October 20, 1916] 



SCIENCE 



567 



It has there been shown that such ideas can 

 be successfully employed for the general treat- 

 ment of the phenomena of lyophobic and lyo- 

 phyllic colloids. In the case of lyophyllic col- 

 loids it is pointed out that in general the sur- 

 face tension for undispersed dispersoid is neg- 

 ative and hence automatic dispersion takes 

 place until the size of particles is reached 

 which have zero surface tension. While for 

 lyophobic colloids large particles have positive 

 surface tension, and this only becomes zero 

 for very small particles. This necessitates a 

 preliminary dispersion by electrical, mechan- 

 ical or chemical means for the artificial prepa- 

 ration of lyophobic colloidal solutions which 

 unlike lyophyllic colloidal solutions are of in- 

 frequent occurrence in nature. The writer 

 has also discussed there the role of the elec- 

 trical charge always present on lyophobic col- 

 loidal particles in producing the state of zero 

 surface tension necessary for permanent sta- 

 bility. 7 



Freundlich 8 is perhaps the principal expo- 

 nent of a theory of colloidal solution which 

 does not take zero surface tension as the neces- 

 sary accompaniment of the stable colloidal 

 state. According to this theory the surface 

 tension at the boundary of the dispersoid par- 

 ticles is always positive and hence there is al- 

 ways a tendency for the particles to unite with 

 decrease of surface. The electrical charges on 

 the particles, however, by mutual repulsion pre- 

 vent such a union and keep the system in a per- 

 manent, although thermodynamically unstable 

 state. Although the writer would not deny 

 that there may be some colloidal solutions 

 which may be in a relatively permanent state 

 without having really reached a condition of 

 minimum of free energy, he believes, however, 

 that the Freundlich theory is entirely inade- 



7 If we wish to extend our considerations to the 

 case of particles so small that they contain only a 

 few ultimate molecules, it may seem somewhat mis- 

 leading to speak of a definite value of the surface 

 tension, and in that case it may seem more desir- 

 able to relate the free energy of the dispersoid 

 directly to the degree of dispersion, without in- 

 termediate considerations as to surface tension. 

 This, however, involves no change in principle in 

 our method of attack. 



8 Freundlich, ' ' Kapillarchemie, ' ' 1909. 



quate for a general treatment of colloidal 

 phenomena. Not only does the absolutely 

 permanent stability of colloidal solutions point 

 to true thermodynamic equilibrium, but the 

 actual growth of particles to a new equilib- 

 rium size on small additions of electrolytes to 

 colloidal solutions and their redispersion to the 

 old size on washing out the electrolyte could 

 only be the case if we have a real thermody- 

 namic equilibrium. Furthermore, Freund- 

 lich's assumption that an actual collision and 

 union of particles is necessary for a decrease 

 in degree of dispersion seems to be entirely un- 

 justified since with positive surface tension, 

 as is well known, the material in the smaller 

 particles would have a higher solubility than 

 that in the larger particles, and the latter 

 would grow at the expense of the former. 

 Indeed, the continuous growth of particles 

 from one equilibrium size to another is 

 evidence that some other process than that of 

 simple union is taking place. Finally, the ex- 

 istence of a definite equilibrium size of par- 

 ticle contradicts his theory since if the sta- 

 bility were due merely to an electrical repul- 

 sion that kept particles apart this would work 

 equally well with particles of all sizes, while 

 microscopic examination shows that in typical 

 lyophobic colloidal solutions all the particles 

 have the same size except for a few very 

 large ones which are floating around with the 

 others and are apparently so large that they 

 lie in the region of positive surface tension 

 which, as we have already seen, characterizes 

 undispersed lyophobic dispersoid. 9 



ElCHARD 0. TOLMAN 



Laboratory of Physical Chemistry, 

 University of Illinois 



9 There are of course in all probability some 

 lyophobic dispersoids in which the surface tension 

 is nearly zero for particles having a considerable 

 range of size, and in such cases even in a stable 

 solution there will be considerable variation around 

 the equilibrium size of particle. Indeed, in col- 

 loidal solutions in equilibrium, we shall expect in 

 general a distribution in the size of the particles 

 according to the laws of probability around that 

 size which has exactly zero surface tension, and the 

 more rapid the change of surface tension with di- 

 mensions the more nearly will all the particles be 

 of the same size. 



