Uncertainty contributions to species specific isotope dilution analysis
Mercury speciation in solid sample matrices has been investigated using high performance liquid chromatography (HPLC) coupled with multicollector sector field (MCSF) and quadrupole (Q) inductively coupled plasma mass spectrometry (ICP-MS) for species specific isotope dilution mass spectrometry (IDMS). 199Hg enriched methylmercurychloride has been synthesised and recovered in the solid form for use as a spike material. The stability of methylmercury during the IDMS procedure was investigated using 199Hg and 13C labelled methylmercury isotopomers and ¹H Nuclear Magnetic Resonance spectroscopy. IntermoIecuJar exchange of the methylmercury halide counter ion was observed, the halide counter ion order of preference was l>Br>Cl. No evidence was found for the decomposition, or formation, of methylmercury during equilibration with soil (NIST2710 SRM) or dogfish muscle (DORM-2 CRM), or during chromatographic separation. The extent of equilibration between the spike and the particulate bound mercury compounds was studied by temporal monitoring of the 200Hg:199Hg isotope amount ratio and determining the amount of Hg species in the liquid phase. For N1ST2710, complete equilibration was only achieved when concentrated HNO3 in combination with a microwave digestion was employed. For DORM-2, complete equilibration was achieved when using 1:1 H2O.CH3OH v\v and 0.01 % 2-mercaptoethanol as the solvent, even though only 47% of the analyte was extracted into the liquid phase. The mass fraction of methylmercurychloride has been determined in E)ORM-2 and BCR464 lobster hepatopancreas CRM by two different procedures, single IDMS and approximate matching double IDMS. Mercury cold vapour generation of the HPLC column eluent allowed isotope amount ratios measurements by MC-SF-ICP-MS. For each CRM the mass fraction of methylmercury determined by the two IDMS methods was not statistically different, within the limits of uncertainty, from the certified values. An uncertainty budget for both IDMS procedures has been formulated to allow the performance of each method to be compared For single IDMS the major uncertainty contribution was derived from the within replicate uncertainty, Uwithin The combined standard uncertainty of each replicate analysis was dominated by two components, the uncertainty associated with the natural isotonic abundance 200Hg: 199Hg isotope amount ratio and the uncertainty associated with the 199Hg enriched methylmercurychloride spike mass fraction. The between blend standard uncertainty, Ubetween, was the major contributor to the expanded uncertainty for approximate matching double IDMS. The combined standard uncertainty for each individual replicate was dominated by the contribution from the standard uncertainty associated with the measured 200Hg:199Hg isotope amount ratios in the spiked sample and the mass bias calibration blend.