﻿

1 lieory of Contact Electromotive Force, 269 



the cases in which the surface-films of gas are removed by 

 heating it is to be expected that the oxide films which remain 

 will be non-conducting, and this attitude is definitely sup- 

 ported by the fact that Greinacher was unable to get rid of 

 the contact electromotive force by any kind of treatment 

 when one of the metals tested was platinum, which is 

 notoriously difficult to oxidize. 



When two of the metals are joined by an elecirolyte there 

 will be no true equilibrium until the battery thus formed has 

 run down, so that the potentials of the metals in the enclosure 

 will not be the same as those they possess in the steady 

 state. The present point of view does not, in fact, affect 

 the thermodynamic theory of electrolytic cells, according to 

 which their electromotive forces are determined by the heat 

 generated by the chemical actions involved. 



The Peltier Effect. 



We shall now return to the case of two metals in contact. 

 Consider the work done in taking an electron from one to 

 the other in two ways : first, through the junction, and 

 second, by crossing the outer surfaces. We thus obtain two 

 equivalent expressions for the work P per unit charge, viz. 



p=! ^-^ 2 Y Y M 1*1. 



The second formula agrees with that given by J. J. Thomson* 

 for the Peltier effect, but differs by a factor of two from the 

 corresponding expression found by Drude f. Thomson's 

 calculation is based on the idea that the electrons are in 

 equilibrium under the opposing action of the electric force 

 and the pressure gradient at the boundary, whereas Drude 

 supposes a kind of convective equilibrium between the dif- 

 fusion of the ions arising from the varying concentration 

 and the electric currents arising from the balancing difference 

 of potential. It seems to the writer that the standpoint 

 taken by Thomson is much the safer of the two. 



It does not seem quite certain, without further considera- 

 tion, that the expression — log— 1 will represent the heating 



effect due to the passage of an electric current across tbe 

 junction of two dissimilar metals, as the energy might con- 

 ceivably be changed, partly or entirely, into some form other 

 than heat. The fact may, however, be established by the 



* 'Corpuscular Theory of Matter,' p. 74. 

 t Ann. dcr P/n/sik, vol. i. p. 590 (1900) 



