Prediction of acid peroxide and TBA values of heat-treated palm oil using a partial least squares“ordinary least squares model based on fourier transform infrared spectroscopy


Citation

Anggraeni R., . and Lioe H. N., . and Pricilia, . and Kusnandar F., . and Faridah D. N., . Prediction of acid peroxide and TBA values of heat-treated palm oil using a partial least squares“ordinary least squares model based on fourier transform infrared spectroscopy. pp. 514-523. ISSN 2811-4701

Abstract

Palm oil is widely used for frying food and is often used for repeated frying up to 40 hr or even longer. Frying causes a gradual quality decrease during heating due to fat oxidation or hydrolysis. The quality of fats and oils is commonly monitored by acid peroxide and thiobarbituric acid values (AV PV and TBAV respectively). This study aimed to use a partial least squares“ordinary least squares (PLS-OLS) model obtained from Fourier transform infrared (FTIR) spectroscopy to predict AV PV and TBAV values of heat-treated palm oil. Commercial palm oil was heated at 180C for 72 hr. The multivariate mathematical models to predict AV PV and TBAV were generated from the percentages of absorbance intensity of significant wavenumbers based on FTIR readings (721.4 871.8 968.3 1033.9 1095.6 1377.2 1462 1751.4 2731.2 2839.2 and 3005.1 cm“). The PLS-OLS mathematical model satisfactorily predicted both AV and PV for palm oil samples heated up to 72 hr (R0.962 and 0.857 respectively) whereas for TBAV the time was 58 hr (R0.845). This approach provides an alternative to monitoring palm oil quality during frying instead of the conventional methods in which the analytical procedures are time-consuming.


Download File

Full text available from:

Abstract

Palm oil is widely used for frying food and is often used for repeated frying up to 40 hr or even longer. Frying causes a gradual quality decrease during heating due to fat oxidation or hydrolysis. The quality of fats and oils is commonly monitored by acid peroxide and thiobarbituric acid values (AV PV and TBAV respectively). This study aimed to use a partial least squares“ordinary least squares (PLS-OLS) model obtained from Fourier transform infrared (FTIR) spectroscopy to predict AV PV and TBAV values of heat-treated palm oil. Commercial palm oil was heated at 180C for 72 hr. The multivariate mathematical models to predict AV PV and TBAV were generated from the percentages of absorbance intensity of significant wavenumbers based on FTIR readings (721.4 871.8 968.3 1033.9 1095.6 1377.2 1462 1751.4 2731.2 2839.2 and 3005.1 cm“). The PLS-OLS mathematical model satisfactorily predicted both AV and PV for palm oil samples heated up to 72 hr (R0.962 and 0.857 respectively) whereas for TBAV the time was 58 hr (R0.845). This approach provides an alternative to monitoring palm oil quality during frying instead of the conventional methods in which the analytical procedures are time-consuming.

Additional Metadata

[error in script]
Item Type: Article
AGROVOC Term: Palm oils
AGROVOC Term: Peroxidation
AGROVOC Term: Analytical chemistry
AGROVOC Term: Spectroscopy
AGROVOC Term: Quality controls
AGROVOC Term: Food processing
AGROVOC Term: Data analysis
AGROVOC Term: Optimization methods
AGROVOC Term: Laboratory experimentation
AGROVOC Term: Heat treatment
Depositing User: Mr. AFANDI ABDUL MALEK
Last Modified: 24 Apr 2025 00:55
URI: http://webagris.upm.edu.my/id/eprint/10260

Actions (login required)

View Item View Item