arXiv subjects

Alexander J. Hoffman

Publications and source records attributed to Alexander J. Hoffman.

3 recordsLinked to original sources

arXiv2026

EquiformerV3: Scaling Efficient, Expressive, and General SE(3)-Equivariant Graph Attention Transformers

As $SE(3)$-equivariant graph neural networks mature as a core tool for 3D atomistic modeling, improving their efficiency, expressivity, and physical consistency has become a central challenge for large-scale applications. In this work, we introduce EquiformerV3, the third generation of the $SE(3)$-equivariant graph attention Transformer, designed to advance all three dimensions: efficiency, expressivity, and generality. Building on EquiformerV2, we have the following three key advances. First, we optimize the software implementation, achieving $1.75\times$ speedup. Second, we introduce simple and effective modifications to EquiformerV2, including equivariant merged layer normalization, improved feedforward network hyper-parameters, and attention with smooth radius cutoff. Third, we propose SwiGLU-$S^2$ activations to incorporate many-body interactions for better theoretical expressivity and to preserve strict equivariance while reducing the complexity of sampling $S^2$ grids. Together, SwiGLU-$S^2$ activations and smooth-cutoff attention enable accurate modeling of smoothly varying potential energy surfaces (PES), generalizing EquiformerV3 to tasks requiring energy-conserving simulations and higher-order derivatives of PES. With these improvements, EquiformerV3 trained with the auxiliary task of denoising non-equilibrium structures (DeNS) achieves state-of-the-art results on OC20, OMat24, and Matbench Discovery.

cs.LG↗

arXiv2025

DiffSyn: A Generative Diffusion Approach to Materials Synthesis Planning

The synthesis of crystalline materials, such as zeolites, remains a significant challenge due to a high-dimensional synthesis space, intricate structure-synthesis relationships and time-consuming experiments. Considering the one-to-many relationship between structure and synthesis, we propose DiffSyn, a generative diffusion model trained on over 23,000 synthesis recipes spanning 50 years of literature. DiffSyn generates probable synthesis routes conditioned on a desired zeolite structure and an organic template. DiffSyn achieves state-of-the-art performance by capturing the multi-modal nature of structure-synthesis relationships. We apply DiffSyn to differentiate among competing phases and generate optimal synthesis routes. As a proof of concept, we synthesize a UFI material using DiffSyn-generated synthesis routes. These routes, rationalized by density functional theory binding energies, resulted in the successful synthesis of a UFI material with a high Si/Al$_{\text{ICP}}$ of 19.0, which is expected to improve thermal stability and is higher than that of any previously recorded.

cond-mat.mtrl-sci↗

arXiv2025

High-Throughput Transition-State Searches in Zeolite Nanopores

Zeolites are important for industrial catalytic processes involving organic molecules. Understanding molecular reaction mechanisms within the confined nanoporous environment can guide the selection of pore topologies, material compositions, and process conditions to maximize activity and selectivity. However, experimental mechanistic studies are time- and resource-intensive, and traditional molecular simulations rely heavily on expert intuition and hand manipulation of chemical structures, resulting in poor scalability. Here, we present an automated computational pipeline for locating transition states (TS) in nanopores and exploring reaction energy landscapes of complex organic transformations in pores. Starting from the molecular structure of potential reactant and products, the Pore Transition State finder (PoTS) locates gas-phase transition states using DFT, docks them in favorable orientations near active sites in nanopores, and leverages the gas-phase reaction mode to seed condensed-phase DFT calculations using the dimer method. The approach sidesteps tedious manipulations, increases the success rate of TS searches, and eliminates the need for long path-following calculations. This work presents the largest ensemble of zeolite-confined transition states computed at the DFT level to date, enabling rigorous analysis of mechanistic trends across frameworks, reactions, and reactant types. We demonstrate the applicability of PoTS by analyzing 644 individual reaction steps for transalkylation of diethylbenzene in BOG, IWV, UTL and FAU zeolites, and in skeletal isomerization of 162 individual reaction steps in BEA, FER, FAU, MFI and MOR zeolites finding good experimental agreement in both cases. Lastly, we propose a path to address the limitations we observe regarding unsuccessful TS searches and insufficient theory in other reactions, like alkene cracking.

cond-mat.mtrl-sci↗