Bacterial Siderophores Promote Dissolution of UO2 under Reducing

Nuclear waste repository designs (2, 3) and groundwater remediation strategies often exploit these properties of uranium by promoting reducing conditi...
0 downloads 0 Views 147KB Size
Download PDF
Environ. Sci. Technol. 2005, 39, 5709-5715

Bacterial Siderophores Promote Dissolution of UO2 under Reducing Conditions

uncertainties about UIV chemistry in the environment remain, and even the solubility of UIV solid phases and UIV hydrolysis constants are still debated (7-12). Moreover, Brainard et al. (13) recently demonstrated that organic ligands can increase the solubility and dissolution kinetics of tetravalent actinide minerals such as PuO2.

SCOTT W. FRAZIER, RUBEN KRETZSCHMAR, AND STEPHAN M. KRAEMER* Institute of Terrestrial Ecology, Swiss Federal Institute of Technology Zu ¨ rich, Grabenstrasse 3, Schlieren CH-8952, Switzerland

A group of organic ligands that may have an important effect on tetravalent actinide mobility are siderophores (13, 14). Siderophores [e.g. desferrioxamine-B (DFO-B), Figure 1] are an interesting and structurally diverse (∼500 known structures) category of biogenic ligands that are common in the soil environment (15, 16). Siderophores are produced and released into the soil by some bacteria, fungi, and plants as part of an Fe acquisition strategy. Siderophores are mainly released in environments where the bioavailability of Fe is limited by the low solubility of iron oxides at circumneutral pH. Siderophore concentrations in soils are usually in the sub-micromolar concentration range, but elevated concentrations have been observed in the rhizosphere (17). Little is known about the speciation of soluble siderophores in terrestrial environments. One-to-one FeIII-siderophore complexes have large stability constants, usually many orders of magnitude greater than those of siderophore complexes of other common divalent and trivalent cations that naturally occur in soils. The dissolution rates and solubilities of iron oxides increase in the presence of siderophores such that the nutritional demands of their producers are met (14, 1719). Interestingly, tetravalent actinides are similar to FeIII in several ways that determine their coordination chemistry (e.g. similar charge to ionic radius ratio and first hydrolysis constants) and their 1:1 complexes with DFO-B can have similarly high stability constants (30.6 for FeIII, 30.8 for PuIV, and 26.6 for ThIV) (20). Hence, these are exciting compounds to study in the context of environmental radionuclide mobility, since siderophores may well mobilize tetravalent actinides as they do iron. The similarity of the stability constants of the ferric and actinide complexes of DFO-B implies that these metals would compete for complexation.

Tetravalent actinides are often considered environmentally immobile due to their strong hydrolysis and formation of sparingly soluble oxide phases. However, biogenic ligands commonly found in the soil environment may increase their solubility and mobility. We studied the adsorption and dissolution kinetics of UO2 in the presence of a microbial siderophore, desferrioxamine-B (DFO-B), under reducing conditions. Using batch and continuous flow stirred tank reactors (CFSTR), we found that DFO-B increases the solubility of UIV and accelerates UO2 dissolution rates through a ligandpromoted dissolution mechanism. DFO-B adsorption to UO2 followed a Langmuir-type isotherm. The maximum adsorbed DFO-B concentrations were 3.3 µmol m-2 between pH 3 and 8 and declined above pH 8. DFO-B dissolved UO2 at a DFO-B surface-saturated net rate of 64 nmol h-1 m-2 (pH 7.5, I ) 0.01 M) according to the first-order rate equation R ) kL[Lads], with a rate coefficient kL of 0.019 h-1. Even at very low siderophore concentrations (e.g. 1 µM), net dissolution rates (16 nmol h-1 m-2, pH 7.5, I ) 0.01 M) were substantially greater than net proton-promoted dissolution rates (3 nmol h-1 m-2, pH 7-7.5, I ) 0.01 M). Interestingly, adding dissolved FeIII had negligible effects on DFO-B-promoted UO2 dissolution rates, despite its potential as a competitor for DFO-B and as an oxidant of UIV. Our results suggest that strong organic ligands could influence the environmental mobility of tetravalent actinides and should be considered in predictions for nuclear waste storage and remediation strategies.

Introduction The most important factor influencing the mobility of uranium in the environment is its oxidation state (1). In aerobic environments the UVI oxidation state predominates, whereas the UIV oxidation state is stable only under sufficiently reducing conditions. UIV is considered less mobile than UVI due to the low solubility of UIV-bearing minerals such as uraninite (UO2), pitchblende (impure and poorly crystalline UO2+x), or coffinite (U(SiO4))1-x(OH)4x). Nuclear waste repository designs (2, 3) and groundwater remediation strategies often exploit these properties of uranium by promoting reducing conditions sufficient enough to stabilize or precipitate UIV minerals [e.g. by promoting bacterial UVI reduction (4-6)]. This presumably prevents extensive aqueous environmental uranium migration. However, large * Corresponding author phone: +41 44 633 6077; fax: +41 44 633 1118; e-mail: [email protected]. 10.1021/es050270n CCC: $30.25 Published on Web 07/02/2005

 2005 American Chemical Society

Iron limitation and siderophore release is more likely to occur in oxic environments than under reducing conditions due to the low solubility of FeIII compared to FeII. The hexavalent redox state of uranium is thermodynamically stable in the presence of oxygen. However, Mironov et al. (21) found that greater than 60% of the uranium spread into the environment by the Chernobyl accident remained in the tetravalent form 12-14 years after the accident, demonstrating that solid UIV-bearing phases can exist in oxic environments for decades as a result of slow oxidation kinetics. On these time scales, siderophores might have a significant effect on UIV mobility in oxic environments. In addition, tetravalent actinides share similar trends in chemical properties. Therefore, general observations about UIV mineral dissolution can be expected to apply to the minerals of other actinides, such as Pu and Th, where the tetravalent state is stable in oxic environments. The purpose of this study is to investigate the effect of bacterial siderophores on UIV mobility. We studied a model system composed of DFO-B and synthetic UO2 in both batch and continuous flow stirred tank reactors (CFSTR) under reducing conditions. UO2 was synthesized and characterized, and DFO-B adsorption to UO2 and DFO-B-promoted UO2 dissolution rates at pH values between 3 and 10 were investigated. Also, the effect of increasing soluble FeIII-DFO-B concentrations on net dissolution rates of UO2 was studied. VOL. 39, NO. 15, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5709

FIGURE 1. Molecular structure for desferrioxamine-B (DFO-B), a microbial siderophore. The three hydroxamate groups chelate FeIII in a hexadentate coordination. The terminal amine group carries a positive charge below pH 8.3.

Materials and Methods Adsorption and dissolution studies were performed in a glovebox with a H2/N2 atmosphere circulated through a Pd catalyst. All solutions were prepared from high-purity water treated for a minimum of 48 h with H2 gas and Pd (wire mesh catalyst) to remove oxygen. All reagents were analytical grade. Anaerobic conditions were monitored using resazurinimpregnated cotton strips (Oxoid BR55). The reduced forms of resazurin [resorufin (pink)/dihydroresorufin (colorless)] are reported to have a formal potential versus a standard hydrogen electrode of -0.051 V at pH 7 (22), which is below the reduction potential for the UVIO2(OH)2(aq)/UO2(s) redox couple (5). This procedure proved successful in maintaining oxygen concentrations low enough for our experimental interests. UO2 Preparation. Uranyl acetate (Fluka) was converted mostly to uranyl perchlorate using an anion-exchange resin (DEAE Sephadex, Sigma). Then the pH of an approximately 1 M uranyl perchlorate solution was adjusted to pH 2-3. Following the procedure of Bruno et al. (23), we reduced the UVI to UIV by bubbling H2 gas through the solution while in the presence of a Pd catalyst. The uranium reduction was monitored by following the disappearance of the UVI peaks in the UV-vis absorbance spectrum (Cary 1E UV-visible spectrophotometer, Varian Inc.). After 2 weeks we precipitated a poorly crystalline uranium oxide phase by neutralizing the solution with the addition of NaOH. This precipitate was then centrifuged, decanted, and rinsed with deoxygenated, high-purity water four times as a desalting step. The large solubility of the uranium oxide phase at this step in treatment (20-30 µM at pH ) 7, I ) 0.01), compared to the known solubility for amorphous UO2 (8, 24), indicated that significant concentrations of UVI remained. Further treatment with H2 gas and a Pd catalyst failed to reduce the residual UVI. We therefore further reduced the uranium oxide by heating to 650 °C in a H2 atmosphere for 7 h. This resulted in a dark brown UO2 solid with a diminished solubility (∼0.1 mM at pH ) 7, I ) 0.01). The UO2 was stored in a H2/N2 atmosphere circulated through a Pd catalyst. UO2 Characterization. We characterized the synthetic UO2 using XRD, gas adsorption (multipoint-N2 BET specific surface area, Gemini 2360, Micrometrics), and electrophoretic mobility measurements. The XRD spectrum of the synthetic uranium oxide was measured using a Cu KR X-ray target source using a step-size of 0.01° and a step-rate of 10.0 s per step (D4 Endeavor, Bruker AXS). To prevent sample oxidation during measurement, the powdered UO2 was sealed in the XRD sample holder using a Mylar film (TF-160, Premier Lab Supply, Inc.) while working in a glovebox and then transferred to the instrument. The Mylar film causes a broad but not intense peak in the spectrum centered at 24.2° 2θ. This broad peak does not overlap with any of the UO2 peaks. The isoelectric point (IEP) of the synthetic UO2 was determined using electrophoretic mobility measurements (Zeta PALS, Brookhaven Instruments Corp.). Briefly, 100 mg L-1 UO2 suspensions (I ) 0.01M) were titrated to the desired pH in 5710

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 15, 2005

the glovebox; for each pH value 3 mL of the suspension was transferred to a cuvette, the electrodes were inserted, and then the sealed cuvettes were removed from the glovebox and immediately analyzed. Adsorption of DFO-B on UO2. The DFO-B was used as received (mesylate salt, Desferal, Ciba Geigy). The adsorption of DFO-B onto UO2 was studied in batch experiments (5-mL batch size in 10-mL polyethylene containers). The UO2 solids concentration was 11.1 g L-1 and the ionic strength was fixed at 0.01 M with sodium perchlorate (Merck). Blanks (without UO2) were prepared to account for DFO-B adsorption to experimental materials. All pH adjustments were made with deoxygenated perchloric acid (Merck) and sodium hydroxide (Merck) solutions. The pH was measured (Metrohm combined glass pH electrode) at the beginning and end of an adsorption measurement (after filtration), and the final pH values are reported. We did not use pH buffers for adsorption experiments. Therefore, the pH fluctuated between 7.3 and 7.7. DFO-B dissolves UO2 during the course of an adsorption experiment, potentially leading to less straightforward interpretations. An investigation of the adsorption kinetics showed that adsorbed concentrations remained constant between 10 and 120 min in batch adsorption experiments but that soluble uranium concentrations increased above 15 µM at DFO-Btot > 100 µM. Therefore, the equilibration time was limited to 20 min. The concentration of dissolved DFOB-UIV complexes remained below 6 µM within this time. After 20 min, the solutions were filtered through 0.025 µm filters (Schleicher & Schuell). Discarding the first several drops of filtrate minimized the error introduced by adsorption of DFO-B to the filters. The DFO-B concentrations in solution after filtration were determined by measuring the UV-vis absorbance of the FeDFO-B complex (18). For 2.5 mL of filtered sample, we added 8 µL of concentrated perchloric acid and 85 µL of a 12 mM FeIII stock solution [prepared from FeIII perchlorate hexahydrate (Aldrich)], waited at least 10 min for the complex to form, and then measured the absorbance at 432 nm. DFO-B-Promoted Dissolution Studied in Batch Reactors. DFO-B-promoted UO2 dissolution experiments were performed in 100-mL batch reactors. The ionic strength was fixed at 0.01 M using sodium perchlorate. The pH was buffered at 7.0 using 0.005 M MOPS (3-[N-morpholino]propanesulfonic acid, Fluka). Samples were taken periodically from the batches, filtered using 0.025 µm filters, and measured for total dissolved U using ICP-MS (7500 Series, Agilent). DFO-B-Promoted UO2 Dissolution Studied Using CFSTR. The CFSTRs have two crucial advantages over batch reactors for studying dissolution kinetics. Primarily, any oxidized surface species or especially reactive UO2 phases can be dissolved before results are collected for calculating dissolution rates. Reactants are constantly replaced and products removed from the CFSTR. Therefore, dissolution rates can be measured at low ligand concentrations. In contrast, working at low ligand concentrations usually leads to quick equilibration in batch reactors. We tested several different reactor designs (Teflon, Plexiglas, Duran, glass) and filter materials (cellulose acetate, cellulose nitrate, polycarbonate, glass fiber) to minimize UO2 adhesion to the reactor surfaces. The 60-mL Duran reactors equipped with polysulfone filter membranes (GR40PP, Dow Chemical) maintained the UO2 as a suspension most successfully. Influent solutions were kept under H2/N2 headpressure, and the flow to the CFSTR was controlled by a multihead peristaltic pump. The effluents from the CFSTR were collected or diverted to waste containers. Though the system is closed from the surrounding atmosphere, the entire apparatus was placed in a glovebox under a H2/N2 atmosphere (circulated through a Pd catalyst) to further ensure that oxygen was excluded. The reactors were stirred with

FIGURE 2. XRD spectrum of UO2. The bars above the peaks indicate the range of UO2 diffraction angles observed for International Center for Diffraction Data (ICDD) reference numbers 05-0550, 13-0225, 201344, 65-0285, 65-0286, 65-0287, 65-0288, 75-0420, 75-0424 and the peak centers for our UO2. magnetic stir bars at 300 rpm. Optimal flow rates were between 0.1 and 0.5 mL min-1 as these resulted in measurable uranium concentrations that were still far from equilibrium (as estimated from batch experiments). The UO2 solids concentration was always 1.0 g L-1. The ionic strength was fixed at 0.01 M with sodium perchlorate. No pH buffers were used. Effluent samples were collected for pH measurements and total dissolved U analyses (ICP-MS). Dissolution rates were calculated from flow rates and total dissolved U concentrations in the effluent using the equation

q R ) [U]eff aV

(1)

where [U]eff is the total dissolved U concentration (µM) in the reactor outflow, q is the flow rate (mL h-1), a is the surface area (m2) of UO2 in the reactor, and V is the reactor volume (mL). R is the surface-area-normalized (µmol h-1 m-2) dissolution rate (25). Once the reactors reached steady-state for a particular set of conditions (constant total dissolved U concentrations in the effluent), samples were taken for several days and the total dissolved U concentrations were averaged to calculate dissolution rates.

Results and Discussion UO2 Characterization. The XRD spectrum of the uranium oxide compares well with the database spectrum for synthetic UO2 (Figure 2). The diffraction peaks fall within the range spanned by that of the reference spectra for UO2 [International Center for Diffraction Data (ICDD) reference numbers 05-0550, 13-0225, 20-1344, 65-0285, 65-0286, 65-0287, 650288, 75-0420, 75-0424]. The variations of published XRD patterns may be due to annealing temperatures, the presence of impurities, and radiation damage (26). The UO2 used in our study most closely resembles ICDD reference number 65-0285. Notably, there is an absence of diffraction peaks that would result from minerals containing significant amounts of UVI relative to UIV. The low peak intensity and broad peaks suggest that this UO2 is poorly crystalline. The N2 BET specific surface area of the synthetic UO2 was 3.94 m2 g-1. The IEP of the synthetic UO2 was at pH 5.4 (Figure 3). This agrees well with previously published point-of-zerocharge values for UO2 (27). The two outlying measurements in Figure 3 (base titration at pH 5.6-6.0) could potentially indicate an alteration of the surface during the course of the titrations (27). DFO-B Adsorption to UO2. The initial steps of a ligandpromoted dissolution mechanism require adsorption of the ligand to the mineral surface (28, 29). Hence, characterizing

FIGURE 3. pH-dependent electrophoretic mobility of UO2. The UO2 exhibited zero electrophoretic mobility (the isoelectric point) at pH 5.4. Both acid and base titrations of the UO2 were performed in a CO2- and O2-free glovebox using 100 mg L-1 UO2 suspensions of 100 mL. Samples were taken from the suspension, sealed in cuvettees with the electrode, and quickly transferred to the instrument. We verified that the electrophoretic mobility was not affected by oxidation during transfer from the glovebox to the instrument. the adsorption of a ligand to the mineral surface is necessary to understand the dissolution mechanism. DFO-B adsorption to UO2 at pH 7.5 initially increased sharply with increasing dissolved DFO-B concentration and plateaued at higher DFO-B concentrations (Figure 4A). The adsorption of DFO-B to UO2 was fit by a least squares approximation to the Langmuir equation

[L] n ) nmaxb 1 + b[L]

(2)

where n is the adsorbed ligand concentration (µmol m-2), nmax is the maximum adsorbed ligand concentration (µmol m-2), b is an affinity parameter (µM-1), and [L] is the ligand concentration in solution (µM). The maximum adsorbed ligand concentration (nmax) for this DFO-B and UO2 system at pH 7.5 was 3.3 µmol m-2 and b was 0.021 µM-1. The maximum adsorbed surface concentration observed here is much higher than that for DFO-B adsorption onto goethite [0.034 µmol m-2, pH 6.5, I ) 0.01 M (18)]. Cheah et al. (18) and Cocozza et al. (30) partly attributed the poor DFO-B adsorption on goethite to steric effects and electrostatic repulsion between the cationic DFO-B and the positively charged goethite surface at their experimental pH. Between pH 5.4 (the IEP of the UO2 surface) and pH 8.3 (pKa1 for DFO-B), the ligand experiences electrostatic attraction to the UO2 surface. DFO-B adsorption to UO2 is constant between pH 3 and 8 (Figure 4B) despite the reversal of UO2 surface charge in this pH range. This implies that electrostatic interactions have only a small effect on DFO-B adsorption to the UO2 surface. DFO-B adsorption to UO2 declines rapidly above pH 8 (Figure 4B). This is the range where the hydroxamate groups of soluble DFO-B species become deprotonated (pKa2 9.06 and pKa3 9.73). Therefore, one could suspect that a net negatively charged surface complex forms at high pH, exerting an electrostatic effect on adsorption. DFO-B-Promoted Dissolution of UO2. The batch dissolution experiments show that DFO-B effectively accelerates UO2 dissolution and increases UO2 solubility (Figure 5). We observed a substantial amount of fast initial dissolution (between 0 and 200 h), which was followed by steady-state dissolution. For example, approximately constant dissolution rates of 30 nmol h-1 m-2 were observed in the presence of 200 µM DFO-B at t > 100 h. Fast initial dissolution of other minerals such as δ-Al2O3 (25) and goethite (31) has been VOL. 39, NO. 15, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5711

FIGURE 4. Adsorption of desferrioxamine-B onto UO2. (A) The adsorption isotherm data be described by a Langmuir equation (see eq 2 in text) with a maximum adsorbed ligand concentration (nmax) of 3.3 µmol m-2 and b ) 0.21 µM-1. The pH-dependent adsorption of DFO-B onto UO2 (B) was constant between pH 3 and 8 (DFO-Btot ) 400 µM). DFO-B adsorption onto UO2 rapidly declined as pH increased above pH 8, in the range where DFO-B changes from positively to negatively charged. Dominant DFO-B solution species are indicated as a function of pH. Experimental conditions were 11.1 g L-1 UO2, pH 7.5, I ) 0.01 M, 20 min equilibration time, and suspension separated using 0.025 µm filters.

FIGURE 5. Batch desferrioxamine-B and UO2 dissolution experiment. Experiments included either no DFO-B (control), 20 µM DFO-B, or 200 µM DFO-B. The 0 M DFO-B and 20 µM DFO-B experiments quickly come to equilibrium as a result of fast initial dissolution. Steady-state dissolution is observed only in the 200 µM DFO-B experiments at t > 200 h. Experimental conditions were 0.35 g L-1 UO2, pH 7.0, I ) 0.01 M, and suspension separated using 0.025 µm filters. observed previously. We did not investigate the mechanism of fast initial dissolution, as steady-state dissolution is more 5712

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 15, 2005

FIGURE 6. Desferrioxamine-B-promoted UO2 dissolution measured using continuous flow stirred tank reactors. The net UO2 dissolution rates are compared as a function of the DFO-B concentration in solution (A) or the adsorbed DFO-B concentration, where there is a linear dependence (B) with a slope of 0.0186 h-1 representing the rate constant kL of DFO-B-promoted dissolution. Experimental conditions were 1.0 g L-1 UO2, pH 7.5, and I ) 0.01 M. Error bars indicate the standard deviation. representative of long-term weathering of UO2 in the environment. However, the fast initial dissolution reactions make it difficult to observe steady-state dissolution rates in batch experiments, particularly at low ligand concentrations that might be environmentally relevant. Therefore, we designed the CFSTR where the reactive species contributing to the fast initial dissolution reactions were depleted before measuring steady-state dissolution rates. Batch experiments demonstrated that oxidative dissolution processes do not dominate the UO2 dissolution kinetics and solubility observed herein, since dissolved U concentrations remained between 0.221 and 0.077 µM during the course of the experiment in the absence of DFO-B (Figure 5). Also, maximum dissolved U concentrations of 10 µM in the presence of 20 µM DFO-B suggest the formation of 1:2 U-DFO-B complexes or of 1:1 complexes of lower stability. The latter hypothesis is unlikely considering that the stability of other 1:1 tetravalent actinide complexes such as ThIV-DFO-B and PuIV-DFO-B approach or exceed the stability of the iron complex (32). The steady-state net dissolution rate (measured using CFSTR) vs dissolved DFO-B concentration (pH 7.5 and I ) 0.01 M) followed a Langmuir-type relationship where net dissolution rates increase sharply with increasing ligand concentration but plateau at higher ligand concentrations (Figure 6A). The surface-saturated (as determined by adsorption experiments) DFO-B-promoted net dissolution rate was 64 nmol h-1 m-2. Dissolution rates measured using CFSTR at pH 7.5 were on the same order of magnitude as

those measured using the batch reactor at pH 7.0 (see above). The surface-saturated DFO-B-promoted net UO2 dissolution rate is substantially greater than the rate for the DFO-Bpromoted dissolution of goethite (0.57 nmol h-1 m-2 at pH 6.5, I ) 0.01 M) (18). The DFO-B-promoted net UO2 dissolution rate is on the same order of magnitude as the observed oxidative dissolution rate for crystalline UO2 of 27 nmol h-1 m-2 between pH 6.6 and 8.8 in solutions that were equilibrated with a 21% O2 atmosphere in the absence of CO2 (33). Significantly higher oxidative dissolution rates were observed at elevated soluble bicarbonate concentrations (34). According to the model for the ligand-promoted dissolution of mineral oxides advanced by Stumm and co-workers (28, 29), the dissolution mechanism consists of several steps: ligands first diffuse to the surface of a metal oxide and adsorb to a metal oxide center on the surface through ligandexchange reactions. The formation of an inner-sphere surface complex weakens bonds between the metal center and its neighboring metal atoms, facilitating detachment of the complexed metal. Finally, the metal-ligand complex diffuses into the solution and the surface protonation state is rapidly reattained. In this model, the dissolution kinetics is controlled by the rate-limiting step, which is considered to be the detachment of the metal center from the surface of the metal oxide. Adsorption and protonation reactions are often relatively fast and an equilibrium adsorbed ligand concentration is established. The ligand controlled dissolution rate is thus dependent on the adsorbed ligand concentration, since the adsorbed ligands (in equilibrium with solution) are the reactants in the rate-limiting reaction (removing the metal center from the bulk oxide). A rate law for the ligandpromoted dissolution of a mineral oxide can then be written as

Rdiss ) kL[>ML]

(3)

Where Rdiss is the ligand-promoted dissolution rate (µmol h-1 m-2), kL is the rate constant (h-1), and [>ML] is the adsorbed ligand concentration (µmol m-2). Here, the ligandpromoted dissolution rate is a linear function of the adsorbed ligand concentration. We do indeed find our UO2 dissolution rates are a linear function of the adsorbed DFO-B concentration (Figure 6B). The slope of the regression line or kL is 0.019 h-1, which is similar to the rate constants reported for the DFO-B-promoted dissolution of goethite (0.015 h-1) (18). pH-Dependent DFO-B-Promoted and Proton-Promoted Dissolution of UO2. Measured net UO2 dissolution rates in the presence of DFO-B are lowest between pH 7 and 8 and increase at both higher and lower pH (Figure 7A). To determine the ligand-promoted dissolution rates, one must subtract the proton- and hydroxide-promoted dissolution rates from the net dissolution rates. We could successfully measure the proton-promoted dissolution rates at and below pH 7.5. Above pH 7.5 it is experimentally very difficult to measure dissolution rates because the solubility of UO2 is so low that the system quickly reaches equilibrium, thus making it impossible to measure kinetics. Measured protonpromoted dissolution rates (2.6-2.8 nmol h-1 m-2) at pH 7.5 correspond to the y-intercept in Figure 6B (2.7 nmol h-1 m-2), indicating that equilibrium was not attained in dissolution experiments without DFO-B. Decreasing proton-promoted dissolution rates of crystalline UO2 between pH 3 and 7 have also been observed previously (35). They observed constant proton-promoted dissolution rates of 6.8 nmol h-1 m-2 between pH 7 and 11. Interestingly, the rate coefficient of ligand-controlled dissolution kL varies with pH (Figure 7B). It has a minimum at pH 7.5 and increases toward higher and lower pH. By comparison between the net dissolution rates (square points in Figure 7A) and the DFO-B-promoted dissolution rates

FIGURE 7. Influence of pH on UO2 dissolution. The pH-dependent net, proton-promoted, and desferrioxamine-B-promoted UO2 dissolution rates (A) and kL (B, see eq 3 in text) are at a minimum around neutral pH. Rates were measured using the continuous flow stirred tank reactors. Experimental conditions were 1.0 g L-1 UO2, [DFO-B] ) 100 µmol, and I ) 0.01 M. Error bars indicate standard deviation. Error bars are smaller than symbols for the protonpromoted dissolution rates. (circular points, net dissolution rates minus proton-promoted dissolution rates, Figure 7A), we see that DFO-B controls the dissolution rate above pH 5. Proton-promoted dissolution rates accounted for 10-30% of the net dissolution rate at pH 3 and 4. The impact of DFO-B on UO2 dissolution kinetics and the solubility of UO2 compared to proton-promoted dissolution is perhaps the most striking feature illustrated in Figure 7. It illustrates that strong organic ligands could have a significant affect on UO2 mobility under reducing conditions. DFO-B-Promoted Dissolution of UO2 in the Presence of FeIII. Siderophores have a high affinity for FeIII and, to a lesser degree, other metal ions. Therefore, it is likely that metal complexes will dominate siderophore speciation in ground and surface waters. Brainard et al. (13) found that Fe enterobactin complexes accelerated PuO2 dissolution rates relative to uncomplexed enterobactin (a bacterial catecholate siderophore). Hence, we performed steady-state UO2 dissolution experiments in the presence of variable concentrations of soluble FeIII. Conceptually, FeIII can affect UO2 dissolution kinetics in various ways. Competition between FeIII and UIV for complexation by DFO-B could potentially reduce the solubility of UIV and decrease the dissolution rates by reducing the surface concentrations of DFO-B. Second, if FeIII is released in a metal-exchange reaction with UIV, it could promote oxidative UO2 dissolution. Finally, the FeIII-DFO-B complex could serve as oxidant, again promoting oxidative dissolution of UO2. VOL. 39, NO. 15, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5713

made by Bruno Nu ¨ ssberger (ETHZ, Physics Department). We are grateful for the advice and polysulfone filters provided by Gerhard Furrer and for thorough revisions by three anonymous reviewers. This work was funded by the Swiss Federal Institute of Technology Zu ¨ rich (TH Project ID 7160).

Supporting Information Available A table of thermodynamic formation constants. This material is available free of charge via the Internet at http:// pubs.acs.org.

Literature Cited

FIGURE 8. Net desferrioxamine-B-promoted UO2 dissolution rates as a function of the dissolved FeIII concentration. Rates were measured using the continuous flow stirred tank reactors. Experimental conditions were 1.0 g L-1 UO2, [DFO-Β] ) 100 µM, pH 7.5, and I ) 0.01 M. Error bars indicate the standard deviation. However, we saw no substantial effect of FeIII on UO2 dissolution kinetics (Figure 8). It is likely that no competitive metal-exchange reaction occurred because UIV concentrations in the reactor effluent were always below 1 µM, which is much less than the minimum free siderophore concentration of 10 µM (at [FeIII] ) 90 µM). Such competitive effects may be more important at equimolar iron and siderophore concentrations. Lowering the free siderophore concentration may have reduced the adsorbed ligand concentration. However, in the absence of Fe, observed dissolution rates at [DFO-B] ) 10 µM were only slightly lower than dissolution rates at [DFO-B] ) 100 µM due to the Langmuir-type relationship between rates and siderophore concentrations. FeIII species can have greater half-cell reduction potentials compared to UVI species (5), suggesting that FeIII could be reduced to FeII while oxidizing UIV to UVI, thus dissolving UO2 through an oxidative dissolution mechanism. However, UVI reduction by FeII adsorbed at hematite surfaces (36) and by pyrite surfaces (37) has been observed. The reduction potential of the FeIII(OH)2+/FeII(OH)2 pair is 0.02 V at pH 7 but can range from 0.33 to 0.52 V for FeIII adsorbed on iron oxide or silicate surfaces (38). Complexation by DFO-B strongly stabilizes the trivalent redox state of iron, resulting in a reduction potential of -0.48 mV for the reaction FeIII-HDFO-B+ + e- T FeII-HDFO-B+ as observed by cyclic voltametry (15). This is much lower than the calculated redox potential for the redox pair UVIO2(OH)2(aq)/UO2(s) of 0.14 V at pH 7 (5), where UVIO2(OH)2(aq) represents the dominant hexavalent uranium species in the pH range between 6.5 and 8.5 (39). Therefore, oxidative dissolution of UO2 by FeIII-DFO-B complexes is not expected. We have shown that siderophores can dramatically increase both the solubility and dissolution kinetics of UO2 and thus potentially increase UIV mobility. This is true for the entire pH range between pH 3 and 10. Surprisingly, these strong biogenic ligands designed for mobilizing Fe can be even more effective at promoting UO2 dissolution than they are at promoting Fe dissolution from goethite, a crystalline iron oxide. We have also shown that Fe had negligible affects on DFO-B-promoted UO2 dissolution. These results demonstrate that it is important to consider biogenic organic ligands when making predictions regarding the stability of UO2 under reducing conditions.

Acknowledgments We thank Kurt Barmettler and Petra Reichard for their ample assistance in the laboratory. We appreciate Jordi Bruno’s advice on preparing synthetic UO2. The CFSTRs were custom5714

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 15, 2005

(1) Silva, R. J.; Nitsche, H. Actinide environmental chemistry. Radiochim. Acta 1995, 70-1, 377-396. (2) Shoesmith, D. W. Fuel corrosion processes under waste disposal conditions. J. Nucl. Mater. 2000, 282, 1-31. (3) Janeczek, J.; Ewing, R. C. Dissolution and Alteration of Uraninite under Reducing Conditions. J. Nucl. Mater. 1992, 190, 157173. (4) Finneran, K. T.; Anderson, R. T.; Nevin, K. P.; Lovley, D. R. Potential for Bioremediation of uranium-contaminated aquifers with microbial U(VI) reduction. Soil. Sediment. Contam. 2002, 11, 339-357. (5) Fredrickson, J. K.; Kostandarithes, H. M.; Li, S. W.; Plymale, A. E.; Daly, M. J. Reduction of Fe(III), Cr(VI), U(VI), and Tc(VII) by Deinococcus radiodurans R1. Appl. Environ. Microbiol. 2000, 66, 2006-2011. (6) Robinson, K. G.; Ganesh, R.; Reed, G. D. Impact of organic ligands on uranium removal during anaerobic biological treatment. Water Sci. Technol. 1998, 37, 73-80. (7) Neck, V.; Kim, J. I. Solubility and hydrolysis of tetravalent actinides. Radiochim. Acta 2001, 89, 1-16. (8) Rai, D.; Felmy, A. R.; Ryan, J. L. Uranium(IV) Hydrolysis Constants and Solubility Product of UO2‚xH2O(am). Inorg. Chem. 1990, 29, 260-264. (9) Rai, D.; Felmy, A. R.; Sterner, S. M.; Moore, D. A.; Mason, M. J.; Novak, C. F. The solubility of Th(IV) and U(IV) hydrous oxides in concentrated NaCl and MgCl2 solutions. Radiochim. Acta 1997, 79, 239-247. (10) Ryan, J. L.; Rai, D. The Solubility of Uranium(Iv) Hydrous Oxide in Sodium-Hydroxide Solutions under Reducing Conditions. Polyhedron 1983, 2, 947-952. (11) Parks, G. A.; Pohl, D. C. Hydrothermal Solubility of Uraninite. Geochim. Cosmochim. Acta 1988, 52, 863-875. (12) Fujiwara, K.; Yamana, H.; Fujii, T.; Moriyama, H. Determination of uranium(IV) hydrolysis constants and solubility product of UO2‚xH2O. Radiochim. Acta 2003, 91, 345-350. (13) Brainard, J. R.; Strietelmeier, B. A.; Smith, P. H.; Langstonunkefer, P. J.; Barr, M. E.; Ryan, R. R. Actinide Binding and Solubilization by Microbial Siderophores. Radiochim. Acta 1992, 58-9, 357363. (14) Ruggiero, C. E.; Matonic, J. H.; Reilly, S. D.; Neu, M. P. Dissolution of plutonium(IV) hydroxide by desferrioxamine siderophores and simple organic chelators. Inorg. Chem. 2002, 41, 35933595. (15) Boukhalfa, H.; Crumbliss, A. L. Chemical aspects of siderophore mediated iron transport. Biometals 2002, 15, 325-339. (16) Kraemer, S. M.; Hering, J. G. Biogeochemical controls on the mobility and bioavailability of metals in soils and groundwaterPreface. Aquat. Sci. 2004, 66, 1-2. (17) Kraemer, S. M. Iron oxide dissolution and solubility in the presence of siderophores. Aquat. Sci. 2004, 66, 3-18. (18) Cheah, S. F.; Kraemer, S. M.; Cervini-Silva, J.; Sposito, G. Steadystate dissolution kinetics of goethite in the presence of desferrioxamine B and oxalate ligands: Implications for the microbial acquisition of iron. Chem. Geol. 2003, 198, 63-75. (19) Romheld, V. The Role of Phytosiderophores in Acquisition of Iron and Other Micronutrients in Gramineous Speciessan Ecological Approach. Plant Soil 1991, 130, 127-134. (20) Hernlem, B. J.; Vane, L. M.; Sayles, G. D. The application of siderophores for metal recovery and waste remediation: Examination of correlations for prediction of metal affinities. Water Res. 1999, 33, 951-960. (21) Mironov, V. P.; Matusevich, J. L.; Kudrjashov, V. P.; Boulyga, S. F.; Becker, J. S. Determination of irradiated reactor uranium in soil samples in Belarus using U-236 as irradiated uranium tracer. J. Environ. Monit. 2002, 4, 997-1002. (22) Tratnyek, P. G.; T. E. R.; Lemon, A. W.; Sherer, M. M.; Balko, B. A.; Feik, L. M.; Henegar, B. D. Visualizing Redox Chemistry:

(23)

(24) (25) (26) (27) (28) (29) (30)

(31)

(32)

Probing Environmental Oxidation-Reduction Reactions with Indicator Dyes. Chem. Educator 2001, 6. Bruno, J.; Casas, I.; Lagerman, B.; Munoz, M. The determination of the solubility of amorphous UO2(S) and the mononuclear hydrolysis constants of Uranium(IV) at 25 °C. Mater. Res. Soc. Symp. Proc. 1987, 84, 153-160. Rai, D.; Yui, M.; Moore, D. A. Solubility and solubility product at 22 degrees C of UO2(c) precipitated from aqueous U(IV) solutions. J. Solut. Chem. 2003, 32, 1-17. Kraemer, S. M.; Hering, J. G. Influence of solution saturation state on the kinetics of ligand-controlled dissolution of oxide phases. Geochim. Cosmochim. Acta 1997, 61, 2855-2866. Janeczek, J.; Ewing, R. C. X-ray powder diffraction study of annealed uraninite. J. Nucl. Mater. 1991, 185, 66-77. Olsson, M.; Jakobsson, A. M.; Albinsson, Y. Surface charge densities of two actinide(IV) oxides: UO2 and ThO2. J. Colloid Interface Sci. 2002, 256, 256-261. Furrer, G.; Stumm, W. The Coordination Chemistry of Weathering 0.1. Dissolution Kinetics of Delta-Al2O3 and BeO. Geochim. Cosmochim. Acta 1986, 50, 1847-1860. Zinder, B.; Furrer, G.; Stumm, W. The Coordination Chemistry of Weathering .2. Dissolution of Fe(III) Oxides. Geochim. Cosmochim. Acta 1986, 50, 1861-1869. Cocozza, C.; Tsao, C. C. G.; Cheah, S. F.; Kraemer, S. M.; Raymond, K. N.; Miano, T. M.; Sposito, G. Temperature dependence of goethite dissolution promoted by trihydroxamate siderophores. Geochim. Cosmochim. Acta 2002, 66, 431-438. Holmen, B. A.; Casey, W. H. Hydroxamate ligands, surface chemistry, and the mechanism of ligand-promoted dissolution of goethite alpha-FeOOH(s). Geochim. Cosmochim. Acta 1996, 60, 4403-4416. Whisenhunt, D. W., Jr.; Neu, M. P.; Hou, Z.; Xu, J.; Hoffman, D. C.; Raymond, K. N. Specific sequestering agents for the actinides. 29. Stability of the thorium(IV) complexes of Desferrioxamine B (DFO) and three octadentate catecholate or hydroxypyridi-

(33)

(34)

(35)

(36)

(37)

(38)

(39)

nonate DFO derivatives: DFMTA, DFOCAMC, and DFO-1,2HOPO. Comparative stability of the plutonium(IV) DFOMTA complex. Inorg. Chem. 1996, 35, 4128-4136. Torrero, M. E.; Baraj, E.; De Pablo, J.; Gime´nez, J.; Casas, I. Kinetics of corrosion and dissolution of uranium dioxide as a function of pH. Int. J. Chem. Kinet. 1997, 29, 261-267. De Pablo, J.; Casas, I.; Gime´nez, J.; Molera, M.; Rovira, M.; Duro, L.; Bruno, J. The oxidative dissolution kinetics mechanism of uranium dioxide. I. The effect of temperature in hydrogen carbonate medium. Geochim. Cosmochim. Acta 1999, 63, 30973103. Bruno, J.; Casas, I.; Puigdomenech, I. The kinetics of dissolution of UO2 under reducing conditions and the influence of an oxidized surface layer (UO2+x): Application of a continuous flow-through reactor. Geochim. Cosmochim. Acta 1991, 55, 647658. Liger, E.; Charlet, L.; Van Cappellen, P. Surface catalysis of uranium(VI) reduction by iron(II). Geochim. Cosmochim. Acta 1999, 63, 2939-2955. Wersin, P.; Hochella, M. F.; Persson, P.; Redden, G.; Leckie, J. O.; Harris, D. W. Interaction between Aqueous Uranium(VI) and Sulfide MineralssSpectroscopic Evidence for Sorption and Reduction. Geochim. Cosmochim. Acta 1994, 58, 2829-2843. Stumm, W. Chemistry of the solid-water interface: Processes at the mineral-water and particle-water interface in natural systems; Wiley: New York, 1992. Murphy, W. M.; Shock, E. L. Environ. Aqueous Geochem. Actinides Rev. Miner. 1999, 38, 221-253.

Received for review February 10, 2005. Revised manuscript received June 3, 2005. Accepted June 9, 2005. ES050270N

VOL. 39, NO. 15, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5715