INVESTIGATION OF SEQUENTIAL EXTRACTIONS METHODS FOR DETERMINATION OF MERCURY SPECIES IN SEDIMENTS.

Chris Sladek¹ and Mae Sexauer Gustin²

¹Department of Geological Sciences ²Department of Environmental and Resource Sciences, University of Nevada, Reno, NV 89557 csladek@scs.unr.edu

ABSTRACT

This project evaluated the efficiency of some procedures commonly used in sequential extraction methods for determining mercury speciation in sediments. Pyrolitic methods were applied for the determination of volatile phases; leaching with chloride solutions was utilized for determination of soluble or mobile species; and acid digestions were used to extract strongly bound or relatively immobile species, and for total mercury analyses. It was determined that pyrolitic extraction can greatly over estimate elemental mercury. Ammonium chloride was found to be more efficient than MgCl₂ in removing soluble species. The organic content of a sediment was found to have a significant influence on sequential extraction results. Inorganic amendments also influenced extraction results but not as strongly as organic matter.

INTRODUCTION

Published methods on mercury (Hg) speciation in sediments include sequential extractions, (Di Guilio and Ryan, 1987; Revis et al., 1990; Lechler et al., 1997; Wallschlager et al., 1998), a pyrolitic extraction (Beister and Scholz, 1997) and x-ray absorption spectroscopy (XAS) (Kim et al., 1999). With sequential extraction procedures Hg is not directly measured, but determined by operationally defined single or sequential extraction methods. Nirel and Morel (1990) stressed that results of sequential extractions for most trace elements have not been validated, and indiscriminate application of a method will produce meaningless results. For example, Hg speciation of contaminated sediments from the Carson River Superfund Site (EPA, 1994) using sequential extraction methods conducted by separate labs showed discrepancies of greater 50 %. Even the analyses of total Hg showed disagreements of up to 82 %. This high degree of uncertainty in Hg speciation methods demonstrates the need for validation of sequential extraction methods and development of digestion protocols for total Hg in soils and sediments.

This paper discusses work done to assess some published sequential extraction methods and preliminary results of an attempt to develop a method that may be used to determine Hg species and mobility in a variety of sediments. The latter involved pyrolitic extraction for elemental mercury (Hg⁰), a weak chloride leach for mobile Hg (Hg²⁺), followed by a strong acid leach for HgS and strongly bound forms of Hg. Several natural and synthetic substrates were used. Pyrolitic extractions were evaluated as a function of temperature and time. Weak leaches using NH₄Cl and MgCl₂ were tested for extraction of HgCl₂. Strong acid leaches were tested for their efficiency in digestion of HgS.

METHODS

Sample and preliminary tests

Two synthetic substrates were used for this study: ground glass amended with pulverized vermiculite, and ground glass coated with iron oxide (FeOOH) generated from rusting iron nails. The pH was 6.7 and 9.2 for the vermiculite and FeOOH amended glass, respectively. Natural sediments consisted of a river bank sand (RB) and back water sediments (BW), collected from the Truckee River, Nevada, USA. Both samples were dominantly fine sand, with 9 % silt and clay (<63 mm) in the RB sediment and 12 % in the BW. Organic contents were approximately 3 % for RB and 5 % for BW. Sediments were sterilized by heating to 120 ^oC for 2 hours. The pH of the RB and BW samples was approximately 6.8. Sediment samples were amended using a solution of HgCl₂ in nanopure water (Millipore^â Milli-Q filtration system) to obtain substrate concentrations of 100 to 200 ppm Hg. After solution addition samples were dried by purging with N₂ at 23 ^oC. Samples containing Hg⁰were prepared by homogenizing a 5 - 6 mg bead of Hg in approximately 20 grams of sediment in a N₂ purged vial. Microcrystalline quartz containing disseminated cinnabar (sinter) collected from the Silver Cloud Mine, Nevada, USA was crushed to <177 mm for testing of strongly bound and total Hg digestion methods.

Individual extraction steps were first evaluated before combining them in a sequence. The pyrolitic extraction targeted Hg⁰ . Pyrolitic extractions were done at 80, 150 and 180 ^oC for 1, 3 and 8 hour intervals. The pyrolitic extraction entails heating the sample in Teflon^â vessels, removing volatilized Hg with prepurified N₂ carrier gas from the sample vessel, and collecting the Hg on an iodated carbon trap (Supelco inc.) connected to the vessel outlet. The Hg collected on the carbon trap was then eluted using 16 ml hot aqua regia. Solutions of 0.5 molar NH₄Cl and 0.5 molar MgCl₂ were tested for their ability to remove Hg from HgCl₂ amended substrates. This entailed leaching the samples with the chloride solution in a closed Teflon^â vial over night, and centrifuging and filtering the lechate through a 0.45 mm syringe filter to separate the lechate from the substrate. The leachate was placed in a 50 ml volumetric flask and acidified with 16 ml aqua regia. Two strong acid leaches (3:7 H₂SO₄ + HNO₃, and 1:3 HNO₃ + HCl (aqua regia)) were evaluated for efficiency in removing HgS and strongly bound Hg. All extractions were stored in 50 ml volumetric flasks and brought to volume before analysis with HCl diluted to maintain 30 % acidity. All analyses were performed using cold vapor atomic adsorption spectroscopy.

At 80 ^oC 76 % of the Hg⁰ amended to pure ground glass was extracted in a 1 hour period, and 99.9 % was extracted after 8 hours; for the same time periods 16 % and 48 % of the Hg that was added as HgCl₂ to pure ground glass was extracted. A NH₄Cl leach extracted 75 % to 80 % of the Hg added as HgCl₂ whereas MgCl₂ extracted only 50 % of the Hg. The cinnabar containing sinter was leached with 16 ml of the two acid solutions heated to 100 - 120 ^oC. A 2 hour leach using aqua regia extracted 428 ± 17.7 ug/g Hg from the sinter, while 3:7 H₂SO₄ + HNO₃ extracted 316 ± 10.2 ug/g. Extraction by 3:7 H₂SO₄ + HNO₃ increased to 379 ± 8.02 ug/g for an overnight extraction (approximately 14 hours).

Evaluated protocol

Once the efficiency of the individual extractions was determined a sequential extraction procedure was tested. This method consisted of a pyrolitic extraction in which Hg was collected on iodated carbon traps for 1 hour, 3 hours and 8 hours. After the pyrolitic extraction the sediment was leached with NH₄Cl over night. The residual sediment was then leached using hot aqua regia for 2 hours to determine strongly bound Hg. Extractions were performed on triplicate 1 gram samples.

RESULTS

Sequential extraction results are shown in Fig. 1 and 2 as the percent extracted with respect to the total Hg extracted. The graphs show the cumulative percentage of Hg extracted in the pyrolitic extraction (at the given temperature) for 1, 3 and 8 hours; the percentage of Hg extracted using NH₄Cl; and the percentage of residual Hg extracted using aqua regia.

For the FeOOH amended glass, at 80 ^oC 7.9 % of the Hg added as HgCl₂, was extracted in a period of 8 hours, 80 % was eluted with the NH₄Cl extraction and 12.1 % Hg was removed in the residual extraction (Fig. 1a). At 180 ^oC, 49.3 % of the Hg was extracted in the first hour and 76.1 % of the Hg was extracted after 8 hours. Following the 180 ^oC pyrolitic extraction 8 % of the Hg was extracted by NH₄Cl and 16.9 % Hg was eluted in the residual extraction. Pyrolitic extractions from the HgCl₂ amended vermiculite (Fig. 1b) showed slightly different behavior than pyrolitic extractions from FeOOH amended glass. More Hg was extracted at lower temperatures, 26.6% in 8 hours at 80 ^oC for the vermiculite verses 7.9 % in 8 hours for FeOOH. At higher temperatures more Hg is extracted from the FeOOH 76 % after 8 hours verses 69 % extracted from the vermiculite amended glass after 8 hours. Pyrolitic extractions from RB amended with HgCl₂ (Fig. 1c) were similar to the extractions from FeOOH amended glass, however slightly less Hg was removed in the NH₄Cl leach and more Hg was left as residual for the RB sediment. Extractions from BW (Fig. 1d) showed that most of the Hg 56 % to 95.8 % remained in the residual fraction and minimal amounts of Hg 0.05 % to 4 % were extracted by NH₄Cl.

Extractions of RB and BW amended separately with Hg⁰ and HgCl₂ are shown in Fig. 2 for pyrolitic extractions conducted at 180 ^oC. Comparison of the extraction profiles for Hg⁰ and HgCl₂ show that sediment type can significantly influence Hg species extraction. The percent extracted at 180 ^oC is approximately the same for Hg⁰ and HgCl₂ for the same sediment type. However, extraction efficiency varied drastically as a function of the sediment type. Approximately 25 % more Hg was strongly bound and left in the residual phase for the BW seqential extractions.

CONCLUSIONS

Mercury added as Hg⁰ and HgCl₂ is removed in a pyrolitic extraction step as a function of increasing temperature and time. This indicates that sequential extraction methods that use pyrolysis could remove a significant amount of Hg species other than Hg⁰ from a substrate. Ammonium chloride although effective at removing HgCl₂ from inorganic sediments is less effective for sediments containing organics. The amount of Hg extracted from natural sediments appears to be governed more by the sediment type, then by the species they were amended with. This suggests that chemical equilibrium within the substrate greatly influences Hg speciation. Complexing by inorganic or organic ligands, and redox processes may change the species of Hg during or subsequent to the spiking process. Organic substances in a soil are strong reducing agents and could reduce Hg²⁺ to Hg⁰. The abiotic production of methylmercury by organic substances in soils has been demonstrated (Chen et al., 1996) and the conversion of some of the Hg to methylmercury could occur during or subsequent to the amendment process. Results of the extractions that were conducted suggest that organic content has a more significant affect on the extraction of Hg species than does inorganic material such as Fe oxides or clay minerals such as vermiculite.

The sequential extraction protocol tested in this study will yield erroneous results as to the Hg species in the substrate. Applied independently the sequential extraction steps may yeild more accurate results. However, the results will be confounded by sediment chemistry and binding strength of Hg species to the substrate.

Figure 1. Sequential extraction results for samples amended with HgCl₂. Plots show sequential extraction for pyrolitic extraction temperatures of 80, 150, and 180 ^oC. Pyrolitic extraction for cumulative periods of 1,3, and 8 hours (1H, 3H, 8H). Hg²⁺ = Hg extracted by NH₄Cl. Res = Hg extracted by aqua regia. Values are the percent of total Hg extracted.

Figure 2. Comparison of sequential extractions for samples of river bank sand (RB) and back water sediment (BW) amended with Hg⁰ and HgCl₂. Pyrolitic extraction temperatures were 180 ^oC.

ACKNOWLEDGMENTS

The authors wish to thank the Nevada Bureau of Mines and Geology for their assistance and the use of their AA during the early phases of these experiments. This research was supported by the U.S. Environmental Protection Agency Star Grant # 2825249.

REFERENCES

Biester H B, Scholz C (1996) Environmental Science and Technology 31: 233-239.

Chen Y, Bonzongo J C, Miller G C (1996) Environmental Pollution 92:281-287.

Di Giulio R T, Ryan E A (1986) Water, Air and Soil Pollution 33:205-219.

EPA, (1994), Revised draft Human Health Risk Assessment and Remedial Investigation report Carson River

mercury site, prepared by Sean Hogan and Stanford Smucker.

Lechler P J, Miller J R, Hsu L-C, Desilets M O (1997) Journal of Geochemical Exploration 58:259-267.

Nirel R M V, Morel F M M, (1990) Water Resources 24:1055-1056.

Revis N W, Osborne T R, Holdsworth G, Hadden C (1989) Archives Environ. Contam. and Tox. 19:221-226.

Kim C S (1999) Abstracts with programs. GSA Cordilleran Section Annual Meeting A-69.

Wallschlager D, Desai M V M, Spengler M, Wilken R-D (1997) Environmental Quality 27: 1034-1044.