Author:

Keywords:

artificial neural networks, real-time rbs, ni silicides, growth kinetics, iba software, rbs analysis, solid-state reaction, thin-films, silicides, algorithm, kinetics, rbs, Science & Technology, Technology, Physical Sciences, Instruments & Instrumentation, Nuclear Science & Technology, Physics, Atomic, Molecular & Chemical, Physics, Nuclear, Physics, Artificial neural networks, Real-time RBS, Ni silicides, Growth kinetics, IBA software, RBS analysis, SOLID-STATE REACTION, THIN-FILMS, SILICIDES, ALGORITHM, KINETICS, ENERGY, RBS, 0202 Atomic, Molecular, Nuclear, Particle and Plasma Physics, 0402 Geochemistry, 0915 Interdisciplinary Engineering, Applied Physics, 5104 Condensed matter physics, 5106 Nuclear and plasma physics

Abstract:

This paper reports on the advantage of using artificial neural networks (ANNs) to analyze large sets of real-time Rutherford backscattering spectrometry (RBS) data. Real-time RBS, i.e. collecting RBS spectra at periodic time intervals during a thermal treatment, probes the full response of a thin film to the annealing in situ. Although very valuable insights can be gained by this technique, the time-consuming analysis of the vast amount of RBS spectra acquired during real-time RBS measurements has so far prevented the widespread use of real-time RBS. Setting up an ANN is quite an intensive process as well, but once trained, these ANNs can handle the analysis of large data sets practically instantaneously. As such, the beneficial combination of real-time RBS and ANN analysis forms a perfect synergy. In this test case, a network was trained and applied to analyze the Ni silicide growth during annealing of a thin 80 nm Ni film on Si(100). The ANN performance was validated by comparing the ANN results with the conventional analysis performed on the same data set. (C) 2010 Elsevier B.V. All rights reserved.