Sorption of Ionic Surfactants to Estuarine Sediment and Their

The sorption of an anionic surfactant (sodium dodecyl sulfate; SDS) and a cationic surfactant (hexadecyl trimethylammonium bromide; HDTMA) to estuarin...
0 downloads 0 Views 168KB Size
Download PDF
Environ. Sci. Technol. 2005, 39, 1688-1697

Sorption of Ionic Surfactants to Estuarine Sediment and Their Influence on the Sequestration of Phenanthrene TRACEY JONES-HUGHES† AND ANDREW TURNER* School of Earth, Ocean and Environmental Sciences, University of Plymouth, Plymouth, PL4 8AA, UK

The sorption of an anionic surfactant (sodium dodecyl sulfate; SDS) and a cationic surfactant (hexadecyl trimethylammonium bromide; HDTMA) to estuarine sediment has been studied in river water and seawater. Sorption isotherms for SDS were essentially linear in both waters, suggesting a nonspecific, hydrophobic interaction between the SDS tail and particle surface. Sorption of HDTMA was considerably greater, more nonlinear, and more sensitive to water composition. These observations were attributed to a combination of both electrostatic and hydrophobic interactions between the surfactant and particle surface, the formation of admicelles, and salinityinduced structural alteration of the hydrophobic tail of the HDTMA molecule. Presence of SDS caused a reduction in the sorption of phenanthrene to estuarine sediment because of the competitive effects of the surfactant tail for hydrophobic sorption sites on the particle surface. Conversely, the presence of HDTMA caused significant enhancement in phenanthrene sequestration because of head-on sorption of surfactant molecules and a resulting, more hydrophobic particle surface. The most persistent feature of our results was an inverse dependence of unit sorption on particle concentration, and an empirical algorithm defining the effect was used to calculate the sediment-water fractionation of realistic concentrations of reactants in the estuarine water column. The results of these calculations, and the more general findings of this study, significantly improve our understanding of both the transport and fate of ionic surfactants in the estuarine environment, and the effects that these surfactants have on the partitioning of hydrophobic organic micropollutants.

Introduction Surfactants are a group of chemicals that comprise both polar and nonpolar regions, and are classified according to the nature of the hydrophile as anionic, cationic, nonionic, or amphoteric. Such amphiphilic properties allow surfactants to dissolve in both oil and water, (ad)sorb at interfaces, and solubilize hydrophobic compounds in micelles and at or within sorbed layers. Surfactants are, therefore, critical in a number of technologies, including detergency, emulsifica* Corresponding author phone: +44 1752 233041; fax: +44 1752 233035; e-mail: [email protected]. † Present address: Astra Zeneca, Brixham Environmental Laboratory, Freshwater Quarry, Brixham, TQ5 8BA, UK. 1688

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 6, 2005

tion, dispersion, coating, petroleum recovery, and adhesion (1). Surfactants are found in a plethora of formulations, but their major use by far is in household products (2). In the aquatic environment, the behavior, fate, and effects of surfactants are largely governed by their rate of degradation (3), tendency to form aggregates (or micelles) (4), and propensity to interact with natural particles (5). The latter is dependent on the nature of the structural groups at the particle surface, the molecular structure of the surfactant (the charge of the hydrophile, and the length and degree of branching of the hydrophobic group), and the composition of the aqueous phase. Regarding ionic surfactants, particlewater interactions may be extremely complex as, for a given chemical, a number of different sorption mechanisms is potentially involved, including ion exchange, ion pairing, polarization of π electrons, and hydrophobic bonding. While the sorption of individual surfactants or surfactant homologues to soils, river sediment, and clay minerals has been studied in some detail (6-8), their sorptive behavior in the estuarine environment has received very little attention (9). This is perhaps surprising because estuaries and coastal waters receive large quantities of surfactants through sewage and industrial effluents (10-12). Moreover, given that hydrophobic organic micropollutants (HOMs) are often codisposed with surfactants, and that the sorption of HOMs to solids may be considerably modified by surfactants (13-15), the presence of the latter in the estuarine environment has added significance. This paper reports on the first systematic investigation into the factors controlling the particle-water interactions of two contrasting but commercially important ionic surfactants in estuaries: sodium dodecyl sulfate (SDS), an anionic, alkyl sulfate, and hexadecyl trimethylammonium bromide (HDTMA), a cationic, quarternary ammonium compound. The physical and chemical properties of these surfactants and their important uses are listed in Table 1. We also examine the influence that these surfactants have on the sequestration of phenanthrene, a three-ringed polycyclic aromatic hydrocarbon. This HOM was selected for study as its physical and environmental properties are well-defined (see Table 1), it occurs at relatively high concentrations in many contaminated aqueous environments (20), and, while not acutely toxic itself, it serves as a useful proxy for more harmful compounds (21).

Experimental Section Sampling and Sample Characterization. Samples were collected from the River Plym and its estuary, in southwest England. The River Plym drains 80 km2 granitic moorland, and the shallow, macrotidal estuary is partly urbanized along its 5 km length. The Plym does not receive significant quantities of chemical contaminants, but samples were taken from locations as remote as possible from any direct anthropogenic inputs. River water from above the tidal limit and seawater from the English Channel (a few kilometers beyond the estuary mouth) were collected in a series of 500 mL Pyrex bottles that had previously been cleaned with Decon-90 detergent (Decon Laboratories, Bryn Mawr, PA) and ashed for 8 h at 500 °C. Samples were vacuum filtered in the laboratory through 0.7 µm pore size GF/F filters (Whatman, Maidstone, UK) within a few hours of collection, and were stored in their original bottles at 4 °C in the dark. Oxidized or “organicfree” river water and seawater were created by UV-irradiating a series of 50 mL filtrate aliquots in quartz cells for 4 h using a 400 W high-pressure Hg-vapor lamp. Previous studies 10.1021/es040077d CCC: $30.25

 2005 American Chemical Society Published on Web 02/05/2005

TABLE 1. Physicochemical Properties and Uses (Surfactants Only) of the Reactants Selected for Study (16-19) CAS #

mol wt, g mol-1

molecular formula

SDS

151-21-3

288.4

CH2(CH2)11OSO3Na

HDTMA

57-09-0

363.9

CH3(CH2)15N(CH3)3Br

phenanthrene

85-01-8

178.2

C14H10

surfactant

a

Aqueous solubility at 25 °C.

b

Cw,sat,a g L-1

CMC,b g L-1

Kowc

100

3.46

50

36

0.34

1500

0.0011

nad

37 000

uses wool washing, laundry, food, toothpaste, shampoo, pharmaceuticals textile softener, disinfectant, emulsifying agent, germicide, hair conditioner, oral antiseptic d na

Critical micelle concentration at 25 °C. Octanol-water partition coefficient.

undertaken by our laboratory have shown that this procedure removes between about 80% and 95% of organic matter (as dissolved organic carbon; DOC) from natural waters (22). The pH, conductivity, and salinity of the water samples were measured on site. Dissolved cation concentrations were determined in an aliquot of sample filtrate by inductively coupled plasma-atomic emission spectrometry (ICP-AES) using a Perkin-Elmer Optima 3000 spectrophotometer calibrated with acidified multi-element standards. For the determination of DOC, 100 mL aliquots of filtered river water and seawater were acidified with 300 µL of concentrated H3PO4 (BDH, Aristar; Poole, Dorset, UK) and stored at 4 °C in the dark before being analyzed using a Shimadzu TOC5000. Surficial, oxic sediment was collected from the intertidal zone of the Plym Estuary in a hexane-cleaned borosilicate jar. The mud fraction was retrieved by wet-sieving the sample through a 63 µm nylon mesh with the aid of a small quantity of filtered river water, and filtering the resulting slurry through 0.7 µm. The sediment was then resuspended in a few mL of river water and stored in a Pyrex vial at 4 °C in the dark. Meanwhile, an aliquot of sieved sediment was freeze-dried and chemically characterized. Thus, to evaluate the content of amorphous oxides, approximately 700 mg were digested in 0.05 M hydroxylamine hydrochloride in 25% v/v acetic acid (both BDH Aristar) for 24 h at room temperature (23). Digests were analyzed for Fe and Mn (and also Ca) by ICPAES as above. Particulate organic carbon (POC) concentration was determined on about 10 mg of sample using the Shimadzu TOC-5000 analyzer, and the specific surface area (SSA) of 200 mg of sample was established by multi-point BET nitrogen gas adsorption. Loss on ignition (LOI) was determined by combusting 300 mg of sample in a Carbolite ESF3 muffle furnace at 550 °C for 5 h. Oxidized or “organic-free” sediment was created by reacting 1 g of freeze-dried sample with 20 mL of H2O2 (BDH, Aristar) in a 50 mL Pyrex centrifuge tube. The mixture was left for about 24 h until effervescing had ceased, and 20 mL of Milli-Q water was then added and the slurry was centrifuged for 30 min at 3000 rpm. The supernatant was removed, and the particles were rinsed with Milli-Q water and recentrifuged. Sediment was then stored in a Pyrex vial at 4 °C in the dark as a suspension in either oxidized river water or oxidized seawater. This procedure has been shown to destroy greater than 90% of estuarine particulate organic matter (24). The Reactants. Unlabeled and radiolabeled (1-14C) SDS were purchased from the Sigma Chemical Co. (St. Louis, MO). A working stock solution for the SDS sorption studies was prepared in a Pyrex vial by the addition of 5 µL of radiolabeled SDS to 2 mL of unlabeled SDS (0.1% w/v in Milli-Q water), resulting in an activity of about 700 Bq per 25 µL. The stock was stored at 4 °C in the dark and was used within 5 days of preparation before being renewed if necessary. Unlabeled and radiolabeled (methyl-14C) HDTMA were purchased from the Sigma Chemical Co. and Tocris

c

d

Not applicable.

Cookson (Avonmouth, UK), respectively. A working stock solution for the HDTMA sorption studies was prepared in a Pyrex vial by the addition of 14 µL of radiolabeled HDTMA to 2 mL of unlabeled HDTMA (0.1% w/v in Milli-Q water), resulting in an activity of about 650 Bq per 25 µL. Because HDTMA precipitates at around 4 °C at concentrations encountered in the working stock, the solution was stored at room temperature and in the dark, and was used within 5 days of preparation before being renewed if necessary. Radiolabeled phenanthrene (9-14C) was purchased from Sigma, and a working stock of 925 Bq per 25 µL was prepared by appropriate dilution in HPLC-grade n-hexane (Rathburn Chemicals, Walkerburn, UK). This stock was stored in a vial at 4 °C in the dark and was used within about 3 weeks of preparation. Surfactant Sorption. Twenty mL of sample filtrate were pipetted into an ethanol-cleaned 30 mL Pyrex centrifuge tube, followed by a spike of sediment slurry to achieve a particle concentration of 50 mg L-1, and 10 µL of 3.5% v/v HgCl2 as a microbial inhibitor. Fifty µL radiolabeled surfactant stock were dispensed into the tube using a glass microsyringe, resulting in a surfactant concentration of 2.5 mg L-1, and the tube was then stoppered and covered with aluminum foil and the contents agitated on a 350 rpm wrist-action shaker at 20 °C and in the dark for 16 h. The contents were then centrifuged at 3000 rpm for 30 min, before 1 mL of supernatant was transferred to a scintillation vial containing 4 mL of Ultima Gold liquid scintillation cocktail (Canberra Packard, Boston, MA). The sediment pellet was resuspended in the remaining water, and a 1 mL aliquot was pipetted into a scintillation vial containing 2 mL of Insta-Gel Plus (Canberra Packard), and the contents were shaken briefly to form a gel. Scintillation vials were counted by liquid scintillation counting (LSC) using a Beckman LS6500 Multi-purpose scintillation system with factory-installed quench curves and luminescence correction. The aqueous concentration of surfactant, Cw (w/v), was derived directly from the activity in the supernatant, and the particulate (sorbed) concentration, Cp (w/w), was calculated from the difference between the activities in the gel and supernatant, and the concentration of sediment in suspension. The effects of particle concentration (between 20 and 600 mg L-1, and representative of turbidities encountered in macrotidal estuaries) and surfactant concentration (between about 0.2 and 3 mg L-1, and representative of total ionic surfactant concentrations encountered in highly contaminated surface waters; 17, 25) were examined by varying the volumes of sediment spike and surfactant stock, respectively. Each experiment was undertaken in quadruplicate, for both SDS and HDTMA, and in both river water and seawater. Selected experiments were also performed using oxidized water and sediment. Phenanthrene Sorption. To study the sorption of phenanthrene, we employed the same approach described above, but added a spike of phenanthrene stock (up to 50 µL, and equivalent to about 12 µg L-1) to the centrifuge tube first and VOL. 39, NO. 6, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1689

TABLE 2. Physicochemical Characteristics of the Samples Used in the Sorption Experiments sample

conductivity, µS cm-1

salinity

pH

Ca, mg L-1

DOC,a mg L-1

Plym river water English Channel seawater

100 43 000

1) of the sorption isotherm. After sample oxidation, sorption of HDTMA is significantly reduced (Figure 1d), and KF values are lower than those reported for SDS under equivalent conditions (Table 3). Removal of particulate and aqueous organic matter limits scope for hydrophobic interactions, which are highly significant for HDTMA. Sorption of HDTMA is, thus, restricted to electrostatic interactions between the cationic headgroup and negatively charged sites of inorganic minerals. As compared to river water, the negative charge on the particle surface in seawater is reduced because of the adsorption of divalent cations, and this is reflected by a lower value of KF under saline conditions. Phenanthrene Sorption. To establish the effects of ionic surfactants on the sorption of hydrophobic organic micropollutants (HOMs), we studied phenanthrene sorption to estuarine particles in river water and seawater, both in the presence and in the absence of SDS or HDTMA. We also examined whether the sequence of reactant introduction was significant to the sorption process by comparing phenanthrene uptake when surfactant was added before and after phenanthrene. The results of these experiments are shown in Figure 2, and estimates of sorption constants, derived from linear or nonlinear regression analyses of the data, are given in Table 3. Sorption in a Surfactant-Free System. Sorption isotherms for phenanthrene in the absence of surfactant are reasonably linear, indicating nonspecific (hydrophobic) interactions with the particle surface. Sorption is greater in seawater than river water because phenanthrene is salted out by the electrostrictive effects of dissolved ions. The effects of dissolved 1692

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 6, 2005

salts on the solubility and sorption of phenanthrene were empirically quantified using a modified form of the Setschenow equation (33). In its original form, this equation defines the change in solubility of a neutral chemical, Cw,sat, as a function of concentration of dissolved salt, Csalt (mol L-1), in terms of a salting constant, σ (L mol-1). Neglecting the possible effects of dissolved organic matter, for river water (rw) and seawater (sw) end-members we may write:

log(Cw,satrw/Cw,satsw) ) σCsalt

(3)

In the presence of suspended estuarine sediment particles, we may rewrite eq 3 in terms of the equivalent change in sorption constant (as the distribution coefficient) between the two end-members:

log(KDsw/KDrw) ) (σ + σsed)Csalt

(4)

Here, an additional constant, σsed, is necessary to account for changes in the sorptive properties (e.g., surface charge) of the suspended particles incurred by the effects of dissolved seawater ions. Equations 3 and 4 were solved for the salting constants by using concentrations of dissolved salt in river water and seawater of 0 and 0.54 mol L-1, respectively. A value of σ ) 0.16 L mol-1 was derived from determining the relative solubility of phenanthrene in river water and seawater in the absence of estuarine sediment according to the method outlined in Rawling et al. (34). In the presence of estuarine sediment, the gradients of the linear sorption isotherm fits, or KD values (Table 3), in river water and seawater were employed for each set of conditions. Sediment particles considerably enhance the salting out of phenanthrene (σ + σsed ) 0.33 L mol-1), presumably because the negative charge on the particle surface is partially countered by the effects seawater cations (27, 29). Effects of SDS on Phenanthrene Sorption. For the sorption of phenanthrene in the presence of SDS (Figure 2a

FIGURE 3. Sorption of phenanthrene (as KD) to 50 mg L-1 Plym estuarine particles suspended in river water as a function of preadded surfactant concentration (9, SDS; 2, HDTMA). Error bars represent the standard deviation about the mean of four separate experimental determinations. and b), we used a surfactant concentration of 2.5 mg L-1. The SDS isotherms indicate that this amount corresponds to a sorbed surfactant concentration of about 5 mg g-1 in both river water and seawater (Figure 1a). This quantity of SDS causes a significant reduction in phenanthrene sequestration, and this reduction exhibits no clear dependence on the order of reactant introduction (i.e., whether phenanthrene was introduced before or after the surfactant). A series of independent experiments involving both reactants in the absence of sediment revealed that this effect was not the result of enhanced solubility of phenanthrene in the presence of SDS. Rather, because the sorption of both anionic surfactants and HOMs proceeds mainly by hydrophobic interactions, the effects of SDS on phenanthrene sorption may be understood in terms of competition for mutually accessible sorption sites on hydrophobic regions of the particle surface. A similar mechanism has been proposed for phenanthrene sorption to soils in the presence of linear alkylbenzene sulfonates (35). Significantly, the greater variability of phenanthrene sorption to estuarine sediment in the presence of SDS implies a greater degree of heterogeneity in the sorption process when anionic surfactant is involved, and that phenanthrene sorption is reduced after addition of SDS indicates phenanthrene may be displaced from the particle surface by surfactant molecules. The effects of SDS on phenanthrene sorption are more pronounced in seawater than in river water. Thus, whereas phenanthrene is salted out in seawater when surfactant is not present, it appears to be salted in when SDS is present. Because SDS itself was not salted out in seawater (Figure 1a), we must assume that the surfactant selectively occupies sorption sites that are only available to phenanthrene in the presence of dissolved seawater ions (that is, when it undergoes salting out). In a separate set of experiments, we examined the sorption of a fixed quantity of phenanthrene as the concentration of (pre-added) SDS was varied. The results, exemplified in Figure 3, indicated a reduction in phenanthrene sorption as the concentration of SDS increased, consistent with a sorption mechanism involving competition between phenanthrene and surfactant molecules. Effects of HDTMA on Phenanthrene Sorption. Sorption isotherms for phenanthrene in the presence of the cationic surfactant, HDTMA, are shown in Figure 2c and d, and corresponding sorption constants are given in Table 3. The concentration of HDTMA employed in these experiments (2.5 mg L-1) corresponds to sorbed surfactant concentrations (at least in the absence of phenanthrene) of about 5 mg g-1 in river water and 40 mg g-1 in seawater (Figure 1c). Sorption data for phenanthrene in the presence of HDTMA are less variable than those involving the phenanthrene-SDS system,

and in both river water and seawater the sequestration of phenanthrene is enhanced by the presence of the cationic surfactant. Because HDTMA sorption proceeds via both hydrophobic and electrostatic interactions, the surfactant may not only compete with phenanthrene for neutral sorption sites on the particle surface, but also create additional sites for hydrophobic bonding by head-on sorption to negatively charged regions of the particle surface. Significantly, because natural organic matter contains a variety of highly polar functional groups, the hydrophobic tails of HDTMA molecules afford a more homogeneous and effective sorptive phase for HOMs (36). The persistent enhancement of phenanthrene sorption in the presence of HDTMA suggests that the effects of a more hydrophobic particle surface are more important than the effects of reactant competition under the experimental conditions employed. In river water, phenanthrene sorption to estuarine particles shows no clear dependency on the sequence of reactant introduction. In other words, regardless of when phenanthrene is added, additional sorption sites are equally available. In seawater, however, phenanthrene sorption is enhanced considerably when HDTMA is added beforehand. Salting out of the hydrophobic tail and the shielding effects of seawater ions both contribute to significantly greater sorption of HDTMA in seawater as compared to river water (Figure 1c). This results in greater and bilayered coverage of the particle surface by HDTMA molecules and, presumably, a more hydrophobic particle surface. These conditions are highly favorable for the sorption of phenanthrene, but are evidently only fully achieved when HDTMA is added prior to phenanthrene. When surfactant is added after phenanthrene, there is less scope for conditioning the particle surface, perhaps because pre-sorbed phenanthrene molecules either occupy sites that are otherwise favorable for HDTMA sorption, or interfere with the formation of HDTMA admicelles at the particle-water interface. The effects of HDTMA on the solubility and sorption of phenanthrene in river water and seawater may be quantified empirically by further modification of the Setschenow equation (eq 3):

log(KDsw/KDrw) ) (σ + σsed + σsurf)Csalt

(5)

where σsurf is a salting constant that accounts for the effects of surfactant on phenanthrene sorption. Using the gradients of the linear isotherm fits for phenanthrene sorption in river water and seawater when HDTMA is added beforehand yields a combined salting constant (σ + σsed + σsurf) of 0.44 L mol-1. An increase in HDTMA concentration was accompanied by an increase in unit sorption of phenanthrene to estuarine particles, as exemplified by the results shown in Figure 3. This is consistent with an increasing hydrophobicity of the particle surface and consequent sorption of HOM as surfactant coverage becomes denser and, in seawater, more bilayered. Particle Concentration Effect. The environmental variable that appears to exert most control on the sorption of organic and inorganic chemicals in the aquatic environment is particle concentration, or the ratio of sorbent to sorbate (31). Thus, an increase in particle concentration is usually accompanied by a reduction in unit sorption (as KD, for instance). The body of experimental and field evidence suggests that the particle concentration effect is related by one or more of the following: (i) the presence of a third complexing or sorbent (e.g., colloidal) phase that is, by operational definition, encompassed by the aqueous phase (37); (ii) particle-particle interactions that result in modification of the particle surface or chemical desorption (38); and (iii) a shift in the composition (and even origin) of the VOL. 39, NO. 6, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1693

as indicated in Figures 4 and 5. These observations are consistent with the assertion that the underlying cause(s) of the effect is related to some ubiquitous factor that is independent of the nature of the sorbate or sorbent and the environmental conditions. For each set of experimental results, the data were fitted with a simple power regression:

FIGURE 4. Distribution coefficients defining the sorption of (a) SDS and (b) HDTMA to Plym estuarine particles suspended in river water (4) and seawater (9) as a function of suspended particle concentration. Error bars represent the standard deviation about the mean of four separate experimental determinations. Parameters defining the regressions of log KD versus log SPM are given in Table 4. suspended particle load during particle resuspension or advection (39). A particle concentration effect was a persistent feature of our results for both surfactant and phenanthrene sorption,

log KD ) -b log SPM + log a

(6a)

KD ) aSPM-b

(6b)

where SPM denotes the concentration of suspended particulate matter (mg L-1), and a and b are data-fitted constants whose estimates are given in Table 4. The magnitude of a defines the extent of sorption normalized to a particle concentration of 1 mg L-1, and the magnitude of b defines the scale of the particle concentration effect. Absolute and relative magnitudes of sorption (as KD or a) in each set of experiments were broadly consistent with the sorption constants derived from the corresponding isotherm experiments (as discussed above). However, it must be appreciated that, in the present experiments, the ratio of sediment concentration to reactant concentration is also a variable, and so precise comparisons between the two sets of experiments are not necessarily appropriate. That said, in some cases the magnitude of the particle concentration effect and the value of b could be partly accounted for by the nature of the inherent sorption mechanism, as indicated by the corresponding sorption isotherm. This may be exemplified by considering the sorption of HDTMA to estuarine particles in river water and seawater, where isotherms were highly nonlinear. At high particle concentrations, there is a relatively low concentration of particulate HDTMA on a w/w basis and, consequently, a relatively low concentration of HDTMA in the aqueous phase. This is equivalent to a point on a sorption isotherm that is relatively close to the origin. Conversely, at low particle concentrations, there are relatively

FIGURE 5. Distribution coefficients defining the sorption of phenanthrene to Plym estuarine particles in the absence of surfactant (9), and in the presence of 2.5 mg L-1 surfactant that was added either before (O) or after (4) phenanthrene. (a) Sorption in river water in the absence and presence of SDS. (b) Sorption in seawater in the absence and presence of SDS. (c) Sorption in river water in the absence and presence of HDTMA. (b) Sorption in seawater in the absence and presence of HDTMA. Error bars are not shown for clarity. Parameters defining the regressions of log KD versus log SPM are given in Table 4. 1694

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 6, 2005

TABLE 4. Empirical Parameters Defining log Distribution Coefficient versus log Particle Concentration (Eq 6) for Each Set of Reactants and Conditions (Figures 4 and 5)a log a

b

r2

SDS river water seawater

3.78 3.18

0.56 0.27

0.717 0.631 (ns)

HDTMA river water seawater

4.75 6.56

0.27 1.01

0.841 0.943

Phenanthrene river water seawater river water + SDS before river water + SDS after seawater + SDS before seawater + SDS after river water + HDTMA before river water + HDTMA after seawater + HDTMA before seawater + HDTMA after

4.69 4.46 3.82 3.93 4.69 4.54 5.01 5.00 4.80 4.99

0.73 0.54 0.42 0.43 0.66 0.60 0.60 0.62 0.53 0.70

0.966 0.867 0.906 0.830 0.975 0.968 0.973 0.984 0.970 0.993

a Regression analyses were performed on mean values only; ns ) result of regression analysis was not significant (p > 0.1).

high concentrations of HDTMA on both the particles and in solution, and this is equivalent to a point on a sorption isotherm that is relatively remote from the origin. In river water, downward curvature (n < 1) of the HDTMA sorption isotherm (Figure 1c) means that the gradient (or the KD) is significantly greater under conditions equivalent to a higher particle concentration than conditions equivalent to a lower particle concentration. Consequently, the particle concentration effect is partly offset, and this is qualitatively consistent with a relatively low value of b in the particle concentration experiments (Table 4). By the same argument, the upward curvature (n > 1) of the isotherm for HDTMA sorption in seawater (and accompanying bilayer formation) is qualita-

tively consistent with a relatively high value of b (or strong particle concentration effect) in the corresponding particle concentration experiments. Partitioning and Retention of Ionic Surfactants and Phenanthrene in Estuaries. The results of the experiments in which particle concentration was varied allow us to predict the net phase partitioning of the surfactants and phenanthrene in the estuarine water column. The aqueous fraction of a chemical in a suspension, fw, is given as follows:

fw )

1 1 + (KDSPM/106)

(7)

where KD represents the sediment-water partitioning of the chemical at that SPM concentration. This equation can be rewritten to empirically account for the particle concentration effect (eq 6) as follows (40):

fw )

1

(8)

1 + (aSPM-(b-1)/106)

The calculated aqueous fraction of ionic surfactant (encompassing any surfactant in a “third” phase) is shown as a function of SPM, up to a concentration of 1000 mg L-1, in Figure 6a and b. With respect to SDS, these calculations indicate a small reduction in fw with increasing SPM, but that greater than 80% occurs in the aqueous phase in both river water and seawater. Given that degradation of SDS in sediment suspensions is reasonably rapid (∼0.3 mg L-1 h-1; 3), we may conclude that there is relatively little scope for its retention or accumulation in estuaries. Regarding HDTMA, there is considerably greater tendency for interaction with suspended estuarine particles. In river water, fw decreases rapidly with increasing particle concentration such that about 10% exists in the aqueous phase at a particle concentration of 1000 mg L-1. In seawater, about 20% HDTMA is predicted to occur in the aqueous phase, regardless of particle concentration, because b has approximately unit value. Based

FIGURE 6. Fraction of reactant in the aqueous phase of the water column, calculated according to eq 8 and using coefficients given in Table 4, as a function of suspended particle concentration. (a) SDS and HDTMA in river water. (b) SDS and HDTMA in seawater. (c). Phenanthrene in river water in the absence and presence of 2.5 mg L-1 pre-added SDS or HDTMA. (d) Phenanthrene in seawater in the absence and presence of 2.5 mg L-1 pre-added SDS or HDTMA. VOL. 39, NO. 6, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1695

on sorptive considerations, therefore, HDTMA is more likely to be retained in an estuarine environment than SDS, and its retention is predicted to be accentuated under saline conditions. Moreover, available data on its decomposition (41) suggest that HDTMA is not readily degraded within typical estuarine flushing times (up to a few months), further favoring its retention. With respect to phenanthrene (Figure 6c and d), the calculated aqueous fraction decreases with increasing particle concentration in both river water and seawater, and is lower in the latter because of the effects of salting out of the solute and sediment organic matter. Depending on particle concentration, the pre-addition of SDS causes either a small increase or a small reduction in the aqueous fraction of phenanthrene, while pre-addition of HDTMA cause a significant reduction in the aqueous fraction of phenanthrene throughout the particle concentration range studied. These calculations, and the more general findings of this study, have afforded a valuable insight into magnitudes, mechanisms, and effects of ionic surfactant sorption in estuaries. Quantitatively, our results are specific to the experimental conditions and reactants employed. However, given that many of the effects we have observed are similar to those observed in independent, surfactant-enhanced remediation studies (4, 14, 42, 43), it is reasonable to assume that, qualitatively, they are sufficiently general to apply to other estuaries, and to different ionic surfactants and HOMs. Our findings are also predicted to be of significance to other estuarine biogeochemical processes. For instance, sorbed surfactants have been shown to influence the bioavailability and rates of degradation of co-sorbed HOMs (44, 45) and may stimulate the attachment of bacteria to the particle surface (46), in turn altering the extents and rates of microbially mediated chemical reactions. Further study would be required to verify and define these processes under estuarine conditions.

Acknowledgments We thank Dr. M. C. Rawling, Mr. N. Crocker, Dr. M. Williams, and Mr. K. Solman (UoP) for technical assistance throughout the study, and Prof. G. Millward (UoP) and Dr. R. Thompson (Astra Zeneca) for useful discussions about the work. The financial contributions from Astra Zeneca and British Maritime Technology were greatly appreciated.

Literature Cited (1) Sharma, R. Small-molecule surfactant adsorption, polymer surfactant adsorption and surface solubilization: an overview. In Surfactant Adsorption and Surface Solubilization; Sharma, R., Ed.; ACS Symposium Series, RSC: London, 1995; pp 1-20. (2) Karsa, D. R.; Bailey, R. M.; Shelmerdine, B.; McCann, S. A. A decade of change in the surfactant industry. In Industrial Applications of Surfactants IV; Karsa, D. R., Ed.; RSC: Cambridge, 1999; pp 1-22. (3) Marshall, S. J.; House, W. A.; White, G. F. Role of natural organic matter in accelerating bacterial biodegradation of sodium dodecyl sulfate in rivers. Environ. Sci. Technol. 2000, 34, 22372242. (4) Ko, S.-O.; Schlautman, M. A.; Carraway, E. R. Effects of solution chemistry on the partitioning of phenanthrene to sorbed surfactants. Environ. Sci. Technol. 1998, 32, 2769-2775. (5) Atay, N. Z.; Yenigu ¨ n, O.; Asutay, M. Sorption of anionic surfactants SDS, AOT and cationic surfactant Hyamine 1622 on natural soils. Water, Air, Soil Pollut. 2002, 136, 55-67. (6) Adeel, Z.; Luthy, R. G. Sorption and transport kinetics of a nonionic surfactant through an aquifer sediment. Environ. Sci. Technol. 1995, 29, 1032-1042. (7) Brownawell, B. J.; Chen, H.; Zhang, W.; Westall, J. C. Adsorption of surfactants. In Organic Substances and Sediments in Water; Baker, R. A., Ed.; Lewis: Chelsea, MI, 1991; pp 127-147. 1696

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 6, 2005

(8) Cano, M. L.; Dorn, P. B. Sorption of two model alcohol ethoxylate surfactants to sediments. Chemosphere 1996, 33, 981-994. (9) Amano, K.; Fukushima, T.; Nakasugi, O. Fate of linear alkylbenzenesulfonates in a lake estuary. Water Sci. Technol. 1991, 23, 497-506. (10) Takada, H.; Ogura, N.; Ishiwatarl, R. Seasonal-variations and modes of riverine input of organic pollutants to the coastal zone. 1. Flux of detergent-derived pollutants to Tokyo Bay. Environ. Sci. Technol. 1992, 26, 2517-2523. (11) Terzic, S.; Ahel, M. Input and behavior of linear alkylbenzenesulphonates (LAS) in a stratified estuary. Mar. Pollut. Bull. 1994, 28, 735-740. (12) Leon, V. M.; Saez, M.; Gonzalez-Mazo, E.; Gomez-Parra, A. Occurrence and distribution of linear alkylbenzene sulfonates and sulfophenylcarboxylic acids in several Iberian littoral ecosystems. Sci. Total Environ. 2002, 288, 215-226. (13) John, W. W.; Bao, G.; Johnson, W. P.; Stauffer, T. B. Sorption of nonionic surfactant oligomers to sediment and PCE DNAPL: effects on PCE distribution between water and sediment. Environ. Sci. Technol. 2000, 34, 672-679. (14) Gao, B.; Wang, X.; Zhao, J.; Sheng, G. Sorption and cosorption of organic contaminant on surfactant-modified soils. Chemosphere 2001, 43, 1095-1102. (15) Park, S.-K.; Bielefeldt, A. R. Aqueous chemistry and interactive effects on nonionic surfactant and pentachlorophenol sorption to soil. Water Res. 2003, 37, 4663-4672. (16) Rosen, M. J. Surfactants and Interfacial Phenomena; John Wiley: New York, 1989. (17) Painter, H. Anionic surfactants. In The Handbook of Environmental Chemistry, vol. 3F. Anthropogenic Compounds - Detergents; Hutzinger, O., Ed.; Springer: Berlin, 1992; pp 1-88. (18) Mackay, D.; Shiu, W. Y.; Ma, K. C. Illustrated Handbook of Physico-Chemical Properties and Environmental Fate of Organic Chemicals, vol. 2: Polynuclear Aromatic Hydrocarbons, Polychlorinated Dioxins and Dibenzofurans; Lewis: Michigan, 1992. (19) Syracuse Research Corp. Database. http://esc.syrres.com/ interkow/ (accessed May 2004). (20) Mackay, D.; Hickie, B. Mass balance model of source apportionment, transport and fate of PAHs in Lac Saint Louis, Quebec. Chemosphere 2000, 41, 681-692. (21) Laor, Y.; Farmer, W. J.; Aochi, Y.; Strom, P. F. Phenanthrene binding and sorption to dissolved and mineral-associated humic acid. Water Res. 1998, 32, 1923-1931. (22) Martino, M.; Turner, A.; Millward, G. E. Influence of organic complexation on the adsorption kinetics of nickel in river waters. Environ. Sci. Technol. 2003, 37, 2383-2388. (23) Turner, A. Trace metal contamination in sediments from UK estuaries: an empirical evaluation of the role of hydrous iron and manganese oxides. Estuarine, Coastal Shelf Sci. 2000, 50, 355-371. (24) Turner, A.; Millward, G. E.; Le Roux, S. M. Significance of oxides and particulate organic matter in controlling trace metal partitioning in a contaminated estuary. Mar. Chem. 2004, 88, 179-192. (25) Boethling, R. S.; Lynch, D. G. Quarternary ammonium surfactants. In The Handbook of Environmental Chemistry, vol. 3F. Anthropogenic Compounds - Detergents; Hutzinger, O., Ed.; Springer: Berlin, 1992; pp 145-177. (26) Armstrong, D. W.; Stine, G. Y. Separation and quantitation of anionic, cationic and nonionic surfactants by TLC. J. Liq. Chromatogr. 1983, 6, 23-33. (27) Hunter, K. A.; Liss, P. S. Organic matter and the surface charge of suspended particles in estuarine waters. Limnol. Oceanogr. 1982, 27, 322-335. (28) Garnier, J.-M.; Martin, J.-M.; Mouchel, J.-M.; Thomas, A. J. Surface reactivity of the Rhoˆne suspended matter and relation with trace element sorption. Mar. Chem. 1991, 36, 267-289. (29) Beckett, R.; Le, N. P. The role of organic matter and ionic composition in determining the surface charge of suspended particles in natural waters. Colloids Surf. 1990, 44, 35-49. (30) Wang, W.; Gu, B. H.; Liang, L. Y. Effect of surfactants on the formation, morphology, and surface property of synthesized SiO2 nanoparticles. J. Dispersion Sci. Technol. 2004, 25, 593601. (31) Schwarzenbach, R. P.; Gschwend, P. M.; Imboden, D. M. Environmental Organic Chemistry; John Wiley: New York, 1993. (32) Xu, S.; Boyd, S. A. Cationic surfactant sorption to a vermiculitic subsoil via hydrophobic bonding. Environ. Sci. Technol. 1995, 29, 312-320.

(33) Turner, A.; Millward, G. E.; Le Roux, S. M. Sediment-water partitioning of inorganic mercury in estuaries. Environ. Sci. Technol. 2001, 35, 4648-4654. (34) Rawling, M. C.; Turner, A.; Tyler, A. O. Particle-water interactions of 2,2′,5,5′-tetrachlorobiphenyl under simulated estuarine conditions. Mar. Chem. 1998, 61, 115-126. (35) Ou, Z.; Yediler, A.; He, Y.; Kettrup, A.; Sun, T. Effects of linear alkylbenzene sulfonate (LAS) on the adsorption behavior of phenanthrene on soils. Chemosphere 1995, 30, 313-325. (36) Lee, J.; Crum, J. R.; Boyd, S. A. Enhanced retention of organic contaminants by soil exchanged with organic cations. Environ. Sci. Technol. 1989, 23, 1365-1372. (37) Koelmans, A. A.; Lijklema, L. Sorption of 1,2,3,4-tetrachlorobenzene to sediments: the application of a simple three phase model. Chemosphere 1992, 25, 313-325. (38) Severtson, S. J.; Banerjee, S. Mechanistic model for collisional desorption. Environ. Sci. Technol. 1993, 27, 1690-1692. (39) Turner, A. Trace metals partitioning in estuaries: importance of salinity and particle concentration. Mar. Chem. 1996, 54, 27-39. (40) Turner, A.; Hyde, T. L.; Rawling, M. C. Transport and retention of hydrophobic organic micropollutants in estuaries: implications of the particle concentration effect. Estuarine, Coastal Shelf Sci. 1999, 49, 733-746.

(41) Li, Z.; Roy, S. J.; Zou, Y.; Bowman, R. S. Long-term chemical and biological stability of surfactant-modified zeolite. Environ. Sci. Technol. 1998, 32, 2628-2632. (42) Yang, L. Y.; Zhou, Z.; Xiao, L.; Wang, X. R. Chemical and biological regeneration of HDTMA-modified montmorillonite after sorption with phenol. Environ. Sci. Technol. 2003, 37, 5057-5061. (43) Kibbey, T. C. G.; Hayes, K. F. Partitioning and UV absorption studies of phenanthrene on cationic-coated silica. Environ. Sci. Technol. 1993, 27, 2168-2173. (44) Crocker, F. H.; Guerin, W. F.; Boyd, S. A. Bioavailability of naphthalene sorbed to cationic surfactant-modified smectite clay. Environ. Sci. Technol. 1995, 29, 2953-2958. (45) Guerin, W. F.; Boyd, S. A. Bioavailability of naphthalene associated with natural and synthetic sorbents. Water Res. 1997, 31, 1504-1512. (46) White, G. F. Multiple interactions in riverine biofilms - surfactant adsorption, bacterial attachment and biodegradation. Water Sci. Technol. 1995, 31, 61-70.

Received for review July 12, 2004. Revised manuscript received December 13, 2004. Accepted December 15, 2004. ES040077D

VOL. 39, NO. 6, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1697