Atomistic Simulations of Calcium Uranyl(VI) Carbonate Adsorption on

May 29, 2012 - Surface chemistry of carbon dioxide revisited. William Taifan , Jean-François Boily , Jonas Baltrusaitis. Surface Science Reports 2016...
16 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/est

Atomistic Simulations of Calcium Uranyl(VI) Carbonate Adsorption on Calcite and Stepped-Calcite Surfaces Slimane Doudou,†,‡ David J. Vaughan,‡ Francis R. Livens,†,‡,§ and Neil A. Burton*,† †

School of Chemistry, University of Manchester, Manchester, M13 9PL, U.K. Williamson Research Centre for Molecular Environmental Science and School of Earth, Atmospheric and Environmental Sciences, University of Manchester, Manchester M13 9PL, U.K. § Centre for Radiochemistry Research, School of Chemistry, University of Manchester, Manchester M13 9PL, U.K. ‡

S Supporting Information *

ABSTRACT: Adsorption of actinyl ions onto mineral surfaces is one of the main mechanisms that control the migration of these ions in environmental systems. Here, we present computational classical molecular dynamics (MD) simulations to investigate the behavior of U(VI) in contact with different calcite surfaces. The calcium-uranylcarbonate [Ca2UO2(CO3)3] species is shown to display both innerand outer-sphere adsorption to the flat {101̅4} and the stepped {3148̅ } and {312̅ 1̅ 6} planes of calcite. Free energy calculations, using the umbrella sampling method, are employed to simulate adsorption paths of the same uranyl species on the different calcite surfaces under aqueous condition. Outer-sphere adsorption is found to dominate over inner-sphere adsorption because of the high free energy barrier of removing a uranyl−carbonate interaction and replacing it with a new uranyl−surface interaction. An important binding mode is proposed involving a single vicinal water monolayer between the surface and the sorbed complex. From the free energy profiles of the different calcite surfaces, the uranyl complex was also found to adsorb preferentially on the acute-stepped {3148̅ } face of calcite, in agreement with experiment.

1. INTRODUCTION The migration of actinides in the environment is a topical aspect of actinide research. Given the generally trace level concentrations of these species in natural systems, sorption of actinyl ions onto mineral surfaces is one of the dominant mechanisms that control the migration behavior of these species in soils and groundwater systems. There have been many experimental studies of the reactions of U(VI) with the calcite mineral surface in aqueous conditions.1−4 The calcite{101̅4} face is the most stable and dominant of calcite surfaces and it is, therefore, the most likely location for adsorption to occur. Here, we investigate the possibilities for U(VI) adsorption on a number of different calcite surfaces using Molecular Dynamics (MD) simulation techniques. MD approaches have previously been used to study uranyl in aqueous solution,5−8 or to study its interaction with quartz,9 and montmorillonite surfaces.10 One of the aims of this study is to validate the already-available classical potential parameters of the uranyl ion7 and apply them to processes not explored previously for this cation. However, the main objective of the present study is to aid experimental interpretation of uranyl adsorption on calcite surfaces by probing, at an atomic level, the complexes that may form on the surface. Experimentally, the coordination environment of the uranyl ion can be ascertained from spectroscopic measurements; however, such measurements do not always have a definitive interpretation. By © 2012 American Chemical Society

performing atomistic simulations, we can obtain a detailed picture of the species involved in uranyl adsorption on calcite, and explore possible mechanisms for this process using free energy calculations along the adsorption pathway. A range of experimental and speciation studies have previously shown that the most likely uranyl species in solution, under the mildly alkaline conditions typical of calcite in carbonate rich aqueous solution (pH ∼9), is the stable [Ca2UO2(CO3)3] neutral complex.2,4,11,12 For the calculations in this article, we focus on modeling the adsorption of this species to different aspects of the calcite surface. Previous analyses of the calcite surface have shown three common topologies which may form the focus for binding studies.13−15 These are the flat {101̅4} terrace sites and the acute or obtuse steps formed with respect to the {314̅8} and {31̅2̅16} calcite planes respectively (highlighted in Figure 1). It has been noted4 that the observed high surface coverage (∼0.05 to ∼0.7)1 of U(VI) would necessitate considerable sorption to the terraces, although this may involve binding to defects such as small etch pits providing Ca vacancy sites, similar to an acute step. Although it is possible that other surface defects may, for Received: Revised: Accepted: Published: 7587

January 5, 2012 May 6, 2012 May 29, 2012 May 29, 2012 dx.doi.org/10.1021/es300034k | Environ. Sci. Technol. 2012, 46, 7587−7594

Environmental Science & Technology

Article

triscarbonate complex incorporated into calcite, but with a 5fold equatorial coordination.17 Rihs et al4 used X-ray standing waves (XSW) to conclude that U(VI) adsorbs as an innersphere complex to calcite at acute (or “−”) steps or calcium vacancies, and proposed that U(VI) can coordinate in a monodentate fashion to a surface carbonate oxygen atom. Their data resolves a distance between the U atom and the calcite (104) plane of 2.55 Å. The experimental studies, taken overall, indicate that U(VI) can adsorb to the surface as a uranyl carbonate complex and some of the data is consistent with the formation of an inner-sphere complex with U−Oeq distances between 2.3−2.5 Å, typical of five- or six- coordinates complexes; however, the data do not clearly distinguish the possible structures. The extent of formation of an outer-sphere complex is also unclear although there is ambiguity arising from the absence of U−Ca backscattering paths in the EXAFS data which may be consistent with more distantly adsorbed complexes or outer-sphere complexation.19 Weakly adsorbed [Ca2UO2(CO3)3] outer-sphere complexes may be difficult to observe directly unless they can be clearly distinguished from the solution species.

Figure 1. Different calcite surfaces considered: a) the flat {101̅4} surface; b) the acute step between {101̅4} and {314̅8} faces; and c) the obtuse step between the {101̅4} and {31̅2̅16} faces.

example, expose surface calcium ions to potentially provide preferential binding sites, such species have not been identified experimentally and so here we shall focus our investigation on uranyl binding to the more ubiquitous unperturbed surfaces of calcite. In this study, where carbonate is both a ligand of the uranyl triscarbonate adsorbate and also a major component of the surface composition of calcite, there can be some ambiguity in the categorization of the potential sorption complexes, particularly when using the definitions of inner- or outersphere; these terms are often used experimentally to distinguish observed structures on the basis of the proximity of a sorbed atom from the surface. In this work we have employed the term inner-sphere to uniquely describe complexes where a uranyl bound carbonate ion forms part of the formally defined unperturbed surface; one or more of the uranyl bound carbonate ions may be as part of the {101̅4} face or one of the {314̅8} and {31̅2̅16} calcite planes. This implies that binding of a uranyl triscarbonate ion must undergo a full or partial ligand-exchange in order to adsorb in an inner-sphere mode and the U(VI) atom will be bound close to the surface. We have reserved the term outer-sphere to describe the case where the carbonate ions of the uranyl triscarbonate adsorbate remain associated with the complex, and this triscarbonate complex is itself adsorbed to an unperturbed surface with no direct interaction of this surface with the U(VI) atom. Here an adsorption mechanism would not require any uranyl ligand dissociation or exchange, although we are mindful that the sorbed complex could equally be considered to result from the binding of uranyl biscarbonate complex to a carbonate surface defect, perhaps by exchange with a water ligand of such a complex. Sorption mechanisms for uptake of the uranyl ion on calcite have been addressed experimentally1−4,16−18 at different pH and uranyl concentrations, and using various experimental techniques. Both Carroll and Bruno1 and Savenko18 performed batch experiments that modeled uranyl sorption on calcite as an exchange of UO22+ with surface Ca2+; here it was also noted that uranyl adsorption decreases with increasing calcium content in solution. Elzinga et al2 used extended X-ray absorption fine structure (EXAFS) spectroscopy and luminescence methods to identify the complex present at the calcite surface as [Ca2UO2(CO3)3]0. Their EXAFS data indicated not only hydroxide and carbonate precipitates with a strongly split equatorial O shell but also U(VI) species with a slightly disordered equatorial shell, suggesting that the uranyl triscarbonate complex could also be more weakly bound to the calcite surface; unfortunately the methodology was unable to clearly discern further detail of the possible structures. Furthermore, the luminescence data suggested the presence of a second species at the calcite surface, which resembles a uranyl

2. COMPUTATIONAL METHODS All calcite slabs simulated in this study were generated using Materials Studio 4.20 Calcite has a rhombohedral crystal structure, a = b = 4.988 Å, c = 17.061 Å, α = β = 90°, γ = 120° and space group R3c.̅ 21 Slabs of three different calcite surfaces were considered for the models: The planar {1014̅ } surface, and the acute {314̅8} and the obtuse {31̅2̅16} vicinal surfaces. Figure 1 shows the typical arrangements of atoms at these surfaces. The {101̅4} and {314̅8} slabs consisted of 240 CaCO3 units, whereas the {312̅ 1̅ 6} slab consisted of 276 units. The slab sizes were chosen so that the middle layers of the mineral along the vertical direction have bulk crystal properties and are not greatly affected by the movements of the top and bottom layers of the surface during MD simulations. As a result, the thickness of the slabs in the z-direction was at least 20 Å. The vacuum gap above the surface has a thickness of ∼45−50 Å, large enough to avoid any interactions between repeating images of the surfaces in the z-direction under the 3D periodic boundary conditions scheme for MD calculations. As noted previously, our simulation conditions aim to be representative of calcite in an aqueous solution typical of a calcium and carbonate rich groundwater. Under such conditions, pH ∼9, we would expect the calcite surface to typically be unprotonated. The potential of zero charge for calcite has been the subject of experimental study and, as summarized by Stipp,15 was not found to be a unique measure which can provide insight into the surface composition of this mineral. The complex [Ca2UO2(CO3)3] was inserted into the system at a distance above the calcite surfaces, described later in section 2.2, and the vacuum gap filled with water molecules using the DLPOLY graphical user interface. Inserting the calcium−uranyl−carbonate species in bulk water far away from the mineral surface should not lead to any adsorption, since that species is very stable in aqueous solution. The potential parameters for the calcite surface developed by Pavese et al.22 were used, along with the Simple Point Charge (SPC) water model23 for the solvent. The force field parameters for the UO22+ cation introduced by Guibaud et al,7 which have previously been developed and fitted to the hydration energies of uranyl in aqueous solution, were employed. The water-calcite interaction parameters were 7588

dx.doi.org/10.1021/es300034k | Environ. Sci. Technol. 2012, 46, 7587−7594

Environmental Science & Technology

Article

obtained from the Kerisit et al.24 study of water adsorption on calcite. Finally, the uranyl-calcite and other nonbonded pair interactions were calculated based on the van der Waals equation Vij =

Aij r

12

−

The statistical data from all simulations can then be pieced together and the probability of the reaction coordinate having a particular value, p(ξ), calculated using the Weighted Histogram Analysis Method (WHAM).27 The free energy (or the potential of mean force, PMF) is then given by the following equation

Bij r6

W (ξ) = −RT ln(p(ξ))

(1)

where Vij is the nonbonded interaction energy between atoms i and j. The A and B values are calculated from the van der Waals parameters for each atom interaction using the standard Lorentz−Berthelot mixing rules. For atoms i and j, the Aij value is calculated as (AiAj), with each A value given in Supporting Information Table 1. The same procedure is followed for the Bij values. The atomic van der Waals parameters for calcite were obtained from the study by Rahaman et al.25 of water adsorption on the calcite surface. A summary of the charges and parameters of the uranyl system used in the simulations are given inTable S1 of the Supporting Information. 2.1. MD Simulation Protocol. The protocol for MD equilibration is as follows: first the system is equilibrated for 2000 steps using the “zero” directive in the DLPOLY package, which is equivalent to performing a MD minimization at a temperature of 10 K. Then, the system was simulated for 50 ps at the NPT ensemble (constant number of particles N, constant pressure P of 1 bar, and constant temperature T of 298 K) to equilibrate the solution phase. The nonbonded cutoff distance used to calculate the potential energy was 10 Å. After this equilibration stage, data collection was performed at constant volume (NVT ensemble), which can be considered to be more appropriate for slab simulations, for 2, 5, and 7 ns for the {101̅4}, {314̅8} and {31̅2̅16} surfaces, respectively. The reason for the differences in production times is that the potential energy for the different systems takes different amounts of time to converge and stabilize, which is a requirement for a well equilibrated system. When the potential energy converges, the uranyl carbonate complex has found a stable configuration on the mineral surface. 2.2. Free Energy Adsorption Profiles of Uranyl on Calcite. To estimate the adsorption path and affinity of the U(VI) complexes and their coordinating carbonate and calcium ions to the calcite surface, free energy simulations to bring a complex from bulk solution to adsorb on the surface were performed. At all stages of the adsorption path, the system is in thermodynamic equilibrium and the free energy minima correspond to stable intermediates or sorbed species. To do this, we have employed a well established method called umbrella sampling.26 In this approach, a reaction coordinate to describe the adsorption event is chosen. For surface adsorption processes, it is appropriate to choose the z-distance of the uranium atom to the surface (or to an atom at the surface). A series of simulations are then performed at different reaction coordinate distances. This distance is harmonically restrained to a particular value to allow the system to sample all possible configurations in phase space. The restraining potential employed has the harmonic form:

V (ξ ) =

1 k(ξ − ξ0)2 2

(3)

where p(ξ) is the unbiased probability of the uranium atom being at a particular reaction coordinate ξ. One important advantage of umbrella sampling is that a path of the U(VI) complex adsorbing on the surface is generated by considering all the simulations together. Hence, the mechanism for adsorption and intermediate structures can be investigated along the chosen reaction coordinate. Using this technique, two free energy profiles of [Ca2UO2(CO3)3] adsorbing on two different sites of the {101̅4} calcite surface were obtained. The two sites were chosen so that the uranyl complex was located either above a carbonate anion or above a calcium cation of the surface. The reaction coordinate distances were calculated as the z-distance between the uranium atom and the carbon atom or the calcium atom on the calcite surface, depending on the path being investigated. For each PMF calculation 24 umbrella sampling simulations were performed spanning a reaction coordinate distance that ranged from 3 to 8.75 Å, with an interval of 0.25 Å between adjacent windows. In each window the simulation was initiated using one of six initial structures at increasing distances (from 4 to 9 Å) from the surface with the calcium−uranyl− carbonate species initially oriented with the UO2 perpendicular to the surface plane, to be consistent with the predictions of Rihs et al.4 The UO2(CO3)3 complex was then allowed to attain a U-surface reaction coordinate distance corresponding to the restraint applied to the window during the equilibration phase of the simulation. Each window was simulated for 2 ns and the last 1000 ps were chosen to produce the PMFs using the WHAM method.27 These choices were made after preliminary assessment of the data to ensure convergence of the PMFs, giving an effective equilibration time of at least 1050 ps per window. The simulation for each window allowed the complex to reorient itself, subject to the restraint potential in the z-direction, and to translate in the x,y-plane without restraint. Thus although the simulations were initially constructed by binding to the calcium or carbonate ions defining the two sites, as sampling progressed the differentiation between the two cases was unbiased; thus at long zdistances, both simulations essentially converge to describe [Ca2UO2(CO3)3] in solution, allowing the two site PMFs to be directly compared. It is important to note that the choice of a reaction coordinate is a restriction of the PMF approach and the z-coordinate is only an approximation to the reaction path of a true adsorption mechanism. Similar procedures were employed to simulate a binding path at the {314̅8} and {31̅2̅16} calcite surfaces above carbonate anions of the surface. In each case, the z-coordinate was measured with respect to the carbon of the carbonate at the vertex of the (acute or obtuse) step (angle) and defined in the direction shown in Figure 1. In this way, the adsorption affinity of the uranyl complex toward different calcite faces relative to aqueous solution could be determined, and a comparison of adsorption paths could be made between the different calcite faces.

(2)

where ξ0 is the target reaction coordinate distance (in Å) and k is the force constant (in kcal mol−1 Å−2). The restraint force constant of the umbrella potential is set to 100 kcal mol−1 Å−2. 7589

dx.doi.org/10.1021/es300034k | Environ. Sci. Technol. 2012, 46, 7587−7594

Environmental Science & Technology

Article

Figure 2. Adsorbed [Ca2UO2(CO3)3] complex on different calcite surfaces: (a) on the flat {101̅4} surface; (b) at an acute step between {101̅4} and {3148̅ } faces; and (c) at an obtuse step between the {1014̅ } and {312̅ 1̅ 6} faces. The calcite surface planes are shown schematically by the ribbon with selected carbonate and calcium ions. The dotted lines represent interactions within the uranyl complex and the dashed lines indicate interaction of the uranyl complex with the surface.

3. RESULTS AND DISCUSSION 3.1. Adsorption of [Ca2UO2(CO3)3] on Different Calcite Faces. Classical molecular dynamics simulations of

Figure 3. Potential of mean force for [Ca2UO2(CO3)3] adsorbing on separate sites of the {101̅4} calcite surface. Solid line is adsorption on a surface calcium cation, and dashed line is adsorption on a surface carbonate anion.

Figure 4. Complex [Ca2UO2(CO3)3] separated by a vicinal water layer on the calcite surface as observed in the adsorption pathways of [Ca2UO2(CO3)3] on calcite. The surface plane is shown schematically by the brown ribbon. The dotted lines indicate hydrogen bond interactions and the ribbons show the mineral surface.

[Ca2UO2(CO3)3] at the interface between a particular calcite surface and water predict both inner- and outer-sphere adsorption complexes for the three faces of calcite considered in this study. A sorbed species is considered inner-sphere when a new and direct interaction between the uranium center and the calcite carbonate anions is formed. An outer-sphere complex forms interactions with the surface through the carbonate ligands already present in the uranyl(VI) complex. Figure 2 shows how the calcium-uranyl-carbonate species is adsorbed on the {101̅4}, {314̅8}, and {31̅2̅16} calcite planes.

Selected average interatomic distances are shown in Figure 2 to illustrate typical configurations of the bound structures; since the computed structures are dynamic, the distances were derived from radial distribution functions (RDFs) taken from the simulations in which they were predominantly observed and, particularly for the uranyl triscarbonate moiety, will reflect the reference values of the empirical force-field parameters 7590

dx.doi.org/10.1021/es300034k | Environ. Sci. Technol. 2012, 46, 7587−7594

Environmental Science & Technology

Article

These simulation results are consistent with experimental observations and interpretations of adsorbed species on calcite. Elzinga et al.2 identified the uranyl species sorbed on calcite as [Ca2UO2(CO3)3] using EXAFS and luminescence data, but could not distinguish between inner- and outer-sphere complexes, although proposed a weakly split U−Oeq shell with oxygens at 2.26 and 2.46 Å. Observations at increased loading further indicated that multiple uranyl species were likely. Our simulations concur with these findings and show that several inner-sphere species can form, as seen at the {101̅4} and {314̅8} faces of calcite in Figure 2a and b. In addition outer-sphere complexes, which may be difficult to distinguish from the inner-sphere complexes, can also be formed as shown in Figure 2c. The study by Rihs et al.4 similarly concluded that U(VI) can adsorb in an inner-sphere mode to calcite, with preferential binding at acute steps, and the possibility of U(VI) binding to surface oxygen atoms in a monodentate manner was proposed based upon a U−Oeq splitting of 2.34−2.43 Å, similar to Elzinger et al. The equatorial splitting in our MD results (∼2.37−2.45 Å) is not as large as the experimental observations due to the force-field approach, but it is quite clear that there is in-equivalent U−Oeq binding with both mono- and bis-coordinated carbonate present. Furthermore, the monodentate binding proposed at the {314̅8} acute step of calcite by Rihs et al. is consistent with the observations from our simulations, as illustrated in Figure 2b. It is also clear that the complexes described in Figure 2 could be categorized as inner-sphere since the U-surface distance is ∼2.5 Å in agreement with the XSW prediction of Rihs et al. of 2.55 Å. 3.2. Adsorption Free Energy Profiles on Different Sites of the {101̅4} Surface. The free energy profiles for the adsorption of the complex [Ca2UO2(CO3)3] on top of carbonate or calcium ions of the calcite {101̅4} flat surface are shown in Figure 3. By comparing the relative energies between the solution uranyl complex and the adsorbed species, it can be observed that this neutral uranyl species has a low affinity toward calcite, and prefers the solution environment

Figure 5. Potentials of mean force for [Ca2UO2(CO3)3] adsorbing to carbonate anions of the {1014̅ }, {3148̅ }, and {312̅ 1̅ 6} surfaces of calcite.

utilized. RDFs corresponding to Figure 2 are available in the Supporting Information (Figures S1 to S3). The simulations show that an inner-sphere complex can be formed on both the {101̅4} and {314̅8} faces, whereas an outer-sphere complex is observed at the obtuse steps of the {31̅2̅16} face. The outer-sphere complex binds to the surface through interactions of carbonate oxygen atoms of the complex [Ca2UO2(CO3)3] with calcium cations on the mineral surface. For inner-sphere adsorption, the uranium atom coordinates directly with a carbonate ligand on the surface in a monodentate fashion, resulting in one of the original carbonates of the complex [Ca2UO2(CO3)3] changing its coordination from bidentate to monodentate, so that the overall 6-fold coordination of uranyl is preserved. There are some configurations in the simulations that predicted this inner-sphere species also to adopt 5-fold coordination at the {101̅4} and {314̅8} faces, where one of the carbonate oxygen atoms in the [Ca2UO2(CO3)3] complex that originally coordinated to uranium shifts toward the calcite surface to form a new interaction with a calcium cation on calcite, as shown in Figure 2a.

Figure 6. Schematic showing the preferential binding of the uranyl(VI) cation to the sides of etch pits in preference to the flat surface at the bottom. 7591

dx.doi.org/10.1021/es300034k | Environ. Sci. Technol. 2012, 46, 7587−7594

Environmental Science & Technology

Article

When the [Ca2UO2(CO3)3] species approaches the calcite surface near a calcium cation, a similar, but somewhat different picture arises. At a reaction coordinate distance of about 5 Å away from the surface calcium atom there is an energy minimum which, after visualization, was found to be similar to the one observed earlier when the same species approaches the surface on top of a carbonate anion (ie. the uranyl species is separated by a single water layer interacting with the mineral surface, as seen in Figure 4). As the uranyl species approaches the surface, a small free energy barrier of about 8 kJ mol−1 is encountered. This barrier is much lower than the previous case (cf., 50 kJ mol−1) because, this time, an outer-sphere complex is formed and no uranyl carbonate interactions are broken. Therefore, this free energy barrier only takes into account the energy needed to displace surface water molecules. The outersphere complex that forms in this adsorption pathway is similar to the one shown in Figure 2c, where the carbonate oxygen atoms of the [Ca2UO2(CO3)3] complex form new interactions with calcium cations of the surface. Overall, it can be concluded that the formation of an outer-sphere adsorbed uranyl species is clearly energetically more favorable than forming an innersphere complex on the calcite surface. Although the data suggest that the intermediate structure may be slightly thermodynamically more stable than the outer-sphere complex we stress that the free energy differences taken from the PMFs are computed directly from probability distributions and are subject to some statistical variation and such small energies are certainly at the limit of the empirical force-field which approximates the electronic effects and neglects polarization. As already noted, it can be hard to distinguish between innerand outer-sphere species in experiments, but these data support the conclusion that uranyl carbonate adsorbs mainly as an outer-sphere complex on calcite.19 This behavior can also be attributed to the fact that [Ca2UO2(CO3)3] in solution is very stable and hence only interacts with the surface when it is possible to do so without distorting its stable structure, and this can be achieved when an outer-sphere complex forms. 3.3. Adsorption Free Energy Paths on the {314̅8} and {31̅2̅16} Faces of Calcite. The free energy profiles for the adsorption of [Ca2UO2(CO3)3] on carbonate anions of both the {314̅8} and {31̅2̅16} faces of calcite are shown in Figure 5, along with the profile on the {101̅4} surface as described previously. All these PMFs display similar behavior, where a minimum in the free energy exists at a reaction coordinate distance of ∼5−6 Ǻ away from the surface. In this free energy minimum, the uranyl−carbonate species interact with the mineral surface via a water layer. Then, as the complex approaches the surface more closely, a free energy barrier has to be overcome to displace the water molecules and form new interactions to the carbonate anions of the surface which leads to the formation of an inner-sphere adsorbed species. The free energy of the adsorbed complex compared to the solvated species is energetically more stable at the {3148̅ } and {312̅ 1̅ 6} steps than on a flat {101̅4} terrace. Therefore, the uranylcarbonate complex would preferably adsorb at these steps when present. The lowest free energy barrier for adsorption is seen at the {314̅8} face (∼23 vs ∼ 44 kJ mol−1 at the {31̅2̅16} steps and ∼50 kJ mol−1 at the {1014̅ } surface), which means that the uranyl species will show a kinetic preference to adsorb at the acute step. The preference for forming surface complexes of uranyl along these acute steps on the calcite surface is in agreement with experimental XSW studies.4

rather than adsorbing onto the surface, with a free energy difference of less than 20 kJ mol−1. Uranyl has been observed experimentally to have weak adsorption on calcite1 and the uranyl-triscarbonate complex has also been found to display weak interactions with the calcite surface.2 For the free energy profile with the uranyl complex approaching the surface toward a carbonate anion, a free energy minimum is present at a reaction coordinate distance of ∼6.4 Å away from the surface, defined here as an intermediate complex. This energy minimum corresponds to the uranyl complex separated from the calcite surface by a vicinal water layer that is in direct contact with the mineral surface. A representative configuration from this state is shown in Figure 4, where water molecules that form hydrogen bonds with the calcite surface also interact with the uranyl complex through other hydrogen bonds with the carbonate anions. As the uranyl complex approaches the surface even closer, there is a high free energy barrier of about 50 kJ mol−1 before it forms an inner-sphere adsorbed species on the mineral surface. This adsorbed complex is similar to the configurations shown in Figure 2a. The high free energy barrier exists because of two factors: one is the need to displace the already-adsorbed water molecules on the calcite surface before they are replaced with the uranyl−carbonate species, and the other, perhaps most significant in this case, is the breaking of one of the uranyl carbonate oxygen bonds and its replacement with a bond to a surface carbonate oxygen atom. Overall, adsorption to form the inner-sphere complex on the {101̅4} surface is unfavorable by ∼16 kJ mol−1 because of the radical change in the environment of the uranyl species; the decrease in solvation of both surface and complex in this case has clearly not been offset by any gain in new or more favorable interactions upon sorption. To shed further light on the thermodynamic contributions in the adsorption mechanism an approximate enthalpy of adsorption can be estimated from the difference in the average potential energies taken from the simulations corresponding to bound and unbound states (although here we compute an internal energy change, ΔU, since we have used a canonical ensemble). Approximate entropies of adsorption were also estimated from the difference between the internal energy and (Helmholtz) free energy change (ΔA) predicted from the PMF according to, ΔA = ΔU − T ΔS

(4)

The internal energy difference between the intermediate species where the uranyl complex is separated from the surface by a water layer (at a reaction coordinate distance of ∼6.4 Å) and the solvated uranyl complex in aqueous solution was thus estimated to be ∼39 kJ mol−1. The corresponding free energy difference from Figure 3 is ∼−3 kJ mol−1. This results in a positive value of ∼42 kJ mol−1 for TΔS indicating that the intermediate state is entropically favored over the fully solvated uranyl species. Such a favorable change can be rationalized as the more ordered water molecules, above the calcite surface and around the uranyl complex, are displaced into bulk solution. This behavior is also observed when the uranyl-carbonate complex moves even closer to the surface to form an innersphere species releasing waters from the vicinal water layer bound to the surface. In this case the enthalpy difference between the inner-sphere adsorbed complex and the intermediate species adsorbed through the vicinal water layer was estimated to be ∼88 kJ mol−1; the free energy difference is ∼21 kJ mol−1 leading to a value of 67 kJ mol−1 for TΔS. 7592

dx.doi.org/10.1021/es300034k | Environ. Sci. Technol. 2012, 46, 7587−7594

Environmental Science & Technology

Article

(5) Doudou, S.; Arumugam, K.; Vaughan, D. J.; Livens, F. R.; Burton, N. A. Investigation of ligand exchange reactions in aqueous uranyl carbonate complexes using computational approaches. Phys. Chem. Chem. Phys. 2011, 13 (23), 11402−11411. (6) Druchok, M.; Bryk, T.; Holovko, M. A molecular dynamics study of uranyl hydration. J. Mol. Liq. 2005, 120 (1−3), 11−14. (7) Guilbaud, P.; Wipff, G. Force field representation of the UO22+ cation from free energy MD simulations in water. Tests on its 18crown-6 and NO3− adducts, and on its calix[6]arene6‑ and CMPO complexes. J. Mol. Struct.: THEOCHEM 1996, 366 (1−2), 55−63. (8) Kerisit, S.; Liu, C. Molecular simulation of the diffusion of uranyl carbonate species in aqueous solution. Geochim. Cosmochim. Acta 2010, 74 (17), 4937−4952. (9) Greathouse, J. A.; O’Brien, R. J.; Bemis, G.; Pabalan, R. T. Molecular dynamics study of aqueous uranyl interactions with quartz (010). J. Phys. Chem. B 2002, 106 (7), 1646−1655. (10) Greathouse, J. A.; Cygan, R. T. Molecular dynamics simulation of uranyl(VI) adsorption equilibria onto an external montmorillonite surface. Phys. Chem. Chem. Phys. 2005, 7 (20), 3580−3586. (11) Bernhard, G.; Geipel, G.; Brendler, V.; Nitsche, H. Speciation of uranium in seepage waters of a mine tailing pile studied by timeresolved laser-induced fluorescence spectroscopy (TRLFS). Radiochim. Acta. 1996, 74, 87−91. (12) Bernhard, G.; Geipel, G.; Reich, T.; Brendler, V.; Amayri, S.; Nitsche, H. Uranyl(VI) carbonate complex formation: validation of the Ca2UO2(CO3)3(aq.) species. Radiochim. Acta. 2001, 89 (8), 511−518. (13) de Leeuw, N. H.; Parker, S. C.; Harding, J. H. Molecular dynamics simulation of crystal dissolution from calcite steps. Phys. Rev. B 1999, 60 (19), 13792−13799. (14) Kristensen, R.; Stipp, S. L. S.; Refson, K. Modeling steps and kinks on the surface of calcite. J. Chem. Phys. 2004, 121 (17), 8511− 8523. (15) Stipp, S. L. S. Toward a conceptual model of the calcite surface: Hydration, hydrolysis, and surface potential. Geochim. Cosmochim. Acta 1999, 63 (19–20), 3121−3131. (16) Kelly, S. D.; Newville, M. G.; Cheng, L.; Kemner, K. M.; Sutton, S. R.; Fenter, P.; Sturchio, N. C.; Spötl, C. Uranyl incorporation in natural calcite. Environ. Sci. Technol. 2003, 37 (7), 1284−1287. (17) Reeder, R. J.; Nugent, M.; Tait, C. D.; Morris, D. E.; Heald, S. M.; Beck, K. M.; Hess, W. P.; Lanzirotti, A. Coprecipitation of uranium(VI) with calcite: XAFS, micro-XAS, and luminescence characterization. Geochim. Cosmochim. Acta 2001, 65 (20), 3491− 3503. (18) Savenko, A. V. Sorption of UO22+ on calcium carbonate. Radiochemistry 2001, 43 (2), 193−196. (19) Heberling, F.; Scheinost, A. C.; Bosbach, D. Formation of a ternary neptunyl(V) biscarbonato inner-sphere sorption complex inhibits calcite growth rate. J. Contam. Hydrol. 2011, 124 (1−4), 50−56. (20) MS Materials Visualizer, release 4.0; Accelrys Software, Inc: San Diego, 2005. (21) Markgraf, S. A.; Reeder, R. J. High-temperature structure refinements of calcite and magnesite. Am. Mineral. 1985, 70 (5−6), 590−600. (22) Pavese, A.; Catti, M.; Parker, S. C.; Wall, A. Modelling of the thermal dependence of structural and elastic properties of calcite, CaCO3. Phys. Chem. Miner. 1996, 23 (2), 89−93. (23) Teleman, O.; Jönsson, B.; Engström, S. A molecular dynamics simulation of a water model with intramolecular degrees of freedom. Mol. Phys. 1987, 60 (1), 193−203. (24) Kerisit, S.; Parker, S. C. Free energy of adsorption of water and metal ions on the {101̅4} calcite surface. J. Am. Chem. Soc. 2004, 126 (32), 10152−10161. (25) Rahaman, A.; Grassian, V. H.; Margulis, C. J. Dynamics of water adsorption onto a calcite surface as a function of relative humidity. J. Phys. Chem. C 2008, 112 (6), 2109−2115. (26) Torrie, G. M.; Valleau, J. P. Monte Carlo free energy estimates using non-Boltzmann sampling: Application to the sub-critical Lennard-Jones fluid. Chem. Phys. Lett. 1974, 28 (4), 578−581.

Using AFM imaging techniques, uranyl carbonate deposits were observed to preferentially build up on the edges of etch pits in the calcite cleavage face.28 Since the {314̅8} and {31̅2̅16} faces are likely to be present as the sides of etch pits, our prediction for favorable sorption at the acute and obtuse steps lead to a possible explanation as to why uranyl prefers to build up on these edge faces rather than on the flat terraces or the surface at the bottom of a pit; the adsorption of the neutral [Ca2UO2(CO3)3] complexes will likely promote further precipitation and remineralization. The schematic in Figure 6 illustrates this preference and emphasizes that the less favorable orientation of surface ions on the flat {101̅4} surface as a possible reason why inner-sphere complexes are disfavored on this face. In summary, we have shown that a simple description of the interaction of U(VI) with its environment using a molecular dynamics simulation approach gives new insights into the sorption behavior of the stable [Ca2UO2(CO3)3] species on different faces of the calcite surface. Our quantitative free energy analysis predicts similar structures consistent with the experimentally observed species on the calcite surface, and a general preference for outer-sphere complex formation by this species over inner-sphere complexation at the calcite surface. Comparison of the full sorption mechanisms indicates that the calcium-uranyl-carbonate complex may also preferentially adsorb to an acute-stepped surface {314̅8} step, in accordance with experimental observations of this species; however thermodynamically [Ca2UO2(CO3)3] prefers to similarly adsorb at both acute and obtuse steps of calcite rather than to a flat surface.

■

ASSOCIATED CONTENT

S Supporting Information *

Empirical force-field parameters and radial distribution functions for the adsorbed states. This information is available free of charge via the Internet at http://pubs.acs.org/.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS Work supported by the Natural Environment Research Council, U.K.. Computational resources from the UK National Supercomputing Service (HECToR) and The University of Manchester Computational Shared Facility.

■

REFERENCES

(1) Carroll, S. A.; Bruno, J. Mineral-solution interactions in the U(VI)−CO2−H2O system. Radiochim. Acta. 1991, 52−53 (3), 187− 193. (2) Elzinga, E. J.; Tait, C. D.; Reeder, R. J.; Rector, K. D.; Donohoe, R. J.; Morris, D. E. Spectroscopic investigation of U(VI) sorption at the calcite−water interface. Geochim. Cosmochim. Acta 2004, 68 (11), 2437−2448. (3) Geipel, G.; Reich, T.; Brendler, V.; Bernhard, G.; Nitsche, H. Laser and X-ray spectroscopic studies of uranium-calcite interface phenomena. J. Nucl. Mater. 1997, 248, 408−411. (4) Rihs, S.; Sturchio, N. C.; Orlandini, K.; Cheng, L.; Teng, H.; Fenter, P.; Bedzyk, M. J. Interaction of uranyl with calcite in the presence of EDTA. Environ. Sci. Technol. 2004, 38 (19), 5078−5086. 7593

dx.doi.org/10.1021/es300034k | Environ. Sci. Technol. 2012, 46, 7587−7594

Environmental Science & Technology

Article

(27) Kumar, S.; Rosenberg, J. M.; Bouzida, D.; Swendsen, R. H.; Kollman, P. A. The weighted histogram analysis method for freeenergy calculations on biomolecules. I. The method. J. Comput. Chem. 1992, 13 (8), 1011−1021. (28) Butchins, L. The chemistry of uranyl and neptunyl in carbonate containing geochemical systems, PhD Thesis. University of Manchester, Manchester, UK, 2005.

7594

dx.doi.org/10.1021/es300034k | Environ. Sci. Technol. 2012, 46, 7587−7594