46 The Electron Microscope 



the contrast will reach this ideal value only if d^ is very large in 

 comparison with the object diameter. Therefore, the contrast 

 will even improve slightly with increasing value of this sort of 

 chromatic deject. 



In microscopes with a small physical aperture this effect is 

 much smaller, but still noticeable. Electrons which have suffered 

 a loss of AV ev will produce a border — dark in the original, 



white .n pnnts or reproducUons-of a wWth 4^^, measured 



in the plane of the object. Such chromatic borders have been 

 observed in many micrographs, especially by E. Ruska and- his 

 collaborators. Sometimes the borders do not show one step only, 

 as in figure 14b, but two, of about equal width. These chromatic 

 fringes can be explained by the discovery of G. Ruthemann ■^^ 

 that fast electrons lose their energy in solids mainly in discrete 

 steps, in carbon for instance in steps of 24 ev. It appears that 

 in most solids one type of collision is dominant, with a char- 

 acteristic energy loss so sharply defined that Hillier "^ could 

 base a method of microanalysis on it. 



Similar borders appear also in numerous micrographs of 

 objects supported by thin organic membranes. This is an indirect 

 proof that a large proportion of the electrons in the illuminating 

 beam traverses the foil without collisions. More direct evidence 

 is contained in Ruthemann's experiments. He found that a 

 large proportion of the electrons suffer no energy losses, and 

 this proportion is larger in very thin foils than could be ex- 

 plained by electrons which have suffered elastic collisions only. 

 A theoretical estimate, based on considering the solid as a 

 highly compressed gas, gives 80-90 per cent as the proportion 

 of electrons which will traverse a collodion foil of 100 A thick- 

 ness without collisions. The theory which has been sketched out 

 is therefore applicable to conditions under which most electron 

 micrographs are obtained. The essence of this theory is in short 

 that almost all electrons which have suffered inelastic collisions 

 in the object zvill produce contrast as if tJiey had been absorbed. 



But the theory can be extended to a certain measure also to 

 elastically scattered electrons, assuming that they are scattered 



