]\gt ; The General Theory of Colloidal Solutions .
By W. B. HARDY , F.R.S. ( Received March 13 , \mdash ; Read IsIay 16 , 1912 .
) A large } ) of observations upon the filterability and the optical properties of colloidal solutions prove that they are coarse grained and consist of particles dispersed throughout a continuous fluid phase .
For this reason these solutions are commonly described as multiphase systems in which the dispersed particles form one of the phases .
This description is wrong unless it be qualified by the further statement that the particles do not constitute a phase in esse , but only a phase in po lise .
Consider a heterogeneous fluid such as is described above in contact with its own vapour , and let the components be two , e.g. elatine and water .
If the fluid really consisted of two phases there would , with the vapour , be three phases .
Let us distinguish the vapour phase by and the exterior and interior fluid ( or solid ) masses by and ' respectiyely .
Then , since the surface enclosing the interior fluid ( or solid ) masses is curved , we have The variables , therefore , are the two components , temperature , and two pressures , or five in all .
Since by hypothesis there are three phases , the system has two degrees of freedom .
Therefore , in order completely to define the system , it will be necessary to fix the temperature and one of the pressures , or the composition of the phases .
But on this hypothesis the two fluid phases are separated by a curved surface , and this introduces a difficulty ; for if the surface be freely permeable by the components the potentials and pressures in the two phases must be the same , and , as neither component is confined to the interior phase or to the interface , that phase is unstable .
* The phase rule is applicable only to contiguous masses which are thoroughly stable , whereas the stability of the interior masses in colloidal solutions is open to question .
From the fact that the state of the solution at any instant is determined not only by temperature , pressure , and composition , but also by previous history , we may conclude that neither the exterior nor the interior fluid masses conform to Gibbs ' criterion of stability , namely , that the value of the expression when , are respectively intrinsic energy , entropy , volume , and mass , and , and constants , * Gibbs , ' Trans. Conn. Acad vol. 3 , p. 406 .
In order to obtain comparable results for osmotic pressure or viscosity , it is necessar ) to start the system from a fixed point and allow it to proceed along a particular path .
Mr. W. R Hardy .
[ Mar. 13 , shall be zero for the particular phase and either zero or positive for any other phase of the same composition .
" " If equilibrium subsists ( in such systems ) without passive resistance to change it must be in virtue of properties which are peculiar to small masses surrounded by masses of different nature and which are not indicated by the fundamental equation.\ldquo ; * There is no intention at this stage of attempting to elaborate a complete theory , and enough has been said in support of the contention that it is almost .
as erroneous to speak of colloidal fluids as multiphase systems without qualification as it would be to ignore their heterogeneity altogether .
The nature of the problem and the particular questions which need investigation may be brought out by considering a simplified diagrammatic scheme of a system consisting of a heterogeneous fluid in contact with its own vapour .
Let it be supposed that a solution of A in is homogeneous and forms a simple phase when the concentration of A is sufficiently small , but becomes eous by the separation of small globules rich in A and poor in when the concentration rises above a point defined by temperature and pressure .
It would obviously be possible to produce the required concentration by compressing the mass in a cylinder with a piston permeable only to the solvent B. Let this be carried out and let the system now consist of vapour of , pure , and a heterogeneous fluid consisting of a weak solution of A in , with globules of solution rich in A dispersed through it , and let us agree to distinguish the vapour phase by , the pure solvent by the globules by ' and the external solution by ' ' .
The piston which separates solvent from the complex solution is maintained in position by a force which is obviously equal to the osmotic pressure of the solution .
Call this pressure Let the volume of the space under the piston be , and let there be only one globule present .
For a small compression which expels of solvent and causes the globule to increase by we have , ( 1 ) and , if be the mass and the potential of the solvent , and and the mass and potential of the solute , and the tension and the area of the surface of the globule , we have for the thermodynamic potential \mdash ; p'dv ' ( 2 ) * Gibbs , , p. 169 .
For the limitation of this supposition in connection with infinitesimal changes of state see Gibbs , pp. 1912 .
] The General Theory of Solutions .
an equation which is true only if the specific volumes of the components be reckoned the same throughout the fluid parts of the system .
But if the piston is completely permeable to the solvent , and the surface of the globule is completely permeabie to both solvent and solute , have ; ( 3 ) nd since and , ( 4 ) all the terms including these quantities disappear .
Also , if be the pressure on the piston\mdash ; that is the osmotic pressure\mdash ; and be the excess pressure in the interior of the lobule due to the tension of the surface , then and Equation ( 2 ) now reduces to ( ta ) .
Dividing through by , and remembering that , we get ( ta : ) , and from ( 5 ) , - In colloidal solutions there is usually a difference of electric potential between the and the exterior fluid due to polarisation of the interface .
The electric density must be a function of the chemical nature and density of the matter on each side of the interface , since the polarisation is due to chemical and osmotic forces acting across it .
Therefore , if the globule be so large as to contain in its interior a phase fully formed ( that is if the radius be greater than the effective range of any of the molecules there present ) electric density will be sensibly independent of the size of the globule , being , like the pressures and potentials of the interfacial film , fixed by the phases on each side supposed constant .
But when the radius is less than the range of molecular action the composition and physical properties of the globule will vary , and therefore the density of the charge on the surface will also vary .
An expression for the energy of an interfacial layer expressed in terms of * The latter part of this assumption probably is not of actual colloidal systems , for if any degree of mechanical stability be iml , arted to the interface by , e.g. , an eIectric charge , the great difference in diffusibility between solute and solvent ( leading to great differences in the rate of transference across the interface ) will operate in finite time as though the surface were only partially permeable to the solute .
Mr. W. B. Hardy .
[ Mar. 18 , tensions may be got by the method of Dupre* : " " If denote the interfacial tension , the energy corresponding to unit of area of the interface also as we see by considering the introduction ( through a fine tube ) of one body into the interior of the other .
A comparison with another method of generating the interface similar to that previously employed when but one body was in question will now allow us to evaluate " " The work required to cleave asunder the parts of the first fluid which lie on the two sides of an ideal plane passing through the interior is , per unit of area , , and the free sur.ace produced is two units in area .
So for the second fluid the corresponding work is .
This having been effected , let us now suppose that each of the units of area of free surface of fluid ( 1 ) is allowed to approach normally a unit area of ( 2 ) until contact is established .
In this process work is gained which we may denote by for each pair .
On the whole , then , the work expended in producing two units of interface is , and this , as we have seen , may be equated to 2 Hence ( 7 ) Experiments with plane surfaces ( see next paper ) suggest that the term consists mainly of the electric energy per unit area of interface due to electric polarisation .
But as this is not established it will be well to proceed in a more general way .
Let represent the tension which the interface would have if no polarisation occurred , then putting as the electric density on the surface of the globule , we have for the actual tension , ( 8 ) a relation which implies that the total energy per unit area of interface fixed by the mass of the components and the temperature .
From ( 6 ) and ( 8 ) we get .
( 9 ) As an approximation , the electric displacement at the interface may be represented by two concentric shells , each infinitely thin , and with electric density and .
The radius of the inner shell being the radius of the globule , and the distance between the shells , the energy due to polarisation is then given by .
* Lord Rayleigh , ' Phil. Mag 1890 , 5 ) , vol. 30 , p. 461 .
Equation ( is rigorously true , for , if account be of any gained or expended in the formation or condensation of vapour , the balance of such work is zero , if Dalton 's law be assumed .
1912 .
] The Theory of Colloidal Solutions .
Putting OIlstant , we have .
( 10 ) If the solute A were an electrolyte , account would have to be taken of the electric work or lost in phases ' and ' 'during the compression .
* Equations ( 9 ) and ( 10 ) show the osmotic pressure of schematir colloidal solution such as we are considering depends upon the functions and When exceeds the range of molecular action , the globule contains in its interior a phase ( the interior phase ) fully formed , but under a pressure greater than the pressure of the external phase by the quantity .
Let us make the probable assumption that , when the lobule has grown to this size , so as to constitute a phase in mass , the pressure has fallen so low as to cease to affect sensibly the composition or properties of the internal phase .
Since the pressure and potentials of the layer of transition between phases ' and ''are fixed when those phases are fixed , and become equal to zero , and the equation becomes .
( 9 ) The point where is equal to the extreme range of nlutual action of any of the molecules present is therefore one of great importance .
Three things therefore need discussion : they are the forms of the functions and and the range of molecular action .
The Surface Tension of the Globul \mdash ; Lord Rayleigh has shown that , according to the Young-Laplace theory of capillarity , the tension of a very thin film of matter should increase as the film thickens , according to the square of the thickness .
Experiment with actual films has failed to confirm this generalisation .
The experiments of Reinold and and of JohannotS prove that , for soap lilms in air , the curve connecting tension and thickness shows a series of maxima and minima ( fig. 1 ) .
The tension in the case under consideration is that of a drop of liquid forming in the interior of a mass of liquid .
It is therefore that of a fluidfluid interface .
If , however , the function has maxima and minima , * In order to simphfy the problem , any kinetic energy which the globule may possess , owing to progressive motion , has been neglected .
This quantity has been experimentally investigated by Perrin for certain suspensions .
'Phil .
Mag 1892 , [ 5 ] , vol. 33 , p. 468 .
'Phil .
Trans 1886 , vol. 17 p. 679 .
S 'Phil .
Mag 1899 , [ 5 ] , vol. 47 , p. 501 .
Mr. W. B. Hardy .
[ Mar. 13 , certain important conclusions follow .
For let phase ' ' be supersaturated with respect to phase ' : by raising the osmotic pressure to , the result would be the formation of a globule of tension and radius .
But if the system contains globules of radius , and the osmotic pressure be reduced by admitting solvent , globules of tension T2 and radius will be iormed .
FIG. 1 .
At pressure , globules of radius can -exist , but those of radius and will probably be unstable .
According to the accepted theory , the of a film is constant so long as the thickness is greater than twice the range of molecular action .
That portion of the curve in the figure which lies between and relates to a system whose globules are so small that the radius is less than the range of molecular attraction .
It follows , therefore , that so long as the globule substance has not the properties of a phase in mass , i.e. so long as , a state in which the interior mass is present as globules is stable , and the globules l1lay be of different radii , also the osmotic pressure of the solution will not be fixed by fixing the ratio of the masses of the components and the temperature .
These are the characteristics of a oolloidal solution , and it is obvious that a determination of the form of the function is of fundamental importance in the theory of such solutions .
The evidence that in certain films tension varies discontinuously with thickness is complete .
Reinold and cker determined the thickness of the black area in a soap film to be about 12 , and found that at the edge of the area the film thickened abruptly to or more .
Since the tension of a horizontal film must be everywhere the same , the tension of the film thick is also that of the thicker film .
The abruptness of the transition between thin and thick regions can only mean that films 1912 .
] The General Theory of Colloidal Solutions .
of intermediate thickness are unstable , that between certain thicknesses tension increases as thickness diminishes .
Johannot 's experiments make it probable that when the tension of the film is considered as a function of the thickness there is more than one maximum .
It is comnlonly held by physicists that these results obtained by the study of films of fluid immersed in air can be applied to variations of the tension of a globule of fluid immersed in its own vapour .
The likelihood of this being true depends upon the view that is taken of the significance of these maxima and minima .
If they are due to something fundamental and fixed in the nature of the field of force about an atom or a molecule , they will appear in all expressions for the of a mass or layer of matter whose dimensions are less than the range of the force .
showed that if the mutual attraction between the molecules of a fluid changes to a repulsion when the distance separating the molecules is less than a certain quantity , the tension of a film would begin to increase with decreasing thickness when the thickness was equal to this quantity , and Reinold and RuckerS suggest that the increase of tension which they found in the films 50 thick is due to this cause .
J. J. Thomson finds in the corpuscle theory of atomic structures justification for the view that the extra atomic field of force has zones of attraction and repulsion .
It is obvious on this theory that the field must be complex , for consider the simplest model of an atom\mdash ; a central core of stationary electrons surrounded by a plane ring of rotating electrons .
There will be a field due to the stationary electrons in which the electric intensity will vary inversely with the square of the distance from the centre of the atom .
And there will also be an electromagnetic field due to the rotating ring in which the 'intensity varies that radius which is at right angles to the plane of the rin inversely with the cube of the radius .
On purely theoretical grounds , therefore , Boscovich 's conception of a field of force about an atom whose sign changes as the distance from the centre of the field increases has some justification , and , as is well known , Kelvinl applied this conception to capillary phenomena .
The assumption * Reinold and Rucker , too .
Thomson , ' Conduction of Electricity through Gases , ' 2nd EditioI ] , p. 183 ; Thomson and Pointing , ' Properties of Matter , ' p. 168 .
' Encyc .
Brit art .
" " Capillarity S , p. 681 .
The Corpuscular Theory of Matter , ' London , 1907 , p. 158 .
'Mathematical and Physical Papers , ' vol. 3 , pp. 398 and 409 Mr. W. B. Hardy .
[ Mar. 13 , that maxima and minima occur in the variation of the tension with the radius of a drop of fluid adso offers a singularly simple and direct explanation of the influence of electrically charged nuclei upon the condensation of supersaturated vapour , , as we have seen , it would furnish an equally direct interpretation of the peculiarities of many colloidal systems .
Unfortunately much may be urged against the view that these maxima and minima in the tension of a soap film have any direct relation to the field about an atom or molecule .
It must be borne in mind at the outset that a drop of water immersed in its own vapour is a system composed of one component , and that variations in the tension of the drop with the radius can , therefore , be due only to the nature of the field of force between the molocules of water .
That is to say , it depends solely upon the form of the function in the Young-Laplace theory of capillarity .
But the soap films with which Reinold and Rucker and Johannot worked contained two components at least , and were composed of colloidal solution , that is , of a peculiarly complex type of fluid .
If maxima and minima did exist in the tension of a drop of pure water of radius less than the range of the mutual attraction of the molecules , it should be possible to form films of pure water just as it is possible to form films of soap solution .
This is not so .
A film of water breaks with very great rapidity ; in fact , it behaves as it should if tension decreased very rapidly wiffi the thickness of the film .
The repulsion suggested by Maxwell , and needed to account for the peculiarities of soap films , can perhaps be found in quite another direction .
In a previous paper I described the existence of a remarkable potential difference between a film of impurity , such as an oil , on the surface of water and the subjacent water .
It occurred to me , when conducting the exlJerimenCs , that if such a surface were lifted up to form a free film , as it can be lifted up by slowly a solid frame from the fluid , it would consist of three layers , two outer ones at a potential higher or lower , as the case may be , than the interior layer .
The arrangement would be such as is shown in the diagram ( fig. 2 ) .
' FIG. 2 .
J. J. Thomson , ' Conduction of Electricity through Gases , ' loc. 'Roy .
Soc. Proc 1911 , , vol. 84 , p. 220 ; compare also Beinold and Rucker , 'Phi .
Trans 1893 , , vol. 184 , p. 527 .
1912 .
] The Theory of Colloidal utions .
, are films of oil which may be from to thick , and at the interfaces , a constant difference of potential is found due to polarisation of the matter 011 either side .
Let the fluid drain away .
The positively charged portion of the fluid on either side is of finite thickness , owing to molecular movements .
When has drained away so that the extreme oscillations of the positively molecules at are into the electric field about the positive molecules at , there will be a lepulsion .
If the film thin further it must first entirely reconstruct itself The first state may be that of soap film showing Newtonian colours ; the second , that is after reconstruction , may be that of a black film .
This theory finds justification in a study of the actual filma The experiments and apparatus will be described llore fully in a later paper .
For the present it will be sufficient tu refer briefly to two typical cases .
A film , 1 ) , of soap , starch , or saponin solution in water , with air on each side , was so arranged that it could be inspected under the microscope while a constant current was pasf ; ed through it .
dust , or specially prepared French chaIk , was dusted on to the film , and also immersed in it , and it was noticed that the unimmersed particles moved in the electric field in a direction opposite to that of the immersed Measuring movement in the divisions of a micrometer eyepiece covered in five minutes , I found:\mdash ; Distance .
Saponin-water film\mdash ; Unwetted particles on the surface \mdash ; Middle of the film -- 12 Starch-water film\mdash ; Unwetted on surface Immersed That is to say , it was possible to observe particles at different levels in the film moving in opposite directions at the same time , the direction of the movement of each particle being reversed on reversing the direction of the electric current .
These observations can , I think , mean only one thing , namely , that such films are composed of layers between which there is a contact difference of potential .
Let us agree to call the various surfaces of such composite films interior and exterior surfaces .
Ihere will then be in the case of a film of fluid in air two exterior and two interior surfaces , and in the case of a film of one fluid on the surface of a mass of another fluid two surfaces , one interior and Mr. W. B. Hardy .
[ Mar. 18 , one exterior .
In the same way there will be in the former case one interior two exterior layers , and in the latter an exterior layer and an interior mass .
When all the layers have a thickness exceeding twice the range of action of any molecules present in the layers , the sum of the tension of the layers and the electric density at the interfaces will be constant and independent a variation in thickness of any or all the layers .
When the thickness of all or of any one layer is less than this magnitude the tension and the electric density vary , and may be considered functions of the thickness .
It of importance , therefore , first of all to determine whether the range of molecular action is a fixed value or whether it is dependent upon the chemical nature of the molecules concerned .
The question is considered in the next paper .
The Tension of Composite Fluiol Surfaces and the Stability of Films of Fluid .
By W. B. , F.R.S. ( Received March 13 , \mdash ; Read May 16 , 1912 .
) The tension of a composite surface composed of a fully separated layer one fluid ( say oil ) spread over a mass of another fluid ( say water ) is given by the equation , ( 1 ) where A and denote respectively the oil layer ( exterior layer ) and the water ( interior mass ) .
If be the depth of the exterior layer , then , for a given temperature the quantity will reach a constant value when is either twice the range of action of the oil molecules on each other , or the sum of two values , namely , the range of action of oil molecuIes on each other , and of the oil molecules on the water molecules .
Let be the lowest value of for which is constant : is equal to the greater of two quantities , the range of action of the molecules of oil , or the mean of this value plus the range of action of these molecules upon the molecules of water .
The attractive force between two molecules in the theory of capillarity is a force which decreases rapidly as the distance between the molecules increases .
Young assumed the force to be negative , that is to say ,

