Electrokinetic Remediation of Tropical Soils : Effect of the Electric Potential Difference

This paper addresses the influence of electric potential difference on the electroosmotic decontamination of two tropical soils from Brazil. The laboratory testing program encompassed: (i) two residual soils of gneiss, a C-horizon saprolite clayey silty sand (soil sample A) and a B-horizon sandy silty clay (soil sample B); (ii) mineralogical and chemical analysis of the soils; (iii) addition of an aqueous solution of cadmium nitrate in the concentration of 100 mg.L to the soil samples; (iv) compaction of the mixtures specimens at the Standard AASHTO; (v) electroosmotic decontamination tests applying 5, 15 and 30 V; (vi) application of a modified sequential extraction analysis adapted to tropical soils. The results support that the PZC of soils influenced the direction of cadmium migration and that the increase in the applied electric potential difference led to the increase in the amount of extractable contaminant.


Introduction
Soil and groundwater contaminations are major problems for public health and the environment.The contaminants present in these polluted areas include metals, volatile and semi-volatile organic compounds which can be found alone, or co-existing with metals and organic contaminants (Cameselle & Reddy, 2012).Because of these environmental problems, techniques of soil decontamination have been developed and applied over the years on a worldwide basis, such as the electrokinetic remediation of soils contaminated with heavy metals.In this case, efforts have been directed to the mobility, interaction and absorption of the contaminant in the soil matrix, mainly in its clayey fraction.Lestan et al. (2008) and Thomé et al. (2013) emphasize that there is a great environmental concern in many parts of the world about soils contaminated with heavy metals, mainly due to fast industrialization, growing urbanization, agricultural practices, and inappropriate waste disposal methods.Regarding the application of the electroosmotic decontamination technique, researchers have been analyzing the mobility of these metals and how they are held in the soil matrix, mainly in the clayey fractions (Christensen, 1989;Boekhold et al., 1993;Naidu et al., 1994;Pombo, 1995;Pierangeli et al., 2005;Velten, 2012;Giannis et al., 2010;Cameselle & Reddy, 2012;Rojo et al., 2014).
Electrokinetic remediation is a technique for treating heavy metal polluted soils by applying a low direct current (in the order of mA.cm -2 of the cross-sectional area) or a low potential gradient (in the order of V.cm -1 of the distance be-tween the electrodes) at electrodes that are inserted into the ground.In the application of this technique, contaminants are transported by electroosmosis and electromigration to either the cathode or the anode, where they can be extracted.
According to Rojo et al. (2014) in the electrokinetic remediation process, water is electrolyzed by the electric field electrodes and oxidation at the anode and reduction at the cathode generate, respectively, an acid and a base front in the soil mass.The acid generated at the anode moves through the soil towards the cathode by electromigration and electroosmosis, while the base generated at the cathode moves towards the anode by electromigration and diffusion.Acid production generally enhances the process, and a high pH zone adversely affects the extraction of heavy metals from soils.In relative terms, the acid front dominates the base front due to the greater mobility of H + ion and backflow due to electroosmosis that retards the base front.
New techniques to improve the extraction of heavy metals of soils have been suggested over the years, as follows (Giannis et al., 2010): (i) the use of an electrolyte solution in order to control the electrolyte pH (Lee & Yang, 2000); (ii) sequential improvement by adding chemical reagents to enhance metal solubility (Reddy & Chinthamreddy, 2003); (iii) use of membrane ion selection to reduce or exclude the OH -ions migration from the cathode to the interior of the soil (Kim et al., 2005); and (iv) conditioning of the pH electrode compartments (Giradakos & Giannis, 2006) through the use of a new technique for electric decontamination of a silty sand with the addition of chelates in order to change the chemical form of heavy metals and thus extract them more easily from the soil through the decontamination process.
More recently, Lu et al. (2012) proposed a new way to increase the efficiency during the electrokinetic soil remediation using the technique named polarity exchange, which basically consists on inversion of electrodes polarities during the process.According to these authors, the technique is easily applied, does not require the addition of chemical elements, is dependent on the distribution of heavy metal in the soil mass, but it has been tested only for contaminated soils with a single contaminant; therefore new studies are needed for multiple contaminants.The results obtained showed that the inversion of polarities greatly increased the extraction of the heavy metals cadmium and chromium, preventing both the absorption of Cr 6+ in acidic conditions and the precipitation of Cd 2+ and Cr 3+ in alkaline media.
On the other hand, it should be emphasized that the inversion of polarities may result in a highly acidic environment that can lead to the destruction of soil minerals.So, Lu et al. (2012) state that researchers have given little attention to adsorption and desorption of heavy metals during electrokinetic remediation and especially to the removal of metal dependency upon its physicochemical form and the chemical prevailing conditions during the remediation process.In order to determine the nature of a system studied in terms of chemical forms present and their relative mobility, it is recommended to use sequential extraction analysis before and after the process, which consists of a sequence of extractions to determine the form of the metal present in the soil.
From a historical perspective, Darmawan & Wada (2002) analyzed the effect of soil clay mineralogy on the feasibility of electrokinetic remediation using the following sequential extraction analysis: (i) water extractors, in order to determine the soluble fractions; (ii) magnesium chloride, to determine the electrostatically fraction weakly adsorbed; and (iii) hydrogen peroxide digestion, followed by hydrochloric acid, to determine the fractions strongly adsorbed or those who presented complexation with the soil minerals.Another sequential extraction methodology was employed by Reddy et al. (2001), which was developed by Tessier et al. (1979), in which the analysis was divided into five phases, namely: (i) soluble or exchangeable forms; (ii) carbonates linked forms; (iii) iron and manganese oxides linked forms; (iv) organic matter linked forms; and (v) residual fraction.With respect to both sequential extraction methods here presented, the first one is usual and more complicated when compared with the second one, however it does not separate the metal strongly adsorbed from the residual fraction; on the other hand the second form has more acceptance by researchers, such as Kim et al. (2002), Chen et al. (2006), Giannis et al. (2010) and Cameselle & Reddy (2012).
In contrast, the two sequential extraction methodologies presented are not generally applied to tropical soils, who frequently do not present significant amounts of manganese oxides and carbonates.In addition, the amount of iron and aluminum oxides present in these soils is very high and they are predominantly in crystalline form, whereas in temperate soils they occur in smaller quantities and in amorphous form.Therefore, considering the limitations of the reported sequential extraction methods, Egreja Filho (2000) developed at the Federal University of Viçosa (UFV), Brazil, a new methodology of sequential extraction comprising four steps in order to extract the elements present in the following forms in soil: (i) soluble metals; (ii) exchangeable or weakly absorbed metals; (iii) specifically adsorbed metals (i.e., strongly adsorbed); and (iv) residual fraction metals.
This methodology comprises the following steps: (i) in the first extraction, distilled water acts only as an agent that washes the soil and removes the metals that are in the soluble form; (ii) in the second extraction, calcium chloride, which is a soluble salt that releases the cation Ca 2+ in solution, is used to displace the other metals or chemicals, in particular cations, which may be linked only electrostatically on the soil surface (weakly adsorbed); (iii) in the third extraction, the fluoride and phosphate anions, which have a strong specific adsorption, are used to compete for adsorption sites and thus release the surface metals present in the soil minerals without dissolution of the oxide; and (iv) in the fourth stage, there is the determination of the residual fraction (i.e., the one that is ingrained in the soil minerals) using digestion with nitric and perchloric acid, which are strong acids that create an opening, leaving only the silicates of the source material.
Therefore, taking into account that tropical soils have geotechnical properties, electrochemical parameters, and mineralogy that differ substantially from temperate soils, this research was directed to the analysis of the influence of the application of electric potential differences to two tropical young soils contaminated with cadmium nitrate using the modified sequential extraction technique proposed by Egreja Filho (2000).

Geotechnical, mineralogical and chemical characterization of soils
Two gneiss residual soil samples were collected on the UFV Campus, located in the North Forest Zone, Minas Gerais state, Brazil, as follows: (i) soil sample A is a clayey silty sand collected in the C-horizon of a soil profile located at the geographic position coordinates 20°46'48.2" of South latitude and 42°52'52" of West longitude, and classified as Saprolite; and (ii) soil sample B is a sandy silty clay soil collected at the B-horizon of a soil profile located at the geographic position coordinates 20°45'23.5" of South lati-tude and 42°50'22.4" of West longitude, and pedologically classified as Red-Yellow Argisol.Tables 1, 2, 3, 4 and 5 summarize data from Velten et al. (2012), regarding, respectively, the geotechnical, mineralogical and chemical soil parameters.

The contaminant
A cadmium mono-species solution with concentration of 100 mg.L -1 of salt was added to the soil samples in different volumes, according to their original moisture contents and considering the water amounts required to make them reach the cadmium concentration of 10 to 12 mg.kg-1 of dry soil mass, which are the minimum guiding values that require intervention in agricultural or residential contaminated areas (Cetesb-SP, 2001).The cadmium salt used was cadmium nitrate tetra hydrated [Cd(NO 3 ) 2 .4H 2 O] 99% from Riedel-de Haën with the molecular weight of 308.48 g.

Electroosmotic cell
Figure 1 introduces the electroosmotic cell used in the study, originally designed and built by Damasceno (2003), and now including graphite in substitution for copper electrodes.

Samples preparation
The soil samples were air-dried and passed through the 4.8 mm sieve.Then, compaction tests were carried out on the AASHTO Standard compaction effort (ABNT, 1986) in order to determine the soils samples optimum water content (w opt ) and maximum dry unit weight (g dmax ).During soil sample preparation, water and cadmium nitrate solution were added to the air-dried samples, in order to reach the optimum moisture content previously determined in the compaction tests, as well as the cadmium concentration of 10 to 12 mg.kg - of dry soil.Later, using incubation  times determined previously by Velten et al. (2012), the contaminated soil samples A and B were incubated, respectively, during 10 and 20 days in a temperature controlled chamber (20 °C ± 1 °C).

Test procedure
After soils samples incubation, a small fraction of each of the contaminated soil sample was taken to the UFV's Laboratory of Mineralogy of the Department of Soils in order to perform sequential extraction test (Egreja Filho, 2000) and pH determination.From each remaining contaminated soil sample, specimens were molded at the AASHTO Standard optimum compaction parameters (w opt and g dmax ) using Plexiglas cylinders to be tested in the electroosmotic cell.After placing each specimen in the electroosmotic cell, its anode and cathode were filled with distilled water up to the desired level.
Decontamination tests were carried out under controlled temperature (20 °C ± 1 °C) for a period of time of 216 h, using electric potential differences of 5, 15 and 30 V, allowing the electric current to vary throughout the test.Considering that the electrodes of the electroosmotic cell were 180 mm apart, electric gradients of 0.28, 0.83 and 1.67 V.cm -1 , respectively, were generated.During the decontamination tests, solutions were collected each 3 days from the anode and cathode reservoirs, in order to determine their cadmium concentrations using the Flame Atomic Absorption Spectrophotometry technique from the UFV' Soils Laboratory.Finally, at the end of the decontamination tests, soils specimens were taken from the cell and subdivided into five equally apart layers, in order to determine their cadmium concentrations.

Sequential extraction analyses
After performing the decontamination tests, the soils specimens were subdivided into five equal parts, which were identified and submitted to the sequential extraction analysis proposed by Egreja Filho (2000) using different extractors in each step, as well as higher extraction power as the process advanced.In this test, the extractor acted by changing the interaction between the heavy metal and the solid phase, promoting solubilization to be dosed by a convenient analytic method.The sequential extraction was carried out in four steps, as illustrated in Table 6 (Velten et al., 2012).

Evaluation of electric potential difference effect in the electroosmotic conductivity coefficient
Table 7 presents the geotechnical parameters of the tested soils before and after performing the electroosmotic decontamination tests in order to evaluate the effect of the electric potential difference in the coefficient of electroosmotic conductivity.From the electroosmotic decontamination tests performed in the specimens of the tested soil samples, it was determined the coefficient of electroosmotic conductivity (k e ).During the tests, it was observed that the electric current decreased with time and the electroosmotic flow increased at the beginning of each test, tending to stabilize with time.In Table 7, it is also noticed that the coefficient of electroosmotic conductivity of soil sample A decreased slowly when increasing the applied electric potential difference, while in soil sample B this parameter increased slowly when increasing the applied electric potential difference.Results obtained by Chang & Liao (2006) using electric gradients of 2, 1 and 0.5 V.cm -1 showed that the electroosmotic flow obtained at the electric gradient of 2 V.cm -1 was two times and six times higher than those obtained at 1 and 0.5 V.cm -1 respectively.However, independently of the applied electric gradient, in the present research the coefficient of electroosmotic conductivity showed values in the same order of magnitude, i.e. 10 -6 cm 2 .s - .V -1 .

Cadmium concentration in the compartments and pH
Figures 2 and 3 present, respectively, the cadmium concentration in the anode and cathode compartments during decontamination tests and pH values determined before and after tests.It should be emphasized that during the decontamination tests, contrary to the behavior of soil sample A, soil sample B presented reverse electroosmotic flow direction, i.e., from the cathode to the anode.In this case, the main explanation for the inversion in the direction of the electroosmotic flow can be related to the behavior of the electrochemical parameter PZC, in this case higher than the pH of the soil sample B, and in accordance with Cameselle & Reddy (2012) previous observation.
The results illustrated in Fig. 2 show that in soil sample A, a sandy soil with pH higher than its PZC, the cadmium concentration in both cell compartments decreased from 3 to 6 days, but increased again from 6 to 9 days, reaching values higher than those determined at 3 days.Regarding soil sample B, a clayey soil with pH smaller than its PZC, in Figure 3 it was observed that the cadmium concentration decreased with time in both cell compartments.
Besides, from Fig. 3 it was observed that the pH of soil samples A and B decreased from cathode to anode, in accordance with Yang & Lin (1998), and, in general, increased after application of the electroosmotic decontamination technique, independently of flow direction.Because of the induced electric potential in the tests, soil pH decreased close to the anode and increased close to the cathode, when H + ions created close to the anode compartment migrated to the cathode compartment (negative electrode) and OH -generated close to the cathode migrated to the anode compartment (positive electrode).As a consequence, pH reached low values in the anode reservoir and high ones in the cathode and, according to Reddy et al. (2001), this pH difference distribution in the specimen created deep and different effects in the contaminant distribution and migration.

Sequential extraction data analysis
Figures 4 and 5 show the results of the sequential extraction analysis performed into the five vertical layers labeled 1 to 5 (2 cm long each) of the contaminated specimens taken before (named initial) and after performing the electroosmotic decontamination tests (named final).Regarding the tested soils and considering all cadmium extracted forms, these figures show that the behavior of the cadmium ion varied in the electroosmotic remediation process and with the applied electric potential difference.Analysis of the data presented in Figs. 4 and 5 support that: • Soil sample A specimens: In general, from Fig. 4 it was noticed an increase in the amount of cadmium in the direction of the electroosmotic flow (anode to cathode) with the increase in the applied electric potential difference.The exception occurred for the electric potential difference of 5 V, because specifically adsorbed and residual cadmium fractions at the end of the decontamination process should not exist in the vicinity of the anode.At 5 V, it was observed that occurred accumulation of cadmium in sections close to the anode, resulting in low 254 Soils and Rocks, São Paulo, 39(3): 249-260, September-December, 2016.
Velten et al. contaminant removal from the soluble fraction.This behavior could be related to a possible accumulation of the metal in this fraction derived from other extractions that were not fully removed, as well as due to both the low intensity of the applied electric potential and the relatively short testing time.In all the contaminant forms in the soil A solid matrix, the contaminant migrated from the anode to the cathode, following the electroosmotic flow.In the vicinity of the cathode, an increase in the applied electric potential difference increased the concentration of cadmium ions, contrary to the behavior observed in the middle sections and in those close to the anode, which highlights the influence of the applied potential difference in removing the contaminant.
• Soil sample B specimens: As referred before, it was observed reversion of the electroosmotic flow direction, i.e. from cathode towards anode, as depicted in Fig. 5.In this case, the fraction was barely detected, the electrostatically attracted fraction showed almost constant behavior in all sections of the specimen, and the cad-  mium concentration was higher than that determined by previous testing in some sections of the specimens, for 15 and 30 V. Regarding the strongly adsorbed and residual fractions, respectively, cadmium ion activity, respectively, increased and decreased in the vicinity of the cathode with the increase in the applied electric potential difference.
Table 8 shows the percentages of cadmium removed in each extraction phase in comparison with its initial value.
Data from this table support that: • At the beginning of the decontamination process, each specimen slice presented the same cadmium concentration that could be represented by an area contained by a horizontal line in a figure that shows cadmium concentration in the ordinate axis (Y) and distance from the cathode in the abscissa axis (X).At the end of each phase of the sequential extraction analysis, the curve of cadmium concentration could be obtained from the amounts of contaminant determined from each slice of the tested specimen.Therefore, from the comparison of these areas, it can be inferred the amount of cadmium extracted from the specimen during the electroosmotic decontamination test.• Decontamination rates of soil samples A and B were strongly influenced by the applied electric potential difference, although the results obtained when applying 15 and 30 V to soil A specimens did not differ significantly when considering the soluble and exchangeable fractions.Therefore, application of an electric gradient of 0.50 V * cm -1 during the remediation process of soils similar to soil A contaminated with cadmium would be a fair estimate when there is no information available on theirs electroosmotic behavior, in agreement with Alshawabkeh et al. (1999) and Mitchell & Soga (2005).
Differences in the behavior of specimens from soil samples A and B during the decontamination tests could be related to the differences in their mineralogy, in particular due to the large amount of iron oxides present in soil sample B, mainly in the form of goethite.The mobility of cadmium in this soil was greatest when applying 30 V, which was the test that presented final pH values generally smaller than the initial one, responding to the least amount of negative charges in the soil and, consequently, greater cadmium mobility.Therefore, the presence of goethite and other iron oxides such as hematite, and also oxides of aluminum, such as gibbsite, can influence considerably the phenomena of adsorption and removal of cadmium from the soils.
From Fig. 5, it was also noticed accumulation of cadmium in the residual fraction of specimens of soil sample B in the vicinity of the cathode, supporting that the strong acidification of the soil at the anode could be responsible for possible destruction of soil minerals in order to favor the removal of contaminants, as well as the strong alkalinization of the cathode may cause precipitation of contaminants in the form of insoluble salts or hydroxides, and may block the porous medium when pH reaches values near to 9.0.Finally, it was observed in the tested soils that the application of electric potential difference of 30 V led to the highest cadmium decontamination rate.

Conclusions
The analysis of the results supports the following conclusions: • The applied electric potential difference did not influence the value of the coefficient of electroosmotic conductivity of the tested soils, for engineering applications.
• The PZC of the tested soils influenced the cadmium direction migration.In soil A, the application of the electric potential differences generated cadmium migration in the electroosmotic flow direction.However, in soil B, the contaminant migrated in the anode-cathode direction, bringing up to attention the importance of the relationship between soil PZC and pH in the electroosmotic decontamination process.
• In general, the values of pH of all tested soils determined after decontamination tests were higher than those determined before, increasing from the anode to the cathode.
• The increase in the applied electric potential difference led to the increase in the amount of extractable contaminant.
• In soil B, the large presence of iron oxides in the form of goethite can be a preponderant factor in the removal of cadmium.
• In all decontamination tests, it was observed that cadmium in the residual fraction migrated from the anode towards the cathode, supporting that the strong acidification of the anode could have resulted in destruction of the constituents of the soil, thus facilitating the removal of chemical elements, as well as that alkalinization of the cathode, resulted in deposition and precipitation of the contaminant.

Figure 2 -
Figure 2 -Cadmium concentration in the anode and cathode compartments during decontamination tests: (a) soil sample A; and (b) soil sample B.

Figure 3 -
Figure 3 -pH values determined in the in the specimens before and after decontamination tests: (a) soil sample A; and (b) soil sample B.

Figure 4 -
Figure 4 -Results from the sequential extraction analysis performed in the specimens from soil sample A before and after performing the electroosmotic decontamination tests: (i) incubation time of 10 days; (ii) application of the electric potential difference of 5, 15 and 30 V; and (iii) specimens divided into five vertical sections 2 cm long after the remediation tests.

Figure 5 -
Figure 5 -Results from the sequential extraction analysis performed in the specimens from soil sample B before and after performing the electroosmotic decontamination tests: (i) incubation time of 20 days; (ii) application of the electric potential difference of 5, 15 and 30 V; and (iii) specimens divided into five vertical sections 2 cm long after the remediation tests.

Table 1 -
Geotechnical characterization of soils.

Table 3 -
Mineralogy of soils A and B -Qualitative results.

Table 7 -
Results of geotechnical parameters of soils specimens submitted to the electroosmotic decontamination tests.
f -water content after test.Sr o -saturation degree before test.Sr f -saturation degree after test.k stable -coefficient of electroosmotic conductivity after flow stabilization.
s -density of solid particles