Real-Time Microfluidic Deconstruction and Predictive Modeling of Lignocellulosic Biomass

• Title: Real-Time Microfluidic Deconstruction and Predictive Modeling of  Lignocellulosic Biomass
• Application deadline: 15 April 2026
• Starting date: October 2026

Contexte: Lignocellulosic biomass (LB) is a strategic renewable carbon feedstock mainly composed of cellulose, hemicelluloses and lignin. While rich in fermentable polysaccharides, these carbohydrates are tightly embedded in a lignin matrix that makes LB highly recalcitrant to enzymatic deconstruction.

Current industrial valorization routes rely on harsh physicochemical pretreatments that are energy-intensive, may degrade sugars, and generate inhibitory by-products. Developing mild, low-energy, enzyme-compatible alternatives therefore represents a major scientific and technological challenge for sustainable biorefineries. 

Microfluidics offers unprecedented control over temperature, pH, residence time and hydrodynamics while drastically reducing reagent consumption. It also enables high-throughput screening of enzymatic conditions and time-resolved monitoring of structural and kinetic phenomena. However, a major bottleneck remains: the lack of quantitative, time-resolved correlations between LB micro-architecture evolution, chemical modifications and hydrolysis kinetics at the cell and tissue scales.

Thesis objectives: This PhD thesis is part of the µLB-Predict project at the interface between biochemical engineering and fluid mechanics funded by the Graduate School SIS of Université Paris-Saclay. The objectives of the study are as follows:

• Develop an integrated microfluidic platform for mild pretreatment and enzymatic hydrolysis of lignocellulosic biomass, including optimization of pretreatment strategies and enzyme cocktails; 
• Characterize biomass deconstruction mechanisms by coupling confocal imaging with saccharification kinetics and establishing correlations between structural evolution and hydrolysis performance; 
• Build predictive digital twin integrating structural and kinetic descriptors to model conversion dynamics and identify rate-limiting steps using MATLAB/Python tools.

Supervisors:

• Ph.D. director: Giana ALMEIDA PERRE (giana.perre@agroparistech.fr)
• Co-supervisor: Kevin LACHIN (kevin.lachin@agroparistech.fr)
• Co-supervisor: Aya ZOGHLAMI (aya.zoghlami@centralesupelec.fr)
• Co-supervisor: Hassan EL ITAWI (hassan.el-itawi@centralesupelec.fr)