Effect of clay mineralogy on the feasibility of electrokinetic soil

Effect of clay mineralogy on the feasibility of electrokinetic soil

Effect of clay mineralogy on the feasibility of electrokinetic soil

decontamination technology

Darmawana, S.-I Wadab,*

a

Department of Bioresources and Environmental Sciences, Graduate School of Kyushu University, Fukuoka 812-8581, Japan

b

Laboratory of Soils, Faculty of Agriculture, Kyushu University, Fukuoka 812-8581, Japan Received 18 January 2001; received in revised form 21 June 2001; accepted 3 July 2001

Abstract

To evaluate the effect of clay mineralogy on the feasibility of electrokinetic soil remediation technology, we contaminated six soils with Cu(II), Zn(II) and Pb(II) and performed electroremediation for 570 h Cation exchange resin saturated with H+ was placed between soil and cathode to prevent soil alkalinization and trap the migrated heavy metal cations After the treatment, the heavy metal cations were sequentially extracted with water, 1 M MgCl2and hot 6 M HCl In soils dominated by crystalline clay minerals, Cu(II) and Zn(II) significantly migrated from anode end and accumulated at the cathode end forming sparingly soluble hydroxides Removal rates of Cu(II) and Zn(II) were highest in a soil dominated with kaolinite and crystalline hematite In humic – allophanic and allophanic soils, the high pH-buffering capacity of allophane kept the soil pH above 5, even

at the anode end, and Cu(II) and Zn(II) did not migrate significantly In all soils, the migration of Pb(II) was infinitesimal due to the formation of insoluble PbSO4and very strong surface complexation at the mineral surfaces These results show that the reactivity of component clay minerals to H+ and heavy metal cations has a crucial effect on the efficiency of the electrokinetic remediation technology and it is not effective for remediation of allophanic soils The results also indicate that allophanic soils may be useful as a barrier material in landfill sites.D 2002 Elsevier Science B.V All rights reserved

Keywords: Clay mineralogy; Electrokinetic remediation; Heavy metal; Soil pollution

1 Introduction

The direct disposal, dumping and tailing of wastes

containing toxic heavy metals often result in

accu-mulation of the heavy metals in soils because they

are strongly retained by some soil components such

as clay minerals and humic substances The heavy

metals retained in soils are, however, gradually released into pore water, resulting in pollution of surface and ground waters In addition, fine-grained soil particles retaining heavy metals are transported

as airborne dust and contaminate agricultural land and crops Thus, the heavy metal-contamination of soils occurring in industrial areas is a great concern due to its direct and indirect harmful effects on human health

Among many technologies developed to decon-taminate heavy metal-polluted soils, the electrokinetic method is regarded as an effective technique partic-ularly for soils having low hydraulic conductivity 0169-1317/02/$ - see front matter D 2002 Elsevier Science B.V All rights reserved.

PII: S 0 1 6 9 - 1 3 1 7 ( 0 1 ) 0 0 0 8 0 - 1

* Corresponding author Tel.: 642-2844; fax:

+81-92-642-2864.

E-mail address: wadasi@agr.kyushu-u.ac.jp (S.-I Wada).

www.elsevier.com/locate/clay

(Acar and Alshawabkeh, 1993) In electrokinetic

remediation, a direct current is passed through a soil

to induce the electromigration and electroosmosis

towards the electrode where the contaminants are

collected According to an analysis by Acar and

Alshawabkeh (1993), electromigration is more

im-portant at least for removal of ionic contaminants

For successful decontamination of heavy metal ions,

therefore, it is important to convert the precipitated

and adsorbed ions into dissolved forms

Heavy metals incorporated into soils take

differ-ent chemical forms including (1) dissolved ionic

form, (2) electrostatically adsorbed form, and (3)

surface complexed form (Darmawan and Wada,

1999) Layer silicate minerals having excess surface

negative charge arising from isomorphous

substitu-tion adsorb heavy metal casubstitu-tions electrostatically Due

to the nonspecific nature of the coulombic force, the

adsorbed heavy metal cations are easily exchanged

by other cations The selectivity coefficient, for

example, for Cu – Na exchange on a montmorillonite

is near unity (Sposito et al., 1981) On the other hand,

oxides and hydroxides of iron and aluminum as well

as humic substances bind heavy metal cations very

strongly, forming surface complexes, in which the

bonding between heavy metal cation and surface

functional groups bears some covalency The surface

complexed heavy metal cations are, therefore, not

exchanged by common ions in the pore water of

soils

Thus, it is expected that the feasibility of the

electro-kinetic remediation method depends strongly on

min-eralogical composition as well as on soil organic matter content Studies by Puppala et al (1997), Reddy et al (1997), and Grundl and Reese (1997) suggested that the electrokinetic method is not necessarily effective for soils with high adsorption capacity and also for those containing calcium carbonate To date, however, many researchers have used only artificially polluted model soils made up of pure clays such as kaolinite (Acar and Alshawabkeh, 1996; Yeung et al., 1996; Dzenitis, 1997) and smectite (Reddy et al., 1997; Grundl and Reese, 1997) or their mixture with quartz (Kawachi and Kubo, 1999) in developing and evaluating the electro-kinetic remediation technology There seems to be no systematic experimental study in which the electro-kinetic remediation method is applied to a series of natural soils that differ in mineralogical composition In the present study, therefore, six soils that differ in clay mineralogy and organic matter content were contami-nated with salts of copper, lead, and zinc, subjected to electrokinetic remediation treatment, and the behaviors

of the heavy metal cations were analyzed through successive extraction

2 Materials Six soil samples from different locations were used in the present study They were air-dried and passed through a 2-mm sieve Some chemical and mineralogical properties determined by standard pro-cedures (Page et al., 1982; Klute, 1986) are listed in Table 1

Table 1

Selected properties of soil samples

carbon (g kg
1)

Clay (g kg
1)

Allophane and/

or imogolite (g kg
1)

DCB a Fe 2 O 3

(g kg
1)

Major cation exchanger

kaolinite

kaolinite

a

Dithionite – citrate – bicarbonate.

b

No measurement.

As shown in Table 1, the Nakajo, Fukuchiyama,

Harumachi, and Pakchong soils are predominated by

crystalline clay minerals Among these four soil

samples, effective CEC of the Pakchong soil is the

lowest irrespective of its extremely high clay content,

reflecting the predominance of kaolinite in its clay

fraction The Goshi and Choyo soils are similar in

that the clay consists exclusively of allophane and

imogolite but the Goshi soil has much higher organic

matter content Table 1 also shows that the amounts

of Fe and Mn oxides dissolved by dithionite –

cit-rate – bicarbonate (DCB) treatment (Mehra and

Jack-son, 1960) All the soil samples have fairly high

amounts of DCB soluble iron oxide with a maximum

amounts of DCB soluble Mn oxides are much lower

Electron microscopy (photos not presented) showed

that the iron oxide in the Pakchong soil exists as fine

pseudo-hexagonal hematite particles, whereas no

dis-crete iron oxide particles was detectable in other

soils

mixing the soil samples with the calculated amounts

of CuSO45H2O, ZnSO47H2O and Pb(NO3)2 and

aged for 1 year at field moisture contents These three

metal cations were employed because they are

expected to behave differently in soils (Darmawan

and Wada, 1999), i.e., Cu(II) and Pb(II) are highly

specifically adsorbed by oxides and humic substances

but Pb(II) tends to form sparingly soluble sulfates, and

Zn(II) is preferred rather by layer silicate minerals

The detailed procedure of the sample preparation is given in Darmawan and Wada (1999) Since Cu(II) and Zn(II) were added as soluble sulfate, sulfate ion from these salts would have reacted with Pb(II) ion

during aging Although this makes the system com-plicated, the present combination of salts was selected because PbSO4is one of the major forms of Pb(II) in heavily polluted soils

3 Methods The apparatus was made from an open plastic box with dimensions of 220 55  55 mm by partitioning

it into four compartments with nylon cloth and attach-ing an inlet for anolyte, as shown in Fig 1 The polluted soil samples were packed into the largest central compartment and a pair of graphite electrodes was placed in the two compartments at the two ends The weight of the packed soil was 320 g for the Nakajo soil, 355 g for the Fukuchiyama soil, 350 g for the Harumachi soil, 335 g for the Pakchong soil, 317

g for the Goshi soil, and 201 g for the Choyo soil The compartment between the soil and electrode was packed with H-saturated cation exchange resin

at the cathode and to trap heavy metal cations (Wada and Ryu, 1999)

The inlet of the apparatus, packed with the soil and resin, was connected to a constant head device containing 10 mM NaCl solution to saturate the

Fig 1 Schematic illustration of the experimental cell.

interstitial pores of the soil After saturation, the

level of the anolyte was set 5 mm below the soil

surface (Wada et al., 1999) and 20-V potential was

applied with a stabilized power supply Throughout

the treatment, the electric current was monitored

with an ammeter inserted in the circuit (Fig 1)

After 570 h, the apparatus was disconnected from

the power supply and the soil column was cut into

six slices of equal length, and air-dried The soil pH

was measured at a soil/water ratio = 1:2.5 Portions

(2 g) of the air-dried samples were placed in

poly-carbonate centrifuge tubes and extracted sequentially

for 1 h at room temperature HCl extraction was

performed by digesting with 10 ml of 30% H2O2to

dryness followed by refluxing with 16 ml of 6 M

HCl (Asami and Kato, 1977) at boiling point The

extracted Cu(II), Pb(II) and Zn(II) were determined

by atomic absorption spectroscopy (AAS) following

the procedures described by Amacher (1996) and

Reed and Martens (1996)

The resin was collected quantitatively and packed

in a column The adsorbed metal cations were

ex-tracted by leaching with 1 M NH4NO3and determined

by AAS

All the measurements were carried out in duplicate

and the results were averaged

4 Results and discussion

The measured electric currents are plotted against

elapsed time in Fig 2 In general, the electric current

decreased rapidly in the first 50-h period followed by

gradual decrease in the subsequent 150 h for all

soils The mode of variation and the magnitude of

the current were, however, significantly different

among soils For the Nakajo, Fukuchiyama and

Harumachi soils, the initial electric current was about

13 – 19 mA and it monotonically decreased to the

final range of 1.6 – 2.7 mA The trend was quite

similar for the Goshi and Choyo soils but both the

initial and final values were much lower, i.e., 5.5 –

6.2 mA and 0.2 – 0.5 mA, respectively For the

Pakchong soil, on the other hand, the electric current

increased up to 28 mA after 7 h and rapidly dropped down to < 1 mA after 200 h

The higher initial values of the electric current are obviously due to the higher electrolyte concentration

in the pore water Since indigenous soluble salts are negligible in all soils and the amount of added heavy metal salts were the same, the cause of the lower initial current for humic – allophanic Goshi and allo-phanic Choyo soils would be that the cations and anions of the added salts were adsorbed on the soil materials to a larger extent (Wada, 1984) The cause of the rapid increase and immediate decrease is not clear but the higher initial electric current in the Pakchong soil supports the contention that the initial contents of water soluble Cu(II), Pb(II) and Zn(II) were the high-est in the Pakchong soil (Figs 3 – 5)

The pH profiles along the soil column after the treatment as well as the initial pH values are pre-sented in Fig 6 Since H+ ion was generated by the electrolysis reaction of water continuously at the anode and it migrated toward the cathode, soil pH significantly dropped, particularly in the section neighboring the anode In the Nakajo, Fukuchiyama, Harumachi and Pakchong soils predominated by crystalline clay minerals, the pH of the soil section next to the anode was below 3 Among these four soils, the Pakchong soil was conspicuous and the pH was far below 3 in the first through fifth sections In the Nakajo, Fukuchiyama, and Harumachi soils, on the other hand, the magnitude of the pH drop de-creased toward the cathode and the pH value of the soil section neighboring the cathode was quite near

Fig 2 Time course of electric current.

the initial one The magnitude of pH drops was much

smaller in the humic – allophanic Goshi and

allo-phanic Choyo soils and the pH was mostly kept

above 5 except in the section neighboring the anode

The plots in Fig 2 were interpolated by a cubic

spline function and integrated over the treatment time

to calculate the amount of the electric charge

trans-ported through the soil samples The calculated

amount of charge was 7.95 kC for the Nakajo, 5.72

kC for Fukuchiyama, 4.90 kC for Harumachi, 4.76 kC

for Pakchong, 1.98 kC for Goshi, and 0.649 kC for

Choyo soils If it is assumed that the electrolytic

generation of H+ ion is proportional to the electric

current, the H+ ion supplies in the Goshi and Choyo

soils were estimated to be much smaller than that in

other soils However, the pH profiles for the Goshi

and Choyo soils are quite similar irrespective of the

difference in the amount of the transported charge,

and the differences in H+ ion concentration estimated

from the pH values are much smaller than those

expected from the differences in the charge transport

(Fig 2) These data suggest that the lower electric currents in the Goshi and Choyo soils are not a major cause for their higher pH values

Alternatively and probably the most important cause for that would be the higher acid-buffering capacity of the soils Allophane and imogolite present in these soils contain large amounts of surface aluminol groups (Wada, 1984) On addition of an acid, HA, the alumi-nol groups uptake H+ ion and the resulting positively charged sites retain the accompanying anion:

With this reaction, the H+ ion generated at the anode would have largely been consumed, and in turn, the pH of the interstitial water dropped to a lesser extent It seems contradictory that the Pakchong soil containing the largest amount of free iron oxide, which can uptake H+ ions and anions, exhibited the lowest pH However, iron oxide in the Pakchong soil was mostly crystalline hematite, whereas it was non-Fig 3 Distribution of three forms of Cu(II) in soils after treatment.

crystalline in other soils Higher crystallinity of free

iron oxide and the predominance of kaolinite with low

cation adsorbing capacity would have favored the

acidification in the Pakchong soil

The contents of three forms of Cu(II), Pb(II), and

Zn(II) in soil slices after electrokinetic treatment are

presented in Figs 3 – 5, together with their initial

concentrations Comparison of these figures indicates

that the distribution pattern of Cu(II) and Zn(II) are

similar for all soils while that of Pb(II) differed

significantly The distribution of Pb(II) was distinctly

different in that the removal rate was significantly

lower and the proportion of the MgCl2-extractable

fraction was comparable to or even higher than the

HCl-extractable fraction (Fig 4)

In the present extraction procedure,

water-extract-able Cu(II), Pb(II) and Zn(II) would come mostly

-extractable fraction would basically represent those

retained on layer silicate clay minerals as

exchange-able cations HCl-extractexchange-able fraction would

corre-spond to those bound as surface complexes on iron oxides, allophane and humic substances (Darmawan and Wada, 1999) as well as hydroxide precipitates

In addition, for Pb, the MgCl2-extractable fraction would include Pb(II) precipitated as PbSO4 (angle-site) in addition to exchangeable Pb(II), due to the increased solubility of PbSO4 in a concentrated Mg

sulfate ion is reduced by the formation of soluble

Pitz-er’s model (Gueddari et al., 1983) showed that the ion activity product of (Pb2 +)(SO4

) would never

solubility product of PbSO4is 1.62 10
7(Lindsay, 1979), the precipitate of PbSO4 is expected to have dissolved during extraction and the MgCl2 -extract-able fraction can be regarded as a sum of exchange-able Pb and PbSO4

Before treatment, copper was found mostly in the strong acid-extractable fraction, except in the Pak-chong soil After the treatment, Cu(II) content re-Fig 4 Distribution of three forms of Pb(II) in soils after treatment.

duced markedly in soil sections near the anode to

less than 100 mg kg
1 and accumulated in sections

near the cathode in the Nakajo, Fukuchiyama, and

Harumachi soils Comparison of Figs 3 and 6

sug-gests that the amount of Cu(II) left in the soil is

inversely correlated with pH The stability of surface complexes of Cu(II) on oxide and humic substances decreases steeply below pH 4 (McKenzie, 1980; Kendorf and Schnitzer, 1980) The distribution pat-tern of Cu(II) shown in Fig 3 suggests that Cu(II) was released from surface functional groups, mig-rated toward the cathode and hydrolyzed to precip-itate when it entered into a region of higher pH, as repeatedly reported by many researchers (Yeung et al., 1996; Puppala et al., 1997; Viadero et al., 1998)

A noticeable feature in Fig 3 is that significant

Cu(II) appeared in the fourth to sixth sections from the anode in the Nakajo, Fukuchiyama, and Haru-machi soils Darmawan and Wada (1999) showed that heavy metal cations loaded to soils are rapidly adsorbed at permanent negative charge on layer sili-cate minerals, and then slowly transferred to oxides and humic substances to form surface complexes in

50 days This and the results presented in Fig 3 suggest that the role of cation exchange sites on Fig 5 Distribution of three forms of Zn(II) in soils after treatment.

Fig 6 Profile of soil pH after treatment.

layer silicate minerals becomes more important in a

nonequilibrium migration process, particularly at low

pH where the stability of surface complexes is low

Thus, longer treatment time and more electrical

power are required for remediation of soils with high

cation exchange capacity (Puppala et al., 1997)

Fortunately, the cation exchange sites of layer silicate

minerals do not show high preference for heavy

metal cations (Sposito et al., 1981) The efficiency

of electrokinetic remediation may be improved by

adding salts of noncontaminant cations, e.g., CaCl2,

that compete with heavy metal cations for the

ex-change site

In the Pakchong soil, the Cu(II) removal rate was

the highest among the soils Total content of the three

forms of Cu(II) was below 100 mg kg
1in the first

and second sections and it was about 200 mg kg
1in

the fifth section, i.e., the removal rate was > 80% The

relatively high removal rate would have resulted from

the relatively low pH of the original soil and low

cation exchange capacity (Table 1) and higher

crys-tallinity, and therefore, lower reactivity of iron oxide

minerals

In contrast, Cu(II) was not removed at all from

the Goshi soil (Fig 3) It is apparent that a major

cause of this is the high acid-buffering capacity of

the soil (Fig 6) In addition, the extremely high

selectivity of humic substances for Cu(II) would

have contributed the result A significant reduction

of Cu(II) content in the section neighboring the

anode in the allophanic Choyo soil that showed

the same pH value to the corresponding section of

the humic – allophanic Goshi soil (Fig 6) supports

this view

Fig 4 shows that the behavior of Zn(II) is basically

similar to that of Cu(II) but Zn(II) is much more

mobile under the applied electric field

The behavior of Pb(II) shown in Fig 5 is distinctly

different from that of Cu(II) and Zn(II) First, the

proportion of MgCl2-extractable fraction is apparently

larger in soils dominated by layer silicate minerals

before treatment This must be due to the presence of

PbSO4and its dissolution in 1 M MgCl2solution as

already discussed Second, the removal rate was

significantly lower than that for Cu(II) and Zn(II)

and half or more of the remaining Pb(II) was in an

MgCl2-extractable form in all soils, except for

hu-mic – allophanic and allophanic soils

Since PbSO4 precipitated in soils does not move either via electrophoresis or via electroosmosis, lower removal rates were expected in soils in which a large proportion of Pb(II) was in a form of PbSO4 Actually the results obtained for the Nakajo, Fukuchiyama, Harumachi and Pakchong soils (Fig 5) follow the expected trend In these soils the HCl-extractable fraction, i.e., surface complexed Pb(II), reduced sig-nificantly in sections near the anode and seems to have accumulated in the following sections

In contrast, most of the Pb(II) was found in HCl-extractable fraction and the proportion of MgCl2 -extractable fractions were low in the Goshi and Choyo soils both before and after the treatment (Fig 5) This indicates that the precipitation of PbSO4 was greatly suppressed in these soils

and Pb2 + adsorption As already discussed, Pb2 + forms stable surface complexes with surface hy-droxyl groups on minerals and carboxyl groups on humic substance The reaction can be expressed, for example,

where SOH stands for surface hydroxyl groups and carboxyl groups In acidic condition, some surface hydroxyl groups, typically AlOH and FeOH groups, are protonated and adsorb anions Among anions, oxo-anions are adsorbed more strongly The reaction can be expressed as:

Since allophanic soils have large amount of the both types of functional groups, the overall reaction is:

With the simultaneous adsorption of Pb2 + and SO4
, the activity of these ions in solution would be main-tained low enough to prevent formation of PbSO4(s)

In the present study, H-saturated cation exchange resin was placed between cathode and soil to prevent alkalinization of soil and to trap the migrated heavy metal cations Wada and Ryu (1999) have examined

this in Cu(II) removal from an artificially

contami-nated soil and found that the alkalinization of the soil

neighboring the anode was successfully suppressed

and the whole soil was acidified The pH profile in

Fig 6, however, shows that the alkalinization was

prevented but the acid front did not reach the cathodic

end of the soil Therefore, the migrated heavy metal

cations hydrolyzed and accumulated in the soil section

nearest to the cathode The reason for the difference

between the present result and those by Wada and Ryu

(1999) is not clear One possible reason is that there

was no drainage of catholyte in the present study

The mass balances of Cu(II), Zn(II) and Pb(II) are

listed in Table 2 The percentages of the total detected

heavy metals were mostly >90 except for Zn in the

Fukuchiyama soil Since there was no noticeable

interference in the AAS determination, the missing

portions of the heavy metals would have associated

with the nylon cloth used for separating the resin and

soils Although there are some uncertainties, Table 2

clearly shows that the percentages of Cu(II) and Pb(II)

trapped in the cation exchange resin were generally

low for all the soil samples than those of Zn(II) Since

the objective of the present study was to examine the

effect of clay mineralogy on the efficiency of

elec-trokinetic remediation technology, any additional enhancement was not performed With some enhance-ment techniques including addition of salts of non-contaminant cations, soil acidification, and drainage

of catholyte, the removal rate would be improved, at least, for nonallophanic soils

The results for the Goshi and Choyo soils sug-gest that the electrokinetic technology would not work for humic – allophanic and allophanic soils, at least for Cu(II) removal Nevertheless, this may not

be a negative finding The extremely low mobility

of heavy metal cations in these soils even under electrokinetic treatment strongly indicates that heavy metal cations incorporated in these types of soils are expected not to be leached into the ground water under the conditions normally encountered in nature This indicates the possible use of these types of soils for low-cost permeable barrier at waste landfill site

5 Conclusions

. Cu(II) and Zn(II) electrically migrated from the anode toward the cathode to a significant extent in

Table 2

Mass balance of Cu, Pb and Zn

Heavy metal Soil name Initial (g) Remained in soil Adsorbed in resin Total detected

soils dominated by crystalline clay minerals But a

lesser drop in pH of soil section near the cathode has

led to the precipitation of Cu(II) and Zn(II) into

hydroxides Among these soils, the electrokinetic

removal of Cu(II) and Zn(II) in the kaolinitic

Pak-chong was the highest due to low initial pH, low

cation exchange capacity, and lower reactivity of iron

oxide minerals of this soil

soils dominated by crystalline clay minerals would

be increased by some enhancement techniques,

including addition of salts of competing

nonconta-minant cations, soil acidification, and catholyte

drainage

allophanic and allophanic soils are the major factors

that have hindered the electromigration of Cu(II) and

Zn(II) in these soils

. Precipitation as sparingly soluble PbSO4 and

stable surface complexation with hydroxyl groups of

minerals and carboxyl groups of humic substances

have resulted in extremely low migration of Pb(II)

from all soils under study

humic – allophanic and allophanic soils and their

extremely low mobility even under electrical field

indicate the possible use of these soils as an ideal

permeable barrier at landfill site

Acknowledgements

We thank Dr Mochizuki of Experimental Farm of

Kyushu University for providing a soil sample This

study was supported in part by a Grant-in-Aid for

Scientific Research (#11660066) from the Japanese

Society for Promotion of Sciences

References

Acar, Y.B., Alshawabkeh, A.N., 1993 Principles of electrokinetic

remediation Environ Sci Technol 27, 2638 – 2647.

Acar, Y.B., Alshawabkeh, A.N., 1996 Electrokinetic remediation: I.

Pilot-scale tests with lead-spiked kaolinite ASCE J Geotech.

Eng 122, 173 – 185.

Amacher, M.C., 1996 Nickel, cadmium and lead In: Sparks, D.L.

(Ed.), Methods of soil analysis: Part 3 Chemical methods Soil

Science Society of America, Madison, WI, pp 739 – 768.

Asami, T., Kato, K., 1977 Comparison of analytical methods for total cadmium, zinc, lead and copper in soils Jpn J Soil Sci Plant Nutr 48, 335 – 336 (in Japanese).

Darmawan, Wada, S.-I., 1999 Kinetics of speciation of Cu, Pb and

Zn loaded to soils that differ in cation exchanger composition

at low moisture content Commun Soil Sci Plant Anal 30,

2363 – 2375.

Dzenitis, J.M., 1997 Soil chemistry effects and flow prediction in electroremediation of soil Environ Sci Technol 31, 1191 – 1197.

Grundl, T., Reese, C., 1997 Laboratory study of electrokinetic ef-fects in complex natural sediments J Hazard Mater 55, 187 – 201.

Gueddari, M., Monnin, C., Perret, D., Fritz, B., Tardy, Y., 1983 Geochemistry of brines of the Chot El Jerid in southern Tune-sia—application of Pitzer’s equations Chem Geol 39, 165 – 178.

Kawachi, T., Kubo, H., 1999 Model experimental study on the migration behavior of heavy metals in electrokinetic remedia-tion process for contaminated soil Soil Sci Plant Nutr 45,

259 – 268.

Kendorf, H., Schnitzer, M., 1980 Sorption of metals on humic acid Geochim Cosmochim Acta 44, 1701 – 1708.

Klute, A., 1986 Methods of soil analysis: Pt 2 Physical and min-eralogical methods, 2nd edn Soil Science Society of America, Madison, Wisconsin.

Lindsay, W.L., 1979 Chemical equilibria in soils Wiley, New York.

McKenzie, R.M., 1980 The adsorption of lead and other heavy metals on oxides of manganese and iron Aust J Soil Res.

18, 61 – 73.

Mehra, O., Jackson, M.L., 1960 Iron oxide removal from soils and clays by a dithionite – citrate system with sodium bicarbonate buffer Clays Clay Miner 7, 313 – 327.

Page, A.L., Miller, R.H., Keeney, D.R., 1982 Methods of soil analysis: Pt 1 Chemical and microbiological properties, 2nd edn Soil Science Society of America, Madison, Wisconsin Puppala, S.K., Alshawabkeh, A.N., Acar, Y.B., Gale, R.J., Bricka, J., 1997 Enhanced electrokinetic remediation of high sorption capacity soil J Hazard Mater 55, 203 – 220.

Reddy, K.R., Parupudi, U.S., Devulapalli, S.N., Xu, C.Y., 1997 Effects of soil composition on the removal of chromium by electrokinetics J Hazard Mater 55, 135 – 158.

Reed, S.T., Martens, D.C., 1996 Copper and zinc In: Sparks, D.L (Ed.), Methods of soil analysis: Part 3 Chemical meth-ods Soil Science Society of America, Madison, WI, pp 703 – 722.

Sposito, G., Holtzclaw, K.M., Johnston, C.T., LeVesque-Madore, C.S., 1981 Thermodynamics of sodium – copper exchange on Wyoming bentonite at 298 K Soil Sci Soc Am J 45, 1079 – 1083.

Viadero Jr., R.C., Reed, B.E., Berg, M., Ramsey, J 1998 A laboratory-scale study of applied voltage on the electrokinetic separation of lead from soils Sep Sci Technol 33, 1833 – 1859.

Wada, S.-I., 1984 Mechanism of apparent salt absorption in Ando soils Soil Sci Plant Nutr 30, 77 – 83.