Molecular signature evolution of dissolved organic matter in wastewater during far-ultraviolet/peracetic acid disinfection: Integrated characterization and machine learning.
Journal:
Water research
Published Date:
Nov 1, 2025
Abstract
A far-ultraviolet/peracetic acid (UV222/PAA) disinfection process was developed in this study to investigate the transformation mechanisms of dissolved organic matter (DOM) in wastewater, via an integrated approach combining comprehensive molecular characterization with machine learning (ML) analysis. The results revealed that DOM was well transformed during the disinfection process, with reductions of 33.8 % in UV254 absorbance, 59.2 % in total organic carbon, and 66.1 % in total fluorescence intensity. These findings indicate that the majority of DOM underwent oxidative evolution, involving structural and compositional changes, while some fraction demonstrated recalcitrance. Integrated analysis using fluorescence excitation-emission matrix-parallel factor analysis and Fourier transform ion cyclotron resonance mass spectrometry revealed that during disinfection, DOM molecular characteristics evolved from highly aromatic, unsaturated, and reduced species to structures with decreased aromaticity, increased saturation, and higher oxidation states. The electron-rich functional groups in DOM were preferentially attacked by the reactive oxidative species generated in the system, as clearly revealed by Fourier transform infrared and two-dimensional correlation spectroscopy and electron paramagnetic resonance. Dealkylation reactions (e.g., -CH2) and oxygen addition reactions (e.g., +O) were identified as the predominant DOM transformation pathways, accounting for 28.3-34.3 % and 23.6-30.0 %, respectively. Furthermore, ML was employed to establish structure-reactivity relationships for DOM components, with molecular characteristics interpreted through Shapley additive explanations analysis. The results revealed the O/C ratio, molecular weight, and nitrogen content as key molecular descriptors affecting DOM reactivity during UV222/PAA disinfection. Notably, lignin/CRAM-like structures preferentially undergo oxidative degradation, while the accumulation of high-O/C tannin-like compounds highlights their resistance characteristics. Furthermore, nitrogen-containing DOM components warrant particular attention due to their potential role as precursors for nitrogenous disinfection byproducts. Our study not only provides a foundation for understanding the molecular-level transformation mechanism of DOM during UV222/PAA disinfection but also offers valuable insights for optimizing disinfection strategies in wastewater reuse scenarios.
