Enhancing biogenic carbon dioxide conversion to acetic acid using support materials and machine learning-based optimization
Year: 2025
Type: article
Abstract: • Conductive and packing materials selection affects CO 2 -to-acetic acid conversion. • No significant interaction effect between the support materials was observed. • BES applied at fixed concentration to inhibit methanogenesis and aid acetogenesis. • Zero-valent iron was the most effective support material (99.8 % efficiency). • ANN-GA showed better predictive accuracy than MCD with under 1 % prediction error. Amid escalating environmental challenges and the global pursuit of carbon neutrality, sustainable pathways for platform chemical synthesis are a critical research priority. Acetic acid, a high-demand compound with extensive industrial applications, is still predominantly produced via petroleum-based processes, raising environmental and economic concerns. Considering these challenges, this study explores the biological conversion of biogenic carbon dioxide (CO 2 ) into acetic acid, promoting a circular carbon economy. Specifically, it investigates the role of packing and conductive materials in influencing biological CO 2 -to-acetic acid conversion, demonstrating that their effects are material-specific, with some enhancing and others inhibiting process performance. To enhance acetogenic activity, 2-bromoethanesulfonic acid was used to inhibit methanogenesis, thereby favoring the Wood–Ljungdahl pathway. The experimental design incorporated a multilevel categorical design (MCD) to assess the individual and combined effects of support materials, namely packing and conductive types, on key process parameters: pH, yield (g acetic acid/g CO 2 ), and efficiency (%). Zero-valent iron was identified as the most effective conductive material, achieving a yield of 0.68 g acetic acid/g CO 2 and an efficiency up to 99.8 %. The statistical models developed through MCD exhibited strong predictive performance (R 2 > 0.80; prediction errors < 3 %) and were validated experimentally. Furthermore, an artificial neural network (ANN) was trained and integrated with a genetic algorithm (GA) to optimize system performance. The ANN-GA model demonstrated slightly higher predictive accuracy compared to MCD. Overall, this study confirms the significant role of support material selection in enhancing biological CO 2 conversion and presents an advanced modeling framework for process optimization.
Source: Bioresource Technology
Language: English
Subfield: Environmental Engineering
Field: Environmental Science
Domain: Physical Sciences
Open Access status: hybrid
Estimated APC paid: $4,880
Award: G2078018860
OpenAlex ID: W4415292276