The Origin of Osmotic Effects .
By Henry E. Armstrong , F.R.S. ( Received and read June 14 , 1906 .
) [ International Catalogue of Scientific Literature .
Author 's title-slip :\#151 ; C. D. Subject slips :\#151 ; C 0310 A theory of osmotic effects .
C 1930 Association in solution .
C 6250 Association of electrolytes with water .
D 7065 Structure in relation to hydration .
v 1 ) 7155 Theory of osmotic effects .
D 7255 Explanation of peculiarities of electrolytes .
Q 0224 Origin of osmotic effects .
] The result of the comparative study of enzymes and acids as hydrolysts may be held to be that an explanation based on an association hypothesis may be given of their action which is simple and rational and in accordance with the facts .
The explanation which has been advocated from the point of view of the ionic dissociation hypothesis , if not inadmissible , is altogether improbable ; moreover , whilst this hypothesis is applicable to the explanation of but a very limited class of chemical interactions and is .in no way a necessary hypothesis , the assumption that association rather than dissociation is the condition precedent of change* appears to afford a sufficient explanation of chemical interchanges in general , of whatever kind .
It is desirable , therefore , to consider somewhat closely what may be the behaviour of salts in solution , in order that their marked activity in comparison with that of nonelectrolytes may be accounted for .
It is well known that salts in solution and indeed electrolytes generally produce effects , in lowering the freezing point or the vapour pressure , for example , which are abnormally large in comparison with those which are produced by non-electrolytes .
It has therefore been assumed that electrolytes are more or less dissociated in solution into ions which play the part of individual molecules .
It is neither desirable to dwell on the inherent improbability of the conception nor to enter into any discussion of the hypothesis , beyond saying that it is difficult to discover any argument of which it is the unavoidable consequence among the reasons put forward in support of its acceptance , as these are inconclusive when not based on uncertain premises ; my object is to consider an alternative explanation .
In the case of osmotic phenomena generally , whatever may be the degree * 'Roy .
Soc. Proc. , ' 1904 , vol. 73 , 537 ; 'Chem .
Soc. Txans .
, ' 1895 , 1171 .
The Origin of Osmotic Effects .
of fortuitous coincidence between calculated and observed values when they are treated as gaseous pressure effects , the effects cannot in reality be due to pressure in any ordinary sense .
In liquids generally attractive forces must come into play between the molecules ; in solutions attractive forces must also come into play as between solute and solvent .
It is necessary to account for two classes of facts , viz.:\#151 ; 1 .
That non-electrolytes all have a similar influence when used in molecular proportions in solutions of equivalent strength .
2 .
That electrolytes have an excessive influence as compared with nonelectrolytes .
In virtue of 1 , it may be supposed that the primary or main osmotic effect is exercised by the solvent\#151 ; not by the solute .
There is apparently no other way in which the constancy of molecular behaviour can be so easily , if at all , accounted for , as the one constituent common to all solutions in a particular liquid is the solvent itself .
In all liquids the molecules must be regarded as in some degree associated .
In the case of water it is generally admitted that not only are the molecules associated , but that a relatively large attractive force operates between the molecules\#151 ; the heat of formation of liquid water at 100 ' from the fundamental gaseous molecules being a high value ( 18 x 536.= 9648 calories ) .
The effect of introducing neutral molecules\#151 ; non-electrolytes\#151 ; into liquid water must be to cause the dissociation of the molecular complexes to an extent corresponding to the proportion in which the neutral molecules are added .
If a substance enter into solution entirely in the form of its fundamental molecules , it will produce its normal effect , provided that its own attractive effect upon the water molecules be inappreciable .
Electrolytes , besides producing the dissociation effect , must be supposed also to exercise an attractive effect on water molecules .
In ordinary water , the state of equilibrium involves the simple dissociative change pictured in the equation : ( H20)n = nB.20 , in which n may have several values , as probably there is a greater number of forms present than two .
The introduction of any substance into solution involves the disturbance of the equilibrium in the direction ( H20)M \#151 ; \gt ; \#166 ; rtH20 .
The osmotic " pressure " is the measure of the extent to which equilibrium is thus disturbed by the liberation of the fundamental molecules or monads These monads are to be regarded as the attracting element in the region of the solution and as conditioning a flow of similar molecules from the region of the pure solvent\#151 ; or from a more dilute solution\#151 ; until the two regions are in equilibrium .
This explanation was definitely advanced by me in the 266 Prof. H. E. Armstrong .
[ June 14 , ' Encyclopaedia Britannica , in the article " Chemistry " * published in 1902 .
The conception has been arrived at from a somewhat different point of view and advocated recently as novel by Mr. Beilby in his Address to Section B of the British Association in South Africa.f Osmotic pressure has been treated as a negative pressure by HulettJ and again quite recently by Hudson .
S It maybe supposed , in the case of non-electrolytes , that the osmotic attraction is exercised upon external water molecules by the water monads present in excess in the solution as the product of dissociation of the water complexes ; in the case of electrolytes , in addition to this attraction , that exercised by the dissolved substance comes also into play : on this hypothesis , electrolytes are substances which are attractive of water practically in proportion to the efficiency of their solutions as electrolytes .
From the point of view here advocated , it is easy to see that , to compare the osmotic efficiency of substances , it is essential to measure their influence on one and the same proportion of solvent ; the use of weight-normal solutions , i.e. , of solutions made by dissolving the presumed molecular proportion ( or fractions thereof ) in a litre of the solvent , as practised by Morse and Frazer , || is thus entirely justified.1T The reason is also clear why only dilute solutions have afforded " normal " results in the hands of the advocates of the dissociation hypothesis and that they have overlooked the important fact discovered by Morse and Frazer , that the osmotic effect varies almost in direct proportion to the concentration , the results obtained with concentrated solutions being , in fact , scarcely less normal than those afforded by dilute solutions .
The presumed distinction between dilute and concentrated solutions almost * * * S * Yol .
26 , p. 739 .
t [ The ' Proceedings ' of the Koninklijke Akademie van Wetenscliappen Te Amsterdam , of May 26 , ' 1906 , issued June 21 , contains a communication , by J. J. van Laar , on " The Osmotic Pressure of Solutions of Non-electrolytes in connection with the Deviations from the Laws of Ideal Gases , " read on the 29th of April last .
Taking Morse and Frazer 's observations into account , van Laar arrives at the conclusion that osmotic pressure does not follow the gas laws ; he is of opinion that it is " no longer possible to uphold the old conception of the osmotic pressure as arising in consequence of a pressure of the molecules of the dissolved substance comparable with the gas pressure .
The molecules of the dissolved substance have nothing to do with the osmotic pressure except in so far as they reduce the water in the solutions to another state of concentration ( concentrated ) , which causes the pure water ( Concentration 1 ) to move towards the water in the solution ( Concentration 1 \#151 ; x ) in consequence of the impulse of diffusion.'1'1\#151 ; Note added Jidy 31 .
] + ' Zeits .
Phys. Chem. , ' 1903 , vol. 42 , p. 361 .
S ' Physical Review , ' 1906 , vol. 22 , p. 257 { re nature of osmotic pressure ; compare Batelli and Stephanini , ' Atti dei Lincei , ' 1905 ( v ) , vol. 14 , ii , pp. 3\#151 ; 14 ; also ' Pliysi-kalische Zeitschrift , ' 1906 , vol. 7 , p. 190 .
|| ' Amer .
Chem. Journ. , ' 1905 , vol. 34 , p. 1 .
IF Compare R. J. Caldwell , post , p. 272 .
1906 .
] The Origin of Osmotic Effects .
disappears .
This is well shown in the following table given by the authors named , in which N is the proportion of a gramme-molecular weight of sugar which is dissolved in 1000 grammes of water ; D the density of the solution at its freezing point ; A the observed depression of the freezing point .
It will be seen that A is very nearly 1*85 NT ) for all concentrations .
N. D. Ax = l -85 ND .
A. o-io 1 -0129 O 0-187 o 0-187 0-20 1-0257 0-379 0 -373 0-30 1 -0380 0-576 0-574 0-40 1 -0497 0-777 0-776 0-50 1 -0611 0-981 0-970 0-60 1 -0717 1 -189 1 -187 0 70 1 -0825 1 -401 1 -398 0-80 1 -0918 1 -616 1 -612 0-90 1 -1016 1 -834 1 -837 1 -oo 1 -1110 2 -060 2 -082 1 '26806 1 -1340 2 -660 2 -660 It should be pointed out , however , that the fact that " normal " values are obtained in any case must not be taken without further evidence as proof that the molecules present in a solution are essentially of the presumed dimensions .
As the presence of associated molecules Mm\#151 ; i.e. , polymorphs of the fundamental molecule or monad M\#151 ; would condition a smaller effect , whilst the attraction of molecules of the solvent by those of the solute would increase the effect , opposing tendencies may be at work unperceived ; especially is this true of electrolytes : the difficulty must always be met with in their case .
It follows from this argument that the " normal " effect of a monad cannot be determined in any absolute manner by a purely physical method .
Thus the value 1*85 used in calculating the molecular depression of the freezing point , which is based on the observation of the depression produced by alcohol and similar non-electrolytes , is presumably too high , inasmuch as alcohol , in all probability , exercises no inconsiderable attractive effect in solution .
The numbers quoted by Morse and Frazer show in the clearest manner possible that such an effect is at work and that it comes more and more into evidence as the solution is made more and more concentrated .
When the values in columns A and D in the above table are inspected , it is obvious that whilst the depression of the freezing point becomes greater at an increasing rate , the density increases at a corresponding diminishing rate\#151 ; the one 268 Prof. H. E. Armstrong .
[ June 14 , compensating the other ; to judge from the depression of the freezing point , the attractive forces within the solution become more and more manifest as the concentration increases.* !
At the conclusion of their memoir , Morse and Frazer express the opinion that " some at least of the abnormalities ( in the depression of the freezing point ) will disappear when the problem is studied from the side of the volume relations of solution and pure solvent " : I venture to think that it is already clear that the apparent abnormalities may easily be explained .
Much capital has been made of the additive nature of the properties of electrolytes and of the possibility of assigning specific mobilities to the ions as evidences of their dissociation .
But it is forgotten that the properties of many organic compounds which no one dreams of representing as dissociated are additive .
My object in referring to this contention here , however , is to lay emphasis on the conclusion which I stated in my paper on " Electrolytic Conduction " presented to the Society in 1886J that hydrogen and the metals generally may be regarded as the analogues of the CnH2n+1 hydrocarbon radicles and that their compounds with negative elements may be likened to unsaturated hydrocarbons of the form CwH2ra+i.CH : CH2 .
Subsequent study has confirmed this view .
1 have little hesitation now in stating the opinion that in cases in which they function as simply as carbon * Compare R. J. Caldwell , post , p. 272 .
t [ In a later paper , which has just been published ( ' Amer .
Chem. Journ. , ' July , 1906 , vol. 36 , p. 39 ) , Morse and Frazer confirm the rule for cane-sugar , Aj = 1'85 ND = A. But it does not hold good , they say , for glucose , the freezing points of glucose solutions being normal in the sense that the magnitude of the depression is strictly proportional to the weight-normal concentration .
On the other hand , they describe a new series of determinations of the osmotic pressures developed in solutions of cane-sugar , the results of which " are probably as precise as we can hope to obtain under present conditions .
" They are certainly in remarkable agreement with those calculated on the assumption that " cane-sugar in aqueous solution exerts an osmotic pressure equal to that which it would exert if it were gasified at the same temperature and the volume of the gas were reduced to that of the solvent in the pure state .
" They surmise , in order to account for the different results arrived at by the two methods , that at the comparatively high temperatures ( 20'\#151 ; 25 ' ) at which the osmotic pressures were determined , the whole of the water plays the part of solvent , while at lower temperatures a portion of it is appropriated by the solvent , giving rise , to an abnormal depression of the freezing point .
Mr. Caldwell 's experiments , which were made at 25 ' , appear to confirm the conclusion arrived at by Lord Berkeley and Mr. Hartley that the osmotic pressures are in excess of the calculated values , although it is true that they worked at 0 ' .
It is difficult to avoid the conclusion that the American observers ' values may be too low , perhaps on account of the lack of rigidity in the rubber stopper used in closing the cell .
The " normal " behaviour of glucose may be the consequence of the formation of associated molecules compensating the dehydration effect.\#151 ; Note added July 31 .
] + ' Roy .
Soc. Proc. , ' vol. 40 , p. 268 .
The Origin of Osmotic Effects .
1906 .
] does in the paraffinoid compounds generally , metallic atoms are characterised as definitely as the paraffinoid or CnH2n+i radicles are by specific physical attributes ; in the one case as in the other the so-called constants are merely algebraic differences .
Variations arise probably only when the metallic atoms become united among themselves by more than single affinities , as in the case of carbon compounds .
But the CnH2\#171 ; +i radicle in the compound C"H2n+i.CH : CH2 is externally inoperative\#151 ; combination is effected solely through the agency of the CH : CH2 radicle .
Applying this argument to metallic salts , the chlorides , for example , the metal may be regarded as inert in solution and the hydration of the salt may be supposed to take place solely through the combination of water with the halogen radicle .
In the language of the ionic dissociation hypothesis , only the negative ion is hydrated .
Probably , in all its interactions , the activity of a salt is primarily traceable to the negative ion.* From the electrolytic and osmotic points of view , mercuric chloride , HgCl2 , is one of the most remarkable of salts , its behaviour being almost that of a non-electrolyte .
It may be contrasted with potassium and calcium chlorides , the former of which has been shown by E. H. Griffiths to depress the freezing point of water at excessively low dilutions to just twice the extent to which it is depressed by the non-electrolyte cane-sugar ; the calcium salt has a still greater effect , approaching to three times that produced by the non-electrolyte .
Whilst mercuric chloride dissolves with difficulty in water and is very volatile , it is very readily decomposed into its elements by heat .
Sodium and calcium chlorides are easily soluble salts , but neither do they volatilise nor do they decompose when heated , except at very high temperatures .
Yet it is confidently argued that potassium and calcium chlorides undergo complete dissociation in solution into free ions , whilst mercuric chloride remains all but unchanged .
It is difficult to consider such a contention seriously .
In what manner then are the peculiarities of mercuric chloride * As to the meaning and use of the term ion and its derivatives , it may be pointed out that\#151 ; pending judgment\#151 ; no idea of dissociation need nor should be involved in the use of the term .
All ions are radicles , but all radicles are not ions : the ion is a radicle which can be moved or separated from a compound by means of an electric current and which is also at once " attackable " under appropriate conditions .
For example , the chlorine radicle is present as an ion in sodium chloride , but not in the chlorates nor in chloracetic acid .
Used in this way , the term is a most valuable accession to scientific language ; it is not until it is coupled with the term dissociation that the radically different conception is introduced about which dispute rages .
Faraday applied the term ion simply to " those bodies which can pass to the electrodes " : he in no way implied that they enjoyed separate existence in the electrolyte .
We may well rest satisfied to use the word in the sense in which he used it .
Prof. H. E. Armstrong .
[ June 14 , and similar chlorides to be explained ?
Conventionally , the three chlorides are represented by the formulae Hg\lt ; Cl Na\#151 ; 01 Ca\lt ; Cl But these symbols are equally misleading and inexpressive of their properties .
Solid mercuric chloride is probably a weak polymorph , but even its monads must differ in constitution from those of the sodium and calcium haloids .
As the mercury haloids all combine readily with other haloids , it cannot be supposed that the halogen is specially exhausted by combining with mercury : therefore , regarding the halogens as potential triads , it may be supposed that in mercuric chloride monads the atoms form a closed system , perhaps thus Sodium and calcium chlorides , on the other hand , may well be open systems .
Na\#151 ; Cl= Ca\lt ; Cl= Cl= The closed system of the mercury salt is but slightly attractive of molecules of its own kind or of water ; the potential activity of the chlorine radicle becomes manifest , however , when the salt is brought into contact with other chlorides which combine with it , forming well-characterised mixed chlorides , e.g. , or it may be / Cl Hg\lt ; H+2KC1- C1 ClmClK = Hg\lt ; XC1\#151 ; C1K / Cl- Hg\lt ; | +2KC1= : Cl- Cl\#151 ; C1K = Hg\lt ; | | NC1\#151 ; C1K Chlorides such as those of potassium and calcium are undoubtedly polymerised in the solid state ; when dissolved , the polymorphs undergo simplification at the instance of the water molecules , becoming more or less hydrated .
The extent to which depolymerisation and hydration take place depends both on the salt and the degree of dilution ; apparently it is only complete in very dilute solutions , judging from the fact that limiting values are only obtained in very dilute solutions .
The formation of hydrates containing several molecules of water may be supposed to involve the association of the salt with water complexes , or it may be that more or less complex ring systems are formed of water 1906 .
] The Origin of Osmotic Effects .
271 monads corresponding to the ring systems which are so common among organic compounds .
But from the point of view advocated in this note , the complexity of the hydrate would seem to be without influence on the osmotic effect exercised by the salt .
To explain the " double effect " produced by potassium chloride monads , it may be supposed that the molecule as a whole has effect\#151 ; such as is exercised by molecules generally in freeing water monads\#151 ; and that , in addition , the chlorine radicle exercises its 'proper attractive effect .
Other chlorides and salts generally may be expected to exercise an influence in proportion to the number of active " acidic centres .
" To judge from the behaviour of the sugars and alcohols , it is not improbable that the hydroxyl group may also function as an acidic centre .
But inasmuch as chlorides such as mercuric chloride have little or no attractive power for water , it is to be expected that the chlorine radicle or any other equivalent acidic centre will not in all cases exert the full unit effect , but that it will exercise an influence varying from unity to zero , according to the nature of the positive radicle with which it is associated ; in other words , the degree of so-called " ionisation " or attackability will vary with the nature of the molecule of which the acidic centre is a component .

