Background: The need for a change from using fossil fuels to using renewable sources of energy is becoming increasingly urgent due to global energy security and the rapidly changing climate. Of the various types of biomass, one of the most abundant and least used biomass sources for the production of second-generation bioenergy is lignocellulosic biomass found in Saccharum officinarum (sugarcane). Sugarcane bagasse and sugarcane leaves provide a large lignocellulosic biomass supply, consisting of 35-50% cellulose, 20-30% hemicellulose, and 15-25% lignin with significant amounts of fermentable sugars available when broken down. However, due to the structural recalcitrance of the lignocellulosic matrix, there is a limitation on enzymatic conversion efficiency at the commercial scale.
Objective: The biomass composition diversity of Saccharum officinarum varieties has been systematically evaluated using the most advanced methods available for pre-treating biomass prior to structural disintegration via enzymatic hydrolysis. This project also provides integrated enzyme-mediated hydrolysis optimization models by Response Surface Methodology (RSM), machine learning techniques, and kinetic modelling techniques to maximize sugar yields while minimizing inhibitory product yields and costs associated with production processes.
Methods: This study used biomass compositional analysis that followed the NREL (National Renewable Energy Laboratory) standard analysis methods. We performed a comparative analysis of alkaline, dilute-acid, steam explosion, and organosolv pretreatments to determine which method most effectively provided cellulose with greater accessibility. Enzymatic hydrolysis was optimized using a Box-Behnken RSM design to evaluate effects of varying enzyme loadings (5 to 30 FPU/g); pHs (4.5 to 5.5); and temperatures (45 to 55 °C). Utilizing the Michaelis-Menton and Langmuir-type competitive inhibition models, we were able to estimate the kinetic parameters for the cellulosic substrates and enzymes used in this study. Finally, techno-economic analysis (TEA) and life cycle assessment (LCA) were integrated to assess the scalability and environmental performance of our methods.
Results: The maximal cellulose accessibility was achieved with an alkaline-steam sequential explosion pre-treatment (87.4% cellulose accessibility) while lignin and hemicellulose were reduced by 73.2% and 61.8%, respectively. Combining RSM-optimized conditions of 20 FPU/g of cellulase, pH 4.8, and 50°C resulted in a glucose conversion efficiency of 92.3%. Gradient boosting machine-learning models had superior predictive accuracy (R2 = 0.98) compared to RSM alone (R2 = 0.91). Bioethanol production via SSF yielded 47.8 g/L, with an 89.1% conversion efficiency of the theoretical maximum. An LCA showed an estimated 76% reduction in net greenhouse gases compared to fossil gasoline.
Conclusion: The goal of this study is to develop a strong overall model optimization framework to optimize the use of sugarcane (Saccharum officinarum) through enzymatic hydrolysis. Integrated pretreatment, utilizing enzyme synergies and advanced optimization methodologies, will enhance the ability to utilize the chemical composition of lignocellulosic materials through enzymatic hydrolysis, therefore overcoming barriers to their use as a biomass feedstock for biofuels. These results will present an economically and environmentally sustainable platform for the production of second generation bioethanol in an integrated circular bioeconomy.