Abstract
Resolving the three-dimensional structure of transition metal oxide nanoparticles (TMO-NPs), upon self-restructuring from solution, is crucial for tuning their structure–functionality. Yet, this remains challenging as this process entails complex structure fluctuations, which are difficult to track experimentally and, hence, hinder the knowledge-driven optimization of TMO-NPs. Herein, we combine high-energy synchrotron X-ray absorption and X-ray total scattering experiments with atomistic multiscale simulations to investigate the self-restructuring of self-assembled Co-NPs from solution under dark or photocatalytic water oxidation conditions at distinct reaction times and atomic length-scales. Using the atomic range order as a descriptor, we reveal that dissolution of a Co-salt in BO3 buffer leads to a self-optimization route forming disordered oxyborate Co3BOx-NPs unveiling a high oxygen yield due to the formation of surface oxo/hydroxo adsorbates. Those Co3BOx-NPs further self-restructure into distorted Co3O4-NPs and, lastly, into distorted CoOOH-NPs through a rate-limiting step integrating Co3+-states during the course of a representative photocatalytic assay. Self-restructuring does not proceed from amorphous-to-ordered states but through stochastic fluctuations of atomic nanoclusters of ≈10 Å domain size. Our key insight into the structure-selection dynamics of TMO-NPs from solution offers a route for tuning their structure–function relationships for wide-ranging emergent technologies.