.[ 647 ] 



LXIV. Precision-measurements in the X-Ray Spectra. 

 Part III. By Dr. Manne Siegbahn and Mr. A. B. Leide *. 



An X-ray Spectrograph for medium ivave-lenqths. 



IN the two former parts of this paper X-ray spectrographs 

 have been described for long waves in vacuo and for 

 very short waves, both allowing a sufficient degree of accuracy 

 in the measurement of wave-lengths. It is not impossible tc 

 let these two apparatus complete one another, and with them 

 to photograph the whole X-ray spectrum-range as yet known. 

 But in this way there will be a region in which the measure- 

 ments are not as reliable as may be wished. For the case 

 of the vacuum-spectrograph this arises from the fact that 

 the X-ray reflexion at a crystal-face is not a surface effect. 

 On the contrary, when shorter waves come into play the 

 pencil may reach a considerable depth in the crystal before 

 being wholly reflected. This, as shown by Blake and 

 Duanef, is very remarkable, when an ionization method is 

 used for registering the reflected ray. When a photographic 

 spectrogram is taken, the breadth of the spectral lines in 

 comparison with the slit-breadth gives a representation of 

 the thickness of the reflecting layer. But there are in spite 

 of this some reasons which make it desirable not to use a 

 spectrograph of the type described in the first part for o the 

 study of! somewhat shorter waves, say between 0*5— 1 A U. 

 An essential condition for the spectrograph giving sharp 

 lines is that the reflecting layer shall coincide with the 

 rotation axis of the crystal-table, a condition which naturally 

 can only be fulfilled when the thickness of the reflecting 

 layer is small. Furthermore, at the small reflexion-angles 

 corresponding to short waves a little fault in the adjustment 

 may have a great influence. 



To illustrate these effects it may be sufficient to take a 

 spectrogram of e. g. the tungsten K-series with a rotating 

 crystal spectrograph. With rock-salt as analysing crystal it 

 is hardly possible in the first order to separate the a-doublet 

 even with the smallest slit. 



A spectrograph intended for more penetrating waves may 

 therefore be constructed on the principle already used i'ov 

 the spectrograph described in the second part, thai is with 

 the slit behind the crystal. A spectrograph where 1 this is 



* Communicated by the Authors. 

 t Blake and Duane, Phys. Rev. 1917. 



2 Y - 



