The severe environmental and health risks posed by perfluoroalkyl substances (PFASs), coupled with the limitations of conventional detection methods, have highlighted the urgent need to develop efficient analytical platforms. However, the simultaneous quantification of multiple PFAS targets in complex systems with easy-to-perform operations remains a challenge. Herein we present a sensing platform that integrates a fluorescence sensor array and deep learning algorithm for the quantitative screening of multiple PFASs in water. The approach leveraged the distinct quenching effects induced by different PFAS species on the fluorescence emission of individual array elements (i.e., fluorescent dyes). Through the feature interpretation of the information-rich three-dimensional fluorescence spectra using a residual neural network algorithm, the platform achieved simultaneous and comprehensive quantification of five types of PFASs in complex water samples. This novel strategy not only offers a facile and rapid method for multiple PFAS analysis but also expands the methodological boundaries of analytical sensing.
An et al. (Thu,) studied this question.