374 



NATURE 



FNOVEMBER 17, I92I 



is a question of fact. I have considered the data 

 very carefully. There is a little uncertainty in 

 some of the items, but u'hen every allowance is 

 made there seems no escape from the conclusion 

 that the photo-electric effect of the whole spectrum 

 is far too small to account for thermionic emission. 

 This question is an important one, apart 

 from the particular case of thermionic emission. 

 The same dilemma is met with when we seek for 

 the actual modus operandi of evaporation, chemi- 

 cal action, and a number of other phenomena. 

 These, so far as we know, might be fundamentally 

 either kinetic or photochemical or a mixture of 

 both. (I am using the term photochemical here 

 in the wide sense of an effect of light in changing 

 the composition of matter, whether the parts 

 affected are atoms, groups of atoms, ions, or elec- 

 trons.) For example, the approximation about 

 boiling points known as Trouton's rule is a fairly 

 obvious deduction from the photochemical point of 

 view. The photochemical point of view has 

 recently been put very strongly by Perrin, who 

 would make it the entire motif of all chemical 

 reaction, as well as of radio-activity and changes 

 of state. In view of the rather minor part it 

 seems to play in thermionic emission, where one 

 would a priori have expected light to be especially 

 effective, this is probably claiming too much for 

 it, but the chemical evidence contains one item 

 which is certainly difficult to comprehend from 

 the kinetic point of view. The speed of chemical 

 decomposition of certain gases is independent of 

 their volume, showing that the decomposition is 

 not due to molecular collisions. The speed does, 

 however, increase very rapidly with rising tem- 

 perature. What the increased temperature can 

 do except increase the number and intensity of 

 the collisions, factors which the independence of 

 A'olume at constant temperature show to be with- 

 out effect, and increase the amount of radiation 

 received by the molecules, is not too obvious. It 

 seems, however, that, according to calculations 

 by Langmuir (Journ. Amer. Chem. Soc, vol. 42, 

 p. 2190, 1920), the radiation theory does not get 

 us out of this difficulty ; for, just as in the ordinary 

 photoelectric case, there is nothing like enough 

 radiation to account for the observed effects. It 

 seems that in the case of these mono-molecular 

 reactions the phenomena cannot be accounted for 

 either by simple collisions, or by radiation, or by a 

 mixture of both, and it is necessary to fall back on 

 the internal structure of the decomposing mole- 

 cule. This is complex enough to afford material 

 sufficient to cover the possibilities ; but, from the 

 point of view of the temperature energy relations 

 of its parts, it cannot at present be regarded as 

 much more than a field for speculation. 



Contact Electricity. 

 A controversy about the nature of the contact 

 potential difference between two metals, similar to 

 that to which I have referred in connection with 

 thermipnic emission, has existed for over a cen- 

 tury. In 1792 Volta wrote: "The metals . . . 

 can by themselves, and of their own proper virtue, 



NO. 2716, VOL. 108] 



excite and dislodge the electric fluid from its state 

 of rest." The contrary position that the electrical 

 manifestations are inseparably connected with 

 chemical action was developed a few years later 

 by Fabroni. Since that time electrical investiga- 

 tors have been fairly evenly divided between these 

 two opposing camps. Among the supporters of 

 thQ intrinsic or contact view of the type of Volta 

 we may recall Davy, Helmholtz, and Kelvin. On 

 the other side we have to place Maxwell, Lodge, 

 and Ostwald. In 1862 we find Lord Kelvin 

 (" Papers on Electrostatics and Magnetism," p. 

 318) writing: " For nearly two years I have felt 

 quite sure that the proper explanation of voltaic 

 action in the common voltaic arrangement is very 

 near Volta's, which fell into discredit because Volta 

 or his followers neglected the principle of the con- 

 servation of force." On the other hand, in 1896 

 we find Ostwald (" Elektrochemie, Ihre Geschichte 

 und Lehre," Leipzig, 1896, p. 65) referring to 

 Volta's views as the origin of the most far-reach- 

 ing error in electrochemistry, which the greatest 

 part of the scientific work in that domain has been 

 occupied in fighting almost ever since. These are 

 cited merely as representative specimens of the 

 opinions of the protagonists. 



Now, there is a close connection between ther- 

 mionic emission and contact potential difference, 

 and I believe that a study of thermionic emission 

 is going to settle this little dispute. In fact, I 

 rather think it has already settled it, but before 

 going into that matter I would like to explain how 

 it is that there is a connection between thermionic 

 emission and contact potential difference, and 

 what the nature of that connection is. 



Imagine a vacuous enclosure, either impervdous 

 to heat or maintained at a constant temperature. 

 Let the enclosure contain two different electron- 

 emitting bodies, A and B. Let one of these, say 

 A, have the power of emitting electrons faster 

 than the other, B. Since they are each receiving 

 as wxU as emitting electrons, A will acquire a 

 positive and B a negative charge under these cir- 

 cumstances. Owing to these opposite charges A 

 and B will now attract each other, and useful 

 work can be obtained by letting them come in con- 

 tact. After the charges on A and B have been 

 discharged by bringing them in contact, let the 

 bodies be quickly separated and moved to their 

 original positions. This need involve no expendi- 

 ture of work, as the charges arising from the elec- 

 tron emission will not have had time to develop. 

 After the charges have had time to develop the 

 bodies can again be permitted to move together 

 under their mutual attraction, and so the cycle 

 can be continued an indefinite number of times 

 In this way we have succeeded in imagining a 

 device which will convert all the heat energy from 

 a source at a uniform temperature into useful 

 work. 



Now, the existence of such a device would, 

 contravene the second law of thermodynamics 

 We are therefore compelled either to deny the 

 principles of thermodynamics or to admit that 

 there is some fallacy as to the pretended facts in 



a 



I 



