Development, Validation, and Application of a Method for

Development, Validation, and Application of a. Method for Quantification of Methylmercury in. Biological Marine Materials Using Gas. Chromatography At...
0 downloads 17 Views 262KB Size
Anal. Chem. 1996, 68, 3859-3866

Development, Validation, and Application of a Method for Quantification of Methylmercury in Biological Marine Materials Using Gas Chromatography Atomic Emission Detection Mary Kate (Behlke) Donais,*,†,‡ Peter C. Uden,‡ Michele M. Schantz,† and Stephen A. Wise†

Analytical Chemistry Division, Chemistry Room B208, National Institute of Standards and Technology, Gaithersburg, Maryland 20899, and Department of Chemistry, University of Massachusetts, Amherst, Massachusetts 01003-4510

A gas chromatographic method was developed for the quantification of alkylmercury species using microwaveinduced plasma atomic emission detection (GC-AED). The column conditioning and analyte derivatization required for previous methods were found to be unnecessary for stable, accurate, and sensitive element-specific detection using GC-AED. Chromatographic and detection parameters such as stationary phase type, stationary phase film thickness, GC column dimensions, helium mobile phase column head pressure, detector makeup gas flow rate, and detector reagent gas type and flow rate were found to significantly affect analyte response. The detection limit for the optimized GC-AED conditions was 0.8 pg (0.1 pg/ s) of methylmercury chloride (as mercury). A solidliquid extraction procedure with preparative gel permeation chromatography cleanup and GC-AED analysis was used to quantify methylmercury in a variety of complex matrix marine materials. The methylmercury quantification method was validated with four marine certified reference materials (CRMs). The method was then applied to 13 standard reference materials, CRMs, and control materials for which no certified reference values for methylmercury have been determined. Four National Institute of Standards and Technology Standard Reference Materials and one control material, which were analyzed using the GC-AED method, were also analyzed by two other laboratories using independent methods to further validate the method. Over the last few decades, extensive damage has been caused to the environment by mercury and organomercury compounds in commonly used products such as paints, wood preservatives, paper, and pesticides.1 The toxicities of these compounds vary considerably, however, necessitating the determination of individual species to accurately assess environmental impact. Methylmercury is one of the most toxic of these mercury species and is found commonly in the marine environment. The routine determination of methylmercury in marine mammals, fish, and * Address correspondence to this author at the National Institute of Standards and Technology. † NIST. ‡ University of Massachusetts. (1) Krenkel, P. A. CRC Crit. Rev. Environ. Control 1973, 3, 303-373. S0003-2700(96)00438-6 CCC: $12.00

© 1996 American Chemical Society

sediments is of growing importance to environmental monitoring programs around the world. Interest in methylmercury as an environmental contaminant first arose in the 1960s with reports of alkylmercury poisoning of marine life and people in Japan2 and of birds and marine life in Sweden.3 Most alkylmercury poisoning incidents have been due to short-term exposure to high levels of these compounds. Acute poisoning effects include kidney damage, damage to the central nervous system, and death.4 Chronic poisoning symptoms include tremors, constriction of visual field, lack of coordination, damage to the central nervous system, and kidney damage, and it can also be fatal. Little is known of the long-term effects of low-level alkylmercury exposure, however. Much concern has been expressed over the potential danger to humans who are exposed to low levels of alkylmercury on a daily basis. The most common mode of human exposure to alkylmercury, predominantly as methylmercury, is through the ingestion of fish and bivalves. Methylmercury is present at extremely low levels in lower forms of marine life such as algae and small fish due to constant exposure to methylmercury in ocean water.5 These lower life forms are consumed by higher life forms, and these higher life forms are consumed by marine mammals and large fish. Over time, methylmercury bioaccumulates in the tissue and fats of organisms due to most species’ inability to efficiently eliminate it from the body. Methylmercury concentrations increase as they move up the food web, often resulting in dangerously high levels in marine species consumed by humans. Therefore, monitoring methylmercury concentrations in marine materials is important to ensure the safety and welfare of both humans and the marine ecosystem. The majority of alkylmercury speciation methods are based on gas chromatographic (GC) separations with electron capture detection6,7 or atomic absorption spectrometric detection.8-10 (2) Kiyoura, R. In Advances in Water Pollution Research; Pearson, E. A., Ed.; Pergamon Press: New York, 1964; pp 291-308. (3) Johnels, A. G.; Westermark, T. In Chemical Fallout; Miller, M. W., Berg, G. G., Eds.; Charles C. Thomas: Springfield, IL, 1969. (4) Peakall, D. B.; Lovett, R. J. Bioscience 1972, 22, 20. (5) Horvat, M.; Liang, L.; Bloom, N. S. Anal. Chim. Acta 1993, 282, 153-168. (6) Westoo, G. Acta Chem. Scand. 1968, 22B, 2277-2280. (7) Zelenko, V.; Kosta, L. Talanta 1973, 20, 115-123. (8) Dressman, R. C. J. Chromatogr. Sci. 1972, 10, 472-475. (9) Bzezinska, A.; Loon, J. C. V.; Williams, D.; Oguma, K.; Fuwa, K.; Haraguchi, I. H. Spectrochim. Acta 1983, 38B, 1339-1346. (10) Robinson, J. W.; Wu, J. C. Spectrosc. Lett. 1985, 18, 47-69.

Analytical Chemistry, Vol. 68, No. 21, November 1, 1996 3859

Many of these GC methods have significant drawbacks and/or are difficult to implement in routine monitoring laboratories. Often, symptoms of significant underlying methodology problems such as analyte peak tailing, gradual deterioration of the chromatographic separation over time, and occurence of interfering chromatographic peaks were addressed without consideration of the cause of the problems. Complicated extraction procedures with multiple cleanup steps,11-13 analyte derivatization prior to chromatographic analysis,14-16 and tedious column conditioning procedures11,17,18 have been used by many to “solve” these problems. A variety of liquid chromatographic (LC) approaches to alkylmercury speciation also have been reported, including LC with ultraviolet detection,15 cold vapor atomic absorption spectrometric detection,19 and cold vapor inductively coupled plasma mass spectrometric detection.20 LC is a much more “gentle” technique than GC for thermally unstable compounds such as alkylmercury halides. The use of LC-based techniques would, therefore, seem a logical solution to column conditioning and degradation problems typical to most GC alkylmercury methods. In general, however, most LC method detection limits are at least an order of magnitude greater than those of GC methods, so they would not be preferable for trace-level measurements. The first goal of this research was to investigate these previously reported separation and detection techniques to develop an improved and simplified approach to organomercury measurements in complex matrix materials. Capillary GC with microwaveinduced plasma (MIP) atomic emission detection is used for this new method due to the high separation efficiencies typical to capillary GC, the wide range of GC stationary phase types and capillary column dimensions that could be investigated, the element-specific nature of the detector, and the extreme sensitivity of the MIP to metallic compounds. The second goal of this research was to develop a simple, selective extraction procedure for alkylmercury species in complex matrix materials. A solid-liquid extraction procedure under acidic conditions for the selective extraction of alkylmercury species from marine tissues and sediments is presented. Interfering sulfurcontaining species are removed during sample extractions with copper powder,21 permitting one extraction procedure to be used successfully for a variety of sample matrices. High molecular weight pigments and lipids are removed from the extracts by preparative gel permeation chromatography (GPC) prior to GC analysis. A helium microwave-induced plasma atomic emission detector is used for sensitive, element-specific detection of the target alkylmercury species. Chromatographic and detection parameters were optimized to maximize analyte signal. Column conditioning and analyte derivatization were found to be unneces(11) Berman, S. S.; Siu, K. W. M.; Maxwell, P. S.; Beauchemin, D.; Clancy, V. P. Fresenius’ Z. Anal. Chem. 1989, 333, 641-644. (12) Lansens, P.; Meuleman, C.; Leermakers, M.; Baeyens, W. Anal. Chim. Acta 1990, 234, 417-424. (13) Horvat, M. Water, Air, Soil Pollut. 1991, 56, 95-102. (14) Horvat, M.; Byrne, A. R.; May, K. Talanta 1990, 37, 207-212. (15) Wilken, R.-D. Fresenius’ J. Anal. Chem. 1992, 342, 795-801. (16) Brunmark, P.; Skarping, G.; Schutz, A. J. Chromatogr. 1992, 573, 35-41. (17) O’Reilly, J. E. J. Chromatogr. 1982, 238, 433-444. (18) Rubi, E.; Lorenzo, R. A.; Casais, C.; Carro, A. M.; Cela, R. J. Chromatogr. 1992, 605, 69-80. (19) Palmisano, F.; Zambonin, P. G.; Cardellicchio, N. Fresenius’ J. Anal. Chem. 1993, 346, 648-652. (20) Bushee, D. S. Analyst 1988, 113, 1167-1170. (21) Behlke, M. K.; Uden, P. C.; Schantz, M. M. Anal. Commun. 1996, 33, 9192.

3860

Analytical Chemistry, Vol. 68, No. 21, November 1, 1996

sary to achieve stable chromatographic separations and quantification of methylmercury at the nanogram per gram level in marine materials. The method was validated by analyzing four certified reference materials (CRMs) that have certified values for methylmercury. The method was then applied to 13 other CRMs and control materials from a variety of sources for which no certified values for methylmercury have been determined.

EXPERIMENTAL SECTION Disclaimer. Certain commercial equipment, instruments, or materials are identified in this report to specify adequately the experimental procedure. Such identification does not imply recommendation or endorsement by the National Institute of Standards and Technology, nor does it imply that the materials or equipment identified are the best available for the purpose. Reagents and Chemicals. Calibration and internal standard solutions were prepared gravimetrically from reagent grade methylmercury chloride (Pfaltz & Bauer, Waterbury, CT), ethylmercury chloride (Johnson Matthey, Ward Hill, MA), and propylmercury chloride (Pfaltz & Bauer) in chromatographic grade toluene. Alkylmercury chlorides are toxic at very low levels and were handled as indicated on the most current material saftey data sheets for each compound. Disposal of all mercury-containing waste from these experiments was performed in accordance with facility guidelines. A 1 mol/L copper sulfate solution (16 g of copper sulfate to 100 mL with chromatographic grade water) and an acidic potassium bromide solution (16 g of potassium bromide and 5 mL of concentrated sulfuric acid to 100 mL with chromatographic grade water) were prepared for use in the extraction procedure. Copper powder was activated in concentrated hydrochloric acid, rinsed with water, methanol, and dichloromethane until neutral, and stored in dichloromethane prior to use. Instrumentation. GPC removal of high molecular weight species from the sample extracts was carried out on a poly(styrene-divinylbenzene) PL-GEL column (10 µm particles, 10 nm pores, 60 cm × 2.5 cm i.d., Polymer Laboratories, Amherst, MA) using a 10 mL/min flow rate of 100% dichloromethane and a 1.5 mL injection loop. Gas chromatographic separations and detection were achieved using a 5890 Series II GC (Hewlett-Packard, Wilmington, DE) with a 5921A atomic emission detector (Hewlett-Packard, GC-AED), monitoring the mercury emission line at 253.652 nm. The GCAED was operated as indicated by the instrument manufacturer with attention to all relevant safety precautions. The following six GC columns were evaluated for the separation of alkylmercury halides: 100% dimethyl polysiloxane column (DB-1, 15 m × 0.53 mm i.d., 1.5 µm film thickness, J & W Scientific, Folsom, CA); poly(ethylene glycol) column (DB-WAX, 15 m × 0.53 mm i.d., 1 µm film thickness, J & W Scientific); two 60 m poly(14%cyanopropyl-86%-dimethylsiloxane) columns with differing film thicknesses (DB-1701, 60 m × 0.23 mm i.d., 0.25 µm and 1 µm film thicknesses, J & W Scientific); a shorter poly(14%-cyanopropyl86%-dimethylsiloxane) column (DB-1701, 10 m × 0.53 mm i.d., 0.5 µm film thickness, J & W Scientific); and a short, thick-film poly(14%-cyanopropyl-86%-dimethylsiloxane) column (OV-1701, 15 m × 0.53 mm i.d., 3.0 µm film thickness, Quadrex, New Haven, CT).

Table 1. CRMs Used To Validate Method CRM name

matrix and sampling location

issuing agency

certified MeHg concn (µg/g as Hg)a

sample state

TORT-122 DOLT-223 IAEA-35024 DORM-225

lobster hepatopancreas, Prince Edward Island dogfish liver, Canada tuna homogenate, Northern Mediterranean dogfish muscle, Canada

NRCC NRCC IAEA NRCC

0.128 ( 0.014 0.693 ( 0.053 3.65 (3.32-4.01) 4.47 ( 0.32

freeze-dried freeze-dried freeze-dried freeze-dried

a

Values based on dry weight of material.

Table 2. Materials Analyzed for Methylmercury Content material names SRM 1646a26 SRM 1941a27 SRM 158828 SRM 270429 SRM 1566a30 SRM 194531 SRM 297632 CARP-233 SRM 297434 SRM 1974a35 fish homogenate36 whale liver control material37 shark control material38

matrix and sampling location

issuing agency

sample state

estuarine sediment, Chesepeake Bay, VA marine sediment, Baltimore Harbor, MD cod liver oil, Norway river sediment, Buffalo River, NY oyster tissue, Bon Secour, AL whale blubber, Cape Cod, MA mussel tissue, Mediterranean Sea carp homogenate, Lake Huron mussel tissue, Boston Harbor, MA mussel tissue, Boston Harbor, MA trout/salmon homogenate, Lake Michigan, Lake Ontario, and Alaska pilot whale liver, Cape Cod, MA bull shark tissue, Gulf Coast of Florida

NIST NIST NIST NIST NIST NIST NIST NRCC NIST NIST NWF-NOAA

freeze-dried freeze-dried room temperature oil freeze-dried freeze-dried frozen freeze-dried room temperature, wet tissue freeze-dried frozen frozen

NIST NIST

frozen frozen

Samples. A description and source of the samples used to validate the method are shown in Table 1.22-25 A description and source of the thirteen additional materials analyzed for methylmercury content are shown in Table 2.26-38 The following (22) Certificate of Analysis for TORT-1, Lobster Hepatopancreas. National Research Council Canada: Ontario, Canada, 1990. (23) Certificate of Analysis for DOLT-2, Dogfish Liver. National Research Council Canada: Ontario, Canada, 1993. (24) Certificate of Analysis for IAEA-350, Trace Elements in Tuna Fish Homogenate. International Atomic Energy Agency: Monaco. (25) Certificate of Analysis for DORM-2, Dogfish Muscle. National Research Council Canada: Ontario, Canada, 1994. (26) Certificate of Analysis for Standard Reference Material 1646a, Estuarine Sediment. National Institute of Standards and Technology: Gaithersburg, MD, 1995. (27) Certificate of Analysis for Standard Reference Material 1941a, Organics in Marine Sediment. National Institute of Standards and Technology: Gaithersburg, MD, 1994. (28) Certificate of Analysis for Standard Reference Material 1588, Organics in Cod Liver Oil. National Institute of Standards and Technology: Gaithersburg, MD, 1989. (29) Certificate of Analysis for Standard Reference Material 2704, Buffalo River Sediment. National Institute of Standards and Technology: Gaithersburg, MD, 1990. (30) Certificate of Analysis for Standard Reference Material 1566a, Oyster Tissue. National Institute of Standards and Technology: Gaithersburg, MD, 1989. (31) Certificate of Analysis for Standard Reference Material 1945, Organics in Whale Blubber. National Institute of Standards and Technology: Gaithersburg, MD, 1994. (32) Certificate of Analysis for Standard Reference Material 2976, Mussel Tissue. National Institute of Standards and Technology: Gaithersburg, MD, 1996 (in preparation). (33) Candidate reference material for CARP-2, Ground Whole Carp Reference Material. National Research Council Canada: Ontario, Candada, 1995. (34) Certificate of Analysis for Standard Reference Material 2974, Organics in Freeze-dried Mussel Tissue. National Institute of Standards and Technology: Gaithersburg, MD, 1996 (in preparation). (35) Certificate of Analysis for Standard Reference Material 1974a, Organics in Mussel Tissue. National Institute of Standards and Technology: Gaithersburg, MD, 1995. (36) Herman, D.; Krahn, P. Trout/Salmon Homogentate Intercomparison Material. Northwest Fisheries Service-National Oceanic and Atmospheric Administration: Seattle, WA, 1995.

abbreviations are used for the sources of the materials: National Research Council Canada, NRCC; International Atomic Energy Agency, IAEA; National Institute of Standards and Technology, NIST; and Northwest Fisheries Center-National Oceanic and Atmospheric Administration, NWF-NOAA. Method Development. (i) Evaluation of GC Columns and Optimization of Column Head Pressure. Alkylmercury species are extracted from biological tissues as halide salts, commonly the chloride or bromide forms. Therefore, initial chromatographic method development focused on the separation of methylmercury chloride (MeHgCl), ethylmercury chloride (EtHgCl), and propylmercury chloride (PrHgCl). Due to the thermal instability of alkylmercury halides, gas chromatographic sample introduction in all investigations was performed using automated, cool oncolumn injection. Recommended operating conditions for detection of the mercury emission line at 253.652 nm were used for all gas chromatographic optimization studies. The six GC columns were evaluated for retention of the target analytes, peak sharpness, peak symmetry, and stability of the chromatographic separation. A mixture of MeHgCl and EtHgCl was separated at constant column pressures from 103 to 241 kPa using the chosen GC column to study the effect of column head pressure on peak symmetry and peak area. (ii) Optimization of AED Parameters. Following chromatographic method development, the detector operating conditions was optimized to obtain the highest sensitivity for the target alkylmercury compounds. The four detector parameters investigated were reagent gas type, reagent gas flow rate, makeup gas flow rate, and discharge tube material. (37) Wise, S. A.; Schantz, M. M.; Koster, B. J.; Demiralp, R.; Mackey, E. A.; Greenberg, R. R.; Burow, M.; Ostapczuk, P.; Lillestolen, T. I. Fresenius’ J. Anal. Chem. 1993, 345, 270-277. (38) Koster, B. J. National Institute of Standards and Technology, personal communication, 1994.

Analytical Chemistry, Vol. 68, No. 21, November 1, 1996

3861

Independent optimization of oxygen and hydrogen as reagent gases and their corresponding flow rates (varied by changing the pressure at the gas cylinders) was conducted in two parts. First, oxygen was held constant at 138 kPa while hydrogen was varied from 0 to 552 kPa to determine the optimum hydrogen flow rate. The hydrogen reagent gas flow rate then was held constant at its optimum pressure while the oxygen flow rate was varied from 0 to 552 kPa. Although possible improvements in instrument sensitivity could have been achieved through further investigations of the interdependence of these two reagent gases, this simplified optimization procedure was judged sufficient for the described goals of this research. Also, each GC-AED system is unique and requires its own optimization in order to obtain maximum sensitivity for a specific element. The optimum reagent gas pressures measured during this research may not be the same as those determined for other GC-AED instrumentation. The optimum makeup gas flow rate was determined next by analyzing a mixture of MeHgCl and EtHgCl by GC-AED using makeup gas flow rates from 60 (no supplemental flow) to 160 mL/ min. Last, silica and alumina discharge tubes were evaluated for alkylmercury species detection. (iii) Extraction and Cleanup Procedures. Four biological tissue CRMs of differing matrix compositions with a wide range of certified methylmercury concentrations were analyzed during development of the alkylmercury extraction procedure to ensure method validity. A modified version of the method reported by Gui-bin et al.39 was used as a starting point in method development. The four major modifications to the Gui-bin method necessary to improve method precision and validate the procedure were as follows: (1) all samples and standards were extracted using the same procedure prior to GC-AED analysis; (2) internal standard was added to all samples and standards to correct for final volume differences; (3) copper powder was added to all solid samples prior to extraction to remove interfering sulfur-containing species; and (4) high molecular weight pigments and lipids were removed from the extracts by preparative GPC prior to GC-AED analysis. The final extraction procedure was as follows: (1) weigh sample and 4 g of copper powder or standard into a screw-cap 50 mL centrifuge tube; (2) add 1 mL of 1 mol/L copper sulfate solution, 8 mL of water (no water added to frozen materials or SRM 2974), and shake until well mixed; (3) add 4 mL of acidic potassium bromide solution and shake until well mixed; (4) add 2 mL of toluene and rotate tube end-over-end for 1 h; (5) centrifuge and separate toluene layer; (6) repeat extraction with a second 2 mL aliquot of toluene; (7) combine toluene fractions, add internal standard, and concentrate to 0.5 mL; (8) remove high molecular weight species by GPC (no GPC cleanup performed for SRM 2976) by collecting a 50 mL fraction; and (9) concentrate fraction to 0.5 mL while switching the solvent back to toluene and analyze by GC-AED. Three subsamples of each CRM and at least two subsamples of each RM were extracted and analyzed by GC-AED together with MeHgCl standard solutions. Response factors were determined relative to an internal standard. EtHgCl was used as the internal standard for all samples but SRM 2976. The internal standard for SRM 2976 was switched to PrHgCl due to a small amount of EtHgCl naturally present in SRM 2976. (39) Gui-bin, J.; Zhe-ming, N.; Shun-rong, W.; Heng-bin, H. Fresenius’ Z. Anal. Chem. 1989, 334, 27-30.

3862

Analytical Chemistry, Vol. 68, No. 21, November 1, 1996

Independent Method Measurements. Subsamples of SRM 1974a, SRM 2974, SRM 1566a, SRM 2976, and the whale liver control material were analyzed by two other laboratories using methods independent from the GC-AED method. SRM 1974a, SRM 2974, and SRM 2976 were analyzed by the Marine Environment Laboratory/IAEA (Monaco) using distillation extraction/ ethylation/room temperature precollection/gas chromatography/ pyrolysis/cold vapor atomic fluorescence spectrometric detection (CV-AFS);40 this method is referred to as independent method one or IM-1. SRM 2974, SRM 1974a, SRM 1566a, SRM 2976, and the whale liver control material were analyzed by the Institute of Applied Physical Chemistry, Research Centre of Ju¨lich (Ju¨lich, Germany) using acid extraction/anion-exchange chromatography/ UV digestion/cold vapor atomic absorption spectrometry (CVAAS);41 this method is referred to as IM-2. Last, in 1988, Horvat and co-workers reported the methylmercury content of SRM 1566a determined using three different analytical approaches;42 these data are included in the independent method comparison. RESULTS AND DISCUSSION Method Development. (i) Evaluation of GC Columns and Optimization of Column Head Pressure. GC Columns. Initial investigations of the DB-1 column quickly demonstrated its inadequacy for use in alkylmercury halide separations. Methylmercury chloride eluted very close to the solvent peak using a 138 kPa constant column pressure. When the column pressure was reduced to 69 kPa in an attempt to more efficiently separate the analyte and solvent, significant peak tailing was observed. Therefore, high column head pressures reduced peak tailing, and a more polar stationary phase was needed to separate the solvent and MeHgCl. The polar DB-WAX stationary phase permitted better separation of the solvent and MeHgCl as compared to the nonpolar DB-1 phase, but very broad peaks were observed for submicrogram per gram alkylmercury chloride concentrations. The broad peak shape on the polar DB-WAX column indicated undesirable analyte/stationary phase interactions. The third column evaluated was the 60 m, 0.25 µm film thickness DB-1701 column. A broad peak of constant size at 2030 min was observed for all MeHgCl injections. This peak is believed to be mercuric chloride, a degradation product produced during the chromatographic separation. The formation of degradation products during separation may cause a downward drift in peak area over time and quick deterioration of the column. The 1.0 µm film thickness, 60 m DB-1701 column was evaluated next. The MeHgCl peak height for the 1.0 µm film thickness column was much higher than that observed for the thinner-film column with identical amounts of MeHgCl injected, suggesting some MeHgCl was irreversibly retained on the 0.25 µm column. However, a gradual increase in MeHgCl peak tailing and a downward drift in peak height were observed for the 1.0 µm film thickness column. The final two columns evaluated were the short 0.5 µm film thickness DB-1701 column and the 3.0 µm film thickness OV1701 column. A 2 m section of 0.53 mm i.d., 0.5 µm film thickness column was attached to the detector end of the 3.0 µm film thickness column to reduce the accumulation of stationary phase (40) Horvat, M.; Liang, L.; Bloom, N. S. Anal. Chim. Acta 1993, 281, 135-152. (41) May, K.; Stoeppler, M.; Reisinger, K. Toxicol. Environ. Chem. 1987, 13, 153-159. (42) Horvat, M.; May, K.; Stoeppler, M.; Byrne, A. R. Appl. Organomet. Chem. 1988, 2, 515-524.

Figure 1. Hg-specific chromatogram of MeHgCl, EtHgCl, and PrHgCl separated on 3.0 µm film thickness OV-1701 capillary column: (1) MeHgCl, (2) EtHgCl, and (3) PrHgCl.

in the detector discharge tube. The thinner-film column showed no evidence of on-column alkylmercury halide degradation, but a lower boiling point solvent (benzene) was required to separate MeHgCl from the solvent. MeHgCl was easily separated from toluene using the 3.0 µm film thickness column due to the increased capacity of the thicker-film column. Toluene is a preferred solvent over benzene due to the gradual phasing out of carcinogenics in industrial laboratories. Therefore, the capillary GC column chosen for use in the quantification of alkylmercury species was the 15 m, 0.53 mm i.d., 3.0 µm film thickness OV-1701 column. The final temperature program used for all subsequent organomercury separations was as follows: hold at 95 °C for 0.4 min, ramp at 45 °C/min to 120 °C, hold at 120 °C for 0.5 min, ramp at 20 °C/min to 230 °C, hold at 230 °C for 10 min. The hold at 230 °C at the end of the separation was used to prevent the possible buildup of alkylHgCl degradation products in the column. Column Pressure. The contant column head pressure that produced the most symmetrical peaks with the highest peak area was 207 kPa. This corresponded to an initial flow rate of approximately 100 mL/min, much higher than typical megabore capillary GC flow rates. This high flow rate significantly reduced peak tailing and minimized the residence times of the analytes in the column, thus reducing the risk of on-column analyte degradation. The separation of MeHgCl, EtHgCl, and PrHgCl on the OV1701 column with 207 kPa constant column pressure is shown in Figure 1. Complete baseline resolution of the three species is obtained in less than 3 min. The same column has been used successfully for over 1 year in the NIST laboratory with periodic removal of short column sections at the injection port end. (ii) Optimization of AED Parameters. Reagent Gas Type and Flow Rate. A maximum in MeHgCl peak area was observed for 276 kPa hydrogen and 0 kPa oxygen, as shown in Figure 2. Oxygen is used as a reagent gas to help prevent the accumulation of elemental carbon in the AED discharge tube. Although the carbon background without oxygen reagent gas was higher than that with oxygen for these alkylmercury chloride separations, no detrimental carbon accumulation in the discharge tube was observed. Makeup Gas Flow Rate. The supplemental helium added to the column flow prior to detection, referred to as makeup gas, can affect the sensitivity and peak shapes of the target compounds. A maximum in peak height for both MeHgCl and EtHgCl was observed for 100 mL/min makeup gas flow rate, as shown in Figure 3. The ratio of MeHgCl to EtHgCl (scaled by multiplying by 10 000 to fit the graph) remained relatively constant over the 60-160 mL/min flow rate range, however, indicating the two compounds have similar detection characteristics. No relationship between makeup gas flow rate and peak symmetry was observed. Discharge Tube Material. Either silica or alumina discharge tubes are available for use in the AED. The silica tube is suitable

for most applications but must be replaced approximately every 4 weeks to avoid tube breakage and resulting water leaks. The alumina tube is much stronger, and its use has been shown to enhance peak shapes and detector response for certain elements.43 Few studies have been performed on the effect of discharge tube material on the detection of individual elements using the AED. Consequently, testing each tube material for a particular application is the easiest way to evaluate which material produces the most desirable chromatographic characteristics. A direct comparison of the two material types was not possible for this application, however, due to dimensional differences between the two tubes. The internal diameter of the alumina discharge tube is slightly smaller than that of the silica tube. This caused the pressure in the detector cavity to increase significantly when the alumina tube was used in combination with a megabore capillary GC column. A column pressure of 207 kPa cannot be used with the alumina tube because of possible damage to the detector cavity. When the column pressure was reduced sufficiently to avoid detector damage, considerable deterioration was observed in the chromatographic separation. No further studies using the alumina discharge tube were performed. (iii) Detection Limit of GC-AED Method. The detection limit (signal-to-noise ratio of 3:1) for the optimized GC-AED method was found to be 0.8 pg of MeHgCl as Hg (0.1 pg/s) injected on-column. This detection limit is a substantal improvement over previously observed detection limits, considering that column conditioning and derivatization were not used. O’Reilly reported a detection limit of 0.2 pg using a conditioned GC column,17 and Wilken reported a detection limit of 1 pg for derivatized alkylmercury species,15 both comparable to the new method. The detection limit for a nonconditioned column and no analyte derivatization was reported by Gui-bin et al. as 100 pg,39 over 100 times higher than that for the new GC-AED method. (iv) Validation of Extraction Procedure and GC-AED Method. The mean methylmercury contents of the four CRMs as determined by the proposed method are shown at the top of Table 3. Methylmercury concentrations are reported as microgram MeHg (as Hg) per gram of dry material. The methylmercury concentrations are converted to a Hg basis to permit the calculation of percentage methylmercury of total Hg in the samples. Corrections for moisture content were made on the basis of material masses before and after freeze-drying.44 All uncertainties except those for IAEA-350 are expressed as 95% confidence intervals. The confidence level for the IAEA-350 uncertainty was not provided on the certificate. The GC-AED results were well within the certified value uncertainties, thus establishing the validity of this method for materials with matrices similar to dogfish muscle, dogfish liver, tuna, and lobster hepatopancreas. Representative chromatograms for each CRM are shown in Figure 4. The excess of potassium bromide used during sample extraction promotes the conversion of the alkylmercury chlorides to their bromide forms, as indicated in the peak labels in Figure 4. This conversion from chloride to bromide forms was advantageous because MeHgBr has a higher partition coefficient.45 The conversion from alkylmercury chloride to bromide was confirmed using (43) Gavlick, W. K.; Stuff, J. R.; Uden, P. C. J. Chromatogr. 1990, 508, 187192. (44) Porter, B. J. National Institute of Standards and Technology, personal communications, 1994-1996. (45) Ealy, J. A.; Shuts, W. D. Anal. Chim. Acta 1973, 64, 235.

Analytical Chemistry, Vol. 68, No. 21, November 1, 1996

3863

Figure 2. Optimization of AED reagent gas flow rates.

Figure 3. Optimization of AED makeup gas flow rate.

bromine-specific AED detection. No peaks other than those of methylmercury, ethylmercury, and propylmercury were observed. The addition of copper powder to the samples prior to extraction was found necessary to ensure the accurate quantification of methylmercury. Sulfur species in the samples were removed by the copper powder during extraction, thus preventing interactions between sulfur and the organomercury analytes. A more detailed description of the use of copper powder to eliminate sulfur interferences during alkylmercury extractions from biological tissues is reported elsewhere.21 GPC was used to remove high molecular weight lipids and pigments from the sample extracts prior to GC-AED analysis to prevent contamination of the GC column and deterioration of the chromatographic separation. Although elimination of this cleanup step would decrease sample preparation time considerably, the removal of these high molecular weight species from the sample extracts significantly increases the lifetime of the GC column. The selectivity of the AED for Hg does not necessitate the use of a cleanup step prior to quantification, however. 3864

Analytical Chemistry, Vol. 68, No. 21, November 1, 1996

Methylmercury and Percent Methylmercury in Samples. The mean methylmercury content, total mercury content where available, and percentage methylmercury of total mercury in the four CRMs and 13 other materials are shown in Table 3. Methylmercury concentrations are reported as micrograms of MeHg (as Hg) per gram of dry material for all samples except SRM 1945, SRM 1588, and the whale liver control material, which are reported on a wet weight basis. Total mercury concentrations were obtained from each material’s certificate of analysis or from interlaboratory comparison exercises or were determined using CV-AAS.37,46,47 The sources of data are indicated in Table 3. The methylmercury concentrations in the marine materials varied considerably from the lowest measured, 0.62 ng/g in SRM 2704 Buffalo River sediment, to the highest, 13.5 µg/g in the shark control material. Methylmercury concentrations increased as the (46) Saraswati, J.; Vetter, T. W.; Watters, R. J., Jr. Mikrochim. Acta 1995, 118, 163-175. (47) Padberg, S.; Burow, M.; Stoeppler, M. Colloquium Atomspektrometrische Spurenanalytik, U ¨ berlingen, Germany, 1991.

Table 3. Summary of Total Mercury and Methylmercury Data for the CRMs and Other Materials Analyzed CRM name TORT-1 DOLT-2 IAEA-350 DORM-2 SRM 1646a SRM 1941a SRM 1588 SRM 2704 SRM 1566a

SRM 1945 SRM 2976

CERT, IM, or LV MeHg concna,b (µg/g as Hg)

GC-AED MeHg concnc (µg/g as Hg)

total Hg concn (µg/g)b

%MeHg of total Hg

CERT: 0.128 ( 0.014 CERT: 0.693 ( 0.053 CERT: 3.65 (3.32-4.01) CERT: 4.47 ( 0.32 nd nd nd nd IM-2: 0.0168 ( 0.0005 LV: 0.0098 ( 0.0010g LV: 0.0088 ( 0.0011h LV: 0.0096 ( 0.0010i nd IM-1: 0.0268 ( 0.0015 IM-2: 0.0281 ( 0.0008

0.125 ( 0.041 0.66 ( 0.15 3.42 ( 0.54 4.34 ( 0.35