Depth Profile Reconstruction from Rutherford Backscattering Data

Abstract

An adaptive kernel method in the Bayesian framework together with a new simulation program for Rutherford backscattering spectroscopy (RBS) have been applied to the analysis of RBS data. Even in the case of strongly overlapping RBS peaks a depth profile reconstruction without noise fitting has been achieved. The adaptive kernel method leads to the simplest depth profile consistent with the data. Erosion and redeposition rates of carbon divertor plates in the fusion experiment ASDEX Upgrade could be determined by RBS-analysis of thin film probes before and after exposition to plasma discharges.