Almond analysis with near-infrared (NIR) spectroscopy

Importance of the Topic

Almonds are a nutritious food commodity whose quality attributes—moisture, protein, and fat—are critical for processing and final product consistency. Conventional assays involve destructive sample preparation, solvents, and lengthy protocols, limiting throughput and increasing operational costs. Near-infrared (NIR) spectroscopy emerges as a rapid, non-destructive alternative for simultaneous multiparameter assessment, streamlining quality control in the almond industry.

Objectives and Study Overview

This study aims to establish and validate NIR calibration models for quantifying moisture content, protein, and fat in whole and ground almonds. By comparing NIR predictions with standard reference methods, the work evaluates accuracy, precision, and suitability for routine analysis during different processing stages.

Methodology

A total of 120 samples (60 whole, 60 ground) were scanned on a Metrohm NIR Analyzer in reflectance mode over 1000–2250 nm. Rotational sampling and spectral averaging minimized heterogeneity. Reference analyses followed AOAC protocols: Kjeldahl digestion for protein, loss on drying for moisture, and Soxhlet extraction for fat. Chemometric models were developed and validated using correlation coefficients (R²), standard error of calibration (SEC), and standard error of cross-validation (SECV).

Instrumentation Used

• Metrohm NIR Analyzer equipped with a large cup accessory
• Wavelength range: 1000–2250 nm in reflectance mode
• Rotating sample holder for enhanced spectral reproducibility
• Metrohm software for data acquisition and multivariate model development

Main Results and Discussion

High correlation between NIR predictions and reference values was achieved:

• Whole almonds: Protein R² = 0.997, SEC = 0.23 %, SECV = 0.28 %; Moisture R² = 0.998, SEC = 0.22 %, SECV = 0.25 %; Fat R² = 0.917, SEC = 1.56 %, SECV = 1.84 %
• Ground almonds: Protein R² = 0.999, SEC = 0.10 %, SECV = 0.14 %; Moisture R² = 0.999, SEC = 0.09 %, SECV = 0.15 %; Fat R² = 0.914, SEC = 1.62 %, SECV = 1.67 %

This performance indicates excellent predictive power for moisture and protein and acceptable accuracy for fat, demonstrating robust model applicability.

Benefits and Practical Applications

NIR spectroscopy provides rapid (seconds per sample), reagent-free analysis, enabling simultaneous measurement of key quality parameters without sample destruction. Implementation supports inline or at-line monitoring during raw material intake, process optimization, and final quality inspection, reducing costs and accelerating decision-making.

Future Trends and Applications

Advancements in sensor integration will facilitate continuous, real-time quality control within processing lines. Further improvements in chemometric techniques and expanded spectral libraries covering diverse almond varieties will bolster model robustness. The approach can be extended to other nuts and food matrices, promoting holistic, non-destructive quality assessment across the food industry.

Conclusion

The application note demonstrates that NIR spectroscopy is a fast, accurate, and cost-effective alternative to traditional wet chemistry for almond analysis. High correlation metrics confirm its suitability for both whole and ground samples, offering significant benefits for industrial quality control workflows.

Reference

• Duduzile Buthelezi, N. M.; Tesfay, S. Z.; Ncama, K.; et al. Destructive and Non-Destructive Techniques Used for Quality Evaluation of Nuts: A Review. Scientia Horticulturae 2019, 247, 138–146. DOI:10.1016/j.scienta.2018.12.008