Deep-Learning-Assisted Fluorescence Sensor Array for Quantitative Screening of Perfluoroalkyl Substances in Water

Area of Science:

• Environmental Chemistry
• Analytical Chemistry
• Sensor Technology

Background:

• Perfluoroalkyl substances (PFASs) pose significant environmental and health risks.
• Conventional methods for PFAS detection are limited, necessitating advanced analytical platforms.
• Simultaneous quantification of multiple PFASs in complex samples is challenging.

Purpose of the Study:

• To develop an efficient sensing platform for the simultaneous quantification of multiple PFASs in water.
• To overcome the limitations of existing methods for PFAS detection.
• To provide an easy-to-perform analytical solution for complex water samples.

Main Methods:

• Integration of a fluorescence sensor array with a deep learning algorithm.
• Utilizing distinct fluorescence quenching effects of different PFAS species on fluorescent dyes.
• Employing a residual neural network for feature interpretation of 3D fluorescence spectra.

Main Results:

• Achieved simultaneous and comprehensive quantification of five types of PFASs.
• Demonstrated effective analysis in complex water samples.
• Validated the platform's ability to interpret information-rich fluorescence spectra.

Conclusions:

• The developed platform offers a facile and rapid method for multiple PFAS analysis.
• This novel strategy expands the methodological boundaries of analytical sensing for environmental contaminants.
• The approach provides a promising solution for monitoring PFAS contamination in water sources.