XISOLUBILITY OF POTASSIUM SULPHATE. 377X.- The Solubility of Potassium S d p h a t e in Concen-tqvrnted Aqueous Solutions o f NowElect7-olytes.By JOHN JACOB Fox and ARTHUR JOSIAH HOFFMEISTER GAUGE.IN a recent communication (Trans., 1909, 95, 885) one of usshowed that the solubility of potassium sulphate in water at 2 5 Owas decreased markedly by the presence of potassium acetate.Since the rate of the decrease is much greater with the moredilute solutions of potassium acetate, which are dissociated electro-lytically to a greater degree than the stronger solutions, th378 FOX AND GAUGE : SOLUBILlTY OF POTASSIUM SULPHATE INpresence of the potassion due to ionised potassium acetate mightbe considered as being the main factor in decreasing the solubilityof potassium sulphate as distinct from the general effect of thesecond substance in solution, in this case non-ionised potassiumacetate.With the view of gaining some further knowledge as tothe action as precipitant of the second substance in solution, itwils thought desirable to determine the effect on the solubility ofpotassium sulphate of a number of non-electrolytes, and to ascertainwhether the nature of the non-electrolyte was to any marked degreeconcerned in the action. I n the case of potassium sulphate a fewdeterminations of this character have already been carfied out,and the general result, both with electrolytes and non-electrolytes,is that the solubility of potassium sulphate in aqueous solution isdepressed. From the point of view of the present communicationthe results of most interest are those of Girard (Bull.SOC. chim.,1885, [ii], 43, 552) for the solubility of potassium sulphate inaqueous ammonia, and of Rothmund and Wilsmore (Zeitsch.physikal. Chem., 1902, 40, 619) for the solubility in aqueous aceticacid and aqueous phenol.While these results are similar to those described below, a strictcomparison cannot be made, since the alteration of solubility byvolume has been used by these observers, whereas we prefer thealteration in solubility referred to a fixed quantity of water. Afair approximation to the depression of solubility by volume can,however, be deduced if it is assumed that the total of the volumesof the potassium sulphate and of the liquid in which it is dissolveddoes not alter.This gives a volume too great by rather more than1 per cent. in the stronger solutions of potassium sulphate, andpractically correct in the weaker solutions. Thus it was found thataqueous alcohol, I):: 0.9913, yielded a solution containing 7 percent. of potassium sulphate, and having a density of 1.0499. Thedensity of finely powdered potassium sulphate was found to be2.656 at 2Oo/2O0. Hence 100 C.C. of the saturated solution should,from the composition, occupy 101.3 C.C. Similarly, a solution ofglycerol and water, containing 7.2 per cent. of potassium sulphate,possessed a density of 1.1029, the original glycerol and waterhaving a density of 1.0420. The calculated volume of 100 C.C. is101-2 C.C. As most of the solutions contain less than 7 per cent.of potassium sulphate, the errors introduced are less. Using thesecalculated volumes, it will be found that the nature of the curvesobtained is similar to that of Rothmund and Wilsmore referredto above.The substances used by us were ethyl alcohol, ethylene glycol,glycerol, mannitol, chloral hydrate, sucrose, acetone, and pyridineCONCENTRATED AQUEOUS SOLIJTIONS OF NON-ELECTROLYTES.379The solutions were examined partly from the point of view ofthe possible formation of definite hydrates, since it was thoughtpossible that if with any mixture a simple hydrate was formed,a change in the solubility curve at this point would be found.With this object, mixtures with water in all proportions were taken,and the solubilities plotted against the percentage composition ofthe aqueous solution.This method of plotting was chosen in preference to the methodof reference to a fixed quantity of water, because of the difficultyof deciding whether water should be considered as solvent or solutein concen trat.ed solu tions.*It is obvious from the curves that as the number of hydroxylgroups in the molecule increases, the precipitating effect of thenon-electrolyte decreases, and if the curves are drawn with moleculesof potassium sulphate as ordinate and non-electrolyte as abscissae,taking 1000 molecules of water as fixed, the result is the same.Whether this would be found to apply to other salts cannot, ofcourse, be decided without further investigation.None of thecurves give any indication of discontinuity, so that on this viewthe existence of definite simple hydrates is negatived.This doesnot, of course, imply that the substances dissolved do not formcomplexes with more or less water, but the most the results setforth here can be said to indicate is that the non-electrolyte andwater exert a material influence on each other, the action prevent-ing the water from dissolving the full amount of salt. There isone consideration, however, which is in a measure opposed to theresults obtained by Jones and Getman from observations of thedepression of the freezing point of aqueous solutions of non-electrolytes (Amer. Chem,. J., 1904, 32, 308). From theseobservations, Jones and Getman conclude that the deviations ofthe observed values of the freezing point from the theoretical valueare due t o the formation of complexes of the solute and water;that in so far as some of the water is used up to form hydrates,less water remains to function as solvent for the hydrate, and thattherefore abnormally high results for depression of freezing pointare obtained.It should follow that if some of the water is pre-vented from acting as solvent in the case of hydrates, the sameeffect should be shown when a second substance (for example,* During the course of this work, a paper by Rothmund appeared (Zeitsch.physikal. Chem., 1909, 69, 523), dealing with a somewhat similar problem, butusing csmparatively dilute solutions of the various alcohols and other organic sub-stances.Their effects as precipitants were studied in the case of lithium carbonateand other sparingly soluble salts. Rothmund used fixed volume, and this is justifiedsince the volume of the original solutions could be altered but slightly by thedissolution of sparingly soluble salts380 FOX AND GAUGE : SOLUBILITY OF POTASSIUM SULPHATE INpotassium sulphate) is dissolved in the solution. Philip hasdemonstrated this to be the case when hydrogen is dissolved inaqueous sucrose solutions (Trans., 1907, 91, 711). Now, accordingto Jones and Getman, alcohol, chloral hydrate, and mannitol do notshow any marked tendency to form hydrates, whereas sucrose, andparticulasly glycerol, show considerable hydration. We shouldtherefore expect alcohol, chloral hydrate, and mannitol to exertless influence on the solubility of potassium sulphate than eitherglycerol or sucrose. As will be seen from the results here given,the reverse is the case, both alcohol and chloral hydrate beingmuch more marked in their action than glycerol or sucrose, whetherthe curves are drawn up on the percentage basis as in the figure,or on the basis of a fixed 1000 molecules of water.It may beargued that the results are in part explicable on the assumptionthat unless ions are hydrated they cannot exist in aqueous solutions,and consequently that the potassium sulphate will not dissolve ifthe ions derived from it are subjected to conditions which tendto dehydrate them. The presence in solution of hydrated non-electrolytes might be supposed to act in the direction of preventingthe ions from obtaining the requisite quantity of water.I n suchcircunistances the ions could only obtain sufficient water a t theexpense of the hydrate of the non-electrolyte, and the final resultwould depend on whether ion or non-electrolyte was most effectivein obtaining water (see Lowry, Trans. Famduy Soc., 1905, 1, 197).It would also follow that with the increasing concentration of thenon-electrolyte the proportion of hydrated non-electrolyte f orniedwould increase, with a corresponding decrease in the hydrated ions.Such an explanation is, however, merely surmise, and does notaltogether apply to the strongest non-electrolyte solutions wherethe water is insufficient t o form any hydrate.EXPERIMENTAL.The solutions used were made up by weighing both the substanceand the water in which it was dissolved. Saturation was obtainedby cooling the saturated solution in contact with solid from asomewhat higher temperature to 2 5 O in a thermostat, and byshaking a t 25O.The amount of potassium sulphate was determinedby direct weighing of the salt after evaporation and ignition, orby estimating the amount of sulphate by means of barium chloride.The percentage composition of the solutions and the number ofmolecules of solutes per 1000 molecules of water froin which thecurves are drawn are as followsCONCENTRATED AQUEOUS SOLUTIONS OF NON-ELECTROLYTES. 381A queous Alcohol-Potassium Sulphate.AlcohoL1-354'807.809 *7012.3414-5115'2620.5026-9135.9743.9069'26Pyridine.4.2313.9024'5134-1946'2955-9375-90Potassiumsulphate.9-176'904'964 *323.572.712.661-830.970 410'220.016Water.89.4888.3087.2485.9884-0982.7882-0877-6772'1263.6255.8830-72Molecules per1000 molecules of water.PotassiumAlcohol.sulphate.5 -9 10.621 '3 8 -135.0 5.944'2 5.257 '4 4'468'6 3'472'7 3 '3103'2 2'4146'1 1'4A/ \- -- -- -Aqueous Pyridilze--Potassium. SulpJLat e.Molecules per1000 molerules of water.Potassiumsulphate.7 *954 -772.751.470 '450.120.006Water.87'82S1'3372'7464 -3453.2643-9524.09cPyridine.11.038.976.8121-1198.0---.PotassiumsulIihste.9.46.13.92.40-9--A queous Ethylene Glycol-Potassium Sulphate.Molecules per1000 molecules of water.E thyleiie Potassium hthylene Potassinni'glycol.sulphate. Water. glycol. sulphate.3-16 9 *67 87-17 10.5 11-59-78 7.69 82.53 34'4 9.618'47 5 '74 75.79 70 '8 7.832.1 1 3.57 64.32 145.0 5.749.03 1'83 49'14 280 7 3.382 FOX AND GAUGE : SOLUBILITY OF POTASSIUM SULPHATE INAqueous Chloral Hydrate-Potassim Sdphate.Chloralhydrate.6 *449.0912-3813'2022.0733.1544'4047'3062'8270.2880.368526Glycerol.8'9613-3620-3424'1533-7340'4043 *5250'18572267'9478-1898 '28Potassiumsulphate.9-138'417 -797 '315'884 *543-362-922.001-751'401-08Water.84'4382'5079.8379.4972.0562.3152.2449.7835.1827 *9718'2413.66Molecules per1000 niolecules of water.Chloral Potassiumhydrate.sulphate.8 '3 11 '212'0 10.516.9 10.118'1 9 *533.4 8 '458-0 7'592'6 6-6108-5 6'1194'5 5 -9273.8 6 -5.A/ \- -- -Aqueous Glycerol-Potassium Sulphate.Potassiumsulphate.8.871-696 '475-884'413.653.382'692 *071-530.980.73Water.82.1778.9573.1970.0261.8355.9553-1047-1340'7130.5330.840.99Molecules per1000 molecules of water.PotassiumGlycerol. sulphate.21.3 11'233.1 10'154'4 9.167 -5 8'6106'8 7'4141'4 6.7160'4 6.6208 '4 5'9275.1 5.3A/ >- -- - - -Aqueous Mannitol-Potassium Sulphate.PotassiumMannitol.sulphate.3 20 10.325 *82 10 -078 *35 9-6111 -26 9.1914-30 8.6617.22 8 -35Molecules per1000 molecules of water.PotassiumWater. Mannitol. sulphate.86.48 3.7 12.384'11 6'8 12'382.04 10'1 12.179.55 14'0 11.977-04 18.4 11'674.43 22 '9 11.6/-hCONCENTRATED AQUEOUS SOLUTIONS OF NON-ELECTROLYTES. 383Sucrose.9'5618'5528-1637 '2247'5557'00Acetone.4.9210.0616'2324.3137'1946-2962'40A queous Sucrose-Potassium Sulphate.Molecules per1000 molecules of water.Potassiumsulphate.9-658.657 '426'355%4'24PotassiumWater. Sucrose. snlphate.80.79 6.2 12.372'80 13'4 12.364-42 23.0 11.956'41 34-8 11%47'24 52.9 11.438.76 77 -5 11.3.4 qu.eous Acetone-Potassium Sulphate.Molecules per1000 molecules of water.Potassiumsulphate.7-205-022-961-500'470 '200.03Water.87-8884-9280-8174-1962'3453-5137-57Potassium'Acetone. sulphate.17.4 8 - 536-7 6.162 -3 3 -8101.7 2 '1185-2 0.8268-5 0'4- -Certain of the curves (p.384) require some consideration. Pyri-dine dissolved in water affords some evidence of the formation of ahydroxide from the fact that it precipitates ferric hydroxide fromaqueous solutions of iron salts. When drawn up on the basis ofa fixed amount of water, this curve cuts the alcohol curve. Itwas observed that above the temperature of 4 5 O two liquid phasesformed at all concentrations above 5 per cent. and below 46 percent. approximately. The position of the chloral hydrate curveclose to the glycerol curve appears to us to demonstrate that thecause of the depression of solubility is similar in both cases, whichdoes not support the deductions of Jones and Getman (Zoc.cit.) asto the remarkable difference in hydration of these two substances.It will be seen that if the chloral hydrate curve is expressed mole-cularly with reference to 1000 molecules of water, the end of thecurve begins to rise slightly, suggesting that potassium sulphateis soluble in absolute chloral hydrate. An actual determinationwith liquefied chloral hydrate at 4 5 O gave the solubility as 0.38per cent. of potassium sulphate.Glycerol of 99.0 per cent. strength dissolved 0-73 per cent. ofpotassium sulphate, an amount which is much greater than wouldbe dissolved by the water present.It must be concluded thatglycerol also dissolves potassium sulphate384 SOLURILZTY OF POTASSIUM SULPHATE.Both the mannitol and sucrose curves are practically straightlines. I n other words, the decrease in the solubility of potassiumsulphate in concentrated solutions of these two substances variesdirectly as the amount of solute present originally, so that thedecrease, if due at all to hydration of the solute, requires thesame degree of hydration a t all concentrations. This is inadmissibleon the usual assumption that the degree of hydration depends uponthe amount of water. The mannitol curve could not be carriedfurther than the point shown, which is very close to the saturationpoint of mannitol.The solubility in aqueous acetone of varying concentrationsexpressed per 1000 molecules of water gave a curve which followed100 2water.60 100 %lion-electrolyte.1.Alcohol, 2. Py?*idine. 3. E'thylmc! glycol. 4 . Glycerol.5 . Cl&wal hyd?vtc. 6. Mftnnitol. 7. Sacrose.the alcohol curve closely, but fell somewhat below it. Acetone wastherefore found to possess the greatest precipitating effect of thenon-electrolytes examined.Both pyridine and absolute alcohol dissolve minute quantitiesof potassium sulphate, but the amount was too small for accurateestimation. Schiff (Annalen, 1861, 118, 362) determined the solu-bility of potassium sulphate in aqueous alcohol at 1 5 O , and gavefour points only. This curve, as far as it goes, runs parallel withand a little below the one given here.It is of interest to compare the curve for solubility in potassiumacetate solutions with the foregoing curves. The position occupiedis well below the alcohol curve, a result which may be accountedfor if to the main action of non-ionised potassium acetate aACTION OF CALCIUM AND LITHIUM ON ORGANIC HALIDES. 385precipitant is added the influence of the potassions from theionised portion.J3ydration of the ions of the salt might be considered as a con-tributory cause of the depression of solubility. .As the rate ofdecrease is for most of the curves greatest with the dilute solutions,this assumption appears to receive some support; but it cannotbe considered quite satisfactory as an explanation, if it is bornein mind that the decrease is continuous even in the strong solutionswhere there is not sufficient water to form hydrates. Dilute solu-tions are, however, the limiting cases, and here again we find, asusual, that the rules deduced from the dilute do not apply to con-centrated solutions. It is hoped that the results of an investigationnow proceeding as to the influence of one non-electrolyte on thesolubility of another may throw some light on the possibility ofthe hydration of ions being a contributory cause of the depressionof the solubility of salts by non-electrolytes.EAST LONDON COLLEGE,UNIVERSITY OF LONDON