Detection of perfluorooctance sulphonic acid in groundwater using an intelligent array of electrochemical sensors

The persistent environmental pollutant family, namely, per-and polyfluoroalkyl substances (PFAS), are widely used materials for manufacturing industrial products, such as non-stick cookware, paints, and firefighting foams. PFAS pose serious concerns because they are bio-accumulative, toxic, and persist in the environment, having a harmful impact on human health and the environment. Current methods to detect PFAS, such as liquid chromatography-mass spectrometry, are very sensitive; however, they are time consuming, based on benchtop equipment, require specialized staff, and cannot be performed outside of an analytical laboratory; hence, they impede rapid assessment of PFAS in the field. Here, we discuss the development of a novel and generic electrochemical sensor that rapidly detects PFAS levels in groundwater in the field. As a proof of concept, we demonstrate the detection of a common form of PFAS, perfluorooctance sulphonic acid (PFOS). This sensor is based on an array of electrodes modified with fluorinated and hydrophobic elastin-like peptides (ELPs). In the presence of a redox mediator, the partially selective interactions of PFOS with the modified electrodes generate a set of cross-reactive electrochemical signals that are then analyzed using chemometrics. First, we characterized the surface modification with ELPs using electrochemical and XPS elemental methods. Then we studied the electrode-electrolyte interface properties of the ELP-modified electrodes in the presence and absence of PFOS. We used a standard addition method to elucidate the sensitivity of each ELP-modified electrode to PFOS, revealing a sensitivity of 4.03 ± 0.25 Ohm/ppb and 22.7 ± 2.7 Ohm/ppb for the fluorinated and the hydrophobic electrodes, respectively. Next, we characterized the effect of the major interference molecule, humic acid, on the analytical performance of the fluorinated and the hydrophobic ELP-modified electrode, which showed no significant difference in the signal. To calibrate the multi-sensor array, we used the intelligent chemometric partial least squares regression model. Using the generated electrochemical signals in the presence of PFOS in buffered solution, we successfully predicted PFOS levels in simulated groundwater solutions, resulting in a limit of detection (LOD) value of 25.2 ± 1.5 ppb. To test our sensor performance in a real-world scenario, we trained a new chemometric model using PFOS in simulated groundwater (LOD 56.7 ± 6.3 ppb, PCC 0.980, and RMSE = 0.0800 ppb) and using the model, we were able to predict the PFOS levels in PFOS-spiked real groundwater. The ability to achieve a selective response to PFOS by an array of sensors modified with functionalized ELPs represents a novel and sustainable approach to rapidly detect PFAS in real-world water samples in the field.