This grant's purpose is to develop the methodology, new algorithms, and computer software for determining local and nanoscale atomic and magnetic order in technological materials using different types…

The National Institute of Standards and Technology (NIST) awarded Theiss Research a Cooperative Agreement totaling $144,342.63 under the Measurement and Engineering Research and Standards (CFDA 11.609) program, effective April 1, 2025, through March 31, 2027. The award funds the development of advanced methodology, algorithms, and software tools for determining local and nanoscale atomic and magnetic order in technological materials using X-ray and neutron scattering data. Key deliverables include implementing new atomistic structural refinement algorithms in the RMCProfile software, developing enhanced Fourier filtering capabilities for three-dimensional diffuse-scattering intensity analysis, creating a machine-learning approach for structural model recovery from experimental total-scattering data, and implementing improved algorithms for neutron magnetic scattering analysis applicable to high-entropy oxides.

The software enhancements and computational tools developed under this agreement will be integrated into RMCProfile, a publicly available platform already used globally by the materials research community. These improvements will expand researchers' capacity to analyze data from modern synchrotron and pulsed neutron facilities, enabling solution of a broader range of structural problems in partially disordered materials with higher fidelity than currently possible. The work is anticipated to benefit experimental crystallographers, molecular dynamics theorists, and advance artificial intelligence applications in structure determination methodologies.

Activities

A new algorithm for atomistic structural refinements using total single-crystal x-RAY and neutron scattering data will be developed and implemented in the rmcprofile software (https://rmcprofile.ornl.GOV/). New capabilities will be added to the software tool for performing fourier filtering of 3D diffuse-scattering intensity to enable adaptive masking of arbitrary-shaped diffuse scattering. A new machine-learning approach to recovering structural models from experimental total-scattering data will be developed and tested for feasibility. A new improved algorithm for fitting neutron magnetic scattering will be implemented in the rmcprofile software and applied to determine magnetic ordering in high-entropy oxides.

Expected outcomes

The proposed developments will lead to major improvements in the existing capabilities for determining atomic order in partially disordered materials, with vastly expanded scope and higher fidelity of the recovered structural models. The rmcprofile software is publicly available and already used worldwide by the materials research community. The new capabilities will permit these researchers to take advantage of the still underutilized data obtainable with modern synchrotron and pulsed neutron facilities and, thus, solve a significantly broader range of structural problems than currently possible. The improved software for the inverse fourier transform of calculated diffuse intensities will be of significant benefit to experimental crystallographers and theorists performing molecular dynamics simulations. The new machine learning approach, if successful, will provide a blueprint for using the most recent developments in AI algorithms for structure determination. Altogether, the proposed developments will significantly advance the existing capabilities in structural characterization available to researchers developing new and improved materials.

Intended beneficiaries

Materials researchers will benefit from the new capabilities which will facilitate the understanding of structure and its relation to function in complex technological systems.

Subrecipient activities

The recipient does not intend to subaward funds.