Abstract
Food fraud is an economically motivated concept that has occurred within the food supply system since trading began. It has been defined as the deliberate and intentional substitution, addition, tampering or misrepresentation of food, food ingredients and food packaging for an economic gain. It is in many cases close to impossible for consumers and players in the food industry to identify. The aim of this thesis was to study whether mass spectrometric platforms, especially ambient mass spectrometry (AMS) techniques, coupled with chemometric modelling could play a prominent role in the detection of food fraud.Firstly, a two-tier approach of Fourier-Transform infrared (FT-IR) spectroscopy and liquid chromatography-high resolution mass spectrometry (LC-HRMS) were used to analyse oregano samples and five potential adulterants (olive leaves, myrtle leaves, cistus leaves, sumac leaves and hazelnut leaves) thought to be used as bulking agents. LC-HRMS detected adulteration of oregano samples through biomarker identification which was achieved using chemometrics. Both analytical techniques were applied to seventy-eight commercially available samples obtained both within and outside the UK/Ireland. There was 100% agreement between the two tests which revealed that 24% of all samples analysed had some form of adulterants present with olive and myrtle leaves being the most commonly found adulterants.
Secondly, rapid evaporative ionisation mass spectrometry (REIMS) was used to determine the feasibility of fish species identification. Five white fish species (cod, coley, haddock, pollock and whiting) were analysed using an electrosurgical knife coupled to a quadrupole time-of-flight mass spectrometer (QTof). Principal component analysis (PCA) and linear discriminant analysis (LDA) models were generated exhibiting clear differences between the five species of fish. They were exported to a recognition software and used as a reference point allowing raw data from a sample unknown to the models to be assigned a species classification near-instantaneously (≈2s). A 98.99% correct classification of ninety-nine validation samples identified that REIMS is capable of both rapid and accurate results. Equally important, the analysis of six suspected mislabelled ‘haddock’ samples were undertaken. Results from REIMS for all six samples was available within fifteen minutes whereas it took twenty-four hours using polymerase chain reaction (PCR), a genomic profiling technique commonly used for such studies.
In a further study, the REIMS technology was applied to four meat species (beef, goat, lamb and pork) to determine the quantitative abilities of the technology. As with the fish study, both PCA and LDA models showed clear differences between the four species. The models were exported to a recognition software to analyse adulterated beef burgers. Adulteration of beef burgers with goat was detectable at levels of 2% adulteration, whilst pork and lamb were detected at 5% and 10% respectively. However, the preparation of burgers made through a serial dilution process impacted the quantitative abilities of the REIMS technology with limits of detection (LOD) for each adulterant being higher compared to those not made through serial dilution. Chemometric analysis of the four-meat species did not result in the identification of unique species-specific markers. However, ions found to occur at more abundant levels in certain species were. They were identified as phospholipids with five different species being assigned; phosphatidic acid (PA), phosphatidylcholine (PC), phosphatidylethanolamine (PE), phosphatidylinositol (PI) and phosphatidylserine (PS). Thus, within the PhD project the potential for ambient mass spectrometry to deliver very rapid and reliable detection of food fraud has been demonstrated.
| Date of Award | Sept 2017 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Sponsors | Waters Corporation & Biotechnology & Biological Sciences Research Council |
| Supervisor | David Kennedy (Supervisor) & Christopher Elliott (Supervisor) |