ARTICLE pubs.acs.org/ac
Measuring Picogram per Liter Concentrations of Freely Dissolved Parent and Alkyl PAHs (PAH-34), Using Passive Sampling with Polyoxymethylene Steven B. Hawthorne,*,† Michiel T. O. Jonker,‡ Stephan A. van der Heijden,‡ Carol B. Grabanski,† Nicholas A. Azzolina,§ and David J. Miller† †
Energy and Environmental Research Center (EERC), University of North Dakota, Box 9018, Grand Forks, North Dakota 58202, United States ‡ Institute for Risk Assessment Sciences (IRAS), Utrecht University, P.O. Box 80177, 3508 TD Utrecht, The Netherlands § Department of Civil and Environmental Engineering, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213, United States ABSTRACT: Passive sampling with nondepletive sorbents is receiving increasing interest because of its potential to measure freely dissolved concentrations of hydrophobic organic compounds (HOCs) at very low concentrations, as well as its potential for both laboratory use and field deployment. However, consistent approaches have yet to be developed for the majority of HOCs of environmental and regulatory interest. In the present study, a passive sampling method was developed which allows the freely dissolved concentrations of 18 parent and 16 groups of alkyl polycyclic aromatic hydrocarbons (PAHs) on the U.S. Environmental Protection Agency (USEPA)’s “PAH-34” target compound list to be measured. Commercially available 76-μm-thick polyoxymethylene (POM) was placed in sediment/water slurries and exposed for up to 126 days, with 28 days found to be sufficient to obtain equilibrium among the sediment, water, and POM phases for the target 2- to 6-ring PAHs. The POM/water partition coefficients (KPOM) necessary to calculate freely dissolved concentrations for parent PAHs were determined in two separate laboratories (one using pure standards, and the other using coal tar/petroleum-contaminated sediments) and agreed very well. Since the so-called “16” alkyl PAHs on the PAH34 list actually include several hundreds of isomers for which no standards exist, sediments impacted by coal tar, or spiked with a coal tar/petroleum nonaqueous phase liquid (NAPL) were also used to measure KPOM values for each alkyl PAH cluster. The log KPOM values ranged from ca. 3.0 to 6.2 for 2- to 6-ring parent PAHs, and correlated well with SPARC octanol/water coefficients (KOW) (correlation coefficient of r2 = 0.986). However, log KPOM values for alkyl PAHs deviated increasingly from SPARC log KOW values with increasing degree of alkylation. A simple empirical model that incorporates the number of carbon atoms in a PAH gave a better fit to the experimental log KPOM values, and was used to estimate log KPOM for alkyl PAHs that could not be directly measured. Detection limits (as freely dissolved concentrations) ranged from ca. 1 part per trillion (ng/L) for the 2-ring PAH naphthalene, down to 0.99) over at least 4 orders of magnitude.
’ INTRODUCTION There is a rapidly growing acceptance that freely dissolved concentrations of hydrophobic organic compounds (HOCs) are superior to sediment-associated concentrations in determining biological uptake and related toxic effects.17 Because of the lack of standard methods to measure freely dissolved concentrations, regulatory models and much of the scientific literature have relied on predicting dissolved concentrations from sediment concentrations using equilibrium partitioning models and conventional sediment organic carbon/water partition coefficients that have been determined by spiking studies or correlative models. However, in recent years, a large volume of experimental data on HOC partitioning into water from historically contaminated sediments has demonstrated that equilibrium partitioning models, which rely r 2011 American Chemical Society
on conventional KOC values and sediment-associated concentrations typically overpredict dissolved concentrations of HOCs by as much as 3 orders of magnitude.811 This overprediction occurs primarily because KOC values normally used in such models are typically much lower than those that exist in historically contaminated sediments.813 Incorporating additional carbon phases (e.g., “anthropogenic,” “soot,” or “black” carbon into partitioning models has not yielded better predictions of dissolved concentrations when applied to large sets of field data,811,14 although recent polyparameter models show promise for improving predictions.1517 Received: June 3, 2011 Accepted: July 15, 2011 Published: July 15, 2011 6754
dx.doi.org/10.1021/ac201411v | Anal. Chem. 2011, 83, 6754–6761
Analytical Chemistry More-accurate models that are based on measured sedimentassociated and freely dissolved concentrations on hundreds of field samples have been proposed,9,11 but as Arp et al.9 concluded in a recent extensive review, “... the only way to accurately obtain accurate pore water concentrations is to measure them directly, and not infer them from sediment concentrations.” Unfortunately, few consistent methods exist to determine freely dissolved HOC concentrations accurately, and such methods typically are calibrated for only a few members of a compound class. To date, only one method is approaching acceptance as a standard regulatory method, i.e., ASTM Standard D7363 for determining freely dissolved concentrations of alkyl and parent PAHs (PAH34), as required for the U.S. Environmental Protection Agency (USEPA)’s hydrocarbon narcosis model.18,19 Although obtaining large volumes of overlying water for conventional solvent extraction is relatively simple, obtaining large volumes of sediment interstitial water can be very difficult, especially for sandy sediments. For practical reasons, methods to determine freely dissolved HOC concentrations should either be able to use small pore water samples, or not require the separation of pore water from the sediment. Such methods must be able to eliminate interferences of colloid-bound HOCs, as well as those associated with the dissolved organic carbon (DOC) phase, which is difficult or impossible with solvent extractions. Methods based on separating the pore water from the sediment and using solid-phase microextraction (SPME) to replace solvent extraction have been developed for PAHs and polychlorinated biphenyls (PCBs) and are capable of low pg/mL detection limits, using only a few milliliters of water.1820 However, lower detection limits for freely dissolved HOCs are often desired (e.g., low fg/mL), especially in uptake/bioconcentration studies. Several passive sampling approaches have been reported to determine freely dissolved HOC concentrations that use nondepletive sorbents such as polydimethylsiloxane (PDMS), polyethylene (PE), and polyoxymethylene (POM), placed directly in the sediment/water slurry.1,3,5,7,2127 These methods rely on several days or weeks of exposure for the sampler to come to equilibrium with the sediment and water phases and/or performance reference compounds (PRCs).22,28 After equilibrium is attained, the HOC concentrations are determined in the samplers, and previously measured sampler/water partition coefficients (Ksampler) are used to calculate the freely dissolved concentrations. Passive sampling methods have the advantage of being useful both for field and laboratory exposures, but suffer from a lack of accurately determined Ksampler values for the large number of HOCs of regulatory interest. In addition, sorbents that have been used for passive sampling come in different forms from different suppliers and are often “home-made” in individual laboratories and, therefore, the resultant methods are difficult to transfer for use among multiple laboratories. In an effort to address these issues, a passive sampling method was recently developed using commercially available 76-μm-thick POM. In that work, KPOM values were directly measured for ca. 60 PCB congeners that were present in significant concentrations in the commercial Aroclor products.29 POM was chosen because it is commercially available, has suitable sampler/water partitioning behavior for a passive sampler, has low cost, has a smooth surface that avoids contamination by colloids, and is physically robust.21 Low pg/L detection limits were obtained, and a correlation of KPOM values with octanol/water partition coefficient (KOW) values was presented that allows KPOM values to be predicted for other PCB congeners with reasonable accuracy.
ARTICLE
Table 1. Test Sediments and NAPL Spiking Concentrationsa PAH Concentration (mg/kg) sediment TOC (wt %) SOC (wt %)
EPA-16
EPA-34
A
3.98
1.44
848
2940
B
1.34
0.43
697
1570
C
4.00
0.66
2
10
NAPL Spike Concentration, Sed. C (mg/kg) low total NAPL
2
total PAH-34
0.7
high 28700 9600
Representative Parent PAHs naphthalene
0.06
950
phenanthrene pyrene
0.03 0.01
440 200
chrysene
0.003
44
benzo[e]pyrene
0.001
22
indeno[1,2,3-cd]pyrene
0.001
14
Parent and Alkyl PAHs 2-ring
0.3
5020
3-ring
0.05
770
4-ring
0.05
760
5+6-ring
0.01
145
a
Concentrations of the alkyl and parent PAHs in the NAPL were determined using gas chromatography/mass spectroscopy (GC/MS), as previously described.32 TOC (total organic carbon) and SOC (soot or “black” organic carbon) were determined as previously described.14
The goal of the present study was to further develop the POM passive sampling approach for the determination of pg/L concentrations of alkyl and parent PAHs (PAH-34) required for the USEPA’s hydrocarbon narcosis model.30 KPOM values for several parent PAHs were independently determined by two laboratories (IRAS, Utrecht University, The Netherlands; and the EERC in North Dakota, USA), following different experimental approaches. Particular effort was focused on determining KPOM values for the alkyl 2- and 3-ring PAHs (which occur in isomeric clusters), because those species are the PAHs most responsible for toxicity to benthic organisms.4,31
’ EXPERIMENTAL SECTION KPOM values for several parent PAHs were determined independently by the laboratories at IRAS and EERC. IRAS determined KPOM values for 13 parent PAHs using pure compounds, while EERC determined KPOM values for parent and alkyl PAHs using two contaminated sediments and one clean sediment, spiked with a mixed coal tar/petroleum NAPL (nonaqueous-phase liquid, presently under certification by the National Institute of Science and Technology (NIST) as SRM 1991). Methods used by each laboratory were developed independently and are described below. POM Preparation and PAH Recovery. The thinnest commercially available POM material (C.S. Hyde Co., Lake Villa, IL, USA) was selected for these investigations so that the same material could be available to other laboratories, and because thinner POM reaches equilibrium more quickly than thicker 6755
dx.doi.org/10.1021/ac201411v |Anal. Chem. 2011, 83, 6754–6761
Analytical Chemistry sheets. POM sheets (76 μm) were cut into coupons using scissors. At the EERC, 100 mg coupons (ca. 2 cm 4 cm) were cleaned by sequential extraction for 2 h in n-hexane, followed by methanol with the aid of sonication.29 IRAS used a similar procedure to clean 5-mg coupons, involving a n-hexane wash followed by three methanol washes. Both procedures yielded blanks with no detectable PAHs. The coupons were air-dried and stored in glass jars until use. After exposure, recovery of the sorbed PAHs from the coupons at the EERC was achieved by sonication in 1:1 acetone/n-hexane (spiked with a d-PAH internal standard solution containing 2- to 6-ring perdeuterated PAHs31) for 3 h. In order to investigate the efficiency of this procedure, triplicate coupons were equilibrated with contaminated sediment (sediment “A” in Table 1) for 28 days, extracted as just described, then extracted a second time for an additional 18 h. A comparison of the first and second extracts showed that all PAH-34 species showed >99% recovery, except for benzo[a]pyrene, which had 98% recovery. Recovery of the parent PAHs at IRAS was achieved by extraction with acetonitrile, which showed similarly high recoveries. IRAS also determined KPOM values for 11 pure PCB congeners so that their values could be compared to KPOM values based on Aroclor mixtures that were reported earlier.29 PCBs were extracted with methanol for 3 h in a Soxhlet apparatus. The methanol extracts were subsequently concentrated and exchanged to n-hexane, cleaned with an aluminum oxide/silica gel column, and analyzed after exchange to isooctane, with an average overall recovery of 96%. KPOM Determinations. KPOM values for alkyl and parent PAHs included on the EPA’s “PAH-34” list were measured at the EERC in a manner similar to that previously used to determine KPOM values for PCBs.26 Two field-contaminated sediments (A and B in Table 1) and one background NAPL-spiked sediment (C in Table 1) were equilibrated with water and POM (typically 28 days as described later). Triplicate sediment/water/POM vials were used for all determinations for sediments A and B, and either duplicate or triplicate vials for the spiked sediment C. Sediments A and B were chosen for their relatively high freely dissolved concentrations, as determined previously,18,19 and their lack of a free NAPL phase. Sediment C was selected for low (nondetectable) freely dissolved concentrations. A large range of NAPL spiking concentrations on the sediment C were used in an effort to get measurable freely dissolved concentrations of the 2- to 6-ring parent and alkyl PAHs in the PAH-34 target compound list, as well as to investigate the linearity of the POM response. Total concentrations of NAPL spikes ranged from ca. 2 to 28700 mg/kg of NAPL, which corresponds to total PAH-34 concentrations from ca. 0.7 to 9600 mg/kg (concentration ranges by ring size are given in Table 1). Higher spiking concentrations were not used to avoid the formation of a separate NAPL phase in the sediment/water/ POM equilibration vials. Each 40-mL sample vial contained ca. 1015 g (dry weight) of sediment, one POM coupon, and ca. 3035 mL of water containing 150 mg/L sodium azide. The vials were continually mixed in a rotating box (ca. six revolutions per minute) during the exposure time. After the exposure was completed, the POM coupons were recovered and rinsed with a gentle spray of distilled water to remove sediment particles. The POM coupons were then extracted and analyzed as described below. Freely dissolved concentrations and concentrations in POM of PAH-34 were determined immediately upon completing the equilibration step as described below.18,19 KPOM values were calculated as the ratio of the concentration in POM to the freely
ARTICLE
Figure 1. PAH sorption kinetics into 76-μm POM. No increases in POM PAH concentrations were observed after 28 days, up to 126 days.
dissolved concentration for each parent and isomeric group of alkyl PAHs. IRAS’ KPOM values were determined using 13 pure PAHs ranging from three to six rings (>98%, Aldrich Chemical Co., Steinheim, Germany) that were spiked into 250-mL amber glass bottles filled with Millipore water containing 0.01 M calcium chloride and 50 mg/L sodium azide. Each bottle was spiked with 40 μL of an acetone solution containing 1.5 mg/L of each of the 13 parent PAH standards. The bottles were placed in darkness on a shaker Table (120 rpm) for 10 weeks to equilibrate. After equilibration, the POM coupons were extracted and analyzed as described below. Each KPOM determination was performed in quadruplicate. KPOM values for individual PCB congeners at IRAS were determined in 5-fold in a similar manner, except that 10-mg POM coupons were used and that 60 μL of a solution containing 2.5 mg/L of each of the 11 PCBs was spiked. Analysis Methods. All PAH analyses for PAH-34 at the EERC were performed using GC/MS in the selected-ion mode. POM extracts were analyzed in the same manner as previously described for sediment PAH-34.32 Freely dissolved concentrations were determined using ASTM Standard D7363.15,16 Both methods utilize several 2- to 6-ring perdeuterated PAH (d-PAH) internal standards, as described in the respective methods. IRAS analyzed PAHs in the POM and water extracts (after three sequential n-hexane extractions) using HPLC with 2-methylchrysene as internal standard, as previously described.33 PCB concentrations (IRAS) were determined using GC/ECD with congener 209 as an internal standard, as previously described.31 PCB KPOM values for the EERC have been previously reported.29
’ RESULTS AND DISCUSSION Equilibration Time. Replicate vials of spiked sediment C and POM were mixed for up to 126 days, with the POM coupons from duplicate vials being collected and analyzed at various times. Figure 1 shows the progress toward equilibrium. As expected, the lower-molecular-weight naphthalene equilibrates faster than the high-molecular-weight benzo[b+k]fluoranthene, but both lowand high-molecular-weight PAHs are essentially at equilibrium after 1421 days, and further exposure times up to 126 days did not measurably increase the PAH concentrations in POM. These 6756
dx.doi.org/10.1021/ac201411v |Anal. Chem. 2011, 83, 6754–6761
Analytical Chemistry
ARTICLE
Table 2. Measured log KPOM Values and Detection Limits for Parent and Alkyl PAHs and PCB Congeners EERC, Contaminated Sediments
IRAS, Pure Compounds mean
SD
mean log KPOM,
freely dissolved
log KPOM
log KPOM
all values
detection limit (pg/L)c
mean log
SD
n
KPOM
log KPOM
naphthalene
17
3.05
0.09
3.05
990
2-methylnaphthalene
17
3.36
0.06
3.36
410
1-methylnaphthalene
17
3.32
0.07
3.32
440
C2 naphthalenes
8
3.59
0.04
3.59
260
C3 naphthalenes
10
3.64
0.11
3.64
230
C4 naphthalenes
10
3.73
0.07
3.73
190
acenaphthylene acenaphthene
6 20
3.78 3.50
0.06 0.04
3.78 3.50
160 310
fluorene
11
3.83
0.12
3.83
140
C1 fluorenes
14
4.09
0.10
4.09
81
C2 fluorenes
14
4.50
0.12
4.50
31
4.74
23
parent and alkyl PAHs
n
C3 fluorenesa phenanthrene
20
4.20
0.07
4
4.18
0.02
4.20
61
anthracene
14
4.31
0.09
4
4.28
0.01
4.30
52
C1 phenanthrenes/anthracenes C2 phenanthrenes/anthracenes
17 10
4.48 4.90
0.10 0.08
4.48 4.90
32 13
C3 phenanthrenes/anthracenes
8
5.31
0.07
5.31
5
C4 phenanthrenes/anthracenes
3
5.41
0.07
5.41
4
fluoranthene
18
4.54
0.09
4
4.64
0.03
4.56
26
pyrene
20
4.55
0.09
4
4.65
0.02
C1 fluoranthenes/pyrenes
14
4.90
0.11
benz[a]anthracene
18
5.47
0.10
4
5.36
chrysene C1 chrysenes/benz[a]anthracenesa
19
5.44
0.12
4
5.43
4.57
26
4.90
13
0.03
5.46
4
0.03
5.43 5.60
3 3
C2 chrysenes/benz[a]anthracenesa
5.89
2
C3 chrysenes/benz[a]anthracenesa
6.18
0.9
C4 chrysenes/benz[a]anthracenesa
6.47
0.5
benzo[b]fluorantheneb
4
5.80
0.03
5.80
1.2
4
5.94
0.04
5.97
1.1
benzo[e]pyrene
4
5.67
0.03
5.67
1.4
benzo[a]pyrene perylene
4
5.96
0.03
2
6.04
0.05
5.96 6.04
0.8 0.9
indeno[1,2,3-cd]pyrene
2
6.16
0.03
4
6.31
0.10
6.26
0.5
4
6.30
0.12
6.30
0.5
benzo[k]fluorantheneb
5
5.99
0.31
dibenz[ah]anthracene benzo[ghi]perylene
2
6.05
0.16
4
6.10
0.09
6.09
0.8
18
15
5.12
0.06
5
5.13
0.06
5.12
8
28
15
5.68
0.10
5
5.45
0.04
5.62
2
52
15
5.65
0.05
5
5.70
0.04
5.66
2
118
5
6.32
0.14
5
6.40
0.06
6.36
0.4
138
4
6.50
0.10
5
6.54
0.04
6.52
0.3
156
3
6.59
0.29
5
6.70
0.05
6.66
0.2
d
PCB congeners
a
No measured values could be obtained. Values were estimated based on eq 2. b The log KPOM values for benzo[b]- and benzo[k]fluoranthene are reported as a single value for the EERC, because of insufficient resolution during gas chromatographic analysis. c The basis for the freely dissolved detection limits is described in the text. Values are given for a single isomer. Detection limits for the alkyl clusters are ca. 5-fold higher. d PCB log KPOM values for EERC are taken from ref 29.
results are similar to those found earlier for PCB congeners having molecular weights similar to the larger PAHs (i.e., up to tri- or tetra-chloro congeners), while higher-molecular-weight congeners required at least 28 days to reach equilibrium.26 These
results are also consistent with previous studies by Cornelissen et al. that reported higher-molecular-weight PAHs were at equilibrium in field-deployed 55 μm POM after one month.33,35 Since one goal of the present study was to develop sampling 6757
dx.doi.org/10.1021/ac201411v |Anal. Chem. 2011, 83, 6754–6761
Analytical Chemistry conditions that apply to several classes of HOCs (and not just PAHs), the longer 28-day equilibrium time was used for all subsequent investigations (except the 10-week KPOM determinations done at IRAS). POM/Water Partition Coefficients for Parent PAHs and Select PCB Congeners. Two major obstacles to widespread use of passive sampling are the lack of consistent sampler materials, and the lack of “standard” partition coefficients for any of the various samplers used. Reported Ksampler values often vary significantly from one laboratory to another, even when the laboratories report using the same sampler material.35 These discrepancies can occur because of different forms (suppliers) of the same polymer being used, differences in exposure time to attain equilibrium (some being insufficient), and other experimental artifacts, such as the use of insufficiently pure water and poor recoveries of sorbed HOCs during the sampler extraction step. Therefore, a major goal of the present study was to compare KPOM values on the same commercially available POM material that were determined independently in two laboratories. Initially, neither laboratory was aware of the others’ activities, and no information was exchanged on methodology (other than the fact that both laboratories were using 76-μm POM) or experimental results until both laboratories had determined their final KPOM values. The ability to measure KPOM using contaminated sediments A and B, and the NAPL-spiked sediment C was limited by the detection limit of the ASTM SPME method used to determine the freely dissolved concentrations after the equilibration of the sediment/water slurry with the POM coupon. Up to 20 values could be determined for the lower-molecular-weight PAHs, but values for the 5- and 6-ring PAHs were limited to only the highest spiking level or could not be determined. However, IRAS was able to obtain KPOM values for the latter PAHs, using pure compounds. A comparison of the KPOM values for parent PAHs determined in each laboratory is given in Table 2. In general, the values determined at IRAS showed smaller standard deviations among the replicates than those determined at the EERC, which might be expected since IRAS used pure PAH standards and the EERC used contaminated (or NAPL-spiked) sediments. However, the agreement between the two laboratories on the log KPOM values is quite good (typically within 0.1 log unit), especially compared to the broad range of values in previous reports.21,29,36,37 The log KPOM values for PCB congeners determined at IRAS with a 10-week exposure also agreed very well with those previously reported by the EERC,28 based on a 28-day exposure (Table 2), again indicating that the 28-day exposure is sufficient for PAH-34 and PCB congener studies. There may be some concern that complex hydrocarbon matrices (e.g., coal tar and petroleum) could cause changes in POM sorption behavior, especially at higher concentrations. However, the good agreement in KPOM values between the IRAS pure PAH determinations and contaminated and coal tar/petroleum NAPL-spiked sediments at the EERC show that bulk hydrocarbon matrix effects were not significant over the wide range of NAPL spike concentrations investigated (228 700 mg NAPL per kg sediment) at the EERC. POM/Water Partition Coefficients for Alkyl PAHs. Sorbent/ water partition coefficients for HOCs are typically based on the use of pure standard compounds, but this approach is not possible for determining Ksampler values for most of the PAHs on the USEPA’s PAH-34 list. It is important to note that the 16 alkyl PAHs included in the PAH-34 list (with the exception of the two methylnaphthalene isomers) are actually composed of isomeric clusters of PAHs that have the same molecular weight.
ARTICLE
Figure 2. Selected ion chromatograms for alkyl naphthalenes. The mass (m/z) is displayed above each set of isomers.
For example, the C4-naphthalene cluster alone consists of ∼70 individual isomers (Figure 2) but is counted as only “one” PAH in the PAH-34 list.32 Therefore, the 16 alkyl PAHs in the PAH-34 list actually consist of several hundreds of individual PAHs. Methods for their determination are particularly important, since the alkyl two- and three-ring PAHs contributes ca. 80% of the PAH toxicity in contaminated sediments (based on mortality to Hyalella azteca.4,31) Since there are only a few pure alkyl PAH standards available, either contaminated field samples or sediments spiked with a “real-world” PAH source are thus required to determine KPOM values for the isomeric clusters that exist in field sediments, as was done in the current study. As noted above, the sediments were spiked with the NAPL sample over a range of 228 700 mg/kg (Table 1), in an attempt to obtain sufficiently high dissolved concentrations and concentrations in POM of all constituent PAHs without causing the production of a separate NAPL phase in sediment/water slurry. Although this approach proved successful for the majority of parent PAHs and most of the alkyl PAHs, it did not succeed for the C3-fluorenes and the alkyl chrysene/ benz[a]anthracenes. When experimental KPOM values are not available, they could be estimated based on a linear correlation of log KPOM with log KOW. As shown in Figure 3 (top), this works well for parent PAHs, since the log KPOM values from both laboratories show a good linear correlation with the log KOW values (correlation coefficient of r2 = 0.9846). However, log KPOM values for the alkyl PAHs do not show an unambiguous relationship with SPARC-modeled KOW values (Figure 3, bottom). Since no experimental KOW values (or values from other models) are available for the alkyl clusters, we investigated simple correlations that include the number of aromatic and aliphatic carbons. The model was fit using both single and multiple regression, where the measured KPOM value (Table 2) was the response variable. The total number of carbons (C) in the PAH molecule was used for the single explanatory variable (eq 1), and the number of aromatic carbons (CAR) and aliphatic carbons (CAL) were used for the multiple regression (eq 2).
6758
log KPOM ¼ 0:132 þ 0:288C
ð1Þ
log KPOM ¼ 0:222 þ 0:284CAR þ 0:246CAL
ð2Þ
dx.doi.org/10.1021/ac201411v |Anal. Chem. 2011, 83, 6754–6761
Analytical Chemistry
Figure 3. Measured log KPOM values compared to SPARC-predicted KOW values (top), and compared to predicted log KPOM from eq 2 (bottom).
Both models explained a significant portion of the variance in measured KPOM values, with adjusted correlation coefficient (r2) values of 0.960 and 0.962, respectively (p < 0.01). A comparison of measured KPOM values to modeled KPOM values shows that the modeled values agree well over the range of PAHs (see Figure 3, bottom). Since both equations yielded almost-identical fits with the experimental log KPOM values, the simpler eq 1 was used to estimate the log KPOM values for the C3-fluorenes and the alkyl chrysene/benz[a]anthracenes given in Table 2. Based on the fit to experimental data shown in Figure 3 (bottom), this correlation should be useful for estimating KPOM values for other PAHs not on the PAH-34 list. Detection Limits and Reproducibility. Attempts were made to use “clean” sediment C spiked with increasingly lower amounts of the NAPL to determine the POM concentrations for each PAH that would yield a reasonable signal-to-noise ratio in the GC/MS selected-ion plots. However, this was not possible, because the sediment yielded significant concentrations in the POM extracts for the majority of PAHs on the PAH-34 list, even when no NAPL was added. Note that, even though the freely dissolved concentrations were below the detection limits of the ASTM method used to select sediment C, the detection limits using the POM method are orders-of-magnitude lower, as discussed below. Therefore, detection limits are conservatively estimated for individual parent PAHs, based on needing a 1 ng/g POM (corresponding to 100 pg extracted from a 100-mg POM coupon, concentrated to 100 μL) and the sensitivity routinely achieved for PAHs using GC/MS with selected-ion monitoring (SIM) (i.e., 1 pg injected in 1 μL of POM extract using splitless injection). The freely dissolved detection limit for each parent PAH was then calculated from the 1 ng/g POM value by dividing by each PAH’s KPOM value. As shown in Table 2, this yields detection limits ranging from ca. 900 pg/L for low-molecular-weight PAHs, down
ARTICLE
to 0.995), demonstrating that the POM method is linear over several orders of magnitude, and that the method should be 6759
dx.doi.org/10.1021/ac201411v |Anal. Chem. 2011, 83, 6754–6761
Analytical Chemistry
Figure 4. POM sorption isotherms for representative parent and alkyl PAHs. POM PAH concentrations are converted to freely dissolved concentrations using the KPOM values in Table 2.
capable of measuring freely dissolved PAH-34 concentrations over a broad range of relevant environmental conditions. These results also demonstrate that the PAH concentration does not affect the sediment/water/POM equilibrium, and that no adjustment for nonlinear POM sorption isotherms is needed. These conclusions are fully in agreement with those previously drawn by others.16,21,22 Nondepletive Sampling. Determining the freely dissolved concentrations using passive sampling requires that the sediment, water, and sampler come to equilibrium, and that the sampler does not deplete the PAHs from the sediment/water slurry enough to significantly change freely dissolved concentrations. Depletion was calculated for the results of the linearity study discussed above, and found to be ca. 2%3% for all PAHs at all spiking levels. No trend with PAH concentration was observed, as expected based on the linearity of the sorption isotherms shown in Figure 4. Similarly, there was no observable trend with molecular weight in percentage depletion, as would be expected since the KOC and KPOM values should show similar increases with increasing PAH molecular weight. The lack of any dependence of depletion on PAH concentrations or molecular weight also agrees with the results for PCB congeners previously reported in a study using the same 76-μm POM material.29 Previous workers have suggested that the sampler should contain no more than 5% of each HOC species at equilibrium in order to maintain “negligible depletion” by the POM coupon36 The depletion values just described were based on the use of 10 g (dry weight) of sediment C and a 100-mg POM coupon (100-to1 sediment-to-POM ratio), and the resultant 2%3% depletion was acceptable. However, sediments with less sorptive capacity (e.g., lower content of organic carbon) would experience more depletion, so these results for PAHs support the suggestion made in a previous study on PCBs29 that the sediment/POM ratio be maintained at ca. 200:1 or greater for routine sampling. Practical Aspects of the Method. As previously noted, a major goal of these investigations was to develop a consistent, robust, and practical approach for determining freely dissolved concentrations of several classes of HOCs, including all members of a class that may be of scientific or regulatory interest. The method described here is fully compatible with that described earlier for PCB determinations,31 and a single POM coupon and extract can be used for both compound classes (and likely as well
ARTICLE
for any HOCs for which KPOM values are known), by simply adding the appropriate internal standards to the POM extracts and analyzing for the HOCs of interest. Since the POM material that we use is commercially available, and since the KPOM values for both PAHs and PCBs have been validated by two independent laboratories, future users of the method should be able to use these methods and the KPOM values that we report for PAHs and PCBs,29 without the need to perform the arduous process of determining KPOM values in every laboratory. Although the 76-μm POM used in this study has the disadvantage of requiring longer equilibration times than some other passive samplers, the very practical advantages of being commercially available, physically robust, and not susceptible to colloidal interferences support its selection. The same approach should apply to both field deployment and laboratory use, and has the advantage that only relatively small (e.g., 100-mL jars) samples must be shipped for laboratory studies, which can significantly reduce costs for large field surveys over approaches that require liters of sample.
’ CONCLUSION Passive sampling has a strong potential to provide high-quality analytical results for determining the freely dissolved concentrations of a wide variety of hydrophobic organic compounds (HOCs) at environmentally relevant concentrations as low as 1 pg/L (parts per quadrillion). The technique is applicable to use both in the field and in the laboratory, but widespread acceptance has been restricted (especially for the regulatory community) by two major factors: (1) the need for a common and reproducible sampler material that is widely available from a commercial source, and (2) accurate sampler/water partition coefficients for the range of target analytes that are relevant to sediments and to regulatory target compound lists. The present study, combined with an earlier report on determining polychlorinated biphenyls (PCBs),29 has attempted to address these objections to passive sampling by utilizing a commercially available sampler material that is easy to prepare and physically robust, by independent validation of the KPOM values we report by two separate research groups, and by characterizing the range, reproducibility, and sensitivity of the method. It is the eventual goal of these studies to provide a single passive sampling method that can be used to determine freely dissolved concentrations of several classes of HOCs with a single polyoxymethylene (POM) coupon, and that is sufficiently robust to be used in a routine manner by multiple laboratories, similar to the way many methods used to determine HOCs in sediments are currently applied. ’ AUTHOR INFORMATION Corresponding Author
*Tel.: 701-777-5256. Fax: 701-777-5181. E-mail: shawthorne@ undeerc.org.
’ REFERENCES (1) Trimble, T. A.; You, J.; Lydy, M. J. Chemosphere 2008, 71, 337–344. (2) Werner, D.; Hale, S. E.; Ghosh, U.; Luthy, R. G. Environ. Sci. Technol. 2010, 44, 2809–2815. (3) Lohmann, R.; Burgess, R. M.; Cantwell, M. G.; Ryba, S. A.; MacFarlane, J. K.; Gschwend, P. M. Environ. Toxicol. Chem. 2004, 23, 2551–2562. 6760
dx.doi.org/10.1021/ac201411v |Anal. Chem. 2011, 83, 6754–6761
Analytical Chemistry (4) McDonough, K. M.; Azzolina, N. A.; Hawthorne, S. B.; Nakles, D. V.; Neuhauser, E. F. Environ. Toxicol. Chem. 2010, 29, 1545–1550. (5) Lu, X.; Reible, D. D.; Fleeger, J. W. Environ. Toxicol. Chem. 2006, 25, 2869–2874. (6) Sun, X.; Ghosh, U. Environ. Toxicol. Chem. 2008, 27, 2287–2295. (7) Van der Heijden, S. A.; Jonker, M. T. O. Environ. Sci. Technol. 2009, 43, 3757–3763. (8) Hawthorne, S. B.; Grabanski, C. B.; Miller, D. J. Environ. Toxicol. Chem. 2006, 25, 2901–2911. (9) Arp, H. P. H.; Breedveld, G. D.; Cornelissen, G. Environ. Sci. Technol. 2009, 43, 5576–5585. (10) Hawthorne, S. B.; Arp, H. P. H.; Grabanski, C. B.; Miller, D. J. Environ. Sci. Technol. 2011, in press. DOI: 10.1021/es200802j. (11) Arp, H. P. H. Environ. Sci. Technol. 2011, 45, 5139–5146. (12) Jonker, M. T. O.; Smedes, F. Environ. Sci. Technol. 2000, 34, 1620–1626. € Environ. Sci. Technol. 2004, 38, (13) Cornelissen, G.; Gustafsson, O. 148–155. (14) Hawthorne, S. B.; Grabanski, C. B.; Miller, D. J. Environ. Toxicol. Chem. 2005, 26, 2505–2516. (15) Endo, S.; Grathwohl, P.; Haderlein, S. B.; Schmidt, T. C. Environ. Sci. Technol. 2009, 43, 3094–3100. (16) Endo, S.; Grathwohl, P.; Haderlein, S. B.; Schmidt, T. C. Environ. Sci. Technol. 2009, 43, 3187–3193. (17) Endo, S.; Grathwohl, P.; Haderlein, S. B.; Schmidt, T. C. Environ. Sci. Technol. 2009, 43, 393–400. (18) Standard Test Method for Determination of Parent and Alkyl Polycyclic Aromatics in Sediment Pore Water Using Solid-Phase Microextraction and Gas Chromatography/Mass Spectrometry in Selected Ion Monitoring Mode. ASTM Standard Test Method D 7363-07. In 2007 Annual Book of ASTM Standards; ASTM International: West Conshocken, PA, 2007. (19) Hawthorne, S. B.; Grabanski, C. B.; Miller, D. J.; Kreitinger, J. P. Environ. Sci. Technol. 2005, 39, 2795–2803. (20) Hawthorne, S. B.; Grabanski, C. B.; Miller, D. J. Anal. Chem. 2009, 81, 6936–6943. (21) Jonker, M. T. O.; Koelmans, A. A. Environ. Sci. Technol. 2001, 35, 3742–3748. (22) Hong, L.; Luthy, R. G. Chemosphere 2008, 72, 272–281. (23) Anderson, K. A.; Sethajintanin, D.; Sower, G.; Quarles, L. Environ. Sci. Technol. 2008, 42, 4486–4493. (24) Allan, I. J.; Booij, K.; Paschke, A.; Vrana, B.; Mills, G. A.; Greenwood, R. Environ. Sci. Technol. 2009, 43, 5383–5390. (25) Fernandez, L. A.; MacFarlane, J. K.; Tcaciuc, A. P.; Gschwend, P. M. Environ. Sci. Technol. 2009, 43, 1430–1436. (26) Witt, G.; Liehr, G. A.; Borck, D.; Mayer, P. Chemosphere 2009, 74, 522–529. (27) Cornelissen, G.; Pettersen, A.; Broman, D.; Mayer, P.; Breedveld, G. D. Environ. Toxicol. Chem. 2008, 27, 499–508. (28) Booij, K.; Smedes, F. Environ. Sci. Technol. 2010, 44, 6789–6794. (29) Hawthorne, S. B.; Miller, D. J.; Grabanski, C. B. Anal. Chem. 2009, 81, 9472–9480. (30) U.S. Environmental Protection Agency (USEPA). Procedures for the Derivation of ESBs for the Protection of Benthic Organisms: PAH Mixtures, EPA/600/R-02/013; Office of Research and Development: Washington, DC, 2003. (31) Hawthorne, S. B.; Azzolina, N. A.; Neuhauser, E. F.; Kreitinger, J. P. Environ. Sci. Technol. 2007, 41, 6297–6304. (32) Hawthorne, S. B.; Miller, D. J.; Kreitinger, J. P. Environ. Toxicol. Chem. 2006, 25, 287–296. (33) Jonker, M. T. O.; van der Heijden, S. A. Environ. Sci. Technol. 2007, 41, 7363–7369. (34) ter Laak, T. L.; Busser, F. J. M.; Hermens, J. L. M. Anal. Chem. 2008, 80, 3859–3866. (35) Difilippo, E. L.; Eganhouse, R. P. Environ. Sci. Technol. 2010, 44, 6917–6925. (36) Cornelissen, G.; Arp, H. P. H.; Pettersen, A.; Hauge, A.; Breedveld, G. D. Chemosphere 2008, 72, 1581–1587.
ARTICLE
(37) McDonough, K. M.; Fairey, J. L.; Lowry, G. V. Water Res. 2008, 42, 575–584. (38) Cornelissen, G.; Wiberg, K.; Broman, D.; Arp, H. P. H.; Persson, Y.; Sundqvist, K.; Jonsson, P. Environ. Sci. Technol. 2008, 42, 8733–8739. (39) Ramos, E. U.; Meijer, S. M.; Vaes, W. J.; Verhaar, H. M.; Hermens, J.L. M. Environ. Sci. Technol. 1998, 32, 3430–3455.
6761
dx.doi.org/10.1021/ac201411v |Anal. Chem. 2011, 83, 6754–6761
