Water Research 37 (2003) 16191627

Adsorption of heavy metals on Na-montmorillonite. Effect of pH and organic substances

O. Abollinoa, M. Acetob, M. Malandrinoa, C. Sarzaninia, E. Mentastia,*
b

Department of Analytical Chemistry, University of Torino, Via Giuria 5, 10125 Torino, Italy Department of Sciences and Advanced Technologies, University of East Piedmont, Corso Borsalino 54, 15100 Alessandria, Italy Received 6 August 2001; received in revised form 3 June 2002; accepted 16 October 2002

Abstract Clays (especially montmorillonite and bentonite) are widely used as barriers in landlls to prevent contamination of subsoil and groundwater by leachates containing heavy metals. For this reason it is important to study the adsorption of metals by these clays. The sorption of seven metals (Cd, Cr, Cu, Mn, Ni, Pb and Zn) on Na-montmorillonite was studied as a function of pH and in the presence of ligands, forming complexes of different stabilities with the metals of interest. The continuous column method was used as it better simulates natural conditions. The total capacity of Namontmorillonite towards these metals was determined. The pH variations inuence to a higher extent the concentrations of Cu, Pb and Cd in the efuent. Moreover the results suggest that complex formation hinders the sorption of the metals on the clay, with an increasing inuence in the order: Mn p Pb p Cd p Zn o Ni o Cu o Cr. The evaluation of the total capacity of Na-montmorillonite shows that this clay is a good sorbent towards all examined metals. r 2002 Elsevier Science Ltd. All rights reserved.
Keywords: Na-montmorillonite; Heavy metals; Adsorption; Metal complexes; Breakthrough

1. Introduction The concentration and the mobility of heavy metals in soils and sediments have been widely studied in the last decades [1,2]. Although many heavy metals are necessary in small amounts for the normal development of the biological cycles, most of them become toxic at high concentrations. Heavy metals are introduced into the environment through natural phenomena and human activities, such as agricultural practices, transport, industrial activities and waste disposal [3]. Clay linings have been used as barriers in landlls to prevent contamination of groundwater and subsoil by leachates containing metals. Generally, these linings are constituted of bentonite and, in particular, montmor*Corresponding author. Tel.:+39-11-6707625; fax: +39-116707615. E-mail address: edoardo.mentasti@unito.it (E. Mentasti).

illonite. These clays are chosen to avoid pollutant release into the environment owing to their high specic surface areas, low cost and ubiquitous presence in most soils [4]. Montmorillonite can adsorb heavy metals via two different mechanisms: (1) cation exchange in the interlayers resulting from the interactions between ions and negative permanent charge and (2) formation of inner-sphere complexes through SiO and AlO groups at the clay particle edges [57]. Both mechanisms are pH dependent because in acid conditions (pH o 4) most silanol and aluminol groups are protonated; therefore, in particular for the latter, an acidication can lead to an increase in mobility of metals bound to soil [8]. For this reason, it is necessary to improve the knowledge of the effect of pH on the sorption capacity of montmorillonite in soilsolution system. The continuous column method, instead of a conventional batch technique, was used to understand the effect of pH and of the presence of ligands able to originate

0043-1354/03/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 0 4 3 - 1 3 5 4 ( 0 2 ) 0 0 5 2 4 - 9

1620

O. Abollino et al. / Water Research 37 (2003) 16191627

metal complexes on the sorption capability of montmorillonite. Batch technique, in fact, is very useful to study the reaction mechanisms happening at the soil/ water interface but has the following limitations: solubilisation of soil components due to soil sample agitation and soil/solution ratios very different from those existing in natural systems [9]. The used technique, not having these disadvantages, better simulates the real natural conditions. With this method, the total capacity of montmorillonite towards the considered heavy metals has been determined under the considered conditions [10].

2. Experimental 2.1. Materials, reagents and instruments The sodium form of montmorillonite was supplied as montmorillonite KSF by Aldrich. The element composition of the clay was determined by ICP-AES (Inductively Coupled Plasma Atomic Emission Spectrometry) after solubilisation by acid digestion in microwave oven and is reported in Table 1. The percentages of SiO2 and Al2O3, principal constituents of montmorillonite, are 48.57 and 34.95, respectively. In order to respect the natural characteristics of montmorillonite and to evaluate the retention behaviour in landll liners, the clay was not treated. The surface area is 2040 m2 g1 and the cation exchange capacity (CEC) is B30 meq/100 g, therefore, they are lower than those generally reported for montmorillonites [11], but this is probably due to the lack of treatment. In fact, it is possible that, in this way, N2 molecules (employed to calculate the specic surface area by BET method) and cations can be introduced less easily in the interlayer regions as indicated by Barbier et al. [12]. All the reagents used were of analytical grade. All metal solutions were prepared from concentrated stock solutions (Merck Titrisol). High-purity water (HPW) produced with a Millipore Milli-Q system was used throughout. A Delta 320 Mettler Toledo pH meter provided with incorporated thermal probe was used for pH measurements; two standard buffer solutions at pH 4.0 and 7.0 were employed for calibration. Metal determinations were performed with a Varian Liberty 100 model Inductively Coupled Plasma Atomic

Emission Spectrometer (ICP-AES). The instrument is provided with a Czerny-Turner monochromator, a Sturman-Masters spray chamber, a V-groove nebuliser and a radio frequency (RF) generator at 40.68 MHz. The instrumental conditions were: plasma power 1.0 kW; nebuliser pressure 150 kPa; sample aspiration rate 15 rpm (B2 ml /min1); argon auxiliary ow 1.5 / min1. A Gilson Minipuls 2 multichannel peristaltic pump was used to drive the sample solutions into the columns with a constant and reproducible ow rate (0.5 cm3 min1). This pump was connected to the columns by polypropylene low-pressure ttings and PVC tubes (i.d. 1.29 mm). The polypropylene columns (Bio-Rad), 4 cm high and 5 mm i.d., were packed with 1 g of Na-montmorillonite. A Milestone MLS-1200 Mega microwave laboratory unit was used for the dissolution of the clay.

3. Procedure The continuous column method was used to study the adsorption of heavy metals on Na-montmorillonite. The apparatus for continuous ion exchange is schematised in Fig. 1. At the bottom of the column, on the porous polymer bed support, a 0.45 mm cellulose acetate lter (Millipore) was placed. The slurry packing technique was used to obtain a homogenous bed in the column. Before each experiment the clay packed in the column was conditioned at the working pH with a buffer solution. Solutions were driven through the columns with the aid of a peristaltic pump. Six replicates of each experiment were performed. Blanks were simultaneously run in order to take account of possible contaminations and eventual releases of the investigated metals from the clay. In the study of pH and ligand effect, 25 ml of the eluate was collected. The Cd, Cu, Cr, Mn, Ni, Pb and Zn concentrations in the inuent and efuent solutions were determined using an atomic emission spectrophotometer (ICP-AES). The amount of every metal adsorbed on to the clay was calculated by the difference between the content of metal in inuent solution and that one in efuent solution, corrected with the blank, and it was expressed in percentage. The pH range studied was 2.58.0. The NaAcetate/ Acetic Acid buffer was used in the pH interval between 2.5 and 5.5, while the HEPES-Na [4-(2-hydroxyethyl)-

Table 1 Element composition of the Na-montmorillonite determined by ICP-AES Element g/kg Cd o 0.01 Cr 0.03 Cu 0.01 Fe 32.50 K 9.71 Mg 12.64 Mn 0.18 Na 1.93 Ni o 0.03 Pb o 0.07 Ti 1.78 Zn 0.01 Zr 0.12

O. Abollino et al. / Water Research 37 (2003) 16191627

1621

Inlet

4. Results and discussion 4.1. Effect of pH The inuence of pH on the adsorption of bivalent and trivalent metal ions (Cd, Cr, Cu, Mn, Ni and Pb) on Namontmorillonite was studied. The results obtained are represented in Fig. 2. As expected, the adsorption of metals decreases with decreasing pH because the aluminol and silanol groups are more protonated and, hence, they are less available to retain the investigated metals. This effect is strongly evident for Cu, Pb and Cd, and less pronounced for the others. The reason of this behaviour is that the surface complexation reactions are inuenced also by the electrostatic attraction between the surface charge and the dissolved ions. In fact, since cadmium and lead have larger ionic radius (0.97 and ( respectively), they have lower charge density and, 1.20 A therefore, are more affected by the protonation of the surface groups that determine a reduction of the adsorption sites on clay. The Cu(II) behaviour, instead, is probably due to the structure of its aquaion. In fact, [Cu(H2O)6]2+ has a tetragonal distortion due to the JahnTeller effect in which the octahedral structure has been contracted along the x and y axes [13]. Therefore, the binding of ligands along the x and y axis is supported whereas the ligands along the zaxis are shielded from the Cu2+ ion by an extra electron. This contraction along the x and y axis results in a structure having four shorter bonds and two longer bonds and this hinders in part the binding of the copper (II) aquaion with the clay surface groups. This effect is more evident when these groups are more protonated because the binding is supported only along particular directions (x and y axis). For these reasons, hence, the adsorption of Cu, Pb and Cd is hindered by cation exchange mechanism whereas this mechanism dominates the adsorption of Cr, Mn, Ni and Zn ions. These

i.d. 5 mm

4 cm

Na-montmorillonite

0.45 m cellulose acetate filter porous polymer bed support polypropylene low pressure fitting

Outlet
Fig. 1. Schematic representation of continuous ion exchange column system.

piperazine-1-ethanesulphonic acid, sodium salt] buffer was used to obtain pH 8.0. All buffer solutions were prepared at 1.0 101 M concentration; the concentration of every metal was always kept to 1.0 104 M in order to minimise the competitions among the considered metals towards the adsorption of montmorillonite. Ligand stock solutions were prepared in water. The following ligands were investigated: nitrilotriacetic acid (NTA), ethylenediaminetetraacetic acid (EDTA), oxalic acid, malonic acid, succinic acid, glutaric acid, tartaric acid and citric acid. The effect of metal complexation on adsorption was studied with sample solutions buffered at pH 5.5 containing 1.0 103 M ligand and 1.0 104 M metal ions. The evaluation of the total capacity of Na-montmorillonite towards every studied metal was performed with the breakthrough technique generally used in chromatography to measure the highest degree of column utilisation in a given process. The solutions containing a single investigated metal at a concentration of 1 103 M were driven, with pH controlled at 5.5, through the column packed with the clay and successive aliquots of 25 ml of the efuent were collected and analyzed.

Fig. 2. Adsorption of Cd, Cr, Cu, Mn, Ni, Pb and Zn on Namontmorillonite as a function of pH (initial concentration of metals 1.0 104 M).

1622

O. Abollino et al. / Water Research 37 (2003) 16191627

factors contribute to a different pH effect on each metal. At pH p 3.5, the studied metals are increasingly adsorbed in the following order: Cu2+ o Pb2+ o Cd2+ o Zn2+ p Mn2+ D Cr3+ D Ni2+. 4.2. Effect of ligands The soils and the leachates coming from rubbish dump generally contain also organic substances beyond heavy metals. For this reason, it is very important to know the behaviour of heavy metals in the presence of organic ligands toward their adsorption on Na-montmorillonite, often used as linings to keep the pollutants. Using again the continuous column method, the simultaneous adsorption of Cd, Cr, Cu, Mn, Ni, Pb and Zn in the presence of ligands having different

complexation constants was evaluated. The equilibrium constants of these ligands are reported in Table 2. The concentration of ligands was chosen so as to be in excess with respect to that of metals and the experiments were executed at pH 5.5 to assure the formation of Meligand complexes and to minimise the one of hydroxo-species. The results are reported in Fig. 3 and Table 3. The equilibrium constants are not all known in the literature [14,20,21]. From the results it is evident that the presence of ligands having high complexation constants, that is EDTA and NTA, strongly hinders the adsorption of all metals on the clay. In fact the adsorption percentages vary between 0.1% for NiNTA complex and 36% for MnNTA complex. This behaviour, moreover, indicates that the adsorption of the free metal ion is supported with respect to that of MeNTA/EDTA complex. This

Table 2 Equilibrium constants of the considered ligands [14,20,21]a,b Cd EDTA pka1=2.00; pka2=2.69; pka3=6.13; pka4=10.19 NTA pka1=1.9; pka2=2.48; pka3=9.65 Meso-tartatic acid pka1=2.97; pka2=4.49 Oxalic acid pka1=1.04; pka2=3.81 A 16:54 B 9:07 Cr A 12:8 B 6:1 Cu A 18:7 B 11:91 C 6:70 A 13:1 B 3:39 A 3:15 B 2:06 A 4:49 B 3:92 D 9:54 Mn A 14:05 B 5:47 Ni A 18:4 B 11:56 Pb A 18:0 B 9:68 C 6:22 A 11:4 B 3:99 A 3:09 B 2:09 A 4:16 B 1:43 D 6:33 Zn A 16:44 B 9:00

A 9:4 B 4:9 A 1:30 B 0:80 C 0:76 A 2:78 B 1:22 C 0:90 D 4:1 E 5:1 A 2:64 B 1:49 C 1:04

A 11:26

A 10:0 B 3:5 A 2:2c B 1:0c

A 2:49c B 1:21c A 2:15 B 1:90 C 1:75

A 3:01 B 2:03 A 3:83 B 3:23 D 6:6

A 3:48 B 2:00 C 1:60 A 2:95 B 0:99

Malonic acid pka1=2.63; pka2=5.28

A 5:04 B 2:08 C 7:8

A 2:5 C 1:24

A 3:29 B 1:04

Succinic acid pka1=4.00; pka2=5.24

Glutaric acid pka1=4.13; pka2=5.01

A 1:47 B 0:82 C 0:45 D 2:29 E 2:74 A 2:0

A 2:7 B 1:85

A 1:26

A 1:6

A 2:4

A 1:13

A 1:6

Citric acid pka1=2.90; pka2=4.35; pka3=5.30

a

A 3:65 B 2:19 C 1:1 D 4:54

A 5:90 B 3:7 C 2:26

A 3:79 B 2:22 C 1:5

A 5:35 B 3:25 C 1:75

A 2:60 B 1:02 C 0:70 D 4:20 E 4:16 A 2:40 B 1:33 C 0:38 D 3:99 E 3:89 A 2:48 B 0:97 C 0:45 D 3:77 A 4:44 B 2:98 C 1:70 D 5:92

A 2:33 B 2:14

A 1:6

A 4:76 B 2:78 C 1:3 D 5:90

Ka=acid dissociation constant; b=cumulative stability constant. A=log b (ML); B=log b (MHL); C=log b (MH2L); D=log b (ML2); E=log b (ML3). c Cumulative stability constant of D-tartaric acid.
b

O. Abollino et al. / Water Research 37 (2003) 16191627

120 100

1623

Adsorption (%)

80 60 40 20 0

EDTA NTA Oxalic Malonic Succinic Glutaric Citric Tartaric

Ni

Mn

Cr

Zn

Cu

Pb

Cd

Metal
Fig. 3. Metal adsorption on montmorillonite in the presence of different ligands (pH 5.5; metal concentrations 1.0 104 M; ligands concentrations 1.0 103 M).

Table 3 Percentages of adsorbed metals on Na-montmorillonite in the presence of a ligand EDTA Cd Cr Cu Mn Ni Pb Zn 9.9 24.7 9.9 10.3 0.3 9.8 8.5 NTA 7.1 26.7 5.4 35.9 0.1 10.0 5.9 Tartaric acid 91.5 29.8 94.9 98.6 49.4 99.1 65.3 Oxalic acid 90.8 25.5 14.7 99.5 63.9 97.7 93.9 Malonic acid 86.1 35.2 24.7 100 81.2 100 100 Succinic acid 79.9 96.9 84.5 96.6 97.4 85.5 96.3 Glutaric acid 83.6 97.4 93.4 99.8 99.3 89.3 98.6 Citric acid 96.4 97.3 91.5 95.0 94.4 96.5 95.3

is probably due to the large sizes of these complexes that hinder their introduction in the interlayers of montmorillonite. Considering four dicarboxylic acids (oxalic acid, malonic acid, succinic acid and glutaric acid) having gradually longer hydrocarbon chains, it can be seen that the adsorption of metal ions generally increases with the increasing of hydrocarbon chain length of these acids and with the decreasing of complex stabilities. In fact the complexes characterised by rings with ve and six atoms are more stable than those formed by more than six atoms. The behaviour of Pb and Cd ions is different because their adsorption on clay decreases in the presence of succinic acid and glutaric acid. This probably occurs because these ions have larger ionic sizes, form more stable and larger complexes with ligands of longer hydrocarbon chain. The adsorption of oxalic acid (OA) on Na-montmorillonite was investigated in order to study the sorption capability of the clay towards an organic acid of small size and establish whether the surface structure SMe OA is possible at the pH of 5.5. Oxalic acid was chosen for this study because it is the smallest considered ligand and, therefore, could enter into the spaces between the

layers of montmorillonite. Besides it better differentiates the adsorption of every metal ion on the clay. The amount of adsorbed acid was determined by redox titration with potassium permanganate both in the inuent and in the efuent solutions of blanks and samples. The percentage of adsorbed oxalic acid is equal to 46% for the blanks, and to 42% for the samples containing also the studied metals. It is evident, therefore, that oxalic acid is adsorbed on the clay by cation bridging, water bridging and H-bond complexation [15]. Its adsorption, however, is partially hindered by the presence of the metal ions in solution; in fact, at pH 5.5, oxalic acid is deprotonated and, then, it preferentially forms complexes with bivalent and trivalent ions in solution. Moreover, the MeOA complex is introduced in the interlayers with greater difculty than the free metal ions because it is characterised by a chelating ring and, then, it has larger size. From the results it is evident that the behaviour of citric acid is similar to that one of glutaric acid; in fact, also in the presence of citric acid all metals were adsorbed in high percentages on montmorillonite. This is probably because, at pH 5.5, only two carboxylic

1624

O. Abollino et al. / Water Research 37 (2003) 16191627

groups of the acid are deprotonated and, then, it forms complexes similar in size and stability to the ones made by glutaric acid. A chemometric study on analytical data was made applying the principal component analysis (PCA) and hierarchical cluster analysis (HCA), in order to obtain a visual representation of the existing correlations among percentages of metal adsorbed on to montmorillonite and ligands present in solution and to evidence groups of ligands and of metal ions having similar characteristics (an introduction to these topics may be found in Kellner et al. [16]; while a more advanced discussion on this chemometric approach may be found in Massart et al. [17]; Otto [18]; Abollino et al. [19]).

From the plot of PC1 vs PC2 and the dendrogram obtained considering the ligands as objects and the metal ions as variables, reported in Fig. 4, it was possible to identify the following groups of ligands: (i) strong ligands: EDTA and NTA; (ii) ligands of intermediate strength: oxalic acid, malonic acid and tartaric acid; (iii) weak ligands: succinic acid, glutaric acid and citric acid. By this approach it was possible to understand that the behaviour of tartaric acid is similar to the one of oxalic acid and malonic acid. Instead, additional information was not obtained from PCA plots and dendrogram computed considering the metal ions as objects and the ligands as variables. Altogether it is possible to deduce that the inuence of ligands on the adsorption of

Glutaric Acid Succinic Acid Cr CitricAcid Cu EDTA NTA Ni Zn Cd MnPb Tartaric Acid

PC2 (15%)

-1 Malonic Acid Oxalic Acid -2 -4 -2 0 2

(a)

PC1 (80%)

Citric Acid Glutaric Acid Succinic Acid Tartaric Acid Malonic Acid Oxalic Acid NTA EDTA 0 (b) 1 2 3 4 5

index

Fig. 4. Combined plot of scores and loadings on PC1 vs PC2 (a) and dendrogram (b) computed considering the ligands as objects and the metal ions as variables.

O. Abollino et al. / Water Research 37 (2003) 16191627

1625

1.0

C/C0

0.5

H+

Men+

0 A

Volume (ml)
Fig. 5. Model of breakthrough curve [10].

investigated metals on Na-montmorillonite increases in the following order: Mn p Pb p Cd p Zn o Ni o Cu o Cr so that the metals are adsorbed in increasing extent in the opposite order. Therefore, it is evident that the afnity of metal ions towards Na-montmorillonite strongly depends not only on pH, but also on the presence of ligands in solution. In fact, for example, copper is much more adsorbed as aquaion than as complex. 4.3. Total capacity By the use of the continuous column method, it was possible to determine also the total capacity of Namontmorillonite towards every considered metal by the breakthrough technique. This procedure is generally used to measure the highest degree of utilisation of a chromatographic column. When the efuent solution concentration results to be equal to the one of inuent solution, that is, when the column inlet-to-outlet concentration ratio (C/C0) attains a value very close to unity, the saturation point is achieved. In this condition all surface adsorption sites of the packing material, in this case Na-montmorillonite, are occupied by the considered ion. If the C/C0 ratio is reported as a function of efuent volume, a curve known as the breakthrough curve is obtained. The model of breakthrough curve is reported in Fig. 5, where B indicates the breakthrough point of the examined ion, that is when the ion rst appears in the efuent solution, and C indicates the saturation point. Calculating the area ABCD, the total capacity is obtained. By this approach the breakthrough curves for all considered metals were obtained.

1.2 1.0 0.8

0.6 0.4 0.2 0 0 25 50 75 100 125 150 175 200 225

Cd Cr Cu Mn Ni Pb Zn

C/C0

Volume (ml)

Fig. 6. Breakthrough curves for the adsorption of Cd, Cr, Cu, Mn, Ni, Pb and Zn. The metal concentration was 1.0 103 M. The pH was controlled at 5.5.

One must consider also that the uptake yield of a metal ion by the investigated clay should depend not only on thermodynamic reasons (extent of the uptake equilibrium towards the substrate) but also on kinetic effects given by the possible low rate of metal uptake in concomitance to the continuous advancement through the column. The almost quantitative uptake for the investigated metals at pHsX5.5 (see Fig. 2) points out, however, that in the present cases kinetic effects should play a very minor role. The breakthrough curves for the adsorption of the seven metals examined on Na-montmorillonite are reported in Fig. 6. Calculating the area ABCD for each experimental curve, the total capacity of Na-montmorillonite towards each considered metal ion was determined and the results are given in Table 4.

1626

O. Abollino et al. / Water Research 37 (2003) 16191627

Table 4 Total capacity of Na-montmorillonite towards Cd, Cr, Cu, Mn, Ni, Pb and Zn Total capacity mg/g Cd Cr Cu Mn Ni Pb Zn 5.20 5.13 3.04 3.22 3.63 9.58 3.61 meq/100 g 9.25 29.6 9.57 11.7 12.4 9.25 11.0

From the experimental curves, it is evident that, in order to saturate the clay, higher volumes of chromium solution and, on the contrary, smaller volumes of cadmium and lead solutions are necessary, therefore, the breakthrough curve of chromium ion has a lower slope than the one of cadmium and lead. Moreover it is interesting to note that the adsorption behaviour of Zn, Mn and Ni, having similar course, seems to be intermediate between that of Cd, Cu and Pb, and the one of Cr. In the case of Cr, for which Fig. 6 does not show a complete levelling-off trend, a volume of 300 ml ensured the attainment of saturation. From the results it is evident that the total capacity of Na-montmorillonite towards the investigated metals increases in order: Pb2+ = Cd2+ o Cu2+ o Zn2+ o Mn2+ o Ni2+ o Cr3+. Cadmium and lead ions are adsorbed less on clay even if they have the same valence as Cu2+, Zn2+, Mn2+ and Ni2+ because they are characterised by larger ionic radii and, then, their introduction in interlayers and complexation with surface sites is limited by steric hindrance and lower electrostatic attraction. As observed in the study of the pH effect, copper (II) ion is the least adsorbed among the ions of the rst series of transition, probably because the different geometry of its aquaion principally hinders the complexation reactions with surface groups. Chromium (III) ion, instead, has a higher charge density and, then, the coulombic attraction towards the supercial sites of the clay is greater. For this reason the total capacity of Na-montmorillonite towards the trivalent ion considered is the highest. In general, however, Na-montmorillonite was found to have a good total capacity towards all considered metals.

towards the supercial sites and, moreover, the SiO and AlO groups are less deprotonated and they form complexes with bivalent and trivalent ions in solution with greater difculty. This effect is particularly evident for Cu2+ that, as aquaion [Cu(H2O)6]2+, has a distorted geometry and for Pb2+, Cd2+ that have a lower electrostatic attraction versus the clay because of their lower charge density. For these reasons, the adsorption of these ions is unsupported by cation exchange mechanism and, hence, they are more inuenced by pH variations. Therefore, the pH effect is different on each metal and at pH p 3.5 the studied metals are adsorbed in increasing entity in the following order: Cu2+ o Pb2+ o Cd2+ o Zn2+ p Mn2+ D Cr3+ D Ni2+. It was also shown that the presence of ligands in solution inuences the adsorption of heavy metals on Na-montmorillonite. This effect is a function of the ligand and metal considered but the formation of Me ligand complexes in solution altogether hinders the adsorption of the metal ions on the clay. In this case the metal adsorption increases in the following order: Cr3+ o Cu2+ o Ni2+ o Zn2+ p Cd2+ p Pb2+ p Mn2+. The results evidenced that the sorption capability of Na-montmorillonite towards each metal ion examined is different in the various conditions, and it varies both as a function of pH and of the ligand present in solution. It is necessary, hence, to consider both these factors to study a real soil/solution system and effectively predict the fate of heavy metals in the environment. Besides, it was possible to determine also the total capacity of Na-montmorillonite towards the considered metals by the breakthrough technique. By this approach it has been veried that the sorption capacity of this clay is remarkable and increases in the following order: Pb2+ = Cd2+ o Cu2+ o Zn2+ o Mn2+ o Ni2+ o Cr3.

Acknowledgements " e della We thank the Ministero dellUniversita Ricerca Scientica e Tecnologica (MURST, COFIN 2000, Rome) for nancial support.

References
[1] Alloway BJ. In: Heavy metals in soils. Glasgow, London: Blackie, Wiley, 1990. p. 740, 100219, 26177. [2] Aller AJ, Deban L. Total and extractable contents of trace metals in agricultural soils of the Valderas area, Spain. Sci. Total Environ. 1989;79(3):25370. [3] Hani H. The analysis of inorganic and organic pollutants in soil with special regard to their bioavailability. In. J. Environ. Anal. Chem. 1990;39(2):197208.

5. Conclusions The adsorption of metal ions on Na-montmorillonite decreases with decreasing pH. At low pH values (2.5 3.5), the hydrogen ion competes with the heavy metals

O. Abollino et al. / Water Research 37 (2003) 16191627 [4] Bailey SE, Olin TJ, Bricka RM, Adrian DD. A review of potentially low-cost sorbents for heavy metals. Water Res. 1999;33(11):246979. [5] Schindler PW, Furst . B, Dick R, Wolf P U. Ligand properties of surface silanol groups. I. Surface complex formation with Fe3+, Cu2+, Cd2+ and Pb2+. J. Colloid Interface Sci. 1976;55(2):46975. [6] Mercier L, Detellier C. Preparation, characterization, and applications as heavy metals sorbents of covalently grafted thiol functionalities on the interlamellar surface of montmorillonite. Environ. Sci. Technol. 1995;29(5): 131823. [7] Kraepiel AML, Keller K. Morel FMM. A model for metal adsorption on montmorillonite. J. Colloid Interface Sci. 1999;210(1):4354. [8] Batchelor B. Leach models for contaminants immobilized by pH-dependent mechanisms. Environ. Sci. Technol. 1998;32(11):17216. [9] Altin O, Ozbelge OH, Dogu T. Effect of pH, ow rate and concentration on the sorption of Pb and Cd on montmorillonite: I. Experimental. J. Chem. Technol. Biotechnol. 1999;74(12):11318. [10] Minczewski J, Chwastowska J, Dtbczynski R. The column process. In: Separation and preconcentration methods in inorganic trace analysis. Chichester: Ellis Horwood Limited, 1982. p. 35564. [11] Evangelou VP. Soil minerals and surface chemical properties. In: Environmental soil and water chemistry: principles and applications. New York: Wiley, 1998. p. 100224.

1627

[12] Barbier F, Duc G, Petit-Ramel M. Adsorption of lead and cadmium ions from aqueous solution to the montmorillonite/water interface. Colloids Surf. A 2000; 166(13):1539. [13] Nicholls D. Complexes, rst-row transition elements, London: Basingstoke. Macmillan. 1974. p. 5672. ! n LG, Martell AE. Stability constants of metal-ion [14] Sille complexes: organic ligands, IUPAC Chemical Data series no. 22, Oxford: Pergamon Press, 1979. [15] Liu A, Gonzales RD. Adsorption/desorption in a system consisting of humic acid, heavy metals, and clay minerals. J. Colloid Interface Sci. 1999;218(1):22532. [16] Kellner R, Mermet JM, Otto M, Widmer HM, editors. Analytical chemistry, the approved text for the FECS curriculum analytical chemistry, Weinheim. Wiley-VCH, 1998 [chapter, 12.5]. [17] Massart DL, Vandeginste BGM, Deming SN, Michotte Y, Kaufmann L. Chemometrics: a textbook. Amsterdam: Elsevier Science Publishers, 1987. [18] Otto M. Chemometrics, Weinheim.Wiley-VCH, 1999. p. 148166. [19] Abollino O, Aceto M, Malandrino M, Mentasti E, Sarzanini C, Barberis R. Distribution and mobility of metals in contaminated sites. Chemometric investigation of pollutant proles. Environ. Pollut. 2002;119:17793. [20] Smith RM. Martell A. Critical stability constants. Vol. 5, rst supplement, New York: Plenum Press, 1982. [21] Smith RM., Martell A. (1989) Critical stability constants. Vol. 6, second supplement, New York: Plenum Press, 1989.