Vol. 6, 1920 
CHEMISTRY: HARKINS AND EWING 
53 
enough to completely fill its pores and voids. The same charcoal dried 
for ten days at 120 to 150° C, and then cooled over phosphorus 
pentoxide, had a smaller heat of adsorption, 7.0 calories, while it gave 
only 3.5 calories when it held a thin film of water equal to 0.04 gram 
per gram of charcoal. Thus the first tenth of the water which penetrates 
the pores gives rise to nearly all of the evolution of heat. Such experi- 
ments as this are now being carried out by precise methods, carefully 
measuring each amount of water added, in somewhat the same way as 
Titoff carried out his experiments on the heat of adsorption of gases, but 
with a number of added precautions. The results indicate that, in accord 
with the theory of Langmuir^ and in agreement with the work of this 
laboratory on surfaces, the forces of adsorption or adhesion act only on an 
extremely thin layer of liquid. 
The heat of adsorption of bone charcoal which had not been outgassed 
was found to be 18*5 calories, and in fullers' earth it was 32 calories, or 
almost the same values as were obtained by Gurvich^ (18.5 and 30.2). 
Thus both of these solids have a very much higher heat of adsorption than 
even the best grades of cocoanut shell charcoal. While commonly known 
as the heat of adsorption, this is more precisely designated as the heat of 
spreading of a liquid on the surface of a solid. If — Q is taken to repre- 
sent the heat given off per square centimeter of surface; 7, the free surface 
energy; /, the latent heat of the surface; —AE, the decrease of surface 
energy in spreading, and the subscript sp represents spreading; s, the solid, 
a and t, the interface, then 
-Qsp = - AE,, = y, + I, - {y, + h). (1) 
The adhesional heat and energy are defined by the equation 
-Qa =E^=~AE,= ys + ls + y, + l,- (t; + h) (2) 
or 
-Qsp = -QA-iyi + h)- (3) 
The heat of adsorption of enough gas to form a liquid in bulk, covering 
the solid surface at constant temperature (Qr), is given by the equation 
-Qt = 7s^ls- {yi + y + (4) 
where X is the latent heat of vaporization of the liquid per unit volume of 
vapor. While we do not have the values of these quantities in the case 
of carbon, they have been determined in this laboratory for the interfaces 
between either mercury or water, and other liquids. The Pouillet effect, 
which is the heat given out when a liquid is adsorbed by a powder and is 
entirely analogous to that described here, has been attributed to the com- 
pression of the surface film of the liquid in contact with the powder. This 
compression, which was merely assumed from the standpoint of ideas on 
molecular attraction, is practically the same as that for which moderately 
direct evidence has been given in this paper. However, the above equa- 
tions, which were developed by Harkins, show that the heat comes from 
