Analysis of Free Fatty Acids (FFA) in Palm Oils Based on the Raman Spectra

Abstract

Palm oil is a stable vegetable oil with various uses in both the food and non-food industries. The quality of edible palm oil is measured by its content of free fatty acids (FFA), which should be below 5% to be safely consumed and used, according to The Palm Oil Refiners Association of Malaysia (PORAM). Current titration methods require that the oil sample be dissolved in neutralized ethanol/diethyl, which is a long process that can damage the sample and cause delays and losses in the process. The focus of this study is to analyze the FFA in palm oils based on the Raman spectra of palm oil samples and oil palm fruitlets. Four samples of commercially available palm oils of different qualities, as well as 47 samples of oil palm fruitlets, were used in this study. The Raman spectra of all samples were collected via point scanning using a confocal microRaman imaging spectrometer with a 532 nm laser source. Manual spectrum pre-processing was performed with baseline correction and data filtering using Savitzky-Golay filters. Possible FFA peaks characteristics on the Raman spectrum were then identified through signal interpolation, signal deconvolution, and curve fitting. Statistical analysis was conducted on the extracted features to remove those with low statistical values, thus helping the machine learning algorithms to train the classifier models more efficiently. From this study, it was found that peaks at 1155 cm−1, 1525 cm−1, 1440 cm−1, and 1655 cm−1 from the Raman spectrum were highly correlated with the FFA levels of palm oils obtained through titration. The peak ratios of 1525 cm−1/1655 cm−1 and 1155 cm−1/1655cm−1 also showed a good and fast relationship for analyzing the FFA content of palm oil.