Assessing the Feasibility of Mid- and Near-Infrared Spectroscopy for Classifying Plant-Based Beverages and Predicting...

Area of Science:

• Food Science
• Analytical Chemistry
• Spectroscopy

Background:

• Rising demand for dairy alternatives necessitates efficient methods for plant-based beverage analysis.
• Assessing composition and authenticity of plant-based beverages is crucial for quality control.

Purpose of the Study:

• To evaluate mid-infrared (MIR) and near-infrared (NIR) spectroscopy for classifying plant-based beverages.
• To predict the nutritional profile of plant-based beverages using spectroscopic data.

Main Methods:

• Analysis of 57 commercial plant-based beverages across five categories (oat, almond, soybean, rice, coconut).
• Application of canonical discriminant analysis for beverage category classification.
• Development of modified partial least-squares regression models using spectral and chemical reference data.

Main Results:

• Both MIR and NIR spectroscopy effectively discriminated among the five beverage categories.
• MIR spectroscopy generally provided more accurate quantitative predictions than NIR.
• Robust models were developed for predicting protein, fat, glucose, ash, and key minerals (P, K), with protein showing excellent accuracy via both methods.
• Amino acid composition was well-predicted, especially using MIR, with satisfactory performance for most amino acids.

Conclusions:

• Infrared spectroscopy is feasible for authenticating and assessing the composition of plant-based beverages.
• MIR spectroscopy offers significant potential for integration into routine quality control workflows for plant-based beverages.
• Spectroscopic methods can complement traditional chemical analyses for rapid assessment.