Sorption of Uranium(VI) onto Lanthanum ... - ACS Publications

concentration, and surface area and sorption site density of the mineral ..... from an ab initio calculated. EXAFS spectrum (FEFF7.02 code)47,48 of th...
3 downloads 0 Views 162KB Size
Download PDF
Langmuir 2002, 18, 7977-7984

7977

Sorption of Uranium(VI) onto Lanthanum Phosphate Surfaces E. Ordon˜ez-Regil,† R. Drot,*,† E. Simoni,† and J. J. Ehrhardt‡ Institut de Physique Nucle´ aire, Universite´ Paris XI, 91406 Orsay, France, and Laboratoire de Chimie Physique et Microbiologie pour l’Environnement, UMR 7564 CNRS - Universite´ Henri Poincare´ Nancy 1, 405 rue de Vandœuvre, 54600 Villers-le` s-Nancy, France Received February 27, 2002. In Final Form: May 28, 2002 The speciation of hexavalent uranium at the solid-solution interface was investigated. To experimentally identify the sorption equilibria, we characterized the structure of the surface complex formed by binding between uranyl ions and surface groups of solid matrixes. To understand the sorption mechanisms at a molecular level, we performed optical and X-ray photoelectron spectroscopies and X-ray absorption spectroscopy on uranyl ion loaded phosphate solids (LaPO4 and La(PO3)3). Two lanthanum phosphates were synthesized. The samples were contacted with aqueous uranyl solutions of pH values ranging from 1.0 to 4.0. Whatever the conditions, uranium surface coverage was always lower than 30% of the monolayer as measured by the proton-induced X-ray emission technique. The U 4f X-ray photoelectron spectra and the lifetime values of uranyl ions sorbed on the lanthanum monophosphate compound clearly evidence that this solid exhibits two different types of sorption sites, as well as lanthanum polytrioxophosphate. The nature of the site which interacts with uranyl ions also depends on the pH value for both solids. Moreover, the interaction of the uranyl ions and the phosphate solids in a nitrate medium leads to two different sorbed species: free aquo UO22+ ions and UO2(NO3)+ ions. The X-ray absorption spectroscopy performed on the sorbed samples gives evidence of the presence of an inner-sphere mononuclear polydentate surface complex.

Introduction Repository of long-term nuclear fuels in deep geological sites is one of the major issues of the nuclear fuel cycle. Thus, definitions of the geological parameters, barrier materials, and repository conditions are some of the problems that should be addressed. One of the most important processes affecting safety during the storage of nuclear waste in underground disposal is migration through the geosphere.1,2 The radionuclides potentially released from the nuclear waste, particularly actinides such as U, Np, Pu, Am, and Cm, could sorb onto nontransportable mineral surfaces which may enhance retardation.3-7 However, a quantitative understanding is rather complicated by the dependence of sorption processes on various geochemical parameters such as pH and ionic strength of the aqueous phase, radionuclide concentration, and surface area and sorption site density of the mineral substrates. To describe and to assess the ion retention, the sorption mechanisms have to be accurately understood. Thus, quantitative models are necessary. Two approaches are mostly used: ion exchange theory or the surface complexation model.8-14 Neverthe†

Institut de Physique Nucle´aire, Universite´ Paris XI. Laboratoire de Chimie Physique et Microbiologie pour l’Environnement, UMR 7564 CNRS - Universite´ Henri Poincare´ Nancy 1. ‡

(1) Degueldre, C.; Ulrich, H.; Silby, H. Radiochim. Acta 1994, 65, 173. (2) Guillaumont, R. Radiochim. Acta 1994, 66/67, 231. (3) Hering, J. G.; Kraemer, S. Radiochim. Acta 1994, 66/67, 63. (4) Nagasaki, S.; Tanaka, S.; Todoriki, M.; Suzuki, A. J. Alloys Compd. 1998, 271/273, 252. (5) Gupta, A. R.; Venkataramani, B. Bull. Chem. Soc. Jpn. 1988, 61, 1357. (6) Payne, T. E.; Davis, J. A.; Waite, T. D. Radiochim. Acta 1996, 74, 239. (7) Degueldre, C.; Wernli, B. J. Environ. Radioactivity 1993, 20, 151. (8) Mishra, T.; Parida, K. M.; Rao, S. B. Sep. Sci. Technol. 1998, 33 (7), 1057. (9) Misak, N. Z. Colloids Surf., A 1995, 97, 129.

less, these two thermodynamical points of view are not able to provide sufficiently accurate information on the reactive surface sites and on the nature and the structure of the sorbed species. To date, only a few studies have been devoted to the structural characterization of the sorption processes as illustrated by the works published on uranyl ion sorption onto clay minerals and natural systems.15-23 However, the main works in this field propose values of sorption constants determined only by direct modeling of the retention data,24-26 without any experimental validation, which yields arbitrary surface complexes. Therefore, to probe directly the structure of (10) Bartlet, J.; Cooney, R. Spectrochim. Acta 1989, 45A, 541. (11) Manceau, A.; Rask, J.; Buseck, P.; Nahon, D. Am. Mineral. 1990, 75, 490. (12) Hayes, K.; Papelis, Ch.; Leckie, J. J. Colloid Interface Sci. 1988, 125, 717. (13) Carroll-Webb, S.; Walther, J. Geochim. Cosmochim. Acta 1988, 52, 2609. (14) Marmier, N.; Fromage, F. J. Colloid Interface Sci. 1999, 212, 252. (15) Morris, D.; Chisholm-Brause, C.; Barr, M.; Conradson, S.; Eller, P. Geochim. Cosmochim. Acta 1994, 58, 3613. (16) O’Day, P.; Brown, G.; Parks, G. J. Colloid Interface Sci. 1994, 165, 269. (17) Drot, R.; Simoni, E.; Alnot, M.; Ehrhardt, J. J. J. Colloid Interface Sci. 1998, 205, 410. (18) Manceau, A.; Charlet, L. J. Colloid Interface Sci. 1992, 148, 426. (19) Spadini, L.; Manceau, A.; Schindler, P.; Charlet, L. J. Colloid Interface Sci. 1994, 168, 73. (20) Ames, L.; McGarrah, J.; Walker, B. Clays Clay Miner. 1983, 31, 321. (21) Tsunashima, A.; Brindley, G.; Bastovanov, M. Clays Clay Miner. 1981, 29, 10. (22) Del Nero, M.; Salah, S.; Miura, T.; Cle´ment, A.; Gauthier-Lafaye, F. Radiochim. Acta 1999, 87, 135. (23) Del Nero, M.; Ben Saı¨d, K.; Made´, B.; Cle´ment, A.; Bontems, G. Radiochim. Acta 1998, 81, 133. (24) Marmier, N.; Dumonceau, J.; Chupeau, J.; Fromage, F. C. R. Acad. Sci., Ser. II 1993, t. 317, 1561. (25) Tochiyama, O.; Endo, S.; Inoue, Y. Radiochim. Acta 1995, 68, 105. (26) Jaffrezic-Renault, N.; Poirier-Andrade, H.; Trang, D. H. J. Chromatogr. 1980, 201, 187.

10.1021/la025674x CCC: $22.00 © 2002 American Chemical Society Published on Web 09/18/2002

7978

Langmuir, Vol. 18, No. 21, 2002

adsorbates, the use of structural methods at a molecular scale should give detailed information on the mechanism of sorption allowing clear-cut discrimination between complexation and exchange and between inner- and outersphere surface complex structure. Among various spectroscopic techniques used,27-35 we performed X-ray photoelectron spectroscopy (XPS), induced laser spectrofluorimetry (emission spectra, decay time measurements), and extended X-ray absorption fine structure (EXAFS) spectroscopy. The metal oxides have been mostly studied as substrates, but among the phosphate minerals, only apatites36 and more recently thorium and zirconium phosphates (Th4(PO4)4P2O7 and ZrP2O7) have been studied.37-39 In this work, we propose to study the following systems: UO22+/ LaPO4 and UO22+/La(PO3)3. This investigation will allow one to study separately the effects of the PO4 and PO3 surface groups on uranyl ion sorption mechanisms. Experimental Section Synthesis and Characterization. All reagents were of analytical grade (Fluka-Aldrich) and were used without further purification. Both solids were synthesized as described in the literature.40,41 A wet route was chosen to produce very homogeneous precursors of the desired final product. Lanthanum chloride and ammonium dihydrogen phosphate solutions were mixed, under continuous stirring, with the following stoichiometries of the final compound: P/La ) 1/1 and P/La ) 3/1 molar ratio for lanthanum monophosphate and polytrioxophosphate, respectively. After addition, the precipitates were dried at 120 °C, ground, and then heated at 600 °C for 2 h (under an argon atmosphere) in order to expel volatile substances. Finally, the compounds were fired at 1200 °C for 8 h for LaPO4 and 800 °C for 2 h in the case of LaP3O9. To obtain homogeneous grain size, the final products were ground and then washed using deionized water until the pH of the supernatant was equal to the pH value of the deionized water. The crystalline structures were checked by recording the X-ray powder diffraction diagrams with a Philips PW 1050/70 apparatus using Cu KR rays and a Ni filter. It was found that both compounds are well crystallized. Diffraction patterns were compared to the JCPDS files no. 2-493 and no. 84-1635 for lanthanum monophosphate and lanthanum polytrioxophosphate, respectively. Both compounds were found as monophasic solids. Infrared analysis was also performed for each solid. Infrared spectra were recorded on a Hitachi I-2001 spectrophotometer. Samples were prepared considering 1-2 wt % of the product in KBr. The obtained spectra for LaPO4 and LaP3O9 were found to be identical to those already reported in the literature.42,43 (27) Glinka, Y. D.; Jaroniec, M.; Rozenbaum, V. M. J. Colloid Interface Sci. 1997, 194, 455. (28) Cavellec, R.; Lucas, C.; Simoni, E.; Hubert, S.; Edelstein, N. Radiochim. Acta 1998, 82, 221. (29) Moulin, I.; Stone, W. E. E.; Sanz, J.; Bottero, J.-Y.; Mosnier, F.; Haehnel, C. Langmuir 1999, 15, 2829. (30) Chung, K. H.; Klenze, R.; Park, K. K.; Paviet-Hartmann, P.; Kim, J. I. Radiochim. Acta 1998, 82, 215. (31) Koretsky, C. M.; Sverjensky, D. S.; Salisbury, J. W.; D’Aria, D. M. Geochim. Cosmochim. Acta 1997, 61 (11), 2193. (32) Manceau, A.; Charlet, L. J. Colloid Interface Sci. 1994, 168, 87. (33) Towle, S. N.; Bargar, J. R.; Brown, G. E.; Parks, G. A. J. Colloid Interface Sci. 1999, 217, 312. (34) Lefevre, G.; Walcarius, A.; Ehrhardt, J.-J.; Bessiere, J. Langmuir 2000, 16 (10), 4519. (35) Scheidegger, A. M.; Lamble, G. M.; Sparks, D. L. J. Colloid Interface Sci. 1997, 186, 118. (36) Yinjie, S.; Hui, Z.; Qiaoling, Y.; Aimin, Z. J. Radioanal. Nucl. Chem. 1995, 198 (2), 375. (37) Drot, R.; Lindecker, C.; Fourest, B.; Simoni, E. New J. Chem. 1998, 1105. (38) Drot, R.; Simoni, E.; Denauwer, Ch. C. R. Acad. Sci., Ser. IIc 1999, t. 2, 111. (39) Drot, R.; Simoni, E. Langmuir 1999, 15, 4820. (40) Pepin, J.; Vance, E. J. Inorg. Nucl. Chem. 1981, 43, 2807. (41) Park, H. D.; Kreidler, E. R. J. Am. Ceram. Soc. 1984, 67, 23. (42) Rulmont, A.; Cahay, R.; Liegeois-Duyckaerts, M.; Tarte, P. Eur. J. Solid State Inorg. Chem. 1991, 28, 207.

Ordon˜ ez-Regil et al. The specific surface area, determined by the N2 BrunauerEmmett-Teller (BET) method, are 10.4 and 3.8 m2 g-1 for LaPO4 and La(PO3)3, respectively. Sorption Procedure. Two types of uranium(VI) stock solutions were prepared depending on the nature of the supporting electrolyte. Nevertheless, whatever the medium and before adding uranyl solid, the background salt was previously acidified at pH ) 3 with HNO3 (or HClO4) in order to avoid cation hydrolysis. Moreover, the initial concentration of all stock solutions was around 5 × 10-3 M. The exact uranyl concentrations were determined by alpha liquid scintillation by using a TRICARB spectrometer supplied by Packard (Camberra Co., Meriden, CT) and the so-called Alphaex scintillation cocktail according to the protocol described in the literature.44 For the nitrate medium, a stock solution was prepared by dissolving a weighed amount of UO2(NO3)2‚6H2O (Merck) in a KNO3 solution (0.5 M, pH ) 3). For the perchlorate medium, the uranyl nitrate solid was dissolved in concentrated perchloric acid (1 M) and then the solution was evaporated. After this first step, the obtained precipitate was dissolved with concentrated perchloric acid and evaporated again. This procedure was repeated three times in order to be sure that all nitrate ions were removed. Finally, an uranyl stock solution was prepared by dissolving a weighed amount of the UO2(ClO4)2‚xH2O solid in a NaClO4 solution (0.5 M, pH ) 3). Sorption experiments were carried out in batch mode, at room temperature, with polypropylene tubes which avoid uranyl sorption onto the tube walls. The solids (200 mg) were fully hydrated in 10 mL NaClO4 or KNO3 solution (0.5 M) for 24 h (LaPO4) or 7 days (LaP3O9), according to previous experiments which have shown that these times are needed to reach hydration equilibrium. After centrifugation (3500 rpm for 30 min), a known amount of the aqueous phase (about 200 µL) was removed and replaced by the uranyl solution adjusted to a desired pH value. The suspension is shaken (45 rpm) for 24 h. Kinetic measurements have already shown that sorption equilibria are reached in these experimental conditions. The suspensions were then centrifuged. The final equilibrium pH value of the supernatant was measured, and the uptake of uranyl ions was determined as the difference between the initial and final uranium concentrations. In addition, the sorbed solids were washed with distilled water and dried at room temperature. Spectroscopic Measurements. The uranyl emission spectra (characteristic strong green fluorescence) were collected, at room temperature, with a 1-m Jobin-Yvon HR 1000 monochromator with a 1200 line/mm grating, equipped with a Hamamatsu R636 photomultiplier tube. The third harmonic of a Quantel Nd:YAG pulsed laser (width around 7 ns) at 355 nm was used for excitation. The time-dependent emission decays were measured with a Lecroy 935M multichannel analyzer. Around 1000 decay measurements were averaged to obtain these decay profiles. The time dependence obtained was fitted to a multiexponential law. All XPS spectra were collected on an electron spectroscopy for chemical analysis (ESCA) apparatus with a multidetection electron analyzer (VSW HA150, fixed analyzer transmission (FAT) mode). A Mg KR1,2 source of photons (1253.6 eV and halfwidth of 0.9 eV) was used for both survey (FAT ) 90 eV) and narrow scans (FAT ) 22 eV). The powdered samples, fixed on a copper plate, were placed in the analytical chamber under a 10-9 mbar vacuum. As the studied compounds are electric insulators, charge effects are expected. Therefore, the binding energy of the C 1s line from adventitious aliphatic carbon (284.6 eV) has been used as an internal reference for calibrating the energy scale. The recorded lines, U 4f, C 1s, and O 1s, were fitted using a Gaussian-Lorentzian peak shape after subtraction of the background (Shirley baseline). The uranium 4f spin-orbit splitting was held at 10.8 eV, and the component ratio (4f7/2/4f5/2) was constrained at 0.8. Moreover, a satellite peak (well-known for UVI) with an intensity around 10% of the main line is present on each spectrum at 3.4 eV toward higher binding.45 The uranium LIII-edge EXAFS spectra (17 167 eV) were recorded at the ROBL experimental station (ESRF, France). (43) Tarte, P.; Rulmont, A.; Sbai, K.; Simonot-Grange, M. H. Spectrochim. Acta 1987, 43A (3), 337. (44) Dacheux, N.; Aupiais, J.; Courson, O.; Aubert, C. Anal. Chem. 2000, 72, 3150.

Sorption of Uranium(VI) onto Lanthanum Phosphates

Langmuir, Vol. 18, No. 21, 2002 7979

Table 1. Characteristics of the Main Experimental Samples substrate LaPO4 LaPO4 LaPO4 LaPO4 LaP3O9 LaP3O9 LaP3O9 LaP3O9

[UO22+]ini (M)

pHeq

% sorbed

medium

code

2 × 10-4 2 × 10-4 2 × 10-4 2 × 10-4 2 × 10-4 2 × 10-4 2 × 10-4 2 × 10-4

1.0 2.0 3.0 3.0 1.0 2.0 3.0 3.0

20 70 100 100 20 30 70 100

NaClO4 NaClO4 NaClO4 KNO3 NaClO4 NaClO4 NaClO4 KNO3

PO4Cl1 PO4Cl2 PO4Cl3 PO4N1 PO3Cl1 PO3Cl2 PO3Cl3 PO3N1

Double-crystal Si(111) located between two Pt-coated mirrors gives a monochromatic beam without higher-order harmonics and an energy resolution at the uranium LIII-edge of around 0.6 eV. All sorbed samples were measured at room temperature in fluorescence mode using a four-element Ge detector. As the uranium concentrations sorbed on the solid are rather low, around five individual scans were necessary in order to get good-quality spectra. EXAFS data reduction has been carried out according to the standard procedures by using in-house codes.46 Fourier transforms of the EXAFS data were achieved in the k3 weighting mode using a Kaiser window. The backscattering phases and amplitudes have been extracted from an ab initio calculated EXAFS spectrum (FEFF7.02 code)47,48 of the reference compound UO2(H2PO4)2‚3H2O.49

Figure 1. Emission spectra of uranyl ions sorbed on LaPO4 (pH ) 3 in NaClO4, 0.5 M): [UVI]ini ) 2 × 10-4 M (straight line); [UVI]ini ) 10-3 M (squares); [UVI]ini ) 10-2 M (dots).

Results The uranyl sorbed samples were prepared according to the following conditions (Table 1): 10-1 M > [UO22+]ini > 2 × 10-4 M; NaClO4 or KNO3 (0.5 M) medium; 1 < pHeq < 3; 1 < % sorbed < 100. The surface sites of the phosphate compounds present mainly both positive and neutral charges, in the above experimental conditions, as the pH value of the isoelectric point is around 4 and 3 for LaPO4 and La(PO3)3, respectively.50 Taking into account the measured specific surface area and the proton-induced X-ray emission (PIXE) measurements for both solids, which give the ratio U/La of around 2 ‰ in atomic composition for higher loaded solids, and assuming that the monolayer corresponds roughly to 5 × 1014 atoms/cm2 (5 sites/nm2),37 these experimental conditions lead to sorbed samples with a surface coverage of less than 30% of a monolayer (considering mononuclear monodentate surface complexes). The XPS method needs samples under high-vacuum conditions which of course do not correspond to an in situ experiment. Therefore, it was first necessary to check whether the drying step in the sample preparation could perturb the structure of the sorbed complexes. We then compared the sorbed uranium fluorescence for both dried and wet samples. It appeared that both spectra were identical, as well as the corresponding lifetimes. Thus, we can consider, in our experimental conditions, that the drying step does not change drastically the surface complex resulting from the sorption process occurring in an aqueous suspension. (45) Teterin, Yu. A.; Teterin, A. Yu.; Lebedev, A. M.; Dementjev, A. P.; Utkin, I. O.; Nefedov, V. I.; Bubner, M.; Reich, T.; Pompe, S.; Heise, K.-H.; Nitsche, H. J. Prakt. Chem. 1999, 341 (8), 773. (46) Michalowicz, A. Ph.D. Thesis, Universite´ du Val-de-Marne, Paris, 1990. (47) Ankoudinov, A. L.; Rehr, J. J. Phys. Rev. B 1997, R1712, 56. (48) Ankoudinov, A. L. Ph.D. Thesis, University of Washington, Seattle, WA, 1996. (49) Mercier, R.; Phan Thi, M.; Colomban, P. Solid State Ionics 1985, 15, 113. (50) Ordon˜ez-Regil, E.; Drot, R.; Simoni, E. Langmuir, submitted for publication.

Figure 2. U 4f XPS spectra of uranyl ions sorbed on LaPO4 (pH ) 3, NaClO4, 0.5 M): [UVI]ini ) 2 × 10-4 M (dots); [UVI]ini ) 10-3 M (circles); [UVI]ini ) 10-2 M (full squares); [UVI]ini ) 10-1 M (open squares).

1. UVI/LaPO4 System. To determine the nature of the sorption site, we have recorded the emission spectra and measured the corresponding lifetimes of uranyl ion sorbed on the phosphate prepared in NaClO4 solution ([U]ini ) 2 × 10-4 M) and at a pH value equal to 3 (samples PO4Cl3 as mentioned in Table 1) for which only free UO22+ ions are present in solution.51 Only one decay time value, corresponding to the fluorescence spectrum shown in Figure 1, was measured at 80 µs for U/LaPO4, and the emission bands are located at 497.6, 518.4, 542.1, and 567.4 nm (this spectrum and this lifetime value are identical for the samples PO4Cl1 and PO4Cl2, prepared at a pH equal to 1 and 2, respectively). According to our experimental setup, the accuracy for the wavelength value is about 0.5 nm. Despite the rather poor signal-to-noise ratio, the corresponding XPS spectrum (Figure 2) has been fitted considering only one component (U 4f7/2) at 382.5 eV with a full width at half-maximum (fwhm) equal to 2.3 eV. This fwhm agrees with data from already published works for uranyl sorbed onto other phosphate materials.17 Nevertheless, the same experiments performed with the samples prepared with a total uranyl concentration equal to 10-3, 10-2, and 10-1 M at pH 3 in the NaClO4 medium give different results. Two different lifetime values were obtained at 75 and 210 µs. Moreover, Figure 1 clearly shows that the position of the emission bands depends on the uranyl concentrations as reported in Table (51) Grenthe, I. In Chemical Thermodynamics of Uranium; Elsevier: North Holland, 1992.

7980

Langmuir, Vol. 18, No. 21, 2002

Ordon˜ ez-Regil et al.

Figure 4. Emission spectra of UVI sorbed on LaPO4 in KNO3 (straight line) and NaClO4 (squares) with [UVI]ini ) 2 × 10-4 M and precipitate (dots). Figure 3. U 4f XPS spectrum decomposition for UVI sorbed on LaPO4: pH ) 3; NaClO4, 0.5 M; [UVI]ini ) 10-2 M. Table 2. Wavelength of the Uranyl Emission Bands: U(VI) Sorbed on LaPO4 and U(VI)-Phosphate Precipitate [UO22+]ini medium 10-4

2× M NaClO4 10-3 M NaClO4 10-2 M NaClO4 2 × 10-4 M KNO3 precipitate

λ1 (nm)

λ2 (nm)

λ3 (nm)

λ4 (nm)

497.6 500.9 497.0 499.8 499.2 502.4

518.4 523.1 518.1 520.9 520.5 523.6

542.1 547.1 541.5 543.8 544.6 547.6

567.4 571.8 569.0 571.0 573.3

2. For 10-3 M, the main emission bands are located at 500.9, 523.1, 547.1, and 571.8 nm but a shoulder appears at a wavelength 5 nm shorter than the main ones. These shoulders seem to correspond to peaks in the spectrum obtained for 2 × 10-4 M which agrees quite well with lifetime results. For an uranyl concentration larger than 10-3 M, the shoulders are not observed but the emission bands are located at averaged wavelengths (Table 2). Moreover, this result is strongly related to XPS measurements as the recorded U 4f7/2 spectra for high uranyl concentrations were decomposed into two components (fwhm ) 2.3 eV) at 382.5 and 381.0 eV. As an example, the typical shape of such a spectrum and its decomposition is presented in Figure 3 (for an initial uranium concentration equal to 10-2 M). The ratio between both components appears to be dependent on the total uranyl concentration. Indeed, the percentage of the peak area corresponding to the lower energy component increases with increasing UVI concentration. If we express this percentage (noted P) as (low energy peak area)/(total U 4f peak area), we obtain the following results: [U]ini ) 10-3 M, P ) 8%; [U]ini ) 10-2 M, P ) 20%; [U]ini ) 10-1 M, P ) 30%. Since the uranyl ions ([U]ini ) 2 × 10-4 M) are not complexed in the NaClO4 medium, at pH ) 3,51 we can assume that the measured values (80 µs, 382.5 eV) correspond to the free aquo uranyl ions sorbed on one type of surface sites of the lanthanum monophosphate. For the samples prepared at pH ) 3 with higher uranyl concentrations, different identifications are possible, but first, investigations have to be performed in order to verify that no secondary uranium phase (precipitate for instance) was formed during our experiments. The lanthanum phosphate presents a monazite structure, and it is a sparingly soluble compound like all the rare-earth element monazite minerals.52 To check that the very weak dissolution of this solid has no influence on the sorption mechanism, we checked if a precipitation of

uranyl phosphate occurs in our sorption conditions. To prepare such precipitates, we performed the hydration of LaPO4 for 7 days in NaClO4 at pH ) 1 (to increase the dissolution process) and after centrifugation the supernatant was kept (called soln A) and uranyl solution ([U] ) 5 × 10-2 M) was added to soln A to obtain the final precipitate which was dried at room temperature. The emission spectrum of the obtained precipitate is shown in Figure 4. This figure leads to the following remarks: (i) The energy positions (502.4, 523.6, 547.6, 573.3 nm) of the emission bands are completely different from those obtained for the sorbed samples. (ii) The width of this spectrum is widely shorter than the ones corresponding to the sorbed samples. Moreover, the decay time for this sample is 140 µs which is very different than the values obtained for sorbed samples. Therefore, we can assume that no uranyl precipitate is withdrawn from the solution with the sorbed solid after centrifugation. Taking into account this result, we can now propose two identifications for the samples prepared at pH ) 3 with uranyl concentrations higher than 2 × 10-4 M: (i) In the NaClO4 medium, at pH ) 3, there are mainly two or three uranyl species in solution depending on the total uranium concentration:53 for [U] ) 10-3 M, free UO22+ (99%) and (UO2)2(OH)3+ (1%); for [U] ) 10-2 M, free UO22+ (92%), (UO2)2(OH)3+ (6%), and (UO2)2(OH)22+ (2%); and for [U] ) 10-1 M, free UO22+ (61%), (UO2)2(OH)3+ (30%), and (UO2)2(OH)22+ (9%). In these conditions, the decay time value at 75 µs and the XPS peak energy at 382.5 eV should correspond to the free uranyl sorbed on the same type of surface sites as in the previous experiment performed at low uranium concentration. Consequently, the second values (210 µs and 381.0 eV) may correspond to the polynuclear species (UO2)2(OH)3+ sorbed on a surface site identical to or different from the first one. However, for the sample prepared with [U]ini ) 10-3 M, the concentration of the polynuclear uranyl complex in solution is around 1%, and the probability of the sorption of these (UO2)2(OH)3+ ions should thus be rather low. Moreover, in such experimental conditions the low-energy component obtained with XPS measurements is rather important (about 8% of the total uranium signal against 1% in solution for (UO2)2(OH)3+) which should imply that even if its steric dimension is larger, the affinity of the polynuclear species is higher than that of the aquo uranyl ion. (ii) One could consider that the uranyl speciation in solution does not play an important role. Especially,

Sorption of Uranium(VI) onto Lanthanum Phosphates

Figure 5. U 4f XPS spectra of UVI sorbed on LaPO4 (squares), UVI sorbed on La(PO3)3 (gray line), and UVI sorbed on La2O3 (straight line); [UVI]ini ) 10-2 M.

considering the rather large size of the polynuclear species, these entities may not be able to sorb onto the surface through the double layer. In that case, the values 210 µs and 381.0 eV should correspond to the free uranyl sorbed on another kind of surface sites. Therefore, we must assume that the lanthanum monophosphate surface presents two different types of sorption sites: strong and weak surface sites. Thus, it appears clearly that even if this first investigation showed that there are only one type of surface complex at low uranium concentration (2 × 10-4 M) and two different ones for higher concentrations, some questions about the nature of the second sorption state observed are still to be clarified. To address this point, we compared the binding energies of U 4f for uranium sorbed on LaPO4 and La2O3. For the U/La2O3 system, the obtained value was 381.4 eV with a fwhm equal to 2.6 eV (Figure 5). This value is larger than the one obtained for phosphate materials but agrees quite well with ones found on the uranyl-loaded oxides.54 This energy is very close to the lower value obtained for the U/LaPO4 system when high uranium concentrations are considered. On one hand, we can safely assume that two types of sorption sites are observed during uranyl sorption processes on lanthanum monophosphate. On the other hand, we can also propose that one of these sites corresponds to hydroxyl groups linked to the surface lanthanum atoms (EB(U 4f 7/2) ) 381.0 eV) and thus the second one should correspond to the phosphate groups (EB(U 4f 7/2) ) 382.5 eV). Furthermore, this interpretation is also supported by previous results obtained considering other phosphate materials.17 As a conclusion, for [U]ini ) 2 × 10-4 M the free uranyl species are sorbed onto phosphate surface groups (strong sites) and when the uranium concentration is increased another surface site interacts which corresponds to the hydrated lanthanum atoms (weak sites). The structure of the sorbed complexes was also investigated via EXAFS experiments. Thus, we looked at possible contributions of U-U distances which could be consistent with the presence of the dinuclear ion (UO2)2(52) Sales, B. C.; White, C. W.; Boatner, L. A. Nucl. Chem. Waste Manage. 1983, 4, 281. (53) Palmer, D. A.; Nguyen-Trung, C. J. J. Solution Chem. 1995, 24 (12), 1281. (54) Lomenech, C.; Simoni, E.; Drot, R.; Ehrhardt, J.-J.; Mielczarski, J. J. Colloid Interface Sci. To be published.

Langmuir, Vol. 18, No. 21, 2002 7981

Figure 6. Fourier transform (not phase shift corrected) of the EXAFS spectrum of uranyl ions sorbed on LaPO4 (a, straight line), uranyl ion in solution (b, dots), and calculated for (UO2)2(OH)3+ (c, open squares).

(OH)3+ on the surface,55,56 especially for the higher uranium concentrations ([U]ini g 10-2 M). Using the ab initio modeling code FEFF7.02, we have calculated the Fourier transform (not phase shift corrected) of the (UO2)2(OH)3+ complex using the U-U distance equal to 3.94 Å and the U-O distance of 2.40 Å.57 This calculated spectrum (Figure 6) exhibits the contribution of the two axial oxygen atoms and an intense peak around 3.7 Å which corresponds to the U-U backscattering interaction. As uranium is a heavy atom, its backscattering intensity should be rather high. Considering the signal-to-noise ratio, no such intensity was found in all Fourier transform spectra obtained with the sorbed samples. Consequently, we can rule out the hypothesis of sorbed polymeric species on the lanthanum monophosphate surface as postulated in the first interpretation. Conversely, this observation confirms the fact that this solid presents both strong and weak sorption sites. Finally, taking into account all these data, we can assume that the species adsorbed onto both surface sites are very close to the structure of free uranyl ions. The X-ray absorption spectra were recorded for sorbed samples prepared at pH equal to 3 with a total uranyl concentration equal to 10-3 M. As EXAFS experiments are much less sensitive than spectrofluorimetry and XPS techniques, the concentrations less than 10-3 M were too low to get good-quality spectra in our experimental conditions. The optical and XPS experiments have evidenced the presence of two different uranyl species on the surface corresponding to strong and weak sorption sites. Nevertheless, due to the low contribution of the weak sorption sites as determined by XPS experiments (around 8%) this contribution was neglected in the processing of the EXAFS data. In contrast, the spectra obtained with the samples prepared with higher uranyl concentrations (more than 10-3 M) were not fitted, because in that case the second contribution cannot be neglected as it represents more than 10% of the total signal. In all spectra, the XANES regions (not shown here) are identical with the classical shape obtained for the uranyl sorbed samples: a white-line maximum energy at 17 176 (55) Allen, P. G.; Bucher, J. J.; Clark, D. L.; Edelstein, N. M.; Ekberg, S. A.; Gohdes, J. W.; Hudson, E. A.; Kaltsoyannis, N.; Lukens, W. W.; Neu, M. P.; Palmer, P. D.; Reich, T.; Shuh, D. K.; Tait, C. D.; Zwick, B. D. Inorg. Chem. 1995, 34, 4797. (56) Clark, D. L.; Conradson, S. D.; Donohoe, R. J.; Keogh, D. W.; Morris, D. E.; Palmer, P. D.; Rogers, R. D.; Tait, C. D. Inorg. Chem. 1999, 38, 1456. (57) Moll, H.; Reich, T.; Szabo, Z. Radiochim. Acta 2000, 88, 411.

7982

Langmuir, Vol. 18, No. 21, 2002

Ordon˜ ez-Regil et al.

Table 3. Best Fit Parameters for Uranyl Ions Sorbed on LaPO4 and La(PO3)3a LaPO4/La(PO3)3 axial oxygens equatorial oxygens equatorial oxygens N σ2 (Å2) R (Å) ∆E (eV) a

2.15/2.2 0.0046/0.0044 1.77/1.77 2.89/7.02

2.4/3.0 0.0020/0.0019 2.31/2.33 8.2/8.2

1.8/1.8 0.0036/0.0034 2.47/2.50 13.0/13.0

[UO22+]ini ) 10-2 M, [NaClO4] ) 0.5 M, pH ) 3.0.

Figure 7. Experimental (symbols) and calculated (straight lines) Fourier transform of UVI sorbed on LaPO4: [UVI]ini ) 10-2 M, pH ) 3, and NaClO4 (0.5 M) medium.

eV corresponding to hexavalent uranium, a shoulder that accounts for multiple scattering along the uranyl rod, and a shoulder due to the mixing of single and multiple scattering. Results of EXAFS data fitting are reported in Table 3 and Figure 7 which shows both radial and imaginary parts of the experimental and calculated Fourier transform. The value at 1.77 Å for the axial oxygen atoms is very close to the one previously obtained for the uranyl ion sorbed on the zirconium phosphate compound (1.76 Å).38 There are two different distances for U-equatorial oxygen atoms (2.31 and 2.47 Å). As the second contribution of the Fourier transform (Figure 6) for uranyl ions in a noncomplexing solution (10-4 M in NaClO4 0.5 M, pH ) 2.5) is due to 5 oxygens at 2.41 Å, we assume that the two oxygen atoms at 2.47 Å belong to hydroxyl groups or water hydration molecules. Therefore, for the U/LaPO4 system prepared in the sodium perchlorate medium, we can assume the presence of a bidentate surface complex as we observe two oxygen atoms at 2.31 Å. The U-P distance for this system, corresponding to the third shell, was determined using the phase and the amplitude corresponding to the U-P bound of the reference compound UO2(H2PO4)2‚3H2O. One phosphorus atom was found at 2.74 Å which corresponds to an angle of around 90° between the bonds U-O (2.31 Å) and P-O (at 1.50 Å in the PO4 group). Therefore, if the two oxygen atoms (U-O, 2.31 Å) belong to the same PO4 surface group, the angle between the two bonds P-O should be around 114°, which is a reasonable value for the PO4 tetrahedron. Thus, we can conclude that the uranyl ion is complexed with only one PO4 surface group. Moreover, as the phosphorus atom is only 2.74 Å away from the uranium, this indicates that there is no layer of water molecules between the sorbed uranyl ion and the surface and thus that it is sorbed as an inner-sphere complex. The sorption experiments were performed in the nitrate medium (KNO3, 0.5 M) as well. For the sample prepared at pH ) 3 with [U]ini ) 2 × 10-4 M (PO4N1), the emission spectrum (Figure 4) exhibits larger bands (fwhm around

117 nm for the 499.3 nm peak) than those obtained for the sample prepared in the NaClO4 solution (around 95 nm for the 496.9 nm peak) considering same uranium concentration conditions. The peak positions (499.2, 520.5, 544.6, and 571.0 nm) are significantly different as well (Table 2). Moreover, two different lifetime values were obtained at 80 and 250 µs, and the XPS spectra (U 4f7/2) were decomposed into two components at EB(U 4f7/2) ) 382.5 and 381.5 eV, respectively. Taking into account the above interpretation and the low uranium concentration, we can assume that only the strong surface sites (EB(U 4f7/2) ) 382.5 eV) are involved in the process. The uranyl speciation leads, in that case, to 81% UO22+ ions and 19% UO2(NO3)+ ions. Therefore, the second lifetime value (250 µs) and the second XPS peak energy (381.5 eV) correspond to the nitrate uranyl complex UO2(NO3)+ sorbed on the lanthanum phosphate. A summary of all these results is presented in Table 4. Spectrofluorimetry and XPS studies indicate that there are two different kinds of surface complexes on the same site, which are the free uranyl ions and the nitrate complexes. The corresponding Fourier transform is the same as the one in the perchlorate medium, which is not surprising because it is not possible to discriminate the backscattering of the oxygens belonging to water molecules and to nitrate ions. Therefore, the obtained distances are averaged, and this result appears to agree with the previous interpretation. In the nitrate medium as for the perchlorate medium, the sorbed uranyl species form innersphere complexes. 2. UVI/La(PO3)3 System. Identical studies were carried out on the UVI/La(PO3)3 system. The fluorescence spectra, the corresponding decay times, and the U 4f XPS spectra were recorded for all samples listed in Table 1. For the samples prepared in the NaClO4 solution at pH values equal to 1, 2, and 3 with a total uranium concentration at 2 × 10-4 M, the emission spectra are quite similar (495.0, 516.1, 539.5, and 566.5 nm, with a width around 90 nm for the first peak). A typical example of these spectra is presented in Figure 8. The corresponding lifetime values (130 and 350 µs) are the same for these three samples as well. In these preparation conditions, free aquo UO22+ ions are the unique species in solution. The two different lifetime values thus indicate clearly that the lanthanum polytrioxophosphate presents two different types of surface sites. This is also supported by the U 4f7/2 XPS spectra, which can be decomposed into two components at EB(U 4f7/2) ) 382.9 and 381.7 eV with fwhm ) 2.3 eV. Furthermore, when the total uranium concentration is increased up to 10-3 M, the lifetime values and the XPS spectra do not change significantly, whereas for uranium concentrations at 10-2 and 10-1 M, one of the two XPS components (EB(U 4f7/2) ) 381.7 eV) diminishes (Figure 9). The percentage of the high-energy XPS component (percentage defined as for LaPO4) increases from 46% to 90% with the initial uranium concentration. In addition, the emission intensity pre-exponential factor for the 130 µs value decreases up to almost zero for the sample prepared with [U]ini ) 10-1 M. This would indicate that the surface site corresponding to values of EB(U 4f7/2) ) 382.9 eV and a lifetime of 350 µs becomes predominant at high uranium concentration. Note that these fluorescence spectra are not at the same position as the one relative to the uranyl phosphate precipitate. For this system, the protocol used to prepare the precipitate was the same as for the previous one except that in that case, lanthanum polytrioxophosphate was

Sorption of Uranium(VI) onto Lanthanum Phosphates

Langmuir, Vol. 18, No. 21, 2002 7983

Table 4. Summary of Optical and XPS Results for the UVI/LaPO4 System sample preparation 10-4

[U]ini ) 2 × M, NaClO4, pH ) 3 [U]ini > 2 × 10-4 M, NaClO4, pH ) 3 [U]ini ) 2 × 10-4 M, KNO3, pH ) 3

lifetime (µs)

XPS EB(U 4f7/2) (eV)

interpretation

80 75 210 80 250

382.5 382.5 381.0 382.5 381.5

tPOH + UVI tPOH + UVI tLaOH + UVI tPOH + UVI tPOH + UVI + nitrate

strong site strong site weak site strong site strong site

in that case, that the surface complex could be rather a tridentate complex linked to the oxygen atoms of the surface PO3 groups. Like the U/LaPO4 system, there is one phosphorus atom at 2.73 Å. Therefore, the sorbed uranyl species form an inner-sphere polydentate complex as in the U/LaPO4 system. Conclusion

Figure 8. Emission spectra of UVI sorbed on La(PO3)3 in KNO3 (straight line) and NaClO4 (squares) with [UVI]ini ) 2 × 10-4 M and precipitate (dots).

Figure 9. U 4f XPS spectra of UVI sorbed on La(PO3)3 for [UVI]ini ) 10-2 M (top) and for [UVI]ini ) 10-3 M (bottom).

hydrated for 7 days. Therefore, for this solid as well, the uranyl precipitation is negligible in our experimental conditions. To check whether the complexant agents, such as NO3ions, have some influence on the sorption process, we prepared the sorbed phosphate in a nitrate solution (KNO3, 0.5 M) at pH equal to 1, 2, and 3 with a uranium concentration at 2 × 10-4 M. The corresponding emission spectra (494.5, 516.0, 539.2, and 566.0 nm, with a width around 90 nm for the first peak) are identical and at the same position as that of the samples prepared in the perchlorate medium. Nevertheless, the decay curves were fitted to three monoexponential laws, which yields three different characteristic lifetime values at 130, 330, and 620 µs. The first two values correspond to the sorption sites already determined, whereas the third one (620 µs) probably accounts for the UO2(NO3)+ ions sorbed on one of the two identified surface sites. The X-ray absorption experiments have given the same distance U-Oaxial at 1.77 Å for all samples. We observed three oxygens at 2.33 Å and two others at 2.50 Å. It seems,

To identify and localize the uranyl sorbed species on the lanthanum monophosphate and polytrioxophosphate, laser spectrofluorimetry, XPS, and X-ray absorption spectroscopy have been carried out. The nature of sorption sites on the two different phosphate surfaces has been characterized as well. The measured lifetime values allow one to assume that there are two different types of reactive sites on the surfaces of both solids. This assumption is corroborated by the interpretation of the U 4f XPS spectra. The variation of the XPS spectra and the decay times as observed in spectrofluorimetry versus the total uranium concentration suggest that one site (the “strong site”) interacts with uranyl ions at low concentration (less than 10-3 M, pH ) 3) and the other one (the “weak site”) is active for an initial uranium concentration bigger than 10-3 M (especially in the case of the UVI/LaPO4 system). Moreover, in the case of the lanthanum polytrioxophosphate, the variation of the ratio between the two observed sites seems to indicate that one sorption site becomes predominant at high uranium concentration. Hence, the surface density of this site is much larger than that of the other one. The effects of the background were studied as well. For both sorbed solids prepared in the nitrate medium (KNO3, 0.5 M), the spectroscopic results show clearly that there are the UO22+ aquo ion and the (UO2)NO3+ complex located on the surface. The interpretation of the EXAFS spectra gave the distances between the uranyl sorbed species and the surface oxygen atoms. According to these results, we can assume that there are no hydration water molecules between the sorbed complexes and the phosphate surface. Moreover, the coordination numbers obtained for the first coordination shells suggest that the uranyl species are sorbed onto both surfaces as an inner-sphere mononuclear polydentate surface complex. In contrast, the discrimination between the two different sites is however difficult. These structural investigations at the molecular level are of primary importance because they allow experimental definition of the different components of the sorption equilibria such as the surface functional groups and the sorbed species. These results, coupled with the uranyl speciation in solution, will then be used to model the retention data and calculate the corresponding sorption constants. Then, the obtained values could be incorporated in radionuclide transport codes as a retardation factor. Acknowledgment. We thank Jacques Lambert from the LCPME at Villers-le`s-Nancy for his assistance in acquiring XPS spectra. EXAFS experiments were per-

7984

Langmuir, Vol. 18, No. 21, 2002

formed at the ESRF in Grenoble on the ROBL beamline, and the authors thank Dr. T. Reich and Dr. C. Hennig for their help during the measurements. We are grateful for the help of Dr. G. Lagarde for PIXE measurements

Ordon˜ ez-Regil et al.

performed at the AGLAE facility at the Laboratoire des Muse´es de France in Paris. LA025674X