Career: machine learning-assisted first-principles modeling of ultrafast photoexcitation in 2D heterostructures

The National Science Foundation Division of Chemistry awarded Clarkson University $697,770 on July 15, 2026, under the Mathematical and Physical Sciences program (CFDA 47.049) to develop machine learning-assisted computational frameworks modeling ultrafast photoexcitation in two-dimensional heterostructures.

Principal investigator Dhara J Trivedi will combine first-principles electronic structure methods (GW-BSE and nonadiabatic molecular dynamics) with machine learning to investigate excited-state dynamics and interfacial charge-transfer processes in two-dimensional hybrid organic-inorganic semiconductors. The research aims to advance understanding of light-driven processes in atomically thin materials and support development of next-generation solar energy conversion, sensing, and optoelectronic devices.

The project integrates research and education through the Q-CHEM Cadets program, a 10-week initiative providing coding and computational chemistry training to high school students and teachers, and a new interdisciplinary course in computational skills. Trivedi's group will also develop open-source software tools and a publicly accessible materials database for the computational materials research community.

Performance occurs in Potsdam, New York. The period of performance runs from the award date through June 30, 2031.

Abstract

Dhara j trivedi of clarkson university is supported to investigate how light interacts with two-dimensional materials. This project will use recent advances in artificial intelligence (AI), particularly machine learning methods, to develop computational frameworks that model how photoexcitation generates and transports energy and charges in next-generation semiconductor materials. The research will advance the fundamental understanding of light-driven processes in atomically thin materials and will contribute to future technologies for more efficient solar energy conversion, sensing, and optoelectronic devices. By combining quantum-level simulations with data-driven materials discovery, the project will help identify new materials for sustainable energy and electronic technologies. The trivedi research program will integrate research, education, and workforce development through the 10-week ?q-chem cadets? Program, which will provide coding and computational chemistry training for high school students and teachers, as well as through a new interdisciplinary course designed to equip students with transferable computational skills that connect fundamental science with emerging technological applications. The project will also support the broader scientific community through the development of open-source software tools and a publicly accessible materials database for computational materials research. The trivedi group will develop integrated first-principles and machine learning frameworks to investigate excited-state dynamics and interfacial charge-transfer processes in two-dimensional hybrid organic-inorganic semiconductors. The project will combine many-body electronic structure methods, including GW-BSE, with nonadiabatic molecular dynamics to model light absorption, exciton formation, and spin-dependent charge and energy transport across materials interfaces. Symmetry-aware graph neural networks and active learning approaches will be employed to establish interpretable structure-dynamics relationships and accelerate the prediction of excited-state processes in complex semiconductor materials while enforcing physically meaningful constraints. The project will deliver open-source computational tools, including the pysocdynamics software package, and the hois-spindb benchmark database to support data-driven materials discovery and excited-state simulations. Broader impacts activities will include interdisciplinary student training, outreach activities for high school students and teachers, and international workshops through the global initiative of academic networks (gian) focused on computational molecular science and AI-enabled materials modeling. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the foundation's intellectual merit and broader impacts review criteria.- subawards are not planned for this award.