Brown rice authenticity evaluation by spark discharge-laser-induced breakdown spectroscopy

Area of Science:

• Analytical Chemistry
• Spectroscopy
• Food Science

Background:

• Rice is a global staple food, making its designation of origin (PDO) crucial for quality assurance.
• Laser-induced breakdown spectroscopy (LIBS) offers rapid, multi-elemental analysis with minimal sample preparation, ideal for food authentication.

Purpose of the Study:

• To evaluate the efficacy of LIBS for the PDO certification of Argentine brown rice.
• To explore the application of machine learning algorithms in spectral data selection for LIBS analysis in food authentication.

Main Methods:

• Analysis of rice samples from two PDOs using LIBS coupled with spark discharge.
• Spectral data selection utilizing the extreme gradient boosting (XGBoost) algorithm.
• Classification of samples using the k-nearest neighbor (k-NN) algorithm.

Main Results:

• Key elemental emission lines (C, Ca, Fe, Mg, Na) were identified for classification.
• The developed LIBS method achieved 84% accuracy, 100% sensitivity, and 78% specificity.
• The k-NN algorithm demonstrated superior performance in classifying rice samples based on LIBS data.

Conclusions:

• LIBS is a viable, simple, and clean technique for the PDO certification of Argentine brown rice.
• The integration of machine learning (XGBoost) enhances spectral data selection for LIBS applications.
• This method provides reliable classification for ensuring the authenticity and origin of rice products.