Metal cans are often coated with a resin barrier to prevent contact between food and the can. Components from these coatings can migrate into the food affecting its safety and quality. Polyepoxyphenol coatings on the inside of cans based on bisphenol A epoxy resins can release the epoxy monomer bisphenol A diglycidyl ether (BADGE) into food1, 2. Bisphenol A and its derivatives are considered as endocrine disruptors3. Both E.U. and U.S. have set regulations on the limit of BADGE migration into food at 1 mg/Kg. Using the quantitative capability of the time-of-flight (TOF) mass spectrometer, we were able to set up a calibration curve and quantitate BADGE in a tuna extract. In addition, the high mass accuracy capability of the TOF along with the proprietary AxION EC ID software, allowed us to identify an unknown impurity cyclo-di-BADGE without having an authentic standard of this compound.
Sample preparation: 10 g of tuna, transferred into a 50 mL tube and spiked with BADGE standard (200 ng). To this 10 mL of acetonitrile was added and shaken. Salts (1 g sodium chloride, 4 g magnesium sulfate, 1 g trisodium citrate, 0.5 g disodium hydrogen citrate) were added to the sample, which was shaken and centrifuged (3700 rpm) for 5 min. The supernatant (1 mL) was transferred to a dispersive SPE micro-centrifuge tube containing primary and secondary amine (PSA, 25 mg) and magnesium sulfate (150 mg) and C18 (25mg). Sample was vortexed and centrifuged at 3000 rpm for 5 min. The supernatant was carefully removed, pH adjusted with 5 µL of 5% formic acid and used for analysis.
Pump: PerkinElmer Flexar FX-15 pump
Flow: 0.4 mL/min
Mobile phase A: Water containing 0.1% formic acid
Mobile phase B: Acetonitrile containing 0.1% formic acid
Gradient conditions: 70% A/30% B to 10%A/90%B in 5 mins in a linear gradient
Injection volume: 5 μL in partial fill mode.
Column used: PerkinElmer Brownlee SPP C-18, 2×50 mm, 2.7 μm, 25 °C
Mass spectrometer: PerkinElmer AxION 2 TOF MS
Ionization source: PerkinElmer Ultraspray 2 (Dual ESI source)
Ionization mode: Positive
m/z range: 90-700
Capillary exit voltage: 100 V
Internal calibration was performed using m/z 118.08625 and 622.02896 as lock mass ions.
Results: The mass spectrum showed BADGE was predominantly observed as the [M+NH4+] ion (Figure 1). We were easily able to detect as low as 2 ppb concentration of BADGE (S/N = 52) standard. Excellent linearity (r2 > 0.995) was observed for the calibration curve generated between 2 to 500 ng/mL (Figure 2) of BADGE standard. The intra assay %RSD for triplicate injections at the 2 ppb concentration was <10%. Tuna extracts spiked with 20 ng/mL of BADGE standard were easily detected by LC-TOF (Figure 3). A 94% recovery of BADGE was observed in the spiked tuna extracts suggesting little or no ion suppression of the analyte in the extracts.
We tried to identify two unknown peaks with same exact masses eluting between 2.5 to 3 min. using the exact mass capability of the AxION 2 TOF. The accurate mass and isotope profile was entered into the AxION EC ID calculator of the software. The software uses this information and searches against the PubChem database and identifies potential molecular formulae matches. The first potential match with the highest score was identified with the elemental composition C36H40O6 within a mass error of < 1ppm. The software also provides a list of possible structures for the given elemental composition and one of the listed structures that related to bisphenol family of compounds was the BADGE.BPA linear structure. However, an isomeric structure cyclo-di-BADGE compound described in the literature could also be possible. Unlike linear BADGE.BPA structure, the cyclo-di-BADGE is known to migrate as cis and trans isomers by HPLC4. Using chemicals that distinguish between epoxide side chains found in BADGE.BPA linear structure and hydroxyl groups in cyclo-di-BADGE one can further confirm the presence of the two structures.