The potential of decision tree application in threshold analysis of hazardous volatile organic compound release from biochar: Implications for environmental risk assessment

The release of hazardous volatile organic compounds (HVOCs) from biochar poses a potential threat to both human health and the environment. This study investigates how low pyrolysis temperature (HTT) and the chemical characteristics of lignocellulosic biomass, expressed as the carbon-relative molar mass (CRMM) index (g·mol^-1), influence the release of hazardous volatile compounds from biochar. A total of 204 biochar samples were produced from pure biomass components: lignin (L), cellulose (C), hemicellulose (H) and their mixtures, and analyzed using stepwise multiple regression, agglomerative hierarchical clustering, principal component analysis (PCA), and regression decision trees. The analysis included both qualitative and quantitative assessments of HVOCs. Qualitative results indicated the presence of HVOCs, but quantitative patterns were nonlinear and inconsistent with expected trends. Decision tree analysis showed that the highest release occurred either at very high CRMM values (>34 g mol^-1) or at low-temperature conditions (200-325 °C) with low CRMM. The lowest release was associated with temperatures 350-425 °C and elevated CRMM. The results emphasize the need to consider both process and substrate parameters when assessing release risks, providing a new approach to evaluating the safety of biochar in the context of environmental and human health protection. To our knowledge, this is the first study to apply decision tree regression to model HVOC release from biochar based on CRMM and pyrolysis conditions.