Prediction of methane yield and pretreatment efficiency of lignocellulosic biomass based on composition

Lignocellulosic biomass is considered a key resource for the future expansion of biogas production through anaerobic digestion (AD), and research on the development of pretreatment technologies for improving biomass conversion is an intensive and fast-growing field. Consequently, there is a need for creating tools able to predict the efficiency of a certain pretreatment on different biomass types, fast and accurately, and to assist in selecting a pretreatment technology for a specific biomass. In this study, seven different types of raw lignocellulosic biomass of industrial relevance were systematically analyzed regarding their composition (carbohydrates, lignin, lipids, ash, extractives, etc.) and subjected to a common pretreatment. The aim of the study was to identify the most important characteristics that make a biomass good receptor of the specific pretreatment prior to AD. A simple ammonia pretreatment was chosen as a case study and partial least squares regression (PLS-R) was used for modeling initially the ultimate methane yield of raw and pretreated biomass. In the sequel, PLS-R was used for modeling the efficiency of the pretreatment on increasing the ultimate methane yield and hydrolysis rate as a function of the biomass composition. The fit of the models was satisfactory, ranging from R² = 0.89 to R² = 0.97. The results showed that the most decisive characteristics for predicting the efficiency of the pretreatment were the lipid (r = -0.88), ash (r = +0.79), protein (r = -0.61), and hemicellulose/lignin (r = -0.53) content of raw biomass. Finally, the approach followed in this study facilitated an improved understanding of the mechanism of the pretreatment and presented a methodology to be followed for developing tools for the prediction of pretreatment efficiency in the field of lignocellulosic biomass valorization.