Machine learning prediction for the removal of As and Cd from co-contaminated water by iron-modified biochar from As-enriched P. vittata

Abstract

The carbonization treatment of As-rich Pteris vittate is still in its infancy, and it is urgent to explore the remediation effects of carbonized products on compound contamination and their potential influencing factors. In this study, the remediation effectiveness of biochars using As-enriched P. vittata after modification by iron (FAs-BC) on As(III) and Cd(II) compound-polluted water and its response to the aging process were explored, and machine learning (ML) was used to screen the decisive factors for its remediation effect. The results showed that the adsorption capacity of FAs-BC for As(III) and Cd(II) in binary solution were significantly higher than that in single solution (42.0 % and 82.6 %). After 5 cycles, FAs-BC still maintained 71.1 % and 73.6 % of its initial adsorption capacities for As(III) and Cd(II). Mineral precipitation was the primary mechanism for the adsorption of As(III) and Cd(II) by FAs-BC, while the generation of type B ternary complexes provided a synergistic removal pathway. After aging, the As-TCLP/As-total and Cd-TCLP/Cd-total within saturated with single As(III), single Cd(II), and binary solution were significantly lower than before aging, but still maintained a considerable adsorption capacity for As(III) and Cd(II). The Random Forest model showed that mass volume ratio (S/L), specific surface area (SSA), As concentration (C(As)), Cd concentration (C(Cd)) and AsCd concentration ratio (As/Cd) significantly affected the adsorption of As (III) and Cd (II) by FAs-BC. Overall, this study delved into the remediation effects and mechanisms of FAs-BC on compound contamination, providing a theoretical foundation for the preparation and application of aquatic remediation materials.