Journal
MICROCHEMICAL JOURNAL
Volume 181, Issue -, Pages -Publisher
ELSEVIER
DOI: 10.1016/j.microc.2022.107746
Keywords
EDXRF; NIR; Milk quality control; Whey adulteration; Cow milk; Goat milk
Categories
Funding
- Fundacao de Amparo `a Pesquisa do Estado do Rio de Janeiro (FAPERJ) Brazil [E-26/200.891/2021, E-26/200.721/2021, E-26/200.358/2021]
- Conselho Nacional de Desenvolvimento Cientifico e Tecnologico (CNPq) [313119/2020-1, 163480/2020-6, 312595/2021-2, 310446/2020-1]
- Coordenacao de Aperfeicoamento de Pessoal de Nivel Superior (CAPES) Brazil [001]
Ask authors/readers for more resources
This study investigates the potential of low-cost portable spectroscopy techniques for discriminating adulterated milk. The results show that energy-dispersive X-ray fluorescence (EDXRF) combined with chemometric tools and near-infrared (NIR) spectroscopy can reliably differentiate authentic cow and goat milk from adulterated ones. Both devices achieved high accuracy rates of 98.4% to 100% when combined with appropriate algorithms.
Milk is one of the main foods present in the human diet. It is a good source of many essential nutrients; however, milk is among the most adulterated food globally. This study investigated the potential of low-cost portable spectroscopy near-infrared (NIR) and energy-dispersive X-ray fluorescence (EDXRF) spectrometry combined with chemometric tools to discriminate authentic cow and goat milk from adulterated ones with whey. The common dimension (ComDim) multi-block analysis enabled us to verify that the EDXRF data had good discrimination between the samples. In contrast, the NIR spectra had no evident discrimination and showed a nonlinear behavior in their data. Both devices had equivalent classification performances of 98.4 to 100% accuracy for the prediction when combined with the appropriate algorithms, such as partial least squares discriminant analysis (PLS-DA) for EDXRF and C-support vector classification (C-SVC) for NIR. Discriminant methods have some limitations for authentication tasks, so class modeling methods such as data driven soft independent modeling of class analogy (DD-SIMCA) were more helpful for this problem, with a prediction accuracy greater than 98.9% for both devices. These results suggest that NIR and EDXRF can be reliably applied to verify milk authenticity. Besides, these devices are portable, cost-effective, user-friendly, fast, robust, do not require extensive sample preparations, and can be easily adapted to desired settings.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available