Chemosphere 83 (2011) 997–1004

Contents lists available at ScienceDirect

Chemosphere

journal homepage: www.elsevier .com/locate /chemosphere

Kinetic characterizing of soil trace metal availability using Soil/EDTA/Chelex mixture

Nastaran Manouchehri ⇑, Stéphane Besançon, Alain BermondAgroParisTech., Laboratoire de Chimie Analytique, 16 rue Claude Bernard, 75231 Paris, cedex 05, France

a r t i c l e i n f o a b s t r a c t

Article history:Received 1 October 2010Received in revised form 2 February 2011Accepted 3 February 2011Available online 5 March 2011

Keywords:SoilTrace metalEDTAKineticsChelex-100

0045-6535/$ - see front matter � 2011 Elsevier Ltd. Adoi:10.1016/j.chemosphere.2011.02.010

⇑ Corresponding author. Tel.: +33 (1) 44 08 18 61;E-mail addresses: manouche@agroparistech.fr, n

Manouchehri).1 According to the glossary of terms used in toxicok

tions 2003), speciation in chemistry is defined as thamongst defined chemical species in a system; source:

The kinetic aspects are not usually tackled when mimicking the soil trace metal mobilization. In thiswork, a simple procedure is developed for measuring the kinetics of Pb, Cu and Cd transfer from the soilsolid phase towards a resin sink. A ternary system of Soil/EDTA/Chelex was employed for mimicking themetal transfer from two agricultural soil samples into the Chelex. Two different kinetic regimes (P1 andP2) were observed. The kinetic profile of Pb was distinctly different from those of Cd and Cu. Basing onkinetic principles, two kinetic models were proposed for estimating the apparent rate constants of leach-ing and removal processes in, respectively, two binary mixtures of soil/EDTA and EDTA extracts/Chelex.Contrary to Pb, solid phase pools of Cd and Cu exchanged with the solution on short time scales. Thekinetic rate of desorption occurred in following order: Pb < Cu 6 Cd. Comparing the kinetics of binary sys-tems, the desorption process was hypothesized to control the metal transfer. The parameters related tothe desorption process were extended to the kinetic regimes of Soil/EDTA/Chelex and revealed that thesoil to Chelex transfer of Pb is under kinetics control whereas the bulk concentration (affected by the sizeof particulate reservoir) is likely to control the Cd transfer. Cu presents an intermediate case.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

Regardless of their origins and the reasons of their concentra-tion increase in soils, trace metals (TMs) are liable to contaminatefood chains by migrating towards groundwater or accumulation inplants. The trace metal bioavailability and the related risk are theresults of both ‘‘capacity’’ factor i.e. the soil solid phase aptitudeto supply the metal to the soil solution and the ‘‘intensity’’ i.e.the concentration in pore solution (Barber, 1995). ‘‘Capacity’’ isstrongly influenced by TM’s reactivity in soil solid phase and theirlocalization in different soil components, usually called ‘‘specia-tion’’1 and ‘‘Intensity’’ is characterized by chemical speciation of me-tal in the solution. This area is subject to active researches in the soilto attempt to characterize the TMs reactivity in soil solid phase andto determine the part of metal in solution, which is available to biota.Chemical extractive schemes have been largely proposed for assess-ing the ‘‘reactive’’ or ‘‘labile’’ pools of metal in the soil solid phase(Young et al., 2006). Single extractive schemes provide a simple ap-proach for estimating the operational mobile pool using chemical re-agent such as EDTA or DTPA (Barona and Romero, 1996; Tippinget al., 2003; Mocko and Waclawek, 2004; Feng et al., 2005; Younget al., 2006; Manouchehri and Bermond, 2009). With regards to

ll rights reserved.

fax: +33 (1) 44 08 72 51.manouchehri@gmail.com (N.

inetics (IUPAC Recommenda-e distribution of an elementPAC, 2004, 76, 1033.

the pore solution, different approaches have been proposed (Zhangand Young, 2006) like as ion-selective electrodes (ISE), Donnanmembrane technique and resin exchange method (Tejowulan andHendershot, 1998). The latter presents a simple technique for iden-tifying the metal species in natural environmental systems howeverrelatively few reports in soil media.

In most of these methods, the metal concentration is measuredat equilibrium whereas natural systems are generally subject tochanging conditions and are practically never at the equilibrium.Indeed, for these dynamic systems, the metal availability may bebelieved to be controlled by kinetic factors (Errecalde et al.,1998; Ma et al., 1999; Fortin and Campbell, 2000; Hasseler andWilkinson, 2003; Slaveykova et al., 2003).

Some kinetically based approaches were then proposed. The ki-netic fractionation is an extractive approach using a first-ordertwo-reaction model dividing the extracted metal by a chemicalagent into ‘‘labile’’ and ‘‘slowly labile’’ pools (Bermond et al.,1998). Labile pools of Cd, estimated by this approach in a soil/EDTAsystem, predicted bioavailable fractions in 11 un-polluted soilsamples (Bermond et al., 2005). DGT (diffusive gradients in thin-films) technique (Zhang et al., 2001) measures the diffusive fluxof labile metal and the kinetic exchange between solid phase andsolution utilizing a simple device that contains a gel resin layer(Chelex for metals) as a sink of potentially labile pools. It has beenrevealed to be a good surrogate of plant uptake (Davison et al.,2000; Ernstberger et al., 2005; Nolan et al., 2005).

In certain kinetic approaches, there is an effort to provide a use-ful tool in order to identify whether metal availability is under the

998 N. Manouchehri et al. / Chemosphere 83 (2011) 997–1004

kinetic control of the soil solid phase or soil solution. This knowl-edge is critical for interpreting the results provided by differentpre-described speciation approaches (Manouchehri et al., 2011).

Owning to the complexity of physical and chemical interactionsin the soil, the modelling of TMs dynamic under natural environ-mental conditions seems still to be out of the capacity of the mostrecent methodologies. One way to try to avoid the complicatedcomplexation, desorption and readsorption of metals is to simplifythe metal desorption from the soil and its removal from the solu-tion. For this purpose, the desorption process could be consideredwithin a single extractive scheme using a single chemical complex-ing agent and the removal process may be modelled using a sinkmimicking the metal uptake by the root membrane handled by achelating resin (such as DGT). In this way, a batch experiment ofsoil/chelant/sink is proposed in the present work as a simple modelfor mimicking the kinetics of soil/plant transfer. The kinetic re-gimes of Pb, Cu and Cd are monitored in a batch system of soil,EDTA (as the chelant) and Chelex (as the uptake sink). Contraryto DGT technique (based solely on diffusive transport), both mech-anisms of metal transport from the bulk–soil to plant roots (massflow and diffusion) are considered in this approach. The kineticsof desorption and removal processes will be also assessed in twoseparate binary systems to determine the associated kineticparameters. The kinetic profiles of Pb, Cu and Cd are thereby as-sessed in three systems:

� Soil/EDTA/Chelex mixture for the metal transfer from the solidphase to Chelex as a simple model for mimicking the soil/planttransfer.� Soil/EDTA mixture, as the single extractive scheme, for simulat-

ing the kinetics of the desorption process.� EDTA extract/Chelex mixture to assess the metal dissociation in

soil solution and its removal by Chelex used here as a sink formetal uptake.

The kinetic trends of desorption/removal processes will be com-pared and extended to the kinetic regimes of each TM in the ter-nary system of Soil/EDTA/Chelex. The objective is to evaluate thelimiting and controlling effect of the kinetics in the transfer of eachTM from the soil to the Chelex.

2. Materials and methods

2.1. Soil samples

The experiments were carried out for two soil samples among aseries of 15 samples for which the origin of natural contents oftrace metals results from geological evolution. These soil samplesbelong to different soil series developed in various parent materialsunder similar climatic conditions. They were collected on 0.3 m2

areas with a spade from the 0 to 25 cm surface-ploughed layer infields at 15 sites across the southern part of the Yonne district, Bur-gundy, France. Two non-calcareous samples were chosen to avoidthe co-extraction effect of calcium cations due to the non-selectivenature of EDTA (Manouchehri et al., 2006). The co-extraction of Caresults in important variation of EDTA concentration with timerendering complex the kinetic simulation (cf. Paragraph 3.2). Somesoil characteristics are briefly described in Supplementary data(SD1).

2.2. Kinetic protocols

2.2.1. Soil/EDTA/Chelex mixtureThe kinetic study of metal transfer in Soil/EDTA/Chelex mix-

tures was made with an initial volume of 100 mL of 0.002 M EDTA

solutions (pH = 6.5) in contact with 3 g of resin and 10 g of soilsample in a polyethylene flask. The mixture was stirred using arotatory stirrer during the reaction time (300 min). At selectedtime intervals, an aliquot of 0.5 mL of the mixture was removedvia a syringe and immediately filtered through a Millipore syringefilter membrane. The filtrates were kept in the polyethylene micro-tubes at 4 �C until analysis of elements.

2.2.2. Soil/EDTA mixtureThe kinetic study of metal extraction by EDTA was performed

on a 10 g soil sample and an initial volume of 100 mL of extractingsolution (0.002 M EDTA at pH = 6.5), which corresponds to a soil/solution ratio of 1:10. The mixture was stirred using a rotatory stir-rer during the reaction time (24 h). At selected time intervals, analiquot of 0.5 mL of the mixture was removed via a syringe andimmediately filtered through a Millipore syringe filter membrane.The filtrates were kept in the polyethylene micro-tubes at 4 �C un-til analysis of major and trace elements.

2.2.3. EDTA extracts/Chelex mixtureThe kinetic study of the metal sorption on the resin from the

EDTA leachates, was made with an initial volume of 60 mL of thesoil extract solutions in contact with 2 g of Chelex-100 in its cal-cium form (Manouchehri and Bermond, 2006), which correspondsto a resin/soil extract ratio of 1:30). The mixture was stirred using arotatory stirrer during the reaction time. The kinetic sampling wasperformed as described in soil/EDTA mixture protocol.

2.3. Determination of extracted cations

Determination of trace metals (Cu, Cd and Pb) and major ele-ments (Al, Ca, Fe and Mg) was performed by Flame Atomic Absorp-tion Spectrometry (FAAS). In certain cases they were determined,by Electrothermal Atomic Absorption Spectrometry (ETAAS). A Hit-achi Z-5000 Polarized Zeeman Atomic Absorption Spectrophotom-eter was used equipped with an air-acetylene flame and a microsampling kit, which has permitted the analysis of all elements inkinetic experiments, where the volume of removed sample at eachtime was small for flame analysis. In this technique, a 80–100 lLaliquot were introduced for each measurement. The repeatabilityof the micro sampling technique has been verified in replicatedmeasurements for standard and soil solutions. The relative stan-dard deviations obtained were satisfactory and ranged from 3%to 5% (n = 6). The reagents used in all experiments were Merck ana-lytical quality. Water of high purity obtained from a Milliporeapparatus (water resistivity = 18 MX cm) was used. The polyethyl-ene tubes and the filtration equipment used in equilibrium exper-iments were cleaned in molar nitric acid and then rinsed with purewater.

2.4. Statistical analysis

The software Sigmaplot 5.0 was used for fitting the kineticexperimental data to the appropriate equation. It was also usedfor data smoothing. In all cases, regression and statistical testswere used to determine the goodness of the fit once all of the re-quired parameters were determined.

3. Results and discussion

The experimental results and the related discussions are di-vided into three parts corresponding to the kinetics of metal trans-fer in the ternary system of Soil/EDTA/Chelex and the kinetics ofdesorption and removal processes in two binary mixtures of soil/EDTA and EDTA extracts/Chelex.

Fig. 1. (a) Kinetic profiles of dissolved concentrations of Pb, Cu and Cd in Soil/EDTA/Chelex mixture for Grimault and Bierry; (b) typical kinetic profiles observed for TM’smonitoring Soil/EDTA/Chelex mixture, P1 for and P2 for Cu and Cd for both soil.

N. Manouchehri et al. / Chemosphere 83 (2011) 997–1004 999

3.1. Monitoring of Pb, Cu and Cd in Soil/EDTA/Chelex mixture

In this part, the kinetics of TM’s transfer from the soil towardsthe resin were studied in Soil/EDTA/Chelex mixture. The experi-ment was carried out in two replicates and the difference betweenthe two replicates did not exceed 5%. Two different kinetic patternshave been observed for Pb, Cu and Cd in both samples (Fig. 1a). Thekinetic profiles of dissolved concentration of Pb and Cu/Cd in Soil/EDTA/Chelex mixture are assigned to P1 and P2, respectively(Fig. 1b).

P1 is characterized by an increase of metal concentration in thesolution reaching a plateau after a given time and no appreciabledepletion in dissolved concentration is observed in the frame timeof the reaction. While in P2, the metal concentration attains a max-imal value after a few minutes and then the dissolved concentra-tion in solution is markedly depleted at higher reaction times.

The different kinetic trends observed in the proposed batch sys-tem is likely affected by the kinetics of metal desorption and re-moval processes. In order to understand how the kinetics of

resupply from the soil solid phase and those of metal-complex dis-sociation in the solution determine the kinetic profiles of dissolvedconcentrations in Soil/EDTA/Chelex system, we have attempted toassess the kinetic parameters associated to the desorption and re-moval processes in separated binary systems of soil/EDTA andEDTA extracts/Chelex.

3.2. Kinetics of metal desorption in soil/EDTA mixture

The extraction of soil trace metal in soil/EDTA system involvesthe following simplified reaction (charges omitted):

S�M þ EDTA)M-EDTAþ S ð1Þ

where S presents the different soil solid phase sites to which tracemetals are bound. The desorption rate of metal (M) is described asfollows:

dMdt¼ ksoil½M�½EDTA� ð2Þ

Fig. 2. (a) Kinetic profiles of Ca extraction in Grimault and Bierry samples; (b) kinetic profiles of EDTAfree in Grimault and Bierry samples.

1000 N. Manouchehri et al. / Chemosphere 83 (2011) 997–1004

EDTA is a non-specific reagent and major cations, such as iron,calcium, and magnesium, are also extracted according to Eq. (3)(Manouchehri and Bermond, 2009). At low concentration(0.002 mol L�1 EDTA in this work), the extractive reagent is nolonger in excess with respect to the total amounts of extractiblecations (Manouchehri et al., 2006).

MðCa;Fe;AlÞ � Sþ EDTA) MðCa;Fe;AlÞ � EDTAþ S ð3Þ

In this case, the leaching reaction (Eq. (1)) could not be consid-ered as a pseudo first-order reaction and the evolution of [EDTA]vs. time should be controlled. A preliminary study was thus per-formed before the kinetic simulation to determine the mass bal-ance of the reagent and its evolution during the reaction timetaking into consideration the extraction of major cations.

3.2.1. EDTA mass balance vs. timeThe monitoring of major cations (Ca, Fe, Mg, etc.) was per-

formed before the kinetic modelling of TM’s. Fig. 2 displays theconcentration of Ca vs. time. Calcium is readily released over thefirst minutes of the extraction and its concentration decreases athigher extraction time due to the slower extraction of iron (notshown), which compete with Ca, according to the following equi-librium (Eq. (4)).

S� Feþ Ca EDTA) Fe-EDTAþ S-Ca ð4Þ

In order to determine the concentration of EDTA at any time,the concentration of TM’s and major cations (Ca, Fe, Al and Mg)in leaching solution was considered. EDTA mass balance may bedefined according to Eq. (5) and EDTAfree is calculated. For bothsamples, the mass balance involved mainly iron and calcium.

½EDTA�total ¼ ½EDTA�free þX½M�TM;Ca;Fe ð5Þ

For both samples, the evolution of [EDTA]free vs. time (Fig. 2)is characterized by a fast decrease at the first minutes of theleaching reaction (about of 20% of initially introduced EDTA isconsumed in less than 10 min) followed by a slowly decreasefor Grimault sample (about 50% of initial EDTA) or a very slowdecrease for Bierry sample. The initial fast decrease may be re-lated to the very fast release of calcium while the slow decreaseis likely more related to the slow leaching of Fe. Conclusively,for non-calcareous soils, [EDTA]free may be considered quasi-constant all over the reaction time. Moreover, the stoichiometricratio of EDTAfree/TM varied between 80 for EDTA/Pb in Grimaultsample to 10 377 for EDTA/Cd in Bierry sample over a 300-minreaction period. EDTA could therefore be considered in excesswith respect to target TM’s in both samples. If a large excessor constant concentration of EDTA is maintained, the leaching

process could be considered as pseudo 1st-order reaction. Thisconclusion is employed for simulating the desorption rate ofTM’s in soil/EDTA system.

3.2.2. Kinetic simulation of TM’s desorptionThe feasibility of a pseudo 1st-order kinetic simulation of

desorption was verified using the classical rate equation, as pro-posed by Labanowski et al. (2008):

DMDt¼ M2 �M1

t2 � t1ð6Þ

where DM/Dt presents the average desorption rate per time unitand is computed from the ratio of desorbed mass evolution (M2–M1) to desorption time evolution (t2–t1) for two consecutivesampling.

Fig. 3a displays the metal desorption evolution with time for Pb,Cd and Cu in Bierry sample. The presented data are mean values oftwo replicates. Differences between replicates did not differ bymore than 10%. It is to notice that the extractable amounts werenear to analytical detection limit in certain cases and the experi-mental data were thus smoothed before computing DM/Dt ratiosfor each lag time. The curves for all three metals seem similar inshape with classical kinetic trends: a fast release at short reactiontime (t < 1 h) followed by a slower release at higher reaction timereaching finally a plateau (1 h < t < 5 h).

Fig. 3b displays the evolution of desorption rate (DM/Dt) withtime over a 300-min of EDTA/soil reaction for Bierry and Grimaultsamples. Two kinetically distinguishable parts could be identifiedcorresponding to two metal pools characterized by two differentdesorption rate: high desorption rates in the beginning of leaching(t < 1 h) followed by a decrease in higher reaction time(1 h < t < 5 h).

Considering the overall two-segments shape of these desorptionrate curves, it could be likely that the extraction of the three cationsobeys a pseudo 1st-order two-rate reaction model as proposed byBermond et al. (1998) and applied in several other works (Songand Greenway, 2004; Wasay et al., 2007; Labanowski et al.,2008). This kinetic model allows dividing the total extractibleamount of TM into labile and less labile pools under the operationalconditions of the leaching experience. The labile pool could be rep-resentative of the potential availability of the target TM. The theo-retical principles of this kinetic model are briefly described below.

For the leaching reaction of a given TM from soil, if it is assumedan excess of reagent and/or its concentration constant vs. time (asverified by EDTA mass balance), a pseudo first-order reaction willbe involved. The desorption rate of the cation M vs. time isdescribed by the following equation:

Fig. 3. Kinetic profiles of Cd, Cu and Pb in Bierry/EDTA mixture and the evolution of desorption rate of Pb, Cd and Cu vs. time in soil/EDTA mixtures for Grimault and Bierrysamples.

N. Manouchehri et al. / Chemosphere 83 (2011) 997–1004 1001

d½M�dt¼ �k1½M�soil ð7Þ

where [M], monitored by the analytical technique, is the extractedmetal and [M]soil is the concentration of metal in the soil solidphase. The equation rate of the simultaneous desorption from labile(lab) and slowly labile (slab) sites will be defined as:

dM=dt ¼ Rklab;slab½M�lab;slab ð8Þ

At time t, the total metal leached out by EDTA corresponds, afterintegration, to the following equation:

MðtÞ ¼ Mlabð1� expð�klabtÞÞ þMslabð1� expð�kslabtÞÞ ð9Þ

where Mlab, Mslab, klab and kslab presents respectively the concentra-tions of metal in labile and slowly labile pools and the associatedrate constants. SigmaPlot 5.0 software was used to fit the

1002 N. Manouchehri et al. / Chemosphere 83 (2011) 997–1004

experimental data to the Eq. (9). An example of kinetic fits is pre-sented in Supplementary data (SD2).

The two-reaction model did give excellent fit to experimentalleaching curves with R2 values generally over 0.95. Other statis-tical parameters like as p values (0.01 < p < 0.05), Durbin Watson(to detect the presence of autocorrelation in the residuals) andt-test were also considered to verify the quality of fittings. Theapplication of Eq. (9) to the experimental data of leachingprocess allows estimating Mlab, Mslab, klab and kslab (Table 1 andsupplementary data; SD3).

Except for Cd in Grimault, for both samples, the labile fractionscorrespond to about 50–60% of the total extractible concentrationsat equilibrium. The high labile fraction of Cd in Grimault could beresulted from the high value of cation exchange capacity (CEC) re-ported for this sample (19.5 for Grimault vs. 9.2 for Bierry). Menchet al. (1997) have already reported the significant influence of CECon Cd extractability. EDTA extracted about 65% of Cd and 60% of Cuand Pb in both samples.

On Table 1 are displayed the labile rate constants estimatedfrom the fitting of leaching experimental data to Eq. (9) andthe characteristic response times (Tc). Tc defines the times neededfor leaching process to attain 63% of its equilibrium values; it iscommonly used as a kinetic proxy of pollutant uptake or desorp-tion (Jannasch et al., 1988; Harper et al., 1998; Lehto et al.,2006). Tc and the apparent rate constants of TM desorption (klab)in soil/EDTA mixtures allow to compare the three TM’s in termsof their lability.

For both samples, the following order occurred in terms ofdesorption rate: Cd > Cu > Pb. Tc ranged from 35 min for Cd to220 min for Pb. Whatever the sample, the most short Tc were ob-served for Cd contrary to Pb presenting relatively high Tc, consis-tent with the desorption rate constants. Whatever the sample, Cdwith high values of klab and short Tc is the most labile metalwhereas Pb is the less kinetically mobile reaching slowly the qua-si-equilibrium concentration in desorption process (high Tc). Cupresents an intermediate case compared to Cd and Pb. The kslab

characterizing the less or slowly labile pools have been revealedmuch lower than the klab values. kslab ranged from 0.008 (min–1)for Cd to 0.001 (min–1) for Pb.

The order of lability observed in these experimental condi-tions is in agreement with the mobility schemes frequentlyreported in the literature (Tejowulan and Hendershot, 1998;Varrault, 2001; Labanowski et al., 2008). Experiments that inves-tigated the two-rate pseudo-first order kinetics of TM desorptionfrom polluted agricultural soils using EDTA 0.05 mol L�1

(Labanowski et al., 2008) provided apparent rate constant valuesof 0.076, 0.118, 0.155 min�1 for Pb, Cu and Cd, in very goodagreement, respectively, with the values of 0.078, 0.12 and0.18 min�1 we found for the same TM at lower concentrationof extractive agent (0.002 mol L�1). Wasay et al. (2007) found avalue of 0.087 (min�1) for the removal rate constant of Pb froma naturally contaminated soil using EDTA 0.05 mol L�1 and thetwo-rate reaction model. However, it is difficult to compare thedesorption rate constants values obtained in different work dueto several influencing elements like as different experimentalconditions, pollution source and soil physicochemical charac-teristics.

Table 1Desorption rate constants for Pb, Cu and Cd for Bierry and Grimault in EDTA mixture.

Sample Pb klab (min�1), Tc

(min)Cu klab (min�1), Tc

(min)Cd klab (min�1), Tc

(min)

Bierry 0.078 (130) 0.12 (90) 0.18 (35)Grimault 0.16 (220) 0.41 (110) 0.81 (40)

3.3. Kinetics of metal transfer in EDTA extracts/Chelex mixture

This second part is devoted to the kinetics of metal transferfrom the soil-EDTA extracts to Chelex-100. The feasibility of metalremoval by Chelex-100 from EDTA soil extracts was previouslyinvestigated (Manouchehri et al., 2006).

The kinetic exchange of metal between EDTA extract and resinmay be simulated for an equilibrium process including a pair offorward and reverse reactions between two species as follows:

MEDTA$kr

k�r

Mres ð10Þ

The reaction rate expression for the above reaction can be ex-pressed as:

R ¼ kr ½M � EDTA� � k�r½M � resin� ð11Þ

where kr is the rate coefficient for the reaction that consumes M-EDTA and k�r is the rate coefficient for backwards reaction. The gen-eralized kinetic equation for the metal uptake by the resin can beexpressed as:

½MEDTA�t ¼ aþ b exp½�ðkr þ k�rÞt� ð12Þ

where a = [MEDTA] (k�r/kr + k�r), b = [MEDTA] (kr/kr + k�r).Both kinetic constants are related to the equilibrium coefficient

of the reaction (K) by the following relationship (set R = 0 inbalance):

K ¼ kr=k�r ð13Þ

SigmaPlot 5.0 software was used to fit the Eq. (12) to the exper-imentally measured of MEDTA during the reaction time (300 min)between EDTA extracts and resin. The statistical parameters usedto access the quality of the fitting process were the standard errorof estimate (S.E.), the coefficient of correlation (r) and the P value.There was a good agreement between the experimental and fitteddata for Cd, Cu and Pb in both samples (example of Cd in Supple-mentary data).

The simulation of kinetic exchange in EDTA extracts/Chelex sys-tem using the Eq. (12) allowed calculating the rate constant (kr) ofmetal uptake by resin from the soil extracts for Cd, Cu and Pb inboth samples (Supplementary data).

In terms of kinetics, the dissociation of Pb, Cd and Cu complexesin EDTA extract occurs at the same apparent rate constants rangingfrom 1.3 � 10�3 to 8.9 � 10�3 (min�1).

Considering the similar dissociation rate constants of metal-complex dissociation in soil solution the desorption step seemsto control the soil to Chelex transfer of metal. In the last part of thiswork, the kinetic informations previously obtained will be ex-tended to the kinetic regimes of metal transfer in Soil/EDTA/Chelexmixture in order to understand how the desorption step couldaffect the metal transfer from the solid phase to the Chelex.

3.4. Interpretation of kinetic regimes (P1 and P2)

As mentioned in part 3.1, different kinetic profiles were ob-served for Pb, Cu and Cd. P1 appears like a classical kinetic patternof the metal desorption from the soil solid phase and P2 looks likethe kinetic of uptake by Chelex (Fig. 1).

Indeed, P1 and P2 regimes observed in Soil/EDTA/Chelex mix-ture are likely related to the degree of the soil solid phase contribu-tion to resupply the soil solution. In order to investigate how theresupply from the soil solid phase could determine the metaltransfer and consequently affect the dissolved concentration pro-files (P1 and P2), we experimentally determined the ratio (r) asfollows:

r ¼ Clab=Csoln ð14Þ

Fig. 4. r plotted against reaction time for Pb, Cu and Cd in Bierry and Grimault.

Table 2Solid–liquid distribution coefficient (Kd) of Pb, Cu and Cd in soil-EDTA system for bothsoils.

Sample Kd Pb Kd Cu Kd Cd

Bierry 4.98 4.55 1.61Grimault 3.11 5.66 2.01

N. Manouchehri et al. / Chemosphere 83 (2011) 997–1004 1003

Clab and Csoln present, respectively, the labile-EDTA measured con-centration (calculated from the two-rate reaction model) and thedissolved concentration of metal in Soil/EDTA/Chelex mixtures. ris used as an indicator for the extent of metal depletion in thesolution in Soil/EDTA/Chelex system. The concentration of metalin labile pool was used as the reference value of soil resupply inthe lack of Chelex. The evolution of r with time was plotted(Fig. 4). The evolution of r during the first 150 min of reactionwas enlarged to show its changes within the time-scale where thelabile pool is likely to be desorbed from the solid phase. As seen,for Cd and Cu, the general trend is a sharp increase in r at shorttimes followed by a decrease after the peak whereas for Pb thechange in r was distinctly different; it presents a slow increasereaching a plateau at higher reaction times.

However, the change in r differs not only in shape but r-valuesvaried between a minimum of 0.17 for Pb and a maximum of 2.18for Cd in Grimault sample. The lowest r-values were observed forPb, Csoln was generally lower than Clab all over the reaction timeindicating a low replenishment from the soil solid phase of bothsamples. Contrary to Pb, nearly equal Clab to Csoln was found forCd at the beginning of the reaction. Cu presents the intermediatecase where r-values are lower than those of Cd with nearly similarprofiles.

It is well recognized that the capacity of soil is controlled by thesize of metal reservoir and the rate of resupply (Ernstberger et al.,2002). Basing on soil capacity and kinetic parameters, each metalsuggests a different resupply regime, detailed below.

3.4.1. Capacity and Kinetics of resupplyThe solid–liquid distribution coefficient (Kd) is an operationally

defined parameter (Degryse et al., 2009), frequently used for esti-mating both ‘‘capacity’’ and ‘‘intensity’’ factors to predict the soilmetal availability. Kd is defined as the concentration in soil solidphase (Cs) to that in the soil solution (CEDTA in this work). In thecontext of this work, the solid–liquid partitioning coefficient is cal-culated as follows:

Kd ¼ Cs=CEDTA ð15Þ

Kd values and the kinetic parameters related to the desorption pro-cess (Table 2) are used to interpret the changes in r and conse-quently the different kinetic regimes (P1 and P2) for the threecations.

3.4.1.1. Cd. Cd is characterized by fast desorption kinetics (short Tc

and high klab), small reservoir of metal in the solid phase (Kd; cf. Ta-ble 2) and high r-values (over the first 1 h of the reaction time). Theinitial peak in r (Fig. 4) is associated with the fast kinetics ofdesorption; the soil solution is then readily replenished and Csoln

is nearly equal to Clab (Bierry) or larger than Clab (Grimault). After1 h of reaction, r declines due to the small size of available reser-voir. This case is characterized by kinetic profile P2 (cf. Fig. 1b) insoil/EDTA/mixture. Kinetics of resupply is presumably not limitingin Cd transfer to Chelex and the depletion in Csoln is associated withthe limited capacity of soil reservoir to replenish the solution.

3.4.1.2. Pb. Pb is characterized by slow desorption kinetics(Tc > 130, klab < 0.16), larger reservoir of metal in solid phase (Kd;cf. Table 2) and lower r-value compared to Cd. The ratio (r) risesto attain a plateau at its maximum value i.e. 0.2 in Grimault and0.45 in Bierry (Fig. 4). Comparing these values with those of Cdin both soils, it could be considered that Csoln is markedly depletedwith respect to the potential concentration in labile pool (Clab) dueto the slow kinetic of desorption. However, a low level of resupplyis sustained due to the large reservoir of Pb in solid phase and no

1004 N. Manouchehri et al. / Chemosphere 83 (2011) 997–1004

appreciable decline in dissolved concentration (Csoln) is observedover 300-min of the reaction time (Fig. 4). The peak of thedepletion may occur in the extended times where the solid phaseconcentration is more likely to become depleted. The case of Pbis associated to the kinetic profile P1 (cf. Fig. 1b) in Soil/EDTA/Che-lex mixture. Hence, Pb transfer to Chelex seems to be controlled bykinetics of desorption.

3.4.1.3. Cu. Comparing with Pb and Cd, Cu presents an intermediatecase. The kinetic of desorption is faster than Pb and slower than Cd(Tc and klab; cf. Table 2). As Pb and Cd, an initial rise in r was ob-served to attain its maximum values ranging from 0.38 in Grimaultto 0.5 in Bierry over the first 1 h of reaction (Fig. 4). The Csoln is thendepleted with respect to the potential labile pool concentration(Clab) all over the reaction time. The ratio (r) declines after 1 h ofreaction despite the large size of solid phase reservoir arising thepossibility of low concentration of Cu in available pool of solidphase. Elsewhere, the kinetic of resupply is likely to be limiting be-cause Csoln attains never Clab in the frame time of the reaction. Thekinetics of resupply and the size of available metal in solid phaseare both presumably involved in the Cu transfer to Chelex.

4. Conclusion

The main feature of the developed approach in this work is thatit tackles with the kinetic aspects of soil TMs availability using asimple chemical extractive scheme in a batch system. In thisway, the kinetics of Pb, Cu and Cd transfer were mimicked in a ter-nary system of Soil/EDTA/Chelex. This simple approach involvesboth convectional and diffusional transport of metal from the soilto the sink model (Chelex). The dissolved concentrations of Pb,Cu and Cd in the ternary mixture revealed two different kinetic re-gimes (P1 for Pb and P2 for Cu and Cd). The kinetic parameters ofmetal desorption from the soil and those of metal-complex disso-ciation in solution, assessed in two binary systems of Soil/EDTAand Soil/EDTA/Chelex were extended to P1 and P2 profiles. Thiscomparative approach, which is not usually tackled when mimick-ing the soil trace metal mobilization, highlighted that transfer ofTMs to a sink like as Chelex is governed by the kinetic parametersrelated to TMs desorption processes and/or the available reservoirin the solid phase.

Conclusively, if the metal supply from the soil solid phase andits dissociation in solution are suggested to mimic the first stepof metal bioavailability process, the kinetic consideration relatedto these processes should be assessed in different techniques ofbioavailability prediction like as extractive schemes and DGT tech-nique. In the continuation of this work, some DGT applications areactually in process for evaluating the limiting role of kinetics in dif-ferent geochemical scenarios.

Appendix A. Supplementary material

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.chemosphere.2011.02.010.

References

Barber, S.A., 1995. Soil Nutrient Bioavailability: A Mechanistic Approach. John Wiley& Sons, New York.

Barona, A., Romero, F., 1996. Fractionation of lead in soils and its influence on theextractive cleaning with EDTA. Environ. Toxicol. 17, 63–70.

Bermond, A., Ghestem, J.P., et al., 1998. Kinetic approach to the chemical speciationof trace metals in soils. Analyst 123, 785–790.

Bermond, A., Varrault, G., et al., 2005. A kinetic approach to predict soil trace metalbioavailability: preliminary results. Plant Soil 275, 21–29.

Davison, W., Hooda, P., et al., 2000. DGT measured fluxes as surrogates for uptake ofmetals by plants. Adv. Environ. Res. 4, 550.

Degryse, F., Smolders, E., et al., 2009. Partitioning of metals (Cd, Co, Cu, Ni, Pb, Zn) insoils: concepts, methodologies, prediction and applications – a review. Eur. J.Soil Sci. 60, 590–612.

Ernstberger, H., Zhang, H., et al., 2002. Determination of chromium speciation innatural systems using DGT. Anal. Bioanal. Chem. 373 (8), 873–879.

Ernstberger, H., Zhang, H., et al., 2005. Desorption kinetics of Cd, Zn and Nimeasured in soils by DGT. Environ. Sci. Technol. 39, 1591–1597.

Errecalde, O., Seidl, M., et al., 1998. Influence of a low molecular weight metabolite(citrate) on the toxicity of cadmium and zinc to the unicellular green algaSelenastrum Capricornutum: an exception to the free ion model. Water Res. 32(2), 419.

Feng, M.H., Shan, X.Q., et al., 2005. A comparison of the rhizosphere-based methodwith DTPA, EDTA, CaCl2 and NaNO3 extraction methods for prediction ofbioavailability of metals in soil to barley. Environ. Pollut. 137, 231–240.

Fortin, C., Campbell, P.G., 2000. Silver uptake by the green alga CHLAMYDOMONASREINHARDTII in relation to chemical speciation: influence of chloride. Environ.Toxicol. Chem. 19, 2769–2778.

Harper, M.P., Davison, W., et al., 1998. Kinetics of metal exchange between solidsand solutions in sediments and soils interpreted from DGT measured fluxes.Geochim. Cosmochim. Acta 62, 2757–2770.

Hasseler, C.S., Wilkinson, K.J., 2003. Failure of the biotic ligand and free-ion activitymodels to explain zinc bioaccumulation by Chlorella kesslerii. Environ. Toxicol.Chem. 22, 620.

Jannasch, H.W., Honeyman, B.D., et al., 1988. Kinetics of trace element uptake bymarine particles. Geochim. Cosmochim. Acta 52 (2), 567–577.

Labanowski, J., Monna, F., et al., 2008. Kinetic extractions to assess mobilization ofZn, Pb, Cu, and Cd in a metal contaminated soil: EDTA vs. citrate. Environ. Pollut.152, 693–701.

Lehto, N.J., Davison, W., et al., 2006. Analysis of micro-nutrient behaviour in therhizosphere using a dynamic plant uptake model. Plant Soil 282, 227–238.

Ma, H., Kim, S.D., et al., 1999. Effect of kinetics of complexation by humic acid ontoxicity of copper to Ceriodaphina dubia. Environ. Toxicol. Chem. 18 (5), 828–837.

Manouchehri, N., Bermond, A., 2006. Study of trace metal partitioning between soil-EDTA extracts and Chelex-100 resin. Anal. Chim. Acta 557, 337–343.

Manouchehri, N., Bermond, A., 2009. EDTA in soil science. a review of itsapplication in soil trace metal investigations. Terrestrial Aquat. Environ.Toxicol. 3, 1–15.

Manouchehri, N., Besançon, S., et al., 2006. Major and trace metal extractionfrom soil by EDTA: equilibrium and kinetic studies. Anal. Chim. Acta 559,105–112.

Manouchehri, N., Besançon, S., Bermond, A., 2011. The interest of kineticconsiderations in soil heavy metal mobilization assessment: An overview.ECG Bulletin of RSC. Available from: <www.rsc.org/ecg>.

Mench, M., Baize, D., et al., 1997. Cadmium availability to wheat in five soil seriesfrom the Yonne district, Burgundy, France. Environ. Pollut. 95 (1), 93–103.

Mocko, A., Waclawek, W., 2004. Three-step extraction procedure for determinationof heavy metals availability to vegetables. Anal. Bioanal. Chem. 380 (5–6), 813–817.

Nolan, A.L., Zhang, H., et al., 2005. Prediction of zinc, cadmium, lead, and copperavailability to wheat in contaminated soils using chemical speciation, diffusivegradients in thin films, extraction, and isotopic dilution techniques. J. Environ.Qual. 34, 496–507.

Slaveykova, V.I., Wilkinson, K.J., et al., 2003. Role of fulvic acid on leadbioaccumulation by Chlorella kesslerii. Environ. Sci. Technol. 37, 1114–1121.

Song, Q.J., Greenway, G.M., 2004. A study of the elemental leachability and retentioncapability of compost. J. Environ. Monit. 6 (1), 31–37.

Tejowulan, R.S., Hendershot, W.H., 1998. Removal of trace metals fromcontaminated soils using EDTA incorporating resin trapping techniques.Environ. Pollut. 103, 135–142.

Tipping, E., Rieuwerts, J., et al., 2003. The solid-solution partitioning of heavy metals(Cu, Zn, Cd, Pb) in upland soils of England and Wales. Environ. Pollut. 125, 213–225.

Varrault, G. 2001. Etude cinétique de l’extraction des métaux traces des sols;application à l’évaluation de l’efficacité de techniques de réhabilitation in situ,Thèse Université Paris 7.

Wasay, S.A., Barrington, S.F., et al., 2007. Kinetics of heavy metal desorption fromthree soils using citric acid, tartaric acid, and EDTA. J. Environ. Eng. Sci. 6, 611–622.

Young, S.D., Zhang, H., et al., 2006. Characterizing the availability of metals incontaminated soils. I. The solid phase: sequential extraction and isotopicdilution. Soil Use Manage. 21, 450–458.

Zhang, H., Young, S.D., 2006. Characterizing the availability of metals incontaminated soils. II. The soil solution. Soil Use Manage. 21, 459–467.

Zhang, H., Zhao, F.J., et al., 2001. A new method to measure effective soil solutionconcentration predicts copper availability to plants. Environ. Sci. Technol. 35,2062–2067.