Use of temperature dependent Raman spectra to improve accuracy for analysis of complex oil-based samples: Lube base oils and adulterated olive oils

Abstract

A simple and effective strategy to improve accuracy for Raman spectroscopic analysis of complex mixture samples by probing a measurement temperature yielding enhanced spectral selectivity has been demonstrated. For the evaluation, the determination of Kinematic Viscosity at 40°C (KV@40) of lube base oil (LBO) samples was initially attempted. Partial least squares (PLS) was used to determine the KV@40 using Raman spectra of the samples collected at 8 different temperatures from 20 to 90°C with 10°C increments. Interestingly, the distinct temperature-induced spectral variation among the samples occurred at 50°C, thereby resulting in the improved accuracy for determination of KV@40. Two-dimensional (2D) correlation analysis was also performed to find an additional supportive rationale for the improved accuracy. The strategy was further evaluated for the identification of soybean oil-adulterated olive oils using linear discriminant analysis (LDA). Similarly, the discrimination accuracy was improved around 80–90°C due to the enhanced spectral selectivity between olive and soybean oils. In overall, these two results successfully demonstrate analytical effectiveness of the strategy.