538 SECTIONAL TRANSACTIONS.— A. 



Tuesday, September 11. 

 Discussion on The Scattering of Electrons by Crystals. (Dr. C. J. 

 Davisson ; Prof. G. P. Thomson; Dr. L. de Broglie.) 



Dr. C. J. Davisson. — Experiments reveal that the interaction between a homo- 

 geneous beam of electrons and a cryst-al of nickel is similar to the interaction between 

 the same cr3'stal and a beam of monochromatic X-rays ; the electron beam is regularly 

 but selectively reflected from the face of the crystal (analogue of the Bragg X-ray 

 reflection phenomenon), and at critical speeds of bombardment beams of electrons 

 issue from the incidence side of the crystal in other directions as well (analogue of the 

 Laue X-ray diffraction phenomenon). Unlike the Laue beams, however, the electron 

 beams do not proceed from the crystal in the direction of regular reflection from its 

 principal sets of Bragg atom planes. That this simple geometrical relation obt-ains 

 in the case of X-ray diffraction is due to the fact that for X-rays the indices of refraction 

 of materials are equal sensibly to unit}'. It is due to this circumstance also that the 

 wave-lengths of Laue beams can be calculated by means of the Bragg formula. If 

 the refractive index were not unity the wave-lengths of Laue beams issuing from the 

 incidence side of the crystal could still be calculated, for, regardless of the value of 

 the index, the wave-length of each such beam satisfies the plane grating formula 

 jiX=D (sin 01 — sin 62) "ndth respect to one or another of the atomic plane gratings to 

 which the surface layer of atoms is equivalent. This formula has been used to 

 calculate equivalent wave-lengths of electron beams of various speeds, and in all 

 cases the values so found are in acceptable agreement with the values calculated from 

 the de Broglie formula 'k=}i/mv. The same procedure cannot be used to calculate 

 equivalent wave-lengths of the beam regularly reflected from the crystal face, for as 

 a plane grating beam it is of zero order and its wave-length is therefore indeterminat*. 

 The observations on this beam are particularlj' suitable, however, for calculating 

 indices of refraction of the crj'stal for electrons of various speeds or wave-lengths. 

 These calculations and others based on the data of the difiraction beams lead to a 

 definite dispersion curve for nickel. For bombarding potentials V greater than 

 100 volts the index [i is given by (jL=(l-|-*/V)i where * is a constant, equal approxi- 

 mately to 18 volts. In terms of the wave theory of mechanic's this means that the 

 average potential inside the crystal is less by 18 volts than that outside. Below 

 . 100 volts the dispersion curve is complicated and exliibits features suggestive of optical 

 anomalous dispersion. 



Prof. G. P. Thomson. — The passage of cathode raj-s through thin films of metals 

 or celluloid results in a difiraction of the de Broglie waves associated with the electrons. 

 This gives rise to a pattern similar to that formed by X-rays under the same conditions, 

 and the whole of the pattern can be predicted with an accuracy of about 1 per cent, 

 from the known crystal structure of the soUd and de Broglie's expression for the 

 wave-length. This has been proved by workers at Aberdeen University for celluloid, 

 aluminium, gold, platinum, silver, copper and tin. The electrons which form the 

 pattern have apparently the same speed as the original cathode rays. 



Rupp in Germany has extended these results to the cases of slow electrons of 

 about 300 volts, and finds a small discrepancy which he ascribes to a refraction of the 

 electron waves. 



A point of great theoretical interest is the length of the train of waves associated 

 with an electron, and the question of the relation to each other of the waves associated 

 with the different electrons in a beam of cathode raj's. The Aberdeen experiments 

 appear to show that each electron has a coherent train of at least fifty waves associated 

 with it. This number probably depends on the conditions of experiment. In the 

 case of the (3 rays from RaE the train is probably a very short one. 



Prof. W. J. DE Haas. — Some New Experiments on Supraconductors. 



The theories on electric conductivity may be divided into two groups. The one 

 treats the metal as containing a gas of electrons — degenerated or not. According 

 to the theories of the other group, the electrons are passing from one atom to the 

 Other. 



The writer thinks that the last point of view is the right one. It is possible 

 to demonstrate that in graphite placed in a magnetic field nothing can be detected 

 of a Lorentz force on the electrons. Other investigations on bismuth contradict the 

 results obtained with graphite. The experiments are not yet conclusive. 



