Monte Carlo simulation based prediction of blended oil composition containing mustard, rapeseed and soybean oil

Abstract

High erucic acid rapeseed and mustard seed (HEARM) oils are popular cooking oil in various southeast Asian countries (e.g. Nepal, India, Bangladesh) due to the preference of the local consumer for its strong pungency. Low erucic acid rapeseed (LEAR) oil having lower pungency is frequently mixed with the pungent HEARM oil to increase its sensory appeal to the local consumer. Moreover, these oils are also prone to be adulterated with the cheaper oil available in the local market like soybean oil for economical reasons. In order to detect this fraud in a quick and easy manner, a Monte Carlo simulation based approach was developed for the estimation of blend composition using only the fatty acid composition of the sample. The limits of detection (LODs) of soybean oil in HEARM oil and LEAR oil were 14% and 13%, respectively. The LODs of LEAR oil in HEARM oil and soybean oil were 11 and 9%, respectively. Similarly, the LODs of HEARM oil in LEAR oil and soybean oil were 9% and 3%, respectively. The prediction from the developed method was evaluated both in real oil blends (prepared in the laboratory) and in theoretically simulated blends. The method was applied on forty-nine samples (labeled as mustard/rapeseed oil) collected from the Nepalese market. Among them, twenty-seven samples were found to be adulterated with soybean oil. The predicted adulteration was further supported by their δ-tocopherol content and trans fatty acid content, as an indicator for the adulteration with refined oil. The developed Monte Carlo simulation method is based on a single analytical run of determining the fatty acid composition of the suspected oil blend and thus useful for a quick segregation of samples in routine analysis.