Étude approfondie de la distillation réactive enzymatique hétérogène : de l'expérimentation à la modélisation et simulation des procédés

In the context of global efforts to improve sustainability and process efficiency in chemical manufacturing, enzymatic reactive distillation (ERD) stands out as a promising candidate. By integrating biocatalytic reactions and separation into a single unit, ERD addresses significant challenges in traditional chemical processes such as improving conversion rates and lowering energy consumptions, all while leveraging an eco-friendly catalyst of elevated specificity. However, the use of ERD presents unique complexities including enzyme stability, limited operating conditions, and the need for effective immobilization strategies. Additionally, the impact of all operating parameters on the process performances are not yet fully understood, the chemical industry hence being hesitant to adopt this technology. As a way to answer these challenges, the present work investigates a continuous and enantioselective ERD setup at the laboratory scale, using both experimental and modeling approaches. The enantioselective transesterification of ethylbutyrate with rac-2-pentanol, catalyzed by the Lipase B from Candida antarctica (CALB), was chosen as a model reaction. The transesterification was studied with two immobilized forms of the lipase: the commercial Novozym®435 resin beads and a xerogel containing the CALB prepared by the sol-gel process. Detailed and simplified kinetic models taking into account internal mass transfer were developed for the two catalysts, addressing the impact of immobilization methods. To integrate the reaction into the ERD column, novel distillation packings such as open-cell foams and 3D printed Tetra Splines were functionalized with the CALB xerogel. The issue of coating adhesion was addressed through an enzyme-friendly approach, and the activated structured then applied to ERD. A parametric study allowed to evaluate the influence of catalytic load, feed composition and column configuration on the steady-state, despite operational issues that affected the system stability. Computational models employing equilibrium (EQ) and non-equilibrium (NEQ) approaches were used to simulate the ERD process. While EQ models provided reliable predictions at low yields, NEQ models were constrained by insufficient mass transfer data. A thorough parametric study through simulations validated experimental observations and gave further insight into ERDs optimization, enabling for the investigation of scaled-up columns through modeling.