Infrared Spectroscopy for Variety Identification and Authenticity Analysis of Tobacco Samples

Area of Science:

• Analytical Chemistry
• Spectroscopy
• Forensic Science

Background:

• Authenticity checking of tobacco products is crucial for regulatory compliance and consumer protection.
• Current detection methods for illegal tobacco rely on package analysis, administrative verification, and laboratory testing.
• There is a need for efficient and reliable methods to identify tobacco varieties in blends.

Purpose of the Study:

• To propose and evaluate an Infrared (IR) spectroscopy approach for identifying tobacco varieties in blends.
• To compare the performance of Mid-Infrared (MID-IR) and Near-Infrared (NIR) spectroscopy for tobacco variety identification.
• To develop quantitative models for estimating the relative amounts of tobacco varieties in mixtures.

Main Methods:

• Preparation of various tobacco blends.
• Acquisition of spectra using MID-IR and NIR spectroscopy.
• Development and validation of Partial Least Squares Discriminant Analysis (PLS-DA) models.
• Evaluation of model performance using cross-validation and an external test set.
• Estimation of the limit of detection (LOD) and development of quantitative models.

Main Results:

• NIR spectroscopy outperformed MID-IR, achieving high classification rates, precision, specificity, and accuracy for most varieties.
• Indicative LOD values ranged from 1% to 8% for the best models.
• Quantitative models demonstrated errors smaller than 5% for estimating variety proportions.
• The approach was successfully applied to five commercial tobacco blends, proving its feasibility.

Conclusions:

• IR spectroscopy, particularly NIR, is a viable and effective tool for identifying tobacco varieties in blends.
• The developed method shows significant potential in combating tobacco fraud and counterfeit products.
• This spectroscopic approach offers a valuable addition to existing authenticity checking strategies.