Strong internal electric field-modulated carbon nitride for tetracycline photodegradation: Guidance by theoretical calculation and machine learning.

Focusing on the limited light absorption and poor charge behaviors of graphitic carbon nitride (CN), this work selected diethyl 2,5-bis(thieno[3,2-b]thiophen-2-yl)terephthalate (DT) as the dopant guided by theoretical calculations to synthesize a fungus-like ultrathin porous carbon nitride (x-DCN) through one-step thermal-induced copolymerization. The embedding of DT units achieves effective spatial separation of the frontier orbitals, establishing a donor-acceptor (D-A) structure with a strong internal electric field (IEF). Under visible light irradiation, the optimized 10-DCN exhibited a tetracycline (TC) degradation rate four times higher than that of pristine CN. The abundant edge active centers, wider visible-light responsiveness and optimized carrier separation-migration kinetics, collectively propelled the photoactivity. Meanwhile, machine learning models, e.g., the Random Forest (RF) model, with high fitting accuracy (R2 = 0.98), enabled precise quantification of the contribution of degradation variables. Finally, the enhanced degradation mechanism, predominant reactive radicals, reaction pathways, and toxicity were investigated. This research provides a new perspective for the precise design and application of CN-based photocatalysts with a strong IEF.