A Test for Polarization of Electron Waves by Reflection ^ 



By C. J. DAVISSON and L. H. GERMER 



A homogeneous beam of electrons is directed at 45° incidence against a 

 {111 i-face of a nickel crystal. The beam regularly reflected from this face 

 impinges upon a second similar face at the same incidence angle. A Faraday 

 collector is set to receive electrons regularly reflected from the second crystal, 

 but only such electrons are accepted into the collector as have survived the 

 two reflections without appreciable loss of kinetic energy. The collector and 

 second crystal are rigidly joined, and may be rotated about the axis of the 

 beam proceeding from the first to the second crystal. Measurements of the 

 intensity of the twice reflected beam have been made at bombarding poten- 

 tials from 10 to 160 volts. Within this range selective reflections (intensity 

 maxima) are observed at 20, 55, 77, 103 and 120 volts. 



These five selectively reflected beams have been separately tested for 

 polarization by measuring the current received by the collector as a function 

 of the azimuth of the movable system. If electron waves are polarized by 

 reflection the intensity of the twice reflected beam should be greatest when 

 the planes of incidence of the two reflections coincide, and least when they 

 stand normal to one another. No such variation of the current to the 

 collector is observed within the limits of error of the measurements — about 

 one half of one per cent of the total current. Our observation is that electron 

 waves are not polarized by reflection. 



THE experiment described in this article was undertaken to deter- 

 mine whether or not a beam of electron waves is polarized by 

 reflection from the surface of a nickel crystal. It is similar in certain 

 respects to the experiment with double Norrenberg mirrors by which 

 one demonstrates the polarization of light by reflection from glass, and 

 in others to the experiment by which Barkla established that X-rays 

 may be polarized. It resembles most closely, however, the variation 

 of the Barkla experiment performed by Mark and Szilard in which the 

 first of the radiators was a crystal and a Bragg reflection beam pro- 

 ceeded to the second radiator. A homogeneous beam of electrons is 

 directed at 45 degrees incidence against a {111} -face of a nickel crystal, 

 and the beam proceeding in the direction of regular reflection from this 

 crystal is then reflected at the same angle of incidence from a second 

 similar crystal. A double Faraday box is placed to receive electrons 

 which have been regularly reflected from the second crystal, but only 

 such electrons are allowed to enter the collector as have retained all or 

 nearly all of their kinetic energy through the two reflections; those 

 which have lost more than a small fraction of their kinetic energy are 

 excluded by a retarding potential of suitable strength. 



The second crystal and the collector are joined rigidly together, and 

 may be rotated about an axis which coincides with the axis of the beam 

 proceeding from the first to the second crystal. It is possible, there- 



^Phys. Rev., Vol. 2>i, May, 1929, pp. 760-772 



466 



