Score-based diffusion models for accurate crystal-structure inpainting and reconstruction of hydrogen positions

Abstract

Generative AI models, such as score-based diffusion models, have recently advanced the field of computational materials science by enabling the generation of new materials with desired properties. In addition, these models could also be leveraged to reconstruct crystal structures for which partial information is available. One relevant example is the reliable determination of atomic positions occupied by hydrogen atoms in hydrogen-containing crystalline materials. While crucial to the analysis and prediction of many materials properties, the identification of hydrogen positions can however be difficult and expensive, as it is challenging in X-ray scattering experiments and often requires dedicated neutron scattering measurements. As a consequence, inorganic crystallographic databases frequently report lattice structures where hydrogen atoms have been either omitted or inserted with heuristics or by chemical intuition. Here, we combine diffusion models from the field of materials science with techniques originally developed in computer vision for image inpainting. We present how this knowledge transfer across domains enables a much faster and more accurate completion of host structures, compared to unconditioned diffusion models or previous approaches solely based on DFT. Overall, our approach exceeds a success rate of 97% in terms of finding a structural match or predicting a more stable configuration than the initial reference, when starting both from structures that were already relaxed with DFT, or directly from the experimentally determined host structures.

References in corpus (19)

• scikit-image: Image processing in Python
• The PseudoDojo: Training and grading a 85 element optimized norm-conserving pseudopotential table
• E(3)-Equivariant Graph Neural Networks for Data-Efficient and Accurate Interatomic Potentials
• Precision and efficiency in solid-state pseudopotential calculations
• AiiDA: Automated Interactive Infrastructure and Database for Computational Science
• Accelerating crystal structure prediction by machine-learning interatomic potentials with active learning
• Materials Cloud, a platform for open computational science
• AiiDA 1.0, a scalable computational infrastructure for automated reproducible workflows and data provenance
• On how good DFT exchange-correlation functionals are for H bonds in small water clusters: Benchmarks approaching the complete basis set limit
• Workflows in AiiDA: Engineering a high-throughput, event-based engine for robust and modular computational workflows
• Playing quantum hide-and-seek with the muon: localizing muon stopping sites
• Quantum states of muons in fluorides
• Modeling lithium-ion solid-state electrolytes with a pinball model
• Identifying Crystal Structures Beyond Known Prototypes from X-ray Powder Diffraction Spectra
• Crystal structure prediction with host-guided inpainting generation and foundation potentials
• Generative Design of Crystal Structures by Point Cloud Representations and Diffusion Model
• Automated computational workflows for muon spin spectroscopy
• MC3D: The Materials Cloud computational database of experimentally known stoichiometric inorganics
• Accurate and efficient protocols for high-throughput first-principles materials simulations