partitioning and speciation of chromium, copper, and arsenic in cca-contaminated soils: influence of...

Ž .The Science of the Total Environment 280 2001 239�255

Partitioning and speciation of chromium, copper, andarsenic in CCA-contaminated soils: influence of soil

composition

Cristina F. Balasoiu, Gerald J. Zagury�, Louise Deschenes´ ˆ´NSERC Industrial Chair in Site Remediation and Management, Chemical Engineering Department, Ecole Polytechnique de

Montreal, Montreal, QC, Canada H3C 3A7´

Received 1 December 2000; accepted 19 March 2001

Abstract

This study focused on the influence of soil composition and physicochemical characteristics on the retention andŽ .partitioning of Cu, Cr and As in nine chromated copper arsenate CCA artificially contaminated soils. A statistical

mixture design was used to set up the number of soils and their respective composition. Sequential extraction and� Ž . Ž .�modified solvent extraction were used to assess Cu and Cr partitioning and As speciation As III or As V . It was

Ž .found that peat had a strong influence on CEC 232 meq�100 g , on buffer capacity and on Cu and Cr retention,Ž .whereas kaolinite’s contribution to the CEC was minor 38 meq�100 g . Average metal retention in mineral soils was

Ž . Ž .low 58% for Cu and 23% for Cr but increased dramatically in highly organic soils 96% for Cu and 78% for Cr .Ž .However, both organic and mineral soils demonstrated a very high sorption of added As 71�81% . Levels of Cu and

Ž .Cr in a soluble or exchangeable form F1 in highly organic soils were very low, whereas the levels strongly bound toorganic matter were much higher. Conversely, in mineral soils, 47% of Cu and 18% of Cr were found in F1. As aresult, Cr and Cu in moderately and highly organic contaminated soils were present in less mobile and lessbioavailable forms, whereas in mineral soils, the labile fraction was higher. The modified method used for selectivedetermination of mineral As species in CCA-contaminated soils was found to be quantitative and reliable. Resultsrevealed that arsenic was principally in the pentavalent state. Nevertheless, in organic soils, arsenite was found in

Ž .significant proportions average value of 29% in highly organic soils . This indicates that some reduction of arsenateŽ .to arsenite occurred since the original species in CCA is As V . � 2001 Elsevier Science B.V. All rights reserved.

Keywords: Copper; Chromium; Arsenic; Chromated copper arsenate; CCA; Soil composition; Synthetic soils; Metal partitioning;Sequential extraction procedure; Solvent extraction; Arsenic speciation

� Corresponding author. Tel.: �1-514-340-4711; fax: �1-514-340-5913.Ž .E-mail address: [email protected] G.J. Zagury .

0048-9697�01�$ - see front matter � 2001 Elsevier Science B.V. All rights reserved.Ž .PII: S 0 0 4 8 - 9 6 9 7 0 1 0 0 8 3 3 - 6

( )C.F. Balasoiu et al. � The Science of the Total En�ironment 280 2001 239�255240

1. Introduction

In recent years, the use of organic wood pre-serving chemicals has been the subject of majorconcerns of regulatory authorities, includingdioxin contamination of the chlorinated phenolsand the potential carcinogenicity to humans of

Žcoal tars and creosote compounds Warner and.Solomon, 1990 . On the other hand, inorganic

waterborne preservatives, such as chromated cop-Ž .per arsenate CCA have proven their ability to

adequately protect wood from bacterial, fungaland insect attack. CCA was developed in 1933and has been widely used throughout the worldsince then. Depending on the relative proportionsof metals, there are three waterborne formula-tions designated as CCA types A, B and C. InNorth America, CCA�C is used almost exclu-

Ž .sively. This CCA type contains w�w 47.5% CrO ,3Ž .18.5% CuO, and 34% As O Cooper, 1994 . The2 5

CCA-treated wood is used in applications such asŽ .utility poles especially electric and telephone ,

wooden playground equipment, foundation wood,garden projects and marine piles.

Despite its broad use, there is increasingconcern about possible environmental contamina-tion from leaching losses of wood preservativesfrom CCA-treated wood. In fact, the active ingre-dients of CCA can leach from wood poles inservice to the detriment of soil organisms depend-ing on rainfalls, pH of aqueous solutions, and

Ž .wood species Cooper, 1994 . Other potentialsources of damage are spillage, deposition ofsludge and dripping from newly impregnated woodat timber treatment facilities.

To assess the environmental impact of con-taminated soils, knowledge of the total concentra-tion of a specific metal without considering itsspeciation is not sufficient. The physicochemicalproperties of soil can widely influence metal spe-ciation and, consequently, its mobility, bioavail-

Ž .ability and toxicity McLean and Bledsoe, 1992 .For example, knowledge of arsenic and chromiumspeciation is very important, since their toxicity isassociated with changes in the oxidation state.Metals may be distributed among many compo-

nents of soil solids and may be associated withŽthem in different ways ion exchange, adsorption,

precipitation, complexation or present in the.structure of minerals . An interesting experimen-

tal approach commonly used for studying parti-tioning and metal mobility in soils is to use se-quential extraction procedures. These proceduresdo not provide a direct characterization of metalspeciation, but rather an indication of its bindingform or its partitioning.

Depending on pH and redox potential of thesoil environment, arsenic can occur in two stable

Ž .oxidation states that form oxyanions: As V asŽ x�3. Ž .arsenate species H AsO , and As III as ar-x 4

Ž x�3.senite species H AsO . In general, trivalentx 3arsenic is of more environmental concern, be-cause it is more mobile and toxic. At high redoxpotential and low pH, pentavalent arsenate

Žspecies tend to dominate Masscheleyn et al.,.1991 . Iron, manganese and aluminium oxides,

clay content, and organic matter content are alsosoil properties that are strongly related to arsenic

Žsorption Thanabalasingam and Pickering, 1986;.Lin and Puls, 2000 . In soils, chromium mainly

Žexists in two stable oxidation states hexavalent.and trivalent . It exhibits a typical anionic sorp-

tion behavior, its adsorption decreasing with in-creasing pH and when competing dissolved an-

Ž .ions are present Khaodhiar et al., 2000 . Wit-Ž .tbrodt and Palmer 1995 reported that oxidized

chromium can be reduced in soils by redox reac-tions with aqueous inorganic species, electrontransfers at mineral surfaces, reactions with non-humic organic substances or reduction by soilhumic substances. Like As, Cu retention and par-titioning in soils is related to the presence oforganic matter, Fe and Mn oxides, and clay min-erals. Cu is mainly retained in soils through ex-change and specific adsorption mechanisms. How-ever, it has a high affinity for soluble organic

Žligands and for humic compounds Alloway, 1990;.McLean and Bledsoe, 1992 .

As seen previously, numerous studies were per-formed to assess the individual behavior of Cr,

Ž .Cu, and As in soils. Yet, Carey et al. 1996 andŽ .Khaodhiar et al. 2000 studied copper, chromate,

Ž .and arsenate CCA sorption in individual or

( )C.F. Balasoiu et al. � The Science of the Total En�ironment 280 2001 239�255 241

mixed-metal systems on natural soils and iron-oxide-coated sand, respectively. Lund and FobianŽ . Ž .1991 and Stilwell and Gorny 1997 evaluatedthe distribution of these metals in various hori-zons of CCA-contaminated soils. However, to ourknowledge, no data are available on the partition-ing and speciation of Cr, Cu and As in varioustypes of CCA-contaminated soils.

Therefore, the objective of this study was todetermine the influence of soil compositionŽ .organic matter, sand and clay content andphysicochemical characteristics on the retentionand partitioning of Cr, Cu and As present in ninesynthetic CCA-contaminated soils. A statisticalmixture design was used to set up the number ofsoils and their respective composition. Cu and Crpartitioning was investigated using a sequential

� Ž .extraction procedure while As speciation As IIIŽ .�or As V was assessed by solvent extraction.

2. Materials and methods

2.1. Soil composition

For the purpose of this study, test soils had torepresent a wide range of soils with differentphysicochemical properties. The soils had to bevery homogenous and available in large quanti-ties. For these reasons, artificially contaminatedsynthetic soils were used. Basic components ofsoils and their proportions were based on a largenumber of studies of the composition of EasternNorth American soils. Four individual soil com-ponents were selected: kaolinite, silt, sand, andorganic matter. These components were chosenbecause soil particle distribution, kaolinite con-tent and organic matter content are expected toinfluence metal partitioning in soils.

2.2. Statistical experimental design

�The following constraints upper and lower per-Ž .�centages wt.% were imposed for the four com-

ponents:

Ž . Ž .5�kaolinite X �30 11

Ž . Ž .30�sand X �69.5 22

Ž . Ž .0.5�organic matter X �15 33

Ž . Ž .silt X �25 44

In addition to these upper and lower percentages� Ž . Ž .�Eqs. 1 � 3 , the following constraint was im-posed:

Ž .X �X �X �X �100 51 2 3 4

All constraints on mixture proportions wereŽprocessed using STATISTICA software StatSoft,

.Inc., 1995 . The experimental space was definedby a three-component system, because the silt

Žproportion was kept constant. Four vertices A, C,. Ž .G, I , four edge centroids B, D, H, F and one

Ž .overall centroid E characterized the three-com-Ž .ponent experimental space Fig. 1 . According to

Ž .Snee 1975 , three component designs based onŽvertices, edge centroids, and overall centroid nine

.points , have a global efficiency of 83%, whereasthe efficiency of a mixture design with 100 pointswould have increased to 97%. Therefore, thisexperimental mixture design with nine syntheticsoils to be tested can be considered satisfactory.

2.3. Soil components

A glacier till, that contained no organic matterand almost no clay-sized particles, from a borrow

Ž .pit in Northern Quebec Canada was used toŽobtain sand and silt. The till was air-dried room

.temperature and manually disaggregated to re-duce clumping. Its particle size distribution was

Ždetermined using the D-2487-98 method ASTM,.2000 . The �2-mm fraction was mixed and split

several times to obtain homogenized samples. Thisfraction was then sieved and the silt and clay

Ž .fraction was retained �75 �m . The particle sizeŽ .distribution dry wt.% of this fraction was 98%

Ž . Ž .of silt �75 �m and 2% of clay �2 �m . Toremove the silt and clay-sized materials, the sand

Ž .fraction 2 mm�75 �m was washed with distilledwater and then allowed to dry at room tempera-

Ž .ture Esposito et al., 1989 .


Fig. 1. Experimental space.

Ž .The kaolinite EPK was purchased from SialŽ .QC, Canada . According to the chemical charac-

Ž .terization wt.% given by the manufacturer, SiO2Ž . Ž .46.2% and Al O 37.7% were the main oxide2 3components. Every other metal oxide proportion

Ž .analyzed including Fe O was below 0.8% and2 3Mn oxides were not detected. The organic mattersource was Spagnum peat moss, purchased from

Ž .Berger Peat Moss QC, Canada . The peat wasmanually disaggregated and sieved. The �1-mmfraction was retained.

All laboratory ware utilized during soil compo-nent preparation was made of non-metallic mate-rial. Stainless steel sieves were used and totalconcentrations of Cr, Cu and As were determinedin natural soil, kaolinite and peat after sieving.Analyses were performed using inductively cou-

Žpled plasma atomic emission spectrometry ICP-.AES � TJA, IRIS�Advantage model after di-

gestion with HNO and HCl according to Clesceri3

Ž .et al. 1998 . Cr, Cu and As concentrationsŽ .mg�kg in soil components were �18, �21 and�0.5, respectively.

2.4. Soil synthesis

Approximately 2 kg of each of the nine syn-thetic soils were constituted. To ensure a goodhomogeneity, batches of 100 g of each soil wereprepared at the same time. After determinationof water content at 105�C using method D-2974-87Ž .ASTM, 2000 , soil components were placed in a

Ž .500-ml polypropylene co-polymerized PPCOscrew cap bottle. Kaolinite was introduced first,then sand, peat, and silt were added. Betweeneach addition, the mixtures were agitated to avoidbinding of components. The synthetic soils were

Ž .agitated at room temperature 22�1�C for 24 hon a customized rotary agitator at 50 rev.�min.


2.5. Physicochemical characterization

The pH of sand, silt, kaolinite, and non-con-taminated and contaminated soils was determinedin distilled water according to method D-4972-95aŽ .ASTM, 2000 using a soil: water ratio of 1:4. PeatpH was measured according to method D-2976-71Ž .ASTM, 2000 with a solid�liquid ratio of 1:16.Tubes were sealed and agitated for 30 min on a

Ž .wrist action shaker Burell model 75 . Measure-Žment of pH Orion Ross 8175 BN Electrode,

.Accumet model 25 pH meter was performedafter 48 h on triplicate samples. The pH of thesynthetic soils had to be constant in order toavoid a bias of the toxicity responses during sub-

Žsequent bioassays this part of the study is not.presented in this paper . The initial pH of min-

eral soils varied between 5.4 and 5.8, whereasŽ .organic soils were more acidic 4.2�4.6 . Their pH

was, therefore, increased to 5.5�0.1 by addingpowdered CaCO . The soils were then agitated3for 24 h on a rotary agitator at 50 rev.�minbefore measurement of adjusted pH. All furthercharacterization and analysis were performed onpH-adjusted soils.

Volatile solids of soil components, non-con-taminated soils and contaminated soils were de-termined on triplicate samples at 550�C according

Ž .to Karam 1993 . The buffer capacity of each soilŽcomponent and of synthetic soils expressed as

the number of moles of H� ions needed to lower.the initial pH of 1 kg of solid by 1 pH unit was

Ž .determined according to Zagury et al. 1997 . TheŽ .cation exchange capacity CEC of each soil com-

ponent, of synthetic soils and of CCA-con-taminated soils was determined on duplicate sam-

Ž .ples using the sodium acetate method pH 8.2Ž .according to Chapman 1965 . The total specific

surface area of each soil component and of syn-thetic soils was measured by single-point BET N2adsorption with a Flowsorb II 230 surface area

Ž .analyzer Micrometrics Inc., Norcross, GA .The alkaline extraction of peat was performed

on duplicate samples according to Anderson andŽ . Ž .Schoenau 1993 . Peat fractionation dry wt.%

resulted in the following: 9.5% humic acid, 2.8%fulvic acid, 80% humin and 7.7% plant debris.Total Cr and Cu concentrations in non-con-

taminated and contaminated synthetic soils weredetermined after digestion with HNO HF and3,

Ž .HClO Clesceri et al., 1998 whereas total As4content was determined after digestion with 10 M

Ž .HCl according to Chappell et al. 1995 . Analyseswere performed using ICP-AES. Detection limitsfor Cu, Cr and As were 0.01, 0.01 and 0.05 mg�l,respectively.

2.6. Soil contamination procedure

Ž .A 60% w�w commercial CCA-C solution con-taining, on an oxide basis, 45.5% CrO , 18.2%3CuO, and 36.3% As O was used to contaminate2 5the synthetic soils. The CCA contamination levelhad to be realistic and compatible with Cr, Cuand As concentrations possibly found in contami-nated soils close to wood treated poles in serviceor at timber treatment facilities. For this reason,

Ž2573 mg of CCA 984 mg Cr, 984 mg As, and 605.mg Cu were added per kilogram of dry soil. All

soils were contaminated with the same concentra-tion of CCA. The contamination procedure in-

Ž . Ž .cluded four important steps: i pre-treatment; iiŽ .contact with agitation; iii contact without agita-

Ž .tion; and iv room temperature drying.As a pre-treatment, 100 g of each synthetic soil

were placed in a 500-ml polycarbonate bottle and100 ml of deionized water were added. The mix-ture was left to stand for 24 h to allow particles toregain their initial physical conditions throughhumidification. A solution containing 257.3 mg of

ŽCCA was then introduced in the bottle final.soil�solution ratio of 1:3 . Samples were then

shaken on a rotary agitator at 50 rev.�min for 24� Ž .�h step ii . This contact period was found to be

sufficiently long to attain equilibrium of Cr, CuŽ .and As in a previous study Carey et al., 1996 .

Various periods of contact without agitation� Ž .�step iii were tested on soil E. Metal retentionas a function of contact time without agitationwas investigated by analyzing metal reductionfrom solution after 6, 24, 48 and 120 h. Tenmilliliters of supernatant were removed, filteredŽ .0.45 �m , and analyzed for Cr, Cu and As. Re-sults showed that after 48 h of contact, 9.1% ofCr, 0.5% of Cu, and 16.8% of As still remained insolution whereas after 120 h, 2.9% of Cr, 0.4% of


Cu, and 12.5% of As remained. Consequently, acontact period of 120 h was selected. The mixture

Ž .was then centrifuged Beckman, model J2-21 atŽ .10 000 rev.�min 4000�g for 30 min. The super-

natant was removed, and the soil was transferredinto a Pyrex container and allowed to stand for 72h in a ventilated hood at room temperature until

Žapparent dryness the water content of the syn-.thetic contaminated soils varied from 0.3 to 8.2% .

ŽSamples were disaggregated, and agitated 50.rev.�min for 24 h to regain the homogeneity lost

during centrifugation.Immediately after contamination, total diges-

tions for Cu and Cr determinations and all se-Ž .quential extraction procedures SEP were car-

ried out. However, due to the large number ofsoil samples, not all characterization could bemade simultaneously. Therefore, the contami-nated soil samples were stored in plastic bags in a

Ž .dark refrigerator 4�C for 2�4 days before de-termination of physicochemical characteristicsand determination of As concentration and speci-ation.

2.7. Sequential extraction procedure

The basic utility of SEP is its use of appropriatechemical reagents in a manner that releases dif-ferent heavy metal fractions from soils by destroy-ing the binding ‘agent’ between the metal and thesoil solids, thus permitting the individual metal tobe detected through appropriate analytical proce-dures. Owing to the inherent lack of selectivity ofextraction procedures, numerical designationshave been assigned to the fractions. Thegeochemical description customarily used is alsogiven, but it should not be interpreted as a pre-cise description.

ŽDespite some criticism limited selectivity and.redistribution of metals dealing with the inter-

Žpretation of these extraction procedures Nirel.and Morel, 1990 , their use has continued to be

recognized as a valuable tool, provided they areŽused with discrimination and care Tessier and

.Campbell, 1991 . These techniques can provide agood indication of metal partitioning in soils,sludge, and sediments and also provide a prag-matic estimation of their potential mobility

Ž .Zagury et al., 1997, 1999; Maiz et al., 2000 .Ž .Moreover, Ho and Evans 2000 recently showed

that the extent of extracted metal redistribution isless than previously suspected and does not inval-idate sequential extraction procedures.

Ž .In this study, the SEP of Tessier et al. 1979was performed on the nine contaminated soilsŽ .duplicate samples . A minor modification of thisprocedure was used for the determination of the

Ž .residual fraction Zagury et al., 1999 . In order toevaluate the recovery of the SEP, a soil samplewas simultaneously digested, following the fifthstep procedure. This latter digestion was per-formed in triplicate. The SEP operationally groupsheavy metals into the following five fractions: F1:

Žsoluble and exchangeable extracted with a mag-.nesium chloride solution , F2: bound to carbo-

Žnates or specifically adsorbed leached by an acetic.acid�acetate buffer , F3: bound to reducible Al,ŽFe and Mn oxides extracted with hydroxylamine

.hydrochloride , F4: bound to oxidizable matterŽreleased by nitric acid, hydrogen peroxide, and

.ammonium acetate , and F5: residual metal frac-Žtion dissolved by acid attack with HNO , HF,3

.and HClO .4The extractions were carried out on 1 g of soil

in a 50-ml PPCO centrifuge tube. All solid�liquidseparations were performed by centrifuging at

Ž .10 000 rev.�min 12 000�g for 30 min. The su-pernatant was removed and analyzed for metalconcentration by ICP-AES. The detection limitfor chromium and copper was 0.01 mg�l. Theresidue was washed with 8 ml of deionized water.After centrifugation for 30 min, this second su-pernatant was also removed and analyzed formetal concentration. The solid residue was usedin the next step.

2.8. Arsenic speciation by sol�ent extraction

The feasibility of using a SEP originally de-signed for transition metals for fractionating soil

Ž .As has been criticized Gruebel et al., 1988 .Moreover, it is well known that As mobility andtoxicity depend on its oxidation state. Thus, inthis study, As speciation was performed using anadaptation of the methods developed by KamadaŽ . Ž .1976 , Huang and Wai 1986 and Chappell et al.


Ž .1995 . This inexpensive, yet effective method al-lows the extraction of all arsenic present in the

Ž .soil with concentrated hydrochloric acid HCl ,without changing its speciation. The speciationŽ .arsenite, arsenate is then performed using sol-vent extractions. Organic As species, which areconsidered less toxic, are not measured with thismethod.

2.8.1. Total arsenic in soilOne gram of contaminated soil was accurately

weighed into a 50-ml PPCO centrifuge tube and10 ml of 10 M HCl were added. The extractionwas assisted by shaking vigorously for approxi-mately 30 min. The resulting slurry was cen-

Ž .trifuged at 15 000 rev.�min 27 000�g for ap-proximately 5 min and the supernatant was fil-

Ž .tered Whatman GF�F, 0.45 �m into a 100-mlvolumetric flask. The extraction procedure wasrepeated twice on the same 1-g sample of soil.When the extraction was complete, the soil waswashed into the filter with deionized water andthe solution diluted. The extraction was per-formed in triplicate. Total arsenic concentrationwas determined by ICP-AES. Detection limit forarsenic was 0.05 mg�l. Two of the three arsenicextracts were used to perform the arsenic specia-tion.

2.8.2. Speciation of arsenic in soilThe principle of arsenic speciation is based on

Ž .the affinity of As III for ammonium pyrrolidineŽ .dithiocarbamate APDC . In the 4.0�5.6-pH

range, the complex APDC-methyl isobutyl ketoneŽ . Ž .MIBK has a great affinity for As III and it can

Ž .be used for the selective separation of As III andŽ . Ž .As V Kamada, 1976 .

2.8.3. ReagentsŽ .Standard As III solution, 1000 ppm: In a

1000-ml volumetric flask, 1.320 g of diarsenictrioxide were dissolved in 10 ml of 10 M sodiumhydroxide, then diluted to 1000 ml with deionizedwater. An aliquot of 50 �l of this solution wasdiluted with water to give a concentration of 50ppb before use.

Ž .Standard As V solution, 10 000 ppb: In a1000-ml volumetric flask, 0.019 g of arsenic pen-

toxide trihydrated was added and diluted withdeionized water. An aliquot of 5 ml of this solu-tion was diluted with water to give a concentra-tion of 50 ppb before use.

Ž .Standard As solution, 50 ppb As III �50 ppbŽ .As V : An aliquot of 50 �l of standard solution

Ž . Ž .of As III 1000 ppm and an aliquot of 5 ml ofŽ . Ž .standard solution of As V 10 000 ppb were

added in a 1000-ml volumetric flask and dilutedwith deionized water.

Ž .APDC solution, 2% w�v : Prepared under ni-trogen conditions, by dissolving 5 g of APDC indeionized water and diluting to 250 ml with water.The solution was immediately used for arsenicspeciation.

Buffer solution, pH 5.2: Prepared by mixing 2M sodium acetate with 2 M acetic acid.

Ž . Ž .EDTA disodium salt solution, 10% w�v .All reagents were prepared daily prior to use

for each experiment.

2.8.4. Arsenic extraction procedureA 20-ml aliquot of the arsenic extract was

transferred to a 250-ml separating funnel.Twenty-five milliliters of the EDTA solution, 25ml of the acetate buffer solution, and 10 ml of theAPDC were added. The EDTA was used to maskthe other metals present in the arsenic extractand to prevent interferences. The mixture wasdiluted to 125 ml with deionized water. After theaddition of 10 ml of MIBK, the funnel was man-ually shaken for 5 s, then agitated on a gyratory

Ž .shaker New Brunswick Scientific for 5 min atŽ .170 rev.�min room temperature . The funnel

was then left to stand for 2 min. An aqueous� Ž .�phase containing As V and an organic phase

� Ž . �containing As III -APDC were formed.After phase separation, the aqueous fraction

was transferred in another separatory funnel anda second extraction with MIBK was realized inorder to achieve higher extraction efficiency. Thesame procedure was followed. After the secondseparation, the aqueous fraction was analyzed for

Ž .As V by ICP-AES-hydride generation and theorganic fractions were combined in a separatoryfunnel. The detection limit for arsenic was 0.001

Ž .mg�l. As III concentration was determined by


the difference between total arsenic and pentava-lent arsenic.

2.8.5. Back-extractionŽ .The organic fraction containing the As III -

APDC complex in MIBK was back-extracted intoHNO for analysis. This step was used in order to3

Ž .compare the As III concentration obtained byextraction with that determined by calculation.

Ž .In the organic fraction, 10 ml of 25% v�vHNO were added. The funnel was agitated for 53min and left to stand for another 2 min. The

Ž .aqueous fraction containing As III -HNO was3separated and the remaining organic fraction wasback-extracted once again. The two aqueous frac-tions were combined and analyzed by ICP-AES-hydride generation.

2.8.6. Accuracy of arsenic speciationThe accuracy of the arsenic speciation method

has been tested by analyzing standard arsenic� Ž . Ž .solutions 50 ppb As V , 50 ppb As III , and 50

Ž . Ž .�ppb As III �50 ppb As V in duplicate. TheŽ .results of the recovery tests Table 1 show a good

reproducibility and quantitative recovery ofŽ . Ž .As III and As V confirming the reliability of the

method.All laboratory ware utilized for the analysis of

Cr, Cu and As was cleaned sequentially with aphosphate-free detergent, soaked twice in 10%

Ž .v�v nitric acid for 12 h, then in deionized waterand finally rinsed three times with deionized

Ž .water 18.2 Mohms . Unless otherwise stated, allŽ .reagents were of analytical grade ACS or better.

3. Results and discussion

3.1. Soil composition

The composition of the nine synthetic soilsgenerated by the statistical mixture design isshown in Table 2. The soils can be classified for

Ž .discussion purposes as mineral soils A�C , mod-Ž .erately organic soils D�F and highly organic

Ž .soils G�I . The silt content is constant in allsoils, but sand and kaolinite content vary con-

Ž .trariwise. In the first category A�C , the organicŽ .matter content is minimal 0.5% , but clay con-Ž . Ž .tent increases from soil A 5% to soil C 30% .

Soil A has the maximum sand content and isconsidered as a sandy soil, while soil C is a clayeysoil. The same pattern applies for moderately

Ž .organic soils 7.75% of peat . The clay contentincreases from soil D to soil F, while the sandcontent decreases. Soil E is the central point ofthe experimental space; its composition is theaverage of all soils. Soils G, H and I are highly

Ž .organic 15% peat content with an increasingkaolinite content and a decreasing sand contentfrom soil G to soil I.

Table 1Recovery tests on various arsenic solutions

Ž . Ž .Standard arsenic solutions Total As Total As As V As IIIŽ .A and B are duplicate samples added found found calculated

Ž . Ž . Ž . Ž .�g�l �g�l �g�l �g�l

Ž .Standard As V 50 ppb 50 51 � �Ž . Ž .Extraction of standard As V 50 ppb A 50 � 44 �Ž . Ž .Extraction of standard As V 50 ppb B 50 � 44 �

Ž .Standard As III 50 ppb 50 45 � �aŽ . Ž .Extraction of standard As III 50 ppb A 50 � ND �aŽ . Ž .Extraction of standard As III 50 ppb B 50 � ND �

Ž . Ž .Standard 50 ppb As III �50 ppb As V 100 100 � �Ž . Ž . Ž .Extraction of 50 ppb As III �50 ppb As V A 100 � 56 44Ž . Ž . Ž .Extraction of 50 ppb As III �50 ppb As V B 100 � 50 50

bExtraction of deionised water � � ND �

a Ž .ND: not detected �6 �g�l .b Ž .ND: not detected �1 �g�l .


Table 2Ž .Composition of synthetic soils wt.%

Soil type Peat Kaolinite Sand Silt

Soil A 0.5 5 69.5 25Soil B 0.5 17.5 57 25Soil C 0.5 30 44.5 25

Soil D 7.75 5 62.25 25Soil E 7.75 17.5 49.75 25Soil F 7.75 30 37.25 25

Soil G 15 5 55 25Soil H 15 17.5 42.5 25Soil I 15 30 30 25

3.2. Soil physicochemical properties

As seen in Table 3, the total specific surfaceŽ .area SSA of the soils depends largely on the

amount of kaolinite. SSA increases with clay con-Ž .tent, the clayey soils 30% kaolinite C, F and I

having the greatest values. In this study, kaoliniteŽ 2 .has the greatest SSA 26.5 m �g , followed by silt

Ž 2 . Ž 2 . Ž 2 .3.5 m �g , peat 1.7 m �g and sand 0.8 m �g .At the same proportion of clay content, SSAappears to slightly decrease with the increase in

Ž .organic matter content. Petersen et al. 1996found that the SSA was highly correlated to the

clay-size fraction of soil and negatively correlatedto the soil organic matter content.

Ž .Generally, the cation exchange capacity CECof any soil is considered to arise principally fromthe organic matter and clay fractions. This isbecause the most negative charges responsible forthe CEC originate from the dissociation of car-boxyl groups in organic matter molecules andboth permanent and variable charges on clayminerals. In this study, peat has the greatest CECŽ . Ž232 meq�100 g , followed by kaolinite 38

.meq�100 g . As expected, silt and sand have aŽ .very small CEC less than 2 meq�100 g . Peat

contribution to the CEC is much more importantthan kaolinite’s. Compared to other clay minerals,kaolinite is known to have a relatively low CECŽ .and a low SSA because very little isomorphous

Ž .substitution has occurred Alloway, 1990 . Hence,Ž .the highly organic synthetic soils G, H and I

Žhave the highest CEC from 37.8 to 42.0 meq�100.g . It is possible to observe the small contribution

of clay content to the CEC in mineral soils.Buffer capacity gives an indication of the soil’s

capacity to resist changes in pH. Peat buffercapacity was 9.5 cmol H��kg, while much lower

Žvalues characterized kaolinite, silt and sand 1.2,� .0.5 and 0.4 cmol H �kg, respectively . Therefore,

Ž .the highly organic synthetic soils G, H, I had the

Table 3Physicochemical characteristics of the synthetic non-contaminated soils

aSoil Specific CEC Buffer Volatile Background level ofbŽ .type surface meq�100 g capacity solids cCr Cu As�Ž . Ž .area cmol H �kg wt.% Ž . Ž . Ž .mg�kg mg�kg mg�kg2Ž .m �g

Soil A 1.8 4.7 0.7 1.3�0.1 19.4 31.2 NDSoil B 5.8 8.3 0.8 2.8�0.1 24.7 27.6 NDSoil C 9.6 12.0 0.9 4.2�0.1 35.6 32.8 ND

Soil D 1.5 21.6 1.1 8.1�0.1 16.7 25.9 NDSoil E 4.2 24.1 1.5 9.5�0.4 25.2 24.3 NDSoil F 7.7 25.0 1.7 10.8�0.2 28.4 26.0 ND

Soil G 1.2 37.8 1.6 15.1�0.5 18.3 27.3 NDSoil H 3.6 38.4 1.9 16.6�0.1 28.0 23.5 NDSoil I 6.9 42.0 2.3 18.6�0.7 25.4 26.6 ND

a Mean values are calculated from two different determinations.b Mean values and S.D. are calculated from three different determinations.c Ž .ND: not detected �4 mg�kg .

()

C.F

.Balasoiu

etal.�T

heScience

oftheT

otalEn�ironm

ent2802001

239�255

248

Table 4Copper retention and partitioning in the CCA-contaminated soils

cSoil Total Cu Fraction 1 Fraction 2 Fraction 3 Fraction 4 Fraction 5 Sum of Recov.type retained Soluble and Carbonates or Reducible Al Oxidizable Residual fractions

exchangeable specifically and Fe oxides matteradsorbed

a bŽ . Ž . Ž . Ž . Ž . Ž . Ž . Ž . Ž . Ž . Ž . Ž . Ž . Ž .mg�kg % mg�kg % mg�kg % mg�kg % mg�kg % mg�kg % mg�kg %

Soil A 283�21 46.7 138 43.9 99.7 31.8 60.8 19.4 8.0 2.6 7.5 2.3 314 111Soil B 379�22 62.6 184 45.3 122 30.0 90.4 22.3 6.3 1.5 2.9 0.7 405 107Soil C 391�17 64.6 222 51.6 117 27.1 74.6 17.3 11 2.6 5.8 1.4 431 110

Soil D 568�37 93.8 99.2 15.7 202 31.9 248 39.4 80.3 12.7 1.9 0.3 632 111Soil E 568�16 93.8 51.9 9.1 194 34.0 211 36.9 108 18.9 6.2 1.1 571 100Soil F 562�24 92.8 59.9 9.4 207 32.4 254 39.6 112 17.4 7.7 1.2 641 114

Soil G 585�21 96.6 48.3 7.3 140 21.3 281 42.8 185 28.1 3.1 0.5 657 112Soil H 572�16 94.5 32.6 5.0 132 20.4 278 43.0 203 31.3 1.9 0.3 648 113Soil I 579�26 95.7 38.8 6.1 124 19.4 281 44.0 190 29.9 3.7 0.6 638 110

a Mean values and S.D. are calculated from three different determinations.b Ž .Calculated as: total Cu�Cu added �100.c Ž .Calculated as: sum of fractions�total Cu �100.

()

C.F

.Balasoiu

etal.�T

heScience

oftheT

otalEn�ironm

ent2802001

239�255

249

Table 5Chromium retention and partitioning in the CCA-contaminated soils

cSoil Total Cu Fraction 1 Fraction 2 Fraction 3 Fraction 4 Fraction 5 Sum of Recov.type retained Soluble and Carbonates or Reducible Al Oxidizable Residual fractions

exchangeable specifically and Fe oxides matteradsorbed

a bŽ . Ž . Ž . Ž . Ž . Ž . Ž . Ž . Ž . Ž . Ž . Ž . Ž . Ž .mg�kg % mg�kg % mg�kg % mg�kg % mg�kg % mg�kg % mg�kg %

Soil A 191�21 19.4 54.3 18.8 39.0 13.5 160 55.4 20.2 7.0 15.4 5.3 289 151Soil B 216�12 21.9 57.9 15.7 49.6 13.4 219 59.4 21.5 5.8 20.8 5.7 369 171Soil C 288�102 29.2 60.6 18.3 46.9 14.2 166 50.0 25.9 7.8 32.0 9.7 331 115

Soil D 638�33 64.8 1.7 0.2 84.9 10.4 542 66.0 173 21.0 19.4 2.4 821 129Soil E 618�63 62.8 3.1 0.5 76.7 11.9 373 58.0 162 25.2 28.0 4.4 643 104Soil F 562�28 57.1 1.4 0.2 93.1 11.0 529 62.5 191 22.5 32.3 3.8 847 151

Soil G 852�47 86.5 1.1 0.1 113 14.0 410 50.8 259 32.2 21.0 2.6 804 94Soil H 725�103 73.6 1.7 0.2 68.5 9.1 397 52.5 262 34.6 27.5 3.6 757 104Soil I 722�71 73.2 1.6 0.2 55.7 7.9 360 51.0 261 36.9 28.8 4.0 707 98

a Mean values and S.D. are calculated from three different determinations.b Ž .Calculated as: total Cr�Cr added �100.c Ž .Calculated as sum of fractions�total Cr �100


highest buffer capacity. The buffer capacity ofpeat is significant, because of its important CEC.

ŽAs shown in Table 3 by comparing soils A, B, C.and soils C, F, I , kaolinite contribution to the soil

buffer capacity is less important than peat con-tribution.

Ž .Total volatile solids wt.% of the soil compo-nents were 96�0.5% for peat, 12�0.7% forkaolinite and �0.5% for sand and silt. Conse-quently, the total volatile solids of the nine syn-thetic soils are primarily correlated to peat con-tent and, to a lesser extent, to kaolinite content.They range from 1.3�0.1 to 4.2�0.1% and from15.1�0.5 to 18.6�0.7% for mineral soils andhighly organic soils, respectively. As expected,metal background levels of the nine syntheticsoils are very low: average Cr concentration is24.6�6.3 mg�kg, while average Cu concentra-tion is 27.2�3.0 mg�kg. Arsenic is not de-tectable.

3.3. Characterization of contaminated synthetic soils

As mentioned in Section 2, synthetic soil char-acteristics were checked after contamination with

CCA. CEC, buffer capacity and total volatilesolids were very similar to the initial values andthe average pH of the nine contaminated soils

Žwas almost unchanged from 5.5�0.1 prior to.contamination to 5.5�0.3 after contamination .

3.4. Total metal content of contaminated soils

The Cu, Cr and As concentrations retained inthe soils following the contamination procedureŽ605 mg of Cu, 984 mg of Cr, and 984 mg of As

.were added per kg of dry soil are shown inŽ .Tables 4�6. The small standard deviations S.D.

obtained show the good reproducibility of thecontamination procedure and highlight the homo-geneity of the soils.

Ž .Total Cu concentrations Table 4 in the con-taminated soils vary from 283�21 to 585�21

Ž .mg�kg 47�97% of retention . Cu retention insoils increases strongly as the organic matter con-tent increases. Its sorption behavior is also corre-

Ž .lated to the increase of CEC Table 3 . This is notsurprising since Cu is mainly retained in soilsthrough ion exchange and has a strong affinity for

Ž .humic compounds. In mineral soils A, B, C ,

Table 6Arsenic retention and speciation in the CCA-contaminated soils

Ž . Ž . Ž . Ž . Ž . Ž .Soil Total as Total As V As III As III As V As III As III Massc dŽ .type retained As calculated back- % calculated back- balance

wash washa bŽ . Ž . Ž . Ž . Ž . Ž . Ž . Ž . Ž . Ž .mg�kg % mg�kg mg�kg mg�kg mg�kg % % % %

Soil A 700�38 71.1 720 684 36 34 95 5 4.7 100Soil B 721�50 73.2 698 634 64 94 90.8 9.2 13.4 104Soil C 736�33 74.7 751 678 73 39 90.3 9.7 5.2 95

Soil D 723�47 73.4 733 539 194 144 73.5 26.5 19.6 93Soil E 795�29 80.7 781 637 144 118 81.5 18.5 15.1 97Soil F 706�8 71.7 710 510 200 192 71.8 28.2 27.0 99

Soil G 765�40 77.7 742 534 208 200 71.9 28.1 26.9 99Soil H 735�7 74.6 734 528 206 235 71.9 28.1 32.0 104Soil I 712�11 72.3 718 493 225 174 68.7 31.4 24.2 93

a Mean values and S.D. are calculated from three different determinations.b Ž .Calculated as: total As�As added �100.c Mean values determined from two different determinations on arsenic extracts used for arsenic extraction.d �Ž Ž . Ž . . �Calculated as: As V �As III back-wash �total As �100.


copper retention increases with the increase inŽkaolinite content. However, in moderately D, E,

. Ž .F and highly organic soils G, H, I with a muchŽhigher organic matter content total volatile solids

.�8% , the influence of kaolinite content and ofCEC is not observable anymore.

Ž .Total Cr concentrations Table 5 also showthat there are considerable variations in the pro-portion of Cr retained depending on the soil typeŽ .from 19 to 86% . Average Cr concentration inmineral soils is low and increases rapidly with theincrease in organic matter content. As expected,kaolinite content does not increase the fraction of

ŽCr retained in Table 5, S.D. must be taken into. Ž .account . Korte et al. 1976 reported that clay

content, CEC and SSA had no significant influ-Ž .ence on Cr VI retention. This is consistent with

the results of a recent investigation on naturalŽsoils contaminated with CCA Stilwell and Gorny,

.1997 . They found that Cr content was high insurface horizons, which coincides with high con-centrations of organic matter, but rapidly de-creased with depth. The low Cr sorption in min-eral soils can also be explained by the presence ofarsenate present in the CCA solution. Khaodhiar

Ž .et al. 2000 found that arsenate significantly de-creased chromate adsorption on an iron oxide-coated sand due to competition for adsorptionsites and electrostatic effects.

Arsenic retention behavior is quite differentŽ .Table 6 . Total As concentrations are similar in

Žmineral and organic soils from 700�38 to 795�. Ž29 mg�kg . The retained proportion is high from.71 to 81% in all soils. This is probably because

the contamination was performed in very favor-able conditions for arsenate retention in soils� Ž .slightly acidic pH 5.5 , presence of organic mat-

� Žter and of kaolinite Masscheleyn et al., 1991;. Ž .Lin and Puls, 2000 . Lund and Fobian 1991

found that As was retained by organic matter andFe, Al and Mn oxides in industrial sites contami-nated with CCA. Under the proportions tested inthis experimental design, organic matter contentŽ . Ž .0.5�15% , kaolinite content 5�30% and, conse-quently, sand content did not influence As sorp-tion in soils. The constant pH and the oxido�re-duction potential were probably the key chemicalparameters influencing As sorption.

3.5. Hea�y metal partitioning in the CCAcontaminated soils

The results of the sequential extraction proce-dures for Cu and Cr are presented in Tables 4and 5. The values represent the average of extrac-tions performed on duplicate samples. The aver-age difference percentage between duplicates for

Ž .the nine soils considering all fractions is 9.3�4.6for Cu and 7.4�2.8 for Cr showing the excellentreproducibility of the experiment. Moreover, the

�Žrecovery of the SEP sum of all fractions�total. �metal concentration �100% is very satisfactory

especially for Cu. Recovery is in the range ofŽ .100�114% for Cu with an average value of 109%

Žand of 94�171% for Cr with an average value of.124% .

As the results show, there are considerablevariations in the proportions of Cr and Cu in thevarious fractions. The metal partitioning of min-eral soils is very different from that portrayed bymoderately and highly organic soils. The onlysimilarity is the metal proportion found in the

Ž .residual fraction F5 . In all soils, regardless oftheir composition and physicochemical character-istics, Cu and Cr proportions found in F5 are verylow. This was expected because the soils wereartificially contaminated, and metals found in thisfraction are considered to be bound within thelattice of minerals. Moreover, the low metal con-centrations found in this fraction are close to thebackground levels of the non-contaminated soilsŽ .Table 3 .

In mineral soils, very high levels of CuŽ .43.9�51.6% are extracted with magnesium chlo-ride, suggesting that Cu is predominantly in asoluble or an exchangeable form in these soilsŽ .Table 4 . The Cu proportion found in F1 in-creases with the increase in kaolinite content and

Ž .the increase in CEC soils A�C . Logically, verylow levels of Cu are bound to oxidizable matter inmineral soils. Even though the CEC of organicsoils is higher than the CEC of mineral soilsŽ37.8�42 meq�100 g compared to 4.7�12

.meq�100 g , the levels of Cu found in a solubleor exchangeable form in highly organic soils arevery low, whereas the levels of Cu bound toorganic matter are much higher. This suggests


Ž . Žthat: 1 suitable reactive groups hydroxyl, car-.boxyl, etc. for the formation of complexes with

Cu are available on the humic compounds ofŽ .peat; and 2 Cu is preferentially retained on

organic matter by complexation rather than byŽ .ion exchange. Wu et al. 1999 , who recently

investigated the sorption of Cu on various clayand clay�organic matter fractions obtained fromsoils, also found that Cu was preferentially re-tained on organic matter. Table 4 indicates thatthe remaining Cu is mainly partitioned between

ŽF2 and F3. The level of Cu found in F2 specifi-. Ž .cally adsorbed is relatively constant 28�6% ,

whereas the level found in a reducible form orŽbound to Al or Fe oxides Mn oxides were not

. Ž .present increases from 17.3 to 44% with theincrease in organic matter content of contami-nated soils.

Ž .Chromium partitioning Table 5 is globallyŽsimilar to copper’s especially regarding the in-

.fluence of organic matter content but presentssome differences. First of all, the level of Cr

Ž .found in a reducible form F3 in the nine con-taminated soils is relatively high and constantŽ .56 � 6% , suggesting that one-half of thechromium retained in the soils kept its original

Ž .oxidation state hexavalent . Hence, if reducingconditions are encountered in CCA-contaminatedsoils, and if Cr partitioning in naturally contami-nated soils is similar to the one found in artifi-cially contaminated soils, a mobilization of Crcould occur. As for Cu, the level of Cr found in

Ž .F2 specifically adsorbed is relatively constantŽ .12�2% but is lower. As the peat content in-creases, the level of Cr strongly bound to oxidiz-

Ž .able matter F4 increases whereas the level of CrŽ .in the soluble or exchangeable form F1 dramati-

cally decreases. Actually, as soon as the peatŽ .content reaches 7.75% soils D�I , the proportion

of Cr extractable with magnesium chloride repre-sents less than 1% of total chromium and this isregardless of the kaolinite content of the con-taminated soils. Likewise, studying the fractiona-tion of heavy metals in sandy and loessial soils,

Ž .Han and Banin 1999 found that, after its addi-tion to the soil, Cr was mostly bound to theorganic matter fraction.

In summary, in moderately and highly organic

soils, Cr and Cu are present in less mobile andless available forms for soil organisms and plants,

Žthe soluble and exchangeable metals labile frac-.tion usually being considered as the most haz-

Ž .ardous Gupta et al., 1996; Maiz et al., 2000 . Onthe other hand, in mineral soils, 47% of Cu and18% of Cr are found in an easily leachable and

Ž .bioavailable form F1 that present a potentialrisk. Globally, Cu is more mobile than Cr.

3.6. Arsenic speciation in the CCA-contaminatedsoils

As previously discussed, total As concentra-tions in all contaminated soils are similar. How-ever, Table 6 also reveals that this element is

Ž . Ž .present in two species, As V and As III in allsoils. Arsenic is principally pentavalent in a pro-portion varying from 68.7 to 95%, depending on

Ž .soils. The greatest proportion of As V is found inŽ .mineral soils A, B, C , with an average value of

Ž .92%, but the As V proportion decreases, as theorganic matter content of soils increases. In fact,

Ž .the average As III proportion in highly organicsoils becomes significant with an average value of29%. Arsenic was introduced into soils as pen-

Ž .tavalent arsenic As O in the wood preserva-2 5tive, but certain conditions were favorable andallowed its reduction to trivalent arsenic.

Ž .As III was calculated by the differenceŽ .between the results for total arsenic and As V .

In this study, a backwash extraction was per-Ž .formed in order to validate the calculated As III

proportion. The difference between the calcu-Ž .lated As III proportion and the one obtained

following the backwash is generally low, varyingŽ .from 1 to 7% Table 6 . Moreover, the mass

balance varies between 93 and 104%. Thus, it canbe concluded that this modified method used forthe determination of arsenic speciation is reliableand effective.

It is generally recognized that arsenate is themajor species present in oxidized acidic environ-ment, while in reducing and alkaline conditions,

Žarsenite becomes significant Sadiq et al., 1983;.Peters et al., 1996 . In this study, arsenite was

found in significant proportions. In the case ofŽ .field soils, several authors also found As III in


large proportions in conditions theoretically fa-Ž .vorable to the presence of As V . In oxidized

Ž . Ž .200�500 mV and acidic conditions pH 5 ,Ž .Masscheleyn et al. 1991 found that the major

Ž .part 65�98% of arsenic in soil solution wasŽ .present as As V , while the remaining part was

Ž . Ž .As III . In acidic soils average pH of 5.3 con-Ž .taminated by mine tailings, Bowell et al. 1994

found that arsenite was present in a large propor-Ž .tion in aerobic soils up to 45% of total arsenic

and it was the major species in anaerobic soilsŽ .79% of total arsenic . Finally, speciation of ar-

Ž .senic investigated by Ng et al. 1998 on arsenicalpesticide contaminated soils showed that trivalentarsenic components were 0.32�56% in nine com-posite samples of surface and subsurface soils.

Some reduction of arsenate to arsenite tookplace in the nine CCA-contaminated soils. Gen-erally, the precise mechanisms controlling thereduction in soils are poorly understood. Accord-

Ž . Ž .ing to Bowell et al. 1994 , the reduction of As VŽ .to As III is predominantly chemically controlled

in both aerobic and anaerobic soils. Yet, PongratzŽ . Ž .1998 and McGeehan and Naylor 1994 reportthat this reduction can occur as a result of bioticprocesses. In our study, the concentration of AsŽ .III is very low in mineral soils and increasesnotably in organic soils. Organic matter could

Ž .have provided suitable conditions for As V re-duction. It could also have provided an energysource for micro-organisms possibly involved inredox transformations.

4. Conclusions

This study has shown that soil composition andphysicochemical characteristics strongly influencemetal retention and partitioning in CCA-con-taminated synthetic soils. The main findings arepresented below:

� Organic matter content had a strong influenceon CEC, on buffer capacity and on Cu and Crretention. The SSA of the soils largely de-pended on the amount of kaolinite, but claycontribution to the CEC was very small.

� Average Cu and Cr retention in mineral soils

was low, but increased dramatically in highlyorganic soils. Cu sorption slightly increasedwith kaolinite content, but this was not truefor Cr. In organic soils, the influence of kaoli-nite content and of CEC on metal sorptionwas no longer observable.

� Arsenic retention behavior was different. Bothorganic and mineral soils demonstrated a veryhigh sorption of As.

� The levels of Cu and Cr found in a soluble orexchangeable form in highly organic soils werevery low, whereas the levels strongly bound toorganic matter were much higher. This sug-gests that Cu was retained on organic matterby complexation rather than by ion exchange.However, the labile fraction was higher inmineral soils, thus presenting a potential risk.Furthermore, the level of Cr found in a re-ducible form in the nine soils was relativelyhigh and constant. In short, metals present inorganic soils contaminated with CCA wereless mobile and less bioavailable than metalspresent in mineral soils. Globally, Cu was moremobile than Cr.

� The modified method used for the selectivedetermination of mineral As species in CCA-contaminated soils was found to be quantita-tive and reliable. Satisfactory results were ob-tained, revealing that the arsenic present wasprincipally in the pentavalent state. Neverthe-less, in organic soils, arsenite was found insignificant proportions indicating that somereduction of arsenate to arsenite occurred.

In order to confirm the likely uptake of thevarious metal species by organisms and plants,bioassays were also performed on the CCA-con-

Ž .taminated soils using earthworm Eisenia foetidaŽ .and barley Hordeum �ulgare L. . These results

will be published in the near future.

Acknowledgements

The authors acknowledge the financial supportof the Chair partners: Alcan, Bell Canada, Cana-dian Pacific Railway, Cambior, Centre d’Expertiseen Analyse Environmentale du Quebec, Gaz de´


´France�Electricite de France, Hydro-Quebec,´ ´Ministere des Affaires Municipales et de la`Metropole, Natural Sciences and Engineering´

Ž .Research Council of Canada NSERC , Petro-Canada, Solvay, Total Fina ELF, and Ville deMontreal. The authors also wish to acknowledge´Prof. Bernard Clement from the Department of´

´Mathematics and Industrial Engineering, EcoleŽ .Polytechnique de Montreal Canada for his col-´

laboration. Thanks are also due to M. Leduc, S.Estrela and J.P. Bertrand for their assistance inthe laboratory.

References

Alloway BJ. Soil processes and the behaviour of metals. In:Alloway BJ, editor. Heavy metals in soils. New York: JohnWiley & Sons, Inc, 1990:7�28.

Anderson DW, Schoenau JJ. Soil humus fractions. In: CarterMR, editor. Soil sampling and methods of analysis. BocaRaton: Lewis Publishers, 1993:391�395.

American Society for Testing and Materials. The AnnualBook of ASTM Standards. Section four: Construction.Volume 04.08: Soil and Rock, 2000.

Bowell RJ, Morley NH, Din VK. Arsenic speciation in soilporewaters from Ashanti mine, Ghana. Appl Geochem1994;9:15�22.

Carey PL, McLaren RG, Adams JA. Sorption of cupric,dichromate and arsenate ions in some New Zealand soils.Water Air Soil Pollut 1996;87:189�203.

Chappell J, Chiswell B, Olszowy H. Speciation of arsenic in acontaminated soil by solvent extraction. Talanta 1995;42:323�329.

Chapman HD. Cation-exchange capacity. In: Black CA, edi-tor. Methods of soils analysis. Part 2: Chemical and mi-crobiological properties. Madison: American Society ofAgronomy Inc, 1965:891�901.

Clesceri LS, Greenberg AE, Eaton AD, editors. Standardmethods for the examination of water and wastewater.Washinghton, DC: American Public Health Association,1998.

Cooper PA. Leaching of CCA: Is it a problem? In: Environ-mental considerations in the manufacture, use and disposalof preservative-treated wood. Madison, Wisconsin: ForestProducts Society, 1994:45�57.

Esposito P, Hessling J, Locke BB, Taylor M, Szabo M, Thur-nau R, Rogers C, Traver R, Barth E. Results of treatmentevaluations of a contaminated synthetic soil. JAPCA1989;39:294�304.

Gruebel KA, Davis JA, Leckie JO. The feasibility of usingsequential extraction techniques for arsenic and seleniumin soil and sediments. Soil Sci Soc Am J 1988;52:390�397.

Gupta SK, Vollmer MK, Krebs R. The importance of mobile,mobilisable and pseudo total heavy metal fractions in soil

for three-level risk assessment and risk management. SciTotal Environ 1996;178:11�20.

Han FX, Banin A. Long-term transformation and redistribu-tion of potentially toxic heavy metals in arid-zones soils: IIIncubation at the field capacity moisture content. WaterAir Soil Pollut 1999;114:221�250.

Ho MD, Evans GJ. Sequential extraction of metal contami-nated soils with radiochemical assessment of readsorptioneffects. Environ Sci Technol 2000;34:1030�1035.

Huang YQ, Wai CM. Extraction of arsenic from soil digestswith dithiocarbamates for ICP-AES analysis. Commun SoilSci Plant Anal 1986;17:125�133.

Ž .Kamada T. Selective determination of arsenic III and ar-Ž .senic V with ammonium pyrrolidinedithiocarbamate,

sodium diethyldithiocarbamate and dithizone by means offlameless atomic-absorption spectrophotometry with a car-bon-tube atomizer. Talanta 1976;23:835�839.

Karam A. Chemical properties of organic soils. In: CarterMR, editor. Soil sampling and methods of analysis. BocaRaton: Lewis Publishers, 1993:459�471.

Khaodhiar S, Azizian MF, Osathaphan K, Nelson PO. Copper,chromium, and arsenic adsorption and equilibrium model-ing in an iron-oxide-coated sand, background electrolytesystem. Water Air Soil Pollut 2000;119:105�120.

Korte NE, Skopp J, Fuller WH, Niebla EE, Aleshii BA. Traceelement movement in soils: influence of soil physical andchemical properties. Soil Sci 1976;122:350�359.

Lin Z, Puls RW. Adsorption, desorption and oxidation ofarsenic affected by clays minerals and aging process. Envi-ron Geol 2000;39:753�759.

Lund U, Fobian A. Pollution of two soils by arsenic, chromiumand copper, Denmark. Geoderma 1991;49:83�103.

Maiz I, Arambarri I, Garcia R, Millan E. Evaluation of heavymetal availability in polluted soils by two sequential extrac-tion procedures using factor analysis. Environ Pollut2000;110:3�9.

Masscheleyn PH, Delaune RD, Patrick Jr. WH. Effect ofredox potential and pH on arsenic speciation and solubilityin a contaminated soil. Environ Sci Technol 1991;25:1414�1419.

McGeehan SL, Naylor DV. Sorption and redox transforma-tion of arsenite and arsenate in two flooded soils. Soil SciSoc Am J 1994;58:337�342.

McLean JE, Bledsoe BE. Behaviour of metals in soils. Techni-cal research document, EPA�540�S-92�018, Environmen-tal Research Laboratory. Ada, Oklahoma, USA: Office ofResearch and Development, U.S. Environmental Protec-tion Agency, 1992.

Ng JC, Kratzmann SM, Qi L, Crawley H, Chiswell B, MooreMR. Speciation and absolute bioavailability: risk assess-ment of arsenic-contaminated sites in a residential suburbin Canberra. Analyst 1998;123:889�892.

Nirel PMV, Morel FMM. Pitfalls of sequential extraction.Water Res 1990;24:1055�1056.

Peters GR, McCurdy RF, Thomas J, Hindmarch JT. Environ-mental aspects of arsenic toxicity. Crit Rev Clin Lab Sci1996;33:457�493.


Petersen LW, Moldrup P, Jacobsen OH, Rolston DE. Rela-tions between specific surface area and soil physical andchemical properties. Soil Sci 1996;161:9�21.

Pongratz R. Arsenic speciation in environmental samples ofcontaminated soil. Sci Total Environ 1998;224:133�141.

Sadiq M, Zaidi TH, Mian AA. Environmental behaviour ofarsenic in soils: theoretical. Water Air Soil Pollut 1983;20:369�377.

Snee RD. Experimental design for quadratic models in con-strained mixture spaces. Technomet 1975;17:149�159.

�StatSoft, Inc. Statistica for Windows Computer program man-�ual . Tulsa, OK: StatSoft, Inc, 2300 East 14th Street, Tulsa,

OK, 1995.Stilwell DE, Gorny KD. Contamination of soil with copper,

chromium, and arsenic under decks built from pressuretreated wood. Bull Environ Contam Toxicol 1997;58:22�29.

Tessier A, Campbell PGC, Bisson M. Sequential extractionprocedure for the speciation of particulate trace metals.Analyt Chem 1979;51:844�851.

Tessier A, Campbell PGC. Comment on ‘Pitfalls of sequentialextraction’. Water Res 1991;25:115�117.

Thanabalasingam P, Pickering WF. Arsenic sorption by humicacids. Environ Pollut 1986;12:233�246.

Warner JE, Solomon KR. Acidity as a factor in leaching ofcopper, chromium, and arsenic from CCA-treated dimen-sion lumber. Environ Toxicol Chem 1990;9:1331�1337.

Ž .Wittbrodt PR, Palmer CD. Reduction of Cr VI in thepresence of excess soil fulvic acid. Environ Sci Technol1995;29:255�263.

Wu J, Laird DA, Thompson ML. Sorption and desorption ofcopper on soil clay components. J Environ Qual1999;28:334�338.

Zagury GJ, Colombano SM, Narasiah KS, Ballivy G. Neutrali-sation of acid mine tailings by addition of alkaline sludgefrom pulp and paper industry. Environ Technol1997;18:959�973 in French, with English abstract.

Zagury GJ, Dartiguenave Y, Setier J-C. Ex situ electrorecla-mation of heavy metals contaminated sludge: pilot scalestudy. J Environ Eng 1999;125:972�978.

partitioning and speciation of chromium, copper, and arsenic in cca-contaminated soils: influence of...

Documents