# Integration of Destructive and Non‐Destructive Analytical Determinations for Evaluating Quality of Fresh and Roasted Hazelnuts Subjected to Different Processing Temperatures

## TL;DR

This study shows that non-destructive tools like FT-NIR and E-nose can quickly and accurately assess the quality of hazelnuts during processing.

## Contribution

The integration of FT-NIR and E-nose with chemometric methods for real-time hazelnut quality monitoring is presented as a novel approach.

## Key findings

- FT-NIR accurately predicted moisture content and other quality parameters with high correlation values.
- E-nose effectively differentiated roasting intensities based on volatile profiles.
- Non-destructive methods provided results consistent with traditional chemical analyses.

## Abstract

The internal quality of hazelnuts (
Corylus avellana
L.), particularly in terms of the degradation of fat components, is widely recognized as a key factor in determining the appropriate type of industrial processing. Additionally, the internal composition and volatile profile of hazelnuts change significantly based on different roasting conditions. The here reported study investigates the efficiency of Electronic Nose (E‐nose) and Near‐Infrared Spectroscopy (FT‐NIR) technologies, combined with multivariate statistical techniques, for the rapid discrimination of hazelnuts subjected to different roasting conditions. Moreover, the study examines the ability of NIR to predict several key quality parameters in fresh and processed hazelnuts. Hazelnut samples were collected throughout the entire industrial processing chain, from delivery to roasting. The influence of three different roasting temperatures (140–150‐160°C) was evaluated, keeping the roasting time constant at 24 min. Partial Least Squares models were computed to estimate moisture content, total soluble solids, protein content, acidity, and peroxide index through correlation with FT‐NIR spectral data. Excellent regression performances were achieved for all quality parameters, except acidity, with correlations ranging between 0.951 and 0.918. Discriminant analysis models, specifically PLS‐DA and Cluster Analysis, were used to assess the ability to discriminate hazelnuts subjected to different roasting conditions using FT‐NIR and the Electronic Nose as non‐destructive tools. Obtained results from these non‐destructive techniques, particularly the volatile characterization GC/MS‐performed, accurately reflected the differentiation of samples observed through traditional chemical analyses, effectively distinguishing different groups of samples based on roasting temperature. The use of non‐destructive tools such as FT‐NIR and E‐nose during the post‐harvest life and processing of hazelnuts offers an excellent solution for monitoring key quality parameters significantly important for the food industry.

Non‐destructive techniques including FT‐NIR spectroscopy and E‐nose detection, combined with chemometric approaches, demonstrated strong potential for real‐time hazelnut quality assessment. FT‐NIR spectroscopy accurately predicted moisture content, while E‐nose differentiated roasting intensities. Their feasibility in process monitoring highlights a significant advantage over traditional, labor‐intensive methods, offering a rapid and efficient solution for quality control in the hazelnut industry.