New Approach To Understanding the Experimental ¹³³Cs NMR Chemical Shift of Clay Minerals via Machine Learning and DFT-GIPAW Calculations

Abstract

Structural determination of adsorbed atoms on layered structures such as clay minerals is a complex subject. Radioactive cesium (Cs) is an important element for environmental conservation, so it is vital to understand its adsorption structure on clay. The nuclear magnetic resonance (NMR) parameters of 133 Cs, which can be determined from solid-state NMR experiments, are sensitive to the local neighboring structures of adsorbed Cs. However, determining the Cs positions from NMR data alone is difficult. This paper describes an approach for identifying the expected atomic positions on clay minerals by combining machine learning (ML) with experimentally observed chemical shifts. A linear ridge regression model for ML is constructed from the smooth overlap of atomic position descriptor and gauge-including projector augmented wave (GIPAW) ab initio data. The constructed ML model predicts the GIPAW data to within a 3 ppm root-mean-squared error. At this stage, the 133 Cs chemical shifts can be instantaneously calculated from the Cs positions on any clay layers using ML. The inverse analysis, which derives the atomic positions from experimentally observed chemical shifts, is developed from the ML model. The input data for the inverse analysis are the layer structure and the experimentally observed chemical shifts. The Cs positions for the targeted chemical shifts are then output. Inverse analysis is applied to montmorillonite, and the resultant Cs positions are found to be consistent with previous results (Ohkubo, T.; et al. J. Phys. Chem. A 2018, 122, 9326−9337). The Cs positions on saponite clay are also clarified from experimentally observed chemical shifts and inverse analysis.

Theoretical Insights into the 27Al NMR Parameters of Binary Aluminosilicate Glass and Their Relationship to the Atomic and Electronic Structure

“…The disordered polyhedral structure of Al is indicated by a variety of spherical harmonic parameters with nonzero values corresponding to their deviations from regular polyhedra. NMR parameters can be well modeled using machine learning (ML) techniques with q l values within an appropriate l range for local structures. − While the description of ML with q l can be useful for predicting NMR parameters from a local structure, “black-boxed” ML models do not allow for discussions of symmetry in Al-coordinated structures. Therefore, the correlation between spherical harmonic parameters ( q lm and q l ) and NMR parameters was investigated without a “black-boxed” ML description of the structure.…”

Crystal Engineering for Achievement of Functional Materials Through Facile Intercalation of Lamella Mineral Kaolinite

“…Actually, theoretical simulations such as DFT calculations and/or molecular dynamics (MD) as well as the high-throughput calculations and AI-assisted methodology (such as machine learning，deep-learning potential, etc.) can be utilized as a suitable and powerful approach [87,88].…”