1910-11.] Temperature Coefficient of Concentration Cells. 375 
XXIV. — On the Temperature Coefficient of Concentration Cells, in 
which the same salt is dissolved in two different solvents. 
By A. P. Laurie, M.A., D.Sc. 
(MS. received February 3, 1911. Read February 20, 1911.) 
In a paper read before the Royal Society of Edinburgh ( Proc . Roy. Soc . 
Edin., xxviii., part v., p. 382 (1908) ; Zeit. phys. Ghem., lxiv. 5) I described 
a new type of concentration cell, in which the one platinum electrode was 
surrounded by a solution of ’025 molecules of KI containing *001 molecules 
of iodine dissolved in absolute alcohol, and the other electrode was 
surrounded by -025 molecules of KI and *001 molecules of iodine, dissolved 
in water. This cell developed a considerable E.M.F. of 198 volts at 25° C. 
in the direction which would transfer the iodine from water to alcohol 
and potassium iodide from alcohol to water. 
The temperature coefficient was positive and amounted to '00037 pr. 
deg., the E.M.F. rising with rise of temperature and the cell being there- 
fore endothermic. 
Cells of this type were originally investigated by Luther (Zeit phys. 
Ghem., xix. p. 567, 1896), who experimented with a cell n which zinc 
electrodes were surrounded by zinc sulphate dissolved in ether and water 
respectively; and he showed that the conditions for equilibrium, and 
therefore for the E.M.F., were determined by the partition coefficient of 
the salt between the two solvents. 
To take for instance Luther’s cell : if a solution of zinc sulphate in 
water is shaken up with ether, the zinc sulphate will share itself between 
the two solvents in a certain ratio. If two zinc electrodes are now 
introduced and the cell connected to an electrometer, no E.M.F. will be 
detected, as is obvious on thermodynamical principles. If there was an 
E.M.F. the zinc sulphate would be transferred during the passage of the 
current from one solvent to the other, again returning by diffusion, and a 
source of energy would be so obtained under conditions which are, as Luther 
has shown, thermodynamically impossible. 
If a solvent is chosen which mixes with water, the conditions are 
complicated by the fact that the two solvents diffuse into each other. 
But there is still a difference of solubility of the salt in the two solvents, 
which means that there is a partition coefficient equilibrium, though it cannot 
be determined in this case by shaking up the two solvents with the salt. 
