A novel optically gated thin-film transistor sensor for real-time chemical differentiation using machine-learning...

Area of Science:

• Materials Science and Engineering
• Sensor Technology
• Analytical Chemistry

Background:

• Existing chemical sensors often require reference electrodes or surface functionalization, limiting their simplicity and portability.
• Optically gated devices offer potential for novel sensing mechanisms based on light-matter interactions.

Purpose of the Study:

• To develop and demonstrate an optically gated thin-film transistor sensor for real-time chemical differentiation.
• To investigate the use of illumination-induced transient electrical responses for chemical signature analysis.
• To apply machine learning for classifying chemicals based on sensor output.

Main Methods:

• Fabrication of a thin-film transistor using a p-type silicon substrate and an amorphous Ge2Se3 photogating layer.
• Application of structured light-pulse sequences to induce time-dependent electrical responses.
• Analysis of transient electrical signatures using supervised machine-learning models.

Main Results:

• The sensor successfully distinguished between different chemical environments based on unique electrical signatures.
• Machine learning models achieved high classification accuracies (95-99%) for alcohols and perfluoroalkyl substances.
• A classifier for total organic fluorine demonstrated 94% accuracy.

Conclusions:

• An optically gated silicon-based transistor provides a compact and generalizable sensor architecture.
• The device enables real-time chemical identification through illumination-induced electrical transients.
• This approach offers a promising platform for developing advanced chemical sensing applications.