Determination of palm oil residue in palm kernel and palm oil methyl esters using near infrared spectroscopy

An analytical method by near infrared spectroscopy using a partial least square regression was developed for the determination of palm oil residue in palm kernel methyl esters (model 1) and palm oil methyl esters (model 2). the wavelength region selected for modeling was between 4500 and 10 000 cm -1. for model 1,the calibration plot had a coefficient correlation (r2) of 0.9996 and a standard error of calibration (sec) of 0.06. validation of the model 1 calibration produced a plot with r2 of 0.9962 and a standard error of prediction (sep) of 0.176. the model 2 calibration plot had r2 of 0.9986 and sec of 0.108. validation of model 2 produced a plot with r2 of 0.9978 and sep of 0.128. both models exhibited good linearity with low standard errors. the differences between the predicted and actual percentages of oil residue were minor. this method was able to predict oil residue contamination ranging from 1% to 10% in the two types of palm-based methyl esters.