Investigation of errors introduced by the species distribution of mercury in organic solutions on total mercury determination by electrothermal vaporisation-inductively coupled plasma mass spectrometry

This paper presents an investigation into the potential and limitations of an ETV system for the determination of total mercury, present as different species, in organic solution. The thermal and chemical stability of four me rcury species in an electrothermal vaporiser, ETV, coupled to an inductivel y coupled plasma (ICP-MS) mass spectrometer was investigated when samples o f hydrocarbon products were introduced. Calibration by the method of isotop e dilution was tested with an isotopically enriched standard of mercury die thyldithiocarbamate, (HgDDTC)-Hg-201, in toluene. Sulfur-containing compoun ds, different palladium solutions and gold metals were investigated for use as matrix modifiers. A combination of Pd(NO3)(2) in propanol with stabilis ing additives applied as a permanent modifier (A. Asheim, poster presentati on, 3rd European Furnace Symposium, Prague, 1998) and thionyl chloride inje cted with each sample gave the most effective stabilisation of mercury spec ies and reproducible signals with a relative standard deviation of 3% for 5 mu g l(-1) HgCl2. The recoveries, by isotope dilution, of HgCl2 and CH3HgC l in natural gas condensates ranged between 100 and 115%, while for Hg-0 an d (CH3)(2)Hg only 48-65% was recovered with optimised conditions. Losses of the latter two species occurred on drying of the sample in the furnace pri or to heating to the pyrolysis temperature. Systematic errors in total merc ury determination were, therefore, found to occur for samples containing Hg -0 or (CH3)(2)Hg, even with isotope dilution calibration. The detection lim it, based on 3s of 10 replicate measurements of the Hg-202(+) signal for to luene spiked with 6 mu g l(-1 201)HgDDTC, was 0.19 mu g l(-1).