SearcharxivSearch

arXiv subjects

Nabankur Dasgupta

Publications and source records attributed to Nabankur Dasgupta.

3 recordsLinked to original sources

Toward Exascale AI for Science: A Scalable AI Skill for Autonomous Microkinetics Discovery

We present a scalable AI-driven framework that advances autonomous scientific discovery by combining agentic workflow automation, high-performance computing, and scientific surrogate models. Using microkinetics discovery as a testbed, the work demonstrates how AI can reduce expert intervention, recover from failed simulations, and systematically evaluate surrogate model reliability. This study shows how AI skills can transform complex domain workflows into robust, scalable capabilities for next-generation materials research.

cs.CE

Emerging Ferroelectric Domains: Stacking and Rotational Landscape of MoS2 Moire Bilayers

The structures and properties of moire patterns in twisted bilayers of two-dimensional (2D) materials are known to depend sensitively on twist angle, yet their dependence on stacking order remains comparatively underexplored. In this study, we use molecular dynamics simulations to systematically investigate the combined effects of stacking order and rotation in MoS2 bilayers. Beginning from five well-established high-symmetry bilayer stackings, we apply twist angles between 1 and 120 to the top layer, revealing a variety of relaxed moire structures. Our results show that the initial stacking significantly influences the moire domain configurations that emerge at a given twist angle. While all five stacking orders are metastable without twist, they form two moire-equivalent classes- AA/AB and AA',A'B,AB', i.e., for a given twist angle, structures within each class relax to the same moire configuration. Specifically, initial AA and AB stackings give rise to triangular ferroelectric domains near 0+/-3, while AA', A'B, and AB' stackings produce triangular ferroelectric domains near 60+/-3. At precisely 60 and 120 twists, the bilayers relax to into pure high-symmetry stackings, highlighting the rotational relationships between these configurations and explaining the shift of 60 in the ferroelectric rotational range. These findings demonstrate the critical role of stacking order in governing the rich moire landscapes accessible in twistronic systems.

cond-mat.mtrl-sci

Allegro-FM: Towards Equivariant Foundation Model for Exascale Molecular Dynamics Simulations

We present a foundation model for exascale molecular dynamics simulations by leveraging an E(3) equivariant network architecture (Allegro) and a set of large-scale organic and inorganic materials datasets merged by Total Energy Alignment (TEA) framework. Thanks to the large-scale training sets, the obtained model (Allegro-FM) is versatile for various materials simulations for diverse downstream tasks covering 89 elements in the periodic table. Allegro-FM exhibits excellent agreements with high-level quantum chemistry theories in describing structural, mechanical, and thermodynamic properties, while exhibiting emergent capabilities for structural correlations, reaction kinetics, mechanical strengths, fracture, and solid/liquid dissolution, for which the model has not been trained. Furthermore, we demonstrate the robust predictability and generalizability of Allegro-FM for chemical reactions using the Transition1x that consists of 10k organic reactions and 9.6 million configurations including transition state data, as well as calcium silicate hydrates as a testbed. With its computationally efficient, strictly-local network architecture, Allegro-FM scales up to multi-billion-atom systems with a parallel efficiency of 0.964 on the exaflop/s Aurora supercomputer at Argonne Leadership Computing Facility. The approach presented in this work demonstrates the potential of the foundation model for a novel materials design and discovery based on large-scale atomistic simulations.

cond-mat.mtrl-sci