Dependence of Compton Line Breadth on Primary Wave-Length with the Multi-Crystal Spectrograph

Abstract

The present report is a continuation of the work reported previously. The main purpose of the research is 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. In the previous paper the natural breadths of the Compton line were observed for different angles of scattering and the breadth was found to obey the functional dependence on scattering angle predicted by DuMond on the assumption that electron velocities cause the breadth. According to DuMond's theory the breadth should also be nearly proportional to the primary wave-length. This point is tested in the present paper. Exposures of about one thousand hours each were made with characteristic $K$ radiation from molybdenum, silver and tungsten target tubes. In each case the radiation was scattered from a graphite scatterer at a very well defined large scattering angle of 156\ifmmode^\circ\else\textdegree\fi{}, the multicrystal spectrograph being used to analyze the scattered radiation. Reproductions of the photographic spectrograms are shown and also microphotometer curves of the spectra. The breadth of the Compton line is found to diminish with shorter primary wave-lengths in complete accord with the predictions of DuMond's theory. Unless some other cause can be found to explain the observed behavior of the breadth the results of this paper and the above mentioned previous paper constitute a complete vindication of this theory and of its basic assumption that the initial velocities of the electrons in the scatterer cause the breadth of the Compton shifted line. If this is correct, then these experiments constitute direct experimental evidence of the dynamic nature of atoms. The Compton shifted line can indeed be thought of as broadened by the Doppler effect of the motion of the scattering electrons. A simplified form of the theory of modified scattering by initially moving electrons is presented to supplement and clarify the more elaborate and exact theory of the previous paper. We call attention to the fact that the theory and experimental results are in no way discordant with wave mechanics or the uncertainty principle but we believe that to translate the present exposition into the language of wave mechanics would tend to obscure rather than clarify the picture. The results of this work are so clearly defined that the reality of the much narrower Compton lines obtained recently by several investigators seems, in our opinion, to be very doubtful.