Environ. Sci. Technol. 2002, 36, 3303-3309
Complexation Studies of Eu(III) with Alumina-Bound Polymaleic Acid: Effect of Organic Polymer Loading and Metal Ion Concentration G . M O N T A V O N , * ,† S . M A R K A I , † Y . A N D R EÄ S , ‡ A N D B . G R A M B O W † Ecole des Mines de Nantes, 4, rue Alfred Kastler, BP 20722, 44307 Nantes Ce´dex France
To contribute to the comprehension of the metal ion sorption properties in mixed mineral-organic matter systems, interaction studies between Eu(III) and polymaleic acid (PMA)-coated alumina colloids were carried out at pH 5 in 0.1 M NaClO4. The studied parameters were the metal ion concentration (1 × 10-10 to 1 × 10-4 M) and PMA loading on alumina (Γ ) 10-70 mg/g). The data were described by a surface complexation model. The constant capacitance model was used to account for electrostatic interactions. The results showed that two sites were necessary to explain the sorption data. At high Eu loading, Eu is surrounded by one carboxylate group and one aluminol group. The existence of this ternary surface site was in agreement with time-resolved laser-induced fluorescence spectroscopy measurements. At low metal ion concentrations, a surface site corresponding to a 1:1 Eu/COOstoichiometry was deduced from modeling. Spectroscopic data did not corroborate the existence of this latter site. This discrepancy was explained by postadsorption kinetic phenomenon: a migration of the metal ion within the adsorbed organic layer.
Introduction In the context of safety of nuclear waste repositories as well as for the assessment of radionuclide mobility in contaminated soils, the interaction between actinide metal ions, and humic substances, HS, has been the subject of various studies (1-3). Studies have been carried out notably with trivalent actinides (Cm, Am) and analogues (Eu), as their solution behavior in the environment is known to be dominated by organic macromolecules (1). The nature of this interaction depends on the chemical state of the organic material which may be soluble or associated with mineral particles. Although numerous studies on the interaction of actinide with HS in solution are available in the literature (1-3), fewer studies have been devoted to the interaction of actinide with HSmineral complexes (4-10). Studies in ternary systems of metal ion/mineral particle/ HS show that the metal ion distribution is mainly governed by the behavior of the organic matter, especially when it forms strong complexes with metal ions. The presence of humic acid (HA) increases the adsorption of Np, Am, Th, Pu, Eu, and U at the mineral surface at acidic pH values but * Corresponding author phone: 33(0)251858420; fax: 33(0)251858452; e-mail:
[email protected]. † Subatech, UMR 6457. ‡ GEPEA. 10.1021/es010312h CCC: $22.00 Published on Web 07/04/2002
2002 American Chemical Society
reduces the metal ion sorption at higher pH values (4-9). The increase of adsorption is explained by the adsorption of HA onto the mineral surface followed by the interaction of the metal ion with surface sorbed HA, whereas the reduction of adsorption is explained by the formation of soluble M-HA complexes which stabilize the metal ion in aqueous solution. The quantitative description of these ternary systems remains on a phenomenological level with the exception of the recent paper of Lenhart et al. (10) who attempted to describe the U(VI)/HA/hematite system using a surface complexation model. However, the uncertainties in these models are high considering the heterogeneous nature of HS, the important number of reactions to take into account (11), and the difficult task to describe the interaction between the metal ion and a mineral particle coated by the organic matter (12). Our purpose is to contribute to the understanding and quantification of the ternary system by using a simplified system. Our ternary system consists of γ-alumina, as mineral phase, Eu, as homologue to trivalent metal actinide, and polymaleic acid, PMA. This polymer, synthesized from pyridine and maleic anhydride, presents similarities to HS as far as its physicochemical characteristics are concerned (13, 14). Furthermore, it is nearly monodisperse, and it has been shown that no fractionation of PMA occurs after adsorption onto goethite (15). This ensures a homogeneous PMA distribution between solid and liquid phases which simplifies the modeling (16). The polymer is immobilized onto the mineral phase prior to Eu adsorption experiments to study only the interaction between Eu and PMA-Al2O3 particles. The experimental procedure intends to simplify the quantitative description. Furthermore, it simulates certain aspects of interaction scenarios concerning radionuclide/HS interactions in natural water systems. The present work aims to study the effect of mineral surface coverage by PMA on the complexation properties of mixed organic-mineral particles and to quantify the interactions using a surface complexation model together with a constant capacitance model. All experiments are carried out at pH ) 5 by varying Eu concentration over a wide range (10-10 to 10-4 M) using the radiotracer 152Eu. The modeling results are compared with speciation measurements performed by time-resolved laser-induced fluorescence spectroscopy (TRLFS).
Materials and Methods Chemicals. Commercially available sodium perchlorate monohydrate (Fluka, purum) was used as received. Eu stock solution was prepared from Eu2O3 (Prolabo, RECTAPUR) after dissolution, evaporation, and redissolution in 10-3 perchloric acid (Fluka, purum). The 152Eu tracer was provided by CEA/DAMRI, and the stock solution was prepared following the procedure used for Eu. Polymaleic acid was synthesized from maleic anhydride (Fluka, purum) and pyridine (Merck, >98% purity) according to the procedure given in the literature (13). It presents a weight-average molecular weight of 2550 Da (13) and a proton exchange capacity of 8.9 meg/g determined by potentiometric titration following the procedure given in ref 17. Poly(acrylic acid) (PAA) at 5000 Da was provided by Aldrich Chemical Company. Stock PAA and PMA solutions were prepared at pH ) 5 in 0.1 M NaClO4 and stored in the dark for a maximum of 1 week. VOL. 36, NO. 15, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3303
TABLE 1. Quantitative Description of the Adsorption of Eu on PMA-Al2O3 at [NaClO4] ) 0.1 M quantitative treatment reactions considered
fixed parameters
tOH + H+ a tOH2+ tOH a tO- + H+ tCOOH a tCOO- + H+ tCOO- + Eu3+ a tCOO - Eu2+ + H+
eq 2 eq 3 eq 4 eq 5
t(COO-)2 + Eu3+ a t(COO)2 - Eu+ + 2H+ t(COO-)3 + Eu3+ a t(COO)3 - Eu + 3H+ tCOO- + tO- + Eu3+ a tCOO - Eu - Ot+ + 2H+
eq 6 eq 7 eq 8
γ-Alumina was provided by Degussa-Huls (Aluminum Oxide C, primary particle size of 20 nm, specific surface of 100 m2/g). All characteristics about this alumina, generally used for interaction studies with organic material or metal ions, are well-documented in the literature (18-20). The number of surface sites as well as the protonation and deprotonation constants taken from ref 20 are given in Table 1. Prior to its use, γ-alumina was washed through a dialysis process with Milli-Q water in order to remove residual HCl. The Milli-Q water bath was frequently changed until the conductivity reached a stable value around 1.4-1.6 µS. The alumina content in the suspension was systematically checked after the washing process. Stock alumina suspensions were prepared at pH ) 5 in 0.1 M NaClO4.
Experimental Procedures All solutions were prepared using Milli-Q water, and all experiments were conducted in polyethylene tubes at room temperature in the presence of 0.1 M NaClO4. Before use, the tubes were washed with 1 M HCl, rinsed with Milli-Q water, and dried at 40 °C. (A) Preparation of the PMA-Al2O3 Colloids. After equilibration of a suspension of PMA and Al2O3 in NaClO4 solution, the free PMA in solution was partially removed by replacing a part of the supernatant by a fresh NaClO4 solution. This step was carried out until only a negligible quantity of PMA remained in solution (below 1 ppm). For the preparation, the concentration of alumina was always kept at 1 g/L. Depending on the degree of flocculation of the suspension, the separation of both colloidal and liquid phases was performed either by centrifugation or by sedimentation. Throughout the experiment, the pH was controlled and readjusted if necessary; the most important variation in pH observed was 0.15 pH units. Results of these experiments showed that no measurable desorption of the initially adsorbed PMA occurred. Furthermore, it was experimentally checked that no desorption occurred in the timescale of Eu adsorption on PMA-Al2O3 colloids, neither after dilution nor in the presence of Eu ([Eu] ) 1 × 10-5 to 1 × 10-8 M). These results are in agreement with literature data which show that, at pH from 4 to 7, the complex formed by ligand exchange between polycarboxylate compounds and a mineral oxide is very strong and that desorption upon dilution at constant pH and ionic strength is so slow that the equilibrium conditions cannot be reached in practice (21, 22). Organicmineral particles prepared for the interaction studies with Eu are characterized by the degree of loading, Γ (mg/g), corresponding to the ratio between concentrations of sorbed PMA (in mg/L) after the washing step and of alumina (in g/L) and by their concentrations, c, expressed in terms of alumina concentration (in g/L). (B) Sorption Experiments. These experiments concern PMA/Al2O3 and Eu/PMA-Al2O3 systems. All isotherms were 3304
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 15, 2002
fitting parameters
surface composition tCOOH: 9 mmol/g PMA tOH: 0.17 mmol/g Al2O3 (20) log K+ ) 7.7 (20) log K- ) -9.65 (20) log K ) -4.75 (24)
log K1 ) 4.35 tCOOH: 0.08 mmol/g PMA log K2 ) 3.3 tCOOH: 0.24 mmol/g PMA
determined by batch experiments. The suspensions were prepared by mixing aliquots of the stock suspensions of the components. For the Eu/PMA-Al2O3 system, Eu was mixed with 152Eu in the perchlorate solution, and the pH was adjusted to 5 before addition of the organic-mineral colloid. The suspensions were equilibrated under slow stirring using a mechanical stirrer. Required equilibration times were determined by kinetic studies to be 24 h for the PMA/Al2O3 system and 72 h for the Eu/PMA-Al2O3 system. pH values were monitored during the equilibration and were readjusted, if necessary, to a value of 5 by addition of aliquots of either NaOH or HClO4. For the Eu/PMA-Al2O3 system, the pH values increased to a maximum value of 5.15 but could decrease to pH ) 4 for the highest metal ion concentrations. After equilibration, the solid phase was separated from the aqueous one by centrifugation for 15 min at 36 000 g. Aliquots of the supernatant were finally analyzed to determine the content of PMA or Eu in solution. For the sorption studies with Eu, the distribution was defined as (3) c
K)
[Eu]tot - [Eu]sol [Eu]sol(PMA)
(1)
where (PMA) corresponds to the concentration of polymer adsorbed on alumina (in g/L), [Eu]tot is the total concentration of Eu (in M), and [Eu]sol is the concentration of Eu in solution at equilibrium (in M). (C) Analytical Methods. The concentration of PMA was analyzed by total organic carbon (TOC) measurements using a Shimadzu TOC 5000A analyzer. Eu analysis was performed by liquid scintillation counting using a Packard 2550 TR/AB liquid scintillation analyzer. The scintillation cocktail was ULTIMA GOLD AB (Packard). The detection limit is 1 × 10-13 M. The PMA-Al2O3 colloids prepared for Eu adsorption experiments were characterized by electrophoretic mobility measurements at 25 °C using a Doppler electrophoretic light scattering analyzer (COULTER DELSA, Coultronics Inc.). Measurements were performed using diluted suspensions (0.1 g/L) at two different angles of detection. Duplicate measurements were performed in a crossed-beam mode of 30 s count time, 3 V applied voltage, and a modulator frequency of 250 Hz. Between measurements, the cell was emptied, cleaned, and filled up again in order to ensure a homogeneous suspension in the cell. In addition to this, heights of sedimentation were measured as a function of Γ in narrow polypropylene test tubes of 16 mL. The samples were stored in the dark for 3 months before sedimentation measurements. The concentration of Al(III) in the supernatant of the various PMA-Al2O3 colloids suspensions was analyzed after centrifugation by ICP-MS. The amount measured was in the order of magnitude of that predicted by the solubility of
TABLE 2. Lifetime Values Obtained for Eu-PMA-Al2O3 Complexes as Well as for Reference Systemsi substrate
conditions
Eu(H2O)9‚(ethyl sulfate)3 ClO4PAA
(39) (23) a b c d e f Γ ) 70 mg/g g Γ ) 30 mg/g g Γ ) 15 mg/g g [Eu3+] ) 2 × 10-5M h
3/97 4/96 3/97 13/87
cs soln soln ppt susp hs soln ppt hs hs hs hs
[Eu3+] ) 1 × 10-5M h
22/78
hs
Al2O3 PMA PMA-Al2O3 series I PMA-Al2O3 series II
SS/WS*
τ (µs)†
nH2O**
111 111 ( 5 270 ( 20 290 ( 20 237 230 ( 11 201 229 ( 11 315 ( 22 268 ( 20 269 ( 22 629 ( 63 (24) 284f (76) 769 ( 77 (16) 284f (84)
9 9.0 ( 0.4 3.3 ( 0.3 3.1 ( 0.3 3.9 4.0 ( 0.2 4.7 4.1 ( 0.3 2.8 ( 0.2 3.4 ( 0.3 3.4 ( 0.3 1.1 ( 0.1
proposed specy Eu(H2O)93+ Eu(COO)3(H2O)3 t(Al-O)2-Eu+‚4H2O, see ref 41 Eu(COO)2+(H2O)5 tCOO-Eu-Ot+‚3H2O
0.8 ( 0.1
a pH ) 5, (PAA) ) 2.4 g/L (PAA at 2000 Da), [NaClO ] ) 0.01 M, and [Eu3+] ) 1.5 × 10-4 M (23). b pH ) 5, (PAA) ) 0.8 g/L, and [Eu3+] ) 10-3 M. 4 pH ) 7-8, [Eu3+] ) 3.2 × 10-6 M, calumina ) 0.57 g/L, [NaClO4] ) 0.1 M (20). d pH ) 6, c ) 20 g/L, and [Eu3+] ) 2 × 10-4 M. e pH ) 5, (PMA) ) 2.5 g/L (commercial PMA at 3000 Da), [NaClO4] ) 0.01 M, and [Eu3+] ) 1.5 × 10-4 M (23). f pH ) 5, (PMA) ) 0.83 g/L, and [Eu3+] ) 10-3 M. g pH ) 5, c ) 5 g/L, and [Eu3+] ) 4 × 10-5 and 8 × 10-5 M for Γ ) 15, 30 and Γ ) 70 mg/g, respectively. h pH ) 5, c ) 5 g/L, and Γ ) 70 mg/g. i (cs) Crystalline solid; (soln) solution; (ppt) precipitate; (susp) suspension; (hs) hydrated solid. (*) Distribution of Eu in % between the “strong” (SS, tCOO-Eu2+) and “weak” (WS, tCOO-Eu-Ot+) sites calculated with the parameters given in Table 1. (†) The number in parentheses are the relative contribution (in %) of a given lifetime on the decay curve. (**) Calculated according to the relation nH2O ) 1.07/τ (ms) - 0.62 (40). (f) Fixed in the fitting procedure. c
γ-Al2O3 (below 10-8 M). This result confirms that the separation between solid and liquid phases is efficient. These concentrations of Al(III) are too low to consider Al(III) as a competitor of Eu(III) for adsorption sites. TRLFS. Lifetimes values of complexed Eu, τ, have been measured in this study. The measurements have been carried out for the Eu/PMA-Al2O3 system as a function of metal ion or PMA loading, to estimate the number and the nature of the surface sites. The goal was to assess the validity of the results deduced from the quantitative interpretation of the interaction data. These data were compared to those obtained for the reference systems Eu/PMA and Eu/Al2O3 to assess the effect of the adsorption of PMA on alumina on its complexation properties toward Eu. The well-characterized Eu/PAA system (23) has already been studied as a reference system for Eu/PMA. All measurements were performed at room temperature. A Nd3+:YAG pulsed laser operating at 355 nm was used as the excitation source. The signal emitted at perpendicular angle from the sample was analyzed by a monochromator (Jobin-Yvon Spex 270M) and a ICCD camera (1024 pixels, Princeton Instruments) cooled at -30 °C. The data collected on the photodiode array were stored on a PC. A programmable OPO-Pulse Generator (Princeton Instruments PG200) allowed for measurements with a delay adjustable from 1 ns to 80 ms during a time of 3.5 ns to 80 ms. Fluorescence lifetime measurements were made at an emission wavelength of 616 nm by varying the temporal delay. After equilibration of Eu with the substrate (PMA, PAA, PMAAl2O3, and Al2O3), the solid/precipitate was removed from the solution and dried at 37 °C. The experimental conditions are given in Table 2. (D) Surface Complexation Modeling. PMA-Al2O3 is considered as a particle presenting carboxylic and aluminol groups homogeneously distributed over its surface. For the quantitative description, we assume that the deprotonation of each carboxylic group can be represented by a single deprotonation constant associated to that of acetic acid (i.e., pK ) -4.75) (24). Intrinsic surface protonation and deprotonation constants for the alumina surface groups are taken from the literature and given in Table 1 (20). The interaction data between Eu and alumina-bound PMA particles is made based on this simple description assuming a 1:1 complex between the metal ion and PMA-Al2O3. The fitting between the calculated and experimental cK values is performed by varying the number of sites, the type, and the
number of interacting groups in each site. Four reactions are considered for the modeling (Table 1, eqs 5-8). For the first three reactions, Eu is considered to interact only with PMA at the PMA-Al2O3 surface through one (eq 5), two (eq 6), or three carboxylate groups (eq 7). The existence of a ternary surface site where Eu interacts with one carboxylate group and one aluminol group was also evaluated (eq 8 in Table 1). At pH ) 5, the adsorption of Eu onto aluminol surface groups is negligible. Using the data of the Eu/Al2O3 system (20) and assuming an additive model, the highest contribution of only 6% to total Eu sorption is predicted for the lowest PMA loading. Therefore, the interaction Eu/Al2O3 is neglected in the present work, and we have ignored the enhancement of its contribution which may occurs, through an electrostatic term, due to the modification of the surface charge (12). Using the thermodynamic database of the geochemical program EQ3/6 (25) without considering the formation of the hydrolysis species Eu2(OH)24+ (26), it is found that Eu exits at 99% as the aquo Eu3+ ion in the conditions of our study for the whole range of Eu concentrations considered (below 1 × 10-3 M). Therefore, the reactions in solution were not considered in the calculations. To account for the interaction of Eu with the electrostatic field created at the charged surface, the constant capacitance model is used. The model is limited in the description of the solid/liquid interface: the charges on the organic-mineral colloidal particles cannot be considered as residing onto a surface but rather in a layer resulting from the adsorption of the polymer chains. On the other hand, it allows the importance of the electrostatic effect to be varied, which is a prerequisite to describe the interaction of metal ions with such a complex surface. The specific surface of PMA-Al2O3 is considered constant irrespective of PMA loading and equal to that of alumina. Furthermore, the conformational variations of PMA at Al2O3 surface are neglected for changing polymer loading (27, 28). This is justified considering the short length of the PMA chains. All calculations were made with the FITEQL 2.0 program (29). The Davies equation is used to calculate the activity coefficients of the components in solution (29).
Results and Discussion PMA-Al2O3 System. (A) Adsorption Isotherms. The adsorption isotherms of PMA on Al2O3 were obtained by varying the VOL. 36, NO. 15, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3305
FIGURE 1. Adsorption isotherms of PMA on 1 g/L alumina in the presence of 0.1 M NaClO4. Representation of Γ as a function of the equilibrium concentration of PMA in solution. The line is a Langmuir model fit (32) with k ) (0.30 ( 0.09) L/mg and Γmax ) (76 ( 4) mg/g.
FIGURE 2. (A) Electrophoretic mobility and (B) sediment height of PMA-Al2O3 particles measured as a function of Γ in the presence of 0.1 M NaClO4. In (A), the solid line corresponds to the potential calculated with the surface complexation model with C ) 8 F/m2 (see text and Table 1). concentrations of PMA from 5 to 150 mg/L in the presence of alumina at 1 g/L. The results are presented in Figure 1. The adsorption Γ increases with increasing total PMA concentration until a plateau is reached. The data are welldescribed by the Langmuir model (30). The best fit gave Γmax ) (76 ( 4) mg/g. This value is close to that found for the HS/Al2O3 system studied in similar experimental conditions (30): Γmax amounted to 99.7 ( 0.6 and 66.9 ( 0.7 mg/g for humic and fulvic acid, respectively. (B) Characterization. The surface charge of PMA-Al2O3 colloids was investigated as a function of the organic matter loading by electrophoretic mobility measurements. With increasing amounts of PMA sorbed onto alumina, the electrophoretic mobility is continuously shifted to more negative values as displayed in Figure 2A. This is related to a surface charge polarity reversal due to an excess of anionic charges in the adsorbed PMA layer (31, 32). The point of zero charge is found at a PMA loading of around Γ ) 25 mg/g. This value is similar to that observed for the PAA/Al2O3 system studied in similar experimental conditions (33). 3306
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 15, 2002
It is well-known that the variation of surface charge, due to electrostatic effects, influences the colloidal stability (32, 34). This was experimentally observed in this study because a flocculation of the system occurred in the presence of PMA. The degree of flocculation/destabilization was dependent on Γ. As described in the literature for the HA/Al2O3 system, this latter may be quantitatively described by the height of sedimentation (21) as shown in Figure 2B. Maximum sedimentation is observed for Γ ) 25 mg/g which corresponds, within experimental error, to the point of zero charge deduced from the electrophoretic measurements, in agreement with data from Klumpp et al. (21). (C) Modeling. The distribution of the charges at the PMAAl2O3 surface has been calculated with the surface complexation model considering reactions 2-4 and the surface composition given in Table 1. The capacitance value is fixed to 8 F/m2 (see the following discussion). The resulting potential values at the surface are plotted in Figure 2A as a function of Γ (line). The model predicts a point of zero charge at a PMA loading of 27 mg/g, which is in agreement with the experimental data. The result supports the validity of the surface description for pH ) 5. This indicates that the number of tCOOH and tOH groups bound by the interaction between PMA and Al2O3 (35) can be neglected. Eu/PMA-Al2O3 System. (A) Interaction Data. The interaction data are shown in Figures 3 and 4; Figure 3 reports the variation of cK with Γ at low Eu concentrations, and Figure 4 presents the isotherms measured for three Γ values. For the isotherm at Γ ) 27.8 mg/g, the concentration of PMAAl2O3 has been varied. In agreement with a 1:1 stoichiometric reaction between Eu and PMA-Al2O3, this variation was shown not to affect the cK values. (B) Modeling. Our first goal was to fit the data presented in Figure 3 and to explain quantitatively the increase of cK with Γ. This increase may be explained either by the variation of the electrostatic term or by the increase of the density of the functional groups at the organic-mineral surface. The importance of this second effect is illustrated in Figure 3A. In this case, the calculations have been carried out without the electrostatic term to take only the second effect into account (C is fixed arbitrarily to 9 × 103 F/m2). One can see that the calculation predicts an increase in cK when considering the sites where Eu interacts with more than one functional group of PMA. For eq 7, the dependency of cK with Γ is too high. On the other hand, a good agreement between calculated and experimental data is found when considering eq 6. This leads to neglect the electrostatic effects, in disagreement with the variation of the surface charge given in Figure 2. Therefore, eqs 6 and 7 were excluded. The experimental data have been fitted by considering eqs 5 and 8 with an electrostatic term. The capacitance value is fixed to that deduced from the proton titration experiments of alumina (20) (i.e., C ) 0.8 F/m2). The results are presented in Figure 3B by the dotted lines. For eq 5, the agreement between calculated and experimental data is poor; the calculation underestimates and overestimates the interaction below and above Γ ) 27 mg/g, respectively. The agreement is better when considering eq 8, but it remains poor below Γ ) 18 mg/g. In conclusion, the electrostatic effect appears to be overestimated. A similar effect has been observed with bacterial surfaces: a capacitance value of 8 F/m2 was necessary to describe the proton titration experiments as well as the interaction data between metal ions and the bacterial surface (36). The use of C ) 8 F/m2 leads also, in our case, to a good description of the data when considering eq 5 (thick solid line in Figure 3B), whereas with eq 8, the agreement remains poor (solid line in Figure 3B). From the dependency of cK with Γ and based on our simple description of the interface, it can be concluded that, at low Eu concentrations, Eu interacts with one functional group
FIGURE 3. Variation of cK values as a function of Γ at low Eu concentration (
