False negatives in food: High resolution mass spectrometry sorts analytes from interferents
- Published: Dec 1, 2011
- Author: Steve Down
- Channels: Base Peak
Photoinitiator migration and measurement
The food that we eat is under constant threat from unwanted contaminants like pesticides, fungicides, veterinary drugs and environmental pollutants in the soil, water or air, to name a few. Even people who eat organic food can be exposed to external compounds which migrate from the packaging.
It was established recently that the components of the inks used on packaging materials have been found in the food that they were holding. It occurs with paper, cardboard, plastic and laminates and one of the associated contaminants is benzophenone, added as a photoinitiator to kick start the printing process.
Unfortunately, most of the added benzophenone remains unreacted so has the potential to migrate to the food. The degree of transfer depends on the type and complexity of the food, as well as the type of packaging.
The EU has set a specific migration limit for benzophenone of 600 µg/L in liquids such as milk and fruit juices and there is an official toxicological tolerable daily intake of 0.01 mg/kg bodyweight/day set by the European Commission Scientific Committee on Food. So, reliable analytical methods must be in place so that organisations can adhere to these limits.
The preferred method of choice for detecting and measuring benzophenone in foods is liquid chromatography-tandem mass spectrometry in selected reaction monitoring mode. Analysts monitor two ion transitions, the first for quantitation and the second for confirmation. The deviations of the relative intensities of these ions, along with the retention time of benzophenone, must be within certain values for confident reporting.
Despite the good performance of this method, there have been some cases where false positives and false negatives have been recorded, probably due to the presence of an interfering compound. In one instance, a team of scientists in Spain could not identify benzophenone unambiguously in packaged foods even though the two ion transitions were monitored. High ion ratio errors greater than 20% prevented positive confirmation.
The scientists were using a triple quadrupole mass spectrometer and it did not possess sufficiently high mass resolution to be able to separate the ions from the suspected contaminant from those of benzophenone. The recommended procedure in these cases is to select a third ion transition but benzophenone only produced two product ions, so that route was a non-starter.
Their solution was to switch to a high-resolution mass spectrometer, operating in full scan mode. Hector Gallart-Ayala, Oscar Nunez, Encarnacion Moyano and Maria Teresa Galceran from the University of Barcelona and Claudia Martins from Thermo Fisher Scientific, Barcelona, described the outcome in Rapid Communications in Mass Spectrometry.
High resolving power saves the day
The instrument was an Orbitrap mass spectrometer, capable of operating at a mass resolving power of up to 100,000. Both this and the triple quadrupole mass spectrometer were coupled to an ultra-high performance liquid chromatography system.
A total of 28 packaged foods were spiked with a deuterium-labelled internal standard and extracted by an established QuEChERS procedure developed previously in the authors' lab for photoinitiators like benzophenone. The extracts were analysed initially by LC-tandem MS on the triple quadrupole system which confirmed the unacceptably high ion ratio deviations of more than 20%.
Further analysis of the same extracts by LC/MS with electrospray ionisation on the high-resolution instrument at three different resolving powers shed some light on the problem. At resolving powers of 10,000 and 25,000, the mass errors remained high for the benzophenone peak, so the compound could not be confirmed.
However, at a resolving power of 50,000, the mass error for benzophenone was 1.1 ppm and two closely eluting peaks were resolved. Alongside the benzophenone peak was a second ion from the unknown interferents. In a typical sample, the difference between the two ion masses was 0.0108, which is too small to allow peak resolution by the low resolution method on the triple quadrupole.
The interfering compound was identified as 1-methyl-beta-carboline, also known as harman. It has been detected in many types of foods and is produced by the cyclisation and oxidation of the amino acid tryptophan.
In this study, 27 of the 28 foods analysed contained harman. It was found in pineapple juices, orange juices, peach juices, sangria, fruit milk, soy milk, milk and baby food.
Once this interferent had been identified and its influence removed by using the high-resolution method, benzophenone could be quantified from its protonated molecular ion. The detection limits were 0.6 µg/kg in fruit juice and baby food and 1.3 µg.kg in milk. It was found in 20 of the food samples at 0.6-5.2 µk/kg.
In practice, the high- and low-resolution methods provided similar results for quantitation but the former is required for unconditional confirmation of benzophenone. This was accomplished not only from the exact mass of its protonated molecular ion but also from its spectrum obtained following all ion fragmentation.
The LC/MS method at high resolution overcomes the confirmatory problems encountered in low-resolution tandem mass spectrometry, allowing the confirmation and subsequent quantitation of benzophenone in foods. This is an important step to ensure that the occurrence of false negative results is minimised, so that the photoinitiator can be detected in foods with confidence.
The views represented in this article are solely those of the author and do not necessarily represent those of John Wiley and Sons, Ltd.