Remediation And Stimulation Of Selected Chlorinated Organic Solvents In Unsaturated Soil By A Specific Enhanced Electrokinetics

Colloids and Surfaces A: Physicochem. Eng. Aspects 287 (2006) 86–93

Remediation and stimulation of selected chlorinated organic solvents in unsaturated soil by a specific enhanced electrokinetics Jih-Hsing Chang a,∗ , Zhimin Qiang b , Chin-Pao Huang c a

Department of Environmental Engineering and Management, Chaoyang University of Technology, 168 GiFeng East Road, WuFeng Hsiang, 413 Taichung County, Taiwan b Environmental Research Center, Department of Civil, Architectural and Environmental Engineering, University of Missouri-Rolla, Rolla, MO 65409, USA c Department of Civil and Environmental Engineering, University of Delaware, Newark, DE 19716, USA Received 7 September 2005; received in revised form 6 March 2006; accepted 14 March 2006 Available online 28 March 2006

Abstract In this study, the remediation performance of a specific electrokinetic (EK) technology for removing chlorinated solvents including tetrachloroethylene (PCE), trichloroethylene (TCE), carbon tetrachloride and chloroform from unsaturated soils is investigated. The EK process is designed by a buffer solution of sodium acetate and a working-solution circulation system to neutralize the pH of the soil matrix. Results indicate that this EK process produces a roughly stable electro-osmotic (EO) flow rate (180 mL/day), pH (around 6.0), and current density (0.26–0.27 mA/cm2 ). All selected chlorinated organic compounds can be effectively removed from the soil with removal efficiency ranging from 85 to 98% after 2 weeks of treatment. The mobility of chlorinated solvents in soils increases with the increase of its water solubility, i.e., chloroform > carbon tetrachloride > TCE > PCE. The transport simulation by a mathematical diffusion–advection–sorption (DAS) model with the linear sorption isotherm is approximately feasible to describe the removal kinetics of chlorinated solvents in unsaturated soils under the enhanced EK conditions by tuning the parameter of mechanical dispersion. © 2006 Elsevier B.V. All rights reserved. Keywords: DNAPLs; Electrokinetics; Soil remediation; Transport simulation

1. Introduction Dissolved chlorinated solvents classified as dense nonaqueous phase liquids (DNAPLs) have been detected in the soil and groundwater in many waste sites around the world. Since the characteristics of high hydrophobicity and density, the solvents will be easily adsorbed in the soil particles and gradually release into the groundwater, which obtains high attention as an environmental concern. There are many treatment methods such as thermal desorption and vapor extraction has been used to clean this kind of contaminated sites. Among them, the electrokinetics (EK) remediation has been considered as a promising technique to treat polluted soils [1]. The major advantages of EK include: (1) removing contaminants from soils with low hydraulic conductivity and heterogeneity;

∗

Corresponding author. Tel.: +886 4 23323000x4210; fax: +886 4 23742365. E-mail address: [email protected] (J.-H. Chang).

0927-7757/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.colsurfa.2006.03.039

(2) controlling the flow direction of contaminants in the subsurface; (3) high removal efficiency for various contaminants; (4) high economic and in situ treatment technique [2]. Moreover, EK can be integrated with different remediation techniques like bioremediation and zero-valent iron to enhance removal efficiency [3]. The EK process applies a DC electric field to a porous medium to invoke some transport mechanisms including electroosmosis, electrophoresis, and electrolytic migration [4]. Electroosmosis is the movement of liquid (water in most cases) induced by excessive surface charge (zeta potential represents its magnitude) on porous media; electrophoresis is the motion of charged colloids in soil–liquid mixture, and electrolytic migration is the migration of ionic species in pore fluid. Based on specific physical-chemical properties of both contaminants and soils, target contaminants can be removed from soils by one or a couple of these mechanisms. The transport of inorganic and organic contaminants in saturated soils has been extensively studied. This process was used to effectively remove heavy metals from

J.-H. Chang et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 287 (2006) 86–93

soils [5–7]. Segall and Bruell [8] demonstrated that benzene, toluene, ethylene, xylene (BETX) and trichloroethylene (TCE) would migrate to the cathode in kaolinite by electro-osmosis. The removal of some polar organic species such as phenol [9] and acetic acid [10] by the EK process was also reported. It is seen that most previous research was focused on the removal of heavy metals, highly soluble non-ionic and ionic organic compounds in saturated soils [11,12]. However, there is scant information on the transport of DNAPLs in unsaturated soils under EK conditions. Therefore, the EK treatment of chlorinated organic solvents in the unsaturated soil is investigated in this study. In addition to the treatment performance of EK process, the quantitative description of the removal behavior by EK is essential for practical applications. For the saturated soil, the transport of non-ionic chlorinated solvents in the EK system can be approximately simulated by the diffusion–advection–sorption (DAS) equation [13]. The diffusion term consists of the molecular diffusion and dispersion, which vary with the heterogeneity of soils and fluid velocity. For the advection term, the electroosmotic (EO) flow ideally has been assumed as a steady velocity to simplify model development. Under this consideration, the zeta potential of soil particles was generally assumed constant, which resulted in a constant EO velocity in the advection term of the DAS equation [14]. The assumption of steady EO is not always valid because the zeta potential may vary dramatically during the EK process. In fact, protons (H+ ) produced at anode will gradually enter the soil–water mixture and thus acidify the soil, which results in the variation of soil’s zeta potential [15]. Since the electrolytic migration dominates the removal process of ionic species, the assumption of steady EO may be applicable for the transport of ionic species [16]. However, the EO flow is the main driving force to remove the non-ionic organic contaminants in the unsaturated soil. To avoid the unstable EO flow rate, a specific EK with a working-solution circulation system (an enhanced EK process) for maintaining soils at neutral pH was developed in this study to treat the contaminated soils. With this specific design, a constant electro-osmotic flow rate was achieved [17], which is expected to lead to a constant advection term in the DAS equation. For the sorption term in the DAS equation, a linear sorption isotherm has been commonly employed to describe the partitioning of organic contaminants to soils and applied to transport models frequently [18]. However, the concentration of contaminant and water content of soil may influence the sorption behavior of the pollutants in the vadose zone, which causes the nonlinear distribution [19]. Another important issue is whether the sorption–desorption of non-ionic organic compounds reaches equilibrium or not during the EK process. To date, the equilibrium status is widely adopted to describe the pollutants transport under the EK conditions [20,21]. On the basis of the above address, the equilibrium-DAS equation is used to simulate the transport kinetics of chlorinated organic solvents (target compounds: tetrachloroethylene (PCE), trichloroethylene, carbon tetrachloride and chloroform) in unsaturated soil by our enhanced electrokinetics. During the simulation, the parameters of diffusion, advection and sorption varied with experimental

87

conditions and treatment time are discussed to understand the transport phenomena. 2. Simulation aspects The transport of nonionic organic contaminants in soil–water system mainly consists of the following major processes: molecular diffusion, hydrodynamic dispersion, advection, sorption–desorption, and chemical or biochemical reactions. Since the experiments are conducted in a relatively short period of time, the chemical and biochemical reactions occurred in the soil–water are neglected. Based on the principle of mass conservation, the contaminant mass accumulation rate in an elementary control volume is equal to the difference between the contaminant mass inflow and outflow rates. A one-dimensional equation for mass conservation was written as follows [22]: ∂Ci ∂ 2 Ci ρ ∂qi ∂Ci = Di∗ 2 + v − ∂t ∂x ∂x φ ∂t

(1)

where Ci is the specific contaminant concentration in water phase (g/cm3 ); D* the sum of dispersion and diffusion coefficients (cm2 /sec); v the electro-osmotic velocity (cm/sec); ρ the soil density (g/cm3 ); φ the soil porosity; qi the specific contaminant concentration in soil phase (g/kg); and x is the distance from the anode of soil column (cm). The diffusion term, Di∗ (∂2 C/∂x2 ) comprises both molecular diffusion and mechanical dispersion. An empirical expression derived from the Stokes–Einstein equation was proposed to estimate the diffusion coefficient of organic compound in the aqueous phase by Wilke and Chang [23]. DA µAB (χMB )1/2 = 7.4 × 10−8 0.6 T VmA

(2)

where DA is the diffusion coefficient of solute A (cm2 /sec); µAB the viscosity of solvent B (water) (cP); T the absolute temperature (K); MB the molecular weight of solvent B (water); VmA the molecular volume of solute A at its normal boiling point (cm3 /g mol); and χ is the association parameter. The association parameter (χ) is solvent specific. The value of χ for water is equal to 2.6 [24]. Accordingly, Eq. (2) can calculate the diffusion coefficients of all target-chlorinated organic compounds applied to the transport simulation. For mechanical dispersion, only longitudinal dispersion is considered in the 1D DAS model. The longitudinal dispersion is approximately proportional to fluid velocity [25]: DL = αv

(3)

where DL is the longitudinal dispersion coefficient (cm2 /sec); and α is the longitudinal dispersity of the soil matrix (cm). If the equilibrium of sorption–desorption is reached, q is a function of solute concentration (C) and is independent of reaction time (t). The linear isotherm is used to describe the contaminant distribution in the soil and liquid phases: qi = Kd Ci (linear isotherm)

(4)

88

J.-H. Chang et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 287 (2006) 86–93

Table 1 Physical–chemical characteristics of the soil sample Physical–chemical characteristics

Result

Method

Sand (%) Silt (%) Clay (%) pH ECEC (meq/100 g)

14.0 38.0 48.0 7.6 20.5

Organic matter (%)

1.7

Hydrometer Hydrometer Hydrometer In 0.01 M CaCl2 The sum of exchangeable K, Ca, and Mg Heating at 105 ◦ C for 2 h, then at 360 ◦ C for 2 h Heating at 105 ◦ C for 24 h Constant-head pH meter and zetameter Coccine dye adsorption

Moisture (%) Hydraulic conductivity (10−8 cm/s) pHzpc Specific surface area (m2 /g)

13.6 2.5 2.2 10.7

ECEC: effective cation exchange capacity.

where Kd is the partition coefficients for the linear isotherms. In this research, the DAS equation with a linear isotherm was solved by MATLAB® based on the computation schemes developed by Sung [26] and on given initial and boundary conditions.

Fig. 1. Schematic of the electrokinetic system for removal of chlorinated organic contaminants from soils.

3. Methodology 3.1. Soil properties and diffusion coefficients of chlorinated organic compounds The soil collected from a specific DOE waste site was airdried and sieved through a no. 10 standard sieve (2 mm openings). Table 1 shows the physical–chemical properties of this soil with corresponding analytical methods adopted. It is noted that the soil had a low hydraulic conductivity (2.5 × 10−8 cm/s), which was suitable for the EK treatment. The pHzpc was determined under a constant ionic strength of 1 mM NaCl by a zetameter (Pen-Kem Inc., Hudson, NY, USA). Table 2 lists the diffusion coefficients of PCE, TCE, chloroform and carbon tetrachloride calculated by Eq. (2). These values would be used to substitute into the DAS diffusion parameters. 3.2. Electrokinetic reactor Fig. 1 shows the schematic of the laboratory EK system. All EK experiments were conducted at a constant voltage of 12 V. The air-dried soil was preheated at 105 ◦ C for 24 h and blended with 0.01 M sodium acetate solution and a desired amount of chlorinated organic contaminants. The soil–solution mixture (water content ca. 20%) was then carefully packed into the central unit of the EK reactor (length 10 cm, diameter 10 cm) with a membrane placed on each side of this unit. There-

after, both sides of the central unit were sealed with parafilm to prevent the chlorinated organics from significant evaporation loss. The unit was allowed to equilibrate overnight so that a homogeneous contaminant distribution was obtained in the soil. The 24-h equilibrium sorption was justified by sorption kinetic tests (data not shown here). The initial soil concentration was measured after the soil specimen was packed in the reactor for 24 h. To control the solution pH, a siphon pipe and a pump were used to automatically neutralize the acid produced at the anode with the base produced at the cathode during the EK operation. Both catholyte and anolyte solutions contained 0.01 M sodium acetate which is biodegradable and nontoxic. A hexane extractor was used to trap the organic compounds released from contaminated soils. To measure the contaminant concentrations in the extractor with time could describe the removal efficiency of the EK system. At pre-selected time, the soil sample in the central unit was sliced into 10 equal sections. In order to extract the target organics from the soil, 2 g of the treated soil was mixed with 2 mL 0.1 M sulfuric acid and 5 mL hexane, and then shaken for 24 h. The contaminant concentrations in the extract were measured by gas chromatography and electron capture detector (GC/ECD) (Hewlett-Packard model 5890, DE, USA). Then, the contaminant concentrations in the soil at various sections could be determined with operation time.

Table 2 The diffusion coefficient of selected organic compounds (Werther, 1976) Organic species

Molar volume (cm3 /mol)

Diffusion coefficient (×10−6 cm2 /sec)

Diffusion coefficient (cm2 /h)

PCE TCE Chloroform Carbon tetrachloride

116.0 98.1 83.3 101.2

8.71 9.63 10.60 9.45

0.0314 0.0347 0.0382 0.0340

J.-H. Chang et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 287 (2006) 86–93

89

4. Results and discussion 4.1. Electrokinetics treatment During the EK remediation, the acidity of the pore fluid is closely related to the electrolysis reactions. If the pH of the anode reservoir is not controlled, upon the application of a direct electric current, the oxidation of water takes place at the anode which will create an acid front in the soil. The reduction of water occurs at the cathode which renders the cathode reservoir solution basic. The oxidation and reduction reactions of water are shown as follows: 2H2 O − 4e− → O2 + 4H+ (anode) 2H2 O + 2e− → H2 + 2OH− (cathode) Accordingly, the pH may drop to below 2 at the anode while increase to above 12 at the cathode depending on the electric current applied. The acid front will advance toward the cathode, which results in acidification of the soil [1]. Under acidic pH condition, the EO flow will decrease and facilitate the dissolution of metal ions from the soil. Some released metal ions such as Al(III) may have potential toxicity to plants and adversely affect nutrient uptake, in addition to groundwater acidification [27]. As a result, it is desirable to control the working solution pH around the neutral range. Fig. 2 shows the change of working solution pH and current density as a function of operation time. Results show that after 3-day operation, a steady state condition is reached: the pH stabilizes at 5.9 and the current density stabilizes at 0.26–0.27 mA/cm2 . This can be attributed to sodium acetate buffer and the circulation system. A constant current implies the electrical resistance of soil specimen maintains stable, which results in the steady electricity consumption. In addition, the soil pH and water content also remain constant during the EK treatment, as shown in Fig. 3. The EO flow rate maintains constant due to the constant pH, current density, and water content. It can

Fig. 2. The solution pH and current as a function of time. Experimental conditions: electrolyte concentration = 10−2 M CH3 COONa; water content = 20% (w/w); voltage = 12 V.

Fig. 3. The average pH and water content profiles as a function of normalized distance from the anode. Experimental conditions: electrolyte concentration = 10−2 M CH3 COONa; water content = 20% (w/w); voltage = 12 V.

be seen that this enhanced EK system has a closely neutral pH range and produces a constant EO flow. The electrical potential gradient of the all experiments is controlled at 1.2 V/cm. The water levels of two reservoirs are kept the same, therefore, no hydraulic gradients occur and only the EO flow velocity ought to be considered for the advection term of DAS transport model. In our previous study, the EO flow velocity based on the identical EK system has been determined around 1.2 cm/day and a semi-empirical mathematical equation has been established [17]. Fig. 4 shows the accumulated pore volume and flow rate of EO flow as a function of time. The pore volume represents the aqueous volume in the soil matrix, that is, one pore volume is around 400 mL because of 20% water content in the specimen. Results show a stable EO flow rate of ca. 180 mL/day and total transport fluid approaches to 4.5 pore volume after the 10-day treatment. Fig. 5 shows the removal efficiency of chlorinated organic compounds as a function of the pore volume. Results show

Fig. 4. The average pore volume and flow rate of electro-osmosis as a function of time. Experimental conditions: electrolyte concentration = 10−2 M CH3 COONa; water content = 20% (w/w); voltage = 12 V.

90

J.-H. Chang et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 287 (2006) 86–93

to close water solubility. Accordingly, this specific EK system can effectively remove the chlorinated organic solvents from the unsaturated soils. 4.2. Transport simulation

Fig. 5. The removal efficiency of chlorinated organic compounds as a function of pore volume. Experimental conditions: electrolyte concentration = 10−2 M CH3 COONa; water content = 20% (w/w); voltage = 12 V.

that chloroform can be removed at 98% after a 2-pore volume treatment. TCE and carbon tetrachloride can be removed after a 3-pore volume treatment, at 85 and 95%, respectively. PCE contaminated soils can be cleaned up to 90% after a 5-pore volume treatment. In general, the organic compound with high water solubility is not preferably sorbable onto soils [28]. Accordingly, TCE is relatively easier than PCE to be removed from the soil phase because the water solubility of PCE (110 mg/L) is much lower than that of TCE (1100 mg/L). Likewise, chloroform having a water solubility of 8000 mg/L can be better removed than all other selected organic contaminants. The removal efficiency of carbon tetrachloride is similar to that of TCE due

The numerical solution on the basis of linear sorption isotherm at equilibrium for solving the DAS model was developed according to Eq. (4). By applying the numerical solution to experimental data, the DAS model can achieve most fitted performance as tuning certain parameters. Fig. 6 shows the input menu for all parameters for solving the DAS solution (written by MATLAB). Parameters such as water content, contamination concentration, weight of soil–solution mixture, soil density, transport distance, and operation time obtained from the EK experimental data are constants. Theta value, θ, is fixed at 0.5 in order to proceed a stable numerical computation. The diffusion coefficients (i.e., molecular diffusion coefficients) of all target-chlorinated organic compounds listed in Table 3 can be calculated by Eq. (2). Since the EO flow requires at least one day reaching a steady state, the average EO flow rate varies with increasing operation time. According to the experimental results, the average value of EO flow rate for 1, 3, 5, and 7 days is 0.05, 0.1, 0.18, and 0.19 cm/h, respectively. Only two parameters, partition coefficient and dispersivity, are not determined. Therefore, these two parameters are tuned to fit and analyze experimental data. Table 3 lists fitted values of partition and dispersion coefficient for chlorinated organic compounds in the EK treatment. The average partition coefficient of PCE, TCE, carbon tetrachloride, and chloroform is 1.2, 0.6, 0.4, and 0.15 L/kg, individually. Bruell reported that the partition coefficient of TCE was

Fig. 6. The parameters input menu of the diffusion–advection–sorption model for chlorinated organic compounds under EK influence.

J.-H. Chang et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 287 (2006) 86–93

91

Table 3 Simulation parameters of the transport for chlorinated organic compounds under electrokinetics Diffusion coefficient (cm2 /h)

Partition coefficient (L/kg)

EO velocity (cm/h)

Dispersion coefficient (cm2 /h)

PCE 24 h 72 h 120 h 168 h

0.0314 0.0314 0.0314 0.0314

1.0 1.2 1.3 1.3

0.05 0.10 0.18 0.19

0.13 0.26 0.47 0.49

TCE 24 h 72 h 120 h 168 h

0.0347 0.0347 0.0347 0.0347

0.4 0.5 0.5 1.0

0.05 0.10 0.18 0.19

0.13 0.26 0.47 0.49

Carbon tetrachloride 24 h 72 h 120 h 168 h

0.034 0.034 0.034 0.034

0.3 0.3 0.4 0.4

0.05 0.1 0.18 0.19

0.13 0.26 0.47 0.49

Chloroform 24 h 72 h 120 h 168 h

0.038 0.038 0.038 0.038

0.15 0.15 0.15 0.15

0.05 0.10 0.18 0.19

0.13 0.26 0.47 0.49

0.85 L/kg in the clay under electrokinetics [21]. That is close to the partition coefficient obtained in this study (0.6 L/kg). The slight different may be due to clay with higher organic contents than the soil used in this study. To the consideration of dispersion coefficients, the value of DL is dependent on the fluid

velocity (v) in the soil according to Eq. (3) (DL = αv). Based on above transport simulation, the fitted value of diversity, α, is 2.6 cm which is a unique value in this case. The diversity may vary with many characteristics of soil layer and experimental scale. Generally speaking, the dispersion coefficient

Fig. 7. Chlorinated organic concentration profile in the soil as a function of transport distance: (a) PCE, (b) TCE, (c) carbon tetrachloride, (d) chloroform. Experimental conditions: electrolyte concentration = 10−2 M CH3 COONa; water content = 20% (w/w); voltage = 12 V.

92

J.-H. Chang et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 287 (2006) 86–93

obtained in the laboratory scale is smaller than that in the real site [29]. It is seen that the fitted values of partition coefficients for PCE and TCE increase with increasing time according to Table 2. The partition coefficient of PCE and TCE increases from 1.0 to 1.3 L/kg and from 0.4 to 1.0 L/kg, respectively. The hypothesis is that the affinity between the soil and organic compounds varies with solute concentration, which results in the various Kd values [15]. When preceding the EK process, the contaminant is removed and its concentration in the soil–water matrix decreases gradually. Based on the above hypothesis, the partition coefficient of organic contaminants in EK system increases with time. However, this kind of concentration effect on the relative polar chlorinated compounds is insignificant. According to Table 3, the partition coefficient of carbon tetrachloride varies with time slightly (from 0.3 to 0.4 L/kg) and that of chloroform maintains a constant value (0.15 L/kg). It needs more precise investigation to discuss and justify the hypothesis in future study. Fig. 7 shows movement profiles of organic contaminants in the soil specimen as a function of the travel distance for PCE, TCE, carbon tetrachloride, and chloroform, respectively. In this figure, the dot points represent experimental data and solid curves are the model simulation. It can be seen that the errors between the simulation and experimental data occur relatively high in the beginning end (anode) of the travel distance. This can be attributed to the incomplete extraction efficiency of the hexane extractor. When the contaminants were washed out to the cathode chamber, the effluent was pumped through the hexane extractor and returned into the anode reservoir. The effluent may still contain few contaminants and enter the soil again after the extraction process. As a consequence, this small amount of contaminants makes the experimental values higher than simulation ones. In addition, there are errors occurring near the cathode end. This results from the low organic concentration in the cathode reservoir. Due to this low organic concentration, the concentration gradient at the cathode boundary increases dramatically in comparison with that at the soil matrix. The driving force of molecular diffusion is proportional to the gradient of concentration; therefore, the contaminants near the cathode are released faster than that in the soil, which is beyond the simulation assumptions. 5. Conclusions According to experimental results and data elucidation, some conclusions can be drawn for this research: 1. The specific enhanced EK system can effectively remove chlorinated organic compounds such as PCE, TCE, carbon tetrachloride, and chloroform from the unsaturated soil at an efficiency ranging from 85 to 98% in a matter of days. 2. The pH and conductivity of working solution, current density, and EO flow rate of the specific enhanced EK system can be operated under a stable status. 3. The higher the water solubility of a chlorinated organic compound, the higher its transport rate by the EK treatment.

4. The diffusion–advection–sorption model based on the linear sorption isotherm in equilibrium is approximately suitable to simulate the movement of chlorinated solvents in the unsaturated subsurface under the EK treatment by tuning the parameter of mechanical dispersion. Acknowledgment This study was supported by the U.S. Department of Energy (Grant no. DE-FG07-96ER14716). Contents of this paper do not necessarily reflect the views and the policies of the funding agency. References [1] Y.B. Acar, A.N. Alshawabkeh, Principles of electrokinetic remediation, Environ. Sci. Technol. 27 (1993) 2638. [2] L. Takiyama, C.P. Huang, In situ removal of phenols from contaminated soil by electroosmosis process, in: Proceedings of the of 27th Mid-Atlantic Industrial & Hazardous Waste Conference, Lehight U., Technomic Pub. Co., Lancaster, PA, 1995. [3] J. Virkutyte, M. Sillanpaa, P. Latostenmaa, Electrokinetic soil remediation—critical review, Sci. Total Environ. 289 (2002) 97. [4] J.K. Mitchell, T.C. Yeung, Electrokinetic flow barriers in compacted clay, Transportation research records; no. 1289, TRB, Washington, D.C., 1991. [5] Y.B. Acar, J.T. Hamed, A.N. Alshawabkeh, R.J. Gale, Removal of cadmium(II) from saturated kaolinite by the application of electrical current, Geotechnique 44 (1994) 239. [6] B.E. Reed, P.C. Carriere, J.C. Thompson, J.H. Hatfield, Electronic (EK) remediation of a contaminated soil at several Pb concentrations and applied voltages, J. Soil Contam. 5 (1996) 95. [7] Z. Li, J.W. Yu, I. Neretnieks, A new approach to electrokinetic remediation of soils polluted by heavy metals, J. Contam. Hydrol. 22 (1996) 241. [8] B.A. Segall, C.J. Bruell, Electroosmotic contaminant-removal processes, J. Environ. Eng. 118 (1992) 84. [9] Y.B. Acar, Electrokinetic cleanups, Civil Eng. 34 (1992) 58. [10] P.C. Renaud, R.F. Probstein, Electroosmotic control of hazardous wastes, in: Proceedings of the 6th International Conference, Dedicated to the Memory of Veniamin Levich, April 6–8, 1987. [11] H.I. Chung, B.H. Kang, Lead removal from contaminated marine clay by electrokinetic soil decontamination, Eng. Geol. 53 (1999) 139. [12] K.R. Reddy, C. Supraja, Electrokinetic remediation of heavy metalcontaminated soils under reducing environment, Waste Manag. 19 (1999) 269. [13] R.F. Probstein, R.E. Hicks, Removal of contaminants from soils by electric fields, Science 260 (1993) 498. [14] B. Narasimhan, R. Sri Ranjan, Electrokinetic barrier to prevent subsurface contaminant migration: theoretical model development and validation, J. Contam. Hydrol. 26 (2000) 291. [15] J.H. Chang, Removal of selected nonionic organic compounds from soils by electrokinetic process. Ph.D. Dissertation, Department of Civil and Environmental Engineering, University of Delaware, United Stated, 2000. [16] E.D. Mattson, R.S. Bowman, E.R. Lindgren, Electrokinetic ion transport through unsaturated soil: 1. Theory, model development, and testing, J. Contam. Hydrol. 54 (2002) 99. [17] J.H. Chang, Z. Qiang, C.P. Huang, D. Cha, Electroosmotic flow rate: a semi-empirical approach, in: P. Gary Eller, W.R. Heineman (Eds.), ACS Symposium Series 778, 2000, pp. 247–266 (Chapter 5 in nuclear site remediation). [18] G. Musso, Transport phenomena in electrokinetic soil remediation, Math. Comput. model. 37 (2003) 589.

J.-H. Chang et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 287 (2006) 86–93 [19] A. Fares, B.T. Kindt, P. Lapuma, G.P. Perram, Desorption kinetics of trichloroethylene from powdered soils, Environ. Sci. Technol. 29 (1995) 1564. [20] A.P. Shapiro, R.F. Probstein, Removal of contaminants from saturated clay by electroosmosis, Environ. Sci. Technol. 27 (1993) 283. [21] C.J. Bruell, B.A. Segall, M.T. Walsh, Electro-osmosis removal of gasoline hydrocarbons and TCE from clay, J. Environ. Eng. 118 (1992) 68. [22] M.T. Van Genuchten, Movement of Solutes in Soils: Computer Simulated and Laboratory Results, Soil Chemistry, Elsevier, New York, 1982. [23] E.L. Cussler, Diffusion-Mass Transfer in Fluid Systems, Cambridge University Press, Cambridge, 1997. [24] T.B. Drew, G.R. Cokelet, J.W. Hoopes JR, T. Vermeulen, Advances in Chemical Engineering, Academic Press, New York, 1981.

93

[25] P.A. Domenico, F.W. Schwartz, Physical and Chemical Hydrogeology, John Wiley & Sons Inc., New York, 1998. [26] M.H. Sung, MATLAB GUI for contaminant transport with different adsorption mechanism, Project for CIEG 667: Computer application in civil engineering, University of Delaware, 1998. [27] E. Matzner, J. Prenzel, Acid deposition in the German soiling area: effects on soil solution chemistry and Al mobilization, Water Air Soil Pollut. 61 (1992) 221. [28] R.P. Schwarzenbach, P.M. Gschwend, D.M. Imboden, Environmental Organic Chemistry, New York, 1993. [29] C.W. Fetter, Applied Hydrogeology, second ed., Merril Publishing, Columbus, Ohio, 1988.