Artificial neural network analysis of RBS data of Er-implanted sapphire

Rutherford backscattering spectrometry (RBS) is a well-established techniqu e for the elemental depth profile of the surface layers of samples, includi ng the determination of the dose and depth of implanted elements. We have d eveloped a code based on artificial neural networks (ANN) to analyse RES da ta. The ANN was trained using the traditional backpropagation algorithm, wh ich is designed to minimise the average error on a training set of generate d data. The algorithm was applied to one important particular case: namely the determination of the amount of Er implanted in sapphire samples, and th e depth at which the Er is located. The Er fluence was between 8 x 10(13) E r+/cm(2) and 2 x 10(16) Er+/cm2, for implant energies of 200 and 800 keV. T he analysis is instantaneous, automated, and requires absolutely no knowled ge from the user aside the experimental conditions. The results obtained ar e hence well-suited for on-line data analysis. (C) 2001 Elsevier Science B. V. All rights reserved.