On the reliability of cellulose hydrolysis models based on the extension of Michaelis-Menten type kinetics

The development of reliable models in the case of biomass feedstocks transformation by the means of hydrolytic enzyme cocktails is of critical importance for the process design, optimization and control. For this, the elementary deconstruction mechanisms related to enzyme/substrate interactions have to be elucidated and mathematically represented. The classical modelling approach is based on Michaelis-Menten type kinetics further modified to account for end-product inhibition and enzyme adsorption. In most cases, the enzyme adsorption is supposed following Langmuir-type isotherm. The transformation process is thus modelled as two consecutive steps (i) the liquefaction of the substrate particles by the action of the adsorbed enzymes (heterogeneous reaction) releasing cellobiose as final product in the case of cellulose transformation (ii) the hydrolysis of the cellobiose into glucose by the b-glucosidase activity modelled as a Michaelis-Menten kinetic since it occurs in a homogeneous system.

In this work, we use this formalism to describe the evolution of glucose and cellobiose during the hydrolysis of Avicel by an enzymatic cocktail (Celluclast). The model parameters are identified by regression of typical saccharification data. Although an accurate fit is obtained, the model is unable to predict the course of hydrolysis after a subsequent addition of fresh substrate or fresh enzymes. These results question the reliability of usual kinetic based model for process design/optimization studies. In order to overcome this major issue, an original model including the evolution of particle sizes and specific area is presented and its potentialities discussed.