A reaction-aware autonomous platform co-optimizing formulation and execution for the synthesis of zeolitic imidazolate framework

Autonomous experimentation is transforming materials discovery, yet most platforms rely on endof-run measurements, ignoring reaction-progress information and imposing rigid execution protocols. Here, we present Axo, a reaction-aware batch autonomous platform that integrates timeresolved in situ optical monitoring into liquid-handling synthesis. Each run is recorded as a reaction trajectory and converted into two kinetic descriptors: the extent of reaction that adapts measurement interval and batch size to match the reaction timescale, and the nucleation rate constant (𝑘𝑛) that guides the Bayesian optimization of formulation variables. Applied to the solution-phase crystallization of zeolitic imidazolate framework-67 (ZIF-67, Axo identified a rapid-nucleation region with 𝑘𝑛 = 2.33 min-1, exceeding the best condition found by static maximin sampling by over five-fold. This work demonstrates how programmable in situ trajectories can be converted into actionable feedback that enables reaction-aware closed-loop optimization to cooptimize what to test and how to measure in solution-phase materials synthesis.