I SORPTION PROPERTIES OF AN ACTIVATED CARBON FOR PROTON AND CADMIUM(II) BY BATCH EQUILIBRATION

Transcription

1 I SORPTION PROPERTIES OF AN ACTIVATED CARBON FOR PROTON AND CADMIUM(II) BY BATCH EQUILIBRATION Maria Pesavento (1) Professor Teresa Soldi Professor Antonella Profuo Professor Fabio Conti Professor Giordano Urbini Professor Address (1) : Dipartiento di Chiica Generale, Università di Pavia, Pavia (Italy) Tel ( ) Fax ( ) e-ail pesaven@unipv.it ABSTRACT The adsorption of cadiu(ii) on an activated carbon was investigated by deterining the adsorption isother fro a 0.1 M NaNO 3 solution. Different acidities, fro ph 5 to 7.5, were separately exained. Increasing concentration of cadiu(ii) were added to the solution, and the concentration of etal ion adsorbed in the carbon phase was deterined after equilibration at 20 C. The data were treated according to the Ruzic linearization ethod, in order to evaluate the concentration of the active sites (c L ) in the carbon phase, able to bind cadiu(ii), and the conditional adsorption coefficient β c. This was found to strongly depend on the acidity of the solution phase, and this dependence was exained on the basis of a odel involving the cobination of cadiu(ii) with protonated sites. Quantities related to the coplexation constant β i and the protonation constant K ai were evaluated. β i and K ai were evaluated fro these quantities on the basis of the Gibbs-Donnan odel for the ion exchange solids. They allow one not only to describe the adsorption process, but also to ake predictions. It has been found that the adsorption of cadiu(ii) on the considered activated carbon can be described by two adsorption equilibria. One involves carbon sites at low concentration ( ol g -1 ) but with a high coplexation constant (log β i =6.75), the other involves carbon sites which are present at uch higher concentration (0.019 ol g -1 ), with a lower coplexation constant (log β i =3.62). The interaction sites of cadiu(ii) are at a uch lower concentration than the protonation sites, which is 0.51 ol g -1. KEYWORDS: Activated carbon, Cadiu adsorption, Ruzic linearization, Active sites on carbon. INTRODUCTION The activated carbons have been recently recognized to have interesting adsorbing properties towards toxic etal ions, so that they can be used to lower the content of these iportant pollutants in water [1,2]. Soe odels have been proposed to understand this, also with the ai of predicting the adsorption at different conditions, for instance at different acidity and ionic coposition of the aqueous phase. In particular the odel of the surface coplex foration has been successfully applied to rationalize the experiental observations [3]. The kinetics of such adsortion processes has been also investigated [4]. In this paper a siple odel is proposed for the adsorption equilibriu, on the basis of which the adsorption of etal ions on activated carbon can be characterized and also predicted, once a few paraeters are known. It assues that the adsorption takes place through cobination of the etal ions with a liited nuber of specific interaction sites in the solid phase. Not any assuption is ade about the structure of the sites and the nature of the bonds, and about the nuber of the sites in the carbon phase, which is deterined experientally. ABES - Associação Brasileira de Engenharia Sanitária e Abiental 1

2 In this paper the adsorption of the toxic etal ion cadiu(ii) on the activated carbon FILTRASORB 300 supplied by CHEMVIRON is exained, to show the possibilities of the proposed characterization ethod. The odel for the adsorption of etal ions on activated carbon It is assued that the adsorption of etal ions in the carbon phase takes place through cobination of the etals with ion exchange interaction sites, which are considered as ligands of the etal ion, and whose cobination with the etal can be described exactly as the coplexation in the solution phase. This odel, which is the Gibbs-Donnan odel, was successfully used by Marinsky and other researchers for the description of the adsorption of etal ions on ion exchange resins and other substances [5, 6, 7]. It is here applied for describing the etal adsorption on the activated carbon, since it is known that this solid is an ionic exchanger. The carbon is probably a uch less hydrophilic phase than, for instance, the ion exchange resins, but the odel could be applied also to hydrophobic ion exchangers [6]. The ost iportant point of the Gibbs-Donnan odel is that it predicts the existence of a potential difference at the interface ion exchangersolution, due to the different obility of the ions through the phases. Then the following equilibriu holds for all the electrolytes able to diffuse through the phases: a CaA = acaa equation (1) where a indicates the activity, the sybols with the over bar are relative to the species in the solid phase, and A and C are an anion and a cation, onovalent for siplicity. The concentration of the interaction sites in the solid phase is experientally deterined by titrating the with the etal ion of interest, which is added at known concentration to an aqueous suspension of the activated carbon. The concentration of etal ion adsorbed in the solid phase after equilibration is easured, obtaining an adsorption function siilar to the Languir adsorption isother, in agreeent with the odel hypothesis that the adsorption takes place through cobination on interaction sites. The experiental data are treated according to a ethod previously suggested to titrate unknown soluble ligands in natural waters, obtaining significant results with a liited nuber of data points [8, 9, 10]. Also the conditional adsorption constant can be deterined by the proposed procedure. This quantity is a easure of the strength of the adsorption, and allows one to quantitatively predict the adsorption. For siplicity, the association etal ion/active site in the carbon phase is also called coplex, and a siple case in which only a 1 to 1 (etal to site) coplex is fored in the solid is considered. It is understood that these interaction sites ay have also different structures, provided that the respective coplexation coefficients with the etal ion are not significantly different considering the experiental uncertainty. The adsorption reaction can be represented as following M + L ML [ML] β c = [M] cl * [ML] β = equation (2) [M][L] Here too the over bar indicates the species in the solid phase, and the concentration is in ol per gra of solid in the solid phase (w g of carbon phase) and in ol per l in the solution phase (V l of solution). β c is a conditional coefficient, depending on the acidity and ionic coposition of the solution, according to the Gibbs-Donnan odel, as discussed below. c L * is the concentration of the interaction site not coplexed with the considered etal, while L is the for of the interaction site which actually cobines with the etal ion. It is possible that it cobines also with the proton, depending on the conditions, and in particular on the acidity of the solution. For this reason β c ay depend on the acidity, as it happens in the case of ligands in aqueous solution. The ligand site is at different protonation states. β is the corresponding adsorption constant, independent of the acidity. The following ass equilibria hold w c M = [M] + [ML] equation (3) V V is the volue of the solution phase (in l) and w the aount of activated carbon (in g) ABES - Associação Brasileira de Engenharia Sanitária e Abiental 2

3 c = c * [ML] equation (4) L L + c L* = α L [L] equation (5) Fro which it is seen that β =β c α L It has been proposed that the value of c L, the concentration of the active groups in the carbon phase, can be evaluated together with β c, using the following relation, which has been previously suggested by Ruzic [8], and is now widely used [M] 1 = [ML] clβc [M] + cl equation (6) Of course this relationship holds at constant ph, since β c is constant at this condition. [M] is the concentration of the etal in the solution phase, in the absence of any coplexing substance in the solution phase. It is deterined experientally by the usual ethods for the deterination of the concentration of etal ions in solution. [ ML] is deterined by eq. 3, once [M] is known. An equation siilar to eq. 6 has been previously suggested for the deterination of the concentration of coplexing substances in aqueous solution, by a procedure which has been defined ligand titration [8, 9, 10]. Of course the function [M]/[ ML] vs. [M] (equation 6) is a straight line only if β c is really constant during the titration of the active groups with the etal ion. It is well known that this can be not true in the case of other ion exchange solids, for instance ion exchange resins [11, 12]. The odel for the effect of the acidity of the aqueous phase on the adsorption of etal ions on activated carbon If the active site can cobine with the etal, and also with proton, the following protonation reaction, which is considered for the sake of siplicity, ust be taken into account: H + L HL [HL] K a = equation (7) [H][L] Where K a is the protonation coefficient of the sites in that activated carbon. Siilarly to what happens with the ligands in the solution phase, the etal ion cobines with the deprotonated coplexing sites. The fraction of the active group in the deprotonated for should be [L] + [HL] α L = = 1+ Ka[H] equation (8) [L] The relation between β c and β is 1 β c = β equation (9) 1+ Ka[H] Thus β and K a can be evaluated fro β c obtained by the ligand titration at different ph, fro eq. 9, by non linear regression or by linearization of the function, according to 1 K = 1 + a[h] βc β β equation (10) According to the Gibbs-Donnan odel (eq. 1) the adsorption constant β depends on the ionic coposition of the solution and solid phases as following ABES - Associação Brasileira de Engenharia Sanitária e Abiental 3

4 [C] β = βi equation (11) [C] Where β i is the intrinsic coplexation constant of the etal with the active sites in the carbon phase. It has been found to be independent of the conditions in the case of ion exchange resins [7, 11, 12]. On the contrary β c is constant only if [ C] is constant, the other quantities, [H] and [C], being constant during a particular titration. In this research [ C] is assued to be constant, but this hypothesis ust be checked. According to the Gibbs- Donnan odel (eq. 1), also K a depends on the ionic copositions of the two phases [C] K a = K a i equation (12) [C] where C indicates the counter ion concentration in the two phases. K ai is the intrinsic protonation constant, and is independent of the ionic copositions of the phases. Considering also the effect of the ionic coposition of the phases, the relationship between the adsorption coefficient experientally deterined, and which ust be used for predictions, β c, and the variables is β ([C]/[C]) β i c = equation (13) 1 + [H] Kai( [C]/[C]) It ust be taken in ind that the activity, not the concentration of the counter ion in the two phases should be considered, when eq. 1 is used. The concentration is reported for convenience. β i is the intrinsic coplexation constant of that etal with that activated carbon. It is independent of the acidity and of the ionic coposition of the two phases, and can be used for the evaluation of the adsorption coefficients β c at particular conditions (acidity and counter ion concentration). EXPERIMENTAL Reagents The activated carbon used in this study was FILTRASORB 300 supplied by CHEMVIRON, 8-30 esh. It was used without further washing and stored in a closed bottle. The surface characteristics are herein suarized. Average specific surface, easured by the nitrogen-bet ethod is /g. Iodine nuber 950 g/g, the iniu ethylene nuber 230 g/g, effective diaeter 0.8-1, wet carbon particles density 1.2 g/l. All the experients were done with the carbon fraction obtained with a esh sieve. Analytical grade reagents were used. Stock solutions of cadiu(ii) were prepared fro weighted aounts of Cd(NO 3 ) 2 by dilution with illi-q water, and checked by FA-AAS with standard ethod. Equipent ph easureents were done with a PHM 84 RESEARCH pheter (Radioeter), and a cobined glass electrode, standardized in H + activity. The etal concentrations were deterined with a Shiadzu odel 6601G Graphite Atoic Absorption Spectrophotoeter, with pyrolitic tubes and with the procedure indicated by the instruent producer. Also voltaetric ethods were used, with a Ael od 433 Polarographic Analyzer. The easureents were done in DPV (starting potential: V, end potential: V, scan rate 50 V s -1, pulse height 50 V) or in ASDPV (deposition potential: V, deposition tie: 120 s) at a static ercury drop electrode (SDME) as working electrode, a satured KCl AgCl/Ag as reference electrode and a platinu electrode as counter electrode. Deterinations were done by the standard additions ethod. ABES - Associação Brasileira de Engenharia Sanitária e Abiental 4

5 Procedure for the deterination of the adsorption isothers The adsorption isothers were obtained in 0.1 M NaNO 3, at roo teperature (20 C), by placing into 30 L polyethylene (PE) tubes constant aounts of carbon (5 g /L) in 0.1 M NaNO 3 solution (20 L) and increasing the etal concentration at constant ph. The experients were perfored at different ph values. ph was never greater than 7.5 for cadiu(ii) in order to prevent hydrolysis and/or hydroxides precipitation. The investigated concentration ranges varied fro a few tens of ppb to a few hundreds of pp. The tubes were equilibrates for at least four hours (usual equilibration tie 10 hours), and then were centrifuged at rp for ten inutes. Proper volues of clear solution were analysed by EA-AAS and/or voltaetry, as descried above. Procedure for the investigation of the protonation of the activated carbon The axiu capacity of the carbon for H + was deterined by adding an excess of acid to the saple and easuring the excess at different conditions. For exaple, 3 l of standard HNO M were added to 25 l of 0.1 M NaNO 3 solution, containing in suspension g of activated carbon. After equilibration (norally 4 hours) an aliquot of the clear solution was separated and the H + excess was deterined by titration with standard NaOH. RESULTS AND DISCUSSION It has been found that the equilibration tie for the adsorption of cadiu(ii) on the considered activated carbon was 4 hours, but the equilibration tie actually used in the experients was always 10 hours, in order to be sure that the equilibriu was attained. The concentration of the active sites in the carbon phase, able to bind cadiu(ii), was deterined by titration of the activated carbon with the etal ion at different conditions, as reported in the experiental part. Each titration was carried out at constant ph and ionic coposition of the solution, by varying the concentration of the etal in the range to M. Different subsaples were added with different concentration of cadiu(ii), and equilibrated with the sae aount of carbon (0.1 g in 20 l of solution). The concentration of the etal ion in the liquid phase after equilibration was deterined by ETA-AAS or voltaetric techniques, and the concentration of adsorbed etal was obtained by difference. For convenience, in the data treatent two different concentration ranges were separately considered: M and M. As an exaple, soe adsorption isothers of cadiu(ii) the highest concentration range of titrant etal, fro to M, and at different ph, are reported in Fig 1. ABES - Associação Brasileira de Engenharia Sanitária e Abiental 5

6 Fig.1 Adsorption isothers of cadiu(ii) on activated carbon in 0.1 M NaNO 3, at different ph, and for titrant concentration fro M to M ol g M ph 5.6 ph 6.5 ph 6.6 ph 7.1 ph 7.4 The subscript 1 indicates the paraeters deterined in this concentration range. A constant concentration of adsorbed etal ion in function of increasing concentration of added etal was never obtained, not even at the highest concentrations, so that the total concentration of the active sites for cadiu in the considered carbon was always deterined using the Ruzic linearization (equation 10). In all the cases the function [M]/[ ML] vs. [M] was an acceptable straight line, indicating that in each titration the conditional adsorption coefficient β c1 is actually constant. Considering equation 13, this eans that [ C], the counter ion concentration in the carbon phase, or also the ratio [ C]/[C], is constant during a particular titration. A reason for this is discussed below. The values of c L1 and β c1 obtained fro the paraeters of the straight line at the different conditions are reported in Table 1. The total concentration of the active sites found by the ligand titration is independent of the conditions, at least for ph lower than around 7.5, the variations being ascribable to the irreproducibility of the ethod. The average value (and the standard deviation) is 0.019(0.004) ol/g. The considerably higher c L obtained at ph=8.2 is probably due to the precipitation of the hydroxide which takes place at the high concentration and ph considered. However such concentrations are not realistic in the environental water saples. The conditional adsorption coefficient significantly increases with the ph, reflecting the fact that the adsorption of cadiu(ii) is prooted at higher ph, as previously observed for exaple in Ref. [3]. Table 1 - Results of the Ruzic linearization of the titration of activated carbon at different ph with cadiu(ii), at titrant concentration M M, in NaNO M. (The standard deviation is in parenthesis). ph c L1 β c1 c L1 logβ c (0.26) (0.27) (0.33) (0.74) (1.10) (0.16) (0.30) (0.50) (0.30) (0.84) (0.25) (0.35) (4.31) (3.31) (0.25) (0.22) (0.30) (0.65) (0.61) (0.31) ABES - Associação Brasileira de Engenharia Sanitária e Abiental 6

7 This is described in the odel by eq. 12, which can be conveniently transfored as following 1 βcc L 1 [H] Kai( [C]/[C]) = + β i cl( [ C]/[C] ) βi cl( [C]/[C]) equation (13a) The paraeters of eq. 13a, obtained by the linear regression, by plotting 1/β c1 c L1 vs [H], are: = 1. 29( ) ( ) 10 [H] (R 2 =0.819) βc1cl1 Fro which it is found that β c ([C]/[C]) i1 L1 = 7.77(2.77) βi1 cl1( [C]/[C]) / K ai1 = 3.55(1.27) 10-5 Ka i1 ([C]/[C]) = 2.19(0.28)10 6 as c L1 is known, it can be evaluated that β ([C]/[C]) i1 = These paraeters are a useful description of the sorbing properties of the considered activated carbon, and they allow the evaluation of β c1 at different conditions by eq. 13. An exaple of active sites titration at uch lower titrant concentration is reported in Fig. 2. The adsorption isothers were obtained at different ph, and for cadiu additions fro to M. The paraeters deterined at these concentrations are indicated by the subscript 2. At these concentrations too a arked effect of the solution ph is observed, causing an increase of the adsorption at increasing ph. In Table 2 the results obtained by the Ruzic linearization are reported. Due to the poor reproducibility of the experiental deterination of the etal concentration in the solution phase at the low concentrations here considered, the concentration of the active sites could be deterined with sufficient precision only at ph=5.7. At the other acidities the slope of the Ruzic function was not significantly different fro 0, so that c L could not be evaluated. The variation of β c2 c L2 with the acidity, according to eq. 13a, was as following = 4. 64( 0. 43) ( ) 10 [H] (R 2 =0.999) βc2cl2 ABES - Associação Brasileira de Engenharia Sanitária e Abiental 7

8 Fig.2 Adsorption isothers of cadiu(ii) on activated carbon in 0.1 M NaNO 3, at different ph, and for titrant concentration fro M to M ol g E-07 1E-06 2E-06 2E-06 M ph=6.9 ph=5.7 ph=4.7 Table 2 - Results of the Ruzic linearization of the titration of activated carbon at different ph with cadiu(ii), at titrant concentration fro M to M, in NaNO M. (The standard deviation is in parenthesis). ph c L2 β c2 c L2 β c (1.10) (1.02) (1.20) (0.48) (0.19) (0.61) Not evaluable 1.86 (0.22) 10 3 Not evaluable Fro the paraeters of the straight line obtained fro eq. 13a, it is found that β c ([C]/[C]) i2 L2 = 2.15(0.20) 10 3 β -1 i2 cl2( [C]/[C]) / Kai2 = 3.59(0.05) 10-3 Ka i2 ([C]/[C]) = 6.01(0.56) 10 5 Assuing that c L2 = ol/g, it is found that β ([C]/[C]) i2 = 5.67(0.54) 10 6 This value is very high, uch higher than that of the sites at higher concentration found above. Moreover the protonation constant of the site at low concentration is one order of agnitude lower than that of the sites at lower concentration, so that the proton copetes with cadiu(ii) at a lower extent when the sites at low concentration are involved. When these strong sites are saturated, at higher etal ion concentration, the sites at larger concentration but with lower conditional coefficient will be occupied. The fact that sites with different affinity for the etal ion are present in the sae sorbing aterial is not surprising, since different density and spatial distribution of the donor atos can be expected. ABES - Associação Brasileira de Engenharia Sanitária e Abiental 8

9 Protonation of the activated carbon The axiu capacity of the activated carbon for H + was deterined as described in the experiental part, at different conditions, and in particular at different excess of acid. Siilar results were obtained, and the average value of the proton exchange capacity was 0.51 ± 0.06 ol/g. This is in agreeent with the values obtained for the surface charge evaluated by other researchers in the case of siilar activated carbons [3]. It ust be noticed that the concentration of the interaction sites of cadiu are at considerably lower concentration in the carbon phase than the protonation sites. The total ion exchange capacity does not correspond to the capacity for cadiu(ii). Probably the sites able to interact with cadiu(ii) not only contain protonable donor groups, for instance carboxyl and hydroxyl groups, but they have also a proper special structure, suitable for the cobination with cadiu. Sites of different structure are present, with different affinity for cadiu. They have also different protonation constants, which have been deterined together with the adsorption of cadiu(ii). The conditional constant β c Soe coents ust be done about the fact that the conditional constant β c obtained at constant acidity and ionic coposition is really constant during the titration. This is an experiental observation obtained fro the Ruzic linearization, and in particular to the fact that the function was always a straight line. β c is correlated to the intrinsic constants β i and K ai through eq. 13, and it should depend on [H + ] and on the concentration of the counter ion in the solid phase and in the solution phase. [H + ] is constant in each titration experient, and the sae ust be true for the counter ion concentrations in the two phases. It is also possible that not the single activity, but the ratio of the activities of the counter ion in the two phases [ C]/[C] is constant. This ay happen when the counter ions in the solid phase are not free. It is not unreasonable to suppose that the alkaline etals are present in the solid phase ainly linked to sites, because of the very lipophilic nature of carbon, involving the presence of only a sall aount of free water. In this case the counter ion is only present in the solid phase as a free ion because of the diffusion. For this reason at the equilibriu its activity ust be equal in the two phases. If this explanation holds, also the effect of the ionic coposition of the solution phase should be liited, and have only an effect on the activity coefficients of the ions in the solution phase. Soe preliinary experients at ionic coposition of the aqueous phase different fro 0.1 M NaNO 3 see to support this interpretation, but further investigation is required to confir this point. An iportant consequence is that the intrinsic coplexation and protonation constants are equal to the paraeters evaluated by the proposed analysis of data, since [ C]/[C]. CONCLUSIONS Two kinds of binding sites for cadiu(ii) are present in the considered activated carbon, indicated by the subscript nubers 1 and 2. They were characterized by deterining their concentration and the corresponding coplexation constants. Also the protonation constant of the site was evaluated, which explains the effect of the acidity. One kind of site, indicated by 2, is at low concentration, but has a very high coplexation constant, and is responsible for the adsorption of cadiu(ii) at concentration as low as about 30 ppb. The other kind of site is at uch higher concentration and has a lower coplexation constant. The knowledge of the paraeters akes it possible to evaluate the adsorption coefficients at the particular conditions considered, and the fraction of etal ion adsorbed onto carbon at these conditions, since the following relation holds 1 f = V 1+ w( β c2[l2] + βc1[l1]) The eaning of the sybols has been explained above. The total concentration of the sites c Lx can be of help to evaluate the concentration of the free sites L x. At concentration of etal ion higher that about M, sites 2 are negligible, while at concentration lower that about M, sites 1 are not occupied, and can be neglected. ABES - Associação Brasileira de Engenharia Sanitária e Abiental 9

10 REFERENCES 1. Corapcioglu, M.O., Huang, C.P. The adsorption of heavy etals onto hydrous activated carbon. Water Res., v. 21, p , Reed, B.E., Matsuoto, M.R. Modelling Cd adsorption in single and binary adsorbent (PAC) systes. J. Environ. Engng., v. 119, p , Gabaldon, C., Marzal, P., Ferrer, J., Seco, A. Single and copetitive adsorption of Cd and Zn onto a granular activated carbon. Water Res., v. 30, p , Chen, J.P., Lin, M. Equilibriu and kinetics of etal ion adsorption onto a coercial H-type granulare activated carbon: experiental and odeling studies. Water Res.,v. 35, p , Eger, C., Anspach, W.M., Marinsky, J.A. The co-ordination behaviour of Co, Ni, Cu, and Zn in a chelating ion exchange resin. J. Inorg. Nucl. Che., v. 30, p , Marinsky, J.A., Kodao, H., Miyajia, T. An approach for assessent of the hydrophobicity/hydrophilicity of charged polyers. J. Phys. Che. B, v. 102, p , Pesavento, M., Biesuz, R. Characterization and application of chelating resins as cheical reagents for etal ions, based on the Gibbs-Donnan odel. Reactive Polyers, v. 36, p , Ruzic, I. Theoretical aspect of the direct titration of natural waters and its inforation yield for trace etal speciation. Anal. Chi. Acta, v. 140, p , Van den Berg, C.M.G., Donat, J.R. Deterination and data evalaution of copper coplexation by organic ligands in sea water using cathodic stripping voltaetry at varying detection windows. Anal. Chi. Acta, v.257, p , Pesavento, M., Alberti, G. Deterination of the coplexing properties of drinking waters toward copper(ii) and aluiniu(iii) by ligand titration. Water Res., v. 34, p , Pesavento, M., Biesuz, R., Gallorini, M., Profuo, A. Sorption echanis of trace aount of divalent etal ions on a chelating resin containing iinodiacetic groups. Anal Che., v. 65, p , Pesavento, M., Biesuz, R., Cortina, J. L. Sorption of etal ions on a weak acid cation-exchange resin containing carboxylic groups. Anal. Chi. Acta, v. 298, p , ABES - Associação Brasileira de Engenharia Sanitária e Abiental 10