Rapid and Cross-Source Detection of Naphthenic Acids and Phenol Using 3d Fluorescence Spectroscopy

There is increasing concern about the growing oil sands process-affected waters (OSPW) produced from bitumen extraction. Naphthenic acids (NAs) are considered the primary toxic compound in OSPW, and other pollutants, like phenol, also contribute to the overall toxicity. The objective of this study was to investigate the capabilities of fluorescence spectroscopy to identify and quantify commercial naphthenic acids and phenol in natural surface waters. Fluorescence signals were quantified using the peak-picking approach, a simple supervised approach to interpreting the high-dimensional fluorescence excitation-emission matrices (EEMs). An unsupervised dimensionality reduction using principal component analysis (PCA) was also investigated to account for overlapping and convoluted fluorescence signals of NAs and phenol from background natural organic matter. The PCA method could detect the contaminants with mean absolute error (MAE) <1.82 mg L⁻¹ and 12.97 µg L⁻¹ for NAs and phenol at concentrations between 0 and 50 mg L⁻¹ and 0–100 µg L⁻¹, respectively. Furthermore, a novel application of model transfer was investigated to transfer a site-specific fluorescence detection model to another water source. Only a few additional unlabeled fluorescence EEMs of the new water source were used to transfer the features of new surface water. Compared to the standard site-specific method used in detecting contaminants, the transferred detection models showed comparable predictability while significantly reducing the burden on sampling, calibration, and quantifying the contaminants in the samples. The results of this study show capability of fluorescence spectroscopy as a promising tool to monitor the contaminants of interest in different water sources. This work could help contaminant detection in surface waters more effective, easier, and less labour-intensive.