Silica Aerogel Synthesis/Process-Property Predictions by Machine Learning

Silica aerogels are mesoporous high surface area materialswithextensive synthetic and processing conditions. To effectively synthesizeaerogels, the impact of synthetic pathways on the resulting aerogelproperties must be understood prior to experimental investigation.We develop an information architecture, the silica aerogel graph database(10(3)), and a supervised machine learning neural networkregression model to examine these relationships. The property graphdatabase enables rapid queries and visualization of the impact ofthe synthesis and processing conditions on the final aerogel properties.The model maps from silica aerogel synthetic and processing conditionsto predict the aerogel BET surface area with an average error of 109 & PLUSMN; 84 m(2)/g. Following a validation experiment, themodel was shown to predict the aerogel surface area from new syntheticand processing conditions with an error of less than 5%. The experimentdemonstrates the usefulness of the model in surface area predictionthrough the compatibility between computational and experimental results.Both in its current form and with further expansion, the developedgraph database could reduce experimental dimensionality, time, andresources, enabling the successful synthesis of high surface areasilica aerogels, which are advantageous for applications includingthermal insulation, sorption media, and catalysis.