Engine oil authentication using near infrared spectroscopy and chemometrics methods

Abstract

Visible/Near Infra-Red Spectroscopy (Vis/NIRS) was used in association with multivariate chemometrics techniques to assess high quality engine oils samples adulterated with lower quality ones. Absorbance spectra were obtained in spectral range from 500 to 1000 nm and from 1000 to 1700 nm. Spectral data were analyzed to establish accurate model allowing the detection and the prediction of the amount of lower quality engine oil added to high quality one. The data set were analyzed using two multivariate analysis methods: Partial Least Square Regression (PLSR) and Principal Component Analysis (PCA). PCA showed a clear separation between three groups of adulterated engine oil samples and the best PLSR model was obtained with a regression coefficient (R) of 0.90, Root Mean Square Error of Prediction (RMSEP) of 4.51 and Standard Error of Prediction (SEP) of 4.55.