Freezing- Point, Be iling- Point, ' HPTI v 7 1 T "V MPT -. Jul.l v ii ! j*M;.;l i. HARRY C JOHKS. (XX >oaco^m4m.''>TMVtVJ*i^gtttaMCWWiwvvijca)o^;mp. ...THE... Freezing-Point, Boiling-Point, -AND Conductivity Methods BY HARRY C. JONKS, INSTRUCTOR IN PHYSICAL CHEMISTRY IN JOHNS HOPKINS UNIVERSITY EASTON, PA.: CHEMICAL PUBLISHING CO. 1897 (All rights reserved.} COPYRIGHT, 1897, BY HOWARD HART. PREFACE I have been impressed, in teaching the physical chem- ical methods in the laboratory, with the fact, that there is no readily accessible place in which they are treated satisfactorily from both the standpoint of theory and of practice. In the text-books, the theoretical side is de- veloped, and usually without sufficient attention to the details of manipulation, to enable them to be applied successfully in the laboratory. In the laboratory man- uals, on the other hand, these methods are often treated largely from the mechanical side, and their theoretical bearing might thus be lost sight of. The physical chemical methods, which find most fre- quent application in the laboratory, are probably those based upon the lowering of the freezing-point, and the rise in the boiling-point of a solvent, produced by a dis- solved substance, and the electrolytic conductivity of solutions of electrolytes. It is my chief object in pre- paring this little work to give an account of the operations involved in carrying out these methods in the laboratory. But since the mere mechanical applica- tion of any scientific method is a matter of comparatively little significance, I have aimed to give, also, enough of the theoretical ground on which each of them rests, to enable the student to work with them intelligently, and to see clearly their scientific significance and use. HARRY C. JONES. CONTENTS PART I THE FREEZING-POINT METHOD PAGE Theoretical Discussion i Early History I Work of Raoult i, 2 Molecular Lowering for Different Solvents 3 Molecular Lowering in Aqueous Solutions 4 Theory of Electrolytic Dissociation 5 Calculation of the Molecular Lowering 6, 7 Experimental Verification 8 Calculation of Molecular Weights from Lowering of Freez- Point 8, 9 The Application of the Freezing-Point Method to the Determina- tion of Molecular Weights in Solution 9 The Apparatus of Beckmann 10, 1 1 Carrying out a Determination 11-13 Correction for the Separation of Ice 13, 14 The Application of the Freezing-Point Method to the Measure- ment of Electrolytic Dissociation 14 The Method of Calculating Dissociation from Lowering of Freezing-Point . 15, 16 The Method of Work 16 The Apparatus of Jones 17, 20 Comparison of the Results with the Dissociation from Con- ductivity Measurements 21 PART II THE BOILING-POINT METHOD PAGE Theoretical Discussion 23 Historical 23, 24 Work of Raoult 24, 25 The Relative Lowering of the Vapor-Tension 26 Calculation of Molecular Weights from Lowering of the Va- por-Tension 27 Beckmann's Work on Rise in Boiling-Point 27, 28 Calculation of Molecular Weights from Rise in the Boiling- Point of Solvents 28 Values of the Constants for Solvents 29 Relations between Boiling-Point and Freezing-Point Meth- ods 29, 30 The Application of the Boiling-Point Method to the Determina- tion of Molecular Weights in Solution 30 The Apparatus of Beckmann 3i~33 The Apparatus of Hite 33~35 The Apparatus of Jones 34-36 Carrying Out a Determination 36-39 Correction for Separation of Vapor 39 Results of Measurements 40, 41 PART III THE CONDUCTIVITY METHOD Two Classes of Conductors 42 Electrolytes and Non- Electrolytes 42 Specific Conductivity 43, 44 Molecular Conductivity 44 Dissociation Measured by Conductivity Method 45, 46 Determination of /* 46-50 CONTENTS vii PAGE The Application of the Conductivity Method to the Measure- ment of Electrolytic Dissociation 50 The Apparatus Employed 50-52 Calculation of the Molecular Conductivity 52, 53 Temperature Coefficient of Conductivity 54 The Ostwald Thermoregulator 55 Calibrating the Wire 56-58 Carrying Out a Conductivity Measurement 58 Determination of the Cell Constant 59 Precautions 60 Correction for the Conductivity of Water 60, 61 The Purification of Water 61-63 Substances to be Used 64 PART I THE FREEZING-POINT METHOD Theoretical It has long been known that when a solid is dissolved in a liquid, the freezing-point of the solution is lower than that of the solvent. The first quantitative relation we owe to Blagden, 1 who pointed out that the lowering of the freezing-point of water, produced by different amounts of the same substance, was proportional to the amount of substance present. This same fact was rediscovered much later by Riidorff. 2 A marked ad- vance was made by Coppet, 3 who dealt with comparable, rather than with equal amounts of different substances. He used quantities of different substances which bore to one another the same relation as their molecular weights, and found that such quantities, of substances which are chemically allied, produce very nearly the same lower- ing of the freezing-point of any given solvent. In a word, the lowering of the freezing-point of a solvent by a dissolved substance, is proportional to the number of parts of the substance present. This is about what was known when the problem was taken up by Raoult, and it is to him more than to any other that we owe the present development of the freez- ing-point method. He investigated water solutions of organic compounds, and found that the lowering pro- duced by molecular quantities was very nearly a con- stant. He used other solvents, such as benzene, and 1 Phil. Trans., 78, 277. 2 Pogg. Annalen, 114, 63; 145, 599. 3 Ann. chim. phys., [4], 23, 366. 2 THE FREEZING-POINT METHOD found that comparable quantities of dissolved substances produced the same lowering of the freezing-point. His investigations included nitrobenzene, ethylene bromide, formic and acetic acids, and in each solvent a large number of substances were dissolved. He was thus in a position not only to compare the lowerings produced by different substances in the same solvent, but the low- erings in different solvents. As the result of this work, Raoult attempted the fol- lowing generalization. One molecule of any complex substance dissolved in one hundred molecules of a liquid, lowers the freezing- point of the liquid by nearly a constant amount, which is 0.62. This has been shown not to hold rigidly. When a gram-molecular weight of any substance is dissolved in say 100 grams of a solvent, the lowering of the freezing-point of the solvent is a constant, regard- less of the nature of the substance, provided that there is no aggregation of the molecules of the substance, and no dissociation. This was shown to hold approximately for a large number of substances, and for several solvents by Raoult. 1 The molecular lowering, which is the low- ering produced by a gram-molecular weight of the sub- stance in 100 grams of the solvent, was calculated by him thus : If g grams of the substance are dissolved in 100 grams of the solvent, if m is the molecular weight of the sub- stance, and A the lowering of the freezing-point of the solvent produced by the presence of g grams of the sub- stance, then the molecular lowering is calculated from the formula : Molecular lowering = - . o 1 Ann. chim. phys. [6J, 2, 66. THE FREEZING-POINT METHOD 3 A few results will show the values of the molecular lowering for different solvents. SOLVENT, ACETIC ACID. Molecular lowering. Methyl iodide 38.8 Aldehyde 38.4 Acetone ,, 38.1 Benzoic acid 43.0 Ethyl alcohol 36.4 Acetamide 36. 1 Stannic chloride . - 41.3 Carbon disulphide 35.6 Sulphuric acid . '. 18.6 Hydrochloric acid 17.2 SOLVENT, BENZENE. Methyl iodide 50.4 Anthracene 51.2 Ether 49.7 Acetone 49.3 Chloral 50.3 Stannic chloride 48.8 Methyl alcohol 25.3 Ethyl alcohol . . . . .\ 28.2 Benzoic acid 25.4 SOLVENT, WATER. Methyl alcohol 17.3 Cane-sugar 18.5 Acetamide 17.8 Chloral hydrate 18.9 Milk-sugar 18. i Acetone 17.1 Hydrochloric acid 39.1 Nitric acid 35.8 Sulphuric acid 38.2 Sodium hydroxide 36.2 Potassium chloride 33.6 4 THE FREEZING-POINT METHOD A careful study of these results will bring out some interesting facts.*" The value of the molecular lowering of acetic acid and of benzene is very nearly a constant for each solvent. This is true for a large num- ber of substances of the general type of most of those given above, i. e., non-electrolytes. There are, however, ex- ceptions for these solvents. In the case of acetic acid, there are a few substances known which, like sulphuric acid, give a molecular lowering of only one-half that pro- duced by the non-electrolytes. In benzene there are also a few exceptions, but in this case, the substances which give only half the molecular lowering of the normal, are either non-electrolytes like the alcohols, or weakly dissociated acids like formic, acetic, benzoic, etc. The probable significance of the small molecular lowering produced by some substances is, that they are in a state of molecular aggregation in the particular solvent. When the molecular lowering, in the case of two undissociated compounds dissolved in a given solvent, is twice as great for one as for the other, it means that twice as many mole- cules of the second are aggregated into a unit, as of the first. If the molecules of the one exist singly in solution, those of the second are combined in twos. This will be seen at once, if we remember that the lowering of the freezing-point of a solvent depends only on the relative number of parts of the solvent and of the dissolved sub- stance. When we come to the results with water as a solvent, we have to deal with an entirely new set of phenomena. The results given above are a few taken from a large number. Compounds like the non- electrolytes, give a molecular lowering for water, which is very nearly a constant, and which is approximately 18.8. This is true for such a large number 1 of substances which have * Ann. chim. phys. [5], 28, 137. THE FREKZING-POINT METHOD 5 been investigated that there is no reason for regarding them as being the exceptions. On the other hand, all the strong electrolytes, including the strong acids, strong bases, and salts of strong acids with strong bases, weak acids with strong bases, and weak bases with strong acids, give molecular lowerings which are greater than the value 18.8. The explanation which has been offered to account for this and related facts by Ar- rhenius, 1 is that the molecules of the electrolytes do not exist as such in water solution. They are dissociated into parts called ions, and the amount of such dissociation depends, for a given substance, chiefly upon the amount of water present on the dilution of the solution. In a very dilute solution of a strongly dissociated electrolyte, we have practically no molecules present, only ions. If the molecule is binary, each yields two ions, and since an ion lowers the freezing-point as much as a molecule, the molecular lowering for such substances, at high dilu- tion, is twice as great as where there is no dissociation. If the molecule dissociates into three ions, and the dilu- tion is such that the dissociation is complete, the lower- ing of the freezing-point will be three times as great as where there is no dissociation as with the non-electro- lytes. It is stated above that in aqueous solutions the molecules break down into parts called ions. It is so easy to confuse ions with atoms, and this is so fre- quently done, that a word of caution here is hardly out of place. An ion is not an atom, but is an atom charged with electricity. The resemblance between the two is far less close than might be imagined, except in weight. The properties of many of the atoms could not be foretold from the properties of the ions, with any de- 1 Ztschr. phys. Chem., i, 631. 6 THE FREEZING-POINT METHOD gree of probability. An atom of potassium has proper- ties so different from an ion of potassium, that one is more impressed by their difference than by their re- semblance. In some cases, as with the non-electrolytes, we have then to deal only with molecules in solutions in water, while with the electrolytes we have both molecules and ions, or only ions, depending on the dilution of the solu- tion. That the true value of the molecular lowering for water, when there is neither molecular aggregation nor electrolytic dissociation, is 18.8, has been shown theo- retically by van't Hoff, 1 and more clearly presented by Ostwald, 2 thus : Given a solution which contains n molecules of the dissolved substance and N molecules of the solvent. L,et T be the temperature of solidification of the solvent and d the lowering of its freezing-point. I^et enough of the solvent solidify to dissolve a molecule of the sub- f N \ stance, ( molecules ) . L,et A. be the molecular heat \ n / of fusion of the solvent, the amount of heat set free = N A.. If the ice is now separated from the solution, warmed to the temperature T t fused, and finally allowed to mix with the solution which has also been warmed to the same temperature, by passing through a semi-per- meable membrane, an osmotic pressure p will be ex- erted. If the volume of the solvent which solidified is N v, the work equals pv, the heat A., 1 Ztschr. phys. Chem., i, 481. a I^ehrbuch allgem. Chem., i, 760. THE FREEZING-POINT METHOD But/z; = XT and R= 2 cal. Substituting we have : n A _ . ' N I,et M be the molecular weight of the solvent, and A7 100 - placing N= JT-J- , we have : 2T* 100 In the Raoult formula m = -j- t m is the molecular jrL weight of the dissolved substance, A is the specific low- A ering of the freezing-point, which equals , in which p is the percentage composition,/ of the solution, and C is a constant, n, the number of molecules of the dis- solved substance in 100 grams of the solvent m = - - mA = Cmn. A = Cn From (i) and (2) we have : /-* (2) M 2T* 100 If L is the latent heat of fusion of a gram of the sol- vent, A. = L,M, and n V 2 C ioo L' The absolute temperature T, for the freezing-point of water, is 273, and L, the latent heat of fusion of a gram of water, was taken by van't Hoff as 79. l When these i Ztschr. phys. Chem., i, 497. 8 THE FREEZING-POINT METHOD values are inserted in the above expression, C= 18.9. The value of L is probably more nearly 79.7 when C becomes 18.8. I have shown experimentally 1 that the value of C for water, as determined with solutions of urea, ethyl and propyl alcohols, is respectively : 18.88 18.76 18.77 The formula of van't Hoff applies to the calculation of the constant for any solvent. The freezing-point method has thus two distinct ap- plications : To determine the molecular weight of com- pounds in solution, which are. not dissociated by the solvent ; and to measure the amount of the dissociation of electrolytes in solutions of different concentrations. The applicability of the freezing-point method to the determination of the molecular weights of substances in solution, was pointed out by Raoult. 2 If we represent the unknown molecular weight of a substance by m, the molecular lowering or constant for the solvent by C, and the lowering of one per cent, of the dissolved sub- stance by S, we have C If the weight of the solvent used is W, that of the dis- solved substance w, and the observed lowering of the freezing-point -^, AW o 100 W 1 Ztschr.phys. Chem., 12, 653. 2 Compt. rend., 101, 1056. THE FREEZING-POINT METHOD 9 Substituting this value of 5 in the above expression, it becomes loo Cw sw~- If the constant C is multiplied by 100 and termed C', the expression becomes Cw = ^w- The values of C 1 for a number of the solvents most commonly used are c\ Water 1880 Benzene 4900 6*t; Phenol 75oo Formic acid 2770 Acetic acid 3880 Nitrobenzene 7070 The Applicotion of the Freezing-Point Method to the De- termination of Molecular Weights in Solution Beckmann 1 has devised a form of apparatus which is both simple and efficient. C (Fig. i) is a small glass battery- jar covered with some poorly conducting sub- stance, and which is filled with the freezing material. A mixture of finely powdered ice and salt is convenient. B is a thick- walled glass tube, into which tube A, con- taining the solution, is inserted. A side tube attached to tube A, is thought to be useful in introducing the substance whose molecular weight is to be ascertained, but can readily be dispensed with. The thermometer, of the Beckmann differential type, is fitted into the tube A, by means of a cork, which can be easily removed. The stirrer S passes through the same cork, and must l Ztschr. phys. Chem., 2, 638. 10 THE FREEZING-POINT METHOD be of such form and dimensions as to move freely up and down between the inner walls of the tube and the bulb of the thermometer. A small glass rod, bent at the bot- tom in the form of a ring, which will easily enter the glass tube A, is quite efficient. A short piece of glass tubing, through which this rod will move freely, is forced through a hole in the cork at the top of tube A, and serves both to hold the stirrer in place, and to allow smoother movement through the cork. The apparatus of the follow- ing dimensions has been found in this laboratory to be convenient. Tube A is 20 cm. in length and 3 cm. in width. B is about 15 cm. long and 5 cm. wide. The glass tube used in constructing the stirrer should be about 2.5 mm. in thickness. A ther- mometer with a short, thick bulb, such as is sometimes furnished on the Fig. i. market, is not as desirable as one whose bulb is longer and of smaller diameter, since it requires a longer time to register the temperature of the liquid. In case the solvent used is hydroscopic, some precau- tion must be taken to protect it from the moisture in the air. An apparatus, satisfying this requirement, has been constructed also by Beckmann, 1 by forcing the air which enters the apparatus to pass over some drying agent, like sulphuric acid. The device is shown in 1 Ztschr. phys. Chem., 7, 324. THK FREEZING-POINT METHOD II Fig. 2. The handle of the stirrer E passes through a glass tube, into which the side tube F, containing a few drops of sulphuric acid, is fused. The air enters through this side tube, is dried, and passes out through the tube receiving the handle of the stirrer. The remainder of the appa- ''Ss^ ratus is of exactly the same "Ck form as shown in Fig. i, |j except that it is provided H with a glass siphon H, for removing the melted freez- ing-mixture. This is really superfluous, since a piece of rubber tubing answers the purpose equally well. Forms of apparatus far more accurate than those Fig. 2. just described, have been devised and used, but since such extra refinement is desirable rather to measure dissocia- tion than to determine molecular weights, reference will be given to them under the second application of the freezing-point method. Carrying: Out a Determination The thermometer must first be so adjusted that the freezing-point of water falls near the top of the scale. To accomplish this, water is poured into the tube A, until the bulb of the thermometer, when placed in posi- tion, is covered. Tube A is placed directly in the freez- ing-mixture in C, and the water allowed to freeze. As 12 THE FREEZING-POINT METHOD soon as fine particles of ice separate, tube A is removed from the freezing-mixture, placed in tube B, and the whole then placed again in the freezing-mixture. The thermometer is then raised out of the water containing ice particles, allowed to remain in contact with the warmer air a moment, and then given a sudden jar. The mercury falls from the top to the bottom of the up- per cup, and leaves the column free at its upper end. The thermometer is then placed again in the ice-cold water, and if the end of the mercury column does not come to rest on the upper half of the scale, the process just described is repeated. A few trials generally suffice to bring the reading approximately where desired. The thermometer being adjusted, tube A is carefully dried, closed at the top and side with wooden stoppers, and weighed. Knough pure water is poured into the tube to cover the bulb of the thermometer when in position, and the tube is again weighed. The weight of the sol- vent employed is thus determined. The stopper is then removed from the top of the tube and the thermometer and stirrer placed in position. Tube A is placed in tube B, and the whole system in the freezing-mixture. Dur- ing the cooling of the solvent the stirrer should be raised and lowered frequently. The water will cool down be- low its freezing-temperature often a degree or more, before the ice will begin to separate. When the undercooling of the solvent or of a solution is very much more than a degree, a small fragment of pure ice should be thrown into the overcooled liquid. This will start the separa- tion of ice, which will continue until the true freezing- temperature is reached. When the ice begins to separate, the mercury column will rise, rapidly at first, then slower, until it reaches the point of equilibrium. While the thermometer is THE FREEZING-POINT METHOD 13 rising, and especially when near the point of rest, it must be tapped gently to prevent the mercury from lag- ging back in the capillary, due to friction against its walls. A lead pencil is convenient to use in jarring the thermometer. The freezing-point of the water is then noted on the thermometer. The reading on the ordi- nary Beckmann instrument can easily be made to 0.001 by means of a small pocket lens. The tube containing the solvent, with the thermome- ter and stirrer in position, is removed from the freezing- mixture, and the ice melted, by seizing the tube for a few moments with the hand. The freezing-point of the water is then redetermined exactly as described above. The two determinations should not differ more than two- or three-thousandths of a degree. The substance whose molecular weight is to be deter- mined, is weighed in a weighing tube, poured into the solvent, and brought completely into solution. If cane- sugar is used, that quantity is taken which will give a solution about one-tenth normal. If urea, or any of the alcohols is used, a more concentrated solution may be employed. A solution of cane-sugar, dextrose, etc., more concentrated than one-tenth normal, gives abnormally large depressions of the freezing-point of water. The reason for this is not entirely clear. The solution is then placed in the freezing-mixture and its freezing-point de- termined, and redetermined, exactly as described for the solvent. All the data are thus available for calculating the molecular weight of the substance from the expres- sion already given. Correction for the Separation of Ice A certain amount of the solvent separates in the solid form in all such determinations, and the solution be- 14 THE FREEZING-POINT METHOD comes concentrated by just this amount. The freezing- point of the solution, as read on the thermometer, is there- fore always lower than would correspond to a solution of the concentration originally used. A correction for the change in concentration, due to the separation of the solid solvent, must be introduced. The amount of the solvent which separates in the solid phase, can easily be determined, knowing the amount by which the solution is undercooled before the ice begins to separate, the la- tent heat of fusion of a unit quantity of the solvent, and the specific heat of the liquid. The fraction of the sol- vent which separates is calculated thus, as was pointed out by the present writer : l If we represent by u the amount of the undercooling of the solution in degrees centigrade, by w the latent heat of fusion of unit weight of the solvent, by s the spe- cific heat of the liquid, and by T the fraction which will solidify, we have r=i, w When water is used as a solvent j= i, and a/ =80. The fraction of this solvent which will separate as a solid, for every degree of undercooling, is therefore -g^, and the concentration of the original solution is increased by just so much. Instead of applying the correction to the concentration, it is simpler to apply it directly to the freezing-point lowering itself. The Application of the Freezing-Point Method to the Measurement of Electrolytic Dissociation An ion lowers the freezing-point of a solvent just as much as a molecule. If a molecule dissociates into two ions it will lower the freezing-point of a given amount of 1 Ztschr. phys. Chetn., 12, 624. THE FREEZING-POINT METHOD 15 a solvent, just twice as much as if it is not dissociated. The lowering of the freezing-point of a given sol- vent by a partially dissociated electrolyte, depends upon the relation between the number of molecules of the sol- vent, and the sum of the molecules plus the ions of the dissolved substance. Thus, it is possible for any given dilution, to determine the amount to which an electro- lyte is dissociated. The calculation of the dissociation from the freezing-point lowering is simple. The molecu- lar lowering of the freezing-point of any solvent by any substance was defined by Arrhenius 1 as the lowering produced by a gram-molecular weight of the substance in a liter of solution. This can be taken as approxi- mately one-tenth of the molecular lowering as defined by Raoult. For our present purpose we accept the definition of Arrhenius, and find that the molecular low- ering of water produced by a gram-molecular weight of a non-electrolyte, like urea, the alcohols, etc., in a liter of solution, is the constant 1.88. If the substance used is dissociated, the molecular lowering is always greater than 1.88. The first step is to calculate the molecular lowering for the solution in question, which is done by dividing the lowering found, by the concen- tration in decimal part of normal. If there were only molecules present the molecular lowering would be 1.88. The molecular lowering found must therefore be divided by 1.88, which gives the value of the van't Hoff coeffi- cient 2, for the solution. 8 Molecular lowering . 1.88 If the molecule breaks down into two ions, the percent- 1 Ztschr. phys. Chem., a, 494. ,501. 1 6 THE FREEZING-POINT METHOD age of dissociation, tf 1 (Arrhenius activity coefficient), is expressed thus a = / i . If the molecule breaks down into three ions, i i a = If into n ions, n i The Method of Work Exactly the same apparatus may be used as was em- ployed in the determination of molecular weights. The method of preparing the solutions is, however, some- what different. The solvent is poured into the innermost vessel in quantity large enough to cover the bulb of the thermom- eter, and its freezing-point upon the thermometer ascer- tained, as in a molecular weight determination. The solvent is then completely removed from the vessel and the solution of known concentration, prepared in a measur- ing flask, introduced. Its freezing-point is then deter- mined exactly as previously described, including the rapid stirring, the tapping of the thermometer, the introduc- tion of a fragment of the solid solvent when necessary, and the correction for the change in concentration due to the separation of the solid solvent. The dilution of the solution is then increased one and a half, two, three, four times, etc., and the dissociation determined for each dilution. It will be found that the value of i, and there- fore of of, always increases with increase in dilution. In 1 Ztschr. phys. Chem., n, 535. THE FREEZING-POINT METHOD 17 this work any of the common chlorides, nitrates, bro- mides, or in general any electrolyte may be used. It is convenient to use a solution of pure sodium or potassium chloride of concentration about 0.5 normal, and then to increase the dilution of this solution in several steps, as indicated above. The chlorides and nitrates break down into two ions each, the sulphates into three. The values of a from the freezing-point method should be preserved and compared with the values of OL for the same solutions, as obtained by the conduc- tivity method. Far more accurate experimental methods have been devised and used for measuring the freezing-point lower- ings, by Loomis, 1 Nernst and Abegg, 2 Ponsot, 3 myself, 4 and others. A form of apparatus, which was found by the writer to give excellent results, is sketched in Fig. 3. A is a large metallic vessel, 25 cm. high and 35 cm. wide. This is surrounded by a mantle of non-conducting material to protect it from the warmer air. B is a ves- sel of zinc, 21 cm. high and 15 cm. wide, which rests upon a tripod, to diminish the surface of contact with the outer vessel. This is provided with a lid of zinc. The vessel B is completely surrounded, except above, with a freez- ing-mixture of ice and a little salt. The space be- tween A and B, filled with the freezing-mixture, was covered with the ring of asbestos, aa, to protect the freezing-mixture from the air. C is a glass vessel, 18 cm. high, 10 cm. wide, and of about 1200 cc. capacity. This rests on a thick felt bottom, which protects it from the zinc vessel beneath. The space between B and C is 1 Ber. d. chem. Ges., 26, 797 ; Wied. Annalen, 51, 500. 2 Ztschr. phys. Chem., 15, 681. Compt. rend., 122, 668. * Ztschr. phys. Chem., n, no, 529. i8 THE FREEZING-POINT METHOD filled with air and covered above with a ring of felt, bb, which rests on a metallic shelf fastened on to the inner side of the vessel B. The air-chamber between B and C is thus closed and remains at nearly the same temperature during a determination. The glass ves- sel was covered with a glass lid. D is a thermometer whose bulb is 14 cm. in length and 1.5 cm. in width. The fine capillary was carefully calibrated. The entire scale, which was 22 cm. in length, corresponded to only 0.6. It was divided into tenths, hundredths, and thousandths of a degree. The finest divisions could be estimated to tenths, through a telescope, so that the scale could be THE FREEZING-POINT METHOD read to o.oooi of a degree. The thermometer was of the Beckrnann type, and the freezing-point of the sol- vent could be adjusted upon the scale, wherever desired. It was fastened firmly in cork c, and passed loosely through g, being suspended in the liquid in C. E is a stirrer, which was constructed as follows : A circular piece of sheet-silver was cut somewhat smaller than the glass vessel, and plated electrolytically with gold. This was cut along the circular lines shown in Fig. 4, and also horizontally, as shown. The ends marked o were bent upwards, those marked u down- wards. S is a small hole which received the handle. P is a large hole in the center, through which the bulb of the thermometer passed. In Fig. 5 is given a section / across one of the openings, to show how the ends are cut and bent. This section corresponds to the dotted line a,b in Fig. 4. A stirrer of this form has the advantage that at every 20 THE FREEZING-POINT METHOD movement up and down the liquid is moved horizontally and vertically, and any currents set up during the stroke in one direction are completely reversed by the opposite stroke. The advantage claimed for this method is that by using a large volume of the solution the temperature can be much better regulated. The comparatively thick layer of air at constant temperature, around the inner- most vessel, makes it far less susceptible to the influ- ence of changes in the temperature of surrounding ob- jects. The large volume of the liquid exposes rela- tively less surface to the cooling mixture, and the rate of cooling is comparatively slow. This makes it possible to determine more accurately the temperature of the liquid in which the ice separates. Since the rate of cooling is slow, the ice which separates during the time required for the thermometer to become constant, is relatively small. A liter of pure water is placed in the vessel C, and its freezing-point determined on the thermometer. A cer- tain volume of this is then removed, and an equal vol- ume of a solution of known concentration, added. Thus the volume of the solution in the vessel always remains a liter, which facilitates the calculation of the results. The same process is repeated in making successive di- lutions. By this method of work the first solution of a series is the most dilute, and these become more and more concentrated to the end of the series. Such accuracy as is reached with this apparatus is not absolutely necessary for laboratory practice, but is very desirable where the problem of the measurement of elec- trolytic dissociation presents itself. The applicability of the method will be seen by com- paring the values of the dissociation of a number of THE FREEZING-POINT METHOD 21 acids, bases, and salts, as obtained by it, 1 with the dis- sociation as determined by the conductivity method of Kohlrausch. 2 Concentration Substance. normal. Dissociation from conductivity. Per cent. Dissociation from lowering of freezing-point. Per cent. NaCl o.ooi 98.0 98.4 o.oio 93.5 90.7 o.ioo 84.1 83.5 K 2 SO 4 0.002 92.2 94.1 o.oio 85.8 88.2 o.ioo 70.1 72.0 BaCl 2 0.002 93.9 94.1 o.oio 87.9 88.4 o.ioo 75.3 76.8 HC1 0.002 100. o 98.4 o.oio 98.9 95.8 o.ioo 93.9 88.6 H 2 SO 4 0.003 89.8 86.0 0.005 85.4 83.8 0.050 62.3 60.7 HNO 3 0.002 loo.o 98.4 o.oio 98.5 96.8 o.ioo 93.5 87.8 H 3 PO 4 0.002 87.8 85.2 o.oio 63.5 68.8 KOH 0.002 100.0 98.4 o.oio 99.2 93.7 o.ioo 92.8 83.1 NaOH 0.002 98.9 98.4 o.oio 99.5 93.7 0.050 90.4 88.4 An absolute agreement between the dissociation values obtained by the conductivity method and by the l Ztschr. phys. Chem., 12, 639, Wied. Annalen, 26, 161. 22 THE FREEZING-POINT METHOD freezing-point method, is not to be expected, since the former method was used at 18 or 25, while the latter was applied at. about o. PART II THE BOILING-POINT METHOD Theoretical The presence of a foreign non-volatile substance diminishes the vapor pressure of the solvent in which it is dissolved. Since the boiling-point of a solvent, or of a solution, is the temperature at which the vapor-pres- sure just overcomes the pressure of the atmosphere, it follows that the solution having a lower vapor-pressure than the solvent, will have a higher boiling-point. There are thus two quantities, either of which we may measure : the depression of the vapor-tension of the sol- vent, caused by the presence of the dissolved substance ; or the rise in the boiling-point of the solvent, due to the same cause. Passing over the work of Faraday, 1 Griffiths, 2 Legrand, 3 and others, along this line, since it all fell short of any very important generalization, we come to that of von Babo, 4 who found that the relation between the amount of salt present and the diminution of the vapor-pressure, was independent of the temperature. The work of Wiillner 5 was of greater significance. He measured the depression of the vapor-pressure of water especially by salts, and arrived at the conclusion that the diminution of the vapor-pressure of water, pro- 1 Ann. chim. phys., ao, 324. 2 Pogg. Annalen. 2, 227. * Ann. chim. phys., 5p, 423. * Jahrb. Chem., 1848-49, 93 ; 1857, 72. 5 Pogg. Anualen, 103, 529 ; io5, 85. 24 THE BOILING-POINT METHOD duced by dissolved non- volatile substances, was pro- portional to the amount of substance present. While this is true only in certain cases, or indeed only for certain classes of compounds, yet it is strictly analo- gous to the earliest generalization reached in connection with the study of the depression of the freezing-point of a solvent by a foreign substance. It will be remembered that Blagden stated that the depression of the freezing- point of a solvent by a dissolved substance was propor- tional to the amount of substance present. When depressions of the freezing-point were measured with a fair degree of accuracy, it was shown that the generalization of Blagden held only approximately, and in some cases. So also, the relation pointed out by Wiillner was shown to be only an approximation under certain conditions, by the work of Pauchon, 1 Tammann 8 and Kmden. 3 It is to Raoult, more than to any other, that we owe the theoretical development of the subject in hand. He employed the first of the two quantities men- tioned, and measured the depression of the vapor- tension of a solvent by foreign substances. A number of rela- tions were brought out by him from a study of solutions in solvents other than water, which would not have been discovered in aqueous solutions, since these are often dis- sociated, and to a different amount for different dilutions. Raoult 4 confirmed the generalization of von Babo, that the relation between the depression of the vapor- pressure and the vapor-pressure of the solvent, was in- dependent of the temperature between o and 20. Also that of Wiillner, that the depression of the vapor-pres- 1 Compt. rend., 89, 752. 2 Wied. Annalen, 24, 523. 8 Ibid, 31, 145. * Compt. rend., 103, 1125. THE BOIIvING-POINT METHOD 25 sure was proportional to the concentration (when there was no dissociation) . If we represent the vapor-pressure of the pure solvent by /, and that of the solution by p\ P-P' P is independent of the temperature and proportional to the concentration. The nature of the substance used was then investi- gated to determine whether the chemical composition of the molecule had any effect on its power to depress the vapor-tension. Solutions containing the same number of molecules of different substances in solution in the same number of molecules of a given solvent, must be compared as to their vapor-pressures. This would be difficult to carry out directly. Convenient concentrations of different substances were used, the vapor-pressures of the solu- tions determined, and the molecular depression of the vapor-pressure calculated for each substance from the expression pp'm ~T X T' in which m is the molecular weight of the substance, and g the number of grams in 100 grams of the solvent. He found that the molecular depression of the vapor- pressure of a solvent is a constant for a given solvent, independent of the nature of the substance which is dis- solved in it. This holds, as we now know, only when the dissolved substances are undissociatgcTby the solvent. Raoult 1 investigated also the relative diminution of the vapor-pressure of different solvents, when the rela- 1 Compt. rend., 104, 1430. 26 THE BOILING-POINT METHOD tion between the number of molecules of the substance and that of the solvent was the same. Below are given the results of the relative diminution of the vapor-pressure for twelve solvents, calculated on the basis of one molecule of the substance to 100 mole- cules of the solvent : Water ..................................... 0.0102 Phosphorus trichloride .................... 0.0108 Carbon bisulphide ......................... 0.0105 Tetrachlormethane ........................ 0.0105 Chloroform ................................ 0.0109 Amylene .................................. 0.0106 Benzene ................................... 0.0106 Methyl iodide ............................. 0.0105 Methyl bromide ........................... 0.0109 Ether ..................................... 0.0096 Acetone ................................ o.oioi Methyl alcohol ............................ 0.0103 The relative diminution of the vapor-pressure of sol- vents produced by a^ molecule of a non- volatile sub- stance, in the same number of molecules of the sol- vents, is very nearly a constant. This relation has been satisfactorily formulated thus : pp'_ n p C N+n> in which n is the number of molecules of the dissolved substance, N that of the solvent, and c a constant, which, from the foregoing table, can be regarded as unity. The expression becomes then p p' _ n or the lowering of the vapor-pressure of the solvent is to the vapor-pressure of the solvent, as the number of molecules of the dissolved substance is to the entire THE BOILING-POINT METHOD 27 number of molecules present. This expression was tested experimentally by Raoult, 1 and found to hold for a large number of substances in ethereal solution. From the foregoing it will be seen that the molecular weight of substances can be determined directly from the depression of the vapor-pressure of a solvent which they produce. L,et the molecular weight of the substance be repre- sented by m, and the amount used by a, then the num- ber of molecules, n = . m Making N= i, and substituting this value of n in the expression PP' = " & p N+ n 1 we have fa m= ^~y Knowing/, p' and a, we calculate m directly. As a practical method for determining molecular weights, this is not used, since such measurements are not easily carried out, and are not very accurate. It has been found to be simpler and more accurate to determine the temperature at which the vapor-pressure of the solution is equal to that of the solvent. Since the boiling-point of a liquid is the temperature at which its vapor-pressure just overcomes the pressure of the atmos- phere, the boiling-points of a solvent and of a solution in that solvent, are the temperatures of equal vapor-pres- sures. The object is then to determine the rise in the 1 Ztschr. phys. Chem., 2, 371. 28 THE BOILING-POINT METHOD boiling-point of a solvent produced by the substance dis- solved in it. The experimental method for carrying out such de- terminations we owe to Beckmann. The forms of appara- tus which seem best adapted to this work, and the details of an experiment will be considered in the second part of this chapter. The rise in the boiling-point is directly proportional to the lowering of the vapor-pressure, and depends upon the relative number of molecules of the solvent and of the dissolved substance. The probability of calculating molecular weights directly from the rise in the boiling-point of solvents produced by substances dissolved in them, is at once ap- parent. The expression by which the molecular weights are calculated, is analogous to that already given for calcu- lating molecular weights from lowerings of the freezing- point of solvents by dissolved substances. If we repre- sent the unknown molecular weight by m, the weight of the substance used by w t the weight of the solvent by W, and the rise in the boiling-point of the solvent by R, we have Cw m =. in which C is a constant of different value for each sol- vent. The analogy between the lowering of the freezing- point and the rise in the boiling-point holds still further, in that the value of C can be calculated from the same formula : c=-ill ioo L THE BOILING-POINT METHOD 29 When applied to calculating the constant for the boil- ing-point method of determining molecular weights, T is the absolute temperature at which the pure solvent boils, and L the latent heat of evaporation of the solvent. The values of C for a number of the solvents more commonly used, are given below. The solvents are arranged in the order of their boiling-points. JC. Boiling-point. Ethyl ether -ino 34.9 Carbon bisulphide 2370 46.0 Acetone 1670 56.3 Chloroform 3.660 60.2 Ethyl alcohol 1 150 78.3 Benzene 2770 9 79.6 Water 520 100.0 Acetic acid 2530 118.0 Ethylene bromide 6320 128.8 Aniline 3220 184.5 Some of the relations between the boiling-point and the freezing-point methods have been mentioned, but others, however, exist. It will be remembered that the freezing-point method can be used to determine the molecular weights of only a limited number of sub- stances the non-electrolytes. The boiling-point method is subject to the same limitation. Those substances which give abnormally great depressions of the freezing- point, due to electrolytic dissociation, give abnormally great depressions of the vapor- tension, or rise in the boiling-point. It was pointed out that in such cases the freezing-point method could be used to measure the amount of the dissociation in the solutions. The boil- ing-point method may be used for the same purpose, but is not capable of the same degree of accuracy as the former. 3O THE BOILING-POINT METHOD Some recent work 1 has shown, however, that it can be applied to the problem of electrolytic dissociation in solution, and when all the precautions specified are taken, it is capable of giving fairly satisfactory results. The Application of the Boiling-Point Method to the De- termination of Molecular "Weights in Solution The precautions which are necessary in making such measurements will be understood best by pointing out the more prominent errors to which the boiling-point method is subject. The method as such, is not capable of that refine- ment to which the freezing-point method has been de- veloped. It is sensitive to barometric changes, which seriously affect the boiling-point of liquids. In this method the vapor escapes quickly from the solution in which its presence is necessary to establish the tem- perature equilibrium. The difference in temperature between the liquid and surrounding objects is generally much greater in this method than in the freezing-point method, so that more precautions are necessary to protect the solution and thermometer from changes in the temperature of external objects. In this method a part of the comparatively pure solvent is constantly separating from the solution as vapor, and is returned as a liquid, at a temperature lower than that of the boil- ing solution. The amount of this liquid cannot, for given conditions, be determined with the same degree of accuracy as was possible in ascertaining the amount of ice which separated in the freezing liquid. The large Beckmann thermometers are more liable to undergo change at the comparatively high temperatures to which they are subjected in this method, than in the i Jones and King : Am. Chem. J., 19, 581. THE; BOILING-POINT METHOD 31 freezing-method, where the temperature of the thermom- eter is at no time widely removed from the ordi- nary. The boiling-point method has this advantage that more solvents can be employed, since compara- tively few solvents freeze within the range of ordinary temperatures ; and further, the solubility of substances is generally increased at the higher temperatures. It is, on the other hand, a misfortune for the boiling-point method that aqueous solutions cannot be used satisfac- torily, partly because of the very small constant for water. A number of forms of apparatus for determining the boiling-point of solvents and solutions have been de- vised by Beckmann, 1 to whom we are as much indebted for the experimental development of the subject as we are to Raoult for its theoretical. That one, which, judged by the results, 2 seems to be on the whole, the most satisfactory, is seen in Fig. 6. The inner glass vessel A, provided with a return condenser K, receives the liquid whose boiling-point is to be determined. The bottom of this vessel is filled to a depth of a few centimeters with glass beads or small garnets, so that the boiling may take place from a number of points and proceed more smoothly. The bulb of the Beckmann thermometer is placed well below the surface of the liquid. Tube A is surrounded on the sides with a double- walled glass jacket B, into which some of the same sol- vent placed in A is poured. The object is to surround the boiling liquid with a liquid as nearly at its own tem- perature as possible. This jacket is provided with a return condenser K a . The whole is supported on a box 1 Ztschr. phys. Chem., 4, 544 ; 8, 224 ; 15, 663 ; 21, 246. a Ibid, 8, 224. THE BOIUNG-POINT METHOD of asbestos C, which is open beneath, as shown in the drawing. Heat is applied Fig. 6. The results obtained by Beckmann 1 with the use of this apparatus were very good. It is, however, not free from objections. It is a question whether the effect of radiation from the bulb of the thermometer outward upon the colder objects, was entirely cut off by the form of i Ztschr. phys. Chem., 8, 226. THE BOILING-POINT METHOD 33 jacket employed. Beckmann 1 used in a later form a porcelain jacket, having abandoned a metallic one, which doubtless cut off the radiation more effectively than the one of glass. But an objection which applies to every form devised by Beckmann, is that the cold solvent from the condenser is returned directly into the hot liquid in which the thermometer is immersed. That the thermometer is affected by this, in that it tends to lag behind the true boiling-temperature of the liquid, is probable. A form of apparatus which largely eliminates this latter source of error, was devised by Hite, 2 and is shown in Fig. 7. The distinctive advance made by Hite is the introduction of an inner glass tube, which prevents the condensed solvent from coming in contact with the thermometer before it is reheated to the boil- ing-point. The cooled liquid must pass through a layer of the boiling liquid between the walls of the inner and outer vessel, some centimeters deep, before it can enter the inner tube which receives the thermometer. The inner vessel is closed at the bottom by means of a glass stopper. Grooves are filed into the edge of the stop- per to allow the vapor to stream through into the inner vessel in fine bubbles, and stir the liquid around the thermometer. I am inclined to lay rather less stress upon the importance of this device, than upon the separation of the condensed solvent from the liquid in which the thermometer is placed, until it has been re- heated to the boiling-point. The apparatus gave ad- mirable results with low-boiling solvents, but could not be used for solvents which boil over 100. The present writer" has devised a form of apparatus 1 Ztschr. phys. Chem., 15, 662. 2 Am. Chem. J., 17, 514. 8/fo'rf, 19, 581. 34 THE BOIUNG-POINT METHOD which aims both at reducing the error from radiation to a minimum, and at preventing the condensed solvent from coming in contact with the thermometer until it is reheated to the boiling-point. It is also one of the sim- plest of the efficient forms thus far devised. It is shown in section in Fig. 8. A is a glass tube 18 cm. high and 4 cm. in diameter, drawn out at the top to a diameter of about 2f cm. and ground to receive a ground-glass stop- per. This tube is filled to a depth of from 3 to 4 cm. with glass beads. P is a cylinder of platinum, 8 cm. high and 2^ cm. in width, made by rolling up a piece of platinum foil, and fastening it in position by wrapping it near the top and bottom with platinum wire. Into the cylinder P, some pieces of platinum foil are thrown. These are made by cutting foil into pieces about f cm. square, bending the corners alternately up and down, to prevent them from lying too closely upon one another, and serrating the edges with scissors, to give a greater number of points from which the boiling can take place. The bulb of the thermometer is thus entirely surrounded by metal at very nearly its own temperature, except directly above. A condenser C, about 40 cm. in length, is attached to the tube A,, which is 2 or 2jcrn. in diame- ter, by means of a cork. When it is desired to protect the solvent from the moisture in the air, the top of the condenser tube should be provided with a tube contain- ing calcium chloride or phosphorus pentoxide. During an experiment, the vessel A is closed above by a cork, through which the Beckmann boiling-point thermometer T passes. M is a jacket of asbestos, 12 cm. high and ij cm. thick, over the top of which the rate of boiling can be observed satisfactorily. It is constructed by bending a thin board of asbestos tightly around the tube A, and fixing it in place by means of a copper Fig. 7. Fig. 8. 36 THU BOILING-POINT METHOD wire. Thick asbestos paper is then wound around this until the desired thickness is reached. The apparatus is supported on a small iron tripod S, 8 cm. in diameter, on which rests an asbestos ring R, about 9 cm. in exter- nal diameter. A circular hole is cut in the center of this ring, about 3^- cm. in diameter, and over this is placed a piece of fine copper gauze. The source of heat is a Bunsen burner B, surrounded by an ordinary metallic cone I, to protect the small flame from air- currents. The glass vessel A is shoved down until it comes in contact with the wire gauze. Under these con- ditions a very small flame suffices when low-boiling sol- vents are employed, and not a large flame is required when a solvent like aniline is used. A number of other forms of apparatus have been con- structed for determining the boiling-points of liquids, but these either do not eliminate error sufficiently for accurate work, or are so complex that they can scarcely hope to find general application in the laboratory for the purpose of determining molecular weights. Carrying Out a Determination The thermometer must first be so adjusted that the top of the mercury thread comes to rest on the lower half of the scale, when the bulb is immersed in the boil- ing solvent. This is accomplished by placing some glass beads in cylinder A, and adding the pure solvent until the bulb of the thermometer will be covered when inserted in place. The solvent is then boiled, and as much mercury as possible is driven out of the lower bulb into the upper cup. The thermometer is then removed from the liquid, inverted for a few moments, when still more of the mercury in the bulb will run down into the cup. The thermometer is then quickly brought into THE BOILING-POINT METHOD 37 normal position and given a sudden tap, when the mer- cury will fall from the top to the bottom of the cup and leave the column free. The bulb is again placed in the boiling solvent, and if the thread comes to rest where de- sired, the apparatus is ready for a determination. If not, the process must be repeated until the desired end is reached, which, however, does not usually require any considerable expenditure of time. When the thermometer is adjusted, it must be re- moved, and the apparatus and beads entirely freed from the liquid. The glass beads are then poured into the glass cylinder, the platinum cylinder inserted, and pressed down into the beads to a distance of from to i cm. The platinum plates are then introduced into the platinum cylinder, the end of the tube A, closed with a cork, and the ground-glass stopper inserted into A. The apparatus is then set into a small beaker glass and weighed, the solvent introduced, and the whole re- weighed. /Great care must be taken that not enough solvent is employed to boil over from one side of the platinum cylinder to the other. ) In case a labora- tory is not provided with a balance capable of weighing accurately 200 or 300 grams, the solvent must be weighed directly and poured into the apparatus. This method of procedure, for low-boiling solvents, is necessarily less ac- curate, due to loss by evaporation. After the solvent is weighed, the glass stopper is re- moved, and the thermometer, fitted tightly into a cork, is placed in position, as shown in the drawing. The appa- ratus is then placed upon the stand in the mantle of asbes- tos, the cork removed from A, and the condenser at- tached. Heat is then applied and the solvent boiled. The size of the flame must be so regulated by means of a screw pinch- cock, that the boiling is quite vigorous, 38 THE BOIUNG-POINT METHOD but not so violent as to be of an irregular or explosive character. A quiet, but very active boiling is absolutely essential to the success of the experiment. The time required to establish the true temperature of equilibrium between the pure liquid solvent and its vapor, is much greater than in the case of a solution^ This is strictly analogous to what is observed with the freezing- point method. Here, the time necessary to establish the temperature of the equilibrium between the solid and liquid phases of the pure solvent, is always much greater than for a solution. Before taking a reading on the Beckmann thermometer, it is always necessary to give if a few sharp taps with a lead pencil, and indeed this should be done occasionally while the mercury is rising, and especially when it is near the point of equilibrium. The use of an electric hammer to accomplish this object is an unnecessary complication. A small hand-lens, magnifying a half dozen times, is quite sufficient to use in making the readings. It is always best to redetermine the boiling-point of the solvent. After this point has been ascertained, a tube containing the substance pressed into pellets, whose molecular weight it is desired to determine, is weighed, and a convenient number of these poured into the solvent, either through the condenser, or directly through the tube A when the solvent is not too volatile, and has ceased to boil. The tube is then reweighed, and the amount of substance introduced, thus ascertained. The boiling-point of the solution is then determined. The carrying out of a determination with a low-boil- ing solvent is a much easier process than with one boil- ing at a considerably higher temperature. Thus : when anisol or aniline is employed, much care and some experience are necessary to determine the rate of boiling which must be adopted. If the boiling is too THE BOIUNG-POINT METHOD 39 slow, the thermometer will never reach the temperature of equilibrium ; if so rapid that it is irregular and explo- sive, the thermometer may rise above the true point, and then suddenly drop below it at the moment when a large amount of the vapor is set free. In a word, for high- boiling solvents, the rate of boiling must be as vigorous as possible, in order to proceed with perfect regu- larity. In all such determinations the barometer must be carefully observed ; but after the boiling-point of the sol- vent has been determined, that of the solution can be as- certained so quickly, that the changes in the barometer during this short interval are usually so slight that they are negligible. Whenever they are of appreciable value, a correction must be accordingly introduced. A portion of the nearly pure solvent is constantly be- ing evaporated from the solution, and condensed on the walls of the apparatus itself, and in the condenser. The solution is thus more concentrated than would be calcu- lated from the amount of substance and of solvent used. A correction must be introduced for the amount of the solvent which separates from the solution, as was neces- sary for the freezing-point method. Unfortunately, we cannot determine the amount in the boiling-point method with even the same degree of accuracy as in the freezing-point method. The amount of the solvent which exists as vapor and condensed liquid, is given by Ostwald 1 as 0.2 gram, and 0.35 gram for water. But this evidently holds only under a special set of conditions, and must be taken, in general, as only a rough approximation. Thus all the data are at hand for calculating the 1 Hand und Hilfsbuch zur Ausfiihrung Physiko-Chemischer Messungen, p. 224. 40 THE BOILING-POINT METHOD molecular weight of the substance in the solvent used, from the formula already given (page 28) : Cw Below are given a few of the results obtained with my apparatus for solvents boiling from 34. 9 to 182.5. SOI/VENT, ETHER : k = 2110 ; BOILING-POINT, 34.9 AT 760 mm. Naphthalene, 128. FIRST SERIES. Ether. Naphthalene. Rise in Molecular Grams. Grams. boiling-point. weight. 1 ......... 57-573 1-2365 0-357 126.9 2 ......... 57-573 2.5155 0.716 128.8 3 ......... 57-573 3-8733 1.110 127.9 Mean, 127.9 % SOLVENT, BENZENE : k = 2670 ; BOILING-POINT, 80.36 AT 760 mm. Naphthalene, 128. f Benzene. Naphthalene. Rise in Molecular Grams. Grams. boiling-point. weight. 1 ......... 70.560 0.7594 0.215 133.7 2 ......... 70.560 2.0548 0.574 135.4 3 ......... 70.560 3-0780 0.850 137.0 4 ......... 70.560 44790 1.234 137.4 Mean, 135.9 >" SOLVENT, ANILINE : k = 3220 ; BOILING-POINT, 182.5 AT 738 mm. Triphenylmethane, 2441 FIRST SERIES. Aniline. Triphenylmethane. Rise in Molecular Grams. Grams. boiling-point. weight. 1 ......... 60.126 0.8017 0.180 238.6 2 ......... 60.126 1.6052 0-353 2 43-5 3 ......... 60.126 2.2914 0.496 247.4 4 ......... 60.126 2.9213 0.654 239.2 Mean, 242.2 THE BOILING-POINT METHOD 41 Diphenylamine ', 169. Aniline. Triphenylmethane. Rise in Molecular Grams. Grams. boiling-point. weight. 1 64.220 0.7780 0.224 I74.I 2 64.220 1.3326 0.391 170.9 3 64.220 1.7832 0.535 l6 7-i Mean, 170.7 For practice in the laboratory it is far better to use solvents with low boiling-points, such as ether or ben- zene. Ethyl alcohol can be employed, but with it the results are liable to be less accurate, since its constant is comparatively small. Naphthalene is easily obtained pure, and may be used in both ether and benzene. In alcohol : benzoic acid, urea, or acetamide may be conveniently used ; while tri- phenylmethane, diphenyl amineor benzanilide, give good results with aniline as a solvent. PART III THE CONDUCTIVITY METHOD Conductors of electricity may, for the sake of conve- nience, be divided into two classes, those which conduct without undergoing any decomposition, such as the metals, carbon, etc., and those which, during the passage of the current, undergo a decomposition or electrolysis at the poles, such as solutions of many substances. It is not at all certain that there is any very fundamental difference between the two classes, and at present, it seems that a resemblance between the two modes of con- duction is becoming clearly recognized. It is by no means true that solutions of all substances conduct. Thus, aqueous solutions of the so-called neu- tsal organic compounds, such as the alcohols, carbohy- drates, urea, and a large number of such substances, do not conduct the current. This furnishes ground for a division of substances into those whose solutions con- duct and are called electrolytes, and those which, in solu- tion, do hot conduct and are called non-electrolytes. The application of the conductivity method in phys- ical chemistry is limited to conductors of the second class, i. e., to solutions of electrolytes, which are chiefly solutions of acids, bases, and salts. The conductivity of any conductor of electricity is the reciprocal of its resistance. The resistance r is, from Ohm's law, expressed thus : 7t i THK CONDUCTIVITY METHOD 43 7t is the difference in potential at the two ends of the conductor, and i is the strength of current. The con- ductivity c is the reciprocal of r. i 7t The unit of resistance, called the ohm, is that of a col- umn of pure mercury 106.3 cm. long and i square mm. in section, at o C. The Siemens or mercury unit is the resistance of a column of mercury 100 cm. in length and with a cross section of one square mm. The two units bear the rela- tion to one another of 106.3 : 100. Specific and Molecular Conductivities The resistance of conductors depends upon their form as well as upon their chemical nature. In order that the resistance of different conductors should be meas- ured in comparable quantities, their dimensions must be taken into account. The dimensions usually chosen afe a cylinder i meter in length and i square mm. in section. The resistance of such forms of conductors, is known as their specific resistance. The reciprocal of this is their specific conductivity. f The conductors of the second class are solutions of some electrolyte in some solvent, and their con- ductivity depends chiefly or wholly upon the presence of the electrolytic substance. That the resistances of such solutions should be comparable, it is clear that we must deal with comparable quantities of the dissolved substances. The most convenient quantities are gram- molecular weights. Given a normal solution which contains a gram-molec- ular weight of the electrolyte in a liter. If this liter of 44 THE CONDUCTIVITY METHOD solution be placed between two electrodes which are i cm. apart, the cross section would be 1,000 square centi- meters. This will have o.ooi of the resistance, or 1,000 times the conductivity of a cube of the same solution whose edge was i cm. in length. If we represent by v the number of cubic centimeters of any solution, which contains a gram-molecular weight of the dissolved sub- stance, and by s the specific conductivity of a cube of the solution whose edge is i cm. in length, the molecu- lar conductivity jw is the product of these quantities : fii = vs. But if we represent by s the specific conductivity of a cylinder of the solution i meter in length and i square mm. in cross section : >u = 10,000 vs. A general expression, where g gram-molecular weights are contained in a liter of the solution, is : j X io 3 \JL ' g when s, the specific conductivity, is referred to a cube of the solution, or : ;* = g when 5 is referred to a cylinder of the solution, i meter in length and a square mm. in cross section. The molecular conductivities of solutions are then the conductivities of comparable quantities of different sub- stances, and when the same dilutions are used, the molecular conductivities are directly comparable with one another. Different substances behave very differently with re- THE CONDUCTIVITY METHOD 45 spect to their power to carry the current, when in solu- tion in a given solvent. The fundamental distinction between substances which conduct, and those which do not conduct at all, has been already mentioned. But among conductors very marked differences exist. Some reach a maximum of conductivity at moderate dilution, while others attain this only at extreme dilution. Take the case of a strong acid like hydrochloric or nitric ; the molecular conductivity increases with the di- lution to about one one-thousandth normal, when it be- comes constant. While, on the other hand, the molecu- lar conductivity of a weak acid like acetic, will increase with the dilution, as far as the dilution can be carried with the conductivity method. The question arises, whence this difference between substances in respect to their power to carry the cur- rent? Here again, the theory of electrolytic disso- ciation comes to our aid. Those substances which give abnormally great depressidns of the freezing-point, abnormally large elevations of the boiling-point, and which show abnormally great osmotic pressures, con- duct the current ; and only such substances conduct. The explanation of the abnormal results with respect to the properties just mentioned, was sought in the dis- sociation of the molecules into ions. From a large amount of evidence from many sources, we seem justi- fied in concluding that only ions conduct the current. Molecules are entirely incapable of carrying electricity through the solvent in which they are dissolved. If only ions conduct, then the conductivity of a solution is proportional to the number of ions present, provided that the ions move with the same average velocity, which is true of ions of the same kind. The conductivity method can then be used to meas- 46 THE CONDUCTIVITY METHOD ure the dissociation of electrolytes in solution, and this is its most important scientific application. When the molecular conductivity attains a maximum constant value, it means that the dissociation is complete, and this value of the molecular conductivity is termed fa , The molecular conductivity at any dilution is written /v in which v is the volume of the solution, i. The value of ^ v is determined directly for any electro- lyte in any solvent, by means of the conductivity method. The determination of ^ for strongly disso- ciated electrolytes is comparatively simple. The value of fa, is determined at a given dilution, the dilution in- creased, the molecular conductivity determined at the new dilution, and this continued until a dilution is reached, which is so great, that when further increased, the value of fa remains the same. It has then attained a constant maximum value, which is the value of //,. The value of //< for strong acids and bases, and for THK CONDUCTIVITY METHOD 47 salts, is usually attained at a dilution between v = 500 and v = 5000. This will be seen from the following ex- amples : Hydrochloric acid. Potassium hydroxide. Potassium chlori v. J* v 18. v. /! 18. V. ft, 18. 2 301 2 184.1 2 95-8 32 335 20 204.5 2O 108.3 128 34i 100 212.4 100 II4.7 1000 346 500 214.0 IOOO II9-3 1667 344 1000 211. 1 5000 120.9 A large number of substances, such as the organic acids and bases, which are only weakly dissociated at any ordinary dilution, present a new problem, when it is desired to determine their maximum molecular con- ductivity. That this is not reached at dilutions to which the conductivity method can be applied, is seen from the following examples : Acetic acid. Ammonia. 1 v. fart*. v. /^i8. 2 1.9 2 1.2 20 6.2 20 4.3 ioo 13.2 100 9.2 1000 38.0 1000 26.0 5000 79.6 5000 50.0 10000 99.5 loooo 61.0 It is evident from these results that the value, of J4*> for such substances cannot be determined by the method given for strongly dissociated compounds. The dilution at which complete dissociation would take place lies far beyond the possibility of applying the conductivity method directly. The method of determining the value of /*> for such substances is as follows. While the weak organic acids are only slightly dissociated, salts of these acids are 1 Ammonia is taken, since work on the substituted ammonias has not gen- erally been carried to very great dilutions. 48 TH CONDUCTIVITY METHOD completely dissociated at moderate dilutions. So also, with respect to the weak bases, which, at ordinary dilu- tions, are only slightly dissociated ; their salts are com- pletely dissociated at dilutions which lie well within the range of the conductivity method. Take an organic acid. Its sodium salt is prepared, and the value of /*> for this salt determined ; or taking an organic base, the nitrate of the base is used, and the value of fa for the ni- trate determined. It remains to see what relation exists between the value of //GO for the sodium salt of an acid, and the acid itself, or between the nitrate of a base and the base. Kohlrausch 1 has shown that the value of /*, for any compound is the sum of two constants, the one depend- ing upon the cation, the other upon the anion. The value of /*oo for sodium acetate is the sum of two con- stants, the one for the cation, sodium, and the other for the anion, CH 3 COO. If the constant for sodium be subtracted, the remainder is the constant for the anion of acetic acid. If to this constant the constant for hy- drogen be added, we have the value of /*> for acetic acid itself. Exactly the same line of reasoning applies to the nitrate of the base. The constant for NO 3 is subtracted from //> for the nitrate, and the remainder is the constant for the cation of the base. To this the constant for hydroxyl is added, and the sum is the value of //> for the base. The value of the constant for sodium is 49.2 at 25, and of hydrogen, 325 at 25. If we add 275.8 to the value of /*> for the sodium salt of an acid, we have the value of //5 for the acid. The value of the constant for NO 3 , at the same temperature, is 65. i , and for (OH) , 170. 1 Wied. Ann., 6, 167. THE CONDUCTIVITY METHOD 49 We must, therefore, add 105 to ^ for the nitrate of a base, in order to ascertain // for the base itself. It is thus possible to determine /*< for compounds which are only slightly dissociated at ordinary dilutions. Since p v can always be determined for any electrolyte, we are able to measure the dissociation of compounds, which, even in water, are only slightly dissociated. The application of the conductivity method to meas- ure the exact dissociation in solvents other than water is not always so successful. Water exercises the strong- est dissociating action of any known solvent. The ion- izing power of many solvents is comparatively so weak that it is impossible to determine the value of ^ x for electrolytes, which are strongly dissociated by water, by the direct application of the conductivity method. In such cases, it is possible to determine the dissociation only approximately. The general applicability of any method to measure electrolytic dissociation is of wide-reaching significance. This will appear, when we consider that man}'' chemical reactions take place between ions, molecules as such not coming into play. The chemical activity of solutions is then a function of the dissociation, and since conduc- tivity is a measure of dissociation, there is a close rela- tion between the conductivity of solutions, and their power to react chemically. Indeed, the former has often been used to measure the latter. In this connection is to be mentioned, especially, the work of Ostwald' on the conductivity of the organic acids, from which he calculated their dissociation con- stants. Knowing the dissociation constant, the chem- ical activity of the acid is known. The work of Bredig 2 1 Ztschr. phys. Chem., 3, 170, 241,369. 2 Ibid, 13, 289. 50 THE CONDUCTIVITY METHOD on the conductivity of organic bases is strictly analogous to that just cited. Ostwald 1 has also shown that it is possible to deter- mine the basicity of acids by determining the conduc- tivity of their sodium salts. The conductivity method has also been extensively applied to determine what we have already called the constants for the ions, or the relative velocities with which the ions move through their solutions. A large number of applications of the conductivity method to special problems in dissociation have been made in the last few years, so that it may be said to be one of the most important of all the physical chemical methods. The Application of the Conductivity Method to the Meas- urement of Electrolytic Dissociation When a continuous current is passed through a solu- tion of an electrolyte, the electrodes become quickly covered with gas, or, as we say, become polarized. This increases the resistance to the passage of the current, and interferes with the measurement of the resistance of the solution. Several devices have been proposed for over- coming the effect of polarization, 2 but none have proved as simple as the use of the alternating current. The effect of polarization, tending to retard the flow of the current in one direction, is counterbalanced by the action in the opposite direction, where the polarization current adds itself to the original. This method of measuring the conductivity of solutions we owe to Kohl- rausch. The apparatus employed is sketched diagramatically 1 Ztschr. phys. Chem., i, 105 ; 2, 902. 2 Stroud and Henderson ; Phil. Mag., 43, 19. THE CONDUCTIVITY METHOD 51 in Fig. 9. J is a small induction coil, with only one or two layers of wire. A larger coil must not be used, since it does not give a sharp tone minimum in the tele- phone. The coil, tuned to a very high pitch, should be inclosed in a box surrounded by a poor conductor of W sound, and placed at some distance from the bridge where the reading is to be made. The coil is driven by a storage cell of medium size. A platinum wire, or bet- ter one of manganese alloy, which has a small tempera- ture coefficient of resistance, is tightly stretched over the meter stick AB, which is carefully divided into millime- ters. A rheostat W, whose total resistance amounts to 1 1, 1 10 ohms, is convenient. The resistance vessel R, containing the solution and electrodes, is shown enlarged in Fig. 10. The electrodes are cut from thick sheet platinum, and into each plate a stout platinum wire, about an inch in length, is welded. Glass tubes are sealed on to the platinum wires and electrode plates, by means of sealing glass, as shown in the drawing. These tubes pass tightly through a rubber cap, which fits over the glass vessel. They are filled to a convenient height with mercury, and electrical connection established by means of copper wires, which dip into the mercury. One arm of the telephone T is thrown into the circuit be- THE CONDUCTIVITY METHOD tween the rheostat and the resistance, and the other arm is connected with the bridge wire, by means of a slider. Fig. 10. This is moved along the wire until that point is found at which the hum of the induction coil ceases to be heard in the telephone. I^et this be some point c, and let us represent Ac by , and Br by b, the resistance of the so- lution in the vessel R by r, and the resistance in ohms in the rheostat by w ; then, from the principle of the bridge, we have : ra = wb. _ wb a ' THE CONDUCTIVITY METHOD 53 But the conductivity of a solution c is the reciprocal of the resistance r ; therefore, a c 7. wb The conductivity of solutions, determined by this ex- pression, would not, in any sense, be comparable with one another, since there is nothing in the expression which takes into account the concentration of the solu- tion. It is most convenient to refer all concentrations to the molecular normal, containing a gram-molecular weight of the electrolyte in a liter. If we represent by i) the number of liters which contains a gram-molecular weight of the dissolved substance, the preceding ex- pression becomes : va c = 7. wb Instead of the conductivity c, we write for the molecu- lar conductivity, /*, and to indicate the concentration at which the /* is determined, we write //z/, in which v has the significance indicated above. va But even this expression does not take into account the dimensions of the cell used. A cell-constant C must be introduced, and determined for each cell, before it can be employed for conductivity measurements. The com- plete expression for the molecular conductivity is then Temperature Coefficient of Conductivity The conductivity of solutions of electrolytes increases rapidly with rise in temperature. The molecular con- ductivity of a solution of hydrochloric acid, which is 301.7 at 18, rises to 331 at 25. This is even more marked in the case of sodium sulphate ; a solution hav- ing a conductivity of 94.8 at 18 has a conductivity of 171.4 at 50.3, of 252 at 82, and of 286.4 at 99.4. From this it is evident, that a definite, constant tern- Fig, ii. perature must be carefully maintained in conductivity work. This is accomplished by placing the vessel con- taining the solution in a large volume of water, which is maintained at a constant, known temperature. A con- venient form of thermostat (Fig. n) for such work has been devised by Ostwald. 1 A metallic vessel, containing from 15 to 20 liters of water, is stirred by paddles driven 1 Ztschr. phys. Chem., a, 565. THK CONDUCTIVITY METHOD 55 by a fan, which is kept in motion by means of a small gas jet beneath. The tube near the bottom of the large vessel is filled with a 10 per cent, solution of calcium chloride. The change in volume of this solution with the temperature, is used to regulate the temperature of the water-bath. The Ostwald regulator (Fig. 12) can be easily ad- justed, so that the temperature of the water-bath will re- main constant, to within one-tenth of a degree, fora day. Tube A is connected with the gas supply. The glass tube C, which opens just above the mercury meniscus, Fig. 12. contains a fine perforation in the side, so as to supply gas enough to keep the flame alive, when the lower end of the tube is closed by the mercury. Tube B connects with the burner, and D with the large tube containing the calcium chloride solution, resting on the bottom of the water-bath. When it is desired to adjust the regulator for a defi- nite temperature, the stop-cock is opened, the flame lighted, and a thermometer, divided into tenths of a de- gree, suspended in the bath. The end of tube C is raised above the mercury surface, and the stirrer is set in motion by means of the small gas jet, placed about a 56 THK CONDUCTIVITY METHOD foot below the fans. When the thermometer registers the desired temperature, the stop-cock is closed, and the end of tube C is pushed down until it just touches the mercury surface. The apparatus will then control the temperature automatically. Calibrating the Wire A stout platinum wire can be used in constructing the Wheatstone bridge, but, as already stated, it is better to use one of an alloy of manganese (mangandraht) . This wire is usually of very nearly uniform resistance, but this can never be taken for granted without testing it. A convenient method for calibrating such a wire has been described by Strouhal and Barus. 1 A piece of German-silver wire about a meter and a half in length, is cut into ten pieces (Fig. 13), which are, as nearly as Fig- 13- possible, of the same length. The insulation is removed from the ends of these wires, and they are soldered on to thick copper wires about an inch in length. Nine holes are made in a board, which is about a meter in length, at equal distances apart. These are partly filled with mercury, and receive the ends of the copper wires, which have been previously amalgamated. The board, with the wires in position, is placed along by the side of the bridge wire, and the two end loops attached to the ex- tremities of the bridge. The current from the small in- ductorium is passed through the bridge, and also 1 Wied. Annalen, 10, 326. THK CONDUCTIVITY METHOD 57 through the series of loops. One of the loops is chosen as the standard of measure, and is suitably marked so as to distinguish it from the others. One end of this standard is attached to one end of the bridge, and the other placed in the first mercury cup. One arm of the telephone is placed in the same mercury cup, and the other attached to the pointer, which moves along the bridge wire. The point of silence on the bridge is as- certained. This is the first reading for point i. The telephone and all other connections remaining un- changed, the standard measuring wire, which was at position i, is moved to position 2, and wire 2 is placed in position i. A reading is again made in the tele- phone, which is the second reading for position i. The arm of the telephone, which was in cup i, is then re- moved to cup 2, and the point of silence ascertained. This is the first reading for cup 2. The standard wire, which is now in position 2, is moved to position 3, wire 3 is taken back to 2, and all other connections are un- changed. The point of equilibrium is again ascertained at 2, which gives the second reading for this position. The standard wire is thus interchanged in position with each of the loops, and two readings obtained on the bridge for each position except the last, for which only one reading is available. It must be observed that in all such work in which the telephone is used it is not advisable to try to ascertain directly, the exact point on the wire at which the coil cannot be heard, or at which the tone is a minimum ; but to find a point on each side of the true zero, at which the intensity of the tone is the same. These two read- ings should, at most, be not more than a centimeter apart. The true zero is then just half-way between these points. 58 THE CONDUCTIVITY METHOD The bridge wire is thus divided into ten lengths. The application of the calibration correction is simple. The ten values are added together, and their sum subtracted from 1,000 mm. The difference is divided into 10 parts and each length is corrected by this amount, so that the sum is 1,000 mm. By adding the parts thus, i, 1 + 2, etc., we obtain the points which correspond to tenths of the wire. The difference between these and 10, 20, etc., gives the correction to be applied. Carrying: Out a Conductivity Measurement After the wire is calibrated, the next step is to deter- mine the value of the constant (C), for the cell which is to be used. The preparation of the cell is a matter of some care. In the first place, the electrodes must be placed at a convenient distance apart, by shoving the glass tubes through the ebonite cover, and these must then be fastened firmly in the rubber plate, so that no further movement is possible. If a fairly concentrated solution is to be studied, the plates must be as much as 2, or 2.5 cm. apart. If a very dilute solution is to be used, a distance of 0.5 cm. is sufficient. The ordinary white platinum plates, such as are furnished by the manu- facturers, cannot be used directly, since they would not give a sharp tone-minimum in the telephone. They must be carefully cleansed by washing in chromic acid, and then in water. A few drops of a solution of pla- tinic chloride are poured into the conductivity cell, (Fig. 10) and the cell filled with pure water until the electrodes are covered. A current from a storage battery is then passed through the solution until the electrodes become more and more deeply blackened. The direction of the current should be frequently altered, so that both elec- trodes may become coated, and that the deposit may be THE CONDUCTIVITY METHOD 59 as uniform as possible. After the plates are completely covered with a layer of the platinum black, the platinic chloride is removed from the cell, a little sodium hydrox- ide added, and the current passed through this solution. The object of the alkali is to remove any chlorine which may have been retained by the platinum black as it was being deposited. The sodium hydroxide is then re- moved by hydrochloric acid, and the acid, by repeated washing with pure redistilled water. In order to determine the value of C, in the expression ~ va for any cell, it is necessary to use some solution for which the value of // is known. Since potassium chlo- ride can generally be obtained in a high degree of purity, by five or six crystallizations, it is convenient to use in standardizing the cell. A one-fiftieth normal solu- tion of potassium chloride has a molecular conductivity (/*,) of 129.7 at 2 5 C. The solution is poured into the cell until the electrodes are covered, and brought to ex- actly 25 C. in the thermostat. The bubbles of air which usually separate on the electrodes with rise in temperature, having been removed, a resistance is thrown into the circuit by means of the rheostat, which will bring the point of tone-minimum not very distant from the center of the bridge, say between 400 and 600 mm. Thus all the quantities in the above expression, except C, are known, and it can therefore be solved at once for the value of C. The constant for any given cell being determined, it is a matter of fundamental importance that its value should not be changed. This would be done if the electrodes were moved with respect to one another, or their surfaces in 60 THK CONDUCTIVITY METHOD any wise altered. It is therefore necessary that the electrodes should never be placed upon a hard surface, but always upon clean, thick, filter paper, and the plates must never be touched with any hard object. Knowing the constant for the cell, the measurement of the conductivity of a solution involves exactly the same procedure as that just described. The difference is in the calculation. C is known, and it is desired to find the value of /* for a given solution. The solution is placed in the cell, brought to 25 C., the resistance introduced in the rheostat, and the bal- ance effected on the bridge. All the values in the above expression are now known except }* Vl which is calcu- lated directly. If the solution used is more concentrated than j^- normal, it is better to use the cell whose electrodes are far apart. If more dilute, the electrodes whose dis- tance from one another is not more than 0.5 cm. should be employed. Precautions are necessary at every turn. The wire, after calibration, must never be touched with the hand, and the point of contact with the wire must be moved over its surface very carefully. The current must not be allowed to flow through the resistance coils for any considerable length of time, or the temperature, and therefore the resistance of the coils will change. The inductorium should be allowed to run only during the ac- tual measurement of the resistance. Especial care should be taken that every connection is clean and well made, otherwise resistance will be introduced at the junctions. Correction for the Conductivity of Water Since water is the most general solvent known, and solutions in this solvent have the greatest conductivity, THE CONDUCTIVITY METHOD 6 1 one is called upon, most frequently to measure the conduc- tivity of aqueous solutions. In all such cases the quan- tity actually measured is the sum of the conductivities of the water and of the dissolved electrolyte. The con- ductivity of the water alone, must, in every case, be de- termined, in order that the conducting power of the electrolyte may be ascertained. It would, at first sight, appear to be possible to use water of only a fair degree of purity, to determine its conductivity, and then to sub- tract this from the conductivity of the solution. Whether this could be done, would depend upon the nature of the impurities. They might easily be of such a character as to react chemically with the dissolved electrolyte, and thus seri- ously affect the nature of the solution. Thus, ammonia, which would neutralize any acid, forming a salt, would materially change the nature of the ions present, and therefore the conductivity. Carbon dioxide would, in like manner, affect the conductivity of any strong base. It is therefore necessary, in all work involving the use of the conductivity method, to prepare water in as pure con- dition as is practicable, and then to introduce a correc- tion for its conductivity, when this is larger than the necessary experimental error. Kohlrausch 1 has prepared the purest water thus far obtained, by distilling the purest water obtainable by other methods, in a vacuum. He determined its con- ductivity without exposure to the air, and found it to be 0.04 X 10 ~ 6 . To prepare water of this degree of purity is not practicable, and indeed is not necessary for con- ductivity work. Nernst 2 has suggested fractional crystallization as a 1 Ztschr. phys. Chem., 14, 317. , 8, 120. 62 THE CONDUCTIVITY METHOD means of purifying water for conductivity purposes, but equally efficient and far more rapid methods have been subsequently devised. Hulett 1 has obtained water of a high degree of purity, by distilling it first from potassium bichromate and sul- phuric acid, and then redistilling from a solution of barium hydroxide. The water purified in this way had a conductivity of from 0.7 to 0.8 X io~ 6 . More recently, Jones and Mackay 2 have used an appa- ratus in which the water is distilled first from acid potassium permanganate or acid potassium bichro- mate, which decomposes any organic matter present and retains the ammonia, and second from alka- line potassium permanganate, which retains any carbon dioxide. The apparatus is shown in Fig. 14. Ordinary Fig. 14. distilled water is introduced into the vessel A, together with a little sulphuric acid and potassium per- manganate, or potassium bichromate, through the 1 Ztschr. phys. Chem., 21, 297. 2 Am. Chem. J., 19, 91 ; Ztschr. phys. Chem., aa, 237. THK CONDUCTIVITY METHOD 63 funnel tube G. It is heated to boiling, the vapor passing into B, which contains distilled water, potas- sium permanganate, and a little potassium or sodium hydroxide. A small flame is sufficient to keep the liquid in B at the boiling temperature. The vapor passes from B along the long neck of the retort, over the glass wool, which is meant to arrest any trace of permanganate car- ried along by the steam, into the tin condenser, and is received in the flask E. Certain precautions must be taken in fitting up and using the apparatus. The glass wool W, introduced into the adapter arm C, must fill only the lower part of the arm, otherwise there is dan- ger that a trace of alkali dissolved from it, will be swept over into the condenser. The glass wool should be washed well with hydrochloric acid. Whenever the ap- paratus is cleaned and refilled, which should be done about once a week when in constant use, the distillate collected at first, must be discarded, since it always con- tains a trace of alkali, probably of ammonia, formed by the action of the alkaline permanganate on the organic impurities in the ordinary distilled water introduced into B. When this is once removed, there is no further escape of ammonia possible, since the organic impuri- ties in the water are destroyed by the permanganic acid in A, and the ammonia combines with the sulphuric acid present in that vessel. The carbon dioxide liber- % ated in A, is absorbed by the alkali in B. The process is thus perfectly continuous for at least a week, or it can be interrupted at any time, by removing the burner. Four or five liters of water can be obtained daily with the use of this apparatus. The water purified by this method gave a conductivity at 25, varying from 1.5 to 2.0 X ic" 6 in mercury units. The correction which must be applied to the values of 64 THE CONDUCTIVITY METHOD /*z,, for the conductivity of the water employed in pre- paring the solutions, is calculated by multiplying the specific conductivity of the water by the molecular vol- ume of the solution in cubic centimeters. This quantity, for water properly purified, is negligible for all ordinary concentrations, and attains an appreciable value only in dilute solutions. In case the substance under investi- gation reacts chemically with the impurities in the water, such a correction would be so uncertain that it is better not to attempt to apply it. Substances to be Used In practice, it is well to use some of the same sub- stances whose freezing-point lowerings have been meas- ured, that the dissociation as determined by conductivity, may be compared with that calculated from the depres- sion of the freezing-point. Prepare say a tenth-normal solution of the substance chosen, determine the value of p v for this dilution, increase the dilution to 3-^-, -5-^-, rcinr> oT5inr> an <* T7mnr norm al, determining in each case the value of /v At about -nnnr> J*v will become constant, and will show no further increase with increase in dilution. This is the value of /*< . To find the per- centage of dissociation, <*, at any dilution, divide the value of fa at that dilution, by the value of // for the substance. a = UNIVERSITY OF CALIFORNIA LIBRARY BERKELEY Return to desk from which borrowed. This book is DUE on the last date stamped below. 8 1947 26Mar'58j N LD 21-100m-9,'47(A5702sl6)476 YC 21607 oil '7*