetric Methods 822 Determination of Ammonia in Ammonium Salts 822 Determination of Nitrous and Nitric Acids 825 Hydrogen Peroxide Methods 826 Standardization of Permanganate Solutions 827 Determination of Cerium in Soluble Salts 828 Silicon Fluoride 828 Determination of Fluorine 829 Determination of Vapor in Gas Mixtures 831 APPENDIX. Specific Gravity of Acids 838 Specific Gravity of Alkalies 840 Tension of Aqueous Vapor 842 Heats of Combustion of Gases 845 Tables for Calculating Analyses 847 International Atomic Weights 849 Table of Chemical Factors 850 Logarithms 854 APPENDIX I. The Influence of Fine Grinding on Composition 859 The Use of Cupferron in Quantitative Analysis 860 Preparation of Cupferron 862 The Determination of Titanium Dioxide in Titanium-Iron Ore 866 Determination of Sulphur in Pyrite: Fusion with Sodium Peroxide 870 A Rapid Method for the Determination of Sulphur in Steel 872 Requisite Solutions 873 The Determination of Chromium in Steel 876 Oxidation by the Barba Method 876 Oxidation by Sodium Bismuthate 878 INDEX . 881 QUANTITATIVE ANALYSIS. INTRODUCTION. THE purpose of a quantitative analysis is to determine the quantity of the constituents present in a compound or in a mix- ture. The methods to be employed depend upon the nature of the substances to be determined, so that in every case a qualitative analysis should precede the quantitative one. In quantitative analy- sis we distinguish two essentially different methods of procedure: I. GRAVIMETRIC ANALYSIS (Analysis by Weight). II. VOLUMETRIC ANALYSIS (Analysis by Volume). In the case of gravimetric analysis we separate the component to be determined from a solution in the form of an insoluble com- pound of known chemical composition, and then determine the weight of this compound; from this we can calculate the amount of the substance present. Suppose, for example, that we have for analysis a sample of barium chloride. The amount of barium present can be determined by dissolving a weighed amount of the chloride in water, precipitating the barium from the solution by the addi- tion of sulphuric acid and weighing the insoluble barium sulphate formed. If we start with a grams of barium chloride and obtain p grams of barium sulphate, the amount of barium present may be calcu- lated as follows: BaSO 4 :Ba=p:s. =weight of barium in a gm. of barium chloride. 2 INTRODUCTION. It is, however, customary to express the results in percentages; therefore in this case we have 100 -Ba p x= ,. .,,. = per cent, banum. BabO 4 a In the case of volumetric analysis the constituents are not weighed, but they are determined by measuring the amounts of reagents of known strength which react with them. Suppose that we have a sample of caustic soda which contains some sodium chloride as an impurity and that we desire to know how much caustic soda there is in 100 gms. of the mixture. A portion of the substance weighing a gms. is dissolved in water, some methyl orange is added and hydrochloric acid of known strength is then run into the solution from a burette until the alkali is just neutralized, this point being reached when the yellow color of the solution changes to pink. If t c.c. of hydrochloric acid were necessary, of which 1 c.c. contained exactly a gms. of HC1, it is evident that to neutralize the caustic soda contained in a gms. of the mixture a-t gms. of HC1 were used up, and it follows : HCl:NaOH=a-*:* s= TJ PI a -2 = gms. NaOH in a gms. of the mixture; in 100 gms. NaOH a: lOO-NaOH-a-* x= - TTTri - = per cent NaOH. We will first consider A. GRAVIMETRIC METHODS. These are divided into (a) Direct Analyses. (b) Indirect Analyses. In the case of a direct analysis the substance to be determined is separated from the solution in the form of an insoluble com- pound and weighed. GRAVIMETRIC METHODS. 3 The determination of barium in barium chloride was an example of a direct analysis. The indirect method depends upon the fact that when two or more substances are made to undergo the same chemical treatment they ex- perience a relatively different change of weight. For example, suppose that we have a mixture of the chlorides of sodium and of potassium and desire to determine the relative amounts of each of the two substances in the mixture. For this purpose a portion of the mixture (a gms.) is weighed, dissolved in water, the chlorine precipitated as silver chloride, and the weight of the latter determined (p gms.). From these data it is possible to calculate the amount of sodium chloride and of potas- sium chloride that was present in the mixture. If we let x represent the amount of the sodium chloride, y the amount of the potassium chloride, a the amount of silver chloride formed from x gms. of sodium chloride, and /? the amount of silver chloride formed from y gms. of potassium chloride, then NaCl KC1 x + y = a AgCl AgCl a + /? = p. We have, therefore, two equations with apparently four un- known quantities, but a and /? can be expressed in terms of x and y: NaCl:AgCl=z:o: KCl:AgCl=i/:/? AgCl Q AgCl ^ and -JL, , however, are known quantities ; they are simply the quotients of the molecular weights in question. If we designate by m the fraction and by n the frao- AgCl tion -T^T, we have x+ y =a mx-\-ny=p 4 INTRODUCTION. and from this we can calculate pn-a x=^ - and y=ax mn or 1 n x= -- p --- a. mn mn All indirect analyses may be calculated by means of this last general equation. In the above example _AgCl_ 143.34 _AgCl_143.34_ ~~~~ ~~"" and If these values are substituted in the above equations we obtain s=1.888-p-3.628-a. Consequently, in order to determine the amount of sodium chloride in the original mixture it is only necessary to determine the values a and p, then multiply them by the coefficients 3.628 and 1.888 respectively, and subtract the first product from the last. Although this method appears so simple and attractive on paper, impossible values are often obtained in practice, so that it is always necessary to be very cautious about using an indirect method. The experimental errors which are unavoidable in such an analysis are multiplied by the value of the coefficients; thus in the above case the actual error in the determination of the weight a is multiplied by 3.63 ... and the error in determining the weight of the silver chloride (p) is multiplied by 1.89 ... It is clear, therefore, that an indirect analysis becomes more accurate in proportion as the coefficients are small and when the error in determining a and p is slight. In the above example the coefficients are relatively small and consequently good results are to be expected, and experiment shows this to be the case. Example: A mixture weighing 0.5480 gm. (a) and consisting of GRAVIMETRIC METHODS. 5 0.4966 gm. sodium chloride (x) and 0.0514 gm. potassium chloride (y) yielded 1.3161 gm. of silver chloride (p), but from the values of a and p we can calculate those of x and y: x= 1.888-1.3161 -3.628-0.5480 = 0.4963 gm. sodium chloride; y = 0.0517 gm. potassium chloride. The calculated values, therefore, show 99.92 per cent, of the true value for the sodium chloride, 100.6 per cent, of the true value for the potassium chloride. Although the above results are satisfactory, it must be borne in mind that the analysis was carried out with chemically pure substances. If this were not so, as is usually the case in prac- tice, the results would be far less accurate. The same analysis may be performed in a much more simple manner than as above described, by weighing the mixture of the chlorides in a platinum crucible, then changing them to sulphates (by treatment with sulphuric acid and evaporating off the excess of the latter) and again weighing. In this case the actual experi- mental error is slight and excellent results might be expected. We have NaCl KC1 x + y = a Na 2 SO 4 KjSO, a + ft = p Na 2 S0 4 m-n- 0.0464 Now (1) x + y =a (2) mx+ny=p and (3) z= -- p v ' mn ^ 6 INTRODUCTION. Substituting the values for m and n in equation (3 we obtain z = 21.547-p-25.181a. In this case the coefficients are very large, so that the analytical error is multiplied enormously in the calculation, so much so that it is impossible to obtain even approximate values except when the mixture is composed of about equal parts of the two chlorides. Example: In a mixture containing about equal parts of the two salts there was found 99.64 per cent, of the sodium chloride present; 100.76 per cent, of the potassium chloride present. In a mixture containing considerable sodium chloride and littk potassium chloride there was found (a) 95.0 per cent, of the sodium chloride present; x 148.0 per cent, of the potassium chloride present. (6) 96.8 per cent, of the sodium chloride present; 129.9 per cent, of the potassium chloride present. The values obtained are, therefore, worthless. In the case of a direct analysis the small unavoidable errors exert a much less influence upon the result, so that a direct deter- mination should always be preferred. Only in those cases where a direct method is unknown should one resort to an indirect analysis! OPERATIONS. The principal operations of quantitative analysis are those of weighing, nitration, and the washing, drying, and ignition of precipitates. Weighing. The balance, as used for purposes of quantitative chemical analysis, is shown in Fig. 1. It consists of a horizontal lever with two arms of equal length, and in order to be serviceable it must be accurate and sensitive. It fulfils the first condition if (1) The arms of the lever are equally long; WEIGHING. 7 (2) The point of support (the fulcrum) lies above the centre of gravity; (3) The fulcrum (a knife-edge) and the knife-edges from which the pans are suspended lie in the same plane and are parallel to one another. The balance is more sensitive the greater the displacement of the position of equilibrium brought about by the addition of a small weight, e.g. one milligram. FIG. 1. The sensitiveness, or sensibility, may be expressed by the equation : tan a * = j- q.d in which p is the weight added, I the length of the balance-arm, q the weight of the beam, and d the distance between the centre of gravity and the point of support. The sensitiveness of the balance is greater, therefore, the heavier the weight added, the longer the beam, the lighter the beam, and the shorter the distance between the centre of gravity and the point of support. * a is the angle through which the pointer moves on the addition of the small weight. B INTRODUCTION. For convenience in determining the position of the balance, a pointer is fastened to the beam which, when the equilibrium is established, rests at the zero of a scale on an ivory plate below. The object to be weighed is placed upon the left scale-pan and the weights upon the right pan; the beam is lowered and the balance set in slight motion, by producing, with the hand, a gentle draft of air upon one of the pans. If the correct weigki has been added, the pointer will swing to the same number of scale divisions to the right of the zero that it does to the left, provided that it does so when there is nothing in either scale-pan, which is usually not the case. It is to be noted that when the zero-point of the balance (i.e., the point of the scale at which the pointer rests when the balance is in equilibrium with nothing in either scale-pan) coincides with the zero of the scale, it may change during the course of the day, so that disregard of this fact may lead to a considerable error. The cause of the displacement of the zero-point is that the first condition for the accuracy of a balance is not fulfilled. On account of unequal warming the arms become of unequal length. In order that accurate weighings may be obtained, it is necessary to make them independent of any inequality in the lengths of the arms, which can readily be done, as the following consideration will show. In the case of a lever, equilibrium takes place when the statical moments are equal. By statical moment is understood the product of the force into the length of the lever-arm, and the length of the lever-arm is the perpendicular distance from the axis of revolution (the fulcrum) to the line of action of the force. If an object, whose weight Q (Fig. 2) is to be ascertained, is placed upon the left balance-pan and equilibrium is established (the pointer rests at zero) by putting weights amounting to P gms. in the right balance-pan, then (1) Ql = Pl r If now the object Q is placed on the right-hand balance-pan and the balance again brought to the state of equilibrium by placing weights upon the left-hand balance-pan, in this case the weights WEIGHING. will not as a rule amount to P gms., but to P l gins. ever, equilibrium has been reached, we have (2) Ql^PJ,. Since, how- FIG. 2. If equation 1 is multiplied by equation 2, we obtain The true weight is obtained, therefore, by taking the geometric mean of the two values. For practical purposes, however, it is sufficiently accurate to take the arithmetical mean, in which case the true weight of the object would be: This method of obtaining the true weight independent of the lengths of the balance-arms is known as that of double weighing. The same end is obtained by Borda's method of substitution. According to this method the object to be weighed <'Q) is coun- terbalanced (tared) by means of shot, sand, weights, etc., the object Q is then removed and equilibrium with the tare is again established by placing weights upon the scale-pan. We havs, then, as a result of the first weighing, 10 INTRODUCTION. and from the second weighing, Pl = Tli from which it follows: Ql = Pl Q = P. The latter method is used 'chiefly in weighing large objects. For ordinary analytical work the weighing is made by the method of swings. First of all the zero-point of the balance is determined by setting the balance in motion (without any load in either pan), observing I /' FIG. 3. and recording the turning-points, or extreme positions, of the pointer on the scale of an uneven number of swings (say five*) and taking the mean of the readings. In order to give the same algebraic sign to all the observed readings it is best to number the divisions on the scale from left to right from to 20 so that the zero-point in case both balance-arms were of equal length would be numbered 10. The next thing to be determined is the sensitiveness of the balance for the object to be weighed. For this purpose the object is placed in the left-hand balance-pan, and by placing weights in the right-hand pan equilibrium is established as nearly as possible * The first two swings are inaccurate on account of the jar in shutting the balance-door, etc., so that they are disregarded. WEIGHING. 1 1 and the point of rest of the pointer on the scale is determined as above. An additional weight of 1 mgm. is added, or removed if the object was too light before, and the point of rest is again deter- mined. The difference (d) between this and the previous point of rest gives the sensitiveness of the balance. Assuming the zero-point to lie at 10.22, the first point of rest, obtained with a load of 19.723 gms., to be at 9.80, and the point of rest with a load of 1 mgm. less (i.e., with a load of 19.722 gms.) to lie at 12.32, then the sensitive- ness of the balance will amount to 12.32-9.80 = 2.52 scale divisions. As the zero-point of the balance was at 10.22 and the point of rest with a load of 19.723 gms. was 9.80, it follows that the object was lighter than the weights in the right-hand pan, and in fact the excess of weights in the pan was sufficient to move the point of rest 10.22 9.80-0.42 divisions on the scale. This amount can be calculated from the determination of the sensitive- ness of the balance as follows: Since 2.52 of the scale divisions correspond to 1 mgm., then 0.42 of the scale divisions correspond to the weight which must be sub- tracted from 19.723 gms. in order to obtain the true weight; there- fore 2.52: 1 = 0.42: x 42 = = 0.17 mgm. ; or about 0.2 mgm. The true weight of the body in air is consequently 19.723 - 0.0002 = 19.7228 * gms. In making a weighing one should always accustom himself to note the observations methodically, as follows: Assume that a platinum crucible is to be weighed. * As most analytical balances will scarcely detect with certainty less than J--Q mgm., the weight is expressed only to the fourth decimal. If the fifth decimal place in a calculation amounts to six or more, the number in the fourth decimal place is increased one. 12 INTRODUCTION. Zero-point. I. Point of Rest with Load of 12.052 gms. II. Point of Rect with Load of 12.053 gms. Left. Right. Left. . Right. Left. Right. 4.2 4.6 5.1 17.6 17.1 5.8 6.2 6.6 18.7 18.3 3.5 3.8 4.2 15.8 15.4 Sum=13.9 Mean =4. 63 34.7 17.35 4.63 18.6 6.2 37.0 18.5 -2 11.5 3.83 31.2 15.60 3.83 Sum of both means= 21 .98 Mean = 10.99 24.7 12.35 19.43 9.71 Sensitiveness=12.35-9.71 = 2.64 scale divisions. 12.35-10.99=1.36 scale divisions. 1.36 : 2.64= 0.5 mgm. Weight of crucible= 12.052 + 0.0005 = 12.0525 gms. The sensitiveness of a balance varies slightly with the load. It is simplest to determine once for all the sensitiveness for 50 gms., 20 gms., 10 gms., 5 gms., and 2 gms., place a card in the balance with the results obtained and use the numbers as required. In this way the sensitiveness of a balance with a load of 50 gms. 20 " 10 " 5 " 2 " was found to equal 2.23 scale divisions 2.28 " 2.64 " 2.66 " 2.66 " The determination of the zero- point, however, must be made with every weighing. If a number of weights are to be made one after another it suffices to determine the zero-point at the beginning and at the end and to use the mean of the two deter- minations. In case of very heavy loads, however, the zero- point should be determined before and after each weighing and the mean value used. REDUCTION OF WEIGHING TO VACUO. 13 Reduction of Weighing to Vacuo. Since most of our weighings are made in the air with brass weights, we are constantly introducing an error due to the dis- placement of air. This error is so small that it can be disregarded in ordinary analyses; in the case of the most accurate work, however, as in atomic weight determinations, calibrations of measuring vessels, etc., it should never be neglected. In such cases the apparent weight must be reduced to vacuo as follows: 1 c.c. of air at 15 C. and 760 mm. pressure weighs 0.0012 gm.= L The specific gravity of brass is 8.0 = s'.* The specific gravity of the substance weight = s. The body that weighs po gms. in vacuo will be balanced by p gms. in the air. The loss in weight of the substance is A gms. The loss in weight of the brass weights is -^ X gms. -1 fpo P \ S ** The total loss therefore, = ( A 4 A). The weight of the substance in vacuo is : ?(i--} , Po/ PA , r \s/ Po=p+ -7 and p = . -. CO 1 s On account of the small values of the fractions and -, s s" sufficient accuracy is obtained by simplifying the expression so that it becomes Po= Instead of making the computation, the following table of Kohlrausch may be used : * Brass has a density of 8.4, but the density of most analytical weights is nearer 8.0. INTRODUCTION. REDUCTION OF A WEIGHING MADE WITH BRASS WEIGHTS TO VACUO. METHOD OF P. KOHLRAUSCH. 5 k s k s k 0.7 + 1.56 2.0 + 0.45 8 -0.00 0.8 1.35 2.5 . 0.33 9 0.017 0.9 1.18 3.0 0.25 10 0.030 1.0 1.05 3.5 0.19 11 0.041 1.1 0.94 4.0 0.15 12 0.050 1.2 0.85 4.5 0.12 13 0.058 1.3 0.77 5.0 0.09 14 0.064 1.4 0.71 5.5 0.07 15 0.070 1.5 0.65 6.0 0.05 16 0.075 1.6 0.60 6.5 0.03 17 0.079 1.7 0.56 7.0 0.02 18 0.083 1.8 0.52 7.5 0.01 19 0.087 1.9 0.48 8.0 + 0.00 20 0.090 2.0 + 0.45 21 -0.093 & = L20( ) mgm. If a substance of specific gravity s \S o.U/ weighs g grams in the air, then g-k mgms. are to be added to the weight in air in order to obtain the weight in vacuo. TESTING Of- WEIGHTS. Testing of Weights. Although it is now possible to buy nearly perfect weights, yet their accuracy should always be tested. For almost all analytical purposes it matters not whether the 50 gm. weight weighs exactly 50 gms., but it is essential that the individual weights represent the corresponding relations to one another. In most sets of weights the following are found: 50 gm., 20 gm., 10 gm., 10' gm., 5 gm., 2 gm., 1 gm., V gm., 1" gm., 0.5 gm., 0.2 gm., 0.1 gm., 0.1' gm., 0.05 gm., 0.02 gm., 0.01 gm., 0.01' gm.; a rider (weighing either 10 or 12 mgms. according as to whether the bal- ance-arm is divided into 10 or 12 equal parts between the fulcrum and the point of suspension of the right-hand balance-pan) ; and usually smaller weights weighing 5, 2, 1, 1, 1 mgms. The weights may be tested in the following manner : * The assumption is first made that the sum of the larger weights is equal to 100 gms. = 100,000 mgms., and the weights of the single pieces obtained by the method of double weighing are compared with one another, e.g. : 50 gm. wt. against 20+ 10+ 10'+ 5+ 2+ 1+ 1'+ 1" and it is found: Left Right (1) 50gm. + 0.31mgm. =(20+10+10'+...) Left Right (2) (20+ 10+ 10'+ . . . ) =50 gm. + 0.61 mgm. from which it follows: , 0.31 mgm. +0.61 mgm. 50 gm.+ - 2 =50 gm.+0.46 mgm, = (20+10+10'+...) or 50 gm. = (20+ 10+ 10'+ . . . ) -0.46 mgm. * Kohlrausch, Leitfaden der prakt. Phys., V. Auflage, p. 34. See also T. W. Richards, Journ. Am. Chem. Soc. (1900) XXII, 144, where the method used at Harvard is described. 1 6 INTRODUCTION. The other weights are compared in the same way, obtaining, for example, 50 gm. = (20+ 10+ 10'+ . . .) gm.+ A mgm. 20 " = 10+10' +B " 10' " = 10 +C " * (5+2+1+...)= 10 +D " in which A, B, C, and D may be eithar positive or negative. The sum of the weights (50+20+10+10'+. . .) was assumed to equal 100 gms., and with the help of the preceding equations each weight is expressed in terms of the 10 gm. weight; then 10XlO + A+2 + 4C + 2Z>=(50 + 20 + 10 + . . .) = 100 gms. and the 10 gm. weight itself: A+2B + 4C + 2D 10=10 10 T , . - A+2B + 4C + 2D . .. . If we let S= , then we obtain 10-10 gm.-S 10'= 10 gm.-S + C 5 + 2 + l + l' + l"=10 gm.-S + D 20 = 20 gm.- 50 = 50 gm.- = 50 gm.+-U The sum %A + B+2C+D 5S should equal 0, which serves as a test for the accuracy of the observations. The 5 gm. weight is now compared with the 2+ 1+1'+ V in exactly the same way, with the result that 2 = 1+1' + b + c * It is well to mark the weights of the same denomination so that they may be distinguished from one another. TESTING OF W 'EIGHTS. According to the preceding work 5+2+ 1+ 1'+ 1" = 10,000-S+D consequently and 10Xl+a+2&+4c+2d=10,000-S+D a+2b+4c+2d+S-D 1 = 1000- 10 , , If we let s , we obtain l = 1000-s i'=iooo-s+c l" = 1000-s+d 2=2000-2s+6+c 5 = 5000-5s+a+6+2c+d In the same way the smaller weights are tested until finally the following correction table is obtained. TABLE FOR CORRECTION OF WEIGHTS. 50 = 50 g. + iA 20 =2Qg.-2S + B + C 10 =10g.-S 10'=10g.-S + C y 10 ). 5 + 0. 2 + 0. 1+0.1' 1 0.1 4 = 0.1 g. s' + ff 5 2 1 =10g. S + D I' 5 = 5g. 5s + a + b + + 2c + d 1' =1 g.-s + c 1"=1 g.-s + d 0.5 =1 g. 5s' + .8 + 0.2 =0.2g. 2s' + r + 0.1^=0.1 g.-s^ 4 = 0.1 g.'s' + ff 0.05 = 0.5g.-5s" + 0.02 = 0.02 g.-2s" +m + n 0.01=0.01 g.-s" 0.01' = 0.01g.-s" Rider = 0.01 g.-s" / + + + Sum = 100 g. 10g.-S + Z> 1 g.-s + a 0.1 g.-s' + ff * 4 = 0.05 + 0.02 + 0.01 +0.01' + Rider. (Rider = 0.01 g.) The milligram weights may be standardized in exactly the same manner. It is much more convenient, however, and the accuracy attained is almost exactly the same, if instead of using these very small weights the rider is hung upon the whole divisions of the balance-arm in order to obtain the weight in milligrams; for the estimation of the fractions of the milligram it is better to calculate them from the sensitiveness of the balance. 1 8 INTRODUCTION. The weights should never be touched with the fingers, but should always be lifted by means of the pincers provided with each set, and nothing should be placed on or removed from the balance-pans with- out arresting the balance, i.e., raising the mechanical supports so that the beam no longer rests upon its knife-edges. Filtration and Washing of Precipitates. How large should the filter be. and how many times should the precipitate be washed? With regard to the latter question it is evident that the pre- cipitate should be washed until the soluble matter is completely removed. It is clear, however, that this point will never be reached because a part of the solution always remains on the filter, but it is not difficult to make the amount of the dissolved substance remaining so small as to be negligible. When the amount of dissolved substance remaining on the filter is so small that it could not be detected by our balance, we consider the precipitate to be completely washed. The aim should be not only to remove all of the soluble matter, but to accomplish this with as little wash water as possible. No precipitate is absolutely insoluble, so that it is clear that every unnecessary excess of wash water causes harm by removing a fraction of the precipitate, and the greater the excess of the wash water the greater the amount of the precipitate dissolved. The amount of wash water to be used depends largely upon the nature of the precipitate itself. Amorphous, gelatinous pre- cipitates always require more washing than crystalline, granular ones.* As a rule, it may be said that the process of washing must be continued until the substance which is being washed out can be no longer detected in the last filtrate. In case the filtrate must be used for another determination, it is obvious that the filtrate should not be tested too soon. When should the filtrate be tested? Let us assume the filter to hold 10 c.c., the solution to drain * The reason why some precipitates require more washing than others is due to the fact that the degree of adsorption varies. (Cf. Ostwald, Die wissenschaftl. Grundl. der analyt. Chem., p. 19.) FILTRATION AND WASHING OP PRECIPITATES. 19 to the last drop from the paper, the amount of the solution held back by the precipitate and filter to be 1 c.c. and to contain 0.1 gm. of the solid substance which is to be removed by wash- ing. The filter is filled to the upper edge with wash water and allowed to drain to the last drop n times, until not more than 5/100 mgm. of the substance to be removed by washing remains. According to our assumption. 9 c.c. drain off and 1 c.c. remains behind; we have consequently: Removed by the There remains after the 1st washing, 0.1-9/10 gm. 1st washing 0.1 1/10 gm. 2d " 0.1- 9/10- 1/10 gm. 2d " 0.1- 1/10- 1/10 gm. 3d " 0.1-9/10- (l/10) 2 gm. 3d " 0.1- 1/10- (1/10) 2 gm. nth " 0.1- 9/10- (l/10)-lgm. nth " 0.1- 1/10- (l/10)n-l gm. After washing n times ; therefore, the amount removed by washing is the sum of the decreasing geometric series of which the first term is 0.1-9/10 and the constant factor is 1/10. In case n = 4, the sum of the series is After washing the precipitate four times, therefore, 0.09999 gm. of the impurity has been removed. According to the assumption that there was originally 0.1 gm. of this substance, there remains in the precipitate only 0.00001 gm., or in other words a negligible amount. Consequently, the filtrate should be tested qualitatively for the substance to be removed only after the precipitate has 'been washed four times. Often the washing will be found to have been complete after the fourth washing, but as a rule this will not be the case, and in many c - H 2 PtCl 6 required. The solution of the two chlorides in water (contained in a platinum or porcelain evaporating-dish) is treated with a few tenths more than the calculated number of cubic centimeters of H 2 PtCl6 and is then evaporated almost to dryness on the water-bath at as low a temperature as possible (the water should not boil). After cooling, the residue is treated with a few c.c. of absolute alcohol (best methyl alcohol *), after which the solid mass is broken up into a fine powder by means of a stirring-rod or a platinum spatula. The liquid is then decanted through a filter moistened with alcohol, and the treatment of the residue with alcohol together with the breaking up into powder, etc., is repeated until the alcohol runs through the filter completely colorless and the salt remaining assumes a pure, gold-yellow color without any orange-colored parti- cles being present (Na 2 PtCl 6 + 6H 2 O) . The precipitate is then carefully transferred to the filter, the alcohol is allowed to com- pletely drain off, and the precipitate is dried in the hot closet at 80-90 C. The greater part of the precipitate is then placed upon a clean watch-glass, the filter is replaced in the funnel, and the precipitate which still adheres to it (and likewise any pre- cipitate adhering to the dish in which the original precipitation took place) is dissolved off by means of a little hot water into a weighed platinum dish or crucible. The precipitate is evaporated * Dupre, Inaugural Dissertation, Halle, 1893. SODIUM: MODIFICATION OF THE CHLOROPLATINATE METHOD. 45 to dryness on the water-bath at as low a temperature as possible, and to it is now added the precipitate from the watch-glass. It is dried at 160 C. and weighed. The calculation of the amount of potassium chloride corresponding to the weight of the precipitate is performed as follows: The weight p of the potassium chloroplatinate is multiplied by 0.3056 and this gives at once the weight of the potassium chloride. Remark. The coefficient 0.3056 is used instead of the true factor (0.3068), because the potassium chloroplatinate precipitate does not exactly correspond to the formula K 2 PtCl 6 .* It con- tains, in fact, a little more chlorine, besides oxygen and hydrogen, which are not given off as water at a temperature of 160 C. We must assume that the chloroplatinic acid is decomposed slightly on evaporation, perhaps according to the following equation: H 2 PtCl 6 + H 2 O^HC1 + H 2 PtCl 5 OH. By this hydrolysis a mixture of the potassium salts (K 2 PtCl 6 + KHPtCl 5 OH) is obtained, but fortunately if the work is always done in the same way these compounds are always formed in the same relative amounts. Innumerable determinations have shown that correct results are obtained if the factor 0.3056 is used in the calculations. Modification of Chloroplatinate Method. Instead of weighing the K 2 PtCl 6 , the dry precipitate may be heated in a stream of hydrogen, when HC1 and H 2 O will be given off and a mixture of platinum and potassium chloride will remain behind. 1. If the amount of hydrochloric acid evolved is determined (p gm.) and from this the calculation of the potassium chloride made according to the following equation, K 2 PtCl 6 + 4H = 4HC1+ Pt+ 2KC1 4HCl:2KCl=p:a? KC1 X ~2HCl' P ' *Cf. Fresenius, Z. anal. Chem., 1882, p. 234. Also F. Dupre, "Die Bestimmimg der Kaliums als Kaliumplatinchlorid," Inaugural Dissert., Halle, 1893. Also W. Dittmar and McArthur, J. Soc. Chem., Ind. 6, 799, and Ber. f 1888, Ref. 412. 4 6 GRAVIMETRIC ANALYSIS. the result will be too low because less HC1 is evolved than corre- sponds to the above equation. 2. If the mixture of platinum and potassium chloride remaining in the dish is weighed (p gm.) and the amount of potassium chloride is calculated according to the equation (Pt+2KCl):2KCl=p:a? 2KC1 ~ too low a result will be obtained. 3. Finally, if the mixture of platinum and potassium chloride is treated with water and, on the one hand, the weight of the plati- num remaining undissolved and, on the other hand, the weight of the potassium chloride which goes into solution (by evaporating the solution and weighing the residue) is determined, then the amount of potassium chloride calculated from the weight p of the platinum Pt:2KCl=p:o? 2KC1 Trr* P again gives a result which is too low; while the amount of potas- sium chloride found in the aqueous solution corresponds to the amount of potassium chloride originally present. Inasmuch as the precipitate of potassium chloroplatinate possesses a constant composition it is possible to determine experi- mentally by working with pure materials the exact ratio which exists between (a) the amount of hydrochloric acid evolved, (b) the mixture of potassium chloride and of platinum remaining; after the ignition, (c) the weight of platinum remaining undissolved after treatment of the residue with water and the amount of potassium chloride originally present. According to Dupre, if the amount of platinum determined according to 3 is multiplied by the factor 0.76142 the true amount of potassium chloride will be obtained. MODIFICATION OF THE CHLOROP LATIN ATE METHOD. 47 As an example of the modified chloroplatinate method we have the Neubauer-Finkener Method. * This method does not require that the sodium and potassium shall be present as chlorides. It depends upon the precipitation of potassium chloroplatinate in the presence of ether-alcohol, igniting the precipitate (K 2 PtCl 6; Xa 2 S0 4 , etc.) in hydrogen, washing out the soluble salts, and weighing the residual plati- num. Procedure. The solution containing about 0.5 gm. of sub- stance is poured into a large porcelain casserole and treated with a few drops of hydrochloric acid. Somewhat more than enough chloroplatinic acid to precipitate the potassium is added, and the solution evaporated on the water-bath until its volume does not appear to diminish perceptibly; an unnecessarily long heating is to be avoided. After cooling the mass is moistened with about 1 c.c. of water and carefully crushed with the end of a flattened stirring-rod; then at least 30 c.c. of alcohol (93-96 per cent, by volume) are added in portions of 10 c.c., each time crushing the mass with the rod. If considerable sodium or potassium is present, the mass toward the end assumes a soft, cheesy consistency but eventually becomes hard and crystalline. The covered casserole is next allowed to stand for half an hour, rubbing the precipitate from time to time. Then the super- natant liquid is poured through a platinum Gooch crucible and the precipitate washed by decantation with alcohol. After each addition of alcohol the crystals are forcibly crushed with the rod. As soon as the filtrate passes colorless through the filter the precipitate is transferred to the crucible with alcohol, the alcohol is removed by washing six times with ether, and the latter by sucking air rapidly through the crucible. The crucible is covered, and through an opening in the cover hydro- gen (or illuminating-gas) t is passed and the crucible is heated, at first very gently, to avoid losses by decreptitation. After five * Z. anal. Chem., 1900, 485. t As in the determination of copper as cuprous sulphide, cf. p. 161. 48 GRAVIMETRIC ANALYSIS. minutes the flame of the burner is turned a little higher so that the bottom of the crucible just shows a faint redness in the center,* and this temperature is maintained for at least twenty minutes. It is then allowed to cool. The contents are next moistened with cold water, and hot water is sucked through the crucible fifteen times to remove the soluble salts com- pletely. To remove calcium sulphate, or other difficultly soluble salts, the crucible is filled with 5 per cent, nitric acid (not HC1), which is allowed to act for about half an hour, from time to time replacing with a little fresh acid. Then wash with hot water, dry and weigh the platinum from which the corresponding amount of KC1 is obtained by multiplying by 0.7612,f or of K^O by using the factor 0.4811. If hydrochloric acid w r ere used instead of nitric acid in the above treatment, the platinum would subsequently run through the filter in a colloidal condition. \ * The crucible should be placed upon a piece of platinum foil so that the flame does not come directly in contact with the perforation in the bottom of the Gooch crucible. f According to Neubauer the coefficient 0.7612 must be used in the presence of sulphates, and Dupre found that the factor 0.7614 holds in the case of chlorides. Similarly, Dittmar and McArthur (Z. anal. Chem., 28, 767) state that the factor 0.7611 applies in the presence of much magne- sium. J Kling and Engels, Z. anal. Chem., 45, 315 (1906). DETERMINATION OF SMALL AMOUNTS OF POTASSIUM. 49 Determination of Small Amounts of Potassium in the Pres- ence of Considerable Sodium. The solution may contain sodium, potassium, calcium, and mag- nesium in the form of their chlorides or sulphates, etc. Hydro- chloric acid gas is conducted into the solution, which has been concentrated as much as possible, until it has become sat- urated with the gas (the lower end of the delivery-tube should be enlarged, as indi- cated in Fig. 23 ? and should not dip into the liquid). To every 100 c.c. of the solu- tion 2 c.c. of water are now added, the precipitated sodium chloride is allowed to settle and the solution poured through a F IGL 23. funnel provided with a platinum filter- cone. The precipitated salt is washed three times by decantation with 95 per cent alcohol, transferred to the funnel, dried by suction and then washed three times more with alcohol. In the solution there remains all of the potassium, some sodium, and possibly calcium, magnesium, and sulphuric acid. The solution is evaporated to dryness on the water-bath if possi- ble (or if sulphuric acid is present the last traces of the free acid are removed by means of the free flame), the residue is weighed, for every decigram of the salt mixture 3 c.c. of double-normal hydrochloric acid are added with more than enough chloroplatinic acid to precipitate all of the potassium, and the liquid is evaporated to a paste. It is then treated with 20 c.c. of absolute alcohol, and well stirred. After standing five minutes 5 c.c. of ether are added, the mixture is allowed to stand half an hour under a bell-jar, and then filtered. As the residue often contains small amounts of other chloroplatinates, it should be purified as follows : The precipitate is allowed to dry in the air, it is dissolved in a little hot water, a few drops of chloroplatinic acid are added, and the above operation is repeated. The precipitate thus obtained contains all of the potassium in the presence of some sodium chloride and possibly sodium sulphate. It is washed with a mixture of ether and alcohol until the liquid runs through $3 GRAVIMETRIC ANALYSIS. the filter completely colorless, after which the precipitate is dried, moistened with hot water, and digested on the water-bath with a few drops of chemically pure mercury,* constantly stirring with a glass rod, until the liquid appears perfectly colorless. By means of this treatment the potassium chloroplatinate is completely decomposed with separation of platinum: K 2 PtCl 6 + 4Hg = 2KC1+ 2Hg 2 Cl 2 + Ft or K a PtCl 6 + 2Hg = 2KC1+ 2HgCl 2 + Pt. The mixture is thoroughly dried on the water-bath and gently ignited until the mercury is all volatilized ; the platinum is changed at the same time to a denser form, which can be readily washed by decantation. After cooling, the mass is treated with water, the aqueous solution is decanted through a filter, and the residual metal is washed with hot water, dried, and cautiously ignited. The filter is ignited in a platinum spiral, its ash is added to the main portion of the platinum in the crucible which is now ignited, and weighed. The weight obtained p multiplied by 0.7612, gives the amount of potassium chloride, or multiplied by 0.3994, gives the corresponding amount of potassium. The above method serves excellently for the estimation of small amounts of potassium in mineral waters. Separation of Potassium from Sodium by the Perchlorate Method. Schlossing-Wense.f Principle. This separation depends upon the insolubility of potassium perchlorate and the solubility of sodium perchlorate in '97 per cent alcohol. The procedure is as follows: The chlorides of the two metals (sulphates must not be present on account of the insolubility of sodium sulphate in alcohol) are dissolved, after weighing, in 20 c.c. of hot water, treated with 1 J * Sonnstadt, Z. f. anal. Chem., 38, 501. f Z. angew. Chem., 1891, 691, and 1892, 233. SEPARATION OF POTASSIUM FROM SODIUM. 51 times as much of the perchlorate solution and evaporated with stirring to a syrupy consistency. A little hot water is added and the liquid evaporated with constant stirring until all of the hydro- chloric acid is expelled and heavy fumes of perchloric acid are given off, when a little more water is poured over the residue and the solu- tion is again evaporated with stirring. The perchloric acid lost by volatilization is replaced from time to time. After cooling, the mass is treated with about 20 c.c.of 97 per cent, alcohol to which 0.2 per cent, by weight of perchloric acid has been added and the mix- ture is vigorously stirred.* It is important, however, not to break up the crystals of potassium perchlorate into too fine a powder as the latter would readily pass through the asbestos filter. After allow- ing the precipitate to settle the alcohol is decanted off through a Gooch crucible, the residue is washed again with the wash-alcohol, and the excess of the latter is removed by gentle warming The residue is dissolved in 10 c.c. of hot water, to which a little per- chloric acid has been added, and the solution is evaporated with stirring until fumes of perchloric acid are given off. One c.c. of the wash-alcohol is added ; and, in order to avoid the employment of an excess of the latter, the precipitate is transferred by meam of a rubber " policeman" to a Gooch crucible, washed with 50-75 c.c. of 97 per cent, alcohol, dried at 130 C., and weighed.f In case sulphuric acid was originally present, it is removed by precipitation with barium chloride. It is not necessary to remove phosphoric acid, but, in case this acid is present, it is advisa- ble to allow the potassium perchlorate to stand for some time with an excess of perchloric acid before treating it with alcohol. Preparation of Perchloric Acid According to Kr eider. { From 100-300 gms. of commercial sodium chlorate (NaClO 3 ) are placed in a round-bottomed flask and gradually heated until oxygen begins to be evolved slowly. This temperature is maintained until the * 100 c.c. of this alcohol dissolve about 5 mg. KC1O 4 whereas 100 c.c. 97.2 per cent, alcohol containing no free perchloric acid will dissolve 0.0158 g.- KC1O 4 (W. Wense). t Using this method, R. Fitzenkam obtained in three experiments, 100.11, 100.04, 100.24, mean 100.13 per cent, of the potassium chloride taken. % Z. anorg. Chem. IX, 342. 52 GRAVIMETRIC ANALYSIS. mass becomes solid (requiring H-2 hours), whereby the chlorate is almost completely changed to perchlorate and chloride. After cooling, the melt is dissolved in water, sufficient hydro- chloric acid is added to decompose any chlorate remaining, and the solution is evaporated to dryness (the liquid being constantly stirred from the time crystals begin to separate out.) The dry mass is broken up with a stirring-rod and then treated in a tall beaker with an excess of concentrated hydrochloric acid, by means of which sodium chloride separates out after a few min- utes. The solution (it now contains perchloric and hydrochloric acids in the presence of small, amounts of sodium chloride) is poured through a Gooch crucible and the residue is washed once or twice by decantation with concentrated hydrochloric acid. The nitrate is evaporated on the water-bath until the hydrochloric acid is completely expelled and heavy white fumes of perchloric acid are evolved. Inasmuch as commercial sodium chlorate is often impure, it is necessary to test the perchloric acid, which has been prepared, for potassium. For this purpose a small amount of the solution is evaporated on the water-bath to dryness and the residue is treated with 97 per cent, alcohol, which will dissolve it readily in the absence of potassium perchlorate. If potassium is found to be present, the melt obtained by heating the sodium chlorate as above de- scribed is treated with HC1 and evaporated to dryness in order to decompose any sodium chlorate remaining. The residue is finely powdered and treated with 97 per cent, alcohol (1 c.c. dissolves 0.2 gm. NaClOa) and filtered, and the process repeated until a little of the alcoholic solution when evaporated to dryness leaves abso- lutely no residue. The alcoholic solution, which is now free from potassium, is distilled from a spacious flask until the perchlorate begins to crys- tallize out, when it is poured rapidly into an evaporating-dish, evaporated to dryness, and treated as previously described, with hydrochloric acid, etc. One c.c. of a potassium perchlorate solution prepared accord- ing to the above directions gave a residue of 0.0369 gm. which was completely soluble in 97 per cent, alcohol, as it should be. In order to ascertain the approximate amount of perchloric DETERMINATION OF LITHIUM, POTASSIUM, AND SODIUM- 53 acid contained in the solution, I c.c. should be treated with an excess of KC1, evaporated to dryness, treated with an excess of 97 per cent, alcohol, filtered through a Gooch crucible, and washed until the nitrate shows no turbidity on being treated with silver nitrate solution. The precipitate is then dried and weighed. LITHIUM, Li. At. Wt. 6.94- Forms: Li 2 S0 4 and LiCl. The determination of lithium in the form of the above salts is carried out in practically the same way as in the case of potassium. It should be mentioned, however, that on evaporating a lithium salt with concentrated sulphuric acid the acid salt, LiHSO 4 , is formed, which on gentle ignition (even without the addition of ammonium carbonate) is changed to difficultly volatile Li 2 S0 4 . Since lithium chloride is a very hygroscopic salt, it is necessary to weigh it out of contact with moist air. To accomplish this, the platinum crucible, after being gently ignited, is placed in a desiccator which is provided with a calcium- chloride tube, and beside the crucible is placed a weighing beaker with ground glass stopper. After both crucible and beaker have assumed the tem- perature of the room, the former is quickly placed within the latter which is then stoppered. It is allowed to stand for 20 min- utes in the balance case and then weighed. The salt is then placed in the crucible and the above process repeated. Determination of Lithium, Potassium, and Sodium in the Presence of One Another, After determining the weight of the combined chlorides, the potassium is determined in one portion as K 2 PtCl 6 , and in a second portion the lithium is determined according to one of the following methods : (a) Gooch's Method* Principle. Anhydrous LiCl is soluble in anhydrous amy' alcohol (15 parts of amyl alcohol dissolve in the cold 1 part 01 LiCl, or 10 c.c. dissolve 0.66 gm. LiCl) while KC1 and NaCl are * Proceedings of the Am. Acad. of Arts and Sciences. 22 [N. S. 14], 177. 54 GRAVIMETRIC ANALYSIS. difficultly soluble in this liquid (solubility of NaCl = 1 : 30,000 of KC1= 1:24,000). Procedure. The solution, after having been concentrated as far as possible, and which should not contain more than 0.2 gm. LiCl, is placed in a 50 c.c. Erlenmeyer flask, 5-6 c.c. of amyl alcohol (boiling point 132 C.) are added and the flask is placed upon an asbestos plate and cautiously heated. The aqueous solution at the bottom of the beaker soon begins to boil and the water vapor escapes through the upper layer *of amyl alcohol.* As soon as all the water has been boiled off. the chlorides of sodium and potassium separate out, while the greater part of the lithium chloride is to be found in the alcoholic solution. During the evaporation of the aqueous LiCl solution, however, some LiOH is formed by hydrolysis, and the latter compound is insoluble in amyl alcohol. In order to bring this completely into solution, the clear amyl alcohol solution is treated with 2-3 drops of concen- trated hydrochloric acid, boiled two or three minutes and filtered while still warm through a small asbestos filter. The crust which remains is composed of sodium and potassium chlorides and is washed with hot amyl alcohol, which has been boiled. The fil- trate is evaporated to dryness, and the residue is dissolved in a little water after the addition of some dilute sulphuric acid The solution is filtered from the carbonaceous residue into a weighed platinum crucible, evaporated as far as possible on the water- bath, the excess of sulphuric acid is removed by gentle heating over a flame (the crucible being held in an inclined position) and it is then weighed. The lithium sulphate thus obtained always contains small amounts of potassium and sodium sulphates ir case these metals were present, so that from the weight obtained, 0.00041 gm. should be deducted for every 10 c.c. of the filtrate {exclusive of the alcohol used in washing the residue") in case only sodium chloride is present, or 0.00051 if only potassium chloride is present, and 0,00092 if both sodium and potassium chlorides are present * To prevent loss by bumping at this point, the flask should be fitted with a cork stopper through which two tubes pass. If air is drawn through the liquid during the boiling, the water evaporates more quickly and without bumping. DETERMINATION OF LITHIUM, POTASSIUM, AND SODIUM. 55 If 10-20 mgm. of lithium chloride were present in the original salt mixture, then the residue obtained after filtering and wash- ing with amyl alcohol is dissolved in a little water and the above treatment is repeated, the lithium being determined in the com- bined filtrates. This method is very accurate, and, in the author's opinion it is to be preferred to all other methods for the determination of lithium. (6) Rammelsberg's Method. Principle. Anhydrous lithium chloride is soluble in a mixture of equal parts alcohol and ether which has been saturated with hydrochloric acid gas, whereas the chlorides of sodium and potas- sium are practically insoluble therein. Procedure. The solution of the chlorides is evaporated to dry- ness in a small flask made of Jena glass and provided with a ground-glass, two-way stopper (p. 34, Fig. 21a) . During the evapo- ration a current of dry air is passed into the flask through the long tube a and out through the short tube b. As soon as the resi- due has become dry the flask is placed in an oil-bath and heated for half an hour at 140-150 C. , during which time dry hydro- chloric acid gas is passed through the flask. The flask and its con- tents are allowed to cool with the hydrochloric acid still passing through the flask, after which the residue is treated with a few cubic centimeters of absolute alcohol, which has been saturated with hydrochloric acid gas and thereupon diluted with an equal volume of absolute ether. The flask is tightly stoppered and allowed to stand with frequent shaking for 12 hours. The solution is then poured through a filter, wet with the ether-alcohol mixture, and the residue is washed three times by decantation with ether-alcohol. A few more cubic centimeters of ether-alcohol are added to the contents of the flask and it is again allowed to stand for 12 hours ; the liquid is then poured off and the residue is washed with ether-alcohol until a trace of the residue tested in the spectroscope shows the complete absence of lithium. The ether-alcohol extract is care- fully evaporated to dryness in a water-bath containing lukewarm water, the residue is dissolved (after the addition of a little dilute sulphuric acid) in as little water as possible, transferred to a weighed 56 GRAVIMETRIC ANALYSIS. platinum crucible and treated with, sufficient sulphuric acid to transform the lithium chloride present completely into sulphate.* The solution is evaporated as far as possible on the water-bath, then cautiously over the free flame, after which it is gently ignited and the residue of lithium sulphate is weighed. Remark. In the presence of considerable sodium and potas- sium salts it is advisable to remove the greater part of these by precipitation with hydrochloric acid gas (cf. p. 49), filtering through asbestos and washing thfc precipitate with concentrated hydrochloric acid until the residue no longer gives the lithium spectrum. The results obtained by this method are satisfactory. Besides the above methods for the separation of lithium from sodium and potassium there are two other methods to be men- tioned; that of W. Mayer f and that of A. Carnot.J According to Mayer the lithium is precipitated in the presence of NaOH as Li 3 P0 4 , which, after being washed with ammonia water, is ignited and weighed. Rammelsberg, however, claims that the Li s PO 4 always contains some sodium, so that the method is inaccurate. A great many experiments tried in the author's laboratory have led to the same conclusion. According to Carnot the lithium is separated as the fluoride and then transformed to the sulphate. Walter claims that this method is accurate but tedious. Example for practice: Lepidolite analysis. (See Index.) Indirect Determination of Lithium and Sodium. The mixture of the two chlorides is weighed and the chlorine determined either gravimetrically or volume trie ally. (See p. 3.) Indirect Determination of Lithium and Potassium. The method is the same. * The above-described method has been modified by the author. Ram- melsberg evaporates the chlorides in the water-bath, heats the residue till it melts and then after cooling extracts with ether-alcohol. By the evapora- tion and fusion of the lithium chloride there is formed some lithium hy- droxide which is changed by the carbonic acid of the air to carbonate. Lithium carbonate is insoluble in ether-alcohol so that the extraction with ether-alcohol is not complete. t Ann. Chem. Pharm., 98, 193, and Merling, Z. anal. Chem., 18, 563. I Z. anal. Chem., 29, 332. The Analyst, 16, 209. AMMONIUM. 57 AMMONIUM NH 4 . Mol. Wt. 18.04. Forms: NH 3 , NH,C1, (NH^PtCl,, Pt, N. We have two cases to distinguish : 1. The ammonium is present as chloride in aqueous solution. 2. The ammonium is present in solution, together with other cations and anions. 1. The solution contains only NH 4 and Cl ions. In this case the solution may be evaporated to dryness and the residue of ammo- nium chloride weighed; or the ammonium can be precipitated as (NH 4 ) 2 PtCl 6 and the precipitate weighed; or the ammonium chloroplatinate can be ignited and the residue of platinum weighed. (a) Determination as NH 4 C1. The aqueous solution is treated with concentrated HC1 and evaporated to a small volume on the water-bath at as low a temper- ature as possible, the solution is transferred to a platinum crucible (or one of porcelain), evaporated on the water-bath to dryness, and the covered crucible is dried to constant weight in a drying-oven. Good results are obtained, but they are always too low. On evaporating the aqueous solution some NH 4 C1 is driven off, and the amount lost increases in proportion to the amount of water used and the temperature at which the evaporation takes place, on account of the NH 4 C1 being partly decomposed according to the equation NH 4 Cl^NH 3 -hHCl into NH 3 and HC1, both of which are volatile.* If, on the other hand, a little hydrochloric acid is added to the solution the dissociation is for the most part prevented eo that the loss is reduced to a minimum. The ammonium chloride must be dried in a covered crucible as otherwise a small amount of the * In cold, aqueous solution the NH 4 C1 undergoes electrolytic dissociation simply : NH 4 C1<=NH* 4 + C1'. , 5 GRAVIMETRIC ANALYSIS. salt will be lost, but this amount is small in comparison with the amount v/hich it is possible to lose during the evaporation. (6) Determination as (NH 4 ) 2 PtCl 6 . On heating (NH 4 ) 2 PtCl 6 to 130 C. the salt is unchanged; it suffers no dissociation, therefore, up to 130 C. In aqueous solu- tion it undergoes only electrolytic dissociation so that the above salt loses no (NH 4 ) during the evaporation of its aqueous solution. The aqueous solution of ammonium chloride, therefore, is treated with an excess of chloroplatinic acid and a little hydro- chloric acid and evaporated at as low a temperature as possible to dry ness. The residue is taken up in absolute alcohol and filtered through a Gooch crucible, dried at 130 C., and weighed. From this weight, the amount of ammonium chloride originally present can be correctly calculated by using the old atomic weight of platinum: Pt= 196.9. If the new value for the atomic weight of platinum (Pt= 195.2) is used, too high a value will be obtained for the amount of ammonium present, as was explained in the case of potassium. . If the weight of the (NH 4 ) 2 PtCl 6 =p, then pX 0.2400 = NH 4 C1 PX 0.08095 = NH 4 pX 0.07643 -NH 3 . (c) Determination as Platinum. Instead of weighing the (NH 4 ) 2 PtCl fi as such, it can be decom- posed by ignition* and the weight of the platinum remaining determined. If the old value for the atomic weight of platinum (196.9) is used in this determination the results obtained will be * As ammonium chloroplatinate decrepitates strongly on being heated, the ignition must take place in a large porcelain crucible which is provided with a close-fitting cover. The precipitate must be heated gradually at first to prevent loss. It is best ignited according to the directions of Rose. The precipitate and filter are placed in the crucible with the filter-paper on top, the crucible is covered and heated over a very small flame until the paper is completely charred without allowing the vapor to escape visibly from the crucible. The crucible is then strongly ignited with free access of air until the charred filter is completely consumed. AMMONIUM PRESENT WITH OTHER CATIONS. 59 about 0.4 per cent too low, while if the new value (195.2)* is used, the results will be about 0.8 per cent too high. Correct results can be obtained by multiplying the weight of platinum (p) by the following factors: pX 0.54527 = NH 4 C1; pX0.18391=NH 4 ; pX0.17364 = 2. The Ammonium is Present Together with Other Cations and Anions in Solution or in Solid Form. (a) The solution is distilled after the addition of a strong base (NaOH Ca(OH) 2 |), the ammonia evolved is absorbed in hydro- chloric acid and the resulting solution is analyzed according to 1. H FIG. 24. Procedure. About 1 gm. of the substance to be analyzed is placed in the 400-500 c.c. Erlenmeyer flask K, it is dissolved in 200 c.c. of water, a few drops of litmus solution are added, and in case the * In analyzing chloroplatinates of organic bases (by weighing the plat- inum) correct results may be obtained by using for the atomic weight of platinum the value 196.9. t MgO is frequently recommended for expelling the ammonia. Accord- ing to a private communication from Herrn Bormann, of Neunkirch, this base is absolutely unsuited for this purpose. 60 GRAVIMETRIC ANALYSIS. solution reacts acid, sodium hydroxide solution (which has been previously boiled to expel traces of ammonia) is slowly added at T with constant shaking until the solution changes to blue, after which ten c.c. more of the caustic soda solution are added.* The liquid is then heated to boiling and 100 c.c. of it is carefully distilled into the receiver V, which already contains 20 c c. of 2 N. hydro- chloric acid. In order to make sure that no ammonia escapes from the receiver it is well to connect it with a small Peligot tube containing 5 c.c. of 2 N. hydrochloric acid and some distilled water. After 100 c.c. of the liquid have distilled over, all the ammo* nia will be found in the receiver and can be determined according to 1 (a) or 1 (6); preferably the latter. The determination can be carried out much more quickly, however, if the receiver contains a measured amount of standardized acid and the excess is deter- mined after the distillation by titrating with alkali (cf . Alkalimetry) . It is also possible to make an accurate determination of the amount of ammonia present by measuring the volume of the gas.f Colorimetric Determination of Ammonium. For the determination of such small amounts of ammonia as occur in drinking-water, the above methods are not suited. In this case the procedure is the same as was described in Vol. I, p. 46. (In the case of mineral waters it is necessary to add more than one drop of the soda solution; the amount necessary is determined by adding litmus to a definite volume of the water and then adding the soda solution until the litmus changes to blue.) The distillate is received in 50 c.c. graduated Nessler tubes (in the fourth one there is usually no ammonia to be detected) and these are Nesslerized. The 50 c.c. of distillate is mixed with 2 c.c. of the Nessler solution and the yellow color produced is compared with the colors produced in the same way from a series of tubes contain- ing known amounts of ammonia. When a standard is found of * The separatory funnel T should be roughly calibrated before setting up the apparatus, by pouring water into it, one cubic centimeter at a time, and marking with a pencil the level of the liquid on the glass. t Cf. Part III, Gas Analysis. COLORIMETR1C DETERMINATION OF AMMONIUM. 61 the same shade as the solution tested, then the two solutions contain the same amount of ammonia. The ammonium chloride solution necessary for preparing the standards is prepared as follows : 3.141 * gms. of ammonium chloride which has been dried at 100 C. are dissolved in 1 liter of water free from ammonia (cf. foot- note, Vol. I, p. 47). The solution now contains 1 mgm. of ammonia (NH 3 ) per cubic centimeter; this, however, is too strong for most purposes, so that 10 c.c. cf it are taken and diluted to 1 liter. Of this solution 1 c.c. contains 0.01 mgm. NH 3 . If the water to be analyzed contains considerable ammonia, a smaller portion should be taken for the analysis than in ordinary cases (500 c.c.) as otherwise the first distillate (50 c.c.) would give too intense a color with the Nessler solution. In such a case only 50 c.c. of the water should be taken for the analysis and this should be diluted to 500 c.c. with water free from ammonia and then distilled. In order to ascertain how much of the water to take for the analysis, the following experiment should be made: About 100 c.c. of the water to be tested are placed in a narrow cylinder (which is provided with a ground-glass stopper), 2 c.c. of a strongly alkaline sodium carbonate solution f are added to pre- cipitate the calcium which may be present, the mixture is vio- lently shaken and allowed to settle. From the clear supernatant liquid 50 c.c. are pipetted off into a Nessler tube, treated with 2 c.c. of Nessler solution and mixed.! If a strong yellow color, or FIG. 25. f 50 gms. NaOH and 50 gins. Na 2 CO 3 (calcined) are dissolved in 600 c.c. of pure distilled water and the solution boiled until the volume is only 500 c.c. % In the case of mineral waters rich in magnesium sulphate, the addition of he 10 c.c. of sodium carbonate solution often fails to prevent a turbidity on jidding the Nessler reagent, which would render a colorimetric determina- tion impossible. In this case 10 c.c. of a boiled BaCl 2 solution (120 gms. BaCl 2 + 2H,O in 500 c.c. H 2 O) should be added before treating the water with the sodium carbonate solution. 62 GRAVIMETRIC ANALYSIS. even a precipitate, is obtained, then only 50 c.c. of the water should be taken for analysis. If, on the other hand, there is not more than a faint coloration apparent, then 500 c.c. must be taken for the determination. For the Nesslerization, the three cylinders each containing 50 c.c. of the distillate are placed over a sheet of white paper, treated with 2 c.c. of the Nessler reagent, and mixed. Beside them are placed a series of similar cylinders containing respectively 0.0, 0.5, 1.0, 1.5, 2.0, 2.5, 3.0 c.c! of the standard ammonium chloride solution diluted to 50 c.c. These are also treated with 2 c.c. of the Nessler reagent and by matching the colors obtained in the test with those obtained from known amounts of ammonia the amount present in the water can be easily estimated. The Nessler reagent should give a distinct coloration with 500 c.c. of water containing 0.005 mgm. NH 3 ; if this is not the case, it must be made more sensitive by the addition of mercuric chloride solu- tion. For mixing the liquid in the cylinders it is convenient to employ a stirrer such as is shown in Fig. 25, the diameter of the bulb on the end being only slightly less than that of the cylinder. By moving this stirrer up and down twice the liquid becomes thor- oughly mixed. Kjeldahl's Method for Determining Nitrogen. The methods which have been described thus far are suitable only for the determination of nitrogen when it is in solution in the form ot NH 4 ions. It is, however, of great importance to be able to determine nitrogen when it is present other than as an ammonium compound (in protein, coal, etc.). Inasmuch as it is possible to determine the amount of ammonia present very accu- rately, and by the employment of volumetric methods, very quickly^ methods were sought for the transformation into ammonia of the nitrogen originally present in some other form. This is readily brought about by the method of Kjeldahl and its modifications. By the oxidation of nitrogenous organic substances with coii- centrated sulphuric acid, potassium permanganate, mercuric oxide, etc., the organic matter is destroyed and the nitrogen is completely changed to ammonium and held as ammonium sulphate, from which the ammonia oan be readily distilled off. KJ ELD AH US METHOD FOR DETERMINING NITROGEN 63 Procedure for KjeldahVs Nitrogen Determination (Wilfarth's Modification*). From 1 to 2 gms. of the substance to be analysed are placed in a 500-600-c.c. flask, made of difficultly fusible potash glass, and to it are added 20 c.c. of sulphuric acid (3 vo-umes of con- centrated acid mixed with 2 volumes of fuming sulphuric acid) and a few drops of mercury, f The flask is then heated in an iron dioh covered with asbestos until its contents gently boil. It is impor- tant, however, to be sure that the substance should be thoroughly moistened by the sulphuric acid before the heating, especially in the case of mealy substances. In order to avoid a loss of nitrogenous matter, it is first heated for half a minute over a very small flame and then over a larger one, but in no case should the flame touch the flask above the part occupied by the liquid. The heating is continued until the solution becomes clear and completely colorless. In the presence of iron compounds, how- ever, the liquid is sometimes slightly yellow. The decomposition is usually complete in two or three hours. The liquid is then allowed to cool, the sides of the flask are washed down and the solution is diluted with about 250 c.c. of water. After it is thoroughly cool, 80 c.c. of caustic soda solution, free from nitrate, are quickly added and sufficient potassium sulphide solution (40 gms. commercial potassium sulphide to the liter) to completely precipitate the mer- cury and cause the liquid to appear black (25 c.c. of the potassium sulphide solution are usually sufficient) . A few grains of powdered zinc are then added and the flask is quickly connected with the distil- ling apparatus. The distilling-tube dips into a 250-300-c.c. Erlen- meyer flask containing a known volume of normal sulphuric acid (10-20 c.c.) and sufficient water to cover the lower end of the con- denser tube. As soon as a noticeable amount of water vapor begins to come over, it is no longer necessary to have the condenser tube dip into the sulphuric acid solution in the receiver. After 100 c.c. of the liquid have distilled over, the receiver is removed and the excess of sulphuric acid is determined by titration with * Chem. Ztg., 9, 502. t Gladding uses a mixture of H 2 SO 4 and KHSO 4 , which usually works very well. Then, as no mercury is added, the subsequent addition of is unnecessary. 64 GRAVIMETRIC ANALYSIS one-tenth normal sodium hydroxide solution, using methyl orange as an indicator. * From the amount of sulphuric acid used, the amount of nitrogen present may be calculated as follows: Let t be the number of cubic centimeters of normal sulphuric acid neutralized by the ammonia evolved from a gms. of the substance, then this corresponds to tX 0.01401 gms. nitrogen, and the percentage of nitrogen in the^ubstance is a : X 0.01401= 100 :z, 1.401X* x = ------- = per cent, nitrogen. If the nitrogen is originally present to a considerable extent in the form of nitrates, oxides, or cyanides, the above modification of Kjeldahl's method will not serve to change all of the .nitrogen into ammonia. In such cases it is best to use the modification proposed by M. Jodlbauenf From 0.2-0.5 gm. of potassium nitrate (or the corresponding amount of another nitrate) are treated with 20 c.c. of concentrated sulphuric acid and 2.5 c.c. of phenolsulphonic acid (50 gms. of phenol dissolved in enough concentrated sulphuric acid of 66 Be. to make 100 c.c. of solution) 2-3 gms. of zinc dust and 5 drops of chloroplatinic acid are added and the mixture heated. After heating the substance with this mixture for four hours, the liquid becomes colorless and is ready to be distilled with the caustic soda solution. * In the titration it is best to add the alkali until the solution turns yellow, and then finish by adding enough acid to get a change to a pale pink. In standardizing the acid and alkali, the volume of the solution and the method of titrating must be the same as during the analysis proper. t Z. anal. Chem., 27, 92. DETERMINATION OF MAGNESIUM. 65 MAGNESIUM, Mg. At. Wt. 24.32. Forms: MgS0 4 , MgO, Mg 2 P 2 O 7 . (a) Determination as MgS0 4 . This method for the determination of magnesium can always be employed when the magnesium is combined with an acid which can be volatilized by heating with sulphuric acid, and when no other metal besides ammonium is present. A weighed amount of the substance is placed in a platinum crucible and treated with a slight excess of concentrated sulphuric acid, * the mixture is evaporated on the water-bath as far as possible, and the excess of free sulphuric acid is removed by cautiously heating the crucible, held in an inclined position, over a free flame. Finally the dry mass is heated just to redness in a covered crucible, and after cool- ing in a desiccator, is weighed as quickly as possible, as the anhydrous magnesium sulphate is hygroscopic. (6) Determination as MgO. This method is seldom used in practice, and then only in case the magnesium is present in a form that can be readily changed to the oxide by ignition i.e., as carbonate, nitrate or salt of an organic acid.f The procedure consists simply of at first carefully heating in a covered crucible, and finally with the full heat of the Teclu burner in a half -covered crucible. (c) Determination as Magnesium Pyrophosphate. This, the most important of a]l the methods for the determina- tion of magnesium, is always applicable and depends upon the following principles : If the solution of a magnesium salt is treated with an alkali orthophosphate solution in the presence of ammonium chloride and ammonia, the magnesium is completely precipitated * Substances which react violently with concentrated H 2 SO 4 should be first treated with water, and dilute sulphuric acid added little by little. f Magnesium chloride can be changed to the oxide by ignition with mer- curic oxide in a porcelain evaporating-dish. Mercuric chloride and the excess of mercuric oxide are volatilized. In this way magnesium is often separated from the alkalies. (Translator.) 66 GRAVIMETRIC ANALYSIS. as magnesium ammonium phosphate, which by ignition is changed to magnesium pyrophosphate : 2MgNH 4 PO 4 = 2NH 3 + H,O+ Mg 2 P 2 O 7 . Formerly it was a common practice to precipitate magne- sium ammonium phosphate in the cold. Neubauer* showed, however, that this sometimes leads to high results while at other times the results are low. The latter is the case when the precipitation takes place in strongly ammoniacal solutions con- taining but little ammonium salts, particularly when the phos- phate solution is added slowly. Tribasic magnesium phosphate, Mg3(PC>4)2, contaminates the precipitate. On the other hand, the results are too high if the precipitation takes place in neutral or slightly ammoniacal solution in the presence of considerable ammonium salts. In this case more or less monomagnesium ammonium phosphate, Mg(NH 4 ) 4^04)2, is formed. This com- pound is changed to magnesium metaphosphate on gentle ignition. and the results are too high. When only a little of the meta- phosphate is present, the temperature of the blast-lamp will eventually lead to volatilization of some phosphorus pentoxide, so that nearly correct results are then obtained. 2Mg(P0 3 ) 2 - Mg 2 P 2 7 + P 2 5 . Neubauer recommends adding an excess of sodium phosphate to the slightly acid solution of the magnesium salt, then stirring in one-third of the solution volume of 10 per cent, ammonia, filtering after four or five hours, washing with 2.5 per cent. ammonia, dissolving the precipitate in a little dilute hydro- chloric acid, adding a few drops of sodium phosphate solution, and precipitating by the addition of one-third volume of 10 per cent, ammonia. This method, however, gives too high results, * Z. angew. Chem., 1896, 439. See also Gooch and Austin, Z. anorg. Chem., 20, 121. DETERMINATION OF MAGNESIUM. 67 e.g., 9.97, 9.95, and 9.98 per cent. Mg in MgS0 4 , 7H 2 instead of 9.88 per cent. Mg. Correct results can be obtained by the method of B. Schmitz,* or of W. Gibbs.f Method of B. Schmitz. The acid solution, containing magnesium in the presence of ammonium salts, % is heated to boiling and then treated with an ex- cess of sodium or ammonium phosphate. One-third the solution's volume of 10 per cent, ammonia is at once added, the solution allowed to cool and filtered through a Munroe crucible after standing for several hours. The precipitate is washed with 2.5 per cent, ammonia, dried and ignited very slowly, gradually increasing the heat until the precipitate is white. After cool- ing, the Mg 2 P2O 7 is weighed: 2MgNH 4 PO 4 = 2NH 3 + H 2 O + Mg 2 P 2 7 . From the weight of the latter, p, the amount of magnesium can be calculated as follows: g -- . p= weight of magnesium. The precipitate can be filtered upon an ordinary filter, but in all cases the ignition must be gradual or it is almost impossible to obtain perfectly white pyrophosphate. * Z. anal. Chem. 1906, 512; cf. Jorgensen, ibid. 1906, 278. t Am. J. Sci. (3), 5, 114. J Ammonium salts do no harm when the precipitation takes place in hot solution, but, on the contrary, cause the formation of a coarsely crystalline precipitate which is easy to filter. 68 GRAVIMETRIC ANALYSIS. Method of W Gibbs. The neutral, not too concentrated, solution of magnesium salt containing ammonium salts is treated at the boiling tem- perature with a normal solution of microcosmic salt, NaHNH 4 PO4 fH 2 O, until no further precipitation takes place. Almost 90 per cent, of the magnesium present is at once thrown down as amorphous magnesium hydrogei* phosphate, MgHP0 4 : NaHNH 4 PO 4 + MgCl 2 = NaCl +. NH 4 C1 + MgHPO 4 . Then, while stirring the hot solution, about one-third volume of 10 per cent, ammonia is added whereby the amorphous pre- cipitate is transformed into crystalline magnesium ammonium phosphate: MgHP0 4 -h NH 3 = MgNH 4 P0 4 . At the same time the magnesium remaining in solution is thrown down almost completely. After standing two or three hours, the supernatant liquid is filtered off, and the precipitate washed three times by decan- tation with 2.5 per cent, ammonia, finally transferred to the filter, washed completely with 2.5 per cent, ammonia and dried in the hot closet. The dried precipitate is transferred as com- pletely as possible to a weighed platinum crucible, the filter- paper burned in a platinum wire spiral, and the ash added to the main portion of the precipitate. The covered crucible is heated very gently at first until the ammonia is all dried off, then more strongly until the mass is snow white. The crucible is cooled in a desiccator and weighed. Separation of Magnesium from the Alkalies. The methods of Gibbs and of Schmitz serve to separate mag- nesium from the alkalies in those cases where the determination of magnesium alone is desired. If it is desired to determine magnesium and the alkalies in one DETERMINATION OF MAGNESIUM. 69 and the same sample, it is best to use the Gooch and Eddy * modifications of the Schaffgotsche Methodf The method is based upon the fact that magnesium can be precipitated quantitatively, by means of an alcoholic solution of ammonium carbonate, as crystalline magnesium ammonium car- bonate, Mg00 3 -(NH 4 ) 2 C(V6H 2 0. Preparation of the Precipitant. A mixture of 180 c.c. con- centrated ammonia, 800 c.c. water, and 900 c.c. absolute alcohol is saturated with commercial ammonium carbonate. The liquid is shaken with the powdered carbonate, and after several hours the excess of the latter is removed by filtration. Procedure. The neutral solution containing only magnesium and the alkalies (lithium must not be present), preferably in the form of chlorides, is treated with an equal volume of absolute alcohol and then with an excess of the ammonium carbonate re- agent. The liquid is vigorously stirred for a few minutes and allowed to stand for at least half an hour, whereupon it is filtered through a Gooch or Munroe crucible. The precipitate is washed with the precipitant, dried, ignited and weighed as MgO. If considerable alkali is present the precipitate always con- tains a small quantity of it. In such cases the precipitate is dis- solved in hydrochloric acid, the solution evaporated to dryness, the residue taken up in a little water, and the precipitation repeated. The combined filtrates are evaporated to dryness and the alkali determined as described on page 43 et seq. If, however, it is desired to separate magnesium from the alka- lies in order that the latter may be determined, the magnesium may be precipitated as magnesium hydroxide, from a solution free from ammonium salts, by the addition of barium hydroxide solution. % The barium is then removed by ammonium carbonate and the alkalies determined .in the filtrate. For the detailed description of this method see Silicate Analysis. Even in this case, however, the use of the Schaffgotsche method of separating magnesium from the alkalies is more satisfactory. * Z. anorg. Chem., 58, 427 (1908). t Pogg. Ann., 104, 482 (1858). J Cf . footnote to page 65. 70 GRAVIMETRIC ANALYSIS. METALS OF GROUP IV. CALCIUM, STRONTIUM, BARIUM. CALCIUM, Ca. At. Wt. 40.09. Forms: CaO, CaCO 3 , CaS0 4 . I. Determination as Calcium Oxide (Lime), CaO. For the determination of calcium as CaO, it is best precipi- tated as the oxalate and converted to the oxide by strong ignition. Procedure. The neutral or slightly ammoniacal solution, which besides magnesium and the alkalies should contain no other metal,* is treated with ammonium chloride, heated to boiling, and precipitated by the addition of a boiling solution of ammonium oxalate. After standing some time, the precipitate becomes coarsely crystalline and settles to the bottom of the beaker, when a little more ammonium oxalate solution is added to make sure that the precipitation has been complete. It is allowed to stand four hours, when the clear liquid is poured through a filter, the pre- cipitate is covered with boiling water containing ammonium oxalate, f allowed to settle, filtered, and the operation repeated three times. " Finally the whole precipitate is transferred to the filter and washed with hot water containing ammonium oxalate until free from chloride. The precipitate is warmed in the hot closet until nearly dry, when it is placed together with the filter in a platinum crucible and ignited wet. It should be heated cautiously at first in order that the too rapid evolution of carbonic oxide will not cause loss. After the filter is burnt the crucible is covered and strongly heated at first over the Teclu burner and finally over the blast-lamp for twenty minutes. The crucible while still quite warm is placed, in the desiccator shown in Fig. 7, near an open weighing-beaker and allowed to re- main there for one hour. During the cooling, the air enters the desiccator, through the U tube, whose outer half is filled with *H magnesium is present, the precipitation should be carried out as described on page 76. t T. W. Richards found that the precipitate was appreciably soluble in pure water but practically insoluble in a dilute solution of ammonium oxalate (Z. anorg. Chem., 28, 85 (1901)). DETERMINATION OF CALCIUM AS SULPHATE. 71 soda-lime and whose inner half contains calcium chloride, in a dry condition and free from carbonic acid gas. The crucible is now placed in the weighing-beaker, quickly covered and allowed to stand for half an hour in the air near the balance, after which it is weighed. The crucible is again heated over the blast-lamp for ten minutes and is cooled in exactly the same way and weighed. Should the weight not be found constant, the process must be repeated. The above directions when carefully followed usually enable one to obtain a constant weight on the second ignition. Example. Calcite: 0.5 gm. of the finely powdered material which has been dried at 100 is placed in a 300-c.c. beaker and moistened with 20 c.c. of water. The beaker is covered with a watch-glass, concentrated hydrochloric acid is added drop by drop, and the liquid is finally heated until all is dissolved. The solution is then boiled for fifteen minutes to expel all carbon dioxide, neutral- ized carefully with ammonia, diluted with 150-200 c.c. of hot water, and precipitated with ammonium oxalate as above described. Remark. If both solutions are not boiling hot during the precipitation, the calcium oxalate forms very fine crystals; it then settles very slowly and passes readily through the filter. Calcium oxalate is almost insoluble in water and acetic acid in the presence of ammonium oxalate, but readily soluble in mineral acids. 2. Determination of Calcium as Sulphate, CaS0 4 . This method is chiefly used for the analysis of calcium salts of organic acids. For this purpose the calcium salt is ignited in a weighed platinum crucible, after which the crucible is covered with a watch-glass, carefully treated with dilute sulphuric acid and warmed upon the water-bath until there is no longer any evolution of carbon dioxide. The under side of the watch-glass is carefully washed and the liquid evaporated as far as possible on the bath. The excess of sulphuric acid is then carefully driven off by inclining the crucible and heating over the free flame (or in an air-bath) (cf. Fig. 11, p. 27). The residue is gently ignited and weighed. By strong ignition, calcium sulphate loses SO 3 .* * 0.2052 gm. CaSO 4 remained unchanged in weight after heating for one hour to dark redness ; but on heating with the full flame of a Teclu burner, it lost 0.0004 gm. in weight. On heating for one hour over the blast lamp it bit 0.0051 gm. (J. Weber.) 72 GRAVIMETRIC ANALYSIS. Calcium, may also be precipitated as calcium sulphate. The solution, which should contain as little free acid as possible, is treated with an excess of dilute sulphuric acid, four volumes of alcohol are added, and the mixture is allowed to stand 12 hours. It is then filtered off, washed with 70 per cent, alcohol, dried, separated from the filter as completely as possible, the filter burned in a platinum spiral, and the ash added to the main part of the precipitate in a platinum crucible, gently ignited and weighed. . 3. Determination of Calcium as Carbonate, CaC0 3 . Only in rare cases is calcium precipitated as carbonate by ammonium carbonate in the presence of ammonia. The filtered and washed precipitate is gently ignited and weighed as carbonate. After weighing it is necessary to moisten the residue with a little ammonium carbonate solution, evaporate to dryness on the water-bath, and again ignite gently. This is done in order to change small amounts of calcium oxide, which may have been formed during the burning of the filter-paper, back to carbonate. In the presence of considerable ammonium chloride the pre- cipitation of calcium by means of ammonium carbonate is not quite complete, whereas the precipitation with ammonium oxalate always is. Consequently it is advisable in all cases to precipitate the calcium as oxalate and weigh it as the oxide, if the oxalate is ignited gently, however, it can be weighed as the carbonate. STRONTIUM, Sr. At. Wt. 87.63. Forms: SrS0 4 , SrCO 3 , SrO. The determination as the sulphate is the most accurate. Determination of Strontium as Sulphate, SrSO 4 . Procedure. To the neutral solution containing strontium, ?, slight excess of dilute sulphuric acid is added and as much alcohol as there is volume of solution. After stirring well, the mixture is allowed to stand twelve hours, filtered and washed, at first with 50 per cent, alcohol, to which a little sulphuric acid has been added, and finally with pure alcohol until the wash water no longer gives the sulphuric acid reaction. The precipitate is DETERMINATION OF STRONTIUM AS OXI^E OR CARBONATE. 73 dried and ignited as described under the determination of cal- cium as sulphate. Solubility of Strontium Sulphate according to Fresenius. 6895 parts of water at the ordinary temperature (17-20) dis- solve 1 part of SrSO 4 . 9638 parts of boiling water dissolve 1 part SrS0 4 . The sulphate is less soluble in water containing sulphuric acid: 12,000 parts of water containing sulphuric acid dissolves 1 part SrSO 4 . The precipitate is soluble in concentrated sulphuric acid, so that the water should not contain very much sulphuric acid. In cold, dilute hydrochloric or nitric acid, strontium sulphate is considerably more soluble, and also in solutions containing acetic acid, magnesium chloride, or alkali chloride. If, therefore, considerable free acid is present, it should be removed by evaporating the solution to dry ness and dissolving the residue in water. 'The strontium is then precipitated as above described. Determination of Strontium as Oxide, SrO, or as Carbonate, SrCO 3 . The strontium is precipitated as carbonate, or in some cases as oxalate, and changed by ignition to the oxide as described under calcium. Strontium carbonate is decomposed by heat more difficultly than calcium carbonate and the determination as carbonate is very satisfactory. It is advisable to treat the precipitate as described under calcium, although it is usually unnecessary to heat with additional ammonium carbonate. Solubility of Strontium Carbonate in Water according to Fresenius. 18,045 parts of water dissolve at ordinary temperatures 1 part of SrC0 3 . In water containing ammonium carbonate the salt is much less soluble; on the other hand, ammonium chloride and ammo- nium nitrate increase its solubility. In case calcium, strontium, magnesium and alkali salts are present together, as in minerals and in mineral waters, the calcium 74 GRAVIMETRIC ANALYSIS. and strontium are both precipitated as oxalates and transformed by ignition into the oxides. Cf. pp. 78, 79. Solubility of Strontium Oxalate in Water. 12,000 parts of water at ordinary temperatures dissolve 1 part of SrC 2 O 4 + 2H 2 O. The solubility is very much less in water containing ammonium oxalate. BARIUM, Ba. At! Wt. 137.37. Forms: BaS0 4 , BaCr0 4 , BaC0 3 . i. Determination as Barium Sulphate. The solution, slightly acid with hydrochloric acid, is heated to boiling and precipitated by the addition of an excess of hot, dilute sulphuric acid. It is allowed to stand on the water-bath until the precipitate has settled, the solution is then poured through a filter, and the precipitate is washed four times with 50 c.c. of water to which a few drops of sulphuric acid have been added. The pre- cipitate is transferred to the filter and washed with hot water until the wash water ceases to give the sulphuric acid reaction. It is then dried somewhat, ignited wet in a platinum crucible, and weighed without previous heating over the blast- lamp. Remark. By the combustion of the filter-paper there is usually a partial reduction of the barium sulphate to sulphide, but the latter, on being gently ignited in an inclined crucible, is readily changed back to sulphate, so that there is no loss to be feared. The procedure for the determination of barium as carbonate is the same as was described under calcium. Solubility of Barium Sulphate in Water. 344,000 parts of water dissolve 1 part of BaS0 4 . 2. Determination of Barium as Chromate. The neutral solution of the barium salt is diluted to about 200 c.c., treated with 4-6 drops of acetic acid (sp. gr. 1.065), heated to boiling, precipitated with a slight excess of ammonium chromate (prepared by adding ammonia to a solution of ammonium bichro- mate free from sulphate, until the color becomes yellow), and DETERMINATION OF BARIUM AS CHROMATE. 75 allowed to cool. The precipitate is filtered off through a Gooch crucible and washed with hot water until 20 drops of the filtrate give scarcely any reddish-brown coloration with a neutral solution of silver nitrate. The precipitate is dried in the hot closet, after which the crucible is fastened to a larger porcelain crucible by means of an asbestos ring, so that there remains a space of about J cm. between the two crucibles (cf. p. 27), and the open crucible is ignited over the free flame until the precipitate becomes a bright yellow.* Solubility of Barium Chromale.-f 86,957 parts of water at ordinary temperatures dissolve 1 part BaCrO 4 . 23,000 " " boiling water dissolve 1 part BaCrO 4 . 49,381] " " a 0.75 per cent, ammonium acetate solution (at 15) dissolve 1 part BaCrO 4 . 45,152 parts of a 0.5 per cent, ammonium nitrate solution (at 14) dissolve 1 part BaCrO 4 . 23,555 parts of a 1.5 per cent, ammonium acetate solution (at 15) dissolve 1 part BaCrO 4 . 22,988 parts of 0.5 per cent, ammonium nitrate solution dissolve 1 part BaCrO 4 . 3,670 parts of 1 per cent, acetic acid solution dissolve 1 part BaO0 4 . 2 618 " " 5 fl " " " " {t 1 " " 1 OSf> " " 10 " '' " " " " 1 " " 1,813 " " 10 " " chromic acid solution dissolve 1 part BaOO 4 . The solubility of barium chromate, therefore, increases con- siderably with increasing concentrations of either acetic or chromic acids; the solubility is affected to a much less degree by solutions containing neutral ammonium salts. By the additions of small amounts of neutral ammonium chromate the solubility becomes lessened to nearly zero. * Oftentimes small amounts of the precipitate are reduced to chromic oxide by traces of organic matter, whereby it appears slightly greenish. By long-continued ignition in an open crucible, the chromic oxide is changed back to chromate, when the precipitate appears a homogeneous yellow throughout. f P. Schweizer, Z. anal. Chem., 1890, p. 414, and R. Fresenius, ibid., 1890, p. 418. 76 GRAVIMETRIC ANA LYSIS. SEPARATION OP THE ALKALINE EARTHS FROM MAGNESIUM AND FROM THE ALKALIES. ic Separation of Calcium from Magnesium (and Alkalies). The separation depends upon the different solubilities of the two oxalates. Calcium oxalate is practically insoluble in hot water, whereas magnesium oxalate is fairly soluble; 1500 parts of cold water, or 1300 parts of boiling water, dissolve 1 part of MgC 2 O 4 -2H 2 0. In an excess of the precipitant, however, mag- nesium oxalate is much more soluble owing to the formation of complex salts. If calcium is precipitated as oxalate from a dilute solution in the presence of magnesium, some magnesium oxalate is occluded by the calcium oxalate precipitate even when the solution is by no means saturated with magnesium oxalate, and high results are obtained for calcium. In such cases the error is usually compensated according to Fresenius, by dissolving the pre- cipitate in hydrochloric acid and repeating the precipitation with ammonia and ammonium oxalate. The work of T. W. Richards * has shown, however, that the quantity of magnesium oxalate occluded is dependent upon the concentration of the undissociated magnesium oxalate in solution, and upon the time in which the calcium oxalate is in contact with the solution. Consequently anything that prevents the dissociation of magnesium oxalate will tend to increase the amount of occlusion and thereby increase the result of the calcium determination. Anything that will cause the ionization of the magnesium oxalate serves to reduce the amount of error. Concentrating the solution and increasing the amount of oxalate ions in solution by the addition of considerable ammo- nium cxalate, both tend to repress the dissociation of the mag- nesium oxalate. The dissociation of this compound is favored by the presence of hydrogen ions and by dilution. A considerable excess of ammonium oxalate is necessary in the quantitative precipitation of calcium and, moreover, mag- * Z. anorg. Chem., 28, 71 (1901). ALTERNATE METHOD. 77 nesium oxalate forms readily soluble complex salts with undisso- ciated ammonium oxalate. It is desirable, therefore, to prevent, the dissociation of the ammonium oxalate as much as possible, and this is accomplished by the addition of another ammonium salt, preferably the chloride. Procedure. Dilute the solution with hot water so that the magnesium is present in a concentration of not over fiftieth normal, and add a considerable quantity of ammonium chloride, if it is not already present. To precipitate the calcium, add a sufficient volume of boiling oxalic acid solution, containing three or four equivalents of hydrochloric acid to lessen the dissociation of the oxalic acid. Introduce a little methyl orange into the boiling solution, and add ammonia until a yellow colora- tion is produced. The ammonia should not be added all at once, but in small quantities from time to time so that about half an hour is consumed in the operation. After the neutralization add a considerable excess of hot ammonium oxalate solution, allow to stand four hours but not longer on account of the fact that the occlusion increases with the length of time, filter, and wash with hot, one per cent, am- monium oxalate until the. filtrate gives no test for chloride after .acidifying a sample with nitric acid. Although this precipitate contains a little magnesium (0.1- 0.2 per cent.) it is also true that an almost equal amount of magnesium passes into the filtrate, so that it is not advisable to repeat the precipitation when made in this way. Alternate Method. W. C. Blasdale * has studied the separation of calcium from different amounts of magnesium and succeeded in getting excellent results by a somewhat simpler method. The acid solution con- taining not more than 0.6 gm. of calcium and magnesium oxides is brought to a volume of 300 c.c., and heated to boiling; 3.5 gms. of ammonium chloride arc added and 1 gm. of oxalic a"H, and the complete precipitation of the calcium is brought about by neutral- * J. Am. Chem. Soc., 31, 917 (1903). 78 GRAVIMETRIC ANALYSIS. izing with 1 per cent, ammonia added during five minutes. The precipitate is allowed to stand for an hour before filtering. If considerably more magnesium than calcium is present, the precipitation is effected in two stages. First only enough oxalic acid is added to combine with all the calcium present. The solution is slowly neutralized as before and allowed to stand ten minutes. Then the balance of the oxalic acid is added, the solu- tion made alkaline and allowed to stand for an hour. With more than ten times as much magnesium as calcium, a double precita- tion is desirable. In the filtrate the magnesium can be precipitated by the addition of ammonia and sodium phosphate. If, however, large amounts of ammonium salts are present it is preferable to evaporate to dryness in a platinum or porcelain dish, to dry the residue at 110-130 for an hour or longer, and to expel the ammonium salts by gentle ignition. The residue is taken up in a little warm hydrochloric acid, diluted with water, a little carbon residue filtered off and the magnesium determined as pyrophosphate. (Pages 66-69.) II. Separation of Strontium from Magnesium. This separation finds practical application in the analysis of almost all mineral waters and of minerals containing strontium. In all of these cases, however, strontium occurs in relatively small amounts in the presence of large amounts of calcium and varying amounts of magnesium, so that it is a question, first, of separating calcium and strontium from magnesium. This separation is effected by the precipitation of the calcium and strontium as oxalates as described on pp. 70 and 73. The filtrate containing magnesium may also contain traces of strontium; hence, after the removal of the ammonium salts by ignition, the residue is dissolved in hydrochloric acid, sulphuric acid and alcohol are added, and the solution is allowed to stand tor twelve hours. Any resulting precipitate, consisting of strontium or barium sulphate, is filtered off and weighed. From this filtrate the magnesium is precipitated as magnesium ammonium phosphate as described on p. 66, and weighed as the pyrophosphate. SEPARATION OF BARIUM FROM MAGNESIUM AND STRONTIUM. 79 III. Separation of Barium from Magnesium. In case it is desired to separate only barium from magnesium, the solution (which must be free from nitric acid) is acidified with hydrochloric acid, heated to boiling, and the barium precipi- tated by the addition of boiling, dilute sulphuric acid (cf. p. 74). The magnesium is precipitated from the filtrate as magnesium ammonium phosphate in the usual way. In most cases, however, a separation of barium, strontium, and calcium from the magnesium is involved. For this purpose the three alkaline earths are pre- cipitated as oxalates, and any barium or strontium remaining in the filtrate is precipitated as described under II. The magnesium is determined in the final filtrate. IV. Separation of the Alkaline Earths from One Another. Principle. The mixture of the dry nitrates is treated with ether-alcohol, which dissolves calcium nitrate alone. The residue is taken up in water, the barium is precipitated as chromate, and the strontium is determined in the filtrate as sulphate. PROCEDURE. (a) Separation of Calcium from Strontium and Barium accord- ing to Rose-Stromeyer-Fresenius. The three metals are assumed to be present together in solution in the form of their nitrates. The solution is evaporated in a small Erlenmeyer flask, as described under lithium, p. 55, in an oil-bath, meanwhile passing a stream of dry, warm air through the flask. When all the water is evaporated, the temperature of the bath is raised to 140 C. and maintained at this temperature for one to two hours, still passing the current of warm air through the flask. After cooling, the dry residue is treated 'with ten times its weight of absolute alcohol, corked up, and allowed to stand with frequent shaking for one to two hours. An equal volume of ether is now added, the flask closed, shaken, and again allowed to stand twelve hours. It is then filtered through a filter moist- ened with ether-alcohol and washed with ether-alcohol until a few drops of the filtrate evaporated on platinum-foil leave no residue. 8o GRAYIIAETRIC ANALYSES. The filtrate is evaporated to dryness in a lukewarm water-bath, the calcium nitrate is dissolved in wate-\ precipitated as the oxalate, and after ignition is weighed as the oxide. Remark. In case only a small amount of calcium is present (not more than about 0.5 gm.) the above separation is complete. With large amounts of calcium, the residue of strontium and barium nitrates almost always contains some calcium. In this case the aqueous solution is again evaporated to dryness in the same way as before and the treatment with alcohol and ether re- peated. The calcium is then determined in the combined nitrates. This separation finds application in the analysis of most min- eral waters. (&) Separation of Barium from Strontium according to Fresenius. * Requirements. 1. A solution of (NH 4 ) 2 CrO 4 (1 c.c. of the solu- tion should contain 0.1 gm. of the salt). The solution is prepared by adding ammonia to a solution of ammonium bichromate (free from sulphate) until the color of the solution be- comes yellow. The solution should be left acid rather than alkaline. 2. A solution of ammonium acetate (1 c.c. containing 0.31 gm. of the salt). 3. Acetic acid of sp. gr. 1.065. 4. Nitric acid of sp. gr. 1.20. Procedure. The residue, consisting of strontium and barium nitrates, is dissolved in a little water and for each gram of salt mixture the solution is diluted until the concentration corre- sponds to 300 c.c., heated to boiling, treated with six drops of acetic acid and about 10 c.c. of ammonium chromate solution (this should be an excess over the theoretical amount necessary) and allowed to stand one hour. The precipitate of barium chromate is washed by decantation with water containing ammonium chro- mate until the wash water nc longer gives a precipitate with ammo- nia and ammonium carbonate; it is then washed with pure hot water until the last washing gives only a slight reddish-brown * Z. anal. Chem., 29, 427 (1890). SEPARATION OF BARIUM FROM STRONTIUM. 81 coloration with neutral silver nitrate solution. The precipitate rn the filter still contains a little strontium. It is carefully washed back into the vessel in which the precipitation took place, while any precipitate remaining on the filter is dissolved in a little warm dilute nitric acid and washed into the dish, finally adding enough nitric acid to the precipitate so that it dissolves completely on warming (about 2 c.c. of nitric acid being usually necessary). The solution is then diluted to 200 c.c., heated to boiling, treated with 5 c.c. of ammonium acetate solution, added little by little, and finally with enough ammonium chromate to cause the disappear- ance of the odor of acetic acid from the solution (usually about 10 c.c. are necessary). After standing one hour the liquid is poured through a Gooch crucible, the residue is treated in the dish with hot water, allowed to cool, then filtered and washed with cold water until the filtrate gives only a slight opalescence with neutral silver nitrate. The precipitate is dried, ignited gently in an air- bath (cf. p. 75), and weighed as BaCrO 4 . The filtrate is treated with 1 c.c. nitric acid, concentrated somewhat, and the strontium precipitated as carbonate by the addition of ammonia and am- monium carbonate. The precipitate, which always contains some chromate, is washed once with hot water, dissolved in hydrochloric acid, and the strontium determined as sulphate, according to p. 72. The results obtained according to this method are very satis- factory. Experiments performed in this laboratory* completely confirm the results obtained by Fresenius. Remark. In the opinion of the author, all other methods for the separation of the alkaline earths give incorrect results; for that reason they will not be discussed in this book. * The results of seven experiments gave (a) for the percentage of barium chromate obtained: 99.9, 99.9, 100.3, 100.4, 100.7, 100.6; mean, 100.3 per cent.; (6) for strontium sulphate, 100.9, 99.73, 99 F6, 99.S4, 99.47, 9C.77- 99.6; mean, 99.75 per cent. (H. Schmidt.) 82 GRAVIMETRIC ANALYSIS. METALS OF GROUP III. ALUMINIUM, CHROMIUM, TITANIUM, IRON, URANIUM, NICKEL, COBALT, ZINC, AND MANGANESE. A. DIVISION OF THE SESQUIOXIDES. ALUMINIUM, CHROMIUM, IRON, TITANIUM, AND URANIUM. ALUMINIUM, Al. At. Wt. 27.1. Form: A1 2 3 . In order to determine aluminium in this form, the metal is pre- cipitated as its hydroxide and converted to its oxide by ignition of the precipitate. It must not be forgotten, however, that aluminium hydroxide exists in a soluble form (hydrosol) and in an insoluble form (hydro- gel) ; and further that the hydrosol is not completely changed by boiling into insoluble hydrogel. To accomplish this the presence of salts in solution (preferably ammonium salts) is also necessary. Since, however, ammonium salts become acid on long boiling (due to the escape of ammonia) there is danger of the aluminium hy- droxide being redissolved . Furthermore, it is true that the hydrogel gradually changes over into hydrosol by standing in a solution containing only a small amount of dissolved salts, or by remaining in a hot solution containing only a small amount of dissolved salts. From these facts the following procedure is derived: The solution containing the aluminium (but no phosphoric acid, or anything but aluminium that is precipitated by ammonia) is treated with considerable ammonium chloride, or ammonium nitrate, heated to boiling in a platinum or porcelain vessel, and a slight excess of ammonia * is added. The precipitate is allowed to settle, after which the clear solution is poured through a filter which rests on a platinum cone, but without applying suction. *The ammonia should be freshly distilled. When it has been kept for any length of time in glass bottles, the ammonia invariably contains silica, and this leads to high results in the case of all precipitates formed by adding ammonia to an acid solution. DETERMINATION OF ALUMINIUM. 83 The precipitate is washed three times by decantation with hot water to which a drop of ammonia and a little ammonium nitrate has been added, and finally transferred to the filter. The pre- cipitate is now washed as quickly as possible with the hot wash liquid (so that the precipitate is thoroughly churned up each time) until the filtrate ceases to give a test for chlorine. The pre- cipitate is then dried as completely as possible by the applica- tion of suction and ignited wet in a platinum crucible. After the precipitate and ash have become white, the covered crucible is heated over the blast-lamp for about ten minutes, cooled in a desiccator and weighed. The process is repeated until a constant weight is obtained. Determination of Aluminium by the Method of Chancel.* In the determination of aluminium in a sample of alum by precipitation with ammonia, there are a number of difficulties to overcome. In the first place the precipitate always contains more or less basic aluminium sulphate, and it requires long ignition to expel the last traces of sulphuric anhydride. It is, to be sure, possible to wash out all the sulphate by means of a solution of ammonia and ammonium nitrate; the operation, however, is very tedious and a large quantity of the wash liquid is required. Another disadvantage arises from the fact that the filtration takes place very slowly, even when the precipitate contains basic sul- phate, on account of the slimy nature of aluminium hydroxide. By the method of Chancel and in the two following methods these difficulties are overcome. The principle of this and the following methods consists in neutralizing, by salts of weak acids, the mineral acid that is set free in the hydrolysis of an aluminium salt; the weak acid is finally removed by boiling or neutralization. In the Chancel method the mineral acid is neutralized by sodium thiosulphate, 2AlCl a + 6H a O=2Al(OH) 8 + 6HCl, 6HC1+ 3Na 2 S 2 3 = GNaCl + 3H 2 + 3S0 2 + 3S. Procedure. The dilute, neutral solution (about 0.1 gm. Al * Compt. rend., 46, $87; Z. anal. Chem., 3, 391. 84 GRAVIMETRIC ANALYSIS. in 200 c.c.) is treated with an excess of sodium thiosulphate and boiled until all traces of SO 2 are expelled. Enough ammonia * is then added so that the odor is barely perceptible after blowing away the vapors, and the boiling is continued a little longer. The precipitate of A1(OH) 3 and S is filtered off, washed with hot water, and ignited in a porcelain crucible. Such a precipitate of A1(OH) 3 is much denser than that produced by direct pre- cipitation with ammonia and is very easy to filter and wash. Remark. This method is often employed for separating alu- minium from iron. Ferric iron is reduced to ferrous salt by the sodium thiosulphate and is not precipitated. In this case, how- ever, the final neutralization with ammonia, as prescribed in the above directions, should not be carried out or a little iron will come- down. Determination of Aluminium by the Method of Alfred Stock.f The aqueous solution of the aluminium salt, which on account of hydrolysis always shows an acid reaction, A1C1 3 +3HOH^A1(OH) 3 +3HC1. is treated in the cold with a mixture of potassium iodide and potassium iodate. This mixture in the presence of acid is decom- posed in accordance with the equation KI0 3 + 5KI + 6HC1 = 3H 2 O + 6KC1 + 3I 2 , whereby the equilibrium which existed in the hydrolysis of the aluminium salt is disturbed and the first reaction continues to take place until all the aluminum is precipitated. If now the iodine in the solution is made to react with sodium thiosulphate and the mixture is heated on the water bath for half an hour, the precipitate collects in such a condition that it can be quickly filtered and washed. Procedure. The solution in which the aluminum is to be determined should be very slightly acid; if more acid is present sodium hydroxide is added until a slight permanent precipitate is obtained which is redissolved by means of a * If the addition of ammonia is omitted, the solution will retain traces of aluminium. t Ber., 1900, 548. DETERMINATION OF ALUMINIUM. 85 few drops of hydrochloric acid. Thereupon equal volumes of a 25 per cent, potassium iodide solution and a saturated potassium iodate solution (about 7 per cent. KIOs) are added. After about five minutes the solution is decolorized by the addition of a 20 per cent, sodium thiosulphate solution and a little of the potassium iodide and iodate mixture is added in order to make sure that enough was added in the first place. One or two c.c. more of sodium thiosulphate are added and the solution heated half an hour on the water bath. The pure white precipitate settles out well, and can be filtered through a filter with relatively wide pores, washed with hot water, ignited and weighed as A1 2 03. Remark. The presence of calcium, magnesium and boric acid does not interfere with the above determination, but if phosphoric acid is contained in the solution, the phosphate of aluminum is precipitated. It is obvious that the method cannot be employed in the presence of organic substances such as tartaric acid, citric acid, sugar, etc., which prevent the precipitation of aluminium hydroxide. Determination of Aluminium by the Method of G. Wynkoop * and of E. Schirm.f Principle. If a neutral solution of aluminium (iron, chro- mium, or titanium) is boiled with an excess of sodium or ammo- nium nitrite until no more fumes are evolved, aluminium hydroxide is precipitated in a form that can be as easily filtered as that of the last methods. 2 A1C1 3 + 6HOH<=2 Al (OH) 3 + 6HC1 ; 6HC1 + 6NaN0 2 = GNaCl + 6HNO 2 ; 6HNO 2 = 3H 2 +3NO +3NO 2 . Procedure. If the solution is acid, enough ammonia is added so that the precipitate first formed dissolves only slowly on * J. Am. Chem. Soc., 19, 434 (1897). f Chem, Ztg., 1909, 877. 86 GRAVIMETRIC ANALYSIS. stirring. An excess of a 6 per cent, solution of pure ammonium nitrite* is then added, the solution diluted to 250 c.c. and boiled until no more fumes of nitrous oxides are evolved (about 20 minutes). The precipitate is filtered, washed with hot water, ignited wet in a platinum crucible, and weighed as A1 2 3 . Remark. In the presence of more than 1 per cent, of ammonium salts, these are hydrolyzed enough so that the solution remains acid and the precipitation of the aluminium is incomplete even after long boiling. In such a case, after the vapors of nitrogen peroxide are no longer visible, ammonia is added drop by drop until the odor is barely perceptible. The precipitate is allowed to settle while the beaker is on the water-bath, and the analysis is finished as above. If the solution contains only aluminium in the form of its chlo- ride, nitrate, or sulphate, it can be determined by evaporating the solution in a platinum crucible on the water-bath with the addi- tion of a little sulphuric acid, the excess of the latter being finally removed by cautious heating over the free flame in an inclined crucible. The residue of aluminium sulphate is then changed to the oxide by strong ignition over the blast-lamp. In the case of organic salts of organic acids, the oxide is readily obtained by careful ignition of the salt in a platinum crucible. * Sometimes the reagent contains a little barium which should be pre- cipitated with ammonium sulphate before using it in an analysis. IRON. IRON, Fe. At. Wt. 55.84. Forms: Ferric Oxide, Fe 2 3 , and Metallic Iron. Determination as Fe 2 3 . (a) By Precipitation with Ammonia. This is the form chiefly used for the gravimetric determination of iron. The solution containing the ferric salt in the presence of ammonium chloride is heated to about 70 C. in a porcelain dish or Jena beaker * and pre- cipitated by means of a slight excess of am- monia. The precipitate is washed by decan- tation with hot water and finally with a strong stream of hot water from the wash- bottle, f It is ignited gradually in a plati- num crucible and finally in the half-covered crucible over the Bunsen burner.! The ferric oxide obtained varies in its appear- ance according to the temperature to which it has been heated. Gently ignited ferric oxide is reddish brown, whereas when strongly ignited it has almost the appearance of graphite. Both forms are difficultly soluble in dilute hydrochloric acid, but can be readily dissolved by digesting with con- centrated hydrochloric acid on the water-bath. (6) By Precipitation with Ammonium Nitrite. The precipitation of Fe(OH) 3 from neutral solutions of ferric salts takes place as described for aluminium, p. 85. * Too high results are invariably obtained in ordinary glass vessels. The ammonia should be freshly distilled or the results will be high. t For this purpose a wash-bottle such as is shown in Fig. 26 is useful. By blowing through the long arm of the U-tube (which is provided with a Bunsen valve) and placing the thumb over the short arm a continuous stream of water is maintained which can be stopped at any time by removing the thumb. I It is not advisable to heat the covered crucible over the blast on account of t 1 e danger of forming some Fe 3 O 4 . The magnetism of such a precipitate can be s' own by placing a magnet outside the platinum crucible and moving it slowly up and down. FIG. 26. 88 GRAVIMETRIC ANALYSIS. If the iron is in solution either as the ferrous or ferric salt of a volatile acid, it can be readily converted into ferric oxide by evapo- ration with sulphuric acid and ignition of the residue. 2. Determination as Metallic Iron. Iron may be determined by electrolysis, but this method offers no advantages over the gravimetric method just described or the following volumetric process, so tht it will not be discussed in this book. In the case of the analysis of oxide iron ores or of mixtures of considerable iron oxide with comparatively little alumina, titanium dioxide, or silica, the following method is accurate and rapid. ^ The finely powdered and weighed f substance contained in a porcelain boat is introduced into a tube of difficultly fusible glass and heated to redness in a stream of dry hydrogen until no more drops of water condense on the cool front end of the tube and the contents of the boat appear gray and not black. By this means the ferric oxide is reduced to metallic iron: Fe 2 O 3 + 3H 2 = 3H 2 + 2Fe. After cooling in the stream of hydrogen, the boat and its contents are again weighed after remaining some time in a desiccator. The loss in weight p represents the amount of oxygen originally com- bined with the iron, from which the amount of iron can be calcu- lated: 2Fe Remark. In attempting to reduce ferric oxide to iron by means of hydrogen, it is very important to heat the oxide to bright * Rivot, Ann. Chem. Phys., 3. Serie, 30, 188 (1850) ; Liebig's Ann., 78, 211 (185). f Ferric oxide after having been powdered and ignited is so hygroscopic that the porcelain boat should be placed within a weighing-beaker with ground -glass top immediately after removing it from the desiccator, and then weighed. VOLUMETRIC DETERMINATION OF IRON. 89 redness. At a dull red heat, the oxide is to be sure reduced to metal, but in such cases black, pyrophoric iron is formed and the latter cannot be exposed to the air and weighed without becoming oxidized. By heating to a bright red heat, however, the iron becomes gray, is no longer pyrophoric, and can, if allowed to cool in the stream of hydrogen, which is subsequently replaced by carbon dioxide, be safely weighed in the air without fear of oxidation. Although this method is extremely simple, and the correspond- ing oxides of aluminium, chromium, titanium and zircon, etc., are not reduced under the same conditions, it should be used with caution and only when the ferric oxide greatly predominates in a mixture of oxides. Otherwise the reduction of the iron is incom- plete on account of some of the ferric oxide being enclosed within the particles of foreign oxide. This has been proved by the work of Daniel and Leberle * and by Treadwell and Wegelin.f It is still more accurate to dissolve the metallic iron produced in dilute sulphuric acid out of contact with the air and determining the amount present volume trie ally by titrating with potassium permanganate solution. 3. Volumetric Determination of Iron, according to Margueritte.J Although the volumetric methods are discussed in the second part of this book, this determination is so important and is so often used to test the purity of the iron oxide produced by a gravi- metric analysis that it seems proper to discuss it at this place. Principle of the Method. Ferrous salts are oxidized by potassium permanganate in acid solution to ferric salts: If, therefore, a potassium permanganate solution of known strength is slowly added to the solution of a ferrous salt, it will * Z. anorg. Chem., 34, 393 (1903). f A table is given, showing the results of twelve experiments, in the German edition of this book. J Ann. de chim. et de phys. [3], 18 (1846), p. 244. 90 GRAVIMETRIC A NA LYSIS. be decolorized as long as there remains ferrous salt to react with it. As soon as all of the ferrous salt has been oxidized, the next drop of the permanganate will impart a permanent pink color to the solution, whereby the end-point of the reaction is determined. Preparation and Standardization of the Permanganate Solution. In most cases a T V normal solution of potassium permanganate is suitable, i.e. one which contains jn one liter enough oxygen to oxidize T V of a gram-atom of hydrogen (1.008 gm. of hydrogen). Potassium permanganate in acid solution reacts according to the equation K 2 0, Mn 2 O 7 = K 2 O+2MnO+5O, 2KMnO 4 so that from two molecules of permanganate five atoms of oxygen (=10 atoms of hydrogen) are available, thus: 2KMnO 4 KMnO 4 158.03 ^ = - -= - =31.61 gm. KMnO 4 = i gm.-atom of 10 5 5 oxygen = 1 gm.-atom of hydrogen. Consequently it is necessary to take -V f a gram-molecule of potassium permanganate (3.161 gms.) for a liter of T V normal solu- tion. Although it is possible to purchase very pure potassium per- manganate, it is not advisable to take the trouble of weighing out just this amount of the substance and dissolving it in exactly the right amount of water, for although we might in this way obtain the correct strength of solution, yet on the following day its value would be different, for the distilled water in which the perman- ganate is dissolved almost always contains traces of organic matter oxidizable by the permanganate. Consequently we weigh out on a watch-glass approximately the right amount of permanganate 3.1-3.2 gms.), dissolve it in a liter of water, and allow it to stand eight to fourteen days * before using it. After this time all of the oxidizable matter in the water will have been completely destroyed. The solution is filtered through an asbestos filter and then stand- ardized. * Cf. Morse, Hopkins, and Walker, Am. Chem. Jour., 18 (1896), p. 401. VOLUMETRIC DETERMINATION OF IRON. 91 Standardization of the Potassium Permanganate Solution. It is possible to standardize the solution by a number of differ- ent methods, as will be discussed in detail under volumetric analysis. In this case we are concerned with the determination of iron only, so that the most natural way for us to standardize the solution will be by means of chemically pure iron. An accurately weighed por- tion of iron is dissolved in dilute sulphuric acid out of contact with the air and permanganate solution is added from a glass-stoppered burette until the solution remains pink for one-half minute after thoroughly stirring or shaking. If for the oxidation of a grams of iron t cubic centimeters of the permanganate solution were necessary, then a 1 c.c. = gm. iron. The value represents the titration value of the solution. Remarks. The chief difficulty lies in the procuring of a suitable standard. It is difficult to prepare iron which is exactly 100 per cent. pure. Moreover, the purity of a sample of iron wire, such as is ordinarily used for standardization, cannot be deter- mined satisfactorily by means .of a gravimetric analysis because the analytical errors are greater than is usually supposed. The ferric hydroxide precipitate is bulky so that it is customary to take only about 0.2 gm. of wire for analysis which yields approximately. 0.283 gm. of Fe 2 3 so that an error of 0.0003 gm. in the final weight corresponds to 0.1 per cent, of iron in the sample.' The precipitate of ferric hydroxide, moreover, often contains silica and alumina when the analysis is carried out in glass beakers, or when the reagents used have been in contact with glass for some time. This leads to high results and the error may amount to 0.003 gm. Again, in igniting a precipitate of hydrated ferric oxide great care must be taken or some of it will be reduced to magnetite as can be shown by a magnet held outside the crucible.* -* * The carbon from the filter-paper causes the reduction when the pre- cipitate and filter are heated together and too much heat is used at the start. Heating the covered crucible over the blast lamp also converts Fe 2 O 3 into Fe 3 04. 92 GRAVIMETRIC ANALYSIS. When magnetite is thus once formed it is practically impossible to change it back to ferric oxide by further heating or by treating the precipitate with concentrated nitric acid. This leads to low results. When thrown down from hydrochloric acid solutions by the addition of ammonia, the precipitate often contains some chloride which it is hard to remove by washing, and on igniting such a precipitate there is danger of losing a little ferric chloride by volatilization. On account of these sources of error it is difficult to carry out a perfect gfavimetric estimation of iron with an accuracy sufficient for standardization purposes. The most satisfactory method is to determine the impurities present; but even although this may give the correct percentage of iron, if there is any carbon, phosphorus, silicon or sulphur present, as is usually the case, compounds may be left in solution which are oxidizable by potassium permanganate so that a sample of iron wire may show against permanganate an apparent iron value of from 100 to 101 per cent., in spite of the fact that it contains only 99.7 per cent, of pure iron. A. Classen * has proposed, therefore, to standardize potassium permanganate against pure electrolytic iron and in the author's laboratory this has been found to be very satisfactory. There is a possibility, however, of such electrolytic iron containing occluded hydrogen, carbon, etc., which may exert some effect upon the titration, although when the electrolysis is properly carried out, these errors cannot be large. G. Lunge, in his report to the Sixth International Congress of Applied Chemists f (Rome, 1907), recommended that permanganate should be standardized against one of four substances : (1) Pure oxalic acid, the exact value of which has been deter- mined by titration against standard barium hydroxide solution. The barium hydroxide is standardized against hydrochloric acid, which is in turn titrated against sodium carbonate. (See Acidimetry.) * Mohr-Classen, Lehrbuch der chem.-analyt. Titriermethode (1896). t See also Z. Angew. Chem. 17, 195 (1904). PREPARATION OF ELECTROLYTIC IRON. 93 (2) Pure iron wire, the exact value of which is known by a comparison with oxalic acid (standardized as above) . (3) Sodium oxalate. (See Acidimetry.) (4) Hydrogen peroxide by the Nitrometer Method. Preparation of Electrolytic Iron. Some commercial ferrous chloride is dissolved in water and a little hydrochloric acid, the solution is saturated with hydrogen sulphide, and filtered. After boiling off the excess of hydrogen sulphide, the ferrous chloride is oxidized by means of potassium FIG. 27. chlorate and hydrochloric acid, the excess of chlorine boiled off,, and the solution neutralized by the careful addition of sodium hydroxide. The iron is then precipitated by the barium carbonate method (see p. 149). The well-washed precipitate is dissolved in hot, dilute hydrochloric acid and freed from barium by a double precipitation of the iron with ammonia. The hydrated ferric oxide thus obtained is dried, ignited, and reduced to metal in a stream of hydrogen as described on p. 88. The metallic iron is dissolved in the calculated amount of dilute sulphuric acid out of contact with the air (passing C(>2 through the solution*). The * For the generation of carbon dioxide, an apparatus similar to that shown in Fig. 30 is used, only the wash-bottle A is filled with permanganate solu- tion, and the tower C contains pumice soaked with copper sulphate solution, above which is a plug of cotton . The potassium permanganate and copper sulphate both serve to remove H 2 S from the CO 2 . 94 GRAVIMETRIC ANALYSIS. solution is then diluted with water until 20 c.c. contain about 0.35 gm. iron. In addition it is also necessary to provide a solution of am- monium oxalate, saturated at the room temperature. For the electrolysis, two electrodes K (Fig. 27) are prepared by taking two pieces of platinum-foil about 25 sq. cm. surface and FIG. 28. fastening a piece of fairly heavy platinum wire to each; they are bent so that they will conveniently pass through the neck of a liter-flask. The electrodes are cleaned by boiling in con- centrated hydrochloric acid and finally igniting them over the free flame. To accomplish the latter purpose, it is convenient to hang them upon a heavy platinum wire which is itself placed PREPARATION OF ELECTROLYTIC IRON. 95 on an iron ring; they are then heated over the non-luminous name of the Teclu burner (Fig. 28).* After the ignition the electrodes are allowed to cool in a desic- cator and weighed accurately by the method of swings (cf. p. 10). About 350 c.c. of the ammonium oxalate solution are now placed in a 400-c.c. beaker and 20 c.c. of the iron solution (about 0.35 gm. Fe) are added. The beaker is covered with a glass plate containing three holes (Fig. 29). At the ends of the plate are fastened two corks which serve to support the two heavy platinum wires a and 6. Through the two side holes are passed from below the bent platinum wires of the FIG. 29. cathodes K, leaving them suspended from a; while through the middle hole the end of the spiral anode passes and is suspended from the cross-wire 6. The wire a is now con- nected with the negative, and wire b with the positive pole of a battery, and the solution is electrolyzed for from one and one-half to two hours at about 60 with a current of 0.5-0.7 ampere. At the end of this time there will be firmly attached to each of the cathodes about 0.15-0.17 gm. of a bright, steel-gray deposit. The circuit is broken, one of the electrodes is removed, and the circuit again closed. The electrode which has been removed is * The electrodes must be above the inner flame mantle. 96 GRAVIMETRIC ANALYSIS. at once plunged into a beaker of distilled water, taken out, the bottom edge touched with a piece of filter-paper to remove the greater part of the adhering water, and then washed with a liberal quantity of absolute alcohol that has been distilled over lime. The lower edge is again touched with filter-paper, then washed with ether which has been distilled over potash, after which it is dried in the hot closet until the ether has evaporated (this takes about half a minute). It is then placed in a desiccator. The sec- ond electrode is now removed from "the circuit and subjected to precisely the same treatment. After the electrodes have been in the desiccator for fifteen minutes they are weighed. While the solution is being electrolyzed the solvent for the iron should be prepared. In the liter-flask K (Fig. 30) are placed 500 c.c. of water and 50 c.c. of chemically pure concentrated sul- phuric acid. The contents of the flask are heated to boiling, meanwhile passing a stream of carbon dioxide through the liquid. After the latter has boiled vigorously -for ten minutes the flask is closed at b, removed from the flame, placed in cold water, and allowed to cool in an atmosphere of carbon dioxide. In this manner a solution of sulphuric acid is obtained com- pletely free from air, so that there is no danger of its oxidizing any of the ferrous salt. One of the weighed electrodes, on which the iron has been de- posited, is thrown into the flask containing the sulphuric acid; the flask is immediately closed and gently heated on the water-bath, or better to boiling, still passing carbon dioxide through the apparatus. The iron dissolves very quickly, leaving no residue.* The flask is then closed at 6, placed in cold water, and titrated with permanganate solution added from a glass-stoppered burette. After noting the burette reading, the permanganate is added drop by drop and the flask is constantly rotated to insure thorough mix- ing of the permanganate with the ferrous solution. When the solu- tion possesses a slight pink color, permanent for half a minute, the * Sometimes a minute trace of carbon remains undissolved, but it is so small in amount that it can safely be disregarded. t See under Volumetric Analysis. PREPARATION OF ELECTROLYTIC IRON. 97 end-point is reached, and after the burette has drained a second reading is taken. A blank test is made with another portion of K FIG. 30. 500 c.c. of water and 50 c.c. sulphuric acid solution (boiled free from air in the same way and allowed to cool in a stream of carbon dioxide), to see how much permanganate is necessary to impart this pink color in the absence of iron. This amount should be subtracted from the total number of cubic centimeters of the per- manganate solution used in titrating the iron. The results obtained by this procedure are excellent. After the strength of the permanganate solution has been accurately determined by the above method the apparent iron value of a sample of iron wire may be determined. When this is known it is possible to determine accurately the strength of a new permanganate solution, or of the same solution at a future date, by titrating against a solution of the same wire. Thus Dr. Schudl obtained the following values by using three methods: 1 c.c. KMn0 4 = 0.005485 g. Fe with electrolytic iron; 1 c.c. " =0.005470 g. " " iodine; lc.c. " = 0.005468 g." " oxalic acid. 98 GRAVIMETRIC ANALYSIS. Determination of the Apparent Iron Value of Iron Wire. The wire is cleaned by rubbing with a piece of emery paper until it is perfectly bright. It is then passed through filter-paper until it no longer leaves a gray mark on the paper. The wire is wound round a dry glass rod, making a spiral, and a portion of 0.15-0.2 gm. is weighed out. This is dissolved, as described on p. 601, in a flask which is fitted with a Bunsen valve * and contains 55 c.c. of dilute sulphuric acid (50 c.c. water + 5 c.c. cone. H 2 SO 4 ) ; the solution is boiled f a few minutes after the iron has all dissolved. It is allowed to cool, and is then titrated with permanganate which has been standardized against electrolytic iron or sodium oxalate. The apparent iron value of the wire is then calculated. Every time a new supply of iron wire is obtained its apparent iron value should be determined. The following results of determinations made with great care by W. A. K. Christie in the author's laboratory illustrate the process : The permanganate solution was standardized against iron wire and it was found that 1 c.c. = 0.005600 gm. iron. The purity of the sample of commercial iron wire was found in three titra- tions to be 99.93, 100.0 and 99.92 per cent., the average of which is 99.94 per cent. The actual purity of the wire was deter- mined to be 99.7 per cent. The apparent purity, therefore, is greater than the actual purity, and if the latter were to be used in the computation the permanganate solution would be found a little too strong. The author recommends the standardization against elec- trolytic iron and compares the value obtained with that found with iron wire, the work all being done on the same day. Then the apparent iron value of the wire will be known for future * Still better is the use of the glass valve as recommended by Contat (Chem. Ztg., 1898, 298) and improved by Gockel (Z. Angew. Caem., 189;), 620). When the boiling is stopped, sodium bicarbonate is sucked back into the solution, and there is no overpressure 01 the outside. Cf. p. 601. t The boiling of the solution is necessary, as otherwise hydrocarbons or other reducing substances remain in solution so that too much perman- ganate is used ANALYSIS OF FERRIC COMPOUNDS. 99 standardizations and it will be necessary to standardize against electrolytic iron only when a new supply of iron is purchased. On the other hand, it must be admitted that the standardization against sodium oxalate is full as accurate and is easier to carry out. See Volumetric Analysis, p. 597. Analysis of Ferric Compounds according to the Method of Margueritte. From what has already been said, it is evident that in order to determine the amount of iron present in a solution by titration with potassium permanganate, it is necessary for the iron to be present entirely in the ferrous condition. In order, therefore, to apply this method to the analysis of ferric compounds, it is first necessary to reduce them completely. To effect the reduction of a ferric sulphate solution we can proceed as follows : The solution is placed in a 200-c.c. flask, acidified with one-tenth its volume of pure, concentrated sulphuric acid, the flask is closed with a stopper provided with two tubes through which gas can enter and leave the flask, the contents of the flask are heated to boiling and hydrogen sulphide is passed through the solution until it is perfectly colorless. The boiling is continued and carbon dioxide is now passed through the solution until the excess of hydrogen sulphide is completely removed. The solu- tion is then allowed to cool in an atmosphere of carbon dioxide and titrated exactly as in the standardization of the solution of permanganate. If t c.c. of permanganate were necessary to completely oxidize the solution and 1 c.c. of the permanganate corresponds to a gm. of iron, then the titrated solution evidently contains a . t gm. of iron. Besides hydrogen sulphide, a great many other substances can be used to reduce the ferric salt, e.g., zinc, sulphurous acid, stannous chloride. The use of these substances w r ill be discussed in the portion of this book devoted to Volumetric Analysis. Remark. The titration of a solution by means of potassium permanganate takes place preferably in a sulphuric acid solution; in the case of hydrochloric acid too high results will be obtained ioo GRAVIMETRIC ANALYSIS. (due to the fact that the permanganate oxidizes some of the acid), unless the oxidation takes place in a dilute solution in the presence of a large excess of manganous sulphate. See Volumetric Analysis. TITANIUM, Ti. At. Wt. 48.1. Titanium, when present in large amounts, is determined as its dioxide, TiO 2 ; but if only small amounts are to be determined, as in the case of many rocks and iron ores, the colorimetric method is preferable. (a) Determination as Titanium Dioxide. The titanium is precipitated from solution either by means of ammonia, or by boiling a solution strongly acid with acetic acid and containing considerable ammonium acetate; or, finally, by boiling the slightly acid solution of the sulphate. In all these cases it is precipitated as titanic acid, from which it is changed by ignition into TiO 2 . The two former methods are preferable to the third. Seo separation of titanium from aluminium. (6) Determination of Titanium Colorimetrically ; Method of A. Weller.* (Suitable for small amounts of titanium.) This determination depends upon the fact that acid solutions of titanium sulphate are colored intensely yellow when treated with hydrogen peroxide; the yellow color increases with the amount of titanium present and is not altered by an excess of hydrogen peroxide. On the other hand, inaccurate results are obtained in the presence of hydrofluoric acid (Hillebrand) ; con- sequently it is not permissible to use hydrogen peroxide for this determination which has been prepared from barium peroxide by means of hydrofluosilicic acid. Furthermore, chromic, vanadic, and molybdic acids must not be present, since they also give colorations with hydrogen peroxide. The presence of small amounts of iron docs not affect the reaction, but large amounts of iron cause trouble on account of the color of the iron solution. If, however, phosphoric acid is added to the colored ferric solution it becomes * Berichte, 15, p. 2593. DETERMINATION OF TITANIUM. 101 decolorized, and from such a solution the determination of titanium offers no difficulty. The solution in which the titanium is to be determined must contain at least 5 per cent, of sulphuric acid; an excess does not influence the reaction. The reaction is so delicate that 0.00005 gm. of TiO 2 present as sulphate in 50 c.c. of solution give a distinctly visible yellow coloration. For this determination a standard solution of titanium sulphate is required. This can be prepared by taking 0.6000 gm. of po- tassium titanic fluoride which has been several times recrystal- lized and gently ignited (corresponding to 0.2 gm. of TiO 2 ). This is treated in a platinum crucible several times with a little water and concentrated sulphuric acid, expelling the excess of acid by gentle ignition, finally dissolving in a little concentrated sulphuric acid and diluting with 5 per cent, sulphuric acid to 100 c.c. One cubic centimeter of this solution corresponds to 0.002 gm. TiO 2 . The determination proper is carried out in the same way as described on p. 60, under the colorimetric determination of am- monium. 50 c.c. of the solution which has been brought to a definite and accurately measured volume is placed in a Nessler tube be- side a series of other tubes, each containing a known amount of the standard titantium solution, filled up to the mark with water and each treated with 2 c.c. of 3 per cent, hydrogen per- oxide * (free from hydrofluoric acid) . The color of the solution in question is compared with the standards. This method is only suitable for the estimation of small amounts of titanium, as the shades of strongly colored solutions cannot be compared accurately. According to J. H. Walton, Jr.,f titanium in the presence of iron may be determined after fusing the finely powdered sub- stance with two or three times as much sodium peroxide. On extracting with water, sodium pertitanate goes into solution and ferric oxide is left behind. The filtered solution is acidified with sulphuric acid, , which is added until 5 per cent, of free acid is *The hydrogen peroxide solution is prepared shortly before using by dissolving commercial potassium percarbonate in dilute sulphuric acid, t J. Am. Chem. Soc., 29, 481 (1907). 102 GRAVIMETRIC ANALYSIS. present. The color of the solution is then compared with that obtained by fusing a known weight of TiO 2 with Na 2 O 2 , etc. CHROMIUM, O. At. Wt. 52.1. Forms: Chromic Oxide, Cr 2 3 ; Barium Chromate, BaCr0 4 . (a) Chromic Compounds. Determination as Chromic Oxide. - 1. By Precipitation with Ammonia or Ammonium Sulphide. If the chromium is present in solution as chromic compound it can be precipitated exactly as described under aluminium, by means of a slightest possible excess of ammonia * in the presence of considerable ammonium salts (or better still, by the addition of freshly prepared ammonium sulphide solution to the boiling solu- tion) . The precipitated Cr(OH) 3 is washed with dilute ammonium nitrate solution and ignited wet in a platinum crucible, being there- by changed to the oxide, Cr 2 03. The results obtained are always a few tenths of a per cent, too high on account of the formation of small amounts of alkali chromate even though the entire opera- tion takes place in platinum vessels. The alkali comes from the reagents. It can be shown that the ignited product contains a little chromate, as the aqueous extraction always possesses a slight yellow color and gives with silver nitrate a red precipitate of silver chromate. If phosphoric acid is present, it will be found in the precipitate. In this case the dried precipitate is fused in a platinum crucible with sodium carbonate and potassium nitrate, whereby sodium chromate and sodium phosphate are obtained. The melt is dis- solved in water, acidified with nitric acid, and the phosphoric acid precipitated by means of ammonia and magnesia mixture, as described under Phosphoric Acid. From the filtrate the chromium is determined as barium chromate in acetic acid solu- tion as described below. * An excess of ammonia prevents the complete precipitation of the chromium hydroxide, the filtrate is then colored pink. In such cases the filtrate must be boiled until the excess of ammonia is expelled, and the chromium is all precipitated. CHROMIUM. 103 2. By Precipitation with Potassium lodideIodate Solution. Method of Stock and Massaciu* The determination is carried out as in the case of aluminium (cf. p. 83). The slightly acid f solution, contained in a porcelain dish, is treated with a mixture of potassium iodide and iodate, decolorized after a few minutes by means of sodium thiosulphate solution, treated with a little more iodide and iodate and then again with a few c.c. of sodium thiosulphate, and heated half an hour on the water-bath. The flocculent precipitate of chromic hydrox- ide settles quickly, and is filtered preferably through a hot water filter under slight suction. The precipitate is ignited wet in a platinum crucible. 3. By Precipitation with Ammonium Nitrite. % If the solution of the chromic salt is acid, it is neutralized with ammonia until a slight permanent precipitate is obtained. This precipitate is dissolved by the addition of a few drops of hydrochloric acid and then an excess of 6 per cent, ammonium nitrite solution is added and the liquid boiled until all nitrous fumes have been expelled. By this means practically all of the chromium will have been precipitated, but in order to throw down the last traces, ammonia is added drop by drop until the odor of free ammonia barely persists in the solution. The precipitate is allowed to settle while the beaker remains on the water-bath, and is finally filtered off, washed with hot water, ignited wet in a platinum crucible, and weighed as Cr 2 3 . (6) Chromates. If the chromium is present in solution in the form of an alkali chromate, free from chloride and large amounts of sulphuric acid, * Ber., 1901, 467. f If the solution is strongly acid, it is neutralized by the addition of pure KOH solution drop by drop, until a faint permanent turbidity is obtained. t E. Schirm, Chem. Ztg., 1909, 877. Cf. p. 111. According to Scholler and Schrauth (ibid., 1909, 1287) iron, chromium, aluminium, and zinc can be precipitated by means of aniline. 104 GRAVIMETRIC ANALYSIS. it may be determined very accurately by precipitation with mer- curous nitrate solution as mercurous chromate; on ignition the latter is changed to Cr 2 3 . Procedure. The neutral or weakly acid solution is treated with a solution of pure mercurous nitrate whereby brown, basic mercurous chromate, (4Hg 2 0,3CrO 3 ) ; is formed. On heating to boiling, the precipitate becomes a beautiful, fiery red, being con- verted into the neutral salt Hg 2 CrO 4 . This red salt settles very quickly, and if the precipitation is complete the solution above the precipitate will be colorless. After cooling, the precipitate is filtered off, washed thoroughly with water containing a little mercurous nitrate, dried and separated from the filter as completely as possible. The filter is burned in a platinum spiral and ignited with the main portion of the precipitate, gently at first and finally strongly, in a platinum crucible under a hood with a good draft, afterwards weighing the residue as O 2 3 . The purity of the mercurous nitrate must be tested before using it. 5 gms. of the salt should leave no residue after being ignited. This excellent method for the determination of chromium unfortunately permits only a very limited application. If the solution contains any considerable amount of chloride, mercurous chloride will be precipitated with the mercurous chro- mate, which, although volatile on ignition, renders the precipitate too bulky and the method inaccurate. If, therefore, it is necessary to determine chromium present as chromate in a solution containing chloride, two other methods are at our disposal. The chromate may be reduced by boiling with sulphurous acid (or by evaporating with concentrated hydro- chloric acid and alcohol) and analyzed according to (a), or it may just as accurately, and much more conveniently, be deter- mined by precipitating as Barium Chromate, which is weighed after gentle ignition. Procedure. The neutral solution, or one weakly acid with acetic acid, is treated at the boiling temperature with a solution of barium acetate added drop by drop,* and after standing for some time, * If the barium acetate solution is added too quickly some of it will be BARIUM CH 'ROM ATE. 105 is filtered through a Gooch crucible (without using very strong suction, as otherwise the filter will soon get stopped up and the solution will filter extremely slowly). The precipitate is washed with dilute alcohol and dried in the hot closet. The crucible is suspended in a larger one of porcelain by means of an asbestos ring (cf. page 27) and heated, at first gently, and finally over the full flame of a good Bunsen burner. After five minutes the cover is removed and the heating is continued until the precipitate appears a uniform yellow throughout, when it is cooled in a desic- cator and weighed. Sometimes the precipitate appears green on the sides of the crucible owing to a slight reduction (by means of dust, traces of alcohol, etc.) of chromic acid to chromic oxide. The latter gradually takes on oxygen from the air during the long-continued heating of the open crucible, so that the green color gradually disappears. If a grams of chromate were taken for analysis, and the barium chromate precipitate weighed p grams, then the amount of chromium present may be calculated as follows: BaCrO 4 : Cr=p : s Cr and 100 Cr p a: = -p ,*. -= per cent. Cr. BaCr0 4 a Example for practice: Potassium bichromate, K^C^Oj, purified and dried as described on pages 33 and 35. Chromium originally present as chromate, or obtained as such after suitable oxidizing treatment, may be determined accurately by volumetric methods described in Part II. carried down with the barium chromate, so that too high results will be obtained. 106 GRAVIMETRIC A 'N 'A LYSIS. UfeANIUM, U. At. Wt., 238.5. Forms: U 3 8 and UO S , (a) Determination as U 3 3 . Uranium is almost always precipitated by means of ammonia as ammonium uranate and changed to U 3 8 by gentle ignition in a platinum crucible with free Access of air. According to Zimmerman * this transformation is only complete when the precipitate is ignited in a stream of oxygen ; the error is, however, so small that for ordinary purposes it can be neglected. According to the temperature of ignition, the U 3 O 8 appears dirty green or black, and is difficultly soluble in dilute hydro- chloric or sulphuric acids; in nitric acid it dissolves gradually. By heating with dilute sulphuric acid (I vol. cone. H 2 SO 4 + 6 vol. H 2 O) in a closed tube at 150- 175 C. for a long time (W. F. Hillebrand),t the U 3 O 8 is completely dissolved with the formation of uranous and uranyl sulphate: U 3 O 8 + 4H 2 SO 4 = 2UO 2 (SO 4 ) +U (SO 4 ) 2 f 4H 2 0. U 3 O 8 is also readily soluble in dilute sulphuric acid in the presence of potassium bichromate. These two last facts are taken advantage of in the volumetric determination of uranium (which see). (6) Determination as UO 2 . The ignited precipitate, obtained in exactly the same way as before, is heated over a good Teclu burner, or over the blast- lamp, in a current of hydrogen, until a constant weight is obtained whereby it is quantitatively changed to UO 2 . This is the most accurate method for the determination of uranium. The UO 2 thus obtained is a brown powder, insoluble in dilute hydrochloric and sulphuric acids, but soluble in concentrated sulphuric acid after long heating, best in a closed tube. This oxide is also soluble in nitric acid. * Ann. d. Ch. und Ph., 232 (1886), p. 287. f Bull. U. S. Geol. Survey, 78, p. 90. SEPARATION OF IRON FROM ALUMINIUM. 107 Separation of Iron, Aluminium, Chromium, Titanium, and Uranium from Calcium, Strontium, Barium, and Magne- sium. The solution containing the above substances in the presence of considerable ammonium chloride is placed in an Erlenmeyer flask and treated with a slight excess of freshly prepared ammo- nium sulphide free from sulphate and carbonate. After stand- ing overnight the precipitate is filtered off and washed with water containing ammonium sulphide. It contains the iron and uranium as sulphides, the aluminium, chromium, and titanium as hydroxides. In case large amounts of magnesium are present, some of it is almost always present in the precipitate, so that it is then necessary to dissolve the precipitate, after filtration, in hydrochloric acid and to reprecipitate with ammonium sulphide. Instead of using ammonium sulphide, the separation can oe accomplished satisfactorily with ammonia; the iron must then be in the ferric condition. Separation of Iron from Aluminium. (1) The solution is treated in a porcelain dish with pure potassium hydroxide solution until strongly alkaline, boiled, diluted with hot water, and filtered. The precipitate contains the iron as hydroxide, while the solution contains the aluminium as aluminate.* For the iron determination the precipitate is dissolved in hydrochloric acid, reprecipitated with ammonia,f dried, and weighed as Fe 2 O 3 (see page 83). The aluminium is precipitated as hydroxide from the filtrate by acidifying with nitric acid and then adding ammonia. (2) The acid solution is treated with tartaric acid (three parts of tartaric acid for each part of the mixed oxides (Fe 2 O 3 +Al 2 O 3 )) ; hydrogen sulphide is passed into the solution until it is saturated, as slight an excess as possible of ammonia is added, and the sulphide of iron is allowed to settle in a closed Erlenmeyer flask. It is then * If the precipitate is large, it should be dissolved in hydrochloric acid and again precipitated with KOH. t It is very hard to wash the KOH precipitate free from alkali so that the first precipitate should not be weighed. io8 GRAVIMETRIC ANALYSIS. filtered, washed with water containing ammonium sulphide, dis- solved in hydrochloric acid, oxidized with a little potassium chlorate or nitric acid, and precipitated as ferric hydroxide by the addition of ammonia. The aluminium is determined in the filtrate by evaporating to dryness with the addition of a little sodium carbo- nate and potassium nitrate. The residue is gently ignited in a platinum dish in order to destroy the tartaric acid, after which it is dissolved in dilute nitric acid,' the carbon filtered off, and the aluminium precipitated from the Solution by the addition of ammonia. (3) The neutral solution of the chlorides or sulphates (not the nitrates) is treated with sodium carbonate until a slight permanent precipitate is formed, which is dissolved by the addi- tion of a few drops of hydrochloric acid. The solution is diluted to about 250 c.c. for each 0.1 or 0.2 gm. of the metals present, an excess of sodium thiosulphate solution is added, and the solution boiled until every trace of SO 2 has disappeared. By this opera- tion the ferric salt is reduced to ferrous salt : 2Na 2 S 2 O 3 + 2FeCl 3 = 2NaCl + Na 2 S 4 O 6 + 2FeCl 2 , and the aluminium is precipitated as the hydroxide : 2 A1C1 3 + 3H 2 O + 3Na 2 S 2 3 = GNaGl + 3SO 2 + 3S + 2 Al (OH) 3 . The precipitate of aluminium hydroxide and sulphur is filtered off, washed with hot water, dried, transferred as completely as possible to a porcelain crucible, the filter burned in a platinum spiral and the ash added to the crucible, which is ignited gently until all the sulphur has been expelled and then more strongly over the blast or a Teclu burner until the weight is constant. To determine the iron, the filtrate may be acidified with hydro- chloric acid, the S0 2 boiled off, the sulphur filtered off, the solu- tion oxidized with nitric acid and precipitated by ammonia as described on page 87. It is still better to precipitate the iron with ammonium sulphide, filter, dissolve in hydrochloric acid, oxidize with nitric acid, and then precipitate with am- monia. SEPARATION OF IRON FROM ALUMINIUM. 109 (4) Both of the metals are precipitated with ammonia, filtered, washed, dried, ignited in a platinum crucible, and the weight of the combined oxides determined. The mixture is then digested with concentrated hydrochloric acid to which a little water has been added (10HC1:1H 2 O) in a covered crucible until the iron is com- pletely dissolved. If ferric oxide predominates, as is frequently the case, the solution is effected in one or two hours. If, on the other hand, a relatively large amount of alumina is present (as is usually the case with silicates), and which can be detected by the color of the precipitate produced by ammonia, the precipitate then dissolves very slowly and in many cases only incom- pletely. In the latter case the ignited oxides are brought into solution by fusing with 12-15 times as much potassium pyrosulphate, K 2 S 2 O7 (cf. Vol. I).* The decomposition of the oxides is usually complete in 2-4 hours. The crucible together with its cover is placed in a beaker, water and a little sulphuric acid are added, and the melt is dissolved by warming gently, and passing a current of air through the solution in order to keep the liquid in motion. A small amount of platinum is always dissolved by this treatment. After removing the crucible and its cover, the solution is heated to boiling and saturated with hydrogen sulphide. The solution is then filtered into a flask and carbon dioxide is passed through it until the excess of hydrogen sulphide is completely removed. The contents of the flask are then cooled by placing the flask in cold water, the carbon dioxide still passing through the flask. The iron is then titrated with potassium permanganate solution as described on page 99. The aluminium is determined by differ- ence, from the weight of the combined oxides. For the determina- tion of iron in silicates the above process is most suitable (Hillc- brand). The reduction of the ferric salt to ferrous salt by means of hydrogen sulphide possesses great advantages over the reduction by means of zinc, for in the former case no foreign element is intro- duced, and furthermore zinc serves to reduce the titanic acid * E. Deussen finds that fusion with KF.HF works better. The platinum is not attached and the solution is effected more readily. Z. angew. Chem., 1905, 815. no GRAVIMETRIC ANA LYSIS. that is almost always present in rocks, and this will be again oxidized by the permanganate, so that too high an iron value will be ob- tained. If the iron is all dissolved by treating the oxides with hydro- chloric acid, the solution is evaporated to dryness and the residue is treated with a few cubic centimeters of dilute sulphuric acid, evaporated on the water-bath as far as possible, and then heated over the free flame until fumes of sulphuric acid are evolved. After cooling, the product is dissolved in water and the ferric sulphate reduced to ferrous sulphate by introducing a piece of zinc, free from iron, into the crucible and covering the latter with a watch-glass.* The reduction is complete in 20-30 minutes. The slight residue of platinumf is filtered off with the excess of zinc into a flask already filled with carbon dioxide. The residue is washed with water that has been boiled, and the solution is titrated with potassium permanganate solution. The latter method is to be recommended for the determination of small amounts of iron in the presence of still less aluminium, as is the case in the analysis of mineral waters. The following procedure leads to the same end, but the results are not quite so reliable: The solution from which the iron and aluminium are to be determined is diluted to a definite volume (e.g., 250 c.c.) and two aliquot portions are taken by means of a pipette (usually 100 c.c.). In one portion the weight of the combined oxides of iron and aluminium is determined by precipitation with ammonia and ignition of the precipitate, while in the other the iron is determined by titration. If the solution contains hydrochloric acid, as is usually the case, the iron is first precipitated with ammonia, filtered, * If titanium is present, the solution is reduced by means of hydrogen sulphide. t Platinum is perceptibly attacked by long digestion with ferric chloride- solution: 4FeCl 3 + Pt + 2HC1= H 2 PtCl + 4FCC12. The chloroplatinic acid is reduced to platinum by the action of zinc. SEPARATION OF IRON, ALUMINIUM, PHOSPHORIC ACID, in washed, and dissolved in dilute sulphuric acid. The solution is then reduced and titrated as previously described.* Separation of Iron, Aluminium, and Phosphoric Acid. Although the determination of phosphoric acid has not yet been considered, we will describe its determination in the presence of iron and aluminium because this highly important separation is necessary in the analysis of almost all minerals containing iron and aluminium as well as in the analysis of many mineral waters. Two cases are to be distinguished: 1 . The solution contains only a small amount (a few centigrams or less) of iron, aluminium, and phosphoric acid. 2. The solution contains large amounts of these substances. 1. In the first case the determination of all three constituents must be undertaken in the same portion, as otherwise errors would be introduced on account of the small amounts to be determined. The solution is first treated with ammonia whereby the iron, alu- minium and phosphoric acid are precipitated.! The precipitate is ignited in a platinum crucible and weighed: The product is then fused with six times its' weight of a mixture consisting of four parts anhydrous sodium carbonate and one part pure silica. The mixture is heated over the blast-lamp, the melt is extracted with water, to which a little ammonium carbonate has been added, and filtered. The filtrate contains all of the phos- phoric acid and a very little silicic acid, while the residue contains all of the iron and aluminium and considerable silica. For the determination of the phosphoric acid, the filtrate is evaporated with hydrochloric acid on the water-bath to dryness, * It is necessary to get rid of the hydrochloric acid on account of its action upon potassium permanganate (cf. Vol. Anal., under Iron). f The phosphoric acid is usually present in such small amounts that the iron and aluminium are more than sufficient to effect the precipitation of all the phosphoric acid, on the addition of ammonia, as phosphates of these metals. U2 GRAy I METRIC ANALYSIS. in order to remove the silica, the residue is moistened with hydro- chloric acid, taken up in a little water, filtered, and the phosphoric acid precipitated in the filtrate by the addition of ammonia and " magnesia mixture." The precipitate of magnesium ammonium phosphate is changed to magnesium pyrophosphate by ignition and from its weight p the amount of phosphoric anhydride, P 2 O 5 , is calculated (=B): Mg 2 P 2 7 :P 2 5 lp:B, PA "ififf,> By subtracting B from A the combined weight of the iron and aluminium oxides is obtained, in which the iron is determined volumetrically and the aluminium by difference. For the deter- mination of the iron, the insoluble residue, obtained after treating the product of the fusion with water and ammonium carbonate, is digested with hydrochloric acid in a small porcelain crucible until the iron oxide is completely dissolved. The solution is treated with dilute sulphuric acid, evaporated on the water-bath as far as possible, and then over a free flame until fumes of sulphuric anhydride are evolved. After cooling, water is added and after digesting on the water-bath for a long time the silica is filtered off, the solution reduced by means of hydrogen sulphide (cf. p c 109, sub. 4), and, after removing the excess of hydrogen sulphide, the iron is titrated with permanganate solution.* From the amount of permanganate used, the amount of ferric oxide (C) can be calculated, and by deducting this amount from the weight of the combined oxides, the weight of the A1 2 O 3 is ascertained : A-(B + C)=A1 2 O 3 . 2. In case the solution contains large amounts of iron, alu- minium, and phosphoric acid, it is divided into three aliquot por- tions and in one the value of " A " is determined by precipitation with ammonia; in the second the phosphoric acid is determined by the molybdate method ; and in the third the iron is determined by titration. * Instead of reducing the iron, the ferric salt may be titrated directly with titanous chloride (cf. p. 699, or iodimetrica'ly (cf. p. 681). SEPARATION OF IRON FROM CHROMIUM. 113 Separation of Iron from Chromium. 1. The chromium is oxidized in alkaline solution by means of chlorine or bromine to a soluble chromate and the insoluble ferric hydroxide is filtered off. Procedure. The solution of the chlorides, which should be placed in an Erlenmeyer flask of Jena glass provided with a ground-glass stopper and tubes by which gas may enter and leave the flask , is treated with potassium hydroxide solution until strongly alkaline, warmed on the water-bath and chlorine gas is conducted through the liquid, or bromine water is added, until it becomes distinctly yellow and the ferric hydroxide has as- sumed its characteristic reddish-brown color. When the oxidation is performed by chlorine gas, 0.5 gm. of the mixed oxides will be completely oxidized in fifteen to twenty minutes. The solution is diluted with water and filtered. The filtrate is carefully acidified with acetic acid, the chromium precipitated by the addition of barium acetate, and the precipitate of barium chromate is treated as described on p. 104. The ferric hydroxide is dissolved in hydrochloric acid, reprecipitated with ammonia and weighed as ferric oxide. Remark. If the chromate is to be determined as barium chro- mate, the solution must contain no sulphuric acid. If the latter is present, the chromate is reduced by evaporating with hydrochloric acid and alcohol; the solution of chromic chloride thus obtained is precipitated with ammonia and the chromium determined as chromic oxide. In the case of a precipitate containing iron and chromic oxides, it is fused with sodium carbonate and a little potassium chlorate, the melt is extracted with water, and the chromium is determined in the solution by precipitating with barium acetate. The insoluble residue from the aqueous extraction of the fusion is dissolved in hydrochloric acid, precipitated with ammonia, and the iron deter- mined as ferric oxide. If it is desired to precipitate the chromium as mercurous chromate, the precipitate containing the iron and chromic oxides is fused with sodium carbonate and potassium nitrate, the melt H4 GRA 1/lME TRIC ANAL YSIS. extracted with water ? the solution neutralized with nitric acid and precipitated with mercurous nitrate solution, as described on p. 104. 2. It has been proposed to analyze the mixture of ferric and chromic oxides by strongly igniting them in a stream of hydrogen whereby the ferric oxide is reduced to metallic iron, while the chromic oxide is unchanged. The iron could then be determined by the loss of weight. This method, although theoretically very simple, seems from experiments carried out in the author's laboratory to be absolutely inadequate, for the ferric oxide is so enveloped in chromic oxide that it is not even approximately re- duced even when heated over the blast-lamp. 3. Iron may be separated from chromium by precipitating the former with ammonium sulphide from a solution containing sufficient ammonium tartrate to prevent the precipitation of the chromium. The separation is the same as was described under aluminium, p. 107, sub. 2. Separation of Aluminium from Chromium. If the chromium is present as chromic salt, it is oxidized by means of chlorine or bromine in a solution made strongly alkaline with potassium hydroxide. The solution is then acidified with nitric acid, and the aluminium precipitated by ammonia as hydrox- ide, being weighed as the oxide. In the absence of sulphuric acid the chromium may be determined in the filtrate as barium chro- mate (cf. p. 104). If sulphuric acid is present, the chromate is reduced to chromic salt again by the action of concentrated hydro- chloric acid and alcohol, precipitated with ammonia, and weighed as the oxide If, however, the chromium is already present as chromate, the aluminium is at once precipitated with ammonia as hydroxide. Separation of Iron from Titanium. It is frequently necessary to determine both iron and titanium in a precipitate produced by ammonia consisting of a mixture of these two oxides alone, but it is more often necessary to determine SEPARATION OF IRON FROM TITANIUM. 115 titanium in the presence of iron, aluminium, and phosphoric acid, all of which are precipitated by ammonia in the analysis of rocks. For the separation of titanium from iron in the absence of alumina, the following methods are suitable : 1. The precipitate produced by ammonia is ignited and then fused with 15-20 times as much of previously dehydrated potas- sium pyrosulphate over a small flame until completely attacked. After cooling, the melt is dissolved in cold water containing sul- phuric acid, and the solution is hastened by keeping the liquid in motion by means of a current of air passed through it. The solution thus obtained is diluted to a definite volume, and after being thoroughly mixed is divided into two portions, one being used for the determination of titanium and the other for the deter- mination of iron. For the iron determination, the acid solution is saturated with hydrogen sulphide in the cold, heated to boiling, and the precipitate of platinum sulphide, sulphur, and a little titanium is filtered off into a flask filled with carbon dioxide, and washed thoroughly with hot water. The filtrate is heated to boiling and carbon dioxide is passed through the solution until the excess of hydrogen sulphide is completely removed, when it is cooled in an atmosphere of carbon dioxide and then titrated with perman- ganate. For the titanium determination, the other part of the solution is treated with sodium carbonate solution until a slight precipitate is formed ; this is dissolved in as little sulphuric acid as possible, saturated with hydrogen sulphide in the cold, and 5 gms. of sodium acetate which has been neutralized with acetic acid * are added. Carbon dioxide is conducted through the solution, it is heated to boiling, filtered hot, washed with water containing hydro- gen sulphide, ignited wet in a platinum crucible, and weighed as Ti0 2 . Remark. If considerable iron is present, the titanic oxide thus obtained is likely to contain iron. It is brought into solution again by fusing with potassium pyrosulphate and the precipitation is repeated exactly as before. In this way a precipitate free from iron is obtained. 2. The Chancel-Stromayer method is also satisfactory. The solution from the pyrosulphate fusion, in this case after being * Cf. footnote t3 a^e ICO. Ii6 GRAVIMETRIC ANALYSIS. neutralized with sodium carbonate, is treated with an excess of sodium thiosulphate, diluted to about 400-500 c.c. and boiled for some time. In this way rnetatitanic acid and sulphur are precipi- tated, while iron remains in solution. During the filtration, how- ever, the finely divided sulphur passes through the filter, so that the first method is preferable. In the presence of considerable iron the metatitanic acid obtained by this method is also contam- inated with iron, so that the separation must be repeated. Separation of Aluminium from Titanium. It has been proposed to effect this separation by diluting to a considerable volume the slightly acid solution of the melt obtained by the potassium pyrosulphate fusion and boiling for some time, thereby precipitating the titanium and leaving the aluminium in solution. This method, however, is useless, for alumina is pre- cipitated with the metatitanic acid unless the solution contains enough acid to prevent this hydrolysis, in which case a considerable amount of titanic acid remains in solution. The best separation is that of Gooch;* it consists of boiling a solution of the two elements containing considerable free acetic acid and alkali acetate ; by this means all of the titanium and none of the aluminium is precipitated. If, however, the amount of aluminium present is large (as is usual in rock analysis), the pre- cipitate will contain some aluminium, so that the separation must be repeated. In no case is there danger of the precipitation of the titanium being incomplete. In practice it is almost always necessary to separate the titanium not from aluminium alone, but from iron and aluminium, so that the method of Gooch will be described for this more general case. The solution obtained by dissolving the pyrosulphate melt in cold water is treated with three times as much tartaric acid as the weight of the oxides, is saturated with hydrogen sulphide gas, and then made slightly ammoniacal. By this means all of the iron is precipitated as ferrous sulphide, while the aluminium and titanium remain in solution. The sulphide of iron is filtered off, the filtrate is acidified with sulphuric acid, heated to boiling, and the precipitate of sulphur and platinum sulphide (the latter from the platinum crucible in which the fusion with pyrosulphate was * Chemical News, 52, pp. 55 and 68. SEPARATION OF ALUMINIUM FROM TITANIUM. 117 made) is filtered off. The filtrate is boiled to expel the last traces of hydrogen sulphide and the tartaric acid is destroyed by adding 2 times as much potassium permanganate as the amount of tar- taric acid present. Sulphurous acid is then added until the precipi- tated manganese dioxide is redissolved, after which a slight excess of ammonia is added and then 7-10 c.c. of glacial acetic acid for each 100 c.c. of solution. The solution is boiled for one minute, the precipitate is allowed to settle, and the filtrate is decanted through a filter,* transferred to the filter, washed with 7 per cent, acetic acid and finally with hot water. The dried precipitate is ignited over a Bunsen burner for from fifteen to twenty minutes and then weighed. The precipitate contains manganese and aluminium, so that it is fused with three times as much sodium carbonate. The melt (colored green by the manganese) is leached with cold water, leaving sodium metatitanate f and some alumina undissolved. The precipi- tate is filtered off by means of a small filter, is ignited in a platinum crucible, and fused again with a little sodium carbonate. After cooling, the contents of the crucible are dissolved in 1.9 c.c. of sulphuric acid (1 vol. cone. H 2 S0 4 :1 vol. H 2 0) diluted to about 150-200 c.c. and treated with 5 gm. sodium acetate and one-tenth of its volume of glacial acetic acid. After boiling one minute and allowing to stand until settled, the precipitate is filtered off, washed with 7 per cent, acetic acid, then with water, dried, ignited, and weighed. This precipitate usually contains aluminium, so that it is again fused with sodium carbonate and the melt again treated with sulphuric acid, etc., exactly as described above. This time the precipitate is usually free from aluminium, but the process should be repeated until a constant weight is obtained. This analysis does not require much time, for usually the amount of titanium present is so small that the precipitates filter and wash quickly. For the determination of very small amounts of titanium, it is advisable to use the colorimetric method proposed by Weller (cf. p. 100). Under the analysis of silicates will be discussed a practical example of this determination. * Schleicher & Schiill's filter-paper No. 589 is satisfactory for this purpose. f The sodium metatitanate undergoes hydrolysis and forms a precipitate containing a much higher percentage of TiO 2 . n8 GRAVIMETRIC ANALYSIS. Determination of Titanium in Rutile and Iron Ores. Method of 0. L. Barnebey and R. M. Isham.* This method is based on the volatilization of the silica by hydro- fluoric acid in the presence of sulphuric acid, evaporation to dryness and fusion with sodium carbonate and a little potassium nitrate (which converts the iron and titanium to insoluble ferric oxide and sodium acid titanate) extraction with hot water to remove the soluble phosphates, sulphates and aluminates, solu- tion of the ferric oxide and sodium titanate in hydrochloric acid, extraction of ferric chloride with ether, reduction of slight traces of iron with sulphurous acid, precipitation of the titanic acid by boiling in acetic acid solution, filtration and ignition to titanium oxide (or the titanium may be determined colorimetrically) . The method is accurate and not long. Procedure. The sample is weighed into a platinum crucible, treated with a little water, 5 to 10 drops of sulphuric acid, and 1 c.c. of hydrofluoric acid, and the mixture heated care- fully until finally no more sulphuric acid fumes are evolved. Five or 10 grams of sodium carbonate and a little potassium nitrate are added and the mixture fused at least thirty minutes. The crucible and cover are cooled, placed in a beaker, covered with hot water, and heated until the melt is disintegrated. Ferric oxide and sodium titanate are left insoluble in hot water. The crucible is removed, washed, and any adhering particles of ferric oxide and sodium titanate are dissolved in hot hydrochloric acid (sp. gr. 1.1). This solution is saved. The residue in the beaker is filtered and washed with hot water. The filter is per- forated and the residue carefully washed into a clean beaker with hydrochloric acid (sp. gr. 1.1). (No water is to be added from this stage of the analysis until after the treatment with ether.) The hydrochloric acid washings from the platinum crucible are transferred to the beaker and the whole heated on the hot plate until solution is complete and the total volume reduced to 15 to 20 c.c. The solution is then cooled, and transferred to a separa- tory funnel, the beaker being rinsed with hydrochloric acid * J. Am. Chem. Soc., 32, 957 (1910). DETERMINATION OF TITANIUM IN RUTILE AND IRON ORES. 119 (sp. gr. 1.1). An equal volume ol ether ; which has been saturated with concentrated hydrochloric acid solution, is added to the solution in the funnel, a rubber stopper is inserted in the top, the funnel is inverted, the stop-cock opened, and the whole thoroughly shaken. The stop-cock is then closed, the funnel placed in an upright position and allowed to stand ten minutes, when the aqueous layer is drawn off into a second separatory funnel. The ether is rinsed twice by shaking well with 5 to 10 c.c. portions of hydrochloric acid (sp. gr. 1.1) and the washings are added to the aqueous solution. The treatment with ether is repeated two or three times until the last portion of ether fails to show any greenish tinge due to the presence of dissolved ferric chloride. The aqueous solution containing all the titanium in the pres- ence of little, if any, iron and aluminium, is heated to expel the dissolved ether, 20 c.c. of sulphuric acid (1 : 1) are added, and the solution evaporated until fumes of sulphuric anhydride are evolved. The cooled solution is diluted to about 100 c.c. and nearly neutralized with ammonia. One or two grams of ammo- nium bisulphite are added and the solution heated on the hot plate for half an hour. Ten to 15 grams of ammonium acetate are now added and the solution boiled for fifteen minutes. The precipitated titanic acid is filtered off, washed with 7 per cent, acetic acid, ignited and weighed as Ti0 2 . Separation of Uranium from Iron and Aluminium. The slightly acid solution, containing considerable quantities of ammonium salts, is treated with an excess of ammonium car- bonate and then with ammonium sulphide, allowed to stand for some time in a closed flask, finally filtered and washed with water containing ammonium sulphide. The precipitate contains the iron as sulphide and the aluminium as hydroxide; in the filtrate is found all of the uranium as UO2(CO3)3(NH4)4. The precipitate is dissolved in hydrochloric acid, its solution freed from hydrogen sulphide by boiling, the ferrous salt oxidized to ferric salt by the addition of potassium 120 GRAVIMETRIC ANALYSIS. chlorate, and the iron and aluminium determined by one of the methods described on pages 107-109. The filtrate containing the uranium is evaporated almost to dryness, acidified with hydrochloric acid, boiled, and the uranium precipitated, by the addition of ammonia, as ammonium uranate. The precipitate is filtered off, washed with 2 per cent, ammonium nitrate solution to which a little ammonia has been added, dried, ignited, and weighed as UaOg. The result obtained may be verified by heating the residue repeatedly in a current of hydrogen in a Rose crucible (see Copper Determination) until a constant weight is obtained; weighing as U0 2 . The purity of the precipitate may also be tested volu- metrically (see Volumetric Analysis). B. DIVISION OF THE MONOXIDES. MANGANESE, NICKEL, COBALT, ZINC. MANGANESE, Mn. At. Wt. 54.93. Forms: MnS0 4r MnS, Mn 3 4 , Mn 2 P 2 7 . I. Determination as Manganous Sulphate, MnS0 4 . This method, first proposed by Volhard,* has recently been tested by Gooch and Austin, f and has been found strictly accurate. Experiments performed by Schudel in the author's laboratory completely confirm Gooch's results. Procedure. The oxide obtained by the ignition of the car- bonate, sulphide, or of manganous manganite, is dissolved in as slight an excess of sulphuric acidj as possible in a porcelain crucible, evaporated as far as possible on the water-bath, after which the excess of acid is removed by heating in an air-bath. A porcelain crucible provided with an asbestos ring (see Fig. 1 1 , p. 27) serves for the air-bath. The walls of the smaller crucible should be separated from those of the larger one by about 1 cm. After the sulphuric acid has been removed, the two crucibles * Ann. d. Chem. u. Pharm., 198, p. 328. t Zeit. f. anorg. Chem. (1898), 17, p. 264. J The manganous manganite (Mn 3 O 4 ) requires the presence of reducing agent (best SO 2 ) or of pure hydrogen peroxide. SEPARATION OF MANGANESE AS CARBONATE. 121 are covered and heated to redness over a good Bunsen burner, allowed to cool in a desiccator and weighed. From the weight of the manganous sulphate, the amount of manganese present may be calculated as follows: MnSO 4 :Mn=p:s Mn (a) Separation of Manganese as Carbonate. This method for the separation of the manganese permits only a limited application, because no other metal that is precipi- tated by alkali carbonates can be simultaneously present. The method, therefore, is only suitable for the determination of man- ganese in solutions of pure manganese salts containing nothing else except alkali and ammonium salts. According to H. Tarnm, * the precipitation is best accomplished by means of ammonium carbonate. For this purpose the neutral solution (which may contain other ammonium salts) is treated with a slight excess of ammonium carbonate, warmed gently, and the beaker containing the solution is allowed to remain in a luke- warm water-bath until the precipitate has settled and the upper liquid has become clear. The precipitate is filtered off, washed with hot water, dried, ignited, and weighed either as sulphate according to 1 or as Mn 3 O 4 according to 2. Remark. If either sodium or potassium carbonate is used to precipitate the manganese, the precipitate will always contain alkali carbonate that cannot be removed by washing. After the precipitate has been ignited, however, the alkali carbonate can be easily extracted by water. Furthermore, the precipitation is not quite quantitative; the filtrate always contains small amounts of manganese. In order to remove this, it is necessary to evapo- rate the aqueous solution to dryness, whereby the manganous car- bonate is completely decomposed hydroly tic ally into carbonic acid and manganous hydroxide, and the latter in contact with the air changes to brown manganic oxide, Mn 2 O 3 . The residue obtained after the evaporation is treated with water, the small amount t Ch. News, 26 (1872), p. 37, and Zeit. f. anal. Chem., 11 (1872), p. 425. 122 GRAVIMETRIC ANALYSIS. of brown manganese compound filtered off, ignited, and added to the main part of the precipitate. (6) Separation of Manganese as Sulphide. This method is employed when it is necessary to separate man- ganese from calcium, strontium, barium, and magnesium. We will distinguish between two different cases : (a) The solution contains, besides manganese, large amounts of the alkaline earths or magnesium. (/?) The solution contains only small amounts of the alkaline earths or magnesium. (a) In case large amounts of the alkaline earths or magnesium are present, the manganese sulphide must be precipitated in the cold in the presence of considerable ammonium salts. The solution is placed in an Erlenmeyer flask of Jena glass and about 5 gm. of ammonium chloride or ammonium nitrate are added. In case the solution reacts acid, ammonia is added until it is slightly alkaline, and a slight excess of freshly-prepared, colorless ammo- nium sulphide solution is added. The flask is now nearly filled with cold distilled water that has been boiled, corked, and allowed to stand twenty-four hours, or, better, still longer. After this time the flesh-colored precipitate will have settled. The clear upper liquid is carefully decanted through a filter,* taking pains not to disturb the precipitate and to keep the filter filled with liquid all the time. If the precipitate is at all bulky, it is washed three times by decan- tation with a 5 per cent, solution of ammonium nitrate to which has been added 1 c.c. of ammonium sulphide. The precipitate is then transferred to the filter and washed with dilute ammonium sul- phide water until 20 drops of the filtrate evaporated to dryness on a crucible-cover leave no residue. Now for the first time the filter is allowed to drain completely and is dried. As much of the pre- cipitate as possible is transferred to a small thin-walled porcelain crucible, the filter-paper is burned in a platinum spiral, and the ash added to the main portion of the precipitate in the crucible. The uncovered crucible is heated over a small flame until the greater part of the sulphur has been burned off, when the flame is increased * Schleicher & Schiill's filter-paper No. 590 can be used to advantage. SEPARATION OF MANGANESE AS MANGANESE DIOXIDE. 123 and the crucible is finally heated over the flame of a Teclu burner, cooled, and weighed as Mn 3 O 4 (cf. p. 125, sub. 3). The heating is repeated until a constant weight is obtained. Manganous sulphide is readily changed to Mn 3 O 4 if the amount of sulphide is compara- tively small. In case more than 0.2 gm. is present there is danger of getting a too high result on account of some manganous sulphate not being decomposed. In this case it is advisable to dissolve the washed precipitate of manganous sulphide in dilute hydrochloric acid, to evaporate the solution to dryness in order to remove all hydrogen sulphide, to dissolve the residue in a little water and to precipitate the manganese as carbonate according to 1; or the manganous sulphide can be weighed as such. (See p. 125.) (5) In case only small amounts of alkaline earths are present, the following procedure can be used : The neutral solution is heated to boiling, an excess of ammonia and some ammonium sulphide is added and the boiling is continued until the manganous sulphide has become a dirty green. The precipitate is allowed to settle for some minutes and is then filtered and washed with water contain- ing a little ammonium sulphide. From this point the procedure is the same as described under (a). (c) Separation of Manganese as Manganese Dioxide. If a dilute solution of a manganous salt is treated with bromine water and boiled, the reaction MnCl 2 + Br 2 + 2H 2 O<=MnO 2 + 2HC1 + 2HBr does not take place unless the halogen acids are neutralized as fast as they are formed. This neutralization can be accom- plished by means of the salt of a weak acid, such as sodium acetate, even when the solution contains free acetic acid, which is scarcely ionized at all in the presence of its alkali salt. Thus in a solution such as is obtained after the removal of iron and aluminium by a basic acetate separation (cf. p. 152), the man- ganese can be precipitated quantitatively by boiling with an excess of bromine water. The oxide does not correspond exactly to MnO 2 , although most of the manganese is in the quadrivalent 124 GRAVIMETRIC ANALYSIS. condition.* When the precipitate has collected together in large flocks, the boiling is discontinued and the precipitate allowed to settle; it is filtered and washed with hot water. Some chemists ignite this precipitate and weigh as Mn 3 O 4 but it is more accurate to dissolve the precipitate in a mixture of HC1 and H 2 SO 3 and to precipitate the manganese finally as manganese ammonium phosphate. (See 4, p. 126.) Chlorine, hydrogen peroxide, l^pochlorites, etc., may be used instead of bromine, but these reagents have no especial advan- tages. When the solution of the manganous salt contains ammonium salts, the precipitation of the manganese does not take place by the above procedure, because the sodium acetate serves rather to neutralize the acid set free by the following reaction: Upon the addition of ammonia, however, the precipitation of the manganese can be effected. In this case, it seems fair to assume that the reaction goes through the following stages: MnCl 2 + 2NH 4 OH<=>Mn (OH) 2 + 2NH 4 C1, Mn (OH) 2 + Br + NH 4 OH = MnO (OH) 2 + NH 4 Br . The precipitation with bromine and ammonia is not so satis- factory as with bromine alone in the presence of acetic acid and sodium acetate and in the absence of ammonia or ammonium salt, because when ammonia is present much of the bromine is used up in oxidizing the ammonia or ammonium salt. In that case there is considerable solution of nitrogen, and, moreover, when an excess of bromine is added the solution may become acid enough to dissolve the precipitated manganese: 2NH 3 + 3Br 2 = 6HBr + N 2 . * The MnO 2 acts as the anhydride of metamanganous acid, H 2 MnO 3 , and some manganous manganite, MnMnO 3 or Mn 2 O 3 , is contained in the pre- cipitate. DETERMINATION OF MANGANESE. 125 It is necessary, therefore, when ammonium salts are present to make sure that the solution is ammoniacal at the end of the operation. This method of precipitating manganese from solutions pos- sesses disadvantages which make ifc useless in many cases. If, besides manganese, the solution contains calcium, zinc, etc., man- ganites of these metals are precipitated with the manganese. In this case the precipitate must be dissolved in hydrochloric acid and the precipitation repeated several times, but even then it is not possible to obtain a precipitate altogether free from these metals. If the other metals are present only in small amounts, the results obtained by this method are sufficiently accurate. The precipitation of manganese as sulphide in the presence of other metals is always satisfactory and should be used in almost all cases. 2. Determination of Manganese as Sulphide. If the manganese has been precipitated, as described on p. 122 as sulphide, the precipitate is separated from the filter as completely as possible, placed in a Rose crucible (of unglazed porcelain), the filter is burned in a platinum spiral, and the ash added to the main portion of the precipitate. Some pure sulphur which has been crystallized from CS 2 is added, after which the crucible and its contents are heated in a current of hydrogen by means of a Bunsen burner exactly as described under the Deter- mination of Copper as Sulphide. After the excess of sulphur has distilled off and been burned, the crucible is cooled in a stream of hydrogen and the precipitate is weighed as MnS. 3. Determination of Manganese as Mn 3 4 . Inasmuch as all the oxides of manganese, as well as those compounds which are converted into oxide on ignition (manga- nous salts of volatile organic and inorganic acids, with the ex- ception of the halogen salts), are converted into Mn 3 O 4 * on being ignited in the air, at temperatures between 940 and 1100, it * Cf. R. J. Meyer and K. Retgers, Z. anorg. Chem., 57, 104 (1908), at 530 the oxides of manganese are slowly but quantitatively changed into Mn 2 O 3 . 126 GRAVIMETRIC ANALYSIS. follows that this method for the determination of manganese is quite generally applicable. It is nearly as accurate as the methods described under 1 and 2, if the ignition of the precipitate takes place in an electric furnace at about 1000, but very good results are obtained if, as recommended by Gooch, * the porcelain crucible (containing the carbonate, manganous manganite, or sulphide) is entirely surrounded by the oxidation flame of a Teclu burner, whereby a moderately high heat is obtained with- out too much free access of air.f After the ignition, the crucible and its contents are cooled in a desiccator and then weighed. From the weight p of the oxide, the amount of manganese can be calculated according to the equation Mn 3 4 :3Mn = p::c 3Mn 4. Determination of Manganese as Manganese Pyrophosphate, Mn 2 P 2 7 . This excellent method was recommended by W. Gibbs J and subsequently studied by Gooch and Austin. The slightly acid solution, containing an amount of manganese corresponding to not over 0.5 gm. Mn 2 P20 7 , and no other metals except alkalies, is treated with 20 gm. ammonium chloride, 5 to 10 c.c. of a cold saturated solution of sodium phosphate, and ammonia, drop by drop, until a slight excess is present. The solution is heated * Zeit. f. anorg. Chem., XVII (1898), p. 268. f To illustrate the accuracy of the three methods just described for the determination of manganese, the following results obtained by H. Weitnauer are given. He obtained after making six determinations by each method the following mean values: 50 c.c. of a pure manganese sulphate solution treated with ammonium carbonate and changing the precipitate to sulphate gave 0.1025 gm. Mn; by precipitating as sulphide and weighing as such, 0.1027 gm. Mn; and by changing the sulphate to Mn 3 O 4 , 0.1029 gm. Mn. J Am. J. Science, 46, 216; Z. anal. Chem., 7, 101 (1868). Z. anorg. Chem., 18, 339 (1S98). CO LORI METRIC DETERMINATION OF MANGANESE. 127 to boiling and kept at this temperature for three or four minutes, or until the precipitate assumes a silky, crystalline appearance. After cooling, the precipitate is filtered through a Gooch or Munroe crucible, washed with cold ammonium nitrate solution, dried, and ignited within a larger crucible or in an electric furnace. The results are excellent. Manganese can be determined very accurately by volumetric methods (see Volumetric Analysis). 5. Colorimetric Determination of Manganese. Small amounts of manganese may be accurately and quickly determined by the colorimetric method. This is chiefly used for the estimation of the manganese present in iron arid steel. If more than 1.5 per cent, of manganese is present, the results are unreliable. The method depends upon the oxidation of the manganese to permanganic acid, bringing the solution to a definite volume and comparing its color with another solution containing a known amount of manganese. If the solutions are colored exactly the same shade, then the amounts of manganese which they contain are the same. Procedure. A standard solution of potassium permanganate is first prepared by dissolving 0.072 gm. of the crystallized salt in 500 c.c. of water; 1 c.c. of this solution contains 0.05 mgm. of manganese. Exactly 0.2 gm. of the iron or steel is dissolved in 15-20 c.c. of nitric acid (sp. gr. 1.2) in a 100-c.c. measuring-flask. The acid is heated to boiling to effect complete solution, after which the solution is allowed to cool and diluted up to the mark with water. After thoroughly mixing, 10 c.c. of the liquid are brought by means of a pipette into a small beaker, 2 c.c. of nitric acid (sp. gr. 1.2) are added, and the liquid is heated until it begins to boil, when the flame is removed, 0.5 gm. of lead peroxide is added, the mixture is shaken and then heated for two minutes to boiling. After standing some time, the warm, violet-colored solution is -filtered through a small asbestos filter * into a glass-stoppered test-tube * The asbestos must have been previously ignited, treated with KMnO< solution, and finally washed with water. 128 GRAVIMETRIC ANALYSIS. about 20 cm. high, and graduated in cubic centimeters. The filter is washed with as little water as possible, the tube is stoppered and shaken until the solution is thoroughly mixed. Into a second tube of the same size, and also graduated in cubic centi- meters, is placed 1-5 c.c. of the standard manganese solution, and this is carefully diluted with water until the two liquids have exactly the same shade when viewed horizontally. The height of the liquid in each tube is then carefully read. Assuming that 1 c.c. of the standard solution were placed in the cylinder and diluted to T c.c. in order to obtain the same shade produced by t c.c. of the other solution, then as the con- centrations of the two liquids are directly proportional to their heights, _ -0.05 mgm. ~~ This amount of manganese is contained in 0.02 gm. of the iron, so that the percentage of manganese present is x = ~ = er cent. Mn. Rather more accurate results are obtained if, instead of using a standard solution obtained from potassium permanganate, a sample of steel is used containing a known amount of manganese and treated in exactly the same way as the sample to be analyzed, a fresh standard being prepared for each analysis. An even better colorimetric method has been devised by M. Marshall * and H. E. Walters, f Although manganese is precipitated as manganous acid, from solutions slightly acid with nitric or sulphuric acid, by the addition of alkali persulphates, the oxidation goes farther and * Chem. News, 83, 76 (1904). f Ibid., 84, 239 (1904). DETERMINATION OF NICKEL AS NICKEL GLYOXIME. 129 permanganic acid is formed within a short time if a catalytic agent is present, ench as silver nitrate. 2Mn(N0 3 ) 2 + 5 (NH 4 ) 2 S 2 O 8 + 8H 2 O = 5 (NH 4 ) 2 S0 4 + 5H 2 S0 4 + 4HNO 3 + 2HMn0 4 . Procedure. 0.2 gm. of steel is placed in a 100-c.c. flask and dissolved in 20 c.c. of cold nitric acid (sp. gr. 1.2). Thereupon 10 c.c. of silver nitrate solution (1.38 gm. AgNOa in a liter of water), are added, the solution made up to exactly 100 c.c. and mixed. Of this solution 10 c.c. are placed in glass-stoppered, graduated test-tube. After adding 2.5 c.c. of ammonium per- sulphate solution (200 gm. in a liter of water) the test-tube is placed in water at 80 to 90 and allowed to remain there until the bubbles of gas arising become more numerous and remain at the top for a few seconds. The solution is then cooled by placing the tube in cold water, and the color is compared with a standard solution containing a known amount of permanganic acid.* NICKEL, Ni. At. Wt. 58.68. Forms: Nickel Dimethyl Glyoxime, NiC 8 H 14 N 4 4 ; Nickel, Ni; and Nickel Oxide, NiO. i. Determination as Nickel Glyoxime, Ni(C 4 H 7 N 2 2 ) 2 . Dimethyl glyoxime, CH 3 - CNOH - CNOH - CH 3 , was recom- mended by L. Tschugaeff f as a reagent for nickel and used by K. Kraut \ for detecting the presence of traces of nickel in ashes. O. Brunck and others have also studied the reaction and found it to furnish a most rapid and accurate method for the quanti- tative estimation of nickel either by itself or in the presence of cobalt, zinc and manganese. If the solution contains tartaric acid enough to prevent the precipitation of iron by ammonia, the * Or better a solution of a steel of known manganese content. t Z. anorg. Chem., 46, 144 (1905); Ber., 38, 2520 (1905). t Z. angew. Chem., 19, 1793 (1906); ibid., 20, 1844 (1907). Ibid., 20, 834 (1907). 130 GRAVIMETRIC ANALYSIS. nickel in a sample of nickel steel can be determined accurately within two hours and without the removal of any other metal. When a dilute, neutral solution of a nickel salt is treated with an alcoholic solution of dimethyl glyoxime, a red, crystalline precipitate of nickel dimethyl glyoxime is formed. Dimethyl glyoxime. Nickel dimethyl glyoxime. . 9 upi} The salt is soluble in mineral acids so that precipitation is incomplete because of the acid set free in the reaction. It be- comes quantitative, however, if the mineral acid is neutralized by ammonia or if sodium acetate is added, whereby the mineral acid is replaced by acetic acid in which the precipitate is prac- tically insoluble. Large quantities of ammonium salts or of alkali acetate do no harm, but an excess of ammonia tends to prevent the formation of the precipitate. The precipitate is distinctly soluble in absolute alcohol, but only traces dissolve in 50 per cent, alcohol, and in more dilute alcohol it is even less soluble. When thrown down in the cold or in the presence of much free ammonia the precipitate is very voluminous and hard to filter. Procedure. The neutral or slightly acid* solution is diluted so that not more than 0.1 gm. of cobalt is present in 100 c.c., heated nearly to boiling and treated with somewhat more than the theoretical amount of an alcoholic 1 per cent, solution of dimethyl glyoxime. f Ammonia is then cautiously added until the solution smells slightly. While still hot, the precipitate is filtered into a Gooch or Munroe crucible, washed with hot water and dried at 110 to 120 for 45 minutes. It contains 20.31 per cent. Ni. The nickel salt of dimethyl glyoxime is red and crystalline. It contains no water of crystallization and sublimes at 250 without decomposition. * If strongly acid, the solution is nearly neutralized with caustic potash, then the reagent is added, etc. f The volume of the alcoholic solution should in no case be more than half that of the nickel solution, as the precipitate is soluble in alcohoL DETERMINATION OF NICKEL AS METAL BY ELECTROLYSIS. 131 2. Determination of Nickel as Metal by Electrolysis. From strongly acid solutions nickel is not deposited upon stationary electrodes by a current of 1-3 amperes. From slightly acid solutions the deposition is not quantitative. From ammoniacal solutions nickel is readily deposited, and on account of its simplicity and accuracy this method is to be strongly recommended for the determination of nickel. (a) Method of Gibbs. * Nickel sulphate or chloride (but not the nitrate) is dissolved in an ammoniacal solution of ammonium sulphate and electrolyzed. The nickel is deposited upon a weighed cathode and, at the end of the electrolysis, the gain in weight represents the quantity of nickel. Requirements and Procedure. For this, as well as for all other electrolytic determinations, the apparatus shown in Fig. 31 may be used. B represents a storage-battery, which is provided with the binding posts MM. The current is led first through the variable resistance W, then through the known resistance W (a resistance of 1 ohm is most convenient to use f)> and from here to the decomposition cell finally back to the battery. If at any time it is desired to measure the voltage between the electrodes of the cell, the voltmeter V is connected with the binding posts of the decomposition cell, by. throwing the switch Q,l Fig. 32, so that c is connected with b and c' with b f . The *Z. anal. Chem., 3, 334 (1864). Cf. Fresenius and Bergmann, Z. anal. Chem., 19, 320 (1880). f The resistance w'=lQ can be made from nickelin wire. The resistance of the wire is measured with the aid of the Wheatstone bridge and a length cut off corresponding to one ohm. This wire is wound round a wooden block and the ends fastened to binding posts. J If a commutating switch is not available, one can be prepared by taking the cover of a pasteboard box, about 2 cm. deep, filling it with melted paraffin, and then, after cooling, making little cavities at aa'bb'cc' by pressing a test-tube, which is filled with hot water, against the cold wax. These cavities are filled with mercury and the switch finished with copper wire. Cf. Fig. 36, p. 178. 132 GRAVIMETRIC ANALYSIS. M M W w I k -oe\ Q QC' FIG. 31. DETERMINATION OF NICKEL /IS METAL BY ELECTROLYSIS. 133 strength of the current, on the other hand, is obtained by placing the switch in the opposite position with a united to 6 and a' to &', as shown in Figs. 31, 32. FIG. 32. Since, according to Ohm's law, the Electromotive force btrength ot current = - =r : - , Resistance then if the strength of the current is expressed in amperes, the electromotive force in volts, and the resistance in ohms, we have In case W = 1 ohm, then A-E, and the voltmeter will show directly the strength of the current (amperes) . * It is arranged so that the current may be taken from different points along MM, and it is thus possible to carry out a number of electrolytic determinations at the same time, and the volt- meter V serves as measuring instrument for all the analyses that are hi progress. By means of the SS it is possible to connect easily the voltmeter with the different cells. While a measure- * With weaker currents, the known resistance can be made W'= 10j2 so that the voltmeter will show ten times the actual current. 134 GRA VIME TRIG ANAL YSIS. ment is being made at any cell, all other switches must be cut out of circuit. The decomposition cell consists of a glass beaker in which is placed as cathode a wire gauze electrode (first recommended by Cl. Winkler) and as anode a platinum spiral. The electrodes o'l FIG. 33. must always reach to the bottom of the beaker and the top of the gauze electrode should be nearly covered by solution. In some cases it is desirable to use a platinum dish as cathode, as recommended by Classen. (See Fig. 36, p. 178.) The electrodes are usually connected with two electrode stands on which metal arms are attached to an upright glass rod (Fig. 33). To prevent serious loss of electrolyte by spattering, DETERMINATION OF NICKEL AS METAL BY ELECTROLYSIS. 135 the beaker is covered with two halves of a watch-glass. This is not entirely satisfactory, as when much gas is evolved a little of the liquid is still carried off mechanically. This method of fastening the electrodes, moreover, has the disadvantage that when the electrolysis is carried out in a hot solution, acid or ammoniacal vapors, as the case may be, condense on the brass FIG. 34. arms of the electrode support and in some cases the liquids thus condensed dissolve some brass and the resulting solution may drop into the beaker, and spoil the analysis. To prevent this misfortune, the author bends the ends of the electrodes to a right angle and connects them with an electrode stand designed as shown in Fig. 34. This electrode holder consists of two brass rods insulated from one another by means of an intervening layer of mica, and the rods are fastened to the ring r through a piece of ebonite, e. 136 GRAVIMETRIC ANA LYSIS. The openings to hold the wires are cut wedge-shaped, so that any shape of wire can be inserted. Since the ends of the electrodes leave the beaker in a horizontal direction, the beaker can be covered tightly by means of a whole watch-glass, and not only are losses by spattering avoided, but there is absolutely no danger of contamination from the outside. The Electrolysis of Nickel. For every 0.25-0.30 gm. nickel, present as sulphate or chloride, b>ut not as nitrate, * 5-10 gm. of ammonium sulphate and 30-40 c.c. of concentrated ammonia are added, and the solution diluted with distilled water to a volume of 150 c.c. This solution is electrolyzed at the room temperature with a current of 0.5-1.5 amperes and an electrode potential of 2.8-3.3 volts. The electrolysis is finished after three hours, as can be shown fairly satisfactorily by adding a little water and allowing the current to pass through the solution for fifteen or twenty minutes longer. If at the end of this time no nickel has deposited upon the electrode surface which was wet for the first time by the last dilution, the determination is finished. If the solution is kept at a temperature of from 50-60 C. only about one hour is necessary for the deposition. The deposited metal adheres firmly to the electrode, is bright, and possesses almost the color of platinum. As soon as the electrolysis is finished, the watch-glass is removed, the electrode holder is raised so that only the bottoms of the electrodes remain in the liquid, and the upper parts of the electrodes are washed thoroughly with water from a wash- bottle. The electrodes are then raised entirely out of the solution and the bottoms washed immediately with water. The current is then turned off and the cathode rinsed with absolute alcohol, after which it is dried by holding it high above a gas flame. After cooling in a desiccator, it is weighed. To clean the cathode, place it in a small beaker, add enough nitric acid (1:1) to wet all the nickel, and heat for at least fifteen minutes. This treatment is absolutely necessary to remove the last traces of nickel. If this is not done, the electrode on being * Page 131. THE ELECTROLYSIS OF NICKEL. 137 ignited becomes discolored, and it is then very difficult to clean the electrode by repeated treatment with acid followed by ignition. The discolored electrode, however, can be used for another electrolytic determination. To make sure that all the nickel has been deposited from the electrolyzed solution, the ammonia, is almost wholly neutralized with hydrochloric acid and a few cubic centimeters are added of 1 per cent, solution of dimethyl glyoxime in alcohol. When less than a tenth of a milligram of nickel is present, it will take several minutes for a yellow coloration to appear, and soon afterward the red crystals of nickel salt will be precipitated. The nickel not deposited by an electrolysis may be estimated accurately by shaking the solution thoroughty and comparing the color produced by the addition of dimethyl glyoxime with that produced with a dilute nickel solution containing a known quantity of nickel. Naturally such a colorimetric test can be used only with very small quantities of nickel. Remark. The electrolysis of nickel from an ammoniacal solution should not be continued for too long a time, because the cathode slowly gains in weight even after all the nickel has been deposited from the solution. The anode is attacked, causing platinum to go into solution, which is deposited upon the cathode, partially, at least. The presence of too little ammonia often results in the forma- tion of black Ni(OH) 3 at the anode; the analysis then comes out too low. Classen's method for depositing nickel from a solution of ammonium oxalate apparently gives too high results * and cannot be recommended. 3. Determination as Nickelous Oxide. The nickel solution is heated in a porcelain dish with bromine water and an excess of pure potassium hydroxide, whereby the nickel is precipitated as brownish-black nickelic hydroxide, Ni(OH) 3 . The precipitate is filtered off, washed by decantation * A. Windelschmidt, Dissertation, Miinster, 1907. W. D. Treadwell, Dis- sertation, Zurich, 1909. 138 GRAVIMETRIC ANALYSIS. with hot water, dried, and, after burning the filter, ignited and weighed as NiO. The grayish-green oxide thus obtained always contains small quantities of silicic acid and alkali,* whereby the results are too high. By treating the ignited mass with hot water, the greater part of the alkali can be removed. Drying and again igniting gives the weight of NiO + Si0 2 . The oxide is treated in a porcelain crucible with hydrochloric acid, evaporated completely to dryness, the dry residue moistened with concentrated hydrochloric acid and then with hot water, filtered through a small filter, washed with hot water, and the filter together with the residue ignited wet in a platinum crucible. The weight of this silica, SiO 2 , subtracted from the former weight of NiO + SiO 2 , gives good results. Remark. It is possible to precipitate nickel quantitatively as Ni(OH) 2 by means of caustic potash alone and the precipitate is changed to NiO by ignition. This method is open to the same objections as the above and, furthermore, Ni(OH) 2 is not so easily filtered and washed as Ni(OH) 3 . These two methods are more tedious to carry out and the results are not as accurate as in the case of the first two methods described and will probably not be used much in the future. Besides the methods described, it has been proposed to pre- cipitate nickel as the sulphide, and weigh it as the oxide by ignition in air.f The method is good but hardly comparable with the dimethyl glyoxime method, the electrolytic method, or the volumetric titration with potassium cyanide. COBALT, Co. At. Wt. 58.97. Forms: Co, CoS0 4 . i. Determination as Metal. (a) By Electrolysis. The most accurate method for the estimation of cobalt is by electrolysis and the details of the process are precisely the same as have been given under nickel, i. e., from a strongly ammoniacal * Cf. A. Windelschmidt, loc. ciL and W. D. Treadwell, loc. cit. f H. Cormimboef, Ann. chim. appl., II, 6 (1906). Cf. A. Windelschmidt, loc. cit. DETERMINATION OF COBALT AS METAL. 139 solution containing ammonium salts and the sulphate or chloride (preferably the former), of cobalt. It is customary to use a little more ammonia than in the determination of nickel, because cobalt has a greater tendency to deposit as black Co(OH) 3 at the anode. The duration of the electrolysis is the same as with nickel, rather than somewhat longer. At the end of the deter- mination, after the electrodes have been removed, the entire solution is tested for cobalt by adding ammonium sulphide or potassium sulphocarbonate. (6) By Reduction of the Oxide in a Stream of Hydrogen. The cobalt solution is heated to boiling in a porcelain evaporat- ing-dish, and the cobalt is precipitated as black cobaltic hydroxide by the addition of caustic potash and bromine water. The pre- cipitate is filtered off,* dried, and ignited. After cooling it is treated with water in order to remove the small amount of alkali which is always present, and then the residue is ignited hi a stream of hydrogen and weighed as metal. After weighing, the metal is dissolved in hydrochloric acid, evaporated to dryness, the dry mass moistened with hydrochloric acid, then treated with water, and the small residue of silicic acid filtered off. This resi- due is ignited and its weight subtracted from that obtained after the ignition in hydrogen. Cobalt may also be precipitated as cobaltous hydroxide by caustic potash alone, but the resulting precipitate is not so easy to filter and wash as the cobaltic hydroxide. The precipitation by means of sodium carbonate is not so satisfactory. The oxides of cobalt when ignited in air yield a mixture of CoO and Co 3 O 4 in varying proportions, so that they are not suited for the quantitative determination of cobalt. Remark. The results obtained by this method are usually a little higher than by electrolysis. * Cobaltic hydroxide, unlike nickelic hydroxide, has the tendency of giving a turbid filtrate on washing. If, however, Schleicher & Schiill's filter-paper No. 589 (blue band) is used, none of the precipitate passes through. 140 GRAVIMETRIC ANALYSIS. 2. Determination as Sulphate. The method is the same as was described under Manganese (p. 104). ZINC, Zn. At. Wt. 65.37. Forms: ZnNH 4 P0 4 , Zn 2 P 2 7 , ZnO, ZnS, Zn. i. Determination as Zinc Ammonium Phosphate or Zinc Pyrophosphate. This excellent method, first recommended by H. Tamm, * has been studied and improved by G. Losekann and T. Meyer,f M. Austin,! and especially H. D. Dakin. Procedure. The cold acid|| solution of the zinc salt is treated with ammonia, in a platinum or porcelain dish, until it is left barely acid. Care is necessary at this point, as zinc ammonium phosphate is soluble both in acids and ammonia. It is then diluted with water to a volume of 150 c.c. and heated on the water- bath. To the hot solution, ten times as much ammonium phos- phate is added as there is zinc present. (If the diammonium phosphate contains some monoammonium phosphate, the salt should be dissolved in cold water and dilute ammonia added until the solution just becomes pink with phenolphthalein.) The precipitate that first forms is amorphous, but it soon changes into fine crystals of zinc ammonium phosphate. The transformation takes place more rapidly in proportion to the quantity of ammo- nium salts present. After the heating has continued for about fifteen minutes, the dish is removed from the water-bath and after being allowed to settle for a short time the precipitate is filtered through a Gooch or Munroe crucible, washed with hot, 1 per cent, ammonium phosphate solution If until free from chlorides * Chem. News, 24, 148. t Chem. Ztg., 1886, 729. % Am. J. Sci., 1899; Z. anorg. Chem., 22, 212 (1900). Z. anal. Chem., 39, 273 (1900). 1 1 If the solution is neutral, 2 or 3 gms. of ammonium chloride are added and then the analysis carried out. If According to Voigt, Z. angew. Chem., 1909, 2282, the precipitate is washed immediately with hot water. DETERMINATION OF ZINC. 141 etc., then twice with cold water, then with 50 per cent, alcohol, dried at 110-120 for an hour and weighed as ZnNH 4 PO 4 , which contains 36.64 per cent. Zn. Or, the precipitate may be weighed as the pyrophosphate, Zn 2 P 2 C>7, in which case the dried zinc ammonium phosphate is heated very slowly in an electric oven to 900-1000. If such an oven is not at hand, the Gooch or Munroe crucible is placed in a larger platinum crucible and heated over the gas flame. The temperature is gradually raised until finally the full heat of the Teclu burner or of the blast lamp is reached. The crucible is heated until its weight is constant. Zn 2 P 2 07 contains 42.90 per cent. Zn. The determination as pyrophosphate is to be recommended when the zinc solution contains a very large quantity of ammonium salts because it requires long washing to remove these and this renders the results a little low. When the precipitate is weighed as pyrophosphate, the ammonium salts are volatilized and it is not necessary to remove them by washing. Remark. In some cases, as when magnesium or aluminium is present, the procedure of K. Voigt is followed. The solu- tion of the zinc salt, containing ammonium salts as well, is treated with an excess of ammonia and then with ammonium phos- phate. After standing some time, the precipitate of magnesium ammonium phosphate and aluminium phosphate is filtered off, the zinc ammonium phosphate being soluble in ammonia. The filtrate is received in a platinum or porcelain dish and is heated on the water-bath until all the free ammonia has been expelled, whereby zinc ammonium phosphate separates out quantitatively in the form of the crystalline precipitate. It is treated as described above. If some of the precipitate should adhere firmly to the sides of the dish, it may be dissolved in a few drops of hydrochloric acid, the solution immediately neutralized with ammonia, and heated a few minutes on the water-bath before filtering. 142 GRAVIMETRIC ANALYSIS. 2. Determination as Zinc Oxide. The carbonate, nitrate, acetate, and oxalate of zinc are readily and quantitatively changed to zinc oxide by ignition in the air; in the case of the sulphate, when present in relatively large amounts, the transformation into oxide is difficult. Small amounts of the sulphate may be changed to oxide by igniting over the blast-lamp. It is advisable, however, in case the zinc is present as sulphate, to precipitate it from the aqueous solution as sulphide and weigh it as such according to 3; or to dissolve the sulphide on the filter in dilute hydrochloric acid, receiving the solution in a weighed plati- num dish, evaporating to dryness on the water-bath, and changing to oxide by the method of Volhard as described below, and Weigh- ing as such. The chloride is readily changed to oxide, according to Volhard, by gentle ignition with pure mercuric oxide. The process is as follows: The neutral solution of the chloride, contained in a platinum dish, is treated with a large excess of pure yellow mer- curic oxide *, suspended in water, and evaporated to dryness on the water-bath, whereby mercuric chloride and zinc oxide are formed, both of which are white substances. Enough mercuric oxide should be used so that the residue obtained after the evaporation is noticeably yellow. The dry mass is ignited under a hood with a good draft (on account of the mercury vapors being poisonous), at first gently and finally strongly, and the residue of zinc oxide is weighed, both mercuric chloride and oxide being volatile. The results are excellent. * The mercuric oxide is prepared by precipitating a solution of mercuric chloride with pure caustic potash. The precipitate is allowed to settle, washed by decantation with water until free from chloride, and kept sus- pended in water in a bottle with a wide neck. A considerable amount of the mercuric oxide, say 5-10 gm., should leave no weighable residue after ignition. DETERMINATION OF ZINC AS SULPHIDE. 143 If the solution contains, besides zinc, also alkalies, the zinc can be precipitated as carbonate and changed to oxide upon ignition. The precipitation of the zinc carbonate should take place in a porcelain dish and the sodium carbonate solution should be added drop by drop to the cold, barely acid solution free from ammonium salts. The sodium carbonate is added until the zinc solution becomes turbid, when it is heated to boiling, whereby the greater part of the zinc is precipitated as granular zinc carbonate. Two drops of phenolphthalein solution are then added and enough sodium carbonate solution to impart a distinct pink color. In this way a precipitate of zinc carbonate is obtained free from alkali, which is not the case if the hot solution is at once precipitated by the addition of an excess of sodium carbonate.* The precipitate is filtered from the hot solution and washed with hot water until 20 drops of the filtrate leave no residue on evaporation. The pre- cipitate is dried, the greater part transferred to a weighed porce- lain crucible, the filter burned by itself in a platinum spiral, and the ash added to the main part of the precipitate in the crucible, which is ignited, at first gently and finally strongly, over a Teclu burner and weighed f after cooling in a desiccator. 3. Determination as Sulphide. This determination is chosen when the zinc is present in a solution containing ammonium salts, or when it is necessary to separate zinc from alkaline earths, alkalies or metals of this group. Zinc sulphide may be precipitated from ammoniacal solutions, cr from solutions containing free acetic, formic, citric, or sulphocyanic acids. *In case considerable amounts of ammonium salts are present there may be no precipitation. Sodium carbonate should then be added until the solution is slightly alkaline and the solution boiled until all the ammonia is expelled. f If the solution contains sulphate, the precipitate produced by sodium carbonate always contains more or less basic zinc sulphate, which may easily lead to high results. In the presence of sulphates, therefore, it is advisable to precipitate the 2inc as sulphide and determine it as such accord- ing to 3. 144 GRAVIMETRIC ANALYSIS. (a) Precipitation ofZnSfrom Ammoniacal Solutions. The slightly acid solution is placed in an Erlenmeyer flask and treated with sodium carbonate solution until a permanent precipitate is obtained. This is dissolved by the addition of a few drops of ammonia, after which for every 100 c.c. of the solution 5 gms. of ammonium acetate (or, better, ammonium thiocyanate) are added, followed by a slight excess of freshly prepared ammo- nium sulphide, the flask is nearly filled with boiled water, stoppered and allowed to stand twelve to twenty-four hours. Without dis- turbing the precipitate, the clear upper liquid is poured through a Schleicher & Schiill's filter No. 590. The precipitate is covered with a solution containing in every 100 c.c. 5 gms. of ammonium ace- tate (or ammonium thiocyanate) and 2 c.c. of ammonium sulphide solution, shaken, allowed to settle, and the turbid upper liquid is poured through the filter, taking care to receive the filtrate in a fresh beaker; in case it comes through turbid it is poured through the filter again. The decantation is repeated three times, after which the precipitate is transferred to the filter and washed com- pletely with the above solution, taking pains to keep the filter full of the wash liquid during the entire operation, finally washing with water containing ammonium sulphide only. The precipitate is then dried, transferred as completely as possible to a weighed Rose crucible, the filter burned by itself and the ash added to the main portion of the precipitate. The precipitate is now mixed with .the aid of a platinum wire, with one-third as much pure sulphur, covered with a layer of sulphur and heated, as described under Manganese (page 125) in a current of hydro- gen. The crucible is finally allowed to cool in the stream of hydrogen and from the weight of the zinc sulphide the weight of zinc present is calculated, Zn ELECTROLYTIC DETERMINATION OF ZINC. 145 and if a is the amount of the original substance, then the per cent, of zinc is 100 Zn (6) Precipitation of ZnS from Acid Solutions. The solution, which has been nearly neutralized with ammonia/ is treated with ammonium chloride or sulphate and a little ammonium or sodium acetate ; and is then saturated with hydro- gen sulphide. After the precipitate has settled completely, the supernatant solution, is poured through a filter, and the pre- cipitate washed with 2 to 4 per cer^t. acetic acid which has been saturated with hydrogen sulphide. When thoroughly washed it is treated as described above. It is to be noted that the zinc sulphide shows less tendency to form colloidal solutions when it is thrown down from a slightly acid solution than when it is precipitated from alkaline solutions. 4. Electrolytic Determination of Zinc. In the presence of acid, zinc is not deposited by an electric current of 0.5 to 1 ampere, although it may be deposited even then by stronger currents. From the solution of potassium or sodium zincate, or from the complex alkali zinc cyanides, it is easy to deposit the zinc quan- titatively. (a) Method of F. Spitzer.* The solution of zinc sulphate (chlorides and nitrates should be absent) is treated with a drop of phenolphthalein and with sodium hydroxide solution until a permanent coloration is obtained. Then 20 to 25 c.c. of normal caustic soda solution are added, the * Z. Elektrochem., 11, 401 (1905). 146 GRAVIMETRIC ANALYSIS. solution is diluted to 150-200 c.c. and electrolyzed, using a plat- inum gauze cathode, with a current of 0.8 to 1 ampere and 3 to 4 volts electrode potential. At the end of three hours the elec- trolysis is finished, provided not more than 0.5 gm. of zinc was present. Without breaking the current, the electrodes are raised nearly out of the bath, the upper portions are washed quickly with water, then the electrodes are taken entirely out of the solution and rinsed with wat^r. The current is then turned off, the cathode washed with absolute alcohol, dried above a flame, cooled in a desiccator, and weighed. When deposited in this way, zinc forms a bluish-gray layer that adheres firmly to the electrode. To make sure that all the zinc was deposited, the electrodes are cleaned and the solution electrolyzed for thirty minutes longer. A slight increase in weight will be obtained in every case because the anode is attacked slightly by the alkaline solution so that the cathode slowly continues to gain in weight from deposited platinum. If at the end of half an hour the gain in weight is not over 0.3 gm. then the deposition of the zinc was complete the first time, as can be shown by testing with sodium sulphide. The results are always a little high.* To clean the electrodes, they are boiled thoroughly with hydrochloric acid (1:2) washed well with distilled water, and ignited. It is not necessary to cover the platinum gauze with a thin coating of copper or of silver, as has been recommended when a platinum dish is used as the cathode. Remark. If too little caustic soda is present, a spongy deposit of zinc is obtained which does not adhere well to the electrode. For this reason the above directions should be followed closely. In the presence of ammonia the determination is not successful. If, therefore, it is desired to analyze a solution containing an ammonium salt, it must be boiled with caustic soda until all the ammonia has been expelled. If, moreover, chlorides or nitrates * Ellwood B. Spear, J. Am. Chem. Soc., 32, 530~(1910). The experiments have been repeated in the author's laboratory by Janini, who obtained as an average from fourteen determinations with 50 c.c. of a zinc sulphate solution, the value 0.1014 gm. Zn instead of 0.1008 gm. Zn/a difference of about 0.6 per cent. SEPARA TION OF MANGANESE, ETC., FROM ALKALINE EARTHS. 147 are present they must be removed by evaporation with sulphuric- acid. The solution is evaporated on the water-bath and finally heated over the free flame until dense vapors of sulphuric acid are expelled. The solution is then diluted and analyzed in the usual manner. (6) The Potassium Cyanide Method. * A drop of phenolphthalein is added to the solution of zinc sulphate, caustic soda solution until a permanent pink coloration is obtained, and then potassium cyanide solution until a clear solution results. This is diluted to a volume of 150-200 c.c. and electrolyzed with a current of 0.5 ampere. At first the electrode potential is about 5.8 volts, but it falls during the analysis on account of the current heating the solution. The electrolysis is finished in two or three hours. Other methods for the electrolytic estimation of zinc are given in A. Classen's book Quantitative Analysis by Electrolysis. SEPARATION OF MANGANESE, NICKEL, COBALT, AND ZINC FROM THE ALKALINE EARTHS. The separation depends upon the insolubility of the sulphides of the metals of this group and the solubility of the sulphides of the alkaline earths. Procedure. The neutral solution of the chlorides, contained in an Erlenmeyer flask, is treated with ammonium chloride (in case it is not already present) and freshly-prepared colorless ammonium sulphide solution is added drop by drop until no further precipitation takes place and the liquid has a distinct odor of ammonium sulphide. The flask is then almost completely filled with boiled water, corked, and allowed to stand twelve hours. The precipitate is filtered and washed as described in tne Deter- mination of Zinc (p. 144). If only a small amount of alkaline-earth metals are present and t Luckow, Z. anal. Chem., 19, 1 (1880). 148 GRAVIMETRIC ANALYSIS. the ammonium sulphide solution is entirely free from ammonium carbonate, the separation is usually complete after one precipita- tion; in the presence of considerable calcium, strontium, barium, or magnesium the sulphide precipitate will always be more or less contaminated with these substances, so that the precipitation must be repeated. For this purpose the washed precipitate is dried, transferred as completely as possible to a porcelain crucible, the filter-paper burned in a platinum spiral and the ash added to the main part of the precipitate in the crucible, which is now covered with a watch-glass, treated with dilute hydrochloric acid, and heated to boiling after the evolution of hydrogen sulphide has ceased, in order to remove all of the hydrogen sulphide. A very little concentrated nitric acid is now added and the mixture warmed until the precipitate is completely dissolved ; the solution is evapo- rated to dryness, treated with a little concentrated hydrochloric acid, and again evaporated to dryness in order to change to chloride any nitrate that may have been formed. The dry mass is moistened with a few drops of concentrated hydrochloric acid, dissolved in hot water, and the slight residue of sulphur filtered off, which, in case barium is present, always contains a small amount of barium sulphate, and is therefore washed with hot water, dried, ignited in a porcelain crucible, and weighed. The filtrate is then precipitated exactly as before by the addition of ammonium sulphide. In case nickel is present, .a too great excess of ammonium sulphide must be carefully avoided, as otherwise the nickel sulphide will pass through the filter (cf. Vol. I). In all cases, however, the filtrate should be tested for nickel by acidifying with acetic acid, heating to boiling, and passing hydrogen sulphide into the solution. If a slight black precipitate is produced by this treatment, it is filtered off and combined with the main precipi- tate (tf. p. 156 et seq). The filtrate containing the alkaline-earth metals is freed from ammonium salts by evaporating to dryness, dissolved in hydrochloric acid, and examined as described on p. 76 et seq. Remark. The ammonium sulphide solution used in the above separation must be free from ammonium carbonate. As, however, all commercial ammonia contains this salt, it must be freed from PREPARATION OF AMMONIA FREE FROM CARBONATE. 149 carbonate before being used for the preparation of ammonium sulphide solution. Preparation of Ammonia Free from Carbonate. * About 10 gms. of freshly slaked lime are added to 500 c.c. of concentrated ammonia contained in a liter flask that is connected with a condenser. The condenser is closed by means of a tube containing soda-lime, and the contents of the flask are allowed to stand for a day with frequent shaking. After this, from 300- 400 c.c. of water are placed in a flask and boiled, meanwhile passing through the water a current of air that has been freed from all traces of carbon dioxido by passing through concentrated caustic potash solution and then through a tower filled with soda-lime. The water is allowed to cool in this air-stream. The flask containing the ammonia is then placed on the water- bath in such a position that the condenser-tube is inclined slightly upward, and this is connected with the delivery-tube, through which the air previously passed into the flask of boiling water. By warming the water-bath the ammonia is now distilled over into the flask containing the boiled water, by which it is completely absorbed. By saturating a part of this ammonia with hydrogen sulphide, a solution of ammonium sulphide is prepared suitable for the above-described separation. SEPARATION OF THE BIVALENT FROM THE OTHER METALS OF THE AMMONIUM SULPHIDE GROUP. This separation is often designated as that of the protoxides from the sesquioxides; this designation is not applicable in the case of titanium and uranyl derivatives. The Barium Carbonate Method. This method depends upon the fact that ferric, aluminium and chromic salts (as well as titanic and uranyl salts) are preci- * The distillation of ammonia also serves to free it from silica, which it always contains when kept in glass bottles for any length of tune. GRAVIMETRIC ANALYSIS. pitated in fche cold by barium carbonate, while manganese, nickel, cobalt, sine, and ferrous salts are not. The trivalent metals are strongly decomposed hydroly tically : ^(HCl) + Fe(OH)Cl 2 . Free acid and a basic salt are. formed by this hydrolysis, the composition of the latter depending upon the quantity of the water and the temperature. If the free acid is removed by the addition of barium carbonate, the equilibrium is disturbed and the hydrolysis goes further until finally the insoluble hy- droxide is formed: Fe(OH)Cl a + 2HOH = 2HC1 + Fe(OH) 3 . The barium carbonate, then, serves only to neutralize the acid set free by the hydrolysis, and the total reaction is expressed by the following equation: .. /// . / // 2FeCl,+ 6HOH + 3BaCO 3 = 3BaCl 2 +2Fe(OH) 3 + The salts of the bivalent metals are not subject to this hy- drolysis in the cold, consequently they are not precipitated by the addition of barium carbonate. On warming, however, they are hydrolyzed to an appreciable extent and are then precipitated by barium carbonate. Procedure. Sodium carbonate solution is added drop by drop to the slightly acid solution of the chlorides or nitrates, but not the sulphates,* of the metals, in an Erlenmeyer flask until a slight, permanent turbidity is produced, which is then redissolved by the addition of a few drops of dilute hydrochloric acid. The solution is diluted and treated with pure barium carbonate f (suspended in water) until after thoroughly shaking an excess of the * Barium carbonate will precipitate the bivalent metals when sulphates are present, e.g.: ZnSO< + BaCO 3 = ZnCO 3 + BaSO 4 . j* The barium carbonate must be free from alkali carbonate. THE BARIUM CARBONATE METHOD. 151 latter remains on the bottom of the flask. The flask is closed and allowed to stand for several hours with frequent shaking. The clear liquid is then decanted off, the residue treated with cold water and again decanted. This decantation is repeated three times, after which the precipitate is transferred to the filter and completely washed with cold water. The precipitate contains all of the iron, aluminium, chromium, titanium, and uranium in the presence of the excess of barium carbonate. The filtrate contains the bivalent metals and barium chloride. The precipitate is dissolved in dilute hydrochloric acid, boiled to remore the carbon dioxide, and the iron, aluminium, chromium (titanium and uranium) are separated from the barium* by double precipitation with ammonium sulphide as described on p. 147. The iron, aluminium, chromium (titanium and uranium) are sepa- rated from one another as described on pp. 107-120. The filtrate from the barium carbonate precipitation is freed from barium by the addition of sulphuric acidf to the boiling solution after it has been made acid with hydrochloric acid. The barium sulphate is filtered off and the monoxides are separated from one another as described on p. 156. Remark. The above separation of the scsquioxides from the protoxides is not absolutely certain in the presence of nickel and cobalt. In this case, particularly when considerable iron is pres- ent, the precipitate produced by barium carbonate contains small amounts of nickel and cobalt. This difficulty can be overcome, however, by adding ammonium chloride to the solution (3-5 gms. for each 100 c.c. of solution) before precipitating with barium carbonate; the separation is then satisfactory. *Most authorities recommend precipitating the barium first with sul- phuric acid and then separating the iron, aluminium, etc. The precipitate of barium sulphate always contains small amounts of the heavy metals, so that the author prefers the above procedure. f Or, better, by double precipitation of the other metals with ammonium sulphide. 252 GRAVIMETRIC ANALYSIS. SEPARATION OF IRON, ALUMINIUM, AND TITANIUM (BUT NOT CHROMIUM AND URANIUM) FROM MANGANESE, NICKEL, COBALT, AND ZINC. Basic Acetate Method. This classic method depends upon the fact that ferric, alu- minium and titanium acetates are hydrolyzed in hot, dilute solutions much more readily thtn the acetates of the bivalent metals. From the equation Fe (C 2 H 3 2 ) 3 + 2HOH<=2HC 2 H 3 2 + Fe (OH) 2 - C 2 H 3 O 2 , it is evident that acid is set free which tends to stop the reaction, due to the solvent action of hydrogen ions. The concentration of free hydrogen ions, however, is kept low by the addition of sodium acetate. Then, as a rule, some manganese is likely to be precipitated, so that it is advisable to dissolve the precipitate and repeat the precipitation. Hydrated manganese dioxide is more insoluble than manganous hydroxide, Mn(OH 2 ), and hence long boiling in the air tends to increase the quantity of manganese precipitated. The method is somewhat tedious, but gives excel- lent results. Procedure. The slightly acid solution of the chlorides, con- tained in a small beaker, is treated with sodium carbonate solu- tion in the cold until a slight permanent opalescence is obtained, which is then redissolved by the addition of a few drops of dilute hydrochloric acid. Meanwhile a boiling, dilute solution of sodium or ammonium acetate is prepared in a large round-bottomed flask, containing for each 0.1 to 0.2 gm. of iron or aluminium, 1.5 to 2 gm. of acetate and 300 to 400 c.c. water. When the iron solution is ready, the lamp is taken away from beneath the flask, the iron solution is added, and then the boiling is con- tinued for one minute, the flame removed (the precipitate becomes slimy on long boiling), the precipitate allowed to settle and filtered immediately while the liquid is hot, through a fluted filter, washing three times by decantation with boiling water containing ammonium or sodium acetate. The filter together with the precipitate is spread upon a glass plate, the bulk of the precipitate rinsed into a porcelain dish, and that remaining on SEPARATION OF IRON FROM MANGANESE. 153 the filter dissolved by alternately treating with concentrated hydrochloric acid and hot water. The resulting solution is evaporated nearly to dryness on the water-bath and the basic acetate precipitation is repeated exactly as before. The filtered and washed precipitate is dissolved in hydrochloric acid and the iron separated from aluminium according to page 107. The combined filtrates containing the protoxides are acidified with 10-20 c.c. of concentrated hydrochloric acid, in order to prevent the precipitation of hydrated manganese dioxide, evaporated almost to dryness, dissolved in a little water, the manganese, nickel, cobalt and zinc precipitated by ammonium sulphide as described on p. 147, and analyzed according to p. 156. Remark. This procedure requires practice. It is especially suited for the separation of iron and titanium from the protoxides; the separation is usually less satisfactory with aluminium, so that in case considerable amounts of the latter are present, the barium carbonate separation is to be preferred. If it is merely a case of the Separation of Iron from Manganese, the following modifications of the basic acetate process give satisfactory separations with only a single precipitation. (a) 0. Brunck's Method. * The acid solution, containing not more than 0.3 gm. of iron, is treated with 0.35 gm. of potassium chloride or 0.26 gm. ammo- nium chloride for each 0.1 gm. of iron present. The solution is evaporated to dryness on the water-bath, the residue pressed with a glass rod, and heated five or ten minutes longer. By this time practically all the mineral acid is expelled. The residual salts are dissolved in 10 to 20 c.c. of water and to the resulting solution there is added 1.5 gm. of sodium acetate for each 0.1 gm. of iron present. f The solution is diluted with boiling water to a volume of 400 to 500 c.c. for each 0.2 gm. of iron present; it is heated, with constant stirring, until boiling begins, and then * Chem. Ztg., 1904, I, 513. Cf. W. Funk, Z. anal. Chem.', 45, 181 (1906). f The sodium acetate crystals often contain a little sodium carbonate, so that they should be dissolved in a little water and the solution made barely acid before adding it to the iron solution. 154 GRAY I METRIC ANALYSIS. the flame is removed and the precipitate allowed to settle. The solution is decanted through a fluted filter and the precipitate washed with hot water. The precipitate is dissolved in as little hydrochloric acid as possible, the iron precipitated by ammonia, filtered, dried, and ignited as described on page 87. The filtrate from the basic acetate precipitation, or better the combined filtrates from both precipitations,* is acidified with hydrochloric acid, evaporated nearly to dry ness, the residue dissolved in a little water and the manganese, nickel, cobalt and zinc precipitated with ammonium sulphide according to the directions on page 147, and separated according to page 156. (6) Method of A. Mittasch.f The slightly acid solution, containing not more than 0.3 gm. of iron and having a volume of not over 100 c.c., is carefully neutralized, while stirring constantly, by adding ammonium car- bonate solution (200 gm. of the commercial salt in 1 liter of water) from a pipette or burette. When the precipitate that is first pro- duced begins to dissolve very slowly, the neutralization is finished with an ammonium carbonate solution, which is prepared by tak- ing 50 c.c. of the first solution and diluting to 1 liter, the dilute reagent being added until the precipitate produced will not dissolve within one or two minutes of stirring. At this point, 3 c.c. of double normal acetic acid are added, and the solution stirred until the precipitate disappears. The solution is diluted with 400 c.c. of hot water and heated until it begins to boil, when the greater part of the iron will have been precipitated. Then 20 c.c. of ammonium acetate solution (60 gm. of the commercial salt in 1 liter of water) J are added and the boiling continued for * The ammoniacal filtrate from the Fe(OH) 3 precipitate is acidified with 5 c.c. of concentrated hydrochloric acid before adding it to the filtrate from the basic acetate precipitation, otherwise manganese is likely to be pre- cipitated when the two filtrates are mixed. t Z. anal. Chem., 42, 508 (1903). % Commercial ammonium acetate has the symbol NH 4 C,H 3 O 2 HC 2 H 3 O 2 . I! none of it is on hand, COO c.c. 2N. solution ammonia are mixed with 50 c.c. 2N. acetic acid; the mixture must be faintly acid. Of this solution 10 c.c. + 5 c.c. 2N. acetic acid are used for the precipitation of the iron, and 10 c.c. of 2N. acetic acid are added to dissolve the precipitate pro- duced by ammonium carbonate. SEPARATION OF IRON AND ALUMINIUM FROM MANGANESE. 155 a minute longer. Without waiting for the precipitate to settle, it is filtered off and washed with hot water until free from chlorides. The small quantity of precipitate adhering to the sides of the vessel in which the precipitation took place is dissolved in a few drops of hydrochloric acid, the iron precipitated by ammonia and the ferric hydroxide filtered off through a separate filter. Both filters are now dried, burned and the iron weighed as Fe 2 O 3 . SEPARATION OF IRON AND ALUMINIUM FROM MANGANESE, NICKEL, COBALT, AND ZINC. Sodium Succinate Method. This method, applicable for the separation of large quantities of iron from small quantities of manganese, nickel, etc., is based upon the fact that ferric iron is quantitatively precipitated from neutral solutions as light-brown ferric succinate by the addition of neutral alkali succinate solution, while manganese, nickel, etc., remain in solution. Procedure. In case the solution contains free acid and all the iron is in the ferric form, it is neutralized with ammonia until a reddish-brown coloration is formed, when sodium or ammonium acetate is added until the color becomes a deep brown, and then the solution of alkali succinate, after which the mixture is warmed gently, allowed to cool, filtered, and washed at first with cold water, then with warm water containing ammonia, until 20 drops of the filtrate leave no residue when evaporated to dryness on platinum. By means of the washing with ammonia, the ferric succinate is changed to ferric hydroxide which is dried and weighed as ferric oxide after ignitioain a porcelain crucible. If aluminium is present, the ignited residue is further analyzed as described on p. 107. The bivalent metals in the filtrate are best precipitated by the addition of ammonium sulphide and analyzed as follows: 156 GRAVIMETRIC ANA LYSIS. SEPARATION OF THE BIVALENT METALS OF THE AMMONIUM SULPHIDE GROUP FROM ONE ANOTHER. Separation of Zinc from Nickel, Cobalt, and Manganese. All methods for this separation rest upon the slight solubility of zinc sulphide and the ready solubility of the remaining sulphides in their state of formation.* At this point it may be well to say a few words with regard to the most recent explanation of the process that takes place in the solution of electrolytes. Solubility Product. Inasmuch as no substance is absolutely insoluble in water, it follows that in every case where a precipitate is produced the solu- tion is saturated with the substance and (according to Ostwald) in the case of difficultly soluble substances the dissolved portion is practically completely dissociated electrolytically. The binary sub- stance A, consisting of the elements B and C, is decomposed in aqueous solution according to this scheme*. If the concentrations of the ions B and C are designated by [B] and [C], and that of the undissociated portion by [A], then according to the mass-action law the following relation holds for any given temperature: *. = constant. Every increase of [B] or [C] causes, therefore, an increase of [A], and, as the solution is already saturated with A, this will produce precipitation of the substance. This product [B.C], which if exceeded causes a supersatura- tion of the solution, and consequently precipitation, is called the * Nickel and cobalt sulphides when once formed are insoluble in dilute acids. These substances probably exist in two allotropic modifications, of which one is soluble and the other insoluble. The soluble form has never been isolated. EXPLANATION OF THE SOLUTION OF SULPHIDES IN ACIDS. 157 solubility product. If, therefore, in any solution the solubility product is already reached, then the solution is saturated with respect to the substance A, and if the solubility product is not reached, then the liquid exerts a solvent action upon the solid substance. Explanation of the Solution of Sulphides in Acids. According to the above theory, the solution of a sulphide (e.g., zinc sulphide) in acid is conceived to take place as follows : On treating the solid sulphide with water, a part of the salt is dissolved until the solubility product is reached. This almost inappreciable amount is practically completely dissociated into ions. On adding acid to the solution, the positive hydrogen ions unite with the negative sulphur ions to form neutral hydrogen sulphide, which being a very weak acid is only dissociated to a slight extent, so that sulphur ions disappear from the solution and the solubility product of zinc sulphide is no longer reached: ZnS + 2HC1 = H 2 S + ZnCl 2 . The liquid, therefore, dissolves more of the solid zinc sulphide and the above reaction again takes place and this process is repeated until all of the zinc sulphide is brought into solution. The solu- bility of a sulphide in acid, therefore, is proportional to its solubility product and to the concentration of the hydrogen ions. If we, then, desire to precipitate zinc by means of hydrogen sulphide from a neutral solution of an inorganic compound, the following consideration shows us how this may be accomplished: If hydro- gen sulphide is conducted into a solution containing zinc com- bined with a mineral acid, the zinc is indeed precipitated, but as the amount of zinc sulphide formed increases, there is an increase in the concentration of the hydrogen ions: /SH* . , ZnCl, + 2HSH = Zn/ +2HC1. NSH The precipitation is, therefore, incomplete. It can be made complete, however, if we can avoid this increase in the concentra- *The Zn(SH) 2 is at once decomposed into ZnS and H 2 S. 158 GRAVIMETRIC A "NA LYSIS. tion of the hydrogen ions. This can take place by replacing the mineral acid formed by a weaker acid, i.e. one which is only slightly dissociated, so that the solution will contain fewer hydrogen ions.* The following methods depend upon this principle. , Method of Smith and Brunner.f Procedure. The hydrochloric acid solution of the four metals is treated with sodium carbonate ujitil a permanent precipitate is formed, which is redissolved by the addition of a few drops of very dilute hydrochloric acid. Into this almost neutral solution hydro- gen sulphide is passed for five minutes, then a few drops of a very dilute solution of sodium or ammonium acetate are added and the solution is saturated with hydrogen sulphide, allowed to stand overnight, filtered, and washed with hydrogen sulphide water which contains in every 100 c.c. 2 gms. of ammonium salt (either the chlo- ride, sulphate, or sulphocyanate). The zinc is then determined either as oxide or sulphide according to the methods described on pp. 142 and 143. Remark. Inasmuch as the exact amount of acid to be set free is unknown, it is impossible to tell exactly how much alkali acetate is necessary, and herein lies the chief difficulty. If too much alkali acetate is added, some nickel or cobalt sulphide may be precipitated (shown by the gray color of the zinc precipitate). If not enough alkali acetate is added, the zinc will not be completely precipitated. The following separation is more certain. Method of Cl. Zimmerman. J Procedure. The weakly acid solution is treated with sodium carbonate solution until a permanent precipitate is formed, which is redissolved by the addition of a few drops of very dilute hydro- chloric acid, then for every 80 c.c. of the solution 10, or at the most 15, drops of double-normal hydrochloric acid, and 10 c.c. o * Concerning the equilibrium conditions in the precipitation of sulphides by hydrogen sulphide, see Bruner and Zawadzki, Chem. Zentr., 1910, 5. t Chem. Centrabl., 1895, 26. j Ann. d. Chem. u. Pharm, 199, (1879) p. 3; 2Q4 (1880), p. 226. The addition of hydrochloric acid is in all cases necessary, because other- wise nickel sulphide will be precipitated with the zinc sulphide, especially when considerable nickel and little zinc are present. METHOD OF CL ZIMMERMAN, 159 ammonium sulphocyanate (1:5) solution are added, after which the solution is heated to about 70 C. and is saturated with hydrogen sulphide. At first the solution becomes only slightly turbid,* but after some time pure white zinc sulphide is thrown down in clouds, constantly becoming denser. After the solution has become sat- urated with hydrogen sulphide, the beaker is covered and allowed to stand in a moderately warm place until the precipitate has set- tled and the upper liquid is clear, after which the precipitate is filtered and washed, as described in the method of Smith and Brunner. From the filtrate nickel, cobalt, and manganese are precipitated by means of ammonium sulphide, filtered and separated according to the following methods. Remark. What is the part played by the ammonium sul- phocyanate in this determination? Certainly it cannot act the same as the ammonium acetate in the Smith-Brunner method, for sulphocyanic acid is not, like acetic acid, a weak acid, but a very strong one, almost as strong as hydrochloric acid itself, and the dissociation of strong acids is only slightly influenced by the addi- tion of their neutral salts. Ammonium sulphocyanate probably simply " salts out" the zinc sulphide (cf. Vol. I). By the action of hydrogen sulphide upon the zinc salt, zinc sul- phide is produced both in the hydrogel and hydrosol forms and the ammonium sulphocyanate changes the latter into the insoluble hydrogel. If this explanation is correct, the separation of zinc from nickel, etc., will succeed equally well if the ammonium sul- * There are at the start but few zinc ions in the solution. The four metals are present for the most part in the 'form of complex thiocyanates of the general formula [R(CNS)J(NH 4 ) 2 . The zinc salt, like carnallite (see Vol. I) is slightly dissociated, [Zn(CNS) 4 ](NH) 2 ^Zn(CNS) 2 and the zinc thiocyanate is converted into insoluble sulphide by the action of hydrogen sulphide. When the zinc begins to precipitate as sulphide, the equilibrium is disturbed and eventually all the zinc becomes precip- itated. i6o GRAVIME 1 RIC ANAL YSIS. phocyanate is replaced by ammonium chloride or ammonium sul- phate. That this is the case is shown by the following method. "Salting-out Method." Experiments were performed by G. H. Kramers in order to determine whether the separation of zinc from nickel and cobalt could be accomplished in weakly acid solutions by hydrogen sul- phide after the addition of any ammonium salt of a strong acid.* The results obtained showed this to* be possible. Procedure. The neutral solution f containing the nickel and zinc either in the form of sulphate or chloride (the sum of the oxides present amounting to about \ per cent, of the weight of the solution) is treated with 8-10 drops of double-normal hydrochloric acid and about 2 per cent, of ammonium sulphate (referred to the total amount of liquid) and the solution is saturated at 50 C. with hydrogen sulphide; the warm solution is allowed to stand until the pure white precipitate of zinc sulphide has settled out and is then treated exactly as described under the Method of Zimmerman. Results. In the following experiments a zinc sulphate solution containing 5.890 gms. zinc to the liter and a solution of nickel sul- phate containing 5.320 gms. nickel to the liter were used. O 02 ' , i 9, Pd o =Q B c N"' %v g's 1 02 8 te -^ ^^ ^PH ^7 6 o 2 o r ^o ^ ^O u f20 20 3 5 0.1188 0.1178 0.1072 . 1066 & I 60 20 3 5 . 3533 0. 3534 1051 0. 1066 S' 20 60 10 10 0.1184 0.1178 0.3206 ).3192 W j 20 60 15 10 0.1182 0. 1 178 fc 120 60 30 10 . 1089 0.1178 * f20 20 5 5 0.1173 0.1178 O 60 20 6 10 . 3536 . 3534 0.1082 ).1066 ^ 20 60 12 10 0.1184 0.1178 :5\ 20 60 12 20 0.1168 0.1178 !NH j ^f) 20 60 8 K 0. 1184 . 1 1 78 0.1064 0.1066 Bleo 20 110 8 10 0.3542 . 3534 60 100 24 20 0.1168 0.1178 'J20 20 60 8 5 0.1182 0.1178 0.1074 0.1066 20 110 8 10 . 3552 0.3534 ( 20 60 100 24 20 0.1190 0.1178 * Or any other salt, e.g., a potassium salt, f If the solution is acid, it is neutralized by sodium carbonate as described under the preceding methods. SEPARATION OF MANGANESE FROM NICKEL AND COBALT. 161 Separation of Manganese from Nickel and Cobalt. The solution of the chlorides or sulphates is treated with an excess of sodium carbonate, strongly acidified with acetic acid, and for each gram of nickel or cobalt present 5 gms. of ammonium acetate are added, the solution is diluted to 100-200 c.c., heated to 70-80 C., saturated with hydrogen sulphide, filtered, and washed with hot water. The manganese is in the filtrate, and the nickel and cobalt are in the precipitate. Remark. The filtrate often contains small amounts of nickel and cobalt. In order to remove these metals, the solution should be concentrated and colorless ammonium sulphide added. It is then made slightly acid with acetic acid, warmed, and filtered. In case a precipitate of nickel or cobalt sulphides is formed by this treatment, the filtrate is again tested in the same way and the process repeated until no further precipitation is produced. Separation of Cobalt from Nickel, (a) Method of Tschugaeff-Brunck.* This method is probably the quickest and most accurate for the estimation of nickel in the presence of cobalt. It depends upon the fact that nickel is quantitatively precipitated by means of dimethyl glyoxime, from a barely ammoniacal solution or from a slightly acid solution containing sodium acetate. Cobalt, under these conditions, is not precipitated. Procedure. If the quantity of cobalt present does not exceed the quantity of nickel, the procedure is exactly the same as when nickel alone is present; with larger quantities of cobalt two or three times as much of the dimethyl glyoxime reagent is added and the precipitation is accomplished exactly as described on page 129. For the determination of both nickel and cobalt, the original solution is divided into two portions. In one por- tion the nickel is determined as outlined above, and in the other the two elements are deposited electrolytically as described on page 136, and the cobalt found by difference. If only a little of the substance is available, the two metals are deposited * O. Brunck, Z. angew. Chem., 1907, 1848. 1 62 GRAVIMETRIC ANA LYSIS. together by electrolysis, the weighed deposit dissolved in nitric acid (the electrodes must be completely immersed in the acid and the solution boiled for at least 20 minutes), the resulting solution concentrated to a small volume, and the nickel deter- mined as described above. The method can be recommended strongly. (6) The Potassium Nitrite Method of N. W. Fischer.* Brunck's Modification.^ The solution containing an excess of acid is evaporated to dryness in a porcelain dish and the residue treated with one or two drops of dilute hydrochloric acid and 5 to 10 c.c. of water. Pure caustic potash solution is then added drop by drop until the reaction is barely alkaline. The resulting precipitate is dissolved in as little glacial acetic acid as possible, half of the solution's volume of 50 per cent, potassium nitrite solution is added, and 10 drops more of acetic acid; the mixture is stirred well and allowed to stand twenty-four hours. At the end of this time the precipitation is almost always complete. It should be tested, however, by removing a little of the undiluted solution with a pipette, adding to it a little more potassium nitrite solution, and allowing to stand a little longer. If at the end of an hour no further precipitation results, then all the cobalt has been precipitated. If a precipitate is formed, the whole solution is treated with more potassium nitrite and again allowed to stand. The clear liquid is poured through a filter, the residue trans- ferred to the filter and washed with a 10 per cent, potassium acetate solution until 1 c.c. of the filtrate on being acidified with acetic acid and boiled with 1 c.c. of a 1 per cent, solution of dimethyl glyoxime will show no test for nickel. This is usually the case after washing four times. As much of the precipitate as possible is now transferred to a small porcelain dish, which is covered with a watch-glass, cautiously acidified with sulphuric acid, and heated on the water-bath until no more brown vapors are evolved. * Pogg. Ann., 71, 545 (1847). } Z. angew. Chem., 1907, 1847. SEPARATION OF COBALT FROM NICKEL. 163 The small quantity of precipitate remaining on the filter is dis- solved* by pouring hot, dilute sulphuric acid through the filter and this acid is added to the main solution of the cobalt. After evaporating as far as possible on the water-bath, the heating is continued on an air-bath until dense vapors of sulphuric acid are evolved. After cooling, the residue is dissolved in water and the cobalt determined electrolytically, as described on p. 138. If it is not convenient to carry out an electrolysis, the nitrite precipitate is dissolved in hydrochloric acid and the cobalt determined according to p. 139 (b). The filtrate containing the nickel can be treated with hydro- chloric acid until the nitric acid is completely decomposed, and the nickel then precipitated as black nickelic hydroxide by caustic potash and bromine water, filtered, washed, and weighed as the oxide, according to p. 137. Remark. This method gives reliable results provided the solution is free from alkaline earths. In the latter case the nickel and alkaline-earth metals are precipitated with the cobalt. (Cf. Vol. I.) (c) Liebig's Potassium Cyanide Method. * This method is based upon the different behavior of the com- plex cyanogen compounds of both metals towards bromine or chlorine in alkaline solution. (Cf. Vol. I.) Procedure. The neutral solution, which may contain only nickel, cobalt, and the alkalies, is treated with an excess of purest 98 per cent, potassium cyanide and 5 gm. of pure potassium hy- droxide, after which bromine water is added, with constant stirring, until no more nickelic hydroxide is precipitated. Care must be taken that the solution remains strongly alkaline until the end of the process; upon this point depends the success of this excellent method. When the precipitation is complete, the solution is diluted with cold water and the nickel determined as oxide, as described on p. 137. The cobalt remains in the filtrate as potassium cobalticyanidc. After the addition ot dilute sulphuric acid, the solution is evapo- * Ann. d. Chem. u. Pharm., 65, 244; 87, 128. 164 GRAVIMETRIC ANALYSIS. rated as far as possible on the water-bath, a little concentrated sulphuric acid is added, and the residue is heated over a free flame until dense, white fumes are evolved and the effervescence has ceased : 2K 3 Co (CN) 6 + 12H 2 SO 4 + 12H 2 O - The cold, blue mass is dissolved in water and the cobalt deposited electrolytically; or, the* cobalt may be precipitated by the addition of bromine water and potassium hydroxide, filtered, dried and determined as metal according to p. 139. (d) Liebig's Mercuric Oxide Method. In this method advantage is taken of the fact that potassium nickelocyanide, like almost all other complex cyanogen compounds, is decomposed by mercuric oxide, whereas potassium cobalti* cyanide, on the contrary, is unaffected : K 2 Ni(CN) 4 +2HgO + 2H 2 - Ni(OH), + 2KOH -f 2Hg(CN) 2 . Procedure. A slight excess of pure potassium cyanide is added to the neutral solution, which is then heated on the water-bath for at least one hour in order to change the potassium cobalto- cyanide to. potassium cobalticyanide (cf. Vol. I). The solution is then treated with a suspension of mercuric oxide in water and heated for a long time, with frequent stirring, upon the water- bath. The decomposition is complete after one or two hours. The solution is diluted somewhat with hot water, and the pre- cipitate, consisting of nickelous hydroxide and the excess of mercuric oxide, is filtered off, dried, ignited under a hood with a good draft, and the residue of nickel oxide weighed as described on p. 137. It is better, however, to dissolve the nickel oxide in sulphuric acid and determine it electrolytically according to p. 136, or as the salt of dimethyl glyoxime, according to p. 129. The filtrate containing potassium cobalticyanide and mercuric cyanide is treated with sulphuric acid exactly as described under (b) and the cobalt determined as metal, preferably by electrolysis. SEPARATION OF NICKEL FROM ZINC. 165 The author has also tested and found satisfactory the method of Ilinsky and Knorre;* but it seems to have no advantages over the above-described procedures. Recently Rosenheim and Huldschinsky f have applied VogePs qualitative test for cobalt (cf. Vol. I, p. 137) to the quantitative separation of this metal from nickel, and have obtained excellent results. Separation of Nickel from Zinc. Method of Tschugaeff-Brunck.J The solution is treated with ammonium chloride, and enough ammonia to make it slightly ammoniacal; no precipitate will be formed if sufficient ammonium chloride has been added. The solution is then just acidified with hydrochloric acid, heated to boiling and the nickel precipitated with an alcoholic 1 per cent, dimethyl glyoxime solution exactly as outlined on p. 129. In the nitrate, it is best to precipitate the zinc as sulphide by acidifying with acetic acid and saturating the hot solution with hydrogen sulphide (cf. p. 145). Remark. When considerable zinc is present it is necessary to add more dimethyl glyoxime to precipitate the nickel. Separation of Nickel from Manganese. Method of Tschugaeff- Brunck. The analysis is carried out exactly as described above with the only difference that the final precipitation takes place in an acetic acid solution. The greater part of any mineral acid present is neutralized carefully with ammonia, the barely acid solution is treated with 1 per cent, dimethyl glyoxime solution and then, after the precipitate has formed, sodium acetate is added and the analysis continued according to p. 129. If the alkali acetate is added before the dimethyl glyoxime, a very voluminous precipitate is formed which, to be sure, can be filtered * Berichte, 18, 669. t Ibid., 34, 2050. j Z. angew. Chem., 1907, 1849. Ibid. 1 66 GRAVIMETRIC ANA LYSIS. with suction, but even then the nitration is tedious. Thus when possible it is best to add the sodium acetate after the dimethyl glyoxime. When, on the other hand, iron has been removed by a basic acetate separation and nickel and manganese are to be determined in the nitrate, the precipitation must take place in a solution already containing sodium acetate. In the nitrate from the nickel dimethyl glyoxime precipitation, the manganese is precipitated with ammonium^ sulphide and determined as described on p. 125. Separation of Nickel from Iron. If the iron is present in the ferrous condition it is oxidized by boiling with nitric acid. Then from 1 to 3 gm. of tartaric acid are added and the solution made slightly ammoniacal in order to find out whether enough tartaric acid has been added (the solution must remain perfectly clear) . After making barely acid with hydrochloric acid, the nickel is precipitated with dimethyl glyoxime, the acid just neutralized with ammonia, and the analysis continued according to p. 129. Determination of Nickel in Steel. * The sample, weighing about 0.5 gm., is dissolved in 10 c.c. of concentrated hydrochloric acid, enough nitric acid is added to completely oxidize the iron, and if there is any separation of silica at this point some hydrofluoric acid is added. Two or three gms. of tartaric acid are introduced, and the solution diluted to a volume of 300 c.c. It is then carefully tested to see whether enough tartaric acid is present to prevent any precipitation of iron when the solution is made alkaline with ammonia, more tartaric acid being added if necessary. The solution, which is left slightly acid, is heated nearly to boiling and treated with 30 c.c. of a 1 per cent, alcoholic solution of dimethyl glyoxime. The acid is finally very carefully neutralized with ammonia, leaving the solution so that it barely smells of this reagent. The precipitate is filtered through * O. Brunck, Stahl und Eisen, 28, 331. REMOVAL OF FERRIC CHLORIDE BY ETHER. 167 a Gooch or Munroe crucible, washed with hot water, dried at 110-120 for 45 minutes and weighed as Ni(C 4 H 7 N 2 O 2 ) 2 . By this method the nickel in a sample of steel can be deter- mined within about two hours. The results are accurate, but lower than is often obtained in practice, because the cobalt is usually determined with the nickel, which is not the case in this method. Removal of Ferric Chloride by Ether, Method of Rothe. The fact that ferric chloride dissolved in hydrochloric acid, sp. gr. 1.1, is more soluble in ether than in this acid is often taken advantage of in the determination of metals such as nickel, copper, vanadium and chromium in samples of steel. It has also been used for the determination of sulphur in steel after oxidation to sulphuric acid, which does not dissolve in the ether. The under- lying principle is the same as that governing the distribution of iodine between water and carbon disulphide (see pp. 658 and 659, footnote). An example will be given of such a process in the Blair method for estimating vanadium, molybdenum, chromium and nickel in steel. (See p. 313.) 1 68 GRAl/IMETRIC ANALYSIS. METALS OF GROUP II. MERCURY, LEAD, BISMUTH, COPPER, CADMIUM, ARSENIC, ANTIMONY, TIN (PLATINUM, GOLD, SELENIUM, TELLURIUM, MOLYBDENUM, GERMANIUM, TUNGSTEN, AND VANADIUM). A. SULPHO-BASES. MERCURY, LEAD, BISMUTH, COPPER, CADMIUM. MERCURY, Hg. At. Wt. 200.6. Forms: HgS, Hg 2 Cl 2 , and Hg. Determination as Sulphide. (a) By Precipitation with Hydrogen Sulphide. The solution containing no oxidizing substances (FeCl 3 , Cl f much HNO 3 , etc.) and the mercury entirely as mercuric salt is saturated with hydrogen sulphide in the cold, the precipitate allowed to settle, filtered through a Gooch crucible, washed with cold water, dried at 105-110 C. and weighed. Remark. This method affords excellent results and should be used whenever possible. Unfortunately, however, it is not always applicable, for in most cases the solution to be analyzed contains strong nitric acid (obtained by the solution of impure mercuric sulphide in aqua regia, by the decomposition of organic mercury compounds by the method of Carius, or by the oxidation of mcrcurous salts). It is not possible to expel the excess of nitric acid by evaporating the solution with hydrochloric acid, because consider- able amounts of mercuric chloride are thereby volatilized with the escaping steam. Thus 50 c.c. of a mercuric chloride solution containing 0.5235 gm. of the salt, treated with 10 c.c. of nitric acid and evaporated on the water-bath five times almost to dryness, with the addition each time of 50 c.c. concentrated hydrochloric acid, yielded in separate experiments 0.3972 gm. mercuric sulphide = 88.56 per cent, and 0.3695 gm. mercuric sulphide = 82.39 per cent., or, in other words, a loss of 11-17 per cent. In such a case the following procedure suggested by Volhard should be used: DETERMINATION OF MERCURY AS SULPHIDE. 169 (b) By Precipitation with Ammonium Sulphide. The acid solution of the mercuric salt is almost neutralized with pure sodium carbonate and is treated with a slight excess of freshly-prepared ammonium sulphide. Pure sodium hydroxide solution (free from Ag, A1 2 3 , and SiO 2 ) is then added, meanwhile rotating the solution until the dark liquid begins to lighten, when it is heated to boiling and more sodium hydroxide is added until the liquid is perfectly clear. The solution now contains the mer- cury as sulpho-salt, Hg/g^ a . Ammonium nitrate is then added and the solution boiled until the ammonia is almost entirely expelled, and the precipitate is allowed to settle, which it will do much more quickly than if it were produced by hydrogen sulphide directly. By means of the boiling with ammonium nitrate, the sulpho-salt is decomposed according to this equation: Hg(SNa) 2 + 2NH 4 N0 3 = 2NaNO, + (NH 4 ) 2 S + HgS. The clear liquid is poured through a Gooch crucible, and the precipitate washed by decantation with hot water until the wash water no longer reacts with silver nitrate solution. The precipitate is then transferred to the crucible, dried at 110C.,and weighed. In case the precipitate contains free sulphur, it should be boiled with a little sodium sulphite before filtering.* H. Rauschenbach tested this method, analyzing pure mercuric chloride with the addition of nitric acid, and obtained as a mean of two experiments 73.80 per cent. Hg instead of the theoretical value, 73.85 per cent. A still better way of removing free sulphur from the precipitate consists of extracting with carbon bisulphide. In this case the mercuric sulphide, together with the sulphur, is filtered through a Gooch crucible, completely washed with water and then three times with alcohol. The crucible is now placed upon a glass tripod in a beaker containing some carbon bisulphide (Fig. 35) ;t * By boiling with sodium sulphite, the sulphur is changed to sodium thiosulphate, Na 2 SO 3 + S = Na 2 S 2 O 3 . t G. Vortmann, Uebungsbeispiele aus der quantitativen chemischea Analyse, p. 28, Vienna, 1899. 170 GRAVIMETRIC ANALYSIS. Cold !=L=-:=rM Water Hot Water the beaker is supported over a vessel filled with hot water and covered with a round-bottomed flask con- taining cold water which serves as a condenser. After about an hour the sulphur will be com- pletely extracted. The carbon bisulphide is removed from the precipitate by washing once with alcohol and once with ether. The ether is driven off by gently warming, and the precipitate then dried at 110 C. and weighed. H. Rauschenbach analyzed pure mercuric chloride by this method and obtained as a mean of eight experiments 73.79 per cent. Hg instead of 73.85 per cent., and in the case of eight further experiments made without FlG 35> removing the sulphur he obtained 74.17 per cent, instead of the theoretical value, 73.84 per cent. Determination of Mercury in Non-Electrolytes. If it is desired to determine mercury in an organic non-electro- lyte, the compound is decomposed by the method of Carius (see Elementary Analysis) by heating in a closed tube with concen- trated nitric acid, and the mercury precipitated as sulphide by the method of Volhard ; or the acid solution is treated with pure sodium hydroxide solution to alkaline reaction and then with pure potas- sium cyanide until the mercuric oxide has dissolved, after which the solution is saturated with hydrogen sulphide, ammonium acetate added, the solution boiled until the ammonia is almost entirely expelled, the precipitate allowed to settle, filtered, and washed first with hot water, then with hot dilute hydrochloric acid, and finally with water. After drying at 110 C. the precipitate of mercuric sulphide is weighed. Determination as Mercurous Chloride. For the analysis of a solution containing a mercurous salt, the solution is treated with sodium chloride, diluted considerably with water, filtered, after standing twelve hours, through a Gooch cruci- ble, dried at 105 C., and weighed. If the solution contains a mer- DETERMINATION OF MERCURY AS METAL. 171 curie salt, it is first reduced, by the method of H. Rose, by means of phosphorous acid in the presence of hydrochloric acid. Procedure. The mercury solution (which almost always con- tains nitric acid) is treated with hydrochloric acid, diluted con- siderably with water, an excess of phosphorous acid is added, and after standing for twelve hours the precipitate is filtered through a Gooch crucible, dried at 105 C., and weighed. Remark The results obtained by this method are always about 0.4 per cent, too low, but in spite of this fact the method is to be recommended. The phosphorous acid necessary for this method is obtained by the oxidation of phosphorus in moist air or by the decomposition of phosphorus trichloride with water, evaporating the solution to remove the hydrochloric acid and dissolving the residue in water. Determination as Metal. Almost all mercury compounds are quantitatively decomposed on heating with lime according to the equation HgX+CaO==CaX + Hg + O. The iodide alone is not readily acted upon. To carry out this determination, a glass tube 50 cm. long and 1.5 cm. wide, open at both ends, is taken and in one end an asbes- tos plug is placed, followed by 8 cm. of pure lime, then an intimate mixture of a weighed amount of substance with lime, finally a layer of lime 30 cm. long and at the other end of the tube another asbes- tos plug. After the tube has been filled, the end nearest this sec- ond asbestos plug is drawn out until it is only 4 cm. wide, and is connected by means of rubber tubing with the empty narrower arm of a Peligot tube. The other wider end of the Peligot tube is loosely filled with pure gold-leaf. The glass tube is placed in a combustion-furnace and illuminating-gas (carbon dioxide is less suited) is passed through it for half an hour. The tube is heated, at first where the 30 cm. layer of lime is, then the other burners are lighted one after another until finally the entire contents of the tube is subjected to gentle ignition. During the whole of the opera- tion illuminating-gas is being passed through the apparatus at the rate of about three bubbles a second. The greater part of the mercury collects in the lower empty end of the Peligot tube and 172 GRAVIMETRIC ANALYSIS. the mercury vapors that are carried further amalgamate with the gold. A small amount of the mercury condenses in the drawn-out tube. After cooling the apparatus (in a current of illuminating-gas) the narrow part of the tube is cut off both sides of the condensed mercury and weighed. It is then heated gently while air is passed through it to volatilize the mercury and again weighed. The dif- ference in weight gives the amount of mercury condensed in the tube. The Peligot tube is usually *noist ; dry air is, therefore, con- ducted through it for some time, after which it is weighed. The results obtained by this method * are perfectly satisfactory. Winteler found in the analysis of pure mercuric chloride 73.81, 73.88, 73.74 per cent, instead of the theoretical value, 73.85 per cent. Experiments made attempting to condense the mercury under water invariably gave too low values (about 1-2 per cent.)- Although it is easy to obtain good results by this method, it can be dispensed with, for the sulphide method affords just as exact results in much less time. In case it is desired to determine the amount of mercury vapor present in a given space, it is only necessary to aspirate the gas through a calcium-chloride tube filled with gold-leaf. The gain in weight of the latter shows the amount of mercury present in the gas. Electrolytic Determination of Mercury.f Mercury can be determined satisfactorily by the electrolysis of acid, neutral, or alkaline solutions. The metal is deposited in the form of little drops, which, when the quantity is small, adhere to the electrode, or, when larger amounts are present, the mercury may collect at the bottom of the platinum dish used as cathode. The use of silver-plated electrodes is also advised. The electrolysis takes place to advantage in solutions slightly acid with nitric acid. * First proposed by Erdmann and Marchand, J. prakt. Chem., 31, 385. fLuckow, Z. anal. Chem., 19, 15 (1880); Smith and Knerr, Am. Chem. J., 8, 206; F. W. Clarke, Ber., 11, 1409 (1878); Riiderff, Z. angew. Chem., 1894, 388; Classen and Ludwig, Ber.; 19, 324 (1886); G. Vortmann, Ber., 24, 2750 (1891). ELECTROLYTIC DETERMINATION OF MERCURY. 173 Procedure. The neutral or slightly acid solution of the mer- curous or mercuric salt is placed in a beaker, diluted with water to 150 c.c., treated with 2 or 3 c.c. of concentrated nitric acid, and electrolyzed with a platinum gauze cathode at the ordinary temperature with a current of 0.055-0.10 ampere. The voltage under these conditions corresponds to 3.5-5 volts. If the electrolysis is started at night, it will be finished next morning, provided the amount of mercury does not exceed 1 gm. By using a current of 0.6-1 ampere the electrolysis is finished at the end of two or three hours. At the end of the electrolysis, the metal is washed with water without interrupting the current, then with alcohol* and dried. The metal is further dried by touching it with filter paper, and then placing it in a desiccator f over fused potassium hydroxide and a small dish of mercury. In this way correct results are obtained. Drying at 100 and then over sulphuric acid in a desiccator gives rise to low results because the acid absorbs considerable mercury vapor. During the electrolysis of mercuric chloride t the solution often becomes turbid in consequence of the formation of insoluble mercurous chloride; this does no harm, however, as the metal is subsequently deposited on the cathode. Mercury can also be electrolyzed from a solution in potassium cyanide in the presence of some caustic alkali, and similarly from * It is usually stated that alcohol is not to be used, but with gauze electrodes it does no harm. t Private communication from A. Miolati, cf. Borelli, Revisto tecnica, V, Part 7 (1905). Even at 20 the tension of mercury vapor is considerable. It amounts to 0.00133 mm. J In the electrolysis of the chloride, it is better to use a platinum dish with dull, unpolished inner surface (Classen) because then any mercurous chloride will certainly be reduced to metal, which is not always the case with gauze electrodes. When a dish is used as cathode, the electrode is washed with water, without breaking the current, by pouring water into it from a wash-bottle while the solution is being siphoned off. As soon as the ammeter (or voltmeter used as an ammeter) reaches the zero mark, the washing is finished. The current is then turned off, the water carefully poured off, the rest of it removed by touching it with filter-paper, and the electrode dried as above and weighed. The drying requires several hours. 174 GRAVIMETRIC ANALYSIS. a solution formed by dissolving mercuric sulphide in 50-60 c.c. of concentrated sodium sulphide solution. The great advantage of the electrolytic determination of mercury lies in the fact that good deposits are obtained irrespective of the nature of the acid radical, or element, which is combined with mercury. LEAD, Pb. A*. Wt. 207.1. Forms: PbO, PbS0 4 , Pb0 2 , and in rare cases PbCl 3 .* i. Determination as Lead Oxide, PbO. If the lead is present as carbonate, nitrate, or peroxide, it is only necessary to ignite a weighed portion in a porcelain crucible over a small flame and weigh the residue. The treatment of the nitrate requires care, because on rapid ignition the mass decrepi- tates. 2. Determination as Lead Sulphate, PbS0 4 . If the lead is present in solution in the form of its chloride or nitrate, it is placed in a porcelain dish, an excess of dilute sulphuric acid is added t and the mixture evaporated on the water-bath as far as possible, then over a free flame until dense white fumes of sulphuric acid are evolved, and aftenvards allowed to cool. A little water is added, the mixture stirred, allowed to stand some hours, filtered through a Gooch crucible, washed at first with 4 per cent, sulphuric acid, then with alcohol, and dried at 100 C. The dried precipitate is placed in a larger porcelain crucible, provided with an asbestos ring, and ignited over the full flame of a Teclu burner. If it is desired to use an ordinary filter, the precipitate is finally washed w r ith alcohol until the wash liquid no longer gives the sul- phuric acid reaction, dried, as much of it as possible is transferred to a weighed porcelain crucible, the filter ignited in a platinum spiral (p. 22), and the ash added to the contents of the crucible. By means of the reducing action of the burning filter, some of the lead sulphate adhering to it is always reduced to lead, which must * See Analysis of Vanadinite. f The solution at the time of filtering should contain about 5 per cent, of free sulphuric acid. DETERMINATION OF LEAD AS LEAD SULPHATE. 1 75 be changed back to sulphate before weighing. For this purpose the precipitate in the crucible is moistened with dilute nitric acid, evaporated on the water-bath to dryness, a few drops of concen- trated sulphuric acid added and the crucible heated over a free flame until no more fumes are given off, when it is gently ignited and weighed. In case the lead is originally present as acetate, the solution is treated with an excess of dilute sulphuric acid and twice its volume of alcohol, filtered after standing some hours, and the precipitate of lead sulphate treated exactly as described above. In order to determine the amount of lead present in organic compounds, the substance can be placed in a large porcelain crucible, treated with an excess of concentrated sulphuric acid, and very cautiously heated in the covered crucible over a free flame until the sulphuric acid is completely expelled. The crucible is then gently ignited, and if the residue is white it is ready to-be weighed; otherwise more sulphuric acid is added and the process repeated until finally a white residue is obtained. In case the organic lead compound is soluble in water, it is preferable to first precipitate the lead by means of hydrogen sulphide and then transform the precipitate into sulphate. For this purpose, as much as possible of the washed and dried pre- cipitate is placed upon a watch glass, the filter and remainder of the precipitate are heated in a large porcelain crucible, which is supported iii an inclined position, and heated carefully over a small flame until the filter-paper is completely consumed. The main part of the precipitate is added to the crucible, which is then covered with a watch-glass and treated with concentrated nitric acid at the temperature of the water-bath. When the main reaction is over, the treatment with fuming nitric acid is repeated until the contents of the crucible are pure white in color. The watch-glass is then removed, five or ten drops of dilute sulphuric acid are added, the liquid is evaporated as far as possible on the water-bath, the excess of sulphuric acid is re- moved by heating on the air-bath (cf. Fig. 11, p. 27) and the lead sulphate is weighed: Should the precipitate be dark colored after the ignition, it is moistened with concentrated sulphuric acid and the excess of acid again expelled. 176 GRAVIMETRIC A 'NA LYSIS. If the lead is present in an organic compound which is not capable of dissociation, the compound should be decomposed in a closed tube with strong nitric acid according to the method of Carius (see page 287), finally washing out the contents of the tube, adding sulphuric acid, and treating the precipitate as above de- scribed. Separation of Lead Sulphate from barium Sulphate and Silicic Acid. In the analysis of sulphide ores containing lead, it is customary to dissolve the finely powdered ore in nitric acid, or aqua regia, and to remove the volatile acids by evaporation with sulphuric acid, eventually heating over the free flame until fumes of sul- phuric acid come off thickly. The sulphuric acid should be diluted with an equal volume of water before adding it to the original solution; usually 5 c.c. of the diluted acid is sufficient. After the evaporation the moist residue is allowed to cool, then water is added and the precipitate filtered and washed with 1 per cent, sulphuric acid. The precipitate contains all the lead as sulphate but often contains silica and barium sulphate (also strontium sulphate and sometimes calcium sulphate). It is purified by redissolving the lead in hot ammonium acetate solution (made by neutralizing acetic acid, sp. gr. 1.04, with ammonia, sp. gr. 0.96, and leaving the mixture barely ammoniacal) . When the precip- itate is large in amount it is best to wash it into a beaker or flask and heat it with about 20 c.c. of the ammonium acetate solution (or enough to dissolve all the lead sulphate), then filter through the original filter and wash with hot ammonium acetate solution, and finally with hot water until the filtrate gives no blackening with ammonium sulphide. Small amounts of lead sulphate are dissolved on the filter. The silica and barium sulphate will all remain behind. In order to obtain lead from the acetate solution, it is precip- itated as sulphide by hydrogen sulphide, and transformed, after drying, into sulphate as described on page 175. Or, the ammonium acetate solution may be treated with 10 c.c. of 50 per cent, sulphuric acid, the acetic acid removed by evap- oration, the residue allowed to cool, diluted with water, and the ELECTROLYTIC DETERMINATION OF LEAD. 177 lead sulphate filtered into a Gooch crucible, washed with dilute sulphuric acid, heated in an air bath and weighed. If the amount of ammonium acetate solution used is not too large, the lead may be precipitated by adding enough sulphuric acid to the acetate solution to make the solution contain from 5-10 per cent, sulphuric acid. Sometimes the precipitate is not pure lead sulphate, in which case it should be redissolved in ammonium acetate and the precipitation as sulphate repeated. 3. Electrolytic Determination of Lead as Peroxide (Pb02). Many neutral solutions of complex lead salts, a neutral solu- tion of lead acetate, also alkaline lead solutions yield deposits of metallic lead on the cathode when subjected to electrolysis; but lead is never determined this way; partly because of the round- about process necessary, and partly on account of the fact that the deposited lead is oxidized so readily. If a neutral or slightly acid (nitric acid) solution of lead nitrate is electrolyzecl, the lead is deposited partly as metal upon the cathode and partly as brown peroxide on the anode. If, however, the solution contains sufficient free nitric acid, it is easily possible to deposit the lead quantitatively upon the anode as firmly -adhering lead peroxide. Procedure. The solution of lead nitrate, containing not more' than 0.5 gm. lead, is placed in a platinum dish whose inner sur- face is unpolished (as recommended by Classen), 20-30 c.c. of pure nitric acid, sp. gr. 1.4, are added, the solution is diluted to 150-200 c.c. and electrolyzed in the cold with a weak current of about 0.5-1 ampere and 2-2.5 volts. When the electrolysis is carried out in the cold, all the lead will be deposited as the peroxide at the end of two hours and a half or three hours. Only an hour or an hour and a half is required if the temperature of the cell is kept at 50-60. If it is desired to let the electrolysis run over night, a current of 0.05 ampere is sufficient. A suitable arrangement of the electrolytic apparatus is shown in Fig. 36, but the dish should serve as anode and the platinum spiral as cathode. The resistance W is made by taking about 10 m. of nickel wire of about 0.5 mm. diameter, fastening it to a board as shown in the drawing and connecting the wires in pairs by means of a brass hook, of which only one is shown in the sketch. By suitably moving these hooks it is possible to vary i 7 8 GRA VIME TRIG ANAL YSIS. the resistance at will. Instead of this arrangement, that shown in Fig. 31, p. 132 ; may be used; such an apparatus is more convenient but also more expensive. At the end of the elec- trolysis, which is shown by the fact that dilution with a little water so as to expose a fresh surface of platinum causes no yel- lowish-brown coating to appear at the end of half an hour, the dish is washed without breaking the current. This is accom- plished by introducing distilled water while the solution is being siphoned off. It is important in this operation to keep the deposit JS innuh FIG. 36. of lead peroxide completely covered with liquid. When the solution that is being siphoned off no longer reacts acid, or at least only barely acid, the washing is complete and the circuit can be broken. The dish is finally washed once more with distilled water, dried at 180 C., and weighed. The results obtained are always slightly high on account of the lead peroxide not being completely anhydrous when dried at this temperature, so that it seems to the author to be advisable to gently ignite the dish before weighing, thereby readily converting the peroxide into lead oxide.* The results obtained in the author's laboratory leave nothing to be desired. * Cf. W. C. May, Z. analyt. Chem., 14, 347 (1875). DETERMINATION OF BISMUTH AS BISMUTH OXIDE, ETC. 179 Results. (a) 10 c.c. lead nitrate solution containing 0.0631 gm. lead yielded deposits of PbO 3 weighing 0.0734, 0.073L 0.0735, 0.0733 gm. ; mean 0.07332 corresponding to 0.0635 gm. lead. After ignition the lead monoxide formed weighed respectively 0.0679, 0.0678, 0.0679, 0.0681; mean 0.0679 gm. corresponding to 0.0630 instead of 0.0631 gm. lead. (b) 10 c.c. of a lead nitrate solution containing 0.1898 gm. lead yielded deposits of PbO 2 weighing 0.2202, 0.2200, 0.2203, 0.2202; mean 0.2202 corresponding to 0.1907 gm. lead. After ignition the weights of lead oxide obtained were 0.2042, 0.2046, 0.2043, 0.2041; mean 0.2041, corresponding to 0.1897 gm. Pb instead of U.I 898 gm. These experiments were performed by M. Stoffel. Remark. By employing a stronger current and keeping the solution warm during the electrolysis, the deposition is complete in much less time, but according to the author's experience the results obtained are not so satisfactory. By rotating one of the electrodes and using a stronger current, the deposition can be made to take place in a short time. If a little lead deposits on the cathode, this is remedied by stopping the current for a short time, toward the end of the electrolysis. Besides the above-mentioned forms, lead is also determine i as the chromate and as the chloride. The latter method is sometimes used in the analysis of bearing metal, cf. p. 252. BISMUTH, Bi. At. Wt. 208.0. Forms: Bi 2 3 , Bi 2 S 3 , Bi. i. Determination as Bismuth Oxide, Bi 2 8 . Solid bismuth nitrate or carbonate is readily changed to the oxide by gentle ignition. When bismuth, however, is present in solution an the nitrate, it should be first precipitated as the basic carbonate and this changed by ignition to the oxide. Procedure. The solution is diluted with water (if a turbidity ensues it makes no difference) a slight excess of ammonium carbonate is added, and after heating to boiling the precipitate i8o GRAVIMETRIC ANALYSIS. is filtered off. washed with hot water, dried, ignited, * and weighed as Bi 2 3 . If the solution from which the bismuth is to be pre- cipitated contains besides nitric acid other acids (HC1,H 2 S0 4 , etc.), the precipitate produced by ammonium carbonate always con- tains basic salts of these acids which cannot be converted to the oxide by ignition. In this case, which is most frequent in analysis, the bismuth should be determined according to one of the follow- ing methods. 2. Determination as Sulphide, Bi 2 S 3 . The slightly acid solution is saturated with hydrogen sulphide, filtered through a Gooch crucible (or a filter that has been dried at 100 C. and weighed), washed with hydrogen sulphide water, then with alcohol to remove the water, and afterwards with freshly- distilled carbon bisulphide f to remove any sulphur that may be mixed with the precipitate. The washing with carbon bisulphide is continued until a few drops of the filtrate leave no residue on being evaporated to dry- ness on a watch-glass. The precipitate is then washed with alcohol to remove the carbon bisulphide and finally with ether, dried at 100 C., and weighed as Bi 2 S3. The distillation of the carbon bisulphide should be performed as follows: Ordinary commercial carbon bisulphide is placed in a long-necked, round-bottomed flask, provided with a closely fit- ting cork (not rubber) stopper which is bored once. Through the hole in the cork is placed a glass tube bent twice at right angles, whose further end leads into a dry flask (without using a stopper for this receiver). Two large beakers are placed upon the table, one filled with water at about 60-70 C. and the other with cold water. If the flask containing the carbon bisulphide is placed in the beaker containing the warm water, and the other flask in the beaker of cold water, the carbon bisulphide will distil rapidly from one flask to the other. Care must be taken during this operation * If the precipitate is large in amount, the greater part is placed on a watch-glass, the remainder adhering to the filter is dissolved in hot, dilute nitric acid, the solution evaporated to dryness in a weighed platinum dish, and the main portion of the precipitate added. The dish and its contents are heated at first gently but finally over the full flame of a Bunsen burner. f As described on p. 169 or on p. 223. DETERMINATION OF BISMUTH AS METAL. 181 that there is no lighted gas-burner in the immediate vicinity, for otherwise there is danger of the vapors of carbon bisulphide taking fire. 3. Determination as Metal. Method of H. Rose.* The bismuth is first precipitated as basic carbonate as described under 1, and the dried precipitate, together with the ash of the filter, is placed in a porcelain crucible and ignited gently. Five tunes as much of 98 per cent, potassium cyanide is added to the con- tents of the crucible and the mixture is fused, whereby the oxide and basic salt are changed to metallic bismuth: Bi 2 3 + 3KCN - 3KCNO + Bi, 2BiOa+4KCN=2KCNO Since bismuth melts at 268 C., but boils at- 1600 C., it is possi- ble to perform this operation with a Bunsen flame of about half the usual height without running, any risk of losing some of the bismuth by volatilization. The reduction is usually complete at the end of twenty minutes. After cooling, the melt is treated with water, which dissolves the salts and leaves the metallic bismuth behind in the form of a fused metallic globule. Frequently, however, the fusion will have loosened some of the glaze of the porcelain crucible, which will remain behind with the bismuth after the treatment with water. Consequently the aqueous solution is filtered through a filter that has been dried at 100 C. and weighed with the empty crucible. After washing first with water, then with absolute alcohol and ether and drying at 100 C., the filter is again placed in the crucible and weighed. The gain in weight represents the amount of metallic bismuth. Bismuth sulphide can also be reduced by potassium cyanide, but in this case a longer and stronger heating is necessary. 4. Determination as Metal. Method of Vanino and Treubert.f In this method the bismuth is precipitated as metal by means of formaldehyde in alkaline solution. The slightly acid bismuth solu- * Pogg. Ann., 110, p. 425. t Berichte, 31 (1898), 1303. 182 GRAVIMETRIC ANALYSIS. tion is treated with formaldehyde and a considerable excess of pure 10 per cent, caustic soda solution and warmed on the water-bath until the liquid above the precipitate has become perfectly clear ; more for- maldehyde and caustic soda solution are then added and the mixture heated over a free flame,* decanted repeatedly with water to which a little aldehyde has been added, again boiled, and by pressing with a glass rod the partly spongy, partly pulverulent precipitate is made to collect together. The precipitate is then filtered through a fil- ter that has been previously dried at 105 C. and weighed, washed with absolute alcohol, dried at 105 C. and weighed. Remark. Results obtained in the author's laboratory by this method were as a rule too high. Thus W. Urech obtained from pure bismuth nitrate solution, as a mean of four experiments, 100.78 per cent, instead of 100 per cent. The high results are caused by the difficulty in removing the last traces of alkali. Absolutely accurate results may be obtained by dissolving the precipitated bismuth in nitric acid,, precipitating by ammonia and ammonium carbonate and weighing as the oxide according to (1). Naturally this roundabout process would only be chosen when the bismuth solution contained other acids (HC1, H 2 S0 4; or H 3 PO 4 ) ; the necessity of fusing with potas- sium cyanide would then be avoided. COPPER, Cu. At. Wt. 63.57. Forms: CuO, Cu 2 S, Cu, Co>(CNS) 2 . i. Determination as Copper Oxide, CuO. The solution, which must be free from organic substances and ammonium salts, is heated to boiling in a porcelain dish and pure caustic potash solution is added, drop by drop, until the precipi- tate becomes dark brown and is permanent, while the solution itself shows an alkaline reaction towards litmus-paper. After the precipitate has settled, the upper liquid is carefully poured through a filter and the precipitate washed by decantation with hot water until the wash water no longer shows an alkaline reaction, when the * Frequently, particularly on long boiling, the liquid becomes colored yellow or I rown. This has no influence upon the results. DETERMINATION OF COPPER AS COPPER OXIDE. 183 precipitate is transferred to the filter and completely washed. Usually a small amount of copper oxide adheres to the porcelain dish so firmly that it can be removed only by vigorous rubbing with a glass rod covered at the end with a piece of rubber tubing, and finally when the precipitate is removed from the dish some will then remain on the rubber. Consequently it is better to proceed as follows : As much of the precipitate as possible is removed by a stream of water from the wash-bottle, then two drops of dilute nitric acid are added, and by inclining the dish and rubbing with the glass rod, the whole of the precipitate remaining on the dish is moistened with the acid. Two drops of the acid are sufficient, with correct manipulation, to dissolve all of the copper oxide. A small fresh filter is prepared and the dish is held in an inclined position, so that the liquid remains near its lip, the sides are washed once with hot water and the contents of the dish (which is continually maintained in this inclined position) are heated to boiling over a .small flame and precipitated by the addition of caustic potash, drop by drop. (A large excess of alkali is to be avoided on account of its solvent action upon the precipitate.) * The whole contents of the dish are then quickly poured through the small filter and the dish is immediately washed once with water. The copper oxide is now all on the filter. The precipitate is washed with hot water, both filters are dried, and the most of the precipitate transferred to a porcelain crucible, the filters ignited in a platinum spiral, and the ash added to the contents of the crucible. The crucible is cov- ered and ignited, at first gently, and finally with the full heat of the Bunsen burner, then weighed. If the process is carried out care- fully, the results obtained are almost the theoretical values but as a rule they are a trifle high. 2. Determination as Cuprous Sulphide, The solution, which contains for every 100 c.c. about 5 c.c. of concentrated acid (best sulphuric acid), is heated to boiling and hydrogen sulphide is introduced until the solution becomes cold. If the right amount of acid was present, the precipitate settles quickly in large flocks and the -upper liquid appears completely * f. Vo\ I. 1 84 GRAVIMETRIC ANALYSIS. colorless. Before filtering, the wash liquid is prepared by passing hydrogen sulphide through the long tube of a wash-bottle for one minute, then closing the short tube with a piece of rubber tubing and shaking vigorously. As soon as no more bubbles pass through the liquid, the water is saturated ; this takes about a minute at the most. A filter is now placed in a funnel containing a platinum cone, the funnel is fitted to a suction-bottle and the filtration is begun at first without using suction, taking care that the filter is con- stantly kept full. When all the precipitate is on the filter, it is washed with the hydrogen sulphide water containing acetic acid, and, at this point also, the filter must be kept full of liquid. The washing is continued until 1 c.c. of the filtrate shows no test for mineral acid.* The filter is now for the first time allowed to drain completely, and it is dried as much as possible by means of gentle suction, then completely by heating in the drying closet at 90-100 C. As much of the precipitate as possible is now transferred to a weighed Rose crucible (of unglazed porcelain),! the filter is burned in a platinum spiral and the ash allowed to fall at first upon an unglazed crucible cover, where it is heated gently till it glows, in order to make sure that it contains no unburned car- bonaceous matter; the ash is then added to the main portion of the precipitate in the crucible. A little sulphur that has been recrystallized from carbon bisulphide is added to the contents of the crucible, the perforated cover is now placed on the crucible (Fig. 37), a stream of hydrogen is passed through it (the wash- bottle shown contains concentrated sulphuric acidj), and the cru- cible is heated at first over a small flame and finally so that the * The test for sulphuric acid is made with barium chloride. To test for hydrochloric acid, the solution is boiled until the hydrogen sulphide is ex- pelled and is then treated with silver nitrate. t A quartz crucible is more desirable, as the transformation of CuS into Cu 2 S can then be watched. t If the hydrogen is prepared from sine and hydrochloric acid, the gas should be passed first through water and then through a wash-bottle con- taining concentrated sulphuric acid. DETERMINATION OF COPPER AS CUPROUS SULPHIDE. 185 bottom of the crucible glows faintly, at which temperature the cupric sulphide is changed to cuprous sulphide, Too strong heating is inadvisable according to Hampe. * When the excess of sulphur has been driven off (which can be readily ascertained by removing the cover of the crucible and FIG. 37. rinding no blue flame to be perceptible and no odor of burning sulphur), the current of hydrogen is increased so that eight bubbles per second pass through the wash-bottle (at first, not more than four bubbles per second should have been the rate), and the flame is removed. The crucible is allowed to cool in the current of hydrogen and weighed after remaining in the desiccator for fifteen minutes. The cuprous sulphide should be brownish black or black, and should show no reddish- brown stains (due to Cu or Cu 2 O) ; this is the case if the cur- rent of hydrogen was too slow during the cooling. In this case t Z. anal. Chem ., 38, 465 (1894). 1 86 GRAVIMETRIC ANA LYSIS. a little sulphur must be added to the precipitate and the process repeated. Remark. It is evident that the sulphur used for this experi- ment should leave on ignition no weighable residue. This is why the sulphur used should be recrystallized from carbon bisulphide. The reason why it is necessary to keep the funnel filled with liquid during the filtration and washing of the cupric sulphide is this : If moist copper sulphide is exposeddto the air it is quickly oxidized and the hydrogen sulphide wash water acts upon the salt formed by the oxidation, (CuS 2 O 3 -CuS0 4 ), and transforms it into colloidal cupric sulphide, which forms a pseudo-solution, passes through the filter, and on coming in contact with the acid filtrate is coagu- lated. If, however, the precipitate is not exposed to the air during the filtration there is no oxidation and the filtrate remains clear. Instead of changing the cupric sulphide into cuprous sulphide, it has been proposed to convert it to oxide by ignition in the air and weighing the copper in this form. If, however, the highest degree of accuracy is desired, this should not be done, for the ignited product always contains some sulphate. When this method is chosen, the cupric sulphide should be heated in a glazed porcelain crucible, at first over a small flame, so that the mass does not melt, and the heat gradually increased until finally a blast-lamp is used and the copper weighed as CuO. The results are about 0.1 per cent, too high when not more than 0.2 gm. of precipitate is present. Holthof* states that copper oxide abso- lutely free from sulphate can be obtained if the precipitate is ignited wet in an inclined porcelain crucible. 3. Determination as Cuprous Sulphocyanate, Cu 2 (CNS) 2 . Method of Rivot.f The solution, slightly acid with sulphuric or hydrochloric acid (oxidizing agents must not be present), is treated with an excess of sulphurous acid,t after which ammonium sulphocyanate is * Z. anal. Chem., 28, 680 (1889). f Compt. Rend., 38, 868; see also R. G. van Name, Zeit. f. anorg. Chem., 26, 230, and Busse, Zeit. f. anal. Chem., 17, 53, and 30, 122. J Instead of sulphurous acid, ammonium bisulphite may be used. The latter is prepared by saturating aqueous ammonia with SO 2 . ELECTROLYTIC DETERMINATION OF COPPER. 187 added drop by drop with constant stirring, whereby at first a greenish precipitate of cupric and cuprous sulphocyanate is pre- cipitated, which after stirring becomes pure white. The precipitate is allowed to settle completely (this requires several hours) ; it is then filtered and washed with cold water until the filtrate shows only a slight reddish coloration when ferric chloride is added, after which it is washed several times with 20 per cent, alcohol, dried at 110-120 C., and weighed. R. Philipp found by this method 99.95 per cent, instead of 100 per cent, copper, as a mean of twelve experiments. The cuprous sulphocyanate can be dried at a temperature as high as 160 C., but at 180 C. it begins to decompose. The Munroe crucible can be used to advantage in this determination. The precipitate permits rapid filtration, and a turbid filtrate is never obtained. After the determination is finished, the greater part of the precipitate is shaken out of the crucible, and the remainder dissolved in hot nitric acid. 4. Electrolytic Determination of Copper. This most accurate and most convenient of all methods for the determination of copper was first proposed by W. Gibbs in 1864.* Copper may be deposited by means of the electric current from acid, alkaline, and neutral solutions, but for analytical pur- poses only the use of acid solutions is of importance. Procedure. The safest way, according to F. Forster,* is to deposit the copper from a sulphuric acid solution. To the neutral solution containing the copper in the form of sulphate, 10 c.c. of twice normal sulphuric acid are added, the solution is diluted to a volume of 100 c.c. and electrolyzed with exactly two volts potential at the electrodes and this potential is kept constant during the electrolysis. These conditions are fulfilled by simply connecting the electrodes with the poles of a single storage cell. The electrolysis requires at least eight hours if done at the ordi- nary temperature, but by keeping the solution at 70-SO, 0.2 gm. of copper is deposited in 60-80 minutes. If, therefore, it * Z. anal. Chem., 3, 334. 1 88 GRAVIMETRIC ANALYSIS. is desired to carry on the electrolysis over night, it is done in the cold. It is very easy to decide when the electrolysis is finished by adding a little water and noticing whether there is any more copper deposited upon the freshly exposed electrode surface. The cathode is then washed with water, without breaking the circuit, exactly as was described under the electrolytic determina- tion of nickel (p. 136). Finally, the cathode is rinsed with alcohol, dried by holding it high above a flame, cooled in a desiccator, and weighed. If these directions are followed closely, the copper is never deposited in a spongy condition. The presence of Ni, Co, Fe, Zn and Cd does not influence the analysis and the copper may be separated from these elements by m^ans of such an electrolysis. If the solution to be analyzed contains copper and some of the above-mentioned base metals, it is evaporated to dryness, heated with a little sulphuric acid until dense fumes are evolved, cooled, treated with 10 c.c. of 2 N. sulphuric acid, diluted to 100 c.c. and electrolyzed as described above. If, however, only copper is present in the solution, it may be deposited very nicely in the following manner. The solution should contain 4-5 c.c. of concentrated nitric acid in 100 c.c. If, originally, i; contained more nitric acid than this, it is either evaporated to dryness or neutralized with ammonia, and then the required quantity of nitric acid added. The solution is heated to 50-60 and electrolyzed with a current of 1 ampere and elec- trode potential of 2-2.5 volts. The electrolysis is over at the end of two hours, when not more than 0.3 gm. of copper is present. The analysis is finished as above but there is more danger of traces of copper being dissolved while the electrodes are being removed. Remark. The copper may be deposited electrolytic ally much more rapidly by the use of a rotating electrode or any stirring arrangement. The use of a gauze cathode has also been rec- ommended. The solution should not be diluted too much, as spongy deposits are obtained from very dilute solutions unless a very weak current is used. As a general rule, the more con- centrated the copper solution, the stronger the current that can be used. ELECTROLYTIC DETERMINATION OF CADMIUM. 189 CADMIUM, Cd. At. Wt. 112.4. Forms: Cd, CdS0 4 , CdO. I. Electrolytic Determination of Cadmium. Of all the methods for the determination of cadmium the electro- lytic method is not only the most convenient, but by far the most accurate, and of the many methods proposed that of Beilstein and Jawein* can be recommended. From the experience obtained in the author's laboratory the best procedure is as follows: To the solution of the sulphate a drop of phenolphthalein is added and then pure caustic soda solution until a permanent red color is obtained. A so- lution of 98 per cent, potassium cyanide is now added with constant stirring until the precipitate of cadmium hydroxide produced by the caustic soda has completely dissolved (an excess of potassium cya- nide should be scrupulously avoided), the solution is diluted with water to 100-150 c.c. and electrolyzed in the cold, using a gauze cathode, for from five to six hours with a current of 0.5-0.7 ampere and an electromotive force of 4.8-5 volts; at the end of this time the current is increased to from 1-1.2 amperes and the solution is electrolyzed for one hour more. If these directions are followed, all of the cadmium (if not more than 0.5 gm. is present) will be deposited as a firmly adhering dull deposit of almost silver- white metal. The current is then stopped, the liquid is quickly poured off and the deposited metal washed first with water, then with alcohol and finally with ether; it is dried and weighed. Experiments performed by von Girsewald gave fault- less results. After the electrolysis is finished, the solution should always be tested for cadmium. For this purpose, it is saturated with hydrogen sulphide. If much cadmium is present, a yellow pre- cipitate is obtained, but if very little, a yellow coloration results. The latter is due to the formation of colloidal cadmium sulphide, and the color is so intense that R. Philip estimates the quantity of cadimum not precipitated, by comparing the shade with that produced in a solution containing a known quantity of cadmium and the same amounts of potassium cyanide and caustic potash as in the solution tested. * Berichte, 12, 446. 190 GRAVIMETRIC ANALYSIS. Remark. If for the electrolysis a current of 0.5 ampere were used, the cadmium will not be all deposited at the end of twelve hours; if, however, the current is increased at the end, as above stated, to 1 ampere, the electrolysis will be surely finished at six to seven hours. To work with the stronger current from the beginning is not to be recommended unless a gauze cathode is used, or one of the electrodes is rotated, for otherwise the metal is deposited in a spongy form and on washing some of it is likely to be lost. A solution containing 0.4568 gm. Cd. 7 3 gm. KCX, 1 gm. NaOH, and diluted to 125 c.c. with water, can be electrolyzed in fifteen minutes with a current of 5 amperes and 5.5 volts if ona of the electrodes be rotated.* From neutral and weakly acid solution:, cadmium can be deposited electrolytically, but not from strongly acid solutions. 2. Determination as Cadmium Sulphate, CdS0 4 . Next to the electrolytic method, the determination of cadmium as the sulphate is the best. If the cadmium is combined with a volatile acid, the compound is treated in a weighed porcelain cru- cible with a slight excess of dilute sulphuric acid, the solution evapo- rated on the water-bath as far as possible, and finally the excess of sulphuric acid is removed by heating in an air-bath (the crucible is placed in a larger crucible that is provided with an asbestos ring).f The heat is applied at first slowly, and the temperature is raised gradually until finally no more fumes of sulphuric acid are evolved. The outer crucible can even be heated with the full flame of a Teclu burner without running any risk of decomposing the cadmium sul- phate ; it is, however, not necessary to heat it so strongly. As soon as the fumes of sulphuric acid cease to come off, the operation is ended and the crucible and its contents are weighed after cooling in a desiccator. The cadmium sulphate should be pure white and should dissolve in water to form an absolutely clear solution. If the cadmium has been precipitated from a solution as the sulphide, the greater part of the precipitate is placed in a large porcelain crucible, covered with a watch-glass, and treated with hydrochloric acid (1:3) on the water -bath. After the precipitate * See Edgar F. Smith's Electro- Analysis, f Cf. Fig. 11, p. 27. THE PRECIPITATION OF CADMIUM AS SULPHIDE. 191 has dissolved and the evolution of hydrogen sulphide has ceased, the lower side of the watch-glass is washed, the crucible is placed under the funnel, and the precipitate which adhered to the filter- paper is dissolved by dropping hot hydrochloric acid (1: 3) upon it, finally washing the filter with hot water, evaporating the solution upon the water-bath, and proceeding as above described. The results obtained by this method are excellent. The Precipitation of Cadmium as Sulphide. The frequently recommended determination of cadmium as the sulphide must be rejected ; it is useless. It is not possible to precipi- tate pure cadmium sulphide from acid solutions by means of hydro- gen sulphide; the precipitate is always contaminated with a basic salt (Cd 2 Cl 2 S Cd 2 SO 4 S, etc.) whether the precipitation takes place in cold or hot solutions, whether under atmospheric pressure or under increased pressure (in a pressure-flask), and in fact the amount of basic salt formed increases with the amount of free acid present. Results are obtained as much as 5 per cent, too high. Follenius * attempted to make the method possible by igniting an aliquot part of the dried and weighed precipitate in a stream of hydrogen sul- phide. If the sulphide was contaminated with sulphate, he suc- ceeded in changing it all to sulphide and obtained results that were acceptable. If, however, chloride was present, a considerable part was lost by sublimation, so that the results obtained were too low. It is, furthermore, not possible to ignite the cadmium sul- phide with sulphur in a current of hydrogen, as was described under Zinc and Copper, for cadmium sulphide is so volatile that some of it is lost. On the other hand, the method of precipitating the cadmium as sulphide from solutions containing 2-7 c.c. of concentrated sul- phuric acid in 100 c.c. is to be recommended, for by this means a precipitate is obtained which can be readily filtered and which by solution in hot hydrochloric acid (1:1) and evaporation with sul- phuric acid can be changed without loss to the sulphate and weighed as such. * Zeit. f. anal. Chem., XIII, 422. 192 GRAVIMETRIC ANALYSIS. 3. Determination as Cadmium Oxide, CdO. Cadmium carbonate and cadmium nitrate can be changed to the oxide by strong ignition. The cadmium is precipitated from its solutions at the boiling temperature by the addition of a slight excess of potassium car- bonate, and after standing for some time on the water-bath, and when the precipitate has completely settled, it is filtered off, washed with hot water, and dried. As much of the dried precipitate as possible is transferred to a watch-glass and set aside for the time being. The filter is washed with dilute nitric acid to dissolve the small amount of the precipitate which still adheres to it and the solution is received in a weighed porcelain crucible and evap~ orated to dry ness. The main portion of the precipitate is now added, and the crucible is at first very gently heated by placing the open crucible high above a small flame from a Teclu burner, until the whole mass has become a uniform brown throughout The temperature is now gradually raised until finally the full heat of the burner is reached. It is important during this operation to take care that the inner flame-mantle does not touch the cruci- ble, for otherwise reducing gases may enter the crucible and reduce a part of the oxide to metallic cadmium, which is volatile at this temperature.* The cadmium oxide is obtained as a brown powder which is infusible, insoluble in water, but readily soluble in dilute acids. f Remark. It is not advisable to precipitate the cadmium by means of sodium carbonate solution, for in that case it is difficult to wash the precipitate free from alkali. SEPARATION OF THE SULPHO-BASES FROM THE METALS OF THE PRECEDING GROUPS. Hydrogen sulphide precipitates only the metals of the "hydro- gen sulphide group" from acid solutions. It is to be noted that zinc precipitates with this group if the solution is not acid enough; * If the cadmium carbonate is filtered into a Munroe crucible, and ignited in an electric oven, the transformation takes place readily without danger of any volatilization. f The oxide after ignition is a black, crystalline powdar. ANALYSIS OF BRASS. 193 while if the solution is too acid lead and cadmium are often incompletely precipitated. A suitable concentration is 5-7 c.c. of concentrated hydrochloric acid to 100 c.c. of liquid. EXAMPLE. Analysis of Brass (Alloy of Copper and Zinc with Small Amounts of Lead, Iron, and Nickel). About 0.4-0.5 gm. of the alloy, in the form of borings,* is dissolved in about 20 c.c. of nitric acid, sp. gr. 1.2, in a 200-c.c. casserole which is covered with a watch-^lass. After the reaction begins to slacken, complete solution is effected by warming on the water-bath. The solution is then evaporated to complete dryness, moistened with a little nitric acid, dissolved in about 50 c.c. of hot water, and any metastannic acid present is allowed to settle, is filtered off, washed with hot water, dried, and the tin determined according to p. 228. To the cold nitrate 3 c.c. of pure, concentrated sulphuric acid are added, the solution is evaporated on the water-bath as far as possible, and then heated cautiously over a free flame until dense white fumes of sulphuric acid are evolved. After cooling the residue is treated with 50 c.c. of water and. 15 c.c. of alcohol, stirred well, filtered, washed, and the lead sulphate determined according to p. 174. The nitrate is evaporated until the alcohol is completely removed, 100 c.c. of water are added, the solution is heated to boiling, and hydrogen sulphide is conducted into it until it becomes cold, when the copper sulphide is filtered off, washed first with hydrogen sulphide water containing in every 100 c.c. 20 c.c. of double-normal sulphuric acid and at the end with 5 per cent, acetic acid, which is saturated with hydrogen sulphide, until the filtrate gives no precipitate on being treated with barium chloride. The copper is determined, accord- ing to p. 183, as Cu 2 S. The filtrate from the copper sulphide is evaporated to a small volume in order to remove completely the excess of hydrogen * The borings are usually somewhat greasy. They should be washed with ether before weighing. Cf. p. 236, foot-note. 194 GRAVIMETRIC ANALYSIS. sulphide, the iron is then oxidized by the addition of bromine water, precipitated by ammonia, and filtered. In order to make sure that the precipitate of ferric hydroxide contains no zinc, it is dissolved in a little hydrochloric acid and the precipitation with ammonia is repeated. The filtered and washed precipitate is ignited in a porcelain crucible and weighed as ferric oxide (cf p. 87). The combined nitrates from the ferric hydroxide are acidified with a little sulphuric acid, heated to about 50 C., and the zinc determined as zinc sulphide according to the " salting-out " method described on p. 160. For the determination of nickel, the filtrate from the zinc sulphide precipitation is boiled to expel the hydrogen sulphide and the nickel determined as the salt of dimethyl glyoxime according to p. 129. SEPARATION OF THE SULPHO-BASES FROM ONE ANOTHER. i. Separation of Mercury from Lead, Bismuth, Copper, and Cadmium. Method of Gerhard v. Rath. Principle. This separation is based upon the insolubility of mercuric sulphide in boiling, dilute nitric acid (sp. gr. 1.2-1.3) and the solubility of the remaining sulphides. Procedure. The solution (containing the mercury entirely in the mercuric form) is precipitated by means of hydrogen sulphide, the precipitate filtered off, washed with hydrogen sulphide water, transferred to a porcelain dish and boiled for a considerable length of time with nitric acid of the above concentration, then diluted with a little water and washed with water containing nitric acid. The residue of mercuric sulphide thus obtained always contains sulphur, and in case considerable lead were present it will also contain lead sulphate. It is, therefore, dissolved in a little aqua regia, diluted with water, filtered from the separated sulphur and lead sulphate and the mercury precipitated according to the method of Volhard, with ammonium sulphide (cf. p. 169). If some of the lead sulphate should go into solution with the mercury on treating SEPARATION OF BISMUTH FROM LEAD. IQ5 with aqua regia, it will be converted by the ammonium sulphide and potassium hydroxide into insoluble lead sulphide, while the mercury will be in the form of its soluble sulpho-salt. In this case the lead sulphide is filtered off, washed with dilute potassium hydroxide solution, and the mercury then precipitated as sulphide, as described on p. 169. 2. Separation of Bismuth from Lead. (a) Method of Lowe. Principle. Bismuth nitrate is changed by the action of water into an insoluble basic salt, while lead nitrate undergoes no such transformation. Procedure. The solution of the two metals in nitric acid is evaporated on the water-bath until it reaches a syrupy consist- ency, water is added, and after thorough stirring with a glass rod the evaporation is repeated and the process continued until the addition of the water fails to produce any further turbidity; a sign that the bismuth has been completely converted into the basic salt Bi 2 O 2 NO 3 OH. A cold solution of ammonium nitrate (1 NH 4 NO 3 :500 H 2 O) is now added, and after standing some time, with frequent stirring, in order to make sure that the lead nitrate is completely dissolved, the solution is filtered. The precipitate is washed with the dilute ammonium nitrate solution and dried. As much of it as possible is transferred to a weighed porcelain crucible and together with the ash of the filter is ignited,* at first gently, and finally with the full flame of a Bunsen burner. It is weighed as Bi 2 O 3 . From the filtrate the lead is precipitated according to p. 174, as sulphate, and weighed as such. It is less satisfactory to pre- cipitate the lead as sulphide and weigh it in this form after gentle heating with sulphur in a Rose crucible. (6) Method oj Jannasch.~\ Principle. The separation depends upon the different vola- tility of the two bromides. Bismuth bromide is fairly readily volatile; lead bromide is only difficultly so. * It is still better to proceed as in the determination of cadmium oxide, p.l 92. f Praktischer Leitfaden der Gewichtsanalyse. 196 GRAVIMETRIC ANALYSIS. Procedure. The solution of the nitrates is evaporated to dry- ness, 100 c.c. of water, sufficient hydrochloric acid to afford a clear solution, and a few drops of fuming nitric acid are added,* after which hydrogen sulphide is introduced. The precipitated sulphides are immediately filtered, the precipitate is dried at 100 C. in a stream of carbon dioxide, after which as much of the precipitate FIG. 38. as possible is placed in an agate mortar and the ash of the filter added to it. The whole of the precipitate is ground fine and transferred without loss to a weighed porcelain boat, which is then introduced into the decomposition tube R^ (Fig. 38), made of difficultly -fusible glass. At first a stream of dry carbon dioxide is passed through the apparatus and the substance is gently heated by means of a small flame, in order to completely dry it. The water condensing in the front part of the tube is driven over into E by careful heating. The bottle A containing bromine J is now connected with the apparatus and the stream of carbon dioxide is passed through it; the gas, carrying bromine vapors with it, is passed through the vertical calcium chloride tube filled with pieces of calcite, * By the addition of the fuming nitric acid the precipitated sulphide is contaminated with considerable sulphur; such a precipitate is more readily decomposed by the action of bromine. f In this determination, the bulb of the tube is unnecessary; it should be replaced by one such as is shown in Fig. 38, II. For other analyses it is better to have the bulb. J For this experiment the bromine used must be absolutely free from chlorine and is prepared as follows: 50-60 c.c. of commercial bromine are treated, in a tightly stoppered separatory funnel, with a 10 par cent, potassium bromide solution. The funnel is shaken vigorously, and the bromine sepa- rated from the aqueous alkali solution. After washing two or three times with water it is ready for use. SEPARATION OF BISMUTH FROM LEAD. 197 then through the concentrated sulphuric acid contained in B, after this through the tube C containing glass beads moistened with sulphuric acid, and finally through the tube D rilled with glass wool, and from this the dry bromine vapors reach the sub- stance. The latter is heated over a small flame (kept in con- stant motion) and the yellow bismuth bromide distils off and condenses partly in the narrow part of the tube and partly in the /eceiver E, which contains dilute nitric acid (1 HNC>3:2 H 2 O). The substance is heated hotter, whereby more bismuth bromide is vola- tilized, and this is again distilled as completely as possible into the receiver. Finally the substance is heated more strongly still, until the lead bromide begins to melt. When no more of the yellow sublimate is formed, the decomposition is shown to be complete and the substance is allowed to cool in a stream of carbon dioxide. The bromine that escapes from the tube K is passed into alcohol contained in the beaker F. When the apparatus has become cold, the bromine bottle is removed, and the bromine is removed from the apparatus by passing carbon dioxide through it for some time. The boat filled with lead bromide is then weighed, and from the weight of the PbBr 2 that of the lead is computed. To check this, the lead bromide is dissolved in freshly-prepared chlorine water, an excess of dilute sulphuric acid is added, and the solution is evaporated to remove the hydrochloric acid, at first on the water-bath and finally over a free flame until dense fumes of sulphuric acid are evolved. After cooling, water and alcohol are added, the precipitate filtered off and the weight of the lead sulphate determined as described on p. 174. For the bismuth determination, the nitric acid solution contained in E and K is poured into a beaker, filtered if necessary from any sulphur, evaporated to a small volume, and the bismuth precipitated by the addition of ammonium carbonate and determined as metal as described on p. 181. There have been many other methods proposed for the separa- tion of lead and bismuth,* all of which are less satisfactory than the two methods just described, so that they will not be discussed in this book. * Cf. O. Steen, Z. angew. Chem., 1895, p. 530. 198 GRAVIMETRIC ANALYSIS. 3. Separation of Bismuth from Copper. The solution is treated with an excess of ammonium carbonate, warmed gently, and filtered. The precipitate of basic bismuth carbonate almost always contains small quantities of copper, so that it is dissolved in nitric acid and the separation by means of ammonium carbonate is repeated. The basic bismuth salt is fused with potassium cyanide and weighed as metal, according to p. 181. For the copper determination, the two filtrates are combined, evaporated to remove the excess of ammonium carbonate, acidified with sulphuric acid, and the copper precipitated by means of hydrogen sulphide, being determined as cuprous sulphide accord- ing to p. 183, or the sulphuric acid solution is subjected to elec- trolysis as described on p. 187. According to Fresenius and Haidlin, bismuth can be separated from copper very nicely by means of potassium cyanide. For this purpose the acid solution is precipitated by the addition of a slight excess of sodium carbonate, potassium cyanide is added, and the solution warmed and filtered. All of the copper is found in the filtrate, while the precipitate contains bismuth oxide con- taminated with alkali. The residue is, therefore, dissolved in nitric acid, the bismuth precipitated by means of ammonium car- bonate and determined as metal according to p. 181. The filtrate containing the copper is evaporated with nitric acid, in order to destroy the cyanide, and the copper determined electro- lytically according to p. 187. 4. Separation of Lead from Copper by Means of Electrolysis. This separation depends upon the fact that the electric current deposits lead quantitatively as PbC>2 upon the anode from solutions containing a definite amount of nitric acid, while the copper is either not deposited at all under these conditions or is found upon the cathode to some extent. After the lead is completely deposited, the copper solution is poured into a second weighed platinum dish, the excess of the acid is neutralized with SEP /I RAT ION OF LEAD FROM COPPER. 199 ammonia, and the solution again electrolyzed. The copper will p.ow deposit quantitatively upon the cathode. Procedure. The solution of the two nitrates is placed in a platinum dish (of the form recommended by Classen wit-i t-.e inner surface unpolished) and 15 c.c. of nitric acid (sp. gr. 1.35- 1.38) are added, after which the solution is diluted to 150 c.c. and electrolyzed at 50 60 C. with a current of 11.5 amperes and an electrode potential of 1.4 volts. After 1-1.5 hours practically all the lead will be deposited upon the anode (dish) in the form of a firmly adhering, brown coating of lead peroxide, Pb0 2 . At the cathode (a plate electrode) a considerable part of the copper will be deposited, but the remainder will still be in solution. The circuit is broken and the solution poured as quickly as possible into a second weighed platinum dish, and the washings added to this dish. After washing the electrodes with water, the first dish with the Pb0 2 deposit is dried at 180 and weighed. The solution in the second dish contains a little lead and some copper, It is made slightly ammoniacal, 4 c.c. of concentrated nitric acid are added, and the solution electrolyzed at 60. The platinum dish now serves as the cathode, while the plate electrode * serves as the anode; in case traces of lead remain in solution after the first electrolysis, it will now be deposited. After an hour or two with a current of one ampere all the remaining copper and lead will be deposited. When the electrolysis is complete the elec- trodes are washed without breaking the circuit and the weight of the copper and PbO 2 is determined. If only small amounts of lead and copper are present, the electrolysis should take place under the conditions described on p. 187, except in this case a weighed plate electrode should be employed as the anode. Under these conditions the lead will be deposited as the peroxide upon the anode, while the copper will separate out upon the dish. * The plate electrode with copper upon it was weighed, cleaned, and then weighed again. 200 GRAVIMETRIC ANA LYSIS. 5. Separation of Lead from Copper and Cadmium. (From Bismuth less satisfactorily.) The solution of the nitrates or chlorides is treated with an excess of sulphuric acid, evaporated to remove the nitric or hydrochloric acid, and the lead determined as sulphate as described on p 174. 6. Separation of Copper from Cadmium. (a) Method of A. W. Hofmann* A. W. Hofmann states that copper and cadmium can be sepa- rated from one another by boiling their sulphides with sulphuric acid (1:5) whereby cadmium sulphide is dissolved while copper sulphide is unacted upon. Hofmann seems to have tested this separation only qualitatively and not quantitatively, but neverthe- less this method is given in all early text-books without submitting any analyses to prove its accuracy. Experiments performed in the author's laboratory showed that in the form proposed by Hofmann this method cannot be used for the quantitative separation of the two metals ; on the other hand, if it is carried out according to the following modifications, excellent results are obtained. Procedure. Sufficient sulphuric acid is added to the solution of the sulphates so that one part of the acid is contained in four parts of the solution. The latter is now heated to boiling, and during the boiling hydrogen sulphide is passed through it for twenty minutes, after which the solution is boiled for fifteen minutes longer. The solution is filtered while hot through a funnel kept filled with carbon dioxide and the precipitate is washed with boiled, hot water to the disappearance of the acid reaction. The copper sulphide thus obtained is easy to filter and wash; it however, always contains small amounts of cadmium, so that the separation must be repeated. The copper sulphide i~, therefore, transferred to a porcelain dish by means of a stream of water from * Ann. d. Chem. und Pharm., 115, 286. SEPARATION OF COPPER FROM CADMIUM. 201 the wash-bottle, where it is dissolved in nitric acid, the solution evaporated to dryness, the dry mass treated with sulphuric acid (1:4) and again evaporated on the water-bath as far as possible to remove the greater part of the nitric acid. After this, without regard to the separated sulphur, the mass is washed with as little water as possible into an Erlenmeyer flask, for every 0.3-0.5 gm. of copper about 150-200 c.c. of sulphuric acid (1:4) are added, and the separation by means of hydrogen sulphide is repeated exactly as above described. The pure copper sulphide that is finally obtained is dried and the copper determined as cuprous sulphide as described on p. 183, or it is dissolved in nitric acid and the solution electrolyzed as described on p. 187. For the cadmium determination, hydrogen sulphide is passed into the cold filtrate, the precipitated cadmium sulphide after being washed is transferred by means of a spatula to a porcelain dish, hydrochloric acid (1:3) is poured over it, the dish covered with a watch-glass and heated on the water-bath until the precipi- tate is dissolved and until the hydrogen sulphide is all expelled. The dish is now placed under the funnel and the cadmium sulphide which remained upon the filter is dissolved by dropping hot hydro- chloric acid (1:3) upon it, finally washing the filter with water. The contents of the dish are evaporated to dryness, the dry mass dissolved in a little sulphuric acid, washed into a weighed porcelain crucible, and treated with 1 c.c. of concentrated nitric acid* and a little more sulphuric acid. After this the contents of the crucible are evaporated as far as possible upon the water-bath, the excess of sulphuric acid removed by heating in an air-bath, and the cadmium determined as sulphate according to p. 190. The above method was tested by Oberer in the author's labora- tory and the following results obtained: * The nitric acid is added to oxidize the fibres of filter-paper; if these are not destroyed they will cause a partial reduction of the cadmium sul- phate. 202 GRAVIMETRIC ANALYSIS. Amount Taken. Found. Difference. Amount Found m Per Cent, of the Theoretical Value. 1. Cu=0.3126gm 0.3130 gm. 4-0.0004 100 12 Cd 2504 " 2506 " + 0002 100 03 2 Cu-0 3126 " 0.3125 " -0 0001 99 97 Cd=0 2504 " . .. . 0.2501 " 0003 99 88 3 Cu=0 3126 " 0.3134 "- + 0.0003 100 25 Cd=0 2504 " 0.2496 " -0.0003 99 63 4. Cu=0.3126 " 0.3120 " -0.0006 99 81 Cd 6 9 59 " 6252 " 0007 C9 88 5 Cu-0 3142 " 3147 " + 0035 100 16 Cd-0 G?59 " 624S ' -0 0011 99 82 6. Cu=0 3142 " 0.3150 " -f 0.0003 100.25 Cd=0.6259 " 0.6240 " -0.0019 99.69 (6) Method of Rivot-Rose. The copper is precipitated as sulphocyanide according to p. 186, and from the nitrate the cadmium is precipitated as sulphide by means of hydrogen sulphide and determined as sulphate according to p. 190. The results are good. (c) Method of Fresenius and Haidlen. (The Potassium Cyanide Method.) The neutral solution containing salts of both metals is treated with potassium cyanide until the precipitate that is first formed redissolves, after which more potassium cyanide is added (about three times as much as was necessary for the precipitation and so- lution of the precipitate) and either ammonium or hydrogen sul- phide is added to the cold solution. The cadmium is precipitated as the yellow sulphide, while the copper remains in solution. * * The copper, however, remains entirely in solution only when more than enough potassium cyanide is present than is required to form the complex salt K 3 Cu(CN) 4 . If the pure potassium cuprocyanide is dissolved in consider- able water and hydrogen sulphide passed into the solution, there is a partial precipitation of Cu 2 S; the more dilute the solution, the more the precipi- tation. By the addition of an excess of potassium cyanide, the precipitation is prevented. A cold, concentrated solution of the above salt is not precipi- tated by hydrogen sulphide (v. Girsewald, Zurich, 1902) SEPARATION OF COPPER FROM CADMIUM. 203 The cadmium sulphide thus precipitated shows a great tendency of passing through the filter-paper even when a "hardened" filter is used, so that it is "salted out." A considerable amount of pure, solid potassium chloride is stirred into the solution, the precipitate is allowed to stand overnight, and in the morning it is filtered through a Schleicher & Schiill "hardened filter." The precipi- tate is washed first by decantation with concentrated potassium chloride solution, it is then transferred to the filter and washed with the same solution. For the cadmium determination this precipitate cannot be used on account of the potassium chloride which adheres to it, and it is not advisable to wash the salt out with water, for in this case a turbid filtrate will be obtained. It is, therefore, dissolved in hot hydrochloric acid (1:3) from a wash- bottle, the solution is evaporated to dryness, the residue dissolved in water, filtered if necessary from separated sulphur, and for every 100 c.c. of the solution 5-7 c.c. of concentrated sulphuric acid are added, and the cadmium is precipitated by passing hydro- gen sulphide into the cold solution. This time the cadmium sulphide is easily filtered. The cadmium is determined as sulphate according to p. 190. The filtrate is evaporated with nitric acid until the odor of hydrocyanic acid can no longer be detected, and the copper is most conveniently determined according to p. 183 as cuprous sulphide. Remark. The results obtained by this method are good, but considerable time and patience are required. (d) By Electrolysis. The experiments of R. Philipp in the author's laboratory show that a very accurate separation can be made, as recommended by Neumann, by electrolyzing the nitric acid solution. The solution, containing not more than 0.2 gm. cadmium, is treated with 4-5 c.c. of concentrated nitric acid or 10 c.c. nitric acid sp.gr. 1.2, and diluted to 150 c.c. in a platinum dish. The anode, a disk electrode, is placed so that it only dips into the liquid a short way. Under these conditions, 0.2 gm. of copper is deposited perfectly free from cadmium, within 12 or 14 hours 204 GRAVIMETRIC ANALYSIS. by a current of 0.2-0.3 ampere and a voltage of 1.9-2.3 volts, with a current of 1 to 1.5 amperes and 2.5-2.6 volts electrode potential, the cadmium is deposited in about five hours. The solution is siphoned off, while pure water is poured into the dish without breaking the current; the dish is finally rinsed with alcohol, dried and weighed with the deposited copper. The solution is treated with sufficient sulphuric acid, evaporated to expel the nitric acid, cooled, diluted and tke cadmium, electrolyzed from cyanide solution as described on p. 189. Remark. If considerably more than 0.2 gm. Cd is present in 150 c.c. of the solution, there is danger of small amounts of cadmium separating out upon the copper during the washing of the deposit, especially when the anode extends well into the solution. This is because the concentration of the acid becomes less during the washing. In analyzing a solution containing a large amount of cadmium and small amount of copper, therefore, it is best to wash at first with 2 per cent, nitric acid rather than with distilled water. The separation requires but a few minutes with a rotating anode or cathode, and a stronger current. DETERMINATION OF ARSENIC AS TRISULPH1DE, ETC. 205 B. DIVISION OF THE SULPHO-ACIDS. Arsenic, Antimony, Tin. SELENIUM, TELLURIUM, GOLD, PLATINUM, TUNGSTEN, MOLYBDENUM, VANADIUM.) ARSENIC, As. At. Wt. 74.96. Forms : A&jSg, As 2 S 5 , Mg 2 As 2 O 7 . I. Determination as Arsenic Trisulphide, As 2 S 3 . For the determination of arsenic in this form, it must be present in its trivalent state, i.e., as arsenious acid or as arsenite. The solution is made strongly acid with hydrochloric acid and the arsenic precipitated in the cold with hydrogen sulphide. The excess of the latter is removed by passing a stream of carbon dioxide through the solution, which is then filtered through a Gooch crucible that has been previously dried at 105 C. The precipitate is washed with hot water, dried at 105 C. to constant weight, and weighed as As 2 S 3 . 2. Determination as Arsenic Pentasulphide, As 2 S 5 , according to Bunsen.* Modified by Fr. Neher.^ The solution, which must contain all of the arsenic as arsenic acid, is treated with hydrochloric acid little by little (it is best to keep the solution cooled by surrounding the flask with ice) until the solution contains at least two parts of concentrated hydro- chloric acid for each part of water. A very rapid stream of hydro- gen sulphide is conducted into this solution (contained in a large Erlenmeyer flask) until it is saturated with the gas, after which * Ann. d. Chem. und Pharm., 192, 305. t Z. anal. Chem., 32, 45; see also Brunner and Tomicek, Monatshefte, 8, 607; McCay, Z. anal. Chem., 27, 682, and J. Thiele, Ann. d. Chem. u. Phann., 265, 65. 206 GRAVIMETRIC ANALYSIS. the flask is stoppered and allowed to stand two hours. The arsenic pentasulphide is then filtered through a Gooch crucible which has been dried at 105 C., and the precipitate is washed completely with water, then with hot alcohol (to hasten the subsequent dry- ing). After drying at 105 C. the precipitate is weighed as As 2 S 5 . It is not necessary to wash it with carbon bisulphide. Remark. If the above directions are conscientiously followed, this method gives faultless result*. If, on the other hand, the directions are deviated from in the slightest respect, the precipitate is likely to contain some arsenic trisulphide, whereby low results will be obtained. If the solution is not kept cool and the hydro- chloric acid is added too rapidly, the heat of the reaction suffices to change a part of the arsenic chloride (this compound probably exists in solution) to arsenious chloride and chlorine, so that on passing hydrogen sulphide into the solution a mixture of arsenic trisulphide and arsenic pentasulphide will be obtained. 3. Determination of Arsenic as Magnesium Pyroarsenate, according to Levol. The solution, which must contain all of the arsenic as arsenate, and have a volume of not over 100 c.c. per 0.1 gm. arsenic, is treated drop by drop, under constant stirring, with 5 c.c. of concentrated hydrochloric acid and then, for each 0.1 gm. of arsenic, there is added 7-10 c.c. of magnesia mixture * and a drop of phenolphthalein solution. !Now, with constant stirring, 10 per cent, ammonia is added from a burette until the phenol- phthalein imparts a permanent red color to the solution, and then enough more of the 10 per cent, ammonia is added to make one- third the volume of the neutralized solution. After standing twelve hours the liquid is filtered through a Gooch or Monroe crucible. The precipitate in the beaker is transferred to the crucible by squirting upon it some of the original solution from a small wash bottle. The precipitate is then washed with 2.5 per cent, ammonia until free from chloride. It is drained as * Prepared by dissolving 55 gms. crystallized magnesium chloride and 70 gms. ammonium chloride in 650 c.c. water and diluting this -to a volume of one liter with ammonia, sp.gr. 0.96. DETERMINATION OF ARSENIC AS MAGNESIUM PYROARSENATE. 207 completely as possible by suction, dried at 100 and heated in an electric oven quite gradually to a temperature of about 400 to 500, until there is no more ammonia evolved. Then the temperature is raised to 800 to 900 and kept there for about 10 minutes. The crucible is then cooled in a desiccator and the precipitate weighed with the precipitate in tjie form of Mg 2 P 2 7 . If an electric oven is not available the crucible with the precipitate is placed in an air-bath (cf. Fig. 11, p. 27), having the bottom of the Gooch crucible [come within about 2-3 mm. of the bottom of the outer crucible. A thin layer of ammonium nitrate powder * is added to the precipitate, which is then heated, at first gently, gradually increasing the temperature until a light-red glow on the outer crucible is obtained, after which the precipitate is allowed to cool in a desiccator and is weighed as MgvAs 2 O 7 . The results obtained are excellent. Remark. The precipitate produced by the magnesia mixture has the formula MgNH 4 AsO 4 +6H 2 O and loses 5^ molecules of water at 102 C.; it has, . therefore, been proposed to dry the precipitate at this temperature and to compute the amount of arsenic present as follows: [MgNH 4 AsO 4 + JH 2 O] : As = p : x. It is, however, impossible to obtain a constant weight at thii temperature, so that the procedure is not to be recommended. If the precipitate is dried at 105-110 C. the salt is obtained almost entirely free from water and at a slightly higher temperature it begins to decompose. The only form in which the precipitate should be weighed is as magnesium pyroarsenate. * Instead of using ammonium nitrate, the crucible may be provided with perforated cover and heated in a current of oxygen. 208 GRAVIMETRIC ANALYSIS. Solubility of Magnesium Ammonium Arsenate, according to Levol. 600 parts of water dissolve 1 part of the salt. In 2 per cent, ammonia it is almost entirely insoluble. Accord- ing to J. F. Virgili,* 1 part of anhydrous magnesium ammonium arsenate dissolves in 24,558 parts of ammonia water. Colorimetric Determination of Arsenic. Small quantities of arsenic, such as are present in wall papers, may be estimated very accurately by means of the Marsh apparatus, comparing the mirror with a series of standards formed with known quantities of arsenic, f It is just as accurate, however, to use the much simpler apparatus used for the Gutzeit test. Treadwell and Comment J allow the arseniuretted hydrogen to react with disks containing silver nitrate and compare the resulting color with a standard which, unfortunately, must be produced freshly with each analysis, as it does not keep very well. Almost equally accurate, and much more convenient is the method of F. Hefti and that of C. R. Sanger and O. F. Black || in which the arseniuretted hydrogen is allowed to act upon mercuric chloride paper. (a) Method of Hefti. In the first place, all the organic matter is destroyed by heat- ing the sample in a tube with fuming sulphuric and nitric acids (see Vol. I), both of which must be free from arsenic. The resulting liquid is evaporated with sulphurous acid on the water- bath in order to reduce the arsenic acid to arsenjous acid and when all the excess of SO 2 has been expelled, the solution is poured into the graduated tube T of the apparatus shown in Fig. 39. In the 100-150 c.c. flask K are placed 6-8 gm. of * Z. anal. Chem., 44, 504 (1905). fC. R. Sanger, Am. Chem. J., 13, 431 (1891); Z. anal. Chem., 38, 137 and 377; G. Lockemann, Z. angew. Chem., 1905, 429 and 491. J This method was given in the former editions of this book. Inaug. Dissert. Zurich, 1907. II Proc. Amer. Acad. Arts and Sciences, No. 8, 1907. COLOR/METRIC DETERMINATION OF ARSENIC. 209 granulated zinc coated with copper* and about 20 c.c. of sul- phuric acid free from arsenic (1 vol. cone, acid + 7 vols. water). At the end of ten minutes all the air should be expelled from the apparatus. The outlet Z)f is now covered with a piece of mercuric chloride paper and kept in place by a small piece of ground glass. According to whether little or much arsenic is present, att or a FIG. 39. part of the solution in T is allowed to flow into the flask K. At the end of twenty minutes the experiment is finished. By com- paring the color of the spot produced on the mercuric chloride * Cf. Vol. I. t For quantities of arsenic under 0.02 mg., the upper diameter of the tube D should be 8 mm. and for larger quantities it should be 16 mm. The upper edge of the tube is ground perfectly flat. 210 GR A *< METRIC ANALYSIS. paper, with the standard spots, the quantity of arsenic present is determined. The disks of mercuric chloride paper are prepared by dipping pieces of pure filter paper into a saturated solution of mercuric chloride and drying them in an oven at a temperature of 60-70. The standards are prepared by carrying out a series of exper- iments with known quantities of arsenic. The spots thus obtained soon lose their color when exposed to moist air, but when dry can be kept in the dark for several days. An older standard is not reliable, but can be used to estimate the approximate quantity of arsenic and then, by making two or three standards with known quantities, the exact amount of arsenic can be determined. The standard solution of arsenious acid used in preparing the scale should contain 20 mg. of As 2 O 3 in a liter. Then 0.05 cm. = 0.001 mg. As 2 3 , 0.1 c.c. =0.002 etc. For smaller amounts of arsenic the above solution is diluted with nine times as much water and thus made one-tenth as strong. (6) Method of C. R. Sanger. Three grams of uniformly granulated, pure zinc is placed in the 30 c.c. evolution flask (Fig. 40) -which is fitted with a stopper holding a thistle tube and a gas delivery tube. The enlargement of the horizontal tube contains a wad of cotton and at the outer end is a piece of thick filter-paper which has been dipped into mercuric chloride solution and dried. Through the thistle tube is poured 15 c.c. of arsenic-free, dilute hydrochloric acid* (1:6), and the hydrogen evolution is allowed to proceed for at least ten minutes to drive out the air from the apparatus. At the end of this time, a measured, or weighed, quantity of the arsenic solution to be tested is added to the flask, which is then nearly filled with water. After a few minutes, the mercuric chloride paper begins to color and at the end of thirty minutes will attain * When hydrochloric acid is used it is unnecessary to plate the zinc with copper. COLORIMETRIC DETERMINATION OF ARSENIC. 211 the maximum coloration. By comparison with standards, the quantity of arsenic present is estimated. Inasmuch as the color of the standards is strongly influenced by moisture, Sanger recommends preserving the strips in a per- fectly dry condition. For this purpose a little phosphorus pent- oxide is placed in a small test-tube, followed by a little cotton, and then the strip of paper- is shoved in with the colored part at FIG. 40. the bottom. The upper end of the paper is moistened with a drop of Canada balsam, after which the tube is closed and sealed. In this way the color can be preserved for several months, although the freshness disappears after a few weeks. See the colored chart at the end of the book, upper row. The color holds a little better if the strips of paper, after being colored, are moistened with strong hydrochloric acid and then dried. Some 6N-HC1 is placed in a small test tube, heated to at least 60, the strips of paper dipped in the acid and allowed to remain there two minutes, washed thoroughly in running water, dried and sealed in tubes as described above. After the drying, the color is a little duller. See the colored chart, middle row. 212 GRAVIMETRIC ANALYSIS. If, however, the strips of colored paper are treated with normal ammonja solution, the spot which was originally red turns black. After drying, such strips are kept in small test-tubes over calcium chloride. These standards are much more permanent than when prepared as above. See the colored chart, bottom row. Remark. To make the strips tmiform as regards the length of the spot and the color, the following conditions must be fulfilled: 1. The evolution flask must be kept the same size and the delivery tubing must be of uniform bore. 2. The same quantity of zinc of the same size must be used in all the tests. 3. The volume and concentration of the acid must remain the same. 4. The wad of cotton must not get too moist. After 10 or 12 experiments it should be renewed. 5. H 2 S, SbH 3 and PH 3 must not be present, as they give a colored spot with the HgCl 2 paper. (c) Electrolytic Determination of Arsenic. * Instead of producing the arseniuretted hydrogen by means of zinc and acid, it may be formed with the aid of cathodic hydrogen. Thorpe passes the arsine through a heated tube and produces an arsenic mirror, but Hefti t allows the gas to react with mercuric chloride paper. In both cases the apparatus devised by Thorpe is used and is shown in Fig. 41. As cathode a perforated cone of thin lead foil is used. This is suspended from the platinum wire that has been fused into the ground-glass stopper of the cathode compartment. The anode consists of platinum foil, two or three centimeters wide, which is wrapped around the porous cell. *Cf. Bloxam., Z. anal. Chem., 1, 483 (1862); T. E. Thorpe, Proc. Chem. Soc., 19, 183 (1903); W. Thomson, Manch. Memoirs, 48, No. 17 (1904); S. R. Tootmann, Chem. Zentr., 1904, 1, 1295; H. J. S. Sand and E. Hackford, Chem. Zentr., 1904, II, 259. t Inaug. Dissert., Ziirich, 1907. ELECTROLYTIC DETERMINATION OF ARSENIC. 213 Procedure. Pure, dilute sulphuric acid (1:7) is poured into the earthenware cell and into the glass outer vessel, E; the level of the acid should be about 2 or 3 cm. from the bottom in the former, and about 0.5 cm. higher in the latter. For the colori- metric determination, the arsenic solution is poured directly into the acid of the inner cell. It must be present as arsenious acid, and, if this is not the case, it must be reduced with sulphurous acid and the excess of the latter expelled by heating. For the FIG. 41. production of mirrors, the air must all be expelled by hydrogen before the arsenic solution is added. The tube C is filled with crystallized calcium chloride. The outlet at D is covered with a disk of mercuric chloride paper (see Method a) and then the circuit is closed. The potential should be about 7 volts and the current about 2 to 3 amperes. The analysis is finished at the end of twenty minutes and the quantity of arsenic estimated by comparing the spot with a standard scale. (See Method a.) If the apparatus is connected with a horizontal delivery tube, Sanger's method can be used. (See Method 6.) Remark. As regards the influence of the cathode material, Thorpe recommends bright platinum foil and Hefti uses lead. 214 GRAVIMETRIC ANALYSIS. Polished platinum does not hold arsenic back, but platinum with a rough surface does, and since bright platinum becomes dull with use, it is easily possible for low results to be obtained. Exper- iments performed by Hefti in the author's laboratory showed that zinc alloyed with a trace of copper or platinum, bright platinum foil and lead did not holdback arsenic when used as the cathode; on the other hand, zinc in the presence of chloroplatinic acid and platinum foil with spongy platinum held back a con- siderable quantity of arsenic. To determine arsenic in a mineral water, 100 c.c., or more if necessary, are evaporated to a small volume in a procelain dish. The resulting solution is acidified with sulphuric acid, reduced with sulphurous acid, the excess of the latter expelled, and the analysis continued by one of the above three methods. Determination of Larger Quantities of Arsenic as Arsine. Method of F. Hefti. In the electrolysis of larger quantities of arsenic it was not possible, in the past, to recover all the arsenic in the form of arseniuretted hydrogen; some arsenic was deposited upon the cathode in the form of the element arsenic and was not trans- formed into arsine by the further action of the electric current. The quantity of arsenic deposited as metal depends upon the potential of the electric current at the electrodes,the temperature, and the concentration of the arsenic solution. At high potentials, low temperature, and low concentration of the solution, the quantity of arsenic deposited becomes zero and the yield of arseniuretted hydrogen is then quantitative. The estimation of the latter is best accomplished iodimetrically. If arseniuretted hydrogen is passed through a solution of iodine in potassium iodide, it is immediately oxidized to arsenic acid in the cold. AsH 3 + 4H 2 O + 4I 2 = SHI + H 3 AsO 4 . If the excess of iodine is titrated with sodium thiosulphate (see lodimetric Methods) it is possible to determine the quantity of iodine that has reacted with the arsine. If T c.c. of 0.1N" iodine were used at the start, and t c.c. of O.lN thiosulphate DETERMINATION OF ARSENIC AS ARSINE. 215 solution were used for the titration, then the quantity of arsenic or arsenic trioxide, present is (T 7 -QX 0.000937 gm. arsenic, (T 7 -0X0.001237 gm. As 2 O 3 . The apparatus necessary is shown in Fig. 42. The decom- position cell is also shown in Fig. 43 and consists of a wide U-tube capable of holding 120 c.c. of solution. The tube is FIG. 42. made in two halves, the edges of the bottom being ground so that they fit tightly together. Between these edges is placed a piece of thin parchment paper, the extending edges of which are folded over one side of the tube. A piece of rubber tubing holds the two halves of the U-tube together and also the parchment paper in place. This tubing is wired tightly in place, taking care that the edges of the parchment paper are also covered by the wire. In one arm of the tube (the anode compartment) which remains open during the whole experi- ment, is suspended a platinum plate electrode as anode, and the other arm (the cathode compartment) is tightly stoppered with a three-holed rubber stopper. Through one hole passes a glass tube containing mercury; at the bottom of this tube 2l6 GRAVIMETRIC ANALYSIS. a platinum wire is sealed in and from this a plate electrode of lead foil is suspended to serve as cathode. The wire from the negative pole of the battery dips into the mercury. Through the second hole in the stopper is passed a gas delivery tube leading to the absorption vessel A. The third hole in the stopper carries a tube that leads to the Erlenmeyer flask E which, in turn, is connected with the empty flask F t and the latter with the rubber tubing shown in Fig. \2. This tubing leads to the hood. Such an arrangement provides for the regulation of the FIG, 43. pressure in the cathode space. If the pressure there exceeds that of the anode space, a part of the arsenic solution will pass into the anode compartment and will be lost in the analysis. If suction is applied at the extreme end of the absorption apparatus, so that bubble after bubble of air passes through the Erlenmeyer, then it is very easy to overcome the pressure in the absorption vessel without having diminished the pressure in the cathode compartment enough to tear the parchment membrane. Procedure. The arsenic solution to be tested must contain all the arsenic in the trivalent condition. In the first place, the anode compartment is filled to within 3 cm. of the top with 10 per cent, sulphuric acid, the arsenic solution is placed in the cathode compartment and this is filled to within 3.5 cm. of the top (in other words, the level in the DETERMINATION OF ARSENIC AS ARSINE. 217 cathode compartment is about 0.5 cm. lower than on the other side of the U-tube) ; the concentration of the arsenic solution in the U-tube, after this dilution with acid, should not ex- ceed 80 mgm. As 2 O 3 in 50 c.c. of solution. The U-tube is placed in ice-water and the gas delivery tube is connected with two ten-bulb absorption tubes, of which only one is shown in the drawing. Into the first absorption tube is now placed an accurately measured volume of tenth-normal iodine solution, and into the second tube, which is not shown in the drawing, 10 c.c. of sodium thiosulphate solution, and about 40 c.c. of water. The purpose of the sodium thiosulphate solution is to catch any iodine that may escape from the first absorption tube. While the apparatus absorption vessels are being filled, the arsenic solution should be in the ice-water, and its tem- perature should be about when the analysis is ready to begin. Gentle suction is started at the end of the second absorption tube, the electric circuit is closed* and the suction is regulated so that bubble after bubble of air slowly streams through the pressure regulator and into the cathode compart- ment throughout the whole duration of the electrolysis. More- over, care is taken that enough ice remains in the cooling bath. When all the conditions are maintained satisfactorily, the liquid in the cell should remain perfectly clear, or at the worst be colored only by a slight brownish turbidity, which eventually disappears. If a black turbidity is formed that settles to the bottom of the U-tube, something has gone wrong and it is useless to continue the experiment. In a normal experiment, the evolution of the arsine is finished in an hour, when not more than 50 mgm. of As 2 O 3 are present. The current is then stopped, the contents of the two absorption tubes (first the iodine and then the thiosulphate solution) are poured into a beaker containing 5 c.c. of a saturated solution of pure NaHCO 3 and the excess of iodine is titrated with O.lN sodium thiosul- phate solution using starch solution as indicator. If on mixing the contents of the two absorption bulbs the solution is decolorized, the titration is finished with O.lN iodine. * A current of 2 to 3 amperes and 7 volts is used. 218 GRAVIMETRIC ANALYSIS. This method can be carried out very easily and gives accurate results in the presence of iron, so that it is suitable for a rapid determination of the arsenic present in iron minerals. Determination of Arsenic in Mispickel. One gram of the finely powdered mineral is fused in a nickel crucible with 6 gm. of sodium c^bonate and 1 gm. of potas- sium nitrate. The resulting melt is extracted with hot water and the residue (Fe 2 O 3 , NiO) washed with hot sodium carbonate solution. To the filtered solution 200 c.c. of water saturated with S0 2 are added to reduce the arsenic, the solution is boiled to expel the excess of SO 2 , allowed to cool, diluted to 500 c.c. with sulphuric acid so that the entire solution contains 10 to 12 per cent, of H 2 SO 4 , and the arsenic is then determined as outlined above, using one-tenth of the solution. Instead of extracting the melt with water, it may be treated with dilute sulphuric acid, whereby all the iron goes into solution. After this solution has been reduced with sulphurous acid, the analysis of an aliquot part gives the same result as when the first procedure is followed. Hefti found 42.67 per cent. As by the former process and 42.73 per cent, by the latter. The mineral analyzed was sup- posed to contain 42.72 per cent, arsenic. ANTIMONY, Sb. At. Wt. 120.2. Forms: Sb 2 S 3 , Sb 2 4 , and Sb. i. Determination as Trisulphide, Sb 2 S 3 . Method of F. Henz. * The best method for the determination of antimony is, in the author's opinion, the following: Hydrogen sulphide is passed for twenty minutes into the cold solution of an antimonite or antimonate, then, without stopping the current of hydrogen sulphide, the solution is slowly heated to * F. Henz, Z. anorg. Chem., 37, 18 (1903). DETERMINATION OF ANTIMONY AS TRISULPHIDE. 219 boiling and the gas passed through it for fifteen minutes more, after which the now dense precipitate is allowed to settle and filtered through a Gooch crucible which has been heated at 280-300 and weighed. The precipitate is washed four or five times by decanta- tion with 50-75 c.c. of hot, very dilute acetic acid into which hydrogen sulphide has been passed, and washed on the filter with the same wash liquid until all chloride is removed. At first the filtrate runs through perfectly clear, but after all the mineral acid has been removed, the filtrate shows a slightly orange tint, owing to an unweighable amount of the antimony sulphide passing through in colloidal solution. As soon as this point is reached the washing is stopped. F. Henz then proceeds as follows: The crucible, after the precipitate has been dried as much as possible by suction, is placed in the tube R, Fig. 44, which is fitted to a drying oven (about 18 cm. long and 10 cm. high; covered with asbestos paper). The tube R is then closed with a rubber stopper that holds a glass delivery tube, and R is pushed into the drying closet until the end of the stopper is reached. To protect the rubber stopper during the subsequent heating, its inner surface is provided with a Rose crucible cover, which is held in place by wrapping the tube a with a strip of asbestos paper. The air is now expelled from the tube by a stream of dry, air-free carbon dioxide* and heated for two hours at 100-130. t In order to obtain accurate results it is neceesary to have the carbon dioxide perfectly free from air. This may be prepared by the use of the Kipp generator as modified by Henz (Chem. Ztg., 1902, 386) see Fig. 45. This differs from the ordinary form of the Kipp apparatus only as regards the siphon tube a; but herein lies a distinct advantage. The apparatus is charged as follows: First of all, pieces of pure marble are placed in the middle compartment, the stop-cock is opened, and water is poured through the upper compartment, until it begins to run out through the stop-cock, which is then closed. By this means all the air has been expelled from the lower parts of the apparatus and it only remains to introduce the hydrochloric acid. To accomplish this, the water is allowed to run out through the siphon while hydrochloric acid (1:4) is poured in at the top of the generator. As soon as carbon dioxide begins to be evolved, the tube a is closed and the apparatus is ready for use. When the acid has become 220 GRAVIMETRIC ANALYSIS. Inasmuch as the tube R extends so far into the drying oven, there is no danger of water condensing in the tube, but it is all expelled as vapor at b. The precipitate is now dry and the air completely expelled from the heating tube. The tube R is now withdrawn a little from the oven, about FIG. 44. FIG. 45. 5 cm., as shown in the drawing, and the temperature is raised to 280-300 and kept there for two hours. Hereby some sulphur is volatilized and collects in the tube R outside the oven. The antimony pentasulphide is also com- pletely changed into the black modification of the trisulphide by this heating.* The crucible is allowed to cool in the stream of too weak, it is removed through the siphon while a fresh supply is poured in at the top; there is no need of taking the apparatus apart during this operation. It is obvious that the same apparatus can be employed to advantage for generating hydrogen or hydrogen sulphide. * According to Paul (Z. anal. Chem., 31, 540 (1892)), the transformation of antimony pentasulphide can be accomplished in his drying oven (shown DETERMINATION OF ANTIMONY AS TRISULPHIDE. 221 carbon dioxide, transferred to the balance case,* and after standing half an hour is weighed. The black antimony trisulphide is not at all hygroscopic. A further heating in the current of carbon dioxide will rarely show any change in weight. (6) Method of Vortmann and Metzel. When antimony is precipitated by hydrogen sulphide from a hot solution which is strongly acid with hydrochloric acid, the sulphide eventually becomes grayish black in color, is crystalline, and can be filtered easily and washed with water without the slightest tendency to pass into the hydrosol condition. The solution, in an Erlenmeyer flask, is treated with concen- trated hydrochloric acid, adding 24 c.c. of the concentrated acid to each 100 c.c. of the neutral solution. It is heated to boiling, and the hot solution subj ected to the action of hydrogen sulphide gas. The Erlenmeyer flask containing the solution is placed in a dish of boiling water and the water in the latter is kept boiling during the precipitation. It is advisable to introduce the hydrogen sulphide gas quite rapidly at first, but towards the end a slow stream is sufficient. The antimony sulphide as it comes down is yellow at first, but as the precipitation proceeds, it becomes redder; grad- ually it becomes heavier and denser, assumes a crystalline form and becomes darker, and finally black in color. The transforma- tion into the crystalline form is hastened by shaking the flask. At first, while the precipitate is of a yellowish color, there is no need of shaking the flask but later on it is very desirable to do so. The shaking, however, should not be too violent, as otherwise some of the precipitate is likely to adhere to the upper portions of the flask and escape the transformation. The duration of the in Fig. 20 of this book) by heating to a temperature of 230. This is per- fectly true, but the transformation takes place more readily at a temperature of 280. It is more difficult to replace the air completely with carbon dioxide in Paul's drying oven and often some white antimony oxide is noticeable in the crucible. * A piece of writing-paper should be rolled up and placed in the tube R, so that the crucible does not come in contact with any of the sulphur sub- limate, on withdrawing it. The crucible is removed with the paper. f Z. anal. Chem., 44, 526 (1905). 222 GRAVIMETRIC ANALYSIS. whole process amounts to from 30 to 35 minutes. Finally a heavy, dense, crystalline precipitate of antimony trisulphide is obtained which settles well and permits a rapid filtration. The solution is diluted with an equal volume of water, which is allowed to flow around the walls of the flask in order to wash down any adhering sulphide. The dilution almost always causes the forma- tion of a slight yellow turbidity. The reason for this is that a little of the antimony is held in solution by the strong acid and as the solution is diluted this is caused to precipitate by the dissolved hydrogen sulphide. The flask is, therefore, once more shaken, placed in the vessel of boiling water and more hydrogen sulphide is introduced. In two or three minutes the solution above the precipitate will become clear. It is filtered through a Gooch crucible, washed with water to remove the acid, then with alcohol, and placed in the drying oven. 2. Determination as Tetroxide, Sb 2 4 (Bunsen). This method is based upon the fact that antimony pentoxide, when ignited at a definite temperature, changes into Sb 2 O4. Bunsen,* who first proposed the method, later abandoned it because his assistant succeeded in volatilizing more than 0.1 gm. of the precipitate by heating it over the blast lamp.f Brunck,{ R6ssing and Henz II have shown, however, that under certain conditions accurate results can be obtained, although they did not specify the exact temperature at which the precipitate should be ignited. If the pentoxide is ignited in a large porcelain crucible over the blast lamp, it is possible to change the antimony pentoxide quantitatively into the tetroxide; if, however, a small, thin-walled, procelain crucible is used, the tetroxide loses oxygen and is transformed into the volatile trioxide, whereby low results are obtained. It is, therefore, purely accidental if exact results are obtained by such a procedure. In 1897, Baubignylf dis- * Ann. Chem. u. Pharm., 106, 3 (1858). t/fcitt, 192, 316 (1878). % Z. anal. Chem., 34, 171 (1895). Ibid., 41, 9 (1902). || Loc. cit. f Compt. rend., 124, 499 (1897) DETERMINATION OF ANTIMONY AS TETROXIDE. 223 covered that antimony pentoxide is converted quantitatively into the tetroxide at a temperature of 750-800 and begins to form the volatile trioxide at a little above 950. The author's assistant, Dr. E. G. Beckett,* has confirmed the work of Bau- bigny. At 750-SOO the transformation is complete and at 1000 it is possible to volatilize 0.35 gm. of the precipitate in about thirty minutes. If these facts are borne in mind, it is possible to get accurate results, although even then the method is less satisfactory than the determination as trisulphide. Procedure. In the majority of cases it is desired to determine the amount of antimony present in a mixture of its tri- and penta- sulphides, or in a mixture of one or the other of the two compounds with sulphur. It is best to proceed as follows: The sulphide of antimony, precipitated from hot solution, is washed first with hot water, then with alcohol, afterwards with a mixture of alcohol and carbon disulphide (in order to remove the sulphur), f again with alcohol and finally with ether, afterwards drying the precipitate. The bulk of the precipitate is separated from the filter and placed upon a watch-glass and the filter is placed in a small porcelain dish and boiled with a little of a freshly prepared solution of ammonium sulphide, stirring constantly with a glass rod. The resulting solution is poured through a small filter into a 30 c.c. porcelain crucible, and the filter is treated repeatedly with ammonium sulphide until it is no longer colored brownish red at the edge of the paper, where it begins to dry; the extrac- tion of the antimony sulphide is then complete. The solution in the crucible is evaporated to dryness and the main part of the precipitate is added. To oxidize the antimony sulphide, Beckett places the crucible, with a dish of fuming nitric acid beside it, under a bell-jar and allows it to stand over night. The vapors of fuming acid slowly oxidize the precipitate in the crucible and in the morning it is possible to complete the oxidation by means of nitric acid (sp.gr. 1.4) without the reaction being too violent. The crucible is then heated on the water-bath until the precipitate becomes white and the greater part of the acid is * Inaug. Dissert. Zurich, 1909. fThiele, Ann. d. Chem. und Pharm., 263, 372. 224 GRAVIMETRIC ANALYSIS. expelled. A little water is added and, with stirring, enough concentrated ammonia to give an alkaline reaction. The con- tents of the crucible are now evaporated to dryness on the water- bath, carefully heated in an air-bath (Fig. 11, p. 27) until no more fumes of sulphuric acid are evolved, and then for half an hour at 800 in an electric oven. After cooling in a desiccator, the crucible is transferred quickly to a glass-stoppered weighing beaker, allowed to stand twenty rfiinutes in the balance case, and then weighed.* The ignition and weighing are repeated until a constant weight is obtained. 3. Determination of Antimony as Metal. Antimony may be deposited from acid solutions by means of the electric current; the metal, however, does not adhere well to the electrode, so that this method cannot be used for its quantita- tive determination. On the other hand, the following method is suitable; it was first proposed by Parrodi and Mascazzini,f then modified by Luckow,| and afterwards improved by Classen and Reiss. According to the experience in the author's laboratory, it is not so accurate as the trisulphide method. If a solution of sodium or ammonium sulphoantimonite, or antimonate, containing not more than 0.3 gm. Sb in a volume of about 140 c.c. is subjected to electrolysis with a current of 1-1.5 amperes at 70 for 90 minutes, the antimony will be deposited upon a platinum dish, which has been gently sand-blasted, as steel- gray, metallic antimony, and the deposit adheres so firmly that it can be dried and weighed without loss. The chief condition for the success of this operation is the absence of polysulphides. In case these substances are present the antimony is incompletely deposited and in some cases not at all, or the deposited antimony may pass into solution, on account of being oxidized to sodium * Sb 2 O 4 is hygroscopic and must be weighed in a weighing beaker, as Finkener, Dexter, and Beckett have all found. t Z. anal. Chem., 18, 587 (1879). $ Ibid., 19, 13 (1880). Berichte, 14, 1629 (1881); 17, 2474 (1884); 18, 408 (1885); 27, 2074 (1894). DETERMINATION OF ANTIMONY AS METAL. 225 antimonite by means of the sodium polysulphide which is formed ut the anode during the electrolysis: 2Sb+3Na 2 S 2 = 2Na 3 SbS 3 . It is necessary, therefore, to prevent the formation .of poly- sulphides during the electrolysis. For this reason Lecrenier* added sodium sulphide to the bath, whereby the polysulphide is transformed into thiosulphate : Na 2 S 2 + Na 2 SO 3 = Na 2 S 2 O 3 + Na 2 S. Ost and Klapproth f carry out the electrolysis with the aid of a diaphragm to keep the polysulphide away from the cathode. It is better, however, to make use of potassium cyanide for this purpose. } Na 2 S 2 + KCN = Na 2 S + KCKS. Procedure. In most cases the antimony is first isolated as the sulphide, which is either precipitated by hydrogen sulphide from acid solution or obtained by acidifying an alkaline solution of the thio-salt. The filtered and washed precipitate, cor- responding to not over 0.2 gm. Sb, is dissolved on the filter by pouring pure sodium sulphide solution (sp.gr. 1.14) over it. * A. Lecrenier, Chem. Ztg., 13, 1219 (1889). f Z. angeW. Chem., 1900, 828. JCf. A. Fischer, Ber., 36, 2048 (1903); Z. anorg. Chem., 42, 363 (1904); Hollard, Bull. Soc. Chem., 23 [3] 292 (1900); F. Henz, Z. anorg. Chem., 37, 31 (1903). A. Inhelder prepares the solution of sodium sulphide as follows: 666 gms. of purest sodium hydroxide (prepared from sodium) are dissolved in 2 liters of water and the solution divided into halves. One-half is placed in a long- necked flask of such a size that the solution just reaches the neck of the flask. The flask is closed with a two-holed rubber stopper and a rapid current of well-washed hydrogen sulphide is introduced through a glass tube 1 cm. wide, keeping out the air as much as possible. When the solution ceases to increase in volume (1000 c.c. of NaOH solution should give 1218 c.c. of sodium NaSH solution). When this is accomplished, the other half of the original sodium hydroxide solution is added. The solution of Na^ thus prepared is colored a pale yellow, and after standing some time, or sooner 226 GRAVIMETRIC ANALYSIS, The solution is caught in a weighed platinum dish with unpolished .inner surface, or in a beaker if a platinum gauze electrode is to be used. After washing the filter with the sodium sulphide solu- tion, the total volume of the liquid in the platinum dish should not be over 80 c.c.; if less than this, enough more sodium sul- phide solution is added. The solution is diluted with 60 c.c. of water and 2-3 gms. of the purest potassium cyanide are added and the liquid is stirred with the ant)de until all the cyanide has dissolved and the solution is well mixed. It is heated to 60-70 and electrolyzed with a current of 1-1.5 amperes and electrode potential of 2-3 volts. After 1.5 to 2 hours all the antimony will be upon the cathode in the form of a firmly-adhering, light- gray deposit.* Now, without breaking the circuit, the electrolyte is siphoned off, while water is added until the current ceases to pass through the liquid (the voltmeter connected as ammeter points to the zero reading). The cathode is removed, washed thoroughly with water, then with absolute alcohol, dried at about 80, cooled in a desiccator, and weighed. The results obtained by this method are invariably too high, as F. Henz f showed in the author's laboratory, the error amount- on shaking, tetragonal crystals of Na 2 S-f 9H 2 O are deposited. The solution keeps indefinitely in a well-stoppered flask; a slight precipitate of black metal sulphide (FeS, NiS, Ag 2 S) is formed after some time. If the sodium sulphide solution is prepared from caustic soda purified by alcohol, the final solution is colored a deep yellow or brown and it contains more foreign sulphides, partly in suspension and partly in colloidal solution. If the suspended sulphides are removed by filtration, a further precipitate will appear on standing. For the antimony determination, the saturated sodium sulphide solution, which has a specific gravity of 1.22, is diluted until the specific gravity is 1.14. According to A. Classen, a satisfactory solution can be prepared by dissolving the purest grade of commercial Na 2 S. Classen formerly recommended that the solution be prepared by boiling a solution of sodium hydroxide which had been saturated with hydrogen sulphide. Such a solution after an hour's boiling, while introducing a stream of hydrogen, contained 69 per cent. NaSH and 31 per cent. Na 2 S (F. Wegelin). * To make sure that the deposition is complete, the liquid may be trans- ferred quickly to a second dish and electrolyzed for half an hour longer. It is seldom that there will be any gain in weight shown by this dish. t Z. anorg. Chem., 37, 31 (1903). DETERMINATION OF ANTIMONY AS METAL 227 ing to about 1.5 to 2 per cent, of the total antimony present. If, however, the antimony deposit is dissolved and precipitated as the trisulphide, the weight of the latter corresponds to the theoretical value, showing that the deposit contained all the anti- mony. If, as A. Fischer* recommends, the deposit is dissolved in alkali poly sulphide, and again electrolyzed with addition of potassium cyanide, the same weight of antimony is obtained as at first, but the antimony is not pure. The error in the analysis is so constant that values not far from the truth will be obtained by subtracting 1.6 per cent, of the weight of antimony deposited electrolytically. This error of the electrolytic antimony determination was first detected by Henz, but has been confirmed by a number of other investigators, including O. M. M. Dormaar,f F. Forster and O. Wolff,| and recently by A. Inhelder. According to Dormaar, Forster and Wolff, the high values are due to the presence of a little sulphur and more oxygen. Forster and Wolff assert that the metal contains from 1 to 1.5 per cent, of oxygen. The error is greater in proportion to the quantity of free sodium hydroxide present in the electrolyte. According to the work of Scheen,[| which was suggested by Classen, the error is due to enclosed mother-liquor and is greater in propor- tion as the electrode surface is rough. Scheen, therefore, rec- ommends a bright electrode surface, or one that is dulled but slightly. A. Inhelder ^f has carefully repeated the experiments of Scheen, using new, polished dishes, but has not been able to confirm Scheen's conclusions, D. Karl Mayr also obtained high values no matter whether the electrode surface was bright or dull. Cleaning the Electrodes. Ost ** recommends heating with a *Berichte, 36, 2348 (1903). t Z. anorg. Chem., 53, 349 (1907). t Z. Elektrochem., 13, 205 (1907). Inaug. Dissert. Zurich, 1910. i| Z. Elektrochem., 14, 257 (1908). [ Inaug. Dissert. Zurich, 1910. ** Z. angew. Chem., 1901, 827. 228 GRAVIMETRIC ANALYSIS. mixture of equal parts concentrated nitric acid and a saturated solution of tartaric acid. The antimony will also dissolve readily by heating with a solution of alkali polysulphide. TIN, Sn. At. Wt. 119.0. Forms: Sn0 2 , Sn. i. Determination of Tin Dioxide, Sn0 2 . Two cases are to be distinguished: (a) The Tin is Present as Metal (in an Alloy). (b) The Tin is Present in Solution. (a) The Tin is Present in an Alloy. Method I. In case the tin is present in an alloy it may be treated accord- ing to the method of Busse* as follows: About 0.5 gm. of the alloy, in the form of borings, is treated hi a beaker with 6 c.c. of nitric acid (sp. gr. 1.5), 3 c.c. of water are slowly added and the beaker is then quickly covered with a watch- glass. As the water is mixed with the acid a violent reaction takes place. When the evolution of nitric oxide (brown vapors on coming hi contact with the air) has ceased, the solution is heated to boiling and diluted with 50 c.c. of boiling water ; the precipitate is allowed to settle completely, then filtered, washed, and dried. After burning the filter, moistening the ash with nitric acid and drying on the water-bath, the precipitate is ignited, at first gently and finally strongly, over the Teclu burner or blast-lamp. It is weighed as Sn0 2 . The tin dioxide thus obtained is never pure; it always contains small amounts of other oxides and must be purified as follows: After weighing, the precipitate is mixed with six times as much of a mixture consisting of equal parts calcined sodium carbonate and pure sulphur, and this mixture is heated in a covered crucible over a small flame until the excess of sulphur is almost entirely removed. This point is easily recognized by there being no longer any odor * Zeit. f. anal. Chem., 17, 53. DETERMINATION OF TIN AS TIN DIOXIDE. 229 of S0 2 and no blue flame of burning sulphur evident between the cover and the crucible. After cooling the melt is treated with a little hot water, whereby the tin goes into solution * as sodium sulphostannate (cf. Vol. I, p. 222), together with some copper and iron. The deep-brown liquid, therefore, is treated with sodium sulphite f solution until it becomes only slightly yellow in color, after which any iron or copper, etc., will be quantitatively precipitated as sulphides. The latter are filtered off and washed, first with water to which a little sodium sulphide has been added and finally with hydrogen sulphide water. As a rule the amount of insoluble sul- phide formed by this treatment is so small that after drying it can be ignited in the air and changed to oxide without introducing any appreciable error. If this weight is subtracted from the original amount of impure stannic oxide, the weight of pure stannic oxide will be obtained. If, however, the amount of impurity present with the residue of metastannic acid should be large, the different metals must be separated according to one of the methods for the separation of the sulpho-bases and the weight of each oxide deter- mined separately and the sum of their weights subtracted from the original weight of the tin dioxide. Instead of determining the amount of impurity present with the tin dioxide, the filtrate from the insoluble sulphides can be acidified with acetic acid and the tin precipitated as yellow stannic sulphide, which, after it has com- pletely settled, is filtered off and changed by careful ignition into tin dioxide, as described on p. 233, and weighed as such. Method II. The alloy is dissolved in nitric acid, the insoluble metastannic acid is filtered off and washed as in the first method. Instead of * Frequently a single fusion with sodium carbonate and sulphur is insuf- ficient; this is recognized by obtaining a sandy residue insoluble in water. In this case the residue is filtered, washed, dried, and the fusion repeated until all the tin is brought into solution. t The sodium sulphite changes the sodium polysulphide to monosulphide, in which copper and iron sulphides are insoluble. 230 GRAVIMETRIC 4N/1LYSIS. drying and igniting the precipitate, however, it is washed into a porcelain evaporating dish, evaporated on the water-bath almost to dryness, and then treated with 1 c.c. of pure sodium hydroxide solution and 10-15 c.c. of concentrated sodium sulphide solution (see foot-note, p. 225). The evaporating dish is covered with a watch-glass, and the dish with its contents is heated for about 45 minutes on the water-bath, whereby all the tintshould pass into solution, and the other metals remain undissolved as sulphides; they are removed by nitration. The filter, upon which the metastannic acid was filtered, still retains some of the precipitate. It is, therefore, laid in a second small evaporating dish, covered with about 1 c.c. of sodium sulphide solution and heated on the water-bath. After half an hour, the tin should all be dissolved. The solution is poured through a small filter and the latter is washed with a little hot water. The two filters are dried, ignited in a platinum spiral, the ash treated with concentrated hydrochloric acid and the resulting solution is added to that obtained by the solution of the alloy in nitric acid. For the determination of the tin, the two solutions containing sodium thiostannate are combined, acidified with acetic acid and boiled to expel the hydrogen sulphide. The precipitated stannic sulphide is filtered off, washed once with water to remove the most of the alkali salts, then transferred back to the original beaker and dissolved in 10 c.c. of 50 per cent, caustic potash, and 1 gm. tartaric acid, these quantities sufficing for 0.1 to 0.2 gm. of tin. (The last traces of precipitate adhering to the filter are dissolved in a very little sodium sulphide solution.) To the solution, pure 30 per cent, hydrogen peroxide (Perhydrol, Merck) is added until the yellow liquid becomes perfectly colorless, then a cubic centimeter in excess. The solution is boiled for about ten minutes to make sure that the oxidation is complete, and that the excess of peroxide is decomposed. As soon as no more bubbles of oxygen are evolved, the solution is allowed to cool somewhat and 15 g. of oxalic acid dissolved in a little hot water are cautiously added. The warm solution is electrolyzed directly as described on page 234. DETERMINATION OF TIN AS TIN DIOXIDE. 231 The precipitated stannic sulphide, as obtained above by acidifying the sodium thiostannate solution, may be ignited in a porcelain crucible and weighed as SnO 2 . The results are usually a little high and the method is not as accurate as the electro- lytic determination. Cf. page 233; /?. Method III. The translator prefers to use a more dilute nitric acid for dis- solving the alloy than was recommended by Busse. It is possible to obtain, in this way, residues of metastannic acid which are fully as pure and the work is not as unpleasant as when the more concentrated acid is employed. Procedure. 0.5 gm. of borings are dissolved in a small beaker with 15 c.c. off nitric acid, sp. gr. 1.2. The solution is evaporated just to dryness on the water-bath, and the beaker removed as soon as this stage is reached. A mixture is prepared of 20 c.c. nitric acid sp. gr. 1.2, and 40 c.c. water and this is divided into three portions. The residue is treated successively with each portion of the above mixture, each time heating to boiling and decanting off the solution through a hardened filter paper. The washing is completed by boiling and decanting with a 1 per cent, solution of ammonium nitrate. As much of the precipitate as possible is allowed to remain in the original beaker, and the total volume of the filtrate should not exceed 150 c.c. The first portions of the filtrate are carefully examined for metastannic acid, refiltered if necessary, and each successive portion removed from below the funnel before new wash water is added. The residue of slightly impure metastannic acid is treated as described under either of the above methods. In case Method II is chosen, however, it is advisable to treat the filters containing the residue from the sodium sulphide treatment with 15 c.c. of hot dilute nitric acid (7 c.c. HNOs, sp. gr. 1.2 and 8 c.c. water) instead of burning and treating with hydrochloric acid as directed above. The resulting solution of the impurities that were originally present in the metastannic acid, is filtered and added to the original nitric acid solution. The filters are dried, ignited 232 GRAVIMETRIC ANALYSIS. and the ash weighed as Sn02. The amount thus found is added to the result obtained from the sodium thiostannate solution. Remark. Sometimes a little metastannic acid is left undis- solved by the treatment with alkaline sulphide. It is not safe, therefore, to discard the filters. (b) Tin is Present in Solution. (a) The Solution Contains Tin only. If the solution contains only tin in the form of stannic salt (chloride or bromide), a few drops of methyl orange are added and then ammonia until the pink color of the indicator is changed to yellow. Ammonium nitrate (obtained by the neutral- ization of 20 c.c. of concentrated ammonia with nitric acid) is added and the solution is diluted to a volume of 300 c.c., heated to boiling, filtered after the precipitate has settled, washed with hot water containing ammonium nitrate,* dried, ignited in a porcelain crucible, and weighed as Sn(>2. Remark. If the solution contains non-volatile organic acids, this method cannot be used for the determination of tin. In this case the tin must be first precipitated as sulphide by means of hydrogen sulphide (cf. p. 233). If the tin is not in solution as stannic salt, but as stannous salt, the solution must be first oxidized by the addition of bromine water until a permanent yellow color is obtained, after which the solution is neutralized with ammonia and treated as above described. According to J. Lowenthal, tin may be precipitated from slightly acid stannic chloride or bromide solutions in the presence of ammo- nium nitrate. Methyl orange is added to the solution and then ammonia until a yellow solution is obtained ; f dilute nitric acid is now added, drop by drop, until the solution just becomes pink again, more ammonium nitrate solution is added (20 c.c. of concentrated ammonia exactly neutralized with nitric acid), the solution is * The ammonium nitrate prevents the formation of soluble, amorphous stannic acid; it "salts out" the precipitate (cf. Vol. I, p. 71). t The excess of acid cannot be removed by evaporation on account of the volatility of stannic chloride. DETERMINATION OF TIN AS TIN DIOXIDE. 233 diluted to 300 c.c., boiled for some time, filtered, washed with water containing ammonium nitrate, dried, ignited, and weighed as SnO 2 . This method is employed when the solution contains small amounts of alkaline earths; they remain in solution. Sodium sulphate can be used instead of ammonium nitrate to "salt out" the tin precipi- tate, but although the tin will be quantitatively precipitated, some sodium sulphate will be also thrown down, so that high results will be obtained. (/?) The Solution Contains, besides Tin, Metals of the Preceding Groups or Organic Substances. In this case, independent of whether the tin is present in the form of stannic or stannous salts, hydrogen sulphide is conducted into the dilute solution until it is saturated with the gas; the solu- tion is allowed to stand until the odor of hydrogen sulphide has almost disappeared and then filtered. The precipitate is washed with a solution of ammonium nitrate (or ammonium acetate), dried, transferred as completely as possible to a porcelain crucible, and the ash of the filter added. The tin sulphide is at first gently heated in a covered crucible to avoid loss by decrepitation, and after- wards in an open crucible until the odor of sulphur dioxide is no longer perceptible. The temperature now is raised gradually until finally the full heat of a good Teclu burner is obtained or the cruci- ble is heated over the blast-lamp. As tin dioxide holds fast to some sulphuric acid with great tenacity, after cooling the crucible somewhat a piece of ammonium carbonate the size of a pea is added, the crucible covered and again heated, after which it is weighed as SnO 2 . The heating with the addition of ammonium carbonate is repeated until a constant weight is obtained. Remark. F. Henz * in testing this method always obtained results which were a little too high. This is due to the fact that it is difficult to wash the stannic sulphide precipitate free from sodium salts. The author recommends, therefore, dissolving the well-washed stannic sulphide precipitate in a little sodium sulphide, transforming this solution into potassium stannioxalate, and determining the tin by electrolysis, according to page 234. * Z. anorg. Chem., 37, 39 (1903). 234 GRAVIMETRIC ANALYSIS. 2. Determination of Tin as Metal. The electrolytic deposition of tin from a solution of the ammonium double oxalate gives excellent results.* It is necessary, however, that some free oxalic acid is always present while the solution is undergoing electrolysis. During the process, ammo- nium oxalate is changed by anodic oxidation into ammonium carbonate and carbon dioxide, (NH 4 ) 2 C 2 4 + O = (NH 4 ) 2 C0 3 + C0 2 , and the solution will smell of ammonia as a result of the hydrolysis of ammonium carbonate. When this point is reached no more tin is deposited. The ammonia often precipitates some stannic acid, which escapes the electrolysis. It is necessary/ therefore, to avoid letting the bath become ammoniacal, and this is best ac- complished by adding a little solid oxalic acid from time to time. Procedure. In the course of an analysis it is usually necessary to precipitate the tin from a solution of alkali thiostannate. This is best accomplished as follows: The thio-salt is decomposed by acidifying with acetic acid, the deposited sulphide dissolved in caustic potash solution, the solution oxidized with hydrogen peroxide, and finally acidified with oxalic acid, all exactly as. described on page 230. The final solution, about 150 c.c. in volume, is heated to 60-70 and electrolyzed with a current of 1 ampere and 3-4 volts potential at the electrodes; from time to time a little solid oxalic acid is added. At the end of about six hours all the tin will have been deposited upon a gauze electrode. The deposit is washed with water, exactly as pre- scribed for nickel on page 136, then with water, dried by holding above a flame, cooled in a desiccator, and weighed. The results are excellent. Remark. If ammonium oxalate is used in place of the potas- sium oxalate, the electrolysis requires more time (eight to ten * Cf. Classen's "Quant. Anal, by Electrolysis" and his " Ausgewahlte Methoden der analytischen Chemie." SEPARATION OF ARSENIC, ANTIMONY, AND TIN. 235 hours) . By the addition of hydroxylamine the duration of the electrolysis is shortened (Engel). F. Henz proposed to prevent the bath becoming ammoniacal by adding sulphuric acid during the course of the electrolysis, but subsequent experiments have shown that it is better to pro- ceed as described above. If too much sulphuric acid is added, all the tin is not precipitated. The chief conditions are the presence of enough oxalate and a slightly acid solution. Separation of Arsenic, Antimony, and Tin from the Members of the Ammonium Sulphide Group. The separation is effected by passing hydrogen sulphide into the acid solution of the above metals whereby arsenic, antimony, and tin are precipitated as sulphides, while the remaining metals remain in solution. From an alloy, or the solid sulpho-salts of the above metals, arsenic, antimony, and tin may be readily volatilized by heating in a stream of chlorine; the chlorides of these three metals are readily volatile, while those of the remaining metals are only difficultly so. Separation of Arsenic, Antimony, and Tin from Mercury, Lead, Copper, Cadmium, and Bismuth. If the metals are all in solution, they are precipitated by means of hydrogen sulphide and the precipitated sulphides after being filtered and washed are treated with alkali sulphide solution. If mercury is present, ammonium polysulphide should be used, but in the absence of this metal sodium polysulphide works better (cf. Vol. I.) If the metals of this group are in the form of an alloy (arsenic and mercury are seldom met with to any extent), the antimony and tin are separated from the remaining metals on treating the alloy with nitric acid. The tin is left behind as meta- stannic acid, insoluble in dilute nitric acid, with the antimony as nearly insoluble antimonic acid. In the presence of tin, all phosphorus and arsenic are thrown down in the insoluble residue as phosphate and arseniate of metastannic acid. The small 2 3 6 GRAVIMETRIC ANALYSIS amount of the latter (and the remaining metals of this group) are JT P-, precipitated by hydrogen sulphide and separated from the copper group by means of alkaline sulphide solution. The separation of tin from the remaining metals of the group can be illustrated by a ^ fSfflU \ practical example. Analysis of Bronzes. Method I. A bronze is an alloy of tin and copper in vary- ing proportions. It almost always contains besides these metals, more or less lead, aluminium, iron, manganese, zinc, and phosphorus. Procedure. About 0.5-1 gm. of the alloy in the form of borings * is placed in a beaker, treated with 6 c.c. of nitric acid, sp. gr. 1.5,f and 3 c.c. of water are added, after which the beaker is immediately covered with a watch- glass. When the reaction begins to diminish, the liquid is heated to boiling, until no more brown fumes are evolved, when 50 c.c. of boiling water are added; the precipitate (con- taining all the tin, the phosphoric acid, and always FIG. 46. small amounts of copper oxide) is allowed to settle completely, * The borings are usually somewhat oily, in which case they should be washed with ether that has been distilled over potash, dried at about 0., and weighed after cooling in a desiccator. The washing with ether is best accomplished in a Soxhlet's fat-extraction apparatus, as shown in Fig. 46. The borings are placed in the extraction-tube, which is filled with ether nearly up to the bend b of the siphon-arm. The tube is then connected with the condenser K. After this from 20 to 30 c.c. of ether are added to the flask and this is heated gently on the water-bath. The ether vapors pass through the wide side-arm to the condenser K, where they are condensed and drop upon the borings. As soon as the ether in the tube has reached the height b f it is siphoned back into the flask, where it is again distilled. All the oil will he removed from the borings in from half an hour to an hour. t See pages 228 and 231. ANALYSIS Of- BRONZES. 237 is filtered, washed with hot water, dried, ignited in a porcelain crucible, and weighed. In this way the weight of the Sn0 2 + P 2 O 5 4- foreign oxide is obtained. In order to obtain the weight of foreign oxide (chiefly copper oxide) the precipitate is fused with a mix- ture of sodium carbonate and sulphur as described on p. 228. The sulphides, remaining after the solution of the melt in hot water, are filtered off, converted into oxides by ignition in the air, and weighed. By subtracting this weight from that pre- viously obtained, the weight of SnO 2 + P 2 5 is obtained. In order to obtain the weight of the SnO 2 a separate portion is analyzed according to the method of Oettel as described below for phos- phoric acid, and the amount of phosphoric anhydride subtracted from the weight of SnO 2 + P 2 O 5 . The oxides obtained by the ignition of the insoluble sulphides are dissolved in a little nitric acid (in case Fe 2 O 3 is present a little hydrochloric acid is also necessary) and the solution of the nitrates is added to the first filtrate from the impure metastannic acid. To this solution an excess of dilute sulphuric acid is added, and it is evaporated on the water-bath as far as possible and then heated over a free flame until dense, white fumes of sulphuric acid are evolved. After cooling, 50 c.c. of water and 20 c.c. of alcohol are added, the precipitate of lead sulphate is filtered off and its weight determined as described on p. 174. The filtrate from the lead sulphate is heated to remove the alcohol and the copper precipitated by means of hydrogen sulphide and weighed as Cu 2 S according to p. 183. In the filtrate from the copper sul- phide the iron, aluminium, and zinc (also manganese) will be found. It is evaporated to a small volume in order to expel the hydrogen sulphide, oxidized by the addition of a few drops of concentrated nitric acid, and the iron and aluminium separated from the zinc by means of a double precipitation with ammonia,* whereby the iron and aluminium are left behind as hydroxides * If considerable zinc is present, the above separation is inexact. In this case the filtrate from the copper sulphide is treated with sodium acetate, heated to 60, saturated with hydrogen sulphide, and the iron and aluminium determined in the filtrate, the zinc in the precipitate. If manganese is present in the alloy, it should be separated from iron and aluminium as described on pp. 149 to 155. 238 GRAVIMETRIC ANALYSIS. and are separated and determined according to p. 107. The zinc is precipitated from the filtrate after acidifying with acetic acid, by passing hydrogen sulphide into the boiling solution. The precipitated zinc sulphide is filtered off, dissolved in hydrochloric acid, evaporated to dryness in a weighed platinum dish, and trans- formed to oxide by heating with mercuric oxide by Volhard's method (cf. p. 142). For the phosphorus determination )ettel * recommends the fol- lowing procedure: From 2-5 gms. of the substance are dissolved, as before, in nitric acid, and the impure metastannic acid with all the phosphorus is filtered off, dried, and transferred as com- pletely as possible to a porcelain crucible. The ash of the filter is added, and the contents of the crucible ignited. After cooling, the substance is mixed with three times as much solid potassium cyanide, the crucible covered, and the contents fused; the stannic oxide is reduced to metal, Sn0 2 + 2KCN - 2KCNO -f Sn, while the phosphoric acid is converted into potassium phos- phate. By skilfully rotating the crucible during the fusion, it is possi- ble to unite the small particles of molten tin into a larger button whereby the subsequent filtration is greatly facilitated. After cooling, the melt is treated with water and filtered. The filtrate is cautiously treated with hydrochloric acid under a good hood and boiled to remove the hydrocyanic acid. It is then saturated with hydrogen sulphide in order to remove traces of copper and tin which almost always remain in the solution. The filtrate is freed from hydrogen sulphide by boiling, made ammoniac al, and the phosphoric acid precipitated as magnesium ammonium phos- phate by the addition of magnesia mixture. After standing for twelve hours, the latter is filtered off, washed with 2J per cent, ammonia water, dried, and changed by ignition to magnesium pyrophosphate, in which form it is weighed. * Chemiker-Zeitung (1896), p. 19. DETERMINATION OF PHOSPHORUS. 239 Ordinary bronzes may be analyzed very nicely in the fol- lowing manner: The alloy is treated with nitric acid as described above, the metastannic acid removed by filtration and the filtrate electrolyzed, using a dull platinum dish as cathode, and a plate as anode, both of which are weighed. The electrolysis is carried out with a current of 1 to 1.2 amperes at about 60 and at the end of two and one-half to three hours the electrodes are washed without breaking the circuit. On the anode will be found all the lead as PbO 2 and on the cathode will be found the copper. The siphoned solution contains the iron, aluminium and zinc, which are determined as above. The phosphorus is determined in a special sample. Remark. The method just outlined will give exact results only when the metastannic acid is purified and the recovered solution of copper and lead nitrates added to the main solution. In the electrolysis, the chief dangers to be feared are having the solution so acid that the copper is not all precipitated, or so dilute that a spongy deposit is obtained. Method II. An excellent method for the analysis of phosphor bronze con- sists in dissolving the alloy as described under Tin, Method III., p. 231, and determining the tin as there described. The copper and lead are determined in the nitric acid solution by electrolysis with a current of 0.2 ampere, the copper being deposited on the cathode and the lead as peroxide on the anode. The electrolysis is usually finished in twelve hours, but it is well to clean the electrodes after weighing the deposits and then to test the solution with the current for an hour or so longer to see whether any lead or copper remains in the solution. Often a little more copper will be found, especially if the solution was a little too acid. During the electrolysis the concentration of the acid gradually diminishes so that eventually all the copper will be thrown down. The iron, aluminium, and zinc remain in solution, and are determined as above outlined. For the phosphorus determination,* one gm. of the borings is weighed into a small beaker and dissolved in 20 c.c. of aqua * Cf. Dudley and Pease, Eng. and R. R. Journ., March, 1894. 240 GRAVIMETRIC ANALYSIS. regia, made by mixing equal volumes of the concentrated acids just previous to use. The beaker is covered with a watch-glass, and, after solution is complete, the contents heated nearly to boil- ing for fifteen minutes. After cooling, 25 c.c. of water are added, and then just sufficient ammonia (sp. gr. 0.90) to redissolve the copper hydroxide* and to produce a deep blue colored solution; thereupon 50 c.c. of colorless ammonium sulphide are introduced. This should be enough*to precipitate the sulphides, and the supernatant liquid should show no blue color. If it does, more ammonium sulphide must be added. The solution is digested at a temperature near the boiling-point for fifteen minutes, the precipitated sulphides of copper and lead allowed to settle, and then filtered into a 300 c.c. Erlenmeyer flask, decanting the clear liquid carefully from the precipitate, and finally throwing the precipitate upon the filter. When the filter has drained the filter and precipitate is returned to the beaker, 50 c.c. of ammonium sulphide wash water (one part colorless ammonium sulphide to three parts of water) are added, and the mixture is heated, and stirred occasionally, for ten minutes; it is then poured upon another filter, washed with 50 c.c. of ammonium sulphide wash water and allowed to drain completely. The total volume should not be over 250 c.c., but it is not necessary to evaporate in case this volume is slightly exceeded. To the filtrate 10 c.c. of mag- nesia mixture are added and the solution shaken. The flask is placed in ice water and allowed to stand with occasional shaking for two hours. The precipitate of magnesium ammonium phos- phate is filtered upon a small filter and washed with ammonia water (one part 0.96 sp. gr. ammonia to three parts water) until nearly free from sulphide. 10 c.c. of hydrochloric acid (one part HC1, sp. gr. 1.20, to four parts water) are placed in the flask, taking care that all of the precipitate adhering to the walls of the flask is dissolved, and then poured through the filter, allowing the solution to run into a No. 1 beaker. The flask and filter are washed with 10 c.c. more of the same acid. 3 c.c. of magnesia mixture are added to the filtrate, which is heated to boiling, removed from the flame, and then treated with ammonia (sp. gr. 0,90 = 10 per cent. NH 3 ) until the latter is present in large excess. * A precipitate of lead or tin hydroxide remains insoluble. SEPARATION OF ARSENIC FROM ANTIMONY. 241 The solution is allowed to stand in ice water for two hours, and is stirred occasionally. The precipitate is then filtered and washed with 2J per cent, ammonia water until free from chlorides, and ignited with the usual precautions, weighing as Mg 2 P2O7. SEPARATION OF THE SULPHO-ACIDS FROM ONE ANOTHER. i. Arsenic from Antimony. (a) Method of Bunsen* Principle. If a slightly acid solution of an alkali arsenate and antimonate is treated with hydrogen sulphide in the cold and the excess of the latter immediately removed by conduct- ing air through the solution, the antimony is quantitatively precipitated as pentasulphide, while the arsenic remains in solu- tion. Procedure. Assume the arsenic and antimony to be present in the solution as arsenious and antimonous acids. Both ele- ments are precipitated by hydrogen sulphide, filtered, and washed with water. The greater part of the precipitate is transferred by means of a spatula to a 200-c.c. porcelain casserole, and the precipitate remaining on the filter is dissolved into the casserole by dropping a solution of hot dilute pure caustic potash upon it. From 3-5 gms. of pure solid caustic alkali are added, and the precipitate dissolves to a clear solution. f The casserole is now covered with a perforated watch-glass. It is placed upon the water-bath, and chlorine is conducted into the solution until all the alkali is decomposed; this takes from one-half to three-quarters of an hour. By this operation the arsenite and antimonite are oxidized to arsenate and antimonate and a small amount of potassium chlorate is formed. Concen- trated hydrochloric acid is now added to the warm solution drop by drop from a pipette until all the chlorate is decomposed and no more chlorine is evolved. The watch-glass is removed, the solu- tion is evaporated to half its volume, and then an equal amount of concentrated hydrochloric acid is added and the solution again * Ann. d. Chem. und Pharm., 192, 305. f If alkaline earths were the only metals present besides the arsenic and antimony, the first precipitation with hydrogen sulphide would be omitted. 242 GRAVIMETRIC ANALYSE. evaporated to half its volume. The contents of the casserole are washed by means of dilute hydrochloric acid into a large beaker, diluted with water to a volume of 600 c.c. and for every decigram, or less of the antimony 100 c.c. of freshly prepared hydrogen sulphide water are added. An orange precipitate of antimony penta sulphide is formed at the end of a short time. A strong current of air (filtered through a wad of cotton) is then passed through the solution without delay until the excess of hydrogen sulphide is completely removed; this usually requires about twenty minutes. In order to avoid loss during this operation a large beaker should be used to contain the solution and it should be covered with a perforated watch-glass. The precipitate of antimony pentasulphide is likely to contain traces of arsenic pentasulphide so that it is dissolved once more in caustic potash and the above operation repeated. The precipitate now obtained will be pure antimony pentasulphide. It is filtered through a Gooch crucible, dried at 280 C. in a stream of car- bon dioxide as described under antimony, and weighed as Sb 2 S 3 .* For the arsenic determination, the combined filtrates are con- centrated somewhat by evaporation, a few drops of chlorine water are added and hydrogen sulphide is passed into the warm solution (being kept on the water-bath) for from six to eight hours, after which it is allowed to cool in a rapid stream of hydrogen sulphide. After allowing the precipitate to settle for twenty-four hours, it is filtered through a Gooch crucible, washed with water, then three limes with alcohol, four times with a mix- ture of pure carbon bisulphide and alcohol (cf. p. 180), and finally three times with pure alcohol. After drying at 110 C., the precipitate is weighed as As 2 S 5 . Remark. If the solution contains no very large excess of hydrogen sulphide, the precipitate will always contain trisulphide, so that it is safer to dissolve it in ammoniacal hydrogen per- * Bunsen weighed the antimony as pentasulphide after washing with car- bon bisulphide. As, however, antimony pentasulphide is likely to be changed to the trisulphide on treating with carbon bisulphide, the above procedure is better. According to Braun, Sb 2 S 3 is reduced to Sb 2 S 2 on long-continued treatment Tfith CS 2 . SEPARATION OF ARSENIC FROM ANTIMONY. 243 oxide * and then to precipitate the arsenic with magnesia mix- ture as magnesium ammonium arsenate, as described on p. 206, weighing it as Mg2As207. Remark. The method gives very accurate results, but con- sumes considerable time. (6) Method of Fred. NeherJ This, in the author's estimation, the best method for the separa- tion of arsenic and antimony, depends upon the fact that arsenic is precipitated from a solution strongly acid with hydrochloric acid by a rapid stream of hydrogen sulphide, while antimony is not. Procedure. Starting with a precipitate consisting of the tri- sulphides of arsenic and antimony, this is dissolved in caustic potash solution and oxidized exactly as described under the previous method. When free from chlorate, the acid solution is washed into an Erlenmeyer flask and cooled by surround- ing the flask with ice. In another flask some concentrated hydrochloric acid (sp. gr. 1.2) is likewise cooled. When both solutions are at C., the arsenic antimony solution is diluted with twice its volume of the strong hydrochloric acid. Into this cold solution a rapid stream of hydrogen sulphide is passed for one and one-half hours. The flask is stoppered up and allowed to stand one to two hours. The As 2 S 5 is filtered through a Gooch crucible and washed with hydrochloric acid (1 vol. water, 2 vols. con- centrated hydrochloric acid) until 1 c.c. of the filtrate after being considerably diluted with water and tested with hydrogen sul- phide shows no precipitation. It is then washed with water, and * For this purpose as much of the precipitate as possible is placed in a beaker, the portion adhering to the filter is dissolved by hot ammonia into the same beaker, and this is warmed until the precipitate has entirely dis- solved. After this, for every 0.1 gm. of As 2 S 5 , 30-50 c.3. of pure 3 per cent. H 2 O 2 are added, the solution heated for some time on the water-bath and then boiled ten minutes. t Z. anal. Chem., 32, 45. 244 GRAVIMETRIC ANALYSIS. finally with hot alcohol. After drying at 110 C., the precipitate is weighed as As2Ss. * The nitrate from the arsenic sulphide is diluted largely with water and saturated with hydrogen sulphide. The Sb2Ss is filtered through a Gooch crucible, dried at 280 C. in a current of carbon dioxide and weighed. (c) The Tartaric *Acid Method. Principle. The separation is based upon the fact that if magnesia mixture is added to a solution of an alkali arsenate and antimonate containing tartaric acid, only arsenic will be pre- cipitated. Procedure. The sulphides are oxidized as described under (a) by solution in aqueous caustic potash and introduction of chlorine. The solution thus obtained is made acid, treated with tartaric acid and an excess of ammonia added. This should not cause any turbidity. If a precipitate is formed, it shows that an insufficient amount of tartaric acid is present. In this case the clear solution is decanted off, the precipitate is dissolved by warming with tartaric acid, and the two solutions are mixed. To the clear, ammoniacal solution, magnesia mixture is added slowly with constant stirring (cf. p. 206. foot-note). After stand- ing twelve hours, the precipitate of magnesium ammonium arsenate is filtered off (it usually contain: 1 a little basic mag- nesium tartrate), washed a few times with 2J per cent, ammonia, dissolved in hydrochloric acid, and reprecipitated by the addi- tion of ah excess of ammonia. After standing for twelve hours mere, the precipitate is filtered, washed with 2J per cent, ammonia, and weighed as magnesium pyroarsenate as described on p. 206, * If the solution was not cold, some arsenic trisulphide will be found in the precipitate. The results are scarcely affected, however, when the precipitate is merely washed with water and alcohol, because the free sulphur is weighed with the sulphide of arsenic. If, however, the precipitate is washed with CS 2 , it is evident that the results will be too low. For the highest degree of accuracy, it is advisable to dissolve the precipitated sulphide in ammoniacal hydrogen peroxide, or in fuming nitric acid, and to deter- mine the arsenic as Mg 2 As 2 O 7 as described on page 206. SEPARATION OF ARSENIC FROM ANTIMONY. 245 Remark. Arsenic can also be separated from tin according to the above method, except that more tartaric acid is necessary to prevent the precipitation of the tin than is the case when an- timony alone is present (cf. p. 255). (d) Method of E. Fischer* Principle. This separation depends upon the ready vola- tility of arsenic trichloride in a current of hot hydrochloric acid gas, under which conditions antimony chloride is not volatile. If the arsenic is present as arsenic acid, which is usually the case, the distillation must take place in the presence of some reducing agent, t Procedure. The apparatus shown in Fig. 47 is used for this determination. In the course of analysis, the arsenic and anti- mony, as a rule, are obtained first in the form of the sulphides, and these are dissolved, as described under (a), in caustic potash solution and oxidized by chlorine. Instead of using chlorine, the alkaline solution may be boiled with hydrogen peroxide or potassium percarbonate. If the latter method is used for the oxidation, the boiling must be continued until there is no further evolution of oxygen. The oxidized solution is transferred, by means of a long- stemmed funnel, to the 500-c.c. distilling flask, A, in which has been placed 1.5 gms. of potassium bromide ;J the solution is diluted in the flask with fuming hydrochloric acid to a volume of about 200 c.c. The receiver, V, consists of a large flask of from * Z. anal. Chem., 21, 266. The process as described is the modification of M. Rohmer, Ber., 34, 33 and 1565 (1901). t Fischer used a ferrous salt, O. Piloty and A. Stock used hydrogen sulphide (Ber., 30, 1649), and Friedheim and Michaelis used methyl alcohol (Ber., 28, 1414). J Instead of the potassium bromide, hydrogen bromide may be used which has previously been prepared by treating 1 gm. of bromine with sulphurous acid. It is not permissible to introduce the bromine into the flask, A, in order to convert it to hydrogen bromide by introducing sulphur dioxide gas into the flask, because it is then possible for bromine vapors to get into the receiver by means of the air which is first expelled from the apparatus, and the bromine would oxidize the volatilized AsCl 3 , and thus interfere with the subsequent determination of the arsenic by pre- cipitation as the trisulphide, or by titration. 246 GRAVIMETRIC ANALYSIS. 1.5-2 liters capacity; it is kept surrounded by a current of cold water coming from the condenser and contains at the start, 800 c.c. of cold distilled water. Then, with the apparatus all connected as shown in the drawing, the distilling flask is heated and its contents partially distilled in a current of hydrogen chloride,* meanwhile constantly passing a little sulphur dioxide into the flask, until at the end of about forty-five minutes, the volume of FIG. 47. liquid in A is reduced to about 40 c.c. The flame is then removed and the T-tube between the two evolution flasks removed in order to prevent liquid from backing up into the wash bottles. The adapter tube which connects the condenser with the receiver is rinsed off and the receiver removed. A new receiver is now placed at the end of the apparatus and a seccond distillation is made in order to make sure that all of the arsenic has been volatilized. t Then, for the determination of the arsenic, the contents of the two receivers are each diluted to a volume of about 1250 c.c., and the excess of sulphurous acid is removed by heating to boiling and passing a stream of carbon dioxide through the liquid as is shown in Fig. 48. When the sulphur dioxide has all been expelled (as can be shown by insert- ing a stopper with delivery tube into the flask so that the escaping vapors can be led into a dilute sulphuric acid solution of deci- uormal permanganate which will be decolorized by sulphur dioxide) , the solution is allowed to cool and the arsenic determined as trisulphide according to the directions on p. 205 and weighed * If there is any tendency to suck back, a little more sulphur dioxide should be introduced. f Rohmer found that as much as 0.15 gm. arsenic was volatilized com- pletely by one distillation. DETERMINATION OF ARSENIC BY TIT RAT ION. 247 as As23 after treatment with 82 (pp. 170, 223), or it may be titrated with iodine. Determination of Arsenic by Titration. The solution is treated with a few drops of phenolphthalem and solid potassium hydroxide is introduced until a permanent pink color is imparted to the liquid. Tho solution is then decolor- ized by the addition of a few drops of hydrochloric acid; 5 gms. of sodium bicarbonate are added, and the solution titrated with decinormal iodine solution as described on p. 688.* The antimony is determined by treating the contents of the distilling flask with 2 or 3 gms. of tartaric acid, washing the FIG. 48. solution into an Erlenmeyer flask, expelling the sulphur dioxide as above, f and determining the antimony gravimetrically by precipitating as the trisulphide according to the directions on p. 218, or it is estimated volumetrically by titration with iodine as described on p. 688. * A blank determination should be made with all the reagents that are to be used, and the iodine solution must be standardized in a solution as dilute as that in which the analysis is made. f The escaping gas will not decolorize a solution of 2-3 c.c. dilute sulphuric acid and one drop of 0.01N. KMnO 4 , when all the SO 2 is expelled. 248 GRAVIMETRIC ANALYSIS. Determination of Arsenic in Commercial Sulphuric Acid. About 30 c.c. of concentrated hydrochloric acid and a little potassium bromide, or hydrogen bromide, are placed in the dis- tilling flask A (Fig. 47), whereupon 50 to 100 gms. of the acid to be tested (the weight is determined by difference) is introduced through a funnel that is fastenecU by means of rubber tubing to the upper end of the delivery tube which enters the flask;* the funnel is rinsed with concentrated hydrochloric acid, and the distillation begun. When the contents of the distilling flask have been concen- trated so that concentrated sulphuric acid remains, the acid is kept hot by means of a small flame until all of the arsenic has been expelled. On account of the high temperature, 1 gm. of arsenic will be driven over in about fifteen minutes. The analysis is finished as described above. Separation of Antimony from Tin. (a) F. W. Clarke' s^ Method. Of all the present known methods for the separation of anti- mony from tin this is probably the most accurate. It depends upon the fact that antimony is completely precipitated from a solution containing oxalic acid, while stannic salts are not. Stan- nous sulphide, however, is decomposed by oxalic acid, forming an insoluble crystalline stannous oxalate, so that the tin must be in the stannic form. * When the concentrated sulphuric acid runs into the flask, it often happens that distillation begins to take place and some of the arsenic would be lost if the flask, A, were left open. t Chem. News, Vol. 21, p. 124. Cf. also Rossing, Zeitschr. fur anal. Chem., XLI, 1. F. Hehz, Z. anorg. Chem., 37, 18 (1903). Vortmann and Metzl, Z. anal, Chem., 44, 525 (1905). SEPARATION OF ANTIMONY FROM TIN. 249 Procedure. In the majority of cases it is a question of sepa- rating antimony from tin after these metals have been separated from the members of the copper group by means of alkaline poly- sulphide; i.e., the tin and the antimony are in the form of their soluble sulpho-salts. The solution of the sulpho-salts, containing not more than 0.3 gm. of the two metals, is placed in a 500-c.c. Jena beaker and treated with a solution of 6 gms. of the purest caustic potash (one-third the sum of the weights of tartaric and oxalic acids to be added) and 3 gms. of tartaric acid (ten times the maximum weight of the two metals). To this mixture 30 per cent, hydro- gen peroxide is added slowly until the yellow solution is com- pletely decolorized; then 1 c.c. in excess is added and the solution is boiled for a few minutes to change any thiosulphate to sulphate and to decompose the greater part of the excess peroxide. As soon as the evolution of oxygen ceases, the solu- tion is cooled somewhat, the beaker covered with a watch-glass, and a hot solution of 15 gms. pure recrystallized oxalic acid is cautiously added (5 gms. for 0.1 gm. of the mixed metals). This causes the evolution of considerable gas (CO 2 + 2 ). Now, in order to completely remove the excess of hydrogen peroxide, the solution is boiled vigorously for ten minutes. The volume of the liquid should amount to from 80 to 100 c.c. After this a rapid stream of hydrogen sulphide is conducted into the boiling solution, and for some time there will be no precipitation, but only a white turbidity formed. At the end of five or ten minutes the solution becomes orange-colored and the antimony begins to precipitate, and from this point the time is taken. At the end of fifteen minutes the solution is diluted with hot water to a volume of 250 c.c., at the end of another fifteen minutes the flame is removed, and ten minutes later the current of hydrogen sulphide is stopped. The precipitated antimony pentasulphide is filtered off through a Gooch crucible which, before weighing and after drying, has been heated in a stream of carbon dioxide at 300 C. for at least one hour. The precipitate is washed twice by decantation with 1 per cent, oxalic acid and twice with very dilute acetic acid before bringing it in the crucible. 250 GRAVIMETRIC Both of these wash liquids should be boiling hot and saturated with hydrogen sulphide. The crucible is heated at 2SO-300 in a current of carbon dioxide (free from air) to constant weight and its contents weighed as Sb 2 S 3 . To determine the tin, the nitrate is evaporated to a volume of about 150 c.c., the excess of oxalic acid nearly neutralized with ammonia and the tin deposited electrolytic ally as described on p. 234. According to Vortmann and Metzl,* antimony may be sepa- rated from tin by passing hydrogen sulphide into a solution containing hydrochloric and phosphoric acids of the proper concentration. (b) Method of H. Rose. Principle. This method is based upon the insolubility of sodium metantimonate and the solubility of sodium stannate in dilute alcohol. Procedure. Both metals are assumed to be present in the form of an alloy. The alloy is treated with nitric acid, whereby the antimony and tin are left in the form of their oxides (cf. pp. 228, 231, and 236). The residue is filtered off, washed with am- monium nitrate water, dried, transferred as completely as possible to a large silver crucible and the ash of the filter added, after which the precipitate is gently ignited. From ten to twelve times as much solid sodium hydroxide and a little sodium nitrate, or better, sodium peroxide, are added and the silver crucible is placed within a larger porcelain one in order to protect it from the flame: the contents are fused and kept liquid for twenty minutes. After cooling, the crucible is placed in a large porcelain dish and its contents treated with hot water until the melt has disintegrated, leaving the insoluble part in the form of a fine meal. One-third of the solution's volume of alcohol (sp. gr. 0.833) is now added, the mixture is well stirred and filtered after standing twelve hours. The residue remaining * Z. anal. Chem., 44, 533 (1905). SEPARATION OF ANTIMONY FROM TIN. 251 L ^ on the sides of the dish is washed onto the filter with dilute alcohol (1 vol. alcohol +2 vols. water). The sodium metanti- monate is washed first with a mixture of 1 vol. alcohol +2 vols. water, then with 1 vol. alcohol +1 vol. water, and finally with 3 vols. alcohol +1 vol. water,* until the filtrate when acidified with hydrochloric acid and tested with hydrogen sulphide no longer gives a yellow coloration (tin sulphide). If considerable tin and little antimony were originally present, a single fusion of the oxides with caustic soda does not afford a complete separation, as the residue of sodium pyroantimonate always contains some tin. It is, therefore, dried, separated from the filter and placed in a silver crucible. The filter is treated repeatedly in a porcelain crucible with fuming nitric acid until the paper is completely destroyed and the excess of acid is then removed by heating in an air-bath. The contents of the porce- lain crucible are subsequently dissolved in a little caustic soda solution and washed into the silver crucible; the water is then removed by heating the silver crucible at first on the water- bath and finally in an air-bath. Ten grams of solid caustic soda are now added, the mixture fused, and the melt treated in the same way as before. The second residue of sodium metantimonate is free from tin. It is dissolved from off the filter by a mixture of hydro- chloric and tartaric acids, f in which it is readily soluble. From this solution the antimony is precipitated by hydrogen sulphide and determined as described on p. 218. For the tin determina- tion, the alcoholic filtrate is gently heated to remove the alcohol, acidified slightly with hydrochloric acid, and the tin precipitated as sulphide by hydrogen sulphide and determined according to p. 233, /?. Remark. If the oxide residue which was first fused with sodium hydroxide and nitre consisted solely of tin and antimony oxides, this method gives very good results. As a rule, however, * A few drops of sodium carbonate solution should be added to all the alcoholic wash liquids. t A mixture consisting of equal volumes dilute hydrochloric acid (1:4) and 5-10 per cent, tartaric acid is used. 252 GRAVIMETRIC ANALYSIS. most antimony and tin alloys contain lead and other metals whose oxides remain to small extent with the tin and antimony on treatment of the alloy with nitric acid, so that the sodium metan- timonate is subsequently rendered impure by the presence of these metals. The antimony determination therefore gives too high re- sults. In this case the method of W. Hampe * should be used. The alloy is dissolved in aqua regia (as described below in the analysis of bearing metal) and the tin and antimony sepa- rated from the remaining metals by means of colorless sodium sulphide. From the solution of the sulpho-salts the tin and anti- mony are precipitated by making barely acid with dilute sulphuric acid; the precipitate is washed and dissolved in a little warm sodium sulphide. After cooling, sodium peroxide is added tc the concentrated solution in small amounts until the liquid becomes colorless, and when treated with more sodium peroxide a distinct evolution of oxygen takes place. By this treatment sodium antimonate is formed; this separates out to some extent, while the tin remains in solution. In order to completely pre- cipitate the antimony from the solution, one-third as much alcohol (sp. gr. 0.833) is added, after which the precipitate is filtered off and treated as above described. Analysis of Bearing Metal. This alloy contains tin ; antimony, lead and a little copper and usually small amounts of iron, bismuth and zinc. One gram of borings are placed in a 400-c.c. beaker and dissolved in 15 c.c. concentrated hydrochloric acid, to which 3 c.c. of concentrated nitric acid are immediately added. The alloy will usually dissolve in the cold after standing a while; when rich in lead, however, it will be necessary to heat for some time on the water-bath. As soon as all has dissolved, the solu- tion (it should be yellow, or greenish yellow if much copper is * Chem. Ztg., 18, p. 1900. ANALYSIS OF BEARING METAL. 253 present) is diluted with 15 times as much alcohol, added in small portions with constant stirring.* After standing for twelve hours, and stirring frequently, the precipitated lead chloride is filtered into a weighed Gooch crucible, washed with absolute alcohol, dried at 150 and weighed. In the nitrate will be found a few milligrams of lead in the presence of antimony, tin, copper, bismuth, iron and zinc.f The alcoholic nitrate is poured into a large, deep porcelain dish | and the alcohol is evaporated off at as low a temperature as possible. It is necessary to avoid evaporating the solution to dry ness as in that case some SnCl 4 will be volatilized. When the alcohol is all gone, 0.1 gm. of potassium chlorate is added, the solution is evaporated to a small volume and then there is added one gm. of tartaric acid and enough caustic potash to make the solution barely alkaline. It is now treated, as recommended by Finkener, with freshly prepared hydrogen sulphide water until no further precipitation takes place. In this way all the Cu, Bi, Fe, Zn and the last of the Pb are pre- cipitated as sulphides (precipitate a) while all the Sn and Sb remain in solution (solution 6). * This stirring is indispensable because lead chloride separates out very slowly from a supersaturated alcoholic solution containing other chlorides. The complete precipitation is best recognized by the fact that no mark is left upon the sides of the beaker when the stirring rod is rubbed against it. f Alloys low in lead are not treated with alcohol in this way. In such cases it is best to decompose the alloy with chlorine gas, as described in the analysis of tetrahedrite on page 359. J In evaporating off the alcohol there is a tendency for the solution to creep over the edges of the dish so that it is advisable to employ a deep dish and to evaporate the liquid in small portions. The separation is complete only when all the tin is in the quadrivalent condition. In driving off the alcohol there is always some stannous chloride formed which must be subsequently oxidized by means of KC1O 3 and HC1. 254 GRAVIMETRIC ANALYSIS. Treatment of Precipitate a. The precipitate is filtered off, washed with hydrogen sulphide water, dissolved in nitric acid (sp. gr. 1.2) and evaporated with hydrochloric acid to remove the nitric acid, and the solution of chlorides diluted so that its acidity corresponds to 1 part HC1 (sp. gr. 1.12) to 25 parts water. The Cu, Pb, and Bi are precipitated as sulphides by hydrogen sulphide, filtered and washed with water containing H 2 S. The filtrate contains the iron and zinc (Filtrate c). The precipitate is dissolved in nitric acid, evaporated with the addition of 4 or 5 drops of concentrated sulphuric acid, and the last of the lead determined as sulphate according to p. 174. Fromjthis filtrate the bismuth is precipitated with an excess of ammonia and determined as Bi 2 O 3 according to p. 179. In the ammoniacal filtrate from the bismuth precipitation, the copper is determined electrolytically, after acidifying with sulphuric acid, according to p. 187, or as cuprous sulphide, according to p. 183. To determine the iron and zinc, the Filtrate c is oxidized by boiling with a few drops of concentrated HNOs and the iron precipitated by an excess of ammonia and weighed as Fe2O3, p. 87. The zinc is determined in this last filtrate by acidifying with acetic acid, precipitating as sulphide and weighing as such, according to p. 143. Treatment of Solution b. To determine the antimony and tin, the alkaline solution is diluted to exactly 250 c.c. in a measuring flask, and after thoroughly mixing, 100 c.c. is withdrawn in a pipette, trans- ferred to a 400-c.c. beaker, acidified with acetic acid, and boiled to expel the hydrogen sulphide. Then 3 gms. of tartaric acid and 6 gms. of purest potassium hydroxide are added, whereby any precipitated sulphide is redissolved. At this point some 30 per cent, hydrogen peroxide is allowed to run slowly into the solution, until the yellow color disappears, then 2 or 3 c.c. in excess are added and the solution boiled a few minutes. Then, SEPARATION OF ARSENIC FROM TIN. 255 for each 0.1 gm. of metal present (Sb + Sn), 5 gms. of pure oxalic acid are added, the solution boiled ten minutes and the antimony separated from the tin as described on p. 248. From the filtrate the tin is determined electrolytically. For this purpose the oxalic acid solution is evaporated to a volume of about 200 c.c. and electrolyzed with a gauze electrode. At the end of six hours the deposition is complete. The electrodes are washed as described on p. 136, dried and weighed. Remark. Rossing's method,* which was recommended in the earlier editions of this book, is not altogether satisfactory. Usually the lead results are too high and the tin too low, on account of the lead sulphide precipitate being contaminated with tin. Separation of Arsenic from Tin. (a) Method of Fred. Neher.^ The moist sulphides are dissolved in freshly-prepared ammo- nium sulphide, evaporated in an Erlenmeyer flask nearly to dry- ness and oxidized with hydrochloric acid and potassium chlorate. From this solution the arsenic is precipitated as sulphide under the conditions described on p. 243. In the filtrate from the arsenic pentasulphide all of the tin is found and can be precipitated as sul- phide after diluting largely with water and passing in hydrogen sul- phide. It is finally changed to the oxide as described on p. 233, /?. (b) Method of W. Hampe. % The precipitated sulphides are dissolved as soon as possible in freshly-prepared ammonium sulphide, the solution is evaporated almost to dryness and oxidized with hydrochloric acid and potassium chlorate in a flask connected with a return-flow condenser. Tar- taric acid and ammonia are then added and the arsenic precipi- tated with magnesia mixture as magnesium ammonium arsenate, as * Z. anal. Chem., 41, 1 (1902). f Ibid. (1893), 32, p. 45. I Chem. Ztg. (1894), 18, p. 1900. So that no arsenic trichloride will be lost by volatilization. 256 GRAVIMETRIC ANALYSIS. described on p. 206, After standing twelve hours, the precipitate is filtered off, washed with 2J per cent, ammonia, and, in order to remove a little magnesia, the precipitate is dissolved in hydro- chloric acid and reprecipitated by the addition of ammonia. After standing another twelve hours, the precipitate is filtered off and again washed with 2J per cent, ammonia. This precipitate can be converted into magnesium pyroar- senate and weighed in this form a described on p. 207. This transformation is somewhat tiresome, however, so that Hampe prefers to dissolve the precipitate in hydrochloric acid once more, to precipitate the arsenic by means of hydrogen sulphide, and then to determine the magnesium in the evaporated filtrate as magnesium pyrophosphate according to p. 66 or p. 67. From the weight of the latter the amount of arsenic can be computed as follows: _2As or gm. arsenc. Separation of Antimony from Arsenic and Tin. (a) Method of Rose. If the metals are present in solution, they are precipitated as sulphides with hydrogen sulphide, heated with fuming nitric acid in a large covered beaker until the sulphur is completely oxidized, washed into a porcelain dish, and the excess of acid removed by evaporation on the water-bath. The almost-dry residue is treated with concentrated sodium hydroxide solution and the contents of the dish are transferred to a silver crucible, after which a little solid sodium hydroxide is added and the contents of the crucible dried in an air-bath. It is then fused * and kept liquid for about twenty minutes by heating over a Teclu burner. After cooling, the melt is disintegrated with water, one-third as much alcohol (sp. gr. 0.833) is added in order to completely precipitate the sodium metantimonate, and after standing twelve hours the * The silver crucible is placed in a larger porcelain orie so as to avoid contact with the flame. GOLD IS PRESENT IN SOLUTION. 257 precipitate is filtered and subjected to the treatment described on p. 251. The filtrate containing all the- arsenic and tin is acidified with hydrochloric acid, whereby stannic arsenate is precipitated. Without filtering, hydrogen sulphide is conducted into the liquid, the precipitated sulphides of tin and arsenic are filtered off, oxi- dized with hydrochloric acid and potassium chlorate, and the arsenic separated from the tin as described on p. 255. (5) Method of Hampe. The moist sulphides are oxidized as described on p. 255, 6, and the arsenic determined in the same way. In the combined filtrates from the magnesium ammonium arsenate the antimony and tin are precipitated by hydrogen sulphide, after making the solution acid. These are separated either according to the method of Clarke (p. 248) or that of Rose (p. 250). SUPPLEMENT TO THE HYDROGEN SULPHIDE GROUP. GOLD, PLATINUM, SELENIUM, TELLURIUM, VANADIUM, MOLYBDENUM, TUNGSTEN. GOLD, Au. At. Wt. 197.2. Gold is always determined as the metal itself. We have three cases to distinguish: 1. The gold is present in solution. 2. The gold is alloyed with copper and silver. 3. The gold is present in an ore. i. Gold is Present in Solution. In almost all cases gold is deposited as metallic gold from its solutions and weighed after filtering and washing. For the deposition of gold the following reducing agents are to be considered: ferrous sulphate, oxalic acid, formaldehyde, and hydrogen peroxide. If the gold is to be precipitated by means of either ferrous sulphate or oxalic add, there must be no free nitric acid present in the solution. If some is present, it must be removed by repeated evaporation with concentrated hydro- chloric acid and the solution then diluted with water. To this 258 GRAVIMETRIC ANALYSIS. dilute solution a large excess of clear ferrous sulphate solution is added, the beaker is covered and its contents are heated for several hours on the water-bath. The precipitate is then filtered off, washed first with water containing hydrochloric acid until the iron is completely removed, and then with pure water. The pre- cipitate is dried, transferred as completely as possible to a porce- lain crucible, the ash of the filter added, and the gold is ignited and weighed. In this way*gold can be separated from almost all metals, even platinum, but not from silver. If silver is present, which is of course never the case in a dilute hydrochloric acid . solution, it is first removed by the addition of hydrochloric acid, the precipitated silver chloride filtered off, and the filtrate treated as above described. For the precipitation of gold by means of oxalic acid, the slightly acid solution is diluted with water, oxalic acid or ammo- nium oxalate is added with a little sulphuric acid, and the covered beaker is allowed to stand forty-eight hours in a warm place. The yellow scales of the deposited gold are filtered off and washed, as above described, with hydrochloric acid and then with water. It is then ignited and weighed. Deposition of Gold by Means of Hydrogen Peroxide (L. Vanino and L. Seemari)* If a gold solution is treated with potassium or sodium hy- droxide solution and then with formaldehyde, or, better still, hydrogen peroxide, the gold is soon precipitated quantitatively, even in the cold. By boiling, the finely-divided gold collects together and assumes a reddish-brown color. The reaction takes place according to the following equation: 2 AuCl 3 + 3H 2 O 2 + 6KOH = 6KC1 + 6H 2 O + 3O 2 + Au 2 . If the gold is deposited by this method from very dilute solu- tions it is obtained in such a finely-divided condition that it passes through the filter. If, however, the solution is boiled until the excess of hydrogen peroxide is completely destroyed, and it is then acidified with hydrochloric acid, the gold can be readily filtered. Gold can be separated from platinum by this method. * Berichte (1C99), 32, p. GOLD ALLOYED WITH COPPER AND SILVER. 259 2. The Gold is Alloyed with Copper and Silver. When gold is present in alloys it is most rapidly and most accurately determined in the dry way. The principle of the method is very simple. If a gold-silver alloy is melted in the air with lead upon a "cupel" (a very porous vessel made of bone-ash)* the lead and copper are oxidized, the oxides fuse and are absorbed by the cupel, while all the gold and silver are left behind in the form of a metallic button, whose weight is obtained. The silver is after- wards separated from the gold by the action of nitric acid which dissolves the silver but leaves the gold behind. If the weight of the gold that is left undissolved is deducted from the weight of the gold-silver button the weight of the silver is obtained. In order to obtain accurate results a number of precautions must be taken. By the cupellation of the alloy some noble metal is always lost and the amount lost increases in proportion to the amount of lead used and the higher the temperature. Furthermore, small amounts of the noble metal are absorbed by the cupel and this amount is greater the smaller the amount of lead used. This second loss amounts to much less than the former one occasioned by the use of too much lead. Consequently, in every gold cupellation an unnecessary excess of had must be avoided. Experience has shown that the richer a gold-silver alloy is in base metal the more lead is necessary for the cupelktion. Furthermore, in the separation of gold from silver by means of nitric acid it is necessary to remember that the separation is only quantitative when the alloy consists of three or more parts of silver to one part of gold. If less than three parts of silver are originally present for one part of gold, it is necessary to add pure silver until this proportion is reached. This operation is known as quartation or inquartation. The separation of the silver from the gold by means of nitric acid is spoken of as parting. If a gold-silver alloy, in the form of foil, which consists of three parts of silver to one of gold, is treated with nitric acid, the latter metal remains behind as a brownish scale; if more silver is present, it is left as a fine powder, unless the acid is made extremely dilute. * According to R. Grund, Oesterr. Z. Berg-Hiittenw., 57, 681, magnesite is better than bone-ash. GRAVIMETRIC ANALYSIS. From what has been said, it is clear that accurate results can be obtained only when the correct amount of lead is present in the alloy that is cupelled, and when the gold and silver are present in the proper proportion; i.e., it is necessary to know the approxi- FIG. 49. mate composition of the alloy before an accurate determination can be made. This is determined by The Preliminary Assay. For this purpose the muffle shown in Fig. 49 is heated to a cherry-red heat, a cupel weighing from 6 to 7 gms.* is placed in the back part of it, the muffle door is closed, and the cupel heated until it has acquired the same color as the muffle. After this 5 gms. of lead are placed upon the cupel, the muffle is closed until * A good cupel will absorb its own weight of litharge. During the cupel- lation about one-tenth of the litharge formed is lost by volatilization, so that the weight of litharge absorbed by the cupel is practically that of the original lead button. Fig. 50 repre- sents a cupel, together with its cross-section. FIG. 50. THE PRELIMINARY ASSAY. 261 the lead is melted and then 0.25 gm. of the accurately-weighed alloy is enveloped in a small piece of lead-foil, placed in the molten lead (with the help of a pair of tongs), and the muffle closed until the alloy has melted and shows a bright upper surface. With the help of an iron hook the cupel is now carefully advanced to about the middle of the muffle and the door should be left open so that there is a ready access of air into the muffle. After about twenty minutes the lead will be all absorbed, which is shown by the "blick." * The hot cupel is then removed from the muffle and after cooling, the color of the button is ob- served. (a) // the button is greenish yellow or darker, it contains less than three parts of silver to one part of gold, in which case from four to six parts of "fine silver" are added (the proper amount can be usually told by the practised eye) and the button is cupelled in a new cupel with 1 gm. of lead. The button now obtained is treated with nitric acid and the residual gold weighed. (b) If the button is pure white, then three or more parts of silver are present to one part of gold. In this case it is imme- diately " parted" and the residual gold weighed. After the approximate amount of gold present has been ascer- tained^ the analysis proper is made, using the amount of lead as indicated in the following table: LEAD TABLE. Amount of Gold Amount of Lead Necessary for the Present in the Alloy. Cupellation of 0.25 gm. of Alloy. 900 " 2 50 gms 800 " 4 00 " 700 " 5.50 " 600 " 6.00 " 500 " 6 50 " 400 or less " ... 8 50 " * The blick is the brightening of the metal which appears when the outer layer of lead oxide that is constantly becoming thinner finally bursts and the bright noble metal shines through. Just before the blick there is a distinct iridescence, so that the point can never be mistaken. f In assay laboratories the approximate gold contents of the alloy is deter- mined by its streak. A fine-grained piece of silicate is blackened with char- -coaL The alloy to be tested is rubbed upon it and the color produced com- 262 GRAVIMETRIC ANALYSIS. The Final Assay. For the definite determination of the gold and silver, two portions weighing exactly 0.25 gm. are taken; the one to serve for the silver determination and the other for the gold. The former is cupelled with the correct amount of lead and the weight of the gold-silver button is determined. If the original alloy was very ^vhite, it contains more than 500 thousandths fine of silver. If the alloy was greenish yellow, it contains 550-750 thou- sandths of noble metal, and silver is present to a considerable extent. If, however, the alloy was a beautiful yellow or reddish yellow, it contains more than 700 thousandths of noble metal and the gold predominates. If, therefore, the alloy was white, once again as much pure silver is weighed out as the amount of gold found to be present by the preliminary assay (inquartated with one part of silver), and this mixture is cupelled with the same amount of lead as the first portion. If the original alloy was greenish yellow, it is inquartated * with two parts of silver; if it was distinctly yellow or reddish yellow it is inquartated with 2J parts of silver. Treatment of the Quartered Gold-Silver Button. The gold-silver button is removed from the cupel with the "button tongs," cleaned with a stiff brush ("button brush"), and hammered upon an anvil to a round disk about 1 mm. thick (Fig. 51, a). This is heated upon a fresh cupel and quickly cooled by placing it upon a piece of brass foil and rolling it between two steel rollers to a long strip (Fig. 51, 6); it is again heated and rolled t up as shown in Fig. 51, c. This little roll is placed in a pared with that obtained from alloys containing known amounts of gold. Afterwards these streaks are tested with dilute aqua regia; alloys containing the same amounts of gold are attacked equally readily. * Cf. p. 259. f By hammering the gold-silver alloy, the metal becomes so brittle that it cannot be converted to a smooth-margined roll, and on the subsequent treatment with nitric acid, little pieces would probably drop off. By again heating the metal and then quickly cooling, it regains its original softness. DETERMINATION OF GOLD IN ORES. 263 little flask (Fig. 52,7), covered with 30-40 c.c. of nitric acid (sp. gr. 1.188) free from chloride, heated to boiling and kept so for ten minutes. The acid is then poured off and replaced by the V FIG. 51. same amount of stronger acid (sp. gr. 1.295) and the above treat- ment repeated. After this acid is poured off, the button is washed by decanting three times with distilled water. The flask is filled with water, covered with an annealing cup (or lack- ing this an ordinary porcelain crucible may be used), and is then quickly inverted (Fig. 51, 77), when the gold will pass into the cup. The flask is removed by first raising its mouth to the level of the water in the crucible and then sliding it off at right angles and skilfully turning the flask right side up. The water is poured off from the gold and the crucible is placed in the back part of the muffle for a short time, whereby the gold is dried and is changed from its former brown and soft condition into a harder, beautiful yellow substance. After cooling, it is weighed. By subtracting the weight of the gold from the weight of the gold and silver together, the amount of silver is obtained. Determination of Gold in Ores. Principle. The very finely ground and sifted ore is mixed in a No. 9 French crucible with lead oxide, charcoal, and some suit- able slag-forming material. The charcoal reduces a part of the* lead oxide to metal which alloys with the noble metal and sinks to the bottom in the form of a button, while the foreign sub- 264 GRAVIMETRIC ANALYSIS. stances should pass into the slag. After cooling, the crucible is broken, the slag is hammered off, the lead button cupelled and the silver-gold button parted in the same way as before. The noble metal should be extracted with as little lead as possible, for with an unnecessarily large amount of lead some gold is lost during the cupellation. The amount of lead reduced from the litharge depends largely upon the nature of the ore. Sulphide ores act strongly reducing, as is shown by the following equations: PbS + 2PbO - SO 2 + 3Pb, FeS 2 + 5PbO = 2SO 2 + FeO + 5Pb. In such cases less charcoal (or in some cases none at all) should be added than would be otherwise necessary to produce the right amount of lead, or in case considerable sulphide is present, it is sometimes necessary to neutralize its action by the addition of oxidizing agents. Reducing ores are recognized by their color: they are gray, bluish-black, or yellow (pyrite, etc.). Reddish-brown ores (Fe 2 O 3 ) usually act oxidizing: Fe 2 O 3 + C = CO-f-2FeO, in which case more charcoal must be added to the charge. The best results .are obtained when the lead button weighs about 18 gms. when obtained from 30 gms. of ore.* In order that such a button may be obtained, it is usually necessary to make a preliminary assay of the ore. But above all, it is neces- sary that the purity of the reagents used should be tested. Testing the Reagents. The ordinary reagents necessary for a gold assay are: i. Litharge (PbO). Litharge, the most important reagent, is a basic flux, for it forms with the silicic acid of the ore a readily fusible silicate; * This amount is usually sufficient; with very rich gold ores 10-15 gms. is enough, while with very poor ores as much as 120 gms. may be used to advantage. Cf. Ricketts and Miller, Notes on Assaying, New York, 1897. REAGENTS FOR GOLD-SILVER ASSAY. 265 at the same time, however, it is a desulphurizing agent, as is shown by the above reaction. The litharge used must be dry and free from minium, for the latter oxidizes silver, carrying it into the slag, so that low results would be obtained in the silver determination. The litharge should be free from silver (which is almost never the case), or its silver contents must be known; this is determined once for all by the following experiment: Litharge 120 gms. Sodium bicarbonate (NaHCO 3 ) 60 " Argols (crude KHC 4 H,O 6 ) 2 " are mixed thoroughly upon a sheet of glazed paper and the mix- ture placed in a No. 9 French crucible and covered with a layer of finely-powdered, dry common salt. The covered crucible is placed in a glowing coke-oven. As soon as the contents of the crucible have reached the state of quiet fusion, the crucible is removed from the fire, its walls are gently tapped by the tongs, and it is lightly tapped upon its bottom in order to knock down any small particles of lead adher- ing to the sides and to make all of the free metal collect together on the bottom in the form of a button. After cooling the crucible is broken, the slag removed from the lead button by hammering it upon an anvil, and it is cupelled upon a cupel weighing only a few grams more than the button itself. The resulting silver button is weighed. The amount of silver obtained must be deducted whenever the corresponding amount of litharge is used in an assay. 2. Sodium Bicarbonate (NaHC0 3 ). 3. Anhydrous Borax (Na^O,). 2 and 3 require no testing. 4. Charcoal. The reducing power is determined as follows: Litharge 60 gms. Sodium bicarbonate 15 " Charcoal 1 gm. 266 GRAVIMETRIC ANALYSIS. are mixed, as in the testing of litharge, in a French crucible No. 9 with a cover of ordinary common salt and fused. After cooling, the weight of the lead button obtained is determined and this ex- presses in terms of lead the reducing power of the charcoal. 1 gm. of charcoal should reduce about 30 gms. lead. 5. Nitre (KW0 3 ) serves as an oxidizing agent. Its oxidizing power expressed in terms of lead is determined: Nitre 3 gms. Litharge .60 " Charcoal 1 gm. Sodium bicarbonate 15 gms. are mixed and fused as before and the weight of the lead button de- termined. If under (4) it was found that 1 gm. charcoal would reduce P gm. lead, and if p gm. of lead were obtained in thi ex- periment, then the difference P p shows the amount of lead that was oxidized by 3 gms. nitre, or the oxidizing power of the nitre. 1 gm. nitre oxidizes about 4 gms. lead. 6. Common Salt. Ordinary table salt is heated in a large Hessian crucible until it melts, and the contents of the crucible are poured into a shal- low iron mould with a raised edge. The solidified crust is finely powdered and preserved in a stoppered flask. After the reagents have all been tested the next step is the Preliminary Assay. Five grams of the finely-powdered and sifted ore are weighed out and mixed with: Litharge 80 gms. Sodium bicarbonate 20 " Borax 5 " placed in a crucible and covered with a layer of common salt. After fusing, cooling, and hammering off the slag, the lead button ob- tained is weighed. Since in an ordinary assay we start with 30 gms. of ore, the GOLD-SILVER ASSAY. 267 weight of the lead button now obtained multiplied by 6 will give the weight of the button from the real assay. We will distinguish four cases : (1) The lead button weighs 3 gms. Consequently the button obtained from 30 gms. of ore would weigh 18 gms. In this case the ore is assayed with the following proportions of flux: Ore 30 gms. Litharge 80 " Sodium bicarbonate 20 " Borax 5 " (2) The lead button weighs less than 3 gms. Evidently the ore acts reducingly, but not enough so to yield a button weighing 18 gms. when 30 gms. of ore are used; it is, therefore, necessary to add charcoal to the flux. Example. Let' us assume that the lead button obtained by the preliminary assay weighed 1 gm. ; then the button obtained from 30 gms. of ore would weigh 6 gms. . In order to obtain a button weighing 18 gms. it is necessary to add enough charcoal to supply 12 gms. of lead. If 1 gm. of charcoal was found to reduce 30 gms. of lead, then it is necessary to add 12 4- 30 gms. = 0.4 gms. of charcoal. (3) The lead button weighs more than 3 gms. In this case the ore has a strong reducing power, and to obtain the lead button of the right weight it is necessary to add some nitre. Example. Suppose the button to weigh 6 gms.; this would mean a 36-gm. button when 30 gms. of ore were used; i.e. 18 gms. too much lead would be produced. We must add, therefore, enough nitre to oxidize this 18 gms. of lead. If the oxidizing power of 1 gm. of nitre was found to be 4 gms. of lead, then 18^4 = 4.5 gms. of nitre must be added to the flux. Remark. Ores which have a very strong reducing power would frequently require the addition of enough nitre to cause the con- tents of the crucible to boil over. In such a case, about 40-50 gms. are placed in a " roasting-dish " and roasted in a muffle, and from this roasted ore the portions are taken for the preliminary and 268 GRAVIMETRIC ANALYSIS. final assays. The results, however, must be expressed in terms of the unroasted ore. (4) There is no lead button formed. The ore is either neutral or possesses an oxidizing action. The assay is repeated, using 1 gm. of charcoal, and from the results now obtained the final assay is based. Final Assay. For the final assay from 30-120 gms. of ore * are taken (accord- ing to the amount of gold present) and the corresponding amount of sodium bicarbonate is added. The amount of litharge also varies with the amount of ore, and in some cases as much as 240 gms. are necessary, although as a rule 80 gms. are sufficient . Otherwise the procedure is exactly the same as in the preliminary assay. The lead button is cupelled and the weighed silver-gold button is parted as described on p. 263. PLATINUM, Pt. At. Wt. 195.0. Platinum is best determined as metallic platinum. The following three cases will be considered : 1. The platinum is present in a hydrochloric acid solution either alone or together with other metals, but other platinum metals are absent. 2. The platinum is present alloyed with gold and silver. 3. The platinum is alloyed with small amounts of the plati- num metals together with small amounts of base metals. I. The Platinum is Present in Hydrochloric Acid Solution Either Alone or Together with Other Metals. The platinum is either precipitated from the solution as ammo- nium chloroplatinate, (NH 4 ) 2 PtCl 6 , which is decomposed by ignition and the residual platinum weighed; or the platinum is precipi- tated as metal by the addition of reducing agents to the solution ; oi* finally the platinum is precipitated as sulphide by conducting hydrogen sulphide into the hot solution and changed to platinum by ignition. The two former methods serve to separate platinum * Usually "assay tons" are used as units in weighing out the ore, and the weights are calibrated in terms of this unit instead of the gram. An "assay ton" contains the same number of milligrams that there are ounces troy to a ton, so that by weighing the button obtained in milligrams, it is ; at once known how many ounces per ton the ore carries. PLATINUM, 269 from most other metals, while the latter serves to separate plati- num only from the metals of the alkali, alkaline earth, and ammo- nium sulphide groups, and not from members of the hydrogen sul- phide group. (a) Precipitation of Platinum as Ammonium Chloroplatinate. The solution, concentrated as much as possible, is nearly neu- tralized with ammonia, an excess of ammonium chloride and consid- erable alcohol are added, and the mixture allowed to stand twelve hours under a glass bell- jar. It is then filtered through an asbestos filter tube 10-15 cm. long, washed with 80 per cent, alcohol until a few drops of the filtrate leave no residue on being evaporated to dryness on a platinum foil. The precipitate is dried by con- ducting, a stream of air warmed to about 90 C. through the tube. After cooling the tube is weighed, a plug of ignited asbestos * is introduced, and the tube is again weighed ; in this way the weight of the asbestos plug is found. A stream of dry hydrogen is now conducted through the tube, and the latter is heated at as low a temperature as possible until no more hydrochloric acid is evolved and all the ammonium chloride has been driven off, after which the tube is cooled in a desiccator and weighed. Instead of filtering the precipitate upon asbestos an un- weighed paper-filter may be used. The moist precipitate is placed together with the filter in a large porcelain crucible so that the apex of the filter-paper points upward, and the covered crucible is then ignited. This ignition must be performed with great care, as otherwise there can be a considerable loss dur- ing the process. At first the precipitate is dried by gently warm- ing the covered crucible, and when the odor of alcohol has disap- peared, the temperature is raised very slowly until the crucible is at a strong red heat. During the whole operation there must be no visible escape of vapors from the crucible. The decomposition is complete when there is no longer a penetrating odor arising from the covered crucible. When this point is reached, the cover (whose under side will be covered with carbon) is removed for the first time and leaned against the crucible and the contents of the latter * Ammonium chloroplatinate decrepitates during the heating. To pre- vent loss of substance it is heated between two asbestos plugs. 270 GRAVIMETRIC ANALYSIS. are ignited with free access of air until the carbon from the filter- paper is completely burned. Often a slight deposit of platinum * will be found in the upper part of the crucible and upon the cover, so that the latter must always be weighed with the crucible. Remark. If it seems likely that the precipitate of ammonium chloroplatinate is contaminated with other substances (e.g. so- dium chloride, etc.) the precipitate can be dissolved in water after it has been washed with alcohol and* dried. The platinum may then be determined, as described on p. 50, by precipitating with mercury, washing with dilute hydrochloric acid and then with water, and finally weighing. The results obtained by this method are satisfactory but some- what lower than the true values; the following process is more accurate : (6) Precipitation of Platinum by Reducing Agents. The solution is freed from any excess of acid by evaporation, placed in an Erlenmeyer flask into the neck of which is ground to fit a return-flow condenser. The solution is neutralized with am- monia, an excess of formic acid and a little ammonium acetate are added, and the contents of the flask after being diluted to a volume of 200 c.c. are heated to about 80 C. on the water-bath until the evolution of carbon dioxide has nearly ceased. The flask is now connected with the return-flow condenser, and its contents boiled for twenty-four hours. The precipitated metal is filtered off, washed with dilute hydrochloric acid, then with water, dried, ignited, and weighed. 2. The Platinum is Alloyed with Gold and Silver. An alloy is seldom found which contains only the above three noble metals ; usually copper is also present. The first step, then, is to separate die noble metals from the others by cupellation with * By means of the dry distillation of the filter, carbon mono >ido is formed, and by the decomposition of the ammonium chloroplatinate chlorine is set free. These two gases act upon the metallic platinum and form volatile com- pounds (PtCl 2 .CO, PtCl 2 .2CO, and 2PtCl 2 .3CO), which, however, are later decomposed by the aqueous vapor. This causes the deposit of piatinum in the upper part of the crucible. In order to avoid loss, a large crucible should be used. PLATINUM. 271 lead as described on p. 259, after which the hammered and rolled button is treated with pure concentrated sulphuric acid. [Nitric acid cannot be used, for some platinum would be dissolved with the silver.] After boiling for ten minutes, the silver will be com- pletely dissolved, provided at least two parts of silver are present for each part of platinum, which is usually the case. If more platinum is probably present than corresponds to the above ratio, pure silver should be added, and the mixture cupelled once more with 1 gm. of lead. After the alloy has been boiled for ten minutes with sulphuric acid it is allowed to cool, filtered, and the treatment with sulphuric acid repeated once again. The metal remaining behind (in the form of a roll or as a powder) is washed three times by decantation with water, ignited, and weighed as described under gold. This gives the weight of the gold and the platinum together, and by sub- tracting this amount from the original weight of the noble metals obtained after cupellation, the weight of the silver is obtained. Separation of Gold from Platinum. Principle. If an alloy of gold and platinum is treated with nitric acid, neither metal is attacked. If, however, the alloy con- tains three parts of silver to one of gold and platinum taken to- gether, and the alloy is treated at first with acid of sp.gr. 1.16 and then with acid of sp. gr. 1.28, the platinum gradually goes into solution with the silver. Procedure. The gold-platinum alloy is cupelled with three times its weight of pure silver and 1 gm. of lead, the resulting button is hammered and rolled, after which it is treated with nitric acid (of the strength stated above), and the residual metal weighed. It is again cupelled with three parts of pure silver, and the same proc- ess repeated. This is continued until a constant, weight is finally obtained for the residual gold ; the third operation usually accom- plishes this. Instead of effecting the separation of the gold from the platinum in this way, the two metals may be dissolved in aqua regia, and the gold precipitated by means of ferrous sulphate, as described on p. 257. This is a good method. 272 GRAVIMETRIC ANALYSIS. According to Vanino, and Seemann,* the separation is effected much more quickly by precipitating the gold from an alkaline solution by means of hydrogen peroxide. In order to determine the platinum, it is precipitated from the boiling acid filtrate by hydrogen sulphide and weighed as metal after ignition in a porce- lain crucible. Analysis of Commercial Platinum, according to Deville and Stas. Five grams of the platinum alloy f are heated for five hours at a temperature of about 1000 C. with ten times as much lead in a crucible of purified gas-carbon; this crucible is embedded in one of clay which is filled with charcoal. After cooling, the lead button is treated with very dilute nitric acid until there is no longer any gas evolved. In this way a solution, A, is obtained, containing about 98.4 per cent, of the lead used, all the palladium and copper, and small amounts of platinum, rhodium, and iron, and a residue, B, consist- ing of a black metallic powder, which is filtered off, and will contain the remainder of the platinum and rhodium with all of the iridium and ruthenium. 1. Treatment of the Nitric Acid Solution A. The lead is precipitated by the addition of slightly more than the theoretical amount of sulphuric acid, and filtered. If the lead sul- phate is pure white, it is washed with dilute sulphuric acid. If it is not absolutely white, it is washed with a solution of ammonium carbonate until it becomes so ; small amounts of lead are dissolved by this operation. This last wash liquid, therefore, is concen- trated, to precipitate the lead carbonate, filtered, and after making acid with hydrochloric acid, added to the main filtrate. The solution is evaporated to about 100 c.c., and when cold is poured into a saturated solution of ammonium chloride. The mixture is heated to boiling and allowed to cool again. The am- monium chloroplatinate is filtered off and washed with a saturated solution of ammonium chloride; in this way, the greater part of the platinum is obtained. * Berichte 1899, p. 1971. f All commercial platinum contains other platinum metals, especially iridium. PLATINUM. 273 The filtrate from the platinum precipitate is boiled with formic acid and ammonium acetate as described on p. 270, b. In this case the remainder of the platinum, the palladium, and the rhodium will be precipitated. These metals are filtered off, and the copper and iron are determined in the filtrate in the usual way. The formic acid precipitate (consisting of a black metallic powder) is dried and fused with potassium bisulphate in a porcelain crucible. The melt is treated with water, the solution decanted from the unattacked platinum and washed alternately with ammonium carbonate and nitric acid (to remove traces of lead sulphate), then with dilute hydrofluoric acid, and finally with water; it is then dried and weighed. The filtrate from the pJatinum contains palladium and rhodium. The former is precipitated by the addition of mercuric cyanide, and boiling until the odor of hydrocyanic acid has disap- peared. The voluminous, yellowish- white precipitate of palladouo cyanide is washed first by decantation, then upon the filter, dried, and ignited at first cautiously and then strongly over the blast until the paracyanide is completely destroyed; finally heating in a current of hydrogen (as in the case of copper sulphide, p. 183) in order to reduce any palladium that has been oxidized by the previous treat- ment. As soon as the flame is removed, the supply of hydrogen is at once cut off in order to prevent its being absorbed by the metal. The palladium is weighed after cooling. The rhodium is precipitated from the filtrate by means of formic acid, as before, and the deposited metal is dried, ignited in a stream of hydrogen, allowed to cool in the gas, and then weighed. 2. Treatment of the Residue B. The washed residue is warmed with dilute aqua regia (in this case 2 vol. nitric acid, 8 vol. hydrochloric acid, and 90 vol. water), and in this way solution C is obtained, which contains the rest of the lead, platinum, and rhodium, and residue D, consisting of lamellae of iridium and ruthenium. 3. Treatment of the Solution C. After evaporating to a small volume, the lead is removed by sul- phuric acid, the solution again evaporated, taken up in hydro- chloric acid, and the platinum present is precipitated by pouring 274 GRAVIMETRIC ANALYSIS. into a cold saturated solution of ammonium chloride exactly as described under 1, p. 272. The platinum precipitate contains a little rhodium, and after washing it with a saturated solution of ammonium chloride, it is placed at one side for the time being. The filtrate, together with the wash water, is evaporated until more platinum and rhodium separate out on cooling, and this pre- cipitate is filtered off and washed as before. Both filters, together with the precipitates, are now placed in a small, weighed porcelain dish, dried, and reduced at as low a tem- perature as possible, in a stream of illuminating gas, and heated somewhat in a muffle so as to remove the carbon from the filter. The metal thus obtained (platinum -f rhodium) is weighed. For the separation of the rhodium from the platinum, the spongy metal is heated in the same dish with potassium bisulphate, gradually rais- ing the temperature until a dull-red heat is obtained. After cool- ing, the melt is extracted with water, the unattacked platinum (it may still contain small amounts of rhodium) is filtered off, w r ashed, and again fused with potassium bisulphate. This operation is repeated until the rhodium is completely extracted, which is known by the melt showing no yellow color after ten minutes. The platinum is washed, ignited, and weighed as described under 1. The combined filtrates from the platinum contain rhodium and still a little platinum. Ammonia, acetic and formic acids, there- fore, are added once more, and the solution boiled for a long time. The precipitated metal is filtered off, ignited, weighed, afterward fused at a distinct red heat with potassium bisulphate, and the cold melt extracted with water. If a residue remains after this treatment, it is filtered off, weighed, and treated with dilute aqua regia. If it dissolves, it is platinum; if it does not, it is rhodium. The filtrate from the ammonium chloroplatinate, which con- tained some rhodium, is diluted, formic acid and ammonium ace- tate are added, and it is gently boiled for two or three days in an Erlenmeyer flask connected with a return-flow condenser. The liquid evaporates somewhat in spite of the condenser, and the evap- orated part is replaced from time to time with a dilute solution of ammonium formate. In this way small amounts of platinum and rhodium are precipitated, which are filtered off and separated by PLATINUM. 275 fusion with potassium bisulphate as before. In the filtrate there are likely to be present traces of platinum, rhodium, and iron. The iron is first removed by the addition of chlorine water and afterward ammonia; the ferric hydroxide is filtered off, ignited, and weighed. In order to remove the last traces of platinum and rho- dium, this last filtrate is evaporated to dryness, the residue heated with nitric acid in order to remove the ammonium chloride com- pletely, and then boiled for a long time with formic acid and am- monium acetate. The traces of metal thus obtained are washed with hydrofluoric acid and added to the main portion of platinum and rhodium. 4. Treatment of the Residue D. The undissolved, gray lamellae consisting of iridium, ruthenium, and small amounts of iron obtained by the action of dilute aqua regia, are filtered off, dried, ignited in an atmosphere of hydrogen or illuminating gas, and weighed. The weighed metal is then fused in a pure gold crucible with potassium nitrate and carbonate. For this purpose, a previously melted mixture of 3 gms. potassium nitrate and 10 gms. potassium carbonate is placed in the crucible, the metal added, and the mix- ture heated for two hours at a dull-red temperature. In this way the ruthenium is changed completely into water-soluble potassium ruthenate, K 2 Ru0 4 , and the iridium is oxidized to Ir 2 O 3 ; the latter forms, to some extent, a soluble compound with the alkali. The melt is treated with water, and the solution, together with the suspended Ir 2 O 3 ,* is poured into a stoppered cylinder, the pre- cipitate allowed to settle, and the clear liquid decanted off into a retort. The residue remaining in the cylinder is covered repeatedly with a dilute solution of sodium hypochlorite and sodium carbonate, until the yellow color is completely removed. The decanted liquid is added to the main solution in the retort. This solution con- tains all the ruthenium and a part of the iridium. It is saturated with chlorine in the cold, distilled, and the distillate received in a mixture of alcohol (distilled over potassium) and pure hydro- chloric acid. * Cf. W.Palmaer, Z. anorg. Chem., 10, 332 (1896). 276 GRAVIMETRIC ANALYSIS. After the distillation is complete, the alcoholic distillate is evap- orated to dryness and the ruthenium chloride thus obtained is reduced to metal by heating in a stream of hydrogen. After weighing, the purity of the ruthenium is tested. It must dissolve completely in a concentrated solution of sodium hypochlorite. The liquid remaining in the retort is evaporated to a small vol- ume, the insoluble residue remaining in the cylinder (that was washed with sodium hypochlorite anc^ sodium carbonate) is added, and the mixture boiled with caustic soda solution, with the addi- tion of a little alcohol, until all of the indium is precipitated. The dark-blue precipitate, consisting of iridium oxide and small amounts of ferric hydroxide, is filtered off, washed, and strongly ignited. The ferric oxide contained in it is then extracted with hydrochloric acid containing some ammonium iodide, and the re- sidual iridium oxide is washed successively with water, chlorine water, and hydrofluoric acid in order to remove gold that came from the crucible and silicic acid from the caustic soda. It is then ignited in hydrogen and the iridium weighed. The iron present in the hydrochloric acid extract is precipitated as ferric hydroxide, ignited, and weighed. Its purity is tested by heating in a stream of hydrogen and hydrochloric acid, to see if it can be completely changed to ferrous chloride and volatilized as such. F. Mylius and F. Forster* have recommended that platinum be tested for small amounts of impurity by taking three separate portions each weighing 10 gms. The first portion is tested for palladium, iridium, and ruthenium according to the lead pro- cedure just described of Deville and Stas. The second portion serves for the iron determination; the metal is dissolved in aqua regia, the platinum metals precipitated by formic acid, and the iron determined in the filtrate. In the third portion, rhodium, silver, copper, and lead are determined by volatilizing the plati- num as PtCl 2 CO at 238 C. (temperature of boiling quinolin) in a stream of carbon monoxide and chlorine, and determining the above substances in the residue. Remark. The determination of the iron in a separate portion is to be recommended, for in the lead procedure some iron is always obtained from the carbon crucible. * Berichte 1892 TJ SELENIUM. 277 SELENIUM, Se. At. Wt. 79.2. Selenium is usually determined as the element itself. Three cases are to be considered: I. The selenium is present as alkali selenite or as selenious acid. II. The selenium is present as alkali selenate or as selenic acid. III. The selenium is present as potassium selenocyanide. I. The selenium is present as selenite or as free seleinous acid. The solution is acidified with hydrochloric acid, saturated with sulphur dioxide gas, boiled, filtered through a Gooch crucible, and washed first with water, then with alcohol. The residue is dried at 105 C. and weighed. Remark. The precipitation of selenium by sulphur dioxide is always quantitative whether the solution is concentrated or dilute, whether it contains much or little free acid. This latter fact is of importance in the separation of selenium from tellurium, for the latter element is not precipitated by sulphur dioxide when consid- erable hydrochloric acid is present (cf. p. 279). Phosphorous acid does not precipitate selenium from cold, dilute, strongly acid solutions; this fact is made use of in the separation of selenium from mercury (cf. p. 281). II. The selenium is present as alkali selenate or as free selenir acid. As selenium in the form of selenic acid is not precipitated by sulphur dioxide, phosphoric acid, or hydrogen sulphide, it must be first reduced to selenous acid by long-continued boiling with hydrochloric acid (cf. Vol. 1); the above procedure is then followed. III. The selenium is present as potassium selenocyanide. The solution, concentrated as much as possible, is treated with hydro- chloric acid, boiled, allowed to settle, and the precipitate filtered through a Gooch crucible, dried at 105 C., and weighed. Remark. From very dilute solutions- of potassium selenocy- anide, selenium separates out only very slowly according to this method; it is therefore advisable to concentrate the solution as much as possible, but when this cannot be done, the boiling with hydrochloric acid should be continued for some time and the liquid allowed to stand before filtering. 278 GRAVIMETRIC ANALYSIS. In practice, selenium is obtained usually in none of the above forms, but as impure selenium (selenium sponge) or as selenide, and by the treatment of these substances one or the other of the above selenium compounds is obtained. If the selenium or selenide is acted upon by concentrated nitric acid,* or aqua regia, all of it is dissolved in the form of selenous acid (not selenic acid). After evaporating the solution several times with hydrochloric acid in order to temove the excess of nitric acid, the selenium is precipitated by sulphur dioxide as described under 1. If the finely powdered selenium or selenide is intimately mixed witji two parts sodium carbonate and one part potassium ni- trate, placed in a nickel crucible, covered with a layer of sodium carbonate and potassium nitrate and heated gradually until it fuses, all the selenium forms alkali selenate and on ex- tracting the melt with water it goes into solution ; in this way it is separated from most of the remaining oxides. The solution, however, often contains small amounts of lead. In order to re- move the latter, the nitrate is treated with hydrogen sulphide, and again filtered; the solution is freed from hydrogen sulphide by boiling, strongly acidified with hydrochloric acid, boiled until no more chlorine is evolved and the selenium is precipitated by sulphur dioxide according to II. Remarks. The mixture must be heated very slowly, as other- wise some selenium is likely to be lost by volatilization. Selenium and very many selenium compounds may be satisfac- torily determined as follows: The dry, finely powdered sponge is fused at as low a temperature as possible in a current of hydrogen f with twelve times as much potassium cyanide. After the mass has fused quietly for about fifteen minutes it is allowed to cool in hydrogen. It is then extracted with water, the solution is heated to boiling, and analyzed according to III. * Mercury cyanide is unacted upon by nitric acid, but is dissolved by aqua regia. f A Rose crucible (Fig. 37, p. 185) is used, or a round-bottomed flask with a long neck made of difficultly fusible glass, from which the air is replaced by hydrogen. In the latter case the delivery-tube must be so wide that the neck of the flask is nearly filled with it. SELENIUM AND TELLURIUM. 279 It is necessary to boil the solution of potassium selenocyanide before acidifying it, for small amounts of potassium selenide (K 2 Se) are almost always present, and on acidifying with hydro- chloric acid this is decomposed with evolution of hdyrogen selenide. On boiling, the potassium selenide is changed to potassium selenocyanide according to the equation: K 2 Se + KCN + H 2 O + O = 2KOH + KCNSe. TELLURIUM, Te. At. Wt. 127.5. Tellurium is usually determined as the element itself. If sulphur dioxide is conducted into a hydrochloric acid solu- tion containing tellurous acid, black tellurium is quantitatively precipitated, provided the solution does not contain too much acid. If tellurous acid is dissolved in 200 c.c. of hydrochloric acid, sp. gr. 1.175, no tellurium will be precipitated on passing sulphur dioxide into the cold solution. If, however, the solution is diluted with an equal volume of water and sulphur dioxide is passed into the boiling solution, all the tellurium will be precipitated. The pre- cipitate is filtered off, washed with water until free from chlorides, then with alcohol, dried at 105 C. and weighed. The oxidation of the tellurium during the drying is so slight that it can be disre- garded.* Separation of Selenium and Tellurium from the Metals of Groups III, IV, and V. By conducting sulphur dioxide into the solution fairly acid with hydrochloric acid, the selenium and tellurium will be quantita- tively precipitated while the other metals remain in solution. * The presence of nitric acid prevents the complete precipitation of the tellurium by means of sulphur dioxide and similarly the presence of sulphuric acid is harmful. To remove nitric acid, sodium chloride is added and the solution evaporated to dryness repeatedly with hydrochloric acid. According to Brauner the addition of sodium chloride is absolutely necessary, as other- wise an appreciable amount of tellurium will be volatilized as chloride. A. Gutbier (Ber. 34, 2724 (1901) ) reports that all these difficulties are over- come by precipitating tellurium from a hot solution by means of hydrazine hydrate or hydrazine hydrochloride, but not the sulphate. See also P. Jannasch and M. Miiller. Ber. 31, 2393 (1898). 28o GRAVIMETRIC ANALYSIS. Separation of Selenium and Tellurium from the Metals of Group II. (a) From Copper, Bismuth, and Cadmium. Sulphur dioxide is passed into the boiling solution, acid with hydrochloric acid, whereby all of the selenium and tellurium and usually some of the bismuth are precipitated. The precipitate after being washed is dissolved in nitric acid, the solution evapo- rated to dryness, taken up in concentrated hydrochloric acid, diluted with a little water and precipitated with hydrogen sul- phide. The precipitate, consisting of the three sulphides, is washed and then treated with sodium sulphide solution whereby selenium and tellurium pass into solution while the bismuth remains behind as its brown sulphide and is filtered off. The solution containing the selenium and tellurium is acidified with nitric acid, carefully evaporated to dryness and the residue boiled with 200 c.c. of hydrochloric acid, sp. gr. 1.175, until there is no longer any evolution of chlorine. The deposited sulphur is then filtered off through a Gooch crucible, and the filtrate satu- rated with sulphur dioxide gas ; all the selenium is in this way pre- cipitated. The latter is filtered off through a Gooch crucible and washed successively with a mixture of 90 vol. HC1 (sp. gr. 1.175) and 10 vol. water, dilute hydrochloric acid, and finally absolute alcohol. The precipitate is dried at 105 C. and weighed. The filtrate is diluted with an equal volume of water and the tellurium precipitated by passing sulphur dioxide into the boiling solution. This precipitate is washed with water until free from chlorides, then with absolute alcohol, after which it is dried at 105 C. and weighed. Remark. The above method is suitable for the separation of selenium and tellurium from small amounts of bismuth, but does not effect the separation of selenium (and tellurium) from copper. In this case, more or less copper selenide is formed according to the conditions, and this compound is not decomposed quantita- tively by sodium sulphide.* In this case, the method of B. Brauner and B. Kuzmaf may be used. * Cf. E. Keller, J. Am. Chem. Soc., 19, 771. t Berichte, 1907, 3362. SELENIUM AND TELLURIUM. 281 The tellurium and selenium are precipitated in a pressure flask, by means of SO 2 , the precipitate, which is contaminated with copper, antimony and bismuth, is filtered (using a Gooch crucible) washed, dissolved in nitric acid, the solution evaporated to dryness and the residue taken up in caustic potash solution (1:5). The alkaline solution is placed in an Erlenmeyer flask upon a water-bath, and little by little 4-6 gm. of ammonium persulphate are added, whereby the potassium tellurite is oxidized to potassium tellurate and the selenite to selenate. When all the persulphate has been introduced, the solution is heated to boiling to decompose the excess of persulphate, then acidified with sulphuric acid and allowed to cool. Now, 100 c.c. of H 2 S- water are added, the excess of the H 2 S expelled by passing C0 2 through the solution, and the precipitated CuS (Bi 2 S 3 , Sb 2 S 3 ) filtered off, and treated as described on p. 235. The filtrate is boiled with hydrochloric acid to reduce the telluric acid to tellurous acid, and the solution is reduced by means of S0 2 and analyzed as described above. The first filtrate from the impure Te and Se will contain the greater part of the Cu, Bi, etc. (b) From Antimony, Tin and Arsenic. If considerable antimony is present, tartaric acid is added to the solution, and the selenium and tellurium are then precipitated by boiling with sulphur dioxide. According to Muthmann and Schroder * this method of separating tellurium from antimony is not quantitative; some antimony is always precipitated with the tellurium. A. Gutbier,f however, finds that a perfect separation can be accomplished by means of hydrazine hydrochloride (not the sulphate). (c) From Mercury. The mercury selenide, or telluride, is dissolved in aqua regia, chlo- rine water is added, and the solution is diluted largely with water. Phosphorous acid is added, { and after twenty-four hours standing, * Z. anorg. Chem., 14, 433 (1897). t Z. anorg. Chem., 32, 263 (1902). J Selenous and tellurous acids are not precipitated by phosphorous acid from dilute hydrochloric acid solution, but are precipitated from hot con- centrated solutions. 282 GRAVIMETRIC ANALYSIS. the mercury is precipitated completely as mercurous chloride, and is determined as such according to p. 170. The nitrate containing selenium and tellurium is concentrated, taken up in water, and the selenium separated from the tellurium according to the method of Keller (see below.) (d) From Gold and Silver. The separation of selenium and tellurium from silver offers no difficulty, inasmuch as the latter can be precipitated by hydro- chloric acid and determined as the chloride. The gold is precipitated as described on p. 257 by oxalic acid and the selenium and tellurium in the filtrate by means of sulphur dioxide. The three metals may also be precipitated together by sulphur dioxide, weighed, and the selenium and tellurium after- ward volatilized by roasting, leaving the gold behind. Tellurium may be separated from gold by precipitating the lat- ter with ferrous sulphate. In the case of selenium, however, it is also precipitated quantitatively by ferrous sulphate from solutions strongly acid with hydrochloric acid. Separation of Selenium from Tellurium. A. Method of E. Keller* Keller's method is based upon the fact that tellurous acid is not precipitated from solutions strongly acid with hydrochloric acid while selenium is precipitated quantitatively. Procedure. The mixture of the two elements precipitated by sulphur dioxide is dissolved in nitric acid and carefully evaporated to dryness. The dry mass is treated with 200 c.c. of hydrochloric acid (sp. gr. 1.175), boiled to remove the nitric acid and saturated with sulphur dioxide. The precipitated selenium is filtered through a Gooch crucible, washed first with a mixture of 90 vol. HC1 (sp. gr. 1.175) and 10 vol. water, then with dilute hydrochloric acid, then with water until free from chloride, finally with absolute alcohol. The selenium is then dried at 105 C. and weighed. The filtrate is diluted with an equal volume of water, * Jour. Amer. Chem. Soc v 19, 771. SEPARATION OF SELENIUM FROM TELLURIUM. 283 heated to boiling, and the tellurium precipitated by sulphur diox- ide and treated in exactly the same way as the selenium. According to Keller, this method gives thoroughly satisfactory results, as long as the amount of tellurium present does not exceed 5 gms. Even then the separation can be effected by increasing the amount of acid to 450 c.c. B. The Potassium Cyanide Method. The precipitate of selenium and tellurium produced by sulphur dioxide is fused with twelve times as much of pure (98 per cent.) potassium cyanide, in an atmosphere of hydrogen, as described on p. 278. The tellurium is almost wholly changed to potassium telluride, K 2 Te (a small amount of pota'ssium tellurocyanide is probably formed), while the selenium is changed for the most part into potassium selenocyanide, and to a slight extent into potassium selenide. The brown melt is dissolved in water, and a slow current of air is conducted through the solution whereby the K 2 Te is quantitatively decomposed according to the equation K a Te+H 2 0+0 = 2KOH+Te. After standing twelve hours the tellurium is filtered off through a Gooch crucible, washed with water, then with absolute alcohol, dried at 105 and weighed. The colorless filtrate is heated to boiling * in order to change any potassium selenide into the double cyanide; it is then acidified with hydrochloric acid under a good hood (hydrocyanic acid!), filtered, and the selenium determined according to p. 277. Remark. This method gives slightly low results for tellurium and high values for selenium. This is due to the fact that a little potassium tellurocyanide is formed by the fusion and this com- pound is not decomposed by the current of air, but is subsequently precipitated with the selenium on acidifying the solution. * Cf. p. 279. 284 GRAVIMETRIC ANALYSIS. Determination of Selenium and Tellurium in Crude Copper. Many copper ores contain selenium and tellurium, so that the crude copper obtained from such ores always contains these ele- ments. The amount present may be determined, according to Keller,* as follows: According to the amounts of selenium and tellurium present, from 5 to 100 gm. of the copper are taken for the analysis The sample is dissolved in nitric acid and an excess of ammonia is added whereby the phosphorus, arsenic, antimony, tin, bismuth, selenium, and tellurium are precipitated together with the ferric hydroxide, while the copper is held in solution by the excess of ammonia. The precipitate is filtered off and washed with dilute ammonia- water until the copper is completely removed. The precipitate is dissolved in hydrochloric acid and this solution satu- rated with hydrogen sulphide in the cold, whereby selenium and tellurium together with arsenic, antimony, tin, and bismuth are thrown down as sulphides and are separated by filtration from the iron and phosphorus. The precipitate thus obtained is treated with sodium sulphide and filtered. The filtrate containing all the selenium and tellurium in the presence of arsenic, antimony, and tin as sulpho salts is acidified with nitric acid and carefully evapo- rated to dryness. The residue is dissolved in 200 c.c. of hydro- chloric acid (sp. gr. 1.175) and treated as described on p. 282, A. MOLYBDENUM, Mo. At. Wt. 96.0. Form: Molybdenum Trioxide, MoO 3 . If the molybdenum is present as ammonium molybdate, a weighed portion is heated in a spacious porcelain or platinum crucible, at first carefully and later to a dull red heat; this leaves the molybdenum trioxide behind in the form of a dense powder, appearing yellow when hot and almost white when cold. There is no danger of losing any of the molybdenum by volatili- zation, provided the dull red heat is not exceeded. If the molybdenum is present as alkali molybdate, it is changed to mercurous molybdate or to its sulphide, and then analyzed as described below. * Jour. Amer. Chem. Soc., 22, 241. PRECIPITATION OF MOLYBDENUM. 285 Precipitation of Molybdenum as Mercurous Molybdate. In the course of analysis it is frequently necessary to determine molybdenum in alkali molybdates obtained by fusion with an alkali carbonate. The greater part of the alkali is neutralized with nitric acid, and to the slightly alkaline solution a barely acid solution of mer- curous nitrate is added until no further precipitation is effected. The liquid is then heated to boiling, the black precipitate, consisting of mercurous carbonate and mercurous rnolybdate, is allowed to set- tle, is filtered and washed with a dilute solution of mercurous nitrate. The precipitate is dried, transferred as completely as pos- sible to a watch-glass, and the precipitate remaining on the filter is dissolved in hot dilute nitric acid into a large porcelain crucible. The solution is then evaporated to dryness, the main portion of the precipitate added to the residue, and the whole is heated very care- fully over a low flame until the mercury is completely volatilized, after which the residual molybdenum trioxide is weighed. Remark. It was formerly customary to add a slight excess of mercurous nitrate solution and then to add mercuric oxide to neu- tralize the excess of nitric acid (the solution of mercurous nitrate contains free nitric acid). According to the above procedure of Hillebrand, the addition of mercuric oxide is wholly superfluous, for the basic mercurous carbonate suffices to remove the slight amount of free nitric acid. Precipitation of Molybdenum as Molybdenum Sulphide. The precipitation of molybdenum as the sulphide can take place in two different ways: either the acid solution may be precipitated by hydrogen sulphide gas, or the solution of ammonium sulpho- molybdate may be acidified with dilute acid. (a) Precipitation of Molybdenum Sulphide from Add Solutions. The molybdenum solution, slightly acid with sulphuric acid,* is placed in a small pressure-flask and saturated in the cold with hydrogen sulphide. The flask is closed, heated on the water- * In many cases, e.g., for the separation of Mo from Ba, Sr, and Ca, it is necessary to effect the separation in a hydrochloric acid solution. 286 GRAVIMETRIC ANALYSIS. bath until the precipitate has completely settled, and filtered after it has become cold. The precipitate is washed with very dilute sulphuric acid and finally with alcohol until the acid has been completely removed. The moist filter is placed in a large porcelain crucible and dried upon the water-bath. The crucible is then covered and very carefully heated over a small flame until no more hydrocarbons are expelled. The cover is then removed, the carbon burned from the sides of *the crucible at as low a tem- perature as possible, and, by raising the temperature gradually, the sulphide is changed to oxide. The operation is finished when no more sulphur dioxide is formed. After cooling, a little mercuric oxide suspended in water is added to the contents of the cr^ible, the mixture is well stirred, evaporated to dry ness on the water- bath, the mercuric oxide is removed by gentle ignition, and the resi- due of molybdenum trioxide is weighed. The mercuric oxide is added in order to remove particles of unburned carbon. It is much easier to transform the molybdenum trisulphide into the oxide as follows : The sulphide is filtered through a Gooch cru- cible, washed with water containing sulphuric acid, and then with alcohol, and dried at 100 C. The crucible is placed within a larger nickel one, covered with a watch-glass, ; and carefully heated over a small flame whereby the sulphide is for the most part changed to the oxide. As soon as the odor of sulphur dioxide can no longer be detected, the watch-glass is removed and the open crucible heated until it is brought to a constant weight. The molybdenum oxide thus obtained always contains traces of SO 3 , and consequently has a bluish appearance. The results, nevertheless, are excellent. (b) Hydrogen Sulphide is passed into the Ammoniacal Molyb- denum Solution until it assumes a bright-red color, when it is acidified with sul- phuric acid and the precipitate treated as described under (a) . The Separation of Molybdenum from the Alkalies can take place by precipitation as mercurous molybdate or as sul- phide, as described above. * To avoid loss by decrepitation. SEPARATION OF MOLYBDENUM. 287 Separation of Molybdenum from the Alkaline Earths. The substance is fused with sodium carbonate, the melt ex- tracted with water and filtered. The solution contains all the molybdenum as alkali molybdate, while the alkaline earths remain undissolved as carbonates. From the aqueous solution the molyb- denum is determined as previously described. Separation of Molybdenum from the Metals of the Ammonium Sulphide Group. The molybdenum is precipitated as sulphide (preferably from a sulphuric acid solution) by treatment with hydrogen sulphide under pressure (see p. 286) . If the solution contains titanium, it is better to first add ammonia and ammonium sulphide, whereby the metals of Group III will be precipitated and the molybdenum will remain in solution in the form of its sulpho salt. After filtration, the molybdenum is precipitated as sulphide by the addition of acid (seep. 286, 6). Separation of Molybdenum from the Metals of Group II. (a) From Lead, Copper, Cadmium, and Bismuth. The solution is treated with caustic soda and then with sodium sulphide, digested some time in a closed flask, and filtered. The molybdenum remains in solution as its sulpho salt, while the other metals are precipitated as sulphides. After filtering, the solution is acidified with sulphuric acid and heated in a pressure-flask until the precipitate has settled and the supernatant liquid appears col- orless. After allowing to cool, the molybdenum sulphide is fil- tered off and converted to oxide, as described on p. 286. (b) From Arsenic. The solution, which must contain the arsenic as arsenic acid, is treated with ammonia, the arsenic precipitated by magnesia mix- ture (see p. 206) and filtered off. The filtrate is acidified with sul- phuric acid and the molybdenum precipitated as sulphide by means of hvdrosen sulphide. 2 88 GRAVIMETRIC ANALYSIS. Separation of Molybdenum from Phosphoric Acid. The phosphoric acid is precipitated from the ammoniacal solu- tion as magnesium ammonium phosphate (see phosphoric acid) and the molybdenum is precipitated as sulphide from the filtrate (cf. p. 285, a). Another way is to saturate the ammoniacal solution with hydrogen sulphide, acidify with hydrochloric acid, and then precipitate the molybdenum^s sulphide. In the filtrate the phosphoric acid is precipitated as magnesium ammonium phosphate under the customary conditions. TUNGSTEN, W. At. Wt. 184.0. Tungsten is determined as its trioxide, WO 3 . If the tungsten is present as ammonium tungstate, as mercu- rous tungstate, or as tungstic acid, it is readily changed by ignition in the air to yellow tungsten trioxide. Since the trioxide is not volatile there is no loss to be feared during its ignition. It is even advisable to heat the crucible finally over a good Teclu burner or over the blast-lamp. If the tungsten is present as alkali tungstate, the tungstic acid may be precipitated as such, o : by means of mereurous nitrate as mercurous tungstate; by ignition the yellow trioxide is obtained and weighed. Precipitation of Tungstic Acid. The aqueous solution of the alkali tungstate is treated with an equal volume of concentrated hydrochloric acid, evaporated to dryness on the water-bath and heated for an hour in the hot closet at 120. The residue is moistened with a little hydro- chloric acid, diluted, boiled, filtered and washed with 6 per cent, hydrochloric acid, or with 10 per cent, ammonium nitrate solu- tion. The yellow tungstic acid is ignited and weighed. In this way the greater part of the tungstic acid is precipitated, but there remains in the filtrate a weighable amount which can be recovered by repeated evaporations with hydrochloric or nitric acid. TUNGSTEN. 289 Remark. The reason why the tungstic acid is not removed completely by the first treatment with nitric or hydrochloric acid is that there is always a small amount of an acid tungstate formed : Na 2 WO 4 + 3WO 3 = Na 2 W 4 O 13 , which it is hard to decompose by acids. If, however, the filtrate is evaporated to dryness with ammonia, then, according to Philipp,* the acid tungstate is converted into ordinary tungstate, Na 2 W 4 13 + 6NH 4 OH = Na 2 WO 4 + 3 (NH 4 ) 2 WO 4 + 3H 2 O, and the latter is decomposed by the nitric acid treatment. This is perfectly true, but on evaporating with acid some more of the acid tungstate is formed, which can be transformed into insoluble tungstic acid only by repeatedly evaporating the filtrates to dryness with hydrochloric or nitric acid. The washing of the precipitate with acid, or with ammonium nitrate solution, is necessary in order to prevent the formation of hydrosol, which would result if pure water were used. Precipitation of Tungsten as Mercurous Tungstate, according to Berzelius.f In the majority of cases it is a question of separating tungstic acid from a solution obtained after fusing with sodium carbonate. The concentrated solution is treated with a few drops of methyl orange, nitric acid is added until the indicator turns pink, the solution is boiled to expel all the carbonic acid, allowed to cool, and then an excess of mercurous nitrate solution is added. The yellow precipitate settles quickly and the supernatant liquid should appear clear as water. After standing three or four hours, the precipitate is filtered off, washed with water containing mer- curous nitrate (5 c.c. saturated mercurous nitrate solution diluted with water to 100 c.c.) dried, ignited in a porcelain crucible under a good hood, using the flame of a Bunsen burner, and weighed as W0 3 . Remark. The tungstic acid is quantitatively precipitated by * Ber., 15, 501 (1882). . t Jahresber., 21, II, 143. Cf. O. v. der Pfordten, Ann. 222, 152 (1883). 290 GRAVIMETRIC ANALYSIS. a single treatment with mercurous nitrate. The process is pref- erable to the above, therefore, when nothing else is present which precipitates under the same conditions. Precipitation of Tungsten as Benzidine Tungstate, according to G. v. Knorre.* If a neutral solution of sodium tungstate is treated with benzidine hydrochloride, a white flcacculent precipitate of ben- zidine tungstate is formed and the precipitate is insoluble in water containing benzidine hydrochloride; when formed in the cold it is hard to filter and, on being washed with pure water, tends to run through the filter. If the precipitate is formed from a hot solution, however, it comes down in a more compact condi- tion and after cooling f can be easily filtered and washed without loss with water containing benzidine hydrochloride. The moist precipitate is ignited in a platinum crucible, heated to 800 in an electric oven, and the residue of WO 3 is weighed. The precipitation can also take place satisfactorily from a cold solution if, before adding the precipitant, a little dilute sulphuric acid or alkali sulphate is added to the solution. In this case H mixture of crystalline benzidine sulphate and amorphous benzidine tungstate is formed which can be filtered after stand- ing five minutes. The benzidine sulphate is entirely volatile, so that equally good results are obtained by either of the above two procedures. If the tungsten is present as tungstate after fusing with sodium carbonate, the melt is dissolved in water, a little methyl orange is added to the clear solution, then hydrochloric acid until the pink color is obtained, and finally 10 c.c. of 0.1 N. sulphuric acid. Benzidine hydrochloride gives a precipitate which can be filtered in five minutes. The washing with dilute benzidine hydrochloride is continued until the evaporation of a few drops of the filtrate on platinum foil leaves no weighable residue. The precipitate is ignited wet as described above. * Ber., 38, 783 (1905). t Since benzidine tungstate is appreciably soluble in hot water contain- ing benzidine hydrochloride, it is necessary in all cases to postpone the filtration until the solution is cold. TUNGSTEN. 291 Preparation of the Benzidine Solution. Twenty grams of commercial benzidine are triturated in a mortar with water, washed with about 400 c.c. of water into a beaker, treated with 25 c.c. hydrochloric acid (sp. gr. 1.2), heated until solution is completed and a brown liquid is formed, which is filtered and diluted to a volume of one liter. * Of this solu- tion, 5.6 c.c. are sufficient to precipitate 0.1 gm. of WOs. If the analysis is carried out with the aid of sulphuric acid, it is necessary to add at least one cubic centimeter of the benzi- dine solution for 10 c.c. of 0.1 N sulphuric acid added. - Preparation of the Wash Liquid. Ten c.c. of the above solution are diluted with distilled water to a volume of 300 c.c. Remark. The method is excellent. W. Kunz found 0.1028 gm., 0.1029 gm., 0.1026 gm. and 0.1033 gm. in aliquot parts of a solution supposed to contain 0.1029 gm. of WO 3 . Determination of Tungsten in Tungsten Steel. (a) Method of G. v. Knorre.f V. Knorre observed that on dissolving tungsten steel in hydrochloric or dilute sulphuric acid, out of contact with the air, all the tungsten remains behind as metal, contaminated with more or less iron. If the finely divided tungsten is filtered off, it oxidizes quickly in the air, forming grayish-yellow tungstic acid, which on further washing with water gives a turbid filtrate; the tungstic acid does not pass through the filter if, instead of using water, the washing is accomplished with a dilute solution of benzidine hydrochloride. By the treatment of the tungsten steel with acids, therefore, the greater part of the iron is separated from the tungsten. To complete the separation, the washed precipitate is ignited wet in a platinum crucible, the residue is fused with four times as much sodium carbonate, the melt * Instead of dissolving the benzidine in hydrochloric acid, 28 gms. of benzidine hydrochloride may be dissolved in 1 liter of water to which 6 c.c. of hydrochloric acid have been added. f Ber., 38, 783 (1905). 292 GRAVIMETRIC ANA LYSIS. extracted with a water, the ferric oxide filtered off, and the tung- sten determined as described on p. 289. It is probable that aluminium-tungsten alloys can be analyzed in a similar manner. (6) L. Welter's Method.* L. Wolter, who has carefully studied the determination of tungsten in tungsten steel, found* that most of the methods are unpractical because they require the metal in a very fine condition. Now a steel with a high percentage of tungsten is so hard that it is practically impossible to get borings or filings without the steel tools becoming badly worn; in such casa.s the sample to be analyzed becomes contaminated with foreign material and the accuracy of the analysis is affected. For this reason if it is required to analyze a steel with a high percentage of tungsten, it should suffice to hammer the sample in a steel mortar until a coarse powder is obtained. Unfortunately a coarse powder dissolves extremely slowly in dilute or concen- trated acids and fusion with sodium carbonate and potassium nitrate, or with sodium peroxide, has little effect upon large particles; it is otherwise with a potassium bisulphate fusion, whereby even large particles of tungsten steel are readily attacked. Procedure. From 0.2-0.5 gm. of the coarse material, which may be in the form of coarse pieces, is fused in a 40-45 c.c. platinum crucible with thirty times as much potassium bisul- phate. To prevent too violent effervescence, at first only one-third of the bisulphate is added and the well-covered, inclined crucible is heated over a small flame until white vapors are evolved. The flame is then removed for half a minute because a fairly violent reaction is now taking place within the crucible. In fact when this point is reached, a strong effervescence can be heard. After cooling somewhat, the remainder of the bisulphate is added in two separate portions. The reaction now takes place more slowly and the temperature is raised until the bottom of the crucible begins to redden and dense white vapors are evolved. The heating is now carefully continued until finally * Chem. Ztg., 1910, 2. TUNGSTEN. 293 the whole crucible is heated to redness. After about fifteen minutes the reaction will be finished. The molten mass in the crucible should show merely a gentle evolution of gas and there should be no black particles of unattacked steel visible. When this point is reached, the mass is allowed to cool, the crucible and cover boiled with water, carefully washed, and the water added is measured so that the volume of the solution will amount to about 60 or 75 c.c. The liquid, which is somewhat turbid with tungstic acid, is treated with 20 c.c. of concentrated hydro- chloric acid and boiled until the precipitated tungstic acid has become a pure yellow. After standing half an hour on the water- bath, the solution is filtered through a small filter and the pre- cipitate washed with 10 per cent, ammonium nitrate solution. The lemon-yellow precipitate is dissolved on the filter in hot, dilute ammonia, the solution caught in a weighed platinum crucible, and the filter washed with ammonium nitrate. The solution of ammonium tungstate in the crucible is evaporated to dryness on the water-bath, then covered with a watch-glass, and very cautiously heated over a free flame, until no more ammonium salts are evolved. After igniting the residue over the full flame of a Bunsen burner, it is weighed as W0 3 . In this way the greater part of the tungsten is obtained. The filtrate from the tungstic acid precipitate still contains tungsten. In order to recover the later, it is treated as described on page 288. If the steel contained silicon, the weighed tungstic acid will always contain silica; it is covered with a few drops of hydrofluoric acid, evaporated to dryness, again treated with hydrofluoric acid, evaporated once more and finally ignited over the flame of a Bunsen burner. The crucible will now contain no silica, but only tungstic acid. Separation of Molybdenum from Tungsten. (a) The Sulphuric Acid Method. This method, proposed by M. Ruegenberg and E. F. Smith,* depends upon the fact that unignited molybdic acid is readily dis- * J. Am. Chem. Soc., 22, 772. 294 GRAVIMETRIC ANALYSIS. solved by warming with sulphuric acid (sp. gr. 1.378), while tung- stic acid is not. W. Hommel* tested this method in the author's laboratory, and could not obtain good results except by digesting the moist oxides with concentrated sulphuric acid, and afterward diluting with three times as much water. Procedure. (a) Both acids are present in a moist, freshly precipitated state. The mixture is covered with concentrated sulphuric acid in a porcelain dish and heated over a free flame. By this means, usu- ally a small amount of the tungstic acid is oxidized to the blue oxide, so that the yellow precipitate of tungstic acid is tinted with green. On adding one or two drops of dilute nitric acid, the green color disappears and the tungstic acid is of a pure yellow color. After digesting for half an hour, the separation is complete. After cooling, the liquid is diluted with three times its volume of water, filtered, washed with water containing sulphuric acid, then two or three times with alcohol, ignited (after burning the filter by itself) in a porcelain crucible, and weighed as W0 3 . The molybdenum is precipitated from the filtrate by passing hydrogen sulphide into the sulphuric acid solution in a pressure- flask, and the precipitate is treated as described on p. 286. If only a little sulphuric acid is used for the separation, the filtrate from the tungstic acid can be evaporated in a platinum dish, the sulphuric acid driven off for the most part, and the residue washed into a weighed platinum crucible with ammonia, and then evaporated, ignited, and weighed. In case large amounts of molyb- denum are present, however, it is always safer to precipitate the molybdenum as sulphide. (/?) Tungsten and Molybdenum are Present in the Form of their Ignited Oxides. These ignited oxides cannot be separated by treatment with sulphuric acid. According to W. Hommel, they can readily be brought into solution by heating for half an hour on the water-bath with concentrated ammonia in a pressure-flask, shaking frequently. After cooling, the contents of the flask, whether dissolved or * Inaug, Dissert., Giessen, 1902. TUNGSTEN. 295 not, are washed into a porcelain dish, evaporated to dryness, and treated as described under (a). It is still better to fuse the ignited oxides with four times as much sodium carbonate, and treat the melt as described under (ct). (b) Sublimation Method.* If a mixture of the trioxides of tungsten and molybdenum, or of their alkali salts, is heated at 250-270 C. in a current of dry hydrochloric acid, the molybdenum is volatilized completely as MoO s .2HCl, which collects on the cooler parts of the tube as a beautiful, white, woolly sublimate, while the tungsten trioxide re- mains behind in the boat. Procedure. The oxides of the two elements, or their sodium salts, are weighed into a porcelain boat, and the latter is placed in a tube made of difficultly fusible glass, of which one end is bent verti- cally downward and is connected with a Peligot tube containing a little water. The horizontal arm of the tube is passed through a drying-oven (to serve as an air-bath) (see Fig. 19, p. 3c), and is connected with apparatus for generating hydrochloric acid gas. The hydrochloric acid bsfore reaching the tube is slowly passed through a flask containing concentrated hydrochloric acid, and then through sulphuric acid. As soon as the temperature has reached about 200 C. the sublimation of the molybdenum begins. From time to time the sublimate collecting in the tube is driven toward the Peligot tube f by carefully heating with a free flame, so that it will be possible to see whether any more molybdenum is being volatilized. After heating for an hour and one-half or two hours, the operation is usually complete. The boat, now containing tung- sten trioxide, or the latter with sodium chloride, is removed, and in case only the former is present, it is weighed after drying in a desic- cator over caustic potash. In case, however, sodium chloride is present (when the tungsten was originally present as sodium tung- state) this is removed by treatment with water, and the filtered WO 3 is weighed. * Pechard, Comptes rendus, 114, p. 173, and 4t>, p. 1101. f By the absorption of the MoO 3 .2HCl in the water of the Peligot tube, the brick-red acid chloride, Mo 3 O.Cl 8 , is often formed: 3[Mo0 3 .2HCl] + 2HC1= 4H 2 O + Mo 3 O 5 Cl<,. This substance is insoluble in hydrochloric acid, but readily soluble in nitric acid 296 GRA yiME TRIG ANAL YSIS. For the determination of the molybdenum, the sublimate in the tube is washed out by means of water containing a little nitric acid, and finally the nitric acid solution of the entire sublimate is care- fully evaporated to dryness in a porcelain dish. The residue is dissolved in ammonia, washed into a porcelain crucible, evaporated to dryness, and changed to the oxide by gentle ignition. (c) The Tartaric Acid Method of H. Rose. The alkali salts of the two metals are dissolved in considerable tartaric acid, an excess of sulphuric acid is added, and the molyb- denum precipitated according to p. 285, by hydrogen sulphide in a pressure-flask. The molybdenum sulphide is filtered off and changed by roasting in the air to the trioxide. For the determi- nation of the tungsten, the tartaric acid is first destroyed by re- peated evaporation with nitric acid, and the precipitated tungstic acid is finally filtered off and changed by ignition to the trioxide. Remark. This method gives correct results, but is not so satis- factory as the preceding one on account of the time consumed in removing the tartaric acid. Analysis of Wolframite (Wolfram). The monoclinic Wolframite is an isomorphous mixture of Ferberite, FeWO 4 , and Hiibnerite, MnWO4, but often contains small amounts of silicic, niobic, tantalic, and stannic acids, besides calcium and magnesium. About 1 gm. of the extremely finely-powdered mineral is fused with 4 gms. sodium carbonate in a platinum crucible over a good burner for from one-half to three-quarters of an hour. After cool- ing, the melt is boiled with water and filtered. The residue con- tains iron, manganese, calcium, and magnesium, and sometimes small amounts of niobic and tantalic acids. The solution contains all the tungstic acid, and silicic acid (stannic acid). The tungstic acid is separated, as above described, either by evaporating with nitric acid or by precipitating with mercurous nitrate. The precipitate is ignited, and weighed as impure WO 3 . The oxide obtained in this way almost always contains silicic acid and sometimes stannic acid. To remove the former the resi- TUNGSTEN. 297 due is heated with hydrofluoric and sulphuric acids, first on the water-bath and finally over a free flame, and the residue is weighed. The difference shows the amount of silicic acid present. Stannic acid is usually present in such small amounts that it is not usually determined. The separation of tungsten from tin, however, may be effected (according to Rammelsberg) by repeated ignition with pure, dry ammonium chloride. The tin is volatilized as stannic chloride, while the tungsten remains behind. This last operation is conducted as follows : The residue obtained after the treatment with hydrofluoric acid is mixed with six to eight times as much ammonium chloride, the crucible is placed within a second larger crucible,* and the latter is covered and ignited until the ammonium chloride is completely expelled. This opera- tion is repeated three times. The inner crucible is then heated with ready access of air until its contents become of a pure-yellow color, after which it is cooled and weighed. The ignition with ammonium chloride and weighing of the residue is repeated until a constant weight is obtained. To determine the iron, manganese, calcium and magnesium, the insoluble residue from the sodium carbonate fusion is dis- solved in hydrochloric acid, the solution evaporated to dry- ness, the dry residue moistened with concentrated hydrochloric acid which is allowed to act for ten minutes, then diluted with water, boiled and any silica filtered off. In the filtrate, the iron is separated from the other metals by the basic acid acetate process described on p. 152. The manganese is precipitated by heating with bromine (see p. 123) and the calcium and magnesium separated as described on p. 77. Remark. For the determination of niobium and tantalum, a larger portion of the substance, about 5 gms., must be taken. The finely-powdered material is treated with hydrochloric acid to which about one-fourth of its volume of nitric acid has been added, and digested on the water-bath until the residue is colored a pure yel- low. The latter is filtered off, washed with water containing acid * This is to prevent any stannic oxide from collecting on the outside of the crucible; the oxide is formed when tin chloride comes in contact with moist air. 298 GRAVIMETRIC ANALYSIS. until the iron reaction can no longer be obtained, when the residue is taken up in ammonia and filtered ; in this way the tungstic acid is removed. The residue is usually dark-colored uid consists of enclosed mineral as well as silicic, stannic, niobic, and possibly tantalic acids. It is treated with aqua regia again, water is added, and the filtered residue is once more treated with ammonia. The final residue is now free from tungsten ; it is ignited, weighed, and freed from silica by treatment with sulphuric and hydrofluoric acids. The residue cf tin dioxide and niobium pentoxide (and perhaps tantalum pentoxide) is placed in a porcelain boat and ignited in a current of hydrogen. The metallic tin is ex- tracted with hydrochloric acid, and the residue, consisting of Nb2O 5 (Ta 2 O5), is weighed. Iron alloys rich in tungsten are acted upon only slowly by aqua regia. If ; however, they are first roasted in the air, they are comparatively easily brought into solution.* Decomposition by a potassium bisulphate fusion is still better (see p. 292) . Analysis of Tungsten Bronzes. The analysis of these alkali salts of complex tungsten acids, discovered by Wohler in 1824,f offered for a long time considerable difficulty on account of the fact that acids do not decompose them very readily. By fusion with alkalies in the air, or better still in the presence of potassium nitrate, the tungsten bronzes can be converted without difficulty into normal alkali tungstate, and the tungsten determined by one of the methods already described. It is obvious that the alkalies cannot be determined in the same sample, so that PhilippJ proceeded as follows: The bronze is treated with ammoniacal silver nitrate solution, whereby the WO 2 is oxidized to WOs with the precipitation of an equivalent amount of silver, whereas the whole of the tungsten remains in solution in the form of alkali and ammonium tungstates. In the filtrate obtained after filtering off the deposited silver, the * Preusser. Zeit. f. anal. Chem., 1889, p. 173. t Pogg. Ann., 2, 350. J Berichte, 15. 500 (1882). TUNGSTEN. 299 tungstic acid is precipitated by treatment \\ ith nitric acid and determined as WOs. After removing the excess of silver, by precipitating it as the chloride, the nitrate is evaporated to dryness with the addition of sulphuric acid, and the alkali weighed as sulphate. Although the above method affords satisfactory results in the analysis of bronzes containing comparatively little tungsten, it is wholly inadequate in the case of bronzes rich in tungsten. The method of B runner,* which follows, is applicable in all cases. It is based upon the fact that the bronzes can be transformed very easily, and without loss of alkali, into normal tungstates by heat- ing them with ammonium persulphate, or ammonium acid sulphate. Procedure. About 0.5 gm. of the finely-powdered bronze is treated in a porcelain crucible with 2 gms. of alkali-free ammonium sulphate and 2 c.c. of concentrated sulphuric acid f and carefully heated over a very small flame. During the heating the contents of the crucible are frequently shaken about a little by cautiously moving the crucible. The escape of gases from the crucible soon ceases and when sulphuric acid vapors begin to be evolved, the decomposition of the bronze results. In the case of sodium and lithium bronzes, the fused mass appears greenish, whereas with a potassium bronze the color is yellowish white. After a part of the ammonium sulphate has been volatilized, the mass in the crucible is allowed to cool, another gram of ammonium sulphate is added and 1 c.c. of con- centrated sulphuric acid, whereupon the contents of the crucible are once more heated as before until sulphuric acid fumes come off thickly; the crucible is then allowed to cooL The greenish or yellowish-white fusion is softened by treat- ment with water and rinsed into a porcelain dish. After adding 50 c.c. of concentrated nitric acid, the contents of the evaporating dish are digested on the water -bath for three or four hours, and * Inaug. Dissert, Zurich, 1903. f If ammonium persulphate is used the addition of sulphuric acid is unnecessary. The only objection to the use of the persulphate lies in the fact that the commercial salt often contains some potassium persulphate. 300 GRA VIME TRIG ANAL YSIS. then, after diluting with water, the residue of pure yellow tungstic acid is filtered off. In order to recover the small amount of tungstic acid remain- ing in the filtrate, it is evaporated as far as possible on the water- bath, allowed to cool, diluted with a little water, carefully treated with an excess of ammonia, again evaporated on the water-bath and treated as described on p. 288. . The final filtrate from the tungstic acid determination is evaporated to dryness, the ammonium salts expelled by ignition and the residue of alkali sulphate weighed (cf. pp. 41 and 42). Separation of Tungsten from Tin. Augenot's Method.* One gram of the finely powdered mineral is intimately mixed in an iron crucible with 8 gms. of sodium peroxide, and the mixture is carefully fused over the Bunsen burner for about fifteen minutes. After cooling, the melt is softened with water and transferred into a 250-c.c. flask (if lead is present, carbon dioxide is conducted into the solution for a few minutes), the solution is diluted to the mark, well mixed and filtered through a dry filter, rejecting the first few c.c. of the filtrate. For the determination of the tungstic acid, Augenot proceeds according to H. Borntrager.f 100 c.c. of the filtrate are allowed to flow into a mixture of 15 c.c. concentrated nitric acid and 45 c.c. of concentrated hydrochloric acid, evaporated to dryness in a porcelain dish, the dry residue treated with 50 c.c. of a solution 1000 c.c. water, 100 g. concentrated hydrochloric acid and 100 g. ammonium chloride, filtered and washed with the same solution. The precipitate, which, besides the tungstic acid, also contains silicic acid and stannic oxide, is dissolved in warm ammonia water, the filter well washed with the same reagent, and the ammoniacal solution allowed to flow into 15 c.c. of con- centrated nitric acid and 45 c.c. of concentrated hydrochloric acid. This time the liquid is evaporated to complete dryness, is taken up with the above mixture of hydrochloric acid and ammonium chloride, filtered and washed finally with dilute nitric * Z. angew. Chem., 19, 140 (1906). fZ. anal. Chem., 39, 361 (1900). TUNGSTEN. 301 acid (using preferably a Munroe crucible), and ignited in an electric oven; or in case the latter is not available, the crucible is placed inside a larger one of platinum and ignited over a Teclu burner to constant weight. The resulting tungstic acid is said to be free from silica and stannic oxide. For the determination of the tin, a second 100 c.c. of the original alkaline solution is used. It is treated with 45 c.c. of concentrated hydrochloric acid, whereby tungstic acid and stannic acid are precipitated. At this point 2 or 3 gms. of pure zinc are added whereby the tungstic acid is changed to the blue oxide and the stannic acid is reduced to metallic tin. The mixture is allowed to stand quietly for an hour at a temperature between 50 and 60. The tin then goes into solution as stannous chloride, and the greater part of the tungsten remains undissolved in the form of the blue oxide. The latter is filtered off, and washed. In this way the whole of the tin is obtained in an acid solution, in the presence of a small amount of tungsten, which does no harm. The blue oxide on the filter, however, -is dissolved in hot, dilute ammonia solution in order to make sure that it contains no trace of metallic tin. If this should be the case, the small particle of tin is dissolved in a little hydrochloric acid and the resulting solution added to the main solution of the tin. The solution is now diluted with water and the tin precipitated as stannous sulphide by the introduction of hydrogen sulphide gas. The precipitate is filtered, ignited in a porcelain crucible and weighed as Sn02 (see p. 233). Or, the moist precipitate of stannous sulphide may be dissolved in potassium hydroxide solution and the tin determined by electrolysis (see p. 234). According to Donath and Miiller * a mixture of stannic oxide and tungstic acid may be separated as follows: The mixture is ignited with powdered zinc for fifteen minutes in a covered porcelain crucible. After cooling, the spongy contents of the crucible are heated in a beaker with hydrochloric acid (1:2) until there is no more evolution of hydrogen perceptible. The solution is then allowed to cool somewhat and some powdered potassium * Wiener Monatshefte, 8, 647 (1887). 302 GRAVIMETRIC ANALYSIS. chlorate is added little by little until the blue tungsten oxide is completely transformed to yellow tungstic acid, when the liquid no longer shows any blue tinge. It is diluted with one and one- half times as much water, and after standing twenty-four hours the tungstic acid is filtered off, washed first with dilute nitric acid and then with one per cent, solution of ammonium nitrate, dried, ignited, and weighed as WO 3 . The tin is determined in the filtrate as above. Separation of Tungstic Acid from Silica. When a mixture of tungstic and silicic acid is at hand, such as is obtained by evaporation with nitric acid, the silicic acid may be removed by treating the ignited residue with hydrofluoric acid and a large excess of sulphuric acid. The separation does not succeed, however, in the mixture of oxides as obtained after precipitation with mercurous nitrate; for in this case the silicic acid is so enveloped with tungstic acid that some of the former is not volatilized as fluoride. In such cases, as Friedheim * has shown, excellent results may be obtained by the Method of Perillon.-f The mixture of the ignited oxides is introduced into a plati- num boat and heated to redness in a stream of dry hydrogen chloride. Thereby the tungsten is volatilized, probably as an acid chloride, which can be recovered in a receiver containing dilute hydrochloric acid; the silica remains behind in the com- bustion tube. Frequently the tungsten is reduced to a blue lower oxide, which is not volatile in a current of hydrochloric acid gas. In such cases, after the apparatus has been allowed to cool, the hydrogen chloride is replaced by air, and the contents of the boat are heated in a current of air. The tube is again allowed to cool, the air replaced by hydrogen chloride, and the tube once more heated *Z. anorg. Chem., 45, 398 (1905). t Bull. soc. 1'industrie miner., 1884. VANADIUM. 303 to redness. The process is repeated if necessary until finally a residue of pure white SiO 2 is obtained. The molybdic acid hydro- chloride, MoO3-2HCl, in the receiver is evaporated to dryness with nitric acid, and the precipitated WOs is filtered, ignited and weighed. Unless the current of hydrogen chloride gas is perfectly free from air, the platinum boat will be strongly attacked. The bisulphate method of effecting the separation is less accurate (See Vol. I). VANADIUM, V. At. Wt. 51.2. Vanadium is determined as the pentoxide, V 2 O 5 . The most convenient method for determining vanadium is a volumetric process, and will be discussed in the chapter on volu- metric analysis. If vanadium is present as ammonium or mercurous vanadate, it can be easily changed to the pentoxide by ignition; the latter is a reddish-brown fusible substance which solidifies as a radiating, crystalline mass. If vanadic sulphide is carefully roasted in the air, it is also changed quantitatively to the pentoxide. In the analysis of most minerals containing vanadium, the vanadium is separated from the other metals present by fusing with a mixture of six parts sodium carbonate and one part potassium nitrate. After cooling, the melt is extracted with water, whereby the sodium vanadate goes into solution while most of the metals are left behind in the form of oxides or carbonates. If phosphorus, arsenic (molybdenum, tungsten), and chromium are present, these elements also dissolve on treating the melt with water in the form of the sodium salts of the corresponding acids. In practice, therefore, the vanadium is usually met with as the sodium salt of vanadic acid, and it is a matter of separating it from the aqueous solution obtained after fusing with sodium carbonate and potassium nitrate, and of separating it from the other acids which are likely to accompany it (phosphoric, arsenic, and chromic acids). 304 GRAVIMETRIC ANA LYSIS. Precipitation of Vanadic Acid from the Solution of Sodium Vanadate. There are two good methods for the separation of vanadic acid from a solution of an alkali vanadate: the Rose method , according to which the vanadium is precipitated as mercurous vanadate, and that of Roscoe, by which it is precipitated as lead vanadate. The Berzelius-Hauer method,* in which the vanadium is precipitated as ammonium metavanadate, was found by Holver- scheidt f to give always too low results, but Gooch and Gilbert, { as well as E. Campagne, obtained correct results by working in an ammoniacal solution, which was saturated with ammonium chloride. 1. The Mercurous Nitrate Method of Rose. The alkaline solution is nearly neutralized with nitric acid and to it is added, drop by drop, a nearly neutral solution of mer- curous nitrate II until, after the precipitate has settled, a further addition of the reagent causes no precipitation. The liquid is then heated to boiling, the gray-colored precipitate is allowed to settle and is filtered and washed with water to which a few drops of mercurous nitrate solution have been added. The precipitate is dried, ignited under a good hood, and the residue of V 2 O 5 is weighed. Remark. In neutralizing the alkaline solution of the vanadate with nitric acid, the solution must on no account be made acid, for in this case nitrous acid (from the nitrate fusion) will be set free and the latter reduces some of the vanadate to a vanadyl salt and the latter is not precipitated by mercurous nitrate. In order to avoid passing over the neutral point, Hillebrand recom- mends fusing with a weighed amount of sodium carbonate and adding the amount of nitric acid that has been found necessary by a blank test to neutralize this. The method gives good results. * Pogg. Ann., 22, 54 and J. prakt. Chem., 69, 388. t Dissertation, Berlin, 1890. J Z. anorg. Chem., 32, 175 (1902). Berichte, 1903, 3164. || The mercurous nitrate used should leave no residue on being ignited. VANADIUM. 305 2. Tine Lead Acetate Method of Roscoe* Principle. If a solution weakly acid with acetic acid is treated with lead acetate, orange-yellow lead vanadate is quantitatively precipitated. The lead vanadate, however, does not possess a constant composition, so that the amount of vanadium present cannot be determined by weighing the precipitate. After being washed, it is dissolved in as little nitric acid as possible, the lead precipitated as lead sulphate, and the vanadium determined in the filtrate by evaporating the latter, driving off the excess of sul- phuric acid, and weighing the residual V 2 O 5 . Procedure. The solution from the sodium carbonate and potassium nitrate fusion is nearly neutralized as before with nitric acid, an excess of lead acetate solution is stirred into it, when the voluminous precipitate will collect together, rapidly settle to the bottom of the beaker, and the supernatant liquid will appear absolutely clear. The precipitate is at first orange-colored, but on standing it gradually becomes yellow and finally perfectly white, It is filtered off, washed with water containing acetic acid until half a cubic centimeter of the filtrate will leave no residue on evaporation. The precipitate is now washed into a porcelain dish, the part remaining on the filter is dissolved in as little as possible of hot dilute nitric acid, and the solution added to the main part of the precipitate, to which enough nitric acid is added to completely dissolve it. An excess of sulphuric acid is added to the solution, and it is evaporated on the water-bath as far as possible, finally heating over the free flame until dense fumes of sulphuric acid are evolved. After cooling, from 50 to 100 c.c. of water are added, the lead sulphate is filtered off and washed with dilute sulphuric acid until 1 c.c. of the filtrate will show no yellow color with hydrogen peroxide. The lead sulphate should be white and free from vanadium ; it will be so provided enough sul- phuric acid was used and the mass was not heated until absolutely dry before diluting with water. The filtrate containing all the vanadic acid is evaporated in a porcelain dish to a small volume, transferred to a weighed platinum crucible, evaporated further on the water-bath, and finally in an air-bath until all the sulphuric * Ann. Chem. Pharm., Suppl., 8, 102 (1872). 306 GRAVIMETRIC ANALYSIS. acid is removed. The open crucible is then ignited for some time * at a faint-red heat and its contents finally weighed as V 2 5 . Remark. Instead of decomposing the lead vanadate by means of sulphuric acid, Holverscheidt recommends precipitating the lead as sulphide- by means of hydrogen sulphide and determining the vanadium in the nitrate. For this purpose the blue-colored filtrate from the lead sulphide precipitate (which contains some vanadyl salt) is boiled to expel the exces^ of hydrbgen sulphide and the depocited sulphur is filtered off. A few drops of nitric acid are added, the solution evaporated to dryness, and the reddish-yellow hydrate of vanadic acid is changed by gentle ignition into the pentoxide of vanadium. Lead may also be separated from the vanadic acid as lead chloride. In this case the procedure recommended in the analysis of vanadinite (p. 308) is followed. The separation of vanadium as the sulphide by acidifying a solution of an alkali vanadate that has been treated with an excess of ammonium sulphide is not admissible, for only a part of the vanadium is precipitated as the brown sulphide, the rest remaining in solution in the form of vanadyl salt. H. Rose called the attention of chemists to the inaccuracy of this method, but this has not pre- vented its being recommended in some of the most recent works on analytical chemistry. The author has carefully tested the method and found it useless. Separation of Vanadium from Arsenic Acid. Most minerals containing vanadium also contain arsenic, and after extracting the melt, obtained by fusion with sodium carbonate and nitre, with water, both elements go into solution. For their separation ; the solution is acidified with dilute sulphuric acid and sulphur dioxide is passed into the hot liquid, whereby the vanadic acid is reduced to vanadyl salt and the arsenic to arsenious acid. After boiling to remove the excess of sulphur dioxide, the solution is saturated with hydrogen sulphide and the precipitate of arsenic trisulphide is filtered off. The filtrate is freed from * On expelling the sulphuric acid, there is finally formed some green and brown crystals of a compound of vanadic acid with sulphuric acid; these are decomposed only at a faint-red heat. VANADIUM. 37 hydrogen sulphide by boiling, evaporated with nitric acid in order to form vanadic acid again, the solution is then made alkaline with sodium carbonate, and the vanadium determined by one of the above methods. Separation of Vanadium from Phosphoric Acid. If the solution of the soda-nitre fusion contains phosphoric as well as vanadic acid, both are precipitated by mercurous nitrate, the precipitate washed with dilute mercurous nitrate solution and weighed. In this way the sum of the V 2 5 + P 2 O 5 is obtained. When P 2 O 5 is present the V 2 5 does not melt, but only sinters together. The ignited oxides are fused with an equal weight of sodium carbonate, the melt is dissolved in water, the solution made acid with sulphuric acid and boiled with sulphurous acid in order to reduce the vanadic acid to vanadyl sulphate; the latter will be recognized by the pure blue color which the solution assumes. Carbon dioxide is passed into the boiling solution until the excess of sulphurous acid is removed, when it is allowed to cool. To the cold solution, now about 100 c.c. in volume, 200 c.c. of a 75 per cent, solution of ammonium nitrate and 50 c.c. of ammonium molybdate solution are added (cf. Remark, below), the solution is warmed to about 60 C., set aside and allowed to stand for one hour. The clear liquid is then decanted through a filter, washed three times by decantation with 50 c.c. of the proper wash liquid (see p. 437), after which the precipitate is dissolved by passing 10 c.c. of 8 per cent, ammonia through the filter into the the beaker containing the bulk of the precipitate and the filter is finally washed with 30 c.c. of water. To this solution 20 c.c. of a 34 per cent, ammonium nitrate solution and 1 c.c. more of ammo- nium molybdate are added, the solution heated until it begins to boil, and the phosphoric acid reprecipitated by the addition of 20 c.c. of hot 25 per cent, nitric acid. The phosphoric acid is determined by the method of Woy (page 440). The amount of phosphoric acid found is deducted from the sum of the oxides and the difference gives the amount of V 2 O 5 . Remark. A. Gressly tested this method in the author's labora- tory and made the interesting observation that if about 15 gm. of V 2 O 5 was present with 0.1 gm. P 2 O 5 , no trace of the latter could 308 GRAVIMETRIC ANALYSIS. be detected according to the procedure of Woy, not even on boiling the solution. On the other hand, an immediate precipitation was produced if a stronger solution of ammonium molybdate were used (75 gms. of ammonium molybdate dissolved in 500 c.c. of water) and this solution poured into 500 c.c. of nitric acid, sp. gr. 1:2. The above-described separation gives correct results only when the vanadium is present as vanadyl sulphate; if vanadic acid is present it is precipitated with the j^Jiosphoric acid. If the solu- tion is allowed to stand after the addition ol the ammonium molyb- date, the vanadyl sulphate is gradually oxidized to vanadic acid; the precipitate therefore should not be allowed to stand long before filtering. Separation of Vanadium from Molybdenum. The solution containing the alkali salts of the two acids is acidified with sulphuric acid and the molybdenum precipitated in a pressure-flask by means of hydrogen sulphide, and the pre- cipitate filtered off through a Gooch crucible, as described on pp. 285 and 286), and weighed as Mo0 3 . After removing the excess of hydrogen sulphide from the filtrate, the vanadium is oxidized with nitric acid and determined as described under the Separation of Vanadium from Arsenic Acid, p. 306. Analysis of Vanadinite, (Pb 5 ( V0 4 ) 3 C1). Besides lead, vanadic acid, and chlorine, the mineral often con- tains arsenic and phosphoric acids. Determination of Chlorine. About 1 gm. of the finely powdered mineral is dissolved in dilute nitric acid (in order to avoid loss of chlorine the solution is kept cold), and the solution is diluted with considerable water. The chlorine is precipitated with silver nitrate and the weight of the silver chloride determined as described on p. 317. Determination oj Lead. The filtrate from the silver chloride is treated with hydrochloric acid in order to precipitate the excess of silver, filtered, washed with hot water, and the solution thus freed from silver is evaporated VANADIUM. 309 to dryness to remove the nitric acid. The dry mass is moistened with hydrochloric acid, 95 per cent, alcohol is added in order to precipitate completely the lead chloride, and the latter is filtered through a Gooch crucible, washed with alcohol, dried at 110 C. and weighed as PbCl 2 . Determination of Vanadium, Phosphoric Acid, and Arsenic. The nitrate from the lead chloride contains the vanadium as vanadyl salt. The alcohol is driven off by careful heating on the water-bath, nitric acid is added to the solution, and the latter is repeatedly evaporated in order to oxidize the blue vanadyl salt to brown vanadium pentoxide. The dry mass is washed by means of as little water as possible into a weighed plati- num crucible, the residue adhering to the sides of the dish is dis- solved in a little ammonia and added to it. The crucible is then heated, at first gradually to expel the ammonia, and afterward more strongly with ready access of air (uncovered crucible) until the reduced, dark-colored oxide is changed over to the brownish- red pentoxide. The temperature is then raised until the vanadium oxide begins to melt. If phosphoric acid is present, its anhydride is weighed with the vanadium and the amount of P 2 5 is deter- mined as described on p. 307; this amount is deducted from the weight of the two oxides. The determination of arsenic is best carried out in a separate portion. For this purpose the mineral is dissolved in hot nitric acid, the greater part of the excess of the acid is removed by evaporation, the solution is diluted with water, and the lead precipitated by the addition of sulphuric acid. From the filtrate, the last portions of lead and arsenic are precipitated by hydro- gen sulphide, after previous reduction with sulphurous acid. The filtered precipitate is digested with sodium sulphide and the arsenic precipitated from the solution thus obtained by the addi- tion of hydrochloric acid. The arsenic sulphide is then changed to arsenic acid, preferably by dissolving in ammoniacal hydrogen peroxide, and is precipitated as magnesium ammonium arsenate and determined according to p. 206. 310 GRAVIMETRIC ANA LYSIS. Determination of Vanadium and Chromium in Iron Ores and Rocks. Method of W. F. Hillebrand* As vanadium often occurs in many ores of iron and in rocks, although in very small amounts, it is often of interest and of im- portance to be able to determine it in such cases. For this purpose it is best to proceed as follows : Five gms. of the finely powdered mineral are mixed with 20 gms. sodium carbonate and 3 gms. potassium nitrate and fused over the blast-lamp. The green fusion (containing manganese) is extracted with water, a few drops of alcohol are added to reduce the man- ganese, and the residue is filtered off.f The aqueous solution contains sodium vanadate and often phos- phate, chromate, molybdate, aluminate, and considerable silicate as well. First of all, the aluminium and the greater part of the silicic acid are removed by nearly neutralizing the alkaline solution with nitric acid.J: It is very important that the solution is not made acid at this point on account of the reducing action of the nitrous acid set free from the nitrite formed during the fusion. The almost neutral solution is evaporated nearly to dryness, taken up in water, and filtered. The cold alkaline solution is now treated with an almost neutral solution of mercurous nitrate until no further precipitation takes place. The somewhat voluminous precipitate contains, besides mercurous carbonate, also its chromate, vanadate, molybdate, arsenate, and phosphate, if the corresponding elements are present in the mineral. If the precipitate is too bulky, a little nitric acid is cautiously added, and then a drop of mercurous nitrate in order to see if the precipitation is complete. * See U. S. Geol. Survey Bull., 411. j- If considerable vanadium is present, the insoluble residue always con- tains vanadium and must be fused with soda-nitre again. J The amount of nitric acid necessary to neutralize 20 gms. of soda is determined by a blank test. The residue of aluminium hydroxide and silicic acid almost never con- tains vanadium, but contains chromium. If it is desired to determine the latter, the residue is evaporated to dryness with hydrofluoric and sulphuric acids, the dry mass is fused with soda and nitre again, and the aqueous solu- tion of the melt added to the main solution. VANADIUM. 3l ; l The liquid is heated to boiling, filtered, the precipitate washed with water containing ammonium nitrate, dried, and ignited in a platinum crucible at as low a temperature as possible. The ignited residue is fused with a little sodium carbonate, the melt extracted with water and, if yellow-colored, it is filtered into a 25-c.c. flask and the amount of chromium determined colorimetrically by com- paring its color with a carefully prepared solution of potassium chromate. The solution is then slightly acidified with sulphuric acid, and the molybdenum, arsenic, and traces of platinum precipitated by hydrogen sulphide in a pressure-flask. The precipitated sulphides are filtered off, the filter together with the precipitate is carefully ignited in a porcelain crucible, a few drops of sulphuric acid are added and the crucible heated again until the acid is almost com- pletely removed. On cooling the mass is colored a beautiful blue if molybdenum is present. The filtrate from the above precipitate is freed from the ex- cess of the hydrogen sulphide by boiling and passing a stream of carbon dioxide through it, and the hot solution is then titrated N to a pink color with potassium permanganate solution (cf . Vol. Anal.). In order to obtain absolutely accurate results, the solu- tion is now reduced by means of sulphur dioxide and the titration repeated. The mean of the two experiments gives the vanadium value. This method gives correct resulcs only when the amount of chromium present is very small, which is true in the majority of cases. In case more than 5 mgm. of chromium are present a correction must be made, for a measurable amount of permanganate is used up in oxidizing the chromium. This is determined by tak- ing a solution containing the same amount of chromate as the analyzed solution, reducing it with sulphurous acid, and titrat- ing with permanganate. The amount of permanganate now used must be subtracted from the amount used in the analysis, and from the difference the amount of vanadium present is calculated. 312 GRAVIMETRIC ANALYSIS. Determination of Vanadium and Chromium in Pig Iron. From 5 to 10 gms. of the iron are dissolved in dilute hydro- chloric acid in a flask, meanwhile passing a current of carbon diox- ide through the liquid. For each gram of iron taken, 5 c.c. of hydrochloric acid, sp. gr. 1.12 and 10 c.c. of water are used. The solution is hastened by warming, finally boiling it until there is no more evolution of gas. It is now dihfted with an equal volume of water, allowed to cool, and, without filtering off the slight residue, an excess of barium carbonate is added and the mixture allowed to stand for twenty-four hours with frequent shaking. The residue is filtered off, rapidly washed with cold water, dried and ignited in a platinum crucible in order to burn off the carbon. Five parts of sodium carbonate and one part of nitre are then added to the con- tents of the crucible and the mixture is heated to quiet fusion. The fusion is leached with water and the solution thus obtained contains all the chromium as chromate, and the vanadium as vana- date in the presence of alkali silicate and phosphate. The aque- ous solution is now nearly neutralized with nitric acid, being care- ful not to make the solution acid as the nitrous acid set free will reduce some of the vanadium and chromium. The barely alkaline solution is then treated with an almost neutral solution of mercu- rous nitrate until no further precipitation takes place, the liquid is heated to boiling, filtered and washed with water containing a little mercurous nitrate. After drying, as much of the precipitate as possible is transferred to a platinum crucible, the filter burned by itself and its ash added to the main portion of the precipitate, which is ignited to remove the mercury. The residue is fused with a little sodium carbonate, the melt extracted with water, and the solution filtered. In case the filtrate is colored yellow, the amount of chromium present is determined colorimetrically * by placing the solution in a graduated cylinder and comparing its color with a potas- * The colorimetric determination is only suitable when small amounts are present. When considerable chromium is present (chrome steel) the results obtained by the colorimetric determination may be as much as 2 per cent, too high. In such cases the chromic acid is titrated with ferrous sulphate (see Vol. Anal.). VANADIUM. . 313 sium-chr ornate solution containing a known amount of chromium. The solution is then acidified with sulphuric acid and hydrogen sulphide is conducted into the boiling solution to precipitate ar- senic and platinum. After filtering, the hydrogen sulphide is re- N moved and the hot solution titrated with -=-= potassium perman- 1UU ganate solution (cf. Volumetric Analysis). Determination of Vanadium, Molybdenum, Chromium, and Nickel in Steel. Method of A. A. Blair.* The method is given in this edition of the book as illustrating the removal of the greater part of the iron from a ferric chloride solution by shaking with ether. Although it is difficult to effect a perfect separation in this way, still when the conditions are right, almost all of the iron can be removed so that a large sample of steel can be taken for analysis. This particular method has not been tested in the author's laboratory and is not given in the German edition. The ether separation, however, is discussed in many other text-books and deserves mention. (TRANSLATOR.) Molybdenum, in this analysis, follows the iron so that when a small amount of the former is present, as is the case in steels, the ether solution may be regarded as containing all of the molybdenum, as well as the greater part of the iron. Procedure. Two grams of the steel are dissolved in nitric acid with the addition of hydrochloric acid if necessary. The resulting solution is evaporated to dryness and the residue dissolved by treatment with hot concentrated hydrochloric acid. If silica is present, the solution is diluted and filtered. The hydrochloric acid solution is evaporated to a sirup, the latter dissolved in a little hydrochloric acid, sp. gr. 1.1, f and transferred with the aid of a little more of the same acid to a separatory funnel of about 250 c.c. capacity which is provided with tightly fitting stop-cock and glass-stopper. About 80 c.c. of ether are added to the cold * J. Am. Chem. Soc., 30, 1228. \ The separation works best with acid of this concentration. 314 GRAVIMETRIC ANALYSIS. solution* and the mixture is shaken vigorously for half a minute. When the two liquids are in equilibrium, the lower layer is transferred to a second separatory funnel. The stopper of the first funnel is carefully rinsed with hydrochloric acid, sp. gr. 1.1, the contents of the funnel once more shaken, and the lower layer added to the contents of the second funnel. The solution in the latter is shaken with 50 c.c. more of ether, and the acid solution containing all the vanadium, chromium, nickel, manganese and copper, is run into a beaker and freed from dissolved ether by evaporating on the water bath nearly to dry- ness. An excess of nitric acid is added and the solution again evaporated to remove all the hydrochloric acid. When the solu- tion is almost sirupy, 20 c.c. of hot water are added and the solution is heated with the addition of a little sulphurous acid to reduce any chromic acid that may have been formed. The hot solution is slowly poured, while stirring vigorously, into a boiling 10 per cent, solution of sodium hydroxide. After boiling a few minutes, the precipitate is allowed to settle, is washed twice by decantation, and finally on the filter until the volume of the filtrate is about 300 c.c. The precipitate contains the hydrated oxides of chromium, nickel and iron with the greater part of the manganese and any copper that may have been in the sample. The filtrate contains the vanadium, some silicate and aluminate (from the reagents) and sometimes a little chromium. It is made barely acid with nitric acid, once more made alkaline with a few drops of sodium hydroxide, boiled and filtered. The vanadium is determined in this last filtrate. It is pre- cipitated as lead vanadate by the addition of 10 c.c. of a 10 per cent, solution of lead nitrate, eventually adding enough acetic acid to make the solution decidedly acid and boiling for several minutes. The lead vanadate is filtered and washed with hot water. The precipitate is dissolved in hot, dilute hydrochloric acid, the solution evaporated nearly to dryness, treated with 50 c.c. of hydrochloric acid, evaporated again, cooled, treated with 10 c.c. concentrated sulphuric acid, and evaporated until fumes * The warm solution would result in the reduction of some of the iron by the ether. DETERMINATION OF VANADIUM, MOLYBDENUM, ETC. 315 of sulphuric anhydride are evolved. When cold 150 c.c. of water are added, the solution heated to between 60 and 70 and titrated with permanganate. The evaporation of the vanadate solution with hydrochloric acid and subsequent treatment with sulphuric acid results in the reduction of the vanadium from the quinque- valent to the quadrivalent condition,* and the titration with permanganate makes the vanadium quinquevalent again. Con- sequently 2 atoms of V are equivalent to 1 atom of oxygen in the titration. The presence of a little iron does not interfere when the vanadium- is reduced in the above manner. The two precipitates obtained by the sodium hydroxide treat- ment contain chromium, nickel and copper besides iron and manganese. The two filters are ignited and the precipitates fused with about 2 gms. of sodium carbonate and half a gram of potassium nitrate. The fused mass is treated with water and the solution filtered. The residue contains the nickel, copper, iron and part of the manganese; the filtrate contains the chromium and the rest of the manganese. To the filtrate, enough ammonium nitrate is added to convert all the sodium salts to nitrates, and the solution evaporated to small volume with the addition of a few drops of ammonia from time to time. The evaporated solution is diluted to 50 c.c. and filtered. The precipitate consists of hydrated manganese dioxide (alumina and silica from the reagents). Tne filtrate is boiled to drive off the ammonia, sul- phurous acid added to reduce any chromic acid, the excess of reducing agent boiled off, and the chromium precipitated by the careful addition of ammonia to the boiling solution. The precip- itate is filtered off, washed and weighed as Cr 2 O 3 . The filter containing the insoluble residue from the above fusion is returned to the same crucible in which the fusion was made and ignited. The ignited oxides are dissolved in hydro- chloric acid, the solution diluted and the copper precipitated by hydrogen sulphide. The filtrate is evaporated with sulphuric acid until the hydrochloric acid is all expelled, whereupon it is *Campagne, Compt. rend., 137, 570 (1903). 316 GRAVIMETRIC ANALYSIS. diluted and treated with a large excess of ammonia and the nickel determined by electrolysis (see p. 136). The ether solution from the two separately funnels is shaken with water, which causes the separation of an ether layer from the solution containing the iron and molybdenum. The lower layer is, therefore, drawn off; the solution of ferric chloride containing all the molybdenum is evaporated nearly to dryness, the cold solution treated ^vith 10 c.c. of concentrated sulphuric acid and evaporated until the sulphuric acid fumes freely. The cold sulphuric acid solution is diluted with 100 c.c. of water, reduced by the careful addition of ammonium acid sulphite, the excess of sulphurous acid boiled off, and the cold solution saturated with hydrogen sulphide in a 200 c.c. pressure bottle. The bottle is stoppered and heated on the water bath for several hours. After slowly cooling, the precipitate is filtered into a Gooch crucible and washed with water containing a little sul- phuric acid and finally with alcohol. The Gooch crucible is placed within a larger porcelain crucible so that the bottom of the former does not touch that of the latter, covered with a watch- glass and heated gradually until there is no more odor of sulphur dioxide. The watch-glass is then replaced by a porcelain crucible cover and the heating is continued until the ignited precipitate becomes bluish white in color. The Gooch crucible is then heated to faint redness, cooled and weighed. The heating and weighing is repeated until the pre- cipitate ceases to lose in weight. The crucible is then placed in the suction bottle and washed with ammonia until the washings are free from molybdenum. The crucible is again heated and weighed. The difference in weight corresponds to the amount of molybdenum trioxide. A small amount of ferric oxide always remains on the felt. f SILVER. 3 X 7 METALS OF GROUP I. SILVER, LEAD, MERCUROUS MERCURY (AND THALLIUM). The determination of lead and mercury has already been considered; it remains for us to discuss the determination of silver. SILVER, Ag. At. Wt. 107.88. Forms: AgCl and Ag. Determination as Silver Chloride, AgCl. The solution, slightly acid with nitric acid, is heated to boiling and the silver precipitated by the addition of hydrochloric acid, drop by drop, until no more precipitate is formed. The precipitate is allowed to settle in a dark place, filtered through a Gooch cruci- ble and washed, first with water containing a little nitric acid until the chloride test can no longer be obtained, then twice with alcohol or water in order to remove the nitric acid. The precipitate is dried first at 100 C. and finally at 130 C. till a constant weight is obtained. If it is not desired to use a Gooch crucible for this determination, the silver chloride can be filtered upon an ordinary washed filter, washed as before and dried at 100 C. As much of the precipitate as possible is transferred to a weighed porcelain crucible, the filter burned (as described on page 21) in a platinum spiral whereby some of the silver chloride adhering to it will be reduced to metal. The ash of the filter is added to the main por- tion of the precipitate. It is moistened with a little nitric acid and a drop or two of concentrated hydrochloric acid, dried on the water- bath and then heated over a free flame until the silver chloride begins to melt. After cooling in a desiccator it is weighed. Solubility of Silver Chloride* One liter of water dissolves 0.00154 g. AgCl at 20 and 0.0217 g. at 100. In water con- taining a little hydrochloric acid, the AgCl is less soluble than in pure water but as the quantity of hydrochloric acid is increased, the solubility of AgCl rises rapidly. Thus one liter of 1 per cent. HC1 dissolves only 0.0002 g. AgCl at 21, but 1 1. of 5 per cent. HC1 dissolves 0.0003 g. and 1 1. of 10 per cent. HC1 dissolves 0.0555 g. AgCl. * G. S. Whitby, Z. anorg. Chem., 67, 108 (1910). 3i8 GRAVIMETRIC ANALYSIS Determination as Metal, Ag. Metallic silver is obtained by the ignition of silver oxido, car- bonate, cyanide or the salt of an organic acid. In the latter case, the substance must be heated very cautiously at first in a covered crucible. When the organic substance is completely charred, the cover is removed from the crucible and the contents are ignited until the carbon is completely burned, and the crucible then weighed. From the chloride, bromide (but not the iodide) and sulphide, the metal may be obtained by igniting in a current of hydrogen. The reduction of the chlorida, bromide, and iodide may be effect- ed very conveniently by passing the electric current through the substance after it has been melted together. The porcelain crucible containing the silver halide is placed in a crystallizing dish and near it is placed a second crucible containing a little mercury and a small piece of zinc. Upon the silver salt is placed a small disk of platinum foil, which is fastened to a platinum wire; the other end of the wire dips in the mercury in the other crucible. The crystallizing dish is filled with dilute sulphuric acid (1:20) so that the crucible is entirely covered with the acid and it is then allowed to stand over night. Next morning all of the silver salt will be found to be reduced. The crucible is removed from the acid, washed with water, dried, ignited, and weighed. By this simple method, E. Lagutt obtained excellent results. If the silver halide has not been fused to a compact mass small particles of the silver precipitate are likely to float around during the opera- tion, and escape reduction. Silver can also be deposited electrolytically, but this method will not be described in this book, for it offers no particular advan- tages over the determination as silver chloride with a Gooch cru- cible. Separation of Silver from Other Metals. ; As almost all metal chlorides * are soluble in dilute hydro- * Thallium chloride is difficultly soluble in water. If thallium is present the silver is precipitated from a nitrate solution by means of H 2 S, ignited in a stream of hydrogen, and weighed as metal. To determine the thallium, the filtrate is evaporated to dryness, the residue dissolved in a little water SILVER. 319 chloric acid, suver is separated from the other metals by the addi- tion of hydrochloric acid to the solution. If the solution contains mercurous salts these are oxidized before the addition of the hydrochloric acid by boiling with nitric acid. For the separation of silver from gold and platinum in alloys consult pages 259 and 270. and the thallium precipitated by the addition of potassium iodide. The thallium iodide precipitate is washed with dilute potassium iodide solution, then with alcohol, dried at 150 and weighed as Til. GRAVIMETRIC DETERMINATION OF THE METALLOIDS (ANIONS). GROUP I. HYDROCHLORIC, HYDROBROMIC, HYDRIODIC, HYDROCYANIC, HYDROFERROCYANIC, HYDROFERRICYANIC, SULPHOCY- ANIC, AND HYPOCHLOROUS ACIDS. HYDROCHLORIC ACID, HC1. Mol. Wt. 36.47. Form: Silver Chloride, AgCl. We can distinguish between two cases: A. The chloride solution is present either as free hydrochloric acid or as a chloride soluble in water. B. It is present in the form of an insoluble chloride. A. The Chloride is Present in Aqueous Solution. If only metals of the alkali or alkaline earth groups are present, the cold solution is made slightly acid with nitric acid, and silver nitrate is slowly added with constant stirring until the precipitate collects together and further addition of the reagent produces no more precipitation. The liquid is now heated to boiling, the pre- cipitate allowed to settle in the dark, filtered through a Gooch crucible, and then treated exactly as described in the determina- tion of silver, page 317. If the aqueous solution contains a chloride of a heavy metal, it is not always possible to follow the above procedure. If, for example, substances are present which on boiling are changed to Insoluble basic salts, it is evident that the precipitate of silver chlo- ride would be contaminated with these substances and too high results would be obtained. This is particularly true of stannic and ferric salts. Ferrous salts, on the other hand, in case only little 320 HYDROCHLORIC ACID. 321 nitric acid is present, reduce silver nitrate to metallic silver on heating the solution; if enough nitric acid is present to prevent the reduction to silver, the danger of forming basic salts still remains to be feared. In such cases the precipitation is effected as before from a cold solution and the subsequent heating is omitted. In all cases, however, it is better to first remove the heavy metal by precipitation with ammonia, caustic soda or sodium carbonate. Example : Analysis of Commercial Tin Chloride. Tin chloride is obtained either as a solid salt corresponding to the formula SnCl 4 + 5H 2 O, or as a concentrated aqueous solu- tion. As both the solid salt and its concentrated solution are very hygroscopic, it is necessary to weigh out the portion for analysis from a stoppared vessel. It is best to proceed as follows: A large sample of the substance (about 10 gm.) is placed in a tared weighing beaker, closed and weighed. About 10 c.c. of water are added, the salt is completely dissolved to a homogeneous syrup by shaking, and the beaker is again weighed. Four more weighing beakers are now tared and into each is placed about 2 c.c. of the syrup. Each beaker is quickly stoppered and then weighed. Determination of Tin. The contents of one of the weighing beakers is washed into a 400-500-c.c. beaker, diluted to about 300 c.c. and a few drops of methyl orange are added, whereby the liquid is colored red. Ammonia solution (free from chloride) is now added until the color of the solution is changed to yellow (an excess of ammonia must be carefully avoided for tin hydrox- ide is somewhat soluble in ammonia) . The solution is then treated with ammonium nitrate (5 c.c. of concentrated ammonia exactly neutralized with nitric acid, sp. gr. 1.2), boiled for one or two min- utes, filtered, and washed with water containing ammonium nitrate, and weighed as SnO 2 . Determination of Chlorine. The filtrate from the tin hydroxide precipitate is acidified with nitric acid, and precipitated in the cold with silver nitrate. The solution is then heated to boiling and, 322 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. after the precipitate has settled, it is filtered through a Gooch crucible, washed with cold water containing a little nitric acid, then with cold water or alcohol, dried at 130 C., and weighed. The amount of tin and chlorine present is computed as follows : Weight of Solid Salt= A. Weight of the Solid Salt + Water =B. Weight of the Solution taken for Analysis = a. Weight of the SnO 2 found=p. Weight of the AgCl found- p'. Since B gm. of the solution contain A gm. of the solid salt, then the amount a of the solution taken for analysis will contain : B:A=a:x x= =-= wt. of substance taken. tt This amount of substance yielded p gm. SnO 2 , corresponding to: SnO 2 :Sn=p:z' Sn0 2 and in percentage: A^.^P_ 100 B 'Sn0 2 ~ lOOSnp.S In the same way the amount of chlorine present is found to be: 100C1 P'-B This analysis may be accomplished much more rapidly by a volumetric process. (Consult Vol. Anal.) If antimony or stannous compounds are present, the above procedure cannot be used. It has been proposed to add tartaric acid to the solution, then dilute with water and precipitate the chlorine with silver nitrate. It is better, however, to proceed as follows: The antimony is precipitated by hydrogen sulphide as its HYDROCHLORIC ACID. 323 sulphide, the excess of the latter is removed by passing carbon dioxide through the solution, after which the precipitate is filtered and washed. The nitrate containing all the chlorine is made slightly ammoniacal, a little hydrogen peroxide or potassium per- carbonate is added (both reagents must be free from chloride) and the solution boiled until the excess of the peroxide is destroyed. By this treatment traces of hydrogen sulphide remaining in the solution are oxidized to sulphuric acid. After cooling, the solu- tion is acidified with nitric acid, and the chlorine determined as silver chloride as described above. According to this method, chlorine may be determined in the presence of large amounts of hydrogen sulphide without difficulty. It is less practical to proceed as follows : The solution is satu- rated with ammonia and the hydrogen sulphide is precipitated by the addition of ammoniacal silver nitrate solution, the deposited silver sulphide is filtered off, washed with ammonia, and the silver chloride precipitated from the filtrate by acidifying with nitric acid. B. Analysis of an Insoluble Chloride. The substance is boiled with sodium carbonate solution * (free from chloride) , and the chlorine determined in the filtrate as before. Many chlorides, e.g. silver chloride, many minerals such as apatite,f sodalite, and rocks containing the latter, are not decom- posed by boiling them with sodium carbonate. In such cases, the substance must be fused with sodium carbonate. Silver chloride should be mixed with three times as much sodium carbonate and heated in a porcelain crucible until the mass has sintered together. The mass is treated with water, the insoluble silver filtered off, and the chlorine determined in the fil- trate as under (a). For the determination of chlorine in rocks, 1 gm. of the finely- * Mercurous chloride is decomposed only slowly by sodium carbonate solution, but readily acted upon by potassium or sodium hydroxide. f According to Jannasch, chlorine in apatite may be determined by treat- ing the finely powdered mineral with nitric acid and silver nitrate on the water-bath. Everything goes into solution with the exception of silver chloride, which is filtered off and weighed. (This does not apply to a sample of apatite contaminated with silica or silicates.) 324 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. powdered material is fused with four or five times as much sodium carbonate (or with a mixture of equal parts sodium and potassium carbonates) at first over a Bunsen burner, afterward over a Teclu burner or the blast-lamp. The melt is extracted with hot water. After cooling, methyl orange is added, the solution is acidified with nitric acid and allowed to stand overnight. If silicic acid has precipitated out by the next morning, a little ammonia is added, the solution is boiled, filtered^ and washed with hot water. The cold filtrate is acidified with nitric acid and the chlorine determined as above. If there is no separation of silicic acid on acidifying the water extraction of the fusion with nitric acid,* the chlorine is precipi- tated at once from the cold solution. Free Chlorine. If it is desired to determine gravimetrically the amount of chlorine in a sample of chlorine water, it is not feasible to simplify add silver nitrate, for all of the chlorine is not precipitated as silver chloride; a part of it remains in solution as soluble silver chlorate: 6C1+ 3H 2 0+ 6AgN0 3 - 5AgCl+ AgClO 3 -}- 6HNO 3 . The chlorine, therefore, must be changed to hydrochloric acid or to one of its salts before attempting to precipitate with silver nitrate. This may be accomplished in several ways: 1 . A definite amount of the chlorine water is transferred by means of a pipette to a flask containing ammonia and after mixing the solu- tion is heated to boiling. After cooling the liquid is acidified with nitric acid and precipitated by silver nitrate. The ammonia con- verts the chlorine partly to ammonium chloride and partly to am- monium hypochlorite. The latter is decomposed partly in the cold and quantitatively on warming into ammonium chloride and nitrogen: (a) 2NH 4 OH + 2C1 - NH 4 C1 + NH 4 OC1 + H 2 O; (6) 3NH 4 OC1 + 2NH 3 = 3H 2 O + N 2 + 3NH 4 C1. * According to W. F. Hillebrand, there is no separation of silicic acid to be feared from 1 gm. of the substance. HYDROCHLORIC ACID. 325 2. The chlorine water is treated with an excess of sulphurous acid, the solution is made ammoniacal, hydrogen peroxide is added, and the liquid boiled until the excess of hydrogen peroxide is re- moved. After cooling the solution is acidified with nitric acid, diluted with water, and the chlorine precipitated by means of silver nitrate. 3. The chlorine water is treated with dilute caustic soda solu- tion, an aqueous solution of sodium arsenite is added (arsenic trioxide dissolved in sodium carbonate) until a dr.op of the liquid will not turn a piece of iodo-starch paper blue. It is then acidi- fied with nitric acid and the chlorine precipitated by a soluble silver salt. If the solution contains both free chlorine and hydrochloric acid, the total chlorine is determined by one of the above methods, while the free chlorine is determined in a separate sample by a volumetric process (see lodimetry). Determination of Chlorine in Non-electrolytes (Organic Compounds). A. Method of Carius* Principle. The method is based upon the fact that all organic compounds are decomposed by heating with concentrated nitric acid at a high temperature under pressure. If the substance con- tains halogen, sulphur, phosphorus, or arsenic, it is first set free as such, but on account of the reducing action of the nitrous acid formed it is then changed over into its hydrogen compound. The latter, however, is partly oxidized by the nitric acid. The reaction is therefore a reversible one. If, on the other hand, the substance is heated under the same conditions with nitric acid in the presence of silver nitrate, the halogen hydride is converted into silver halide as fast as it is formed and the halogen is in this case quantitatively changed into its silver salt. Sulphur, phosphorus, and arsenic are oxidized in the same way to sulphuric, phosphoric, and arsenic acids and any metals present form nitrates. * Ann. d. Chem. u. Pharm. (1865), 136, p. 129, and Zeit. f. anal Chem. t, p. 451. 326 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. Procedure for the Halogen Determination. A glass tube made of difficultly fusible glass (about 50 cm. long, 2 cm. in diameter, with walls about 2 mm. thick) is sealed at one end, thoroughly cleaned and dried by sucking air through it. About 0.5 gm. of powdered silver nitrate (or in the case of substances rich in halogen as much as 1 gm. may be used) is trans- ferred to thff tube by pouring the powder through a cylinder made by rolling up a piece of glazed paper and shoving the paper into the tube until it reaches about the middle of it. About 40 c.c. of pure nitric acid (sp. gr. 1.5) free from chlorine are poured into the tube through a funnel whose stem is about 40 cm. long. In this way only the lower half of the tube is wet with the acid. The tube is then inclined to one side and from 0.15-0.2 gm. of the substance contained in a small glass tube closed at one end is introduced into it (this smaller tube should be about 5 cm. long and 5 mm. wide). As soon as the tube containing the substance has- reached the acid, it remains suspended (Fig. 53, a). It is very important that the substance should not come in contact with the acid before the tube is closed at the upper end, as other- wise there is likelihood of some halogen escaping. The upper end of the tube is now heated very cautiously in the flame of the blast-lamp until the tube begins to soften and thicken (Fig. 53, b). It is then drawn out into a 3-5 cm. long, thick-walled capillary and the end fused together (Fig. 53, c). After the tube has become cold, it is enveloped in asbestos paper, carefully shoved into the iron mantle of a "bomb furnace, " and gradually heated. Aliphatic substances are usually decom- posed by heating four hours at 150-200 C; substances of the aromatic series usually require from eight to ten hours' heating at 250-300 C., while in some cases an even longer heating at a higher temperature is necessary. The time and temperature must be a FIG HYDROCHLORIC ACID. 327 found out for each substance by experiment. The decomposition is complete when on cooling the contents of the tube neither crystals nor drops of oil are to be recognized.* The heating is so regulated that after three hours the temperature of about 200 C. is reached, after three hours more 250-270 C., and finally after another three hours a temperature of about 300 C. is attained, j After the heating is finished, the tube is allowed to cool completely in the furnace, the iron mantle together with the tube is then removed, and by slightly inclining the mantle the capillary of the tube is brought out into the open air. In most cases a drop of liquid will bs found in the point of the latter. In order not to lose this, the outer point of the capillary is carefully heated with a free flame, and by this means the liquid is driven back into the other part of the tube. The point of the capillary is now more strongly heated J until the glass softens, when it will be blown out in consequence of the pressure within the tube. The gas escapes with a hissing sound. When the contents of the tube are at the atmospheric pressure, a scratch is made upon it with a file just below the capillary, and this is touched with a hot glass rod, whereby the tube usually breaks and the upper part can be removed. The contents of the tube are then carefully poured into a fairly large beaker without breaking the little tube in which the substance was weighed out, and the inner part of the tube as well as its capillary is washed out with water. The liquid in the beaker is diluted to about 300 c.c. and heated to boiling. After cooling, the insoluble silver halide is filtered off through a Gooch crucible, and after washing and drying at 130 C. its weight is determined. If it is thought that the precipitate is contaminated by frag- ments of broken glass, as is often the case even with careful work, the clear liquid is decanted through a filter, the residue washed by * Sometimes, with substances rich in sulphur, crystals of nitrosyl sulphuric acid are formed and adhere to the sides of the tube. They are easily distin- guished from crystals of the undecomposed substance. f Such a high pressure is often attained that the tube bursts as soon as it is heated very hot. In such cases it should be heated to only 200 C., allowed to cool, the capillary opened and the gas set free. It is then fused together again and heated to the desired temperature. J Before heating, the tube and the hand should be wrapped in a towel to avoid accidents. 328 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. decantation with very dilute nitric acid to the disappearance of the silver reaction, and the residue (except when it is silver iodide) is dissolved in warm ammonia water. The solution is filtered through the same filter, but the filtrate is this time collected in a fresh beaker. After washing the filter with dilute ammonia, the filtrate is acidified with nitric acid, heated to boiling, and after allowing the silver chloride or bromide to settle in the dark, it is filtered through a Gooch crucible, dried at 130 C., and weighed. In the case of silver iodide, it cannot be dissolved in ammonia and in this way separated from splinters of glass. In this case the substance, together with the glass, is filtered through an ordinary washed filter (not a Gooch crucible), completely washed with dilute nitric acid, then once with alcohol in order to remove the nitric acid, and dried at 100 C. As much of the precipitate as possible is transferred to a watch-glass, the filter burned, and its ash placed in a weighed porcelain crucible. A little dilute nitric acid is added (in order to change any reduced silver into the nitrate), the liquid is evaporated on the water-bath, a few drops of water and a drop of pure hydriodic acid are added, and the contents of the crucible are again evaporated to dryness, when the main part of the precipitate is added, heated until it begins to fuse, and then weighed. The mass in the crucible is then covered with pure dilute sulphuric acid, a piece of chemically pure zinc is added, and the crucible allowed to stand overnight. After this time the silver iodide will be com- pletely reduced to metallic silver. The zinc is removed, and the residue washed by decanting several times with water until the iodine reaction can no longer be detected. The residue is then warmed with dilute nitric acid upon the water-bath, in order to dissolve the silver, the solution is filtered through a small filter; and the latter is washed with water and dried. This filter is ig- nited in a crucible and the residue (the glass) is weighed. This second weight deducted from the former gives the amount of silver iodide present. This method is also suitable for obtaining lead and mercury from organic compounds in a form which can be precipitated by hydrogen sulphide. The method of Carius is by far the best for the determination of halogens in organic substances when only one of the halogens is HYDROBROMIC ACID. 329 present. If two or three of them are present at the same time, the " lime method " is to be preferred. The Lime Method. Into a glass tube made of difficultly fusible glass (about 40 cm. long, 1 cm. wide and closed at one end), a layer of lime (free from chloride) from 5 to 6 cm. long is introduced, then about 0.5 gm. of substance, and finally 5 cm. more of lime. The substance is then mixed thoroughly with the lime by means of a copper wire whose end is wound into a spiral. The tube is nearly filled with lime, placed on its side, and gently tapped so that a small canal is formed above the lime. The tube is then placed in a small combustion furnace (cf . Carbon) and heated. First of all the front end of the tube, free from substance, is heated to a dull redness, then the back end, and afterward the other burners are lighted one after another until finally the whole tube is at a dull-red heat. After cooling, the contents of the tube are transferred to a large beaker and the lime dissolved in dilute nitric acid free from chlorine. The carbon is filtered off, and the halogen precipitated with silver nitrate. If the lime contains calcium sulphate, this is reduced to sul- phide, so that some silver sulphide is likely to be precipitated with the silver halide. In this case the solution is treated with hydrogen peroxide (free from halogen) before enough nitric acid has been added to make the solution acid, the liquid is boiled to remove the excess of the reagent, then acidified, filtered, and precipitated with silver nitrate.* In the analysis of substances rich in nitrogen, it is possible that some soluble calcium cyanide will be formed. In this case care must be taken that the silver precipitate con- tains no silver cyanide (cf. Separation of Cyanogen from Chlorine- Bromine, and Iodine, p. 339). HYDROBROMIC ACID, HBr. Mol. Wt. 80.93. Form: Silver Bromide, AgBr. Hydrobromic acid is determined exactly the same as hydro- chloric acid. This is also true of the determination of free bro- mine, and bromine in non-electrolytes. * W. Biltz (Chem. Ztg. , 1903, Rep. 142) , separates the halides from sulphide by treating the precipitated silver salts with an ammoniacal sodium thiosul- phate solution, whereby the silver halide goes into solution, from which the silver is precipitated as silver sulphide, by adding ammonium sulphide, and determined as silver. 330 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. HYDRIODIC ACID, HI. Mol. Wt. 127.93. Forms: Silver Iodide, Agl, and Palladous Iodide, PdI 2 . (a) Determination as Silver Iodide. The determination of hydriodic acid is carried out in exactly the same way as the analysis of hydrochloric acid. If it is de- sired to 'filter the silver iodide through an ordinary washed filter instead of through a Gooch crucible, the procedure described on p. 328 is used, converting the reduced metal to iodide by dissolv- ing in nitric acid and adding hydriodic acid. In case there is no hydriodic acid at one's disposal, the main portion of the precipitate is placed in a weighed porcelain crucible and heated until it begins to melt and then weighed. The filter ash is placed in another crucible, and treated with nitric and hydrochloric acids, whereby the silver and any unreduced iodide are changed to silver chloride. The silver chloride is weighed and the equivalent amount of silver iodide is added to the weight of the main part of the precipitate. Example. Suppose a grams substance gave p grams silver iodide and p r grams silver chloride, then -Mv "AgCl p - AgT We have, therefore, in a grams substance p-f . p' grams silver Agbi iodide, and the amount of iodine present may be calculated in the usual manner. (6) Determination as Palladous Iodide. This important method for the separation of iodine from, bromine and chlorine is carried out as follows: The solution is acidified with hydrochloric acid, and palladous chloride solution is added until no more precipitate is formed. After standing one or two days in a warm place, the brownish- black precipitate of palladous iodide is filtered through a Gooch SEPARATION OF IODINE FROM CHLORINE. 331 crucible, or through a tared filter that has been dried at 100 C. f washed with warm water, dried at 100 C., and weighed as PdI 2 . According to Rose, the PdI 2 may be changed to palladium by igniting in a current of hydrogen, and from the weight of the palla- dium the amount of iodine calculated. SEPARATION OF THE HALOGENS FROM ONE ANOTHER. i. Separation of Iodine from Chlorine. (a) The Palladous Iodide Method. The iodine is determined as above as palladous iodide, and in a second sample the sum of the chlorine and iodine is determined from the weight of their insoluble silver salts. (6) Method of Gooch. This method depends upon the fact that in a dilute solution of the three halogens, nitrous acid sets free iodine alone : 2KI+ 2KNO 2 + 4H 2 S0 4 = 4KHSO 4 + 2NO+ 2H 2 0+ T 2 , which escapes from the solution on boiling. In one sample, there- fore, the halogens are precipitated together in the form of their silver salts, in a second sample the amount of the chlorine is deter- mined after setting free the iodine by means of nitrous acid, and the amount of iodine determined by difference. In order to obtain correct results by this method, the solution must be very dilute when it is boiled to expel the iodine; otherwise some chlorine escapes. Procedure. The mixture of the halogen salts (about 0.5.gm. of the substance should be dissolved in 600-700 c.c. water in a liter flask) is treated with 2-3 c.c. of dilute sulphuric acid, 0.5-1 gm. of solid potassium nitrite (free from halogen) is added, and the solution is boiled until entirely colorless ; in most cases this is accomplished in about three-quarters of an hour. The contents of the flask are now treated with silver nitrate solution, and the resulting precipitate is allowed to settle. It is filtered through a Gooch crucible, and weighed. 332 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. (c) Method of Jannasch* Jannasch proceeds in exactly the same way as Gooch, but in- stead of letting the iodine escape, he collects in it a mixture of caustic soda and hydrogen peroxide, whereby it is transformed to sodium iodide and is subsequently determined as silver iodide. In the other solution the chlorine is determined in the usual way. Procedure. The solution containing the two halogens is placed in a IJ-liter round-bottomed flask and diluted to a volume of 600-700 c.c. Like a wash-bottle, this flask is provided with one glass tube reaching to the bottom, through which vapor can be conducted into the flask, and with another shorter tube for the escape of gas. This second tube is connected with an Erlenmeyer flask for a re- ceiver, and this is in turn connected with a Peligot tube. About 50 c.c. of pure 5 per cent, caustic soda solution are placed in the Erlenmeyer flask, an equal volume of hydrogen peroxide free from chlorine is added, and the mixture cooled by surrounding the flask with ice or snow. The Peligot tube is likewise filled with a suitable amount of caustic soda and hydrogen peroxide. From 5-10 c.c. of dilute sulphuric acid (1:5) and 10 c.c. of 10 per cent, sodium nitrite solution are now added to the solution containing the halo- gens, the flask is immediately closed, and the contents of the flask are heated over a free flame while water vapor is at the same time conducted into it. As soon as the liquid begins to boil, the space above is filled with violet vapors of iodine, which are gradually driven over into the Erlenmeyer flask, where, with evolution of oxygen, they are completely absorbed by the hydrogen peroxide solution. The iodine is changed into sodium iodide and sodium hypoiodite by means of the dilute alkali : 21 + 2NaOH = Nal+ NaOI+ H 2 O. The sodium hypoiodite. however, is reduced by the hydrogen peroxide to sodium iodide: NaOI + H 2 O 2 = H 2 + O 2 + Nal. When all the iodine is driven over into the receiver (which is always the case after the solution in the flask has become colorless * Zeit. fiir anorg. Chem. I, p. 144, and Prakt. Leit. der Gewichts-analyse, p. 182 et seq. SEPARATION OF IODINE FROM CHLORINE. 333 and has been boiled for twenty minutes longer), the delivery-tube between the distilling-flask and the Erlenmeyer flask is removed, the liquid within it is washed with hot water into the Erlenmeyer and the current of steam is stopped. The contents of the Peligot tube are added to the Erlenmeyer flask and the solution heated to boiling in order to remove the excess of hydrogen peroxide. After cooling, the liquid is acidified with a little sulphuric acid; this always causes a yellow coloration due to free iodine.* The solution, there- fore, is treated with a few drops of sulphurous acid, whereby it is completely decolorized. An excess of silver nitrate and a little nitric acid are then added, the liquid is boiled, and the silver iodide filtered through a Gooch crucible and weighed. For the chlorine determination, the contents of the distilling- flask are transferred to a beaker and the chlorine determined as silver chloride. Remark. This method has been carefully tested in the author's laboratory by O. Brunner, and in the above form has been found to give very exact results. Jannasch recommends a slightly different procedure. He adds silver nitrate directly to the alkaline solution in the Erlen- meyer flask. In this way accurate results are obtained provided there is no iodate formed by the absorption of the iodine. In the latter case, due to insufficient cooling of the contents of the receiver, the addition of silver nitrate results in the formation of some silver iodate, and this amount of iodine escapes determination, for silver iodate is soluble (though difficultly so). In such cases the results obtained are too low. If the solution is acidified, how- ever, the presence of the iodate is shown by the separation of a little iodine, and this can be changed by sulphurous acid to iodide, and accurate results will be obtained. * If the above directions are closely followed, there should not be much separation of iodine. It may be caused by the presence of a small amount of nitrous acid which is not oxidized to nitric acid by hydrogen peroxide or if the contents of the Erlenmeyer flask are not kept cool, appreciable amounts of sodium iodate (NaIO 3 ) are formed, and the latter is not reduced by hydro- gen peroxide. In this case there is a separation of a considerable amount of iodine on acidifying the solution, but the addition of sulphurous acid changes it to iodide without loss. 334 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. Determination of the Halogens by Indirect Analysis. (a) Determination of Bromine together with Chlorine. Principle. In this method the sum of the weights of the silver salts of the two halogens is first determined and afterwards the silver bromide converted to silver chloride by heating in a current of chlorine. Procedure. The solution containing about 0.5 gm. of the halogen salt is acidified with a little nitric acid (free from chlorine) and precipitated in the cold by the addition of a slight excess of silver nitrate. The liquid is heated to boiling, with frequent stir- ring, and after cooling again, the precipitate is filtered through a 15 cm. long, asbestos filter-tube made of difficultly fusible glass. The precipitate is dried at 150 C. and weighed after cooling. For the transformation of the bromide into chloride, the asbes- tos is shoved forward a little in the tube by means of a glass rod (in order that the gas may pass through it more readily), the tube is fastened in a slightly inclined position, and a current of dry chlorine gas is passed through it. At the same time the tube is heated cautiously by moving a small flame back and forth. During the first half hour the precipitate should not be heated hot enough to melt it; finally, however, the temperature is raised until it begins to melt, after which the chlorine is replaced by air, and after cooling the residue is again weighed. If p represents the combined weight of the two silver salts, and q the weight after the silver has been completely changed to chloride, then AgCl AgBr 1. x + y = p 2. x + my = g (AgCl) and from this it follows: In this equation m=g-. = 0.7633. If this value is substituted in equation (3), we obtain (AgBr)4, s, can be computed according to the proportion. Mg 2 P 2 O 7 :2PO 4 = p:s 2P0 4 M g2 P 2 7 ' P ' Solution and Reprecipitation of the Ignited Magnesium Pyrophosphate. If it is desired to dissolve the ignited precipitate and to repre- cipitate the phosphoric acid, the crucible together with its cover, is placed in a beaker, enough water is added to cover the crucible, and then an excess of concentrated hydrochloric acid. The beaker is covered with a watch-glass and its contents are heated on the water-bath, the liquid in the beaker being occasionally rotated. When the precipitate has dissolved, the heating is continued for three or four hours longer in order to make sure that the pyro- phosphoric acid is completely changed to orthophosphoric acid. This change is always complete at the end of this time if the \\ eight of the magnesium pyrophosphate was not over 0.2 gm. The time necessary to effect this transformation is proportional to the amount of nitric acid used After the liquid has been sufficiently heated, the crucible and its cover are removed, washed oft", from 2 to 5 c.c. of magnesia mixture are added, and the solution is treated, as described above, with 2J per cent, ammonia, etc. Th3 method descried on page 434 for the precipitation of phosphoric ac.J i; not app icable when tlie substance contains alkaline eartas or heav / met .Is. In such cases the phosphoric 43 6 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. acid should be precipitated first as ammonium phosphomolybdate and the phosphoric acid in this precipitate determined by one of the following methods. i. Determination of Phosphoric Acid as Magnesium Pyro- phosphate after Previous Precipitation as Ammonium Phosphomolybdate. This method, first proposed by*Sonnenschein, has experienced, in the course of time, a great many modifications, and of these, that of Woy * will be described, for it is one of the quickest and most accurate. It may be mentioned that the molybdate method is always applicable when the phosphoric acid is present as ortho- phosphate, irrespective of what metals are in solution. Principle. If a solution containing phosphoric acid, in the pres- ence of ammonium nitrate and sufficient nitric acid, is treated with a slight excess of ammonium molybdate and heated just to the boiling-point, all of the phosphoric acid is immediately precipitated as yellow ammonium phosphomolybdate. According to Hunde- shagen, the precipitate possesses the following composition: (NH 4 ) 3 PO 4 - 12MoO 3 2HNO 3 H 2 O, and always contains, when sufficient molybdic acid is present 24 mols. of MoO 3 to 1 mol. P 3 5 . It never contains more molybdic acid than corresponds to the above formula, but is always some what contaminated with small amounts of the bases in solution, even when only alkalies are present. If, however, after decanting off the supernatant liquid, the precipitate is dissolved in ammonia, a little more ammonium molybdate added, and the boiling solution re- precipitated by the addition of nitric acid, it is then obtained pure. It must also be noted that the solution may contain neither silicic acid nor organic substances f and only a small amount of * Chem. Zeit., 21, p. 442. t According to Hundeshagen, (Zeit. f. anal. Chem., 28, p. 164) and Eggertz (Jour. f. prak.Chem.,79, p. 496) the presence of tartaric and oxalic acids hin- ders the formation of the yellow precipitate, and in some cases prevents it entirely. According to Hans v. Jiiptner (Oesterr. Zeit. fur Berg- u. Hiittenw., 1894, p. 471) this is not the case; he even recommends that tartaric acid be added for the determination of phosphorus in iron, on the ground that it prevents the precipitate being contaminated with molybdic acid and ferric oxide. PHOSPHORIC ACID. 437 chloride (best none at all), but there must be considerable free nitric acid present; 1 gm. of P a O 5 requires 11.6 gms. of HNO 3 , but as much as 35.5 gm. of the latter acid does no harm.* The precipitate will dissolve somewhat if more nitric acid than the above quantity is used, but the addition of ammonium molyb- date decreases the solubility of the precipitate in nitric acid; 1 gm. of ammonium molybdate makes 55.7 gms. of nitric acid inactive. The presence of ammonium nitrate not only facilitates the forma- tion of the precipitate, but its presence is absolutely necessary, although about 5 per cent, is sufficient. Solutions Required. 1. A 3 per cent, solution of ammonium molybdate obtained by the solution of 120 gms. commercial ammonium molybdate, (NH 4 ) 6 Mo 7 O 24 +4H 2 0, in 4 liters of water (1 c.c. of this solution will precipitate 0.001 gm. P 2 O 5 ). 2. A solution of ammonium nitrate, obtained by dissolving 340 gms. of ammonium nitrate in 1 liter of water. 3. Nitric acid, sp. gr. 1.153 (containing 25 per cent. HNO 3 ). 4. As wash liquid, 200 gms. ammonium nitrate and 160 c.c. of nitric acid dissolved in 4 liters of water. Woy's Method of Precipitation. In all cases 50 c.c. of the solution are taken, containing at the most 0.1 gm. P2O 5 . If the solution contains more than this amount of phosphoric acid, an aliquot part is used for the analysis. This amount of the neutral or slightly acid (HNOs) solution is placed in a 400-c.c. beaker and to precipitate 0.1 gm. of P2Os 30 c.c. of ammonium nitrate solution and 10-20 c.c. of nitric acid are added and the solution is heated until bubbles begin to rise. At the same time the required amount of ammonium molybdate * These figures are taken from experimental data furnished by Hun- deshagen. They do not refer to the formula on p. 436 given the yellow precipitate. [Translator.] 43 8 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. solution (in this case 120 c.c.*) is likewise heated until it begins to boil, and then transferred to a separatory funnel and allowed to run in a thin stream into the middle of the phosphate solution, which is rotated while the molybdate solution is being added. The yellow ammonium phosphomolybdate is at once thrown down and the separation is quantitative. The contents of the beaker are kept in motion for about one minute more and then allowed to stand for fifteen*minutes ; when the clear liquid is poured through a filter, the precipitate is washed once by decanta- tion with 50 c.c. of the wash liquid and then dissolved in 10 c.c. of 8 per cent, ammonia. To this solution 20 c.c. of the ammonium nitrate solution, 30 c.c. of water, and 1 c.c. of ammonium molybdate are added. It is heated, as before, until bubbles begin to rise, when the phosphoric acid is reprecipitated by the addition of 20 c.c. of hot nitric acid, added drop by drop through the same funnel that was used for the molybdate solution, the solution being rotated as before. The precipitate is immediately formed and is now pure. After standing ten minutes it is filtered off and dissolved in warm 2J per cent, ammonia, after which hydrochloric acid is added until the yellow precipitate produced dissolves only slowly on being mixed with the solution. Now, according to Schmitz,f an excess of an acid solution of " magnesia mixture " is added, and the solution heated to boiling. A few drops of phenol- phthalein are added, and an approximately 2.5 per cent, ammonia solution introduced as quickly as possible from a burette, while stirring the solution, until the liquid becomes slightly red in color. It is allowed to cool and then one-fifth of its volume of concen- trated ammonia is added. After ten minutes, the precipitate of magnesium ammonium phosphate is ready to filter. *AMOUNTS OF REAGENTS REQUIRED. Amoun f of PoOs Present in Grams. Ammonium Molybdate. Ammonium Nitrate. Nitric Acid. 0.1 120 c.c 30 c.c 19 C.C. 0.01 15 " 20 " 10 " 005 15 " 20 10 " 002 10 " 15 " 5 0.001 10 " 15 " 5 " jLoc. cit. PHOSPHORIC ACID. 439 2. Direct Determination of Phosphoric Acid as Ammonium Phosphomolybdate (Finkener) . * The precipitate produced as described under 1, having the following composition, (NH 4 ) 3 P0 4 12Mo0 3 - 2HN0 3 H 2 O, is transformed by heating for a long time at 160-180 C. into pure ammonium phosphomolybdate of the composition (NH 4 ) 3 P0 4 .12Mo0 3 . Theoretically this substance contains 3.782 per cent, of P 2 5 . If, therefore., the amount of yellow precipitate (dried until its weight is constant) is multiplied by 0.0378, the actual amount of P 2 O 5 present should be obtained. The results obtained by Finkener, however, were accurate only when the factor 0.03794 f was used. Hundeshagen ; | on the other hand, found that the factor 0.03753 should be used, and this has been confirmed by experiments per- formed in the author's laboratory.! Procedure. The phosphoric acid is precipitated twice, accord- ing to the directions of Woy (p. 437), with ammonium molyb- date; the precipitate is filtered through a Gooch crucible, washed with the prescribed mixture until no further brown coloration is produced by K 4 Fe(CN) 6 , and dried in a current of air at 160 C. in a Paul's drying oven, until a constant weight is obtained. If the precipitate should become slightly greenish, a small crystal of ammonium nitrate and one of ammonium carbonate are added and the contents of the crucible again heated, whereby the pre- cipitate will at once assume a homogeneous yellow color. Remark. The results of Hundeshagen and Steffan show that this method gives very exact results. Steffan worked precisely according to the directions of Finkener, precipitating the phos- phoric acid in the cold with a 33J per cent, solution of ammonium * Berichte, 11 (1878), p. 1640. f-Loc. cit. J Zeit. f. anal. Chem., XXXII (1893), p. 144. A. Steffan, using 50 c.c. of a potassium phosphate solution containing 0.0989 gm. P 2 O 5 , in four experiments found 0.0994, 0.0994, 0.0995, 0.0992 gm. P 2 5 . 44 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. molybdate and filtering after standing twenty-four hours. It is, however, not necessary, as Hundeshagen has shown, to work with such a concentrated solution of ammonium molybdate; the pre- cipitation from a hot solution with a 3 per cent, molybdate solu- tion yields just as accurate results and the solution does not have to stand so long before filtering. Even when iron is present this method gives good results, so that it is to be recommended for the determination of phosphorus in iron and steel. 3. Determination of Phosphoric Acid as Phosphomolybdic Anhydride (Woy). The precipitate, produced in the same way as before, is gently ignited, whereby a greenish-black residue remains of the com- position 24Mo0 3 -P 2 O 5 , wi th 3.946 per cent, of P 2 5 . The pre- cipitate is ignited as follows: Upon the bottom of a nickel cru- cible is placed a disk of ignited asbestos paper about 2 mm. thick, or the porcelain plate of a Gooch crucible may be used. Upon this is placed the Gooch crucible containing the precipitate, which is covered with a watch-glass and heated at first gently and finally until the bottom of the nickel crucible is at a dull-red heat. When the precipitate has become of a homogeneous, bluish-black color, it is allowed to cool in a desiccator, after which the covered cru- cible is weighed. This method is rapid and gives good results in the presence of iron and aluminium.* Determination of Phosphorus and Silicon in Iron and Steel. The determination of these two elements is often effected in the same sample, and in all cases it is best to remove the silicic acid before precipitating the phosphoric acid. Since phosphorus and silicon are present in the iron as phos- phide and silicide, a too dilute nitric acid must not be used for dis- solving the sample or there will be a loss of volatile phosphides and silicides. * Steffan found, in the analysis of SOc.c.of a potassium phosphate solu- tion containing 0.0989 gm P 2 O 5 , 0.0988, 0.0992, 0.0986 gm. P 2 O 5 ; and in a solution of 5 gms. of iron in the form of its nitrate, 0.0099 gm. P 2 O 5 , this method gave 0.0099 gm. and the same result was obtained by the method of Finkener. PHOSPHORUS AND SILICON IN 7/?O.V AND STEEL. 441 Determination of the Silicon. About 5 gms. of the iron borings, after having been washed with ether (cf. p. 236, foot-note), are placed in a 500-c.c. beaker under a good hood, covered with 60 c.c. of nitric acid (1 vol. concentrated acid, sp. gr. 1.4, and 1 vol. of water) and a watch-glass placed upon the beaker. A violent reaction at once takes place and brown vapors are evolved. As soon as the action slackens, the beaker is placed upon wire gauze and its contents boiled gently until all the iron is dissolved and no more brown vapors are evolved. The contents of the beaker are then washed into a 250-c.c. porcelain casserole, evaporated on the water- bath to a syrupy consistency, and then heated over a free flame to dryness, constantly stirring with a glass rod. Care is taken during this operation that a cake of basic ferric nitrate does not adhere to the bottom of the dish, as in this case the latter will surely break during the subsequent ignition. The dry mass should at the end be reduced to a loose powder. When this point is reached, the contents of the dish are ignited until all of the ferric nitrate is changed to oxide, which is accomplished when no more brown fumes are expelled. By this procedure all organic matter formed by the oxidation of hydrocarbons is destroyed and the silicic acid is dehydrated. After cooling, the residue is covered with 50 c.c. of concentrated hydrochloric acid and heated with constant stirring almost to the boiling-point. This dissolves the ferric oxide and phosphate, while the silicic acid remains behind.* When all of the iron oxide has dissolved, the solution is evap- orated to dryness, moistened with 2-3 c.c. of hydrochloric acid, allowed to stand for twenty minutes, after which water is added. After heating the liquid to boiling the silicic acid is filtered off through a small filter, washed with water containing hydrochloric acid and finally with pure hot water. The silica is ignited wet in a platinum crucible and weighed. The silica thus obtained usually contains ferric oxide, so that its purity must be tested in all cases. For this purpose it is covered with 1 c.c. of water, a drop of dilute sulphuric acid and 2 c.c. of pure hydrofluoric acid are added, and after evaporating on the water-bath as far as possible, the excess of * If graphite were originally present it remains with the silica. 442 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. sulphuric acid is removed by placing the crucible on a triangle in an inclined position and carefully heating by means of a moving flame. As soon as no more vapors of sulphuric acid are given off, the contents of the crucible are more strongly ignited and the residue of ferric oxide is weighed. This amount deducted from the weight of impure silica gives the amount of pure silica, p t from which the amount of silicon, x, can be calculated as follows: Si /y _ _ /yi " p and in per cent., where a is the amount of iron taken for the analysis, 100 Si p x' = -i- = per cent. Si. SiO 2 a Remark. If the impure silica was grayish colored (as is always the case when graphite is present) it is not weighed, but a little pure sodium carbonate and potassium nitrate are added to the contents of the crucible and, by fusing, the graphite is completely oxidized. The melt is placed in a small porcelain dish and dissolved in water. The solution is acidified with hydrochloric acid, evap- orated to dryness on the water-bath, moistened with a little con- centrated hydrochloric acid, diluted with water and filtered. The residual silica is ignited wet; a further purification of the silica is unnecessary. Drown Method for Determining Silicon in Iron and Steel. This method has come into very general use, and is much more rapid than the above method, though quite as exact. It is recommended by the American Foundrymen's Association for the analysis of cast iron and has been used by the Bureau of Standards at Washington, D. C., for analyzing samples of steel. One gram of borings is treated in a platinum or porcelain dish with 20 c.c. of nitric acid, sp. gr. 1.2. When all action has ceased PHOSPHORUS AND SILICON IN IRON AND STEEL. 443 20 c.c. of 50 per cent, sulphuric acid are added and the solution evaporated until copious fumes are evolved. The liquid is then cooled, diluted with 150 c.c. of water, and heated until all the ferric sulphate has dissolved. The hot solution is at once filtered, washed with dilute hydrochloric acid, sp. gr. 1.1, and then with hot water. The residue is placed in a platinum crucible without drying, ignited and weighed. The contents of the crucible are then treated with 4 or 5 c.c. of hydrofluoric acid and a few drops of sulphuric aicd, evaporated "to dryness, and the crucible again ignited and weighed. The difference in the two weights is the silica. Determination of Phosphorus. In the hydrochloric acid filtrate from the silicon determina- tion (p. 441) all the phosphorus is present in the form of phosphoric acid. The latter is determined according to (a) The Acetate Method or (&) The Molybdate Method. Both methods give equally good results, judging from experiments performed in the author's laboratory. (a) The Acetate Method. The filtrate from the silicic acid is diluted in a beaker to a volume of about 400 c.c. and ammonia is added until a permanent precipi- tate of ferric hydroxide is produced. The liquid is then treated with 200 c.c. of a saturated, aqueous solution of sulphurous acid and slowly heated to boiling. The precipitate of ferric hydrox- ide soon dissolves and the liquid assumes a dark reddish-brown color, which on further heating becomes a light green, or almost colorless. As soon as this point is reached, 10-20 c.c. of concen- trated hydrochloric acid are added and a current of carbon dioxide is conducted into the colorless solution until the excess of sul- phurous acid is removed. The solution is now cooled by placing the beaker in cold water, after which 1 or 2 c.c. of chlorine or bromine water is added to oxidize a part of the iron. To this solution ammonia is added very slowly with constant stirring until the greenish precipitate of ferrous-ferric hydroxide dissolves with difficulty. The addition is then continued drop by drop until a distinct red or brown precipitate is formed and then, on adding another drop of ammonia, the whole precipitate appears green. 444 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. If before this occurs the precipitate does not appear decidedly red in color, it is dissolved in a drop of two or hydrochloric acid and 1 or 2 c.c. more of chlorine or bromine water is added, and the addition of ammonia is repeated until the red and then the permanent green precipitate is obtained. Acetic acid (sp. gr. 1.04) is now added drop by drop until the green precipitate is dissolved and the remaining precipitate is quite red in color. About 1 c.c. of acetic acid in excess is added and the solution is boiled for one minute. All the phosphoric acid is then pre- cipitated as white ferric phosphate and the excess of ferric iron as red basic acetate. The greater part of the iron remains in solution as ferrous salt. The solution is filtered promptly through a large filter and washed once with hot water. The pre- cipitate filters readily and the filtrate is at first clear, but becomes turbid on standing in the air. The precipitate adhering, to the sides of the beaker is dissolved by warming with a mixture of hydrochloric acid (1:1) and 10 c.~. of bromine water. Should this not be sufficient to effect com- plete solution (as is usually the case) enough concentrated hydro- chloric acid is added to accomplish this. The solution is then poured upon the filter containing the precipitate and the filtrate received in a small beaker. The filter is washed well with hot water and the solution is evaporated nearly to dryness to get rid of the excess of hydrochloric acid, 5 c.c. of a 50 per cent, citric acid solution are added, an equal amount of magnesia mixture and enough ammonia to make the solution faintly alkaline. When perfectly cold, one-half of the liquid's volume of strong ammonia is added and' the mixture well stirred. After standing twelve hours, the precipitate is filtered off and washed with 2J per cent, ammonia containing 2.5 gms. of ammonium nitrate in each 100 c.c. This precipitate of magnesium ammonium phosphate always con- tains a small amount of iron and silicic acid (the latter from the glass) so that it is dissolved in hydrochloric acid, the solution evapo- rated to dryness, the residue moistened with concentrated hydro- chloric acid, taken up in a little water, filtered through a small filter and the residual silica washed with hot water. The filtrate, amounting to not over 20 c.c. at the most, is treated with 1 c.c. of the citric acid solution and two drops of magnesia mixture and the precipitation with ammonia is repeated as above. In DETERMINATION OF PHOSPHORIC ACID IN SILICATES. 445 this way a precipitate is obtained which yields pure magnesium pyrophosphate on ignition. Remark. A. A. Blair* recommends the use of ammonium bisulphite (NH^HSOs) instead of sulphurous acid for the reduction of the ferric salt. Much of the ammonium bisulphite of commerce, however, contains phosphoric acid, so that it seems safer to use sulphurous acid for this purpose. Again, Blair suggests that hydrogen sulphide be passed into the solution after the excess of the sulphurous acid has been removed,in order to precipitate any arsenic as the trisulphide. The filtrate from the arsenic precipitate is heated to boiling, the excess of hydrogen sulphide expelled by means of a current of carbon dioxide, and the solution then partly oxidized as above described. (6) The Molybdate Method. The filtrate from the silica (see p. 441) is evaporated to dry- ness in a porcelain dish, the dry residue is dissolved in as little nitric acid as possible, 30 c.c. of ammonium nitrate solution and 10 c.c. of nitric acid are added, and the phosphoric acid is precipi- tated according to the procedure of Woy, p. 437, by the addition of 75 c.c. of ammonium molybdate. After decanting off the clear liquid, the precipitate is washed once by decantation with 10-20 c.c. of the prescribed wash liquid and redissolved in a little ammo- nia. To this solution 6 c.c. of molybdate solution and 30 c.c. of water are added; it is heated just to the boiling-point and re- precipitated by the addition of 20 c.c. of hot nitric acid. The precipitate is then analyzed by the method of Finkener (p. 439) or by that of Woy (p. 440) . 1 gm. Mg 2 P 2 O 7 = 0.27848 gm. P (NH 4 ) 3 P(Vl2MoO 3 = 0.01639 " " P 2 O 5 '24MoO 3 =0.01723 " " Remark. According to the above directions, some difficulty is likely to be encountered at the stage where the dry residue is taken up in nitric acid. If the residue is overheated at all, it dissolves very slowly in the nitric acid owing to the formation of basic ferric salts. For this reason many chemists prefer to carry out the analysis in accordance with the directions of the American Foundrymen's Association, which are as follows: * The Chemical Analysis of Iron, 7th edition, 1908. 446 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. A 2-gm. sample of the borings is dissolved in 50 c.c. of nitric acid, sp. gr. 1.13, and 10 c.c. of hydrochloric, sp. gr. 1.2. In case the sample contains a fairly high percentage of phosphorus, it is advisable to use half the above quantities of sample and reagents. The solution is evaporated to dryness and the residue baked until free from acid, at a temperature of about 200. This baking serves to oxidize carbonaceous matter which otherwise interferes with the precipitation of the prTosphorus. The residue is dis- solved by heating it with 25-30 c.c. of concentrated hydrochloric acid; the solution is diluted to about 60 c.c. and filtered. The filtrate is evaporated to about 25 c.c., 20 c.c. of concentrated nitric acid are added, and the evaporation is repeated until a film begins to form. At this point 30 c.c. of nitric acid, sp. gr. 1.2, are added and once more the solution is evaporated until a film forms. It is then diluted with hot water to a volume of about 150 c.c. and allowed to cool somewhat. When at a temperature between 70 and 80 C., 50 c.c. of ammonium molybdate solution are added and the solution agitated for a few minutes. The precipitate is then filtered on a tared Gooch crucible which has a paper disc at the bottom. The precipitate is washed three times with 3 per cent, nitric acid and twice with alcohol. It is then dried at a temperature between 100 and 105 to constant weight. The precipitate contains 1.63 per cent, of phosphorus. The ammonium molybdate solution used in these last directions is prepared by dissolving 100 gms. of molybdic acid in 250 c.c. of water and 150 c.c. of concentrated ammonia, stirring until all is dissolved, whereupon 65 c.c. of nitric acid, sp. gr. 1.42, are added. Another solution is prepared containing 400 c.c. of the concentrated nitric acid and 1100 c.c. of water. When the two solutions are cold, the first is poured slowly into the second with constant stirring and a few drops of ammonium phosphate solution are added. After a little ammonium phosphomolybdate precipitate has settled out, the reagent is decanted off and is ready for use. The solution does not keep very well, so that the analysis should always be carried out with a reagent that has not stood very long. The phosphorus in iron and steel is very conveniently analyzed by a volumetric method. See Volumetric Analysis. SEPARATION OF PHOSPHORIC ACID FROM THE METALS. 447 Determination of Phosphoric Acid in Silicates. In the analysis of silicates (see p. 491) the phosphoric acid is found in the precipitate produced by ammonia in the filtrate from the silica together with iron and aluminium hydroxides. It is analyzed according to p. 111. Determination of Phosphoric Acid in Mineral Waters. The contents of a 56 liter flask is acidified with hydrochloric acid and evaporated to dryness, the residue is moistened with concentrated hydrochloric acid, taken up with water, and the silicic acid filtered off. The filtrate is precipitated with ammonia, by which means the phosphoric acid is usually completely thrown down in the form of phosphate of iron, aluminium, or alkaline earth. The filtered and washed precipitate is dissolved in nitric acid and the phosphoric acid present determined according to one of the molybdate methods (pp. 436-440) . Remark. If the mineral water does not contain much iron, aluminium, or alkaline-earth metal, but is rich in phosphoric acid and the alkalies, the precipitate produced by ammonia will not contain all of the phosphoric acid. In such a case the hydrochloric acid solution from the silica is evaporated several times to dryness with nitric acid, the residue is dissolved in as little nitric acid as possible, and the phosphoric acid determined by one of the molybdate methods. Recovery of Molybdenum Residues (H. Borntrager).* In practice the great majority of phosphoric acid determina- tions are carried out according to p. 436. The acid and am- moniacal filtrates containing molybdenum are saved, and the molybdenum is recovered as follows : Into a large, wide-mouthed flask 250 c.c. of strong ammonia are placed and the molyb- denum filtrates are added to this. Either immediately or after standing some time a crystalline deposit of almost pure molybdic acid is formed. When the flask is nearly full, the solution is made almost neutral, the precipitate allowed to settle, and the *Zeit. f. anal. Chem., XXXTII (1894), p. 341. 448 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. upper liquid containing only a small amount of molybdenum is poured off. The residue is poured upon a suction plate, washed once with water (not more, or the molybdic acid will dissolve) and sucked as dry as possible. The precipitate is dissolved by warming with as little ammonia as possible, leaving behind a residue of iron and aluminium hydroxides, magnesia, and silicic acid. These are filtered off and the solution diluted with dis- tilled water until at 17 C. it has a specific gravity of 1.11 = 14 Be. It then contains 150 gms. of ammonium molybdate in a liter. If this solution is diluted with four times as much water, a 3i per cent, solution will be obtained. Determination of Phosphorus in Organic Substances. The substance is decomposed by the method of Carius. By the action of the nitric acid in the closed tube the phosphorus is oxidized to phosphoric acid and this is determined as usual. SEPARATION OF PHOSPHORIC ACID FROM THE METALS. i. Separation from the Metals of Groups I and II. Hydrogen sulphide is conducted into the hydrochloric acid solution,* by which means all the members of these groups are precipitated as sulphides while the phosphoric acid remains in solution. 2. Separation from the Metals of Group III. (a) The phosphoric acid is first precipitated as ammonium phosphomolybdate according to p. 436. In order to determine the metals, the solution containing molybdenum, but free from phosphoric acid, is evaporated with the addition of sulphuric acid to a syrupy consistency, and carefully heated over a free flame until the nitric acid is expelled. After cooling, the residue is moistened with hydrochloric acid and taken up in water. The solution is placed in a pressure-flask, saturated with hydrogen sulphide, the flask stoppered and heated for some time on the water-bath, when the molybdenum is precipitated in large flocks. After cooling, the pressure-flask is slowly opened and the molyb- * When silver is present it is precipitated as silver chloride, filtered off, and the filtrate treated with hydrogen sulphide. SEPARATION OF PHOSPHORIC ACID FROM THE METALS. 449 denum sulphide is filtered off. The filtrate, now free from phos- phoric acid and molybdenum, is analyzed for the metals as described on pages 82 to 167. (b) The phosphoric acid is separated as before, the filtrate is made slightly ammoniacal and saturated with hydrogen sul- phide. After standing for some time the solution becomes red- dish yellow in color, when the precipitate is filtered off. The metals of this group will be found in the precipitate while the molybdenum is in the filtrate in the form of its sulpho-salt. Remark. If nickel is present, some of it will remain in the filtrate with the molybdenum on account of the solubility of nickel sulphide in ammonium sulphide, so that method (a) will then give more accurate results. 3. Separation of Phosphoric Acid from Iron, Cobalt, Manganese, and Zinc. In case the solution contains iron in the ferric form, it is acidi- fied with hydrochloric acid, saturated with hydrogen sulphide, and for each gram of the mixed oxides 3 gms. of tartaric acid are added; the solution is made slightly ammoniacal and allowed to stand overnight in a stoppered flask. The precipitate con- tains the metals as sulphides free from phosphoric acid. It is filtered, washed with water containing ammonium sulphide, dis- solved in acids, and analyzed according to pp. 150 and 156. 4. Separation from Chromic Acid. If the solution contains free alkali or alkali carbonate it is acidi- fied with nitric acid, then made slightly alkaline with ammonia and the phosphoric acid precipitated with " magnesia mixture " as described on page 434. 5. Separation from Calcium, Strontium, Barium, Magnesium, and the Alkalies. Ammonium carbonate is added to the hydrochloric acid solu- tion until a slight permanent turbidity * is produced, which is * If only alkalies are present there will be no turbidity, and the ammo- nium carbonate is added until the solution is neutral. 45 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. redissolved by a few drops of hydrochloric acid. Ferric chloride is then added drop by drop until the liquid above the yellowish- white precipitate of ferric phosphate becomes distinctly brown in color. The solution is diluted with water to a volume of 300 to 400 c.c., boiled for one minute, filtered and washed with water containing ammonium acetate. In the filtrate are now found the alkaline earths and alkalies, which, after expelling the am- monium salts by igniting the residue obtained after evaporating to dryness, is analyzed in the usual way (seepages 43 and 76 ff.). THIOSULPHURIC ACID, II 2 S 2 O 3 . Mol. Wt. 114.16. Form: Barium Sulphate, BaS0 4 . The aqueous solution of the alkali thiosulphate is treated with an ammoniacal solution of hydrogen peroxide, or with am- moniacal per carbon ate solution, heated for some time on the water- bath, and then boiled to destroy the excess of the reagent. This solution is acidified with hydrochloric acid and the sulphuric acid formed by the above treatment is precipitated as barium sul- phate. Two mols. BaSO 4 correspond to 1 mol. H 2 S 2 3 . A much better procedure for the estimation of thiosulphuric acid will be discussed under lodimetry, Part II. The remaining acids of this group, arsenious, arsenic, vanadic, and chromic, have been discussed under the respective metals, while periodic acid is analyzed precisely in the same way as iodic acid. DETERMINATION OF NITRIC ACID AS NITRON NITRATE. 451 GROUP V. NITRIC, CHLORIC, AND PERCHLORIC ACIDS. NITRIC ACID, HNO 3 . Mol. Wt. 63.02. Forms: Nitron Nitrate, C 2 oH 16 N 4 . HN0 3 , Nitrogen Pentoxide, N 2 O 5 ; Ammonia, NH 3 ; Nitric Oxide, NO, and Volumetrically. i. Determination of Nitric Acid as Nitron Nitrate.* The base diphenyl-endo-anilo-hydro-triazole, C2oHi6N4, or C 6 H 5 -N called " nitron " for the sake of brevity, forms a fairly insoluble, crystalline nitrate, C2oHi6N 4 -HN03, which can be used for the separation and quantitative estimation of this acid. Procedure. Enough of the substance is taken to furnish about 0.1 gm. of nitric acid, and dissolved in 80-100 c.c. of water with the addition of 10 drops of dilute sulphuric acid. The solution is heated nearly to boiling and treated with 10-12 c.c. of nitron acetate solution!, which is added all at one time. The beaker containing the solution and precipitate is kept surrounded by ice- water for about two hours. The precipitate is then transferred to a Munroe crucible and drained as completely as possible from the * M. Busch, Ber. 38, 861 (1905). A. Gutbier, Z. angew. Chem., 1905, 494. f The reagent is prepared by dissolving 10 gm. of nitron (which can be obtained of Merck) in 100 c.c. of 5 per cent, acetic acid. The solution usually has a reddish color, but can be kept for a long 'time in a dark-colored bottle without its undergoing any change. 452 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. pale yellow mother-liquor. It is washed with 10 or 12 c.c. of ice-water, added in small portions, and the precipitate drained well after each washing. The precipitate is dried at 110 to constant weight. It contains 16.53 per cent, of NO 3 . Remarks. The following acids interfere with the determination of nitric acid by the nitron method, hydrobromic, hydriodic, nitrous, chromic, chloric, perchloric and the less common thio- cyanic, hydroferrocyanic, hydrofejricyanic, picric and oxalic acids. All of the above acids form salts with nitron which are not very soluble; these acids must, therefore, be removed from the solution before precipitating the nitric acid. Hydrobromic acid is decomposed by adding chlorine water drop by drop to the neutral solution and boiling, until the yellow coloration entirely disappears. Hydriodic acid is removed by adding an excess of potassium iodate to the neutral solution, and boiling until the iodine is all expelled. Nitrous acid is removed by dropping finely powdered hydrazine sulphate into the concentrated solution (0.2 gm. of substance in 5 or 6 c.c. of water). Chromic acid is reduced by hydrazine sulphate. Some idea as to the relative solubilities of the various salts of nitron is obtained from the following table : 100 c.c. of slightly acid water dissolve at ordinary temperatures about 0.0099 gm. of nitron nitrate, corresponding to 0.0017 gm. HNO 3 0.61 0.017 0.19 0.06 0.12 0.008 0.04 bromide, iodide, nitrite, chromate, chlorate, perchlorate, thiocyanate, 0.125 0.005 0.022 0.011 0.022 0.002 0.007 HBr HI HNO 2 H 2 O 2 O 7 HC1O 3 HC1O 3 HCNS These values are only approximate. The solubility of the nitrate is given a little too high and that of the other salts a little too low. DETERMINATION OF NITRIC ACID. 453 On account of the apprec'able solubiity of the nitrate, it was to be expected that the results wouM be a little low. This is not the case, however, as Busch and Gutbier have proved. It is probable that the precipitate occludes a little nitron acetate and in his way the error caused by amount left in solution is compensated. 2. Determination of Nitric Acid as Nitrogen Pentoxide.* This method is based upon the fact that when an intimate mixture of a dry nitrate is heated with an excess of silica, nitrogen pentoxide is evolved and the amount is determined by the loss in weight. 2NaNO 3 + SiO 2 = Na 2 SiO 3 + N 2 O 5 . This method cannot be used when there is any other volatile substance present, which is usually the case. 3. Determination of Nitric Acid as Ammonia. The usual method for the determination of nitric acid is to reduce it in alkaline solution to ammonia by means of aluminium, zinc, or, best, Devarda's alloy (cf. Vol. I, page 6) : NaNO 3 + 8H - 2H 2 O + NaOH + NH 3 . After the reduction, the solution is distilled into a known amount of acid and the excess of the acid is found by titration, or the ammonia is determined as ammonium chloroplatinate or as platinum (cf . page 58, b and c) . * Reich. Z. Chem., 1, 86 (1862). 454 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. Procedure of Devarda* About 0.5 gm. of the nitrate is placed in a 600-800-c.c. Erlenmeyer flask (Fig. 78) and dissolved in 110 c.c. of water. To this solution 5 c.c. of alcohol, 50 c.c. of caustic potash (sp. gr. 1.3), and 2 to 2J- gms. of powdered Devarda's alloy are added. After K B A FIG. 7! this the flask is immediately connected with the distilla- tion apparatus as shown in the figure. The Peligot tube, A, of about 250 c.c. capacity, is constructed as proposed by F. Pan- nertz.f Its left arm is connected by a curved tube with the mid- dle bulb, so that a spurting back of the liquid is avoided. The delivery-tube (of potash glass) connecting the flask K with the tube A is about 1 cm. in diameter and is provided with a small opening at o, inside the flask, to prevent any of the alkaline solution being carried over with the ammonia. Twenty cubic centimeters *Zeit. f. anal. Chem., XXXIII (1894), p. 113. t Ibid., XXXIX (1900), p. 318. DETERMINATION OF NITRIC ACID AS AMMONIA. 455 of half -normal sulphuric acid are added to the tube A * and diluted so that the solution just reaches to each of the bulbs 0:1 the side, while 5 c.c. of the acid are placed in B, with a few drops of methyl orange, and diluted in the same way. The tubes A and B are connected by means of a T tube, of which the upper end is closed by a pinch-cock upon a piece of rubber tubing, so that a piece of red litmus paper may be introduced here. When all is ready, the contents of the flask K are gently heated in order to start the reaction, then the flame is removed and the reaction allowed to proceed by itself. After an hour this will be shown to be complete by the cessation of the hydrogen evolution. The liquid in K is then slowly heated to boiling, and kept at this temperature until about half of the liquid has distilled over into A ; this requires about half an hour. During the last ten minutes a slow current of air is passed through the tube r. If the distillation has been correctly performed, all of the ammonia will now be found in A ; no trace should reach B, and the red litmus paper in the T-tube should show no tinge of blue. When the distillation is finished, the pinch-cock at r is opened and the flame removed. A little methyl orange is added to A where- by the liquid is colored red, the contents of B are poured in, the latter tube is washed with water that is added to A, and the excess of the sulphuric acid is titrated with half-normal caustic potash solution until the solution is changed to yellow. The amount of nitric acid is computed as follows: The tubes originally contained 25 c.c. of half-normal acid, and t c.c. of half -normal caustic potash solution were used up in the titration; consequently the ammonia formed from 0.5 gm. of the nitrate was neutralized by 25 t c.c. of half-normal sul- phuric acid. Since 1 mol.f of HNO 3 (63.02 gms.) on being reduced yields 1 mol. of NH 3 , and one liter of half-normal sulphuric acid con- tains enough sulphuric acid to neutralize ^ mol. of NH 3 , it is evi- /^o cyy dent that 1 c.c. of the acid is equivalent to - - = 0.03151 gm. * The tube A has a capacity of about 250 cubic centimeters. t Ostwald has proposed that the molecular weight in grams of a sub- stance be designated by the word "mol." 45 6 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. of nitric acid so that 25- 1 c.c. = (25-0X0.03151 gm. of HNO 3 or (25 0-03101 gm. NO 3 , and the per cent, of NO 3 present is 0.5;(25-OvQ.03 101= 100:3 s = 6.202-(25-0=per cent. NO 3 . Determination of Nitric Acid as Nitric Oxide. Method of Schlosing and Grandeatl, modified by Tiemann and Schulze.* Principle. If a nitrate is heated with ferrous chloride and hydrochloric acid, the nitric acid is reduced to nitric oxide : NaNO 3 + 3FeCl 2 + 4HC1 = NaCl+ 3FeCl 3 + 2H 2 O-f NO. From the volume of the nitric oxide its weight is calculated. The method of Schlosing in its original formf was not much used on account of the apparatus required ; but after being modi- fied by Grandeau J it has become one of the best methods for the determination of nitric acid. The apparatus necessary is shown in Fig. 79 and consists of a 150-c c. flask K fitted with a double-bored rubber stopper. Through one of the holes is passed the tube b, which reaches into the flask just to the lower surface of the stopper; through the other hole passes the tube a, ending in a restriction about 1 mm. wide and reaching 1J cm. below the stopper. The tube b is con- nected by means of a piece of rubber tubing 5 cm. long, which is wired on to the tube, and is provided with a pinch-cock, with a second tube whose lower end reaches up into the measuring-tube and is covered with rubber tubing as is shown in the figure. In the same way the tube a is connected with a straight tube. Solutions required, 1. A nitrate solution of known strength, prepared by dissolving in one liter of water 2.0222 gms. of recrystal- lized potassium nitrate that has been dried at 160 C. Fifty c.c. * Zeit. f. anal. Chem., IX (1870), p. 401, and Berichte, VI (1873), p. 1041. f Annales de chim. et de phys., [3], 40 (1853), 479. J Grandeau, Analyse chimique applique & I'agriculture. Grandeau used a separatory funnel instead of the tube a; the lattel was proposed by Tiemann and Schulze. DETERMINATION OF NITRIC ACID AS NITRIC OXIDE. 457 of this solution evolve at C. and 760 mm. pressure 22.41 c.c. of NO. 2. A ferrous chloride solution obtained by dissolving 20 gm. of iron (nails) in 100 c.c. of concentrated hydrochloric acid. 3. Hydrochloric acid, of specific gravity 1.1. Procedure. First of all, 10 c.c. of water are poured into K and its upper level is marked on the outside of the flask by means of FIG. 79. a colored pencil, then 40 c.c. more are added and its position is also marked. The water is now poured out and exactly 50 c.c. of the standard nitrate solution is added to K, the stopper fitted with the two tubes is placed in the flask, and the pinch-cocks h f and h" are opened. The contents of the flask are heated to boiling with a free flame (a wire gauze is not used) until finally no more bubbles of air escape from the lower end of b into the bath containing boiled water. To make sure that the air is all expelled from the apparatus, the rubber tubing at h f is pinched with the thumb and finger, when, if no air is present, the liquid will quickly rise in 6, exerting: a noticeable pressure. The pinch-cock hf is then closed and the boilirg is continued until the 50 c.c. has been re- 458 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. duced to a volume of 10 j.c.,when the flame is removed and the pinch-cock A" is immediately closed. The lower end of a, which dips into distilled water, is immediately filled with the latter up to the pinch-cock. The vapors in the flask condense, forming a vacuum, as shown by the closing together of the rubber tubing at h' and Ji" . 30 c.c. of the ferrous chloride solution are poured into a beaker and the upper level is marked on the outside with a colored pencil, then 20 c.c. more are added and the position in the beaker is again marked. The lower end of the tube a is placed in the ferrous chloride solution so that it reaches below the lower mark on the beaker, and, by opening h", 20 c.c. of the solution are allowed to pass into the flask K. The beaker containing the ferrous chloride is then replaced by one containing boiled water. The tube a should not extend vertically into the water, but should be inclined as much as possible. The specifically heavier ferrous chloride solution in the tube passes into the water, while the latter takes its place. When the lower end of a has become filled with pure water in this way, it is dipped into a beaker containing hydrochloric acid (sp. gr. 1.1) and about 20 c.c. of the acid are allowed to flow into K, and finally 3-4 c.c. of water are added to replace the acid in a. A 50-c.c. measuring-cylinder is now filled with boiled water, placed over the lower end of b as shown in the figure, and the con- tents of the flask K are heated fifteen minutes on the water-bath,* then heated to boiling with a free flame. As soon as the com- pressed rubber tubing begins to expand h' is opened, but the rubber tubing is at the same time pinched between the thumb and finger. As soon as the liquid no longer rises in b, the hand is removed from the rubber tubing and the nitric oxide begins to collect slowly in the measuring-tube. After half * of the liquid has evaporated there is no further evolution of nitric oxide to be noticed, although the brown color of the solution shows that the gas has not been completely expelled. In order to accomplish this, the flame is re- moved, h' is closed, and the liquid in K allowed to cool. By means of the vacuum thus produced the remainder of the nitric oxide is expelled from the solution and the boiling is once more repeated, with the same precautions as before, until the lower mark is * The heating on the water-bath is necessary, as otherwise a little nitric will distil over and not be reduced. A. Wegelin, Inaug. Dissert. Zurich, 1907. DETERMINATION OF NITRIC ACID AS NITRIC OXIDE. 459 reached. The flame is removed, h f is closed, and the measuring- tube containing the nitric oxide is placed in a cylinder containing pure water at the temperature of the room. To prevent the tube containing the gas from sinking, its upper end is encased in a large cork so that it floats on the water. After standing fifteen to twenty minutes the tube is raised by means of the cork until the level of the liquid within stands at the same height as that in the cylinder without, and the volume of the gas is read. At the same time the temperature of the water is taken and the barom- eter reading is noted. The volume thus obtained is reduced to C. and 760 mm. pressure. If the temperature was t, the barometer reading B millimeters, and w the tension of aqueous vapor at t, then the reduced volume is _V(B-w}273 * 760(273+0" Now 50 c.c. of the standard potassium nitrate solution con- tain 0.1011 gm. of KNO 3 corresponding to 0.06201 gm. of NO 3> so that the volume V of the nitric oxide corresponds to 0.06201 gm. NO 3 . The same procedure is now followed with 50 c.c. of the solu- tion of the unknown nitrate, which should be prepared so that the amount of nitric oxide evolved will be about the same as that from 50 c.c. of the standard solution. If at t' C. and B L mm. pressure the volume V of nitric oxide is obtained, and w l is the tension of aqueous vapor at t, then the reduced volume of the nitrogen will be as before 760(273+/ /0 ) ' The following proportion now holds: 7 : 0.0620 l=r o :x T^'0.06201 , r = o;m. NO 3 in 50 c.c. of solution. * Three or four experiments are performed with the standard solution, and the mean value is used. 460 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. Remark. It is not permissible to compute directly the weight of NO 3 which corresponds to the volume of nitric oxide obtained, for some nitric oxide always remains in the flask, so that low values would result. This error is eliminated by the above pro- cedure. L. L. de Koninck * has devised an apparatus which prevents the liquid from sucking back into the decomposition-flask and at the same time permits the earring out of a number of deter- minations one after the other without cleaning the apparatus or boiling it free from air in the meantime. Determination of Nitric Acid in a Drinking-water. From 100 to 300 c.c. of the water are evaporated to 40-50 c.c. in a porcelain dish, a few drops of methyl orange are added, fol- lowed by dilute hydrochloric acid, free from nitrate, until the solu- tion is pink in color. Sodium carbonate solution is now added until the liquid is barely alkaline (it becomes yellow) and the con- tents of the flask are washed into the decomposition-flask K, Fig. 79, and analyzed as desi-ribed on page 457 with the difference that, instead of collecting the gas over water, a 10 per cent, solution of sodium hydroxide is used, to make sure that the carbonic acid which is set free is completely absorbed. After the experiment has been performed with the water to be analyzed, it is repeated with an amount of the standard solu- tion sufficient to evolve about the same quantity of nitric oxide. The analysis is then computed as before. Remark. In drinking-water the neutralization of the evap- orated sample is not absolutely necessary, except in the case of alkaline mineral waters ; in that case the introduction of the hydro- chloric acid would otherwise cause such a violent evolution of carbon dioxide that the flask might crack. CHLORIC ACID, HC10 3 . Mol. Wt. 84.47. Forms: Silver Chloride, AgCl, besides volumetric and gasometric methods. In order to determine chloric acid as silver chloride it must previously be reduced to chloride by means of ferrous sulphate or zinc. * Z. anal. Chem., 33, 300 (1894). See also Liechti and Hitter, ibid., 42 205, (1903). CHLORIC ACID. 461 Reduction by means of Ferrous Sulphate. About 0.3 gm. of the salt is dissolved in 100 c.c. of water, treated with 50 c.c. of a 10 per cent, solution of crystallized ferrous sulphate, heated with constant stirring till it begins to boil, and kept at this temperature for fifteen minutes. After cooling, nitric acid is added until the deposited basic ferric salt is dis- solved, when the chloride is precipitated by means of silver nitrate and weighed after the usual treatment. One gram of silver chloride corresponds to 0.8550 gin. KC1O 3 . Reduction with Zinc. Although chlorates are reduced in neutral solution by means of zinc or Devarda's alloy, it is not advisable to effect the reduc- tion in this way for quantitative purposes. The same end is reached more expeditiously by adding zinc-dust to the acetic acid solution. The dilute chlorate solution is treated with acetic acid until it reacts distinctly acid, an excess of powdered zinc is added, and the solution boiled for one hour. After cooling, nitric acid is added in sufficient quantity to dissolve all of the excess of zinc, after which the solution is filtered if necessary and the chloride precipitated and determined as silver chloride. Remark. Both methods afford exact results, but the former is to be preferred, for it is accomplished in less time. Chlorates are not quantitatively decomposed into chlorides by ignition in open vessels or in a current of carbon dioxide. Some chlorine and a little alkali is always lost, so that even when the residue is evaporated with kydrochloric acid, too low results are obtained. L. Blangey, working in the author's laboratory, ob- tained results which were from 0.3 to 1.1 per cent, below the theo- retical value. According to the two following methods, the decomposition of alkali chlorate into chloride is quantitative. (a) By Evaporation with Hydrochloric Add. The chlorate contained in a weighed porcelain crucible is covered with hydrochloric acid (1:3). A watch-glass is placed upon the crucible, and the contents of the latter are heated on the water- 462 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. bath until the evolution of chlorine ceases. The liquid on the lower surface of the watch-glass is then washed into the crucible, and its contents are evaporated to dryness on the water-bath. The cover is placed upon it and it is then gently ignited over a free flame until the decrepitation ceases. After cooling in a desic- cator, the crucible is again weighed. In this way L. Blangey obtained, as a mean of four experiments, 100.02 per cent, of the theoretical value. (b) By Ignition with Ammonium Chloride. The alkali chlorate is mixed in a porcelain crucible with three times as much pure ammonium chloride, covered with a watch- glass, and heated over a free flame, kept in constant motion, until the ammonium chloride is completely removed. The crucible is then weighed. As a mean of two experiments, L. Blangey obtained 100.06 per cent, of the theoretical value. PERCHLORIC ACID, HC1O 4 . Mol. Wt. 100.47. Form: Silver Chloride, AgCl. Perchlorates cannot be reduced to chloride by means of ferrous sulphate, zinc, or by repeated evaporation with concentrated hydrochloric acid.* On ignition, some chlorine and alkali chloride are lost, so that an error amounting to as much as 1 per cent, may be expected. On the other hand, Winteler has shown that per- chlorates may be changed to chlorides by heating with concen- trated nitric acid and silver nitrate in a closed tube (see Carius' method for determining chlorine in organic substances, page 325), while L. Blangey found that ignition with ammonium chloride would accomplish the same result. Decomposition of Perchlorates by Ignition with Ammonium Chloride. By twice igniting an intimate mixture of 0.5 gm. potassium perchlorate with 1 \ to 2 gms. of ammonium chloride f in a platinum * On evaporating with hydrochloric acid there is a loss without any evo- lution of chlorine; it must be due to the volatilization of small amounts of perchloric acid. j- When 2 gms. of NH 4 C1 are used, one and one-half to two hours are nec- essary. PERCHLORIC, CHLORIC, AND HYDROCHLORIC ACIDS. 463 crucible covered with a watch-glass, the former is completely changed to chloride. Care should be taken not to melt the residual chloride, for in that case the platinum is attacked, although the accuracy of the results is not affected. Blangey obtained in two experiments 100.06 and 100.08 per cent, of the theoretical values. It is worth mentioning that complete decomposition could not be effected by igniting three times in a porcelain* crucible; the platinum evidently plays the part of a catalyser, as was proved by the following experiment: 0.4767 gm. of KC1O 4 was mixed in a porcelain crucible with 1J gm. of NH 4 C1, and 1 c.c. of hydro- chlorplatinic acid (containing 0.0918 gm. Pt) was added. After evaporating to dryness on the water-bath, the ammonium chloride was completely expelled and the residue was ignited twice more with the same amount of the latter. The residue of potassium chloride then weighed 0.2572 gm., corresponding to 100.24 per cent, of the theoretical amount, f Determination of Perchloric Together with Chloric Acid. In one portion the chlorate is reduced, as described on page 461, with ferrous sulphate, and the chloride formed determined as silver chloride. A second portion is ignited in an old platinum crucible (or in one of porcelain) with the addition of 1 c.c. of hydro- chlorplatinic acid and three times as much ammonium chloride (as described above). In this way the total amount of chlorine is obtained and from these data the amount of each acid can be calculated. Determination of Perchloric, Chloric, and Hydrochloric Acids in the Presence of One Another. The three acids are assumed to be present, in the form of their alkali salts. *Thus on 'rrMting 0.4395 gm. KC1O 4 with 2 gms. NH 4 C1 a residue of 0.3205 gm. was obtained instead of one weighing 0.2365 gm. f There is often a slight deposit of alkali chloride upon the cover-glass. To determine this, the glass together with the deposit is weighed, then the glass 13 washed, cried, and again weighed : the difference between the two weights represents the amount of alkali chloride. This rarely amounts to more than a fraction of a milligram, and if the ignition was performed with care, there will be no deposit at all upon the glass. 464 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. In one portion the chloride-chlorine is determined by precipita- tion with silver nitrate. In a second sample the chlorate and chloride-chlorine are determined after the former has been reduced bo chloride by means of ferrous sulphate. The total amount of chlorine present is determined in a third portion after ignition with ammonium chloride. GROUP VI. SULPHURIC, HYDROFLUORIC, AND HYDROFLUOSILICIC ACIDS. SULPHURIC ACID, H 2 SO 4 . Mol. Wt. 98.09. Form: Barium Sulphate, BaS0 4 . Theoretically the gravimetric determination of sulphuric acid is extremely simple, it being only necessary to precipitate with barium chloride, filter and weigh the barium sulphate. Prac- tically, however, it is a process connected with many difficulties. According to the manner of precipitating barium sulphate, the composition of the precipitate varies in such a way that some- times the results are too high and sometimes too low. Errors which may Occur in the Precipitation of Barium Sulphate.* /. In the Precipitation of Barium Chloride with Pure Sulphuric Acid. If a dilute, slightly acid solution of barium chloride is treated at the boiling temperature with an excess of dilute sulphuric acid, the precipitate contains all of the barium except a very small, negligible amount. If, however, the precipitate is weighed, the result is invariably too low; and this is true even when the solution is evaporated to dryness in order to recover the last traces of barium. The precipitate always contains barium chloride in a form which cannot be removed by washing. A mixture, there- fore, of barium sulphate and barium chloride is weighed, and as the molecular weight of the latter is less than that of the former, the result must be too low. In order to obtain accurate results the chlorine combined with barium in the precipitate must be replaced * See the interesting article by M. J. van't Kruys on the determination of sulphuric acid in the presence of various salts which affect the result. Z. anal, Chem., 1910 ,393. SULPHURIC ACID. 465 by 804; and this is easily accomplished by moistening the precip- itate with concentrated sulphuric acid, and heating until the excess of the latter is removed by volatilization. Not only is barium chloride carried down with barium sulphate, but all barium salts as well, especially the chlorate and nitrate. These are, however, readily changed to sulphate by the above treatment with concentrated sulphuric acid. It is immaterial in the estimation of barium how the precipitation is effected; whether the sulphuric acid is added quickly, or drop by drop, the results are always the same. II. In the Precipitation of Pure Sulphuric acid with Barium Chloride. This is the reverse process, but in this case it is not a matter of indifference whether the barium chloride is added slowly, drop by drop, or rapidly all at one time. In the first instance, the results are very near the truth without applying any correction whatsoever; in the latter instance, too high results are obtained, because by the rapid addition of the reagent much more barium chloride is carried down with the precipitate than when the reagent is added very slowly. In order to obtain the true weight of barium sulphate, it is necessary to make a deduction for the amount of barium chloride contained in the precipitate and to add the weight of barium sulphate remaining in solution. The chlorine contained in the precipitate can be determined in several different ways. 1. The precipitate is fused with four times as much pure sodium carbonate, the melt extracted with hot water, the solution filtered, acidified with nitric acid, and the chlorine precipitated with silver nitrate which is filtered off and weighed. (Cf. p. 320.) 2. Still more accurate is the process of Hulett and Duschack * The greater part of the ignited precipitate of barium sulphate is placed in a U-tube of which one arm is drawn out into a thin, right-angled, gas delivery tube. Concentrated sulphuric acid is added to the precipitate and the mixture is heated by placing the U-tube in hot water. The barium sulphate dissolves readily in * Z. anorg. Chem., 40, 196 (1904). 466 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. the hot concentrated sulphuric acid and the barium chloride present is decomposed. In order to determine the amount of hydrochloric acid set free, a slow stream of air, which has pre- viously passed through caustic potash solution, is led through the tube, which has the drawn-out end of the Matter dipping into a stout test-tube containing 0.01 N. silver nitrate solution. After two or two and one-half hours all of the hydrochloric acid will have been expelled from the sulphuric a*cid. The decomposition apparatus is then removed, the gas delivery tube washed out with a little water, and the silver remaining in solution is determined volumetrically (cf . pp. 702-05) . For the determination of the dissolved barium sulphate the filtrate from the first precipitation is evaporated to dry ness,* the residue moistened with a few drops of concentrated hydro- chloric acid, taken up with water and the slight precipitate of barium sulphate filtered off and weighed. Calculation of the true weight of Barium Sulphate. If the weight of the first precipitate of crude barium sulphate is a, the weight of the barium chloride contained in this precipitate, as deter- mined by titration of the amount of chlorine, is b } and the amount of barium sulphate in solution is c, then a b + c represents the weight of pure barium sulphate. Experience has shown, however, that when pure sulphuric acid is precipitated by means of dilute barium chloride solution added drop by drop, the errors b and c are approximately equal and counterbalance one another so that the weight a is very close to that of the pure barium sulphate. ///. In the Precipitation of Sulphates with Barium Chloride. Here the relations are far more complicated than in the pre- cipitation of pure sulphuric acid, partly because the barium sulphate is much more soluble in salt solutions than in water containing a little acid, and partly because of the tendency of barium sulphate to occlude not only barium chloride but many other salts as well. Solutions of chromium sulphate are either * During all such work care should be taken to prevent sulphuric acid contamination from the air in the laboratory. The evaporation should therefore take place on the steam bath or steam table. SULPHURIC ACID. 467 violet or green. From the boiling-hot green solution only one- third of the sulphuric acid is precipitated, the remainder probably being present in the form of a complex chromium sulphate cation; * on cooling there is a tendency for the green solution to become violet and after some time all of the sulphuric acid is precipitated. The precipitation of barium sulphate in the presence of ferric iron has been much studied. In the boiling hot solution all of the sulphuric acid is not precipitated and considerable iron is thrown down with the barium sulphate and furthermore the precipitate then loses S0 3 on ignition. Since ferric oxide weighs much less than an equivalent weight of barium sulphate sometimes the results are as much as 10 per cent, too low. On the other hand, Kiister and Thiel,t were able to get satisfactory results (1) by precipitating the sulphuric acid from such a solution in the cold, or (2) by slowly adding the ferric chloride and sulphuric acid solution to the hot solution of barium chloride, or (3) by precipitating the iron by an excess of ammonia, heating, and adding barium chloride to the solution without filtering off the ferric hydroxide, and finally dissolving the latter in dilute hydrochloric acid. Most chemists, however, deem it advisable to remove trivalent metals before attempting to determine the sulphuric acid. This is accomplished in the case of ferric iron by adding a liberal excess of ammonia to the dilute slightly acid solution which is at a temperature of about 70. If from 5-7 c.c. of concentrated ammonia (sp. gr. 0.90) is added in excess of the amount required for neutralization,! the precipitate is not likely to contain any basic ferric sulphate. If, on the other hand, the solution is barely neutralized with ammonia, the precipitate produced will invariably contain some sulphate. The bivalent metals are occluded to a much less extent, so that it is not, as a rule, necessary to remove them. On the other hand, in the presence of considerable amounts of bivalent metal with relatively small amounts of sulphuric acid, the error arising from occlusion is likely to be large, so that it is better to remove the bivalent metals in all such cases. * Recoura, Comptes rendus, 113, 857; 114, 477. f Z. anorg. Chem., 22, 424. j Pattinson, J. Soc. Chem. Ind., 24, 7. 468 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. ' In the presence of nitric or chloric acid the barium sulphate precipitate will contain considerable quantities of barium chlorate and nitrate which it is impossible to remove by washing with hot water. These acids, therefore, must be decomposed by evapora- tion with hydrochloric acid before attempting to precipitate the sulphuric acid. In ordinary chemical practice it is usually a question of deter- mining sulphuric acid in a solution containing considerable amounts of ammonium or alkali chloride, ammonium or alkali sulphate, and some free hydrochloric acid. Now ammonium and alkali sulphates are also occluded by barium sulphate, and the amount of occulsion increases as the solution is more concentrated with respect to these substances. For this reason it is evident that barium sulphate should always be precipitated in a very dilute solution. Furthermore, a small amount of free hydrochloric acid is indispensable, but larger amounts have a solvent effect upon the precipitate. For an amount of sulphuric acid corresponding to between 1 and 2 gms. of barium sulphate, the precipitation should take place in a volume of between 350 and 400 c.c. and in the presence of hydrochloric acid amounting to 1 c.c. of sp. gr. 1.17. If a neutral solution is at hand, it is diluted to a volume of 350 c.c. and 1 c.c. of concentrated hydrochloric acid is added. An alkaline solution is carefully neutralized with hydrochloric acid, using methyl orange as indicator, 1 c.c. of concentrated hydrochloric acid is added in excess, and the solution is diluted to 350 c.c. Finally, in the case of an acid solution, it is either evaporated to dryness, the residue moistened with 1 c.c. of concentrated hydrochloric acid and 350 c.c. of water added, or, with methyl orange as indicator, the solution is neutralized with ammonia, treated with 1 c.c. of concentrated hydrochloric acid, and diluted, to 350 c.c. After the solution has been prepared in accordance with the above directions it is ready for the SULPHURIC ACID. 469 Precipitation of Sulphuric Acid in the Presence of Ammonium or Alkali Salts according to E. Hintz and H. Weber. The solution is heated to boiling, and then for each gram of barium sulphate precipitate 10 c.c. of normal barium chloride solution are taken, diluted to 100 c.c., the solution heated to boiling, and added all at one time to the hot sulphate solution which is being stirred continuously. After the solution has stood for half an hour, best in a warm place, it is filtered, washed with hot water and ignited (cf . p. 74) . The use of a Gooch or Munroe crucible is to be recommended. Remarks. In the presence of ammonium salts the pre- cipitation of the barium sulphate should not be effected, as is otherwise desirable, by the cautious addition of the barium chloride, for, as Hintz and Weber have shown,, this leads to low results whereas the occlusion caused by the rapid addition of the barium chloride counterbalances this error. Under no circumstances should a precipitate of barium sulphate be heated over a blast lamp, for in that case sulphuric anhydride would be evolved from the barium sulphate itself. To explain the occlusion of barium chloride by barium sulphate, Hulett and Duschak * have suggested that perhaps the precipitate may contain salts such as BaCl.HSO 4 , (BaCl) 2 SO 4 , and Ba(HSO 4 ) 2 and Folin f believes that such is this case because some of his precipitates have lost SOa on ignition while others have lost HC1. He also suggests the possibility of salts such as Ba(KSO 4 )2 being precipitated. Determination of Sulphuric Acid in Insoluble Sulphates. Calcium and strontium sulphates are decomposed by long digestion with ammonium carbonate solution, but barium sulphate is not. The latter is mixed with four times as much sodium car- bonate, fused in a platinum crucible, the melt extracted with water, and the barium carbonate residue washed with sodium carbonate solution. After acidifying the nitrate with hydro- * Z. Anorg. Chem., 40, 196 (1904). t J. Biol. Chem., 1, 131 (1905). 470 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. chloric acid and boiling off the carbon dioxide, the sulphuric acid is precipitated as usual. Lead sulphate is boiled with sodium carbonate solution; af^er cooling, the solution is saturated with carbon dioxide and filtered. The lead remains behind as carbonate, while the filtrate contains all of the sulphuric acid. For the determination of sulphuric acid in silicates, the finely powdered substance is fused with six times as much sodium car- bonate, the melt is extracted with water, the filtrate acidified with hydrochloric acid and evaporated to dryness in order to dehydrate the silica. The residue is moistened with a little concentrated hydrochloric acid, taken up in hot water, and the silicic acid filtered off; the sulphuric acid is determined in this filtrate. Determination of Sulphuric Acid in the Presence of Soluble Sulphides. The substance is placed in a flask, the air replaced by carbon dioxide, dilute hydrochloric acid is added, and the solution boiled while carbon dioxide is passed through it until all of the sulphide has been expelled. The sulphuric acid is then precipitated from the solution. This determination is used for the analysis of cements. In this case, however, the hydrochloric acid solution will contain much calcium as well as iron and aluminium, so that these metals are precipitated by the addition of ammonia and ammonium -carbonate and the sulphuric acid determined in the filtrate. If it is desired to determine the amount of sulphide-sulphur, ~the substance is covered with bromine water until the color of the bromine is permanent, hydrochloric acid is added, and the solution boiled to expel the excess of the bromine. The iron, alu- minium, and calcium are precipitated by ammonia and ammonium carbonate, and the total sulphur is determined in the filtrate. The difference between the two results represents the amount of sul- phur present as sulphide. For the volumetric determination of sulphuric acid consult Part II. HYDROFLUORIC ACID. 47 l HYDROFLUORIC ACID, HF. Mol. Wt. 20.01. Forms: Calcium Fluoride, CaF 2 ; Silicon Fluoride, SiF 4 , besides volumetric and gasometric methods. i. Determination as Calcium Fluoride. If the solution contains free hydrofluoric acid or an acid fluoride, sodium carbonate is added until the reaction is alkaline and from one-fourth to one-fifth as much more in excess.* To solutions of neu- tral fluorides about 1 c.c. of double-normal sodium carbonate solution is added. The alkaline solution is heated to boiling, precipitated by means of an excess of calcium chloride solution, filtered, and thoroughly vrashed with hot water. The precipitate consisting of the fluoride and carbonate of calcium is dried, as much of it as possible is transferred to a platinum crucible, the ash of the filter is added, and the contents of the crucible are ignited, f After cooling, the mass is covered with an excess of dilute acetic acid, by which the lime is changed to the soluble acetate, while the fluoride is unaf- fected. Aftsr evaporating to dryness on the water-bath, the mass is taken up in water, filtered, washed, and dried. t After transfer- ring as much of the dried precipitate to the crucible as possible, the tilter-paper is burned, its ash added, and after ignition the crucible is again weighed. To confirm the result the substance is treated with a little concentrated sulphuric acid (added cautiously), and after evaporating off the excess of the latter and once more igniting, the contents of the crucible are weighed as calcium sulphate. 1 gm. CaF 2 yields 1.7436 gms. CaSO 4 . Remark. The results are usually a little low on account of the solubility of calcium fluoride; 100 c.c. water dissolves 0.0016 gm., and 100 c.c. 1.5 N. acetic acid dissolves 0.011 gm. CaF 2 at the temperature of the water bath. * By the addition of the excess of sodium carbonate the precipitate of calcium fluoride will contain calcium carbonate, and presence of the latter renders the precipitate easy to filter. A pure precipitate of calcium fluoride is so slimy that the pores of the filter become so clogged that it is almost impossible to complete the filtration. f The ignition makes the CaF 2 denser and hence easier to filter. J The calcium fluoride is not volatilized in an open platinum crucible heated over a Bunsen burner. Heated over the blast, there is appreciable volatilization. 47 2 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. Example: Determination of Fluorine in Calcium Fluoride. As was stated in Vol. I, page 391,* calcium fluoride is not com- pletely decomposed by fusing with sodium carbonate; but if the fluoride is mixed with 2J times as much silicic acid and then fused with 6 times as much sodium-potassium carbonate, the greater part of the silicic acid and all of the fluorine will be changed to sol- uble alkali salts, while the calcium will be left as insoluble calcium carbonate. The mixture must be Cheated gradually, as otherwise the evolution of carbon dioxide may cause the contents of the crucible to run over its sides. The thin liquid fusion soon changes to a thick paste or only sinters somewhat. On raising the tempera- ture, it is almost impossible to further melt this mass, and it is not necessary. In fact too high a temperature is to be avoided on account of the danger of losing some alkali fluoride by volatilization. The reaction is complete when there is no further evolution of car- bon dioxide. After cooling, the melt is treated with water, the insoluble residue is filtered off and thoroughly washed. The alkaline solution containing all the fluorine and considerable silicic acid is freed from the latter by the addition of considerable ammonium carbonate f (about 4 gms. of the solid salt) . The liquid is heated for some time at about 40 D., allowed to stand over- night, and in the morning the voluminous precipitate is filtered off and washed with ammonium carbonate water. The filtrate now contains only a small amount of silicic acid. It is evaporated almost to dryness on the water-bath, J diluted with a little wa^er and a few drops of phenolphthalein are added. The liquid is colored pink by the indicator and enough hydrochloric acid is now added to make it colorless. The solution is heated to boiling, and this causes the reappearance of the pink color. After cooling the color is again discharged with hydrochloric acid, and this operation is repeated until finally the addition of 1-1J c.c. of double- normal hydrochloric acid is sufficient to effect the decolorization. It is best to perform the operation in a platinum dish, but if this is lacking one of porcelain may be used. * Second edition. t Before adding the ammonium carbonate, the greater part of the alkali carbonate should be neutralized with dilute hydrochloric acid, but care should be taken not to make the solution acid. J The liquid foams during the evaporation owing to the decomposition of the excess of ammonium carbonate; the evaporating-dish is covered with a watch-glass until the evolution of carbon dioxide ceases. HYDROFLUORIC AUD. 473 The solution still contains traces of silicic acid, which are re- moved, as recommended by Berzelius, as follows : The solution is treated with 1 or 2 c.c. of ammoniacal zinc oxide solution,* boiled until the ammonia is completely expelled and the precipi- tate of zinc silicate and oxide is filtered and washed with water. An excess of calcium chloride is added to the filtrate and the re- sulting precipitate, consisting of calcium carbonate and fluoride, is treated as described on page 471. The calcium fluoride finally obtained should be tested for fluorine, for the addition of calcium chloride will almost always cause a precipitation (cf. page 471). which may consist of calcium fluoride and phosphate, or the latter only. After weighing the precipitate, it is treated with a few drops of concentrated sulphuric acid and covered with a watch-glass whose convex surface is coated with a thin layer of beeswax with a few lines scratched in the latter. The crucible is allowed to stand this way for twelve hours at the ordinary temperature. A little water is then poured upon the watch-glass and the crucible is heated over a tiny flame until the vapors of sulphuric acid begin to be evolved. If fluorine is present there will be a distinct etching of the glass where the wax coating was removed. The weight of the calcium fluoride obtained should stand in the same relation to that of calcium sulphate obtained after treatment with concentrated sulphuric acid, as CaF 2 (78.09) :CaSO 4 (136.16). This relation does not hold exactly in practice, for it is almost impossible to obtain a precipitate of calcium fluoride absolutely free from silica. Remark. By this method the fluorine present in all fluorides can be determined, e.g., in topaz, lepidolite, cryolite, etc. With a silicate containing much silica, the addition of silicic acid is * Moist zinc oxide is dissolved in ammonia water. The oxide is best preparel by dissolving chemically pure zinc in hydrochloric acid, and pre- cipitating the zinc with potassium hydroxide ; the precipitate is filtered and washed. 474 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. unnecessary, and the substance is at once fused with 4-5 times as much sodium-potassium carbonate; with silicates containing little silica, from ^-1 part of silicic acid is added. If the substance contains phosphoric acid, the fluorine cannot be determined by the above method, because the calcium fluoride precipitate is then contaminated with calcium phosphate. It is then necessary to effect a Separation of Phosphoric and Hydrofluoric Acids. The following method is that of Rose as modified by the author and A. A. Koch.* It is based upon the fact that silver phosphate is insoluble in water whereas silver fluoride is soluble. Procedure. The alkaline solution of the two acids f is carefully neutralized with nitric acid and then transferred to a 250 c.c. calibrated flask. A slight excess of silver nitrate is added, the solution diluted to the mark, thoroughly mixed and the precipitate allowed to settle completely. The solution is then filtered through a dry filter, but the first 10 c.c. of the filtrate are rejected, and the rest allowed to run into a dry beaker. Of this filtrate, exactly 200 c.c. are transferred to a 250-c.c. flask again, the excess of the silver nitrate precipitated by the addition of some dissolved sodium chloride, the solution made up to the mark, well shaken and the precipitate allowed to settle completely. This solution is filtered, using the same precautions as before and the fluorine is determined in 200 c.c. of the filtrate as calcium fluoride according to the directions on p. 471. If the weight of the calcium fluoride precipitate is p, that of the original substance is a, and x is the per cent, of fluorine origi- nally present, then 76.03 p_ (Tf - /o r , * Z. anal. Chem., 43, 469 (1904). t It is usually a matter of analyzing an aqueous solution of a sodium carbonate fusion. DETERMINATION OF FLUORINE AS SILICON FLUORIDE. 475 Determination as Silicon Fluoride. This method, proposed by Fresenius, depends upon the fact that many fluorides are decomposed by the action of concentrated sulphuric acid and silica, while the fluorine escapes as silicon fluoride, which can be absorbed and weighed. Procedure. The same reagents and a very similar apparatus to that described on p. 477 is required for this determination, except that in place of the Peligot tubes (Fig. 80) two weighed U-tubes are used,* of which the first is filled with moistened pieces ofjpumice, and the second has one arm filled with soda lime and the other with calcium chloride. The analysis is carried out in exactly the same way as is described for the Penfield method (see below) but at the end of the experiment the two U-tubes are weighed. The increase in weight represents the amount of SiF 4 , and from this the amount of fluorine present is calculated as follows: Assume that a gms. of calcium fluoride yielded p gms. of SiF 4 . The treatment with the concentrated sulphuric acid caused the following reaction to take place: 2CaF 2 + 2H 2 S0 4 + SiO 2 * 2CaSO 4 + 2H 2 O+ SiF 4 , consequently the following proportion holds: SiF 4 :4F = p:s 4F s = p = gms. iluonne and in per cent. 400F p_ SiF 4 'a or = 72.87- = per cent, fluorine. * It is best to use U-tubes for the absorption which are provided with ground-glass stoppers (see Fig. 59, p. 381). 470' GRAVIMETRIC DETERMINATION OF THE METALLOIDS. Remark. This method is suitable for the determination of fluorine in all fluorides which are decomposed by sulphuric acid. The analysis can be carried out in the presence of phosphates, but if carbonates are present they should be decomposed by ignition before the treatment with sulphuric acid. According to K. Daniel,* exact results are obtained only when the decomposition of the fluoride takes place at the temperature at which sulphuric acid boils. For the determination of fluorine in topaz and micas, the method is not suitable. Determination of Fluorine as Hydrofluosilicic Acid, according to S. L. Penfield.t Modified by Treadwell and Koch.% Principle. Penfield expels the fluorine as silicon fluoride in exactly the same way as in the method of Fresenius (page 475), but the gas is absorbed in 50 per cent, alcoholic potassium chloride solution. By contact with water the silicon fluoride is decomposed into hydrofluosilicic and silicic acids. The former unites with the potassium chloride, forming potassium silicofluoride, insoluble in 50 per cent, alcohol: H 2 SiF 6 + 2KC1 = K 2 SiF e + 2HC1, and sets free an equivalent amount of hydrochloric acid; the latter is titrated with N/5 sodium hydroxide solution, with cochineal is an indicator. For the calculation the following pro- portion holds: 1000 c.c. N/5 HC1= j=V mol. CaF 2 =fF. .'. 1 c.c. N/5 NaOH = 0.0234 gm. CaF 2 or 0.0114 gm. F. *Z. anorg. Chem., 38, 257 (1904). fChem. News, 39, p. 179; also Am. Chem. Jour., 1, p. 27 ^. anal. Chem., 43, 469 (1904). DETERMINATION OF FLUORINE AS HYDROFLUOSIL1C1C ACID. 477 Requirements. 1. Pure Quartz Powder. Pieces of pure rock crystal are placed in a platinum crucible, heated strongly over the blast lamp, and then thrown into cold water. After this treatment it is very easy to reduce the quartz to a fine powder by grinding in an agate mortar. The powder is ignited, and while still warm is transferred to a flask, fitted with ground-glass stopper. The open flask and its contents are allowed to cool in a desiccator, after which the flask is stoppered and set aside. 2. Sea Sand. The purest sea sand is treated with boiling, concentrated sulphuric acid, washed, dried, ignited, and cooled in a desiccator. 3. Anhydrous Sulphuric Acid. Chemically pure concentrated sulphuric acid is heated in a porcelain crucible until it has fumed strongly for some time and then allowed to cool in a desiccator over phosphorus pentoxide. Procedure. The weighed sample of the fluoride is intimately mixed in an agate mortar, which is placed upon black glazed paper, with 1.5-2 gms. of the quartz powder and then transferred through the cylindrical arm A of the perfectly dry decomposition apparatus to the pear-shaped compartment B shown in Fig. 80. Then 1.5-2 gms. of the sea sand are added, and mixed with the rest of the material by shaking the apparatus, which is then connected with the dry U-tube containing glass beads. The two Peligot tubes P and PI each contain between 10 and 15 c.c. of alcohol which is saturated with potassium chloride. When the apparatus is all connected as shown in the drawing, a dry current of air,f free from carbon dioxide, is allowed to enter at h, and pass through the apparatus at the rate of 2 or 3 bubbles per second. Then without stopping the air current, about 20 c.c. of anhydrous sul- phuric acid is allowed to enter the decomposition apparatus through the funnel T. By introducing the sulphuric acid in this way, while maintaining the air current, the sulphuric acid and the * This tube serves to keep back any sulphuric acid that is carried over mechanically. t The air is passed through caustic potash solution, then over calcium chloride, and finally through concentrated sulphuric acid before entering the apparatus. 478 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. greater part of the silica and fluoride mixture is made to pass directly into the compartment B. After adding the sulphuric acid, the decomposition vessel is placed in a paraffin bath which is slowly heated to a temperature, of 130 to 140 C. The evolution of silicon tetrafluoride at once begins to take place, as is evident from the formation of foam. The passage of the air and heating of the bath is allowed to continue for five hours, at the end of which time the flame under the bath is turned down and air is passed through the apparatus for half an hour longer at the rate FIG. 80. of 3 to 4 bubbles per second. During the heating the apparatus should be frequently shaken in order that the sulphuric acid is brought into contact with all portions of the solid mixture. It is not necessary, however, with this arrangement of the apparatus, to shake as frequently as in the forms of apparatus described by Penfield and by Fresenius, because the air in its passage through the narrow connecting tube between A and B of the decomposition apparatus serves of itself to effect a good mixing. In order to accomplish this end, however, it is necessary to construct the apparatus exactly as shown in Fig. 80; the connecting tube DETERMINATION OF FLUORINE IN HYDROFLUOSILICIC ACID. 479 between A and B must be so narrow that it is completely filled with the bubbles of air passing, and furthermore the parts marked c e b must form an inclined plane upon which the substance can readily pass back and forth. If there is a hollow in the apparatus at c e b, in which some of the substance can collect, the sulphuric acid may not come in contact with some of the fluoride so that the decomposition will be incomplete. Similarly it is necessary to guard against making the connecting tube c e too narrow, as other- wise the air will not pass in a uniform stream, but in spurts, so that in spite of the long tube D some of the sulphuric acid fumes are likely to reach the Peligot tubes and thereby give rise to high results. If not more than 0. 1 gm. of the fluoride was present, the action is over at the end of five and one -half hours, and this is evident, as Daniel * was the first to discover, from the fact that the foaming in the apparatus ceases; the hydrochloric acid which has been set free in the Peligot tubes can now be titrated. To this end a few drops of fresh cochineal f solution are added to each tube and the contents are titrated with fifth-normal potassium hydroxide solution with frequent shaking, until the indicator changes from yellow to red. This is, however, by no means the correct end- point, because as Penfield observed, the gelatinous silicic acid encloses very appreciable amounts of hydrochloric acid. The silicic acid, therefore, must be thoroughly worked over with a stirring rod and the addition of the alkali continued until the color change is permanent. The results obtained by this method, using 0. 1 gm. of substance, appear to be about 0.4 per cent, too high. The method can be used in the presence of phosphoric acid, but carbonates are first removed by ignition before the treatment with sulphuric acid. * Z. anorg. Chem., 88, 257 (1904). f Instead of cochineal, methyl orange may be used, although it is neces- sary then to add an equal volume of alcohol before titrating the hydro- chloric acid. 480 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. Determination of Fluorine in Mineral Waters. From 1 to 10 liters of the water (according to the amount of salts present) are evaporated in a large platinum or porcelain dish to a small volume, with the addition of enough sodium carbonate to keep the solution slightly alkaline. Then an excess of calcium chloride is added, the liquid boiled, and the precipitate filtered and washed with hot water until free from chlorides. The precipitate is dried, transferred as completely as possible to a platinum dish, and the ash of the filter added to the main precipitate which is then gently ignited. This residue contains all of the fluorine as calcium fluoride; besides considerable calcium (possible strontium) and magnesium carbonates; iron, aluminium and manganese oxides ; often barium sulphate ; and almost invariably some calcium phosphate. It is treated with an excess of dilute acetic acid, allowed to stand for some time with frequent stirring, and then evaporated to dryness on the water bath. This residue is treated with water, filtered, and washed with hot water. As much of it as possible is transferred to a platinum crucible, the ash of the filter added, and the contents of the crucible gently ignited. From ^-2 gms. of ignited quartz powder are then intimately mixed with the residue in an agate mortar. The mixture is transferred to the decomposition vessel A, Fig. 80, and treated with concentrated sulphuric acid exactly as described on p. 477 by the method of Penfield. As only very little fluorine is present in this case, two small U-tubes are used instead of the large Peligot tubes shown in Fig. 80. Remark. The formation of a precipitate in the first U-tube at the place marked a a in Fig. 80 indicates the presence of fluorine. It is well to confirm it by the etching-test. After carrying out the titration of the hydrochloric acid set free, the contents of the U-tube are transferred to a platinum dish, a few drops of double normal sodium carbonate added, and the solution evaporated to dryness. Ammoniacal zinc oxide is then added (cf. p. 473), the liquid again removed by evaporation, the residue taken up in water, and the DETERMINATION OF FLUORINE. 481 zinc oxide and silicate filtered off. The filtrate is treated with calcium chloride as described on p. 473 and the etching test applied. In the former editions of this book it was recommended to determine the fluorine in mineral waters by evaporating a very much larger volume of the water nearly to dryness, filtering, and then examining the insoluble residue alone for fluorine. In this way the fluorine content of many mineral waters was entirely overlooked. Gas- Volumetric Determination of Fluorine according to Hempel and Oettel. See Part III, Gas Analysis. Separation of Fluorine. (a) From the Metals. For the determination of the metals present, the fluorine usually can be removed by heating with concentrated sulphuric acid; in the case of many silicates containing fluorine, however, e.g., topaz, lepidolite. and other micas, this treatment will not accom- plish the desired result. In such cases the mineral is fused with 4 to 6 times as much sodium-potassium carbonate, the melt is extracted with water, the silica and aluminium precipitated from the solu- tion obtained by means of ammonium carbonate (see page 472). and these two substances determined in the residue, while the filtrate is used for the fluorine analysis. The metals and the remainder of the silicic acid are determined in the residue obtained on extracting the melt with water (cf. p. 493). The estimation of the alkalies must be undertaken in a separate portion of the substance (pp. 497-501). 482 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. (6) Separation of Fluorine from the Acids. 1. Determination of Hydrochloric and Hydrofluoric Acids in the Presence of One Another. In the case of soluble alkali ealts, the fluorine is first pre- cipitated from the solution by means of a little sodium car- bonate and an excess of calcium nitrate solution, as described on p. 471. The filtrate is acidified with nitric acid and the chlorine determined by precipitation with silver nitrate, according to p. 320. It is simpler to treat the solution containing hydrochloric and hydrofluoric acids in a platinum evaporating dish with nitric acid and silver nitrate. Silver chloride is alone precipitated and can be filtered off, using a funnel of hard rubber, or a glass one coated over with wax. The precipitate is washed and weighed as described on p. 320. When phosphoric acid also is presenc, this is precipitated with the hydrochloric acid by the addition of silver nitrate to the slightly alkaline solution, the precipitate is filtered off, washed with as little cold water as possible, and the precipitate treated with dilute nitric acid. By this means the silver phosphate goes into solution, while the silver chloride is unaffected. In order to determine the amount of phosphoric acid present, the silver is removed from the solution by the addition of hydrochloric acid, and the phosphoric acid is precipitated in the filtrate by addition of magnesia mixture and ammonia (cf. p. 434). In the filtrate from the silver phosphate and silver chloride precipitate, the excess of silver nitrate is removed by the addi- tion of sodium chloride and the fluorine is determined as calcium fluoride. In the case of an insoluble compound containing chlorine and fluorine, the melt obtained after fusing with sodium- potassium carbonate is extracted with water, the silica is removed with ammonium carbonate and zinc-ammonium hydroxide as HYDROFLUOSILICIC ACID. 483 described on pp. ^72-3, and the chlorine and fluorine determined as above. In a majority of cases it is more convenient to determine the two acids in separate portions of the substance. 2. Determination of Boric and Hydrofluoric Acids. The solution containing the alkali salts of these two acids is precipitated at the boiling temperature by means of an excess of calcium chloride ; the precipitate is filtered off and washed with hot water. The precipitate, consisting of calcium carbonate, calcium fluoride, and some calcium borate, is gently ignited, treated with dilute acetic acid, evaporated to dryness, and more acetic acid and water are added. By this means the calcium acetate and cal- cium borate go into solution, while the calcium fluoride is left behind and is determined as described on p. 471. For the boric acid determination a second portion of the solution is taken, made barely acid with acetic acid, and treated with a slight excess of calcium acetate solution in order to precipitate the fluorine. The solution, together with the calcium fluoride, is placed in the Gooch retort and subjected to distillation as described on p. 428. HYDROFLUOSILICIC ACID, H 2 SiF 6 . Mol. Wt. 144.32. Forms: Calcium Fluoride, CaF 2 ; Potassium Silicofluoride; or volumetrically. i. Determination as Calcium Fluoride. Prindple. Alkali fluosilicates are decomposed on heating with sodium carbonate solution into fluoride and silicic acid: Na 2 SiF 6 + 2Na 2 CO 3 + H 2 O = 6NaF + H 2 SiO 3 + 2CO 3 . If a solution is to be analyzed containing free hydrofluosilicic acid or its sodium salt, it is treated with sodium carbonate 484 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. tion until the reaction is alkaline, a considerable amount of ammo- nium carbonate is added, the solution heated to about 40 C., and, after standing twelve hours, the precipitated silicic acid is filtered off. The solution now contains all the fluorine as sodium fluoride, in the presence of small amounts of silicic acid, which are pre- cipitated by the addition of zinc-ammonia hydroxide (see p. 473). In the filtrate the fluorine is determined as calcium fluoride, as described on p. 471. An insoluble fluosilicate is fused with four times as much sodium-potassium carbonate, the melt extracted with water, and the solution subjected to the above treatment. 2. Determination as Potassium Silicofluoride. This analysis is only applicable for the determination of free hydrofluosilicic acid in aqueous solution. Procedure. The solution is treated with potassium chloride and an equal volume of absolute alcohol. The barely-visible potassium silicofluoride is filtered through a tared filter which has been dried at 100 C. After washing with 50 per cent, alcohol the precipitate is dried at 100 C. and weighed as K 2 SiF 6 . The volumetric determination of hydrofluosilicic acid will be discussed in Part II. Analysis of Salts of Hydrofluosilicic Acid. For the determination of the metal present, the salt is treated with concentrated sulphuric acid in a platinum dish and heated until dense fumes of sulphuric anhydride are given off; silicon fluoride and hydrofluoric acid volatilize, while the metals are left behind as sulphates (cf. Vol. I). Determination of Water Present in Fluosilicates, (Rose-Jannasch)* The water cannot be determined by ignition, because all fluo- silicates, even topaz, evolve silicon fluoride when subjected to this treatment (cf. Vol. I, p. 355). If, as proposed by Rose, the * Rose-Finkener: Lehrbuch der analyt. Ch., Bd. II; and Jannasch, Prak- tischer Leitfaden der Gewichtsanalyse, Leipzig, 1897, p. 243. SILICIC ACID. 485 substance is fused with six or eight times as much lead oxide, all the water is evolved, while the fluorine remains behind; R a SiF 6 + 3PbO = 2RF + 2PbF 3 + PbSiO 3 . The analysis is best performed according to the directions of Jannasch: A bulb with a capacity of about 25 c.c. is blown near one end of a tube of difficultly fusible glass which is 26 cm. long and 1 cm. wide. Near the middle of the longer side of the tube is placed, between asbestos plugs, a layer 3 to 5 cm. long of pulver- ized, anhydrous lead oxide, and this end of the tube is connected with two weighed calcium chloride tubes. The substance is placed in the bulb, after which six or eight times as much lead oxide is added and mixed with the substance by carefully revolving the tube. A dry current of air is now conducted through the apparatus and the contents of the bulb are slowly melted. All of the water and often some of the fluorine is thereby expelled, and the latter is absorbed by the layer of lead oxide. At the end of the operation this layer is cautiously heated with a moving flame until no more water condenses in the cooler part of the tube. When all of the water has been driven over into the calcium chloride tubes the latter are weighed with the customary precautions. GROUP VII. SILICIC ACID (ALSO TITANIC, ZIRCONIC, TANTALIC, AND NIOBIC ACIDS). SILICIC ACID, H 2 SiO 3 . Mol. Wt. 78.32. Form: Silicon Dioxide, Si0 2 . Two cases must be considered: (a) The silicate is decomposed by acids. (6) The silicate is not decomposed by acids. (a) Silicates Decomposed by Acids. These are treated with hydrochloric acid in a porcelain dish and evaporated upon the water-bath with frequent stirring until the residue is obtained in the form of a dry powder. In many cases the decomposition is shown to be complete by the fact that 486 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. no gritty particles can be felt with the stirring- rod on the bottom of the dish. If, however, the substance contained quartz or some silicate that is not decomposed by hydrochloric acid, this is not the case and the procedure described on p.. 507 may be followed. The dry powder is moistened with concentrated hydrochloric acid and the covered dish is allowed to stand for 10 or at the most 20 minutes at the ordinary temperature, in order that basic salts and oxides formed during the evaporation and drying may be once more changed to chlorides. Then 100 c.c. of water are added, it is heated to boiling, and after the silicic acid has been allowed to settle, the clear liquid is decanted through a filter sup- ported upon a platinum cone placed in the apex of the funnel. The residue is washed 3 or 4 times with hot water by decantation, then transferred to the filter and washed with hot water until free from chloride.* The precipitate is then dried by means of suction, placed in a platinum crucible, and set aside for the time being. The separation of the silicic acid is now by no means quantitative; as much as 5 per cent, of the total amount may remain in the filtrate. In order to remove this, the solution is once more evap- orated to dryness on the water-bath, kept at this temperature for one or two hours (or more), moistened with a few cubic cen- timeters of concentrated hydrochloric acid, and allowed to stand not more than fifteen minutes. t Hot water is then added, the residue is filtered through a new and correspondingly small filter, and washed with hot water. The amount of silicic acid now remaining in the filtrate amounts to not more than 0.15 per cent, of the total amount, and for most purposes can be neglected. It can be removed, however, by a third evaporation to dryness. The filters containing the silica are ignited wet in a platinum crucible and finally over the blast-lamp, and weighed.J The silica obtained is only slightly hygroscopic. * If the precipitate is not perfectly white, but somewhat brownish owing to the presence of a basic ferric salt, concentrated hydrochloric acid is allowed to run around the upper edge of the filter and is immediately washed down through the funnel by means of a stream of hot water. This is repeated until the nitrate comes through perfectly colorless. t By being kept in contact with the acid for too long a time some silicic acid will go into solution. I With regard to the temperature at which silica is completely dehydrated, SILICIC ACID. 487 Testing the Purity of the Silica. The silica thus obtained is never absolutely pure, except in the analysis of a water-glass. Its purity must always be tested. For this purpose it is covered with 2 or 3 c.c. of water,* a drop of concentrated sulphuric acid is added, and 3 to 5 c.c. of pure hydro- fluoric acid (distilled from a platinum retort). The crucible is then placed in a platinum-lined cone (Fig. 16, p. 31) on the water-bath and evaporated under a good hood until no more vapors are expelled. The excess of sulphuric acid is then removed by heat- ing over a free flame. The temperature is raised and the crucible is finally heated over a blast-lamp, after which it is again weighed. This process is repeated until the contents of the crucible (usually A^Os and Fe 2 O 3 ) are at a constant weight, and this amount is deducted from the weight of impure silica. Remark. In order to make the separation of silicic acid more nearly quantitative it has been proposed to heat the residue obtained by evaporation at 110-120 C.f This hastens the dehydration, but the temperature of 120 should not be exceeded on account of the danger of the silicic acid being contaminated with basic salts, and because of the tendency for there is a difference of opinion. Lunge and Millberg (Zeit. f. angew. Chem., (1897), p. 425) state that the temperature of the Bunsen burner is sufficient, but they operated with silica obtained by the hydrolysis of silicon tetra- chloride, in order to obtain a product absolutely free from alkalies. Hille- brand .(Ami Chem. Soc., XXIV (1902), p. 362) confirmed the results of Lunge and Millberg with regard to the ignition of a silica obtained in this way, but positively asserts that silicic acid when obtained by the decom- position of an alkali silicate with acid must be ignited over the blast-lamp in order to dehydrate it completely. The results of Hillebrand have been confirmed in the author's laboratory by A. Schroter. Jordis and Kanter (Z. anorg. Chem., 35, 20 (1903) ) state that hydrated silica forms a small amount of a chlorine compound with hydrochloric acid. This compound is decomposed on heating only with great difficulty, unless it is treated with water and evaporated to dryness several times. This treatment is recom- mended by Jordis and Kanter in all accurate silica determinations. * If the water is not added, the mass will effervesce so strongly that there is danger of losing some of the impure silica. t James P. Gilbert, Tech. Quarterly, III, p. 61, and Zeit. fur. anal Chem., XXIX (1890), 688. 48 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. it to recombine with these.* It is, therefore, not advisable to attempt to dehydrate the silica at a temperature higher than that of the water-bath. (&) Silicates Not Decomposed by Acids. These must be fused; this can be effected by (a) The Sodium Carbonate Method. One gram of the very finely powdered substance is placed in a spacious platinum crucible together with 4 to 6 parts of calcined sodium carbonate (or a mixture of equal parts sodium and potas- sium carbonates) and fused. The powdered silicate should be intimately mixed with the flux and a little sodium carbonate sprinkled on top, the crucible covered and heated for some time over a small flame in order to drive out any moisture present. The temperature is raised gradually until finally the highest heat of a good Teclu burner is obtained; or, lacking the latter, a blast- lamp should be used. As soon as the mass melts quietly and there is no further evolution of carbon dioxide, the decompo- sition is complete. The crucible is seized with a pair of crucible tongs having platinum points and placed in cold water, but so that the water does not enter the crucible. By means of this rapid cooling the -melt is usually detached from the sides of the crucible and can be removed by simply turning the crucible up- side down and gently tapping its sides, f The melt is received in a good-sized beaker, covered with water, a sufficient quantity of strong hydrochloric acid is added, and the beaker covered with a watch-glass. A lively evolution of carbon dioxide at once takes * When considerable magnesium was present, more silica was found in the filtrate after igniting at 280 than when dried on the water-bath. This is due to the fact that magnesia formed by hydrolysis reunites with the silica to form magnesium silicate, and the latter is decomposed by hydrochloric acid with the formation of soluble silicic acid. t Instead of the cold water a blast of air may be used. Then, when the melt is still warm but not hot enough to cause spattering, the addition of water enough to cover it will usually help to loosen the melt from the sides of the crucible. Hillebrand recommends the rotation of the contents of the crucible while still molten so that a relatively thin layer solidifies on the sides and bottom of t v c crucible. SILICATES NOT DECOMPOSED BY ACIDS. 489 place, but in proportion as silicic acid separates out, the inner part of the cake gradually becomes coated with a film of silicic acid which protects it from the further action of the acid. Con- sequently it is necessary to break up the cake from time to time by means of a glass rod until finally there is no further evolution of a gas and no more hard lumps remain. When manganese is present the melt is colored green and the solution is pink. The latter is heated until this pink color disappears and is then transferred to a platinum dish (or lacking this, one of porcelain may be used). The small amount of the melt adhering to the sides of the crucible is transferred to the contents of the dish by means of water and hydrochloric acid. The solution is then analyzed as described on page 485. Remark. If the fusion cannot be removed from the crucible, it is placed, together with its cover, in the beaker and treated as above. In this case, if the melt was very green-colored, it should not be decomposed with hydrochloric acid, but with nitric acid, for the chlorine evolved by the action of the hydrochloric acid upon the manganate would attack the platinum. Substances containing considerable fluorine cannot be treated as above, for silicon fluoride will be lost by volatilization. In this case it is necessary to use the old method of Berzelius, The melt from the sodium carbonate fusion is extracted with water, as in the determination of fluorine (p. 472), and the greater part of the silica removed by means of ammonium carbonate. The precipitate is filtered off, ignited, and weighed. The silicic acid remaining in the filtrate is precipitated by means of ammoniacal zinc hydroxide. The precipitate thus ob- tained, consisting of zinc oxide and zinc silicate, is decomposed with hydrochloric acid and the silica obtained by evaporation on the water-bath as usual. As a rule, the insoluble part of the melt contains silicic acid, and this must also be removed by evapo- ration with hydrochloric acid. All three silica precipitates are ignited together and the purity of the silica tested. Besides the sodium carbonate method for the analysis of silicates not decomposable with acids a great number of other methods have been proposed, but of these only the following will be mentioned here. 490 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. (/?) Lead Oxide Method of Jannasch* This analysis is interesting because it permits of an exact determination of the alkalies and of silicic acid in the same sample. Inasmuch as commercial lead oxide (litharge) is not free from impurities, it is prepared for the analysis by the ignition of pure lead carbonate. The lead carbonate is prepared by adding the theoretical amount of ammonium carbonate to a boiling solution of lead acetate. The precipitate is washed several times by decantation with hot water, then transferred to a hardened filter, and com- pletely washed, using suction. The mass is finally carefully re- moved from the filter-paper and dried on the water-bath. Procedure. For each gram of the silicate 10-12 gms. of lead carbonate are used. First of all a little lead carbonate is placed in the crucible, then the very finely powdered substance, and after mixing thoroughly with a platinum spatula the covered crucible is heated for fifteen to twenty minutes over a flame which is not more than 3-4 cm. high, by which means the greater part of the carbon dioxide is expelled. The contents of the crucible are then more strongly heated until fusion is effected, taking care that the flame used is strictly non-luminous; the lower third of the crucible, and no more, may ,be heated to redness. After fusing for ten to fifteen minutes the decomposition is complete, and the covered crucible is quickly touched into cold water, but so that. its contents remain dry. The melt is placed in a platinum dish, covered with hot water and a sufficient quantity of concentrated nitric acid and evaporated on the water- bath, breaking up the melt with a stirring-rod as much as possible. When the cake is completely disintegrated, as is shown by there remaining no more hard yellow pieces and only slightly colored flocks of silicic acid floating in the liquid, the latter is evaporated on the water-bath until a dry powder is obtained ; this is moist- ened with concentrated nitric acid and once more evaporated *Gaston Bong, Zeit. fur anal. Chem , XV11I (1879), p. 270, first pro- posed that silicates be decomposed by fusion with red lead (Pb.,O 4 ), but Jannasch in his Praktisches Leitfaden der Gewichtsanalyse has greatly improved thj" method. ANALYSIS OF SILICATES ORTHOCLASE. 491 to complete dryness. The dry residue is moistened with 20 c.c. of concentrated nitric acid, and allowed to stand fifteen minutes; 100 c.c. of water are added, and the liquid is heated for twenty minutes on the water-bath. The residue of silicic acid is filtered off, washed first with hot water containing nitric acid, then with pure water, and weighed after the usual ignition. Remark. In the analysis of minerals containing fluorine, e.g. topaz, Jarmasch finds that the results obtained are about 0.5-1 per cent, lower than when the Berzelius method is used. In such a case this method of decomposition is used only for the deter- mination of the metals and of the alkalies, after introduction of hydrogen sulphide and removal of the lead. ANALYSIS OF SILICATES. Orthoclase. Constituents : silicic acid (63-70 per cent.) ; aluminium oxide (16-20 per cent.) ; ferric oxide (0.3 per cent.) ; potassium oxide (8-16 per cent.); sodium oxide (1-6 per cent.); and often small amounts of calcium oxide, magnesium oxide, and in rare cases barium and ferrous oxides. Preparation of the Substance for Analysis. The substance is placed upon a thick steel plate within a steel ring (about 2 cm. high and 6 cm. in diameter) and broken into small pieces by means of a hardened steel hammer ; the pieces are then reduced to a coarse powder. The latter is placed in an agate mortar in small portions and ground as fine as possible and preserved in a glass-stoppered bottle. In this way from 5-6 gms. of powder are obtained. By this means, as proposed by Hillebrand, there is less danger of contaminating the substance with small particles of iron than when a so-called steel mortar is used, especially after the latter has been worn rough on its inner surface. Further, the practice of passing the powder through bolting-cloth is to be avoided when possible, as in this way the substance becomes contaminated with 492 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. fibres of cloth and too large an amount of ferrous iron will be found. Weighing the Substance. It is customary to dry the powder before weighing at 100- 110 C. until a constant weight is^obtained. If there is danger of losing combined water by this procedure, it has been recommended to dry the powder in a vacuum over concentrated sulphuric acid. The practice of drying the substance in either of the above ways is, however , to be discountenanced. It is far better to use the air- dried substance for the analysis, and to determine the moisture in a separate sample. This is more accurate, because the dry silicate powder is hygroscopic, so that a portion weighed out to-day is likely to contain a different amount of moisture than one taken to-morrow, and this is not the case when the air-dried powder is taken for the analysis. Further, as Hillebrand has conclusively shown, chemically combined water is not only likely to be ex- pelled by heating at 100 C., but also by drying in a vacuum over sulphuric acid. This is particularly true of the zeolites. In the case of orthoclase, however, only about 0.1 per cent, of moist- ure is present, so that in this particular case accurate results will be obtained by either method. For the analysis two portions must be taken, each amounting to about 1 gm. in weight. The first serves for the determination of SiO 2 , AlA+FeA, CaO, and MgO; the second for that of the alkalies. i Determination of Silica, Aluminium, etc. About 1 gm. of the air-dried substance is placed in a spacious platinum crucible, dried for one hour at 105-7 C., cooled in a desiccator, and weighed. The difference in weight represents the amount of hygroscopic moisture. The dry substance is mixed with 4 to 5 grns. of calcined sodium carbonate by means of a platinum spatula, and the silicic acid is DETERMINATION OF ALUMINIUM AND FERRIC OXIDES. 493 determined exactly as described on p. 488.* The silica obtained is treated with sulphuric and hydrofluoric acids, as described on p. 487, and the residue of A1 2 O3 in the crucible is placed at one side for the present. Determination of Aluminium and Ferric Oxides. The nitrate from the silicic acid contains, besides the chlorides of aluminium, iron, calcium, and magnesium, weighable amounts of platinum, partly coming from the crucible in which the fusion was made, and partly from the action of the ferric chloride and hydrochloric acid upon the platinum dish in which the evapora- tion took place (cf. p. 110, foot-note). To remove the platinum, the solution is heated to boiling and hydrogen sulphide is passed into it. The mixture of platinum sulphide and sulphur is filtered off and the solution is boiled to expel the excess of hydrogen sulphide. The iron is then completely oxidized back to the ferric state by the addition of bromine water and boiling until the excess of the latter is expelled. After this about 10 c.c. of double-normal ammonium chloride solution are added and the boiling-hot solution is precipitated by the addition of a slight excess of ammonia, free from carbonate (cf . p. 149, Remark). * Formerly a single evaporation of the melt with hydrochloric acid was made, and it was assumed that the silica remaining in solution was quan- titatively precipitated with the iron and aluminium by the addition of ammo- nia. After obtaining the weight of the ignited ammonia precipitate it was fused with potassium pyrosulphate and the melt taken up in the dilute sul- phuric acid; the residual silica was filtered off and weighed. The filtrate was analyzed as above described. Hillebrand has recently shown that this procedure is inaccurate. In the first place, the silica remaining in solution is not completely thrown down with the iron and aluminium precipitate, and in the second place the silicic acid is not absolutely insoluble in dilute sulphuric acid. Hillebrand found that from a solution containing 0.20 gm. A1 2 O 3 and 0.0101 gm. SiO 2 as much as 0.0007 gm. SiO 2 could be detected in the filtrate from the ammonia precipitate. From the potassium pyro- sulphate melt he succeeded in obtaining, according to the old method, only 0.0033 gm. SiO.,, while he obtained, by evaporating the solution until fumes of sulphuric acid came off and subsequently diluting with water, as much as 0.0060 gm. SiO 2 , or about twice as much as was at first insoluble. 494 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. The precipitate is allowed to settle, after which it is filtered, and washed twice by decantation with hot water. It is then dissolved by running hot dilute hydrochloric acid through the filter into the beaker containing the greater part of the precipitate. The precipi- tation with ammonia is repeated as before, and after filtering and washing by decantation, the precipitate is transferred to the filter and washed until free from chloride with water containing am- monium nitrate. The precipitate is allowed to drain as completely as possible, and is ignited wet in the crucible containing the residue obtained from the treatment of the impure silica with sulphuric and hydrofluoric acids. After igniting strongly over a good Teclu burner (or the blast-lamp) the crucible is weighed; its contents represents the sum of the aluminium and ferric oxides. For the determination of the ferric oxide, the mixed oxides are fused with potassium pyrosulphate as described on p. 109. The decomposition is complete after two to four hours. The melt is dissolved in water containing a little sulphuric acid and the iron is determined, after previous reduction with hydrogen sulphide, by titration with potassium permanganate (cf. p. 99). If the weight of the Fe 2 O 3 is deducted from the weight of Fe 2 O 3 +Al 2 O 3 , the weight of A1 2 3 is obtained.* Determination of Calcium. The combined filtrates from the ammonia precipitate are evaporated to a small volume, heated to boiling, and precipitated by means of a boiling solution of ammonium oxalate. After standing twelve hours the calcium oxalate is filtered off, and with small amounts of calcium this precipitate is ignited wet in a platinum crucible and weighed. If, however, considerable calcium is present,! * The amount of iron and aluminium can be determined more quickly, though less accurately, as follows: The moist ammonia precipitate is dis- solved in hot dilute sulphuric acid and diluted to a volume of exactly 250 c.c After thoroughly mixing, 100 c.c. are removed by means of a pipette into a beaker and a second portion of the same volume is placed in a 200- c.c. flask. In the first portion the sum of Fe/) 3 + Al 2 O 3 is determined by precipitating with ammonia, filtering, igniting, and weighing; in the other portion the iron is reduced by hydrogen sulphide and then Crated with permanganate t Cf. pp. 76-78. DETERMINATION OF MAGNESIUM. 495 the moist precipitate is redissolved in hydrochloric acid, and again precipitated by the addition of ammonia and a little more am- monium oxalate. The precipitate is ignited strongly, and weighed as CaO. (Cf. p. 70.) Testing of the Calcium Oxide Precipitate for Barium. Although it is usually unnecessary to make either a qualitative or quantitative test for barium in a sample of orthoclase, yet it is likely to be present in traces so that it may be well to show how this can be done. As far as the author knows strontium has never been found in orthoclase. On account of the solubility of barium oxalate in a solution of ammonium oxalate, the barium will rarely be found in the calcium precipitate when a double pre- cipitation is made, except when it is present to an extent of more than 3 or 4 mgms.* To test the calcium precipitate for barium, it is dissolved in nitric acid, evaporated to dryness, and heated for some time at 140 C. The calcium nitrate is dissolved out by ether-alcohol (p. 79, a), and any residue remaining behind is tested in the spectroscope for barium. If an appreciable amount of the latter is found, the calcium must be determined in the ether-alcohol extract. It is carefully evaporated to dryness, the residue dissolved in a little water and precipitated as before by the addition of ammonium oxalate. After standing twelve hours the precipitate is filtered off, washed, ignited, and weighed. If no barium is found with the lime, it is by no means safe to conclude that the former is absent; it can very well have gone into the filtrate from the double precipitation of calcium. This amount will be precipitated with the magnesium as barium phosphate unless it is removed as indicated below. For the quantitative determination of barium a separate portion of the substance is taken (see below). Determination of Magnesium. The combined filtrates from the calcium oxalate are evapo- rated to dryness, ignited in a porcelain dish, and the residue dis- * W. F. Hillebrand, Journ. Am. Chem. Soc., 16 (1894), p. 83. 49 6 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. solved in water to which a few drops of hydrochloric acid have been added. The carbonaceous residue is filtered off, a drop of sulphuric acid added, and the solution is allowed to stand twelve hours to see if any precipitate of barium sulphate will form. In the latter case, the precipitate is filtered off and tested for barium according to Vol. I, p. 62; in the nitrate from the barium sulphate the magnesium is determined as described on page 65. Determination of Barium. If the qualitative tests have shown the presence of barium, a larger sample of the substance is weighed out (about 2 gms.) moistened in a platinum dish with 10 c.c. of sulphuric acid (1:4) and 5 c.c. of hydrofluoric acid are added. The liquid is evapo- rated on the water-bath, with frequent stirring, until the mineral is completely decomposed, which is recognized by there no longer being any sandy particles perceptible on stirring with a platinum spatula. Frequently a further addition of hydrofluoric acid is necessary. When the decomposition is complete, the greater part of the sulphuric acid is removed by heating the contents of the dish in an air-bath. After cooling, the residue is taken up in water, and the barium sulphate is filtered off, and ignited wet in a platinum crucible. The precipitate thus obtained always contains small amounts of calcium sulphate which must be elim- inated. To accomplish this, the residue in the crucible is dis- solved in a little hot concentrated sulphuric acid, and after cool- ing the solution is diluted with cold water. The barium sulphate is now completely free from calcium; it is filtered off, ignited, and weighed. Determination of the Alkalies. (a) Method of J. Lawrence Smith.* Principle. The substance is heated with a mixture of 1 part ammonium chloride and 8 parts calcium carbonate. By this means the alkalies are obtained in the form of chlorides, while the remaining metals are for the most part left behind as oxides, * Am. Jour. Science [2], 50, 269, and Ann. d. Chem. u. Pharm., 159, 82 (1871). DETERMINATION OF THE ALKALIES. 497 and the silica is changed to calcium silicate, as represented by the following equation: 2KAlSi 3 O 8 + 6CaC0 3 + 2NH 4 C1= = 6CaSiO 3 + 6C0 2 + A1 2 3 + 2KC1+ 2NH 3 + H 2 O . The alkali chlorides together with calcium chloride can be removed from the sintered mass by leaching with water, while the other constituents remain undissolved. Preparation. The ammonium chloride necessary for the de- termination is prepared by subliming the commercial salt; the calcium carbonate by dissolving the purest calcite obtainable in hydrochloric acid and precipitating with ammonia and ammonium carbonate. This last operation is performed in a large porcelain dish. After the precipitate has settled, the clear solution is poured off and the precipitate is washed by decantation until free from chlorides. The product thus obtained contains traces of alkalies, but tfie amount present is determined once for all by a blank test and a corresponding deduction made from the results of the analysis; it is usually sodium chloride and amounts to 0.0012- 0.0016 gm. for 8 gms. calcium carbonate. The decomposition was performed by Smith in a finger-shaped crucible about 8 cm. long and with a diameter of about 2 cm. at the top and 1^ cm. at the bottom. Such a crucible is suitable for the decomposition of about 0.5 gm. of the mineral. A larger quantity can be analyzed in a somewhat wider crucible. Filling the Crucible. About 0.5 gm. of the mineral is mixed with an equal quantity of sublimed ammonium chloride by trit- uration in an agate mortar, then 3 gms. of calcium carbonate are added and intimately mixed with the former. The mixture is transferred to a platinum crucible with the help of a piece of glazed paper, and the mortar is rinsed with one gram of calcium carbonate, which is added to the contents of the crucible. The Ignition. The covered crucible is placed in a slightly inclined position and gradually heated over a small flame until no more ammonia is evolved* (this should take about fifteen * During this part of the operation the heat should be kept so low that ammonium chloride does not escape. The latter is dissociated into ammonia and hydrochloric acid by the heat, and the acid unites with the calcium carbonate to form calcium chloride. [Translator.] 49 8 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. minutes), then the temperature is raised until finally the lower three-fourths (and no more) of the crucible are brought to a dull red heat, and this temperature is maintained for 50-60 minutes. The crucible is then allowed to cool and the sintered cake usually can be removed by gently tapping the inverted cru- cible. Should this not be the case, it is digested a few minutes with water, which serves to soften the cake so that it can be readily washed into a large porcelain, of, better, platinum dish. The covered dish is heated with 50-75 c.c. of water for half an hour, replacing the water lost by evaporation, and the large particles are reduced to a fine powder by rubbing with a pestle in the dish. The clear solution is decanted through a filter and the residue is washed four times by decantation, then transferred to the filter and washed with hot water until a few cubic centimeters of the washings give only a slight turbidity with silver nitrate. To make sure that the decomposition of the mineral has been ^com- plete, the residue is treated with hydrochloric acid; it should dissolve completely, leaving no trace of undecomposed mineral. Precipitation of the Calcium. The aqueous solution is treated with ammonia and ammonium carbonate, heated and filtered. As this precipitate contains small amounts of alkali, it is redis- solved in hydrochloric acid and the precipitation with ammonia and ammonium carbonate is repeated. The combined filtrates are evaporated to dryness in a porcelain or platinum dish, and the ammonium salts are removed by careful ignition over a moving flame.* After cooling, the residue is dissolved in a little water and the last traces of calcium are removed by the addition of ammonia and ammonium oxalate. After standing twelve hours, the cal- cium oxalate is filtered off and the filtrate is received in a weighed platinum dish, evaporated to dryness, and gently ignited. After cooling the mass is moistened with hydrochloric acid in order to transform any carbonate into chloride, the evaporation and igni- tion is repeated, and the weight of the contents of the dish is deter- mined; this represents the amount of alkali chloride present. To determine potassium, the residue is dissolved in water, and the * Before igniting, it >s well to heat the contents of the dish in a drying-oven at 110. By this means there is no danger of loss by decrepitation. [Translator.] DETERMINATION OF THE ALKALIES. 499 potassium is precipitated as potassium chloroplatinate, as de- scribed on p. 44. The sodium is determined by difference. (b) The Hydrofluoric Add Method of Berzelius. About 0.5 gm. of the mineral is weighed into a platinum dish, 2 c.c. of water and 0.5 c.c. of concentrated sulphuric acid are added, and mixed with the substance by means of a platinum spatula; after cooling about 5 c.c. of pure, concentrated hydro- fluoric acid, which has been distilled from a platinum retort with the addition of a little potassium permanganate, are added.* The liquid is evaporated on the water-bath, frequently stirring with the platinum spatula, until no more hydrofluoric acid is expelled and no more hard particles can be felt at the bottom of the dish. The dish is heated in an air-bath until the greater part of the sulphuric acid is removed; this is necessary to make sure that the hydrofluoric acid is completely expelled. It is not advisable, however, to remove all of the sulphuric acid, on account of the danger of forming insoluble basic salts. The mass is allowed to cool, covered with 200 c.c. of water, and digested until all of the residue has gone into solution. f The sulphates are now transformed to chlorides by precipitation with as slight an excess of barium chloride as possible; and then, without stopping to filter off the barium sulphate, the aluminium, calcium, and excess of barium are precipitated by the addition of ammonia and ammonium carbonate. The precipitate is allowed to settle, washed four times by decantation, then transferred to the filter and washed free from chloride. The nitrate is evaporated to dryness, and the ammonium salts removed by gentle ignition. A few drops of hydrochloric acid are added, and the magnesium is removed by adding barium hydroxide solution until slightly alkaline, boiling and filtering. The filtrate is treated with ammonia and am- monium carbonate, boiled, and the precipitated barium carbonate filtered off. This filtrate is again evaporated to dryness, the ammonium salts are expelled, the residue is dissolved in a * The permanganate serves to destroy organic matter that is likely to be present in commercial hydrofluoric acid. f If barium was present, it is left behind as the sulphate. 500 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. little water, and a little more barium carbonate is pre- cipitated by the addition of ammonia and ammonium car- bonate. This treatment is repeated until finally the addition of ammonia and ammonium carbonate produces no further pre- cipitation. The last nitrate is evaporated to dryness, gently ignited, moistened with hydrochloric acid, again evaporated, ignited and weighed; this represents the weight of the alkali chlorides together with a small amount of magnesium chloride. The chlorides .are dissolved in a little water, and the potassium precipitated as potassium chlorplatinate (p. 44). If the cor- responding amount of potassium chloride is deducted from the first weight, the amount of sodium chloride plus the small amount of magnesium chloride will be obtained. In order to determine the latter, the alcoholic filtrate from the potassium chlorplatinate precipitate is evaporated to dryness' on the water-bath (the water in -the bath must not boil), and the residue is dissolved in a little water and washed into a small flask. The latter is now fitted with a rubber stopper containing two holes, and through these, two right-angled pieces of glass tubing are introduced, one reaching to the bottom of the stopper and the other until it almost touches the liquid in the flask. The solution is now heated to boiling so that steam escapes from both of the tubes. After boil- ing two minutes we can assume that the air is completely ex- pelled from the flask ; the short tube is connected with a hydrogen generator and a rapid current of hydrogen is conducted through the apparatus, while at the same time the flame is removed from beneath the flask and the long tube is closed by means of a piece of rubber tubing containing a glass rod. The liquid is allowed to cool completely, and the air-space above will be entirely filled with hydrogen. As the hydrogen is absorbed by the liquid, the sodium and magnesium chlorplatinates are reduced to chloride with the deposition of metallic platinum, which floats on the liquid in the form of deridrites : Na 2 PtCl 6 + 2H 2 = 4HC1 + 2NaCl + Pt MgPtCl 6 + 2H 2 = 4HC1 + MgCl 2 + Pt The flask is placed in a lukewarm water-bath, frequently ANALYSIS OF LEPIDOLITE. 501 shaken, and the hydrogen is allowed to act upon the solution until the reduction is shown to be complete by the liquid becoming per- fectly colorless. The connection with the hydrogen generator is now broken and a rapid current of carbon dioxide is conducted through the solution for two minutes through the longer tube in order to remove the hydrogen. This is necessary, as otherwise on opening the flask there is likely to be an explosion between the hydrogen and oxygen, owing to the catalytic action of the plati- num. The platinum is filtered off, the filtrate concentrated, and the magnesium precipitated by the addition of ammonia and sodium phosphate. After standing twelve hours, the magnesium ammonium phosphate is filtered off and the magnesium deter- mined as magnesium pyrophosphate. The corresponding weight of MgCl 2 is deducted from the weight of NaCl+MgCl 2 , and 'in this way the amount of NaCl is determined. Remark. This method is in very general use, and the results obtained agree closely with those by the J. Lawrence Smith method. Many silicates, such as the feldspars, are readily decomposed by the action of sulphuric and hydrofluoric acids; * others, such as certain specimens of tourmaline, only with difficulty. According to Jannasch the members of the andalusite group are not com- pletely decomposed by hydrofluoric acid, but this can be effected by strongly igniting with ammonium fluoride. For this purpose the ignited mineral is placed in a platinum dish, covered with 10 c.c. of ammonia, evaporated to dryness, diluted with water, strongly acidified with concentrated hydrofluoric acid, and again evaporated to dryness. The dish is placed in a nickel beaker and ignited quite strongly, until finally the excess of ammonium fluoride is driven off. The residue is now treated with sulphuric acid (1:2) in order to decompose salts of hydrofluosilicic acid, evaporated on the water-bath as far as possible, and then the greater part of the sulphuric acid is removed. From this point the procedure is the same as in the regular Berzelius method. The Smith method is always applicable and has the advan- tage that the magnesium is practically completely removed at the start. * Many silicates can be decomposed by evaporation with hydrofluoric and hydrochloric acids. F. Hinder, Z. anal. Chem., 1906, 332. 502 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. Analysis of Lepidolite. Lepidolite is a member of the mica group and contains lithium and fluorine with the following composition: Si 3 9 Al 2 (Li,K,Na) 2 (F,OH) 2 SiO 2 = 40 to 45 per ct. ; A1 2 O 3 = 19 to 38 per ct. ; MnO = to 5 per ct. ; MgO = to 0.5 per ct. ; KjO = 4 to*l 1 per ct. ; Li 2 O = 1 to 6 per ct . ; Na 2 O = to 2 per ct.; F=l to 10 per ct.; H 2 O = 1 to 3 per ct. Besides the above, calcium, iron, phosphoric acid, and chlorine are frequently found, and in rare cases small amounts of caesium and rubidium are present. The determination of the silicic acid, aluminium, iron, man- ganese, and magnesium is effected as in the case of the orthoclase analysis, except that in this case the manganese must be separated from the iron and aluminium as described on p. 149 or 152. Determination of the Alkalies. The weight of NaCl+ KC1+ Lid is determined by one of the methods given under the analysis of orthoclase, and the potassium weighed as potassium chloroplatinate. The platinum is then removed by the treatment with hydrogen, or the solution is heated to boiling and the platinum is precipitated as the sulphide by the introduction of hydrogen sulphide. The filtrate free from platinum is evaporated to dryness and the lithium separated from the sodium as described on p. 53 or p. 55. Determination of Fluorine. This determination is the same as an the case of analysis of fluorine in calcium fluoride (p. 471), except that it is unnecessary to add any silica, for the mineral itself already contains a sufficient quantity. Determination of Water. This is effected by the method of Rose-Jannasch (p. 484). Determination of Ferrous Iron in Silicates and Rocks. The very finely powdered, but not bolted, mineral contained in a platinum dish is covered with 5 to 10 c.c. of dilute sulphuric acid (1:4) and placed upon the little triangle (a) Fig. 81, made of glass or, better, platinum. This is placed in the lead vessel C and the latter rests in a paraffin bath (B) . After the cover is placed upon DETERMINATION OF FERROUS IRON. 53 C, a rapid current of carbon dioxide is passed through A, whereby the air within the apparatus will be replaced in about three minutes. The cover is quickly removed, and 5 to 10 c.c. of con- centrated hydrofluoric acid are added; The cover is immediately replaced, and the current of carbon dioxide continued, while the contents of the dish are repeatedly stirred during the whole oper- ation by means of a platinum spatula or piece of coarse wire introduced through the other hole in the cover.* At the same time the paraffin bath is heated to 100 C. and kept at this tem- perature for about an hour. As soon as no more gritty particles FIG. 81. are to be felt, the temperature of the bath is raised to about 120 C. in order to remove the large excess of hydrofluoric acid. This requires about another hour. The dish is then allowed to cool in the carbon dioxide atmosphere and its contents are finally washed into 400 c.c. of cold distilled water, 10 c.c. of concentrated sulphuric acid are added, and the solution is titrated with a potas- sium permanganate solution of known strength until a pink color is obtained which is permanent for several seconds. This end- point is fugitive in proportion to the amount of hydrofluoric acid remaining in the solution. Remark. The above method has been used in the author's laboratory with success for several years. It is a modification of Cooke'sf method in which the decomposition with hydrofluoric * In Fig. 81, this second opening is incorrectly shown. It should really be in tbe middle of the cover directly over the evaporating-dish. f J. P. Cooke, Am. J. Science [2], 44, p. 347 (1867). 504 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. acid took place under a glass funnel upon the water-bath. In this case a large amount of hydrofluoric acid remains in solution and it is difficult to obtain a sharp end-point. Another method for the determination of the amount of ferrous iron present in insoluble silicates is that of Mitscherlich. The silicate is decomposed in a closed tube with sulphuric acid (8 H 2 SO 4 : 1 H 2 O) under pressure, and the resulting solution titrated with potassium permanganate. Tihis method usually gives good results in the case of a silicate analysis, but it is worthless for the analysis of rocks containing pyrite or other sulphides, which on treatment with sulphuric acid are decomposed with evolution of S0 2 .* The latter serves to reduce iron that was originally present in the ferric form, so that a too high result will be obtained. Determination of Small Amounts of Titanium in Rocks. The colorimetric method of A. Weller is best suited for this purpose, and is to be preferred over all gravimetric methods. Procedure. The silicic acid is removed exactly as in the analysis of orthoclase (p. 491) and in the filtrate the iron, aluminium, titanium, zirconium (chromium and vanadium) are separated from the manganese, magnesium, and calcium, by the acetate method. The precipitate thus obtained still contains traces of manganese, so that it is dissolved in dilute hydrochloric acid and reprecipitated by ammonia. The precipitate is ignited in the same crucible in which the residue from the impure silica is con- tained (small amounts of titanium are likely to be in this residue) fused with potassium pyrosulphate, and the melt dissolved in water containing sulphuric acid. Any insoluble silicic acid is filtered off and the titanium determined in the filtrate as described on p. 100 by treatment with hydrogen peroxide. Remark. In rock analysis it is convenient to determine the titanium after the determination of the total iron. For this purpose the solution of the potassium pyrosulphate melt is satu- rated with hydrogen sulphide in order to precipitate the platinum * L. L. de Koninck, Zeit. fiir anorg. Chem., 26 (,1901). 125, and Hille- brand and Stokes, J. Am. Chem. Soc., XXII (1900), p. 625. See also Stokes, Am. J. Sci., Dec., 1901. ZIRCONIUM AND SULPHUR IN ROCKS. 55 and reduce the iron, and the filtrate from the platinum sulphide is titrated with potassium permanganate after expelling the excess of hydrogen sulphide, as described on p. 109. The solution is afterwards concentrated to about 80 c.c., and the titanium de- termined as above. Of the gravimetric methods, that of Gooch is best suited (p. 116), but even this fails in the presence of zirconium (Hillebrand), so that it is in all cases better to employ the colorimetric method. If it is desired to analyze a rock for titanium alone, about one gram should be treated with hydrofluoric and sulphuric acids (see p. 499), the greater part of the sulphuric acid removed by volatilization, in order to make sure that the hydrofluoric acid is expelled, and the residue taken up in water. From this solu- tion the titanium is determined as above. Determination of Zirconium and Sulphur in Rocks. W. F. Hillebrand.* About 2 gms. of the substance are fused with 5 or 6 times as much sodium carbonate (free from sulphur) and 0.5 gm. potassium nitrate in a large platinum crucible. The crucible should be placed through a hole in a piece of asbestos and held in an in- clined position so that none of the sulphur from the flame can come in contact with the contents of the crucible. The melt is taken up in water, a few drops of alcohol are added in order to reduce any manganate to manganous salt, the solution is filtered, and the precipitate washed with dilute soda solution. The filtrate contains all the sulphur in the presence of sodium silicate, f while the residue contains all the barium and zirconium together with the remaining oxides which were present in the rock. (a) Treatment of the Filtrate. This should amount to 100-250 c.c. in volume; it is acidified with hydrochloric acid, heated to boiling, and precipitated with hot barium chloride solution. After standing twelve hours the barium sulphate is filtered off and weighed. * Bulletin of the U. S. Geolog. Survey (1900), p. 73. f Besides sulphuric and silicic acids the filtrate may contain chromic (yellow color), vanadic, molybdic, phosphoric, arsenic, and tungstic acids. 506 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. According to Hillebrand it is not necessary to evaporate the solution to remove the silicic acid before precipitating the sul- phuric acid, for from a dilute solution silicic acid is never precipi- tated with the barium sulphate. (b) Treatment of the Residue. The residue is washed by means of a stream of dilute sulphuric acid (1:20) into an evaporating-dish, and, after digesting for some time, it is filtered through the original filter. The filtrate con- tains aluminium, iron, and the greater part of the zirconium. The residue contains the rest of the zirconium together with barium sulphate and some silicic acid; after being washed, it is ignited in a platinum crucible and freed from silica by evaporation with sulphuric and hydrofluoric acids. The residue in the crucible is then taken up in hot dilute sulphuric acid and filtered. The insoluble portion can be used for the determination of barium (see below). The two sulphuric acid filtrates, containing at the most only 1 per cent, of this acid, are treated with hydrogen peroxide and a few drops of disodium phosphate. Aluminium and iron are not precipitated on account of the acid present, and only traces of titanium are thrown down, while all of the zirconium is precipitated as phosphate, after standing 24 to 48 hours. If the yellow color of the solution should fade away, a little more hydrogen peroxide is added; the precipitate is filtered off, and, even when it is small in amount, it is purified from the titanium as follows: The filter, together with the precipitate, is ignited, fused with a little sodium carbonate, the melt extracted with water and filtered. This residue is likewise ignited, but it is now fused with potassium pyrosulphate, and the fusion dissolved in hot water containing a few drops of dilute sulphuric acid. The solution is poured into a small Erlenmeyer flask of about 20 c.c. capacity, a few drops of 4 per cent, hydrogen peroxide and a few drops of sodi- um phosphate solution are added, and after standing 1 or 2 days the precipitate is filtered off. The latter is now free from titanium in nearly every case, and after ignition it is weighed as zirconium phos- phate. Although zirconium phosphate theoretically contains 51.8 per cent. ZrO.,, there will be no appreciable error introduced if it is SEPARATION OF SOLUBLE FROM INSOLUBLE SILICIC ACID. 507 assumed that one-half the weight of the precipitate represents the amount of this oxide present. Determination of Barium. The above-mentioned precipitate containing all the barium as sulphate, in the presence of calcium and perhaps strontium, always contains a little silicic acid. In order to remove the latter, it is heated with hydrofluoric and sulphuric acids and the residue is fused with sodium carbonate. The melt is treated with water and the carbonates of barium and calcium are filtered off, washed, and then dissolved in hot dilute hydrochloric acid. From this solution the barium is precipitated by the addition of a slight excess of sulphuric acid and ignited wet in a platinum crucible. The pre- cipitate thus obtained contains a small amount of calcium sulphate, which must be eliminated. For this purpose the residue is dis- solved in the crucible by hot concentrated sulphuric acid, and after cooling the solution is poured into water. In this way a precipitate of barium sulphate free from calcium is obtained. It is ignited and weighed. Separation of Soluble from Insoluble Silicic Acid: Lunge and Millberg.* Frequently a mixture of silicates is to be analyzed which is partly decomposed on treatment with acids, with the separation of gelatinous silicic acid, and partly unaffected. The silicic acid deposited from solution by the addition of acids is soluble in 5 per cent, sodium carbonate solution, while quartz and feldspar are not appreciably attacked by the latter (cf. Vol. I. pp. 356, 357). If it is desired to separate the deposited silicic acid from the unattached silicate (usually feldspar and quartz), the substance is treated with acid (hydrochloric or nitric) and evaporated on the water-bath until a dry powder is obtained. This is moistened with acid, diluted, boiled, and filtered. After washing, the residue is digested with 5 per cent, sodium carbonate solution on the water-bath, in a porcelain dish for fifteen minutes. It is then filtered, washed first with soda solution and finally with water. * Zeitschr. f. angew. Chemie, 1S97, pp. 393 and 425 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. If a turbid filtrate should be obtained, a little alcohol is added to the wash water, after which the filtrate will at once run through clear. The alkaline filtrate contains the soluble silicic acid; this can be determined by acidifying and evaporating to dry ness. The residue from the sodium carbonate treatment, consisting of quartz and feldspar, is weighed. In order to determine the quartz, the mixture is acted upon by sulphurfe and hydrofluoric acids, the excess of the latter is removed by heating with sulphuric acid, and the cold residue is dissolved in water, precipitated with ammonia, and the alumina weighed. If this weight is multiplied by 5.41,. the corresponding amount of feldspar is obtained, and if this is deducted from the weight of the quartz -f feldspar, the weight of the quartz will be found. Determination of Soluble Silicic Acid in Clay. Clay contains besides alumina, sand (quartz + breccia) and small amounts of calcium and magnesium carbonates. About 2 gms. of the substance, after having been dried at 120, and being in the form of a not-too-fine powder, are moistened with water, and a mixture of 100 c.c. water and 50 c.c. concentrated sul- phuric acid * is added. The porcelain dish is covered with a watch- glass and heated over a free flame until dense fumes of sulphuric acid vapors are evolved. The contents of the dish are allowed to cool, 150 c.c. of water and 3 c.c. of concentrated hydrochloric acid are added, the solution boiled for fifteen minutes, filtered , washed completely, and the mixture of soluble silicic acid, quartz, and insoluble silicate is treated as above. Remark. It was formerly the custom to separate the soluble silica from the insoluble silica by boiling with potassium hydroxide solution. According to the experiments of Lunge and Millberg, however, this is not permissible because quartz is perceptibly soluble in caustic potash solution. If, on the other hand, the substance is obtained in a very finely-divided condition, even sodium carbonate solution cannot be used for the same reason. * Alexander Subech, Die chem. Industrie 1902. p. 17. ANALYSIS OF CHROM1TE. 509 Analysis of Chromite. Although chromite (chrome iron ore) is not a silicate, it is insoluble in all acids, and can be brought into solution by fusion with alkali carbonates, or borates, so that its analysis will be dis- cussed at this place. Chromite contains 18 to 39 per cent. FeO, to 18 per cent. MgO, 42 to 64 percent. Cr 2 O 3 , to 13 per cent. A1 2 3 , and to 11 per cent. SiO 2 . Calcium, manganese, and nickel are also occasionally present. Of the finely-powdered and bolted mineral, 0.5 gm. is fused in an inclined, open platinum crucible with 4 gms. of pure sodium carbonate * for two hours over a good Teclu burner. After cool- ing, the melt is leached with water, acidified with hydrochloric acid,f evaporated in a porcelain dish until a dry powder is obtained, moistened with hydrochloric acid, taken up in water, and the silica filtered off. The latter is ignited, weighed, and its purity tested with hydrofluoric acid (p. 487). The filtrate from the silicic acid is precipitated hot with hydrogen sulphide and the precipitate of platinum sulphide and sulphur is filtered off. It is then placed in an Erlenmeyer flask, 10 c.c. of ammonium chloride, enough ammonia (free from carbonate) to make the solution alkaline, and a little freshly-prepared ammonium sul- phide are added, after which the flask is corked up and allowed to stand over night. In the morning the precipitate is filtered off, washed twice with water containing a little ammonium sul- phide, then dissolved in hydrochloric acid, and the precipitation by means of ammonium sulphide is repeated. The ammonium salts are removed from the filtrate and the calcium and magnesium determined as described on p. 76-8. The ammonium sulphide precipitate is dissolved in dilute hydrochloric acid, any residue of nickel or cobalt sulphide is fil- * Bunsen fused the chromite with one-third as much SiO 2 and 6 to 8 parts NaaCOs and then subtracted the amount of silica added from the total amount found. This makes the decomposition take place more read- ily, but the author prefers not to add the silica on account of the possibility of thereby introducing an error. t If a dark residue of undecomposed mineral should remain, it is filtered off and again fused with sodium carbonate. 510 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. tered off and dried. This residue is then ignited first in air, then in a current of hydrogen, and finally weighed as metal. It is not worth while to attempt the separation of the nickel from the cobalt on account of the small amount present. The filtrate from the sulphides of nickel and cobalt is freed from hydrogen sulphide by boiling, the iron present is oxidized by evaporating with potassium chlorate and hydrochloric acid, and the iron, chromium, and aluminium are sepatated from the manganese by means of the barium carbonate method (p. 149) and from one another as described on p. Iu7 et seq. In the filtrate from the barium carbonate precipitate, the manganese is separated from the barium as described on p. 122, 6, and determined as sulphide or as sulphate. Remark. If it is desired to determine the chromium alone, this is best accomplished by means of a volumetric process (see Part II). Determination of Thorium in Monazite, according to E. Benz.* Monazite is a phosphate of the rare earths [PO 4 (Ce,La,Di,Th)J. It occurs in so-called "monazite sand" mixed with quartz, rutile, zircon, tantalates, etc., and is at present the raw material used for the preparation of thorium (used in the Welsbach mantle). The value of a sample of monazite sand depends upon the amount of thorium present, and its determination is best effected as follows: Of the bolted monazite sand, 0.5 gm. is intimately mixed with 10 gms. of potassium pyrosulphate in a spacious platinum cruci- ble; the latter is covered and slowly heated until its contents are at a gentle fusion. This is best accomplished by placing the platinum crucible within a larger porcelain one which is provided with an asbestos ring. After no more gas is given off, the cruci- ble is gently ignited over the free flame, and, after cooling, its contents are treated with water and a little hydrochloric acid until it is completely disintegrated. After allowing the residue to settle, it is filtered, treated with a little concentrated hydro- chloric acid, diluted with water, and again filtered. f In the com- * Zeit. f. angew Chem. 15 (1902), p. 297. t This residue is free from thorium, and consists chiefly of silicic and tantalic acids. DETERMINATION OF THORIUM IN MONAZITE. 511 bined filtrates the hydrochloric acid is nearly neutralized with ammonia (the formation of a permanent precipitate is to be avoided, for it will be difficult to redissolve it), the solution is heated to boiling, and 3 to 5 gms. of solid ammonium oxalate are added while the liquid is vigorously stirred. The oxalates of the rare earths are immediately deposited in the form of a coarse powder. To make sure that the precipitation is complete, a little ammonium oxalate solution is added. After standing twelve hours the precipitated oxalates are filtered off, washed once by means of water acidified with nitric acid, then transferred to a porcelain dish, and the last portions of the precipitate are eventually washed from the filter by repeated additions of hot, concentrated nitric acid and water ; the liquid is evaporated almost to dryness. Ten cubic centimeters of concentrated nitric acid (sp. gr. 1.4) and 20 c.c. of fuming nitric acid are then added, the dish covered with a watch-glass anc. heated on the water-bath. After a short time the nitric acid oegins to decompose the oxalic acid, shown by the lively evolution of gas. After no more gas is given off, the watch-glass and sides of the dish are washed down and the solution evaporated to dryness. In order to remove the free nitric acid, a little water is added and the solution evaporated once more; after this the filter fibres present are removed by filtration. It is now necessary to separate the thorium from the remaining earths. This is effected by precipitating the former with hydrogen peroxide as thorium peroxide. On ignition the latter is changed into ThO 2 , in which form it is weighed. The precipitation with hydrogen peroxide takes place as follows: The neutral solution of the nitrates is diluted with 10 per cent, ammonium nitrate solution to a volume of 100 c.c.. heated to 60-80 C., and precipitated by the addition of 20 c.c. of pure 3 per cent, hydrogen peroxide solution. The precipitate, which is colored yellow by traces of cerium peroxide (at the most T 3 3- mg. of the latter is present), is immediately filtered, washed with hot water containing ammonium nitrate, ignited wet in a platinum crucible, and weighed as ThO 2 . If it is desired to obtain an absolutely pure thorium oxide, .'he moist precipitate is dissolved in nitric acid and the above precipitation with hydrogen peroxide is repeated. By this method 512 GRAVIMETRIC DETERMINATION OF THE METALLOIDS. E. Benz obtained in the analysis of a South American monazite sand the following results: 4.72, 4.58, 4.50 per cent. ThO 2 . Remark. The above process for the determination of thorium in monazite is quicker and more accurate than either that of Glaser * or that of Hintz and Weber, | so that it is to be recom- mended for both technical and scientific purposes. The determination of thorium oxide in thorite is carried out in the same way with the difference that instead of fusing the mineral with sodium fluoride and potassium pyrosulphate. it is decomposed by treatment with hydrochloric acid, and the silica removed as usual. The filtrate from the silica is analyzed as above. J For the Analysis of Incandescent Mantles consult the work of T. B. Stillman, Chem. Zeit., 1906, 60. Determination of Water in Silicates. If the mineral on ignition loses nothing but water, the amount of the latter can be determined by the loss in weight. In the great majority of cases, however, other constituents (e.g. C0 2 . SO 2 , Cl. F, etc.) are lost and the substance may undergo an oxidation (FeO is changed to Fe 2 3 . PbS to PbS0 4 . etc.). In such cases the procedure recommended by Jannasch can be used to advan- tage. The substance is heated with lead oxide, the water vapor conducted over a heated mixture of lead oxide and lead peroxide and absorbed in a weighed calcium chloride tube (see p. 484). If the substance on ignition loses simply water and carbon dioxide the former may be accurately determined by the method of Brush and Penfield. The substance is introduced by means *Chem. Ztg.. 1896. p. 612. f Zeitschr. fur anal Chemie (1897). XXXVI. p. 27. % As members of the hydrogen sulphide group are usually present, it is advisable to first remove them and to effect the precipitation of the rare earths with ammonium oxalate from the slightly acid filtrate from the hydro- gen sulphide precipitate. Amer. Journ. Sci. [3], XLVIII (1894), p. 31, and Zeit. fur anorg. Chem, VII (1894). p. 22 DETERMINATION OF WATER IN SILICATES. 513 of a long funnel into a bulb blown on the end of a narrow tube made of difficultly-fusible glass, and the tube is provided with a second bulb about 2 or 3 cm. from the end one. The open end of the tube is connected by means of a short piece of rubber tubing with a short tube drawn out into a capillary, and the substance is heated in the flame of a good Teclu burner. The water is expelled and condenses in the colder portion of the tube, and as a precaution the latter is enveloped in moist blotting-paper. As soon as no more water can be expelled, the end of the tube is heated until it softens and the tube is drawn out between the two bulbs. The front end of the tube now contains the water in the presence of a little carbon dioxide, and the latter must be removed. For this purpose the tube is inclined at an angle of 40, so that the heavier carbon diox- ide will run out of it. The weight of the tube slowly diminishes, but at the end of about three hours it becomes constant, losing about 0.0003 gm. per hour, due to the evaporation of water. If therefore, the tube is allowed to stand three hours before weighing, 0.0009 gm. must be added to the weight of the water. If the substance contained a large amount of carbonate, the escaping carbon dioxide will carry aqueous vapor with it, so that a further correction must be made. One gram of CC>2 at an average baro- metric pressure (760 mm.) and temperature (20 C.) will cause a loss of 0.0096 gm. water vapor. If the amount of carbon dioxide present is known, it is, therefore, only necessary to multiply its weight by 0.0096 to obtain the amount of water that would other- wise escape the determination. Determination of Silicon. See Steel Analysis, p. 441. Determination of Silicon in the Presence of Silicic Acid. Gf. M. Phillips, Z. angew. Chem. ; 14, 1969 (1905). PART II. VOLUMETRIC ANALYSIS. A GRAVIMETRIC analysis is accomplished by adding to the solution of the substance to be analyzed a reagent of only approxi- mately-known strength, separating one of the products of the reac- tion from the solution and weighing it. On the other hand, a volumetric analysis is made by causing the reaction to take place by means of a measured amount of a solution of accurately- known strength and computing the amount of substance present by the volume of the solution which reacts with it (cf. p. 2). For the latter sort of analysis accurately-calibrated measuring instruments are necessary, as will be briefly described. Measuring Instruments. 1. Burettes are tubes of uniform bore throught the whole length; they are divided into cubic centimeters and are closed at the bottom, as shown in Fig. 82, by means of a glass stop-cock, or with a piece of rubber tubing containing a glass bead h. The latter form was devised by Bunsen and is used as follows : The tubing is seized between the thumb and forefinger at the place where the glass bead is, and by means of a gentle pressure a canal is formed at one side of the bead through which the liquid will run out. Instead of the glass bead an ordinary pinch-cock is frequently used. Besides the above forms of burettes, a great many others are in use, but it is unnecessary to describe them here. 2. Pipettes. A distinction must be made between a " full " pipette and a " measuring " one. A full pipette has only one 514 MEASURING INSTRUMENTS. 5*5 mark upon it, and serves for measuring off a definite amount of liquid. They are constructed in different forms; usually they consist of a glass tube with a cylindrical widening at the middle. The lower end i drawn out, leaving an opening about J-l mm. FIG. 82. wide. Pipettes of this nature are constructed which will hold respectively 1, 2, 5, 10, 20, 25, 50, 100, and 200 c.c. Measuring pipettes are burette-shaped tubes graduated into cubic centimeters and drawn out at the lower end as before. They serve to measure out any desired amount of liquid and are obtained with a total capacity of 1, 2, 5, 10, 20, 25, and 50 c.c. 3. Measuring-flasks are flat-bottomed flasks with narrow necks provided with a mark, so that when they are filled to this 516 VOLUMETRIC ANALYSIS. point they will contain respectively 50, 100, 200, 250, 300, 500, 1000, and 2000 c.c. They serve for the preparation of standard solutions and for the dilution of liquids to a definite volume. 4. Measuring-cylinders are graduated into cubic centimeters and are used only for rough measurements. It is clear that accurate results can be obtained by a volu- metric analysis only when the instruments used are accurately calibrated. It should never be tafcen for granted that a pur- chased instrument is correct, but it should always be carefully tested. In the case of measuring-flasks and "full" pipettes, it is best for each one to etch for himself the position on the flask or tube up to which they should be filled with liquid. Normal Volume and Normal Temperature. A liter, which is the volume of a kilogram of water at its maximum density, is taken as the normal volume. If it is desired to mark on the neck of a liter-flask the point to which this volume reaches, the position of the mark depends upon the temperature of the vessel. It is necessary, therefore, to choose for the vessel itself a definite temperature, the so-called normal temperature* At present the temperature of 4- 15 G. is almost universally taken as the normal temperature. According to this, then, the flask should be marked at 15 with the volume occupied by a kilogram of water at +4, and as the kilogram is the unit of mass, the weighing should also take place in a vacuum. This experimental impossibility can be avoided inasmuch as the weight of a liter of water is known accurately at temperatures other than +4, also the expansion of the glass with rise of temperature, and the buoyancy which the weights and the water experience as a result of weighing in the atmosphere. The weights which must be placed upon the balance pan in order to determine the space occupied by a true liter of water, therefore, depend upon the temperature of the water and of the vessel, as well as the density of the air at the time of the experiment. The density of the air varies somewhat from day to day and depends upon the barometric pressure, the temperature, and the amount of moisture. MEASURING INSTRUMENTS. 517 It suffices in most cases, however, to assume the average values of these factors corresponding to the locality. As regards the density of water at different temperatures, this is given in the following table: DENSITY OF WATER AT DIFFERENT TEMPERATURES * t Density. t Density t Density. 0.999867 14 0.999271 28 0.996258 1 9926 15 9126 29 0.995969 2 9968 10 8969 30 5672 3 9992 17 8801 31 5366 4 1 . 000000 18 8021 32 5052 5 0.999992 19 8430 33 0.994728 6 9968 20 8229 34 4397 7 9929 21 8017 35 4058 8 9876 22 0.997795 36 0.993711 9 9808 23 7563 37 3356 10 9727 24 7321 38 0.992993 11 9032 25 7069 39 2022 12 9524 26 0.996808 40 0.992244 13 9404 27 6538 * Thiesen, Scheel and Diesselhorst, 1901. With the aid of this table it is possible to tell what the weight of a liter of water will be at any temperature. It was stated on page 13 that if po is the weight of a body in a vacuum and p that of the same body in the air, then in which expression A denotes the density of the air under the prevailing conditions, s that of the body and si that of the brass weights at t. At t, however, the volume occupied by the mass compared with that at 1^ is where a is the coefficient of cubical expansion. The weight of 5 l8 VOLUMETRIC ANALYSIS. water contained in the mass at t in a vacuum (disregarding quantities of the second order) is: *! * If, therefore, it is desired to determine the volume of a liter by weighing water at 17.35 with brass weights, the computation is carried out as follows: The density of water at 17.35 is given by interpolation in the above table as 0.9987 = s, the density of the brass weights can be taken as 8.0 = si, the density of the air as 0.001214 and the coefficient of cubical expansion of glass as 0.000027, so that by inserting these values in the above equation : 0.9987[1+O.OOOQ27(17.35-15)1 _ Q P 0.001214 0.001214 0.9987 8.0 or in other words the volume occupied by 997.7 gms. of water under the above conditions represents one liter in glass at 15. Invariably the temperature of the laboratory is such that somewhat less than 1000 gms. is used for the calibration in true cubic centimeters. It is convenient, therefore, to place the 1000-gm. weight on one side of the balance together with the empty flask, and then place a tare on the opposite side of the balance. Then the 1000-gm. weight is removed and in its pl,ace 1000 p gms. are placed, after which equilibrium is restored by filling the flask with water. To avoid the somewhat tedious calculation of the value of p, W. Schlosser * has calculated the following table in which the values are given for 1000 p at different temperatures. * Z. angew. Chem., 1903, 960; Chem. Ztg., 1904, 4. MEASURING INSTRUMENTS. Correction Table. 5*9 This table shows in milligrams how much less than 1000 gm. is the weight of water which occupies a volume of one liter, on the assumption that the coefficient of cubical expansion for the glass is 0.000,027 per degree Centi- grade, the normal temperature is 15, the barometric pressure 760 mm. the temperature of the air 15, and the tension of aqueous vapor is normal. The table reads from a temperature of 5.0 to one of 30.9. t 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 5 1341 1340 1339 1338 1338 1338 1338 1338 1338 1338 6 7 8 9 1338 1350 1376 1417 1339 1352 1380 1421 1340 1354 1384 1426 1341 1356 1388 1431 1342 1358 1392 1436 1343 1360 1396 1442 1344 1363 1400 1447 1345 1366 1404 1452 1346 1369 1408 1458 1348 1372 1412 1464 10 1471 1477 1483 1489 1496 1503 1510 1517 1524 1531 11 12 13 14 1539 1619 1713 1819 1547 1628 1723 1830 1555 1637 1733 1841 1563 1646 1743 1853 1571 1655 1753 1865 1579 1664 1764 1877 1587 1673 1775 1889 1595 1683 1786 1901 1603 1693 1797 1913 1611 1703 1808 1925 15 1937 1949 1962 1975 1988 2001 2014 2027 2040 2053 16 17 18 19 2066 2208 2360 2525 2080 2223 2376 2542 2094 2238 2392 2559 2108 2253 2408 2576 2122 2268 2424 2593 2136 2283 2440 2610 2150 2298 2457 2627 2164 2313 2474 2645 2178 2328 2491 2663 2193 2344 2508 2681 20 2699 2717 2735 2753 2771 2789 2807 2826 2845 2864 21 22 23 24 2883 3078 3283 3498 2902 3098 3304 3520 2921 3118 3325 3542 2940 3138 3346 3564 2959 3158 3367 3586 2978 3178 3388 3609 2998 3199 3410 3632 3018 3220 3432 3655 3038 3241 3454 3678 3058 3262 3476 3701 25 3724 3747 3770 3793 3816 3839 3862 3886 3910 3934 26 27 28 29 3958 4202 4455 4716 3982 4227 4481 4743 4006 4252 4507 4770 5041 4030 4277 4533 4797 4054 4302 4559 4824 4078 4327 4585 4851 4102 4352 4611 4878 4127 4377 4637 4905 4152 4403 4663 4932 4177 4429 4689 4959 30 4987 5014 5069 5097 5125 5153 5181 5210 5239 If it be desired to take into consideration the deviation of the temperature and barometric pressure from that assumed in the above table, it is sufficient to add (or subtract) to the figure given 5 20 VOLUMETRIC ANALYSIS. in the table 1.4 mgm. for each millimeter that the barometer reads above (or below) 760 mm., and to subtract (or add) 4 mgm. for each degree that the temperature of the air is above (or below) 15 C. If, for example, the temperature of the water is 17.35, the barometer reading 720 mm., and the temperature of the air 23. 7, then the correction is computed as follows: According to the table the value of 1000 p is 2260 mgm., this number, therefore should be diminished by (760 -720) 1.4 = 56 mgm. (23.7-15)0.4 = 35 91 mgm. The correction becomes 2260-91=2169 mgm. = 2.169 gms. In order to simplify the matter still further, Schlosser recommends preparing a special table for localities where the average barometric pressure is considerably less than 760 mm. Thus the following table applies to Zurich and can be used for other places where the average barometric pressure is correspondingly low. CORRECTION IN GRAMS FOR 1000 C.C. under the assumption that the coefficient of cubical expansion for glass is 0.000027 per degree Centigrade, the normal temperature of glass is 15, the temperature of the water between 5 and 30.5, the barometer reading 720 mm., the temperature of the air 15 and the vapor tension normal. t Correction in Grams. t Correction in Grams. t Correction in Grams. t Correction in Grams. 5.0 1.284 11.5 1.522 18.0 2.303 24.5 3.552 5.5 1.281 12.0 1.562 18.5 2.383 25.0 3.667 6.0 1.281 12.5 1.607 19.0 2.468 25.5 3.782 6.5 1.286 13.0 1.656 19.5 2.553 26.0 3.901 7.0 .293 13.5 .707 20.0 2.642 26.5 4.021 7.5 .303 14.0 .762 20.5 2.732 27.0 4.145 8.0 .319 14.5 .820 21.0 2.826 27.5 4 . 270 8.5 .339 15.0 .880 21.5 2.921 28.0 4.398 9.0 .360 15.5 .944 22.0 3.021 28.5 4.528 9.5 .385 16.0 2.009 22.5 3.121 29.0 4.659 10.0 .414 16.5 2.079 23.0 3.226 29.5 4.794 10.5 .446 17.0 2.151 23.5 3.331 30.0 4.930 11.0 .482 17.5 2.226 24.0 3.441 30.5 5.068 MEASURING I.-.STRUMENTS. 5" To calibrate a 500 c.c. flask for the normal temperature of the glass at 15 by means of water at 19.5 at Zurich, the correction 2 553 is taken from the above table and divided by 2, ' = 1.276 gms. In most cases it is not necessary to take into consideration slight changes in the barometer reading or in the temperature of the air. The Mohr Liter. Before the above tables had been worked out by Schlosser it was customary to avoid the computation otherwise necessary by adopting a standard other than that of the true liter, and the practice is still adhered to by many chemists. Thus for vol- umetric work the liter was taken as the volume of a kilogram of water at either 15 or 17.5 C. as weighed in the air. For all titra- tions this standard is perfectly satisfactory, but it is not suitable for the measurement of the volume of gases in which it is necessary to estimate the weight of a gas from the volume, because the density of gases is always referred to true liters. A Mohr liter measured with water at 15 is 1.0019 and one measured with water at 17.5 is 1.0023 times the volume of a true liter. In other words the former is 1.9 cm. and the latter 2.3 cm. too large. When in the course of this book a liter is mentioned, the true liter is to be understood in all cases. Since, however, many instruments are still graduated at normal temperatures of 15, 17.5 and 20, a table will be given (see p. 522) which can be used for testing such apparatus. Thus, to determine the volume of a Mohr liter for the normal temperature of 15, the liter flask and 1 kg. in brass weights should be counterpoised against a tare. The kilogram weight is then removed, the flask is filled with water at 15 and the position of the meniscus in the neck of the flask is marked. If, however, the temperature of the water is not 15, but say 25. 5 , then evidently the Mohr liter will weigh, according to the above table, 998.095 gms. 522 VOLUMETRIC ANA LYSIS. TABLE FOR PREPARING A MOHR LITER AT THE NORMAL TEMPERATURES OF 15, 17.5 AND 20 C. ACCORDING TO W. SCHLOSSER. Temperature of Water. Normal Temperatures. 15 17.5 20 Grams. Grams. Grams. 15 1000.000 1000.345 1000.763 16 999.871 .217 .634 17 .728 .075 .491 18 .576 999.023 .339 19 .413 .760 .175 20 .237 .584 1000.000 21 .053 .400 999.816 22 998.858 .204 .620 23 .652 998.999 .414 24 .437 .783 .199 25 .212 .558 998 . 973 26 997.977 .323 .739 27 .733 .078 .494 28 .479 997 . 825 .240 29 .218 .563 997.978 30 996.946 .292 .707 Calibration of Measuring-flasks. A flask is chosen with a long neck, as cylindrical as possible, the diameter of which should not exceed a certain value. GREATEST PERMISSIBLE DIAMETER OF THE NECK Contents 2000 1000 500 250 200 100 50 25 c.c. Diameter 25 18 15 15 12 12 10 6 mm. The flask is very carefully cleansed, and dried, after which it is placed upon an accurate balance and counterpoised by a tare. Beside the tare weights are placed corresponding to the volume of the flask, and on the opposite side of the balance weights corresponding to the correction obtained from the table on page 519 corresponding to the temperature of the water to be used, after which equilibrium is again established by pouring distilled water into the flask. Care is taken that no drops of water are left suspended from the sides of the neck above the water-level; if any are present, they are removed by touching with a piece of filter- paper wrapped around the end of a glass rod. An exact equi- librium is finally established by adding or removing a little water MEASURING INSTRUMENTS. 523 by means of a capillary tube. The flask is then placed upon a level surface and a piece of gummed paper with a straight edge is fast- ened around the neck of the flask so that its upper edge is just tangent to the deepest point of the water meniscus. The flask is now emptied, dried, its neck covered with a uniform layer of bees- wax, and allowed to cool; this usually requires about fifteen minutes. The flask is then held, as is shown in Fig. 83, against the piece of wood s, the blade of a pocket-knife is placed firmly against the upper edge of the thick paper ring, and the flask is revolved through 360 around its horizontal axis; in this way a circle is cut in the wax layer. By means of a feather (Fig. 6, p. 22) a drop of hydrofluoric acid is placed along this circle while the flask is held in the horizontal position. By turning the flask around its axis, the drop of hydrofluoric acid is allowed to act upon the glass where the wax coating has been cut. At the end of two minutes the excess of hydrofluoric acid is washed off, the neck of the flask dried by means of filter-paper and heated until the wax melts, when the latter can be readily wiped off. The last traces of wax are removed by rubbing with a cloth wet with alcohol. As it is possible that the etched circle will not exact 1 XT FIG. 83. coincide with the upper edge of the paper, the flask should always be tested.* * The attempt should not be made to test the correctness of the calibra- tion by filling the flask with water which has been brought to a definite temperature; it is important, on the other hand, that the flask and water should be allowed to remain for some time in the same place in order that the temperature of the two may be nearly the same. 524 VOLUMETRIC ANALYSIS. Testing Calibrated Flasks. The flask is counterbalanced with a tare and then weights are added to the tare corresponding to the volume of the flask. The flask is filled with distilled water up to the mark and equi- librium is restored by adding small weights. Thus in testing the calibration of a liter flask, it was filled three times with water at 21.5 after it had been counterbalanced with a tare when perfectly dry. It was found necessary to place small weights on the side with the filled flask amounting to 2.987, 2.893 and 3.122 gms.; average 3.001 gms. If the flask had been perfectly accurate it is found from the table on page 520 that the small weights should have been equal to 2.921 gms. for water at 21.5. The flask is, therefore, 3.001-2.921 = 0.080 cm. too small. This is, however, an unusually good agreement. According to the Royal Commission of Berlin, the allowable error in cal- ibrated flasks is shown by the following table : PERMISSIBLE ERROR FOR FLASKS CALIBRATED FOR CONTENTS (The allowable error for flasks calibrated for delivery is twice as large.) Cpntents . . 2000 1000 500 400 300 250 200 100 50 c.c. Error 0.5 0.25 0.14 0.11 0.11 0.08 0.08 0.08 0.05 c.c. Liter flasks which are calibrated for contents are marked in 1 t^O Germany 1 1. -^- (E) in case they are calibrated in terms of true 15 cubic centimeters, and 1 1. y^ (E),* in case they are calibrated in Mohr liters. Flasks calibrated for delivery are marked with an A instead of the E. Calibration of Pipettes. It is best to have pipettes prepared by the glass-blower and to etch them for one's self. First of all, the pipette must be scrupulously clean; no trace of fat should be left on the inner sides of the tube, for it will cause drops of moisture to adhere and escape 17.5 20 * Or - (E), or E, according to the normal temperature chosen. JL / .O ' ' CALIBRATION OF PIPETTES. 525 measurement. The pipette is, therefore, cleaned by placing it in a tall beaker containing a little soap solution and the latter ig drawn to the top of the pipette by sucking through a rubber tube fastened to its upper end and which is provided with a pinch-cock. The solution is allowed to remain in the pipette for about fifteen minutes. The alkali is then allowed to run out, the pipette washed with water and filled with a warm solution of chromic acid in con- centrated sulphuric acid.* This is allowed to remain from five to ten minutes in the pipette and is then removed, the tube washed first with water from the tap and finally with distilled water. The pipette is now clean and ready to be calibrated. A long strip of paper is fastened upon the upper part of the tube, the lower end is closed with the finger, and the pipette is filled with water which has stood for some time in the balance room, from another of the same size or from a burette. The position of the bottom of the mensicus is noted with a lead-pencil upon the paper which was fastened to the side of the pipette. Assume, for example, that it is desired to calibrate a 10-c.c. pipette, and that the water to be used is at a temperature of 18. According to the table on p. 480 one liter of water at 18 weighs in the air exactly 1000-2.303 = 997.70 gms., consequently 10 c.c. should weigh 9.9770 gms. The point of the pipette is dipped in water, and this is sucked up into the pipette by placing the mouth at the upper end until the water is above the pencil marking. The top of the pipette is then closed with the finger, the water adhering to the outside carefully wiped off, and that inside is allowed to run into a beaker, with the point of the pipette against the walls, until the upper meniscus in the stem is exactly on the mark. The contents of the pipette are then allowed to run into a tared beaker which is covered with a watch glass, or into a glass-stoppered weighing- beaker, allowing the water to flow along the walls of the beaker. Now on weighing the beaker again it is perhaps found that the gain * A solution of potassium dichromate in concentrated sulphuric acid can be used. 526 VOLUMETRIC ANALYSIS. in weight is 9.9257 gms. or 9.9770-9.9257 = 0.0513 gms. too little, A second mark is therefore made a little higher up on the paper attached to the stem of the pipette, and the above process is repeated. If necessary a third mark is made until finally the weight of the water does not vary more than 5 mgms. from that computed. The strip of paper is then cut off at exactly the correct mark, a strip of gummed paper is placed roifnd the pipette at this point, and, after the gum has dried, it is covered with a layer of beeswax and etched with hydrofluoric acid as described on p. 523. After the mark has been etched upon the pipette, it is filled with water up to the mark and emptied into the tared flask. This operation is repeated three times and the mean value is taken as correct. Pipettes may be emptied in several ways: 1. By allowing the contents to run out freely with the pipette held vertically. At the end the end of the pipette is touched to the sides of the beaker. A drop of the liquid will then always remain in the pipette. 2. The solution is allowed to run out while the point of the pipette is held against the side of the vessel into which the liquid is being delivered. All other methods of emptying pipettes, especially that of blowing at the last, are to be abandoned. At all events, it is always necessary to use the pipette in the same way as in the calibra- tion. The " kaiserl. Normaleichungskommision " allows the fol- lowing error in pipettes. Contents of pipette . 100 50 25 20 10 2 1 c.c. Error in c.c 0.07 0.05 0.025 0.025 0.02 0.006 0.006 Error in per cent 0.07 0.1 0.10 0.125 0.2 0.3 0.6 It is possible, however, to prepare pipettes which are mors accurate than this; thus the author by using pipettes as recom- mended above obtained the following values: CALIBZA -ION OF BURETTES. 527 50 c.c. Pipette: 49.9904, 49.9910, 49.9926. Mean 49.9913. /* = 0.002%, ^ = 0.001%. 20 c.c. Pipette: 20.0059, 20.0068, 20.0055. Mean 20.0061. /=0.003%, F = 0.002%, and in the same way: 10 c.c. Pipette: /= 0.008%, F = 0.004. 5 c.c. Pipette: /= 0.011, F = 0.006%. Calibration of Burettes. In volumetric titrations it is advisable to begin each titration with the solution at the zero point of the burette. It is proper, therefore, to calibrate burettes in the same way. The burette is filled to the zero point and a definite volume, e.g., 5 c.c., is allowed to run into a tared beaker, as described for pipettes on p. 525, allowing the tip of the burette to touch the side of the beaker. After determining the weight of the water, the burette is filled again to the zero point and then 10 c.c. are withdrawn in exactly the same way. This process is repeated for each 5 c.c. until finally the 50 c.c. mark is reached, each time determining the weight of the water withdrawn. In withdrawing liquid from a burette until a given point is reached, without waiting for the burette to drain, evidently the amount actually withdrawn depends upon the rate at which it flows from the burette. It is advisable, therefore, to have the tip so narrow that it will take 80 seconds for 50 c.c. to run out. It is true that the burette is not * By / is understood the average error of the single determination. It is computed by the formula /= \ - (cf. Kohlrausch: n 1 Leitfaden der prakt. Physik.), in which n represents the number of deter- minations made, and d, d l} d 2 . . . represent the deviation of each from the arithmetical mean and S (d 2 + d 1 2 + d 2 2 + . . .) the sum of the squares of the errors. -F=\/ and represents the probable error n(n-l) of the mean. 528 VOLUMETRIC ANALYSIS. drained completely in this time,* but according to Wagner f it is sufficiently so for practical purposes. Burettes with rubber tubing at the bottom gradually change with regard to the amount delivered on account of the rubber losing its elasticity. For this reason the tubing should be made quite short and when it begins to get old it should be renewed. The corrections obtained as above are best tabulated by means of a plot in which the bifrette readings are taken as abscissae and the corrections as ordinates. By connecting the points, a curve is obtained by means of which the correct reading of all parts of the burette can be obtained at a glance. Method of Reading Burettes. Although in the case of graduated flasks and pipettes the marks are carried around the whole circumference of the tube, in b the case of burettes this is not usually done,! so that it is a matter of some difficulty to determine with certainty the exact position of the lowest part of the meniscus. To avoid a parallax error a number of means have been devised. Thus floats are often used such as are shown in Fig. 84, a and b ; the former represents that of Beuttel and the latter that of Key. Around the bulb of a a circle is etched, and if the eye is in the correct position, it appears to the observer as a straight line. The liquid in the burette is at the zero-mark, when the projection line from the circle on the float exactly coincides with the line at the zero-point on the burette. In the case of dark-colored liquids it is difficult to see the circle in the case of the float a, but this difficulty is overcome in b by the circle being etched upon the upper bulb (in the figure the latter is drawn too small) . Such floats are weighted so that the upper bulb rises above the level of the liquid in the burette; it is, therefore, easier to make a reading with the float devised by Rey than with that of Beuttel. In refilling the burette, the former float assumes an inclined position; it must, therefore, * W. Schlosser, Chem. Ztg., 1904, 4. f Habilitationsschrift, Leipzig, 1898. J Such burettes can be purchased, however, and very accurate readings can be made with them. CALIBRATION OF BURETTES. 529 be removed, dried off, and again carefully introduced into the liquid.* Schellbach has in vented another method of avoiding the parallax error, by providing the back of the burette with a dark vertical a b c FIG. S5 line upon a background of milky glass as is shown in Fig. 85. When the eye is in the correct position, this dark line is apparently drawn out into two points as shown in b, whereas if the eye is too low the appear- ance a is obtained, or c if the eye is too high. Kreitling f has proved, however, that the use of floats is likely to lead to error, and experiments have also shown that the Schell- bach burettes are not alto- gether reliable. Better than these is the Bergmann's screen as improved by Gockel.J If burettes * This difficulty is overcome by Diethelm by placing below the large bulb a second " flattened-out " bulb, and in this case the float will not attach itself to the sides of the burette, so that it is not necessary to remove it in refilling the burette. |Z. angew. Chem., 1900, 829, 990; 1902, 4. J Chem. Ztg., 1901, 1084. Z. angew. Chem., 1898, 1856. 530 VOLUMETRIC ANALYSIS. are used on which the divisions extend at least half around the tube, and the eye is placed so that the line on the back of the burette coincides with that on the front, then with the use of this screen very exact results are obtained. Bergmann's screen, which can be used to advantage with all kinds of vessel, consists of a wooden test-tube holder painted a dull black (Fig. 86). The reading is made easier if a piece of ground glass or a strip of oiled paper is held behind the burette, or fastened to the screen itself. The " kaiserl. Normalaichungskommission " gives as the ALLOWABLE ERROR FOR BURETTES. Contents 100 75 50 30 10 2 c.c. 0.08 0.06 0.04 0.03 0.02 0.008 c.c. Normal Solutions. By a normal solution is understood one which contains one "gram-equivalent" of the active reagent dissolved in one liter of solution. By " gram-equivalent " is meant the amount of substance corresponding to one gram-atom (1.008 gms.) of hydrogen. For convenience in computation the concentration of solutions used for volumetric purposes are expressed in terms of their normal- ity; i.e., a solution is 2 normal, normal, ^ normal, etc. The letter N is used as an abbreviation for normal. A normal solution of the following substances will contain dissolved in 1000 c.c.: Hydrochloric acid, 1 gm.-mol. HC1 = 36.468 gms. = l gm.-atom H. Nitric acid (as an acid), 1 gm.-mol. HN03 = 63.018 gms. = 1 gm.- atom H. Sulphuric acid, gm.-mol. H 2 S04= =49.043 gms. = l gm.- & atom H. Potassium hydroxide, 1 gm.-mol. KOH = 56.108 gms. = 1 gm.- atom H. Sodium carbonate, ^ gm.-mol. N^ 2 QO 3 = ^- = 53.00 gms.= l & gm.-atom H. Silver nitrate, 1 gm.-mol. AgNOa= 169.89 gms. 1 gm.-atom H. NORMAL SOLUTIONS. 531 How much potassium permanganate or potassium dichromate must be dissolved in 1 liter of solution to obtain a normal solu- tion of each salt for oxidation purposes? First the oxidation equations must be written: 1. In acid solutions potassium permanganate is reduced to salts of potassium and manganous oxides with loss of oxygen ; the latter is taken up by the substance oxidized: 2KMnO 4 = K 2 O + 2MnO + 5O. Two gram-molecules of KMnO 4 , therefore, correspond to 5 gm.- atoms of oxygen, or 10 gm. -atoms of hydrogen. For the normal ICO QO solution jr molecular weight in grams of KMnO 4 , ^ = 31.61 5 gms., should be dissolved in 1 liter, 2. Potassium dichromate is reduced to salts of potassium and chromic oxides: K 2 O 2 7 = K 2 O+ Cr-A-f 30. One gram-molecule of K 2 O 2 O 7 loses 3 gm.-atoms of oxygen, corresponding to 6 gm.-atoms of hydrogen. One liter of normal solution will contain -J- molecular weight in grams of K 2 Cr 2 O 7 , If, however, the potassium dichromate is not used as an oxi- dizing agent, but as a precipitant, e.g. for the precipitation of barium as barium chromate, the case is different: K 2 2 7 + 2BaCl 2 + 2NaC 2 H 3 O 2 +H 2 O = = 2KC1+ 2NaCl+ 2HC 2 H 3 O 2 + 2BaCrO 4 . * The same results will be obtained by considering the change of valence which the manganese and chromium undergo on reduction. The former is reduced (in acid solutions) from a valence of seven to a valence of two, corresponding to five atoms of hydrogen. In the dichromate, each of the two atoms of chromium is changed from a valence of six to a valence of three, equivalent to six atoms of hydrogen. Consequently from this point of view it is evident that | of the molecular weight of permanganate and I of a molecular weight of bichromate must be used to prepare one liter of normal solution. [Translator.] 532 VOLUMETRIC ANALYSIS. One gram-molecule of K 2 Cr 2 7 , therefore, precipitates 2 gm.- atoms of barium, and this corresponds to 4 gm. -atoms of hydrogen, so that \ gm.-molecule of K 2 Cr 2 O 7 will now be sufficient to make 1 294 4 liter of normal solution, ^ = 73.6 gms. It is evident, then, that the amount of substance necessary to make a normal solution depends upon the purpose for which it is to be used. Preparation of Normal Solutions. The required amount of substance should be dissolved in \vater at 15 and diluted to a volume of 1 liter while at this tem- perature. In most cases, however, the water is not at the normal temperature of 15, so that it is customary to dissolve the substance in water at the laboratory temperature and then dilute the solution. up to the mark in a liter flask. After thoroughly mixing the solution, its temperature is taken by a sensitive thermometer. If the temperature is above 15, as is usually the case, the volume of the solution would be less than 1 liter if it were cooled to exactly 15, so that the solution as made up is a little too strong. The error can be computed as follows: Not only the solution, but the glass of the flask should have been at the normal temperature of 15. The coefficient of cubical expansion for glass may be taken as a, and that of the solution as /?. The volume of the flask, and that of the solution is equal to 1000[1 -{-( 15)] c.c., but this volume of solution at t would assume at 15 a volume of U - 15) C Schlosser * has worked out the following table to show how much greater or less a given volume is at different temperatures than it would be at exactly 15. * Chem. Ztg., 1904, 4; 1905, 510. PREPARATION OF NORMAL SOLUTIONS. 533 TABLE FOR THE REDUCTION OF THE VOLUME OF WATER. NORMAL, AND TENTH NORMAL SOLUTIONS TO THE NORMAL TEMPERATURE OF 15 C. Tem- perature. W ater and Vio n - Solutions. Viii. HC1. Viii. H 2 C 2 04. 'An. MO* Via. HNOs Vin. Na 2 C0 3 'An. NuOil 5 + 0.60 + 1.26 + 1.33 + 1.94 + 2.00 + 2.03 + 2.18 6 0.60 1.18 1.25 1.79 1.84 1.87 1.99 7 0.59 1.10 1.16 1.63 1.68 1.69 1.80 8 0.56 1.00 1.05 1.46 1.50 1.50 1.60 9 0.52 0.88 0.94 1.28 1.31 1.31 1.39 10 0.46 0.76 0.81 1.09 1.11 1.11 1.18 11 0.40 0.63 0.67 0.89 0.91 0.90 0.96 12 0.33 0.48 0.52 0.68 0.69 0.69 0.73 13 0.22 0.33 0.35 0.46 0.46 0.47 0.50 14 + 0.12 + 0.17 + 0.18 + 0.23 + 0.23 + 0.24 + 0.25 15 0.00 0.00 0.00 0.00 0.00 0.00 0.00 16 -0.13 -0.18 -0.20 -0.24 -0.25 -0.24 -0.25 17 0.27 0.36 0.40 0.49 0.50 0.49 0.51 18 0.42 0.56 0.61 0.75 0.76 0.75 0.78 19 0.59 0.76 0.82 1.02 1.03 1.02 1.05 20 -0.76 -0.97 -1.05 -1.30 -1.30 -1.29 -1.33 21 0.95 1.19 1.29 1.58 1.58 1.57 1.62 22 1.94 1.41 1.54 1.86 1.87 1.85 1.92 23 1.35 1.64 1.80 2.15 2.17 2.14 2.23 24 1.56 1.88 2.07 2.45 2.47 2.44 2.54 25 1.79 2.14 2.34 2.76 2.78 2.75 2.85 26 2.02 2.40 2.62 3.08 3.10 3.06 3.17 27 2.27 2.67 2.90 3.41 3.43 3.38 3.50 28 2.52 2.95 3.19 3.75 3.76 3.70 3.83 29 2.75 3.23 3.49 4.09 4.10 4.04 4.17 30 -3.06 -3.52 -3.82 -4.43 -4.44 -4.38 -4.52 The use of the tables on pp. 533-535 can be best explained by a few examples : 1. A liter flask is calibrated to contain exactly 1000 true c.c. at 15. A normal solution of sodium hydroxide is prepared at 25. The table shows that the solution at 25 would occupy at 15 2.85 c.c. less, so that in order to make the solution exactly normal, 2.85 c.c. of water should be added. 2. In a titration 47.35 c.c. of normal sodium hydroxide solution were used which was at a temperature of 19; this amount of solution would at the normal temperature occupy a volume of 47.35X0.76 . 7Q _ 47 ' 35 " 1000 534 VOLUMETRIC ANALYSIS. TABLE FOR THE REDUCTION OP THE VOLUME OF A N/10 (Correction is given in Burette Ivciidiiig, 6 8 y u 10 11 12 13 14 15 16 in c.c. 1 + + + + + +0 + + + + + -0 2 3 4 5 . o u 6 7 8 9 10 + + + + + + + + + + + -0 11 1 1 1 1 12 1 1 1 1 1 13 1 1 1 1 1 1 14 1 1 1 1 1 1 15 1 1 1 1 1 1 1 16 1 1 1 1 1 1 1 17 1 1 1 1 1 1 1 18 1 1 1 1 1 1 1 19 1 1 1 1 1 1 1 1 20 + 1 + 1 + 1 +1 + 1 +1 + 1 + 1 + +0 + -0 21 1 1 1 1 1 1 1 1 22 1 1 1 1 1 1 1 1 23 1 1 1 1 1 1 1 1 24 1 1 1 1 1 1 1 1 25 1 1 1 1 1 1 1 1 26 1 1 1 1 1 1 1 1 27 2 2 1 1 1 1 1 1 28 2 2 2 1 1 1 1 1 1 29 2 2 2 2 1 1 1 1 1 30 + 2 + 2 + 2 + 2 + 1 +1 +1 +1 +1 + + -0 31 2 2 2 2 2 1 1 1 1 32 2 2 2 2 2 1 1 1 1 33 2 2 2 2 2 1 1 1 1 34 2 2 2 2 2 2 1 1 1 35 2 2 2 2 2 2 1 1 1 36 2 2 2 2 2 2 1 1 1 37 2 2 2 2 2 2 1 1 1 38 2 2 2 2 2 2 1 1 1 39 2 2 2 2 2 2 1 1 1 40 + 2 + 2 + 2 + 2 + 2 + 2 + 2 +1 +1 + + -0 41 2 2 2 2 2 2 2 1 1 42 2 2 2 2 2 2 2 1 1 43 2 2 2 2 2 2 2 1 1 44 3 3 3 2 2 2 2 1 1 45 3 3 3 2 2 2 2 1 1 1 46 3 3 3 2 2 2 2 1 1 1 47 3 3 3 3 2 2 2 1 1 1 48 3 3 3 3 2 2 2 1 1 1 49 3 3 3 3 2 2 2 1 1 1 50 + 3 + 3 + 3 + 3 + 3 + 2 + 2 + 2 + 1 + + -1 : Computed according to PREPARATION OF NORMAL SOLUTIONS. 535 SOLUTION TO THE NORMAL TEMPERATURE OF 15 C. nro cubic centimeters.*) 17 18 19 20 21 22 23 24 25 26 27 28 29 30 -0 -0 -0 -0 .-0 -0 -0 -0 -0 -0 -0 -0 -0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 2 2 1 1 1 1 1 1 1 2 2 2 1 1 1 1 1 1 2 2 2 2 2 1 1 1 1 1 2 2 2 2 2 3 -0 -0 -0 -1 -1 -1 -1 -1 -2 2 -2 -2 -3 -3 1 1 1 1 1 2 2 2 2 3 . 3 3 1 1 1 1 2 2 2 3 3 3 3 4 1 1 1 1 2 2 2 3 3 3 3 4 1 1 1 2 2 2 2 3 3 3 4 4 1 1 1 1 2 2 2 3 3 3 4 4 4 1 1 1 1 2 2 2 3 3 4 4 4 5 1 1 1 2 2 2 3 3 3 4 4 5 5 1 1 1 2 2 2 3 3 4 4 4 5 5 1 1 1 2 2 2 3 3 4 4 5 5 6 -0 -1 -1 -1 -2 -2 -3 2 -4 -4 -4 -5 -5 -6 1 1 2 2 2 3 3 4 4 5 5 6 6 1 1 2 2 2 3 3 4 4 5 5 6 7 1 1 2 2 3 3 4 4 4 5 6 6 7 1 1 2 2 3 3 4 4 5 5 6 7 7 1 2 2 3 3 4 5 5 6 6 7 8 1 2 2 3 3 4 5 5 6 6 7 8 1 2 2 3 4 4 5 5 6 7 7 8 2 2 3 3 4 4 5 6 6 7 8 8 2 2 3 3 4 4 5 6 6 7 8 9 -2 -2 ^ -3 -4 -5 -5 -6 n -7 -8 g 2 2 3 3 4 5 5 6 7 8 9 9 2 2 3 4 4 5 6 6 7 8 9 10 2 2 3 4 4 5 6 7 7 8 9 10 2 2 3 4 4 5 6 7 8 8 9 10 2 3 3 4 5 5 6 7 8 9 10 11 2 3 3 4 5 6 6 7 8 9 10 11 1 2 3 3 4 5 6 7 7 8 9 10 11 2 2 3 4 4 5 6 7 8 9 9 10 12 2 2 3 4 4 5 6 7 8 9 10 11 12 _ -2 -2 -3 -4 -4 -5 -6 -7 -8 -9 -10 -11 -12 2 2 3 4 5 5 6 7 8 9 10 11 12 2 2 3 4 5 6 6 7 8 9 11 12 13 2 2 3 4 5 6 7 8 9 10 11 12 13 1 2 2 3 4 5 6 7 8 9 10 11 12 13 1 2 3 3 4 5 6 7 8 9 10 11 13 14 1 2 3 3 4 5 6 7 8 9 10 12 13 14 1 2 3 4 4 5 6 7 8 9 11 12 13 14 1 2 3 4 4 5 6 7 8 10 11 12 13 14 1 2 3 4 5 6 6 8 9 10 11 12 14 15 -1 2 -3 -4 -5 -6 -6 -8 -9 -10 -11 -12 -14 -15 the table on page 493. 53 6 VOLUMETRIC ANALYSIS. 3. If a normal solution of common salt is prepared at 25, it is evident that the volume of one liter when reduced to the normal temperature would be 1000-1.79 = 998.21 c.c. The solution is, therefore, too strong, for it contains as much sodium chloride as should be present in 1000 c.c. at the normal temperature. 1 c.c. of this solution is equivalent to 1.0018 c.c. of a normal solution prepared at 15. This number is called the factor of the solution, for if the number of cubic centimeters actually used is multiplied by it, the result represents the corresponding number of cubic centimeters of exactly tenth-normal solution. Similarly in all future experiments, the actual volume should be reduced to the normal temperature when the greatest accuracy is desired. This reduction can be accomplished by means of the tables on pages 534 and 535 for tenth-normal solutions, and by the table on page 533 for more concentrated solutions. If 20 c.c. of a tenth-normal solution is used at 25 this would correspond at the normal temperature of 15 to 20-0.04 = 19.96 c.c. (cf. p. 535). Now calibrated vessels serve not only for the measurement of water but for other dilute and concentrated liquids. The question arises as to whether the volumes as determined by weighing water are accurate for those liquids which differ greatly from water as regards viscosity, adhesion and capillarity. In the case of vessels calibrated for contents, the only difference is that arising from the different nature of the meniscus; but even in the most unfavorable instances no appreciable error is occasioned. Schlosser and Grimm * have carefully studied the amount of error in vessels calibrated for delivery. According to them there is no correction needed for tenth-normal solutions except in the cai3 of iodine. With normal solutions a correction is needed at the most only with hydrochloric and oxalic acids and with liquids in the nature of alkalies and ferric chloride (1 c.c. =0.012 gm. Fe). In the case of concentrated liquids it is those containing alcohol in which the deviations are most marked. Thus Boutron and Boudet found that with an alcohol soap solution, 0.255 c.c. less was delivered * Chem. Ztg., 1906, 1071. PREPARATION OF NORMAL SOLUTIONS. 537 from a 100 c.c. pipette than of water, and from a 25-c.c. pipette, 0.103 c.c. less. With concentrated alkalies and acids the devia- tions were also quite marked; thus with 95% sulphuric acid, 0.442 c.c. less were delivered from a 100-c.c. pipette and 0.085 c.c. less from a 10-c.c. pipette than when water was used. The amount of time 'allowed for the draining of the pipette exerts an important effect in this connection. If the pipette is allowed to drain for a long time the negative correction becomes smaller and may even become positive. It would be well, then, for the chemist to determine the amount of time required to drain a pipette with water and with any other liquid, and if the difference exceeds two seconds to then determine the contents of the pipette for the other liquid. If it be desired to avoid this difficulty, the pipettes may be graduated both for contents and for delivery. The pipette is then filled with the special liquid to the mark corresponding to the former graduation, it is allowed to drain, and then the remaining solution is carefully washed out. SUBDIVISIONS OP VOLUMETRIC ANALYSIS. I. Acidimetry and Alkalimetry. II. Oxidation and Reduction Processes. III. Precipitation Processes. I. ACIDIMETRY AND ALKALIMETRY. This covers the analysis of acids and bases. In order to deter- mine the amount of acid present, an alkaline solution of known strength is required; and conversely, in the analysis of a base, an acid solution is required. In both cases the "end-point" of the reaction is determined with the help of a suitable indicator. The accuracy of the result depends largely upon the choice of the indicator, so that at this place a few words will be said with regard to the indicators most frequently used for detecting the presence of acids or alkalies. yOLU METRIC ANALYSIS. INDICATORS. The indicators used in acidimetry and alkalimetry are dyestuffs which are of one color in acid solutions and another color in dilute alkali. They are, as a rule, weak acids; though some of them are weak bases. It has been found that in organic compounds the color can usually be traced to a particular arrangement of atoms called a chromophor. The change in color, therefore, is caused by a slight rearrangement of the atoms in the molecule. Thus, if the salt of an indicator acid is yellow and on treatment with acid it turns red, this is due to the fact that when the free indi- cator acid is set free by the action of the stronger acid, it under- goes a change whereby a slight change takes place in the way the atoms are linked together in the molecule, and in fact thereby loses temporarily the ability to dissociate electrolytically as an acid. It is not sufficient, however, to assume that this change of color is caused solely by the fact that the ions have a color othjer than that of the un dissociated molecule; on the contrary it has been shown in certain cases that the ions have the same color that the undissociated molecule has before the rearrangement of the atoms in the molecule has taken place. On the other hand, as regards the proper use of indicators it is necessary simply to bear in mind how salts of weak acids behave in the presence of stronger acids and how the acids themselves behave in the pre- sence of alkali. The number of indicators which have been discovered and used more or less is very large, but it will be sufficient here to consider only methyl orange, methyl red, lacmoid, litmus, and phenolphthalein. INDICATORS. 539 I. Methyl Orange.* Under methyl orange, Lunge, t who first proposed the use of this indicator, understood either the free sulphonic acid of dimethyl- amido-azo-benzene or its sodium or ammonium salt. In the free state the free sulphonic acid is obtained in the form of reddish-violet scales, soluble in considerable water. If some of the solid is dissolved in as little water as possible, a distinct reddish- orange colored solution is obtained ; but on the further addition of water this color gradually changes to yellow. If a trace of an acid is added to the yellow solution, it becomes red again and on further dilution with water the color changes to orange and finally to yellow once more, if too much acid was not added. This color change can be easily explained. In the sensitive neutral solution there is a condition of equi- librium between two isomeric forms of methyl orange as expressed by the equation: HS0 3 C 6 H 4 . N : N - The formula on the left represents the yellow substance and the color is due to the azo group N:N, whereas the formula on the right represents the red substance which has for its chromophor the quinoid group iCeH^:. The formula on the left has a sul- phoniq group which imparts acid properties to the molecule and at the other end is an N(CH 3 )2 group which has weakly basic properties. The formula on the right, therefore, represents an * This dyestuff is known commercially as helianthin, orange III, tro- paolin D, Poirrier's orange III, dimethylaniline orange, mandarine orange, and gold orange. fBerichte, II (1878), p. 1944; Zeitschr. f. ch. Industrie, 1881, p. 348; Handbuch fur Sodaindustrie, I (1879), p. 52; II (1893), p. 151. 540 VOLUMETRIC ANA LYSIS. inner salt inasmuch as the acid and base forming groups are here united. The sodium salt of methyl orange is yellow and has the formula NaS0 3 - C 6 H 4 N : NC 6 H 4 X (CH 3 ) 2 and when decomposed by acids the free sulphonate at once reverts to the red form: SO 3 .C 6 H 4 .NH.N:C 6 H 4 :N(CH 3 )2 Methyl orange is an excellent indicator for weak bases, but cannot be used for the titration of weak acids.* If it is desired to titrate a solution containing sodium hydrox- ide with a tenth-normal acid, a little methyl orange is added to the alkaline solution and the acid is added until the solution is colored a distinct red. The latter color will not appear, however, until an excess of the acid has been added. This causes a slight error in the analysis which is greater in proportion to the amount of indicator employed, and the more dilute the solution. It is apparent that the weaker the acid character of the indi- cator the more sensitive it will be, and the opposite is true of in- dicators which are bases. From what has been said the following rule holds : In any titration the smallest amount possible of indicator should be used, and inasmuch as the change of color is propor- tional to the concentration and not to the absolute amount of acid present, the titrated solution should have as nearly as pos- sible the same concentration as was the case in the standardiza- tion of the normal solution. When a normal acid is used for the titration, the change of color is very sharp when the volume of the solution titrated *Cf. Stieglitz, J. Am. Chem. Soc., 25, 1117. INDICATORS. 541 amounts to about 100 c.c. Even with a fifth normal solution the change of color is very distinct, but less so with tenth-normal solutions, but these can be titrated provided the standardization was made at the same dilution as that used in the analysis. How is it with the end-point in the titration of an acid with an alkaline hydroxide solution? If a few drops of methyl orange are added to 100 c.c. of water, the latter will be colored distinctly yellow. If we imagine that the solution contains the same amount of gaseous hydrochloric acid as is contained in 10 c.c. of a tenth-normal solution of this acid, the solution will be colored a deep red. In order that the solution shall assume its original yellow color, it is only necessary N to add exactly 10 c.c. of alkali hydroxide solution, but no ex- cess of alkali, because the water is itself sufficient to decompose the dyestuff sufficiently to produce the yellow color. It is evident, then, that it is not a matter of indifference in the analysis whether the titration is completed by the addition of acid or by the addition of alkali. In the former case, for the N N titration of T c.c. of alkali solution, T-M c.c. of acid would be necessary. Methyl orange is more sensitive toward alkali than it is toward acid, but many prefer to finish the titration by the addition of acid, for most eyes can detect the change from yellow to red with greater accuracy. In principle it is more accurate to accomplish the titration the other way, as was recommended by F. Glaser. Preparation of Methyl-orange Solution. The solution of 0.02 gm. of solid methyl orange * dissolved in 100 c.c. of hot water is allowed to cool, and any deposited meta-sulphonic acid is filtered off. Use. Methyl orange is suitable for the titration of strong acids (HC1, HNO 3 , H 2 SO 4 ) as well as phosphoric and sulphurous acids. Hydrochloric and nitric acids can be titrated with this * If the free acid is not at hand, 0.022 gm. of the sodium salt is dissolved N in 100 c.c. of water, 0.67 c.c. r^r HC1 is added, and after standing some time an> deposited crystals are filtered off. 542 VOLUMETRIC ANALYSIS. indicator with a sharper end-point than is the case with sulphuric acid If free phosphoric acid is titrated with sodium hydroxide using this indicator, the solution changes from red to yellow when one-third of the phosphoric acid has been neutralized : H 3 PO 4 + NaOH = NaH 2 PO 4 + H 2 O. The primary phosphates are neutral toward methyl orange, while the secondary and tertiary phosphates react alkaline toward it. With half -normal solutions, the end-point of the reaction is fairly sharp, with tenth-normal solutions it is less so; in the latter case an excess of about 0.3 c.c. of the tenth-normal alkali is nec- essary to cause the change from red to yellow. Sulphurous Acid. In titrating sulphurous acid with sodium hydroxide, the yellow color is obtained when half the acid has been neutralized, H 2 SO 3 + NaOH = NaHSO 3 + H 2 O, so that NaHSO 3 is neutral toward this indicator. The weak acids HCN, CO 2 , H 2 S, As 2 O 3 , B 2 O 3 , Cr0 3 when pres- ent in considerable amount do not act upon the indicator. CO 2 and H 2 S produce an orange-red coloration only when present in large amounts. For this reason the alkali salts of these acids can be titrated with accuracy by means of this indicator. Organic acids cannot be titrated with methyl orange. The strong and weak bases NaOH, KOH, NH 4 OH, Ca(OH) 2 , Sr(OH) 2 Ba(OH) 2 , and Mg(OH) 2 can be titrated with great accu- racy by means of this indicator, and the same is true of the amine bases (methyl and ethyl amines, etc.): on the other hand, such weak bases as pyridine, aniline, and toluidine cannot be titrated. Nitrous acid ordinarily cannot be titrated with this indicator because the acid destroys it. If, however, an excess of alkali is first added to the solution of nitrous acid, then the methyl orange, the titration can be accomplished with accuracy. B. THE SODIUM SALT, N(CH 3 ) 2 C 6 H 4 N=N C H 4 S0 3 Na. This sodium salt can be used as an indicator in the same way as the free acid ; it should be mentioned, however, that the com- INDICATORS. 543 mercial salt often contains small amounts of sodium carbonate as impurity, which causes it to be slightly less sensitive than the free acid. The change of the yellow color of the solution in this case takes place when the salt has been decomposed by the addition of an equivalent amount of a stronger acid, and the dissociation of the free acid diminished by increasing the concentration of the hydrogen ions. As a matter of fact, however, the amount of acid necessary to effect this change in a solution containing a drop of the indicator solution is inappreciable. 2. Methyl Red. ( CH 3 ) 2 N C 6 H 4 N = N C 6 H 4 COOH. Para-dimethyl-amido-azo-benzene-o-carboxylic add. This valuable indicator is suitable for titrating weak organic bases and ammonia. The aqueous solution of methyl red is orange, but if a few drops are added to 50-100 c.c. of water, the latter is colored a pale yellow. The addition of a drop of 0.1 N. HC1 at once turns the liquid a violet red without passing through any intermediate shade and by the addition of a drop of ammonia the solution becomes nearly colorless again. Methyl red is not very sensitive toward carbonic acid, but more so than is methyl orange, so that it is less suitable for the tit ration of carbonates. The chief advantage of this indicator lies in the sharp color change from a very pale yellow to a violet red, even in titrating ammonia. Preparation of the Indicator. About 0.02 g. of the free acid is dissolved in 100 c.c. of hot water, the solution allowed to cool, and then filtered. Two or three drops of this solution are added for every 100 c.c. of the solution to be titrated. H. W. Langbeck f recommends the use of ortho-nitrophenol as indicator, but it has no advantages over methyl orange and methyl red. It is not at all sensitive towards carbonic acid. It is turned yellow by alkalies and colorless by acids. * E. Rupp and R. Loose, Berichte, 41, 3905 (1908). f Chem. News, 43, 162. 544 VOLUMETRIC ANA LYSIS. 3. Lacmoid, or Resorcin Blue, C 12 H 9 8 N. Lacmoid is prepared by heating resorcin with sodium nitrite at not too high a temperature. The constitution of the dye has not been completely established. Pure lacmoid is soluble in water (the impure product is difficultly soluble) , but more soluble in alcohol, glacial acetic acid, actoife, and phenol, and less so in ether. To determine whether a sample of commercial lacmoid is suitable for use as an indicator, a little of it is boiled with water; if the water is colored an intense and beautiful blue, it can be used. In this case the alcoholic solution will be of a pure blue color, and not with a tinge of violet, as is the case with the impure substance. Preparation of Pure Lacmoid. The solution of the good com- mercial product in hot 96 per cent, alcohol is filtered and allowed to evaporate in vacuo over concentrated sulphuric acid. Preparation of the Indicator. A solution is used containing 0.2 gm. of the purified lacmoid in 100 c.c. of alcohol. Behavior of Lacmoid toward Acids and Bases. If the solution after it has been colored reddish by acid is treated with a solu- tion of an alkali hydroxide, the red color is gradually changed to a violet-red, and on further addition of alkali, it suddenly changes to a pure blue. If the violet solution is diluted with considerable water, it becomes blue. Uses. Lacmoid is suitable for the titration of strong acids and bases as well as for ammonia, but is not suited for the titration of nitrous acid or weak acids. 4. Litmus. The chief coloring principle of litmus, the azolitmin, is a dark- brown powder only slightly soluble in water and insoluble in alcohol and ether. With alkalies it forms a readily soluble blue salt. Besides the azolitmin, there are other dyestuffs present in litmus which are soluble in alcohol with a red color. Commercial litmus is obtained in small cubes mixed with con- siderable calcium carbonate; the dyestuffs are then in the form of their calcium salts, soluble in water. If the commercial mate- INDICATORS. 545 rial is dissolved in water, a solution of blue and reddish-violet coloring matter is obtained, which becomes red on the addition of acid. On making alkaline again, a pure blue color is not obtained at first, but a reddish-violet, which becomes blue on the addition of considerable alkali. Such a solution, therefore, is far from being a sensitive indicator and cannot be used for accurate work. A number of different methods have been proposed for obtaining a sensitive litmus solution, and that of F. Mohr* will be de- scribed. Purification of Litmus. The cubes of litmus are placed in a porcelain dish (without powdering), covered with 85 per cent, alcohol, and digested on the water-bath for some time with frequent stirring. The solution is decanted off and the operation is repeated three times. By this means the undesired coloring matter is removed. The residue is now extracted with hot water, and as it is very difficult to filter the solution, it is poured into a tall cylinder, and after standing several days the clear liquid is siphoned off. The solution is concentrated to about one-third of its volume and acidified with acetic acid in order to decompose the potassium carbonate present. It is then evaporated to a syrupy consistency upon the water-bath and the mass covered with a large amount of 90 per cent alcohol. By this means the blue coloring matter is precipitated, while the remainder of the violet substance remains in solution with the potassium acetate. The residue is filtered off and dis- solved in sufficient hot water so that three drops of the solution will be necessary to impart a distinct color to 50 c.c. of water. Use. Litmus can be used for the titration of inorganic and strong organic acids, alkali and alkaline-earth hydroxides, and ammonia, as well as for the titration of carbonates in hot solm- tion. 5. Phenolphthalein. Phenolphthalein is a very weak acid forming red salts which contain the strongly chromophoric quinoid group rCeH^:. The free acid, however, is unstable and when set free from one of its colored salts reverts instantly into a colorless lactoid form, con- taining no chromophor group : HOOC C 6 H 4 C(C 6 H 4 OH) : C 6 H 4 : O^OOC - CoH 4 C(C 6 H 4 OH) 2 _ I I _ * Lehrbuch der Chemisch-Analytischen Titrirmethode. 546 VOLUMETRIC ANALYSIS. In the case of the free acid, therefore, the condition of equi- librium favors the lactoid form and only minimal traces of the quinoid acid are present. This trace of quinoid acid is ionized and is in equilibrium with its ions : HOOG G 6 H 4 C (C 6 H 4 OH) : C 6 H 4 : O^ <=>H' + OOC C 6 H 4 C (C 6 H 4 OH) : C 6 H 4 : 0' The addition of an alkali causes the hydrogen ions to disappear, so that more of the quinoid molecules must be ionized to preserve equilibrium, and the quinoid molecules in turn be reproduced from the lactoid as fast as the former are converted into the salt. Phenolphthalein is a very sensitive indicator towards acids, but on account of being such a weak acid it does not form stable salts with weak bases. Preparation of the Indicator. One gram of pure phenolphthalein is dissolved in 100 c.c. of 86% alcohol. Uses. P'.ionolphthalein is particularly suited for the titration of -organic and inorganic acids and strong bases, but not for the titration of ammonia. ' If the red-colored solution containing phenolphthalein and a little alkali is treated with an excess of concentrated alkali hydrox- ide solution, the red color disappears, but returns on diluting the solution with water. Phenolphthalein, therefore, cannot be used as an indicator for the titration of concentrated alkali without previous dilution with water. Phenolphthalein is the most sensitive indicator we possess toward acids, far more sensitive than methyl orange, for in this case not only can the presence of weak acids be detected, but very small amounts can be titrated with accuracy. Ordinary distilled water usually contains carbon dioxide, as N can be shown by slowly adding barium hydroxide solution, drop by drop, to 100 c.c. of water containing a drop of the indicator solution. Where the alkali first meets the water, a red color is produced which disappears on stirring, so that often as much as 0.5 to 1.8 c.c. of the alkali must be added before a permanent red color is obtained. The disappearance of the red shows the pres- INDICATORS. 547 ence of acid (in this case carbonic' acid), and its amount corre- sponds to the alkali neutralized. Phosphoric Acid. If a solution of phosphoric acid containing phenolphthalein is titrated with normal sodium hydroxide solu- tion, a permanent coloration is produced when two-thirds of the phosphoric acid is neutralized : H 3 PO 4 + 2NaOH = Na 2 HPO 4 + 2H 2 O. Apparently Na 2 HPO 4 reacts neutral toward phenophthalein, but this is not quite correct, for a pure solution of disodium phos- phate is colored by phenolphthalein a pale pink, and on diluting with water the intensity of the color increases owing to progres- sive hydrolysis: Na 2 HPO 4 + H OH <= NaOH' + Na H 2 PO 4 . During the titration of phosphoric acid with sodium hydrox- ide, a pale-pink color is obtained somewhat too soon, and this color gradually increases in intensity until finally a maximum is reached ; the latter point is taken as the end-point. It is possible that this hydrolysis could be prevented by the addition of a large excess of sodium chloride and cooling to about zero Centigrade. Carbonic Acid. If the solution of a neutral alkali carbonate is treated with phenolphthalein a red color is obtained, showing the presence of hydroxyl ions in the solution, due to hydrolysis : Na Na CO 3 + H OH<=>Na OH+ Na HCO 3 . If hydrochloric acid is added to such a solution which is not too dilute and is at a temperature of C., decolorization is effected when half of the soda has been neutralized. At ordi- nary temperatures* a sharp end-point cannot be obtained; the color gradually fades. Pure sodium bicarbonate dissolved in ice-cold water is not colored by the addition of phenolphthalein; if it is warmed to the temperature of the room it turns red, but on cooling the color disappears (Kiister). Silicic acid seems to be without influence upon phenolphtha- lein, for alkali silicates (the water-glasses) can be titrated with accuracy. Chromic Acid and Acid Chromates are changed by the addi- tion of alkali to neutral chromates and the latter have no action upon phenolphthalein. 548 VOLUMETRIC A "NA LYSIS. Alkali aluminates can be titrated accurately with this indicator, tor aluminium hydroxide does not affect it. Almost all the problems involved in acidimetry and alkalim- etry can be solved by the use of one or the other of these two indicators: methyl orange and phenolphthalein. For further information with regard to the countless other indicators which have been proposed, the student is referred to Glaser's "Indica- toren der Acidimetrie und Alkalimetfie," Wiesbaden, 1901.* NORMAL SOLUTIONS. For the standardization of the solutions used in acidimetry and alkalimetry, a great many different methods have been pro- posed, all of which more or less satisfactorily answer the purpose. It was Gay-Lussac who first proposed the use of chemically-pure, calcined sodium carbonate, and in simplicity and accuracy this, method has never been excelled,f so that we will content ourselves with its description. The chemically-pure sodium carbonate must form a clear solu- tion with water and should contain neither sulphuric nor hydro- chloric acids. It is possible to obtain the pure substance com- mercially, but as a rule it must be purified. For this purpose about 300 gms. of crystallized sodium carbonate are dissolved in 250 c.c. of water at 25-30 C., and quickly filtered into a two-liter flask of Jena glass. After replacing the air by carbon dioxide, J the flask is closed by means of a perforated rubber stopper through which a short, right-angled glass tube is passed, and the latter is connected by means of a long piece of rubber tubing with a Kipp-carbon dioxide generator. The contents of the flask are shaken until no more carbon dioxide will be absorbed; this usu- ally takes from half to three-quarters of an hour. In proportion as carbon dioxide is absorbed, sodium bicarbonate is deposited. The solution is cooled to C., while the carbon dioxide is continu- ally passed through it; the thick mass of crystals is transferred to a * See also J. Wagner, Zeitschr. fur anorg. Chem., XXVII (1901), p. 138. t According to Sorensen, the standardization takes place with equal accuracy by means of anhydrous sodium oxalate, which after weighing is heated until the carbonate is fonned; cf. page 597. % The carbon dioxide is passed through a solution containing sodium bicarbonate before it reaches the flask. NORMAL HYDROCHLORIC ACID. 549 filter-plate which is covered with a piece of hardened filter paper and sucked as dry as possible. The sodium bicarbonate thus obtained often contains considerable chloride and sulphate. It is washed back into the flask by means of 50 c.c. of distilled water (that has been cooled to C. and saturated with carbon dioxide), vigor- ously shaken, and the mother-liquor once more removed by suc- tion. This operation is repeated until finally 3 gms. of the salt will no longer give the test for chlorides or sulphates. The pure sodium bicarbonate thus obtained is dried on the water-bath and preserved for further use in a tightly-stoppered bottle. Normal Hydrochloric Acid. 1000 c.c. contain 1 HC1 =36.468 gms. Pure, concentrated hydrochloric acid is diluted until its spe- cific gravity is about 1.020, and in this way a solution is obtained that is slightly more than normal in strength. To obtain an ex- actly normal solution, it is titrated against a weighed amount of chemically-pure sodium carbonate, and from the result obtained the amount of water to be added can be computed. About 8 gms. of the pure, dry sodium bicarbonate are placed in a large plati- num crucible, and the latter is inserted in an inclined position within a hole in a piece of asbestos board and over a small flame (cf. p. 358). The contents of the crucible are stirred frequently with a short piece of heavy platinum wire, and only the bottom of the crucible is heated to redness. The mass must not be al- lowed to sinter together or fuse, for in that way an appreciable amount of the normal carbonate would be decomposed. After heating for about half an hour the crucible is cooled in a desicca- tor, weighed, and to make sure that a constant weight has been obtained, the heating is repeated once or twice more.* * If it is feared that some of the carbon dioxide may be expelled from the normal carbonate, the bicarbonate may be heated for half an hour at 270- 300. This can be easily accomplished by embedding the platinum crucible, which contains the bicarbonate, in sand, so that the latter extends up on the outside of the crucible as high as the bicarbonate on the inside, and then heat slowly to 230. The heating is then continued for about half an hour, but taking care that the thermometer in the sand beside the crucible does not register above 300. 55 VOLUMETRIC ANALYSIS. The amount necessary to neutralize 35-40 c.c.* of normal acid (about 2 gms.) is weighed out from a glass-stoppered weighing- tube into a beaker, dissolved in about 100 c.c. of distilled water, and enough methyl orange is added (from 5-6 drops) to impart a pa'e-yellow color to the solution. The hydrochloric acid at 17- 18 C. is added from a burette, with constant stirring, until the color of the solution is changed from yellow to orange. The burette is then read and a drop mor of the acid is added to see whether this will produce a pure pink color. If this is not the case, more hydrochloric acid is added until this point is reached, and in this way the number of cubic centimeters of the acid that are required to neutralize the weighed amount of the sodium carbo- nate is determined. Assuming that for the neutralization of 2.1132 gms. of Na 2 CO 3 , 39.20 c.c. of hydrochloric acid at 19 were necessary, how strong is the acid? If the acid were exactly normal, according to definition (p. 530) 1000 c.c. would neutralize Na2 ^ 3 = ^^P = 53.00 gms. of 2i 2i sodium carbonate, so that the amount weighed out would require for neutralization at 15 53-00:1000 = 2.113:a; 2113 53.00 = 39.87 c.c. This, would be equivalent to 39.90 c.c. at 19.f As, however, only 39.20 c.c. were necessary it is evident that our solution is too strong, and for each 39.20 c.c. of the acid, 39.90 39.20 = 0.70 c.c. of water must be added to make it normal, and to 1 liter: 39.2: 0.70= 1000 :x 700 39.2 = 17.86 c.c. water. * It is best not to weigh out more substance than can be titrated with one buretteful, and not too small an amount should be taken, for in the latter case the error in reading is too great. t According to the table on page 533, 1000 c.c. N. HC1 at 19 = 1000-0.76 NORMAL HYDROCHLORIC ACID 551 A perfectly dry liter flask is, therefore, filled exactly to the mark with acid, and 17.9 c.c. of water are added from a burette (or measuring-pipette), the solution is thoroughly mixed, and the strength of the solution is verified by a second titration with a weighed amount of sodium carbonate. Further, it is to be rec- ommended that the beginner should convince himself of the accu- racy of the result by determining the amount of chlorine present gravimetrically as silver chloride. 10 c.c. of normal acid yield 1.4338 gms. AgCl. For practical purposes it is quite unnecessary to spend the time necessary for the preparation of an exactly normal solution, but its normality * is determined, and if the number of cubic centi- meters used is multiplied by this factor, the corresponding amount of normal solution will be obtained. Thus in the above case 39.20 .c. of acid were used to do the work that would require 39.90 c.c. OQ qrv of normal acid. The solution is, therefore, - -=1.018 N. Or, oU .^u if instead of using 40.10 c.c. it was found that 40.15 c.c. of acid 40 10 were required, the solution would be ~r. '-^-= = 0.9987 N. Whatever 40.15 the normality may be, it is written upon a label and pasted upon the bottle containing the acid. For most purposes, a normal solution is too strong, so that J, | and T V N solutions are used. Obviously a tenth-normal solu- tion can be prepared by diluting 100 c.c. of a normal solution to 1 liter, etc. N In order to titrate a acid solution with sodium carbonate. about 0.2 gm. of the salt is placed in a white porcelain dish and dis- solved in 50 c.c. of water, methyl orange is added until a pale-yellow color is obtained, and acid is added until the color becomes orange. The carbon dioxide is then expelled by heating to boiling, after which the solution is cooled and once more titrated until an orange color is obtained; the second titration requires but about 0.1-0.2 c.c. more, but in this" way the correct end-point is obtained. At this * dilution the carbon dioxide exerts an imperceptible action upon the indicator. * By normality is understood the relation to a normal solution. 55 2 VOLUMETRIC ANA LYSIS. Normal Nitric and Sulphuric Acid Solutions. These are prepared in the same way as was described in the preparation of normal hydrochloric acid. Oxalic Acid. 10 1000C.C. contain . = 6 . 30 3 gms. An oxalic acid solution of this strength can be prepared by dissolving exactly 6.303 gms. of pure, crystallized oxalic acid in water at 17.5 and diluting to a volume of 1 liter. The com- mercial acid, however, must always be purified. The chief impurities found in the commercial product are cal- cium and potassium oxalates. In order to remove these salts, 500 gms. are dissolved in 500 c.c. of pure, boiling hydrochloric acid of specific gravity 1.075 in a porcelain dish. If an insoluble resi- due should be obtained, the solution is filtered through a hot-water funnel and the filtrate received in a porcelain evaporating-dish, the latter placed upon ice and cooled as quickly as possible. The fine crystals thus obtained are placed in a funnel provided with a platinum cone and the mother-liquor completely removed by suc- tion. The above process is repeated, and the crystals obtained the second time are washed with a little ice-cold water, re- crystallized three times from hot water, and their purity tested. A solution of 2 gms. of the purified acid should give no sign of a turbidity with silver nitrate, and another portion of 5 gms. should leave no weighable residue after ignition in a weighed platinum dish. After having been dried as completely as p'os- sible by suction, the crystals are spread out upon several layers of blotting-paper and allowed to stand in the air for several days; they then have the formula H 2 C2O4 + 2H 2 O. The strength of the N solution is tested by titration with sodium hydroxide solution using phenolphthalein, as indicator (see p. 553), or with ^ potassium permanganate solution (see p. 598). NORMAL SODIUM HYDROXIDE SOLUTION. 553 Normal Sodium Hydroxide Solution. 1000 c.c. contain 1 NaOH = 40.01 gms. About 45 gms. of the commercial caustic soda are roughly weighed out, the carbonate on the surface is washed off as much as possible by a stream of water from the wash-bottle, and the alkali is dissolved in a little more than a liter of water. The solution is then allowed to stand for about one hour beside the hydrochloric acid against which it is to be titrated, in order that both solutions may be at the same temperature. About 40 c.c. of the solution are measured off from a burette, and titrated with normal hydrochloric acid after the addition of a few drops of methyl orange solution. The titration is repeated several times with fresh amounts of the sodium hydroxide and from the mean of the results the amount of water to be added is calculated. If, for example, 40 c.c. NaOH = 41.23 c.c. N. HC1, it is evident that 1.23 c.c. of water must be added to each 40 c.c. of the alkali to make the solution exactly normal, and for one liter 40:1.23 = 1000:^ 1230 x= = 30.75 c.c. water. 40 After the solution has been diluted with water until it is exactly normal, it must be tested once more with the hydrochloric acid, and from it tenth-normal and fifth-normal hydroxide solutions can be prepared. The solutions thus obtained always contain carbonate, so that they are not suitable for titration with phenolphthalei'n, but with methyl orange the results obtained are the same as if all of the sodium was present as the hydroxide. With phenolphthalei'n accurate results can be obtained from a boiling-hot solution, or by preparing a solution of alkali free from carbonate. 5^ VOLUMETRIC ANA LYSIS. Titration of Alkali containing Carbonate with Phenolphthalein in Hot Solutions. The alkali is measured into a porcelain dish, a drop of phenol- phthale'in added, and hydrochloric acid of approximately the same strength ,s run into the solution until the red color disappears. The solution is then heated to boiling, when the red color soon reappears; it is cooled by placing the dish in cold water,* hydrochloric acid is again added until decolorized, and the process is repeated until finally the red color does not reap- pear on boiling. This method of titration is tedious, but the N results obtained are accurate. On titrating acids with methyl orange as indicator, there is no sharp change from yellow to pink, as is the case with normal and half -normal solutions, but first a brownish-orange color is obtained which becomes pink on the addition of more acid. The correct end-point is the change from yellow to yellowish brown. Only when considerable car- bonate is present will this change occur before enough acid has been added, for in this case the carbon dioxide exerts an action upon the methyl orange. The disturbing action of carbon dioxide is best prevented by first titrating in the cold, then heating to remove the carbon dioxide, again titrating the cold solution with acid. If only a small amount of carbonate is present, it exerts no appreciable effect upon methyl orange. The titration of oxalic acid with alkali which contains carbonate is best effected with phenolphthalein in hot solution. The process is carried out as follows: About 40 c.c. of the sodium hydroxide are accurately measured into a porcelain dish, a few drops of phenolphthalein added, and oxalic acid run in from a burette until the solution is decolorized. The solution is then heated upon the water bath until the red color reappears, whereupon it is decolorized by oxalic acid * With phenolphthalein the titration can be finished in the hot solu- tion, but the end-point is not so sharp. NORMAL SODIUM HYDROXIDE SOLUTION. 555 and the process continued until finally the color does not reappear on heating the solution. This point is reached, however, only after the solution has been evaporated to dryness and the residue taken up with a few cubic centimeters of distilled water. A slight red color will appear after this first evaporation, but it will be discharged by the fraction of a drop of oxalic acid and will not reappear upon -a second evaporation. Remark. Formerly the author was accustomed to heat the oxalic acid solution over a free flame, but since Christie has found in this laboratory that it was impossible to reach an end point in this way, the use of a free flame has been avoided. Sorensen met with the same difficulty and attributed the reappearance of the red color to the following reaction having taken place: 2Na 2 C 2 O 4 + H 2 O = Na 2 CO 3 + 2HCOONa + CO 2 Sodium formate It seems more probable, however, that this reappearance of the red color after the alkali is all neutralized is not due to the decomposition of sodium oxalate in the solution but to its being overheated on the sides of the dish, whereby it is decomposed into sodium carbonate and carbon monoxide: Na 2 C 2 O 4 - Na 2 CO 3 + CO Such a decomposition does not occur when the heating takes place upon a water bath. Preparation of Sodium Hydroxide Solution Free from Carbonate. This is best effected as proposed by Kiister.* About 40 c.c. of pure alcohol is placed in a small round-bottomed flask, heated to boiling on the water-bath, and little by little 2.5 gms. of bright metallic sodium are added, the latter being freed from petroleum * Zeit. f. anorg. Chem., 13, 134. 556 VOLUMETRIC ANALYSIS. rubbing between pieces of blotting-paper. The reaction be- tween the boiling alcohol and the sodium is at first very violent and large amounts of hydrogen and alcohol vapors are evolved. Dur- ing this time the flask is, there- fore, kept covered with a watch- glass. Gradually the reaction be- gins to diminish and finally stops. In the flask there will be a deposit of sodium alcoholate and some undissolved sodium on account of the insufficient amount of alcohol. Small amounts of water free from carbon dioxide * are now added, a test-tube full at a time. The alcohol is almost all boiled away, and in order to completely remove it, a current of air free from car- bon dioxide is passed through the solution until the odor of alco- hol can no longer be detected. The solution is then quickly cooled by the addition of water free from carbon dioxide, imme- diately placed in a liter flask, and diluted to the mark with pure water at 17-18 C. This solution will give the same value when titrated with phenolphthalein in a cold solution as when the latter is hot.f With methyl orange correct results are also obtained if the orange color is taken as the end-point. Such a solution quickly absorbs carbon dioxide from the air. Tn order to prevent this, it is placed in a bottle as shown in Fig. * This is accomplished by boiling the water while a current of air free from carbon dioxide is passed through it. t Provided the hydrochloric acid solution was prepared with water free /rom carbonate, otherwise too little acid will be necessary when the titration takes place in the cold. FIG. 87. BARIUM HYDROXIDE SOLUTION. 557 87 which is connected with a soda-lime tube, N, and with the burette by means of the tubes p and r. The burette is filled by squeezing the tube at a. In this way a solution can be kept free from carbon dioxide for a long time. In order to determine whether the solution is free from carbonate, two parallel titrations are made with phenolphthalein as an indicator, one in the cold and the other in the hot solution. If the results agree the solution is free from carbonate. Otherwise it is necessary either to prepare a fresh solution or to make a corresponding correction in each analysis after determining the amount of carbonate present as described on p. 563. In many cases it is better to use a ^ normal barium hydrox- ide solution; as long as it remains clear it is free from carbonate. ivr Preparation of Barium Hydroxide Solution. Ba(OH),+ 8H 2 O 315.51 1000 c.c. contain ~ - = =15.776 gms. The crystallized barium hydroxide of commerce always con- tains barium carbonate, so that the solution cannot be prepared by simply weighing out the necessary amount and diluting to 1 liter. About 20 gms. of the commercial product are dissolved in the necessary amount of distilled water within a large flask. The flask is closed and shaken until the crystals have completely dis- appeared and a light, insoluble powder of barium carbonate remains. The solution is allowed to stand for two days, until the barium carbonate has completely settled, when it is siphoned into a bottle through which a current of air free from carbon dioxide has been passed for two hours previous, after which the bottle is con- nected with a soda-lime tube and with the burette as shown in N Tig. 87. For the titration, 50 c.c. hydrochloric acid are placed in an Erlenmeyer flask, a little phenolphthalein is added, and the solution titrated by the addition of the barium hydroxide solution. The normality found should be written upon the label. N It is not advisable to make the solution exactly . for it usu- allv becomes turbid on dilution. 55 8 VOLUMETRIC ANALYSIS. A. ALKALIMETRY. I. Determination of Alkali Hydroxides. Rule. If the substance to be analyzed is a solid, an accurately weighed amount is dissolved in enough watei so that the solution is at about the same concentration as that of the. acid to be used in the titration. If, on the other "hand, a solution of an alkali hydroxide in water is to be analyzed, the specific gravity of the solution is determined by weighing in. a pycnometei or by means of an areometer, and then diluted 'to the amount desired. (a) Determination of Sodium Hydroxide in Commercial Caustic. Soda, NaOH= 40.01. N For the titration a hydrochloric acid solution can be N used. Consequently in this case an approximately normal solution of the alkali is prepared. As sodium hydroxide absorbs water and carbon dioxide from the an\ the sample for analysis is weighed out in a tared watch-glass and dissolved in water to a definite volume. After thoroughly mixing the solution a pipetted portion is treated with methyl orange and titrated in the cold with N hydrochloric acid. Example. 4.6623 gms. sodium hydroxide were dissolved in 1000 c.c. of solution and 25 c.c. of the latter, corresponding to N 0.11656 gm. sodium hydroxide, required 28.66 c.c. hydro- chloric acid for neutralization. N Since 1000 c.c. of acid correspond to 4.001 gms. NaOH, N 4 001 it is evident that 1 c.c. ~r acid =T^ =0.004006 gm. NaOH, and lU N 28.66 c.c. acid correspond to 0.004001 X 28.66 = 0.1 147 gm. NaOH. ALKALIMETRY. 559 This amount of NaOH was contained in 25 c.c. of solution, equivalent to 0.1166 gm. of the solid substance, so that the per cent, of sodium hydroxide present can be calculated: 0.1166 :0.1147=100:z 11.48 = 98.38 per cent. NaOH. 0.1166 (6) Determination of Sodium Hydroxide Present in Caustic Soda Solution. N For the titration assume that a solution is at hand; 1000 c.c. =^5=20.00 gms. NaOH. The alkali solution to be analyzed has a specific gravity of 1.285 at 15 C. ; and by consulting the table (see the supplement) we find that the solution should contain 25.80 per cent. NaOH by weight; i.e., 100 gms. of the solution should contain 25.80 gms. NaOH. Usually instead of weighing out the solution it is meas- ured and the per cent, by volume is computed. As 128.5 gms. of the alkali occupy a volume of 100 c.c. we have 100:25.8- 128.5 :x z = 33.153 gms. NaOH in 100 c.c. N Now as 1 liter of sodium hydroxide contains 20.00 gms. NaOH, ja we can compute how much of the alkali must be taken to be di- N luted to 1000 c.c. in order to make a solution: 100:33.153 = x : 20.00 We measure, therefore, 60 c.c. into a liter flask, dilute with water just to the mark, shake thoroughly, and by means of a pi- petle 25 c.c. of the solution are removed and titrated with half- normal acid, using methyl orange as an indicator. 560 yOLU METRIC ANALYSIS. N Assume that 25 c.c. of the solution require 24.3 c.c. of acid for neutralization. As 1000 c.c. ^ acid contain - =20.0 gms. NaOH, it is evident that 1 c.c. of the acid corresponds N to 0.02000 gm. NaOH, and 24.3 c.c. acid is equivalent to 0.02000X24.3 = 0.4860' gm. NaOH.. 25 c.c. of the dilute alkali, therefore, contain 0.4860 gm. NaOH, and 1000 c.c. of the dilute solution, or 60 c.c. of the original alkali, contain 0.4860X40=19.44 gms. NaOH, and 100 c.c. of the original solution contain 60 : 19.44= 100 :x 32.40 gms. NaOH. In order to obtain the per cent, by weight, this number must be divided by the specific gravity. In the assumed case we have: = 25.21 per cent. NaOH. Remark. The titration of alkali hydroxides with methyl orange as an indicator will only give correct results when the alkali hydroxide is free from carbonate, which with commercial material is never the case. The above results are too high, for they represent the total amount of alkali, i.e. the amount of NaOH+Na 2 CO 3 , though the latter is expressed in terms of NaOH. For an accurate determination of alkali hydroxide in the presence of alkali carbonate, see p. 563. (c) Determination of Ammonia in Aqueous Ammonia. The procedure is the same as under (6). (d) Determination of Ammonia in Ammonium Salts. A weighed amount of the ammonium salt is placed in the flask K (Fig. 24, p. 59),* dissolved in about 200 c.c. of water, and * Or better, the apparatus shown in Fig. 78, p. 454, may be used. TITRATION OF PYRIDINE BASES. 561 treated with 10 c.c. of a boiled solution of 10 per cent, caustic soda. The solution is distilled, and the distillate received in a known amount of normal acid in the receiver V, as described on p. 59. The excess of acid is titrated with normal alkali, using methyl orange as an indicator and the ammonia calculated from the difference between the amount of acid now found and that originally in the receiver. Example. The amount of ammonia in a sample of commer- cial ammonium sulphate is to be determined. As the technical product is never entirely pure, a large amount of the substance is weighed out, and for the sake of convenience this can amount to the gram-equivalent of ammonia, i.e. 17.03 gms. This quantity of the salt is dissolved in 500 c.c. of distilled water, and for the analy- sis, 50 c.c. of this solution are taken (1.703 gms. of salt). This is placed in the flask K (Fig. 23, p. 59), diluted with 150 c.c. of water, and distilled after the addition of 10 c.c. of 10 per cent, caustic soda solution. The distillate is received in 60 c.c. of half-normal hydrochloric acid, the excess of the latter titrated with half-normal alkali, and from the difference the amount of ammonia calculated. N For the titration, t c.c. of alkali are necessary; consequently the amount of ammonia in 1.703 gms. of the substance neutral- N ized 60 -t c.c. -^ acid. This corresponds to (60 t) X 0.008515 gm. NH 3 and in per cent. 1.703:(60-0x0.008515 = 100:z (60-00.8515 60- 1 X = 17Q3 ~2~ = per ' 3 * (e) Titration of Pyridine Bases. Method of K. E. Schulze.* 1000 c.c. N. acid = C 5 H 6 N = 79.05 gms. pyridine. The pyridine bases are so weak that they cannot be titrated with ordinary indicators. If, however, an aqueous pyridine solu- tion is treated with an aqueous solution of ferric chloride, the iron is precipitated as ferric hydroxide: FeCl 3 + 3QH 5 N + 3HOH = 3 (C 6 H 5 N, HC1) + Fe(OH) r * Berichte, 20 (1887), p. 3391. 562 VOLUMETRIC A 'NA LYSIS. If normal sulphuric acid is very carefully added with constant stirring until the precipitate redissolves, each cubic centimetre of the acid required will correspond to - =0.07905 gm. pyridine. lUUU Procedure. 5 c.c. of pyridine are dissolved in 100 c.c. of water, 25 c.c. of the resulting solution are treated with 1 c.c. of 5 per cent, aqueous ferric chloride solution, and the precipitate of reddish- brown ferric hydroxide is titrated ^vith normal sulphuric acid until completely dissolved. 2. Determination of Alkali Carbonates. Alkali carbonates can be titrated in the cold by using methyl orange as an indicator, the end-point being taken as the change from yellow into reddish orange. When fifth-, half-, and normal acids are used this is the correct end-point, but with tenth-normal acids this change is obtained a little too soon, for large amounts of carbonic acid exert a slight action upon the indicator. In this case the difficulty is best overcome by titrating the solution until the orange color is obtained, then heating to boiling to expel the car- bon dioxide, cooling, and again titrating until the now yellow solution becomes orange again.* With phenolphthalein, accurate results may be obtained by titrating the hot solution (cf. p. 554). According to Warder,f sodium bicarbonate solution reacts neu- tral toward phenolphthalein in the cold, so that when a sample of sodium carbonate is titrated in the cold, with phenolphthalein as an indicator, an end-point is obtained when the carbonate is changed to bicarbonate: Na 2 CO 3 + HC1 - NaCl + NaHC0 5 .J If the acid is allowed to run upon the carbonate solution, a part of the carbon dioxide from the sodium bicarbonate is lost, so that too much acid must be added before the end-point is reached. On the other hand, correct results may be obtained if * Kiister recommends in titrating carbonates with methyl orange, that a blank experiment be made to see how much effect an equal amount of water saturated with carbon dioxide has upon the same amount of indicator solution. (Zeitschr. fur anorg. Chem., XIII, p. 140.) t Zeitschr. f. analyt. Ch., 21, p. 102. j Zeitschr. f. anorg. Ch., XIII, p. 140. DETERMINATION OF ALKALI CARBONATES. 563 the titration is carried out at in the presence of NaCl (cf. p. 547). This is important, for in this way a convenient method is obtained for determining the amount of hydroxide in the presence of carbonate. 3. Determination of Alkali Carbonate and Hydroxide in the Presence of one Another. (a) Method of C. Winkler. Of the many methods which have been proposed for this deter- mination that of Winkler is the best. In one portion the total amount of alkali present is determined by titration with acid, using methyl orange as an indicator, and the hydroxide in a second portion is determined as follows: The solution is treated with barium chloride solution, when the follow- ing reaction take place : Na 2 CO 3 + BaCl 2 = 2NaCl + BaCO 3 (insoluble) . 2NaOH + BaCl 2 = 2NaCl + Ba (OH) 2 (soluble) . The sodium of the carbonate is transformed into neutral sodium chloride, while insoluble barium carbonate is precipitated from the solution; the sodium hydroxide, however, yields an. equivalent amount of barium hydroxide. If the solution contain- ing phenolphthaleiin is slowly titrated with hydrochloric acid with constant stirring, decolorization is effected as soon as the hydroxide is neutralized. The amount of acid used corresponds to the amount of hydroxide originally present. Example: 1. 20 c.c. (Na 2 CO 3 + NaOH) require T c.c. acid for Na^Og- 2. 20c.c.(Na 2 C0 3 + NaOH) " t " ". " NaOH alone, so that 20 c.c. (Na 2 C0 3 +NaOH) require T-t c.c. ^ acid for Na 2 CO 8 ; 20 c.c. of the solution, therefore, contain (a)*X0.004001 gm. NaOH, (6) (T-t) X0.005300 gm. Na 2 CO 3 . Remark. It has been proposed to add an excess of barium chloride solution to the mixture of alkali carbonate and hydroxide 564 1/OLUMETRIC ANALYSIS. contained in a measuring-flask, then dilute to the mark, thoroughly mix, and filter through a dry filter; for the titration an aliquot part of the nitrate is taken. This method, however, will only give accurate results when the water used for the dilution is absolutely free from carbon dioxide, and this will be the case only when it is previously boiled with a current of air free from carbon dioxide passing through it. Further, no attention is paid to the volume occupied by the precipitated barmm carbonate, and in the case of a large amount of the latter, a considerable error is introduced. The method of Winkler does not have these disadvantages. Care must be taken, however, with regard to the addition of the hydro- chloric acid in the titration; unless it is added very slowly some of the barium carbonate will be acted upon before the end-point is reached. (b) Method of R. B. Warder. To the cold * solution containing phenolphthalei'n, hydro- chloric acid is added and the liquid is gently stirred with a glass rod. Decolorization takes place when all of the hydroxide and half of the carbonate are neutralized: NaOH + HC1 = NaCl+ H 2 O, Na 2 C0 3 + HC1 = NaCl+ NaHCO 3 . To the colorless solution, methyl orange is added, and the solution is again titrated with acid until the other half of the car- bonate is neutralized, when the solution turns brownish-red. If the amount of acid used for the titration with phenolphtha- lein is represented by T, and that necessary for the titration with methyl orange by -t, then 2t c.c. corresponds to the amount of carbonate present, and T t represents the amount of hydroxide. 4. Determination of Alkali Bicarbonates. The solution is titrated in the cold unuij an orarge color is obtained with methyl orange, or untL a tciorless solution is obtained by titrating hot with phenolphthalei'n. (See page 553.) * The results are accurate only when the solution is at and NaCl is present. Cf. Kuster. Zeitschr. f. anorg. Chem., XIII, p. 134 (1897). DETERMINATION OF ALKALI CARBONATES. 565 5. Determination of Alkali Carbonates in the Presence of Alkali Bicarbonates. (a) Method of C. Winkler. The total alkali is determined in one portion by titration with hydrochloric acid, using methyl orange as an indicator, and in a second portion the amount of bicarbonate is determined as follows : A definite volume of the solution is treated with an excess of sodium hydroxide, by which means the bicarbonate is changed to neutral carbonate: NaHCO 3 + NaOH = Na 2 CO 3 + H 2 O. The solution now contains sodium carbonate with the excess of sodium hydroxide, and the amount of the latter is determined as described under 3. In other words, barium chloride is added, then phenolphthalein, and the solution is titrated until colorless. The amount of acid now used corresponds to the excess of sodium hydroxide added, and if this amount is deducted from the total sodium hydroxide, the corresponding amount of bicarbonate will be obtained. Example: 25 c.c. Na 2 CO 3 +NaHCO 3 required T c.c. ~ acid for Na 2 CO 3 +NaHCO 3 ; 25 c.c. Na.COa+NaHCOg+T 7 ! c.c. ^ NaOH+BaCl a N required t c.c. acid for the excess of NaOH; N 25 c.c. Na 2 CO 3 + NaHCO 3 , therefore, require T l - t c.c. ~ acid for the NaHCO 3 and T - ( T, - 1) c.c. ^ acid for the Na 3 CO s . 25 c.c. of the original solution contain (a) (T,-t) X 0.008401 gm. NaHCO 3 , (6) (T-Ti+t) X 0.005300 gm. 566 VOLUMETRIC A 'NA LYSIS. Remark. In order to make sure that enough sodium hydrox- N ide solution is present, the same amount of the alkali is added N as there were cubic centimeters used of acid in the deter- mination of the total alkali ; in this case, then, T = T v and t y the excess of alkali, corresponds at the same time to the amount of Na 2 C0 3 present. The caustic alkali solutions, even when origi- nally free from carbonate, gradually absorb it from the air, so that in every case the amount of carbonate in the alkali should be deter- mined before making the above analysis and a corresponding correction applied to the calculation. (b) Method of Warder* Using phenolphthale'in as indicator, the solution is titrated with hydrochloric acid until colorless, and in this way half of the carbonate is determined. Methyl orange is then added and the solution titrated until a brownish-red color is obtained, and in this way the total amount of the bicarbonate together with half of the carbonate is determined. If t represents the amount of acid used in the first titration, and T the total amount used, then: 2t c.c. of acid correspond to the amount of carbonate and (T 2t) c.c. correspond to the bicarbonate. 6. Determination of Alkaline-earth Hydroxides. The solution containing phenolphthale'in is titrated until color- 7. Determination of Alkaline-earth Carbonates. The carbonate is dissolved in an excess of the standard acid, boiled to remove the carbon dioxide, and the excess of acid titrated with alkali, using methyl orange as indicator in cold solution. * Cf. Am. Ch. Journ., 3, No. 1, and Chem. News, 43, 228. DETERMINATION OF ALKALINE-EARTH OXIDE. 567 8. Determination of Alkaline-Earth Oxide together with Alkaline-Earth Carbonate. Suppose, for example, it is desired to determine the amount of oxide and carbonate in a sample of " quicklime." The lime is broken up into pieces about the size of a pea, exactly 14 gms. are accurately weighed out and slaked with boiled water, the paste is washed into a 500-c.c. flask and diluted to the mark with water free from carbon dioxide. After thoroughly mixing, 50 c.c. of the turbid liquid is transferred to a second 500-c.c. flask and again diluted to the mark. Determination of the Total Calcium. 50 c.c. (0.14 gm. of sub- N stance) of the last solution are treated with 60 c.c. of hydro- chloric acid and heated until there is no further evolution of car- bon dioxide, the solution is cooled, and the excess of the acid N titrated with caustic soda solution, using methyl orange as an indicator. For this purpose t c.c. of the latter are required; N consequently 60 c.c. acid were necessary to neutralize the calcium hydroxide and calcium carbonate in the 50 c.c. of the solution taken for analysis. Determination of the Calcium Oxide. A second portion of the N freshly-shaken solution is titrated with ^ hydrochloric acid added drop by drop to the cold solution, using phenolphthalein as an indicator. Assume that ^ c.c. of the acid were necessary to neutralize the calcium oxide. Consequently, for the neutralization of the CaCO 3 +CaO = N N 60 t c.c. acid were required, and for the CaO, ^ c.c. acid were necessary. For the neutralization of the CaCO 3 , therefore, N 60 (2 -MJ) c.c. acid were necessary. 50 c.c. solution (0.14 gm. lime) contain: (a) and from the difference the total SO 3 present is computecL . With regard to the SO 2 , the following reactions take place during the titratjons: H 2 S0 3 + NaOH = NaHSO 3 + H 2 O *G. Finch (Chem, ' Ztg., 1916, 297) and R. H.' Vernon (ibid., 1910, 702) use a different apparatus and larger samples, thus getting more accurate results. 57 8 VOLUMETRIC ANALYSIS. It is to be noted in the first reaction that, although sulphurous acid is a dibasic acid, the end-point is reached, with methyl orange as an indicator, when the first hydrogen atom has been neutralized. From the two equations, then, it is evident that 2 gm. atoms of N iodine are equivalent to 1 gm. molecule of NaOH, or 1 c.c. ^ iodine N solution is equivalent to J c.c. sedium hydroxide. N N Since, in general, 5 c.c. -^ solution= 1 c.c. -5- solution, then N N 1 c.c. TTT solution =-J c.c. solution, ID 2i N N and in the given case 1 c.c. iodine= T Vc.c. sodium hydroxide; N so that if T c.c. ~ alkali were used in the first titration of the 2i N total acid present, and t c.c. of =/* iodine solution for the oxidation of the sulphurous acid, it is plain that T represents the amount of alkali necessary for the neutralization of the total sulphuric acid. The S0 3 is determined by an indirect analysis. We will assume that the fuming acid consisted of H 2 S0 4 =* S0 3 =2/ S0=a 100 then 100-a=a;+2/. In order to determine x and y a second equation is necessary, and this is found from the titration of the total sulphuric acid. Assume that, after the deduction corresponding to the amount of S0 2 has been made, the total amount of H 2 S0 4 was found to be p per cent., then: 1. x+y *=100-a 2. x+my= p p+a 100 ~~ y=^ i =per cent. SO, ACID CONTENTS OF FUMING ACIDS. S79 24 . In equation 2, m= gQ - = =1.2250 z=100-(a-f?/) = per cent. H 2 SO 4 H 2 S0 4 98.086 and Example* 3.5562 gms. of fuming acid were diluted to 500 c.Co, and of this amount 100 c.c. = 0.7112 gm. were taken for analysis. N 1. 100 c.c. required 5.40 c.c. iodine = 5.4 X 0.003203 = 0.01730 gm. S0 2 =2.43 per cent. S0 2 =a. N 2. 100 c.c. required 34.40 c.c. sodium hydroxide. From the latter must be deducted 0.54 c.c. to correspond to the amount of alkali necessary for the S0 2 . 34.40 - 0.54 = 33.86 c.c. 33.86X0.02452=0.8305 gm. H 2 SO 4 = 116.7 per cent.p. If these values are introduced in the above equations we obtain 119.16-100 19.16 y== 0.2250 = 02250 = 85 - 15 per C6nt ' S 3 and x= 100 -(85. 15 + 2.43) = 12.42 per cent. H 2 S0 4 . The acid contains, therefore: H 2 SO 4 = 1.2.42 f SO 3 = 85.15 SO 2 = 2.43 100.00 * Lunge, Zeitschr. f. angew. Ch., 1895, p. 221. t Like all indirect analyses, the results obtained are not absolutely accu- fate. Almost all fuming acids contain solid constituents which are neg- lected in the above calculation. It would be more accurate to determine the amount of the latter in a separate portion, by weighing the residue OD ignition. 580 VOLUMETRIC ANALYSIS. Preparation of Concentrated Sulphuric Acid Mixtures (M. Gerster.) It is often necessary to prepare fuming sulphuric acid of definite concentration. Given : (a) Fuming sulphuric acid (A) with a per cent, free SO 3 . (6) Sulphuric acid (B) with c per cent. H 2 SO 4 and 100 c per cent, water. A fuming acid containing 6 per cent, free SO 3 is desired. To obtain the latter, 100 gms. of the acid A are mixed with x gms. of the acid B. It must be remembered, however, that the water in the acid B requires SO 3 in order to form 100 per cent. H 2 SO 4 : H 2 0+S0 3 =H 2 S0 4 . The acid B requires for the water present in, each 100 gms H 2 O:SO 3 =(100-c):2/ If 100 gms. of the acid B require 4.44 (100 c) gms. SO 3 from A> then x gms. of the acid B require 0.0444 (100 c) x gms. SO 8 from A. Now A + B (100 + z): [a- 0.0444 ( 100 -c)z] = 100:6 100 (a- 6) ^=777 ~T A AA ' 444+6 4.44 c Example. The fuming acid A contains 25.5 per cent, free S0 3 =a. Sulphuric acid B contains 98.2 per cent. H 2 S0 4 =c. The acid desired is to contain 19.0 per cent. 80^=6. If these values are inserted in the above equations, we obtain 100(25.5-19.0) 650 TITRATION OF HYDROXYLAMINE SALTS, ETC. 581 We must add, therefore, 24.07 gms. of the 98.2 per cent, sulphuric acid to 100 gms. of the fuming acid in order to obtain an acid containing 19.0 per cent, of free S0 3 . Titration of Hydroxylamine Salts. Hydroxylamine hydrochloride reacts neutral towards methyl orange and acid towards phenolphthalein. If the latter is added to an aqueous solution of the salt, and the titration is made with N alkali, the end-point will be obtained when the total amount of acid present has been neutralized by the alkali. It is impossible to determine the amount of free hydrochloric acid present when phenolphthalein is used, but it can be done with methyl orange. Romijn * recommends for the titration of the acid a r^r borax solution. Hydrofluoric Acid. 1000 c.c. normal alkali =HF =20.01 gms. HF, Hydrofluoric acid can be titrated with phenolphthalein as an indicator, but not with litmus or methyl orange. The acid is measured out into a platinum dish by means of a pipette which is coated with beeswax, an excess of sodium hydroxide free from alkali is added, and the excess of the latter is titrated in hot solution with an acid of known strength.f Hydrofluosilicic Acid. The titration of this acid may take place 'according to either of the following reactions : I. H 2 SiF 6 + 2KOH = K 2 SiF 6 + 2H 2 O, II . H 2 SiF 6 + 6KOH = 6KF + 2H 2 O + Si (OH) 4 . * Z. anal. Chem., 36 (1897), 19. This method has not been tested in the author's laboratory. t Cf. Winteler, Z. angew. Chem., 1902, p. 33. 582 yOLU METRIC ANALYSIS. According to Equation I. 1000 c.c. normal KOH or Ba(OH) 2 = 72.16 gms. H 2 SiF 6 . Author's Method. If hydrofluosilicic acid is titrated in the cold with caustic potash, using phenolphthalein as indicator, a red color appears after a time, but disappears later on account of the excess alkali reacting according to the equation: K 2 SiF 6 +4KOH=6KF + Si(OH) 4 . This last reaction, however, takes place so slowly that it is impossible to obtain a distinct end point. If, however, the solution is diluted with an equal volume of alcohol, then 2 or 3 drops of phenolphthalein added, it can be titrated with tenth- normal potassium or barium hydroxide. The insoluble potassium or barium fmosilicate separates out and is not acted upon by an excess of the alkali, so that a sharp end point is obtained. Sodium hydroxide forms a soluble salt so that the titration cannot be made with this reagent. Indirect Method of Penfield* Penfield treats the solution to be titrated with an excess of KC1, dilutes with an equal volume of alcohol, and then titrates the hydrochloric acid set free in the reaction, H 2 SiF 6 +2KCl= K 2 SiF 6 + 2HCl, with tenth-normal sodium hydroxide solution, using cochineal as indicator. Methyl red is preferable to the cochineal. According to Equation II. 1000 c.c. normal NaOH= 24.05 gm. H 2 SiF 6 . (a) Method of Sahlbom and Hinrichsen.-f The solution is titrated at the temperature of the water-bath with tenth-normal sodium hydroxide solution, using phenol- phthalein as indicator. * Chem. News, 39, 179. t Ber., 39, 2609 (1906). DETERMINATION OF ORGANIC ACIDS. 583 (b) Method of Schucht and Mcller* The solution to be titrated is treated with an excess of neutral calcium chloride solution (25 c.c. of 4N. CaCl 2 ) and titrated with tenth-normal sodium hydroxide, using methyl orange as indi- cator. The following reaction takes place in the cold: H 2 SiF 6 + 3CaCl 2 + GNaOH = 3CaF 2 + 6NaCl + Si (OH) 4 + 2H 2 O . During the titration the solution remains clear, for the CaF 2 and the Si(OH) 4 remain in colloidal solution. Phenolphthalein should not be used as indicator, as it is hard to decide upon the correct end point. In the titration of salts of hydrofluosilicic acid, however, the titration must always be carried out with phenolphthalein as indicator: Na a SiF 6 + 3CaCl 2 + 4NaOH = 3CaF 2 + GNaCl + Si (OH) 4 . In this case 1000 c.c. of normal NaOH = 47.08 gms. of Na 2 SiF 6 . Determination of Organic Acids. Methyl orange cannot be used for the titration of organic acids, but either phenolphthalein or litmus may be employed. If car- bonic acid is present at the same time, the titration is made in a hot solution (cf. p. 554). It is best to dilute the organic acid with water free from carbon dioxide, add phenolphthalein, and titrate with half-normal barium hydroxide in the cold. To illustrate. It is desired to analyze a sample of acetic anhy- dride. The only impurity that the distilled product is likely to contain is acetic acid, so that it is a question of determining the amount of acid and anhydride in the presence of one another. Such a problem can be solved only by an indirect analysis. The mixture is weighed out in a small glass bulb and then thrown into an accurately-measured amount of standard barium hydroxide solution. The latter is contained in a flask which is * Ber., 39, 3693. This method has not been tested by the author. 584 VOLUMETRIC A NA LYSIS. connected with a return-flow condenser and at the top of the condenser a soda-lime tube is fitted. The contents of the flask are warmed gently until the oil has completely dissolved; it is thereby changed to acetic acid, and the latter is neutralised by thfj alkali. After the reaction is complete, a drop of phenolphthalein is r.dded and the solution is decolorized by the addition of a titrated acid. From the amount of the latter used, tho excess of the alkali is known, and if this is deducted from the total amount of alkali in the flask, the amount necessary for the complete neutralization of the acetic acid, whether originally present as the free acid or in the form of its anhydride, can be calculated: 4 C 4 H G 3 C 2 H 4 2 1. x + y = p (original weight) ; 2. mx + y = q (weight acetic acid after the action of water); and from this x can be calculated. 2C 2 H 4 O 2 120.06 and in these equations m= ~ u . = 1AO n _ = 1 . 17bo and L^rieUa LUZ.UO - = 5.665. ra 1 Example. The absolutely clear preparation of acetic anhy- dride from a well-known firm gave the following results, 0.9665 gm. being taken for the analysis: N 200 c.c. of barium hydroxide solution required 187.79 c.c. HC1; 200 c.c. of barium hydroxide -h 0.9665 gm. of substance N required 6.03 c.c. HC1; so that the 0.9665 gm. of substance was equivalent to 181.76 c.c. DETERMINATION OF ORGANIC ACIDS. 585 N N HC1, and this amount of Ba(OH) 2 solution would have been required to neutralize it. This corresponds to 181.76X0.006003 = 1.0911 gms. acetic acid=g. If, now the values of p and q are introduced in the previous equations, we obtain z = 5.665(1.0911-0.9665) =0.7059 gm. anhydride, and in per cent. 0.9665: 0.7059 =100: x # = 73.04 per cent, acetic anhydride. The preparation, therefore, contained Acetic anhydride = 73.04 per cent. 26.96 per cent. Acetic acid 100.00 per cent. Remark. Acetic acid anhydride is also hydrolyzed by water at the ordinary temperature. If a weighed amount of the sub- stance is shaken with water in a flask until no more drops of anhydride are to be recognized, and the acetic acid formed is then titrated with barium hydroxide, using phenolphthalem as indicator, correct results are obtained if the water used is entirely free from carbon dioxide. It is always safer, however, to carry- out the determination as outlined above. In some factories the analysis of acetic acid anhydride is carried out by the method of Menschutkin and Wasiljeff. This is based upon the fact that when acetic acid anhydride is treated with freshly distilled aniline, acetanilide is formed in accordance with the following equation. pTT pA\ CH ' CO/ + C6HsNH2 = CeH5N (2 H 30) H + CH 3 COOH 586 VOLUMETRIC ANALYSIS. whereas acetic acid itself does not form acetanilide under the same conditions. Two or three grams of commercial acetic anhydride are shaken in a dry weighing beaker with from 4 to 6 c.c. of freshly distilled aniline. The anhydride immediately begins to combine with the aniline, liberating considerable heat. After cooling, the solidified contents of the weighing beaker are rinsed by means of absolute alcohol into an ordinary beaker, phenolphthalein is added and the total amount of acetic acid present titrated with half-normal alkali. We have then x + y = p- } mx + y = q (acetic acid) ; from which can be computed z=^i = 2.428. 1 n.. In this equation ^C 2 H 4 O2 = 60.03 It is true that concordant results are obtained by this method, but they are much too high; in fact as much as 14 to 16 per cent, too high. This is due to the fact that although acetic acid itself does not react with aniline in the cold, it does react very readily when heated. When, therefore, a mixture of acetic anhydride and acetic acid are allowed to remain in contact with aniline, there is so much heat liberated from the reaction between the anhydride and the analine that a part of the acetic acid itself reacts and takes part in the formation of acetanilide : so that evidently too little acetic acid is found in the subsequent DETERMINATION OF SULPHUROUS ACID. 587 titration and consequently too high values are obtained for the amount of anhydride present. Determination of Sulphurous Acid. For the determination of sulphurous acid by itself, the analy- sis is always accomplished, as recommended by Volhard, by an iodimetric process, i.e., it is oxidized to sulphuric acid. In many cases, however, it is necessary to titrate the sulphurous acid with alkcli (cf. p. 577), and here the choice of an indicator is important. for the end-point is very different in the case of methyl orange from that obtained when phenolphthalein is used : H 2 SO 3 + 2NaOH = Na 2 SO 3 + 2H 2 O (with phenolphthalein) , H 2 SO 3 + NaOH = NaHSO 3 + H 2 O (with methyl orange). NaHSO 3 reacts acid toward phenolphthalein, but neutral toward methyl orange, so that twice as much alkai would be added in the first case. The most accurate results are obtained with the use of methyl orange, for the carbon dioxide which is almost always present does not exert much of an effect upon this indicator, whereas it does upon phenolphthalein. Determination of Orthophosphoric Acid. NaH 2 P0 4 reacts acid toward phenolphthalein, and neutral toward methyl orange, while Na 2 HPO 4 is neutral toward the former indi- cator and basic toward the latter. Therefore, on titrating free phosphoric acid with alkali one of the following reactions will take place: = Na 2 HPO 4 + 2H 2 (phenolphthalein). 2. H 3 PO 4 + NaOH=NaH 2 POi+H 2 (methyl orange). The first reaction is not sharp, because pure Na 2 HPO 4 is disso- ciated to a slight extent, so that it becomes alkaline to phenol- phthalein : Na 2 HPO 4 + H OH<=*Na H 2 PO 4 + Na OH. 588 VOLUMETRIC ANA LYSIS. To prevent this hydrolysis, the titration is best effected in a cold, concentrated solution containing sodium chloride. Alkalimetric Determination of Phosphorus in Iron and Steel * 1000 c.c. normal NaOH = 1.348 gm. P. The phosphorus is precipitated in a 2 gm. sample with ammonium molybdate according to p. 437, filtered, washed with 1 per cent, nitric acid and then with 1 per cent, potassium nitrate solution until the washings no longer react acid. The filter and precipitate are then transferred back to the Erlenmeyer flask in which the precipitation took place, covered with an excess of tenth-normal NaOH (T c.c.), stirred until solution is complete, and then the excess of alkali titrated with tenth-normal nitric acid, (t c.c.) using phenolphthalein as indicator; The reactions taking place are as follows : 2[(NH 4 ) 3 P0 4 . 12Mo0 3 ] + 46NaOH = 2(NH 4 ) 2 HPO 4 + (NHJ 2 MoO 4 + 23Na 2 MoO 4 +22H 2 O, from which it is clear that 46 gms. mols. of NaOH are equivalent to 2 gm. atoms of P and 1 c.c. of the tenth-normal NaOH= 0.0001348 gm. P. Since the precipitate was produced from 2 gm. steel, and T c.c. of NaOH and t c.c. of HNOs were used, the percentage of phos- phorus is (T-t)X 0.01348 ^ p 2~ ~ fcf Remark. To obtain accurate results, it is advisable to deter- mine the percentage of phosphorus in a steel gravimetrically, then to standardize the alkali against this steel, carrying out the titration exactly as described above. Determination of Boric Acid. Free boric acid has no action upon methyl orange, conse- quently alkali borates may be titrated with hydrochloric and nitric acids, using this indicator; with sulphuric acid the results * See Blair, Analysis of Iron and Steel. The method was proposed by J. O. Handy. DETERMINATION OF BORIC ACID. 589 are not as satisfactory, for there is in this case no sharp color change. If phenolphthalein is used as the indicator, the red color fades gradually and the end point cannot be determined with certainty. If, on the other hand, sodium hydroxide is slowly run into an aqueous solution of boric acid containing phe- nolphthalein, after some time a pale-pink color is noticeable which becomes deeper on the addition of more alkali. Tha first pink color is formed before all of the boric acid has been neutralized, for sodium borate is perceptibly hydrolyzed. Free boric acid cannot be titrated by itself, but if, as proposed by Jorgensen,* a sufficient amount of glycerol f (or mannitolt) is added to the solution, the hydrolysis is prevented, so that when 1 mol. of NaOH is present for 1 mol. of H 3 BO 3 the solution suddenly changes from colorless to red; probably a stronger acid is formed by the addition of the glycerol, the glyceryl-boric acid (C 3 H 5 O 2 OH)B(OH). If the solution does not contain sufficient glycerol the color change takes place too soon, as can be shown by the addition of more glyoerol. If the red color disappears on adding the lat- ter, more alkali is added until it reappears. The right end-point is reached when the red color no longer disappears on the addition of glycerol. Inasmuch as commercial glycerol reacts acid, it must be just neutralized with alkali before being used for this determination. Furthermore, in order to obtain accurate results it is necessary that the solutions should be absolutely free from carbonate. Application. Determination of Boric Acid in an Alkali Borate Free from Carbonate.^ About 30 gms. of the borate are dissolved in water free from carbon dioxide, diluted to 1 liter, and the total alkali is determined in an aliquot part by titration with ~ hydrochloric acid, using methyl orange as an indicator. A fresh portion of the borate is * Zeitschr. f. Nahrungsm. IX, p. 389, and Zeitschr. f. angew. Ch., 1897, p. 5. t Zeitschr. f. angew. Ch., 1896, p. 549. j Jones, Am. J. Sci. [4] 7, 147 (1899). M. Honig and G. Spitz, Zeitschr. f. angew. Ch., 1896, p. 549. 590 VOLUMETRIC ANALYSIS. taken and exactly neutralized by the amount of hydrochloric acid found necessary by the previous titration; by this means the solution will contain free boric acid. After adding about 50 c.c. of glycerol for each 1.5 gms. of the borate, the solution is titrated N with sodium hydroxide, using phenolphthalei'n as indicator. 2i After the end-point is reached, 10 c.c. more of glycerol are added, and this usually causes the solution to become colorless. The end- point with sodium hydroxide is again obtained and the process repeated until finally the addition of glycerol causes no further action upon the end-point. If the borate contained carbonate, the portion taken for analysis is neutralized with acid as before, then boiled for a few minutes, taking the precaution of connecting the flask containing the solu- tion with a return-flow condenser.* After the carbon dioxide is expelled, the sides of the condenser are washed down with water and the titration with sodium hydroxide made as before.f For the Determination of Boric Acid in Insoluble Silicates. seeE. T. Wherry and W. H. Chapin, J. Am. Chem. Soc., 30, 1687 (1908). Determination of Carbonic Acid. (a) Determination of Free Carbonic Acid. To determine 'the amount of free carbonic acid present in a dilute aqueous solution, an excess of titrated barium hydroxide N solution is added, and the excess is determined by means of HC1, using phenolphthalei'n as an indicator: H 2 CO 3 + Ba(OH) 2 = BaCO 3 + 2H 2 O 1 c.c. ^ HC1 = 0.0022 gm. CO 2 . * The condenser serves to keep back 'any boric acid escaping with the steam. t Instead of the glycerol, about one gram of mannitol may be used to advantage. DETERMINATION OF CARBONIC ACID. 59 1 (6) Determination of Carbon Dioxide Present as Bicarlonate. N The solution is titrated with HC1 in the presence of methyl orange: NaHCO 3 + HC1 = NaCl+ H 2 CO 3 1 c.c. ^- HC1 = 0.0044 gm. C0 2 . (c) Determination of Carbon Dioxide Present as Carbonate. N The titration is effected with HC1 and methyl orange:* Na 2 C0 3 + 2HC1 = 2NaCl+ H 2 C0 3 1 c.c. HC1 = 0.0022 gm. C0 2 . (d) Determination of Free Carbonic Acid in the Presence of Bicarbonate. N One portion is titrated with HC1, using methyl orange as indicator, and the amount of bicarbonate is determined as under (6). A second portion is treated with an excess of barium chloride, f then with an excess of barium hydroxide, and the excess of the latter titrated back with HC1, using phenoiphthalei'n as indicator. N If the amount of acid used for the first titration is deducted N from the amount of barium hydroxide solution found to be necessary by the last titration. the difference multiplied by 0.0022 will give the amount of free carbonic acid.t * Alkaline-earth carbonates are dissolved in an excess of standard acid and the excess titrated back with standard alkali. f The addition of barium chloride is only necessary when free carbonic acid is titrated in the presence of alkali bicarbonates. Without it free alkali would then be formed: NaHCO 3 + Ba(OH) 2 = BaCO 3 + H 2 O + NaOH. J This method cannot be used when magnesium salts are present. 592 VOLUMETRIC A NA LYSIS. (e) Determination of Bicarbonate in the Presence of Carbonate. Method of C.. Winkler. In one portion the total alkalinity is determined by titration N with HC1, using methyl orange as indicator. This requires T c.c. of -| HC1. In a second portion the bicarbonate is determined by adding N an excess of -^ NaOH, then neutral barium chloride solution, and afterward titrating the excess of the former with phenol- N phthalein and HC1. We will assume that for this purpose N N TI c.c. NaOH and t c.c. HC1 were used, then evidently N (T l f) c.c. ^ NaOH were necessary to convert the bicarbonate into carbonate: NaHC0 3 + NaOH = Na 2 CO 8 + H 2 0. INaOH corresponds, consequently, to 1C0 2 , or 1 c.c. -^ NaOH = 0.0044 gm. CO 2 , and therefore (7\-0-0.0044 = CO 2 as bicarbonate. For the decomposition of the normal carbonate -T l ) c.c. j^- were necessary, and from the equation Na a CO, + 2HC1 = 2NaCl + H 2 O + CO, it is evident that 2HC1=1CO 2 and 1 c.c. -^ HC1= 0.0022 gm. CO 2 . The carbon dioxide as carbonate =( T+tTJ- 0.0022 gm. DETERMINATION OF CARBONIC ACID IN THE AIR. S93 Remark. It has been proposed to determine volumetrically the free and bicarbonate carbonic acid in drinking and mineral waters; with the former accurate results can be obtained, but with the latter this is not the case. In the determination of the total alkalinity not only the bicarbonate but also the ever-present silicate and borate are likewise determined, so that this in many cases causes considerable error in the analysis of mineral waters, Thus in analyzing a sample of mineral water containing in reality 4.63 gms. of carbonic acid as bicarbonate per kilogram, the titra- tion showed 5.42 gms., a difference of 0.61 gm, C0 2 1 Determination of Carbonic Acid in the Air. Method of Pettenkofer. Principle. A large, measured volume of air is treated with an excess of titrated barium hydroxide solution whereby the car- bon dioxide is quantitatively absorbed, forming insoluble barium carbonate. Phenolphthalem is added, and the excess of barium hydroxide is determined by titration with hydrochloric acid until the solution is colorless. From the amount of alkali used to absorb the carbon dioxide, the amount of the latter is calculated. Requirements. 1. A calibrated flask of 5 liters capacity. 2. Standard solutions of barium hydroxide and hydrochloric acid. The acid is prepared so that 1 c.c. = 0.25 c.c. CC>2 at C. N and 760 mm. pressure; this is accomplished by diluting 224.7 c.c. = hydrochloric acid to 1 liter. The barium hydroxide solution should be of about the same strength. Procedure. The flask, with its capacity etched upon it, is placed in the space from which the air is to be taken, and by means of a bellows, the mouth of which is connected with a piece of rubber tubing, the air in the flask is changed; about 100 strokes are made with the bellows. The flask is then stoppered with a rut> ber cap, and at the same time the temperature and barometer readings are noted. 594 VOLUMETRIC AN 'A LYSIS. By means of a pipette, 100 c.c. of barium hydroxide solution are run into the flask, the rubber cap replaced on the bottle, and the solution is gently shaken back and forth in the flask for fif- teen minutes. The turbid liquid is then poured into a dry flask, 25 c.c. are pipetted out, phenolphthalei'n is added, and hydrochloric acid slowly run in with constant stirring until the solution is colorless. This requires n c.c., so that for the 100 c.c. of alkali solution, 4X^ c.c. would be necessary. The strength of the barium hydroxide in terms of acid is now accurately determined; 25 c.c. of barium hydroxide require N c.c. of the standard hydro- chloric acid, or 100 c.c. would neutralize 4X^V c.c. of acid. Calculation. Assume the contents of the flask to be V c.c. at t C. and B mm. pressure. By the introduction of 100 c.c. barium hydroxide solution the same volume of air was replaced, so that the amount of air taken for analysis amounts to (V 100) c.c. at t C. and B nun, pressure. At C. and 760 mm. pressure the volume is r (7-100)5 100 c.c. of barium hydroxide solution require 4 N c.c. HC1, while 100 c.c. of the alkali after treatment with 7 c.c. of air require 4 n c.c. of the acid and this corresponds to 4 (N ri)'Q.25 = (N n) c.c. C0 2 at C. and 760 mm. pressure. The amount of C0 2 present in 1 liter of air measured at standard conditions amounts to *- - r - - gms. C0 2 ., PERSULPHURIC ACID. 595 Per sulphuric Acid. N 1000 ex. Potassium Hydroxide K 2 S 2 8 270.34 = on = 2fT~ = 13. 517 gms. K 2 S 2 O 8 . If an aqueous solution of either potassium, sodium, or barium persulphate is boiled for some time, the salt is decomposed in accordance with the equation: 2K 2 S 2 O 8 + 2H 2 O = 2K 2 SO 4 + 2H 2 SO 4 + O 2 into neutral sulphate and free sulphuric acid. The latter can be titrated with tenth-normal potassium hydroxide solution. Procedure. About 0.25 gm. of the persulphate is placed in an Erlenmeyer flask of Jena glass, dissolved in about 200 c.c. of water, and the solution boiled for twenty minutes. It is then cooled, methyl orange added, and the solution titrated with tenth-normal potassium hydroxide. Or, an excess of the alkali may be added and the amount of excess titrated with tenth-normal acid. The results correspond with those obtained by the ferrous sulphate method (cf. p. 629) provided the persulphate is not con- taminated with potassium bisulphate. Remark. Ammonium persulphate cannot be analyzed by the above method because when a solution of this salt is boiled, two reactions take place. The principal reaction, to be sure, is 2 (NH 4 ) 2 S 2 O 8 + 2H 2 O = 2 (NH 4 ) 2 SO 4 + 2H 2 S0 4 + O 2; but the oxygen is evolved to some extent in the form of ozone and the latter oxidizes a part of the nitrogen, so that besides sulphuric acid, the solution will contain more or less nitric acid. 8 (NH 4 ) 2 S 2 O 8 + 6H 2 = 7 (NH 4 ) 2 S0 4 + 9H 2 SO 4 + 2HNO 3 . II. OXIDATION AND REDUCTION METHODS. All processes considered under this heading are those in which the substance analyzed is either oxidized or reduced by means of the solution with which the titration is made. As a standard for measuring the normality, we Consider the oxidation of two gm.-atoms of hydrogen by 1 gm.-atom of oxygen. The normaj solution, therefore, will be one which for each 1000 c.c. gives up or requires -77 = 8 gms. of oxygen = 1 gm. of hydrogen. 2i OXIDATION METHODS. A. The Permanganate Methods. These are based upon the fact that 2 gm.-molecules of potas- sium permanganate in acid solution give up 5 gm.-atoms of oxygen, equivalent to 10 gm.-atoms of hydrogen: 2KMnO 4 = K 2 O+ 2MnO+ O 5 ( = 10H). Or, what amounts to the same thing, each atom of manganese is reduced from a valence of 7 to a valence of 2, so that !KMn0 4 will oxidize the equivalent of 5H. The solution must always contain enough sulphuric acid in order that the metals will be left in the form of sulphates and not as oxides; otherwise less oxygen is available from the perman- ganate (cf. p. 613). The amount of potassium permanganate required for the prep- aration of a liter of normal solution is shown by the above equa- KMn0 4 158.03 Q1 R1 tion to be ^ * - =31.61 gms. N N For the great majority of oxidation analyses and rarely -^ solutions are used. TWT The Preparation of Potassium Permanganate Solution 10 was described on p. 90. 6 OXIDATION METHODS. 597 Standardization of Permanganate Solution. 1. Against Sodium Oxalate (Sorensen).* 1000 c.c. of normal permanganate solution = 67. 00 gms. Na 2 C 2 O 4 . Sodium oxalate can be purchased in a very pure condition. The traces of moisture present may be expelled by heating the oxalate for two hours at 130 and cooling in a desiccator; but for ordinary work this is usually unnecessary. A weighed amount of sodium oxalate is dissolved in 200 c.c. of distilled water at 70, about 20 c.c. of double-normal sulphuric acid are added, and the hot solution titrated with permanganate. At the start the titration should proceed very slowly, waiting after the addition of each drop until the color has disappeared before adding more permanganate 2KMnO 4 + 5Na 2 C 2 O 4 + 8H 2 SO 4 = = K 2 SO 4 + 2MnSO 4 + 5Na 2 S0 4 -f 10C0 2 + 8H 2 O. The purity of the sodium oxalate may be tested by heating a weighed sample in a covered platinum crucible for thirty minutes over a small flame, so that the bottom of the crucible is barely red. It is best to use an alcohol lamp or else insert the crucible in a disk of asbestos as in a sulphur determina- tion. Otherwise the sulphur in the illuminating gas may cause the formation of some sodium sulphate in the crucible. By the heating the oxalate is converted quantitatively into car- bonate, but there is a separation of some carbon which should be removed for the most accurate work. This may be accom- plished by heating the contents of the crucible to a much higher temperature with free access of air, or more readily by adding a few cubic centimeters of water, evaporating the solution to dry- ness on the water bath, and then very carefully heating the crucible over a free flame. In about ten minutes the carbon will all disappear without the carbonate being melted. The crucible is then allowed to cool, its contents dissolved in hot * Z. anal. Chem., 42, 352, 512 (1903); 45, 272 (1906). 59 8 VOLUMETRIC ANALYSIS. water, the crucible and cover thoroughly washed, and the cold solution titrated with tenth-normal hydrochloric acid, using methyl orange as indicator. 1000 c.c. tenth-normal hydrochloric acid = 6. 700 gms. Na 2 C2O 4 . Remark. Sodium oxalate crystallizes without water of crys- tallization, is not hygroscopic, and is especially suited for the standardization of permanganate solutions. Sorensen, in fact, has strongly recommended this substance as a standard for acidimetry, although it has no advantage over the standardization by means of sodium carbonate. The titration of the carbonate may be carried out with methyl orange as an indicator, but Sorensen recommends phenolphthalein as somewhat more reliable. 2. Against Oxalic Acid. Tenth-normal oxalic acid solution is excellent for the standard- ization of a permanganate solution. By means of a pipette 25 c.c. are measured into a beaker, 10 c.c. of dilute sulphuric acid (1:4) are added, the solution is diluted with water at about 70 C. to a volume of 200 c.c., and the permanganate is run into it, with constant stirring, from a glass-stoppered burette. At first the solu- tion is colored red for several seconds, then it becomes colorless but after the reaction is once started the permanganate is rapidly decolorized until an excess is present. The permanent pink color is Imparted to the solution by the permanganate as soon as al) the oxalic acid is oxidized ; this is taken as the end-point. The oxidation is expressed by the following equation: 2KMn0 4 + 5H 2 C 2 4 + 3H 2 SO 4 = K 2 S0 4 + 2MnSO 4 + 8H 2 O+ 10C0 2 . Since for the oxidation of 1 gm. molecule of oxalic acid, COOH | + 0=2C0 2 +H 2 0, COOH N 1 gm.-atom of oxygen is necessary, and 1 liter oxalic acid con- N tains -gV gm.-molecule of the acid, it is evident that 1000 c.c. of OXIDATION METHODS. 599 oxalic acid are equivalent to ^V gin. -atom of oxygen =0.8 gm. and 1 c.c, of the solution = 0.0008 gm. O. N If for the oxidation of 25 c.c. oxalic acid 2^.3 c.c. of per- manganate solution were required, these 24.5 cc correspond to 25X0.0008 = 0.0200 gm. oxygen or 1 c.c. KMn0 4 = -= 0.0008230 gm. O. Instead of expressing the concentration of the permanganate solution in terms of oxygen, it has been the custom to express it in terms of iron. The following consideration will show how the calculation may be made. "From the oxidation equation 2FeO+0 = Fe 2 O 3 it follows that 1 gm.-atom of oxygen is necessary for the oxida- tion of 2 gm. -atoms of iron; consequently ^ gm.-atom of oxygen N (= 1H= 10,000 c.c. = oxalic acid) corresponds to 1 gm.-atom iron,, N so that 25 c.c. -^oxalic acid = 24.3 c.c. permanganate = 25X0. 00559 1398 = 0.1398 gm. iron; or 1 c.c. of the permanganate solution = f ^X,tJ = 0.005751 gm. Fe. Remark. Against the use of oxalic acid solution for the stand- ardization of a permanganate solution is the fact that the con- centration of the aqueous solution is not permanent; for this reason, E. Riegler * proposed the addition of 50 c.c. of concen- trated sulphuric acid to each liter of the oxalic acid, by which means the solution can be kept unchanged for a much longer length of time. That this is the case is shown by the following experiments: A solution of oxalic acid in water was prepared, and also one in dilute sulphuric acid. Both solutions were titrated on the same day with permanganate solution which had been standardized against electrolytic iron. At the end of eight months * Z. anal. Chem., 1896, p. 522. 600 VOLUMETRIC ANALYSIS. the same solutions were titrated against a freshly-standardized permanganate solution, with the following results: Aqueous Oxalic Acid. Oxalic Acid containing Sulphuric Acid. Freshly-prepared . After 8 months. . . 1000 c.c.= 1000.6 c.c.^ sol. 1000 c.c. = 994.9 c.c. " *" 1000 c.c. =1002.5 c.c. ^ sol 1000 c.c. =1001.8 " " " At the end of eight months, therefore, the aqueous solution had depreciated 0.56 per cent, in strength, while the solution con- taining the sulphuric acid had only weakened to an extent of 0.12 per cent, of its original concentration. From this it is evident that a solution of oxalic acid contain- ing sulphuric acid can be used for the standardization of a per- manganate solution, provided the former has not stood more than eight months since it was prepared. The use of old aqueous solutions of oxalic acid is to be discouraged. 3. Against Metallic Iron. It has been a favorite practice to standardize permanganate solutions against iron wire. The wire, however, is never abso- lutely pure, and there is a chance of the impurities reducing permanganate so that the actual iron content of the wire does not suffice to show exactly how much oxidizing agent it will need. Classen,* therefore, has recommended that pure iron be prepared by the electrolyis of ferrous ammonium oxalate. This iron is dissolved in dilute sulphuric acid out of contact with the air and the solution titrated with permanganate (cf. p. 93). The standardization of a potassium permanganate solution can be correctly accomplished by means of iron wire, provided the apparent iron content of the wire has been determined by a com- parison of the values obtained in a titration with a standardization by either electrolytic iron or sodium oxalate. Every time a new supply of iron wire is purchased, the comparison should be made. * Mohr Classen, Lehrbuch der chem. anal. Titrirmethode. OXIDATION METHODS. 601 Determination of the Apparent Iron Value of Iron Wire. The wire is cleaned as described on p. 98, and a weighed portion of about 0.2 gm. is introduced into a flask of not more than 250 c.c. capacity as shown in Fig. 89. The air is displaced by the intro- duction of a stream of carbon dioxide, which has passed through a bottle containing water and another containing copper sulphate solution (cf. p. 93, foot-note); the wire is then dissolved in 55 c.c. FIG. 89. FIG. 90. of dilute sulphuric acid (1 part concentrated acid to 10 of water). During the solution of the wire, the flask is supported somewhat as shown in the drawing, and is closed by a rubber stopper which carries a bulb tube connected with a Bunsen valve.* The con- * A Bunsen valve consists of a short piece of rubber tubing with a cut along a few centimeters of one side, and the outer end of the tubing is closed by a glass rod. This valve prevents the entrance of air from without. A flask larger than 250 c.c. capacity is likely to be so thin as to break during the cooling of the iron solution. In Fig. 80a, instead of using a glass rod at the end of the valve, a glass tube is used which is sealed at one end, and has a hole on one side. This tube serves to prevent the collapse of the rubber tubing at the place where the slit is formed. 6o2 VOLUMETRIC ANALYSIS. tents of the flask are heated by means of a low flame until the wire has all dissolved, after which the solution is boiled gently for a short time. It is then allowed to cool, the stopper is removed, and the permanganate added until a color is obtained which is per- manent for thirty seconds. Instead of using a Bunsen valve, the Contat-Gockel valve may l^e used as shown in Fig. 90. The funnel contains a cold, saturated solution of sodium bicarbonate, tifl-ough which the hydrogen from the flask passes. When the flame is removed sodium bicarbonate solution is drawn into the flask, and this causes the evolution of carbon dioxide, which prevents the entrance of more of the solution. S. Christie, by following the above procedure, found the apparent iron content of a wire to be 99.985 per cent., and Dr. Schudel found 100.21 for another wire. It must be. mentioned, however, that the apparent iron value varies considerably with the way in which the solution of the wire is effected. If the volume of the liquid is large (cf. p. 96), there is more chance of hydrocarbons remaining in solution, and the same is true if the solution is not boiled as in the above direction, but merely heated upon the water bath. Remarks Concerning the Standardization by Means of Electro- lytic Iron. The objection has been raised that electrolytic iron is contaminated with hydrocarbons. According to Avery and Benton Dales,* the iron obtained by the electrolysis of ferrous ammo- nium oxalate contains from 0.2 to 0.4 per cent, carbon on an average; according to Skrabal f considerably more. Verwer and Groll,J however, assert that electrolytic iron contains no carbon provided the bath still contains an excess of iron at the end of trie electrolysis. Christie has carried out extensive experiments in the author's laboratory and found that the electrolytic iron pre- pared by the Classen method does often contain carbon, but the amount is so small that it may be disregarded. Christie, further- more, standardized a solution of permanganate by four different methods and obtained the following values: *Ber., 32, 64 (1899). fZ. anal. Chem., 42, 395 (1903). j Ber., 32, 806 (1899). See also H. Verwer: Chem. Ztg., 25, 792 (1901). 5 TAN DARDIZA TION OF PERMANGANA IE SOLUTION. 603 Against. Value 1 c.c. in Terms of Oxygen. Electrolytic iron 0.0007972 0.0007960 Iodine 0.0007977 0.0007982 Oxalic acid 0.0007978 0.0007967 Sodium oxalate . 0.0007970 0.0007975 4. Against Sodium Thiosulphate. See lodimetry. 5. Against Hydrogen Peroxide. See Gasometric Methods. Permanence of Potassium Permanganate Solutions. As mentioned on p. 90, a permanganate solution will keep indefinitely, provided it is kept free from dust and reducing vapors. In order to test the permanence -of such a solution,* it was stand- ardized against electrolytic iron and after eight months it was again tested, f It had lost only 0.17 per cent, of its original value and could be used for all ordinary analyses. For very accurate work, however, it is advisable to standardize the solution fre- quently. USES OF PERMANGANATE SOLUTION. i. Determination of Iron (Margueritte 1846). N r 0.005585 gm. Fe i c.c. -rr KMnO 4 corresponds to -j 0.007185 gm. FeO ( 0.007985 gm. Fe 2 O 3 In this determination the iron is oxidized from the ferrous to the ferric condition: 2KMnO 4 + 10FeSO 4 + SH 2 SO 4 = K 2 SO 4 + 2MnS0 4 + 5Fe 2 (S0 4 ) 3 + 8H 2 O The solution of the "ferrous salt is strongly acidified with sul- phuric acid (about 5 c.c. of concentrated sulphuric acid should be present for each 100 c.c. of the solution), diluted with boiled water to a volume of 400 to 500 c.c., and titrated in the cold by * The solution was already three months old. fin June, 1899, 1 c.c. of the KMnO 4 solution =0.0054853 gm. Fe; in March, 1900, 1 c.c of the KMnO, solution = 0.0054 7G1 gm. Fe. See also Morse, Hopkins and Walkor, Am. Chem. Jour., 18, 401. 604 VOLUMETRIC ANALYSIS. the addition of potassium permanganate from a glass-stoppered burette until a permanent pink color is obtained. If the perman- ganate solution is tenth-normal, the number of cubic centimeters used multiplied by 0.005585, 0.007185, or 0.007985 will give respectively the amounts of iron, ferrous or ferric oxide. This determination affords very accurate results and is un- questionably one of the best methods for determining iron. Remark. The titration of iroft in hydrochloric acid solution gives high results unless particular precautions are taken. If dilute permanganate solution is allowed to run into a cold dilute solution of ferrous chloride containing hydrochloric acid, the former is decolorized and the iron is oxidized, but there is a noticeable evolution of chlorine.* More permanganate is used up than is necessary to oxidize the ferrous salt to the ferric condition. If, however, permanganate is run into cold, dilute hydrochloric acid, in the absence of ferrous salt, there is no evolution of chlorine. Furthermore, tne presence of a ferric salt does not cause evolution of chlorine. The chlorine, therefore, is not a result of the direct action of the permanganate upon the hydrochloric acid, but is due to the intermediate formation of a peroxide. When permanganate is run into a dilute hydrochloric acid solution containing ferrous chloride and considerable manganous salt, the ferrous iron is quantitatively oxidized to ferric iron and there is no evolution of chlorine. This was shown by Kessler f in 1863 and by Zimmermann J in 1881. It has since been con- firmed by many other chemists. This can be explained as follows: Permanganate reacts with manganous salt and forms, as Volhard || found, MnC>2. This MnO2 oxidizes the ferrous iron to ferric iron. 2FeO + MnO 2 = Fe 2 O 3 + MnO more quickly than it is able to react with hydrochloric acid. * Lowenthal and Lenssen, Z. anal. Chem., 1863, 329. f Fogg. Ann., 118, 779, and 119, 225. j Ber., 14, 779, and Ann. Chem. Pharm., 213, 302. For example, J. A. Friend, J. Chem. Soc., 95, 1228 (1909). C.C. Jones and J. H, Jeffery. The Analyst, 34, 306 (1909). 1 1 Ann. Chem. Pharm., 198, 337, STANDARDIZATION OF PERMANGANATE SOLUTION. 605 Zimmermann * suspected that, in the presence of manganous salts, ferrous iron is converted into a peroxide which immediately breaks down into ferric iron and oxygen, and the latter acts upon the hydrochloric acid. This has more recently been confirmed by the interesting work of Manchot.f According to Manchot, there is formed in all oxidation processes a " primary oxide " which has the character of a peroxide. These primary oxides are as a rule unstable compounds which cannot be isolated and which constantly tend to give up oxygen and pass to a more stable lower state of oxidation. When an acceptor { is present, it will take up the oxygen which is lost by the primary oxide as it passes to a lower and more stable state of oxidation; in the absence of an acceptor this oxygen is evolved as gas. According to the method of oxidation, iron tends to form different primary oxides. Thus, in the direct oxidation by means .of oxygen, the primary oxide is Fe02; in the oxidation by means of permanganate, chromic acid, or hydrogen peroxide, it is Fe 2 O 5 ; whereas FeOs is probably formed in the oxidation by means of hypochlorous acid. The oxidation of ferrous oxide to ferric oxide, therefore, does not take place directly, as has been usually assumed, but the primary oxide FeO 2 is first formed, and this reacts with more of the ferrous oxide, which therefore plays the part of an acceptor, to form ferric oxide: 2FeO + O 2 = 2FeO 2 , 2FeO + 2FeO 2 = 2Fe 2 O 3 . Potassium permanganate causes the formation of Fe 2 Os as primary oxide: 2FeO + Mn 2 7 -^Fe 2 5 + 2MnO 2 . * Ber., 14, 779 and Ann. Chem. Pharm., 213, 302. f Ann. Chem. Pharm., 325, 105 (1902). J An acceptor is a substance which is not oxidized by oxygen alone, but can be thus oxidized by the aid of some other substance present called an auto- oxydator. A substance which tends to be peroxidized may play the part of an acceptor. Cf. Engler, Ber., 33, 1097 (1900). 606 VOLUMETRIC ANALYSIS. The MnO2 at once oxidizes more ferrous oxide to ferric oxide 2FeO + MnO 2 = Fe 2 3 + MnO and the Fe2O 5 converts the MnO to Mn0 2 again. If, however, the concentration of the manganous salt in the solution is too low, hydrochloric acid begins to play the part of an acceptor, so that a part of the Fe 2 0s is used up in the oxidation of hydrochloric acid: Fe 2 O 5 + 10HC1 = 2FeCl 3 + 5H 2 O + 2C1 2 The action of the manganese sulphate is twofold. On the one hand it regulates the velocity of the reaction between ferrous oxide and permanganic aicd, for, according to Volhard, the HMn04 acts upon the manganous salt with the formation of manganese peroxide, which then reacts with the ferrous salt; on the other hand it takes up the oxygen from the iron peroxide and carries it to the unoxidized ferrous salt. In both cases it is essential that manganese peroxide does not react with hydro- chloric acid very rapidly, and it is necessary, too, that the amount of manganous salt shall greatly exceed the amount of iron present. Zimmermann suggested a similar explanation, but it seemed to meet with but little approval, so that the hypothesis of Wagner * was quite generally adopted. The latter claimed that the excess of permanganate required for the titration of ferrous chloride in the absence of manganous sulphate was due to the intermediate formation and rapid oxidation of a ferrous-hydrochloric acid, FeCl 2 -2HCl. Manchot's explanation, however, seems to be the better one. Although it is possible, then, to titrate iron in hydrochloric acid solutions in the presence of manganous sulphate, the method possesses the disadvantage that the end-point cannot be seen so distinctly as when no chloride is present, since ferric chloride forms a much more yellow solution than does ferric sulphate. * Zeitschr. f. physikal. Chem., 28, 33. STANDARDIZATION OF PERMANGANATE SOLUTION. 607 This difficulty can be overcome by the addition of phosphoric acid, as suggested by C. Reinhardt.* TITRATION OF FERROUS SALTS IN HYDROCHLORIC ACID SOLUTION. METHOD OF ZIMMERMANN-REINHARDT. From 20 to 25 c.c. of the manganese sulphate solution f pre- pared as described below are added to the solution, and after diluting with boiled water to a volume of 500 c.c. it is titrated with potassium permanganate which is added so slowly that the drops can be counted. Care is taken toward the last not to add a drop of permanganate until the color of the preceding oils has disappeared. The manganous sulphate solution is prepared as follows : 67 gms. of crystallized manganous sulphate (MnSO 4 + 4H 2 O) are dissolved in 500 to 600 c.c. of water, 138 c.c. of phosphoric acid (of specific gravity 1.7) and 130 c.c. of concentrated sulphuric acid (sp. gr. 1.82) are added, and the mixture is diluted to 1 liter. If the iron is present as ferric salt, it must be reduced com- pletely to the ferrous condition before it can be titrated with potassium permanganate. THE REDUCTION OF FERRIC SALTS TO FERROUS SALTS can be accomplished in a number of different ways. 1. By Hydrogen Sulphide. This reduction has already been described on page 99. 2. By Sulphur Dioxide. The solution containing the ferric salt is neutralized with sodium carbonate, % an excess of sulphurous acid is added, the solution boiled, and a current of carbon dioxide is passed * Stahl und Eisen, 1884, p. 709, and Chem. Ztg., 13, 323. f It is well to add one cubic centimeter of manganese sulphate for each cubic centimeter of HC1 (sp. gr. 1.12) present. Cf. J. A. Friend or Jones and Jeffery, loc. cit. t Ferric salts are not completely reduced by sulphurous acid in the pres- ence of considerable hydrochloric or sulphuric acid. 6oS VOLUMETRIC ANALYSIS. through it until the excess of the reagent is completely removed.* The reduced solution is then cooled in an atmosphere of carbon dioxide and titrated. 3. By Metals. The acid solution of the ferric salt, contained in a small flask fitted with a Bunsen valve, is reduced by heating on the water- ' bath with the addition of small piece* of chemically-pure zinc until the solution is completely colorless and a drop of it, removed by means of a piece of capillary tubing, will no longer give any color with potassium sulphocyanate solution. After cooling, the solu- tion is poured through a funnel containing a platinum cone (no paper) , and the undissolved zinc remaining in the funnel is washed several times with boiled water, f Remark. Since zinc often contains iron, a blank experiment must be made by dissolving 3 to 5 gms. in the saTne way and titra- ting the solution with permanganate. If iron is present, as shown by the fact that a measurable amount of potassium permanga- nate is decolorized, the reduction of the ferric salt must be effected by means of a weighed amount of zinc and a correction made for the iron. It is self-evident that in this case the titration must not take place until all of the zinc has dissolved. Instead of zinc, cadmium and aluminium are frequently used. Remark. Against this method objections can be raised. In the first place, the fact that a foreign metal is introduced into the solution is in many cases unfortunate. Furthermore, by means of zinc, titanic acid is reduced to Ti 2 O 3 , only to be oxidized again by the permanganate solution, so that more permanganate solution will then be required than corresponds to the amount of iron present. By means of H 2 S or SO 2 , titanic acid is not reduced and there is no foreign metal introduced into the solution. Con- * It is not advisable to depend upon the sense of smell. The escaping gas is tested by passing it through dilute sulphuric acid containing a few drops of KMnO 4 solution. If the latter is not decolorized at the end of two or three minutes, the excess of sulphurous acid has been removed. t The reduction by means of zinc may be satisfactorily accomplished with a " Jones redactor." Cf . Fig. 91, p. 637. REDUCTION OF FERRIC SALTS TO FERROUS SALTS. 609 sequently, for accurate mineral analyses, it is necessary to use one of these methods, and in fact the reduction by means of hydrogen sulphide is to be preferred. By means of the latter the ferric salt is completely reduced, independent of how little or how much free acid is present in the solution; again, any metals of the hydrogen sulphide group are precipitated at the same time; while finally it is easy to recognize the fact that the excess of the gas has been removed by the use of the sensitive lead acetate paper test. 4. By Stannous Chloride. This method proposed by Zimmermann and Reinhardt* is especially suited for metallurgical purposes, because it can be accomplished most rapidly. Principle. The method depends upon the fact that ferric chloride in hot solution is easily reduced by stannous chloride: SnCl 2 + 2FeCl 3 = SnCl 4 + 2FeCl 2 . The complete decolorization of the solution shows the end- point of reduction. The excess of stannous chloride is afterwards oxidized by means of mercuric chloride: SnCl 2 + 2HgCl 2 = SnCl 4 + Hg 2 Cl 2 . After this treatment, which consumes but a few minutes, some manganese sulphate solution is added and the solution imme- diately titrated with potassium permanganate, which is added slowly. Requirements. (a) Stannous chloride solution. 50 gms. of stannous chloride are dissolved in 100 c.c. of concentrated hydrochloric acid and diluted with water to a volume of one liter. (6) Hydrochloric acid, sp. gr. 1.12. ' * Loc. cit. 610 VOLUMETRIC ANALYSIS. (c) Mercuric chloride solution. A saturated solution of the pure commercial salt in water is used. (d) Manganese sulphate solution. See p. 607. Procedure. The ferric salt is dissolved in 20 to 25 c.c. of the hydrochloric acid (b) heated to boiling, the flame removed, and the stannous chloride solution (a) is added drop by drop until the iron solution just becomes colorless. The solution is cooled to at least the room temperature and 10 .c. of mercuric chloride (c) are quickly added, whereby a slight silky precipitate of Hg2Cl2* is formed. After ten minutes the solution is diluted to about 500 c.c., 20 to 25 c.c. of the manganese sulphate solution (d) are added, and the mixture is titrated (very slowly) with potassium per- manganate until a pink color permanent for one minute is obtained. Example: Determination of Iron in Hematite, Fe 2 O3. About 0.25 to 0.3 gm. of the finely-powdered mineral is weighed out into a beaker, 3 c.c. of the stannous chloride solution (a) f are added and 25 c.c. of the acid (b). The beaker is covered with a watch- glass and its contents heated nearly to boiling until all of the iron oxide has dissolved and a white sandy residue is obtained. This operation seldom requires more than ten minutes. The slightly yellow colored solution thus obtained is carefully treated with stannous chloride drop by drop until it becomes colorless and the reduced solution is analyzed as above. * If the precipitate produced by mercuric chloride is at all grayish in color, the portion must be thrown away; too large an excess of stannous chloride was used. Moreover, the end point with permanganate is difficult to see if the solution contains much precipitate. f The stannous chloride greatly facilitates the solution of the hematite. If too much is used, strong permanganate should be added drop by drop until the yellow color of ferric chloride appears, and the solution then care- fully decolorized again. DETERMINATION OF METALLIC IRON. 611 Determination of Metallic Iron in the Presence of Iron Oxide. This method is useful for testing ferrum reductum which is ob- tained by the reduction of Fe 2 O 3 in a stream of hydrogen. Usually the reduction is not complete and the preparation contains, besides the metallic iron, some oxide, usually assumed to be Fe 3 O 4 . The value of the preparation depends upon the free iron content. (a) Method of Wilner *-Merck.} Principle. The method is based upon the fact that a neutral solution of mercuric chloride dissolves iron according to the equa- tion Fe + HgCl 2 =Hg+FeCl 2 while the Fe 3 4 is not attacked. The solution of ferrous chloride is titrated with permanganate solution. Procedure. About 0.5 g. of ferrum reductum, in the form of a fine powder, { is placed in a 100 c.c. graduated flask, from which the air is replaced by CO 2 , 3 gms. of solid mercuric chloride 'are added and 50 c.c. of water. The contents of the flask are then heated to boiling, by means of a small flame, and the liquid boiled for a minute. The flask is then filled up to the mark with boiled water. After cooling to 15 the solution is again carefully brought to the mark, well shaken, and then allowed to stand in the stop- pered flask until the precipitate has settled. The liquid is then poured through a dry filter and the filtrate caught in a flask filled with carbon dioxide. Of this filtrate, 20 c.c. are taken, acidified with 20 c.c. of sulphuric acid (1:4), treated with 10 c.c. of man- ganese sulphate solution, diluted to 200 c.c., and treated with tenth-normal permanganate solution. * Farm. Tidskrift, 1880, 225. t Z. anal. Chem., 41, 710 (1902). J A coarse powder is not decomposed quantitatively. See page 607. 612 VOLUMETRIC ANALYSIS. The Ferric Chloride Method* Principle. A neutral solution of ferric chloride dissolves metallic iron with the formation of ferrous chloride : Fe + 2FeCl 3 = 3FeCl 2 and the ferrous chloride formed is titrated with permanganate solution. One-third of the iron th^s found corresponds to the weight of metallic iron present in the sample. Procedure. About 0.5 g. of ferrum reductum are placed in a 100 c.c. graduated flask, which has been filled with CO 2 , and 50 c.c. of ferric chloride are added (1 gm. anhydrous ferric chloride in 20 c.c. water). f The flask is stoppered and its contents frequently shaken during the next fifteen or twenty minutes. The solution is then brought to the mark with cold, boiled water, mixed, the flask stoppered, and allowed to stand over night. Of the clear supernatant liquid, 20 c.c. are removed by a pipette and titrated, as in the previous method, with tenth-normal perman- ganate solution .{ 2. Determination of Manganese. Method of Volhard. 1000 c,c, N. KMn0 4 |[ = ^ 3 X ^ 4 ' 93 = 16.48 gms. Mn. If an almost boiling, slightly acid solution of manganese sul- phate is slowly treated with a solution of potassium permanga- * A. Christensen, Z. anal. Chem., 44, 535 (1905). t The ferric chloride must give a clear solution in cold water. As it often contains a little ferrous chloride, a blank test must be made and a correction, corresponding to the amount of iron found, applied to the analysis proper. J For other methods of analyzing ferrum reductum, see E. Schmidt, Chem. Ztg , 21, 700 (1897). A. Marquardt, Ibid., 45, 743 (1901). L. Wolfrum, Inaug. Dissert., Erlanger, 1896. F. Forster and V. Herold, z. Elektrochefn. 1910, 461. Ann. d. Chem. und Pharm., 198, 318. i| Strictly speaking, the normality of the permanganate is different when used to oxidize manganese in slightly acid or neutral solution. In this case DETERMINATION OF MANGANESE. 613 nate, each drop will cause the formation of manganous acid (H 2 MnOs), which is formed under certain conditions, as described below, according to the following scheme: K 2 Mn 2 O 7 + 3MnO - K 2 O + 5MnO 2 2KMnO 4 According to this equation, therefore, 2KMnO 4 will oxidize 3 gm.-atoms of manganese, and as 1000 c.c. of N. KMnO 4 con- tain \ gm.-mol. KMnO 4 , evidently this amount of permanganate corresponds to = 16.48 gms. Mn. A. Guyard, who first determined manganese by this method, assumed that the oxidation took place according to the following equation : 2KMn0 4 + 3MnSO 4 + 7H 2 O = 2KHSO 4 + H 2 SO 4 + 5H 2 MnO 3 . In reality, however, the reaction does not take place in this way, but instead of pure manganous acid being precipitated, dif- ferent acid manganites of varying composition are formed; e.g., /OH 4KMn0 4 + 1 !MnS0 4 + 14H 2 = 4KHSO 4 + 7H 2 SO 4 + 5 %> Mn. .OH Volhard has shown that if calcium/barium, or, better still, zinc salts, are present, manganites of these metals are precipitated the manganese of the permanganate is reduced to the quadrivalent form instead of to bivalent manganese, so that a normal solution would now contain KMnO 4 KMn0 4 gms. instead of gms. Inasmuch as it is customary to stand- o o ardize permanganate as outlined on pages 597-603, we shall understand by normal KMnO 4 , a solution containing one-fifth of the molecular weight. [Translator.] 6i 4 VOLUMETRIC ANALYSIS. The precipitate, although varying in composition, contains all of the manganese in the quadrivalent form; e.g., /OH 4KMn0 4 + 5ZnS0 4 + 6MnS0 4 + 14H 2 = = 4KHS0 4 + 7H 2 S0 4 + 5 ~ > Zn. Mn^O \OH In case iron is present, the reaction does not take place quan- titatively in the direction from left to right, so that a different procedure is then necessary. (A) PROCEDURE WHEN IRON IS ABSENT. N Requirements. 1. A potassium permanganate solution. 2. A manganese sulphate solution, obtained by dissolving 4.530 gms. of anhydrous manganous sulphate in one liter of solution: N 1 c.c. of this solution= 1 c.c. of KMnO 4 .* 3. A zinc sulphate solution obtained by dissolving 200 gms. zinc sulphate in one liter of water. 4. Zinc oxide suspended in water, obtained by precipitating pure zinc sulphate by means of caustic potash solution in such a way that the solution does not react alkaline. The residue is washed several times with hot water, then transferred to a tightly- stoppered bottle, and kept suspended in water. Standardization of the Permanganate Solution. 20 c.c. of the manganese sulphate solution are placed in an Erlenmeyer flask, 40 c.c. of zinc sulphate solution and 2 or 3 drops of nitric acid f are added, after which the mixture is diluted to * Strictly speaking, this solution is ^ normal. By definition, a ~ solu- of manganese sulphate contains - 4 = 7.550 gms. MnSO 4 in one liter. Such a solution, however, would not be equivalent to a KMnO 4 solution which is tenth normal in acid solution. Cf. foot-note to page 527 (Translator). f The addition of the nitric acid causes the precipitate to settle much more quickly. DETERMINATION OF MANGANESE IN STEEL. 615 200 c.c., heated to boiling, and treated with potassium permanga- nate solution, added with constant shaking, until the supernatant liquid remains a permanent pink. Titration of Manganese. If a neutral solution of manganese sulphate is to be analyzed, the same procedure is used as in the above standardization. If the solution contains manganous chloride, it should be freed from hydrochloric acid by evaporation with an excess of sulphuric acid. The acid solution thus obtained is neutralized with the zinc oxide until a little of the latter remains suspended in the liquid. From this point the procedure is the same as before. (B) PROCEDURE WHEN IRON IS PRESENT. If a hydrochloric acid solution is to be analyzed containing all of the iron in the ferric form, it is evaporated to dryness with the addition of sulphuric acid, the dry mass is moistened with nitric acid and warmed until complete solution is effected. The greater part of the acid is neutralized with sodium hydroxide solution, the solution placed in a measuring-flask, and an excess of the zinc oxide is added whereby all of the iron is precipitated as hydroxide. The liquid is diluted up to the mark with water, filtered through a dry filter, and an aliquot part of the filtrate is titrated as before with potassium permanganate solution.* Determination of Manganese in Steel. (a) Volhard Method. The solution is prepared for titration by dissolving the steel borings f in nitric acid (sp. gr. 1.2), evaporating the solution, after the addition of 20 c.c. of 50 per cent, sulphuric acid, allowing the residue to cool, and then adding 150 c.c. of cold water. The water is boiled until the ferric sulphate is all dissolved, the solu- tion filtered, the filtrate nearly neutralized with sodium carbonate and the zinc oxide added exactly as described above. * The first few cubic centimeters of the filtrate should be discarded, for the dry filter absorbs some of the dissolved substance. t Three gms. of steel are used when the manganese content is about 1 per cent, less when the content is higher. The process is not well suited for low manganese steels. 616 VOLUMETRIC ANALYSIS. (6) The Bismuthate Method. This method originated with Schneider, f who used bismuth tetroxide as the oxidizing agent, but as the oxide is difficult to prepare free from chlorides and traces of chloride interfere with the end point of the titration, it was abandoned by Reddrop and Ramage,t who proposed the use of sodium bismuthate, NaBiOs. The product sold under this name is of more or less indefinite composition. It may be prepared by heating 20 parts of caustic soda nearly to redness in an iron or nickel crucible, adding in small quantities from time to time ten parts of dry basic bismuth nitrate, followed by two parts of sodium peroxide, pouring the yellow fused mass on an iron plate to cool. When cold, the fusion is extracted with water, collected on an asbestos filter, washed five times by decantation with water, and dried in the hot closet at 110. After grinding and sifting the product is ready for use. The process is based on the fact that a manganous salt in the presence of an excess of nitric acid is oxidized to perman- ganic acid by sodium bismuthate. The permanganic acid formed is very stable in nitric acid of 1.135 sp. gr. when the solution is cold, but in hot solutions the excess of bismuthate is rapidly decomposed and then the nitric acid rea'cts with the permanganic acid; as soon as a small amount of manganous salt is formed the remainder of the permanganic acid is decom- posed, manganous nitrate dissolves and manganese dioxide precipitates. In the cold, however, the excess of the bismuth salt may be filtered off and to the clear filtrate an excess of ferrous sul- phate added; the excess of the latter is determined by titrating with permanganate. The end-reactions are very sharp and the method is extremely accurate. 1000 c.c. N. KMn0 4 = 10.99 gms. Mn. * A. A. Blair, J. Am. Chem. Soc. 26, 793. t Ding. poly. J. 269, 224. J Trans. Chem. Soc. 1895, 268. DETERMINATION OF MANGANESE IN STEEL 617 Procedure for Steels. Dissolve 1 gm. of drillings in 50 c.c. of nitric acid (sp. gr. 1.135*) in an Erlenmeyer flask of 200 c.c. capacity, cool, and add about 0.5 gm. of bismuthate. The bismuthate may be measured in a small spoon and experience will soon enable the operator to judge of the amount with sufficient accuracy. Heat for a few minutes or until the pink color has disappeared, with or without the precipitation of manganese dioxide. If the solution now shows precipitated manganese dioxide, add crystals of ferrous sulphate free from manganese, sulphurous acid or sodium thiosulphate until it becomes clear. Heat for two minutes to remove oxides of nitrogen and cool to about 15. Now add 2 or 3 gms. more of sodium bismuthate and agitate the contents of the flask for several minutes. Dilute with 50 c.c. of 3 per cent, nitric acid and filter through an asbestos filter, using gentle suction. Wash the asbestos with 50 to 100 c.c. of cold 3 per cent, nitric acid.f Run into this solution 50 c.c. of standardized ferrous sulphate solution and titrate back to pink color with potassium permanganate. The value of the ferrous sulphate solution in terms of potas- sium permanganate must be determined in the following manner: Measure into a 250-c.c. Erlenmeyer flask 50 c.c. of cold nitric acid (sp. gr. 1.13), add about 0.5 gm. of sodium bismuthate, agitate, dilute with 50 c.c. of 3 per cent nitric acid and filter through asbestos. To the filtrate add 50 c.c. of ferrous sulphate solution and titrate with permanganate solution to pink color. Having determined the value of the permanganate solution in terms of ferrous sulphate, the manganese in the sample is represented by the difference between the amounts of perman- ganate solution actually used in the determination and in the * Concentrated nitric acid mixed with three times as much water. This is often called 25 per cent nitric acid (by volume). It contains 22.5 per cent nitric acid by weight. t 30 c.c. HNO 3 , sp. gr. 1.42, in one liter of water. J An approximately 0.03 N solution made by dissolving 9 gms. crys- tallized ferrous sulphate, FeSO 4 -7H 2 O or 12 gms. of ferrous ammonium sulphate, FeSC^NH^SCVGHzO in 950 c.c. water and 50 c.c. of concen- trated sulphuric acid. 1 gm. KMnO 4 to the liter. 6i8 VOLUMETRIC ANALYSIS. titration of a volume of ferrous sulphate equivalent to that used in the determination. Pig Iron. Dissolve 1 gm. in 25 c.c. of nitric acid (sp. gr. 1.135) in a small beaker and as soon as the action has ceased filter on a 7-cm. filter into a 200-c.c. Erlenmeyer flask, wash with 30 c.c. of the same acid as proceed as in the case of steels. In the analysis of white irons it may be necessary to treat the solution several times with bismulhate to destroy the combined carbon. The solution, when cold, should be nearly colorless; if not, another treatment with bismuthate is necessary. Iron Ores Containing Less than Two Per Cent, of Manganese. Treat 1 gm. in a platinum dish or crucible with 4 c.c. of strong sulphuric acid, 10 c.c. of water and 10 to 20 c.c. of hydrofluoric acid. Evaporate until the sulphuric acid fumes freely. Cool and dissolve in 25 c.c. of 1.135 nitric acid. If no appreciable residue remains, transfer to a 200-c.c. Erlenmeyer flask, using 25 c.c. of 1.135 nitric acid to rinse the dish or crucible and proceed as usual. If there is an appreciable residue, filter on a small filter into a beaker, wash with water, burn the filter and residue in a crucible and fuse with a small amount of potassium bisul- phate. Dissolve in water with the addition of a little nitric acid, add to the main filtrate, evaporate nearly to dryness, take up in 1.135 nitric aci.d and transfer to the flask as before. Manganese Ores and Iron Ores High in Manganese. Treat 1 gm. as in the case of iron ores, using a little sulphurous acid, if necessary. Transfer the solution to a 500-c.c. flask, dilute to the mark, mix thoroughly and measure into a flask from a care- fully calibrated pipette such a volume of the solution as will give from 1 to 2 per cent, of manganese and enough strong nitric acid (sp. gr. 1.4) to yield a mixture of 1.135 acid in a volume of 50 to 60 c.c. Ferro-manganese. Treat 1 gm. exactly like steel. Dilute to 500 or 1000 c.c. and proceed as in manganese ores. Ferro-silicon. Treat 1 gm. with sulphuric and hydrofluoric acids and proceed as with iron ores. Special Steels, Steels containing chromium offer no special difficulties, except that it must be noted that while in hot solu- tions the chromium is oxidized to chromic acid, which is reduced by the addition of sulphurous acid, the oxidation proceeds so DETERMINATION OF\ MANGANESE IN STEEL 619 slowly in cold solutions that if there is no delay in the nitration and titration the results are not affected. Steels containing tungsten are sometimes troublesome on account of the necessity for getting rid of the tungstic acid. Those that decompose readily in nitric acid may be filtered and the filtrate treated like pig iron, but when it is necessary to use hydrochloric acid it is best to treat with aqua regia, evaporate to dryness, redissolve in hydrochloric acid, add a few drops of nitric acid, dilute, boil, and filter. Get rid of every trace of hydrochloric acid by repeated evaporations with nitric acid and proceed as with an ordinary steel. (c) Williams Methid* 1000 c.c. KMn0 4 = jy=p = 2.747 gms. Mn. Principle. If a nitric acid solution of a manganous salt is heated with potassium chlorate, all of the manganese is pre- cipitated as the dioxide: The Mn0 2 is dissolved in a measured volume of acid ferrous sul- phate, and the excess is titrated with tenth-normal permanganate. Procedure. From 2 to 3 gms. of an ordinary steel, about 1 gm. of " spiegel " or 0.3 to 0.5 gm. of ferromanganese are weighed into a 600 c.c. Erlenmeyer flask and dissolved in 60 c.c. of nitric acid, sp. gr. 1.2. To prevent loss by spattering, a small funnel is placed in the neck of the flask. After evaporating the solu- tion to a volume of about 15 c.c., 50 c.c. of concentrated nitric acid, sp. gr. 1.42, and 3 gms. of solid potassium chlorate are added, and the solution is boiled for fifteen minutes. It is then removed from the source of heat and the treatment with 50 c.c. concentrated nitric acid and 3 gms. of potassium chlorate is repeated, after which the solution is boiled for fifteen minutes longor. The solution is cooled quickly by placing the flask in cold water, and the precipitated manganese dioxide is filtered on asbestos,t * Trans. Inst. Min. Eng., 10, 100. See also W. Hampe, Chem. Ztg., 7, 73 (1883), 9, 1478 (1885), and Ukena, Stahl imd Eisen, 11, 373 (1891). f A satisfactory filter is obtained by placing a little glass wool in a funnel, and on this a little asbestos, such as is used for Gooch crucibles. 620 . VOLUMETRIC ANALYSIS. washed with concentrated nitric acid till free from iron, and with water till free from acid. The asbestos pad and precipitate is transferred to the original flask, covered with 50 c.c. of standard- ized ferrous sulphate solution,* and diluted with water to a volume of 200 c.c. The contents of the flask are shaken with glass beads until all the precipitate is dissolved, and the solution is then titrated with tenth-normal permanganate. The amount of manganese present is computed as follows : N 50 c.c. of ferrous sulphate solution require T c.c. KMnO 4 50 c.c. of ferrous sulphate solution +a g. of iron require i c.c. KMnO 4 . 10 N Consequently a g. of the substance = (7 7 t) c.c. KMnC>4 and 0.002747(7 7 -OX100 - = % Mn. a (d) G. v. Knorre's Persulphate Method.^ Principle. If a solution of manganous sulphate containing a little free sulphuric acid is treated with ammonium persulphate, the manganese is precipitated quantitatively as hydrated man- ganese dioxide. MnS0 4 + (NH 4 ) 2 S 2 8 + 3H 2 = MnO 2 H 2 O + (NH 4 ) 2 SO 4 + 2H 2 SO 4 and the latter can be estimated as in the above determination. Procedure. In the case of the harder alloys of iron and manganese, the sample is pulverized as much as possible in a steel mortar. The weights of sample taken correspond to those recom- mended for the previous determination. The weighed substance is treated in a beaker with sulphuric acid (1:10) at the boiling temperature, using 50 c.c. of the dilute acid for the harder alloys and 60 c.c. for the softer ones. As soon as the evolution of * 10 gm. FeSO 4 -7H 2 O. 50 c.c. concentrated H 2 SO 4 , and 950 c.c. water, t Z. angew. Chem., 14, 1149_(1901). ADDITIONAL METHODS FOR DETERMINING MANGANESE. 621 hydrogen ceases, the solution is filtered through a small filter which is washed with cold water until the washings give no test for iron with potassium ferricyanide. Frequently, especially in the case of ferro-manganese rich in silicon, the insoluble residue still contains a little manganese, so that for an accurate analysi? Ledebur ignites the filter and precipitate in a platinum crucible, treats the residue with hydrofluoric acid and about 0.5 c.c. of concentrated sulphuric acid, and evaporates in an air bath until sulphuric anhydride vapors are evolved. After cooling the contents of the crucible are added to the main solution. Then from 150 to 250 c.c. of ammonium persulphate solution are added (60 gms. per liter) the solution diluted to about 300 c.c. and heated to boiling. After boiling for fifteen minutes, the precipitate is allowed to settle and is filtered, washed, treated with an excess of ferrous sulphate solution and titrated exactly as in the previous Williams method. 3. Determination of Uranium. Method of Belhoubek,* Zimmermann,t Hillebrand.J 1000 c.c. N. KMnO 4 = -= =119.3 gms. U. This method is especially suited for testing the purity of a precipitate of U 3 O 8 obtained in the analysis of uranium minerals. It is based upon the fact that when U 3 O 8 is heated in a closed tube ivith dilute sulphuric acid at 150 to 175 C. it is readily decom- posed according to the equation U 3 8 + 4H 2 S0 4 2U0 2 S0 4 + U (SO J 2 + 4H 2 O, forming uranyl and uranous sulphates. The latter compound is oxidized to the former by means of potassium permanganate, 2KMnO 4 + 5U(SO 4 ) 2 +2H 2 O = = 2KHS0 4 + 2MnSO 4 + H 2 SO 4 + 5UO 2 S0 4 . * Journ. f. prakt. Chem., 99, 231. t Ann. der Chem. u. Pharm., 232, 285. % U. S. Geol. Survey, No. 78, 90 (1889). 622 VOLUMETRIC ANALYSIS. From this equation it follows that 2 gm.-mols. of KMnO 4 are equiva- lent to 5 gm.-atoms of uranium, and 1000 c.c. N. KMnO 4 solution U 238 5 (=KMnO 4 ) = } gm.-atom of uranium = = = 119.25 gms. U. 2i Procedure. The weighed amount of U 3 8 is placed in a tube closed at one end, 10 to 15 c.c. of dilute sulphuric acid (1:6) are added, and the open end of the tube is made narrower by heating in a blast-lamp and drawing it* out somewhat. The air in the tube is removed by inserting a long capillary so that it reaches to the bottom of the tube containing the substance, and conducting a current of carbon dioxide through it; the larger tube is finally sealed without removing the capillary. The tube is then heated in a "bomb furnace" at 150-175 C. until everything has dissolved to a clear green liquid. After cooling, the tube is opened by making a scratch with a file and touching it with a hot glass rod, The contents are poured into a large porcelain dish, diluted with N distilled water to 500-700 c.c., and titrated with KMnO 4 solu- tion until a permanent pink color is obtained. L c.c. ~ KMnO 4 = 0.01 1925 gm.U = 0.01352, Remark. The above method gives very exact results. 1 c.c. ~ KMnO 4 = 0.01 1925 gm.U = 0.013525 gm. UO 2 oxidized.* 4. Determination of Oxalic Acid. 1000 c.c. N. KMnO,= H A 2H q IH5 = 63 . The procedure is exactly the same as was described under the standardization of permanganate by means of oxalic acid (page 598) * It must be remembered, however, that only one-third of the total uranium in U 3 O 8 has been oxidized by the KMnO, (U 3 O 8 = 2UO 3 + UO 2 ) . Consequently, with regard to the total uranium, 1 c.c. ^- KMnO 4 = 0.03578 gm. 17 = 0.04218 gm. U 3 O 8 . [Translator]. ANALYSIS OF RED LEAD. 623 5. Determination of Calcium. Ca 40.09 1000 c.c. N. KMnO 4 = =- - = 20.05 gms. Ca. The calcium is precipitated as described on p. 70 in the form of its oxalate, filtered, and washed with hot water. The still moist precipitate is transferred to a beaker by means of a stream of water from the wash-bottle, and the part remaining on the filter is removed by allowing warm dilute sulphuric acid to pass through it several times. To the turbid solution in the beaker, 20 c.c. of sulphuric acid (1:1) are added, and after dilution with hot water to a volume of from 300 to 400 c.c. the oxalic acid is titrated with N ~ KMnO 4 solution. 1 c.c. ^ KMnO 4 =0.002005 gm. Ca. 6. Determination of Pb0 2 in Minium [Red Lead, Pb 3 OJ . Method of Lux.* 1000 c.c. N. KMn0 4 =^p = = 119.55 gms. PbO r Principle. If lead peroxide (PbO 2 ) is treated with oxalic acid in acid solution, the latter is oxidized according to the following equation: PbO 2 + H 2 C 2 4 = PbO-f H 2 O+ 2CO 2 . If the decomposition takes place with a measured amount of titrated oxalic acid solution and the excess of the latter is titrated by means of potassium permanganate solution, the difference shows the amount of oxalic acid necessary to effect the reduction of the lead peroxide. Procedure. About 0.25 gm. of minium (red lead) is weighed into a porcelain dish and heated with 20 to 30 c.c. of double normal nitric acid.f The original oxide is thereby changed into soluble lead nitrate and brown, insoluble N After solution is effected, 50 c.c. -= oxalic acid are added, the o * Z. anal. Chem., 19, p. 153. t Nitric acid sp. gr. 1.2 diluted with two volumes of water. 621 VOLVM2TRIC ANALYSIS. N solution is heated to boiling, and titrated hot with -^ 5 N N If t c.c. KMn0 4 solution were used, then 50 t c.c. -= H 2 C 2 O 4 o o were necessary for the reduction of the amount of Pb0 2 contained in the minium (a gm.) taken for analysis. Since 1000 c.c. N. H 2 C 2 O 4 = 119.55 gms. PbO 2 , then 1000 c.c. - -i^p' =23.91 gms. PbO 2 and 1 c.c. = 0.02391 gm. PbO 2 . Consequently N (50-0 c.c. oxalic acid correspond to (50-0X0.02391 gm. Pb0 2 . The per cent, of the latter is a:(50-0x0.02391 = 100:s (50-OX2.391 x=~ - per cent. PbO 2 .* a 7. Determination of Mn0 2 in Pyrolusite. 1000 c.c. N. KMnO 4 = =43.47 gms.MnO 3 . (a) Method of Levol and Poggiale, Modified by G. Lunge.~\ After drying at 100 to constant weight, 1.0866 gms. of the finely powdered pyrolusite are placed in a 250 c.c. flask which is provided with a Contat valve (see page 602). The air is expelled by conducting CO 2 into the flask, and then 75 c.c. of the ferrous sulphate solution, prepared as described below, are added, the flask closed, and its contents heated over a small flame until there is no longer any dark-colored residue. The flask is cooled quickly, the contents diluted with 200 c.c. water, and the excess of ferrous sulphate titrated with 0.5N KMnO 4 solution. Immediately be- fore the analysis, the titer of the ferro-sulphate solution is deter- mined by taking 25 c.c. of it, diluting to 200 c.c. and titrating with permanganate. * To express the results in per cent. PbsO^ the number 2.391 should be replaced by [6.853. [Translator.] f Chem.-techn. Untersuchungsmethoden. Edition 6, Vol. I., p. 569. DETERMINATION OF MnO 2 IN PYROLUSITE. 25 By the treatment of the pyrolusite with ferrous sulphate, the following reaction takes place: MnO 2 + 2FeS0 4 + 2H 2 SO 4 = Fe 2 (S0 4 ) 3 + MnSO 4 + 2H 2 O. The computation of the percentage of MnO 2 is as follows: 75 c.c. FeSO 4 solution require T c.c. 0.5N KMn0 4 75 c.c FeSO 4 + 1.0866 gms. pyrolusite . . " t c.c. 0.5N KMnO 4 .'. 1.0866 gms. pyrolusite require T -t c.c. 0.5N KMnO 4 corresponding to (T t) X 0.02173 gm. Mn0 2 and in percentage (T-t)X 0.02173X100 1.0866 = 2(T-t)% Mn0 2 . The ferrous sulphate solution is prepared as follows: 200 c.c. of concentrated' sulphuric acid are slowly poured, with stirring, into 500 c.c. of water and while the mixture is still hot 100 gms. of powdered FeSO 4 -7H 2 O crystals are added; on stirring, solution should take place within a few minutes. The solution is finally diluted to one liter, and when cold is ready for use. (6) The Oxalic Acid Method of Fresenius-Will, Modified by Mohr* About 0.4 gm. of finely powdered pyrolusite, which has been N dried at 100, is heated on the water-bath with 50 c.c. -=- oxalic o acid and 20 c.c. sulphuric acid (1:4) until no more black particles remain undissolved. The solution is diluted with 200 c.c. of hot N water and titrated with KMn0 4 solution. The reaction which o takes place between the manganese dioxide and the oxalic acid is expressed by the following equation: MnO 2 + H 2 S0 4 + H 2 C 2 4 = MnSO 4 + 2CO 2 + 2H 2 O. 1 c.c. ^ KMnO 4 = 0.0087 gm. MnO 2 . o * Fresenius-Will carried out the analysis in an alkalimeter and determined the CO 2 evolved by loss in weight. 626 VOLUMETRIC A 'NA LYSIS. 8. Determination of Formic Acid (Lieben).* 1000 c.c. N. KMnO,=**OH= **- 2 = 13 .80 gms. HCOOH.f In cold acid solutions permanganate reacts only slowly with formic acid, while in a hot solution the latter is lost by volatiliza- tion, so that the titration in open vessels is impossible; in alkaline solutions, on the other hand, the oxidation takes place readily and quantitatively in the cold : KMnO 4 + 3HCOOK = K 2 CO 3 + KHCO 3 + 2MnO 2 + H 2 O. Procedure. The formic acid is neutralized by an excess of sodium carbonate, and permanganate is run into the hot % sodium formate solution until the clear liquid above the precipitate is colored reddish. 9. Analysis of Nitrous Acid (Lunge). 1000 c.c. N. KMnO^^l^^^i =23.51 gms. HNO r On account of the volatility of nitrous acid, the aqueous solu- tion of the nitrite, or the solution of nitrous acid in concentrated sulphuric acid (nitrose), is measured from a burette into a known amount of permanganate solution, which has been made acid with sulphuric acid, diluted to a volume of about 400 c.c. and warmed to 40 C. The nitrous acid is thereby oxidized to nitric acid : 2KMnO 4 + 5HNO 2 + 3H 2 S0 4 = K 2 SO 4 + 2MnS0 4 + 3H 2 O+ 5HNO 3 , and the decolorization of the solution shows the end-point. Toward the end the nitrous acid must be added slowly, for the change from red to colorless requires some time. 10. Analysis of Hydrogen Peroxide. 1000 c.c. N.KMn0 4 = H & 8 =? = 17.01 gms. H 2 O t . Ten cubic centimeters of commercial 3 per cent, hydrogen peroxide are placed in a 100-c.c. measuring-flask, diluted up * Monatshefte, XIV, p 746, and XVI, p. 219. f In reality, the normal solution of formic acid would contain i (not T 8 ^) the molecular weight. See foot-note to page 612. t The titration is made in hot solution because the manganous acid formed does not settle well from a cold solution. ANALYSIS OF HYDROGEN PEROXIDE. 627 to the mark with water, and, after thoroughly mixing, 10 c. c. ( = 1 c.c. of the original solution) are placed in a beaker, and diluted with water to a volume of 300 to 400 c.c. After adding 20 to 30 N c.c. of sulphuric acid (1:4), the solution is titrated with KMnO 4 until a permanent pink color is obtained. The following reac- tion takes place: 2KMn0 4 + 5H 2 O 2 + 4H 2 S0 4 = 2KHSO 4 + 2MnSO 4 + 8H 2 O+ 5O 2 . Frequently it happens that the first drop of the permanganate causes a permanent coloration of the solution. This shows that either not enough sulphuric acid is present, or else there is no more hydrogen peroxide left in the solution. In this case a little more sulphuric acid is added, when if the coloration still remains the preparation is surely spoiled, as can be shown by the titanic or chromic acid tests (cf. Vol. I, p. 42). The amount of hydrogen peroxide is expressed either as per cent, by weight or as per cent, by volume. Example. 10 c.c. of the above-mentioned dilute solution of hydrogen peroxide ( = 1 c.c. of the original solution) required N 17.86 c.c. KMnO 4 solution, corresponding to 17.86X0.001701=0.03038 gm. H 2 O 2 . As the specific gravity of the original hydrogen peroxide solution can be assumed to be 1, it therefore contains 3.04 per cent. H 2 O 2 . When expressed in "per cent, by volume" the result shows how many cubic centimeters of oxygen can be obtained from 100 c.c. of the solution. In this case 100 c.c. of the hydrogen peroxide solution con- tain 3.04 gms. of H 2 2 and, on being decomposed, 1 gm.-mol. H 2 O 3 sets free 1 gm.-at. O : H 2 2 =H 2 0+0 34.02 = 18.02+16, or 11195 c.c. of oxygen at C. and 760 mm. pressure; conse- quently 3.04 gms. H 2 O 2 will evolve 34.02: 11195 = 3.04:x 3.04X11200 x = oT~no = 1000 c.c. oxygen measured under standard con- o4.U.Z ditions of temperature and pressure. 628 VOLUMETRIC ANALYSIS. 100 c.c. of the commercial hydrogen peroxide, therefore, will evolve 1000 c.c. of oxygen, i.e., ten times its own volume. This is somewhat anomalously designated as hydrogen peroxide of 10 per cent, by volume, 100 c.c. 3 per cent, hydrogen peroxide = 10 per cent, by volume. 100 c.c. 6 " " =20 " " " 100 c.c. 9 " " " =30 " " il " ii. Analysis of Barium Peroxide. 1000 c.c. N. KMnO 4 = -- 1 = - = 84.70 gms. B a O 2 . About 0.2 gm. of the substance is weighed into a 400 c.c. beaker, covered with 300 c.c. of cold water, and treated, under constant stirring, with 20-30 c.c. of hydrochloric acid (1:5). When all the BaCb has dissolved, the solution is titrated with 0.1 N. KMn0 4 . The addition of H 2 SO 4 is not advisable, as the precipitated BaSC>4 is likely to enclose some BaO2 which will then escape the titration. Another method for the analysis of BaC>2 has been proposed by Kassner.* 12. Analysis of Potassium Percarbonate. 1 Q8 2 1000 c.c. N. KMnO 4 = - 1 - A - 6 = - - =99.10 gms. K 2 C 2 O fl . 0.25 gm. potassium percarbonate is weighed out into 300 c.c. of cold, dilute sulphuric acid (1:30), in which it dissolves with * Arch. Pharm., 228, 432. ANALYSIS OF PERSULPH4TES. 629 violent evolution of carbon dioxide and formation of an equiva- lent amount of hydrogen peroxide : and the latter is titrated with potassium permanganate. 13. Analysis of Persulphates (Persulphuric Acid, H 2 S 2 Og). ^ ( 97.08 gms. H 2 S 2 O 8 1000 c.c. N.KMn0 4 = - 1H.4 " (NH 4 ) 2 S 2 O 8 135.2 " K 2 S 2 8 A solution of persulphuric acid does not reduce permanga- nate, nor does it react with titanic acid ; on the other hand it oxi- dizes ferrous salts immediately in the cold to ferric salts, and by means of this behavior it can be easily determined. The ammo- nium and potassium salts are now commercial products, and are analyzed as follows : About 0.3 gm. of the salt is weighed out into a flask fitted with a Bunsen valve, the air is replaced by carbon dioxide, 30 c.c. of a freshly titrated solution of ferrous sulphate are added and then 200 c.c. of hot water; the flask is closed and its contents rotated. The salt dissolves without difficulty, and the ferrous sulphate is oxidized: Fe 2 (S0 4 ) 3 + H 2 SO 4 . After all of the salt has dissolved, the contents of the flask are cooled by placing the flask in cold water, and the excess of N ferrous salt is titrated with KMn0 4 .* * The ferrous sulphate must be added to the persulphate, and then the hot water. If the hot water is added first, the persulphate is decomposed somewhat and the results obtained will be low. 630 VOLUMETRIC ANALYSIS. In this way it is found that: N 30 c.c. ferrous sulphate solution require T c.c. KM"nO 4 solution. N 30 c.c. ferrous sulphate+agm. persulphate require t c.c. KMnO 4 solution. Consequently a gm. of persulphate correspond to (T t) c.c. In the case of the potassium salt, since 1000 c.c. N. KMn0 4 = 135.2 gms. K 2 S 2 O 8 , and 1 c.c. ^KMnO 4 = 0.01352 gm. K 2 S 2 8 , we have: (Tt) XO. 01352 gm. K 2 S 2 O 8 in a gm. of the commer- cial salt, or in per cent. : 1.352(7^-0 - =per cent. K 2 S 2 O 8 . With the ammonium salt the factor becomes 0.01141 instead of 0.01352. The ferrous sulphate necessary for this determination is pre- pared by roughly weighing out 30 gms. of crystallized ferrous sulphate (FeSO 4 + 7H 2 O), dissolving it in 900 c.c. of water, and diluting to 1000 c.c. with pure concentrated sulphuric acid. Persulphates may also be analyzed very satisfactorily by means of oxalic acid.* When a sulphuric acid solution of a persulphate is treated with oxalic acid alone, there is no perceptible reaction. On adding a small amount of silver sulphate as catalyzer, however, a lively evolution of carbon dioxide takes place, and at the water- bath temperature the reaction is soon completed. H 2 S 2 8 + H 2 C 2 O 4 = 2H 2 SO 4 + 2CO 2 . * R. Kempf, Ber., 38, 3965 (1905). DETERMINATION OF HYDROXYLAMINE. 031 The excess of the oxalic acid can be titrated with perman- ganate. Procedure. About 0.5 gm. of the persulphate is placed in a 400-c.c. Erlenmeyer flask, 50 c,c. of tenth-normal oxalic acid solu- tion, and a solution of 0.2 gm. silver sulphate in 20 c.c. of 10 per cent, sulphuric acid are added, and the mixture is heated on the water bath until the evolution of carbon dioxide ceases; this requires not more than 15 or 20 minutes. The solution is then dilated to about 100 c.c. with water at about 40 and titrated with tenth-normal permanganate. I4 Determination of Hydroxylamine (Raschig).* 1000 c.c. N. KMnO 4 = = ^ = 16.52 gms. NH 2 OH. Principle. Hydroxylamine is oxidized in hot acid solutions by means of ferric salts to nitrous oxide and water: 2NH 2 OH + O 2 = N 2 O + 3H 2 O, and an equivalent amount of ferrous salt is formed: 2NH 2 OH + 2Fe 2 (SO 4 ) 3 = 4FeSO 4 + 2H 2 S0 4 + H 2 O + N 2 O. The amount of ferrous salt is determined by titration with - potassium permanganate. Procedure. About 0.1 gm, of the hydroxylamine salt is placed in a 500-c.c. flask and dissolved in a little water, 30 c.c. of a cold saturated solution of ferric-ammonium alum are added, and 10 c.c. of dilute sulphuric acid (1:4). The contents of the flask are heated * Ann. d. Chem. und Pharm., 241, p. 190. 632 VOLUMETRIC ANALYSIS. to boiling and kept at this temperature for five minutes, after which the solution is diluted with distilled water to a volume of about 300 c.c. and immediately titrated with permanganate solution. Remark. If only slightly more than the theoretical amount of the ferric salt is added, the oxidation of the hydroxylamine does not take place entirely in accordance with the above equa- tion, but part of the substance is oxi3ized to nitric oxide: 2NH 2 OH+ 30 = 3H 2 O + 2NO, so that it is then impossible to obtain exact results. 15. Determination of Hydroferrocyanic Acid (de Hae'n).* 1000 c.c. N.KMnO 4 = l mol. K 4 Fe(CN) 6 = 368.3 gms. K 4 Fe(CN) 6 . Principle. By oxidation in acid solution, hydroferricyanic acid is formed from hydroferrocyanic acid: 2H 4 Fe(CN) 6 + O = H 2 O+ 2H 3 Fe(CN) 6 . This procedure is chiefly used for the analysis of potassium ferrocyanide (yellow prussiate of potash), so that the concentra- tion of the permanganate solution is expressed in terms of this salt. Procedure. 0.9 gm. of the salt to be analyzed is dissolved in 100 c.c. of water, 10 c.c. of dilute sulphuric acid are added, and this solution is titrated . in a porcelain dish with permanganate until a permanent pink color is obtained. It is not easy to determine the end-point. On acidifying, the solution of the ferrocyanide becomes milky with a bluish tinge, and on the addi- tion of permanganate at first a yellow shade is obtained, after- wards becoming green, and finally on the addition of more perman- ganate the color changes to pink. On account of the difficulty in determining this point, de Hae'n recommends that the perman- ganate be standardized against pure potassium ferrocyanide solu- tion (K 4 Fe(CN) 6 +3H 2 0). * Ann. d. Chem. und Pharm., 90, p. 160. DETERMINATION OF HYDROFERRICYANIC ACID, ETC. 633 1 6. Determination of Hydroferricyanic Acid. 1000 c.c. N. KMnO 4 = l mol. K3Fe(CN) 6 = 329.2 gms. K 3 Fe(CN) v Principle. The potassium ferricyanide is reduced in alkaline solution to potassium ferrocyanide, and the latter is titrated with permanganate. Procedure. In a 300-c.c. flask, 6.0 gms. of the ferricyanide are dissolved in water, the 'solution made alkaline with potassium hydroxide, heated to boiling, and an excess of a concentrated ferrous sulphate solution is added. At first yellowish-brown ferric hydroxide is precipitated, later black ferrous-ferric hydrox- ide is formed, and this shows the completion of the reaction. After cooling, the contents of the flask are diluted with water up to the mark, filtered through a dry filter (after thoroughly mix- ing), and 50 c.c. of the filtrate * ( = 1 gm. of the substance) are taken N for the titration with KMnO 4 solution. 17. Determination of Chloric Acid. XT T ^ RC1O, \ 20.44 gms. KCIO. )0 c.c. N. KMn0 4 = ' - j 1?J4 * About 5 gms. of potassium chlorate, or 4 gms. of the sodium salt, are dissolved in water, and the solution diluted to 1 liter. After thoroughly mixing, 10 c.c. are placed in a flask fitted with a Bunsen valve and the air expelled from the flask by a current of carbon dioxide. After this 50 c.c. of a freshly-standardized solution of ferrous sulphate (prepared as described on p. 630) are added, and the solution boiled ten minutes. The following reaction takes place: KC1O 3 + 6FeS0 4 + 3H 2 SO 4 = KC1+ 3Fe 2 (SO 4 ) 3 + 3H 2 O. After cooling the solution is diluted with cold distilled water, 10 c.c. of manganous sulphate solution are added (cf. p. 607), and the excess of the ferrous sulphate is titrated with potassium permanganate. We find that: *The first ten or fifteen cubic centimeters of the filtrate should be discarded. 634 VOLUMETRIC ANALYSIS. N 50 c.c. ferrous sulphate required T c.c. KMnO 4 sol. 50 c.c. " " + 10 c.c. chlorate sol. " t c.c. ^ " ~a ~N 10 c.c. chlorate solution = - gm. substance = (T t)c.c. KMnO 4 " J- \)\) .LvJ For the analysis of potassium chlorate a gm. of the substance contain (T t)X 0.2044 gm. KC1O 3 , and 'the per cent, present is 20.44 X(T-t) = per cent. The calculation for sodium chlorate is analogous. 18. Determination of Nitric Acid (Pelouze-Fresenius). RNO { 21.01 gms. HNO, 1000 c.c. N. KMn0 4 = ^^ s = \ 28.34 " NaNO, L S3.70 " KNO 3 This method depends upon the fact that on heating a nitrate in the presence of considerable hydrochloric acid and ferrous chloride the latter is oxidized to ferric chloride and the nitric acid is reduced to nitric oxide: 2KNO 3 + 6FeCl 2 + 8HC1 = 2KC1+ 2NO+ 4H 2 O+ 6FeCl 3 . As a measure for the amount of nitrate reduced we have : 1. The excess of ferrous salt. 2. The ferric salt produced. 3. The nitric oxide formed. The method of Schlosing-Grandeau described on p. 456 is based upon the measurement of the nitric oxide formed. C. D. Braun * estimates the amount of ferric salt formed, while Pelouze and Fresenius determine the amount of ferrous salt not used up in the reduction of the nitric acid. Procedure. A weighed amount of iron wire (about 1.5 gms.) is placed in a long-necked flask, and the air expelled by passing a current of pure carbon dioxide through it for two or three minutes. After this 30 to 40 c.c. of pure, concentrated hydro- chloric acid are added and the flask is placed in an inclined posi- tion and closed by means of a rubber stopper through which tubes pass so that a current of carbon dioxide can be conducted * Journ. f. prakt. Chem., 81 (1860), p. 421. DETERMINATION OF NITRIC ACID. 635 through the flask. The solution is heated on the water-bath in this atmosphere of carbon dioxide until the iron has com- pletely dissolved, when the solution is allowed to cool in a cur- rent of the gas. Meanwhile about 0.25 to 0.3 gm. of the nitrate is weighed out in a small glass tube closed at one end; this is thrown into the acid solution of the ferrous sulphate and the flask quickly closed again. The flask is then once more placed in its inclined position upon the water-bath and heated for fifteen minutes ; while the current of carbon dioxide is continually passed through it. The tube through which the gas leaves the flask, during the whole operation, dips into a beaker filled with water so that there is no chance of any air getting back into the flask. After this the solution is heated to boiling and kept there until its dark color disappears and the yellow color of the ferric chloride becomes apparent. In order to make sure that the nitric oxide is entirely removed, the contents of the flask are boiled five minutes longer and then allowed to cool in the atmosphere of carbon dioxide. When cold the solution is poured into a beaker, the flask washed out with a little boiled water, the solu- tion is diluted to a volume of about 400 to 500 c.c., 10 c.c. of manganese sulphate solution are added, and the unoxidized iron N is titrated with KMnO 4 solution. 2 The amount of pure iron present in the wire used is deter- mined under the same conditions as prevailed during the pre- vious operation, using a smaller portion of wire but the same amount of acid, manganese sulphate, etc. The calculation is as follows: If a gm. of potassium nitrate and p gm. of the wire were taken for N the analysis, t c.c. of KMnO 4 were required to oxidize the excess 2i N of iron, and further p gm. of the wire require T c.c. of KMnO 4 solution, we have, then: N p gm. iron require T c.c. -= KMnO 4 solution N p gm. iron + a gm. saltpeter t c.c. KMnO 4 N and a gm. saltpeter = (T t) c.c. KMnO 4; 636 VOLUMETRIC ANALYSIS. so that a gm. of saltpeter contain (Tt) X0.01685 gm. KNO 3 , and in per cent. (!F-OX 1.685 = per cent. Remark. This method gives results just as accurate as those obtained by the method of Devarda, but the latter determination is much easier to carry out. The determination becomes sknpler if the contents of the iron wire is assumed to be 99.7 per cent. Fe and the second titra- tion thus done away with. It does not take long to make the analysis of the wire, however, and it is advisable to do it. Instead of titrating the excess of the ferrous salt with potassium perman- ganate solution, a solution of potassium dichromate may be used. For the determination of the ferric salt formed, cf. p. 681. 19. Determination of Vanadium. N V O 1 89 4. 1000 c.c. ^ KMn0 4 = -^ 6 ==^=9.12 gms. V 2 O 5 - Sulphur dioxide is conducted into the boiling solution of an alkali vanadate containing sulphuric acid until the solution appears a pure blue; by this means the vanadic acid is reduced to vanadyl salt: The boiling is continued and a current of carbon dioxide is passed through the solution until the escaping gas will no longer decolorize a solution of potassium permanganate, showing that the excess of the sulphur dioxide has been expelled. The hot solution is then titrated with potassium permanganate until a permanent pink color is obtained. The end-point is easily recog- nized only when the solution is hot. This accurate determina- tion is used for the analysis of vanadium in iron and steel, or in ores. (Cf. p. 310.) * Of course the calculation can be made from the amount of iron oxi- ized. In that case: Fe:iKNO, = (p-*X0.02793) :x (p-tX 0.02793). KN0 3 x= - -- - gms. KNO 3 in a gm. of substance, and in per cent. 100(- so that after shaking n times with fresh portions of carbon bisulphide, the amount of iodine remaining in the water would be : (3) Xn=x 1 -^- T - .... ) gms. iodine. Assuming that in the analysis 0.005 gm. of iodine was dissolved in 10 c.c. of water and that this solution was shaken once with 1 c.c. of carbon bisulphide, then according to equation (1 0.005--=0.0001 gm. iodine l+ 400 would r.< HT -,r,ived in the water, or an amount that can be neglected. 660 VOLUMETRIC ANALYSIS. just above the surface of the hot bromide solution, so that as little chlorine as possible is lost by evaporation. 7. Determination of Iodine and Bromine in Mineral Waters. According to the amount of halogen present, from 5 to 60 liters of water are taken for the analysis. The amount of bromine and iodine present is usually small compared with the chlorine, so that the residue obtained by the evaporation of a large amount of water cannot be used directly for the analysis, but by partial crystallization a mother-liquor rich in bromide and iodide must first be obtained. Procedure. The water is placed in a large porcelain evaporating- dish, a liter at a time, and if not already alkaline,* enough pure sodium carbonate solution is added to make it distinctly so, and the water is evaporated to about one-fourth of its original volume. This causes the separation of some calcium and magnesium car- bonates in the presence of hydroxides of iron and manganese, while all of the halogen salts remain in solution. The residue is filtered off and thoroughly washed with water. The filtrate is further concentrated until salts begin to crystallize out, and the hot solution is then poured into three times its volume of absolute alcohol ; this causes the greater part of the sodium chloride and other undesired salts to precipitate. After standing twelve hours, the alcoholic liquid is filtered and the residue washed five or six times with 95 per cent, alcohol. The alcoholic solution, which contains all of the iodine and bromine with considerable chlorine in the form of the alkaline salts, is treated with five drops of concentrated potassium hydroxide solution and almost all of the alcohol distilled off, while a current of air is passed through the solution by means of a capillary tube reaching to the bottom of the liquid in the distilling-flask. The residue from the distillation is further concentrated until salts again begin to crystallize out and the precipitation with alco- hol is repeated. The alcohol is again distilled off, but this time with the addition of only one or two drops of potassium hydroxide * The solution is alkaline if after the addition of phenolphthalei'n the eolution turns red on boiling. DETERMINATION OF IODINE AND BROMINE. 66 1 solution. According to the amount of salts present in solution this operation is repeated from three to six times. The final filtrate, after the alcohol has been distilled, off, is placed in a plati- num dish, evaporated to dryness, the dish covered with a watch- glass, and the residue gently ignited to destroy organic matter. The residue from the ignition is dissolved in a little water, the carbonaceous material filtered off,* the solution slightly acidified with dilute sulphuric acid, the iodine liberated by the addition of one or two drops of "nitrose," and titrated with sodium thio- sulphate, after shaking with chloroform, as described on p. 657.f The bromine is determined in the aqueous solution obtained after the extraction of the iodine with chloroform. The acid solution is made alkaline by the addition of sodium carbonate solution, two drops of a saturated sugar solution are added, and the solution evaporated to dryness in a platinum dish. With a watch-glass upon the dish, the residue is gently ignited in order to destroy the sugar and the excess of nitrite. J After this has been accomplished the residue is dissolved in water, filtered, acidified slightly with sulphuric acid, and the bromine titrated with chlorine water as described on p. 659. Remark. If sufficient mineral water is available it is better to divide the mother-liquor containing the bromide and iodide into two portions; in one portion the iodine i determined as before, while in the other the bromine and iodine are determined by titration with chlorine water. 8. Analysis of Peroxides (Bunsen). All peroxides of the heavy metals, which evolve chlorine on treatment with hydrochloric acid, can be determined with great * If the filtrate is not completely colorless, it is evaporated and again ignited. f Lecco determines the iodine colori metrically (Zeitschr. f. anal. Chem., XXXV, p. 318). t The addition of the sugar causes the nitrite to be dertroyed at a lower temperature than would otherwise be the case, and the danger of losing bromine by volatilization is avoided. As the chlorine water was standardized against bromine, an amount of the latter equivalent to the iodine present is deducted from the amount represented by the chlorine water used; the difference shows the bromine present. 662 VOLUMETRIC ANALYSIS. accuracy by conducting the chlorine into potassium iodide solu- tion and titrating the deposited iodine with sodium thiosulphate or arsenious acid solution. It is only necessary to make sure that the chlorine is allowed to act upon the potassium iodide without loss. For all such determinations, Bunsen employed the apparatus shown in Fig. 94. The small decomposition-flask of about 40 c.c. capacity has a ground-glass connection with the delivery- FIG. 94. tube * and is held firmly in place by means of rubber rings, as at a. The lower end of the bent delivery-tube is drawn out into a not- too-small capillary. Procedure. The finely-powdered substance is placed in the small glass-stoppered weighing-tube (Fig. 94 5), which has a small piece of glass fused on the end, and weighed. The tube is then taken hold of by means of the glass at the bottom ? t intro- duced into the neck of an absolutely dry decomposition-flask, and the required amount of the substance is allowed to fall into it by carefully revolving the weighing-tube. On again weighing the tube, the amount of substance taken is determined. Hydro- chloric acid is now added (its concentration depends upon the nature of the substance), the delivery tubing is at once connected with the flask and introduced into the retort containing potassium iodide solution. By means of a tiny flame, the contents of the flask are heated to boiling and from half to two-thirds of the liquid * Instead of the ground-glass connection, Bunsen used a tube of the same size as the neck of the flask and connected them with rubber tubing, the two glass tubes being against one another. t By holding the tube in this way, deviations of weight, due to unequal warming, are avoided. DETERMINATION OF MANGANESE DIOXIDE. 66} is distilled over into the retort. In order to prevent the iodide solution from sucking back into the flask, the delivery-tube is taken out of the retort before removing the flame; the contents of the tube are then washed into the retort. The potassium iodide solution is poured into a large beaker, the retort washed out several times with a little water, and then with potassium iodide solution in order to remove any iodine which may remain adhering to the glass. The iodine is titrated with N sodium thiosulphate solution. In this way pyrolusite, chro- mates, lead peroxide, minium, eerie oxide, selenic, telluric, and molybdic acids may be analyzed. (a) Determination of Manganese Dioxide in Pyrolusite. 1000 c.c. L Na^SjO, solution = >2 = = 4.347 gms. MnO 2 . . How much pyrolusite shall be taken for the analysis? * If possible, an amount should be taken for analysis which will N not require more than one buretteful of the -^ Na 2 S 2 O 3 solution. We assume that the sample contains 100 per cent, of MnO 2 , and calculate how much of the latter would correspond to 50 c.c. of IffNa&O,: N 1 c.c. solution: 0.004347 gm. Mn0 2 = 50:rc; x = 50X0.004347 = 0.2173 gm. MnO 2 . Consequently for the analysis about 0.2 gm. of the substance is taken, which has been dried at 100 C. To this 25 c.c. of hydro- chloric acid (1:2) are added and the analysis is made as described above. * This is applicable to almost every volumetric analysis. To insure the most accurate results, the concentration of the standard solution and the weight of substance taken for analysis should be so chosen that between 35 and 50 c.c. of the reagent are used in the final titration. In this way the errors in determining the end-point, reading the burette, etc., will not influence the result appreciably. [Translator.] 664 VOLUMETRIC ANALYSIS. The calculation is based upon the following equations; MnO 2 + 4HC1 = 2H 2 O + MnCl 2 f CL, 2Cl=2I=lMnO 2 , lCl=lI = iMnO 2 =43.47 gms. The amount of substance taken for analysis = a gms., and the N Na-jSjjOj solution used for the titration of the iodine =t c.c. Then a:*X0.004347 = 100:.c; 0.4347 -Z x = -- =per cent. MnO2. . The determination of chromates, lead peroxide, and selenic acid is carried out in the same way, except that concentrated hydrochloric acid is used for the decomposition. (b) Determination of Telluric Acid. If the telluric acid is present as the hydrous acid (H 2 TeO 4 +2H 2 O) or as tellurate, the analysis is performed in the same way as with selenic and chromic acids. If, however, the tellurium is present as the anhydrous acid or as the anhydride, the method must be modified, for these substances are scarcely attacked by concen- trated hydrochloric acid. They are placed in the decomposition- flask, dissolved in a little concentrated potassium hydroxide,* and to the tellurate solution thus obtained the concentrated hydrochloric acid is added ; the reduction then is accomplished without difficulty : K 2 TeO 4 -f 4HC1 = 2KC1 -f H 2 TeO 2 + H 2 O + C1 2 . According to this equation TV 1 97 ^ 1C1=1I = = i = 63.75 gms. Te. * The solution could not be effected by using sodium hydroxide. DETERMINATION OF CERIC OXIDE AND VANADIC ACID. 665 , c) Determination of Ceric Oxide. 1000 c.c. -j iodine solution =^ 2 = ^p = 17. 225 gms. CeO 2 . Ceric oxide when mixed with considerable lanthanum and di- dymium oxides is reduced by distillation with concentrated hydro- chloric acid: 2CeO 2 + 8HC1 = 4H 2 O + 2CeCl 3 + C1 2 . If, however, the mixture contains but little of the two last substances, or if it is pure eerie oxide, the heating with concen- trated hydrochloric acid is of no avail; the eerie oxide will not dissolve. In the presence of hydriodic acid, however, the reduction takes place readily, so that it is only necessary to add 2 gms. of potassium iodide to a weighed amount of the substance (0.67-0.68 gm.) in the decomposition-flask, and then, after the addition of hydrochloric acid, violet vapors of iodine can be dis- tilled from the solution: 2CeO 2 + 2KI + 8HC1 = 2KC1 + 2CeCl 3 + 4H 2 + 21. Often there will be so much iodine given off that the solid is likely to stop up the tube and the flask will often explode. To prevent this, the end of the delivery-tube is not drawn out into a capillary, but at the bottom an opening of about 4 mm. in diame- ter is left. During the operation, the flame must be protected from air-currents, for otherwise there is danger of liquid sucking back from the retort. (d) Determination of Vanadic Acid.J 1000 c.c. iodine solution = -| -*= ' =9.12 gms. V 2 O 5 . By boiling vanadic acid, or one of its salts, with concentrated hydrochloric acid, the vanadium is reduced with evolution of chlorine. Unfortunately, this reaction cannot be used for the determination of vanadic acid, for the amount of chlorine evolved 666 VOLUMETRIC AN 'A LYSIS. depends upon the concentration of the vanadium solution; the vanadium is not reduced to a definite oxide. On the other hand, by means of hydrobromic acid,* vanadic acid is reduced to a blue vanadyl salt: If the free bromine is absorbed in potassium iodide, and the liberated iodine titrated with sodiurrf thiosulphate, a sharp deter- mination of the vanadium will be obtained. To carry out this analysis, about 0.3-0.5 gm. of the vanadate, together with 1.5 to 2 gms. of potassium bromide, is placed in the decomposition- flask of the Bunsen apparatus (Fig. 94, p. 662), 30 c.c. of con- centrated hydrochloric acid are added, and distillation is effected as before. The decomposition is always complete when the liquid in the flask is a pure blue. If hydriodic acid is used instead of hydrobromic acid, the vanadic acid is reduced still further, almost to V 2 O 3 .f In fact, a complete reduction to the latter oxide can be accomplished if potassium iodide, concentrated hydrochloric acid, and 1 or 2 c.c. of syrupy phosphoric acid are added and the liquid distilled until no more vapors of iodine are evolved. According to Steffan, this will always be the case when the liquid is reduced to one-third of its original volume. (e) Determination of Molybdic Acid. 1000 c.c. ^ Na 2 S 2 3 =^= ~ : =14.4 gms. MoO 3 . The determination depends upon the fact that molybdic acid is reduced to molybdenum pentoxide by means of hydriodic acid with liberation of iodine: 2MoO 3 + 2HI = H 2 O + Mo 2 5 + 1 2 . Remark. This method finds no practical application on account of the fact that it is difficult to obtain a quantitative reduction in * Holverscheidt, Dissertation, Berlin, 1890. t Friedheim and Eulcr, Berichte, 28 (1895), 2067. j Ibid., 28 (1895), 2067, and ?9 (1896), 2981. DETERMINATION OF V//NAD1C AND MOLYBDIC ACID. 667 accordance with the above equation. Gooch and Fairbanks * found that if a solution containing molybdic acid is distilled in the Bunseri apparatus with potassium iodide and hydrochloric acid, until iodine vapors are no longer visible and the solution is a light green ; too little iodine is obtained. On the other hand, if the distillation is continued still further, they found that the reduction goes on and more iodine is obtained than corresponds to the above equation. Steffan,| who tested the method in the author's labo- ratory, obtained results agreeing with those published by Gooch and Fairbanks. By means of hydrobromic acid, molybdic acid is not reduced. (/) Determination of Vanadic and Molybdic Acids in the Presence of One Another. According to StefTan, these two acids may be determined very accurately when present together. The vanadic acid is determined, according to Holverscheidt, by distillation with potassium bromide and concentrated hydrochloric acid, absorp- tion of the bromine in potassium iodide solution, and titration of the liberated iodine (cf. p. 666). The contents of the distilla- tion flask, in which the vanadium is present as vanadyl salt and the rcolybdenum as molybdic acid, are treated with hydrogen sulphide in a pressure-flask, and the precipitated molybdenum sulphide is filtered through a Gooch crucible, and weighed as MoO 3 , as described on p. 286, The results obtained by this method are perfectly satisfactory. As molybdie acid is unattacked by hydrobromic acid, but is reduced to Mo 2 O 5 with separation of iodine by means of hydri- odic acid, Friedheim and Euler proposed the following method for the determination of vanadic and molybdic acids when present together: The mixture of the two acids is distilled as before with potassium * Gooch and Fairbanks, Zeitschr. f. anorg. Chem., XIII (1897), 101, and XIV, 317. t Steffan, Inaug. Dissertation, Zurich, 1902. 668 VOLUMETRIC ANALYSIS. bromide and hydrochloric acid and the vanadium thereby reduced to the tetroxide compound V 2 5 + 2HBr - H 2 + V 2 O 4 + Br 2 , with separation of two atoms of bromine which are determined iodimetrically. To the cold solution remaining in the distilling- flask, potassium iodide, hydrochlorjp acid, and syrupy phosphoric acid are added, and the distillation continued until no more iodine is given off and the solution is a light green. By means of this second reduction the vanadium tetroxide is supposed to be reduced to V 2 O 3 , V 2 4 +2HI-H 2 0+V 2 3 +I 2 , and consequently more iodine is liberated by the vanadium. Furthermore, according to Friedheim and Euler, the molybdenum is reduced to Mo 2 O 5 : 2Mo0 3 + 2HI = H 2 O + Mo,0 6 + 1^ If, therefore, the amount of iodine corresponding to the first titration' is deducted from the amount obtained in the second, the difference should correspond to the amount of molybdenum present. But Gooch and Fairbanks have shown that this is not the case.* The error in the method lies in the fact that the vanadic acid is only reduced completely to V 2 3 when the solution is distilled to one-third of its original volume. In this case, however, the molybdenum is reduced further than corresponds to the formation of Mo 2 O 5 ; too much iodine is liberated and too high a value is obtained for the molybdic acid present. On the other hand, if after the addition of the potassium iodide the liquid is only distilled until the iodine vapors cease to appear and the solution is a light green, the vanadium is not completely reduced to V 2 O 3 , and then a too low value for the molybdenum is obtained. * The results of Gooch and Fairbanks have been confirmed in every re- spect by Steffan. DETERMINATION OF HYPOCHLOROUS ACID. 669 9. Analysis of Chlorates. This is carried out the same way as the analysis of pyrolusite (cf . p. 663) : KC1O 3 -f 6HC1 = KC1 + 3 H 2 O + 3C1 2 r N gm. = at. iodine= 1000 c.c. y^ Na 2 S 2 O 3 solution KC1O 3 122.6 = 2.043 gms.KClO 3 60 " 60 Many oxidizing agents can be determined iodimetrically with- out previous distillation with hydrochloric acid. For other methods of analyzing chlorates iodimetrically, consult H. Ditz, Chem.-Ztg. 1901, 727 and Luther and Rutter, Z. anal. Chem. 46, 521 (1907). 10. Determination of Hypochlorous Acid. This determination is made use of in the analysis of chloride of lime. Procedure. Into a tared weighing-tube about 5gms.of " chloride of lime " are introduced, and the stoppered tube is weighed. Its contents are washed into a porcelain dish, rubbed to a paste by means of a pestle, and then transferred without loss to a 500-c.c. measuring-flask, diluted up to the mark with water and well shaken. Of this turbid solution, 20 c.c. are run into 10 c.c. of 10 per cent, potassium iodide solution, and after acidifying with hydrochloric N acid the iodine set free is titrated with Na 2 S 2 O 3 . The result is expressed in per cent, of chlorine. Remark. If the " chloride of lime " contained calcium chlo- rate it will be partially reduced by hydrochloric acid and potassium iodide with liberation of iodine, and consequently the results obtained for hypochlorite chlorine (bleaching chlorine) will be too high. In this case the hypochlorite is best determined by a chlorimetric process with arsenious acid (see p. 701). 670 VOLUMETRIC ANALYSIS. ii. The Analysis of lodates. 1000 c.c. Na 2 SA= = l = 2.932 gms. HIO 3 . The solution of the iodate is allowed to run into an acid solution containing an excess of potassium iodide. Iodine is set free according to the equation and the iodine is titrated with thiosulphate solution as described on p. 647. 12. The Analysis of Periodates. 1000 c.c. Na 2 SA= 4 = 2.399 gms. HI0 4 . The analysis of periodates is carried out exactly as with iodates; the reaction that takes place is KIO 4 + 7KI + 8HC1 = 8KC1 + 4H 2 O + 4I a . 13. Analysis of a Mixture of Iodate and Periodate.* If a neutral or slightly alkaline solution of an alkali periodate is treated with a solution of potassium iodide, the following reaction takes place : The liberated iodine is titrated with tenth-normal arsenious acid (not with sodium thiosulphate) ; in a neutral solution the iodate does not react with potassium iodide. For the analysis of a mixture of iodate and periodate, the following procedure is used: In one sample the iodate -f periodate is determined by adding 'the solution of the substance to an acid solution containing an excess of potassium iodide and the liberated iodine is titrated with sodium thiosulphate solution; A second sample of the substance is dissolved in water, a drop *E. Mullet- and O. Friedberger, Berichte, 1902, 2655. ANALYSIS OF IODIDES. 671 of phenolphthalein added, and the, solution is made just alkaline enough to give the pink color with phenolphthalein, adding alkali if the solution is acid and hydrochloric acid if the solution is strongly alkaline. To the barely alkaline solution, 10 c.c. of a cold, saturated solution of sodium bicarbonate are added and then an excess of potassium iodide ; the liberated iodine is at once titrated with tenth-normal arsenious acid.* Example. In a mixture of KIO 3 and KIO 4 weighing a grams, the iodine liberated on treatment with an acid solution of KI reacts with T c.c. of 0.1 N Na 2 S 2 3 and the same weight of sample liberates in alkaline solution only enough iodine to react with t c.c. of O.lN As 2 3 solution. By comparing the equations given under 12 and 13, it is evident that the periodate alone would react with 4t c.c. of O.lN Na 2 S 2 O 3 in acid solution. The amount of KI0 4 and KIO 3 present will be *X0.01150gm.= %KIO 4 , CF-40X0.003567 gm.= ( y - 4 *)><0.3567 % 14. Analysis of Iodides, f Method of H. Dietz and B. M. Margosches. 1000 c.c. KI0 3 = = '- 10.58 gm. iodine. The solution of the iodide is treated with an excess of tenth- normal potassium iodate solution, acidified with hydrochloric acid, a piece of calcite added, as suggested by Prince, { and boiled until all the iodine is expelled. The solution is allowed to cool, then an excess of potassium iodide is added, and the iodine now * The iodine cannot be titrated in the alkaline solution with sodium thiosulphate, and the iodine in the acid solution cannot be titrated with the arsenious acid. t Chem. Ztg., 1904, II, 1191. J Inaug. Dissert. Zurich, 1910. 672 VOLUMETRIC ANALYSIS. liberated, which corresponds to the excess of potassium iodate used, is titrated with tenth-normal Na 2 S 2 O 3 solution. From the equation KIO 3 + 5KI + 6HC1 = 6KC1 + 3H 2 O + 3I 2 , it is evident that five-sixths of the iodine liberated comes from the iodide. If, therefore, T c.c. of O.lN KIO 3 solution were added and t c.c. of O.lN Na 2 S 2 O Vere used for titrating the excess of KIO 3 , then there is present (T-t) X0.01058gm. iodine as iodide. 15. Determination of Copper with Potassium Iodate.* Potassium iodate in dilute hydrochloric acid solution is reduced by potassium iodide to free iodine (cf. p. 670) : KIO 3 + 5KI + 6HC1 = 6KC1 + 3I 2 + 3H 2 O, but if the solution is strongly acid with hydrochloric acid and an excess of the iodate is added, the iodine is oxidized to IC1: 2I 2 + KI0 3 + 6HC1 = KC1 + 5IC1 + 3 H 2 O, and in this case the whole reaction may be expressed by the equation : The ICl is not very stable, and is at once reduced to free iodine in the presence of any oxidizable substance. L. W. Andrew t has shown that quite a number of reducing substances, such as free iodine, iodides, arsenites, and antimonites, can be titrated with potassium iodate very exactly, by taking advantage of the fact that when the reducing agent is present in excess free iodine is formed, which is oxidized quantitatively by more iodate, provided the proper amount of hydrochloric acid is present. Copper solutions are precipitated quantitatively * Jamieson, Levy, and Wells, Jour. Am. Chem. Soc., 30, 760 (1908). t /&*., 25,756 (1903). DETERMINATION OF COPPER. WITH POTASSIUM IODATE 673 by potassium thiocyanate and sulphurous acid as cuprous thio- cyanate, CuSCN, and Parr* has estimated copper quantita- tively by titrating this precipitate with permanganate. The oxidation is, however, simpler and more accurate when the titration is effected by potassium iodate, or biiodate. The reaction goes through the stage in which iodine is set free, but the latter is oxidized completely to iodine chloride upon the addition of more iodate : (a) 2CuSCN + 3 KIO 3 + 4HC1 ~ 2CuSO 4 + 1 2 + IC1 + 2HCN + 3KC1 + H 2 O. (6) 2I 2 +KIO 3 +6HC1=KC1+5IC1 + 3H 2 0, and the whole reaction is (multiplying (a) by 2 and adding (6)), (c) 4CuSCN + 7KI0 3 + 14HC1 = 4CuSO 4 + 7IC1 + 4HCN + 7KC1 + 5H 2 0. The potassium iodate solution is very stable and can be preserved for years if protected from evaporation. The standard solution used can be prepared by weighing out a known amount of the pure salt and dissolving to a definite volume, or the solution may be standardized against pure copper, carrying out the process as in an analysis. A convenient concentration is one-fifth of the formula weight. Procedure. To 0.5 gm. of the ore in a 200 c.c. flask, add 6 to 10 c.c. of strong nitric acid, and boil gently, best over a free flame, keeping the flask in constant motion and inclined at an angle of about 45, until the larger part of the acid has been removed. If this does not completely decompose the ore, add 5 c.c. of. strong hydrochloric acid and continue the boiling until the volume of liquid is about 2 c.c. Now add gradually and carefully, best after cooling somewhat, 12 c.c. of sulphuric acid (1 : 1), and continue the boiling until sulphuric acid fumes are evolved copiously. Allow to cool, add 25 c.c. of cold water, heat to * J. Am. Chem. Soc., 22, 685 (1900). 674 VOLUMETRIC ANALYSIS. boiling, and keep hot until the soluble sulphates have dissolved. Filter into a beaker, and wash the flask and filter thoroughly with cold water.* Nearly neutralize the filtrate with ammonia and add 10 to 15 c.c. of strong sulphur dioxide water. Heat just to boiling and add 5 to 10 c.c. of a 10 per cent, solution of ammonium thiocyanate, according to the amount of copper present. Stir thoroughly, allow the precipitate to settle for 5 or 10 minutes, filter on paper, and^wash with hot water until the ammonium thiocyanate is completely removed. Place the filter with its contents in a glass-stoppered bottle of about 250 c.c. capacity, and by means of a piece of moist filter-paper transfer into the bottle also any precipitate adhering to the stirring-rod and beaker. Add to the bottle about 5 c.c. of chloroform, 20 c.c. of water and 30 c.c. of concentrated hydro- chloric acid (the two latter liquids may be previously mixed). Now run in standard potassium iodate solution, inserting the stopper and shaking vigorously between additions. A violet color appears in the chloroform, at first increasing and then diminishing, until it disappears with great sharpness. The rapidity with which the iodate solution may be added can be judged from the color changes of the chloroform. In order to make another titration it is not necessary to wash the bottle or throw away the chloroform. Pour off two-thirds or three-fourths of the liquid in order to remove most of the pulped paper, too much of which interferes with the settling of the chloroform globules after agitation, add enough properly diluted acid to make about 50 c.c. and proceed as before. In this case, where iodine monochloride is present at the outset, the chloroform becomes strongly colored with iodine as soon as the cuprous thiocyanate is added, but this makes no difference with the results of the titration. * With substances containing appreciable amounts of silver a few drops of hydrochloric acid should be added before making this filtration, but not enough to dissolve any considerable amounts of the lead sulphate or antimonis oxide that may be present. DETERMINATION OF CHROMIUM IN CHROMITE. 675 1 6. Analysis of Soluble Chromates. A concentrated, acid solution of potassium iodide is treated with a weighed amount of the chromate, diluted with water, and the liberated iodine titrated. (Cf. standardization of sodium thiosulphate against potassium dichromate, p. 649.) i6a. Determination of Chromium in Chromite. About 0.2 gm. of the finely powdered chromite is intimately mixed with 2 gms. of sodium peroxide in a porcelain crucible. This crucible is placed inside a larger porcelain crucible and heated for fifteen or twenty minutes over a small flame.* At the end of this time, all the chromium will be converted into soluble sodium chromate. The crucible and its contents are placed in 100-200 c.c. of water, which is heated to boiling and kept at this temperature until the melt is completely disintegrated. The ferric oxide is then filtered off, the filtrate evaporated in a porce- lain dish nearly to dryness,f the residue taken up in as little water as possible, one or two grams of potassium iodide added, the solution diluted to about 400 c.c., and the free iodine titrated with tenth-normal thiosulphate solution: 1 c.c. Na 2 S 2 O 3 = 0.001733 gm. Cr. 17. Determination of Lead Peroxide. Method of Diehl, modified by Topf.% The analysis depends upon the fact that lead peroxide is re- duced by means of potassium iodide in acetic acid solution when considerable alkali acetate is present: Pb0 2 + 4HI = PbI 2 + 2H 2 + 1 2 . N After diluting with water the iodine is titrated with y^ Na 2 S 2 O 3 solution. * If the crucible is heated too hot, it is likely to be strongly attacked by the sodium peroxide. With care, a single crucible may be used for four or six determinations. t The evaporation to dryness is necessary to remove the last traces of per- oxide. t Diehl, Dingl. polyt. Journ., 246, p. 196, and Topf, Zeitschr. f. analyt. Chem., XXVI (1887), p. 296. 676 YOLU 'METRIC ANALYSIS. Procedure. About 0.5 gms. of the substance are dissolved with 1.2 gms. of potassium iodide and 10 gms. of sodium acetate in 5 c.c. of 5 per cent, acetic acid. The solution is diluted with water to a volume of 25 c.c. and titrated with sodium thiosulphate. Remark. Moist lead peroxide reacts almost instantly on undergoing the above treatment; thoroughly dried material, on the other hand, dissolves after a few minutes provided it is finely ground. If, however, the dry peroxide is in the form of coarse grains, it may be several hours before the reaction is finished, or the decomposition may be incomplete. Furthermore, too much potassium iodide should not be used, as otherwise lead iodide will separate out. In that case from 3 to 5 gms. more of sodium acetate are added and a few cubic centi- meters of water. The mixture is shaken until the lead iodide has dissolved completely and not till then diluted to a volume of 25 c.c. The solution must remain perfectly clear and there should not be a trace of lead iodide precipitate. This excellent method may also be used by the analysis of minium (red lead). 18. Determination of Ozone in Ozonized Oxygen. 1000 c.c. ^ Na 2 S 2 3 = ^ = ^ = 2.4 gms. O 3 . (a) Schonbeiris Method. The most accurate method for estimating ozone consists in allowing th.e ozonized oxygen to act upon potassium iodide solution whereby free iodine is formed : 2KI + O 3 + H 2 O - 2KOH + 1 2 + O 2 , and the iodine may be titrated, after acidifying the solution with dilute sulphuric acid, by means of N/10 sodium thiosulphate. It is not, however, immaterial whether the ozone reacts with a neutral or with an acid solution. In the latter case far too much iodine is liberated, although in the former case exactly the right amount is set free. Sir B. C. Brodie * called attention io this t Phil. Trans., 162, 435-484 (1872). DETERMINATION OF OZONE IN OZONIZED OXYGEN. 677 fact in his classic researches on ozone. Brodie confirmed the results obtained of his titrations by weighing the amount of ozone used in the experiments. This work of Brodie's appears to have been forgotten,* for many other chemists have since that time attempted to work out an iodimetric method for estimating ozone, some using acid solutions of potassium and iodide and some neutral solutions to absorb the gas, although for a long time it occurred to no one else that the results could be checked by weighing out a definite amount of ozone for test experiments. In 1901, however, this was done in a very simple way by R. Ladenburg and R. Quasig, f who were without knowledge of Brodie's work. Their method consisted in weighing a glass bulb of known capacity which was provided with glass stop-cocks, filling it with oxygen and then weighing. The oxygen was then replaced by ozone, so that the gain in weight multiplied by three represented the amount of ozone present. In order, now, to titrate the ozone, Ladenburg and Quasig expelled the gas from the bulb by distilled water, and conducted it slowly through a neutral solution of potassium iodide which was subsequently treated with an equivalent amount of sul- phuric acid and the liberated iodine titrated with N sodium thio- sulphate. The results of Ladenburg and Quasig have been carefully tested in the author's laboratory { and the method improved somewhat by absorbing the ozonized oxygen by potassium iodide solution in the glass bulb itself rather than expelling the gas from the bulb and passing it into the iodide solution. The estimation of ozone by weighing is a much too round- about process to permit a practical application, particularly on account of the fact that the measurement and weighing of the gas must take place in a room at constant temperature, a condition which cannot in many cases be readily fulfilled. Consequently the volumetric titration of the gas is far more practical. Procedure. A glass bulk of about 300 to 400 c.c. capacity, of the form shown in Fig. 95, is procured and its volume accurately * Luther and Inglis, Z. phys. Chem., 48, 208 (1903). fBer. 34, 1184(1901). J Treadwell and Anneler, Z. anorg. Chem., 48, 86 (1905). VOLUMETRIC ANALYSIS. determined by weighing it empty and then filled with water, applying the correction for temperature as described on p. 517 et seq. The bulb is then connected with a gas delivery tube, making use of Babo flanged joints (Fig. 95, c and d) which are pressed together by means of a steel clamp, lined with cork. The delivery tube is connected with the supply of ozone and oxygen, with which the water in the bulb in replaced. During the filling of the bulb, but little of the ozone is absorbedby the water. When the \ FIG. 95. FIG. 96. FIG. 97. tube is filled, the lower stop-cock is closed first and the upper one a few seconds later. The bulb is then disconnected with the gas delivery tube, inverted, the upper-stop cock opened quickly for an instant in order to establish atmospheric pressure in the bulb, i-nd then connected by means of rubber tubing with the gas reservoir N which is filled with double-normal potassium iodide solution (Fig. 97). The air imprisoned in the rubber tubing is allowed to escape through the three-way stop-cock b and after properly setting the cock, about 20 to 30 c.c. of the iodide solution DETERMINATION OF OZONb IN OZONIZED OXYGEN. 679 are introduced into the bulb. Finally the stop-cock 6 is closed and the rubber tubing disconnected^ The contents of the bulb are vigorously shaken and allowed to stand for half an hour; at the end of this time the absorption of the ozone will be complete. An Erlenmeyer flask is then placed under the stop-cock &'; this is opened and immediately afterwards the upper stop-cock also. The bulb is washed out first by introducing some potassium iodide solution through a and finally with pure water. The contents of the flask are then acidified with dilute sulphuric acid and the liberated iodine titrated with tenth-normal sodium thiosulphate. The computation takes place as follows: Contents of the bulb = V c.c. Ozone found by titration = p gms. Temperature = t, barometer reading = B, aqueous tension w. The volume of the bulb at and 760 mm. pressure is = , v ^-i0)273 ~ 760(273 +"0 ' When filled with oxygen this would weigh: 32- 7 22,391 gm Therefore the weight of oxygen and ozone in the bulb is 32- V p 22,391 3' and the per cent, of ozone in the mixture is 100- p 6,717,300- p 32- V p 96 7 + 22,391 22,391 + 3 = per cent, ozone. 68o VOLUMETRIC ANALYSIS. (b) Method of Soret-Thenard* Ozone is absorbed quantitatively by means of sodium arsenite solution in accordance with the following equation: Na 3 AsO 3 + O 3 = Na 3 AsO 4 + O 2 , although A. Ladenburg f finds that the absorption takes place much more slowly than by means of potassium iodide. When, therefore, the ozone is passed through the arsenite solution, there is danger of getting too low results. If the absorption takes place in a glass bulb, however, the results are good. Ozone is also absorbed by alkali bisulphite | solutions and may be estimated in this way, by titrating the excess of bisulphite with iodine. Ladenburg, however, has shown that the method is not as accurate as the potassium iodide one, so that it will not be considered further here. 19. Determination of Hydrogen Peroxide. Kingzstt's Method.!) 1000 c.c. ^ Na 2 S 2 3 solution = ^^ = ^51? =1.7008 gms. H 2 O 2 . The hydrogen peroxide solution is diluted until its H 2 O 2 content corresponds to about 0.6 per cent, by weight and of this solution 10 c.c. are used in the analysis. Procedure. About 2 gms. of potassium iodide are placed in an Erlenmeyer flask and dissolved in 200 c.c. of water, 30 c.c. of sulphuric acid (1:2) are added, and then, with constant stirring, 10 c.c. of the hydrogen peroxide solution are added from a pipette. After standing five minutes, the iodine liberated in accordance with the equation H 2 O 2 + 2KI + H 2 SO 4 = K 2 SO 4 + 2H 2 O + 1 2 is titrated by means of tenth-normal thiosulphate solution. * Compt. rend., 38, 445 (1854), 75, 174 (1872). tBer., 36, 115 (1903). t Neutral alkali sulphite is not suitable here, because it is not oxidized quickly by pure oxygen alone. Loc. cit. || J. Chem. Soc., 1880,792. DETERMINATION OF IRON. 681 Remark. This method is rather better than that described on p. 627 because the titration can take place in the presence of glycerol, salicylic acid, etc., which are sometimes used as pre- servatives in commercial hydrogen peroxide preparations. These substances will render the results obtained by the permanganate titration less accurate. 20. Determination of Iron. This method was first proposed by Carl Mohr * and is based upon the following reaction: FeCl 3 +HI < HCl + FeCl 2 +I. As the reaction is reversible, it is necessary to have an excess of hydriodic acid present in order that it may take place quanti- tatively in the direction from left to right. Procedure. The hydrochloric acid solution containing a weighed amount of the ferric salt is placed in a 300-c.c. glass- stoppered bottle, the greater part of the acid is neutralized by means of sodium hydroxide, and the air removed by means of a current of carbon dioxide. After this about 5 gms. of potas- sium iodide are added, the bottle closed, shaken, and allowed to stand in the cold for twenty minutes. The liberated iodine N is then titrated with sodium thiosulphate solution. As soon as the blue color has disappeared f more carbon dioxide- is con- ducted through the solution, the bottle is stoppered and allowed to stand for a few minutes to see whether the blue color will re- appear. Should this be the case, more thiosulphate is added, the flask again stoppered and allowed to stand. If a blue color again appears, the solution contains too little potassium iodide, so that it is necessary to repeat the entire analysis, using 1-2 gms. more of it. With sufficient potassium iodide and only little free hydrochloric acid, the reaction is always com- plete at the end of twenty minutes. The results obtained are satisfactory. * Ann. d. Chem. u. Pharm., 105, p. 53. t Starch is added in all these titrations. 682 VOLUMETRIC ANALYSIS. 21. Determination of Copper. Method of Haen*-Low.f 1000 c.c. Na 2 S 2 3 solution =^ = 6.357 gms. Cu. Principle. If an acid solution of a cupric salt is treated with an excess of potassium iodide, all the copper is precipitated as cuprous iodide, 2CuSO 4 + 4KI -. Cu 2 I 2 + 2K 2 SO 4 + 1 2 , and there is liberated one atom of iodine for each atom of copper present. The iodine is titrated with sodium thiosulphate solution. This method has found extensive application, especially at the copper mines in the United States and, if the proper conditions are fulfilled, the results are just as accurate as those obtained by electrolysis. The method has been studied by Gooch and Heath, if who find that the above reaction takes place preferably in a solution containing not over 3 c.c. concentrated sulphuric acid, hydrochloric acid, nitric acid free from lower oxides, or 25 c.c. of 50 per cent, acetic acid. Obviously the solution must not contain ferric iron or any other oxidizing agents which will react with potassium iodide under the same conditions. Standardization of the Thiosulphate Solution. In technical work it is customary to standardize the thio- sulphate solution against pure copper. About 0.2 gm. of pure copper is weighed into a 200-c.c. Erlenmeyer flask and dissolved in 5 c.c. of a mixture of equal parts nitric acid (sp. gr. 1.42) and water. The solution is diluted with 25 c.c. water and boiled a few minutes to get rid of the greater part of the reduced oxides of nitrogen. To remove the last of the nitrous oxides, 5 c.c. of bromine water are added and the solution boiled until the excess bromine is expelled. The flask is then removed from the flame and strong ammonia added to its contents until a slight excess is present. After boiling off the excess of ammonia, * Ann. Chem. Pharm., 91, 237 (1854). f Technical Methods of Ore Analysis. J Z. Anorg. Chem., 55, 129 (1907). DETERMINATION OF COPPER IN ORES. 683 7 c.c. of strong acetic acid are added, which dissolves any copper oxide that has deposited. After cooling to room temperature, 3 gms. of potassium iodide are added and the brown solution titrated with sodium thiosulphate until nearly colorless. Then starch solution is added and the titration finished. In making the titration for the first time, one is bothered somewhat by the fact that 'the cuprous iodide is of a light-brown color but after a little practice there is no difficulty in getting the correct end- point. If t c.c. of the solution were used in titrating a gms. of copper then 1 c.c. of thiosulphate = gm. Cu. t Determination of Copper in Ores. Low's Method. Principle. The ore is dissolved in acid, the copper separated from iron, etc., by precipitating it upon metallic aluminium, the deposit dissolved in nitric acid and treated as in the standardiza- tion. Procedure. To 0.25-0.50 gms. of finely ground ore weighed into a 250-c.c. Erlenmeyer flask, add 6 c.c. of nitric acid (sp. gr. 1.42), and boil gently until nearly to dryness. Add 5 c.c. of strong hydrochloric acid and heat again. As soon as the in- crusted matter has dissolved add 7 c.c. of concentrated sulphuric acid and heat until the sulphuric acid fumes freely. Cool and add 25 c.c. of water. Then heat until any anhydrous ferric sulphate is dissolved, and filter to remove insoluble sulphates and silica. Wash the flask and filter-paper until the volume of the filtrate amounts to about 75 c.c., receiving it in a No. 2 beaker. Place on its edge in the beaker a piece of aluminium- foil bent into triangular shape. Cover the beaker and boil gently for seven to ten minutes, which will be sufficient to pre- cipitate all the copper, provided the solution does not much exceed 75 c.c. Avoid boiling to very small bulk. The alu- minium should now appear clean, the copper being detached or loosely adhering. Remove from the heat and wash down the cover and sides of the beaker with cold water. There is danger of finely-divided copper being oxidized and dissolved. To prevent this, and at the same time to remove any traces of copper remaining in solution, add 15 c.c. of strong hydrogen- 684 VOLUMETRIC ANALYSIS. sulphide water. The next step depends upon the amount of copper in the ore. (a) When there is apparently less than 20 per cent. Cu present, decant the liquid through a filter and then without delay transfer by means of a jet of hydrogen-sulphide water from a wash-bottle, the copper to the filter leaving the foil as clean as possible in the beaker. Wash the copper and filter thoroughly with this hydrogen-sulphide wat5r, being careful not to allow the filter to stand empty until the washing is finished. (The filtrate should not show a brown tinge of copper sulphide.) Now place the original flask under the funnel. Pour over the alu- minium in the beaker 5 c.c. of a mixture of equal parts concen- trated nitric acid and water. Heat just to boiling and pour the hot acid very slowly upon the filter, lifting the fold if neces- sary. Now, before washing, pour 5 c.c. of bromine water into the filter and wash the beaker and filter with hot water. Finally remove the filter and wash any residue upon it into the flask. If the bromine was not sufficient to give a slight tinge to the filtrate more of it must be added. Boil the filtrate, which does not exceed 75 c.c., to expel the excess of bromine, but do not concentrate to small volume. Remove from the heat and add a slight excess of strong ammonia (usually 7 c.c.). Boil off the excess of ammonia and add 3 or 4 c.c. of strong acetic acid. Cool to room temperature, add 3 gms. of potassium iodide, and titrate with thiosulphate, adding starch toward the end of the reaction. (b) With high percentages of copper it is better to wash the copper by decantation instead of on the filter. Transfer the liquid .and copper to the original flask, and set the beaker and aluminium aside temporarily. Allow the liquid in the flask to settle, decant through a filter and wash three or four times by decantation with hydrogen-sulphide water, using about 20 c.c. each time. Now place the flask containing the copper under the funnel. Heat the 5 c.c. of nitric acid (1 vol. cone. HN0 3 to 1 vol. water) in the beaker with the aluminium-foil and pour it through the filter. Remove the flask, place the beaker in its place under the funnel, and heat the acid until all the copper has dissolved and the red fumes are mostly expelled. Now return the flask under the funnel, add the bromine as in the above method of analysis (a), and continue as described above. DETERMINATION OF ANTIMONY TRIOXIDE COMPOUNDS. 685 22. Analysis of Arsenious Acid. The titration is effected in the same way as in the standard- N ization of the iodine solution, described on p. 650. 23. Determination of Antimony Trioxide Compounds. 1000 c.c. ^iodine solution = - 3 == 7.21 g. Sb 2 O 3 = 6.01g. Sb. The titration is carried out exactly as in the case of arsenious acid (cf. p. 650) except that tartaric acid, or Rochelle salt, must be added to the solution in order to prevent the precipitation of antimonous acid, or antimony oxychloride, as a result of hydrolysis. Examples: (a) Determination of Antimony in Tartar Emetic. If an aqueous solution of tartar emetic be treated with iodine in the presence of starch, the first few drops of reagent will impart a permanent blue color to the solution. If, however, a little sodium bicarbonate is added to the solution, the trivalent antimony is oxidized quantitatively to the pentavalent condition. K(SbO) C 4 H 4 O 6 + 6NaHCO 3 + 1 2 = = Na 3 SbO 4 + 2NaI + KNaC 4 H 4 O 6 + 3H 2 O + 6C0 2 . innrk N . .. K(SbO)C 4 H 4 6 + ^H 1000 c.c. -T iodine solution = 9Q ^332.34 20 16.617 gms. 8.309 gms. of tartar emetic are dissolved in water, the solution diluted to exactly 500 c.c. and well mixed. Of this solution, 20 c.c. are removed by a pipette, diluted to 100 c.c., treated with 20 c.c, of 2 per cent, sodium bicarbonate solution, and titrated with 686 VOLUMETRIC ANALYSIS. tenth-normal isodine solution, using starch as an indicator. If t c.c. are used for the titration, the salt contains: 1.6617X25XJ ~~ = 5- = per cent, tartar emetic, or = 1.809' = per cent, antimony. (6) Determination of Antimony in Stibnite. Not over 0.5 gm. stibnite is dissolved in a small covered beaker by means of 10 c.c. concentrated hydrochloric acid (sp. gr. 1.2). The acid is allowed to act in the cold for about ten minutes, after which the contents of the covered beaker are heated gently on the water bath for ten or fifteen minutes. Three gms. of powdered tartaric acid are then added and the heating is continued for ten minutes longer, but care is taken not to allow the liquid to evapor- ate sufficiently to expose any part of the bottom of the beaker. When this precaution is taken, there is no volatilization of the antimony, and all of the hydrogen sulphide is expelled. Sb 2 S 3 4- 6HC1 = 2SbCl 3 + 3H 2 S. The solution is now removed from the water bath, allowed to cool to the room temperature, and very cautiously diluted with water, which is added at first drop by drop, until a volume of about 100 c.c. is obtained. If, in the meantime, a red coloration due to antimony sulphide appears during the dilution, the solution should be at once heated until it disappears, and the diluting then continued. The diluted solution is nearly neutralized with ammonia, but is left slightly acid. The cold, slightly acid solution is poured into a 700 c.c. beaker containing 3 gms. of sodium bicarbonate dissolved in 200 c.c. of water, starch paste is added, and the solution titrated with iodine to the appearance of a permanent blue. Remarks. Antimony chloride is volatile v/ith steam from its concentrated solutions, so that the solution should not be boiled until it has been diluted. The heating on the water-bath can be DETERMINATION OF ANTIMONY PENT OXIDE COMPOUNDS. 687 carried out, however, without fear of losing antimony provided the acid is not allowed to evaporate to any extent. This heating serves to remove all the hydrogen sulphide which would otherwise precipitate the antimony as trisulphide upon diluting the solution. If insufficient tartaric acid is present, antimony oxychloride, SbOCl, precipitates and if the solution is titrated in this condition it is impossible to obtain a permanent end-point. Such a pre- cipitate may be filtered off, dissolved in concentrated hydrochloric acid and the solution treated by itself as above described. The value of the iodine solution in terms of Sb is given in the previous process (a). 24. Determination of Antimony Pentoxide Compounds (A. Weller).* By heating a pentavalent antimony compound with concen- trated hydrochloric acid and potassium iodide in the Bunsen apparatus (Fig. 94, p. 662), the antimonic acid is reduced to antimonous acid with separation of iodine: Sb 2 O 5 4- 4HI =Sb 2 O 3 + 2H 2 O + 2I 2 . The iodine is distilled over into potassium iodide solution and N titrated with Na 2 S 2 Oa solution. The results are a little low. 25. Determination of Hydrogen Sulphide. 1000 c.c. Na 2 S 2 5 solution = = =1.704 gms. H 2 S. If a solution of hydrogen sulphide is treated with iodine, it is oxidized with separation of sulphur: For the determination of the amount of the gas present in hydrogen sulphide water, a measured amount is transferred by * Annal., 213, 364. 688 VOLUMETRIC ANALYSIS. X means of a pipette to a known amount of ^ iodine solution and the excess of the latter is titrated with thiosulphate solution.* If the amount of hydrogen sulphide present is not very large, correct results are obtained without difficulty. With consider- able hydrogen sulphide, on the other hand, the deposited sulphur is likely to enclose some of the iodine solution, as shown by its brown color; this iodine escapes the titration with thiosulphate. In such a case, the film of sulphur floating on the surface of the liquid is removed with a glass rod after the completion of the thiosulphate titration, transferred to a glass-stoppered cylinder, and shaken with 1-2 c.c. of carbon bisulphide. The latter dis- solves the iodine with a violet color and the color is discharged by the addition of sodium thiosulphate solution. t In this way the total amount of the iodine that remains can be titrated. Remark. This method can be used to advantage for deter- mining the sulphur present in soluble sulphides. The sulphides are decomposed as described on p. 350 by means of acid, and the hydrogen sulphide evolved is conducted into a definite amount N of iodine solution. The excess of the latter is titrated as above with sodium thiosulphate solution. Determination of Hydrogen Sulphide in Mineral Waters. N A measured amount of iodine solution and 2 gms. of potas- -I \J\J sium iodide are placed in a tall liter cylinder, 1000 c.c. of the water to be analyzed are added, and after thoroughly shaking, the excess N of the iodine is titrated with r- thiosulphate. The standardiza- * Correct results cannot be obtained by titrating directly with iodine, cf . O. Brunck, Z. Anal. Chem., 45, 541 (1906). fThe separation of the sulphur into a coherent film can be prevented by sufficiently diluting the solution with boiled water. 0. Brunck (Z. anal. Chem., 45, 541) therefore, recommends using hundredth-normal iodine instead of tenth-normal solution, and this is certainly advisable in the case of small quantities of hydrogen sulphide as, for example, in a mineral water. On the other hand, when a relatively large volume of hydrogen sulphide is liberated from a sulphide by means of acid (see page 350) it is advisable to use tenth-normal iodine, as otherwise the volume of solution will be too large unless a very small weight of substance is used in the analysis. ANALYSES OF ALKALI SULPHIDES. 689 tion of the iodine solution used is accomplished by measuring off 10 c.c. of the solution, adding 2 gms. of potassium iodide, N diluting to 1 liter with boiled water, and titrating with thio- 1UJ sulphate solution. 26. Analysis of Alkali Sulphides. 'M' T> Q Cl QO f)7 1000 c.c. JQ iodine solution^- or =^- = 1.604 gms. S. A measured volume of the alkali sulphide solution is allowed to run slowly, with constant stirring, into a very dilute solution of iodine which is acid with hydrochloric acid. The excess of iodine is titrated with sodium thiosulphate solution. 27. Analysis of a Mixture of Alkali Sulphide and Alkali Sulphydrate. If a solution of alkali sulphide and alkali sulphydrate is treated with an acid solution of iodine, the following reactions take place: (a) Na 2 S + 2HCl = 2NaCl+H 2 S >> (6) (c) It is evident from these equations that in the case of the sulphide, the quantity of hydriodic acid formed by the oxida- tion of the hydrogen sulphide is equivalent to the quantity of hydrochloric acid required to decompose the sulphide; in this case, therefore, the acidity of the solution remains unchanged. In the case of the sulphydrate, however, which is the acid salt of hydrosulphuric acid, the quantity of hydriodic acid formed is equivalent to twice the quantity of hydrochloric acid required to decompose the sulphydrate. Thus the acidity of the solution is a measure of the quantity of sulphydrate present. Moreover, if t c.c. of tenth-normal alkali is required to titrate this acid then 690 VOLUMETRIC ANALYSIS. 2t c.c. of tenth-normal iodine must have been required to oxidize the hydrogen sulphide from the sulphydrate. Procedure. A known volume of tenth-normal iodine together with a known volume of tenth-normal hydrochloric acid * is diluted in a beaker to a volume of about 400 c.c. and the solution containing the sulphide and sulphydrate is added slowly from a burette, with constant stirring, until the solution becomes pale yellow. Starch indicator is added and the excess of iodine titrated with tenth-normal thiosulphate solution. Finally, the acid in the solution is titrated with tenth-normal sodium hydroxide solution. Computation. In the analysis, the volumes of standard solutions used were: T c.c. tenth-normal iodine; ^ c.c. tenth- normal thiosulphate;, t 2 c.c. tenth-normal hydrochloric acid; t z c.c. of tenth-normal sodium hydroxide: and V c.c. of the sulphide mixture. Then (TtJ c.c. = total volume of iodine required, and (t 3 t 2 ) c.c. = the volume of sodium hydroxide required to neu- tralize the acid formed by the reaction with the sulphydrate. Then a n NaSH_ V^o v<) I -t f^ ^ ,<-. ^ V^o *">/ /N ^ v/Vy tJ\J\JQ ylill. XiCtOJ-J-, 2 10,000 and are present in V c.c. of solution. * The quantity of hydrochloric acid present must be sufficient to decompose all the sulphide and sulphydrate; an excess does no harm. DETERMINATION OF SULPHUROUS ACID. 691 28. Analysis of a Mixture of Hydrogen Sulphide and Alkali Sulphydrate. The analysis is carried out exactly as in the case just described and the computation is similar. N N N" Let r = c.c. iodine, ^-=c.c. thiosulphate, 2 = c.c. ^ acid, N and 3 = c.c. alkali. Then -y-( ? -y ]X0.00560S gm. NaSH and 2y]X0.001701 gm. H 2 S. Remark. The last two methods of analysis are applicable only to solutions containing no other compounds decomposable by hydrochloric acid than sulphide and sulphydrate, and no other substance that will react with iodine. The analysis may be carried out without the addition of any hydrochloric acid. In this case the solution of sulphides is diluted to about 400 c.c., starch added, and the titration with iodine carried out directly. The hydriodic acid formed is titrated with caustic soda, using lacmoid * as indicator. The reactions are 29. Determination of Thiosulphate in r the Presence of Sulphide and Sulphydrate. A measured volume of the solution is treated in a 200-c.c. graduated flask with an excess of freshly-precipitated cadmium carbonate. After shaking thoroughly, the liquid is diluted to the mark, filtered through a dry filter and 100 c.c. of the filtrate * Methyl orange can be used, but it is not so easy to distinguish the end- point. Phenolphthalein works well but no better than the lacmoid. Care should be taken to titrate the acid' against the alkali during the standardiza- tion at the same dilution as to be used in the analysis and an appreciable amount of carbonate should not be present. 692 YOLU METRIC ANALYSIS. titrated with iodine solution. By shaking with cadmium car- bonate, the sulphide and sulphydrate are removed and the thiosulphate remains in solution. 30. Determination of Sulphurous Acid. 1000 c.c. ^ iodine solution =^- 2 = ^?= 3. 203 gms. SO 2 . The determination is based upon the following reaction: SO 3 + H 2 O + 21 = 2HI + S0 3 , the sulphurous acid being oxidized to sulphuric acid. If starch is added to a solution of sulphurous acid, and a titrated iodine solution is run into it from a burette, the blue color will not be obtained until all of the sulphurous acid has been acted upon. Bunsen, however, in 1854 showed that this sensitive reaction, which was first used by Dupasquier, will only take place quanti- tatively according to the above equation when the solution does not contain more than 0.04 per cent, by weight of SO 2 . With greater concentrations uniform results are not obtained. This irregularity was ascribed to the reversibility of the reaction, so that it was suggested that the titration be performed in alkaline solution,* thus removing the hydriodic acid as fast as it is formed. But the results then obtained are still inaccurate. f Finkener,f o/i the other hand, states that correct values will be obtained if the sulphurous acid is allowed to run into the iodine solution. J. Volhard has confirmed the results of Finkener and shown that the anomalous results obtained on titrating sulphurous acid with iodine are not due to the reversibility of the reaction, for the direct addition of 20 per cent, sulphuric acid is without * Addition of MgCO 3 or NaHCO 3 (Fordos and Gelis). t E. Rupp, Ber., 35, 3694 (1902), states that it is possible to obtain good results by the method of Fordos and Gelis if the sulphurous acid is allowed to act for at least half an hour upon an excess of iodine in the presence of sodium bicarbonate. The solution is then titrated with sodium thiosulphate. According to E. Miiller and O. Diefenthaler, however, this is theoretically in- correct, for the iodine tends to form a little hypoiodite: I 2 + H 2 O^HI + HIO, and the latter reacts with sodium thiosulphate: Na 2 S 2 O 3 +4HIO-|-H.,O = N a2 S0 4 +H 2 S0 4 +4HI. % Finkener- Rose, Quantitative Analyse (1871), p. 937. Ann. d. Chem. u. Pharm., 242, 94. DETERMINATION OF SULPHUROUS ACID. 693 influence. The incomplete oxidation of the sulphurous acid is caused by the fact that the hydriodic acid reduces a part of the sulphurous acid to free sulphur: * (1) SO a +4HI = 4I + 2H a O+ S. If sulphurous acid, whether dilute or concentrated, is allowed to run into a solution of iodine with constant stirring, there is complete oxidation of the SO 2 : (2) S0 2 + 2I + 2H 2 0=2HI+H 2 SO 4 . If, on the contrary, iodine solution is run into the solution of sulphurous acid, both reactions will take place: (3) 3SO 2 + 4HI + 2H 2 O = 2H 2 SO 4 + 4HI f + S. According to Raschrg 7 { however, Volhard's explanation is a\so incorrect, for he finds that no separation of free sulphur takes place if the iodine is allowed to act upon sulphur dioxide in a dilute solution. Raschig believes that the error that results when iodine is added to the sulphurous acid solution is due to a loss of SO 2 by evaporation. Correct results are always obtained if the sulphurous acid is added slowly, with constant stirring, to the iodine solution until the latter is decolorized. In the analysis of sulphites, the sulphite solution is added from a burette to the solution of iodine and hydrochloric acid. * If iodine solution is added slowly to a not too-dilute sulphurous acid solution, a distinct separation of sulphur is soon apparent, t The HI acts as a catalyser according to Volhard. J Z. Angew. Chem., 1904, 580. 694 VOLUMETRIC A "NA LYSIS. 31. Determination of Formaldehyde (Formalin). Method of G. Romijn.* 1000 c.c. N. iodine solution = - =~ =15.01 gms. formaldehyde. Principle. Formaldehyde is quantitatively oxidized to formic acid by remaining in contact with iodine for a short time in alkaline solution: HCHO + H 2 Q + 1 2 = 2HI + HCOOH. Procedure. The aqueous solution of formaldehyde, known com- mercially as " formaline," contains about 40 per cent, of for- maldehyde. For analysis, 10 c.c. of the formaldehyde solution are diluted to 400 c.c., and of this 1 per cent, solution, 5 c.c. ( = 0.125 c.c. N of the original solution) are taken for analysis. 40 c.c. of iodine solution are added, and immediately afterwards strong sodium hydroxide solution, drop by drop, until the color of the solution is a light yellow; it is then placed one side for ten minutes. The solution is then acidified with hydro- N chloric acid, and the unused iodine is titrated back with sodium thiosulphate solution. N 1 c.c. TT iodine solution= 0.001501 gm. formaldehyde. 32. Determination of Hydroferricyanic Acid.f 1000 c.c. iodine solution=^^= ? =32.92 gms. K 3 Fe(CN) 8 . Principle. If a neutral solution of potassium ferricyanide is treated with an excess of potassium iodide, the ferricyanide ion is reduced to ferrocyanide ion with separation of free iodine: * Zeitschr. f. anal. Chem., 36 (1897), p. 19. f Lenssen, Ann. Chem., 91, 240. Mohr., Ibid., 105, 60. DETERMINATION OF PHENOL. 695 Lenssen titrates the liberated iodine with sodium thiosulphate, but the results are not concordant, because the reaction is a reversible one. The reaction is quantitative, however, as Mohr first showed, if the ferrocyanide is removed from the solution as fast as it is formed. This is accomplished, according to Mohr, by adding an excess of zinc sulphate, free from iron, to the solution. According to the experiments of E. Miiller and O. Diefenthaler, * the titration should take place in a solution which is as nearly neutral as possible, but not in one made alkaline by the addition of sodium bicarbonate (see p. 692). Miiller and Diefenthdler' s Procedure. About 0.7 gm. of the ferrocyanide is weighed into a glass-stoppered flask, dissolved in about 50 c.c. water, and treated with 3 gms. potassium iodide and 1.5 gms. of zinc sulphate free from iron. If an acid solution of ferricyanide is to be analyzed, it is carefully neutralized with caustic soda until barely alkaline and then just acidified with a drop of sulphuric acid. Alkaline solutions must always be neutralized with acid. 33. Determination of Phenol. Method of W. Koppeschaar.f 1000 c.c. Na 2 S 2 3 = ^S = =L567 g^. C 6 H 5 OH. Principle. If an aqueous solution of phenol is treated with an excess of bromine, the phenol is converted quantitatively into tribromophenol : C 6 H 5 OH+3Br 2 =3HBr + C 9 H 2 Br 3 (OH). The tribromophenol is a pale yellow, crystalline substance which is quite insoluble in water (43,700 parts of water dissolve 1 part of tribromophenol). If, after the reaction has taken place, potassium iodide is added to the solution, iodine is liberated corresponding to the excess of bromine, and by titrating this * Z. anorg. Chem., 1910, 418. t Z. anal. Chem., 15, 233 (1876). 696 VOLUMETRIC ANALYSIS. iodine with sodium thiosulphate solution, it is easy to find how much bromine reacted with the phenol. Requirements. A tenth-normal solution of bromine and a tenth-normal solution of sodium thiosulphate. On account of the volatility of free bromine, Koppescharr uses a solution of potassium bromate and bromide which, upon being acidified, gives a known amount of bromine in accordance with the equation: KBrO 3 + 5KBr + 6HC1 = 6KC1 + 3H 2 O + 3Br 2 . Thus, to obtain a tenth-normal solution of bromine which will keep indefinitely, exactly 2.784 gms. of pure potassium bromate (dried at 100) and about 10 gms. of potassium bromide are dissolved in water and the solution diluted to one liter. An excess of bromide does no harm : KBr0 3 _ 167.02 60 60 5KBr 5X119.02_ ~60~ : ~60~~ Procedure. About 0.5 gm. of phenol is weighed out in a weighing beaker, dissolved in a little water, the solution rinsed into a liter flask and well shaken. Of this solution, 100 c.c. are withdrawn in a pipette, transferred to a second liter flask, diluted with water up to the mark, mixed and 170 c.c. this solution transferred to a stoppered bottle of about 250-c.c. capacity, treated with 50 c.c. of the bromate solution, shaken, acidified with 5 c.c. concentrated hydrochloric acid, shaken again, and allowed to stand fifteen minutes. At the end of this time, 2 gms. of potassium iodide are added and the liberated iodine, corre- sponding to the excess of bromine, is titrated with tenth-normal thiosulphate solution, using starch as indicator. Then if t c.c. of the last solution are used and the weight of phenol was a gms. : (60-0X0.1567 phenol REDUCTION METHODS. 697 Remark. Before making an analysis, a blank experiment should always be made with the bromate solution to make sure that its strength corresponds to the theoretical value. This method is suitable for the analysis of pure preparations of phenol (carbolic acid) but not for crude phenol, creosote oil, etc.* B. REDUCTION METHODS. i. Determination of Ferric Iron (Fresenius).f In the case of all methods previously discussed, it was necessary to reduce the iron to the ferrous condition before it could be determined volumetrically. In the following method, first suggested by Penny and Wallace,! but improved by Fresenius, the iron in the ferric condition may be determined with accuracy and rapidity. The hydrochloric acid solution containing ferric chloride is titrated hot with stannous chloride solution until the former becomes colorless. By this means the ferric salt will be reduced to ferrous salt: 2FeCl 3 + SnCl 2 - SnCl 4 + 2FeCl 2 . Inasmuch as it is not very easy to determine the end-point with accuracy, because the last part of the iron is reduced very slowly, it is customary to run over the end-point and to titrate the excess of the stannous chloride, with iodine solution. Solutions Required. 1. A Ferric Chloride Solution Containing a Known Amount of Iron. It is prepared by dissolving exactly 10.03 gms. of bright iron wire in hydrochloric acid within a long- necked flask held in an inclined position; the iron is oxidized with potassium chlorate and the excess of chlorine is completely * J. Toth, Z. anal. Chem., 25, 160 (1886). t Z. anal. Chem., 1, p. 26. j Dingl. polyt. J., 149, 440. 698 YOLU METRIC ANALYSIS. expelled by boiling. The solution of ferric chloride is washed into a liter flask and diluted up to the mark with water; 50 c.c. of this solution contain 0.5 gm. of pure iron.* 2. A Stannous Chloride Solution. 25 gms. of tin-foil are heated for two hours on the water-bath with 50 c.c. of hydrochloric acid of specific gravity 1.134 and a few drops of chloroplatinic acid in a porcelain dish which is covered with a watch-glass. After this, 150 c.c. of hydrochloric acid and aft equal volume of water are added, the solution filtered and diluted up to 1 liter. As stan- nous chloride is oxidized by contact with the air, it is placed in a flask which on one side is connected with the burette as shown in Fig. 87, p. 556 and on the other side with a Kipp carbon di^ oxide generator. 3. An Iodine Solution Approximately Tenth-normal. Procedure. (a) Standardization of the Solutions. First of all, the stannous chloride and iodine solutions are titrated against one another. About 2 c.c. of the former are measured from the burette, diluted to about 60 c.c., a little starch solution added, and the mixture titrated with iodine until a blue color is obtained. Next, 50 c.c. of the acid ferric chloride solution containing a known amount of iron are titrated against the stannous chloride solution. (6) Determination of Iron in Hematite. 5 gms. of the finely-divided ore are ignited in order to destroy any organic matter which may be present, then placed in a long-necked flask and boiled with concentrated hydrochloric acid and a little potas- sium chlorate until the iron oxide is all dissolved, leaving behind nothing but a white sandy residue. After this 20 c.c. more of hydrochloric acid are added and the boiling is continued while a current of air is passed through the solution, until all the excess of chlorine is completely removed and the escaping vapors will no longer set free iodine when passed into a potassium iodide solution. The solution thus obtained is diluted to exactly 500 c.c. and 50 c.c. of it are taken for the analysis. * The assumption being made that the iron wire contained 99.7 per cent pure iron. DETERMINATION OF HYPOCHLOROUS ACID. 699 Example. 1. Standardization of the reagents: 2 c.c. of stannous chloride solution require 7.2 c.c. of iodine solution. 1 c.c. iodine solution =0.278 c.c. SnCl 2 50 c.c. ferric chloride solution ( 0.5 gm. iron) require for decolorization 30.34 c.c. SnCl 2 and for the titration of the excess 0.51 c.c. of iodine solution = 0.51X0.28 0.14 c.c. SnCl 2 Consequently, 50 c.c. ferric chloride solution = 0.5 gm. iron =30.20 c.c. SnCl 2 and 1 c.c. SnCl 2 = ^|_ =0.01656 gm. Fe. 2. Titration of the solution to be analyzed: 50 c.c. ( = 0.5 gm. of iron ore) require 18.96 c.c. SnCl 2 and for the titration of the excess, 0.64 c.c. of iodine = 0.64X0.28 = 0.18 c.c. SnCl 2 so that 0.5 gm. of ore corresponds to 18.78 c.c. SnCl 2 and contain, therefore, 18. 78 X 0.01656 = 0.31 10 gm. Fe, and in per cent. : 0.5:0.3110 = 100:* a: = 62.20 percent. Fe. 2. Determination of Ferric Iron by Means of Titanous Chloride (Knecht and Hibbert).* 1000 c.c. TiCl 3 solution ==0.8 g. oxygen = = 5.585 g. Fe. Principle. If an acid solution of a ferric salt is treated with titanous chloride, the iron is immediately reduced in the cold to the ferrous condition: FeCl 3 + TiCl 3 = TiCl 4 + FeCl 2 . Preparation of Titanous Chloride Solution. A concentrated solution of titanous chloride, prepared by the electrolysis of TiCl4, can now be obtained on the market. Such a solution is treated *Ber. 36, 1551 (1903). ;oo VOLUMETRIC ANALYSIS. with an equal volume of concentrated hydrochloric acid, boiled,* and then diluted with ten times as much boiled water. The solution is maintained in contact with an atmosphere of hydrogen, or carbon dioxide, and kept in a bottle such as is shown in Fig. 87, p. 556 which is connected with a burette, and in this case with a Kipp hydrogen, or carbon dioxide, generator instead of the soda lime tube. Standardization of the TitanouS Chloride Solution. A ferric chloride solution known of strength is prepared as described on p. 697, and of this solution 50 c.c. are measured out into a beaker, and the titanium trichloride is slowly added with constant stirring, while a current of carbon dioxide is constantly being passed into the beaker. After the solution is Dearly decolorized, a drop of potassium sulphocyanate solution i? introduced, and the addition of titanous chloride is continued to the disappearance of the red color. The analysis proper is carried out in exactly the same manner. 3. Determination of Ferrous and Ferric Iron by the Titanium Method. The ferrous iron is first titrated by means of permanganate in the presence of manganous sulphate (cf. p. 607) and then the total iron is determined as above with titanous chloride. The method can be carried out very rapidly and the results are accurate. 4. Determination of Hydrogen Peroxide.f If titanous chloride is run into an acid solution of hydrogen peroxide, the latter is colored first yellow, then a deep orange, and as soon as the maximum depth of color is produced, it begins to fade upon the further addition of titanous chloride until finally the solution becomes colorless, which is taken as the end-point. The reaction takes place in two stages: Ti 2 O 3 +3H 2 O 2 = 2TiO 3 + 3H 2 O 2TiO 3 + 2Ti 2 O 3 = 6TiO 2 or combining the two equations : 2TiCl 3 + H 2 2 + 2HC1 = 2TiCl 4 + 2H 2 O. * The boiling serves to expel any hydrogen sulphide that is present. t Knecht and Hibbert, Ber., 38, 3324 (1905). DETERMINATION OF FERRIC IRON BY TiCl,. 701 On account of the fact that the value of the titanous chloride solution is not very permanent, it is standardized against ferric chloride before each series of experiments. If t c.c. of titanous chloride solution of which 1 c.c. =a gms. Fe were required for the reduction of 1 c.c. of hydrogen peroxide, then the amount of the latter is 34.02 Xat ,. x= gms. H 2 O2 and in per cent. 30. 46a=per cent. H 2 0. If it is desired to express the per cent, in per cent, by volume of active oxygen (cf. p. 628) the following proportion holds: 10023 -at= per cent, oxygen by volume. According to Knecht and Hibbert,* persulphuric acid may likewise be estimated by titration with titanous chloride. The solution of the persulphate is treated with titanous chloride solution and the excess of the latter is titrated with ferric chloride in an atmosphere of carbon dioxide. 5. Determination of Hypochlorous Acid by Means of Arsenious Acid. 1000 c.c. ^ As 2 O 3 = 3.546 gms. chlorine. On adding arsenious acid to a solution of a hypochlorite, the former is oxidized to arsenic acid ; while the latter is reduced to chloride : 2NaOCl + As 2 O 3 = As 2 O 5 + 2NaCl. The end-point is reached when a drop of the solution added to a piece of iodo-starch paper will cause no blue coloration. Alkali hypochlorites and chloride of lime may be analyzed by this method and the results obtained are more reliable than in the case of those obtained by the iodimetric method described on p. 669, for the presence of chlorate has no effect in this case. * Knecht and Hibbert, Ber. 38, 3324 (1905). 702 VOLUMETRIC ANALYSIS. III. PRECIPITATION ANALYSES. I. Determination of Silver. Method of Gay-Lussac. This exceedingly accurate determination, which is extensively used for testing silver alloys, depends upon the precipitation of silver chloride from nitric acid solution. Common salt is used as the precipitant. Solutions Required. 1. Sodium Chloride Solution of Known Concentration. For convenience, it is customary to make the solution of such a strength that 1000 c.c. correspond to exactly 5 gms. of silver. It is more practical, however, to use a some- what weaker solution, consequently 2.700 gms. of chemically pure salt are dissolved in distilled water and diluted to 1 liter. 2. Decimal Solution of Sodium Chloride. 100 c.c. of the above solution are diluted with distilled water to 1 liter. In laboratories where silver determinations are frequently made, the above solutions are made up in much larger quanti- ties and kept in bottles similar to the one shown in Fig. 87, p. 556. The stronger solution is connected with a 100-c.c. pipette and the decimal solution with a burette. Standardization of the Sodium Chloride Solution. Exactly 0.5 gm. of chemically pure silver is weighed into a 200-c.c. flask provided with a well-ground glass stopper, and dissolved in 10 c.c. of nitric acid of specific gravity 1.2, free from chlorine. The solution is hastened by heating on a sand- bath. When the silver has dissolved, the solution is heated to boiling in order to expel the nitrous acid formed. The brown vapors collecting in the flask are removed by blowing in air. As soon as no more of these are formed, the flask is removed from the sand-bath, and allowed to cool. To the silver solution exactly 100 c.c. of the stronger salt solution are added, the flask stoppered, and vigorously shaken until the pre- cipitated silver chloride collects together, and the supernatant liquid appears clear. . -< compose the gas. In this case it is best to a ^ use the water-jacketed iron tube devised by St. Claire Deville and shown in Fig. 100. Cold water is run into the outer condenser at a and allowed to run out at 6, and the gas is col- lected as described above through the tube c. It is important that the water should run through the tube fast enough to keep the inner tube cold, otherwise the gas will be decom- posed. By this means there is no difficulty in collecting gas samples from different heights of the glowing layers of coal in blast-furnaces or producers. Collection of Gases Arising from Mineral Springs. The receiver R is connected with the funnel T by means of the rubber tubing q (Fig. 101). All these parts of the apparatus are FIG. 100. COLLECTION AND CONFINEMENT OF GAS SAMPLES. 733 filled with spring-water and the gas is allowed to ascend up through the funnel as shown in the illustration. In order that the gas rray pass from the funnel into the receiver, R is raised so that only the tubing p remains in the water while the funnel is lowered as deep as possible, causing pressure enough to drive the gas over. Th3 tubing is then closed just above a by means of a screw-cock, a beaker filled with spring- water is placed under p, the apparatus removed from the spring, and both ends of R are fused together FIG. 101. with the blowpipe. If the gas is to be analyzed within two or three days, the receiver may be closed by pieces of short rubber tubing each containing a short piece of glass rod with rounded ends. All of such connections must be fastened by means of wires where the glass comes in contact with the rubber. According to the above method the gas arising from the thermal springs of Baden, Switzer- land, was collected and analyzed. * The results obtained showed that it makes but little difference which method is used for closing the receiver, provided the analysis is made within a short time.f 100 c.c. of the gas contained: * " Chemische Untersuchung der Schwefeltherme von Baden (Kanton Aargau)," by F. P. Treadwell, 1896. t The following analyses illustrate the importance of analyzing the gas soon after it has been collected. Both samples were taken at the same time and one was analyzed promptly but the other only after two years and a half had elapsed. The gas receiver was closed by fresh rubber tubing con- taining a piece of stirring rod with rounded edges, and precaution was taken to wire the connections tightly. The rubber remained soft and flexible. I. II. CO 2 = 8.52% 0.43% O 2 = 10.77 14.82 CO = 0.17 CH 4 = 0.15 N 2 = 80.39 84.75 100.00% 100.00% 734 GAS ANALYSIS. Nitrogen 69. 13 Carbon dioxide 30.81 Hydrogen sulphide . 05 Oxygen 0.00 99.99 100.10 Sample I was collected and the ends of the receiver fused to- gether, while with sample II the ends*were closed by means of rub- ber tubing and glass rods, and analyzed five days later. Collection of Gases Absorbed in Spring -water. Of the many different methods which have been proposed for the analyses of the absorbed gases in spring-water, the author has found the following to give the best re- sults. The flask A, Fig, 102, is filled with spring-water up to its upper edge and the rubber stopper con- taining the tube L, which is fused to- gether at the bottom but has an opening on the side at /, is im- mediately placed in the neck and pressed down to the mark. The tube L is raised so that the opening I is within the stop- per, thus making an air-tight connection. FIG. 102. I M 1U ^ . The bulb K is now connected with L, filled half full with distilled water and connected with the capillary tubing C, although the latter is not yet connected with COLLECTION AND CONFINEMENT OF GAS SAMPLES. 735 the measuring-tube B, as shown in the illustration. The levelling- tube N is next raised until mercury begins to flow out of the right-angled capillary tube, when the stop-cock H is closed. After this the water in the bulb K (which is held in an inclined position) is boiled for three minutes, meanwhile warming the capillary tub- Ing connected with the measuring-tube. Unless this last precau- tion is taken, the capillary tubing is likely to break, particularly in winter. After the water in K has boiled vigorously for three minutes, the flame is removed, C is quickly connected with the measuring-tube B and the rubber connection is securely fastened with wire. By boiling the water in K, a complete v^uium is produced in the bulb, so that the gas can be at once collected from the spring-water. For this purpose the tube L is pressed down through the rubber stopper until the opening I comes just below its lower edge, the levelling-tube N is lowered, and the stop-cock H is opened. At once there is a lively evolution of gas from the water in A and this is subsequently maintained by warming the water. As soon as the eudiometer is full the stop- cock is closed and the volume of the gas read after bringing the mer- cury to the same level in N that it is in B. At the same time the temperature of the water in the condenser M is taken by the ther- mometer T and the barometer is read. The gas is then driven over into the Orsat tube containing potassium hydroxide solution (1 : 2) and allowed to remain there for the time being. Meanwhile the boil- ing of the water in A; measurement of the gas in B, etc., are con- tinued until finally no more gas is evolved from the spring-water. All of the gas is driven over into the Orsat tube after its volume has been noted and by means of the caustic potash, 'the carbonic acid is quantitatively absorbed. The unabsorbed gas is again driven over into B and its volume read. By correctly regulat- ing the velocity of the current of water flowing through the con- denser, it is easily possible to maintain a constant temperature throughout the whole of the experiment. The residual gas remain- ing after the absorption of the carbon dioxide consists usually of nitrogen, oxygen, and in some cases methane. It is trans- ferred to the apparatus of Hempel, and analyzed according to methods which will be described further on. According to this method, the determination of nitrogen, 73 6 GAS ANALYSIS. oxygen, and methane gives exact results, but the apparent amount of carbon dioxide is sometimes too much and sometimes too little. If the water contains large amounts of bicarbonate in solution, the carbonic acid found will represent more than was originally present in the free state, for such substances are partly decom- posed by boiling their aqueous solution. On the other hand if only a little bicarbonate is present, the result will be too low, for it is not possible to remove all of ohe free carbonic acid from a solution by boiling it in a vacuum. Consequently, in all cases the free carbonic acid must be deter- mined by computation. For this purpose, in a fresh sample of the water, the total carbonic acid is determined according to p. 393, and then if the composition of the solid constituents present is known, the volume of the free carbonic acid can be calculated. Example. 1000 gms. of Tarasper - Lucius water contain 7.8767 gms. of total carbonic acid. Of this amount, a part of it is present in the water as carbonate (" combined " carbonic acid), an equal amount as " half-combined " carbonic acid, and the re- mainder is free carbonic acid. If from the total amount of carbonic acid the " combined " and " half-combined " acid is deducted (or what is the same thing, double the amount of the " combined " carbonic acid), the difference represents the amount of free carbonic acid present. Calculation of the " Combined " Carbonic Acid. This is obtained by multiplying the difference between the number of. cations and anions (expressed in univalent ions) by the molecular weight of carbonic-acid (COa) ions and dividing by two,* because the sum of the cations in every salt solution is equal to that of the anions present when both are expressed in univalent ions. The " univalent ions " are obtained by dividing the amount in grams of each element (or radical) present by its atomic (or molecular) weight and multiplying by the valence. * For the CO 3 ion is bivalent. COLLECTION AND CONFINEMENT OF GAS SAMPLES. 737 (a) Calculation of the Cations. 1000 gms. Lucius water contain: Sodium Grams. 3.90610 Combining v . Univalent Weight. Valence. I(mg 23.00 = 0.16983 Xl=0. 16983 Potassium 16603 39 10 = 00425 Xl = 00425 Lithium 00914 7 00 = 00131 Xl = 00131 Ammonium 0.01298 18.04 = 0.00072 XI = 0.00072 Calcium 62691 40 09 = 0.01564 X2 = 0. 03128 Strontium ... 00879 87 62 = 00010 X2 = 00020 Magnesium 19040 24 32 00783 X2 01566 Iron 00603 55 85 = 0.00011 X2 = 0. 00022 Manganese . . 00021 54 93 = 000004X2 = 00001 Aluminium 00064 27 1 =0 00002 X3 = 00006 Sum of the cations = . 22354 (6) Calculation of the Anions. Chlorine (Cl) 2.40000 ; Bromine (Br) . 02890 ; Iodine (I) 0.00086: Sulphuric acid (SO 4 ) 1 . 72098 : Boric acid (BO 2 ) . 57600 ; Phosphoric acid (PO 4 ) .... . 00008 Silicic acid (SiO 3 ) . 01421 35.46 = 0.06768 Xl = 0. 06768 79.92 = 0.00035 Xl = 0. 00035 126.92 = 0.00001 Xl = 0. 00001 96.07 = 0.01701 X2 = 0. 03582 43.0 =0.01340 Xl = 0. 01340 95.0 =0.00000 X3 = 0. 00000 76.3 =0.00019 X2 = 0. 00038 Sum of the anions =0.11764 Sum of the cations =0.22354 Sum of the anions =0.1 1764 expressed in univalent ions. CO 3 anions remaining =0.1 0590 As CO 4 is a bivalent ion, half of this amount represents the actual amount of CO 3 ions present: 0.10590 = 0.05295. This corresponds to 0.05295X60=3. 177 gms. CO S Or the " combined " carbonic acid =2.330" CO, 738 GAS ANALYSIS. Calculation of the Free Carbonic Acid. Total amount of carbonic acid (CO 2 ) present 7.877 gms. per liter Amount of "combined" carbonic acid. . . 2.330 " " " Amount of free + " half-combined " carbonic acid . 5 . 547 Amount of " half-combined " carbonic acid 2 . 330 Amount of free carbonic acid. . . T 3 . 217 This weight of carbon dioxide occupies 1637 c.c. under standard con- ditions. By boiling 828.3 gms. of the water, 1868.9 c.c. of C0 2 were obtained at 8.4 C. and 651 mm. pressure, containing only traces of nitrogen. This corresponds to 1851.4 c.c. at C. and 760 mm. pressure, per liter, which is more than the calculated amount, because the carbonic acid gas consisted partly of free and partly of " half-combined " carbonic acid. In cases where the amount of bicarbonate present is very small, the total amount of carbonic acid obtained by boiling the water is always too small. Thus in the case of the thermal water of Baden, by boiling there was obtained : Nitrogen 14. 43 c.c. per liter Carbon dioxide 112.12 " " " 126.55 " " " while from the analysis, the free carbonic acid was computed to be 180.52 c.c. The absorbed gas in the thermal water of Baden is, therefore, Nitrogen 14 . 43 c.c. per liter Carbon dioxide . , , 180.52 " " " 194.95 " " Remark. With the above method of collecting the gas, it is difficult to prevent some water getting into the measuring-tube B, by means of which a small amount of the gas will be reabsorbed. This difficulty is avoided, however, if the flask shown in Fig. 103 is used to contain the water. COLLECTION AND CONFINEMENT OF GAS SAMPLES. 739 FIG. 103. This flask is provided with a short tube blown into its neck near the top and connected by means of thick-walled rubber tubing w^th the mercury reservoir R. In order to determine the con- tents of the flask, a scratch is made on the small tube about 4 cm. from the neck of the flask, tho mercury is driven over just to this mark, and the rubber tubing tightly closed by means of a screw- cock. The reservoir is then emptied of mercury, and the flask is weighed together with the stopper, glass tube L, rubber tubing, and what mercury remains above Q. The flask is then filled with water, the stopper pressed down to the mark m ^ ne nec k f * ne flask, and the tube L is raised until the lower opening I comes within the stopper. After drying the tube L with blotting- paper, the flask and its contents are weighed. Its capacity is then etched upon it. For the determination of the gases absorbed in a liquid, the flask A is filled with it in the same way as in the determination of its capacity, the bulb-tube K, half filled with distilled water is connected with L, and the latter is connected with a capillary tube as shown in Fig. 102. The air is removed from K and the capillary tubing by boiling the water .in the former, as described on p. 735, and the capillary is then connected with the measuring- tube B, Fig. 102. The heavy rubber tubing is now connected with the reservoir as shown in Fig. 103, and the latter is placed in a beaker of hot water. The tube L is introduced into the neck of the flask until the opening / can just be seen, and the gas is expelled in the same way as described on p. 735, except that in this case the liquid is not allowed to rise so high in K. After three-quarters of an hour the gas will be completely expelled from the liquid. The last portions of the gas are driven over into B by lowering the levelling tube N (Fig. 102), raising the mercury 740 GAS ANALYSIS. reservoir R (Fig. 103), and carefully opening the screw-cock Q. A warm stream of mercury will then flow into the flask, expelling the gas into the measuring- tube. As soon as the liquid in A has been driven over as far as the stop-cock H, this is immediately closed. Otherwise the procedure is the same as was described on p. 735. In order to test the accuracy of this method, the author made a few determinations of the oxygen absorbed in the lake-water at Zurich, and the results were compared with those obtained by E. Martz in this laboratory by means of the method of L. Winkler (see p. 760). OXYGEN IN 1 LITER OF ZURICH LAKE-WATER. Modified Pettersson Method. Method of L. Winkler. I II I II 7.66c.c. 7 74 c.c. 7. 67 c.c. 7. 75 c.c. Collection of Gases Absorbed by Defibrinated Blood. The author has used for this purpose the apparatus shown in Fig. 104. The experiment is carried out as follows : First, the rubber tubing which connects N r and A is filled with mercury by raising the levelling tube N', and the pinch- cock h" is closed. T^jien the gas burette C is likewise filled with mercury by raising D and the stop-cock h is turned to the position shown in the drawing. By raising the leveling tube TV, the bulb K and the vessel A are filled with mercury, which is allowed to flow into the funnel M up to the line a and then the stop- cock h is closed. The blood to be examined is poured into M, N' is lowered, and, by carefully opening h, the mercury is allowed to fall from M until it just reached the stop-cock h, which is then closed. M is now filled with blood to the mark 6, h te opened, and the blood is sucked down until its upper level is exactly at the line a, when h is closed once more. The levelling tube N is lowered until a good vacuum is produced in A and the mercury level falls to near the bottom of K. When this is COLLECTION OF GASES ABSORBED BY DEF1BR1NATED BLOOD. 74* FIG. 104. 742 GAS ANALYSIS. accomplished, the gas escapes from the blood so rapidly that all of A is filled with foam; after a few minutes, however, the effervescence subsides and all the blood collects in K. Then, the leveling tube N is raised until the blood reaches the cock In! and the latter is then closed. In this way the greater part of the gas absorbed is separated from the blood. The leveling tube N' is raised, h" opened so that the gas in A is under pressure, and by properly opening h and the burette stop-cock, the gas is driven over into the measuring burette C; when this is accom- plished, h is closed, N' lowered until the mercury reaches the cock h" which is then closed and In! opened. The blood again effervesces, but not so vigorously as before. The bulb K is now surrounded by water at 55, which causes further effervescence from the blood. As soon as the foam subsides, the gas is again driven over into C and the process of evacuating . is continued until the blood ceases to effervesce. The volume of the gas in C is finally read, the temperature and pressure noted, and the analysis carried out as described on p. 775 or 786. The Transference of Gases in Sealed Tubes to the Apparatus Used for the Analysis. We will assume the gas to be contained in R, Fig. 105. Over one of the short tubes connected with the three-way stop-cock H is placed a piece of thick- walled rubber tubing which contains a short piece of heavy glass tubing r. The stop-cock is then revolved so that the rubber tubing is above it and the latter is filled with mer- cury. H is then turned 180 toward the left so that the left and upper tubes communicate with one another. As soon as the mercury begins to run out, the stop-cock is closed. One end of R is then introduced into the rubber tubing containing the mercury so far that its drawn-out point reaches within r, and the rubber tubing is securely fastened by wiring.* In a similar way, the other end of R is connected with the rubber tubing filled with mercury of the levelling-tube N, and after this the stop-cock H is connected with the measuring apparatus W by means of the capillary tubing * Annealed iron wire is used. Copper or brass wire would be likely to become amalgamated with mercury. CALIBRATING GAS MEASURING VESSELS. 743 E. By raising the levelling bulb K, the air is expelled from W and the capillary E, and the mercury is allowed to rise hi the fun- FiG.105. nel T. The stop-cock H is turned so that communication is estab- lished between R and W, and the ends of the former are opened by pressing the capillaries against r and r' '. Then, by raising N and lowering K, the gas is readily driven over into W. Calibrating Gas Measuring Vessels. When vessels are purchased to be used in measuring gases, the correctness of the calibrations should always be tested; the testing can be done with water or with mercury. The calibration with water is carried out in exactly the same wa,y as was described for vessels to be used in measuring liquids (cf. pp. 522-530). 744 GAS ANALYSIS. The calibration by means of mercury will be illustrated by an example. We shall assume that it is desired to calibrate the apparatus shown in Fig. 106. The vessel must be thoroughly cleaned and then placed in a vertical position as shown in Fig. 106 II. The lower capillary a is connected by means of thick-walled rubber tubing with a leveling vessel containing mercury, and the mercury is made to rise slowly in the vessel to a little above the upper mark. The stop-cock is then closed, the leveling tube together with the rubber tubing is removed, and the mercury al- lowed to flow out slowly until the highest point in the meniscus is exactly tangent to the horizontal plane through a'a'. To avoid a parallax error, the reading is taken by means of a telescope placed 2 or 3 mm. away from the glass. The whole contents of the vessel, including the space in the stop-cock, is next al- lowed to run into a tared flask, which is then weighed to the nearest centigram. After determining the temperature of the mercury, its volume can be found by means of the table (top of page 745) pre- pared by Schlosser.* If the weight of the mercury at 20.3 amounted to 2025.26 gms., then its volume corresponds to 2025.26 . = 149.41. feince, however, mercury torms a convex mems- lo.54oo cus and the volume is desired up to the plane aa' , it is evident that the volume of mercury weighed did not include the space a'a-aa.' and, moreover, since the instrument is to be used in the reversed position, the error is really twice as much, as is evident from the inspection of Fig. 106 I. This is called the * Schlosser and Grimm, Z. Chem. App.-Kunde, 2, 201 (1907). I II FIG. 106. CALIBRATING GAS MEASURING VESSELS. 745 WEIGHT OF 1 C.C. OF MERCURY IN AIR AT TEMPERATURES BETWEEN 15 AND 20. Normal temperature 15. Temperature Weight of Mercury. Temperature Weight of Mercury. Temperature Weight of Mercury. Deg. C. Gms. Deg. C. Gms. Deg. C. Gms. 15 13 . 5593 20 13 . 5489 25 13.5385 15.5 13 . 5583 20.5 13 . 5479 25.5 13.5374 16 13.5573 21 13 . 5468 26 13.5364 16.5 13.5562 21.5 13.5458 26.5 13.5353 17 13.5552 22 13.5447 27 13.5343 17.5 13.5541 22.5 13.5437 27.5 13.5332 18 13.5531 23 13 . 5426 28 13.5322 18.5 13.5520 23.5 13.5416 28.5 13.5312 19 13.5510 24 13.5405 29 13.5301 19.5 13 . 5499 24.5 13.5395 29.5 13 . 5291 1 30 13.5280 double meniscus correction. Its value is dependent upon the bore of the tube, as is shown by the following table : TABLE OF MENISCUS CORRECTIONS.* Diameter of Tube in mm. Double Meniscus, Correction for Hg in mgs. Double Meniscus, Correction for thO in mgs. = cubic millimeters Simple Meniscus Cor- rection (H 2 O Hg) in cubic millimeters. 3 76 12 3 4 108 20 6 5 174 31 9 6 314 44 10 7 550 61 10 8 790 81 11 9 1038 106 15 10 1288 134 20 11 1540 167 27 12 1796 204 36 13 2058 245 46 14 2326 289 59 15 2596 336 72 16 2872 387 88 17 3152 441 104 18 3436 499 123 19 3726 560 143 20 4016 624 164 21 4314 691 187 22 4614 757 208 23 4920 821 229 24 5230 881 247 25 5544 938 264 26 5864 991 279 27 6185 1042 293 28 6515 1090 308 29 6845 1135 315 30 7182 1179 324 * W. Schlosser. Private Communication. 746 GAS ANALYSIS. If the diameter of the vessel in question is 20 mm., then the correction, according to the table, would be 4.016 gms. and the true volume will be 2025.26+4.016^2029.276 = 13.5473 " 13.5483 ~~ The volume of this instrument, therefore, is 0.22 cm. less than the intended 150 c.c. The volume of the narrower parts of the tube can be found in a similar manner. The diameter of the tube is best determined by filling with mercury up to a mark, then allowing it to run out until a lower mark is reached, weighing the escaped mercury, and measuring the distance between the two marks with a millimeter rule. If the weight of the mercury is p, the distance between the marks h, the temperature of the mercury 20.3, then the diameter = h XTTX 13.5483* In many cases it is sufficiently accurate to compute the diameter from the circumference of the tube and then subtract twice the thickness of the glass. If it is desired to determine the total volume of a tube provided with stop-corks at both ends, the apparatus is weighed empty and then filled with mercury. In this case, it is obvious that no meniscus correction is necessary. For a measuring vessel calibrated with water, when in a reversed position, the meniscus correction is obtained from the table on page 745. If an instrument calibrated with water is to be used subsequently with mercury, the water meniscus in calibrating the reversed tube occupies a similar position to that of the mercury meniscus when the instrument is in use (see Fig. 107) but the mercury meniscus is not so deep as that of the water. The volume of the gas is therefore found as much too large as the difference between the simple meniscus corrections (H 2 O Hg). Thus if the volume of a gas measuring instrument of 70 mm. diameter is found by weighing with water to be 10.167, according PURIFICATION OF MERCURY. 747 to the table on page 745, then if the instrument is to be used with mercury, the gas volume will be 10.167-0.020=10.147 c.c. Purification of Mercury. Lothar Mayer's Method. * The mercury used for gas-analytical operations must be purified. The principal impurities are copper, cadmium, zinc and sometimes silver and gold. The base metals are removed FIG. 107. FIG. 108. most readily by allowing the mercury to run in a fine stream through about a meter of 8 per cent, nitric acid. This is done in the apparatus shown in Fig. 108. The bottom of the tube B is first filled with impure mercury and the nitric acid is added. The mercury is then poured through the funnel A, the stem of which is drawn out to a capillary and bent to an angle of 60. This causes the mercury to take a zig-zag course as it flows slowly through the nitric acid. The dry mercury that first passes over into the flask C is impure and must be poured into the funnel and allowed to flow through the acid. In this way a fairly pure mercury *Z. anal. Chem., 2, 241 (1863). C. J. Moore (Chem. Ztg., 1910, 735) has used a similar apparatus for purifying large quantities of mercury, but filters through buckskin before allowing it to fall through the acid. 74 8 GAS ANALYSIS. is obtained which can be used as it is for most purposes. If the mercury is to be used for calibrating apparatus, it must be distilled. For this purpose, Hulett's apparatus, shown in Fig. 109, may be used. The mercury .is placed in the long-necked flask K which is connected with the receiver V. The flask is covered with a hood of asbestos paper and heated on the sand bath. Through the arm a, the receiver is connected with a suction pump and at FIG. 109. b a slow current of nitrogen (or carbon dioxide) which has been dried by passing over calcium chloride, is introduced into the flask through the long glass tube that ends in a capillary. The distillation is regulated so that the mercury condenses in the glass arm c, where it leaves the hood of -asbestos paper. About 150-200 c.c. of mercury can be distilled in an hour with this apparatus. Frequently, especially when the nitrogen used contains a little oxygen, the distilled mercury is covered with a thin coating of oxide. This may be removed by filtration. To filter the mercury, the point of a paper filter is perforated several times with a needle, the filter pla'ced in a funnel and the mercury poured SUBDIVISIONS OF GAS ANALYSIS. 749 through. The pure metal runs through the holes in the paper while the impurity remains behind. Subdivisions of Gas Analysis. According to the manner of determining the amount of gas, we distinguish between : 1. Absorption Methods. 2. Combustion Methods. 3. Volumetric Methods. In the case of an absorption ^method -the- mixture of gases is treated with a series of absorbents. The difference in the volumes of the gas before and after it has been acted upon by each absorbent represents the amount of gas absorbed. The absorption of the gas may take place in the measuring-tube itself, or, what is better, in separate absorption vessels. In this way, the amount of carbon dioxide, heavy hydrocar- bons (ethylene, benzene, acetylene, etc.), oxygen, and carbon mon- oxide may be determined in illuminating-gas, producer gas, water- gas, or Dowson gas. After the constituents capable of absorption have been re- moved, a gas residue is left consisting of hydrogen, methane, and nitrogen; the two former constituents are determined by combustion, while the latter is always determined by subtracting the total amount of other gases found from 100 per cent. For a combustion analysis the unabsorbed constituents of the gas mixture are mixed with air, or oxygen, in more than sufficient amount to ensure complete combustion, and burnt in a suitable apparatus ; the amount of combustible gas is determined by measuring the contraction, the amount of carbon dioxide formed, and the excess of oxygen. Finally, if the gas evolved by means of a chemical reaction is measured and from the volume of the latter the weight of the body producing it is calculated, we have made use of what is called a gas-volumetric method. (Cf. Determination of Carbonic and Nitric Acids, pp. 384 and 456). 750 GAS ANALYSIS. DETERMINATION OF THE GASES. Carbon Dioxide, C0 2 . Mol. Wt. 44. Density = 1.5290 * (Air = 1) . Weight of 1 liter = 1.9767 gms. Molar volume =22.26 1. Critical temperature = +31.5 C. Carbon dioxide is absorbed to a* considerable extent by water; 1 vol. water absorbs: At C , . . _. . . .,, . ,, . . 1.7967 c.c. CO 2 " 15 C ... . ; . . : . ... ... ... .... .;. . . 1.0003 " " " 25 C ,. ... ... .,. .,. . r . ~ _ . . 0.8843 " " or in general /?t = 1.7967 -0.07761 X* +0.0016424 X* 2 , Absorbent. Potassium Hydroxide Solution 1 : 2. 1 c.c. of caustic potash of the above strength will absorb at least 40 c.c. of CC>2. Sodium hydroxide solution is not used on account of the difficult solubility of sodium bicarbonate. Small amounts of CC>2 may be absorbed by means of a definite amount of standardized Ba(OH) 2 solution, and the excess of the N latter titrated with HC1, using phenolphthalein as indicator. (Seep. 593). * This number is the mean from the observations of Lord Rayleigh (1897) = 1.52909, Leduc (1898) = 1.52874, and Christie (1905) = 1.52930. f /? is called the absorption coefficient of the gas. This signifies the volume of gas, measured at and 760 mm. pressure, which 1 c.c. of a liquid at t will absorb when the pressure upon the surface of the liquid is 760 mm. If h c.c. of liquid, at t and B mm. pressure, absorb Vt c.c. of the gas, then the absorption coefficient can be computed by the equation: Vt THE HEAISY HYDROCARBONS. 751 The Heavy Hydrocarbons. Ethylene (Ethene) ; C 2 H4; Benzene, C 6 H 6 ; Acetylene (Ethine), Ethylene, C 2 H 4 . Mol. Wt. 28.03. Density =0.9739* (Air = 1) . Weight of 1 liter = 1.2590 gms. Molar volume =22.27. Critical Temperature = + 9 C. Preparation of Ethylene. One of the most satisfactory methods consists in treating an alcoholic solution of ethylene bromide with zinc dust f : % C 2 H 4 Br 2 + Zn =ZnBr 2 + C 2 H 4 . A round-bottomed flask of about 200 c.c. capacity, and having a short wide neck, is chosen for the experiment. In the neck is inserted a rubber stopper with three holes, carrying respectively a safety tube provided with mercury seal, a gas delivery tube, and a dropping funnel. A sufficient amount of zinc dust, moistened with alcohol, is placed in the. flask and gently heated at the start by placing the flask in a bath of warm water at about 50 C. A mixture of 1 part ethylene bromide and 20 parts absolute alcohol is placed in the dropping funnel, and allowed to flow slowly upon the zinc dust. The escaping gas is passed first through olive oil, to remove a little ethylene bromide which is carried over mechanically, then through caustic potash solution, and finally through water; it is collected over mercury, perhaps in the Drehschmidt pipette (Fig. 105, p. 743). The gas thus obtained is almost pure, par- ticularly when the mixture of ethylene bromide and alcohol has stood for some time over anhydrous sodium carbonate to remove traces of hydrobromic acid. A sample of the gas prepared by W. Misteli was found to con- *M. Bretschger (Inaug. Dissert. Zurich, 1911) found the density of ethylene to be 0.9724, but M. Stahrfoss and P. A. Guye (Arch. soc. phys. et nat, 28 1909) found the val,ue 0.9758. In the text the mean of these two values is used. t Gladstone and Tribe, Berichte, 7, 364 (1874). 752 VOLUMETRIC ANALYSIS. tain 98.84 per cent, of ethylene, 1.00 per cent, of hydrogen and 0.16 per cent, of nitrogen. Absorption Coefficient for Water. I volume of water absorbs at C 0.256 c.c. C 2 H 4 15 C , 0.161 " " 20 C 0.149 " " or in general, P =0.25629 =0.00913631^+0.000188108^. Alcohol absorbs more ethylene; the general formula is 0=3.594984 -0.077162 * + 0.0006812 t\ Absorbents. 1. Fuming sulphuric acid* (with 20 to 25 per cent, free SO 3 ), 1 c.c. absorbs 8 c.c. of C2rLt. 2. Bromine water, f Ammoniacal cuprous chloride solution will also absorb ethy- lene. By means of bromine, the ethylene is absorbed with the for- mation of ethylene bromide, C2rl4Br 2 . If the absorption is effected with a titrated bromine water, the amount absorbed can be determined by titrating the excess of bromine. This excellent method, proposed by Haber,{ is at present the best known for the determination of ethylene in the presence of benzene. (See p. 818.) Benzene, C 6 H 6 . Mol. Wt. 78.04. 78.04 gms. of benzene vapor occupy a volume of 22.391 liters under standard conditions. Benzene is readily soluble in alcohol, ether, carbon bisulphide, caoutchouc, ethylene bromide, bromine, and fuming sulphuric acid. * Ethionic acid, C 2 H 6 S 2 O 7 , is formed. tTreadwell and Stokes, Berichte, 21 (1888), p. 3131. J Haber and Oecheihauser, Berichte, 29, p. 2700. BENZENE. 753 Absorbents. Fuming sulphuric acid* and bromine water con- taining an excess of bromine. Inasmuch as benzene is neither brominated nor oxidized by bromine at ordinary temperatures, it was difficult to under- stand why bromine water should absorb it quantitatively. In fact, Berthelotf and Cl. Winklerf disputed it, but the results of Treadwell and Stokes have recently been confirmed by Haber. He suggested that the absorption of benzene by bromine was of a purely physical nature, and M. Korbuly|| has shown that such is the case. Just as bromine can be removed from aqueous solution by shaking with benzene, so benzene can be removed by shaking with bromine, or even ethylene bromide and other like solvents. By means of highly concentrated nitric acid (specific gravity 1.52) benzene is also absorbed; this solvent cannot be used in the analysis of gases containing carbon monoxide, for the latter is quantitatively oxidized to carbon dioxide by nitric acid of this strength, and is therefore removed with the benzene If when the acid vapors are neutralized by caustic potash solution. Behavior of Benzene to Water. Benzene vapors are absorbed to a considerable extent by water and all aqueous salt solutions, a circumstance which must be considered when an exact gas analysis is to be made. In order to determine how much benzene is absorbed by water, M. Korbuly performed the following experiments: Different amounts of air containing 3.16 per cent, of benzene vapor were shaken in a Drehschmidt's pipette with the same amount of water (5 c.c.) until no more benzene was absorbed. He obtained the following results: * Benzene sulphonic acid is formed, C 6 H 6 SO 3 . t Compt. rend., 82, p. 1255. j Zeitschr. f. anal. Chem., 1889, p. 281. Treadwell and Stokes, loc. cit. |l Inaug. Dissertation, Zurich, 1902. H Treadwell and Stokes, loc. cit. 754 GAS ANALYSIS. Experiment. Gas Taken in c.c. Per Cent. Benzene Present by Volume. Amount of Benzene Absorbed at the End of Three Minutes. 1 58.92 3.16 1.28c.c. = 2.17% 2 61.14 3.16 0.80 " =1.31% 3 58.32 3.16 0.52 " =0.89% 4 59.86 3.16 0.44 " =0.73% 5 60.78 3.16 0.28 " =0.46% 6 59.88 3.16 0.08 << =0.01% 7 60.20 3.16 0.02 " =0.00% Potassium hydroxide behaves similarly. In the analysis of a mixture of carbon dioxide and benzene, it is customary to first remove the carbon dioxide by means of potassium hydroxide solution and then the benzene with fuming sulphuric acid or bromine. It is evident, then, that both of the results obtained will be inaccurate if a fresh solution of potassium hydroxide is used for the absorption of the carbon dioxide, for this will absorb not only the whole of the carbon dioxide, but in many cases nearly all of the benzene. Accurate results may be obtained by using a solution of potassium hydroxide which has been saturated with benzene vapors. Acetylene, C 2 H 2 . Mol. Wt. 26.02. Density =0.9134 * (Air = 1) . Weight of one liter = 1.1808 gms. Molar volume =22.03 1. Critical temperature = +37 C. Boiling Point =-80.6 C. Acetylene is quite soluble in water; 1 volume of water at the ordinary temperature absorbs an equal volume of this gas. In amyl alcohol, chloroform, benzene, glacial acetic acid, and acetone it is much more soluble; thus 1 volume of acetone absorbs 31 volumes of acetylene. f * M. Bretschger (Inaug. Dissert. Zurich, 1911) found the density of acetylene = 0.9157, whereas M. Stahrfoss and P. A. Guye (Arch. sci. phys. et nat., 28, 1909) found it -0.9120. The mean of these two values is 0.91335. f Hempel, Gasanalytische Methoden (1900), p. 206. PREPARATION OF PURE ACETYLENE. 755 Preparation of Pure Acetylene. (a) Method of M. Bretschger.* The crude acetylene, prepared from calcium carbide, is passed through an acid solution of copper sulphate, then through aqueous chromic acid, caustic potash, and finally over slaked lime; it is then subjected to a fractional distillation. The gas is passed through a small bulb cooled by liquid air which causes the acetylene to solidify. By gentle warming, the acetylene is then evaporated and is caused to pass through calcium chloride tubes. (6) Method of M. Stahrfoss and P. A. Guyej The impure acetylene, prepared from calcium carbide, is passed through a solution of potassium permanganate, then through caustic potash solution and finally over phosphorus pentoxide. It is frozen by means of liquid air and then fractionated. The method of preparing acetylene by decomposing copper acetylide cannot be recommended, because the gas is then strongly contaminated with ethylene (C2H4) and vinyl chloride (C2H3C1). Thus M. Bretschger { found from 5 to 10 per cent, of ethylene in such gas. Absorbents: Fuming sulphuric acid. By bromine water acetylene is absorbed extremely slowly in the cold, a fact which permits the titration of ethylene in the presence of acetylene (see page 821). By means of ammoniacal cuprous chloride, acetylene is absorbed and forms red copper acetylide (Cu2C2H2)O. This reaction is so characteristic that it is used for the Qualitative Detection of Acetylene in gas mixtures. This test is best performed by the method of L. Ilosvay von Nagy Ilosva.|| * Loc. tit. f Private Communication from Professor Guye. J Loc. cit. C 2 H 4 SO 4 is formed. || Berichte, 32 (1899), p. 2698. 756 GAS ANALYSIS. Preparation of the Reagent. One gram of copper nitrate (chloride or sulphate) is placed in a 50-c.c. measuring-flask and dissolved in a little water. To the solution 4 c.c. of concentrated ammonia (20-21 per cent. NH 3 ) and 3 gms. of hydroxylamine hydrochloride are added, and the liquid is shaken until it becomes colorless, when it is immediately diluted with water up to the mark. The Qualitative Test. A few cubic centimeters of the reagent are placed in a 500-c.c. glass-stoppered cylinder and the gas to be tested for acetylene (illuminating-gas) is passed over it until the color of the reagent becomes pink. The cylinder is then stoppered and its contents thoroughly shaken. If acetylene is present, a beautiful red precipitate is immediately formed. Another method of making the test is to pass the gas through a small bulb-tube containing glass-wool moistened with the reagent. Remark. If the reagent is placed under petroleum it can be kept for about one week, but if copper wire is added to the solu- tion, it can be kept for a much longer time, as L.Pollak has shown. Such a solution gave a distinct reaction after it had been kept for one year, but the precipitate obtained, instead of being a bright red, was more the color of sealing-wax. The solution is much less permanent when it is prepared from the chloride or sulphate, even when copper is added to it. Without the copper, the chloride would give no reaction after being a week old, and with the addition of copper it was spoiled at the end of two weeks. The sulphate behaved about the same. Separation of the Heavy Hydrocarbons from One Another. It has been attempted repeatedly to separate ethylene from benzene, but usually in vain. The separation as proposed by Berthelot, of absorbing the ethylene with bromine water and afterwards removing the benzene by means of concentrated nitrio .acid, is erroneous in every respect.* The method of Harbecl: and Lunge f is correct in principle but very tedious, and the original modification of Pfeiffer % always gives too high results. * Treadwell and Stokes, loc. cit. f Zeit. f. anal. Chem., XVI (1898), p. 26. J J. f. Casbeleuchtung und Wasserversorgung, 1899, p. 697, and Berichte, 29, p. 2700. OXYGEN. 757 Recently Pfeiffer * has improved' his method so that it gives the same results as that -of Harbeck and Lunge. Haber and Oechelhauser,f on the other hand, have devised a method which is accurate and to be recommended. Principle. In one portion of the gas, the sum of the ethylene and benzene is determined by absorption with bromine water or fuming sulphuric acid, while in a second portion the gases are absorbed in titrated bromine water, and the excess of the latter is determined iodimetrically. From the amount of bromine re- quired the ethylene is calculated: N 1 c.c. 1=1.114 c.c. C 2 H 4 at C. and 760 mm. pressure. As this analysis is performed in the Bunte burette, it will not be described in detail until we have become acquainted with this important form of apparatus. (See p. 798.) Oxygen, O = I6. Mol. Wt. 32. Density =1.1053 (Air=l). Weight of 1 liter =1.4289 gms. Molar volume = 22.39 1. Critical temperature = -119 C. Oxygen is only slightly soluble in water; according to the experiments of L. W. Winkler,J one liter of water at 60 mm. pressure will absorb the following quantities of air : ABSORPTION COEFFICIENTS OF ATMOSPHERIC AIR. Temperature. (1000 c Oxygen 10 . 24 c .c. absorbed) Nitrogen. c. 18.57c 16.45 14.67 13.29 12.19 11.31 10.59 9.92 9.35 8.93 8.59 8.31 Air. c. 28.81c 25.43 22.64 20.45 18.69 17.24 16.06 15.03 14.18 13.51 12.97 12.53 c. 5 8 . 98 10 7 97 15 7 16 20 . 6 50 25 . . . 5 93 30 . . 5 47 35 5 11 40 4 . 83 45 4 . 58 50 4 38 55 4 22 * See Chem. Ztg., 1904 (884). t J. f. Gasbeleuchtung und Wasserversorgung, 1896, p. 804, and Berichte, p. 2700. JBerichte, 34, 1410 (1901). GAS ANALYSIS. From these data, the absorption coefficient of pure oxygen for water at to 55 can be computed. ABSORPTION COEFFICIENTS OF OXYGEN FOR WATER. Temperature. /? Temperature. /? 0., 0.04890 30 0.02608 5 0.04286 35 0.02440 10 0.03802 40 0.02306 15 0.03415 * 45 0.02187 20 0.03102 50 0.02090 25 .02831 55 .02012 Oxygen can be determined by combustion or by absorption. The Determination of Oxygen by Combustion. The determination of oxygen by combustion may be effected by exploding it with hydrogen (Bunsen) or by conducting a mixture of the two gases through a glowing platinum capillary (Drehschmidt), exactly as in the determination of carbon mon- oxide (cf. p. 765). In both cases the combustion takes place in accordance with the equation : O + H 2 = H 2 1 vol. 2 vols. vol. Three volumes of gas, therefore, disappear for each volume of oxygen present. If the contraction resulting from the com- bustion of a mixture of oxygen and an excess of hydrogen is designated by V c , then the amount of oxygen present =JVc. The Determination of Oxygen by Absorption. The absorbents of oxygen are : 1. Alkaline Pyrogallol Solution (Liebig). One volume of a 22 per cent, aqueous solution of pyrogallol is mixed with five or six times as much potassium hydroxide solution (3:2). 1 c.c. of this solution absorbs 12 c.c. of oxygen. OXYGEN. 759 At a temperature of 15 C., or higher, the absorption takes place quickly; the oxygen in 100 c.c. of air will be absorbed in three minutes or less. At lower temperatures the absorption takes place less readily and at C. the above quantity of oxygen cannot be absorbed completely in half an hour. A pyrogallol solution of the above concentration will not evolve carbon monoxide during the absorption. 2. Phosphorus (Lindemann). The absorption of oxygen by means of phosphorus takes place by simply allowing the gas containing the oxygen to remain over moist phosphorus. The formation of white clouds indicates the presence of oxygen, and their disappearance shows that the absorption is complete. A temperature of 15 to 20 C. is best suited for the absorption. The oxygen is completely absorbed at the end of three min- utes from 100 c.c. of air at this temperature. At lower temper- atures the absorption requires more time and at more than an hour is necessary. If the gas contains more than 60 per cent, of oxygen, moist phosphorus will absorb none of it at the ordinary atmospheric pressures. In this case the gas must be diluted with nitrogen or hydrogen until a mixture is obtained containing less than 60 per cent, oxygen, or the gas must be allowed to act upon the moist phosphorus under diminished pressure. In the latter case, how- ever, the phosphorus easily becomes heated enough to melt it. Further, oxygen is not absorbed by moist phosphorus if the gas contains traces of heavy hydrocarbons, ethereal oils, alcohol, or ammonia. According to Hempel* 0.04 per cent, of ethylene, and according to Haberf 0.17 per cent., suffices to prevent com- pletely the absorption of oxygen. * Gasanalytische Methoden. t Experimental-Untersuchung liber Zersetzungen und Verbrennungen von Kohlenwasserstoffen, Habilitationschrift, Munich, 1896. 760 GAS ANALYSIS. 3. Chromous Chloride. Consult the paper by Otto von der Phordten, Anhal. Chem. Phys. 228, 112. 4. Copper. The gas is either conducted over glowing copper, or it is intro- duced into a Hempel pipette containing rolls of copper gauze and an ammoniacal solution of ammonium carbonate. 5. Sodium Hydrosulphite,* N a 2 2 04 (Franzen |). An -alkaline solution of sodium hydrosulphite, which can now be obtained commercially at a low price, is an excellent absorbent for oxygen. The reagent may be prepared for use in the Hempel pipette by dissolving 50 gms. of the salt in 250 c.c. water and 40 c.c. caustic potash solution (500 KOH:700H 2 O). For absorption in the Bunte burette, the above solution is too concentrated; in this case 10 gms. hydrosulphite in 50 c.c. water plus 50 c.c. of 10 per cent, caustic soda, may be used. The absorption takes place in accordance with the equation : 2Na 2 S 2 O 4 + 2H 2 + 2 = 4NaHS0 3 . Sodium hydrosulphite has the advantage over all other absorbents that the absorption is always complete at the end of five minutes. Determination of Absorbed Oxygen in Water. Method of L. W. Winkler.J 1000 c.c. NazSjOj, solution = = =0.8 gm. or 559.8 c.c. oxygen at and 760 m.m pressure. Principle. If water containing dissolved oxygen be heated in a closed vessel with manganese hydroxide, the latter is oxidized to manganous acid according to the following equation : Mn(OH) 2 + 0=H 2 MnO 3 . * This is really sodium hyposulphite, although sodium thiosulphate, Na 2 S 2 O 3 , is commonly called "hyposulphite." t Berichte, 39, 2069 (1896). % Ibid, 21 (1888), p. 2843. DETERMINATION OF ABSORBED OXYGEN IN H/ATER 761 The amount of oxygen taken up is determined iodimetrically by adding hydrochloric acid and potassium iodide to the manga- nous acid and titrating the liberated iodine, H 2 MnO 3 +4HCl = MnCl 2 +2H 2 0+Cl 2 and 2KI+C1 2 = 2KC1+I 2 . Hence 1 gm.-at. 1 = 8 gms. = 5597.8 c.c. oxygen at C. and 760 mm. pressure. Reagents Required. 1. An approximately 4N - MnCl 2 solution obtained by dissolving 400 gms.of MnCl2+4H2O in water and dilut- ing to 1000 c.c. The manganese chloride must be free from iron. 2. Sodium Hydroxide Solution Containing Potassium Iodide. On account of the nitrite usually present in commercial sodium hy- droxide, the alkali solution is prepared from sodium carbonate and calcium hydroxide. The clear liquid is siphoned off and con- centrated in a silver dish until its specific gravity is 1.35. In 100 c.c. of this solution, 10 gms. of potassium iodide are dissolved. A portion of the alkaline potassium iodide solution on being acidified with hydrochloric acid should not immediately turn starch paste blue, and, furthermore, large amounts of carbonate must not be present. N 3. Sodium Thiosulphate Solution. Procedure. A glass-stoppered flask of about 250-c.c. capacity is taken and its exact capacity is determined by weighing it first- empty and then filled with water at 17.5 C. If the water to be analyzed is saturated with air, it is simply poured into the flask, otherwise the water is conducted through it for ten minutes. Then, by means of a pipette reaching to the bottom of the flask, 1 c.c. of the alkaline potassium iodide solution is introduced and immediately afterwards 1 c.c. of the manganese chloride solution. The flask is closed, shaken, and allowed to stand until the precip- itate has settled. Then, by means of the long-stemmed pipette, about 3 c.c. of concentrated hydrochloric acid are introduced and the contents of the flask once more shaken. The precipitate dis- solves readily with liberation of iodide and the latter is titrated with sodium thiosulphate in the usual way. Remark. The results obtained by this method agree closely with those obtained by boiling the water as described on p. 739. 762 GAS ANALYSIS. Carbon Monoxide, CO. Mol. Wt. 28. Density -0.96702 (Air = l). Weight of 1 liter 1.2502 gms. Molar Volume = 22. 3 97 liters. Critical temperature = 136 C. Preparation. Some concentrated sulphuric acid is heated in a fractionating flask to a temperature of 140 to 160 C. upon an oil bath, and formic acid (sp. gr. 1.2) is allowed to drop into it: HCOOH = In order to free the escaping gas from water and acid vapors, it is conducted first through a Liebig condenser, which leads to an empty flask to receive the condensed water, and from thence into a concentrated caustic potash solution. This method * yields about 60 liters of carbon monoxide in half an hour, using about 500 c.c. of concentrated sulphuric acid. The method of Wade and Panting, f according to which very pure carbon monoxide can be prepared by allowing concentrated sulphuric acid to drop upon potassium cyanide, is not, according to Allner, a suitable process for preparing large quantities of the gas; because considerable potassium cyanide becomes enveloped in pyrosulphuric acid during the reaction, so that there is con- siderable danger involved in working with the residues. By the action of hot concentrated sulphuric acid upon oxalic acid, it is very easy to prepare a mixture of equal volumes carbon monoxide and carbon dioxide; on account of the large amount of the latter, however, this method is less satisfactory than that of Allner. The gas is only very slightly soluble in water; ABSORPTION COEFFICIENTS OF CARBON MONOXIDE FOR WATER. f Temperature. Temperature. /? ............ 0.03537 30 ............ 0.01998 5 ............ 0.03149 35 ............ 0.01877 10 ............ .02816 40 ............ .01775 15 ............ 0.02543 45 ..... ....... 0.01690 20 ............ .02319 50 ............ .01615 25 ............ .02142 55 ............ .01548 * W. Allner, Inorg. Dissert. Karlsruhe, 1905. t J. Chem. Soc., 73, 255. j L. W. Winkler, Berichte, 34, 1414 (1901). CARBON MONOXIDE. 763 In alcohol the gas is about ten times more soluble than it is in water. Its determination is effected either by absorption or by com- bustion. Absorbents. Ammoniacal Cuprous Chloride. 200 gms. of com- mercial cuprous chloride are shaken in a closed flask with a solu- tion of ammonium chloride (250 gms. in 750 c.c. water), and to every three volumes of this mixture 1 vol. of ammonia, specific gravity 0.91, is added. In order that the solution may remain active, a spiral of copper wire is introduced into the flask long enough to reach from the bottom up to the stopper. 1 c.c. of this solution will absorb 16 c.c. of carbon monoxide. Formerly it was the almost universal custom to absorb this gas by means of a hydrochloric acid solution of cuprous chloride, but to-day this is not done on account of the following reasons. The absorption of carbon monoxide by means of cuprous chloride takes place according to the. following equation: Cu 2 Cl 2 + 2CO^Cu 2 Cl 2 .2CO.* The compound Cu 2 Cl 2 .2CO is extremely unstable and can only be formed when there is a certain pressure exerted by the carbon monoxide, so that when the acid solution is used the absorption will never be quantitative. Further, if a gas free from carbon monoxide (nitrogen or hydrogen) is shaken with such a solution after it has been used several times, a part of the Cu 2 Cl 2 .2CO in solution will be decomposed according to the above equation in the direction of right to left, until the partial pressure of the carbon monoxide set free is sufficient to restore equilibrium. Consequent^ the volume of the gas will appear greater after it has been treated with the cuprous chloride solution than it was originally. When an ammoniacal cuprous chloride solution is employed.. the absorption of the carbon monoxide is almost quantitative, but after such a solution has been used repeatedly it will readily give up some of the gas, although not so readily as is the case of the 'solution of cuprous chloride in hydrochloric acid or calcium chlo- * The compound has been isolated in the solid state, according to W. A. Jones (Am. Chem. J., 22, 287) its formula is CuCl 2 -2CO-4H 2 O, but according to the experiments of C. v. Girsewald in the author's laboratory, the formula isCu 2 C! 2 .2CO.2H 2 O. 764 GAS ANALYSIS. ride.* It is advisable, therefore', to adopt the suggestion of Dretn schmidt, and first absorb the greater part of the gas by means of an old solution of cuprous chloride, afterwards removing the last traces by means of a freshly-prepared solution, or one which has beer, used but a few times. Besides carbon monoxide, the ammoniacal cuprous chloride solution will absorb acetylene, ethytene, etc., so that these gases must be removed previously by means of fuming sulphuric acid or bromine water. By long shaking with concentrated nitric acid (specific gravity 1.5), carbon monoxide is completely oxidized to carbon dioxide, and the latter can be removed by treatment with potassium hydroxide solution.f Determination of Carbon Monoxide by Combustion with Air or Oxygen. The following reaction shows how carbon monoxide may be determined by combustion: CO + O = C0 2 . 2 vols. 1 vol. 2 vols. From the reaction we can make the following deductions: 1. The difference in the volume of the gas mixture before and after the combustion is for 2 vols. CO; 3 2=1 and for 1 vol. CO = ^. This difference is designated as the contraction. The contraction caused by the combustion of carbon monoxide is, there- fore, equal to half the original volume of CO. 2. The volume of the carbon dioxide formed is equal to the volume of the carbon monoxide originally present. If. then, the carbon dioxide is determined by absorption with caustic potash, the volume of the carbon monoxide is at once obtained, provided no other combustible gas containing carbon is present at the same time. * Cuprous chloride is soluble in a concentrated solution of calcium chlo- ride. 1 c.c. of this solution absorbs 12 to 15 c.c. of CO. tTreadwell and Stokes, Berichte, 21, p. 3131. CARBON MONOXIDE. 765 3, For the combustion of 2 vols. of CO, 1 vol. of oxygen is neces- sary, and consequently the amount of oxygen consumed is equal to half the volume of the carbon monoxide. Methods of Effecting the Combustion. The combustion of the carbon monoxide can be carried out in several different ways: 1. By explosion. 2. By conducting the gas over glowing palladium or platinum. 3. By conducting the gas over copper oxide. 1. Combustion by Explosion. The gas is mixed with a sufficient amount of air in a measuring vessel, such as is shown in Fig. 105, and the latter is connected by means of the capillary E with the Hempel's explosion pipette shown in Fig. 110. The gas is completely driven over into the latter so that the capillary is entirely filled with mercury, the stop-cocks of the capillary and of the explo- sion pipette are both closed, and an electric spark is made to pass be- tween the two platinum points which are fused into the glass walls of the pipette; this immediately causes an explosion to take place. After- wards the gas is once more driven FIG. 110. , . . x back into the measuring burette, and its volume again determined. The difference in volume before and after the explosion represents the contraction. This most excellent method can in some cases lead to erroneous results. In practice, it is almost always a question of determining the amount of combustible gas in a mixture containing nitrogen obtained after treatment with the different absorbents. If the amount of combustible gas present is too small in proportion to the amount of non-corr.bustible gas, there will be no combustion what- soever; while on the other hand, if this relation is too large, a part of the nitrogen will be burnt to nitric acid (hydrogen is usually 766 GAS ANALYS S. present). According to Bunsen, the combustion is complete when 30 parts of combustible gas are present for each 100 parts of non- combustible gas. Consequently, if the explosion method is to be used for the analysis, the approximate composition of the gas must be known. 2. Combustion by Conducting the Gas ovei Glowing Palladium. This is the most certain of all methods for effecting the com- bustion, because it is entirely independent of the proportion of combustible gas present,* and there is no danger of any of the nitrogen being oxidized. The combustion is best effected, as proposed by Drehschmidt, by passing the gas through a thick- walled platinum capillary tube containing three palladium wires. The platinum capillary (Fig. 105, V) is placed between the gas burette and the Drehschmidt pipette S (Fig. 105), and it is heated by means of the non-luminous flame of a Teclu burner. The gas is repeatedly passed through the glowing capillary until there is no further diminution in volume, showing the combustion to be complete. There is no danger to be feared from explosions even when pure detonating gas is passed through the platinum tube, and by this method CO, H, and CH 4 are completely oxidized. In the analysis of gases containing only small amounts of the above gases (e.g. exhaust gases from gas-motors) the so-called fractional combustion is employed. By this means either hydrogen and carbon monoxide are oxidized while methane is not, or car- bon monoxide is alone burned. Fractional Combustion. If, according to Haber,f an abso- lutely dry gas mixture, consisting of considerable nitrogen and oxygen with little carbon monoxide, hydrogen, and methane, is slowly conducted (at the rate of about 700- 800 c.c. per hour) through a glass U tube 3 mm. in diam- eter which contains a palladium wire 55 cm. long, folded into three lengths of about 18 cm., then if the temperature of boiling sulphur is maintained, the hydrogen and carbon monoxide will be completely burned, while methane will escape from the tube * It is only necessary to make sure that a large excess of oxygen is pres- ent (cf. Hempel, Zeitschr. f. anorg. Chem., XXXI (1902), p. 447. t Loc. cit. CARBON MO\OXIl)E. 7 6 7 in an unchanged condition. By connecting the U tube with a weighed calcium chloride tube and then with two weighed soda- lime tubes (see p. 380) the increase in the weight of the former will show the amount of water formed from the hydrogen, and the gain in weight shown by the soda-lime tubes corresponds to the amount of carbon dioxide formed from the carbon monoxide. If, after passing through the soda-lime tubes, the gas is passed through a combustion-tube filled with platinized asbestos, or copper oxide, which is heated to a dark-red heat, the methane is quantitatively burned to water and carbon dioxide; the former is absorbed in a calcium chloride tube and the latter in two soda- lime tubes, all three tubes being weighed before the gas is passed through them. In this way a check is obtained upon the accuracy of the determination, for the proportion of carbon to hydrogen found should be 1:4. The combustion of carbon monoxide alone from a mixture of this gas with hydrogen, methane, and air can be effected satis- factorily as follows: After the gas has been freed from CO 2 , unsaturated hydrocarbons, and aqueous vapor, it is conducted through a U tube * containing 60_70 gms. of pure iodine pentoxide f heated to 160 C. ; by this means the carbon monoxide is alone oxidized with liberation of iodine according to the equation I 2 o 6 +5co=5co 2 -r-i 2 4 If the gas is now conducted through two Peligot tubes containing potassium iodide solution, the iodine will be absorbed and can N be titrated at the end of the experiment with sodium thiosulphate solution. 1 c.c. under standard conditions. * The U tube is heated in a small paraffin bath. f Iodine pentoxide is prepared by heating iodic acid in a current of dry air at 10 until the water is completely removed. J Nicloux, Compt. rend., 126, p. 746, and Kinnicutt, Journ. of the Am, Chem. Soc., XXII, p. 14. N 1 c.c. Na 2 S 2 O 3 solution corresponds to 5.6 c.c. CO, measured / ..N A LYSIS. If, after the carbon dioxide and water have once more been removed from the gas, it is passed through a combustion-tube half filled with copper oxide and half with platinized asbestos, both heated to dark redness, the hydrogen and methane will be completely burned to water and carbon dioxide, which can be absorbed and weighed as before. From the amounts of each, the hydrogen and methane present in the gas can be calcu- lated. Qualitative Detection of Traces of Carbon Monoxide in the Air. If blood be diluted with water until the solution shows only a slight red color, it will give a characteristic absorption spectrum; two dark absorption bands appear between the D and E lines. If to this dilute blood solution a few drops of a concentrated, freshly-prepared ammonium sulphide are added, the dark bands disappear, and instead a single broad band will appear at a place between the positions of the previous bands. Blood containing carbon monoxide behaves quite differently. When the latter gas is present, the blood takes on a rose color and the solution gives almost the same absorption spectrum as pure blood (the bands shift slightly toward the violet) but in this case the two bands do not disappear on the addition of ammonium sulphide. To detect traces of carbon monoxide in the air, Vogel directs that a 100-c.c. bottle, filled with water, be emptied in the room containing the gas, .and that 2 to 3 c.c. of blood, highly diluted with water, and showing only a very faint red color (although still giving the blood spectrum in a column as thick as a test-tube) be poured into the bottle and shaken for some minutes. To the solution a few drops- of ammonium sulphide solution are added and the liquid is examined by means of the spectroscope. If the two bands are now visible, carbon monoxide is present. Ac- cording to Vogel as little as 0.25 per cent, of CO can be detected in this way. Hempel has improved this method to a marked degree. He found that it was not possible to completely remove small amounts of carbon monoxide by shaking with the dilute solution CARBON MONOXIDE. 769 of blood, and furthermore concentrated blood solutions could not be used because they foam so much. By using a living animal, its lungs furnish a better means of absorption, for the gas then comes in contact with undiluted blood. A mouse is placed between two funnels which are joined together by means of a broad band of thin rubber and the gas to be tested is passed through the funnels at a speed of ten liters per hour. At the end of two or three hours the mouse is killed by immersing the funnels in water and a few drops of its blood are taken from the region near the heart. In this way Hempel was able to detect with certainty as little as 0.032 per cent. CO. With such small amounts of CO the live mouse showed no symptons of poisoning; this was first apparent when 0.06 per cent, of the gas was present. In the latter case after half an hour the mouse^breathed with difficulty and lay exhausted on its side. Potain and Drouin detect small amounts of carbon monoxide by passing the gas through a dilute solution of palladous chloride, when metallic palladium is precipitated: 2HCl+CO 2 +Pd. The solution is decolorized, or turns a pale gray, when large amounts of CO are present, but appears a light yellow in color when only traces are present. In order to estimate better the decrease in color, Potain and Drouin filter off the deposited palladium and compare the color of the filtrate. For the detection of small amounts of carbon monoxide, C. Winkler recommends a method which, as the author has found, will often lead to error. According to Winkler, the gas to be tested is conducted through a solution of cuprous chloride in a saturated solution of sodium chloride, afterwards diluting with four to five times as much water, causing the precipitation of snow-white cuprous chloride. If this turbid solution is treated with a drop of sodium palladous chloride, a black precipitate of metallic palladium is obtained. Unfortunately, however, the palladium is often precipitated even in the absence of a trace of carbon 77 ^ GAS monoxide; for cuprous chloride itself will readily reduce salts ol palladium. It is true, on the other hand, that at a definite concentration the reduction of the palladous chloride is only effected by means of carbon monoxide, but it is difficult to always obtain the right conditions, and herein lies the inaccuracy of the method. If the solution be too concentrated with respect to sodium chloride, even large amounts of carbon monoxide will fail to precipitate a trace of palladium, because in that case the solution contains not only copper but also palladium in the form of complex sodium salts : [Cu 2 ClJNa 2 and [PdClJNa 2 . The sodium palladous chloride is not reduced by carbon mon- oxide and there is even less likelihood of the two sodium salts acting upon one another. If the solution be diluted with water, both salts break down according to the equations 2Nad+Cu a C] |PdCl 4 ]Na 2 <=> 2NaCl + PdCl 2 , and only when the palladium is in the ionic condition is it capable of entering into the reaction. The fact that the reduction of the palladous chloride is effected by means of CO at a concentration at which Cu 2 Cl 2 is incapable of causing any reduction is easy to understand, for the gas, CO, comes in contact more readily with a sufficient number of palladium ions than does the difficultly soluble cuprous chloride. Hydrogen, H. Mol. Wt. 2.016. Density =0.06960 * (Air = 1) . Weight of 1 liter =0.089978 gm. Molar volume =22.405 1. Critical temperature = -238 C. Hydrogen is practically insoluble in water. * Lord Rayleigh, Proc. Roy. Soc., 53, 1134 (1893). HYDROGEN. 771 ABSORPTION COEFFICIENTS OF HYDROGEN FOR WATER.* Temperature. /? Temperature. ............ 0.02148 30 ...... ....... 0.01699 5 ............ 0.02044 35 ............ 0.01666 10 ............ 0.01955 40 ............ 0.01644 15 ............ 0.01883 45 ......... 0.01624 20 ............ 0.01819 50 ............ 0.01608 25 ............ 0.01754 55 ............ 0.01604 The usual way for determining this gas by absorption is by means of metallic palladium, f but in the majority of cases it is determined by combustion with oxygen and observing the con- traction : H 2 + = H 2 2 vols. 1 vol. vol. It is evident that by the combustion of two volumes of hydro- gen, three volumes of gas will disappear (the water formed occupies a negligible volume). The contraction, therefore, is equal to f the volume of the hydrogen consumed. If the contraction is denoted by V c and the volume of the hydrogen by V H , then and consequently V H =Wc- In many cases the weight of the water formed is determined by absorbing the latter in weighed calcium chloride tubes, and from the gain in weight the volume of hydrogen is computed as follows : 18.02: 22,405 = p:x, 22 405 x = ^~Xp = 1243.6 Xp c.c. hydrogen under standard conditions. * L. W. Wihkler, Berichte, 24, 99 (1891). f The absorption can also be accomplished very satisfactorily by means of a one per cent, solution of palladous chloride. Campbell and Hart, Am. Chem. J., 18, 294. [Translator.] 772 GAS ANALYSIS. Combustion of Hydrogen, according to Cl. Winkler. The following method is employed frequently in technical analyses for the separation of hydrogen from methane. A mixture of hydrogen and air is conducted over gently-ignited palladium-asbestos, by which means the hydrogen is quantita- tively burned to water and the methane is not affected. Fig. Ill FIG. 111. represents the apparatus required. A is the eudiometer and is connected by means of the capillary E, in which is found a short fibre of palladium-asbestos, with a Hempel pipette filled with water. The capillary, E, is heated by means of the small flame F, at the place where the palladium-asbestos rests, to a temperature of about 300 to 400, but not hot enough to soften the glass. After the gas, which is mixed with air,* has been passed back and forth * If oxygen is used instead of air, some of the methane is sure to be oxidized. Cf. O. Brunck, Zeit. f. angew. Chem., 1903, p. 195. HYDROGEN. 773 through the capillary three times, the combustion is complete,, If the above-specified temperature is not exceeded, no trace of meth- ane will be burned and the hydrogen determination will be accu.- rate. It is, however, difficult to regulate this temperature closely enough to prevent the combustion of some methane unless, as recommended by Haber, the tube is heated by means of sulphur vapor; the results are usually from 0.5 to 1 per cent, too high. Preparation of Palladium-asbestos. Three gms. of sodium palladous chloride are dissolved in as little water as possible, 3 c.c. of a cold saturated solution of sodium formate are added and enough sodium carbonate solution to make the solution alkaline. Then about 1 gm. of soft, long-fibred asbestos is added, which sucks up the whole of the liquid, and the mixture is dried on the water bath; by this means finely- divided palladium is deposited uniformly through the asbestos: Na 2 PdCl 4 + HCOONa = SNaCl + HCi + C0 2 + Pd. The hydrochloric acid formed by the above reaction is neutral- ized by the sodium carbonate. In acid solutions formic acid hardly reduces palladous chloride at all. After the asbestos has thoroughly dried, the mass is softened with hot water, placed in a funnel and washed with hot water until the soluble salt is completely removed. It is then dried once more and preserved in a well-stoppered bottle. The palladium-asbestos fibre is introduced into the capillary tube as follows : The fibre is rolled between the fingers to a little round wad, the latter is placed in the opening of the unbent capil- lary tubing and by gentle tapping upon the table it is made to pass along to the centre of the tube. The latter is then bent as shown in the figure. Remark, inasmuch as the palladium-asbestos is likely to become shoved into the capillary, it is perhaps more satisfactory to use instead a palladium wire which is wound into a spiral.* * Private communication from Dr. Leutold of Hamburg. 774 GAS ANALYSIS. Methane, CH 4 . Mol. Wt. 16.03. Density =0.55297. (Air = 1.) Weight of 1 liter =0.71488 gms. Molar volume =22.43 1. Critical temperature = 82 C. Preparation. Methane is conveniently prepared by a process analogous to that used in making ethylene * (cf. p. 751). X mixture of equal parts methyl iodide and alcohol (sp. gr. 0.805) is allowed to act upon a zinc-copper couple which has been washed with alcohol. The zinc-copper couple is obtained by pouring a 2 per cent. copper sulphate solution four times over granulated zinc, then washing with water, and finally with alcohol. By allowing the mixture of methyl iodide and alcohol to drop upon the copper-coated zinc, a steady stream of methane is obtained at the ordinary temperature. The gas is purified by shaking it with fuming sulphuric acid, and then with caustic potash solution. It then contains nearly 99 per cent of CH 4 and alDOut 1 per cent, of nitrogen. Methane, also called marsh-gas or fire damp, is only slightly soluble in water. ABSORPTION COEFFICIENTS OF METHANE FOR WATER, f Temperature. Temperature. .......... 0.05563 30 ............ 0.02762 5 .......... 0.04805 35 ............ 0.02546 10 ........... 0.04177 40 ............ 0.02369 15 ............ .03690 45 ............ .02238 20 ............ 0.03308 50 ............ 0.02134 25 ............ 0.03006 55 ............ 0.02038 In alcohol, the gas is about ten times as soluble as it is in water. Inasmuch as no satisfactory absorbent for methane is known, it is always determined by combustion. From the equation representing the combustion, C0 2 +2H 2 0, 2 vols. + 4 vols. 2 vola. vol. we can make the following deductions: 1. Contraction. The contraction caused by the combustion of methane is equal to twice its original volume. 2. Carbon Dioxide. By the combustion of methane an equal volume of carbon dioxide is produced. 3. Oxygen Consumed. For the combustion of one volume of methane two volumes of oxygen are necessary. * Gladstone and Tribe, J. Chem. Soc., 45, 154. t L. W. Winkler, Berichte, 34, 1419 (1901). f ILLUMINATING AND PRODUCER GASES. 775 ANALYSIS OF ILLUMINATING AND PRODUCER GASES. The analysis of all such gases is best performed either by the method of Hempel * or that of Drehschmidt.t Hempel's Method. HempePs apparatus is shown in Fig. 105, p. 743. It consists of a eudiometer, W, divided into -J c.c. and connected by means of rubber tubing with the levelling-bulb K. The eudiometer is also connected with the compensation-tube D and the latter is connected with a manometer C; both the tubes W and D are sur- rounded by a cylinder containing water. Calibration of the Apparatus. First of all the manometer-tube is filled with mercury by raising the levelling-bulb K with the stop- cock p in the position shown in Fig. 105, so that there is an open connection between W and c ; the mercury is allowed to pass over into C until the mark mm is reached. The volume of the manom- eter-tube from the mark m to the point a (Fig. 105) is now deter- mined as follows: By carefully lowering the bulb K the mercury is drawn over into C exactly to the point a when the stop-cock p is closed. A little air is allowed to enter into the eudiometer through the right- hand capillary tube above p (the tube E should be withdrawn as in Fig. 112), the levelling-bulb K is placed upon a solid support at about the same height as the mercury in W, and with the stop- cock p still open the position of the mercury in W is read. The stop-cock is closed, K is raised a little and p is turned to the position shown in Fig. 105. By raising K still higher, the air is driven over into the manometer-tube C until the mercury has exactly reached the mark m, when the stop-cock A (Fig. 105) is closed. The exact position of the mercury is then adjusted by turning the stop-cock p one way or the other, and the position of the mercury in W is once more read. The difference between the * Gasanalytische Methoden (1900), p. 48 ff. t Berichte, 21, p. 3242 (18881. 77 6 GAS ANALYSIS. two readings represents the volume of the tube between the marks m and a, an amount which must be added to all subsequent readings. A drop of water is now introduced at c, by means of a fina FIG. 112. pipette, into the compensation-tube D and the end of the tube c is either fused together, or closed with a cork stopper and made air-tight with sealing wax. Procedure for the Analysis. If the analysis is to be carried out on the spot, a large sample of the gas is collected in a Dreh- schmidt pipette (Fig. 105 S). To accomplish this the capillary tube E' is connected by means of rubber tubing with the source ILLUMINATING AND PRODUCER GASES. 777 of the gas, and the stop-cock M is turned so that the tube E' is in connection with the bulb of the pipette, the levelling-bulb being in a low position and the stop-cock s left open. The pipette is entirely filled with the gas, then the stop-cock M is turned so that it communicates with the outer air, and the gas is com- pletely expelled from the pipette. The gas is in this way drawn in and out of the pipette at least three times in order to make sure that all foreign gas (air) is removed from the rubber tubing. The sample of gas is then taken and the two stop-cocks M and s are closed. In order to bring the gas to be tested from the Drehschmidt pipette into the eudiometer, the two instruments are connected by means of the capillary E f (imagine the capillary E in Fig. 105 to be replaced by E r ) and the rubber connections are firmly wired to the glass. The stop-cock M is turned to the posi- tion shown in Fig. 105, the levelling-bulb K is raised (after previously causing the mercury in the manometer- tube to reach to the point a) and the burette is entirely filled with mercury until the latter begins to flow from out of the tip of the key at M, when the cocks A and p are closed. The cock M is then turned so that the pipette S and the burette W are in con- nection, K r is raised, s opened, K lowered, and both p and A are opened. After about 40 c.c. of the gas have passed over into the eudiometer, the cocks A and M are closed, the key of the stop- cock M (which must be entirely filled with mercury) is dipped into a beaker containing mercury, and the gas in the capillary is sucked into W by lowering K and opening A and p. As soon as the capillary E is entirely filled with mercury, A, p and finally M are closed. The volume of the gas in W is now determined as follows: A is opened and K raised so that the mercury in the bulb is a little higher than it is in W. After this p is opened and the gas is driven over towards C until the mercury in both arms of the manometer-tube is at about the same height, when A is immediately closed. The last fine adjustment of the mercury levels within the tubes is made by closing or opening 778 GAS ANALYSIS the screw-cock Q;* the volume is now read, and to the reading the correction corresponding to the volume between the marks M and a is added. From this point begins the analysis. i. D3termination of Carbon Dioxide. With the stop-cock p closed, the cock M is turned as shown in Fig. 105 the Drehschmidt pipette is removed and replaced by a second, clean pipette completely filled with mercury. On connecting the stop-cock M with the rubber connector of the capillary E f , it should be in the position shown in the drawing. By this means the mercury in the rubber tubing can flow out through the key. After wiring the rubber tightly to the glass, from 3 to 5 c.c. of caustic potash solution (1:2) are introduced through the key into the pipette M and the alkali in the capillary is washed out with about 2 c.c. of distilled water and then with a little mercury; after this the gas itself is driven over into the pipette. When the mercury has filled the whole capillary, both to the right and left of M, then A, p, and M are closed. The bulb K' is raised so that extra pressure is placed upon the gas in the pipette and s is closed. The pipette is now gently shaken for three minutes with- out disconnecting it from the eudiometer, after which the gas is returned to W as follows: M, p, and A are opened, K is lowered, K f raised, and s opened. When almost all of the gas has been driven out of the pipette, M, p, A, and Q are closed, the levelling-bulb is placed on the table below, and K' is placed upon the support (missing from Fig. 105, but shown in Fig. 112) upon which the pi- pette itself rests. M, p, A, and s are now opened and Q screwed up a little so that the gas is very slowly sucked into the burette. As soon as the caustic potash solution has reached M the latter is closed. The gas remaining in the capillary to the left of M is now removed by sucking mercury through the key of M into W. Finally the volume of the unabsorbed gas is read in the * The reading is best made with the help of a small telescope, the ocular of which is provided with cross-hairs. For this purpose the telescope con- nected with a Bunsen spectroscope is suitable. ILLUMINATING AND PRODUCER GASES. 770 same way as before. The difference between the two readings represents the amount of CO 2 . 2. Determination of the Heavy Hydrocarbons. The pipette containing the caustic potash solution is removed and replaced by another containing fuming sulphuric acid.* The gas is driven over into the latter, shaken with the acid for three minutes,f and the pipette emptied in precisely the same way as before. The gas is now returned to the pipette containing the caustic potash in order to remove the acid vapors, and finally trans- ferred to the burette W and its volume read. The difference before and after the treatment with fuming sulphuric acid repre- sents the sum of the heavy hydrocarbons (C 2 H 4 , C 6 H 6 , C 2 H 2 , etc.). It is not usually customary to attempt to separate the Benzene from the ethylene. 3. Determination of Oxygen. This part of the analysis is carried out in exactly the same way as the determination of the CO 2 , except that in this case the absorp- tion pipette contains an alkaline solution of pyrogallol (cf. pp. 758-9). 4. Dstsnnination of Carbon Monoxide. The determination of carbon monoxide may be effected either by absorption with ammoniacal cuprous chloride or by simul- taneous combustion with hydrogen and methane. For the absorption method, the procedure is the same as in the case of the determination of the heavy hydrocarbons, i.e., the absorption is effected in a pipette containing only ammoniacal cuprous chloride (no mercury). The gas is shaken for three minutes with a solution of cuprous chloride which has already been used frequently, and then the same length of time with a fresh, * In this pipette the bulb-tube K' is fused on to the absorption-bulb, so that it is a little higher than the latter, in the same way as in the Hempel pipette (Fig. 115). Mercury is acted upon by fuming sulphuric acid. f From the experience of the Massachusetts Gas Inspectors it would seem as if more time were necsssary for the complete absorption of the heavy hydrocarbons perhaps thirty minutes instead of three. [Translator.] GAS ANALYSIS. or little used, solution (cf. pp. 763-4). B3fore reading the volume of the unabsorbed gas it must be freed from ammonia vapors, which is accomplished by shaking with hydrochloric acid (It 2) in a Drehschmidt pipette. 5. Determination of Hydrogen and Methane. After the removal of the carbdh monoxide, the gas may con- sist of hydrogen, methane, and nitrogen. An excess of oxygen is added to this mixture (with illuminating-gas twice its volume is added, while with Dowson, water, and producer gas only a little more than half as much oxygen is necessary). The eudiom- eter W is connected with a Drehschmidt pipette entirely filled with pure mercury * by means of a Drehschmidt platinum capil- lar} (Fig. 105, V), and the latter is heated to bright redness with the non-luminous flame of a Teclu burner, taking care that the inner flame mantle does not come in contact with the platinum. The gas mixture is conducted three times in a slow stream through the hot platinum tube, but taking care that no mercury enters the latter. The volume of the unconsumed gas is then measured without removing the platinum capillary, and the carbon dioxide is determined by introducing some caustic potash into the pipette and then shaking the gas with it; after three minutes' shaking, the unabsorbed gas is returned to the eudi- ometer, closing the stop-cock M as soon as the caustic potash solution reaches it. Calculation of Hydrogen and Methane. Assume V c.c. of gas to be taken for the analysis. The residue remaining after the absorption of the CO 2 , C n H 2n , O, and CO was mixed with oxygen and burned. The contraction pro- duced was V c and the CO a formed amounted to V K . We saw on p. 774 that the volume of the methane is equal * There must be no trace of caustic potash in the pipette, because in that case CO ? would be absorbed and an inaccurate result would be obtained. To make sure that all the alkali is removed, the pipette is washed first with water, then with hydrochloric acid, and finally with water once more. ILLUMINATING AND PRODUCER GASES. 781 to the volume of the CO 2 formed, V K , and in per cent. ; V:V K =WO:x z=-rrlOO = per cent. CH 4 . Since by the combustion of one volume of CH 4 two volumes of gas disappear, it is evident that by the combustion of V K c.c. of CH 4 the contraction will amount to 2V K . If the latter value be subtracted from the total contraction V c , the difference represents the contraction caused by the combus- tion of the hydrogen present (V c 2V K } and two- thirds of the lat- ter represents the amount of hydrogen, 2(V C -2V K ) T" =H ' and in per cent.: C - X X= - ^- - ' ^er cent. H. Determination of Carbon Monoxide, Methane, and Hydrogen by Combustion. After the absorption of the CO 2 , C n H 2n , and O, the residual gas consists of CO, CH 4 , H, and N. To it a measured volume of oxygen * is added, the mixture burned, and both the contraction, V C: and the carbon dioxide formed, V K , are estimated. After this the unused oxygen is determined by absorption with alkaline pyrogallol solution. If the excess of oxygen is subtracted from the amount originally added, the difference will give the amount of oxygen necessary for the combustion, V . * The purity of the oxygen must be tested before the analysis, because the commercial product almost always contains nitrogen. For the analysis a measured volume of nitrogen is added to a definite amount of oxygen, as otherwise the amount of the residual gas might be too small to fill the manometer-tube between the marks a and m (Fig. 105). The nitrogen is prepared by allowing air to stand over phosphorus in a Hempel pipette. (Cf. p. 759). 7 2 GAS ANALYSIS. If the amount of CO is denoted by x, the CH 4 by y, and finally the hydrogen by z, we have the following three independent equa* tions: 3. V = and from these equations we find that * According to A. Wohl (Berichte, 1904, 433) the results are not quite accurate when obtained in this way because the molecular volume does not always equal the theoretical value of 22.41 liters. Nernst, in his book on Theoretical Chemistry, gives the following molecular volumes: For 1 gm.-mol. of the gas, or referred to oxygen. H 2 = 22.431. H 2 = 1.0017 O 2 = 22.391. O 2 = 1.0000 CO = 22.391. CO = 1.0000 CH 4 = 22.441. CH 4 = 1.0020 CO 2 = 22.261. CO 2 =0.9939 Taking these values into consideration, A. Wohl obtains for x CO, y CH 4 , and z H 2 , the following formulas: x =0.3329 V c -V + l .3394 Ffc, y= - 0. 3336 V c + 1. 0020 F -0. 3340 Vk; 2 = 1.0005F c -1.0017Fo-0.0060FA;. F. Haber (Thermodynamik techn. Gasreaktionen, p. 289) sees no reason for modifying the Bunsen formulas in this way, for when a combustion analysis is carried out by explosion, the volume of gas after the explosion is so poor in carbon dioxide that the partial pressure of the latter does not vary much from that of an ideal gas, and, therefore, follows Avogadro's Rule. It is quite another matter in the case of mixtures rich in carbon dioxide, as often occur in gas- volumetric analyses. In that case the weight of carbon dioxide (or of carbonate) is computed from the volume of the gas and accurate values are obtained by using the observed molecular volume of 22.26 for this gas (see p. 386). The necessity of using the observed molecular volume instead of the theoretical value has been shown by Tread well and Christie (Z. angew. Chem., 1905, 1930) for chlorine. With other vapors (NH 3 , HC1, SO 2 , N 2 O) the observed molecular volume should be used unquestionably. ILLUMINATING AND PRODUCER GASES. 783 In order to illustrate the accuracy of the method, the results obtained in the analysis of the gas from a Dowson gas generator with the help of the Deville tube (Fig. 100, cf. p. 732) will be given. Two samples of the gas were taken, one 35 cm. and the other 45 cm. above the grate. The height of the coal layer in the producer amounted to 45 cm. DOWSON GAS. Sample I (35 cm. above the grate). I. II. CO 2 = 8.54 8.48 CH 2n = 0.30 0.30 O = 0.36 0.27 CO = 20.79 20.81 CH 4 - 1.32 1.26 H = 21.84 22.27 N = 46.85 46.61 100.00* 100.00 100.00 The above analysis was performed by Korbuly in the author's laboratory, and the carbon monoxide was determined by absorp- tion in ammoniacal cuprous chloride, but in the following analysis this gas was determined, as described above, by simul- taneous combustion with hydrogen and methane DOWSON GAS. Sample II (45 cm. above the grate). I. II. Mean. CO 2 = 8.58 8.55 8.56 C n H 2 n= 0.48 0.48 0.48 O = 0.17 0.26 0.21 CO = 20.79 20.59 20.69 CH, = 0.43 0.43 0.43 H = 19.31 19.22 19.26 N = 50.24 50.47 50.37 100.00 100.00 100.00 * These analyses add up to exactly 100 per cent, simply because the nitrogen is determined by difference. [Translator.] 784 GAS ANALYSIS. Obviously, the above results are perfectly satisfactory; it is worth mentioning, however, that according to the former method (absorption of the CO and combustion of the residue) the value obtained for the methane is almost invariably somewhat higher, .and that for hydrogen a trifle lower than according to the second method. To illustrate this, the results of a third analysis * will be given, which was also made by Korbuly in the sample of gas taken 35 cm. above the grate Sample I (Dowson Gas, 35 cm. above the grate). CO determined CO determined by absorption. by combus'.ion. CO, = 8.51f 8.43 C n H 2 n = 0.30 0.33 O * - 0.31 0.27 CO = 20.80 20.91 CH 4 = 1.29 0.79 H = 22.05 23.38 N = 46.74 45.89 100.00 100.00 Of the two methods, the author decidedly prefers the latter. Analysis according to H. Drehschmidt. J The apparatus of Drehschmidt; like that of Hempel, con- sists of the gas-burette B and the compensation-tube C, both of which are contained in a] cylinder filled with water (Fig. 113). Through the stop-cocks a and b, B and C are connected by means of capillary glass tubing in which a drop of a colored solution (indigo and sulphuric acid) is placed ; in order to deter- mine the position of the latter, the capillary is provided with * The gas came from the same tube as in the case of the other analyses. The gas was removed from the tube, as described on pp. 731-2. t This is the analysis given on p. 783. $ Berichte, 21 (1888), p. 3242. ILLUMINATING AND PRODUCER GASES. 785 a millimeter graduation. The three-way cock a can be turned so that C connects with the outer air or with the capillary, or so that the capillary is in connection with the air; it has an FIG. 113. opening through the top of the key. The cock b has a right- angled boring like H, Fig. 105. The burette is divided into milli- meters and must be calibrated with mercury before using. The apparatus is used in the same way as described under the Hempel method, p. 775. 786 GAS ANALYSIS. TECHNICAL GAS ANALYSIS. Method of Hempel. The apparatus necessary is depicted in Fig. 114. It consists of a long measuring-tube ending at the top in a thick-walled capillary tube and connected at. the bottom by means of rubber tubing about a meter long with the levelling tube. The gas is confined over water which has been saturated with the gas to be examined, and the absorption is effected in HempePs absorption pipettes such as are shown in Figs. 115, 116, 117, and 118.* Fig. 100 represents a simple pipette for liquid absorbents, while Fig. 101 shows a compound absorption pipette. The latter is used for solutions which undergo change on exposure to the air, e.g., an alkaline solution of pyrogallol, or an ammoniacal cuprous chloride solution. The liquid in the two right-hand bulbs serves to protect the solutions on the left. Fig. 117 shows the pipette used for fuming sulphuric acid. The small bulb is filled by the glass- blower with glass beads, which serve to give to the sulphuric acid the largest possible surface, so that the absorption is effected much more readily. Fig. 118 is a pipette used for solid absorbents, such as phosphorus, etc. In order to fill it with phosphorus, the pipette is placed upside down, the cylindrical part is filled with distilled water, and small sticks of colorless phosphorus are introduced. After filling the pipette, the rubber stopper is inserted, the apparatus is placed right side up, water is poured into the bulb, and any air-bubbles in the cylindrical part of the pipette are removed by blowing through the bulb until the water flows out from the top of the left-hand capillary, which is then closed by means of rubber tubing and a pinch-cock. Analysis of Illuminating-gas. First of all the confining liquid is prepared by conducting the gas through distilled water in a wash-bottle for several minutes with constant shaking. * These wooden pipette stands are no longer much used ; iron ones are preferred. [Translator.] ANALYSIS OF ILLUMINATING GAS. 787 The gas-burette is filled entirely full with this liquid and then the upper rubber tubing is closed with a pinch-cock. In order to fill the burette with gas, the receiver is connected with FIG. 114. FIG. 116. the burette by means of a piece of rubber tubing through which the gas has been flowing for two or three minutes, the levelling-tube is lowered, the pinch-cock opened and a little more than 100 c.c. of the gas are allowed to flow into the burette. The upper cock is now closed, the levelling-tube raised until the lower meniscus of the confining liquid is exactly at the 100-c.c. mark, when the rubber between the levelling-tube and the burette is closed near the burette with a pinch-cock. The apparatus is allowed to stand until the water no longer rises in the burette; this requires 788 GAS ANALYSIS. two or three minutes. When the water is stationary, the lower pinch-cock is carefully opened (for there is extra pressure in the burette), which causes the water-level to sink. When the 100-c.c. mark is again reached, this cock is closed, the upper pinch-cock is opened an instant in order to allow the excess of gas to escape and then immediately closed. Then, to make sure that the burette contains exactly 100 c.c. of the gas,the lower pinch-cock is opened and after bringing the water in the levelling-tube to the same height as in the burette, the reading is taken; the lowest point of the meniscus should coincide exactly with the 100-c.c. mark of the burette. Finally the lower pinch-cock is closed. i. Determination of Carbon Dioxide. The burette is connected with a pipette containing caustic potash solution by means of a capillary filled with water, as shown in Fig. 114, the levelling-tube is raised, first the lower pinch-cock and then the upper one * is opened and the gas is driven over into the pipette. The confining liquid should now fill the entire capil- lary. The upper pinch-cock is closed, the pipette taken up and shaken for three minutes, f and the gas is returned to the burette, taking care that none of the alkali enters with it. The liquid in the levelling-tube is brought to the same level as that in the burette; the lower pinch-cock is closed and after the water has completely drained from the sides of the tube, the volume of the unabsorbed gas is read. 2. Determination of the Heavy Hydrocarbons, CnH 2n . The burette is connected by means of a dry, empty capillary with sulphuric acid pipette (Fig. 117) and the gas is passed back and forth four times, taking care that no water enters the pipette and that the sulphuric acid does not reach the rubber connection. Before the experiment the position of the sulphuric acid is * In the figure this pinch-cock is lacking. t The absorption takes place more rapidly with one of Hempel's new pipettes, which is similar to the one shown in Fig. 117, except that the right- hand bulb is replaced by a movable levelling-bulb, as in Fig. 105. The latter is filled with mercury, upon which the liquid absorbent floats. For the absorption of CO 2 , it is only necessary to pass the gas back and forth once. DETERMINATION OF OXYGEN. 789 marked upon the milk-glass plate back of the pipette and at the end of the experiment the acid must come to the same mark. The gas in the burette is now contaminated with acid vapors which FIG. 117. FIG. 118. are removed by passing it into the potash pipette, afterwards returning it to the burette. 3. Determination of Oxygen. This can be effected by shaking the gas in the compound pipette with alkaline pyrogallol solution, but far preferably by means of phosphorus. In the latter case, the gas is driven over into the phosphorus pipette and allowed to remain there until the white vapors disappear; this usually requires but three or four minutes (cf. p. 759). If no white vapors can be detected, this shows con- clusively that the absorption of the heavy hydrocarbons was incomplete (cf. p. 759). In such a case, the gas must be again" treated with sulphuric acid and afterwards with phosphorus. If no white fumes are then formed, no oxygen is present, a case which practically never occurs, for in the determination of the hydrocarbons a little air containing oxygen always reaches the gas from the small capillary. 4. Determination of Carbon Monoxide. The gas is shaken three minutes with an old solution of ammoniacal cuprous chloride and then the same length of time with a fresh solution. (See pages 763, 779.) 790 GAS ANALYSIS. 5. Determination of Hydrogen and Methane. After the absorption of the carbon monoxide the residual gas is placed in the hydrochloric acid pipette, while the burette is washed out with hydrochloric acid in order to remove traces of alkali, and then filled with distilled water. About 15 to 16 c.c. of the gas in the hydrochloric acid pipette are transferred to the burette> ancT after reading its volume it is driven over into an explosion pipette containing mercury (Fig. 1 10) . 100 c.c. of air (containing 20.9 c.c. of oxygen) are accurately meas- ured off in the burette and added to the contents of the explosion pipette. The latter is then closed by means of a pinch-cock, the contents of the pipette are mixed by shaking, the levelling- tube is lowered so that the gas is placed under reduced pressure, and the glass stop-cock of the pipette is closed. The platinum wires which are fused in the upper part of the bulb are now con- nected with the poles of a small induction coil so that sparks pass between the platinum points within the pipette. The explosion at once occurs with a flash without ever breaking the pipette. The gas is returned to the burette. It would seem natural to read the volume of the gas and then determine the amount of carbon dioxide formed, the latter being a measure of the amount of methane burned. This is not advisable, however, because the gas in the burette is confined over water which absorbs ap- preciable quantities of carbon dioxide.* Consequently without reading the volume of the gas, it is transferred to the potash pipette, the carbon dioxide removed, and the volume of the gas then read ; this gives the contraction V c . Finally, the amount of unused oxygen is determined by means of absorption with phosphorus. If the excess of oxygen is subtracted from the total amount added (20.9 c.c.), the amount of oxygen required for the combustion is determined (V \ so that we have two equations from which the amount of hydrogen and methane can be computed. * Subsequent experiments have shown that the error caused by absorption of CO 2 by the water in so slight, during the short time of waiting, that it is better to determine the CO 2 with caustic potash after the explosion, as Hempel also recommended. The results thus obtained are usually more concordant than those by the method described in the text. (See p. 792.) A 'N 'A LYSIS OF ZURICH ILLUMINATING-GAS. 791 If we represent by x the volume of the hydrogen, and by y the volume of the methane we have 2. V = and from these equations we find The values thus obtained are referred to the total gas residue and in this way the amount of hydrogen and methane present in the illuminating gas is determined. Great accuracy is naturally not to be expected by such an analysis, but the procedure is very satisfactory for an approxi- mate estimation. In order to illustrate this point, the results of analyses made by two different students in the author's labora- tory at the same time will be given. Analysis of Zurich Illuminating-gas by Hempel's Technical Method. i. II. Gas taken ............... 100 c.c. 100 c.c. -*1.8% C0 2 -1.8% CO, After removal of CO 2 ...... 98 . 2 98 . 2 " " " CH2n ... 94.6 94.6 ->0.6% O -0.6% O " " " O ....... 94.0 94.0 ->8.6% CO -8.8% CO " " "CO ...... 85.4 85.2 For the H and CH 4 deter- mination were taken of gas ................... 16.0 15.6 +air .................... 116.0 115.6 Alter the explosion ........ 86 . 85 . 8 >5 . 2 excess oxygen 5 . 6 excess oxygen " removal of excess of O 80.8 80.2 V = 20. 9-5. 2= 15. 7. F =20.9-5.6=15.3. 792 GAS ANALYSIS If the values of Vc and Vo are inserted in the above equations, we have: Hydrogen =8.6 #=9.1 Methane y=5.7 y=5A and in per cent. : x=45.9% H 2/=30.42%CH 4 z=49.7% H 2/=29.5% CH 4 SUMMARY OP THE TWO ANALYSES. I. II. Difference. CO 2 1.8 1.8 0.0 C n H n 3.6 3.6 0.0 O 0.6 0.6 0.0 CO 8.6 8.8 0.2 H 45.9 49.7 3.8 CH 4 30.4 29.5 0.9 . N 9.1 6.0 3.1 From the results obtained, it is obvious that in each case the values obtained by absorption agree closely; on the other hand, the two determinations of hydrogen differ by almost 4 per cent. while that of methane shows a divergence of nearly 1 per cent. It is possible to obtain a much closer agreement than the above in the determination of hydrogen and methane, but the analysis is inaccurate on account of the fact that only one-fifth of the residual gas is taken for the explosion; thus every error is multiplied by five. As was mentioned in the foot-note on page 790, the gas residue may be analyzed as under (a) with the exception that the C(>2 obtained by combustion is measured. Then, if the volume of the hydrogen =x and that of the methane =y, the contraction = V c = f + 2y. from which the hydrogen (x) can be computed as follows: ANALYSIS OF ZURICH ILLUMINATING-GAS. 793 As an example of this kind of an analysis, two of the author's students analyzed independently a sample of illuminating gas from Montbeliard. Taken -CO 2 - CnH 2n -O 2 -CO Of the residual gas there was taken for H and CH 4 determination + I 100 c.c. 97.4c.c. = 2.6%CO 2 92 . 7 c.c. = 4 . 7% C n H 2 92.4 c.c. = 0.3% O 2 83.1 c.c. = 9.3% CO air After the explosion -CO 2 O 2 15. 4 c.c. 115. 4 c.c. 90 . 6 c.c. = 24 . 8 = V c 84.5c.c. = 6.1 = Vjfc=CH 4 II 100 c.c. 97.2 c.c. = 2. 8% CO 2 92 . 4 c.c. = 4 . 8% CnH 92.0 c.c. = 0.4% O 2 83.0 c.c. = 9.0% CO 2 15. 6 c.c. 115. 6 c.c. 90 . 2 c.c. = 25 . 4 V c 83.8 c.c. = 6.4 = CH 4 80.3 c.c. = 4. 2 = excess oxy- 79.8 c.c. =4.0 = excess gen of O 2 If the values of V c and Vk are inserted in the above equa- tions: Hydrogen Methane a; =8.4 = 45.3 per cent H2 i/ = 6.1 = 32.9 " CH 4 x = 8.4 44.7 per cent H 2 t/ = 6.4 34.0 " CH 4 SUMMARY OF THE TWO ANALYSES CO 2 = CO 2 = H 2 = CH 4 = H 2 = I. II. Difference. 2.6 2.8 0.2 4.7 4.8 0.1 0.3 0.4 0.1 9.3 9.0 0.3 45.3 44.7 0.6 32.9 34.0 1.1 4.9 4.3 0.6 100.0 100.0 794 GAS ANALYSIS Much better results are obtained by the (6) Method of W inkier-Dennis. In this method, the entire gas residue is transferred to a Hempel pipette containing mercury and connected with a leveling bulb (Fig. 119). Through the rubber stopper at the bottom two steel needles are inserted (knitting needles), the longer of which is enveloped throughout its whole length by a glass tube, and the upper end is connected, at about three-quarters FIG. 119. the height of the cylindrical part of the pipette, with a thin platinum spiral. The pipette is now connected with a Hempel burette containing 100 c.c. of oxygen * over water, a low pressure is produced in the * The oyxgen used for experiments in gas analysis should preferably be prepared in the laboratory by heating potassium chlorate in a small retort, which is prepared by blowing a bulb (of about 20 c.c. capacity) at the end of a narrow piece of glass tubing; after introducing about 5 gms. of potassium chlorate, the tubing is bent to a right angle close to the bulb. The end of the tube is connected with a short piece of rubber tubing and the bulb heated over a free flame. As soon as oxygen begins to come off freely (lighting a glowing splinter) the rubber tubing from the retort is connected with a Drehschmidt absorption pipette, which contains a little caustic potaeh Solution and is filled with mercury (cf. Fig. 105, p. 743). The oxygen is not ANALYSIS OF ZURICH ILLUMINATING-GAS. 795 oxygen burette by lowering the leveling tube and then closing the rubber tubing with a screw-cock, after which the leveling tube is placed in a high position. The bottom ends of the two needles of the pipette are now connected with the wires of a small storage battery of such a strength that the platinum spiral is heated to dull redness. By lowering the leveling bulb, a slightly- lower pressure is produced in the pipette, and by opening the FIG. 120. two upper screw-cocks between the pipette and the oxygen burette, and gradually opening the lower screw-cock on the burette, a very slow stream of oxygen is conducted into the pipette. Since a large excess of the gas residue is present at the start, the com- bustion takes place quietly; explosions never occur. During the introduced at once into the pipette, but is allowed to pass through the cock M into the air. After about a minute, one can assume that the air from the retort and rubber tubing has been entirely replaced by oyxgen. The leveling bulb K of the piptte is lowered, the cock s opened, and the cock M turned 90 so that the pipette fills with oxygen. When the filling is accom- plished, M is closed and the retort removed. By shaking the pipette, any carbon dioxide formed by the burning of dust, etc., is absorbed. 790 GAS ANALYSIS. combustion the platinum spiral begins to glow more brightly; to prevent its melting, a resistance * must be placed in the circuit by means of which the strength of current, and thus the glowing of the platinum, may be regulated as desired. As soon as all the oxygen is in the pipette, the spiral is allowed to glow two or three minutes longer, the electric current is then stopped, and the gas allowed to remain in the pipette for fifteen minutes so that it will assume the room temperature. It is then transferred to a Hempel burette and its volume measured; the carbon dioxide is determined in the usual manner. To illustrate the accuracy of this technical method, the follow- ing three analyses were 'carried out independently by three of the author's students. ANALYSIS OF ZURICH ILLUMINATING-GAS ON JULY 14, 1909. CO 2 2 CO CH 4 H 2 N 2 I II III 2.0% 2.2% 1.9% 4.4 4.4 4.6 0.7 0.5 0.6 9.2 9.2 9.3 27.6 28.2 27.9 49.8 49.4 49.3 6.3 6.1 6.4 100.0% 100.0% 100.0% Remark. By this method it is possible to burn pure acetylene without any explosion. The oxygen, however, must not be conducted, as above, into the acetylene because in that case the combustion of the acetylene will be incomplete and considerable carbon will deposit. If the oxygen is first placed in the combustion pipette, the platinum wire brought to glowing, and then the acetylene introduced, the combustion takes place nicely without deposition of any carbon. The Winkler-Dennis pipette is open to the objection that the rubber stopper eventually leaks; for this reason the author prefers the form of apparatus devised by his assistant, M. Bretsch- ger, as shown in Fig. 120. * The resistance mentioned on page 178 is suitable to use here. ORSATS APPARATUS. 797 Instead of burning the gas residue according to the Winkler- Dennis method, it may be conducted over glowing cupric oxide.* Orsat's Apparatus. For the analysis of flue gases, Orsat has constructed the appa- ratus shown in Fig. 121. It consists of the 100 c.c. measuring-tube a.- FIG. 121. B surrounded by a cylinder containing water, and connected on the one hand with three Orsat tubes by means of the cocks /, //, and ///, and the other hand with the outer air through the stop-cock h. The Orsat tube 777 contains caustic potash, 77 alka- line pyrogallol solution, and 7 ammoniacal cuprous chloride solution. * Jager, J. Gasbeleuchtung, 1898, 764. G. v. Knorre, Chem. Ztg., 1909, 717. 798 GAS ANALYSIS. Manipulation. By raising the leveling-bottle N and open- ing the stop-cock h, the measuring -tube B is filled with water. As soon as the water is above the mark in the widened part of the measuring-tube, the rubber tubing between the levelling-bottle and the measuring-tube is closed by means of a pinch-cock, a is connected with the source of the gas, and the gas is sucked into the measuring-tube by lowering the levelling-bottle and opening the pinch-cock. The U tube on tfce outside of the apparatus is filled with glass-wool and serves as a filter; any smoke being removed from the gas to be examined. The sample thus col- lected is naturally contaminated with the air from the rubber tubing, the U tube, and the capillary, which must be removed. The cock h serves for this purpose and is provided with a T boring. The cock is turned so that the burette communicates with the outer air through a small tube (not shown in the illustration) and the gas is expelled by raising the bottle N. This process of filling and emptying is repeated three times, and the fourth filling of the tube B is taken for the analysis. The gas in the burette is brought to the mark, and it is placed under atmospheric pres- sure by quickly opening and then closing h. After this the gas is driven over into the potash-bulb and back again to the meas- uring-tube several times, until there is no further absorption, after which the volume of the gas is again read. In the same way the gas is successively passed into the pyrogallol and the cuprous chloride tubes, thus obtaining the amount of C0 2 , O, and CO in the gas. Bunte's Apparatus. This apparatus, shown in Fig. 122, differs from those previously described, inasmuch as the absorption takes place in the measuring vessel itself, whereas in the other cases the absorption takes place in the pipettes. , The Bunte burette has a capacity of about 110 to 115 c.c. between a and 6; a is a three-way cock, while b is bored only once. Manipulation. The burette is connected with the levelling- bottle N, as shown in the illustration, a and b are opened, and the water is allowed to run up to the mark in the funnel above a. The key of the stop-cock a is connected with the source of the gas, N is lowered, a turned to the proper position, and BUNTE'S APPARATUS. 799 the gas is sucked into the burette. After about 101 to 103 c.c. of the gas have entered the burette, a and b are closed, N is raised, and by opening b the gas in the burette is compressed until the confining liquid has exactly reached the zero mark. The cock a is now cautiously opened, when some of the gas in the burette will escape through the water in the funnel. The gas in the burette is now under a pressure equal to that of the atmosphere plus the pressure from the column of water in the funnel, and all subsequent measurements are taken under the same conditions. Absorptions. In order to introduce the different absorbents into the burette, its lower end is connected by means of the rub- ber tubing h with the bottle F containing a little water, the water having been blown up into the rubber tubing. The cock 6 is opened, as is the screw-cock at h, and the water in the burette is allowed to run out until it exactly reaches the cock 6, which is then closed. The absorbent is placed in a small dish, the lower tip of the burette is introduced into the liquid, and the cock b is opened. Inas- much as the gas in the burette is under less than atmospheric pressure, the ab- sorbent is sucked up into the burette. The cock b -is now closed, the burette grasped above a and below 6 (in order not to warm the gas), and its contents well shaken, after which the burette is again dipped into the absorbent in the dish and a little more of the latter drawn up into the burette. This process is repeated until no more of the ab- sorbent is sucked up into the burette. It would now be incorrect to read the volume of the unabsorbed gas, for it is under quite a different pressure than at the beginning of the analysis; namely, the atmospheric pressure less the pressure of the column of FIG. 122. 8oo GAS ANALYSIS. liquid remaining in the burette with the cock b open. Further- more the vapor tension of the liquid in the burette is different from that of the water originally present. In order to obtain the original conditions, the burette is connected with the bottle F, which now only contains enough water to fill the rubber tubing and the glass tube, and the absorbent is sucked from the burette into the bottle until the upper level of the liquid reaches the cock &.* The* end of the burette is then dipped into a dish containing water, which rises into the burette on opening b. The latter is closed and water is allowed to run into the burette from the funnel until the original pressure is established, when the volume of the gas is once more read. The difference gives at once the per cent, of absorbed gas. By means of this excellent method the carbon dioxide can be removed by caustic potash, heavy hydrocarbons by bromine water, oxygen by alkaline pyrogallol solution, and carbon monoxide by cuprous chloride. ANALYSIS OF GASES WHICH ARE ABSORBED CONSIDERABLY BY WATER. Under this heading belong N 2 O, SO 2 , H 2 S, Cl, SiF 4 , HF, NH 3 , etc. Nitrous Oxide, N 2 O. Mol. Wt. 44.02. Density =1.5297 f (Air = 1) . Weight of 1 liter = 1.9766 gms. Molar volume =22.26 1. Critical temperature = +36 C. This gas is best prepared according to the method of Victor Meyer, J by allowing sodium nitrite to act upon a concentrated solution of a salt of hydroxylamine : NH 2 OH HC1 + NaNO 2 = NaCl + 2H 2 O + N 2 O. * The absorbent is now by no means exhausted, so that it is returned to the proper bottle, and can be used for several other determinations. t Lord Rayleigh, Proc. Roy. Soc., 74, 181 (1904). J Ann. Chem. Pharm., 157, 141. BUNTE'S APPARATUS. 80 1 It is best to proceed as follows: A concentrated, aqueous solution of sodium nitrite is added drop by drop from a separatory funnel, with constant cooling, to a concentrated solution of hydroxylamine hydrochloride, which v is contained in a small evolution flask; in this way the gasj evolved is pure and escapes in a regular stream. It is not advis- able to proceed in the opposite way, namely, to add the hydroxyl- amine solution to a concentrated nitrite solution, for in the latter case the decomposition is likely to take place with explosive vio- lence; it is still less advisable to add one of the reagents in the solid form. In a very dilute condition the solutions scarcely act upon one another. Nitrous oxide is never pure when it is prepared by heating ammonium nitrate; it is always contaminated with nitrogen and nitric oxide, but the latter may be removed by washing the gas with a solution of ferrous sulphate. According to L. Pollak the solubility of nitrous oxide between and 22 C. is expressed by the formula /?=!. 13719-0.042265- + 0.000610 - 2 , while according to Bunsen its solubilty is greater, being expressed by the formula /? = 1.3052 - 0.045362 t + 0.0006843 - 1\ The gas is absorbed to a much greater extent by alcohol than by water. According to Pollak, the absorption coefficient for alcohol is /? = 3.22804-0.04915-*+0.00023- 2 , wnile according to Bunsen it is somewhat greater: /?=4.17805-0.069816-*+0.000609- 2 . The determination of nitrous oxide can be effected with accu- racy by combustion, and this may be carried out in two different ways: &02 GAS ANALYSIS. 1. According to Bunsen, by exploding with hydrogen, or accord- ing to Knorre, by means of the Drehschmidt capillary. The con- traction produced is equal to the original volume of the nitrous oxide : N 2 + H 2 = H 2 O + N 2 . 2 vols. 2 vols. vol. 2 vols. 2. According to Pollak, by combustion with pure carbon monoxide, eithe^ by explosion or with the help of the Drehschmidt capillary; the volume of the CO 2 formed, which is measured, is equal to the volume of the nitrous oxide: N 2 O + CO = CO 2 + N 2 . 2 vols. 2 vols. 2 vols. 2 rols. There is no contraction in this case. Nitric Oxide, NO. Mol. Wt. 30.01. Density =1.0366,* (Air = l). Weight of 1 liter = 1.3402 gms. Molar volume =22.39 1. Critical temperature = 94 G. Preparation of Pure Nitric Oxide. The best way to prepare pure nitric oxide is the method of A. Deventer,f in which a solution of potassium ferrocyanide and potassium or sodium nitrite is acidified with acetic acid and shaken : + 2H 2 O+2NO. According to Emich % a very pure gas is obtained by shaking a nitrate with concentrated sulphuric acid in a nitrometer con- taining mercury. * Computed from observations of Gray (1905), Guye and Davila (1906). t Berichte,' 26, 589 (1893). J Monatshefte, 13, 73 (1892). NITRIC OXIDE. 803 ABSORPTION COEFFICIENTS OF NITRIC OXIDE FOR WATER. Tempetature. 0.07381 Temperature. 30 ft 0.04004 5 0.06461 35 0.03734 10 0.05709 40 0.03507 15 0.05147 45 0.03311 20 0.04706 50 0.03152 25 0.04323 55 0.03040 Although nitric oxide is only slightly soluble in water, its determination will be discussed at this place because this gas frequently occurs with nitrous oxide, and must therefore be determined at the same time. Nitric oxide may be determined by absorption with a con- centrated solution of ferrous sulphate or an acid solution of potas- sium permanganate, likewise, according to E. Divers, f by an alka- line solution of sodium sulphite (40 gms. Na2S0 3 + 4 gms. KOH in 200 c.c. H 2 O) with the formation of Na 2 N 2 O 2 SO 3 .:|: It is better, however, to carry out a combustion by the method of Knorre and Arndt ; in which the gas is mixed with hydrogen and very slowly passed through a Drehschmidt's platinum capillary heated to bright redness. Under these conditions the nitric oxide is quantitatively burned according to the equation. 2NO + 2H 2 = 2H 2 O + N 2 . 4 vols. 4 vols. vol. 2 vols. The contraction produced by the combustion of one volume of nitric oxide is equal, therefore, to f the original volume of the gas. Remark. If the gas mixture is passed too quickly through a platinum capillary heated to bright redness, or slowly through a less strongly heated platinum capillary, an appreciable amount of ammonia is formed and the results obtained are inaccurate. By explosion with hydrogen it is not possible to burn NO * L. W. Winkler, Berichte, 34, 1414 (1901). f Journ. Science Coll. Imp. University; Tokio, Vol. XI (1893), p. 11. J Nitric oxide is only partially absorbsd by an alkaline solution of pyro- gallol, where alkali nitrite, N 2 O, and N 2 are formed. (C. Oppenheim, Berichte, 36, 1744 (1903). Berichte, 21 (1889), p. 2136. 804 GAS ANALYSIS. when it is pure; when it is mixed with considerable nitrous oxide, violent explosions take place, yet the combustion of the NO is even then not quantitative. The gas may be determined, however, by combustion with carbon monoxide in the Drehschmidt capillary. According to Henry a mixture of carbon monoxide and nitric oxide is not explosive. On the other hand, according to Pollak, by conducting a mixture of these -gases through a Drehschmidt platinum capillary heated to bright redness, the combustion is quantitative if at the same time the carbon dioxide formed is removed by means of caustic potash; * otherwise the oxidation is not quantitative. According to the equation 4 vols. 4 vols. vol. 2 vols. the contraction produced is equal to f the volume of the nitric oxide. Remark. If considerable nitrous oxide is present at the same time, the combustion in the Drehschmidt capillary takes place quantitatively without the removal of the carbon dioxide. In this case the contraction is \ the volume of the nitric oxide. Analysis of a Mixture of Nitrous and Nitric Oxides. I. Combustion with Hydrogen. The gas is mixed with an excess of hydrogen and oxidized according to Knorre in the Drehschmidt platinum capillary heated to bright redness. If the volume of the N20 = x and that of the = 2/, we have: N 2 O. NO. 2. x-\-y=%V c (contraction) from which can be calculated: z = 3V-2V c y = 2(V c -V). * The mercury in the Drehschmidt tube is covered with caustic potash solution, by which means the CO 2 is absorbed immediately after its formation. DETERMINATION OF NITROUS OXIDE, ETC. 805 II. Combustion with Carbon Monoxide. The gas mixture is treated with an excess of carbon monoxide and burned in the red-hot platinum capillary; the contraction, V c , and the carbon monoxide, Vk, are both determined: N 2 O. NO. x + y = V k to = Vc from which we can compute: *-n-2F. Determination of Nitrous Oxide, Nitric Oxide, and Nitrogen in the Presence of One Another. I. By Combustion with Hydrogen in a Drehschmidt Capillary. After noting the contraction formed by the combustion with hydrogen, an excess of oxygen is added to the gas residue and the mixture is burned in the Drehschmidt capillary; two-thirds of the contraction which now takes place is equal to the amount of unused hydrogen in the first oxidation. If this quantity is deducted from the amount of hydrogen originally added, the difference, V w , represents the amount of hydrogen necessary. We have now: N 2 O. NO. N. 1. x + y + z = V 2. x + ly = V c 3. x + y = V w from which we can compute x=3V w -2V c 8o6 GAS ANALYSIS. II. By Combustion with Carbon Monoxide in ike Drehschmidt Capillary. We have: N 2 O. NO. N. x + y + z=V \y =Vc (contraction) x + y =7*(CO) 2 from which it follows : Determination of Nitrous Oxide, Nitric Oxide, and Nitrogen in the Presence of Carbon Dioxide. The accurate determination of nitrous oxide in the presence of carbon dioxide offers certain difficulties. It is not possible to determine the former by combustion with hydrogen in the Drehschmidt capillary, because when the carbon dioxide is present it takes part to some extent in the combustion, C0 2 +H 2 =H 2 + CO, and the previous absorption of the carbon dioxide by means of a large quantity of caustic potash is equally unsatisfactory, because a considerable amount of nitrous oxide will be absorbed by the reagent. The only way which can be recommended to effect this determination consists in absorbing the carbon dioxide by means of the smallest possible quantity of caustic potash, in which case the error introduced by the absorption of the nitrous oxide is reduced to a minimum ; the residual gas is examined as described above. Nitrogen, Mol. Wt. 28.02. Density = 0.9673 (Air=l). Weight of 1 liter =1.2505 gms. Molar volume 22.41 liters. Critical temperature = 149 C. Pure nitrogen is best prepared by heating a concentrated solution of potassium nitrate and ammonium chloride, present in amounts proportional to their molecular weights, and then NITROGEN. 807 conducting the escaping gas over glowing copper to reduce traces of nitric oxide. Nitrogen is but slightly soluble in water. According to L. Winkler ; * ABSORPTION COEFFICIENTS OF NITROGEN FOR WATER. Temperature. ft Temperature. ft 0.02348 30 0.01340 5 . 0.02081 35 0.01254 10 0.01857 40 0.01183 15 0.01682 45 0.01129 20 0.01542 50 0.01087 25 0.01432 55 0.01051 Nitrogen cannot be determined by any of the ordinary methods of gas analysis. It is always estimated by determining all the other constituents present in a mixture and subtracting the sum of the percentages found from 100. Technical preparations Oi nitrogen, prepared from the air, always consist of nitrogen and small amounts of rarer elements. According to Cavendish these latter may be obtained by adding oxygen and allowing a strong electric spark to pass through the mixture. In this way the nitrogen is completely oxidized to nitric acid, which can be removed by means of caustic potash solution. Then, by absorbing out the oxygen, the rarer gases are obtained. A still better process is that of Hempel, in which the nitrogen is absorbed by passing the gas over a glowing mixture of 1 gm. magnesium, 5 gm. freshly burnt lime, and 0.25 gms. sodium. The rare gases are not absorbed by this treatment. According to Bunsen, there is no combustion of nitrogen when detonating gas explodes in the presence of air, provided net more than 30 volumes of combustible gas are present for each 100 volumes of non-combustible gas. There is no oxidation of nitrogen during a combustion of a gas mixture which is passed through a Drehschmidt platinum capillary. * Barichte, 24, 3506 (1891). 8o8 GAS ANALYSIS. Analysis of Gases by Titration of the Absorbed Constituents. If a mixture of gases contains several constituents, of which two are removed by the same absorbent, and one of these can be determined by titration, it is a matter of no difficulty to determine the amount of each. The diminution in volume after treatment with the absorbent represents the amount of the two constituents, the titration value represents the amount of ons of them, and the difference shows the amount of the other. Such problems can be solved in a variety of ways, and only a fsw examples will be mentioned. Chlorine, Cl. Mol. Wt. = 70.92. Density =2.488 (Air = 1) .* Weight of 1 liter =3.2164 gins. Molar volume = 22.049 liters. Critical temperature = + 146 C. Determination of Carbon Dioxide in Electrolytic Chlorine.t The author has used the apparatus shown in Fig. 123 with the best success for this purpose. The absolutely dry eudiometer, B, . the capacity of which between the two stop-cocks is accurately known, and for con- venience may be 100 c.c., is rilled through the lower cock, after the gas has been dried by passing it through a long calcium chloride tube.J After five or ten minutes it is safe to assume that the air has been completely replaced by the gas. The lower three-way cock is now closed and then the upper one. The temperature and barometric pressure are both noted at this point. The tip of the burette is connected by rubber tubing with the reservoir N, the three-way cock is turned so that the reservoir communicates with the outer air, and then the lower * Leduc, Compt. rend., 116, 968 (1893) and Treadwell and Christie, Z angew. Chem., 47, 446 (1905). t Treadwell and Christie, Z. angew. Chem., 47, 1930 (1905). J If the burette and gas are not perfectly dry, some chlorine will be absorbed by the water. This will not affect the gas reading, but will be harmful in the subsequent titration. CHLORINE. 809 tip of the burette and the stop-cock are thoroughly washed, after which the latter is closed. A solution of potassium arsenite is prepared by dissolving 4.95 gms. As 2 0s in dilute potassium hydroxide, adding dilute sulphuric aoid until the solution is neu- tral to phenolphthalein and then diluting to 1 liter.* 100 c.c. of this solution are placed in N and any air in the rubber tubing is expelled by pinching it with the thumb and finger. By raising N and opening the stop-cock, a little of the arsenite solution is made to flow into the burette, which is inclined from side to side in such a way that the walls are thoroughly wet with the arsenite solution. The chlorine is slowly absorbed, as is evident from the fact that the solution slowly rises in the burette. As soon as there is no further absorption, the lower stop-cock is closed and the solution in the burette is made to flow back and forth in the burette several times, by inverting the burette and then turning it back again. After one or two minutes all of the chlorine will have been absorbed. Then in order to absorb all the carbon dioxide present, the tube N is lowered, 10 c.c. of potassium hydroxide solution (1:1) are poured in the funnel, and carefully made to flow into the burette. The stop-cock is again closed and the alkali solution poured back and forth in the burette. After the liquid in the burette and in the leveling tube has been brought to the same height, the reading is taken. The unabsorbed gas residue on being deducted from the original volume of the gas gives the volume of the chlorine plus that of the carbon dioxide. For the determination of the chlorine, the contents of the burette leveling tube, N, are emptied into a large Erlenmeyer flask and the stop-cock is turned to the position shown in the drawing, so FIG. 123. * An ordinary solution of arsenite prepared with sodium bicarbonate cannot be used here. 8io GAS ANALYSIS. that the liquid in the rubber tubing can flow out. The tubing is then removed from the burette and washed out with distilled water which is also allowed to run into N. The contents of the burette are added and the burette itself is rinsed with distilled water. The contents of the Erlenmeyer flask are treated with two drops of phenolphthalem solution and then with hydrochloric acid until the red color just disappears, 60 c.c. of sodium bicarbonate solu- tion are added (35 gms. in 1000 c.c. water), a little starch solution, N . and the excess of the arsenious acid is titrated with iodine solu- tion. It will be assumed that n c.c. are used in the titration. The ratio of the arsenite solution to the iodine is then established in the same way as in the above titration. 100 c.c. of arsenite solution are placed in an Erlenmeyer flask, 10 c.c. of caustic-potash solution (1:2) are added, two drops of phenolphthalem, hydro- chloric acid to decolorization, and then 60 c.c. of sodium bicar- bonate solution. The solution is then diluted to the same volume as that of the original experiment and titrated with tenth-normal iodine. Hereby n' c.c. are required. The difference n' n multiplied by 1.102 * gives the number of cubic centimeters of chlorine gas at C. and 760 mm. pressure. In other words, F'o = (n'-n)XL102. As, however, the original gas was measured at the temperature t C. and under the pressure B mm., it follows according to page 666, that F'--273 ' o~ 760 -(273+0' * This number is derived from the observation that the density of chlorine is 2.488 at 20, 70.92 = 0.001293X2.488 Therefore, 35.46 gms. of chlorine at and 760 mm. occupy a volume of 11,020 c.c. In the paper just cited, the value 22,039.2 c.c. is taken for the molecular volume of the chlorine and 0.001293 for the density of air. In the analyses of gases rich in chlorine, correct values are obtained by using the observed molecular volume of chlorine, which is 22.049 liters, and according to the experiments of N. Busvold in the author's laboratory, correct results are also obtained with this value in the analysis of gases containing little chlorine. CHLORINE. 811 from which can be computed F , _ FV760 (273 + Q .8-273 If V is the original volume of the gas used and R that of the residual gas in the burette, then Cl 2 +CO 2 -f Residue = V Residue = R C\ 2 +CO 2 =V-R -C\ 2 =Vi and in per cent. z-- u y ' = per cent. C02. Remark. In the first edition of this book the mixture of chlorine and carbon dioxide was absorbed by means of 5 per cent, caustic soda solution and the solution titrated with arsenious acid. This method is not entirely correct, however, because it is based upon the assumption that the chlorine is absorbed by the alkali in accordance with the equation, 2NaOH + C1 2 = Nad + NaCIO + H 2 O, whereas in reality there is always some chlorate formed and escapes the titration.* Offerhaus,| therefore, uses two burettes * The error here is practically constant and amounts to 0.77 per cent, chlorine. The work can be carried out according to the original method, adding, for the sake of accuracy, 0.7 per cent, to the value of chlorine found. Cf. O. Sterner, loc. tit., and Treadwell and Christie, loc cit. If the weight of chlorine fou^J by titration is used in connection with the theoretical molecular volume, a considerable error will be introduced for, in spite of the formation of chlorate, about 0.9 per cent, too much chlorine will be found in electrolytic chlorine. t Cl. Winkler (Industriegase II, 318) and Offerhaus (Z. angew. Chem., 1903, 1033), also Lunge-Berl, Chem. techn. Untersuchungsmethoden, 6th edition, Vol. I, p. 582, and O. Steiner, Z. Elektrochemie, 1904, 327. Si 2 GAS ANALYSIS. for the determination, absorbing the chlorine and carbon dioxide in one by means of caustic-potash solution, and absorbing the chlorine in another sample of gas by means of potassium iodide and the titrating with tenth-normal thiosulphate solution. This is, however, an unnecessary complication, for besides requiring an additional burette it involves the use of more of the expensive potassium iodide. It is possible, however, to carry out the analysis in one burette by absorbing first the chlorine with 10 per cent, potassium-iodide solution, and then intro- ducing caustic potash solution from the top of the burette. Hereby the carbon dioxide is absorbed and the iodine is trans- formed into iodide and iodate (the solution becomes nearly colorless) : C1 2 + 2KI=2KC1 + I 2 , 3I a + 6KOH=5KI + KI0 3 +3H 2 C. In order to determine the amount of iodine originally set free, the contents of the burette are allowed to flow into a potassium iodide solution which is acid with hydrochloric acid: <*KI + KIO 3 + 6HC1=6KC1+3H 2 0+3I 2 , and the liberated iodine is titrated with tenth-normal sodium thiosulphate solution. The method has no advantage over that described above and is not quite as accurate. Recently Schloetter has described another method for the examination of electrolytic chlorine gas. The chlorine is absorbed by means of hydrazine sulphate, whereby two volumes of chlorine set free one volume of nitrogen. The carbon dioxide is then absorbed by means of caustic soda solution. P. Ferchland* determines the chlorine by absorption with mercury in the residual gas after the CO 2 has been absorbed with caustic potash. This last method, according to the experiments of Busvold,t gives good results; it is to be recommended especially for the analysis of chlorine gas from the Deacon process. Examination of the Unabsorbed Gas Residue. Usually the residual gas is too small in amount (as in the above case) to * Z. Elektrochem., 13, 114. t Inaug. Dissert. Zurich, 1909, also P. Philosophoff. Chem. Ztg., 1907, 959. CHLORINE. 813 examine quantitatively, so that for this part of the analysis a larger sample of the gas is taken. The author has used the apparatus shown in Fig. 124 for this purpose with good results FIG. 124. The thick-walled filter-bottle A has a capacity of about 1.5 liters. It contains about 500 c.c. of strong caustic potash solution and the absorption-tube with stop-cock H is fastened air-tight within it. Manipulation. First of all, the absorption-tube is entirely filled with the caustic potash solution by suction through H, finally closing the latter.' The patent cock is then turned to the position //, and by suction through the left side-arm, the glass tube is filled with the alkali up to the cock. The latter is then turned to the position 7, the left side-arm is connected, by means of a * This apparatus has been used often by the author in the study of elec- trolytic chlorine gas and was described in the first edition of this text-book. Since then a similar apparatus has been recommended by Thiele and Deckert, Z. angew. Chem., 20, 437 (1907). 8 14 GAS ANALYSIS. short piece of rubber tubing and a long piece of glass tubing, with the source of the gas and several liters of gas are drawn through this tube by connecting the right side-arm with an aspirator. As soon as it is safe to assume that all of the air has been driven out from the tubing, the cock is turned to the position //, the aspirator is connected at a with the flask A in which a slight vacuum is produced, whereby the gas begins to collect in the absorption- tube. Chlorine and carbon dioxide are completely absorbed, while the residual gas collects in the upper part of the absorption- tube. The gas is allowed to enter the tube until from 50 to 70 c.c. of the gas residue are obtained; the cock / is then closed, the aspirator removed, and the gas driven over into a HempePs gas-burette and analyzed according to the methods already de- scribed. 60.9 c.c. of the gas residue from the above-mentioned elec- trolytically prepared chlorine * gave : Oxygen =40.71 f 2 = 66.9 Carbon monoxide = 2.6 \ and in per cent. | CO= 4.3 Nitrogen =17.6 J IN = 28.8 60.9 100.0 At the carbon electrode (the anode) not only chlorine but also a small amount of oxygen is liberated. The latter attacks the carbon of the electrode, forming carbon monoxide, the greater part of which in turn combines with the chlorine, forming phosgene gas, COC1 2 , but the latter is decomposed by water with the forma- tion of CO 2 and HC1: COC1 2 + H 2 O = C0 2 + 2HC1. This accounts for the presence of the CO 2 and CO in chlorine which has been prepared electrolytically. Hydrochloric Acid HC1. Mol. Wt. 36.47. . Density = 1.2686 (Air = l).f Weight of one liter = 1.6400 gms. Molar volume =22.24 liters. Critical temperature = +52 C. Hydrochloric acid is determined in gas mixtures by absorbing with standardized alkali. * Consisting of 99.0 per cent. C1 2 and 0.6 per cent. CO 2 . t Ledue, Compt. rend., 125, 571 (1897), found the density of hydrogen to be 1.2692 and Busvold, Inaug. Dessert. Zurich, 1910, obtained the value 1.2680. The above figure is the mean of these two determinations. SULPHUR DIOXIDE. 815 Sulphur Dioxide, Mol. Wt. 64.07. Density =2.2639 (Air = l).* Weight of one liter =2.9267 gms. Molar volume = 2 1.89 liters. Critical temperature =+ 155 C. For the determination of sulphur dioxide from pyrite burners, F. Reich recommends that the gas should be sucked by means N of an aspirator through a measured amount of iodine solution, colored blue with starch, until the latter is decolorized. The amount of the gas is equal to the quantity of water which has flowed from the aspirator + the volume of the absorbed SC>2. N For example, 10 c.c. of iodine solution were decolorized after V c.c. of water had flowed from the aspirator; the gas was at t C. and 760 mm. pressure. Since in the absorption of the SO2 by the iodine the following reaction takes place, SO 2 + H 2 + 1 2 = 2HI + SO 3 , it is evident that the amount of 862 absorbed, measured dry at N and 760 mm. pressure, will be 10.95 c.c., for 1 c.c. -^- iodine solution corresponds to 1.095 c.c. S02. It follows, then, that the volume of gas taken for the analysis equals ' 760- (273+0 and from this the per cent, of SO 2 in the gas can be calculated: FI; 10.95 1003 Y^ = per cent. S(>2. Other examples of gas analyses in which the absorbed con- stituent is estimated by titration are found in the determination * Leduc, Compt. rend., 117, 219 (1893). 8i6 GAS ANALYSIS. of the hydrogen sulphide in gas mixtures (see below) and in the determination of carbonic acid in the atmosphere by the method of Pettenkofer (cf. p. 593). Hydrogen Sulphide, H 2 S; Mol. Wt. =34.09. Density = 1.1895 (Air = 1) .* Weight of one liter = 1.5378 gms. .Molar volume = 22. 16 liters. (Critical temperature = 100 C. Determination of Hydrogen Sulphide in Gas Mixtures. Hydrogen sulphide, when present in the gases escaping from mineral springs, is estimated as follows: A large funnel, of from 2 to 3 liters capacity, is lowered into the spring and held in place by means of a wooden frame B weighted with stones, s (Fig. 125). The rubber tubing d is FIG. 125. removed from the flask a, the stop-cock h is opened, and the funnel T is filled, by means of suction, with water up to the stopper, and then h is closed. As soon as the water in the funnel has been replaced by the ascending bubbles of gas, the flask a is connected on the one side with the stop-cock tube h and on the other side directly with the aspirator A by means of a long rubber tubing. Suction is then started by opening H and is * Leduc. Compt. rend., 125, 571 (1897). HYDROGEN SULPHIDE. 817 continued with h open until the water in the funnel T again reaches the stopper, when h is closed. The funnel is allowed to fill with gas again, and this is eventually removed through a by means of suction. This operation is repeated twice more. In this way the neck of the funnel, the glass tube h } the rubber tubing d, and the flask a, all have the air originally present in them replaced by gas from the spring; a few drops of water are carried along mechanically into a. Ten c.c. of hundredth- normal iodine solution are then introduced into the ten-bulb tube b and 10 c.c. of hundredth-normal thiosulphate solution are placed in the tube c. The flask a is now quickly connected with b by means of a short piece of rubber tubing /, and c is con- nected with the aspirator A by means of a longer piece of tubing. Meanwhile the funnel T is again filling with gas. A measuring cylinder C is placed under the outlet tube of the bottle A, H is opened, and then the stop-cock h is cautiously turned. The gas is allowed to bubble slowly through 6 until the iodine solution becomes light yellow, but is not decolorized. H is now closed and after about two minutes h also. The contents of c are poured into b, starch is added, and the excess of the thiosulphate is titrated with hundredth-normal iodine solution (cf. page 650). The number of cubic centimeters, n, of iodine required for the titration, represents the amount of iodine which reacted origin- ally with the hydrogen sulphide. The position of the water in the graduate is noted (V c.c.), the temperature of the room t, and the barometer reading J5; w is the tension of aqueous vapor at the temperature t. In computing the amount of hydrogen sulphide in the gases escaping from the spring, it is to be remembered that the volume of gas which is taken for ike analysis is equal to the amount of water which has flowed into c plus the volume of hydrogen sulphide which has been absorbed by the iodine in b during the experiment. Inasmuch, however, as the amount of the latter is small in comparison with the total amount of gas taken for analysis, it may be neglected here. Furthermore, it is neces- sary to call attention to the fact that the volume of gas escaping from the spring is at a different temperature than that of the analysis; all the volumes, therefore, should be reduced to corre- 8i8 GAS ANALYSIS. spond to the temperature at the spring. The amount of hydrogen sulphide present per liter in the spring gases at the temperature t and the barometric pressure B is = c.c. H 2 S per liter.* Determination of Ethylene^ according to Haber. The principle of this method was discussed on p. 752. The determination is effected in the Bunte burette (cf. p. 799, Fig. 122). First, the contents of the lower portion of the burette from the lowest scale division to the cock is determined by weighing the water drawn from between these points, after allowing the burette to drain. Then about 90 c.c. of the gas to be examined are placed in the burette and the thermometer and barometer readings are taken. Then, exactly, as described on p. 799, the liquid is sucked down to the stop -cock, f a little bromine water is poured into a small evaporating-dish, about 10 c.c. of the liquid are allowed to rise into the burette, and in order to wash the bromine \vater from the tip into the burette, 2 or 3 c.c. of water are added. The walls of the burette are now thoroughly wet with the bromine water by suitably turning and inclining the tube, and in this way the ethylene is quickly absorbed. In order to de- termine the excess of bromine, a strong solution of potassium iodide is allowed to rise into the burette, and the contents of the latter are vigorously shaken. The liquid is then run out into an Erlenmeyer flask, the burette is carefully washed out N with water and the deposited iodine is titrated with -^ sodium thiosulphate solution. The titre of the bromine water added is next determined by pouring a little into a porcelain dish, * In this formula the temperature of the spring does not come into con- sideration because the gas is at the laboratory .temperature when measured. t After about one minute liquid will collect above the stop-cock, owing to the drainage of the liquid from the sides of the burette; this is removed before adding the bromine. DETERMINATION OF ETHYLENE. 819 pipetting off 10 c.c. of it, allowing this amount to run into a solution of potassium iodide and titrating the liberated iodine with N - sodium thiosulphate solution. The method of calculating the results will be illustrated best by means of a single example. Example. A gas consisting of 90 volumes of air and 10 volumes of ethylene was used for the analysis. Taken for analysis, 91.2 c.c. of the mixture. Temperature, 18.3 C. Barometer reading, 725 mm. Tension of aqueous vapor at 18.3 C. = 15.6 mm. mercury. Volume of the ungraduated portion of the burette. . . 6.10 c.c. Reading of the bromine water in the graduated part. . 10.00 " Bromine water used 16.10 c.c. Titre of the bromine water: N 10 c.c. of the bromine water correspond to 12.0 c.c. yp- sodium thiosulphate solution, so that 16.10 c.c. of bromine water are N equivalent to 19.32 c.c. of sodium thiosulphate. We have now: 16.1 c.c. bron 16.1 c.c. bromine water+ ethylene =12.23 " " 16.1 c.c. bromine water 19.32 c.c. -r~ solution. The ethylene corresponds to ..... 7.09 " " " Since the absorption of the ethylene by the bromine water takes place according to the equation it follows that N 2Br =21=20 .000 c.c. sodium thiosulphate solution = 22,270* N c.c. ethylene, and since 1 c.c. sodium thiosulphate corresponds *Cf. page 751. 820 GAS ANALYSIS. N to 1.100 c.c. C 2 H 4 , the 7.09 c.c. of solution used represent 7.09X1.100 = 7.94 c.c. C 2 H 4 at C. and 760 mm. pressure, or 9.10 c.c. C 2 H 4 at 18.3 C. and 725 mm., measured moist. The gas consists, therefore, of: C ' H <= 9 -!i and in per cent. JC 2 H 4 =10.0 percent. Air =82.1) (Air =90.0 " " 91.2 100.0 per cent. This method is especially suited for the determination of ethylene present with benzene fin illuminating-gas. In one sample the sum of the two gases is determined by absorption with fuming sulphuric acid or bromine water, and in a second sample the ethylene is determined as described above. This method is suitable for determining ethylene mixed with benzene vapors in illuminating gas. In one sample the sum of the two is obtained by absorbing with fuming sulphuric acid or bromine and in a second sample the ethylene is determined as above. Remark. Instead of using bromine water, which changes its strength so rapidly, the author uses a tenth-normal solution of potassium bromate; on acidifying an equivalent quantity of bromine is obtained. The experiment is carried out as follows : Exactly as described above, 90 c.c. of the gas are led into the Bunte burette, the water withdrawn till the lower cock is reached, and then some potas- sium bromate solution is placed in a small porcelain dish and about 10 c.c. of it is sucked up into the burette and the volume deter- mined. Then, after wiping off the lower capillary, an excess of concentrated potassium bromide solution and finally an excess of dilute hydrochloric acid is introduced. After shaking eight minutes, all the ethylene will be brominated. At the end of this time, 10 per cent, potassium iodide solution is allowed to enter the burette, the contents shaken, and emptied into an Erlenmeyer flask. The iodine thus liberated is titrated with tenth-normal sodium thiosulphate solution. The calculation is carried out as before. Using this modification of Haber's method, the author's DETERMINATION OF ETHYLENE. 821 assistant, M. Bretschger, analyzed a mixture containing air and known quantities of ethylene with the following results: Ethylene taken. Found. 1 5.20% 5.23% 2 5.20 5.16 3 5.20 5.15 4 3.90 3.88 5 3.70 3.70 6 3.70 3.68 Determination of Ethylene in the Presence of Acetylene. Since acetylene is not attacked in the cold by either bromine or iodine, whereas ethylene is readily attacked, the author has caused his assistant to carry out a number of experiments in thus analyzing gas mixtures. In one sample of the mixture, the sum of the ethylene-acetylene was determined by absorption with fuming sulphuric acid and in another sample the ethylene was determined by the above bromate-bromide method. The accuracy is attested by the following analyses of M. Bretschger: Ethylene taken. Ethylene found. Acetylene found. 4.38% 4.33% 0.61% 4.17 4.23 0.32 9.96 9.71 0.35 9.83 9.84 0.43 7.85 7.73 4.65 7.83 7.82 4.60 2.69 2.80 19.18 2.64 2.74 19.60 822 GAS ANALYSIS. i GAS-VOLUMETRIC METHODS. If in consequence of a chemical reaction a gas is evolved, from the volume of the latter the weight of the original substance may be determined. Examples of this sort of an analysis were given under the determination of CO 2 in carbonates (pp. 384, 388, 393), the carbon contents of iron and steel (pp.* 402 and 404), and the NO 3 in nitrates (p. 456). At this place a few more important determinations of the same nature will be described. Determination of Ammonia in Ammonium Salts. The following method, first proposed by Knop* and later modified by P. Wagner, f depends upon the fact that ammonia is oxidized by sodium hypobromite with evolution of nitrogen: 2NH 3 +3NaOBr=3H 2 O+3NaBr+N 2 . The nitrogen is collected in an azotometer and measured. If the amount of the ammonia be calculated from the volume of the nitrogen, too low results will be obtained, and this fact was formerly explained by t*he assumption that a part of the nitrogen was absorbed as such by the alkaline bromine solution. To-day, however, we know that such is not the case. At ordinary temperatures all of the ammonia is not oxidized according to the above equation to water and nitrogen, but a small amount of ammonium hypobromite is formed; for this reason too little nitrogen is obtained in the azotometer. If, on the other hand, the decomposition takes place at 100 C., the reaction goes quanti- tatively according to the equation. It is inconvenient to work at such a high temperature, so that it. is more practical to make a correction to the volume of nitrogen obtained at ordinary tem- peratures. * Chem. Centralbl., 1860, p. 243. t Zeitschr. f. anal. Chem., XIII (1874), p. 383; XV (1876), p. 250. DETERMINATION OF ETHYLENE. 823 Reagents and apparatus required : 1. An ammonium chloride solution, obtained by dissolving 8.3544 gms. of the pure sublimed salt in water and diluting to 500 c.c. 10 c.c. of this solution evolve at C. and 760 mm. pressure 35 c.c, of nitrogen if the reaction takes place according to the equation. 2. Sodium hypobromite solution. 100 gms. of sodium hy- droxide are dissolved in water, diluted to 1250 c.c. and after cooling by placing the flask in cold water, 25 c.c. of bromine are added, the contents of the flask vigorously shaken and again cooled. This solution must be preserved hi a stoppered bottle and protected from the action of light. 3. An azotometer. Instead of the azotometer of Wagner,* Lunge's Universal Apparatus (Fig. 61, 6, p. 387), or any such meas- uring instrument may be used. Procedure. Ten c.c. of the standard ammonium chloride solu- tion are placed in the small Wagner decomposition-bottle (Fig. 61, a, p. 387) while 40 to 50 c.c. of the hypobromite solution are poured into the glass L (which is fused to the bottom of the bottle H). The bottle is then connected with the measuring- tube A t "\ which is entirely filled with mercury, 6 opened and the le veiling-tube B lowered. The bottle H is inclined so that some of the hypobromite solution comes in contact with the solution of ammonium chloride and the two liquids are mixed by gentle shaking. A lively evolution of nitrogen at once takes place and the liquid becomes heated. As soon as the action ceases, more of the hypobromite solution is allowed to act upon the ammonium salt and the process is repeated until finally all of the hypobromite is in the outer part of H. As soon as no more gas is evolved by shaking, the decomposition-bottle is placed in water at the room temperature and after allowing it to stand ten minutes, the volume of the nitrogen is read under the conditions de- scribed on p. 389 The volume of nitrogen at and 760 mm. * Loc. cit. f The contents of the decomposition-bottle are previously cooled to the room temperature before the cock b is connected with it. 824 GAS ANALYSIS. thus found will be smaller than the theoretical value of 35 c.c., but it corresponds to the amount of ammonia contained in 10 c.c., of the ammonium chloride solution, i.e., 0.05320 gm. NH 3 . A number of such determinations are carried out and the mean of the results obtained is taken for the correct value. After this, some of the ammonium salt to be analyzed is weighed out, dissolved in water, and diluted so that 10 c.c. of the solution will contain approximately the same amount of ammonia as in the case of the standard solution. Then if, for example, from a gms. of an ammonium salt, V c.c. of nitrogen at C. and 760 mm. pressure were found, we have: V: 7=0.05320: x _ViX 0.05320 X y^ and in per cent.: FiX 5.320 =7 - =per cent. V-a Remark. The results obtained by this method agree exactly with those obtained by the distillation method described on p. 560. Only with substances containing sulphocyanates are the results obtained too high; in this case the sulphocyanate is decomposed by the alkaline hypobromite solution with evolution of nitrogen and carbon monoxide. f Consequently, the above method affords uncertain results in the analysis of the ammonia in gas liquors. Urea is decomposed by the alkaline hypobromite solution according to the equation: * Lunge (Lunge-Berl, Chem. techn. Untersuchungsmethoden, 6th edition, Vol. I, p. 155) does not standardize against the solution of ammonium chloride of known strength, but adds 2.2 per cent, more ammonia to correspond to the loss of nitrogen. Then FX0.001558 = gm. ammonia. f Donath and Pollak, Zeitschr. f. angew. Chem., 1897, p. 555. DETERMINATION OF NITROUS AND NITRIC ACIDS. 825 so that it can be determined in the same way as ammonium salts, the carbon dioxide produced by the decomposition being kept back by means of caustic soda solution. Determination of Nitrous and Nitric Acids. Principle. If a solution of a nitrite or nitrate be shaken with mercury and an excess of sulphuric acid, all of the nitrogen is set free as nitric oxide: 2HN0 2 +2Hg+ H 2 S0 4 =2H 2 0+ Hg 2 SO 4 +2NO, 2HN0 3 +6Hg+3H 2 S0 4 =4H 2 0+3Hg 2 S0 4 +2NO. From the volume of the nitric oxide, the weight of the nitrate or nitrite is computed. The analysis is best performed in a Lunge ni-trometer.t The latter is a Bunte burette, in which the lower stop-cock is lacking and the lower end of which is connected with a levelling-tube containing mercury. By raising the latter, the nitrometer (which need not be graduated) is entirely filled with mercury and the two-way cock under the funnel is closed. Then a weighed amount of the substance dissolved in a little water is placed in the funnel, the levelling-tube lowered, and the solution introduced into the nitrometer by carefully opening the cock, the funnel being finally washed out four times with two or three c.c. of concentrated sulphuric acid. The decomposition-tube is now taken out of the frame, it is placed several times in a nearly horizontal position and then quickly changed to a vertical position. By this means the mercury becomes intimately mixed with the acid and the decomposition at once begins. The shaking is continued one or two minutes until there seems to be no further increase in the volume of the liberated gas. The decomposition vessel is then * This reaction does not take place as completely as with ammonium salts. Lunge finds hi the determination of urea in urine that the nitrogen deficit is 9 per cent. If, therefore, the volume of nitrogen after being reduced to and 760 mm. is multiplied by 2.952, the correct urea value is obtained. t Berichte, 1890, p. 440, and Zeitschr. f. angew. Chem., 1890, p. 139. J See also Lunge-Berl, Chem. techn. Untersuchungsmethoden, 6th edition, Vol. I, p. 156. 826 GAS ANALYSIS. connected by means of a short piece of rubber tubing with the gas-burette filled with mercury, the nitric oxide is transferred to the latter, and its volume read after reducing it to the standard conditions by means of the gas-compensation tube. (Cf. p. 387, Fig. 61, 6.) If in an analysis a gms. of a nitrate were taken and F c.c. of NO were obtained, we have: 22,391 c.c. : 62 ^7 X 62.01 22391 and in per cent. : A9ni v* V* - = 0.2769 X= per cent. N0 3 . 22391 N a " a Remark. For the analysis of "nitrose," * the author knows of no method which affords such exact results. For the determination of nitrous acid in the presence of nitric acid by a gas-volumetric method, P. Gerlingerf treats the neutral solution of the two salts with a concentrated solution of ammonium chloride, whereby the following reaction takes place : NH 4 C1 +KN0 2 = 2H 2 O + KC1 + N 2 . Half of the nitrogen evolved, therefore, comes from the nitrous acid present. For the details of this determination, the original article should be consulted. Hydrogen Peroxide Methods. Hydrogen peroxide is oxidized by means of a number of different substances, giving up all of the oxygen; twice as much oxygen is evolved as is obtained from the oxidizing agent: XO + H 2 O 2 = X + H 2 O + O 2 . * Cf. Vol. I. fZ. angew. Chem., 1901, 1250. See also J. Gaihlot, J. Pharm. Chim., 1900, 6th Series, Vol. XII, p. 9. STANDARDIZATION OF PERMANGANATE SOLUTIONS. 827 Since, however, hydrogen peroxide itself slowly decomposes on standing (the decomposition being measurable on shaking and quite considerable by shaking in the presence of solid sub- tances (sand, etc.)), it follows that in the following methods no great excess of hydrogen peroxide should be used, and long-con- tinued shaking must be avoided. Standardization of Permanganate Solutions. The determination is best made according to Lunge in a gas volumeter (p. 387, Fig. 61). In order to obtain correct results; however, it is absolutely necessary that no excess of hydrogen peroxide be present. Consequently it is necessary to determine by means of a preliminary experiment the exact value of the permanganate solution in terms of the H 2 O 2 solution used (cf. p. 626). Then a measured amount of the latter is placed in the outside part of the Wagner decomposition-bottle (Fig. 61a, p. 387) , and 30 c.c. of dilute sulphuric acid (1:5) are added. After this, the exact amount of hydrogen peroxide required for the decom- position of the permanganate is introduced into the inner part or of the bottle and the latter is connected with the measuring-tube, which is filled with mercury, the cock b being removed for the time being, but it is replaced at the end of two or three minutes and turned to the position shown in the figure. The two liquids are then mixed, taking care to hold the decom- position-flask so that its contents will not be warmed by the heat of the hand, inclining it to an angle of about 90 with the vertical, and shaking for exactly one minute. While the oxygen is being evolved, care must be taken that the gas in the eudiometer is under reduced pressure. At the end of the decomposition, the gas is placed under atmospheric pressure, b is closed, and by means of the compensation-tube, the volume of the gas is reduced to what it would be at C. and 760 mm. pressure, as described on p. 388. One-half of the observed volume corresponds to the amount cf oxygen given up by the potassium permanganate. This number multiplied by 0.001429 gives the weight of the oxygen obtained from the permanganate. 828 GAS ANALYSIS. Remark. The amount of permanganate to be taken for the experiment is determined by the size of the measuring-tube. If N ' the latter has a capacity of 150 c.c., 15 c.c. of a solution o N or 40 to 50 c.c. of a -^ solution should be taken. The hydrogen peroxide used should not be too concentrated; it should be about a 2 per cent, solution. The active oxygen present in a sample of pyrolusite * may be determined by the same procedure. Determination of Cerium in Soluble Ceric Salts. If hydrogen peroxide is added to an acid solution of a soluble eerie salt, the latter is reduced with evolution of oxygen: 2CeO 2 + H 2 O 2 = Ce 2 O 3 + H 2 O + O 2 . The determination is effected in precisely the same way as was described above in the standardization of the permanganate solution. If half the volume of liberated oxygen is multiplied by 0.03077, the product represents the corresponding amount of CeO 2 .f Remark. If a large excess of hydrogen peroxide is avoided in the above analysis, satisfactory results will be obtained. For other determinations of this sort, consult " Lunge's Alkali Makers' Handbook" and "Hempel's Gas-Analytical Methods." Silicon Fluoride, SiF 4 . Mol. Wt. = 104.3. Density = 3.605 (Air = 1) . Weight of one liter = 4.660 gms. Molar volume = 22.40 liters. * Lunge's Alkali Makers' Handbook. t Assuming that the atomic weight of Ce= 140.25. DETERMINATION OF FLUORINE AS SILICON FLUORIDE. 829 Determination of Fluorine as Silicon Fluoride (Hempel and Oettel).* Principle. If a mixture of calcium fluoride and powdered quartz is treated with concentrated sulphuric acid in a glass vessel, all of the fluorine will be expelled as silicon fluoride : 2CaF 2 + SiO 2 + 2H 2 SO 4 = 2CaSO 4 + 2H 2 O + SiF 4 , and this gas can be collected and measured. One c.c. SiF 4 at and 760 mm. pressure corresponds to 0.006978 gms. CaF 2 , or 0.003395 gms. F 2 . Procedure. A weighed amount of the very finely-powdered substance, which must not contain any other acid that can be expelled by treatment with concentrated sulphuric acid,f is mixed with 3 gms. of ignited finely-powdered quartz and intro- duced into the dry decomposition flask K (Fig. 126). The latter is then evacuated somewhat by twice lowering the leveling-tube N with the stop-cock H open, closing the cock and expelling the air. At the beginning of the experiment, the burette H is not connected with the Orsat tube 0. By raising the ground- glass tube R, about 30 c.c. of concentrated sulphuric acid are allowed to flow into the flask. This acid must have been pre- viously heated in a porcelain crucible for some time at a tem- perature near the boiling-point, in order to detsroy every trace of organic matter, and allowed to cool in a desiccator over phosphorus pentoxide. The acid in K is heated to boiling with the stop-cock H open and the flask is frequently shaken. During the entire experiment, the mercury level in the tube N is kept a little lower than that of the mercury in the measuring-tube M . J At first the sulphuric acid foams considerably, but soon ceases, which is a sure sign that the decomposition is complete. The flame is then removed, the sulphuric acid allowed to cool and all the gas in K is expelled by introducing through V sulphuric * Gasanalytische Methoden, III, p. 342. f Cf. p. 479. J In order to keep the inside of N perfectly dry, 2 or 3 c.c. of concen- trated sulphuric acid are placed on top of the mercury. 830 GAS SNA LYSIS. acid, which has been previously heated and cooled as described above. As soon as the sulphuric acid reaches the stop-cock H, this is closed. After waiting ten minutes more, the gas is placed under atmospheric pressure, by suitably raising N, and the volume and temperature are noted. The gas is now driven over into the Orsat tube 0, containing caustic potash solution (1:2). The^silicon tetrafluoride is imme- diately absorbed. The residual gas is carried back to the tube M, and after waiting fifteen minutes the volume is read. The differ- FIG. 126. ence between the two readings gives the volume of silicon tetra- fluoride. Remarks. A. Koch tested this method in the author's labor- atory and obtained results varying from 98.97 to 102.63 per cent, with pure calcium fluoride. To obtain this accuracy, however, it is necessary to carry out the decomposition under approx- imately atmospheric pressure. When working under a vacuum the results were always too low. Thus in one case only 85.70 per cent, of the theoretical value was obtained. During these experiments with reduced pressure, a white sublimate forms at the lower part of the condenser, which on coming in contact with the sulphuric acid that is introduced at DETERMINATION OF VAPOR IN GAS MIXTURES. 831 the last, causes a strong effervescence. Since, however, all the gas in the burette was replaced, we believed that the low results could be traced to the absorption of silicon fluoride by the sul- phuric acid. This idea proved to be false, for a measured volume of silicon fluoride does not change when allowed to stand for twenty-four hours over concentrated sulphuric acid. The error, therefore, must be caused by the strange deposit that has con- densed in the lower part of the condenser. If the work is carried out under atmospheric pressure as described above, the white deposit is never obtained. The method can be used for the estimation of fluorine in the presence of carbonates. In this case the silicon fluoride is absorbed by means of a little water and the carbon dioxide by means of caustic potash. As, however, a little of the carbon dioxide is dissolved by the water, the gas residue which has been freed from carbon dioxide is shaken with this water again, whereby this dissolved carbon dioxide is removed and can be absorbed by a further treatment with caustic potash solution. For further details, consult the original paper by Hempel and Schefner.* Determination of Vapor in Gas Mixtures. It is often desired to estimate the weight, and from this the volume, of vapor present in a mixture of gases and vapors. This will be illustrated by one or two examples. Let it be assumed that in a unit of volume of a given gas mixture there are v' parts of gas and v" of vapor. If the whole mixture is under the pressure P, then the partial pressures of the gas and vapor will be respectively v'P and v"P r . The following equation then holds: P=v'P + v"P. The total pressure is therefore equal to the sum of the partial pressures. If now the partial pressure of either constituent is known, the volume of this constituent can be found by dividing by the total pressure. If, for example, v"P=w, then v^P_ = w = P P Applications. 1. Reduction of Volumes of Moist Gases to a Dry Condition at C. and 760 mm. Pressure. * Z. anorg. Chem., 20, 1 (1897). 832 GAS ANALYSIS. A gas saturated with water vapor occupies a volume v t at t C. and P mm. pressure. A unit of volume of the gas consists of v' volumes of dry gas and v" volumes of water-vapor. Now the tension of aqueous vapor at t = w, a value which can be obtained from the tables on p. 842. This value, w, represents in fact the partial pressure of the water-vapor. Consequently Pw is the partial pressure of the dry constituents and the volume of the latter, v' , is, as explained above: v' = at t and P mm. pressure. This volume reduced to and 760 mm. pressure is (P-w)P P-w v'o P-7QO-(l+at) 760(1+ at)' If the original volume of the gas and vapor is not 1, but V t , then, V = P ~ w V 760(1 +at) Similarly, the volume of the water-vapor at and 760 mm. pressure is 7" = Vt * 760(1 +at) 2. Calculation of the Moisture in the Air at Normal Pressure (760 mm.) and Temperature t. What is the volume of water-vapor contained in 100 c.c. of moist air at 0, 25, and 35 C.? According to the tsible: 100 = 4.6 mm.; 1025 = 23.5 mm.; and 1035 = 41.8 mm. Hence the volume of water-vapor present in a unit of volume is WQ 4.6 1025 23.5 7,035 41.8 = 760 ; 760 = 760 ; 760 = 760' * In this formula is it assumed that the water- vapor also follows Boyle't law, which is not strictly true. DETERMINATION OF YAPOR IN GAS MIXTURES. 833 and the percentage of moisture is 0.61 per cent at 0; 3.09 " " 25; 5.50 " " 35. If the gas is not saturated with moisture, the relation of the dry gas to the moist one is not so easy to determine, unless the degree of saturation is known. The humidity of a gas, or the degree of saturation, expresses the amount of moisture present as compared with the total amount which the gas can take up when perfectly saturated with moisture. Thus if the humidity is 50 per cent, the gas could take up as much again moisture at the prevailing temperature. If the degree of humidity of a gas is expressed by r, then the volume of the given gas at this temperature and 760 mm. pressure, is V(P-r-w) ~760 ' and the volume of the water-vapor present is V-r-w 760 ' 3. Calculation of the Weight of Water-vapor in a Given Volume of Air which is Saturated with Moisture at the Temperature 1? and the Pressure P mm. One c.c. of water-vapor weighs 0.000801 gm. at C. and 760 mm. pressure. One c.c. of water-vapor at t and P mm. pressure occupies a volume of l.P ^~ 760(1 +at)' and weighs If now w the vapor pressure of water at t and P mm, 834 GAS ANALYSIS. pressure, then the volume of water-vapor present in the volume V of the gas is, under these conditions, w-V t P ' and the weight of the water-vapor amounts to 760(1 +at) P 780(1 -Kofi If the weight of the water-vapor is g, then the volume of the moist air is y. 760(1 ' 0.000801 -w' If the gas is not saturated with water vapor, but the degree of saturation (the humidity) is known, then the following formula gives the weight of water-vapor present in the volume v t of the gas, 760(1 Or, if the weight of vapor present in a given gas volume is known, then from the last equation the humidity, r, of gas may be computed: _ g 760(1 +Q ~~ 0.000801- wv't 4. Calculation of the Weight of One Liter of Air at and 760 mm. when Free from Moisture and Carbon Dioxide. 1 c.c. of pure, dry air weighs 0.0012928 gm. at and 760 mm. 1 cc. of water-vapor weighs 0.000801 gm. at and 760 mm. 1 cc. of water-vapor is therefore 0.62 times as heavy as 1 c.c. of air 1 c.c. of C0 2 weighs 0.001977 gm. at and 760 mm. * Similarly, the volume of a gas saturated with any other vapor can be computed if the weight of vapor present and its density are known. DETERMINATION OF VAPOR IN GAS MIXTURES. 835 Air contains, on an average, 0.03 per cent of CO 2 . One liter of dry air, containing the average amount of CO 2 consists of 999.7 c.c.air and weighs 999.7X0.001293 = 1.2924 gms. 0.3 " CO 2 0.3X0.001977 = 0.0006 " 1000.0 " 1.2930 " =a The corresponding volume of dry air and of water-vapor at and 760 mm. pressure is 760- w w 760 + 760 and weighs 760 -w w ~760~ or, in other words, 1 liter of moist air weighs at and 760 mm. pressure, 0.38wA -76CTJ gmS - 5. Calculation of the Weights of One Liter of Moist Air Con- taining Carbon Dioxide at t and P mm. Pressure. If the tension of aqueous vapor = wt, then the volume of the moist gas is ~ H p ! = 1 at t and P mm. pressure, and at and 760 mm. w t 760 w t 760(1 +at) 760(1 +*)' and, as shown under 4, this weight, P-w w a(P-O.ZSw) 760(1 +oO 760(1+0 760(l+aO or gms. 836 GAS ANALYSIS. 27o -}-t P mm. pressure. liters of moist air would therefore weigh , 0.46445 (P- 0.38w) 9= 27* + t ~ V * APPENDIX. 8 3 8 APPENDIX. SPECIFIC GRAVITY OF STRONG ACIDS AT IN VACUO. (According to LUNGE, ISLER, NAEF, and MARCHLEWSKY.)* Specific Gravity at 15 Per Cent, by Weight. S >ecific Gravity Per Cent, by Weight. lt > (Vacuo). HC1. HN0 3 H 2 S0 4 . at 40 (Vacuo). HNO 3 . H 2 S0 4 . 1.000 0.16 0.10 O.iO 1.235 37.51 31.70 1.005 1.15 1.00 0.95 1.240 38.27 32.28 1.010 2.14 1.90 1.57 1.245 39.03 32.86 1.015 3.12 2.80 2.30 1.250 39.80 33.43 1.020 4.13 3.70 3.03 1.255 40.56 34.00 1.025 5.15 4.60 3.76 1 . 260 41.32 34.57 1.030 6.15 5.50 4.49 1.265 42.08 35.14 1.035 7.15 6.38 5.23 1.270 42.85 35.71 1.040 8.16 7.26 5.96 1.275 43.62 36.29 1.045 9.16 8.13 6.67 1 . 280 44.39 36.87 1.050 10.17 8.99 7.37 1.285 45.16 37.45 1.055 11.18 9.84 8.07 .290 45.93 38.03 1.060 12.19 10 67 8.77 .295 46.70 38.61 1.065 13.19 11.50 9.47 .300 47.47 39.19 1.070 14.17 12.32 10.19 .305 48.24 39.77 1.075 15.16 13.14 10.90 .310 49.05 40.35 1.080 16.15 13.94 11.60 .315 49.88 40.93 1.085 17.13 14.73 12.30 .320 50.69 41.50 1.090 18.11 15.52 12.99 1.325 51.51 42.08 1.095 19.06 16.31 13.67 1.330 52.34 42.66 1.100 20.01 17.10 14.35 1.335 53.17 43.20 1.105 20.97 17.88 15.03 .340 54.04 43.74 1.110 21.92 18.66 15.71 .345 54.90 44.28 1.115 22.86 19.44 16.36 .350 55.76 44 . 82 1.120 23.82 20 . 22 17.01 .355 56.63 45.35 1.125 24.78 20.99 17.66 .360 57.54 45.88 1.130 25.75 21.76 18.31 1.365 58.45 46.41 1.135 26.70 22.53 18.96 1.370 59.36 46.94 1.140 27.66 23.30 19.61 1.375 60.27 47.47 1.145 28.61 24.07 20.26 1.380 61.24 48.00 1.150 29.57 24.83 20.91 1.385 62.21 48.53 1.155 30.55 25.59 21.55 1.390 63.20 49.06 1.160 31.52 26.35 22.19 .395 64.22 49.59 1.165 32.49 27.11 22.83 .400 65. 2/ 50.11 1.170 33.46 27.87 23.47 .405 66.37 50.63 .175 34.42 28.62 24.12 .410 67.47 51.15 .180 35.39 29.37 24.76 .415 68.60 51.66 .185 36.31 30.12 25.40 1.420 69.77 52.15 .190 37 . 23 30.87 26.04 1.425 70.95 52.63 .195 38.16 31.60 26.68 1.430 72.14 53.11 .200 39.11 32.34 27.32 1.435 73.35 53.59 .205 33.07 27.95 1.440 74.64 54.07 .210 33.80 28.58 1.445 75.94 54.55 .215 34.53 29.21 1.450 77.24 55.03 .220 35.26 29.84 1.455 78.56 55.50 .225 36.01 30.48 1.460 79.94 55.97 1.230 36.76 31.11 1.465 81.38 56.43 * Lunge-Berl, Chem. techn. Untersuchungsmethoden, 6th edition, Vol. I, Tables, 61, 39, and 49. SPECIFIC GRAVITY OF STRONG ACIDS. 839 SPECIFIC GRAVITY OF STRONG ACIDS AT ~ (According to ISLER, NAEF, an IN VACUO. ConL Specific Gravity atf Per Cent. 1 Dy Weight. Specific Gravity atf Per Cent. bv Weight. Specific i Gravity 15 at ~7^ Per Cent, by Weight. (Vacuo). HNO 3 . H 2 S0 4 . (Vacuo). H,S0 4 . 4 (Vacuo). H 2 S0 4 . .470 82.86 56.90 1.610 69.56 .750 81.56 .475 84.41 57 . 37 1.615 70.00 .755 82.00 .480 86.01 57.83 .620 70.42 .760 82.44 .485 87.66 58.28 .625 70.85 .765 83.01 .490 89:86 58.74. .630 71.27 .770 83.51 1.495 91.56 59.22 .635 71.70 .775 84.02 1.500 94.04 5*9.70 1.640 72.12 .780 84.50 1.505 96.34 60.18 1.645 72.55 .785 85.10 1.510 98.05 60.65 1.650 72.96 .790 85.70 1.515 99.02 61.12 1.655 73.40 .795 86 . 30 1.520 99.62 61.59 1.660 73.81 .800 86.92 1.525 62.06 1.665 74.24 .805 87.60 1.530 62.53 1.670 74.66 .810 88.30 1.535 63.00 1.675 75.08 .815 89.16 1.540 63.43 .680 75.50 .820 90.05 .545 63.85 .685 75.94 .825 91.00 .550 64.26 .690 76.38 .830 92.10 .555 64.67 .695 76.76 .835 93.56 .560 65.20 .700 77.17 .840 .95.60 .565 65.65 .705 77.60 .8405 95.95 .570 66.09 .710 78.04 .8410 96.38 .575 66.53 .715 78.48 .8415 97.35 .580 66.95 .720 78.92 .8410 98.20 .585 67.40 .725 79.36 .8405 98.52 1.590 67.83 .730 79.80 .8400 98.72 1.595 68.26 .735 80.24 1.8395 98.77 1.600 68 . 70 .740 80.68 1.8390 99.12 1.605 69.13 .745 81.12 1 . 8385 99.31 840 APPENDIX. SPECIFIC GRAVITY OF POTASSIUM AND SODIUM HYDROXIDE SOLUTIONS AT 15 C. Specific Gravity. Per Cent. KOH. Per Cent. NaOH. Specific Gravi.y. Per Cent. KOH. Per Cent. NaOH. 1.007 0.9 0.59 1 . 252 27.0 22.50 1.014 1.7 1.20 1.263 23.2 23.50 1.022 2.6 1.65 1.274 28.9 24.48 1.029 3.5 2.50 9 1.2 5 29.8 25.50 1.037 4.5 3.22 .297 30.7 26.58 1.045 5.6 3.79 .304 31.8 27.65 1.052 6.4 4.50 .320 32.7 28.83 1.060 7.4 5.20 .332 33.7 30.00 1.067 8.2- 5.86 .345 34.9 31.20 1.075 9.2 6.58 .357 35.9 32.50 1.083 10.1 7.30 .370 36.9 33.73 1.091 10.9 8.07 .3-3 37.8 35.00 1.100 12.0 8.78 .397 38.9 36.36 1 . 108 12.9 9.50 .410 39.9 37.65 1.116 13.8 10.30 .424 40.9 39.06 .125 14.8 11.06 .438 42.1 40.47 .134 15.7 11.90 .453 43.4 42.02 .142 16.5 12 . 69 .463 44.6 43.58 .152 17.6 I'.oO .483 45.8 45.16 .162 18.6 14.35 .498 47.1 47.73 .171 19.5 15.15 .514 48.3 48.41 .ISO 20.5 16.00 .530 49.4 50.10 .190 21.4 16.91 .546 50.6 - .200 22.4 17.81 .563 51.9 .210 23.3 18.71 1.580 53.2 1.220 24.2 19.65 1.597 54.5 1.231 25.1 20.69 1.615 55.9 1.241 26.1 21.55 1.634 57.5 ~ SPECIFIC GRAVITY OF AMMONIA SOLUTIONS. 841 SPECIFIC GRAVITY OF AMMONIA SOLUTIONS AT 15 C. (According to LUNGE and WIERNIK.) * Specific Gravity. Per Cent. NH 3 . Specific Gravity. Per Cent. NH 3 . 1.000 0.00 0.940 15.63 0.998 0.45 0.938 16.22 0.996 0.91 0.936 16.82 0.994 1.37 0.934 17.42 0.992 1.84 0.932 18.03 0.990 2.31 0.930 18.64 0.988 2.80 0.928 19.25 0.986 3.30 0.926 19.87 0.984 3.80 0.924 20.49 0.982 4.30 0.922 21.12 0.980 4.80 0.920 21.75 0.978 5.30 0.918 22.39 0.976 5.80 0.916 23.03 0.974 6.30 0.914 23.68 0.972 6.80 0.912 24.33 0.970 7.31 0.910 24.99 0.968 7.82 0.908 25.65 0.966 8.33 0.906 26.31 0.964 8.84 0.904 26.98 0.962 9.35 0.902 27.65 0.960 9.91 0.900 28.33 0.958 10.47 0.898 29.01 0.956 11.03 0.896 29.69 0.954 11.60 0.894 30.37 0.952 12.17 0.892 31.05 0.950 12.74 0.890 31.75 0.948 13.31 0.888 32.50 0.946 13.88 0.886 33.25 0.944 14.46 0.884 34.10 0.942 15.04 0.882 34.95 * Lunge-Berl, Chem. techn. Untersuchungsmetho den, 6th edition, p. 531. 842 APPENDIX. TENSION OF WATER VAPOR ACCORDING TO REGNAULT. Degrees, Tension in Millimeters. Degrees, Tension in Millimeters. Degrees, C. Tension in Millimeters. -2.0 3.95 + 2.0 5.302 + 6.0 6.998 1.9 3.985 2.1 5.340 6.1 7.047 1.8 4.016 2.2 5.378 6.2 7.095 1.7 4.047 2.3 5.416 6.3 7.144 1.6 4.173 2.4 5.454 6 4 7.193 1.5 4.109 2.5 5.491 6.5 7.242 1.4 4.140 2.6 5.530 6.6 7.292 1.3 4.171 2.7 5.569 6.7 7.342 1.2 4.203 2.8 5.608 6.8 7.392 1.1 4.235 2.9 5.647 6.9 7.442 1.0 4.267 3.0 5. 687 7.0 7.492 0.9 4.299 3.1 5.727 7.1 7.544 0.8 4.331 3.2 5.767 7.2 7.595 0.7 4.364 3.3 5.807 7.3 7.647 0.6 4.397 3 4 5.848 7.4 7.699 0.5 4.430 3.5 5.889 7.5 7.751 0.4 4.463 3.6 5.930 7.6 7.804 0.3 4.497 3.7 5.972 7.7 7.857 0.2 4.531 3.8 6.014 7.8 7.910 0.1 4.565 3.9 6.055 7.9 7.964 0.0 4.600 4.0 6.097 8.0 8.017 + 0.1 4.633 4.1 6.140 8.1 8.072 0.2 4.667 4.2 6.183 8.2 8.126 0.3 4.700 4.3 6.226 8.3 8.181 0.4 4.733 4.4 6.270 8.4 8.236 0.5 4.767 4.5 6.313 8.5 8.291 0.6 4.801 4.6 6.357 8.6 8.347 0.7 4.836 4.7 6.401 8.7 8.404 0.8 4.871 4.8 6.445 8.8 8.461 0.9 4.905 4.9 6.490 8.9 8.517 1.0 4.940 5.0 6.534 9.0 8.574 1.1 4.975 5.1 6.580 9.1 8.632 1.2 5.011 5.2 6.625 9.2 8.690 1.3 5.047 5.3 6.671 9.3 8.748 1.4 5.082 5.4 6.717 9.4 8.807 1.5 5.118 5.5 6.763 9.5 8.865 1.6 5.155 5.6 6.810 9.6 8.925 1.7 5.191 5.7 6.857 9.7 8.985 1.8 5.228 5.8 6.904 9.8 9.045 1.9 5.265 5.9 6.951 9.9 9.105 TENSION OF WATER VAPOR. TENSION OF WATER VAPOR. Continued. 843 Degrees, C. Tension in Millimeters. Degrees, C. Tension in Millimeters, Degrees Tension in Millimeters + 10.0 9.165 + 14.0 11.903 + 18.0 15.357 10.1 9.227 14.1 11.986 18.1 15 454 10.2 9.2S8 14.2 12.064 18.2 15.552 10.3 9.350 14.3 12.142 18.3 15.650 10.4 9.412 14.4 12.220 18.4 15.747 10.5 9.474 14.5 12.298 18.5 15.845 10.6 9.537 14.6 12.378 18.6 15.945 10.7 9.601 14.7 12.45S 18.7 16.045 10.8 9.665 14.8 12.538 18.8 16.145 10.9 9.728 14.9 12.619 18.9 16.246 11.0 9.792 15.0 12.699 19.0 16.346 11.1 9.857 15.1 12.781 19.1 16.449 11.2 9.923 15.2 12.864 19.2 16.552 11.3 9.989 15.3 12.947 19.3 16.655 11.4 10.054 15.4 13.029 19.4 16.758 11.5 10.120 15.5 13.112 19.5 16.861 11.6 10.187 15.6 13.197 19.6 16.967 11.7 10.255 15.7 13.281 19.7 17.073 11.8 10.322 15.8 13.366 19.8 17.179 11.9 10.389 15.9 13.451 19.9 17.285 12.0 10.457 16.0 13 . 536 20.0 17.391 12.1 10.526 16.1 13 . 623 20.1 17.500 12.2 10.596 16.2 13.710 20.2 17.608 12.3 10.665 16.3 13.797 20.3 17.717 12.4 10.734 16.4 13.885 20.4 17.826 12.5 10.804 16.5 13.972 20.5 17 935 12.6 10.875 16.6 14.062 20.6 18.047 12.7 10.947 16.7 14.151 20.7 18.159 12.8 11.019 16.8 14.241 20.8 18.271 12.9 11.090 16.9 14.331 20.9 18.383 13.0 11.162 17.0 14.421 21.0 18.495 13.1 11.235 17.1 14.513 21.1 18.610 13.2 11.309 17.2 14.605 21.2 18.724 13.3 11.383 17.3 14.697 21.3 18.839 13.4 11.456 17.4 14.790 21.4 18.954 13.5 11.530 17.5 14.882 21.5 19.069 13.6 11.605 17.6 14.977 21.6 19.187 13.7 11.681 17.7 15.072 21.7 19.305 13.8 11.757 17.8 15.167 21.8 19.423 13.9 11.832 17.9 15.262 21.9 19.541 8 4 4 APPENDIX. TENSION OF WATER VAPOR. Confirm**. Degrees. C. Tension in Millimeters. Degrees, C. Tension in Millimeters. Degrees, Tension in Millimeters. + 22.0 19.659 + 26.0 24.988 + 30.0 31.548 22.1 19.780 26.1 25.138 30.1 31.729 22.2 19.901 26.2 25.288 30.2 31.911 22.3 20.022 26.3 25.438 30.3 32.094 22.4 20.143 26.4 25.588 30.4 32.278 22.5 20.265 26.5 25.738 30.5 32.463 22.6 20.389 26.6 25.cS91 30.6 32.650 22.7 20.514 26.7 26.045 30.7 32.837 22.8 20.639 26.8 26.198 30.8 33.026 22.9 20.763 26.9 26.351 30.9 33.215 23.0 20.888 27.0 26.505 31.0 33.405 23.1 21.016 27.1 26.663 31.1 33.596 23.2 21 . 144 27.2 26.820 31.2 33.787 23.3 21.272 27.3 26.978 31.3 33.980 23.4 21.400 27.4 27.136 31.4 34.174 23.5 21.528 27.5 27.294 31.5 34.368 23.6 21.659 27.6 27.455 31.6 34.564 23.7 21.790 27.7 27.617 31.7 34.761 23.8 21.921 27.8 27.778 31.8 34.959 23.9 22.053 27.9 27.939 31.9 35.159 24.0 22.184 28.0 28.101 32.0 35.359 24.1 22.319 28.1 28.267 32.1 35.559 24.2 22.453 28.2 28.433 32.2 35.760 24.3 22.588 28.3 28.599 32.3 35.962 24.4 22.723 28.4 28.765 32.4 36.165 24.5 22.858 28.5 28.931 32.5 36.370 24.6 22.996 28.6 29.101 32.6 36.576 24.7 23.135 28.7 29.271 32.7 36.783 24.8 23.273 28.8 29.441 32.8 36.991 24.9 23.411 28.9 29.612 32.9 37.200 25.0 23.550 29.0 29 . 782 33.0 37.410 25.1 23.692 29.1 29.956 33.1 37.621 25.2 23.834 29.2 30.131 33.2 37.832 25.3 23.976 29.3 30.305 33.3 38.045 25.4 24.119 29.4 30.479 33.4 38.258 25.5 24.261 29.5 30.654 33.5 38.473 25.6 24.406 29.6 30.833 33.6 38.689 25.7 24.552 29.7 31.011 33.7 38.906 25.8 24.697 29.8 31.190 33.8 39.124 25.9 24.842 29.9 31.369 33.9 39.344 HEAT OF COMBUSTION OF GAS. TENSION OF WATER VAPOR. Continued. 845 Degrees, C. Tension, Millimeters. Degrees, Tension , Millimeters. Degrees, Tension , Millimeters. + 34.0 39.565 + 34.4 40.455 + 34.8 41.364 34.1 39.786 34.5 40.680 34.9 41.595 34.2 40.007 34.6 40.907 34.3 40.230 34.7 41.135 35.0 41.827 HEATS OF COMBUSTION OF 1 LITER OF GAS MEASURED AT AND 760 MM. BAROMETRIC PRESSURE. Gas. Weight of One Liter. Referred to Gaseous Water Calories. Liquid Water Calories. Carbon monoxide Hydrogen 1.25016 0.09004 0.71488 1.25899 1.93660 3.48428 1.18080 3,034 2,595 8,505 14,018 21,226 (?) 33,750 13,582 about 900 3,386 1,400 5,000 3,034 3,077 9,469 14,989 22,720 (?) 35,198 14,073 about 1,000 3,700 5,500 Methane . . Ethylene Propylene Benzene-gas Acetylene Generator-gas Water-gas Dowson gas Illuminating-gas The values in the above table are based upon Thomson's measurements and only in the case of benzene is the theoretical density used.* * Julius Thomsen, Thermochem. Untersuchungen (1882), Vol. II, pp. 56, 85, 107, and Vol. IV, p. 254. CO + O = CO 2 + 67,960 cals. H 2 + O = H 2 O +68,357 cals. CH 4 +4O =2H 2 0+ CO 2 + 211,930cals. C 2 H 4 + 6O - 2H 2 O + 2CO 2 + 333,350 cals. C 3 H 6 + 9O = 3H 2 O + 3CO 2 + 492,740 cals. C 6 H 6 + 15O = 3H 2 O + 6CO 2 + 787,950 cals. C 2 H 2 + 5O = H,O + 2CO 2 + 310,450 cals. * TABLES FOR CALCULATING ANALYSES. DIRECTIONS FOR USING THE TABLES. Assume that it is desired to find the per cent, of arsenic present in an ore. For this purpose a gms. of the ore are taken for the analysis and the arsenic is determined as Mg 2 As 2 7 ; the ignited precipitate weighed p gms. The calculation of the arsenic is made as follows: Mg 2 As 2 O 7 :As 2 =p:s As 2 ~Mg 2 As 2 7 XP and in per cent. S A/r A r> Mg 2 As 2 O 7 x _ 10Q.As 2 p Mg 2 As 2 O 7 a The last equation is solved for x, after the proper values for p, a, etc., have been inserted, with the aid of logarithms. From 100 As the tables, the logarithm of M 7^ * s taken, to it the logarithm of p is added, and from the sum the logarithm of a is deducted. The number corresponding to this difference (found in the table of antilogarithms) represents the desired result in per cent. Example. 0.5 gm. of mispickel gave 0.4761 gm. Mg,As 2 Oi 7 . What is the per cent, of arsenic in the ore? In the following table under the heading Sought we find As, and under Found we look for Mg 2 As 2 O 7 , the form in which * The tables in this book can be purchased separately in flexible cloth binding. Price, thirty-five cents. 847 848 TABLES FOR CALCULATION OF THE ANALYSIS. the arsenic was determined, and then from the column headed Log we take the logarithm of this factor multiplied by 100; log. factorX 100= 1.6837 flog. 0.4761 (p)=9.6777-10 1.3614 - log. 0.5 (q) =9.6990 -10 1.6624 Number = 45. 96 per cent, arsenic. For most analytical computations the accompanying four- place logarithms are sufficiently accurate; where greater accuracy is desired tables of five- or even seven-place logarithms may be used. INTERNATIONAL ATOMIC WEIGHTS. 849 International Atomic Weights, 1915. Symbol. Atomic Weight. Symbol. Atomic Weight. Aluminium Al 27.1 Molybdenum Mo 96 Antimony Argon Sb A 120.2 39.88 Neodymium Neon Nd Ne 144.3 20 2 Arsenic As 74 96 Nickel Ni 58 68 Barium Ba 137 37 Nitrogen N 14 01 Bismuth Bi 208 Osmium. Os 190 9 Boron B 11.0 Oxygen. . o 16 00 Bromine Br 79 92 Palladium Pd 106 7 Cadmium Cd 112 40 Phosphorus. . . p 31 04 Caesium Cs 132 81 Platinum Pt 195 2 Calcium Ca 40.07 Potassium K 39 10 Carbon. c 12 00 Praseodymium. . Pr 140 6 Ceriurr Ce 140 25 Radium Ra 226 4 Chlorine Cl 35.46 Rhodium Rh 102 9 Chromium Cr 52 Rubidium Rb 85 45 Cobalt Co 58 97 Ruthenium . Ru 101 7 Columbium Cb 93 5 Samarium Sa 150 4 Copper. . , Cu 63 57 Scandium Sc 44 1 Dysprosium . Dy 162 5 Selenium Se 79 2 "Erbium Er 167 7 Silicon Si 9 8 3 Europium Eu 152.0 Silver Ae 107 88 Fluorine F 19 Sodium Na 93 00 Gadolinium Gd 157.3 Strontium Sr 87 63 Gallium Ga 69.9 Sulphur s 32 07 Germanium Ge 72 5 Tantalum Ta 181 5 Glucinum * Gl 9 1 Tellurium . . . Te 127 5 Gold Au 197 2 Terbium Tb 159 2 Helium He 3 99 Thallium Tl 204 Hydrogen H 1.008 Thorium Th 232 4 Indium In 114 8 Thulium Tm 168 5 Iodine .... I 126 92 Tin. Sn 119 Iridium Ir 193 1 Titanium Ti 48 1 Iron Fe 55.84 Tungsten W 184 Krypton . . . Kr 82 92 Uranium u 238 5 Lanthanum. La 139 Vanadium v 51 Lead Pb 207 10 Xenon Xe 130 2 Lithium Li 6 94 Ytterbium Lutecium Magnesium Lu Me 174.0 24 32 (Neoytterbium) . Yttrium. Yb Yt 172.0 89 Manganese Mn 54.93 Zinc '. . . Zn 65 37 j Mercury He 200.6 Zirconium Zr 90 6 * Also called Beryllium, Be. 850 TABLE OF CHEMICAL FACTORS. Sought. Found. Factor. Log.* Sought Found. Factor. Log. Ag AgCl AgBr Agl 0.75262 0.57444 0.45946 1.87658 1.75925 1.66224 BaO BaSO 4 BaCrO 4 BaSiF 6 0.65700 0.60532 0.54840 1.81757 1.78199 1.73909 Ag 2 AgCl 0.80843 1.90764 Bi BiA BiA BiAsO 4 Bi 0.89654 0.59942 1.1154 1.95258 1.77773 2.04743 Al A1A A1PO 4 0.53033 0.22195 1.72455 1 . 34625 Br Ag AgBr AgCl 0.74082 0.42556 0.55755 1.86971 1.62896 1.74629 A1 2 3 A1PO 4 0.41851 1.62171 As As 2 S 3 As 2 S 6 Mg 2 As 2 O 7 MgaPaOr 0.60911 0.48319 0.48269 0.67313 1.78470 1.68412 1.68371 1.82810 C CO 3 COo CC 2 . 27273 1.3636 1 . 43573 2.13470 Ca CaO CaCO 8 CaSO 4 CaF. 0.71464 0.40045 0.29440 0.51326 1.85409 1.60252 1.46893 1.71034 AssO, As 3 S 3 AfeS, Mg 2 As 2 O 7 Mg 2 P 2 C 7 0.80405 0.63783 0.63724 0.88865 1.90528 1.80471 1.80430 1.94873 CaO CaCO, CaSO 4 CaF 2 0.56031 0.41195 0.71820 1.74843 1.61485 1.85625 AsC 3 As 2 S 3 As 2 S 6 Mg 2 As 2 O 7 Mg 2 P 2 7 0.99915 0.79260 0.79186 1 . 1042 1.99963 1.89905 1.89805 2.04304 Cd CdS CdO CdS0 4 0.77801 0.87539 0.53919 1.89099 1.94220 1.73174 As 2 O 6 As 2 S 3 As 2 S 5 Mg 2 As 2 O 7 Mg 2 P 2 7 0.93414 0.74103 0.74034 1.0323 1.97041 1.86984 1.86943 2.01382 CdO CdS Cd CdSO 4 0.88877 1 . 14235 0.61591 1.94879 2.05780 1 . 78952 2.10901 2.05121 1.84073 AsO 4 As 2 S 3 AszSs Mg 2 As 2 O 7 Mg 2 P 2 7 1.1292 0.89574 0.89490 1.2477 2.05276 1.95218 1.95178 2.09616 CdS Cd CdO CdSO 4 1.2853 1 . 1252 . 69300 B BO 2 BO 3 BA B 2 3 B 2 3 BA BA 0.31428 1.2286 1.6857 1.1143 1.49732 2.08940 2.22678 2.04700 Cl AgCl Ag 0.24738 0.32870 1.39337 1.51680 C1H AgCl Ag 0.25442 0.33804 1.40555 1.52898 Ba BaSO 4 BaCrO 4 BaSiF 6 0.58846 0.54217 0.49119 1 . 76973 1.73414 1.69125 C10 3 AgCl KC1 NaCl 0.58225 1.1194 1.4276 1.76511 2 . 04897 2.15462 * In this column the logarithm of the factor multiplied by 100 is given. The logarithms are given to five decimal places, but it should be borne in mind that the fourth decimal place is in most cases doubtful. Four-place logarithms are accurate enough for nearly all chemical analyses. The atomic weights for 1915 are used in these tables. TABLE OF CHEMICAL FACTORS. 851 Sought. Found. Factor. Log. Sought. Found. Factor. Log. 1 . 78349 C1O 3 K AgCl KC1 0.85503 1.6438 1.93198 2.21584 SiF 6 CaF 2 0.60742 Fe FeO Fe 2 3 Fe 2 3 0.69940 0.89980 1.8447. 1.954io ClO 3 Na AgCl NaCl 0.74271 1.8211 1.87082 2.26033 H H 2 0.11190 1.04883 C10 4 AgCl KC1 NaCl 0.69388 1.3339 1.7013 1.84128 2.12514 2.23079 Hg Hg 2 Cl 2 0.84979 0.86216 1.92931 1.93559 C10 4 K AgCl KC1 0.96665 1.8584 1.98527 2.26913 I Agl PdI 2 AgCl 0.54055 0.70408 0.88545 1 . 73283 1.84761 1.94716 C10 4 Na AgCl NaCl 0.85433 2.0948 1.93163 2.32114 K KC1 K 2 S0 4 KC1O 4 K 2 PtCl 6 Pt 0.52441 0.44873 0.28219 0.16085 0.40061 1.71937 1.65199 1.45054 1 . 20643 1.60273 1.93231 1 . 73087 1.48676 1.88306 CN AgCN Ag 0.19426 0.24110 1.28839 1.38220 CNS AgCNS CuCNS BaSO 4 0.34996 0.47744 0.24880 1.54402 1 . 67891 1.39585 KC1 K 2 SO 4 KC1O 4 K 2 PtCl 6 Pt 0.85569 0.53811 0.30673 0.76397 HCNS AgCNS CuCNS BaS0 4 0.35604 0.48572 0.25312 1.55150 1.68639 1.40332 K 2 KC1 K 2 S0 4 KC1O 4 K 2 PtCl Pt 0.63170 0.54054 0.33993 0.19376 0.48258 1.80052 1.73283 1.53138 1.28727 1.67357 Co CoO CoSO 4 Co CoSO 4 0.38035 1.2713 0.48355 1.58019 2.10426 1.68444 Cr Cr 2 3 PbCr0 4 BaCrO 4 0.68421 0.16094 0.20523 1.83519 1.20667 1.31248 Li Li 2 SO 4 LiCl 0.12624 0.16368 1.10119 1.21399 Li 2 O LiCl LizSO* 0.35236 0\ 27176 1.54699 1.43419 1.78044 1.30537 1.33924 1.52493 1.55879 Cr 2 3 PbCrO 4 BaCrO 4 0.23522 0.29996 1.37148 1.47706 Mg MgO MgS0 4 Mg 2 P 2 O 7 0.60318 0.20201 0.21839 CrO 8 Cr 2 O 3 PbCrO 4 BaCrO 4 1.3158 0.30950 0.39468 2.11919 1.49066 1.59625 MgO MgS0 4 Mg 2 P 2 7 0.33491 0.36207 Cu CuO Cu 2 S CuCNS 0.79892 0.79857 0.52256 1.90250 1.90231 1.71814 Mn MnSO 4 MnS MiXiO* Mn 2 P 2 O 7 0.36378 0.63138 0.72030 0.38672 1.56083 1.80029 1.85751 1.58740 CuO Cu 2 S CuCNS Cu 0.99956 0.65408 1.2517 1.99981 1.81563 2.09750 MnO MnS0 4 MnS MnsO 4 Mn 2 P 2 O 7 0.46973 0.81529 0.93011 0.49961 1.67185 1.91131 1.96853 1.69863 F CaF 2 CaSO 4 0.48662 0.27908 1.68719 1.44574 852 TABLE OF CHEMICAL FACTORS. Sought. Found. Factor. , Log. Sought Found. Factor. Log. Mo MoO 3 ,0. 66667 1.82391 NO 2 NO 1.5332 2.18559 N NH 3 NH 4 C1 (NH 4 ) 2 PtCl 6 Pt 0.82247 0.26187 0.06310 0.14354 1.91512 1.41808 0.80005 1 . 15699 N 2 3 NO 1.2666 2.10263 P Mg2P 2 7 P ? O 5 ,24MoO 3 (NH 4 ) 3 P0 4 , 12MoO 3 . 27874 0.01725 0.01654 1.44519 0.23688 0.21842 Na NaCl Na 2 SO 4 0.39343 0.32378 1.59487 1.51026 P0 4 Mg,P 2 O 7 P 2 O 5 ,24Mo0 3 (NH 4 ) 3 P0 4 , 12MoO s 0.85345 0.05283 0.05063 1.93118 0.72287 0.70440 Na 2 O NaCl NasSO 4 0.53028 0.43640 1.72450 1.63989 NII 3 NH 4 C1 (NH 4 ) 2 PtC: 6 Pt 0.31821 0.07670 0.17449 1.50271 0.88483 1.24177 P 2 5 MgsPjOr P 2 O 5 ,24Mo0 8 (NH 4 ) 3 P0 4 , 12MoO 3 0.63793 0.03947 0.03785 1.80477 0.59628 0.57800 NH 4 NH 3 NH 4 C1 (NH 4 ) 2 PtCl 6 Pt 1.0592 0.33723 0.08125 0.18484 2 . 02497 1 . 52793 0.90985 1.26679 Pb PbO PbO 2 PbS PbSO 4 PbCr0 4 PbClo 0.92828 0.86616 0.86591 0.68312 0.64098 0.74491 1.96768 1.93760 1.93747 1.83449 1.80684 1.87210 NH 4 C1 NH 3 (NH 4 ) 2 PtCl 6 Pt 3.1409 0.24097 0.54815 2.49705 1.38196 1.73890 Ni NiO NiC 8 H 14 N 4 4 0.78576 0.20316 1.89529 1 . 30785 PbO Pb0 2 PbS PbSO 4 PbCrO 4 PbClo 0.93308 0.93281 0.73589 0.69050 0.80246 1.96992 1.96979 .86681 .83916 .90442 NiD Ni NiC 8 H 14 N 4 O 4 1.2727 0.25856 2.10471 1.41256 NO 3 NO NH S NH 4 C1 (NH 4 ) 2 PtiCle Pt CzoHnNsOj, 2.0663 3.6404 1.1591 0.27930 0.63535 0.16528 2.31520 2.56115 1.06411 1.44607 1.80301 1.21823 S SO 2 S0 2 S0 4 SO 4 H 2 H 2 S FeS 2 BaSO 4 BaSO 4 BaSO 4 BaS0 4 BaSO 4 BaSO 4 BaS0 4 0.13738 0.27446 0.34300 0.41154 0.42018 0.14602 0.25700 . 13792 .43848 .53530 1.61441 1.62343 1 . 16443 1.40994 NO 3 H NO NH 3 NH 4 C1 (NH 4 ) 2 PtCl 6 Pt C 2 oH 17 N 5 3 2.0999 3.6995 1.1779 0.28385 0.64570 0.16797 2.32220 2.56815 2.07111 1.45309 1.81003 1.22523 Sb Sb 2 4 Sb 2 S 3 0.78975 0.71418 1.89749 1.85381 Si SiO 3 SiO 2 SiO 2 0.46931 1 . 2653 1.67147 2.10221 N 3 6 NO NH 3 NH 4 C1 (NH 4 ) 2 Pt0 6 Pt C 20 H 17 N 5 3 1.7997 3.1707 1.0095 0.24327 0.55338 0.14396 2.25521 2.50116 2.00412 1.38608 1.74302 1 . 15824 Sn SnO 2 Sn0 2 Sn 0.78808 1.2689 1.89657 2.10343 Sr SrO SrC0 3 SrS0 4 Sr(N0 3 ) 2 0.84560 0.59358 . 47703 0.41403 1.92717 1.77348 1.67854 1.61703 TABLE OF CHEMICAL FACTORS. 853 Sought. Found. Factor. Log. Sought. Found. Factor. Log. SrO SrCO 3 SrSO 4 ^rf^VO "U 0.70196 0.56413 4806*3 1.84631 1.75138 1 68Q87 w WO 3 WO 3 w 0.79310 1 . 2609 1.89933 2.10067 01 ^Mvy^ 7n 7nO 0&A007 QflAQI Th ThO 2 0.87898 1.94398 ZnS 7nNH PO 0.67087 OQAAQO .82664 CfiOOf! Ti TiO 2 0.60051 1.77852 Zn 2 P 2 7 0.42891 .63237 U U 3 8 UO 2 (UC 2 ) 2 P 2 7 0.84824 0.88170 0.6*6703 1.92852 1.94532 1.824U ZnO' ZnS ZnNH 4 PO 4 Zn 2 P 2 O 7 0.83508 . 45598 0.53390 .92173 .65895 .72746 V V 2 6 0.56044 1 . 74853 Zr ZrC 2 0.73899 1.86864 854 LOGARITHMS. Natural Numbers. 1 1 2 3 4 5 6 7 8 9 PROPORTIONAL PARTS. 1 2 3 4 5 6 7 8 9 10 0000 0043 0086 0128 0170 0212 0253 0294 0334 0374 4 8 12 17 23 25 2 sr 37 11 0414 0453 0492 0531 0569 0607 0645 0682 0719 0755 4 8 11 1C 23 26J30 34 12 0792 0828 0864 0899J0934 0969 1004 1038 1072 1106 3 7 10 , 17 21 24 28 31 13 1139 1173 1206 123911271 1303 1335 1367 1399 1430 3 6 10 ie 19 23 2e 29 14 1461 1492 1523 1553 1584 1614 1644 i673 1703 1732 3 6 9 2 15 182124 27 15 1761 1790 1818 1847 1875 1903 1931 1959 1987 2014 3 6 8 14 17 2C 22 25 16 2041 2068 2095 2122 2148 2175 2201 2227 2253 2279 3 5 8 13 16 182ll24 17 2304 2330 2355 23SO 2405 2430 2455 2480 2504 12529 2 5 7 10 12 15 17 2022 18 2553 2577 2601 2625 264S 2672 2695 2718 2742 2765 2 5 7 g 12 14'l6 19 21 19 2788 2310 2833 2856 2878 2900 2923 2945 2967 2989 2 4 7 9 11 13 16 18 20 20 3010 3032 3054 3075 3096 3118 3139 3160 3181 3201 2 4 6 8 11 13 15 17 19 21 3222 3243 3263 3284 3304 3324 3345 3365 33853404 2 4 6 8 10 12 14 16 18 22 3424 3444 3464 34833502 3522 3541 3560 35793598 2 4 6 8 10 12 14 15 17 23 3617 3636 3655 36743692 3711 3729 3747 37663784 2 4 6 7 9 Il'l3 15 17 24 3802 3820 3838 3856 3874 3892 3909 3927 39453962 2 4 5 7 9 11 12 14 16 25 3979 3997 4014 4031 4048 4065 4082 4099 4116 4133 2 3 5 7 9 10 12 14 15 26 4150 4166 4183 4200 4216 4232 4249 4265 4281 4298 2 3 5 7 8 10 11 13 15 27 28 4314 4472 4330 4487 4346 4502 4362 4378 4518 4533 4393 4548 4409 4425 456414579 4440 4456 4594 4609 2 2 3 3 5 fi 6 6 8 8 911 911 13 14 1214 29 4624 4639 4654 4669 4683 4698 4713 4728 4742 4757 1 3 4 6 7 9 10 1213 30 4771 4786 4800 4814 4829 4843 4857 4871 4886 4900 1 3 4 6 7 9 10 11 13 31 4914 4928 4942 49554969 4983 4997 5011 5024 5038 1 3 4 6 7 s'lOll 12 32 5051 5065 5079 50925105 5119 5132 5145 5159 5172 1 3 4 7 8 91112 33 5185 5198 5211 5224 5237 5250 5263 5276 5289 5302 1 3 4 6 8 91012 34 5315 5328 5340 5353 5366 5378 5391 5403 5416 5428 1 3 4 6 8 9 10 11 35 5441 5453 5465 5478 5490 5502 5514 5527 5539 5551 1 2 4 6 7 9 1011 36 5563 5575 5587 5599 5611 5623 5635 5647 5658 5670 1 2 4 6 7 8 1011 37 56 S2 5694 5705 5717 5729 5740 5752 5763 5775 6786 1 2 3 6 7 8 9 10 38 5798 5809 5821 5832 5843 5855 5866 5877 5888 5899 1 2 3 6 7 8 910 39 5911 5922 5933 5944 5955 5966 5977 5988 5999 6010 1 2 7 8 9 10 40 6021 6031 6042 6053 6064 6075 6085 6096 6107 6117 1 2 6 8 910 41 6128 6138 6149 6160 6170 6180 6191 6201 6212 6222 1 2 6 7 8 a 42 6232 6243 6253 6263 6274 6284 6294 6304 6314 6325 1 2 6 7 8 9 43 6335 6345 6355 6365 6375 6385 6395 6405 6415 6425 1 2 6 7 8 9 44 6435 6444 6454 6464 6474 6484 6493 6503 6513 6522 1 2 t 4 6 7 8 9 45 6532 6542 6551 6561 6571 65SO 6590 6599 6609 6618 1 2 c 4 6 7 8 9 46 6628 6637 6646 6656 6665 6675 6684 6693 6702 6712 1 2 ^ 4 6 7 7 5 47 6721 6730 6739 6749 6758 6767 6776 6785 6794 6803 1 2 2 4 r 6 7 48 6812 6821 6830 6839 6848 6857 6866 6875 6884 6893 1 2 3 4 4 5 6 7 3 49 6902 6911 6920 6928 6937 6946 6955 6964 6972 6981 1 2 3 4 4 c 6 7 8 50 6990 6998 7007 7016 7024 7033 7042 7050 7059 7067 1 2 3 3 4 5 6 7 51 7076 7084 7093 7101 7110 7118 7126 7135 7143 7152 1 2 3 3 4 5 6 7 8 52 7160 7168 7177 7185 7193 7202 7210 7218 7226 7235 1 2 2 3 4 5 6 7 7 53 7243 7251 7259 7267 7275 7284 7292 7300 7308 7316 1 2 2 3 4 5 6 6 7 54 7324 7332 7340 734 ? 7356 7364 7372 7380 7388 7396 1 2 2 3 4 5 6 6 7 g . LOGARITHMS. 855 | Natural ! J Numbers. I 1 2 3 4 5 6 7 8 9 PROPORTIONAL PARTS. 1 2 n 4 5 6 7 8 6 55 7404 7412 741S 7427 7435 7443 7451 7459 7466 7474 1 2 2 3 4 5 5 6 7 56 7482 7490 7497 7505 7513 7520 7528 7536 7543:7551 1 2 9 3 4 5 5 6 7 57 7559 7566 7574 7582 7589 7597 7604 7612 7619 7627 1 2 2 3 4 5 5 6 7 58 7634 7642 7649 7657 7664 7672 7679 7686 76947701 1 1 2 3 4 4 5 6 7 59 7709 7716 7723 7731 7738 7745 7752 7760 7767 7774 1 1 2 3 4 4 5 G 7 60 7782 7789 7796 7803 7810 7818 7825 7832 7839 7846 1 1 2 3 4 4 5 6 6 61 7853 7860 7868 7875 7882 7889 7896 7903 7910 7917 1 1 2 3 4 4 5 6 6 62 7924 7931 7938 7945 7952 7959 7966 7973 7980 7987 1 1 2 3 3 4 5 6 6 63 7993 8000 8007 8014 8021 8028 8035 8041 8048 8055 1 1 2 3 3 4 5 5 6 64 8062 8069 8075 8082 8089 8096 8102 8109 8116 8122 1 1 2 3 3 4 5 5 6 65 8129 8136 8142 8149 8156 81628169 8176 8182 8189 1 1 2 3 3 4 5 5 6 66 8195 8202 8209 8215 8222 8228'8235 8241 8248 8254 1 1 2 3 3 4 5 5 6 67 8261 8267 8274 8280 8287 8293,8299830683128319 1 1 2 3 3 4 5 5 6 68 8325 8331 8338 8344 8351 8357 8363 8370 8376 8382 1 1 2 3 3 4 4 5 6 69 8388 8395 8401 8407 8414 8420 8426 8432 8439 8445 1 1 2 2 3 4 4 5 6 70 71 8451 8513 8457 8519 8463 8525 8470 8531 8476 8537 8482 8488 8543 8549 8494 8555 8500 8561 8506 8567 1 1 1 1 2 2 2 2 3 3 4 4 4 4 5 5 6 5 72 8573 8579 8585 8591 8597 8603 8609 8615 8621 8627 1 1 2 9 3 4 4 5 5 73 8633 8639 8645 8651 8657 8663 8669 8675 8681 8686 1 1 2 2 3 4 4 5 5 74 8692 8698 8704 8710 8716 87228727 8733 8739 8745 1 1 2 2 3 4 4 5 5 75 76 8751 8808 8756 8814 8762 8820 8768 8825 8774 8831 8779 8837 8785 8791 8842 8848 8797 8802 8854 '8859 1 1 1 1 2 2 2 2 3 3 3 3 4 4 5 5 5 5 77 8865 8871 8876 8882 8887 8893 8899 8904 89108915 1 1 2 2 3 3 4 4 5 78 79 8921 8976 8927 8982 8932 8987 8938 8993 8943 8993 8949 9004 8954 8960 9009 9015 8965 8971 9020 9025 1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 4 5 5 80 9031 9036 9042 9047 9053 9058 9063 9069 9074 9079 1 1 2 2 3 3 4 4 5 81 82 9085 9138 9090 9143 9096 9149 9101 9106 9154 9159 911291179122 916591709175 91289133 91809186 1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 4 5 5 83 9191 9196 9201 9206 9212 9217 9222 9227 9232 9238 1 1 2 2 3 3 4 4 5 84 9243 9248 9253 9258 9263 9269 9274 9279 92849289 1 1 2 2 3 3 4 4 5 85 9294 9299 9304 9309 9315 9320 9325 9330 9335 9340 1 1 2 2 3 3 4 4 5 86 9345 9350 9355 9360 9365 9370 9375 9380 9385 9390 1 1 2 2 3 r ] 4 4 5 87 9395 9400 940594109415 9420 9425 9430 9435 9440 1 1 2 3 3 4 4 88 9445 9450 9455 9460 9465 9469 9474 9479 9484 9489 1 1 2 3 g 4 4 89 9494 9499 9504 9509 9513 9518 9523 9528 95339538 1 1 2 3 3 4 4 90 9542 9547 9552 9557 9562 9566 9571 9576 95S19586 1 1 2 3 3 4 4 91 9590 9595 9600 96059609 9614 9619 9624 96289633 1 1 2 3 3 4 4 92 93 9638 9685 9643 9647 9689 9694 96529657 9699! 9703 9661 9708 9666 9713 9671 9717 9675 9722 9680 9727 ) 1 1 1 2 2 3 3 3 3 4 4 4 4 94 9731 9736 9741 9745 9750 9754 9759 9763 9768 9773 1 2 3 3 4 4 95 9777 9782 9786 9791 '9795 9800 98059809 9814 9818 1 2 3 3 4 4 96 9823 9827 9832 98369841 9845 9850 9854:9859 9863 1 2 3 3 4 4 97 9S 9868 9912 9872987798819886 9917992199269930 9890 9894 9899 9903 990S 9934 9939 9943 9948 1 9952 ) ) 1 1 2 2 3 3 3 4 4 4 4 99 9956 9961 9965 9969 9974 9978 9983 9987 9991 9996 1 2 3 3 3 4 ) 1 856 ANTILOGARITHMS. i 1 2 8 4 5 6 7 8 9 PROPORTIONAL PABT&. 1 2 3 4 5 6 7 8 9 .00 1000 1002 1005 1007 1009 1012 1014 1016 1019 1021 1 1 1 1 2 2 2 .01 1023 1026 1028 1030 1033 1035 1038 1040 1042 1045 1 1 1 1 2 2 2 .02 1047 1050 1052 1054 1057 1059 1062 1064 1067 1069 1 1 1 1 2 2 2 .03 1072 1074107610791081 10841086 1089 1091 1094 1 1 1 1 2 2 2 04 1096 10991102 11041107 1109 1112 1114 1117 1119 1 1 1 1 2 2 2 2 .05 1122 11251127 1130 1132 1135 1138 1140 1143 1146 1 I 1 1 2 2 2 2 .06 1148 1151 11531156 1159 11611164 1167 1169 1172 1 1 1 2 2 2 2 .07 1175 117811801183 1186 11891191 1194 1197 1199 1 1 1 2 2 2 2 .08 1202 120512081211 1213 12161219 1222 1225 1227 1 1 1 2 2 2 3 .09 1230 1233 1236 1239 1242 1245 1247 1250 1253 1256 1 1 1 2 2 2 3 .10 1259 1262 1265 1268 1271 12741276 1279 1282 1285 n 1 1 1 2 2 2 3 .11 1288 1291 1294 1297 1300 1303 1306 1309 1312 1315 1 1 2 2 2 2 3 .12 1318 1321 1324 1327 1330 1334 1337 1340 1343 1346 1 1 2 2 2 2 3 .13 1349 1352 1355 135S 1361 1365 1368 1371 1374 1377 1 1 2 2 2 3 3 ,14 1380 1384 1387 1390 1393 1396 1400 1403 1406 1409 1 1 1 2 2 2 3 3 .15 1413 1416 1419 1422 1426 1429 1432 1435 1439 1442 1 1 1 2 2 2 3 3 .16 1445 1449145214551459 14621466 1469 1472 1476 1 1 1 2 2 2 3 3 .17 1479 1483 1486 14S9 1493 14961500 1503 1507 1510 1 1 1 2 2 2 3 3 .18 1514 1517 1521 1524 152 Q 1531 1535 1538 1542 1545 1 1 1 2 2 2 3 3 .19 1549 15521556^560 1563 1567 1570 1574 1578 1531 1 1 1 2 2 3 3 3 .20 1585 1589 1592 1596 1600 1603 1607 1611 1614 1618 () 1 1 1 2 2 3 3 3 .21 1622 1626 1629 1633 1637 1641 1644 1648 16521656 1 1 2 2 2 3 3 3 .22 1660 1663166716711675 16791683 1687 1690 1694 1 1 2 2 2 3 3 3 .23 1698 1702 1706 1710 1714 1718 1722 1726 17301734 1 1 2 2 2 3 3 4 .24 1738 1742 1746 1750 1754 1758 1762 1766 1770 1774 1 1 2 2 2 3 3 4 .25 1778 178217861791 1795 1799 1803 1807 1811 1816 1 1 2 2J2 3 3 .26 1820 1824 1828 1832 1837 1841 1845 1849 18541858 1 1 2 2 3 3 3 .27 1862 18661871 18751879 1884 1888 1892 1897 1901 1 1 2 2 3 3 3 .28 1905 1910 1914 1919 1923 1928 1932 1936,1941 1945 1 1 2 2 3 3 4 .29 1950 1954 1959 1963 1968 1972 1977 1982 19861991 1 1 2 2 3 3 4 .30 1995 2000 2004 2009 2014 2018 2023 2028 2032 2037 1 1 2 2 3 3 4 .31 2042 2046 2051 2056 2061 2065 2070 2075 2080 20 ]4 1 1 2 2|3 3 4 .32 2089 2094 209921042109 2113 2118 2123 2128 2133 1 1 2 2 3 3 4 4 .33 2138 2143 214821532158 2163,2168 2173 2178 2183 1 1 2 2 3 3 4 4 .34 2188 2193 2198 2203 2208 2213 2218 2223 2228 2234 1 1 2 2 33 4 4 5 .35 2239 2244 2249 2254 2259 2265 2270 2275 2280 2286 1 1 2 2 3 3 4 5 .36 2291 2296 2301 2307 2312^317 2323 2328 2333 2339 1 1 2 2 3 3 4 5 .37 2344 2350 2355 2360 2366 2371 2377 2382 2388 2393 1 1 2 2 33 4 5 .38 2399 2404 2410 2415 2421 2427 2432 2438 2443 2449 1 1 2 o 3 3 4 5 .39 2455 2460 2466 2472 2477 2483 2489 2495 2500 2506 1 1 2 2 3 3 5 5 .40 2512 2518 2523 2529 2535 2541 2547 2553 2559 2564 1 1 2 2 3 4 4 5 5 .41 2570 2576 2582 2588 2594 2600 2606 2612 2618 2624 1 1 2 2 3 4 4 5 5 .42 2630 2636 2642 2649 2655 2661 2667 267312679 26 5 1 1 2 2 3 4 4 5 6 .43 2692 2698 2704 2710 2716 2723 2729 27352742 2748 1 1 2 3 3 4 4 5 6 .44 2754 2761 2767 2773 2780 2786 2793 2799 2805 2812 1 1 2 3 3 4 4 5 6 .45 2818 2825 2831 2838 2844 2851 2858 2864 2871 2877 1 1 2 3 3 4 5 5 ft .46 2884 2891 2897 2904 2911 2917 2924 2931 293S 2944 1 1 2 3 3 4 5 5 6 .47 2951 2958 2965 2972 2979 2985 2992 2999 3006 3013 1 1 2 3 3 4 5 5 6 .48 3020 3027 3034 3041 3048 3055 3062 3069 3076 3083 1 1 2 3 4 4 5 6 S .49 3090 3097 3105 3112 3119 31263133 3141 3148 3155 1 l j? 3 4 4 5 6 6 ANTILOGARITHMS. 857 1 1 2 3 4 5 6 7 8 PROPOR 9 TIONAL, PARTS. 123- 156789 .50 316231703177 31843192 319S 3206 3214 3221 3228 i i 2 3 44567 .51 3236 3243 3251 3258 3266 3273 3281 329 3296330 i 2 2 3 45567 .52 331133193327 33343342 3350 33573365 33733381 i 2 2 3 45567 .53 338833963404 34123420 342^ 34363443 34513459 l 2 2 45667 .54 3467 34753483 3491 3499 3508 3516 3524 35323540 122, 45667 .55 3548 3556 3565 3573 35S1 3589 35973606 3614'3622 122: 45677 .56 3631 3639 364S 3656 3664 3673 36813690 36983707 123; 45678 .57 371537243733 3741 3750 375S 3767 3776 37843793 123; 45678 .58 3802 381 1J3819 3828 3837 3846 3855 3864 38733882 1 2 3 ; 45678 .59 3890 3899 3908 39173926 3936 39453954 3963 3972 i 2 3 ^ 55678 .60 3981 3990 3999 4009 4018 4027 4036 4046 40554064 1234 56678 .61 4074 4083 4093 4102:4111 4121 4130414041504159 1234 56789 .62 4169 4178 4188 4198 4207 4217 4227 4236 4246 '4256 1234 56789 .63 4266 4276 4285 4295 4305 4315 4325 433543454355 1234 56789 .64 4365 4375 4385 43954406 4416 4426 443644464457 1234 56789 .65 4467 4477 4487 4498 4508 4519 4529 4539 45504560 1234 56789 .66 .67 4571 4581 467714688 4592 4699 4603 4613 47104721 4624 4732 4634 4742 464546564667 1234 475347644775 1234 567 9 10 57 8 9 10 .68 4786;4797 4808 4819|4831 4842 4853 486448754887 1234 67 8 9 10 .69 4898 4909 4920 49324943 4955 4966 4977 4989 5000 1 2 3 s 6 7 8 9 10 1 70 5012 5023 5035 5047 5058 5070 5082 5093 5105 5117 1245 67 8 911 .71 5129^5140 5152 5164 5176 5188 5200 521252245236 1245 6 7 8 10 11 .72 5248 5260 5272 5284 5297 5309 5321 5333 5346 5358 1245 67 9 10 11 .73 5370 538315395 540S 5420 5433 5445 545854705483 1345 6 8 9 10 11 .74 5495 5508 5521 5534 5546 5559 5572 5585 5598 5610 1345 6 8 9 10 12 .75 5623 5636 5649 5662 5675 5689 5702 571557285741 1345 7 8 9 10 12 .76 5754 5768 5781 5794 5808 5821 5834 5848 5861 5875 1345 7 8 9 1 12 .77 5888 5902 5916 5929 5943 5957 5970 598459986012 1345 7 8 1C 1 12 .78 6026 6039 6053 6067 60S1 6095 6109 612461386152 1346 7 8 10 11 13 .79 6166 6180 6194 6209 6223 6237 6252 6266 6281 6295 1346 7 9 10 11 13 .80 6310 6324 6339 6353 6368 383 6397 6412 6427 6442 1346 7 9 10 12 13 .81 6457 6471 6486 6501 6516 531 65466561! 6577 6592 2356 8 9 11 12 14 .82 6607 6622 6637 6653 6668 683 6699,671467306745 2356 8 9 11 12 14 .83 6761 6776 6792 6803 6S23 S39 6S55|6871 6887 6902 2356 8 911 13H4 .84 6918 6934 6950 6966 6982 998 7015 7031 7047 7063 2356 8 10 11 13 15 .85 .86 7079 244 7096 7261 7112 7278 7129 7295 7145 7311 161 328 717871947211 7345:7362 7379 7228 2357 7396 2357 8 10 12 13 15 8 10 12 13 15 .87 7413 7430 7447 7464 7482 499 751675347551 7568 2357 9 10 12 14 16 .88 586 7603 7621 7638 7656 674 7691 7709 7727 7745 2457 911 121416 .89 762 7780 7798 7816 7834 852 7870 7889 7907 7925 2457 9 11 13 14 16 .90 7943 7962 7980 7998 8017 035 8054 8072 8091 8110 2467 9 11 13 15 17 .91 128 8147 8166 8185 8204 222 8241 8260:8279 8299 2468 9 11 13 15 17 .92 S3 18 8337 8356 8375 8395 414 843384538472 8492 2468 10 12 14 15 17 .93 511 8531 8551 8570 8590 610 8630 8650 8670 8690 2468 10 12 14 16 18 .94 8710 8730 8750 8770 8790 810 8831 8851 8872 8892 2468 10 12 14 16 18 .95 913 8933 8954 8974 8995 016 9036 9057 9078 9099 2468 10 12 15 17 19 .96 9120 9141 9162 9183 9204 226 9247 9268 9290 9311 2468 11 13 15 17 19 .97 333 9354 9376 9397 9419 441 9462 9484 9506 9528 2479 11 13 15 17 20 .98 9550 9572 9594 9616 9638 9661 9683 97059727 9750 2479 11 13161820 .99 9772 9795 9817 9840 9863 9886 9908 9931 9954 9977 2579 11 14 16 1820 APPENDIX I* The Influence of Fine Grinding on Composition. THE rate at which a substance dissolves increases as the amount of surface exposed to the solvent is increased and for this reason solid substances always dissolve more quickly when reduced to a fine powder. Moreover, when a material undergoes chemical attack an insoluble substance is often formed and, during the process of solution, the insoluble substance may form a protective coating over particles of material that have been unacted upon. This danger is diminished if the material is in the form of a fine powder. For these reasons the chemist usually prefers to grind a solid substance to an impalpable con- dition before attempting to analyze it. This practice, while desirable in most cases and absolutely necessary in others, is accompanied by certain disadvantages. If the material is hard there is always some contamination from the material of which the grinding apparatus is constructed. Thus when the sample is ground in a steel mortar or in a steel ball-mill, it will be contaminated with a little iron and if ground in an agate mortar with a little silica.t Again, if the sample readily undergoes slight decomposition, such a chemical change is likely to take place during the operation of grinding. In this way the determination of moisture, of ferrous iron, and of sulphur may be influenced very appreciably. * The material in this appendix is added temporarily in this form at a time when the book is out of print and there is not time to get out a complete new edition. It will be introduced eventually into the main portion of the book. (Editor.) f Hempel (Z. angew. Chem., 1901, 843) found that an agate mortar and pestle lost 0.052 gm. in grinding 10 gms. of glass to a fine powder. 859 86o APPENDIX I A number of investigators have pointed out the effect of grinding upon the moisture content of a sample. If the sample is practically dry, it is likely to absorb considerable moisture when dried in the air. Thus Hillebrand * found that a piece of unglazed porcelain contained no moisture originally, but showed 0.62 per cent, of water when ground. If the substance is very hygroscopic, this danger becomes greater. On the other hand, grinding often causes loss of moisture. This is notably true in the cass of substances containing water of crystallization or superficial moisture. Thus grinding can easily reduce the moisture content of a sample of gypsum from twenty to five per cent, and a sample of coal may show several per cent of moisture when large lumps of it are tested and very little moisture after it is reduced to a fine powder. The heat produced by grinding may not only serve to expel moisture from the sample, but it may cause chemical change. Thus Mauzelius f has shown, and the experiment has been repeated by Hillebrand,! that the ferrous iron content of a rock becomes smaller on account of grinding it to a fine powder. It has also been found that some sulphur may be lost by long grinding of a sample of pyrite. The effect of grinding, therefore, accounts for many cases of divergent results obtained in the hands of different chemists who have analyzed the same original material. The Use of Cupferron in Quantitative Analysis. Cupferron, CeHs N N ONH 4 , is a short name for the O ammonium salt of nitrosophcnyl hydroxylamine. O. Baudisch first suggested its use as a reagent for the quantitative precip- itation of cupric and ferric ions. By means of cupferron, it is * The Analysis of Silicate and Carbonate Rocks, Bull. 422, U. S. Geol. Survey. f Sveriger Geol. Undersokning, Arsbok 1 (1907). t J. Am. Chem. Soc., 30, 1120 (1908). O. Baudisch, Chem.-Ztg., 33, 1298 (1909); Baudisch and King, J. Ind. Eng. Chem., 3 t 629 (1911). USE OF CUPFERRON. 86 1 possible to precipitate quantitatively copper, iron and titanium from strongly acid solutions and in this way a number of sep- arations may be accomplished which otherwise involve con- siderable difficulty. The copper salt is gray, the ferric salt red and the titanium salt yellow. As a precipitant for copper, the reagent apparently offers no special advantages, and when silver, lead, mercury, tin or bismuth are present, these elements contaminate the precipi- tate to some extent. On the other hand, in the case of iron and titanium it is very useful to possess a reagent which will precipitate these elements quantitatively from acid solutions without contamination from aluminium, chomium, manganese, nickel, cobalt, zinc or alkaline earth metal. Zirconium is also precipitated by cupferron.* The advantages of cupferron for effecting separations has been pointed out by a number of chemists. f Cupferron is readily soluble in water and the ammoniac al solution keeps well. The reagent is not very stable in acid solutions, particularly in hot solution or when an oxidizing agent is present. By oxidation, nitrosobenzene is formed and its presence can be detected by the peculiar, sweetish odor in nearly every precipitation with cupferron. When much nitro- sobenzene is formed it separates out in the form of white needles. Precipitation with cupferron is always effected in cold, acid solution and an excess of the reagent must be used. The pre- cipitate is stable as long as an excess of cupferron is present. When iron and copper are precipitated together by means of cupferron, washing with ammonia serves to remove the copper and the excess of cupferron; it also serves to convert the iron precipitate into ferric hydroxide, in which form it is more readily converted into ferric oxide on ignition. In the ammoniacal solution of the copper precipitate, the copper is again precipitated upon the addition of acetic acid and by washing with one per *Schroeder, Z. anorg. Chem., 72, 89 (1911). fNissenson, Z. angew. Chem., 23, 969 (1910); Biltz and Hodtke, Z. anorg. Chem., 66, 426 (1910); Hanus and Soukup, ibid., 68, 52 (1910); R. Fresenius, Z. anal. Chem., 50, 35 (1911); Bellucci and Grassi, G&zz. chim. ital., 43, I, 570; Thornton, Am. J. Sci., 37, 173 and 407 (1914). 862 APPENDIX /. cent, sodium carbonate solution the cupferron may be removed from the copper. The manner of preparing the reagent and four methods illus- trating its use will be described. Preparation of Cupferron.* Sixty gms. of nitrobenzene, lf)00 cc. of water and 30 gms. of ammonium chloride are stirred rapidly to an emulsion with the aid of an efficient mechanical stirrer. To the emulsion (stirred constantly) 80 gms. of zinc dust are added in small portions, while keeping the temperature of the liquid between 15 and 18 by the occasional addition of shaved ice. The addition of the zinc dust is continued until the odor of nitro- benzene disappears and the precipitate of zinc hydroxide appears gray. Usually about half an hour is required for the reduction. The solution is then filtered and the precipitate washed a few times with ice water. The filtrate is cooled to nearly and saturated with sodium chloride, which causes the separation of phenylhydroxylamine crystals in the form of white needles. These are filtered immediately with the aid of suction and dried between filter paper. As phenylhydroxylamine solutions are active skin poisons, the hands and face should be washed with water and alcohol in case any of the solution spatters upon them. The crystals are dissolved in 300-500 c.c. of ether and the ethereal solution is filtered through a dry filter, cooled to 0, and saturated with dry ammonia gas. While still at some- what more than the theoretical quantity (one mole) of fresh amyl nitrate is added all at once. The solution will suddenly become hot and the whole beaker will be filled with snow-white crystals of cupferron. After filtration and washing with ether, the crystals are dried by filter paper and kept in a closely-stoppered bottle in which a small lump of ammonium carbonate is also placed. * The salt may be purchased from Elmer & Amend, New York. USE OF CUPFERRON. 863 1. THE DETERMINATION OP IRON IN MANGANESE ORES. Principle. After effecting the solution of the ore, the iron is precipitated in acid solution by cupferron. The filtrate then contains all the aluminium, chromium, manganese, nickel, cobalt, zinc and alkaline earths. After the removal of any of these elements that may be present, by washing the precipitate with cold water, the organic matter is removed and the iron con- verted into ferric hydroxide by washing the precipitate with ammonia, and the iron is weighed as Fe2Os. Procedure. About 1 gm. of the finely pulverized ore is dis- solved in concentrated hydrochloric acid, sp. gr. 1.2, and the solution evaporated to dryness. The residue is moistened with strong hydrochloric acid, the solution diluted with water, boiled and filtered. The residue is fused with sodium carbonate in a covered platinum crucible, and after the fusion, the melt is taken up in water and hydrochloric acid. The solution is evap- orated to dryness, and after the removal of the silica in the usual manner, the filtrate is added to the main solution. The cold solution is then treated with a solution of about 3 gms. cupferron in 50 c.c. of cold water, added in a fine stream down the sides of the beaker while the solution is being vigorously stirred. A brownish red, partly amorphous and partly crystal- line precipitate of the ferric salt separates out. As soon as a drop of the reagent causes the formation of a snow-white pre- cipitate of nitrosophenylhydroxylamine, all the iron is precip- itated. A slight excess of the reagent is added and the solution allowed to stand about 10 minutes, after which it is filtered through an ashless paper, using gentle suction. In case the last particles of the precipitate cling tenaciously to the beaker, a little ether is added to loosen them and the ether removed by adding a little boiling water. The precipitate is now washed with cold water until the washings are no longer acid to litmus and then with ammonia (sp. gr. 0.96) to remove the excess of the reagent and form ferric hydroxide. The filter is finally washed once more with cold water. The precipitate is ignited carefully with the filter, avoiding reduction, and weighed as Fe 2 O 3 . 864 APPENDIX I. Remark. Following this procedure, R. Fresenius * obtained the values 18.23 per cent., 18.56 per cent., and 18.03 per cent. Fe in the analysis of three manganese ores which when analyzed volumetric ally for iron, showed respectively, 18.25 per cent., 18.49 per cent., and 17.93 per cent. Fe. The manganese may be determined in the nitrate. 2. THE DETERMINATION OF MANGANESE IN FERRO-MANGANESE. About one gram of the material is dissolved in strong hydro- chloric acid and the residue is fused with sodium carbonate in a platinum crucible. The product of the fusion is extracted with water and alcohol is added to reduce the manganate, formed by the fusion, to hydrated manganese dioxide. The aqueous extract from the fusion is filtered, the residue is dissolved in hydrochloric acid, and the solution added to that obtained in the first place. The iron is precipitated with cupferron, as in the previous method, but it is not advisable to add the ammonia washings to the solution on account of the organic material they contain. In the filtrate the manganese is precipitated as described on page 122 and determined as sulphide according to page 125. Remark. In testing this method R. Fresenius * obtained duplicate values of 81.04 and 81.01 per cent. Mn as compared with 8.1.01 and 81.17 by the basic acetate method. The cupferron method is much shorter and easier to carry out. 3. THE DETERMINATION OF NICKEL AND COBALT IN ARSENICAL SULPHIDE ORES.f One gram of the sample is dissolved in concentrated hydro- chloric acid, which is saturated with bromine, and the solution evaporated somewhat to volatilize arsenic. Ten cubic centi- meters of sulphuric acid (1 : 1) are added and the solution evap- orated until dense fumes are evolved. The solution is diluted, heated to boiling, and hydrogen sulphide is introduced to pre- cipitate the metals of the copper group and any remaining arsenic. In the filtrate the iron is precipitated by cupferron * Z. anal. Chem., 50, 35 (1911). t H. Nissenson, Z. angew. Chem., 23, 969 (1910). USE OF CUPFERRON. 865 as in Method 1, and after evaporating the nitrate until fumes of sulphuric acid are evolved, the nickel and cobalt are deter- mined electrolytically according to page 136. In case it is desired to determine the arsenic as well, the ore is dissolved in 15 c.c. of concentrated sulphuric acid, and the arsenic is precipitated as sulphide by passing hydrogen sul- phide into the hot solution. The arsenic is separated from cop- per according to page 235 by means of alkaline sulphide, the arsenic sulphide precipitated in the nitrate upon making it acid, the arsenic sulphide dissolved in concentrated hydrochloric acid and potassium chlorate, and the arsenic determined as magnesium pyroarsenate according to page 206. 4. THE DETERMINATION OF TITANIUM AND ITS SEPARATION FROM IRON, ALUMINIUM AND PHOSPHORIC ACID.* Principle. The iron may be reduced completely to ferrous salt by passing hydrogen sulphide into an acid solution, and then, in the presence of tartaric acid which prevents the pre- cipitation of titanium, precipitated as ferrous sulphide in ammoni- acal solution. After acidifying and boiling off the hydrogen sulphide, the titanium can be precipitated quantitatively by means of cupferron while the aluminium and phosphoric acid remain in the tartaric acid solution. It is unnecessary to remove the organic matter before igniting the yellow titanium precipitate. Procedure. To the solution, which should have a volume not greater than 100 c.c., at least four times as much tartaric acid is added as corresponds to the weight of the oxides of iron, titanium, aluminium, aad phosphorus. The solution is neutral- ized with ammonia, acidified with 3 c.c. of sulphuric acid (1 : 1) and hydrogen sulphide is introduced until the solution becomes colorless. Unless the iron is all reduced, the subsequent pre- cipitate of iron sulphide will contain some titanium. Ammonium hydroxide is now added in considerable excess and the iron is completely precipitated as ferrous sulphide by introducing * W. M. Thornton, Jr., Am. J. Science, 37, 407 (1914). 866 APPENDIX I. more hydrogen sulphide gas; the solution should remain alkaline to litmus paper. The ferrous sulphide is filtered off and washed with water containing a little colorless ammonium sulphide. To the filtrate, 40 c.c. of H 2 S0 4 (1 : 1) are added and the lib- erated hydrogen sulphide is expelled by boiling. When this is accomplished, the solution is cooled to room temperature, di- luted to 400 c.c., and treated with an, excess of 6 per cent, cupferron solution, which is added slowly down the sides of the beaker while the solution is being well stirred. After the precipitate has subsided, the supernatant liquid is tested by adding more of the cupferron solution. A white precipitate of nitroso- phenylhydroxylamine indicates that an excess of the reagent is present but a yellow turbidity shows that the precipitation of the titanium is incomplete. It is also well to test the filtrate in the same way. The precipitate of the titanium precipitate is collected on filter paper, using gentle suction, and washed twenty times with normal hydrochloric acid solution. The precipitate is then ignited cautiously in a platinum or quartz crucible until the organic matter is all consumed. Finally, it is heated to constant weight over a Me*ker burner and weighed as TiO 2 . The Determination of Titanium Dioxide in Titanium-Iron Ore. Most titanium ores are very difficultly soluble in the usual solvents. The analysis of these ores is also complicated by the fact that aqueous solutions of titanium salts readily undergo hydrolysis, partly on account of the amphoteric character of titanium dioxide and partly because of the insolubility of this oxide. If sodium titanate, Na 2 TiC>3, is prepared by fusion with sodium carbonate, it undergoes hydrolysis when treated with water and an acid titanate of sodium is precipitated, e.g., 2Na 2 O 9TiO 2 5HO 2 . This acid titanate is soluble in hydrochloric acid, sp. gr. 1.12, but if the solution is diluted and heated to boiling, hydrolysis takes place, as with any other salt of titanium, and metatianic acid, H 2 TiC>3, is precipitated. A method for determining titanium after fusing with sodium carbonate and dis- DETERMINATION OF TITANIUM. 867 solving the sodium titanate in hydrochloric acid has already been given (p. 118). One of the best methods for attacking titanium dioxide, or an insoluble titanium ore, is based upon the fusion of the ore with potassium acid sulphate. This converts titanium into titanium sulphate,* (TiSO^, which is soluble in cold water. Potassium acid sulphate melts at about 200 and, on further heating, is easily decomposed into potassium pyrosulphate, K 2 S 2 O 7 : 2KHS0 4 = K 2 S 2 O 7 + H 2 0, which melts at a little over 300. Upon raising the temperature sufficiently, it is easy to transform the potassium pyrosulphate into normal potassium sulphate, which melts at about 1050. Between these last two temperatures, sulphuric acid is avail- able for attacking insoluble oxides of iron, aluminium and titanium : = Ti(SO 4 ) 2 +2K 2 S0 4 . If the temperature is raised too high, however, Ti0 2 may be formed again, as is the case in heating the residue after the removal of silica with hydrofluoric and sulphuric acids : Ti(SO 4 ) 2 = TiO 2 +2SO 3 . Moreover, if the heating is continued long enough to convert all the pyrosulphate into molten potassium sulphate, there is some danger of the crucible bursting upon the solidification of the melt.f When potassium acid sulphate is decomposed into potas- sium pyrosulphate and water, the latter is liberated at a tem- perature far above its boiling-point. There is danger of the contents of the crucible boiling over if the reaction takes place * Titanium sulphate shows a tendency to form complex ions. Thus Warren (Pogg. Ann., 102, 449 (1857) obtained the salt K 2 Ti(SO 4 ) 3 after fusing with potassium acid sulphate. See also Schutte, Z. anorg. Chem., 26, 239 (1901). t H. P. Talbot, Quantitative Chemical Analysis, 5th Edition, p. 4. 868 APPENDIX I. too quickly. Before making the fusion, therefore, it is advisable to heat the potassium sulphate by itself until vapors of sul- phuric anhydride begin to be evolved. A low flame of the Bunsen burner is sufficient. The following procedure is given, not because it gives the most convenient or the most accurate method for determining titanium in an insoluble ore, but because it is an especially good method for teaching the art of fusing with potassium acid sul- phate to advanced students in quantitative chemical analysis. The method will give excellent results when properly carried out, but is likely to give too low results if the ore is not entirely decomposed by the acid sulphate fusion, or too high results if the titanic metatitanic acid is not purified sufficiently and washed free from alkali salts. Procedure. From 0.4 to 0.6 gm. of the finely-ground ore is weighed into a platinum crucible and fused with 6 to 8 times its weight of sodium carbonate for at least 30 minutes. The fused mass is extracted with hot water and, when the melt is all disintegrated, the solution is decanted through an ashless filter paper. The residue is boiled with 25 c.c. of normal sodium carbonate solution, transferred to the filter, and washed twice with dilute sodium carbonate solution and several times with hot water. This serves to remove the phosphoric acid as sodium phosphate,* and to convert titanium into acid sodium titanate; the iron is left with the titanium as insoluble ferric oxide. The residue and filter paper are ignited in a platinum cru- cible at a low temperature until the carbon of the filter paper is all consumed. From 12 to 15 parts of potassium pyrosul- phate, which has just been prepared by heating potassium acid sulphate until all the water is expelled, are added to the contents of the crucible and the latter is very carefully heated. The temperature should be kept low until all the potassium pyro- sulphate has melted, then the heat should be gradually increased until sulphuric acid fumes are noticed on removing the cover * Phosphoric acid interferes with the precipitation of metatitanic acid by hydrolysis. It tends to precipitate with the titanium to some extent and also to render the precipitation incomplete. DETERMINATION OF TITANIUM. 869 of the crucible. At no time should dense fumes be given off from the covered crucible, but it will be necessary to raise the temperature from time to time in order to keep the flux melted. Occasionally the cover of the crucible should be raised to see if the residue has dissolved in the flux. Sometimes the fusion can be accomplished in half an hour, but more often nearly an hour is required. Finally, when the contents of the cruci- ble are at a dull red heat and there is no sign of undissolved material, a coil of platinum wire is suspended in the melt and it is allowed to cool. When cold, the fusion is then withdrawn by heating the sides of the crucible. The melt is suspended in 200 c.c. of cold water and 100 c.c. of saturated sulphurous acid solution, and allowed to stand in a cool place until solution is complete, stirring from time to time. The solution is usually effected by allowing the dilute acid to stand over night with the melt suspended near the top of the liquid. The process may be hastened by stirring with a mechanical stirrer. If necessary the solution is filtered, and the residue treated with sulphuric and hydrofluoric acids to expel silica, fused with more potassium pyrosulphate, and the melt added to the solution previously obtained. Finally the solution is diluted to 800 c.c. in a large beaker, 125 c.c. of acetic acid, sp. gr. 1.04, and 20 gms. of ammonium acetate are added and the solution is boiled five minutes. Just before boiling begins, 25 c.c. more of sulphurous acid solution are added. When the precipitation is complete, the beaker is allowed to stand in a warm place for at least half an hour, and is then filtered through a 9-cm. filter paper. A siphon is meanwhile prepared which is long enough to extend to the bottom of the beaker, with the glass tubing bent upward so that the water will suck into the siphdh from above without drawing much precipitate with it. The longer end of the siphon is connected by means of rubber tubing with a short piece of glass tubing which rests justs below the upper edge of a 9-cm. washed filter paper. When the precipitated metatitanic acid has settled to the bottom of the beaker, the supernatant solution is siphoned through the filter, regulat- ing the flow by means of a small screw clamp placed on 8j APPENDIX I. the short piece of rubber tubing, a little way above the- funnel. The precipitate is washed with 5 per cent, acetic acid until most of the sulphate has been removed. The filter and its contents are ignited at as low a temperature as possible and the fusion with potassium pyrosulphate is repeated. This time there should be no difficulty in getting a good fusion. The product of the fusion is treated exactly as before, but the residue is now washed more carefully, until all the sulphate has been removed. The filtered precipitate is ignited and weighed as TiC>2. Determination of Sulphur in Pyrite: Fusion with Sodium Peroxide. The Fresenius method (page 357) has long been accepted as a standard method for determining sulphur in insoluble sul- phides, but the procedure is long on account of the necessity of removing all the nitrate before precipitating the sulphuric acid ion. If the analysis is carried out in a platinum crucible the oxidizing mixture attacks the platinum quite badly, which is a serious objection, especially when the analysis is frequently made. If the Fresenius method is carried out in a nickel or iron crucible, it is difficult to raise the temperature sufficiently without using the blast lamp. It has been repeatedly shown* that equally satisfactory results can be obtained by using sodium peroxide as the oxidizing flux. Moreover, the analysis can be accomplished in about two hours. Procedure. About 0.5 gm. of pyrite is mixed with 5 gms. of pure sodium peroxide and 4 gms. of sodium carbonate in a nickel or iron crucible. An opening is cut in a piece of asbestos board (at least four inches square) sufficiently large to allow two-thirds of the crucible to project below the asbestos. The purpose of this shield is to keep the products formed by the combustion *W. Hempel, Z. anorg. Chem., 3, 193 (1893); J. Clark, J. Cher,i. Soc., 63, 1079 (1893); Hohnel, Arch. Pharm., 232, 222; C. Glaser, Chem.-Ztg., 18, 1448; Fournier, Revue generate de chimie, pure et appliquee, 1903, 77; List, Z. angew. Chem., 1903, 414. DETERMINATION OF SULPHUR IN PYRITE. 871 of the gas from reaching the mouth of the crucible. Illuminating gas often contains a little sulphur. The fusion mixture is heated gently for ten minutes so that the mass softens and bakes together and then the temperature is raised until the crucible is exposed to the full heat of the Tirrill burner for twenty minutes. The contents of the crucible are allowed to cool and are treated with 150 c.c. of hot water. When the sodium salts are all dissolved, the crucible is removed and the solution is treated with 5 c.c. of hydrochloric acid, sp. gr. 1.2, to which liquid bromine has been added till the acid iss aturated. The pur- pose of the bromine is to make sure that the oxidation of the sulphur is complete.* It is necessary to add acid as otherwise the hot, caustic soda solution is likely to destroy the filter paper. After heating to boiling, the solution is filtered and the residue of ferric hydroxide is washed free from sulphate. The nitrate is carefully neutralized with 6-normal hydro- chloric acid, sp. gr. 1.12, and 2 c.c. of this acid are added in excess. The solution is heated till all the bromine is expelled, diluted to 250 c.c. and the sulphuric acid precipitated by adding very slowly, while stirring, a slight excess of hot barium chloride solution (10 c.c. of normal barium chloride solution diluted to 100 c.c.). The barium sulphate precipitate is washed three times by decantation with hot water, then transferred to the filter and washed free from chloride, dried, ignited and weighed. If, in the above analysis, the aqueous extract of the sodium peroxide fusion shows manganate or permanganate by the color, these substances may be reduced by the addition of a few drops of alcohol and boiling. Often the reduction of the manganate to manganese dioxide causes a turbidity in the fil- trate. This does no harm, as it dissolves without difficulty on making the solution acid. * A black residue may denote ferrous sulphide or nickelic oxide. It may be tested for sulphur by dissolving in hydrochloric acid and bromine and adding barium chloride to the diluted solution. 872 APPENDIX I. A Rapid Method for the Determination of Sulphur hi Steel. Principle. The method depends on the evolution of the sulphur as hydrogen sulphide and its absorption in an ammoniacal cadmium chloride solution, followed by an iodometric titration. As already stated,* it has been shown that in many cases some of the sulphur in iron and steel is not evolved as hydrogen sul- phide upon treatment of the sample with hydrochloric acid. This fact has received much attention in the literature and various attempts have been made to overcome the difficulty. Thus T. G. Elliot, f mixes 5 gms. of the iron or steel with 0.25 gm. of anhydrous potassium ferrocyanide, wraps the mixture in filter paper, places it in a covered porcelain crucible, and anneals for twenty minutes in a muffle furnace heated to 750-850. The sample is allowed to cool slowly, is broken up in a mortar and then introduced into the evolution flask. The fact that this annealing of the sample serves to convert practically all the sulphur into a form in which it is readily evolved as hydrogen sulphide has been confirmed by other chemists. It has been shown, { however, that if concentrated hydro- chloric acid is used for dissolving the sample, and the sample is dissolved quickly without letting air enter the evolution flask, there is rarely any difficulty in obtaining accurate values by the rapid evolution process which is to be described. Besides hydrogen and hydrogen sulphide, various other gases are evolved to some extent when a sample of iron or steel is dissolved in hydrochloric acid. To prevent the influence of these gases upon the volumetric determination of the sulphur one of the most satisfactory expedients is to absorb the hydrogen sulphide in an ammoniacal solution of cadmium chloride, to filter off the resulting precipitate of cadmium sulphide, to dissolve * Cf. page 352. fChem. News, 104, 298 (1911). }Cf. C. Reinhardt, Stahl u. Eisen, 10, 430 (1890); W. Schindler, Z. angew. Chem., 1893, 11; W. Schulte, Stahl u. Eisen, 26, 985 (1906) and H. Kinder, ibid., 28, 249 (1908). DETERMINATION OF SULPHUR IN STEEL. 8^3 it in acid in the presence of iodine, and to titrate the excess of the iodine with sodium thiosulphate solution. The reactions involved are expressed by the following equations : MnS + 2HC1 = MnCl 2 + H 2 S H 2 S + Cd(NH 3 ) 4 Cl 2 = CdS + 2NH 4 C1 + 2NH 3 2KMnO 4 + 10KI + 8H 2 SO 4 = 6K 2 SO 4 + 2MnSO 4 + 8H 2 + 5I 2 CdS + H 2 S0 4 + 1 2 = CdSO 4 -f 2HI + S I 2 +2Na 2 S 2 O 3 = Na 2 S 4 O 6 +2NaI. Requisite Solutions. 1. Sodium Thiosulphate Solution, approximately 0.02 normal. This is prepared by dissolving 5 gms. of sodium thiosulphate crystals in one liter of freshly boiled water. 2. Permanganate Solution, approximately 0.08 normal. It is prepared by dissolving 5 gms. of potassium permanganate in 2 liters of distilled water, allowing the solution to stand some time, and then filtering through asbestos. 3. Potassium Iodide Solution. This is prepared by dissolving 30 gms. of pure potassium iodide and 10 gms. of sodium bicar- bonate in a little water, filtering the solution if necessary and diluting to 1 liter. 4. Soluble Starch Indicator Solution. About 0.5 gm. of soluble starch is dissolved in 25 c.c. of boiling water. This solution is more stable than solutions of ordinary potato starch, and it is much easier to prepare. The solution is ready for use as soon as it is cold. It usually keeps a week or longer. 5. Ammoniacal Cadmium Chloride Solution. 20 gms. of cadmium chloride are dissolved in 400 c.c. of water, and 600 c.c. of ammonia, sp. gr. 0.96, are added. The solution of potassium permanganate is standardized against sodium oxalate (p. 597) using about 0.2 gm. of the salt. The solution of sodium thiosulphate is standardized by taking 20 c.c of the potassium iodide solution, adding 25 c.c. 874 APPENDIX I. of sulphuric acid (1:3) and allowing exactly 10 c.c. of the per- manganate to run in while stirring. The liberated iodine is then titrated with sodium thiosulphate, adding 2 c.c. of the starch solution toward the last. Apparatus. A simple form of apparatus consists of a 250-c.c. round-bottomed flask connected in series with three 150-c.c. Erlenmeyer flasks. Each of the flasks is fitted with two-holed rubber stoppers. The second hole in the stopper of the evolu- tion flask carries a dropping funnel, the end of which reaches nearly to the bottom of the flask. Each delivery tube is bent twice at right angles, starts just below the lower bottom of the rubber stopper in one flask and leads nearly to the bottom of the next. In the first Erlenmeyer flask is placed 50 c.c. of water to remove most of the hydrochloric acid which is distilled over. It does not absorb an appreciable amount of the hydrogen sul- phide gas because it becomes heated nearly to boiling toward the last. The second Erlenmeyer contains 50 c.c. of the cad- mium solution and the third 25 c.c. of the same. The hydrogen sulphide is usually absorbed completely in the second flask, but the third flask with its cadmium solution is added as a precaution. Procedure. About 5 gms. of the iron or steel is weighed into the evolution flask. The apparatus is connected together and it is made sure that all the joints are tight. 50 c.c. of concen- trated hydrochloric acid, sp. gr. 1.19, are placed in the dropping funnel and about half of it is allowed to flow into the flask. If the action is not too vigorous, the remainder of the acid is at once added, but the stop-cock in the dropping funnel is closed while the tube is still full of acid. This is necessary, because if air enters the solution there is danger of some loss of sulphur owing to oxidation. The evolved hydrogen serves to prevent this oxidation, provided air does not enter through the funnel. The dropping funnel is again filled with 50 c.c. of HC1, and all of this, except enough to fill the stem of the funnel, is added. As soon as the rate of flow of the gas through the Erlenmeyer flasks becomes less than about 3 bubbles per second, a very low flame is placed below the flask and the flow of gas is maintained steadily at this rate. The burner should be furnished with a DETERMINATION OF SULPHUR IN STEEL. 875 flame protector, as otherwise a sudden draft will cause trouble with the small flame. The steel sample should dissolve in about thirty minutes, but it is not advisable to heat too quickly as this will result in too much acid being carried over into the first flask. At first the flame beneath the evolution flask should be only about 0.7 cm. high and it should not be much over 4 cm. high until all the iron has dissolved. When this has taken place, the height of the flame is raised to about 7 cm. and the solution is boiled gently for ten minutes. When the solution begins to boil, the stop-cock in the separatory dropping funnel may be opened. This boiling serves to expel all the hydrogen sulphide from the evolution flask and from the first Erlenmenyer flask. It is unnecessary to pass hydrogen or carbon dioxide through the apparatus, as has been advocated by some chemists. Finally, making sure that the stop-cock in the dropping flask is open, the flame is removed and the apparatus promptly dis- connected. The precipitated cadmium sulphide is filtered off and washed once or twice with water. The filter and precipitate are placed in a 250-c.c. Erlenmeyer flask containing 20 c.c. of the potas- sium iodide solution, 25 c.c. of sulphuric acid (1 : 3) and 10 c.c. of the standardized permanganate solution. When all the cadmium sulphide has dissolved, the excess of iodine is titrated with the sodium thiosulphate solution. Computation. If 1 c.c. of the permanganate solution = a gm. of pure sodium oxalate, TI c.c. of thiosulphate solution were used in titrating 10 c.c. of permanganate, and T% c.c. for titrating the same amount of permanganate used in the analysis of a sample of iron or steel weighing s gms., then the per cent, of sulphur is given by the equation of: S (Pi -!T 2 ) X 100 _ 23.930(7*1 -T 2 ) _ X Na 2 C 2 4 X s s Remark. Massenez * has recommended the direct titration of the cadmium sulphide precipitate with potassium perman- * Stahl u. Eisen, 32, 2089 (1912). 8; 6 APPENDIX I ganate. The cadmium solution, without filtering off the pre- cipitate, is boiled thirty minutes to expel hydrocarbons and is then rinsed into a beaker containing 600 c.c. of water and enough permanganate to give a pink tint. Twenty-five c.c. of sul- phuric acid (1:1) are added and the liberated hydrogen sulphide is titrated with permanganate. The Determination of Chromium in Steel. A small quantity of chromium is a common constituent of iron and steel and in some of the alloy steels it is present to the extent of several per cent. The simplest and most rapid method for determining this element consists in oxidizing it to chromate, adding a known quantity of ferrous salt and deter- mining the excess of the latter by titrating with standard potas- sium permanganate. A great many different methods have been proposed for carrying out the oxidation, two which are very rapid and give accurate results will be described. Oxidation by the Barba Method.* Principle. The steel is dissolved in dilute sulphuric acid, the iron is oxidized to the ferric state by means of nitric acid, and the chromium is oxidized by the addition of strong per- manganate solution. The excess of the latter is destroyed by boiling in ammoniacal solution, the solution is acidified again, the precipitated manganese dioxide filtered off, a known volume of standard ferrous sulphate solution is added to an aliquot part of the filtrate, and the excess of the latter is titrated with permanganate. Procedure. Exactly 1.25 gm. of steel is dissolved in 20 c.c. of 6-normal sulphuric acid, sp. gr. 1.2, and when the solution is com- plete, nitric acid is added drop by drop until the iron is all oxi- dized to the ferric state, five cubic centimeters of nitric acid, sp. gr. 1.2, is sufficient in most cases. The solution is boiled to remove * J. Iron and Steel Institute, 1893, ii, 536; The Iron Age, 52, 153. DETERMINATION OF CHROMIUM IN STEEL. 877 nitrous fumes, diluted to 150 c.c. and treated with-5 c.c. of a sat- urated solution of potassium permanganate (6 gms. per 100 c.c.). The solution is boiled briskly for 15 to 20 minutes. It is then removed from the hot plate, the sides of the beaker are washed down with water and 25 c.c. of strong ammonia are poured down the sides of the beaker. The liquid is stirred vigorously and then placed on the cooler part of the hot plate, for if it is heated too rapidly there is likely to be loss by " bumping.'' The diges- tion is continued with occasional stirring for fifteen minutes, or until the permanganate is all decomposed as shown by the disappearance of the pink color. Then 20 c.c. of dilute sul- phuric acid (1:1) are added cautiously and the solution is heated to boiling. It is transferred to a 250-c.c. calibrated flask, cooled to room temperature, and diluted up to the mark. After thoroughly mixing by pouring back and forth several times into a dry beaker, the solution is filtered and exactly 200 c.c. (corresponding to 1 gm. of metal) is taken for the rest of the analysis. Fifty c.c. of standard ferrous sulphate solution* is added, and the excess titrated with 0.08-normal perman- ganate. Computation, When the ferrous sulphate solution is added, the chromium is reduced from chromic acid to chromic salt in accordance with the following equation : 2CrO 3 4- 6FeSO 4 + 6H 2 S0 4 = O 2 (SO 4 ) 3 + 3Fe 2 (SO 4 ) 3 + 6H 2 O. The ferrous sulphate solution is not very stable and should be titrated against the permanganate at the same time the analysis is made. If 50 c.c. of the ferrous sulphate solution are equiv- alent to TI c.c. of permanganate, of which 1 c.c. = a gm. of pure sodium oxalate, and T 2 c.c. of permanganate were used in titrating the excess of ferrous sulphate in the analysis of 1 gm. of steel, then 2Cr.a(7 7 i-7 7 2 ).100 3Na 2 C 2 O, = 25.86a(7 7 1 -7 7 2 ) = % Cr. * Page 617, footnote. This is sufficient in case the steel does not con- tain more than 2.5 per cent, of chromium. If more chromium than this is present, a smaller aliquot part of the solutions should be taken. 878 APPENDIX I. Oxidation by Sodium Bismuthate. 1000 c.c. of tenth-normal permanganate =1.733 gms. Cr Principle. Sodium bismuthate, NaBiOa, has already been used for oxidizing manganese from the bivalent to septavalent condition.* This oxidation takes place best in a cold solution containing 20 to 40 per cent, of Citric acid (free from nitrous acid, in a volume of 50 to 100 c.c.f Moreover, an excess of sodium bismuthate should be present and the solution must not stand long. Under these conditions scarcely any chromium is oxidized so that the manganese may be determined by the bismuthate method even when chromium is present. If the solution is heated to boiling, the permanganate is decomposed and the manganese is precipitated as manganese dioxide. Chro- mium, on the other hand, is oxidized in hot solutions from the trivalent to the sexavalent condition and the chromic acid is not decomposed by boiling, unless some reducing agent is present. In hot solutions, therefore, it is possible to oxidize the chromium by sodium bismuthate, to filter off the precipitated manganese dioxide and to determine the chromium in the filtrate by adding a known volume of standard ferrous sulphate solution and titrating the excess of the latter with permanganate. Procedure. One gram of the steel is dissolved in a 250-c.c. Erlenmeyer flask with 100 c.c. of 30 per cent, nitric acid (30 c.c. of concentrated acid and 70 c.c. water). If there is any carbonaceous residue, as in the analysis of cast irons, it should be filtered off and examined for chromium by fusion with an alkaline oxidizing flux.{ If the metal is difficultly soluble in nitric acid, it is sometimes necessary to add sulphuric acid to hasten the solution. When the sample is all dissolved, the solution is cooled to between 65 and 75 and 2 gms. of sodium bismuthate are added. The contents of the flask are agitated for a few minutes * Cf. p. 617 of this book. t William Blum, J. Am. Chem. Soc., 34, 1395. t Cf . p. 675. If the carbonaceous residue is not removed it will inter- ere with the oxidation of the chromium. DETERMINATION OF CHROMIUM IN STEEL 879 and then the sides are washed down with a little water. The solution is heated and gently boiled until the permanganate formed from the manganese in the steel is all decomposed as shown by the color. This usually requires about fifteen minutes. . Fifty c.c. of nitric acid (3 per cent, by volume) are added and any precipitated manganese dioxide or undissolved sodium bis- muthate is filtered off on an asbestos filter. The residue is washed three times with 50 c.c. portions of the dilute nitric acid and then diluted with distilled water to 500 c.c. A measured excess of ferrous sulphate solution is added and the excess titrated with permanganate. The ferrous sulphate is titrated against the permanganate on the same day that the analysis is made and the same quan- tities of ferrous sulphate and nitric acid are used as in the analysis. The computation is the same as in the previous analysis. The chemical reactions involved are the following: 2Cr + 3H 2 SO 4 = O 2 (SO 4 ) 3 + 3H 2 Cr 2 (SO 4 ) 3 + 3NaBi0 3 + 6HNO 3 = 2CrO 3 + Bi 2 (SO 4 ) 3 + Bi(N0 3 ) 3 + 3NaNO 3 +3H 2 O 2CrO 3 + 6FeS0 4 + 12HN0 3 = 2Fe 2 (SO 4 ) 3 + 2Fe(NO 3 ) 3 + 2Cr(NO 3 ) 3 + 6H 2 O KMnO 4 + 5FeSO 4 + 9HN0 3 = Fe 2 (SO 4 ) 3 + 3Fe (NO 3 ) 3 + + KHSO 4 + MnSO 4 + 4H 2 O. 2 5 10 15 20 25 30 35 40 50 60 70 2 5 10 15 20 25 30 35 40 50 60 70 2 5 10 15 20 25 30 35 40 50 60 70 ARSENIC STANDARDS. See pages 211 and 212. INDEX OF AUTHORS. A. PAGE Alefeld, E., Determination of chlorine 708 Allner, Determination of carbon monoxide 762 Andrew, L. W., Use of potassium iodate 672 Andrews, Volumetric estimation of nickel 720 sulphuric acid 716 Anneler, Determination of ozone 677 Arndt, Combustion of nitric oxide 803 Arthur, Volumetric estimation of nickel 720 Augenot, Separation of tungsten from tin 300 Austin, M., Determination of magnesium 66 manganese 120, 126 zinc . 140 B. Bamber, Method for determining sulphur in iron and steel 354 Barnebey, O. L., Determination of titanium 118 Baubigny, Determination of antimony 222 Beckett, E. G 223, 224 Beilstein, Electrolytic determination of cadmium 189 Belhoubek, Volumetric determination of uranium 621 Benz, Determination of thorium 510 Bergmann, Electrolysis of nickel solutions 131 Screen for reading burettes 529 Berthelot, Absorption of benzene 753 Distribution of iodine between two solvents 658 Berzelius, Alkalies in silicates 499 Mercurous tungstate precipitation 289 Silicic acid precipitation 473, 489 Vanadium determination 304 Biltz, W., Separation of halides from sulphides 329 Black, Colorimetric determination of arsenic 2Q 881 882 INDEX OF AUTHORS. PAGE Blair, A, A,, Condition of sulphur in iron and steel 352 Direct combustion of steel 413 .Manganese in steel, iron, ores, etc 616 Phosphorus in iron and steel 637 Vanadium, molybdenum, chromium, and nickel in steel. . . 313 Blasdale, W. C., Separation of calcium and magnesium 77 Bloxam, Electrolytic determination of arsenic 212 Bockmann, Sulphur in insoluble sulphides w 358 Bong, G., Decomposition of silicates 490 Borchers, W., Sulphocyanic acid 341 and hydrocyanic acids 342 Borda, Method of weighing 9 Boreli, Electrolytic determination of mercury 173 Bormann, Expelling ammonia with magnesium oxide 59 Borntrager, H., Determination of tungsten 300 Recovery of molybdenum residues 447 Boudet, Error in measuring alcoholic soap solutions 536 Boutron, Error in measuring alcoholic soap solutions 536 Braun, Effect of carbon disulphide on antimony sulphide 242 Brauner, B., Separation of selenium and tellurium 279, 280 Brearley, Volumetric estimation of nickel 720 Bretschger, M., Density of acetylene 754 ethylene 751 Pipette for gas analysis 796 Preparation of acetylene . . 754 Brodie, B. C., Ozone 676 Bruhns, Volumetric estimation of sulphuric acid 716 Brunck, O., Combustion of hydrogen 772 Determination of antimony 222 nickel 129 Separation of iron and manganese 153 nickel 166 nickel and cobalt 162 zinc 165 Volumetric estimation of hydrogen sulphide 688 Brunner, Analysis of tungsten bronze 299 Determination of arsenic 205 Separation of zinc from nickel, cobalt, and manganese 158 Brush, Water in silicates 512 Bunsen, Alkalimeter 376 Analysis of chromite 509 Combustion of nitrogen 807 Determination of antimony arsenic 205 Igniting precipitates 22 INDEX OF AUTHORS. 883 EAGE Bunsen, lodimetric analysis of peroxides 661 Separation of arsenic and antimony 241 Solubility of nitrous oxide 801 Type of burette 511 Valve 87, 98, 6C1 Volumetric estimation of bromides 659 sulphurous acid 692 Bunte, Apparatus for gas analysis 798 Burgstaller, A., Determination of chlorine 708 Busch, M., Determination of nitric acid 451 Busvold, N., Analysis of gases rich in chlorine 810 Density of hydrogen chloride 814 Experiments with chlorine , 812 C. Campagne, Chromium in steel 315 Campbell, Absorption of hydrogen 771 Determination of nickel 720 Direct combustion of steel 413 Carius, Method for decomposing organic substances 325, 370 Carnot, A., Determination of lithium 56 Cavendish, Rare gases in commercial nitrogen 807 Chancel, Determination of aluminium 83 Separation of iron and titanium 115 Chapin, W. H., Boric acid in silicates 590 Christensen, A., Metallic iron in the presence of oxide 612 Christie, W. A. K., Computations in gas analysis 782 Density of carbon dioxide 750 chlorine 808 Standardization of permanganate 98 Clarke, F. W., Electrolytic determination of mercury 172 Separation of antimony and tin 248 Classen, A., Determination of antimony 224, 226, 227 carbon dioxide 380 zinc 147 Electrolytic apparatus 134 determination of mercury 172 tin 234 Standardization of permanganate 92, 600 Comment, Colorimetric determination of arsenic 208 Contat, Valve 98, 602 Cooke, J. P., Ferrous iron in silicates 503 Corleis, Carbon in iron and steel 399 Cormimboef , H., Nickel determination 138 884 INDEX OF AUTHORS. D. PAGE Dakin, H. D., Determination of zinc 140 Daniel, D., Determination of iron 89 Davila, Density of nitric oxide 802 Deckert, Apparatus for chlorine determination 813 de Hae'n, Titration of ferrocyanic acid 632 de Koninck, Determination of ferrous iron 504 nitric acid . 460 Dennis, Technical gas analysis 794 Dennstedt, M., Combustion in open tube 421 Deussen, E., Fusion with potassium acid fluoride 109 Devada, Determination of nitric acid 454 Deventer, Preparation of nitric oxide 802 Deville, St. Claire, Analysis of commercial platinum 272 Water- jacketed tube 732 Dexter, Determination of antimony 224 Diefenthaler, O., Volumetric estimation of ferricyanic acid 695 Diehl, Volumetric analysis of lead peroxide 675 Diethelm, Burette float 529 Dietz, H., Analysis of iodides 671 Dittmar, Determination of potassium 45, 48 Ditz, H., Analysis of chlorates 669 Divers, E., Absorption of nitric oxide 803 Donath, Determination of ammonia 824 Separation of tungsten and tin 301 Dormaar, O. M. M., Determination of antimony 227 Drechsel, G., Volumetric estimation of chlorine 707 Drehschmidt, Analysis of gas mixtures 784 Apparatus for gas analysis 785 Pipette 743, 766 Platinum capillary 743, 766 Drouin, Qualitative detection of carbon monoxide 769 Drown, T. M., Determination of silicon in iron and steel 442 Dudley, Determination of phosphorus in bronze 239 Dulin, Determination of copper 724 Dumas, Method for determining nitrogen 422 Dupasquier, Volumetric estimation of sulphurous acid 692 Dupre, Determination of potassium 44, 45, 46, 48 Durkes, Benzidine hydrochloride 714 Duschak, Determination of sulphuric acid 465, 469 E. Eddy, Determination of magnesium 69 Eggertz, Precipitation of ammonium phosphomolybdate 436 Emich, Determination of nitric oxide 802 INDEX OF AUTHORS. 885 PAQB Engel, Electrolytic determination of tin 235 Engels, Determination of potassium 48 Engler, Theory of oxidation 605 Erdmann, Determination of mercury 172 Euler, Separation of vanadic and molybdic acids 666, 667 F. Fairbanks, Volumetric estimation of molybdenum 667, 668 Feldhaus, Gravimetric estimation of cyanogen 337 Ferchland, P., Determination of chlorine 812 Finch, G., Analysis of fuming sulphuric acid 577 Finkener, Determination of antimony 224 ferrocyanic acid 343 phosphoric acid 439 Precipitation of potassium 47 Volumetric estimation of sulphurous acid 692 Fischer, A., Determination of antimony .* 225, 227 E., Separation of arsenic and antimony 245 N. W., Separation of nickel and cobalt 162 Fitzenkam, R., Determination of potassium 51 Folin, O., Determination of sulphuric acid 469 Follenius, Determination of cadmium 191 Fones, Determination of boric acid , 428 Fordos, Volumetric estimation of sulphurous acid 692 Forster, F., Analysis of ferrum reductum 612 Electrolytic deposition of copper 187 Determination of antimony 227 Testing of commercial platinum 276 Franzen, Use of hydrosulphite . . . . 760 Fresenius, Alkalimeter 376 Determination of carbon dioxide 380 potassium 45 sulphur in sulphides 357 Drying precipitates 21 Electrolysis of nickel solutions 131 Separation of barium, calcium, and strontium 79 and strontium 80 bismuth and copper 198 calcium and magnesium 76 copper and cadmium 202 Solubility of barium chromate 75 strontium carbonate 73 sulphate 73 Volumetric estimation of ferric iron 697 886 INDEX OF ALTHORS. Fresenius, Volumetric estimation of iodides 657 nitrates 634 pyrolusite 625 Friedberger, O., lodates and periodates 670 Friedheim, Determination of sulphuric acid 714 vanadium 666, 667 Separation of arsenic and antimony 245 vanadic and molybdic acids 666, 667 Friend, J. A., Titration with permanganate 604, 607 Funk, W., Separation of iron and molybdenum 153 G. Gaihlot, Determination of nitrous and nitric acids 826 Gay-Lussac, Volumetric estimation of silver 702 Geissler, Alkalimeter 376 Bulb for absorbing gases 416 Gelis, Volumetric estimation of sulphurous acid 692 Gerlinger, Determination of nitrous and nitric acids 826 Gerster, M., Preparation of fuming sulphuric acid mixtures . . .580 Gibbs, W., Electrolytic determination of copper 187 nick< 1 131 Determination of magnesium 68 manganese 126 Gilbert, Determination of silica 487 vanadium 304 Gladding, Modification of Kjeldahl's method 63 Gladstone, Density of methane 774 Preparation of ethylene 751 Glaser, Determination of sulphur in sulphides 358 Indicators 548 Gockel, Screen for reading burettes 529 Valve 98, 602 Gooch, F. A., Crucible for asbestos filters 24, 25 Determination of boric acid 428 magnesium 66, 69 manganese 120, 126 phosphoric acid 434 vanadium , . . 304 Separation of aluminium and titanium 116 iodine and chlorine 331 lithium from sodium and potassium 53 Volumetric estimation of copper 682 molybdenum 667, 668 Gott, Direct combustion of steel 413 Goutal, Determination of nickel 720 INDEX OF AUTHORS. 887 PAGE Grandeau, Determination of nitric acid 456 Gray, Density of nitric oxide 802 Gressly, A., Separation of vanadic and phosphoric acids 307 Griess, Peter, Colorimetric determination of nitrous acid 344 Grimm, Error in measuring instruments 536 Meniscus corrections 744 Groger, M., Determination of sulphur in sulphides 367 Grossmann, Determination of nickel 720 Grund, R., Magnesite cupels 259 Gutbier, Determination of nitric acid 451 tellurium 279 Separation of tellurium and antimony 281 Guye, P. A., Density of acetylene , 754 ethylene 751 nitric oxide 802 Preparation of acetylene 755 H. Haber, Absorption of oxygen 759 Combustion of carbon monoxide ,., . 766 Computations in gas analysis 782 Determination of benzene 753 ethylene. . . 752, 818 Separation of benzene and ethylene 757 Hackford, E., Electrolytic determination of arsenic 212 Hae'n, Volumetric estimation of copper 682 Haidlen, Separation of bismuth and copper 198 copper and cadmium 202 Hampe, Determination of copper 185 manganese in steel 619 Separation of antimony, arsenic, and tin 257 antimony and tin 252 arsenic and tin 255 Handy, J. O., Phosphorus in steel -. 588 Harbeck, Separation of benzene and acetylene 756 Harding, Hydrogen sulphide from insoluble sulphides 368 Hart, Absorption of hydrogen 771 Hauer, Determination of vanadium 304 Heath, Volumetric estimation of copper 682 Heberlein, Determination of sulphocyanic acid 340, 341 Hefti, F., Colorimetric estimation of arsenic 208 Determination of arsenic as arsine 214 Electrolytic determination of arsenic 212 Results obtained in determining arsenic 218 Helmer, Determination of sulphur in iron and steel 364 888 INDEX OF AUTHORS. X PAGE: Hempel, Absorption of oxygen 759 Analysis of illuminating gas 775 Apparatus for gas analysis 743, 785 Combustion of carbon monoxide 766 Detection of carbon monoxide 768 Determination of carbon in iron and steel 404 fluorine as silicon fluoride 829 Rare gases in commercial nitrogen 807 Solubility of acetylene in acetone 754 Technical gas analysis 786 Henry, Non-explosibility of carbon monoxide and nitric oxide mixtures . . . 804 Henz, Determination of antimony 218, 222, 225, 226, 227 boric acid 432 tin 233,235 Separation of antimony and tin 248 Herold, Analysis of ferrum reductum 612 Hibbert, Volumetric estimation of hydrogen peroxide 700 iron 699 persulphuric acid 701 Hill, A. E., Volumetric estimation of chlorine 707 Hillebrand, W. F., Analysis of silicates 487, 491 Dehydration of silica 493" Determination of chlorine in minerals 324 ferrous iron in silicates 504 iron in silicates 109 sulphur and zirconium 505 titanium 100 uranium 106, 621 vanadium and chromium 310 Precipitation of vanadium 304 Removal of melts from crucibles : 488 Testing silicates for barium 495 Hinder, Decomposition of silicates 501 Hinman, Volumetric estimation of sulphuric acid 716 Hinrichsen, Titration of hydrofluosilicic acid 582 Hintz, Determination of sulphuric acid 469 Hofmann, A. W., Separation of copper and cadmium 200 Hoitsema, Determination of silver 706 Hollard, Determination of antimony 225 Holliger, Volumetric estimation of sulphuric acid .716 Holthof, Determination of copper 186 Holverscheidt, Determination of vanadium 304, 306, 666 Hommel, W., Separation of molybdenum and Tungsten. . . .294 Honig, M., Titration of boric acid . 589 Hopkkis, Permanence of permanganate solution 90 INDEX OF AUTHORS. 889 PAGE Huldschinsky, Separation of cobalt and nickel 165 Hulett, Apparatus for distilling mercury 748 Determination of sulphuric acid 465, 469 Hundeshagen, Precipitation of ammonium phosphomolybdate 436 I Ilinsky, Separation of nickel and cobalt 165 Ilosvay von Nagy Ilosva, Determination of Acetylene 755 Nitrous acid 345 Inglis, Determination of ozone 677 Inhelder, A., Determination of antimony 227 Preparation of sodium sulphide reagent 225 Isham, R. M., Determination of titanium 118 Isler, Specific gravity of strong acids 838, 839 J Jager, Oxidation of gases 797 Jamieson, Volumetric estimation of copper 672 Janini, Determination of zinc 146 Jannasch, P., Decomposition of silicates 490 Determination of alkalies in silicates 501 chlorine in apatite 323 water in fluosilicates 484 silicates 512 Separation of bismuth from lead : 195 iodine from chlorine 332 selenium and tellurium 279 Jarvinen, K. K., Determination of phosphoric acid 434 Jarvis, Volumetric estimation of nickel 720 Jawein, Electrolytic determination of cadmium 189 Jeffery, J. H., Titration with permanganate 604, 607 Johnson, Volumetric estimation of nickel 720 Johnston, Determination of sulphur in iron and steel 366 Jones, Reductor 608, 637 Jones, C. C., Titration with permanganate 604, 607 Jones, H., Titration of boric acid 589 Jones, W. A., Carbon monoxide-copper compound 763 Jordis, Dehydration of silicic acid 487 Jorgensen, G., Determination of magnesium 67 phosphoric acid 434 Titration of boric acid 589 Jungfleisch, Distribution of iodine between two solvents 658 K Ivanter, Dehydration of silicic acid . 487 Kassner, Titration of barium peroxide 628 890 INDEX OF AUTHORS. PAGE Keen, W. H., Electric furnace 414 Keller, E., Determination of selenium and tellurium in copper 284 Separation of selenium and tellurium 282 from copper 280, 282 Kempf, R., Titration of persulphates 630 Kerner, Separation of halogens and cyanogen 339 Kessler, Titration with permanganate 604 Kingzett, Determination of hydrogen pergxide 680 Kinnicutt, Oxidation of carbon monoxide 767 Kjeldahl, Method for determining nitrogen 62 Klapproth, Determination of antimony 225 Kling, Determination of potassium 48 Knecht, Volumetric estimation of hydrogen peroxide 700 iron 699 persulphuric acid 701 Knerr, Electrolytic determination of mercury 172 Knop, Determination of ammonia 822 Knorre, Separation of nickel and cobalt 165 Koch, A. A., Determination of fluorine 476, 830 Separation of phosphoric and hydrofluoric acids 474 Kohlrausch, F., Reduction of weights to vacuo 13, 14 Testing of weights 15 Koppeschaar, W., Determination of phenol 695 Korbuly, M., Absorption of benzene 753 Kramers, G. H., Precipitation of zinc sulphide 160 Kraut, K., Determination of nickel 129 Kreider, Preparation of perchloric acid 51 Kreitling, Use of burette floats 529 Krug, Determination of sulphur in iron and steel 365 Kiister, Determination of sulphuric acid 467 Sodium hydroxide solution free from alkali carbonates 555 Titration of alkali carbonates 562, 564 Kuzma, B., Separation of selenium and tellurium from metals 280 L Ladenburg, R., Determination of ozone 677, 680 Lagutt, E., Determination of silver 318 Langbeck, H. W., Ortho-nitrophenol as indicator 543 Leberle, Determination of iron 89 Lecco, Calorimetric determination of iodine 661 Lecrenier, Determination of antimony 225 Leduc, Density of carbon dioxide 750 chlorine 308 hydrogen chloride 814 hydrogen sulphide 348, 816 INDEX OF AUTHORS. 891 PAGE Leduc, Density of sulphur dioxide 815 > Lefort, J., Aqua regia for dissolving sulphides 362 Lenssen, Titration with permanganate 604 Volumetric estimation of ferricyanic acid 694 l.eutold, Combustion of hydrogen 773 Levol, Determination of arsenic 206 manganese in pyrolusite 624 Solubility of magnesium ammonium arsenate 208 Levy, Volumetric estimation of copper 672 Lieben, Volumetric estimation of formic acid 626 Liebig, J., Determination of carbon and hydrogen in organic compounds. 414 sulphur in organic compounds 371 Separation of nickel from cobalt 163, 16i Titration of cyanogen 711 Liechti, Determination of nitric acid , 460 Lindetnann, Absorption of oxygen by phosphorus 759 Lockemann, Determination of arsenic 208 Loose, R., Methyl red 543 Losekann, G., Determination of zinc 140 Low, A. H., Analysis of copper ores 725 Determination of copper bj iodide method 682 potassium cyanide 724 lead 726 Lowe, Separation of bismuth and lead 195 Lowenthal, J., Precipitation of tin 232 Titration with permanganate 604 Luckow, Determination of antimony 224 zinc 147 Electrolytic determination of mercury 1 72 Lunge, G., Analysis of fuming acids 579 pyrolusite 624 Dehydration of silicic acid 487 Determination of carbon dioxide 388, 391 nitrous acid 345, 626 sulphur in pyrite 362 -Rey pipette 575 Separation of ethylene and benzene ^ 756 Soluble and insoluble silicic acid 507 Standardization of permanganate 92 Specific gravities of ammonia solutions 841 strong acids 838 Universal apparatus 387, 823 Luther, Analysis of chlorates 669 Ozone 677 Lux, Analysis cf red lead 623 892 INDEX OF AUTHORS, M PAGE Manchot, Titration with permanganate : 605 Marchand, Determination of mercury 172 Mar chlewski, Determination of carbon dioxide 388 Specific gravities of strong acids 838, 839 Margosches, B. M., Analysis of iodides . . . 671 Margueritte, Volumetric estimation of iron 89, 99, 603 Marquardt, Analysis of ferrum reductum 612 Marshall, M., Colorimetric determination of manganese 128 Martz, E., Oxygen in sea water 740 Mascazzini, Determination of antimony 224 Massaciu, Determination of chromium 103 May, W. C., Drying lead peroxide deposits 178 Mayer, L., Determination of lithium 56 Purification of mercury 747 Mayr, D. K., Determination of antimony 227 McArthur, Determination of potassium 45, 48 McKay, Determination of arsenic 205 Mensel, Volumetric analysis of iodides 656 Merck, Determination of metallic iron in the presence of oxide 611 Merling, Determination of lithium 56 Metzl, Determination of antimony 221 Separation of antimony and tin 248, 250 Meyer, J., Determination of manganese 125 Meyer, V., Nitrous oxide 800 Michaelis, Separation of arsenic and antimony 245 Millberg, Dehydration of silicic acid 487 Separation of soluble and insoluble silicic acid 507 Miller, Notes on assaying 264 Miolati, A., Electrolytic determination of mercury 173 Misteli, W., Preparation of ethylene 751 Mitscherlich, Determination of ferrous iron in silicates 504 Mittash, Separation of iron and manganese 154 Mohr, Definition of liter 521 Titration with pyrolusite 625 Mohr, C., Volumetric estimation of iron 681 ferricyanic acid 694 Mohr, F., Determination of bromine 709 chlorine 708 Preparation of litmus solution 545 Moller, Titration of fluosilicic acid 583 Moore, Volumetric determination of nickel 720 Moore, C. J., Purification of mercury 747 Morse, Permanence of permanganate solutions 90 Morton, Volumetric determination of sulphuric acid 716 INDEX OF AUTHORS. 893 PAOB Miiller, E., Analysis of iodates and periodates 670 Determination of ferricyanic acid 095 Miiller, M., Separation of tellurium and selenium 279 tungsten and tin 301 Miiller, W., Use of benzidine hydrochloride 714 Munroe, C. E., Crucible with platinized felt 27 Muthmann, Separation of tellurium and antimony 281 Mylius, Testing commercial platinum 276 N. Naef , Specific gravities of strong acids 838, 839 Neher, F., Determination of arsenic 205 Separation of arsenic and antimony 243 tin 255 Neubauer, Determination of cyanogen and halogens 339 phosphoric acid 434 Precipitation of potassium 47, 48 Nernst, Molecular volumes , 782 Neuman, Electrolytic separation of copper and cadmium 203 Nicloux, Oxidation of carbon monoxide 767 Noyes, W. A., Determination of sulphur in iron and steel 364 Nydegger, Determination of sulphuric acid with benzidine 714 O. Oberer, Separation of copper and cadmium 010 Oechelhauser, Determination of ethylene 752 Separation of benzene and ethylene 757 Oettel, Determination of fluorine as silicon fluoride 829 phosphorus in bronze 238 Gfferhaus, Determination of chlorine gas by titration 811 Ohm, Law of 133 Oppenheim, C., Absorption of nitric oxide 803- Orsat, Apparatus for gas analysis 797 Osann, Volumetric determination of silver 706 Ost, Determination of antimony 225, 227 Ostwald, Definition of mol . 455 solubility product 156 Washing precipitates 18 P. Palmera, W., Iridium in commercial platinum 275 Panting, Determination of carbon monoxide 762 894 INDEX OF AUTHORS. Parr, Volumetric determination of copper 673 Parrodi, Electrolytic determination of antimony 224 Pattinson, J., Determination of manganese in iron and steel 642 sulphuric acid 467 Paul, Drying oven 33, 34 Heating antimony pentasulphide 220 Pease, Determination of phosphorus in bronze 239 Pechard, Analysis of wolframite . 296 Pelouze, Titration of nitrates 634 Penfield, S. L., Determination of fluorine 476 water in silicates 512 Titration of hydrofluosilicic acid 582 Pennock, Volumetric determination of sulphuric acid 716 Penny, Determination of iron by dichromate 641 stannous chloride 697 Perillon, Separation of tungsten and silicon 302 Pettenkofer, Determination of carbon dioxide in air 593 Pettersson, O., Determination of carbonic acid 384 carbon in steel 405 Pf eiffer, Separation of benzene and ethylene 756, 757 Miilipp, R., Analysis of tungsten bronzes 298 Determination of sulphocyanic acid 340 Colorimetric determination of cadmium 189 Separation of copper and cadmium 203 Phillips, Condition of sulphur in iron and steel 352 Determination of silicon in presence of silicic acid 513 Philosophoff, P., Analysis of chlorine gas 812 Piloty, O., Separation of arsenic and antimony 245 Pincus, Determination of phosphoric acid 718 Poggiale, Titration of pyrolusite 624 Pollak, L., Explosibility of carbon monoxide nitric oxide mixtures 804 Permanence of ammoniacal copper solution 756 Solubility of nitrous oxide 801 Potain, Determination of carbon monoxide in air 769 Preusser, Analysis of iron-tungsten alloys 298 Prince, Analysis of iodides 671 Q. Quasig, Ozone determination 677 R. Rammage, Determination of manganese 616 Rammelsberg, Analysis of wolframite 297 Separation of lithium, sodium and potassium 55 L\DEX OF AUTHORS, 895 PAGE Raschig, Titration of hydroxylamine 631 sulphuric acid by benzidine 714 sulphurous acid 693 Rayleigh, Density of carbon dioxide 386, 750 hydrogen 770 nitrous oxide 800 Recoura, Determination of sulphuric acid 467 Reddrop, Determination of manganese 616 Regnault, Tension of aqueous vapor 842-845 Reich, Determination of nitric acid 453 sulphur dioxide 815 Reinhardt, Titration with permanganate 607, 609 Retgers, K., Determination of manganese 125 Reuter, M., Determination of sulphuric acid 716 Richards, T. W., Precipitation of calcium in presence of magnesium .... 76 Solubility of calcium oxalate 70 Testing weights 15 Ricketts, Notes on assaying 264 Riegler, E., Standardization of permanganate 599 Riess, Determination of antimony 224 Ritter, Determination of nitric acid 460 Rittener, Determination of carbonic acid 391 Rivot, Analysis of iron ores . 88 Determination of copper 186 Separation of copper from cadmium 202 Rohmer, M., Separation of arsenic and antimony 245, 246 Romijn, Titration of formaldehyde 694 hydroxylamine salt 581 Rosanoff, M. A., Determination of chlorine 707 Roscoe, Precipitation of vanadium 304, 305 Rose, H., Determination of ammonium as chloroplatinate 58 bismuth , 181 cyanogen 338 f erricyanide . 343 sulphur in sulphides 359 water in fluosilicates 484 Precipitation of vanadium 304, 306 Reduction of mercuric salts 171 Separation of antimony and tin 250 , arsenic, and tin 256 barium, calcium, and strontium 79 copper and cadmium 202 molybdenum and tungsten 296 phosphoric and hydrofluoric acids 474 Volumetric estimation of sulphurous acid 692 896 INDEX OF AUTHORS. PAGl Rosenbladt, Determination of boric acid 428 Rosenheim, Separation of copper and nickel 165 Rossing, Determination of antimony 222 Separation of antimony and tin 248, 255 Rothe, Solution of ferric chloride in ether 167 Rothmund, Determination of chlorine 708 Ruegenberg, M., Separation of molybdenum and tungsten 293 Ruderff, Electrolytic determination of merury 172 Rupp, Methyl red 543 . Volumetric estimation of sulphurous acid 692 Rutter, Analysis of chlorates 669 S. Sahlbom, Titration of hydrofluosilicic acid 582 Sand, H. G. S., Electrolytic determination of arsenic 212 Sanger, C. R., Colorimetric determination of arsenic 208, 210 Sarnstrom, Determination of carbon in steel 399 Schaff gottsche, Precipitation of magnesium ammonium carbonate 69 Scheen, Determination of antimony 227 Schellbach, Burette 529 Schiff, II., Azotometer 423 Schirm, E., Determination of aluminium 85 chromium 103 Schloetter,. Examination of electrolytic chlorine 812 Schlosser, W., Correction tables 519, 533 ...... Draining of burettes 528 ,,.,.. Error in measuring vessels 536 Meniscus corrections 745 Schlossing,. Separation of potassium and sodium 50 Schmidt, E.^ .Analysis of ferrum reductum 612 .EL,, Separation of barium and strontium 81 Schmitz, B., Method of precipitating magnesium 67 . , phosphoric acid 434 Schneider, Direct combustion of steel 413 Manganese determination 616 Scholler, Determination of chromium 103 Schonbein, Estimation of ozone 676 Schrauth, Determination of chromium 103 Schroder, Separation of tellurium and antimony 281 Schroter, A., Dehydration of silicic acid 487 Schrotter, Alkalimeter 376 Schucht, Titration of hydrofluosilicic acid 583 Schudel, Determination of manganese 120 Schudl, Standardization of permanganate 97 Schulze, Determination of nitric acid 45& INDEX OF AUTHORS. 897 PAGE- Schulze, Titration of pyridine bases 561 Schweitzer, P., Solubility of barium chromate 7 Seeman, L., Precipitation of gold by hydrogen peroxide 258- Separation of gold and platinum 272 Shimer, P. W., Direct combustion of steel 413 Smith, Method for separating zinc from nickel, etc 158 Smith, E. F., Electrolytic determination of cadmium 190 mercury 172 Separation of molybdenum and tungsten 29& Smith, J. L., Determination of alkalies in silicates 49(> Smitt, A., Determination of carbon in steel 405 Snelling, W. O., Use of a Munroe crucible 27 Sonnenschein, Determination of phosphoric acid 436- Sonnstadt, Separation of potassium and sodium 50 Sorensen, Standardization of acids 54& permanganate ' . . 597 Soret, Determination of ozone . 680 Soxhlet, Fat extraction apparatus 236 Spear, E. B., Electrolytic determination of zinc 146 Spitz, G., Titration of boric acid 589 Spitzer, Determination of zinc 145- Stahrfoss, M., Density of acetylene 754 ethylene 751 Preparation of acetylene 755- Stas, Analysis of commercial platinum 272" Steen, O., Separation of lead and bismuth 197 Steffan, Determination of phosphoric acid 439, 440 Volumetric separation of vanadium and molybdenum 667, 668 Stieglitz, J., Theory of indicators 540 Steinbeck, Determination of copper 724 Steiner, O., Determination of chlorine by titration 811 Stillman, T. B., Analysis of incandescent mantles 512 Stock, A., Determination of aluminium 84 chromium 103 Separation of arsenic and antimony 245 Stoffel, M., Determination of lead 179 Stokes, Absorption of benzene 753 ethylene 752 Determination of carbon monoxide 764 ferrous iron 504 Separation of ethylene and benzene 756 Stromayer, Separation of barium, calcium, and strontium 79 iron and titanium 115 Subech, A., Silicic acid in clay 508 Swett, O. D., Use of Munroe crucible 27 898 INDEX OF AUTHORS. T. PAGE Talbot, H. P., Removal of melt from crucible 488 Tamm, H., Determination of manganese 121 zinc 140 Than, C., Standardization of sodium thiosulphate solution 647 The"nard, Ozone 680 Thiel, Determination of sulphuric acid 467 Thiele, J., Apparatus for chlorine determination 813 Determination of arsenic 205 antimony 223 Thomsen, J., Heats of combustion 845 Thomson, W., Electrolytic determination of arsenic 212 Thorpe, T. E., Electrolytic determination of arsenic 212 Tiemann, Determination of nitric acid 456 Tomicek, Determination of arsenic 205 Tootmann, S. R., Determination of arsenic 212 Topf , Volumetric determination of lead peroxide 675 Toth, J., Titration of phenol 697 Treadwell, F. P., Absorption of benzene 753 ethylene 752 Apparatus for chlorine determination 808 Collection of gas samples 733 Colorimetric determination of arsenic ' 208 Computations in gas analysis 782 Density of carbon monoxide 764 chlorine 808 Determination of carbon monoxide 764 ferrous iron in silicates 503 hydrofluoric acid 476 iron 89 nickel 137, 138 ozone 677 Gases from defibrinated blood 740 Separation of ethylene and benzene 756 Titration of hydrofluosilicic acid 582 Treubert, Determination of bismuth 181 Tribe, Density of methane 774 Preparation of ethylene 751 Tromsdorff, Determination of nitrous acid in water 346 Tschugaeff, L., Determination of nickel 129 Separation of nickel and cobalt 161 manganese 165 zinc . . 165 INDEX OF AUTHORS. 899 U. PAGE Ukena, Determination of manganese in steel 619 Urech, W., Determination of bismuth 182 V. Vanino, L., Determination of bismuth 181 Precipitation of gold by hydrogen peroxide 258 Separation of gold and platinum 272 Van Name, R. G., Determination of copper 186 van't Kruys, M. J., "Determination of sulphuric acid 464 Vernon, R. H., Analysis of fuming acids 577 Viogili, J. F., Solubility of mangesium ammonium arsenate 208 Vogel, Detection of carbon monoxide 768 cobalt 165 Voigt, Determination of zinc 140 Volhard, Analysis of sulphocyanate and halogens 342 Determination of manganese as sulphate 120 Precipitation of mercuric sulphide 168 Standardization of sodium thiosulphate 648 Titration of bromine 709 chlorine 707 cyanogen 710 iodine 709 manganese 612 silver 705 sulphurous acid 692 sulphocyanide 712 with permanganate 604 Transformation of chloride into oxide 142 von Girsewald, Electrolytic determination of cadmium 189 Separation of copper and cadmium 202 v. Jiiptner, H., Determination of phosphorus 436 v. Knorre, Combustion of gases using copper oxide 797 nitric oxide 803 Determination of manganese in steel 620 sulphuric acid 714 tungsten 290, 291 v. der Pfordten, O., Absorption of oxygen 760 Precipitation of mercurous tungstate 289 v. Rath, G., Precipitation of mercuric sulphide 194 Vortmann, G., Electrolytic determination of mercury 172 Determination of antimony 221 Removal of sulphur from precipitates 169 Separation of antimony and tin 248, 250 900 INDEX OF AUTHORS. W. PAGE Wade, Determination of carbon monoxide 762 Wagner, Determination of ammonia in ammonium salts 822 Draining of burettes 528 Indicators 548 Standardization of sodium thiosulphate solution 647 Walker, Permanence of permanganate solutions 90 Wallace, Volumetric estimation of iron . . 9 697 Walters, J. H,, Jr., Determination of titanium 101 H. E., Colorimetric determination of manganese 128 Warder, R. B., Behavior of sodium bicarbonate solutions toward phenolphthalein 562 Titration of alkali carbonate and bicarbonate 566 carbonate and hydroxide 564 Washburn, E. W., lodimetric titration of arsenic 650 Weber, H., Determination of sulphuric acid 469 Weber, J., Ignition of calcium sulphate 71 Wegelin, A., Composition of sodium sulphide solution 226 Determination of iron 89 nitric acid 458 Weitnauer, H., Determination of manganese 126 Weller, A., Determination of titanium 100, 504 Titration of antimony 687 Wells, Determination of copper by potassium iodate 672 Wense, Separation of potassium and sodium 50, 51 Wherry, E. T., Determination of boric acid in minerals 590 Whitby, G. S., Solubility of silver chloride 317 Wiernik, Specific gravity of ammonia solutions 841 W T iborgh, J., Determination of carbon in steel 405 sulphur in iron and steel 354 Wilfarth, Modification of Kjeldahl method 63 Will, Alkalimeter 376 Titration of pyrolusite 625 Wilner, Determination of metallic iron in the presence of oxide 611 Windelschmidt, A., Determination of nickel 137, 138 Winkler, C., Absorption of benzene 753 Combustion of hydrogen 772 Detection of carbon monoxide 769 Determination of chlorine by titration 811 Titration of alkali carbonate and bicarbonate 565 hydroxide 563 bicarbonate in the presence of carbonate 592 Use of gauze electrodes 134 Winkler-Dennis, Method of technical gas analysis 794 Winkler, L. W., Absorption coefficient of hydrogen 771 INDEX OF AUTHORS. 901 PAGE Winkler, L. W., Absorption coefficient of methane 774 nitric oxide 803 nitrogen 807 Determination of absorbed oxygen 760 carbon monoxide 762 Solubility of oxygen 757 Winteler, Determination of mercury 172 Titration of hydrofluoric acid 581 Wohl, A., Computations in gas analysis 782 Wohler, Determination of carbon in steel 406 Tungsten bronzes 298 Wolff, O., Determination of antimony 227 Wolfrum, L., Analysis of ferrum reductum 612 Wolter, L., Determination of tungsten in steel 292 Woy, Determination of phosphorus 436 as phosphomolybdic anhydride 440 Method of precipitating ammonium phosphomolybdate 437 Wynkoop, G., Determination of aluminium 85 Y. Young, S. W., Standardization of iodine solution 651 Z. Zawadzki, Precipitation of sulphides 158 Zimmermann, Determination of uranium 106 Precipitation of zinc as sulphide 158 Reduction of ferric salts 609 Titration with permanganate 604, 605, 609 Volumetric determination of uranium 621 INDEX OF SUBJECTS. A. Acceptor, definition of 605 Acetic acid, 371,583 anhydride 583 Acetylene 754 determination in presence of ethylene 821 preparation of 755 Acidimetry 537, 571 Acids, specific gravities of 838-389 volumetric determination of 571, 575, 583 Adsorption 18 Air bath 27, 32 Air, determination of carbon dioxide in 397, 593 Alkali bicarbonates, volumetric determination of 564 in presence of carbonates 565 Alkali carbonates, volumetric determination of 562 in presence of bicarbonates 565 hydroxides 563 Alkalies 38 determination in lepidolite 502 silicates 496. 499 separation from metals of Group III 107, 147 molydebnum 286 (caustic), specific gravities of 840-841 volumetric determination of 558 Alkali hydroxides, volumetric determination of 558 in presence of carbonates . . 563 Alkalimeter 376 Alkalimetry 537, 558 Alkaline earths 70 separation from the alkalies, and magnesium 78 metals of Group III 107, 147 molybdenum 287 one another 79 Alkaline earth bicarbonates, volumetric determination of 568 carbonates, volumetric determination of 566, 567 903 904 INDEX GF SUBJECTS. PAGE Alkaline earth hydroxides 566 salts, titration of 570 Alkali sulphides, analysis of 689 and hydrogen sulphide 691 sulphydrates, titration of 689 Aluminium 82 determination in bronzes 236 silicates 492 separation from alkalies and alkaline earths 107 chromium 114 iron 107 iron and phosphoric acid Ill manganese, nickel, cobalt and zinc 149, 152 metals of Group II 192, 235 titanium 116 uranium 119 Ammonia (reagent), necessity for redistilling 82 preparation free from carbonate 149 specific gravity of 841 titration of 560 Ammonium 57 colorimetric determination 60 gas-volumetric determination 822 in drinking \vater 60 volumetric determination 560 Ammonium molybdate, preparation of the reagent 437, 638 Analysis 1 direct 2 gravimetric 1, 38 indirect 2 volumetric 2 Anions, gravimetric determination of 320 calculation of 737 Antimony 218 determination in bearing metal 252 electrolytic determination 224 removal from electrodes 227 separation from arsenic 241 arsenic and tin 256 mercury, lead, etc 235 metals of Group III ' 235 selenium and tellurium 281 tin 248 volumetric determination 685, 687 Apparent iron value of wire for standardization, 98, 601 INDEX OF SUBJECTS. 905 PAG". Arsenic ^ 205 colqrimetric determination 208 determination as arsine 214 in commercial sulphuric acid 248 mispickel 218 vanadinite 30 electrolytic determination 212 separation from antimony 241 antimony and tin 256 mercury, lead, etc 235 metals of Group III 235 molybdenum 287 selenium and tellurium 281 tin 255 Asbestos niters ; . . 26 Assay for gold and silver 25& Assay ton 268- Atomic weight table 849- Azotometer 423 B. Balance, accuracy of 6 sensitiveness of 7 Baking powders, determination of carbonic acid in 377 Barium 74 detection in calcium precipitates 495 determination in silicates and rocks 496, 507 separation from calcium and strontium 79 magnesium 79 strontium . .. ._..... 80 metals of Group III. , , 107, 147 metals of Group II 192 volumetric determination 566 Barium carbonate method 149 hydroxide, normal solution of , 557 Basic acetate method 152 Bearing metal, analysis of 252 Benzene, determination in gas mixtures 752 separation from ethylene 756 Benzidine hydrochloride, reagent 291, 714, 715 Bicsrbonates, in presence of carbonates 592 volumetric estimation of 591 Bichromate, determination of chromium in 105 Bismarck brown 344 Bismuth. . . 179> 906 INDEX OF SUBJECTS. PACE Bismuth, determination in bearing metal 252 separation from arsenic, antimony, and tin 235 copper and cadmium 198 lead 195 mercury 194 metals of Groups III, IV, and V 192 molybdenum 287 selenium anc^ tellurium 280 Bitter-almond water 337 Blick, the 261 Blood, gases from defibrinated 740 Boric acid 428 in mineral waters 431 silicates and enamels 431 volumetric determination of 588 Brass, analysis of 193 Bromine, determination in mineral waters 660 non-electrolytes 325, 329 soluble bromides 659 gravimetric determination of 329 separation from chlorine 334 chlorine and iodine 336 iodine 335 volumetric determination 655, 660, 709 Bronzes, analysis of 236 Burettes .... 514 allowable error in 530 calibration of 527 draining of 527 floats for 528 reading of 528 C. Cadmium 189 colorimetric determination 189 electrolytic determination 189 precipitation as sulphide 191 separation from arsenic, antimony, and tin 235 copper 200-204 lead 200 mercury 194 metals of Groups III, IV, and V 192 molybdenum 287 selenium and tellurium . . . 280 INDEX OF SUBJECTS. 907 PAG 3 Calcium 70 determination in silicates 494 separation from barium and strontium 79 magnesium 76 metals of Group III 107, 147 metals of Group II 192 volumetric estimation 566, 623 chloride, presence of free lime in 377 precipitates, testing for barium in 495 Calibrated flasks, testing of 524 Calibration of burettes 527 gas-measuring instruments 743 flasks ta 522 pipettes 524 Carbon, determination in iron and steel 398 nitrogenous organic substances 419 organic substances 414 dioxide (see Carbonic acid), determination of 750, 778, 788 in electrolytic chlorine 808 Carbonates, in presence of bicarbonate 592 volumetric estimation of 591 Carbonic acid 375 combined, calculation of 736 determination in air 397, 593 baking powders 377 carbonates 591 chlorine 397, 808 illuminating gas 778, 788 mineral waters 382, 736 presence of bicarbonate 591 cyanic and hydrocyanic acids 371 N 2 O, NO, and N 806 free, calculation of 738 determination of 590 Carbon monoxide, determination of 762, 779, 781, 789 qualitative detection in air 768 Cathodes, cleaning of 136, 227 Cations, calculation of those present in water 737 gravimetric determination of 38 Ceric oxide, iodimetric titration of 665 Cerium, determination in soluble salts 828 Charcoal, testing the reducing power of 265 Chemical factors, table of 850 Chlorates, analysis of 460, 633, 669 Chloric acid, gravimetric determination 460 98 INDEX OF SUBJECTS. PAGE' Chloric acid, in the presence of perchloric acid 463: and hydrochloric acids 463 reduction of 461 volumetric determination of 633, 669 Chlorine, free, determination by titration 809- gravimetric determination 324 volumetric determination 654 gravimetric determination 320 in aqueous solutions 320 commercial tin chloride 321 insoluble chlorides 323 organic substances 325 vanadinite 308 separation from bromine 334 carbon dioxide 397 chlorate and perchlorate 463 cyanogen 339, 71 1 fluorine 482 iodine 331, 335 iodine and bromine 336 volumetric determination 654, 707, 708 Chlorine gas, examination of 808 Chloroplatinates, conversion of chlorides into 43 Chromates, gravimetric analysis of 105 volumetric analysis of 641, 649, 664, 675 Chromic acid. See chromates. compounds 102 Chromite, analysis of 509 1 determination of chromium in 675 Chromium 102 determination in iron ores and rocks 310 chromite 510, 675 pig iron 312 steel 313 separation from alkaline earths and magnesium 107 aluminium 114 iron 113 nickel, cobalt, manganese and zinc 149 volumetric determination of 664, 675 Clay, soluble silicic acid in 508 Closet, drying 24, 25 Cobalt. . 138 electrolytic determination 138 separation from alkaline earths and magnesium 147 manganese 161 INDEX OF SUBJECTS. 909 PAGE Cobalt, separation from metals of Group III 149, 152 metals of Group II 192 nickel 161-164 zinc 156 Coil for heating crucibles by steam 32 Collection and confinement of gases 730 Combustion (elementary analysis) 414 of organic substances containing halogen 421 metal 422 sulphur 422 furnace 415 tube 416 of gases 764 (a) by explosion 765 (6) method of Dreschmidt 766 (c) by method of Winkler-Dennis 794 (d) fractional 766 Cone for holding crucible on water bath 31 Copper 182 analysis of for selenium and tellurium 284 determination in bearing metal 252 brass 193 bronze 236 ores ; . 673, 683, 725 electrolytic determination 187 separation from arsenic, antimony and tin 235 bismuth 198 cadium 200 lead 198 mercury 194 metals of Groups III, IV and V 192 volumetric determination 672, 682, 724 Crucible, Gooch 24, 25 Crystals, size of 36 Cupellation 259 Cyanic acid 371 in the presence of carbonic and hydrocyanic acids 371 Cyanogen 337 in the presence of halogens 339, 711 sulphocyanogen 342, 712 chlorine and sulphocyanogen 713 volumetric determination 710 D. Defibrinated blood, gases from 740 Desiccator 23 INDEX OF SUBJECTS. PAGE Bichromate methods 641 Dimethyl glyoxime, reagent 129 Direct methods of analysis 2 Distribution coefficient 658 Double weighing 9 Drying apparatus 416 closet 24, 25, 33 of precipitates ^. 21 of substances in currents of gases 33, 220 ovens 21, 24, 25, 28, 33, 34 E. Electric furnace 28 Electrodes 93, 134 cleaning of 136, 227 Electrolytic determination of antimony 224 arsenic 212 cadmium 189 cobalt 138 copper 187 lead 177 mercury 172 nickel 131 tin 234 zinc 145 iron, for standardizing permanganate solution 600 preparation of 93 outfit, 132, 178 Elementary analysis 414 Ethylene, determination of 751, 818 separation from acetylene 821 benzene 757, 820 Evaporation of liquids Explosion pipette F Factors, table of chemical Fahlerz, analysis of 359 Ferric chloride, solution in ether iron, volumetric determination of 99, 681, 697, 699 Ferric salts, reduction of solutions of by hydrogen sulphide 99, 112 metals stannous chloride 609 sulphurous acid 607 INDEX OF SUBJECTS. 911 PACK Ferricyanic acid, gravimetric estimation of 344 volumetric estimation of 633, 694 Ferrocyanic acid, gravimetric estimation 342 volumetric estimation 632 Ferrous iron, volumetric determination of 89, 603, 607, 641 Ferrum reductum, analysis of 611, 612 Filters 18 of asbestos 26 size of 20 Filtration and washing precipitates 18 Flasks, calibrated, permissible error in 524 calibration of 522 Floats for burettes 528 Fluorine, determination as calcium fluoride 471 hydrofluosilic acid 476 silicon fluoride 475, 829 in calcium fluoride 472 lepidolite 502 mineral waters 480 presence of phosphoric acid 474 separation from acids 482 metals 481 Formaldehyde (formalin) volumetric determination of 694 Formic acid, volumetric determination of 626 Fractional combustion of gases 766 Fuming acid, analysis of 575 Fuming sulphuric acid 577 Furnace, electric 28 G. Gas analysis 729 exact 775 technical 786 Gas-combustion pipette 792, 794 Gases, collection and confinement of 730 transference of 742 Gas measuring instruments, calibration of 743 pipettes 786 Gas-volumetric methods 822 Gauze platinum electrodes 134 Generator gas, analysis of 775, 783 Gold, determination in solutions 257 ores 263 precipitation by hydrogen peroxide 258 separation from platinum 271 912 INDEX OF SUBJECTS. PAGE Gold, separation from selenium and telluefum 282 silver 262 Gram-equivalent, definition of 530 Graphite in cast iron 414 Gravimetric analysis, methods of 2, 38 H. Halogens, determination by indirect analysis 334 gravimetric determination in presence of cyanide 339 separation from cyanide and sulphocyanide 342 hydrofluoric acid 482 one another 331 Hardness of water 568, 569 Hematite, determination of iron in 610, 698 Heats of combustion of gases 845 Hood 30, 31 Hydriodic acid 330 separation from hydrobromic and hydrochloric acids, 331, 336 hydrocyanic acid 339 volumetric determination 709 Hydrobromic acid 329 determination in mineral waters 660 separation from hydrochloric acid 334 and hydriodic acids 336 hydrocyanic acid 339 hydriodic acid 335 volumetric estimation 655, 659 Hydrocarbons, heavy 751, 779, 788 separation of 756 Hydrochloric acid 320 gas, determination of 814 normal solution of 549 separation from chloric and perchloric acid 463 hydriodic acid 335 hydrobromic acid 334 hydrogen sulphide 329 hydrocyanic acid 339, 711 and sulphocyanic acids . 713 sulphocyanic acid 342, 713 volumetric determination 571, 707, 708 Hydrocyanic acid 337 determination in bitter almond water 337 separation from cyanic and carbonic acids 371 halogen hydride 339, 711 hydrochloric and sulphocyanic acids. . 713 INDEX OF SUBJECTS. 913 PAGE Hydrocyanic acid, separation from sulphocyanic acid 342, 712 volumetric determination of 710, 711 Hydroferricyanic acid 344, 633, 694 Hydroferrocyanic acid 342, 632 Hydrofluoric acid 471 determination as calcium fluoride 471 hydrofluosilic acid 476 silicon fluoride 475, 829 in calcium fluoride 472 lepidolite 502 mineral waters 480 separation from boric acid 483 hydrochloric acid 482 metals. 481 phosphoric acid 474 volumetric determination 581 Hydrofluosilic acid 483 analysis of its salts 484 determination as calcium fluoride 483 potassium silicofluoride 484 volumetric determination 581 Hydrogen 770 determination in nitrogenous organic substances 419 organic substances 414 Hydrogen peroxide, iodimetric titration 680 methods 826 titration by permanganate 626 titanous chloride 700 sulphide, colorimetric determination 354 expulsion from insoluble sulphides 367 evolution and absorption 350 determination in mineral waters 349, 688 gas mixtures 816 gravimetric determination 347 titration of 687 separation from alkali sulphydrate 691 thiosulphate 691 Hydrosulphite of sodium 760 Hydrosulphuric acid (see Hydrogen sulphide) 347 Hydroxylamine 581, 631 Hypochlorous acid 344, 669, 701 determination in the presence of chlorine 655 Hypophosphorous acid 372 separation from phosphorous acid 374 QI4 INDEX OF SUBJECTS. I PAGE Igniting precipitates, method of 21, 28 Illuminating gas, analysis of 775, 786 Incandescent mantles, analysis of 512 Indicators 538 Indirect analysis 2, 56 determination of halogens by 334 SO 3 content of fuming sulphuric acid . 579 error in 6 Inquartation 259 International atomic weights 849 lodic acid 433 determination in the presence of periodates 670 Iodides, analysis of (see Iodine) 671 lodimetry 644 Iodine, determination by gravimetric methods 330 in mineral waters 660 . non-electrolytes 328 soluble iodides 656 free, titration of 654 preparation of pure 646 separation from bromine 335 bromine and chlorine 336 chlorine 331 solution, standardization of 649 volumetric determination 654, 709 lodo-starch reaction, sensitiveness of 652 Iron 87 electrolytic 93, 603 determination in bearing metal 252 brass 193 ferrum reductum 611 hematite 610, 698 presence of oxide 611 silicates 493, 502 separation from alkaline earths and magnesium 107 aluminium 107 and phosphoric acid ' Ill chromium 113 manganese 153 nickel, cobalt and zinc 149, 152, 155 nickel 166 metals of Group II 192 titanium 114 uranium 119 INDEX OF SUBJECTS. 915 PAGE Iron, volumetric determination by permanganate dichromate method 641 iodimetric method 681 permanganate 603, 607 stannous chloride 697 titanous chloride 699 ore, determination of vanadium in 310 and chromium in 312 wire, determination of apparent iron value 98, 601 L. Lacmoid .' 544 Lead 174 determination in bearing metal 252 brass 193 bronze 236 vanadinite 308 electrolytic determination as peroxide ^177 separation from arsenic, antimony, and tin 235 bismuth 195 cadmium 200 copper 198 mercury 194 metals of Groups II, IV, and V 192 molybdenum 287 peroxide, analysis of 675 sulphate, separation from barium sulphate and silica 176 volumetric determination by molybdate . . . 726 Lepidolite, analysis of 502 Lime method for halogens in organic substances 329 Liquids, evaporation of 30 Liter, definition of 516, 521 Litharge, testing of 264 Lithium 53 determination in lepidolite 502 indirect determination of 56 separation from sodium and potassium 53 Litmus 544 Logarithms, tables of 854, 855 M. Magnesia mixture, preparation of 206 Magnesium 65 determination in silicates 495 separation from alkalies 68 barium 79 916 INDEX OF SUBJECTS. Magnesium, separation from calcium ................................ 7fr metals of Group II ...................... 192 Group III ................. 107, 147 strontium .............................. 78 Manganese ...................................................... 120 colorimetric determination .............................. 127 determination in bronze ................................ 237 iron and steel : by bismuthate method .............. 616 Volhard's method ................ 615 v. Knorre's method .............. 620 Pattinson's method .............. 642 Williams' method ............. x . 619 pyrolusite ......................... 624, 663 separation from alkaline earths and magnesium ........... 147 iron ............................... : . . . 153 metals of Group II ...................... 192 nickel and cobalt ........................ 161 trivalent metals ..................... 149, 155 zinc ................................... 156 volumetric determination .............. 612, 616, 619, 620, 663 Measuring flasks ................................................. 515 calibrating ...................................... 522 instruments ..... ...................................... 514 for gas analysis .............................. 743 Melt, removal from the crucible ................................... 488 Meniscus corrections ............................................. 745 Mercury ........................................................ 168 determination in organic substances ........................ 170 electrolytic determination ................................. 172 purification of ........................................... 747 separation from arsenic, antimony and tin .................. 235 lead, bismuth, copper, and cadmium ......... 194 metals of Groups II, IV, and V ............. 192 selenium and tellurium ..................... 281 Metallic iron, determination in presence of oxide ..................... 611 Metalloids, gravimetric determination of ............................ 320 Metaphosphoric acid ............................................. 433 Methane ....................... - 774 determination in gas mixtures .................... 780, 781, 790 separation from hydrogen ................................. 790 carbon monoxide and hydrogen ............ 781 Methods, gravimetric ............................................. 2 volumetric ............................................. 514 Methyl orange ......................................... .......... 539 INDEX OF SUBJECTS. 917 PAGE Methyl red 543 Minium (red lead), volumetric determination of 623, 675 Mispickel, determination of arsenic in 218 Mol, definition of 455 Molybdenum 284, 666 determination in steel 313 residues, recovery of molybdenum from . . . . 447 separation from the alkalies 286 the alkaline earths 287 the metals of Groups II and III 287 phosphoric acid 288 tungsten 293 vanadium 308, 667 Molybdic acid, gravimetric determination 284 volumetric determination 666 Moment, statical 8 Monazite, determination of thorium in 510 Munroe crucible 27 N Nickel 129 determination in brass 193 steel 166, 313, 723 electrolytic determination 131, 136 separation from alkaline earths and magnesium 147 cobalt 161, 162, 163, 164 iron 166 , aluminium, titanium, and uranium 149 , aluminium, chromium, titanium, and uranium 149 manganese 161, 165 zinc 156, 165 volumetric determination 720 Nickel-chromium alloy for crucible triangles 29 Niobium, determination in wolframite 297 Nitre, testing the oxidizing power of 266 Nitric acid 451 determination as ammonia 453 nitric oxide 456, 825 nitrogen pentoxide 453 nitron nitrate 451 in drinking water 460 normal solution of 552 volumetric determination 571, 634 918 INDEX OF SUBJECTS. PAGE Nitric oxide 802 separation from nitrous oxide 804 and nitrogen 805 nitrogen, and carbon dioxide . 806 Nitron 451 Nitrous acid, colorimetric determination 344 determination as nitric oxide 825 volumetric determination . . . . . 626 oxide 800 determination in the presence of nitric oxide 804 and nitrogen . . 805 nitrogen, and carbon dioxide 806 Nitrogen, determination by Dumas method 422 Kjeldahl method 62 in organic substances 422 properties and method of preparation 806 separation from nitrous and nitric oxides 805 oxide and carbon dioxide . . 806 Nitrophenol as indicator 543 Normal solutions 530, 548 of barium hydroxide 557 hydrochloric acid 549 nitric and sulphuric acids 552 oxalic acid 552 sodium hydroxide 553 preparation of 532 standardization in acidimetry and alkalimetry 548 volume and temperature 516 O. Oil, removal from borings 236 Oleum, analysis of 575 Operations 6 Organic acids, titration of 583 substances, determination of carbon in 414, 419 chlorine in 325, 329 hydrogen in 414, 419 nitrogen in 62, 422 sulphur in 370 Orthoclase, analysis of 491 Orthophosphoric acid 434 volumetric estimation of 718 Oven for drying 24, 25, 28, 33, 34, 220 INDEX OF SUBJECTS. 919 PAGE Oxalic acid 427 normal solution of 552 volumetric determination of 622 Oxidation methods 596 Oxygen 757 determination in illuminating gas 779, 789 presence of hydrogen in 417 Ozone, determination of 676 P. Partition, law of (see law of distribution). Paul's drying oven 33, 34 Percarbonates, analysis of 628 Perchloric acid 462 determination in the presence of chloric acid 463 hydrochloric acid 463 preparation of 51 Perhydrol 230 Periodic acid (and periodates) 670 Permanence of ammoniacal copper solution 756 permanganate solutions 90, 603 sodium thiosulphate solution 649 oxalic acid solutions 58>9 Permanganate methods 596 solution, permanence of 90, 603 preparation of 596 standardization of 91, 597, 827 uses of 603 Peroxides, analysis of 627, 680, 661 Persulphuric acid (and persulphates), analysis by permanganate 629 potassium hydroxide. . 595 titanous chloride 701 Phenol, volumetric estimation of 695 Phenolphthalein 545, 554 Phosphoric acid 434 determination in calcium phosphate 720 mineral water 447 silicates 447 vanadfnite 309 separation from alkaline earths and alkalies 449 chromic acid 499 iron and aluminium .. Ill metals of Groups I, II, and III 448 vanadium. . . 307 920 INDEX OF SUBJECTS. PAGE Phosphoric acid, volumetric determination 587, 588, 718 Phosphorous acid 374 determination in the presence of hypophosphorous acid. 374 Phosphorus, determination in bronze 238, 239 iron and steel 440, 443, 445, 588, 637 organic substances 448 Pipettes 514 calibration of ^ 524 permissible error in 526 Platinum 268 action of ferric chloride on 110 analysis of commercial platinum 272 brass cone lined with 31 capillary for use in gas analysis 743, 766 determination in alloys 270 separation from gold and silver 270, 271 Potassium 38 determination in mineral water 50 indirect determination 56 separation from lithium 53 sodium 43, 49, 50 Potassium bichromate, determination of chromium in 105 potassium in 40, 41 biiodate solution 647 dichromate solution 532, 641, 649 percarbonate, analysis of 628 permanganate solution 90, 531, 597, 827 persulphate, analysis of 629 Precipitates, drying and igniting of 21 filtration and washing of 18 method of igniting when wet 28 Precipitation analyses, (volumetric) 702 Preparation cf the substance for analysis 35 Primary oxides 605 Producer gas, analysis of 775, 783, 784 Protoxides, separation from the sesquioxides 149 Prussian blue, analysis of 343 Prussic acid (see Hydrocyanic acid). Pyridine bases, titration of 561 Pyrite, determination of sulphur in 362 Pyrogallol solution, preparation of 758 Pyrolusite, analysis of 624, 663 Q. Quartation 259 INDEX OF SUBJECTS. 921 R. PACE Recrystallization 35 Red lead (minium), analysis of 623, 676 Reduction methods of volumetric analysis 697 of ferric salts 607 weighings to vacuo 13 Resorcin blue as indicator 544 Rutile, determination of titanium in 118 S. Salting out method for precipitating zinc 160 Selenium 277, 374 determination in crude copper 284 separation from gold and silver 282 metals of Groups II, III, IV, and V 280, 281 tellurium 279, 282 Selenous acid 277, 374 Sensitiveness, or sensibility, of the balance 7 Sesquioxides of Group III 82 separation from the protoxides 149 Silica, separation from tungsten 302 testing the purity of 487 triangles for platinum crucibles 29 Silicates, analysis of 491 decomposable by acids 485 determination of alkalies in 496 ferrous iron in 502 water in 484, 512 not decomposable by acids 491 Silicon, determination in iron and steel 441, 442 the presence of silica 513 Silver 317 determination in alloys 259, 706 ores (see Gold) 268 separation from other metals 318 selenium and tellurium 282 volumetric determination 702, 705 Silver chloride, solubility of 317 Sodium 43 determination in silicates 496 indirect determination of 56 separation from lithium 53 potassium 43-50 9 2 2 INDEX OF SUBJECTS. PAGE Sodium hydroxide, determination in commercial caustic soda 558 caustic soda solution 558 normal solution of 553 preparation of a solution free from carbonates 555 sulphide, reagent, preparation of 225 succinate method of separation 155 thiosulphate solution, normal solution of 645 permanent of 649 Solubility product, definition of 156 Solution of sulphides, explanation of the process 157 Specific gravity tables of acid and alkali 838-841 Stannic chloride, analysis of .' 321, 573 Starch solution 652 Statical moment of a balance 8 Steel, determination of carbon in 398 manganese in 615, 616, 619, 620, 642 nickel 166, 723 nickel, manganese, chromium, and vanadium. . . . 313 phosphorus 440, 443, 445, 588, 637 Stibnite, determination of antimony in 686 Streak of gold alloys 261 Strontium 72 separation from barium 80 calcium 79 magnesium 78 metals of Group II 192 metals of Group III 107, 147 Substitution, weighing by 9 Sulpho-acids 205 separation from one another 241 Sulpho-bases 168 separation from one another 194 Sulphides, determination in the presence of sulphates 470 theory of their solubility in acid 157 titration of 689 Sulphocyanic acid 339, 712 separation from halogen hydrides 342 hydrocyanic and hydrochloric acids . . 713 Sulphydrates, analysis of 689, 691 Sulphur, colorimetric determination 354 determination in insoluble sulphides 357, 367, 368 iron and steel 352, 354, 364, 365 mineral waters 688 organic substances 370 p>rite 357, 362, 716 INDEX OF SUBJECTS. 923 PAOB Sulphur, determination in rocks 505 sulphides soluble in acids 350 water 349 dioxide, gravimetric determination 373 volumetric determination 587, 692, 815 removal from precipitated sulphides 169, 180, 223 volumetric determination in gas mixtures 687, 816 'Sulphuretted hydrogen (see Hydrogen sulphide and Sulphur) . Sulphuric acid 464 determination of arsenic in 248 in the presence of soluble sulphides 470 preparation of concentrated acid of definite strength 580 normal solution of 551 volumetric determination of 571, 577, 714, 716 Sulphurous acid (see Sulphur dioxide) 373 volumetric determination 587, 692 Swings, weighing by the method of 10 T. Tables * 517, 519, 520, 522, 533, 534, 535, 838-857 Tantalum, determination of wolframite 297 Tare 9 Tartar emetic, analysis of 685 Tartaric acid 433 Teclu burner 94 Telluric acid (see Tellurium) 664 Tellurium, determination in crude copper 284 precipitation with sulphurous acid 279 separation from antimony, tin, and arsenic 281 copper, bismuth, and cadmium 280 gold and silver 282 mercury 281 metals of Group II 280 Groups III, IV, and V 279 selenium 282 volumetric estimation 664 Tellurous acid (see Tellurium) 374 Temperature, taken as normal in volumetric work 516 Testing of weights 15 Tetrahedrite, analysis of 359 Thallium, determination of 318 Thiocyanic acid (see Sulphocyanic acid) . * The tables in this book can be purchased separately in flexible cloth binding. Price, thirty-five cents. 924 INDEX OF SUBJECTS. PAGF Thiosulphuric acid (thiosulphate) 450, 645 determination in presence of sulphide 691 Thorium, determination in monazite 510 Tin 228 determination in bearing metal 252 bronze 236 tin chloride 321, 573 separation from alkaline earths and magnesium 235 antimony 248, 256 arsenic 255 mercury, lead, bismuth, copper, and cadmium 235 metals of Group III 235 phosphoric acid 238, 239 silicic acid 298 tungsten 297, 300 Titanium 100 Titanium, colorimetric determination. 100 determination in rocks 504 rutile and iron ores 118 separation from alkaline earths 107 aluminium 116 iron 114 manganese, nickel, cobalt and zinc 149, 152 Titanous chloride, as reagent in volumetric analysis 700 Total carbon in iron and steel 399 Tungsten . . 288 determination in steel 291 wolframite 296 separation from molybdenum 293 silica 302 tin 297, 300 Tungsten bronzes, analysis of 298 U. Uranium 106 separation from alkaline earths and magnesium 107 aluminium and iron 119 metals of Group II 192, 235 nickel, cobalt, manganese and zinc 149 volumetric determination 621 V. Vacuo, reduction of weighings to 13 Valve,Bunsen 87, 98, 601 Contat-Gockel 304, 60-2 Vanadic acid (see Vanadium). INDEX OF SUBJECTS. 925 PAGE Vanadium 303 determination in ores and rocks 310 pig iron 312 steel 313 vanadinite 309 separation from arsenic 306 molybdenum 308, 313, 667 phosphoric acid 307 volumetric determination 636, 665 Vanadinite, analysis of 308 Vapors, determination in gases 831 Volume, normal 516 Volumetric analysis 1, 514 W. Water bath 31 Water, density at different temperatures 517 determination of absorbed oxygen in 760 in fluosilicates 484 lepidolite 502 silicates 512 hardness of 569, 570 vapor, tension of 842 Washing precipitates 18 Weighing 6 double 9 by substitution 9 swings 10 reduction to vacuo 13 Weights, testing of 15 Wet precipitates, method of igniting 28 Welsbach mantles, analysis of . \ 512 White lead, analysis of 379 wolframite (Wolfram) analysis of 296 Z. Zinc 140 electrolytic determination 145 separation from alkaline earths and magnesium 147 metals of Group II 192, 235 nickel cobalt, and manganese 156 trivalent metals of Group III 149-155 determination in bearing metal 252 brass 193 bronze 237 Zirconium, determination in rocks 505 \ YC 1 VI 12