Optimization of Colloidal Nanomaterials Synthesis with AI-Driven Experiments and Accessible Automation

The synthesis of nanomaterials (e.g. SiO₂, ZnO) using sol-gel and nucleation-growth processes is strongly affected by a large number of parameters (e.g. pH, concentrations, temperatures, mixing, surfactants) in a complex and large design space. This can be an advantage, due to the high-tunnability of the desired material structure and properties, and a curse since it can lead to the slow and costly exploration of vast design spaces, and can cause confusion in the interpretation of relationships between synthesis parameters and outcomes. Artificial intelligence (AI), when paired with accessible laboratory automation, can greatly accelerate materials optimization and scientific discovery in complex systems such as these. For example, it can efficiently map phase-diagrams with intelligent sampling along phase boundaries, or in ‘retrosynthesis’ problems where a material with a target structure is desired but a synthetic route is not yet known. In this work, we employ an open, accessible and flexible automation platform called ‘Jubilee’ and interface it with advanced characterization tools such as spectroscopy, dynamic light scattering (DLS), and small angle x-ray scattering (SAXS) to help accelerate the efficient exploration of large design spaces. Advanced data science and AI tools are also integrated and used to accelerate the interpretation of results and to propose optimized searching of the unkown design space to maximize the information value of future sampling campaigns. We demonstrate this integrated approach using examples of sol-gel synthesis of ceramic nanomaterials, metal nanoparticles, and new metal chalcogenides. This approach is generalizable and can help accelerate nanomaterials research in every laboratory through the democratization of automation and the adoption of open source practices.