8-06: Mechanistic modeling of lignocellulosic biomass pretreatment and enzymatic hydrolysis to enable virtual engineering

Mechanistically based mathematical models incorporating relevant chemical and physical phenomena are being used to aid scientific discovery and technology development. With computing power expanding rapidly, increasingly more sophisticated computational models are being implemented to enable “virtual” science and engineering to be carried out. For example, computational fluid dynamics (CFD) coupled with reaction kinetics models are routinely used in the development of combustion and fermentation processes, allowing much faster scale-up, design, and optimization of these technologies compared to traditional approaches that rely on experimentation and empirical modeling. Existing reaction models from the chemical process industries do not apply well to biochemical biomass conversion, however, due to the complex, heterogeneous nature of lignocellulosic feedstocks. Moreover, many proposed empirical models are quite limited in their range of predictive capabilities.

Here, we present recently developed, mechanistically based kinetic models for dilute-acid pretreatment and enzymatic saccharification of corn stover as a prerequisite to establishing a virtual engineering capability for biochemical conversion. The heterogeneous structures of milled biomass particles are modeled as grains of substrate distributed within polydisperse solid particles. The pretreated biomass particles are represented as cellulosic cores partially covered by xylan and lignin that impede access of glucanase enzymes. Distributed species (particles and cellulose polymers) are accounted for with population-balance equations. The use of these models to simulate conversion of corn stover to sugars over a broad range of substrate properties is presented and compared with experimental data. Progress toward coupled CFD and kinetics models and integrated-process simulations is also discussed.