Exploration of key factors in the preparation of highly hydrophobic silica aerogel from rice husk ash assisted by machine learning

To expand the applications of hydrophobic silica aerogels derived from rice husk ash (HSA) through simple traditional methods (without adding special materials or processes), this paper employs machine learning to establish mathematical models to identify optimal conditions for extracting water glass and investigates how preparation conditions and heat treatment temperatures affect properties such as the porosity and hydrophobicity of HSA. The results indicate that the decision tree regression model provides the most accurate predictions for the extraction rate and modulus of water glass. Notably, the water contact angle of HSA produced using nitric acid as a catalyst can reach as high as 159.5°, classifying it as a superhydrophobic material. Additionally, while moderately increasing the concentration of the hydrophobic modifier enhances HSA’s hydrophobicity, it concurrently reduces its porosity. The HSA maintained hydrophobicity until 500 °C. The pore structure of HSA collapsed gradually with the increase in heat temperature. After treatment at 700 °C, HSA lost its hydrophobicity and the porous structure was severely damaged. Compared with silica aerogel using traditional silicon sources, the damage to pore structure and the crystallization occurred at lower temperatures, but the hydrophobicity remained at higher temperatures.