Decomposing chemical warfare agents (CWA) into chemically inactive forms is a complex multi-aspect problem. Due to restrictions on using CWA, experiments are limited. Currently, there is active research on increasing the adsorption capacity using metal-organic frameworks (MOF), which can also act as catalysts. We present a framework to screen potential MOF candidates to capture and decompose CWA, beginning with aminated MOF-808, a Zr-based MOF functionalized with an amine group. This functionalization makes MOF reactive towards nerve agents (reactive towards Lewis acid sites) and blistering agents (reactive towards Lewis base sites). The framework assesses the diffusivity and adsorption capacity using empirical methods, which MOF-808 fulfilled. We develop reaction-aware machine-learning potential using reactive active learning (RAL) protocol on selected materials to identify novel reaction intermediates and CWA decomposition pathways. The framework is extensible to other class of reactions in different nanoporous materials.
