Abstract
Food authentication serves a crucial role in the food industry, as it ensures safe consumption
of food and may prevent sicknesses. Food authentication has been a growing concern and
techniques that enable its conduction are being developed by researchers and scientists. For
this study, three analytical techniques, i.e., GC-FID, TGA and FTIR, were employed in the
authentication of various edible oils, including the oil blends. The triglycerides in oils were
converted to fatty acid methyl esters through two transesterification methods, i.e., the
automated flow and manual batch methods, which were conducted using two catalyst
concentrations, i.e., 5.0 g/L and 12.5 g/L CH3ONa. Statistical analysis was performed and the
results revealed that reactions conducted in an automated flow system using a 12.5 g/L
CH3ONa catalyst produced the best outcomes. A new screening approach was developed and
the peak area ratios were used to construct boxplots and the results showed EVOO, avocado
and macadamia oil to be outliers, meaning that these oils had values that were significantly
different from other oils in question. This approach was then proposed to be a possible way
of identifying unknown oils. Furthermore, thermal analysis was also performed to evaluate
the thermal stability of the oils and the results revealed that avocado oil was the most
thermally stable oil compared to the other oils. Finally, spectroscopic analysis was performed
to investigate the functional groups of the oils. The fingerprint region was used in this case
and the results showed no significant difference between the oils. It is suggested that
employing chemometrics for future work could be useful in distinguishing between the oils.
Overall, this study employed the automation of sample preparation, e.g., transesterification
of edible oils, which was achieved by incorporating 4IR related technology. From the thermal
analysis outcomes, interesting synergistic and antagonistic effects were observed in terms of
the thermal properties of the oil blends.