Available online at www.sciencedirect.com

www.elsevier.com/locate/chemosphere

Chemosphere 70 (2008) 1329–1337

Removal of heavy metals and arsenic from contaminated soils usingbioremediation and chelant extraction techniques

Katerina Vaxevanidou, Nymphodora Papassiopi *, Ioannis Paspaliaris

School of Mining and Metallurgical Engineering, National Technical University of Athens, GR 157 80 Zografos, Greece

Received 14 August 2007; received in revised form 12 October 2007; accepted 15 October 2007Available online 26 November 2007

Abstract

A combined chemical and biological treatment scheme was evaluated in this study aiming at obtaining the simultaneous removal ofmetalloid arsenic and cationic heavy metals from contaminated soils. The treatment involved the use of the iron reducing microorganismDesulfuromonas palmitatis, whose activity was combined with the chelating strength of EDTA. Taking into consideration that soil ironoxides are the main scavengers of As, treatment with iron reducing microorganisms aimed at inducing the reductive dissolution of soiloxides and thus obtaining the release of the retained As. The main objective of using EDTA was the removal of metal contaminants, suchas Pb and Zn, through the formation of soluble metal chelates. Experimental results however indicated that EDTA was also indispens-able for the biological reduction of Fe(III) oxides. The bacterial activity was found to have a pronounced positive effect on the removal ofarsenic, which increased from the value of 35% obtained during the pure chemical treatment up to 90% in the presence of D. palmitatis.In the case of Pb, the major part, i.e. approximately 85%, was removed from soil with purely chemical mechanisms, whereas the biolog-ical activity slightly improved the extraction, increasing the final removal up to 90%. Co-treatment had negative effect only for Zn, whoseremoval was reduced from 80% under abiotic condition to approximately 50% in the presence of bacteria.� 2007 Elsevier Ltd. All rights reserved.

Keywords: Soil remediation; Iron reducing bacteria; EDTA; Arsenic; Heavy metals

1. Introduction

Contamination of soils with heavy metals and metal-loids, such as lead, cadmium, arsenic, etc., represents a seri-ous threat for the ecosystem and human health andrequires the implementation of appropriate remedial mea-sures. Multi-contaminated soils usually require more thanone single treatment option so that all contaminants whichmay be present in the soil are efficiently removed. A typicalcase is the simultaneous presence of cationic heavy metalsand anionic metalloids, which cannot be treated with thesame techniques due to their different chemical characteris-tics. For instance, cations, such as Pb, Zn or Cd, can begenerally removed from the soil using either acids or che-lating reagents, as it has been demonstrated by several

0045-6535/$ - see front matter � 2007 Elsevier Ltd. All rights reserved.

doi:10.1016/j.chemosphere.2007.10.025

* Corresponding author. Tel.: +30 210 7722298; fax: +30 210 7722168.E-mail address: papasiop@metal.ntua.gr (N. Papassiopi).

studies (Neale et al., 1997; Wasay et al., 1998; Papassiopiet al., 1999; Theodoratos et al., 2000; Tampouris et al.,2001; Sun et al., 2001; Manouchehri et al., 2006). Acidicenvironments favour desorption of metal cations as wellas the dissolution of oxides or carbonates, which are com-mon mineral phases found in contaminated soils. On theother hand, various chelating reagents, such as EDTA (eth-ylene diamine tetraacetic acid), DTPA (diethylene triaminepentaacetic acid) etc., can form strong soluble complexeswith most divalent and trivalent metal contaminants,resulting thus in the recovery of these contaminants inthe aqueous solution. Arsenic however, which occurs inthe form of oxyanions, AsO4

�3 or AsO3�3, can not be

mobilised with the same mechanisms. The techniques,which have been investigated for the removal of arsenicfrom soils take into consideration the strong associationof As oxyanions with soil iron oxides and try to obtainthe release of arsenic either through the selective desorption

Table 1Typical reactions taking place during the treatment of soil withCaEDTA2� and D. palmitatis

Typical reactions

8FeOOHþ CH3COO� þ 15Hþ !D:palmitatis8Fe2þ þ 2HCO�3 þ 12H2O (1)

Fe2+ + CaEDTA2� �! FeEDTA2� + Ca2+ (2)�Pb + CaEDTA2� �! PbEDTA2� + Ca2+ (3)�Zn + CaEDTA2� �! ZnEDTA2� + Ca2+ (4)Ca2+ + HCO3

� �! CaCO3 + H+ (5)

�Pb, �Zn: denote Pb and Zn species in soil.

Table 2Selected properties of soil material

pH 8.65CaCO3 equ. (%) 39.5Ctotal (%) 4.80Corg (%) <0.1

mg kg�1 mmol kg�1

Ca 340600 8494Mg 8100 333Al 7300 270Mn 1960 36Fe 25000 448Pb 19600 95Zn 23900 365Cd 357 3As 946 13

1330 K. Vaxevanidou et al. / Chemosphere 70 (2008) 1329–1337

of arsenic species from oxides surface or through the com-plete dissolution of iron oxides.

Desorption of arsenic can be obtained by treating thesoil with alkaline solutions or with solutions containingcompetitive oxyanions, like phosphates, which can replacearsenates on the surface coordination sites of oxides (Jack-son and Miller, 2000; Alam et al., 2001; Jang et al., 2005).Acidic treatment, when applied for the remediation ofarsenic contaminated soils, has as main target the dissolu-tion of iron oxides (Tokunaga and Hakuta, 2001), in orderto obtain indirect release of arsenic. Use of acids however isnot appropriate for the treatment of calcareous soils due totheir high acid buffering capacity. Another alternative isthe dissolution of Fe(III) oxides by reductive mechanisms,which can take place under less aggressive pH conditions,using either chemical or biological processes. Removal ofAs from contaminated soils applying anaerobic bioremedi-ation techniques has been recently investigated by Chatainet al. (2005) and Ignatiadis and Battaglia-Brunet (2005).Both research teams showed that it is possible to obtainthe mobilization of arsenic by stimulating the activity ofindigenous microorganisms with addition of appropriatecarbon and energy sources under anaerobic conditions.The degree of As mobilization was found to depend onthe specific characteristics of contaminated soil samples,e.g. the crystallinity of iron oxides, the organic matter con-tent, etc. (Ignatiadis and Battaglia-Brunet, 2005).

The soil sample used in the current experimental work isheavily contaminated both with metal cations, mainly Pband Zn, as well as with As oxyanions, due to past miningactivities. Previous studies (Papassiopi et al., 1999; Theo-doratos et al., 2000) have demonstrated that treatment ofthis soil with the strong chelator EDTA can be very effi-cient for the extraction of cationic metal contaminants,but has rather limited efficiency in the removal of arsenic.It was also shown, that due to the calcareous matrix ofthe soil it is better to use the calcium salt of EDTA,Na2CaEDTA, instead of the commonly used sodium salts,in order to avoid the dissolution of calcite.

In this work EDTA treatment is coupled with the use ofthe iron reducing microorganism Desulfuromonas palmitatis

in order to enhance the reductive dissolution of iron oxidesand obtain simultaneous release of arsenic in the aqueoussolution. D. palmitatis is an anaerobic microorganism,which is able to use Fe(III) as terminal electron acceptor inits respiration process with simultaneous oxidation of a widevariety of organic compounds, such as acetate, fumarate,palmitate, stearate and succinate (Coates et al., 1995).

The typical reactions, which are expected to take placeduring this treatment, are given in Table 1. The main con-tribution of D. palmitatis is the reductive dissolution ofFe(III)-oxides in order to obtain the release of the retainedAs (reaction (1) in Table 1). The addition of Na2CaEDTAhas a double objective: (a) to maintain biogenic Fe(II) insolution (reaction (2) in Table 1) and (b) to remove the cat-ionic metal contaminants, i.e. Pb, Zn, from soil (reactions(3) and (4) in Table 1). In the presence of carbonates, the

decomplexed calcium cations precipitate as CaCO3, asshown in reaction (5) of Table 1.

Parameters examined during the experimental workinclude (a) the effect of NaHCO3, which was added tothe slurries in order to facilitate precipitation of CaCO3

and increase the availability of EDTA for complexationwith the target metals Pb, Zn and Fe(II) and (b) the liquidto solid (L/S) ratio, which was modified from the value of20:1 to that of 2:1 ml g�1.

2. Materials and methods

2.1. Soil material

Selected properties of the contaminated soil material arepresented in Table 2. It is a calcareous soil containing39.5% of calcite equivalent and exhibiting alkaline pH of8.65. It contains 4.8% total carbon, which is all inorganic.Chemical analysis indicated that the material containshigh concentrations of contaminants, such as Pb and Zn,up to 19600 and 24000 mg kg�1, respectively, As up to950 mg kg�1 and Cd 360 mg kg�1. Speciation analysisshowed that all soil As is in the pentavalent state, As(V).Examination of this soil material with scanning electronmicroscopy and microanalysis techniques revealed the pres-ence of several discrete carbonate and arsenate mineralssuch as PbCO3, Pb3(CO3)2(OH)2, ZnCO3, Pb5(AsO4)3Cl,Zn2AsO4OH and FeAsO4Æ2H2O, as well as contaminantsassociated with Fe(III) oxy-hydroxides containing 2–5%PbO, 2–6% ZnO and 1–2% As2O5 (Xenidis et al., 2003).

Table 3Summary of experimental conditions

A. Comparison of chemical and biological treatment – effect ofNa2CaEDTA and NaHCO3 addition

Standard conditions:Liquid/solid ratio: L/S = 20:1 (80 ml solution: 4 g soil)Temperature: 30 �CpH = 7.0 (buffered by the presence of NaHCO3 and pCO2 = 0.2 atm in

the headspace)Slurries autoclaved before inoculationVariable conditions:

Composition of aqueous solution:– Basic mediuma (with Na-acetate 10 mM)– Basic mediuma (with Na-acetate 10 mM) + Na2CaEDTA 100 mM– Basic mediuma (with Na-acetate 10 mM) + Na2CaEDTA

100 mM + NaHCO3 100 mMInoculated and non inoculated tests

B. Effect of L/S ratioStandard conditions:Temperature: 30 �CComposition of aqueous solution:– Basic mediuma (with Na-acetate 20 mM) + Na2CaEDTA 100 mM +

NaHCO3 100 mMpH = 7.0 (buffered by the presence of NaHCO3 and pCO2 = 0.2 atm in

the headspace)Slurries not autoclavedVariable conditions:L/S ratios: L/S = 20:1, 10:1, 5:1, 2:1 (80 ml solution: 4 g, 8 g, 16 g and

40 g soil)

a Constituents of basic medium as described in Section 2.2 withoutaddition of Na-fumarate.

K. Vaxevanidou et al. / Chemosphere 70 (2008) 1329–1337 1331

2.2. Microorganisms and media

The microorganism used for the biological treatmentof soil, i.e. D. palmitatis (DSM 12391), was purchasedfrom the German Collection of Microorganisms DSMZ(Deutschen Sammlung von Mikroorganismen und Zellk-ulturen, Braunschweig, Germany). The initial growthwas carried out in the medium suggested by DSMZ(Medium 837), which contained (mM): NaCl 267,NaHCO3 30, KCl 10, NH4Cl 19, NaH2PO4 H2O 0.72,MgSO4 7H2O 0.81, CH3COONa 3H2O 10, Na-fumarate50, selenite-tungstate solution 1 ml l�1, vitamin solution10 ml l�1, trace element solution 10 ml l�1. In this med-ium fumarate (50 mM) is used as electron acceptor andacetate (10 mM) as electron donor. Incubations were car-ried out at 30 �C under anaerobic conditions (N2-CO2

atmosphere of 80:20 ratio). It is noted that during growththe pH of the medium is buffered close to the value of 7.0by the presence of NaHCO3 in the aqueous solutions andthe partial pressure of CO2 (20.3 kPa) in the headspace ofserum bottles. Growth of bacteria was determined bymeasuring the corresponding total protein with Bradfordassay (Bradford, 1976).

2.3. Experimental procedure during the treatment of soil

Standard anaerobic techniques were used throughout.All the tests were carried out using 125 ml serum bottlessealed with butyl rubbers and aluminium crimps. For thepreparation of experimental slurries, a predeterminedamount of soil was placed in the serum bottles. The aque-ous phase constituents were added from concentrated stocksolutions and the volume was adjusted with deionisedwater to 72 and 80 ml for the biological and control tests,respectively. The slurries were bubbled for approximately20 min with N2:CO2 (80:20) to remove dissolved oxygenand then sealed and autoclaved. After cooling, the slurriescorresponding to the biological tests were inoculated with8 ml from a freshly developed D. palmitatis culture contain-ing a cell concentration equivalent to approximately70 ± 5 mg l�1 proteins and the serum bottles were incu-bated horizontally on a shaking waterbath.

During the tests, aliquots of 1 ml were withdrawn fromthe serum bottles using a syringe. The samples were filteredthrough a 0.2 lm RC-membrane filter. The filtrate wasdiluted to 10 ml with deionised water and analysed forFe, As, Pb, Zn, Cd, and Ca with atomic adsorption spec-trophotometry. The pH, the redox potential, Eh, and theconcentration of EDTA anions (Wallace and Hinton,1970) were determined only in the final solutions.

2.4. Experimental conditions

Two series of tests were conducted and the experimen-tal conditions are summarised in Table 3. The first seriesaimed at comparing the biological and the chemical treat-ment of soil using the iron reducing microorganism

D. palmitatis and the chelating reagent Na2CaEDTA,when these two treatment options are separately appliedon the soil or when they are combined in order to maxi-mize the overall effectiveness of the treatment. The effectof NaHCO3 addition was also investigated during theseinitial tests. The tests were carried out using a L/S ratioof 20:1 ml g�1 and the operation temperature was con-trolled at 30 �C. The main parameter that was varied dur-ing these experiments was the composition of the aqueoussolution. Two tests (one inoculated and one similar asabiotic control) were conducted with an aqueous solutioncontaining only the constituents of the basic culturemedium for D. palmitatis growth. The medium was pre-pared with 10 mM of Na-acetate as electron donor butwithout addition of fumurate in order to compel themicroorganism to use the Fe(III) oxides of soil as electronacceptor. Two other tests were carried out under similarconditions but adding 100 mM of Na2CaEDTA in theaqueous solution. In the last two experiments the solutionwas spiked with both Na2CaEDTA 100 mM andNaHCO3 100 mM.

The second series of tests aimed at investigating theeffect of L/S ratio. Experiments were carried out usingboth Na2CaEDTA and NaHCO3 in the aqueous solu-tion. The volume of the solution was kept constant,80 ml, and the amount of soil was increased from 4 gup to 40 g to obtain four levels of L/S ratios, i.e. 20:1,10:1, 5:1 and 2:1 ml g�1.

1332 K. Vaxevanidou et al. / Chemosphere 70 (2008) 1329–1337

3. Results and discussion

3.1. Evaluation of chemical and biological treatment – effect

of Na2CaEDTA and NaHCO3 additions

Results concerning the dissolution of iron and theextraction of contaminants are shown in Fig. 1. Whenthe treatment was carried out using only the medium ofD. palmitatis growth (dotted lines with diamonds inFig. 1), either under abiotic or biological conditions, therewas no detectable dissolution for none of the four elementsexamined. It is thus evident that the various constituents of

0

20

40

60

80

Fe, %

dis

solu

tion

Medium

Medium + CaEDTA

Medium + CaEDTA + NaHCO3

(a) Fe, abiotic tests

0

20

40

60

80

100

As,

% e

xtra

ctio

n

(b) As, abiotic tests

0

20

40

60

80

100

Pb, %

ext

ract

ion

(c) Pb, abiotic tests

0

20

40

60

80

100

0 50 100 150t, days

Zn,

% e

xtra

ctio

n

(d) Zn, abiotic tests

Fig. 1. Comparison of abiotic (a)–(d) and biological (e)–(h) tests. Effect ofdissolution of Fe; (b), (f) extraction of A; (c), (g) extraction of Pb, (d), (h) ext

growth medium cannot mobilise chemically the contami-nants. Moreover, as seen in Fig. 1e, inoculation with D.

palmitatis had no apparent effect on Fe. Despite theabsence of any increase on the levels of aqueous Fe duringthe biological treatment of soil, reduction of Fe(III) oxidesand reprecipation of biogenic Fe(II) cannot be excluded.Precipitation of insoluble ferrous compounds, such as mag-netite, green rust, siderite, vivianite or ferrous arsenates,during the biological reduction of ferric oxides, have beenreported in several studies (Fredrikson et al., 1998; Cum-mings et al., 1999; Papassiopi et al., 2003). To examine thispossibility, the solid residues of the biological tests were

0

20

40

60

80

Medium

Med. + CaEDTA

Med. + CaEDTA + NaHCO3

(e) Fe, biological tests

0

20

40

60

80

100(f) As, biological tests

0

20

40

60

80

100(g) Pb, biological tests

0

20

40

60

80

100(h) Zn, biological tests

0 50 100 150t, days

Na2CaEDTA and NaHCO3 presence in the aqueous solution. (a), (e)raction of Zn.

K. Vaxevanidou et al. / Chemosphere 70 (2008) 1329–1337 1333

subjected to an acidic treatment with 6 M HCl and theextract was analysed for Fe(II) with the phenanthrolinemethod (Harvey et al., 1955). This treatment is reportedto dissolve most ferrous iron compounds including therefractory mixed Fe(II)–Fe(III) oxide, magnetite (Donget al., 2000). The concentration of ferrous iron, Fe(II),was found to be below the detection limit of the assay inall solid residues. It was thus concluded that D. palmitatis

were not able to reduce the Fe(III) oxides of soil duringthe experiment that was carried out without addition ofEDTA.

The addition of EDTA had very limited effect on the dis-solution of iron in the abiotic test. As seen in Fig. 1a, only1.4% of Fe was found in solution, which indicates thatCaEDTA does not dissolve chemically the Fe(III) of soiloxides. It should be noticed that EDTA forms very strongcomplexes with trivalent iron, as indicated by the value oflog stability constant, logKFe(III)-EDTA = 27.7, which is sev-eral orders of magnitude higher than that of CaEDTA,logKCaEDTA = 12.4 (at 25 �C and I = 0; Gustafsson,2003). However most common iron(III) oxides, such asgoethite, have very low solubility at neutral pHs and can-not be dissolved by chelation with EDTA. Visual Minteq2.2 software (Gustafsson, 2003) was used to compare thesolubility of goethite in water with its solubility in the pres-ence of CaEDTA2�. Calculations indicated that at pH = 7the solubility of FeOOH increases from 7.6 · 10�13 M inwater to 1.1 · 10�3 M in the presence of CaEDTA2�

(0.1 M). The concentration of aqueous Fe measured in thisexperimental work was of the same order of magnitudewith the calculated value, i.e. 0.3 mM. This amount of dis-solved Fe(III) in the presence of EDTA corresponds to animportant increase in solubility but represents only a smallpercentage of soil iron oxides, i.e. 1.4%.

Comparison between Fig. 1a and e clearly illustratesthat appreciable dissolution of iron up to 60% is obtainedonly when the soil is subjected to the biological treatmentwith D. palmitatis and more particularly when the activityof D. palmitatis is combined with the presence of EDTA.When the aqueous solution contains only the constituentsof culture medium, the bacteria have no apparent effecton soil iron oxides. Dissolution of iron is observed onlywhen the activity of D. palmitatis is supported by the pres-ence of EDTA.

It was not possible to determine directly the oxidationstate of dissolved iron during the experiments that were car-ried out in the presence of EDTA, due to the strong interfer-ence of EDTA with the phenanthroline method. However,an indirect indication that iron dissolution took place viareductive mechanisms was provided by the redox potentialvalues, which were determined in the final solutions. Moreparticularly, the Eh values were found to be 45 mV (atpH = 7.78) and 53 mV (at pH = 7.82) in the two experi-ments that were carried out in the presence of D. palmitatisand EDTA. This combination of Eh, pH values indicate theestablishment of mild reducing conditions, which are typicalin ecosystems with rich iron reducing microbial activity.

On the contrary, the aqueous solution resulting from thebiological treatment without EDTA had a high redoxpotential, Eh 385 mV (at pH = 7.46), providing an addi-tional evidence that D. palmitatis were not able to developtheir reducing activity in the absence of EDTA.

The role of chelating reagents during the microbialreduction of Fe(III)-oxides has been the subject of severalstudies (Urrutia et al., 1999; Roden and Urrutia, 2002;Nevin and Lovely, 2002. Urrutia et al. (1999) and Rodenand Urrutia (2002) attribute the enhancing effect of chelat-ing reagents to the complexation of biogenic Fe(II). Theabove researchers suggest that the ferrous iron, which isgenerated during the initial stages of bioreduction, inhibitsthe continuation of the microbial reductive activity by twomain mechanisms, i.e. sorption on the surface of Fe(III)-oxides and biosorption on the cell surfaces of the microor-ganisms. According to this mechanism, the inhibitory effectof Fe(II) is decreased when bioreduction takes place in thepresence of Fe(II) complexants. By complexing Fe(II) assoon as it is produced, the surface of Fe(III) oxides, as wellas the surface of bacteria cells, remain free and bioreduc-tion can freely proceed.

Nevin and Lovely (2002) propose a different mechanism.They suggest that Fe(III) chelators may alleviate the needfor the iron reducing microorganisms to establish directcontact with the solid Fe(III) oxide, by dissolving throughchelation a small amount of Fe(III). In this case the termi-nal electron acceptor in the microbial metabolic process is asoluble form of Fe(III), which is generally much more bio-available compared to the Fe(III) on the surface of crystal-line ferric oxides.

In the case of our experiments, the positive role ofEDTA in the bioreduction of Fe(III)-oxides is evident,but it is not possible to distinguish, which one of the abovesuggested mechanisms is preponderant.

The release of As during the abiotic and biological treat-ment of soil is shown in Fig. 1b and f, respectively. Treat-ment of soil only with culture medium under abioticconditions had no effect on As. The addition of Na2CaE-DTA under abiotic conditions resulted in 40% As extrac-tion. This fraction of As corresponds probably to Pb andZn arsenate compounds, which can be dissolved chemicallyvia the chelating action of EDTA. The biological treatmentwithout EDTA caused only a slight mobilisation of As inthe order of 3%. In the presence of Na2CaEDTA, theremoval of As was more than 70% and the addition ofNaHCO3 further improved the extraction up to 90%. Theobserved increase of As extraction in the biological testsis apparently related to the bacterial dissolution ofFe(III)-oxides.

Lead extraction is only slightly affected by the presenceof D. palmitatis. As shown in Fig. 1c and g, lead is dis-solved up to 80% in 120 days and this extraction is mainlyobtained through the chelating action of EDTA. The addi-tion of NaHCO3 slightly increases Pb concentration insolution to 89% in the inoculated sample. It is worthwhileto mention that a high percentage of Pb, i.e. more than

1334 K. Vaxevanidou et al. / Chemosphere 70 (2008) 1329–1337

65%, was dissolved during the preparation of the slurries,which involved sterilization at 120 �C for 20 min.

Different results were obtained for Zn (Fig. 1d and h)since biological treatment seems to hinder the dissolutioncompared to simple chemical treatment. In the control test90% of Zn was extracted in the aqueous phase, while only70% was detected in the corresponding test with D. palmi-

tatis. Without the addition of NaHCO3, zinc extraction inthe inoculated test was further less, i.e. 62%. Immobiliza-tion of zinc in highly insoluble mineral phases during themicrobial reduction of Fe(III) oxides has been reportedby Cooper et al. (2000). These researchers have found thatduring the reduction of goethite (a-FeOOH) and lepidocro-cite (c-FeOOH) with the iron reducing microorganismShewanella putrefaciencs, Zn(II), which has the same ionicradius as Fe(II), was incorporated in the structure of char-acteristic Fe(II) mineral phases, such as siderite (FeCO3),magnetite (Fe3O4), and a new unidentified mineral. Zinc-substituted siderite was soluble in 0.5 M HCl, but Zn-enriched magnetite as well as the new unidentified mineralwas insoluble when subjected to this treatment. Our exper-iments were conducted in the presence of CaEDTA2�,which can dissolve many ferrous compounds. Calculationsthat were carried out using VMinteq 2.2 software indicatedthat FeCO3 can be quantitatively dissolved according toreaction: FeCO3 + CaEDTA2� ! FeEDTA2� + CaCO3.In contrast with siderite, magnetite, Fe3O4, cannot be dis-solved by CaEDTA. The solubility of this compound wascalculated to be very low, i.e. 0.32 mM at pH = 7 in thepresence of 0.1 M CaEDTA2�. Based on the above discus-sion a potential explanation for the lower degree of Zn dis-solution during the biological treatment of soil with theFe(III) reducing bacteria could be the sequestration of Znin insoluble ferrous compounds, similar to that reportedby Cooper et al. (2000). It should be mentioned howeverthat it was not possible to detect any magnetite or otherknown ferrous compound, when the residue was subjectedto XRD analysis. Even if such compounds have beenformed, their relative percentage in the soil material wouldbe rather low to be detected by XRD techniques.

3.2. Effect of L/S ratio

Experiments investigating the effect of L/S ratio wereconducted without prior sterilization of soil slurries. Itwas thus possible to evaluate whether any indigenousmicroorganisms can be stimulated to perform the microbialreduction of Fe(III)-oxides, as observed in other experi-mental studies (Chatain et al., 2005; Ignatiadis and Batta-glia-Brunet, 2005). During the control tests the soil wasmixed with aqueous solutions containing all micro andmacro nutrients that were also used for the experimentscarried out in the presence of D. palmitatis. However noneof the control tests gave any indication of native ironreducing microbial activity. In all the cases, iron dissolu-tion was less than 3% and was mainly due to the chemicalaction of EDTA. It is reminded that this soil is very poor in

organic matter as seen in Table 2, a fact which justifies theabsence of any apparent indigenous microbial activity.

This series of tests was conducted using constant volumeof aqueous solution (80 ml), constant initial concentrationof Na2CaEDTA in the aqueous solution (100 mM) andvariable amount of soil. Modifying the quantity of soil thatwas mixed with the constant solution volume had as conse-quence a proportional change in the amount of EDTAwhich is available for the chelation of target metals. In thistreatment scheme the target metals are Pb and Zn, as maincationic contaminants, as well as biogenic Fe(II), whichmust be maintained in solution to avoid secondary precip-itation of low solubility ferrous compounds. It must benoted that the initial amount of Fe + Pb + Zn in the soilis 909 mmol kg�1 and that of Pb and Zn 463 mmol kg�1.Considering the constant initial concentration 100 mM ofNa2CaEDTA, it can be calculated that the molar ratiosof EDTA versus the sum of target metals, i.e. EDTA/(Fe + Pb + Zn), are equal to 0.22, 0.55, 1.1 and 2.2, whenthe L/S ratio is 2, 5, 10 and 20 ml g�1, respectively.

The experimental results at variable L/S ratios are pre-sented in Fig. 2. Dissolution of Fe (Fig. 2a) was observedonly in the biological tests that were carried out with L/Sratio greater than 2 ml g�1. Percentage of dissolved Fewas 62% at L/S = 20 ml g�1, 47% at L/S = 10 ml g�1 and14% at L/S = 5 ml g�1. At the lower L/S ratio 2 ml g�1,there was no Fe dissolution either in the control or in thebiological test.

In the dense slurries, i.e. L/S = 2 ml g�1, all the chelat-ing capacity of EDTA is used for the extraction of Pband Zn and there are no more available chelant anions toinduce the reduction of Fe(III)-oxides. This is illustratedin Fig. 3a and b presenting the distribution of divalentmetal cations, which form complexes with EDTA, in thefinal solutions of the tests carried out at various L/S ratios.In the case L/S = 2 ml g�1, the sum of Pb and Zn concen-trations in the aqueous solution is 98 mM, which is veryclose to the total EDTA concentration of 100 mM.

As seen in Fig. 2b, the removal of As has the maximumvalue, 92%, at the higher L/S = 20 ml g�1 ratio and isslightly lower, i.e. 87%, at L/S 10 ml g�1. Decrease of L/Sto 5 ml g�1 depresses arsenic extraction to 62%. Finallyoperation at the lower L/S ratio 2 ml g�1 reduces furtherarsenic removal down to the value of 15%.

Lead is the element whose extraction is less affected by thevariation of L/S ratios (Fig. 2c). At L/S 20 and 10 ml g�1 Pbremoval is close to 90%, at L/S 5 ml g�1 is 80% and only atthe lower ratio L/S 2 ml g�1 Pb extraction is reduced to 62%.This is due to the fact that EDTA ligand has a higher affinityfor Pb than for Zn or Fe(II), as suggested by the relative sta-bility of the corresponding 1-1 metal chelates, whose log sta-bility constants (logKS) are equal to 19.7, 18.0 and 16.0 forPb-, Zn- and Fe(II)-EDTA, respectively (at 25 �C andI = 0; Gustafsson, 2003). As a consequence, when theamount of EDTA per gram of soil decreases at the lowerL/S ratios, Pb is bound preferentially to the available EDTAto the expense of other competing elements.

Inoculated

Control0

20

40

60

80

100

% F

e in

sol

utio

n

(a) Fe

Control

Inoculated

0

20

40

60

80

100

% A

s in

sol

utio

n

(b) As

0

20

40

60

80

100

0 20 40 60t, days

0 20 40 60t, days

% P

b in

sol

utio

n

(c) Pb

Control

Inoculated

0

20

40

60

80

100

% Z

n in

sol

utio

n

(d) Zn

Legend

L/S EDTA/(Fe+Pb+Zn) L/S EDTA/(Fe+Pb+Zn)

A 2/1 ml g-1 0.22 mol mol-1 A 10/1 ml g-1 1.1 mol mol-1

B " " B " "A 5/1 ml g-1 0.55 mol mol-1 A 20/1 ml g-1 2.2 mol mol-1

B " " B " "

A: Abiotic (control), B: Biological (inoculated with D. palmitatis)

Fig. 2. The effect of L/S ratio on metals extraction. Tests carried out at 30 �C.

0

20

40

60

80

100

120

L/S=2 L/S=5 L/S=10 L/S=20

Other

Ca

Fe

Zn

Pb

0

20

40

60

80

100

120

L/S=2 L/S=5 L/S=10 L/S=20

Div

alen

t cat

ions

in s

olut

ion

mM

Fig. 3. Distribution of Pb, Zn, Fe and Ca in the final solutions for the tests carried out at various liquid to solid ratios. (a) Abiotic tests, (b) Biologicaltests.

K. Vaxevanidou et al. / Chemosphere 70 (2008) 1329–1337 1335

The extraction of zinc during the abiotic tests variesbetween 75 and 80% at the L/S ratios 20, 10 and 5 ml g�1

(Fig. 2d) and is reduced to 40% at the lowest L/S ratio2 ml g�1. In the presence of D. palmitatis, the extractionof Zn is adversely affected by the microbial generation of

Fe(II). This effect was observed in the previous experimen-tal tests and is also apparent in this series.

The pH and redox potential were measured in the aque-ous solutions after the completion of the tests. The pH val-ues were found to increase from the value of 7.94 to 8.29,

1336 K. Vaxevanidou et al. / Chemosphere 70 (2008) 1329–1337

8.35 and 8.80 as the L/S ratio was reduced from 20 ml g�1

to 10, 5 and 2 ml g�1, respectively. It should be remindedhowever that the treatment was carried out under aN2:CO2 (80:20) atmosphere, which contributes in main-taining the pH close to neutral values. When the serum bot-tles were opened, there was no more CO2 atmosphere andsolution pHs were shifted to more alkaline values, bufferedmainly by the CaCO3 content of treated soils. The redoxpotentials, Eh, were close to zero, i.e. 12 mV, �7 mV and�6 mV, in the final solutions of the biological tests thatwere carried out at L/S 20, 10 and 5 ml g�1, in accordancewith the observed reductive dissolution of Fe. On the con-trary, the redox potential of the final solution at L/S2 ml g�1 was oxidic 212 mV, confirming the absence ofany appreciable iron reducing microbial activity in this test.

The results of theses tests indicate that the optimum dos-age of EDTA for this type of combined biological andchemical treatment is an amount of the chelant stoichio-metrically equivalent with the sum of the three target met-als Fe, Pb and Zn. Applying this dosage As, Pb and Znextraction was 87, 90 and 50%, respectively. Doubling thedose of EDTA had no effect on Pb and Zn extractionand increased only marginally the removal of As from 87to 92%. Several other studies investigating the effect ofEDTA dosage on the removal of contaminants result insimilar conclusions. Di Palma and Mecozzi (2007) havestudied the removal of Cu, Pb and Zn from contaminatedsediments and found that increasing the dose of EDTAdoes not improve the removal of contaminants but resultsin the increased dissolution of competing cations, such asCa and Fe, from sediment major components. Similarly,Finzgar and Lestan (2007) pointed out that above a certainlevel higher dosages of EDTA do not result in appreciablegain in Pb and Zn removal.

Another important aspect in this type of combined treat-ment is the fate of EDTA and its eventual biodegradation bythe microorganisms which are used in this specific remedia-tion scheme. Although not easily biodegradable, EDTA hasbeen reported to decompose under the action of several pureor mixed microbial populations (Thomas et al., 1998). Toevaluate this effect, EDTA concentration was measured inall final solutions after completion of the tests. The concen-tration was found to range between 90 and 80 mM, indicat-ing a loss of 10–20 mM. This loss cannot be attributedexclusively to the action of D. palmitatis, because it has beenalso observed in the abiotic tests. Probable reasons for thisloss may be adsorption of EDTA on the surface of soil par-ticles or partial photocatalytic decomposition during theexecution of the tests, due to the known photosensitivityof iron chelates (Lambert et al., 1963).

4. Conclusions

A treatment scheme combining the action of the ironreducing microorganism D. palmitatis with the chelatingstrength of EDTA was investigated in the framework ofthis study. This treatment resulted in high extraction rates

both for the cationic metal contaminants and for the met-alloid arsenic.

More specifically, the removal of As was greatlyenhanced by the bacterial activity with a final extractionup to 90% whereas only 35% was obtained when purechemical treatment was applied. The major part of Pb,approximately 85%, was removed from soil with purelychemical mechanisms, i.e. via the chelating action ofEDTA. However, the biological activity was found toimprove by 10% the extraction of Pb, increasing the finalremoval up to 95%. The removal of Zn was limited to50% in the presence of bacteria, due probably to thesequestration of Zn in low solubility compounds with bio-genic Fe(II). Under abiotic conditions the extraction of Znwas almost 80%. A sequential treatment, consisting of apure chemical step followed by the addition of bacteria ata second stage could be possibly applied to avoid this effect.

Acknowledgments

This work received the financial support of theEuropean Commission in the framework of METALBIO-REDUCTION project (Contract No. EVK1-CT-1999-00033). Part of the research was conducted with financialsupport by the European Social Fund (75%) and NationalResources (25%) – (EPEAEK II) – PYTHAGORAS.

References

Alam, M., Tokunaga, S., Maekawa, T., 2001. Extraction of arsenic in asynthetic arsenic-contaminated soil using phosphate. Chemosphere 43,1035–1041.

Bradford, M.M., 1976. A rapid and sensitive method for the quantifica-tion of microgram quantities of proteins utilizing the principle ofprotein dye binding. Anal. Biochem. 72, 248–254.

Chatain, V., Bayard, R., Sanchez, F., Moszkowicz, P., Gourdon, R., 2005.Effect of indigenous bacterial activity on arsenic mobilization underanaerobic conditions. Environ. Int. 31, 221–226.

Coates, J.D., Lonergan, D.J., Lovley, D.R., 1995. Desulfuromonas

palmitatis sp. Nov., a marine dissimilatory Fe(III) reducer that canoxidize long-chain fatty acids. Arch. Microbiol. 164, 406–413.

Cooper, D.C., Picardal, F., Rivera, J., Talbot, C., 2000. Zinc immobili-zation and magnetite formation via ferric oxide reduction by Shewa-

nella putrefaciens 200. Environ. Sci. Technol. 34, 100–106.Cummings, D.E., Caccavo Jr., F., Fendorf, S., Rosenzweig, R.F., 1999.

Arsenic mobilization by the dissimilatory Fe(III)-reducing bacteriumShewanella alga BrY. Environ. Sci. Technol. 33, 723–729.

Di Palma, L., Mecozzi, R., 2007. Heavy metals mobilization from harboursediments using EDTA and citric acid as chelating agents. J. Hazard.Mater. 147, 768–775.

Dong, H., Fredrickson, J.K., Kennedy, D.W., Zachara, J.M., Kukk-adapu, R.K., Onstott, T.C., 2000. Mineral transformation associatedwith the microbial reduction of magnetite. Chem. Geol. 169, 299–318.

Finzgar, N., Lestan, D., 2007. Multi-step leaching of Pb and Zncontaminated soils with EDTA. Chemosphere 66, 824–832.

Fredrikson, J.K., Zachara, J.M., Kennedy, D.W., Dong, H., Onstott,T.C., Hinman, N.W., Li, S.-M., 1998. Biogenic iron mineralizationaccompanying the dissimilatory reduction of hydrous ferric oxide by agroundwater bacterium. Geochim. Cosmochim. Acta 62, 3239–3257.

Gustafsson, J.P., 2003. Visual MINTEQ, A geochemical AssessmentModel for Environmental Systems. www.lwr.kth.se/English/OurSoft-ware/vminteq.

K. Vaxevanidou et al. / Chemosphere 70 (2008) 1329–1337 1337

Harvey, A.E., Smart, J.A., Amis, E.S., 1955. Simultaneous spectropho-tometric determination of Fe(II) and total Fe with 1-10-phenanthro-line. Anal. Chem. 27, 26–29.

Ignatiadis, I., Battaglia-Brunet, F., 2005. Applicability of anaerobicbacterial leaching as remediation technique for As-contaminated soils:batch, column and pilot experiments and economic assessment. In:ConSoil 2005, Uhlmann, O., Annokkee, G.J., Arendt, F. (Eds.),Proceeding of the 9th International FZK/TNO Conference on Soil-Water Systems. Bordeau, October 2005, pp. 1612–1621.

Jackson, B.P., Miller, W.P., 2000. Effectiveness of phosphate andhydroxide for desorption of arsenic and selenium species from ironoxides. Soil Sci. Soc. Am. 64, 1616–1622.

Jang, M., Hwang, J.S., Choi, S.I., Park, J.K., 2005. Remediation of arsenic-contaminated soils and washing effluents. Chemosphere 60, 344–354.

Lambert, J.L., Godsey, C.E., Seittz, L.M., 1963. Preparation andphotodecomposition of the complex acid, hydrogen aquo ethylenedi-aminetetraacetoferrate(III). Inorg. Chem. 2, 127–129.

Manouchehri, N., Besancon, S., Bermond, A., 2006. Major and tracemetal extraction from soil by EDTA: Equilibrium and kinetic studies.Anal. Chim. Acta 559, 105–112.

Neale, C.N., Bricka, R.M., Chao, A.C., 1997. Evaluating acids andchelating agents for removing heavy metals from contaminated soils.Environ. Prog. 16, 274–280.

Nevin, K., Lovely, D., 2002. Mechanisms for Fe(III) oxide reduction insedimentary environments. Geomicrobiol. J. 19, 141–159.

Papassiopi, N., Tambouris, S., Kontopoulos, A., 1999. Removal of heavymetals from calcareous contaminated soils by EDTA leaching. WaterAir Soil Pollut. 109, 1–15.

Papassiopi, N., Vaxevanidou, K., Paspaliaris, I., 2003. Investigating theuse of iron reducing bacteria, as an option for the remediation of Ascontaminated soils. Water Air Soil Pollut.: Focus 3, 81–90.

Roden, E., Urrutia, M., 2002. Influence of biogenic Fe(II) onbacterial crystalline Fe(III) oxide reduction. Geomicrobiol. J. 19,209–251.

Sun, B., Zhao, F.J., Lombi, E., McGrath, S.P., 2001. Leaching of heavymetals from contaminated soils using EDTA. Environ. Pollut. 113,111–120.

Tampouris, S., Papassiopi, N., Paspaliaris, I., 2001. Removal of contam-inant metals from fine grained soils, using agglomeration, chloridesolutions and pile leaching techniques. J. Hazard. Mater. B84, 297–319.

Theodoratos, P., Papassiopi, N., Georgoudis, T., Kontopoulos, A., 2000.Selective removal of lead from calcareous polluted soils using the Ca-EDTA salt. Water Air Soil Pollut. 122, 351–368.

Thomas, R., Lawlor, K., Bailey, M., Macaskie, L., 1998. Biodegradationof metal-EDTA complexes by an enriched microbial population. Appl.Environ. Microbiol. 64, 1319–1322.

Tokunaga, S., Hakuta, T., 2001. Acid washing and stabilization of anartificial arsenic-contaminated soil. Chemosphere 46, 31–38.

Urrutia, M., Roden, E., Zachara, J., 1999. Influence of aqueous and solid-phase Fe(II) complexants on microbial reduction of crystalline Fe(III)oxides. Environ. Sci. Technol. 33, 4022–4028.

Wallace, R.M., Hinton, J.F., 1970. Determination of small amountsof aminecarboxylic acid complexing agents in the presence ofexcess trivalent lanthanides and actinides. Anal. Chim. Acta 51,540–543.

Wasay, S.A., Barrington, S.F., Tokunaga, S., 1998. Remediation of soilspolluted by heavy metals using salts of organic acids and chelatingagents. Environ. Technol. 19, 369–380.

Xenidis, A., Papassiopi, N., Komnitsas, K., 2003. Carbonate rich miningtailings in Lavrion: Risk assessment and proposed rehabilitationtechniques. Adv. Environ. Res. 7, 207–222.