Speciation of mercury by chromatography coupled with atomic spectrometry
A commercial GC-AFS instrument has been developed and optimised for the speciation of organomercury. This instrument couples a GC oven to a modified atomic fluorescence detector via a ceramic pyrolyser. Organomercury compounds in dichloromethane solvent were directly injected through a Programmable Temperature Vaporiser Injector onto a DBl Megabore column. Once separated, the compounds eluted from the column and were atomised in the pyrolyser then detected by AFS. The direct injection technique, ceramic pyrolysis design and argon purged detector have improved previous instrument designs by enhancing and maintaining sensitivity. The instrumental limit of detection was determined to be 0.25 pg Hg absolute. Methods were developed for the extraction of methylmercury from a variety of marine samples. The techniques were validated using mussel homogenate and dogfish liver (IAEA 142, SRM 8044 and DOLT-2) certified reference materials. An interlaboratory comparision exercise was participated in and a method was developed for the detemination of methylmercury in Fucus sea plant (IAEA 140). A concentration of 0.63 ± 0.006 ng g-1 was reported. The material is now certified at 0.626 +0.139 ng g-1. Of all the participating laboratories, this was the closest result to the certified value. The instrument and methods were also applied to soil and sediment samples. Once again validation was performed with a CRM sediment, IAEA 356. Although this material has been reported to give positive artifact formation when using a steam distillation sample preparation procedure, good agreement and no artifects were observed upon analysis. A further contaminated land, an uncontaminated soil and sediment sample were also studied. For all the samples studied by GC-AFS total mercury measurements were also made following an appropriate digestion procedure and CV-AFS. A gas chromatograph was also coupled with ICP-MS and HPLC was coupled to CV-AFS as comparative techniques. Both approaches were optimised and validated with CRM's. The GC-ICP-MS had the advantage of providing additional element information and confirmed the presence of methylmercury bromide in the final mussel homogenate extract. The HPLC approach found to be much less sensitive than the GC techniques and also suffered from vapour generation interferences. The PTV injector was considered for large volume injection and thermal desorption techniques. Injector breakdown problems were overcome by optimising the conditions and solid phase adsorbent for cold splitless injection. A recovery of 70% was achieved for a 50 ul large volume injection of methylmercury chloride in DCM. This technique indicated the possibility that LVI may in the future offer increased method sensitivity.