Understanding Dynamics of Nanocluster-Organic Frameworks and Gas Diffusion from Machine Learning Potential-Based Simulations

Nanocluster-organic frameworks (NOFs) are emerging materials with broad applications in sensing, photocatalysis, and optoelectronics. These photofunctional materials have an excellent luminescence switching response to gases, such as oxygen (O₂) and volatile organic compounds. However, their atomistic structural evolution and the mechanism governing the diffusion of small molecules inside them remain elusive. In this study, we developed machine learning potentials to accurately model an experimentally synthesized NOF, [Ag₁₂(S^tBu)₈(CF₃COO)₄(bpy)₄)]_n, and we investigated its structure and dynamics using molecular dynamics simulations. Furthermore, we used on-the-fly probability-enhanced sampling simulations to study O₂ gas diffusion inside NOF pores and construct the free energy surface. The O₂ gas predominantly localizes around the bipyridine linker, consistent with previous experimental observations. The pore-to-pore diffusion barrier of ∼16-18 kJ/mol for O₂ suggests feasible diffusion of the O₂ at room temperature. This work presents the first-ever integrated approach for modeling a fully flexible NOF and gas diffusion within it with DFT-level accuracy.