Short Course

BASICS AND APPLICATIONS OF ELECTROKINETIC REMEDIATION

Handouts Prepared for a Short Course Universidade Federal do Rio de Janeiro (COPPE-UFRJ) Federal University of Rio de Janeiro By Akram N. Alshawabkeh, Ph.D., PE Department of Civil and Environmental Engineering 400 Snell Engineering Center Northeastern University 360 Huntington Avenue Boston, MA 02115 Phone (617) 373-3994 Fax (617) 373-4419 Email: [email protected] 19-20 November 2001 Rio de Janeiro, Brazil 1

Short Course Basics and Applications of Electrokinetic Remediation Akram N. Alshawabkeh, Ph.D., PE

TABLE OF CONTENTS

SECTION 1: INTRODUCTION AND OVERVIEW

3

SECTION 2: THEORETICAL BASIS FOR ELECTROKINETIC REMEDIATION

18

SECTION 3: PRACTICAL ASPECTS OF 1D AND 2D ELECTROKINETIC REMEDIATION

35

SECTION 4: FIELD DEMONSTRATION

59

SECTION 5: POTENTIAL ENHANCEMENT OF BIOREMEDIATION BY ELECTROCHEMICAL METHODS

75

LIST OF REFERENCES:

2

SECTION 1 INTRODUCTION AND OVERVIEW

The use of electric fields is an innovative method for in situ restoration of contaminated hazardous waste sites. Direct currents (DC) are applied across electrodes inserted in the soil to generate an electric field for mobilization and extraction of contaminants and for bio-geochemical modifications of polluted soils and slurries. The driving mechanisms for this technique, known as electrokinetic remediation, are transport under electric fields (in particular electroosmosis and ionic migration) coupled with electrolysis and geochemical reactions. Extraction and removal are generally achieved by electrodeposition, precipitation or ion exchange (for heavy metals) and collection and treatment of organics in external systems. Contaminants that could be treated by electric field applications include inorganic, organic, and radioactive compounds that are charged (ionic) or non-charged (polar and non-polar). The major advantages of the technology include: (a) it can be implemented in situ with minimal disruption, (b) it is well suited for fine-grained, heterogeneous media, where other techniques such as pump-and-treat can be ineffective, and (c) accelerated rates of contaminant transport and extraction may be obtained. This short course provides a review of the fundamentals and applications of electrokinetic remediation. The course describes the role of clay mineralogy, general electrokinetic and transport phenomena in soil under electric fields, followed by identification of electrolysis and geochemical reactions associated with application of electric fields in soils. Current technology status and considerations for practical implementation of the technology are presented. The short course also describes the potential for enhancing bioremediation of organic contaminants by eletrochemical methods

Soil and Groundwater Contamination in the US Soil contamination is a major threat to groundwater resources in the US and around the world. Soil and groundwater contamination is a result of decades on misuse and mishandling of earth resources. Many federal and state agencies, and private companies have contributed to the problem. In the US, activities by the Department of Defense and the Department of Energy caused significant soil contamination with organics (such as chlorinated solvents), heavy metal (such lead and chromium), and radionuclides. Contamination problems caused by other parties include pesticides, heavy metals and petroleum contamination among others. It estimated that there are more than 40,000 contaminated sites in the US, out of which, about 1,450 sites are listed on the National Priority List. The estimated cost for remediation is more than $ 350 Billion and the estimated cost for cleaning the NPL sites is about $ 40 Billion. Management and remediation of is challenging due to the difficulties in locating contamination, site characterization, evaluation of the extent of contamination, and risk assessment. In situ remediation of low permeability clayey soils that exhibit sorption capacity is a difficult task in managing contaminated sites. The low permeability of such soils limits extraction of the pore fluid or contaminants by hydraulic gradients. The sorption capacity prevents solubilizaion of the contaminants and therefore their bioavalibility or their transport. The microstructure and composition controls transport and reactive processes of contaminants in clays. A brief description 3

of clay mineralogy and its impact on transport and contaminant fate in the subsurface is provided.

Clay Mineralogy Clay minerals occur in small particle size and are formed of repetitive layers of unit cells. The cells consist of 2,3 or 4 structured sheets of tetrahedral and octahedral units. The sheets are stacked on top of each other such that the stacking is repetitive in each layer. The two basic units (Figure 1.1) of layer silicates are (a) Silica Tetrahedron – a silicon ion is tetrahedrally coordinated with four oxygens, and (b) Aluminum (Al) or Magnesium (Mg) Octahedron – Al or Mg ion is octahedrally coordinated with 6 oxygens or hydroxyls. Table 1.1 lists properties of three common minerals (kaolinite, montmorillonite and illite) in geotechnical engineering. Generally, three criteria can be considered for classification of clay minerals: 1. The height of the unit cell or “thickness of layer” 2. Composition, whether “dioctahedral” or “trioctahedral” and ionic content of layer. 3. Stacking segments of layers and degree of orderliness of stacking. Factors that define the geoenvironmental engineering behavior of clays include fabric/structure, cation exchange capacity, organic content and specific surface area. Fabric is the arrangement of particles, particle groups and pore spaces in a soil. Structure is the combined effect of fabric, composition and interparticle forces. For example, the variation of the hydraulic conductivity of clay is an outcome of the type of clay microstructure. Generally, discrete clay particles have a negative surface charge that influences and controls the particle environment. This surface electric charge can be developed in different ways, including the presence of broken bonds and due to isomorphous substitution1 (Mitchell 1993). The quantity of this surface charge varies from mineral to mineral and is affected by pore fluid environment. Figure 1.2 displays the effect of pH on the surface charge of kaolinite and montmoriloonite. Decreasing the pH reduces the value of the negative surface charge (and may also cause surface charge reversal). The clay particle-water-electrolyte system is usually considered to consist of three different zones; the clay particle with negatively charged surface, pore fluid with excess positive charge, and the free pore fluid with zero net charge. The net negative charge on the clay particle surfaces requires an excess positive charge (or exchangeable cations) distributed in the fluid zone adjacent to the clay surface forming the diffuse double layer (Figure 1.3). The quantity of these exchangeable cations required to balance the charge deficiency of clay is termed the cation exchange capacity (CEC), and expressed in milliequivalents per 100 grams of dry clay. Several theories have been proposed for modeling charge distribution adjacent to clay surface. The Gouy-Chapman diffuse double layer theory has been widely accepted and applied to describe clay behavior. A detailed description of the diffuse double layer theories for a single flat plate is found 1 Isomorphous substitution is the substitution of ions of one kind by ions of another type, with the same or different valence, but with retention of the same crystal structure. (e.g., Al in place of Si; Mg instead of Al and Fe(II) for Mg) (Mitchell 1993)

4

in Hunter (1981), Stumm (1992), Mitchell (1993), and Yeung (1993). The Gouy-Chapman theory (1910-1913) is considered the most successful theory that describes the diffused double layer. A brief description of the theory is provided. Poisson equation describes the relationship between the potential and charge, given by d2 =− dx 2 Where x is the distance from the surface (m), ψ is the electric potential (V), ρ is charge density (C/m2) and ε is the dielectric constant of the medium (C2 J-1 m-1). The charge density of the ions in the double layer is = e ∑ zi n i

where e is the electronic charge (1.6 x 10-19 C), zi and ni are the charge and concentration of ith ion, respectively. The concentration of the cations and anions in the double layer is given Boltzmann distribution,

n + = n 0e n − = n 0e

−(

(

ze ) kT

ze ) kT

where no: ionic concentration (ions/m3) in the free pore fluid, k: Boltzmann Coefficient (1.38x10-23 J ko) T: Temperature (ko) Accordingly, for a simple case of monovalent ions,   − e   e   e  = en 0 e  kT  − e  kT   = −2 en 0 sinh    kT   

Substituting ρ into the Poisson Equation, 2n e d2 e  = 0 sinh   2 dx  kT 

Solving the potential distribution equation for boundaries (ψ=ψ0 at x=0 and ψ =0 as x →∞) and then 5

simplifying yields the following distribution, =

0

e (− Kx )

which is known as the Debye-Huckel Equation and the normalized value of (1/K) represents the characteristic thickness of the double layer (known as Debye Length),

1  N7   =  K  8 Q o z 2 e 2 

1/ 2

The Debye length is inversely related to the cations valence and concentration of free pore fluid. As shown in Figure 1.4, cations of higher valence will reduce the thickness of the double layer (Debye length). Also increasing the pore fluid electrolyte concentration will reduce the Debye length (Figure 1.5). Stern (1924) evaluated the impact of ion size and pH on the double layer thickness. As pH increases, the thickness (1/K) increases. As ion size increases, (1/K) increases. An important characteristic of the double layer is the value of plane of shear potential, known as zeta potential (Figure 1.6). The value of zeta is less than the surface potential of the particle and represents the value at the slip plane, which located at a small unknown distance from the clay surface.

Electrokinetic Phenomena in Soils Electrokinetics is defined as the physicochemical transport of charge, action of charged particles, and effects of applied electric potentials on formation and fluid transport in porous media. The presence of the diffuse double layer gives rise to several electrokinetic phenomena in soil, which may result from either the movement of different phases with respect to each other including transport of charge, or the movement of different phases relative to each other due to the application of electric field. The electrokinetic phenomena include electroosmosis, electrophoresis, streaming potential, and sedimentation potential. Electroosmosis is defined as fluid movement with respect to a solid wall as a result of an applied electric potential gradient. In other words, if the soil is placed between two electrodes in a fluid, the fluid will move from one side to the other when an electromotive force is applied. Electrophoresis is the movement of solids suspended in a liquid due to application of an electric potential gradient. Streaming potential is the reverse of electroosmosis. It defines the generation of an electric potential difference due to fluid flow in soils. Sedimentation (or migration) potential, known as Dorn effect (Kruyt 1952), is an electric potential generated by the movement of particles suspended in a liquid. Figure 1.7 displays the electrokinetic phenomena in soils.

6

Under certain conditions (such as presence of appropriate mineral, high water content and low ionic strength of pore fluid), electroosmosis will have a significant role in electrokinetic soil remediation. Several theories are established to describe and evaluate water flow by electroosmosis including Helmholtz-Smoluchowski theory, Schmid theory, Spiegler friction model, and ion hydration theory. Descriptions of these theories are given in Gray and Mitchell (1967) and Mitchell (1993). Helmholtz-Smoluchowski model is the most common theoretical description of electroosmosis and is based on the assumption of fluid transport in the soil pores due to transport of the excess positive charge in the diffuse double layer towards the cathode. The rate of electro-osmotic flow is controlled by the coefficient of electro-osmotic permeability of the soil (ke), which is a measure of the fluid flux per unit area of the soil per unit electric gradient. The value of ke is assumed to be a function of the zeta potential of the soil-pore fluid interface, the viscosity of the pore fluid, soil porosity, and soil electrical permittivity. When the soil pores are treated as capillary tubes, the coefficient of electroosmotic permeability is given by, ke =

n

where ζ is the zeta potential (V), n is the porosity, and η is the viscosity (FT/L2). While hydraulic conductivity, kh, is significantly influenced by the pore size and distribution in the medium (Acar and Olivieri 1989), the electroosmotic coefficient of permeability, ke, according to the Helmholtz-Smoluchowski theory is dependent mainly on porosity and zeta potential. The value of ke has been assumed to be constant during the electrokinetic process as long as there is no change in the concentration of ions or pH of the pore fluid. Extensive research has been carried out on the zeta potential of the glass-water interface. There is a good qualitative agreement in the results of different studies. Hunter (1981) in a detailed description of theoretical and experimental treatise of the zeta potential in colloid science, displays the effect of pH and ion concentration in the pore fluid on zeta potential.. The effect of electrolyte chemistry on zeta potential could therefore represented by (Kruyt 1952), = A - B log C o where A, and B are two constants that are evaluated experimentally, and Co (M/L3) is the total electrolyte concentration. It is hypothesized that the drop in pH of the soil due to electrokinetic processing will cause a decrease in the coefficient of electro-osmotic permeability associated with the drop in zeta potential; hence, the electroosmotic flow will start to decrease and eventually stop at later stages of the process. The results of Acar et al. (1989), Hamed et al. (1991), and Acar et al. (1993) demonstrate the decrease and cessation of electroosmosis upon continued testing. Consequently, the ke value determined in one--dimensional tests is time--dependent and controlled by the chemistry generated at the electrodes.

7

Vane and Zang (1997) investigated the effect of pore fluid properties on electroosmostic permeability. The results displayed that the effect of pH on zeta potential and electroosmostic flow vary significantly depending upon the mineral type. Lockhart (1983) demonstrated that high electrolyte concentration in the pore fluid causes strong electrolyte polarization which limits electroosmotic flow. At a certain pH value and pore fluid ionic strength, the soil surface charge could drop to zero rendering a zero zeta potential or what is called the iso-electric point (Lorenz 1969). Negative surface charge of clay particles (negative zeta potential) causes electroosmosis to occur from anode to cathode while positive surface charge causes electroosmosis to occur from cathode to anode (Eykholt 1992; Eykholt and Daniel 1994). The electro-osmotic flow can virtually be eliminated at the iso-electric point.

Ion Migration Electric currents occur in soils due to ion migration, which is the transport of charged ions in the pore fluid toward the electrode opposite in polarity. The ionic mobility is a term used to describe the rate of migration of a specific ion under a unit electric field. A similar term is used in soils, but to account for soil porosity and tortuosity, the term is modified to “effective” ionic mobility. Rates of contaminant extraction and removal from soils by electric fields are dependent upon the values of the effective ionic mobilities of contaminants. Heavy metal ionic mobilities at infinite dilution are in the range of 10-4 cm2/Vs. Accounting for soil porosity and tortuosity, the effective ionic mobilities are in the range of 10-4 to 10-5 cm2/Vs. Accordingly, the rate of heavy metals transport in clayey soil is about few centimeters per day under a unit electric gradient (1 V/cm). As a result of ion migration in the soil pores, cations are collected at the cathode and anions at the anode. In summary, application of electric gradients in soil will result in two significant transport mechanisms; electroosmosis and ion migration. Electroosmosis draws contaminants with the flowing water under electric fields. Ion migration transports ions to the electrode opposite in polarity under electric fields. Electroosmosis and any other hydraulic flow will usually carry all types of solutes from one location to another, depending on flow direction. However, ion migration separates negatively and positively charged ions and cause their migration to opposite electrodes. Consequently, hydraulic flow might enhance the migration of certain ions, but retard migration of other ions (with opposite charge). The relative contribution of electroosmosis and migration to ion transport under electric fields varies for different soil types, water content, type of ion, pore fluid chemistry, and boundary conditions.

Electrolysis Application of direct electric current through electrodes immersed in water induces electrolysis reactions at the electrodes. Oxidation of water at the anode generates an acid front while reduction at the cathode produces a base front as described by the following electrolysis reactions, 2H2O - 4e4H2O + 4e-

O2 2H2

+ 4H+ + 4OH-

(anode) (cathode)

The prevailing of electrolysis reactions at the electrodes depends on the availability of chemical 8

species and the electrochemical potentials of these reactions. Although some secondary reactions might be favored at the cathode because of their lower electrochemical potential, the water reduction half reaction (H2O/H2) is dominant at early stages of the process. Within the first few days of processing, electrolysis reactions drops the pH at the anode to below 2 and increase it at the cathode to above 10, depending upon the total current applied (Acar et al., 1989; Acar et al., 1990; Acar and Alshawabkeh 1993). Studies (Hamed et al., 1991; Probstein and Hicks 1993; Eykholt and Daniel, 1994; Yeung and Datla, 1995) showed that while acid production enhances the process development of a high pH zone at the cathode adversely impacts extraction of heavy metals from soils.

Soil pH and Geochemical Reactions While the acid generated at the anode advances through the soil toward the cathode by ionic migration and electroosmosis, the base developed at the cathode initially advances toward the anode by diffusion and ionic migration. However, the counterflow due to electroosmosis retards the back-diffusion and migration of the base front. The advance of this front is slower than the advance of the acid front because of the counteracting electro-osmotic flow and also because the ionic mobility of H+ is about 1.76 times that of OH-. As a consequence, the acid front dominates the chemistry across the specimen except for small sections close to the cathode (Acar et al., 1990; Alshawabkeh and Acar 1992; Probstein and Hicks 1993; Acar and Alshawabkeh 1994). Geochemical reactions in the soil pores significantly impact electrokinetic remediation and can enhance or retard the process. Precipitation and sorption of heavy metals prevent their transport and thus limit extraction. Complexation could reverse the charge of the ion and might reverse direction of migration. These geochemical reactions, including precipitation/dissolution, sorption, redox and complexation reactions are highly dependent upon the pH condition generated by the process. The advance of the acid front from anode toward the cathode assists in desorption and dissolution of metal precipitates. However, formation of the high pH zone near the cathode results in immobilization to precipitation of metal hydroxides. Limitations of electrokinetic remediation due to catholyte high pH required seeking innovative methods to enhance the technique and prevent immobilization of metals close to the cathode.

9

Table 1.1. Characteristics of Kaolinite, Illite and Montmorillonite Kaolinite (1:1 mineral) Octahedral Sheet: Bonding: Basal Spacing: Particle size: Cation Exchange Capacity: Specific Surface Area: Shape: Atterberg Limits Compression Index (Cc): Coeff. of Consolidation (Cv) Occurrence: Isomorphous Substitution:

Gibbsite (dioctahedral, Al) Hydrogen bond and Van der waals bond 7.2 D ~ 0.2 – 2 micrometer 3 – 15 meq/100g 10 – 20 m2/g 6-sided thick (bulk) plates LL (30 – 110); PL (25 – 40) 0.19 – 0.28 12 – 90 x 104 cm2/s Common One Si in each 400 by Al

Smectite (Montmorillonite) (2:1 mineral) Octahedral Sheet: Gibbsite (dioctahedral, Al) Bonding: Van der waals bond and exchangeable cations Basal Spacing: 9.6 - 4 D Particle size: ~ 0.1 micrometer Cation Exchange Capacity: 80 – 150 meq/100g Specific Surface Area: up to 700-840 m2/g Shape: Thin flakes like films Atterberg Limits LL (100 – 900); PL (50 – 100) Compression Index (Cc): 1.0 – 2.6 Coeff. of Consolidation (Cv) 0.06 – 0.3 x 104 cm2/s Occurrence: Common Isomorphous Substitution: In every 6 Al by one Mg; In 15% of Si by Al Illite (2:1 mineral) Octahedral Sheet: Bonding: Basal Spacing: Particle size: Cation Exchange Capacity: Specific Surface Area: Shape: Atterberg Limits Compression Index (Cc): Coeff. of Consolidation (Cv) Occurrence: Isomorphous Substitution:

Gibbsite (dioctahedral, Al) K ions (R= 1.33 D, R. of Hexagonal hole = 1.32 D) ~ 10 D ~ 0.1 – 0.2 micrometer 10 – 40; if no K: up to 150 meq/100g 65 - 100 m2/g Flakes LL (60 – 120); PL ( 30 – 60) 0.50 – 1.1 0.3 – 2.4 x 104 cm2/s Common 1/4 Si replaced by Al; charge is balanced by K; 9-10% K2O

10

Figure 1.1 Basic Unit of Layer Silicates (Mitchell, 1993)

11

0

pH 4

2

6

0.2 Montmorillonite Kaolinite Surface Charge (C/m2)

0.1

0

-0.1

-0.2

Figure 1.2 Effect of pH on the Surface Charge of Montmorillonite and Kaolinite

12

8

Clay Particle surface

Conc.

+++-+-+ ++-++++-+ +-++++-+ Double layer thickness Cations

Anions Distance from clay surface

Figure 1.3 A Schematic of Charge Distribution at the Clay Surface

13

Figure 1.4 Effect of Cation Valence on Charge Distribution in the Diffuse Layer

14

Figure 1.5 Effect of Free Pore Fluid Concentration on the Double Layer Potential Distribution

15

Potential (mV) ψ0

Moving particle plane

ζ

Distance Figure 1.6 Potential Distribution Showing the Slipping Plane (Zeta) Potential

16

E ( DC )

E ( DC )

Saturated Clay

Saturated Clay

(a) Electric Gradient Induces Water Flow

(c) Electrical Gradient Induces Particle Movement

Particle Movement

E

E

Saturated Clay

(b) Water Flow Induces Electric Potential E

(d) Particle Movement Induces Electric Potential

Figure 1.7 Electrokinetic Phenomena in Soil

17

SECTION 2 THEORETICAL BASIS FOR ELECTROKINETIC REMEDIATION

Theoretical understanding and simulation of the electrokinetic treatment demand a grasp of the mathematical formulation of transport processes, which are controlled by such variables as electrolysis reactions at the electrodes, pH and soil-surface chemistry, equilibrium chemistry of the aqueous system, electrochemistry of the contaminants, and geotechnical/hydrological characteristics of the porous medium. The complexity of transport processes necessitates simplifying assumptions which would allow adequate numerical simulation. The following assumptions are employed in the theoretical development presented in this section: (1) The soil medium is isotropic and saturated, (2) The porous medium is a solid framework of cation exchange surfaces with the pore space occupied by chemically reactive species in aqueous solution, (3) All fluxes are linear homogeneous functions of all driving forces (or potential gradients), (4) Isothermal conditions prevail (coupled heat transfer is neglected), (5)All the applied voltage is effective in fluid and charge transport, (6) Electrophoresis is not present,(7) The chemicoosmotic coupling is negligible, (8) Soil particles are treated as electrically nonconductive (insulators), (11) Surface conductance and streaming potential are negligible. Based on the assumptions, theoretical formulations are provided for transport mechanisms under electric fields. Contaminant transport mechanisms include hydraulic or fluid flow, species or mass transport, and charge transport. Following is a description of these transport mechanisms.

Fluid Flux Fluid flux result from application of a hydraulic gradient (Darcy’s law), an electric gradient (electroosmosis) and/or a chemical gradient (chemicoosmosis). Chemicoosmosis could be neglected because it is significant only in the presence of large chain molecules and in very active clay deposits (activity describes plasticity of the soil and equals plasticity index divided by clay fraction in percentage). Fluid flux per unit area of the porous medium due to hydraulic and electric gradients, Jw (L3L-2T-1), is given by,

J w = k h ∇(− h) + k e ∇(−Φ) where kh is the coefficient of hydraulic conductivity (LT-1), ke the coefficient of electroosmotic permeability (L2 V-1 T-1 KLVWKHK\GUDXOLFKHDG/ DQG LVWKHHOHFWULFSRWHQWLDO9  Extensive research has been carried out on the hydraulic conductivity of fine-grained soils with a relatively good understanding of the fundamental factors affecting its value (Mitchell 1956; Olson and Daniel 1981; Boynton and Daniel 1985; Acar and Olivieri 1989; Daniel 1989). These studies indicate that microstructure and fabric are the factors that highly influence fluid transport in fine-grained deposits. Other factors affect kh include soil porosity and pore size distribution. Presence of uniformly distributed fine size pores result in lower hydraulic conductivity while presence of few macro pores result in higher hydraulic conductivity even if soil porosity is the same 18

in both cases. Figure 2.1 shows two different microstructures and pore size distributions that produce two different hydraulic conductivities even at the same porosity. Electrokinetic soil remediation induces changes in the pore fluid chemistry, diffuse double layer, soil fabric and consequently the hydraulic conductivity. Furthermore, electroosmotic consolidation is expected to take place and influence the hydraulic conductivity value. In attempting to provide a mathematical formulation of electrokinetic soil remediation, hydraulic conductivities are generally assumed to be constant in time and space because; (a) there is no clear mathematical formalism that can describe the effect of pore fluid chemistry on soil fabric and consequently the hydraulic conductivity, and (b) the uncertainties in evaluating the hydraulic conductivities are more significant than the changes expected in its values. Contribution of each component due to hydraulic and electric gradients is dependent upon the ratio of coefficient of electroosmotic permeability relative to hydraulic conductivity (ke/kh). Soil type, microstructure, and pore fluid conditions are the factors that impact this ratio. In course-grained soils this ratio is very small and goes to zero due to almost nonexisting electroosmotic flow and relatively high hydraulic conductivities ( > 10-3 cm/sec) of such soils. On the other hand, in soft, fine-grained soils the ratio of (ke/kh) becomes significant as ke is usually in the order of 10-5 cm2/V.sec while kh is less than 10-5 cm/sec (10-7 cm/sec for clayey soils).

Mass Flux Mass flux of different chemical species relative to pore fluid is a result of different coupled potential gradients. Hydrodynamic dispersion is mass transport due to a chemical concentration gradient. Migrational mass flux is mass transport of charged species to electrode opposite in polarity due to an electric potential gradient. Filtration or ion sieving is mass transport due to a hydraulic gradient. Total mass flux of dissolved species also includes the advective component due to species transport by the flowing fluid. Hydrodynamic dispersion is a result of two basic phenomena; mechanical dispersion and molecular diffusion. While mechanical dispersion occurs as a result of velocity variation within the porous medium, molecular diffusion is mass transport due to the difference in thermal kinetic energy of the molecules. Mechanical dispersion is a significant mechanism in contaminant transport in ground water (Perkins and Johnston 1963, Bear 1972) because of the relatively high hydraulic conductivity and advective hydraulic flow in such deposits (higher than 10-5 cm/sec). On the other hand, molecular diffusion is the primary process that controls hydrodynamic dispersion in clay deposits due to the low advective hydraulic flow in these deposits. Therefore, only molecular diffusion is considered in mass transport under concentration gradients. The total mass transport of chemical species per unit cross-sectional in a saturated soil medium under hydraulic, electric and chemical concentration gradients is described by, J i = Di ∇(- ci ) + ci (u i + k e) ∇(-Φ ) + ci k h ∇(- h) *

*

19

where Ji (ML-2T-1) is the total mass flux of the ith chemical species per unit cross sectional area of the porous medium, ci (ML-3) is the molar concentration of the ith chemical species, Di* (L2T-1) is the effective diffusion coefficient of the ith chemical species, ui* (L2 T V-1) is the effective ionic mobility of the ith species. Transport mechanisms included in are diffusion, ion migration, electroosmotic advection and hydraulic advection. A schematic of mass transport of cationic and anionic species is provided in Figure 2.2. Transport profiles are based on the assumptions that water advection components (electroosmosis and hydraulic) act from anode to cathode. The advective flow enhances transport of cationic species, which migrates from anode to cathode, and retards transport of anionic species, which migrates from cathode to anode. The effective diffusion coefficient in the porous medium, Di*, is related to the respective diffusion coefficient in free solution, Di, by (Bear, 1972; Gillham and Cherry, 1982; Shackelford and Daniel, 1991), * Di = Di n

ZKHUH GLPHQVLRQOHVV LVDQHPSLULFDOFRHIILFLHQWDFFRXQWLQJIRUWKHtortuosity of the medium. 9DOXHVRI VSDQRYHUDZLGHUDQJHIRUdifferent saturated and unsaturated soils. Experiments are often necessary to determine its value for a specific soil type. Shackelford and Daniel (1991) VXPPDUL]HUHSRUWHG YDOXHVIRUGLIIHUHQWVRLOW\SHV7KHVHYDOXHVare as low as 0.01 and as high as 0.84, mostly ranging between 0.2 to 0.5. Diffusion coefficient for different ions at infinite dilution are available in most electrochemistry handbooks and references. These values represent the maximum values attained under ideal conditions. Many factors might affect the molecular diffusion coefficient such as the electroneutrality requirement, concentration, and electrolyte strength (Shackelford, 1991). Shackelford and Daniel (1991) investigate the effective diffusion coefficients, D*, of different inorganic chemicals in compacted clay. Their results demonstrate that molding water content and compaction methods have little effect, if any, on the effective diffusion coefficients. Generally, changing molding water content in compaction tests result in significant changes in the compacted soil microstructure (Mitchell 1993). According to results of Shackelford and Daniel (1991) the change in compacted clay microstructure due to different molding water contents will have little effect on the effective diffusion coefficient of different chemicals. The effective ionic mobility, ui*, define the velocity of the ion in soil pores under unit electric field. There is no method yet devised to measure the effective ionic mobility (Koryta 1982); however, ui* can be theoretically estimated by assuming that Nernst-Townsend-Einstein relation between Di, the molecular diffusion coefficient, and ui, holds for ions in the pore fluid of soils (Holmes, 1962); *

Di zi F u = n ui = RT * i

where ui is the ionic mobility of species i at infinite dilution, zi is the charge of the ith species, F is Faraday’s constant (96,485 C/mol electrons), R is the universal gas constant (8.3144 J/K.mol), and T is the absolute temperature. Note that each of ui in this case has a value and a sign that reflects the 20

charge of species i (i.e. the ionic mobilities and effective ionic mobilities, ui and ui* will have negative values for anions and positive values for cations). The signs are included in the ionic mobilities of cations and anions to simplify the mathematical equations. Mass transport equation demonstrates that electrical gradient results in two mass transport mechanisms, ion migration and electroosmotic advection. The relative contribution of these components to species transport depends upon soil and contaminant characteristics. Both mechanisms require high degree of saturation. However, electroosmosis require specific conditions, including presence of clay minerals. It is also affected by pore fluid ionic strength and pH. Acidic pore fluid conditions could reverse the charge on the clay mineral surface, and consequently the electroosmostic flow (Lorenz 1969; Hunter 1981; Stumm 1992; Eykholt and Daniel 1994). At specific pH, the soil could have a zero net charge (pzc: point of zero charge, Stumm 1992) and electroosmosis could cease. Ionic migration on the other hand, occurs in all soil types, including course grained soils (sand and gravel). Acar and Alshawabkeh (1993) compared the role of electroosmosis and ionic migration for different conditions. The results showed that in most cases ionic migration is more significant than electroosmosis. Best conditions for electroosmosis can result in a flow in the order of 10-5 cm/sec per unit voltage gradient (V/cm). Average effective ionic migration rates of different ions are usually in the order of 10-5 cm/sec, but also could be in the order of 10-4 cm/sec.

Charge Flux Applying a DC current through a soil-water-electrolyte medium generates an electric field causing charge transport. It is assumed that the soil pore fluid has a relatively high ionic strength that makes the contribution of the free pore fluid dominate the other charge transport mechanisms. Therefore, the contribution of the soil solids and the diffuse double layer ions on charge transport is neglected. The simplest form of electrical conductance of the soil is governed by Ohm’s law describing the current density (charge transport) in the pore fluid due to electrical gradients, *

I=



where I is the electric current density (CL-2T-1 DQG LVWKHHIIHFWLYHHOHFWULFDOFRQGXFWLYLW\RIWKH soil. However, charge flux has another component due to ion diffusion. Therefore, the total charge flux can be evaluated by using Faraday’s law for equivalence of mass flux and charge flux, N

I = ∑ zj F Jj j=1

substituting the value of Ji, the total charge flux is given by,

21

N

I = F ∑ z j D*j c j) +

*



j=1

ZKHUHWKHHIIHFWLYHHOHFWULFDOFRQGXFWLYLW\ RIWKHVRLOEXONGXHWRFKDUJHIOX[LVJLYHQE\ N

*

=

∑Fz

*

i

u i ci

i =1

It should be noted that the advective components (electroosmosis and hydraulic advection) of species mass flux will not result in any charge flux due to preservation of the electrical neutrality of free pore fluid, given by, N

∑c z =0 j

j

j=1

Conservation of Mass and Charge To evaluate changes per unit pore volume of the soil, it will necessary to apply the conservation equations to transport equations. Conservation of mass and charge in a unit volume of the soil under the set of assumptions employed require that,

∂ v = mv ∂t

w

∂h = − ∇ Jw ∂t

∂ n ci = − ∇ Ji + n R i ∂t

i = 1,2,..., N

∂ Te ∂ = Cp = − ∇I ∂t ∂t where v is the volumetric strain of the soil mass, mv is the coefficient of volume compressibility of the soil (L2F-1); Te is the volumetric charge density of the soil medium (CL-3); Cp is the electrical capacitance per unit volume (farad L-3); Ri (ML-3T-1) is the production/consumption rate of the ith aqueous chemical species per unit fluid volume due to geochemical reactions. The consolidation equation describes change in hydraulic head due to soil volume change. This equation becomes significant in the case where hydraulic gradients are used to enhance transport under electric 22

gradients. Significant changes in the hydraulic head and development of negative pore water pressure (suction) could occur due to nonuniform pH and voltage gradient distributions. The mass conservation equation describes transient reactive transport of i chemical species under hydraulic, electric and chemical concentration gradients. For zero electrical gradients, this equation will render the advective-diffusive equation. The last PDE of the system describes conservation of charge in the porous medium. The equation describes the rate of change in the electrical potential required to maintain electrical neutrality of the medium. For zero net change in charge, one should assume zero electric capacitance (Cp) of the soil.

Geochemical Reactions Reactive transport of charged species is controlled by rates of geochemical reactions (Ri). Electrolysis reactions at the electrodes usually produce extreme pH and chemistry conditions at the boundaries (unless amendments are used). The chemistry boundary conditions results in a complex system of geochemical reactions that include sorption (surface complexation and ion exchange), redox and precipitation-dissolution reactions. The term Ri could then be expanded to account for each reaction type, e.g., s aq p Ri = Ri + Ri + Ri

where Ris is a term for sorption, similarly Riaq is a term for aqueous reactions and Rip for precipitation/dissolution reactions. Two approaches have been developed and used to describe chemical reactions; instantaneous equilibrium approach and kinetics approach. In instantaneous equilibrium reactions species concentrations reach equilibrium instantaneously whereby in kinetic reactions approach concentrations in solution vary with time till they reach equilibrium. For several species, chemical reactions, e.g., dissolution, sorption and redox, have been found to vary with time before reaching equilibrium. Kinetics approach is expected to be more realistic for modeling these reactions. However, one could assume chemical reactions reach equilibrium at a very short time (relative to transport time) and can use equilibrium. Geochemical models (using kinetics or equilibrium) exist and one could incorporate these models for predicting rates of geochemical reactions during electric field applications. However, an overview of these reactions is provided as most modelers attempt to develop their own geochemical codes. Sorption Reaction The following general term has been considered for sorption evaluation: s

Ri =

∂ si ∂ si ∂ ci = n ∂ t n ∂ ci ∂ t

i = 1,2,..., N

ZKHUH LVWKHEXONGU\GHQVLW\RIWKHVRLOsi is the adsorbed concentration of the component j per unit mass of the soil solids (MM-1). The reversible term (∂si / ∂t) is often used to describe the 23

sorption rate. The equilibrium partitioning between the adsorbed phase and the aqueous phase of the chemical components are commonly measured under controlled temperature and applied pressure, and the resulting correlations of si versus ci are called adsorption isotherms. Several equilibrium models (linear, Freundlich, and Langmuir models) have been used to describe sorption of heavy metals on soils. Assuming instantaneous equilibrium in sorption reactions and linear isotherms,

∂ si = K di ∂ ci where Kdi is called the distribution coefficient, Kd, of species i. A retardation factor, Rdi, have been introduced and used in modeling species transport accounting for linear sorption as,

R di = 1 +

K di n

The retardation factors of species i, Rdi, define the relative rate of transport of a nonsorped species to that of a sorped species. For a nonsorped species, Rdi = 1. Simple isotherm sorption models ignore the potential effects of variations in pH, solute composition and ionic strength, redox potential, or processes such as competitive adsorption. Alternative, more robust (and complicated) sorption models include ion- or ligand-exchange, mass action models, and surface complexation models as described by Langmuir (1997), Kirkner and Reeves (1988), Yeh and Tripathi (1989) Davis and Kent (1990), Stumm and Morgan (1995), Bethke (1996). Aqueous Reactions In aqueous phase reactions, any complex j is the product of i’s reactant components, i.e.: Nc

∑a

ji

ci ↔ x j

j = 1,...., N x

i =1

where c‘i is the chemical formula for component i, x‘j is the chemical formula for the complex j, aji is the stoichiometric coefficient in complex j for component i. The law of mass action implies that: Nc

eq x j = K j ∏ cia ji

j = 1,...., N x

i =1

where Kjaq is the equilibrium constant for aqueous reaction j. The rate of accumulation of component i due to aqueous reaction j, Rjiaq, is:

24

aq R ji = a ji R j

The total rate of accumulation of component i due to all aqueous reactions is: Nx

aq R = ∑ R ji = aq i

j=1

Nx

∑a

ji

Rj

j=1

Precipitation/Dissolution Reactions It is necessary to account for the precipitation/dissolution reactions in the formulation of mass transport equations. In precipitation reactions, the chemical components are assumed to be composed of products,

pj →

Nc

∑b

ji

j = 1,...., N

ci

p

i =1

where p‘j is the chemical formula for precipitate j, bji is the stoichiometric coefficient in precipitate j for component i, and Np is the number of precipitates for component i. The production of the precipitate will not occur until the solution is saturated. Therefore, the law of mass action is written as, Nc

sp K j ≥ ∏ cib ji

j = 1,...., N p

i =1

where Kjsp is the solubility product equilibrium constant for precipitate j. By the same rationalization of previous formulations, the total rate of production of component i due to precipitation/dissolution reactions, Rip, is, Np

p R = ∑ R ij = p i

j=1

Np

∑b

p

ji

Rj

j=1

where Rjp is the rate of production of precipitate j.

General System for Modeling Transport The theoretical formalism results in a mathematical system of equations describing the transient reactive multi-component species transport under hydraulic, electric, and chemical gradients. The resulting system consists of partial differential equations for transport and algebraic equations for geochemical reactions. The transport PDEs are divided into three types. The first consists of one equation that describes transient fluid flow. The second consists of N number for equations that 25

describe reactive transport of N species. The third type is described by one equation for charge transport. Substituting fluid flux into volume change equation results in the electroosmotic consolidation given by,

∂h = cv ∇ 2 h + k e ∇2 Φ ∂t mv γ w This equation is necessary to describe changes in the hydraulic head across the soil, which impacts the advective component of mass transport. The transient reactive PDEs for mass transport are derived by substituting mass flux equation into mass conservation equation,

∂ n ci = D*i ∇ 2 ci + ∇(ci [ (u*i + k e) ∇Φ + k h ∇ h ]) + n R i ∂t where i=1,2,.....,N. Note that for the case of nonreactive solute transport (Ri=0), steady state fluid flux ( ∂h/ ∂x =const.) and no electrical gradient, the equation becomes the advectivediffusive solute transport equation widely used to describe nonreactive solute transport,

∂ n ci = D*i ∇ 2 ci - v ∇ ci ∂t where v = - kh ∇ h. Changes in the electric potential distribution across the soil as a result of changes in the geochemistry is formulated by substituting the charge flux equation in the charge conservation equation, N ∂Φ = F ∑ z j D*j ∇2 c j + ∇ (σ * ∇Φ ) Cp ∂t j=1

The total number of differential equations described for this system are N+2 which are N equations for mass transport, one equation for charge conservation, and one equation for fluid flow. The unknowns described in this system are N species concentrations, ci, one electric SRWHQWLDO RQHK\GUDXOLFSRWHQWLDOKDQG1XQNQRZQVIRUWKHUDWHRIi chemical reactions. Therefore 2n+2 unknowns are described by N+2 differential equations. The other N number of equations required for this system are the mass balance equations for the chemical reactions.

26

Preservation of Electrical Neutrality The following derivation is used to demonstrate that the charge transport equation preserves the electrical neutrality of the porous medium. For a unit volume of the soil, the rate of change in the electric charge equals the total rate of change of chemical species concentrations multiplied by their charge and Faraday’s constant, i.e., N ∂ n cj rate of change in electric charge = ∑ z j F ∂t j=1 Preservation of electrical neutrality requires that the total change in electric charge per unit volume equals zero. N

∑ zj F j=1

∂ n cj = ∂t

N

N

j=1

j=1

∑ z j F ∇(− J j) + n ∑ z j F R j

The total rate of change of all chemical species under chemical reactions times their charge is zero, N

∑z R j

j

=0

j=1

In other words, consider the following chemical reaction,

A ↔ m B+ l + l D m one mole of A will produce m moles of B+l and l moles of D+m. The total change in B+l concentration times its charge is m moles of B+l * (+l) = ml. The total change in D+m concentration times its charge is l moles of D+m * (-m) = -ml. Therefore, the total change in each one concentration times its charge is ml-ml=0. Substituting flux equations into charge conservation equation: N

∑ zj F j=1

N ∂ n cj N = ∑ z j F D*j ∆ 2 c j + ∑ z j F ∆ (c j [ (u*j + k e ) ∆Φ + k h ∆ h ] ) ∂t j=1 j=1

Simplifying and substituting Σ zj cj = 0, N

∑ F zj j=1

N ∂ n cj N = ∑ F z j D*j ∇2 c j + ∑ F z j ∇ ( c j u*j ∇Φ ) ∂t j=1 j=1

Accordingly, for Cp=0, we get:

27

N

∑z F j

j=1

∂ n cj =0 ∂t

Boundary Conditions Boundary conditions are required for hydraulic head, N chemical species concentrations, and electric potential. Usually hydraulic heads are controlled either to provide constant head difference or constant flow rates. Thus, hydraulic head boundary conditions are easily identified. In most cases, zero head difference is applied between the cathodes and anodes. Boundary conditions for this case are,

h |S1 = h |S 2 = 0 where S1 is the boundary surface at the anode and S2 is the boundary surface at the cathode. Boundary conditions for charge conservation equation are developed from the current value at the boundary. Two type of boundary conditions could be applied. If the voltage difference is kept constant between the anode and cathode (current density changes depending upon electric conductivity) then the following boundary conditions are used,

Φ |S1 = Φ max

Φ |S 2 = 0

However, if the current is maintained constant (voltage changes depending upon electric conductivity), then the following boundary conditions are used, N

[ F ∑ z j D*j∇ c j +

*

∇

] |S1 = I

j=1

|S 2 = 0 Identifying boundary conditions for the partial differential equations describing species transport is calculated based on electrolysis reactions at the electrodes. When inert electrodes are used in groundwater, oxidation of water at the anode generates an acid front, while reduction at the cathode produces a base front by the following electrolysis reactions: 2 H 2 O - 4 e- - > 4 H + + O 2 2 H2 O + 2 e- - > 2 OH- + H 2

Anode Cathode

While these electrolysis reactions are the most likely to occur (if no amendments are used), the prevailing electrolysis reactions at the electrodes depend upon the availability of other species, pH, 28

and the electrochemical potentials of their reactions. In the presence of multi-species at the cathode or anode, electrolysis reactions will occur depending on the electrochemical potential of each reaction. It is necessary to evaluate these potentials for different species. In general, the overall electrolysis (cell) reaction comprises two independent half-reactions at the anode and at the cathode. Consider the following half-cell reaction vo O + n e → v r R

where O is the oxidized form, R is the reduced form, and vo and vr are the stoichiometric coefficients. From basics of thermodynamics, the free energy of this cell reaction is given by ∆ G = ∆ G o + RT ln

(R )vr (O )vo

where G is the Gibbs free energy and parenthesis represent activities. Since ∆G = - nFE E = Eo +

RT (O )vo ln nF (R )vr

This is the Nernst Equation (Bard and Faulkner 1980) and it provides the potential of the O/R electrode verses the natural hydrogen electrode (NHE) as a function of the activities of O and R. Equation 42 is useful in identification of the type of electrolysis reactions expected at the anode and at the cathode in the presence of multi-species. The electrolysis reaction that has the highest positive E will occur at the cathode, while the electrolysis reaction with the most negative E will occur at the anode. As a result of these electrolysis reactions, the chemistry at the electrodes will undergo continuous changes. These changes either enhance the electrokinetic process (e.g., generation of acid at the anode) or retard the process (e.g., generation of the base at the cathode). Chemical reagents are introduced at the electrodes to enhance electrokinetic process. Once the type of electrolysis reactions are identified, then the boundary conditions can be evaluated from mass equilibrium in electrodes compartments. The rate of concentration change of a specific species i in the electrode compartment will be equal to the net mass flow rate of the species. Figure 2.3 displays a schematic of the anode well showing the mass fluxes of a specific species, i. The in and out fluxes of species i in this well are (1) generation or electroplating at the electrode due to electrolysis reactions, (2) in/out flux at the soil boundary due to transport, (3) injection of any enhancement agent in the compartment, (4) extraction of the anolyte from the electrode well and (5) chemical reactions of species i. Accordingly, the rate of change of species i concentration in the anode well is given by,

∂ cia ~ = Ji dS - q E cia + q A CiA ± R iA V a ± R ielect Va V ∂t ∫ a

SA

29

where Va is the volume of water in the anode well, cia is the concentration of species i in the anode well, Ji is the flux of species i in/out of the anode/soil boundary, SA is the boundary surface area of the soil/anode well, qE is the flow rate of fluid extracted from the anode, qA is the flow rate of the (enhancement) fluid injected into the anode well, CiA is the concentration of species i in the enhancement fluid, RiA is the production/sink rate of species i in the anode well due to chemical reactions, Rielect is the production/sink rate specie i due to electrolysis reactions at the electrode. . Numerical Strategies Three approaches for numerical simulation are encountered for the developed system of differential and algebraic equations: 1. Differential and Algebraic Equations Approach (DAE): this approach consists of providing a solution to the mixed differential and algebraic equations in which the transport equations and chemical equilibrium reactions are solved simultaneously as a system (Miller and Benson 1983; Lichtner 1985). 2. Direct Substitution Approach (DSA): this approach consists of direct substitution of the algebraic chemical equilibrium equations into the differential transport equations to form a highly nonlinear system of partial differential equations (Vallocchi et al. 1981; Jennings et al. 1982; Rubin 1983; Lewis et al. 1987). 3. Sequential Iteration Approach (SIA): this approach consists of iterating between the sequentially solved differential and algebraic equations (Kirkner et al. 1984, 1985; Yeh and Tripathi 1991).

Electrokinetic Remediation Models Researchers have attempted to model contaminant transport under electrical gradients. Shapiro et al. (1989) and Shapiro and Probstein (1993) describe a 1-D model accounting for species transport under electric fields. The model accounts for ion diffusion, migration, and electroosmotic advection in predicting species transport rate. The model solves the charge flux equation in order to evaluate the nonlinear electric field distribution. The model assumes incompressible soil medium and thus constant hydraulic head distribution. Water electrolysis reactions are used to calculate constant flux boundary conditions for hydrogen ion at the anode and hydroxyl ion at the cathode. A steady state electroosmotic flux is calculated by averaging the electrical gradient and zeta potential across the soil sample. The results are compared with the experiments for the case of acetic acid extraction with constant voltage at the boundaries. Species incorporated in the code include acetate, hydrogen, hydroxyl, sodium, and calcium ions. Geochemical reactions include are first order sorption and water and acetic acid dissociation. Numerical solution is achieved using finite element method in the spatial domain and Adams-Bashforth integration in time. Comparisons show good agreement in one case of acetic acid removal from 0.4 m length kaolinite sample. Jacobs et al. (1994) followed the model described in Shapiro and Probstein to predict 1-D transport 30

of zinc under electric fields. The model uses an averages electroosmotic flow rate across the soil (i.e. electroosmotic flow is assumed independent of location). The model accounts for zinc precipitation and dissolution reactions and demonstrated the role of background ion concentrations on the process. Jacobs and Probstein (1996) further modify the code to model 2-D species transport under electric fields. They applied the 2-D code for the case of electroosmotic extraction of phenol from kaolinite. The model solves three PDEs for transport of phenol, sodium ion, and chloride ion. Hydrogen and Hydroxyl ion concentrations were calculated using zero net charge equation and water equilibrium equation. The model demonstrates 2-D phenol between one anode and one cathode. Limited geochemical reactions were incorporated (water and phenol dissociation) due to the complex 2-D simulation of the process. Mitchell and Yeung (1991) propose a model in a study of the feasibility of using electrical gradients to retard or stop migration of contaminants across earthen barriers. Principles of irreversible thermodynamics are employed and a one dimensional model is developed for transport of contaminants across the liner. Integral finite difference method was used to solve the problem and the model reasonably predicted the transport of sodium and chloride ions across the liner. Geochemical reactions were not incorporated in this model. Eykholt (1992) attempts to model the pH distribution during the process using mass conservation equation accompanied by empirical relations to account for the nonlinearity in the parameters controlling the process. One transport differential equation is formed assuming that hydrogen and hydroxyl ions have the same diffusion coefficients and ionic mobilities. In this model, the development of negative pore water pressure is modeled using the modified Smoluchoweski equation of Anderson and Idol (1986); and the complexity in electrical potential distribution is modeled using proposed empirical relations. Haran et al. (1997) present a 1-D model for extraction of hexavalent chromium from soils using electric fields. The model accounts for transport of H+, OH-, CrO4-2, K+, Na+, and SO4-.Geochemical reactions included sorption (described by a retardation coefficient) and water equilibrium. Acar et al. (1988) and Acar et al. (1989) present a one dimensional model to estimate pH distribution during electrokinetic soil processing. The model demonstrates the impact of electrolysis reactions on pH distribution during electrokinetic remediation. Alshawabkeh and Acar (1992) describe a modified formulation and present a system of differential/algebraic equations for the process and accounting for the chemical reactions of adsorption/desorption, precipitation/dissolution, and acid/base reactions. Acar and Alshawabkeh (1994) model the change in soil and effluent pH during electrokinetic soil processing. Two transient transport equations for hydrogen and hydroxyl ions are used together with water autoionizaion equation. This attempt assumes linear electric and hydraulic gradients throughout the process and disregards the coupling of these components. Alshawabkeh and Acar (1996) and Acar and Alshawabkeh (1996) enhance the model and modify the code for stimulating reactive extraction of heavy metals by electric fields. The model predicts reactive transport of hydrogen, lead, hydroxyl, and nitrate ions. The charge conservation equation is used to solve for electric field distribution and electroosmotic consolidation equation is used for predicting hydraulic head profile. The model accounts for lead hydroxide precipitation-dissolution, lead sorption (assuming linear pH dependent isotherm) and water equilibrium reactions.

31

Frequency (%)

0.01

0.1

1

10

Pore Diameter (µm)

Wet of optimum Dry of optimum

Figure 2.1 A schematic of microstructure and pore size distribution of clay at wet and dry of optimum

32

Co nc ent rati on

Initial Concentration

Co 2 An ode

1

3

Xh

4

Xe

Xm

1. Hydraulic Advection X1 = (kh ih) T 3. ElectroosmosisXe = (ke ie) T

Distance 2. Diffusion

4. Ion MigrationXm = (u* ie) T

Xm

Initial Concentration

Co 1

Xh

4

Cathode

2 Anode

Concentration

Cat ho de

3

Distance

Xe

1. Hydraulic Advection X1 = (kh ih) T 3. Electroosmosis Xe = (ke ie) T

2. Diffusion

4. Ion Migration Xm = (u* ie) T

Figure 2.2 Transport Mechanisms of (a) cations and (b) anions

33

Boundary S1

OHOHOHOHOHFigure 3.2

Figure 3.2 Boundary condition at the anode

34

SECTION 3 PRACTICAL ASPECTS OF 1D AND 2D ELECTROKINETIC REMEDIATION

Although significant research has been conducted in the 80’s and 90’s on electrokinetic remediation, one can argue that the fundamentals of the process are not yet well understood. The reason is that the objective of most of these studies was to demonstrate the feasibility of the remediation process and not to investigate the fundamental physicochemical and geochemical processes. For example, rates of electroosmosis and ionic migration in soils under heterogeneous, anisotropic, and partially saturated soils are not easy to establish. Even in homogenous and saturated conditions, the complex geochemical reactions make it difficult to predict transport rates. In most cases these reactions are assumed instantaneous while in fact they are time dependent and their rates significantly impact time and energy requirements. In any case, a discussion is provided for practical applications of the process based on the current understandings of the process. Few studies have looked into practical considerations for field implementations. Schultz (1997) provided an economic modeling and calculations of optimum spacings, time and energy requirements of one dimensional field applications based on electroosmotic transport.

Soil type Bench-scale and pilot-scale tests indicated that the technology can be successful in clayey to fine sandy soils. However, contaminant transport rates and the efficiency of the process depend heavily on soil type, mineral composition and pore fluid conditions. Reddy et al. (1997) demonstrated that presence of iron oxides in glacial till creates complex geochemical conditions that retards Cr(VI) transport. On the other hand, the same study showed that presence of iron oxides in kaolinite and Na-montmorillonite did not seem to significantly impact Cr(VI) extraction. Pamukcu and Wittle (1992) and Wittle and Pamukcu (1993) demonstrated removal of Cd2+, Co2+, Ni2+, and Sr2+ from different soil types at variable efficiencies. The results showed that kaolinite, among different types of soils, had the highest removal efficiency followed by sand with 10% Na-montmorillonite, while Na-montmorillonite showed the lowest removal efficiency. The results indicated that soils of high water content, high degree of saturation, low ionic strength and low activity (soil activity describes soil plasticity and equals plasticity index divided by % fines, clay and silt, in the soil) provide the most favorable conditions for transport of contaminants by electroosmotic advection and ionic migration. Highly plastic soils, such as illite, montmorillonite, or soils that exhibit high acid/base buffer capacity require excessive acid and/or enhancement agents to desorb and solubilize contaminants before they can be transported through the subsurface and removed (Alshawabkeh et al., 1997), thus requiring excessive energy. Hydraulic conductivities of different soil types can vary many orders of magnitude within a heterogeneous deposit. Consider a contaminated deposit containing interlayers of sand and clay, typical values of hydraulic conductivities of these soils are 1x10-3 and 1x10-9 m/s, respectively. If pump-and-treat is used to remediate such a heterogeneous deposit, most of the fluid flow induced will occur in the sandy layer and the clayey layer will be practically untreated. On the other hand, electric conductivities of these soils are within an order of magnitude. As a result, the electric field strengths in the different soil layers will be similar when an externally electric potential is applied across the deposit. Similar ionic migration rates of contaminant transport can be generated in 35

different soil layers within the heterogeneous deposit resulting in a more homogeneous cleanup. The ability to remove contaminants uniformly from a heterogeneous natural deposit is another distinct advantage of the technology. A combination of hydraulic and electric gradients could also enhance the process. Runnells and Wahli (1993) showed the use of ion migration combined with soil washing for removal of Cu2+ and SO42- from fine sands. A field study reported by Banerjee et al. (1990) also investigated the feasibility to use electrokinetics in conjunction with pumping to decontaminate a site from chromium. While soil chromium profiles were not evaluated in this study, the results showed increase in effluent chromium concentrations. One of the factors that should be considered carefully is the degree of saturation. Depending upon the target transport mechanism, the degree of saturation impacts both electroosmosis and ionic migration.

Contaminants type and concentrations Studies showed that removal of heavy metals, radionuclides, and selected organics by electrokinetics is feasible. Hamed (1990) and Hamed et al. (1991) demonstrated electrokinetic remediation using kaolinite samples mixed with Pb2+ at various concentrations below and above the soil cation exchange capacity. The process removed 75% to 95% of lead at concentrations of up to 1,500 mg/kg across the test specimens at reported energy expenditure of 29 to 60 kWh/m3 of soil processed. However, since no enhancement procedure was used, most of the removed lead was found deposited at section close the cathode. Acar et al. (1994) demonstrated 90% to 95% removal of Cd2+ from kaolinite specimens with initial concentration of 99-114 mg/kg. Other laboratory studies reported by Runnels and Larson (1986), Lageman et al. (1989), Eykholt (1992), and Acar et al.(1993) further substantiate the applicability of the technique to a wide range of heavy metals in soils. The process can potentially remove radionuclides from clayey soil samples (Ugaz et al., 1994). Bench-scale tests displayed that uranium at 1,000 pCi/g of activity is efficiently removed from kaolinite. A yellow uranium hydroxide precipitate was found in sections close to cathode. Enhanced electrokinetic processing showed that 0.05M acetic acid was enough to neutralize the cathode reaction and overcome uranium precipitation in the soil. Other radionuclides such as thorium and radium showed limited removal (Acar et al., 1992). In the case of thorium, it was postulated that precipitation of these radionuclides at their hydroxide solubility limits at the cathode region formed a gel that prevented their transport and extraction. Limited removal of radium is believed to be either due to precipitation of radium sulfate or because radium strongly binds to the soil minerals causing its immobilization (Acar et al., 1992). Lageman et al. (1989) showed that the process can migrate a mixture of different contaminants in soil simultaneously. Lageman (1993) reported 73% removal of Pb at a concentration of 9000 mg/kg from fine argillaceous sand, 90% removal of As at 300 mg/kg from clay and varying removal rates ranging between 50% to 91% of Cr, Ni, Pb, Hg, Cu, and Zn from fine argillaceous sand. Cd, Cu, Pb, Ni, Zn, Cr, Hg, and As at concentrations of 10 to 173 mg/kg also were removed from a river sludge at efficiencies of 50 to 71%. The energy expenditures ranged between 60 to 220 kWh/m3 of soil processed. Therefore, one can conclude that the type of contaminant does not pose a significant limitation on the technology provided it does not exist in an immobile form, e.g., sorbed on the soil particle surface or precipitated in the soil pore.

36

Regarding contaminant concentrations, existing experimental data indicate that removal of Cu(II) of concentration up to 10,000 mg/kg of soil and Pb(II) concentration up to 5,000 mg/kg are possible. Acar and Alshawabkeh (1996) demonstrated extraction of lead at 5300 mg/kg from pilot-scale kaolinite samples. However, high ion concentrations in the pore fluid increase the electrical conductivity of the soil and thus reduce the efficiency of electroosmotic fluid flow (Gray and Mitchell 1967; and Lockhart 1983). Moreover, the strength of electric field applied may have to be reduced to prevent excessive power consumption and heat generation during the process. The results of Hamed (1990), Pamukcu and Wittle (1992) and Wittle and Pamukcu (1993) demonstrated that lower initial concentrations of cadmium result in higher electro-osmotic efficiency; however, removal efficiencies were higher for samples with higher initial concentrations. Alshawabkeh et al. 1997 investigated electrokinetic extraction of heavy metals from clay samples retrieved from a contaminated Army Ammunition site. The soil contained cations at the following concentrations: calcium: 19,670 mg/kg; iron: 11,840 mg/kg; copper: 10,940 mg/kg; chromium: 9,930 mg/kg; zinc: 6,330 mg/kg; and lead: 1,990 mg/kg. The high calcium concentration hindered extraction of the metals. However, the results showed that metals with higher initial concentration, less sorption affinities, higher solubilities, higher ionic mobilities are transported and extracted faster than other metals. Electrokinetic remediation is also effective for the removal of organic pollutants such as phenol, gasoline hydrocarbons, and TCE from contaminated soils. Successful application of the process has been demonstrated for extraction of the BTEX (benzene, toluene, ethylene and m-xylene) compounds and trichloroethylene from kaolinite specimens at concentrations below the solubility limit of these compounds (Bruell et al., 1992; Segal and Bruell, 1992). High degrees of removal of phenol and acetic acid (up to 94%) were also achieved by the process (Shapiro et al., 1989; Shapiro and Probstein 1993). Acar et al. (1992) reported removal of phenol from saturated kaolinite by the technique. Two pore volumes were sufficient to remove 85% to 95% of phenol at an energy expenditure of 19 to 39 kWh/m3. Wittle and Pamukcu (1993) investigated the feasibility of removal of organics from different synthetic soil types. Tests were conducted on kaolinite, Na-montmorillonite, and sand samples mixed with different organics. Their results showed the transport and migration of acetic acid and acetone towards the cathode. Samples mixed with hexachlorobenzene and phenol are reported to show accumulation at the center of each samples. The results of some of these experiments were inconclusive, either because contaminant concentrations were below detection limits or because the samples were processed for only 24 hrs, which might not be sufficient to demonstrate any feasibility in electrokinetic soil remediation. Recently, the Department of Energy (DOE), Environmental Protection Agency (EPA), Monsanto, General Electric, and Dupont have also applied electric fields for electroosmotic extraction using layered horizontal electrodes in what is called the "Lasagna" process. Ho et al. (1997) reported successful extraction of TCE from a site in Paducah, Kentucky using the LasagnaJ process. They also reported 98% removal efficiency of p-nitrophenol, as a model organic compound, from soil in a pilot-study. Although removal of free phase non-polar organics is questionable, Mitchell (1991) stated that this could be possible if they would be present as small bubbles (emulsions) that could be swept along with the water moving by electroosmosis. Acar et al. (1993) stated that unenhanced electrokinetic remediation of kaolinite samples loaded up to 1,000 mg/kg hexachlorobutadiene has been unsuccessful. However, Acar et al. (1993) reported that hexachlorobutadiene transport was detected only when surfactants were used.

37

Electrolyte Enhancement Several procedures have been proposed to enhance electrokinetic remediation of heavy metals and radionuclides. Some of these procedures attempt to control production of hydroxyl ions at the cathode. Other procedures attempt to enhance complexation of heavy metals to enhance extraction at the anode. Some of these procedures are presented. Catholyte neutralization One way on controlling the catholyte pH is to neutralize the hydroxyl ions produced by electrolysis using weak acids or catholyte rinsing. The advantages of using weak acids include (a) they form soluble metal salts; (b) their low solubility and migration rates will not increase the electric conductivity of the soil; and (c) they are biodegradable and, if properly selected, environmentally safe. However, improper selection of some acids may pose a health hazard. For example, the use of hydrochloric acid may pose a health hazard because: (a) it may increase the chloride concentration in the groundwater; (b) it may promote the formation of some insoluble chloride salts, e.g., lead chloride; and (c) if it reaches the anode compartment, chlorine gas will be generated by electrolysis. Rødsand et al. (1995) and Puppala et al. (1997) demonstrated that neutralization of the cathode reaction by acetic acid can enhance electrokinetic extraction of lead. Hicks and Tondorf (1994) indicated that development of a pH front could cause isoelectric focusing, which retards ions transport under electric fields. They showed that this problem can be prevented simply by rinsing away the hydroxyl ions generated at the cathode. They demonstrated 95% zinc removal from kaolinite samples by using the catholyte rinsing procedure. Ion-selective membranes Another procedure to control hydroxyl ions and enhance metals transport toward the cathode is the use of membranes. Ion selective membranes, which are impermeable to hydroxyl ions, could be used to separate the catholyte from the soil and thus prevent or minimize the transport of hydroxyl ions into the soil. These membranes are insoluble in most solvents and chemically resistant to strong oxidizing agents and strong bases. Rødsand et al. (1995) and Puppala et al. (1997) showed that this technique has limited success when compared to catholyte neutralization. The reason is that heavy metals accumulate and precipitate on these membranes resulting in a significant increase in the electrical resistivity of membrane. Unless these membranes are continuously rinsed and cleaned, the energy cost of this technique will substantially increase. Chelating or complexing agents An acid front may not develop in soils of high buffer capacity and in soils that produce reverse electroosmosis, i.e., from the cathode toward the anode (Yeung et al., 1996). Reddy et al. (1997) showed that soils that contain high carbonate buffers, such as glacial till, hinder the development and advance of the acid front. On the other hand, acid advance in soils with low buffering capacity may cause uncontrolled dissolution of soil minerals resulting in an excessive release of some of their constituents, such as Al and Si. Under these circumstances, it is necessary to use enhancement agents to solubilize the contaminants without acidification. Chelating or complexing agents, such as citric acid and EDTA, have been demonstrated to be feasible for the extraction of different types 38

of metal contaminants from fine-grained soils. The enhancement agents should form charged soluble complexes with the metal contaminants. Cox et al. (1996) demonstrated the feasibility of using iodine/iodide lixivant to remediate mercury-contaminated soil. The use of EDTA as an enhancement agent has also been demonstrated for the removal of lead from kaolinite (Yeung et al., 1996) and lead from sand (Wong et al., 1997). Enhancement of anolyte pH While acidification of the soil causes dissolution of the soil minerals, it also increases the ionic strength and electric conductivity of the soil, which may hinder contaminant transport (Acar and Alshawabkeh 1993). If hydrogen ion generation and supply at the anode is not controlled, most of the energy may be consumed by generation and migration of this ion between the electrodes rather than the transport of charged contaminants. Therefore, it may be necessary to neutralize the anode reaction and/or control acid production and introduction into the soil mass.

Voltage and current levels Electric current intensities used in most reported studies are in the order of a few Amps per square meter. Although high current levels generate more acid that will work for the process, it increases the total ionic concentration that will decrease the overall electroosmotic flow. Selection of the most appropriate current density and voltage gradients depends on the soil electrochemical properties, especially electric conductivity. Soils with higher electric conductivities require more charge and higher currents than lower conductivity soils. A voltage gradient in the order of 100 V/m can be used as an initial estimate for initial processing. Increasing the current densities (or voltage gradients) will increase transport rates under ionic migration. However, increasing current densities will increase energy expenditure and cost of the process. An optimum current density or voltage gradient could be appropriately selected based on the soil properties, electrode spacing, and time requirements of the process. A procedure is provided for selection of appropriate current densities.

Electrode Requirements The number of electrodes required for 1-D applications depends upon spacing between electrodes of the same-polarity (e.g. anode-anode or cathode-cathode spacing). Decreasing spacing between same-polarity electrodes minimizes the area of inactive electric field, but increases the cost of the process. The same situation applies for 2-D configurations. In general, the goal of 2-D applications of electric fields is to achieve axi-symmetrical (or radial) flow towards a center electrode. The derivations will focus on extraction of positively charged heavy metals, thus locating the cathode as the center electrode allows accumulation of the cationic contaminants in a smaller zone around the cathode. Outer electrodes (anodes) are placed at specific distances from the center cathode to achieve relatively radial flow. The electrodes can be placed in a hexagonal or square configuration. Hexagonal (honeycomb) electrode configuration consists of cells, each contains a cathode surrounded by six anodes. The square configuration consists of a cathode and four (or possibly eight) anodes surrounding the cathode. Hexagonal and square grids generate two-dimensional, non-linear electric fields. Areas of inactive electric fields will develop depending on the configuration selected. Since this impacts the cost of electrodes, it will be necessary to select the configuration with the 39

optimum number of electrodes per unit area while minimizing the area of the ineffective electric fields. Table 3.1 provides a comparison of number of electrodes required per unit surface area for 1-D, 2-D hexagonal, and 2-D square configurations. Three cases are provided for 1-D arrangement where the same-polarity electrode spacing equals (a) the anode-cathode spacing, (b) one half the anode-cathode spacing, and (c) one third the anode-cathode spacing. Number of electrodes is calculated based on a unit surface (plane) area. For each configuration, the number of electrodes per unit area is calculated considering a unit cell. Accordingly, the number of electrodes required per unit plane area of site can be given by,

N

'

[

F1 2 LE

]1D flow ' [

F1

] 2 2D flow

RE

where N (L-2) is the number of electrodes per unit surface area of site to be treated, LE (L) and RE (L) are 1-D and 2-D anode-cathode spacing, respectively, and F1 (dimensionless) is a shape factor depending on electrode configuration. F1 is calculated by adding the number of electrodes serving a unit area (or a unit cell). If an electrode is serving more than one cell, then a fraction of the electrode is serving each cell. Table 1 shows the values of F1 for selected configurations. The results show that 1-D configurations with same-polarity electrode spacings of one half and one third anodecathode spacings require 100% and 200% increase in number of electrodes, respectively, when compared to the 1-D case of equal electrode spacings. Hexagonal configuration requires 15% increase in number of electrodes when compared to 2-D square configuration. Electrode requirements clearly affect the uniformity of electric fields and development of ineffective areas. Thus, it is necessary to evaluate the impact of increasing the electrodes on the development of ineffective spots.

Electric Field Distribution Numerical simulations of contaminant transport under electric fields use nonlinear, 1-D electric-field distributions for predicting species transport (Alshawabkeh and Acar, 1992; Shapiro and Probstein, 1993; Alshawabkeh and Acar, 1996, Jacobs et al., 1996). Similar numerical procedure can be used for 2-D applications. One also could assume uniform steady-state conditions and use the Laplace equation to describe the 2-D electric-field distribution. Analytical solutions, numerical methods, or conformal mapping can be used for solving the Laplace equation. However, these solutions do not provide a mechanism for comparing the effectiveness of different configurations. An approximate and practical approach is adopted in this paper for evaluating the area of ineffective electric fields. Electric field distributions show that the ineffective area for each cell has the shape of a curvilinear triangle with the base being the distance between electrodes of the same polarity (Figure 3.1). The height of this triangular area is approximate and depends on processing time, electrode spacings and alignment. The height of this triangle is expected to be larger in the case of 1-D compared to 2-D applications due to electrode alignment. This height is assumed as half the length of the triangle base for 1-D applications and a quarter the length of the base for 2-D applications. This assumption 40

provides a practical method for comparing the efficiency of different configurations. Figure 3.1 shows approximate distributions of the resulting inactive spots for 1-D (Figure 3.1a) and 2-D (Figure 3.1b) configurations. Approximate calculations of the percentage of ineffective area for each configuration are summarized in Table 1. For the 1-D configurations, the ineffective area is half the total area when the same-polarity electrode spacing equals the anode-cathode spacing. Thus, it is not practical to use such a scheme unless remediation is implemented in two stages where electrode polarity can be changed. This is also the case for the 2-D square arrangement.

Electrode configuration and time requirements 1-D Transport Electric currents could be applied in soils to generate one-dimensional (1D) or two dimensional electric fields. 1D electric fields could be generated by electrode sheets with specific spacing between the anodes and cathodes. Using sheets requires trenches and is not expected to be costeffective in most cases. Placing electrodes in bore holes is expected to provide an optimum and cost effective method for in situ application of electric fields. However, problems associated with bore hole configuration include development of inactive (dead) electric field spots between the anodes and between the cathodes. Several factors will affect electrode configuration, spacing and time requirements for electrokinetic remediation. These factors include (a) location and size of any inactive electric field spots, (b) number and costs of electrodes per unit area to be treated, and (c) time requirements of the designed remediation process. Large electrode spacings reduce the number of boreholes and installation costs, but increase the processing time and operation costs. It will be necessary to assess these variables prior to selecting the required configuration and spacing. In general, the processing time required is a function of the rate of transport and electrode spacing. As electroosmotic advection and ionic migration are the prominent transport mechanisms, hydrodynamic dispersion and retardation can be neglected in preliminary analysis. Accordingly, rate of species transport under an electric field is given by, v = ( u* + k e )

(-

)

where v = rate of species transport (or velocity) assuming the soil is a homogeneous medium (m/s); u* = effective ionic mobility of the ion (m2/V-s); ke = coefficient of electroosmotic conductivity (m29V DQG LVWKHHOHFWULFSRWHQWLDO9ROW ,IWKHVSDFLQJEHWZHHQHOHFWURGHVRIRSSRVLWHSRODULW\ is chosen to be L, the time (T) required for remediation can be estimated by dividing the spacing (L) over the velocity of species transport (v), i.e., T =

L ( u + k e ) (*

)

Where T is the time required for clean-up. However, this is a simplified estimation of time requirements where the contaminant of interest is assumed to be readily available for transport in 41

the soil pore fluid. This is probably the exception rather than the rule in real-life field implementation of the technology. Heavy metals are usually either sorbed on the soil particle surface or precipitated in the soil pore. Therefore, their transport is retarded by sorption and precipitation. A delaying factor similar to the retardation factor in advection-dispersion contaminant transport can be introduced to account for the extra time required for acid transport, metal desorption and dissolution, etc. Therefor, total time requirements should account for this retardation and can then be modified to T =

Rd L ( u + k e ) (*

)

where Rd = delaying factor (dimensionless). The value of Rd depends on soil type, pH, and type of contaminant. Sorption retardation factor can be used as an initial estimate of Rd and it equals unity for non-reactive contaminants which are readily available for transport. If enhancement agents are used to solubilize heavy metals, this factor should be modified accordingly. Alshawabkeh et al.  GHILQHGDQHZSDUDPHWHU WRFDOFXODWHWKHUHDFWLYHWUDQVSRUWUDWHRIDVSHFLHVUHODWLYHWRWKH electric conductivity of a medium, i.e. *

=

u + ke * Rd

:KHUH LVWKHHIIHFWLYHHOHFWULFFRQGXFWLYLW\RIWKHVRLOPHGLXP$FFRUGLQJO\WKHWLPH required for remediation in 1D applications is given by,

T =

1 *

L (-

)

This equation indicates that for a given electric field strength, the time required for remediation is linearly related to the spacing between the electrodes. If time is to be calculated using current

TR

'

1

LE Id

(1)

density, then we get, where Id (Amps L-2) is the electric current density. 2-D Radial Transport The approach provided below for 2-D applications assumes radial electric fields distribution (Figure 3.2). Rw represents the radius of center well (assumed cathode), RE represents the cathode-anode spacing, and Z is the depth of the site. The difference between this case and the 1-D case is that the current density in radial flow is a function of the radial distance (r); however, in both cases the total 42

current is constant. The electric current per unit depth for the radial transport is given by, Iz ' (2 r) ( ir

(1)

where Iz (amp L-1) is the current per unit depth and ir (VL-1) is the radial voltage gradient. Contaminant transport rate is dependent upon the voltage gradient, which is a nonlinear function of the radial distance. Ignoring diffusion and accounting for ion migration and electroosmosis, the radial velocity of ions transport is given by

v(r) '

( u ( % ke ) Rt

ir

(1)

where v(r) ( L T-1) is the velocity of species transport. Substituting ir leads to, IZ

v(r) '

(1)

2 r

The velocity of contaminant transport is a nonlinear function of the radial distance (even if the soil is homogeneous and isotropic). The time required for the contaminants to transport from the outside anodes to the center cathode is calculated by integrating dt = dr / v(r) from RW to RE, leading to 2

TR '

2

(RE & RW) IZ

(1)

Since (RW)2<< (RE)2, this equation can be simplified as, 2

T'

RE IZ

(1)

In order to provide time evaluation as a function of the voltage, a modified variable (rN) is introduced to simplify the formulation,

43

r ) ' ln

r RW

dr ) '

dr r

(1)

The transformed voltage gradient with respect to rN (irN) is independent of rN, and is given by, ir '

Iz

)

2

(1)

(

Substituting the value of Iz from,

TR '

1

2

RE

(1)

( i) r

The form of this equation for radial transport is similar to the equation of for 1-D transport. In both cases, $ and F* are soil properties and the gradients ie and irN are constants. However, comparing the two equations shows that while T is a function of the linear distance between the electrodes for 1-D case, it is a function of square the radial distance for 2-D applications. This is important for selection of electrode spacings. Selection of radial spacing for 2-D radial cases is much more critical than for 1-D cases as time and cost of remediation will significantly increase with increased radial spacing.

Energy expenditure Several factors impact energy requirements and cost for electrokinetic remediation at a specific site. These factors include soil properties, contaminant properties, electrode configuration and processing time. Energy consumption changes during processing due to changes in electric conductivity. However, energy calculations could by simplified by averaging soil electrical conductivity throughout the process. Accordingly, energy expenditure per unit volume of

W

'

(

max I d T R

L

)1D&flow ' (

max IZ TR 2 2 (RE&RW)

)radial&flow

(1)

contaminated soil is given by the following equation for both 1-D and radial applications where W (J L-3) is energy expenditure per unit volume of soil and Nmax (V) is the applied voltage. Substituting the equations for TR results in the following equation for both 1-D and radial cases,

44

'

W

max

(1)

This equation shows that same energy expenditure is expected for both 1-D and radial applications, assuming that same total voltage is applied and each case is processed for specific required time. This indicates that energy requirement could be considered independent of electrode configuration if energy source (maximum voltage) is the controlling factor. In other words, two 1-D schemes (for one specific site) with different spacings should result in same energy expenditure if same total voltage was used in both schemes. The difference between the two schemes would be in terms of time requirements. However, electrode configuration is a design factor if the energy source is not the limiting factor.

Cost The total costs for full-scale in situ implementation of electrokinetic remediation can be divided into five major components (Alshawabkeh et al., 1999): (1) cost of electric energy,(2) cost for fabrication and installation of electrodes, (3) cost of enhancement agents, if necessary, (4) costs of any posttreatment, if necessary, and (5) fixed costs. Impacts of electrode configuration and spacing on these cost components are addressed separately. Electric energy cost Based on energy expenditure evaluation, the following energy cost equation provided by

Cenergy

'

C2

max

3,600,000

(1)

Alshawabkeh et al. (1999) for 1-D conditions is also valid for 2-D, radial conditions, where Cenergy ($ L-3) is electric energy cost per unit volume of soil treated, and C2 ($/kWh) is electric energy cost. Costs for fabrication and installation of electrodes If acid production at the anode is not controlled, then inert electrodes, such as graphite or coated titanium, should be used to prevent dissolution of the electrode and generation of undesirable corrosion products in an acidic environment. If necessary, sacrificial electrodes can also be used as anode. Any conductive materials that do not corrode in a neutral or basic environment can be used as cathode. Important considerations for the choice of electrode material are (Alshawabkeh et al., 1999): (1) electrical conduction properties of the material, (2) availability of the material, (3) ease of fabrication to the form required for the process, (4) ease of installation in the field, and (5) 45

material, fabrication, and installation costs. The electrodes can be installed horizontally or vertically. The costs of each electrode depend on the material used, complexity of installation, and dimensions. The number of electrodes per unit volume of soil to be treated depends on electrode configuration and spacing. The installation costs depend on the method of installation, depth of the electrodes to be installed, and number of electrodes to be installed. The total electrode costs per unit volume of soil to be treated can be calculated by evaluating the number of electrodes per unit cell of an area, i.e,

'

Celectrode

C1 N

(1)

where Celectrode ($ L-3) is electrode costs per unit soil volume, N ( L-2) is the number of electrodes per unit surface area, and C1 ($ L-1) is the cost of electrodes per unit length. Values for N are provided in Table 3.1 for different configurations. Increasing electrode spacings decreases the value of N and hence decreases total electrode costs. For the 1-D case, spacings between electrodes of same polarities significantly impact this cost component. Cost for enhancement agent Enhancement agents and chemicals are used to improve the efficiency of electrokinetic remediation Chemicals cost is a significant component of the total cost of the processes. Chemicals are used for either neutralizing pH conditions, or enhancing solubility of target contaminants, or both. The cost of chemicals required for pH neutralizing is considered in this evaluation. This cost is dependent upon the electric current and is given by the following equations, Cn&chemical

'

'

C3

Cn&chemical

C3

Id MW TR L F

Iz

MW TR 2 2 F (RE&RW)

for 1&D

for Radial

(1)

(2)

where Cn-chemical ($ L-3) is the cost of chemicals required to neutralize electrolytes per unit soil volume, C3 ($ M-1) is the cost of the chemical agent, MW (MM-1) is the molecular weight of the neutralizing chemical, " (dimensionless) is a factor depending upon the stoichiometry of the neutralizing reaction, and F is Faraday’s constant (96,485 C/mol-electron). Substituting the time required for remediation in both 1-D and radial cases will result in the following equation;

46

Cn&chemical

'

C3 MW F

(1)

Accordingly, the chemicals cost is independent of electric current or spacing and is dependent on soil characteristics. This is due to the fact that electric current and electrode spacings impact time requirements. For example, increasing the current will decrease the time required for remediation, such that the same total charge is introduced for any electric current value.

Cost of post-treatment Post treatment costs should also be considered if effluent treatment is required. These costs are highly site and contaminant specific. An estimate of effluent treatment costs could be evaluated per unit volume of the soil for 1D case as follows,

k e  T Cpost-treat = C4 (L/n) where Cpost-treat is the post treatment cost per unit volume of the soil ($ L-3) and C4 is the cost of treatment per unit volume of the electrolyte (effluent) collected ($ L-3). Substituting for the value of T (time required for remediation), then effluent treatment cost is given by;

Cpost- treat = C4

n ke *

= C4

n R d ke * u + ke

Derivations for radial transport post treatment costs also result in similar equation. Volume and cost of effluent treatment depends on the ratio of transport under electroosmosis relative to total transport rate. In order to minimize the volume collected, it is necessary to maximize transport by ionic migration and minimize transport by electroosmosis. If contaminant transport occurs only due to migration then this cost component will be zero and one needs only to treat electrolyte in electrode well. However, if electroosmosis is the only mechanism used for contaminant transport (e.g., for non charged contaminants) then cost of treatment will be equal to (C4 n Rd) which, indicates that the cost depends upon the number of pore volumes required for remediation. If the contaminant is readily available for transport, then Rd =1 and one pore volume is enough for remediation. However, if extraction is retarded due to geochemical reactions then it is obvious that the pore volumes required will increase depending upon the value of Rd. Sometimes catholyte recycling is used which will add another component that should be considered for evaluation of total volume of water collected.

47

Other fixed and variable costs Other costs for full-scale implementation include mobilization and demobilization costs of various equipment, site preparation, security, progress monitoring, insurance, labor, contingency, and miscellaneous expenses. The equipment will not be consumed in a particular project. However, there are capital, depreciation, or rental costs involved. These cost components will be divided into fixed (e.g. mobilization and demobilization) and variable (e.g. monitoring, insurance, rentals) components. Variable costs are simply evaluated by multiplying the cost rate by the total time required for remediation, i.e.

Cvariable

Cvariable '

C5

'

L (i

C5

for 1&D

(1)

e

2

RE

( i)

for radial

(1)

r

where Cvariable ($ L-3) is the total variable cost per unit soil volume, and C5 ($ L-3 T-1) is the variable cost rate per unit soil volume. C5 is evaluated by estimating the variable daily cost (for monitoring, insurance, rentals, ... etc) and dividing by the total volume of site. C5 is highly dependent upon the size of the site and decreases as volume of contaminated soil increases. Total costs The total costs per unit volume of soil to be treated are thus given by

Ctotal = Celectrode + Cenergy + Cchemical + Cpost- treat + Cfixed + Cvariable where Ctotal = total costs per unit volume of soil to be treated ($ L-3); and Cfixed = fixed costs per unit volume of soil to be treated ($ L-3). Cost evaluation indicates that electrode configuration will impact electrode, energy, and variable costs. Other costs (chemicals, fixed, and posttreatment) independent on electrode configuration and spacing.

Optimum Electrode Spacing 1D Applications Assuming post-treatment, chemicals, and fixed costs are independent of electrode spacing, the optimum electrode spacing can be obtained for 1D applications by equating the partial derivative of Ctotal with respect to L to zero, which renders the following equation, 48

3 L optimum =

* 7,200,000 C1 F1 3,600,000 C5 + C2 (∇ )2

*

where Loptimum = optimum electrode spacing (L). This equation provides an estimate for the optimum electrode spacing that minimizes the total costs of 1D applications as a function of the properties of the contaminated soil and electric field strength. The optimum spacing is also dependent upon electrode costs (C1), energy costs (C2), variable costs (C5) and electrode configuration (F1). This equation demonstrates the impact of the cost ratios C2/C1 (energy cost / electrodes cost) and C5/C1 (variable cost / electrodes cost) on optimum electrode spacing. In some cases, time requirements might be the limiting factor where remediation needs to be FRPSOHWHGZLWKLQDVSHFLILFWLPHSHULRG)RUVXFKFDVHVRQHFDQVXEVWLWXWHWKHYDOXHVRI VRWKDW the optimum electrode spacing could be evaluated based on time requirement by the following equation for 1D applications, 7,200,000 2 C1 F1 * T 2 = L optimum 2 3,600,000 * 2 T 2 C5 + C2 Loptimum Thus, the procedure for design of field implementation is to decide whether the limiting factor is time or energy. Radial Applications Assuming post-treatment, chemicals, and fixed costs are independent of electrode spacing, the optimum electrode spacing can be obtained for 2-D applications by equating the partial derivative of Ctotal with respect to RE to zero, which renders the following equation,

RE2&opt

'

7,200,000 3,600,000

C1 F1 ir

C5 RE2&opt %

)

(

C2 ir2 (

(1)

)

This equation provides an estimate for the optimum electrode spacing that minimizes the total costs of 2-D applications as a function of the properties of the contaminated soil and electric field strength. The optimum spacing is also dependent upon electrode costs (C1), energy costs (C2), variable costs (C5) and electrode configuration (F1). The optimum spacing equation provides a formulation that demonstrates the impact of the cost ratios C2/C1 (energy cost / electrodes cost) and C5/C1 (variable cost / electrodes cost) on optimum electrode spacings. If time is the limiting factor, then substituting the value of irN yields the following equation,

49

RE2&opt

'

2

7,200,000 3,600,000

(

2

C1 F1 ( TR

T 2 C5 %

C2 RE2&opt

(1)

Thus, the procedure for design of field implementation is to decide whether the limiting factor is time or energy.

Example Cost Evaluations The following example is considered to show the role of each cost component. In this example, a contaminated area of 30m × 60m is assumed. The depth of contamination is assumed to be 3 m. The soil is a saturated, silty clay. The porosity, tortuosity factor, electrical conductivity, and coefficient of electroosmotic permeability of the soil are determined from preliminary laboratory analyses to be 0.4, 0.3, 0.02 S/m, and 2×10-9 m2/V-s, respectively. The ionic mobilities of target contaminants are taken to be 5×10-8 m2/V-s. The value of Rt is assumed to be 4 and the value of $ is thus calculated to be 1 ×10-7 m3/C. The mobilization cost of a drilling rig and the labor cost of a two-man operating crew are taken to be $1000 per day to evaluate C1 (Alshawabkeh et al., 1999). For boreholes to be drilled without installation of casing and sampling, a continuous flight auger can achieve approximately 65 m a day. Therefore, the drilling cost is estimated to be $15 per linear meter. Costs for fabrication and installation of electrodes are approximately $5 per linear meter as the electrodes are reusable. Therefore, C1 is taken to $20 per linear meter. The electricity cost C2 is assumed $0.04 per kWh. Cost of enhancement reagent (molecular weight assumed 100 g/mole) is taken as $ 2 / kg. Variable cost rate is calculated based on two-man crew, 14 hours per week, at a rate of 25 $/ man-hr. Accordingly, the variable rate will be around 0.001 $/m3-hr, including 30% increase for insurance. Total cost of other components, including post treatment and fixed costs, is taken as $25 / m3. Two configurations will be considered: (a) 1-D, where the same-polarity electrode spacing equals one-third the anode-cathode spacing and (b) 2-D hexagonal configuration. In both cases F1 = 3. Impact of electrode spacing on electrode, energy and variable costs is demonstrated in Figure 3.3 for the 1-D configuration and in Figure 3.4 for the hexagonal configuration. In these figures, a 1-D electric gradient ie of 100 V/m and a transformed radial electric gradient irN of 100 volt are assumed for cost evaluations. For the two schemes provided, increasing electrode spacing showed a significant decrease in electrode costs. For 1-D case, increasing electrode spacing from 1 m to 2 m resulted in a more than 80% decrease (from $60 /m3 to about $10 /m3) in electrode costs. However, increasing electrode spacing beyond 2 m did not result in a comparable decrease in electrodes cost. Similar behavior is noted in the hexagonal case. Increasing electrode spacing results in a linear increase in energy and variable costs in the 1-D case. As indicated earlier, increasing electrode spacing will increase the total voltage applied (assuming constant voltage gradient), increase energy expenditure and cost. Similarly, increasing electrode 50

spacing increases time requirements and, consequently, variable costs. For the hexagonal case, the increase of energy and variable costs is a nonlinear function of electrode spacing, as shown in Figure 3.4. Impact of electrode spacing on total costs and time requirements for both the 1-D and hexagonal cases is displayed in Figures 3.5 and 3.6, respectively. Increasing electrode spacing has more impact on the time requirements for the hexagonal case when compared to the 1-D case. This is again because time requirements are related to square of the radial electrode spacing. Accordingly, selection of the optimum electrode spacing is more critical for 2-D applications when compared to 1-D applications. A minimum exists in the cost versus electrode spacings relationship. It is necessary to provide a mechanism for selecting the optimum electrode spacing that will minimize cost, time, or both.

51

Table 3.1. Impact of configuration on electrode requirements and size of ineffective areas Config.

Electrode spacing

Opp. Charge

No. of Electrodes per cell (F1)

Area of cell

No. of electrodes per unit area

Ineffective Area

(Acell)

Same

N

% increase

Aineff

% of

charge

Acell

1

LE2

1/LE2

0

LE2/2

50%

1-D

LE

LE

1-D

LE

LE/2

2

LE2

2/LE2

100%

LE2/4

25%

1-D

LE

LE/3

3

LE2

3/LE2

200%

LE2/6

17 %

Square

RE

¥5E

2

2 RE2

1/R2

0

RE 2

50%

Hex.

RE

RE

3

3(¥ 5E2/2

¥ 5E2

15.5%

3RE2/4

29%

52

LE 6

LE 4

LE 2

LE LE 2

LE 4

LE 6

LE LE 2

Area of Cell = L2E

Area of Cell = L2E Ineffective Area =

LE 3

LE 2

2 E

L 6

Ineffective Area =

LE 3

Area of Cell = L2E 2 E

L 4

Ineffective Area =

(a)

2R E 4

RE

RE RE

2R E

2R E

Area of Cell = 2R

LE 3

RE 4

3 3R 2E 2 3R 2E Ineffective Area = 4 Area of Cell =

2 E

Ineffective Area = R 2E

(b)

Area of effective electric field

Cathode

Area of ineffective electric field

Anode

Figure 1. Approximate evaluation of ineffective areas for (a) 1-D and (b) 2-D electrode configurations.

53

L2E 6

r

A

A

RE

RW

DE

r Z

Section A-A

N max

N(r)

Anode (+)

Cathode (-)

Anode (+)

Figure 3.2 A schematic showing radial electric field distributions.

54

Figure 3.3

55

Figure 3.4

56

Figure 3.5

57

FIGURE 3.6

58

SECTION 4 FIELD DEMONSTRATION

Field demonstration of electrokietic remediation is site specific and the design, implementation and performance varies from site to site. Successful field studies at various conditions are needed to support the hypothesis and laboratory experiments that claim the potential of the processes. The objective of this study is to evaluate the potential of electrokinetic remediation of chromium at representative field conditions. In particular, this section evaluates the effect of saline water on the efficiency of chromium extraction by EK under field conditions. The demonstration was conducted by the Environmental Laboratory of the Engineering Research and Development Center (ELERDC), US Army Corps of Engineers.

Description of Site, Equipment and EK Testing System The study is conducted at NAS Point Mugu, which is located in Ventura County, California approximately 50 miles northwest of Los Angeles. Established in 1944, the main base comprises approximately 4,500 acres. NAS Point Mugu is bordered by Highway 1 on the north and east, the Pacific Ocean on the south and west, and a Ventura County game reserve on the West and Northwest. The Old Area 6 shops are located along Beach Road just west of the south end of Laguna Road (Figure 4.1). Site 5 is a large area where electroplating and metal finishing operations disposed of their effluent. The waste disposal locations are less than one acre in size. They are the result of lab and shop waste disposal practices in the Old Area 6 shops between 1947 and 1978. The area of initial study is approximately ½ acre in and around two former waste lagoons located in the center of the site. The lagoons were unlined and were used between 1947 and 1978 to receive wastewater discharge from the electroplating and metal finishing activities. The largest waste generator in the area was plating shop, which reportedly disposed of up to 95 million gallons of plating rinsate between 1948 in 1965. In the late 1940's and early 1950's, the photo and rocket fuel chemical shops disposed of waste photo fixer and developer, organic solvents, chemical wastes, and rocket fuel into septic tanks which eventually emptied into Mugu Lagoon. In 1994, an emergency removal action was performed during which approximately 117 cubic yards of material were excavated from the waste lagoons. The removal action was executed to limit exposure of several residential and migratory birds and to reduce the source of contamination that could impact surface and groundwater. The initial EK test site consists of the two waste lagoons (pits) and its adjacent tidal marsh area. The original demonstration/validation (Dem/Val) plan was to address the heavy metals contamination in the lagoon area with an additional EK cleanup effort in the adjacent tidal marsh (Figure 4.2) thick by approximately 18-ft deep high density polyethylene (HDPE) barrier wall was installed surrounding the two waste lagoons. This area was given the designation of Test Cell #1 (TC1) while the adjacent salt water march area was designated Test Cell # 2 (TC2). The barrier was installed to limit tidal influences and groundwater flow into and out of the demonstration area. It would also prevent the movement of mobile species outside of the test cell treatment area in the event the EK 59

process could not control the movement of mobilized metals. Because of the presence of the lightfooted clapper rail, a federally- and state-listed endangered species, a 7-ft high frame covered with a dark mesh screen was installed to enclose the site. This enclosure served to shield the site from access by the bird. It also allowed free movement by personnel inside the enclosure for normal system operations without disturbing the bird. However, due to technical difficulties, particularly the effect of brackish water on conductivity and transport, the study was transitioned from a Dem/Val to a pilot-scale project. The transition and the pilot-study were operated by ERDC EL (WES) at the end of October 1998. ERDC EL (WES) restarted the EK system with a reduced electrode well field in January 1999. Figure 4.2 shows the location of the original Dem/Val study (Cell # 1) and the area for the pilot-study. This paper focuses on the results of the pilot study to demonstrate the potential for chromium and cadmium removal from the site by the application of EK technology. The main EK system components are: a process control trailer, power supplies, electrode wells, chemical storage tanks, and fluid control and distribution system. The electrokinetic remediation system consists of an array of electrodes, a power distribution and control system, automated process monitoring equipment, and process piping to distribute chemical amendments to the electrode wells and to extract contaminants from the electrode wells. Other equipment necessary to support system operations due to the site conditions at NAS Point Mugu included off-gas extraction and treatment equipment to treat the gases that were generated in the electrode wells during system operation. A pH control system was used at the cathode wells by regulating the addition of citric acid to the cathode wells. The pH control system contains several pumps and solenoids that are controlled and operated via the computer program. If there is a failure in the pH system, the system shuts itself down. Operation is suspended until a technician corrects the problem. Citric acid was selected because it is inexpensive and environmentally acceptable. The EK system operation required electricity (220-volt, 100 amp, 3-phase power), freshwater supply, and telephone lines to be installed at the site. The electrical supply was needed to provide power to energize the field of electrodes, pumps, and other control equipment. The fresh water was required for make up water for the electrode wells. The two telephone lines were for voice communication and system monitoring/remote control. The distance between the anode and cathode wells was 15 feet. The anode and cathode wells were 10 feet deep. The distance between like electrode wells was 5 feet. The electrodes were labeled Anode Well (AW#) and Cathode Well (CW#), depending on the electrode type and its location (Figure 4.2). ERDC EL (WES) operated the pilot-scale study at reduced electrode field in the north end of Pit 1 (electrodes AW9-11, AW18-20 and CW8-10). The anode electrode, which operate under highly corrosive conditions, consisted of 3 feet long, 1-in. diameter titanium hollow tubes (Eltech) with an iridium oxide coating. The cathode electrodes were constructed of 1/8 inch x 2-inch x 10-feet long stainless steel mesh. The anode and cathode wells were capped with PVC couplings for easy removal. The anode wells were sealed with a system attached to vent the gases produced during system operation (e.g., chlorine and oxygen,). Gas from the wells passed through the scrubber unit prior to release to the atmosphere. At first, the cathode wells were vented to the atmosphere, since only small volumes of hydrogen sulfide gas were expected to be produced. During the demonstration at NAS Point Mugu, hydrogen sulfide gas (H2S) was detected in the 60

cathode wells. Upon determination that H2S was being produced during system operation, the cathode wells were sealed and the gases were vented to a scrubber unit prior to release to the atmosphere. The wells were installed using standard well drilling and casing installation practices. The only notable exception was the extra care required installing the cathode well ceramic casings. These casings are brittle and easily broken. Electrical power is applied to the electrode array via three 10 kilowatt (kW) power supplies (EMI part number ESS 30-333). Each of these power supplies is capable of delivering up to 30 volts (V) at up to 333 amps (A). The power supplies were wired in series to deliver up to 90 V at up to 333 A. Application of the electric power to the electrode array was controlled by an on-site computer system with customized LabView 4.0 software. ERDC EL (WES) operated the system with constant voltage beginning in January 1999. The applied current fluctuated with the varying resistivity of the soil in the treatment area. The total power supply current as well as each individual electrode well current was logged by the computer system. The data were collected by computer data acquisition/control and manually recorded by hand. The data were stored in four separate file types in text from and are viewable with most spreadsheet software. The data from the field (electrode wells and other sensors) and the power supplies were collected approximately every 30 minutes to multiple computer files. Frequent and extensive monitoring of the electrokinetic remediation process was necessary. To conduct this monitoring, 18 two-inch diameter piezometer wells were installed in and around the test cells. A series of piezometer wells were also installed between AW10 and CW9 to determine transport profiles and pH front (acid front) development. The wells were screened at various depths and located in positions where contaminant transport or electric field effects would most likely occur if the electrokinetic process control could not be maintained. The installation data for the wells are shown in Figure 4.2. Sampling and chemical analysis were conducted by an independent certified laboratory. Soil cores were taken from 20 coring points to a depth of 12 feet in TC1 for baseline analysis. The sampling analysis results are presented in isopleths by depth. The baseline data presented in this section were extracted from the isopleths by creating a transparency grid overlay for the reduced treatment area at the north end of TC1 Pit 1. Metal analysis included Cr, Cd, Pb, Cu, and Zn. No significant concentrations of Pb, Cu, and Zn in the liquid or the soil were detected. The two major soil contaminants, chromium and cadmium, are presented here. Processing conditions are summarized in Table 4.1. The volume of the treated soil is estimated about 64 m3. Initially a total voltage of 60 volts was applied, but it was reduced to 45 volts after 20 days of processing. The reason was the difficulty in controlling the catholyte pH under 60 volt application. After 118 days of processing, the electrodes were shortened to 4 ft to facilitate extraction at shallow depths. Further discussion regarding the changes is provided in the results section.

61

Results and discussion pH profile and Control The piezometers were sampled before treatment and each month with the regulatory compliance sampling. Figure 4.3 shows the pH contours before processing (Figure 4.3a) and after 4 months of treatment (Figure 4.3b) across wells AW10 and CW9. The figure shows the development and advance of the acid front from the anode towards the cathode. Acidification seems to occur at lower depths (below 70 in.) and moves towards the cathode. The first few inches close to the cathode were acidified because of citric acid injection to control the cathode pH at 4. The upper middle section in the cathode side showed a relatively high pH (6.5 - 7.5). In the first 20 days, the pH in CW9 was difficult to maintain below 4.0 with electric power applied at 60 volts and 50 amps. The current level at this applied power resulted in the citric acid cathode amendment being consumed faster than it could be physically pumped into the cathode wells. To maintain a pH of 4.0 in CW9, the applied voltage was lowered to 45 V where it remained until the June suspension of treatment. Occasional pH spikes (pH 12) occurred when the computer failed to respond to the software or when the software is turned off (Figure 4.4). This action trips the failsafe relay circuit that allows a power supply to maintain 9 volts in the test area. Since no citric acid is being pumped into the cathode wells, the well pH increases. Maintaining electrical energy to the field was considered a safety precaution to prevent soluble chromium from migrating outside the confined test area. Chromium Extraction The results from the baseline soil characterization indicate that the majority of the chromium contaminant is located in the top sections of soil nearest the cathode (Figure 4.5a). The lowest chromium concentration (180 mg/kg) was located near the anode at a depth of 5-ft. This analysis shows that the entire soil volume in pilot treatment area was completely contaminated with chromium from the surface to a depth of 10 ft. The initial average contaminant mass of chromium in the treatment area was estimated to be 2,319 grams. The soil sampling results after treatment indicate that most of the chromium contamination has moved upward toward the cathode (Figure 4.5b). The chromium movement in the direction of the cathode was expected because positively charged trivalent chromium is attracted to the cathode. Figure 4.5b reveals that the chromium contaminant was not detected or was below the natural Point Mugu background levels (109 mg/kg) in the sections closest to the anode to over 8-ft toward the cathode. The total mass of chromium remaining in the soil was estimated to be 1,621 grams.

62

Extracted chromium was measured in the effluent storage tank and in the electrode wells. At the end of the treatment, the effluent storage tank held approximately 800 gallons of liquid collected from the anode wells plus some rinse water from the NAS water distribution system. NAS water was used to rinse out the well fluid from the field pumps, pipes, and valves. The effluent liquid contained 12 mg/L of chromium. This corresponds to a total of 33 grams of chromium removed from the site. Figure 4.6 illustrates the cumulative total chromium concentration with time from the cathode (Figure 4.6a) and the anode wells (Figure 4.6b). These wells were sampled monthly and quarterly. The highest concentrations of chromium were discovered in the anode wells not in the cathode wells. This indicates that a fraction chromium may have existed in negatively charged Cr(VI) oxyanions. The slope of the chromium concentration lines was flat after the process was terminated. A total mass of 126 grams of chromium was moved from the soil into the electrode well liquid. After 4 weeks of treatment, the soil sampling analysis along the piezometer profile between AW10 and CW9 indicated that most of the chromium contamination had risen to shallower depths near the cathode. The chromium movement can be seen when comparing the pretreatment seen in Figure 4.5a to post treatment seen in Figure 4.5b. The upward movement can be attributed to the electrode modifications. The field data results reveal that approximately five percent of the chromium mass had been removed from the soil to the electrode wells. The soil sampling results indicate that 78 percent of the soil volume has been cleared of the chromium contaminant or treated below the Point Mugu natural background levels. The mass balance error or chromium removal was 24.7 percent. Cadmium Extraction The baseline characterization indicates that the majority of the cadmium contaminant was located in the top sections of soil evenly distributed between the anode and the cathode (Figure 4.7a). The minimum cadmium concentrations were 5 mg/kg, located at 5-ft and 10 ft depths. The maximum cadmium concentrations were approximately 20 mg/kg, near the surface. This analysis illustrates that the entire soil volume was contaminated with cadmium. The initial average cadmium mass in the treatment area was estimated to be 70.9 grams. Soil sampling results after treatment indicates that most of the cadmium contamination has moved upward toward the cathode. Figure 4.7b reveals that cadmium was not detected in most of the sections closest to the anode to over 7 feet toward the cathode. The total calculated mass of cadmium remaining in the treatment area was estimated to be 36.8 grams. These results follow the same pattern as the chromium analysis. EK treatment is effective in removing the cadmium contaminant from the soil. However, additional testing is required to achieve the study goals of removing the cadmium contaminant from the soil at Site 5. In comparing the pretreated cadmium analysis in to post treatment cadmium analysis in Figure 4.7, the results show that most of the cadmium contamination had moved toward the cathode and was rising to shallower depths. The analysis revealed that beginning from AW10 to over 7 feet toward CW9 the soil was cleaned of cadmium contamination or treated below the natural background level. The data analysis from the six-month sampling revealed that approximately two percent of the cadmium mass had been moved from the soil to the electrode well liquid. The mass balance error 63

for cadmium removal was 46.3 percent. The analysis indicates that 70 percent of the soil between AW10 and CW9 had been cleared of cadmium contamination. The bulk of the remaining chromium and cadmium contaminants are located in the upper soil sections closest to the cathode wells. This data conclusively verifies the effectiveness of electrokinetic treatment at NAS Point Mugu Site 5. Electric Current and Energy Expended The current density reached a peak of 1.79 mA/cm2 on the 17th day of operation. At 60 volts, the current in AW-11 went to a high of 45 amps. This current level consumed the citric acid cathode amendment faster than it could be physically pumped into the wells. When the voltage was dropped to 45 volts, the total current density was 0.97 mA/cm2. Assuming a linear voltage gradient of 9.8 V/m, the apparent electric conductivity of the soil between the electrodes is about 1 S/m. This is a very high electric conductivity, which is reflective of the site’s saline water. The total cumulative energy expended was collected from two different sources, the electric energy meter to the site and the calculated energy expended at the electrodes (Figure 4.8). The readings from the electric meter represent the total energy supplied to the site while the calculated energy represents the energy expended at the electrodes to treat the contaminated soil. The cumulative energy from the electric power meter was 25,798 kWh (208 kWh/m3), and the reading from the data acquisition system was 24,978 kWh (200 kWh/m3). The difference between the meter and the data acquisition was 820 kWh. The energy applied to the electrode field was not as high as the total energy supplied to the site. The computer controlled hardware (pumps, sensors, solenoids, lights, and refrigerator) consumed energy that contributes to the total energy supplied to the site. To provide a conservative cost estimate of the energy expended, the higher cumulative energy number was employed. At a cost rate of $0.08 per kWhour, the total energy expended from January through June 1999 was $2064 ($16.5/m3). For comparison, the difference in cost between the two cumulative energy calculations was $66. Impact of Brackish Condition A critical component of this study is to assess the impact of brackish water on transport of contaminants. In general, the total current is carried by species in the solution depending upon their concentrations and mobilities. The total current can be divided into positive charge current and negative charge current: I = I+ + I-, where I+ is the fraction of the current carried by positively charged ions and I- is the fraction of the current carried by negatively charged ions. To simplify, and using conservation of electrical neutrality, I+ should be equal to I-. Which means that half the current is carried by the positively charged ions and half is carried by the negatively charged ions in solution. The positively charged ions carrying the charge include calcium, hydrogen, chromium, cadmium, lead, ... etc. The negatively charge species carrying the current include chloride, hydroxyl, chromium complexes, etc. Accordingly, a fraction of the current is used for chromium (or cadmium) transport. This fraction will depend upon chromium form, concentration and mobility relative to electric conductivity of the medium. Increasing specific ion concentration and mobility will increase its impact on current. However, increasing electric conductivity of the soil will minimize the amount of charge carried by this specific species. The high ionic strength and electric conductivity at Point Mugu significantly decreases the contribution of chromium transport in total electric current. To 64

overcome this problem, a relatively high current density (~ 10 A/m2) was used in this project. The high electric conductivity of the soil rendered a relatively small voltage gradient (~ 10 V/m). The outcome is a reasonable energy expenditure (~ 400 kWh/m3) and efficient extraction of chromium (Figure 4.5b) and cadmium (Figure 4.7b).

65

Table 4.1. Testing Conditions at Point Mugu Site Cathode:Anode Spacing: Same Electrode Spacing: Electrode length: Soil Volume: Initial Conditions Initial Cr: Initial Cd: Initial pH:

4.57 m (15 ft) 1.5 m (5 ft) 3 m (10 ft shortened to 4 ft after 118 days) 9.14 m x 4.57 m x 3 m = 125 m3 (30 ft x 15 ft x 10 ft)

180 to 1100 mg/kg 5 to 20 mg/kg 4 to 8

Electric Energy: (Constant voltage, varying current) Day 1: 60 volts Day 20: 45 volts Day 118: 45 Volts Electrode lengths were shortened to 4 ft.

66

“C” Av

Ox na rd

7th

St.

8th St.

Duck Pond

Draina ge

o. 3

11th St.

Di tc h

Main

Dr ain ag eD itc hN

7th St.

“F ”A ve N. . Mu gu Rd .

Runway 9-27

Ox na rd

Calle guas Cree k

1

e.

Duck Pond

ugh Slo

Peri me

ter R

d.

on vol Re

Ventura County Game Reserve

Pt. Bo Mug un da u ry

No .2 s La

d. sR sa Po

Santa Monica Mountains

Laguna Peak

Laguna

So .“ L” Av e

Rd.

13th St.

Mugu

Beac h Rd .

Lagoon Pt. Mugu State Park

Beach Rd.

Site

1

20th St.

Scale in Feet 0

2000

Pacific Ocean

4000

LEGEND Water

Drainage Ditch

Private/Public Lands

Base Property

LB&M ASSOCIATES, INC.

Source: Modified from Fugro-McClelland, 1991

Figure 4.1. NAS Point Mugu General Location Map

67

Figure 4.2. TC1 Electrode Well Field

68

Figure 4.3a. Pilot Study Initial pH Profile Results between AW10 and CW9

Figure 4.3b. Pilot Study pH Profile Results after Treatment

69

14

12

Cathode Well 9 (CW9) pH

10

8

CW9 pH 6

4

2

0 0

20

40

60

80

100

120

140

Time Since January 1999 Restart (Days)

Figure 4.4. CW9 pH changes during processing

70

160

180

15 feet (+)

(-)

200

240

260

300

300

340

370

400

650

800

850

900

900

950

1000

1100

185

190

190

190

200

215

230

245

260

285

300

320

340

360

380

400

180

185

190

190

195

200

215

230

250

275

300

325

350

375

400

410

550

560

580

600

600

600

600

600

600

575

550

525

500

490

485

475

200

210

210

200

200

200

200

205

205

205

210

210

215

215

220

220

10 feet Concentration > 109 mg/kg from 1997 Natural Pt. Mugu Background Levels Concentration < 109 mg/kg from 1997 Natural Pt. Mugu Background Levels

Figure 4.5a. Pilot Study Initial Site Characterization Chromium Results. 15 feet (+)

(-)

1

0

0

0

147

5

0

0

318

1351

1935

3222

5163

5052

8226

2072

0

0

90

0

25

0

0

0

236

122

340

510

216

278

344

2419

0

0

0

0

0

0

0

0

37

0

0

47

30

79

0

58

76

0

0

0

0

0

0

0

0

0

0

0

40

33

0

0

43

0

13

0

13

0

542

81

35

1

8

0

14

2

0

0

10 feet Concentration > 109 mg/kg from 1997 Natural Pt. Mugu Background Levels Concentration < 109 mg/kg from 1997 Natural Pt. Mugu Background Levels

Figure 4.5b. Pilot Study Soil Chromium Results after Treatment

71

Cathode Wells Cumulative Total Chromium Concentration (mg/L)

160

CW8

140

CW9 CW10 120

100

80

60

40

20

0 13-Jan-99

12-Feb-99 14-Mar-99

13-Apr-99 13-May-99 12-Jun-99

12-Jul-99

11-Aug-99 10-Sep-99

Date

Figure 4.6a. Pilot Study Cumulative Total Chromium in Cathode Wells

Anode Wells Cumulative Total Chromium Concentration (mg/L)

300 AW09 AW10 250

AW11 AW18 AW19 AW20

200

150

100

50

0 13-Jan-99 12-Feb-99 14-Mar-99 13-Apr-99 13-May-99 12-Jun-99

12-Jul-99

11-Aug-99 10-Sep-99

Date

Figure 4.6b. Pilot Study Cumulative Total Chromium in Anode Wells

72

15 feet (+)

(-)

20

20

20

20

20

20

20

20

20

20

20

20

20

20

20

20

10

10

10

10

10

10

10

10

10

10

10

10

10

10

10

10

5

5

5

5

5

5

5

6

7

8

9

10

10

12

13

14

20

19

18

16

15

15

16

16

16

15

15

14

13

12

12

11

5

5

5

5

5

5

5

5

5

5

5

5

5

5

5

5

10 feet Concentration > 3.1 mg/kg from 1998 N atural Background Levels _ 3.1 mg/kg from 1998 N atural Background Levels Concentration <

Figure 4.7a. Pilot Study Cadmium Results from Initial Site Characterization

15 feet (+)

(-)

0

0

2

0

3

0

0

0

0

12

27

44

0

49

78

36

0

0

1

0

0

1

0

0

0

4

14

32

8

13

17

47

0

0

0

0

0

3.5

0

0

0

0

0

2

4

4

4

16

0

0

0

0

0

0

0

0

0

2

0

0

3.3

3.4

3

0

2

1

0

0

0

0

28

0

0

1

0

0

4

5

2

3

10 feet Concentration > 3.1 mg/kg from 1998 Natural P t. Mugu Background Levels Concentration <_ 3.1mg/kg from 1998 Natural P t. Mugu Background Levels

Figure 4.7b. Pilot Study Soil Cadmium Results after Treatment

73

Maintance and Adjustments to Offgas System

20,000

pH Controller Failure

Cummulative Energy (kWh)

25,000

Pumps Clogged With Biomass From Acid Tank

Voltage Lowered for Better pH Control

30,000

15,000

10,000 Electric Meter Data Acquistion

5,000

0 0

20

40

60

80

100

120

140

Time Since January Restart (days)

Figure 4.8. Pilot Study Cumulative Energy Used

74

160

SECTION 5 POTENTIAL ENHANCEMENT OF BIOREMEDIATION BY ELECTROCHEMICAL METHODS

Organic contaminants have been known to be present in many hazardous waste sites (USEPA, 1991; DOE, 1995; Kelsh and Parsons, 1997). Economical restoration of these contaminated sites to environmentally acceptable conditions is an important challenge facing the scientific and technical community. Current in situ soil remediation technologies depend on hydraulic and air flow for effective remediation of soils and are not as effective in the clean-up of lower hydraulic conductivity soils (less than 10-5 cm/sec) such as fine sands, silts and clays. In situ bioremediation is an attractive and often cost-effective option to remediate soil and groundwater contaminated with organics. Successful implementation of in situ bioremediation is dependent upon presence, or effective injection, of electron acceptors and nutrients into the porous medium. Microbial processes require an electron donor, macronutrients (e.g, nitrogen and phosphates), micronutrients, trace nutrients and an electron acceptor. Microbially-mediated cometabolite transformations require an additional external electron donor as well. In some cases, inducers may be needed to trigger the microbial transformation reaction. Effective introduction and transport of these additives is hindered by low soil permeability, preferential flow paths (channeling), biological utilization, and chemical reactions in the soil. Complications of site geohydrology, additive transport and associated reactions, coupled with the observed inefficiency in the field, have been mostly approached by gross overinjection of the additives. Excessive dosing coupled with the shortcomings of the hydraulically-driven transport processes can result in nutrient rich areas with excessive biological growth (biofouling). Biofouling adversely impacts system implementation due to reduced conductivity by microbial growth plugging the flow paths. Several surveys have concluded that ineffective transport of remediation additives is the primary cause of system failure for some in situ bioremediation efforts (Zappi et al., 1993; NRC 1993). A technology for uniform introduction of nutrients and electron acceptors/donors has been the principal bottleneck in the successful field implementation of in situ bioremediation (Suflita and Sewell, 1991; Zappi et al., 1993). An emerging technology for treating such hazardous waste sites using in situ bioremediation methods is through the application of electric fields to transport the nutrients as well as bacteria. Electric fields could be used to overcome problems associated with additives injection into heterogeneous and/or low permeability soils. Uniform and accelerated delivery of nutrients, and electron acceptors may be achieved using electric injection instead of hydraulic injection. Uniform transport of ions under electrical fields is controlled by the charge and ionic mobility of available species in the pore fluid. Accelerated and uniform transport rates of nutrients, electron acceptors and microorganisms could be achieved in heterogeneous or low hydraulic conductivity soils by electric fields. The technique could also employ electroosmotic flow into fine-grained soils to enhance additive injection. This method can be used to stimulate bioremediation under aerobic or anaerobic conditions. The conditions generated under dc electric fields and their impact on the porous media are discussed. This section describes the effect of dc fields on ions transport in soils, electrolysis and geochemical 75

reactions, microbial adhesion and transport, and microbial activity. The interest in these in these processes is derived from the potential to develop strategies to enhance in situ bioremediation. A review of ion transport mechanism in porous media under dc fields is provided to address the potential of injecting and transporting nutrients and biostimulants by ion injection and electroosmosis and also the possibility of injecting bioaugmentation innoculants by electrophoresis. A discussion is provided for electrolysis reactions and their effects on pH and dissolved oxygen (DO) values, which in turn affect microbial survival. Electrolysis reactions are also important as they cause the production of oxygen and hydrogen and also may cause abiotic degradation of contaminants. A discussion is provided for the factors that may impact microbial adhesion and transport under dc fields. A review of studies that investigate electrokinetic extraction of organic contaminants is also provided. Finally, the results of studies that evaluate the general impacts of electric fields on anaerobic and aerobic cultures are summarized.

Nutrients Transport under Electric Fields Injection and transport of bioremediation additives by electric fields are controlled by electrokinetic transport mechanisms in an ion exchange medium. These transport mechanisms include electroosmotic advection (or electroosmosis), electromigration (or ionic migration), electrophoresis and to a lesser extent diffusion. Acar et al. (1996; 1997) showed that ionic migration could be used for injection and transport of anionic and cationic additives. In a bench-scale experimental setup, ammonium hydroxide (NH4OH) was introduced at the anode compartment and sulfuric acid (H2SO4) at the cathode compartment. The electric field caused migration of nitrate ion from anode towards the cathode and sulfate ion from cathode towards the anode. The study reported transport rates of 5 to 20 cm/day in fine sand and kaolinite soil specimens and consequent soil saturation of ammonium and sulfate ions. The study concluded that ion migration under dc fields can be used to inject nutrients, electron acceptors/donors to enhance in situ bioremediation.

Redox by Electrolysis Application of direct electric currents in saturated soils results in redox reactions at the electrodes. If inert electrodes (such as graphite) are used, water oxidation generates an acid (H+) and oxygen gas at the anode while water reduction produces a base (OH-) and hydrogen gas at the cathode (Acar and Alshawabkeh, 1993). Based on Faraday’s law for equivalence of mass and charge, the rate of electrolysis reactions depends on the total current applied. For a specific reaction, e.g., water oxidation at the anode, the rate of electrolysis is given by,

J=

I zi F

where J is the rate of oxidation or reduction by electrolysis (MT-1), I is the current (A), z is the 76

charge of the ion (for hydrogen z = 1) and F is Faraday’s constant (96,485 C / mole). Assuming that water electrolysis occurs at the electrodes with 100% efficiency, one amp current will oxidize 1/2F mole H2O into 1/F mole of H+ and 1/4F mole O2 at the anode per second. The same current will reduce 1/F mole H2O into 1/F mole OH- and 1/2F mole H2 per second at the cathode. It should be indicated that other electrolysis reactions may occur and limit the water electrolysis reaction. The type of electrolysis reaction depends upon the chemistry of the electrolyte, pH and the standard electrode potential for ions in the electrolyte. Oxidation and reduction generate complex boundary conditions at the electrodes that can either enhance or retard microbial activity and contaminant degradation. Oxidation at the anode generates oxygen that stimulates aerobic degradation. An acid (H+) is also produced, which could drop the pH at the anode to below 2. The acid migrates under the applied electric field toward the cathode and may acidify the soil. The soil resistance to pH changes depends upon the soil buffering capacity. However, a pH drop is not favored for microbial growth. Optimum pH for bacterial growth is near neutrality with minimum and maximum pH values for growth near 5 and 9, respectively (Gaudy and Gaudy, 1988). A relatively neutral pH is needed to optimize biological activity in the soil. Soil acidification may also cause dissolution of the soil minerals. Electrolytes conditioning that neutralize acid generation and enhance delivery of proper additives for bioremediation should be considered. As an example, the use of ammonium hydroxide at the anode and sulfuric acid at the cathode was successful in (a) neutralizing the acid and base at the anode and cathode and (b) injection of ammonium and sulfate ions into the soil (Acar et al. 1996). Oxygen production is another critical boundary condition generated at the anode. The oxygenated water at the anode may be introduced into the soil at efficient rates by electroosmotic flow to enhance aerobic conditions. Preliminary results indicated that the efficiency of gas production by electrolysis is 75%. Other mechanisms, such as air sparging, in-well aeration, hydrogen peroxide, oxygen releasing solid compounds, or even cryogenic oxygen generators, can be used to supplement oxygen requirement. However, the advantage of dc fields is that they serve for both oxygen generation and transport in low hydraulic conductivity soils. It should be noted that while this oxygen will enhance conditions for aerobic microorganisms, it is likely to adversely impact the anaerobic microorganisms. If anaerobic bioremediation is the target treatment method, measures to remove oxygen at the anode using chemical scavengers or gas-stripping (eg. nitrogen gas) are needed to prevent adverse impacts on anaerobic microorganisms. Alternatively, an aerobic-anaerobic treatment could be designed to take advantage of the availability of oxygen in a limited subsurface zone

Organics Extraction by dc Fields Electric fields have been used for extraction of the contaminants from the subsurface. Bench-scale tests and limited field studies demonstrated the use of electric fields for extraction of heavy metal and radionuclides from soils (Lageman et al., 1989; Hamed et al., 1991; Pamukcu and Wittle, 1992; Acar and Alshawabkeh, 1993; Probstein and Hicks, 1993; Runnels and Wahli, 1993; Eykholt and Daniel, 1994; Alshawabkeh and Acar, 1996; Pamukcu et al., 1997). Electric fields have also been used for electroosmotic extraction of organic contaminants from soils (Shapiro et al., 1989; Acar et 77

al., 1992; Bruell et al., 1992). Recently, the Department of Energy, Environmental Protection Agency (EPA), Monsanto, General Electric, and Dupont applied electric fields for electroosmotic extraction of TCE from a site in Paducah, Kentucky using layered horizontal electrodes by the “Lasagna” process (Ho et al., 1997; 1999a; 1999b). The Lasagna process uses mainly electroosmosis for extraction of TCE out of the soil into electrode or treatment wells. Electric fields use in soil restoration has been focused on contaminant extraction by their transport under electroosmosis and ionic migration. Contaminant extraction by electric fields is a successful technique for removal of ionic or mobile contaminants in the subsurface. However, this technique might not be effective in treatment of soils contaminated with immobile and/or trapped organics, such as dense non aqueous phase liquids (DNAPLs). For such organics, it is possible to use electric fields to stimulate in situ biodegradation under either aerobic or anaerobic conditions. It is necessary to evaluate the impact of dc electric fields on the biogeochemical interactions prior to application of the technique. It is not clear yet how dc electric fields will impact microbial adhesion and transport in the subsurface. Further, the effect of dc fields on the activity of microorganisms in a soil matrix is not yet well understood.

Microbial Adhesion and Transport The presence of appropriate microorganisms (depending on electron-acceptor and nutrient availability) at the actual site of contamination (sometimes at a micro-scale) has long been recognized as a key factor in determining if biodegradation/biotransformation will occur, as well as in influencing the rate of biodegradation. Microorganisms can be present in the subsurface in suspension (in the pore fluid), as microcolonies or as a biofilm. Bacterial adhesion to porous media often influences the nature and extent of colonization of a subsurface medium. The biofilm development model (Characklis and Wilderer, 1988), summarized below, can provide some understanding into bacterial transport and adhesion in subsurface environments. The "first three steps" in biofilm development are: (1) Surface conditioning, (2) Transport of microoganisms to the conditioned surface and, (3) Sorption of microorganisms to the surface. In biofilm literature, it is recognized that the nature of the bacterial cell surface is a key parameter in determining its adhesion to any media and eventually the biofilm architecture (Reynolds et al., 1989; Characklis and Wilderer, 1988; van Loosdrecht et al., 1987). The sorption of microorganisms is due to reversible sorption (governed by charge/electrostatic interactions) and irreversible sorption (due to production of extracellular exoploymers and formation of matrix). The dependency of reversible sorption of bacteria on charge/electrostatic interactions indicates that an applied electric field may play a significant role in bacterial adhesion and transport in subsurface environments. Further, the role of bacterial surface (eg., lipid content and surface charge) itself could influence the electrophoretic mobility of bacteria in porous media. Data available in literature indicate that the presence of heavy metals alter the electrokinetic properties of bacteria, although the context of these researchers was not hazardous waste decontamination (Collins and Stotzky, 1992). The imposed dc electric field is also expected to affect the electrokinetic properties, adhesion and transport of microorganisms. It is not clear what will be the extent of this effect; however, the authors noted that microorganisms tend to stick and attach themselves to the electrodes in experiments conducted using diluted sludge samples under dc fields. The type or activity of microorganisms attached to the electrodes was not 78

evaluated. Electrode polarity did not seem to have an effect on adhesion as microorganisms were attached not only to the anode but also to the cathode. This might indicate that the adhesion to electrode surface is not necessarily due to the electric attraction between the negatively charge microrganisms and the positively charged anode. Further evaluation is needed to verify if this adhesion is due to reversible sorption, irreversible sorption, or due to electrode charge. Electric fields will not only impact microorganisms adhesion but will also affect their transport in porous media. As microbes are generally negatively charged, dc fields will cause their transport towards the anode. DeFlaun and Condee, (1997) demonstrated electrokinetic transport of a pure bacterial culture in bench-scale soil samples. Generally, the rate of transport is related to the effective electrophoretic mobility in soils and is affected by soil physical parameters such as porosity, pore size distribution and tortousity. The results of DeFlaun and Condee (1997) demonstrate the potential of using microorganisms electrophoresis for the purpose of bioaugmentation to enhance in situ bioremeidation. It is also necessary to note that microorganisms transport in porous media under electric field might not be strictly governed by electrophoresis alone, since microbes, as living entities, may be subject to other influences or "attractors" and also tend to form colonies and attach to the soil particle surface.

Microbial Activity under dc Fields Recently, there has been an increasing interest in the bioelectrochemical processes in medicine and biosensor fields. Studies showed that low level ac (alternating currents) and dc electric fields (in the range of volts/cm and up to few hundred Hz) stimulate the metabolic processes (Berg and Zhang, 1993; McLeod et al., 1992, Blank et al., 1992) in a nonlinear way so that only a specific range (or so-called window) of field strength and frequencies can cause a significant impact (Tsong, 1992; Fologea et al., 1998). This “electrostimulation” process has been explored in areas that include enzyme activation, biopolymer synthesis, membrane transport, and proliferation (Berg, 1993). Furthermore, bioelectrochemical devices (or biosensors) are being developed for manipulation of bacteria, viruses and genetic material using sophisticated microelectrodes (Buerk, 1993; Ramsey, 1998). Electrostimulation and biosensors research fields are at the micro-scale (might reach the nano-meter) level, focus on ac fields, and have not yet been employed in the soil bioremediation area. It is necessary to evaluate the dc field intensities that microorganisms can sustain and also the “window” of dc fields that may stimulate microbial activity. Electric fields also introduce environmental changes that affect microbial growth. As discussed earlier, electrolysis reaction impact pH, Dissolved Oxygen (DO) and other geochemical conditions. Furthermore, electric fields may produce an increase in temperature. Most microorganisms grow rapidly at temperatures between 20 and 45 oC and are capable of growing over a range of 30 to 40oC (Gaudy and Gaudy, 1988). Temperature increase to above 45 oC will significantly limit the growth of most microbes (some could survive high temperatures). Temperature increase due to current application will depend upon field strength and resistivity of the medium. Acar and Alshawabkeh (1996) reported 10 oC increase in temperature in an unenhanced (no additives were used to control electrolyte pH) large-scale test on extraction of lead from kaolinite. 79

Anaerobic Microbial Activity The impact of electric current on the environmental conditions and the anaerobic microbial activity in completely mixed fed-batch reactors was studied at various electric field strengths (Maillacheruvu and Alshawabkeh, 1999). Experiments were conducted in bio-electrokinetic (BioEK) reactors, which consist of plexiglass boxes with titanium-coated mesh electrodes mounted at both ends. Electric fields of 1.5 V/cm through 6 V/cm were applied. Unacclimated anaerobic cultures obtained from a mesophilic anaerobic digester were used in these experiments. Limited pH changes may occur if water electrolysis reactions (Equations 3 and 4) occur at the same rate and efficiency. In a completely mixed reactor, the proton produced at the anode should neutralize the hydroxyl ion produced at the cathode. However, the results indicated that the pH decreased to less than 5.5 even under completely mixed conditions in fed-batch reactors. The pH drop indicate less hydroxyl production at the cathode, either because different electrolysis reactions occurred (other than Equation 4) or because of biochemical reactions in the reactor. The type and concentrations of ions in the solution will impact the pH changes and require further investigation. Sodium bicarbonate was used and was effective in buffering the system for the range of electric field strengths studied. Dissolved oxygen was produced in the experimental reactors, proportional to electric field strength used, where no oxygen scavengers were used. Dissolved oxygen was, however, controlled by addition of sodium sulfite to the system. In an actual soil, there may exist niches, which are devoid of oxygen where anaerobic bacteria may survive even if some dissolved oxygen was produced and not eliminated completely using oxygen scavengers. Further studies in soils are needed to evaluate this hypothesis. Some anaerobic bacteria (particularly sulfate-reducing bacteria or SRB) may indeed have survived even in the presence of relatively high concentrations of oxygen. Electric current in the reactor generally increased slightly with time (5% at 1.5 V/cm to about 15%18% at 4.5 V/cm for an exposure period of about 140 hours) at different electric field strengths. While this suggests that the ionic composition of the electrolyte medium changed over time, results from this preliminary study indicate that dissolved organic carbon (DOC) removal efficiency and microbial activity do not appear to be significant, especially at the lower end of electric field strengths. Microbial activity was estimated as the capacity of the culture to recover from exposure to electric currents. Microbial activity was measured as a function of the ability of the anaerobic microorganisms to consume readily degradable acetate (measured as DOC). A sample of culture was withdrawn from a fed-batch experimental reactor at regular intervals. One part of the sample was immediately analyzed (designated the "to sample") for DOC concentration while the other part allowed to "recover" for a period of 24 hours in an evacuated glass vial to preserve anaerobic conditions. After 24 hours had elapsed the DOC was measured again. The second part of the sample was designated the "t24 sample". If the t24 sample shows a decrease in DOC as compared to the to sample, it is indicative of an active culture. Figure 5.1 shows the microbial activity data for experiments at electric field strengths of 1.5 V/cm and 4.5 V/cm. These data indicate that, before 80

the electric current was applied, the t24 sample showed more removal (about 20%) of DOC than the to sample -- indicative of an active culture as noted earlier. From Figure 5.1, it is apparent that the initial "shock" exposure to the electric current results in a decrease in activity for the t24 sample as compared to the to sample. This trend continued for several hours during which period the total DOC percent removal in the experimental reactor also dropped by about 10 to 20% in all the experiments tested. However, after a period of several hours of exposure, the t24 sample gradually showed an increase in microbial activity by exhibiting higher DOC removal than the to sample -even during the application of the electric field. This is an interesting phenomenon since it indicates a certain degree of acclimation of the culture to electric current, and a tendency for the anaerobic culture to recover from the initial shock load in terms of changes in environmental conditions due to application of the electric current. Once the electric current application was stopped, the rate of DOC percent removal in the experimental reactors gradually improved. Eventually, after the removal of the electric current, the 24 hour sample showed about 20% higher removal of DOC as compared to the to sample. These data also suggested that there was essentially no difference in recover of microbial activity 1.5 V/cm and 4.5 V/cm experiments. Results from 3.0 V/cm and 6.0 V/cm showed the same trends. Aerobic Microbial Activity Another preliminary study was conducted to evaluate the impact of dc fields on aerobic microbial activity in a completely mixed and aerated reactor. The study used a sludge sample from the aeration tank of Deer Island Wastewater Treatment Plant, MA. The sludge was placed in two identical polyethylene carboys and aerated by gas diffusing stones connected to an air source. A batch-fed system was used to maintain the sludge. Percent increment of volatile suspended solids (VSS) was used to measure growth rate and the sludge was maintained until steady state growth rates were achieved after about 20 days. Samples were then taken from the sludge and placed in Bioelectrokinetic (BioEK) reactors (Figure 5.2). The reactors consist of acrylic boxes (14-cm length × 14-cm width × 10-cm height), which have two holes on the top: one was used for sampling and for measurement probes (DO, pH, conductivity and temperature probes), while the other hole was used for the aeration tubing. Two titanium-coated mesh electrodes were fixed on the inner side of the box as anode and cathode (Fig. 3). Samples were then taken from the sludge and placed in the BioEK reactors (Figure 5.2). The reactors were modified to allow aeration of sludge. Electric dc fields of 4, 8, and 16 volts, which reflect 0.28, 0.57, and 1.14 V/cm dc electric fields, respectively, were applied. These electric field strengths were less than those used in the anaerobic tests, which makes it easier to control the environmental changes, such as pH, produced by electrolysis. Initial testing conditions are summarized in Table 5.1. Complete aeration and mixing maintained dissolved oxygen around saturation during exposure to electricity. Temperature measurements did not show any significant change during testing. Although there was a slight decrease in pH, sludge mixing allowed electrolysis reactions at the electrodes to neutralize each other, thus minimizing significant pH changes. Anaerobic tests showed a drop in the pH to about 5.5. The slight drop in pH in the aerobic tests compared to the anaerobic tests may be 81

related to the difference in processing time (which was shorter in the aerobic tests) and the constituents of the anaerobic and aerobic cultures. It is possible that another electrolysis reaction, other than water reduction, has occurred at the cathode in the anaerobic tests thus limiting OHproduction. Another possible reason is that the currents used in the aerobic tests are smaller than those used in the anaerobic tests, which may limit the changes due to electrolysis in the aerobic tests when compared to the anaerobic tests. In any case, this is an issue that needs further evaluation. With most variables (pH, DO and temperature) were controlled, any changes in microbial activity can only be attributed to electric currents. Percent change in chemical oxygen demand (COD) of the sludge was used as an indication of microbial activity. Other measures of substrate utilization may be used, such as TOC, DOC, BOD, and/or VSS. However, COD was used because it is easier and faster than some of these measures and because of our interest in the general response. Some concerns may rise because microorganisms are organic in nature and COD might not accurately reflect microbial activity. However, microorganisms consume more organic matter than they synthesize and the ratio of consumed to synthesized organic matter is about 10:1 under normal aerobic condition (Gaudy and Gaudy, 1988). Therefore, COD values provide a good indication of microbial activity. Furthermore, COD values were used for comparison with control test (no electric field) results and not as absolute values. Average COD changes are summarized in Figure 5.3. The behavior can be divided into 2 groups: one describing the low voltage (LV) of 0 and 4 V tests and the other describing high voltage (HV) of 8 and 16 V tests. After 24 hours, the LV group showed less drop (about 17% difference) in COD, when compared to the HV group. This is an indication that tests with higher voltage gradients (8 and 16 V) resulted in more degradation of the organic matter. After 48 hours of exposure to electricity, both LV tests showed the same behavior, where COD dropped to around 42-45% of the initial value. This is an indication that application of 4 volts (0.28 V/cm) may not be high enough to produce any changes, when compared to tests with no electricity. On the other hand, the test with 8 volts seems to cause the most significant drop in the COD value (61% after 48 hours exposure). The impact of the highest voltage (16 V) seems to diminish after the first 24 hours. While the 16 V tests showed less COD drop when compared to the 8 V test, it still showed more COD drop when compared to the LV tests. The change in COD drop between the tests may be attributed to an increased microbial activity. The results indicate that dc electric fields (up to 1.14 V/cm) do not have an adverse effect on mixed aerobic cultures. In fact, an increase in the degradation rate may occur due to increasing the voltage gradient up to a certain value (0.58 V/cm in this study), beyond which this increase may diminish. This behavior is similar to electrostimulation by ac electric fields reported in the bioelectrochemistry research area (Fologea et al., 1998). For the conditions of the preliminary study, the window of significant response of bacteria stimulation is somewhere between 0.27 and 1.14 V/cm. Further, the results show that this window of dc range is also affected by time of exposure. This can be concluded as the impact of the 1.14 V/cm field diminished after 24 hours. However, to confirm this conclusion, it is necessary to separate the effects of any abiotic processes. The tests were conducted in an open, mixed and aerated reactor to maintain constant values of pH, DO, and temperature. Thus the difference in COD drop may not be related to pH, temperature. 82

Aeration and mixing maintained DO around saturation in all tests, thus the effect of oxygen production at the anode is minimized. The only other process (other than microbial activity) that may relate to COD drop is abiotic transformation by electrolysis reactions at the electrodes. If abiotic redox of the organic content occurs in this study, then increasing the current density should increase the COD drop. The results do not show this pattern, but show a peak followed by a decrease in COD drop with increasing current density (Figure 5.3). Accordingly, microbial activity may be the major factor in the difference in COD drop obtained in tests with different current densities.

83

Table 5.1. Initial testing conditions for aerobic tests. Sludge Volume Volatile Suspended Solids (VSS) Chemical Oxygen Demand (COD) Dissolved Oxygen (DO) pH Conductivity Voltage (No. of Tests)

84

~ 1.5L ~ 280 mg/l 390~414 mg/l ~ 8 mg/l ~8 0.5~1.0 mS/cm 0 (3); 4V (3); 8V (2); 16V (2)

50 40 30

% Removal

EF Applied EF Removed

20 10 0 -10

1.5 V/cm + Sulfite 4.5 V/cm + Sulfite

-20 -30 -40 -50 0

100

200

300

400

500

600

Time (Hours)

Figure 5.1. Variation in anaerobic microbial activity under electric fields (Maillacheruvu and Alshawabkeh, 1999).

Aeration Tubing Air

dc Power Supply Probe

Electrode Reactor Porous Stone

Sludge Sample

Magnetic Stirrer

Electrode

Stir Bar

Figure 5.2. A schematic of BioEK reactor

85

70 0V 4V 8V 16 V

% Drop of COD

60 50 40 30 20 10 0 0

20

40

60

Time (hour) Figure 5.3. Average COD drop in aerobic tests (electrode spacing =14 cm)

86

References Acar, Y. B. and Alshawabkeh, A. N. (1993), “Principles of Electrokinetic Remediation,” Environmental Science and Technology, 27(13): 2638-2647. Acar, Y. B. and Alshawabkeh, A. N. (1994) "Modeling conduction phenomena in soils under an electric current," Proceedings of XIII International Conference on Soil Mechanics and Foundation Engineering (ICSMFE), New Delhi, India. Acar, Y. B. and Alshawabkeh, A. N. (1996), “Electrokinetic Remediation: I. Pilot-scale Tests with Lead-Spiked Kaolinite,” ASCE, Journal of Geotechnical Engineering, Vol. 122, No. 3, pp. 173-185. Acar Y. B. and Alshawabkeh, A. N. (1997) Electrochemical Decontamination of Soil and Water, Special Issue of Journal of Hazardous Material, Vol. 55, No’s. 1, 2 and 3, 322p. Acar, Y. B. and Olivieri, I (1989) "Pore Fluid Effect on the Fabric and Hydraulic Conductivity of Laboratory Compacted Clay," Transportation Research Record, No. 1219: pp 144-159. Acar, Y. B., Gale, R. J., Putnam, G. and Hamed, J. (1989), "Electrochemical processing of soils: its potential use in environmental geotechnology and significance of pH gradients," 2nd International Symposium on Environmental Geotechnology, Shanghai, China, May 14-17, Envo Publishing, Bethlehem, PA, Vol. 1: 25-38. Acar, Y. B., Gale, R. J., Putnam, G. A, Hamed, J., and Wong, R. L. (1990). "Electrochemical processing of soils: Theory of pH gradient development by diffusion, migration, and linear convection." J. Envir. Sci. Health, A25(6), 687-714. Acar, Y. B., Hamed, J. T., Alshawabkeh, A., and Gale, R. J. (1994). "Cd(II) removal from saturated kaolinite by application of electrical current." Géotechnique, 44(3), 239-254. Acar, Y. B., Li, H., and Gale, R. J. (1992). "Phenol removal from kaolinite by electrokinetics." J. Geotech. Engrg., ASCE, 118(11), 1837-1852. Acar, Y. B., Alshawabkeh, A., and Gale, R. J. (1993). "Fundamentals of extracting species from soils by electrokinetics." Waste Management, 13(2), 141-151. Acar, Y. B., Ozsu, E., Alshawabkeh, A. N., Rabbi, F. M., and Gale, R. (1996) “Enhanced Soil Bioremediation with Electric Fields,” CHEMTECH, ACS, 26(4), pp. 40-44. Acar, Y. B., Rabbi, M. F. and Ozsu, E. (1997) “Electrokinetic Injection of Ammonium and Sulfate Ions into Sand and Kaolinite Beds,” Journal of Geotechnical and Geoenvironmental Engineering, ASCE, 123(3), pp. 239-249.

Alshawabkeh, A. N., Puppala, S. K., Acar, Y. B., Gale, R. J., and Bricka, R. M. (1997) “Effect of Solubility on Enhanced Electrokinetic Extraction of Metals” In Situ Remediation of the Geoenvironment (In Situ Remediation >97), Minneapolis, Minnesota, October 5-8, 1997.

87

Alshawabkeh, A. N., Yeung, A. and Bricka, R. M, (1999) “Practical aspects of in situ electrokinetic remediation,” ASCE Journal of Environmental Engineering, accepted, in press. Alshawabkeh, A. N. and Acar, Y. B. (1992), “Removal of Contaminants from Soils by Electrokinetics: A Theoretical Treatise,” Journal of Environmental Science and Health, A27 (7): 1835-1861. Alshawabkeh, A. N. and Acar, Y. B. (1996), “Electrokinetic Remediation: II. Theory,” Journal of Geotechnical Engineering, ASCE, 122(3): 186-196. Anderson, J. L. and Idol, W. K. (1986) “Electroosmosis through pores with non uniformly charged walls,” Chemical Engineering Communications, Vol. 38, pp. 93-106. Bard, A.J. and Faulkner, L.R. (1980), Electrochemical Methods: Fundamentals and Applications, John Wiley and Sons, Inc. 718p. Banarjee, S., Horng, J., Ferguson, J., and Nelson, P. (1990),"Field scale feasibility of electrokinetic remediation," report presented to USEPA, Land pollution Control Division, PREL, CR 811762-01, 122p. Bear, J. (1972), Dynamics of Fluids in Porous Media, American Elsevier, New York. Bethke, C.M. (1996). Geochemical Reaction Modeling: Concepts and Applications, Oxford Univ. Press, New York, NY, 1-397. Berg, H. (1993) “Electrostimulation of Cell Metabolism by Low Frequency Electric and Electromagnetic Fields,” Bioelectrochemistry and Bioenergetics, 31, pp. 1-25. Berg, H. and Zhang, L. (1993) “Electrostimulation in Cell Biology by Low Frequency Electromagnetic Fields,” Electro Magnetbiology, 12, 147-163. Blank, M., Soo, L., Henderson, A. S., Goodmann, R. (1992) “Changes in Transcription in HL60 Cells Following Exposure to Alternating Currents from Electric Fields,” Bioelectrochemistry and Bioenergetics, 28, pp. 301-309. Boynton, S. S. and Daniel, D. E. (1985) “Hydraulic conductivity tests on compacted clay,” Journal of Geotechnical Engineering, ASCE, 111(4): 465-478. Bruell, C. J., Segall, B. A., and Walsh, M. T. (1992). "Electroosmotic removal of gasoline hydrocarbons and TCE from clay." J. Envir. Engrg., ASCE, 118(1), 68-83. Buerk, D. G. (1993), Biosensors: Theory and Applications, Technomic Publishing Co., 221p. Characklis W. and Wilderer, P., (1988), Biofilms, John Wiley and Sons, New York. Collins, Y.E., and Stotzky, G. (1992). Heavy metals alter the electrokinetic properties of bacteria, yeasts, and clay minerals. Applied and Environmental Microbiology, vol. 58, p 88

1592-1600. Cox, C. D., Shoesmith, M. A., and Ghosh, M. M. (1996). "Electrokinetic remediation of mercurycontaminated soils using iodine/iodide lixivant." Envir. Sci. & Technol., 30(6), 1933-1938. Daniel, D. E. (1989) “In Situ hydraulic conductivity tests for compacted clays,” Journal of Geotechnical Engineering, ASCE, 115(9): 1205-1226. Davis, J.A., and Kent, D.B. (1990). “Surface complexation modeling in aqueous geochemistry,” in M.F. Hochella and A.F. White (eds.) Mineral-Water Interface Geochemistry, Reviews in Mineralogy, 23, 177-260. DeFlaun, M. F. and Condee, C. W. (1997) “Electrokinetic Transport of Bacteria,” Journal of Hazardous Materials, Special Edition on Electrochemical Decontamination of Soil and Water, Edited by Yalcin B. Acar and Akram N. Alshawabkeh, pp. 263-278. Denisov, G., Hicks, R. E., and Probstein, R. F. (1996). "On the kinetics of charged contaminant removal from soils using electric fields." J. Colloid and Interface Sci., 178(1), 309-323. DOE/EM-0232 (1995) “Estimating the Cold War Mortgage, Volume I” The 1995 Baseline Environmental Management Report, US Department of Energy, March 1995. Eykholt, G. R. (1992). "Driving and complicating features of the electrokinetic treatment of contaminated soils." PhD thesis, Dept. of Civ. Engrg., Univ. of Texas at Austin, Tex. Eykholt, G. R., and Daniel, D. E. (1994). "Impact of system chemistry on electroosmosis in contaminated soil." J. Geotech. Engrg., ASCE, 120(5), 797-815. Fologea, D., Vassu-Dimov, T., Stoica, I., Csutak, O., and Radu, M. (1998) “Increase of Saccharomyces Cerevisiae Plating Efficiency After Treatment with Bipolar Electric Pulses,” Short Communication, Bioelectrochemistry and Bioenergetics, 46, pp. 285-287. Gaudy, A. F. and Gaudy, E. T. (1988), Elements of Bioenvironmental Engineering, Engineering Press, Inc., San Jose, California, 592p. Gray, D. H., and Mitchell, J. K. (1967). "Fundamental aspects of electro-osmosis in soils." J. Soil Mech. Found. Div., Proc. ASCE, 93(6), 209-236. Gilham, R. W. and Cherry, J. A. (1982) "Contaminant Migration in Saturated Unconsolidated Geologic Deposits," Geological Society of America, Special Paper 189: 31-61. Hamed, J. (1990), "Decontamination of Soil Using Electro-osmosis," A Dissertation, Louisiana State University in Partial Fulfillment of the Degree of Doctor of Philosophy. Hamed, J., Acar, Y. B., and Gale, R. J. (1991). "Pb(II) removal from kaolinite using electrokinetics." J. Geotech. Engrg., ASCE, 117(2), 241-271.

89

Haran, B. S., Popov, B. N., Zheng, G. and White, R. E. (1997) “Mathematical modeling of hexavalent chromium decontamination from low surface charged soil,” Electrochemical Decontamination of Soil and Water, Special Issue of Journal of Hazardous Material, 55 (1-3): 93108. Hicks, R. E., and Tondorf, S. (1994). "Electrorestoration of metal contaminated soils." Envir. Sci. & Tech., 28(12), 2203-2210. Ho, S. V., Athmer, P.W., Sheridan, P. W. and Shapiro, A. (1997) “Scale-up Aspects of the Lasagna Process for in-situ Soil Remediation,” Journal of Hazardous Materials, Special Edition on Electrochemical Decontamination of Soil and Water, Edited by Yalcin B. Acar and Akram N. Alshawabkeh, pp. 39-60. Ho, S. V., Athmer, C., Sheridan, P. W., Hughes, B. M., Orth, R., McKenzie, D., Brodskey, P. H., Shapiro, A. M., Sivavec, T. M., Salvo, J., Schultz, D., Landis, R., Griffith, R., and Shoemaker, S. (1999a) “The Lasagna Technology for In Situ Soil Remediation. 1. Small Field Test,” Environmental Science and Technology, 33(7), pp. 1092-1099. Ho, S. V., Athmer, C., Sheridan, P. W., Hughes, B. M., Orth, R., McKenzie, D., Brodskey, P. H., Shapiro, A. M., Thornton, R. Salvo, J., Schultz, D., Landis, R., Griffith, R., and Shoemaker, S. (1999b) “The Lasagna Technology for In Situ Soil Remediation. 2. Large Field Test,” Environmental Science and Technology, 33(7), pp. 1092-1099. Holmes, P. J. (1962), The Electrochemistry of Semiconductors, Academic Press, London, 396p. Hunter, R. J. (1981). Zeta potential in colloid science. Academic Press, New York, N.Y. Jacobs, R. A., Sengun M. Z., Hicks, R. E., and Probstein, R. F. (1994), “Model and Experiments on Soil Remediation by Electric Fields,” Journal of Environmental Science and Health, A29(9): 19331955. Jacobs, R. A. and Probstein, R. F. (1996) “Two-dimensional modeling of electromigration,” AICHE Journal, 42(6): 1685-1696. Jennings, A. A., Kirkner, D. J., and Theis, I. L. (1982), “Multicomponent Equilibrium Chemistry in Groundwater Quality Models,” Water Resources Research, 18(4): 1089-1096. Kelsh, D. J. and Parsons, M. W. (1997) "Department of energy sites suitable for electrokinetic remediation" Journal of Hazardous Materials, Special Edition on Electrochemical Decontamination of Soil and Water, Vol 55, No. 1—3. Kirkner, D. J., Theis, I. L., and Jennings, A. A. (1984), “Multicomponent Solute Transport with Sorption and Soluble Complexation,” Advances in Water Resources, No. 7: 120-125. Kirkner, D. J., Theis, I. L., and Jennings, A. A. (1985), “Multicomponent Mass Transport withChemical Interaction Kinetics,” Journal of Hydrology, 76: 107-117.

90

Kirkner, D. J. and Reeves, M. (1988), "Multicomponent Mass Transport with Homogeneous and Hetrogeneous Chemical Reactions: Effect of the Chemistry on the choice of Numerical Algorithm, 1, Theory," Water Resources Research, 24(1): 1719--1729. Koryta, J. (1982), Ions, Electrodes, and Membranes, John Wiley and Sons, New York. Kruyt, H. R. (1952), Colloid Science (I): Irreversible Systems, Elsevier Publishing Company, 389p. Lageman, R., Wieberen, P., and Seffinga, G. (1989) “Electro-Reclamation: Theory and Practice,” Chem. Industry London, 9: 585-590. Lageman, R. (1993) “Electro-Reclamation,” Journal of Environmental Science and Technology, 27(13): 2648-2650. Langmuir, D. (1997). Aqueous Environmental Geochemistry, Prentice-Hall, Inc., Upper Saddle River, NJ; 1-600. Lewis, F. M., Voss, C. I., and Rubin, J. (1987), “Solute Transport with Equilibrium Aqueous Complexation and Either Sorption or Ion Exchange; Simulation Methodology and Applications,” Journal of Hydrology, 90: 81-115. Lichtner, P, C. (1985), “Continuum Model for Simultaneous Chemical Reactions and Mas Transport in Hydrothermal Systems,” Geochem. Cosmochim. Acta, No. 49: 779-800. Lockhart, N. C. (1983), "Electro-osmotic dewatering of clays I, II, and II," Colloids and Surfaces, 6, pp. 238-269. Lorenz, P. B. (1969), "Surface Conductance and Electrokinetic Properties of Kaolinite Beds," Clays and Clay Minerals, Vol. 17: 223--231. Maillacheruvu K. and Alshawabkeh, A. N. (1999) “Anaerobic Microbial Activity under Electric Fields,” Book Chapter in “Emerging Technologies in Hazardous Waste Management VIII, Kluwer Academic/Plenum Publishers, in press. McLeod, B. R., Liboff, A. R., and Smith, S. D. (1992) “Biological Systems in Transition: Sensitivity to Extremely Low Frequency Field, Electro Magnetbiology, 11, 29-42. Miller, C. W. and Benson, L. V. (1983), “Simulation of Solute Transport in a Chemically Reactive Heterogeneous System: Model Development and Application,” Water Resources Research, 19., No. 2, pp. 381-391. Mitchell, J. K. (1956), “The fabric of natural clays and its relation to engineering properties,” Proceedings of the Highway Research Board, Vol. 35, pp. 693-713. Mitchell, J. K. (1991). "Conduction phenomena: from theory to geotechnical practice." 91

Géotechnique 41(3), 299-340. Mitchell, J. K. (1993), Fundamentals of Soil Behavior, John Wiley and Sons, New York. Mitchell, J. K. and Yeung, T. C. (1991), "Electro--kinetic Flow Barriers in Compacted Clay," Transportation Research Record, No. 1288, Soils Geology and Foundations, Geotechnical Engineering 1990, pp. 1--10. NRC (1993), In Situ Bioremediation: When Does it Work?, National Research Council, National Academy Press, Washington D.C., 207 p. Olson, R. E. and Daniel, D. E. (1981) “Measurements of the hydraulic conductivity of fine-grained soils,” ASTM STP 746 Pamukcu, S., Khan, L., and Fang, H. (1990),"Zinc detoxification of soils by electroosmosis," Electro-kinetic Phenomena in Soils, Transportation Research Record, TRB, Washington, D.C. Pamukcu, S., and Wittle, J. K. (1992). "Electrokinetic removal of selected heavy metals from soil." Envir. Progress, AIChE, 11(4), 241-250. Pamukcu, S., Weeks, A. and Wittle, J. K. (1997) “Electrochemical Extraction and Stabilization of Selected Inorganic Species in Porous Media,” Journal of Hazardous Materials, Special Issue on Electrochemical Decontamination of Soil and Water, Edited by Y.B. Acar and A.N. Alshawabkeh, V. 55, No’s. 1—3, pp. 1-22. Perkins, T. K., and Johnston, O. C. (1963) “A review of diffusion and dispersion in porous media,” Journal of Society of Petroleum Engineering, Vo. 19, 70-84. Probstein, R. F. and Hicks, R. E. (1993), “Removal of Contaminants from Soils by Electric Fields,” Science, 260, pp. 498-504. Puppala, S., Alshawabkeh, A. N., Acar, Y. B., Gale, R. J, and Bricka, R. M. (1997) “Enhanced electrokinetic remediation of high sorption capacity soils,” Electrochemical Decontamination of Soil and Water, Special Issue of Journal of Hazardous Material, Vol. 55, No’s. 1-3, pp. 203-220. Ramsey, G. (1998), Commercial Biosensors: Applications to Clinical, Bioprocess, and Environmental Samples, John Wiley and Sons, Inc., 304p. Reddy, K. R., Parupudi, U. S., Devulapalli, S. N., and Xu, C. Y. (1997) “Effect of soil composition on removal of chromium by electrokinetics,” Electrochemical Decontamination of Soil and Water, Special Issue of Journal of Hazardous Material, Vol. 55, No=s. 1-3, pp. 135-158. Reynolds, P. J., Sharma, P., Jenneman, G. E., McInerney, M.J. (1989). Mechanisms of microbial movement in subsurface materials. Applied and Environmental Microbiology, vol. 55, p 2280-2286. Rødsand, T., Acar, Y. B., and Breedveld, G. (1995). "Electrokinetic extraction of lead from spiked 92

Norwegian marine clay." Characterization, containment, remediation, and performance in environmental geotechnics. Geotech. Spec. Publ. No. 46, ASCE, New York, N.Y., 2, 1518-1534. Rubin, J. (1983), “Transport of Reacting Solute in Porous Media; Relation Between Mathematical Nature of Problem Formulation and Chemical Nature of Reactions,” Journal of Water Resources Research, 19(5): 1231-1252. Runnels, D. D. and Larson, J. L. (1986), "A laboratory study of electromigration as a possible field technique for the removal of contaminants from ground water," Ground Water Monitoring Review, pp. 81-91, Summer 1986. Runnels, D. D. and Wahli, C. (1993) “In Situ Electromigration as a Method for Removing Sulfate, Metals, and Other Contaminants from Groundwater,” Groundwater Monitoring Review, Winter 1993, pp. 121-129. Schultz, D. S. (1997) “Electroosmosis technology for soil remediation: laboratory results, field trial and economic modeling,” Electrochemical Decontamination of Soil and Water, Special Issue of Journal of Hazardous Material, Vol. 55, No’s. 1-3, pp. 81-92. Segall, B. A., and Bruell, C. J. (1992). "Electroosmotic contaminant removal processes." J. Envir. Engrg., ASCE, 118(1), 84-100. Shackelford, C. D. and Daniel, D. E. (1991), "Diffusion in Saturated Soil (I): Background," ASCE, Journal of Geotechnical Engineering, 117 (3): 467--484. Shackelford, C. D. (1991) “Diffusion of contaminants through waste containment barriers,” Transportation Research Record, No. 1219, pp. 169-182. Shapiro, A. P., Renauld, P., and Probstein, R. (1989), “Preliminary studies on the removal of chemical species from saturated porous media by electro-osmosis,” Physicochemical Hydrodynamics, 11, No. 5/6: 785-802. Shapiro, A. P. and Probstein, R.F. (1993), “Removal of contaminants from saturated clay by electroosmosis”, Environmental Science and Technology, 27(2): 283-291. Stumm, W., and Morgan, J.J. (1995). Aquatic Chemistry: Chemical Equilibria and Rates in Natural Waters, 3rd ed., J. Wiley & Sons, Inc., New York, NY. 1-1022. Stumm, W. (1992) "Chemistry of the Solid--Water Interface, Processes at the Mineral--Water and Particle--Water Interface in Natural Systems," A Wiley-Interscience Publication, John Wiley & Sons, Inc., 428p. Suflita, J. M. and Sewell, G. W. (1991) “Anaerobic Biotransformation of Contaminants in the Subsurface,” EPA, Robert S. Kerr Environmental Laboratory, Ada, OK, EPA/600/M-90/024. 93

Tsong, T. Y. (1992) “Molecular Recognition and Processing of Periodic Signals in Cells: Study of Activation of Membrane ATPases by Alternating Electric Fields,” Biochem. Biophys. Acta, 1113, pp. 53-70. Ugaz, A., Puppala, S., Gale, R.J., and Acar, Y. B. (1994). " Electrokinetic soil processing: complicating features of electrokinetic remediation of soils and slurries: saturation effects and the role of the cathode electrolysis," Chem. Engineering Communications, vol. 129: 183-200. USEPA, 1991, Groundwater Issue. Report No. EPA/540/4-91/002, 21 p. Vallocchi, A. J., Street, R. L., and Roberts, P. V. (1981), “Transport of Ion-Exchange Solutes in Groundwater: Chromatographic Theory and Field Simulations,” Journal of Water Resources Research, 17(5): 1517-1527. Van Loosdrecht, M.C.M, Lyklema, J., Norde, W., Schraa, W., and Zehnder, A.J.B. (1987). Electrophoretic mobility and hydrophobicity as a measure to predict the initial steps of bacterial adhesion. Applied and Environmental Microbiology, vol. 53, p 1898-1901. Vane, M. L. and Zang, G. M. (1997) “Effect of aqueous phase properties on clay particle zeta potential and electroosmostic permeability: Implications for electrokinetic remediation processes,” Electrochemical Decontamination of Soil and Water, Special Issue of Journal of Hazardous Material, Vol. 55, No’s. 1-3, pp. 1-22. West, L. J, and Stewart, D. I. (1995). "Effect of zeta potential on soil electrokinesis." Characterization, containment, remediation, and performance in environmental geotechnics. Geotech. Spec. Publ. No. 46, ASCE, New York, N.Y., 2, 1535-1549. Wittle, J. K. and Pamukcu, S. (1993) “Electrokinetic treatment of contaminated soils, sludges, and lagoons,” Final Report, Contract No. 02112406, DOE/CH-9206, Argonne National Laboratory, Chicago, IL. Wong, J. S., Hicks, R. E., and Probstein, R. F. (1997) “EDTA- Enhanced Electroremediation of Metal Contaminated Soils,” Journal of Hazardous Materials Special Edition on Electrochemical Decontamination of Soil and Water, Vol. 55, No. 1-3. Yeh, G. T. and Tripathi, V. S. (1989), “A critical evaluation of recent developments in hydrogeochemical transport models of reactive multicomponent components,” Water Resources Research, 25(1): 93-108. Yeh, G. T. and Tripathi, V. S. (1991), “A Model for Simulating Transport of Reactive Multispecies Component: Model Development and Demonstration,” Water Resources Research, 27(12) 30753094. Yeung, A. T., and Datla, S. (1995), "Fundamental formulation of electrokinetic extraction of contaminants from soil." Can. Geotech. J., 32(4), 569-583. Yeung, A. T., Hsu, C., and Menon, R. M. (1996). "EDTA-enhanced electrokinetic extraction of lead." J. Geotech. Engrg., ASCE, 122(8), 666-673. 94

Yeung, A. T., Hsu, C., and Menon, R. M. (1997). "Physicochemical soil-contaminant interactions during electrokinetic extraction." Electrochemical Decontamination of Soil and Water, Special Issue of Journal of Hazardous Material, Vol. 55, No=s. 1-3, pp. 221-238. Zappi, M, Gunnison, D., Pennington, J., Cullinane, J., Teeter, C. L., Brannon, J. M., and Myers, T. (1993) “Technical Approaches for In Situ Biological Treatment Research: Bench-Scale Experiments,” US Army Corps of Engineers, Waterways Experiment Station, Vicksburg, MS, August 1993, Technical Report No. IRP-93-3.

95