Unraveling Complexities of Nanostructured Sol-Gel SiO₂ Synthesis with Automation and ML-Accelerated Experimentation

Silica nanoparticles can be synthesized through relatively straightforward sol-gel processes that involve common precursors and ambient synthesis conditions. Developments in this process have uncovered appropriate synthesis conditions to create spherical and mesoporous nanoparticles of sizes spanning nanometers to microns. However, this relatively accessible synthesis process belies the complexity of the parameter space that must be navigated to attain a desired level of morphological control. Particle diameter, polydispersity, and internal structure or ‘porosity’ are all controlled by many interdependent and often competing factors including precursor types, concentrations, pH, and temperature amongst others. This large parameter space makes it challenging to achieve retrosynthetic planning for the synthesis of particles with a specific morphology. Accelerated experimentation via automation and artificial intelligence integration can aid in selecting appropriate synthesis conditions to achieve target particle morphologies for use in applications including catalytic supports and drug delivery, as well as many others. We have developed an automated experimentation platform and workflow for the autonomous high-throughput synthesis and characterization of silica nanoparticles. This system uses the Jubilee laboratory automation platform to create nanoparticles via sol-gel synthesis processes, and a custom sample loading system to integrate small angle x-ray scattering structural characterization. This system has enabled us to extensively explore the relationships between synthesis parameters and morphological outcomes for silica nanoparticles. This automation platform also enables the application of closed-loop optimization techniques to develop synthesis conditions for nanoparticles with specific target diameters and porosity properties.