THE DETERMINATION OF DEPTH PROFILES FROM ANGLE-DEPENDENT XPS USING MAXIMUM-ENTROPY DATA-ANALYSIS

The principles of maximum entropy have been used to determine elemental depth profiles from simulated angle-dependent XPS data. The theory of maximum entropy data analysis as applied to angle-dependent XPS is presented, and brief details are given of the novel encoding needed for this application. The method overcomes the ambiguity of solutions that is encountered when reconstructions are attempted from real data with noise. This is done by maximising the overall logarithmic probability function obtained by optimising the balance between maximum entropy and minimum chi-squared. The overall scaling of the noise in the data is determined, confidence limits are assigned to reconstructions from individual data sets, and the relative probabilities of different prior models are found.