Environmental Forensics: Contaminant Specific Guide

journals.

1

Mercury

Gary N. Bigham, Betsy Henry and Brad Bessinger

Mercury (Hg) is a naturally occurring element that exists in three oxidation states. Of the three forms, only elemental mercury and mercuric mercury contribute to the global mass balance compounds can be explained by the instability of Hg-Hg bonds that characterize the Hg. Chemical transformations between forms govern the fate and transport of elemental mercury in the atmosphere. The production of mercuric methyl mercury in aqueous environments is similarly dependent. From a forensic science perspective, the most important transformation of elemental mercury is its conversion to mercuric, a species that is readily deposited and potentially transported to shallow-sediment aquatic environments that are conducive to methylation. The oxidation of elemental mercury occurs primarily at the solid or liquid interface in fog and cloud droplets, but some gas phase oxidation also occurs. The relatively high mobility of elemental mercury is reflected in atmospheric residence times that are estimated to be in the order of several months to one year. This longevity allows atmospheric emissions to contribute to terrestrial and oceanic mercury loads on both regional and global scales.

Contents

1.1. INTRODUCTION

1.2. NATURALLY OCCURRING SOURCES OF ELEMENTAL MERCURY

1.3. SOURCES OF MERCURIC MERCURY AND METHYLMERCURY

1.4. ANTHROPOGENIC SOURCES

1.5. DETECTING MERCURY IN INDOOR AIR

1.6. MERCURY FORENSICS

REFERENCES

1.1

INTRODUCTION

Mercury is a naturally occurring element that exists in three oxidation states: Hg(0) (Hg⁰, elemental mercury), Hg(I) (Hg+, mercurous mercury), and Hg(II) (Hg²+, mercuric mercury). Of the three forms, only elemental mercury and mercuric mercury contribute to the global mass balance (the rarity of Hg(I) compounds can be explained by the instability of Hg–Hg bonds that characterize the Hg(I) forms). Although mercury is nondegradable, it readily converts between the oxidation states and chemical species listed in Table 1.1.1. Chemical transformations between forms govern the fate and transport of elemental mercury in the atmosphere. The production of mercuric methylmercury (CH3Hg+) in aqueous environments is similarly dependent.

Table 1.1.1

Typical Mercury Species and Concentrations of Mercury Species in Gas and Water Phases in the Atmosphere

aAtmospheric concentrations in gas and water phases from Seigneur et al. (1994) (where μg represents micrograms).

bTypical species reported in air from Lin and Pehkonen (1999), water from Hintelmann et al. (1997), Benoit et al. (1999, 2001), and Haitzer et al. (2002, 2003) (polysulfides in groundwater not included [Paquette and Helz, 1997: Jay et al., 20001), and soil/sediment from API (2004).

cNote that air, water, and sediment each contain multiple phases (gas. water, and solid).

d> refers to surface adsorption sites for mercury (e.g., carboxyl and sulfide sites on soot and organic material).

eR designates a natural organic ligand such as a humic and fulvic acid.

fHgS(s) is the most abundant naturally occurring form.

gHgO may be present in the solid particulate phase (Seigneur et al., 1994).

hNot significant in concentration.

iA large fraction of the mercury concentration dissolved in water droplets in the atmosphere is associated with soot particles (Petersen et al., 1995).

1.1.1 Elemental Mercury

Zero-valent elemental mercury (Hg(0)) occurs as a liquid (Hg⁰(l)), gas (Hg⁰ (g)), and dissolved constituent in water (Hg⁰(aq)) at ambient temperatures (Table 1.1.1). Owing to its relatively high (for a metal) saturation vapor pressure, the gaseous concentration in equilibrium with the liquid is predicted to be 13, 000 μg/m³ (Winter, 2003). This concentration is three orders of magnitude higher than the time-weighted average threshold limit value of 25 μg/m³ used for occupational exposure (ACGIH, 2000). It is also six orders of magnitude higher than the reported background concentration of 0.002–0.005 μg/m³, reported in Table 1.1.1 (Seigneur et al., 1994). Although slow evaporation kinetics and ventilation will necessarily reduce concentrations in indoor environments where mercury spills have occurred (Winter, 2003), measurements between 0.5 (Carpi and Chen, 2001) and 140 μg/m³ (Smart, 1986) have been observed.

The fate and transport of elemental mercury released to the atmosphere are conceptualized in Figure 1.1.1. Consistent with Table 1.1.1, Figure 1.1.1 shows that gaseous elemental mercury is the predominant species in outdoor air, accounting for as much as 98 percent of the total elemental mass. This speciation is partially attributed to mercury’s high saturated vapor pressure. (Because gaseous elemental mercury concentrations are significantly below saturation with respect to the liquid form, atmospheric releases are thermodynamically restricted from converting to Hg⁰(l).) Other chemical properties that contribute to the stability of gaseous elemental mercury in the atmosphere include its low water solubility (which prevents it from being quickly removed by precipitation via the chemical species Hg⁰(aq)) and its high ionization potential (which inhibits its oxidation to reactive gaseous mercury (Hg(II)), a species with a higher water solubility and rate of sequestration, by cloud droplets and precipitating aerosols) (Mason et al., 1994).

Figure 1.1.1 Global mercury budget and fluxes. Notes: Lamborg et al. (2002) calculated a net increase in mercury in the oceanic mixed layer on the order of 1.5% per year. Also, Mason et al. (1994) calculated a rever source to the ocean of 200 tons/yr (most of which is deposited near shore).

From a forensic science perspective, the most important transformation of elemental mercury depicted in Figure 1.1.1 is its conversion to mercuric mercury, a species that is readily deposited and potentially transported to shallow-sediment aquatic environments that are conducive to methylation. The oxidation of elemental mercury shown in the figure occurs primarily at the solid/liquid interface in fog and cloud droplets (Munthe, 1992), but some gasphase oxidation also occurs (Hall, 1995). The primary oxidants involved in the conversion include ozone (Lin and Pehkonen, 1999), hydroxyl radicals (Bergan and Rodhe, 2001), and/or halogens such as chlorine and bromine (Hedgecock and Pirrone, 2001; Ariya et al., 2002; Temme et al., 2003). Although mercuric mercury can also be reduced back to elemental mercury by sulfite (Munthe et al., 1991) or by photoreduction (Xiao et al., 1994), oxidation typically exceeds reduction. The relatively high mobility of elemental mercury is reflected in atmospheric residence times that are estimated to be in the order of several months (Sommar et al., 2001) to one year (Fitzgerald and Mason, 1997). This longevity allows atmospheric emissions to contribute to terrestrial and oceanic mercury loads on both regional and global scales.

Although not shown in the figure, liquid elemental mercury (Hg(l)) that is released directly to soil can also be converted to methylmercury. Once spilled onto soil, the dense liquid travels downward in the soil column, especially through cracks and fissures (Turner, 1992). Counteracting this process is mercury’s tendency to form balls or beads and become trapped in soil pore spaces due to its high surface tension (Turner, 1992). Elemental mercury can be oxidized to more soluble forms of mercury (i.e., Hg(II) species) in soil, which can then be transported by leaching; however, the process of oxidation is usually slow, and elemental mercury is long-lived at contaminated sites.

1.1.2 Methylmercury

The predominant oxidation state of mercury in terrestrial and aquatic environments is Hg(II), of which methylmercury (composed of ionic Hg(II) [Hg+2] and a single methyl group [CH3−] to form CH3Hg+) is a subset (Table 1.1.1). Although methylmercury is created in aqueous environments, it rarely accounts for more than a few percent of the total mercury concentration (Krabbenhoft et al., 2003). Methylmercury bioaccumulates because it is a cation, is lipophilic, and has a high affinity for sulfur-containing organic matter. These properties cause it to readily sorb to phytoplankton and particulate humic substances, which are subsequently consumed by fish or benthic macroinvertebrates. Within higher trophic levels, methylmercury tends to concentrate in muscle tissue (muscle tissue is where sulfhydryl groups of protein are located).

The production of methylmercury involves both a methylation (i.e., addition of a methyl group to a mercuric ion) and demethylation (i.e., removal of a methyl group from a methylmercury ion) aqueous reaction (Figure 1.1.1). Methylation is mediated primarily by sulfate-reducing bacteria, which are present in anoxic sediment of freshwater and estuarine environments (Compeau and Bartha, 1985; Winfrey and Rudd, 1990; Gilmour et al., 1992). Demethylation is dominated by abiotic processes in many waters, effectively removing methylmercury from the system by transforming it into elemental mercury, which volatilizes to the atmosphere (Figure 1.1.1). (Gardfeldt [2003] discussed evidence that Hg(II) is the more probable product of photolysis.) In aquatic sediment, demethylation is apparently accomplished by sulfidogens (sulfate-reducing bacteria) and methanogens (methane-producing bacteria) (Oremland et al., 1991, 1995; Pak and Bartha, 1998). Because of demethylation, methylmercury is relatively short-lived in surface water in the presence of oxygen and light; however, because methylmercury is formed continually, concentrations remain high enough to be of concern.

Net rates of methylmercury production in sediments depend on a number of factors that mainly affect the activity of mercury-methylating bacteria and/or the availability of inorganic mercury for methylation. These factors include temperature, dissolved oxygen, and sulfate concentration. In littoral (i.e., located in shallow water) sediments of lakes, temperature is positively correlated with net methylmercury production (Matilainen et al., 1991). High rates of mercury methylation in profundal surficial sediments during late summer may also result from elevated temperatures (Callister and Winfrey, 1986; Korthals and Winfrey, 1987). Dissolved oxygen is another critical factor. Numerous studies of both freshwater (Callister and Winfrey, 1986; Korthals and Winfrey, 1987; Regnell, 1990; Matilainen et al., 1991) and estuarine (Olson and Cooper, 1976; Compeau and Bartha, 1984) sediments have shown higher rates of net methylmercury production under anoxic, rather than oxic, conditions. Finally, sulfate additions have been found to stimulate mercury methylation in freshwater sediments (Gilmour et al., 1992).

The factors that primarily affect the availability of inorganic mercury for methylation include concentrations of sulfate/sulfide, dissolved and particulate organic matter, and total mercury. High sulfate concentrations may limit mercury methylation in estuarine sediments (Blum and Bartha, 1980; Compeau and Bartha, 1984, 1987; Gilmour and Capone, 1987), either due to complexation and precipitation of mercury as mercuric sulfide or because of a change in aqueous speciation that affects the supply of neutral, bioavailable mercury sulfide species to methylating bacteria (Benoit et al., 1999). Elevated concentrations of total organic matter may also reduce the net methylation potential of sediment, due to adsorption of Hg+2 on the organics and a subsequent reduction in dissolved ionic mercury available for methylation (Hammerschmidt and Fitzgerald, 2004). Finally, an increase in the total mercury concentration may be related to an increase in methylmercury (Krabbenhoft et al., 2003), although the strength of this correlation likely varies among sites. Bloom et al. (2003) found that the relative degree of mercury solubility and methylation depends on its chemical form, with sorbed or organo-chelated compounds exhibiting higher solubility than mercuric sulfide (HgS(s)).

1.2 NATURALLY OCCURRING SOURCES OF ELEMENTAL MERCURY

Terrestrial Emissions—Sources of mercury in the environment are both natural and anthropogenic. A third category (or sub-category), termed re-emission, refers to mercury that was originally deposited in water or soil by natural or anthropogenic processes but is re-emitted to the environment (Ebinghaus et al., 1999). Natural mercury emission and re-emission cannot be quantified separately, and together account for a substantial portion of the global mercury cycle.

Natural terrestrial sources of elemental mercury to the atmosphere include volcanic emissions, evasion from the subsurface crust via geothermal activity, and volatilization of mercury from soil (Table 1.2.1). Geologic regions, specifically those associated with volcanically active areas, that contain elevated concentrations of mercury (termed mercuriferous belts) contribute measurably more mercury to the atmosphere through degassing (Lindqvist et al., 1991; Ferrara et al., 2000; Coolbaugh et al., 2002), compared to areas that are not enriched in mercury (Gustin et al., 2000). This volatilization is most likely a combination of elemental mercury volatilization and reduction of Hg(II) in soil water followed by volatilization (Lindberg et al., 1998), and is positively correlated with soil and air temperatures (Gustin et al., 1995; Lindberg et al., 1995), solar radiation and soil moisture (Carpi and Lindberg, 1997), and weather conditions such as wind speed, relative humidity, and turbulence (Kim, 1995).

Table 1.2.1

Global Budget Estimates for Atmospheric Mercury (tons/yr)

aUsed reported base value for Year 1998 of Seigneur et al. (2004).

bUsed data from best-fit solution of global mass balance model of Lamborg et al. (2002).

cEstimate does not include passive, degassing volcanoes with estimated emissions of 30 tons/yr or geothermal sources with emissions of 60 tons/yr (Varekamp and Buseck, 1986).

dEstimate does not include 200 tons/yr from unmineralized soil.

eThe emission from this source category is estimated to be between 3000 and 6000 tons/yr (Rasmussen, 1994).

fTotal natural terrestrial emissions.

gTotal natural ocean emissions.

hTotal anthropogenic land emissions.

According to Seigneur et al (2004) and Lamborg et al (2002) (Table 1.2.1), total natural emissions and reemissions account for approximately one-third of the total annual mercury load to the atmosphere. Although the natural terrestrial component is estimated to constitute one-half to two-thirds of that value, there are large uncertainties in the estimates. For example, Rasmussen (1994) discussed evidence that natural emissions from terrestrial bedrock are between 3000 and 6000 tons/year. If this subsurface crustal emission were included in the global budgets of Seigneur et al (2004) and Lamborg et al (2002), natural terrestrial sources of mercury would be predicted to be equal to, or greater than, anthropogenic sources.

Oceanic Evasions—The reduction of mercuric mercury to elemental mercury and its subsequent volatilization is an important pathway by which mercury enters or re-enters the atmosphere (Table 1.2.1). The concentration of elemental mercury in surface waters tends to be supersaturated relative to the atmosphere (Vandal et al., 1991), suggesting that elemental mercury is formed in surface water. The reduction reaction itself is apparently photo-induced (Amyot et al., 1997a,b; Krabbenhoft et al., 1998). In temperate lakes, the flux of mercury from water to the atmosphere by volatilization is estimated to be 10 percent of the annual input of mercury to the lakes from atmospheric deposition (Vandal et al., 1991). Seigneur et al (1994) estimated that natural oceanic evasions account for approximately one-sixth of the total current global air emissions, with the fraction that is remobilized mercury accounting for more than half of that value (Table 1.2.1). These estimates are likely to be revised downward in the future due to evidence that mercury is cycled more rapidly in the marine boundary layer relative to the continents (Hedgecock and Pirrone, 2004).

1.3 SOURCES OF MERCURIC MERCURY AND METHYLMERCURY

Natural atmospheric sources of mercury to aquatic environments include the following: (1) mercury originally introduced as the gaseous elemental form but transformed to Hg(II) and deposited in wet precipitation or adsorbed to aerosols and (2) particulate-bound Hg(II) derived from the erosion of soil and rocks by wind. Mercury deposited from the atmosphere is subject to methylation in acquatic environments. Air Deposition—Methylmercury forms predominantly in anoxic sediment and waters of estuarine, lake, and riverine environments where sulfate-reducing bacteria actively metabolize inorganic mercuric mercury to methylated forms. (Methylmercury formation is usually negligible in surface soil, because the process requires anoxia and moisture, conditions rarely found in upland surface soil.)

Most mercuric mercury that is deposited to the earth’s surface is dissolved in wet precipitation or adsorbed to aerosols such as soot (Mason et al., 1994; Ebinghaus et al., 1999). Annual rates of atmospheric deposition of mercury (primarily from rain and snow) are a function of both the concentration of mercury in deposition (which is affected by geologic and geographic variables) and the annual rate of deposition. Estimates of the total mercury flux from wet deposition include 15 μg/m²-year in northeastern Minnesota (Glass et al., 1991), 6.8 ± 2 μg/m²-year over Little Rock Lake in Wisconsin (Fitzgerald et al., 1991), 14 μg/m²-year in Scandinavia (Iverfeldt, 1991; Bindler, 2003), and 10–20 μg/m²-year around Chesapeake Bay and western Maryland (Mason et al., 1997). The fraction of the total deposition rates attributable to natural sources is purported to be small. For example, Bindler (2003) measured mercury concentrations in 500- to 4000-year-old peat cores in south-central Sweden and determined that preindustrial natural deposition rates were in the range 0.5–1 μg/m²-year. These values are at least an order of magnitude lower than air measurements in Sweden over the past 30 years (5–30 μg/m²-year) (Munthe et al., 2001). Recent modeling studies have similarly concluded that deposition rates have increased by a factor of 2–10 during the past 200 years (Bergan et al., 1999).

Of the mercury that is deposited from wet and dry precipitation, approximately 10 to 25 percent is believed to be transported from drainage basins to lakes, either through direct deposition or run-off (Lorey and Driscoll, 1999; Swain et al., 1992). This 75–90 percent estimated retention of mercury by terrestrial soils is consistent with loads for riverine (Quemerais et al. 1999) and oceanic environments (Mason and Sheu, 2002).

Upstream Discharges—In addition to direct atmospheric deposition and run-off, natural sources of mercury to aquatic ecosystems include erosion of bedrock, leaching of soil, and discharge from wetlands. For example, mercury occurs as a trace element in geologic formations and can be eroded with fine particles by surface waters (Plouffe, 1995; Quemerais et al., 1999). In addition, although Hg(II) is strongly adsorbed to soil particles, soil organic matter, and sulfides (Schuster, 1991), a small fraction can be leached over time by groundwater (Hultberg et al., 1995; Krabbenhoft et al., 1995). Finally, wetlands discharge is of particular importance, because wetlands have been found to be natural locations of methylmercury production (St. Louis et al., 1994; Hurley et al., 1995; Brianfireun et al., 1996). Driscoll et al. (2002), for example, investigated watershed ecosystems in the Adirondack region of New York and concluded that riparian wetlands and lakes were sources of methylmercury to downstream surface waters.

The fate of upstream mercury discharges in a receiving water body is shown in Figure 1.3.1 and includes sedimentation and burial, methylation, volatilization, and downstream transport/loading. Because mercuric mercury (Hg(II)) is a weak Lewis acid, it preferentially forms stable complexes with Lewis bases, such as sulfur functional groups, that are present on the surfaces of particulate humic material (Haitzer et al., 2002, 2003). Consequently, mercury partitions strongly to organic-rich particles. Because mercury also tends to adsorb to mineral colloids (Schuster, 1991), the primary mass transport process controlling mercury transport from upstream sources is believed to be sedimentation and sediment resuspension/burial (Le Roux et al., 2001).

Figure 1.3.1 Mercury cycling in aquatic systems.

In moving waters, such as rivers, creeks, and marshes, surface sediments are often resuspended and transported, resulting in transport of mercury downstream. In quiescent waters, sediment remains in place, and the ultimate fate of mercury associated with particles in these systems is burial. Mercury in buried sediments is not remobilized by processes such as cation exchange or redox reactions that can mobilize other metals such as iron (Hurley et al., 1991; Watras et al., 1994). However, mercury in surface sediments in shallow water may be involved with the redox cycling of iron, resulting in flux into overlying water (Gill et al., 1999).

Although mass transfer mechanisms such as dissolved fluxes from sediment and volatilization likely contribute to the cycling of mercury and methylmercury (Figure 1.3.1), measurements of dissolved fluxes of mercury from sediment are highly variable, and their relevance to the overall transport and fate of mercury in aquatic systems is still unclear. Fluxes of methylmercury from Lavaca Bay, Texas, sediments, for example, varied across a range of three orders of magnitude, and fluxes at a single site varied seasonally (maximum fluxes occurred in late winter to early spring) and diurnally (Gill et al., 1999). Methylmercury flux from sediment was six times higher during dark periods (at night) than during light periods (Gill et al., 1999). The authors suggest that chemical or biological processes at the sediment/water interface, such as migration of the oxic/anoxic boundary downward during the day as oxygen is produced by photosynthesis, control methylmercury flux. Lavaca Bay is a shallow estuary (average water depth of 1 m), so sediments there would be more responsive to changing light conditions than sediments in deeper waters would be. As regards the production and volatilization of dissolved elemental mercury in fresh water, its transport is also variable, depending on photolysis rates (Gardfeldt et al., 2001) and microbial mercury oxidase activity (which additionally varies inversely with hydrogen peroxide [H2O2] diurnal patterns) (Siciliano et al., 2002).

Total mercury concentrations in upstream rivers tend to be highly variable, because suspended particles can contribute to the total mercury concentration. A survey of 60 streams in Maine that were unaffected by any known mercury sources found a median total mercury concentration of 1.5 ng/L and a 95th-percentile concentration of 4.5 ng/L (Maine Department of Environmental Protection, 1999). In Wisconsin, analysis of mercury in water samples from seven rivers found median total mercury concentrations ranging from 1.13 to 8.70 ng/L (Babiarz et al., 1998). In some cases, however, single high-flow events (Babiarz et al., 1998) or storm events in general (Mason and Sullivan, 1998; Scherbatskoy et al., 1998) can carry a major fraction of the annual load of mercury from a watershed. High flows tend to transport particulate-bound mercury. As with lakes, methylmercury is generally a small fraction of total mercury in rivers, except in watersheds with high proportions of wetlands (i.e., sites of methylmercury production) and high dissolved organic carbon (DOC) concentrations.

Background concentrations of mercury in receiving water and sediment can also vary widely depending on the type of watershed, the type of water body, water quality characteristics, and sediment quality characteristics. Total mercury concentrations in lake water not affected by point sources of mercury range from less than 1 ng/L in surface waters to greater than 40 ng/L in anoxic bottom waters (Watras et al., 1994; Jacobs et al., 1995). Total mercury concentrations in surface waters of lakes in Wisconsin range from 0.3 to 2.9 ng/L (Babiarz and Andren, 1995). Methylmercury usually constitutes a small fraction of the total mercury in surface waters (e.g., 5–15 percent in Wisconsin lake surface waters (Watras et al., 1994]), but the portion can increase to as much as 50 percent in anoxic bottom waters (Watras et al., 1994; Jacobs et al., 1995).

Background mercury concentrations in sediment depend a great deal on sediment characteristics. Given similar loading and depositional patterns, the strongest factor controlling sediment mercury concentration appears to be the organic content of the sediment (Langston, 1986; Benoit et al., 1998; Mason and Lawrence, 1999). Mercury concentration is also positively correlated with silt content in estuarine sediments (Bartlett and Craig, 1981). In a study of 34 lakes in Newfoundland that were unaffected by industrial discharge, the mean total mercury concentration was 0.039 mg/kg, and the range was 0.003 to 0.156 mg/kg (French et al., 1999). In 24 of Massachusetts’ least affected water bodies, mercury concentrations in sediment ranged from 0.008 to 0.425 mg/kg (Rose et al., 1999). Methylmercury concentrations are generally low (i.e., from less than 0.01 to 2 μg/kg) in sediment and make up less than 0.1 to 16 percent of total mercury (Gilmour and Henry, 1991). Methylmercury concentrations tend to be highest near the sediment surface, in the zone of active mercury methylation.

1.4 ANTHROPOGENIC SOURCES

1.4.1 Combustion

Anthropogenic sources of mercury in the atmosphere are varied; however, most mercury emissions are attributed to combustion. For example, combustion sources accounted for 87 percent of emissions in the United States (US EPA, 1997) and 77 percent of emissions in the world (Pirrone et al., 2001; UNEP, 2002) in 1994–1995. The principal sources of combustion include the following; coal-fired utility boilers, municipal waste combustors, coal- and oil-fired commercial/industrial boilers, and medical waste incinerators. Table 1.4.1 (Schierow, 2004) summarizes the estimated annual contributions of anthropogenic sources to the atmosphere in the United States between 1995 and 1999. As shown in the table, electric utilities are currently the largest combustion source. By contrast, state-mandated emission reductions by solid waste incinerators have dramatically reduced the contribution from this source category.

Table 1.4.1

Major Anthropogenic Mercury Emission Sources in the United States (1995–1999) (Schierow, 2004)

aUS EPA (1997).

bSeigneur et al. (2004).

cUS EPA (2004).

dMobile sources not included.

Pacyna et al. (2004) provide estimates of global total anthropogenic mercury emissions over the period 1990 to 2000 by continent (Table 1.4.2). Global mercury emissions have increased over this period while the contributions from Europe and North America have decreased. As of 2000, Asian countries contributed about 53 percent of the global mercury emissions to the atmosphere.

Table 1.4.2

Anthropogenic Mercury Emissions by Continent (1990–2000) (Pacyna et al., 2004)

Coal-Fired Power Plants—Coal contains only trace concentrations of mercury, but the high volume combusted annually results in emissions of mercury into the atmosphere that are significant relative to other US combustion sources. Of the 48 tons/year of mercury emitted from utilities, 45 tons/year is attributable to coal-fired power plants (EPRI, 2000).

The factors that affect the speciation and concentration of mercury emitted from power plants include the chemical characteristics of flue gases (CO, hydrocarbons, H2O, O2, NO, SOx), the physical characteristics of unburned carbon fly ash (abundance, size, surface area) (Niksa et al., 2001 and 2002), and numerous operational considerations (temperature, residence time, and air pollution control devices) (Senior, 2001; Kilgore and Senior, 2003). The primary release of mercury from coal combustion is Hg⁰(g); however, it can be oxidized through homogeneous (i.e., gas phase) and heterogeneous reactions (i.e., on activated carbon substrates) to produce more water-soluble species such as HgCl2(g) and solids like HgO(s) (Johansen, 2003) prior to emission. Elemental mercury and Hg(II) can also sorb onto fly ash to produce particulate-bound phases (Hg(p)). Finally, particulate control devices such as fabric filters and flue gas desulfurization systems (i.e., scrubbers) can remove up to 90 percent of mercury from bituminous coals before it enters the atmosphere (Kilgore and Senior, 2003).

Table 1.4.3 reports typical speciation profiles for coal-fired power plant emissions relative to other sources. The fraction of total emissions that is gaseous elemental mercury is important because particulate-bound and Hg(II) are more likely deposited on a local or regional scale (within 100 km of the source) (Judson, 2002), whereas elemental mercury is transported globally (Constantinou et al., 1995). Although Pacyna and Pacyna (2002) estimated that gaseous elemental mercury accounts for approximately one-half of the total atmospheric mercury releases from power plants, the United States Environmental Protection Agency’s (EPA’s) Information Collection Request has shown that elemental mercury may constitute anywhere between 0 and 95 percent of the emissions for a particular facility (Senior, 2001). In fact, the relatively low total mercury concentrations in Lake Velenje, Slovenia, have been attributed to a low abundance of oxidized mercury (5–20 percent) in the emissions from the adjacent coal-fired power plant (Kotnik et al., 2000, 2002).

Table 1.4.3

Mercury Emission Profiles from Anthropogenic Sources in 1995, in Fractional Abundance (Pacyna and Pacyna, 2002)

In addition to direct air emissions, coal-fired power plants also generate wastewater and fly ash that require disposal. Mercury in these sources can potentially be converted to methylmercury after being transported and deposited in methylating environments. An estimate of the relative global mass of mercury released as solid and aqueous waste is shown in Figure 1.4.1. Although the data are more than 15 years old (and deserve greater quantification [Trip et al., 2004]), power plant releases to soil and water likely exceed air emissions. Also, because the releases are necessarily more localized than atmospheric emissions, they merit consideration in evaluating methylmercury production in nearby water bodies. For example, without distinguishing between anthropogenic and natural sources, Rolfhus et al. (2003) concluded that methylmercury in Lake Superior was more strongly influenced by watershed sources than wet deposition.

Figure 1.4.1 Global anthropogenic inputs of mercury to the environment.

Other Combustion Sources—The second-largest combustion source in the United States is waste incineration in 1998 and 1999. The information in presented in Table 1.4.1, with emissions predominantly characterized by reactive gaseous mercury (Hg(II)) (Table 1.4.3). The relative importance of incinerator emissions to local air deposition has been demonstrated by Dvonch et al. (1999), who used modeling of multi-element tracers and analysis of wet deposition samples in southern Florida to conclude that 71–73 percent of the mercury in wet deposition at 17 sites could be accounted for by local anthropogenic sources such as municipal waste incineration and oil combustion. The majority of aerosols in the study area were similarly attributable to municipal waste incinerators (Graney et al., 2004).

Fate and Transport—The most common approach to tracking the fate of combustion emissions in the atmosphere is the implementation of numerical atmospheric mercury models that include anthropogenic sources, chemical speciation, and meteorological inputs (US EPA, 1997; Bergan et al., 1999; Seigneur et al., 2001; Ryaboshapko et al., 2002; Lin and Tao, 2003; Munthe et al., 2003; Dastoor and Larocque, 2004; Seigneur et al., 2004). For example, the Northeast Mercury Study (NESCAUM, 1998) performed model simulations with and without local, regional, and global anthropogenic sources, and concluded that approximately one-third of the total mercury depositing in New Jersey originated within the state. More recently, Cohen et al. (2004) used a computer model to conclude that coal combustion is the most important source of atmospheric mercury deposition to the Great Lakes.

Whereas methods have been developed to understand atmospheric releases, the fate of mercury introduced from combustion sources to aquatic ecosystems is more problematic. In addition to being subjected to the numerous transport and transformation processes, the methylation potential of the mercury depends on its chemical form within the methylating environment. For example, Bloom (2001) investigated the methylation of various dissolved and sediment-bound mercury species added to sediment slurries. It was found that the smallest net change in methylation occurred for mercuric sulfide, and the largest occurred for dissolved mercury spikes.

1.4.2 Manufacturing

The second-largest atmospheric mercury emission source category in the United States is manufacturing, which includes chlor-alkali (at 7 tons/yr), Portland cement (at 5 tons/yr), and pulp and paper production (at 2 tons/yr) (US EPA, 1997). In chlor-alkali plants, elemental mercury is used as a cathode in the electrolysis of sodium chloride to produce chlorine and caustic soda. In cement manufacturing, mercury is introduced as a trace element in raw materials such as limestone and fuels for cement kilns (Johansen and Hawkins, 2003). As with combustion sources, the load contribution and speciation data presented in Tables 1.4.1 and 1.4.2 should be interpreted cautiously. For example, fugitive air emissions from chlor-alkali plants likely vary with factory operating conditions (Southworth et al., 2004). Also, the 30 percent reactive gaseous mercury concentration (Hg(II)) in Table 1.4.3 may instead be closer to 2 percent (Landis et al., 2004). Finally, emissions from manufacturing sources are likely to change dramatically in the future, because the US industry has largely switched to a mercury-free membrane-cell process in chlor-alkali plants, and the European Union is poised to either decommission existing plants or convert them to mercury-free facilities by 2020 (European Commission 2004).

In contrast to fuel combustion, the largest potential releases of mercury from manufacturing sites are estimated to occur through direct or indirect discharges into waterways (Table 1.4.2). Numerous industries such as mercury-cell chlor-alkali and electroplating facilities may discharge mercury to surface water bodies under National Pollutant Discharge Elimination System (NPDES) regulations. Also, municipal treatment plants are recipients of mercury from common sources such as hospitals, dental laboratories, and university laboratories.

Typical effluent concentrations of mercury in treated wastewater are not well characterized, but current information shows that they can vary from plant to plant and within a plant. Analysis of total mercury in effluent grab samples collected at 75 municipal sewage treatment plants in Maine showed that concentrations in secondary treated effluent averaged 11.3 ng/L and ranged from 0.74 to 99.23 ng/L (Maine Department of Environmental Protection 1999). Monthly sampling (24-hour, flow-weighted, composite samples) of Metropolitan Syracuse (New York) sewage treatment plant discharge showed total mercury concentrations ranging from 17.7 to 104 ng/L and methylmercury concentrations ranging from 0.60 to 3.1 ng/L (Bigham and Vandal, 1996). In general, sewage treatment removes most of the mercury that enters the plant (Glass et al., 1990; Balogh and Liang, 1995; Mugan, 1996), primarily by sedimentation of solids. However, certain configurations such as combined sewer overflows and outside sewage lagoons may become anoxic, resulting in increased discharge of mercury and methylmercury (Bodaly et al., 1998).

Wastewater releases to waterways are particularly important, because the mercury is predominantly dissolved and/or complexed with organic material (Hsu and Sedlak, 2003), and these forms can be more readily converted to methylmercury than sediment-bound forms. For example, the higher methylation potential of dissolved species may explain mercury concentrations in sunfish from Reality Lake (Oak Ridge, Tennessee), a site highly contaminated by mercury in the sediment. After the primary source of dissolved mercury to the lake was eliminated in 1998, fish-tissue mercury concentrations decreased dramatically (Southworth et al., 2002).

Although leaching and transport of contaminated soils is a potentially important source of mercury to aquatic environments, load estimates are inherently uncertain and are likely variable among sites. Mercury is widely used in many industries, and anthropogenic sources, such as former battery recycling facilities, mercury retorting facilities, manufactured gas plants, mercury-cell chlor-alkali facilities, metering stations on natural gas pipelines, and sewage sludge applications to soil, can produce locally elevated mercury concentrations in environmental media surrounding the facilities. At mercury-contaminated sites, concentrations as high as 3000 mg/kg in soils have been reported (Turner and Southworth, 1999). This is significantly higher than the background concentration (<1 mg/kg).

1.4.3 Mining and Refining

Another important mercury emission source category is mining and refining. The specific source identified in Table 1.4.1 for the United States is gold mining, which uses mercury amalgam to recover gold during ore processing. Other mining- and refining-related atmospheric emission sources include smelting of nonferrous metals (in which mercury occurs as a trace element) (US EPA, 1997), direct mercury mining (or emissions from historical tailings containing cinnabar (HgS(s)) (Gustin et al., 2003), and petroleum refining (with an estimated contribution of 1.5 tons/yr in the United States) (Wilhelm, 2001). Although Table 1.4.1 indicates that combustion is the largest anthropogenic emission in the United States, nonferrous metal smelting operations were the largest source category in Mexico and Canada for the year 1990 (31 and 24 tons/yr, respectively) (Pai et al., 2000).

Similar to manufacturing sources, Figure 1.4.1 indicates that the largest releases of mercury from mining sites are to the soil or water. Examples include mercury mines at Mt. Amiata in central Italy (Ferrara, 1999), Almaden in Spain (Ferrara, 1999), Idrija in Slovenia (Miklavcic, 1999), Pinchi Lake in British Columbia (Turner and Southworth, 1999), and Clear Lake in California, United States (Turner and Southworth, 1999). Mercury contamination of soil has also resulted from gold and silver mining, with about 500,000 tons of mercury used from 1540 to 1900 for silver and gold production in the Americas (Nriagu, 1994). Soil contamination associated with these areas resulted from tailing deposits and erosional transport of fine particles containing mercury (Lacerda and Salomons, 1999). Not only has mercury amalgamation resulted in elevated mercury and methylmercury levels in the Sacramento River Basin, California (Domagalski, 1998), but continued use in gold mining in South America has resulted in widespread mercury contamination in the Amazon region of Brazil (Barbosa et al., 1994).

1.4.4 Miscellaneous Sources of Mercury

The primary sources of elemental mercury to indoor air are occupation-related contamination and consumer products that contain either mercury salts or liquid elemental mercury (Hg⁰). The clothing and personal gear of employees in mercury-related industries can be a source of contamination to indoor air in the home (ATSDR, 1990). In a study conducted in homes of thermometer plant workers, 10 percent of 39 randomly selected homes had indoor mercury vapor concentrations that exceeded the EPA reference concentration (RfC) (Hudson et al., 1987). More commonly, spills from the following household devices that contain liquid elemental mercury are a source of mercury to indoor air (ATSDR, 1999):

• Fluorescent light bulbs

• Thermostats

• Glass thermometers

• Barometers

• Switches in large appliances

• Natural gas pressure regulators.

Consumer products such as latex paint and chlorine-based detergents and cleansers that can decompose and volatilize to produce gaseous elemental mercury (Beusterien et al., 1991) are also potential sources. In addition to these sources, elemental mercury is sometimes released intentionally during cultural and religious practices (Riley et al., 2001).

The level of exposure to mercury vapor via inhalation is the critical health issue related to elemental mercury in indoor environments (ATSDR, 1999). Potential toxic effects include damage to the respiratory tract, heart, gastrointestinal tract, reproductive system, brain, and kidneys (ATSDR, 1999); however, the primary adverse effects are usually limited to the brain and kidneys, which have been identified as the most sensitive target organs. Either acute or chronic exposure apparently elicits a similar toxicological syndrome. The lowest-observed-adverse-effect level (LOAEL), established on the basis of central nervous system (CNS) effects during human occupational exposure, is estimated to be 26 μg/m³ (Ashe et al., 1953; Fawer et al., 1983).

Existing regulatory guidance on mercury vapor air concentrations and corresponding exposure durations that are protective of human health in occupational and residential environments is presented in Table 1.4.4. As shown in the table, the Occupational Safety and Health Administration (OSHA) has set a legally enforceable concentration ceiling for workplace exposure at 100 μg/m³. In addition, two recommended time-weighted average concentrations have been established by the National Institute for Occupational Safety and Health (NIOSH) and the American Conference of Governmental Industrial Hygienists (ACGIH). The corresponding values of 50 and 25 μg/m³, respectively, are designed to be protective of occupational exposure during an 8- or 10-hour workday (and a 40-hour work week). The 25- μg/m³ concentration is based on the LOAEL noted above.

Table 1.4.4

Summary of Mercury Vapor Guidance Values

aAgency for Toxic Substances and Disease Registry (ATSDR, 2005) unless otherwise indicated.

bOSHA recommendation.

As illustrated in Table 1.4.4, permissible concentrations in residential environments are lower than occupational settings, owing to the longer expected exposure durations. In response to concerns about mercury vapor exposure in homes, schools, or businesses due to accidental releases during the removal of mercury-seal gas-pressure regulators, the Agency for Toxic Substances and Disease Registry (ATSDR) established residential relocation and occupancy levels of 10 and 1 μg/m³, respectively (ATSDR, 2005). The upper value is used to determine whether residents can remain in a house or apartment during a spill. The lower value is used as an acceptable re-occupancy level after cleanup, and is deemed protective of pregnant women and children under the age of six. The last two values in Table 1.4.4 are the EPA and ATSDR reference minimal risk levels, which were calculated using the LOAEL studies as a basis. ATSDR considers the RfC and chronic minimum risk level to equivalently represent the daily human exposure concentration that can occur without appreciable risk of adverse, non-cancer health effects.

In 1999 and 2000, ATSDR’s Hazardous Substances Emergency Events Surveillance system reported more than 300 incidents of mercury spills per year in the northeastern United States. Evidence shows that very small quantities of elemental mercury spilled from consumer products such as thermometers can result in indoor air mercury concentrations that exceed the standards and action levels presented in Table 1.4.4 (Smart, 1986). In addition, a random sampling of indoor sites in the New York metropolitan area revealed that 11 of the 12 investigated residences had mercury vapor concentrations that were higher than outdoor levels (Carpi and Chen, 2001). In one case, a concentration of 0.045 μg/m³ was observed in a bathroom following a spill that had occurred 16 years prior.

Examples of litigation involving indoor mercury releases include the NICOR Power Company in Chicago, which was ordered to inspect more than 200,000 households following the discovery that liquid elemental mercury was accidentally spilled during the routine replacement of natural gas pressure regulators (NY Times, 2000). No systematic environmental forensic technique was employed in the related investigations. Instead, breathing-zone mercury vapor levels were measured, without direct assessment of all possible sources.

1.5 DETECTING MERCURY IN INDOOR AIR

Several methods are available for assessing mercury vapor concentrations in indoor air. Some of the more common techniques are summarized in Table 1.5.1. According to procedures developed by ATSDR for assessing indoor air mercury contamination from natural gas mercury pressure regulators (ATSDR, 2005), air monitoring should occur during initial characterization, confirmation, cleanup, and post-remediation. The methods described below each offer advantages in terms of portability, detection limits, and response time during the investigation.

Table 1.5.1

Summary of Residential Mercury Vapor Measurement Techniques

Jerome®—The Jerome® 431-X mercury vapor analyzer is a portable, hand-held device that allows rapid, screening-level assessments. The instrument uses a gold film sensor to collect mercury from a known volume of indoor air. The measured change in electrical resistance in the sensor is compared to a reference sensor to produce a quantifiable mercury vapor concentration. The circuit in the instrument is automatically zeroed after each reading, and the gold film is periodically cleaned of mercury by electric heating. Although the Jerome® can be used to rapidly characterize indoor air, its detection limit is just below the threshold concentration for residential occupancy (Table 1.4.4). Because the Jerome® is also subject to interferences from volatile gases that can adsorb to the gold film, Bigham (2004) used a Lumex® for confirmation when the Jerome® recorded values equal to or greater than 3 μg/m³ in samples collected 5 in. above the floor.

Lumex®—The Lumex® RA-915+ mercury analyzer is a portable device, with many advantages over the Jerome® 431-X. Among its advantages, the instrument is an atomic absorption spectrometer with a Zeeman background correction for removing interferences. It offers a lower detection limit than the Jerome® and also generates real-time data.

NIOSH 6009—This method involves the collection of an 8-hour air sample by passing room air drawn by a pump through a solid sorbent material in a glass collection tube. NIOSH recommends that the sampling height be approximately 3 ft (corresponding to the height of a small child), and the sample pump be located in the area of the home where the maximum exposure to mercury contamination is expected to occur. Analysis of the sorbent in the laboratory involves digesting the sorbent in concentrated acids (HCl and HNO3), reducing the mercury using stannous chloride, and finally quantifying the mercury vapor concentration using cold-vapor atomic absorption spectroscopy.

Dosimeters—Both NIOSH Method 6009 and passive dosimeters are OSHA-approved methods for measuring mercury vapor levels in workplace environments (OSHA, 2005). The collection medium is typically hopcalite (oxides of manganese and copper). Although no interferences are listed, disadvantages of passive dosimeters include a sampling rate dependence on face velocity. Consequently, dosimeters are not recommended where the air velocity is greater than 229 m/min.

Tekran®—Ultra-low detection limits are possible with the Tekran® Model 2537A mercury vapor analyzer by first preconcentrating mercury from the air using a pure gold adsorbent that amalgamates the mercury. Cold-vapor atomic fluorescence spectroscopy is then used to quantify the desorbed vapor using a method developed by Temmerman et al. (1990). Tekran® instruments are potentially useful tools in mercury source evaluations. This is particularly true in outdoor air, where background mercury vapor concentrations are typically in the range of 0.002 − 0.005 μg/m³ (Seigneur et al., 1994).

Biological Sampling Methods—An additional method for identifying exposure is measuring mercury concentrations in biological samples. Of the various biological media, the concentration of mercury in urine is believed to be the most accurate indicator of chronic exposure to mercury vapor (blood better approximates short-term exposure and is more sensitive to dietary pathways) (ATSDR, 1999). Tsuji et al. (2003) found that mercury levels in air and urine are correlated below 0.050 mg/m³; however, chronic exposure to mercury vapor concentrations below 0.010 mg/m³ corresponds to urinary mercury concentrations that are indistinguishable from background urinary mercury levels.

1.6 MERCURY FORENSICS

Investigation of mercury in an environmental forensic context has had limited application. Perhaps the most well-developed methods are the use of radioisotopes to track the fate of mercury in terrestrial and aquatic environments. Mercury has seven naturally occurring (nonradioactive), stable isotopes, listed in Table 1.6.1. Different stable isotopes can be added to soils, sediment, or water to follow its movement, including chemical transformation. A major experiment, known as METAALICUS (Mercury Experiment To Assess Atmospheric Loading In Canada and the United States), is underway in Ontario, Canada, to evaluate the behavior of mercury contributed by atmospheric deposition. The study is being conducted in two phases. Background studies (Phase 1) began in 1999. Phase 2, the artificial loading of mercury isotopes to the whole ecosystem, began in 2001 and will continue through at least 2005 (EPRI, 2005). Different mercury isotopes (upland, ²⁰⁰Hg; wetland, ¹⁹⁸Hg; and lake, ²⁰²Hg) have been applied at a rate of about three times background level (25 μg/m²-yr) (METAALICUS, 2003). The different isotopes can be detected in sediments and as methylmercury in fish tissue to distinguish it from natural background and to determine the original source of the mercury. The objective of the drainage-basin-scale experiment is to determine the response of the ecosystem to a change in atmospheric mercury loading.

Table 1.6.1

Naturally Occurring Mercury Isotopesa

aIUPAC (1998).

A second application of environmental forensics to mercury is to evaluate the rate of volatilization of spilled elemental mercury. Spillage of elemental mercury in homes, schools, or the workplace is relatively common, and most state health departments provide advice on actions to take in response to a spill. Spillage of mercury has been a particular problem during removal of natural gas regulators that contain mercury as part of a pressure relief mechanism. Originally installed, typically in basements, in the 1940s–1960s, mercury-containing regulators have been replaced by regulators with a spring pressure-release mechanism. As noted earlier, the issue of mercury spilled from regulators became a much-publicized issue for Nicor Gas, EPA, and the State of Illinois in 2000. About the same time, similar problems developed for other gas utilities in the Chicago area and in Detroit. Most recently, a class action lawsuit related to spillage of mercury in homes was filed against a gas utility in Philadelphia in 2005 and was subsequently settled.

While the action levels for mercury in air have been set (Table 1.4.4), the relationship between the mass of mercury spilled and the resulting concentration of mercury vapor in air has not been well established. The typical problem related to mercury spilled from gas pressure regulators is to hindcast the historical mercury vapor concentrations in the basement or entire home for a spill that occurred up to ten years ago. Commonly, an effort would have been made to recover the spilled mercury, but small amounts of mercury remained in floor cracks, under carpet, and under or absorbed into wood molding. While the evaporation of mercury vapor from a spherical drop or bead can be readily calculated, it is unclear how applicable these calculations are to the typical form of spilled mercury. Winter (2003) evaluated the evaporation of mercury from two beads, one about 0.2 g and 3 mm in diameter, and the other a larger, approximately 2 g, nonspherical drop. Based on theory, Winter (2003) calculated the evaporation rate of the smaller drop to be about 11 μg/hr at 20°C and estimated that the drop would evaporate completely in about 3 years. Such rapid evaporation seems inconsistent with observations in homes where spills occurred (Bigham, 2004). Through repeated weighing of the two mercury drops, Winter (2003) found that the rate of evaporation decreased significantly, from an initial rate of 7 μg/hr for the small drop to negligible weight loss after 2.5 months. Evaporation of the larger drop decreased from an initial rate of 6.2 μg/hr to 0.5 μg/hr six months later.

The explanation for the decrease in the evaporation rate comes from an unusual source—astronomy. Telescopes with mirrors of elemental mercury have been in use for more than 20 years (Borra, 1982). Known as liquid mirror telescopes (LMTs), these instruments take advantage of the fact that spinning liquid takes on the form of a parabola, ideal for telescope mirrors. Current LMTs are about 2–4 m in diameter. An LMT constructed by the National Aeronautics and Space Administration (NASA) is

3 m in diameter and contains 14 L of elemental mercury (http://www.astro.ubc.ca/LMT/Nodo/index.html). The mercury spreads over the parabolic-shaped support to a thickness of 1.6 mm as the support spins at a rate of 10 rpm. LMTs are limited to pointing straight up or tilting only a few degrees from vertical, but are much less expensive to build and provide superior reflectivity than conventional telescopes.

Extensive evaluation of the stability of the reflective surface has shown that elemental mercury oxidizes in the presence of water vapor and impurities on the time scale of a few hours. The mercuric oxide surface layer reduces the evaporation rate by five orders of magnitude, so that after two weeks, evaporation is negligible (Mulrooney, 2000).

Based on the work of Mulrooney (2000) and Winter (2003), it is reasonable to expect that mercury spilled in homes or the workplace will remain indefinitely if not cleaned up or periodically disturbed. Because of oxidation, the evaporation rate of elemental mercury hidden in floor cracks or elsewhere will decline to negligible levels over the course of a few months. If disturbed by foot traffic, for example, the oxidized crust can be destroyed and the oxidation process must begin again.

A further consequence of mercury oxidation and the resulting negligible long-term evaporation is that forensic techniques based on isotope fractionation are unlikely to be useful for determining the timing of a mercury spill. The rates of evaporation of mercury’s stable isotopes (Table 1.6.1) are slightly different because of the slight difference in atomic mass. If the isotopic ratio of the mercury originally spilled was known, it might be possible to determine how long the mercury had been evaporating by determining the degree of isotopic fractionation in the spilled sample. Unfortunately, this technique cannot be used when the rate of evaporation (fractionation) is variable or negligible.

References

ACGIH. Mercury, elemental and inorganic. Threshold limit values for chemical substances and physical agents and biological exposure indices. Cincinnati, OH: American Council of Governmental and Industrial Hygienists; 2000.

Amyot, M., Lean, D. R.S., Mierle, G. Photochemical formation of volatile mercury in high arctic lakes. Environ. Toxicol. Chem.. 1997; 16:2054–2063.

Amyot, M., Mierle, G., Lean, D. R.S., McQueen, D. J. Effect of solar radiation on the formation of dissolved gaseous mercury in temperate lakes. Geochim. Cosmochim. Acta. 1997; 61:975–988.

Unpublished API Publication API. Mercury: Chemistry, fate, toxicity, and wastewater treatment options. Washington, DC: American Petroleum Institute; 2004.

Ariya, P., Khalizov, A., Gidas, A. Reactions of gaseous mercury with atomic and molecular halogens: Kinetics, product studies, and atmospheric implications. J. Phys. Chem. A. 2002; 106:7310–7320.

Ashe, W. F., Largent, E. J., Dutra, F. R., Hubbard, D. M., Blackstone, M. Behavior of mercury in the animal organism following inhalation. Ind. Hyg. Occup. Med.. 1953; 17:19–43.

ATSDR. Final report. Technical assistance to the Tennessee Department of Health and Environment. Mercury exposure study Charleston, Tennessee. Atlanta, GA: Agency for Toxic Substances and Disease Registry, US Department of Health and Human Services, Public Health Service, Centers for Disease Control; 1990.

ATSDR. Toxicological profile for mercury. Agency for Toxic Substances and Disease Registry, US Department of Health and Human Services, Public Health Service; 1999.

Available at ATSDR, Suggested action levels for indoor mercury vapors in homes. Agency for Toxic Substances and Disease Registry, 2005. http://www.publichealth.columbus.gov/Asset/iu_files/Indoor_Air_Table.pdf [Accessed May 11, 2005].

Babiarz, C. L., Andren, A. W. Total concentrations of mercury in Wisconsin (USA) lakes and rivers. Water Air Soil Pollut.. 1995; 83:173–183.

Babiarz, C. L., Hurley, J. P., Benoit, J. M., Shafer, M. M., Andren, A. W., Webb, D. A. Seasonal influences on partitioning and transport of total and methylmercury in rivers from contrasting watersheds. Biogeochemistry. 1998; 41:237–257.

Balogh, S., Liang, L. Mercury pathways in municipal wastewater treatment plants. Water Air Soil Pollut. 1995; 83:173–183.

Barbosa, A. C., Boischio, A. A., East, G. A., Ferrari, I., Goncalves, A., Silva, P. R.M., da Cruz, T. M.E., Mercury contamination in the Brazilian Amazon: Environmental and occupational aspects. Water Air Soil Pollut. 1994; 80:109–121

Bartlett, P. D., Craig, P. J. Total mercury and methyl mercury levels in British estuarine sediments-II. Water Res.. 1981; 15:37–47.

Benoit, J. M., Gilmour, C. C., Mason, R. P., Reidel, G. S., Reidel, G. F. Behavior of mercury in the Patuxent River estuary. Biogeochemistry. 1998; 40:249–265.

Benoit, J. M., Gilmour, C. C., Mason, R. P. Sulfide controls on mercury speciation and bioavailability to methylating bacteria in sediment pore water. Environ. Sci. Technol.. 1999; 33:951–957.

Benoit, J. M., Mason, R. P., Gilmour, C. C., Aiken, G. R. Constants for mercury binding by dissolved organic matter isolates from the Florida Everglades. Geochim. Cosmochim. Acta.. 2001; 65:4445–4451.

Bergan, T., Rodhe, H. Oxidation of elemental mercury in the atmosphere: Constraints imposed by global scale modeling. J. Atmos. Chem.. 2001; 40:191–212.

Bergan, T., Gallardo, L., Rodhe, H. Mercury in the global troposphere: A three-dimensional model study. Atmos. Environ.. 1999; 33:1575–1585.

Beusterien, K. M., Etzel, R. A., Agocs, M. M., Egeland, G. M., Socie, E. M., Rouse, M. A., Mortensen, B. K. Indoor air mercury concentrations following application of interior latex paint. Arch. Environ. Contam. Toxicol.. 1991; 21:62–64.

Bigham, G. N., Assessment of exposure to mercury vapor in indoor air from spilled elemental mercury. Abstract: Proc. 7th International Conference on Mercury as a Global Pollutant, Ljubljana, Slovenia, June 27–July 2, 2004, 2004.

Bigham, G. N., Vandal, G. M. A drainage basin perspective of mercury transport and bioaccumulation: Onondaga Lake, New York. Neurotoxicology. 1996; 17(1):279–290.

Bindler, R. Estimating the natural background atmospheric deposition rate of mercury utilizing ombrotrophic bogs in Southern Sweden. Environ. Sci. Technol.. 2003; 37:40–46.

Bloom, N. S. Assessment of ecological and human health impacts of mercury in the Bay-Delta watershed. CALFED Bay-Delta Mercury Project, Final Report. 2001.

Bloom, N. S., Preus, E., Katon, J., Hiltner, M. Selective extractions to assess the biogeochemically relevant fractionation of inorganic mercury in sediments and soil. Anal. Chim. Acta.. 2003; 479:233–248.

Blum, J. E., Bartha, R. Effect of salinity on methylation of mercury. Bull. Environ. Contam. Toxicol.. 1980; 25:404–408.

Bodaly, R. A., Rudd, J. W.M., Flett, R. J. Effect of urban sewage treatment on total and methyl mercury concentrations in effluents. Biogeochemistry. 1998; 40:279–291.

Borra, E. F. The liquid-mirror telescope as a viable astronomical tool. Royal Astron. Soc. Can. J.. 1982; 76:245–256.

Brianfireun, B. A., Heyes, A., Roulet, N. T. The hydrology and methylmercury dynamics of a Precambrian Shield headwater peatland. Water Resources Res. 1996; 32:1785–1794.

Callister, S. M., Winfrey, M. R. Microbial methylation of mercury in upper Wisconsin River sediments. Water Air Soil Pollut.. 1986; 29:453–465.

Carpi, A., Chen, Y-F. Gaseous elemental mercury as an indoor air pollutant. Environ. Sci. Technol.. 2001; 35:4170–4173.

Carpi, A., Lindberg, S. E. Sunlight-mediated emission of elemental mercury from soil amended with municipal sewage sludge. Environ. Sci. Technol.. 1997; 31:2085–2091.

Cohen, M., Artz, R., Draxler, R., Miller, P., Poissant, L., Niemi, D., Ratte, D., Deslauriers, M., Duval, R., Laurin, R., Slotnick, J., Nettesheim, T., McDonald, J. Modeling the atmospheric transport and deposition of mercury to the Great Lakes. Environ. Res.. 2004; 95:247–265.

Compeau, G., Bartha, R. Methylation and demethylation of mercury under controlled redox, pH, and salinity conditions. Appl. Environ. Microbiol.. 1984; 48:1203–1207.

Compeau, G., Bartha, R. Sulfate-reducing bacteria: Principal methylators of mercury in anoxic estuarine sediment. Appl. Environ. Microbiol.. 1985; 50:498–502.

Compeau, G., Bartha, R. Effect of salinity of mercury-methylating activity of sulfate-reducing bacteria in estuarine sediments. Appl. Environ. Microbiol.. 1987; 53:261–265.

Coolbaugh, M. F., Gustin, M. S., Rytuba, J. J. Annual emissions of mercury to the atmosphere from natural sources in Nevada and California. Environ. Geol.. 2002; 42:338–349.

Constantinou, E., Wu, X. A., Seigneur, C. Development and application of a reactive plume model for mercury emissions. Water Air Soil Pollut.. 1995; 80:325–335.

Dastoor, A. P., Laroque, Y. Global circulation of atmospheric mercury: A modeling study. Atmos. Environ.. 2004; 38:147–161.

Domagalski, J. Occurrence and transport of total mercury and methylmercury in the Sacramento River Basin, California. J. Geochem. Explor.. 1998; 64:277–291.

Preprints of Extended Abstracts Driscoll, C. T., Kalicin, M., McLaughlin, E., Liussi, C., Newton, R., Munson, R., Yavitt, J., Chemical and biological control of mercury cycling in upland, wetland, and lake ecosystems in the Adirondack region of New York. Mercury in the Environment: Assessing and Managing the Multimedia Risks; 42. American Chemical Society, Orlando, FL, 2002:809–812.

Dvonch, J. T., Graney, J. R., Keeler, G. J., Stevens, R. K. Use of elemental tracers to source apportion mercury in south Florida precipitation. Environ. Sci. Technol.. 1999; 33:4522–4527.

Ebinghaus, R., Tripathi, R. M., Wallschlager, D., Lindberg, S. E. Natural and anthropogenic mercury sources