Louisiana State University LSU Digital Commons
LSU Historical Dissertations and Theses Graduate School
1990 Decontamination of Soil Using Electro-Osmosis. Jihad Tawfiq aH med Louisiana State University and Agricultural & Mechanical College
Follow this and additional works at: https://digitalcommons.lsu.edu/gradschool_disstheses
Recommended Citation Hamed, Jihad Tawfiq, "Decontamination of Soil Using Electro-Osmosis." (1990). LSU Historical Dissertations and Theses. 5054. https://digitalcommons.lsu.edu/gradschool_disstheses/5054
This Dissertation is brought to you for free and open access by the Graduate School at LSU Digital Commons. It has been accepted for inclusion in LSU Historical Dissertations and Theses by an authorized administrator of LSU Digital Commons. For more information, please contact [email protected]. INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.
The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book.
Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.
• University Microfilms International A Bell & Howell Information Company 300 North Zeeb Road. Ann Arbor. Ml 48106-13-16 USA 313 761-4700 800 521-0600
Order Number 9123197
Decontamination of soil using electro-osmosis
Hamed, Jihad Tawfiq, Ph.D.
The Louisiana State University and Agricultural and Mechanical Col., 1990
UMI 300 N. ZeebRd. Ann Arbor, MI 48106
DECONTAMINATION OF SOIL USING ELECTRO-OSMOSIS
A DISSERTATION
Submitted to the Graduate Faculty of the Louisiana State University and Agricultural and Mechanical College in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy
in
The Department of Civil Engineering
BY
Jihad T. Hamed M.S., University of Texas at EL Paso, 1986 December 1990 DEDICATED TO MY PARENTS, WHOSE LOVE,
SUPPORT, UNDERSTANDING AND ENCOURAGEMENT NEVER
CEASED TO INSPIRE ME ACKNOWLEDGEMENTS
The author wishes to express his sincere thanks to his major advisor Dr.
Y?Jcin B. A car for his patient supervision during the period in which this study
was carried out.
My appreciation is also extended to Dr. Robert J Gale for his valuable
suggestions and for serving in my committee. I would like to extend my thanks to Dr. Mehmet T. Tumay, Dr. Joseph N. Suhayda, Dr. Dipak Roy, Dr. J.
Dorr oh, and Dr. Bengston for serving in my committee.
I would also like to extend my appreciation to the Board of Regents for awarding this study which provided me the financial assistance necessary to conduct this work.
Finally, the author wishes to express his sincere thanks to his parents, brothers, and sisters for their love, understanding and encouragement throughout his graduate education. Table of Contents
ACKNOWLEDGEMENTS ...... iii
TABLE OF CONTENTS ...... iv
LIST OF TABLES ...... vii
LIST OF FIGURE ...... ix
ABSTRACT ...... xiii
CHAPTER I INTRODUCTION ...... 1 1.1 Scope of Study ...... 3
CHAPTER H LITERATURE REVIEW ...... 4 2.1 Electrokinetic Phenomena in Soils ...... 4 2.2 Review of Theories for Electro-Osmosis ...... 10 Water Transport 2.2.1 Double Electric Layer Theory(Helmholtz) . . . 11 2.2.2 Schmid Theory ...... 11 2.2.3 Spiegler Friction Model ...... 15 2.3 Potential Uses of Electrokinetics in Waste ...... 18 Management 2.4 Synthesis of Available Data ...... 19 2.5 Removal of Contaminants by Electrokinetic ...... 34 Soil Processing
CHAPTER HI MODELLING OF ACID/BASE DISTRIBUTIONS 37 3.1 Analytical Modeling ...... 37 3.1.1 Chemistry by Electrode Reactions and pH Gradients ...... 38 3.1.2 Mass Transfer in the System ...... 42 3.2 Predictions of The Analytical Model ...... 48
CHAPTER IV METHODOLOGY ...... 53 4.1 Testing Program ...... 53 4.2 Permeation Fluid . 58 4.3 Contaminants ...... 58 4.4 Testing Equipm ent ...... 58 4.5 Specimen Preparation ...... 63 4.6 Testing Procedure ...... 64
CHAPTER V ANALYSIS OF RESULTS ...... 67 5.1 Feasibility Tests ...... 67 5.1.1 Pb(II) ...... 67 5.1.1.1 Electrical Potential Gradients and Flow ...... 69 5.1.1.2 Effluent pH ...... 73 5.1.1.3 pH Profiles ...... 73 5.1.1.4 Conductivity ...... 78 5.1.1.5 Efficiency of Pb(II) Removal ...... 85 5.1.1.6 Energy Expenditure ...... 91 5.1.2 Cd(II) ...... 95 5.1.2.1 Electrical Potential Gradients and Flow 95 5.1.2.2 Effluent pH ...... 102 5.1.2.3 pH Profiles ...... 105 5.1.2.4 Conductivity ...... 107 5.1.2.5 Efficiency of Cd(II) Removal .... 109 5.1.2.6 Energy Expenditure ...... 113 5.1.3 CrfTII) ...... 115 5.1.3.1 Electrical Potential Gradients and Flow 115 5.1.3.2 Effluent pH ...... 125 5.1.3.3 pH Profiles ...... 125 5.1.3.4 Conductivity ...... 128 5.1.3.5 Efficimcy of Cr(DI) Removal .... 131 5.1.3.6 Energy Expenditure ...... 134 5.2 Concentration and Current Density Effects ...... 136 5.2.1 Efficiency of Removal ...... 140 5.2.2 Energy Expenditure ...... 149
CHAPTER VI SUMMARY AND CONCLUSIONS ...... 163 6.1 Implications ...... 163 6.1.1 Electrochemistry ...... 163 6.1.2 Efficiency ...... 164 6.2 Conclusions ...... 168
REFERENCES ...... 172
APPENDICES ...... 178
v Experimental Results for Pb(II), Cd(II), and Cr(H) .. 180 APPENDIX A: i APPENDIX B: Two-Dimensional Finite Element Method Formulation for the Analytical Model ...... 208
APPENDIX C: Two-Dimensional Finite Element Method Code for the Analytical Model ...... 218
APPENDIX D: Reprint Permission from Hamed, et al. 1991 .... 230
APPENDIX E: A two Dimensional Finite Element Program for Advective/Dispersive Migration of Reactive Contaminants ...... See Packet
Vita ...... 231
vi List of Tables
Table 2.1 Analysis of Laboratory Data Reported for Removal of Chemicals by Electrokinetics ...... 21
Table 2.2 Synthesis of Field Data Reported for Removal of Chemicals by Electrokinetics ...... 30
Table 4.1 Typical Physical and Chemical Properties ...... 54
Table 4.2 Characteristics of Georgia K aolinite ...... 55
Table 4.3 Summary of Testing Program ...... 59
Table 5.1 Summary of Testing Program and Processing Data ...... 68
Table 5.2 pH Distributions Across the Specimens ...... 75
Table 5.3 Measured Conductivity Values Across the cell ...... 83
Table 5.4 The Ratio of Final to Initial Pb(II) Concentration ...... 86
Table 5.5 Pb(II) Mass Balance in the Specimens ...... 89
Table 5.6 Summary of Testing Program and Processing Data for Cd(H) ...... 96
Table 5.7 Effect of Concentration and Valences of ions on Thickness of the Diffuse Double L a y e r ...... 103
Table 5.8 pH Distributions across the Specimens ...... 106
Table 5.9 Measured Conductivity Values across the Cell ...... 110
Table 5.10 The Ratio of Final to Initial Concenliaticfss, Cf/Cj . . . . I ll
Table 5.11 Mass Balance in the Specimens ...... 112
vii Table 5.12 Summary of Testing program and Processing Date for Cr(III) ...... 116
Table 5.13 Electrical Potential Difference ...... 120
Table 5.14 pH Distributions Across the Specim ens ...... 127
Table 5.15 Measured Conductivity Values Across the Cell ...... 130
Table 5.16 , The ratio of Final to Initial Concentrations, Cf/Cj ...... 132
Table 5.17 Mass Balance in the Specimens ...... 133
Table 5.18 Summary of Testing Program to Study the Effects of Current Density and Concentration ...... 137
Table 5.19 Summary of Testing Program And Processing Data for Pb(II) ...... 138
Table 5.20 The ratio of Final to Initial Concentration, C{/Cl ...... 141
Table 5.21 Mass Balance in the Specimens ...... 143
Table 5.22 Results of Tests conducted to Study the Effects of Current Density and Concentration ...... 154
Table A .l Concentration of Cd(II) ...... 181
Table A.2 Concentration of Cr(H[) ...... 182
Table A.3 pH Distributions Across the Specim ens...... 183
Table A.4 Measured Conductivity Values Across the Cell ...... 186
Table A.5 Concentration of Pb(II) ...... 188
viii List of Figures
Figure 2.1 Electrokinetic Phenomena in Soils ...... 5
Figure 2.2 A Schematic Diagram of the Matrix Potentials ...... 7
Figure 2.3 A Schematic Diagram of One Dimensional Laboratory Tests in Electrokinetic Soil Processing ...... 9
Figure 2.4 Comparison of Electro-osmotic Flow with Hydraulic Flow in a Single Capillary ...... 12
Figure 2.5 Schmid Model for Electro-osmosis ...... 16
Figure 2.6 Schematic View of Different Applications of Electrokinetic Phenomena in Remediation ...... 20
Figure 2.7 Cation Removal Versus Electro-osmosis Flow ...... 26
Figure 2.8 Pb(II) Removal from Kaolinite and Corresponding Energy Expenditure ...... 29
Figure 2.9 A Schematic Diagram of the Field System Reported by Lageman (1 9 8 9 ) ...... 33
Figure 3.1 The pH Gradients Predicted by the Model ...... 49
Figure 3.2 pH Gradients in Electro-osmosis ...... 51
Figure 4.1 Pb(D) Adsorption Isotherm for Georgia kaolinite ...... 57
Figure 4.2 Electro-osmosis c e ll ...... 61
Figure 4.3 A Schematic Diagram of the Consolidation Frame ...... 62
Figure 4.4 A Schematic Diagram of the Electro-osmosis Cell and the Test Set-up ...... 65
Figure 5.1 Electrical Potential Gradient and Total How ...... 70
ix Figure 5.2 Changes in Electro-osmotic Coefficient of Permeability and Coefficient of Water Transport ...... 72
Figure 5.3 Effluent pH versus the Pore Volumes of Flow ...... 74
Figure 5.4 A Comparison of Insitu and Pore Fluid of pH Profiles Across the Cell in Test 1 and 5 76
Figure 5.5 The change in Apparent Conductivity with Time ...... 80
Figure 5.6 A Comparison of Conductivity Profiles across the Cell . . . 84
Figure 5.7 Concentration Ratio Profiles across the Specimens ...... 87
Figure 5.8 Comparison of Pb(H), pH, and Conductivity profiles in Tests No. 1 and 5 90
Figure 5.9 Pore Fluid/In situ Pb(TT) Concentration Ratio after Electrokinetic Processing of Kaolinite ...... 92
Figure 5.10 Energy Expended/Unit Volume versus Pore Volumes of Flow ...... 94
Figure 5.11 The Electrical Potential Gradient ...... 97
Figure 5.12 Total Flow versus Time ...... 99
Figure 5.13 Changes in Electro-osmosis Coefficient of Permeability . . 100
Figure 5.14 Change in Coefficient of Water Transport Efficiency . . . 101
Figure 5.15 Effluent pH versus the Pore Volumes of F lo w ...... 104
Figure 5.16 The Change in Apparent Conductivity with T im e ...... 108
Figure 5.17 Energy Expended/Unit Volume versus Pore Volumes of Flow ...... 114
Figure 5.18 Electrical Potential Gradient .... 117
Figure 5.19 Voltage Across the Cell versus Normalized
x Distance from Anode ...... 119
Figure 5.20 The Total Flow versus Time ...... 122
Figure 5.21 Change in Electro-osmosis Coefficient of Permeability . . 123
Figure 5.22 Change in Coefficient of Water Transport Efficiency ...... 124
Figure 5.23 Effluent pH versus the Pore Volumes of Flow ...... 126
Figure 5.24 The Change in Apparent Conductivity with T im e ...... 129
Figure 5.25 Energy Expended/Unit Volume versus Volumes of Flow . 135
Figure 5.26 Concentration Ratio Profiles across the Specimens ..... 145
Figure 5.27 Energy Expended/Unit Volume versus Pore Volumes of Flow ...... 150
Figure 5.28 Maximum Electro-osmosis Coefficient of Permeability . . . 157
Figure 5.29 Average Pore Fluid Conductivity at the end of each Test . 158
Figure 5.30 Maximum Electro-osmosis Water Transport Efficiency . . 160
Figure 5.31 The Power per Unit Volume of Soil versus Current Density 162
Figure A .l Voltage across the Electrodes versus T im e ...... 190
Figure A.2 Total Flow versus Time ...... 193
Figure A.3 Change in Electro-osmosis Coefficient of Permeability versus Time ...... 196
Figure A.4 Change in Coefficient of Water Transport Efficiency versus Time ...... 199
Figure A.5 Effluent pH versus the Pore Volumes of Flow ...... 202
Figure A.6 The Change in Apparent Conductivity with Time ...... 205
xi Figure B .l Element Formulation and the Shape Function * . 211
Figure C .l Flow Chart of the Main Program ...... 221
Figure C.2 Flow Chart for Subroutine INPUT ...... 223
Figure C.3 Flow Chart for Subroutine ELSTIF ...... 224
Figure C.4 Flow Chart for Subroutine GENER ...... 226
Figure C.5 How Chart for Subroutine SOLVES ...... 227 Abstract
Electrokinetic soil processing is an innovative in situ technique to remove
contaminants from soils and groundwater. The process is an alternative to
conventional processes, with significant economic and technical advantages.
A theory is presented for pH gradient development by diffusion and
convection in electrokinetic processing of soils. The premises and consequences
of the theory are discussed. Based upon this theory, a mathematical model is
developed to predict the pH gradient development in electro-osmosis.
Electro-osmosis tests are conducted on saturated kaolinite specimens loaded
with Pb(H), Cd(D), and Cr(HI) to investigate the feasibility, efficiency, and energy
requirements of the process in removal of heavy metals. The test results pertaining
to the flow and the associated electrochemistry (voltage, current, pH gradients , and conductivity) are presented. Tests indicated that the flow in electro-osmosis with open electrodes is time-dependent and it is strongly influenced by electrochemistry generated as a result of the prevailing pH gradients.
High removal efficiency, up to 95 %, was achieved on kaolinite specimens loaded with Pb(II) up to 1000 /xg/g of soil. Also the removing efficiency was very high (92 to 100 %) for kaolinite specimens loaded with Cd(TI) up to 120 mg/g of dry soil. However, tests conducted on kaolinite loaded with Cr(HI) (120 /xg/g of dry soil) indicated that only 60 to 70 % of adsorbed Cr(lll) was removed from the anode zone. The influence of fundamental variables affecting the chemistry and efficiency of the process such as the current density, the voltage gradient and
contaminant concentration are also investigated.
The results indicated that the process can be used efficiently to remove ions from saturated soil deposits. The cost of the process will be a function of the current density and the processing time. In the tests conducted, the energy expenditure varied between 11 to 306 kW-hr/m3.
The study demonstrates the feasibility of using this process in decontamination and provides a fundamental understanding to the underlying mechanisms controlling the process.
xiv CHAPTER I
INTRODUCTION
Contaminants migrating from unengineered facilities, hazardous waste sites, accidental spills, and industrial operations threaten the environment and pose a hazard to the public as potential sources of contamination of groundwater. Such contamination often involves large volumes of soil underlying several acres of surface area. The remediation of these areas is very costly and time consuming; therefore there exists an everlasting need to find new economical and efficient technologies of remediation for rapid reclamation and rehabilitation of such sites.
The electrokinetic phenomena in soils is envisioned to be used for removal/separation of organic and inorganic contaminants and radionuclides, barriers and leak detection systems in clay liners, diversion schemes for waste plumes, and for injection of grouts, microorganisms and nutrients into subsoil strata (Acar and Gale, 1986; Mitchell, 1986; Renauld and Probstein, 1987; Acar, et al., 1989; and Steude, et al., 1989).
In the five decades since its first application and use (Casagrande, 1947), the mechanics of consolidation by electro-osmosis has been extensively investigated by geotechnical engineers. However, studies investigating removal
1 2 of ions from soils by the electrokinetic phenomena are limited, possibly due to insufficient understanding of the electrochemistry associated with the process. The need to utilize the process in removal/separation of contaminants necessitates a good understanding of electrochemistry and its relation to the mechanical behavior. The purpose of this study is to provide an understanding to the electrochemistry associated with the process and to assess the feasibility of removing inorganic contaminants adsorbed on fine-grained soils.
The following objectives are identified for this study.
1- A preliminary analytical model will be developed for the development of acid/base distributions in electrokinetic soil processing. The predictions of this model for pH gradients across soil specimens will be compared with available experimental results ,
2- The feasibility of electrokinetic removal of selected inorganic contaminants adsorbed on fine-grained soils will be investigated under controlled, one dimensional laboratory experiments,
3- The influence of fundamental variables affecting the chemistry and the efficiency of the process such as the current density, the voltage gradient, and the contaminant concentration will be investigated. Their effect on the feasibility and efficiency of the process will be evaluated. 3 1.1 SCOPE:
Three inorganic contaminants are selected Pb(II), Cd(II), and Cr(III). These inorganic ions are typical of heavy metal contaminants in the environment. In this study, Georgia kaolinite is used because it is a low activity soil and it has a high electro-osmotic flow efficiency. The study is focused on only one dimensional laboratory tests since a fundamental understanding of the process is not yet developed. The study is also focused on studying the effect of fundamental variables (current density and concentration) influencing the efficiency of the process. Tests are conducted on soils contaminated with different concentrations of Pb(II) ranging from 7.0 percent to 100.0 percent of the cation exchange capacity of the kaolinite. The influence of current densities ranging from 0.012 to
0.37 mA/cm2 is also investigated. The results of tests conducted in this study together with the understanding developed on the fundamentals of the process demonstrate the feasibility and efficiency of the soil processing technology. CHAPTER H
LITERATURE REVIEW
In order to carry out the objectives of this study, available data on electrokinetics, laboratory experiments, and field studies are reviewed. Evaluation of the past studies on electrokinetic soil decontamination yields inconclusive and sometimes inconsistent results. However, recent studies by Lageman (1989) and early results of this study ( Hamed, et al. 1991) demonstrate that electrokinetics can be efficiently used in soil decontamination. A summary of all literature and research studies related to the fundamentals of electrokinetics phenomena and its application in soil processing is presented below ( Acar and Hamed, 1991).
2.1 Electrokinetic Phenomena in Soils
Coupling between electrical, chemical and hydraulic gradients is responsible for different types of electrokinetic phenomena in soils. These phenomena include electro-osmosis, electrophoresis, streaming potential and sedimentation potential, as shown in Figure 2.1. Electro-osmosis and electrophoresis are the movement of water and particles, respectively, due to application of a low DC current.
Streaming potential and sedimentation potential are the generation of a current due to the movement of water under hydraulic potential and movement of particles under gravitational forces, respectively. The effect of coupling becomes more important in fine-grained soils with lower coefficients of permeability.
4 5
+ ' - AE AE Water Movement Particle Movement »-
/Saturated Clay;- Cloy Suspension
ELECTRO-OSMOSIS ELECTROPHORESIS
SL r AH c (U E a> > Water Movement o Q.
a> CO o :';;:';'S a tu rated ClayV.lv o CL
MIGRATION POTENTIAL
STREAMING-POTENTIAL
Figure 2.1 Electrokinetic Phenomena in Soils (adapted from Mitchell, 1976) 6 In electro-osmosis, electrodes can be placed in an open or closed flow arrangement. Open flow arrangement constitutes the case when an electrode is sufficiently permeable to admit ingress and egress of water. In the closed flow arrangement, the electrode is not permeable or porous. Different electrode configurations ( open or closed) result in substantial variations in the total matrix potentials across the soil specimen.
Figure 2.2 presents a schematic diagram of the matrix potentials (pressure head) developed in a soil specimen under constant current (or voltage) conditions.
In Case I of Figure 2.2, the water in a reservoir behind the anode would permeate through the soil to displace the pore fluid. In Case II, where ingress of water is prevented at the anode, negative matrix potentials are generated in the soil increasing the confining stress proportionally until equilibrium conditions are reached by the flow of water from the anode to the cathode. This arrangement has often been used by geotechnical engineers to consolidate soil deposits. In
Case ID, positive pore pressures are developed reducing the confining stress which may potentially lead to swelling. In Case IV, the matrix potential drops at the anode and increases at the cathode. The water level will rise in the cathode side, and drop at the anode side
The electro-osmotic flow rate, qc, is defined with an empirical relationship,
qe = KeieA = Kjl = K,. I 7
Cl Valve C2 1
Free 0*0 V-saturated S Water
Anode + - Cathode
ELECTRICAL POTENTIAL,
MATRIX POTENTIAL, \f/m
J. A-Open, B-Open , C-Closed m . A-Open , B-Closed , C-Closed '//m2
JH.A-Closed, B-Open, C-Closed U. A-Closed, B-Closed, C-Open 'Kni
Figure 2.2 Schematic Diagram of Matrix Potential Generated With Different Electrode Configurations ( Acar, et al., 1989 ) 8 where K,, = coefficient of electro-osmotic permeability (cm2/sec-V), k; = electro- osmotic water transport efficiency (cm3/amp-sec), I = current (amp), a = conductivity (siemens/cm), i,. = electrical potential gradient (V/cm), A = cross- sectional area (cm2). Estimates of electro-osmotic flow rates can be made using equation (1). K,. varies within one order of magnitude for all soils; 1 x 10'5 to
10 x 10*5 (cm/sec)/(V/cm), the higher values being at higher water contents.
Figure 2.3 presents a schematic diagram of one-dimensional laboratory tests in electrokinetic soil processing. The prevailing electrical gradients, and the ion flow are also depicted. A comparison of flow under electrical and hydraulic gradients in clays is also provided. The coefficient of electro-osmotic permeability is independent of the size and distribution of pores (fabric) in the soil mass.
However, hydraulic conductivity is most affected by the fabric. In general, hydraulic conductivity decreases by five to six orders of magnitude (10*3 cm/sec to 10*® cm/sec) from fine sands to clays. Figure 2.3 indicates that under equal gradients, electrical potentials in fine-grained soils would result in three orders of magnitude larger flows than hydraulic potentials. Therefore, electro-osmosis induced flow can be considered to be an efficient pumping mechanism in saturated, low permeability, fine-grained soils.
The efficiency and economics of electro-osmotic dewatering is governed by the amount of water transferred per unit charge passed, which is quantified by the electro-osmotic water transport efficiency, K;. The parameter may vary over 9
Gos Venl Gos Venl t || Outflow
Soturoled Specimen
<♦) EO Flow (-) Anode Cothode Electro-Osmotic Flow, Q0 Inflow Qe " ke * ‘e ‘ 'e I * Constont ke « electro-osmotic permeability
CURRENT,I i0 = electrical gradient A = area
ELECTRIC V, Hydraulic Row, Qh POTENTIAL ,
m otru) ♦ '/'j, (tuition ) ih « gradient — ( comtoni), ^ (vonobie ) Ratio of Two Flows HYDRAULIC POTENTIAL, i/' i Qe V e Qh kh 'h A Comparison of Two Flows in Clays C olloid
ION PLOW k0 « 1 x 10 '5 (crtVsec)/(v/cm) (IN ITIAL) kj, « 1 x 10-* cm/sec I ■ 0 i0 =* 1 v/cm (typical for field application) i^ - 1 (selected for comparison) ------(HjO Oe V - ' p jo h ij ^ S - 2 - 1,000 ION FLOW
Figure 2.3 A Schematic Diagram of Electrokinetic Processing, Ion Row and Comparison of Row in Clays ( Hamed, et al., 1991) 10 a wide range from 0 to 1.2 cm3/amp-sec depending upon the electrical
conductivity of the porous medium. The conductivity changes with water content,
cation exchange capacity and free electrolyte content in the soil and also due to the
prevailing chemistry during electrokinetic processing. An approach to quantifying
this amount in soil-water-electrolyte systems is proposed by Gray and Mitchell
(1967). Their study indicated that electro-osmotic efficiency decreases with a
decrease in water content and an increase in activity of the soil. The electro-
osmotic dewatering efficiency is independent of variations in electrolyte
concentration (sodium ion) for active clays, while an increase in electrolytes tends
to decrease the efficiency in inactive clays.
Recent studies by Lockhart (1983) substantiate the conclusions of Gray and
Mitchell (1967). IQ increased from 0.32 to 1.20 cm3/amp-sec with the decrease of NaCl and HC1 concentrations from 10'1 M to 10-3 M. IQ decreased by an increase in electrical gradients, possibly due to a higher influx of H+ ions
(Hamed, et al., 1991). Lockhart (1983) shows that a higher electro-osmotic efficiency is recorded with H and Cu clays.
2.2 Review of Theories for Electro-Osmotic Water Transport:
Reuss in 1809 was the first to discover that when an electric potential is appbed to a porous medium, water moves through the capillaries from the positive side (anode) to the negative side (cathode); and that when the electric potential is 11 switched off, the flow of water stops. This phenomenon has been called " electro osmosis
Electro-osmosis technique have been used since the 1930’s in solving a variety of soil stabilization and foundation problems. It has been of interest to geotechnical engineers for the last five decades as a technique for ;1- improving stability of excavations (Chappell, et al., 1975), 2- increasing pile strength
(Butterfield, 1980), 3- stabilization of fine grained soils (Mitchell, et al., 1977),
4- removal of metallic objects from the ocean bottom ( Esrig, et al., 1966), 5- decreasing pile penetration resistance (Begemann, 1953; Thompson, 1971), 6- increasing petroleum production (Amba, 1964), 7-determination of volume change and consolidation characteristics of soils (Baneijee, etal., 1980), 8- separation and filtration of materials in soils and solutions (Yukawa, et al., 1976).
Over the years, a number of theories have been proposed to explain the mechanism by which the pore water is transported; namely, the Ionic transport theory, the Helmholtz-Smoluchowski model, Schmid model, and the Spiegler friction model. These theories provide a rational explanation of how the pore fluid is transported in electro-osmosis. A brief review is presented for background.
2.2.1 Double Electric Layer Theory (Helmholtz)
In a cylindrical capillary tube filled with water, Helmholtz assumed the existence of opposite electric charges on two layers forming together the double 12 layer, as shown in Figure 2.4a. A mathematical theory for electro-osmosis flow in a cylindrical capillary tube was derived. The following assumptions have been made:
1. The liquid containing excess positive ions is oppositely charged to the rigid wall of the capillary tube, forming an electric double layer along the wall,
2. The thickness of the double layer is very small compared with the linear dimensions of the cross section of the tube,
3. The outer part of the double layer which is in contact with the wall is very thin and it is not movable. It is fixed to the wall regardless of the force imposed on the liquid in the tube,
4. The much thicker inner part of the double layer which carries the positive charges was termed by Helmholtz the movable part; movable, because if an electric potential is applied to the capillary , the positive charges move toward the negative pole. In doing so, they drag along the water molecules of the movable part of the double layer,
5. Only a laminar flow of liquid can occur, so that the line of flow is parallel to the axis of the tube,
6. The wall is an insulator and the contained liquid possesses the property of electrolytic conductance. If we assume that no other forces are acting upon this cylinder of free water its rate of flow will be constant and the velocity distribution will be as shown in Figure 2.4a where ; Resisting Force
+ + + + + i> Double Layer Double Layer Resisting Force
Moving F orce
Free Water Velocity Moving F ree W a te r V elocity Force
Double Layer Double Layer Resisting Force
Resisting Force
(a) Electroosmotic Flow (b) Hydraulic Flow
Figure 2.4 Comparison of Electro-osmotic Flow With Hydraulic Flow in A Single Capillary ( Casagiande, 1951) 14
(2.2) 4tct]
Therefore in a capillary tube, the quantity of liquid moved in a unit of time by electro-osmosis can be computed from the following equation.
E = the electric potential, D = the dielectric constant of the liquid, r = the radius of the capillary, f = the zeta potential ,which is defined as the potential existing between the rigid and the movable parts of the double layer, rj = the viscosity of die liquid, and L = the length of the capillary between the electrodes.
If we let ~D£/4r)ir =cl ; where q is a constant, and xr2 = a cross section area of capillary, ie= E/L = electric gradient, then the electro-osmotic flow will be given by ;
(2.4)
(2.5)
In a laminar flow regime, the rate of electro-osmotic flow for a bundle of N straight capillaries in the total cross sectional area A ,and with void ratio e, would be ; 15
„ , r A e (2.6) Q^ Nqm r e c'''a
Q rci-rz‘A =kJA (2.7) l+e
from the above equation it follows that electro-osmotic coefficient of permeability,
Ke , is independent of the size of the capillaries. The dimensions of Ke are
(cm2)/(volts)(sec.), Since IQ is independent upon the pore size of the soil, it follows that the zeta potential does not change very much within the ordinary range of concentration of electrolytes as found in soils. However, for high concentrations of electrolytes, it is possible that the zeta potential can drop to zero leading to a decrease in electro-osmotic water transport.
5.2.2 Schmid Theory :
Schmid (1857) ignored the existence of the double layer and proposed that the counter ions are uniformly distributed throughout the entire liquid volume of the pore or capillary. As a result of this assumption the electrical forces act uniformly over the entire pore cross section area rather than just along the double layer as in the Helmholtz theory. The resulting flow distribution is parabolic with zero velocity at the wall as shown in Figure 2.5. Therefore, the Schmid theory brings only an update to the velocity profile in the capillary. 7& //77?//777///777/,777//777//777//777//7777/S777//777///Z
Velocity Profile Net + Force
?22Z77ZZZ772ZZ77ZZZ77ZZZ?7ZZZZ77ZZZZ7ZZZZ7ZZZZ77ZZZ77ZZZ?7ZL
Figure 2.5 Schmid Model for Electro-osmosis 17
2.2.3 Spiegler friction model
A completely different model is presented by Spiegler (1958) which describes the flow of cations, anions, and water through a permselective porous plug.
Spiegler derived the following equation for electro-osmotic transport of water across an ion exchange membrane.
Q=W-H= C 3- Cj+C3* 3 4 <2 -8 > * 3 1
where ;
0 = the true electro-osmotic flow ( mole/Faraday )
W = total moles of water transported per Faraday
H = ion hydration, moles/Faraday
C3 = the total (analytical) water concentration in the membrane .
Q = concentration of mobile counter ions in the membrane, mole/sec.
X34 = friction coefficient between water and solid wall.
X13 = friction coefficient between cation and water.
The assumptions Spiegler used in derivation of equation (2.8) were as follows :
1. Exclusion of co-ions; that the medium admitting ions of only one sign.
Therefore the friction coefficient between cations and anions are neglected, 18
2. Complete dissociation of pore fluid ions.
The coefficients X34 ,X31 can be calculated by independent measurements of self diffusion of cation and anion respectively, the transference number, the specific conductance and electro-osmotic water transport. So, the above equation is limited in its usefulness as a predictive equation. But the real value of the above equation lies in providing a relatively simple physical picture for complex processes that occur during electro-osmosis. So, the following conclusion can be made regarding the qualitative predictions of this model;
- At high water contents and for large pores, the friction term will vanish by X34 tending to zero, then equation 2.8 became;
0 = C3/C! (2.9)
Equation (2.9) indicates that as the ratio of water to cations increases, the electro- osmotic flow will increase.
A major limitation of the above three theories is their exclusion of the changing chemistry due to time-dependent changes due to electrode reactions.
2.3 Potential Uses of Electrokinetics in Waste Management
The four electrode configurations given in Figure 2.2 could potentially be 19 used in the following ways in waste disposal (Acar, et al. 1989): (1) Dewatering of waste sludge slimes, dredged spoil by first concentrating the solid particles using electrophoresis and subsequent consolidation by electro-osmosis (Mitchell,
1986), (2) Electro-osmotic flow barriers (Mitchell and Yeung, 1991), (3) Leak detection systems for disposal facilities, (4) Injection of grouts to form barriers
(Esrig, 1967), (5) Provide nutrients for biodegrading microcosm, (6) Insitu generation of reactants such as hydrogen peroxide for cleanup and/or electrolysis of contaminants (AFESC, 1985), and (7) Decontamination of soils and groundwater. Figure 2.6 conceptualizes electrokinetic clay barriers, waste plume diversion schemes, and electro-osmotic injection.
Insitu remediation methods often necessitate the use of hydraulic charge and recharge wells to permeate the decontaminating liquid or stabilization agent through the soil deposit, or to provide nutrients for the biodegrading microcosm.
Although such systems may effectively be used in highly permeable soils, they become inefficient and uneconomical in low permeability silts and clayey deposits.
Electrokinetic soil processing with open electrode configuration could also be used to achieve an efficient seepage and decontamination method in such soils.
2.4 Synthesis of Available Data
Studies investigating removal of ions from soils by electro-osmosis are rare possibly due to difficulties in understanding the chemistry. Table 2.1 provides a FLOW BY HYDRAULIC AND CHEMICAL POTENTIALS DIVERSION 1 1 I I I rmSr^iriv/r ^ ^ ------tt>roQmwr CLAY LIN ER f t M I OPPOSING FLOW BY ELECTROKINETIC POTENTIALS WASTE PLUME MIGRATION tUidi ' DIRECTION
(a) Flow Barriers CHARGE (b) Plume Diversion Scheme GROUT DISCHARGE NUTRIENTS CHEMICALS 4 © | © Trrirmr jjmvmt rtiwfrr rvrmrTT7rr7T
7 i
E.O FLOW
(c) Electro-osmotic Injection
Figure 2.6 Schematic View of Different Applications of Electrokinetic Phenomena in Remediation
to o Table 1. Analysis of laboratory data reported for removal of chemicals by electrokinetics
Concentration Current Density Charge Soil Typsi l/rg/gl and/or Voltage Duration Energy Remarks Chemical ImA/cm'l or (hr) amp-hr kWh/m3 Initial Final IV/cm) ' m3
1. Puri and Anand (1936) A Bucher funnel was used in testing. The diameter w as 18 in. Cathode is circular brass plate. Anode High pH Soil - Na* N/A N/A 9.3-14.9 8 N/A N/A consists of five cylindrical bars arranged symmetri (20.01 (inter cally along the circumference. Na normality of mittent) percolate increased up to 0.6 N. The effluent was 90% NaOH. 10% Na3C03.
2. Jacobs and Mortland (1959) 1-D tests. Cylindrical specimens ID = 0.75 in.. L = 1 in.). Circular platinum electrodes. Rate of 5% Bentonite/95% Sand removal of monovalent ions were directly related to the amount remaining in the specimens. The rate
Na-Ca . i of removal w as in the order of Na* > K* > Mg3* Na-Ca-Mg .5 9 /5 7 0.0/IN/A) 0 .32-0.64 10-140 2-20 N/A > Ca3*. Na* was removed more efficiently than all Na-K .4 4 /.3 /.4 0.0/IN/AI IN/A] other ions. Tests were discontinued when most Na .65/.41 0.0/IN/A) Na* w as removed. Concentrations reported are in Ca 0.78-1.11 0 .0 symmetry units. K 0.8 6 -1 .0 .26-0.30 0.79 0 .0
3. Krizek, et al. (1976) Slurries (w = 100 - 142%) from discharge pipes and contaminated bottom sediments are tested. Slurry/Sediment Cylindrical 1-D consolidation tests (D = 14 cm, L = 25 cm). The concentrations are the initial and No. 1 - Na 320 330 I0.5I 150 N/A 20 final effluent values. K 7 290 Ca 65 640 n h 3-n 138 320
No. 2 - Na 360 205 11.01 150 N/A 65 K 15 160 Ca 340 5800 n h 3-n 172 456 Table 1 (continued)
Concentration Current Density Charge Soil Type/ (mj/oI and/or Voltage Duration Energy Remarks Chemical (mA/cm'l or (hr) amp-hr kWh/m3 Initial Final IV/cml „3
4. Hamnet (1980) A 3-D laboratory model. 35 cm by 8 cm, with 5 cm depth. Carbon rods were used as electrodes. Silica Sand - NaCI 3% N/A 6.25-16.25 12-20 N/A N/A After 12 hours, concentration at anode was a third Heavy Day 10.61 of the cathode side. The data shows movement of ions to respective electrodes.
5. Runnels and Larson (198S) 1-D tests. Cylindrical specimens (D = 0.75 in.. L = 6 in.). Square platinum electrodes (1 in-). Silty Sand - Cull!) 617 290-543 0.01-0.05 24-72 N/A N/A Quartz sand washed with HCI to remove impurities. 1.165)
6. Renauld and Probstein (1987) 1 -D cylindrical spacm ens. 30 cm in length. 8 cm in width, and 5 cm in depth. Tests were conducted Ksolicits - Acetic Acid to assess electro-osmotic water transport efficien 0.5-1.3 N/A 1.47 6.7 328 11.5 cy. Efficiency increased with increasing concentra (1.161 tion of acetic acid.
7. Thompson (1989) A tubular, 3 section test set-up is used. The middle section contained the chemical in solution, the Ottawa Sand other two sections contained the soil specimen, and contaminant movement into the cathode section SiO, flour N/A was monitored. Transport was a function of pH of Cu (N03j, 0.01 N N/A 1251 336 N/A the soil.
8. Lageman (1989) Details of experiments are not available. From a presentation by Lageman (1990). it is understood Peat that 1-D tests are conducted with rectangular electrodes. The pore fluid at anode/cathode com Pb 9000 2400 N/A N/A N/A 101 partment w as flushed with a conditioning fluid Cu 600 200 N/A N/A N/A 101 intermittently in order to control the chemistry et the electrodes and sustain the advection in decon Pottery Day - Cu 1000 100 N/A N/A N/A 25 tamination.
Fine Dayey Sand - Cd 275 40 N/A N/A N/A 198
tot o Table 1 (continued)
Concentration Current Dentity Charge Soil Type/ (po/ol and/or Voltage Duration Energy Remarkt Chemical (mA/cm’l or (hr) amp-hr kWh/m3 Initial Final IV/cml m3
S. Lageman (1989) (cont.)
Clay • At 300 30 N/A N/A N/A 207
Fine Clayey Sand
Cd 319 <1 N/A N/A N/A 54 Cr 221 20 Ni 227 34 Pb 638 230 Ha 334 110 Cu 570 50 Zn 937 180
River Sludge
Cd 10 5 N/A N/A N/A 180 Cu 143 41 Pb 172 80 Ni 56 5 Zn 901 54 Cr 72 26 Ha 0 .5 0 0 .2 0 At 13 4 .4
U>N) Table 1 (continued)
Concentration Current Density Charge Soil Type/ (po/s ) and/or Voltage Duration Energy Remarks Chemical ImA/cm’l or (hr) amp-hr kWh/m3 Initial Final IV/cml m3
9. Banerjae, et al. (1990) Eight cylindrical 1-D tests were conducted on specimens brought from the field (D = 5.1 cm. L = Silty/Silty Clay - Cr 2460 22 N/A 24-168 N/A N/A 2.5 cm to 6.7 cm). Electrodes used were Ni-Cu 2156 12 10.1-1.0) wire mesh. Hydraulic and electrical potentials were 870 50 applied simultaneously in orderto facilitate removal. 704 19 642 22 532 37 234 3 148 10
10 Hamed. et al. (1990) 1-D tests. Cylindrical specim ens (D = 4 in.. L = 4in. and 8 in.). Circular graphite electrodes. Initial Georgia Kaolinite - Pb(li) 118-145 7-40 0.0 3 7 100-1285 362-2345 29-60 conductivity of specimens 75-86 //s/cm. Rose up I s 2.5) to 1000 //s/cm at the anode, dropped to 22 //s/cm at the cathode after the process. Pb(ll) movement and electrochemistry across the specimens are reported.
11 Mitchell and Yeung (1990) N/A N/A N/A N/A N/A N/A Investigated the feasibility of using electrokinetics to stop migration of contaminants. Electric field slowed down the migration of cations and increased the movement of anions, k. did not display a marked change by an increase in backpressure, molding water content and dry density.
N>-P>- 25 synthesis and analysis of laboratory studies which reported some form of data related to ion removal from soils. One of the earlier studies is by Puri and Anand
(1936), where leaching of Na+ ions are detected in the effluent in electro-osmotic consolidation. Puri (1949) suggested that in electro-osmosis monovalent ions will move faster than divalent ions due to the former’s higher dissociation from the clay surface. It is also noted that movement of ions was low at low water contents and significantly increases by an increase in water content.
Jacobs and Mortland (1959) demonstrated that Na+, K+, Mg++ and Ca+ + ions can be leached out of Wyoming bentonite by electro-osmosis. Figure 2.7 presents the amount of the ions removed versus the electro-osmotic flow in bentonite-sand mixtures. Monovalent ions are removed at a faster rate.
Krizek, et al. (1976) showed that the soluble ions content substantially increased in effluent in electro-osmotic consolidation of polluted dredgings, while they also noted that heavy metals were not found in the effluent during the period they applied the process.
Hamnet (1980) studied the reclamation of agricultural soils by removal of unwanted salts by electro-osmosis. Hamnet’s tests demonstrated that Na+ ions move toward the cathode, while Cl' and S03'2 ions move toward the anode.
Shmakin (1985) notes that the method has been used in the Soviet Union since the early 1970’s as a method for concentrating metals and exploring for minerals in deep soil deposits. Shmakin (1985) mentions its use in prospecting for ELECTROOSMOSIS FLOW ml 0 - 50 60 20 - 30 - 40 0 h 10 0 02 . 06 . 10 . 14 . 1 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0 Figure 2.7 Cation Removal versus Electro-osmotic Flow Flow Electro-osmotic versus Removal Cation 2.7 Figure A IN REMOVED,CATIONS me. Jcb adMrln, 1959) Mortland, and (Jacobs 8s Ca.86s .s Ca 1.0s / .9 K 0.79s .8 Na 0.78s 8 26 27 Cu, Ni, Co, Au. A porous ceramic probe with HN03 is placed at the cathode.
The migrating ions are extracted with this probe. The quantity of the extracted metal at the cathode and the rate of accumulation is correlated with the composition of the ore and the distance of the sampling locations to the ores.
The potential of the technique in waste remediation resulted in initiation of several recent studies. Runnels and Larson (1986) have investigated the potential use of electromigration to remove contaminants from groundwater. The amount of copper removed increased with processing time (total charge passed).
However, the current efficiency decreased as the processing time increased, possibly due to the increase in conductivity as noted by Hamed, et al. (1991).
Renauld and Probstein (1987) investigated the change in the electro-osmotic water transport efficiency of kaolinite specimens loaded with acetic acid and sodium chloride. This study indicated that the current efficiency increased with higher concentrations of this weak, organic acid. This implies organic acids may increase the efficiency of electrokinetic soil processing.
A better understanding of the chemistry in electrokinetic soil processing is achieved by studies at LSU. Putnam (1988) investigated the development of acid/base distributions in electro-osmosis. Acar, et al. (1990) and Acar, et al.
(1991) present the theory for acid/base distributions in electro-osmosis and compare the predictions of this model with the results of the tests conducted by
Putnam (1988). A good correlation was noted. This theory and the model 28 presented by Acar, et al. (1991) describe the movement of different species in electrokinetics. The study demonstrates the movement of the acid front by advection and diffusion and provides the fundamental basis of the chemistry developed during the process.
A comprehensive subsequent study on removal of Pb(II) from kaolinite is reported by Hamed, et al. (1991) (this study is a part of the results of the work presented in this dissertation). Kaolinite specimens were loaded with Pb(II) at 118
Hg/g to 145 /xg/g of dry kaolinite, below the cation exchange capacity of this mineral. As presented in Figure 2.8, electro-osmosis removed 75 to 95 percent of Pb(H) across the test specimens. The study clearly demonstrated that the removal was due to migration and advection of the acid front generated at the anode by the primary electrolysis reaction. The energy used in the study to decontaminate the specimens was 29 to 60 kwh per cubic meter of soil processed.
This study also demonstrates and explains the complicated electrochemistry associated with the process. An interesting finding of this study is electroplating of Pb(II) at the carbon cathode.
Other laboratory studies conducted by Lageman (1989) and Baneijee, et al.
(1990) further substantiate the applicability of the technique to a wide range of contaminants and soils.
While the above laboratory studies display the feasibility of using electro osmosis to decontaminate soils, limited field studies are available. Table 2.2 Figure 2.8 Pb(II) Removal from Kaolinite and Corresponding Corresponding and Kaolinite from Removal Pb(II) 2.8 Figure
ENERGY EXPENDED PER UNIT VOLUME OF SOIL ( kw-hr/m3) 20 30 50 60 <02 z •v 04 I < £00 £00 O <* O u I J 2 J 0 nry xedtr (Hmd e l, 1991) al., et Hamed, ( Expenditure Energy 06 A - • - - “ ■ - “ . . . . . - - o o (b) Energy Expenditure Energy (b) 01 '1 1 "'■1 at No Tatt
5 7 1 1 a 3 4 4 © 3 © 2 • 1 a bI) Removal Pb(II) (a) o IE0 ITNE RM ANOOE FROM DISTANCE 0 LIZE A M R O N 0 9 9 - 4 02 OE OUE O FLOW OF VOLUMES PORE 1 60
et I Te»t o t i 9 C , , C 03 0 5 4 1 9 2 2 1 6 3 3 1 3 4 1 ifiq/q _ i _ ■
7 O 9 1 04 )
0 • 03 ■ ■
ft 0 • 06
A e 0 29 • 07 *>/ j 1
,.111 s A •
08 1 —o— o 1—
I A • 09 ■
4 0 A • 0 10 . . Table 2. Synthesis of field data reported for removal of chemicals by electrokinetics
Concentration Current Density Soil Type/ and/or Voltage Duration Energy Remarks Chemical (mA/cm2) or (hr) kWh/m3 Initial Final IV/cml
1. Puri and Anand (1936) An area of 4.5 m x 4.5 m was first trenched all around, 1.05 m in depth and 0 .3 0 m in High pH Soil - NaOH 4.8 2.8 1.35 N/A width. Anode and cathode were laid hori 12.441 zontally within this area. Anode was a shot of iron, 0.9 m x 1.8 m laid at top. Cathode was a perforated iron tube, 0.10 cm in dia meter and 1.8 m long laid at 0.3 m depth. The concentrations reported are for the exchangeable Na in the fop 7.5 cm of the soil.
2. Case and Cutshall (1979) An area of 11 m by 5 m was investigated. 1.7 m stainless steel rods were driven in an Alluvial deposits - ®°Sr 50 0 .3 5,352 N/A arc-plus-center point array. The arc con 10.051 sisting of 25 anodes and a central cathode. The concentrations reported are for the effluent in a monitoring well.
3. Segall, et al. (1980) Concentrations noted are for the effluent in electro-osmotic consolidation of dredged Dredged Material material compared to that of water leached specimens. The distance between electrodes Cd, Zn, Pb, As, Fe <1 0 .2-20 [0 .01 - 1.01 N/A N/A is 3-5 m. Na 61 5 0 16300 K 350 510 OH 0 5 9 5 0 h c o 3 ’ 0 9 79 Organic Nitrogen 4 15 Ammonia Nitrogen 67 128 TOC 3 2 0 0 0
UiO Table 2 (continued)
Concentration Current Density Soil Type/
4. Lageman (1989)
Sandy Clay - Zn 70-5120 30-4470 0.8 1344 287 An area of 15 m by 6 m studied. Contami 10.4-0.2! nation depth: 0.40 m. Temperature rose from 12°C to 40°C. Conductivity increased from 2000 //s/cm to 4000 //s/cm. Voltage gradient decreased. 2 cathodes (vertical) at 0.5 m depth. 33 anodes (vertical), 3 rows at 1.0 m depth. Distances: cathode-anode = 1.5 m; anode-anode = 1.5 m.
An area of 10 m by 10 m was studied within Heavy Clay - As 90-385 20-240 0 .4 1200 270 a depth of 2 m. Cathodes (vertical), 2 rows: 10.4-0.2] 1 row at 0.5 m depth; 1 row at 1.5 m depth. 36 anodes (vertical), 3 rows at 2 m depth; 2 rows of 14; 1 row of 8, Distances: cathode- anode = 3 m; anode-anode = 1.5 m. Tem perature rose from 7°C to 50°C.
An area of 70 m by 3 m was studied to a Dredged Sediment depth of 0.2 m to 0.5 m. Cathode is laid horizontal, anode (vertical). Cathode-anode Pb 3 40-500 9 0 -3 0 0 N/A 430 N/A 3 m; anode-anode 2 m. Cu 3 5-1150 15-580
5. Banerjee, et al. (1990) Nine field experiments were conducted in an array of electrodes. Combined Hydraulic and Silt/Silty Clay - Cr N/A N/A 2-4 < 7 2 Na electrical potentials were applied. Steel 10.2-0.24] reinforcing bars were used and replaced after each experiment. The results are inconclu sive. 32 provides a synthesis of field tests investigating and/or reporting some form of chemical removal from soils. Segall, et al. (1980) present the chemical characteristics of the water accumulated at the electrodes in electro-osmotic dewatering of dredged soil. Their study indicated that:
(1) There was a significant increase in heavy metals and organic
materials in electro-osmosis effluent above that recorded in the
original leachate. The concentrations of zinc, lead, mercury and
arsenic were especially high,
(2) Total organic carbon content of the effluent was two orders of
magnitude higher than the original leachate. It is postulated that
highly alkaline conditions resulted in dissolution and release of the
organic material. Pesticides came out at the cathode.
Case and Cutshall (1979) describe a field study for control of radionuclide migration in soil by application of DC current. This study demonstrates that it is possible to migrate rationuclides with the technique. Lageman (1989) reports the results of field studies conducted in the Netherlands to decontaminate soils by electrokinetic soil processing. Figure 2.9 presents a schematic diagram of the reported field process. An electrode fluid conditioning and purification system is noted. The conditioning is for the control of the influent/effluent chemistry, while purification (such as ion exchange resin columns) is for removing any excess ions in the effluent. A recent study investigating the application of the process to 33 ER CONTAINER
PURIFICATION PURIFICATION AC/DC CONVERTER L CONDITIONING CONDITIONING GENERATOR OR MAIN
CONTAMINATED SOIL;
Figure 2.9 A Schematic Diagram of the Field System Reported by Lageman ( 1989 ) 34 decontaminate a chromium site is reported by Banerjee, et al. (1990).
The laboratory studies reported by Hamnet (1980), Runnels and Larson
(1986), Lageman (1989) and Hamed, et al. (1991) together with the pilot-scale field studies of Lageman (1989) display the feasibility of using the process in site remediation. Further pilot-scale studies are necessary to improve the technology and establish the field remediation scheme for different site conditions and chemistry.
2.5 Removal of Contaminants by Electrokinetic Soil Processing
The results of this study and a scrutiny of the above presented synthesis indicate that the following processes prevail in removal of contaminants by electrokinetics soil processing.
Upon application of low-level DC current (in the order of miiliamps per cm2 of electrode area) to the saturated porous medium, the following processes occur:
(1) The water in the immediate vicinity of electrodes is electrolyzed. An
acid front is generated at the anode while a base front is created at
the cathode. A car, et al. (1989) formalized the development of these
fronts. The pH at the anode can drop to below 2.0 and can increase
at the cathode to above 12.0.
(2) The acid front will advance across the specimen in time towards the
cathode by: 35 (a) advection of the pore fluid due to the prevailing electro-osmotic
flow,
(b) advection of the pore fluid due to any hydraulic differences,
(c) diffusion due to concentration gradients,
(d) migration due to the electrical gradients. Development of these
acid and base distributions and movement of ionic species are
formalized by A car, et al. (1991).
(3) The migration, diffusion and advection will also result in
movement of cations and anions to respective electrodes in the
porous medium (Hamed, et al., 1991; Mitchell and Yeung, 1991).
(4) The acid advancing across the specimen exchanges with adsorbed
cations in the diffuse-double layer, resulting in their release into the
pore fluid and advance towards the cathode by advection and
diffusion (Hamed, et al., 1991).
(5) In case the generation of the acid front is not controlled at the
anode, the electrolyte concentration inside the porous medium will
gradually rise, resulting in increased conductivity in the vicinity of
the anode, decrease in electro-osmotic flow (Hamed, et al., 1991),
and a corresponding decrease in bulk-flow movement by advection.
(6) The anion depletion at the cathode region may lead to an increase
in voltage and an increase in energy expenditure during the process. 36 The chemistry at the anode and the cathode should ideally be
controlled to achieve continued advection while providing sufficient
H+ ions for desorption of contaminants and/or solubilization of
salts. This is possible either by decreasing the current to levels
where pH is at a desirable level or by frequent flushing at both ends
by a fluid of controlled pH and chemistry.
The development of acid/base distributions across the soil specimen is important in electrokinetic soil processing. An analytical model is developed in . this study in order to predict the pH profile across the soil specimens. This analytical model is presented in the subsequent chapter. CHAPTER HI
MODELLING OF ACID/BASE DISTRIBUTIONS
3.1 Analytical Modelling
The potential usefulness of electrokinetics in hazardous waste site remediation depends, to a large extent, on the pH of the solution that moves across the soil. This is mainly due to the solubility of most heavy metals such as
Pb(H), Cd(II), Cr(IIl) in a low pH environment.
Electrokinetic processing of soils by use of a low DC current results in development of electrical, chemical and hydraulic gradients across the saturated porous media. Consequently, the process results in the movement of ions and water molecules under the direct influence of and also due to coupling effects generated by these gradients. An analytical model has been developed to predict the pH gradients in the electrokinetic processing of soils. It is assumed that a constant current and a hydraulic potential difference is applied initially across a saturated, homogenous specimen. It is also assumed that the porous medium is made up of capillary tubes.
Formation of chemical gradients will depend upon the electrolysis and coupled chemical reactions between the soil and the electrolyte products.
However, the importance of pH gradients in electrochemical processing of soils
37 38 has been noted. Several investigators have recorded the time rate of changes in pH
of the pore fluid and mostly only at the electrodes. The general trend is that the pore fluid gets acidic (pH =1-4) at the anode and basic at the cathode (pH= 12-
13). In assessing the chemistry of the medium, for example the adsorption and desorption of species on soil substrate, prediction of the development of the pH
gradient therefore becomes essential. The following model is presented by A car, et al. (1990).
3.1.1 Chemistry by Electrode Reactions and pH Gradients
In the absence of all but traces of electrolyzable solutes in the pore fluid
(i.e., metal ions, Cl', organic matter, pollutants, dissolved oxygen, etc.) and for the case of inert electrodes, 100% faradaic efficiency can be assumed for water electrolysis by electrodes inserted in a saturated soil mass. The primary electrode reactions are:
2H20 - 4e- -* 02\ + 4H* (anode) (3.1)
AH20 + 4e~ -* 2H2f + 40H~ (cathode) (3.2)
Secondary reactions may exist depending upon the concentrations of available species, e.g.,
H* + e" - 1 H2 1, Pb2* + 2e- -» Pb 2 2 39 The production of H+ ions at the anode decreases the pH by the reaction
(1), while the reaction (2) increases the OH' concentration at the cathode
increasing the pH. Thus, as a consequence of reactions (1) and (2), in addition
to the migration of existing anions and cations in the pore fluid of the specimen
under the electrical field, two supplemental ionic species, H+ and OH', are
generated which can have a significant influence on the local conductance.
In order to estimate the boundary conditions generated by the electrode
reactions, it was necessary to assess the product distributions generated by the
electrolysis. Acar, et al. (1989) demonstrated that for currents of the order of
milliamperes used in electro-osmosis, there will be an ample amount of H20 flow
to sustain the electrolysis and local convection will be present at electrodes which
produce gases.
In electro-osmosis, the water flow rate, qe5 is related empirically to the
current by (Mitchell, 1976);
qe =Ke(cm 2/V-sec )i^(cm 2)
jf£ qe =K{cm3famp -sec)I(amp)=— I (3 • 3) o
where a is the conductivity of the pore fluid in siemens/cm, K; is given in L/amp-
sec, I in amp, and IQ is the electro-osmotic coefficient of permeability. K, values range between 0 to 1.0 x 10'3 L/amp-sec for soils. 40 If the applied current is I, 2 moles of H20 needs 4F coulombs at anode
(equation 3.1), and then the rate at which H20 is lost;
I (amp) % ,0 (mole/seC) = (2)(96485 (amp-sec/mole))
= I x (5.18 x 10"6) moles/sec (3.4) Considering an incremental element at the anode, the instantaneous steady state pH will be estimated as follows:
Q = rate of H+ production = 2xRHO H+ 2 = (1.036 x 10'5) I (moles/sec) (3.5)
The net steady state concentration will then be,
Q _ H+ (moles/sec) Molar Concentration = qc (L/sec) _ (1.036 x 10'5) (moles/amp/sec) kj (L/amp/sec) 1.036 x 105 M (3.6) K For maximum Kj of 1.0 x 10'3 L/amp-sec, equation 3.6 renders a maximum pH of 2.0 at the anode which is in conformity with recorded values in previous studies. It is interesting to note that as the efficiency of the process decreases
(smaller K, values), the pH at the anode would also decrease. At the cathode
(equation 3.2), 4 moles of H20 need 4F coulombs of electrical charge to produce
4 moles of OH' ion and hence the [OH ] ion concentration of the solution at the 41
electrode will have a pH of 12.0.
The above deliberation was for an aqueous solution in which the
decomposition of water is the primary result and so for an anode which does not
decompose. It was noted that several types of electrodes (aluminum, steel, iron,
graphite, bronze, copper, platinum, gold) are used in processing studies. The
electrolyzed products of some of these electrodes will introduce other ions in
conjunction with the H+ and OH" ions and thus can change the chemistry
generated at the boundaries. These ions will complicate the chemistry further by diffusion into the specimen and/or precipitating, depending upon the pH of the pore fluid.
In order to provide a first approximation to the development of pH gradients across the specimen, it is assumed that, due to their high transference numbers and hydration numbers, H+ and OH' ions will be the major carriers of the current and water molecules in a system where the concentrations of cations and anions in the free pore fluid are relatively negligible. This assumption may not be valid in cases where cations with high transference and hydration numbers are present (such as Mg++) in the pore fluid. However, the concentration of different anions and cations present in the laboratory prepared kaolinite specimens used in this study indicates that under the controlled conditions used in this study, the assumption may be valid since Na+ is the predominant cation and it has a low hydration value and transference number when compared with H+. 42
3.1.2 Mass Transfer in the System:
The total material flux into an incremental element of thickness dx, qlc, is composed of three components:
Qtc = qCc + qch + qce (3-7) where qcc = material influx due to chemical gradients, qch = material influx due to hydraulic gradients, and qce = material influx due to electrical gradients. The concentration of the material (solute), Cj, is defined as the mass of solute per unit volume of solution. The mass of solute per unit volume of porous media is therefore (nCj) where n is the porosity of the medium (the volume of voids over the total volume of the element). The total mass transfer due to electrical and material gradients in an electrochemical system, in general, is described by the
Nemst-Planck equations (lb, 1983)
_ [- Dj dCj + Vx Q - ZF Dj Cj d<}>] n (3.8) Qtc ” dx R T dx where z = charge on the ion, F = Faraday’s constant (96,485 coulombs), R =
Universal gas constant, T’ = temperature (°K),
The material (solute) flux due to chemical gradients is given by Fick’s First
Law; 43
(3.9) where Dx is called the longitudinal dispersion coefficient in a saturated porous media. It is composed of two components (Gillham and Cherry, 1982):
(3.10)
The first term represents the dispersion of the species due to the average linear seepage velocity in the pores of the medium and the second term describes the diffusion of the chemical in the pores. The term ax is the longitudinal dispersivity, which is dependent upon the size and frequency of the pores in the medium (fabric). It is noted that in fine-grained soils, the contribution of the first term is negligible due to very low seepage velocities and hence the longitudinal dispersion coefficient may be taken equal to the diffusion coefficient (Gillham and
Cherry, 1982). The molecular diffusion coefficient D* in the pore fluid is related to the diffusion coefficient in the free solution by
D* = p Dc (3.11) where p depends upon porosity and tortuosity of the medium and varies between
0.13 to 0.49 (Rowe, 1987).
The solute flux due to hydraulic gradients (convection) is directly related to the seepage velocity and hence the Darcy’s law,
( kx ah ) n C| (3.12) 9ch ax 44 where k* is the hydraulic conductivity of the medium and h is the hydraulic potential.
These expressions show that the material flux depends on the diffusion gradients, the convective flow conditions, and the migration contributions of species, respectively. The average seepage velocity of species will be taken as linear, uniform and equal to the average macroscopic fluid velocity. It is noted that the migration contribution presented in Eq. (3.8) represents all the ionic species in the medium. For conditions in which a dilute aqueous electrolyte is slowly passed through an electro-osmotic cell at a small but constant electrical current, the migration term will be nonlinear in concentration if the electric field d4>/dx is not constant or if the species in question carries a significant but varying fraction of the current anywhere in the bulk solution. Migration effects are usually avoided in electrochemical cells by providing an excess of an inert, supporting electrolyte.
In the following derivation, one further simplifying assumption was made.
The electrical field (3^/dx) is assumed to be constant in time across the specimen.
It was noted that this assumption may not entirely be valid since the electrical field in electro-osmotic experiments depends upon the concentration gradients imposed by the generated H+ and OH' ions and hence the specific conductance of the pore fluid in time and space. However, this assumption considered valid in order to provide a first order approximation to chemical gradients. 45
These conditions will simplify the total material flux as:
qtc = [k Q - D* 9C, ] n (3.13a) dx ZF D* d$_ + kx 9h. _ k . (3 13b) k ‘ RT' dx dx '* » + “k where k is a constant with units of [L/T] representing the velocity of the pore fluid. It is noted that the first term represents the contribution to the flow of water molecules due to the migration of H+ ion for example andits hydrated water molecules under an electric potential gradient and the second term Kh depicts the contribution to flow due to the induced or applied hydraulic gradients as a consequence of the electro-osmosis phenomena.
The mass balance across the element requires,
- _ p aC n (3.14) qtc at where (dC/dt) is the rate of mass change, A is the area and R is called the retardation coefficient, a term representing the adsorption and desorption of different species on clay surfaces and/or any interactions between the different species leading to precipitation or other chemical reactions,
R = 1 + Pd kp/n (3.15) where pd is the dry bulk density of the soil, and kp is the partition coefficient.
The gradient operator, V, in one-dimensional conditions renders,
(dqtc) = (dC_) (3.16) dx dt 46
When equations 3.8, 3.9, 3.10 and 3.13 are substituted in equation 3.16,
Dx C,xx - k C x — R C t (3.17) where C xx, C x, C t represent partial derivatives in space and time. If the experiments are conducted with a constant hydraulic potential gradient across the specimen, or if it is also assumed that the total flow due to water molecules being transported through the migration of H+ ions is constant throughout the system, k simplifies to Vx,
(3.18)
Equation (3.19) could be further non-dimensionalized by taking,
X = x/L (3.19a)
C* = 1 - C_ (3.19b) Q
P = VXL/DX (3.19c)
T = Dxt/RL2 (3.19d) where L is the length of the specimen and Q is the initial concentration of the specific species in the bulk solution. In solute transport, P is often called the
Peclet number. T is the non-dimensional time. When equations 3.19 are substituted in equation 3.18, one obtains the non-dimensional form,
(3.20)
Equation (3.20) must be solved with appropriate boundary conditions for the initial conditions and process rate in order to have a first estimate of the pH 47 gradients in electrochemical processing of soils.
It is noted that the species generated at the inlet and the outlet are different.
A constraint exists at the boundary, therefore, when they meet at a location where
the following neutralization reaction occurs.
H+ + OH H20 (3.21)
Initial and Boundary Conditions:
It was assumed, that the mass transfer chemistry is dominated by the
concentration profiles of acid and base, therefore, the initial conditions throughout
the system will be the initial pH of the pore fluid in the soil specimen.
CjfoO^pHj
under constant conditions,it is necessary to ensure that a sufficient flux of reactant is supplied at the electrodes to generate a constant supply of H+ ions. This
condition is readily met in electro-osmosis experiments and this is give a constant boundary condition, at the inlet;
(dc*/dx).m = (V C*/Dx L)m and at the outlet,
(dC*/dx)0Ut = (V C*/Dx L)out
Based on the above equation and initial and B.c’s, a two dimensional finite element model has been developed.
A computer program (DIFFUSE) was written in FORTRAN based upon 48 the finite element formalism (Acar, et al., 1988). Two dimensional finite element method formulation for the analytical model is presented in Appendix B. Flow chart for the program and the subroutines are presented in Appendix C.
3.2 Predictions of The Analytical Model:
Figure 3.1 presents different pH gradients predicted by the model. Figure
3.1a is for the case when the pH at the anode and the cathode are 2.0 and the initial pH of the porous medium is 7.0. The front presented is for diffusion only and hence the acid-base front meets in the middle. It is noted that a neutralization reaction will occur when these fronts meet. Figure 3.1 .b depicts the case when the specimen is initially acidic with a pH of 5.0. In this case, the OH' ion front moves through the specimen. Figure 3.1.C is for the case when the convective flux provides a P value of 30. The acid front moves further toward the cathode in time and is expected to flush the specimen eventually. The three cases presented indicate that the pH gradient would be a function of:
(1) the pH values established at the electrodes by the rate of
electrolysis; extreme values occur at low K; (see equation 3.7) and
hence high electrolyte concentrations of products ( equation 3.3);
(2) the initial pH of the porous medium;
(3) whether the flow would be governed by solely diffusion of the
chemical or by the advection of the bulk fluid flow (P value); 49 Vx L 12 = 30
10
9 001 — 8 .0 0 3 — .0 0 4 2 5 7 pH 6
5 Vx L = 0 4
3
2
0.1 0.2 0.3 04 0.5 0 6 07 08 0.9 0 1 0 2 0.3 0.4 0.5 0.6 0.7 0.8 0 9 NORMALIZED DISTANCE FROM ANODE n o r m a l iz e d distance from anode (x/L) (x/L)
12
10
9
8
7
6
5
4
3
2
0.1 0 2 0.3 0.4 0.5 0 6 0.7 0 8 0 9 NORMALIZED DISTANCE FROM ANODE (x/L)
Figure 3.1 The pH Gradients Predicted by die Model 50 (4) any adsorption/desorption and precipitation of H+ and OH" ions.
The effects are included in the retardation coefficient R given in
equation 3.15.
Figure 3.2 compares the predictions of the analytical model with experimental results. The acid front generated at the anode advances towards the
cathode by advection and migration. In electro-osmosis, with open electrode configuration, this front will flush the specimen and reach the cathode (Putnam,
1988, Acar et al., 1990a). The analytical model provides a good parametric assessment of the acid/base distributions. Recent studies further substantiate the potential of this model in predicting the progress of the acid front (Thompson,
1989).
As a result of the pH gradients, the following physico-chemical interactions would be expected:
(1) Dissolution of the clay minerals beyond a pH range of 7-9,
(2) Dissolution of certain salts available in the medium,
(3) Adsorption/desorption and exchange of cations by replacement by
H+ and OH ,
(4) Precipitation of metal ions at pH values corresponding to their
hydroxide solubility products leading to cementitious products,
(5) Changes in the structure and hence the engineering characteristics
of the soil due to variations in the pore fluid chemistry. 14
13 GEORGIA HAOlfNiTE CURRENT I m A 13 DIAMETER 5 0cm pH (cothode) ■12 0- LENGTH 20 0 cm 12 fc, * 0 3 - 0 0 3 giji/omp-hf 12 3 0 to QBcm2/*»i#c 1 I P « v.L/D, ■ 30 w * 0 73 - 0 90 T« Time focto' » Dft/RL^ II TIME Ihf) ...... T« 0 002 10 ------T > 0 007 ------T- 0 010 10 27 ------T * 0 017 9 ------T « 0 043 9 120 8 0 pH 7 7 6 6 3 Imho1 pH
4
3 2 pH ( onode) ■ 2 0
0 01 02 03 04 05 08 07 0 8 0 9 >0 0 1 0 2 0 3 0 4 0 5 0 8 0 7 0 8 0 9 10 NORMALIZED DISTANCE FROM ANODE, X/L NORMALIZED DISTANCE from ANODE, y/L
Figure 3.2 pH Gradients in Electro-osmosis ( Acar, et al., 1991 ) 52 A review of previous studies indicates that all of these interactions were reported to be manifested by changes in the characteristics of the soil and the geotechnical properties in electro-osmosis (Acar, et al., 1989). Laboratory and field studies show that changes in consistency limits of soils correlated only with the pH gradients (Bjerrum, et al., 1967; Esrig and Gemeinhardt, 1968).
The movement of the low pH front by migration and advection leading to the desorption of inorganic cations from the clay surfaces together with the concurrent electro-osmotic flow process will constitute the fundamental mechanisms by which inorganic cations could be removed from fine-grained soils.
The predictions of this model led to the conclusion that electrokinetics has the potential to be used in soil remediation of contaminated soil. Therefore, a testing program has been implemented to verify the hypothesis. CHAPTER IV
METHODOLOGY
The objectives of this study require an assessment of the feasibility of removing inorganic contaminants adsorbed on fine-grained soils. A survey of frequently encountered contaminants has led to selection of Pb(II), Cd(D), and
Cr(m). These heavy metals are common sources of contamination. In accordance with the objectives of this study, it is decided to select conditions which will most favor electro-osmotic flow. Gray and Mitchell (1967) indicate that low activity, low electrolyte concentration and high water content soils promote high electro- osmotic efficiency. Therefore, a kaolinitic soil is selected for this study. Kaolinite, being a low activity soil mineral, results in high electro-osmotic flow efficiencies and furthermore, it has an adsorption capacity for cations. A testing program is planned using this soil and the above mentioned heavy metals.
4.1 Testing Program
Georgia kaolinite from the Thiele Kaolin Company in Wrens, Georgia is used in this study. Typical physical and chemical properties of the soil provided by Thiele Kaolin Company are presented in Table 4.1. The compositional and engineering characteristics of this mineral, determined by tests at LSU, are presented in Table 4.2. This soil has a low activity (A=0.32). It is noted that the
53 Tables 4.1 Typical Physical and Chemical Properties ( Thiele Kaolin Company )
Moisture Content % 0 .5 - 1.5
Surface Area m2/g 20 - 26
pH (20% solids) 3.6
Silica (Si02) % 4 3 .5 - 4 4 .5
Alumina (Al20 3) % 3 8 .0 - 40 .5
Iron Oxide (Fe20 3) % 0.9 - 1.3
Titanium Dioxide (Ti02) 1 .4 - 3.5 55
Table 4.2 Characteristics of Georgia Kaolinite
Mineralogical Composition {% by weight)
Kaolinite 98 lllite 2
Index Properties (ASTM D 4318)*
Liquid Limit <%) 64 Plastic Limit (%) 34
Specific Gravity (ASTM D 854)b 2.65
% Finer than -2 /jm Size 90
Activity 0.32
Cation Exchange Capacity 1.06 (milliequivalents/100 gm of dry clay)
Proctor Compaction Parameters
Maximum Dry Density, tons/m3 1.37 Optimum Water Content, % 31.0
Initial pH of Soilc 4.7-5.0
Compression Index (Cc) 0.25
Recompression Index (Cr) 0.035
Permeability of Specimens Compacted at the Wet of Standard Proctor Optimum (x 10'8 cm/sec)d 6-8
•ASTM Method for Liquid Limit, Plastic Limit, and Plasticity Index of Soils (D 4318- 83) bASTM Method for Specific Gravity of Soils [D854-58 (1979)] cpH Measured at 50% Water Content dFlexible wall permeability at full saturation 56 pH ranges between 4.7 to 5.0 at 50 % water content. The cation exchange
capacity ( CEC ) of this soil is determined by adsorption tests. CEC is expressed
in milliequivalents per 100 gm of dry mineral (Equivalent weight of a species is
the molecular weight divided by the valence and is a quantity which represents the
weight of that species per valence). Therefore, CEC would provide the adsorption
capacity of the soil.
Adsorption tests are conducted on Georgia kaolinite, by adding 20 mL of
Pb(N03)2 solution at different concentrations to 5 grams of dry clay in a 50.0 mL
centrifuge tube. The centrifuge tubes are mechanically shaken for 24 hours, then
the mixture is centrifuged up to 3500 rpm to separate the solids from the liquid
phase. A liquid sample of about 10 mL is analyzed by ICP for Pb(II). The initial
Pb(II) present minus the equilibrium concentration in the liquid phase provided the
amount adsorbed on the clay.
The adsorption isotherm for Lead on Georgia kaolinite is presented in
Figure 4.1. It is observed that this mineral can adsorb about 1100 ng of Pb(II) per
gram of dry clay. This limiting quantity is defined as the cation exchange capacity
of this mineral. Therefore, Pb(D) exchange capacity of Georgia kaolinite is 1.06 milliequivalents per 100 gm of this mineral. Cations in excess of this quantity would be largely present in the pore fluid.
It was decided to conduct four different sets of tests. The objectives of the first three sets of tests were to investigate the feasibility of removing adsorbed Amount Adsorbed on Clay (ug/g) ,3 2 C1 1 1z 0 104 103 10z 10 1 ICf1 Figure 4.1 Pb(II) Adsorption Isotherm for Georgia Kaolinite Georgia for Isotherm Adsorption Pb(II) 4.1 Figure qiiru Cnetain (ppm) Concentration Equilibrium 57 58
Pb(H), Cd(II) and Cr(III) by electrokinetics. Therefore, the concentration of these heavy metals loaded on the kaolinite is selected to be below the exchange capacity of this mineral. The fourth set of tests is conducted to determine the influence of the increase in concentration of the contaminant and current density on the operational variables and the feasibility of the process. Pb(n) was selected for this set of tests. Table 4.3 presents a summary of the testing program planned for this study.
4.2 Permeation fluid
The clay is washed with deionized water several times and the pore fluid collected is used as the influent. This pore fluid will have some of the mineral salts and other soluble material available in the kaolinite and therefore it is considered to be a representative pore fluid.
4.3 Contaminants
Pb(N03)2, Cr(N03)3 , and Cd(N03)2 are used as the salts which provided the necessary ionic forms of the desired inorganic cations. In each experiment, these salts were ionized in a required amount of deionized water to prepare the solution used. Well developed and accurate chemical analysis schemes exist using Atomic
Adsorption (AA) or Inductively Coupled Plasma (ICP). Tests have been conducted at low current density to avoid heating and other electrochemical effects which inhibit flow. Table 4.3 Summary of Testing Program 59
Dimensions of Specim en Number Current Concentration Contaminant of Density (meq./100 gm) T est |m A /cm 21 Length Area (cm) (cm2)
Pb (ll)a T ests 2,4,5 3 0 .0 3 7 0 .1 4 1 0.2 81 1 1 0 .0 1 2 0 .1 4 2 0 .3 81 3 1 0 .0 3 7 0 .1 4 2 0 .3 81
Cd(ll)8 Cd#1 1 0 .0 3 7 0 .2 0 10.2 81 Cd#2 1 0 .0 3 7 0 .1 8 10.2 81 Cd#3 1 0 .0 3 7 0 .1 9 1 0.2 81 Cd#4 1 0 .0 3 7 0 .1 8 10.2 81
Cr(lll)a Cr#1 1 0 .0 3 7 0 .3 0 1 0.2 81 Cr#2 1 0 .0 3 7 0 .2 9 1 0.2 81 Cr#3 1 0 .0 3 7 0 .2 6 10.2 81 Cr#4 1 0 .0 3 7 0 .3 0 10.2 81
Pb(ll)b Set #1 Pb#h3 i 0 .0 1 2 0 .5 2 1 0.2 81 Pb#h2 1 0.037 0.60 10.2 81 Pb#h1 2 0 123 0.51 1 0.2 81 Pb#hr 2 0 .1 2 3 0 .5 7 1 0.2 81
Set #2 Pb#h5 1 0 .0 1 2 0 .8 3 10.2 81 Pb#h9 1 0.037 1.02 1 0.2 81 Pb#h6 1 0 .1 2 3 0 .7 2 10.2 81 Pb#h8 1 0.123 0.81 10.2 81
Set #3 Pb#h10 1 0 .0 1 2 0 .0 8 10.2 81 Pb#h13 1 0 .0 3 7 0 .0 7 10.2 81 P b # h 1 1 1 0 .1 2 3 0 .0 7 1 0.2 81 P b # h 1 2 | 1 0 .1 2 3 0 .0 7 10.2 81
Cd#1 = Cadmium test #1 , Cr#1 = Chromium test #1 a = Tests conducted to study the feasibility of the process, b = Tests conducted to study the effects of current density and concentration 60
4.4 Testing Equipment
The following equipment was used in conducting this study:
1. Four electro-osmosis cells are designed and constructed. The cells are made of plexiglass to facilitate visual observation. Each cell consists of a cylindrical body ( 4 in. diameter and 4 in. to 8 in. length) and two caps as presented in Figure 4.2. Graphite electrodes are used in this study to prevent corrosion and introduction of secondary corrosion products. The electrodes are held in place with polyacrylate end caps, connected with threaded rods. A liquid reservoir of 110 ml capacity is available at each end. Holes are drilled into the top of each end cap above these reservoirs to allow venting of gaseous electrolysis products.
2- In order to achieve full saturation in specimen preparation, it is found necessary to consolidate the specimens from slurry. Two consolidation frames were designed. A view from these frames is presented in Figure 4.3.
3- It is found necessary to monitor the outflow, the effluent pH, the current, the potential difference across the electrodes and in some cases, the potential difference across the specimen. Therefore, two personal computers with data acquisition boards (DAS AD Boards) , and a data acquisition and monitoring software (Labtech Notebook) were assembled. The conductivity measurements are conducted separately by a CDM 83 conductivity meter.
4- Four HP power supplies (HP Model 6121C) with a maximum range of Figure 4.2 View of Electro-Osmosis Cell. 62
a
Figure 4.3 View of the Consolidation Frames Used in This Study. 63
100 mA and 120 volts are used to supply the constant current throughout the tests
conducted in this study.
5- Two Beckman $31 and $32 pHmeters and several combination glass pH
electrodes are used.
4.5 Specimen Preparation
Different concentrations of Pb(ll), Cd(II), and Cr(III) ( Table 4.3 )
solutions are prepared with ACS reagent grade nitrates [ Pb(N03)2, Cd(N03)2, and
Cr(N03)3 ] and distilled water. A few drops of HN03 were added to the solutions
to avoid any precipitation of the metals. One kg (2.2 lb) samples of air-dry
kaolinite are mixed with 1.5 liter portions of different concentrations of nitrate
solutions. The resulting slurry was stirred for 30 minutes to allow the ions to
equilibrate between the solid and the liquid phases. The Pb(H), Cd(II) or Cr(lH)
loaded kaolinite slurry was kept in a container for at least two weeks before it was
used. This provision was taken with the anticipation that a better homogenization
would be achieved through this wait period. The slurry was then placed in a series
of the plexiglass cylinders and consolidated. The consolidation process consists of a self-weight consolidation for at least two days, followed by successive increase
of the load increment from 25 kPa to 200 kPa. The pore fluid extracted due to
consolidation of the slurry was collected for chemical analysis. This provided an assessment of the initial pore fluid concentration. Upon completion of the consolidation process, the cylindrical cell was removed and soil samples were 64 collected from the top and the bottom of the specimen. Chemical analyses were conducted and water contents were determined to provide the initial contaminant distribution throughout the specimen. The cylindrical specimen was directly placed in the electro-osmosis set-up for testing.
4.6 Testing Procedure
Electro-osmosis test specimens were assembled as presented in Figure 4.4.
Two sheets of 8 /xm filter papers were placed at both ends of the specimen.
Uniform flow across the electrodes was arranged by drilling fifty 0.3 cm (0.12 in) holes into the electrodes.
Specimens were placed and tested in a horizontal configuration. This configuration allows the gases generated by the electrolysis reactions to escape through the gas vents at both ends. The current applied was maintained in an open flow arrangement in which a Mariotte bottle was used at the anode side to maintain the external hydraulic gradient small (< 0.1) and constant.
Influent/effluent pH, conductivity of effluent, the outflow, the current and the electric potential difference were continuously monitored by the data acquisition system.
After a selected period of current application, the specimen was promptly removed from the cell and sliced into 10 sections. Each segment was analyzed for water content, pH, conductivity, and chemical analysis. Furthermore, the pore fluid in each section was separately extracted under pressure applied by a jack to COMPUTER
□D
TERMINAL BOARD
POWER SUPPLY WATER pH METER RESERVOIR
CONDUCTIVITY METER
CONDUCTIVITY CELL CARBON ELECTRODES VENT — FILTER PAPER -----
CATHODE
CLAY SPECIMEN CJ
GLASS ANODE BURET
END CAP END CAP
SUL TRANSDUCER TRANSDUCER POWER SUPPLY
Figure 4.4 A Schematic Diagram of the Electro-Osmosis Cell and the Test Set-up 66 a 5 cm (1.97 in.) stainless steel die containing 5-10 gm of the specimen, or by centrifuging. In later tests, only centrifuging the specimens was used. This pore fluid was also analyzed for chemical content, conductivity and pH. A pH meter and universal indicator paper were used to determine the pH values of the pore fluid and sliced segments of the specimen. Conductivity measurements were also made of this pore fluid and each individual slice with a conductivity meter.
As a result of electrolysis, a large amount of heavy metal was found to be deposited at the graphite electrodes. To remove the metal from the electrodes, about 650.0 mL of HN03 with concentration of 8.0 M was added to the electrode in a beaker. The solution was heated to boiling over a hot plate for two hours.
After cooling, the solution was filtered and diluted for chemical analysis.
All chemical analysis was by flame atomic absorption (AA) (in the first two tests), or inductively coupled plasma (ICP) for all other tests. Contaminant was extracted from clay specimens by adding 20 ml of 1.6 M HN03 to 5 gm soil taken from each slice and leaving the slurry in the shaker for 48 hours at room temperature. ■ - ,
Constant current conditions were used in all tests to keep constant the net rates of the electrolysis reactions at all times and minimize complicated current boundary conditions. Secondary temperature effects have been reported to decrease the efficiency of electro-osmotic flow when the current density is greater than 5 mA/cm2 (Gray, 1970). In order to avoid such effects, the current density used was selected about one to two orders of magnitude lower than 5 mA/cm2. CHAPTER V
ANALYSIS OF RESULTS
In light of the understanding developed for the creation of acid/base
distributions in the electrokinetic soil processing, tests have been conducted to
investigate the feasibility of using electrokinetics to remove Pb(II), Cd(II), and
Cr(III) from saturated kaolinite under controlled laboratory conditions. The effect
of concentration and current density on the process was also investigated.
The results of these tests together with an analysis of the process efficiency,
an evaluation of the energy expenditure, the effect of concentration , and the role
of the current density are presented in this chapter.
5.1 Feasibility tests:
5.1.1 Pb(ID
Five tests were conducted to evaluate the feasibility of lead removal using the technique. The dimensions of the specimens, duration of tests, current used, and other parameters in these tests are presented in Table 5.1. The initial Pb(II) concentration ranged from 117 to 144 fxg/g of dry soil, which was below the cation exchange capacity of the soil.
The predictions of the analytical model presented in Chapter 3 and subsequent laboratory tests on blank specimens, indicated that the efficiency of the
67 68
Table 5.1 Summary of Testing Program and Processing Data
T est No.
Param eter 1 2 3 4 5
Current (mA) 1 3 3 3 3
Dimensions Area (cm2) 81.1 81.1 81.1 81.1 81.1 Length (cm) 20.3 10.2 20.3 10.2 10.2
Duration (hr) 638 100 1285 409 545 (763)
Current Density (mA/cm2) 0.012 0.037 0.037 0.037 0.037
Total Charge (amp-hr) 0.638 0.300 3.86 1.227 1.640 (2.89)
Total Charge/ unit volume of soil 388 362 2345 1483 1982 (amp-hr/m ) (1390)J
Porosity of the specimen 0.68 0.76 0.68 0.72 0.71
Initial Pb(ll)1 Concentration 144 r 1.4 134 c 1.5 145 = 0.5 123 ~ 0.2 117 = 5 : (ng/g of dry soil)
Pore Volume 2 cm2 1120 629 1120 596 587
Total Flow (cm3) 1304 110 11833 1419 1956
Pore Volumes of Flow 1.16 0.17 ■ 1.06 2.38 3.33
(1) Mean and one standard deviation
(2) Pore Volume « Total Volume ’ Porosity of Specimen
(3) T est continued until 1285 hr. Flow w as unm easurable at 763 hrs. 69 Process in Pb(II) removal will depend upon the extent to which the acid front
reaches the cathode. Therefore, the duration of tests 3 to 5 continued until the pH
at the cathode decreased. Tests 1 and 2 are interrupted earlier in order to assess
the conditions across the cell before complete acid flushing. In tests 2 to 5 the
current was increased to 3 rnA in order to achieve the breakthrough in a shorter
duration. Similarly, the length of the specimens were decreased in tests 2, 4 and
5 to decrease the acid breakthrough time.
5.1.1.1 Electrical Potential Gradients and Flow
Figure 5.1 presents the development of the net average electrical potential
gradient across the cell electrodes and the total flow in time. It is noted that the
total electrical potential across the electrodes is slightly higher than the potential
gradients recorded in the specimen, due to polarization of the electrodes by the electrolysis reactions. The potential losses at the electrodes were not found to be
significant at the current levels used in this study. The following observations are made:
(1) Under constant current conditions employed in these tests, the average cell electrical potential gradients increases to relatively constant values from an initial average value of about 0.3 to 0.9 V/cm. The initial and the final values attained are approximately in proportion to the applied current; i.e., in test 1, the current was 1 mA and the final electrical potential gradient was 0.8 to 1.0 V/cm and in 70
3.0
> 2.5
UJ 9 2.0
UJ
Test No I
£ 0.5
UJ
100 20C 3 0 0 4 00 500 600 700 800 Time ( hr )
2000
1800
1600
1400 r 1200
1000
Test No 600 / / / Test Current (m4) Length(cm) * 1 1.0 2 0.3 2 3 .0 1 0 2 200 3 3 .0 20 3 4 5 0 10 2 5 3.0 10.2
100 200 300 400 500 600 700 800 TIME ( hr )
Figure 5.1 Electrical Potential Gradient and Total Flow 71 the other tests, the current was 3 mA and the mean gradient becomes 2.3 to 2.5
V/cm,
(2) There was no measurable flow in the first 10 to 20 hrs of current application.
Subsequently, the flow rate increased rapidly and then nonlinearly decreased in
time,
(3) The total flow varied considerably in tests at 3 mA, possibly due to variations
in the initial water content of the specimens, or other unknown factors.
Similar mean potential gradient and flow data were obtained in the previous blank tests conducted under constant current conditions (Putnam 1988, Acar, et al., 1991). These results demonstrate convincingly that flow in electro-osmosis is
not continuous and that the flow rate is time-dependent.
Figure 5.2 presents the change in the empirical electro-osmotic coefficient
of permeability, K^, and the electro-osmotic water transport efficiency, K,, in the tests. The tests indicate that,
(1) Kc increases to 1.8 to 2.2 x 10"5 cm2/V-sec at the start of each test and decreases by an order of magnitude in time (note the flow data in Figure 5.1); and
(2) similarly, Kj increases to 0.8 cm3/amp-sec at the start of the tests and decreases to less than 0.1 cm3/amp-sec in conformity with the flow data.
Similar trends have been reported by Casagrande (1983). Time dependent changes in Kc and K, demonstrate that significant changes occur in the overall cell resistance and hence the chemistry across the cell during the process. Therefore, 72
> O 2 .0 v oE m Test No. I O K 0 5 200 400 600 600 TIME ( hrs ) 0.8 Test No t 0.6 0.2 200 «00 600 600 T IM E ( h r s ) Figure 5.2 Changes in Electro-Osmotic Coefficient of Permeability and Coefficient of Water Transport Efficiency 73 the coefficients Kc and are not constants for a specific soil but they are time- dependent variables presumably controlled by the chemistry generated. 5.1.1.2 Effluent pH Figure 5.3 presents a plot of the effluent pH at the cathode with respect to the number of pore volumes of flow. Upon the start of the experiment, the pH at the cathode compartment rose to values of 11 to 12. Subsequently, the pH remained relatively constant until 0.8 to 0.9 pore volumes of flow, after which it decreased with further processing. It is interesting to note that the decrease in pH was a companied by the decrease in flow rate (Figure 5.1). 5.1.1.3 pH Profiles Table 5.2 compares the pH distributions across the specimens upon completion of the tests. The pH values obtained with the pH probe inserted directly into the soil mass compare favorably with the ones obtained using the Indicator paper. Tests No. 1 and 5 represent those with the least and maximum amount of total charge passed per volume of the specimen, respectively. Figure 5.4 compares the in situ and pore fluid pH values across the specimen in tests 1 and 5. It is noted that the pH across the specimen decreases with an increase in the processing time, or with the total charge passed per volume of specimen. The increase in the pH from the anode to the cathode is a direct consequence of the advance of the acid front by migration, diffusion and advection. These repetitive iue . Efun Hvru h oeVlms f Flow of Volumes Pore the versus pH Effluent 5.3 Figure pH OF EFFLUENT 14 ! 0 12 4 6 8 05 . 15 . 25 3.0 2.5 2.0 1.5 1.0 0.5 0 OE OUE O FLOW OF VOLUMES PORE Test# 3 Test# 74 Table 5.2 pH Distributions Across the Specimens Normalized Blank Test T e st 1 T est 3 T est 4 T est 5 Distance from A node I II I II Ill I II III I III I II III 0 .0 - 0.1 3 .6 4 .5 2 .8 2 .0 3.1 2.7 3 .5 2 .8 2 .7 2 .6 2 .4 1.5 2 .3 0.1 - 0 .2 3 .7 4 .5 2 .8 2 .0 3.2 3.2 3.0 3.1 2 .7 2 .6 2 .3 1.5 2 .4 0 .2 - 0 .3 3 .7 4 .5 3 .3 2 .0 3.2 3.6 3.5 3.5 2 .7 2 .7 2.3 1.8 2 .4 0.3 - 0.4 3.9 4.5 4.0 3.5 5 .7 3 .9 3.8 3.9 2.8 2.6 2 .2 2 .3 2 .4 0.4 - 0.5 3.9 4 .5 4.1 3 .5 5.7 4 .0 3 .8 4.1 2 .8 2.7 2 .9 2 .5 3 .4 0 .5 - 0 .6 3.9 4.5 4.2 3.8 5 .6 4 .0 3 .5 4.1 3 .0 2 .9 3 .0 2 .5 4 .3 0.6 - 0.7 3.9 4.8 4 .2 4 .0 5 .4 4 .0 3 .8 4.1 3 .0 3 .4 3 .3 2 .8 4 .5 0 .7 - 0 .8 4.0 4.5 4.3 4.0 5.6 4.1 3 .0 4 .8 3 .4 3 .9 3 .0 2 .7 5 .0 0 .8 - 0 .9 4 .0 4 .5 4 .4 4 .0 5.7 4.0 3.0 4.6 3.8 4.3 3.3 2.5 4.9 0 .9 - 1.0 4 .0 5 .0 4 .4 - 6 .2 4.1 3 .0 4.9 4.2 5.9 2.8 2.5 5.0 I) In-situ with pH meter II) Litmus paper {Accuracy is 0.5 pH units) III) Pore fluid with pH meter 76 7 Test No. Total Charge Initial Pore Fluid pH (amp-hr/m3) I 368 5 1982 6 5 Test 5 —y Pore Fluid\ / Initiol Insitu 4 Test I Insitu Test I Pore Fluid 3 ■Test 5 Insitu 2 — i------1------1------1______i______i______i______i 0.2 0.4 0.6 0.8 1.0 NORMALIZED DISTANCE FROM ANODE Figure 5.4 A Comparison of Insitu and Pore Fluid of pH Profiles across the Cell in Tests 1 and 5 77 results showing the evolution of pH gradients in the Pb(II) loaded specimens of tills study and the blank specimens of a previous study (Acar et al., 1991) reconfirm the predictions of the analytical model for the acid/base distributions in time. The acid front generated at the anode flushes across the cell, neutralizing the base at the cathode decreasing the effluent pH. It is interesting to note that as the cathode compartment is approached, the pore fluid pH is considerably larger than the in-situ pH values. The pH is usually taken as a quantitative indication of the activity of hydrogen ions in aqueous homogeneous solutions. Although the insitu pH values at the anode in test 1 and at the cathode in test 5 are relatively identical (pH = 2.7 - 3.0), the pore fluid pH measurements show a substantial difference near the cathode. This observation was repeated in all tests conducted. This implies that insitu pH measurements depend on the degree of saturation of the clay surface by H+ ions. An explanation of this observation is possible by considering the production of acid and base at the boundaries of the cell. As the process continues, the pore fluid pH close to the anode decreases more rapidly with the decrease in lq. In other words, the dispersion of acid at the anode decreases. Some of the H+ ions are readily adsorbed by the negatively charged clay surfaces to satisfy the cation exchange capacity of the mineral. When the adsorption process reaches equilibrium, any excess ions are free to flow through the cell; however, the initial dispersion appears to be a non-equilibrium phenomenon. As this acid front moves 78 through the specimen, it is dispersed by both advection and diffusion. The adsorption mechanism increases the concentration of H+ ions in the double layer, decreasing the pore fluid H+ ion concentration. The pore fluid with an H+ ion concentration lower than that at the anode finally meets and neutralizes the OH' ions generated at the cathode which have back-diffused upstream. Thereby, the OH' ion concentration in the vicinity of the cathode is eventually lowered. Consequently, the pH boundary conditions and flow rates eventually dominate the pH of the pore fluid at the cathode, while the insitu pH is controlled by the rate at which the H+ ions adsorb on the soil. 5.1.1.4 Conductivity Conductivity is inversely related to the resistance offered to current flow. This resistance would change due to variations in pore sizes (porosity), tortuosity in the porous medium, and variations in pore fluid and double layer electrolyte concentrations. Conductivity is measured in siemens (mhos) per distance between the measurement points in a medium. Conductivity rises from a very low value (~ 1 /us/cm) for deionized, distilled water to orders of magnitude higher for a fluid containing electrolytes, depending on their types and concentrations. In electrokinetic soil processing, conductivity can either be estimated from the electrical potential drop across the supplied electrodes and the current across the electrodes (apparent conductivity), or it could be directly measured by probes 79 in the soil or the pore fluid with a conductivity meter (overvoltages associated with the electrode reactions should be made small in comparison to the cell iR voltage drop). In the relatively simple electrolyte system (118 to 145 /xg/g of Pb 2+ in the soil and 10"4 equivalents of Na+ and S042' in the pore fluid) present in the specimens of this study, conductivity values are sensitive indicators of H+ and OH' concentrations if they exceed the pore fluid ionic concentrations. The apparent conductivity, K,, calculated from the electrical potential drop across the electrodes, is defined as, Ka(siemens/cm) = I,(amp) L(cm)/Vt(volt) A (cm2) (5-1) where It is the current, V, is the voltage difference between the current generating electrodes, L is the specimen length and A is the cross-sectional area of the specimen. Equation 5.1 normalizes the conductivity across the cell and allows comparison of conductivity in specimens having different lengths and areas with different current or voltage conditions. Figure 5.5 presents the change in apparent conductivity across the electrodes in these experiments. Since the tests were performed under constant current conditions, the decrease in K* in time reflects an increase in the resistance offered and an increase in the voltage required to maintain the desired current. In all tests, the apparent conductivity steadily decreased from 60 to 80 fis/cm to | | APPARENT CONDUCTIVITY, KA ( /xs/ cm ) 100 80 90 60 TO 40 20 50 30 Figure 5.5 The Change in Apparent Conductivity with Time. with Conductivity Apparent in Change The 5.5 Figure nta specimen were measured to be 75 to 86 /xs/cm and 186 to 202 /xs/cm respectively. The apparent conductivity is a good approximation of the actual soil conductivity only during the early phases of electro-osmosis, before significant polarization takes place and when the electrolyte conux. jumon is relatively constant across the cell. Therefore, the initial apparent conductivity value of 60 ixs/cm to 80 /xs/cm calculated across the electrodes compares favorably with the initial conductivity of 75 ix s/cm to 85 /xs/cm measured in segments across the cell. However, at longer times of processing, the K, values calculated by the potential difference across the cell and the current supplied may significantly underestimate the conductivity across the cell due to the following reasons: (1) the electrochemical overpotentials required to drive the electrode reactions may contribute to the measured voltage increasing the resistance and hence lowering the apparent conductivity (these variations are not found to be significant for the current levels used in these tests), (2) in case of significant variations in electrolyte concentrations across the cell, K, would be more representative of those locations that would offer the highest resistance to current flow, or render the lowest conductivity (series resistors). Therefore, when the conductivity at a limited section of the cell decreases significantly due to a decrease in available electrolyte concentration, the voltage across the specimen would increase and the calculated KB would decrease. 82 In order to develop a better understanding of the electrochemistry associated with electro-osmosis, it is necessary to evaluate the conductivity across the cell. At the end of each test, conductivity measurements were done across each segment and also in the pore fluid extracted from the segment. Table 5.3 presents the insitu and pore fluid conductivity profiles across the cell in the blank specimen and upon completion of each test. Figure 5.6 compares the profiles in tests 1 and 5. Pore fluid conductivity is higher than the insitu conductivity and the trend in both profiles is similar across the cell. Conductivity is directly related to the mobility and concentration of the ions carrying the charge. Therefore, ions which are immobile due to their association with the double layers in the clay mineral would contribute less significantly to the conductivity of a porous medium of this type. Consequently, for a first order approximation for dilute systems the in-situ conductivity is less than the pore fluid conductivity by a factor that depends upon the porosity and tortuosity of the medium. Conductivity at the anode compartment of the cell increases at least by an order of magnitude from the initial conductivity, while it decreases by almost an order of magnitude in the cathode compartment. The H+ ions in the pore fluid significantly affect the conductivity of the system once their concentrations are of the order of, or are greater than, the initial electrolyte concentrations. In fact, in the longer duration tests (tests 3, 4, and 5) the conductivity profiles follow the 83 Table 5.3 Measured Conductivity Values Across the Cell Conductivity, a (//s/cm) Normalized Blank Test Insitu Pore Fluid Distance Insitu Pore Fluid Test No. Test No. from Anode 3 4 5 1 3 4 6 0.0 - 0.1 75 202 500 _ 956 994 1346 1420 3305 0.1 - 0.2 77 202 255 638 770 700 810 986 2800 0.2 - 0.3 77 202 117 600 653 486 272 900 800 0.3 - 0.4 81 187 50 465 362 123 107 875 880 0.4 - 0.5 82 187 17 347 111 139 46 655 59 0.5 - 0.6 82 186 14 283 24 46 34 448 36 0.6 - 0.7 86 186 14 158 23 49 71 240 49 0.7 - 0.8 86 202 12 128 20 124 66 - 148 35 0.8 - 0.9 84 202 14 157 20 146 26 59 57 0.9 - 1.0 84 202 104 27 22 194 33 50 23 CONDUCTIVITY, tr ( /tis/cm) iue56 Cmaio fCnutvt rfls cos h Cell the Across Profiles Conductivity of Comparison A 5.6 Figure 0 0 0 5 1000 IOOO 100 100 0.0 T OMLZD ITNE RM ANODE FROM DISTANCE NORMALIZED nl K inal F ia K Final 0.2 T *=Cluae ars Electrodes across K* =Calculated * & .4 0 I “ oa Cag * 92 hrm3 r/m -h p m a 1982 * Charge Total oa Chre 38 m3 /m r h - p m a 388 * harge C Total oe li er Fluid Pore Measured in Segments Segments in Measured Pore Fluid Fluid Pore ' nsitu nsitu ------. 0.8 0.6 et 5 Test o t I st e T 1 ------Pr Fluid) (Pore (Insitu) or Initial Initial ' ------Pr Fluid) (Pore nta cr Initial nto cr Initiol (Insitu) a o 1 ------ ' ---- 1.0 84 85 reverse trend observed in the pore fluid pH profiles, i.e., at the anode, low pore fluid pH renders high conductivity values. At the cathode, the neutralization acid/base reaction makes the pore fluid conductivity a more complex function of elapsed time. Finally, in the shorter duration test (test 1) pore fluid conductivity drops below the initial conductivity in the mid section of the cell, possibly indicating depletion due to migration of the initial electrolyte ions that were present. Upon further processing, this decrease stretches across die cathode section (test 5) suggesting that anion migration or acid/base neutralization has depleted die available ion carriers. 5.1.1.5 Efficiency of Pb(D) Removal Table 5.4 provides profiles of final Pb(II) concentrations, Cf, in the specimens as a percentage of the initial concentrations, C;, given in Table 5.1. Figure 5.7 presents a comparison of these results. It is noted that Pb(II) was only partially removed along the anode section and it was accumulated at die cathode section in the shorter duration tests 1 and 2, while in the longer duration tests (tests 3, 4 and 5), 75 to 95 percent of Pb(II) was removed across the cell. The mass balance in each test is provided in Table 5.5. It is interesting to note that a significant amount of the removed Pb(II) was electroplated at the cathode in tests 3, 4 and 5 with the higher total charge input. These results 86 Table 5.4 The Ratio of Final to Initial Pb(ll) C oncentrations Normalized Distance Concentration Ratio, C,/Cj from Anode ______ T est No. 1 2 3 4 5 0 .0 0 -0 .1 0 0.21 0.60 0.05 0.09 0 .0 5 0 .1 0 -0 .2 0 0 .3 0 0 .37 0 .0 7 0 .1 2 0 .0 6 0 .2 0 -0 .3 0 0.48 0.31 0.08 0.11 0 .0 8 0 .3 0 -0 .4 0 0.52 0.92 0.10 0.12 0.08 0 .4 0 -0 .5 0 0.79 0.69 0.10 0 .1 6 0 .1 0 0 .5 0 -0 .6 0 1.69 1.22 0.11 0 .1 4 0 .1 2 0 .6 0 -0 .7 0 1.35 0 .7 6 0 .1 0 0 .1 7 0.11 0 .7 0 -0 .8 0 1.11 1.10 0 .1 2 0 .2 9 0 .1 3 0 .8 0 -0 .9 0 1.14 2.09 0 .1 8 0 .2 0 0 .1 4 0 .9 0 -1 .0 0 0 .8 9 2.02 0 .1 5 0 .2 8 0.21 FINAL/ INITIAL Pb ( H ) CONC E NT RAT ION , Cf /Cj 0.2 0.4 0. B 0. 0.6 2.0 1.0 1.2 1.4 1.6 1.8 0 Figure 5.7 Concentration Ratio Prifiles Across Prifiles Ratio Concentration 5.7 Figure - - - - — ” - - _ - _ - - . D A • 0 . 0. 3 0. 5 0. 7 0. 9 1.0 .9 0 .8 0 :7 0 .6 0 .5 0 .4 0 .3 0 .2 0 0.1 et No. Test 1 1 1 1 5 A 4 3 3 o 2 1 0 • OMLZD DISTANCE FROMNORMALIZED ANODE Q A o a • h Specimens the ...... B o o • Cj {/ig/g ) {/ig/g Cj 133.B 145.0 11 7.5 122.9 143.7 1 i i 0 • 1 i f A 0 « — i r 6 o # ...... a A o • . i i B A • o I 6 A • — 4 A □ • o k - * “ - - m, 87 - w - . 88 indicate that the process flushes the adsorbed Pb(13) across the clay to the cathode where it is electroplated by receiving two electrons. In shorter duration tests (tests 1 and 2), Pb(II) is redistributed across the specimens with the advancing acid front and very little reaches the cathode because of the pH in this region which can cause precipitation. Figure 5.8 compares the concentration ratio, Cf/C;, the pH and the conductivity data across the cell for tests 1 and 5. This Figure clearly shows that during the early stages of the process (test 1), as the acid front generated at the anode advances into the cell, Pb(II) is removed at the low pH environment in the first half of the cell (anode section), while it accumulates at the higher pH environment in the other half of the cell (cathode section). Conductivity is increased at the anode by the advancing H+ ions, sharply decreasing to values below the initial conductivity in the mid-section and rising back at the cathode. With further processing (test 5), pH values are consistently lowered and Pb(II) is removed across the whole cell. In this test, the conductivity at the anode section rises to higher values than in test 1 and lowers substantially below the initial values across the cathode section. This lowering of the conductivity in the cathode section is consistently observed in all longer duration tests (tests 3 to 5). This effect is also externally displayed by the decrease in the apparent conductivity calculated across the electrodes. In order to have a better understanding of the chemistry associated with the process, Pb(II) concentrations in the pore fluid also are measured in each section. 89 Table 5.5 Pb(ll) Mass Balance in the Specimens Description Total Amount of Pb(ll), mg T e st No. 1 2 3 4 5 Specim en: Initial 1B6.3 56.8 184.0 65.9 82.0 Final 160.2 54.4 19.4 12.6 8.6 Electrodes C athode 0.1 2.1 149.4 42.3 72.7 Anode <0.1 <0.1 <0.1 <0.1 <0.1 Effluent 0.1 0.1 1.0 0.6 0.5 Influent (*) (*) n C) n Error1 25.9 0.2 14.2 10.4 0.3 (Initial-Final) (’) Influent reservoir was not sampled for Pb(ll) concentration. (1) Most of the error is possibly due to the variations in water contents of blank and test specimens resulting in differences in the total lead amount calculated in each slice. Some Pb(ll) may have diffused into the influent reservoir. It is also possible that HNOg did not fully desorb the Pb(ll) in the processed specimens. PORE FLUID CONDUCTIVITY, ___ Test 5 et I Test 8 0 li 5—v 5 Fluid T 1- a ___ — : l c 1.0 h 90 91 Figure 5.9 presents the ratios of the Pb(II) concentration in the pore fluid, CP, to the final Pb(II) concentration in the soil, Cf. It is noted that a higher fraction of the Pb(II) is in the pore fluid in the anode section than in the cathode section, where the pore fluid generally contained little Pb(H); it is either precipitated in the higher pH environment as Pb(OH)2, or adsorbed on the clay surfaces. These results are in conformity with the conductivity and pH data which show an increase in the pore fluid pH and a decrease in conductivity in the cathode section. In the light of these findings, the author hypothesizes that the decrease in conductivity in the cathode section is both due to an ion depletion by precipitation of Pb(II) and migration of other ions, as well as the formation of water by the H+ and OH' ions. 5.1.1.6 Energy Expenditure Energy expenditure per unit volume of soil processed, E„, is given by: (5.2) where E is the energy (w a tt-h r) , V8 is the volume of soil mass processed (m3), R is the resistance (ohm), and t is the time (hr). In tests with constant current conditions, the energy expended is directly related to the time integral of the resistance across the cell, or it is inversely related to the time integral of the CONCENTRATION RATIO, Cp/Cf .4 0 0.2 0.6 0.8 iue . Pr Fudi-iu bI) Concentration Pb(II) Fluid/in-situ Pore 5.9 Figure ANODE FROM DISTANCE NORMALIZED p = ia P (I i te Pore Fluid the in (II) Pb Cp Final= 0.2 Ratio Profile after Electrokinetic after Profile Ratio rcsig fGoga Kaolinite Georgia of Processing .4 0 et o Symbol No. Test 0.6 4=0 0.8 1.0 92 93 apparent conductivity. This implies that the energy expenditure is substantially increased due to the evolution of the low resistance zone near the cathode . Figure 5.10 presents a plot of the energy expended per unit volume of soil versus the pore volumes of flow recorded. The total flow was only 1.06 pore volumes in test 3 and the reason for this difference is unknown. The energy expenditure in this test is clearly noted to be different than the other three tests which exhibited similar trends. In spite of this difference, Pb(H) was efficiently removed in the longer duration tests 3, 4 and 5. These tests were terminated at a time after the pH level at the cathode started decreasing. The energy expenditure at the time these tests were terminated was 60, 29 and 37 kW-hr/m3, respectively. From the test data presented, it is not possible to discern the necessary minimum energy needed to flush Pb(II); however, the results demonstrate that the energy expenditure for Pb(II) removal will be below or in the range of 29 to 60 kW-hr/m3. Figure 5.10 Energy Expended/Unit Volume versus Pore Volumes of Flow of Volumes Pore versus Volume Expended/Unit Energy 5.10 Figure ENERGY EXPENDED PER UNIT VOLUME Of SOIL ( ( kw-hr/m3) 60 40 50 20 30 10 0 Test 2 Test PORE VOLUMES FLOW OF 60 et I Test M 29 37 94 95 5.1.2 Cd(II) Cadmium metal (Cd) has two electrons in its outer shell and eighteen in the underlying shell. Cadmium is a silvery, crystalline metal resembling zinc. It has an atomic weight of 112.41 grams. A large part of cadmium produced is used in electroplating metals in protection against corrosion. Cadmium dissolves slowly in either hot, moderately dilute hydrochloric or sulfuric acid, with evolution of hydrogen. It is also dissolved by nitric acid. Four tests are conducted for removal of cadmium adsorbed on saturated Georgia kaolinite. The dimensions of the specimens, duration of tests, current density, initial concentration, porosity, total flow, and other parameters are presented in Table 5.6. 5.1.2.1 Electrical Potential Gradients and Flow Figure 5.11 presents the change in voltage across the electrodes in time for the four tests. In all these tests, the current and the dimensions of the specimens are kept constant. Therefore, any change in the potential difference implies a change in the resistance of the cell. During each test, it is noticed that the electric potential difference increases gradually to almost 20 volts to 25 volts (except in test Cd#2). It is believed that the fluctuations in electric potential difference in test Cd#2 was a result of a loose connection between the electric wires and the electrodes. It is noted that as the pH at the cathode increases the electric potential 96 Table 5.6 Summary of testing program and processing data for Cd(ll). Test Number Param eter Cd #1 Cd #2 Cd #3 Cd #4 Current (mA) 3.0 3.0 3 .0 3 .0 Dimensions Area (cm 2) 81.1 81.1 81.1 81.1 Length (cm) 10.2 10.2 10.2 10.2 Duration (hr.) 801 1027 9 5 6 71 6 Current Density (mA/cm2) 0 .0 3 7 0 .0 3 7 0 .0 3 7 0 .0 3 7 Total Charge (amp.hr.) 2.40 3 .08 2.87 2.15 Total Charge/Unit Volume of Soil 2905 3 7 2 5 34 6 7 25 9 5 (amp.-hr./m3) Porosity of the Specimen 0 .58 0.61 0 .6 2 0 .6 0 Initial Concentration 114 100 105 99 (jjg/g of dry soil) Pore Volume (cm3) 4 7 9 505 5 1 4 5 0 0 Total Flow (cm3) 1072 1090 2051 136 Pore Volumes of Flow 2 .24 2 .16 3 .98 0 .2 7 50 CD# 1 CD#2 40 30 20 10 0 0 300 600 900 1200 0 200 400 600 600 1000 1200 50 CD#3 CD#4 40 30 20 10 0 0 300 600 900 1200 0 300 600 900 1200 TIME ( hours ) TIME (hrs) Figure 5.11 Electrical Potential Difference for Cd(II) Tests 98 difference also increases. Figure 5.11 also shows that 30 hours are required to reach an electric potential difference of 20 volts in test Cd#3 while test Cd#4 required more than 200.0 hours to reach the same electric potential under the same conditions. The only difference in these tests was that, in test Cd#3 the initial pH was 4.4 compared to only 1.70 in test Cd#4 (the difference in the initial pH resulted from adding different amount of HN03 to the initial solution in order to avoid any precipitation at the sides of the mixer ). As the initial pH decreases, the conductivity of the medium increases resulting in a lower resistance for ion flow. However, in electrokinetic soil processing the pH of the cathode section increases with time resulting in a higher resistance to ion flow. Figure 5.12 presents the flow in time for the four tests. It is observed that there was no flow for the first 10 to 25 hours. The flow started at an electric potential gradient of about 1.0 v/cm or more. Figure 5.12 also indicates that the flow rate decreases in time. This phenomenon is analogous to the flow behavior observed for the Pb(II) tests. Figures 5.13 and 5.14 present the change in the empirical electro-osmotic coefficient of permeability, Kg, and electro-osmotic water transport efficiency, Kj, in these tests. A curve fitting was performed on the data of Kg and to eliminate the effects of the fluctuating voltage, current, and computer hardware and software sensitivity. These tests indicate that, (1) increases to 1.0 x lCf5 cm2/V-sec at the start of each test and FLOW ( mL ) 2100 1800 1200 1500 600 300 900 iue .2 oa Eetoomss lw C(I Tss ) Tests Cd(II) ( Flow Electro-osmosis Total 5.12 Figure 200 0 600 400 IE hus ) hours ( TIME et o 1 No. Test 800 s N. 4 No. est T et o 3 No. Test et o 2 No. Test 1000 1200 99 100 - 4 10 10"‘ CD#1 CD#2 *io- » o » a> 0) o 10'7 300 600 900 1200 300 600 900 1200 10‘4 CD#3 10'1 —110‘7 0 300 600 9001200 0 300 600 900 1200 TIME ( hours ) TIME ( hours ) Figure 5.13 Changes in Electro-osmosis Coefficient of Permeability ( Cd(II) Tests ) 101 1 Cd#l Cd#2 » -i —110'* 300 600 900 1200 0 300 600 900 1200 Cd#3 Cd#4 0 300 600 900 1200 0 300 600 900 1200 TIME ( hours ) TIME ( hours ) Figure 5.14 Changes in Electro-osmosis Coefficient of Water Transport Efficiency ( Cd(II) Tests ) 102 decreases by an order of magnitude in time (note the flow data in Figure 5.12). (2) Similarly, k, increases to 0.5 cm3/amp-sec at the start of the tests and decreases to less than 0.1 cm3/amp-sec in conformity with the flow data. The change in Kc and Kj are significantly different in test Cd#4 and Cd#3 . All specimens are loaded with the same quantity of Cd(II) and they are prepared using a similar procedure. The only difference was in the initial pH of these specimens. The author believes that because the initial pH of test Cd#3 was higher than test Cd#4, the pore fluid electrolyte concentration of Cd#3 was lower than Cd#4. Therefore, the thickness of the diffuse double layer for Cd#3 was larger than Cd#4. Table 5.7 supports this conclusion. As the thickness of the double layer increases the electro-osmotic flow would increase. Thus, the flow in test Cd#3 is much higher than test Cd#4. 5.1.2.2 Effluent pH Figure 5.15 presents a plot of the effluent pH at the cathode with respect to the number of pore volumes of flow. As a result of electrolysis, the pH at the cathode compartment rises to values of 11.0 to 12.0. Subsequently the pH remains relatively constant until 0.8 to 0.9 pore volumes of flow, then decreases with further processing. It is interesting to note that the decrease in pH at the cathode is accompanied by the decrease in flow rate, as noted in the specimens loaded with Pb(H). Table 5.7 Effect of Concentration and Valences of Ions on Thickness of the Diffuse Double Layer. Thickness of Diffuse Double Layer Electrolyte (1/K), cm Concentration mol/l Monovalent ions Divalent ions 1 x 1 0 5 1 x 10'5 0 .5 x 10 5 1 x 1 0 3 1 x 1 0 ® 0.5 x 10 6 1 x 101 1 x 10'7 0.5 x 10'7 Source : Tan, 1982 pH OF EFFLUENT pH OF EFFLUENT 14 10 12 Figure 5.15 Effluent pH versus the Pore Volumes of Flow ( Cd(II) Tests ) Tests Cd(II) ( Flow of Volumes Pore the versus pH Effluent 5.15 Figure 10 12 14 B 6 0 2 4 4 6 8 0 2 0 0 PORE VOLUMES FLON OF 1 1 2 2 3 3 CD#3 CD#1 4 4 10 12 14 14 10 12 0 6 2 4 B 0 2 6 B 4 12 3 4 3 2 1 0 0 PORE VOLUMES FLOW OF 21 3 104 4 105 5.1.2.3 pH Profiles The pH profiles across the specimens are presented in Table 5.8. These profiles are obtained upon completion of the tests. The determination of the hydrogen-ion activity or pH of soils is by far the most commonly made soil test. Conventionally, soil pH is defined by the following equation, Soil pH = - log aH+ (5.3) where aH+ is the activity of H+ in the soil suspension. The "effective" concentration of hydrogen ions includes all sources such as those arising by dissociation of soluble acids and those dissociated from soil particles. In these tests the pH values are obtained by 1- inserting pH probe directly into the soil mass ( in-situ ), 2- by using indicator paper (litmus paper with accuracy of.5 pH unit), and 3- by extracting the pore fluid out and measuring its pH value directly. It is noted that the pH across the specimen decreases with increasing the processing time, or with increasing the total charge passed per volume of specimen. The increase in pH from the anode to the cathode is a direct consequence of the advance of the acid front by migration, diffusion, and advection. These repetitive results showing the evolution of pH gradients in specimens loaded with Cd(II) as well Pb(D) reconfirm the predictions of the analytical model for the acid/base distributions in time. The acid front generated at the anode flushes the cell with time, neutralizes the base at the cathode, and decreases the effluent pH. A comparison between insitu and pore fluid pH indicates the following; Table 5.8 pH distributions across the specimens Normalized Test Distance from Cd #1 Cd #2 Cd #3 Cd #4 Cathode IS LP PF IS LP PF IS LP PF IS LP PF 0.0 -0.1 3.79 4.25 7.22 3 .03 4 .75 5.90 4.32 4.00 7.50 4 .3 4 6.00 7.95 0.1 - 0 .2 3 .60 4 .0 0 7.73 2.56 4.25 5.28 4.28 3.50 7.90 4 .3 2 5.50 7.55 0 .2 - 0 .3 3.63 3 .50 7.15 2.38 4.00 5.53 4 .2 2 3.25 7.35 4.28 3.75 7.68 0 .3 - 0 .4 3.66 3 .00 7.35 2.36 4 .00 5.20 4 .30 3.25 6.90 3.05 3.50 7.60 0.4 - 0.5 3.32 3.00 6.93 2.41 4 .0 0 4 .8 0 4 .2 4 3.25 7.20 2.50 3.25 2.90 0.5 - 0 .6 3.43 2.75 7.21 2.33 3.75 3.55 4 .13 3 .50 6.60 2 .04 2.50 2.48 0.6 -0.7 3.41 2.75 3.75 1.45 3.75 1.54 3.79 3 .00 4 .1 6 1.79 2.00 2.23 0.7 - 0.8 3.05 2.75 2.84 1.50 3.00 1.20 2.86 1.75 2.72 1.80 1.75 2.03 0.8 - 0.9 2.38 1.75 2.30 1.31 2 .00 1.21 2.62 1.50 2.62 1.54 1.50 1.88 0.9 - 1.0 2.10 1.50 1.82 1.43 1.75 1.47 2.51 1.50 2.08 1.49 1.25 1.68 IS - In-situ with pH Meter LP - Litmus Paper (accuracy is .5 pH units) PF - Pore Fluid with pH Meter 107 1. In the anode zone, the insitu and the pore fluid pH values are relatively identical and this can be explained by realizing that, " in acid solutions, the surface will be positively charged; the cation exchange capacity of kaolimte will be small and it will have a significant anion exchange capacity", (Drever, 1988). Therefore, as the acid zone advances across the soil specimen, the soil will reach a saturated point of hydrogen ions. Consequently, any insitu measurements of pH will be the measurements of activity of hydrogen ions in the pore fluid. 2. In the cathode zone, there is a substantial difference between the insitu and the pore fluid pH values. This phenomenon can be explained as follows; As long as the soil is negatively charged, the adsorption and desorption of cations will continue resulting in a high concentration of hydrogen ions in the diffuse part of the double layer. Therefore, these protons are not free to leave the soil when the water is forced out of the soil pores. This implies that higher pH will be obtained for the pore fluid than the insitu values. 5.1.2.4 Conductivity The electrical conductivity phenomena arise from the movement of ions or electrons through a conducting system under the influence of an electric field (E). Figure 5.16 presents the change in apparent conductivity across the electrodes in the experiments conducted to remove Cd(II) loaded on Kaolinite. Equation 5.1 is used to calculate the electrical conductivity depicted in Figure 5.16. The same APPARENT CONDUCTIVITY. Ka (us/cn) APPARENT CONDUCTIVITY. Ka (us/cs) 200 240 200 Figure 5.16 The Change in Apparent Conductivity with Time ( Cdfll) Tests ) Tests Cdfll) ( Time with ConductivityApparent in Change The 5.16 Figure 240 120 160 120 160 300 IE hus TM ( or ) hours ( TIME ) hours ( TIME 0 900 600 0 900300600 CD#3 CD#1 0 — o J - 1200 1200 200 240 160 200 120 240 120 160 BO 40 BO 40 0 0 300 0 0 900 600 300 0 900 600 CD#4 CD#2 1200 108 1200 109 trends are identical to the ones presented in Figure 5.5 for soils loaded with Pb(II). Table 5.9 presents the insitu and the pore fluid conductivity profiles across the cell upon completion of each test. It is clear that the calculated conductivity at the end of the tests in Figure 5.16 and the measured insitu conductivity in the cathode zone are compatible. Pore fluid conductivity is higher than the insitu conductivity as presented in Table 5.9. The same trend was observed in the blank test and in tests conducted on soil loaded with Pb(II). For more discussion the reader is referred to section 5.1 . 5.1.2.5 Efficiency of Cd(II) Removal Table 5.10 provides profiles of the ratio of final concentration of Cd(II), (Ct), in the specimens to the initial concentration, (Cj). It is interesting to note that for the four tests conducted on soil loaded with Cd(II), 92 to 100 percent removal of Cd(H) was achieved across the cell except in the test section close to the cathode. It is noted that even in test Cd#4 which had a very low KKj, and high pH, Cd(H) was removed by the process. This shows that the removal efficiency is not only due to the advection generated by electro-osmotic flow but also due to migration of contaminants by electrical gradients. The mass balance in each test is provided in Table 5.11. A significant amount of removed cadmium is electroplated at the cathode. 110 Table 5.9 Measured conductivity values across the cell Normalized Test Distance Cd #2 Cd #3 Cd #4 from Cd #1 Anode ISC PFC ISC PFC ISC PFC ISC PFC 0.0 - 0.1 61 6660 75 6060 1370 4450 1280 16300 0.1 - 0.2 51 2350 73 4530 1010 2700 1710 12940 0.2 - 0.3 55 487 70 3210 503 1070 1620 10680 0.3 - 0.4 18 397 66 560 105 81 1280 6980 0.4 - 0.5 11 360 21 40 21 67 1050 4280 0.5 - 0.6 11 108 13 49 20 70 598 1680 0 .6 - 0 .7 8 103 12 57 18 59 88 87 1 11 123 00 0 o 9 133 15 57 17 67 0.8 - 0.9 12 146 16 56 18 99 8 138 0.9 - 1.0 15 116 15 66 19 71 12 95 ISC - In-situ Conductivity PFC • Pore Fluid Conductivity Table 5 .1 0 The ratio of final to initial concentrations, Cf/Cj. Normalized Test Number Distance from Cd #4 Anode Cd #1 Cd #2 Cd #3 0 .0 - 0.1 0 .0 0 0 .0 0 0 .0 0 0 .0 7 0.1 - 0.2 0.00 0 .0 0 0 .0 0 0 .0 8 0.2 - 0.3 0.00 0.00 0 .0 0 0 .0 8 0.3 - 0.4 0.00 0.00 0 .0 0 0 .0 7 0.4 - 0.5 0.00 0 .0 0 0 .0 0 0 .0 7 0.5 -0.6 0.01 0.00 0 .0 0 0 .0 3 0 .6 - 0 .7 0 .0 0 0 .0 0 0 .0 0 0 .0 2 0 .7 - 0.8 0.01 0 .0 0 0 .0 0 0.01 0.8 - 0.9 0.01 0.00 0 .0 0 0 .0 2 0.9 - 1.0 0.13 0.01 0 .1 7 1.74 Table 5.11 Mass balance in the specimens Total Amount of Cadmium, mg Description Cd #1 Cd U2 Cd #3 Cd # 4 Specimen: Initial 102.0 9 6.7 9 5.9 8 8 .4 Final 1.6 0 .2 2.1 21.1 Electrodes Cathode 92.1 8 4 .7 83 .0 4 7 .5 Anode 0 .0 N/A N/A N/A Effluent 0 .0 0.1 0.1 0.1 Influent N/AN/A N/A 0 .7 Error* 8 .2 11.8 10.8 19 * Error = (initial - final - electrodes - effluent - influent) 113 5.1.2.6 Energy Expenditure Energy expenditure per unit volume of soil processed, Eu, is given by equation 5.2 where Eu is the energy (watt-hr/m3). This implies that the energy expenditure increases at a constant rate because, in all tests conducted, the electric potential difference across the specimen tends to stabilize at a range of 2.0 to 2.5 v/cm . Figure 5.17 presents a plot of the energy expended per unit volume of soil versus the pore volumes of flow recorded. The energy expenditure in test Cd#2 (106 kW-hr/m3) is clearly shown to be higher than the other three tests which exhibited similar trends. This is due to the longer duration of test Cd#2 than the other tests and that the electric potential was higher than the other tests, probably due to a loose connection between the electric wires and the electrodes. Tests are terminated at a time after the pH level at the cathode started to decrease. The results demonstrate that the energy expenditure at the time these tests were terminated ranged between 50 to 106 kW-hr/m3. ENERGY ( kw-hr/« ) ENERGY ( kw-hr/n ) 120 100 120 100 Figure 5.17 Energy Expended/Unit Volume versus Pore Volumes of Flow of Volumes Pore versus Volume Expended/Unit Energy 5.17 Figure 60 80 20 40 20 60 80 40 0 0 1 0 0 O E OUE O FO PORE VOLUMES FLOW OF PORE VOLUMES FLOW OF 1 2 2 CD#1 CD#3 3 3 ( Cd(H) Tests ) Cd(H)Tests ( 4 4 120 100 100 120 20 20 0 1 0 2 C D#4 CD#2 3 ItL4 4 115 5.1.3 Cr(EI) Chromium (Cr) is a transition metal which shows several oxidation states but only three are important; Cr(0), Cr(IIl), and Cr(VI). It has an atomic number of 24 and an atomic weight of 52.0 grams. The adsorption of Cr(VI) and Cr(III) by kaolinite has been investigated by Griffin, et al., (1977). They reported that the adsorption of Cr(m) and Cr(VI) species are pH dependent. The adsorption of Cr(IH) increases as pH increases, while the adsorption of Cr(VI) decreases as pH increases. Four tests are conducted for removal of Cr(III) adsorbed on saturated Georgia kaolinite. The dimensions of the specimens, current density, duration of tests, and other parameters in these tests are presented in Table 5.12. Test 1 for chromium removal was a preliminary one. Problems were encountered in measurements of electric potential, flow, and pH due to difficulties with the data acquisition software and the power supply. The current density was not constant throughout this test. 5.1.3.1 Electrical Potential Gradients and Flow Figure 5.18 presents the voltage across the electrodes in time for the three tests. In the three tests the electric potential increases gradually from 0.15 V/cm to 2.5 V/cm in less than 30.0 hours and then it stays relatively constant. As the electric potential increases to its maximum value, the pH value also increases to 116 Table 5.12 Summary of testing program and processing data for Cr(lll). Test Number Param eter Cr #1 Cr #2 Cr #3 Cr #4 Current (mA) 3.0 3.0 3.0 3.0 Dimensions Area (cm 2) 81.1 81.1 81.1 81.1 Length (cm) 10.2 10.2 10.2 10.2 Duration (hr.) 4 8 3 699 1 60 2 1191 Current Density (mA/cm2) 0.037 0 .0 3 7 0 .0 3 7 0 .0 3 7 Total Charge (amp.hr.) N/A 2 .1 0 4.81 3 .5 7 Total Charge/Unit Volume of Soil N/A 2 5 3 9 5811 4 3 2 0 (amp.-hr./m3) Porosity of the Specimen 0.62 0.62 0.62 0.64 Initial Concentration 136 123 108 134 (/yg/g of dry soil) Pore Volume (cm3) 537 5 1 1 .5 5 13 5 3 4 Total Flow (cm3) 869 70 6 7 6 8 351 Pore Volumes of Flow 1.62 1.38 1.50 0 .6 6 VOLTAGE ( volts ) 3C 50 SO 40 10 0 30 0 90 20 1500 1200 900 600 300 0 A UI I I J I l.I n.l I nh..l t .li..I II iJ I I I I JU.I i Figure 5.18 Electrical Potential Gradient ( Cr(III) Tests ) Tests Cr(III) ( Gradient PotentialElectrical 5.18 Figure IE hus ) hours ( TIME i U > « > o 20 30 50 40 10 0 30 0 90 20 1500 1200 900 600 300 0 Cr#2 IE hus ) hours ( TIME 20 30 50 40 10 0 30 0 90 20 1500 1200 900 600 300 0 Cr#4 IE hus ) hours ( TIME Cr#3 117 118 its maximum value of about 11.0 and stays relatively constant for about 0.7 to 0.8 pore volumes of flow before it starts decreasing ( Figure 5.23). Similar trends were observed in the tests conducted on soils loaded with Pb(II) and Cd(II). Figure 5.19 shows the development of the electrical potential difference across the cell in test Cr#3 in time. In this test, holes were drilled across the specimen in order to monitor the electrical potential gradients. This was done in order to assess any potential drop at the electrodes. The conductivity profiles obtained in the Pb(H) and Cd(II) tests suggested that the electrical potential gradients will be nonlinear across the cell. This test was conducted to verify the earlier hypothesis that the nonlinearity in conductivity across the cell would lead to nonlinearity in electrical potential drop. Table 5.13 indicates that the electrical potential difference measured across the electrodes is identical to the values measured across the specimen. This implies there were not significant overpotentials developed at the electrodes due to any gas generation and other secondary factors. Therefore, the current levels used in this study did not result in any significant overpotential. The electrical potential developed across the specimen is an approximate measure of the resistance ( or conductivity ) offered to ion flow under electrical gradients. Figure 5.19 depicts that as the electrical conductivity increases in the anode section, a corresponding decrease accompanies in the cathode section. As a consequence of the high conductivity in the anode section ( due to the H+ ions ), there is small electrical VOLTAGE (VOLTS) 30 24 IB 12 6 0 0 Figure 5.19 Voltage Across the Cell versus Normalized Normalized versus Cell the Across Voltage 5.19 Figure 0.1 4 hr 54. 2 . hr125. 95 hr19.5 366 hr 0.2 OMLZD ITNE FROM ANODE DISTANCE NORMALIZED itne rm Anode from Distance 0.3 0.4 0.5 0.6 0.7 0.8 0.9 119 1.0 Table 5.13 Electrical Potential Difference Voltage Difference In the Specimen Time Across the at 1.0 cm from (hours) Electrodes the Electrodes (volts) (volts) 4 2.3 2.4 13 2.8 3.3 20 3.3 3.6 40 4.8 5.7 70 15.2 16.6 100 22.2 23.7 140 22.2 23.8 170 23.0 2 4 .4 2 1 0 2 5.0 26.5 260 24.7 26.7 450 26.0 28.0 600 25.3 2 7.4 9 3 0 24.7 2 8.4 121 potential difference in this section. In conformity with the decreasing conductivity in the cathode section, the majority of the electrical potential difference across the cell is realized in this section. Nonlinearity in electrical potential gradient imply nonlinearity in advective flux due to electrical gradients. This would result in coupling between flow due to differences in hydraulic and electrical potentials. Figure 5.20 presents the flow in time for tests Cr#2, Cr#3, and Cr#4. There was no flow in the first 20.0 hours. Then the flow rate increased up to 0.8 to 0.9 pore volumes before it started to decrease. Figures 5.21 and 5.22 present the change in electro-osmotic coefficient of permeability, Kg, and electro-osmotic water transport efficiency, K; respectively. These figures indicate that: (1) Kg increases to 1.0 x 10'5 cm2/V-sec at the start of each test and then decreases by more than an order of magnitude in time (note the flow data in Figure 5.20). (2) Similarly, k, increases to 0.6 cm3/amp-sec at the start of the tests and then decreases to less than 0.1 cm3/amp-sec in conformity with the flow data. Similar trends are presented in section 5.1.1 for Pb(II) tests, in section 5.1.2 for Cd(H) tests, and by Casagrande (1983). Time dependent changes in kg and lq demonstrate that the coefficients kg and lq are not constants for a specific soil but they are time-dependent variables. FLOW ( mL ) 300 900 0 0 200 Figure 5.20 Total Flow ( Cr(III) Tests ) Tests Cr(III) ( Flow Total 5.20 Figure 400 600 IE hus ) hours ( TIME O 10 10 10 1600 1400 1200 1000 BOO Cr#4 Cr#3 122 123 - 4 Cr#2 Cr#3 -B 0 300 600 900 1200 1500 0 300 600 900 1200 1500 TIME ( hours ) TIME ( hours ) 10"4 Cr#4 ’io-9 u 10V u -10"* » S* 10'7 0 300 600 900 1200 1500 TIME ( hours ) Figure 5.21 Changes in Electro-osmosis Coefficient of Permeability ( Cr(III) Tests ) 124 1 Cr#3 » » - 3 0 300 600 900 1200 1500 300 600 900 1200 1500 TIME (hours ) TIME ( hours ) Cr#4 o ao-‘ 6 a a _ 2 _10 10-3 300 600 900 1200 1500 TIME (hours ) Figure 5.22 Changes in Electro-osmosis Coefficient of Water Transport Efficiency ( Cr(III) Tests ) 125 5.1.3.2 Effluent pH Figure 5.23 presents a plot of the effluent pH at the cathode with respect to the number of pore volumes of flow. As a result of electrolysis, the pH at the cathode compartment rises to values more than 11.0. The advance of the acid front from the anode to the cathode affects the pH values at the cathode compartment. This effect is observed to occur at 0.8 to 0.9 pore volumes of flow. It is interesting to note that the decrease in pH value is accompanied by a decrease in the flow rate. This can be explained by the fact that as the pH of the cathode compartment rises to its maximum value of 11 to 12, the pH at the anode drops to its minimum value. By advection, diffusion, and migration, the acid front advances through the cell from the anode to the cathode resulting in the acidification of the anode zone. As the pore fluid becomes more and more acidic, the clay particles become less and less negatively charged. This results in a decrease of the electro-osmotic flow rate. When the clay particles reach a zero point of charge, the electro-osmotic flow stops. 5.1.3.3 pH Profiles pH distribution across the specimens upon the completion of the tests is presented in Table 5.14. The pH values are obtained for each section by using the three methods: 1- inserting pH probe directly into the soil mass (insitu), 2- by using the indicator paper (litmus paper with accuracy of 0.5 pH unit) and 3- by pH OF EFFLUENT 10 14 12 Figure 5.23 Effluent pH versus the Pore Volumes of Flow ( Cr(III) Tests ) Tests Cr(III) ( Flow of Volumes Pore the versus pH Effluent 5.23 Figure 2 4 6 8 0 0 O E OUE O FO PORE VOLUMES FLOW OF PORE VOLUMES FLOW OF 1 UJ IL UJ £ u. 10 0 PORE VOLUMES OF FLOW 2 10 12 14 0 2 4 6 e 1 2 10 Cr#3 126 2 ■ able 5.14 pH distributions across the specimens. Normalized Test Distance from Cr* #1 Cr #2 Cr* #3 Cr #4 Cathode IS LP PF IS LP PF IS LP PF IS LP PF 0 .0 -0 .1 2.72 N/A 5.10 4.00 4.50 4.08 4.38 4.00 7.30 4.28 4.50 5.60 0.1 -0 .2 2.75 N/A 4.99 3.95 4.50 4.08 4.06 4.00 6.55 4.15 4.50 5.84 0.2 - 0.3 3.52 N/A 3.95 3.95 4.75 3.77 4.00 4.00 7.28 3.88 4.75 7.05 0.3 - 0.4 2.96 N/A 3.51 3.80 5.00 3.77 4.10 3.75 6.30 3.98 4.75 6.84 0 .4 - 0 .5 2.27 N/A 3.17 3.80 4.50 3.53 3.95 3.75 6.26 3.81 3.50 6.42 0.5-0.6 2.26 N/A 2.78 3.65 4.50 2.59 3.65 2.50 5.05 3.55 4.75 4.72 0.6-0.7 2.39 N/A 2.69 2.70 3.75 2.26 2.56 2.00 2.71 2.85 2.50 2.59 0.7 - 0.8 2.45 N/A 2.56 2.27 2.00 1.94 2.34 1.75 2.44 2.80 1.75 2.46 0.8-0.9 2.38 N/A 2.60 2.07 1.50 1.91 2.04 1.50 2.24 2.43 1.75 2.27 0.9 - 1.0 2.07 1.50 1.86 N/A 2.41 1.75 2.10 *Cr #1 and Cr #2: the cell divided into 9 segments. , 128 measuring the pH in the extracted pore fluid. The pH across the specimen decreases with an increase in the processing time or with the total charge passed per volume of specimen. The increase in pH from the anode to the cathode is a direct consequence of the advance of the acid front. A comparison between insitu and pore fluid pH indicates the following; 1. In the anode zone, the insitu and the pore fluid pH values are relatively identical. 2. In the cathode zone, there is a substantial difference between the insitu and the pore fluid pH values. The reasons for these observations were discussed in section 5.1.2. 5.1.3.4 Conductivity A discussion of electrical conductivity in electrokinetic soil processing was previously presented in section 5.1. Figure 5.24 presents the change in apparent conductivity across the electrodes in the chromium experiments. The trend is identical to the ones obtained for Pb(II) and Cd(II) loaded specimens. Table 5.15 presents the insitu and the pore fluid conductivity profiles across the cell upon completion of each test. It is clear that the calculated conductivity in Figure 5.24 and the measured insitu conductivity in the cathode zone are comparable to each other. In Table 5.15, the pore fluid conductivity is higher than tine insitu conductivity. The same trend has been observed in blank tests and in APPARENT CONDUCTIVITY. Ka (us/ca) Figure 5.24 The Change in Apparent Conductivity with Time ( Cr(III) Tests ) Tests Cr(III) ( Time with Conductivity Apparent in Change The 5.24 Figure 200 240 280 120 160 40 60 0 ■ ■ Q 30 0 90 20 50 30 0 90 20 1500 1200 900 600 300 0 1500 1200 900 600 300 0 • Cr#2 - ’ . k...... IE hus TM I or ) hours I TIME ) hours ( TIME ...... OL cr UJ f- u > cr z < u 3 M s I I 1 1 I J. I X I. I I I I I . I X I I . .J I I . 11 1 I I I 0 200 240 260 120 160 30 0 90 20 1500 1200 900 600 300 0 IE hus ) hours ( TIME 200 240 260 120 160 80 40 0 • • 1 Cr#3 ■ ’ • Cr#4 129 130 Table 5.15 Measured conductivity values across the cell Normalized Test Distance Cr #1 Cr #2 Cr #3 Cr # 4 from Anode ISC PFC ISC PFC ISC PFC ISC PFC 0 .0 - 0.1 111 6 9 0 0 530 N/A 4 5 5 7 2 5 0 60 7 5 2 0 0.1 - 0 .2 111 4 7 6 0 4 2 0 N/A 4 4 0 4 1 4 0 55 5 2 6 0 0 .2 - 0.3 111 2 5 9 0 130 N/A 3 7 0 136 0 52 2 6 2 0 0 .3 - 0.4 111 1530 12 N/A 29 8 1 .2 47 1350 0 .4 - 0.5 45 3 0 3 9 N/A 21 106 10 75 0.5 - 0.6 27 31 5 N/A 18 68 7 73 0 .6 - 0.7 25 36 16 N/A 16 111 7 75 0.7 - 0.8 31 29 18 N/A 12 98 7 82 o o O) 00 t 33 27 14 N/A 9 90 5 71 0 .9 - 1.0 11 N/A 3 53 ISC - In-situ Conductivity PFC - Pore Fluid Conductivity Cr*#1 and Cr‘#3: the cell was divided into 9 segments. 131 tests conducted on soil loaded with Pb(II) and Cd(II). For more discussion, the reader is referred to section 5.1 . 5.1.3.5 Efficiency of Cr(IH) Removal Table 5.16 provides profiles of the ratio of final concentration (Cf) of Cr(III) on the soil specimen to the initial concentration (C). It is interesting to note that in the four tests conducted on specimens loaded with Cr(HI), only 65 to 70 percent of Cr(XII) was removed in the anode zone while a large amount of Cr(III) precipitated at the cathode zone. An understanding of the behavior of Cr(III) at different pH levels is necessary for a better assessment of these lower removal levels than those for Pb(II) and Cd(II). Griffin et. al, (1977) concluded that above pH 5.0, Cr(HI) will become immobile due to precipitation. Below a pH of 4.0, Cr(III) species will be highly adsorbed by the kaolinite and therefore will have low mobility through the soil. At a pH range of 4.0 to 5.0, a combination of precipitation and adsorption processes will take place. In these tests, the acid front generated at the anode would drop the pH in the anode zone to about 2.0 ( Table 5.13). Cr(III) species will be mobile and they will migrate towards the cathode zone at this pH value. However, at the cathode zone where the pH values are higher than 3, the processes of precipitation and adsorption will take place and therefore Cr(III) species will not migrate to the cathode as in the case for Pb(II) and Cd(II). Table 5 .1 6 The ratio of final to initial concentrations, C,/Cj. Normalized Test Number Distance from Cr #1 Cr #2 Cr #3 Cr # 4 Anode t o o o 0 .3 0 0 .6 5 0 .3 0 0 .3 4 0.1 - 0.2 0.35 0.76 0 .3 8 0 .3 7 0 .2 - 0.3 0 .4 2 0 .8 2 0 .4 2 0 .4 6 0.3 - 0.4 0.55 0.83 0.51 0 .5 0 0.4 - 0.5 0.66 0.91 0.65 0 .5 6 0.5 -0.6 0.78 0.99 0 .8 7 0 .7 5 0.6 - 0.7 0.95 0.96 0.89 0.88 0.7 - 0.8 1.08 1.03 1.20 0 .9 3 1 0 bo o CD 1.69 1.08 2.76 1.05 0.9 - 1.0 2.25 1.80 Table 5.17 Mass balance in the specimens Total Amount of Chromium, mg Description Cr #1 Cr #2 Cr #3 Cr #4 Specimen: Initial 89.7 93.6 81.5 94.6 Final 73.8 94.9 71.1 74.9 Electrodes Cathode 28.7 0 .95 18.8 15.7 Anode N/D 0 .1 3 N/D 0.1 Effluent N/D N/D N/D N/D Influent N/D N/D N/D N/D Error* -12.8 -2.4 -8.4 3.9 •Error - (initial - final - electrodes - effluent - influent) N/D = Not D etectable 134 The mass balance for each test is provided in Table 5.17. A significant amount of removed chromium is again electroplated at the cathode. 5.1.3.6 Energy Expenditure The energy expenditure per unit volume of soil processed is given by equation 5.2 . Figure 5.25 presents a plot of the energy expended per unit volume of soil versus the pore volumes of flow recorded. The energy expenditure in tests Cr#2 and Cr#3 was 57, and 42 (Kw-hr/m3), respectively. The energy expenditure in test Cr#4 (118 kW- hr/m3) is higher than the other two tests which exhibited similar trends. The author believes that the reasons for this behavior are as follow; 1- the electric potential in test Cr#4 was higher than the other tests , 2- the initial pH for Cr#4 was 4.0 while the initial pH value for Cr#2 and Cr#3 were 6.5, and 3.0, respectively. 3- the duration for Cr#4 was 1191 hours while the periods for Cr#2, and Cr#3 were 619, and 1600 hours, respectively. All tests are terminated at a time after the pH level at the cathode started to decrease. The results demonstrate that the energy expenditure at the time these tests were terminated range between 42 to 118 kW-hr/m3. 120 120 Cr#2 Cr#3 100 100 80 60 40 20 0 0 1 2 0 1 2 PORE VOLUMES OF FLOH PORE VOLUMES OF FLOW 120 Cr#4 100 e c. £ i s >- octo UJ z UJ 0 1 2 PORE VOLUMES OF FLOW Figure 5.25 Energy Expended/Unit Volume versus Pore Volumes of Flow ( Cr(III) Tests ) 136 5.2 Concentration and Current Density Effects While the above tests demonstrate the feasibility of removing adsorbed inorganic cations from kaolinite by the electrokinetics technology, the complicated chemistry prevailing during the process and the behavior of different species in the changing pH environment do not allow an assessment of the feasibility under higher current densities and for higher ion concentrations. Bench-scale studies are deemed necessary to develop a better understanding of the influence of increased current densities and concentrations on the efficiency of removal. Pb(II) was selected for the purpose of studying these effects. Table 5.18 summarizes all the tests conducted in this study. Three different concentration levels are used. These concentrations were selected to be less than the cation exchange capacity of the mineral as movement of adsorbed species is within the scope of this study. The current density was selected as 0.012, 0.037, and .123 mA/cm2. In essence, the current is increased by an order of magnitude. The processing conditions for these tests are presented in Table 5.19. The reduced data pertaining to these experiments ( flow, voltage gradients, effluent pH, pH profiles, apparent conductivity, and conductivity profiles) are presented in Appendix A. An analysis of the data obtained is presented. The tests were continued until pH levels at the cathode compartment decreased to relatively steady values ( Figure A.5 ). In test Pb#hl0 the flow decreased substantially (probably due to the low initial pH value Table 5.18 Summary of Testing Program to Study Effect of Current Density and Concentration Current Concentration Concentration Test No. Density (m eq./100 (mA/cm2) (AQ/g) gm) Set #1 Pb#h3 0 .0 1 2 538 0 .5 2 Pb#h2 0 .0 3 7 625 0 .6 0 Pb#h1 0 .1 2 3 530 0.51 Pb#h4 0 .1 2 3 589 0 .5 7 Set #2 Pb#h5 0 .0 1 2 8 6 4 0 .8 3 Pb#h9 0 .0 3 7 1064 1.02 Pb#h6 0 .1 2 3 7 53 0 .7 2 Pb#h8 0 .1 2 3 835 0.81 Set #3 Ph#h10 0.012 82 0 .08 Pb#h13 0 .0 3 7 75 0 .0 7 P b # h 1 1 0 .1 2 3 68 0.0 7 Pb#h12 0 .1 2 3 75 0 .0 7 Table 5.19 Summary of Testing Program and Processing Data for Pb(ll). Test Number Param eter Pb #h1 Pb #h2 Pb #h3 Pb #h4 Pb #h5 Pb #h6 Current ImA) 10.0 3.0 1.0 10.0 1.0 10.0 Dimensions Area (cm2) 81.1 81.1 81.1 81.1 81.1 81.1 Length (cm) 10.2 10.2 10.2 10.2 10.2 10.2 Duration (hr.) 4 2 4 717 1204 411 . 1684 350 Current Density (mA/cm2) 0 .1 2 3 0 .0 3 7 0 .0 1 2 0 .1 2 3 0 .0 1 2 0 .1 2 3 Total Charge (amp.hr.) 4 .2 4 2.15 1.20 4.11 1.68 3 .50 Total Charge/Unit Volume of Soil 5120 2598 1455 4 9 6 3 2036 4231 (amp.-hr./m3) Porosity of the Specimen 0.59 0.62 0.61 0.58 0.59 0 .6 0 Initial Concentration 530 625 538 589 8 6 4 753 (pglg of dry soil) Pore Volume (cm3) 4 8 9 511 507.5 581 488 493 Total Flow (cm3) 713 604 3 0 4 4 6 3 142 584 Pore Volumes of Flow 1.46 1.18 0.60 0.80 .29 1.19 Table 5.19 (continued) Test Number Param eter Pb #h8 Pb #h9 Pb #h10 Pb #h11 Pb #h12 Pb #h13 Current (mA) 10.0 3.0 1.0 10.0 10.0 3.0 Dimensions Area (cm2) 81.1 81.1 81.1 81.1 81.1 81.1 Length (cm) 10.2 10.2 10.2 10.2 10.2 10.2 Duration (hr.) 301 567 1463 213 339 6 1 4 Current Density (mA/cm2) 0 .1 2 3 0 .0 3 7 0 .0 1 2 0 .1 2 3 0 .1 2 3 0 .0 3 7 Total Charge (amp.hr.) 3.01 1.70 1.46 2.13 3.39 1.84 T-^tal Charge/Unit Volume of Soil 3639 2 0 5 6 .4 1 7 6 8 .6 2575 4 0 9 8 2227 (amp.*hr./m3) Porosity of the Specimen 0.57 0 .58 0 .6 0 0.57 0 .57 0 .65 Initial Concentration 835 1064 82 68 75 75 (/yg/g of dry soil) Pore Volume (cm3) 4 7 4 4 8 3 4 9 3 469 4 7 4 537 Total Flow (cm3) 648 357 26 560 8 18 4 2 8 Pore Volumes of Flow 1.37 0 .7 4 0 .0 5 1.20 1.73 0 .8 0 140 of 1.8) and therefore, the experiment was terminated earlier than the decrease in pH values at the cathode compartment. Efficiency of Removal Table 5.20 provides the dimensionless concentration ratios across the specimens upon termination of the tests. Under constant current conditions and at a given initial concentration of Pb(II), the removal is expected to proceed from the anode to the cathode. Pb(H) is removed across the cell and it is precipitated in the cathode zone. The concentrations used were 7 % to 100 % of the cation exchange capacity of kaolinite and ranged between 68 /xg/g to 1060 mg/g of soil. The mass balance in the tests is presented in Table 5.21. The mass was balanced within 15 % of the initial amount of Pb(H) loaded into the specimen. This error is found reasonable for an assessment of removal efficiency. It is observed that a significant amount of removed Pb(II) is electroplated at the cathode electrode. Most of the unremoved Pb(D) was found in the cathode section particularly in the segment adjacent to the cathode electrode. Figure 5.26 shows the ratio of final to initial concentration across the specimens. There are two major factors affecting the process in removal of contaminants; namely, the time and the current density; Time: time is a very important factor because it is not yet possible to predict the timetable for the acid to breakthrough. In this study, tests were terminated when 141 Table 5 .2 0 The ratio of final to initial concentration, Cf/Cj. Normalized Test Number Distance from Pb #h1 Pb #h2 Pb #h3 Pb #h4 Pb #h5 Pb #h6 Anode 0 .0 - 0.1 0.01 0 .0 2 0 .0 7 0 .0 2 0 .1 0 0 .0 2 0.1 -0.2 0.01 0.03 0 .0 6 0.01 0 .0 9 0 .0 2 0 .2 - 0 .3 0.01 0 .0 4 0 .0 7 0.01 0.09 0.01 0 .3 - 0 .4 0.03 0.05 0.08 0 .0 2 0 .0 4 0 .0 2 0.4 - 0.5 0.12 0.05 0.07 0 .0 2 0 .0 9 0 .0 3 0 .5 - 0 .6 0 .1 0 0 .0 6 0 .0 7 0 .0 2 0 .0 7 0 .0 3 0.6 -0.7 0.05 0.06 0.06 0 .0 2 0 .0 7 0 .0 4 0 .7 - 0 .8 0 .0 4 0 .0 5 0 .0 6 0 .0 2 0 .0 5 0 .0 4 0 .8 - 0 .9 0.11 0 .2 2 0 .1 0 0.02 0.06 0.03 0 .9 - 1.0 2.75 7.22 7.97 3 .0 7 9 .9 4 .3 8 142 Table 5.20 (continued) Normalized Test Number Distance from Pb #h8 Pb #h9 Pb #h10 Pb #h11 Pb #h12 Pb #h13 Anode 0 .0 -0 .1 0.01 0 .0 5 0 .2 3 0 .2 4 0 .0 5 0.11 0.1 - 0 .2 0 .0 5 0 .0 6 0 .1 9 0 .1 7 0 .0 5 0 .1 8 0 .2 - 0 .3 0.02 0.08 0.19 0 .1 4 0 .0 6 0 .1 4 0.3 -0.4 0.03 0.10 0 .1 8 0 .2 4 0 .0 8 0 .1 5 0 .4 - 0 .5 0 .0 3 0 .1 2 0 .1 7 0 .2 7 0 .0 7 0 .2 0 0 .5 - 0 .6 0 .0 5 0 .1 3 0 .1 8 0 .2 2 0.11 0 .2 3 0.6 -0.7 0.06 0 .1 2 0 .1 6 0 .1 7 0.11 0 .2 3 0.7 -0.8 0.06 0.08 0 .1 5 0 .1 7 0 .1 3 0 .2 2 0 .8 - 0 .9 0 .0 4 0 .0 7 0 .1 6 0 .24 0 .1 2 0 .3 0 0 .9 - 1.0 6 .95 8 .4 5 5 .4 4 1.32 1.16 5 .1 5 143 Table 5.21 Mass balance in the specimens. Total Amount of Pb(ll), mg Description Pb #1 Pb #2 Pb #3 Pb #4 Pb #5 Pb #6 Specim en: Initial 5 8 7 .0 6 8 2 .0 610.1 5 92.5 1143.3 1064.0 Final 180.8 6 4 4 .4 4 2 3 .3 2 2 1 .5 1140.7 496.0 Electrodes Cathode 323.5 3 5.7 180.0 2 7 3 .4 7 6.5 3 8 5 .0 Anode 0.0 0.5 0.7 1.3 3.3 2 2.5 N/D 0.1 N/D N/D N/D N/D Influent N/D N/D N/D N/D N/D N/D Error* 8 2.7 1.3 6.1 96.3 0.7 160.5 * Error = (initial - final - electrodes * effluent - influent) 144 Table 5.21 (continued) Total Amount of Pb(ll), mg Description Pb #8 Pb #9 Pb #10 Pb #11 Pb #12 Pb # 1 3 Specimen: Initial 1154.0 1237.3 93.1 86.0 96.8 100.7 Final 863.2 1069.0 63.5 26.5 19.7 67.9 Electrodes Cathode 2 7 5 .0 192.3 26.3 5 0.4 69.7 4.3 Anode 10.3 6.3 1.2 4 .4 N/D 2.0 Effluent N/D N/D 1.2 N/D N/D N/D Influent N/D N/D 0.3 N/D N/D N/D Error* 4.5 -24.3 0.6 9.1 7.4 2 6 .4 •Error = (initial - final - electrodes - effluent - influent) 145 103 I TE o .012 1453 13 & A .037 614 24 • .123 339 155 ■ .123 213 108 * 102 is C (i) = 68-82 ug/g r»») rij ^ IS •5 io CJ 0.2 0.4 0.6 0.Q 1.0 NORMALIZED DISTANCE FROM ANQpE Figure 5.26 Concentration Ratio Profiles Across the Specimens for Tests 10, 11, 12, and 13 (1= Current Density mA/cm2, T= Time (hr.), E= Energy kW-hr/m3 ) 146 103 • i T E £L i : o .012 1200 27 . A .037 717 45 ft .123 424 240 w .123 411 310 1 » * 102 • » * c (i) = 530-625 ug/g e *rt A CJ ■ <0 • •Sio ft e o O O : ° $ o A ^ 0 A A ft A A f t A • • * ft ft ft ft ft ft — . — ...... J...... A. 0.2 0.4 0.6 0.8 1.0 NORMALIZED DISTANCE FROM ANODE Figure 5.26 Concentration Ratio Profiles Across the Specimens for Tests 1, 2, 3, and 4 (1= Current Density mA/cm2, T= Time (hr.), E= Energy kW-hr/m3 ) 147 103 I T E X ■ o .012 16P4 11 • A .037 567 28 • .123 350 165 ■ .123 301 108 *10a 0 O A A ■ ■ O ■ m o O • • ■ ■ ■ • • i — * i ^ ■ * ■ * ■ » 0 0.2 0.4 0.6 0.8 1.0 NORMALIZED DISTANCE FROM ANODE Figure 5.26 Concentration Ratio Profiles Across the Specimens for Tests 5, 6, 8, and 9 (1= Current Density mA/cm2, T= Time (hr.), E — Energy kW-hr/m3 ) 148 the pH at the cathode compartment decreased below a pH of 7. The drop in pH was due to the advance of the acid front from the anode to the cathode by diffusion and advection. Current Density: Tests with higher current densities have higher removing efficiencies than tests with lower current densities in the same period of processing. The flow rate increases as the current density increases. As a result of this increase, the acid front moves at a faster rate to the cathode by the advection, reducing the processing time. The total amount of charge input per unit volume of soil, E (kW-hr/m3), is a parameter which combines the influence of current density together with the time of processing. It is noted that as the current density increases, the time of processing decreases but the total energy expended is substantially increased. Figure 5.26 also provides the following implications; 1- High removal efficiencies were achieved in tests with concentrations varying from 68 fxg/g to 1064 ug/g of Pb(II). 2- In all these tests, a large amount of Pb(II) was found precipitated adjacent to the cathode. The amount of precipitation of Pb(II) at the last segment close to the cathode was higher for tests with lower current densities. This indicates that, further processing is needed even when the effluent pH decreases below 7.0. 3- The relationship between the current density and the time needed to remove the contaminant is not linear,( i.e. it took about 420 hours to remove 99 % of Pb(II) 149 in test Pb#hl and Pb#h4 where the current density was 0.123 mA/cm2, w'hile it took only 716 hours to remove 98% in test Pb#h2 with a current density of 0.037 mA/cm2.). The same result is observed in the other tests. Energy Expenditure Energy expenditure per unit volume of soil processed is given by equation 5.2 for tests with constant current conditions. It is directly related to the time integral of the electric field across the cell. Figure 5.27 presents the plots of the energy expended per unit volume of soil versus time. The energy expenditure in tests with high current densities was higher than the tests with lower current densities by more than one order of magnitude (i.e., the energy expended for Pb#h4 was 300 kW-hr/m3 in 400.0 hours while the energy expended for Pb#h2 was 27 kW-hr/m3 for the same period of time). The slope of the energy expenditure plots given in Figure 5.27 would give the power consumption. The power consumption is similar in tests conducted at the same level of current density, regardless of the level of concentration. This implies that power consumption will only change with current density rather than concentration of Pb(II). However, the author notes that such a generalization would only be true for concentrations lower than the cation exchange capacity of the mineral. For example, the energy expended in test Pb#h6 with an initial concentration of about 800 /xg/g is about 130 kW-hr/m3 in 300 hours while the ENERGY ( kw-hr/m3 ) ENERGY ( kw-hr/»3 ) 100 150 200 250 300 350 50 SO 30 40 60 10 0 0 0 0 iue .7 nry xeddUi Vlm vru Time versus Volume Expended/Unit Energy 5.27 Figure 300 200 IE hus TM ( or ) hours ( TIME ) hours ( TIME IE hours ( TIME 600 0 600 400 o et 1 2 3 ad 4 and 3, 2, 1, Tests for 900 Pb#h3 ) 1200 BOO 200 250 100 300 350 150 20 30 50 10 60 0 0 20 0 60 BOO 600 400 200 0 200 IE hus ) hours ( TIME 600400 Pb#h4 150 BOO ENETOY * k*i-hr/»3 ) ENERGY ( kw-hr/«3 ) 1B0 100 120 160 20 40 60 BO 20 30 25 35 10 15 0 0 5 0 30 0 90 20 1500 1200 900 500 300 0 Figure 5.27 Energy Expended/Unit Volume versus Time versus Volume Expended/Unit Energy 5.27 Figure 200 IE hus TM ( or ) hours ( TIME ) hours ( TIME IE hus TM ( or J hours ( TIME ) hours ( TIME 400 o et , , , n 9 and 8, 6, 5, Tests for 600 O 0 BOO 100 120 160 180 140 20 10 25 30 15 35 0 5 0 200 200 400 400 Pb#hB 0 800 600 0 BOO 600 151 152 35 35 Pb#h10 Pb#hl3 30 30 m ■ 25 25 20 20 15 15 CD cc 10 10 UJ 5 5 0 0 0 300 600 900 1200 1500 200 400 600 BOO TIME ( hours ) TIME ( hours i 1B0 180 160 Pb#hll 160 140 140 > 120 120 100 CD UJ 0 200 400 600 BOO 0 200 400 600 BOO TIME ( hours ) TIME ( hours ) Figure 5.27 Energy Expended/Unit Volume versus Time for Tests 10, 11, 12, and 13 153 energy expended for test Pb#hl2 with an initial concentration of about 72 jug/g is 130 Kw-hr/m3 for the same period of time. The results demonstrate that the energy expenditure at the time these tests were terminated ranges from 11 to 310 kW-hr/m3. Table 5.22 provides a summary of the results of the tests conducted to study the effects of current density and concentration. These results indicate that, (1) K,. vales for twelve tests conducted in this section ranging from 1.0 x 10 5 to 6 x 10'7 cm2/V-sec. It increased at the start of each test and decreased by an order of magnitude or more during the process. (2) Similarly, K* increased to 0.5 cm3/amp-sec at the start of the tests and decreased to less than 0.01 cm3/amp-sec in conformity with the flow data. The maximum values of Kc were plotted versus the current density in Figure 5.28. The figure indicates that changing the current density by one order of magnitude does not significantly affect the values of maximum Kg. It shows that maximum Kc ranges between 3X10"6 to lOxlO"6. Therefore, the change in the initial concentration does not affect Ke values and any change can be ignored for concentrations below the cation exchange capacity of the soil. Figure 5.29 presents the average pore fluid conductivity at the end of each test versus the current density. This figure clearly demonstrates that the final conductivity of the medium is a function of the current density. It increases by an 154 Table 5.22 Results of Tests Conducted to Study the Effects of Current Density and Concentration. Test No. Parameter Pb#h3 Pb#h2 Pb#h1 Pb#h4 Concentration 538 625 530 589 Cug/g) Current Density 0.012 0.037 0.123 0.123 (mA/cm2) Duration (hr.) 1200 717 424 411 Energy 27 45 240 310 (Kw-hr/m3) Effluent pH 6.4 3.7 4.5 6.0 (initial) Effluent pH 7.2 7.5 5.8 5.8 (final) Voltage (volts) 2 2 - 1 8 2 2 - 1 8 58 - 40 98 - 44 (V(max) - V(final)) E.O Permeability 50 - 5 30 - 10 30 - 10 20-6 (K.x 1 0 7) E.O Efficiency 0.35 - 0.08 0.2 - 0.06 0.11 -0.04 0.1 - 0.08 (K,) Flow (cm3) 304 604 713 463 155 Table 5.22 (continued) Test No. Parameter Pb#h5 Pb#h9 Pb#h6 Pb#h8 Concentration 864 753 835 1064 U/g/g) Current Density 0.012 0.037 0.123 0.123 (mA/cm2) Duration (hr.) 1684 567 350 301 Energy 11 28 165 301 (Kw-hr/m3) Effluent pH 5.2 N/A 6.0 N/A (initial) Effluent pH 9.5 7.0 6.2 6.2 (final) Voltage (volts) 5 - 5 14 - 10 50 - 25 - 65 40 - 25 - 35 (V(max) - V(final)) E.O Permeability 30 - 20 60-20 40 - 6 30-15 (K. x 10'7) E.O Efficiency 0.30 - 0.08 0.3 - 0.04 0.15 - 0.04 0.1 - 0.04 (K.) Flow (cm3) 142 357 584 648 156 Table 5.22 (continued) Test No. Parameter Pb#h10 Pb#h13 Pb#h11 Pb#h12 Concentration 82 75 68 75 {(JQ/Q) Current Density 0.012 0.037 0.123 0.123 (mA/cm2) Duration (hr.) 1463 614 213 339 Energy 13 24 108 155 (Kw-hr/m3) Effluent pH 1.8 3.8 5.0 4.0 (initial) Effluent pH 11.6 7.0 6.2 6.2 (final) Voltage (volts) 1 - 18* 13 - 12 58 - 40 60 - 28 - 38 (V(max) - V(final)) E.O Permeability 3 95 - 10 30-20 35-18 (K„ x 10-7) E.O Efficiency 0.01 0.5 - 0.05 0.12 - 0.06 0.15 - 0.05 (Kj) Flow (cm3) 26 428 560 818 •Voltage continuously increased 157 10* ♦ o u>OJ a i E CJ. „-a " - iO x (O C (initial), ug/g Q) ♦ 735 to 1064 A 530 to 625 • 68 to 82 10 -7 0.05 0.10 0.15 Current Density. mA/cm2) Figure 5.28 The Maximum Electro-Osmosis Coefficient of Permeability versus Current Density av. Pore Fluid Conductivity, us/cm 2000 3000 1000 4000 5000 6000 7000 8000 Figure 5.29 Avreage Pore Fluid Conductivity at the End of Each of Endthe Conductivityat Fluid Pore Avreage 5.29 Figure 0 00 00 00 01 0.15 0.12 0.09 0.06 0.03 0 ♦ ▲ Test vresus Currentvresus Density Test urn Dniy (mA/ Density, Current .J...... 8 5 o 1064 to 753 ♦ 8t 82 to 68 • A (i ti ), ug/g , l) ia it in ( C 3 o 625 to 530 cnf) ♦ ▲ ▲ ♦ 158 159 increase in the current density, which increases the rate of production and introduction of H+ and OH' ions into the medium. As a result of these increases the electrolyte concentration of the medium increases. The increase in the electrolyte concentration would result in a decrease in the thickness of the double layer, reducing the electro-osmotic water transport efficiency. Figure 5.30 presents the maximum values of the electro-osmotic water transport efficiency Kj, (cm3/amp.sec), as a function of current density. The substantial decrease in K, with increasing the current density is depicted in this figure. A better understanding to this decrease is provided by the empirical equation given in 2.1. ( Kg I )/a = IQ I (5.4) Since Kg is relatively constant in tests with different current densities, (Fig. 5.28), and that the conductivity of the medium increases by increases in the current density, (Fig. 5.29), then K; is inversely related to the current density. The power consumption is given by the following equation; P = dEu/dt = V I = I2 R = V-la (5.5) Therefore, in tests conducted at constant current densities, it is expected that the power consumption change with the conductivity of the cell. However, the overall conductivity of the cell decreases at early stages of the process and reaches a relatively constant value (Figure A.6). Thus, any change in the power Ki (max), (cm /amp.sec) Figure 5.30 Maximum Electro-osmosis Water Transport Efficiency, Transport Water Electro-osmosis Maximum 5.30 Figure 0.1 0.2 . I ♦ ♦ ♦ I- 0.3 0.4 .5 0 ▲ urn Dniy (mA/cm2) Density. Current .5 .0 0.15 Density Current versus 0.10 0.05 (initial), ug/g , ) l a i t i n i ( C 5 t 1064 to 753 ♦ 3 t 625 to 530 ± 8 o 82 to 68 • I 160 161 consumption should be function of I2. Figure 5.31 presents the current density versus the power per unit volume of soil used in these tests. It indicates that increasing the current density from to I2 will increase the power consumption per unit volume of soil by a factor of^/Ii)2, (where I is DC current, amp). It is possible to predict the cost of the process for a certain period of time by using Figure 5.31. However, this data is valid for only the tests conducted in this study. The above results demonstrate the feasibility of removing contaminants at different current densities and at concentration up to the CEC of the clay mineral. The study shows that lower energy expenditure will be achieved by keeping the current densities low and by extending the processing time. Power, P, (kW/m3) iue .1 oe osmto prUi oue fSi versus Soil of Volume Unit per Consumption Power 5.31 Figure 00 00 00 01 0.15 0.12 0.09 0.06 0.03 0 the Current Density Current the II ■ ■ ■ ■ ■ ■ I I * urn Dniy (mA/cm ) Density, Current 5 t 1064 to 753 + 3 t 625 to 530 A 6 o 82 to 66 • (i tal ug/g . l) a it in ( C 1 —■ - 162 CHAPTER VI SUMMARY AND CONCLUSION 6.1 IMPLICATIONS Electrochemistry The data enable the author to suggest the following reasons for the decrease in electro-osmotic flow rate and the increase in the electrical potential gradient across the cell during the electro-osmosis: (1) The significant increase in the electrolyte concentration due to the diffusion and advection of the H+ ions into the specimen at the anode and the saturation of the clay surface by H+ ion is the fundamental reason for the decrease in flow rate in time. The resistance offered to current flow will be very low in zones with high hydrogen ion concentrations, (2) The decrease in (the mobile) electrolyte concentration due to the increase in pH and due to the removal of anions by migration and cations by precipitation at the cathode section increases the resistance to current flow. As a consequence, the electrical potential across the cell increases. The major implication of these findings is that the electrical potential gradients are nonlinear across the cell. This nonlinearity has been noted in previous studies (Wan and Mitchell, 1976). This hypothesis is further verified by figure 5.19. Most of the electrical potential drop is in the cathode section of the cell where the conductivity is low. If the permeability of the specimen is low, 163 there will be an insufficient supply of pore fluid to the cathode zone from the anode zone leading to suction and a consolidation process in the anode region. Previous studies have reported suction pressures (Esrig and Gemeinhardt, 1968) and consolidation (Acar, et ai., 1991) in tests conducted with an open electrode configuration, supporting this hypothesis. Lockhart (1983) demonstrated that the efficiency of flow in the process can depend on the pH of the porous medium. The results of the present study clearly substantiate this earlier observation. When low pH conditions are promoted at the anode, electrolyte concentration rises rapidly at the anode section of the specimen and gradually across the cell, leading to a decrease in potential drop, decreasing flow from the anode to the cathode. The boundary conditions for flow (open or closed electrode configuration) and current density, together with the initial electrolyte concentration in the soil, will result in the development of significantly different chemistry across the electrodes. Efficiency This study indicates that several variables affect the efficiency of removing contaminants from soils by electro-osmosis: (1) Chemistry Generated at Electrodes: Low-pH conditions generated at the anode cause desorption and ionization of most heavy metals and inorganic chemicals. 165 However, a flux of high H+ ion concentration also results in the increase in conductivity, decrease in the thickness of the double layer, and decrease in electro-osmotic flow. The pH conditions at the anode and the cathode may be strictly controlled and adjusted for continued flow. This adjustment depends on the cation exchange capacity of the soil, the type and concentration of the chemicals in the soil and the initial pH of the medium. (2) Water Content of the Soil: Electro-osmotic flow is promoted at higher water contents (Gray, 1970). Therefore, at this state of knowledge regarding this process, high moisture content deposits should be favored. (3) Saturation of the Soil: Recent data on compacted soils (Mitchell and Yeung, 1991) indicate that k,. changes slightly with backpressure. Although there is no other factual data, this suggests that the technique can be used in partially saturated deposits by supplying a pore fluid at the anode. (4) Behavior of Primary Chemicals in the Soil at Different pH Conditions: The chemistry in the system is governed by the pH gradients across the soil mass. A knowledge of the behavior of the primary chemicals in different pH environments is necessary for a better understanding of the efficiency and the decision on the required processing conditions and time. (5) Type and Concentration of Chemicals in the Soil: An increase in the conductivity of the specimen results in a decrease in the resistance offered to current flow and hence the ion flow is governed more by diffusion and migration. 166 Soils with low initial ionic strengths favor high electro-osmotic efficiencies. (6) Current Density: At an inert anode and for 100 % Faradaic efficiency for water oxidation, the density of the current controls the flux of H+ ions. The control of current dehsity is critical to ensure the optimal electro-osmotic efficiency and contaminant removal. The cathodic current density and species available in its vicinity establish the efficiency of the reduction processes. These vary to a greater extent than the anode process because the pH and the species reaching the cathode will vary with processing time. (7) Conditioning: Similar to the changes in current density, the pore fluid at the anode and cathode compartments can be conditioned to a specific pH or chemistry to increase the efficiency of the process. This study demonstrates that electrokinetic soil processing can be used efficiently to remove inorganic cations from soils and explains the fundamental underlying mechanisms by which the process occurs. The results also lead the author to suggest that it is essential to control the extent of the electrode reactions and influent/effluent chemistry in order to improve the efficiency of the electro- osmotic flow when the technique is used as a remediation process. Although the concentrations of Pb(D) used in this study varied between 70 fig/g to 1060 /*g/g, Cd(II) concentration was 100 to 114 ng/g, and Cr(III) concentration was 100 /xg/g of dry kaolinite, a recent pilot-scale field study demonstrates that the process effectively decreased Cu(IH), Pb(II), Zn(II), and As(IlI) at concentrations of 200 167 ppm to 5000 ppm by more than 50 percent (Lageman, 1989). These preliminary studies confirm that the process has potential for large scale applications in remediation. The soils processed by Lageman (1989) had conductivities of 2000 to 3000 /*siemens/cm, significantly higher than the Pb(H), Cd(II) , or Cr(m) loaded kaolinite used in this study ( 80 to 600 /jsiemens/cm). Lageman (1989) used a very high current density of 0.8 mA/cm2. This caused excessive heating of the soil mass from 12°C to 40°C. Lageman notes that conductivity increased by a factor of two during processing. The energy expended during an eight week period of processing increased to 288 kW-hr/m3. However, the author observed that the removal process is not necessarily made more efficient by merely increasing the current densities. At this state of knowledge regarding this innovative remediation technology, it is apparent that a better understanding of the relation between the electrochemistry presented and the electro-osmotic flow is necessary for efficient use and optimization of the process. Electro-osmotic flow and energy expenditure are controlled by the variations in conductivity across the cell. An understanding of the reasons for the conductivity decrease at the cathode section and its relation to flow conditions is essential. It is noted that the process temporarily acidifies the soil mass. There also exists the need to establish the time and conditions required for restoration of a particular site to its original pH conditions. 168 6.2 CONCLUSIONS The following conclusions are obtained from this study which investigated Pb(II), Cd(II), and Cr(TII) removal from saturated kaolinite specimens. (1) The acid front generated at the anode in electro-osmosis flushes across the specimen ultimately, decreasing the pH of the effluent, (2) In response to the advance of the acid front across the cell, significant variations occur in chemistry and conductivity across the cell. The pore fluid and specimen conductivities increase in time at the anode section of the cell, (3) The conductivity in the cathode section decreases in time in response to changing electrolyte concentrations. This decrease in conductivity is displayed externally by the increase in net resistance and the voltage needed across the electrodes to maintain the constant current, (4) The coefficient of electro-osmotic permeability, and electro- osmotic water transport efficiency, k,, decrease in time as a response to an increase in voltage across the cell and/or the decrease in flow, (5) Experiments indicate that the analytical model for acid/base distributions provides an understanding for the development of pH gradients in electrokinetic processing of soils, 169 (6) Pb(II) at levels of 70 jxg/g to 1060 jxg/g of dry kaolinite is efficiently removed by electrokinetic soil processing. The removed Pb(lI) is found mostly electroplated at the cathode. The removal efficiency exceeded 95 % of the initial concentration in the anode zone. (7) Cd(II) at levels of 100 /xg/g to 114 /xg/g of dry kaolinite is efficiently removed by electrokinetic soil processing. The removed Cd(II) is found mostly electroplated at the cathode. The removal percentage of Cd(II) in the anode zone was more than 98%, (8) Cr(lH) at levels of 63 /xg/g to 81 /xg/g of dry kaolinite is partially removed by electrokinetic soil processing. Part of the removed Cr(III) is found electroplated at the cathode. The rest was found at accumulated in the cathode segments of the specimen, (9) The removal of heavy metals is due to flushing of the acid generated at the anode across the specimen resulting in desorption of inorganic metals together with its migration and advection by electro-osmotic flow, (10) In short duration tests in which the acid front does not completely flush across the specimen, Pb(II) is removed at the acidified anode section, while it accumulated at the cathode section, 170 (11) The energy expended in tests at higher current densities was higher by cne order of magnitude than the tests conducted at lower current densities, (12) The energy expended in tests in which Pb(II) is effectively removed, varied between 10 kW-hr/m3 to 306 kW-hr/m3 mainly depending upon the current density, (13) The energy expended in tests in which Cd(II) is effectively removed, varied between 50 kW-hr/m3 to 106 kW-hr/m3 mainly depending upon the duration of the test, (14) The energy expended in tests in which Cr(III) is partially removed, varied between 39 kW-hr/m3 to 118 kW-hr/m3 mainly depending upon the duration of the test, (15) Similar degrees of removal was achieved in tests conducted at lower densities and longer times when compared with tests conducted at higher densities and shorter times. Substantially higher energy expenditure was necessary in tests at higher current densities. (16) Initial concentrations of contaminants below the cation exchange capacity have no major affect on the efficiency of electrokinetic soil processing. The chemistry and flow in electro-osmosis are dependent upon the electrolytes generated at the electrodes and the initial chemistry in the specimen. 171 The physico-chemical interactions, ion flow associated with the current, and the geotechnical properties of soils after electro-osmotic flow are strongly influenced by the chemistry developed in the system. The process could be used in practice for flushing soils with acid, or other types of electrolytes and also for prospecting minerals in deposits. Electrokinetic soil processing has the potential to be used in soil remediation in the field as well as in the laboratory. REFERENCES Acar, Y. B., Gale, R. J. (1986) "Decontamination of Soils Using Electro- Osmosis," A proposal submitted to the Board of Regents of the State of Louisiana, LEQSF Research Development Program Office of Research Coordination, Louisiana State University. Acar, Y. B., Haider, L. (1990), "Transport of Low Concentration Contaminants in Saturated Fine - Grained Earthen Barriers," Journal of Geotechnical Engineering, ASCE, pp. 1031-1052. Acar, Y. B., Gale, R. J., Putnam, G., Hamed, J. (1988), "Development of pH Gradients in Electrochemical Processing of Soils," Annual Report presented to the Board of Regents of Louisiana, 119 p. 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, pp. 25-38. Acar, Y. B., Gale, R. J., Putnam, G., Hamed, J., Wong, R. (1990), "Electrochemical Processing of Soils: Theory of pH Gradient Development by Diffusion and Linear Convection," Journal of Environmental Science and Health, Part (a); Environmental Science and Engineering, Vol. 25, No. 6 ,pp. 687-714. Acar, Y. B., Hamed, J., Gale, R. J., and Putnam, G. (1991), "Acid/Base Distributions in Electro-Osmosis," Transportation Research Record, No. 1289, (in print). Acar, Y. B., Hamed, J. " Electrokinetic Soil Processing in Waste Remediation/Treatment: Synthesis of Available Data," Bulletin of the Transportation Research, Record No., (Submitted, 1990). AFESC 0985), "FY86 - R&D Project Descriptions ESL Environics Division," Engineering and Services Laboratory, Air Force Engineering and Services Center, Tyndall Air Force Base, July 1985, 87 p. Amba, S. A., Chilingerian, G. V., and Beeson, C. M. (1964), "Use of Direct 172 173 Electrical Current for Increasing the Flow Rate of Reservoir Fluids during Petroleum Recovery," Journal of Canadian Petroleum Technology, vol. 3, No. 1, pp. 8-14. Baneijee, S., and Mitchell, J. K. (1980), "In-situ Volume Change Properties by Electro-Osmosis Theory," JGED, ASCE, Vol. 106, No. GT4, pp. 347-365. Baneijee, S., Homg, J., Ferguson, J. F., Nelson, P. O. (1990), "Field Scale Feasibility of Electro-Kinetic Remediation," Unpublished Report Presented to USEPA, Land Pollution Control Division, RREL, CR811762-01, 122 p. Begeman, H.K.S. (1953), "The Influence of a Direct Current Potential on the Adhesion Between Clay Particles and Metal Objects - Laboratory and Full Scale Tests", Proceedings of the Third International Conference on Soil Mechanics and Foundation Engineering, Vol. 1, pp. 89-93. Bjerrum, L., Mourn, J., and Eide, O. (1967), "Application of Electroosmosis on a Foundation Problem in a Norwegian Quick Clay," Geotechnique, Vol. 17, No. 3, pp. 214-235. Butterfield, R., and Johnston, I. W. (1980), "The Influence of Electro-osmosis on Metallic Piles in Clay," Geotechnique, vol. 30, No. 1, pp. 17-38. Casagrande, L. (1947), "The Application of Electro-osmosis to Practical Problems in Foundations and Earthwork," Department of Scientific and Industrial Research, Building Research, London, England, Technical Paper No. 30, 22 p. Casagrande, L. (1952), "Review of Past and Current Work on Electro-Osmotic Stabilization of Soils," December 1953, with supplement of June 1957, Harvard Soil Mechanics Series No. 38. Casagrande, L. (1983), "Stabilization of Soils by Means of Electro-Osmosis - State of the Art," Journal of Boston Society of Civil Engineers, ASCE, V. 69, 2, pp. 255-302. Case, F. N. and Cutshall, N. H. (1979), "Oak Ridge National Lab., TN (USA)," Conf. 791112-28, 5 p., Symposium on the Scientific Basis for Nuclear Waste Management, Boston, MA, USA, Nov. 26-29. Chappell, B. A., Burton, P. L. (1975), "Electro-Osmosis Applied to Unstable Embankment," JGED, ASCE, Vol. 101, No. Gt8, pp. 733-740. 174 Drever, J. I. (1988), The Geochemistry of Natural Waters. Prentice Hall, New Jersey. 2nd Ed, 437 p. Esrig, M. I., Henkel, D. J. (1966), "The Use of Electrokinetics in the Raising of a Submerged, Partially Buried Metallic Objects," Department of the Navy, Research Project No. RR-004-01-01. Esrig, M. I. (1967), "Application of Electrokinetics in Grouting," Reprint No. 480, ASCE National Meeting on Structural Engineering, Seattle, Washington. Esrig, M. I., Gemeinhardt, J. P., (1967) "Electrokinetic Stabilization of an Illitic Clay," JSMFE, ASCE, 93, SM3, Proceedings Paper 5235, pp. 109-128, May, (with discussions by D. Edwards and R. Lewis, No. SMI, pp. 332-334, Man., 1968). Esrig, M. I. (1968), "Pore Pressures, Consolidation and Electrokinetics, "Journal of the Soil Mechanics and Foundation Division ASCE, Vol. 94, No. SM4, pp. 899-921. Gillham, R. W., Cherry, J. A., (1982) "Contaminant Migration in Saturated Unconsolidated Geological Deposits," Geological Society of America, Special Paper 189, pp. 31-62. Gray, D. H., and Mitchell, J. K. (1967), "Fundamental Aspects of Electroosmosis in Soils," Journal of the Soil Mechanics and Foundation Division, ASCE, Vol. 93, No. SM6, pp. 209-236. Gray, D. H. (1970), "Electrochemical Hardening of Clay Soils," Geotechnique, 20, 1, pp. 81-93. Griffin, R. A., Anna, K. Au., and Frost, R. R. (1977), " effect of pH on Adsorption of Chromium from Landfill- Leachate by Clay Minerals. J Environ. Sci. Health, A12(8), pp. 431-449. Hamed, J., Acar, Y. B., Gale, R. J. (1991), "Pb(II) Removal from Kaolinite by Electro-kinetics," ASCE, Journal of Geotechnical Engineering February, 1991 (in print). Hamnet, R. (1980), "A Study of the Processes Involved in the Electro- Reclamation of Contaminated Soils," MS Thesis, University of Manchester, England, 84 p. 175 Helmholtz, H.L.F., " Studies of Electric Boundary Layers", Annalen d. Physik, Leipzig, Germany, Vol. 7, pp. 337-382, 1879, Translated by P. E. Bocquet, Two Monographs on Electrokinetics, University of Michigan Engineering Research Institue, Ann Arbor, Bulletin of Department of Engineering Research, No. 33, pp. 5-47, October, 1951. Ibl, N. (1983) "Comprehensive Treatise of Electrochemistry," Vol. 6 - Electrodics: Transport, Chapt. 1, E. Yeager, J. O’M. Bockris, B. E. Conway, and S. Sarangapani (eds.), Plenum Press, NY. Jacobs, H. S., and Mortland, M. M. (1959), "Ion Movement in Wyoming Bentonite During Electro-Osmosis," Proceedings of Soil Science Society, Vol. 23, No. 2342,Lafayette, In. Krizek, R. J., Gularte, F. B., and Hummel, P. B. (1976), "Stabilization of Polluted Dredgings by Electro-osmosis," ASCE National Water Resources and Ocean Engineering Convention, San Diego, CA, April 5-8, 1976, Preprint 2641. Lageman, R. (1989) "Theory and Practice of Electro-Reclamation, "NATO/CCMS Pilot Study, Demonstration of Remedial Action Technologies for Contaminated Land and Ground Water, Copenhagen, Denmark, May 9, 1989, 18 p. Lockhart, N. C. (1983a), "Electroosmotic Dewatering of Clays: I. Influence of Voltage," Colloids and Surfaces, 6, No. 3, pp. 229-238. Lockhart, N. C. (1983b), "Electroosmotic Dewatering of Clays: II. Influence of Salt, Acid and Flocculants," Colloids and Surfaces, 6, No. 3, pp. 239-251. Lockhart, N. C. (1983c), "Electroosmotic Dewatering of Clays: HI. Influence of Clay Type, Exchangeable Cations and Electrode Materials, "Colloids and Surfaces, 6, No. 3, pp. 243-269. Mitchell, J. K., (1970) "In-Place Treatment of Foundation Soils," JSMFD, ASCE, 96, SMI, pp. 73-109. Mitchell, J. K. (1976), Fundamentals of Soil Behavior. John Wiley and Sons, New York, 422 p. Mitchell, J. K., and Wan, T-Y (1977), "Electro-osmotic Consolidation—Its Effects on Soft Soils," Proceedings of the IX ICSMFE, Tokyo, Vol. 1, pp. 219-224. 176 Mitchell, J. K., and Baneijee (1980), "In-situ Volume Change Properties by Electro-osmosis Evaluation," ASCE, JGE, Vol. 106, GT4, pp. 367-384. Mitchell, J. K. (1986), "Potential Uses of Electro-kinetics for Hazardous Waste Site Remediation," Position paper prepared for USEPA-University of Washington Workshop on Electro-Kinetic Treatment and its Application in Environmental- Geotechnical Engineering for Hazardous Waste Site Remediation, Seattle, WA, August 4-5, 1986, 20 p. Mitchell, J. K., Yeung, T-C. (1991), "Electro-kinetic Flow Barriers in Compacted Clay," Transportation Research Records, 1289 (in print). Probstein, R. F., and Renauld, P. C. (1986), "Quantification of Fluid and Chemical FLow in Electrokinetics," Position paper prepared for USEPA- University of Washington Workshop on Electro-Kinetic Treatment and its Application in Environmental-Geotechnical Engineering for Hazardous Waste Site Remediation, Seattle, WA, August 4-5, 1986, 20 p. Puri, A.N. and B. Anand. (1936), "Reclamation of Alkali Soils by Electrodialysis", Soil Science, Vol. 42, pp.23-27. Puri, A. N. (1949), Soils, Their Physics and Chemistry, Reinhold Publishing Corporation, New york, pp. 134-140. Putnam, G. (1988), "Development of pH Gradients in Electrochemical Processing of Kaolinite," MS Thesis presented to the Department of Civil Engineering, Louisiana State University, 256 p. Renauld, P. O., Probstein, R. F. (1987), "Electro-osmotic Control of Hazardous Waste," Physicochemical Hydrodynamics, v. 9, No. 1/2, 1987, pp. 345-360. Rowe, K. R., (1987) "Pollutant Transport through Barriers," Geotechnical Practice for Waste Disposal, Geotechnical Special Publication No. 13, ASCE, Edited by R. D. Woods, pp. 159-189. Runnels, D. D., 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, v. 6, pp. 81-91, Summer 1986. Schmid, G. (1950), "Zur Electrchemie Feinporiger Kapillarsystems", Zhumal fur Elektrochemie, Vol. 54, p. 424. 177 Schmid, G. (1951), "Zur Electrchemie Feinporiger Kapillarsystems", Zhumal fiir Elektrochemie, Vol. 55, p. 684. Segall, B. A., O’Bannon, C. E., and Matthias, J. A. (1980), "Electro-Osmosis Chemistry and Water Quality," Journal of the Geotechnical Engineering Division, ASCE, Vol. 106, No. GT10, Oct. 1980, pp. 1143-1147. Shmakin, B. M. (1985), "The Method of Partial Extraction of Metals in a Constant Current Electrical Field for Geochemical Exploration," J. Geochem. Explor., Vol. 23, No. 1, pp. 27-33. Smoluchowski, Von, M., "Electric Endosmosis and Streaming Currents", Handbuch De Elektrizitat und der Magnetismus, Vol. 2, Leipzig, pp. 366-428, 1921, Translated by P.E. Bocque, University of Michigan, Engineer Research Bulletin, Vol. 33, pp. 47-158. Spiegler, K. S. (1958), ""Transport Processes in Ionic Membranes," Transactions of the Faraday Society, Vol. 54, pp. 1408-1428. Steude, J., Viani, S., Baker, K. (1989), "Emerging Technologies for the Remediation of Radioactive Soils," Roy F. Weston, Inc., Walnut Creek, CA, Technical Report for the USEPA Office of Radiation Programs, 64 p. Tan, K. H., (1982), Principles of Soil Chemistry. M. Dekker, New York, 267 p. Thompson, R. P. (1971), "The Effect of Electro-Osmosis on the Penetration and Extraction Resistance of Model Piles in Clay," Master of Science Report, University of Southampton, England. Thompson, R. T. (1989), "The Effect of Secondary Reactions on the Electrokinetic Treatment of a Silty-Sand Soil," M.S. Thesis, The University of Texas at Austin, Civil Engineering Department, 115 p. Wan, T-Y., Mitchell, J. K. (1976), "Electro-Osmotic Consolidation of Soils," JGE, ASCE, 102, GTS, pp. 473-491, (With discussions by I. W. Johnston and T. V. Pavate, 103, GT4, pp. 335-357, and Closure, pp. 289-291, Feb 1978). Wan, Tai-Yeu, Mitchell, J. K. (1976,1978), "Electro-Osmotic Consolidation of Soils," Journal of the Geotechnical Engineering Division, ASCE, Vol. 102, No. GT5, pp. 473-491, May, 1976, (With discussions by I. W. Johnston and T. V. Pavate, Vol. 103, No. GT4, pp. 335-357, and Closure, pp. 289 291, Feb 1978). 178 Yukawa, H., Chigira, H., Hoshiro, T., and Iwata, M. (1971), "Fundamental Study on Electro-osmotic Filtration; Effect of Electric Field on Flow Rate of Permeation," Journal of Chemical Engineering of Japan, Vol. 4, No. 4, pp. 370- 376, Nagoyal, Japan. APPENDICES 179 APPENDIX A EXPERIMENTAL RESULTS FOR Pb(II), Cd(II), AND Cr(III). 180 Table A.1 Concentration of Cd(ll), pg/g Normalized Test Number Distance from Cd #1 Cd #2 Cd #3 Cd #4 Anode Initial Final Initial Final Initial Final Initial Final 0.0 0.1 79.4 11.2 74.7 6.7 72.9 .1 66.8 4.4 0.1 0.2 77.6 39.9 74.0 4.3 72.2 .1 65.5 5.1 0.2 0.3 77.6 12.3 74.4 5.1 71.9 .1 65.7 5.5 0.3 0.4 78.5 19.4 74.2 5.5 71.6 .1 66.0 4.3 0.4 0.5 77.9 22.4 74.0 8.0 72.3 .1 66.0 4.6 0.5 -0.6 76.8 79.9 75.0 8.2 71.5 .1 67.4 2.0 0.6 0.7 77.3 21.4 74.5 9.9 71 .1 66.6 1.0 0.7 0.8 77.8 53.4 73.9 9.5 71.5 .2 65.8 .6 0.8 0.9 78.1 112 74.6 11.8 71.7 .4 65.8 1.0 0.9 1.0 79.1 1267 75.0 88.4 72.9 12.6 67.5 117.4 Table A.2 Concentration of Cr(lll), ng/g Normalized Test Number Distance Cr #4 from Cr #1 Cr #2 Cr #3 Anode Initial Final Initial Final Initial Final Initial Final 0.0 - 0.1 71.9 21.9 75.4 48.7 64.2 19.2 81.2 28 0.1 - 0.2 71.7 25.4 73.8 56.2 63.4 24.4 78.1 29.3 0.2 - 0.3 72.3 30.6 72.7 59.5 63.1 26.7 76.4 35 0.3 - 0.4 73.4 40.5 72.0 59.8 64.4 32.9 75.3 37.4 0.4 - 0.5 74.4 48.9 73.0 66.5 65.9 43 76 42.8 0.5 - 0.6 75.5 58.6 73.0 72.5 65.3 56.6 76.5 57.5 0.6 - 0.7 76.6 73 74.9 71.7 66.0 58.8 78.5 69.2 0.7 - 0.8 78.9 84.9 74.7 76.8 66.8 79.9 75.9 70.8 0.8 - 0.9 81.3 137.4 74.4 80.5 66.8 184.4 76.2 80.2 0.9 - 1.0 75.1 169.3 76.8 138.4 Table A.3 pH distributions across the specimens. Normalized Test Distance from Pb #h1 Pb #h2 Pb #h3 Pb #h4 Cathode IS LP PF IS LP PF IS LP PF IS LP PF 0.0 - 0.1 4.86 4 .7 5 7.60 4.40 4.25 7.70 5.00 5.00 8.44 4.31 4.25 7.45 0.1 - 0.2 4.65 3.50 6.70 4.50 4.25 7.80 4.62 3 .75 7.07 4.15 3.75 6.68 0.2 - 0.3 3 .52 3 .0 0 5.97 4.50 3.25 7.30 4.01 3 .0 0 6.00 4.01 3.5 0 6.70 0.3 - 0.4 2.95 2 .0 0 3 .2 4 4.50 3.00 6.45 3.29 2.25 3.29 3 .70 3.00 6 .60 0 .4 - 0 .5 2 .34 2 .0 0 2.59 2.70 2.25 3.03 2.96 2.00 3 .04 2.23 1.75 2.54 0.5 -0.6 2.00 2.00 1.77 2.50 1.50 2.22 2.71 1.50 2.71 1.68 1.25 1.53 0.6-0.7 1.92 1.75 1.83 2.22 1.50 1.90 2.57 1.50 2.56 1.51 1.25 1.35 0.7 - 0.8 1.64 1.50 1.73 2.00 1.50 1.66 2.48 1.50 2.42 1.34 1.25 1.30 0.8 - 0.9 1.66 1.25 1.52 2.82 1.50 1.72 2.41 1.50 2.35 1.40 1.25 1.21 0.9 - 1.0 1.57 1.00 1.63 1.62 1.25 1.47 2.34 1.50 2.27 1.18 1.50 1.20 Table A.3 (continued) Normalized Test Distance from Pb #h5 Pb #h6 Pb #h8 Pb #h9 Cathode IS LP PF IS LP PF IS LP PF IS LP PF 0.0-0.1 4.35 5.5 0 6.90 4.30 5.50 8.15 4.70 5.00 7.95 4.48 5.00 7.69 0.1 - 0 .2 3.7 5 5 .0 0 7.90 3.90 4.00 8.50 3.35 4.25 8.07 4.06 3.25 7.69 0.2 -0.3 3.83 N/A 7.44 3.81 3.75 3.32 2.70 3.80 3.28 3.05 2.25 3.79 0.3 - 0 .4 2.83 N/A 3.83 2.95 2.75 2.54 2.13 1.75 2.31 2.34 1.75 2.65 0.4-0.5 2.59 N/A 3.02 1.95 2.50 2.04 2.09 1.50 2.08 2.18 1.50 2.36 0.5 -0.6 2.42 N/A 2.68 1.81 1.75 1.88 2.02 1.50 1.99 1.96 1.50 2.17 0.6-0.7 2.29 N/A 2.55 1.62 1.50 1.62 1.97 1.50 1.92 1.90 1.50 2.13 1 0 o bo 2.25 N/A 2.40 1.68 1.50 1.62 1.90 1.50 1.36 1.86 1.50 2.05 0 .8 - 0.9 2.23 N/A 2.22 1.67 1.50 1.77 1.87 1.25 1.77 1.80 1.50 1.93 0.9 - 1.0 2.50 1.50 2.20 1.76 1.50 2.20 1.91 1.50 2.08 1.61 1.25 2.07 oo4^ Table A.3 (continued) Normalized Test Distance from Pb #h10 Pb # h 1 1 Pb #h12 Pb *h13 Cathode IS LP PF IS LP PF IS LP PF IS LP PF 0 .0 - 0 .1 4.82 5.00 8.41 4.25 5.00 8.49 4.52 N/A 8.50 3.96 N/A 8.13 0.1 - 0.2 4.64 3.50 8.51 40.5 4.50 8.19 4.15 N/A 8.41 3.49 N/A 4.83 0 .2 - 0 .3 4.50 3.25 8.44 4 .1 0 3.75 7.86 3 .3 4 N/A 3.41 2.59 N/A 2.70 0 .3 - 0.4 3.40 3 .2 5 3.49 3 .5 0 3.00 5.80 2.37 N/A 2.25 2.28 N/A 2.28 0 .4 - 0 .5 2.76 2.75 2.79 3.10 2.00 2.67 2.15 N/A 2.04 2.17 N/A 2.15 0 .5 - 0 .6 2.52 2 .5 0 2.51 2.49 1.75 2.67 2.07 N/A 1.82 2.14 N/A 2.09 0 .6 - 0 .7 2.33 2.25 2.26 2.05 1.50 1.97 2.07 N/A 1.89 2.13 N/A 2.01 0 .7 - 0 .8 2.20 2.25 2.10 1.99 1.50 1.91 2.08 N/A 1.89 2.06 N/A 1.95 } o 00 o CO 2.08 2.0 2.03 1.76 1.25 1.91 2.00 N/A 1.83 2.01 N/A 1.74 0 .9 - 1.0 1.93 1.50 1.96 1.93 1.75 1.72 1.78 N/A 1.84 2.05 N/A 1.83 oo Table A.4 Measured conductivity values across the cell. Normalized T est D istance from Pb #h1 Pb #h2 Pb #h3 Pb #h4 Pb #h5 Pb #h6 A node ISC PFC ISC PFC ISC PFC ISC PFC ISC PFC ISC PFC 0 . 0 - 0 . 1 118 1 1 6 8 0 68 8 7 4 0 3 6 3 5 1 7 0 5 1 9 1 1 0 2 0 87 6 2 0 0 8 4 1 7 6 0 0 0.1 - 0.2 38 1 1 5 3 0 68 7500 358 4440 510 1 1 5 2 0 87 5 3 1 0 8 4 1 6 6 0 0 0 .2 - 0 .3 59 9 7 2 0 67 5750 338 3600 505 1 0 0 1 0 9 0 4330 84 11320 0 .3 - 0 .4 57 7 7 3 0 66 4 0 7 0 3 0 5 2 8 3 0 481 8 9 1 0 91 3 1 5 0 8 5 1 1 2 2 0 0 .4 - 0 .5 58 5 9 0 0 65 2600 255 2070 420 6100 90 2380 83 9000 0 .5 - 0 .6 12 2 9 8 0 56 461 115 1120 230 782 82 1220 7 4 6 4 1 0 0 .6 - 0 .7 2 0 941 2 5 51 89 343 26 64.4 77 341 12 1 8 6 0 0 .7 - 0 .8 10 167 12 42 27 5 0 12 115 6 2 7 8 14 1 4 2 0 .8 - 0 .9 10 87 11 5 5 9 28 10 4 6 16 78 9 4 5 0 .9 - 1.0 11 101 11 55 9 4 7 15 4 7 12 158 6 6 4 oo Os Table A.4 (continued) Normalized Test D istance from Pb 0 . 0 - 0 . 1 6 4 1 4 3 0 0 87 8 1 5 0 N/A 9 4 6 0 9 2 1 4 1 0 0 51 1 0 1 3 0 1 3 5 0 1 1 7 7 0 0.1 -0.2 63 15700 87 9 7 0 0 N/A 7 7 0 0 97 1 0 7 5 0 4 9 11850 63 10760 0 .2 - 0 .3 97 1 4 7 0 0 9 0 7 8 5 0 N/A 6 4 0 0 9 6 1 0 9 4 0 51 9 9 5 0 6 5 9 6 5 0 0 .3 - 0 .4 103 1 3 3 0 0 91 6 5 0 0 N/A 4 3 9 0 9 0 9 0 0 0 53 9 9 0 0 1 0 0 0 8 2 4 0 0 .4 - 0 .5 9 8 1 0 9 7 0 • 9 0 5 5 0 0 N/A 3 0 4 0 86 7340 50 8940 1225 6820 0 .5 - 0 .6 98 8 9 6 0 8 2 3 9 7 0 N/A 1 5 5 0 8 5 2 3 5 0 4 7 6100 1095 5540 0 .6 - 0 .7 95 4960 77 2090 N/A 349 71 124 46 4 8 4 0 4 2 3 4 9 0 0 .7 - 0 .8 8 0 7 7 0 6 2 3 5 7 N/A 4 0 19 7 4 4 4 7 9 6 4 2 1171 0 .8 - 0 .9 19 98 16 5 8 N/A 30 17 1 22 2 0 18 4 11 6 8 0 .9 - 1.0 10 123 12 58 N/A 73 14 152 18 2 0 0 12 83 Table A.5 Concentration of Pb(ll), fjglg. Normalized T est D istance from Pb #h1 Pb #h2 Pb #h3 Pb #h4 Pb #h5 Pb #h6 A node Initial Final Initial Final Initial Final Initial Final Initial Final Initial Final 0 . 0 - 0 . 1 4 3 5 .4 4.1 5 2 2 .0 1 1 .0 4 9 3 .4 3 3 .4 4 8 8 .9 1 0 .0 8 7 0 .0 84.0 797.0 13.0 0.1 -0.2 431.1 5.3 521.0 14.0 491.9 30.4 487.4 6.9 870.0 76.0 799.0 19.0 0 .2 - 0 .3 4 3 1 .4 6 .4 5 1 7 .0 1 9 .0 4 9 1 .4 3 3 .4 4 8 5 .0 6 .7 8 7 0 .0 7 5 .0 8 0 0 .0 1 1 .0 0 .3 - 0 .4 4 3 0 .7 11.5 515.0 24.0 290.8 37.0 489.3 7.6 870.0 36.0 800.0 14.0 0 .4 - 0 .5 4 3 0 .8 5 1 .2 5 1 5 .0 2 4 .0 4 9 0 .2 36.6 486.3 10.4 8 7 0 .0 7 4 .0 8 0 0 .0 2 6 .0 0 .5 - 0 .6 4 3 1 .6 4 2 .4 5 1 4 .0 3 0 .0 4 9 1 .6 3 2 .0 4 8 6 .6 8 .7 87.0 64.0 800.0 25.0 0 .6 - 0 .7 4 3 2 .0 20.1 514.0 31.0 491.9 2 9 .7 4 8 8 .3 11.6 870.0 57.0 7 9 9 .0 3 2 .0 0 .7 - 0 .8 43 1 .1 17.9 5 1 8 .0 24.0 493.8 28.7 488.1 11.0 870.0 46.0 799.0 33.0 0 .8 - 0 .9 4 3 2 .0 4 8 .0 5 1 5 .0 1 1 4 .0 4 9 4 .0 4 9 .6 4 8 6 .0 11.3 870.0 49.0 802.0 24.0 0 .9 - 1 .0 4 3 6 .6 1200 521.0 3768.0 497.9 3703 489.0 1500.0 870.0 8660.0 7 9 8 .0 3 5 0 1 .0 Table A.5 (continued) Normalized T est D istance from Pb #h8 Pb 0h9 Pb #h10 Pb #h11 Pb #h12 Pb #h13 Anode Initial Final Initial Final Initial Final Initial Final Initial Final Initial Final 0 . 0 - 0 . 1 8 9 3 .0 13.0 971.0 51.0 7 1 .0 16.0 66.0 16.0 7 2 .0 3 .5 7 2 .8 8 .0 0.1 -0.2 896.0 42.0 966.0 56.0 71.0 13 .0 6 6 .0 1 2 .0 7 2 .0 4 .0 7 2 .6 9 .0 0 .2 - 0 .3 8 9 7 .0 2 2 .0 966.0 81.0 70.0 13.0 66.0 9.0 72.0 4.0 7 2 .6 1 0 .0 0 .3 - 0 .4 8 9 8 .0 2 4 .0 9 6 5 .0 9 5 .0 70.0 13.0 66.0 1 6 .0 7 2 .0 5.5 72.6 11.0 0 .4 - 0 .5 8 9 8 .0 2 8 .0 965.0 120.0 70.0 12.0 66.0 177.0 72.0 5.0 7 2 .7 1 4 .0 0 .5 - 0 .6 8 9 8 .0 4 9 .0 9 6 5 .0 1 2 4 .0 7 0 .0 1 2 .0 6 6 .0 1 4 .0 7 2 .0 8.0 72.6 17.0 0 .6 - 0 .7 8 9 6 .0 53.0 966.0 117.0 70.0 11.0 6 6 .0 1 1 .0 72.0 8.0 72.7 1 7 .0 0 .7 - 0 .8 8 9 7 .0 5 7 .0 9 6 8 .0 7 6 .0 70.0 10.0 66.0 1 1 .0 7 2 .0 9 .0 7 2 .7 1 6 .0 0 .8 - 0 .9 8 9 6 .0 3 7 .0 9 6 8 .0 7 2 .0 7 0 .0 1 1 .0 66.0 16.0 72.0 9.0 7 2 .7 1 5 .0 0 .9 - 1.0 8 9 5 .0 6 2 2 0 .0 974.0 8227.0 71.0 383.0 66.0 88.0 72.0 83.0 72.8 375.0 oo vo VOLTAGE ( volts I VOLTAGE (volts) 100 20 50 60 90 30 40 70 BO 10 Figure A.l Electrical Potential Gradient ( Tests, 1, 2, 3, and 4) and 3, 2, 1, Tests, ( Gradient Potential Electrical A.l Figure 0 0 \r^\ Pb#h3 f 200 IE hus ) hours ( TIME 400 600 i h # b P BOO 100 \ r i * j ■ « » 1 ■ ■ 1 » * « ■ » * » j * j 0 40 600 400 200 IE hus ) hours ( TIME Pb#h4 Pb#h2 800 190 VOLTAGE ( volts ) VOLTAGE ( volts ) 30 20 25 20 50 60 70 30 40 10 15 10 iue . Eetia Ptnil rdet(Tss 5 6 8 ad 9) and 8, 6, 5, Tests, ( Gradient Potential Electrical A.l Figure 5 0 0 0 0 400 200 IE hus TM ( or ) hours ( TIME ) hours I TIME IE hus ) hours ( TIME 0 10 1600 1200 600 400 600 Pb#h5 Pb#h6 800 30 20 50 60 40 70 10 25 30 10 15 0 0 5 0 0 40 0 BOO 600 400 200 200 IE hus ) hours ( TIME 0 0 BOO 600 400 at Pb#h8 191 192 30 Pb#hl0 Pb#hl3 25 *• 20 20 15 iu CD £ 10 - 10 5 ■ * 1 1 ‘ 1 ‘ 1 1 ‘ 1 1 « -«—J Q 0 300 600 900 1200 1500 0 200 400 600 BOO TIME ( hours ) TIME ( hours ) Pb#hll 70 60 a so 50 40 30 ^ 20 20 10 0 0 200 400 600 BOO 0 200 400 600 BOO TIME ( hours ) TIME ( hours ) Figure A.l Electrical Potential Gradient (Tests, 10, 11, 12, and 13) FLOW ( mL ) 900 BOO 700 600 500 400 300 200 100 Figure A.2 Total Electro-osmosis Flow ( Tests, 1, 2, 3, and4) 3, 2, 1, TotalTests, Electro-osmosis ( FigureA.2 Flow 100 200 Pb#hl 300 IE hus ) hours ( TIME Pb#h4 400 500 Pb#h3 600 700 BOO 193 FLOW ( ML ) 900 BOO 700 600 500 400 300 200 100 iue . oa lcr-soi Fo (Tss 5 6 8 ad 9) and 8, 6, 5, Tests, ( Flow Electro-osmosis Total A.2 Figure 100 Pb#h8 200 300 IE hus ) hours ( TIME Pb#h6 Pb#h9 400 500 600 Pb#h5 700 BOO 194 FLOW ( ML ) iue . oa lcr-soi Fo (Tss 1, 1 1, n 13) and 12, 11, 10, TotalTests,Electro-osmosis ( Figure A.2 Flow 300 200 400 600 700 500 BOO 900 100 0 0 0 200 100 300 IE hus ) hours ( TIME 0 500 400 Pb#hl2 l3 h # b P 0 700 600 lO h # b P 800 195 196 10' Pb#h3 Pb#h2 10' 10' 10-7 200 400 600 BOO 10- 4 Pb#hl *lo' u w V*10 cu UB -10’ » 10‘7 200 400 600 BOO 0 400 600 BOO TIME ( hours ) TIME ( hours ) Figure A.3 Changes in Electro-osmosis Coefficient of Permeability ( Tests, 1, 2, 3, and 4) 197 10' Pb#h5 u •> cu » um » -10* » 10-7 200 400 600 200 400 600 BOO TIME ( hours } TIME ( hours ) Pb#h8 Pb#h6 cu 5* 0 200 400 600 600 200 400 600 BOO TIME I hours ) TIME ( hours ) Figure A.3 Changes in Electro-osmosis Coefficient of Permeability ( Tests, 5, 6, 8, and 9) 198 10’ Pb#nlO *i0* i u » a> » » v. » cu ■ u ' “I O'* 0 10'7 200 400 600 200 400 600 800 TIME ( hours ) TIME ( hours I Pb#ll Pb#hl2 u cu 200 400 600 800 0 200 400 600 800 TIME ( hours ) TIME ( hours ) Figure A.3 Changes in Electro-osmosis Coefficient of Permeability (Tests, 10, 11, 12, and 13) 199 1 Pb#h3 Pb#h2 10° 0 200 400 600 BOO 200 400 600 BOO 1 Pb#hl Pb#h4 So- io - £ 10 y .» SO'* 200 600400 BOO 0 200 400 600 800 TIME (hours ) TIME ( hours ) Figure A.4 Changes in Electro-osmosis Coefficient of Water Transport Efficiency ( Tests, 1, 2, 3, and 4) 2 0 0 Pb#h5 Pb#h9 So-1 £ 10 m 6 U j ^JO*8 10* 200 400 600 600 200 400 600 BOO TIME ( hours ) TIME ( hours ) Pb#hB d o m B U .2 JO* 10” 0 200 400 600 BOO 2000 600400 BOO TIME i hours ) TIME ( hours ) Figure A.4 Changes in Electro-osmosis Coefficient of Water Transport Efficiency ( Tests, 5, 6, 8, and 9) 2 0 1 l Pb#hlO P b # h l3 u »o-‘ i 10 V. m ■ 3 o - 10- 10- io- 200 400 600 BOO 200 400 600 BOO TIME ( hours ) TIME I hours ) 1 P b # h l2 10‘* * en IB •JO* 10 200 400 600 BOO 0 200 400 600 BOO TIME ( hours ) TIME ( hours ) Figure A.4 Changes in Electro-osmosis Coefficient of Water Transport Efficiency (Tests, 10, 11, 12, and 13) pH of Effluent pH of Effluent 14 12 14 12 0 2 6 6 6 B 2 0 0 0 oe oue o Fo Pr Vlms f Flow of Volumes Pore Flow of VolumesPore Figure A.5 Effluent pH versus the Pore Volumes of Vlow of Volumes Pore the versus pH Effluent A.5 Figure 0.5 0.5 1.0 1.0 ( Tests, 1, 2, 3, and 4) and 3, 2, 1, Tests, ( Pb#hl 1.5 1.5 2.0 2.0 10 12 14 10 12 14 0 2 6 e 4 0 2 6 4 6 0 0 .5 0 .5 0 1.0 1.0 Pb#h4 1.5 1.5 2.0 2.0 2 0 2 14 ------14 Pb#h5 Pb#h9 12 ■ 12 *0 10 B ■ B 6 * 6 4 * 4 2 • 2 Q ■ 1 « « 1 --- -» 0 0 0.5 1.0 1.5 2.0 0 0.5 1.0 1.5 2.0 14 Pb#h8 12 10 B 6 4 2 0 0 0.5 1.0 1.5 2.0 0 0.5 1.0 1.5 2.0 Pore Vo 1 lines of Flow Pore Volumes of Flow Figure A.5 Effluent pH versus the Pore Volumes of Vlow { Tests, 5, 6, 8, and 9) 204 14 12 Pb#hl3 10 8 6 a . 4 2 1 ■ 1 • ‘ ‘ Q 0.025 0.050 0.075 0.100 0 0 .5 1.0 1.5 2.0 Pore Volumes of Flow Poi 14 Pb#hl2 12 £ 10 10 8 as 6 4 2 0 0 .5 2.0 0 0 .5 1.0 1.5 2.0 Pore Volumes of Flow Pore Volumes of Flow Figure A.5 Effluent pH versus the Pore Volumes of Vlow (Tests, 10, 11, 12, and 13) APPARENT CONDUCTIVITY. Ka (us/cn) APPARENT CONDUCTIVITY. Ka (ua/cii) 350 200 250 300 300 350 250 200 100 150 100 150 50 50 Q 0 0 30 0 90 20 20 0 60 800 600 400 200 0 1200 900 600 300 0 0 » . i' ■ j . L K (initial) Ka ■ Figure A.6 The Change in Apparent Conductivity With Time With Conductivity Apparent in Change The A.6 Figure a iiil-9 ua/ca (initial)-294 Ka 200 IE hus TM ( or ) hours ( TIME ) hours ( TIME IE hus TM I or I hours I TIME ) hours ( TIME 0 0 ■ -399 ua/cn ( Tests, 1, 2, 3, and 4) and 3, 2, 1, Tests, ( 600400 Pb#h3 Pb#hl . . BOO 300 350 200 250 250 350 200 300 150 100 150 100 50 0 b I m » • » .I ■ ,.JI a iiil -2 ua/ca -322 (initial) Ka a (initial) Ka 200 JL l I I B I I I -l. J 400 -242 Pb#h4 0 800 600 Pb#h2 ua/ca 205 Figure A.6 The Change in Apparent Conductivity With Time With Conductivity Apparent in Change The A.6 Figure APPARENT CONDUCTIVITY. Ka (us/cm) APPARENT CONDUCTIVITY. Ka (us/ca) 200 250 300 350 5 450 400 450 250 350 150 200 300 100 150 50 o 50 0 o 0 30 0 90 20 50 20 0 60 800 600 400 200 0 1500 1200 900 600 300 0 • «42S us/es (initial) Ka ’ Pb#h6 ^ » • - ■ m » • » ■ Ka 0 40 600 400 200 IE hus ) hours ( TIME IE hus TM ( or ) hours ( TIME ) hours { TIME (Initial) ua/ca *315 (Initial) ( Tests, 5, 6, 8, and 9) and 8, 6, 5, Tests, ( Pb#h5 BOO 400 200 250 300 350 100 150 200 250 300 350 100 150 50 50 0 0 » » • m ► • » • » m m » 0 40 0 BOO 600 400 200 IE hus ) hours ( TIME Pb#hB Pb#h9 206 207 I » I 500 ; PbihlO 500 1 P b#hl3 m ! Ka (initial)* 452 ua/ca ! Ka (initial)- 422 ua/ca 400 £ 400 » » S 300 » 300 » » ► 200 • 200 100 :V 100 >------. . 0 0 300 600 900 1200 1500 200 400 600 BOO TIME ( hours ) TIME ( hours ) u "S. m P b # h l2 P b # h ll 500 m 500 » » * 400 • Ka (initial)-396 ua/ca 400 • Ka (Initial)- 448 ua/ca » £ • > 300 a 300 • 200 200 m » 100 100 » 0 0 200 400 600 BOO 200 400 600 BOO I TIME ( hours ) TIME ( hours ) Figure A.6 The Change in Apparent Conductivity With Time ( Tests, 10, 11, 12, and 13) ■i APPENDIX B TWO-DIMENSIONAL FINITE ELEMENT METHOD FORMULATION FOR THE ANALYTICAL MODEL 208 209 APPENDIX B TWO-DIMENSIONAL FINITE ELEMENT METHOD FORMULATION FOR THE ANALYTICAL MODEL B.l VARIATIONAL FORMULATION The differentia] equation to be solved is given by, Dx C „ + Dy C yy - V, C, - Vy C , = R C (B-l) with flux or constant concentration boundary conditions. The equation was developed in one-dimension; however, the terms in y direction are added for a complete numerical treatment. This equation contains a strong variational principle. The integral equation corresponding to equation B.l is given as, 1 = lv |D, C „ + D„ Cyy - V, C , - v y C , - R C] dV (B. 2 ) where V constitutes the volume integral over the problem domain. The first variation of this integral equation is given by, 6,1 = 5 C Jv [Dx C xx + Dv Cyy - Vx C>x - Vy Cy - R C] dV (B.3) and since the energy must be minimized, the first variation must give, 5,1 = 0 (B.4) Equation B.3 can be further simplified to decrease the order of differentiation by integration by parts. This operation is carried out in order to select a lower order approximation function. Taking only the diffusion term in x-direction. lv 5 C D xx dV = - Jv Dx 5 C x 5 C x dV + D x 8 C C x | b (B.5) The last term in equation (B.5) represents the boundary condition. It is noted that for zero flux conditions, the term C x drops out, while for constant concentration, the term 8C drops out. However, in case the flux at a boundary is specified, this term will not drop out and contribute to the solution. When a similar formulation is done for the Y term and it is substituted in equation B.3, 6,1 = - Jv Dx 8 C x C x dV - Jv Dy 8 Cy C y dV - Jv Vx 8 C C x dV - Jv Vy 6 C Cy dV - Jv R 8 C C dV + JA D x 8 C Cx dA + JA D y 8 C Cy dA = 0 (B.6) where A represents the area integral. At this point, it is necessary to discretize the domain, select local approximation functions, form local stiffness matrices representing each integral and add the contribution of each to a master stiffness matrix and solve the problem. The procedure used will be explained in the subsequent sections. B.2 DOMAIN DISCRETIZATION AND LOCAL APPROXIMATION FUNCTIONS Two dimensional 4-node quadrilaterals are selected in order to discretize the domain. Figure B.l depicts the quadrilateral and the selected shape functions which also give the local coordinate system. It is noted that since the shape functions and the local coordinate system are identical, the element used is isoparametric. Hence, (x) = (N)t (x,) (B.7a) 211 N = _L (1 - 1) 0 - s) 1 4 n = j_ a + o (i - s) 2 4 N = 1 (1 + 0 (1 • s) 3 4 N = JL (1 - 1) (I + s) 4 4 Note: A counter clockwise system is adopted in element stiffness matrix generation Figure B.l Element Formulation and the Shape Functions. 212 {y} = (N}T {y,} (B.7b) {c} = {N}T {c,} (B.8) where Xj, are the global values of the nodal coordinates and Cj is the nodal concentrations in the specific element. The first variation of the integral equation given in equation B.6 requires that Cx and Cy values be calculated. Since the concentrations in one element are discretized by equation B.8, the spatial derivatives x and y are now on the shape functions. {CJ = {N J T {c,} (B.9a) {CJ = {N JT {Cj} (B.9b) However, the shape functions are dependent upon the new local coordinates t and s. Consequently, by chain rule of differentiation: dN, = 9 A/, dx + dNt dy dt dx dt dy dt dN, _ dN, dx + dN, dy 9s dx ds dy ds In matrix notation, equation B.10 could be expressed as where [J] represents the Jacobian matrix. The requested x and y derivatives can then be evaluated as: ( d } dx dt = M-1 d d W [a sj where dy By c>S dt (B.13) M-’ - PT dx dx ds dt and the determinant | J I is called the Jacobian which is given by, ■ . dx dy dx dy 1 1 It ds" ds dt The Jacobian matrix is evaluated by (B.15) * = £m'\ N,*, (B.16) y = E x- /= i where x,, are the nodal coordinates. B.3 LOCAL MATRICES The variational formulation given in equation (B.6) can now be formed at the local element k as, 5,1* = - Jv Dxk 5 Cxk C xk dV - Jv Dyk 5 C / C j dV - J Vxk 8 Ck CIxk dV - J Vy 8 Ck Cjxk dV - J Rk 8 Ck C* dV + Dix JA 8 C* C xk dA + D y J 8 Ck C yk dA (B.l7) It is noted that 5,Ik would not equal to zero at the element level. Assuming that Dx, Dy, Vx, Vy and R are constants for the specific element and they are not a function of time and when equations B.8 and B.9 are substituted in B.17 S,!k = {SC)T [- D,k [- D.k Jv [N J T [N J dV - Dyk Jv [Ny]T [Ny] dV - V,k Jv [N]T [N J dV - V,k Jv [N]T [N.y] dV] (C) + {8C)T [- R Jv [N f [N]] (C) + {8C]T [JA [N] C, dA + JA[N ]C y dA] (B.18) Note that C x and Cy values in the last two terms of equation (B.18) are the flux values supplied on the boundaries of the element. The time derivative {C} is evaluated by using a finite difference scheme, {£} - ~ {C}> (B.19) At The volume and area integrals in equation (B.18) are calculated numerically by Gauss-Legendre polynomial integration routine. The four nodes of the quadrilateral results in a 4 x 4 matrix in all integrals where the shape function matrix appears twice while a vector is formed in the ones with single shape functions. The resulting formulation can be represented by, 6, /* = {5C1r \{SKF]* { C } ^ , + I B 1 {CJ*A, - (C l? + { O H (B.20) A t A t where [SKF] matrix represents all terms at t+At and [E] matrix constitutes the time derivative. B.4 GLOBAL MATRIX The connectivity of each element is used to generate the global stiffness matrix as 8,I8 = {8C]T {At [SKF]8~{C]l+A8 + [EJ« {C}l+Atg - [E]g {C]t8 +At {Q}} = 0 (B.21) It is noted that when the global stiffness matrix is formed and the contributions from each element is accounted for, the first variation of the integral equation 5,1s must equal zero to minimize the energy input into the system. Since the variations {8C} are arbitrary, equation B.21 renders, At [SKF]8 {C]l+Al8 + [E]8 [C]t+Al8 = [E]8 {C}8 = At {Q}8 (B.22) when further condensed will give, [Al]8 {C},+Alg = [E]8 {€}, - {RQ}8 (B.23) B.5 SOLUTION IN TIME The matrix equations given in B.23 are solved by using the available initial and boundary conditions. At any time step, both [Al]8 and [E]8 matrix are multiplied by the constant boundary conditions and the elements corresponding to the unknown values are transformed to the right hand side to form a new known vector due to the boundary conditions, {RB}8. [Al]8 [C]1+At8 = [E]8 [C], - {RQ}8 - {R}8 (B.24) Next, the initial conditions [C]l=0 are multiplied by the [E]8 matrix to form a third known vector {RI}8, [Al]8 [C]l+A, = [RI]8 - [RQ]8 - [RB]8 = [R]8 (B.25) The stiffness matrix [Al]8 is then condensed and inverted. (Note: in coding, this operation is actually conducted at the element stiffness matrix level and only the matrix elements corresponding to unknown concentrations are left. This operation is done to minimize the required space to hold the Al matrix.) ( C U = [Al]1 [R]8 (B.26) The operation in B.26 would give the concentrations at the first time step. The concentrations in the second time step are evaluated as [C]l=2Al = [Al]'1 {[E] [ C U - [RQ] - [RB]} (B.27) Therefore, the concentrations at a previous time step is used as initial values, while the constant flux and concentration boundary conditions do not change the values of [RQ] and [RB]. It is noted that once the [Al]'1 matrix is inverted in the first time step, this operation is no further needed for subsequent time steps. Choleski decomposition method is used in this inversion in coding of the above scheme. It should be noted that since the first derivatives of concentrations appear in the velocity terms of the integral equation, the resulting matrix would be nonsymmetrical and hence an appropriate nonsymmetric matrix solution routine needs to be utilized. The stability and accuracy considerations for one-dimensional problems are provided by Acar, et al. (1989). APPENDIX C TWO-DIMENSIONAL FINITE ELEMENT METHOD CODE FOR THE ANALYTICAL MODEL 218 A 219 APPENDIX C TWO-DIMENSIONAL FINITE ELEMENT METHOD CODE FOR THE ANALYTICAL MODEL C.l INTRODUCTION The differential equation solved is, Dx C xx + C y Cyy - Vx C, - Vy C,y = R C (C.l) Two types of boundary conditions are handled by the code: (1) Neumann B.C. (C .2a) dx = B (C.2b) ay (2) Drichlett B.C. C = y (C3) It should be noted that in solution either the first or the second type of boundary condition could be specified in solving a second degree parabolic partial differential equation. The computer program is written in FORTRAN 77 and for an IBM compatible microcomputer. It is recommended to have 560K RAM memory and if large problems are solved, the availability of a fixed disk. This is mainly due to the fact that the element stiffness matrices are stored in the disk and sufficient space is needed to perform this operation for a large scale problem. It is further noted that in the current version (88.01), the master stiffness matrix is inverted in the RAM memory and as the problem size increases beyond the capacity of the RAM memory erroneous results may be encountered. The size of the compiled executable file is approximately 360 Kb which leaves about 200 Kb of space for the solution. Assuming that each storage takes about 4 bytes of memory, and considering that a certain portion of this memory is taken by the operating system, approximately 40,(XX) spaces are available for the solution. In a square discretization, this corresponds to about 30 elements each in x and y direction. C.2 FLOW CHART FOR THE PROGRAM AND THE SUBROUTINES The flow chart for the main program is given in Figure C.l. It is noted that the data input could be furnished either from the screen or from a file. The screen input is stored into a file in the disk using which parametric studies could be done in further runs. It is recommended that the data is first input from the screen. This procedure avoids the labor of entering the data with a specific format, since it is readily stored in a specified file. The total dimension of the one dimensional array where the problem is solved is currently set at 40,000 by the two statements in the main program: D(40000) MTOT = 40000 221 CALL INPUT ALL ELEMENT ELSTIF CALL FORMBCALL CALL GENER CALL SOLVES CALL MODIFY CALL CHOLES END Figure C.l Row Chart of the Main Program In case the problem size requires a larger dimension, these two statements should be changed. The flow chart for SUBROUTINE INPUT is presented in Figure C.2. This routine stores the nodal point data, generates them if required by the user, gives equation numbers to each node according to the specified boundary conditions, and receives the initial concentrations at each node. The SUBROUTINE ELEMENT calls another one, ELSTIF. The flow chart for ELSTIF is presented in Figure C.3. This routine receives each constant for the number of materials used in the problem. The element data, connectivity is provided. It is noted that a counterclockwise notation is used in the formulation and hence node numbers should be provided in a counterclockwise manner. The flux boundary conditions and the specified fluxes are input in this routine. The routine, afier receiving all this data, performs a Gauss-Legendre polynomial integration scheme to calculate the volume integrals. A four point scheme is used for plane problems and a 9-point scheme is used in an axisymmetric problem. The need to use a higher order scheme in axisymmetric problem arises since the order of the function to be integrated increases by one in the axisymmetric problems. After conducting the integration by calling another subroutine, FORMB, the routine finally stores the [SKF], [E] and [RB] matrices in a file on disk and calls this file XXX.STF. This procedure is necessary if the concentration gradients are required throughout the domain in any problem (e.g., in dissipation of pore SUBROUTINE INPUT NODAL POINT DATA x,y COORDINATES AND CONSTANT BOUNDARY CONDITIONS GIVE EQUATION NUMBERS TO EACH NODE GENERATE NODAL POINT DATA ACCORDING TO SPECIFIED BOUNDARY CONDITIONS (IF REQUIRED) INPUT INITIAL CONDITIONS AT EACH NODE RETURN Figure C.l Flow Chart for Subroutine INPUT. SUBROUTINE ELSTIF READ DISPERSION COEFFICIENT, VELOCITY, RETARDATION CONSTANT FOR THE NUMBER OF MATERIALS, READ PROBLEM TYPE INPUT ELEMENT DATA CONNECTIVITY AND ANY FLUX BOUNDARY CONDITIONS ON THE ELEMENT GENERATE ELEMENT DATA IF REQUIRED GENERATE LOCAL STIFFNESS MATRIX AND FLUX VECTORS BY CALLING FORMB AND USING GAUSS-LEGENDRE INTEGRATION SCHEME STORE SKF, E, AND RD IN FILE 3 FOR EACH ELEMENT CALCULATE BANDWIDTH RETURN Figure C.3 Flow Chart for Subroutine ELSTIF. 225 pressure in a consolidation problem it may be necessary to find the gradients to calculate the flow). Subroutine FORMB calculates the values of the integrals in equation (B.20) (except the constant vectors) at the specified Gauss-Legendre polynomial integration points and returns the values to ELSTIF for incremental addition routine. The Jacobian and the differentiation described in equations (B.9) through (B.16) are conducted discretely in this subroutine. Subroutine GENER (Figure C.4) generates the master stiffness matrices [Al] and [E] and calculates {RB} matrix due to boundary condition, and adds the {RQ} matrix due to flux condition to the above two to generate the constant matrix which will not change all throughout the solution. These vector values are then stored to the bottom of both [Al] and [E] matrices. The solution is accomplished in SUBROUTINE SOLVES. The flow chart for this subroutine is given in Figure C.5. This routine first calculates the {RI} matrix due to the initial conditions and adds to the constant vector calculated due to the boundary conditions. The constant vector is stored below the [Al] matrix in a one dimensional array. The Choleski decomposition inversion routine, SUBROUTINE CHOLES is then called to invert the [Al] matrix and also get the solution by multi plying the inverted matrix with the constant vector at the bottom of the [Al] matrix. The solution is then placed to the bottom of the same matrix. The results at the first time step are printed. Subsequently, depending upon the user requested number of SUBROUTINE GENER READ [SKF] FOR EACH ELEMENT [E], {RB} MATRICES AND CORRESPONDING NODAL POINT NUMBER FROM FILE 3 FROM MASTER STIFFNESS MATRICES Al AND E IN A ONE-DIMENSIONAL ARRAY AND THE ELEMENTS WITHIN THE BANDWIDTH MULTIPLY THE BOUNDARY CONDITION WITH Al AND {E} AND FORM {RS} MATRIX ADD {RQ} + {RB} AND PLACE TO THE BOTTOM OF [Al] AND [E] MATRICES RETURN Figure C.4 Flow Chart for Subroutine GENER. SUBROUTINE CALL MODIFY WITH [E] FOR EACH INITIAL CONDITION (TO INCLUDE {RI} VECTOR) CALL CHOLES WITH CONDITION 1 AND INVERT THE Al MATRIX GET THE RESULTS IN THE FIRST STEP PRINT RESULTS, STORE RESULTS IN AN ARRAY CHECK NO. STOP OF CYCLES OF ITERATION CALL MODIFY WITH [E] AND PERVIOUS STEP’S INITIAL CONDITIONS CALL CHOLES WITH CONDITIONS TO MULTIPLY THE INVERTED MATRIX WITH THE NEW CONSTANT VECTOR Figure C.5 Flow Chart for Subroutine SOLVES. 228 cycles of iteration in time, first MODIFY is called to get the new {RI} matrix using the one previous solution as initial condition and CHOLES is used to get the solution. It is noted that in these call requests for CHOLES, it is no more necessary to invert the matrix since the inversion has been accomplished in the first time step. The solution is then read from the bottom of [Al] and is printed at output intervals requested by the user. C.3 SPECIAL CONDITIONS It should be noted that if the program is modified to include changes in velocity in time (nonlinearity in coefficients), the master stiffness matrix [Al] will be different in time. Therefore, it will be necessary to invert this matrix at each time step. Furthermore, the master stiffness matrix has to be regenerated at each time step and the boundary condition "vectors has to be recalculated. APPENDIX D Reprint Permission from Hamed, et al. 1991 229 230 Dt'iwtnu'nlof C:vil l u y ihvrmy Louisiana State Univer S1TV WPACRM.ITI.KA1 ASP MK HAM, At vOLIH.I BATON ROUGE • LOUISIANA . 7HK0?. September 5, 1990 Ms. All i p, Marabella ASCE 345 East 47th Street New York, NY 10017 Re: P b(II) Removal by E lectrokinetics Hamed/Acar/Gale rS Dear Ms. Marabello: The above following paper submitted by us is accepted for publication."Ve have been informed that the paper will appear in the February 1991 issue. The work in this paper constitutes a part of my dissertation thesis which I w ill defend in December 1990. I would lik e to use some of the fig u res, tables and analysis in my dissertation. I would like to request your per mission to include this material in the manuscript. Sincerely,1 Jihad Hamed Graduate Research Assistant e n g in e : 't i l STREET N.Y. 10017 VITA Jihad Tawfiq Abed El Hadi Hamed was bom in Beit Hanon. Gaza Strip, on February 23, 1957, the son of Tawfiq Abed El Hadi Hamed. After graduating from Beit Hanon High School, Beit Hanon, Gaza Strip, he entered the University of Technology at Baghdad, Iraq. He received the degree of Bachelor of Science in Civil Engineering from the University of Technology at Baghdad. Iraq in June. 1982. In the Fall of 1984, he entered the Graduate School of the University of Texas at EL Paso. He received the degree of Master of Science in Civil Engineering in Spring 1984. In Spring 1987, he entered the Graduate School at Louisiana State University. Permanent address: 34/6 Beit Hanon Gaza Strip 231 DOCTORAL EXAMINATION AND DISSERTATION REPORT Candidate: Jihad Tawfiq Hamed Major Field: C iv il E ngin eering Title of Dissertation: Decontamination of Soil using Electro-Osmosis Approved: Major Professor and Chairman Dean of the Graduate School EXAMINING COMMITTEE: Pi a, n a a 1' / I Date of Examination: 11-21-1990