Bayesian Analysis of Ion Beam Diagnostics

Ion beam diagnostics are routinely used for quantitative analysis of the surface composition of mixture materials up to a depth of a few m. Unfortunately , advantageous properties of the diagnostics, like high depth resolution in combination with a large penetration depth, no destruction of the surface, high sensitivity for large as well as for small atomic numbers, and high sensitivity are mutually exclusive. Among other things, this is due to the ill-conditioned inverse problem of reconstructing depth distributions of the composition elements. Robust results for depth distributions are obtained with adaptive methods in the framework of Bayesian probability theory. The method of adaptive kernels allows for distributions which contain only the signiicant information of the data while noise tting is avoided. This is achieved by adaptively reducing the degrees of freedom supporting the distribution. As applications for ion beam diagnostics Rutherford backscattering spectroscopy and particle induced X-ray emission are shown.