ABSTRACT
Liquid chromatography/mass spectrometry (LC/MS) experiments are described, leading to a reliable method for the measurement of perfluorooctanesulfonic acid (PFOS) and other perfluorinated chemicals (PFCs) in foods. Separations were performed on new fluorinated stationary phases, RP Octyl (-C(8)F(17)) or propyl-perfluorobenzene (-C(3)H(6)-C(6)F(5)), to ensure resolution of PFOS and interfering taurohydroxycholate isomers. Aqueous ammonium formate (5 mM) and methanol were used as the mobile phases. The mass spectrometer was operated in negative electrospray ionisation mode, recording two transitions for each analyte and one for each internal standard. The purities of the analytical standards for the eleven target perfluoro analytes (C(7) to C(12) carboxylic acids, C(4), C(6) and C(8) sulfonic acids, and octanesulfonamide (PFOSA)) were found to be in close agreement with the supplied values; the lowest purity was 91%. Five candidate internal standards were investigated, (13)C(4)-PFOS, (13)C(4)-perfluorooctanoic acid, (13)C(2)-perfluorodecanoic acid, D(9)-n-ethylperfluorooctanesulfonamidoethanol (D(9)-n-Et-FOSE) and D(3)-n-methylperfluorooctanesulfonamide (D(3)-n-Me-FOSA); the purities were all >98%. The use of tetrahydro-PFOS generated backgrounds (>1 microg/kg) for perfluoroheptanoic acid and perfluorobutanesulfonic acid. Similarly D(9)-n-Et-FOSE was unacceptable and D(3)-n-Me-FOSA was volatile, leaving no clear candidate for normalisation of the measurement of PFOSA. Severe matrix-induced suppression and enhancement effects influenced ionisation, making external calibration and quantification problematic. This was addressed by a parallel standard addition and matrix-matching approach, comparing ionisation in methanol, in procedural blanks and in food-based extracts. The limits of detection (LODs) of 0.001-0.01 microg/kg in solvent and 0.01-1 microg/kg in foods demonstrate that this method is suitable for the determination of PFCs in all food to the required 1 microg/kg reporting level.
