APPLICATION OF AN ITERATIVE MAXIMUM-LIKELIHOOD ALGORITHM IN PIXE DEPTH PROFILING OF TRACE-ELEMENTS

We have applied an iterative maximum-likelihood algorithm to extact tr ace element depth profiles from PIXE yields data for various incident proton energies. This algorithm is based on the exact stochastic model of X-ray yield measurements. We have tested the algorithm using calcu lated X-ray yields generated from a PIXE simulation program. Several p rofiles of phosphorus with depth extensions of a few micrometers in a silicon matrix are used to generate the X-ray yields. The maximum-like lihood algorithm is able to reconstruct the depth profiles from these yields successfully even when Poisson noise is added to the yield data . The reconstructed profiles are smooth and do not suffer from erratic fluctuations normally encountered in other conventional deconvolution methods.