Transition Probabilities in the L Shell of Heavy Atoms.

Measurements of L shell transition probabilities have been performed in response to a need for accurate experimental determination of fluorescence and Coster-Kronig yields. The scatter in literature data is significant, particularly for Coster-Kronig transition probabilities. Sophisticated relativistic self-consistent field models of the atom have been developed, incorporating corrections for relaxation, radiative, and electron-electron correlation effects. With an accurate set of experimental data, the accuracy of theory can be established so that a complete and reliable set of data can then be calculated. A number of techniques, which were pioneered in earlier work in this field, are exploited. Of basic concern was the removal of experimental errors which became evident as techniques evolved. Particular attention was given to analytical modelling of spectra and efficiency calibration of solid-state detectors. A K x-ray-L x-ray coincidence technique is the primary means by which measurements were performed. Using radionuclide sources, the technique was used to evaluate the Coster -Kronig transition probability f_{23 } and the fluorescence yields omega _2 and omega_3 for elements with Z = 78, 80, 81, 82, 92, and 94. In the region of Z = 80, predictions of Dirac-Hartree-Slater theory were in excellent agreement with measured values of omega_2 and omega_3 but results indicated a consistent 5%-10% overestimate of f_{23}. For Z >=q 92, the opposite situation arose whereby theoretical and experimental values for f_ {23} demonstrated reasonable agreement while theory had a tendency to overestimate omega _2 and omega_3. This coincidence approach was also adapted to provide further results of f_{23} for Z = 67 and 68 using gamma-ray fluorescence as a means of x-ray production; theory agreed quite well with these results. Methods for evaluating fluorescence and Coster-Kronig yields associated with all three L subshells were explored. One which involves accurate deconvolution of singles L x-ray spectra was tested, with promising results, using a ^{210}Pb source. A second method was an L conversion electron-L x-ray coincidence technique which requires the use of an electron spectrometer. Investigation led to the conclusion that a modified version of a particular design of electrostatic analyzer would provide a feasible means of accomplishing such measurements.