Environ. Sci. Technol. 1996, 30, 2942-2947
Partitioning of Polynuclear Aromatic Hydrocarbons between Sediments from San Francisco Bay and Their Porewaters K E I T H A . M A R U Y A , * ,† ROBERT W. RISEBROUGH,‡ AND ALEXANDER J. HORNE† Department of Civil & Environmental Engineering, University of California, Berkeley, California 94720, and Bodega Bay Institute, Berkeley, California 94705
The in situ sediment-porewater partition coefficients (Koc′s) of a suite of 3-6 ring polynuclear aromatic hydrocarbons (PAHs) increased with the organic carbon content and with the amount of fines in surface sediments collected from a mudflat in San Francisco Bay along an intertidal gradient during the dry and wet seasons in 1993-1994. The Koc′s were an order of magnitude or more higher during the wet period of high surface runoff than during the dry season; the in situ measurements bracketed the corresponding values of Kow. Moreover, Koc′ decreased from the high- to the low-intertidal zone during the dry season, indicating heterogeneity in partitioning behavior along this spatial gradient. A plausible explanation for these trends in the partitioning behavior is a heterogeneity of the sediment organic carbon matrices, principally the content of highly aromatic soot particles; an increase in Koc′ might be anticipated in association with the lower activity coefficients of the aromatic PAHs in a highly aromatic solid phase matrix. If correct, this interpretation would qualify the utility of equilibrium partitioning models in formulating sediment quality criteria for PAHs.
Introduction A recent study of the distribution of polycyclic aromatic hydrocarbons (PAHs) in porewaters and sediments of three cores from Boston Harbor by McGroddy and Farrington (1) showed that the measured porewater PAH concentrations were lower than those predicted by either two- or three-phase equilibrium partitioning models. Values of the in situ sediment-porewater partition coefficient (Koc′), * Corresponding author present address: Skidaway Institute of Oceanography, 10 Ocean Science Circle, Savannah, GA 31411; telephone: 912-598-2327; fax: 912-598-2310; e-mail address:
[email protected]. † University of California, Berkeley. ‡ Bodega Bay Institute.
2942
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 30, NO. 10, 1996
defined as
Koc′ ) Kp′/foc
(1)
Kp′ ) cs/caq
(2)
where cs is the solid-phase concentration, caq is the aqueousphase concentration, and foc is the sediment fraction of organic carbon (1, 2), suggested that only a fraction of the sediment PAHs was available to partition rapidly into sediment porewaters. A high affinity of pyrogenicallyderived PAHs with soot particles, which have been shown to have highly aromatic matrices (3), was proposed as the probable cause. Such an interpretation would explain earlier observations of differences in partitioning and bioaccumulation behavior between pyrogenically- and petroleum-derived PAHs (4-7). PAHs in sediments from almost all of more than 150 sites in the San Francisco Bay region sampled in 19901992 had ratios of the sum of methylphenanthrene isomers to phenanthrene that were less than unity (8, 9), indicating pyrogenic origins (10). Only at several contaminated sites did petroleum-derived PAHs predominate (8, 9). Pyrogenic PAHs are believed to reach San Francisco Bay sediments principally from two input pathways: atmospheric fallout of combustion particles (i.e., soot) throughout the year and surface runoff during the rainy season. The local rainy season typically begins in November and extends into April, after which there is little or no rain until the following November. This study examines the spatial and temporal variation in partitioning behavior of PAHs between surface sediments and their porewaters of an intertidal marsh of the San Francisco Bay estuary. The local environment is very different from that of Boston Harbor, where industrial sources may have contributed to the PAH content of local sediments (1). Sampling periods of October and January were chosen to represent the peaks of the dry and wet seasons, respectively. In October there has been no input of PAHs from surface runoff for 6 months or more, whereas January is typically a period of high surface runoff.
Materials and Methods Study Site and Design. The study site was the mudflat section of Hoffman Marsh, a small intertidal marsh on the northeastern shore of central San Francisco Bay (Figure 1). This marsh has been historically subjected to both direct and runoff contamination, including disposal from lead battery and munitions manufacturing facilities. Runoff is currently received through two channels of Potrero Creek, a local urban waterway that drains a combination commercial, industrial, and residential landscape. These mudflats (∼4 ha) support many migratory and resident birds and are rich in epibenthic and infaunal invertebrates, including species of crustaceans, molluscs, and polychaetes. Whole sediments (∼25 kg) were collected from the top 5 cm of a 10-m2 quadrat and placed into a large, precleaned stainless steel bucket. Pieces of debris were discarded. The sediment was protected from external contamination by kiln-fired aluminum (Al) foil and was immediately transported back to the laboratory for refrigeration and further processing. This collection process was completed for the
S0013-936X(95)00909-6 CCC: $12.00
1996 American Chemical Society
FIGURE 1. Study area showing sampling locations along an intertidal gradient in Hoffman Marsh near Richmond, CA.
six sites shown in Figure 1sthree each collected in the low-, middle-, and high-intertidal zones (or Outer, Middle, and Inner Marsh, respectively) during October 1993 (OM-2, MM2, and IM-2) and again in January 1994 (OM-3, MM-3, and IM-3). Sediment and Porewater Extractions. Upon return to the lab, a several hundred gram subsample of homogenized sediment from the bucket was placed into a precleaned Teflon or glass jar and immediately frozen. After thawing and re-homogenization, 10-50-g duplicates of this subsample were weighed in 500-mL Teflon jars and oven-dried at 100 ( 2 °C for 24 h. Dried sediment was weighed again, homogenized with an equal amount of kiln-fired Na2SO4, and extracted with 500 mL of methylene chloride (CH2Cl2) in a glass Soxhlet apparatus for 8 h minimum. The CH2Cl2 extract was reduced by rotary evaporation, exchanged to hexane, re-reduced to ∼4 mL, and stored in clear borosilicate glass vials with Teflon-lined screw caps in the dark at 4 °C. Total organic carbon was measured as the difference between total and inorganic carbon in selected subsamples (∼1 g) of dried sediment by Huffman Analytical Laboratories (Golden, CO). Particle size distribution (14 φ size intervals including silt and clay fractions) was determined from the remaining frozen whole sediment aliquots using the pipet method (11) by Toxscan, Inc. (Watsonville, CA). Porewaters (PWs) were isolated from homogenized sediment in the steel bucket in a cold, dark room using a modified vacuum extraction method (12). Vacuum was applied for a period ranging from 24 to 72 h. Collected porewater was then transferred into a kiln-fired amber volumetric flask and stored in the dark at 4 °C. Within a period of days, raw porewater was glass-fiber filtered (Gelman type A/E, 47 mm diameter) under gentle vacuum. The filtrate was then divided into one to three subsamples (50-100 mL each); each subsample was vigorously shaken with three separate volumes (20-25 mL) of CH2Cl2 in a separatory funnel. CH2Cl2 extracts of each subsample were combined and prepared for fractionation in the same manner described previously for sediment Soxhlet extracts.
Extract Fractionation and GC/MS Analysis. All sediment and PW extracts in hexane were fractionated using packed column chromatography (13). Two sorbent systems were used: 18.0 g of 0.5% H2O-deactivated Florisil (60-100 mesh; Fisher Scientific) for the October 1993 samples (F1, 60-80 mL of hexane; F2, 200 mL of 30-50% (v:v) CH2Cl2 in hexane) and 8.0 g each of 5.0% H2O-deactivated alumina (80-200 mesh; Matheson, Coleman and Bell) on top of silica gel (80-200 mesh; J. T. Baker; F1, 25 mL of hexane; F2, 40 mL of 20% (v:v) CH2Cl2 in hexane). PAHs were recovered from the second (F2) fraction, which was treated with acid-activated copper filings to remove sulfur (sediment extracts only), exchanged to hexane, and reduced to 1-4 mL prior to analysis by GC/MS. The unsubstituted PAHs acenaphthylene, acenaphthene, fluorene, phenanthrene, anthracene, fluoranthene, pyrene, benz[a]anthracene, chrysene, benzo[b]fluoranthene, benzo[k]fluoranthene, benzo[e]pyrene, benzo[a]pyrene, indeno[1,2,3-cd]pyrene, dibenz[a,h]anthracene, and benzo[g,h,i]perylene were identified and quantified by GC/ion trap MS using authentic standards. The sum of these 16 compounds are hereafter referred to as total PAHs (or TPAH). The four methylphenanthrene isomers (m/z 192) eluting between anthracene and fluoranthene were quantified based on injections of authentic standards of the 1-methyl isomer. Quality Assurance Provisions. All organic solvents used in this study were of high purity (Optima grade; Fisher Scientific). All glassware was exhaustively hand-washed and rinsed, kiln-fired at 500 °C for 8 h minimum, and protected with kiln-fired Al foil prior to use. Stainless steel and Teflon containers and implements were carefully washed and rinsed thoroughly with acetone. Duplicate samples of kiln-fired Na2SO4 (∼50 g) were extracted and fractionated using the same methods described previously. PAHs in these sediment system blanks were generally nondetectable. In the rare cases where PAHs were detected, their levels were 2 orders of magnitude lower than the least contaminated field sediment. PAHs in PW procedural
VOL. 30, NO. 10, 1996 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2943
TABLE 1
Sediment Parameters, Total PAHs, and MP/P Ratios location
date
%SAND
%SILT
%CLAY
foc
TPAHoca (mg/kg)
MP/Pb
Inner Marsh (IM-2) Middle Marsh (MM-2) Outer Marsh (OM-2)
10/13/93 10/15/93 10/27/93
n/ac n/a 93.0
n/a n/a 4.9
n/a n/a 2.1
0.0072 0.0039 0.0033
47 13 7.9
0.47 0.53 0.72
Inner Marsh (IM-3) Middle Marsh (MM-3) Outer Marsh (OM-3)
01/28/94 01/24/94 01/26/94
79.2 80.7 n/a
13.7 15.1 n/a
7.1 4.2 n/a
0.0076 0.0059 0.0044
79 87 39
0.43 0.50 0.68
a Sum of 16 unsubstituted 3-6 ring PAHs on a sediment organic carbon (oc) basis. 192) to phenanthrene (m/z 178). c n/a, not analyzed.
b
Ratio of the sum of four methylphenanthrene isomers (m/z
blanks (CH2Cl2 rinses of the separatory funnel) were also nondetectable. PAH detection limits were variable primarily due to differences in extract volume at injection. However, typical detection limits for sediments and PWs were on the order of 0.5 ng/g on a sediment dry weight basis and 0.1 ng/L, respectively. Prior to extraction, the deuterated PAHs acenaphthene-d10, phenanthrene-d10, and chrysene-d12 were added to each sediment and PW sample to measure PAH recovery. Mean recoveries of these surrogates in sediment extracts were 82, 92, and 88% for acenaphthened10, phenanthrene-d10, and chrysene-d12, respectively. The corresponding percent recoveries for PWs were 70, 78, and 88. Individual PAHs were quantified only if the corresponding surrogate recovery exceeded 50%. In addition, a 1.0-g aliquot of a standard reference marine sediment (HS-6; NRC Canada) was extracted and subdivided into four subaliquots for PAH quantitation. The mean percent deviation from previously determined values for 14 of the 16 unsubstituted PAHs analyzed in this study was 24 ( 19.
Results and Discussion Sediment Parameters and PAHs. The sediments collected in this marsh ranged from sandy to silty mud and had relatively low sediment organic carbon (foc) content (Table 1). Using additional data obtained in a parent study (14), total PAHs in sediment on a dry weight basis were positively correlated with both the fines content (%FINES ) %SILT + %CLAY; n ) 13, r2 ) 0.86, p < 0.01) and foc (n ) 16, r2 ) 0.75, p < 0.01). As a result of this covariation, %FINES, foc, and (for the most part) sediment PAHs increased from the low to the high intertidal for both the wet and dry season. Unexpectedly, total PAHs on an organic carbon basis (TPAHoc) also increased with %SILT and %CLAY (Figure 2) and with foc (Figure 3). Four principal PAHs (phenanthrene, fluoranthene, pyrene, and chrysene) showed the same behavior as TPAHs; linear regressions between cs,oc ()cs/ foc) and foc were all significant (n ) 6, p < 0.05). These relationships indicate that PAHs are associated with sediment fines and, perhaps more importantly, that organic carbon normalization does not account for the heterogeneous distribution of PAHs in these urban marsh sediments. Porewater PAHs. Filtered PWs exhibited measurable but very low levels (ng/L or parts per trillion range) for 12 of the 16 unsubstituted PAHs (Table 2). Total PAHs in porewaters ranged between 13 ng/L (Middle Marsh, January 1994) and 71 ng/L (Middle Marsh, October 1993). Many of the higher MW compounds (g5 rings) were not detected (
