Investigation on the co-pyrolysis of bamboo sawdust and low-density polyethylene via online photoionization mass spectrometry and machine learning methods

Abstract

The high oxygen content of biomass is the limitation for its widespread utilization, and the co-pyrolysis of biomass feedstock with plastics has gotten much attention by researchers in recent years. Herein, bamboo sawdust (BS) and low-density polyethylene (LDPE) are selected to investigate the synergistic effects during co-pyrolysis via photoionization mass spectrometry (PI-MS). Based on the statistical analysis, for enhancing the synergetic effects of co-pyrolysis, the suitable LDPE/BS ratio should be at least 1:1. During the co-pyrolysis, LDPE promotes the conversion of BS-derivatives to low molecular weight oxygenated compounds (e.g., furfurals and simple phenolic compounds), while the decomposition of LDPE could also be enhanced by BS pyrolysis products. The synergetic effect of LDPE and lignin is more notable than that of LDPE and cellulose or hemicellulose. Meanwhile, for overcoming the drawback of PI-MS analysis in obtaining quantitative information, the RF (Random Forest) and LSTM (Long Short-Term Memory) algorithms are successfully used to predict the TG curves of BS/LDPE mixtures via the temperature-evolved profiles of representative products. This work could provide a reliable tool for monitoring co-pyrolysis process at different LDPE/BS ratios and combine with machine learning, which is helpful to reduce the cost of industrial application by shortening the experimental cycle.