Experimental investigation and prediction of chemical etching kinetics on mask glass using random forest machine learning

Chemical etching on the surface of glass is an essential program to improve its anti-reflective properties of glass. Developing a model for chemical etching kinetics is crucial for improving and refining the etching process. In this study, we investigated the impact of reaction temperature, reaction time, the viscosity and additives of the chemical etching solution on the kinetics of chemical etching by experiment. Random forest was trained using 400 chemical etching reaction rates under different operating conditions. Base on machine learning model training, the random forest demonstrated strong predictive capability with an R-2 exceeding 0.9. Additionally, the impacts of chemical etching kinetics were analyzed and the machine learning model was evaluated by etching experiments. The relative importance of chemical etching kinetics conditions was reaction time > the viscosity of solution > the amount of thickener added > reaction temperature > the amount of sodium gluconate added > the amount of water reducer added. Finally, a high-accuracy chemical etching kinetics model was established.