Comparison of Vegetable Oils and Monitoring of Rancidity and Lard Adulteration in Palm Olein Using Electronic Nose

Flavour analysis is typically performed by organoleptic panel, which is often expensive and less objective. A novel approach using a surface acoustic wave (SAW) sensing-based electronic nose (zNose™) for flavour analysis was explored in this study for determination of some parameters of edible oils and fats. In the characterization of sixteen types of vegetable oils, the high resolution olfactory image, called VaporPrint™, was unique for each type of vegetable oil studied. The score plot from principal component analysis (PCA) indicated that 97% of the total variance in the zNose™ measurement data was described by PC 1 and PC 2. The loading plot revealed five compounds (m, k, n, s, and p) that were important to differentiate the vegetable oils. In determining oxidative stability of RBD palm olein, the results of zNose™ showed significant difference (P<0.05) between fresh oil and rancid oil. VaporPrint™ provided the operator with a visually recognised pattern for rapid identification of rancid off-flavour. By using Pearson correlation analysis, high correlation (r>0.90) was observed between electronic nose responses and chemical test data; as well as between electronic nose responses and sensory evaluation scores. The zNose™ technique was also employed to monitor the presence of lard as an adulterant in RBD palm olein. As the adulteration level increased from 1% to 20%, a few distinct peaks were found to gradually increase in size in the zNose™ chromatograms. These peaks dramatically increased in 3% lard, while the corresponding VaporPrint™ was obviously an exception to the normal RBD palm olein pattern. Qualitative identification of adulterated RBD palm olein samples was possible by the characteristic VaporPrint™. The most significant relationship occurred between percent lard and adulterant peak number 6 (R2 = 0.906). An ideal correlation was also observed between the electronic nose response and chemical analyses (r > 0.90).

Preview PDF 549726_FSMB_2004_20.pdf Download (108kB)

Actions (login required)

View Item