376 



NATURE 



[November 17, 192 1 



let us say, n. Owin^ to the fact that the emitted 

 electrons may originate from different depths in 

 the metal, and may undergo collisions at irregular 

 intervals, it is only the maximum kinetic energy 

 of those which escape which we should expect to 

 exhibit simple properties. As a matter of fact, 

 it is found that the maximum kinetic energy is 

 equal to the difference between the actual fre- 

 quency n and the threshold frequency n^ multiplied 

 by Planck's constant h. In mathematical sym- 

 bols, if V is the velocity of the fastest emitted 

 electron, m its mass, e its charge, and V the 

 opposing potential required to bring it to rest, 



eV = ^ VI \^ = h (w — no). 



From this equation we see that the threshold fre- 

 quency has another property. It is evidently that 

 frequency for which kinetic energy and stopping 

 potential fall to zero. This suggests strongly, I 

 think, that the reason the electron emission ceases 

 at tiq is that the electrons are not able to get 

 enough energy from the light to escape from the 

 metal, and not that they are unable to get any 

 energy from the light. 



The threshold frequencies have another simple 

 property. If we measure the threshold fre- 

 quencies for any pair of metals, and at the same 

 time we measure the contact difference of poten- 

 tial K between them, we find that K is equal to 

 the difference between their threshold frequencies 

 multiplied by this same constant h divided by the 

 electronic charge e. 



These results, as well as others which I have 

 not time to enumerate, admit of a very simple 

 interpretation if we assume that when illuminated 

 by light of frequency n the electrons individually 

 acquire an amount of energy hn. We have 

 seen that in order to account for thermionic 

 phenomena it is necessary to assume that the elec- 

 trons have to do a certain amount of work w to 

 get away from the emitter. There is no reason 

 to suppose that photo-electrically emitted electrons 

 can avoid this necessity. Let us suppose that 

 this work is also definite for the photoelectric 

 electrons and let us denote its value by hriQ. Then 

 no electron will be able to escape from the metal 

 until it is able to acquire an amount of energy at 

 least equal to hn^ from the light — that is to say, 

 under the suppositions made — until n becomes at 

 least as great as rig. Thus Wq ^^^^ be identical 

 with the frequency which we have called the 

 threshold frequency, and the maximum energy of 

 any electron after escaping will be h (n — no). 



The relation between threshold frequencies and 

 contact potential difference raises another issue. 

 We have seen that the contact potential difference 

 between two metals must be very nearly equal to 

 the difference between the amounts of work w for 

 the electrons to get away from the two metals by 

 thermionic action, divided by the electronic charge 

 e. The photo-electric experiments show that the 

 contact electromotive force is also nearly equal to 

 the differences of the threshold frequencies multi- 

 NO. 2716, VOL. 108] 



plied by '•/«. It follows that the photo-electric 

 work huQ must be equal to the thermionic work w 

 to the same degree of accuracy. The photoelec- 

 tric and thermionic works are known to agree to 

 within about one volt. To decide how far they 

 are identical needs better experimental evidence 

 than we have at present. The indirect evidence 

 for their substantial identity is stronger at the 

 moment than the direct evidence. 



I do not think that the complete identity of the 

 thermionic work w and the phpto-electric hriQ is a 

 matter which can be inferred a priori. What we 

 should expect depends to a considerable extent on 

 the condition of the electrons in the interior of 

 metals. We cannot pretend to any real know- 

 ledge of this at present ; the various current 

 theories are mere guesswork. Unless the elec- 

 trons which escape all have the same energy when 

 inside the metal we should expect the thermionic 

 value to be an average taken over those which get 

 out. The photo-electric value, on the other hand, 

 should be the minimum pertaining to those internal 

 electrons which have most energy. The apparent 

 sharpness of the threshold frequency is also sur- 

 prising from some points of view. There seems 

 to be scope for a fuller experimental examination 

 of these questions. 



I have spoken of the threshold frequency as 

 though it were a perfectly definite quantity. No 

 doubt it is when the condition of the body is or 

 can be definitely specified, but it is extraordinarily 

 sensitive to minute changes in the conditions of 

 the surface, such as may be caused, for example, 

 by the presence of extremely attenuated films of 

 foreign matter. For this reason we should accept 

 with a certain degree of reserve statements which 

 appear from time to time that photo-electric action 

 is some parasitic phenomenon, inasmuch as it can 

 be made to disappear by improvement of vacuum 

 or other change in the conditions. What has 

 generally happened in these investigations is that 

 something has been done to the illuminated surface 

 which has raised its threshold frequency above 

 that of the shortest wave-length in the light 

 employed in the test. Unless they are accom- 

 panied by specific information about the changes 

 which have taken place in the threshold frequency, 

 such statements are of little value at the present 

 stage of development of this subject. 



Light and X-rays. 



One of the great achievements of experimental 

 physics in recent years has been the demonstration 

 of the essential unity of X-rays and ordinary light. 

 X-rays have been shown to be merely light of par- 

 ticularly high frequency or short wave-length, the 

 distinction between the two being one of degree 

 rather than of kind. The foundations of our 

 knowledge of X-ray phenomena were laid by 

 Barkla, but the discovery and development of the 

 crystal diffraction methods by v. Laue, the 

 Braggs, Moseley, Duane, and de Broglie have 

 established their relations with ordinary light so 



