Drug classification with a spectral barcode obtained with a smartphone Raman spectrometer

Area of Science:

• Spectroscopy
• Analytical Chemistry
• Materials Science

Background:

• Spectral information serves as a unique fingerprint for material identification.
• There is a growing demand for accessible, portable spectral analysis tools for both scientific and consumer use.

Purpose of the Study:

• To demonstrate the feasibility of drug classification by chemical components using a smartphone-based Raman spectrometer.
• To introduce and validate the concept of 'spectral barcodes' for material identification.

Main Methods:

• Development of a smartphone Raman spectrometer utilizing the device's CMOS image sensor and band-pass filters.
• Acquisition of 2D Raman spectral intensity maps, termed 'spectral barcodes'.
• Application of convolutional neural networks (CNNs) for data analysis and drug classification.

Main Results:

• Accurate classification of 11 major drug components with 99.0% accuracy.
• Demonstration that spectral barcodes can distinguish between drug brands and identify major components of unknown drugs.

Conclusions:

• Smartphone-based Raman spectroscopy, using spectral barcodes, offers a powerful and accurate method for drug classification.
• This technology has significant potential for fundamental research and commercial applications in material identification and labeling.