Controlling metal-organic framework crystallization via computer vision and robotic handling

Abstract

Traditional experimental techniques for metal–organic frameworks (MOFs) crystal growth are often time-consuming due to the need for manual bench chemistry and data analysis. In this study, we integrated laboratory automation with computer vision to accelerate the synthesis and characterization of Co-MOF-74, a framework containing coordinatively unsaturated Co(II) sites. By utilizing a liquid-handling robot, we significantly improved the efficiency of precursor formulation for solvothermal synthesis, saving approximately one hour of manual hands-on labor per synthesis cycle. We developed an accelerated characterization strategy using high-throughput optical microscopy and computer vision to identify the quality of crystallization outcomes. Our computer vision framework, Bok Choy, enables automated feature extraction from microscopic images, improving the analysis efficiency by approximately 35 times compared to manual analysis methods. Using this integrated workflow, we systematically performed a rapid screening of synthesis parameters and examined how each parameter influences the crystal morphology. Furthermore, by varying solvent compositions, we rapidly screened synthesis conditions that modulate crystal formation, identifying regimes that promote isolated crystallization, clustering, or lack of growth. The resulting structured dataset linking synthesis conditions to crystal morphology provides a scalable foundation for machine learning-driven materials discovery. The combination of high-throughput experimentation and automated data analysis establishes a cost-effective and widely applicable platform for accelerating the discovery and optimization of functional materials, with broad applications in catalysis, energy storage, and beyond.