Environ. Sci. Technol. 2006, 40, 4274-4280
Enhanced Soil Washing of Phenanthrene by Mixed Solutions of TX100 and SDBS K U N Y A N G , †,‡ L I Z H O N G Z H U , * ,† A N D BAOSHAN XING‡ Department of Environmental Science, Xixi Campus, Zhejiang University, Hangzhou 310028, China, and Department of Plant, Soil, and Insect Sciences, University of Massachusetts, Amherst, Massachusetts 01003
Increased desorption of hydrophobic organic compounds (HOCs) from soils and sediments is a key to the remediation of contaminated soils and groundwater. In this study, phenanthrene desorption from a contaminated soil by mixed solutions of a nonionic surfactant (octylphenol polyethoxylate, TX100) and an anionic surfactant (sodium dodecylbenzenesulfonate, SDBS) was investigated. Phenanthrene desorption depended on not only aqueous surfactant concentrations and phenanthrene solubility enhancement but also the soil-sorbed surfactant amount and the corresponding sorption capacity of sorbed surfactants. The added surfactant critical desorption concentrations (CDCs) for phenanthrene from soil depended on both sorbed concentrations of surfactants and their critical micelle concentrations (CMCs). Phenanthrene desorption by mixed solutions was more efficient than individual surfactants due to the low sorption loss of mixed surfactants to soil. Among the tested surfactant systems, mixed TX100 and SDBS with a 1:9 mass ratio exhibited the highest phenanthrene desorption. Mixed micelle formation, showing negative deviation of CMCs from the ones predicted by the ideal mixing theory, was primarily responsible for the significant reduction of soil-sorbed amounts of TX100 and SDBS in their mixed systems. Therefore, mixed anionic-nonionic surfactants had great potential in the area of enhanced soil and groundwater remediation.
Introduction Desorption of hydrophobic organic compounds (HOCs), such as polycyclic aromatic hydrocarbons (PAHs), from soils and sediments is a key to successful remediation of contaminated soils and groundwater. Surfactants, especially surfactant micelles, can enhance the mobility and water solubility of HOCs, thus, increasing desorption efficiency (1-3). Therefore, surfactant-enhanced remediation (SER) has been suggested as a promising technology for removal of HOCs from contaminated soils and groundwater (4-6). Surfactant selection is important for SER technology. Anionic and nonionic surfactants have widely been used for remediation application because of their great solubilization, high efficiency in HOC mobilization, and their high volume of production in industry (1, 3, 6-8). However, further * Corresponding author phone: (+86)571-88273733; fax: (+86)571-88273450; e-mail:
[email protected]. † Zhejiang University. ‡ University of Massachusetts, Amherst. 4274
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 13, 2006
evidence has indicated that successful surfactant application to enhance remediation goes beyond selection of a surfactant (or a surfactant system) that will efficiently solubilize or mobilize HOCs; the surfactant must also be matched to the subsurface conditions so that it remains an active concentration (5, 9-11). Several aspects of single-surfactant behavior may reduce surfactant active concentration according to subsurface conditions (5, 10): (i) sorption of surfactants by soils and sediments; (ii) precipitation of anionic surfactant with inorganic salts such as Ca2+; (iii) partitioning of nonionic surfactant to nonaqueous phase liquids (NAPLs); and (iv) coacervation of surfactants to form a separate surfactantrich phase when the temperature is below the Krafft point of anionic surfactant or above the cloud point of nonionic surfactant, resulting in the surfactant concentration in water below the critical micelle concentration (i.e., no micelles form in aqueous phase). Recently, new surfactant types such as gemini surfactants and surfactant alternatives such as extracellular polymers have been developed, which are compatible with the contaminated medium, meanwhile having high solubilization and mobilization for HOCs (10, 12-16). Limited by production, the practical application and operative cost of these alternatives for remediation may make them not very realistic soon. Mixed surfactants are of practical and fundamental interest in industrial applications (2). They can be easily obtained because most commercial and industrial surfactant applications involve several surfactant types or several isomers of a particular surfactant type. Generally, surfactant mixtures exhibit a number of synergistic advantages in their practical applications over the use of single surfactant type because of the formation of mixed micelles (2, 7-19). For example, mixed anionic-nonionic surfactants exhibit cloud points higher than those of the pure nonionic surfactants, along with Krafft points lower than those of the pure anionic surfactants (17, 18). Furthermore, addition of nonionic surfactants to anionic surfactant solutions can decrease precipitation between anionic surfactant and multivalent electrolyte (e.g., Ca2+) (19). Therefore, such mixed surfactant systems could be expected to be more applicable over a wide range of subsurface conditions such as temperature, salinity, and hardness than individual surfactants. Moreover, our recent work showed that mixed anionic-nonionic surfactants had advantageous solubility behavior (20) and low partitioning loss to NAPLs (21) as compared to individual surfactants. Consequently, in the surfactant-enhanced remediation of contaminated soils and groundwater, anionic-nonionic mixed surfactants may be more efficient than single surfactants. Washing experiments showed that anionic-nonionic surfactant mixtures were more effective in displacing gasoline and perchloroethylene than individual ones (2225). Anionic-nonionic surfactant mixtures were also observed to give better desorption efficiency than both anioniccationic surfactant mixtures and cationic-nonionic surfactant mixtures (26). Synergistic solubilization has been suggested to be responsible for enhanced desorption by mixed surfactants in these studies. However, surfactant sorption by soils and sediments may be significant and thus affect the washing efficiency (5, 9-11). To date, sorption of such mixed surfactants by soils and sediments and their performance in enhanced desorption of HOCs are still unclear. Therefore, in this study we explore (i) sorption of the mixtures of a nonionic surfactant (octylphenol polyethoxylate, TX100) and an anionic surfactant (sodium dodecylbenzenesulfonate, SDBS) by a phenanthrene-contaminated soil and (ii) its effect on phenanthrene desorption to evaluate 10.1021/es060122c CCC: $33.50
2006 American Chemical Society Published on Web 05/25/2006
the performance of anionic-nonionic mixed surfactants. Phenanthrene was selected as a representative of polycyclic aromatic hydrocarbons (PAHs), which is difficult to desorb from subsurface media because of slow desorption kinetics from soils and sediments (27). TX100 and SDBS were employed in this study because they are widely used in industrial application and studied for subsurface remediation applications (7, 10, 11).
Materials and Methods Chemicals and Analytical Methods. Phenanthrene (purity > 98%) was obtained from Aldrich Chemical Co. Its molecular weight and log Kow value are 178 g mol-1 and 4.57, respectively (20). Anionic surfactant, sodium dodecylbenzenesulfonate (SDBS, C12H25C6H4SO3-Na+), with purity of 95%, was purchased from Tokyo Kasei Kogyo Co. and nonionic surfactant, octylphenol polyethoxylate (TX100, C8H17C6H4O(CH2CH2O)9.5H), with purity of 98%, from Sigma Chemical Co. Surfactants were used without further purification. HPLC analyses were conducted for surfactants and phenanthrene with a flow rate of 1.0 mL min-1 and a mobile phase of 80% methanol and 20% water. A wavelength of 224 nm was used for SDBS and TX100 analysis and 250 nm for phenanthrene. Before measurements, phenanthrene solutions were diluted to contain 50% methanol by volume. HPLC analyses were performed (Hitachi Instruments, L-7000 series, Japan) with a C18 reverse-phase column (Wakosoil, 250 mm × i.d. 4.6 mm, Japan). Preparation of Phenanthrene-Contaminated Soil. A vadose zone soil was collected from Hangzhou, China. It was air-dried and ground to pass through a 0.28 mm sieve. Its organic carbon and clay contents were about 1.43 and 47.8%, respectively. This phenanthrene-free soil was then spiked with phenanthrene dissolved in acetone and stirred vigorously for 30 min to promote homogeneous distribution of phenanthrene in soil (16). Then acetone was evaporated by leaving the sample resting for 3 days at 30 °C under a hood. The contaminated soil, with final phenanthrene concentration of 202 mg kg-1, was aged for 1 month before desorption experiments. Measurement of Surface Tension and Critical Micelle Concentration. Surface tension of TX100, SDBS and their mixtures were determined with a Model 20 surface tensiometer (Fisher Scientific Co.) using the method described by Kile and Chiou (1). Surfactant critical micelle concentration (CMC) values were estimated by their surface tension curves over a wide concentration range. The determined CMC for SDBS was 963 mg L-1, which was higher than 732 mg L-1 reported by Saiyad et al. (17) and lower than 1400 mg L-1 reported by Rouse et al. (10). The determined CMC for TX100 was 167 mg L-1, in agreement with 130 mg L-1 reported by Kile and Chiou (1) and 157 mg L-1 reported by Saiyad et al. (17). Solubilization Tests. Apparent solubilization experiments of phenanthrene by surfactant solutions were performed using a batch equilibration technique at 25 ( 1 °C. Phenanthrene solids and 20 mL of surfactant solutions were added to 25 mL Corex centrifuge tubes with Teflon-lined caps. The amount of phenanthrene added was slightly over that required to saturate the water. These centrifuge tubes were then equilibrated on a reciprocating shaker for 48 h (150 rpm). Then, the solid and aqueous phases were separated by centrifugation (4000 rpm) for 60 min, and the phenanthrene concentrations were determined by HPLC. Surfactant Loss to Soil and Phenanthrene Desorption from Soil. All desorption experiments were also performed using the batch equilibration technique at 25 ( 1 °C. A total of 20 mL of surfactant solution was mixed with 2 g of contaminated soil in 25 mL Corex centrifuge tubes. These tubes were shaken on a reciprocating shaker for 1 h at 150
FIGURE 1. Enhanced solubility curves of phenanthrene by mixed TX100 and SDBS solutions. The insert is for a low surfactant concentration range (0-1500 mg L-1, showing the CMCs). rpm (16), and the solid and aqueous phases were separated by centrifugation (4000 rpm for 30 min). Then, the concentrations of SDBS, TX100, and phenanthrene were determined by HPLC. The sorbed amount of SDBS and TX100 were calculated by the difference from initial and equilibrium surfactant concentrations. The residual concentrations of phenanthrene in soil were calculated by the difference from initial concentrations in soil and the concentrations in washed solutions.
Results Solubilization and Partitioning of Phenanthrene in Surfactant Solutions. Enhanced solubility of HOCs is always used as an index for evaluating the surfactant-enhanced remediation process. Apparent solubility (Sw*) of phenanthrene in mixed solutions of TX100 and SDBS was examined and compared with that in individual surfactant solutions. Sw* of phenanthrene increased slightly at surfactant concentrations below the CMC and increased significantly at surfactant concentrations above the CMC (Figure 1). The experimental surfactant CMCs are listed in Table 1. At surfactant concentration of 20 000 mg L-1, phenanthrene Sw* values were in the range of 180-680 mg L-1 (Figure 1), 2 orders of magnitude higher than its intrinsic water solubility (Sw, 1.08 mg L-1). Solubilizing efficiency of mixed surfactant solutions for phenanthrene increased with the fraction of TX100 (Figure 1). Among the tested surfactant systems, TX100 showed the highest solubilizing efficiency and affinity for phenanthrene, while SDBS with the lowest. The solubility enhancement is related to surfactant monomer and micelle concentrations and the corresponding solute partition coefficients (1), which can be expressed by the following equation:
Sw*/Sw ) 1 + KmnXmn + KmcXmc
(1)
where Sw* and Sw are explained as above; Xmn is the concentration of surfactant monomers in water; Xmc is the concentration of surfactant micelles in water; Kmn is the partition coefficient of a solute between surfactant monomers and water, and Kmc is the solute partition coefficient between the aqueous micellar phase and water. All concentrations are expressed in moles per liter. An alternative method to quantify surfactant solubilizing performance in the ranges of surfactant concentrations above CMCs is to measure the molar solubilization ratio (MSR) (3). VOL. 40, NO. 13, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4275
TABLE 1. Surfactant CMCs and the Scmc, MSR, Kmn, and Kmc of Phenanthrene in Surfactant Solutionsa mass ratios (TX100-SDBS) CMC (mmol L-1) SCMC (mmol L-1) MSR log Kmn log Kmc
10:0 0.2679 0.01306 0.1193 3.63 5.66
9:1 0.2703 0.01324 0.0908 3.64 5.54
7:3 0.2756 0.01172 0.0608 3.52 5.43
5:5 0.3725 0.01242 0.0418 3.45 5.25
3:7 0.5243 0.01138 0.0266 3.22 5.10
1:9 0.6177 0.00859 0.0208 2.83 5.12
0:10 2.764 0.01600 0.0182 2.77 4.79
a CMC is the critical micelle concentration of surfactants; S cmc is the apparent solubility of HOCs at the CMC; MSR is the molar solubilization ratio; Kmn is the partition coefficient of a solute between surfactant monomers and water; and Kmc is the solute partition coefficient between the aqueous micellar phase and water.
FIGURE 2. Mutual effect of SDBS and TX100 on their sorbed loss to soil: (A) TX100 loss by SDBS and (B) SDBS loss by TX100. MSR is defined as the number of moles of compound solubilized per number of moles of micellized surfactant and can be calculated as follows:
MSR ) (Sw* - SCMC)/(Cs - CMC)
(2)
where Cs is the surfactant concentration at which S w* is evaluated and SCMC is the compound apparent solubility at CMC. MSR may be obtained from the slope of solubilization curves in the ranges of surfactant concentrations above the CMCs. Based on the molar volume of water (1.8 × 10-2 L mol-1 at 25 °C), Kmc can also be calculated from MSR (20),
Kmc ) 55.4 × MSR/[SCMC(1 + MSR)]
(3)
The experimental SCMC, MSR, Kmn, and Kmc of phenanthrene in surfactant solutions are listed in Table 1. MSR and log Kmc of phenanthrene in TX100 solutions were higher than that in mixed surfactant solutions and SDBS. The experimental MSR and Kmc values were lower than the predicted ones using the ideal solution theory (2); i.e., no positive synergistic solubilization was observed in this case (Figure S1 of the Supporting Information). Sorption Loss of Surfactants to Soil. Figure 2A shows the effect of SDBS on the sorption loss of TX100 to soil, while the effect of TX100 on sorbed SDBS is shown in Figure 2B. Both TX100 and SDBS were highly sorbed by soil. The loss of TX100 is attributed to sorption (9). Hydrogen bonding (28) and electrostatic attraction between the negatively charged soil surface and the TX100 molecule with slight positive charges (29) may be responsible for TX100 sorption. Due to the negatively charged soil surface, anionic surfactants are expected to be sorbed less than nonionic surfactants because of electrostatic repulsion. However, precipitation of anionic surfactant ion (DBS-) with divalent cations (e.g., Ca2+) would be significant and result in SDBS loss (10, 30). Since precipitates of anionic surfactants can be redissolved by micelles in aqueous solutions due to counterion binding of Ca2+ into these micelles (19, 30, 31), the large decrease in 4276
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 13, 2006
FIGURE 3. Total sorbed loss of mixed surfactants to soil. SDBS loss at equilibrium concentrations above its CMC (Figures 2B and 3) indicates that SDBS loss was mainly attributed to precipitation. The maximum amount of sorbed TX100 decreased with increasing fraction of SDBS (Figure 2A), while that of SDBS declined in an order of 0:10 > 5:5 > 1:9 > 3:7 ) 7:3 > 9:1 (Figure 2B). The anomaly of SDBS sorption could be attributed to the enhanced positive charge on the soil surface by sorbed TX100 and thus decreased the electrostatic repulsion between SDBS and the soil surface (32), but reduction of SDBS sorption by adding TX100 was significant (Figure 2B). Total sorption loss of mixed TX100 and SDBS surfactants was lower than individual ones at special conditions (Figure 3). For example, the total sorbed amount of the mixtures, at a mass ratio of 1:9 and an initial concentration of 7000 mg L-1, was about 15 mg g-1, significantly lower than that of TX100 (about 70 mg g-1) and SDBS (about 50 mg g-1). The
FIGURE 4. Residual phenanthrene concentrations in soil after washing as a function of initial surfactant concentrations (A) and aqueous equilibrium surfactant concentrations (B). solubilization. Sorption effects of soil organic matter and soil-sorbed surfactants on phenanthrene desorption are discussed below.
FIGURE 5. Aqueous phenanthrene concentrations in surfactant solutions after washing as a function of equilibrium surfactant concentrations. above results indicate that mixed anionic-nonionic surfactants can decrease not only the sorption loss of nonionic surfactant but also the precipitation loss of anionic surfactant. These properties are desirable, implying that mixed surfactants will maintain their active concentrations in aqueous solutions with low loss to soil, and thus the lower dose will be needed when applied for subsurface remediation. Desorption of Phenanthrene by Mixed TX100 and SDBS Surfactants. Residual phenanthrene concentrations in soil after washing by surfactant solutions are displayed in Figure 4A.B. Residual phenanthrene concentration was about 199 mg/kg after washing by surfactant-free water. No enhanced desorption but slightly increased retardation of phenanthrene in soil was observed at relatively low surfactant concentrations (Figure 4A,B), accompanied with decreased aqueous phenanthrene concentrations in surfactant solutions (Figure 5). For example, residual concentration of phenanthrene in soil increased from 199 mg kg-1 by water to 202 mg kg-1 by TX100 solution (Figure 4A,B), while the aqueous phenanthrene concentrations decreased from 0.371 to 0.049 mg L-1 (Figure 5). To compare with the solubilization behaviors in Figure 1, the equilibrium surfactant concentrations in Figures 4B and 5 were used to eliminate the effect of surfactant loss on its aqueous concentrations. The phenanthrene concentrations in Figure 5 were significantly lower than those in Figure 1 at a given surfactant concentration, showing phenanthrene desorption cannot be explained simply by
There were critical desorption concentrations (CDCs) for added initial surfactant dose (Figure 4A), above which phenanthrene began to desorb and the desorption amount increased sharply with increasing surfactant concentrations. CDCs for tested surfactant systems changed with the mass ratios of TX100 to SDBS, showing an order 1:9 < 3:7 < 0:10 < 5:5 < 7:3 < 9:1 < 10:0 (Figure 4A). Corresponding to the CDCs (Figure 4A), the equilibrium surfactant concentrations in water were about their CMCs (Figures 4B and 5). Among the tested surfactant systems, mixed TX100 and SDBS surfactants with a 1:9 mass ratio had the highest phenanthrene desorption at the added dose up to 20 000 mg L-1. At 7000 mg L-1, for example, residual phenanthrene concentrations were about 90, 120, and 202 mg kg-1 for 1:9 TX100SDBS, SDBS, and TX100, respectively. The 1:9 TX100-SDBS surfactant mixture outperformed the other mixtures and individual surfactants tested (Figures 4B and 5), which could not be explained by solubilization alone (Figure 1). Interpretation of this anomaly on the basis of the surfactant sorption by soil is given below. Interaction between TX100 and SDBS in Solution. To interpret the behaviors of mixed surfactants in soil and their effects on desorption of phenanthrene from contaminated soil, it is essential to know the mutual interaction between TX100 and SDBS in solution. Surfactant mixtures in solution, in general, form mixed micelles, exhibiting strong deviations from idealitysthe experimental CMCs of mixed surfactants (CMCexp) are considerably lower than that predicted from ideal mixing theory (CMCideal) (2, 19). Ideal CMCs of mixed surfactant solutions at any molar ratio of TX100 to SDBS can be computed with ideal solution theory (2),
1 R 1-R ) + CMCideal CMCTX100 CMCSDBS
(4)
where R is the molar ratio of TX100 in mixed solutions and 1 - R is the molar ratio of SDBS. CMCTX100 and CMCSDBS are the critical micelle concentrations of TX100 and SDBS, respectively. CMCideal is the predicted critical micelle concentration of mixed surfactants with ideal solution theory. CMCexp listed in Table 1 was obviously lower than CMCideal at any ratio of TX100 to SDBS (Figure S2 of the Supporting Information), indicating the formation of mixed micelles (2, 17). VOL. 40, NO. 13, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4277
Discussion The equilibrium surfactant concentrations corresponding to the CDCs were about their CMCs (Figures 4 and 5), showing the enhanced desorption of phenanthrene is attributed to the enhanced phenanthrene solubility by surfactant micelles. However, CDCs were higher than CMCs, due to the significant sorbed loss of surfactants to soil (Figure 3). Therefore, CDCs depended on both the lost concentrations of surfactants and their CMCs, which can be calculated by the following equation,
CDC ) CMC + Closs ) CMC + Qloss(m/V)
(5)
where Closs (mg L-1) is the lost concentrations of surfactants to soil; Qloss (mg g-1) is the maximum amounts of surfactants sorbed in the soil; V (L) is the volume of surfactant solutions in washing; and m (g) is the added mass of the contaminated soil. In this case, CDCs were mainly determined by the lost concentrations of surfactants because they were much higher than CMCs, showing that CDC values were close to the initial concentrations at which the maximum sorbed amounts of surfactants by soil were reached (Figures 3 and 4A). Sorption and desorption are the results of phenanthrene distribution in the interface between soil and aqueous solutions. In the surfactant-free system, distribution of HOCs is mainly determined by soil organic matter (SOM) (33) and can be estimated with the soil-water distribution coefficient Kd according to partition theory (33). However, the addition of surfactants will change the distribution of HOCs because of additional partition of HOCs into both aqueous surfactant micelles and soil-sorbed surfactants (34, 35). The experimental phenanthrene concentrations were significantly lower than the predicted ones by solubilization theory using the parameters listed in Table 1 (Figure S3 of the Supporting Information), showing the significant effects of SOM and soil-sorbed surfactants on phenanthrene distribution. The apparent soil-water distribution coefficients (Kd*), expressed by the following equation, were often used to evaluate these complex HOC distribution behaviors (34).
Kd* ) (Kd + Csorb Ksf)/(1 + Kmn Xmn + Kmc Xmc)
(6)
where Csorb (mg g-1) is the soil-sorbed surfactant amount; Ksf (mL g-1) is the solute distribution coefficient normalized by soil-sorbed surfactants; (Kd + Csorb Ksf) describes the sorption capability of SOM and soil-sorbed surfactants for HOCs; and (1 + KmnXmn + KmcXmc) as explained in eq 1 describes surfactant solubilization for HOCs. Equation 6 can be used to calculate the Ksf values, and the calculated values varied largely (Figure S4 of the Supporting Information). The great variability in Ksf implies a fundamental weakness in understanding the effect of soil-sorbed surfactants on HOC desorption and the limitation of eq 6 in predicting application. However, eq 6 with parameters of solubilization (Table 1) can help in understanding the distribution behaviors of HOCs between soil and water in the presence of surfactants. According to eq 6, the soil-water distribution of HOCs in the presence of surfactants depends not only on the soilsorbed surfactant amount (Csorb) and their sorption capacity for HOCs (Ksf) but also on aqueous surfactant concentrations (Xmn and Xmc) and their solubility enhancement for HOCs (Kmn and Kmc). When added surfactant doses are lower than CDCs, i.e., the corresponding aqueous surfactant concentrations (X) lower than CMCs, no surfactant micelles can form in solution (Xmn ) X and Xmc ) 0). In this condition, Csorb increased sharply (Figure 3) and the sorbed surfactants were very effective for HOCs sorption (Figure S4 of the Supporting Information), while the value of (1 + KmnXmn) increased slightly because of low Kmn (Table 1) and Xmn (
