Advances and perspectives of response surface methodology for product optimization, prediction and synergy identification in biomass and plastic co-pyrolysis
Shengyu Xie, Qingting Yang, Chuan Ma, Shogo Kumagai, Hao Wu, Changsong Zhou, Qiangqiang Ren, Hongmin Yang, Toshiaki Yoshioka
Abstract
Co-pyrolysis of biomass and plastic waste has emerged as a promising thermochemical route for sustainable fuel and chemical production. However, the complex interplay of biomass composition, plastic types, and operating conditions causes uncertainties in product distribution and ambiguous pyrolytic synergy, hindering process optimization and prediction. Compared with conventional methods, response surface methodology (RSM) can simultaneously assess multi-variable interactions, reduce experimental numbers, and visualize synergistic effects via intuitive response surface plots. This review discusses recent progress in applying RSM to the co-pyrolysis of biomass and plastic, particularly in modeling the influence of critical variables. Temperature and feedstock mixing ratio are important factors in co-pyrolysis, and understanding the independent and interactive effects of pyrolysis parameters is essential for maximizing the yield of target products. Additionally, RSM has demonstrated high predictive accuracy (R 2 > 0.9) in numerous studies for key pyrolysis outputs, including yields of bio-oil, char, gas, and specific pyrolyzates (levoglucosan, furfural, phenols, and hydrocarbon oil). The potential pyrolytic synergy between biomass and plastics can be visualized through quadratic and cubic models, arising from radical reaction mechanism, heat transfer, and catalytic effects. More importantly, integrating hierarchical cluster analysis has further enhanced the understanding of how interactions influence low-yield pyrolyzates. The combination of RSM and machine learning algorithms has shown promise in improving predictive accuracy and optimizing co-pyrolysis process parameters. Future research should focus on improving model adaptability, enhancing predictive accuracy, strengthening mechanism interpretability, and enabling the intelligent design of co-pyrolysis processes.