competitive adsorption behavior of heavy metals on kaolinite

cular metalII), Cu(II),lution ofsingle-

us metalsled using

formation of

pacitance

Journal of Colloid and Interface Science 290 (2005) 28–38www.elsevier.com/locate/jcis

Competitive adsorption behavior of heavy metals on kaolinite

Prashant Srivastavaa,∗, Balwant Singha, Michael Angoveb

a Faculty of Agriculture, Food and Natural Resources, The University of Sydney, NSW 2006, Australiab La Trobe University, P.O. Box 199, Bendigo, VIC 3552, Australia

Received 1 December 2004; accepted 8 April 2005

Available online 1 June 2005

Abstract

Polluted and contaminated soils can often contain more than one heavy metal species. It is possible that the behavior of a partispecies in a soil system will be affected by the presence of other metals. In this study we have investigated the adsorption of Cd(Pb(II), and Zn(II) onto kaolinite in single- and multi-element systems as a function of pH and concentration, in a background so0.01 M NaNO3. In adsorption edge experiments, the pH was varied from 3.5 to 10.0 with total metal concentration 133.3 µM in theelement system and 33.3 µM each of Cd(II), Cu(II), Pb(II), and Zn(II) in the multi-element system. The value of pH50 (the pH at which50% adsorption occurs) was found to follow the sequence Cu< Zn < Pb< Cd in single-element systems, but Pb< Cu< Zn < Cd in themulti-element system. Adsorption isotherms at pH 6.0 in the multi-element systems showed that there is competition among variofor adsorption sites on kaolinite. The adsorption and potentiometric titrations data for various kaolinite–metal systems were modean extended constant-capacitance surface complexation model that assumed an ion-exchange process below pH 7.0 and theinner-sphere surface complexes at higher pH. Inner-sphere complexation was more dominant for the Cu(II) and Pb(II) systems. 2005 Elsevier Inc. All rights reserved.

Keywords: Heavy metals; Cadmium; Copper; Lead; Zinc; Kaolinite; Competitive adsorption; Surface complexation modeling; Extended constant-ca

vi-encelin-redtive

a-arins iner bThemorl

ltitu-ve ad de-

nd-

avyviorvar-

, Cu,s inncan

edsCd,

model

1. Introduction

Growing concern about the quality of the natural enronment has stimulated increasing interest in the occurrand behavior of heavy metals in soils and water. Kaoite is the most abundant phyllosilicate in highly weathetropical soils[1], and possesses a small permanent negacharge[2–4]. It is a 1:1 aluminosilicate comprising a tetrhedral and an octahedral sheet bonded through the shof oxygen atoms between silicon and aluminum atomadjacent sheets. Successive 1:1 layers are held togethhydrogen bonding of adjacent silica and alumina sheets.permanent negative charge is produced because of isophic replacement of Si4+ by Al3+ in the silica tetrahedrasheet or of trivalent metal ions (such as Al3+) by divalent

* Corresponding author. Fax: +61 2 9351 5108.E-mail address: [email protected](P. Srivastava).

0021-9797/$ – see front matter 2005 Elsevier Inc. All rights reserved.doi:10.1016/j.jcis.2005.04.036

g

y

-

ions (such as Fe2+ and Mg2+) in the alumina octahedrasheet, leaving a single negative charge for each substion. Both the alumina sheet and the crystal edges hapH-dependent variable charge caused by protonation anprotonation of surface hydroxyl (SOH) groups[5]. Hence,the kaolinite surface is expected to have two kinds of biing sites that could interact with metal ions.

Contaminated soils often contain more than one hemetal, which can potentially impact the adsorption behaof each metal present as a result of competition amongious ions present in the system. Heavy metals such CdPb, and Zn occur commonly at elevated concentrationcontaminated soils[6]. Metals such as Cu and Zn are knowto be essential to plants, humans, and animals, but theyalso have adverse effects if their availability in soils excecertain threshold values. Other heavy metals, such as
Pb, and Hg, which are not essential to plants or animals, areknown to be hazardous to health, even at low concentrations.Among the important heavy metals, Pb is reportedly the least
http://www.elsevier.com/locate/jcis

mailto:[email protected]

id and

solare

aternic

able

co-anic

ndefeled

tleetals

dnce

firstreag tom-

onto-reeeac-ontoper-lex-

the

viorcur

sandWeatio

nerr ad

d-Ther-

orp.,r-

was

as

ion

entsthetiontra-ib-0.0m-(pHad-ur.at-

yzedFS

ountbe-

ns.pH

µMµMtra-heMberlin-

there-

ite(s) atsorp-etryen-ere10.0a-orpHas

aser-l

-

P. Srivastava et al. / Journal of Collo

mobile. Copper has been shown to remain bound up as inuble complexes in soil and sediment, whereas Zn and Cdconsiderably more mobile. Cadmium and Zn have a gretendency to dissociate from insoluble inorganic and orgacomplexes to form soluble ionic species that remain stat neutral or slightly alkaline pH[7].

Adsorption is arguably the most important of the physichemical processes responsible for the retention of inorgand organic substances in the soil environment[8]. Factorssuch as pH[9,10], nature and concentration of substrate aadsorbing ion[11,12], ionic strength[13], and the presencof competing and complexing ions[14], affect the extent oadsorption. Adsorption of heavy metals has been modon various minerals such as goethite[15,16], smectite[17–19], and illite[20] in single-element systems. However, litwork has been done to model the adsorption of heavy monto kaolinite in multi-element systems.

Schindler et al.[21] proposed that binding of Cu, Cd, anPb on kaolinite could be described by constant-capacitamodel that assumes two kinds of binding sites. Theadsorption sites adsorb metals by ion exchange, whethe second adsorption sites involve inner-sphere bindinampholytic SOH groups. This view was supported by coprehensive studies on adsorption of transition metalskaolinite [22]. Angove et al.[12,23]applied a surface complexation model to adsorption results obtained from thdifferent types of experiments to test possible surface rtions. Their results suggested that the Cd(II) adsorbskaolinite by two distinct processes: ion exchange at themanently charged sites on the silanol faces, and compation to aluminol and silanol groups, which occur atcrystal edges.

In the present work, we have studied adsorption behaof four divalent metals (Cd, Cu, Pb, and Zn), which ocmost commonly in contaminated soils[24] and have differ-ent hydrolysis behavior[25]. The objective of this study wato examine the adsorption of heavy metals (Cd, Cu, Pb,Zn) onto kaolinite in single- and multi-element systems.used an extended constant-capacitance surface complexmodel to describe the sorption, as it allows for both insphere and outer sphere complexation, but requires fewejustable parameters than triple layer models.

2. Materials and methods

Acid-washed kaolinite supplied by Ajax Chemicals, Syney, Australia was used without any further treatment.BET-N2 surface area[26] determined by a Quantasorb suface area analyzer (Model Autosorb-1, Quantachrome CNY) was 14.4 m2 g−1. The XRD pattern showed characteistic peaks of kaolinite, and no other mineral componentdetected.

Adsorption experiments as a function of pH (adsorptionedge) and concentration (adsorption isotherm) and potentio-metric titrations were conducted in a borosilicate reaction

Interface Science 290 (2005) 28–38 29

-

s

n

-

vessel at controlled room temperature (22± 1◦C) undernitrogen atmosphere. A sufficient mass of kaolinite wadded to give a surface area concentration of 96.3 m2 L−1

in 0.01 M NaNO3 background electrolyte. The suspenswas stirred overnight for 16 h at its natural pH (∼5.0) tohydrate the mineral surface. Adsorption edge experimwere conducted at 133.3 µM metal concentration insingle-element systems and at 33.3 µM metal concentraeach of Cd, Cu, Pb, and Zn (i.e., total metal concention of 133.3 µM) in multi-element systems. The equilrium pH of the suspension was varied from 3.5 to 1with a 0.5-pH unit increment using a Radiometer cobined pH electrode with Radiometer standard buffers4 and 7). The suspension was equilibrated for metalsorption for 1 h and an aliquot was collected every hoThe aliquots were centrifuged and filtered through Whman no. 1 filter paper and the supernatant was analfor the respective metal(s) using a Varian SpectrAA-220flame atomic absorption spectrophotometer. The amof metal adsorbed was calculated as the differencetween the initial- and equilibrium metal concentratioAdsorption isotherm experiments were conducted at6.0 with the metal concentrations from 16.7 to 950.0in the single-element systems and from 4.2 to 237.5each of Cd, Cu, Pb, and Zn (i.e., total metal concention from 16.7 to 950.0 µM) in multi-element system. TpH was maintained by addition of 0.1 M HCl and 0.1NaOH. Proton stoichiometry was measured as the numof protons released per metal ion adsorbed onto kaoite. The amount of 0.1 M NaOH added to maintainpH was used to account for the number of protonsleased.

Potentiometric titrations were performed on kaolinsuspensions in the absence and presence of metal ionthe same mineral and metal concentrations as in the adtion edge experiments to determine the proton stoichiomof the various surface reactions. After the kaolinite suspsion was equilibrated overnight, the metal ion(s) was/wadded and the system was titrated between pH 3.0 andusing 0.1 M HCl and 0.1 M NaOH, respectively, with a Rdiometer TIM800 autotitrator. After each addition of acidbase, the suspension pH was allowed to stabilize until thedrift was less than 0.01 units per minute. This criterion wtypically achieved within 30 min.

2.1. Aqueous speciation of metals

Aqueous speciation of metals as a function of pH wstudied using the computer program Visual MINTEQ, Vsion 2.30[27], which is a modified version of the originaMINTEQA2/PRODEFA2 program[28]. The solution spe
ciation of the metals was modeled because the hydrolysisbehavior of metal ions has been found to influence sorptionprocesses[29].

30 P. Srivastava et al. / Journal of Colloid and Interface Science 290 (2005) 28–38

Table 1Extended constant-capacitance surface complexation model parameters for adsorption of Cd, Cu, Pb, and Zn onto kaolinite using GRFIT[31]

Site density, XH (µmol m−2) 1.04Site density, SOH (µmol m−2) 6.93Specific inner capacitance,κ (F m−2) 7.0Specific outer capacitance,κ (F m−2) 3.0

Protonation reactions logK

XH + Na+ = XNa + H+ −3.30SOH+ H+ = SOH+

2 3.81SOH = SO− + H+ −6.16

Hydrolysis reactionsa Cd(II) Cu(II) Pb(II) Zn(II)

M2+ + H2O = MOH+ + H+ −9.60 −8.00 −7.71 −8.962M2+ + 2H2O = M2(OH)2+

2 + 2H+ −10.36M2+ + 2H2O = M(OH)2 + 2H+ −18.80 −17.30 −17.12 −16.9M2+ + 3H2O = M(OH)−3 + 3H+ −28.06 −28.4

3M2+ + 4H2O = M3(OH)2+4 + 4H+ −23.88

Surface reactions2XNa+ M2+ = X2M + 2Na+ 4.13 4.45 4.28 4.73

2+ +
2SOH+ M = (SO)2M + 2H −8.79 −7.06
ge,us-

atioiteon

inedith-to-

fixednts

theeseheneri-

atneg

toteergo

rp-

nt to

x

f thepo-per-ner

H

ave

us-

SOH+ MOH+ = SOMOH+ H+a From Baes and Mesmer[25].

2.2. Modeling the adsorption data

The data from potentiometric titration, adsorption edand adsorption isotherm experiments were modeleding an extended constant-capacitance surface complexmodel [30]. The values for equilibrium constants and sdensities (Table 1) were estimated by modeling adsorptiand titration data using the computer program GRFIT[31].Parameters for surface protonation reactions were obtaby modeling the titration data of kaolinite suspensions wout metal ions. The equilibrium constants for surface pronation reactions and site densities were then used asvalues for modeling the data from adsorption experimeand potentiometric titrations of kaolinite suspension inpresence of metal ions. The equilibrium constants for thadsorption reactions were deemed acceptable only wthese constants closely fitted the data from all three expments.

Earlier modeling studies by Schindler et al.[21] andIkhsan et al.[22] suggest that transition metals adsorbpermanent- and variable-charge sites. The permanent,atively charged sites, represented by X−, can undergo anexchange reaction,

X−- - -Na+ + H+ = X−- - -H+ + Na+.

In the absence of metal ions, the X− sites are assumedbe occupied by Na+ ions from the background electrolyat higher pH values. The variable charge sites can undboth protonation and deprotonation reactions,

SOH+ H+ = SOH+2 ,

SOH= SO− + H+,

where SOH represents a surface hydroxyl group. Althoughno distinction is made between aluminol and silanol surface

−8.13 −8.65

n

-

groups, it is likely that the SOH groups involved in adsotion are mostly aluminol (AlOH)[21].

For permanent charge sites, one reaction was sufficiemodel adsorption of all metals on kaolinite (Reaction 1),

M2+ + 2X−- - -Na+ = X2M + 2Na+, (1)

where M2+ represents Cd(II), Cu(II), Pb(II), or Zn(II), X− isa permanent charge site and X2M is an outer-sphere complein the model.

For variable charge sites, inner-sphere complexes oform SOMOH were required to give adequate fits totentiometric titration, adsorption edge, and isotherm eximents involving Cu and Pb (Reaction 2). Bidentate insphere complexes of the stoichiometry (SO)2M were re-quired for Cd and Zn uptake (Reaction 3),

M2+ + 2SOH= SOMOH+ 2H+, (2)

M2+ + 2SOH= (SO)2M + 2H+. (3)

Adsorption of various other species such as CdO+,Cu2+, Cu2(OH)2+

2 , Pb2+, Pb3(OH)2+4 , and ZnOH+ were

also tested, but only the stoichiometries shown above gthe best fit to all three sets of experimental data.

3. Results

3.1. Aqueous speciation of metals

The aqueous speciation of metals was determineding the hydrolysis constants from Baes and Mesmer[25].
The speciation of each of the metals in solution is simi-lar in single- (Fig. 1) and multi-element (Fig. 2) systems.Cadmium occurs predominantly as Cd2+ species over the

P. Srivastava et al. / Journal of Colloid and Interface Science 290 (2005) 28–38 31

the s ective met

ntras

O

tlyu-

cies,

O

pHbe-rvedpor-

enb

nt

sys-

irtire

s

n

so-

Znhe

ionh ofM)

avythe

Fig. 1. Aqueous speciation of Cd, Cu, Pb, and Zn as affected by pH inin the nitrate form and the background solution was 0.01 M NaNO3.

entire pH range studied in both the systems. The concetion of CdOH+ increases up to∼25% after pH 8.5, whereathe concentration of Cd(OH)2(aq) increases up to∼18% atpH 10.0. The concentration of other Cd species (CdN+

3 ,Cd(NO3)2(aq), Cd(OH)−3 , Cd(OH)2−

4 , and Cd2(OH)3+) isvery small in solution and does not change significanover the pH range. Copper occurs predominantly as C2+species up to pH∼ 6.0, after which its concentration decreases. Concentration of the various Cu hydroxyl speviz., CuOH+, Cu2(OH)2+

2 , Cu(OH)2−4 , Cu(OH)2(aq), and

Cu(OH)−3 , increases above pH∼ 6.0 in both the sys-tems. The concentration of other Cu species (CuN+

3 ,Cu(NO3)2(aq), Cu(OH)2−

4 , and Cu2(OH)3+) is negligiblein solution and does not significantly change over therange investigated. The difference in the Cu speciationtween the single- and the multi-element systems is obsein the concentrations of hydroxyl species, as greater protions of CuOH+, Cu(OH)−3 , and Cu(OH)2 (aq) are formed inthe multi-element system, as compared to the single-elemsystem. Similarly, Pb also occurs predominantly as P2+species up to pH∼ 6.0 in both systems. Besides Pb2+, a ni-trate species PbNO+3 is also present in a significant amou
up to pH 6.5, after which its concentration starts decreas-ing. Concentration of PbOH+ and Pb(OH)2 (aq) increasesin the multi-element system, as compared to the single-
ingle-element system. The solution consisted of 133.33 µM of the respal

-

t

element system. The proportion of Pb3(OH)2+4 is higher in

the single-element system than in the multi-elementtem. Small amounts of other Pb species (Pb(NO3)2(aq),Pb4(OH)4+

4 , and Pb2(OH)3+) also occur in solution, but theconcentration does not significantly change over the enpH range. Zinc occurs predominantly as Zn2+, and the con-centrations of ZnOH+ and Zn(OH)2 (aq) are more or lessimilar in both systems. The concentration of Zn2+ startsto decrease after pH∼ 7.5 in both the systems. Other Zspecies (ZnNO+3 , Zn(NO3)2(aq), Zn(OH)−3 , Zn(OH)2−

4 , andZn2(OH)3+) occur at negligible concentrations under thelution conditions of our experiments.

3.2. Adsorption edges

We investigated the adsorption of Cd, Cu, Pb, andonto kaolinite as a function of pH (adsorption edges). TpH was varied from 3.5 to 10.0 with a metal concentratof 133.3 µM in single-element systems and 33.3 µM eacCd, Cu, Pb, and Zn (i.e., total metal concentration 133.3 µin the multi-element system. The adsorption of the hemetals on kaolinite increased with increasing pH, with
shape of the curves dependent on the metal (Figs. 3 and4). In single-element systems, the adsorption edges for Cuand Pb were sigmoidal, whereas those for Cd and Zn were


the m u, Pb, and

le-

em

ache pHheddle

the

wathe-ed

Pbile

tionandre-s ofeaterCu

vari-the

theuilib-le-

ad-aredec-was.t at

Fig. 2. Aqueous speciation of Cd, Cu, Pb, and Zn as a function of pH inZn in the nitrate form and the background solution was 0.01 M NaNO3.

Table 2Values of pH50 for adsorption of Cd, Cu, Pb, and Zn onto kaolinite in singand multi-element systems

Metal Single-element system Multi-element syst

Cd 6.40 6.10Cu 5.30 5.30Pb 5.40 4.85Zn 5.35 5.80

characterized by two distinct adsorption stages, with estage in Cd adsorption edge separated by a plateau in thrange∼6.0 to ∼6.5. The plateau was more evident in tCd adsorption edge compared to Zn. A plateau in the miof adsorption edges was not observed for any metal inmulti-element system. The pH50 (i.e., the pH at which 50%of metal is adsorbed) shows that the selectivity sequenceCu> Zn > Pb> Cd in single-element systems, whereassequence was Pb> Cu> Zn > Cd in the multi-element system (Table 2). The pH50 for each of the metals also showthat, in the multi-element system, adsorption edges forand Cd were shifted to lower pH, Zn to a higher pH, whthat for Cu remained unchanged.

The surface speciation of the adsorption edges shows thaat lower pH, uptake of all four metals first occurred on per-manent charge sites (X−), and thereafter on variable charge

ulti-element system. The solution consisted of 33.33 µM each of Cd, C

s

sites in both single- and multi-element systems. Adsorpat permanent charge sites predominated up to pH 5.76.0 for Cu and Pb and pH 7.3 and 8.0 for Zn and Cd,spectively, in the single-element systems. The proportionCd and Zn adsorbed on permanent charge sites were grin the multi-element systems. The opposite was true forand Pb, where a greater proportion of metal sorbed atable charge sites in multi-element systems compared tosingle-element systems.

3.3. Adsorption isotherms

The adsorption isotherms at pH 6.0 indicate thatamount of each metal adsorbed increased as the eqrium concentration of the metal increased in both singand multi-element systems (Figs. 5 and 6). The isothermsshowed that Cu and Pb had similar but relatively highersorption capacities compared to Cd and Zn, which appeto have a lower affinity for the kaolinite surface. The seltivity sequence in the adsorption isotherm experimentsCu� Pb> Zn � Cd in single- and multi-element systems

The surface speciation of the isotherms shows tha
tpH 6.0, adsorption of Cu and Pb occurred predominantly onthe variable charge sites in both single- and multi-elementsystems; with a greater proportion being adsorbed in single-


Fig. 3. Surface speciation showing adsorption of Cd, Cu, Pb, and Zn onto kaolinite as a function of pH in the single-element system (metal concentratione ac alculated

accor

e in

on-Zn

bothwas

f Cd.

-face

n in

facethe

ab-r

romce

tra-ns,umedtto-e ofwaserexyl

133.33 µM and background solution 0.01 M NaNO3). The dots represent thfrom the extended constant-capacitance surface complexation model

Table 3Proton stoichiometry for adsorption of Cd, Cu, Pb, and Zn onto kaolinitsingle- and multi-element systems

Metal–kaolinite system Proton stoichiometry (χ)

Cd system 0.02Cu system 0.43Pb system 0.34Zn system 0.03Multi-element system 0.47

element systems than in the multi-element system. In ctrast to Cu and Pb adsorption, the adsorption of Cd andoccurred predominantly on permanent charge sites insystems. The adsorption of Zn on variable charge sitesgreater in the single-element system, while adsorption oon variable charges sites was negligible in both systems

3.4. Proton stoichiometry

The proton stoichiometry (Table 3) is the number of protons released per metal ion adsorbed on the kaolinite sur
It shows that more protons were released from the adsorptionsites upon Cu and Pb adsorption compared to the case foCd and Zn adsorption in single-element systems. The proton
tual experimental data and the lines represent the modeled adsorption cding to the parameters inTable 1.

.

stoichiometry was greater in multi-element systems thasingle-element systems.

3.5. Model parameters

Surface complexation parameters for the kaolinite sur(site densities and acidity constants) were obtained frompotentiometric titrations on kaolinite suspensions in thesence of heavy metals (Table 1). The titration data, togethewith lines showing the fit of the model, are shown inFig. 7.The hydrolysis constants for the metals were obtained fBaes and Mesmer[25]. The extended constant-capacitansurface complexation model fitted the potentiometric tition data for kaolinite closely. For kaolinite suspensiowhere no metal was added, the surface reactions asswere ion exchange between H+ and Na+ on permanencharge (X−) sites, together with protonation and depronation of variable charge (SOH) sites. In the presencmetal(s), a bidentate exchange reaction (Reaction 1)required to account for the sorption below pH 6.5, whas reactions involving surface complexation with hydro

rgroups (Reactions 2 and 3), were required to model the ad-sorption of the metals at higher pH. For the Zn system anadequate fit could be obtained assuming the same reaction


Fig. 4. Surface speciation showing adsorption of Cd, Cu, Pb, and Zn onto kaolinite as a function of pH in the multi-element system (metal concentrationsent sorption

tion m

inede Cd/Pbom-fitslti-

tedila

for

atm-

t ansible

is

isd tobe-

use,

sig-e Pbact

gle-urredrac-ater0. In

Pbb in-

33.33 µM each and background solution 0.01 M NaNO3). The dots reprecalculated from the extended constant-capacitance surface complexa

set as for Cu and Pb; however, the best fit was obtaassuming the scheme proposed for the Cd system. Thsystem could not be modeled effectively using the Cureaction scheme (Reaction 2). The fit of the surface cplexation model was generally good, but less satisfactorywere obtained for the Cd isotherms, particularly in the muelement system.

4. Discussion

4.1. Adsorption of metals in single-element system

The adsorption edges for the four metals investigahave a similar shape and occur across a pH range simto those reported previously for various surfaces[11]. Theadsorption edges (Fig. 3) for Zn and Cd, in particular, showevidence of a plateau region occurring between pH∼ 6.0and∼6.5. This behavior has been described previouslythe adsorption of various cations onto kaolinite[11,12,20–23] and is thought to result from initial uptake of metallow pH onto exchange sites of the kaolinite prior to co
plexation with hydroxyl edge sites at higher pH.
In modeling these systems we have adopted the approacof previous studies[12,20–23] and assumed that ion ex-

the actual experimental data and the lines represent the modeled adodel according to the parameters inTable 1.

r

change is the dominant mechanism at low pH, and thaexchange process represented by Reaction 1 is responfor metal uptake. The surface species X2M is an outer-spherecomplex, with the best fit obtained when the complexbidentate; a monodentate species XM+ could not model thedata as effectively. It is probably unrealistic to think of thspecies as a divalent metal ion specifically coordinatetwo sites; rather it represents the electrostatic attractiontween a divalent ion that neutralizes an opposite, but diffcharge at the surface. The surface speciation given inFig. 3shows that interaction with permanent charge is morenificant in the Zn and Cd systems. This may be becausand Cu hydrolyze at lower pH and are more likely to interwith edge hydroxyl sites than Cd or Zn.

The adsorption edges for the metal ions in the sinelement systems showed that the Cu and Pb edges occat lower pH than the Cd and Zn edges. The stronger intetion of Cu and Pb at lower pH is also reflected by the greamounts adsorbed in the isotherm experiments at pH 6.addition, the proton stoichiometry (Table 3) measured in theisotherm experiments is significantly higher for Cu andand supports the suggestion that adsorption of Cu and P

hvolves significant hydrolysis and/or interaction with surfacehydroxyls at pH values where Cd and Zn interact only withexchange sites. The selectivity series for these metals fol-


Fig. 5. Surface speciation showing adsorption of Cd, Cu, Pb, and Zn onto kaolinite as a function of concentration at pH 6.0 in the single-element system.sent t itance s

Pbter-orsve aisanditerp-

e-minus

hadityeand.ti-oilsan

ttionncemest an

t oc-anuter-

in a

theith

orp-et-

tion.d to

ingsur-phichere

The dots represent the actual experimental data and the lines reprecomplexation model according to the parameters inTable 1.

lows the tendency for the metal to hydrolyze. Copper andhydrolyze more readily and hence are more likely to inact with a hydroxylated surface. A number of investigathave shown that metals that hydrolyze more readily halower pH50 [9,31,32]. The selectivity sequence found in thstudy generally agrees with previous research work. PulsBohn[33], for example, found that Zn sorbed to the kaolinsurface at a lower pH than Cd. A study of metal adsotion onto goethite by Forbes et al.[34] showed a selectivitysequence Cu> Pb> Zn > Cd, which corresponds to the rsults obtained in this study. On the other hand, Benjaand Leckie[35], who studied metal uptake by amorphoiron oxyhydroxide, and Bidappa et al.[36], in a study ofmetal sorption by various Japanese soils, found that Pba higher affinity for the surface than Cu, with the selectivseries Pb> Cu> Zn > Cd for both studies. The differencobserved in the selectivity between this present studythe work of Benjamin and Leckie[35] and Bidappa et al[36], is probably a function of the different surfaces invesgated, although the interpretation of the study on whole sby Bidappa et al.[36] is complicated by the presence of
organic matrix.
The modeling generally provided satisfactory fits to theadsorption edges, isotherms and potentiometric titrations.

he modeled adsorption calculated from the extended constant-capacurface

Schindler et al.[21], Angove et al.[12,23], and Ikhsan eal. [22] proposed a bidentate binding scheme for adsorpof metal ions onto kaolinite, using a constant-capacitamodel. However, the constant-capacitance model assuthat all complexes are inner-sphere and is perhaps noideal representation of the ion exchange process thacurs on the kaolinite surfaces. In this study we appliedextended constant-capacitance model that assumes osphere complex formation in the exchange reactionsmanner analogous to Lackovic et al.[20]. While the modelproposed here is simple, it provides an adequate fit tothree different sets of experiments for all metals studied wminimum adjustable parameters. The fit for the Cd adstion isotherm was not as good as observed for other mals, as the model slightly underestimated the Cd adsorpThere are undoubtedly other models that could be usemodel these data. The CD-MUSIC approach[37,38], for ex-ample, potentially provides a more intellectually satisfydescription of oxide surfaces, where site densities andface acidity constants can be estimated from crystallograand PZC data. In the case of clay systems, however, w
surfaces are weathered and quite heterogeneous, we haveopted for a simple approach, where the surface reactions pro-posed represent an “average” of all the processes involved.


Fig. 6. Surface speciation showing adsorption of Cd, Cu, Pb, and Zn onto kaolinite as a function of concentration at pH 6.0 in the multi-element system.sent t itance s

ac-ntalns.enttion

d by-5.0of

.0.e ofCuxes

tementateatem

6.5em,the

ptakeent

r ofys-ore

fromy toelyore

n in-s are

e inakeu islex-f the

The dots represent the actual experimental data and the lines reprecomplexation model according to the parameters inTable 1.

In our modeling we have attempted to determine the retion stoichiometries that best fitted the pooled experimedata, and hence to identify the dominant surface reactio

An interesting feature of the modeling was the differreaction schemes that were required to model the sorpof the metals. For Cu and Pb, the best fits were obtaineassuming a complex SOMOH.Fig. 3 shows that this surface species becomes significant at pH values aboveFor Zn and Cd, the model required a surface complexthe form SO2M, which becomes important above pH 7This difference in the surface speciation is again reflectivthe different hydrolysis behavior of the metals, with theand Pb surface species occurring as hydrolyzed complewhile Zn and Cd adsorb as divalent ions.

4.2. Adsorption of metals in multi-element systems

The adsorption of the metals in the multi-element sysdiffered from their corresponding behavior in single-elemsystems. Most notable was the disappearance of the plregion in the Cd sorption edge in the multi-element sys
(Fig. 4). The surface speciation curve for Cd adsorption re-veals that the adsorption in single-element systems predominantly occurred on permanent negatively charged sites in
he modeled adsorption calculated from the extended constant-capacurface

.

,

u

the first stage up to pH∼ 6.0 (Fig. 3). The adsorption of Cdonto the variable charged hydroxyl edges began at pHin the single-element system. In the multi-element systhowever, Cd adsorption took place predominantly onpermanent charge surface up to pH 7.5. The increased uof Cd by the permanent charge surface in the multi-elemsystem might be explained by considering the behaviothe other metals present. Modeling in the multi-element stem reveals that metals that form hydrolysis products mreadily (Pb and Cu) adsorb to variable charge surfacesabout pH 5.0 and above. Because of its lower tendencform hydrolysis products, Cd does not compete effectivfor variable charge surfaces, and so its adsorption is mrestricted to permanent charge sites; hence we see acreased uptake of Cd at these sites when other metalpresent. A similar effect is observed for Zn. Atanassova[39]reported that, in a multi-component system, an increasthe Cu concentration resulted in a reduction in the uptof other heavy metals such as Ni, Cd, and Zn. Since Cpredominantly specifically adsorbed (inner-sphere compation), it can be expected that increasing the amounts o

-more strongly bonded Cu reduced the number of sites avail-able for Cd and Zn adsorption. Just as the presence of Cu andPb suppressed the uptake of Cd and Zn by variable charge


Fig. 7. Potentiometric titrations of kaolinite suspension, alone (a) and with added metals (Cd (b), Cu (c), Pb (d), and Zn (e)) [(133.33 µM in the single-elementts re the exrameT

ptakd ar, it

ackistrye, itrmsowsrely

fuln inse-geolv-

systems) and (33.33 µM each in the multi-element system) (f)]. The doconstant-capacitance surface complexation model according to the pa

surfaces, the presence of Cd and Zn decreased the uof Cu and Pb onto exchange sites. The model providegood fit to the adsorption edge data for all metals; howevefailed to adequately fit the Cd adsorption isotherm. The lof fit in the multi-element system suggests that the chemof the sorption process for Cd is not simple. For examplcould be that a multi-element sorbate or co-precipitate foat the surface. Inclusion of such species in the model allus to fit the adsorption data, but we feel that this is me
a curve-fitting exercise, and without other experimental ev-idence, the introduction of additional model parameters isunjustified. Direct methods such as extended X-ray absorp-
present the actual experimental data and the lines are calculated fromtendedters inable 1.

etion fine structure spectroscopy (EXAFS) might be useto ascertain the mechanism of heavy metal adsorptiomulti-element systems. This technique might provide uful information on how the surface speciation might chanwhen competing metals were present, and if species inving two or more different metal cations also formed.

5. Conclusions

The adsorption behavior of Cd, Cu, Pb, and Zn ontokaolinite in single- and multi-element systems differs. The

d and

met,

-was

edosethento

sorpby

andandxescurs

ationad-theentom-ermEX-iredtals

pordu-ia.

35.

oilof

nt:n En-ical

lants,

33.204

ch-

1)

im.

im.

oc.

. J.

ette,

er-

87)

217

7)

20

rk,

-ds,

forring,

2,Ver-

ncy,

s,hys-

6)

9.art

99)

ce

38 P. Srivastava et al. / Journal of Colloi

selectivity sequence of the adsorption edges of theseals was Cu> Pb > Zn > Cd in single-element systemwhile it was Pb> Cu> Zn > Cd in the multi-element system. For adsorption isotherms, the selectivity sequenceCu > Pb > Zn�Cd in single-element systems and Cu>

Pb > Zn > Cd in multi-element systems. The extendconstant-capacitance surface complexation model propthree kinds of possible reaction mechanism to explainadsorption behavior. At low pH, all these metals adsorb opermanent charge sites by ion exchange reactions. Adtion onto variable charge sites takes place at higher pHforming inner-sphere complexes at the crystal edgesoctahedral alumina faces. The hydroxyl species of CuPb adsorb by forming monodentate inner-sphere complewhereas adsorption of Cd and Zn on variable charges ocby forming bidentate complexes.

The extended constant-capacitance surface complexmodel fitted the data adequately to different kinds ofsorption and potentiometric titration experiments withexception of the Cd adsorption isotherm in the multi-elemsystem. The model probably needs to include for more cplex speciation to account for the Cd adsorption isothin the multi-element system. Direct methods such asAFS and other spectroscopic techniques might be requto ascertain the mechanism for adsorption of heavy meon kaolinite particularly in multi-element systems.

Acknowledgments

Prashant Srivastava acknowledges the financial supfrom the University of Sydney and the Department of Ecation, Science and Technology, Government of Austral

References

[1] B. Singh, R.J. Gilkes, J. Soil Sci. 43 (1992) 645.[2] R.K. Schofield, H.R. Samson, Discuss. Faraday Soc. 18 (1954) 1[3] M.E. Sumner, Clay Miner. Bull. 5 (1963) 218.[4] M.B. McBride, Clays Clay Miner. 24 (1976) 88.[5] G.N. White, J.B. Dixon, in: J.B. Dixon, D.G. Schulze (Eds.), S

Mineralogy with Environmental Applications, Soil Science SocietyAmerica, Madison, WI, 2000.

[6] B. Singh, in: D. Smith, S. Fityus, M. Allman (Eds.), GeoEnvironmeProceedings of the 2nd Australia and New Zealand Conference ovironmental Geotechnics, Newcastle, NSW, Australian GeochemSociety, Newcastle, 2001, pp. 77–93.

[7] A. Kabata-Pendias, H. Pendias, Trace Elements in Soils and PCRC Press, Boca Raton, FL, 1984.

[8] M.E. Essington, Soil and Water Chemistry: An Integrative Approach,CRC Press, Boca Raton, FL, 2004.

Interface Science 290 (2005) 28–38

-

d

-

,

t

[9] R.D. Harter, Soil Sci. Soc. Am. J. 47 (1983) 47.[10] J.J. Msaky, R. Calvet, Soil Sci. 150 (1990) 513.[11] K.M. Spark, J.D. Wells, B.B. Johnson, Eur. J. Soil Sci. 46 (1995) 6[12] M.J. Angove, B.B. Johnson, J.D. Wells, J. Colloid Interface Sci.

(1998) 93.[13] J.A. Dyer, P. Trivedi, N.C. Scrivner, D.L. Sparks, Environ. Sci. Te

nol. 37 (2003) 915.[14] M.M. Benjamin, J.O. Leckie, Environ. Sci. Technol. 15 (198

1050.[15] L. Lovgren, S. Sjoberg, P.W. Schindler, Geochim. Cosmoch

Acta 54 (1990) 1301.[16] L. Gunneriusson, L. Lovgren, S. Sjoberg, Geochim. Cosmoch

Acta 58 (1994) 4973.[17] J.M. Zachara, S.C. Smith, J.P. McKinley, C.T. Resch, Soil Sci. S

Am. J. 57 (1993) 1491.[18] C. Arpa, R. Say, N. Satiroglu, S. Bektas, Y. Yurum, O. Genc, Turk

Chem. 24 (2000) 209.[19] D.G. Strawn, N.E. Palmer, L.J. Furnare, C. Goodell, J.E. Amon

R.K. Kukkadapu, Clays Clay Miner. 52 (2004) 321.[20] K. Lackovic, M.J. Angove, J.D. Wells, B.B. Johnson, J. Colloid Int

face Sci. 257 (2003) 31.[21] P.W. Schindler, P. Leichti, J.C. Westall, Neth. J. Agric. Sci. 35 (19

219.[22] J. Ikhsan, B.B. Johnson, J.D. Wells, J. Colloid Interface Sci.

(1999) 403.[23] M.J. Angove, B.B. Johnson, J.D. Wells, Colloid. Surf. A 126 (199

137.[24] R. Naidu, M.E. Sumner, R.D. Harter, Environ. Geochem. Health

(1998) 5.[25] C.F. Baes, R.E. Mesmer, Hydrolysis of Cations, Wiley, New Yo

1976.[26] D.L. Carter, M.M. Mortland, W.D. Kemper, in: A. Klute (Ed.), Meth

ods of Soil Analysis. Part I. Physical and Mineralogical Methosecond ed., American Society of Agronomy, Madison, WI, 1986.

[27] J.P. Gustafsson, Visual MINTEQ Ver. 2.30: A Computer ProgramSpeciation, Department of Land and Water Resources EngineeKTH, Sweden, 2004.

[28] J.D. Allison, D.S. Brown, K. Novo-Gradac, MINTEQA2/PRODEFAa Geochemical Assessment Model for Environmental Systems,sion 3.0, EPA/600/3-91/021, U.S. Environmental Protection AgeAthens, GA, 1991.

[29] R.O. James, T.W.J. Healy, Colloid Interface Sci. 40 (1972) 42.[30] S. Goldberg, Adv. Agron. 47 (1992) 233.[31] C. Ludwig, GRFIT: A Program, for Solving Speciation Problem

Evaluation of Equilibrium Constants, Concentrations and Their Pical Parameters, The University of Berne, Switzerland, 1992.

[32] H.A. Elliott, M.R. Liberati, C.P. Huang, J. Environ. Qual. 15 (198214.

[33] R.W. Puls, H.L. Bohn, Soil Sci. Soc. Am. J. 52 (1988) 1289.[34] E.A. Forbes, A.M. Posner, J.P. Quirk, J. Soil Sci. 27 (1976) 154.[35] M.M. Benjamin, J.O. Leckie, J. Colloid Interface Sci. 79 (1981) 20[36] C.C. Bidappa, M. Chino, K. Kumazawa, J. Environ. Sci. Health P

B 16 (1981) 511.[37] T. Hiemstra, W.H. Van Riemsdijk, J. Colloid Interface Sci. 210 (19

182.[38] P. Venema, T. Hiemstra, W.H. Van Riemsdijk, J. Colloid. Interfa

Sci. 183 (1996) 515.[39] I.D. Atanassova, Environ. Pollut. 87 (1995) 17.

competitive adsorption behavior of heavy metals on kaolinite

Documents