Moh, Felix Mee Ho (1998) Development of Infrared Spectroscopic Techniques for the Determination of Some Quality Parameters of Palm Oil. Masters thesis, Universiti Pertanian Malaysia.
Several parameters for good quality crude palm oil (CPO) and its fractions such as free fatty acid (FFA), oxidative value, moisture and impurities, iodine value and antioxidant composition are very crucial in palm oil industry. For trading purposes in Malaysia, the standard of these quality parameters have to be accomplished by all the refiners Association of Malaysia (PORAM) standard specifications. Conventionally, the analytical techniques used to analyze the quality parameters in palm oil are by wet chemical methods based on the American Oil Chemists' Society (AOCS) procedures which involve titration. Normally this chemical techniques require large amount of chemicals and reagents which may be unfriendly to environment and hazardous to health. Moreover, most analyses are time consuming and laborious especially for large number of samples. To overcome the problem, numerous studies on instrumental techniques have been carried out over the years. From this standpoint, near infrared (NIR) spectroscopy and Fourier transform infrared (FTIR) spectroscopy are considered the instrumental approaches of choice. 1bree parameters were selected in the study. For the determination of FFA, NIR transflectance and FTIR transmittance techniques were investigated. Range of FF A concentration were prepared by hydrolyzing oil with 0.15% (w/w) lipase in an incubator at 60°C (200 rpm). In NIR approach, sample preparation was performed in Dutch cup and optimized calibration models were constructed with multiple linear regression (MLR) analysis based on C=O overtone regions. The best wavelength combinations were found to be 1882,2010 and 2040 nm for CPO and refined bleached deodorized (RBD) oils. For FTIR technique, partial least squares (PLS) analysis was used to replace the previous simple linear regression model developed for predicting FFA. Both IR techniques showed excellent results in predicting the FFA values of the unknown samples. The determination of peroxide value (PV) was based on NIR transmittance spectroscopy. Calibration standards (2 to 11 PV) were prepared by oxidized CPO at 90°C. The best NIR region for predicting unknown was from 1350 to 1480 nm with referenced to single-point baseline at 1514 nm. The standard error of performance (SEP) was similar to the calibration to within 0.1% PV. The final aspect studied was the determination of β-carotene in CPO using different techniques. Both FTIR and NIR spectroscopy were compared with the UV-visible spectrophotometric reference method. PLS calibration model for predicting β-carotene was developed for the FTIR spectral region from 976 to 926 cm⁻¹] due to trans -CH=CH- conjugated deformation modes. The study showed that NIR spectra did not correlate well with the component of interest. However, highest correlation was found in the visible spectral region from 546 to 819 nm. Statistical results showed that both FTIR and NIR spectroscopy can accurately predict β-carotene to 5% error; this value was 4% higher as compared to the reference method. Results from this study indicated that both NIR and FTIR spectroscopy were useful techniques for measuring palm oil quality parameters. Compared to the conventional wet chemical method, NIR and FTIR spectroscopy are capable of measuring hundreds of samples in a day and yet meet the trading specification. In addition, by applying these techniques, the amount of hazardous solvents can be reduced dramatically as well as the cost of labor.
|Item Type:||Thesis (Masters)|
|Subject:||Infrared spectroscopy - Malaysia.|
|Subject:||Palm oil - Quality - Malaysia.|
|Chairman Supervisor:||Professor Yaakob Bin Che Man, PhD|
|Call Number:||FSMB 1998 5|
|Faculty or Institute:||Faculty of Food Science and Technology|
|Deposited By:||Nurul Hayatie Hashim|
|Deposited On:||15 Nov 2010 02:58|
|Last Modified:||02 Jul 2012 07:13|
Repository Staff Only: item control page
Document Download Statistics
This item has been downloaded for since 15 Nov 2010 02:58.