Analytical speciation of mercury in fish tissues by reversed phase liquid chromatography-inductively coupled plasma mass spectrometry with Bi3+ as internal standard

In this work, the quantification of two mercury species (Hg2+ and CH3Hg+) in fish tissues has been revisited. The originality of our approach relies on the use of Bi3+ as internal standard (IS) and on the modification of typical extraction conditions. The IS (125 μl, 1000 μg l−1 Bi3+) was added to the aliquot of fresh fish tissue (400-500 mg). A high-speed blender and ultrasound-assisted homogenization/extraction was carried out in the presence of perchloric acid (1.5 ml, 0.6 mol l−1), l-cysteine (500 μl, 0.75 mol l−1) and 500 μl toluene:methanol (1:1). Perchloric acid was used for protein denaturation and precipitation, toluene helped to destroy lipid structures potentially sequestering CH3Hg+, l-cysteine was used to form water-soluble complexes with Bi3+, Hg2+ and CH3Hg+. The excess of perchloric acid was eliminated by addition of potassium hydroxide (pH 5 with acetic acid). The obtained extract, was diluted with the mobile phase (1:1) and introduced (20 μl) to the reversed phase HPLC-ICP-MS system. The separation was achieved by isocratic elution (2.5 mmol l−1 cysteine, 12.5 mmol l−1 (NH4)2HPO4, 0.05% triethylamine, pH 7.0:methanol (96:4)) at a flow rate 0.6 ml min−1. Column effluent was on-line introduced to ICP-MS for specific detection of 202Hg, 200Hg and 209Bi. Analytical signal was defined as the ratio between 202Hg/209Bi peak areas. The detection limits evaluated for Hg2+ and CH3Hg+ were 0.8 and 0.7 μg l−1. Recovery of the procedure, calculated as the sum of species concentrations found in the sample with respect to total ICP-MS-determined Hg was 91.9% for king mackerel muscle and 89.5% for red snapper liver. In the standard addition experiments, the recovery results were 98.9% for Hg2+ and 100.6% for CH3Hg+. It should be stressed that the use of Bi3+ as IS enabled to improve analytical performance by compensating for incomplete extraction and for imprecision of sample handling during relatively non-rigorous protocol.