NSLS-II Friday Lunchtime Seminar

Mat3ra.com is a data centric materials R&D platform that integrates physics-based approaches, such as density functional theory and molecular dynamics, and machine learning techniques with experimental data streams through a common set of FAIR, JSON based data standards and repeatable workflows. This presentation will first outline the architecture of Mat3ra.com and show use cases where automated high throughput simulations and interoperable data formats accelerate property prediction for heterogeneous interfaces and semiconductor relevant materials, building on prior work on data centric online ecosystems and model ontologies for computational materials science . I will then describe a collaborative theory–experiment case study with Prof. Sergei Kalinin's group at the University of Tennessee, Knoxville, where Mat3ra was used to propagate modeling-based predictions into the interpretation of STEM measurements, demonstrating the potential for a closed loop refinement between simulations and laboratory data. Finally, I will discuss our open source Mat3ra Data Convention and related tooling for model and workflow categorization, and how these standards can interoperate with Bluesky and other experimental data infrastructures at Brookhaven National Laboratory to support reproducible multiscale simulations for future semiconductor and quantum materials devices.