]\gt ; On the Ionic Solubility-product .
By JAMES KENDALL , M.A. , B.Sc. ( Vans Dunlop Scholar in Chemistry , University of Edinburgh ) .
ommunicated by Prof. James Walker , F.R.S. Received February 16 , \mdash ; Read March 9 , 1911 .
) The theory of a constant ionic solubihty-product was first advanced Nernst , * from the analogy of the laws governing dissociation in gasea According to the principle of mass action , if the dissociation pressure of a system in equilibrium be changed by the introduction of an excess of of the dissociated gases , combination takes place between them unhl equilibrium is restored .
Thus , on the addition of either or HC1 to a system consisting of ammonium chloride in equilibrium with its dissociation products , ammonium chloride is produced .
For each temperature , the product of the pressures of and HC1 is a constant .
Similarly , : Nernst found that the solubility of an electrolyte in water was reduced by addition of any salt containing a common ion .
The variations in the solubility of silver acetate ( a sparingly soluble salt ) in water knoWn amounts of silver nitrate or sodium acetate were investigated , results obtained confirmed the view that a corresponding equilibrium law was here applicable .
If we confine ourselves to substances of the simplest , type , the law may be stated thus : " " At a given temperature the solubility .
of a sparingly soluble electrolyte is dependent upon a constant , hich is proportional to the product of the of the ions of the electrolyte The fundamental assumption made in the application of this law is that in a saturated solution of an electrolyte the undissociated salt in solution plays the part of an intermediary even the ions and the solid , being equilibrium with the ions on the one hand and with the solid on the othgr. .
Consequently , so long as the solvent remains the same , the concentration the undissociated salt is constant , however much the ionic lations in subsequent experiments .
With picric and methyl picric the end point was at firsb difficult to distinguish , owing to the disturbiqg colour , but after some practice the first permanent change of tint could be detected quite accurately by using a comparison solution .
In most cases the saturation point was soon reached ; methyl picric acid , however , attained its full value only after two months , picric acid not even then .
It was found that the presence of a few drops of ligroin or benzenr in the solution shortened the necessary time for methyl picric to a few days , but the end points reached differed from that obtained with pure water .
Ligroin , in which the acid is moderately soluble , increased its solubility by 2 per cent. ; benzene , in which it is freely soluble , by 6 per cent. A considerable difference was thus caused by the presence of the extremely small amounts of each in solution in the water .
This is note worthy in view of results afterwards obtained .
In Table I below the concentrations of the saturated solutions of the acids are given in normalities ; from these , by means of the dissociation constant , the solubility-product and the " " constant value\ldquo ; of each acid are obtained .
The above values for the molecular conductivities of the acids are in general greement with those given for picric acid by Ostwald* and Bothmund ; f but the degree of error may be fully 1 per cent. This would affect degree of dissociation to the same extent ; but , by trial from the equations , it was found that an errot of 5 per cent. was required to affect the solubility determinations , afterwards made with the acids , beyon.d the limits of , experimental errol .
The conductivities were determined with a rotating commutator and galvanometer , more satisfactory results being obtained with this apparatus than with the usual induction coil and telephone , especially for the more dilute solutions .
In the determination of the solubility of one acid in the presence of. .
another , the method of procedure was exactly similar to that already described for single acids .
The experiments were of two classes : in the first , two solid acids were taken with water ; in the second , one acid was present as a solid , the other as a solution in water of known strength .
In all cases the end point was quickly reached , even methyl picnc , in the presence of another acid , attaining the steady state in a few days .
The first investigations were of the simplest type , the two aoids obeying Ostwald 's dilution law .
Of these , salicylic and -nitrobenzoic acids are good examples , and a full series of experiments with these two acids was carried out , the results of which are shown in Table II ( p. 207 ) .
Since they serve to illustrate the other cases , they are commented upon in detail .
It will be seen that the amount of acid in solution is in every case in excess of the theoretical .
The divergences in Sections and increase with the strength of the original solution taken , and the greatest divergence is obta.ined in Section , where both acids are present in the soIid form .
All the differences are far beyond the limits of experimental error .
There are , however , some small corrections to be made in the above The results are given in normalities , whereas to be strictly comparable 208. .
The " " effective\ldquo ; correction is thus , even in the last case , only .
This is less than per cent. of the whole titre .
In Sections and of Table II it will be seen that not only does the divergence from theory increase with increase of concentration , but .
that , the actual amount of acid dissolved , instead of decreasing , at first rapidly and then more slowly , as demanded by the theory , either shows an increase or becomes practically constant .
The divergences from theory are therefore fundamental .
The result of another experiment will show this more A saturated solution of -nitrobenzoic acid , of strength , free from solid , was added to solid salicylic acid and shaken in the thermostat until constant .
strength was now N. the concentration of the hydrogen ion had been increased , the solution should be supersaturated with respect to -nitrobenzoic acid .
But on solid -nitrobenzoic acid to the solution , and shaking again until the titre was constant , the concentration rose to , i.e. the solution was not supersaturated , but unsaturated .
A comparison of the three results of Sections , and is int.eresting .
The concentrations of the solutions originally taken in and are , as nearly as possible , those shown by A to be ' theoretically , saturated in presence of the other acid .
The final titre should therefore be the same in all three cases .
Instead of the theoretical value of however , the three values obtained are , and respectively , which differ widely not only from the theoretical value bul also from each other .
The presence of solid acid allows excess beyond theory to be dissolved ; this is allowed in the case of both acids in , in the case of -nitrobenzoic only in , and in the case of salicylic only itive divergences .
in all cases within error limits of the theoretical result- A typical loriea ot corrections is given in Table IIIA .
Table IIL\mdash ; Remaining Experiments with Two Weak Acids .
The " " effective\ldquo ; correction is only at the greatest concentration .
This is less than per cent. of the whole titre .
In the next table there are given the results of experiments with two strong acids .
For the calculation of the theoretical results the following table of dissociation values for hydrochloric acid was used : , The dissociation values of picric and methyl picric acids at various dilutions have already been given .
Table \mdash ; Experiments with Two Strong Acids .
Here also divergences from theoretical results are obtained ; negative in the case of methyl picric and hydrochloric , positive in the case of methyl picric and picric acids .
Table V gives examples of weak acids ( salicylic and -nitrobenzoic ) Mr. J. Kendall .
The differences in Table VI are all large and positive .
that methyl picric acid is more soluble in salicylic acid solutions than pure water .
In none of the above tables is the theory exactly followed , and a further experiment was performed , to ascertain whether the divergences obtained were superposable .
A known solution of hydrochloric acid was taken , and two weak acids ( salicylic and -nitrobenzoic acids ) added : The differences to be expected from theory estimated by interpolation from the results of the preceding tables , and compared with the difference experimentally 'obtained .
Table \mdash ; Two Weak Acids in Hydrochloric Acid Solution .
The following divergences were to be expected ( see Tables II and ) ( o ) Due to solid salicylic and solution of -nitrobenzoic N. Due to solid -nitrobenzoic and solution of salicylic N. ( c ) Due to solid salicylic and solution of hydrochloric N. ( d ) Due to solid -mitrobenzoic and solution of hydrochloric N. The solution should therefore be approximately of theoretical strength if the various differences are superposable .
Seeing that the accuracy of a large number of experiments is inyolved in the values of the expected divergences given above , the experimental result agrees with this supposition well within the limits of error .
A general consideration of the results in Tables II to VII shows that the theory holds strictly in none of the possible combinations of acids .
It makes no difference whether they are both weak , both strong , or one of each type ; similar ergences of several per cent. obtained in every oaae .
And these divergences , though usually numerically small , are that of the acid dissolved , in which latter small variations were impossible to measure .
The method employed was that of evaporation .
The solutions were .
prepared in the same manner as already described , 25 .
of the liquid pipetted out into a weighed evaporating dish , and evaporated down to dryness .
The amount of the dissolved solid was then found by weighing , and its titre in the solution directly calculated .
The case investigaM were :\mdash ; ( a ) In solutions of formic acid : Salicylic acid , hippuric acid .
( b ) In solutions of acetic acid : Salicylic acid .
( c ) In solutions of : Salicylic acid , -nitrobenzoic acid .
utions c hippuric acid could be evaporated down on steam-bath without loss of solid by volatilisation ; those containing salicylic .
and -nitrobenzoic acids , since these proved to be more or less volatile with steam , were allowed to evaporate slowly at the ordinary temperature .
The residues obtained weighed from to of a gramme ; the limit in the latter case is per cent. Experiments performed in duplicate gave results agreeing together within this limit , and the results of this method , as tested with some of the more dilute solutions , were closely concordant with those obtained volumetrically .
The results obtained are shown in both tabular and graphic form .
Some volumetric determinations are also given in the tables , both for the purpose .
of comparison with the gravimetric results and for use in the accompanying curves .
Results are expressed , as before , in normalities .
The corrections in the volumetric determinations , although much here owing to thu greater concentrations of the solutions used , are not apply \amp ; That due to .
oduct .
\fnof ; g Acid .
Acid .
( a ) An undissociated portion ( below the line ) , increasing or decreasing : regularly , according as the acid is more or less oluble in the solvent acid : than in water .
( b ) A dissociated portion above the ecreasing more or kss rapidly .
according to the ionic strength of the solvent acid .
This decrease is not regular , but is proportionally greater the less concentrated the solution is , falling off as the dissociated portion is suppressed by increase of concentration .
The solvent effect , acting directly upon the undissociated portion , will also indirectly affect the dissociated part of the substance in solution , since this must remain in equilibrium with the undissociated part .
In the curves given , the acids exist in solution mainly in the undissociated state , hence this indirect effect is but small , and soon becomes negligible with increase of concentration , to the almost total suppression.of the ionised part .
In solutions of strong electrolytes , however , the indirect may be greater than the direct effect ; hence the solubility curves in such cases will not..-finally approximate to lines which cut the axis , when produced backwards , at or very near to the " " constant undissociated value The case of salicylic acid in solutions of acetic acid is another example of the inapplicability of the rule in the eYent of the ionised portion persisting appreciable quantity at all concentrations of the solvent acid , The experimental diyergences obtained from the .
values indicated by the theory of the constant solubility-product can thus be accounted all cases by this solvent effect .
The divergences , positive or negative , are :

