Environ. Sci. Technol. 2005, 39, 7529-7534
Comparative Sorption and Desorption of Benzo[a]pyrene and 3,4,3′,4′-Tetrachlorobiphenyl in Natural Lake Water Containing Dissolved Organic Matter JARKKO AKKANEN,* ANITA TUIKKA, AND JUSSI V. K. KUKKONEN Laboratory of Aquatic Ecology and Ecotoxicology, Department of Biology, University of Joensuu, P.O. Box 111, FIN-80101 Joensuu, Finland
The sorption and desorption of two model compounds, benzo[a]pyrene (BaP) and 3,4,3′,4′-tetrachlorobiphenyl (TCBP), were studied in natural lake water with high dissolved organic matter (DOM) content using the equilibrium dialysis and Tenax extraction methods. The sorption of TCBP was lower and reached steady value more slowly than did BaP. Tenax extraction revealed at least two differently desorbing fractions for both model compounds, which also supported the conclusion that DOM-HOC associations may involve several mechanisms. The rapidly desorbing fraction may be attributed to freely dissolved and loosely sorbed compound, whereas the more strongly sorbed fraction may indicate the presence of specific binding sites. The data indicated that the association between hydrophobic organic contaminants (HOC) and DOM is not simply absorption that is solely driven by the lipophilicity of the sorbates. Although contact time had a rather negligible effect on the sorption of BaP, the proportion of desorption resistant fraction increased with time, whereas the desorption of TCBP was less affected by contact time. Steric factors may be the cause of the lower sorption and smaller desorption resistant fraction of TCBP. The results indicate potential differences in the behavior of PAHs and PCBs in the aquatic environment.
Introduction Sorption to dissolved natural organic matter (DOM) influences the environmental fate of many contaminants in aquatic systems. DOM may, for example, increase the apparent water solubility of various hydrophobic organic contaminants (HOCs) (1-4) and promote mobilization of contaminants from soils and sediments (2, 5). Thus, both aqueous concentration and transport of HOCs can be affected depending on the type and concentration of DOM as well as on the characteristics of the contaminants (2, 5-7). In contrast to the enhanced concentration and mobilization of HOCs, DOM has been shown to decrease the bioavailability and toxicity of several contaminants (8, 9). The magnitude of the effects on bioavailability is also controlled by the type and concentration of DOM as well as by the characteristics of the contaminants. * Corresponding author phone: +358-13-251 3549; fax: +35813-251 3590; e-mail:
[email protected]. 10.1021/es050835f CCC: $30.25 Published on Web 08/27/2005
2005 American Chemical Society
As indicated above, the key factors controlling the magnitude of sorption are the characteristics of both DOM and the contaminants. Sorption of polycyclic aromatic hydrocarbons (PAHs) to DOM follows mostly the lipophilicity (octanol-water partitioning) of the compounds; that is, the higher is the lipophilicity, the higher is the sorption. Within and between other compound classes, additional factors also play a role. For instance, the sorption of polychlorinated biphenyls (PCB) to DOM appears to be lower than that of PAHs of similar lipohilicity (10-12). It has also been shown that a non-ortho-substituted PCB congener can be more extractable from natural lake water containing DOM with organic solvent than a PAH compound despite the slightly higher lipophilicity of the PCB congener (13). Furthermore, ortho-substituted PCB congeners demonstrate even lower sorption than non-ortho-substituted (i.e., coplanar) congeners (3, 14). This is probably due to the fact that orthosubstituted congeners cannot attain planar conformation, which hampers their interaction with DOM. However, this does not explain the lower sorption of non-ortho-substituted PCBs as compared to that of PAHs. Most of previous studies suggest that the sorption of PAHs to various types of dissolved humic substances is completed within minutes (15-17). These results are obtained by the fluorescence quenching method with relatively short contact times (up to 3 h). However, there are some indications that the sorption of PAHs to natural DOM takes more than a few hours to reach equilibrium (18). More polar compounds, such as chlorophenols, have been shown to require even days to reach equilibrium sorption (19). Several studies also indicate that short contact times (1-24 h) are enough to establish an interaction between DOM and HOC that reduces the bioavailability of the HOC in laboratory test systems (8, 13, 20). On the other hand, Haitzer et al. (21) showed decreasing bioavailability with increasing contact times up to 12 days for one DOM sample out of three tested. Longer contact times have rarely been studied, and it is still unclear what prolonged contact times, that is, “aging”, would mean for the sorption, desorption, or bioavailability of HOCs in natural waters. The desorption of sorbed HOCs has been studied with increasing interest during recent years. However, these studies have focused on sediments and soils. The stability of the interaction and desorption of DOM-sorbed compounds are still largely unknown. McCarthy and Jimenez (15) suggested that the sorption of benzo[a]pyrene to Aldrich humic acid was almost totally reversible and that a contact time up to 7 days had no influence on this. This suggests that sorption is simply an absorption (also referred to as phase partitioning) process to flexible organic matter. The above argument is supported by the fact that chemical concentration does not affect partition coefficients (10, 15, 22); that is, sorption isotherms are often found to be linear (15, 23). However, with highly hydrophobic chemicals, the water solubility limit can be reached before DOM becomes saturated by the sorbate. Laor and Rebhun (24) showed that a site-specific mechanism may also be important for sorption to DOM, observing nonlinearity in sorption isotherms for pyrene to humic acids. Schlebaum et al. (25) provided further evidence for “dual-mode” sorption, which encompasses both absorption and site-specific sorption types, showing a biphasic desorption for pentachlorobenzene in humic acid solution. Tenax-TA resin has been applied to the study of the desorption of organic contaminants from soils and sediments (26-29). In these experiments, the resin is considered to act VOL. 39, NO. 19, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7529
as an infinite sink for freely dissolved HOCs, thus enabling trapping and quantification of the release of compounds from soils and sediments. Because the resin floats on water, it can easily be removed from the experimental solutions. Recently, Zhao and Pignatello (30) verified the usefulness of Tenax-TA as a sink for organic compounds in solutions by studying the uptake of various organic compounds by the resin in water. In this study, we determined sorption and desorption of benzo[a]pyrene and 3,4,3′,4′-tetrachlorobiphenyl in natural lake water containing DOM. Equilibrium dialysis was used to determine the sorption, and the Tenax extraction method was used to quantify desorption. Furthermore, the effect of contact time between DOM and the model compounds on sorption and desorption was tested up to 60 days.
Experimental Section Model Compounds. Radiolabeled 14C-benzo[a]pyrene (BaP; Sigma, St. Louis, MO; 54.0 mCi/mmol) and 14C-3,4,3′,4′tetrachlorobiphenyl (TCBP; Sigma; 104.0 mCi/mmol) were used as model compounds. BaP is a five-ring PAH compound with a log Kow of 6.1 (31). TCBP (PCB 77) is a coplanar PCB with a log Kow of 6.6 (32). The radiochemical purity of the compounds was determined by a combination of thin-layer chromatography (plates developed in pentane:diethyl ether 9:1 v:v) and liquid scintillation counting. The radiochemical purity of the model compounds was proven to be >96%. To avoid photodegradation of the compounds, all experiments were conducted in the dark or under yellow fluorescent lights. Experimental Waters. Natural surface water was collected from Lake Kontiolampi in Eastern Finland (62°,43′,46′′ N, 29°,51′,18′′ E) and stored at 4 °C until use. Lake Kontiolampi is a small, natural freshwater lake with low ionic content ([Ca] + [Mg] 0.07 mmol L-1; conductivity 42 µS m-1) and low pH (5.1). Before use, the water was filtered through a 0.45µm membrane filter (Schleicher & Schuell, Dassel, Germany). After filtering, the concentration of total dissolved organic carbon (DOC) and inorganic carbon was 15.4 and 0.3 mg C L-1, respectively. In addition, soft synthetic freshwater (DOMfree; DOC < 0.2 mg C L-1) was prepared in Milli-Q-grade water (Millipore, Bedford, MA) by adding the following salts: CaCl2‚2H2O 11.8 mg/L, MgSO4‚2H2O 4.9 mg/L, NaHCO3 2.6 mg/L, and KCl 0.2 mg/L. The hardness of the synthetic water was 0.1 mmol L-1 expressed as [Ca] + [Mg], conductivity was 30 µS m-1, and pH was adjusted to 5. Sorption. The equilibrium dialysis method was used to compare the sorption of the model compounds to DOM and for quantification of the DOM-sorbed fraction. The method was adopted from previous studies (13, 33). Briefly, a 6-mL sample of the lake water was added to the dialysis tubes (Spectra/Por, cut off 1000 Da; Spectrum, Houston, TX), and the tubes, which were sealed with clips, were placed in glass jars containing the model compound (nominal aqueous concentration 2.5 nmol L-1) dissolved in 200 mL of synthetic freshwater. HgCl2 was added (50 mg L-1) to minimize biological activity. The jars were closed with screw caps and additionally sealed with laboratory film. Verification of the free diffusion of the model compounds through the dialysis membranes was carried out in a corresponding way, except that synthetic freshwater was added to the tubes instead of lake water. Four replicates were made for each sample. The jars were placed on a shaker (45 rpm) at 20 °C in the dark until the time of analysis. The concentration of the model compounds inside and outside the tubes was measured after 4, 7, 30, and 60 days. The concentration was measured by introducing 5-mL samples into scintillation vials, adding an equal volume of scintillation cocktail (Insta-Gel Plus, Packard, Groningen, The Netherlands) and analyzing by liquid scintillation counter (Wallac Winspectral 1414, Wallac, Turku, Finland). An external standard method was used, and the 7530
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 19, 2005
results were corrected for background and quenching. On this basis, the organic carbon normalized partition coefficients (KDOC) were calculated as follows:
KDOC ) Cs/(Cf‚DOC)
(1)
where Cs is the concentration of DOM-sorbed compound, that is, the difference between the aqueous concentrations inside and outside the dialysis bags. Cf is the concentration of freely dissolved compound, that is, aqueous concentration outside the dialysis bags. DOC is the concentration of dissolved organic carbon (kg L-1) in the sample. One of the drawbacks of equilibrium dialysis is the possible leakage of low molecular weight DOM through the dialysis membrane. This is especially true for bulk water samples that contain a variety of organic compounds of different molecular sizes. Thus, equilibrium dialysis could underestimate the sorbed fraction. On the other hand, molecules smaller than 1000 Da, which is the cutoff size of the dialysis membrane, are probably too small to carry other molecules such as PAHs (34). Equilibrium dialysis is therefore considered to be accurate enough for estimation of the sorbed fraction. Tenax Extraction. Before experiments, the resin was cleaned up with acetone, pure water, and hexane. The resin was shaken with water to remove water-soluble material and minimize any possible addition of soluble carbon from Tenax to the test solutions. Furthermore, all nonfloating material was removed in the water treatment by decanting. The resin was dried at 60 °C. The capacity of Tenax-TA-resin to take up the model compounds from DOM-free water was also tested. Sampling and analyses were carried out as described below for the desorption studies with the natural lake water. Tenax resin was shown to efficiently remove the model compounds from the DOM-free synthetic freshwater. After 20, 60, and 120 min of shaking with a single portion (0.25 g) of Tenax, 82%, 92%, and 94%, respectively, of BaP was detected in the resin, which verified that the amount of the resin was sufficient for the experiments. Recovery of BaP in the DOM-free tests was 97.1 ( 2.5%. In the case of TCBP, the tests in DOM-free media indicated some problems, because after 20 min over 100% of radioactivity was detected in Tenax, and therefore the recovery was well over 100% on most occasions. This may be due to the higher hydrophobicity of TCBP as compared to BaP. Thus, TCBP may be sorbed to the glass walls of the test tubes and may also become concentrated on the surface film of the water to a greater extent than BaP. Consequently, lower aqueous concentration is detected, and the amount of the compound added to the test tubes is underestimated. This, in turn, would result in recoveries higher than 100%, because it appears that the resin also efficiently scavenges the chemical from the surface film and the glass walls. After Tenax extraction, the empty test tubes were extracted with hexane, but the extraction yielded less than 0.05% of the total radioactivity added to the test tubes. In lake water, this problem was not obvious, probably because DOM increases the solubility of the model compounds. Lake water was spiked with the model compounds (individually) to give a nominal aqueous concentration of 2.5 nmol L-1. The concentration of the carrier (ethanol) was less than 0.0025% (v:v). To suppress microbiological activity, HgCl2 was added (50 mg L-1) to the test solutions. After 2 h of stirring, the spiked solutions were stored at +4 °C in the dark until commencement of the desorption experiments. Tenax extraction was performed after storage times (aging) of 1, 7, 30, and 60 days for BaP. For TCBP, the corresponding storage times were 1 and 60 days. After aging of the solutions, the concentration of the model compounds was measured by liquid scintillation counting to quantify the total amount of compounds to be placed in the test vessels. Duplicate
1-mL samples were taken, and 10 mL of liquid scintillation cocktail (Insta-Gel Plus) was added. The samples were analyzed with a liquid scintillation counter (Wallac Winspectral 1414). All liquid scintillation analyses were conducted by an external standard method and corrected for background and quenching. 55-mL glass test tubes (duplicate) were filled with 53 mL of the test solution, leaving a 2-mL headspace, 0.25 g of Tenax-TA resin (60/80 mesh, Buchem, Apeldoorn, The Netherlands) was added, and the test tubes were closed with screw caps. Tenax-TA-resin was used to remove the freely dissolved fraction of the model compounds from the solution. The tubes were placed on a rotary mixer for slow continuous mixing at 6 rpm. To follow the desorption of the model compounds from DOM, the resin was replaced 14 times within a period of 360 h. At each sampling time, the tubes were taken from the shaker and the resin was collected from the surface of the test solution with a stainless steel strainer. After that, a new of batch of clean resin was added to the tubes. For quantification of the desorbed compounds, every resin sample was extracted once with 5 mL of acetone and twice with 5 mL of hexane, after which the extracts were combined. The solvents were evaporated to approximate volume of 2 mL, 12 mL of scintillation cocktail was added, and the samples were analyzed with the liquid scintillation counter. This method was modified from the method developed for the measurement of desorption from sediment samples (29, 35). When using Tenax to study desorption from DOM, at least two kinds of phenomena that may distort the results are possible. First, DOM may be sorbed by the resin, and it may either reduce the measured desorption by competing with the model compounds for sorption sites or it may increase the measured desorption by removing any DOMassociated chemical from the system. It is quite likely that a fraction of DOM is removed by the resin; however, the lake water used in this study was colored even after the experiments. Zhao and Pignatello (30) also estimated that the sorption of DOM was insignificant also with considerably higher DOC concentrations than those used in the current study. Second, the resin might introduce DOC (not floating on the surface), which would sorb the model compounds and retain them in the test solutions, and this interaction would then erroneously be detected as an effect caused by the studied DOM sample. This is quite unlikely, because the resin was shown to efficiently remove the model compounds from DOM-free water. Furthermore, the contact time between the model compounds and DOM increased the desorption resistant fraction, which should not be the case if Tenax originating DOC was responsible for the dynamics of the system. The data obtained were fitted by using a two (rapid and slow)-phase model (28) to distinguish fractions desorbing at different rates. The model fitted the data well and gave high regression coefficients (coefficient of determination (COD) varied between 0.95 and 0.99). The two-phase model was considered to be sufficient to show that HOC-DOM association was based on more than one type of interaction and that contact time may have an effect on the association. However, this does not suggest that the number of the mechanisms of the interaction between DOM and HOC are limited to two. The model merely indicates that there are at least two fractions that differ with regards to the strength of their interaction with DOM:
St/S0 ) Frapide-krapidt + Fslowe-kslowt
(2)
where St and S0 indicate the amount of the studied compound in sediment at time t and in the beginning, respectively. F indicates the proportions of the fractions, which are rapidly (Frapid) or slowly (Fslow) extracted by Tenax from water, and
FIGURE 1. Sorption of the model compounds to DOM in equilibrium dialysis.
TABLE 1. Partition Coefficients between Water and DOM after Different Equilibration Periods log KDOC
equilibration time (d)
BaP
TCBP
4 7 30 60
5.91 ( 0.02 6.16 ( 0.04 6.08 ( 0.06 6.10 ( 0.02
5.03 ( 0.14 5.06 ( 0.06 5.41 ( 0.05 5.44 ( 0.08
k is a rate constant (h-1) for each corresponding fraction. In addition, the data were also modeled after subtraction of the freely dissolved fraction of the model compounds. The subtraction was done by estimating the freely dissolved fraction from the equilibrium dialysis and assuming that the freely dissolved fraction was removed totally in the first Tenax sampling. This was done to evaluate desorption of the truly sorbed fractions.
Results and Discussion Sorption. Equilibrium dialysis showed that 92.5 ( 1.1% of BaP was sorbed to DOM after 4 days of equilibration (Figure 1). A minor increase (up to 95.7 ( 0.3%) was found for sorption of BaP between days 4 and 7, but thereafter the fraction sorbed by DOM remained constant until the termination of the experiment (60 days). In principle, this was expected on the basis of previous studies using the fluorescence quenching method, which indicated rapid sorption of PAHs to humic and fulvic acids (16, 17, 36). The results obtained in the current study also agree with our previous study, in which 4 days was enough to reach equilibrium in dialysis for BaP with another water sample from the same lake (13). Note that in the previous study neither HgCl2 nor any other biocide was used in the dialysis. Corresponding KDOC values are listed in Table 1. TCBP behaved differently in the equilibrium dialysis. Sorption of TCBP to DOM increased up to 30 days from 61.7 ( 7.6% to 83.1 ( 1.4%; however, it was still lower than that of BaP at the end of the test period (Figure 1). This was also in line with our previous study, which showed an increasing trend for sorption of TCBP up to 7 days of dialysis with a sample from the same lake (13). The reason for the slower sorption is still unclear. Diffusion of TCBP through the dialysis membrane may be slower, although this is unlikely because the controls showed similar equilibrium for both model compounds. In addition, TCBP may establish the interaction with DOM more slowly than BaP, but there are no studies showing differences in sorption kinetics. Previous studies support the findings made in the present study about the lower sorption of halogenated HOCs as compared to that of PAHs. This has been observed with different types of aromatic DOM including humic acids (4), reverse osmosis isolates of DOM from natural waters (12), and natural water samples (11). The same observation has been made when using different techniques, such as reverseVOL. 39, NO. 19, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7531
phase separation, solubility enhancement, liquid-liquid extraction, and equilibrium dialysis methods (4, 10, 13). As already stated, studies concerning DOM-HOC interactions contact times are usually days at most, even though prolonged contact times might increase the environmental relevance of the results. However, in the case of equilibrium dialysis, the study compounds are usually added outside the dialysis bags (as in the present study), and therefore the compounds have to pass through the dialysis membrane via diffusion before coming to contact with DOM. Thus, the short-term contact time (first 2 or 3 days) in the equilibrium dialysis method is not directly comparable with methods in which the compounds come into contact with DOM immediately (e.g., fluorescence quenching, solubility enhancement, and reverse-phase separation). Various explanations have been suggested for the lower sorption of halogenated HOCs as compared to that of PAHs of equal or even lower lipophilicity. Cho et al. (4) hypothesized that the lower sorption of p,p′-DDT to humic acid as compared to that of pyrene is attributable to the threedimensional structure of the sorbates. It is assumed that planar pyrene fits more easily to sorption sites than the bulkier p,p′-DDT. However, it has to be kept in mind that TCBP is considered to be coplanar. In addition, it has been suggested that different electron densities around the molecules and different DOM fractions may play a role in the unequal sorption of PCBs and PAHs (37). In a previous study, streamwater containing DOM was fractioned with XAD-8 resin into three fractions, hydrophobic acids, hydrophobic neutrals, and hydrophilics (14). In this study, electrophilic TCBP had greater affinity toward the electron-rich fraction of DOM (hydrophobic neutrals) than to the other fractions, whereas electron-rich BaP had a high affinity to the hydrophobic acids fraction, which has low electron density. On the other hand, BaP appeared to have a relatively high affinity to hydrophobic neutrals fraction as well (14). Thus, the relative proportion of the different fractions of DOM may influence the magnitude of the sorption of PCBs. For example, a different sample from the same lake as used in this study was fractioned in a previous study, and 9% of DOM was shown to be hydrophilic neutrals (38). The fact that TCBP has a high affinity to that particular fraction, which constitutes only 9% of the total DOM, may offer an explanation for the lower sorption as compared to that of BaP. However, it has to be kept in mind that the exact proportion of hydrophilic neutrals in the current sample is not known. In any case, there appear to be sterical or other factors involved in causing the lower sorption of halogenated HOCs as compared to that of PAHs (39, 40). Tenax Extraction. A fraction of BaP was retained in water even after 360 h of Tenax extraction (Figure 2), which indicated that BaP was partly sorbed to desorption resistant domain (i.e., slowly desorbing fraction) within DOM. This in turn suggested that, in addition to simple absorption, DOM may also have specific sorption sites to which HOCs can be strongly sorbed. The presence of specific sorption sites or microenvironments that sequester HOCs in dissolved humic substances has also been anticipated in previous studies (16, 41). Furthermore, Schlebaum et al. (25) showed a biphasic desorption of pentachlorobenzene from humic acid (isolated from peat) solution and hypothesized that the reason for the existence of the slowly desorbing fraction is the fact that the sorbate is trapped by a change in the structure of the humic acid. On the other hand, McCarthy and Jimenez (15) concluded that sorption of BaP to Aldrich humic acid was almost totally reversible. In the present study, the strongly sorbed fraction was smaller for TCBP than for BaP (Figure 2). In fact, for the contact time of 1 day, the slowly desorbing fraction was quite negligible (Table 2). This may be due to 7532
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 19, 2005
FIGURE 2. Desorption curves for the model compounds after different contact periods. the same steric factors as those discussed in the case of lower sorption. Increasing the contact time from 1 to 30 days, the proportion of desorption resistant fraction of BaP in the water increased from 32% to 61% (Table 2). Correspondingly, the rapidly Tenax extractable fraction, which includes both free and rapidly desorbing compound, decreased. We can only speculate on reasons behind the effects of contact time on desorption. After the fast sorption, which may be mainly absorption, slower diffusion to specific sorption sites may occur. The alternative would be that after the rapid sorption the structure of DOM changes, trapping a fraction of the BaP molecules as proposed by Schlebaum et al. (25). Previously it has been demonstrated that the availability of PAHs in liquid-liquid extraction in the presence of DOM decreases over time (18). In that study, the time scale was comparable to that used in the present study. On the other hand, Schlebaum et al. (25) showed that contact time had no effect on desorption of pentachlorobenzene. In the current study, increasing contact time from 1 to 60 days increased the Fslow of TCBP in the water from 9% to 21% (Table 2) at the same time the proportion sorbed by DOM increased more (Figure 1). However, the Fslow remained still lower for TCBP than for BaP. When considering the rate constants presented in Table 2, a few things should be noted. On one hand, the Frapid modeled with the total chemical concentration includes freely dissolved and rapidly desorbing compound, and therefore the k-values are higher than they are when the freely dissolved fraction is subtracted before modeling. This increases the krapid for TCBP because equilibrium dialysis showed that a considerable fraction was freely dissolved. On the other hand, it may be that adsorption to the resin is not sufficiently fast to accurately quantify the rate constant for the Frapid on all occasions because of the presence of a freely dissolved fraction and the high desorption rate of the rapid fraction. The pretest showed that Tenax scavenged over 80% of BaP from the DOM-free synthetic freshwater in the first 20 min, whereas in the lake water tests the corresponding value was only 8-20% depending on the contact time. Correspondingly, Tenax scavenged 59-75% of TCBP from the lake water in 20 min. These data implied that the freely dissolved fraction in
TABLE 2. Desorbing Fractions and Corresponding Rate Constants for Both Model Compounds at Different Contact Timesa
a
contact time (d)
Frapid
krapid
1 (total) 1 (sorbed) 7 (total) 7 (sorbed) 30 (total) 30 (sorbed) 60 (total) 60 (sorbed)
0.66 ( 0.03 0.64 ( 0.03 0.54 ( 0.03 0.54 ( 0.02 0.35 ( 0.02 0.35 ( 0.01 0.37 ( 0.02 0.37 ( 0.01
0.86 ( 0.08 0.75 ( 0.07 0.52 ( 0.07 0.45 ( 0.06 0.33 ( 0.04 0.26 ( 0.03 0.31 ( 0.04 0.25 ( 0.02
1 (total) 1 (sorbed) 60 (total) 60 (sorbed)
0.91 ( 0.03 0.86 ( 0.05 0.79 ( 0.03 0.75 ( 0.04
5.15 ( 0.53 3.31 ( 0.49 4.09 ( 0.47 3.20 ( 0.43
BaP
TCBP
Fslow
kslow
COD
0.32 ( 0.02 0.34 ( 0.02 0.41 ( 0.02 0.42 ( 0.02 0.61 ( 0.01 0.63 ( 0.01 0.58 ( 0.01 0.60 ( 0.01
0.003 ( 0.0004 0.003 ( 0.0004 0.002 ( 0.0003 0.002 ( 0.0003 0.001 ( 0.0001 0.001 ( 0.0000 0.001 ( 0.0001 0.001 ( 0.0000
0.98 0.98 0.97 0.98 0.98 0.99 0.99 0.99
0.09 ( 0.01 0.13 ( 0.02 0.21 ( 0.01 0.25 ( 0.01
0.007 ( 0.003 0.006 ( 0.003 0.003 ( 0.001 0.003 ( 0.001
0.98 0.95 0.97 0.96
Note that the data were modeled with (total) and without (sorbed) freely dissolved fraction; COD ) coefficient of determination.
the lake water was removed in the first sampling. Table 2 also presents desorbing fractions and rate constants for data from which the freely dissolved fraction is subtracted. The changes were minor especially for BaP due to the low percentage of freely dissolved compound. The data suggested that the differences observed in desorption of the model compounds do not depend on the differences in the proportion of the sorbed fraction. Thus, certain issues concerning the desorption rates merit discussion. The krapid values were clearly higher for TCBP than for BaP (Table 2). This was partly due to a larger fraction of freely dissolved compound. With short contact times, over 50% of TCBP may be freely dissolved, which constitutes over half of the Frapid, and at 60 days, over 15% of TCBP may still be in freely dissolved form. The decrease of krapid with increasing contact time for TCBP presented in Table 2 can, at least partly, be attributed to the decrease in the freely dissolved fraction over time as measured by equilibrium dialysis. This was supported by the fact that the subtraction of freely dissolved fraction leveled the krapid values for TCBP (Table 2). For BaP, the decrease in krapid over time is harder to explain, because sorption does not increase in the same way as that of TCBP. Either the equilibrium dialysis method was not capable of detecting the changes in sorption or there may be a true change in the nature of interaction between DOM and BaP within the rapidly desorbing fraction. The kslow values for both model compounds were within a same range as those measured for laboratory-spiked sediments (29, 42). The curves for sediments have been modeled via a three-phase model, which divides the desorption into rapid, slow, and very slow fractions. In those sediments that were modeled by the triphasic model, the rate constant for the very slow fraction of BaP is an order of a magnitude lower than the kslow in the sediment and in the lake water in the present study. On the basis of the sorption and desorption data, KDOC values for the desorbing fractions can be calculated using the desorbing fraction instead of total sorbed fraction in the formula. The log KDOC values of the rapidly desorbing fraction are 5.7 ( 0.02 and 5.3 ( 0.08 for BaP and TCBP, respectively. Correspondingly, the KDOC values of the desorption resistant fraction are 5.9 ( 0.02 and 4.8 ( 0.08 for BaP and TCBP, respectively. This indicates that the largest difference in the sorption of the model compounds is in the desorption resistant fraction, indicating that the specific sorption sites are less accessible to TCBP. However, it has to be kept in mind that the experimental concentration of the hydrophobic model compounds is probably by far higher than the environmental concentrations at most sites. Therefore, the
specific sorption sites may be saturated in the experimental solution, and in natural situations at lower contaminant concentrations, the desorption resistant fraction could be more dominant in the case of TCBP.
Acknowledgments This study was financed by the European Commission (contract EVK1-CT-2001-00094) and the Academy of Finland (project no. 208783). We wish to thank Prof. Ismo J. Holopainen for his comments on earlier versions of the manuscript. The anonymous reviewers are acknowledged for their valuable comments.
Literature Cited (1) Chiou, C. T.; Malcolm, R. L.; Brinton, T. I.; Kile, D. E. Water solubility enhancement of some organic pollutants and pesticides by dissolved humic and fulvic acids. Environ. Sci. Technol. 1986, 20, 502-508. (2) Johnson, W. P.; John, W. W. PCE solubilization and mobilization by commercial humic acid. J. Contam. Hydrol. 1999, 35, 343362. (3) Uhle, M. E.; Chin, Y.-P.; Aiken, G. R.; McKnight, D. M. Binding of polychlorinated biphenyls to aquatic humic substances: The role of substrate and sorbate properties on partitioning. Environ. Sci. Technol. 1999, 33, 2715-2718. (4) Cho, H.-H.; Park, J.-W.; Liu, C. C. K. Effect of molecular structures on the solubility enhancement of hydrophobic organic compounds by environmental amphiphiles. Environ. Toxicol. Chem. 2002, 21, 999-1003. (5) Johnson, W. P.; Amy, G. L. Facilitated transport and enhanced desorption of polycyclic aromatic hydrocarbons by natural organic matter in aquifer sediments. Environ. Sci. Technol. 1995, 29, 807-817. (6) McCarthy, J. F.; Zachara, J. M. Subsurface transport of contaminants, Mobile colloids in the subsurface environment may alter the transport of contaminants. Environ. Sci. Technol. 1989, 23, 496-502. (7) Dunnivant, F. M.; Jardine, P. M.; Taylor, D. L.; McCarthy, J. F. Cotransport of cadmium and hexachlorobiphenyl by dissolved organic carbon through columns containing aquifer material. Environ. Sci. Technol. 1992, 26, 360-368. (8) Haitzer, M.; Ho¨ss, S.; Traunspurger, W.; Steinberg, C. Effects of dissolved organic matter (DOM) on the bioconcentration of organic chemicals in aquatic organisms. Chemosphere 1998, 37, 1335-1362. (9) Kukkonen, J. V. K. Toxicity and bioavailability of contaminants. In Limnology of Humic Waters; Keskitalo, J., Eloranta, P., Eds.; Backhuys Publishers: Leiden, The Netherlands, 1999; pp 117129. (10) Landrum, P. F.; Nihart, S. R.; Eadie, B. J.; Gardner, W. S. Reversephase separation method for determining pollutant binding to aldrich humic acid and dissolved organic carbon of natural waters. Environ. Sci. Technol. 1984, 18, 187-192. VOL. 39, NO. 19, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7533
(11) Kukkonen, J.; Oikari, A. Bioavailability of organic pollutants in boreal waters with varying levels of dissolved organic material. Water Res. 1991, 25, 455-463. (12) Akkanen, J.; Vogt, R. D.; Kukkonen, J. V. K. Essential characteristics of natural dissolved organic matter affecting the sorption of hydrophobic organic contaminants. Aquat. Sci. 2004, 66, 171177. (13) Akkanen, J.; Kukkonen, J. V. K. Measuring the bioavailability of two hydrophobic organic compounds in the presence of dissolved organic matter. Environ. Toxicol. Chem. 2003, 22, 518524. (14) Kukkonen, J.; McCarthy, J. F.; Oikari, A. Effects of XAD-8 fractions of dissolved organic carbon on the sorption and bioavailability of organic micropollutants. Arch. Environ. Contam. Toxicol. 1990, 19, 551-557. (15) McCarthy, J. F.; Jimenez, B. D. Interactions between polycyclic aromatic hydrocarbons and dissolved humic material: Binding and dissociation. Environ. Sci. Technol. 1985, 19, 1072-1076. (16) Schlautman, M. A.; Morgan, J. J. Effects of aqueous chemistry on the binding of polycyclic aromatic hydrocarbons by dissolved humic materials. Environ. Sci. Technol. 1993, 27, 961-969. (17) Perminova, I. V.; Grechishcheva, N. Y.; Petrosyan, V. S. Relationships between structure and binding affinity of humic substances for polycyclic aromatic hydrocarbons: Relevance of molecular descriptors. Environ. Sci. Technol. 1999, 33, 37813787. (18) Johnsen, S. Interactions between polycyclic aromatic hydrocarbons and natural aquatic humic substances: Contact time relationship. Sci. Total Environ. 1987, 67, 269-278. (19) Peuravuori, J.; Paaso, N.; Pihlaja, K. Sorption behaviour of some chlorophenols in lake aquatic humic matter. Talanta 2002, 56, 523-538. (20) Nikkila¨, A.; Kukkonen, J. V. K. Effects of dissolved organic material on binding and toxicokinetics of pyrene in the water flea Daphnia magna. Arch. Environ. Contam. Toxicol. 2001, 40, 333338. (21) Haitzer, M.; Burnison, B. K.; Ho¨ss, S.; Traunspurger, W.; Steinberg, C. E. W. Effects of quantity, quality, and contact time of dissolved organic matter on bioconcentration of benzo[a]pyrene in the nematode Caenorhabditis elegans. Environ. Toxicol. Chem. 1999, 18, 459-465. (22) Krop, H. B.; Van Noort, P. C. M.; Govers, H. A. J. Determination and theoretical aspects of the equilibrium between dissolved organic matter and hydrophobic organic micropollutants in water (Kdoc). Rev. Environ. Contam. Toxicol. 2001, 169, 1-122. (23) Laor, Y.; Rebhun, M. Complexation - Flocculation: a new method to determine binding coefficients of organic contaminants to dissolved humic substances. Environ. Sci. Technol. 1997, 31, 3558-3564. (24) Laor, Y.; Rebhun, M. Evidence for nonlinear binding of PAHs to dissolved humic acids. Environ. Sci. Technol. 2002, 36, 955961. (25) Schlebaum, W.; Badora, A.; Schraa, G.; Van Riemsdijk, W. H. Interactions between a hydrophobic organic chemical and natural organic matter: Equilibrium and kinetic studies. Environ. Sci. Technol. 1998, 32, 2273-2277. (26) Pignatello, J. J. Slowly reversible sorption of aliphatic halocarbons in soils. I. Formation of residual fractions. Environ. Toxicol. Chem. 1990, 9, 1107-1115. (27) Pignatello, J. J. Slowly reversible sorption of aliphatic halocarbons in soils. II. Mechanistic aspects. Environ. Toxicol. Chem. 1990, 9, 1117-1126. (28) Cornelissen, G.; Van Noort, P. C. M.; Govers, H. A. J. Desorption kinetics of chlorobenzenes, polycyclic aromatic hydrocarbons,
7534
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 19, 2005
(29)
(30) (31)
(32)
(33) (34) (35)
(36)
(37)
(38) (39)
(40) (41) (42)
and polychlorinated biphenyls: sediment extraction with tenax and effects of contact time and solute hydrophobicity. Environ. Toxicol. Chem. 1997, 16, 1351-1357. Kukkonen, J. V. K.; Landrum, P. F.; Mitra, S.; Gossiaux, D. C.; Gunnarsson, J.; Weston, D. Sediment characteristics affecting desorption kinetics of select PAH and PCB congeners for seven laboratory spiked sediments. Environ. Sci. Technol. 2003, 37, 4656-4663. Zhao, D.; Pignatello, J. J. Model-aided characterization of TenaxTA for aromatic compound uptake from water. Environ. Toxicol. Chem. 2004, 23, 1592-1599. De Maagd, P. G. J.; ten Hulscher, D. T. E. M.; van den Heuvel, H.; Opperhuizen, A.; Sijm, D. T. H. M. Physicochemical properties of polycyclic aromatic hydrocarbons: aqueous solubilities, n-octanol/water partition coefficients, and Henry’s law constants. Environ. Toxicol. Chem. 1998, 17, 251-257. Sabljic, A.; Gu ¨ sten, H.; Hermens, J.; Opperhuizen, A. Modeling octanol/water partition coefficients by molecular topology: Chlorinated benzenes and biphenyls. Environ. Sci. Technol. 1993, 27, 1394-1402. Carter, C. W.; Suffet, I. H. Binding of DDT to dissolved humic materials. Environ. Sci. Technol. 1982, 16, 735-740. Gustafsson, O ¨ .; Nilsson, N.; Bucheli, T. D. Dynamic water-colloid partitioning of pyrene through a coastal Baltic spring bloom. Environ. Sci. Technol. 2001, 35, 4001-4006. Cornelissen, G.; Van Noort, P. C. M.; Parsons, J. R.; Govers, H. A. J. Temperature dependence of slow adsorption and desorption kinetics of organic compounds in sediments. Environ. Sci. Technol. 1997, 31, 454-460. Gauthier, T. D.; Shane, E. C.; Guerin, W. F.; Seitz, W. R.; Grant, C. L. Fluorescence quenching method for determining equilibrium constants for polycyclic aromatic hydrocarbons binding to dissolved humic materials. Environ. Sci. Technol. 1986, 20, 1162-1166. McCarthy, J. F.; Strong-Gunderson, J.; Palumbo, A. V. The significance of interactions of humic substances and organisms in the environment. In Humic substances in the global environment and implications on human health; Senesi, M., Miano, T. M., Eds.; Elsevier: Amsterdam, The Netherlands, 1994; pp 981997. Akkanen, J.; Penttinen, S.; Haitzer, M.; Kukkonen, J. V. K. Bioavailability of atrazine, pyrene and benzo[a]pyrene in European river waters. Chemosphere 2001, 45, 453-462. Chin, Y.-P.; Aiken, G. R.; Danielsen, K. M. Binding of pyrene to aquatic and commercial humic substances: The role of molecular weight and aromaticity. Environ. Sci. Technol. 1997, 31, 1630-1635. Chiou, C. T.; McGroddy, S. E.; Kile, D. E. Partition characteristics of polycyclic aromatic hydrocarbons on soils and sediments. Environ. Sci. Technol. 1998, 32, 264-269. Engebretson, R. R.; von Wandruszka, R. Microorganization in dissolved humic acids. Environ. Sci. Technol. 1994, 28, 19341941. Leppa¨nen, M. T.; Landrum, P. F.; Kukkonen, J. V. K.; Greenberg, M. S.; Burton, G. A., Jr.; Robinson, S. D.; Gossiaux, D. C. Investigating the role of desorption on the bioavailability of sediment associated 3,4,3′,4′-tetrachlorobiphenyl in benthic invertebrates. Environ. Toxicol. Chem. 2003, 22, 2861-2871.
Received for review May 2, 2005. Revised manuscript received July 24, 2005. Accepted July 29, 2005. ES050835F
