800 THE BELL SYSTEM TECHNICAL JOURNAL, OCTOBER 1951 



results in a somewhat different shading distribution. The distributions are 

 mathematically formulated in Appendix I. 



1.1 Shadows 



In certain regions the local surface is not exposed to the source, resulting 

 in "shadows" within which te is zero. These shadows are bounded by two 

 lines, the shadow-casting profile and the shadow edge. The shape of the 

 shadow edge depends both on the shadow profile and on the topography in 

 the vicinity of the edge and in consequence interpretation of shadowing is 



DIRECTION OF ELECTRON 

 SOURCE WHEN FILM IS LATER 

 MOUNTED IN MICROSCOPE 



Fig. 1 — Diagrammatic representation of thin film replicas produced when every atom 

 sticks where it strikes. 



difficult unless one of the surfaces is smooth.'" With multiple or extended 

 sources, partial shadows and shading due to partial shadowing occur. 



1.12 Negative shadows 



There may be other regions of the surface which are exposed to the atom 

 source, but not to the subsequently incident electron beam. When the 

 replica film is mounted in the microscope, these regions appear reentrant 

 to the electron beam, which must pass through three rather than one layer 

 of replica. These regions appear as very hghtly exposed areas in micro- 

 graphs, and since they are essentially ''negative shadows" are subject to 

 the same considerations of shape and interpretation as ordinary shadows. 



»» "Physical xMethods in Chemical Analysis," Vol. 1, 1950. On pp. 571-3, R. D. Heiden- 

 reich reports some unpublished work of S. G. Ellis and W. G. Gross, showing great differ- 

 ences in appearance of shadow-cast replicas as the azimuth of the atom source is varied. 



