Evaluation of Multivariate Calibration with RADIAN ASAP Data for the Quantitation of Edible Oil Blends

Importance of the Topic

The blending of edible oils is essential to enhance functional properties, stability, and sensory characteristics in food products. However, it also creates opportunities for adulteration of higher-value oils with cheaper alternatives. Rapid, accurate analytical methods are therefore vital for quality control and authenticity verification in the food industry.

Objectives and Study Overview

This study investigates the use of RADIAN ASAP direct mass spectrometry in combination with multivariate calibration to quantify the proportions of extra virgin coconut oil and soybean oil in binary blends. An orthogonal partial least squares (OPLS) regression model was developed using a comprehensive training set and validated with blind samples.

Methodology and Instrumentation

RADIAN ASAP employs ambient solids analysis probe ionization coupled to a compact mass spectrometer, eliminating the need for sample preparation or chromatographic separation. Data quality was assessed by principal component analysis (PCA) before building an OPLS model with one predictive and two orthogonal components. The training set included 130 samples across six blend ratios, and 57 independent measurements were used for validation.

Main Results and Discussion

PCA separated coconut and soybean oil samples, capturing 98.6% of variance in the first three components. The OPLS model achieved R2Cum = 0.986 and Q2Cum = 0.97, with a strong linear calibration curve (R2 = 0.9853) between actual and predicted blend percentages. Validation demonstrated mean prediction errors within ±5% of true values. Loadings analysis identified discriminating m/z regions corresponding to diacylglycerols and triacylglycerols characteristic of each oil type.

Benefits and Practical Applications

• Instant, preparation-free analysis for real-time quality control.
• Accurate quantitation of oil blends within a ±5% error margin.
• Simple implementation with minimal instrument footprint and user training.

Future Trends and Possibilities

• Coupling with high-resolution MS for definitive lipid identification.
• Extension to diverse oil matrices and complex food formulations.
• Incorporation of automated and AI-based data analytics for enhanced predictive accuracy.

Conclusion

Combining RADIAN ASAP data with multivariate calibration delivers a fast, reliable, and user-friendly approach for quantifying edible oil blends. The validated OPLS model demonstrates robustness and high accuracy, supporting its adoption for routine quality assurance and authentication workflows.

References

1. Senanayake SPJN, Shahidi F. Modification of Fats and Oils via Chemical and Enzymatic Methods. Bailey’s Industrial Oil and Fat Products. 2005.
2. Srivastava Y, Semwal AD, Sajeevkumar VA, Sharma GK. Melting, Crystallization and Storage Stability of Virgin Coconut Oil and Its Blends by DSC and FTIR. J Food Sci Technol. 2017;54(1):45–54.
3. Bro R. Multivariate Calibration: What Is in Chemometrics for the Analytical Chemist? Anal Chim Acta. 2003;500(1):185–94.
4. Forina M, Lanteri S, Casale M. Multivariate Calibration. J Chromatogr A. 2007;1158(1–2):61–93.
5. Ferreira JA, Santos JM, Breitkreitz MC, et al. Characterization of the Lipid Profile From Coconut Oil by ESI-MS with PCA and ICA. Food Res Int. 2019;123:189–97.
6. Bataglion GA, Santos JM, da Silva FMA, et al. Comprehensive Characterization of Lipids From Amazonian Vegetable Oils by MS. Food Res Int. 2014;64:472–81.
7. Pizzo JS, Galuch MB, Manin LP, et al. Direct Infusion ESI-MS for Detection of Coconut Oil Adulteration With Palm Kernel Oil. Food Addit Contam Part A. 2019;36(11):1597–604.