TRE EFFECT OF SOME SIMPLE ELECTROLYTES ETC. 119 XIIT.-The Efect of some Simple Electrolytes on the Temperature of Maximum Density of Wuter. By ROBERT WRIGHT. ROBETTI (Ann. Chim. Phys. 1867 [iv] 10 461; 1869 17 370) has given a fairly exhaustive account of the early work carried out B f 120 WRIGHT THE EFFECT OF SOME SIMPLE ELECTROLYTES on the determination of the temperature of maximum density of water and of a few salt solutions. A considerable portion of this work is clue t o Despretz ( A i ) i 7 ? . C‘him. Phys. 1839 [ii] 70 49; 1840 73 296) and the most important result is embodied in the following law iianied after that investigator “ The lowering of the temperature OP the point of maximum density of water caused by the addition of a solute is directly proportional t o the concen-tration of the latter.” An attempt was made by Rosetti to coiinect t’he lowering of the temperature of maximum density brought about’ by the addition of a solute with the lowering of the freezing point pro’duced by the same cause but it was found impossible t o formulate any general law f o r although the ratio of the two lowerings was csmtant.for any given solute at different’ concentrations still a different ratio was obtained by the use of a second solute. In other words whilst the lowering of the freezing point-being con-iiected with the osmotic pressure o l the solution-depends oaly on the concentration of the solute molecules the lowering of the point of maximum densit-y depends 011 the iiature as well as on the number of dissolved molecules.Coppet in a series of researches (A1212. Cliim. I.’hys. 1894 [vii], 3 246 268; Conzpt. rend. 1897 125 533; 1899 128 1559; 1900, 131 178; 1901 132 1218; 1902 134 lZO8) has determined the molecular lowering of the temperature of maximum density for a number of salts of the alkalis that is the lowering produced by a gram-molecule of salt per litre and the following table contains the more important of his results : TABLE I. Chloride. Bromide. Iodide. Rubidium ............ 11.7 13.2 15.6 Potassium.. ............. 11.6 12.8 15-4 Sodium .................. 13.2 14.5 17.0 Lithium ............... 6.0 7.0 8.3 Ammonium ............ 7.2 8.7 11.1 From an examination of these Lignres Coppet points out’ that of the three acid radicles the iodide has the greatest and the chloride the least effect and as a general conclusion states that’: “Le rapport entre les abaisserrients produit par le chlorure et le bromure (ou le bromure e t le iodure) du m6me mQtal est sensible-ment le meme pour tous les mdtaux du groupe.” The ratio varies between the values 0.78 and 0.91.From the results of the present investigation carried out wit ON THE TEMPERATURE 0%' MAXIMUM DENSITY OF WATER. 121 monobasic inorganic acids and their sa1t.s with univalent metals it will be shown that the lowering produced by any given salt con-forms to a simple general rule and can in fact be calculated from the known lotwerings produced by other salts. The results of the melasurements are given in table 11 which contains the figures obtained for solutions varying in strength between semi- and one-sixteenth-molecular the normal br molecular lowerings being calculated from those of lower concentration TABLE 11.M/16. Jl/S. A! 14. .................. 0-7 1.3 HC1 -1.4- LiCl ................... .................. 1.G 3.1 NaCl -. &>. 8 .................. 1.4 KC1 -_ ............... 2 . 0 1.8 _- -NH,CI __ llensity of 8dt Molecular M 12. lowering. 2.6 6.2 2.8 5.6 G-2 12-4 5-6 11.0 3.B 7.2 .................. 0.9 1.8 3.7 7.4 HBr .................. 1.9 3.8 7.6 LiBr .................. 1.8 3-7 7.4 14.5 NaBr __ .................. 1.6 3.2 6.5 13.0 KBr -_ ............... 1.2 2.8 4.7 9.4 8.8 .................. 1.2 i d HI ................... 1.8 3.3 9.2 Lil.. -NaI ..................1.0 2.0 4.0 16.4 14.8 KI .................. 0.9 1.8 3.7 10.8 HNO ............... 0.8 1.6 3-1 12.4 ............... 1.6 3.1 12.4 20.0 NaNO ............... 1-3 3.5 5.0 18.0 14-4 ___ - -NH,Br -_ .). -> -- .-NH,I 1.4 9.7 _. LiNO -KNO ............... 1.1 8.2 4.5 NH,NO ............ 0.9 1-8 3-6 I t will be seen a t once that the results agree with the law of Despretz the semi-molecular solutions giving twice the depression of the corresponding fourth-molecular. Further it is obvious that the lowering is not connected with the osmotic pressure as the values shown for the molecular lowerings of differeni; solutes vary greatly; nor is a consideration of the difference in the degree of ionisation sufficient to account f o r this abnormality since the various solutions of any given colncentration are practJcally ionised to the same extent.The regularity running through all the measurements can readily be seen if the difference between the lowering shown by any acid and say its sodium salt is considered. This difference for the four .................. 0.7 -_ _ _ _ 122 WRIGHT THE EFBEOT OB SOME SIMPLE ELECTROLYTES acids tabulated has the values 7*2,7.4,7*6 and 7.6 ; thus thereplace-ment of the hydrogen ion by sodium causes a practically constant increase in the molecular lowering. A similar increase is found when potassium is used instead of sodium the average value being 5-75 whilst for ammonium the value is 2.0. Further the same effect is observed in the case of the acid radicle; thus the replace-ment of chlorine by bromine Tncreases the molecular lowering by 2.1 whilst the sibstitution of iodine for chlorine causes an increase of 3'7.From a consideration of these results it is evident that each acidic or basic radicle has its own effect on tho lowering of the point of maximum density and that the effect produced by a salt is equal to the sum of the lowerings caused by the metallio and acidic radicles. Hence if we take the molecular lowering of hydrochloric acid-which gave the smallest effect of all the sub-stances examined-as standard we can obtain the molecular lower-ing of any salt or acid by the addition of two numbers one corre-sponding with the acidic and hhe other Fith the basic radicle of the salt. It will a t once be seen that there is a close resemblance between the above conclusion and Valson's law of moduli which states that the density of a normal salt solution is the sum of an acidic and a basic effect and can in fact be Calculated by adding to the density of a normal solution of a standard s u b s t a n c e ammonium chloride-two figures or moduli one characteristic of the acidic and the other of the basic radicle of the salt.The moduli for the lowering of the point of maximum density are given in table 111 and the molecular lowering of any salt can be found by adding to the molecular lowering of hydrochloric acid (5.2) the two moduli corresponding with the given salt. For ex-ample the calculated lowering for potassium nitrate would be 5-2(hydrochloric acid) + 5*75(potassiurn) + 7*2(nitrate) =18*15 the actual value found being 18.0.Several values for each modulus calculated from different salts are shown in the table together with the mean value derived from them. TABLE 111. C1. Br. I. NO,. Average. ............ Li 0.6 0.2 0.4 0.0 0-3 Na ............ 7.2 7.4 7.6 7.6 7-45 K ............ 5.8 5.6 6.0 5.6 5.75 NH ......... 2.0 2.0 2.0 2.0 2.0 Br 2.2 2.0 2.4 2.0 2.2 2.2 I 3.6 3.6 4.0 3.8 3.6 3.7 NOa ......... 7.2 6.8 7.6 7.0 7.2 7.2 H. Li. Ne. K. NH,. Average. ............ .............. ON THE TEMPERATURE OE’ MAXIMUM DENSITY OF WATER. 123 It should be noted that a similar set of moduli could be calcu-lated from the molecular lowerings given by Coppet although as a rule his values would not be identical with those tabulated; the results however approximate to one another fairly closely con-sidering the difference in the experimental methods employed.We may next consider the results obtained with the weak mane basic organic acids in comparison with their highly ionised salts. Formic acetic and propionic acids together with their sodium and ammonium salts have been examined. TABLE IV. N 18. - Formic acid ............ Na salt .................. 1.6 Acetic acid ............ N a salt .................. 1-5 Na salt .................. 1.5 - NH salt ............... NH salt -Propionic acid ......... -NH salt --............... ............... N/4. 1.7 3.2 1.7 1.8 3.0 1.6 2.0 3.0 1.7 N/2. 3.6 3.6 3.7 3.1 4.0 3.4 --N. 7.2 12.8 7.2 7.4 12.0 6.2 8.0 12.0 6.8 The results do not show the normal change 7.6 which was obtained with strong acids when the hydrogen atom was replaced by sodium but the difference between the values for the sodium and ammonium salts is constant in all three cases and is identical with that obtained in the case of the inorganic acids.In other words the highly ionised salts of organic acids behave in the normal manner whilst the feebly ionised acids themselves are abnormal. The dibasic acids with their acid aiid neutral salts are also of interest. TABLE V. Sulphuric acid ......... NaH salt .................. Na salt .................. Oxalic acid ............ NaR salt.. ................ Na salt .................. Succinic acid ............ NaH salt ..................Ne salt .................. MI1 6. M 18. 3.0 2.0 4-0 2.0 4.0 -1.7 - 2.5 1.5 2.9 -M 14. M. 6.1 24.4 - 32.0 - 32.0 3.0 12.0 I 19.2 - 32.0 3.4 13.6 - 20.0 I 23.2 It will be seen that the replacement of one hydrogen atom b 124 WRIGHT THE EFBECT OF SOME SIMPLE ELECTROLYTES sodium in the two stronger acids gives values approximating to the normal whilst succinic acid gives a slightly lower value thus re-sembling the weak monobasic acetIc and propionic acids. I n all cases, the replacement of the second hydrogen atom is quite abnormal and differs widely in the three cases. From the normal behaviour of the acid salt it may be colncluded thafu the ions of sodium hydrogen sulphate consist mainly of Na' and HSO,' and not H and NaSO,'.It should be noticed that as the greatest concentration examined in these acids was iV/4 the results are not so accurate as with the monobasic acids. The results obtained for the salts of the bivalent metals show great irregularities probably on account cf the complex ions which are present. For eixarnple the molecular lowering obtained for barium chloride was 24.6 and for barium nitrate 32.8 from which the two values 14.2 and 8.0 are obtained for t.he modidus of barium.' Similar varying results can be obtained from the figures given by Coppet and Miiller (Compt. reird. 1902 34 1208) f o r the loweriiigs shown by the halogen salts of barium and calcium. TABLE VI. Blolocular lowering. Modulus. Barium bromide .................. 25-14 26-28 10.9 Bariuiz- iodide .....................29.24 29.42 11.7 Calcium chloride ............... 18.0 18.3 7-8 Calcium bromide ............... 20.12 20.93 5-7 Calcium iodide ..................... 26.09 26-63 8.7 It will be seen at once thatl there is no regularity comparable with the case of the univalent metals. E X P E R I M E N T A L . The apparatns employed is shown in the diagram; it consists of a dilatometer with a capacity of about 50 C.C. and fitted with a stem 25 cm. long and of 0.5 mm. bore. To co'mpensate f o r tho change in volume with temperature a portion of the bulb is filled with mercury; the fraction of the total dilatometer volume thus filleld is equal to the ratio between the coefficients of cubical es-pansion of glass and mercury so that on a change of temperature, the expansion or contraction of the metal exactly compensates the expansion or contraction of the bulb t'he volume of the latter unoccupied by the mercury thus remaining constant.The dilahineter is filled by rneanr of a tap funnel and a vacuun ON THE TEMPERATURE OF MAXIMUM DENSITY OF WATER. 125 pump. The stem of the dilatometer passes through a rubber cork fitted into the opening of the tap funnel which contains the liquid to be int'roduced into the dilatometer. The funnel and the attached dilatometer is now inverted and connexim made between it and a filter pump. As the pressure is lowered the air in the dilatometer bulb bubbles through the liquid in the funnel and on detaching from the pump the liquid is forced into1 the dilatometer bv the action of the atmospheric pressure.A second and a third exhaustion are usually necessary and the last trace of air in the bulb is removed by heating. The dilatometer with t.he funnel still attached is now placed in a vacuum-jacketed vessel filled with brine a t about 5O and allotwed to cool. When the apparatus has attained the temperature of the surrounding liquid i t is disconnected from the tap funnel and a few shavings of ice are added to the liquid in the vacuum flask; air is then driven through the cooling mixture so as to stir it until the ice has melted. The apparatus is now left for a quarter of an hour to allow the dilatometer to aswme the temperature of the bath; this tempera-ture is then not'ed and the level of the liquid in the dilatometer tube measured ; the apparatus is again left' and readings are taken at five-minute intervals until the liquid in the dilatometer ceases t o contract; this precaution is necessary in order to ensure that the whole apparatus is in a state olc thermal equilibrium.A furbher small quantity of ice is now added so as to lower the temperature a fraction of a degree and the process repeated. After several additions of ice the liquid in the dilatometer reaches its poii?t of maximum density, and on further cooling i t expands As this point. is approached the coefficient of expansion of the liquid diminishes so that i t is difficult t.0 deter-iniiie the exact temperature of maximum density, and the readings given are only accurate t o 0 - 2 O . After a measurement the instrument is warmed within about so as to expel a little of the contents and is then inverted so that the mercury runs out.This mercury is dried and reserved for the next deter-mination whllst the rest of the contents of the dilatometer are removed by means of t,he filter pump. The instrument is then rinsed with the next liquid t o be investigated and after the re 126 THE EFFECT OF SOME SIMPLE ELECTROLYTES BTC. introduction of the mercury it is filled with the solution and the measurement made as before. The coefficient of expaasion of the glass was calculated between the ordinary temperature And looo by filling the instrument with mercury and weighing the quantity expelled when heated in a steam-jacket whilst the volume of the bulb was measured by filling with water and weighing a t a known temperature.The stem of the instrument was not graduated but the level of the liquid below the upper end was determined by means of a depth gauge fitted with a vernier; by !this means a change of level of 0.1 mm. could be detected and, moreover the labour of regraduation of the stem after an accident was avoided. A few other points may be noted. Conclusions. (1) The lowering of the temperature of the maximum density of water produced by the addition of a solute is directly proper-tional to the concentration of the latter (law of Despretz). (2) The lowering produced by a highly ionised binary elecicro-lyte is composed of two separate independent effects one due to the acid radicle and the other due to the basic radicle of the electr ol yt e . (3) The lowering produced by a highly ionised binary electrolyte of molecular concentration can therefore be calculated by the addi-tion of two moduli to the lowering produced by a molecular soln-tion of a chosen standard substance. The chosen standard was N-hydrochloric acid which gives a lowering of 5'2. (4) The acid salts of the dibasic acids behave normally but the neutral salts of such acids and the salts of the bivalent metals do not follow any simple rule in their effect on the temperature of the maximum density. (5) Feebly ionised organic acids show abnormal effects but the highly ionised salts derived from them behave in the normal manner. CHEMICAL LABORATORY, QUEEN'S UNIVERSITY BELFAST. [Received December 2184 191 8.