Machine learning for analysis of experimental scattering and spectroscopy data in materials chemistry

Abstract

We discuss how machine learning methods can be applied to advance analysis of spectroscopy and scattering data in materials chemistry. We give examples illustrating the state-of-the-art and identify current challenges in the field.

Gaining More Insights from Synchrotron-Based X-ray Spectroscopy for Alkali Ion Rechargeable Batteries

“…The core idea of the ML approach lies in the assumption that there must be a hidden linkage between the electronic/structural/compositional descriptors of a given material and its XAS spectra, and ML is effective in solving these intricate nonlinear relationships, either in a supervised or unsupervised way. 118 Diverse machine learning algorithms have been successfully used to identify the structural model from the X-ray spectra in the inverse process, which is used to predict the spectra from an input structure or to directly extract key electronic or local structural parameters from the measured spectra. Based on over 40,000 materials from the open-science Materials Project database, Zhang et al calculated more than 800,000 K-edge XANES spectra as references, resulting in a large XANES database, XASdb.…”

“…In such cases, manually analyzing each spectrum is not feasible, especially considering the computational expense of the least-squares algorithm. Benefiting from advances in data science, the machine learning (ML) approach has emerged as an unprecedented tool for automatically handling large data sets. − For example, the combined use of attribute extraction and clustering algorithms allows the identification of unanticipated minority phases from over 10 million Co K-edge XANES spectra covering more than 100 LiCoO 2 particles collected in the STXM experiment (Figure b). Beyond the battery community, the field of condensed matter and catalysis has witnessed the significant power of ML in X-ray spectroscopy. ,,, For instance, the direct conversion of XANES data into the radial distribution function (RDF) was accomplished using an artificial neural network (ANN) ML model (Figure c).…”

“…More recently, the incorporation of explainability into the graph neural network (GNN) architecture not only facilitates accurate predictions of XANES spectra but also enhances our current comprehension of XANES (Figure d). The core idea of the ML approach lies in the assumption that there must be a hidden linkage between the electronic/structural/compositional descriptors of a given material and its XAS spectra, and ML is effective in solving these intricate nonlinear relationships, either in a supervised or unsupervised way . Diverse machine learning algorithms have been successfully used to identify the structural model from the X-ray spectra in the inverse process, which is used to predict the spectra from an input structure or to directly extract key electronic or local structural parameters from the measured spectra.…”

Enhancing antioxidant properties of CeO2 nanoparticles with Nd3+ doping: structural, biological, and machine learning insights

“…Currently, Nd 3+ is used to dope TiO 2 as a nanothermometer, Gd 2 O 3 as a magnetic resonance imaging agent, and ZnO as an antibacterial agent against multidrug-pathogenic bacteria. 22–24 However, its effect on CeO 2 -NPs has not been unveiled.…”

Computational Reverse Engineering Analysis of the Scattering Experiment Method for Interpretation of 2D Small-Angle Scattering Profiles (CREASE-2D)

“…Additionally, the surge in high-throughput measurements and the quest for artificial intelligence (AI)-driven manufacturing demand analysis methods that can be fast and automated in interpreting scattering profiles and complementary characterization results, as and when the measurement is done. We direct readers to a recent perspective by Anker et al that covers many ongoing developments and studies within this topic of fast computational analysis of scattering and spectroscopic measurements in materials sciences . The challenges for computational methods being developed for fast or automated scattering analyses in the area of synthetic soft materials are different from inorganic hard materials or biological molecules.…”