344 



J. H. L. MCAUSLAN AND K. C. A. SMITH 



Fig. 2 a. Needle crystal of silver azide — partly decomposed 

 by heat. 



Fig. 2 b. End of needle crystal of lead styphnate. Note crystal- 

 lographic break-up due to dehydration. 



directly in the microscope by means of a hot stage 



(fig. 1). 



Bombardment of the specimen in the conventional 

 and scanning,' microscopes. — In order to extract a 

 given amount of information from the image in the 

 electron microscope a certain minimum number of 

 electrons, as determined by quantum and fluctuation 

 considerations, must interact with the specimen. 

 This minimum number will be determined, among 

 other factors, by the quantum efficiency of the trans- 

 mission process between the specimen and the brain 

 of the observer; the quantum efficiency being defined 

 as the ratio of the number of electrons or quanta 

 associated with the point of minimum quanta trans- 

 fer to the number falling upon the specimen. The 

 bombardment of the specimen is smaller in the scan- 

 ning instrument mainly because of its superior quan- 

 tum efficiency. 



The conventional instrument operating in trans- 

 mission will have a quantum efficiency, during re- 

 cording of the image, not far short of unity since 

 virtually all of the electrons passing through an 

 element of the specimen will fall on the correspond- 

 ing element of the plate and be recorded. However, 

 when observing the image directly the quantum level 

 falls because of the low efficiency of the fluorescent 



screen and the small angle subtended at the screen 

 by the eye. It may be estimated that under these 

 conditions the quantum efficiency is about 4 10~^ 

 (4), that is, for every single visual stimulus which 

 the observer receives, about 250 electrons must pass 

 through the specimen. 



The quantum efficiency of the conventional in- 

 strument operated in reflexion is very much worse 

 because, under the usual conditions of operation, 

 only about 1 in 10^ of the electrons hitting the speci- 

 men will pass through the objective aperture. This 

 factor will, of course, vary widely according to the 

 nature of the specimen, the observation angle and 

 the objective aperture, but for the purposes of this 

 paper will be assumed constant at 10"-'. Thus the 

 quantum efficiency will be of the order of 10"* when 

 recording and lO^-lO"** when viewing directly. 



In the scanning instrument an electron multiplier 

 is used to collect the electrons passing from the 

 specimen and the input of the multiplier may be 

 arranged to collect electrons of all energies over a 

 very large solid angle. Since the total emission ratio 

 may, under certain conditions, exceed unity the 

 transfer efficiency at this stage of the process may 

 well approach unity. The high gain of the electron 

 multiplier ensures that the quantum level is 



