Af

APa

b

c

d

e

f

g

1

a

ARR2AA

KMSGFBI

1

oiCepmbpmime[

0d

Journal of Hazardous Materials 176 (2010) 913–918

Contents lists available at ScienceDirect

Journal of Hazardous Materials

journa l homepage: www.e lsev ier .com/ locate / jhazmat

ssessment of the BCR sequential extraction procedure for thalliumractionation using synthetic mineral mixtures

les Vaneka,∗, Tomás Grygarb, Vladislav Chrastny c,d, Václav Tejneckya,etr Drahotae,f, Michael Komárekg

Department of Soil Science and Soil Protection, Czech University of Life Sciences Prague, Kamycká 129, 16521 Praha 6, Czech RepublicInstitute of Inorganic Chemistry, Analytical Laboratory, Academy of Sciences of the Czech Republic, 25068 Rez u Prahy, Czech RepublicCzech Geological Survey, Geologická 6, 15200 Praha 5, Czech RepublicDepartment of Applied Chemistry, University of South Bohemia, Studentská 13, 37005 Ceské Budejovice, Czech RepublicInstitute of Geochemistry, Mineralogy and Mineral Resources, Charles University in Prague, Albertov 6, 12843 Praha 2, Czech RepublicInstitute of Geology, Academy of Sciences of the Czech Republic, Rozvojová 269, 16500 Praha 6, Czech RepublicDepartment of Agro-Environmental Chemistry and Plant Nutrition, Czech University of Life Sciences Prague, Kamycká 129,6521 Praha 6, Czech Republic

r t i c l e i n f o

rticle history:eceived 10 September 2009eceived in revised form3 November 2009ccepted 23 November 2009vailable online 27 November 2009

a b s t r a c t

This work focused on the specific behavior of Tl-bearing phases in the BCR (Community Bureau of Ref-erence) sequential extraction (SE) scheme, namely Tl-bearing ferrihydrite, goethite, birnessite, calcite,illite, sphalerite and feldspar in their simple model mixtures with quartz. Several significant discrepanciesbetween the obtained and expected behaviors of these phases in the BCR SE were observed. The amountof Tl released as the exchangeable/acid-extractable fraction (55–82% of the total Tl content) showed asubstantial H+-promoted dissolution of all Fe(III) and Mn(III, IV) oxides (corresponding to up to 61% of

eywords:etal

equential extractionoethiteerrihydriteirnessite

solid Fe dissolved) and incongruent (increased) extraction of Tl from ferrihydrite and goethite. Reductiveconditions of the second SE step were insufficient to complete goethite dissolution with correspondingTl amount retained in the solid phase. Similarly, insufficient oxidation of sphalerite and lower Tl recov-ery of the oxidisable fraction was identified. In contrast, the BCR SE seems to produce well predictableresults of Tl leaching from Tl-bearing calcite and feldspar. Only 70% of total Tl content was extracted fromTl-modified illite in the exchangeable/acid-extractable step, while 30% was associated with the reducible

., Tl w

llite and residual fractions, i.e

. Introduction

Thallium (Tl) is a toxic metal [1,2] included in the US EPA listf priority toxic pollutants. Because of its acute and chronic toxic-ty (for most living organisms) that is higher compared to, e.g., Hg,d and Pb [2,3], Tl can be thought as one of the most dangerouslements in the environment. Its anthropogenic sources includerimarily emissions or solid wastes from ferrous and non-ferrousining/smelting and coal combustion [3,4]. Moreover, there has

een evidence of Tl contamination as a consequence of cementroduction [5,6]. Environmental availability as well as mobility ofetals and metalloids depends on their chemical speciation [7];

.e., on Tl bonding to individual soil/sediment components. Infor-ation about Tl retention in soils is generally attributed to simple

xchange reactions on oxides, specific silicates or organic matter5]. According to Jacobson et al. [8], mineral phases responsible

∗ Corresponding author. Tel.: +420 224 382 633; fax: +420 234 381 836.E-mail address: vaneka@af.czu.cz (A. Vanek).

304-3894/$ – see front matter © 2009 Elsevier B.V. All rights reserved.oi:10.1016/j.jhazmat.2009.11.123

as strongly fixed to the illite matrix.© 2009 Elsevier B.V. All rights reserved.

for Tl immobilization include primarily illite clays and Mn(III, IV)oxides (mainly �-MnO2). Due to its lithophilic and chalcophiliccharacter, Tl can substitute/exchange K (e.g., in silicates and oxides)as well as associate with sulfides (e.g., pyrite and sphalerite). Fur-thermore, its affinity for thiol groups in organic compounds is wellknown [4].

To our best knowledge, the standardized or optimized BCR(Community Bureau of Reference) sequential extraction (SE) is oneof the most frequently used methods for predicting Tl fractiona-tion in polluted soils and sediments [9–12]. However, the use ofSE methods has been often criticized for the lack of selectivenesscaused by possible redistribution and readsorption of some ele-ments during the extraction procedure, inadequate efficiency ofthe extraction agents used or even absence of specific soil/sedimentphases [7,13,14]. Identification and quantification of specific pollu-

tant associations within the geological samples may be thereforewrongly evaluated.

For this reason, the aims of present study are (i) to test theefficiency of the optimized three-step BCR SE [15] for Tl frac-tionation in samples of artificially prepared mineral mixtures

9 rdous

as

2

2

GiMwmFir[sT(ma

2

wr(uiTt

pctTrcc

Cw115tw

�

TM

14 A. Vanek et al. / Journal of Haza

nd (ii) to assess its application in natural Tl-rich soils orediments.

. Experimental

.1. Chemicals and samples

Chemicals of analytical grade (Lach-Ner, Czech Republic; Merck,ermany) were used for the preparation of synthetic phases and for

ndividual extraction steps. All the solutions were prepared usingilliQ-Plus (Millipore, USA) deionised water. The Tl source was aater solution of dissolved Tl2SO4 (analytical grade; Fluka, Ger-any) which was added to the reaction mixtures for synthesis of

e/Mn oxides, calcite (CaCO3) and Tl-enriched illite. Regarding Tlncorporation into the structures of mineral phases observed withineal/natural systems (with varying redox potentials and pH values)16], co-precipitation performed by all synthetic phases was cho-en as a more suitable method for Tl enrichment [7] compared tol adsorption on the surface of “pure” phases. Potassium–feldsparKAlSi3O8) and sphalerite (ZnS) were used as primary Tl-bearing

inerals. Quartz (SiO2) represented the Tl free mineral phasepplied in the experiment.

.2. Synthetic mixtures

Mixtures containing different amounts of Tl-bearing phasesere prepared to approximate mineral associations of Tl occur-

ing in soils and sediments. The proportion of Tl-bearing phases1–50 wt.%) and solid phase Tl concentrations within the individ-al mixtures was chosen in accordance with their usual abundance

n contaminated (or non-contaminated) soils/sediments [5,6,9,10].he mineralogical composition and corresponding Tl concentra-ions are listed in Table 1.

Thallium-enriched illite ((KH3O)Al2(SiAl)4O10(OH)2) was pre-ared from a circum-neutral water suspension of illite in solutionontaining 50 mL of 600 mg Tl L−1 (Tl2SO4, Fluka, Germany), so thathe total Tl concentration corresponded to ∼2000 mg kg−1 illite.he suspension was aged at room temperature for two days, theninsed and dried at 105 ◦C to constant weight. Illite with the final Tloncentration 1980 ± 20 mg kg−1 was used in mixture A, a modellay-rich material.

Thallium-enriched calcite (CaCO3) was prepared with 47.2 ga(NO3)2·4H2O dissolved in 500 mL deionized water and enrichedith Tl solution, so that the total concentration was ∼1000 mg Tl inkg CaCO3. The carbonate was precipitated after adding of 200 mLM Na2CO3·10H2O solution. The suspension was aged at 30 ◦C for

days, then dried at 105 ◦C to constant weight. The final Tl concen-

ration associated with CaCO3 was 958 ± 12 mg Tl kg−1. This phaseas used in mixture B, a model carbonate-rich material.

Synthetic Fe(III) oxides (ferrihydrite, 5Fe2O3·9H2O, goethite,-FeOOH) were prepared by the KOH–Fe(NO3)3 method of Schw-

able 1ineral composition of the synthetic mixtures (wt.%) with corresponding thallium conce

Mixture A Mixture B Mixture C Mix

% Tl % Tl % Tl %

Ferrihydrite – – – – 1 9.64 ± 0 –Goethite – – – – – – 1Birnessite – – – – – – –Calcite – – 2 19.2 ± 0.2 – – –Illite 1.5 29.7 ± 0.3 – – – – –Sphalerite – – – – – – –K–feldspar – – – – – – –Quartz 98.5 0.1 ± 0.02 98 0.1 ± 0.02 99 0.1 ± 0.02 99∑

– 29.8 – 19.3 – 9.74 –

Materials 176 (2010) 913–918

ertmann and Cornell [17]. Birnessite (K4Mn14O27·9H2O) wasprepared by the HCl–KMnO4 method of McKenzie [18]. Prior tothe addition of KOH or HCl reaction solutions, dissolved Fe(NO3)3or KMnO4 were enriched in Tl so that ∼1 mg Tl could be availablefor synthesis of 1 g of the Fe or Mn oxide formed. Thallium con-centration in ferrihydrite, goethite and birnessite reached 964 ± 0,635 ± 14 and 555 ± 13 mg Tl kg−1, respectively. The oxides wereused in mixtures C, D and E and represent Fe(III) or Mn(III, IV)oxide-rich materials.

Mixtures F and G containing Tl-bearing sphalerite (ZnS,9 mg Tl kg−1) and K–feldspar (KAlSi3O8, 1.3 mg Tl kg−1), respec-tively, represent mixtures with increased lithogenic Tl content.

In order to preserve chemical homogeneity of the mineral con-stituents, Tl-bearing phases (natural minerals as well as artificialTl-enriched phases) were finely crushed or pulverized in an agatemortar (KM 1, Janetzki, Germany). Quartz was crushed in a labora-tory stone grinder (Fritsch, Germany).

2.3. BCR sequential extraction

The selectivity of the BCR SE for specific Tl-bearing species wastested using the optimized method according to Rauret et al. [15].The fractions determined were as follows: (i) exchangeable/acid-extractable fraction (0.11 M CH3COOH-extractable); (ii) reduciblefraction (0.5 M NH2OH·HCl-extractable); (iii) oxidisable fraction(8.8 M H2O2/1 M CH3COONH4-extractable); (iv) residual fraction(total digestion of the residue using a mixture of concentratedHF/HClO4/HNO3). The extraction was performed in three replicatesand the sum of individual extraction steps was in a good agree-ment with total Tl concentration (recovery differences were lessthan 10%; Table 2).

2.4. Selective dissolution procedures

In order to better understand the stability of synthetic oxides,mixture D (containing Tl-doped goethite) was selectively testedfor dissolution with acetic acid (CH3COOH) and hydroxylaminehydrochloride (NH2OH·HCl). Two extraction procedures were con-ducted as follows:

(i) 1 g of solid was extracted with stirring in 40 mL 0.11 MCH3COOH at pH 2.9 (equivalent to the first step of the BCR SE)for 8, 16 and 24 h;

(ii) 1 g of solid was extracted with stirring in 40 mL 0.5 MNH2OH·HCl at pH 1.5 (equivalent to the second step of the BCRSE) for 8, 16 and 24 h. After stirring, the suspensions were fil-

tered at 0.2 �m. The supernatants were consequently analyzedfor Fe using ICP-OES. Both the extractions were performed inthree replicates; recovery differences between the individualreplicates were less than 10%. All the mineral residues wereexamined by X-ray diffraction (XRD).

ntrations (mg kg−1 ± SD).

ture D Mixture E Mixture F Mixture G

Tl % Tl % Tl % Tl

– – – – – – –6.35 ± 0.14 – – – – – –– 1 5.55 ± 0.13 – – – –– – – – – – –– – – – – – –– – – 30 2.70 ± 0.01 – –– – – – – 50 0.63 ± 0.020.1 ± 0.02 99 0.1 ± 0.02 70 0.07 ± 0.01 50 0.05 ± 0.01

6.44 – 5.64 – 2.77 – 0.68

A. Vanek et al. / Journal of Hazardous

Tab

le2

Thal

liu

md

istr

ibu

tion

betw

een

the

oper

atio

nal

ly-d

efin

edfr

acti

ons

ofth

eB

CR

sequ

enti

alp

roce

du

rein

the

stu

die

dm

iner

alm

ixtu

res

(con

cen

trat

ion

ofTl

inm

gkg

−1±

SD).

Frac

tion

Mix

ture

AM

ixtu

reB

Mix

ture

CM

ixtu

reD

Mix

ture

EM

ixtu

reF

Mix

ture

G

Val

ues

exp

ecte

dV

alu

esob

tain

edV

alu

esex

pec

ted

Val

ues

obta

ined

Val

ues

exp

ecte

dV

alu

esob

tain

edV

alu

esex

pec

ted

Val

ues

obta

ined

Val

ues

exp

ecte

dV

alu

esob

tain

edV

alu

esex

pec

ted

Val

ues

obta

ined

Val

ues

exp

ecte

dV

alu

esob

tain

ed

Exch

ange

able

/ac

id-e

xtra

ctab

le29

.7±

0.3

19.3

±0.

319

.2±

0.2

18.9

±0.

30

6.43

±0.

540

5.10

±0.

130

2.76

±0.

120

<DL

0<D

L

Red

uci

ble

05.

19±

0.62

00.

41±

0.03

9.64

±0

2.91

±0.

256.

35±

0.14

0.27

±0.

015.

55±

0.13

2.26

±0.

010

0.05

±0.

020

0.02

±0.

002

Oxi

dis

able

00.

27±

0.03

0<D

L0

<DL

0<D

L0

<DL

2.70

±0.

011.

76±

0.11

0<D

LR

esid

ual

0.1

±0.

023.

01±

0.45

0.1

±0.

02<D

L0.

1±

0.02

0.12

±0.

170.

1±

0.02

0.83

±0.

050.

1±

0.02

<DL

0.07

±0.

010.

95±

0.11

0.68

±0.

020.

61±

0.05

∑29

.8±

0.3

27.8

±0.

819

.3±

0.2

19.3

±0.

39.

74±

0.02

9.45

±0.

626.

44±

0.14

6.19

±0.

145.

64±

0.13

5.02

±0.

122.

77±

0.02

2.76

±0.

160.

68±

0.02

0.63

±0.

05

Materials 176 (2010) 913–918 915

2.5. Mineralogical and chemical analyses

X-ray powder diffraction analysis (X’Pert Pro diffractometer,PANalytical, the Netherlands) was used to characterize the min-eralogical composition of the experimental samples. The analyseswere performed under the following conditions: CuK� radiation,40 kV, 30 mA, step scanning at 0.02◦/150 s in the range 3–80◦ 2�.Qualitative analysis was performed with XPert HighScore software1.0d, equipped with the JCPDS PDF-2 database [19].

In order to determine the chemical composition of the min-eral phases, the amount of 0.3–0.5 g of the homogenized samplewas digested in mixtures of HNO3/HF/H2O2 using a microwavedigestion unit (Mars 5, Matthews, USA). Thallium and metals (e.g.,Fe, Mn, Zn, Ca, K) concentrations in all the minerals, SE digestsas well as solutions from selective extractions were determinedusing ICP-MS (PQExCell, ThermoElemental, UK) and ICP-OES (iCAP6500, Thermo Fisher, Germany) under standard analytical condi-tions, respectively. Aqueous multielement solutions of Merck IVand Merck VI (CertiPUR, Merck, Germany) were used for externalcalibration of ICP-MS and ICP-OES, respectively. The accuracy ofTl determination was controlled by the method of spiking with anaqueous solution (ICP-MS-200.8-CAL1R-1, AccuStandard, Inc., USA)added periodically to each different sample matrix. The differencesobtained between spiked and unspiked solution was less than 5%.

Quality control of Tl and metal analysis was evaluated usingthe SRM 2711 (Montana II soil) reference material certified by theNational Institute of Standards and Technology (USA) (certified Tlvalue 2.47 ± 0.15 �g g−1; found Tl value 2.43 ± 0.07 �g g−1).

3. Results and discussion

3.1. Exchangeable/acid-extractable (“labile”) fraction

This operationally-defined fraction can be regarded as thefraction with the highest potential for metal mobilization and avail-ability. The amount of Tl released during the first extraction stepof SE comprises adsorbed, exchangeable and easily acid-solublespecies. Due to the acidic character of the extraction agent (0.11 Macetic acid, CH3COOH), the interaction of soil/sediment with rootexudates can be approximately simulated.

For mixture A, where Tl is associated with illite, the major-ity of Tl (∼70%) was extracted in the “labile” fraction (Table 2).Approximately 30% of Tl was relatively insoluble (included in otherfractions of SE) indicating a stronger Tl incorporation into the sil-icate structure. For mixture B with Tl associated with calcite, agood agreement between the expected and obtained amount ofthe “labile” fraction was found. Up to 98% of total Tl (correspond-ing to ∼90% of total Ca, data not shown) was released during thefirst extraction step (Table 2), proving practically complete carbon-ate dissolution. Contrarily to our expectations, “labile” Tl extractedfrom the mixtures C, D and E containing Tl-bearing Fe(III) andMn(III, IV) oxides was considerably large (55–82% of the total Tlcontent) (Table 2). The concentrations of dissolved Fe and Mnaccounted for 47% (ferrihydrite), 61% (goethite), and 56% (birnes-site) of total Fe/Mn (Fig. 1); a substantial dissolution of all thethree oxides is thus suggested. Such poor oxide stability must beattributed to the acidic character (pH ∼2.9) of CH3COOH applied.Iron and Mn complexation by acetate ions, as another mobilizationmechanism, is not as significant as H+-promoted dissolution (logK Fe(III)acetate = 4.0–7.6; log K Mn(II)acetate = 1.4, [20]). Whalley

and Grant [21] recorded that the majority of metallic contaminantsbound to ferrihydrite was released during the first step of the BCRSE, indicating oxide degradation in such acidic environment. Sim-ilarly, Slavek and Pickering [22] found, on the basis of dissolvedFe concentrations, that significant amounts of ferrihydrite (up to

916 A. Vanek et al. / Journal of Hazardous Materials 176 (2010) 913–918

F erent(

8tFdmp

onwgiiaipMasssTM(f

3

fTrfE2etpbOrmstXo

ig. 1. Chemical fractionation of Fe and Mn in mixtures C, D and E enriched with diffn = 3).

0 wt.%) were dissolved in the presence of acetic acid. Accordingo Raksasataya et al. [23], dissolution of the synthetically-preparede oxides in the first step of the SE may be enhanced by possibleegradation of amorphous oxide films which are unstable. Such aechanism can, nevertheless, only partially contribute to the total

rocess of Fe/Mn oxide dissolution.Thallium desorption from the exchangeable positions on the

xide surface, as can be expected mainly in the case of bir-essite [24], cannot also be omitted. The calculated Tl/Fe ratiosere approximately 1.4 and 1.3 times higher for ferrihydrite and

oethite, respectively, compared to their initial values. Therefore,t can be concluded that the dissolution of both Fe(III) oxides wasncongruent (i.e., Fe and Tl were not released in the same proportions found in the bulk mineral) because the oxides’ surface is enrichedn Tl. The opposite behavior was revealed for birnessite with com-arable (initial and calculated) Tl/Mn ratios indicating congruentn oxide dissolution and limited Tl retention within the exchange-

ble/surface positions. Thallium incorporation into the birnessitetructure (in non-exchangeable K-exchanged layers) is likely. Thistatement is consistent with our previous findings [25] and thetudy of Jacobson et al. [26] who evaluated Mn oxides as effectivel “scavengers” in mineral soils. No “labile” Tl (undetectable by ICP-S) was determined in mixtures F and G with lithogenic Tl content

Table 2) proving adequate efficiency of the extraction agent usedor that fraction.

.2. Reducible fraction

Hydroxylamine hydrochloride (NH2OH·HCl) is one of the mostrequently employed extraction agents with reducing properties.he application of that reagent is expected to cause the completeeductive dissolution of the synthetic Fe(III) and Mn(III, IV) oxidesrom mixtures C, D and E. Thallium recoveries for mixtures C and

(containing Tl-enriched ferrihydrite and birnessite) were 3 andtimes lower (Table 2), respectively, than expected values in this

xtraction step resulting from the previous H+-promoted dissolu-ion. In mixtures C and E, the remaining Tl (present in the solidhase) was extracted in this step; i.e., the residual ferrihydrite andirnessite were efficiently attacked by reductive acid dissolution.n the contrary, extremely low Tl amount (4% of total Tl, 24% of Tl

emaining in the solid phase after the first step) was released from

ixture D. Regarding reducible Fe extracted (220 mg kg−1, corre-

ponding to 4% of total Fe) (Fig. 1), these results reveal a low rate ofhe residual goethite decomposition. Such prediction is favored byRD indication of goethite within the mineral residue after the sec-nd step (Fig. 2). Pénilla et al. [14] highlight the lack of selectivity of

Fe(III) and Mn(III, IV) oxides. Data of operationally-defined fractions are means ± SD

0.1 M NH2OH·HCl (applied in the standardized BCR SE) which doesnot allow quantitative dissolution of goethite. Likewise, Xiao-Quanand Bin [27] and Neaman et al. [28] reported the same incompletedissolution of crystalline Fe oxides (goethite and hematite). There-fore, it is probable that more stable/crystalline Fe(III) oxides arenot completely reduced with 0.5 M NH2OH·HCl (see Section 3.5)which can lead to overestimation of Tl fractionation in subsequentSE steps.

Interestingly, 19% of total Tl was associated with the reduciblefraction in mixture A (Table 2). Because only some structural Fe(III)ions can be reduced in illite, further ion exchange, selective Tlrelease from some specific adsorption sites, or acid leaching ofillite in consequence of low pH (∼1.5) are more probable expla-nations. Further ion exchange due to the “weakness” of previousion exchange agent or insufficient reaction time (of the previousstep) cannot be excluded. The remaining amount of Tl (∼2%) wasextracted from mixture B (containing Tl-enriched calcite) and wascaused by late carbonate dissolution. Small Tl portions (<2%) werereleased from mixtures F and G (associated with Tl-bearing spha-lerite and K–feldspar, respectively) (Table 2) and can be linked toacidic attack (although in reducing conditions) and slight dissolu-tion of these relatively “insoluble” phases.

3.3. Oxidisable fraction

The oxidisable fraction represents metals tightly bound/incorporated into sulfides and/or organic matter. Hydrogen per-oxide (H2O2) in acidic medium is commonly used as the extractionagent for displacing metals from such components. For mixture F(containing Tl-bearing sphalerite), the recovery of Tl bound to theoxidisable fraction was significantly underestimated (∼36%), sug-gesting insufficient oxidation of that sulfide. Besides the mixture Awith a minimum amount of oxidisable Tl (<1%), no further Tl wasdetected within this fraction (Table 2).

3.4. Residual fraction

The residual Tl fraction obtained by the total dissolution of theresidual matter indicates stable mineral associations. The portionof residual Tl is commonly described by its presence within thestructure of crystalline/amorphous silicates (feldspars, micas, clay

minerals, etc.). Although the residual metal content is not includedin the standardized BCR extraction scheme and is also environmen-tally insignificant, knowledge of Tl release in the last extractionstep is essential for assessment of the SE recovery. For mixture A,high recovery of Tl (11%) was determined in the residual fraction

A. Vanek et al. / Journal of Hazardous Materials 176 (2010) 913–918 917

F ond (r

(Tiii

it(ifs

3

(oowwaab32wttTfor

4

asnc

lh

ig. 2. X-ray diffraction pattern of the mineral residue of the mixture D after the sec

Table 2). This fact may be interpreted by rapid incorporation ofl into the illite structure. Such Tl behavior was already observedn other works and attributes to Tl–K replacement within the illitenterlayers [5,25,26,29] resulting from the same valency and similaronic radius of Tl+ to K+ [4].

In the case of mixture D, the relatively high Tl amounts (13%)ncluded in the residual fraction attributes to late goethite dissolu-ion. This is in accordance with residual Fe concentration dissolved1972 mg kg−1, corresponding to 34% of total Fe) (Fig. 1). Approx-mately 98% of total Tl content was associated with the residualraction in mixture G. Adequate efficiency of previous extractionteps on K–feldspar dissolution is thus confirmed (Table 2).

.5. Selective dissolution of goethite

More attention was paid to the kinetics of goethite dissolutionincluded in mixture D) in the first two SE steps. The dissolutionf goethite by 0.11 M CH3COOH was considerably large; up to 58%f total Fe concentration (corresponding to 3418 ± 288 mg Fe kg−1)as released after 8 h of leaching. A slight increase of dissolved Feas identified with increasing reaction time and accounted for 61

nd 63% (corresponding to 3582 ± 370 and 3688 ± 210 mg Fe kg−1)fter 16 and 24 h, respectively. Higher amounts of Fe were releasedy 0.5 M NH2OH·HCl reaching 64, 69 and 70% (corresponding to804 ± 392, 4060 ± 334 and 4116 ± 495 mg Fe kg−1) after 8, 16 and4 h, respectively. After 16 h only a small increase of dissolved Feas observed, indicating approximate Fe saturation. No precipita-

ion of dissolved Fe has been observed. X-ray diffraction revealedhe presence of the remaining goethite in all the residual samples.hese findings combined with data obtained from the BCR SE, thus,avored (i) substantial degradation of goethite and other Fe/Mnxides in the presence of 0.11 M CH3COOH and (ii) insufficienteduction of goethite by 0.5 M NH2OH·HCl.

. Conclusions

The use of synthetic mineral mixtures in the BCR SE is a reason-ble way to interpret results of this procedure, although naturaloils or sediments may behave in an even more complex man-

er due to the presence of organic and complex organo-mineralonstituents.

The BCR SE scheme seems to produce well predictable results ineaching of Tl from Tl-bearing calcite and feldspar. On the otherand, several significant differences between the obtained and

educible) step of the BCR SE with detected diffraction peaks of goethite (�-FeOOH).

expected behaviors of Tl-bearing phases in the BCR SE scheme wereobserved.

Both the synthetic Fe(III) and Mn(III,IV) oxides were subjectedto H+-promoted dissolution in the first step of the BCR SE andsubstantial amounts of oxide-associated Tl were released in theexchangeable/acid-extractable fraction. The second SE step, reduc-tive conditions, were insufficient to complete goethite dissolutionwith corresponding Tl amount retained in the solid phase; incom-plete reduction of other well crystalline Fe(III) oxides is suggested.The third SE step, oxidation, was insufficient to dissolve sphaleriteand a portion of Tl passed to the residual fraction.

Therefore, the BCR SE seems to be efficient for predicting (i)adsorbed and easily soluble Tl species (e.g., bound to carbonates),and (ii) Tl species associated with primary silicates (feldspars,micas, etc.). On the contrary, a limited selectivity of the BCR SE wasfound for oxide- and sulfide-associated Tl species. The behavior ofTl-enriched illite indicates that further research should be focusedon the interaction of Tl with clay minerals.

Although the present study is focused only on Tl fractionation,the SE procedure is based on specific solubility and dissolutionkinetics of specific mineral phases and results of chemical fraction-ation could be thus possibly generalized also for other metals (e.g.,Pb, Zn and Cd). In real soil/sediment systems, the BCR SE resultsmust be evaluated very carefully and always supported by min-eralogical investigations (e.g., XRD or EPMA). Its application forcomparison purposes may be problematic in geological sampleswith contrasting phase composition.

Acknowledgements

This research was funded by the grant of Czech Science Founda-tion (GACR 526/08/P428) and the projects of Ministry of Educationof the Czech Republic MSM 6046070901 and AV0Z4032918.The ICP-MS/OES and XRD analyses were supported by MSM6007665806 and MSM 002162085, respectively. The authors wouldlike to thank Dr. David Hradil from the Institute of Inorganic Chem-istry (ASCR) for providing pure illite used in the experiment. Twoanonymous reviewers are gratefully thanked for improving theoriginal manuscript.

References

[1] V. Zitko, Toxicity and pollution potential of thallium, Sci. Tot. Environ. 4 (1975)185–192.

9 rdous

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

18 A. Vanek et al. / Journal of Haza

[2] G. Kazantis, Thallium in the environment and health effects, Environ. Geochem.Health 22 (2000) 275–280.

[3] A.L. John Peter, T. Viraraghavan, Thallium: a review of public health and envi-ronmental concerns, Environ. Int. 31 (2005) 493–501.

[4] J.O. Nriagu, Thallium in the Environment, John Wiley & Sons Inc., New York,USA, 1998.

[5] A. Tremel, P. Masson, T. Sterckeman, D. Baize, M. Mench, Thallium in Frenchagrosystems: I. Thallium contents in arable soils, Environ. Pollut. 95 (1997)293–302.

[6] H. Al-Najar, R. Schulz, V. Römheld, Plant availability of thallium in therhizosphere of hyperaccumulator plants: a key factor for assessment of phy-toextraction, Plant Soil 249 (2003) 97–105.

[7] M. Mihaljevic, M. Ponavic, V. Ettler, O. Sebek, A comparison of sequential extrac-tion techniques for determining arsenic fractionation in synthetic mineralmixtures, Anal. Bioanal. Chem. 377 (2003) 723–729.

[8] A.R. Jacobson, M.B. McBride, P. Baveye, T.S. Steenhuis, Environmental factorsdetermining the trace-level sorption of silver and thallium to soils, Sci. Tot.Environ. 345 (2005) 191–205.

[9] C. Yang, Y. Chen, P. Peng, C. Li, X. Chang, C. Xie, Distribution of natural andanthropogenic thallium in the soils in an industrial pyrite slag disposing area,Sci. Tot. Environ. 341 (2005) 159–172.

10] M. Jakubowska, A. Pasieczna, W. Zembrzuski, Z. Swit, Z. Lukaszewski, Thal-lium in fractions of soil formed on floodplain terraces, Chemosphere 66 (2007)611–618.

11] C.R.M. Rao, M.J. Ruiz-Chancho, A. Sahuquillo, J.F. López-Sánchez, Assessment ofextractants for the determination of thallium in an accidentally polluted soil,Bull. Environ. Contam. Toxicol. 81 (2008) 334–338.

12] J. Medved’, M. Kalis, I. Hagarová, P. Matús, M. Bujdos, J. Kubová, Thalliumfractionation in polluted environmental samples using a modified BCR three-step sequential extraction procedure and its determination by electrothermalatomic absorption spectrometry, Chem. Pap. 62 (2008) 168–175.

13] V.H. Kennedy, A.L. Sanchez, D.H. Oughton, A.P. Rowland, Use of single andsequential extractants to assess radionuclide and heavy metal availability fromsoils for root uptake, Analyst 122 (1997) 89–100.

14] S. Pénilla, F. Bordas, J.C. Bollinger, Sequential heavy metal extraction from pol-luted solids: influence of sulfate overconcentration, J. Colloid Interf. Sci. 292

(2005) 20–28.

15] G. Rauret, J.F. López-Sánchez, A. Sahuquillo, E. Barahona, M. Lachica, A.M. Ure,C.M. Davidson, A. Gomez, D. Lück, M. Bacon, M. Yli-Halla, H. Muntau, Ph. Que-vauviller, Application of a modified BCR sequential extraction (three-step)procedure for the determination of extractable trace metal contents in a sewagesludge amended soil reference material (CRM 483), complemented by a three-

[

[

Materials 176 (2010) 913–918

year stability study of acetic acid and EDTA extractable metal content, J. Environ.Monit. 2 (2000) 228–233.

16] A. Vanek, V. Chrastny, M. Komárek, I. Galusková, P. Drahota, T. Grygar, V. Tej-necky, O. Drábek, Thallium dynamics in contrasting light sandy soils – Soilvulnerability assessment to anthropogenic contamination, J. Hazard. Mater.173 (2010) 717–723.

17] U. Schwertmann, R.M. Cornell, Iron Oxides in the Laboratory, John Wiley & SonsInc., Germany, 2000.

18] R.M. McKenzie, The synthesis of birnessite, cryptomelane, and someother oxides and hydroxides of manganese, Mineral. Mag. 38 (1971)493–502.

19] ICDD (International Center for Diffraction Data), JCPDS PDF-2 Database, New-town Square, PA, USA, 2003.

20] A.E. Martell, R.M. Smith, R.J. Motekaitis, NIST SRD 46 Critically Selected StabilityConstants of Metal Complexes, Version 5.0, NIST, Gaithersburg, MD, USA, 1998.

21] C. Whalley, A. Grant, Assessment of the phase selectivity of the European com-munity bureau of reference (BCR) sequential extraction procedure for metalsin sediment, Anal. Chim. Acta 291 (1994) 287–295.

22] J. Slavek, W.F. Pickering, Extraction of metals ions sorbed on hydrous oxides ofiron(III), Water Air Soil Pollut. 28 (1986) 151–162.

23] M. Raksasataya, A.G. Langdon, N.D. Kim, Assessment of the extent of lead redis-tribution during sequential extraction by two different methods, Anal. Chim.Acta 332 (1996) 1–14.

24] G. Bidoglio, P.N. Bibson, M. O’Gorman, K.J. Roberts, X-ray absorption spec-troscopy investigation of surface redox transformations of thallium andchromium on colloidal mineral oxides, Geochim. Cosmochim. Acta 57 (1993)2389–2394.

25] A. Vanek, V. Chrastny, M. Mihaljevic, P. Drahota, T. Grygar, M. Komárek,Lithogenic thallium behavior in soils with different land use, J. Geochem. Explor.102 (2009) 7–12.

26] A.R. Jacobson, S. Klitzke, M.B. McBride, P. Baveye, T.S. Steenhuis, The desorptionof silver and thallium from soils in the presence of a chelating resin with thiolfunctional groups, Water Air Soil Pollut. 160 (2005) 41–54.

27] S. Xiao-Quan, C. Bin, Evaluation of sequential extraction for speciation of tracemetals in model soil containing natural minerals and humic acid, Anal. Chem.65 (1993) 802–807.

28] A. Neaman, B. Waller, F. Mouélé, F. Trolard, G. Bourrié, Improved methods forselective dissolution of manganese oxides from soils and rocks, Eur. J. Soil Sci.55 (2004) 47–54.

29] H.W. Martin, D.I. Kaplan, Temporal changes in cadmium, thallium, and vana-dium mobility in soil and phytoavailability under field conditions, Water AirSoil Pollut. 101 (1998) 399–410.