A Study of the Short Wave-Length Limit of the Continuous X-Ray Spectrum

Author: Bailey, Howland Haskell

Year: 1941

Degree: Dissertation (Ph.D.)

Advisor: DuMond, Jesse William Monroe

Committee Member: Unknown, Unknown

Option: Physics

DOI: 10.7907/dw87-7z48

Abstract

Because of the troublesome discrepancy among the natural atomic constants e, h and mo which has existed since 1935, a precision measurement of h/e by the short wave-length limit of the continuous x-ray spectrum is undertaken. The method used is the customary one of observing an isochromat, but several changes over previous practice are made in the apparatus and the experimental technique, and these are described in detail. Chief among these are: (a) an improved filament supply circuit and an electronic robot for stabilizing the voltage, (b) an elaborately calibrated precision resistance for measuring the voltage, (c) improved sensitivity in the ion chamber measurements of intensity, (d) the use of freshly cleaved crystals in the two-crystal monochromator to improve its resolving power and thereby reduce the fillet at the foot of the isochromat (made possible with the high intensity available from the Watters Memorial power supply and x-ray tube), (e) the use of Ross balanced filters to suppress the effect of non-coherently scattered radiation and to eliminate the integrated effect of the wings of the spectral transmission "window" of the monochromator, (f) an indirect cathode to preclude the existence of excess energy secondary electrons, and (g) a very clean target of high atomic number to insure a good intercept on the isochromat. The calibration of the voltmeter and of the spectrometer are not yet in a satisfactory state, but preliminary results give h/e= (1.38054 ± .00115) x 10-17 erg sec/esu. This is to be compared with the indirect value (obtained from combining the Rydberg constant, the x-ray value of e and the best value of e/mo ) of (1.37929 ± .00023) x 10-17 erg sec/esu. The overlapping of these probable errors demonstrates the resolution of the discrepancy. The experiment is quite capable of five times the stated precision after certain checks are made, and better results are expected soon.

Files