14 



B. V. BORRIES ti G. LANGNER AND W. SCHEFFELS 



D 



Fig. 4. Impurities in beryllium foil, x-ray collection. A, all 

 radiation recorded; B. C, D, characteristic emission from dif- 

 ferent impurities separately selected — manganese, nickel and 

 calcium respectively. 



or AgL characteristic radiation to form the picture. 

 The distribution of the copper can thus be shown 

 in one picture and the distribution of the silver in 

 the other. Another way of displaying this same 

 result is to regard the two pictures as components of 

 a colour picture and photograph them through 

 different filters in register on the same piece of colour 

 film. In this way the copper was made to appear red 

 and the silver green. In principle one can attach a 



given colour to a given x-ray wavelength, which may 

 be a useful technique for presenting in one picture 

 the relative positions of several elements. 



As an example of the identification of inclusions 

 in a surface, fig. 4 shows some impurities in a piece 

 of beryllium foil (3). In A all the emitted x-rays 

 were used to form the picture and the impurities 

 show up as emitting more strongly than the beryl- 

 lium. In 5, C, and D the pulse analyser was set to 

 pass the characteristic radiation from three different 

 types of impurity in turn, which were identified as 

 manganese, nickel and calcium respectively. The in- 

 formation may again be presented on one picture in 

 colour. 



The energy resolution of the proportional counter 

 is not good enough for the separation of character- 

 istic radiation from elements which differ by less 

 than about four in atomic number. However, it 

 seems that there would be sufficient intensity reflected 

 from a curved crystal in the spectrometer to form 

 the picture, in which case adjacent elements would 

 easily be separable. 



The instrument, therefore, can show up either the 

 topography of a surface by electron scattering with 

 a resolution of well below I //, or the distribution 

 of an element over a surface by x-ray emission with 

 a resolution of about I //. A selected point in the 

 surface can then be analysed quantitatively by plot- 

 ting the characteristic x-ray emission spectrum by 

 means of a crystal spectrometer. 



References 



1. Castaing, R., Thesis. Paris, 1951. 



2. Castaing, R. and Descamps, J., J. pliys. 16, 304 (1955). 



3. COSSLETT, V. E. and Duncumb, P., Nature 177, 1172 



(1956). 



4. CossLETT, V. E. and Nixon, W. C, /. Appl. Pins. 24, 616 



(1953). 



5. McMuLLAN, D., Proc. Inst. Elec. Engrs., Pt. 1, 100, 245 



(1953). 



6. Nixon, W. C, Proc. Roy. Soc. A 232, 475 (1955). 



7. Smith, K. C. A. and Oatley, C. W., Brit. J. Appl. Phys. 6, 



391 (1955). 



Imaging Elements Operating with Permanent Magnets 

 B. V. BoRRiES t, G. Langner and W. Scheffels 



Rheinisch-Westfdlisches Iiistitut fiir Vbermikroskopie, Diisseldorf 



Electron lenses operating with permanent magnets 

 can be employed in electron microscopes, as several 

 authors (I, 8) have shown. The advantage of such 

 lenses is that no source for a highly stabilized lens 

 current is needed. As the main disadvantage during 

 operation must be named the impossibility to switch 

 off the magnetic field. Moreover there must be at 

 least two gaps in a system excited by permanent 



magnets to avoid stray field. This does not neces- 

 sarily mean, however, that these gaps must act as 

 separate lenses. Permanent magnetic einzel-lenses 

 with variable focal length have been used as con- 

 densor lenses by v. Borries (1954), the magnetic 

 circuit of which has been calculated by v. Borries 

 and Lenz (4) and the properties of which have been 

 investigated experimentally by Langner (6). Also as 



