Molecular Transformation of Dissolved Organic Matter at the Coal-Water Interface: Insights from High-Resolution Mass Spectrometry

This study employed untargeted high-resolution mass spectrometry combined with multivariate statistical analysis to elucidate the detailed molecular-geochemical differentiation patterns and underlying driving mechanisms during the migration and transformation of coal-derived dissolved organic matter (DOM) into mine groundwater. Results indicate that the coal DOM molecular library was highly complex and diverse, with a total of 40,158 molecular formulas identified, primarily composed of heteroatom-rich, nitrogen- and sulfur-containing compounds (e.g., CHON, CHONS classes) and condensed aromatic structures. In contrast, the total molecular formulas in mine groundwater DOM markedly decreased to 8,829, and its composition underwent fundamental restructuring, shifting toward a predominance of CHO-type compounds, a notable increase in phosphorus-containing constituents, and overall higher hydrophilicity and oxidation state of molecules. Both Van Krevelen diagrams and principal component analysis statistically confirmed substantial differences and clear separation between the two DOM types in terms of elemental composition and molecular characteristics. Further analysis revealed that this differentiation was driven by stringent environmental screening, potentially influenced by synergistic processes including physicochemical selective dissolution, microbial degradation and transformation, and adsorption at the water-rock interface. The identification of 1,529 shared molecules common to both media, predominantly consisting of aliphatic and lignin-like CHO/CHON compounds, provides putative molecular evidence of coal-water interaction. These findings offer essential molecular-level insights and a scientific basis for accurately tracing organic pollutant sources in mine groundwater, assessing pollutant migration risks, and understanding the transformation and stability of the subsurface carbon pool.