596 BELL SYSTEM TECHNICAL JOURNAL 



disappearance of X-ray diffraction beams under similar conditions. 

 In other words, the wave-length resolving power of the crystal is not 

 so good for electron waves as for X-rays of comparable wave-lengths. 

 This is due to the fact that the penetration of electron waves into the 

 metal is very much less than the penetration of X-rays of the same 

 wave-length.^ 



It is, of course, just this slight penetrating power of electron waves 

 which made the diffraction patterns arising from our nickel crystal 

 sensitive to the presence of adsorbed gas, and caused the extraordinary 

 complexity in the observed phenomena. The circumstance, that 

 electron waves are scattered very efficiently by the surface atoms of the 

 crystal and are consequently extinguished on penetrating into the 

 crystal at a very rapid rate, opens up to us the possibility of the use of 

 electron diffraction as a means of studying surfaces. 



The first application of this new method of surface study was the 

 analysis (Phys. Rev., loc. cit.) of the two transient diffraction patterns 

 which existed for only a short time after the crystal surface was cleaned 

 by heating. The first of these patterns to appear after the heating was 

 found as soon as the crystal was comparatively cool. This pattern I 

 shall refer to as the electron diffraction pattern of the third type. It 

 consisted of electron beams emerging near to grazing the surface in the 

 principal azimuths of the crystal, occurring in each azimuth just as if 

 the surface of the crystal were a plane diffraction grating. For each 

 diffraction beam the grating constant was equal to the separation be- 

 tween the rows of atoms on the surface of the crystal normal to the 

 azimuth of the beam. Figures showing beams of this type in the prin- 

 cipal crystal azimuths were exhibited in our original paper (loc. cit., 

 Figs. 14-16). 



The change in intensity with time of a typical beam of this third 

 type after the crystal surface was cleaned by heating is shown by the 

 curve marked "Type-3" in Fig. 5. (The crystal was not cool until 

 nearly twenty minutes after the heating.) From the positions of these 

 "plane grating" (Type-3) beams near to grazing the crystal surface 

 and from their behavior with time after the surface was cleaned, we 

 concluded that, when the beams of this type were most intense, the 

 crystal surface was completely free from adsorbed gas. We concluded 

 also that, until these beams had become quite weak, the surface was 

 not covered by so much as a single layer of gas atoms. The condition, 



^ The lower curves in Fig. 3 represent the data from which we calculated {Phys. 

 Rev., loc. cit.) the rate of extinction of 54 volt electrons (X = 1.67 A.) on penetratins; 

 into the metal. A similar calculation cannot readily be made from the data of the 

 lower curves in Fig. 4. These matters were considered in detail in our original paper 

 and need not be discussed here. 



