The cellulose hydrolysis into glucose with carbon-based solid acid catalyst via machine learning, life cycle assessment and bibliometric analysis

Abstract

The cellulose hydrolysis into glucose with carbon-based solid acid catalyst contributes to sustainable and low-carbon development. Determining the quantitative impact of key variables on glucose yield and environmental sustainability is a notoriously time-consuming and labor-intensive process. In this study, a bibliometric analysis was carried out to examine the current research on cellulose hydrolysis into glucose using solid acid catalyst. The application of machine learning (ML) has demonstrated that the random forest (RF) model possesses optimal predictive performance concerning the glucose yield (Training R2 = 0.89, Testing R2 = 0.80, MAE = 6.52, RMSE = 8.29). The SO3H acid densities was emerged as a most crucial influential parameter. The interactions relationship of influenced factors for glucose production was clarified by the partial dependence plot analysis (PDPs). Furthermore, the life cycle assessment (LCA) further indicated that the carbon solid acid catalysts used in cellulose hydrolysis to produce glucose have lower emissions of CO2, NOx, and SO2 compared to traditional acid catalysts, which is 85.5 kg ep of CO2, 0.17 kg ep of NOx, 0.23 kg ep of SO2 for carbon solid acid catalysts, while 91.84 kg ep of CO2, 0.24 kg ep of NOx, 0.61 kg ep of SO2 emission for acid hydrolysis. Thus, carbon solid acid catalysts have a comparably lower negative environmental impact than traditional acid hydrolysis. This study provides crucial insights into improving the prediction accuracy of ML models and reducing the carbon footprint involved in glucose production.