Description
Engine (or motor) oil is a lubricant used in internal combustion engines, produced by the blending of 80% (w/w) base oil and 20% (w/w) of different additives1. The American Petroleum Institute (API) has categorized base oils into five categories2; the first three groups are mineral stocks, refined from crude oil with different processes, Group IV are full synthetic (polyalphaolefin) and Group V is for all other base oils not included in Groups I through IV. To date, it is possible to identify the base oil of a lubricant only by looking at the combination of physical properties (such as viscosity index, density, colour, flash point, pour point, aniline point, thermal stability) but the measurement of these parameters is expensive and time consuming. The aim of the present study was to investigate, for the first time, the capability of near infrared (NIR) spectroscopy as low-cost, green and non-destructive method in order to identify the type of base oil into engine oil. In order to reach this goal, 53 pure base oils belonging to different API groups and 43 engine oils were analysed without any pre-treatments.
NIR spectra were acquired in transmittance mode with a FT-NIR spectrophotometer (Buchi NIRFlex N-500) in the 4000-10000 cm-1 range and 4 cm-1 resolution, using a quartz cuvette with 10 mm pathlength.
Principal component analysis performed on the NIR spectra (as the average of 3 spectra) showed that samples form clusters according to their API groups and so to their chemical composition. Thanks to the interesting results obtained in PCA; PLS-DA, as a multivariate classification tool, was applied in order to distinguish among different API groups and satisfactory results were achieved: the prediction abilities in the external test set composed of 20% of samples randomly selected was 87%.
| Consider for full paper in JNIRS | No, thank you |
|---|
