AI-guided biorefinery optimization for the production of lignin-carbohydrate complexes with tailored properties

Lignin-carbohydrate complexes (LCCs) present a unique prospect for exploiting the synergy between lignin and carbohydrates in high-value products. To date, the production of LCCs in high yields is still an open challenge. Herein, we address this challenge with a novel approach for the targeted production of LCCs. With the help of artificial intelligence (AI), we optimized our AquaSolv Omni (AqSO) biorefinery toward the synthesis of LCCs with high carbohydrate content (up to 60/100 Ar) and high yields (up to 15 wt%). Our AI approach was essential for biorefinery fine-tuning toward maximum performance, while keeping the number of experiments within reasonable limits. More specifically, we followed a Bayesian Optimization approach that allowed us to iteratively collect data and explore the effect on yield and carbohydrate content of selected processing conditions: temperature, process severity, and liquid-to-solid ratio. By means of a Pareto front analysis, we identified optimal trade-offs between the LCC yield and carbohydrate content. We discovered sizeable regions of processing conditions that yield LCCs in 8-15 wt% with carbohydrate content in the range of 10-40/100 Ar. To evaluate the utility of the produced LCCs for future high-value applications, we measured key properties: the glass transition temperature (Tg), the surface tension, and the antioxidant activity. Intriguingly, we found that LCCs with high carbohydrate content are generally related with low Tg and surface tension. The presented biorefinery concept, in conjunction with its AI-guided optimization, is a first step toward the scalable production of LCCs tailor-made for high-value applications.