Cm (III) with Alumina-Bound Poly (acrylic acid

ions M(III) (M ) Eu, Cm) and poly(acrylic acid) (PAA)- coated alumina colloids at pH 5 in 0.1 M NaClO4. Under these conditions, M(III) interaction wit...
0 downloads 0 Views 175KB Size
Download PDF
Environ. Sci. Technol. 2004, 38, 4312-4318

Interaction of Eu(III)/Cm(III) with Alumina-Bound Poly(acrylic acid): Sorption, Desorption, and Spectroscopic Studies G . M O N T A V O N , * ,† T . R A B U N G , ‡ H. GECKEIS,‡ AND B. GRAMBOW† Laboratoire SUBATECH, 4 Rue A. Kastler, BP 20722, 44307 Nantes Cedex 03, France, and Institut fu ¨ r Nukleare Entsorgung, FZK, Postfach 3640, D-76021 Karlsruhe, Germany

This paper contributes to the comprehension of kinetic and equilibrium phenomena governing trace metal ion sorption on organic matter coated mineral particles. Sorption and desorption experiments were carried out with trivalent metal ions M(III) (M ) Eu, Cm) and poly(acrylic acid) (PAA)coated alumina colloids at pH 5 in 0.1 M NaClO4. Under these conditions, M(III) interaction with the solid is governed by sorbed PAA. The results were compared with spectroscopic data obtained by time-resolved laserinduced fluorescence spectroscopy (TRLFS). Within less than 30 s, a state of local equilibrium is reached between M(III) and adsorbed poly(acrylic acid). M(III) bound to the organic-mineral surface and to dissolved PAA keeps five water molecules in its first hydration sphere. Interaction of M(III) with alumina-bound PAA appears to be stronger than with dissolved PAA. With increasing contact time, a change of the metal ion speciation occurs at the organic-mineral surface. This change is explained quantitatively by kinetically controlled reactions, which succeed a rapid local equilibrium. The experimental findings suggest, in agreement with model calculations, that a part of the initially sorbed M(III) is slowly converted to a kinetically stabilized species, thereby losing water molecules from the first coordination sphere as indicated by TRLFS. This species might be assigned as a ternary Al2O3-M(III)PAA complex. The second part of the initially bound M(III) appears to experience as well kinetically controlled reactions, however, without showing changes in the first coordination sphere. We assume that the kinetic stabilization is the consequence of rearrangement processes of the PAA at the alumina surface.

Introduction In the context of safety evaluations of nuclear waste repositories as well as for the assessment of radionuclide mobility in contaminated soils, the interaction between actinide metal ions (M) and humic substances (HS) has been the subject of various studies (1-3). These studies have been carried out with trivalent actinides (Cm, Am) and chemical homologues (e.g., Eu(III)), as their solution behavior in the environment under given conditions is known to be dominated by organic macromolecules (1). The nature of this * Corresponding author phone: +33(0)251858420; fax: +33(0)251858452; e-mail: [email protected]. † Laboratoire SUBATECH. ‡ Institut fu ¨ r Nukleare Entsorgung. 4312

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 16, 2004

interaction depends on the chemical state of the organic material, which can be soluble or associated with mineral particles. Although numerous studies on the interaction of actinides with HS in solution are available in the literature (1-4), fewer studies have been devoted to the interaction of actinides with HS-mineral complexes (5-13). Moreover, within these studies, almost no mechanistic insight into the reaction and no information on the formed species are provided. In most cases, the simple description of the metal ion interaction with the organic-mineral complex cannot be simply described by combining the behavior observed in the individual binary systems M/HS and M/mineral phase (14). To understand the M/HS-mineral interactions, simplified systems are used. Polycarboxylic acids (15-17) were selected to model the behavior of humic substances as far as their polyelectrolyte character and functional groups are concerned. In a recent study using polymaleic acid (PMA) (17), it was postulated that the interaction of Eu(III) at trace concentrations with the organic polymer is not the same for dissolved PMA and PMA sorbed to alumina. In the latter case, a migration of Eu(III) within the adsorbed organic layer after an initial rapid adsorption was postulated to explain the spectroscopic results. The goal of this paper is to gain further insight into the kinetic phenomena of metal ion sorption. As in ref 17, the ternary system is composed of γ-alumina, trivalent metal ions M(III) (M ) Eu and Cm), and a polycarboxylic acid. Poly(acrylic acid) (PAA) is preferred over polymaleic acid because the Eu(III)/PAA interaction can easily be described and is reversible (18). To simplify the quantitative description and to focus on reaction kinetics, M(III) interaction with the organic-mineral complex is studied by immobilizing the polymer on the mineral phase prior to Eu(III) adsorption. To compare the complexing properties of sorbed and free PAA, experimental conditions were chosen where M(III) interaction with aluminol surface sites can be neglected in the ternary system: all experiments were carried out at pH 5 in 0.1 M NaClO4 (7) with the alumina surface being saturated with PAA. To characterize the interaction mechanisms between M(III) and the sorbed PAA (PAAads), sorption as well as desorption experiments were carried out as a function of the contact time between M(III) and PAAads (from 30 s to 26 d). For the desorption studies, dissolved PAA is used to promote Eu(III) dissociation from the Eu-PAAads complex. The modeling results are compared with speciation measurements performed with Cm(III) using time-resolved laserinduced fluorescence spectroscopy (TRLFS).

Materials and Methods Chemicals. Commercially available sodium perchlorate monohydrate (Fluka, purum), europium oxide (Prolabo, RECTAPUR), and poly(acrylic acid) at 5000 Da (Aldrich Chemical Co.) were used as received. The proton exchange capacity of PAA amounts to 11.3 ( 0.4 mequiv/g (18). An 152Eu(III) tracer solution was provided by CEA/DAMRI. γ-Alumina was purchased from Degussa-Huls (aluminum oxide C, primary particle size of 20 nm, specific surface area of 100m2/g, site density of 1 nm-2; 19) and was used without further purification. All characteristics of this alumina, generally used for interaction studies with organic material or metal ions, are well documented in the literature (19-21). Stock Eu(III) and 152Eu(III) solutions were prepared in 1 × 10-3 M HClO4 after several dissolution/evaporation cycles in 2 M perchloric acid. Stock suspensions/solutions of 10.1021/es0301626 CCC: $27.50

 2004 American Chemical Society Published on Web 07/08/2004

FIGURE 1. Results of the sorption of Eu(III) on PAAads. (A) Sorption kinetic experiment with [Eu]tot ) 1 × 10-7 M and c ) 0.5 g/L. The inserted figure is a zoom of the gray zone. (B) Sorption isotherm for tcon < 3 min (full symbols) and for tcon ) 7 d (open symbols). Data were measured as a function of PAAads and Eu(III)tot concentration with two methods of colloid/solution separation (filtration at 0.02 µm and centrifugation). (C) Eu(III)/PAA/ PAAads system for tcon < 3 min. The figure presents Eu(III) concentration complexed with free PAA as a function of PAA concentration. [Eu]tot ) 5 × 10-8 M and c(PAAads) ) 0.35 g/L. (O) Eu(III) equilibrated with PAA before PAAads addition. (b) Eu(III) contacted with PAAads before PAA addition. The lines in the panels correspond to the calculations made with the parameters given in Tables 1 and 2. In panel B, PAAadsK and PAAadsS are the fitting parameters whereas in panel C only the PAAK value is fitted. alumina and PAA at 10 g/L were prepared at pH 5 in 0.1M NaClO4. The alumina content in the suspension was systematically checked by drying at 105 °C. PAA stock solutions were stored for a maximum of 1 week. Their concentrations were checked by total organic carbon (TOC) measurements. Experimental Procedures. All solutions were prepared using Milli-Q water, and all experiments were conducted in polyethylene tubes at room temperature under atmospheric conditions. Before use, the tubes were washed with 1 M HCl and 1 M NaOH, rinsed with Milli-Q water, and dried. (a) Preparation of Alumina-Bound Poly(acrylic acid) Colloids. A suspension containing PAA at 0.1 g/L and Al2O3 at 1 g/L was equilibrated overnight. Then, the free PAA in solution was partially removed by replacing a part of the supernatant by a fresh NaClO4 solution. The separation of both colloidal and liquid phases was performed by centrifugation at 11000g. This step was carried out until only a negligible quantity of PAA remained in solution (below 1 ppm). Throughout the experiment, the pH was controlled and readjusted if necessary; the highest variation in pH observed was 0.10 pH units. The suspension was finally concentrated by a factor of 6 (6 g/L) and stored in the dark for a maximum of 1 month. The concentration of the suspension (c), expressed in terms of alumina concentration (in g/L), was determined by drying at 105 °C. The degree of PAA loading, corresponding to the ratio between concentrations of sorbed PAA (in mg/L) and of alumina (in g/L), was equal to Γ ) 59 ( 3 mg/g. This value corresponds to the maximum of adsorption deduced from isotherm data (22). This confirms that the alumina surface is saturated with PAA.

It was checked that no significant desorption of the initially adsorbed PAA occurred (i.e., below 6%) neither during the preparation of the stock suspension nor in the experiments with Eu(III). Furthermore, it was shown that no further PAA adsorption on organic-mineral particles occurred in the experiments carried out in the Eu(III)/PAA/PAAads system. (b) Sorption and Desorption Experiments. These studies were performed as batch experiments with Eu(III). Eu(III) and 152Eu(III) were mixed at pH 5 in 0.1 M NaClO4 before the addition of PAA and/or PAAads. The colloidal phase (PAAads) was separated from the aqueous one (containing PAA or not) by centrifugation for 15 min at 36000g or by filtration through 0.02 µm Anatop 25 inorganic membrane filters (Whatman). The distribution coefficient for dissolved and sorbed/ complexed Eu(III) is defined as

Kd )

[Eu]tot - [Eu]sol

(1)

[Eu]sol(P)

where (P), [Eu]tot, and [Eu]sol correspond to the concentration of adsorbed or free polymer (in g/L), the total concentration of Eu(III) (in M) in the system, and the concentration of all Eu(III) species in solution at equilibrium (in M), respectively. When needed, the Kd values were corrected for the amount of Eu(III) sorbed on the filters. The different types of experiments performed in this study are described below together with the experimental conditions. Kinetic Experiments (Figure 1A). In these experiments, the rate of adsorption of Eu(III) on PAAads was followed as VOL. 38, NO. 16, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4313

FIGURE 2. Interaction studies between Eu(III) and PAAads as a function of the contact times (log scale). The inserted figures are zooms of the gray zones. (A) Evolution of the amount of exchangeable fraction at the organic-mineral surface as a function of contact time (rearrangement curve). [Eu]tot ) 1 × 10-7 M and c(PAAads) ) 0.5 g/L. (B) Dissociation of Eu-PAAads complexes as a function of the contact time in the presence of PAA, with [Eu]tot ) 5 × 10-8 M, c(PAAads) ) 0.25 g/L, and (PAA) ) 5 g/L. (0) tcon ) 0; (b) tcon ) 21 min; (O) tcon ) 2.5 h; (2) tcon ) 1 d; (4) tcon)7 d. All the lines in the panel correspond to model calculations made with the parameters given in Tables 1 and 2. a function of contact time. The experiment was carried out at a low Eu(III) concentration ([Eu]tot ) 1 × 10-7 M) with c ) 0.5 g/L. Determination of the Sorption Isotherm (Figure 1B). Organic-mineral particles and Eu(III) concentrations were varied from 0.1 to 4.9 g/L and from 1 × 10-10 to 2 × 10-5 M, respectively. The suspensions were contacted for tcon < 3 min and tcon ) 7 d before separation. The pH values were monitored during the equilibration and were readjusted, if necessary, to a value of 5 by addition of either NaOH or HClO4. Competition Experiments (tcon < 3 min) (Figure 1C). In these experiments, the Eu(III) solution was mixed with PAA and PAAads to evaluate the interaction strength of the two forms of PAA for Eu(III). Eu(III) and organic-mineral particle concentrations were fixed at 5 × 10-8 M and 0.35 g/L, respectively, whereas PAA concentration was varied between 0.05 and 1.5 g/L. The addition order of the reagents was varied: in a first case, Eu(III) was equilibrated with PAA prior to the addition of the organomineral colloid whereas in the second case, Eu(III) was contacted with PAAads for 30 s before the addition of PAA. In both cases, the mixing time between Eu(III), PAA, and sorbed PAA was fixed to 1 min before separation. 4314

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 16, 2004

Rearrangement Experiments (Figure 2A). This experiment intends to show the evolution of the speciation of sorbed Eu(III) at the organic-mineral surface, as a function of the contact time between Eu(III) ([Eu]tot ) 1 × 10-7 M) and sorbed PAA (c ) 0.5 g/L). The evolution was followed by regarding, at a given contact time, the ability of free PAA (5 g/L) to desorb Eu(III) in association with PAAads. The contact time between Eu-PAAads complex and PAA was fixed to 1 min. Dissociation Experiments (Figure 2B). The experiment is dedicated to the evaluation of the Eu(III)-PAAads desorption kinetics. Suspensions containing 1 × 10-7 M Eu(III) and 0.5 g/L PAAads were left for 21 min, 2.5 h, 1 d, and 7 d. After these contact times, the Eu(III) suspensions were mixed with an equivalent volume of a 10 g/L PAA solution. To simulate the contact time ) 0, Eu(III) is mixed with PAA before the addition of PAAads. Then, Eu(III) distribution between PAA and PAAads was followed as a function of the time, t. Analytical Methods. PAA concentration was analyzed by TOC measurements using a Shimadzu TOC 5000A analyzer. Eu(III) analysis was performed through the 152Eu(III) radiotracer by liquid scintillation counting using a Packard 2550 TR/AB liquid scintillation analyzer. The scintillation cocktail was ULTIMA GOLD AB (Packard).

FIGURE 3. Results of the spectroscopic studies. [Cm]tot ) 1.4 × 10-7 M and c(PAAads) ) 1 g/L. (A) Emission spectra of Cm(III) complexed with PAAads as a function of the contact time. (B) Decay curves obtained as a function of the contact time. The lines correspond to the fittings carried out according to the following expression: I ) A1 exp(-t/τ1) + A2 exp(-t/τ2) with A1 ) 5.1 × 106, τ1 ) 122 µs, and A2 ) 0 for tcon ) 15 min and A1 ) 4.55 × 106, τ1 ) 120 µs, A2 ) 5.4 × 105, and τ2 ) 226 µs for tcon ) 9 d. TRLFS. The goal of these spectroscopic investigations is to observe the evolution of metal ion speciation at the organic-mineral surface as a function of the contact time and to compare ternary M(III)/PAAads with binary M(III)/ PAA and M(III)/Al2O3 systems. Cm(III) is chosen instead of Eu(III) because it allows to work at very low concentrations and provides more spectroscopic information. It is considered that Eu(III) and Cm(III) have similar interaction properties with PAA. The Cm(III) concentration was fixed at 1.4 × 10-7 M; PAAads and PAA concentrations were fixed at 1 and 0.01 g/L, respectively. Data on the Cm(III)/Al2O3 system can be found in the literature (23). The chemical environment of complexed Cm(III) is reflected by the lifetime values and emission spectra (Figure 3). Details concerning the measuring device as well as details on how spectroscopic data are obtained can be found elsewhere (23). The samples were excited at a constant wavelength of 396.6 nm, lifetime values being measured by integration of the fluorescence peak at an emission wavelength of 600.5 nm. Lifetime measurement was started from the longest delay time down to 1 µs. The laser energy was fixed at 3.5 ( 0.5 mJ. Under these conditions, no degradation of PAA occurs. This was checked by monitoring the fluorescence intensity before and after lifetime measurement. It was also verified that laser irradiation did not induce any PAA desorption from the alumina. Fluorescence decay curves were fitted with SigmaPlot software using the Marquardt-Levenberg algorithm (version 2.0, Jandel Co.). Uncertainties associated to the fitting parameters were given by the software. Modeling. In the paper, the “P” abbreviation is used to denote PAA and PAAads. Considering the experimental conditions, M(III)/Al2O3 interaction is neglected in the ternary system. On the basis of the parameters deduced from binary systems (18, 19) and neglecting the electrostatic contribution which may enhance M(III)/Al2O3 interaction (14), the highest M(III)/Al2O3 contribution to total M(III) sorption in the conditions of the study amounts to 3%. The different models used to describe the data are presented in the following. In all the models, constants are considered as conditional. (a) Model 1. This model describes M(III)/PAA and M(III)/ PAAads interaction by the simplest way using a Langmuirtype isotherm (3, 18). The model assumes the formation of a 1:1 complex between the metal ion and P according to the reaction: Pk

+

M + P {\} MP Pk

-

(2)

where Pk+ and Pk- are the forward and backward rate constants, respectively. In the model, M(III) species interacting with P are not specified. But considering the experimental conditions, the Eu3+ aquo ion appears to be the most probable interacting species (17). Within the polyelectrolyte, M(III) interacts with one binding site consisting of a certain number of functional groups. The formed species, composed of the metal ion and the interacting ligands (carboxylic, aluminol, hydration water molecules), is described by a conditional complexation constant PK: P

K)

[MP] [M]sol[P]

(3)

with P

P k+ K)P k-

(4)

The total concentration of available binding sites for M(III) (in sites/L) is given by

[P]tot,a ) (P)PS ) [P] + [EuP]

(5)

with (P) and PS being the concentration of the polyelectrolyte (in g/L) and the number of active binding sites per unit mass of P (mol/g), respectively. In the following, the species formed according to eq 2 will be qualified as “initially formed species”. (b) Model 2 (Figure 2A). The model intends to account for the rearrangement of the initially formed species that occurs when the contact time between M(III) and PAAads varies. This rearrangement is described through successive reversible kinetically controlled reactions (24), which follow the rapid metal ion adsorption (eq 2) according to the general reaction scheme: Pk

+

Pk

1

Pk

n

M + P {\} MP {\} MP1 ‚‚‚ {\} MPn Pk

-

Pk

-1

Pk

(6)

-n

(c) Model 3 (Figures 1A, 2, and 4). In addition to model 2, model 3 includes one more M(III) species. It arises from the passage of M(III) from a fraction of the initially formed species toward a kinetically stabilized species (eq 7). In this VOL. 38, NO. 16, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4315

TABLE 1. Quantitative Description of the Interaction of Eu(III) with PdPAA and PAAads According to Model 1 system

PK (M-1)

PS (mmol/g)

Pk + (M-1 s-1)

Pk (s-1)

Eu(III)/ (3.2 ( 1.8) × 105 1.3a g2 × 104 g6.2 × 10-2 PPA (1.7 ( 0.8) × 105 a Eu(III)/ (5.8 ( 0.7) × 106 0.39 ( 0.03 g3 × 106 g0.52 PAAads a

FIGURE 4. Comparison of M(III) speciation determined by spectroscopy (dots) and by modeling (dashed and solid lines). The modeling is done according to model 3 with the parameters given in Tables 1 and 2. case, the kinetically controlled reaction is considered as irreversible, at least in the time scale explored in this paper:

(d) Modeling Strategy. Parameters characterizing the initially formed species Pk-, Pk+, and PS for PAAads and Pkand Pk+ for PAA were deduced from experiments carried out at short contact times (i.e., sorption isotherm and competitive experiments in the presence of PAA). PS for dissolved PAA was fixed to that given in the literature (18). Models 2 and 3 were used to describe the data obtained for tcon < 26 d. The kinetic parameters were determined as follows. For the dissociation experiments, the experimental rate law is given by (25)

[MPAAads] dt

)

∑A

0,i[1

- exp(-Pk-it)]

(8)

i

where A0,i represents the initial M(III) concentration bound to the ith site. Under the conditions of the experiments, MP dissociates instantaneously in the presence of PAA. Therefore, the dissociation curves give the information about the number of kinetically controlled reactions and the Pk-n values. The values obtained according to eq 8 were thus used as input parameters to fit the rearrangement curve according to eq 7. This curve is sensitive to both kinetic parameters: the slope of the curve gives Pkn parameters, and the ratio Pk /Pk P n -n gives the equilibrium conditions. Obtained k-n parameters were then used to re-simulate the dissociation experiments. This was done till the set of parameters described all experimental data. All fitting results based on eq 8 were realized with the fitting procedure of SigmaPlot software using the Marquardt-Levenberg algorithm (version 2.0, Jandel Co.) All simulations, taking eq 7 into account, were done with the MAKSIMA program (26).

Results Figure 1A depicts the kinetics of Eu(III) adsorption onto PAAads. After a fast initial adsorption within the first 30 s the 4316

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 16, 2004

Taken from ref 18.

Kd values increase to reach a plateau after about 100 h. This indicates a rapid sorption process followed by a kinetically controlled one. To characterize the first one separately, experiments are carried out for 30 s < tcon < 3 min (competition experiments and sorption isotherm). In this time scale, the Kd value is nearly constant indicating a “local equilibrium” where the kinetically controlled process can be neglected. The whole interaction mechanism is then derived from the experiments carried out for tcon < 26 d (dissociation experiments, rearrangement experiments, and sorption isotherm). Results obtained are presented in the first part of the section. They will be compared in a second part with TRLFS results. Finally, the results will be discussed in a context. Rapid Contact Time. Figure 1B presents the sorption isotherm of Eu(III) on the organomineral colloids obtained after 3 min contact time. Experimental data are well described by the model 1 (solid line in Figure 1B). The existence of a state of local equilibrium for this rapid initial adsorption reaction described by eq 2 is confirmed by the experiment realized in the Eu(III)/PAAads/PAA system (Figure 1C) since (i) whatever the addition order of P in the system is, the values describing the distribution of Eu(III) between free and sorbed PAA are similar and (ii) data are well described by the model 1 with adjusted PAAK values close to that obtained by another method (18) (see Table 1). Long Contact Time. (a) Experimental Results. The rearrangement experiment is presented in Figure 2A. The experimental conditions are chosen to remove the initially formed species from the organic-mineral surface, at a given contact time between Eu(III) and PAAads. The results show that the amount of this species decreases by a factor of about 3 after 80 h. After about 100 h, an equilibrium is attained. Figure 2B shows the results of the dissociation experiments (i.e., the desorption rate of Eu(III) from PAAads after different aging times (0-7 d)). The experiment shows that with increasing contact time, the desorption of Eu(III) becomes more and more difficult. It is interesting to notice that for contact time ) 0 (i.e., PAAads is added after equilibration between Eu(III) and PAA), Eu(III) starts to move from the PAA complex in solution into PAAads after about 3 h even though free PAA is in large excess over the adsorbed PAA (about a factor of 1000 when comparing the [P]tot,a values). These experimental observations clearly indicate a modification of Eu(III) speciation at the organic-mineral surface as the contact time increases. This change is consistent with the increase of the Kd values in Figure 1A. Sorption isotherms measured at conditions close to metal ion saturation of the PAAads for tcon ) 7 d showed that the number of available sites for Eu(III) does not change with the contact time (Figure 1A). Therefore, the increase in the Kd value has to be explained by an increase of the PAAadsK constant as suggested in eqs 6 and 7. (b) Modeling. The experimental data are relatively well described by applying model 2 and considering a minimum of three kinetically controlled reactions. The contribution of each reaction kinetics in the curve in Figure 2A is schematically represented by dashed and solid lines. The model remains however limited since it does not describe the data for the dissociation experiments measured close to equi-

TABLE 2. Quantitative Description of M(III)/PAAads Interaction According to Models 2 and 3a MP1

model 2 model 3

MP2

MP3 Pk

MPni

Pk

1 (s-1)

Pk -1 (s-1)

Pk

2 (s-1)

-2 (s-1)

Pk

3 (s-1)

Pk

-3 (s-1)

Pk ni (s-1)

fraction 1 (%)b

fraction 2 (%)c

4 × 10-4 3 × 10-4

8 × 10-4 1 × 10-3

3.5 × 10-5 3 × 10-5

3.2 × 10-5 2.5 × 10-5

6 × 10-6 1 × 10-5

3 × 10-6 5 × 10-6

1 × 10-3

100 87

0 13

a Recapitulation of the parameters associated to kinetically controlled reactions. b Fraction of sorbed M(III) described by eq 6 including MP species (see Table 1). c Fraction of sorbed M(III) in the kinetically stabilized environment.

TABLE 3. Spectroscopic Data Obtained in Ternary and Binary Systems in the Conditions of the study (pH 5, [NaClO4] ) 0.1 M) analysis of the decay curves fit 1a substrate PAAads

PAA Al2O3 (23) a

tcon

peak position (nm)

τ1 (µs)

I1 (%)

15 min 3h 1d 9d 1h

600.5-600

122 ( 1 112 ( 3 113 ( 5 120 ( 6 109 ( 3 110

100 88 ( 5 83 ( 10 82 ( 10 100 100

600.5 600.6

Decay curves were fitted with all the parameters free (i.e., τ1, I1, τ2, and I2).

librium (Figure 2B). Model 2 predicts a nearly complete Eu(III) dissociation, while a certain fraction of Eu(III) appears to remain associated to PAAads at long contact times. Therefore, model 3 appears to be more appropriate. The parameters associated to the transformation into a kinetically stabilized species (i.e., the percentage of the initially sorbed fraction being transformed and the Pkni parameter) could be selectively determined based on the dissociation curve for tcon ) 0. As indicated by the lines in Figure 2B, the agreement obtained between the experiment and the calculation on the basis of model 3 is rather good. As a result, Eu(III) is distributed to 13% in the kinetically stabilized fraction and 87% in the reversible fraction (eq 6). With respect to model 2, a slight change of the parameters had to be made (see Table 2). The model explains also the increase of the Kd values with the contact time (solid line in Figure 1A). TRLFS Measurements. The results are presented in Table 3. For all samples, irrespective of the contact time, emission spectra are almost identical (Figure 3A). A slight peak shift to higher wavelength is noticeable. A difference appears also with regard to the lifetime values, τ. These data are particularly interesting as they are linked to the number of hydration water molecules, nH2O, in the first coordination shell of M(III) (27). In the present work, nH2O is calculated from τ using the empirical law of Kimura et al. (28). The analysis of the data will be based on the following considerations: (i) The excitedstate reactions are neglected (29) and the general reaction scheme with about 100% of Cm(III) sorbed is described as follows:

In that case, every excited species Cm(III)Pi* is characterized by its own lifetime τi. The number of obtained lifetimes gives the minimum number of Cm(III) species at the organicmineral surface, provided that for these species the metal ion is surrounded by a different number of water molecules in the first hydration sphere and assuming that water molecules are the only fluorescence quenchers in the system.

b

fit 2b τ2 (µs)

I2 (%)

I1 (%)

I2 (%)

210 ( 53 208 ( 59 226 ( 85

12 ( 7 17 ( 18 18 ( 20

94.1 ( 1.0 90.9 ( 1.9 89( 1 81.2 ( 1.0

5.9 ( 0.2 9.1 ( 0.5 11.0 ( 0.3 18.8 ( 0.4

Decay curves were fitted with τ1 ) 117 and τ2 ) 215 µs.

(ii) We assume as an approximation that sorbed species have similar fluorescence intensity factors. This allows us to compare directly the intensity associated to each lifetime with the quantity of the different species. The results of the fitting procedure, leaving all the parameters free (fit 1 in Table 3), give similar lifetime values irrespective of the contact time. They amount to 117 ( 5 and 215 ( 10 µs. This result indicates that there are two sorbed species where Cm(III) is coordinated to 4.7 ( 0.5 (117 µs) or 2.1 ( 0.5 (215 µs) water molecules in the first coordination sphere. The proportion between the two species varies with the contact time. To get a more precise description of this variation (i.e., to diminish the errors associated to the intensities), the decay curves were re-fitted by fixing the lifetime values (fit 2 in Table 3). For tcon ) 15min, the species characterized by a lifetime of τ ) 215 µs appears in a small quantity of about 6%. This species is not observable when letting all the parameters free (see Figure 3B). Therefore, the species characterized by τ ) 117 µs dominates at short contact times. Consequently, this component is attributed to the initially sorbed species observed in the batch experiments. The values for the lifetime as well as the peak position correspond to spectra obtained for the Cm(III)-PAA complex. This result indicates a similar chemical environment for M(III) in both M(III)/PAA and M(III)/PAAads systems. On the other hand, it is not possible to rule out the presence of aluminol-bound Cm(III) species based on the TRLFS results because the spectra for Cm(III)Al2O3 are almost identical (same peak position and same lifetime value). With increasing contact time, the contribution of the longer lifetime increases also. This is shown in Figure 3B displaying the decay curves for tcon ) 15 min and tcon ) 9 d. The results show also that the species distribution derived from the lifetime analysis does not vary significantly after 1 h of contact: the Cm(III) fractions attributable to the two species characterized by 117 and 215 µs are 87 ( 5% and 13 ( 5%, respectively. This latter value coincides with the amount of the MPni species present at the organic-mineral surface at equilibrium (model 3, Table 2). Furthermore, the evolution of the Cm(III) species characterized by the longer fluorescence lifetime with increasing contact time in Figure 4 can be compared with the appearance of the MPni species derived from the kinetic batch experiments. It is therefore reasonable to conclude that this Cm(III) species can be VOL. 38, NO. 16, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4317

attributed to the kinetically stabilized M(III) species. Figure 4 shows as well that the TRLFS could not distinguish between MP, MP1, MP2 and MP3 species (see eq 6). One may conclude that the first M(III) coordination sphere is similar in all these cases.

would be thus explained by a difference in the electrostatic term contribution in the conditional PK values. This great PK augmentation with respect to the slight diminution of PS leads to the fact that interaction strength for M(III), characterized by the product PKPS, is higher for adsorbed PAA than for free PAA.

Discussion In agreement with results described in our previous paper (18), the interaction of M(III) with mineral-adsorbed PAA is different from the interaction with dissolved PAA. The M(III)/ PAA interaction can be described by considering PAA as a classical organic ligand (18). The interaction of M(III) with PAAads is more complex: after a rapid attachment of M(III) at the organic-mineral surface, a modification of M(III) speciation occurs. The origin of this discrepancy may be explained by the difference in PAA “structure” in the two systems. When PAA is free in solution, in the range of PAA concentrations studied (18), the polymers are independent linear chains. When PAA is adsorbed on alumina, the polymer chains cannot any more be treated as independent chains. One may rather assume formation of aggregates that form an organic layer at the mineral surface, which can be viewed as a penetrable, gel-like structure offering the possibility of kinetically controlled reactions. Concerning the natural system, humic substances are assumed to behave toward trivalent metal ions as a penetrable gel-like structure even when they are free in solution (30). The effect observed in the simplified system may therefore be difficult to transfer to the natural system. The variation in metal ion speciation observed in our system can be explained quantitatively according to eq 7 considering that the adsorbed Eu(III) is divided into two fractions in the proportions 87:13%. The first one undergoes a reversible change characterized by successive kinetically controlled reactions (eq 6). TRLFS measurements indicate that this change induces no significant modification of the chemical M(III) environment in the first coordination sphere. Hence one may invoke changes in the second coordination sphere. This would be the consequence of polymer conformational changes at the alumina surface allowing sorbed M(III) to move together with its inner-sphere coordination cage to a less water-accessible environment within the organic layer. The second M(III) fraction undergoes a kinetically controlled reaction toward a kinetically very stable species. TRLFS results indicate that this transformation is accompanied by an expulsion of water molecules from the hydration sphere of M(III). This stabilization may be due to an additional coordination in the first coordination sphere of M(III), either with carboxylate or aluminol surface groups. Studies are in progress to identify the nature of this species and to see which parameters can affect its formation (i.e., metal ion and organic loading and PAA conformation at the organicmineral surfaces). On the other hand, it appears from the study that the species formed in the M(III)/PAA and M(III)/PAAads systems can be compared when the contact time in the ternary system is rapid (i.e., below 3 min). Under these conditions, large similarities are observed: (i) the quantitative description can be done for both systems irrespective whether PAA is adsorbed or not according to eq 2, and (ii) TRLFS results indicate a similar chemical environment for M(III) in both M(III)/PAA and M(III)/PAAads systems. If SP values are compared, it appears that the number of sites diminishes by a factor of about 3 when PAA is sorbed on alumina. This is coherent with the fact that, in the ternary system, a part of the carboxylic groups interact with aluminol sites by ligand exchange reactions to form chemical bindings (31). The interaction constant PK is 18 times higher in the Eu(III)/ PAAads system as compared to the Eu(III)/PAA one. Considering that the species are identical in both systems, similar intrinsic binding constants are expected. The difference 4318

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 16, 2004

Literature Cited (1) Choppin, G. R. In Chemical Separation Technologies and Related Methods of Nuclear Waste Management; Choppin, G. R., Khankhasayev, M. Kh., Eds.; Kluwer Academic Publishers; Dordrecht, 1999; p 247. (2) Czerwinski, K.; Kim, J. I. Mater. Res. Soc. Proc. 1997, 465, 743. (3) Hummel, W. In Modelling in Aquatic Chemistry; Grenthe, I., Puigdomenech, I., Eds.; OECD Nuclear Energy Agency; Paris, 1997; Chapter V. (4) Glaus, M. A.; Hummel, W.; Van Loon, L. R. Appl. Geochem. 2000, 15, 943. (5) Samadfam, M.; Sato, S.; Ohashi, H. Radiochim. Acta 1998, 82, 361. (6) Rabung, Th.; Geckeis, H.; Kim, J. I.; Beck, H. P. Radiochim. Acta 1998, 82, 243. (7) Fairhurst, A. J.; Warwick, P.; Richardson, S. Colloids Surf. A 1995, 99, 187. (8) Allard, B.; Moulin, V.; Basso, L.; Tran, M. T.; Stammose, D. Geoderma 1989, 44, 181. (9) Ho, C. H.; Miller, N. H. J. Colloid Interface Sci. 1985, 106, 281. (10) Righetto, L.; Bidoglio, G.; Azimonti, G.; Bellodono, I. R. Environ. Sci. Technol. 1991, 25, 1913. (11) Lenhart, J. J.; Honeyman, B. D. Geochim. Cosmochim. Acta 1999, 63, 2891. (12) Reiller, P.; Moulin, V.; Casanova, F.; Dautel, C. Appl. Geochem. 2002, 17, 1551. (13) Labonne-Wall, N.; Moulin, V.; Vilarem, J.-P. Radiochim. Acta 1997, 79, 37. (14) Wermeer, A. W. P.; Mcculloch, J. K.; Van Riemsdijk, W. H.; Koopal, L. K. Environ. Sci. Technol. 1999, 33, 3892. (15) Takahashi, Y.; Kimura, T.; Kato, Y.; Minai, Y. Environ. Sci. Technol. 1999, 33, 4016. (16) Floroiu, R. M.; Davis, A. P.; Torrents, A. Environ. Sci. Technol. 2001, 35, 348. (17) Montavon, G.; Markai, S.; Andre`s, Y.; Grambow, B. Environ. Sci. Technol. 2002, 36, 3303. (18) Montavon, G.; Grambow, B. New J. Chem. 2003, 27, 1344. (19) Rabung, Th.; Stumpf, Th.; Geckeis, H.; Klenze, R.; Kim, J. I. Radiochim. Acta 2000, 88, 711. (20) Ochs, M.; Cosovic, B.; Stumm, W. Geochim. Cosmochim. Acta 1994, 58, 639. (21) Davis, J. A. Geochim. Cosmochim. Acta 1982, 46, 2381. (22) Montavon, G.; Markai, S.; Grambow, B. Radiochim. Acta 2002, 90, 689. (23) Stumpf, Th., Rabung, Th.; Klenze, R.; Geckeis, H.; Kim, J. I. J. Colloid Interface Sci. 2001, 238, 219. (24) Bryan, N.; Griffin, D.; Regan, L. Third Technical Progress Report, EC Project F14W-CT96-007; FZKA Report 6524; FZK-INE: Karlsruhe, Germany, 2000. (25) Langford, C. H.; Gutzman, D. W. Anal. Chim. Acta 1992, 256, 183. (26) Carver, M. B.; Hanley, D. V.; Chaplin, K. R. In MAKSIMACHEMISTsA Program for Mass Action Kinetics Simulation by Automatic Chemical Equation Manipulation and Integration Using Stiff Techniques; AECL: Chalk River, 1979. (27) Horrocks, W. D.; Sudnick D. R. J. Am. Chem. Soc. 1979, 101 (2), 334. (28) Kimura, T.; Choppin, G. R.; Kato, Y.; Yoshida, Z. Radiochim. Acta 1996, 72, 61. (29) Billard, I.; Lu ¨ tzenkirchen, K. Radiochim. Acta 2003, 91, 285 (30) Bryan, N. D.; Jones, D.; Griffin, D.; Regan, L.; King, S.; Warwick, P.; Carlsen K.; Bo, P. Second Technical Progress Report, EC Project F14W-CT96-0027; FZKA Report 6324; FZK-INE: Karlsruhe, Germany, 1999. (31) Schlautman, M. A.; Morgan, J. J. Geochim. Cosmochim. Acta 1994, 58, 4293.

Received for review December 11, 2003. Revised manuscript received April 29, 2004. Accepted May 25, 2004. ES0301626