Dissolved organic matter effects on the performance of a barrier to polycyclic aromatic hydrocarbon transport by groundwater

Journal of Contaminant Hydrology 60 (2003) 307 – 326 www.elsevier.com/locate/jconhyd

Dissolved organic matter effects on the performance of a barrier to polycyclic aromatic hydrocarbon transport by groundwater Jung-Won Moon a, Mark N. Goltz b, Kyu-Hong Ahn c, Jae-Woo Park d,* a

National Subsurface Environmental Research Laboratory (NSERL), Ewha Womans University, 11-1 Daehyon-dong, Seodaemun-gu, Seoul 120-750, South Korea b Air Force Institute of Technology, Wright-Patterson AFB, Ohio 45433-7765, USA c Korea Institute of Science and Technology, PO Box 131, Cheongryang, Seoul 130-650, South Korea d Department of Civil Engineering, Hanyang University, 17 Haengdang-dong, Seongdong-gu, Seoul 133-791, South Korea Received 17 July 2001; received in revised form 24 June 2002; accepted 26 June 2002

Abstract In order to contain the movement of organic contaminants in groundwater, a subsurface sorption barrier consisting of sand or clay minerals coated with a cationic surfactant has been proposed. The effectiveness of such a sorption barrier might be affected by the presence of dissolved organic matter (DOM) in the groundwater. To study the impact of DOM on barrier performance, a series of batch experiments were performed by measuring naphthalene and phenanthrene sorption onto sand coated with cetylpyridinium chloride (CPC) and bentonite coated with hexadecyltrimethylammonium bromide (HDTMA) in the presence of various concentrations of DOM. The overall soil – water distribution coefficient (K*) of naphthalene and phenanthrene onto CPC-coated sand decreased with increasing DOM concentration, whereas the K* of the compounds onto HDTMAcoated bentonite slightly increased with increasing DOM concentration. To describe the overall distribution of polycyclic aromatic hydrocarbons (PAHs) in the systems, a competitive multiphase sorption (CMS) model was developed and compared with an overall mechanistic sorption (OMS) model. The modeling studies showed that while the OMS model did not explain the CPC-coated sand experimental results, a model that included competitive sorption between DOM and PAH did. The experimental results and the modeling study indicated that there was no apparent competition between DOM and PAH in the HDTMA-coated bentonite system, and indicated that in

*

Corresponding author. E-mail address: [email protected] (J.-W. Park).

0169-7722/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 9 - 7 7 2 2 ( 0 2 ) 0 0 0 8 4 - 0

308

J.-W. Moon et al. / Journal of Contaminant Hydrology 60 (2003) 307–326

groundwater systems with high DOM, a barrier using HDTMA-coated bentonite might be more effective. D 2002 Elsevier Science B.V. All rights reserved. Keywords: Surfactant; Immobilization; Dissolved organic matter; Polycyclic aromatic hydrocarbons; Distribution coefficient

1. Introduction The widespread occurrence of polycyclic aromatic hydrocarbons (PAHs) in soil and groundwater has become an important environmental concern. A number of studies have focused on the fate and transport of these pollutants and application of remedial technologies to manage them. One remediation strategy to manage PAH-contaminated soil is to mobilize the contaminants by flushing the area with a surfactant solution (Ko et al., 1998; Guha et al., 1998; Finkel et al., 1999). The surfactant causes the PAH to be desorbed into the aqueous phase, which is then removed from the subsurface. This strategy, however, has several disadvantages. Not only must the carrier solution be pumped out of the ground and treated, but also, because the contaminant is mobilized, there is a risk that areas that were not contaminated prior to treatment may become contaminated. An alternative remediation strategy is to use immobilization to prevent the spread of contamination. One immobilization technique is to use a barrier of sand or clay minerals coated with a cationic surfactant to promote the sorption of organic pollutants, thereby preventing their transport. Specific applications for immobilization barriers that have been suggested include: water purification, industrial wastewater treatment, remediation of contaminated groundwater, and landfill liners (Jaynes and Vance, 1996; Zhao et al., 1996). Sorption barriers may also be used in conjunction with permeable reactive barriers (PRBs), serving as a pretreatment barrier, and hence reducing the pollutant loading on the PRB. Various factors, such as pH, extent of surfactant coverage on the mineral surface, presence of dissolved organic matter (DOM), and ions, can control the effectiveness of the technique (Boyd et al., 1988; Lee et al., 1989, 1990). DOM, which is ubiquitous in aquatic systems, is largely composed of humic substances. Due to its high organic content, DOM serves as a sorbent to hydrophobic organic compounds such as PAHs (Zimmer et al., 1989). Owing to the prevalence and relative stability of DOM in natural water, it is likely that it can significantly affect PAH sorption. The hydrophobic sites in DOM can form a complex with hydrophobic organic compounds, such as PAHs. Many studies have demonstrated that DOM enhances the apparent water solubility and mobility of highly hydrophobic contaminants (e.g. McCarthy and Jimenez, 1985). Conversely, due to its hydrophobicity, DOM can be also sorbed onto soil organic material, thus increasing the sorption and reducing the mobility of contaminants that may be associated with the DOM (Kile and Chiou, 1989). Therefore, as a mobile phase, DOM can enhance the mobility of hydrophobic organic compounds such as PAHs, or as a sorbed phase, it can increase PAH sorption and decrease mobility (Karickhoff et al., 1979; Chiou et al., 1987; McCarthy and Zachara, 1989; Magee et al., 1991; Liu and Amy, 1993; Johnson and Amy, 1995).

J.-W. Moon et al. / Journal of Contaminant Hydrology 60 (2003) 307–326

309

Previous studies have shown that the presence of DOM facilitated PAH mobilization in soil –water systems (Rebhun et al., 1996; McGinley et al., 1996; Lu¨hrmann and Noseck, 1998). However, the sorbents that were used in these studies were relatively weak, such as natural soil or aluminum oxide. These systems have been modeled assuming the distribution of PAH in the sorbed and aqueous phases is dependent upon the DOM distribution (Lee and Kuo, 1999). For a strong sorbent, such as the surfactant-coated minerals that would be used in an immobilization barrier, the effect of DOM on PAH transport has not been studied. In this work, in order to determine how the PAH sorption capacity of surfactantmodified sorbent is influenced by the presence of DOM, a series of batch experiments were performed with cetylpyridinium chloride (CPC)-coated sand and hexadecyltrimethylammonium bromide (HDTMA)-coated bentonite. Experimental results were described and an apparent overall sorption distribution coefficient (K*) for PAH sorption on surfactant-modified sorbent in the presence of DOM was predicted using a competitive multiphase sorption (CMS) model which was developed based upon ideal adsorbed solution theory (IAST). Predictions of the CMS model were compared with predictions of an existing model, the overall mechanistic sorption (OMS) model, which was developed for weak sorbent systems (Lee and Kuo, 1999; Lee et al., 2000).

2. Theory The conceptual model of PAH sorption onto a surfactant-modified sorbent in the presence of DOM is depicted in Fig. 1. Lee and Kuo’s (1999) OMS model assumes linear equilibrium partitioning between (1) DOM and sorbent, (2) PAH and DOM, and (3) PAH

Fig. 1. Conceptual model.

310

J.-W. Moon et al. / Journal of Contaminant Hydrology 60 (2003) 307–326

and sorbent. With these assumptions, the OMS model expresses the distribution coefficient describing overall PAH partitioning between the sorbed and aqueous phases as: * ¼ KOMS

DOM P DOM P K2;3 K1;2 CW þ K1;4 P C DOM 1 þ K1;2 W

ð1Þ

DOM where K2,3 (l3 M  1) is the distribution coefficient describing DOM partitioning P between the sorbed and aqueous phases, K1,2 (l3 M  1) is a coefficient describing P partitioning of aqueous phase PAH into aqueous phase DOM, K1,4 (l3 M  1) is the distribution coefficient describing PAH partitioning between the sorbed and aqueous phases, and CwDOM (M l  3) is the equilibrium concentration of DOM in the aqueous phase. Eq. (1) implicitly assumes that the distribution between PAH and sorbed DOM P (K1,3 (l3 M  1) in Fig. 1) is the same as that between PAH and DOM in the aqueous phase P P P DOM (K1,2 ). The OMS model also assumes that K1,2 , K1,4 , and K2,3 are constants, DOM independent of either DOM (Cw ) or PAH concentration. The competitive multiphase sorption (CMS) model proposed in this work does not make this assumption. Specifically, as will be shown below, the CMS assumes that the distribution coefficient describing PAH P partitioning between the aqueous and sorbed phases (K1,4 ) is dependent on CwDOM. The mass balance of DOM can be:

DOM M DOM ¼ CW VW þ qDOM mads S

ð2Þ

DOM where MDOM is the total mass of DOM in the system and CW (M l  3) and qSDOM (M 1 M ) denote the aqueous and sorbed phase concentrations of DOM, respectively. Vw is water volume and mads is the mass of adsorbent. By these definitions, the distribution DOM coefficient describing DOM partitioning between the sorbed and aqueous phases, K2,3 3 1 (l M ), is:

DOM K2;3 ¼

qDOM S DOM CW

ð3Þ

After substituting Eq. (3) into Eq. (2), the following equation is obtained: DOM DOM DOM VW þ CW K2;3 mads M DOM ¼ CW

ð4Þ

The overall mass balance of PAH (P) can be expressed as: MP ¼ MW þ MS

ð5Þ

MW ¼ ðCfr þ Cb ÞVW

ð6Þ

DOM DOM MS ¼ ðqP3 qDOM þ qP4 Þmads ¼ ðqP3 K2;3 Cw þ qP4 Þmads S

ð7Þ

where MP denotes total mass of P, Mw and Ms denote mass of P in the aqueous and sorbed phases, respectively, Cfr (M l  3) is the concentration of P in the aqueous phase that is not associated with DOM (Phase 1), Cb (M l  3) is the concentration of P in the aqueous phase

J.-W. Moon et al. / Journal of Contaminant Hydrology 60 (2003) 307–326

311

that is partitioned into DOM (Phase 2), q3P (M M  1) is the concentration of sorbed P that is associated with DOM (Phase 3), q4P (M M  1) is the concentration of sorbed P that is not associated with DOM (Phase 4). The distribution of PAH between Phases 1 and 2 is defined as: Cb P K1;2 ¼ DOM ð8Þ CW Cfr while partitioning between Phases 1 and 3 can be expressed as: P ¼ K1;3

qP3 Cfr

ð9Þ

and partitioning between Phases 1 and 4 is: P K1;4 ¼

qP4 Cfr

ð10Þ

The ideal adsorbed solution theory (IAST) has been used to successfully describe sorption competition between similar solutes (McGinley et al., 1996; Li and Werth, 2001). For the CMS model, it is hypothesized that a modified version of the IAST may be useful to describe competition between DOM and PAHs. The IAST assumes that the equation below is true for a system consisting of two solutes, DOM and PAH, competing for sorption sites: qP4 qP;O 4

þ

qDOM S qDOM;O S

¼1

ð11Þ

where q4P,O and qSDOM,O denote sorbed concentration of PAH without DOM present, and sorbed concentration of DOM without PAH present, respectively. For the CMS model, let us define x = qSDOM/qSDOM,O and assume the presence of sorbed PAH negligibly impacts DOM sorption. Thus, x will be a number slightly less than 1, and we can rewrite Eq. (11) as follows: qP4 qP;O 4

¼1x¼e

ð12Þ

where e is a small number, except when there is no DOM present, and e = 1. If we further assume that e is inversely proportional to qSDOM, that is, the more sorbed DOM, the closer x approaches unity and the smaller e, we can express e as follows: 1 ec ð13Þ 1 þ bqDOM S where b is a parameter that quantifies the reduction of PAH sorption due to the presence of sorbed DOM. Using the definitions in Eqs. (12) and (13), we can rewrite Eq. (7) as: ! qP;O P DOM 4 þ ð14Þ mads M S ¼ q3 qS 1 þ bqDOM S

312

J.-W. Moon et al. / Journal of Contaminant Hydrology 60 (2003) 307–326

Since the distribution coefficient describing overall PAH partitioning between sorbed and aqueous phases is defined as K* ¼

MS =mads MW =VW

ð15Þ

we can use Eq. (14) and Eqs. (3), (6), and (8) –(10) to derive the distribution coefficient for overall PAH partitioning between sorbed and aqueous phases defined by the CMS model (K*CMS ) as: * KCMS ¼

P DOM DOM DOM DOM P K1;2 K2;3 Cw ð1 þ bK2;3 Cw Þ þ K1;4 P C DOM Þð1 þ bK DOM C DOM Þ ð1 þ K1;2 w w 2;3

ð16Þ

Note that the equation above assumes linear equilibrium sorption of both DOM and PAH and that if the presence of sorbed DOM does not affect PAH sorption, b = 0 and K*CMS (Eq. (16)) equals KOMS * (Eq. (1)). If we assume DOM sorption is described by a nonlinear Langmuir isotherm, the equation below can be used to describe the relation between DOM in the aqueous and sorbed phases (Jones and Tiller, 1999): qDOM ¼ S

DOM qmax Kads CW DOM 1 þ Kads CW

ð17Þ

where qmax (M M  1) is the maximum capacity of the sorbent to sorb DOM and Kads (l3 M  1) is an adsorption constant. For nonlinear sorption of DOM, the following equation can be obtained by replacing Eq. (3) with Eq. (17) when deriving K*CMS   qmax Kads CwDOM P P Kads CwDOM K1;2 qmax 1 þ b 1þK ð1 þ Kads CwDOM Þ þ K1;4 DOM C ads w * KCMS ¼ ð18Þ P C DOM Þð1 þ K C DOM þ bq DOM Þ ð1 þ K1;2 ads w max Kads Cw w

3. Materials and methods 3.1. Materials Both sand (425 – 825 Am, 0.068% organic carbon content, specific surface area by N2BET: 1.94 m2/g) and bentonite (from Donghae Chemical, CEC: 63.9 meq/100 g, surface area by N2-BET: 73.94 m2/g) were used as sample sorbents. Cetylpyridinium chloride (CPC) and hexadecyltrimethylammonium bromide (HDTMA) were purchased from Aldrich. Due to the negligible CEC of the sand, a preliminary experiment to determine surfactant sorption to the sand was performed. The experiment showed that CPC had a much greater affinity for the sand than HDTMA. Though both surfactants have identical 16-carbon hydrophobic tails, their hydrophilic groups are different. Whereas the hydrophilic group of HDTMA contains only a methyl group, that of CPC has a benzene ring, which perhaps results in a greater affinity of the CPC to sand

J.-W. Moon et al. / Journal of Contaminant Hydrology 60 (2003) 307–326

313

(Kibbey and Hayes, 1993). To prepare the CPC-sand sorbent, 0.25 l of CPC solution (2.79  10  4 M) was added to 5 g of sand, equilibrated for 48 h, and the treated sand was then air-dried for 72 h. To prepare the HDTMA-bentonite sorbent, aqueous HDTMA (0.016 M) was added in a mass equivalent to the available CEC of 20 g of sorbent, and then equilibrated for 24 h. The modified sorbents were subsequently washed with distilled water, oven-dried for 40 h at 80 jC, and then ground. The surfactant sorption varied from 0 to 1.75 mg/g for CPC on sand and from 0% to 100% of the CEC for sorption of HDTMA on bentonite. Total organic carbon content was measured for both sorbents, and XRD analysis was performed for the HDTMA-bentonite. Results of these analyses are presented in Table 1. Naphthalene (Aldrich, 99+%) and phenanthrene (Aldrich, 98+%) were used as model PAHs. In order to demonstrate the effect of DOM on PAH sorption, it was desirable from the standpoint of experimental quantification that sorption of the PAH be within 30 – 80% of the maximum sorption. Therefore, considering the different sorptive capacities of the modified sorbents, different PAHs were tested. Based on preliminary experimental results, phenanthrene was chosen as the model PAH for the CPC-sand sorption studies and naphthalene was chosen as the model PAH for the HDTMAbentonite sorption studies. For the purposes of these studies, it was assumed that the two PAHs qualitatively behaved similarly. The solubility of naphthalene and phenanthrene reported in the literature is 31 –34 and 1.6 mg/l at 25 jC, respectively (Edwards et al., 1991). Stock solutions were prepared by dissolving the PAHs in methanol and injecting the PAH/methanol solution into deionized water. The presence of minimal residual methanol in solution (less than 0.3%) is not expected to affect sorption (Schlautman and Morgan, 1993). Soil humic acid (SHA) and extracted dissolved organic matter (eDOM) were used as surrogates for dissolved organic matter (DOM). The SHA (Reference 1R102H) was obtained from the International Humic Substance Society (IHSS). The eDOM was extracted from the soil in Chang-Rang Brook (in Go-Yang City, South Korea) according to the procedures recommended by IHSS (Chiou et al., 2000). Stock solutions of DOMs were prepared by dissolving dry DOMs in distilled water with 0.1N NaOH.

Table 1 Sorbent characteristics Sorbent

Surfactant concentration

Organic carbon (%)

XRD d001-spacing ˚) (A

CPC-sand

0.00 mg/g 0.65 mg/g 1.09 mg/g 1.75 mg/g 0% of CEC 25% 50% 100%

0.068 0.110 0.131 0.179 1.89 5.48 8.97 15.08

ND ND ND ND 12.3 14.2 14.1 20.0

HDTMA-bentonite

ND: not determined.

314

J.-W. Moon et al. / Journal of Contaminant Hydrology 60 (2003) 307–326

3.2. Experimental procedures Sorption experiments were performed to obtain the following sorption equilibrium relationships: (1) sorption of DOM onto sorbent, (2) partitioning of PAH into DOM, (3) sorption of PAH onto sorbent with no DOM present, and (4) sorption of PAH onto sorbent in the presence of DOM. Half a gram of CPC-sand and 0.1 g of HDTMA-bentonite were put into separate Corex-II glass centrifuge tubes that were then filled with 25 ml of each DOM solution at concentrations ranging from 0 to 20 mg/l. The PAHs (naphthalene and phenanthrene) in methanol were then added to the tubes so that the overall PAH concentration in the tube was 80% of the aqueous solubility of the PAH. These same procedures were repeated, but without the addition of the PAH solution, to measure DOM sorption. The tubes were prepared at least in duplicate and closed with aluminum-foillined screw caps and agitated at 25 jC for 48 h. The suspension was centrifuged at 4302  g for 10 min and the DOM in the supernatant was quantified by measuring total organic carbon (TOC) with a TOC analyzer (Shimazu-5000A). PAH in the supernatant was analyzed using an HPLC (Waters 5890) with UV detector and a Waters 3.9  300 mm Abondapak C18 reverse phase column. The HPLC was operated under the following conditions: a flow rate of 1.8 ml/min, an injection volume of 15 Al, wavelength of 254 nm and a mobile phase of acetonitrile to water of 80 to 20, and with isocratic flow conditions. Retention time of naphthalene and phenanthrene were 2.5 and 3.5 min, respectively. Fluorescence quenching experiments (Hitachi F-3010) were also performed to measure the dissolved concentration of PAHs not associated with DOM. The excitation wavelength/ emission wavelength used was 290 nm/365 nm for the PAH probe. If we assume linear equilibrium partitioning of PAH between water and aqueous phase DOM (Eq. (8)), we can P determine K1,2 in Fig. 1 using the Stern – Volmer equation for linear quenching data (Schlautman and Morgan, 1993): FO P DOM ¼ 1 þ K1;2 CW F

ð19Þ

where FO denotes the fluorescence intensity of the sample in the absence of DOM and F represents the intensity in the presence of DOM.

4. Results and discussion 4.1. Effect of dissolved organic matter Fig. 2 is an isotherm for DOM sorption onto CPC-sand. The sorption isotherms exhibit Langmuir-type behavior, such as has been observed in prior studies of humic sorption on organics (Zhao and Vance, 1998). The Kads and qmax for DOM sorption onto CPC-sand was estimated by fitting Eq. (17) to the Fig. 2 data. Results are tabulated in Table 2. As can be seen from the table, both Kads and qmax increased with increasing concentrations of CPC onto sand. The sorption of DOM to inorganic sorbents has been explained by ligand exchange, electrostatic interactions, van der Waals forces, hydrophobic interactions, and

J.-W. Moon et al. / Journal of Contaminant Hydrology 60 (2003) 307–326

315

Fig. 2. Sorption of (a) SHA and (b) eDOM to 0.00, 0.65, 1.09, and 1.75 mg/g CPC-sand. Lines indicate Langmuir model fits to data.

hydrogen bonding (Sposito, 1984). As little sorption to untreated sand was observed, the dramatic increase of DOM sorption to CPC-modified sand can be explained as hydrophobic partitioning of the DOM into the hydrophobic tail of the CPC surfactant that is attached to the sand. DOM sorption to HDTMA-bentonite is shown in Fig. 3. As the DOM sorption isotherm appears linear, the K2,3 for DOM sorption onto HDTMA-bentonite was estimated by fitting Eq. (3) to the data (Table 2). Considering its large size (20 –500 ˚ ), DOM is assumed to preferentially sorb to external surfaces of the modified clay rather A

316

System

DOM

Surfactant concentration

P K1,2 (l/kg)

qmax (mg/g)

Kads (l/kg)

CPC-sand with phenanthrene

SHA

0.00 mg/g 0.65 1.09 1.75 0.00 mg/g 0.65 1.09 1.75 0% of CEC 25% 50% 100% 0% of CEC 25% 50% 100%

8.77E + 04 8.77E + 04 8.77E + 04 8.77E + 04 7.72E + 04 7.72E + 04 7.72E + 04 7.72E + 04 8.57E + 04 8.57E + 04 8.57E + 04 8.57E + 04 2.30E + 04 2.30E + 04 2.30E + 04 2.30E + 04

0.15 0.29 0.44 1.50 0.086 0.45 0.67 0.80

110 110 140 140 160 170 200 310

eDOM

HDTMA-bentonite + with naphthalene

SHA

eDOM

a

DOM K2,3 (l/kg)

P K1,4 (l/kg)

bqSDOMa

* K OMS (l/kg)

K *CMS (l/kg)

Kexp (l/kg)

41.6 123 240 968 29.4 51.3 151 3490

30.9 33.5 76.8 198 30.9 33.5 76.8 198 3.40 98.2 333 744 3.4 98.2 333 744

2.01 0.095 0.51 975.7 1981.6 766.4 290.4 3671.6 2700 3.65 0 0 2800 0 0 0

57.9 97.4 171.2 468.8 48.2 180.4 302.5 449.1 20 106 304 779 7.84 91.7 308 803

22.9 13.4 64.0 66.4 2.8 4.7 20.7 23.1 28.8 87.4 303 780 16.6 69.8 313 947

14.2 18 67.1 59.7 0.77 17.7 42.8 34.7 5.50 77.9 546 931 11.7 87.7 528 839

CPC-sand results determined by Eq. (18); HDTMA-bentonite results determined by Eq. (16).

J.-W. Moon et al. / Journal of Contaminant Hydrology 60 (2003) 307–326

Table 2 Summary of experimental and model results for K* and model parameter values

J.-W. Moon et al. / Journal of Contaminant Hydrology 60 (2003) 307–326

317

Fig. 3. Sorption of (a) SHA and (b) eDOM to 0%, 25%, 50%, and 100% of CEC HDTMA-bentonite. Lines indicate linear model fits to data.

than within the interlamellar region, which, as seen in Table 1, has d001 spacings of less ˚. than 20 A Figs. 4 and 5 illustrate how the overall soil –water distribution coefficients (K*) for PAH sorption onto CPC-modified sand and HDTMA-modified bentonite, respectively, are affected by the presence and absence of DOM. After constructing PAH sorption isotherms (plotting sorbed versus aqueous concentrations of PAH), lines passing through the origin were fit through the data points and the slopes of the lines, which represent the distribution

318

J.-W. Moon et al. / Journal of Contaminant Hydrology 60 (2003) 307–326

Fig. 4. Distribution coefficient (K*) for phenanthrene sorption on CPC-sand with increasing concentration of (a) SHA (b) eDOM.

coefficients (K*), were determined. Fig. 4 shows how K*, the distribution coefficient for overall phenanthrene partitioning between the sorbed and aqueous phases, varies with increasing concentrations of DOM for phenanthrene sorption onto CPC-sand. The figure shows that K* decreases with increasing DOM concentrations. That is, phenanthrene strongly partitions into the DOM, and either (1) the DOM partitions less into the CPC-sand than phenanthrene does, thereby reducing the overall distribution coefficient of phenanthrene, K*, or (2) the DOM competes with phenanthrene for sorption sites, thereby reducing K*. In support of the second alternative, Li and Werth (2001) noted that when a

J.-W. Moon et al. / Journal of Contaminant Hydrology 60 (2003) 307–326

319

Fig. 5. Distribution coefficient (K*) of naphthalene sorption on HDTMA-bentonite with increasing concentration of (a) soil humic acid and (b) eDOM.

sorbate mixture is present, sorption sites with the highest adsorption potential are preferentially occupied by the more favorable sorbate, resulting in fewer available sites for the less favored sorbate. In support of the first alternative, PAHs have been shown to associate with DOM dispersed in the solution phase, and thus effectively enhance the solubility of the PAH and reduce sorption from solution (Chin et al., 1990). The degree of association is a function of the hydrophobicity of the PAH, and of the concentration and the reactivity of DOM.

320

J.-W. Moon et al. / Journal of Contaminant Hydrology 60 (2003) 307–326

Fig. 5 shows how K* varies with increasing concentrations of DOM for naphthalene sorption onto HDTMA-bentonite. The figure indicates that for HDTMA-modified bentonite, K* is relatively unaffected by increasing DOM concentration, while for unmodified bentonite, K* decreases with increasing DOM concentration. This finding is similar to results reported by Zhao and Vance (1998). These observations may be due to the fact that the HDTMA-modified bentonite has an excess of sorption sites. Thus, addition of DOM neither increases nor decreases naphthalene sorption since DOM and naphthalene partition into the HDTMA-bentonite to approximately the same extent. On the other hand, like the

Fig. 6. Experimentally determined versus model simulated distribution coefficients (K*) for phenanthrene sorption on CPC-sand in the presence of 10 mg/l (a) SHA (b) eDOM.

J.-W. Moon et al. / Journal of Contaminant Hydrology 60 (2003) 307–326

321

sand, unmodified bentonite has limited sorption sites, so K* decreases with increasing DOM due to either competition for sorption sites, or the fact that DOM sorbs less than naphthalene onto the unmodified bentonite. 4.2. Modeling results DOM P P As described above, values for K2,3 , K1,2 , and K1,4 were experimentally deter* . The mined. These values were then used in Eq. (1), the OMS model, to calculate KOMS OMS model accounts for the fact that DOM sorption to a sorbent may be different than

Fig. 7. Experimentally determined versus model-simulated distribution coefficients (K*) for naphthalene sorption on HDTMA-bentonite in the presence of 10 mg/l (a) SHA (b) eDOM.

322

J.-W. Moon et al. / Journal of Contaminant Hydrology 60 (2003) 307–326

PAH sorption. However, the model does not account for possible competition for sorption sites between PAH and DOM. Thus, we also used the CMS model (Eqs. (16) and (18)) to calculate K*CMS . Since a linear sorption isotherm described DOM sorption onto HDTMAbentonite and a Langmuir isotherm described DOM sorption onto CPC-sand (Table 2), Eqs. (16) and (18) were used to fit data from the HDTMA-bentonite experiments and CPC-sand experiments, respectively. The parameter b in Eqs. (16) and (18) was used as a * fitting parameter to minimize the sum of squares of the differences between the KCMS values predicted by the models and the experimental values determined for K*.

Fig. 8. Experimentally determined versus model-simulated distribution coefficients (K*) for phenanthrene sorption on 1.75 mg/g CPC-sand for increasing concentrations of (a) SHA (b) eDOM.

J.-W. Moon et al. / Journal of Contaminant Hydrology 60 (2003) 307–326

323

Model calculations were then compared with the values of K* measured experimentally. Results are shown in Table 2 and Figs. 6 and 7. Table 2 shows that PAH strongly P partitions into DOM (that is, K1,2 >1), as was speculated earlier. These large values of P K1,2 may partially be attributed to the analytical technique used in this research, as the application of fluorescence quenching to measure PAH-DOM binding constants has been reported to result in the overestimation of the partition coefficients (Laor et al., 1998). The OMS model consistently overestimated sorption for phenanthrene onto CPC-sand. This overestimation by the model may be attributed to the fact that the model neglects competitive sorption between PAH and DOM. It appears, as discussed earlier, that sorption

Fig. 9. Experimentally determined versus model-simulated distribution coefficients (K*) for naphthalene sorption on 50% of CEC HDTMA-clay for increasing concentrations of (a) SHA (b) eDOM.

324

J.-W. Moon et al. / Journal of Contaminant Hydrology 60 (2003) 307–326

sites onto CPC-sand are limited, so PAH sorption may be inhibited by the presence of DOM. As seen in Fig. 8, the CMS model, which accounts for this competition, does a much better job simulating the experimental values of K*. Table 2 shows that for the CPCsand sorbent, values of bqSDOM (see Eqs. 14 and 15) are either much greater than or comparable to 1. Table 2 and Fig. 7 show that for naphthalene sorption onto HDTMAbentonite, both models do a reasonably good job of matching the experimental data. In this case, the OMS adequately describes the effect of the presence of DOM on PAH sorption, and we may infer that for the HDTMA-bentonite sorbent, competition for sorption sites between the DOM and PAH is not significant. Table 2 shows that bqSDOM is typically 0 for the HDTMA-bentonite sorbent, and even in those few instances where it is not 0, the model fit to the data is not significantly improved over the fit of the OMS model which neglects competition. The reason the CMS model helps explain the results found for the CPC-sand sorbent but not for the HDTMA-bentonite sorbent may be seen from Figs. 8 and 9. Fig. 8 shows how the OMS model, which neglects sorption competition, fails to describe the decrease in phenanthrene sorption to CPC-sand with increasing DOM concentrations, while the CMS model, which incorporates competition, describes the data. Conversely, Fig. 9 shows how the OMS model describes naphthalene sorption onto HDTMA-bentonite in the presence of increasing concentrations of DOM, and how consideration of competition by the CMS model fails to improve the model fit to the data.

5. Conclusions and future work Another modification that can be made to the OMS model, which perhaps can be used to explain the CPC-sand data, is to assume that the distribution coefficients describing P P PAH sorption to aqueous and sorbed DOM are different (that is, K1,2 p K1,3 ). Since DOM is a mixture of complex macromolecules, it is possible that fractionation may occur and the larger molecular weight DOM particles may preferentially be sorbed, resulting in the DOM attached to the surface not being the same as the original bulk DOM, and therefore, also different from the DOM remaining in solution. This fractionation may affect the PAH binding capacity of the DOM (Jones and Tiller, 1999). The differences in the physical and chemical nature of the aqueous DOM compared with sorbed DOM might affect partitioning of the PAH. Rav-Acha and Rebhun (1992) and Rebhun et al. (1992) reported an organic carbon partition coefficient value that was 10 times higher for fluoranthene partitioning into dissolved humic material than into mineral-associated humic material. They suggested that dissolved humic substances are more accessible for incorporation of hydrophobic compounds than humic material that is associated with a solid. The current work did not consider the above assumption, however, since it was not experimentally possible to distinguish between the distribution coefficients of PAH to sorbed versus aqueous DOM. Further work is needed to characterize and quantify the differences between sorbed and aqueous DOM, and incorporate those differences into a revised OMS model. Another area deserving of further study is to determine how the sources and origins of DOM may influence the effect of DOM on PAH sorption to a contaminant transport

J.-W. Moon et al. / Journal of Contaminant Hydrology 60 (2003) 307–326

325

barrier. Previous studies showed that natural organic matter originating from different sources (e.g. soil, sediment, and aquatic) exhibit different relationships between their chemical composition and PAH sorption (Chefetz et al., 2000), though in the current work, there was little difference seen between how SHA and eDOM influenced PAH sorption in the systems that were studied. In this study, it was found that distribution coefficients for sorption of PAH onto HDTMA-modified bentonite were higher than the coefficients quantifying PAH sorption onto CPC-modified sand, and that the distribution coefficients for the HDTMA-bentonite sorbent were relatively unaffected by the presence of DOM, whereas, the distribution coefficients for the CPC-sand sorbent decreased with increasing DOM concentration. Consequently, HDTMA-bentonite, from the aspect of sorptive capacity, appears to be more appropriate for use as a barrier to PAH transport by groundwater, especially in the presence of DOM. This work also demonstrated that the OMS model can be used to describe sorption of PAH in the presence of DOM to a high capacity sorbent like HDTMA-bentonite, while sorption to a less sorptive material like CPC-sand is better described by a model that accounts for sorption competition between PAH and DOM. Acknowledgements This research was supported by both the National Research Laboratory program and the Green 21 program of the Korean Ministry of Science and Technology. Mark N. Goltz was on a sabbatical to the NSERL, supported by the Air Force Office of Scientific Research and the Air Force Institute of Technology. References Boyd, S.A., Sun, S., Lee, J.F., Mortland, M.M., 1988. Pentachlorophenol sorption by organo-clays. Clays Clay Miner. 36 (2), 125 – 130. Chefetz, B., Deshmukh, A.P., Hatcher, P.G., Guthrie, E.A., 2000. Pyrene sorption by natural organic matter. Environ. Sci. Technol. 34, 2925 – 2930. Chin, Y.P., Weber, W.J., Eadie, B.J., 1990. Estimating effects of dispersed organic polymers on the sorption of contaminants by natural solids: 2—sorption in the presence of humic and other natural macromolecules. Environ. Sci. Technol. 24, 837 – 842. Chiou, C.T., Kile, D.E., Brinton, T.I., Malcolm, R.L., Leenheer, J.A., MacCarthy, P., 1987. A comparison of water solubility enhancements of organic solutes by aquatic humic materials and commercial humic acids. Environ. Sci. Technol. 21, 1231 – 1234. Chiou, C.T., Kile, D.E., Rutherford, D.W., Sheng, G., Boyd, S.A., 2000. Sorption of selected organic compounds from water to a peat soil and its humic-acid and humin fractions: potential sources of the sorption nonlinearity. Environ. Sci. Technol. 34, 1254 – 1258. Edwards, D.A., Luthy, R.G., Liu, Z., 1991. Solubilization of polycyclic aromatic hydrocarbons in micellar nonionic surfactant solutions. Environ. Sci. Technol. 25, 127 – 133. Finkel, M., Liedl, R., Teutsch, G., 1999. Modeling surfactant-enhanced remediation of polycyclic aromatic hydrocarbons. Environ. Model. Softw. 14, 203 – 211. Guha, S., Jaffe, P.R., Peters, C.A., 1998. Solubilization of PAH mixtures by a nonionic surfactant. Environ. Sci. Technol. 32, 930 – 935. Jaynes, W.F., Vance, G.F., 1996. BTEX sorption by organo-clays: cosorptive enhancement and equivalence of interlayer complexes. Soil Sci. Soc. Am. J. 60, 1742 – 1749.

326

J.-W. Moon et al. / Journal of Contaminant Hydrology 60 (2003) 307–326

Johnson, W.P., Amy, G., 1995. Facilitated transport and enhanced desorption of polycyclic aromatic hydrocarbons by natural organic matter in aquifer sediments. Environ. Sci. Technol. 29, 807 – 817. Jones, K.D., Tiller, C.L., 1999. Effect of solution chemistry on the extent of binding of phenanthrene by a soil humic acid: a comparison of dissolved and clay bound humic. Environ. Sci. Technol. 33, 580 – 587. Karickhoff, S.W., Brown, D.S., Scott, T.A., 1979. Sorption of hydrophobic pollutants on natural sediments. Water Res. 13, 241 – 248. Kibbey, T.C.G., Hayes, K.F., 1993. Partitioning and UV absorption studies of phenanthrene on cationic surfactantcoated silica. Environ. Sci. Technol. 27, 2168 – 2173. Kile, D.E., Chiou, C.T., 1989. Water solubility enhancements of DDT and trichlorobenzene by some surfactants below and above the critical micelle concentration. Environ. Sci. Technol. 23, 832 – 838. Ko, S., Schlautman, M.A., Carraway, E.R., 1998. Partitioning of hydrophobic organic compounds to sorbed surfactants: 1. Experimental studies. Environ. Sci. Technol. 32, 2769 – 2775. Laor, Y., Farmer, W.J., Aochi, Y., Strom, P.F., 1998. Phenanthrene binding and sorption to dissolved and to mineral-associated humic acid. Water Res. 32, 1923 – 1931. Lee, C.-L., Kuo, L.-J., 1999. Quantification of the dissolved organic matter effect on the sorption of hydrophobic organic pollutant: application of an overall mechanistic sorption model. Chemosphere 38 (4), 807 – 821. Lee, J.F., Crum, J.R., Boyd, S.A., 1989. Enhanced retention of organic contaminants by soils exchanged with organic cations. Environ. Sci. Technol. 23, 1365 – 1372. Lee, J.F., Mortland, M.M., Chiou, C.T., Kile, D.E., Boyd, S.A., 1990. Adsorption of benzene, toluene, and xylene by two tetramethylammonium-smectites having different charge densities. Clays Clay Miner. 38 (2), 113 – 120. Lee, C.-L., Huang, H.-T., Kuo, L.-J., 2000. Experimental validation of an OMS model for the sorption behaviors of PAHs onto aluminum oxide coated with humic acids. J. Environ. Sci. Health A35 (4), 515 – 536. Li, J., Werth, C.J., 2001. Evaluating competitive sorption mechanisms of volatile organic compounds in soils and sediments using polymers and zeolites. Environ. Sci. Technol. 35, 568 – 574. Liu, H., Amy, G., 1993. Modeling partitioning and transport interactions between natural organic matter and polynuclear aromatic hydrocarbons in groundwater. Environ. Sci. Technol. 27, 1553 – 1562. Lu¨hrmann, L., Noseck, U., 1998. Model of contaminant transport in porous media in the presence of colloids applied to actinide migration in column experiments. Water Resour. Res. 34 (3), 421 – 426. Magee, B.R., Lion, L.W., Lemley, A.T., 1991. Transport of dissolved organic macromolecules and their effect on the transport of phenanthrene in porous media. Environ. Sci. Technol. 25, 323 – 331. McCarthy, J.F., Jimenez, B.D., 1985. Interactions between polycyclic aromatic hydrocarbons and dissolved humic materials: binding and dissociation. Environ. Sci. Technol. 19, 1072 – 1076. McCarthy, J.F., Zachara, J.M., 1989. Subsurface transport of contaminants. Environ. Sci. Technol. 23, 496 – 502. McGinley, P.M., Katz, L.E., Weber Jr., W.J., 1996. Competitive sorption and displacement of hydrophobic organic contaminants in saturated subsurface soil systems. Water Resour. Res. 32 (12), 3571 – 3577. Rav-Acha, Ch., Rebhun, M., 1992. Binding of organic solutes to dissolved organic substances and its effect on adsorption and transport in the aquatic environment. Water Res. 26, 1645 – 1654. Rebhun, M., Kalabo, R., Grossman, L., Manka, J., Rav-Acha, Ch., 1992. Sorption of organics on clay and synthetic humic-clay complexes simulating aquifer processes. Water Res. 26, 79 – 84. Rebhun, M., Smedt, F.D., Rwetabula, J., 1996. Dissolved humic substances for remediation of sites contaminated by organic pollutants. Water Res. 30, 2027 – 2038. Schlautman, M.A., Morgan, J.J., 1993. Binding of a fluorescent hydrophobic organic probe by dissolved humic substances and organically-coated aluminum oxide surfaces. Environ. Sci. Technol. 27, 2523 – 2532. Sposito, G., 1984. The Surface Chemistry of Soils, vol. 256. Oxford Univ. Press, New York. Zhao, H., Vance, G.F., 1998. Sorption of trichloroethylene by organo-clays in the presence of humic substances. Water Res. 32, 3710 – 3716. Zhao, H., Jaynes, W.F., Vance, G.F., 1996. Sorption of the ionizable organic compound, dicamba, (3, 6-dichloro2-methoxy benzoic acid) by organo-clays. Chemosphere 33 (11), 2089 – 2100. Zimmer, G., Brauch, H.J., Sontheimer, H., 1989. Activated-carbon adsorption of organic pollutants. In: Suffet, I.H., MacCarthy, P. (Eds.), Aquatic Humic Substances: Influence on Fate and Treatment of Pollutants. Advances in Chemistry Series, vol. 219. American Chemical Society, Washington, DC, pp. 579-596.