Particle-Size Dependent Sorption and Desorption of Pesticides within

Switch Switch View Sections. All; My List ... My Content. Favorites; Downloaded. Settings. My Account · Help; Full Site. Hide Menu Back. Please wait w...
0 downloads 0 Views 492KB Size
Environ. Sci. Technol. 2008, 42, 3381–3387

Particle-Size Dependent Sorption and Desorption of Pesticides within a Water-Soil-Nonionic Surfactant System PENG WANG AND ARTURO A. KELLER* Bren School of Environmental Science and Management, University of California, Santa Barbara, 93106

Received October 30, 2007. Revised manuscript received February 17, 2008. Accepted February 20, 2008.

Although nonionic surfactants have been considered in surfactant-aided soil washing systems, there is little information on the particle-size dependence of these processes, and this may have significant implications for the design of these systems. In this study, Triton-100 (TX) was selected to study its effect on the sorption and desorption of two pesticides (Atrazine and Diuron) from different primary soil size fractions (clay, silt, and sand fractions) under equilibrium sorption and sequential desorption. Soil properties, TX sorption, and pesticide sorption and desorption all exhibited significant particle-size dependence. The cation exchange capacity (CEC) of the bulk soils and the soil fractions determined TX sorption capacity, which in turn determined the desorption efficiency. Desorption of pesticide out of the clay fraction is the limiting factor in a surfactantaided washing system. The solubilization efficiency of the individual surfactant micelles decreased as the amount of surfactant added to the systems increased. Thus, instead of attempting to wash the bulk soil, a better strategy might be to either (1) use only the amount of surfactant that is sufficient to clean the coarse fraction, then separate the fine fraction, and dispose or treat it separately; or (2) to separate the coarse fractions mechanically and then treat the coarse and fine fractions separately. These results may be applicable to many other hydrophobic organic compounds such as polyaromatic hydrocarbons (PAHs) and polychlorinated biphenyls (PCBs) strongly sorbed onto soils and sediments.

1. Introduction Pesticide spills and accidents involving pesticide handling take place each year on farms and pesticide formulating and manufacturing plants resulting in high pesticide concentration at these sites (1). There are many other situations where soils or sediments are contaminated with hydrophobic organic compounds (HOCs) that are strongly sorbed and not bioavailable for natural or enhanced biodegradation. The ability of surfactants to enhance the water solubility of HOCs provides a potential means of remediating pesticidecontaminated soils and sediments by surfactant-aided soil washing (2, 3). Loss of anionic surfactants (e.g., linear alkylbenzene sulfonate (LAS) and sodium dodecyl sulfonate (SDS)) by complexation with divalent cations in soils (e.g., Ca2+, Mg2+) can be so significant that the use of anionic surfactants for * Corresponding author phone: 01-805-453-1822; fax: 01-805-4563807; E-mail: [email protected]. 10.1021/es702732g CCC: $40.75

Published on Web 04/02/2008

 2008 American Chemical Society

remediating contaminated soils rich in divalent cations is typically ineffective (4, 5). Cationic surfactants, due to their positive charge, tend to significantly sorb onto the soil particles via cation exchange (6, 7), resulting in significant surfactant loss. Thus, the use of nonionic surfactants for soil washing has received considerable research attention (e.g., refs 1, 2, 8–10,). However, nonionic surfactants can also sorb onto soil to some extent (2, 8–10). Surfactant loss via sorption onto soils results in reduced solubilization of HOCs and, more importantly, the sorbed surfactants can serve as an effective partitioning medium for the HOCs (3, 6, 7, 11) which further complicates the sorption and desorption behavior of the HOC within these systems. Thus, understanding surfactant sorption is critical in understanding the mechanism by which HOCs can be removed within surfactant-aided soil washing systems. Most soil washing processes remove contaminants from soils either by dissolving them in a washing solution and/or concentrating them into a smaller amount of soil through particle size separation (12, 13). By separating the fine from the coarser particles, the soil washing process effectively concentrates the contaminants into a smaller amount of soil that can be further treated or disposed of. Therefore, understanding soil particle-size dependent surfactant sorption behavior and its effect on HOC partitioning among the size fractions is highly relevant to soil washing. Although HOC sorption onto different soil size fractions has received considerable attention in the literature (14) there has been little work on understanding the effect of adding surfactants to these pesticide-soil systems (3, 15, 16). To our knowledge, the sorption and desorption of HOC onto and from primary soil size fractions (i.e., clay, silt, and sand size fractions) in the presence of surfactants has not been reported elsewhere. Equilibrium HOC sorption experiments within soil–water–surfactant systems have been conducted by numerous researchers (e.g., refs 9, 10, and 17,), but there have been relatively few desorption experiments reported (1, 8, 18). Although the results of equilibrium sorption experiments can shed valuable insights into the desorption behavior of HOC within surfactant-aided soil washing systems, there are still gaps in understanding the parameters that control HOC desorption from the different size fractions. Also, it is uncertain whether the results of equilibrium sorption experiments can be used for predicting HOC desorption in a soil washing system. The nonionic surfactant Triton-100 (TX) was considered in this study since it has been intensively studied within surfactant-aided soil washing systems (e.g., refs 1, 3, 8, 10, 16–19,) in the past few years. This project focused on exploring the effect of TX on the sorption and desorption of two pesticides (Atrazine and Diuron), as an example of a more general application for removing HOCs from contaminated soils by ex situ surfactant-aided soil washing. More specifically, the objectives of the study were (1) to investigate equilibrium soil particle-size dependent surfactant sorption and pesticide partitioning behavior and (2) to examine particle-size dependent pesticide desorption behavior under sequential desorption.

2. Materials and Methods 2.1. Chemicals. Atrazine (2-chloro-4-ethylamino-6-isopropylamino-1,3,5-triazine) was purchased from Supelco Inc. (Bellefonte, PA) with a reported purity >97%, and Diuron (3-(3,4-dichlorofenyl)-1,1-dimethylurea) was purchased from ChemService Inc. (West Chestnut, PA) with a reported purity VOL. 42, NO. 9, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3381

TABLE 1. Measured Soil Properties soil Ag #1 (sandy loam) bulk clay silt sand Ag #2 (loam) bulk clay silt sand Ag #3 (loam) bulk clay silt sand clayey (sandy clay loam) bulk clay silt sand sediment (loamy sand) bulk clay silt sand

CEC BET surface weight % OC (%) (cmol/kg) area (m2/g) pH

100% 5.3% 17.1% 77.6%

1.51 4.95 1.82 0.50

6.20 40.2 13.0 3.0

3.5 30.8 4.5 1.3

7.3 6.7 8.5 8.3

100% 6.8% 47.6% 45.6%

1.50 4.36 1.29 0.15

15.2 59.0 16.0 6.0

9.4 53.1 8.4 1.4

7.4 7.2 8.4 8.3

100% 13.8% 41.0% 45.2%

1.52 4.50 1.13 0.11

15.4 54.4 19.0 3.0

14.1 63.1 8.0 2.1

7.6 7.9 8.2 8.2

100% 14.0% 28.4% 57.6%

1.37 1.80 0.66 0.50

15.7 50.4 18.0 8.0

13.1 71.0 7.0 1.0

7.1 7.5 6.9 8.5

100% 2.7% 21.1% 76.2%

1.12 6.02 1.28 0.27

5.4 42.2 13.0 3.0

2.0 36.6 4.5 0.9

8.1 6.4 8.6 8.5

>99%. Triton-100 (t-octylphenoxypolyethoxyethanol) was purchased from Sigma-Aldrich (St. Louis, MO). The chemicals were used as received. Selected physicochemical properties of these compounds can be found in the Supporting Information. 2.2. Bulk Soils and Soil Particle Size Fractions. Four soils and one sediment (denoted as Ag#1, Ag#2, Ag#3, Clayey, and Sediment) were collected from Santa Barbara, California. The water dispersible clay (50 µm) fractions were separated using a low energy method, which involved using only water as dispersant, gentle mixing, repeated wet sedimentation, dialysis desalination, and freezedrying. The details of the particle size separation can be found in the Supporting Information. The organic carbon (OC), cation exchange capacity (CEC), and pH of the soils were measured using the standard methods described by Carter (20). BET surface area (SA) measurement was conducted on a TriStar 3000 gas adsorption analyzer (Micromeritics Inc., Norcross, GA) using N2. 2.3. Equilibrium Sorption of Pesticide and TX. The sorption experiments were conducted in duplicate by the batch equilibration technique. The TX sorption isotherm was determined in the absence of the pesticides. Deionized (DI) water was used for preparing all aqueous solutions. Pesticide partitioning was determined in the absence and presence of TX. Pesticide concentrations used were 15.00 mg/L for Atrazine and 15.95 mg/L for Diuron. Initial TX concentration spanned over a large range (0∼20 g/L), below and above its critical micelle concentration (CMC ) 0.12 g/L). A 0.01 M CaCl2 background electrolyte was used to minimize change in ionic strength, and 0.02% NaN3 was used as microbial growth inhibitor in all cases. Aliquots of 2.00 g of a bulk soil, 0.30 g clay fraction, 0.70 g silt fraction, or 1.00 g sand fraction were treated with 10 mL of solution containing pesticide and surfactant with varying concentrations in 15 mL glass centrifuge tubes. The amount used for each particle size fraction was based on its average weight fraction within the bulk soil. 3382

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 9, 2008

FIGURE 1. TX sorption isotherms onto (a) clay fractions and (b) silt fractions. Note: The x axis presents the equilibrium TX aqueous concentration. The tubes with soil and the pesticide and/or surfactant solution were then rotated at 60 rpm for 24 h to reach sorption equilibrium in an end-over-end shaker at 22 ( 2 °C, and then centrifuged at 5000g for 30 min at the same temperature. A number of researchers have used 24 h as a mixing period for studying the partitioning of HOCs within soil–water–surfactant systems. For example, Li and Bowman (6), Hayworth and Burris (7), Sanchez-Camazano et al. (8), Sheng et al. (11), Edwards et al. (17), Rodriguez-Cruz et al. (18), Ko et al. (21), Zhu et al. (22), Park and Jaffe (23), Sun et al. (24), Kibbey and Hayes (25), and Abu-Zreig et al. (27), all found that the sorption equilibrium of the HOC and surfactant (cationic, anionic, or nonionic) within soil–water–surfactant systems was reached within a 24 h mixing period. Additional results from this study are provided in the Supporting Information, and they indicate that 24 h are more than sufficient to reach equilibrium sorption for both pesticide and surfactant. Pesticide and surfactant concentrations in the supernatant were analyzed using high performance liquid chromatography (HPLC). The sorption of pesticide and surfactant on the centrifuge tubes was determined to be negligible, and the amount of the pesticide and surfactant blank (with no soils) did not show any significant change before and after mixing. Thus, the amount of pesticide and surfactant sorbed was calculated as the difference between the initial and final mass in the aqueous phase. 2.4. Pesticide Desorption in the Presence of TX. Pesticide desorption experiments were conducted using the same experimental procedure as for the sorption experiments except that the desorption experiments consisted of one

TABLE 2. TX sorption Capacities in Different Soils and Soil Fractions, Amount of Pesticide Presorbed before the Desorption Experiments, And Break-Even Concentrationsa soils

Diuron

Atrazine

TX sorption capacity (mg/kg)

presorbed (mg/kg)

break-even conc. (g/L)

presorbed (mg/kg)

break-even conc. (g/L)

bulk

Ag#1 Ag#2 Ag#3 clayey sediment

4,100 ( 340 28,000 ( 940 28,000 ( 1640 31,000 ( 2800 3,700 ( 310

27.4 ( 1.2 28.5 ( 1.9 33.7 ( 1.2 24.6 ( 1.0 18.3 ( 0.9

3.1 ( 0.4 9.3 ( 0.5 9.0 ( 0.7 9.3 ( 0.5 3.0 ( 0.4

12.8 ( 0.9 14.0 ( 0.8 15.0 ( 0.9 13.8 ( 0.7 10.9 ( 0.5

5.2 ( 0.5 10.7 ( 0.6 11.0 ( 0.8 10.5 ( 0.5 5.9 ( 0.4

clay

Ag#1 Ag#2 Ag#3 clayey sediment

43,000 ( 3200 147,000 ( 7500 116,000 ( 9200 148,000 ( 6700 58,000 ( 4200

145.6 ( 3.2 144.0 ( 5.0 164.1 ( 2.1 117.9 ( 7.3 194.3 ( 3.2

6.6 ( 0.5 10.9 ( 0.6 11.5 ( 0.7 10.5 ( 0.8 6.4 ( 0.4

55.3 ( 2.2 52.9 ( 2.2 49.3 ( 0.8 38.6 ( 1.9 52.9 ( 1.5

8.9 ( 0.6 16.4 ( 0.8 18.5 ( 0.7 16.8 ( 0.7 9.3 ( 0.3

silt

Ag#1 Ag#2 Ag#3 clayey sediment

9,200 ( 890 18,000 ( 1200 34,000 ( 3200 29,000 ( 2300 9,900 ( 980

92.2 ( 5.2 55.3 ( 4.2 56.5 ( 3.5 47.5 ( 2.9 68.6 ( 3.8

1.8 ( 0.2 2.8 ( 0.1 5.2 ( 0.5 4.5 ( 0.3 2.3 ( 0.2

25.2 ( 1.1 15.8 ( 0.5 15.1 ( 0.4 16.2 ( 1.0 17.4 ( 0.9

2.5 ( 0.2 4.5 ( 0.3 8.9 ( 0.5 7.5 ( 0.5 2.9 ( 0.3

sand

Ag#1 Ag#2 Ag#3 clayey sediment

2,200 ( 190 7,800 ( 620 4,000 ( 380 8,600 ( 790 2,400 ( 220

10.5 ( 0.3 13.4 ( 0.6 8.5 ( 0.5 12.3 ( 1.0 12.9 ( 0.9

0.9 ( 0.1 2.0 ( 0.1 2.4 ( 0.3 1.9 ( 0.2 1.0 ( 0.1

6.1 ( 0.3 9.2 ( 0.6 5.2 ( 0.5 7.2 ( 0.6 7.5 ( 0.7

1.1 ( 0.1 2.8 ( 0.2 3.3 ( 0.2 2.8 ( 0.2 1.2 ( 0.1

a

Note: data were presented as average ( standard error of the duplicate measurements.

sorption step followed by five consecutive desorption steps. Only the bulk soils and their clay and silt size fractions were used for the desorption experiments. During the sorption cycle, aliquots of 2.00 g of a bulk soil, 0.30 g clay fraction, or 0.70 g silt fraction were treated with 10 mL of a solution containing only pesticide in 15 mL glass centrifuge tubes. The concentrations were 15.00 mg/L for Atrazine and 15.95 mg/L for Diuron in all cases. During each desorption step, 5 mL of the supernatant were replaced with 5 mL of surfactant solution. The concentration of these 5 mL of surfactant solution was adjusted so that the 10 mL of supernatant would be at the desired surfactant concentration for each case. Three TX concentrations were employed: 1.00, 2.00, and 3.00 g/L. The blank experiments were conducted with water only containing the corresponding background electrolytes. No significant change in pH was observed before and after the sorption/desorption experiments. 2.5. HPLC Analysis. A Shimadzu HPLC system (Shimadzu, Nakagyo-ku, Kyoto, Japan) was equipped with two LC-10AT VP pumps, a Sil-10AF autosampler, a DGU-14A degasser, and a SPD-M10AVP diode array detector. A Shimadzu Premier C18 5 µ reverse phase column was used with a length of 250 mm and an inner diameter of 4.6 mm. The analyses were performed at a constant flow rate of 1.0 mL/min. The solvent concentration gradients were designed so that all the homologues of the same surfactant eluted under the same peak. The UV detector monitored the absorbance at 222 nm for Atrazine, 247 nm for Diuron, and 225 nm for TX. Some samples were diluted as needed. The calibration was conducted daily and R2 was greater than 0.98 in all cases. The retention time of Atrazine, Diuron, and TX was 5.3, 5.6, and 8.0 min, respectively.

3. Results and Discussion 3.1. Characterization of the Bulk Soils and Their Size Fractions. The measured properties of the bulk soils and their size fractions are presented in Table 1. Generally speaking, the soil properties were highly particle-size dependent. As the size of the fractions decreased, their OC, SA, and CEC increased. The clay fractions had consistently higher

OC, SA, and CEC than the bulk soils and the silt and sand fractions. The correlation coefficient between OC and CEC was 0.83 and that between OC and SA was 0.72 (c.f. Supporting Information). 3.2. TX Sorption Isotherms. For all soils and soil fractions, TX sorption first increased with increasing aqueous surfactant concentrations until a saturation sorption capacity was reached, which is consistent with other researchers (1, 3, 16). Figure 1 presents the TX sorption isotherms onto the clay and silt fractions as examples. The average of duplicate measurements was used in preparing these graphs. The standard errors were smaller than 10% of the averages; error bars are all less than the symbol size and thus are not presented. The TX saturation sorption capacities, defined here as the average of the plateau points in the TX sorption isotherms, can be found in Table 2. The analysis indicated that the sorption capacity for TX in different soil fractions was highly related to CEC with a correlation coefficient of 0.90, whereas the correlation coefficient between OC and TX sorption capacities was low (0.43), indicating that soil organic matter (SOM) was not the dominant phase for TX sorption, which is consistent with other studies focused on bulk soil behavior (1, 27). However, the strong correlation between TX sorption capacity and CEC has not been reported before. Unlike cationic surfactant sorption, which takes place via cation exchange reaction (6, 7), the sorption of nonionic surfactant is less clear. Edward et al. (17) proposed a three-stage sorption model for TX sorption onto sediment, which states that at high bulk solution surfactant concentrations, a patchy bilayer of the sorbed TX is formed, with the base layer consisting of the “head-on” sorbed TX held on the hydrophilic (i.e., charged) patches of the sediment surface. The high correlation between TX sorption capacities and CEC in this study seems to support this hypothesis. Since CEC and surface area (SA) are strongly correlated, the sorption capacity of TX is also highly related to SA, with a correlation coefficient of 0.93. The correlation coefficient between the TX sorption capacity for bulk soils and their clay content was 0.83. VOL. 42, NO. 9, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3383

FIGURE 2. Diuron sorption in the presence of TX with (a) clay fractions; (b) silt fractions; (c) sand fractions; (d) all size fractions of Ag#1 and Ag#3. Note: (1) The error bars shown for soils Ag #1 and Ag #3 are representative of the experimental error for all soils considered; (2) x and y axes are equilibrium TX and Diuron concentrations respectively. 3.3. EquilibriumPesticideSorptionwithinSoil–Water-TX Systems. Figure 2presents the relationship between equilibrium aqueous Diuron concentrations and aqueous TX concentrations for the various soil fractions (Figure 2a-c) of two bulk soils (Ag#1 and Ag#3 in Figure 2d). In all cases, the standard errors were within 15% of the averages of the duplicates, so the error bars are presented only for Ag#1 and Ag#3 for clarity. At first the aqueous pesticide concentration (i.e., the sum of water-dissolved concentration plus the concentration present in the micelles) decreased with increasing aqueous surfactant concentration before the CMC was reached, due to sorption of surfactant onto the soil particles and subsequent partitioning of the pesticide into the sorbed surfactant. After the CMC was reached, the aqueous pesticide concentration increased with increasing aqueous surfactant concentration due to the presence of surfactant micelles. Atrazine showed similar behavior to Diuron, although less pronounced due to lower partitioning into both sorbed TX and TX micelles (Supporting Information), given its lower hydrophobicity. From Figure 2a-c, it can be seen that the partitioning behavior of the pesticide to the sorbed surfactant and to the micelles is quite different for the different soil particles within the same size group, based on their CEC. This effect can also be most clearly seen in Figure 2d for the different size fractions of the same soils (Ag#1 and Ag#3). For these equilibrium partitioning systems, we define a break-even concentration as the aqueous equilibrium surfactant concentration at which the aqueous equilibrium pesticide concentration in the presence of the surfactant is equal to that in the absence of the surfactant. As an example, the determination of the break-even concentration for the sand fraction of Ag#1 is depicted in Figure 2c. The breakeven concentrations were all determined experimentally. In order to clearly compare the slopes of the rising limbs of 3384

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 9, 2008

Diuron aqueous concentrations curves across the different size fractions, the scale of the x-axes in Figure 2 have been constrained to TX concentrations smaller than 3.0 g/L. However, the experiments were conducted for aqueous TX concentrations up 20 g/L, to determine the break-even concentrations. Figures for the entire range of TX concentrations for Diuron equilibrium sorption with the clay and silt fractions are presented in the Supporting Information. The measured break-even concentrations of the bulk soils and their size fractions are presented in Table 2. Thus, if sorption/ desorption equilibrium is assumed, the break-even concentration can be used to predict the pesticide desorption behavior within soil washing systems. Within a surfactantaided soil washing system, an enhanced pesticide desorption will not occur until the surfactant break-even concentration is reached in the aqueous phase. As can be seen, the sand fractions have the lowest break-even concentrations and the clay fractions have the highest break-even concentrations, by a factor of 5-6. Thus, desorption of the pesticide associated with the clay fractions will require significantly more surfactant than the other fractions. As shown in Figure 2d, above the CMC of TX, at any given equilibrium TX aqueous concentration, the equilibrium Diuron aqueous concentrations of the sand fractions are greater than those of the clay and silt fractions. Thus, in a soil washing system where active mixing occurs, pesticide molecules may transfer from the coarser to the finer fractions, possibly via micelles as an intermediate, resulting in additional pesticide accumulation on the clay fractions. Although this study was not designed to explicitly test this transfer mechanism, there is some evidence in the literature to support it (16). At the saturation surfactant sorption the aqueous Diuron concentrations decreased to the same level for the clay and silt fractions of each of the two soils (Ag#1 and Ag#3) (Figure

FIGURE 3. Kmc,dl as a function of equilibrium aqueous TX concentration for Ag #1 clay and silt particles. 2d). However, beyond the saturation surfactant sorption, with increasing micelle concentrations, the aqueous pesticide concentrations in equilibrium with the silt fractions were always higher than the ones with the clay fractions, suggesting higher affinity of the pesticide onto the clay fractions under the same aqueous equilibrium surfactant concentration. From a soil washing perspective, it indicates greater difficulty associated with desorbing pesticide sorbed onto clay fractions than silt fractions. Since the interaction between TX monomers and either pesticide was found to be negligible (c.f. Supporting Information), a dimensionless pesticide partitioning coefficient for this system can be defined as follows: R )

Pestoc + Pestss Pestw + Pestmc

(1)

Where R is the ratio of the mass of sorbed pesticide in solid phase to the mass of pesticide in aqueous solution (mg/mg); Pestoc, Pestss, Pestw, and Pestmc are the mass of the pesticide associated with original OC, sorbed surfactant, water, and surfactant micelles, respectively (all in mg). Based on a mass balance Pestt ) Pestoc + Pestss + Pestw + Pestmc

(2)

where Pestt is the total mass of pesticide added to soil–water–surfactant system (mg). Rearranging Pestoc + Pestss ) Pestt - (Pestmc + Pestw)

(3)

At equilibriumKmc ) 1000 × (Kmc, dl/ C pest, w) is a constant, where Kmc is a dimensional pesticide partitioning coefficient into the TX micelles (L/g) as reported in the Supporting Information, Kmc,dl is a dimensionless pesticide partitioning coefficient into the TX micelles (mg/mg), Cpest,w is the aqueous pesticide concentration (mg/L), and 1000 is a unit conversion factor. Also since Pestw ) Cpest,wVw, where Vw is the volume of water (L), then Pestw ) 1000 × (Kmc, dl/Kmc) Vw. Pestmc ) MmcKmc,dl, where Mmc is the mass of the TX molecules present as micelles (mg). Thus, the following expression can be derived: R )

Pestt - (MmcKmc,dl + 1000 × (Kmc,dl ⁄ Kmc)Vw) MmcKmc,dl + 1000 × (Kmc,dl ⁄ Kmc)Vw

(4)

Since R, Mmc, and Kmc can be measured and Pestt and Vw are known for the current experiments, the above equation can be solved for Kmc,dl. The relationship between Kmc,dl and the equilibrium aqueous TX concentration of Diuron, in the presence of Ag#1 clay and silt fractions, is presented in Figure 3. The x-axis in Figure 3 starts from an aqueous TX concentration 0.12 g/L, which is the CMC of TX. As can be seen, Kmc,dl decreases as the TX micelle concentration

FIGURE 4. Diuron desorption isotherm from the clay fractions in the presence of (a) water only and (b) TX (2.00 g/L), and (c) TX sorption during the desorption cycles. Note: the y axes of (a) and (b) are the Diuron sorbed concentration while that of (c) is the sorbed TX concentration. increases, since the capacity of the micelles to store pesticide molecules decreases with increasing loading. Also, the Kmc,dl of the clay fraction is always smaller than the Kmc,dl of the fraction, reflecting the stronger partitioning of the pesticide to the clay particles even in the presence of TX micelles. In sum, based on the results of the equilibrium sorption experiments the largest amount of surfactant is needed to desorb the pesticide associated with the clay fractions. Thus, it is the clay fractions that may determine the amount of surfactant needed for a surfactant-aided soil washing application. 3.4. Pesticide Desorption in the Presence of TX by Sequential Washing. Figure 4 presents Diuron desorption isotherms from the clay fractions in the absence (Figure 4a) and presence (Figure 4b) of surfactant. In the absence of surfactant, pesticide desorption proceeds monotonically (Figure 4a). Adding TX (Figure 4b) results in an abrupt increase in the amount of pesticide sorbed onto the soil particles during the first and sometimes even the second desorption steps, followed by a decrease in sorbed pesticide thereafter. The initial increase is caused by the sorption of the surfactant onto to the soil (Figure 4c), which is followed by partitioning of more dissolved pesticide onto the sorbed surfactant phase. Once the aqueous TX concentration exceeds VOL. 42, NO. 9, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3385

TABLE 3. Percentage of Pesticide Remaining Sorbed in the Presence (DS) and Absence (DW) of Surfactant, and Desorption Efficiency Coefficients (E) after Five Desorption Cycles with TX Diuron TX conc. (g/l)

0.00

1.00

Atrazine

2.00

3.00

0.00

1.00

2.00

3.00

Dw

Ds

E

Ds

E

Ds

E

Dw

Ds

E

Ds

E

Ds

bulk

Ag#1 Ag#2 Ag#3 clayey sediment

35% 32% 42% 28% 29%

36% 86% 84% 82% 22%

0.98 0.20 0.27 0.24 1.10

20% 53% 60% 46% 16%

1.24 0.69 0.70 0.74 1.18

16% 40% 42% 40% 14%

1.29 0.88 1.00 0.82 1.20

41% 33% 35% 32% 38%

20% 44% 52% 49% 18%

1.34 0.84 0.73 0.75 1.33

18% 31% 30% 33% 15%

1.38 1.04 1.08 0.98 1.37

13% 23% 24% 23% 11%

1.46 1.16 1.16 1.13 1.43

clay

Ag#1 Ag#2 Ag#3 clayey sediment

43% 41% 47% 39% 35%

53% 94% 80% 100% 54%

0.82 0.09 0.28 0.01 0.71

28% 77% 58% 84% 34%

1.25 0.39 0.56 0.26 1.01

27% 48% 45% 62% 32%

1.27 0.88 0.84 0.62 1.05

45% 37% 38% 35% 42%

68% 97% 98% 99% 69%

0.57 0.05 0.04 0.01 0.52

44% 86% 88% 92% 51%

1.00 0.21 0.2 0.12 0.84

40% 102% 99% 104% 32%

1.09 -0.0 0.01 -0.1 1.16

silt

Ag#1 Ag#2 Ag#3 clayey sediment

46% 53% 57% 53% 56%

40% 63% 58% 77% 51%

1.12 0.79 0.98 0.48 1.12

30% 46% 41% 60% 38%

1.29 1.15 1.36 0.85 1.39

26% 38% 40% 50% 30%

1.37 1.32 1.39 1.06 1.60

28% 36% 49% 52% 45%

40% 30% 36% 40% 37%

0.84 1.10 1.27 1.24 1.14

32% 36% 51% 51% 33%

0.95 1.00 0.96 1.01 1.23

24% 26% 46% 43% 22%

1.06 1.16 1.07 1.17 1.43

its CMC, micelles are present in the aqueous phase. The pesticide molecules partition into the micelles in the aqueous phase. A desorption efficiency coefficient, E, can be defined as follows: E )

(1 - Ds) (1 - Dw)

(5)

where Ds and Dw are the fractions of pesticide remaining sorbed after a given number of desorption steps in the presence (Ds) and absence (Dw) of surfactant relative to the initial amount of pesticide presorbed. In this study we considered five desorption steps. An E > 1 indicates enhanced pesticide desorption, while E < 1 represents an inhibited pesticide desorption. If at the end of five desorption steps the amount of pesticide remaining sorbed is greater than the initial amount of pesticide presorbed, a negative E value is generated, in which case, instead of desorption, an overall enhanced sorption occurs. The measured Ds, Dw, and E are presented in Table 3. From the results of the equilibrium sorption experiments, the correlation coefficients between the break-even concentrations and TX sorption capacities are 0.89 and 0.81 for Atrazine and Diuron, respectively, while those between the break-even concentrations and the amount of pesticide presorbed in the absence of TX are 0.69 and 0.55 for Atrazine and Diuron. On the other hand, the correlation coefficients between TX sorption capacity, at TX concentrations of 1.0, 2.0, and 3.0 g/L, and Ds are 0.83, 0.89, and 0.80 for Diuron, and 0.78, 0.90, and 0.88 for Atrazine. Therefore, the results of both equilibrium sorption and sequential washing indicate that the amount of TX sorbed determines the overall pesticide desorbability. Since TX sorption capacity is determined by CEC, this soil property controls surfactant-enhanced pesticide desorption. Also, these results indicate that to determine the amount of surfactant to be used in a surfactant-aided soil washing system, desorption of pesticide sorbed onto sorbed surfactant is a more significant factor than the amount of the pesticide presorbed in the absence of surfactant. These findings establish the usefulness of characterizing equilibrium sorption for predicting pesticide desorption behavior within a soil washing processes. In fact, the correlation coefficient between the break-even concentrations and E is -0.91, -0.88, -0.87 for TX concentrations of 3386

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 9, 2008

E

1.0, 2.0 and 3.0 g/L for Diuron, and -0.84, -0.85, and -0.87 for Atrazine for the same TX concentrations. The highly negative correlation between the two suggests that the measured break-even concentrations from the pesticide equilibrium sorption experiments serve to predict well the relative desorbability of an HOC in surfactant-aided soil washing systems using a particular surfactant. The results point to the difficulty in desorbing Atrazine from the clay fractions of the Ag#2, Ag#3, and clayey soils. This finding is consistent with the results of the equilibrium sorption experiments which showed the clay fractions of Ag#2, Ag#3, and clayey soils had the highest break-even concentrations among all the size fractions. This reflects the fact that the clay fractions of these soils have the highest CEC and thus highest TX sorption capacities. An additional factor is the lower affinity of Atrazine for TX micelles, relative to the more hydrophobic (higher Kow) Diuron. The clay fractions showed a statistically lower E than the bulk soils and silt fractions for either pesticide using TX (Table 3), indicating that pesticide desorption from the clay fractions using TX is more difficult than from the bulk soils and other fractions. It has been reported that pesticide desorption from clay fractions in the absence of surfactant exhibited higher desorption hysteresis than the bulk soils and other soil fractions (14). Presumably, the difficulty associated with desorbing pesticide out of the clay fractions is attributable to the higher structural hysteresis of the clay fractions, which is related to the difficulty of pesticides sorbed inside micropores diffusing out of the pore space (14). Thus, desorption of the pesticide from the smallest particles will require the largest amount of surfactant, which is consistent with the break-even concentrations determined by equilibrium sorption experiments. The results showed consistently lower desorption efficiency (E) for Atrazine than for Diuron (Table 3) with the clay fractions under the same conditions, due to its lower affinity to TX micelles than Diuron. As a result, more surfactant is required to desorb Atrazine molecules out of the sorbed phase than to desorb Diuron. Thus, one can expect a higher desorbability for more hydrophobic organic compounds than Diuron and Atrazine. Further work may result in a generalizable relationship between E and Kow. Thus, in view of the greater difficulty in desorbing pesticide out of the clay fractions, for a contaminated soil with high clay content, desorbing all the pesticide from all size fractions

might be less efficient. Instead of attempting to wash the entire bulk soil, a better strategy might be to either (1) use only the amount of surfactant that is sufficient to clean the coarse fraction, then separate the fine fraction, and dispose or treat it separately; or (2) to separate the coarse fractions mechanically and then treat the coarse and fine fractions separately.

Supporting Information Available Additional information on pesticide properties, soil size separation, pesticide and TX sorption kinetics, pesticide solubility enhancement by TX, calculation of sample correlation coefficient of two variables, and actual correlation between the measured parameters. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Mata-Sandoval, J. C.; Karns, J.; Torrents, A. Influence of Rhamnolipids and Triton X-100 on the desorption of pesticide from soil. Environ. Sci. Technol. 2002, 36, 4669–4675. (2) Chu, W. Remediation of contaminated soils by surfactant-aided soil washing. Prac.t Period. Hazard., Toxic, Radioact. Waste Manage. 2003, 1, 19–24. (3) Zhu, L. Z.; Yang, K.; Lou, B. F.; Yuan, B. H. A multi-component statistic analysis for the influence of sediment/soil composition on the sorption of a nonionic surfactant (Triton X-100) onto natural sediments/soils. Water Res. 2003, 37, 4792–4800. (4) Jafvert, C. T.; Heath, J. K. Sediment- and saturated-soil-associated reactions involving an anionic surfactant (dodecylsulfate) 1. precipitation and micelle formation. Environ. Sci. Technol. 1991, 25, 1031–1039. (5) Jafvert, C. T. Sediment- and saturated-soil-associated reactions involving an anionic surfactant (dodecylsulfate) 2.partition of PAH compounds among phases. Environ. Sci. Technol. 1991, 25, 1039–1045. (6) Li, Z. H.; Bowman, R. S. Sorption of perchloroethylene by surfactant-modified zeolite as controlled by surfactant loading. Environ. Sci. Technol. 1998, 32, 2278–2282. (7) Hayworth, J. S.; Burris, D. R. Nonionic surfactant-enhanced solubilization and recovery of organic contaminants from within cationic surfactant-enhanced sorbent zones. 1. experiments. Environ. Sci. Technol. 1997, 31, 1277–1283. (8) Sanchez-Camazano, M.; Rodriguez-Cruz, S.; Sanchez-Martin, M. Evaluation of component characteristics of soil-surfactantherbicide system that affect enhanced desorption of Linuron and Atrazine preadsorbed by soils. Environ. Sci. Technol. 2003, 37, 2759–2766. (9) Lee, J. F.; Liao, P. M.; Kuo, C. C.; Yang, H. T.; Chiou, C. T. Influence of a nonionic surfactant (Triton X-100) on contaminant distribution between water and several soil solids. J. Colloid Interface Sci. 2000, 229, 445–452. (10) Chu, W.; Chan, K. H.; Choy, W. K. The partitioning and modeling of pesticide parathion in a surfactant-assisted soil-washing system. Chemosphere 2006, 64, 711–716.

(11) Sheng, G. Y.; Xu, S. H.; Boyd, S. A. Mechanism(s) controlling sorption of neutral organic contaminants by surfactant-derived and natural organic matter. Environ. Sci. Technol. 1996, 30, 1553–1557. (12) Anderson, R.; Rasor, E.; Van Ryn, F. Particle size separation via soil washing to obtain volume reduction. J. Hazard. Mater. 1999, 66, 89–98. (13) Sheets, R. G.; Bergquist, B. A. Laboratory treatability testing of soils contaminated with lead and PCBs using particle-size separation and soil washing. J. Hazard. Mater. 1999, 66, 137– 150. (14) de Jonge, L. W.; de Jonge, H.; Moldrup, P.; Jacobsen, O. H.; Christensen, B. T. Sorption of prochloraz on primary soil organomineral size separates. J. Environ. Qual. 2000, 29, 206– 213. (15) Urum, K.; Pekdemir, T.; Copur, M. Surfactants treatment of crude oil contaminated soils. J. Colloid Interface Sci. 2000, 456– 464. (16) Yeh, C. K.; Young, C. C. Effects of soil fines and surfactant sorption on contaminant reduction of coarse fractions during soil washing. J. Environ. Sci. Health, Part A 2003, 38, 2697–2709. (17) Edward, D. A.; Adeel, Z.; Luthy, R. G. Distribution of nonionic surfactant and phenanthrene in a sediment/aqueous system. Environ. Sci. Technol. 1994, 28, 1550–1560. (18) Rodriguez-Cruz, M. S.; Sanchez-Martin, M. J.; Sanchez-Camazano, M. Surfactant-enhanced desorption of Atrazine and linuron residues as affected by aging of herbicides in soil. Arch. Environ. Contam. Toxicol. 2006, 50, 128–137. (19) Adeel, Z.; Luthy, R. G. Sorpton and transport kinetics of a nonionic surfactant through an aquifer sediment. Environ. Sci. Technol. 1995, 29, 1032–1042. (20) Carter, M. R. Soil Sampling and Methods of Analysis; Lewis Publishers: Boca Raton, FL, 1993. (21) Ko, S. O.; Schlautman, M.; Carraway, E. Partitioning of hydrophobic organic compounds to sorbed surfactants 1. Experimental studies. Environ. Sci. Technol. 1998, 32, 2769–2775. (22) Zhu, L. Z.; Chen, B. L.; Tao, S. Interactions of organic contaminants with mineral-adsorbed surfactants. Environ. Sci. Technol. 2003, 37, 4001–4006. (23) Park, J.; Jaffe, P. R. Partitioning of three nonionic organic compounds between adsorbed surfactants, micelles, and water. Environ. Sci. Technol. 1993, 27, 2559–2565. (24) Sun, S. B.; Inskeep, W. P.; Boyd, S. A. Sorption of nonionic organic compounds in soil-water-systems containing a micelle-forming surfactant. Environ. Sci. Technol. 1995, 29, 903–913. (25) Kibbey, T. C. G.; Hayes, K. F. A multicomponent analysis of the sorption of polydisperse ethoxylates nonionic surfactants to aquifer materials: equilibrium sorption behaviors. Environ. Sci. Technol. 1997, 31, 1171–1177. (26) Abu-Zreig, M.; Rudra, R. P.; Dickinson, W. T.; Evans, L. J. Effect of surfactant on sorption of Atrazine by soil. J. Contam. Hydrol. 1999, 36, 249–263. (27) Cano, M. L.; Dorn, P. B. Sorption of an alcohol ethoxylate surfactant to natural sediments. Environ. Toxicol. Chem. 1996, 15, 684–690.

ES702732G

VOL. 42, NO. 9, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3387