Toward Healthcare Diagnoses by Machine-Learning-Enabled Volatile Organic Compound Identification

Area of Science:

• Analytical Chemistry
• Spectroscopy
• Biomarker Detection

Background:

• Volatile organic compounds (VOCs) are crucial biomarkers for disease diagnosis and health monitoring.
• Current VOC sensors (semiconductor, optical, electrochemical) lack selectivity and efficiency, measuring only total concentration with slow responses.
• Traditional infrared (IR) spectroscopy is limited by weak light-matter interaction, requiring large optical paths for VOC detection.

Purpose of the Study:

• To develop a novel plasma-enhanced IR absorption spectroscopy technique for sensitive and selective VOC detection.
• To improve the light-matter interaction in IR spectroscopy for enhanced VOC analysis.
• To demonstrate the potential of this technique for healthcare monitoring and disease mimicry.

Main Methods:

• Utilized plasma-induced ultrahigh electric fields to enhance molecular vibrations and light-matter interaction.
• Employed a triboelectric nanogenerator for multiswitched manipulation, achieving kilovolt voltages.
• Applied IR absorption spectroscopy to quantify VOC species and concentrations based on spectral signals.

Main Results:

• Achieved fast response, accurate quantification, and good selectivity in VOC detection.
• Successfully quantified VOC species and concentrations using wavelength and intensity of spectral signals enhanced by plasma.
• Demonstrated the feasibility of VOC identification in mixtures using machine learning for potential patient mimicry.

Conclusions:

• Plasma-enhanced IR absorption spectroscopy offers significant advantages over traditional methods for VOC detection.
• The technique shows promise for accurate, selective, and rapid analysis of VOC biomarkers.
• Machine learning integration further enhances the potential for clinical applications in disease diagnosis.