Possibilities of Future Development 101 



of present-day magnetic objectives are already practically sat- 

 urated, and the magnetic intensities cannot be further increased 

 without the field spreading out over a larger region, thus frus- 

 trating the reduction of the pole piece dimensions. In electro- 

 static lenses, reduction of the dimensions by one order of 

 magnitude is still possible, if the technique of precision watch- 

 makers is used, but the problem is how to prevent breakdown 

 of the voltage by auto-electric discharge. It has been suggested 

 in German and American patent applications to avoid this by 

 operating the microscope with surges of very short duration. 

 (A.E.G., 17th Dec, 1938; S. Ramo and G.E. Co., 1st April, 

 1941, and 11th Sept., 1942.) But as / figures in equation (24b) 



only with the — power even a reduction by a factor of y^r would 



reduce the resolution limit only in the ratio T^yo , and this would 



not be sufficient to bring the best electrostatic microscopes in 

 line with the best magnetic ones. Surge operation is likely to 

 produce potential differences by oscillations between the elec- 

 trodes, therefore, the gain, if any, is likely to be even less. 



E. G. Ramberg ^" was the first to point out that more could 

 be gained by increasing the strength of existing lenses than by 

 scaling them down, as this would reduce C as well as f, though 

 he left it open how to achieve this in face of the limitations dis- 

 cussed above. But from the material computed by Ramberg it 

 is easy to conclude that appreciable progress could be achieved 

 not by scaling dow^n present-day lenses, but, rather surprisingly, 

 by scaling them up. Ramberg calculates a magnetic lens formed 

 by two cylinders of infinite permeability, of diameter D, sepa- 

 rated by a narrow gap, and gives / and C as functions of a 

 parameter 



R 



_ (DH)2_ 



V 



(lens diameter X maximum magnetic field on axis) 2 cm^gauss^ 



accelerating potential volt 



Some values calculated from Ramberg's data are: 



