Formula‐Level Turnover and Homologous‐Series Patterns in Dissolved Organic Matter Associated With Sunlight‐Exposed PLA, PET, PE, and PBAT by Negative‐Ion ESI FT‐ICR Mass Spectrometry
Jun Dong, Qingzhai Meng, Ruipeng Zhang, Chao Ma, Yulin QiABSTRACT
Rationale
Sunlight exposure can alter plastic‐associated material and generate a filterable dissolved organic matter (DOM) pool containing polymer fragments, oxygenated products, and additives. This study compared the molecular composition and time‐dependent spectral differences of DOM associated with four widely used plastics.
Methods
Particles of poly(lactic acid) (PLA), poly(ethylene terephthalate) (PET), polyethylene (PE), and poly(butylene adipate‐co‐terephthalate) (PBAT) were dispersed in water and exposed to natural sunlight for 6 or 48 h. After particle removal, DOM was enriched by PPL solid‐phase extraction and analyzed by negative‐ion electrospray ionization Fourier transform ion cyclotron resonance mass spectrometry. Assigned formulas were compared using elemental classes, intensity‐weighted descriptors, van Krevelen distributions, Jaccard similarity, Bray–Curtis dissimilarity, and homologous‐series screening.
Results
Between 2247 and 4661 molecular formulas were assigned per sample, with maximum absolute mass errors below 0.60 ppm. Relative to corresponding 6‐h spectra, formula numbers were lower at 48 h for PLA (−37.5%) and PET (−39.3%) but higher for PE (+41.8%) and PBAT (+14.0%). Formula‐set similarity was lowest for PET (Jaccard 0.329; Bray–Curtis 0.787) and highest for PBAT (0.712; 0.260). In PLA, a C 3 H 4 O 2 ‐spaced oxygen‐rich series increased from 13.4% to 20.9% of the assigned intensity. A C 10 H 8 O 4 ‐spaced PET‐related series contributed 8.70% at 6 h and 0.17% at 48 h. PET‐48 h was dominated by CHSO formulas, whereas PE showed greater CHNO abundance and higher relative contribution of lower mass oxygenated CH 2 series. PBAT retained the greatest formula‐set overlap between spectra.
Conclusions
The 6‐ and 48‐h spectra showed distinct polymer‐dependent compositional differences. Homologous‐series analysis provided information beyond van Krevelen statistics, whereas nonexclusive N‐ and S‐containing signals highlighted the need for process blanks, dark controls, and tandem mass spectrometry before source‐specific assignments. Because replicate and matched control spectra were not available in the analyzed dataset, these differences are interpreted descriptively and do not demonstrate chemical formation, disappearance, or sunlight‐specific transformation.