.png){kind=icon}
.png){kind=icon}
Return
Charge-Regulated Surface-Enhanced Raman Spectroscopy (SERS) Fingerprints Combined with Data Fusion for Source Identification of Domestic Wastewater under Environmental Dilution
.png){kind=icon}
PRE
AI
DOI：10.1007/s41664-026-00481-3.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Reliable source-resolved detection of domestic wastewater in receiving waters remains challenging due to strong dilution effects and variable background matrices. Here, we introduce a charge-regulated surface-enhanced Raman spectroscopy (SERS) strategy that employs negatively charged citrate-reduced Ag nanoparticles and positively charged poly(diallyldimethylammonium chloride) (PDDA)-modified Ag nanoparticles to fractionate dissolved organic matter and generate complementary spectral fingerprints. Seven domestic wastewater sources were profiled and diluted in river and lake waters up to 1000× to evaluate the feasibility of this approach for wastewater sources tracking in real receiving waters. SERS spectra acquired from both substrates were embedded using principal component analysis (PCA) and integrated through a multilevel data fusion framework, including feature-level fusion of low-dimensional representations and probability-level fusion of classifier outputs. Model performance was systematically evaluated using multiple conventional machine-learning classifiers to assess robustness across algorithms. The fused models consistently outperformed single-substrate approaches, achieving cross-validated classification accuracies of up to 99.8% in river-water dilutions and maintaining reliable predictive performance, even at extreme dilution. The charge-regulated design improved sensitivity and selectivity by capturing oppositely charged components that are under-represented on individual substrates, while data fusion combined non-redundant information to stabilize decisions across variable matrices. Overall, this measurement-plus-chemometrics framework provides a rapid, label-free complement to high-resolution mass spectrometry (HRMS) for wastewater fingerprinting and establishes an apportionment-ready pathway for quantifying source contributions in complex aqueous mixtures. This approach can also be extended to other multicomponent effluent systems, showing potential for source tracking in water pollution incidents.
Keywords:
Surface-enhanced Raman spectroscopy (SERS)
Wastewater source identification
Surface charge regulation
Data fusion
Machine learning
Journal
J
IF:
7
Papers:
480
Citations:
1.2K