Concentration of PAH in sediment pore water
Polycyclic aromatic hydrocarbons in solution in ambient water or pore water of sediments are much more bioavailable and toxic than those adsorbed to particles (particularly combustion soot) (Gustafsson et al. 1997) or associated with a nonaqueous phase liquid (NAPL; e.g., petroleum, creosote, or coal tar) (Pastorok et al. 1994). The dissolved phase of PAHs in sediments can be estimated based on equilibrium partitioning theory (Hansen et al. 2003).
Nonpolar organic chemicals, such as PAHs, have low aqueous solubilities and high affinities for adsorption to sediment and organic particles and absorption (bioconcentration) by living organisms (Neff 2002). Most of the higher molecular weight pyrogenic PAHs entering aquatic environments are sorbed to soot (the particulate fraction of smoke and engine exhaust). Petrogenic lower molecular weight PAHs may enter the water from the vapor phase in rainfall or dry fallout. They quickly adsorb to the organic phase of suspended particles and are deposited with them in sediments (Neff 2002).
Petrogenic PAHs from petroleum and pyrogenic PAHs from creosote and coal tar in sediments may be complexed with the colloidal and particulate organic fraction of sediment or associated with a NAPL, an oil phase, or an oil coating on sediment particles. Because the affinity of hydrocarbons is higher for the oil phase than for the sediment organic matter and sediment porewater phases, partitioning of hydrocarbons into sediment porewater is controlled primarily by the affinity of the hydrocarbons for the NAPL phase (Zemanek et al. 1997). Thus, in estimating the partitioning of PAHs between the NAPL phase (petroleum, coal tar, creosote) and dissolved phase, PAH concentration should be normalized to some measure of total hydrocarbons.
The PAHs in sediments are distributed between the dissolved (porewater) and particulate and NAPL phases of the sediment according to their relative affinities for the three phases. This distribution can be expressed as an organic carbon/water partition coefficient (Koc) or an oil/water partition coefficient (Koil) (Lee et al. 1992a; Neff and Sauer 1995; Di Toro and McGrath 2000; Hansen et al. 2003). Both partition coefficients are similar to the octanol/water partition coefficient (Kow) that is used frequently to model bioconcentration of nonpolar organic compounds from water by aquatic animals (Connell 1993; Neff 2002). The Koc for most nonpolar organic chemicals and colloidal/particulate organic matter in sediments is lower than the Kow (Karickhoff 1981; Di Toro et al. 1991; Neff 2002), whereasthe ow Koil for PAHs in most refined petroleum products and liquid coal tars is about the same as or higher than the Kow and tends to increase with average molecular weight of the NAPL material (Shiu et al. 1990; Lee et al. 1992a, 1992b).
However, high molecular weight, petrogenic PAHs in coal particles and asphalt, pyrogenic PAHs in soot or coal tar, and related viscous liquids often are bound to sediment particles more strongly than predicted by equilibrium partitioning theory. Mitra et al. (1999) reported high, invariant log Kocs for PAHs in sediments from the Elizabeth River (VA, USA), which is heavily contaminated with creosote-contaminated wood particles. Polycyclic aromatic hydrocarbons in sediments of urban estuaries—such as the Tamar River (UK) (Readman et al. 1987); Boston Harbor (MA, USA) (McGroddy and Farrington 1995); and San Francisco Bay (CA, USA) (Maruya et al. 1996)—often are more tightly bound to sediment particles (have higher log Kocs) than predicted. The desorption rate of PAHs from sediments decreases with duration of sediment contamination (Kraaij et al. 2002). Polycyclic aromatic hydrocarbons also are tightly bound to coal (Ghosh et al. 2001). These tightly bound PAHs do not partition effectively into the aqueous phase of porewater.
Table Table 3.. Log Kow' freshwater solubility, and estimated acute and chronic toxicity of PAH frequently found in crude and refined petroleum. Solubility and toxicity values are micrograms per liter (μg/L ppb). Log Kow values and solubilities are from Mackay et al. (1992), Neff and Burns (1996), and Ran et al. (2002)
The use of Kow or Koc tends to overestimate concentrations of dissolved PAHs in porewater of sediments contaminated primarily with pyrogenic PAHs but should give a reasonable upper-limit estimate of dissolved-phase PAHs in porewater of petroleum-, creosote-, or coal tar-contaminated sediments if the oil or other NAPL phase is still liquid and in physical contact with sediment porewater. The NAPL, particularly if it is crude oil or coal tar, may develop a surface “skin” of resins-asphalthenes or other high molecular weight polar compounds, decreasing NAPL/water partitioning (Ghoshal et al. 2004). A NAPL or oil-filled pores also may substantially decrease the permeability of the soil or sediment, decreasing the effective NAPL/water interface and limiting accessibility of the PAH to partitioning into sediment porewater. Empirically determined Kd (particle/water partition coefficients) are best for estimating sediment/water partitioning of pyrogenic PAHs or PAHs from weathered crude oil.
Values for the octanol/water partition coefficient (Kow) have been published for a large number of PAHs (Mackay et al. 1992; Durell et al. 2004). The most accurate current values for log Kow for several PAHs of environmental concern are summarized in Table 3. Koc can be estimated from Kow (Karickhoff et al. 1979), but Kd must be determined empirically on a site-specific basis. The concentration of a PAH in sediment porewater in equilibrium with its concentration in the bulk sediment can be estimated by the simple equation
where Cw is the concentration of the PAH in solution in sediment porewater, Cs is the concentration of the PAH in bulk sediment (measured as concentration per unit mass of sediment organic carbon or concentration per unit mass of total petroleum hydrocarbons [TPH]), and Kx is the sediment organic matter/water partition coefficient (Koc) or sediment particle/water partition coefficient (Kp) for the PAH. The K for pyrogenic PAHs associated primarily with combustion soot requires a variation on Equation 1 to account for the high affinity of soot particles for PAHs (Bucheli and Gustafsson 2000; Cornelissen and Gustafsson 2004).
Concentrations of individual PAHs in bulk sediment, expressed as μg/g dry sediment, should be normalized to the concentration of total extractable (C8+) petroleum hydrocarbons (TPH) in sediments, determined by gas chromatography/flame ionization detection (Sauer and Boehm 1995), if the source of the PAHs in sediments is primarily petroleum. If the PAHs in the sediment are primarily pyrogenic, then concentrations of PAHs should be normalized to sediment total organic carbon if the sediments contain high concentrations (several percent) of particulate organic matter or if a pyrogenic NAPL (e.g., creosote or coal tar) is present. TPH or total extractable organic matter often is the best parameter for normalizing PAH concentrations in urban or industrial sediments, even when the PAHs are primarily from pyrogenic sources because the total extractables analysis quantifies mainly the nonpolar organic fraction in bulk sediment that often is the most important in adsorbing PAHs. This calculation is repeated for all PAHs analyzed in sediment and is the basis for an estimate of the maximum concentration of total PAHs in solution in sediment porewater. Where the estimated concentration of a PAH exceeds its aqueous solubility, the aqueous solubility is used as the water concentration.
Toxicity of dissolved PAH mixtures to aquatic organisms
A search of the U.S. EPA Toxicity Information Retrieval (USEPA 1997) database identified more than 300 values for the acute toxicity (median lethal concentration, LC50) of aromatic hydrocarbons to freshwater and marine invertebrates and fish. The search excluded LC50 concentrations greater than the aqueous solubility of the particular hydrocarbon. Suitable aquatic toxicity data were found for 25 aromatic hydrocarbons, including 14 PAHs (Table 4). Log geometric mean acute toxicity values (in mM/L) for the aromatic hydrocarbons were regressed against log Kow. The regression has a high correlation (r2 = 0.885) and the form
This equation was used to estimate the acute toxicity of each of the PAHs analyzed in sediment (Table 3). Equation 2 is similar to that developed by McCarty et al. (1992) to estimate the toxicity to freshwater fish of a large number of nonpolar organic compounds. Equation 2 considers toxicity data for freshwater and marine invertebrates and fish and applies only to aromatic hydrocarbons.
The chronic toxicity of each PAH was estimated by dividing the acute value by an acute/chronic ratio of 5. An acute/chronic ratio of 5 represents a conservative estimate of the acute/chronic ratio for aromatic hydrocarbons. For example, Suter and Rosen (1988) evaluated the comparative acute and chronic toxicity of several chemicals to marine fish and crustaceans. Acute/chronic ratios for aromatic hydrocarbons calculated from their data are between 2 and 4.
The estimated concentration of each PAH in solution in sediment porewater was divided by its chronic toxicity value to derive a HQ. Hazard quotients for all of the PAHs detected in sediment were summed to produce a HI for total PAHs:
Equations 3 and 4 are based on the reasonable assumptions that the dissolved PAHs are much more bioavailable and toxic than adsorbed PAHs (Neff 2002) and the toxicities of individual PAHs in a mixture in solution are additive (Warne et al. 1989; Di Toro and McGrath 2000; Hansen et al. 2003; Landrum et al. 2003).
Toxicity of PAH assemblages in sediments
Log Kow values for the PAHs most frequently analyzed in freshwater and marine sediments increase with molecular weight from 3.37 for naphthalene to 8.0 for C4-chrysenes (Table 3). Log Koc values estimated by the log Koc/log Kow regression of Karickhoff (1981) are slightly lower than the log Kow values in Table 3, ranging from 2.94 for naphthalene to 6.68 for benzo[ghi]peylene. Values for log Koil, based on the regression of Lee et al. (1992a) for diesel fuel PAHs, are slightly higher than the log Kow values, ranging from 3.81 for naphthalene to 6.72 for benzo[ghi]perylene. The actual value of log Koil varies with the “average molecular weight” of the oil and, therefore, is different for different crude and refined petroleum products and changes with oil weathering (Lee et al. 1992a, 1992b; Shiu et al. 1990). The value of log Koil for a particular PAH decreases as the average molecular weight and density of the bulk oil increases, in agreement with Raoult's Law (Lane and Loehr 1995). Thus, Kow is a reasonable, conservative coefficient to use for estimating the dissolved concentrations of PAHs associated with sediments, most of which contain both weathered petrogenic and pyrogenic PAHs. The use of log Kow may result in approximately 2-fold under- or overestimation of the true concentration of a PAH in solution in equilibrium with the NAPL phase (Shiu et al. 1988).
Table Table 4.. Geometric mean toxicity (LC50 with 48-h or longer exposure) for PAH, based on available data from the AQUIRE database (USEPA 1997)
Polycyclic aromatic hydrocarbon solubility in freshwater decreases with increasing PAH molecular weight (Table 3). The solubility of some alkyl-PAHs is greater than that of the parent PAH, possibly reflecting steric effects of the alkyl carbons. Solubility tends to decrease with increasing seawater salinity and decreasing water temperature (Neff 2002). For example, the solubility of phenanthrene in fresh and salt water at 25°C is 1,080 and 644 μg/L, respectively (Eastcott et al. 1988). The solubility of anthracene in freshwater decreases from 56.5 μg/L at 29.1°C to 15.5 μg/L at 8.9°C (Reza et al. 2002). At all salinities and temperatures, anthracene is much less soluble than its isomer, phenanthrene. These physical properties of PAH affect their bioavailability and toxicity to freshwater and marine organisms.
The measured acute toxicity of aromatic hydrocarbons increases (LC50 decreases) with increasing PAH molecular weight and log Kow (Table 4; Hansen et al. 2003). The estimated acute toxicity values follow the same trend (Table 3). Estimated chronic values for PAHs range from 970 μg/L (ppb) for naphthalene to 0.01 ppb for C4-chrysenes (Table 3). For anthracene and PAHs with molecular weights of 228.3 (chrysene and benz[a] anthracene) or higher, the acutely lethal concentration approaches, or is higher than, the single phase aqueous solubility. Saturated solutions of these highly nonpolar PAHs are not acutely toxic to aquatic organisms.
We selected sediment PAH data from a recent study of the sources of PAHs in sediments of the Wycoff/Eagle Harbor Superfund site in the state of Washington, USA, to demonstrate the method described above for estimating the aquatic toxicity, measured as HI, of sediment-bound PAHs (Table 5). Two sediment samples from Eagle Harbor in Puget Sound were used, one heavily contaminated with creosote and the other contaminated with PAHs from urban runoff and deposition of pyrogenic PAHs from combustion sources. The PAH data for the two sediment samples used in this example are plotted in the PCA plot for the site data (Figure 2) and show that one sample has a clear creosote signature and the other has a clear urban runoff signature.
The creosote-contaminated sediment contained 27,441 μg/g (ppm) of TPH and 17,283 ppm of total PAH. The urban runoff sediment contained 212 ppm of TPH and 25 ppm of total PAH. Total PAH concentrations were much higher than the “high” value reported in the National Status and Trends database of 2.18 ppm of total PAH (24 parent PAH and alkyl homologue groups in sediments) (Daskalakis and O'Connor 1995), indicating that sediments from Eagle Harbor were highly contaminated with PAHs. Daskalakis and O'Connor (1995) identified Eagle Harbor as the location of some of the most heavily contaminated sediments in Puget Sound.
The sediment PAH concentrations were normalized to TPH concentration for calculation of HIs. TPH-normalized PAH concentrations were 629,808 μg PAH/g TPH and 115,655 μg/g in the creosote-contaminated and urban runoff-contaminated sediments, respectively. Estimated concentrations of total PAH in solution in sediment porewater in equilibrium with the two sediments were 17,190 μg/L (ppb) and 10,216 μg/L, respectively (Table 5). However, estimated concentrations of several PAHs in solution in water in equilibrium with the creosote-contaminated sediment were in excess of their single-phase aqueous solubilities (Table 3). Actual concentrations of these PAHs in solution in sediment pore water would not exceed their aqueous solubilities. Therefore, the aqueous solubilities of these PAHs were used as the exposure concentrations.
The estimated concentration of each PAH in solution was divided by its chronic toxicity value (Table 3) to obtain an HQ. The HQs for the PAHs in the creosote-contaminated sediment pore water ranged from 0.1 to 32.1 (C4-phenanthrenes). The sum of HQs (the HI) for this sediment was 250.
The estimated concentration of only benzo[a] pyrene in solution exceeded its water solubility for the urban runoff sediment sample. The difference was small, so no adjustment was necessary. Estimated HQs for PAH in the urban runoff-contaminated sediment porewater ranged from 0.1 to 6.9 (naphthalene), and the estimated HI was 64.
An HI value greater than 1 indicates that the porewater contains in solution a concentration of total PAHs in excess of its estimated chronic toxicity to aquatic animals (Ozretich et al. 2000). Both sediments had HIs substantially greater than 1, suggesting that both sediments would be toxic to the benthic fauna of Eagle Harbor.
Table Table 5.. Hazard quotients (HQ) and hazard indices (HI) for two sediment samples collected from Eagle Harbor, Washington, USA. The weathered creosote-contaminated sediment contained 17,283 μg/g dry wt (ppm) total polycyclic aromatic hydrocarbons (PAH) and 27,441 ppm total petroleum hydrocarbons (TPH). The sediment sample contaminated with urban runoff/fallout PAH contained 25 ppm total PAH and 212 ppm TPH. Concentrations in parentheses are freshwater solubilities. PAH data from Stout et al. (2001a)
|Total PAH||629,808||17,190|| ||115,655||10,216|| |
|HI|| || ||250|| || ||64|
Sediment quality guidelines, based on toxicity to sediment-dwelling marine animals, have been developed for total PAH in marine sediments (Long et al. 1995). The effects range low (ERL) and effects range median (ERM) concentrations for total PAH are 4.022 μg/g dry wt and 44.792 μg/g, respectively. The ERL is the concentration in bulk sediment below which toxicity to benthic organisms is unlikely; the ERM is the concentration above which effects are likely. Concentrations between the ERL and ERM may be toxic and may require additional evaluation. At Eagle Harbor, the creosote-contaminated sediment contained more than 17,000 μg/g of total PAHs, more than 384 times higher than the ERM concentration. The high HI value and substantial exceedence of the ERM value for this sediment indicates that it is likely to be highly toxic to benthic fauna.
Approximately 34% of the HI for the creosote-contaminated sediment at Eagle Harbor was attributable to 4-ring+ (fluoranthene and higher) PAHs, most of which are pyrogenic (Table 5). Other predominantly pyrogenic PAHs also contributed to the HI, including acenaphthene, dibenzofuran, and anthracene. Dibenzothiophenes, which usually are considered primarily petrogenic, make a substantial contribution to the HI. Some of the creosote from the Wycoff wood treatment facility may have been distilled from a high-sulfur petroleum tar, which would contain high concentrations of dibenzothiophenes. These creosote-associated PAHs, particularly the higher molecular weight ones, have much higher log Kocs than predicted (Mitra et al. 1999), indicating a low accessibility and bioavailability. Thus, it is likely that the sediments are much less toxic to benthic animals than predicted by the high HI value and exceedence of the ERM value.
The urban runoff sediment contained 25 ppm of total PAHs, about 55% of the ERM concentration and about 6 times the ERL concentration. The HI of the PAH assemblage in this sediment was 64, indicating a hazard (risk of toxicity) to benthic organisms if the PAH are accessible and bioavailable. This sediment would be toxic to benthic animals if the PAH associated with the sediment particles are accessible. As indicated in Table 5, this sediment sample was enriched in parent PAHs and several 4-ring+ PAHs characteristic of pyrogenic sources. There also is evidence of some petrogenic PAH contributions, particularly the dibenzothiophenes (DBT) that are much more abundant in petrogenic than pyrogenic PAH assemblages (Neff 2002). More than 40% of the HI for this sediment was attributable to 4-ring and higher PAHs (mostly pyrogenic) that tend to sorb to sediment particles much more strongly than predicted (Neff 2002). However, there may be enough alkyl naphthalenes, phenanthrenes, and dibenzothiophenes (mostly petrogenic) in the sediments to elicit effects in some sensitive benthic organisms.
The toxicity of two heavily contaminated Eagle Harbor sediments was evaluated with a sensitive sand dollar embryo test (Meador et al. 1990). The sediment samples contained 33.6 and 37.0 μg/g total PAH. There was 100% mortality of the echinoderm embryos during exposure to both sediments. Ozretich et al. (2000) evaluated the toxicity of 30 creosote-contaminated sediments from nearby Elliott Bay (WA, USA) with two amphipod species. Mean amphipod mortality was less than 10% (sediments were not toxic) in seven sediments containing 12 to 140 μg/g total PAH (34 parent and alkyl PAH groups). There was 100% mortality in eight sediments containing 500 to 25,000 μg/g PAH. Mean amphipod mortality ranged from 13 to 78% in the remaining sediments, containing 19 to 480 μg/g PAH. Thus, the creosote-contaminated sediment used in the present investigation, containing 17,000 μg/g total PAH, probably also would be toxic. The urban runoff sediment containing 25 μg/g PAH probably were either nontoxic or moderately toxic to benthic animals.