Experimental Evidence for Electron Velocities as the Cause of Compton Line Breadth with the Multicrystal Spectrograph

The main purpose of this research was to test the correctness of the assumption that the initial velocities of electrons in the scattering body cause the observed breadth of the Compton shifted line. The test consists in observing the natural breadths of the Compton line for different scattering angles and primary wave-lengths and comparing these with the functional dependence of breadth on scattering angle and primary wave-length deduced theoretically on the basis of the assumption under test. It is shown in this paper that if electron velocities are the cause of the breadth then this breadth should increase with the scattering angle according to the approximate formula =Kcos12 where is the breadth, the scattering angle and K a constant depending on the primary wave-length and the scattering substance. For the same scattering angle and substance the breadth should be proportional to the primary wave-length. The experimental test was made with the multicrystal spectrograph of fifty units herein briefly described. Three scattering angles were used, 63^, 90^, 156^, the inhomogeneity in each case being less than one degree. The spurious breadth due to this inhomogeneity is negligible compared to the observed breadths. Three very clear cut spectrograms are reproduced representing MoK radiation scattered from graphite together with microphotometer curves taken from them. The increase of shifted line breadth with scattering angle is clearly visible and compares favorably with the theoretical prediction. The increase of line breadth with increasing primary wave-length comparing the breadths of shifted ₁ and shifted ₁ lines seems just detectable. The unshifted lines are very sharp, the doublet being clearly resolved. Incidentally the cause of the heavy background so frequently observed on Compton effect spectrograms has been found to be non-selective scattering at the crystals and a great reduction of background and improvement in contrast has been effected by the use of baffles to diminish this effect. The observed shift of the Compton line supports Compton's formula = (hmc) (1-cos) where hmc=24. 2X. U. The shifted line breadths are greater than those reported by Gingrich and Bearden and possibly Ross but seem to be in general accord with previous breadths obtained at this laboratory and with those of H. M. Sharp and of F. L. Nutting. The reason for these discrepancies is unknown. The possibility of double or higher multiplicities of scattering is being investigated. (See note added in proof at the end of this article. )