Artificial neural network in Mossbauer mineralogy

• Main Authors: de Souza, P.A., Salles, E.O.T., Garg, V.K.
• Format: Conference Proceeding
• Language: English
• Subjects: Artificial neural networks; Computer networks; Crystallization; Iron; Magnetic fields; Minerals; Paramagnetic materials; Software; Spectroscopy; Vector quantization
• DOI: 10.1109/MWSCAS.1995.504500

The first version of the Mossbauer Effect Assistant program composed of an Artificial Neural Network (ANN) that is associated with Mossbauer parameters and references data bank of the iron containing minerals reported in the literature is presented. Mossbauer spectroscopy is a useful technique for characterizing the valencies, electronic and magnetic states, coordination symmetries and site occupancies of the cation. The Mossbauer parameters of isomer shift and quadruple splitting are useful to distinguish paramagnetic ferrous and ferric iron in several substances, while the internal magnetic field provides information on the crystallinity. A computer software named as Mossbauer Effect Assistant has been developed using learning vector quantization (LVQ) neural network linked to a Mossbauer data bank that contains Mossbauer parameters of isomer shift, quadrupole splitting, internal magnetic field and the references of the Mossbauer published data of the substances. The program identifies the substance under study and/or its crystalline structure when fed with experimental Mossbauer parameters. It can also list the references of the literature that are stored in the data bank by feeding the name of the substance or the author of the publication. Application of Mossbauer Effect Assistant for iron mineral Mossbauer spectroscopy is in user friendly Microsoft Windows environment.