Determination of the adulteration of benzyl alcohol and Dipropylene glycol in sage, mint, and oregano essential oils belonging to the Lamiaceae Family by FTIR using Chemometry and 2D-COS

Spectrochimica Acta Part A Molecular and Biomolecular Spectroscopy · Mart 2026

Makale Bilgileri

DergiSpectrochimica Acta Part A Molecular and Biomolecular Spectroscopy

Yayın TarihiMart 2026

Cilt / Sayfa348

DOI10.1016/j.saa.2025.127085

Scopus ID2-s2.0-105020973518

Özet In the presented study, fast and easy Fourier transform infrared (FTIR) methods were tried to be developed to detect benzyl alcohol (B), and dipropylene glycol (D) adulteration in the range of 1 to 12 % in essential oils (EOs) of sage (SEO), mint (MEO), and oregano (OEO), which belong to the Lamiaceae family and have high commercial value. The potential of two different infrared spectral acquisition modes, attenuated total reflection (ATR) mode FTIR (AF) and transmittance mode FTIR (TF) to identify and quantitatively determine the adulterant was assessed by employing multivariate spectral analysis techniques. Two dimensional-correlation spectroscopy (2D-COS) was shown to be a useful approach for monitoring the changes in SEO, MEO, and OEO due to B and D adulteration and is useful in selecting the spectral areas to be used for the principal component analysis (PCA) and partial least squares (PLS). PCA modeling performed on selected EOs at the studied concentrations effectively identified B and D. According to the PCA modeling, B was successfully detected in all EO samples using the = CH rocking corresponding to 760–650 cm−1 by TF. Qualitative analysis was performed for the detection of D using the = CH rocking corresponding to 846–650 cm−1 by TF, the C-O-C ether bands attributed to 1180–920 cm−1 by AF, and the C-O-C ether bands corresponding to 1150–1010 cm−1 by TF for SEO, MEO, and OEO, respectively. The cumulative variance explained by PC1 + PC2 of the obtained PCA models vary between 88 and 99 %. In the PLS models developed for the quantitative analysis of B, =CH rocking corresponding to 760–650 cm−1 was used for all EOs and successful models were demonstrated with root mean square error of prediction (RMSEP) values of 0.5308–0.6788 by AF, 1.0753–0.8123 by AF, and 1.2536 by TF for SEO, MEO, and OEO, respectively. For the quantitative analysis of D, PLS models were developed with C-O-C ether bands corresponding to 1152–986 cm-1 by AF and 1167–1015, 1014–876, 846–650 cm-1 by TF for groups 1 and 2 of SEO, respectively. A similar grouping strategy was successful for the quantitative analysis of D in MEO, while no grouping strategy was used for the analysis in OEO. It was concluded that a better quantitative prediction could be achieved for the B and D they contain when a pre-grouping strategy was determined by paying attention to the biochemical compounds alkane-pinene, and linalyl acetate for SEO and alkane, and menthol for MEO.