Conventional heterogeneous catalysts are constrained by static atomic structures, which are ill-suited for the diverse requirements of multistep reactions. This article introduces Field-Programmable Dynamic Catalysis (PDC), a paradigm enabling real-time, reversible programming of atomic structures via external fields and intelligent feedback. Using copper-based liquid metal catalysts, we demonstrate adaptive restructuring that optimizes methane oxidation performance. With magnetically tunable iron-based catalysts, we achieve reversible switching between atomically dispersed and clustered states, thus validating on-demand control over reaction pathways. This paradigm holds great promise for inert small molecule activation, such as nitrogen conversion, yet key challenges remain in boosting liquid metal surface area and designing scalable reactors. To address this multidimensional complexity, artificial intelligence is indispensable for accelerating the discovery and development of advanced PDC catalysts. Ultimately, PDC transforms catalysts from passive materials into active information-processing systems, enabling intelligent and precise chemical control at the atomic scale.
Zhang et al. (Tue,) studied this question.