New Analytical Screening Method for Fast Classification of Hemp Oil Based on THC Content

Area of Science:

• Analytical Chemistry
• Spectroscopy
• Machine Learning

Background:

• Accurate quantification of Δ9-tetrahydrocannabinol (THC) in hemp oil is crucial for regulatory compliance.
• Traditional analytical methods for THC detection can be time-consuming and require complex sample preparation.
• Developing rapid and efficient screening methods for THC content is an ongoing challenge.

Purpose of the Study:

• To present a novel analytical approach for classifying commercial hemp oil samples based on their THC content.
• To develop and validate machine learning models for rapid THC screening.
• To compare classification performance based on different regulatory thresholds.

Main Methods:

• Analysis of 204 commercial hemp oil samples using mid-infrared (MIR) spectroscopy.
• Application of Partial Least-Squares-Discriminant Analysis (PLS-DA) for classification.
• Development of two classification models based on regulatory thresholds (0.2% and 0.3% THC w/w).

Main Results:

• Both classification models achieved high accuracy (>88.50%).
• Model B (threshold >0.3% THC) demonstrated improved sensitivity (STR) compared to Model A (threshold >0.2% THC), particularly for the test set (95.00% vs 77.27%).
• The MIR spectroscopy and machine learning approach provided accurate classification with simplified sample handling.

Conclusions:

• The developed analytical approach offers a viable, rapid, and simple alternative to conventional laboratory methods for THC screening in hemp oil.
• This method facilitates efficient compliance with varying international regulatory standards for hemp products.
• The combination of MIR spectroscopy and PLS-DA provides a powerful tool for quality control and regulatory monitoring in the hemp industry.