"1 REESE LIBRARY [ OK THK UNIVERSITY OF CALIFORNIA. ;' rs* IT. Accessions F. E. BECKER & CO, importers anfc i&anufacturer0 of CHEMICAL AND SCIENTIFIC INSTRUMENTS FOR GAS ENGINEERS, &c. CONTRACTORS TO THE GAS LIGHT AND COKE COMPANY, &c, Ac. ILLUSTRATED PRICE LIST ON APPLICATION TO F, E. BECKER & CO. 33, 35 and 37 Hatton Wall, Hatton Garden, LONDON, B.C. JAMES M C KELVIE & Co (tone! $ das (0al (B#p0rtcrs, 37 & 38, MARK LANE, LONDON, E.G. Head Office: HAYMARKEJ, EDINBURGH. Prices and Analyses of all the principal Cannels and Gas Coals forwarded on application. SOLE AGENTS FOR THE DUKE OF HAMILTON'S LESMAHAGOW CANNEL COAL. ESTABLISHED 1840. Telegraphic Address: "IGNITIBLE, LONDON." "MACKELVIE, EDINBURGH.' iii W U cc -I CC o O ', E-i s K> u O s cp HM O CO s DRY, WET, EXPERIMENTAL, TEST, AND OTHER METERS, THOMAS GLOVER & CO. 214 to 222 St. John St., Olerkenwell Green, London, E.G. "GOTHIC, LONDON." Telephone No. 6725, Meters, Test Holders, Pressure Gauges, Pressure Registers, King's Gauges, Main & Ordinary Thermometers, Standard Baro- meters, Standard Photometers, Jet Photometers, Standard Ar- gand Burners, Chemicals, Balances, Standard Testing Apparatus for Ammonia, Sul- phur, Harcourt's Colour Test, Chemicals, &c. &c. THOMAS GLOVER & CO. GAS APPARATUS AND APPLIANCES, 214 to 222 ST. JOHN ST., CLEflKENWELL GREEN, " Gothic, London. ' Telephone 6725. [To face Title. THE GAS ENGINEEB'S lABORATOKY HANDBOOK BY JOHN HOKNBY, F.I.C. HONOURS MEDALLIST IK GAS MANUFACTURE, CITT AND GUILDS OF LONDON INSTITUTE HontJon: E. & F. N. SPON, 125 STEAND flefo SPON & CHAMBEKLAIN, 12 COKTLANDT STKEET !894 UNIVEB8IV1 Of & bis little SStork is gritt V TO MY CHIEF JOHN METHVEN, ESQ., A.M.I.C.E AS A SLIGHT RECOGNITION OF HIS UNIFORM COURTESY AND KINDNESS PBEFACE. T IE object of this volume, as its title implies, is to describe tl e various analytical operations which are required in Gas- \v >rks. In order to make it useful to those who may not have the 0] portunity of access to other works on analysis, the general operations incidental to the carrying out of analytical o] erations are described in detail, and further, as a means oi assisting those who are dependent on self study, the in stbods of analysing certain substances obtainable in a p( rfectly pure state, and consequently having an established composition, are minutely described ; these examples are chosen, as being typical of the more commonly occurring ai alyses required in Gasworks, as they involve the same methods of manipulation. By comparing the results obtained in the analysis of these simple substances with the figures demanded by theory, the operator can satisfy himself as to his competency to perform the same operations when they occur in the estimations required in his daily work. For example, the analysis of barium chloride serves as a type of the methods of manipulation necessary in the determination of sulphur in coal, coke, and gas, and in ascertaining the accuracy of the solutions employed in the Keferees' Ammonia Test, &c. ; the estimation of calcium serves as a type of the analysis of lime ; and so on. The subject matter has been arranged as follows : Part I. treats of various preliminary operations necessary vi THE GAS ENGINEERS LABORATORY HANDBOOK. in quantitative work, such as the use of the balance, the pre- paration of substances for analysis, the ordinary operations employed in analysis, &c. Part II. describes the method of estimating gravimetri- cally certain pure substances which are typical of the more commonly occurring analyses required in Gasworks. Part III. gives an account of various operations con- nected with volumetric analysis, a description of the vessels employed for measuring liquids, and the method of calibrating the same ; some simple volumetric processes applicable to gas testing being afterwards described. Part IV. describes the more complex determinations required in Gasworks, some being gravimetric, others volu- metric, while some embrace both methods. These include complete analysis of coal, fresh and spent purifying material, (lime, oxide of iron and Wei don Mud), ammoniacal liquor, sulphate of ammonia, fire bricks and clay, tar, crude gas for C0 2 , SH 2 and NH 3 , and purified gas for NH 3 and sulphur compounds. Part V. treats of technical gas analysis, analysis of coal gas, furnace gases, &c. In the Appendix a variety of tables of use in gas testing will be found. In a book of this description there is not much scope for originality, and my work has principally consisted in collecting together, and arranging in a convenient and easily accessible form, the scattered literature dealing with the various subjects treated on, and in explaining by means of numerous practical examples the chemical reactions and calculations upon which the various tests depend. I am greatly indebted to the excellent text- book on * Quantitative Analysis ' by Prof. Clowes and Mr. Coleman, which is published by Messrs. J. & A. Churchill. Many of the engravings and descriptions of analytical methods and processes have been, by permission, copied from that treatise. Keference may with advantage be made to Messrs. Clowes PREFACE. Vll and Coleman's work for additional methods, as well as for full }r details concerning many of the processes included in my text. C have also derived considerable assistance from the tre- tise on ' Quantitative Analysis ' by Prof. Thorpe, pub- lisl ed by Messrs. Longmans & Co. This is also well suited for general reference. The description of Hempel's Gas Ap laratus and its applications is taken, by permission of Mo ;srs. Macmillan & Co., from Hempel's ' Gas Analysis,' tra islated by Dennis. The last mentioned work should be coi suited for fuller details, and for a description of Gas Ai ilysis generally. In addition to the above, the following works, which have al^ > been consulted, will give further information on some of he subjects treated on : Sutton's 'Volumetric Analysis.' Allen's ' Organic Analysis.' Streatfield's * Organic Chemistry.' Cohen's * Organic Chemistry.' Arnold's 'Ammonia and Ammonium Compounds.' Lunge's ' Tar and Ammonia Distillation.' Hartley's ' Ammonia Liquor Tests.' Winkler and Lunge's ' Technical Gas Analysis.' Lunge and Hurter's * Alkali Maker's Pocket Book.' ' The Journal of the Society of Chemical Industry.' ' The Journal of Gaslighting.' J. H. GASWOKKS, BECKTON : July 1893. CONTENTS. PAET I. SECTION I. INTRODUCTORY THE BALANCE WEIGHTS AND WEIGHING. PAR. . PAGB Introductory .. .. .. .. .. .. .. .. 1 1 The Balance 2 2 The Weights 5 3 The Eider 6 4 Testing the Balance 7 5 Testing the Weights 9 6 The Process of Direct Weighing 9 7 Weighing by Substitution 9 8 Correction for Weighing in Air .. .. .. .. .. 10 9 Directions for Weighing 10 SECTION II. SAMPLING MECHANICAL DIVISION DRYING AND DESICCATION SOLUTION AND EVAPORATION PRECIPITATION FILTRATION AND TREATMENT OF PRECIPITATES. 10 Sampling 14 11 Mechanical Division .. .. .. .. .. .. 14 12 Desiccation .. .. .. .. .. .. .. .. 16 13 Solution of Solids 19 14 Evaporation 20 15 Precipitation .. .. .. .. .. .. .. .. 22 16 Filtration and Washing of Precipitates 23 17 Accelerated Filtration 28 18 Drying of Precipitates .. .. 32 19 Ignition of the Precipitate 33 20 Collection of Precipitates on Weighed Filters 38 THE GAS ENGINEER'S LABORATORY HANDBOOK. PAKT II SIMPLE GRAVIMETRIC ESTIMATIONS. PARA. PAGE Introductory Remarks .. .. .. .. .. .. 40 21 Preparation of Pure Substances by Crystallisation .. .. 40 22 Precipitation .. .. 42 23 Sublimation .. .. 42 24 Estimation of S0 4 in a Sulphate as BaS0 4 43 25 Barium as BaSO 4 47 26 Arsenic as Sulphide .. .. .. .. 48 27 Cadmium as Sulphide .. .. .. .. 50 28 Calcium as Carbonate, and as Oxide .. .. 51 29 Iron as Oxide 52 30 Aluminium as Oxide .. .. ., .. 54 31 Separation of Iron and Alumina .. .. .. .. .. 55 32 Estimation of Magnesium as Pyrophosphate .. .. .. 56 33 Silicate as Silica 57 34 Carbonate as CO 2 , by Direct Weighing, and by Loss 59 PART III. VOLUMETRIC ANALYSIS. Introductory Remarks . . . . . . . . . . . . 68 Example of Volumetric Analysis . . . . . . . . . . 68 35 Notes on the Measurements of Liquids as applied to Volumetric Analysis .. .. 70 36 Notes on Measuring Vessels .. .. .. .. .. 71 37 Measuring Flasks and their Calibrations .. .. .. .. 72 38 Test Mixers and their Calibrations 74 39 Pipettes and their Calibrations .. .. .. .. .. 74 40 Burettes and their Calibrations .. .. .. ..76 NOTES ON THE PREPARATION OF SOLUTIONS. 41,42,43 Standard Solutions 7879,80 44 Application of Normal Solutions .. .. .. .. .81 45 Empirical Solutions used in Gasworks .. .. .. . 82 46 Precautions to be Observed in making Standard Solutions . 82 47 Storage of Standard Solutions .83 48 Indicators ...... 84 CONTENTS. XI I: DICATORS USED IN THE VOLUMETRIC ESTIMATION OF ACIDS AND ALKALIES. PARA 49 3ochineal Solution .............. 84 50 Methyl Orange .............. 85 51 Litmus Solution .. .. .. 85 52 Pheuol-phthalein .. .. ..86 ANALYSIS BY SATURATION. (ALKALIMETRY AND ACIDIMETKT.) Preparation of Normal Acid and Alkaline Solutions. 53 General Eemarks .. .. .. .. .. .. 87 54 Normal Sodium Carbonate Solution ........ 88 55 Normal Sulphuric Acid Solution .. .. .. .. .. 90 56 Normal Hydrochloric Acid Solution . . . . . . . . 93 57 Normal Oxalic Acid Solution .. .. .. .. .. 93 58 Normal Sodium Hydrate Solution .......... 93 59 Mode of Using Normal Solutions in the Estimation of Ammonia 94 60 Appli2atioii of Normal Solutions in the Analysis of Lime .. 97 PROCESSES OF OXIDATION AND KEDUCTION. General Kemarks . . . . . . . . . . . . . . 98 61 Preparation and Use of Potassium Permanganate Solution .. 99 62 Application of Potassium Permanganate Solutiou in the Analysis of Weldou Mud .......... 103 63 Use of Potassium Bichromate Solution ........ 104 64 Application of Potassium Bichromate Solution in the Analysis of Iron Ore .............. 108 65 Use of Standard Iodine Solution .......... 110 PAKT IV. SPECIAL ANALYSES REQUIRED IN GASWORKS. I. ANALYSIS OF COAL AND COKE. 66 Proximate Analysis of Coal 113 67 Elementary Analysis of Coal .. .. .. .. ..117 68 Specific Gravity of Coal 124 xii THE GAS ENGINEER'S LABORATORY HANDBOOK. II. ESTIMATION OP THE IMPURITIES IN CRUDE GAS. PARA. PAGK 69 Estimation of Sulphuretted Hydrogen and Carbonic Acid (Wright's Method) 125 70 Estimation of Carbonic Acid (Sheard's Method) 132 71 Estimation of Sulphuretted Hydrogen by means of Cadmium Chloride .. .. .. 137 III. TESTING PURIFIED GAS TOR SULPHUR COMPOUNDS, &o. 72 General Instructions.. .. .. .. .. .. .. 138 73 Preliminary Stage in Testing for Sulphur Compounds and NH 3 139 74 Solutions required in the Estimation of Ammonia . .. 139 75 Preparation of Standard Sulphuric Acid Solution for Ammonia Test 140 76 Adjusting Sulphuric Acid Solution to Correct Strength .. 141 77 Further Particulars Eelative to Sulphuric Acid Solution .. 143 78, 79 Preparation of Standard Ammonia Solution .. .. 143, 145 80 Storage of Solutions used in the Ammonia Test .. .. .. 145 81 Description of Apparatus employed in Ammonia Test .. .. 145 82 Description of Apparatus employed in Sulphur Test .. .. 145 83 Fitting up of the Apparatus 147 84 Method of Making Test 147 85 Determination of the Ammonia .. .. .. .. ..148 86 Calculations for Obtaining Amount of Ammonia .. .. 149 87 Method of Analysing the Liquor produced in the Sulphur Test Apparatus .. .. .. .. .. .. .. 150 88 Explanation of Rule for obtaining Amount of Sulphur by Referees' Method 150 89 Correcting the Amount of Gas Burnt in Referees' Sulphur Test for Temperature and Pressure .. .. .. .. 151 90 Explanation of the Chemical Reactions which take place in the Referees' Sulphur Test 15.2 91 Estimation of Sulphur by Evaporating the Liquor obtained in the Referees' Test .. 153 92 Estimation of Ammonia in Crude Gas Preparation of Solu- tions 15 93 Mode of Testing for NH 3 in Crude Gas 154 94 Harcourt's Colour Test, General Remarks 155 Description of Apparatus .. .. 155 Method of using .. .. .. 156 97 Method of Determining Amount of Sulphur 158 98 Table showing Amount of Sulphur according to Harcourt's Colour Test 158 99 Remarks on the Standard and Lead Solution employed in Harcourt's Colour Test 159 CONTENTS. Xlll IV. ANALYSIS OP AMMONIACAL LIQUOR. PARA. PAGE 100 Composition of Ammoniacal Liquor .. .. .. .. 159 101 Causes Influencing the Composition of Ammoniacal Liquor .. 160 102 Llemarks on the Method of Determining the Commercial Strength of Ammoniacal Liquor .. .. .. .. 161 103 Description of the Acid or Saturation Test for Determining the Value of Ammoniacal Liquor .. .. .. .. .. 162 104 Vlethod of Determining the Value of Ammoniacal Liquor by Distillation Preparation of Standard Sulphuric Acid Solution " 163 105 Checking the Strength of Acid Solution employed in Liquor Testing Gravimetrically . . . . . . . . 1 64 106 Jheckini: the Strength of Acid Solution Volumetrically .. 165 107 Preparation of Standard Soda Solution for Liquor Test .. 166 108 Arrangement of Apparatus and Mode of Testing Liquor by Distillation Test 167 109 ritration of Solution in Liquor Test 168 110 jteneral Notes in Connection with Liquor Testing .. .. 168 111 Testing Liquor by Means of Normal Solutions .. .. .. 169 112 Estimation of Sulphuretted Hydrogen in Gas Liquor Gravi- metrically .. .. 169 113 Sstimation of Sulphuretted Hydrogen in Gas Liquor Volume- trically 170 114 Estimation of Carbonic Acid in GHS Liquor .. .. .. 171 115 Complete Analysis of Ammoniacal Liquor .. .. .. 172 116 Analysis of Sulphate of Ammonia .. .. .. .. 175 V. ANALYSIS OF LIME. 117 ' Jeneral Notes on Lime .. .. .. .. .. ..176 118 Theoretical Purifying Value of Lime 177 119 impurities in Lime .. .. .. .. .. .. .. 177 120 information Commonly Kequired with Eespect to Lime .. 178 121 Obtaining the Valuo of Lime from its Behaviour on Slaking .. 178 122 Data Necessary in Order to Arrive at the Actual Purifying Value of Lime .. .. .. .. .. .... 179 123 Estimation of Lime Volumetrically 17!) 124 another method 180 VI. ANALYSIS OF LIMESTONE. 125 Determination of the Constituents in Dolomite 180 126 Kapid Method for the Analysis of Limestone .. .. .. 18(3 VII. ANALYSIS OF OXIDE OF IKON. 127 Analysis of Oxide Gravimetrically and Volumetrically .. 187 xiv THE GAS ENGINEEE'S LABORATORY HANDBOOK. VIII. ANALYSIS OP SPENT OXIDE OF IRON. PARA. PAGR 128 Method of Determining the Amount of Sulphur in Spent Oxide 189 129 Estimation of Sulpho-cyanides in Spent Oxide .. .. 191 130 of Ferro-cyanides in Spent Oxide 192 IX. ANALYSIS OF FIRE-CLAY AND FIRE-BRICKS. -, qi / General Notes on Fire-clay .. .. .. .. .. 194 (Method of Analysis of Fire-clay .. ..194 X. ASSAY OF COAL TAR. 132 Method of Conducting an Assay of Coal Tar 200 XL FRACTIONAL DISTILLATION. 133 Kectification of Benzene 203 XII. DETERMINATION OF THE SPECIFIC GRAVITY OF GAS. 134 Method of Ascertaining the Specific Gravity of Gases.. .. 207 PART V. TECHNICAL GAS ANALYSIS. Introductory Kemarks .. .. .. .. .. .. 213 135 Notes on the Measurement of Gases .. 214 136 Eeduction of Gases to Normal Temperature and Pressure 137 Comparison of G as Volumes 138 Calibration of Measuring Tubes Employed in Gas Analysis 139 Hempel's Burette 140 Gas Analysis by means of HempeFs Apparatus 141 Description of Hempel's Gas Burette 142 Manipulation of the Gas Burette 143 Description of Hempel's Simple Absorption Pipette 144 The Double Absorption Pipette 145 Method of filling the Simple Pipette 146 Double Pipette .. 147 Manipulation of the Absorption Pipette.. 148 Treatment of Pipettes when not in use 214 216 216 217 219 219 220 221 222 223 223 224 227 149 Precautions to be Observed to Secure good Kesults with Hempel's Apparatus .. .. .. .. .. .. 227 CONTENTS. XT PARA. PAGE 150 LLt of Gases most frequently met with in Gasworks .. .. 227 151 Properties and Method of Determination of Oxygen ., .. 227 1 52 Preparation of Solution of Potassium Py rogallate tor Absorption of Oxygen 228 153 Preparation of Phosphorus for Absorption of Oxygen .. .. 228 154 Conditions under which Phosphorus may be used for the Determination of Oxygen .. . .. .. .. 229 155 Determination of Hydrogen by Combustion with Oxygen .. 229 156 Effect of the Presence of Nitrogen on the Combustion of Hydrogen 230 157 Combustion of Hydrogen with an Excess of Air and Pure Hydrogen 230 158 Combustion of Hydrogen by means of Palladium Sponge .. 231 159 Determination of the Amount of Hydrogen in a Mixture of Hydrogen, Nitrogen and Marsh Gas .. .. .. 231 160 Kegeueration of Palladium Employed in the Combustion of Hydrogen 233 161 Method of Preparing Palladium for Use 233 162 Absorbents for Carbonic Oxide 234 163 Preparation of Ammoniacal Cuprous Chloride Solution .. 234 164 Preparation of Acid Cuprous Chloride Solution .. .. 235 165 Absorbent for Carbon Dioxide (CO 2 ) 235 16G Estimation of Ethylene (C 2 H 4 ) by Means of Fuming H 2 S0 4 . . 236 167 Precautions to be Observed in the Determination of Ethylene by Fuming H 2 SO 4 . . . . . . . . 236 168 Absorption of Ethylene by Bromine 237 161) Determination of Marsh Gas (CH 4 ) 237 170 Nitrogen 238 171 Gases and Gaseous Mixtures Capable of being Analysed by means of Hempel's Apparatus . . . . . . . 238 172 Method of Analysing Coal Gas by means of Hempel's Appa- ratus .. .. .. 238 173 Determination of Hydro- carbon Vapours in Coal Gas .. .. 239 174 CO 2 in Coal Gas . . . . . . 239 175 Heavy Hydro-carbons in Coal Gas .. .. 240 176 Oxygen iu Coal Gas 240 177 Carbonic Oxide in Coal Gas 240 178 Treatment of the Kesidue for Amount of H 2 ,CH 4 ,N 2 . . . . 241 179, 180 Combustion of the Kesidue for Amount of H 2 ,CH 4 ,N 2 .. 241 181 Treatment of the Kesidue after Combustion .. ..241 182 Determination of the Marsh Gas from the result of the Combustion .. .. .. .. .. .. .. 241 183 Determination of the Hydrogen from the result of the Com- bustion 242 184 Summary of the Kesults in the Analysis of Coal Gas .. .. 242 185 Determination of Hydrogen in Coal Gas by Absorption with Palladium 242 186 Details of Estimation of Hydrogen by Absorption with Palla- dium ' 243 187 Determination of Marsh Gas when Hydrogen is Burnt by Palladium 244 xvi THE GAS ENGINEER'S LABORATORY HANDBOOK. PARA. PAGE 188 Calculations to Determine Amount of Marsh Gas .. ... 244 189 Result of Analysis of Coal Gas 244 ANALYSIS BY MEANS OP BUNTE'S APPARATUS. 190 Analysis of Furnace Gases, General Description of Apparatus 244 191 Determination of CO 2 245 192 Oxygen 247 193 CO and Hydrogen 248 194 Calculation of the Eesults Obtained by the Combustion of CO andH .. .. 249 195 Determination of CO by Absorption with Cuprous Chloride Solution 250 196 Method of Sampling Furnace Gases 250 APPENDIX. GAS REFEREES' INSTRUCTIONS. As to the Times and Mode of Testing for Purity 251 As to the Maximum Amount of Impurity in each form with which the Gas shall be allowed to be charged 257 Meters 257 Gas Referees' Cubic Foot Measure 258 TABLES 261 INDEX 297 Errata. Page 79, line 2 from top, for single read simple. 185, 9 from top, for precipitates read nitrates. 186, 15 from top, for K 2 Pt HC1 6 read K 2 Pt C1 8 . 189, 10 from bottom, for grams read grains. 242, 4 from top, after marsh read gas. THE GAS ENGINEER'S LABORATORY HANDBOOK. PAET I. SECTION I. INTRODUCTORY THE BALANCE WEIGHTS AND WEIGHING. THI chemical work of a gasworks laboratory chiefly con- sist ; in the quantitative determination of various substances wh( >se composition is already known. The gas engineer is therefore principally concerned with " quantitative " analysis. By " quantitative " analysis is meant that branch of practical chemistry which treats of the methods by which are determined the relative amounts of the constituents of a body, " qualitative " analysis simply informing us of the nature of the constituents, and how they may be separated. The methods of quantitative analysis are divided into two branches, gravimetric and volumetric analysis. By gravimetric analysis is meant the estimation of a substance by converting its known constituents into forms or combinations which admit of the most exact determination of their weight, and of which the composition is already accurately known. By volumetric analysis is meant the method of deter- mining the quantity of a substance (in solution) by the use of a measured volume of a reagent of known strength. B 2 THE GAS ENGINEER'S LABORATORY HANDBOOK. The Balance. 1. The first essential for the correct performance of gravimetric analyses is the balance, which for careful quantitative work should turn easily and quickly, without too much oscillation, to T V of a milligram, or -^^ or ^^ of a grain, when 50 or 60 grams, or 1000 grains, are placed on each scale-pan. Fig. 1 shows one of the forms of this instrument. It consists of a perforated brass beam, suspended at the centre on a triangular piece of hardened steel or agate termed a " knife-edge," which oscillates on a polished plate of agate, fixed on a brass support. At each end of this beam, is fixed a similar knife-edge of steel or agate, on which rest two agate plates, imbedded in brass settings. From each of these a hook depends, to which the pans are attached by light wires or thin straps. Figs. 2 and 3 give enlarged views of these terminal knife-edges and planes. When the instrument is not in use, the beam rests upon a frame connected with a rod which descends through the pillar of the balance, and which can be raised or lowered by an eccentric. The same movement also lifts up the plates connected with the pans, so that the knife-edges and agate planes are only in contact when the balance is oscillating. A second eccentric, connected with two bent levers, raises or depresses a support under each of the pans, so that they may be kept stationary whilst their load is being placed on them, or their vibrations arrested before the beam is lowered down on the centre plate. In some balances, all these movements are so connected that a single turn of the eccentric first rapidly releases the supports of the pans, then gently lowers the beam upon the centre knife-edge, and finally drops the pan suspensions upon the terminal knife-edges. By these arrangements the durability of the instrument is greatly increased. To the beam is attached a long pointer, moving in front THE BALANCE. of a graduated ivory scale fixed to the pillar of the balance. This renders visible the slightest oscillations of the beam. So long as the instrument is in equilibrium, the pointer hangs perfectly vertical, and opposite to the centre (zero) point of the scale, or it oscillates over an equal number of B 2 THE GAS ENGINEER'S LABOKATOEY HANDBOOK. divisions on each side. The least preponderance on one side is indicated by a permanent deflection of the pointer from the zero, or by its vibrating through an unequal number of divisions on either side of the zero point. Tf from any cause the instrument itself is not in equilibrium, it may be adjusted by means of a little movable vane fixed on the top of the beam, by moving which, either to the right or to the left, the preponderance may be counterpoised. In some balances, in which the pans are suspended by a simpler method, this adjustment is effected by means of a little milled-head screw travelling along a thread, Fig. 3. FIG. 2. FIG. 3. The accuracy of the balance depends (1) upon the equality of its arms, (2) upon the position of the point of support in relation to the centre of gravity of the instrument, from which also follows (3) that the points of suspension and support must be in the same plane, and (4) that the beam must be perfectly rigid. The sensibility of the instrument depends (1) upon the freedom from friction of the knife-edges and planes, (2) upon the near approximation of the centre of gravity of the instrument to its point of support, and (3) upon the lightness of its beam. The relation of the centre of gravity to the point of support is, of course, adjusted as accurately as possible by the balance-maker, but by means of a small weight (termed a gravity bob) working on a thread, it may be THE BALANCE. 5 altered at pleasure, and the stability of the balance thus regulated. On account of the injurious effects of acid fumes and moisture, which rapidly impair the delicacy of a balance, tin instrument is enclosed in a wooden case, having counter- ba anced sliding glass sashes at the front and back. The ca^e stops the entrance of dust, diminishes the rusting of the kn ife-edges, and prevents air-currents from acting upon the set le-pans during the process of weighing. The balance- ca e is supported on levelling screws, by which it can be ail usted to a perfectly horizontal position by means of a sp rit-level. The balance, if possible, should be kept from acid fumes, ai; I, as a further precaution, a shallow vessel containing dr ed carbonate of potash or quicklime should be placed in th 3 balance-case, being renewed when it becomes deli- qi escent by the absorption of moisture from the air. It is fu?ther necessary that the balance should be placed in a pc sition free from extremes of temperature, and its support sb ould be quite free from vibration. The Weights. 2. The weights generally employed in chemistry are those of the French or metric system, of which the gram is the unit ; in the chemical operations incidental to gasworks, however, it is often necessary to use weights based on the English grain. The metric system is in every respect the most con- venient, for the following reasons : (1) that the unit is extremely small; (2) that the different denominations in weight being related by multiples and submultiples of ten, they fit in with the system of numerical notation in general use, and consequently make chemical calculations more simple; and (3) that it bears a very simple relation to the measures of capacity and length. THE GAS ENGINEER S LABOKATORY HANDBOOK. For general use a set of weights ranging from 50 grams to a milligram on the metric system, and from 1000 grains to T V grain on the grain system, will be found convenient. They are generally preserved in a box, Fig. 4, the larger weights being placed in receptacles lined with velvet to pre- vent their being scratched, the smaller ones being also kept in separate compartments, covered by a plate of glass. The box is provided with a pair of small forceps with which the FIG. 4. weights are always to be lifted, never on any account being touched by the fingers. The gram weights down to one gram, and the grain weights down to ten grains, are generally made of brass, cylindrical in shape, and provided with handles for lifting by the forceps. The lower weights in both systems are usually made of platinum or aluminium foil, one corner of each being turned up for holding in the forceps. 3. Accompanying the weights are the riders. The rider THE WEIGHTS. 7 is a piece of fine wire bent as in Fig. 5. It is generally ma le either of gilt brass or of aluminium wire, and by reason of i ts form it is capable of " sitting " across the balance, and remaining on whatever part of the beam it may be placed. Tli 3 object of the rider is to weigh the lowest fractions, such as milligrams and tenths of a milligram, and hundredths an< 1 thousandths of a grain. The rider in the metric system weighs one centigram, and in the grain sy* tern a tenth of a grain. The arm of the beam of th balance, from the centre to the terminal edge, is di^ ided into ten equal parts, and each of these is (gc uerally) further subdivided into five equal parts, Tl e centigram rider, when placed just above the ter- mi lal knife-edge of the beam, weighs its normal weight of one centigram, or 10 milligrams ; when placed on a point on the an i exactly in the middle of the two knife-edges it exerts on y half this effect, that is to say, it is now equal to 5 mi lligrams ; when placed at a fourth of the distance from the centre knife-edge it is equal to 2' 5 milligrams; at three- qu arters the distance 7 5 milligrams, and so on. The rider is noved from one division to another by means of an arm parsing through the balance-case, and operated from the on :side. Testing the Balance and Weights. 4. Before commencing a course of quantitative work, it is necessary to test the accuracy of the balance, which may be done by means of the following experiments : (a) The floor of the balance is first of all brought to the horizontal position, by the aid of the levelling screws and spirit-level ; it is then accurately adjusted, if necessary, either by the regulating screws, or by means of pieces of tin foil, and a milligram weight placed in one of the scale-pans. A balance adapted for every-day work should turn very dis- tinctly with this weight, whilst one adapted for delicate determinations should indicate the y 1 ^ of a milligram with 8 THE GAS ENGINEER'S LABORATORY HANDBOOK. distinctness. It may be here pointed out, that the mere fact of the index needle pointing to zero on the scale, is not sufficient to prove that the balance is in equilibrium. The better plan is to observe the oscillations of the pointer, which can, if necessary, be caused to vibrate by moving the hand near one of the scale-pans, so as to create a slight current of air. The pointer should then traverse very nearly the same space on each side of the zero mark, travelling a lesser dis- tance with each oscillation, and finally stop at the zero point. (6) Both scale-pans are loaded with the maximum weight the balance is constructed to carry. The instrument is then accurately adjusted, and a milligram added to the weight in one of the scale-pans. This should cause the balance to oscillate to about the same extent as in (a), but in the majority of cases it is somewhat less. (c) The balance is accurately adjusted (in the event of it being necessary to bring the scale-pans into equilibrium by adding tin foil to one of them, this tin foil must be left on the scale-pan during the experiment); both pans are then equally loaded with, say, 50 grams each, the balance, if neces- sary, being again adjusted (by adding small weights) ; the load on the two pans is then reversed, so as to transfer that of the right-hand pan to the left-hand, and vice versa. If the balance has perfectly equal arms it should maintain its absolute equilibrium. (d) The balance is again accurately adjusted ; it is then arrested, and afterwards set in motion again, the same operation being repeated several times. A good balance should always resume its original equilibrium; faulty construction of the knife-edges will cause differences to appear. (e) The speed of oscillation of the index pointer should be observed, for the slower the needle oscillates, the greater is the sensitiveness of the balance to small differences of weight, and vice versa. It should not oscillate too slowly however, or it will cause the process j}f weighing to become wearisome. TESTING THE BALANCE AND WEIGHTS. 9 Tho oscillations are slower the nearer the centre of gravity of the beam approaches the axis of its suspension. 5. The weights should be tested in the following manner to see if they agree among themselves. One pan of the balance is loaded with a 1-gram weight, am then counterbalanced with small pieces of brass, and, finally, tin foil. The weight is then removed, and replaced by another 1-gram weight, any difference being noted ; the rei laining gram weight is then treated in the same manner. Th 3 2-gram weight is then weighed against two single l-ram weights, the 5-gram weight against three single gr; m weights, and the 2-gram weight, and so on. In com- pa] ing the smaller weights among themselves in this manner, the y should not show the least difference on a balance turn- in <. with T L milligram. The Process of Weighing. 6. There are two methods for determining the weight of a s ibstance the direct method, and the method of substitu- tion . When weighing a substance by the direct method, the substance is placed upon one scale-pan, preferably the left, and the weight required to counterpoise it upon the other. Ov\ing to this method being the more expeditious of the two, it is the one commonly employed. If, however, the balance is not in perfect equilibrium, or the arms of the beam are of unequal length, the true weight is not obtained by this method. This is of no consequence in ordinary analytical operations, since, in any one series of weighings, it is the relative weights only of the different substances which are commonly required, and these will be correctly obtained, provided the different substances be placed always upon the same scale-pan, and the condition of the balance does not undergo any alteration between the successive weighings of the series. 10 THE GAS ENGINEER'S LABORATORY HANDBOOK. 7. In the process of weighing by substitution, both the relative and absolute weights of substances are obtained. In this method, the substance is counterpoised by weights in the usual way ; it is then removed from the scale-pan, weights being substituted in its place exactly balancing those already placed in the opposite scale-pan. The weights substituted for the substance will then represent the true weight of the substance. Operating in this way, it would not make any difference if the arms of the balance were not of equal length, or if the equilibrium of the instrument had not previously been adjusted. 8. The volume of the body weighed is almost invariably different from that of the weights which counterpoise it, consequently different volumes of air are displaced by the substance and the weight ; it follows from this, that a body when counterpoised by weights in air would not be in equilibrium with the same weights in vacuo. The magnitude of the error from this cause is proportional to the difference between the volume of the weights and of the body weighed. If the volume of the body is greater than that of the weight, its weight, as indicated by the balance, is too small ; and con- versely, when the volume is less than that of the weights, the indicated weight is too great. In the generality of cases the error thus introduced is too small to be taken cognizance of, but in excessively accurate determinations it is necessary to take it into account. The absolute weight of a substance in vacuo can be obtained, if necessary, as follows. The weight of the volume of air displaced by the weights, is subtracted from the weight of the volume of air displaced by the substance, and the weight thus obtained is added to the weight on the scale-pan. Directions for Weighing. 9. The following directions describe the method of con- ducting the operation of weighing. DIRECTIONS FOE WEIGHING. 11 In the first place, see that the scale-pans and floor of the balar ce-case are quite clean, and free from dust. Kext, see that the rider hangs upon the projecting arm of th ) brass rod, and not upon the beam of the balance. Then see if the instrument is in equilibrium, by gently lowe ing the supports, and setting the beam swinging. If the bean does not commence to oscillate on being released, start it, I) ,* gently waving the hand over one of the scale-pans ; the jointer should then oscillate through equal spaces on each side of the zero of the scale ; if this is not the case, brin ; ; it into equilibrium by moving the brass vane or altering the terminal screws. If, how- ever it is simply required to obtain the weight of a body by difference, this adjustment is not necc isary. Thus, supposing, for instance, we wish to \v jigh out a certain quantity of powdered coal for malysis. It may be put into a weighing bott e, Fig. 6, and the weight of bottle and coal F ~ toge her determined. About as much of the coal as if- required (say 1 gram) is then emptied out and the bottle again weighed; the difference between the first and secoi id weighings gives the weight of coal taken. The substance whose weight is required, should invari- ably be placed on the left-hand scale-pan, the weights being placed by the forceps on the right-hand pan. The weights shou d be placed on the pan in a systematic manner, and not taken at random from the box. The better plan is to start with a weight which is a little too heavy, and then to try the effect of lower weights of the same denomination, in succession, until the balance assumes a state of equilibrium, or the weights are a little too small ; in the latter case, the weights of the next lower denomination should be tried in the same way. For example, supposing we wish to deter- mine the weight of a platinum crucible, which was ulti- mately found to weigh 16*715 grams. We place on the weight pan the 10-gram weight and release the beam, this 12 THE GAS ENGINEEK'S LABOKATOKY HANDBOOK. proves too little ; we then add another 10-gram weight, this proves too much ; we substitute the 5-gram weight for the 10-gram, the weight is again too little; we add the 2-gram weight, this is again too much ; we substitute one of the 1-gram pieces for the 2-gram piece, the weight is too little ; we add the 0*5 gram, still too little; also the 0'2 gram, still too little, although it will now be noticed that the rate of the vibration of the beam becomes much slower, indicat- ing that the balance is rapidly approaching equilibrium. We have found that 16*7 grams is not quite sufficient to balance the crucible, we therefore add 1 gram, too much (the pointer swings with increased energy in the opposite direction); we substitute 0*05 for the 0*1 grain, still too much ; we then try the 02, the pointer now vibrates much more slowly, but still indicates that the weight is too great ; we substitute 0*01 for the 0'02 gram, the pointer swings with the same slowness, but shows an equal deflection to the opposite side of the zero; we add 0'005 gram, the pointer now makes excursions of equal amplitude, that is, the balance is in equilibrium. The position of the pointer at the end of the vibrations, should be read off several times in succession, in order to make sure that the balance is in equilibrium ; the beam is then arrested, and the aggregate value of the weights needed determined, both from the vacant spaces in the box, and from the denominations marked on the weights them- selves. This method of double reading should always be employed, so as to prevent false entries. It should be particularly noted, that whenever an ex- change of weights is necessary, the motion of the beam must be arrested. Under no circumstances must anything be added to, or taken away from the pans when the balance is in motion. By not attending to this precaution, the knife- edges would soon become worn, and the balance would consequently be inaccurate. The balance-case should be closed while the rider is being used, and the final observa- DIEECTIONS FOR WEIGHING. 13 tion of equilibrium must be always conducted in the closed case. In weighing out substances for analysis, they should not be placed direct on the scale-pan, but weighed in vessels of glas-i, porcelain, or platinum. Powders may be weighed out fron a stoppered weighing bottle, the difference in weight between the first and second weighings representing the anu ant of powder weighed out. Hygroscopic substances ma} be weighed in watch-glasses with ground edges, and hel< together by a brass clip. Hygroscopic, efflorescent and vol; tile substances in fact, all substances which gain or lose weight on exposure to air, should be weighed in closed vest els. i substance should never be weighed when warm. By plac ing a warm object on the balance-pan, the indications of 1 he instrument are greatly affected. The ascending air cur -ent produced by the heat of the warm substance, acts aga nst the pan containing the substance, and also against the beam above it, and the substance appears to weigh less tha:i it ought to weigh. The warm air, after a time, also affects the portion of the beam against which it impinges, and by increasing its length disturbs the equality in the arms. Further, all substances condense upon their surface a certain amount of air and moisture, the weight of which depends upon the temperature : this also is an error in the same direction, and causes a body when warm to weigh less than when cold. Apart from the question of temperature, a vessel, say a platinum crucible, when rubbed with a dry cloth and weighed immediately afterwards, always weighs sensibly less than after half-an-hour's exposure to the air of the balance-case, owing to the condensation of the air, with its accompanying aqueous vapour, upon its surface. It is ad- visable, therefore, to allow the crucible, if newly wiped, to remain in the balance-case some little time before weighing. 14 THE GAS ENGINEER'S LABORATORY HANDBOOK. SECTION II. SAMPLING MECHANICAL DIVISION DRYING AND DESICCATION SOLUTION AND EVAPORATION PRECIPITATION FILTRATION AND TREATMENT OF PRE- CIPITATES. Sampling. 10. The first operation in the conduct of an analysis is that of sampling, to which too much attention cannot be paid, as on the sample being a true representative of the whole bulk, depends the value of the analysis. For instance, in taking a sample of oxide of iron from a heap, if we were to simply take a sample from the outer portion, this would not give a fair indication of the bulk ; by reason of that part of the heap being acted upon by the atmosphere, it would probably be found to be much drier as compared with the interior portion, and thus, at any rate as far as the amount of moisture was concerned, the sample would not represent the average of the heap. The proper mode of procedure in the case above cited, is to dig well into the heap, and taking small portions from every part, mix these well together, afterwards selecting still smaller portions from this for the final sample ; these latter portions being also turned over and well mixed. The sample should be immediately transferred to a tightly stoppered bottle, and labelled with the description and date of sampling. In sampling coal or cannel, all pieces should be broken up into lumps not larger than walnuts, and well turned over in the manner just described. Mechanical Division. 11. The next operation is that of mechanical division. In order to prepare a substance for analysis, that is to say, to make it in a suitable condition for the action of solvents, MECHANICAL DIVISION. 15 &c., it is generally necessary to have it in a minute state of subdivision, as this will create abundant points of contact for the action of the solvent. Substances are obtained in thi^ fine state by being pounded or triturated. The opera- tion of pounding is generally performed in mortars, the ess* ntial factor in the operation being, that the material of wh ch the mortar is composed, is harder than the substance it s wished to reduce to powder, otherwise the substance ma 7 get contaminated with the material composing the mo -tar. For the ordinary salts, and substances not possess- ing any very considerable degree of hardness, a porcelain, gl; .38, or Wedgwood-ware mortar may be used, but for mi lerals, and siliceous materials such as fire-bricks, an agate me rtar is indispensable. In pounding coal, an iron mortar wi;l be found useful. In the operation of pulverisation, as in most others, much tin e and trouble may be saved by a systematic method of we rking, and an equally great loss incurred by careless manipulation. Some substances fall into small fragments wl en simply touched by the pestle, and are easily reduced to fine powder by a grinding motion, while others require a grc at deal of pounding, and necessitate care in order to avoid dispersion of the fragments. This may be prevented by striking with the pestle straight down, carefully avoiding lataral blows, as such blows chiefly cause the particles to fly about. After the substance has been tolerably reduced, it is more advantageous to substitute a movement between that of a blow and a grind, for the directly downward motion of the pestle. This is effected by grasping the handle firmly in the hand, the striking portion of the pestle pointing somewhat inwards, and then striking downwards, drawing the pestle towards the operator's body. When by this means a certain degree of fineness has been obtained, it is advisable to entirely substitute trituration for blows, the pestle then travelling in turn, over the whole of the lower portion of the inside of the mortar. 16 THE GAS ENGINEER'S LABORATORY HANDBOOK. Where the matter adheres to the sides of the mortar, it must be scraped down towards the centre with a spatula. It is necessary to have only a moderate amount in the mortar at one time, as otherwise the fragments are protected by each other from the action of the pestle. The operation of pounding may be considerably assisted by that of sifting. For this purpose, the sets of sieves sold by dealers in chemi- cal apparatus will be found useful. The following device may also be employed for the same purpose. A piece of clean and dry muslin is tightly tied round the top of a small beaker. On this is lightly poured some of the powdered substance ; on gently tapping the muslin with a glass rod a portion falls through, the remainder, which is too large for the pores of the muslin, is returned to the mortar, and again pounded until it is likewise fine enough to pass through. Impurities may be removed from porcelain and Wedg- wood-ware mortars by rubbing in them a little sand saturated with a strong acid or alkali. It is seldom that any dirt is found to resist the action of nitric or sulphuric acids, or caustic potash. Desiccation. 12. Before an analysis can be proceeded with, it is necessary that the substance under examination is free from all unessential constituents. The most frequent of these is moisture, by which is meant, the moisture in excess of that which is necessary to the constitution of the body. In drying a substance, therefore, the aim generally is, to get rid of the adherent moisture without interfering with the chemi- cally combined water, or with any other constituent. The methods of effecting the operation in each particular case, will depend upon the temperature to which the substance can be exposed without being decomposed. Substances con- taining water of crystallisation may frequently have their adherent moisture removed by pressing repeatedly between DESICCATION. 17 folds of filter paper. Another easy but rather slow method of dosiccation, is to expose the substance in a closed vessel, calkd a "desiccator," over a pan containing calcic chloride, strong sulphuric acid, or other hygroscopic body. A partial vacvum in the desiccator accelerates the operation. Many siibctances may be freed from water at a temperature of 100 C. (212 F.). A thin layer of the substance (powdered if soli* L), spread on a watch-glass, is exposed to a gentle current of ; ir in a steam -jacketed oven, known as a water oven, FIG. 7. Fig. 7. The apparatus is made of copper, and is doubly- cased throughout with the exception of the door. The method of using it is to fill it to about three-fourths of its height with water ; it is then heated by means of a Bun sen burner, with the effect that when the water boils, the upper part of the hollow casing becomes filled with steam, and the temperature inside the oven approaches 100 C. A gentle current of air at the same time passes through the oven. The distillation of water may frequently be combined with the heating of water ovens. When a temperature higher than 100 C. is required for 18 THE GAS ENGINEER'S LABOR ATOBY HANDBOOK. the drying of a substance, an air oven is employed. An air oven is almost identical in construction with a water oven, the main difference being, that heated air takes the place of boiling water and steam within the jacket. Since the bottom of the oven comes into direct contact with the flame, it follows that it is usually hotter than the air in the interior, consequently the substance must not rest directly on the bottom. A convenient stand may be made by turning down the wire ends of a pipe-clay triangle at right angles to the plane of the triangle. A thermometer passing through a cork in the top of the oven should be provided, so as to register the temperature of the interior. With a little ex- perience any required temperature may easily be obtained by regulating the tap on the gas supply, and after the heat has once been regulated the temperature will remain con- stant for several hours. Kegulators are also sold by dealers in chemical apparatus, by attaching which the temperature may be kept constant for any length of time. Many substances after drying readily take up moisture again, consequently some means must be adopted to prevent them from absorbing moisture from the air while cooling, or waiting to be weighed. Dried apparatus also which is left for -weighing, should be preserved from the deposition of water upon its surface. The usual method of preventing the access of atmospheric moisture is to enclose them in an air-tight vessel, the air in which is kept dry by exposure to a hygroscopic substance, sulphuric acid or calcic chloride being the reagent commonly used. One form of this instru- ment is shown in Fig. 8. It consists of a broad glass jar contracted in the middle. 1 he upper portion contains the substance to be weighed, the lower portion the desiccating agent. A circular piece of perforated zinc forms a bottom to the upper portion, and on this a bent pipe-clay triangle or other support is laid. The body to be cooled and weighed is placed on this support. The upper rim of the desiccator is DESICCATION. 19 ground, and a flat lid, also ground, fits truly on to it, so that when the upper rim is lightly greased the flat lid fits air- tight. By this device, an air-tight chamber is obtained, in wl ich a body may cool in dry air, and be kept from contact wi :h moisture for an indefinite period. FIG. 9. Solution of Solids. 13. It is usually necessary before analysing a solid substance to have it in a state of solution, and in order to dis,solve it the substance should be in a fine state of sub- division, being powdered, if necessary, as described in (11). If the substance is soluble in water, a convenient method of dissolving it is shown in Fig. 9. The substance is put into a beaker containing distilled water, and the beaker is then placed on a piece of wire gauze on an iron tripod over a Bunsen burner. Should " bumping " ensue, the contents of the beaker may be stirred occasionally with a glass rod. In order to guard against loss by spurting, the beaker is covered with a clock-glass. o 2 20 THE GAS ENGINEERS LABOEATORY HANDBOOK. If the substance is likely to effervesce whilst dissolving, as is sometimes the case, particularly on adding an acid to carbonates, the following methods may be employed : (1) The substance is placed in a conical flask, the solvent being added through a small funnel which is left in the mouth of the flask, closing it during the process of solution ; by this arrangement loss by spurting is prevented, but free passage of the gas given off is allowed. (2) Another method is to place the substance in a round flask inclined at an angle of about 45. In this case also loss by spurting is prevented, as the drops thrown up during effervescence will be prevented" from escaping by striking the inside of the flask, at the same time the gas will have a free exit. It is necessary to exercise care in the selection of vessels intended to be used for purposes of solution or evaporation, so that they will not be acted upon by the solvent. Evaporation. 14. The process of evaporation is employed either to concentrate a solution, or to drive off the whole of the liquid by the aid of heat. It is usually conducted in basins made of porcelain, platinum, nickel, or silver, although some- times a beaker or crucible is employed. Liquids are usually evaporated over boiling water in vessels known as water- baths ; the heating agent in this case being steam. A simple form consists of a cylindrical copper vessel, rather more than half filled with water, and heated by a Bunsen burner. The upper portion of the vessel can be fitted with round rings of flat sheet copper, the diameters of the rings gradually diminishing in size. The top of the vessel may thus be made to fit and support vessels of different sizes. A more elaborate water-bath is shown in Fig. 10. The top of this arrangement is provided with holes of different sizes, so as to fit vessels of varying dimensions. These holes when not required for use are covered with lids. Should a vessel EVAPOKATION. 21 employed for purposes of evaporation require to be after- w irds weighed, it should rest on a glass ring, and not come in to direct contact with the copper edge of the water-bath as al ove described, as such contact is likely to stain it. The u] >per part of a beaker, provided it is of the proper diameter, m ikes a very good support. A broken or cracked beaker may easily be converted into si ch a support, by cutting round it with a diamond, or k iding a crack round it about an inch below its upper edge. A water-bath of a simple form may be easily improvised by FIG. 10. partially filling a beaker with water and heating it over a Bunsen burner. Should the water in the beaker be in- clined to bump while boiling, the bumping may be prevented by placing a few small pieces of paper into the water. When a higher temperature than 100 C. is required in order to evaporate a liquid, one of the following methods may be employed. The vessel may be placed in a sand-bath, which is simply a shallow, saucer-shaped vessel of thin sheet iron filled with fine dry sand, heated by a Bunsen burner. By immersing the vessel to a depth of an inch or so in the sand, a uniform heat is imparted to the bottom of the vessel ; or rf'V, ' rt Kl\ 22 THE GAS ENGINEER'S LABORATORY HANDBOOK. the vessel may be heated by direct contact with a naked flame, the rose head being placed on the Bunsen during the operation. Another method is to support the vessel on a piece of wire gauze, or on a sheet-iron plate. The latter plan possesses the advantage that the rate of evaporation may be modified by moving the vessel, say a beaker, to a greater or lesser distance from the part of the plate directly in contact with the flame. Where effervescence is likely to occur, it is advisable to use an evaporating dish, covering it with a large inverted funnel. When conducting an evaporation it is advisable to observe the following precautions : (1) To prevent the liquid from actually boiling, or there will be a risk of loss by spurting. (2) When conducting an evaporation over a naked flame or sand-bath, towards the end of the operation, the vessel should, if possible, be transferred to a water-bath, as a liquid, even when heated below its boiling point, is likely to spurt when it is in a thick and pasty state, by the separation of solid matter. (3) Sometimes it happens that the liquid creeps up the side of the vessel during the process of evaporation. This particularly applies to solutions containing salts of ammonia. In the event of this occurring, the incrustation forming near the edge of the vessel should be detached, and pushed down by the aid of a glass rod. Precipitation. 15. The operation of precipitation is frequently made use of in quantitative analysis, and is employed when it is required to convert a substance in solution into an insoluble form, in which condition it may be collected and weighed. The operation is usually performed in glass beakers, on account of the facility with which the contents of the same may be transferred either to the filter, or to the vessels in PRECIPITATION. 23 which, they are to be weighed. In the event of having to de il with strongly alkaline liquids, or where the latter have to be heated for some time, porcelain basins are more suitable. Precipitation is usually effected with hot solutions, and but ra -ely in the cold. If the solution is boiling, the beaker co itaining it is covered with a clock-glass, so as to avoid any loss by spurtirg. The separation of the precipitate fi ided. The solution to be filtered should never be poured dii ectly into the funnel, but down a thin glass rod, Fig. 13, FIG. ]3. the stream of liquid being so directed as to fall against the side of the filter, and not into the apex, as that would en- danger the bursting of the paper, and cause loss by splashing. A liquid retaining a precipitate in suspension, should never be poured higher than within half-an-inch of the margin of the filter. The rim of the vessel containing the liquid to be filtered should be slightly greased ; this prevents the liquid running down the outside of the vessel. When not in use, the rod employed in pouring the liquid into the filter is placed in the vessel containing the precipitate, or, if it is 26 THE GAS ENGINEER'S LABORATORY HANDBOOK. necessary that the latter should not be disturbed, in a little flask or beaker, which is afterwards rinsed out into the filter so soon as the whole of the precipitate has been trans- ferred on to the filter. During the progress of the filtration, it is advisable to cover the various vessels with glass plates, so as to prevent dust from falling into them. A special plate with a small hole at the side will be necessary for covering the beaker receiving the filtrate, so as to admit the stem of the funnel. It occasionally may happen that some of the precipitate passes through with the filtered solution, barium sulphate and calcium oxalate, when freshly precipitated, being es- pecially liable to do so. Should this occur, it may be necessary to pass the precipitate through the filter two or three times before it comes through quite clear. A clear filtrate may, however, be obtained in most cases by well boiling the mixture before filtering, which causes the finely divided particles of the precipitate to coalesce together. The presence of particular saline matters in solution also some- times prevents a precipitate passing through ; ammonium chloride, for example, exercises this property with barium sulphate. As previously remarked, the operations of filtration and decantation are usually combined in the following manner. After decanting the clear liquid, a fresh quantity of distilled water, or other washing liquid, is poured upon the precipitate in the beaker, and the contents of the vessel agitated. The mixture is then allowed to subside until the liquid is clear again, and the clear liquid is poured through the same filter. These operations are then repeated several times. Finally the precipitate is transferred to the filter-paper by means of the spray from the wash-bottle, Fig. 14, the portions of the precipitate adhering to the beaker being removed by means of a glass rod capped with india-rubber tubing. Should it be found impossible to remove the last traces of the pre- cipitate by means of the latter contrivance, the precipitate FILTRATION AND WASHING OF PRECIPITATES. 27 usually be dissolved and re-precipitated. When the wh( le of the mixture has been poured on the filter, all the solution should be allowed to pass through before com- mercing to wash the precipitate on the filter. The object of \ -ashing, is to remove all soluble matters from the filter, and the operation is of great importance in analysis, as a pre< ipitate imperfectly washed would weigh too high. For inst ince, in the determination of sulphuric acid by precipi- tati >n with an excess of barium chloride, if the excess of bar um chloride is not removed by washing, we should be wei piling, not the weight of barium sulphate alone, but a nm ture of barium sulphate and chloride, whi 3h would, of course, give an erroneous resi It. Precipitates are generally washed by me; ns of the wash-bottle, shown in Fig. 14. Thi 3 is simply a flask made of thin glass for the purpose of standing heat, fitted wit i a sound cork, through which two glass tubos pass. One, which is bent at an angle of 45 , has a piece of india-rubber tube attached to it, and enters the flask only a short distance ; the other, which reaches nearly to the bottom of the flask, has its upper end drawn out to a fine point, and bent at a suitable angle for directing the stream of water. In order to hold the wash-bottle with convenience, a coil of thick string may be wrapped round its neck. On filling the wash-bottle with water and blowing through the india-rubber tube, a fine stream of water may be directed on any part of the filter. The precipitate remaining on the filter is now well washed (generally with hot distilled water, as hot water filters more rapidly than cold), the current of water being applied, first towards the upper part of the filter, and then directed gradually downwards. When the filter has in this manner been nearly filled up with water, the whole should 28 THE GAS ENGINEER'S LABORATORY HANDBOOK. be allowed to run through before adding any more ; the operation is then repeated until the precipitate is sufficiently washed, which is known by testing some of the final washings. If the precipitate, while standing in the filter, cakes together into lumps, these must be broken up by directing upon them a strong current of water from the wash-bottle, as otherwise the water would not penetrate them, and some of the soluble matter would escape removal. Accelerated Filtration. 17. The process of washing the precipitate as described above is frequently very tedious, hence various expedients have been devised in order to accelerate the process. One of the simplest means of attain- ing this result is to attach a bent tube to the funnel, as shown in Fig. 15. When the liquid descends and fills the longer limb of the tube, a considerable " pull or vacuum " is exerted on the contents of the funnel, consequently increasing the rate of filtration. The most effective method, however, for inducing rapid fil- tration is by means of the filter-pump, or other appliances which remove the atmo- spheric pressure from beneath the filter more or less completely, and yet leave that pressure in its entirety to act on the surface of the liquid in the funnel. A handy form of filter-pump is that invented by Geissler, and shown in Fig. 16. It is made of glass. The most con- venient place for fixing this apparatus is over the sink of the water supply. The upper vertical tube of the filter-pump should be attached, by means of a stout india-rubber tube, to a high-pressure water service, and in order to withstand the FIG. 15. ACCELERATED FILTRATION. 29 pressure of the water, the india-rubber tube should be firmly bon nd with strong copper wire to the water tap and filter- pump. The action of the pump is as follows. As shown in the drawing, there are three tubes fixed into the chamber in the upper portion of the apparatus. The upper vertical tube conveys the water from a high- prc ssure water service. This, in passing through thi contracted end of the tube into the opening im nediately below it, entangles air from the cham- ber , and carries it away through the lower vertical oui let pipe. Air is consequently pulled through th( side horizontal tube, and a vessel placed in coi nection with this tube will consequently be gr; dually exhausted of its air. The side tube of th( pump is connected with a tube from a small W< >ulffe's bottle. This tube passes through the cork of the bottle nearly to the bottom ; the other tube of the bottle passes just through the cork in its FIG. 16. ne< 'k. This tube is connected in its turn by means of india-rubber tubing, either with another Woulffe's bottle containing the furmel, or with the side tube of a filtering flask. The first Woulffe's bottle is employed in order to prevent any of the water from the water service from entering the Woulffe's bottle containing the filter, or filtering flask, when the water supply is suddenly checked. In the event of this occurring, the water from the filter- pump would be drawn back into the side tube by the partial vacuum created ; but instead of entering the filtering flask it is arrested in the first Woulffe's bottle. In this way the water is prevented from mixing with the filtrate in the second Woulffe's bottle or filtering flask, and as soon as the filter-pump is re-started, the water in the Woulffe's bottle is rapidly sucked out. By reducing the pressure underneath the filter in the manner indicated, the pressure acting on the upper surface of the paper is frequently greater than the unsupported 30 THE GAS ENGINEER'S LABORATORY HANDBOOK. paper can bear without breaking; it is therefore necessary to support the apex of the filter-paper by a platinum cone, which is obtained of the desired shape in the following manner. A circular piece of writing paper, 10 or 12 centi- metres in diameter, is folded in the shape of a filter, and placed in the funnel so as to fit accurately to the sides of the same, particularly near its apex. It is kept in position by a few drops of sealing-wax, and is saturated with oil by means of a feather, care being taken that no drops of the oil remain at the apex of the paper cone. A thin cream of plaster of Paris is then poured into the paper mould, and a small wooden handle is inserted into it just before the plaster solidifies. In a few hours the plaster cast will be dry enough to be removed from the funnel, together with the oiled paper. The outside of the same paper cone is now thoroughly oiled and inserted into a small capsule, or similar shaped vessel, of 4 or 5 centimetres in height, filled with a cream of plaster of Paris. As soon as the outer mould is dry, the plaster cone is removed, and the paper rubbed off FIG. 17. it. In this way a solid cone fitting into a hollow cone is obtained, both of which have the angle of inclination of the funnel. A piece of platinum foil, of such a thickness that one square centimetre weighs 0*15 gram, is cut into the shape and size shown in Fig. 17 ; it is divided by a pair of scissors along the line a 6, as far as the centre a. The foil is then held in the Bunsen flame for a few minutes to render it pliable, and placed against the plaster cone, so that the point a is at the end of the cone; the side a, 6, d, is folded against the cone, and over this is folded the remainder, a, 6, c, so that the foil also becomes a cone, the sides of which have exactly the same inclination as those of the plaster cast, and also of the funnel. The shape of the platinum cone may be completed by dropping it into the hollow mould, and pressing it down by ACCELERATED FILTRATION. 31 meons of the plaster cone. If, after use, the cone should happen to get out of shape, the correct shape may at any tim 3 be obtained again, by simple pressure between the two con 38. The platinum cone should not allow any light to pas 5 through its apex. When properly made it will support a fi ter filled with liquid under the pressure of an atmosphere wit bout the paper breaking. The small space between the fol< s of the foil is quite sufficient to allow of the passage of a rap Id stream of water from the filter. The platinum cone is fitt -d for use in the following manner. A funnel is selected ex;; 3tly fitting the cone, which is placed in the funnel, and thi ti a dry filter-paper, folded in the ordinary manner, inserted. The filter-paper is pressed so that the cone and pa} er fit the sides of the funnel closely. The paper is then we ;ted by pouring in a little distilled water. The over- lapping edges of the filter-paper are pressed down by the finders, and any air-bubbles between the funnel and paper art also carefully removed by pressure. The funnel is then fitt ed, by a perforated india-rubber cork, either into the neck of i clean, strong, conical filtering flask, or into one tubulure of i Woulffe's bottle of suitable size ; the side tube of the filtering flask, or the remaining tubulure of the Woulffe's bottle being placed in connection with the filter-pump. Afier making the connection between the pump and the vessel carrying the funnel, but before commencing filtration, it is advisable to fill the filter with distilled water and to start the pump. If any air is observed to be sucked down between the filter and funnel, this should be stopped by gently pressing the moistened edges of the paper against the funnel. In commencing filtration the pump should be started very gently, particularly if the precipitate is a fine one, as in that case particles of the precipitate may be at first easily drawn through the paper. To filter a precipitate by this apparatus, the supernatant liquid is cautiously poured on to the filter, the action of the pump is gently started, and as the liquid flows through into the filtering flask, fresh portions 32 THE GAS ENGINEER'S LABOKATOEY HANDBOOK. are added until the whole has been decanted. The precipitate is then transferred in the ordinary manner, and washed by the addition of water from an open mouthed vessel and not by a jet from a wash-bottle. The liquid in which the pre- cipitate was originally formed, together with that necessary to transfer the precipitate to the filter, should be allowed to flow away completely before the operation of washing is commenced. As soon as the precipitate is drained, but before any channels begin to form in it, the filter should be filled up with distilled water, poured cautiously down the side of the funnel. When this lot of wash- water has drained away, the suction is continued until the precipitate is observed to shrink, when the filter is again filled up. This process is to be twice repeated, after which the precipitate may be drained almost dry by using the filter-pump a few minutes longer. This method of filtration and washing is exceedingly rapid as compared with the old plan. A further advantage attends the use of the filter-pump, that is, the condition of the precipitate after filtration. Precipitates such as ferric oxide and alumina are left so dry that, without further drying, the precipitate wrapped in the filter-paper may be placed in the crucible over the Bunsen flame, and, after cautiously charring the paper, the whole may be ignited without risk of loss by particles being projected out of the crucible. Drying of Precipitates. 18. The next operation is that of drying the precipitate, which is effected as follows. The water remaining in the stern of the funnel is first removed by means of a filter-paper, and the mouth of the funnel is then covered with a moistened filter ; on drying, the paper adheres to the rim of the funnel, and in this way protects the precipitate from dust. The funnel is then placed in the water oven (12) and kept there until the paper and DRYING OF PRECIPITATES. 33 precipitate are quite dry. This method of drying the pre- cip-.tate possesses the advantage that, when the precipitate is (ontracting, it detaches itself from the filter-paper, so that precipitates of a curdy or gelatinous nature, sue h. as silver chloride, or the oxides of iron and chromium, may be almost completely shiken out of the funnel, direct into the crr.oible in which they are to be weighed. Tli is easy separation of the dried precipitate fro n the paper is a material help to accuracy in obtaining the weight of the precipitate. Ai other method of drying, is to place the fu] nel in a tin cone, supported on a piece of wi -Q gauze over a Bun sen burner, Fig. 18 ; but thi s method is open to the objection that it may cause charring of ihe paper, or fracture of the funnel, besides possessing the further disadvantage of causing the precipitate, by reason of the way it is heated, to adhere to the paper. Ignition of the Precipitate. 19. Since the precipitate is required to be in a perfectly dry state before it is in a fit condition for weighing, it is usually necessary to ignite it. The filter-paper itself is also buint and converted into ash, and the weight of the result- ing ash, which is usually extremely small, is subtracted from the total weight of the ignited precipitate and ash ; it is therefore necessary to determine the average weight of the ash of the filter-papers employed, which is done as follows: Six filter-papers are selected from different parts of a packet of filters ; these are then carefully folded and burnt, one by one, to a perfectly white ash in a weighed platinum crucible, the crucible and ash being afterwards weighed. The weight of the ash thus obtained, divided by six, will give the amount of ash from one filter-paper. The -vessels employed in the ignition of precipitates are crucibles 34 THE GAS ENGINEER'S LABORATORY HANDBOOK. made of platinum, nickel, or porcelain. A platinum crucible is to be preferred whenever it is practicable to use it, as it is heated more rapidly and uniformly than one made of porcelain, and is better adapted to resist the action of a coal- gas flame than one made of nickel ; but platinum cannot be used for the ignition of the chlorides of silver or lead, and some oxides, sulphides, and phosphides are liable to injure the metal also. The crucible is usually sup- ported on a pipe-clay triangle, Fig. 19. The temperature to which a precipitate should be exposed, and the duration of the FIG. 19. ignition, varies with the nature of the pre- cipitate, but in all cases the heat should be applied gradually. This precaution is especially necessary when the substance is not perfectly dry, or is in the form of a very light powder, such as silicic acid. When heating precipitates which are capable of reduc- tion, care must be exercised that the flame does not come into contact with the interior of the crucible, or possibly reduction of the substance may take place. A simple method of igniting precipitates requiring a comparatively low temperature, is to place the crucible upon a triangle, resting upon the chimney of a Wallace Argand burner minus the gauze top. When the gas as it comes from the ring at the bottom of the burner is lit, it causes a current of heated air to play against the sides of the crucible, and in this way the crucible may be uniformly heated to any tem- perature desired, up to a red heat. This method of heating is specially useful in the determination of calcium as CaC0 3 . The ordinary method of ignition, however, is by means of a Bunsen burner, supplemented in some cases by the blow-pipe, the crucible being placed on a pipe-clay triangle, resting on the ring of an iron tripod, or retort-stand. Precipitates may be ignited with, or apart from, the filter-paper, according as to whether the combustion of the IGNITION OF THE PRECIPITATE. 35 filt T-paper has any reducing action on the precipitate or not If :he combustion of the filter-paper has no reducing action upi n the precipitate, the latter may be ignited as follows. Th 3 filter and precipitate, first being thoroughly dried in the wa er oven, the filter-paper (which still contains the pre- cip tate) is folded together, and the free edge turned down so as o enclose the precipitate in a small compass. The filter is tin i deposited in the crucible, and the lid placed on the latter so ts to nearly cover it, leaving only a small space for the esc ,pe of the gases which are given off by the ignition and COD bustion of the filter-paper, Fig. 21. Heat is then applied, verf moderately at the commencement, As soon as the flai ie due to the combustible gases ceases to appear, the lid ma r be removed, and the heat increased until the tempera- tur ) required is obtained. When the filter-paper has been eot rely converted into a white ash, the burner is taken aw; y, and as soon as the crucible has cooled down a little, it is i emoved to the desiccator by means of a pair of clean cru< -ible tongs. The lid of the crucible is then ignited until it if- clean, placed upon the crucible, and the whole is weighed wh( n cold. .:~t sometimes happens that the filter-paper is difficult to completely incinerate, black particles of carbon stubbornly rem lining. In this event, the combustion may be accelerated by wringing the portions of unconsumed filter-paper into contact with the hottest portion of the crucible, by means of a stout platinum wire fused into a piece of glass tubing; or a small blow-pipe flame may be used, so as to produce a higher temperature. Another way of helping the combus- tion of the carbon, is to moisten the ash with a drop of a saturated solution of ammonium nitrate. Precipitates such as A1 2 3 , 3H 2 O ; or Cr 2 3 , 3H 2 ; which have been washed and drained by means of the filter-pump, may be ignited without being previously dried in the water oven. In these cases, it is necessary to cover the crucible with its lid, until all the moisture and gases have been driven off, and to apply 36 THE GAS ENGINEER'S LABORATORY HANDBOOK. the heat very gradually, commencing with a small flame, placed at some distance below the crucible. When the combustion of the filter-paper is liable to exert a reducing action on the precipitate, and also in cases in which it is desirable to keep the precipitate apart from the ash, the precipitate is ignited apart from the filter, in one of the following ways. (a) The dry precipitate and paper are taken from the funnel, and the precipitate is carefully shaken out of the paper into the crucible standing on a piece of glazed paper, so as to catch any particles of precipitate which may fall outside the crucible. The glazed paper should be dark for light-coloured, and white for dark - coloured precipitates. The portions of the precipitate adhering to the filter are loosened by rubbing its sides together, care being taken that FIG. 20. the surface of the paper is not destroyed by so doing, other- wise filaments of paper are liable to contaminate the preci- pitate, and may impair the accuracy of the result, by escaping being burnt in the subsequent ignition, or if burnt, may alter the composition of the precipitate. The precipi- tate, detached as above described, is then added to that already in the crucible. The filter-paper is then folded up as in Fig. 20, and afterwards rolled into a coil. The paper should be folded in such a manner that the soiled half of the filter is in the centre ; there is then less risk of loss from projection, or from the precipitate fusing to the platinum wire, which is wound round the paper in a spiral form. The platinum wire employed should be about 10 inches long and fairly stout, and, for convenience of handling, should be fused into a piece of glass tube. The filter-paper is then ignited by means of a Bunsen IGNITION OF THE PKECIPITATE. 37 burner, which, is held in one hand, the paper being held in th ) other hand directly over the crucible standing on a piece of glazed paper. The paper is maintained in a glowing con- di ion by being occasionally placed in the flame of the burner, ui til the carbon of the paper is entirely consumed. A slight ta ) of the platinum wire on the edge of the crucible will ca ise the ash to fall into the latter, when the ignition may b 1 completed in the usual way by means of a Bunsen burner, FIG. 21. at first gently with the lid on, and afterwards more strongly with the lid removed. The crucible is then placed in the desiccator and weighed when cold. It is then heated a second time and again weighed, to ascertain that its weight is constant. In every case before proceeding to use a crucible for the purpose of igniting a precipitate, it should first be heated, placed in the desiccator, and weighed, so as to obtain the weight of the empty crucible. 38 THE GAS EXGINEEK'S LABOKATOKY HANDBOOK. (b) Another method of igniting the filter-paper and pre- cipitate, is to burn the paper on the lid, and the precipitate in the crucible, both being ultimately weighed together. The details of the method are as follows. The precipi- tate is shaken out into the crucible, the last traces being detached from the paper as completely as possible as in the previous method. The filter-paper is then cut up into small strips by means of a clean pair of scissors ; these strips are then incinerated, one by one, on the inverted lid of the crucible, which is supported on a pipe-clay triangle. When all the strips have been burnt, the resultant ash is added to the precipitate in the crucible, and the ignition completed as already described. As previously remarked, the intensity and duration of heat to which a precipitate should be exposed, depends on the nature and quantity of the precipitate, but as a general rule from five to ten minutes at a low red heat will be sufficient. A few words on the care of platinum vessels are necessary. Platinum crucibles, when red hot, should never be touched with brass tongs, or placed in brass rings, as these produce black stains on the metal. Should a platinum crucible become soiled, it may be cleaned by rubbing it with moist sea-sand, the particles of which, being rounded, will not scratch the metal. The nature of the sand should first be ascertained by examining a few grains of it through a magnifying glass ; if they appear to be rounded, resembling pebbles, the sand is suitable, but if angular the sand should not be used. Should the stains not disappear on rubbing with sand, a little bisulphate of potassium may be fused in the crucible ; this, when cold, is removed by boiling it with water, and the surface of the metal finally polished with sea- sand. Collection of Precipitates on Weighed Filters. 20. It occasionally happens that it is not possible to expel the moisture from a precipitate by ignition, before COLLECTION OF PRECIPITATES. 39 v, eighing. It is therefore necessary, in such a case, to weigh tl .e precipitate on the filter on which it has been deposited ; consequently, we require to know the weight of the filter piper itself. The paper is folded in the usual manner, p aced in a stoppered weighing tube, or between a pair of vi ell-ground watch glasses with clip, Fig. 22, and heated in t i.e water oven for an hour or so. The stoppered tube or v atch-glasses containing the filter paper are then allowed t> * cool in the desiccator, and weighed when cold. The filter i: then placed in the funnel, and the precipitate deposited < i it in the usual manner, the tube or watch-glasses FIG. 22. r jmaining in the balance case while this operation is pro- c ceding. When ready for drying, the paper and precipitate a re first of all dried in the funnel ; the filter is then taken o it of the same and placed in the weighing tube or between the watch-glasses, dried for some hours in the water oven, aid repeatedly weighed until a constant weight is obtained. Another plan is to weigh two filters of equal size, A and B, against each other, and mark the difference in weight on B. The precipitate is collected on A, the filtrate and washings being allowed to pass through B ; both are dried and weighed against each other, and the original difference in weight allowed for. 40 THE GAS ENGINEER'S LABORATORY HANDBOOK. PAKT II. SIMPLE GKAVIMETKIC DETEKMINATIONS. THE following simple gravimetric determinations are selected principally, with a view to their applicability to the esti- mations most frequently required in gasworks. In most of the examples, the substance selected for analysis is of known and definite composition. The accuracy of the result ob- tained can therefore be checked, by calculating from the formula of the substance, the percentage of the constituent present. The substances recommended for analysis in these exer- cises, are generally soluble in water, but should a substance requiring estimation be found to be insoluble in water, it will be necessary then to ascertain, by experimenting on a small portion of the finely powdered substance, what are the most suitable solvents. These will usually be found to be either hydrochloric or nitric acids in the order named, or possibly a mixture of the two. If treatment with acids does not dissolve it, the substance may generally be decomposed by fusing it with about six times its weight of a mixture of sodium and potassium carbonates ; it is then extracted with water, and the insoluble residue dissolved by treatment with acid. 21. In order to make the following exercises of any value, it is necessary that) the substances chosen for the ex- periments should be quite pure ; a few hints on the prepara- tion of pure substances will therefore be of service. The purity of a soluble and crystallisable chemical com- pound can usually be ensured by repeated crystallisation SIMPLE GKAVIMETKIC DETERMINATIONS. 41 fr )m a suitable medium. Owing to the fact that the in ipurities are left behind in the mother liquor by this mode of treatment, many commercial salts may be purified by one 01 two crystallisations from water. The following is the m jthod of procedure in the case of some substances which a i e employed in the examples given for practice. Potassium aluminium sulphate (A1K(S0 4 ) 2 ,12II 2 0); cop- p r sulphate (CuSO 4 ,oH 2 0) ; barium chloride (BaCl 2 ,2H 2 0) ; n agnesiurn sulphate (MgS0 4 ,7H 2 0). The coarsely crushed commercial salt is dissolved in a 1> -aker or evaporating basin, the contents of the vessel being h :ated and well stirred, until the solid no longer dissolves, S' as to obtain a hot and nearly saturated solution of the solid. r l he hot liquid is then decanted ft om the excess of solid, into a c ystallising dish. In the event, h jwever, of any suspended solid n atter remaining in the liquid, tJ le latter must be passed through a filter - paper contained in a funnel, which is kept hot by n eans of a jacket of hot water. The apparatus employed to at- tain this result is shown in Fig. 23. It consists of a copper vessel, somewhat resembling a funnel in shape, and double walled. In communication with the interspace of the funnel-shaped portion of the apparatus is a sheet copper projection, a. The apparatus is provided with a handle for convenience of holding. The apparatus before being used is about half-filled with water. The water is then heated to boiling by placing a Bunsen burner beneath the projecting portion, a, which will be filled with water communicating with that in the jacket. In this way hot water will circulate throughout the copper jacket, and FIG. 23. 42 THE GA.S ENGINEER'S LABORATORY HANDBOOK. consequently, the funnel placed in it may be kept hot during nitration. After filtering, the clear hot solution (filtrate) is surrounded with a vessel containing cold water, so as to rapidly cool down the solution, and further the formation of small crystals. When the crystals have ceased to form, the supernatant liquid is poured off, and the crystals are placed upon porous tiles to drain. After draining for several hours, the crystals are broken up, and are finally dried by means of repeated pressure between folds of filter or blotting paper. They are then placed in a tight-fitting stoppered bottle. 22. Many chemicals may be prepared in a pure state by means of precipitation, and subsequent washing and drying. The preparation of pure calcium carbonate is an example. Dissolve 100 grams of calcium chloride (CaCl 2 ) in 250 cc. of distilled water, heat the solution to boiling, filter if necessary, and then add a strong solution of pure ammonium carbonate (NH 4 ) 2 C0 3 , as long as any precipitate is thrown down. Pour the precipitate on to a filter- paper, and wash it with hot water until the latest washings cease to give a precipitate with a solution of nitrate of silver (AgN0 3 ). Then dry the precipitate in the water oven, detach it from the filter-paper, and preserve in a stoppered bottle. 23. The process of sublimation is sometimes employed for the purpose of separating volatile from non-volatile sub- stances, the purification of such substances being in this way effected. In order to purify a volatile substance, the sublimation may be effected by placing the substance in a porcelain crucible, and inverting over it a similar crucible. The lower crucible is then gently heated, when the substance will volatilise and condense in the upper one, which must be replaced by a cold one if it becomes too hot. Another method is to place the substance into a porcelain dish, and condense its vapour by means of an inverted funnel, placed so as to cover the substance. Unless the SIMPLE GRAVIMETRIC DETERMINATIONS. 43 yi bstance which is to be sublimed is known to be quite dry, i r is advisable to allow the first portion of its vapour to pass a-vay, and carry off any moisture there may be in the sub- is 1 ance with it. The sublimation of iodine is an example of the first 11 ethod of purification by means of sublimation. Partly fill a porcelain crucible with a mixture of three ]>irts of iodine and one part of iodide of potassium (KI), 1i;iely powdered, and well mixed together. This is then < jvered with a second crucible, and a gentle heat applied 1 > the lower crucible. At the close of the operation the s iblimed crystals are removed and stored in a clean dry 1 ottle. The sublimation of arsenious oxide (As 2 3 ) is an example o ' the second method of purification by means of sublima- t on. Place in a small porcelain dish a thin layer of As 2 3 . r l hen invert over it a clean glass funnel, and gently heat t le dish, The sublimed As 2 3 is detached from the funnel a ; the end of the operation and preserved for future use. ESTIMATION OF S0 4 IN A SULPHATE (SOLUBLE). 24. The S0 4 is precipitated as barium sulphate and weighed IP. tliis form. The manipulation necessary in the determination of the tibove, finds frequent application in the analytical work required in gasworks. The determination of the amount of sulphur in coal and coke, the sulphur compounds in purified gas, the testing of the solutions employed in the Keferees' ammonia test, and in the testing of ammoniacal liquor and sulphate of ammonia, are typical instances. The substance chosen for the estimation is copper sul- phate (CuSO 4 ,5H 2 0). 44 THE GAS ENGINEEK'S LABOKATORY HANDBOOK. In order to obtain the substance in a state of purity, it is necessary to purify it by recry stall isation (21) ; but even this treatment is not always sufficient to purify the com- mercial salt (blue vitriol) ; as it not unfrequently contains ferrous sulphate, which cannot be removed even by repeated crystallisation, as the two sulphates tend to crystallise, to- gether. By heating the solution, with a few drops of nitric acid, however, the ferrous salt is oxidised to ferric sulphate, and on concentrating the liquid, crystals of pure copper sulphate are easily obtained. In order to prepare a quantity of the pure salt, about 200 grams of clean, well-formed crystals of commercial blue vitriol are dissolved in about a quai ter of a litre of hot water, a few drops of nitric acid are added, the solution is filtered if necessary, and boiled for half an hour ; on cooling, the liquid deposits crystals of pure CuS0 4 . After standing for a few hours the solution is poured off, and the mother liquor is drained as far as possible from the crystals. The crystals remaining are then broken up by means of a glass rod, and dried by pressure between folds of filter or blotting paper. It is advisable not to press the crystals with too great a force, as the sulphate may consequently become mixed with filaments of filter-paper, which would interfere with the accuracy of the after analytical operations. When the greater portion of the moisture has in this way been got rid of, the salt is wrapped in a fresh sheet of dry filtering paper, and the folds are placed under a heavy weight for an hour or two. When dry the salt is placed in a stoppered weighing tube. About one gram of the dried salt is now weighed out in the following manner. The stoppered weighing tube containing the CuS0 4 is carefully weighed, and about a gram of the salt is shaken out into a clean dry beaker of about 18 ounces capacity. On replacing the stopper and again weighing the tube, the loss of weight resulting, gives the exact amount taken for analysis. Care must of course be taken that all the salt shaken out of the tube is deposited in the beaker. SIMPLE GEAVIMETKIC DETERMINATIONS. 45 The salt is now dissolved in 30-40 cubic centimetres of d istilled water, a few drops of hydrochloric acid are added, a ad the solution, covered with a clock-glass, is heated to 1 oiling. The burner is then removed from underneath the 1 eaker, and a saturated solution of barium chloride added ( rop by drop. In order to ascertain whether an excess of 1 lie precipitant has been added; the barium sulphate is left 1 ) settle down, and when the supernatant liquid is suf- 1 eiently clear, a drop of the barium chloride solution is 1 oured down the side of the beaker. If an increased tur- i idity ensues, the liquid is again heated, and a further < uantity of barium chloride solution added ; the precipitate i 3 once more allowed to settle, and the liquid again tested 1 >y cautiously adding barium chloride solution. When it is uite evident that the precipitation is complete, cover the I eaker and set it aside in a warm place for the precipitate o settle. If it is attempted to filter the turbid liquid mmediately, the finely divided precipitate will inevitably pass through the pores of the paper, but on standing, especially after precipitation from a hot solution slightly acidified with hydrochloric acid, the barium sulphate becomes denser, and assumes a granular form. When the precipitate has completely settled, add one more drop of barium chloride to be quite certain that the precipitation is complete, and proceed to collect the BaS0 4 on a filter; the supernatant liquid is first poured through, care being taken not to disturb the precipitate. The precipitate is then washed twice by decantation through the filter, and finally transferred to the latter by means of 1he wash-bottle with a fine jet, and a glass rod tipped with india-rubber tube. The precipitate is washed free from chloride, as shown by a portion of the latest washings not giving a precipitate with a solution of nitrate of silver. When thoroughly washed, the filter and precipitate are dried in the water oven, and, when dry, the precipitate is ignited, apart from the filter, in the following manner. The 46 THE GAS ENGINEER'S LABOR ATOKY HANDBOOK. precipitate is first detached as completely as possible from the filter, and placed in a platinum crucible. The filter is then folded up with the soiled surface in the middle, and bound round by a spiral of platinum-wire, and burnt to a white ash in the flame of a Bunsen burner. The ash is then added to the precipitate in the crucible, and the whole is ignited for about fifteen minutes at a dull red heat. It is then allowed to cool in the desiccator and weighed. The precipitate is ignited again for ten minutes, then allowed to cool, and weighed; this operation being repeated until the weight is constant. From the weight of the BaS0 4 thus obtained, the percent- age of S0 4 may be calculated in the manner described below. It is necessary to note that any BaS0 4 adhering to the filter paper will be reduced to BaS. If this is present in any appreci- able quantity it will be necessary to reconvert it into BaS0 4 by the following procedure. The filter-ash which has been dropped into the crucible is 'moistened with two drops of dilute HC1, and one drop of dilute H 2 S0 4 . The contents of the crucible are then very gently heated until the excess of acid has been evaporated, great care being taken to avoid loss by spurting. The crucible and its contents are then strongly ignited. The following shows the method of entering the results, and making the necessary calculations. Example. Determination of the amount of sulphuric acid in copper sulphate, by precipitation as barium, sulphate. Grams. Weight of tube + CuS0 4 (before) 9-7612 (after) 8-6066 Weight of CuS0 4 taken 1-1546 Weight of crucible + BaS0 4 + ash 17 "t 024 alone 16-5216 1-0808 Weight of filter ash '0022 BaS0 4 1-0786 SIMPLE GRAVIMETRIC DETERMINATIONS. 47 When the weight of a precipitate has been ascertained, it is necessary to calculate that of the constituent the \v )ight of which we wish to learn this is done by the help of the atomic weights given in the Appendix. In the present example, we first require to know the ai lount of (S0 4 ) which is equivalent to '0786 grams, 15 iS0 4 . The molecular weight of BaS0 4 = 233 (Ba = 137, S = 32, . = 64 = 233), and this contains 96 parts by weight of SO 4 (> "= 32, 4 = 64 = 96). Then, as 233 : 96 :: 1-0786 : x. x = 0-4444 gram S0 4 . We thus have 0*4444 gram S0 4 , from 1*1546 grams C iS0 4 ,5H 2 0, so, to obtain the percentage, we say As 1-1546 : 100 : : 0-4444 : x. x = 38-49 per cent. Tiat is, we have found 38-49 per cent, of S0 4 in the CuS0 4 5 ,1 2 operated on. We have now to see if the analysis is correct or not, by cl .ecking it against the numbers demanded \)y theory. The molecular weight of CuS0 4 ,5H 2 O = 249 5 (Cu= 63 5, S = 32, 4 = 64, 5H 2 = 90 = 249 -5), and this contains 96 pcjts by weight of S0 4 , consequently. As 249-5 : 100 : : 96 : x. x = 38-48 per cent., hence the error in the estimation = 0-01. A difference of 2 per cent, from the numbers demanded by theory, or from the result of a duplicate analysis, is the mean maximum error which is usually allowed in the case of simple quantitative estimations. ESTIMATION OF BARIUM IN A SOLUBLE COMPOUND. 25. Inasmuch as the estimations involving the same methods of manipulation as required in the previous ex- ample, are amongst those most frequently occurring in gas- works, it will be advisable to gain further experience in the 48 THE GAS ENGINEER'S LABORATOKY HANDBOOK. same, by estimating the amount of barium in a soluble salt, the operations being almost identical with those described in (24). The barium is precipitated as barium sulphate, and weighed in this form. Weigh out from a weighing bottle about 1 gram of pure recrystallised barium chloride (Ba01 2 ,2H 2 0) into an 18- ounce beaker; dissolve the salt in about 100 cc. of distilled water to which a few drops of dilute hydrochloric acid have been added. Heat the solution to boiling, and add an ex- cess of dilute sulphuric acid ; keep up a boiling heat until the precipitate becomes compact and readily settles down. Allow the precipitate to settle down completely, decant the clear liquor through a No. 5 Swedish filter, wash the precipi- tate twice with hot water by decantation, filter, and finish the washing on the filter. The last washings must give no precipitate with BaCl 2 solution. Dry the precipitate in the water oven, and ignite it as described under the estimation of(S0 4 ). (24.) From the weight of BaS0 4 thus obtained, the percentage of Ba may be calculated. The following expresses the reaction : Ba01 2 + H 2 S0 4 = BaS0 4 + 2 HC1. Consequently every 233 parts of BaS0 4 are equivalent to 137 parts of Ba. ESTIMATION OF AESENIC AS ARSENIOUS SULPHIDE. 26. This method is applicable to the determination of sul- phuretted hydrogen in gas liquor or in crude coal gas. Weigh out accurately a little more than half a gram of pure resublimed arsenious oxide, place it in an 18-ounce flask, and add about 50 cc. of dilute hydrochloric acid. Heat the flask on a water bath until all the oxide is dis- solved, being careful not to allow the temperature of the solution to attain 100 C., or some of the arsenic will be SIMPLE GRAVIMETRIC DETERMINATIONS. 49 volatilised as chloride. A few drops of sulphurous acid solution are then added, in order to convert any As 2 5 that may be present into As 2 O 3 , and the heating on the water bath is continued until all smell of S0 2 has ceased. The arsenic is then precipitated as As 2 S 3 , by passing SH 2 gas th rough the solution in the following manner. The flask co itaining the solution of As 2 3 is provided with a doubly pi rf orated india-rubber cork, through one of the holes in w iich a glass tube passes, bent at right angles, and reaching ne irly to the bottom of the flask. A glass tube passes through th 3 other hole also, one end of this tube terminating just be low the bottom of the cork, the other end being slightly cc atracted in order to prevent air from diffusing into the flask. Si Iphuretted hydrogen gas is then passed through the first tu be into the liquid in a gentle stream, until the liquid is sa ;urated with the gas. The flask is then left for some hours in a warm place, in order to ensure complete precipitation. In very accurate experiments, C0 2 gas is next passed through th 3 liquid until the excess of SH 2 is removed. During the pa ssage of the C0 2 , the liquid is gradually heated to boiling ; th s makes the precipitate more dense, and consequently more easy to wash. If care is taken that the precipitated As 2 S 3 is kept covered, and is not much exposed to the air, the passage of C0 2 may be omitted. Filter through a double tared filter which has been dried at 100 C. (20), and wash the precipitate with hot wa ter containing a little H 2 S, until it is free from HC1. Test the filtrate and washings to see if all the As is precipitated, by saturating the liquid with H 2 S gas, and allowing it to stand. If any further precipitate is thrown down, add it to tho main portion on the filter. Dry the precipitate at 100 C. in the water oven until the weight is constant. Since the precipitate may contain a little free sulphur, it is necessary to treat it with carbon bisulphide in the same manner as described under cadmium (27). After this treatment, the precipitate is again dried at E 50 THE GAS ENGINEER'S LABORATORY HANDBOOK. 100 C., and is finally weighed. From the weight of the As 2 S 8 obtained, the percentage of As is calculated. As 2 S 3 As 2 As 246 : 150 : : wei ^ P hide : weiht f arseni<3 ' 198 parts of As 2 O 3 should yield 246 parts of As. 2 S 3 . ESTIMATION OF CADMIUM, AS CADMIUM SULPHIDE (CdS). 27. This method may be employed in the estimation of SH 2 in crude coal gas by means of cadmium chloride. Weigh out accurately about 1 gram of pure recrystallised cadmium sulphate (CdS0 4 ,4H 2 0). Dissolve it in about 250 c.c. of distilled water, and add a few drops of dilute HC1. Pass a current of SH 2 into this solution, until it smells strongly of the gas. Filter off the precipitated cadmium sulphide, making sure that all the cadmium has been pre- cipitated, by diluting a portion of the filtrate and passing SH 2 through it again. If no further precipitate is thrown down, collect on a tared filter (20) and wash the precipitate, first with diluted sulphuretted hydrogen water mixed with a little HC1, and finally with pure water. Dry the filter and precipitate in the water oven, and weigh. The pre- cipitate sometimes contains free sulphur. In very accurate determinations, therefore, the dried precipitate should be repeatedly washed with carbon bisulphide, as long as any residue is observed on evaporating a few drops of the wash- ings upon a clean watch-glass. The precipitate, after this treatment, is again dried and weighed, any difference be- tween the two weighings being due to the removal of free sulphur. From the weight of CdS obtained the percentage of cadmium may be calculated. CdS Cd As 144 : 112 : : Weig P hide : weight of cadmium. SIMPLE GRAVIMETRIC DETERMINATIONS. 51 ESTIMATION OF CALCIUM. 28. This method is applicable to the determination of lime. Tl e calcium is precipitated as calcium oxalate, and weighed eit \er as carbonate or as oxide. Weigh out accurately into an 18-ounce beaker about 1 ^ram (not more) of pure calcium carbonate. Moisten it with distilled water, and dissolve it in dilute HC1. The be iker should be covered with a clock-glass, and the HC1 ad led very cautiously, so as to prevent loss by the effer- vc >cence set up by the addition of the acid. The resulting so ution is then diluted with distilled water to about 100 c.c., an 1 the liquid heated nearly to boiling ; a slight excess of so ution of ammonia is then added, and finally a moderate ex 3ess of ammonium oxalate solution, after adding which the li( uid is allowed to stand. As soon as the precipitate has completely subsided, pour oif the supernatant liquid through a 1 liter, being careful not to disturb the precipitate. Wash th $ precipitate in the beaker with hot water two or three titles by decantation; and finally transfer the precipitate to tho filter by means of a glass rod tipped with india-rubber tu "'ring, and the jet of the wash-bottle. The washing of the precipitate on the filter is then continued, until the last por- tion of the wash-water no longer turns a solution of nitrate of silver milky. When that stage is reached, the filter and precipitate are dried in the water oven. The next operation is to convert the precipitate of calcium oxalate into calcium carbonate. The precipitate is transferred, as completely as possible, to a platinum crucible, and the filter-paper burnt in a platinum wire spiral (19&), the ash also being allowed to fall into the crucible. The crucible is then heated with the lid on, very gently at first, and finally to very faint redness for ten minutes. The bottom of the crucible must only appear of a faint red colour when shaded from direct light. The crucible is then placed in the desiccator, and when cold weighed. During the process of ignition of the above E 2 52 THE GAS ENGINEER'S LABOEATOKY HANDBOOK. precipitate, especially if it is at all over-heated, a portion of the calcium carbonate is often converted into calcium oxide (CaO); hence, after the first weighing, the contents of the crucible should be treated with a little strong solution of ammonium carbonate, which converts any CaO back again into CaC0 3 . It is then placed again in the water oven to dry, heated to faint redness for a few minutes, allowed to cool, and weighed once more. These operations are all repeated, until the weight no longer increases by the treat- ment. From the weight of CaC0 3 finally obtained, the percentage of calcium may be calculated. Mol. weight Atomic weight of CaCO 3 of calcium As 100 : 40 :: weight of CaC0 3 weight of obtained calcium. The precipitate of calcium oxalate is, however, more readily converted into CaO than into CaC0 3 , especially if it is small in amount, say not much more than 1 gram. In order to convert the precipitate into CaO, the crucible and its contents are heated to bright redness for 15 minutes over the blow-pipe flame. The crucible is then allowed to cool, and weighed, and then heated again for 5 minutes in the same way, cooled, and weighed. These operations are then repeated, until the crucible ceases to lose weight. The weight of substance finally obtained when the weighings are steady, is that of the calcium oxide (CaO) plus the filter ash. From the weight of CaO thus obtained, the percentage of calcium may be calculated. Mol. weight Atomic weight of CaO of Ca As 56 : 40 :: weight of CaO . weight of obtained calcium. ESTIMATION OF IRON. 29. This method of determination is applicable to the analysis of Bog Ore. SIMPLE GRAVIMETEIC DETERMINATIONS. 53 The iron is precipitated as ferric hydrate (Fe 2 3 , 3H 2 0), This is then ignited and weighed as ferric oxide (Fe 2 O 3 ). Weigh out accurately about 1'5 grams of pure ferrous a nmonium sulphate Fe(NH 4 ) 2 (S0 4 ) 2 6H 2 into an 18-ounce 1) iaker, or preferably into a platinum or porcelain basin. I issolve this in about 200 cubic centimetres of distilled v ater, containing a few drops of dilute sulphuric acid. Heat t.ie solution with sufficient strong nitric acid to convert the i on, which originally was in the ferrous form, into the ferric s ate. This change may be known to be nearly effected when t ie further addition of a drop of strong nitric acid does not c 'eate any brown colour in the solution ; but the change to t ie ferric condition should be proved to be complete by add- i ig a drop of the dilute solution to a drop of a freshly made s )lution of potassium ferricyanide, placed on a white porcelain t Ie, when no blue colour should appear. While the solution i; still hot, add an excess of solution of ammonium hydrate, a ad boil. The precipitate is washed once by decautation 1 y boiling it with water, and is then transferred to a filter, t ie process of nitration being accelerated by the use of the filter pump. (17) The washing is continued until the latest \vashings do not give any precipitate when boiled with a s nail quantity of BaCl 2 solution. The precipitate is dried in the water-oven and transferred to a crucible. The filter is burnt apart from the precipitate, bsing added to the latter, which is then ignited at a bright red heat over a Bunsen burner, until a constant weight is ob- tained. From the weight of Fe 2 3 obtained, the percentage of iron may be calculated. Of course, if the iron is already in the ferric condition it will not require the preliminary treatment with HN0 3 , but may be precipitated at once with AniHO solution. It is necessary to note that the presence of certain non-volatile organic substances prevents the complete precipitation of iron from its solutions by AmHO. The organic substance should therefore be completely destroyed before precipitating with ArnHO, by adding carbonate 54 THE GAS ENGINEEK'S LABOKATOKY HANDBOOK. of soda and nitrate of potash to the solution, evaporating to dry- ness, and fusing the residue. (If the iron is in the solid state, the organic matter may be at once got rid of by ignition at a red heat.) The iron is then dissolved out by heating with dilute HC1, or a mixture of HC1 and HN0 3 , and the filtered solution treated as above. Fe 2 O 3 Fe 2 As 160 : 112 :: wei ? ht f : * = 78 ht precipitate of Fe. ESTIMATION OF ALUMINA (A1 2 3 ). 30. The estimation of A1 2 3 occurs in the analysis of Bog Ore and in fire-clay. The A1 2 O 3 is precipitated by AmHO as A1 2 3 , 3H 2 0. This is converted into A1 2 3 by ignition, and is weighed in this form. Weigh, out accurately about a gram and a half of pure recrystallised potash alum A1K(S0 4 ) 2 12H 2 into a platinum or porcelain dish. Dissolve the salt in about 150 c.c. of distilled water, add a solution of ammonium chloride in moderate quantity, and afterwards a slight excess of am- monium hydrate solution (AmHO). A white gelatinous precipitate of (A1 2 3 , 3H 2 0) will form. Cover the vessel with a clock-glass, and bring the solution to the boiling point, continuing the boiling, until the liquid has only a faint smell of ammonia. Wash the precipitate by decantation and transfer it to a filter, employing the filter-pump to accelerate filtration. Continue washing the precipitate on the filter until it is free from dissolved sulphate. Dry in the water-oven, and ignite the precipitate with the filter-paper in a platinum crucible. The crucible should be heated gently at first, and gradually brought to a white heat over the blow- pipe flame. It should then be cooled and weighed, the igni- tion being repeated until a constant weight is obtained. From the weight of precipitate we can get the percentage SIMPLE GKAVIMETKIC DETERMINATIONS. 55 of aluminium, but as a general rule aluminium is generally shown in analyses as combined with oxygen, in the form of al imina A1 2 3 . SEPARATION OF IRON AND ALUMINA. 31. Weigh out into a platinum basin about one gram of potash, alum and half a gram of pure ferrous sulphate. Dissolve in a little distilled water, add hydrochloric acid ai id a few drops of nitric, and boil, in order to convert the ii .m into the ferric state. Then dilute to about 20 c.c. and a< .d sufficient pure potash to produce a strong permanent p -ecipitate. Now add excess of strong caustic potash, and k >ep at a boiling heat, constantly stirring with a platinum r< -d to prevent actual boiling, until the A1 2 3 can be assumed t< have passed into solution, and dilute, first as far as p )ssible in the platinum basin, then by instalments, more krgely, in a Berlin ware basin provided with a spout, and filter off the ferric hydrate (Fe 2 3 ,3H 2 0). Wash the pre- cipitate thoroughly with hot water; the result is that sub- si an tially the iron and alumina are separated from each o ;her. In general, however, what should be ferric oxide r< stains a more or less considerable portion of the alumina. Ir, is necessary, therefore, to extract this alumina by re- dissolving the precipitate with hydrochloric acid, and repeating the treatment with potash. The crude ferric hydrate obtained cannot be ignited and weighed as it is, as it contains a certain amount of combined fixed alkali (KHO), which it carries down with it in the act of precipitation. It is requisite to dissolve it in hydrochloric acid, re-precipitate it by ammonia, and ignite it as in (29). From the alkaline liquors precipitate the A1 2 3 , by adding an excess of NH 4 C1, and heat the mixture until the NH 3 liberated is practically expelled, then treat the precipitate as in (30). For a check, determine the iron and alumina conjointly by precipitating with NH 3 and NH 4 C1. 56 THE GAS ENGINEERS LABORATORY HANDBOOK. ESTIMATION OF MAGNESIUM. 32. Weigh out about one gram of pure, dry, crystallised magnesium sulphate, place it in a half-pint beaker, and dissolve it in about 50 c.c. of distilled water. Add enough solution of ammonium chloride to prevent the precipitation of any magnesia on adding ammonia (about one gram of solid NH 4 C1 will be sufficient), the ammonia being added in considerable excess. A solution of phosphate of soda is next added (about 1'5 of crystallised phosphate of soda dissolved in water) as long as it appears to produce a precipitate, and the solution is briskly stirred with a glass rod, but without rubbing the sides of the beaker ; the glass rod may be left in the beaker. When the precipitate has subsided, a few drops more sodium phosphate are added to the clear liquid, in order to be quite sure that no further precipitate is produced. The beaker is now covered, and left for about twelve hours. The precipitate is then collected upon a filter, of about 4 inches in diameter before being folded, and is rinsed on the filter with distilled water containing one-eighth of its bulk of strong ammonia, the same solution being subsequently employed for washing the precipitate, the washing being continued until the washings, after being acidulated with nitric acid, give only a very slight turbidity with a solution of nitrate of silver. The filter is then dried with the precipitate, and when dry, as much of the precipitate as possible is transferred to a platinum crucible, which is then gently heated (the cover being loosely placed on) and finally ignited at a red heat. In this way, the precipitate of phosphate of magnesium and ammonium (MgNH 4 P0 4 ) is converted into magnesium pyrophosphate (Mg 2 P 2 7 ), which is allowed to cool, and weighed. It may then be emptied out of the crucible (pre- serving it for subsequent examination if the result should not prove correct) and the filter may be incinerated by igniting it, and allowing the ash to drop into the crucible SIMPLE GRAVIMETRIC DETERMINATIONS. 57 containing the precipitate. When cool, the crucible is agaiii w 3ighed, when the total weight of (Mg 2 P 2 07) will be obtained. (The above method of incineration is recommended on ac count of the filter-paper in this case being very difficult to burn, from its retaining traces of phosphate, and it can sc ircely be heated sufficiently to burn off the carbon on the tc p of the mass of precipitate.) The amount of magnesium is calculated by the following p: oportion : "Sum" Two atomic weights Weight .of ignited pyrophosphate of magnesium precipitate 222-6 : 48-6 : : x. ESTIMATION OF SILICA IN SILICATES. 33. This method of estimation is used in the analysis of fir e- biicks and fire-clay. The silica is converted into the insoluble form by evaporating tlw silicate to dryness with excess of acid. It is then rendered co npletely anhydrous by ignition, and weighed as silica (Si0 2 ). If a silicate is insoluble in acids it is first fused with six times its weight of fusion mixture (Xa 2 CO 3 -f K 2 C0 3 ) in a covered platinum crucible until all effervescence ceases ; the fused mass is then separated from the crucible by boiling with hc-t water in a covered evaporating basin, HC1 being added if required, and then treated as below. If the silicate is soluble in water, or is capable of being decomposed by strong hydrochloric or nitric acids, it may be treated at once as below. A suitable substance for an experimental determination is soluble glass, or sodium silicate, which is one of the soluble silicates, and consequently will not require the operation of fusion. Weigh out accurately into a platinum basin about one gram of soluble glass in very fine powder, and moisten the 11 C&i 58 THE GAS ENGINEER'S LABORATORY HANDBOOK. powder with a little distilled water. Having covered the basin with a clock-glass, place it upon a water-Lath, and gradually add strong HC1, at the same time keeping the contents of the basin constantly stirred up by the aid of a glass rod rounded at the end. When the powder is com- pletely decomposed, which is known to be the case as soon as no gritty particles are felt with the glass rod, remove the clock-glass, and continue the heating until the liquid is evaporated to dryness. The mass should be continually stirred the whole time, any lumps which form being broken up by means of the glass rod. When the powder appears to be perfectly dry it may be removed to a sand-bath. It is then gently heated with a small Bunsen flame until, on placing a cold clock-glass on the basin for a few seconds, there is no appearance of any moisture on the clock-glass. All the silica will now be insoluble. When the contents of the dish are cold, moisten the powder with strong HOI, and warm it upon the water-bath ; then add hot distilled water, stir the mixture well, allow the solid portion to settle, and decant the liquid through a filter. This treatment of the residue is repeated three times. The residue itself is then transferred to the filter, washed thoroughly with hot distilled water, and dried in the water-oven. The silica is next de- tached from the filter as completely as possible, and the filter is incinerated in a coil of platinum wire, and added to the precipitate. The crucible, covered with its lid, is then gently heated, care being taken that none of the fine, light silica is lost by being carried off with the escaping moisture ; the best way of preventing this being to apply heat to the crucible very gradually, and only after some little time, to increase the tem- perature until the crucible attains a bright red heat, at which it must be kept for about a quarter of an hour. The crucible is then cooled in the desiccator, and when cold weighed in the usual manner, being re-heated and weighed, until a constant weight is obtained. The percentage of Si0 2 may then be calculated. In all determinations of silica, it is necessary to test the Si0 2 in order to see if it is pure. Th SIMPLE GRAVIMETRIC DETERMINATIONS. 59 ignited silica is heated with about 50 c.c. of a saturated solution of sodium carbonate for about 30 minutes. The silica, if pure, will then have completely dissolved. Should ai y residue remain, this is again treated with sodium car- be nate solution, and if now there should still a residue remain, it should be washed, dried, and weighed. This \v jight, subtracted from the weight of the silica obtained in tl e first instance, will give the corrected weight of the silica. T ie dried residue is then fused with alkaline carbonates, the n: iss when cool is treated with HC1, and the Si0 2 is separated and estimated as above described. The weight of Si0 2 thus found should be added to the c< rrected weight obtained above. From the total weight of Si0 2 , the percentage of Si0 2 may b> calculated. ESTIMATION OF C0 2 IN A CARBONATE. 34. Two methods of estimating the above are given. The first m 3thod is by weighing the C0 2 evolved on treating the carbonate w'th an acid ; the second by finding the loss of weight which the cc rbonate suffers by the removal of C0 2 . These methods may be ertployed in the testing of the fresh and spent lime met with in gasworks. (1) By direct weighing. The C0 2 is driven off on treating the carbonate with an acid, and then absorbed by means of soda lime in U tubes and weighed. Pure calcium carbonate, or crystals of Iceland spar, may bo employed in the estimation. The apparatus used is shown in Fig. 24. A conical flask a, having a capacity of from 100 to 150 cubic centimetres, is fitted with an india-rubber cork pierced with two holes. Through one of the holes a 50 c.c. pipette b passes, this being filled with dilute HC1 for the decomposition of the carbonate before the experiment is started. The lower end of the pipette should de.-cend to within about 1 inch from the bottom of the flask, and be drawn out to a fine point. To the other end of the pipette 60 THE GAS ENGINEERS LABORATORY HANDBOOK. a piece of india-rubber tubing, about 8 or 9 inches long is attached, and this connects the pipette with the U tube c filled with soda lime. This U tube is open to the atmo- sphere at one end, and is intended to remove the C0 2 from the air, when it is required to aspirate the same through the apparatus. The india-rubber tube is provided with a screw pinch-cock, so as to open or close the connection with the U tube c at pleasure. Through the other hole in the cork of the flask a piece of glass tubing, bent at right angles, is FIG. 24. inserted to the depth of about an inch. This serves to connect the outlet of the flask with the remaining portions of the train of apparatus, consisting of bulb and U tubes, connected by means of india-rubber tubing as shown. These are intended for drying, purifying, and absorbing the CO 2 gas given off in the flask a. The bulb tube d contains strong sulphuric acid, and its object is to free the gas from moisture ; but it is also useful in denoting the speed at which the gas or air issues from the flask a. Liebig's potash bulbs answer SIMPLE GKAVIMETRIC DETERMINATIONS. 61 we 11 for the purpose. The acid should about half fill the three bottom bulbs. The U tube e has the limb nearest to the bulb tube filled with fragments of calcium chloride, the other limb being filled with fragments of pumice impreg- na ced with dehydrated sulphate of copper, in order to absorb th j vapours of HC1 given off during the heating of the flask a. The stoppered tube / is the one intended for the absorption of the C0 2 from the carbonate, and has the limb nc xt to e, and half the other limb, filled with crushed soda- lii le, the remaining portion of the limb being filled up with fr, .gments of calcium chloride. During the process of al sorbing C0 2 the soda-lime becomes heated, and in con- se [uence moisture is evolved, this is taken up by the calcium cHoride in the upper part of the limb, otherwise loss of w ight would occur by reason of the escape of moisture. T ie remaining tube g is filled with fragments of calcium cb loride, and is connected with a water aspirator A, which is en iployed to draw a current of air through the flask and tubes at the end of the experiment, so as to cause all the C0 2 to be sent forward to the absorption tube /. The U tubes employed may be of two kinds, those not requiring to be weighed being of the form shown in Fig. 25. They are prepared for use as follows. In the first place they are washed perfectly clean with ordinary water, and finally rinsed with distilled water. They aro then dried in the water-oven, and when quite dry they are ready for filling. Before filling them, however, it is necessary to see that the absorb- ent is in a proper condition, by crush- ing any large lumps and by rejecting any fine powder that may be present. This is necessary for the reason that as the absorbent takes up the gas by surface contact, the smaller the lumps, the larger is the absorbent surface presented to the current of the gas. On the other hand, if the ab- FU FIG. 25. 62 THE GAS ENGINEER'S LABOKATORT HANDBOOK. sorbent were in the state of fine powder, this would more or less completely fill up interstices, and in this way impede, or even entirely stop the passage of the gas. An absorbent may be prepared in a proper condition for use in the following manner. The substance is first coarsely powdered in a porcelain mortar, and the resultant powder is shaken upon a fine brass wire sieve, of about twenty strands to the inch, the sieve being perfectly clean and dry. The portion passing through the sieve is rejected ; the remainder is then passed through a coarser sieve of only eight strands to the inch, so as to separate out the larger lumps. The smaller particles which pass through this sieve may then be used for filling the absorption tubes, and the larger lumps remaining behind are broken up again in the mortar, and again sifted as above. The U tube is then filled with the absorbent up to the level of the side tubes, and a loose plug of cotton-wool placed on the top, so as to prevent small particles of the absorbent from being carried through the side tubes by the issuing gas current. Two sound corks are then selected so as to tightly fit the open ends of the U tubes, and when fitted into the tubes, the top of the cork is cut off level with the top of the tube ; finally the top of the cork is coated with melted sealing wax, so as to make it perfectly gas-tight. When the absorbent is in the tube, it is most essential to protect it from contact with the atmosphere, and this may be effected by pushing over the side tubes short pieces of india-rubber tubing, closed with small pieces of glass rod. These plugs are only removed when the U tubes are being connected up for use, or while they are being weighed. Tubes c and g are filled in the manner above described. The copper sulphate on pumice contained in the tube e is prepared in the following manner. Lumps of pumice are broken up and sifted as just de- scribed, and the sifted particles are then heated with a moderately strong solution of copper sulphate. The excess of liquid is poured off, and the pumice, which is now thoroughly saturated with the solution, is heated in an air-oven to a SIMPLE GKAVIMETRIC DETERMINATIONS. 63 temperature of 200 C., until dehydrated, as shown by the bl le colour of the salt disappearing. It may then be used for partially filling the tube e. For the absorption tube /, which contains soda-lime for the prrpose of taking up the C0 2 , as this tube is weighed, it will be more convenient to use the modern form of stoppered IT tube. These stoppered tubes differ from those ordinarily en ployed, by having a hollow stopper, ground into the u] per part of each limb. At the side of the hollow stopper, ar cl on a level with the opening of the side tube, is a per- fo -ation, so that by simply turning the stopper round, co nmunication between the U tube and the side tube may be opened or closed as required ; and as the stoppers, when gi 3ased with a little resin cerate or vaseline, close the tube qi ite tightly, contact of the absorbent with the atmosphere ni ly be prevented by simply turning the stopper ; con- se [uently it is not necessary to cover the side tubes, as in the or linary form of U tube. A plug of cotton- wool is loosely in verted in these stoppers, so as to prevent small particles of ths fragments from being carried out of the tube. The so la-lime for filling the tube / may either be purchased, or prepared in the following manner. Slake 200 grams of the best flare lime to a fine powder. Add 400 cubic centimetres of syrupy caustic soda solution. Make into a paste, dry quickly in a lar*>e copper dish protected from the access of C0 2 , and granulate by constant stirring. The tube is then filled with this material in the manner already described. The aspirator h may be improvised from a Winchester quart bottle in the following manner : A sound cork is selected to fit the neck of the bottle, and pierced with two holes. Into one of the holes a piece of glass tube, bent at right angles, is inserted, so as to just pass through the cork. Through the other hole, another glass tube bent at right angles passes, but in this case it passes to the bottom of the bottle. When in use, a piece of india- rubber tubing is attached to the longer glass tube, reaching below the level of the bottom of the bottle, in order that it 64 THE GAS ENGINEER'S LABORATORY HANDBOOK. may act as a syphon. At the bottom of the india-rubber tube there is a screw pinch-cock, in order to regulate the flow of the water. When required for use, the bottle is filled with water, and suction is applied to the syphon tube until it is also filled with water. As soon as this tube is filled, a continuous stream of water will continue to flow. After the suction is stopped air will replace the water in the bottle, and in this way a continuous current of air can be drawn through the train of apparatus when they are attached to the shorter tube in the bottle. The rate of aspiration is con- trolled by the screw pinch-cock. The next operation is to test the apparatus for soundness. The tubes and flask are connected up by means of short pieces of india-rubber tubing, the glass tubes " butting " against each other. To the end tube g a piece of glass tubing is attached, dipping into a beaker of water. Gentle suction is then applied at the outlet of the tube c until the water rises to about the height of 6 inches in the glass tube attached to g. The screw pinch- cock on b is then closed. If the level of the water in the glass tube does not alter during the space of ten minutes, we may safely conclude that the apparatus is air-tight, but if the level of the water in the tube falls the apparatus leaks somewhere. The position of the leakage must be discovered and remedied by binding the leaky joints with fine copper wire, and, if necessary, replacing any defective india-rubber tubes with new ones. When all the joints have been made tight, the soda-lime absorption tube /, previously wiped clean and dry, is carefully weighed, and replaced in the train of appa- ratus. About 1 gram of calcium carbonate is then accurately weighed and placed in the flask a. The pipette 6 is then filled by suction with dilute hydro- chloric acid (1 : 3), the acid being retained in the pipette by the screw pinch-cock. The cork carrying the pipette is then inserted in the flask a, and is bound down to the neck of the flask, if necessary, by means of thin copper wire. The aspirator and the U tube c, not being required until a later stage of the process, are not connected for the present. SIMPLE GRAVIMETRIC DETERMINATIONS. 65 The pinch-cock on the pipette is now slightly opened, so as t j allow the acid to slowly fall on the CaC0 3 in the flask. Carbonic acid gas is given off, and drives forward the air in t ie flask through the train of apparatus. The bubbles of CO. should not pass through the bulbs containing sulphuric acic , at a greater speed than that at which the bubbles can be ( asily counted. When all the acid has passed out of the pip< tte into the flask, the pinch-cock on the pipette is closed, and as soon as gas has entirely ceased being evolved, the poii t of the pipette is gently pushed down beneath the sur: ace of the liquid in the flask. The U tube c, and the asp rator 6, are now connected to the rest of the apparatus, the pinch-cock on the pipette is opened, and the aspirator gen ly started. A small Bunsen burner is then placed under the flask , so as to cause the contents of the flask to be in a s tat 3 of incipient ebullition, in order to drive off all dis- soh ed C0 2 . The current of air drawn through by the asp rator replaces the C0 2 gas in the apparatus in front of the weighed soda-lime tube /, and carries it forward into the latt 3r tube. The speed of the air passing through should be such that the bubbles can be counted as they pass through the drying bulb d, the speed being regulated by the pinch- cocl: on the syphon pipe leading from the aspirator. In ordt r to expel the whole of the C0 2 , a volume of air equal to about six times the capacity of the flask should be drawn through the apparatus. A guide as to whether the C0 2 has all been absorbed, may be found in the temperature of the soda- lime tube. At the commencement of the experiment it will be noticed that the soda-lime tube becomes very hot, owing to the heat generated by the chemical combination of the C0 2 and the soda lime ; but as the absorption ceases, and air alone passes through, the tube becomes cool ; hence, when the soda-liine tube attains its normal temperature again, the greater part of the C0 2 gas will have been absorbed. When sufficient air has been drawn through, the aspirator is shut off, and the stoppers of the soda-lime tube are turned so as to close the F 66 THE GAS ENGINEER'S LABORATORY HANDBOOK. inlet tubes, the soda-lime tube is disconnected, and after standing for about half an hour is weighed. Any increase in weight over the weight at the commencement of the experiment is the amount of C0 2 given off from the calcium carbonate experimented on, and from these weights, the percentage of C0 2 in the CaCOg can be calculated. (2) Estimation of C0 2 by difference. The C0 2 gas is ex- pelled by treating the carbonate with an acid in a suitable apparatus, the loss of weight observed being the amount of C0 2 driven off. This method is more adapted for carbon- ates which are completely decomposed by H 2 S0 4 , since, when more volatile acids (HC1, &c.) are employed, the results are generally too high, owing to some of the acid being driven off by the heating at the latter stage of the experiment. The apparatus known as " Schrotters," Fig. 26, may be employed in the determina- tion. It is made of thin blown glass, and its total weight should not exceed 50 grams. It consists of a flask a, in which the decomposition of the carbonate takes place. Communicating with the flask there are two compart- ments, 6, and c, also an opening d, provided with a glass stopper, for the purpose of introducing the carbonate into the flask. The compartment b is filled with dilute acid, which is subsequently gradually allowed to fall into the flask a for the purpose of decomposing the carbonate by turning the stop-cock. The compartment c is half filled with strong sulphuric acid, for the purpose of drying the C0 2 gas before it leaves the apparatus. The gas passes up through the vertical inner tube, then down again, and out through the two holes at the bottom of the second tube. It then bubbles up through the strong sulphuric acid FIG. 26. SIMPLE GRAVIMETRIC DETERMINATIONS. 67 before escaping through the tipper exit tube into the air. The experiment is made as follows. About 1 gram of the oarbonate (say sodium carbonate) is accurately weighed and introduced into the flask a, by the opening d. The compartment b is then nearly filled with dilute sulphuric acid (1 : 5), and the compartment c is half filled with strong su\i huric acid. The whole apparatus is then accurately weighed. The dilute acid is now gradually allowed to fall into a by gently turning the stop-cock. As soon as the acid begins to act on the carbonate, bubbles of gas wiE be observed passing through the strong acid in c, and it ij requisite to regulate the speed at which they pass by Adjusting the rate of admission of the dilute acid, so that the separate bubbles can be easily counted. If the bubbles pass more rapidly than this, the gas will be imperfectly driel. When all the dilute acid has flowed into the flask, and bubbles of gas are no longer observed, the apparatus is plac ad upon a piece of wire gauze on an iron tripod, over a sma 1 Bunsen burner. At the same time, the aspirator employod in the previous experiment is attached by means of a piece of india-rubber tubing to the exit tube of c, and a gent le current of air is drawn through the flask. The liquid is tl en kept in a state of incipient ebullition, until a volume of air equal to about three times the capacity of the appa- ratus has been drawn through it. The flask is then allowed to cool, and when cold weighed. The loss of weight sustained gives the weight of C0 2 in the sodium carbonate operated on. As the results by this method are usually too high, it is not- adapted for very accurate determinations. The error is due to the loss of water and acid during the heating and aspiration. Note. If sulphites or sulphides are present, in addition to a carbonate, as is sometimes the case with spent lime, their injurious influence is best ob\iated by adding to the subsiance a solution of yellow chromate of potash, in more, than sufficient quantity to effect their oxidation. F 2 68 THE GAS ENGINEEE'S LABORATORY HANDBOOK. PAKT III. VOLUMETRIC ANALYSIS. Introductory Remarks. THE main principle upon which the system of volumetric analysis depends, is that of submitting the substance to be estimated to certain characteristic reactions, such reactions being produced by the employment of solutions of known strength ; and from the volume of solution necessary to produce such a reaction, the weight of the substance is determined by the aid of the laws of chemical equivalence. The following conditions are therefore essential to the successful carrying out of analyses on the volumetric system, as applied to liquids and solids. (1) The provision of a solution of a suitable reagent, the chemical power of which is accurately known, called the " standard solution." (2) The provision of a graduated vessel, from which portions of the standard solution may be exactly delivered, such vessel being known as a " burette." (3) The reaction produced by the standard solution with any given substance must, either in itself or by an indicator, be such that its termination is strikingly apparent to the eye of the observer, and the quantity of the substance with which it has combined accurately determined thereby. The following example of the estimation of the amount of chlorine in a substance, by means of a solution of nitrate of silver, will serve as an illustration of the method of volumetric analysis. The estimation depends upon the fact, that if we add to a solution of a chloride a solution of silver VOLUMETRIC ANALYSIS. 69 ni :rate, in the presence of a solution of potassium chromate, as soon as the silver nitrate is added in excess, it produces a permanent dark red precipitate of silver chromate; con- sequently the appearance of a slight red coloration th roughout the whole of the liquid, is an indication of the co inpletion of the reaction. The following equation shows what takes place : NaCl + AgN0 3 = AgCl + NaN0 3 . Consequently, if we have a solution of nitrate of silver, prepared by adding a definite weight of that substance to a k] .own volume of water, on adding this solution to a solution of sodium chloride, until the reaction mentioned above takes p] ice, every 170 parts of nitrate of silver required, indicates tl 9 presence of 35 5 parts of chlorine, or the equivalent w sight of any chloride. In an estimation of the amount of chlorine in a substance by this method, a solution of silver nitrate was employed w lich contained one-tenth of the molecular weight of silver ni :rate, or 17 00 grams per litre (= 1,000 cubic centimetres) ; co tisequently, from the equation previously given, each c.c. of tha solution was equivalent to 0*00355 gram of chlorine. 2 -6223 grams of the substance were weighed out and dissolved, and required 27 '5 c.c. of the silver solution for neutralisation. Consequently the weight of chlorine in the substance = 0-00355 x 27-5 = 0-097625 gram, and the percentage . , _ . 0-097625 X 100 . _ weight of chlorine - The operation of determining the amount of pure sub- stance in any solution by means of standard solutions is termed " titration." The term is more applicable to the operation than testing or analysing, as "these expressions may apply equally to qualitative and quantitative examina- tions, whereas titration can only relate to quantitative examinations. 70 THE GAS ENGINEER'S LABORATORY HANDBOOK. Volumetric determinations may be broadly classified into three methods : (1) The determination of the substance is effected by saturation with another substance of opposite properties. This class embraces the analysis of acids and alkalies, and alkaline earths, and is the one which is the most frequently used in gasworks. (2) The determination of the substance is effected, by a reducing or oxidising agent of known strength, the principle oxidising agents being potassium permanganate, potassium bichromate, and iodine, and the corresponding reducing agents, ferrous and stannous compounds, and sodic thiosul- phate. This method of analysis is employed in the analysis of oxide of iron, and of Weldon Mud. (3) The determination of the substance is effected by precipitating it in an insoluble and definite combination, as in the case of the determination of chlorine, previously de- scribed. NOTES ON THE MEASUREMENTS OF LIQUIDS AS APPLIED TO VOLUMETRIC ANALYSIS. 35. The metric system of weights and measures is the one usually adopted in systematic volumetric analysis, and the capacities of the different measuring vessels employed are based on the volume occupied by the weight of 1 gram of distilled water at a temperature of 15- 5 C. = 60 F., this volume being called a cubic centimetre. A litre weighs 1000 grams, and has a capacity of 1000 cubic centimetres. It is necessary to point out, however, that this is not the true, or normal cubic centimetre, as the latter contains 1 gram of distilled water at its greatest density, viz. 4 C., or 39 F., but as this is a temperature not often attainable in this country, it is better to adopt the temperature most easily attained, viz. 15 5 C., or 60 F. The true cubic centimetre (i. e. = 1 gram at 4 C., or 39 F.) contains only . 998981 gram at 60 F. ; but as the measuring vessels are used for VOLUMETRIC ANALYSIS. 71 ascertaining relative, and not absolute volumes of liquids, if all the vessels employed are graduated on the same basis no err >r can arise. If necessary, the contents of a vessel can be reduced to absolute volume by means of a correction based on the ex} ansion of water at different temperatures, as given in the Ap )endix. It is very essential to see that the temperature is cor stant when graduating measuring vessels, since the vol ime of a liquid alters very considerably under different ch; nges of temperature. Glass measuring vessels are always ma '/ked with the temperature at which the liquid is to be me isured in them. 36. Measuring vessels are required either to measure or to leliver liquids, and the capacity of a glass vessel is indi- cat 3d by a horizontal line scratched upon its surface, which she ws its capacity to that particular level. Some vessels are int 3nded both to measure and to deliver liquids. In this case it s necessary to have two marks; one of these is for the pu: pose of measuring the volume of a liquid, the other serves to deliver the same volume. The surface of liquids contained in narrow tubes is always curved, by reason of the capillary attraction exerted by the sides of the tube, and in consequence, it is rather difficult to obtain a distinct level in the liquid to be measured. If, however, the lowest point of the curve or meniscus is caused to coincide with the graduation mark, a correct read- ing is always obtained. While the volume of a liquid is being read off it is very essential that the measuring vessel is placed in a perfectly vertical position. The accuracy of all measuring vessels should be carefully gauged or " calibrated," by weighing definite amounts of distilled water in them, at a temperature of 15*5 C. The vessels usually required in volumetric analysis are measuring flasks, test mixers or measuring cylinders, pipettes, and burettes. 72 THE GAS ENGINEER'S LABORATORY HANDBOOK. Measuring Flasks. 37. A convenient series of measuring flasks is the follow- ing : (1) 1000 c.c., (2) 500 c.o., (3) 250 c.c., and 100 c.c." The flasks should be fitted with well-ground glass stoppers, and the graduation marks of each flask should be near the middle of the neck, which should be narrow, in order to ensure accurate readings. The space between the marks and stopper allows the liquid to be more readily mixed by agita- tion. The flasks should be made thin enough to stand being heated without risk of fracture. Fig. 27 shows a convenient form of litre flask. Measuring flasks are used both for measuring and for delivering liquids, consequently two marks are required. When the flask is filled to the lower mark it contains a certain volume, and when it is filled to the upper mark, the same volume can be poured out, or delivered from the flask. FIG. 27. Before a measuring flask is used, it is neces- sary to ascertain whether it contains or delivers exactly the volume marked upon it at the given tempera- ture, this operation being known as "calibration." In order to calibrate a flask for containing or measuring, it is first made quite clean and dry. It is then placed upon the pan of a sensitive balance, and gram weights equal in value to the number of cubic centimetres which the flask should contain, are placed on the pan beside the flask. The flask and weights are then exactly counterpoised by means of lead shot placed upon the other pan. The flask and weights are then removed. The flask is placed upon a level surface, and distilled water, at a temperature of 60 F., is poured into it, until the lower graduation mark on the neck is exactly coincident with the lowest point of the meniscus of the sur- face of the water. The neck of the flask above the mark is then carefully dried, and the flask (without the weights) is re- VOLUMETRIC ANALYSIS. 73 pLiced upon the pan of the balance. If the balance is again in exact equilibrium, the flask is correctly graduated for measuring or containing. If the balance is not in equi- lil rium, then the flask may be regraduated correctly, or a co -rection may be applied each time the flask is used. In order to regraduate the flask correctly, all that is nc cessary is to add to, or take away water from that already in the flask, until the balance is in equilibrium. The flask is then placed on a level surface, and a scratch made on the n< ck with a triangular file, or diamond, on a level with the lo vest point of the meniscus. This will be the true level \v len measuring liquids. In order to calibrate a flask for delivering, the flask is filled with distilled water to the mark on its neck. The w iter is then poured out, the flask inverted, and allowed to d] ain for fifteen seconds. The flask, with any remaining w iter adhering to it, is now placed upon one pan of the to lance, the neck having been previously closed by the st >pper, in order to prevent loss of weight by evaporation. A\ eights are then placed on the same scale-pan, and by the si le of the flask, the value of the weights in grams being equal to the number of cubic centimetres which the flask is in irked to deliver. The flask and weights are then counter- pc ised by means of shot placed upon the other pan of the balance. The flask and weights are now removed from the sc lie-pan, the flask placed on a level surface, and accurately filled to the delivering mark with distilled water. The stopper is then inserted, and the flask replaced on the balance. If the balance is still in equilibrium, the flask was correctly graduated in the first instance. If the balance is not in equilibrium, the flask must be regraduated in the manner described for graduating the flask for measuring. When the corrected volume is obtained, that quantity will be delivered by the flask when it is filled exactly to the delivering mark, and is allowed to drain for fifteen seconds after pouring out the liquid. 74 THE GAS ENGINEER'S LABORATORY HANDBOOK. Test Mixers for Measuring Cylinders. 38. Test mixers are tall cylindrical vessels, standing on a broad base, and provided with stoppers, as shown in Fig. 28. They are used to prepare test acids, test alkalies, C^i and similar solutions, by the dilution of strong solu- JTP^' tions to others of a fixed strength. These vessels are usually graduated throughout the greater part of their length, the graduations varying from 1 to 10 c.c. each, according to the size of the vessel. A measuring cylinder may be calibrated by running distilled water into it from a previously calibrated burette. Pipettes. 39. Pipettes are glass tubes for transferring specific quantities of liquid from vessel to vessel. They are of two kinds ; one has a bulb, and one FIG. 28. mar k on the neck above the bulb, as shown in Fig. 29. This form is used for the accurate mea- surement and delivery of one specific quantity of a liquid, such as 5, 10, 20, 50, or 100 cubic centimetres. The other form is shown in Fig. 30, and consists of a narrow tube having a scale engraved upon it lengthwise, similar to the scale of a burette, and is used for delivering a varying quantity of a liquid. Pipettes are chiefly used to measure with exactness any specified quantity of a solution which is to be subjected to volumetric analysis. The lower end of a pipette from which the solution flows out ought not to be more than ^ inch in diameter. The upper end, which is closed by the finger when in use, should be narrowed by melting and thickening the glass, and then be cut square across, and ground or fused flat, so that the finger can press firmly upon it. In order to fill a pipette, it is held near the top by the thumb and middle VOLUMETRIC ANALYSIS. 75 \ n finger of tlie right hand, and is then placed in the liquid it is wished to measure out. The mouth is then applied to the uy per end, and the liquid gently sucked up into the pipette. During this operation, the progress of the rise of th 3 liquor in the pipette should be carefully w; tched, and as soon as it is slightly higher th tn the top of the graduation, the instrument is removed from the mouth, and firmly closed w th the top of the first finger. During the pro- ce ts of filling, it is necessary to see that the point of the pipette always remains below the surface of the liquid, otherwise air will enter the pipette, ai d may force the liquid up into the mouth, which, w th corrosive liquids, might lead to serious consequences. An effectual means of preventing STI 3h an accident is to commence by dipping the p(. int into the liquid, and holding it in the same b} the left hand during the process of filling. The pipette having been filled in this way, it is held over the vessel from which it has been filled, FIGS. 29, 30. th 3 pressure of the finger on the top is relaxed, and th 3 liquid allowed to flow slowly out, until the bottom of the moniscus of the liquid in the pipette exactly coincides with th>3 graduation mark. The finger is once more firmly placed on the top of the pipette, which is then held over the vessel intended to receive its contents, and the latter is delivered into such vessel. There are several methods of delivering the contents of a pipette. (1) The contents are allowed to run freely in to the vessel, without permitting the point of the pipette to touch the sides of the vessel, and without blowing out the last drop from the pipette. (2) After delivering the bulk of the liquid, the point of the pipette is made to touch the inside of the vessel, in order that capillary attraction may carry away the drop adhering to the point of the pipette. 76 THE GAS ENGINEER'S LABORATORY HANDBOOK. (3) While holding the pipette as last described, the mouth is applied to the top of it, and the drop adhering to the point blown out. The method generally adopted is to allow the liquid to flow out by its own momentum, and then to leave the pipette to drain for fifteen seconds with its point touching the inside of the glass vessel. The correctness of a pipette is tested by filling it up to the mark with distilled water at 15-5 C., allowing the water to flow out as just described into a tared beaker and weighing. The Burette. 40. This instrument, a conveni- ent form of which is shown in Fig. 31, is used for delivering an accurately measured quantity of any standard solution. It consists of a glass [/tube of equal diameter throughout its length, its upper end being slightly enlarged. The lower end is drawn out, and is closed, preferably by a glass stop- cock ; or the contracted end is connected by a short piece of india- rubber tubing with a small glass jet. The india-rubber tube is clamped with a brass pinch- cock, which closes or opens the tube at will. The burette is graduated throughout the greater part of its length into cubic centimetres, and tenths of a cubic centimetre. Burettes usually contain 50 or 100 c.c. when filled to the zero or highest graduation mark. The burette is filled by closing the lower end, and FIG. 31. VOLUMETRIC ANALYSIS. 77 ponring in the solution from the top, until it rises above the highest graduation ; the stop-cock below is then opened, an 1 the liquid is allowed to run out until the bottom of the im niscus just touches the graduation mark. It is necessary to be careful when filling the burette, to se<- that no air-bubbles are left in the stop-cock, since, when the burette is in use, it is assumed, when the stop-cock is jpened, that only liquid flows out, and the volume due to ar y air-bubble which might pass out, would be counted as so ni ich liquid. The volume of liquid used in any determina- tii n is measured by the amount the liquid surface in the burette has been lowered. In order to fa militate exact readings, an Erdmann's float, Fig. 32, a little smaller in diameter than the burette, is ot ;en used. The float sinks with the liquid, and tte reading on the burette, corresponding with the horizontal mark on the centre of the float, is tl e one taken. When an Erdmann's float is not employed, the eye may be assisted in reading the graduations on tie burette by using a small card, the lower half oi which is blackened, the upper portion being left white. On holding the line of division between the black and white about an eighth of an inch below the surface of the liquid, and bringing the eye on a level with it, the meniscus can then be seen by transmitted light, bounded below by a sharply defined black line. A burette may be calibrated by filling it with distilled water at 15 5C., the tube below being carefully emptied from air-bubbles. Successive quantities of water, in separate instalments of 5 c.c., are run off" into a tared stoppered flask, which is then re; weighed, and from which it is ascertained whether the weight in grams of water run out is exactly equal to the volume in c.c. as indicated by the burette. If the reading of the burette is found to be incorrect, a 78 THE GAS ENGINEER'S LABORATORY HANDBOOK. table should be made out showing the true capacity up to each graduation, the error between any two successive weighings being equaHy distributed between the gradua- tions included in the 5 c.c. For example, in calibrating a burette from to 10 c.c. the following were the weights of water obtained. From to 5 c.c. 4-9672 grams; and from 5 to 10 c.c. 4-9850 grams. The first c.c. division on the burette corres- 4* 9672 ponds, therefore, to = 0-99344 c.c.; the second divi- sion corresponds to 0*99344 x 2 = 1-98688 c.c.; and so on for the first five divisions. 4* 985 In the same way, the sixth division will contain o = 0-99700 c.c. Consequently the first six divisions of the burette will contain 4-9672 -f 0-99700 = 5-9642 c.c., and seven divisions will contain 5-9642 -f 0-9970 = 6-9612 c.c. All measuring vessels should be in relative corres- pondence ; that is, although they may not exactly contain the required quantity of water at standard temperature, they should relatively agree. For instance, two quarter- litres, when poured into the half-litre flask, should fill it to the graduation mark. The 50 c.c. pipette, when its contents are delivered into the half-litre flask ten times in succes- sion, should also fill it to the graduation mark. 20 c.c. delivered into the burette from a 20 c.c. pipette, should in- crease the volume of liquid already in the burette by 20 c.c., and so on. NOTES ON THE PREPARATION OF SOLUTIONS. Normal Solutions. 41. The solutions employed in volumetric analysis are termed standard solutions. The connecting link between measurement and weight, VOLUMETRIC ANALYSIS. 79 as applied to volumetric analysis, is that of the combining weight of the reagent used, or of some single fraction of the co nbining weight. The following short table gives the combining weight of a few reagents, which are in common use in volumetric an ilysis. Name of reagent. Symbol. Molecular weight. Combining weight or hydrogen equivalent. Ai monia Ba 'ium hydrate NH 3 BaO, H 2 O BaCO 3 17-0 171-0 197 17-0 85-5 98-5 Ca cium hydrate , oxide CaO,H 2 CaO 74-0 56-0 37-0 28-0 carbonate CuC0 3 I 100-0 127-0 50-0 127-0 Po assium hydrate So- lium hydrate . KHO NaHO 56-0 40-0 56-0 40-0 Na 2 CO 3 106-0 53-0 H} drochloric aoid Niricacid Ox xlic acid (anhydrous) (crystallised) .. Su phuric acid HCl HN0 3 H 2 C 2 4 H 2 C 2 4 ,2H 2 H 2 S0 4 36-5 63-0 90-0 126-0 98-0 36 5 68-0 45-0 63-0 49-0 Standard solutions are made up in terms of the com- bining weight or hydrogen equivalent of the active reagent, as shown in the last column of the above table. Solutions are termed " normal " (N) when one litre, at 15-5 C. = 60 F., contains the hydrogen equivalent of the active reagent, weighed in grams (H = 1). 42. Semi-normal, deci-normal, and centi-normal solutions are also used, being respectively J, T ^, or T ^, of the strength of a normal solution. They are usually designated N N . N . .. 2 ' ib' and ioo solutlons - It will be seen from the above table, that in the case of univalent substances, such as iodine, hydrochloric acid, 80 THE GAS ENGINEER'S LABORATORY HANDBOOK. s6dium hydrate, &c., the equivalent and the atomic (or, in the case of salts, molecular), weights are the same ; thus a normal solution of nitric acid must contain 63 grams of acid in a litre of the solution, and sodic hydrate 40 grams. In the case of bivalent substances, such as barium hydrate, oxalic acid, sodium carbonate, sulphuric acid, the equivalent is one-half of the atomic (or, in the case of salts, molecular) weight ; thus a normal solution of sulphuric acid would be made by adding 49 grams of pure strong acid to distilled water, cooling the mixture, and then diluting to 1 litre. Also, in the case of trivalent substances, such as phos- phoric acid, a normal solution of sodic phosphate would Q t Q be made by weighing - = 119-3 grams of the substance, dissolving in distilled water, and diluting to the meabure of 1 litre. 43. In the preparations of solutions for volumetric analy- sis, the value of a reagent, as expressed by its equivalent hydrogen weight, must not always be regarded, but rather its particular reaction in any given analysis. Thus, when using a solution of potassium permanganate (KMn0 4 ), as an oxidising agent, it is the available oxygen which has to be taken into account ; and consequently, when preparing a normal solution, one-fifth of its molecular weight, 1 *)& = 31 * 6 grams, must be contained in the litre. The great convenience of this equivalent system is, that the numbers used as coefficients for calculation in any analysis are familiar, and for technical purposes the plan allows the use of all solutions of systematic strength, and simply varies the amount of substance tested, according to its equivalent weight. Thus, normal solutions of the following contain, per litre VOLUMETRIC ANALYSIS. 81 Gram?. Sulphuric acid 49 Hydrochloric acid 36' 5 Nitric acid 63 Anhydrous carbonate of soda 53 Sodic hydrate 40 Ammonia 17 100 c.c. of any of the normal acid solutions in the above lisi should exactly neutralise 100 c.c. of any of the normal allx-ilies, or the corresponding amount of the pure substance wHch the 100 c.c. contain. This system is especially use- ful in determining the amount of pure substance in com- me -cial samples. For example, supposing we are buying cos unercial sulphuric acid, and wish to know the exact pei 3entage of pure hydrated acid in it ; 4*9 grams are we ghed out, diluted with water, a little methyl-orange sol ition added, and normal alkali run in from a 100 c.c. bui ette until saturated ; the number of c.c. of normal alkali usc:l will show the percentage of real acid present. Sup- posing 62-6 c.c. are required = 62 '6 per cent. 44. In order to obtain the percentage of pure substance in uny commercial article, such as alkalies, acids, and various sails, by means of the normal solutions just described, the following method is adopted. With normal solutions, y 1 ^- or -2^ (according to the atomicity) of the molecular weight in gra ns of the substance to be analysed, is weighed out for titration, and the number of c.c. required to produce the desired reaction will represent the percentage of the sub- star.ce whose atomic weight has been used. With decinormal solutions, y^ or ^J^ of the molecular weight in grams is taken, and the number of c.c. will, in a similar manner, represent the percentage. It may sometimes happen, however, that from the nature of the substance, or from the fact of its being in solution, that the above per- centage method cannot be employed. For instance, supposing we have a solution containing an unknown quantity of caustic soda, and we desire to know G 82 THE GAS ENGINEER'S LABORATORY HANDBOOK. the strength of the same. We weigh, or measure out, a definite quantity of the solution, add a few drops of methyl- orange solution, and run in from a burette, normal acid until the solution is saturated, as shown by the change of colour. Supposing 38 c.c. of normal acid are required, then the molecular weight of sodic hydrate being 40 : (Na = 23, H = l,0 = 16, = 40), 100 c.c. of normal acid will saturate 4*0 gram ; therefore, as 100 c.c. are to 4*0 gram so are 38 c.c. to #, =1-52 gram NaHO. The simplest way is to 100 multiply the number of c.c. of test solution required in any particular determination by the y^^ ( r nrVtr i" ^ e case f a bivalent substance) of the molecular weight of the substance sought. This will at once give the amount of substance present. 45. Although solutions based on the above mentioned systematic principle, are those commonly adopted in analy- tical work, we require in addition, in the carrying out of some of the special determinations necessary in gas works, empirical solutions, based on the English system of weights and measures, and which are prepared to suit particular cases, not being based on any broad general principle. The solutions employed in the determination of the amount of ammonia in gas by the Eeferees' method, and in the com- mercial valuation of ainrnouiacal liquor, are cases in point. These are described in the Special Part. 46. In the preparation of standard solutions, it is neces- sary to bear in mind the fact, that saline substances on being dissolved in water, have a considerable effect upon the volume of the resulting liquid. The same thing occurs when mixing solutions of various salts or acids with each other. In the case of strong solutions, the condensation in volume is, as a rule, very considerable, and therefore in pre- paring such solutions for volumetric analysis, or in diluting such solutions to a given volume, for the purpose of remov- ing aliquot portions subsequently for examination, sufli- VOLUMETKIC ANALYSIS. 83 cien ; time must be afforded for the solutions to attain their constant volume at the standard temperature. TABLE FOR THE SYSTEMATIC ANALYSIS OF ALKALIES, ALKALINE EARTHS, AND ACIDS. (Sutton.) Substance. Formulae. Atomic weight. Quantity to be weighed so that 1 c.c. normal solution = 1 per cent, of sub- stance. Normal factor.* gram. Soda Na O 62 o'l 0-031 Sodic hydrate NaHO 40 4-0 040 carbonate .. Na 2 CO 3 106 5-3 0-053 bicarbonate NaHC0 3 84 8-4 0-084 Pota,- i K 2 O 94 4-7 0-047 Potat sic hydrate . . KHO 56 5-6 0-056 carbonate K,C0 3 138 6-9 0-069 bicarbonate KHCOg 100 10-0 o-ioo A mm >nia NH 3 17 1-7 0-017 ,. carbonate (NH 4 ) 2 C0 3 96 4-8 0-048 Lime CaO 56 2-8 0-028 Calcic hydrate CaO,H 2 74 3-7 0-037 ,, carbonate . . CaCO 3 100 5-0 0-050 Barium hydrate .. BaO,H 2 171 8-55 0-0855 i arbonate BaCO 3 197 9-85 0-0985 Nitric acid .. HN0 3 63 6-3 0-063 Hydrc chloric sicicl HC1 36-5 3-65 0-0365 Sulpli me acid Oxalic acid H 2 S0 4 H 2 C 2 4 ,2H 2 98 126 4-9 6-3 0-049 0-063 * This is the coefficient by which the number of c.c. of normal solution used in any analysis is to be multiplied, in order to obtain the amount of pure substance present in the material examined. Storage of Standard Solutions. 47. Standard solutions may be kept in any tightly stoppered bottle of a suitable size. A " Winchester quart " is a handy size for most purposes. The bottle should be clean and dry, and after the solution has been poured in the bottle should be immediately tightly stoppered, and labelled with the name, strength, and date when the solu- G 2 84 THE GAS ENGINEER'S LABORATORY HANDBOOK. tion was made. It is necessary that the bottle should be quite full, otherwise internal evaporation and condensation will cause drops of pure water to form on the upper part of the inside of the bottle. When using the solution therefore, it is necessary to well shake the bottle before pouring out, so as to mix this water with the solution from which it originated. This loss by evaporation may be lessened by storing the solutions in a cool place. Some solutions are affected by light, and in consequence should be kept in a dark cupboard, rather than upon open shelves where they would be exposed to the light. INDICATORS. 48. In many analyses it is necessary to add a substance which shall act as an indicator of the end of the process ; such, for instance, is cochineal or methyl-orange in alkali- metry, potassium chromate in silver and chlorine estimations, and starch in determinations where iodine is employed. There are other processes, the final termination of which can only be determined by an indicator separate from the solution. An example is found in the estimation of iron by potassium bichromate, in which case a drop of the solution is brought into contact with another drop of a solution of ferricyanide of potassium, on a white tile ; as soon as a blue colour ceases to appear when the two liquids are brought into contact, the end of the process is reached. INDICATORS USED IN THE VOLUMETRIC ESTIMATION OF ACIDS AND ALKALIES, Cochineal Solution. 49. This indicator possesses the advantage of not being affected in colour by C0 2 . In order to prepare it, digest about 10 grams of powdered cochineal for several hours at a gentle heat in a litre of weak spirit, composed of 200 c.c. VOLUMETRIC ANALYSIS. 85 of n.ethylated spirit, with 800 c.c. of water. The clear liquid, when decanted off, is ready for use. Its normal colour is yellow, and this is changed to reddish- violet by alkalies. This reddish-violet solution of cochineal is changed to yellow again by mineral acids, but it is not so easily acted upoa by weak organic acids. It should not be used in the presence of compounds of iron or aluminium, or of acetates. It is well adapted for the estimation of ammonia in gas liqi or (Will's Test), and in the analysis of ammonium sul >hate. Methyl-Orange, Tropseolin D, or Orange III. 50. This indicator also is not affected by C0 2 . Alkalies change its colour to yellow, and this changes to red on the addition of an acid. The solution does not change or de- compose on keeping. It is prepared by dissolving 1 gram of methyl-orange in ->owder (procurable from the operative chemist) in a small quantity of methylated spirit, and then making the solution up to 1 litre with methylated spirit diluted with its own vol iime of water. Methyl-orange is employed in the estimation of ammonia in _^as, and in the estimation of free ammonia in gas liquor. Litmus Solution. 51. This solution is used both for acids and alkalies. The solution is most sensitive when it is of a purple colour. The presence of an excess of acid, causes the litmus to assume a bright red tint; excess of alkali produces a pure blue tint. In order to prepare a solution of this indicator, all the colouring matter is first extracted from solid litmus by re- peatedly digesting it with hot water. The solution thus obtained, is evaporated to a moderate bulk, and a slight excess of acetic acid added, in order to 86 THE GAS ENGINEER'S LABORATORY HANDBOOK. convert all carbonates present into acetates. The solution is then again evaporated over a water-bath, until it assumes a pasty condition, when an excess of methylated spirit is added. The spirit precipitates the blue colouring matter, red colouring matter, together with the alkaline acetates remaining in solution. The precipitate is transferred to a filter, and washed with spirit. The pure colouring matter is then dissolved in warm water, and the solution, after being rendered purple by the addition of a little dilute nitric acid, is ready for use. Litmus solution must always be kept exposed to the air or it loses its colour, regaining it on exposure. A convenient method of securing this exposure, is to have the cork which closes the bottle containing the litmus solution, made with several deep grooves. The great defect of litmus as an indicator is that the presence of free car- bonic acid interferes with the production of the blue colour. Accordingly, in titrating an acid solution by means of an alkaline carbonate, the litmus solution changes from blue to purple before the point of neutralisation by the acid is arrived at. This is caused by the liberated C0 2 acting upon the litmus. The C0 2 can be removed by boiling the liquid, and the neutral point is just passed, and the reaction is con- sequently complete, when the solution has assumed a bright red colour, which does not alter on boiling. PJienol-phthalcin. 52. This indicator, when in the form of an alcoholic solution, is without colour, but on adding a few drops to an alkaline solution the liquid changes to a beautiful red colour, which is destroyed by acids. Phenol-phthalein cannot be employed in the presence of C0 2 , or of salts of ammonia. Phenol-phthalein is employed in the estimation of C0 2 in coal gas (Sheard's method). VOLUMETRIC ANALYSIS. 87 It is necessary to bear in mind that in judging the end of a re-action by means of an indicator, or of any final change of colour, it is requisite in all cases, that the one liquid should be run into the other in the same order. The liqi lids must not be reversed, or the results will fail to corres- poi d. For instance, if cochineal is used as an indicator in sta idardising the acid, by running the acid into a standard sol ition of caustic soda, in all processes of estimating acid by the alkali, where cochineal is used as an indicator, the aci I must be run into the alkali, and not the alkali into the aci I. ANALYSIS BY SATURATION. (ALKALIMETRY AND ACIDIMETBY.) Preparation of Normal Acid and Alkaline Solutions. 53. In this particular branch of volumetric work, it is vei y essential to have at least one standard acid and alkali prt pared with the greatest accuracy, and from which all others can be standardised. Sulphuric acid is best adapted for preparing the normal acid, as it can easily be obtained pure. Normal sulphuric acid is i ,ot affected by boiling. For the standard alkali, sodic carbonate is to be preferred, inasmuch as it can readily be obtained in a perfectly pure state, or can easily be made by igniting the pure bi- carbonate. The chief difficulty in times past with sodic carbonate has been, that, when using litmus as indicator, it was neces- sary to carry on the titration at a boiling heat, in order to get rid of the C0 2 , which hindered the pure blue colour of the indicator, notwithstanding the alkali may have been in excess. This difficulty is now got rid of by the employment of methyl-orange as indicator. But in case methyl-orange is 88 THE GAS ENGINEER'S LABORATORY HANDBOOK. not available, litmus may be made to give accurate results, if the saturation is carried on by rapidly boiling the liquid in a thin flask for a minute after each addition of acid, until the point is reached when one drop of acid in excess gives a pink red colour, which does not disappear on further boiling. Normal Solution of Sodium Carbonate. (53 grams Na 2 CO 3 per litre.) 54. This solution is prepared by dissolving pure Na 2 C0 3 in distilled water, so as to make a solution containing 53 grams Na 2 C0 3 per litre. It is consequently a semi- molecular solution. If we dissolve Na 2 C0 3 in water, the volume of the liquid contracts, therefore, we cannot obtain the correct weight per litre by simply dissolving the salt in the measured volume of water. After adding the salt to a quantity of water contained in the litre flask, the solution must be allowed to attain the normal temperature, and then suffi- cient distilled water added as will bring the volume up to the litre mark, the contents of the flask being well mixed. The easiest way of obtaining pure Na 2 C0 3 , is to ignite the bicarbonate. It is necessary in the first place to see that the bicarbonate is pure, by testing it for chlorides and sulphates. This is effected by shaking a portion of the powder with distilled water in a stoppered bottle, and then testing the clear liquid. Should traces of these impurities be found, about 100 grams of the powder are shaken up with three or more separate quantities of distilled water, the salt being allowed to settle, and the supernatant liquid decanted off. As soon as the washings are free from sulphates or chlorides, the salt is drained on a porous tile, and dried by pressing it between sheets of dry blotting or filtering paper. About 85 grams of the salt prepared as above are spread in a thin layer inside a weighed platinum or VOLUMETRIC ANALYSIS. 89 porcelain dish. The dish and its contents are then heated by :neans of a Bunsen burner to dull redness for about 10 mil utes, care being taken that the salt does not fuse or frit. The dish is now placed in the desiccator to cool, and wh 3n cold weighed ; it is then heated, cooled, and weighed agf in, until the weight is constant. The weight of pure Na C0 3 is then obtained by subtracting the weight of the dis i. The standard solution of Na 2 C0 3 may now be prc pared from the above by either of the two following me .hods. (1) The volume of the normal solution which should coi tain the weight of Na 2 C0 3 obtained as above is found by calculation. For example, if the sodium carbonate weighed 54 -6 grams, it will require to be dissolved in 54-6JL1000 53 The salt is transferred without loss to a beaker, the dish be ng afterwards rinsed out with successive small quantities of distilled water into the same beaker. The salt is well eti -red to assist solution, and when all has dissolved, the solution is poured into the litre flask, the beaker being also well rinsed out several times with small quantities of distilled water, and the rinsings added to the contents of the flask. The solution is now nearly made up to the mark on the flask with distilled water at a temperature of 60 F., the contents of the flask being well mixed by shaking. The stopper of the flask is then inserted, and the solution is again vigorously shaken to ensure thorough mixing. The contents of the flask are now made up to the exact volume by placing the flask on a level surface, and slowly pouring in water until the bottom of the meniscus just touches the litre mark on the flask. The additional volume of water required, amounting in the present example to 30*2 c.c., is then run into the flask from a burette, and the liquid is again thoroughly mixed by inverting the flusk and giving it a rotary motion. DO THK CMS KNOTNKKH'S LAMOUATOKY HANDBOOK. (2) Exactly 58 grams of Na 2 C0 3 , prepared as above, are weighed out and dissolved to a litre. The weight mav ho eonvoniently obtained by placing the dish on one seale-pan of Hio halanee, and adjusting the uoight. on the other pan so that tiny a iv equal to the weight of the dish -f 63? small portions ol' the sail, are then removed with lln |>oint of a penknife or spatula until equilibrium occurs. The salt is then dinBolved in a beaker as in (1), and on adding water to the litre mark a normal solution will be obtained. Normal Sulphuric Add. 56. This solution is a semi-molecular one (~r) (sulphuric being a bibasio acid) and should contain 40 grams of pure I I.,S( ), per lit.ro. H is prepared l>y adding a suitable <|iianl it yof pure strong oil of vitriol to distilled water contained in ft thin jjlaNH voNsol, thon eooliiii;- th(> hot. solution, and sulso|iiont.ly determining its strength by titrating a portion of it against the normal sodium carbonate solution. When the strength has been accurately determined, the amount of water re- quired to dilute it, in order to reduce it to the 'proper strength, may be easily calculated. Measure out ,'50 e.o. of pure, strong oil of vitriol, and run it into about. 'JOO o.o. of distilled w.'itor contained in a thin ^lass fl>ek. On adding the aoid great heat will b* given out; mix the solution woll, and eool it, by shaking it round in tli> tlask. at th(^ SMIHO timo allowing a stream of water from a water tap to flow over the exterior. When cooled down to the normal tempera! ure, the liquid ill the tlask is made up to exactly 1 litre with distilled water ; it is then thoroughly mixed again, and a. portion is titrated by normal sodium carbonate solution, as follows. A clean 20 0.0. pipette is partially filled with the normal odium carbonate solution by suction j the liquid is well shaken round in the pipette, and is then allowed to run away into the sink, or other receptacle for waste. The pipette is again charged with the VOLUMETRIC ANALYSIS. 91 s< lution, Imt this lime exactly to tho 20 c.o. "mark, by (hawing up more than tho required volume, rapidly placing tl o first finger on tho top of tho pipette, and slowly allowing 11 o excess to flow out. The 20 c.c. of solution are then delivered into a clean Leaker. Of course, in tho event of II o interior of tho pipette being perfectly clean and dry a the commencement, there is no occasion for tho prolimi- n iry rinsing with tho solution* inasmuch as in this caso tlero would not bo any danger of tho standard solution l> -ing diluted with tho water adhering to tho interior of llio pipette. A burette is next charged with tho acid prepared as (1 'Scribed, care being taken that the interior of tho instrument it perfectly dry ; if not, it should bo first rinsed out with a portion of tho solution, which is afterwards thrown away. I is also necessary to see that there arc no air bubbles i' maining in tho tap at tho bottom of the burette; should t tore be any, they may bo got rid of by rapidly running a p irtion of the solution out of tho burette, and then refilling II 10 sumo. Tho solution in tho beaker, previously diluted \\ ith distilled water, is placed underneath the burette, upon .i white tile, or a clean filter paper, a small quantity of n ethyl-orange solution is added, and tho acid run slowly in fiom the burette, tho contents of tho beaker at tho same time being constantly stirred. Tho acid i run in until a single drop causes the golden yellow colour of tho solution to suddenly ehango to a reddish-brown tint, shewing that tho solution in tho beaker has been exactly neutralised by tho acid. Tho volume of acid required to effect the neutra- lisation JH then read off on tho burette, the result being chocked by repeating tho titration in exactly tho same manner with a fresh quantity of tho sodium carbonate solution. Tho two titrations should not differ by more than 0-1 c.o. If more than 20 c.o. are required, the acid is too woak ; if less, too strong. It will generally be found that the acid is too strong, and will require diluting to tho 92 THE GAS ENGINEER'S LABORATORY HANDBOOK. normal strength in the following manner. Supposing that the 20 c.c. of the alkaline solution required 18*6 of the acid solution to neutralise it; then each 18 '6 c.c. of acid would require to have sufficient water added to it to make it up to 20 c.c., in order to produce a standard acid exactly equivalent to the standard alkali. Therefore, to make a litre (1000 cubic centimetres) of normal acid, 930 c.c. of the acid, prepared as above, are measured into a litre flask, and then made up to 1 litre with distilled water. After thoroughly mixing the acid diluted in this manner, it should again be titrated against the alkaline solution, as already described, in order to be certain of its accuracy. It is necessary to bear in mind the fact that standard solutions, even when accurately prepared in the first instance, will gradually become more or less altered in strength after being stored for some time. A standard acid solution will therefore require to be tested occasionally, by titration against a freshly prepared standard sodium carbonate solu- tion, as just described. In the event of the acid solution having altered in strength, it is not absolutely necessary (if the error is small) to bring it back to the correct standard. In place of doing this, a correction is made use of when working with the solution. For instance, if 19*7 c.c. of acid are needed instead of 20, in order to neutralise 20 c.c. of a freshly prepared sodium carbonate solution, the number of c.c. of acid needed in any 20 titration will require to be multiplied by = 1-015, in JL y * i order to find the number of c.c. which would have been needed if the acid had been of the correct strength. The " factor " required for this correction should be marked on the label of the bottle, together with the date. The operation of determining the amount of acid in a solution in the above manner is termed " acidimetry." VOLUMETKIC ANALYSIS. 93 Normal Hydrochloric Acid. (36-5 grams per litre.) 56. Pure strong hydrochloric acid is diluted with dis- ti led water until it has a specific gravity of I'lO at 60 F., at tested by a hydrometer. About 165 c.c. of this acid are d luted to 1 litre. Its exact strength is then determined by titration against normal sodium hydrate solution (58), and the solution is brought to the standard strength by dilution, a ; in the previous examples. Hydrochloric acid is useful on account of its forming sol- u ble compounds with the alkaline earths, but it has the dis- a Ivantage of volatilising at a boiling heat. The hydrochloric a iid from which standard solutions are made, should be free f 'om chlorine gas or metallic chlorides, and when evaporated i i a platinum vessel should leave no residue. Normal Oxalic Acid. (63 - grains per litre.) 57. This solution is prepared by dissolving 63 '0 grams of jure, not effloresced, crystallised oxalic acid in distilled water, and making up to 1 litre, the solution being after- wards checked with normal alkali. The solution is apt to deposit some of the acid at low temperatures, but keeps well if preserved from direct sunlight, and may be heated without volatilising the acid. It cannot be employed in alkalimetry when methyl-orange is used as the indicator. This solution may be used in testing for caustic lime in the presence of a carbonate* Normal Sodium Hydrate Solution. (40 grams per litre.) 58. Caustic soda suitable for the preparation of the above may be obtained from two sources, either by the action of metallic sodium on boiled and cooled distilled water, or from caustic soda purified by alcohol. 94 THE GAS ENGINEER'S LABORATORY HANDBOOK. In order to obtain it from metallic sodium, about 25 grams of clean metallic sodium in small pieces are weighed out, and placed in distilled water contained in a platinum or silver dish. A clock glass is placed over the dish after the addition of each piece of sodium, in order to prevent loss of the NaHO by spurting, the chemical action set up by each separate piece of sodium being allowed to subside before a fresh piece is introduced. In this way a strong solution of NaHO is obtained. If caustic soda purified by alcohol is used, then about 44 grams should be weighed out, and dissolved in distilled water. If either of these solutions are diluted until they have a specific gravity of about 1 * 05, then the solution will contain about 50 grams per litre. The water employed in diluting should be distilled, and recently boiled and cooled, so as to be free from C0 2 . 20 c.c. of the solution thus prepared are then titrated against 20 c.c. of the normal acid (55), em- ploying methyl-orange as indicator, and the solution is reduced to the exact normal strength, by diluting it with the necessary quantity of distilled water. ESTIMATION OF AMMONIA. 59. This determination comes under the head of alkali- metry. The principles upon which it is based apply to the determination of the ammonia in gas, and to the analysis of ammoniacal liquors and sulphate of ammonia. By alka- limetry is meant the method of determining the amount of alkali in a substance by saturation with a standard acid. The quantity of free ammonia in a solution of ammonia and water may be determined by means of standard acid and methyl-orange soluti' -n. A definite quantity of the solution, say 10 c.c., is transferred to a small tared flask and weighed ; in this way its absolute weight and specific gravity are ob- tained in one operation. If the 10 c.c. weighed 9 -449 grams, its specific gravity would of course be 0*9449, water = 1. VOLUMETKIC ANALYSIS. 95 The weighed quantity of the ammonia, is then diluted with 6 or 8 times its bulk of distilled water, and titrated direct with the standard acid. Ammonia in combination may be determined by expel- ling it by means of caustic soda or lime, collecting the ev >lved ammonia in a known volume of standard acid, and de ermining the excess of acid by standard soda solution. The following shows the method of determining the an ount of ammonia in ferrous ammonium sulphate: Ft(NH 4 ) 2 (S0 4 ) 2 6H 2 0. The apparatus employed is shown in Fig. 33. It consists of a large flask A, having a capacity of about 30 fluid ounces, fit ,ed with a cork pierced with two holes, into one of which is inserted a tube B about 4 inches long, and half an inch in ernal diameter, containing pieces of broken glass, and cL -sed at the top with a cork, through which passes a small tuoe the diameter of a goose-quill, and about 2 inches long. I:u to the other hole is fixed a bent tube, C, one end of which pa sses nearly to the bottom of the flask A, the other being co inected, by an india-rubber joint, to a similar tube passing through a cork adapted to a second flask E, of about 15 ounces capacity ; a small funnel D is also passed through the co :k so as to nearly reach to the bottom of the flask E. A sand bath F, supported on a tripod G, and a Bunsen burner H, com- pletes the apparatus. There are various modifications of this apparatus employed, but the one described is about the most simple. Having connected the whole apparatus together, as shown, 100 cubic centimetres of normal sulphuric acid are poured, by means of a pipette, through the tube B into the flask A, so as to thoroughly moisten with acid the pieces of glass in the tube B. About 1 gram of the pure salt is accurately weighed out, all crystals showing a yellowish tinge being re- jected. The quantity weighed out is dissolved in about 50 c.c. of distilled water, and the solution is poured into the flask E. About 50 c.c. of a strong solution of sodium hydrate is then poured i..to E, the funnel tube rinsed out with a little 96 THE GAS ENGINEER'S LABORATORY HANDOOOK. distilled water, and the stop-cock closed. The burner is then lit beneath the sand-bath, and the contents of the flask E brought to incipient boiling. Ammonia will then be disen- gaged, and will be absorbed by the acid in the flask A, and in the tube B. The tube C should only just touch the acid in the flask A. When the ammonia ceases to be absorbed, the FIG. 33. flame should be raised, and the contents of the flask E kept briskly boiling for ten minutes. The apparatus is then disconnected, and the pieces of glass in B rinsed with a little distilled water into the flask A. The liquid in A is then made up to exactly 200 cubic centimetres, and divided into two portions of 100 c.c. VOLUMETRIC ANALYSIS. 97 each. A 50 c.c. burette is then filled with normal soda sol ition, and one portion of the acid solution, coloured with a fev - drops of solution of cochineal, is titrated with the soda sol ition, the remaining portion being similarly treated. The tw ) titrations should not differ by more than -^ c.c. Example: 1 3723 grams of crystallised ferrous ammonium sul phate were weighed out, and treated as above described, thi ammonia evolved being absorbed by 100 c.c. of normal su' phuric acid. After absorption, the acid solution was made up to 200 c.c., and divided into two portions. 100 c.c. of this liq aid ( = 50 of the original solution) required 46 5 c.c. of no 'mal soda solution to neutralise it, consequently, the whole of the solution would require 46*5x2 = 93 c.c. of the soda sol ation. The number of c.c. neutralised by the ammonia = 10 ) 93 = 7-00, and since each c.c. of acid corresponds to 0- )18 gram of NH 4 , the weight of NH 4 in the salt = 0-018 X 7 = 0-126 grams, . . the percentage of NH 4 = > = 9-18, which agrees with the percentage demanded by theory, as. under. The molecular weight of Fe(NH 4 \(S0 4 ) 2 ,6H 2 = 392,. and this contains 36 parts by weight of NH 4 , consequently, As 392 : 100 :: 36 : x. x = 9-18. 60. Another example of the process of alkalimetry as apr plied to the requirements of gasworks, is that of the analysis of lime. In this case the lime, either in the form of carbonate or of caustic lime, is dissolved in an excess of standard hydrochloric acid, the volume of acid noted, and retitrated with a standard solution of caustic soda. The volume of caustic soda now re- quired to neutralise the acid is deducted from the amount of acid originally added ; the difference represents the amount of acid which has entered into combination with, the H. 98 THE GAS ENGINEER'S LABORATORY HANDBOOK. Each c.c. of acid used represents 0'028 gram of caustic lime, or 037 gram of calcic hydrate, or * 050 gram of CaC0 3 . This method of determination is termed the " residual " method. Full details are given under the heading "Analysis of Lime " in the Special Part. PROCESSES OF OXIDATION AND EEDUCTION. The principles involved in analytical processes based on the methods of oxidation and reduction are extremely simple ; substances capable of absorbing oxygen are dissolved, and titrated with another substance of known oxidising power, as for example in the estimation of iron in ferrous compounds by means of permanganate of potash. The iron is eager to receive oxygen, and the permanganate is willing to give it up. During the process of absorption, the permanganate instantly loses its colour on being added to the iron solution, and the whole mixture is colourless, but as soon as the iron has taken up its complement of oxygen, or has been converted into the higher state of oxidation as a ferric salt, the rose colour of the permanganate solution no longer disappears, as there is no more oxidisable iron to operate on. The follow- ing equation expresses what has taken place in the above reaction : lOFeO + 2MnK0 4 = 5Fe 2 3 + 2MnQ + K 2 0. The oxidising agents commonly employed are potassium permanganate, potassium bichromate, and iodine, the common reducing agents being sulphurous acid, sodic hyposulphite, stannous chloride and zinc. A great many combinations may be arranged with this variety of materials, so as to make this system of analysis very comprehensive. The following, however, are the most typical, and serve to show the applicability of the system to the requirements of a gasworks. (a) Permanganate of potash and ferrous salts, the rose VOLUMETKIC ANALYSIS. 99 col< >ur of the permanganate acting as indicator (may be used in ' he analysis of Weldon Mud). (fe) Bichromate of potash and ferrous salts, the end of the rea3tion being shown by the cessation of the blue colour wh 3n brought in contact with a solution of ferricyanide of pot ussium as indicator (may be used in the estimation of iron oxi les for purifying purposes, bog ore). (c) Iodine with starch as indicator (may be used in the est mation of SH 2 in gas liquor). PREPARATION AND USE OF POTASSIUM PERMANGANATE SOLUTION. 61. The molecular weight of potassium permanganate (1VJ n 2 K 2 8 ) = 316 0, and as shown in the equation previously given, 1 equivalent, or 316 parts by weight of the salt, is eq> lal to 10 equivalents, or 560 parts of iron. In order to pre- pa 'e a decinormal solution of permanganate, therefore, 3 '16 gr; ms of the pure salt may be dissolved in a litre of water. 17 85 c.c. of this solution should then peroxidise O'l gram of iroa; for 1000 c.c. being equal to 5*16 grams Fe, 17*85 c.c. arc equal to 0*1 gram. The above method, theoretically, should give a solution of the exact strength, but practically we cannot calculate the strength of such a solution, i.e. the amount of oxygen that it is capable of yielding, beforehand, on account of its instability. It is absolutely necessary, therefore, to determine the power of the solution by direct experiment before using it. It is consequently not essential to make the solution strictly decinormal, as its power is accurately determined by experiment before use. About 5 grams of the pure crystallised salt are dissolved in a small quantity of distilled water, and the resulting solution is diluted to 1 litre. The solution should be kept in a dark blue Winchester quart bottle, and when not in use should be stored in a cool dark place, as the liquid, if exposed to light, becomes weaker H 2 100 THE GAS ENGINEER'S LABORATORY HANDBOOK. owing to the decomposition of the salt; but a solution pre- served in the above manner may be kept for a long period without experiencing much alteration. The most accurate method of estimating the strength of the solution consists in determining the amount required to transform a known weight of iron from the condition of ferrous oxide to that of ferric oxide. This is effected as follows : About 1 gram of fine iron wire (flower wire), quite free from rust, is accurately weighed out into a 250 c.c. flask, and dissolved in about 100 c.c. of dilute sulphuric acid (1 part of acid to 6 of water). A small quantity of carbonate of soda is thrown into the liquid at the same time, in order that the air within the flask may be displaced by carbon dioxide. The flask is provided with a cork and bent glass tube, furnished with a brass pinch-cock ; the end of the tube dips beneath the surface of about 25 c.c. of water contained in a small flask, Fig. 34. Whilst the iron is dissolving, the pinch-cock closing the india-rubber connecting tube is taken off the latter, and slipped over the glass tube; by so doing there is a free communication between the two flasks. The solution of the iron may be accelerated by a gentle heat. The liquid is gradually brought to the boiling point, and kept briskly boiling for a minute or two, so as to expel the mixture of C0 2 and hydrogen ; the india-rubber tube is then immediately closed, and the burner removed. In the course of a minute or so the pinch-cock is again opened, when the water from the small flask is driven over into the solution of iron. In proportion as it passes over, boiling distilled water is poured into the smaller flask, until the larger one is almost filled. The india-rubber tube is then once more closed, the flask and its contents allowed to cool, and the volume of the liquid made up to the containing mark ; the stopper of the flask is inserted, and the liquid well shaken so as to cause it to be thoroughly mixed. Whilst the solution VOLUMETBIC ANALYSIS. 101 cools, a Mohr's burette fitted with a glass stop-cock (the permanganate solution gradually attacks india-rubber) is fi] st rinsed out with, and then filled with the permanganate solution prepared as above. When cool, 50 c.c. of iron solution are taken out of the flask, and poured into about 2 1 '0 c.c. of distilled water, contained in a beaker standing 01 i a sheet of white paper. The permanganate solution is tl en added, drop by drop, to the liquid, with constant FIG. 34. stirring, until the pink colour imparted to the solution remains. The permanganate is at first decomposed with great rapidity, but as the iron becomes oxidised the colour disappears more slowly, the rapidity of the change indicating the progress of the oxidation. The operation of standardising thus described should be repeated once or twice on successive portions of 50 c.c. of iron solution, and the mean of the results taken as representing the true value of the per- manganate solution. 102 THE GAS ENGINEER'S LABORATORY HANDBOOK. The reactions occurring in the foregoing are expressed by the following equations : 10FeS0 4 + 2KMn0 4 -f 8H 2 S0 4 = 5Fe 2 (S0 4 ) 3 + 2MnS0 4 + K 2 S0 4 + 8H 2 0, or, shortly, lOFeO + Mn 2 7 = 2MnO + 5Fe 2 3 . As an illustration, let us assume that we have weighed out 1 10 grams of iron, and dissolved it in 250 c.c. of liquid ; 50 c.c. of the solution, or a fifth part, would be equivalent to 0*2200 gram of iron. The iron employed, however, is not chemically pure, but if we assume that its impurities amount to 0*4 per cent, we shall not be far out. The amount of pure iron in the 50 c.c. is therefore 1 : 0*996 : : 0-2200 : x. x = 0-21912. We will further suppose that it has required 20 c.c. of the permanganate solution to produce a permanent coloration with the 50 c.c. of iron solution. Then, as 20 c.c. perman- ganate solution converts 0-21912 grams iron from protoxide to peroxide, 100 c.c. permanganate are equivalent to 1*09560 grams iron, or 1 c.c. to 0*010956 gram. The strength of the permanganate solution may also be determined by means of pure ferrous sulphate, precipitated from its aqueous solution by means of alcohol. Ferrous sulphate so prepared keeps unchanged for very long periods. Instead of FeSO , the double sulphate of iron and am- monium FeS0 4 (NH 4 ) 2 S0 4 -f 6H 2 may be employed. This contains exactly one-seventh of its weight of iron : 0- 7 gram of salt is equivalent to 1 gram of iron. The strength of the solution may also be determined by means of oxalic acid. Of the several substances which may be used for titrating the permanganate solution metallic iron is the best, but whichever method is adopted, it is absolutely necessary that VOLUMETRIC ANALYSIS. 103 th-3 solution to be titrated should contain free sulphuric acid. If there is a deficiency of free acid, the solution turns brown an d eventually a precipitate collects. It is necessary to exer- cii- e discretion as to the acid employed for acidulation. Nitric ac^d cannot well be employed under any circumstances. H /drochloric acid is liable to be oxidised and give off cl lorine, unless it is very considerably diluted, and this cl ange interferes with the accuracy of the result. Organic m itter also reduces the permanganate, and destroys the ac- CT racy of the determination. 62. Permanganate of potash may be employed in deter- n: ining the amount of Mn0 2 in the Weldon Mud employed ii gas purification. The method depends upon the fact than when Mn0 2 is ti sated with an acid, it is decomposed and oxygen liberated, w liich combines with any substance present ready to receive it In the present instance the substance is a ferrous salt in a] > acid solution. The amount of permanganate required to o: :idise a definite amount of the ferrous salt being known, on a< iding a weighed amount of the Mud to this solution, on again ti trating with permanganate, it will be found that we do not require so much of that solution as we did in the first ii stance, due to the fact that the oxygen of the Mn0 2 in the Weldon Mud, has supplied the place of the oxygen of the pormanganate solution. About 1-5 to 2 grams of "flower wire," perfectly free from rust, are accurately weighed out and dissolved in 100 c.c. of dilute H 2 S0 4 (1 of acid to 4 of water) in the apparatus shown in Fig. 34. About the same weight of the finely powdered Mud is then added to the iron solution, and the liquid is gently heated until the Mud is dissolved. The solution is then boiled, the water allowed to flow back, and the contents of the flask made up to 250 c.c. When cold, the amount of residual ferrous sulphate is determined by the permanganate solution. 104 THE GAS ENGINEER'S LABORATORY HANDBOOK. The following equation shows what takes place : 2FeS0 4 + Mn0 2 + 2H 2 S0 4 = Fe 2 (S0 4 ) 3 + MnSO + 2H 2 0. 112 parts of iron correspond to 87*0 of Mn0 2 . Example : 1*756 grams of flower wire were weighed out = 1-756 x 996 = 1-749 grams pure iron. This was dis- solved in the requisite quantity of dilute sulphuric acid, and 2.12 grams Weldon Mud added to the iron solution. When cold, the solution required 86-2 of permanganate solution. 1 c.c. permanganate = 0-010956 Fe; consequently 0-010956 X 86 2 = 944407 Fe. 1 749 - 944407 = 8046 gram of iron, has been oxidised by the manganese dioxide. Then as 112 : 87 : : 0-8046 : x. x= -625, and 2-12 : 0-625 : : 100 : 29-48. The Weldon Mud therefore contained 29-48 per cent, of manganese dioxide. USE OF POTASSIUM BICHROMATE SOLUTION. 63. When used as an oxidising agent in volumetric analysis, potassium bichromate always yields the whole of its oxygen to the hydrogen of the accompanying acid, a corresponding quantity of acidulous radical being set free, four-sevenths of this radical immediately combining with the potassium and chromium of the potassium bichromate, three-sevenths becoming available. Advantage is taken of this property in the estimation of iron (ferrous) salts. Thus, if a solution of a ferrous salt is acidified with a dilute acid (sulphuric) it is at once converted into a ferric salt on adding a solution of potassium bichromate in sufficient quantity, the reaction being as under. 6FeS0 4 + K 2 Cr A + 7H 2 S0 4 = 3Fe 2 (S0 4 ) 3 +Cr 2 (S0 4 ) 3 + K 2 S0 4 4-7H 2 0. VOLUMETRIC ANALYSIS. 105 From this it will be seen that one equivalent or 294 '42 parts of potassium bichromate will convert 6 equivalents, or 33-) parts of iron, from the ferrous into the ferric state. The fin il completion of the change of ferrous into ferric salt in th< > above reaction is ascertained by bringing a drop of tho solution in contact with a drop of solution of freshly prepared potassium ferricyanide, when no blue coloration w 11 be produced. So long as the faintest trace of ferrous su iphate remains, a drop of the liquid will give a blue colour w th the solution of potassium ferricyanide. The solution of bichromate can be standardised syn- tl otically. -^ of 294-42 grams = 4 '907 grams of the pure di y reagent, dissolved in distilled water to 1 litre, should gi ve a solution of which each c.c. peroxidises ^ Fe = 0'0056 gi am of iron. The principal impurity in the salt to be avoided is sr Iphate. In order to test for it, dissolve 1 gram in 10 per c( nt. hydrochloric acid, and heat with a few drops of alcohol, to reduce the Cr0 3 to Cr 2 Cl 6 . Dilute the dark-green solution, add chloride of barium, and allow to stand for twelve hours. Then decant off all except a few drops into another vessel, and dilute the slight residue with distilled water. The least quantity of sulphate of barium becomes visible. In order to dehydrate the salt it is powdered, and after having been kept near its fusing point for about 10 minutes, it is fused at the lowest sufficient temperature. On cooling it breaks up into numerous little fragments, and thus spontaneously assumes a convenient form for being weighed out. The titre, as calculated from the synthesis, ought to be perfectly correct ; but unfortunately there is no method for proving the absence of surplus chromic acid in the salt, which, if present, would alter the strength of the solution. It is necessary, therefore, to test the strength of the solution by direct experiment. For this purpose a solution of ferrous sulphate of known strength is required. This is prepared 106 THE GAS ENGINEER'S LABORATORY HANDBOOK. from a weighed quantity either of metallic iron, or of ferrous ammonium sulphate. In order to prepare the ferrous sulphate from metallic iron, about 1 gram of flower wire, free from rust, and in pieces of about an inch long, are accurately weighed out. Then fit up the apparatus shewn in Fig. 35. This consists of a 10-ounce round flask, provided with a perforated cork, through which passes a bent glass tube, one end of which passes just through the cork, while the other end dips below the surface of a little water contained in a small beaker. The flask is clamped in a slanting position upon a retort stand as shewn. The flask is now half filled with dilute sulphuric acid, a small piece of pure calc spar dropped in, and the cork and tube replaced. The calc spar will dissolve, giving off carbon dioxide, which will replace the air in the flask. When the effervescence has been kept up for some time, add the weighed wire, fit in the cork and tube, and arrange the apparatus as shewn in the figure. Heat the flask gently, and continue the heating until the wire is dis- solved. While the iron is dissolving, some cold water free from dissolved oxygen may be prepared by boiling some distilled water for a few minutes, and then rapidly cooling it by immersing the flask in cold water. As soon as the iron wire is all dissolved, cool the solution, transfer it rapidly to a 250 c.c. flask, and rinse out the flask in which the iron was dissolved several times with the air-free water prepared as above, into the measuring flask, finally making up the solution to the 250 mark with the air-free water. The titration is conducted as follows : A 50 c.c. burette is filled with the solution of potassium bichromate. 25 c.c. of the iron solution are measured out into a 10-ounce flask, and diluted with about 50 c.c. of air-free water. A number of drops of a very dilute solution of a freshly prepared solution, of potassium ferricyanide are then placed on a white tile by means of a glass rod. (The solution of ferricyanide is prepared when required from a compact crystal of the salt, VOLUMETKIC ANALYSIS. 107 previously washed with water in order to remove any ferro- cyanide. The solution does not keep long, especially if ex posed to the action of light. Before use it must be tested w th pure ferric chloride, with which it should give a brown colour, free from all shade of green.) The solution of bichromate is now run from the burette ii to the iron solution, mixing well after each addition. As tl e solution of bichromate is being run in, a drop of the iron st lution is taken out of the flask from time to time, by d pping a clean glass rod into the liquid, and bringing the d op adhering to the rod into contact with one of the drops o1 potassium ferricyanide solution on the tile. With the fi -st few drops a deep blue coloration will be produced, but a; the addition of the bichromate proceeds, the colour produced by the ferricyanide will gradually become fainter, a id will finally disappear. When this stage is reached the v )lume of bichromate solution used is read off. Two more similar titrations should then be made, the \v hole volume of bichromate solution required, as obtained ft om the first or preliminary titration, being run in at once \\ ithin half a c.c., and the last portion being added gradually, d 'op by drop. The results of these separate titrations should not differ by more than the tenth of a cubic centi- metre. It is necessary to multiply the weight of iron contained in the solution employed in this titration by 0'996, in order to correct it for impurities present in the wire orginally used, and from the weight of pure iron thus found, the exact strength of the standard solution of bichromate may be calculated. Example: 1*338 grams of flower wire were dissolved as above, and the solution made up to 250 c.c. 25 c.c. of this solution required 23 '6 c.c. of the standard bichromate solution. Therefore 236 c.c. of bichromate solution correspond to 1 % 338 grams of wire. 108 THE GAS ENGINEERS LABOKATORY HANDBOOK. 1-338 X 1000 Consequently 1000 c.c. will correspond to - ^og = 5-6695 grams of wire, which are equal to 5-6695 x 0-996 = 5*6468 grams of pure iron. Each c.c. of the bichromate solution is therefore equivalent to 0-005646 gram of Fe. In order to standardise the bichromate solution by means of ferrous ammonia sulphate, Fe(NH 4 ) 2 (S0 4 ) 2 6H 2 0, weigh out exactly seven grams of the substance, dissolve it in air- free water, and make up the solution to 250 c.c. as already described. The above weight of the salt will make a solution containing four grams of iron per litre, since the salt contains one-seventh its weight of iron. The solution is titrated in exactly the same manner as the solution of iron wire just described, with the exception that it is necessary to add a little dilute sulphuric acid to the iron solution, before running in the bichromate. The above method may be applied in the determination of the amount of iron contained in the oxide of iron employed in gas purification. 64. The following example shows the method of pro- cedure in the case of haematite, which is an iron ore very rich in ferric oxide. Weigh out about 2'5 grams of haematite iron ore in the state of very fine powder. Place the ore into a beaker, and add hydrochloric acid diluted with its own volume of water. Cover the beaker with a clock-glass, and apply a gentle heat for about half-an-hour. Then add a small quantity of distilled water, allow any portion not dissolved to settle down, and decant the solution through a filter. Treat the residue again with a small quantity of dilute acid, and decant the liquid through the same filter as before. By these operations all the iron will generally be extracted from the ore, but it is necessary to see whether this is the case or not by testing a drop of the solution with a solution of potassium sulphocyanide. Should a decided red coloration ensue, it will be necessary to treat the ore a VOLUMETEIC ANALYSIS. 109 tliLrd time with, acid, and continue this treatment until no more iron is extracted. When this stage is arrived at. the residue is transferred to the filter, and washed free from acid, using as small a quantity of water as possible during tl e operation. The iron which is present in the solution in tl e ferric state is now converted into the ferrous state. In order to effect this conversion, the solution and TV ashings are transferred to the flask shown in Fig. 35. S iould the whole of the liquid more than half fill the flask, FIG. 35. it will be necessary first to evaporate it down to the bulk required. Strong hydrochloric acid is then added, and fragments of granulated zinc free from iron are dropped into the flask. The acid will dissolve the zinc, with the evolution of hydrogen. A portion of the nascent hydrogen thus evolved will act upon the ferric chloride and reduce it to ferrous chloride, the colour of the solution changing at the same time from yellow to pale green. After a time the action will become rather sluggish ; it will then be necessary 110 THE GAS ENGINEERS LABORATORY HANDBOOK. to gently heat the liquid in order to promote the action, continuing the heating until all the zinc is dissolved. Before proceeding with the titration, it is necessary to see that the whole of the iron is converted into the ferrous state, by taking out a drop of the solution with a clean glass rod, and bringing it into contact with a drop of potassium sulphocyanide deposited on a white tile. No red colour, or only an exceedingly faint pink tint, should result. If a distinct red colour is produced, it will be necessary to add more zinc and hydrochloric acid, and the operation of re- duction continued as above until all the zinc has dissolved. When the zinc has dissolved, the solution is then again tested with sulphocyanide, and the same operations repeated until all traces of the ferric salt has disappeared. When the whole of the iron has been converted to the ferrous state, the solution is rapidly cooled out of contact with the air, and immediately made up to 250 c.c. with air- free water. The titration is then at once effected as in (63), and com- pleted as soon as possible. The value of the bichromate solution being known, from the volume of that solution used, the quantity of iron present in the ore may be calculated, and from that the percentage. USE OF STANDARD IODINE SOLUTION. 65. Iodine dissolved in iodide of potassium may be used for numerous volumetric determinations, amongst others that of H 2 S in aqueous solution, and upon this, a method for determining the H 2 S in gas liquor is based. In the present example we require a decinormal solution of iodine, and a solution of starch. 127 The solution of iodine should contain - = 12*7 grams of pure iodine per litre. 12-7 grams of pure resublimed VOLUMETKIC ANALYSIS. Ill iodine are accurately weighed out from a stoppered bottle in1 o a litre flask. About 20 grams of pure potassium iodide ar( then added,' together with about 250 c.c. of distilled wster. The contents of the flask are well shaken until the ioc ine is completely dissolved, when the flask is filled up to th ) mark with distilled water. Owing to the volatility of iodine, it is difficult to weigh out exactly the quantity requisite to make the standard so ution. In order to get over this difficulty, about 13 grams m ty be weighed out by difference, this is then dissolved in 1jl 3 manner above described, and made up to the volume n< cessary to furnish a solution of decinormal strength. T ms, supposing 12 '826 grams were weighed out, the ... 1000x12-826 vc lume oi the iodine solution required would be ToTy = 1009-9 c.c. This solution must be kept in a well stop- pi red bottle, and in a cool place. In order to prepare the starch solution, about 1 gram of st irch free from acid is made up into a thin cream with a si lall quantity of cold water. This mixture is then poured into about 100 c.c. of boiling water contained in a porcelain dish, the boiling being kept up for a few minutes after ac ding the starch. The liquid is then allowed to stand until cold, when the clear solution is poured off, being then ready for use. Starch solution freshly prepared is much more sensitive to the action of iodine than that which has been kept for some time, consequently, wherever practicable, the solution should be prepared immediately before use. When a dilute solution of sulphuretted hydrogen is brought into contact with free iodine the following reaction takes place. H 2 S + I 2 = 2HI + S. In tbis determination it is necessary that the solution con- taining the H 2 S to be determined does not contain more than 04 per cent, of the gas, otherwise there is a possibility 1]2 THE GAS ENGINEER'S LABORATORY HANDBOOK. of the change not taking place entirely according to the above equation. Measure out a portion of sulphuretted hydrogen water made by passing the gas into water, and titrate this with the decinormal iodine solution. This titration is made in order to give a rough idea of the quantity of SH 2 in the solution, should the latter contain more than 04 per cent, of the gas, the rest of the solution is measured out and diluted with air-free water to the requisite volume, and a measured volume of this diluted solution titrated with the iodine solution. This titration will indicate very nearly the proper amount of iodine solution required. During the process of titration as conducted above, there is a liability of loss of gas by its escaping into the air, and also by the oxidising action of the same, and these tend to make the results rather low. In order to avoid this source of error the following method of procedure may be adopted. Introduce into a flask almost but not quite as much iodine solution as was required in the last titration, and add to this the quantity of diluted sulphuretted hydrogen solution employed in that particular titration. This amount will destroy the colour of the iodine solution. Now add a little starch solution, and carefully run in more iodine solution until a permanent blue colour is just obtained. From the total volume of iodine solution used the amount N of sulphuretted hydrogen can be calculated. Each c.c. iodine solution = 0'0017 gram H 2 S. 113 PART IV. SPECIAL ANALYSES KEQUIRED IN GASWORKS. I. PROXIMATE ANALYSIS OF COAL. 66. THE proximate analysis of coal consists of the f Howing determinations : Moisture lost at some stated temperature, volatile matter, fixed carbon, ash, sulphur. The actual amounts of carbon, hydrogen, and nitrogen, are determined by means of elementary organic analysis. Estimation of Moisture. The actual temperature at which coal should be dried is r.ither a disputed point. It is known that coal does not lose ail its moisture at 100 C., consequently some analysts recommend drying at a higher temperature than that of billing water, but as 100 C. is a temperature easily obtained a ad kept constant, it is a convenient temperature to adopt as a standard. Powder about two grams of the coal very finely, weigh the same between two watch-glasses, and dry in the water-oven at a temperature of 100 C. Weigh every two hours. After drying for some time, it will be noticed that tlie coal increases in weight, consequently no constant weight will be obtained, the lowest weight observed should therefore be taken. The increase of weight, after drying for a period, has been attributed as being probably due to the slow oxidation of some matters in the coal (pyrites, &c.), or it may be caused by the absorption of gases into the pores of the coal which the expelled moisture left vacant. i 114 THE GAS ENGINEER'S LABORATORY HANDBOOK. Estimation of Volatile Matter and Coke. Take about two grains of finely powdered coal, and spread it in an even layer on the bottom of a thin platinum crucible. Weigh without the cover. Cover the crucible with its lid, and place it in an upright position on an iron ring, round which, sufficient platinum wire has been wound to prevent any con- tact between the iron and the platinum crucible. Then apply a powerful gas flame (to obtain comparable results, the length of the flame and the position of the crucible over it, should be the same in all experiments). Notice when the gases cease issuing from beneath the lid, continue the heating one minute longer, then take away the gas flame, place the crucible and cover for about five minutes in a desiccator to cool, and weigh it without the lid as soon as possible. The loss of weight observed represents the volatile matter which has been driven off by the heating. Estimation of the Ash. Weigh out accurately about two grams of finely powdered coal or coke in a platinum boat, made by bending a piece of platinum foil over a glass rod. The boat should be capable of being easily pushed into a piece of combustion-tube of Q FIG. 36. about one inch in diameter. The combustion tube and boat are then heated to redness in a gas-furnace ; at the same time a gentle current of air should be drawn through the tube by means of an aspirator. Care should be taken that the tem- perature does not become high enough to soften the glass, or the boat and tube may be fused together. The apparatus is shown in Fig. 36. SPECIAL ANALYSES. 115 As soon as the combustible matter is entirely consumed, allow the tube to cool, then withdraw the boat and weigh it. By this method of working, a beautifully clean ash can be obtained in a very short time. It is advisable to take a sc parate portion of coal for the determination of ash, and not tc weigh a portion of the coke obtained in the determination ol the volatile matter, as coke absorbs moisture very rapidly. T he ash should be preserved for the further determination of tl.e sulphuric acid which it contains. Estimation of Sulphur. In determining the sulphur in coal, it is necessary to d stinguish between the sulphur which goes over in the v )latile matter, the sulphur remaining in the coke, but which c; ,n be converted into S0 2 by combustion, and the sulphur \v hich is finally left in the ash combined as sulphate. For this purpose three determinations will be necessary. (1) Total Sulphur. This is best obtained by the method d3vised by Nakamura, which consists in heating the coal bilow a red heat in contact with alkaline carbonates, by which the coal, whether bituminous or not, rapidly undergoes without evolution of smoke, complete atmospheric oxidation, in a manner hardly to have been expected. The details of the method are as follows : Weigh out accurately about two grams of the very finely, powdered coal, and mix it with four times its weight of dry fusion mixture (sodium and potassium carbonates mixed in their molecular proportions). The mixing is effected by adding the fusion mixture gradually to the coal, contained in a platinum crucible, and stirring constantly with a dry glass rod. The crucible is then loosely covered with its lid, and the mixture at first gently heated, so as not to volatilise the hydrocarbons, that is, so that no smell, or only a faint aromatic odour is given off. An Argand spirit lamp should be used as the source of heat, in order to avoid the possible absorption of sulphur from the flame of coal gas, which might i 2 116 THE GAS ENGINEER'S LABORATORY HANDBOOK. occur if a Bunsen or other gas burner were used. Keep at a low temperature for some time, then slowly increase the heat without letting it reach that of visible redness, until the sur- face of the mass, which is at first of a dark grey colour, becomes only faintly grey. No smoke or odorous gases should escape during the whole of the oxidation. When the surface becomes only faintly grey, raise the temperature to a faint red heat, and keep it stationary for about forty to sixty minutes ; at the end of this time the mass will have become almost white, owing to the complete combustion of the coal, but if iron be present in the coal the tint will be reddish. Tlje crucible is now allowed to cool, and then placed in a beaker of distilled water, the beaker covered with a clock- glass, and gently heated for some time, when it will generally be found that the mass is wholly detached from the crucible, and the greater part dissolved. When that stage is reached, the crucible is carefully lifted out of the beaker with the crucible tongs, both tongs and crucible being well rinsed with distilled water from the wash-bottle into the beaker. Pure hydrochloric acid is now added to the contents of the beaker, in order to drive off the carbonic acid (due to the fusion mixture) ; the acid should be added very cautiously, and in small portions at a time, the clock-glass being immediately replaced after each addition, otherwise there is a risk of loss by spurting. When all the C0 2 appears to have been driven off, the solution should be raised to incipient ebullition, and maintained at that temperature for some time. It is then filtered, with the object of removing sand, &c., which need not be weighed. It is necessary to be careful not to lose any of the filtrate, and the filter-paper should be washed until it ceases to redden blue litmus-paper. The filtrate is then treated with a few drops of bromine, brought to the boiling point, an excess of solution of barium chloride added, and kept boiling for about ten minutes longer, the precipitated BaS0 4 is allowed to settle, filtered, washed free from acid and excess of BaCl 2 , dried, ignited, and weighed, in the same SPECIAL ANALYSES. 117 m? inner as in the typical analysis of barium. From the w( ight of BaS0 4 , the amount of sulphur in the coal operated on (233 parts BaS0 4 = 32 of S) is obtained, from which the percentage is easily calculated. To secure the complete oxidation of the coal by the above method, it is most im- pc rtant that it should be powdered as finely as possible ; the best plan is to sift it through cotton cloth. Stirring the m xture during the oxidation, impedes the process instead of h; stening it, as might have been expected, probably by cl >sing the passages left by the unconsumed coal. (2) Volatile Sulphur in the Coke. The sulphur, which is converted into sulphur dioxide on combustion of the coke in ai r, is determined as follows : A quantity of coke, equivalent to a known weight of coal, is roasted in a platinum boat, in a combustion tube in a current OJ air, as in the determination of ash, Fig. 36 p. 114. The g tseous products of combustion are aspirated through 10 c.c. 01 standard solution of iodine placed in a bulbed U-tube. The following reaction then occurs : S0 2 -+- I 2 +2H 2 = I3 2 S0 4 + 2HI. The amount of free iodine remaining in sdution is determined by standard solution of sodium thio- siilphate, and the amount acted on by the S0 2 is obtained by difference. From this the weight of S0 2 and of S may be calculated. (3) Sulphur in the Ash. This can be done in most cases by boiling the ash with a little hydrochloric acid, filtering, and washing the residue. The sulphate is determined in the filtrate and washings as in (1), or the ash could be fused with alkaline carbonates. 67. The following are methods for ascertaining the amount of carbon, hydrogen, nitrogen and oxygen in coal. Organic bodies (of which coal is one) are always composed of hydrogen and carbon, generally in combination with oxygen, and often with nitrogen, the other elements, such as chlorine, sulphur, &c., occurring much less frequently. When any organic substance containing carbon and 118 THE GAS ENGINEER'S LABORATORY HANDBOOK. hydrogen is burnt in air or in oxygen, these elements com- bine with the oxygen, and are converted into carbon dioxide and water respectively. Therefore, by making suitable arrangements, it is possible to collect and weigh the C0 2 and H 2 O, and from the weight obtained, to deduce the amount of carbon and hydrogen in the substance operated ou. The operation of burning is termed making a " combus- tion," and is generally effected in a current of air or oxygen ; but a common method is to heat the organic substance with some material which will readily part with its oxygen, such as oxide of copper or chromate of lead. FIG. 37. As combustions are always made in tubes (generally of glass) a furnace for heating the same is needed. There are a number of different forms of furnaces in use ; the one shown in Fig. 37 is known as " Hoffman's." In this form of furnace the combustion of coal gas, supplied by a series of vertical gas jets, is effected by means of a number of per- forated clay burners in which the gas is mixed with air. These burners are arranged so as to form a channel for the combustion tube, a system of stop-cocks serving to regulate the heat, or to confine it within any desired limit. The tube in which the combustion is made, known as the " com- bustion tube," should be of very hard glass, so as not to soften at a red heat, or, if it does so, not sufficiently to allow SPECIAL ANALYSES. 119 it to be blown out by the internal pressure of the gases. In cider to make a combustion tube, procure a piece of good combustion tubing, about ^ inch in diameter and 3 feet long, tl loroughly clean it inside, by passing a piece of cotton wool a -tached to a string through it several times; next heat the t*< ibe strongly in the centre over the blow-pipe flame, turning it c mstantly round while doing so ; then, when it is quite soft, ts 'parate the two halves in a direction at right angles to the 1 'ngth of the tube ; let it cool, and when cold cut it in two at t ae narrowest part, and seal up the points in the blow-pipe so ; s to make two tubes of the form shown in Fig. 38, and 1 nally round off the edges of the open end in the blow-pipe i ame. A soft cork which fits the tube tightly, and pierced FIG. 38. FIG. 39. with one hole will also be required. The water is collected n a calcium chloride tube, Fig. 39. This tube is filled with granulated calcium chloride in small fragments (but no dust), a loosely fitting plug of cotton- wool being placed over the CaCl 2 at D, in order to prevent any small particles of CaCl 2 from being carried over. The open end of the tube is closed with a tight-fitting cork pierced with one hole, through which a piece of glass tube bent at right angles passes. This cork should be made quite gas-tight, by coating- it with sealing-wax or melted paraffin. The apparatus should be provided with a loop of platinum wire, to enable it to be hung on the hook of the balance when weighing. When not in use, the ends of the tubes should be closed by small corks, made by cutting down the portions of cork left 120 THE GAS ENGINEER'S LABORATORY HANDBOOK. in the cork-borer when boring corks. The C0 2 is absorbed by the potash bulbs, shown in Pig. 40. The bulbs are filled, to the extent indicated by the dotted line in the figure, with strong potash solution, prepared by dissolving three parts of potash free from carbonate, in two parts of water. The bulbs are readily filled with this liquid contained in a porcelain dish, by dipping the end of the tube connected with the larger bulb beneath the surface of the potash solution, and gently aspi- rating at the other tube, until the re- quired amount has been drawn in. The tube is then carefully dried, inside and out, with filter-paper, and the ends of FIG. 40. the tube closed with small corks. A piece of platinum wire is twisted round the tubes where they touch, in the manner shown in the figure ; this serves to suspend the bulbs from the hook of the balance-pan. In order to make a combustion of a sample of coal, the combustion tube, Fig. 38, is made perfectly clean and quite dry. It is then filled for about 14 inches of its length with pure chromate of lead, which has previously been made hot. A sample of the coal is reduced to fine powder, and about five grains are weighed, and carefully dried, at a temperature of 240 F. About a fourth of the lead chromate in the com- bustion tube is then placed in a hot porcelain mortar, the coal is added, and the two are intimately mixed. The re- mainder of the chromate in the tube, with the exception of a couple of inches is then added, the whole well mixed, and quickly transferred back without loss to the combustion tube. The mortar is then rinsed with a little fresh hot chromate, which is also placed in the tube, which is then filled up to within 3 or 4 inches of its mouth with pure hot chromate. Some pure copper turnings are then pressed into the remaining space, so as to effect the decomposition of any SPECIAL ANALYSES. 121 oxides of nitrogen which may be formed during the com- bustion. The combustion tube is then rapped horizontally on the bench, so as to shake down the mass, in order to allow a froe passage above by which the evolved gases can escape ; it may then be placed in the furnace, the open end of the combustion tube projecting about an inch from the furnace. T le calcium chloride tube, Fig. 39, previously weighed, is attached to the combustion tube by means of its cork, ai.d to the end of this are connected, by a short piece of ii dia-rubber tubing, the previously weighed potash bulbs, F ig. 40. After making certain that all the connections are quite tight, the combustion is started, by lighting the gas burner n Barest to the open end of the tube. As soon as the front o ]' the tube attains a dull red heat, the next gas tap is opened, a id the same operation gradually performed along the whole 1( ngth of the tube, until the whole length attains a dull red hoat. (It is necessary to effect the heating gradually, in o -der that the gases evolved may pass away slowly, as, if they were driven off too rapidly, a portion might escape absorption.) As soon as bubbles cease to pass through the potash bulbs, the point of the combustion tube is broken off by n eans of a pair of pincers, and a current of dry atmospheric a r, free from C0 2 , is slowly drawn through the apparatus, by means of a suction tube connected to the end of the potash bulbs. In this way, the whole of the C0 2 and water vapour remaining in the apparatus is swept through, into the receptacles intended for their detention. The chloride of calcium tube and potash bulbs may now be disconnected, allowed to cool, and when cold re-weighed. The gain in weight of the CaCl 2 tube is due to water, and of the potash bulbs to C0 2 . As an example, supposing that we had weighed out 5' 5 grains of coal, and we found that the CaCl 2 tube had gained 3 '2 grains in weight, and the potash bulbs 15' 7 grains, 122 THE GAS ENGINEER'S LABORATORY HANDBOOK. then, in order to find out how much hydrogen the 3*2 grains of H 2 0, are equivalent to, we should say Mol. weight Mol. weight of water of hydrogen As 18 : 2 :: 3-2 : x. x - -355; and to obtain the percentage As 5-5 : 100 :: -355 : x. x = 6-46. To obtain the amount of carbon the 15-7 grains of C0 2 are equal to, we should say Mol. weight Atomic weight of CO 2 of carbon As 44 : 12 :: 15-7 : x. x = 4-277, and to obtain the percentage As 5-5 : 100 :: 4-277 : x. x =77-77. In order to determine the amount of nitrogen, the latter element is converted into ammonia by a combustion with soda lime, in a similar combustion tube to the one employed in the determination of hydrogen and carbon, but it need not be so long. A convenient length is about 12 inches. About 9 inches of the length is filled with dry soda lime. A portion of this (about a third) is then shaken out into a porcelain mortar, and intimately mixed with about ten grains of finely powdered coal previously dried. More soda lime is then shaken out (leaving an inch in the tube), mixed, and the whole replaced in the tube. The mortar is then rinsed with a little soda lime, which is added to the contents of the tube, which is now nearly filled with fresh soda lime. A plug of asbestos is placed in the mouth of the tube, the contents of which are then shaken down so as to leave a space for the passage of the gas ; the tube is then placed in the combustion furnace. The bulbs, Fig. 41, are now connected with the com- SPECIAL ANALYSES. 123 bustiontube by means of a perforated cork. The bulbs are filled with dilute sulphuric acid, made by adding 10 c.c. of normal sulphuric acid and diluting with distilled water. The combustion is then conducted as in the determination o ' carbon and hydrogen, but it maybe effected more rapidly. Y'hen the evolution of gas has ceased, the bulbs are discon- nected, and the contents separated from tarry matters by fi Itration. The solution is then coloured with a few drops of c )chineal solution, and titrated with a standard solution of c lustic soda, of such a strength that 3 c.c. of the soda solu- t on neutralise 1 c.c. of the acid solution. From the number FIG. 41. c f cubic centimeters of the soda solution required, the amount of ammonia may be determined as in (59). The amount of oxygen in an organic compound is always estimated by difference, that is, having ascertained the weight of all the other elements entering into the com- position of the body under examination, the amount required to make up the 100 parts is put down as oxygen. The above description is only intended to give a general idea of the method of performing an organic analysis, it is almost impossible to describe the various precautions neces- sary to be observed in order to make a reliable determina- tion by this method : the only way of obtaining the requisite knowledge is to see an analysis actually performed by a skilful operator. UNIVERSITY ] 124 THE GAS ENGINEER'S LABORATORY HANDBOOK. Specific Gravity. 68. It is occasionally requisite to know the weight of a cubic foot of the coal, or the number of cubic feet corre- sponding to a ton. This is easily arrived at when the specific gravity of the coal is known. The specific gravity of a body expresses the relation which exists between the weights of equal bulks of one substance compared with another, chosen as the standard. For solids and liquids, the standard is distilled water at the temperature of 60 F. In order to take the specific gravity of a sample of coal, a small piece weighing about an ounce is selected, and suspended by means of a horse-hair, or piece of fine silk, from the specific gravity pan of the balance. It is then accurately weighed, while hanging in the air. It is then immersed in a beaker of distilled water, as shown in Fig. 42, and again weighed. Before the weight in water is recorded, any adhering air bubbles must be removed from the surface of the coal. This is effected by removing the sample from the water after it has been im- mersed for a short time, and brush- ing the surface with a moistened camel's-hair brush. It can then be put back again into the water, and the weighing completed. It is also desirable that the coal be sufficiently soaked ; this is effected by immersing the sample, after attaching the horse-hair to it, in water in the flask of the filter pump, and exhausting the air within the apparatus as far as practicable. The difference between the weight of the coal when weighed in air, and when weighed in water, is the weight FIG. 42. SPECIAL ANALYSES. 125 of a bulk of water equal to the bulk of the coal immersed, hence the weight of the coal in relation to that of water ; in ot ler words its specific gravity, is found by dividing the weight in air by the loss of weight in water, as under. Grains. Weight of coal in air 526'7 water 140-6 Loss of weight in water, or difference representing ^ the weight of an equal bulk of water / 386 1 hen = 1 346, the specific gravity of the coal. The specific gravity being ascertained, the remaining d ita are easily calculated from the same. A cubic foot of v ater at 60 F., weighs exactly 1000 ounces, hence a cubic f >ot of coal having the specific gravity given above, will veigh 1364 ozs.,or 8^*25 Ibs. avoirdupois. Having obtained t lis information, the measurement or space occupied by a t m of the same coal is arrived at as under 2240 86^25 = 26 ' 28 ' t he measurement of a ton of the coal in cubic feet. II. ESTIMATION OF THE IMPURITIES IN CRUDE GAS. Sulphuretted Hydrogen and Carbonic Acid. 69. These impurities may be determined with accuracy by the method devised by Mr. L. T. Wright, F.C.S., and described by him in a paper read before the Nottingham Section of the Society of Chemical Industry in the year 1885. This method, which is founded upon the increase in weight of absorption tubes, admits of the use of a tolerably large quantity of gas, collected regularly during an interval 126 THE GAS ENGINEER'S LABORATOEY HANDBOOK. of time sufficiently long to give an idea of the average composition of the gas supply to be tested. The reagent employed for the absorption of SH 2 in absorption tubes, may be described as an impure di- tri- ortho-phosphate, and may be prepared in the following manner. Solutions of 2 Ibs. of hydrogen disodium phosphate in 1 gallon of water, and 2J Ibs. of cupric sulphate crystals in 1 J gallons of water are mixed with vigorous stirring, and the resulting bright blue precipitate washed by decantation, and then dried in a water bath to about 100 C. The material sometimes is light and powdery, sometimes hard, and then requires crushing in a mortar. Its action with SH 2 is very sharp. The U-tubes used in these experiments are preferably of the modern form, with hollow round glass stoppers, which also serve as stop-cocks; the inlet and outlet con- nections being small pieces of glass tubing, fused into the sides of the U-tubes. A small hole in the stopper, corre- sponding to the hole in the side of the tube where the glass tubing is fixed on, makes the connection. A slight turn of the stopper opens or closes the tube. The tubes are weighed when charged with clean coal gas, which, on the same gas supply, varies but little in specific gravity. The stoppers are lubricated with vaseline or resin cerate, and with this exception are perfectly clean. A small plug of cotton wool in each stopper, prevents the mechanical carrying forward of fine particles of the reagents by issuing Closed tubes charged with coal gas do not suffer any appreciable variations in weight during a period of two days. Mr. Wright found that such tubes, charged with cupric phosphate, gain in weight under the action of clean coal gas, but the increase soon reaches a limit, and the phosphate becomes "saturated." In order to attain this saturation point, it is necessary to pass 3 cubic feet of clean dry coal gas slowly through the tubes. In order to charge the phosphate tubes, they are first SPECIAL ANALYSES. 127 cleaned and dried, great attention being paid to the ground portions of the tubes and stoppers. A plug of cotton wool is placed in the bend of the tube, then one leg of the same is charged with the cupric phosphate in a suitable state of su bdivision, care being taken that the rough and fine portions of the powder are regularly distributed, for if all the rough pc rtions are on one side and the fine on the other, a passage oi the gas down the side where the coarse portions are takes plice, and consequently the fine portions never come into a< tion. The other leg is charged with powdered calcium cl loride. The ground portions of the tubes are wiped free from dnst, and the stoppers, fitted with plugs of dry cotton wool ii the hollow places, and lubricated with a little resin cerate, &. e put in position. A current of dry coal gas to the extent of 3 cubic feet is p issed through, and the tubes are then ready for weighing. A 6-inch U-tube charged in this manner will be capable of a )sorbing 20 grains of SH 2 ; and if, in the case of gas containing 10 grains of SH 2 per cubic foot, the experiments are made upon quantities of 0'5 cubic foot, each will serve for four analyses without being recharged. The calcium chloride will last much longer. The C0 2 is absorbed by soda lime, in a similar way, one half of a U-tube being filled with soda lime, and the other half with calcium chloride. Mr. Wright found it necessary to use the soda lime tubes in a moist condition, for when quite dry, soda lime has a much feebler absorptive power for C0 2 . The requisite degree of moisture may be attained by exposing the soda lime to a moist atmosphere from twelve to eighteen hours. There are no inconveniences attending it in this condition. These soda lime tubes remain very constant in weight, when clean, pure, coal gas is passed through them, and therefore do not require saturating in the same manner as the phosphate tubes. A small quantity of dry, pure, coal gas should, however, be passed through, to expel air previous to their being weighed . 128 THE GAS ENGINEER'S LABORATORY HANDBOOK. A 6-incli tube, charged with soda lime in the manner above described, serves for three analyses of 0*5 cubic foot each, of gas containing from 10 to 12 grains of C0 2 per cubic foot. In cases where there is also ammonia in the gas to be tested, as in the inlet to the scrubbers, it will be requisite to effect its removal before the gas reaches the weighing tubes. This might be done by means of the drying cylinder con- taining calcic chloride, which effectually removes NH 3 from coal gas ; but in order to maintain the neutrality of the calcium chloride, it is better to remove the NH 3 from the gas before drying, by passing the gas through a 12-inch U-tube filled with pieces of broken pumice saturated with syrupy phosphoric acid. It should be noted that in drawing off samples of coal gas for analysis, certain precautions are necessary, since vulcanised and iron tubing, when new, remove SH 2 from crude coal gas, so that it is advisable to have the services conducting the gas from the source to be tested to the absorption tubes as short as possible, and previously saturated with crude coal gas. It is further of importance to keep a tolerably quick current of gas through the iron pipe, in order to prevent the gas from becoming " stale." The iron service should be fitted with a four-way piece and three cocks. One of these serves to take the gas to the train of apparatus, another to be left open during the experiment, blowing gas away, and thus main- taining a rapid current in the iron pipe, while the third can be ubed for other purposes, such as the examination of gas for ammonia, &c. Eefore making an analysis it is necessary to blow away a little gas, in order to clear stale gas out of the iron service pipe and bring along a fresh supply. The complete apparatus can now be described. The first piece in the train is the syrupy phosphoric acid tube, which can, however, be dispensed with in cases where the gas has been previously thoroughly washed free from ammonia. This tube is directly connected with the cock on SPECIAL ANALYSES. 129 the iron pipe by means of a small piece of " saturated " vulcanised tubing. r . 1 he outlet of this tube is in communication with a large dryi ng cylinder, filled with small pieces of calcium chloride quit 3 free from alkalinity. The outlet of the drying cyli ider is provided with a T-piece, which carries on one arm a small piece of vulcanised tubing and a screw pinch- cocl , for the purpose of blowing away a little gas before and during each experiment. The other arm of the T-piece is c mnected to the inlet of the cupric phosphate tube, the out et of the latter being connected to the inlet of the soda Km , tube. A gas meter connected to the outlet of the soda lim i tube completes the apparatus. The meter most suit .ible for the purpose, is the test -meter employed in the Ga^ Eeferees' Sulphur Test. It is advisable to bring the poi: iter attached to the meter up to within a division or so of the zero mark by means of the key provided for that purpose, and then to blow gas through until the pointer exactly covers the mark : by operating in this manner, any error due to slackness in the gearing of the index is avoided. Before staiting an experiment, gas should be blown away at the outlet of the drying cylinder, as well as at the cock on the iron service. When the tubes have been weighed full of clean coal gas, and all the pieces of the train joined tightly together, the stop-cocks are turned on, and the gas is allowed to traverse through the apparatus at a moderate rate, which is often more dependent on the pressure of the gas, than on the will of the operator. From 0'25 to 0*50 cubic feet per hour is a convenient rate, and one well within the limit at which the complete action of the reagents is obtained. The quantity of gas used for each experiment is a matter of choice, and will be determined by the amount of impurity in the gas to be examined. When the required quantity of gas has passed through the tubes, the inlet stopper of the phosphate tube, and the outlet stopper of the soda lime tube are turned off, and the two tubes are taken off together, and 130 THE GAS ENGINEER'S LABORATORY HANDBOOK. connected in their proper order to a supply of desiccated clean gas, in order to drive the gas in the phosphate tube through the soda lime tube. The stoppers can then be turned off, and the tubes weighed. Of course, before weighing the tubes, they must be wiped perfectly clean and dry, but as this operation causes them to be a little light, they should be left for about five or ten minutes in the balance case before being weighed. The results in grains, corrected for temperature and pressure, are best calculated to a cubic foot of clean coal gas at 60 F. and 30*0" Bar. Since washed, but otherwise unpurified, coal gas contains about 6 grain of cyanogen per cubic foot, the results obtained by this method are slightly high. The absorption of cyanogen in the first tube is not complete about one-half going forward into the soda lime tube. The error falls equally on each tube. Cyanogen may be esti- mated by passing a measured quantity of gas free from ammonia, but otherwise unpurified, through a U-tube filled with soda lime, and then making a combustion of the residue, as in an ordinary nitrogen determination. The tubes may be conveniently supported by suspending them from hooks on a light wooden framework or " gallows " by means of fine copper wire. Having obtained the weights of SH 2 and C0 2 contained in a definite volume of gas, it is an easy matter to convert the same into volumes and from these data to arrive at the percentage by volume, in which form the results are gener- ally expressed. The following examples will show the method of performing the necessary calculations: (1) For SH 2 . Volume of gas passed, corrected to 60 F. and 30-0" Bar. = 0-426 cubic feet. Grains. Weight of copper-phosphate tube after experiment 1593' 7 before 1590-6 SH 2 absorbed 3'1 A cubic foot of SH 2 (dry) at 60 F. and 30-0" Bar. may SPECIAL ANALYSES. 131 be taken as weighing 631 '5 grains, consequently 1 grain will be equal to 0-0015835 of a cubic foot. Now in the example, as there are 3-1 grains of SH 2 in 0*426 cubic foot of ijas, in 100 cubic feet there will be 727 7 grains ; therefore O-i 015835 X 727-7 = 1-1523 cubic feet of SH 2 in 100 cubic fee : of gas, or 1 1523 per cent, by volume. (2) For C0 2 . Volume of gas passed corrected to 60 F. and 30-0" Bar. =0-426 cubic feet. Grains. Weight of soda lime tube after experiment .. .. 1763 '4 before .... 1757 '2 C0 2 absorbed 6*2 A cubic foot of C0 2 (dry) may be taken as weighing 817 3 grains, consequently 1 grain will be equal to 0012235 of a cubic foot, and as there are 6-2 grains of C0 2 in 4:: 6 cubic feet of gas experimented on, in 100 cubic feet there will be 1455-4 grains; therefore 0-0012235 x 1455 '4 = 1-780 cubic feet of C0 2 in 100 cubic feet of gas, or 1 r , 80 per cent, by volume. It is necessary to notice that, as the gas is measured aftor the H 2 S and C0 2 are taken out, that the original volume will be greater than that actually recorded by the me :er. In the above example it would be : Gas measured 100 'cubic feet H 2 S 1-152 C0 2 1-780 102-932 Consequently to get the true percentage it will be neces- sary to make the following calculations : ForH 2 S. As 102-932 : 100 : : 1-152 : 1-120 = 1-120 per cent. ForC0 2 . As 102-932 : 100 : : 1-780 : 1-729 = 1-729 per cent K 2 132 THE GAS ENGINEER'S LABORATORY HANDBOOK. 70. The carbonic acid in crude gas may also be~ estimated by a volumetric method devised by Mr. J. T. Sheard, F.C.S., and communicated by him to the Journal of Gas Lighting. The apparatus employed is shown in Fig. 43, and is a form of Volhard's absorption apparatus, as modified by Mr. Sheard ; it may be said to consist of three parts, viz., the reservoir or body of the apparatus, the bulbs, and the scrubber, which is the long tube filled with pieces of broken glass or glass beads. The apparatus is made of blown glass, and has a total capacity of from 250 to 300 c.c. Not more, however, than about 40 c.c. of liquid, or the contents of rather more than one of the bulbs, can be employed in it for one estimation. When in use, the pressure of the gas forces the liquid out of the reservoir into the bulbs, and there bubbles through it. The method of using the absorption tubes is to run into the reservoir, a known quantity of the absorbing liquid of known value, and then to cause a measured quantity of gas to bubble through it ; the scrubber is afterwards rinsed with distilled water to wash down any portion of the absorbent that may have been mechanically carried into it by the gas, and the whole of the resultant liquid is titrated in the reser- voir. The method adopted for the estimation of C0 2 , consists in absorbing it in a solution of barium hydrate of known strength, the resultant liquid being afterwards titrated, in order to ascertain the amount of free hydrate remaining in solution, from which the amount of C0 2 absorbed is deduced. The indicator employed is phenol-phthalein, prepared by dis- solving 0*5 gram of the substance in 1000 c.c. of 50 per cent, alcohol. In the presence of the hydrates of the alkalies and alkaline earths, this indicator gives a splendid purple-red colour, which is destroyed on neutralising the alkali by an acid, the end reaction being exceedingly sharp and distinct ; carbonate of barium precipitated in a solution of the hydrate does not develop a red colour with phenol-phthalein, but remains neutral. It follows that the value of the solution of SPECIAL ANALYSES. 133 bar um hydrate being known, a simple titration of the liquid after absorption of C0 2 , will show the diminution in the ami >unt of free hydrate present, and inferentially the amount of ( )0 2 absorbed. The solution of barium hydrate is made by lissolving the ordinary crystallised hydrate of baryta in wa er until the solution is saturated, and then syphoning off the clear liquid into a well-stoppered bottle. The strength of ;he clear liquid is then determined very accurately by tit: ation with decinormal hydrochloric acid, and should be approximately equal in strength to that acid; i.e. if (say) mr ah more than 10 c.c. of the acid are required to neutralise th( same quantity of the baryta solution, a little water shi uld be added to the latter to reduce its strength. It is ne< essary that the strength of the solution should be checked by titration from time to time, as, unless very carefully pre- sei ved from contact with the air, it gradually loses power th ough the absorption of C0 2 . In addition to the gas ab- soi ption tubes, an aspirator, two 500 c.c. flasks, two 50 c.c. bu -ettes (graduated into 1-1 Oth c.c. and provided with glass stc p-cocks), a supply of distilled water, and a decinormal sol ation of hydrochloric acid are required. Each cubic centi- me tre of this solution is equal to 0022 gram of C0 2 . In order to make a test for C0 2 (the gas being previously freed from H 2 S by passing it through a small oxide purifier), a measured quantity varying from 15 to 30 c.c. according to tho amount of C0 2 expected to be met with in the gas of the barium hydrate solution is run from a burette or pipette into the reservoir of one of the absorption tubes, and from 10 to 20 c.c., also accurately measured, into another tube. Should tho smaller quantities mentioned be employed, an equal amount of distilled water is added, as the best results are ob- tained when the lower bulb of the first absorption tube is quite filled with liquid as the current of gas bubbles through it. About 30 c.c. is therefore the proper amount of liquid ; if much more than this is employed, it is liable to be carried forward into the scrubber. The absorption tubes are 134 THE GAS ENGINEER'S LABORATORY HANDBOOK. placed in the stand, Fig. 43, and the apparatus connected np by means of india-rubber tubing and stoppers pierced by glass tubes, the inlet of the first tube being connected to the gas supply, and its outlet to the inlet of the second tube ; the outlet of the latter being connected in turn, to the inlet of the aspirator. In connecting the gas-supply tube to the inlet of the first absorption tube, gas must first be allowed to blow away, and the india-rubber tube pushed over the FIG. 43. bent glass tube while the gas is flowing, care being taken that the pressure of the gas supply is not so great as to force its way through the liquid in the bulbs without the aspirator being in operation. Assuming everything to be in readiness, the measuring flask is placed under the tap of the aspirator, and water allowed to flow, so as to draw the gas in bubbles through the liquid in the bulbs. It is necessary at this stage to see that the connections are quite gas-tight, which may be done by turning off the gas supply, and the tap of the SPECIAL ANALYSES. 135 aspirator, and observing whether the level of the liquid in the bulbs remains stationary or not. If the connections are tig! t, the current of gas is continued until 500 c.c. have passed, as ^ hown by the water that is run out of the aspirator. (Mr. Shcird found that the best speed of current was obtained wh ;n the flow of water from the aspirator is adjusted to the poi it at which it breaks from a succession of drops to a con- tin lous stream.) The gas supply is then disconnected from the inlet to the first tube which is done without stopping the flow of water, by slipping off the india-rubber gas-tube as the water in the flask passes the mark a fresh flask is brought un ,er the aspirator, and a further quantity of 500 c.c. of air drr, wn through the apparatus, in order to drive forward the ga^ remaining in the reservoirs of the two absorption tubes at he end of the first aspiration. The tubes are then dis- coi nected, and a little water (free from C0 2 ) run through the scrubbers, in order to wash down any of the barium hy> Irate solution that may have been carried forward me- chi nically by the gas. A few drops of phenol-phthalein sol ition are now added, and the liquid is titrated with deci- noi mal hydrochloric acid, the point of saturation being shown by the disappearance of the purple colour imparted to the liq aid by the phenol-phthalein. In titrating the carbonated liquid, the acid should be adeled a little at a time, and allowed to run down the sides of the reservoir, not dropped directly into the liquid, the liquid being gently agitated after each addition of acid. The dif- ferance between the quantity of free barium hydrate found at the close of the experiment, and that run into the tubes at the commencement of the same, shows the amount of C0 2 absorbed ; and from this, the amount of the latter impurity existing in the gas may be calculated into grains per cubic foot, or percentage by volume. Example: Two gas absorption tubes were charged in the manner described, the first with 25 c.c. and the second with 18 c.c. of barium hydrate solution, 500 c.c. of gas 136 THE GAS ENGINEEK'S LABOKATORY HANDBOOK. followed by an equal quantity of air, being afterwards drawn through them. At the end of the experiment, the scrubbers were rinsed with water, phenol-phthalein solution added, and the liquid in the tubes titrated with decinormal hydrochloric acid, 16 '60 c.c. being required to neutralise that in the first tube, and 19-00 c.c. for the second. The titre of the barium hydrate N solution used being 1 c.c. = 1-09 c.c. acid, therefore, First Tube. Second Tube. Equivalent of barium hydrate employed . . 27 25 c.c. 19 62 c.c. acid required to neutralise the resultant ) 10 } ib'bu ,, iy uu ,, liquid J Equivalent of BaO,H 2 neutralised .. .. 1065,, 0'62 From these data, the percentage by volume, or grains, of C0 2 per 100 cubic feet of gas may be easily obtained, but in order to make it quite clear to those who are not used to such calculations, the process is described in detail. (1) To obtain the percentage by volume. The total equivalent of BaO, H 2 0, neutralised as above, is 10-65 + 0*62 = 11-27 c.c., and as each c.c. is equal to 0-0022 gram C0 2 , 11-27 X 0-0022 = 0-024794 gram C0 2 . Now the weight of 500 c.c. of C0 2 saturated with moisture = 0'914 gram, consequently As 0-914 : 0-024794 :: 500 : 13-56, that is, in the above experiment there were 13*56 c.c. of C0 2 in 500 c.c. of gas. Therefore, in order to obtain the percentage we say As500 : 100 :: 13-56 : 2-71 = 2-71 percent, by volume of C0 2 . The above might be more shortly written thus : 11-27 C.C.X 0-0022 gram X 100 = 2 . ?1 cent . b 0-914 gram SPECIAL ANALYSES. 137 and more shortly still, by employing the factor 241 (ob- 0022 grm. X 100 taujod thus: TTTTiJ -=0-241) 0*914 gram thei. 11-27 x 0-241 = 2-71 per cent, by volume of C0 2 as beft re. 2) To obtain the weight in grains of C0 2 per cubic foot of < as. \.s before, 11-27 XO '0022 = 0-024794 grain C0 2 , and as i lere are 15-432 grains in a gram, 15-432 X 0-024794 = O-o ^262 grains C0 2 in 500 c.c. of gas. Now 1 cubic foot = 28,: 15 c.c., consequently As 500 : 28315 : : 0-38262 : 21-66 = 21-66 grains of C0 2 per cubic foot of gas. This might be expressed more shortly thus : 1L-27 c.c. X 0-0022 gram x 15-432 grains X 28,315 c.c. 500 c.o. = 11-66 grains C0 2 per cubic foot of gas, and from the 0-0022 x 15-432 X 28,315 ,, , ,000- figures 2: '- the factor 1-922 is 500 obtained, which shortens the calculation materially thus: 1-92 X 11-27 = 21-66 grains C0 2 per cubic foot of gas as before. The amount of sulphuretted hydrogen in crude gas may also be estimated by means of an acid solution of cadmium chloride. 71 The gas is slowly passed through a solution of the above contained in a couple of Woulffe's bottles, and is then measured by means of an aspirator or meter. After passing a quantity of gas at the rate of about half a cubic foot per hour, the bottles are disconnected, and the precipitated cadmium sulphide is washed on to a filter ; the contents of the filter are then placed in a beaker with distilled water, and bromine added in excess so as to oxidise the CdS to CdS0 4 . The 138 THE GAS ENGINEER'S LABORATORY HANDBOOK. "bromine is driven off by heat, and the resulting solution filtered and washed, the filtrate (CdS0 4 ) is then treated with an excess of EaCl 2 which precipitates BaSO^ This is washed, dried, ignited and weighed in the usual manner. 233 parts BaS0 4 correspond to 32 of sulphur, and 32 parts of sulphur are equal to 34 ot sulphuretted hydrogen, from which the quantity per cubic foot can readily be cal- culated. The method of determining the amount of ammonia in crude gas is described in (92). III. TESTING PURIFIED GAS FOR SULPHURETTED HYDROGEN, AMMONIA, AND SULPHUR COMPOUNDS ACCORDING TO THE INSTRUCTIONS OF THE GAS REFEREES, &c.* 72. The testings for the above impurities extend over twenty hours of each day, and are made on 10 cubic feet of gas, which is tested successively for each of the following : (1) Sulphuretted Hydrogen. In testing for this impurity, the gas, as it leaves the service-pipe, passes through an apparatus in which are suspended slips of bibulous paper impregnated with basic acetate of lead. The test paper from which these slips are cut, is prepared by moistening sheets of bibulous (filter or blotting) paper with a solution of one part of sugar of lead in eight or nine parts of water, and holding each sheet, while still damp, over the surface of a strong solution of ammonia for a few moments. As the paper dries, all free ammonia disappears. It is hardly necessary to state, that if there be any SH 2 in the gas, the slips of lead paper will be discoloured. (2) Ammonia. The gas, on leaving the above mentioned apparatus, passes through a glass cylinder filled with glass beads, which have previously been moistened with a measured quantity of standard sulphuric acid ; any ammonia in the gas * The full instructions of the Gas Referees as to the method of testing for the above impurities will be found in the Appendix. SPECIAL ANALYSES. 139 is absorbed by the acid, and from the amount of acid neu- tralised, the amount of ammonia in the gas is determined. (; ) Sulphur Compounds. The gas which has passed through the sulphuretted hydrogen and ammonia testing apparatus, pass* s next through a meter, by means of which the rate of flow can be adjusted to half a cubic foot per hour, the said meter being provided with an automatic arrangement for shut ing off the gas when 10 cubic feet have passed through the f ame. ] 'rom the meter, the gas passes to the sulphur test, where it is burned in air impregnated with the vapour of ammonia, the esulting water vapour is condensed and collected, and cont lined therein will be found the sulphuric acid and other conci ansable products resulting from the combustion. 73. The operation of testing the gas for ammonia and sulpiur, is effected in two stages. The first or preliminary stag >, consists in arresting the ammonia by means of acid, and bun ing the gas as described; the second stage consists in find; ng out how much acid has entered into combination with the ammonia in the gas, and the weight of barium sulphate obtained from the condensed products of combustion on adding a solution of barium chloride. 74. In estimating the quantity of ammonia in purified gas, two solutions are required, one containing one grain of ammonia in 100 measures of the solution, and the other the equivalent amount of sulphuric acid in 25 measures of the solution, which will exactly combine with one grain of ammonia, that is to say, 25 measures of the acid solution should exactly neutralise 100 of the alkaline solution. In order to prepare these solutions we shall require various glass vessels graduated on the septem system. The following is a list of the same : A decigallon mixer divided into 100 parts ; pipettes to deliver septems ; two burettes on stands, one for use with the acid and the other with the alkaline solution, each burette being graduated into 100 parts (septems), each part being further subdivided into fifths. (A septem is equal 140 THE GAS ENGINEER'S LABORATORY HANDBOOK. to the one ten-thousandth part of a gallon, or to 7 grains weight of distilled water at 62 F., consequently a gallon contains 10,000 septems, and a decigallon 1000 septems). 75. The first operation is to prepare the sulphuric acid solution of such a strength, that 25 measures (septems) shall be equal to 1 grain of ammonia. Sulphuric acid combines with ammonia, in the proportion of one molecule of the acid to two of ammonia, as in the following equation ; and, consequently, in the proportion by weight of 98 to 34, therefore, as the Eeferees' calculations are worked out in grains, 2-8824 grains of acid are required to neutralise 1 grain of ammonia. Now the specific gravity of the strongest sulphuric acid obtainable is about 1-845, consequently a septem of such acid would weigh 12-915 grains. The decigallon mixer contains 1000 septems, and, as the equivalent of one grain of ammonia requires to be contained in 25 septems of solution, the total contents of the mixer should contain forty times the equivalent of 1 grain of ammonia, or 115*2960 grains of acid, and as one septem of acid weighs 12-915 grains, theoretically we should require 8*93 septems of acid in the decigallon. Sulphuric acid of the specific gravity 1 845 is not generally obtainable however, owing to its absorbing moisture so readily ; it is better, there- fore, to add an excess of acid, as it is easier to reduce, than to increase the strength of the solution. If we add 9 4 septems of acid it is probable that we shall obtain a solution of about the right strength. Therefore, having filled the mixer with distilled water until the centre of the meniscus of the surface of the water, coincides with the top, or hundredth line of the measure, 9 4 septems of acid are added from one of the two burettes previously mentioned, which should always be reserved for the use of the acid solution. The best way of dealing with the strong sulphuric acid, is to fill the burette with it by means of a small lipped beaker, and having poured SPECIAL ANALYSES. 141 out 1 lie 9*4 septems into the mixer, empty the burette, and immediately wash it out with water until it ceases to redden blue litmus-paper, and afterwards turn it mouth downwards to d am. After adding the acid, the stopper of the mixer should be inserted, and the vessel well shaken for several min ites. It should then be allowed to stand in a cool place for jme time, so as to attain the normal temperature, and allow tim< for contraction. When the vessel and its contents are quil a cool, 50 septems are measured out into a beaker by mea .is of a pipette, diluted with distilled water, neutralised wit L ammonia, made acid with hydrochloric acid, raised to the boiling point, the sulphuric acid precipitated by barium chli ride, and the precipitate filtered, washed, dried, ignited and weighed as in (24). The net weight of barium sulphate produced should be 13 '8 grains, which is arrived at in the foil iwing manner : Ninety-eight parts by weight of sulphuric acid, give 233 par s of barium sulphate as under : H 2 S0 4 + BaCl 2 = 2HC1 + BaS0 4 ; 98 208 73 233 and as 25 septems of the acid solution contain 2 8824 grains of sulphuric acid, 50 septems will contain 5*7648 grains, consequently, if we make the following proportion, we obtain the theoretical quantity of barium sulphate which the 50 septems should give : As 98 : 5-7648 : : 233 : 13-76, which may be expressed in round numbers as 13*8. Pro- bably the weight obtained will be somewhat greater than this, due to the excess of strong acid taken ; and, according to the Referees' Instructions, if the weight exceeds 13-9 grains, or falls below 13*7 grains, more water or sulphuric acid must be added, until the weight falls within these limits. 76. In order to determine the quantity of water which must be added to the solution in order to reduce the strength, if too strong, divide the number of septems of test-acid 142 THE GAS ENGINEER'S LABORATORY HANDBOOK. contained in the mixer by 13 8, and multiply the quotient, by the difference between 13 "8 grains, and the number of grains of sulphate obtained. The following is an example, the excess obtained being purposely exaggerated. On adding 9*4 septems of strong acid to a decigallon of water, the volume of the mixture will be 1006*9 septems (the decreased volume being due to the fact that when sulphuric acid combines with water contraction ensues). On taking out the 50 septems for the determination of the amount of BaS0 4 , there will remain 1006-9 - 50 = 956 '9 septems, which, being divided by 13*8, gives a quotient of 69*34. Now assuming the excess of sulphate to be 2 grains, then 69 '34 X 2 = 138 '68, as the number of septems of water necessary to be added. This may be more clearly explained as under. 50 septems of dilute acid, if of correct strength, should give 13 '8 grains BaS0 4 ; but if the BaS0 4 weighs 15-8 grains as in the example, we have as much acid present in the 50 septems as should be contained in 57*25 septems. 15-8 x 50 _ K-.O- 13*8 and consequently in 100 septems, we have as much acid as should be contained in 114*5 septems, therefore in the whole contents of the mixer, we have as much acid present, as should be contained in 1095* 6 septems. 956 ' 9 >< 114 ' 5 = 1095-6, 100 consequently, if we subtract 956 * 9 from 1095 6 we get 138 7 as the number of septems of water necessary to be added, which confirms the result in the preceding example. After adding the necessary quantity of water, and well shaking up the contents of the mixer, on operating on another 50 septems of dilute acid, we ought to get very near to the 13*8 grains required, but in case the solution is still too strong, the additional quantity of water necessary to be SPECIAL ANALYSES, 143 adchd may be found by the before mentioned rule, due notice being taken of the volume of the solution remaining in he mixer after the abstraction of the portion taken out, and the water added, thus: the 956-9 septems -f- 138-7 septems of water added, will be equal to about 109 5- 6 septems, and, after abstracting 50 septems for a sect nd determination, there will remain 1045*6 septems to be livided by 13 '8, in order to obtain a quotient to be mu tiplied by the excess weight of sulphate obtained. 77. In case it is wished to prepare a larger quantity of the acid at one time, " measure a gallon of distilled water int< a clean earthenware jar or other suitable vessel. Add to 1 his 94 septems of pure concentrated sulphuric acid and mi> thoroughly." This solution should of course give the san e weight of BaS0 4 as the one just described, viz. 13' 8 grains, but assuming that the weight is greater. "Add now to the diluted acid a measured quantity of wa~er which is to be found by subtracting 13 '8 from the we ght of BaS0 4 obtained in the experiment and multiplying the difference by 726. If these operations have been accu- rately performed, a second precipitation and weighing of the BaS0 4 obtainable from 50 septems of the test-acid, will give nearly the correct number of 13 '8 grains." The only item requiring explanation in the above, is that relating to multiplying by 726. A gallon being equal to 10,000 septems, on adding 94 septems of strong acid to that quantity of water, we obtain 10,069 septems (the decrease as in the previous case being due to contraction) and on taking out 50 septems from this, we have remaining 10,019 septems, which, divided by 13-8, will give as a quotient the number 726, as in the case when preparing a decigallon of the solution, on dividing 956 '9 by 13 '8 we obtained the quotient 69 '34. 78. Having obtained the sulphuric acid solution of the correct strength we now proceed to prepare the ammonia solution of equivalent strength for use with the same. The 144 THE GAS ENGINEER'S LABORATORY HANDBOOK. strongest solution of ammonia obtainable, has usually a specific gravity of about 88, and contains about 2 grains of absolute ammonia (NH 3 ) per septem, consequently on adding 5 septems to the decigallon, we ought to obtain a solution of nearly the correct strength. This quantity of strong am- monia is added to a decigallon of distilled water, the stopper of the mixer inserted, and the contents well shaken ; trials are afterwards made, in the manner described below, to see if 100 septems of the ammonia solution just neutralise 25 septems of the standard acid ; if not, more water or ammonia must be added, as the case may be, until 100 sep- tems will so neutralise 25 septems of the acid solution. Having placed the acid burette perfectly upright in its stand, charge it with the acid solution, and pour from it exactly 25 septems into a clean dry beaker ; the contents of the beaker are then largely diluted with distilled water, and coloured with a few drops of methyl orange solution, which will turn the solution pink. The burette reserved for the alkaline solution is then rilled with the dilute ammonia solu- tion, and the beaker containing the acid, is placed beneath it, upon a white tile or clean filter-paper. The stop-cock of the burette is then opened, and the ammonia solution allowed to gradually fall into the acid solution, which is kept constantly stirred by means of a thin glass rod with rounded ends. The colour of the acid solution will gradually change from pink, to a golden yellow or sherry colour, and when this change is effected by the last drop or two of the solution, we may assume that the latter is correct, otherwise the strength will have to be increased or diluted, by adding strong ammonia or distilled water, as the case may be, until the last drop does so change the colour. 79. Should it be considered necessary to make a larger quantity of the ammonia solution, "measure a gallon of distilled water into a clean glass or earthenware vessel and add to it 50 septems of strong ammonia (sp. gr. 0*88)." This should give a solution of nearly the correct strength, SPECIAL ANALYSES. 145 wh ch must be tested and adjusted, if necessary, in the ma mer just described. 30. Both the acid and ammonia solutions should be kept tig itly stoppered and in a cool place : this particularly apj lies to the ammonia solution, which is very apt to become we ik owing to the volatility of ammonia. Ammonia Apparatus. 81. The glass cylinder filled with glass beads in which the acid is placed, is shown in Fig. 44, immediately in front of he meter ; it has a movable nose-piece at one end, and a gla 3s stop-cock at the other. It is supported in a perfectly ho] izontal position upon a stand. The meter employed indicates the hourly rate of con- sui iption by the observation of one minute, and in addition registers up to 10 cubic feet, when the flow of gas is capable of >eing automatically shut off. The Sulplmr Test Apparatus. 82. Connected to the outlet of the meter is the sulphur tes:, Fig. 44, which, in the Eeferees' form of apparatus, corsists of a small Bunsen burner with a steatite top, which is mounted on a short cylindrical stand A, perforated with holos for the admission of air, and having on its upper surface a deep circular channel to receive the wide end of a glass trumpet-shaped tube B. On the top of the stand, between the narrow stem of the burner and the surrounding trumpet tube B, pieces of commercial sesqui-carbonate of ammonia, weighing in all about two ounces, are placed. The products of the combustion of the gas, and of the gradual volatilisation of the sesqui-carbonate of ammonia, go upwards through the trumpet tube, into a vertical glass vessel C, which is filled with glass marbles, in order to break up the current of the ascending gases and promote condensation. From the top of the cylinder a long glass chimney-pipe D IP'"" SPECIAL ANALYSES. 147 proceeds; this serves to effect some further condensation, as we 1 as to regulate the draught, and affords a means of exit for the uncondensable gases. In the bottom of the cylinder a s mall glass tube is fixed ; through this, the liquid formed dui ing the testing, drops into a beaker placed underneath. Tli 3 Gas Keferees give various precautions to be observed in sel< cting and setting up the apparatus, which will be found in the Appendix, but in addition to those precautions, it is nee sssary to be careful not to place the apparatus where it is t xposed to very low temperatures, or the results obtained wi. I be inaccurate, owing to the fact that there will not be suf icient heat to thoroughly volatilise the sesqui-carbonate of immonia. In the event of there being no chance of av< iding placing the apparatus in such a situation, then it will be necessary to enclose the trumpet tube in a casing. Th s may be made of tin, care being taken not to interfere wit h the air supply, otherwise there will be imperfect corn- bus tion. Fitting up of the Apparatus. 83. The whole train of apparatus is shown in Fig. 44. The bench or table upon which the instruments are placed must be level and firm, and not be liable to vibrations. The sulphuretted hydrogen test is fixed upon a stand at one end of the bench, a metal gas service pipe is connected to i ts inlet, and its outlet is connected by a short piece of red or black, but not grey, india-rubber tube to the ammonia cylinder. The whole train is then connected up as shown, by short pieces of the same kind of india-rubber tube. 84, In order to make a test we proceed as follows : Suspend slips of lead paper in the sulphuretted hydrogen test. Remove the glass cylinder from its stand, taking out the ground glass nose-piece and closing the glass stop-cock. Remove also the glass chimney tube from above the sulphur test burner. Charge the ammonia cylinder from a burette or pipette, with 50 septems of the test acid, which is greatly in L 2 148 THE GAS ENGINEER'S LABORATORY HANDBOOK. excess of any quantity of ammonia likely to be found in the gas, unless the gas contains more than about 6 grains of ammonia in 100 cubic feet. In such case, it will be necessary to employ two cylinders in line. The whole interior surface of the cylinder should be wetted with the acid, as well as the glass beads ; this is effected by holding the cylinder in a perfectly horizontal position, and gently rotating it for a minute or so. The glass cylinder is then replaced on its stand, and reconnected with the gas supply and meter. Place about 2 ounces of fresh sesqui-carbonate of ammonia around the sulphur test burner. The index pointer of the meter is then brought to zero, the gas supply turned on, the sulphur test burner lit, the trumpet tube replaced, and connected with the vertical glass cylinder or condenser, the chimney tube being afterwards placed on top of same. Before leaving the apparatus, see that it is burning at the proper rate of half a cubic foot per hour. Determination of the Ammonia. 85. At the end of each period of testing, the glass stop- cock of the ammonia cylinder is closed, the nose-piece removed, and the cylinder fixed in a vertical position in a suitable support ; a clean beaker is then placed beneath, and the contents of the cylinder allowed to drain into the beaker, taking care to avoid any loss by spurting. The cylinder is then well washed out with distilled water, the washings being collected in the beaker containing the acid drained from the cylinder; the washing should be con- tinued until the liquid ceases to turn blue litmus paper red. The contents of the beaker are then well stirred round by means of a thin glass rod, half the contents (accurately measured) being put aside for analysis, and the remaining half transferred to a well-stoppered bottle, for future reference if required. The burette used for the alkaline solution is then filled SPECIAL ANALYSES. 149 with the standard ammonia solution, and the liquid to be tested, having been diluted with distilled water and coloured w th a few drops of methyl-orange solution, is placed on a white surface beneath the burette. The alkaline solution is n< w added, by small portions at a time, until the colour cl anges, showing that the acid is neutralised, when it will IK. found (unless there is no ammonia whatever in the gas) tl at the acid solution will require less than the 100 septems oi alkali which it originally needed in order to neutralise it, tl e difference, being due to the quantity of ammonia in the g .s, which has been absorbed by the acid. The quantity of standard alkali remaining in the burette a: ter the neutralisation of the acid is noted, and from this, tl e quantity of ammonia in the gas is found as per the E sferees' Instructions (which, however, apply to 10 cubic feet o: gas only). 86. "To find the number of grains of ammonia in 1 )0 cubic feet of the gas, multiply by two the number of &( ptems of test alkali remaining in the burette, and move the decimal point one place to the left." As previously stated, the rule for finding the amount of ammonia, given above, is only applicable when 10 cubic feet oj' gas have passed through the cylinder, but, by means of the following method, the amount may be calculated on any quantity of gas. In order to understand the method it is necessary to note, that the burette containing the alkaline solution is divided into 100 parts, and the whole is equal to one grain of ammonia, consequently each division is equal to one-hundredth of a grain of ammonia. Therefore if we note the number of divisions remaining after the acid has been neutralised, we obtain the number of hundredths of a grain of ammonia corresponding to half the acid employed, and oil multiplying this result by two (because the acid was halved) we obtain the number of hundredths of a grain of ammonia in the quantity of gas operated on, and from which the quantity per 100 cubic feet can easily be calculated. As an 150 THE GAS ENGINEER'S LABORATORY HANDBOOK. example, suppose after filling the cylinder with 50 septems of test acid, and passing 12-6 cubic feet of gas through the same, we found that it took 86 divisions of the ammonia solution to neutralise 25 (or 50 -- 2) divisions of the acid solu- tion in place of 100 as it originally did, then 100 86 = 14 divisions remain, and as each division is equal to one-hun- dredth of a grain of ammonia, 14 divisions are equal to 14 hundredths of a grain, then 14 x 2 = 28 grain of ammonia in the gas which has passed through the cylinder, and ^8_X__100 = 2-22 grains per 100 cubic feet. 12* 6 Determination of the amount of Sulphur. 87. The following is the method of analysing the liquor produced in the sulphur test apparatus, so as to obtain the amount of sulphur. At the expiration of the test, the contents of the beaker containing the products of combustion, are placed in a graduated glass measure, and the trumpet tube, glass cylinder, and chimney tube are washed well out with distilled water, and the washings added to the liquid in the measure. This is thoroughly well mixed, and half the quantity poured into a beaker for analysis, the remainder being placed in a well-stoppered bottle for future reference if required. The contents of the beaker are then acidified with hydrochloric acid (preferably charged with free bromine to secure complete oxidation of the products in the condensed liquid). The beaker is now covered with a clock- glass, its contents raised to the boiling point, and an excess of a saturated solution of barium chloride added, the pre- cipitated BaS0 4 is allowed to settle, and then filtered, washed, dried, ignited and weighed as in (24) ; the weight obtained multiplied by 11 and divided by 4 = the grains of sulphur per 100 cubic feet of gas. 88. This rule is based on the following: 233 parts by weight of BaS0 4 correspond to 32 parts of sulphur, and as grain weights are employed, 233 grains of BaS0 4 , correspond SPECIAL ANALYSES. 151 to 32 grains of sulphur, consequently one grain of BaS0 4 cor- re*- ponds to -13734 of sulphur, then, as half the liquor only w; s used, and exactly 10 feet of gas is assumed to have been burnt, -13734 x 2 x 10 = 2-7468, which is equivalent to multiplying by 11 and dividing by 4, = 2-75. As previously stated, the Keferees' rule for calculating the anount of sulphur per 100 cubic feet, only applies to the pi Dducts resulting from the combustion of exactly 5 cubic fe :t of gas (half the quantity obtained in the combustion of 1< cubic feet), consequently, if any other quantity of liquid tl. in that due to 5 cubic feet of gas be operated on, the w sight of sulphur in that quantity of gas will be found by rn Liltiplyin^ the weight of sulphate obtained by '13734, ai d then finding by proportion the number of grains per 1( cubic feet. It may be of interest to state, that if the gas burnt ai lounts to 2-12 cubic feet, or if the portion of the condensed liquid taken, corresponds to 2 '12 cubic feet, then the BaS0 4 w sighed in centigrams, gives at once the grains of sulphur per 100 cubic feet without any calculation. 89. It is necessary in every experiment to note the temperature and pressure, and if necessary to make a correction for the same. When working to the Keferees' Instructions on 10 cubic feet of gas, the number of grains of sulphur obtained as described above " is to be corrected for tlie variations of the temperature and atmospheric pres- sure," and "the readings of the barometer and thermometer are to be taken for the day on which the testing commenced, and also the day on which it closed, and the mean of the two is to be used. " This correction may be made most simply, and with sufficient accuracy in the following manner : " When the Tabular Number (Appendix) is between 955-965, 966-975, 976-985, 986-995, increase the number of grains of sulphur by -^ths, T f ^ths, T f ths, T ^th of the weight found. 152 THE GAS ENGINEER'S LABORATORY HANDBOOK. " When the Tabular Number is between 996-1005 no correction need be made. " When the Tabular Number is between 1006-1015, 1016- 1025, 1026-1035, diminish the number of grains of sulphur The following is an example : Grains. Grains of BaS0 4 from half the liquid produced in the combustion of 10 cubic feet of gas Multiplied by 11 and divided by 4 = grains of sul- \ jg.^ phur in 100 cubic feet of gas (uncorrected) or / Barometer mean (30 -2)" Thermometer 76 Tabular number 964 Add T ^thsof 15-4 grains = .......... -61 Grains of sulphur in 100 cubic feet of gas (corrected) 16^01 When operating on quantities of gas other than 10 cubic feet, the gas is corrected by multiplying direct by the tabu- lar number, and then making a proportion for the amount of sulphur per 100 cubic feet. 90. The Referees' sulphur test depends upon the follow- ing chemical reactions : Among other substances, coal gas contains carbon, hydrogen, and sulphur, which, in the pro- cess of burning, combine with the oxygen of the air to form water, carbonic acid, and sulphurous acid (S -}- 2 = S0 2 ), but this latter is speedily oxidised, more or less completely into sulphuric acid (S0 8 ). Ammonia being present, the acids unite with it, forming ammonium carbonate and sulphate, which are carried down in solution, by the condensed water resulting from the combustion. Barium chloride combines with both ammonium car- bonate and ammonium sulphate, the acid radicals of those salts uniting with the barium of the BaCl 2 , to form barium carbonate and barium sulphate, while the chlorine of the BaCl 2 unites with the ammonium (NH 4 ) of the ammonium carbonate and sulphate, to form soluble ammonium chloride. SPECIAL ANALYSES. 153 Prom the foregoing it will be seen that if to the solution, as obtained from the sulphur test, we were at once to add a solution of BaCl 2 , we should get a precipitate of both barium carl ionate and barium sulphate, so, in order to prevent this, the solution is acidified with HC1 and boiled, which has the effe-jt of driving oif all the carbonic acid, and completely uni es the NH 4 , which at first was in solution as a carbonate wit i the hydrogen acid, so as to form the other soluble salt Nil ,01. Then on adding the BaCl 2 , we obtain the insoluble Ba; 4 . )1. If desired, the amount of sulphur in the gas can be arr ved at by means of the liquor produced from the combustion of t le gas in the Referees' apparatus, without going through the processes of precipitation and filtering, in the following mai ner : The liquid obtained from the combustion in the Eeferees' sul] 'hur test of a specified quantity of gas, is evaporated in a smiill porcelain basin whose weight has been accurately det< ;rmined, and when the contents are nearly dry, the whole is -ransferred to an air-oven, and the complete drying effected at a temperature of about 240 F. The basin and its contents are then allowed to cool, and subsequently weighed on the balance ; after noting the weight, the basin is replaced in the air-oven for a few minutes, and then cooled and re-weighed, in order to make sure that all adhering moisture has been expelled: this operation should be re- peated until the weight is constant. The increase in weight over the weight of the empty basin is due to sulphate of ammonia, with an infinitesimal quantity of tarry matter, and from this, the amount of sulphur may be obtained as under. Supposing that 24 -6 grains have been obtained, Equivalent Equivalent weight of weight of (NH 4 ) 2 SO 4 Sulphur. As 132 : 32 :: 24-6 : x. x = 5 '97 grains. 154 THE GAS ENGINEER'S LABORATORY HANDBOOK. Estimation of Ammonia in Crude Gas. 92. The amount of ammonia in crude gas is estimated in exactly the same manner as the amount in purified gas, but by means of a stronger acid solution, as crude gas generally contains from 200 to 300 grains per 100 cubic feet. A convenient strength of acid is 10 times that of the Referees' solution, and it is made by adding 10 times the quantity of acid which is employed in making the Referees' solution previously described, to the decigallon of distilled water; we should then obtain a solution of 10 times the strength. It must, however, be tested in the same manner as the Referees' solution was tested, but in place of estimat- ing the amount of BaS0 4 in 50 septems, it is better to take, say, 10 septems, and multiply the result obtained by 5, which should equal 10 times 13*8 grains. The reason for taking 10 septems is that the 50 septems would give such a heavy precipitate as to make it unwieldy to manipulate. The ordinary ammonia solution, 100 septems of which is equal to one grain of ammonia, will serve for titrating. 93. The following is the method of making an experi- ment : An ammonia cylinder is charged with 50 septems of the acid solution, and 5 feet of gas passed through, any tar that may be in the gas, being kept back by cotton wool placed in a calcium chloride tube. The cylinder and con- tents are washed out until free from acid, and the washings, after being well mixed, are placed in a graduated measure, which may conveniently be 20 ounces. If the washings do not amount to that volume, they must be diluted with dis- tilled water to that extent. A twentieth part, or 1 ounce of the solution, after being largely diluted with distilled water, is now titrated with the ammonia solution as pre- viously described, and, after neutralisation, the alkali re- maining in the burette noted. Supposing this to be 60 septems, then, as each septem represents 01 grain of NH 8 , the 60 septems would represent 0-60 grains of ammonia; SPECIAL ANALYSES. 155 but as only a twentieth part of the liquor received was tab n for analysis, this result must be multiplied by 20, which would give 12 grains of ammonia in the 5 cubic feet 12 x 100 of >-as, and therefore = 240 grains in 100 cubic feet. Harcourfs Colour Test. )4. The Referees' method for the estimation of sulphur con pounds, although giving accurate results, is not adapted wh re rapidity is an essential, as, say, when it is required to knc w the condition of the sulphide vessels for the elimination of 1 he sulphur compounds at any particular time ; in this case the colour test devised by Mr. Vernon Harcourt, F.R.S., wil be found more suitable. The test is based on the fact that when coal gas con- taii ing bisulphide of carbon is passed over heated platinised pun ice, the bisulphide is more or less decomposed into sul] huretted hydrogen and methane, and the sulphuretted hydrogen being led into a solution of a lead salt, the intensity of colour imparted to the lead solution from a given quantity of gas, as compared with a standard tint corresponding to a definite quantity of lead sulphide, at onco enables us to determine the amount of CS 2 in the gas under examination. The general arrangement of the test is shown in Fig. 45. 95. The glass bulb, which is filled with platinised pumice, is to be so adjusted that it may be about an inch above the burner, and in the middle of the cylinder. In order to use the apparatus, first turn on the upper stop-cock, thu&- sending gas through the bulb at the rate of about half a cubic foot per hour, as may be judged by lighting the gas for an instant at the end of the horizontal arm, when a flame about an inch long should be produced. Eaise the cylinder, which will be supported by the pressure of the wires, light the burner, and turn down the flame until it forms 156 THE GAS ENGINEER'S LABORATORY HANDBOOK. a blue non-luminous ring. Lower the cylinder, and place the small fire-clay pieces upon it round the neck of the bulb. 96. A testing may be made five minutes after the burner is lit, except when the apparatus is used for the first time, when the gas should be allowed to flow through the bulb for a quarter of an hour or a little longer, and any number of testings may be made consecutively, so long as the heat is maintained. The manner of conducting a test is as follows : FIG. 45. Lay a piece of white paper on the table by the side of the burner, and fix a piece of cardboard upright in the brass clip provided for that purpose the cardboard serves as a background against which to observe the colour of the contents of the test glasses, and should receive a side light, and be as clear as possible from shadows. Place the " standard " glass, " day light " or " gas light " according to circumstances, in its receptacle, and dilute some of the con- SPECIAL ANALYSES. 157 cent rated lead syrup with about twenty times its volume of distilled water, and fill one of the test glasses provided, up 10 the mark scratched on same, with a portion of the liqti'd thus prepared. ] nsert the india-rubber plug with capillary and elbow tub* s, and connect up, as in Fig. 45, with the bulb and aspi 'ator, placing tbe standard glass and test glass side by side The capillary tube should descend very nearly to the boti .>m of the glass, but must not press on the bottom, or it will probably be broken. At starting, the aspirator should be 1 ill of water and the measuring cylinder empty. Turn the ap of the aspirator gradually, a stream of bubbles will rise thn ugh the lead solution. Turn off the tap for a minute, and observe the liquid at the bottom of the capillary tube. If t gradually rises, the india-rubber connections are not tigl t, and must be made so before proceeding further. It is nec ssary to avoid pressing the plugs into the test glass, or the aspirator, while they are connected, which would drive up he lead solution into the inlet tube. )7. When the connections are air-tight, allow the water to run into the measuring cylinder in a slender stream until the lead solution has become as dark as the standard. As the ascending bubbles interfere with the observation of the tint, it is best to turn off the tap when the colour seems almost deep enough; compare the two, turn on the tap if necessary, for a few moments ; then compare again, and so on unt il the colour of the two liquids is the same. The volume of water which the measuring cylinder now contains, is equal to the volume of gas which has passed through the lead solution. This volume of gas contained a quantity of sulphur as carbon bisulphide, which, as lead sulphide, has coloured tbe liquid in the test glass to the same depth as the standard tint. The standard has been made such, that to impart this tint to the solution of liquid contained in the glass, 0*0187 grains of lead sulphide must be present, con- taining 0-0025 grains of sulphur. Hence, supposing the 158 THE GAS ENGINEER'S LABORATORY HANDBOOK. measuring cylinder, each division of which corresponds to TOTHT * n f a cubic foot, to have been filled to the thirtieth division, ^g^ ths of a cubic foot of gas contained 0025 grains of sulphur. From these figures the number of grains of sulphur existing as bisulphide of carbon in 100 cubic feet of gas can easily be calculated. 98. The following table gives the relation between (V) 1;he divisions of the measuring cylinder filled with water, and (S) the grains of sulphur existing as bisulphide of carbon in 100 cubic feet of gas. Since gas contains, besides carbon bisulphide, some other sulphur compounds, which are not decomposed by the action of heat into sulphuretted hydrogen, and which contain sulphur amount- ing on an average to seven or eight grains per 100 cubic feet of gas, it is necessary to add this quantity to that found by the colour test if it is wished to know approxi- mately the total amount of sulphur in the gas. 500 V. S. V. S. V. S. V. S. V. S. 10 50-0 29 17-2 48 10-4 67 7-5 86 5-8 11 45-4 30 16-7 49 10-2 68 7-4 87 5-7 12 41-7 31 16-1 50 10-0 69 7-2 88 5-7 13 38-5 32 15-6 51 9-8 70 7-1 89 5-6 14 35-7 33 15-1 52 9-6 71 7-0 90 5-6 15 33-3 34 14-7 53 9-4 72 6-9 91 5-5 16 31-3 35 14-3 54 9-2 73 6-9 92 5-4 17 29-4 36 13-9 55 9-1 74 6-8 93 5-4 18 27-8 37 13-5 56 8-9 75 6-7 94 5-3 19 26-3 38 13-2 57 8-8 76 6-6 95 5-3 30 25-0 39 12-8 58 8-6 77 6-5 96 5-2 21 23-8 40 12-5 59 8-5 78 6-4 97 5-2 22 22-7 41 12-2 60 8-3 79 6-3 98 5-1 23 21-7 42 11-9 61 8-2 80 6-2 99 5-1 24 20-8 43 11-6 62 8-1 81 6-2 100 5-0 25 20-0 44 11-4 63 7-9 82 6-1 150 3-3 26 19-2 45 11-1 64 7-8 83 6-0 27 18-5 46 10-9 65 7-7 84 6-0 28 17-9 47 10-6 66 7-6 85 5-9 SPECIAL ANALYSES. 159 09. For another experiment, the test glass is to be disc< nnected and recharged ; the water in the measuring cylii der being poured back into the aspirator. The colour of the ; tandard is unaffected by exposure to light, but deepens if tl e liquid is warmed, resuming its original shade as the liqu d cools. If, therefore, the glass containing the standard has >een kept in a warm place, it should be allowed to cool befo -e testing. r J'he lead solution used in the experiments becomes colo' .rless after being exposed to the light for a few hours, and nay therefore be used over and over again for twenty time 5 or more, provided that it does not absorb carbonic acid from the air. The best mode of working is to have two well-corked bottles, into one of which the coloured liquid is er iptied, while the test glass is recharged from the other. ] t is hardly necessary to remark, that in the event of mak ng an experiment for CS 2 on crude gas containing SH 2 , the 1 itter impurity should previously be removed by passing it th rough oxide of iron. IV. ANALYSIS OF AMMONIACAL LIQUOR. 100. Gas liquor may be said to consist of a solution of the following ammonium salts. Volatile Salts : Ammonium carbonate (mono- sesqui- bi-), sulphide (NH 4 ) 2 S. hydrosulphide NH 4 HS. cyanide. ,, acetate. Fixed Salts . Ammonium sulphate. sulphite. thiosulphate (hyposulphite). thiocarbonate. chloride. ,, sulphocyanide (thiocyanate). ferrocyanide. 160 THE GAS ENGINEER'S LABORATORY HANDBOOK. Also the salts of organic bases, especially those of the pyridine series, phenols, and other matters of a tarry character. The term " free " as employed here, is not to be taken in its ordinary sense, but rather that the compounds of ammonia with such acids as carbonic and sulphydric, can be removed from their solutions by simply boiling them for a sufficiently long period, or by acting on them with dilute acid in the cold : the ammoniacal salts called " fixed " are not removed from their solutions by boiling, or by treatment with acid, so it is necessary to heat such solutions in the presence of a caustic alkali or lime, which sets the ammonia free from its combination with the acids. Free ammonia in the ordinary sense of the word means a solution of caustic ammonia. , 101. The chief and most often recurring of the volatile salts are the sulphide and the carbonate ; of the fixed salts the chloride and the sulphocyanate. The percentage of the total ammonia combined as fixed salt varies greatly in different samples of liquor. The occurrence of these different ammonium salts and the proportion in which the ammonia is combined with each acid depend on (1) the description of coal distilled, (2) the tem- perature at which it is distilled, (3) the temperature to which the liquor has been subjected, and whether it has been exposed to the air or not, and (4) the point in the condensing or scrubbing plant from which the sample is taken, and the general treatment to which the gas is subjected during the operation of washing and scrubbing. For example if a coal is distilled which contains an appreciable quantity of sodium chloride (common salt), hydrochloric acid is given off, and we therefore find a large percentage of fixed ammonium chloride in the liquor. On the high or low heats of the retorts depends in a great measure the larger or smaller percentage of fixed ammonium sulphocyanate in the liquor. Exposing liquor to a high temperature, or to the air at ordinary tern- SPECIAL ANALYSES. 161 * pera hires, causes it to lose ammonia, thereby converting its amn onium sulphide (NH 4 ) 2 S as well as ammonium mono- carl jnate, (NH 4 ) 2 C0 3 respectively, into ammonium hydro- sulphide NH 4 HS, and ammonium bicarbonate NH 4 HC0 3 . It also has the effect of oxidising its ammonium sulphide to ami; ionium thiosulphate, and in some very slight degree to ami ionium sulphate. 02. The value of ammoniacal liquor is commercially ex- pret;ed by the number of ounces by weight of mono-hydrated sulj 'turic acid, H 2 S0 4 , which is required to saturate one gallon. For instance, if a liquor is said to be of eight-ounce strength, it mea is that a gallon of the liquor would require eight ounces of pur( , strong, sulphuric acid to neutralise it. The crudest method by which the strength of liquor is deti rmined, is based on the assumption that the quantity of acic required to saturate the liquor, bears a certain relation to t le specific gravity of the latter. In this method a No. 1 Tw-iddell's hydrometer is employed, having a scale divided and numbered from zero at the top to 25 at the lower part of its t tern. A glass jar of a suitable size having been filled with liquor, which should be at the temperature of 60 F., the hydrometer is immersed, and the division on the scale to which it sinks in the liquor noted. Each degree upon the scab is supposed to indicate the presence of as much NH 3 in one gallon of the liquor as will require two ounces by weight of pure strong sulphuric acid to neutralise it. For instance, if the hydrometer sinks to the division 5 on the scale, the liquor is said to be of ten-ounce strength. This method of determining the strength of ammoniacal liquor is not at all an accurate one, and is only suitable for general guidance in the working of the scrubbing or washing plant in gasworks; it should never be used when the actual amount of ammonia in the liquor is required, for, supposing a liquor to contain an excess of caustic ammonia, its strength would be under-estimated, owing to ammonia being lighter than water, and thus reducing the specific gravity of the M 162 THE GAS ENGINEER'S LABORATORY HANDBOOK. liquor; on the other hand, if a liquor had been repeatedly passed through the scrubbers, and thus become charged with an excess of carbonic acid, it would probably be over-esti- mated, as C0 2 would raise the specific gravity. 103. Another method of estimating the value of liquor is to find by an experiment the number of ounces of pure strong sulphuric acid which are required to neutralise the ammonia in a gallon of liquor, but as it would not be practicable to operate on so large a quantity as one gallon, or to employ strong sulphuric acid, the following modification is em- ployed : A graduated glass vessel called an alkalimeter, Fig. 46, is provided, and this is graduated into 1 6 parts. A solution of sulphuric acid is made by adding one pound of pure sulphuric acid (sp. gr. 1845) to distilled water, and making up the solution to one gallon at a temperature of 60 F. The specific gravity of this solution should be 1064-4. The testing is conducted as follows : The alkalimeter is first rinsed out with a portion of liquor, and is after- wards charged with liquor to the top of the scale ; this is then poured carefully out into a beaker or porcelain ba^in. The instrument is then well rinsed out with distilled water, and the rinsings added to the liquor in the basin. In order to guard against loss whilst pouring, it is advisable to grease the edge of the beak of the alkalimeter. A small quantity of the test acid is now poured into the alkalimeter and the instrument well rinsed out with the same ; this is then poured away ; the alkalimeter is next charged with the acid to the top of the scale. Having largely diluted the contents of the beaker with distilled water, and coloured the FIG. 46. SPECIAL ANALYSES. 163 solution with a few drops of methyl orange indicator, which will give it a yellow tint, the contents of the alkalimeter are ca -efully poured into the liquor. This operation should be performed very gradually and by sn all portions at a time, the contents of the beaker being well stirred between each addition of the acid, so as to allow of the disengagement of the carbonic acid and sulphuretted h} drogen in the liquor. When all the liquor has been neu- tr; Used, the solution will change to a pinkish tint, and on reading oft the number of divisions of acid required for ne itralisation, the ounce strength is obtained. Supposing it required nine divisions, then the liquor would be of nine- ou ice strength. By this method the same results are ob ained as if we had added strong acid to a gallon of liquor, foi the alkalimeter being divided into sixteen parts, each part represents such a proportion of acid solution as would contain ono ounce of the strong acid, while its total contents represent on", gallon. It is first used to measure the liquor, and after- wa rds to measure the acid solution, and under these condi- tioas it does not signify, within certain limits, what its true capacity may be. This method of determining the strength of ammoniacal liquor is also unreliable, as it does not take any cognisance of Jie fixed ammonia. The only reliable method of obtaining the value of a sample of liquor, is to heat it with an excess of potash or soda, which will liberate the whole of the ammonia, from its combination with acids ; the evolved ammonia is led into a solution of sulphuric acid, which is afterwards titrated with an alkaline solution of equal strength, and from the amount of acid neutralised by the ammonia, the amount of the latter is determined as in (59). 104. The solution of sulphuric acid is of the same strength as that already mentioned, and is commonly known as ten per cent. acid. It is prepared by pouring about three pints of distilled water into a half-gallon graduated measure, and adding to this, half a pound (4* 34 ounces by measure) of pure M 2 164 THE GAS ENGINEER'S LABORATORY HANDBOOK. concentrated sulphuric acid (sp. gr. 1845). The solution is well mixed by stirring with a glass rod, the vessel is covered, and the mixture allowed to cool down to the temperature of 60 F. When this temperature is attained, the additional quantity of water necessary to bring the volume to exactly half a gallon is added, and the solution again well mixed. Its specific gravity may now be tested by means of a delicate hydrometer, and if more than half a degree above or below 1064*4, it will be necessary to add small quantities of water or of acid, until the correct specific gravity is closely approached. The following table, which is due to Mr. L. T. Wright, will be found useful when there is a difficulty in obtaining the temperature of 60 F. ; it shows what the specific gravity of the acid should be at different temperatures below and above 60 F., in order to correspond to the specific gravity 1064 -4 at 60 F. Temp. F. Specific Gravity. Temp. Specific Gravity. Temp. F. Specific Gravity. Temp. F. Specific Gravity. 35 1068-85 54 1065-64 61 1064-19 68 1062-72 40 1068-10 55 1065-45 62 1063-98 69 1062-51 45 1067-30 56 1065-24 63 1063-77 70 1062-30 50 1066-40 57 1065-03 64 1063-56 75 1061-20 51 1066-21 58 1064-82 65 1063-35 80 1060-05 52 1066-02 59 1064-61 66 1063-14 85 1058-95 53 1065-83 60 1064-40 67 1062-93 90 1057-80 105. The strength of the solution may also be checked gravimetrically or volumetrically. In order to check it by the gravimetric method, exactly one ounce by measure of the solution is poured into a beaker, and having been largely diluted with distilled water, neutra- lised by ammonia, and afterwards rendered acid by hydro- chloric acid, it is raised to the boiling point, and an excess SPECIAL ANALYSES. 165 of solution of barium chloride added. The precipitate of BaS0 4 obtained, is treated in the usual manner, and should wtigh, after deducting the filter ash, 104*017 grains, which is arrived at as under. Half a gallon = 80 fluid ounces, and J Ib. of acid (3500 gr tins troy) are contained in J gallon of the solution, con- se juently each ounce by measure of the acid solution cc itains 43 '75 grains of sulphuric acid, consequently 4- ' 75 x 233 = 104-017 grains of BaS0 4 , which is equal to 4: *75 grains of sulphuric acid. If the weight of BaS0 4 ol tained differs from 104*017 grains, the solution must be m ide stronger or weaker, as the case may be, until the correct w :ight is obtained. 106. The solution may be checked volumetrically by IK utralising chemically pure Na 2 C0 3 , using methyl orange as indicator. It is essential that the Na 2 C0 3 be pure ; this is ascertained by dissolving about 80 grains of it in distilled \v iter, when a clear colourless solution should be obtained, ai d testing for the presence of sulphates or chlorides. To dc this, add an excess of HN0 3 in order to expel all C0 2 , ai d to one half of the acidified solution add a few drops of pure BaCl 2 solution, and to the other half a few drops of A^N0 3 solution. The solutions should remain clear in both cases. Should there be any difficulty in procuring pure Na 2 C0 8 it may be obtained in the following way : Shake up a quantity of the best commercial sodium bicarbonate (NaHCOg) with successive small quantities of distilled water, until the washings when tested for Cl or H 2 S0 4 in the manner described above, give negative results. The NaHC0 3 so treated, is then dissolved in boiling water, the solution filtered, evaporated over a Bunsen burner almost to dryness, stirring constantly as the mass becomes pasty, and finally to complete dryness on the water-bath. The salt is then roughly powdered and preserved in a stoppered bottle. To standardise the acid solution, take about 200 grains of 166 THE GAS ENGINEER'S LABORATORY HANDBOOK. the NaHC0 3 an( ^ l ]ea ^ it earefulLy in a platinum crucible over a small Bunsen flame for about 20 minutes, stirring constantly so as not to fuse the Na 2 C0 3 , until all moisture is expelled ; place the crucible in the desiccator to cool, and when cold, weigh out 47 32 grains on a watch-glass, this being the quantity which will be neutralised by 1 fluid ounce of the standard acid if it is correctly made up. The weight 47*32 is thus obtained: 106 parts by weight of Na 2 C0 3 , are neutralised by 98 parts of H 2 S0 4 , and as 1 fluid ounce of the solution contains, as previously stated, 43*75 grains of H 2 S0 4 , this would require 47*32 grains of "Na 2 C0 3 for neutralisation : - ^g - = 47*32. The 47*32 grains of Na 2 C0 3 are washed off" the watch- glass, by means of distilled water from the wash bottle, into a deep white porcelain dish, a few drops of methyl orange solution added, and the standard acid run in from a gradu- ated burette, the solution being constantly stirred while this is effected, until a drop of the acid suddenly changes the colour of the solution from a golden yellow to a reddish- brown tint. A second weighed quantity of Na 2 C0 3 is then treated in the same way, in order to check the result ob- tained with the first quantity weighed out. Should the acid solution prove either too weak or too strong, the cal- culated quantity of water must be added, and the strength of the solution again tested with Na 2 CO 3 . When the acid solution has been obtained of the correct strength we may compare it with the alkaline (soda) solu- tion, which should exactly match it. 107. The latter solution is made by dissolving " stick " caustic soda, " not by alcohol," in water, filtering the solution through asbestos, and diluting until it has a specific gravity of about 1086 at 62 F. ; this will about match the acid solution ; in order to ascertain whether it does so, a burette having a capacity of 2 fluid ounces, and divided into 32 parts is provided; this is charged in the ordinary way with SPECIAL ANALYSES. 167 2 fl lid ounces of the caustic soda solution. Exactly 1 fluid ounoe of the acid solution is then measured out by means of a p^ pette into a beaker, diluted with distilled water, coloured wit i a few drops of tincture of cochineal solution, and placed on a white tile or filter paper, beneath the burette con raining the soda solution. The latter solution is then slcrvly run in to the acid (which is kept constantly stirred) um il the colour suddenly changes to purple. If more or lest than 1 ounce of the soda solution be required to effect thi ; change, small additional quantities of caustic soda or of wa er must be added to the remainder of the ^ gallon, and fre ;h experiments made until the correct strength is attained. Th j solution must then be immediately bottled, tightly corked, and kept in a cool place. 108. The apparatus employed is the same as was de- scr bed in the analysis of ferrous ammonium sulphate, and is she wn in Fig. 33. Having connected the whole apparatus together as shown, into the flask A place, by means of a pipette, 2 fluid ounces of ;he test acid ; this should be poured through the tube B, in order to thoroughly moisten with acid the pieces of brc ken glass contained therein. Then pour into the flask E, one measured ounce of the ammoniacal liquor, closing the stop-cock of the funnel immediately after so doing ; a little distilled water should be added to the measure which con- tained the liquor, and the rinsings added to the contents of the flask E. One ounce of caustic soda solution, made by dissolving some " stick soda " in the proportion of 1 ounce of soda to 9 ounces of distilled water, is also added to the contents of the flask E. The Bunsen burner beneath the sand-bath is then lit, and the contents of the flask E are brought to the boiling point, and maintained in that condi- tion for about 30 minutes, by which time all the ammonia contained in the liquor will have passed over into the flask A. The stop-cock on the funnel is then opened, the Bunsen burner turned off, and the apparatus allowed to cool. The CNIVEBSITt 168 THE GAS ENGINEER'S LABORATORY HANDBOOK. point of the tube in the flask A must not touch the surface of the acid during the experiment. The apparatus is now disconnected, and the pieces of glass in B rinsed with a little distilled water into the flask A. The whole contents of the flask A are then transferred without loss to a beaker, and coloured with a few drops of cochineal solution. 109. The burette having been charged with the standard solution of soda (107) the beaker is placed underneath, and the soda solution run in, until the colour of the acid solution changes to purple, when the division at which the solution now stands in the burette must be noted. The strength of the liquor under examination is indicated by the quantity of soda solution remaining in the burette, which represents the quantity of ammonia which has been given off from the liquor and absorbed by the acid in the flask A. For example, if 12 divisions of soda solution be left in the burette, the liquor contains sufficient ammonia in 1 gallon to neutralise 12 ozs. of pure strong sulphuric acid. The reason of this is obvious. In the flask A, 2 ounces of acid were placed, and the burette contained originally 2 ozs. of soda solution, which exactly matched the 2 ozs. of acid. In the flask E 1 oz. of the liquor was placed, consequently the volumes of the two standard solutions are exactly double the volume of the liquor operated on. Now if the liquor had been exactly sixteen-ounce strength, its ammonia would have neutralised just 1 fluid ounce, or half the acid employed ; the remaining half would then have been neutralised by half the soda in the burette, or by 16 divisions, and therefore 16 parts would be left and indicate the strength of the liquor. The excess of acid is employed in order to ensure that no ammonia escapes. 110. The following are useful notes in connection with this subject : Each ounce of acid (H 2 S0 4 ) combines with 0-347 ounce of ammonia, forming neutral ammonium sulphate (NH 2 S0 4 ). In order to find the quantity of sulphate of ammonium SPECIAL ANALYSES. 169 cap ible of being produced from any liquor, the ounce strength of which has been determined by distillation in the manner described, multiply the ounce strength by I*c47, and by the number of gallons. If the strength of the liquor is given in terms of the percentage of ammonia, mn tiply the total percentage ounces by 4- 61, then by 1-347, anc finally by the number of gallons. The product of the wh )le will be the total weight in ounces of sulphate. In. oru 3i- to calculate direct into pounds per gallon, multiply the oui ce strength by *841 in place of 1'347, moving the de< imal point one place to the left, thus, liquor ten-ounce str sngth 10 x 841 = 8 410, moving the decimal point = * 8410 poi nds sulphate per gallon. 111. The value of a liquor may also be determined in the same apparatus as described above, using the ordinary systematic normal solutions : 10 c.c. of the liquor are distilled wrh caustic soda, and the evolved ammonia led into 20 c.c. of normal acid. The acid is then titrated with normal caustic soda solu- tion. As an example, 8 c.c. of normal soda solution were re- quired to neutralise the 20 c.c. of acid, then the quantity of ammonia is 12 x 0-017 x 10 = 2 '04 per cent. 112. In order to be able to tell whether the scrubbers are doing their proper amount of work in taking out the impurities C0 2 and SH 2 from the gas, it is desirable to make periodical analyses of the liquor for the amounts of these gases. The following are methods for determining the same : Estimation of Sulphuretted Hydrogen in Gas Liquor. Weigh out- 100 grains of the gas liquor into a light glass flask, and wash the contents of the flask into a beaker ; then add an excess of solution of arsenite of soda, and afterwards hydrochloric acid to acid reaction, and allow the precipitate to settle. This, which is As 2 S 3 , should be collected on a 170 THE GAS ENGINEER'S LABORATORY HANDBOOK. weighed filter, and treated in exactly the same manner as in the case of As 2 3 previously given (26). Eveiy 246 parts of As 2 S 3 obtained correspond to 102 parts of H 2 S, and from this the percentage of H 2 S, by weight, in the sample is arrived at. The cubic inches per gallon may be obtained from the percentage by weight in the following manner : First, take, by a hydrometer, the specific gravity of the liquor, from which the weight in grains of one gallon of the liquor is obtained. Thus, supposing the gravity of the liquor to be 1030, we say Sp.gr. H,0 Bp.gr. As 1000 : 1030 : : 70,000 : 72,100. We next require to find the grains of H 2 S in one gallon of the liquor, and, supposing the percentage by weight was found to be 0*58, we say Weight in Percentage of Weight of gallon Weight of H 2 S grains of liquor ' g 5 ^of liquor in in a gallon of analysed. grains. liquor in grains. As 100 : 0-58 :: 72-100 : 418-1. We next require to find the cubic inches of H 2 S, corre- sponding to 418-1 grains ; and to do this we say Weight of one One cubic Weight of H 2 S Cubic inches cubic inch of H 2 S inch in a gallon per gallon As 0-365 : 1 :: 418-1 : 1145. 113. A method of determining volumetrically the amount of H 2 S in a sample of gas liquor is also given. It is based on the fact that when a solution of iodine is brought into contact with sulphuretted hydrogen, the follow- ing reaction takes place : H 2 S -f I 2 = 2HI + S. The first operation is to prepare a standard solution of iodine, which is performed as follows : Dissolve 149-4 grains of pure iodine with 300 grains of potassium iodide. This is then diluted with distilled water SPECIAL ANALYSES. 171 until it exactly measures 2 deci gallons, when 100 septems of the solution will equal 1 grain of H 2 S. The following is the method of making an experi- men : : Weigh out, into a small flask, 100 grains of the gas liquor, and wash this out with distilled water into a large beaker ; dilirie this solution with distilled water until it measures aboi t a pint ; add some starch paste, and afterwards acetic acid to acid reaction. Then, having placed the beaker on a \\ bite filter-paper, run in the iodine solution until the consents of the beaker turns a blue colour (showing that all :he H 2 S has been neutralised), when the number of divi sions required should be noted, each septem representing T
= CaO.H 2 0). A good idea of the value of a lime may be ormed from the way it behaves- on slaking ; a good lime sho ild slake easily, and form, when thus treated, a light uni orm powder, free from lumps, which, on the addition of dili te hydrochloric acid, should not give any marked effer- ves 'ence, indicating that all the C0 2 has been driven off in the operation of calcining the limestone. When quicklime is exposed to the action of the atino- spli ^re for a considerable length of time it slowly absorbs wai 3r therefrom, and falls to powder. In that condition it fori is an exceedingly light dry powder, similar to calcined nia< -nesia in appearance. Such lime which has " fallen," or bec< ime " air-slaked," as it is termed, is partially hydrated, its probable formula being, according to Prof. Wanklyn, (Ca 3) 2 H 2 ; lime in this condition is very caustic and cor- rosi/e. Lime is placed in the purifiers after being slaked. 118. The theoretical purifying value of perfectly pure hydrate of lime is as follows : Seventy-four parts by weight are capable of uniting with 44 parts of C0 2 , or 34 parts of H 2 S. Adopting a theoretical unil of 1 Ib. of pure calcium hydrate, such amount is capable of i niting with 0*594 Ib. of carbonic acid, equal to about 5 cubic feet, or to '46 Ib. of sulphuretted hydrogen equal to the same bulk. The actual value of any particular sample of lime will, of course, be less than the theoretical value, according to the amount of impurities which the lime contains, 119. The impurities existing in lime may be derived from the following sources : The chalk or limestone from which the lime is obtained may be impure, and therefore yield a correspondingly 178 THE GAS ENGINEER'S LABORATORY HANDBOOK. inferior lime on being calcined ; the "burning may not have been efficient, the consequence of which would be to leave an undue percentage of undecomposed carbonate in the lime ; the lime may have been of good quality in the first instance, but may have become subsequently deteriorated by exposure to the weather (air-slaked). 120. In ascertaining the commercial value of a lime for purifying purposes, it is not generally necessary to make an exhaustive analysis with the idea of ascertaining the amount of the various impurities present, all that is required is to determine the amount of available lime (CaO). One method of arriving at the value of a sample of lime is to note the amount of water it takes up on slaking, as the weight of water taken up represents the quantity which has entered into chemical combination with the caustic lime, as previously explained. 121. A fair sample of the lime is taken, and from this between 6 and 7 grams are weighed out into a porcelain basin. Water is then added, in sufficient quantity to thoroughly slake the lime. The basin and its contents are then placed in an air-oven, and dried at a temperature of 250 F., and when quite dry, weighed. The total weight will have increased by the amount of water necessary to convert the caustic lime into hydrate, and the amount of such increase is, therefore, an indication of the purity or impurity of the sample. As an example, 6 '84 grams of lime were weighed out into a basin, slaked, dried and weighed as above described ; the gain of weight observed in the second weighing amounted to 2 '05 grams = 30 per cent., con- sequently, as 18 parts of water absorbed, represent 56 parts of caustic lime, the percentage of the latter is obtained from the following equation : As 18 : 30 : : 56 : x. x 93-33 per cent, of caustic lime. 122, The amount of caustic lime alone, however, does not SPECIAL ANALYSES. 179 always fairly represent the actual purifying value of any par- tic alar sample, for a portion of the real lime present may have abeady been converted into hydrate by the absorption of m< isture from the atmosphere, and consequently, it would not be able to take up any more water. It is necessary, therefore, in order to ascertain the actual purifying value of any sample, to determine the amount of the whole of the lime present in at available condition. This is effected as follows : 123. A portion of the slaked lime obtained as above, eq lal to a definite amount of the original lime, is placed in a be iker, and dissolved in an excess of a measured quantity of nc rmal hydrochloric acid (56), should it be necessary to assist th ) solution by heat employing only a very moderate tempera- te *e. When dissolved, a little distilled water is added, and the so ution coloured with a few drops of methyl orange solution. T] ie liquid is then titrated with normal soda solution (58), ui til the pink colour of the solution changes to yellow. On de lucting the number of c.c. of the soda solution which are re [uired to neutralise the acid from the number of c.c. of ac ^d added to the lime, the number of c.c. of acid which it has taken to combine with the lime, in accordance with the fo lowing equation, is arrived at : CaO,H 2 + 2HC1 = CaCl 2 + 2H 2 0, and from this the percentage amount of lime can readily be calculated. The following is an example on another sample of lime : A portion of lime, weighing 5*462 grams, on being slaked, absorbed 1'44 grams of water = 26*36 per cent., which would correspond to 82 per cent. CaO. A fifth part of this slaked lime, weighing 1-380 grams, corresponding to 1'0924 grams of the original lime, was then dissolved in 50 c.c. normal hydrochloric acid. On titrating this solution with normal soda solution, it required 16*4 c.c. of the latter to neutralise it; therefore 50 16 '4 = 33*6 c.c. of the acid were required to combine with the lime in the sample N 2 178 THE GAS ENGINEER'S LABORATORY HANDBOOK. inferior lime on being calcined ; the burning may not have been efficient, the consequence of which would be to leave an undue percentage of undecomposed carbonate in the lime ; the lime may have been of good quality in the first instance, but may have become subsequently deteriorated by exposure to the weather (air-slaked). 120. In ascertaining the commercial value of a lime for purifying purposes, it is not generally necessary to make an exhaustive analysis with the idea of ascertaining the amount of the various impurities present, all that is required is to determine the amount of available lime (CaO). One method of arriving at the value of a sample of lime is to note the amount of water it takes up on slaking, as the weight of water taken up represents the quantity which has entered into chemical combination with the caustic lime, as previously explained. 121. A fair sample of the lime is taken, and from this between 6 and 7 grams are weighed out into a porcelain basin. Water is then added, in sufficient quantity to thoroughly slake the lime. The basin and its contents are then placed in an air-oven, and dried at a temperature of 250 F., and when quite dry, weighed. The total weight will have increased by the amount of water necessary to convert the caustic lime into hydrate, and the amount of such increase is, therefore, an indication of the purity or impurity of the sample. As an example, 6*84 grams of lime were weighed out into a basin, slaked, dried and weighed as above described ; the gain of weight observed in the second weighing amounted to 2*05 grams = 30 per cent., con- sequently, as 18 parts of water absorbed, represent 56 parts of caustic lime, the percentage of the latter is obtained from the following equation : As 18 : 30 : : 56 : x. x = 93-33 per cent, of caustic lime. 122. The amount of caustic lime alone, however, does not SPECIAL ANALYSES. 179 alv- T ays fairly represent the actual purifying value of any par- tic ilar sample, for a portion of the real lime present may have already been converted into hydrate by the absorption of me isture from the atmosphere, and consequently, it would not be ;ible to take up any more water. It is necessary, therefore, in order to ascertain the actual purifying value of any sample, to determine the amount of the whole of the lime present in an available condition. This is effected as follows : 123. A portion of the slaked lime obtained as above, eq lal to a definite amount of the original lime, is placed in a be iker, and dissolved in an excess of a measured quantity of no :mal hydrochloric acid (56), should it be necessary to assist th s solution by heat employing only a very moderate tempera- tu 'e. When dissolved, a little distilled water is added, and the so ution coloured with a few drops of methyl orange solution. Tl e liquid is then titrated with normal soda solution (58), UL til the pink colour of the solution changes to yellow. On de lucting the number of c.c. of the soda solution which are required to neutralise the acid from the number of c.c. of ac d added to the lime, the number of c.c. of acid which it ha 3 taken to combine with the lime, in accordance with the following equation, is arrived at : CaO,H 2 + 2HC1 = CaCl 2 + 2H 2 0, anil from this the percentage amount of lime can readily be calculated. The following is an example on another sample of lime : A portion of lime, weighing 5*462 grams, on being slaked, absorbed 1*44 grams of water = 26*36 per cent., which would correspond to 82 per cent. CaO. A fifth part of this slaked lime, weighing 1*380 grams, corresponding to 1*0924 grams of the original lime, was then dissolved in 50 c.c. normal hydrochloric acid. On titrating this solution with normal soda solution, it required 16*4 c.c. of the latter to neutralise it ; therefore 50 16*4 = 33*6 c.c. of the acid were required to combine with the lime in the sample N 2 Therefore 0-028 x 33-6 = 0-9408 gram of lime present, 0-9408 x 100 consequently - - =86-1 per cent, of lime in all forms, caustic, hydrated and carbonated. If there is much carbonate of lime present, shown by the lime effervescing strongly on the addition of the acid, it will be necessary to estimate the amount as in (34). Should there be much difference between the result of the slaking and of the acid test as just described, the lime is either not sufficiently burnt, and will therefore contain a quantity of useless carbonate, which is shown by the effer- vescence just spoken of, or else it has been exposed to the air for some time, and become air-slaked to some extent. 124. The following is another method of determining the value of a sample of lime : For the amount of free CaO, weigh 100 grams of an average sample, slake it completely, put the resulting milk of lime into a half-litre flask, fill up to the containing mark, and shake well. Then take out 100 c.c., run it into a half- litre flask, fill up with distilled water, mix well, and employ 25 c.e. of the contents, equal to 1 gram of quicklime (CaO), for the test. Titrate with normal oxalic acid (57) and phenol- phthalein as an indicator. The colour is changed when all free lime has been saturated and before the CaC0 3 is acted on. For C0 2 , the CaO and CaC0 3 are titrated together by dissolving in an excess of normal HC1 (56) and titrating back with normal soda (58). By deducting the CaO first obtained, the quantity of CaC0 3 is arrived at. For very accurate estimations the C0 2 is expelled by HC1, and ab- sorbed in soda lime as in (34). VI. ANALYSIS OF LIMESTONE. 125. As mentioned under the analysis of lime, the lime employed in gasworks is usually met with as quick, or caustic lime, but as it may sometimes be requisite to know SPECIAL ANALYSES. 181 tl e composition of the chalk or limestone from which the li ne is obtained, the following method may be employed. Tlie substance selected for an example is dolomite or n Mgnesium limestone, which contains both calcium and 11 agnesium carbonates. Limestone also contains moisture, silicious matter, iron and aluminium, and sometimes man- and the spent oxide consequently contains 2 x 1 9 = 3 8 per cent, of ammonium sulphocyanide. 130. Estimation of ferrocyanides in spent oxide. The residue from the extraction of the sulphocyanide, corresponding to 50 grams spent oxide, is warmed for some time with an excess of caustic soda. There should not be too great an excess, and the whole should be well stirred. The mixture is then filtered, the residue washed with hot water, and the filtrate boiled with pure ammonium chloride or sulphate solution until no further smell of ammonia is per- ceptible. In this manner the A1 2 3 dissolved by the NaHO is precipitated. Its elimination is necessary, as it is precipitated, together with Prussian blue, even in slightly acid solutions ; more- over, by its precipitation, dark tarry matters are likewise carried down, and the solution is thus partially clarified. After filtering off the A1 2 O 3 , the solution is made slightly acid with HC1, heated to boiling, and poured into a hot dilute solution of ferric chloride. The Prussian blue thus precipitated is allowed to settle, the dark red liquid poured off, and the precipitate well washed by decantation. It is then poured on to a filter, and washed with hot water until the filtrate is free from iron, dried and ignited in a platinum crucible to which the air has access. It is thus converted into ferric oxide, which is weighed, and the quantity of ferrocvanide calculated from the weight found. SPECIAL ANALYSES. 193 560 parts of ferric oxide correspond to 636 parts ferro- cyanogen (Fe (CN) 6 ) and to 860 parts of Prussian blue. Dr. Arnold, in his 'Ammonia and Ammonium Com^ I ounds,' describes a volumetric process which he is in the 1 abit of using. 42 '2 grams of pure dry potassium ferrocyanide (K 4 Fe ( CN) + 3H 2 = 422) are dissolved in water and the soln- t ion diluted to one litre. A dilute, slightly acid, ferric ( hloride solution is placed in a burette and 10 c.c. of the \ 3rrocyanide solution titrated with it. So long as an excess < f ferrocyanide remains in solution, no Prussian blue is precipitated. If a drop of the dark blue liquid be placed < n filter-paper, it spreads out to a uniform blue spot. As ^oon, however, as all ferrocyanide is converted into Prussian 1 lue, the spots do not spread uniformly, but give a blue tatch, in which the particles of precipitated Prussian blue ; ,re easily seen. The end of the titration may be accurately detected by employing filter-paper previously soaked in a solution of ammonium sulpho-cyanide. The slightest xcess of iron is then recognised by the fact that the blue ] >articles no longer appear on a white, but on a yellow or red ground. The strength of the iron solution having thus been ascer- tained, it is diluted to correspond with the ferrocyanide solution, and again compared with the latter in the manner above described. With this iron solution, all liquids contain- ing ferrocyanides can be titrated with fairly accurate results. The liquids must, however, be neutral and free from alumina. Let it be supposed that 20 c.c. of the neutralised caustic soda extract (diluted previously to 500 c.c.) require 30 c.c. of ferric chloride solution, then as 1000 c.c. ferric chloride solution corresponds to 21*2 ferrocyanogen (Fe(CN) 6 ) the whole 21-2 x 10 extract contains 1000 = 5'3 grams = 2 x 5-3 20 = 10 '6 per cent, of ferrocyanogen. 194 THE GAS ENGINEER'S LABORATORY HANDBOOK. IX. ANALYSIS OF FIRE-CLAY AND FIRE-BRICKS. 131. Fire-clay is the name given to any clay which is capable of standing a high temperature without melting ; such clays are said to be refractory. A refractory fire-clay will contain nearly pure hydrated silicate of alumina. The more alumina that there is in proportion to the silica, the more infusible will be the clay. The composition of dif- ferent fire-clays necessarily varies however. They contain From 59 to 96 per cent, silica (Si0 2 ) 2 36 alumina (A1 2 3 ) 2 5 oxide of iron (Fe 2 O 3 ), and a very small percentage of lime, magnesia, potash and soda. The fire-resisting properties of the clay depend chiefly upon the relative proportions of these constituents. If the oxide of iron or alkalies are present in large proportions, they act as a flux, and cause fusion, the clay is then no longer refractory. The following is the method of analysis. A quantity of the substance (fire-clay or brick) is reduced to an impalp- able powder in an agate mortar and placed in a stoppered weighing tube. About 2 grams of the sample are dried in a platinum crucible or platinum dish at a temperature of 100 C. until the weight is constant ; the loss in weight gives the moisture. In the case of clay it is then ignited, at first gently, and afterwards strongly, and for a tolerably long time ; the loss in weight corresponds with the combined water, together with the organic and volatile constituents of the clay if such are present. 1'5 grams of the powdered sample are then weighed accurately out into a platinum crucible and about four times its weight of a fusion mixture added, consisting of sodium SPECIAL ANALYSES. 195 and potassium carbonates mixed in molecular proportions. The whole is intimately mixed by means of a smooth rounded i. lass rod. It will be found convenient to add the fusion nixture by small portions at a time, since in this way a nore thorough mixture is obtained. The mixture should nly half fill the crucible. The lid is then placed on the crucible, and the latter : ;ently heated over the Bunsen flame ; the temperature is gradually increased, care being taken that no loss ensues by he frothing due to the evolution of C0 2 . When the mass s fused, the crucible is transferred to the blow-pipe flame, md the whole is kept at a bright red heat until effervescence leases and the fused mass becomes tranquil. The flame is :hen removed, and the crucible is allowed to cool just below edness, when it is placed on a cold surface, such as a clean ;>lock of iron, so as to assist it to cool rapidly. When cold, :he crucible and its contents are placed in a deep evaporating lish or in a shallow beaker; this is covered with a clock- ^lass, and tolerably strong hydrochloric acid added to the 3on tents, which should be gently agitated after each addition }f the acid, and kept covered during the operation. When affervescence has ceased, and the crucible is free from all adherent solid, remove the crucible by means of the crucible tongs, carefully rinsing off any adhering liquid by means of the jet from the wash-bottle, into the main portion of the liquid. On treating the fused mass with HC1 as above described, most of the Si0 2 will separate out as a gelatinous mass. If any gritty particles are felt, on stirring the bottom of the vessel with a glass rod, the fusion is imperfect. This is generally due to the original substance not having been powdered sufficiently finely. In this event, it is usually more satisfactory to make a fresh fusion, taking care that no coarse particles are present in the portion of the sample for the fresh fusion. (a) Estimation of tlie Silica. The liquid containing the gelatinous silica is then transferred (if necessary) to an, 196 THE GAS ENGINEER'S LABORATORY HANDBOOK. evaporating basin, preferably of platinum, and evaporated to dryness upon a water-bath. When the contents of the basin become pasty, they should be continually stirred with a rounded glass rod to prevent the formation of lumps. When all the liquid has been driven off, the contents of the dish should be in the state of a fine powder. In order to expel the last trace of HC1, the dish should now be placed upon a sand-bath, and heated with a small Bunsen flame until no moisture is deposited on a cold clock-glass when placed upon the dish for a few seconds. The dish is then allowed to cool, and its contents are moistened with strong HC1. It is then heated on a water-bath for about half an hour, a small quantity of hydrochloric acid being occasionally added with stirring. Hot distilled water is now added, and the silica is filtered off and is washed free from dissolved chlorides. The precipitate is ignited apart from the filter (196), the precipi- tate being transferred to the platinum crucible cautiously, since it consists of a very light powder, which is easily blown away. The lid is placed on the crucible, and the latter heated, exceedingly gently at first, and the tempera- ture raised very gradually, or the escaping steam will carry some of the fine light powder away with it. The crucible is finally raised to a full red heat over the Bunsen flame. (fc) Estimation of A1 2 3 and Fe 2 3 . The filtrate from the Si0 2 is mixed with NH 4 C1 solution and then with NH 3 in slight excess, the hydrates of iron and aluminium, of which the precipitate consists, being separated as under.. The precipitate is washed, and dissolved upon the filter in hot dilute hydrochloric acid, and the solution allowed to flow into a nickel or porcelain dish containing about 50 c.c. of pure strong KHO solution. Wash out the acid which remains adhering to the filter-paper with a small quantity of distilled water, and allow these washings to also run into the dish, and boil the contents of the latter for a few minutes. The iron will be precipitated as ferric hydrate, while the hydrate of aluminium will remain in solution. SPECIAL ANALYSES. 197 The iron precipitate is filtered off, it is then again dis- solved in HC1, and reprecipitated by ammonium hydrate, in order to free the ferric hydrate from potash. It is then washed and ignited apart from the filter, at a red heat, and weighed as Fe 2 3 . The solution of the aluminium hydrate in the potassium hydrate solution is treated with a slight excess of strong hydrochloric acid, and then with a very slight excess of ammonium hydrate. It is then filtered off, washed, dried, ignited, and weighed as A1 2 3 . Another method of separation is, after weighing the mixed hydrates of Fe and Al, to dissolve them in KHS0 4 and about 5 c.c. H 2 S0 4 , add about 1 gram hyposulphite of soda, boil, and titrate the solution with yi^- normal bichromate of potash; this will give the amount of iron. (c) Estimation of the Calcium. If the volume of the filtrate from the iron and alumina precipitate is very large, evaporate it down to a convenient bulk, add a little ammonium hydrate if not already alkaline, and then a slight excess of ammonium oxalate. Allow the liquid to stand, filter off, ignite, and weigh the precipitate as oxide (28). (d) Estimation of the Magnesium. Evaporate the filtrate and washings from the calcium oxalate precipitate to dryness, ignite the residue, and treat it with a little strong HC1, add water and filter, if necessary. To the clear solution add ammonium hydrate in moderate excess, and then an excess of sodium hydrogen phosphate solution. Allow the liquid to stand for a few hours, or shake it vigorously in a stoppered bottle, filter off, wash the precipitate with dilute ammo- nium hydrate solution, then ignite it, and weigh the Mg as Mg 2 P 2 7 . (e) Estimation of the Alkali Metals. Since sodium and potassium carbonates have been employed in the fusion, the alkali metals cannot be estimated in the filtrate from the magnesium. A separate portion of the substance is accord- ingly used for their determination. Lawrence Smith's method for the determination of the 198 THE GAS ENGINEER'S LABORATORY HANDBOOK. alkali metals will be found the most convenient. The following is the mode of procedure. Weigh out accurately about 1 5 grams of the finely powdered substance into a platinum crucible, intimately mix this with a mixture of 1 *5 grams of pure recrystallised ammonium chloride, and 9 grams of pure calcium carbonate. Then heat the crucible to bright redness for an hour over a good Bunsen or blow-pipe flame, or preferably as follows. Place the platinum crucible in a clay crucible containing a little calcined magnesia or lime at the bottom and round the sides, and heat the clay crucible in a gas furnace, which is capable of maintaining it at a bright red heat. When the crucible has been heated for an hour, allow it to cool, and place the platinum crucible and its contents in hot water in a covered platinum or porcelain dish, and boil for a time. This procedure will dissolve out the alkaline chlorides together with some calcium hydrate. Filter and mix the filtrate with ammonium hydrate and ammonium carbonate solutions in excess, and with a few drops of ammonium oxalate solution. Allow the liquid to stand, filter into a platinum or porcelain dish, evaporate the filtrate to dry- ness, and heat the residue just below redness, but sufficiently strongly to drive off the ammoniacal compounds. Dissolve the residue in water ; add a few drops of AmHO and of ammonium oxalate solution to precipitate any trace of calcium still in solution ; filter, and evaporate the filtrate to redness, with a few drops of HC1, in a weighed dish. Gently ignite the residue and weigh, repeating the ignition until the weight is constant. The weight of the residue thus obtained gives the combined weight of the potassium and sodium chlorides. The residue is then dissolved in water and the potassium chloride is precipitated by platinic chloride in the following manner. To the solution of the residue a few drops of HC1 are added, then an excess of platinic chloride solution, the liquid being afterwards evaporated on the water-bath until a semi-solid crystalline mass is obtained. SPECIAL ANALYSES. 199 The platinic chloride is seen to be in excess by the super- is atant liquid being of an orange colour, after the liquid has been concentrated to a small bulk. When it is certain that there is an excess of platinic c hloride, we may then proceed according to a or b as below. (a) Pour alcohol upon the mass ; gently shake the liquid i ound in the dish so as to mix the contents of the same well ogether, allow the precipitate to settle completely and pour ' >ff the liquid through a tared filter. Repeat these operations wice, and finally transfer the undissolved double salt to the liter by means of a small wash-bottle filled with alcohol. Continue washing the precipitate upon the filter with alcohol intil the washings are no longer coloured. Dry the filter md its contents at 100 C., and weigh as 2KClPt01 4 . (fe) A rather quicker method of treating the precipitated louble salt is to wash it with alcohol by decantation until the alcohol is no longer coloured, the alcohol being decantated through an untared filter-paper. Care must be taken that as little as possible of the precipitate is poured off with the alcohol. The double salt, freed from the excess of PtCl 4 is aow washed into a platinum crucible, dried at 100C. and weighed. The filter, which will contain a little of the double salt, is then incinerated, and the ash is dropped into the crucible and weighed. By deducting from this weight the weight of the filter ash, the approximate weight of platinum left by ignition is found ; this is calculated into double salt and the weight is added to that of the double salt already found in the crucible. If the quantity of precipitate left on the filter is appreciable, the weight of KC1 left in the filter ash, not being allowed for, will introduce an error. The filter in this case should be ignited in a separate crucible, the KC1 washed out from the ash by hot water, and the dried residue weighed. The true weight of the platinum in the ash is thus ascertained, and is made use of as mentioned above. The weight of the sodium chloride in the mixed chlorides of potassium and sodium is then ascertained by difference. The chlorides are finally calculated into oxides. 200 THE GAS ENGINEER'S LABORATORY HANDBOOK. X. ASSAY OF COAL TAR. 132. The assay of coal tar is usually limited to a labora- tory operation, in which the various fractions are collected as nearly as possible under the same conditions as pertain to the large scale, though the details will necessarily vary with circumstances. With practice, good results are obtained with as small a quantity as 10 oz. of the tar, the yield corresponding closely with those given on a large scale ; but the chief value of such laboratory operations is for comparing different samples of tar. The following is the method of making the assay : 250 c.c., or 10 oz., measure of the tar to be examined should be placed in a glass retort, of such a capacity that the tar only fills about one-third of it, the object of this being to prevent the distillate being spoilt in the event of there being much frothing during the distillation. The retort may be con- veniently supported on a cup-shaped piece of wire gauze, placed in an aperture in a sheet-iron plate. Over the retort a dome is placed : this may be made by removing the bottom from a tin can or bottle, and cutting out a piece of the side to allow the neck of the retort to pass through. The object of this contrivance is to confine the heat, and prevent the distillate or heavy vapour from falling back : without some arrangement of this kind, a satisfactory assay of coal tar in a glass vessel is nearly impossible. The products obtained by the distillation are : (1) amrnoniacal liquor ; (2) total light oils ; (3) creosote oil ; (4) anthracene oils ; and (5) pitch. In obtaining the above fractions the character of the distillate is quite sufficient to indicate the point at which the receiver should be changed. It is not necessary to insert a thermometer, nor to connect the retort to any condensing arrangement. The lamp being lighted underneath the retort (a powerful SPECIAL ANALYSES. 201 ]>unsen should be used, as towards the close of the experiment it is necessary to maintain the wire gauze at a red heat), the i. inmoniacal liquor and naphtha should be collected together in a graduated cylinder, which should be changed when a i rop of the distillate collected in a test-tube of water begins 10 sink. In this fraction the ammoniacal liquor and the i otal " light oils " will be found : the expression " light oils " 3 aeans oils lighter than water. After standing for some time i o allow perfect separation of the ammoniacal liquor and light - ils, the volume of each should be observed, and, if deemed lecessary, the strength of the liquor ascertained in the usual vay by distillation with caustic soda, and titration of the i iistillate. The quantity of light oils obtainable will not be ; ufficient to allow of any further fractionation for benzols, &c. The next fraction of the distillate will consist of creosote < 'il. At the commencement it will contain much naphthalene, ; ,nd will very probably solidify in white crystals on cooling ; >ut later on a more fluid distillate will be obtained. At a t till later stage, a drop of the distillate collected on a cold h.teel spatula will be found to deposit amorphous solid matter, of a yellow, or greenish-yellow colour. When this action v,akes place, the receiver should again be changed, and the fraction obtained measured. If required, this fraction, which :s called "carbolic and creosote oils," may be assayed for carbolic acid and naphthalene, as described later on. With modern London tars it is generally found that this fraction is semi-solid ; it will therefore be necessary to measure it while quite warm. The next fraction of the distillate will be found rich in anthracene, and very often condenses in the neck of the retort as a yellow, waxy substance, which may, however, be melted out by the local application of a small Bunsen flame. The collection of the anthracene oil is known to be complete when 110 more distillate can be obtained, and the pitch remaining in the retort inturnesces and gives off heavy, yellow fumes. 202 THE GAS ENGINEER'S LABORATORY HANDBOOK. The distilled fraction containing the anthracene oil should then be measured and cooled thoroughly, and the resultant pasty mass pressed between folds of blotting paper, weighed, and assayed for real anthracene by the anthraquinone test. The result should be calculated into crude anthracene of 30 per cent., a standard which is generally adopted by manufacturers. When the distillation for anthracene oil is complete, the retort should be allowed to cool, and when almost cold, the body of it should be immersed in cold water. This opera- tion produces a rapid surface-cooling and shrinking of the pitch from the glass, which may then be broken away and removed by gentle tapping, leaving the cake of pitch clean, and in a fit state for weighing. The following figures by Mr. B. Nickels, F.I.C., show the result obtained by the assay of four representative samples of London tar. The apparent excess is due to the fact that the samples of tar experimented on were measured into the retort, while the resultant pitch was weighed : A. B. c. n. Ammoniacal liquor .. 2-5 3*7 8-0 5'0 Total light oils .. .. 2-5 3'4 0'5 3'2 Carbolic and creosote oik 21-3 17 '0 23-0 20-0 Solid. Solid. Solid. Solid. Anthracene oils .. .. 17'0 17'0 13-0 13-0 Semi-solid. Semi-solid. Semi-solid. Semi- solid. Pitch (grams per 100 c.c.) 59'4 60-0 58-0 62-0 102-6 101-1 102-2 103-2 Pressed anthracene.. .. 4-0 .,. 1*5 Containing real anthra- \ 13 . 4 per cent 2 5 68 per cent. cene = 30 per cent crude an- } g ^ 1>3 ia2 thracene m tar / The following table shows the mean composition of the SPECIAL ANALYSES. 203 fo -egoing samples in juxtaposition with the general yield of E iglish country tar : London. Country. Ammoniacal water 4*5 per cent. 4 per cent. Total light oils 2-4 3 Carbolic and creosote oils .. 20-3 22 Anthracene oils 15' 4 Pitch (grams per 100 c.c.) .. 59 & 67 102-1 100 In the above assay the total light oils were collected fr gether, but in practice, on the large scale, the distillate is g 3nerally divided into two fractions, which are known as " first light oils " and " second light oils." When they are n 3t collected separately, the fraction is known as " crude u aphtha " or " light oil." The crude naphtha is generally re-distilled without any I revious chemical treatment, the resultant distillate being known as " once run naphtha " and " last runnings." On the large scale, once run naphtha, before being re- distilled, is purified by treating it with sulphuric acid of sp. gr. 1-845. By this means the hydrocarbons of the ethylene and crotonylene series are removed, as well as some of the higher homologues of benzene. A further treatment with caustic soda eliminates phenols and other bodies having an acid character. The oil is afterwards washed with water and again distilled. XI. FRACTIONAL DISTILLATION. Rectification of Benzene. 133. It is often possible to separate almost completely, by a simple distillation, two liquids occurring together in a mixture, when their boiling points lie widely apart. The more volatile liquid first passes over, the temperature sud- denly rises, and the higher-boiling liquid distils. It is otherwise when we have a liquid consisting of a mixture of 204 THE GAS ENGINEER'S LABORATORY HANDBOOK. bodies boiling very near each other, especially in the case of homologous compounds such as occur in petroleum and coal- tar naphtha; one distillation only effects a very imperfect separation; a portion of the less volatile liquid is carried over by the vapour of the more volatile substance, the temperature rising throughout the distillation. In order to effect separation of the several substances in a case of this kind, recourse is had to the method of fractional distillation. It will be evident that if the vapours arising from a mixture of boiling liquids are somewhat cooled before reaching the condenser proper, the less volatile portion, carried upwards by the vapour of the more volatile liquid, will be partially condensed. If this is permitted to flow back into the retort, only the lower boiling portion passes into the condenser, and the separation is consequently more rapid and effective. In order to carry out this partial condensation during the process of distillation, several forms of apparatus have been devised. A very cheap and effective apparatus is that of Hempel, which consists of a long glass tube containing glass beads, or pieces of broken glass, Fig. 48. Fig. 49 shows the apparatus of Le Bel and Henninger for the same purpose ; this form of apparatus has side tubes down which the con- densed liquid flows. At the narrow parts of the tube a, 6, c, d, are fixed small cups of wire gauze. Little pools of condensed liquid form FIG 48 * n t ^ ese CU P 8 > an( i *hi s liquid, may be said to wash the vapour passing upwards; in fact, a process of fractionating is carried on in these cups by the ascending vapours. In carrying out a fractional distillation, the apparatus is arranged as in Fig. 50. The flask is heated over wire gauze, or, in the case of a very volatile liquid, in a water-bath. If wire gauze is used, the burner should be placed in a I SPECIAL ANALYSES. 205 deep tin basin containing sand, in order to absorb the liquid in the event of the flask or retort cracking. The frac- tionating bulb or tube is fitted with a thermo- meter, the bulb of which is well below the e> it tube. A number of clean dry flasks or b( ttles, fitted with corks and having labels a1 tached, are required for recording the boiling pi >int of the fraction. We will show the method of applying this p inciple in the rectification of a sample of b ;nzene. 200 c.c. of the sample of benzene are in- ti oduced into an 8-oz. flask ; a spiral of plati- n im wire or two or three pieces of broken pipe s* ems are added, in order that the liquid may b )il without bumping. The flask is connected t( the condenser by means of the fractionating FIG. 49. FIG. 50 apparatus, and heat so applied to the flask that the dis- tillate comes over into the receiver in drops; it musst not be allowed to come over in a continuous stream. The 206 THE GAS ENGINEER'S LABORATORY HANDBOOK. indications of the thermometer must be carefully watched and noted, and the distillate between every five degrees collected in separate flasks. These different fractions are then redistilled in order, adding the next to the residue of the previous one in the distilling flask. The portions boiling below 85 C. and above 105 being collected between every two or three degrees. The fractions below 85 and above 105 are re- distilled as before and collected between every two degrees. A. 7i- 5-5 c. B. 85-90. 90-95. D. 95-100. E. 100-105. F. 105-110. G. 110-115. Residue. 19 c.c. 53 c.c. 26 c.c. 15 c.c. 13 c.c. 17 c.c. 21 c.c. 33 c.c. A'. below 7 u. V. 79-81. C'. 81-85. D'. 85-105 E'. 105-108. F'. 108-110. Residue. A 5 c.c. Added B 42 c.c. (10 c.c.*) Added C (9 c.c.*) Added DE 50 c.c. Added F (11 C.C.*) Added G 22 c.c. 42c.c. * Refrac- tioned C' 12 c.c. 7 c.c. E' 6 c.c. 5 c.c, 5 c.c. 54 c.c. 7 c.c. 50 c.c. 6 c.c. 27 c.c. , 42 c.c. The fraction 79-81 C., which is nearly pure benzene, can be further purified by freezing, pressing, and redistilling. SPECIAL ANALYSES. 207 It will be found that by repeating the process the liquid is gradually separated into two large, fractions, consisting chiefly of benzene and toluene, and a number of smaller intermediate fractions. The table by Dr. Cohen on page 206 gives the volume in c.< . and the boiling points of the fractions obtained by this ni 3thod from 200 c.c. 50 per cent, benzene, each table d( noting a complete series of fractionations. XII. DETERMINATION OF THE SPECIFIC GRAVITY OF GAS. 134. This may be most conveniently performed by a n ethod devised by Mr. Greville Williams, F.K.S., a descrip- ti 3n of which appeared in the Transactions of the Gas Institute f< r 1882. The following is taken from the above mentioned s< urce : "The specific gravity of a gas may be defined as the c< >inparative weights of equal volumes of air and the gas uider the same conditions as to temperature and pressure. Vhether the gas should be dried or not before its density I)-) taken is a matter of no importance as regards the present piper. I, however, invariably weigh the gas and air in a i absolutely dry state. It is quite unnecessary for me to describe the' processes devised by Eegnault, Bunsen, and others, as I have already done so very fully in Watt's ' Dictionary of Chemistry,' and in my 'Handbook of Chemical Manipulation.' . . . The sole object of this paper IB to show that, by a simple modification of well-known processes, and without the use of an air pump, an accuracy may be obtained which leaves nothing to be desired for technical purposes ; and the formula given for the calcula- tions, which is of the utmost possible simplicity, does not introduce an error of any importance. " If, however, perfect accuracy be desired, the corrections mentioned further on can easily be made. As with the 208 THE GAS ENGINEER'S LABORATORY HANDBOOK. method described by Banister,* only a globe with two stop- cocks is used, which is represented in Fig. 51, one-quarter the actual size. Instead, however, of weighing the balloon, as Banister directs, against weights, I weigh against another balloon of as near as possible the same size, but a little lighter a device very generally adopted in the determina- tion of specific gravities of gases and vapours. " The determination of the specific gravity of a gas is by no means a difficult operation ; but, nevertheless, to obtain accurate results, attention must be paid to the minutest precautions, the neglect of any one being fatal to success. In the first place, the balance should be of the highest class; both planes and edges must be of agate or rock crystal ; and it should distinctly indicate one deci-milli- gramme with the globe and its counterpoise hanging from the beam. If this be the case, the capacity of the globe need not be greater than 400 cubic centimetres. " The usual methods employed involve exhaustion of the balloon by the air pump, either at each experiment, or, as with Letheby's, a preliminary exhaustion, which serves once for all. By the method here described, not only is no exhaustion required, but all corrections are avoided by the simple device of selecting a day when the barometer is steady, and working in a room where the temperature can be made the same when the globe is weighed with air as when it is weighed with gas. Both these conditions, I find, can be readily obtained in practice. " The room in which I worked formed the centre one of a series of three, and had no doors ; notwithstanding this, by means of a gas stove, I found it easy to regulate the temperature to 0*2 of a degree centigrade. If the room be too cold, the heat may be raised ; if it be too hot, cooler air may be introduced by the ordinary means employed for ventilation. By working in this way, it will be seen that * ' Gas Manipulation,' Banister and Sugg, pp. 18, 19, 20. SPECIAL ANALYSES. 209 th 3 air and the gas may be weighed under the same cou- di ions of temperature and pressure. It is proper, however, to remark that the air must be rendered absolutely free from ca i-bonic anhydride and moisture. For this purpose it was di iwn slowly through one bottle of potassium hydrate so ution, two of sulphuric acid, six U-tubes of very active so la-lime, and four U-tubes of calcic chloride, before entering tl 9 globe. One U-tube of calcic chloride was placed, as a g' ard, after the globe. " The drawing through of the air is effected by means of di aspirator, and the current is kept up until every trace ol any gas previously experimented on is removed ; this FIG. 51. FIG. 52. m-iy be known by the balloon remaining constant in weight. \\ hen this is arrived at, and the temperature at which it is desired to make the experiment attained (and which should be as near as possible the average of the room), the tube attached to the aspirator is to be slipped off, and the tap of the globe on that side closed, and then the other tap. The balloon, having been carefully wiped with a clean chamois leather, or a soft silk handkerchief kept for that purpose, is to be hung on the balance by a stirrup of platinum wire, as represented in Fig. 51. The counterpoise, Fig. 52, is also to be hung on the balance by the glass hook. The counterpoise being somewhat lighter than the balloon, a very few grammes will be required to bring them into equi- librium. 210 THE GAS ENGINEER'S LABORATORY HANDBOOK. " Having adjusted the last deci-milligramme, by means of the rider apparatus, the balloon is to be left for five minutes, and the weight taken again; the last weight is to be called 'weight of balloon and air.' The balloon is then to be attached to a tube evolving the gas, which is to be passed slowly through six U- tubes of soda lime, to remove the last traces of carbonic acid, and four of calcic chloride, for one hour ; a guard tube, as before, being employed at the exit. The tap next the gas supply is to be turned off first, the other immediately after ; the balloon is then to be weighed with the same precautions as were taken with air. It is as well, in a first experiment, to send the gas a second time through the globe for another half hour ; when, if it be found that the globe has not diminished in weight, it is, of course, certain that all the air has been expelled. The last weight is to be entered in the laboratory book as * weight of balloon and gas.' " From this time a specific gravity may be determined every hour without reweighing the balloon and air, as long as the barometer and thermometer remain the same. The weights of the balloon and air at various barometric pres- sures and temperatures should be tabulated, as when that pressure and temperature occur again, the previous deter- mination of weight of air will render another one unneces- sary. " It is obvious that the temperature of the room in winter will be different to that in summer, but this will have no effect upon the result, as all that is necessary is that the temperature should be the same when both gas and air are weighed. " Care must be taken that the calcic chloride and soda- lime are renewed at proper intervals. " In order to subject the process to the severest possible test, I made some determinations of the specific gravity of hydrogen in the globe, Fig. 51, having a capacity of 443 cubic centimetres. The specific gravity of hydrogen SPECIAL ANALYSES. 211 be ng only 0*0693, it will be seen that an error of only 0*01 is very conspicuous. In four experiments, the following va lues were obtained : 1 K>. p. Grammes. V. c.c. T. c. n t D. Theory. Difference. r. 0-5038 443 Deg. 16 0-001221 0-0685 0-06930 -0-0008 n. 0-5045 443 14 0-001230 0-07412 0-06930 +0-00482 ui. 0-5020 443 16 0-001221 0-07192 0-06930 +0-00262 IT. 0-5040 443 16 0-001221 0-0682 0-06930 -0-0011 Mean. 0-0707 Theory. 0-0693 Difference. + 0-0014 " If all the corrections are made, the mean density be- co nes a little higher, but even then the error is only 0*0126, wJdch is well within the limits laid down at starting. Bi nsen's standard experiments with his apparatus gave 0- )79; difference 0*010. " In the above table, P is the difference between the w( ight of the balloon and air, and the balloon and gas, in gr immes. V is the capacity of the balloon in cubic centi- metres ; T the temperature, centigrade ; n t the weight of one cubic centimetre of air at T ; D the experimental specific gravity. As care has been taken that the barometer shall no: vary during the experiment, it does not enter into the formula, which for gases lighter than air is " The value of n t may, when extreme exactitude is not required, be taken direct from the following table ; other- wise n t should be corrected for the elastic force of the atmosphere at the time of the experiment, as the table is calculated for a barometric pressure of 760 mm., or 29*922 inches. p 2 210 THE GAS ENGINEER'S LABORATORY HANDBOOK. " Having adjusted the last deci-milligramme, by means of the rider apparatus, the balloon is to be left for five minutes, and the weight taken again; the last weight is to be called 'weight of balloon and air.' The balloon is then to be attached to a tube evolving the gas, which is to be passed slowly through six U- tubes of soda lime, to remove the last traces of carbonic acid, and four of calcic chloride, for one hour ; a guard tube, as before, being employed at the exit. The tap next the gas supply is to be turned off first, the other immediately after ; the balloon is then to be weighed with the same precautions as were taken with air. It is as well, in a first experiment, to send the gas a second time through the globe for another half hour ; when, if it be found that the globe has not diminished in weight, it is, of course, certain that all the air has been expelled. The last weight is to be entered in the laboratory book as * weight of balloon and gas.' " From this time a specific gravity may be determined every hour without reweighing the balloon and air, as long as the barometer and thermometer remain the same. The weights of the balloon and air at various barometric pres- sures and temperatures should be tabulated, as when that pressure and temperature occur again, the previous deter- mination of weight of air will render another one unneces- sary. " It is obvious that the temperature of the room in winter will be different to that in summer, but this will have no effect upon the result, as all that is necessary is that the temperature should be the same when both gas and air are weighed. 44 Care must be taken that the calcic chloride and soda- lime are renewed at proper intervals. *' In order to subject the process to the severest possible test, I made some determinations of the specific gravity of hydrogen in the globe, Fig. 51, having a capacity of 443 cubic centimetres. The specific gravity of hydrogen SPECIAL ANALYSES. 211 be ng only 0*0693, it will be seen that an error of only 0*01 is very conspicuous. In four experiments, the following va lues were obtained : 1 IS, P. Grammes. V. c.c. T. c. n t D. Theory. Difference. E, 0-5038 443 16" 0-001221 0-0685 0-06930 -0-0008 1L 0-5045 443 14 0-001230 0-07412 0-06930 + 0-00482 HE. 0-5020 443 16 0-001221 0-07192 0-06930 +0-00262 IV, 0-5040 443 16 0-001221 0-0682 0-06930 -0-0011 Mean. 0-0707 Theory. 0-0693 Difference. + 0-0014 "If all the corrections are made, the mean density be- co nes a little higher, but even then the error is only 0*0126, which is well within the limits laid down at starting. BT nsen's standard experiments with his apparatus gave 0- )79; difference 0-010. " In the above table, P is the difference between the w( ight of the balloon and air, and the balloon and gas, in gr immes. V is the capacity of the balloon in cubic centi- me -tres ; T the temperature, centigrade ; n t the weight of on 3 cubic centimetre of air at T ; D the experimental specific gravity. As care has been taken that the barometer shall no; vary during the experiment, it does not enter into the foimula, which for gases lighter than air is " The value of n t may, when extreme exactitude is not required, be taken direct from the following table ; other- wise n t should be corrected for the elastic force of the atmosphere at the time of the experiment, as the table is calculated for a barometric pressure of 760 mm., or 29-923 inches. p 2 212 THE GAS ENGINEER'S LABORATORY HANDBOOK. TABLE FOR THE CALCULATION OF Temp. n t Temp. n t Temp. n t Grammes. Grammes. Grammes. C. F. C. F. C. F. Deg. Deg. I*g. Deg. Deg. l>eg. 1 33-8 0-001288 11 51-8 0-001243 21 69-8 0-001201 2 35-6 0-001284 12 53-6 0-001239 22 71-6 0-001197 3 37-4 0-001279 13 55-4 0-001234 23 73-4 0-001193 4 39-2 0-001275 14 57-2 0-001230 24 75-2 0-001189 5 41-0 001270 15 59-0 0-001225 25 77-0 0-001185 6 42-8 0-001266 16 60-8 0-001221 26 78-8 0-001181 7 44-6 0-001261 17 62-6 0-001217 27 80-6 0-001177 8 46-4 0-001257 18 64-4 0-001213 28 82-4 0-001173 9 48-2 0-001252 19 66-2 0-001209 29 84-2 0-001169 10 50-0 0-001248 20 68-0 0-001205 30 86-0 0-001165 " It is as well to repeat that the flow of a light gas should "be continued until the balloon ceases to lose weight ; and if the gas be heavier than air, until it ceases to gain in weight. " The following numbers were obtained in the determina- tion of the density of carbonic anhydride : y T No. Grammes. C.C. C "fc D. Theory. Difference. I. 0-2833 443 17 0-001217 1-5254 1-5224 + 0-0030 II. 0-2855 443 17 0-001217 1-5295 1 5224 + 0-0071 Mean. 1-5274 Theory. 1-5224 Difference. + 0-0050 " The gas being heavier than air, the formula becomes D = Y ^ + P . " In conclusion, it may be said that, if the method above given be carefully carried out, the specific gravity of coal or any other gas, of which a cubic foot or two can be spared, may be readily and accurately determined, without the use of an air-pump." 213 PAET V. TECHNICAL GAS ANALYSIS. Introduction. T HE measurement of gases, and the quantitative analysis of g iseous mixtures, involve special methods and constitute a S] ecial branch of chemical analysis, frequently involving the u e of special apparatus. Further, when dealing with the volume of a gas, special corrections and calculations are fi ^quently required in order to arrive at comparative results. These corrections are necessitated by the very considerable el feet which changes of temperature and pressure, and ai nount of moisture, exert upon the volume of a gas. The most exact methods of gas analysis necessitate the ^je of mercury as the confining liquid, and also require a vory considerable amount of manipulative skill on the part of the operator. They further involve the expenditure of a considerable amount of time, and in addition, require very expensive apparatus; consequently such methods are not adapted to the every-day requirements of a gasworks. For such requirements, something involving less time and care, even at the sacrifice of a slight degree of accuracy, is neces- sary. These conditions are complied with in a number of designs for apparatus for Technical Gas Analysis lately devised; amongst them, those of Hempel and Bunte aro specially adapted for use in gasworks. 214 THE GAS ENGINEER'S LABORATORY HANDBOOK. TJie Measurement of Gases. 135. The volume of a gas can be found directly or in- directly ; it is estimated either volurnetrically, by titration, or gravimetrically. The quantity found is in all cases expressed in per cent, by volume. The volume of a gas is influenced by pressure, temperature, and the tension of the liquid over which it is measured. Gases are measured in their condition at the time at which the measurement is made, that is at the pressure of the air as indicated by the barometer, and at the tempera- ture as indicated by the thermometer. As water, in the case of technical gas analysis, is used as the trapping liquid, the gas is always in a state of complete saturation with moisture. From the above remarks it will, be seen that the condi- tions under which gases are measured may be very different, and may vary during the progress of an analysis even from one observation to another. Every such change, unless duly taken notice of, might introduce very considerable errors. It is consequently necessary, in many cases, to make a correc- tion, which consists in reducing the volume of gas, which is measured under known but varying conditions, to the volume which it would measure at the normal barome- tric pressure of 760 ' millimetres, at the normal tempera- ture of C., and in the dry state. Gases, when reduced to these standard conditions, are said to be in their normal state. 136. The reduction of the volume of a gas to the normal state is made by the aid of a formula based on the following data. (1) Pressure. According to Boyle's law the volume of a gas is in inverse proportion to the pressure to which it is bubjected. TECHNICAL GAS ANALYSIS. 215 Consequently, if v = the observed volume of the gas, Vp = the volume of the gas at 760 mm. pressure, p the observed barometric pressure, (2) Temperature. A gas expands by one-273rd = 0' 003665) of its volume, for each degree centigrade rise n its temperature. Therefore, if v = the observed volume of the gas, V = its volume when reduced to C., t = the observed temperature, , 273 X v __ ^_ v ~ 273 -f-$ ~1 -f-0-003665 x t' (3) Correction for Moisture. When a gas is saturated with moisture by contact with water, under the same conditions of temperature and pressure, it always contains the same quantity of moisture. This water vapour mingled with the gas exerts a certain pressure, which is commonly termed the tension of aqueous vapour, and increases with the tempera- ture. It is usually expressed in millimetres of mercury, and has been accurately determined for each degree of tempera- ture (see Table in Appendix). It is necessary to deduct the pressure due to the tension of aqueous vapour from the baro- metric pressure. Thus if w is the tension of aqueous vapour in a volume of gas under pressure p, the true pressure on the gas = (p to). The three preceding corrections are embraced in the following formulas, in which V represents the volume of the gas corrected for pressure and temperature, and for moisture if necessary. For a dry gas V = 760 (273~+~0 = 760(1 + 0-00365*) * 216 THE GAS ENGINEER'S LABORATORY HANDBOOK. __ ( p - w) x 273 x v For a moist gas V = - 760 (27 3 + *). w X ~ 760(1 +0036650' 137. Only those volumes can be directly compared with one another, that have been reduced to equal pressure and temperature, the tension of the confining liquid being also allowed for. Parallel gas measurements can be carried on as under (1) Varying pressure, varying temperature, and varying volume. (2) Constant pressure, constant temperature, and varying volume. (3) Constant temperature, varying pressure, and constant volume. (4) Constant pressure, varying temperature, and vary- ing volume. In the first case, the gas volumes found must be reduced to like temperature and pressure. In the second and third, the resulting volumes can be directly compared since density and pressure are directly proportional. The reductions of the volumes of gases to the normal state, may be omitted in analyses which are performed rapidly, as material changes of pressure and temperature are not likely to be then experienced ; also in cases in which only approximately correct results are required. Calibration of Measuring Tubes. 138. The graduations of measuring tubes for gases represent either absolute cubic centimetres, and fractions of cubic centimetres capacity, or millimetres length on the tube. In the latter case, it will be necessary to determine the value of the graduations in terms of cubic centimetres. TECHNICAL GAS ANALYSIS. 217 Since glass tubes always vary in diameter in different parts throughout their length, it follows that equal lengths on different parts of a graduated tube do not correspond to equal capacities; it is necessary, therefore, to calibrate a measuring tube throughout its length. The two liquids commonly used for calibrating tubes are mercury and water. Calibration of the Hempel Burette. 139. In this form of tube the gas is measured over water, consequently it is advisable to calibrate it by means of that liquid. The measuring tube, Fig. 53, is disconnected from the pres- sure tube, and has then attached to its lower end, by means of a piece of india-rubber tubing, a glass tube about four inches long, with the end drawn out into a fine jet, and having a stop-cock in the centre. The stop-cock and tube must be firmly attached to the measuring tube by tightly binding the india-rubber tubing with thin copper wire, the glass ends butting against one another within the joint. The jet should be bent in a downward direction, so as to deliver the water drawn off, into a vessel placed beneath it. The measuring tube and stop-cock tube are completely filled with distilled water at the temperature of the room, which should be care- fully noted. All air-bubbles must be excluded. Successive portions of water, of 5 c.c. each, are then drawn off into a stoppered weighing bottle and weighed, the weight of water corresponding to each reading on the tube, being noted, and the bottle carefully dried between each weighing. The weight of water is then calculated into ab- solute volume, by applying the correction for the expansion of water given in the Table in the Appendix. For example, if the first 5 c.c. weighed 4 '92 grams at 17 C., its absolute volume would be 4-92 x 1' 00101 = 4-925 c.c. When the top of the tube is being calibrated, it is FIG. 53. TECHNICAL GAS ANALYSIS. 219 advisable to draw off smaller volumes of water than 5 c.c., if the measuring tube is to be used for small volumes of gas. In all measurements in which water and similar liquids are employed, it is necessary that a certain time elapse for the liquid to drain down the sides of the measuring-tube before taking a reading. The time will vary from about 30 seconds to five minutes according to the liquid employed. On an average, however, about two minutes will be found sufficient. GAS ANALYSIS BY MEANS OF HEMPEL'S GAS APPARATUS. 140. The absorption of a gaseous constituent is frequently not carried out in the measuring tube itself, but in a separate vessel which serves for holding the absorbing liquid, and for bringing the gas into contact with it after being measured. When the absorption has been completed, the remaining gas is again carried over into the measuring tube, and its volume is read off. The volume of the gas absorbed follows from the difference of the two readings. This process admits of thoroughly utilising the absorbent, and dispenses with washing out the measuring tube after each estimation. Hempel's apparatus is based upon the above principle, and consists of two portions ; the burette, or measuring vessel, and the pipette or absorption vessel. 141. The gas burette is shown in Fig. 53. It consists of two glass tubes, a and fr, which are cemented into slots in semicircular weighted wooden feet, and are con- nected by about a yard of thin india-rubber tubing. To facilitate the cleaning of the burette, this india-rubber tube is divided in the centre, and the two ends are joined by a piece of glass tubing. Inside the feet, the tubes a and b are bent at right angles, and conically drawn out. The project- ing ends are of about 4 millimetres external diameter, and are corrugated, to enable the india-rubber connecting tube to be tightly attached by means of thin copper wire. The 220 THE GAS ENGINEER'S LABORATORY HANDBOOK. measuring tube fc, terminates at the top in a thick- walled capillary tube c, of from i to 1 millimetre internal diameter, and about 3 centimetres long. Over this a short piece of new black india-rubber tubing is firmly fastened on by means of thin copper wire. The india-rubber tube is closed by means of a strong brass pinch-cock which is placed close to the end of the capillary tube. The pinch- cock is always taken off from the india-rubber tube after using, as this helps to keep the latter in good condition. Notwithstanding the fact that readings cannot be made under the rubber tube, and that the pinch-cock cannot always be put on above ithe tube in ex- actly the same position, no practical error arises, since the glass tube c is so small in diameter. Hempel found that the differences in volume are much less than y 1 ^ c.c., a varia- tion which, in determinations not made over mercury, may be entirely disregarded. The graduated measuring tube 6, contains 100 c.c., the lowest mark being slightly above the wooden foot. The cubic centimetres are divided into fifths, and the numbers run both up and down. The tube a, which may be called the level or pressure tube, is not graduated, and is somewhat widened at the upper end A, to facilitate the pouring in of liquids. Manipulation of the Gas Burette. 142. Fill the tubes a and b with water, taking care to drive all air out of the connecting rubber tube by suitably raising or lowering the tubes ; then join the measuring tube to the gas supply by means of a glass or rubber tube con- taining water (this connecting tube can be easily filled with water by raising the pressure tube). To fill the burette with the gas to be examined, grasp the tube a in the left hand, close the rubber tube at e by pressing it between the little finger and the palm of the hand, and pour out the water in a. Place the pressure tube on the floor, and open the pinch- TECHNICAL GAS ANALYSIS. 221 cock/. The water will now flow into the pressure tube and the gas will be drawn into the measuring tube. When the latter is filled with the gas, close the pinch-cock/, disconnect 1) from, the gas supply, and, after the liquid has run down the walls of the burette, take up the tubes by their wooden feet, and by raising or lowering, bring the water in the tubes to the same level. The gas is now under the pressure of the atmosphere, and its volume is read off. To measure off exactly 100 c.c., bring rather more than 100 c.c. of the gas into the burette, close the latter with its pinch-cock, and let the water run down. Now compress the gas to less than 100 c.c. by raising the pressure tube, close the rubber tube at g with the thumb and first finger of the left hand, set the pressure tube on the table, and raising the burette in the right hand to the level of the eye, carefully open the rubber tube, and let the water run back until the meniscus stands at the 100 c.c. mark. Keeping the rubber tube still com- pressed, open the pinch-cock for a moment. The excess of gas will escape, and there remains in the burette exactly 100 c.c. of gas under atmospheric pressure. 143. The next portion of Hempel's apparatus is the absorption pipette, which is used in connection with the burette or measuring tube. By using a series of these pipettes, a gaseous mixture can be submitted to the succes- sive action of suitable absorbent reagents, and the pro- portions of the constituent gases thus ascertained. The " simple absorption pipette " is shown in Fig. 54. It consists of two large bulbs c and d, joined together as shown, and of a thick-walled glass tube a, bent as in the figure, called the capillary tube. The bulb c holds about 150 c.c., and the bulb d about 100 c.c., so that when 100 c.c. of gas is brought into c sufficient space for the absorbing liquid will remain. To protect the pipette from being broken and to facilitate manipulation it is screwed on to a wooden stand. A short piece of india-rubber tubing is fastened by means of thin copper wire to the free end of the capillary tube. 224 THE GAS ENGINEER'S LABORATORY HANDBOOK. Manipulation cf the Absorption Pipette. 147. To analyse a gas -with, the apparatus described, the burette is filled with distilled water which has been pre- viously saturated with the gas to be analysed. If simple pipettes are used, these are so filled with the absorbent that the bulb d remains empty. The absorbent FIG. 56. also must be saturated, by shaking with the gases which are but slightly soluble in it. The saturation of liquids is most conveniently effected in a flask half filled with the liquid, a rapid stream of gas being led through the liquid, and the flask vigorously shaken. If the reagents in the absorption pipettes are at the temperature of the room, as can easily be ascertained by placing a thermometer at k, Fig. 56, the analysis is started by aspirating the gas into the measuring tube as already described. It will be found convenient to take exactly 100 c.c. so that the results may be read off directly in percentages. TECHNICAL GAS ANALYSIS. 225 The apparatus is then arranged as in Fig. 57. The pipette is placed on the wooden stand Gr, and is connected with the burette by the capillary tube F, which is a piece of FIG, 57. 226 THE GAS ENGINEER'S LABORATORY HANDBOOK. thermometer tubing of about 0'5 millimetre internal dia- meter. To avoid the enclosing of air bubbles, the rubber tube d is first filled with water by means of a capillary funnel, and the capillary F is then introduced; F is thus completely filled with water. The rubber tube i of the pipette is squeezed between the thumb and the first finger of the right hand, and while thus compressed and free from air, the capillary connecting tube is inserted. Upon raising the pressure tube a, and opening the pinch- cock, the gas passes through the connecting tube into the absorption pipette. Any small air bubbles which may have been enclosed when F was inserted into ', are, at the begin- ning, separated from the gas by the water in F. If these bubbles do not take up more than 5 to 10 mm. space in the capillary tube of the pipette they may be neglected, since the error arising therefrom is only about 03 cubic centimetre. If the bubbles are larger, although with a little dexterity this may always be avoided, the gas is brought back into the burette by lowering the pressure tube and the operation is repeated. When the gas has passed over into the pipette, about % c.c. of water is allowed to follow, this water serving to rinse the capillary tube, and to free it sufficiently from the absorbing liquid which it previously contained. The gas is now enclosed between two columns of liquid, the absorbent on the one side, and the water in the capillary on the other. The burette having been closed by the pinch-cock, the pipette is disconnected and shaken, and the absorption of the gas thus effected. The burette and pipette are then reconnected, the pres- sure tube is placed on the floor, and the gas is brought back into the burette, care being taken that the absorbing liquid does not pass farther than the connecting capillary F. The pinch-cock is closed, the pipette removed, and the reading of the remaining volume is made as before described. The manipulation of the pipettes filled with solid absorbents is simpler still, for in this case no shaking is TECHNICAL GAS ANALYSIS. 227 r ecessary, because of the large amount of surface contact, I etween the solid and the gas. 148. A separate pipette is used for each absorbent. After rsing, the pipettes are closed at i with a piece of glass rod, ond at k with a small cork. 149. In order to obtain good results with the above ;i pparatus, it is necessary to watch that the apparatus and r .^agents do not change in temperature during the progress f f an analysis. A rise of temperature of 1 would cause an c rror of 0*3 per cent, in a total volume of 100 c.c. It is also of importance that the confining liquids be ; llowed to flow down from the walls of the burette in exactly t le same manner after each absorption, otherwise an error i lay be caused by the adhesion of more or less liquid to the <_lass of the instrument. In the case of gases which are < :>nfined over water, readings which are made one minute after the gas has been shaken with water in the burette, ( iffer by several tenths of a c.c. from readings made five ninnies later. Distilled water will run down completely i i five minutes. 150. The method of analysing the more common gases vhich have to be dealt with in gasworks by means of Hempel's apparatus, will now be described. The following is a list of the same : Oxygen, hydrogen, carbonic oxide, carbonic anhydride, marsh gas, defiant gas. Oxygen. 151. This gas is but slightly soluble in water, and is determined either by combustion with an excess of hydrogen, or by absorption. In the combustion with hydrogen, f of the volume turned consist of hydrogen, and i of oxygen. The volume of oxygen present is therefore found by dividing by 3 the decrease in volume resulting from the combustion. When oxygen is mixed with combustible gases, it is necessary to determine it by absorption. Q 2 228 THE GAS ENGINEER'S LABORATORY HANDBOOK. The best absorbents for oxygen are a strongly alkaline (potash) solution of pyrogallic acid, or phosphorus. 152. The solution of potassium pyrogallate is made by mixing together, either directly, or in the absorption pipette, 5 grams of pyrogallic acid dissolved in 15 c.c. of water, and 120 grams of KHO dissolved in 80 c.c. of water. It is necessary to note that KHO purified with alcohol should not be used, since this preparation even after strong ignition may cause erroneous results in the analysis. The absorptions should not be carried on at a temperature under 15 C., for it has been observed that the potassium pyrogallate used for absorption is very much less active at a temperature under 7 C. According to Hempel, at a temperature of 15 C., or higher, the last trace of oxygen can be removed with certainty in the space of three minutes, by shaking with the solution of potas- sium pyrogallate, while at lower temperatures the absorption is not complete after six minutes. 153. In order to prepare phosphorus in the necessary stick form for use in oxygen determinations, it is melted under water, in a test-tube placed in a water-bath. Enough phosphorus is used to form a column about six cm. high. A slightly conical glass tube of two to three mm. internal diameter is then dipped into the molten phosphorus, the upper end of the tube is closed with the finger, and the tube is lifted out, and dipped immediately into a tall beaker full of water. A peculiar movement takes place in the phos- phorus enclosed in the tube at the moment when it solidifies, and since the phosphorus undergoes a marked decrease of volume when it becomes solid, the stick usually falls out of the tube upon gentle tapping ; if it sticks, it can easily be pushed out with a wire. Phosphorus sticks, as thus prepared, are used in the absorption pipette, Fig. 55. The cylindrical part is filled as full as possible with the sticks, the remaining space being filled with distilled water. In order to make an absorption TECHNICAL GAS ANALYSIS. 229 with this reagent, the gas is driven over into the pipette, thereby displacing the water, and coming into contact with the moist sticks of phosphorus. A bright glow is visible when the reaction proceeds normally ; the phosphorus burns 1 o phosphoric acid, phosphorous acid, &c., at the expense of i he oxygen. After three minutes at the longest the absorption is com- pete. The end of the absorption is sharply shown by the Lisappearance of the glow, when the pipette is in a dark oom. Since the different products resulting from the oxidation if phosphorus are all soluble in water, the surface of the ticks of phosphorus is kept fresh by the action of the con- ining water alone, if that be renewed from time to time. Vnd, further, since these oxidation products, as solid and iquid substances, have a very small tension, no error is ;aused by the white cloud which may be present in the gas -esidue after the absorption. The phosphorus can be used br a very large number of analyses, if it is protected from ;he action of the light. To do this, the cylindrical part of ;he pipette is covered with a small box, or the whole pipette :s covered, when not in use, by a box of wood or cardboard which is impervious to the light. 154. Unfortunately this neat and accurate method is not universally applicable ; the following are the conditions under which it can be used. The percentage of oxygen in the gas must not be more than 50 ; and the gas must be free from ammonia, olefiant gas, and other hydrocarbons. Con- sequently in gaseous mixtures such as coal gas, which contain olefiant gas, the phosphorus method would not be applicable unless the hydrocarbons were first removed. Hydrogen. 155. Hydrogen can be determined with great accuracy by burning with oxygen. The oxygen may be used mixed with nitrogen (atmospheric air), or pure oxygen made in 230 THE GAS ENGINEER'S LABORATORY HANDBOOK. retorts blown from a glass tube, as proposed by Bunsen, may be employed. These retorts are half filled with dried and powdered chlorate of potash, and the end of the delivery tube is heated and bent in an upward direction. On heating the bulb of this improvised retort, the air is first driven out by a rapid evolution of oxygen, and the gas is then led directly into the endiometer, care being taken that the volume of oxygen does not amount to more than three or four times that of the hydrogen to be determined. The quantity of hydrogen present is two-thirds of the volume disappearing in the combustion. If the mixture contains absorbable constituents also, these are first absorbed, and the residual gas is then used for the analysis. 156. When nitrogen is present, a considerable error may be caused by not avoiding, in the combustion, the tempera- ture at which nitric acid is formed. It is consequently always necessary to calculate, after the experiment, the proportion of nitrogen to the oxy hydrogen gas burned. If this was less than six to one, the analysis must be repeated, with the addition of so much air that this proportion, or a still greater amount of nitrogen will be present. If, on the other hand, the proportion of hydrogen to incombustible gas is very small, such an amount of electro- lytic oxyhydrogen gas is added that complete combustion will result. The oxyhydrogen gas disappears completely on the combustion, and hence need not be exactly measured. 157. An accurately measured amount of pure hydrogen, mixed with an excess of air, may be used instead of the oxyhydrogen gas. The contraction resulting from the hydrogen added must then be allowed for. The combustion is made either in the explosion pipette, Fig. 58, by ignition with an electric spark, or in a glass tube filled with palladium black, or palladium sponge. The advantage of the combust- tion with palladium is, that in a mixture of hydrogen, marsh gas, and nitrogen, the hydrogen alone may be burned ; this is known as fractional combustion. TECHNICAL GAS ANALYSIS. . 231 158. The arrangement of the apparatus for carrying out tie reaction (fractional combustion) is shown in .Fig. 59. 'J 'he gas burette A and the gas pipette B are joined together 1 y means of the capillary tubes E, and the tube H. This tube H is of about four mm. internal diameter and 20 cm. i 3tal length, and it contains four grams of palladium sponge. r 'he gas pipette upon the stand G is filled with water, and i is only use is so as to render it possible to repeatedly pass ' he gas through the palladium tube. 159. To determine the amount of hydrogen present in- a [Fio. 58. mixture of hydrogen, nitrogen and marsh gas, from which, S40 far as possible, the absorbable constituents have already been removed, measure the gas in the burette, join it in the manner described to the pipette B, which is filled with water nearly to t, place the tube H in a large beaker containing warm water of from 90 to 100 C., and, after opening the pinch-cock d, drive the gas three times backward and for ward | through the palladium, by raising and lowering 232 THE GAS ENGINEER'S LABORATORY HANDBOOK, FIG. 59. TECHNICAL GAS ANALYSIS. 233 the tube a. Then replace the hot water with water of tae temperature of the room, and lead the gas residue twice backwards and forwards through the tube, in order to completely cool the gas. It is in this manner possible to absorb with certainty every particle of hydrogen. Upon c rawing the gas so far back into the measuring tube that the water in the pipette again stands near e, the difference 1 etween the two measurements made before and after the i bsorption, corresponds to the hydrogen plus the amount of ( xygen in the air enclosed in the U-tube, when the ap- ] aratus was put together. This air volume, and there- N dth its oxygen contents, may be determined with sufficient ( xactness, once for all, by closing, with a piece of rubber tubing and glass rod, one side of the tube filled with } alladium, cooling the tube to about 9 C. by placing it in cool water, and then, after connecting it by a capillary tube with a gas burette completely filled with water, warming it 1 3 100 C. by placing it in boiling water. The expansion of tie enclosed air volume corresponds to a difference of tem- jerature of 91, i.e., to a third of the enclosed volume of gas. 160. The palladium is regenerated after the reaction by first leading air over it, whereby it becomes quite hot, removing any drops of moisture which may collect, so that tie palladium may easily be shaken out of the tube in the form of a dry powder, and then superficially oxidising the metal by heating it on the lid of a platinum crucible. 161. Before using the palladium, it should be heated, in portions of about one gram at a time, nearly to redness upon the cover of a platinum crucible, so that it is covered with a larger quantity of palladium oxide than would be formed by merely leading air over it. The warm water in which the tube stands serves in the beginning to give the gases the temperature necessary to start the combustion, and at a later stage it prevents the temperature inside the palladium tube being raised too high by the reaction. 234 THE GAS ENGINEER'S LABORATORY HANDBOOK. Carbon Monoxide, or Carbonic Oxide (CO). 162. For absorbing CO, either an aramoniacal, or a hydrochloric acid solution of cuprous chloride is used. 163. The ammoniacal solution of cuprous chloride is prepared by dissolving 10*3 grams of copper oxide in 100 to 200 c.c. of concentrated common hydrochloric acid, and then allowing the solution to stand in a flask of suitable size, filled as full as possible with copper wire or copper wire gauze, until the cupric chloride is reduced to cuprous chloride, and the solution is completely colourless. The clear hydrochloric acid solution thus prepared is poured into a large glass beaker containing Ij to 2 litres of water, in order to precipitate the cuprous chloride formed. After the precipitate has settled, the dilute HC1 is poured off as completely as possible. The cuprous chloride is then washed into a 250 c.c. flask with about 100 to 150 c.c. of distilled water, and ammonia is led into the solution, which is still slightly acid, until the liquid assumes a pale blue colour. Since the tension of very concentrated ammonia solutions renders the absorption difficult, no more ammonia than is necessary should be added. While the ammonia is being led in, it is advisable to protect the contents of the flask from the oxidising influence of the air. This may be done by fitting the flask containing the cuprous chloride to be dissolved with a cork pierced with two holes, through one opening of which passes the delivery tube from the ammonia flask, while through the other hole is inserted a bent glass tube which dips into a little mercury. If a flask with a funnel tube is used for the evolution of NH 3 , hydrogen may first be led through this tube, and the apparatus be thus completely freed from air. For the evolution of am- monia, about 200 c.c. of a concentrated ammonia solution of 9 sp. gr. is used. The solution of cuprous chloride thus prepared is diluted with water to 200 c.c,, and since the HC1 is not completely TECHNICAL GAS ANALYSIS. 235 washed out, there is, of course, some ammonium chloride pi esent. 100 c.c. of this solution contain 7 3 grams of cuprous cl loride ; it is impracticable to use more dilute solutions of ci prous chloride. 1 64. For the preparation of the hydrochloric acid solution of cuprous chloride, 86 grams of copper scale are mixed with 1" grams of copper powder, prepared by reducing copper o: ide with hydrogen, and this mixture is slowly brought ii to 1086 grams of hydrochloric acid of 1-124 sp. gr. ; while tl e mixture is being added the vessel containing the HC1 si ould be constantly shaken. A spiral of copper wire r< aching from the bottom to the neck of the bottle is then p aced in the solution, and the bottle is closed with an india- rubber cork. The solution is dark at first, but it becomes o< -mpletely colourless on standing. On contact with the air it assumes a dark brown colour, due to the formation of some c ipric chloride. In the analysis of a mixture containing a n imber of gases, if CO and H are present, the latter being d jtermined by combustion with palladium after the elimina- tion of CO, it will be necessary to use the ammoniacal cuprous chloride. If the amount of CO alone is to be ascertained the hydrochloric acid solution may be used. These cuprous chloride solutions are used in the double pipette, Fig. 56. Solutions of cuprous chloride have not much tension, so that this may be disregarded in analyses which are not required to be strictly accurate. In exact determinations, however, the gases which have been in contact with the reagent must be freed from the gaseous HC1 or from the IS H 3 . This can be effected in the burette itself, or in a pipette filled with distilled water. Carbon Dioxide (C0 2 ). 165. To absorb C0 2 a solution of caustic potash is used, consisting of one part of commercial caustic potash in two parts of water. 236 THE GAS ENGINEER'S LABORATORY HANDBOOK. This solution is put into the simple pipette for solid and liquid reagents, Fig. 55, the cylindrical part being first closely filled with very short rolls of iron wire gauze. The gauze has a mesh of 1 to 2 mm. and the rolls are from 1 to 2 cm. long and about 5 mm. thick. When the per- centage of C0 2 is not too high, it can be completely ab- sorbed by simply passing the gas once into the pipette. n Efhylene (C 2 H 4 ). 166. Is either absorbed by fuming sulphuric acid, or by bromine water. It is advisable to use sulphuric acid so concentrated that when the temperature is slightly lowered, crystals of pyro- sulphuric acid will separate. The acid is used in a simple pipette pro- vided with the three bulbs, Fig. 60. The small bulb is filled at the time of blowing with glass beads, which serve to give to the sulphuric acid the largest possible surface. With this arrangement, the complete ab- sorption of the heavy hydrocarbons, and of ethylene in particular, is effected by passing the gas into the pipette once only. In this re- action some S0 2 is usually formed ; and, further, the vapour of fuming sulphuric acid has a very high tension, so that the gas residue, before being measured, must be freed from the acid vapours in the caustic potash pipette, passing the gas once into the latter pipette being sufficient. 167. To avoid the rubber connections between the pipette and burette being attacked by the fuming sul- phuric acid, the apparatus is put together in such a manner FIG. 60. TECHNICAL GAS ANALYSIS. 237 that the acid does not quite fill the capillary tube of the pipette, and the connecting capillary is allowed to remaiii ei ipty ; the short rubber tube of the burette is also freed fr)m liquid by means of a narrow tipped suction pipette, ai y reagent remaining in the rubber tube being first w ished out by water with the same pipette. If care be taken that the sulphuric acid is stopped, after tl e absorption, at the same point in the capillary at which ii stood when the burette and pipette were first put together, t- ien the small volume of air contained in the empty capillary t ibes in the beginning causes, of course, no error in the deter- n ination of the heavy hydrocarbons or other gases with the e :ception of nitrogen. In the nitrogen determination allow- a ice may be made for this air volume, but as each c.c. of the e npty capillary corresponds to only O'OOS c.c., this value is lalow the limit of the usual unavoidable experimental e rrors. After the absorption, the rubber tube is taken off from tie pipette, and the capillary and the larger tube are closed a r-tight by small pieces of glass rod, which are pushed over narrow rubber rings placed upon the glass tubes. 168. Bromine is also a good absorbent for ethylene. It is used in a similar pipette to the one just described. It is not necessary to completely fill the pipette with bromine ; it will be quite sufficient if a few c.c. of bromine lie under water in tie pipette. There is thus formed a saturated solution of bromine in water which absorbs the ethylene. Marsh Gas or Methane (CH 4 ). 169. There is no absorbent for marsh gas, which is always determined by combustion. One volume of methane unites with two volumes of oxygen, and one volume of C0 2 is formed. To avoid the burning of nitrogen in the explosion, 100 volumes of incombustible gas are taken for from 25 to 37 volumes of the mixture of methane and oxygen (Bunsen). 238 THE GAS ENGINEER'S LABORATORY HANDBOOK. Nitrogen. 170. This gas is but slightly soluble in water. No method of directly determining nitrogen is known. The residue of gas mixtures which cannot be directly determined is hence calculated as nitrogen. It follows from this, that all the errors of the preceding determinations fall upon the nitrogen, and the more complicated the gas mixture the more inexact are the results. 171. Hempel's apparatus may be used for determining the following. (1) Estimation of C0 2 in Air or in Furnace Gases. A simple absorption pipette, Fig. 54, is employed, containing a solution of KHO of sp. gr. 1 '20. (2) Estimation of Oxygen in Atmospheric Air. A compound absorption pipette, Fig. 56, filled with a strong solution of potassium pyrogallate is employed. (3) Estimation of C0 2 , 2 , and N 2 in a Furnace Gas. The gases are absorbed in the following order by the reagent mentioned. (a) C0 2 by KHO solution. (6) 2 by alkaline pyrogallate solution. (c) Measure the residual nitrogen. (4) Estimation of C0 2 , 2 , CO, and N 2 in a Furnace Gas. Absorb the gases in the following order by the reagents mentioned. (a) C0 2 by KHO solution. (6) 2 by alkaline pyrogallate solution. (c) CO by a saturated solution of cuprous chloride in hydrochloric acid. (d) Measure the residual nitrogen. The following is a description of the method of making an analysis of coal gas by means of Hempel's apparatus. 172. A quantity of water for use in the measuring tube is first saturated with the gas. by taking a flask half full of TECHNICAL GAS ANALYSIS. 239 water, leading a rapid stream of gas through the liquid, and v gorously shaking the flask. The hydrocarbon vapours are first absorbed with alcohol, then C0 2 with KHO, then the heavy hydrocarbons with fuming sulphuric acid, then oxygen with phosphorus, and h stly CO with ammoniacal cuprous chloride solution. The r> 'sidue, which consists of marsh gas, hydrogen, and nitrogen, ic measured, and is then led back into the cuprous chloride j ipette, and a portion is taken for the explosion analysis. With ordinary illuminating gas 12 c.c. of the residue suffice f >r the explosion. These 12 c.c. are measured off in the gas burette and e lough air is drawn in to bring the mixture to about 100 c.c. ] i all these measurements, the running down of the liquid i iust be carefully waited for, because the amount of gas t iken is so small, that any errors that may be made are <. reatly multiplied. The gas mixture is now burned in the c splosion pipette. The gas is then transferred to the i leasuring tube and the total contraction is measured. Then 1 16 C0 2 is absorbed with KHO, and finally the oxygen in e xcess is absorbed with phosphorus. The last determination i 5 made merely to be sure that a sufficient excess of oxygen was present in the combustion. 173. The following is an illustration. 100 c.c. of illu- riinating gas measured off. The hydrocarbon vapours are frst absorbed in the following manner. Two explosion I ipettes, Tig. 58, are used, containing 1 c.c. of alcohol and water respectively. These explosion pipettes are first filled completely with mercury, and the alcohol and water are run in at the capillary tube from burettes. Both liquids must be saturated with the illuminating gas before being used for the absorption. To do this about 50 c.c. of gas is passed into each pipette, and the pipettes are shaken for several minutes. The gas is then driven out and the pipettes are ready for use. 174. The measuring tube containing the 100 c.c. of gas is joined, by a fine-bore capillary tube, to the pipette containing 240 THE GAS ENGINEER'S LABORATORY HANDBOOK. the 1 c.c. of alcohol, the gas is driven over, the pipette is closed by a pinch-cock and disconnected, and shaken for three minutes. The gas is then drawn back into the burette, and passed into the pipette containing the 1 c.c. of water, and is there shaken for three minutes. The gas is then drawn back into the measuring tube and measured = 7 c.c. or per cent, hydrocarbon vapours. Passed into KHO pipette and drawn directly back into the measuring tube, measured after three minutes (time allowed for the running down). Mea- surement gave 4*1 c.c. ; hence there was present 4*1 0'7 = 3 4 c.c. or per cent. C0 2 . 175. Burette now connected by means of a dry piece of rubber tube and a dry capillary with the pipette containing fuming sulphuric acid. Gas driven over and drawn back at once into the measuring tube. Gas now passed again into KHO pipette, and after being drawn back into the measuring tube and allowed to stand three minutes again measured. The measurement gave 8 4 c.c. ; hence there were 8*4 4*1 = 4*3c.c., or 4*3 per cent, of heavy hydrocarbons present. 176. The gas now passed into phosphorus pipette, Fig. 55, and allowed to stand for three minutes, then drawn back into burette and measured at the end of three minutes. Reading gave 8*4 c.c., hence no oxygen was present. 177. The gas was then passed into the pipette containing ammoniacal cuprous chloride, Fig. 56, which had been repeatedly used, and was shaken therein for two minutes. It was then drawn back into the burette and transferred at once to a second pipette containing ammoniacal cuprous chloride which had been used but little, and it was here shaken for three minutes. Drawn back into the burette and measured after three minutes ; the reading was 18 c.c. hence there was 18 c.c. 8'4 = 9*6 c.c. or per cent, of CO present. 178. The remaining 82 c.c. of gas were then passed back into the cuprous chloride pipette, and the pipette was closed with an ordinary pinch-cock. The water in the burette is poured out, the burette TECHNICAL GAS ANALYSIS. 241 washed with hydrochloric acid and then with distilled water, ard then filled with water which is saturated not with illuminating gas but with air. 12 to 15 c.c. of the gas resi- dre are now measured off into the burette. In this case lc *2 c.c. were taken. 170. So much air is then drawn in that the total volume of the gas residue taken and the air amounts to about 1( c.c. In this case it was 99* 6 c.c. This mixture is now brought into an explosion pipette, F g. 58, filled with mercury, care being taken that the capil- la y remains full of water. The rubber connecting piece is cl )sed by a strong pinch-cock, and a piece of glass rod is si pped into the end of the india-rubber tube. The pipette is then vigorously shaken, the glass stop-cock is closed, the pi pette is connected with the poles of an induction coil and tl e mixture is exploded. The glass stop-cock is at once O] ened and the remaining gas is transferred without delay to the measuring tube, and after three minutes measured. T le result here was 78 c.c. The total contraction was therefore 99-6 78 = 21-6 c.c. 180. The gas remaining from the combustion is now pe ssed into the KHO pipette, drawn directly back into the l)i rette, and after three minutes measured. The reading was 73*2 c.c. Hence by the combustion 78 73 '2 = 4' 8 c.c. of C() 2 were formed. 181. Although this gave all the data necessary for the ca .culation of the analysis, the remaining gas was neverthe- less passed into the phosphorus pipette in order to be sure that an excess of oxygen was present in the combustion, or in other words that the gas was completely burned. The measurement gave 70 '2 c.c. Hence there were 73-2^ 70-2 = 3 c.c. of oxygen in excess. 182. In the combustion of the marsh gas its own volume of C0 2 is formed, so that in the 13*2 c.c. of the gas residue taken for the explosion there were 4= 8 c.c. of marsh gas. 242 THE GAS ENGINEER'S LABORATORY HANDBOOK. The marsh gas in the total gas residue of 82 c.c. is found by the proportion 13-2 : 82 :: 4-8 : x. x = 29 '8 per cent, marsh. Since marsh gas in burning unites with twice its volume of oxygen, the contraction which has resulted from the combustion of the hydrogen is found by subtracting twice the volume of the C0 2 found from the total contraction. 21-6 - (2 x 4-8) = 12 c.c. contraction due to the burning of hydrogen. 183. One volume of hydrogen unites in burning with one-half its volume of oxygen ; hence the volume of the hydrogen is found by multiplying 12 x Thus the 13*2 c.c. of the gas residue taken for the explosion contained 8 c.c. of hydrogen. The total amount of hydrogen is given by the proportion 13-2 : 82 :: 8 : x. x = 49 6 per cent, hydrogen. 184. The nitrogen is found by subtracting the sum of all the other constituents from 100. This gives 2 6 per cent. Hence the illuminating gas contained 0*7 per cent, hydrocarbon vapours 3-4 C0 2 4*3 heavy hydrocarbons 0-0 oxygen 9-6 CO 29-8 CH 4 49-6 H 2 ^6 N 2 100-0 185. The determination of hydrogen may be made more exactly by absorption with palladium. The following analy- sis is given to illustrate the calculation of the analysis when the hydrogen is fractionally burned. TECHNICAL GAS ANALYSIS. 243 The direct absorption gave 6 per cent- hydrocarbon vapours 3-4 C0 2 4-4 C 2 H 4 0-3 Oxygen 10-1 CO 186. The residue of hydrogen, methane and nitrogen, amounted to 81 '2 c.c. This was transferred to a pipette and 4( -5 c.c. were measured off in a burette for the fractional co nbustion of the hydrogen. To this was added air (in this ca ;e 58 7 c.c.), and the mixture was passed into a pipette fil ed with water. Then more air was measured off in the bi rette and transferred to the pipette, so that the total an ount of air added would, without doubt, be sufficient for th 5 combustion of the hydrogen. In this second measurement 16*1 c.c. of air were taken, so th it the total amount of gas taken for the fractional com- bustion was 40 '5 + 58-7 + 16-1 = 115-3 c.c. The gases we re vigorously shaken in the pipette to thoroughly mix th sm, and were then fractionally burned by leading them ov ir 5 gram of palladium black. [Palladium black is made by reducing palladious chloride with alcohol in a strongly all aline solution, the same method being used in preparing pli tinura black with platinum chloride. The palladium black is placed in a small U-tube, Fig. 59, plunged into water at a tenperature of 100 C., and the gas and air is passed slowly through the tube two or three times. The tube at the time mi st be connected with an ordinary absorption pipette filled wi ;h water, or else with the KHO pipette, which in this case, of course, simply acts as a kind of receiver.] The volume after the combustion was 81 c.c., hence the contraction was 115-3 - 81 = 34-3 c.c., corresponding to 22-9 c.c. of hydrogen, the total amount of hydrogen being found by the proportion As 40-5 : 81-2 :: 22-9 : x. x = 45-9 per cent, hydrogen. B 2 244 THE GAS ENGINEER'S LABORATORY HANDBOOK. 187. To determine the marsh gas, 19*9 c.c. of the residue of hydrogen, marsh gas and nitrogen were taken, and, to- gether with 110 c.c. of air, were transferred to the explosion pipette. The gases were well mixed and were then exploded, freed from C0 2 in the KHO pipette, and measured. There remained 90*5 c.c. The contraction was 110+ 19-9 90-5= 39-4 c.c. 188. From the determination of the hydrogen by the fractional combustion, we know that in 19 c.c. of the residue the hydrogen would cause a contraction of 16 -9 c.c. (40*5 : 19 '9 : : 34 '3 : x. x = 16 '9), hence the contraction due to the methane is equal to 39*4 - 16*9 = 22*5 c.c., and the volume of the methane itself is 7 5 c.c. The percentage of methane is 19-9 : 81-2 :: 7-5 : x. x = 30 6 per cent, methane. 189. The nitrogen, determined by difference as before, is 4' 7 per cent., so that the composition of the gas was as follows : 0-6 per ceiit. hydrocarbon vapours 3-4 4.4 0-3 10-1 45-9 30-6 4-7 100-0 C 2 H 4 2 CO H 2 CH 4 N 2 GAS ANALYSIS BY MEANS OF BUNTE'S APPARATUS. Analysis of Furnace Gases. 190. These may be conveniently determined by means of the apparatus devised by Dr. Biinte, late of the gasworks, Munich. The following is a description of the apparatus and method of using it : TECHNICAL GAS ANALYSIS. 245 A and B, Fig. 61, are two burettes fitted with three-way co jks, each graduated in fifths of a c.c., and capable of holding 113 c.c. ; T, a one-gallon tubulated bottle, serving as a water re ;ervoir, and F an aspirator. The burette is filled by oj ening the stop-cocks, b and a, and allowing water from the be ttle, T, to enter the burette until it nearly fills the fuinel, t. The stop-cock should then be closed and the in lia-nibber tube detached from the bottom of the burette. T ie longitudinal bore of the stop-cock, a, is now connected w th the tube supplying the gas to be examined, and the gas aspirated by running the water out of the burette by m :ans of the stop-cock, b. Rather more than 100 c.c., sa f, 105 c.c., of gas should be allowed to enter the bi rette, and the exact adjustment to the zero mark then m ide as follows : By means of the bottle, T, sufficient witer should be forced into the burette to compress tl 3 gas to about 95 c.c. ; then b is closed, the bottle, T, is d( tached, and by cautiously turning the tap, b, the water is rr. n out again exactly to the zero mark. The gas is still ui der a plus pressure, and now that pressure has to be e& ;ablished at which every reading off has to take place with this apparatus. For this purpose the funnel, t, should be filed with water up to the mark, when, on opening the tap, a, for an instant the excess of gas will escape through the water. The burette now contains exactly 100 c.c. of gas at tli3 pressure of the atmosphere, plus the pressure of the co umn of water standing in the funnel, /. Determination of Carbonic Acid. 191. The tube, r, of the aspirator, F, is connected to the bottom of the burette, suction applied at S, stop-cock, 6, opened, and as much water as can be aspirated allowed to run out. The stop-cock should then be closed, the tube removed, and the end of the burette dipped into a strong solution of caustic potash. On opening the stop-cock, &, a quantity of 246 THE GAS ENGINEER'S LABORATORY HANDBOOK. FIG. 61. TECHNICAL GAS ANALYSIS. 247 tho solution will enter the burette. (Care should "be taken th.-it the bottom of the burette is always immersed in the solu- ti< n, or air would enter and so spoil the experiment.) When tho solution has entered the burette, the stop-cock, b, should be closed, the burette taken from its support, and the hand of th 3 operator being firmly placed on t, the contents of the bu- re ;te should be vigorously shaken up with the solution, when, if the solution of potash employed has been strong enough, it w 11 be found that the whole of the C0 2 has been absorbed ; tl erefore, on opening the stop-cock, a, and allowing water to fl< w down until the normal atmospheric pressure is reached (^ rhich is shown by the water ceasing to flow), and filling tl e funnel up to mark, on reading off the amount absorbed b T the caustic potash = percentage of C0 2 . Determination of Oxygen. 192. The caustic potash solution should be drawn off by n eans of the aspirator, and an alkaline solution of pyrogallic a ;id applied in the same manner as in the estimation of C0 2 ; il oxygen is present, the solution immediately turns black. The g is should be well shaken up with the solution, but in this case the absorbing liquid is not so active as the potash. It v ill thus be necessary to see if the gas will take up any more o i" the solution. Therefore, having replaced the burette in its s ipport, the solution should again be applied to the burette, thereupon some of the liquid will enter in the place of the absorbed gas. The burette should then be well shaken, and the solution applied again, when, if the liquid remains at the same level as it previously stood at, the absorption is complete; otherwise, the process must be continued until there is no further absorption. After the reduction of the contents of the burette to normal pressure, the difference between the volume now observed and that at the com- mencement, after the abstraction of the C0 2 , will give the percentage of 0. 218 THE GAS ENGINEER'S LABORATORY HANDBOOK. Determination of Carbonic Oxide. 193. If oxygen has been proved to be present, carbonic oxide will most likely be absent, unless the gases have been brought together at a temperature insufficient to promote their combination. Carbonic oxide and hydrogen are most conveniently estimated by combustion, as under. The burette should first be cleared from the solution of alkaline pyrogallate, by opening the cocks, a and &, and allowing a stream of water to run through the burette until clean, taking the precaution to always keep the funnel, , sealed. It is now necessary to mix the gases with an ex- cess of air, but as the burette would not hold sufficient to combine with the whole of the gases generally present, it is requisite to expel a portion and work on, say, half the volume. For instance, if 100 c.c. were originally taken, and there were found 10 per cent of C0 2 , on driving out gas until it measured 45 c.c. we would get the equivalent of half the volume originally present. The air is admitted by placing the cock, a, in communication with the burette, opening the pinch-cock at c, and allowing water to flow out at b. This should be continued until the water level is two or three divisions below zero, when the cock, 6, and the pinch-cock should be closed, the contents of the burette brought to the normal pressure, and the reading of the burette taken, the contents of the burette having been pre- viously well shaken up so as to mix the gases with the air. Connection should then be made between the burettes, A and B, uniting the two at the india-rubber tubes, C and C', by means of a piece of glass combustion tubing containing a small coil of palladium wire. The burette, B, should pre- viously have been filled with water in the usual manner. The palladium wire in the combustion tube should be brought to a red heat by means of a Bunsen burner, and the gas in the burette is caused to pass from A over the heated wire into B by opening b and c, connecting the water-supply, TECHNICAL GAS ANALYSIS. 249 T, with "bottom of burette, A. When all the gas from A has passed over (shown by the burette being full of water) the o] >eration is reversed, the gas being again collected in A. It is tl en allowed to stand until the normal temperature is attained, a< Ljusted for pressure, and the volume then read off, and this reading kept for reference. A solution of caustic potash should then be applied in the manner described for the e .timation of C0 2 , and the diminution in volume also noted. 194. In order to work out the calculations, it will be n >cessary to notice what action takes place during the com- b istion. The gases to be dealt with are hydrogen, carbonic o :ide and nitrogen, mixed with an excess of air; by passing -er the red hot palladium wire, the oxygen of the air com- b nes with the carbonic oxide CO, to form carbonic anhydride, ( J 2 ; (CO + = C0 2 ) and with the hydrogen to form A\ ater, H 2 O ; (H 2 + = H 2 0) the nitrogen of course not b dng affected. Now, supposing that after igniting the g ises and treating the residue with caustic potash, a diminu- tion in volume of 10 c.c. was observed, this would be equal to 1 > c.c. of C0 2 , for each volume of CO produces an equal volume of C0 2 , and, as only half the original volume taken was treated, on multiplying the result by two, the true per- centage of C0 2 will be obtained. Supposing, also, that after ignition, but before treatment with KHO, there was a diminution in tbe volume of 10 c.c., this would be due partly to the combination of the hydrogen with the oxygen, and partly to the combination of the carbonic oxide with the oxygen. The CO would require half its volume, or 5 "c.c. of oxygen ; therefore, on deducting 5 c.c. from the 10 c c. due to combustion, we get 5 c.c. as a result of the combination of the hydrogen with the oxygen. Hydrogen combines with oxygen to form water, in the proportion of two volumes to one; consequently, on multiplying 5 by we get 3*3 c.c. as the number of c.c. of hydrogen, and this multiplied by 2, gives 6 6 as the percentage of hydrogen. Nitrogen is calculated by difference. 250 THE GAS ENGINEER'S LABORATORY HANDBOOK. 195. In order to effect the estimation of CO by absorption, it is necessary to remove every trace of the pyrogallate solu- tion by the use of the funnel and the suction bottle. This done, a concentrated solution of cuprous chloride in HC1 is applied to the bottom of the burette in the usual manner. When absorption is complete the Cu 2 Cl 2 is drawn off, the tube washed, and treated with a solution of KHO, for the purpose of absorbing any HC1 vapour which may have been liberated in the reaction. After bringing to correct pressure, the reading shows the percentage of CO. 196. The gas may be conveniently taken off by tightly inserting a wrought iron tube, or preferably an inner pla- tinum tube encased by an outer tube of wrought iron, into the furnace or flue, and then aspirating into a suitable vessel, taking every precaution to prevent any air mixing with the gases, by aspirating a sufficient quantity. The sample may then be transferred to the laboratory for analysis. APPENDIX. GAS EEFEBEES' INSTRUCTIONS. As TO THE TIMES AND MODE OF TESTING FOR PURITY. r HE testings for purity shall extend over twenty hours of each day, r id shall be made upon ten cubic feet of gas, which shall be tested > iccessively for each of the following impurities : I. Sulphuretted Hydrogen. The gas shall be passed, as it leaves the service-pipe, through an : pparatus in which are suspended slips of bibulous paper, impregnated ^ ith basic acetate of lead. The test-paper from which these slips are cut is to be prepared from t me to time by moistening sheets of bibulous paper with a solution of ] part of sugar of lead in 8 or 9 parts of water, and holding each sheet A 'hile still damp over the surface of a strong solution of ammonia for a f 3W moments. As the paper dries all free ammonia escapes. If any discoloration of the slip of test-paper is found to have taken place, this is to be held conclusive as to the presence of sulphuretted 1 ydrogen in the gas. Fresh test-slips are to be placed in the apparatus tvery day. In the event of any impurity being discovered, one of the test-slips shall be placed in a stoppered bottle, and kept in the dark at the testing place ; the remaining slips shall be forwarded with the daily report. II. Ammonia. The gas which has been tested for sulphuretted hydrogen shall pass next through an apparatus consisting of a glass cylinder filled with glass beads, which have been moistened with a measured quantity of 252 THE GAS ENGINEER'S LABORATORY HANDBOOK. standard sulphuric acid. A set of burettes, properly graduated, is pro- vided. The maximum amount of ammonia allowed is 4 grains per 100 cubic feet of gas ; and the testings shall be made so as to show the exact amount of ammonia in the gas. Two test-solutions are to be used one consisting of dilute sulphuric acid, of such strength that 25 measures (septems) will neutralise 1 grain of ammonia ; the other, a weak solution of ammonia, 100 measures of which contain 1 grain of ammonia. The correctness of the result to be obtained depends upon the fulfil- ment of two conditions : 1. The preparation of test-solutions having the proper strength. 2. The accurate performance of the operation of testing. To prepare the test-solutions the following processes may be used by the gas-examiner : Measure a gallon of distilled water into a clean earthenware jar or other suitable vessel. Add to this 94 septems of pure concentrated sulphuric acid, and mix thoroughly. Take exactly 50 septems of the liquid and precipitate it with barium chloride in the manner prescribed for the sulphur test. The weight of barium sulphate which the test acid should yield is 13*8 grains. The weight obtained with the dilute acid prepared as above will be somewhat greater, unless the sulphuric acid used had a specific gravity below 1 '84. Add now to the diluted acid a measured quantity of water, which is to be found by subtracting 13 '8 from the weight of barium sulphate obtained in the experiment, and multiplying the difference by 726. The resulting number is the number of septems of water to be added. If these operations have been accurately performed, a second precipi- tation and weighing of the barium sulphate obtainable from 50 septems of the test acid will give nearly the correct number of 13 8 grains. If the weight exceeds 13 '9 grains, or falls below 13 '7 grains, more water or sulphuric acid must be added, and fresh trials made, until the weight falls within these limits. The test acid thus prepared should be trans- ferred at once to stoppered bottles which have been well drained, and are duly labelled. To prepare the standard solution of ammonia, measure out as before a gallon of distilled water, and mix with it 50 septems of strong solu- tion of ammonia (sp. gr. 0*88). Try whether 100 septems of the test alkali thus prepared will neutralise 25 of the test acid, proceeding according to the directions given subsequently as to the mode of testing. APPENDIX. 253 1 ' the acid is just neutralised by the last few drops, the test alkali is of t le required strength ; but if not, small additional quantities of water of strong ammonia solution, must be added, and fresh trials made, uitil the proper strength has been attained. The bottles in which the solution is stored should be filled nearly full, and well stoppered. The mode of testing is as follows : Take 50 septems of the test- a 3id (which is greatly in excess of any quantity of ammonia likely to 1 .? found in the gas), and pour it into the glass cylinder, so as to well A - et the whole interior surface, and also the glass beads. Connect one t >rminal tube of the cylinder with the gas supply, and the other with t le meter, and make the gas pass at the rate of about half a cubic foot ] er hour. Any ammonia that is in the gas will be arrested by the : .ilphuric acid, and a portion of the acid (varying with the quantity of ; mmonia in the gas) will be neutralised thereby. At the end of each criod of testing, wash out the glass cylinder and its contents with i istilled water, and collect the washings in a glass vessel. Transfer ( ne-half of this liquid to a separate glass vessel, and add a quantity of ; neutral solution of hgematoxylin or litmus just sufficient to colour he liquid. Then pour into the burette 100 septems of the test alkali, ; nd gradually drop this solution into the measured quantity of the ' v-ashings collected, stirring constantly. As soon as the colour changes I indicating that the whole of the sulphuric acid has been neutralised), : ead off the quantity of liquid remaining in the burette. To find the number of grains of ammonia in 100 cubic feet of the gas, multiply i he number of septems of test alkali remaining in the burette by 2, and uove the decimal point one place to the left. The remaining half of the liquid is to be preserved in a bottle, duly labelled, for a week. III. Sulphur Compounds other than Sulphuretted Hydrogen. The gas which has been tested for sulphuretted hydrogen and am- monia shall pass next through a meter, by means of which the rate of low can be adjusted to half a cubic foot per hour, and which is pro- vided with a self-acting movement for shutting off the gas when ten oubic feet have passed. The testing shall be made in a room where no gas is burnt other :han that which is being tested for sulphur and ammonia. The apparatus to be employed is represented in Fig. 62, and is of the following description : The gas is burnt in a small Bunsen burner 254 THE GAS ENGINEER'S LABORATORY HANDBOOK. with steatite top, which is mounted on a short cylindrical stand, per- forated with holes for the admission of air, and having on its upper surface a deep circular channel to receive the wide end of a glass trumpet tube. On the top of the stand, between the narrow stem of FIG. 62. the burner and the surrounding glass trumpet tube, are to be placed pieces of commercial sesqui-carbonate of ammonia, weighing in all about two ounces. The products, both of the combustion of the gas and of the gradual volatilisation of the ammonia salt, go upwards through the trumpet tube into a vertical glass cylinder, packed with balls of glass, to break APPENDIX. 255 ur_ the current and promote condensation. From the top of the cylinder there proceeds a long glass pipe or chimney, serving to effect sone further condensation, as well as to regulate the draught, and afiord an exit for the uncondensable gases. In the bottom of the cy Under is fixed a small glass tube, through which the liquid (formed di ring the testing) drops into a beaker placed beneath. The following cautions are to be observed in selecting and setting up tl 3 apparatus : See that the inlet pipe fits gas-tight into the burner, and that the h< les in the circular stand are clear. . If the burner gives a luminous fi; me, remove the top-piece, and having hammered down gently the ri zzle of soft metal, perforate it afresh, making as small a hole as will g ye passage to half a cubic foot of gas per hour at a convenient pres- si re. See that the tubulure of the condenser has an internal diameter of n >t less than f inch, and that its outside is smooth and of the same size a; the small end of the trumpet tube. See that the short piece of india-rubber pipe fits tightly both to the ti timpet tube and to the tubulure of the condenser. The small tube at the bottom of the condenser should have its lower e: id contracted, so that when in use it may be closed by a drop of v ater. The india-rubber pipe at the lower end of the chimney tube should fi ', into, and not simply rest upon the mouth of the condenser, and the u )per extremity of this tube may with advantage be given a downward c-irvature. At the end of each period of testing, the cylinder and trumpet tube are to be well washed out with distilled water. Fresh pieces of scsqui-carbonate of ammonia are to be used each day. The gas examiner shall then proceed as follows : The liquid in the beaker and the water used in washing out the apparatus shall be put into the same vessel, well mixed, and measured. C ne-half of the liquid so obtained is to be set aside, and preserved for a week, properly labelled, in case it should be desirable to verify the correctness of the testing. The remaining half of the liquid is to be put into a flask, or beaker, covered with a large watch-glass treated with hydrochloric acid, sufficient in quantity to leave an excess of acid in the solution and tnen raised to the boiling point. An excess of a solution of barium chloride is now to be added, and the boiling continued for five minutes. 256 THE GAS ENGINEER'S LABOEATORY HANDBOOK. The vessel and its contents are to be allowed to stand till the barium sulphate settles at the bottom of the vessel, after which the clear liquid is to be as far as possible poured off through a paper filter. The remain- ing liquid and barium sulphate are then to be poured on to the filter, and the latter well washed with hot distilled water. [In order to ascertain whether every trace of barium chloride and ammonium chloride has been removed, a small quantity of the washings from the filter should be placed in a test-tube, and a drop of solution of silver nitrate added. Should the liquid, instead of remaining perfectly clear, become cloudy, the washing must be continued until, on repeating the test, no cloudiness is produced.] Dry the filter with its contents, and transfer it into a weighed platinum crucible. Heat the crucible over a lamp, increasing the temperature gradually, from the point at which the paper begins to char, up to bright redness. When no black particles remain, allow the crucible to cool ; place it when nearly cold in a desiccator over strong sulphuric acid, and again weigh it. The difference between the first and second weighings of the crucible will give the number of grains of barium sulphate. Multiply this number by 11 and divide by 4 ; the result is the number of grains of sulphur in 100 cubic feet of the gas. This number is to be corrected for the variations of temperature and atmospheric pressure in the manner indicated under the head of " Illuminating Power," with this difference, that the readings of the barometer and thermometer are to be taken for the day on which the testing commenced, and also the day on which it closed ; and the mean of the two is to be used. This correction may be made most simply, and with sufficient accuracy, in the following manner : When the tabular number is between 955-965, 966-975, 976-985, 986-995, increase the number of grains of sulphur by y When the tabular number is between 996-1005, no correction need be made. When the tabular number is between 1006-1015, 1016-1025, 1026-1035, diminish the number of grains of sulphur by -^th, T oths, APPENDIX. 257 Example. Grains of barium sulphate from 5 cubic feet of gas . . 4 3 Multiply by 11, and divide by 4 11 4)47-3 Grains of sulphur in 100 cubic feet of gas (uncorrected) 11 * 82 Add 11-8 XT^= '24 Grains of sulphur in 100 cubic feet of gas (corrected) .. 12 '07 Barometer (mean) 29 '4 Thermometer (mean) 58 Tabular number 985 Result 12-1 grs. L s TO THE MAXIMUM AMOUNT OF IMPURITY IN EACH FORM WITH WHICH THE GAS SHALL BE ALLOWED TO BE CHARGED. Sulphuretted Hydrogen. By the Acts of Parliament all gas sup- pi ed must be wholly free from this impurity. Ammonia. The maximum amount of this impurity shall be 4 grains per 100 cubic feet. Sulphur Compounds other than Sulphuretted Hydrogen, The m iximum amount of sulphur with which gas shall be allowed to be cl arged shall be 17 * grains of sulphur in every 100 cubic feet of gas. METERS. The meters used for measuring the gas consumed in making the vciious testings, having been certified by the Eeferees, shall, at periods of not less than seven days, be proved by the Gas Examiners by means of the Referees' Cubic Foot Measure a description of which apparatus, w th directions how to use it, is given below. Should a meter show ary variation, water must be added or withdrawn until the meter is correct. Every testing place shall have the above-mentioned * The amount allowed in the winter is 22 grains. 258 THE GAS ENGINEERS LABORATORY HANDBOOK. apparatus, so that the Gas Examiner may employ it whenever he thinks necessary. No meter other than a wet one shall be used in testing the gas under these instructions. GAS REFEBEES' CUBIC FOOT MEASUBE. This instrument is a vessel of a cylindrical form with rounded ends, made of hardened tin about one quarter of an inch thick, fitted at each end with a narrow glass tube the joints being made sound with india-rubber packing. The instrument stands in a vertical position firmly fixed to a strong plank. At the top of the instrument is a three-way cock, marked on the head of the key with a T> each arm of which shows the direction of a way through the plug; and attached to the side of this cock there is a small pipe, which serves to admit gas when the measure is to be charged, When the T is in its ordinary position communication is closed between the measure and the tube leading from it to the meters, but is open to the gas supply; but with the 1 in this position (i.e. with the stem pointing to the opposite side of the cock to that connected with the gas supply) the measure is open to the meters and shut to the gas. At the bottom of the instrument there are three cocks, one large and two small ones. One of the small ones, when opened, admits water into the measure ; the other small one is used for verifying the adjustment of the measure, by discharging the water into a standard mea- sure. The large cock, when opened, allows the water to run off. < 63. The cock at the top of the measure is con- tinued by a tin pipe, which rises above the level of the water in the cistern, and then returns downwards, passing to the two meters. Affixed round the glass tubes fitted to the upper and lower part of APPENDIX. 259 the measure there are narrow strips of paper, which indicate the exact me; sure of one cubic foot. The instrument should be in communication with a tank of water in t he same room. To verify the meters employed in ascertaining the consumption of gas in the photometers, or those employed in the sulphur and ammonia test s, the mode of procedure is as follows : Starting with the bottle full of water, turn the three-way cock at the top of the instrument so as to place the measure in communication wit n the gas supply from the main, and open the large cock at the bot ;om of the apparatus. When the bottle is rilled with gas to below the water-line, close the large cock, and turn the upper three-way cock so is to close the inlet for gas, while opening the way to the meters. Ne ct turn the cock of the meter which is to be tested so as to cut off th( ordinary gas supply and to place the meter in communication with th( cubic foot measure, taking care that all the other meters are closed to :he cubic foot measure. See that the tap is open which allows gas to pass from the meter through the governor to the burner. Then pn ceed gently to open the small lever cock and allow water to flow int > the measure, sufficient to fill the lower glass tube up to the water- lin !. Note the exact position of the index hands. Now turn on the le\ 3r cock, watching the pressure-gauge attached to the measure, until ga,^ is passing through the meter at about the normal rate of five cubic fee ; an hour. When the measure is nearly full, stand by it to check thi rate of flow, and finally shut the water off as it reaches the upper wa:er-line. Eead the meter. Should the meter have completed more than the prescribed number of revolutions when the measure is full, then some water must be rer loved from the meter ; if the contrary is the case, then water must be added to the meter. The testing is then to be repeated until the meter is found to register correctly. The dial of the photometer meters is divided into 50 divisions; and as each revolution indicates one-twelfth of a foot each division consequently represents the e^o tn P art f a cubic foot ; and therefore 6 cf these divisions represent i^th part of a cubic foot. The dial of the sulphur meters is divided into 100 parts ; and as each complete revolution indicates one foot, each division consequently represents 1 per cent, of the volume sent through. For accurate measurement it is essential that the temperature of the gas in the measure and that in the meter should be the same. If the temperature s 2 260 THE GAS ENGINEER'S LABORATORY HANDBOOK. of the water with which the measure has been filled is different from that of the meter, a sufficient time must be allowed to elapse after the measure has been filled with gas for the measure to have gained the temperature of the surrounding air. The measure and the meters must be so placed that the temperature of the air which surrounds them is the same. The difference of about an ounce in the quantity of water in the meters when they are charged nearly to the water level will make them register 1 per cent, either fast or slow. When the temperature of the cubic foot measure is higher than the temperature at the outlet of the meter, the meter (if correct) will register a smaller quantity than has actually passed to it from the measure ; when the temperature of the measure is lower than that of the meter, the quantity registered ought to be greater. To find the volume which the meter ought to register when such a difference of temperature exists, divide the tabular number corresponding to the barometric pressure and the temperature of the cubic foot measure, by the tabular number corresponding to the barometric pressure and: the temperature of the meter (see page 261}. APPENDIX. 261 OO. I, Ir-ICNCXC^COCOCO^-^-^lOtfJ^CDcOCO 050505050505050505050505050505C505050505 05 IN CO 01 CO * J 1~ 005( US 05 M in O> (M to 05 CO (O 05005050> 00000000 05 05 C5 05 05 i^ooco*~j--t-aoaoooc5C5O5OOO' 'r-ir-HOloic^coccco-^^^Oioco 05050505050505050405050S0500OOOOOOOOOOOOOOOO 050505050505050505050505000000000 o222c5Scoco-*^vnu5SSS ooooooooooooooooo 0505050050502000 ^HiMcoroco-^-^-^irainmcoto*- 0000000000OOOOOOOO ^sslliilllllssslllil 00000000000 ^cor-'iftaoc < imO5C3coo5 coco o ti*-oocioaoo;.Hr-iMMWHi-*'*iie*oteoce*-t- CN I O5O5O5O5O5O5O6O5OOOOOOOOOOOOOOOOOOOOOO ooooooooooooo ooooooooooo ^ J O^HacDt-COO5O .s > II Z II li 1 1 I! *CJ *S S II; A S3 a sli . 111 I a .s 262 THE GAS ENGINEER'S LABORATORY HANDBOOK. iCOOSC- CD GO OS CO t*~ C^ !> O CO tO CO !M CD CO CM CO r-l OS O CM OS ) r> Arsenious oxide. As 2 O 3 .. 0-80488 j j> j Arsenic anhydride, As 2 O 5 0-93496 I arium : Barium sulphate, BaSO 4 Barium oxide, BaO . . 0-65665 carbonate, BaC0 3 >j > 0-77655 ilcium : Calcium sulphate, CaSO 4 carbonate, CaC0 3 .. Calcium oxide, CaO .. > 0-41176 0-56000 C arbon : Carbonic anhydride, C0 2 '^ .. 0-27273 Calcium carbonate, CaC0 3 .. Carbonic anhydride . . . . 0-44000 Barium Carbonate, BaC0 3 .. > .... 0-22335 ( hlorine : Silver Chloride, AgCl .. .. Chlorine, Cl 0-24739 jj j> jj .... Hydrochloric acid, HC1 .. 0-25435 j .... Sodium chloride, NaCl . . 0-40767 I :on : Ferric Oxide, Fe 2 3 .. .. Iron, Fe 0-7000 > .... Ferrous oxide, FeO .. 0-9000 1 [ydrogen : Water, H 2 O Hydrogen, H 0-11111 I lagnesium : Magnesium pyrophosphate,\ Mg 2 P 2 7 / Magnesium, Mg 0-2160 Sulphur : Barium sulphate, BaSO 4 Sulphur, S 13734 Sulphuric anhydride, S0 3 0-3433 H f> Sulphurous anhydride, S0 2 0-2746 These factors are employed in the following manner : The weight of a precipitate having been obtained, on multiplying this weight by the factor given, the weight of the substance to be determined is arrived at. Thus, on multiplying any weight of ferric oxide by the factor 0*7 the equivalent weight of iron (Fe) will be obtained ; or any weight of BaSO 4 , on being multiplied by - 13734, will give the equivalent weight of sulphur. 264 THE GAS ENGINEER'S LABORATORY HANDBOOK. SOLUBILITY OF DIFFBEENT SALTS. Eemark. The solubility is given in parts of the anhydrous salt dissolved by 100 parts of water. 100 Water Dissolve. Cold. Boiling. 9 422 potash 9-5 857 Aluminium sulphate Ammonium oxalate nitrate sulphate Barium chloride 33 4-5 199 66 35 89 40-8 100 60 5 10 nitrate g 35 2 21 3 Calcium carbonate 0-0036 400 hydrate 0*128 0-079 400 sulphate 0-23 0-21 7 19-8 nitrate 127 sulphate 21 75 Iron protosulphate Lead acetate 20 46 178 71 3 5 ,, nitrate 48 139 sulphate 008 Magnesium oxide .. carbonate chloride .. Manganous chloride .. .. .. .. 002 02 200 62 11-5 002 400 123 100 Potassium hydrate .. .. .. .. chromate (neutral) bichromate oxalate (acid) .. sulphite hyposulphite .. ... bitartrate .. tartrate (neutral) . .. cyanide .... ... ferrocyanide .. ... ferricyanide .. ... iodide 200 48 10 2-5 100 hygroscopic 0-4 133 122 28 40 141 102 10 10-5 296 9*i 82 221 APPENDIX. 265 OP DIFFERENT SALTS continued. 100 Water Dissolve. Cold. Boiling. Sod um acetate (borax) 35 4 150 55 61 hyposulphite , phosphate 50 12 25 more than 200 100 Stri ntium hydrate nitrate 1-6 20 34-8 113 chloride Tat aric acid 53 76 102 200 Tii (stannous) chloride 270 300 ,, sulphate .. 50 95 BOILING POINTS. Centigrade. Fahrenheit. All ohol, absolute An monia, anhydrous nitrate, satur. solution Ba: ium chloride, satur. solution BiMlphide of carbon Bci zol .. o 78 -38-5 164 104-4 47-0 80-4 172-4 -37-3 327 220 116-6 177 Br< mine . . . 63-0 145-4 Cal jiuni chloride, satur. solution 66 per cent, solution .. 33 nitrate, satur. solution Caibonic acid 179-5 156 128 152 78 355-1 312-8 262-4 305-6 108 Ether 35 95 Hydrochloric acid, 20-2 per cent. HC1 .. Iodine above 110 200 230 392 Methvlic alcohol . . .. . 60 140 357 674-6 Naphthalene 217 422-6 Nitric acid, most concentrated ,., specific gravity 1 42 Niirous anhydride ,, oxide . . . . 86 121 -2 88 186-8 249-8 28-4 126 28 82-4 266 THE GAS ENGINEER'S LABORATORY HANDBOOK. BOILING POINTS continued. Centigrade. Fahrenheit Potassium chloride, satur. solution chlorate acetate carbonate nitrate Sodium chloride acetate carbonate phosphate nitrate Sulphur Sulphuric acid anhydride o & Sulphurous anhydride Turpentine, spirits of 110 105 169-4 135 118 108-4 124-4 106 106-6 122 448 326 15 50 -10 160 230 221 336-9 275 244-4 227-1 255-9 222-8 223-8 251-6 838 618-8 59 122 14 320 FUSING POINTS. Centigrade. Fahrenheit. Aluminium 700 Antimony 432 Asphalt 100 Bismuth 260 Boric acid 186 Brass 900 Bromine 22 Bronze 900 Cadmium 316 Cobalt 1500 Colophonium 135 Copper 1100 Cupric chloride 498 Cuprous chloride 434 Fat, oxen 40 sheep 42 pig 27 Fluorspar 902 Glass 1200 Glass containing lead 1000 1292 809 212 500 367 1652 -7-6 1652 600 2732 275 2012 928 813 104 107-6 80-6 1655 2192 1832 APPENDIX. FUSING POINTS continued. 267 Centigrade. Fahrenheit. G>ld 1075 Ir >n, cast, white 1075 ,, grey 1275 ,, wrought 1550 I. dine 113 L;ad 326 oxide 954 chloride 498 IV agnesium 500 ft ercury 39 ft ercuric chloride 293 > aphthalene 79 Mckel 1500 lilmoil 29 I araffiu 45-60 1 latinum 1775 I itch (coal tar) 150-200 1 hosphorus 44 1 otassium chlorate 359 iodide 634 carbonate 834 nitrate .. .. 329 Hearicacid 70 Heel 1375 Silver, metallic 960 ,. chloride 451 nitrate 217 trontium chloride 825 Selenium 217 Sodium chloride 772 sulphate 861 nitrate 316 chlorate 302 carbonate 814 Spermaceti 45-50 Thallium 290 Tin 230 Wax, bee's 62-70 Zinc 412 1967 1967 2327 2822 235 '4 618 1749 928 932 -38-2 560 174-2 2732 84-2 113-140 3227 3JO-400 111-2 678 1173 1533 624 158 2507 1760 843-8 422 1517 422 1421 1581 600 575 1497 113-122 554 446 143-158 773 268 THE GAS ENGINEER'S LABORATORY HANDBOOK. SPECIFIC GBAVITT OP NITRIC ACID AT 60 F. (15 C.), CALCULATED FBOM KOLB'S RESULTS. (Lunge and Hurter.) Degrees Twaddell. Specific Gravity. Percentage 1 by Weight. Grammes per Litre. Degrees Twaddell. Specific Gravity. Percentage by Weight. Grammes per Litre. HN0 3 N 2 5 HN0 3 N 2 5 HN0 3 N 2 5 HN0 3 N 2 e 1 1-005 -88 75 8-8 7-5 41 1-205 33-02 28-30 397-9 341-1 2 1-010 1-75 1-50 17-6 15-1 42 1-210 33-79 28-96408-8 350-4 3 015 2-62 2-25 26-6 22-8 43 1-215 34-55 29-61419-8 359-9 4 020 3-50 3-00 35-7 30-6 44 220 35-32 30-27430-9 369 '3 5 025 4-37 3-75 44-8 38-4 45 225 36-08 30 -93' 442-0 378-9 6 030 5-25 4-50 54-0 46-3 46 230 36-85 31-58453-2 388-5 7 035 6-12 5-25 63-3 54-2 47 235 37-61 32-24 464-5 398-1 8 040 6-95 5-95 72-3 62-0 48 240 38-38 32-90 475-9 407-9 9 045 7-77 6-66 81-2 69-6 49 245 39-15 33-55 487-4 417-8 10 1-050 8-59 7-36 90-2 77-3 50 1-250 39-91 34-21 498 '8 427-6 11 1-055 9-41 8-07 99-3 85-1 51 1-255 40-68 34-86 510-5 437-6 12 1-060 10-23 8-77 108-4 97-3 52 1-260 41-44 35-52522-1 447-6 13 1-065 11-06 9-48117-8 101-0 53 1-265 42-21 36-18532-8 456-6 14 1-070 11-88 ilO-18127-l 108-9 54 1-270 42-97 36-83 545-7 467-8 15 1-075 12-70 10-89136-5 117-0 55 1-275 43-74 37-49 557-7 478 16 1-080 13-52 ill-59 146-0 125-1 56 280 44-50 38-15 569-6 488-2 17 1-085 14-34 12-29155-6 133-4 57 285 45-27 38-80581-7 498-6 18 1-090 15-16 12 99165-2 141-6 58 290 46-04 39-46593-9 509-1 19 1-095 15-98 13-70175-0 150-0 59 295 46-80 40-11606-0 519-5 20 1-100 16-80 U'40184-8 158-4 60 300 47-57 40-77618-4 530-1 21 1-105 17-57 J15-06 194-1 166-4 61 305 48-33 41-43630-7 540-6 22 1-110 18-35 15-72203-7 174-6 62 310 49-10 42-08643-2 551-3 23 1-115 19-12 16-39213-2 182-7 63 1-315 49-86 42-74655-6 562-0 24 1-120 19-89 17.05222-7 190-9 64 1-320 50-63 43-40668-3 572 -8 25 1-125 20-67 17-71 232-5 199-3 65 1-325 51-40 44-06681-0 583-8 26 1-130 21-44 18-38242-3 207-7 66 1-330 52-24 44-78694-8 595-6 27 1-135 22-21 19-04252-0 216-0 67 1-335 53-09 45-51708-7 607-5 28 1-140 22-98 |19-70262-0 224-6 68 1-340 53-94 46-24722-8 619-6 29 1-145 23-76 20-36272-0 233-2 69 345 54-79 46-96736-9 631-6 30 1-150 24-53 ; 21-03282-1 241-8 70 350 55-64 47-69751-1 643-8 31 1-155 25-30 21-69292-2 250-4 71 355 56-53 48-45 766-0 656-6 32 1-160 26-08 22-35302-5 259-3 72 360 57-42 49-22 780-9 669-4 33 1-165 26-85 23-01312-8 268-1 73 365 58-31 49-98 795-9 682-2 34 1-170 27-62 23-68>323-l 276-9 74 37059-21 50-75 811-2 695-3 35 1-175 28-40 24-34 333-7 286-0 75 375 60-10 51-51 826-4 708-4 36 1-180 29-17 25-00 344-2 295-0 76 380 61-00 52-29 841-8 721-6 37 1-185 29-94 25-66354-8 J304-1 77 1-385 61-97 53-12858-3 S735-8 38 1-190 30-71 26-33365-4 313-2 78 1-39062-95 53-96875-0 750-0 39 1-195 31-49 26-99376-3 322 -6 79 1-39563-92 54-79891-7 764-2 40 1-200 32-26 27-65387-1 331-8 80 1-400 64-90 55-66908-6 778-8 I APPENDIX. 269 i- PECTFIO GRAVITY OP NITRIC ACID AT 60 F. (15 C.) continued. Percentage' Grammes Percentage Grammes Degrees Specific by Weight. per Litre. Degrees Specific by Weight. per Litre. Twaddell. * ravity. HN0 3 N 2 5 HN0 3 N 2 5 Twaddell. jrfitvity. HN0 3 N a O HN0 3 N 2 5 81 ; 1-40566-01 56 '61 927 -4 794-9 94 470 82-9071-05 1219 1045 82 1-410 67- 12 57-56946-4 811-2 95 475 84-2872-24 1243 1066 83 1-41568-23 58-51965-4 827-4 96 480 85-6673-43 1268 1087 84 1-42069-34 59-46984-6 844-0 97 485 87-0574-61 1293 1108 85 1-42570-45 60-39 1004 860-6 98 490 88-43 75-80 1318 1130 86 1-43071-83 61-57 1027 880-4 90 495 89-8276-98 1343 1151 87 1-43573-21 62-76 1050 900-0 100 500 91-2078-18 1368 1173 88 1-44074-59 63-94 1074 920-6 1 101 505 92-6679-43 1394 1195 89 1-44575-98 65-13 1097 940-4 102 510 94-1380-69 1421 1218 90 1-45077-36 66-31 1122 961-8 103 1-515 95-5981-94 1448 1241 91 1-45578-75 67-50 1146 982-4 104 1-520 97-0683-20 1475 1264 92 1-46080-13 68-68: 1170 1003 105 1-525 98-5384-46 1502 1287 93 1-465.81-52 69-87 1195 1824 106 1-530 100-00 85-71 1530 1311 PERCEX 'AGE OF PURE HC1 GRAMMES PER LITRE 60 F. (15 C.), CAI CULATED FROM KOLB'S EssuLTS. (Lunge and Hurter, Alkali M( ker's Pocket Book.) Degrees Twaddell. Specific Gravity. Percentage of HC1. Grammes per Litre. Degrees Twaddell. Specific Gravity. Percentage of HC1. Grammes per Litre. 1 1-005 1-12 11-32 21 1-105 21-06 232-68 2 1-010 2-12 21-45 22 1-110 22-06 244-80 3 1-015 3-12 31-67 23 1-115 23-05 257 02 4 1 020 4-11 41-99 24 1-120 24-05 269-34 5 1-025 5-11 52-41 25 1-125 25-05 281-76 6 1-030 6-11 62-93 26 1-130 26-04 294-28 7 1-035 7-10 73-55 27 1-135 27-04 306-90 8 1-040 8-10 84-27 28 1-140 28-04 319-62 9 1 045 9-10 95-09 29 1-145 29-03 332-44 10 1 050 10-09 106-01 30 1-150 30-03 345-36 11 1 055 11-09 117-02 31 1-155 31-03 358-34 12 1-060 12-09 128-14 32 160 32-02 371-44 13 1-005 13-08 139-36 33 165 33-02 384-64 14 1-070 14-08 150-68 34 170 34-02 397-94 l.'i 1 075 15-08 162-10 35 175 35-01 411-34 16 1 080 16-07 173-62 36 180 36-01 424-84 17 1 085 17-07 185-24 37 1-185 37-01 438-44 18 1-01)0 18-07 196-96 38 1-190 38-01 452-14 19 1095 19-07 208-78 39 1-195 39-00 466-00 20 1-100 20-06 220-70 40 1-200 40-00 479-84 270 THE GAS ENGINEEB'S LABOBATOBY HANDBOOK. TABLE OF THE PERCENTAGES OF EEAL SULPHURIC ACID (S0 4 H 2 ) CORRESPONDING TO VARIOUS SPECIFIC GRAVITIES OF AQUEOUS SULPHURIC ACID. Bineau ; Otto. Temp. 15. Specific Gravity. Per cent. Specific Gravity. Per cent. Specific Gravity. Per cent. 1-8426 100 1-568 66 1-2476 33 1-842 99 1-557 65 1-239 32 1-8406 98 1-545 64 1-231 31 1-840 97 1-534 63 1-223 30 1-8384 96 523 62 1-215 29 1-8376 95 512 61 1-2066 28 1-8356 94 501 60 1-198 27 1-834 93 490 59 1-190 26 1-831 92 480 58 1-182 25 1-827 91 469 57 1-174 24 1-822 90 4586 56 1-167 23 1-816 89 448 55 1-159 22 1-809 88 438 54 1-1516 21 1-802 87 428 53 1-144 20 1-794 86 1-418 52 1-136 19 1-786 85 1-408 51 1*129 18 1-777 84 1-398 50 1-121 17 1-767 83 1-3886 49 1136 16 1-756 82 1-379 48 106 15 1-745 81 1-370 47 098 14 1-734 80 1-361 46 091 13 1-722 79 1-351 45 083 12 1-710 78 342 44 0756 11 1-698 77 333 43 068 10 1-686 76 324 42 061 9 1-675 75 315 41 1-0536 8 1-663 74 306 40 1-0464 7 1-651 73 2976 39 1-039 6 1-639 72 289 38 1-032 5 1-627 71 281 37 1-0256 4 1-615 70 272 36 1-019 3 1-604 69 1-264 35 1-013 2 1-592 68 1-256 34 1-0064 1 1-580 67 APPENDIX. 271 TAB uE OP SPECIFIC GEAVITT AND PERCENTAGE OP CAUSTIC POTASH IN AQUEOUS SOLUTION. Sp gr. at i > C. Per- centage of KOH. Sp. gr. at 15 C. Per- centage of KOH. Sp. gr. at 15 C. Per- centage of KOH. Sp. gr. at 15 C. Per- centage of KOH. 1 309 1 1-166 19 374 37 1-590 54 317 2 1-177 20 387 38 1 604 55 J25 3 1-188 21 400 39 1-618 56 :)33 4 1-198 22 412 40 1-630 57 341 5 1-209 23 425 41 1-642 58 1)49 6 1-220 24 438 42 1-655 59 .358 7 1-230 25 450 43 1-667 60 065 8 1-241 26 462 44 1-681 61 074 9 1-252 27 475 45 1-695 62 083 10 1-264 28 488 46 1-705 63 1 092 11 1-276 29 499 47 718 64 1 101 12 1-288 30 511 48 729 65 1 110 13 1-300 31 1-525 49 740 66 1 119 14 1-811 32 1-539 50 754 67 1 128 15 1-324 33 1-552 51 7H8 68 1 137 16 1-336 34 1-565 52 780 69 1 146 17 1-349 35 1-578 53 790 70 1 155 18 1-361 36 T.BLE OP SPECIFIC GEAVITT AND PERCENTAGE OP SODA (Na 2 0) IN AQUEOUS SOLUTION. Tunnermann. Sp.gr. Per cent. Sp.gr. Per cent. Sp.gr. Per cent. Sp.gr. Per cent. 1-4285 30-220 1-3198 22-363 1 2392 15-110 1042 7-253 1-4193 29-616 1-3143 21-894 1-2280 14-500 0948 6-648 1-4101 29-011 1-3125 21-758 1-2178 13-901 0855 6-044 1-4311 28-407 : 1-3053 21-154 1-2058 13-297 0764 5-440 3923 27-802 l 1-2982 20-550 1 1948 12-692 0675 4-835 3836 27-200 i 1-2912 19-945 1-1841 12-088 0587 4-231 3751 26-594 1-2843 19-341 1-1734 11-484 0500 3-626 3668 25-989 1-2775 18-730 1-1630 10-879 0414 3-022 3586 25-385 i 1-2708 18-132 1-1528 10-275 1-0330 2-418 3505 24-780 1-2642 17-528 1-1428 9-670 1-0246 1-813 3426 24-176 1-2578 16-923 1-1330 9-066 1-0163 1-209 3349 23-572 1-2515 16-379 1-1233 8-462 1-0081 0-604 3273 22-967 1-2453 15-714 1-1137 7-857 1-0040 0-302 , Of rut r UNIVERSIE1 272 THE GAS ENGINEER'S LABORATORY HANDBOOK. SPECIFIC GRAVITY OF CAUSTIC AMMONIA CONTAINING DIF- FERENT QUANTITIES OF NH 3 , TEMPERATURE 14: C. (57 P.). (Caritw.) Specific Gravity. Percentage NH 3 Specific Gravity. Percentage NH 3 Specific Gravity. Percentage NH 3 0-8844 36-0 0-9021 28-2 0-9239 20-4 0-8848 35-8 0-9026 28-0 0-9245 20-2 0-8852 35-6 0-9031 27-8 0-9251 20-0 0-8856 35-4 0-9036 27-6 0-9257 19-8 0-8860 35-2 0-9041 27-4 0-9264 19-6 0-8864 35-0 0-9047 27-2 0-9271 19-4 ', 0-8868 34-8 0-9052 27-0 0-9277 19-2 , 0-8872 34-6 0-9057 26-8 0-9283 19-0 0-8877 34-4 0-9063 26-6 0-9289 18-g 0-8881 34-2 0-9068 26-4 0-9296 18-6 0-8885 34-0 0-9073 26-2 0-9302 18-4 0-8889 33-8 0-9078 26-0 0-9308 18-2 0-8894 33-6 0-9083 25-8 0-9314 18-0 0-8898 33-4 0-9089 25'6 0-9321 17-8 0-8903 33-2 0-9094 25-4 0-9327 17-6 0-8907 33-0 0-9100 25-2 0-9333 17-4 0-8911 32-8 0-9106 25-0 0-9340 17-2 0-8916 32-6 0-9111 24-8 0-9347 17-0 0-8920 32-4 0-9116 24-6 0-9353 16-8 0-8925 32-2 0-9122 24-4 0-9360 16-6 0-8929 32-0 0-9127 24-2 0-9366 16-4 0-8934 31-8 0-9133 24-0 0-9373 16-2 0-8938 31-6 0-9139 23-8 0-9380 16-0 0-8943 31-4 0-9145 23-6 0-9386 15-8 0-8948 31-2 0-9150 23-4 0-9393 15-& 0-8953 31-0 0-9156 23-2 0-9400 15-4 * 0-8957 30-8 0-9162 23-0 0-9407 15-2 0-8962 30-6 0-9168 22-8 0-9414 15-0 0-8967 30-4 0-9174 22-6 0-9420 14-8 0-8971 30-2 0-9180 22-4 0-9427 14*6 0-8976 30-0 0-9185 22-2 0-9434 14-4 0-8981 29-8 0-9191 22-0 0-9441 14-2 0-8986 29-6 0-9197 21-8 0-9449 14-0 0-8991 29-4 0-9203 21-6 0-9456 13-8 0-8996 29-2 0-9209 21-4 0-9463 13-6 0-9001 29-0 0-9215 21-2 0-9470 13-4 0-9006 28-8 0-9221 21-0 0-9477 13-2 0-9011 28-6 0-9227 20-8 0-9484 13-0 0-9016 28-4 0-9233 20-6 0-9491 12-0 APPENDIX. SPECIFIC GKAVITY OF CAUSTIC AMMONIA continued. 273 ' specific < ravity. Percentage NH 3 Specific Gravity. Percentage NH 3 Specific Gravity. Percentage NH 3 9498 12-6 0-9654 8-4 0-9823 4-2 9505 12-4 0-9662 8-2 0-9831 4-0 9512 12-2 0-9670 8-0 0-9839 3'8 9520 12-0 0-9677 7-8 0-9847 3'6 9527 11-8 0-9685 7-6 0-9855 3-4 9534 11-6 9693 7-4 0-9863 3-2 9542 11-4 9701 7-2 0-9873 3-0 9549 11-2 0-9709 7-0 0-9882 2-8 -9556 11-0 0-9717 6-8 0-9890 2-6 -9563 10-8 0-9725 6-6 0-9899 2-4 9571 10-6 0-9733 6-4 0-9907 2-2 9578 10-4 9741 6-2 0-9915 2-0 9580 10-2 0-9749 6-0 0-9924 1-8 95!)3 10-0 0-9757 5-8 0-9932 1-6 9601 9-8 0-9765 5-6 0-9941 1-4 9608 9-6 0-9773 5-4 0-9950 1-2 9616 9-4 0-9781 5-2 0-9959 1-0 9623 9-2 0-9790 5-0 0-9967 0-8 9631 9'0 0-9799 4-8 0-9975 0-6 9639 8-8 0-9807 4-6 0-9983 0-4 >-9647 8-6 0-9815 4-4 0-9991 0-2 SOLUBILITY OF AMMONIA IN WATER BY WEIGHT. 1 GRAMME WATER DISSOLVES AT 760 MM. PRESSURE. (Eoscoe and Dittmar.') Temperature C- ntigrade. Grammes NH 3 Absorbed. Temperature Centigrade. Grammes NH 3 Absorbed. Temperature Centigrade. Grammes NH 3 Absorbed. o o o 0-875 20 0-526 40 0-307 2 0-833 22 0-499 42 0-290 4 0-792 24 0-474 44 0-275 6 0-751 26 0-449 46 0-259 8 0-713 28 0-426 48 0-244 10 0-679 30 0-403 50 0-229 12 0-645 32 0-^82 52 0-214 14 0-612 34 0-362 54 0-200 16 0-582 36 0-343 56 0-185 18 0-554 38 0-324 274 THE GAS ENGINEEK'S LABOKATOKY HANDBOOK. CONVERSION OF DEGREES ON THE CENTIGRADE THERMOMETER INTO DEGREES OF FAHRENHEIT'S SCALE. Cent. Fahr. Cent. Fahr. Cent. Fahr. Cent. Fahr. o o o - 50 -58-0 - 6 21-2 38 100-4 82 179-6 - 49 -56-2 - 5 23-0 39 102-2 83 181-4 -48 -54-4 -4 24-8 40 104-0 84 183-2 -47 - 52-6 -3 26-6 41 105-8 85 185-0 -46 - 50-8 - 2 28-4 42 107-6 86 186-8 -45 -49-0 - 1 30-2 43 109-4 87 188-6 - 44 -47-2 32-0 44 111-2 88 190-4 - 43 -45-4 + 1 33-8 45 113-0 89 192-2 -42 - 43-6 2 35-6 46 114-8 90 194-0 -41 -41-8 3 37-4 47 116-6 91 195-8 - 40 -40-0 4 39-2 48 118-4 92 197-6 - 39 -38-2 5 41-0 49 120-2 93 199-4 -38 -36-4 6 42-8 50 122-0 94 201-2 -37 - 34-6 7 44-6 51 123-8 95 203-0 - 36 - 32-8 8 46-4 52 125-6 96 204-8 - 35 - 30-0 9 48-2 53 127-4 97 206-6 - 34 - 29-2 10 50-0 54 129-2 98 208-4 - 33 -27-4 11 51-8 55 131-0 99 210-2 - 32 -25-6 12 53-6 56 132-8 100 212-0 -31 -23-8 13 55-4 57 134-6 101 213-8 - 30 -22-0 14 57-2 58 136-4 102 215-6 - 29 - 20-2 15 59-0 59 138-2 103 217-4 - 28 - 18-4 16 60-8 60 140-0 104 219-2 - 27 - 16-6 17 62-6 61 141-8 105 221-0 - 26 -14-8 18 64-4 62 143-6 106 222-8 -25 13-0 19 66-2 63 145-4 107 224-6 - 24 -11-2 20 68-0 64 147-2 108 226-4 - 23 - 9-4 21 69-8 65 149-0 109 228-2 - 22 - 7-6 22 71-6 66 150-8 110 230-0 - 21 - 5-8 23 73-4 67 ,52-6 111 231-8 - 20 - 4-0 24 7-v2 68 154-4 112 233-6 - 10 - 2-2 25 77-0 69 156-2 113 235-4 - 18 - 0-4 26 78-8 70 158-0 114 237-2 - 17 + 1-4 27 80-6 71 159-8 115 239-0 - 10 3-2 28 82-4 72 161-6 116 240-8 - 15 5-0 29 84-2 73 163-4 117 242-6 - 14 6-8 30 86-0 74 165-2 118 244-4 - 13 8-6 31 87-8 75 167-0 119 246-2 - 12 10-4 32 89-6 76 168-8 120 248-0 - 11 12-2 33 91-4 77 170-6 121 249-8 - 10 14-0 34 93-2 78 172-4 122 251-6 - 9 15-8 35 95-0 79 174-2 123 253-4 - 8 17-6 36 95-8 80 176-0 124 255-2 - 7 19'4 37 98-6 81 177-8 125 257-0 APPENDIX. 275 CONVERSION OP DEGREES, ETC. continued. Cei t. Fahr. Cent. Fahr. Cent. Fahr. Cent. Fahr. c o o o 126 258-8 175 347-0 224 435-2 273 523 4 Ii7 260 6 176 348-8 225 437-0 274 525-2 iLS 262-4 177 350-6 226 438-8 275 527-0 ll!) 264-2 178 352-4 227 440-6 276 528-8 11:0 266-0 179 354-2 228 442-4 277 530-6 i; i 267-8 180 356-0 229 444-2 278 532-4 i: 2 269-6 181 357-8 230 446-0 279 534-2 i: 3 271-4 182 359-6 231 447-8 280 536-0 i: t 273-2 183 361-4 232 449-6 281 537-8 i: .) 275 184 363-2 233 451-4 282 539-6 IJ'6 276-8 185 365-0 234 453-2 283 541-4 i: 7 278-6 186 366-8 235 455-0 284 543-2 l: 8 280-4 187 368-6 236 456-8 285 545-0 l: i) 282-2 188 370-4 237 458-6 286 546-8 1 284-0 189 372-2 238 460-4 287 548-6 1 1 285-8 190 374-0 239 462-2 288 550-4 1 -2 287-6 191 375-8 240 464-0 289 552-2 1 :-) 289-4 192 377-6 241 465-8 290 554-0 1 i 291-2 193 379-4 242 467-6 291 555-8 1 5 293-0 194 381-2 243 469-4 292 557-6 1 (i 294-8 195 383-0 244 471-2 293 559-4 17 296-6 196 384-8 245 473-0 294 561-2 1-iS 298-4 197 386-6 246 474-8 295 563-0 149 300-2 198 388-4 247 476-6 296 564-8 150 302-0 199 390-2 248 478-4 297 566-6 151 303-8 200 392-0 249 480-2 298 568-4 152 305-6 201 393-8 250 482-0 299 570-2 153 307-4 202 395-6 251 483-8 300 572-0 i5i 309-2 203 397-4 252 485-6 301 573-8 155 311-0 204 399-2 253 487-4 302 575-6 153 312-8 205 401-0 254 489-2 303 577-4 157 314-6 206 402-8 255 491-0 304 579-2 158 316-4 207 404-6 256 492-8 305 581-0 lot) 318-2 208 406-4 257 494-6 306 582-8 160 320-0 209 408-2 258 496-4 307 584-6 1G1 321-8 210 410-0 259 498-2 308 586-4 162 323-6 211 411-8 260 500-0 309 588-2 103 325-4 212 413-6 261 501-8 310 590-0 161 327-2 213 415-4 262 503-6 311 591-8 1(55 329-0 214 417-2 263 505-4 312 593-6 16S 330-8 215 419-0 264 507-2 313 595-4 167 332-6 216 420-8 265 509-0 314 597-2 163 334-4 217 422-6 266 510-8 315 599-0 1C,,) 336-2 218 424-4 267 512-6 316 600-8 173 338-0 219 426-2 268 514-4 317 602-6 171 339-8 220 428-0 269 516-2 318 604-4 172 341-6 221 429-8 270 518-0 319 606-2 173 313-4 222 431-6 271 519-8 320 608-0 174 345-2 223 433-4 272 521-6 276 THE GAS ENGINEER'S LABORATORY HANDBOOK. SOLUBILITY OF SULPHUR IN CARBON BISULPHIDE AT 15 C. (59 P.). Specific Gravity. Per cent. S. Specific Gravity. Per cent. S. Specific Gravity. Per cent. S. 1 1 Specific Gravity. Per cent. S. Specific Gravity. Per cent. S. 1-271 1-296 6-0 1-321 12-1 1-346 18-1 371 25-6 1-272 0-2 297 6-3 322 12-3 1-347 18-4 372 26-0 1-273 0-4 298 6-5 323 12-6 1-348 18-6 373 26-5 1-274 0-6 299 6-7 324 12-8 1-349 18-9 374 26-9 1-275 0-9 300 7'0 325 13-1 1-350 19-0 375 27-4 276 1-2 301 72 326 13-3 1-351 19-3 376 28-1 277 1-4 302 7-5 327 13-5 1-352 19-6 377 28-5 278 1-6 303 7'8 328 13-8 1-353 19-9 378 29-0 279 1-9 304 8'0 329 14-0 1-354 20-1 379 29-7 280 2-1 305 8-2 330 14-2 1-355 20-4 380 30-2 281 2*4 '306 8-5 331 14-5 1-356 20-6 381 30-8 282 2-6 307 8-7 332 14-7 1-357 21-0 382 31-4 283 2-9 308 8-9 333 15-0 1-358 21-2 383 31-9 284 3-1 309 9-2 334 15-2 1-359 21-5 384 32-6 285 3-4 310 9-4 335 15-4 1-360 21-8 385 33-2 286 3 6 311 9-7 336 15-6 1-361 22-1 386 33-8 287 3-9 312 9-9 337 15-9 1-362 22-3 387 34-5 288 4-1 313 10-2 338 16-1 1-363 22-7 388 35-2 289 4-4 314 10-4 339 16-4 1-364 23-0 389 36-1 290 4-6 315 10-6 340 16-6 1-365 23-2 390 36-7 291 4-8 316 10-9 341 16-9 ' 1-366 23-6 1-391 37-2 292 5-0 317 11-1 342 17-1 1-367 24-0 293 5-3 318 11-3 343 17-4 1-368 24-3 294 5-6 319 11-6 344 17-6 1-369 24-8 Saturated. 1-295 5-8 320 11-8 -345 17-9 1-370 25-1 WEIGHTS AND MEASURES. The corresponding values of the French and English weights and measures are here given. The use of the French or decimal system is strongly recommended by its extreme simplicity. The smaller denominations are obtained by taking a tenth, hundredth, thousandth, &c., of the unit chosen; they are designated by the Latin prefixes deci-, centi-, milli-, &c. The higher denominations are 10 times, 100 times, 1000, &c., times the unit, and are named by the Greek prefixes deca-, hecto-, kilo-, &c. Examples of this will be found in the Tables given below. The starting-point of the French system is the " metre" (= 39 '37 inches) ; this is the " unit of length." The " unit of measure " is the " litre," which is 1 cubic decimetre. The " unit of weight "is the gram, which is the weight of 1 cubic centimetre of distilled water at 4 C. APPENDIX. 277 The chief conveniences arising from the use of this system are : L That all the different denominations can be written as one ; since they are either multiples by ten, or are decimal fractions, of the unit. Thrs 5 decagrams, 3 grams, 4 decigrams, 8 milligrams would be wri ten 53 '408 grams. .j. That since 1 cubic centimetre of water at 4 C. weighs 1 gram, we may obtain the weight of water to be used from the measure, by sin) ply converting the measure into cubic centimetres ; the number thin obtained will represent at once the corresponding weight of water in r,rams. Of coarse this conversion js strictly accurate only when the water is measured at 4 C. But for ordinary purposes the error int oduced, when the water is at the temperature of the air, is too sm ,11 to be of any importance in the preparation of solutions. The weights and measures most frequently used for chemical pur- poi- ?s are the gram, the millimetre (mm.), the litre, and the cubic cei timetre (c.c.) which is -^^Q of a litre. ENGLISH WEIGHTS AND MEASURES. Apothecaries Weight. Avoirdupois Weight. 11 . oz. drms. scruples, grains lb. oz duns. grains. J = 12 = 96 = 288 = 5760 1 = 16 = 256 = 7<'00 1 = 8 = 24 = 480 1 = 16 = 437-5 1 = 3 = 60 1 = 27-343 1 = 20 Imperial Measure. gallon. pints. fluid oz. fluid drms. 1 = 8 = 160 = 1280 1 = 20 = 160 1 = 8 1 gallon = 70,000 grains of water at 16- 7 C. 1 fluid ounce = J n pint = 437-5 M 1'gallori = 277 -280 cubic inches. 1 iJ__ * 1 - . I . (TOO Millimetre Ceitimetre Decimetre Mitre Decametre Hectometre Kilometre Mriometre = 0*001 = O'Ol = 0-1 = 1-0 = lO'O = lOO'O = lOOO'O = 10000 '0 1 inch = 1 foot = 1 fluid ounce = 1-733 FRENCH WEIGHTS AND MEASURES. Measures of Length. inches. mile. fur. yds. ft. = -03937= ........ = -39371 = ........ = 3-937i>8 = ........ = 39-37079=: ...... 3 = 393-70790= .. .. 10 2 = 3937-07900= .. .. 109 1 = 39370-79000= .. 4 213 = 393707-90000 =6 1 156 -0254 metre = 2-5399 centimetres. -3048 inches. 03937 .. -39371 .. 3-9371 3 3-371 2 9-7 1 1 4 10-2 6 1 sq. inch = 6 '45 14 sq. centimetres. 278 THE GAS ENGINEER'S LABORATORY HANDBOOK. FRENOH WEIGHTS AND MEASURES continued. Measures of Capacity. 1 litre = 1 cubic decimetre. litre. cubic inches. Millimetre, or Cubic} Centimetre (c.c.) . . / 001 = 06103 Centilitre . . ight of 1 c.c. of dry hydrogen at C. and 760 mm. = 0*0000896 gram. \V sight of 1 c.c. of dry air at C. and 760 mm. = 0' 0012937 gram. Si scific gravity of hydrogen, air as unit = 0*0693. Si ecific gravity of air, hydrogen as unit = 14*43. \\ eight of 1 c.c. of mercury at C. = 13 * 596 grams. C< efficient of expansion for gases = ^ = * 003665. A -erage percentage of oxygen in air, by volume = 20 '96. PERCENTAGE BY VOLUME, CORRESPONDING TO THE WEIGHT IN GRAINS OF C0 2 PER CUBIC FOOT OF GAS, AT 60 FAHR. AND 30 IN. BAR. J ercentage by Volume. Grains of C0 2 per cubic foot. Percentage by Volume. Grains of C0 2 per cubic foot. Percentage t>y Volume. Grains of C0 2 pei- cubic foot. 0-1 0-817 1-1 8-987 2-1 17-157 0-2 1-634 1-2 9-804 2-2 17-974 0-3 2-451 1-3 10-621 2-3 18-791 0-4 3-268 1-4 11-438 2-4 19-608 0-5 4-085 1-5 12-255 2-5 20-425 0-6 4-902 1-6 13-072 2-6 21-242 0-7 5-719 1-7 13-889 2-7 22-059 0-8 6-536 1-8 14-706 2-8 22-876 0-9 7-353 1-9 15-523 2-9 23-693 1*0 8-170 2-0 16-340 3-0 24-510 280 THE GAS ENGINEER'S LABORATORY HANDBOOK. PERCENTAGE BY VOLUME, CORRESPONDING TO THE WEIGHT IN GRAINS OF NH 3 PER CUBIC FOOT OF GAS, AT 60 FAHR. AND 30 IN. BAR. Percentage by Volume. Grains of NH 3 per cubic foot. Percentage by Volume. Grains of NH 3 per cubic toot. Percentage by Volume. Grains of NH 3 per cubic foot. o-i 0-315 1-1 3-471 2-1 6-627 0-2 0-631 1-2 3-787 2-2 6-943 0-3 0-946 1-3 4-102 2-3 7-258 0-4 1-262 1-4 4-418 2-4 7-574 0-5 1-578 1-5 4-734 2-5 7-890 0-6 .1-893 1-6 5-049 2-6 8-205 0-7 2-209 1-7 5-365 2-7 8-521 0-8 2-524 1-8 5-680 2-8 8-836 0-9 2-840 1-9 5-996 2-9 9-152 1-0 3-156 2-0 6-312 3-0 9-468 SOLUBILITY OF GASES IN WATER (BY VOLUME) AT A PRESSURE OF 760 MM. = 29-92 IN. (Bunsen.) 1 Vol. of Water dissolves atC. Am- monia. Atmo- spheric Air. Carbon Dioxide. Carbon Mon- oxide. Hydro- gen. Nitrogen. Oxygen. Sulphu- retted Hydro- gen. o 1049-6 0-02471 1-7967 0-032870-0193 0-02035 0-04114 4-3706 1 1020- 8 0-02406 1-7207 03207 0193 01981 JO 04007 4 2874 2 993- 3 0-02345 1-6481 0-03131 0-0193 0-01932 0-03907 4-2053 3 967 02287 1 5787|0 03057^0 0193 01884 0381 4-1243 4 941-90-02237 1 -5126 ,0-0298710 -0193 0-01838 0-03717 4-0442 5 917 9 02179 1 4497 02920 ( 0193 01794 03628 3 9652 6 895 02128 1 3901 02857 0193 01752 03554 ! 3 8872 7 873- 10 02080 1 3339 02796 0193U 01713'0 03465 ! 3 8103 8 852 10 02034 1 2809 02739 01 93 01675 03389 3 7345 9 832-0 0-019921-2311 0-02686 0-0193 0-01640 0-033173-6596 10 812-80-019531-1847 02635 0193 01607 03250 3 5858 11 794-3 0-01916 1-1416 0-02588,0- 0193 0-01577 0-03189 3- 5132 12 776 -60- 01882 l-lOlSiO- 02544 0-0193 0-01549 0-03133 3 -4415 13 759-60-01851 1 0653'0 02504 : 0193.0 01523 03082 3 3708 14 743- 10-01822 1-0321 0-02466 0-0193 0-01500 0-03034 3-3012 15 727 -210 -01795 1 0020 02432 0193 01478 02989 3 2326 16 711 -810 -01771 0-9753 0-02402 0-0193 0-01458 0-02949 3 -1651 17 696 9,0 01750 9519^0 02374 0193 01441 0-029143-0986 18 682 -3 0-01732 0-9318 0-02350 0-0193 0-01426 0-02884 3- 0331 19 668 01717 9150 02329 0193 01423 02858 2 9687 20 654-0 0-01704 0-9014 0-02312 0-0193 0-014030-02838 1 2-9053 APPENDIX. 281 VOLUME AND DENSITY OP WATER AT DIFFERENT TEMPERATURES. (Mieaw results of the observations of Kopp, Pierre, Despretz, Hagen, Matthiessen, Weidner, Kremers, and RossettL] T mp. Sp. gr. of Water Volume of Water Sp. gr. of Water (at 4 = 1). Volume of Water (at 4 = 1). 1-000000 i-oooooo 999871 1-000129 1 1-000057 0-999943 999928 1 000072 2 1 000098 999902 999969 1-000031 3 1-000120 999880 999991 1-000009 4 1-000129 999871 1-000000 1 000000 5 1-000119 999881 0-999990 1-000010 6 1-000099 999901 999970 1 000030 7 1-OOU062 999938 999933 1-000067 8 1-000015 999985 999886 1-000114 9 0-999953 1-000047 999824 1-000176 10 999876 1-000124 999747 1-000253 11 999784 1-000216 999655 1 000345 12 999678 1-000322 999549 1-000451 13 999559 1-000441 999430 1-000570 14 999429 1 000572 999299 1-000701 15 999289 1-000712 999160 1 000841 16 999131 1-000870 999002 1-000999 17 998970 1-001031 998841 1-001160 18 91)8782 1-001219 998654 1-001348 19 998588 1-001413 998460 1-001542 20 998388 1-001615 998259 1-001744 21 998176 1-001828 998047 1-001957 22 997953 1 002049 997826 1-002177 23 997730 1-002276 997601 1-002405 24 997495 1-002511 997367 1-002641 25 997249 1 ' 002759 997120 1-002888 26 996994 1-003014 996866 1-003144 27 996732 1-003278 996603 1-003408 28 996460 003553 996331 1-003682 29 1)96179 003835 996051 1 003965 30 995894 004123 995765 1-004253 35 99431 00572 99418 1-00586 40 91(248 00757 99235 1-00770 50 98833 01181 98820 1-01195 60 98351 01677 98338 1-01691 70 97807 1-02243 97794 1-02256 80 97206 1-02874 97194 1-02887 90 96568 1-03554 96556 1-03567 100 95878 1-04300 95865 1-04312 282 THE GAS ENGINEER'S LABORATORY HANDBOOK. COMPARISON OF TWADDELI/S HYDROMETER WITH SPECIFIC GRAVITY. Twaddell. = wi'o Twaddell. c 11 00 & Twaddell. f o> cS o>3 Twaddell. s CO Twaddell. ,s >> 11 P.S o-O Twaddell. "o "> t&o 000 25 1-125 52-6 1-263 81 1-405 110 1-550 141 1-705 1 005 26 1-130 53 265 82 1-410 111 1-555 142 1-710 1-4 007 26-8 1-134 54 270 83 1-415 112 1-560 143 1-715 2 010 27 1-135 54-8 274 84 1-420 112-6 1-563 144 1-720 2-8 014 28 1-140 55 275 84-8 1-424 113 1-565 145 1-725 3 015 28-4 1-142 56 280 85 1-425 114 1-570 146 1-730 4 020 29 1-145 57 285 86 1-430 115 1-575 146-4 1-732 4-4 1-022 30. 1-150 58 290 87 1-435 116 1-580 147 1-735 5 1-025 30-4 1-152 59 295 87-6 1-438 117 1-585 148 1-740 5-8 1-029 31 1-155 59-4 297 88 1-440 118 1-590 149 1-745 6 1-030 32 1-160 60 300 89 445 119 1-595 150 1-750 7 1-035 32-4 1-162 61 305 90 450 119-4 1-597 150-6 1-753 7-4 1-037 33 1-165 61-6 1-308 90-6 453 120 1-600 151 1-755 8 1-040 34 1-170 62 1-310 91 455 121 1-605 152 760 9 1-045 34-2 1-171 63 1-315 92 460 122 1-610 J153 765 10 1-050 35 1-175 64 1-320 93 465 123 1-615 154 770 10-2 1-052 36 1-180 65 1-325 93-6 468 124 1-620 155 775 11 1-055 37 1-185 66 1-330 94 470 125 1-625 156 780 12 1-060 38 1-19H 66-4 1-332 95 475 126 1-630 157 785 13 1-065 39 1-195 67 1-335 96 480 127 1-635 158 790 13-4 1-067 40 1-200 68 1-340 96-6 483 128 1-640 159 1-795 14 1-070 41 1-205 69 1-345 97 485 129 1-645 160 1-800 15 1-075 42 1-210 70 1-350 98 490 130 1-650 161 1-805 16 1-080 43 1-215 71 1-355 99 1-495 130-4 1-652 162 1-810 16-6 1-083 44 1-220 71-4 1-357 99-6 1-498 131 1-655 163 1-815 17 1-085 45 1-225 72 1-360 100 1-500 132 1-660 164 1-820 18 1-090 46 1-230 73 1-365 101 1-505 133 1-665 165 1-825 18-2 1-091 46-2 1-231 74 1-370 102 1-510 134 1-670 166 1-830 19 1-095 47 1-235 75 1-375 103 1-515 134-2 1-671 167 1 835 20 1-100 48 1-240 76 1-380 104 1-520 135 1-675 168 1-840. 21 1-105 48-2 1-241 76-6 1-383 105 1-525 136 1-680 168-4 1 842 21-6 1-108 49 1-245 77 1-385 106 1-530 137 1-685 169 1-845 22 1-110 50 1-250 78 1-390 107 1-535 138 1-690 170 1-850 23 1-115 50-4 1-252 79 1-395 108 1-540 138-2 1-691 171 1-855 23-2 1-116 51 1-255 79-4 1-397 109 1-545 139 1-695 172 1-860 24 1-120 52 1-260 80 1-400 109-2 1-546 140 1-700 173 1-865 To convert degrees Twaddell into specific gravity (water = 1-000) multiply the number of degrees by 5, and add 1-000 to the product. To reduce specific gravity (water = 1-000) to degrees Twaddell: deduct 1 000, and divide the remainder by 5. APPENDIX. 283 BOILING POINTS OF WATER AT DIFFERENT BAROMETRICAL PRESSURES. oiling int F. Barometer Inches. Boiling Poi.it F. Bnrometer Inches. Boiling Barometer Point F. Inches. o 184 16-670 195 21-124 o 206 26-529 185 17-047 196 21-576 207 27-068 186 17-421 197 22-030 208 27-614 187 17-803 198 22-498 209 28-183 188 18-196 199 22-965 210 28-744 189 18-593 200 23-454 211 29-331 190 18-992 201 23-937 212 29-922 191 19-407 202 24-441 213 30-516 192 19-822 203 25-014 214 31-120 193 20-254 204 25-468 215 31-730 194 20-687 205 25-992 216 32-350 TEMPERATURE OF STEAM AT HIGH PRESSURES. Pressure in | Pressure in Pressure in atmospheres Temp. :>f 30 inches Fahr. Atmospheres of 30 inches Temp. Fahr. Atmospheres of 30 inches Temp. Fahr. of Mercury. of Mercury. of Mercury. 1 212-0 8 339-4 15 390-0 2 249-5 9 348-4 16 395-4 3 273-3 10 356-5 17 400-8 4 291-2 11 364-3 18 405-9 5 306-0 12 371-1 19 410-7 6 318-2 13 377-8 20 415-4 7 329-5 14 384-0 284 THE GAS ENGINEER'S LABORATORY HANDBOOK. GRAINS OF SULPHUR (CORRECTED) IN 100 CUBIC FEET OF GAS, CALCULATED FROM BARIUM SULPHATE OBTAINED BY GAS KEFEREES' METHOD. (Sugg.) TABULAR NUMBER. 1 1060 j 1040 1020 1000 980 960 940 920^ g jl ' 11 AEROUTHOMETER NUMBER. 1 943 962 980 1000 1020 1042 1064 1087 1 2-59 2-64 2-70 2-75 2-81 2-86 2-93 2-99 2 3-11 3-17 3-24 3-30 3-37 3-44 3-51 3-59 4 3-63 3-70 3-77 3-85 3-93 4-01 4-10 4-18 6 4-15 4-23 4-31 4-40 4-49 4-58 4-68 4-78 8 4-67 4-76 4-85 4-95 5-05 5-16 5-27 5 38 2 5-19 5-29 5-39 5-50 5-61 5-73 5-85 5-98 2 5-71 5-82 5-93 6-05 6-17 6-30 6-44 6-58 4 6-23 6-35 6-47 6-60 6-73 6-88 7-02 7-17 6 6-74 6-87 7-01 7-15 7-30 7-45 7-61 7-77 8 7-26 7-40 7-55 7-70 7-86 8-02 8-19 8-37 3 7-78 7-93 8-09 8-25 8-42 8-59 8-78 8-97 'r2 8-30 8-46 8-63 8-80 8-98 9-16 9.36 9-57 4 8-81 8-98 9-17 9-35 9-54 9-73 9-95 10-16 6 9-33 9-51 9-71 9-90 10-10 10-30 10-53 10-76 8 9-85 10-04 10-24 10-45 10-66 10-88 11-12 11-36 4 10-38 10-58 10-78 11-00 11-22 11-46 11-70 11-96 2 10-90 11-11 11-32 11-55 11-78 12-03 12-28 12-56 4 11-41 11-63 11-86 12-10 12-31 12-60 12-87 13-16 6 11-93 12-16 12-40 12-65 12-90 13-17 13-45 13-75 8 12-45 12-69 12-94 13-20 13-46 13-74 14-04 14-35 5 12-97 13-22 13-48 13-75 14-03 14-32 14-63 14-95 2 13-49 13-74 14-02 14-30 14-59 14-89 15-21 15-54 4 14-00 14-27 14-56 14-85 15-15 15-46 15-80 16-14 6 14-52 14-80 15-10 15-40 15-71 16-03 16-38 16-74 8 15-04 15-33 15-64 15-95 16-27 16-60 16-97 17-33 6 15-56 15-86 16-18 16-50 16-84 17-18 17-55 17-93 2 16-07 16-38 16-72 17-05 17-40 17-76 18-14 18-53 4 16-59 16-91 17-25 17-60 17-97 18-33 18-72 19-12 6 17-11 17-44 17-79 18-15 18-53 18-90 19-31 19-72 8 17-63 17-97 18-33 18-70 19-09 19-47 19-89 20-32 7 18-16 18-51 18-87 19-25 19-64 20-05 20-48 20-92 2 18-67 19-04 19-41 19-80 ; 20-20 20-62 21-06 21-52 4 19-19 19-57 19-95 20-35 20-76 21-19 21-65 22-12 6 19-71 20-10 20-49 20-90 21-32 21-77 22-23 22-72 8 20-23 20 -C3 21-03 21-45 i 21-88 22-34 22-82 23-31 APPENDIX. GRAINS OF SULPHUR, ETC. continued. 285 TABULAR NUMBER. . 1 1060 1040 1020 1000 980 960 1 940 920^ if AERORTHOMETER NUMBER. !-. 3 ^ 1 943 962 980 1000 1020 1042 1064 1087 V 20-75 21-15 | 21-57 22-00 .22-45 22-91 23-40 23-91 2 21-27 21-68 22-11 22-55 23-01 23-49 23-99 24-51 4 21-79 22-21 22-65 23-10 23-57 24-06 24-57 25-11 6 22-31 22-74 23-19 23-65 24-13 24-63 25-16 25-71 8 22-83 23-27 23-72 24-20 24-69 25-20 25-74 26-30 23-35 23-80 24-26 24-75 25-25 25-78 26-33 26-90 2 23-87 24-33 24-80 25-30 25-82 26-35 26-91 27-50 4 24-38 24-85 25-34 25-85 26-37 26-92 27-50 28-10 6 24-90 25-38 25-88 26-40 26-94 27-50 28-08 28-70 8 25-42 25-91 26-42 26-95 27-50 28-07 28-67 29-29 1 ) 25-94 26-44 26-96 27-50 28-06 28-64 29-25 29-89 2 26-46 26-97 27-50 28-05 28-62 29-21 29-84 30-49 4 26-98 27-50 28-04 28-60 29-18 29-79 30-42 31-09 6 27-50 28-03 28-58 29-15 29-74 30-36 31-01 31-68 8 28-02 1 28-56 29-12 29-70 30-30 30-93 31-59 32-28 1 l 28-53 29-08 29-66 30-25 30-87 31-50 32-18 32-88 2 29-05 j 29-61 30-19 30-80 31-43 32-08 32-76 33-48 4 29-57 \ 30-14 30-73 31-35 31-99 32-65 33-35 34-08 6 30-09 30-67 31-27 31-90 32-55 33-22 33-93 34-67 8 30-61 31-20 31-81 32-45 33-11 33-80 34-52 35-27 12 31-13 31-73 32-35 33-00 33-67 34-37 35-10 35-87 2 31-65 32-26 32-89 33-55 34-23 34-94 35-69 36-47 4 32-17 32-79 33-43 34-10 34-79 35-52 36-27 37-07 6 32-69 33-32 33-97 34-65 35-35 36-09 36-86 37-66 8 33-20 33-85 34-51 35-20 35-91 36-66 37-44 38-26 13 33-72 34-37 35-05 35-75 36-48 37-23 38-03 38-86 2 34-24 :;4-90 35-59 36-30 37-04 37-81 38-61 39-46 4 34-76 35-43 36-13 36-85 37-60 38-38 39-20 40-06 6 35-28 35-96 36-66 37-40 38-16 38-95 39-78 40-65 8 35-80 3;-49 37-20 37-95 38-72 39-53 40-37 41-25 J4 H6-32 37-02 37-74 38-50 39-28 40-10 40-95 41-85 2 36 -8i 37-55 38-28 39-05 39-84 40-67 41-54 42-45 '4 37-35 38-07 38-82 39-60 40-40 41-25 42-12 43-05 6 37-87 38-60 39-36 40-15 40-96 41-82 42-71 43-64 8 38-39 39" 13 39-90 40-70 41-52 42-39 43-29 44-24 15 38-91 39-66 40-44 41-25 42-09 42-96 43-88 44-84 286 THE GAS ENGINEER'S LABORATORY HANDBOOK. TENSION OP AQUEOUS VAPOUR IN INCHES OF MERCURY FROM 1 TO 100 F. Temperature Fahrenheit. Inches of Mercury. Temperature Fahrenheit. Inches of Mercury. Temperature Inches of Fahrenheit. Mercury. o o o 1 046 36 212 71 759 2 048 37 220 72 . 785 3 050 38 229 73 812 4 052 39 238 74 840 5 054 40 247 75 868 6 057 41 257 76 897 7 060 42 267 77 927 8 062 43 277 78 958 9 065 44 288 79 990 10 068 45 299 80 1-023 11 071 46 311 81 1-057 12 074 47 323 82 1-092 13 078 48 335 83 1-128 14 082 49 348 84 1-165 15 086 50 361 85 1-203 16 090 51 374 86 1-242 17 094 52 388 87 1-282 18 098 53 403 88 1-323 19 103 54 418 89 1-366 20 108 55 433 90 1-401 21 113 56 449 91 1-455 22 118 57 465 92 1-501 23 123 58 482 93 1-548 24 129 59 500 94 1-596 25 135 60 518 95 1-646 26 141 61 537 96 1-697 27 147 62 556 97 1-751 28 153 63 576 98 1-806 29 160 64 596 99 1-862 30 167 65 617 100 1-918 31 174 66 639 32 181 67 661 33 188 68 685 34 196 69 708 35 204 70 733 APPENDIX. 287 TENSION OF AQUEOUS VAPOUR FOE EACH TENTH OF A DEGREE CENTIGRADE FROM TO 30 C. (Regnault.^) TCJ .p. C. Tension in mm. of Mercury. Temp. C. Tension in mm. of Mercury. Temp. C. Tension in mm. of Mercury. Temp. C. Tension in mm. of Mercury. < 4-6 3-0 5-7 6-0' 7-0 9-0 8-6 1 4-6 1 5-7 1 7-0 1 8-6 2 4-7 2 5-8 *2 7-1 2 8-7 3 4-7 3 5-8 3 7-1 3 8-7 4 4.7 4 5-8 4 7-2 4 8-8 5 4-8 5 5-9 5 7-2 5 8-9 6 4-8 6 5-9 6 7-3 6 8-9 7 4-8 7 6-0 7 7-3 7 9-0 8 4-9 8 6-0 8 7-4 8 9-0 9 4-9 9 6-1 9 7-4 9 9-1 4-9 4-0 6-1 7-0 7-5 10-0 9-2 1 5-0 1 6-1 1 7-5 1 9-2 2 5-0 2 6-2 2 7-6 2 9-3 3 5-0 3 6-2 3 7-6 3 9-3 4 5-1 4 6-3 4 7-7 4 9-4 5 5-1 5 6-3 5 7-8 5 9-5 6 5-2 6 6-4 6 7-8 9-5 7 5-2 7 6-4 7 7-9 7 9-b 8 5-2 8 6-4 8 7-9 8 9-7 9 5-3 9 6-5 9 8-0 9 9-7 ::-0 5-3 5-0 6-5 8-0 8-0 11-0 9-8 1 5-3 1 6-6 1 8-1 1 9-9 2 5-4 2 6-6 2 8-1 2 9-9 3 5-4 3 6-7 3 8-2 3 10-0 4 5-5 4 6-7 4 8-2 4 10-1 5 5-5 5 6-8 5 8-3 5 10-1 i ; 5-5 6 6-8 6 8-3 6 10-2 7 5-6 7 6-9 7 8-4 7 10-3 8 5-6 8 6-9 8 8-5 8 10-3 9 5-6 9 7-0 9 8-5 9 10-4 288 THE GAS ENGINEER'S LABORATORY HANDBOOK. TENSION OF AQUEOUS VAPOUR continued. Temp. C. Tension in mm. of Mercury. Temp. C. Tension in mm. of Mercury. Temp. C. Tension In mm. of Mercury. Temp. C. Tension in mm. of Mercury. 12-0 10-5 15-0 12-7 18-0 15-4 21-0 18-5 1 10-5 1 12-8 1 15-5 1 18-6 2 10-6 2 12-9 2 15-6 2 18-7 3 10-7 3 12-9 3 15-7 3 18-8 4 10-7 4 13-0 4 15-7 4 19-0 5 10-8 5 13-1 5 15-8 5 19-1 6 10-9 6 13-2 6 15-9 6 19-2 7 10-9 7 13-3 7 16-0 7 19-3 8 11-0 8 13-4 8 16-1 8 19-4 9 11 -1 9 13-5 9 16-2 9 19-5 13-0 11-2 16-0 13-5 19-0 16-3 22-0 19-7 1 11-2 1 13-6 1 16-4 1 19-8 2 11-3 2 13-7 2 16-6 2 19-9 3 11-4 3 13-8 3 16-7 3 20-0 4 11-5 4 13-9 4 16-8 4 20-1 5 11-5 5 14-0 5 16-9 5 20-3 6 11-6 6 14-1 6 17-0 6 20-4 7 11 ? 7 14-2 7 17-1 7 20-5 8 11-8 8 14-2 8 17-2 8 20-G 9 11-8 9 14-3 9 17-3 9 20-8 14-0 11-9 17-0 14-4 20-0 17-4 23-0 20-9 1 12-0 1 14-5 1 17-5 1 21-0 2 12-1 2 14-6 2 17-6 2 21-1 3 12-1 3 14-7 3 17-7 3 21-3 4 12-2 4 14-8 4 17-8 4 21-4 5 12-3 5 14-9 5 17-9 5 21-5 <) 12-4 6 15-0 (.) 18-0 6 21-7 7 12-5 7 15-1 7 18-2 7 21-8 8 12-5 8 15-2 8 18-3 8 21-9 9 12-6 9 15-3 9 18-4 9 22-1 APPENDIX. 289 TENSION OP AQUEOUS VAPOUR, ETC. continued. amp. C. Tension in mm. of Mercury. Temp. C. Tension in mm. of Mercury. Temp. C. Tension in mm. of Mercury. >4'0 22-2 26-0 25-0 28-0 28-1 1 22-3 1 25-1 1 28-3 2 22-5 2 25-3 2 28-4 3 22-6 3 25-4 3 28-6 4 22-7 4 25-6 4 28-8 5 22-9 5 25-7 5 28-9 6 23-0 6 25-9 6 29-1 7 23-1 7 26-0 7 29-3 8 23-3 8 26-2 8 29-4 9 23-4 9 26-4 9 29-6 25-0 23-5 27-0 26-5 29-0 29-8 1 23'7 1 26-7 1 30-0 2 23-8 2 26-8 2 30-1 3 24-0 3 27'0 3 30-3 4 24-1 4 27-1 4 30-5 5 24-3 5 27-3 5 30-7 6 24-4 6 27-5 6 30-8 7 24-6 7 27-6 7 31-0 8 24-7 8 27-8 8 31-2 9 24-8 9 27-9 9 31-4 290 THE GAS ENGINEER'S LABORATORY HANDBOOK, SPECIFIC GRAVITY OF VARIOUS GASES. Name of Gas. Molecular Weight. Experimental Specific Gravity. Air = 1. Weight of 1 cubic foot in grains at 60 F. and 30-0 Bar. Number of cubic feet equal to lib. Hydrogen, H 2 Marsh gas, CH 4 Ammonia, NH 3 Carbonic oxide, CO Olefiant gas, C Z H 4 .. .. Nitrogen, N 2 2 16 17 28 28 28 0-06926 0-558 0-597 0-9678 0-971 0-97137 37-15 297-20 315-77 520-10 520-10 520-10 188-42 23-55 22-17 13-46 13-46 13-46 Air .. 1-000 535-96 13-03 Oxygen, O 2 Sulphuretted hydrogen, SH 2 Carbonic acid, CO 2 Water vapour, H 2 O Bisulphide of carbon, CS 2 . . 32 34 44 18 76 1-10563 1-1912 1-529 0-615 2-640 594-40 631-54 817-30 334-35 1411-70 11-77 11-09 8-56 20-93 4-95 In order to obtain the weight of a cubic foot of any gas whose molecular formula is known, calculate the molecular weight, and multiply the weight of 1 cubic foot of hydrogen (37 '15 grains) by half the molecular weight of the gas ; this will give the weight in grains of 1 cubic foot of the gas. For example, the molecular weight of marsh gas = 16, divide this by 2 = 8, which multiplied by 37-15 = 297 -20 grains, as in the table. When it is not requisite to be strictly accurate, a cubic foot of hydrogen may be taken as weighing 37 grains ; this is a number easily remembered, and by means of the same the weights of different gases are easily calculated. The same rules also apply to the metric system of weights and measures, the standard in this case being the weight of a litre of hydrogen at C. and 760 mm. ; this weighs 08958 gram. APPENDIX. 291 } 'RODUCTS OF THE DESTRUCTIVE DISTILLATION OF COAL. (Mills.) Name. Formula. Boiling Point. Melting Point. 1 Lydrogen H 2 . a C- ? Eethylic hydride .. ^ texylic CH 4 68 t >ctylic C H 119 . ' >ecylic 'araffin CnR 2 n + 2 171 400 (?) Ithylene C 2 H 4 ' - 102-5 'ritylene 'etrylene C 4 H 8 -'5 'entylene 31 texylene C 5 6 Hl 2 71 Leptyleue C 7 H 14 97 Acetylene C 2 H 2 .. ' 'rotonylene C 4 H 6 25 ' 'erene C H tt " lexoylene CgH^ 80 1 tyrolene C 8 H 8 145 .. 'hiophene C 4 H 4 S 84 ' 'hiotoluene C 5 H 6 S 113 r 'hioxene C 6 H 8 S 137 1 5enzene C H 80 5 '17 J 'arabenzene* C 6 H 6 97 '- 'oluene C 7 H 8 111 mf < "rthoxylene 143 1 'aroxylene C H 137 mt Metaxylene cXo 137 Cumene C 9 H 14 166 co 2 C arcoal (wood) C. al, anthracite , bituminous best C ke from ditto C 'al, average quality C ke from ditto O efiant gas 6,152 12,906 2,495 4,478 12,455 15,600 15,504 14,375 13,600 12,800 21,500 6-37 13- 3 1> 2-58 4-63 12-90 16-14 16-05 14-88 14-08 13-25 22-25 ]M irsh , 21,020 24-86 20,272 20-98 P opylene 21,200 21-94 Pi Iphur 4,102 4-25 "NN ood, dry 7,824 8-10 cub. ft. per Ib. C. 'al gas (17 candles) . . 32 227 \\atergas 22-862 P/oducer gas 14*369 Producer water gas .. 13-478 21,696 6,649 1,897 ,983 22-46 6-88 1-96 1-02 The British standard unit of heat (thermal unit) is the amount of IK at required to raise the temperature of 1 Ib. avoirdupois of water l c Fahrenheit. The French standard unit of heat (calorie) is the amount of heat required to raise the temperature of 1 kilogram of water l c Centigrade. 296 THE GAS ENGINEER'S LABORATORY HANDBOOK. WEIGHT OF 1000 CUBIC FEET OP GAS OF DIFFERENT SPECIFIC GRAVITIES AT 60 FAHR., AND 30 INCHES BAB. (Newbigging.) Specific Gravity Air =1-000. Weight per 1000 cubic feet. Specific Gravity Air = 1-000. Weight per 1000 cubic feet. Ibs. Ibs. 380 29-146 515 39-500 385 29-529 520 39-884 390 29-913 525 40-267 395 30-296 530 40-651 400 30-680 535 41-034 405 31-063 540 41-418 410 31-447 545 41-801 415 31-830 550 42-185 420 32-214 555 42-568 425 32-597 560 42-952 430 32-981 565 43-335 435 33-364 570 43-719 440 33-748 575 44-102 445 34-131 580 44-486 450 34-515 585 44-870 455 34-898 590 45-253 460 35-282 595 45-636 465 35-665 600 46-020 470 36-049 605 46-403 475 36-432 610 46-787 480 36-816 615 47-170 485 37-200 620 47-554 490 37-583 625 47-937 495 37-966 630 48-321 500 38-350 635 48-704 505 38-733 640 49-088 510 39-117 645 49-471 INDEX. A SSOKPTION pipettes 221 double 222 n single 221 single, for solid and liquid reagents . . . 222 A -celerated filtration 28 A idimetry 92 A r oven 18 A (kali metals, estimation of, in silicates 197 ,, separation of 198 A Ikalimeter, description of 162 Alkalimetry 94 A lumina, estimation of 54 separation from iron 55, 183 A mmonia, estimation of, in crude gas 154 purified gas 139 solubility of, in water 273 specific gravity, and strength of aqueous solutions of . . 272 A mmoniacal liquor, complete analysis of 172 composition of 159 determination of, the commercial strength of . 161 total ammonia in . . . 172 carbonates in . , . 171, 173 chlorides in . . . . 173 ferro-cyanides in . . . 175 sulphides in . . 169, 170, 173 sulphates in . . . . 174 thiocyanates in . . .174 thiosulphates in ... 174 Ammonium, estimation of, in ferrous ammonium sulphate . . 95 sulphate, analysis of 175 Aqueous vapour, tension of 286, 287, 288, 289 Arsenic, gravimetric estimation of 48 Ash, determination of filter 33 298 THE GAS ENGINEER'S LABORATORY HANDBOOK. PAGE Ash, estimation of, in coal and coke . 114 Assay of coal tar . . 200 Atomic weights, table of 262 BALANCE, description of .... * 2 testing the . . . . . 7 Barium, estimation of 47 Benzene, rectification of 203 Bog ore, analysis of 188 Boiling points of different substances, table of 265 water at different barometrical pressures . . 283 Burette 76 Hempel's 219 Bunte's 244 CADMIUM, estimation of 50 Calcium, estimation of 51 Calibration of burettes 77 flasks 72, 73 Hempel's burette 217 pipettes 76 test mixers 74 Carbon and hydrogen in an organic substance, estimation of . .117 Carbonates, gravimetric estimation of 59 Carbonic acid in ammoniacal liquor, estimation of . 171, 173 in crude gas, estimation of 125, 132 in furnace gases, estimation of 245 in purified gas, estimation of 239 oxide, absorbents for 234 in coal gas, estimation of 248 in furnace gases, estimation of 250 Caustic ammonia, specific gravity and strength of aqueous solutions of 272, 273 ,, potash, specific gravity and strength of aqueous solutions of 271 soda, specific gravity and strength of aqueous solutions of . 271 Centinormal solutions 79 Changes of volume when gases are burnt in oxygen . . . . 293 Coal, elementary analysis of 117 proximate analysis of 113 specific gravity, determination of 124 gas, analysis of, by means of Hempel's apparatus . . . 238 Cochineal as indicator 84 INDEX. 299 PAGB Combustion furnace 118 heats of, with oxygen 295 Con parison of Twaddell's Hydrometer with specific gravity . . 282 Con rersion of thermometric scales 274, 275 Cop ier phosphate for absorption of SH 2 , preparation of . . . 126 Cop >er sulphate, estimation of SO 4 in 43 Cor ection of gases to normal temperature and pressure . . .214 Cry ^tallisation of salts 40 Cuj rous chloride, mode of preparation of 234, 235 Di: ANTATION 26 Dt i normal solutions 79 DC location 16 DC iccator 19 Dt I ructive distillation of coal, products of .... 291,292 Du illation test for determining the value of ammoniacal liquor . 163 Do omite, analysis of 180 E> ^LISH weights and measures 277 Eq livalent weights 262 Er .mann's float 77 Et. .ylene, estimation of, by bromine 237 by fuming H 2 S0 4 236 Ev iporation 20 Ex pansiou of water 281 FA :TOBS for calculating gravimetric analyses 263 Fe -rocyanides in ammoniacal liquor, estimation of . . . . 175 in spent oxide, estimation of 192 Fe rrous ammonium sulphate, estimation of ammonium in . . . 95 Fi- ter ash, determination of 33 Fi ters, incineration of 33 Filter papers, preparation of 24 ., pump 28 Fi iers, use of weighed 38 Fi tration, precautions to be observed in 24 Fire-clay, analysis of 194 French weights and measures 277 Furnace gases, estimation of 244 Fusing points, table of 266 GAS analysis, technical, remarks on 213 .,, by Hempel's apparatus 219 300 THE GAS ENGINEEK'S LABORATORY HANDBOOK. PAGE Gas analysis, Bunte's apparatus . . i. . . 244 Gases, determination of specific gravity of , . . . . 207 specific gravity of various . . . . ' , . 290 Grains of sulphur per 100 cubic feet corresponding to the BaSO 4 obtained by Gas Keferees' method 284 Gravimetric analysis, definition of . , . . . . I determinations, simple . , 40 HAEMATITE ore, analysis of . .... . . . .108 Harcourt's colour test 155 Hempel's gas apparatus . . .219 Hydrochloric acid, normal, preparation of 93 specific gravity and strength of aqueous solutions of 269 Hydrogen, determination of, in coal gas . . . _., . 2 *2 in furnace gases ..... 248 IGNITION of precipitate 33 apart from the filter 36 with the filter 34 Indicators , 84 Iodine, standard solution of 110 Incineration of filter 33, 34, 35, 36 Iron, estimation of, by potassium bichromate 104 permanganate . . . .100 in bog ore 188 in haematite 108 gravimetric estimation of 52 separation from alumina 55, 183 LAWRENCE SMITH'S method for estimation of alkali metals . . 197 Liebig's potash bulbs 120 Lame, data necessary in order to arrive at actual purifying value of . 179 estimation of, volumetrically 179, 180 general notes on 176 impurities in 177 information commonly required concerning . . . .178 obtaining the value of, from its behaviour on slaking . . 178 purifying value of 179 theoretical purifying value of 177 Limestone, analysis of 180 rapid method for the analysis of 186 Litmus as indicator . 85 INDEX. 301 PAGE M; QNESIUM limestone, analysis of 180 estimation of 56 M: nganese dioxide, estimation of, in Weldon Mud .... 103 IV] ; rsh gas, estimation of 237 separation of, from hydrogen 231 M iasurement of gases 214 liquids 70 ]M 3asuring cylinders or test mixers 74 flasks 72 tubes for gases, calibration of 216 vessels 71 IV omoranda, useful 279 1\ elting points, table of 266 ]\ ethyl-orange as indicator 85 IN oisture in coal, determination of 113 3 ortars 15 ? AKAMUKA'S method for determining sulphur in coal . . . 115 J itric acid, specific gravity and strength of aqueous solutions of . 268 } itrogen, properties of 238 determination of, in organic analysis 122 ? "ormal hydrochloric acid 93 oxalic acid 93 sodium carbonate 88 hydrate .93 solutions 81 sulphuric acid 90 volume of gases 214 ORGANIC analysis, determination of carbon and hydrogen . .117 n nitrogen 122 Oven, air 18 water 17 Oxalic acid, normal solution of 93 Oxygen, absorption of, by phosphorus 228 by potassium pyrogallate .... 228 change of volume when bodies are burnt in ... 293 ,, determination of, in furnace gases 247 heats of combustion with 295 properties of 227 302 THE GAS ENGINEERS LABORATORY HANDBOOK. PAGE PALLADIUM, combustion of hydrogen with . , . ..-.* . 231 Phenol-phthaleiu as indicator . . . .... .86 Pipettes . . . ." . . . ' . . ' A . , .' . 74 Phosphorus, alisorption of oxygen by 228 Platinum cone for filter . . . 30 crucibles, precautions to be observed in using ... 38 Potash, caustic, specific gravity and strength of aqueous solutions of 271 Potassium bichromate solution, standard 104 Precipitates, drying of 32 filtration of 23 ignition of 33 apart from the filter . . . . 36, 37 with the filter . . . . 35 washing of 23 Precipitation 22 Percentage by volume corresponding to the weight in grains of ammonia per cubic foot of gas 280 Percentage by volume corresponding to the weight in grains of C0 2 per cubic foot of gas 279 Preparation of pure substances, precipitation 42 crystallisation 40 sublimation 42 Products of the distillation of coal 291 Pure substances by crystallisation 40 precipitation 42 sublimation 42 RECTIFICATION of benzene 203 Referees' ammonia test 138 instructions 251 sulphur test . . \ 138 table for reduction of gases to normal temperature and pressure . 261 Eelationship between the metrical system and English weights and measures .-.-.*. 279 Eider, use of 6 SAMPLING, remarks on 14 Saturation test for determining the value of ammoniacal liquor . 162 Sifting solids, method of . 16 Silica, estimation of, in silicates 57 INDEX. 303 PAGE Silica, estimation of, in fire-clay 195 in limestone 182 S< da lime, preparation of 63 ,, caustic, specific gravity and strength of aqueous solutions of . 271 8 'dium carbonate, normal solution of 88 hydrate, normal solution of 93 8 'lids, powdering of 15 sifting of 16 S >lubility of ammonia in water 273 gases in water 280 sulphur in CS 2 . . * . * . . . . 276 ,, various salts 264 ^ tlution of solids 19 ! pecific gravity of coal, method of determining .... 124 gases, method of determining .... 207 table of the 290 ,, some tar products 293 heats of gases 294 ., ,, solids and liquids 294 > tandard solutions 78, 79, 80 storage of 83 'tarch solution, preparation of Ill ! itorage of standard solutions 83 ! iublimation of arsenic 43 iodine 43 : kilphate of ammonia, analysis of 175 oulphates, estimation of 43 ! Sulphur, determination of, in gas, by Referees' method . . .139 estimation of, in spent oxide 189 in gas, by Harcourt's apparatus . . . 155 solubility of, in CS 2 276 Sulphuretted hydrogen in ammoniacal liquor, estimation of 169, 170, 173 crude gas, estimation of . . . 125, 137 Sulphuric acid, gravimetric estimation of 43 preparation of normal solution of . . .90 standard solution of, for estimation of NH 3 in crude gas .... 154 standard solution of, for estimation of NH 3 in purified gas. . . . 140 standard solution of, for estimation of strength of ammoniacal liquor. , 163 304 THE GAS ENGINEER'S LABORATORY HANDBOOK. PAGE Sulphuric acid, volumetric estimation of ...... 91 specific gravity and strength of aqueous solutions of 270 TAB, assay of coal . .200 Temperature of gas, correction for . . v 214 steam at high pressures 283 Tension of aqueous vapour . . . . . 286, 287, 288, 289 Thermometric scales, conversion of . . . . . 274,275 Thiosulphates, estimation of, in ammoniacal liquor .... 174 Twaddell's hydrometer, comparison of, with specific gravity . . 282 method of determining the value of am- moniacal liquor by 161 TJ TUBE, filling of .......... 61 stoppered . . . ." ; . . . . . 126 Useful memoranda 279 VAPOUR, aqueous, tension of 286, 287, 288, 289 Volatile matter in coal, determination of 114 Volumetric analysis, definition of ........ 1 principles of 68 example of 68 WASH bottle . .. 27 Washing precipitates 27 Water aspirator . 63 bath . 20 expansion of 281 ove* 17 vapour, tension of . > . .. . . . 286,287,288,289 Weighing, operation of 9 rules for . . 10 Weight of 1000 cubic feet of gas of different specific gravities . . 296 Weights, atomic, table of 262 Weights and measures, English 277 French . 278 description of 5 method of testing 9 Weldon Mud, method of determining MnO 2 in .... 103 IjONDON: PKINTBD BT WUJJAM CLOWES AND SON^, LMITBD, 8TAMFOED STREET ANT> (JHAJIDJ& CKOSS.) ADVERTISEMENTS. GRIFFIN & SONS, Supply all the Apparatus, Balances, Chemicals, and Test Solutions required for GRAVIMETRIC ESTIMATIONS and VOLUMETRIC ANALYSIS of COAL GAS, &c. The GAS REFEREES' Apparatus for Sulphur, Ammonia, and Sulphuretted Hydrogen. For t ie Analyses of Ammoniacal Liquor, Coal and Coke, Lime, Spent Oxide, Coal Tar, &c. Also Hempel's Burettes, Absorption Pipettes, and Bunte's Apparatus for TECHNICAL GAS ANALYSIS. The above are fully described in GRIFFIN'S "CHEMICAL HANDICRAFT," SRD EDITION. the mo3t complete & cheapest Catalogue of Modern Apparatus and Reagents extant. Post free, 3g. JOHN J, GRIFFIN & SONS, LTD. 2 GARRICK STREET, COVENT GARDEN, LONDON, W.C. [To follow mailer. X ADVERTISEMENTS. THE GAS METER COMPANY LIMITED, PATENTEES AND MANUFACTURERS OF WET AND DRY GAS METERS. STATION METERS, GOVERNORS, GAS APPARATUS, &C, &c. DRY METER. PREPAYMENT METERS. 1. Valon's Stop Meter. 2. The "Mutual" Boon Penny-in-the-Slot Meter. For Prices and Particulars apply to R. L. ANDREWS, General Manager, 238 KINGSLAND ROAD, LONDON. LONDON, W O R K S : OLDHAM, DUBLIN, ADVERTISEMENTS. 1 1 1 u_r CO LXJ cc. s ^^ Cj <;D CO Lu OUGHT 1 o in Q CO Z rv* 2 CD H z LJ cc UJ LLJ LLJ x 'NOIldlJ: -SE cc Q L CQ 1 O 1 Q g >T CO CQ D OC LLJ | CO UJ Q ^ Qu Z ~ZL D 1 I O CO CO cc -j^ cc: ^ZL H UJ Q. cc *^ UJ ' H H co < o H SE 3 O co" 111 LLJ I CO ^ UJ err iii Q O ID z u. O ^ co" ce LU 5 ? ^ LL O 1 4^ji i CO LJ of UJ LU CO ID Q CO Z cc O 1- H cc 2 cT UJ 3 URIFIEI CO o LIPSE" UJ PLANT "T" "^ ^ o t Z < ^] co" s CO CO O C z cc (_ Z . z cc LU N f ^ O a r UJ Q. C UJ 1 _J Q LU CO* or LLJ Q. CO CO z UJ Q Z 1- z UJ < (__J J '^MB CO < ^ < cc LU O UJ cc DC h O 2 CO ^ co" cc *^C o fe Z LU UJ Q Z DC CO LL CO Q LU LU n I CD CO UJ Q OC ^ [lll _ HE: ^ Q < Q CO Q. .^ o ^^ < Z o ^=- Z _J l^^ 1 co~ "*T" LU *^ UJ -^ ^ ^ o CO LL J_a Lu DC CO "1 *" L O ^_ cr_ZD u. D Q. "T~ J 1 Q ^ ^" O CO 1 fe D Q - < CO ^ QC I CC co ^^ LL ^ _l co" 1 ^ o DC LU CO <: Q. Q. Q_ cT rj^ QC UJ Q z E S UJ < J S O I CO ^ ^ l.^ CO ^ 1 Q ^ Z UJ UJ s J . CO ** rrs CD o ADVERTISEMENTS. JAS. MENZIES & GO. Phoenix Tube Works, Glasgow, MANUFACTURERS OF WJ*OUG)