74 The Electron Microscope 



amplified voltage is impressed on the screen grids of the output 

 stage of the oscillator which supplies the rectifier with 32 kilo- 

 cycles power. The sign of this voltage is so chosen that it will 

 reduce the high tension if it is too high, and vice versa. (Nega- 

 tive feedback.) The smoothing efifect of this circuit is so strong 

 that what fluctuations remain must be ascribed to other causes. 

 As A. W. Vance has shown, they are mostly due to variations in 

 the potentiometer, caused in some way, not yet properly under- 

 stood, by the circulation of the oil in which it is immersed. This 

 unit maintains the output voltage constant within 0.004 per cent 

 or better over 30 seconds, which is shown by experience to be 

 the longest exposure required. 



As mentioned in chapter 5, this microscope has a physical 

 objective diaphragm of about 0.001 in. diameter. This extremely 

 fine hole is drilled into a gold leaf with a finely pointed needle. 

 Under optimum conditions, it improves the resolution limit of 

 the R.C.A. microscope to about 24 A, which is half of what is 

 obtainable without the diaphragm. As the performance of the 

 microscope with a diaphragm tends to fall.ofif, and at the end of 

 two months of operation is not better than 100 A, it is not sur- 

 prising that many users of the R.C.A. microscope prefer to 

 discard the aperture and are satisfied with a steady resolution of 

 about 50 A. 



The preparation of specimens for investigation is simpler than 

 in ordinary microscopy, as no dyeing is necessary. Figure 22 

 shows four stages of the ordinary routine. First a drop of 

 collodion (nitrocellulose) solution is placed on a water surface. 

 This spreads out rapidly, and after the solvent has evaporated 

 a solid film of about 100 A thickness will float on the water.* 

 Next, a small circular disk of very fine wire mesh is placed on 

 top of the collodion film, and pressed down somewhat, so that it 

 adheres to the film. The wire mesh is fished out (last figure) and 

 its edges are freed from the collodion. The mesh which is now 

 covered with a film is dried, and a drop of the suspension or 



* Experiments of the author indicate that the artificial resin, "Petrex 

 5," manufactured by the Hercules Powder Company, Inc., lends itself 

 even better for the production of extremely thin films than nitrocellulose. 



