Proffen, Thomas [ORNL] (ORCID:0000000214086031)

DOE OSTI2026

CoSn-type NiIn1–xSbx (0 ≤ x ≤ 0.17): Site-Selective Substitution, Electronic Structure, Chemical Bonding, and Structural Transformation

CoSn-type intermetallic compounds have emerged as a model platform for Kagome-derived flat-band physics, where subtle chemical perturbations can strongly influence electronic structure and phase stability. Here, we present a combined experimental and theoretical study of Sb-substitution in CoSn-type NiIn1–xSbx (0 ≤ x ≤ 0.17) to elucidate the interplay between site selectivity, solubility limit, chemical bonding, and electronic structure. Rietveld refinements on Neutron powder diffraction data confirmed the selective Sb-substitution at the electron-rich In2 (2d) site forming the honeycomb substructure, while the In1 (1a) site within the Kagome layer remains exclusively occupied by In. Density functional theory (DFT) calculations revealed that pristine CoSn-type NiIn hosts Ni 3d-dominated flat bands near the Fermi level (EF), originating from the Kagome-like Ni substructure. Partial replacement of In by Sb within the honeycomb layer alters these flat-band features below EF, reducing the density of states and suppressing the flat-band topology near the Fermi level. Orbital-resolved electronic structure and chemical-bonding analyses show that Sb-substitution enhances Ni-p-block (In/Sb) covalency and optimizes charge compensation, stabilizing the CoSn-type structure up to the solubility limit x ≈ 0.17. Beyond the limit, the higher-Sb compositions show satellite reflections consistent with an incommensurately modulated phase. These results establish a link between site-selective chemical substitution, bonding optimization, and flat-band electronic structure evolution, providing fundamental insights into how chemical substitution influences the electronic properties of Kagome-based intermetallic compounds.

Roy, Nilanjan [National Institute of Technology Si↗

DOE OSTI2024

An Active Learning-Based Streaming Pipeline for Reduced Data Training of Structure Finding Models in Neutron Diffractometry

Structure determination workloads in neutron diffractometry are computationally expensive and routinely require several hours to many days to determine the structure of a material from its neutron diffraction patterns. The potential for machine learning models trained on simulated neutron scattering patterns to significantly speed up these tasks have been reported recently. However, the amount of simulated data needed to train these models grows exponentially with the number of structural parameters to be predicted and poses a significant computational challenge. To overcome this challenge, we introduce a novel batch-mode active learning (AL) policy that uses uncertainty sampling to simulate training data drawn from a probability distribution that prefers labelled examples about which the model is least certain. We confirm its efficacy in training the same models with ∼ 75% less training data while improving the accuracy. We then discuss the design of an efficient stream-based training workflow that uses this AL policy and present a performance study on two heterogeneous platforms to demonstrate that, compared with a conventional training workflow, the streaming workflow delivers ∼ 20% shorter training time without any loss of accuracy.

Wang, Tianle [Brookhaven National Laboratory (BNL)↗