Implementation of an artificial neural network in the identification of the Mössbauer spectral shape of hematite and magnetite

Artificial intelligence (AI) is widely used for analyzing large data systems. Artificial neural networks (ANN) are a specialized AI method used for fitting and classifying patterns in data sets. This work presents the use of an artificial neural network as a recognition system in Mössbauer spectroscopy. Using a Mössbauer spectra database (Dyar, DEVAS Database, 2022) and generated spectra, the ANN was trained to identify whether the shape of a Mössbauer spectrum corresponds to hematite (α-Fe2O3), magnetite (Fe3O4), or neither. A precision percentage of 99% on the training data set and 100% on the validation set were achieved. An accuracy of 97% on the test set was achieved. These results confirm the efficiency of the ANN for identifying these compounds. The proposed method can be easily extended to identify other iron-containing compounds and for use with other characterization techniques. © 2023, The Author(s), under exclusive licence to Springer Nature Switzerland AG.