This paper (open access) is an example of the growing trend of “data fusion” from multiple techniques to aid authenticity classification. Contributing techniques can be extremely simple. In this case the researchers including Differential Scanning Calorimetry (DSC), which is a routine way to measure the melting point of a solid (or the crystallisation point of an oil). They combined DSC with GC-MS profiling of fatty acid methyl esters (FAMEs) to detect adulteration of walnut oil.
The researchers prepared cold-pressed walnut oils in an normal manufacturing setting, in compliance with the Romania legal specification constraints. They produced binary mixtures adulterated with sunflower oil between 5 – 50 %.
They report that, as the proportion of adulterant oil increased, thermal parameters changed progressively. Samples with 40% and 50% sunflower oil exhibited low crystallisation enthalpy values, reflecting the dominant thermal characteristics of sunflower oil at higher substitution levels. Chemometric analysis demonstrated strong discrimination between authentic and adulterated oils, with support vector machine models achieving complete classification under the conditions of this study.
Photo by Pranjall Kumar on Unsplash
