Structural Correspondence between Uranyl Chloride Complexes in

ARTICLE pubs.acs.org/JPCA

Structural Correspondence between Uranyl Chloride Complexes in Solution and Their Stability Constants L. Soderholm,* S. Skanthakumar, and Richard E. Wilson Chemical Sciences and Engineering Division, Argonne National Laboratory, Argonne, Illinois 60439, United States

bS Supporting Information ABSTRACT: Pair-distribution functions (PDF)s were obtained from highenergy X-ray scattering (HEXS) data on a series of uranyl solutions as a function of chloride ion concentration. Analyses reveal that chloride forms only innersphere complexes with the uranyl, replacing inner-sphere waters such that the total uranyl coordination number decreases from 4.7 waters at [Cl
] = 0 m to 4.4 (1.7 water and 2.7 Cl
) at [Cl
] = 6.8 m. Some of the second-coordination sphere waters reorient upon uranyl inner-sphere chloride complexation in order to hydrogen bond with the bound anion. Similar data obtained on a series of solutions maintained at constant ionic strength are used to confirm structural assignments through determining stability constants for the addition of chloride to uranyl and comparison with published values. The stability constants, β1 = 1.5(10) m
1, β2 = 0.8(4) m
2, and β3 = 0.4(1) m
3, obtained in a series of solutions with constant ionic strength of 5.3 m, are in reasonable agreement with previously published results determined by solvent extraction. The agreement of stability constants supports our peak assignments for the PDF and thus our structural model for uranyl chloride complexes in solution. Using coordination numbers and speciation determined here as a function of chloride ion concentration, the monochloro species is found to have four coordinating waters in the uranyl equatorial plane, the dichoro species is found to be an equilibrium of three and two coordinating waters, and the trichloro species has only a single water in the equatorial plane. These values correspond to total average coordination numbers of 5, 4.3, and 4 for the mono-, di-, and trichlorouranyl complexes. From the equilibrium value of the dichloro species, it can be further estimated that ΔG =
0.5 kcal/mol for the conversion of five to four coordinate species. Overall, the HEXS data support the assertion that uranyl chloride correlations do exist and the results are not simply the result of solvent
ion effects.

’ INTRODUCTION Quantifying a dissolved metal ion’s speciation is the first step toward gaining a predictive understanding of its chemistry under the wide range of solution conditions that span laboratory to geosphere. Taken together, oxidation state and coordination environment are the critical features ultimately determining stability, reactivity, and solubility. Accurate prediction of solution chemistry involves the capability to model the relative stability of a number of competing species with high precision. Critical to these calculations is correctly determining the number and identity of all ligands that contribute to the energetics of a metal ion complex, including those associated in both first and second coordination spheres. In part because of the absence of metrical data and in part because of technical modeling limitations, it is often assumed that interactions other than inner-sphere coordination do not contribute significantly to the overall energetics of interest and thus can be ignored or modeled with a generalized assumption. For example, a continuum solvent approach is often used to model lanthanide(III) ions in solution, despite the high charge on the metal ion.1 A recent demonstration that this simplification may lead to erroneous conclusions is found in the detailed examination of Er
Cl speciation in acidic aqueous solution.2 A series of r 2011 American Chemical Society

constant ion strength solutions with varying chloride ion concentration were used to demonstrate that Cl
forms both innerand outer-sphere complexes with Er3þ and that they both contribute equally to the free energy of the aqueous Er
Cl complex. The stability constants for ErCln (n = 1, 2) were determined from high-energy X-ray scattering (HEXS) data and confirm values obtained independently from phase-transfer separations experiments.3 This example highlights the need to correlate experimentally determined energetics, such as free energies and stability constants, with molecular level structures before they are used to compare directly with calculations meant to predict solution behaviors. The purpose of this study is to provide structural information on uranyl chloride complexes in solution for comparison with computational studies of relative energetics and stabilities. The solution chemistry of uranium is dominated by the highly soluble uranyl ion, the linear dioxo UO22þ moiety. Uranyl solution chemistry has been well studied because of its central role in nuclear energy.4 It is known to form strong complexes with Received: December 4, 2010 Revised: April 12, 2011 Published: April 28, 2011 4959

dx.doi.org/10.1021/jp111551t | J. Phys. Chem. A 2011, 115, 4959–4967

The Journal of Physical Chemistry A inorganic ligands including phosphate, carbonate, and fluoride, moderately strong complexes with sulfates, and only weak complexes with nitrate and the halogen ions.5 The structure of sulfate solution complexes have been well studied because of the geological legacy left from their use in acid processes for milling uranium ores in eastern Germany.6 Structures of uranyl chloride solution complexes have received much less attention, despite the brine conditions found in relevant groundwaters, in part because the small stability constants expected for these complexes, log10 β01 = 0.17(2) and log10 β02 =
1.1(4),5 indicate that they play only a minor role in geologic fate and transport of the uranyl ion. On the basis of the small published stability constants and the discrepancy between these values and EXAFS results, which found almost no UO22þ
Cl
correlations until very high anion concentration,7 we assumed that outer-sphere complexes were the primary species present. Instead, using HEXS data and the approach previously reported for Er
Cl complexation,2 we have found that Cl
binds to UO22þ solely as an inner-sphere ligand and that it significantly reorganizes the outer-sphere correlated solvent waters upon complexation.

’ EXPERIMENTS Sample Preparation. Two sets of uranyl chloride samples were prepared, (1) a set in which the concentration of chloride ion was varied between 2.5 and 6.8 m and (2) a set which were approximated to a constant ionic strength of 5.3 m by varying the HCl/HClO4 ratio, with the chloride ion concentration varying across the entire range from 0 to 4.8 m. Perchloric acid was chosen for this series of solutions because it is known to be a noncomplexing ion.8 Samples of the desired composition were prepared by the dissolution of UO3 in either concentrated HCl or HClO4 which were brought to their constant boiling azeotropes before use. Accounting for the water of reaction upon dissolution of the oxide, dilutions of this stock solution were prepared for measurement with a constant uranium concentration of 0.5 m and the desired anion concentrations. Background solutions for each sample were prepared omitting the uranyl addition. The mole fractions of all elements excluding the uranium were matched as closely as possible between sample and background, generally to significantly less than 1% error. This accuracy in matching sample and background molalities is a requisite for our approach to the analysis of HEXS data. X-ray Scattering. Data Collection and Reduction. High-energy X-ray scattering data were collected at undulator beamline 11-ID-B at the Advanced Photon Source, Argonne National Laboratory. The incident beam energy was fixed at 91 keV, which corresponds to a wavelength of 0.13702 Å, with the scattered intensity measured using a General Electric amorphous silicon flat panel X-ray detector (GE Healthcare) mounted in a static position (2θ = 0°) providing detection in momentum transfer space Q up to 32 Å
1 with this fixed geometry. Samples were enclosed in Kapton capillaries with epoxy plugs and further contained as required for actinide samples. HEXS data were obtained on solutions at room temperature. The X-ray data were treated as described previously.9 The data are equivalent to standard powder patterns, intensity versus scattering angle, except that the high incident X-ray energy enables data collection out to large momentum transfers (Q). For example, a powder pattern obtained with a copper tube as the X-ray source has a maximum Q of about 8 Å
1. This is important

ARTICLE

because the scattering data are Fourier transformed, to provide g(r), a pair-correlation or pair distribution function (PDF), as a function of distance r, the resolution of which is dependent on the Q range used in the Fourier transform (FT).10,11 The scattering function S(Q) is obtained from the raw data by correcting for background (with empty sample holder), polarization corrections, Compton scattering, fluorescence, and a variety of geometric corrections. These data were normalized to a cross section per formula unit. S(Q) contains all the correlations present in the sample; in this case these include solvent
solvent, solvent
solute, and solute
solute correlations. Because of our focus on the uranyl correlations, we subtract background solution scattering, which includes all solvent
solvent interactions. The solutions used in this background subtraction procedure need to reproduce the ion ratios to better than 1%. The resulting difference scattering, SΔ(Q) includes only correlations involving uranyl and are the data used for all Fourier transforms (gΔ(r)). Because we are interested in small changes in metal
cation coordination environments, we have chosen to reduce the data by a background subtraction procedure similar to that used in neutron scattering experiments, except that we cannot use isotopic substitution.11 Instead, we have matched the atomic mole fractions of the samples and backgrounds to better than 1%, omitting only the uranyl in the background samples. This approach has proven successful2,12
15 but only when the concentrations of the targeted and background solutions can be matched to an accuracy of 1% or less. The data shown in the figures are background subtracted, which means that peak positions in the gΔ(r) vs r plots represent distances between uranium and other correlated ions in solution. The peak intensities are related to the number of electrons involved in the correlation;9 in this case determined by the number and identity of the ion correlations with uranium. The coordination number requires independent information about the ligating species and its electron count. Because the X-rays probe all species in solution, the coordination number represents an average number of ligands per uranyl. Data Fitting. Individual peaks are modeled using Gaussians, which are subsequently integrated to determine the number of electrons contributing to the correlation. This approach compares favorably to a direct integration, usually within one to two electrons for well-separated peaks. Its advantage lies in the ability to integrate unresolved peaks such as those seen for the U
water and U
Cl correlations in the first coordination sphere. In addition to the Gaussian representation of the correlation peaks, a model for the scattering arising from the noncorrelated component of the solution is also required. This disorder is manifested in the data as a rise in background that begins at the second coordination sphere, which masks its details. The disorder must be accounted for when treating the ordered component of the second coordination sphere. We have chosen to model this disorder in the two series of solutions under study herein using the assumption that the disordering does not change as a function of changing solution conditions. The background is modeled assuming it has the form of an integrated Gaussian with an invariant position and width. These two parameters are optimized by minimizing the shift in the peaks used to fit the ordered solution components. Slight changes in either the position or width of the background Gaussian will introduce the need to shift the positions of the Gaussians representing the ordered component. The parameters thus optimized are 4.48 Å and 0.05 for the position and 2σ2 Gaussian terms for both series 4960

dx.doi.org/10.1021/jp111551t |J. Phys. Chem. A 2011, 115, 4959–4967

The Journal of Physical Chemistry A

Figure 1. The Fourier transformed (FT) X-ray scattering data from a background subtracted aqueous 0.5 m uranyl perchlorate solution. The peaks correspond to UdO (1.76 (2) Å), U—O equatorial (2.41(3) Å), U—H water (2.9(1) Å), and broad peaks centered at 4.4(1), 6.5(1), and 8.6(3) Å attributed to second and higher coordination spheres.

of solutions. The overall uncertainty generated by the assumptions used in this model were estimated and included in the reported uncertainties.

’ RESULTS AND DISCUSSION Chloride-Free Solution. HEXS data from a uranyl ion dissolved in an aqueous perchoric acid solution has been described previously.16
18 The background-subtracted, FT, scattering pattern from our current study is shown in Figure 1. Similar to previous reports, a peak is observed at 1.76(2), consistent with the linear dioxo coordination consistently observed for uranium(VI) in aqueous solution, and the peak at 2.41(3) Å is attributable to the oxygens coordinating in the equatorial plane of the uranyl ion.19 The U
O interaction at 2.41 Å has been previously seen in EXAFS studies, where it is attributed to about 5 waters in the equatorial coordination sphere perpendicular to the linear dioxo moiety.7,19
25 Consistent with our current analysis, an integration of the equatorial water peak associated with an aqueous uranyl ion previously reported from HEXS data was integrated to yield 47(2) electrons, corresponding to 4.7 waters.18 Although calculations tend to favor predominately 5-fold equatorial waters,26
28 this slightly reduced coordination has been reproduced with calculations that included second-coordination sphere waters explicitly.29 Increased resolution and statistics of the data presented in Figure 1 over our previously published HEXS data allows the separation of the O and H scattering arising from their independent correlations with U, with the peak at 2.9(1) Å assigned to the Hþ ions of the coordinating water molecules. Similar metal
H correlation peaks have been observed by HEXS in acidic, aqueous solutions of Er3þ,2 Th4þ,14 and Cm3þ.12 In addition to the peaks associated with the uranyl first coordination sphere, there are also peaks attributable to uranyl
solvent interactions at 4.4(1) Å, 6.5(3) Å, and 8.6(3) Å. The position and intensities of the latter two peaks appear insensitive to solution conditions but are of insufficient quality to analyze quantitatively. A correlation has been previously reported at 4.5 Å in scattering data from uranyl perchorate solutions, where it was attributed to a U
O correlation corresponding to a second

ARTICLE

Figure 2. FT of the background subtracted X-ray scattering data obtained from an aqueous 0.5 m uranyl solution as a function of chloride ion concentration. Visible changes are attributed to the coordination of chloride in the first coordination sphere.

hydration sphere of about 1018 to 14 waters.16 With the exception of our previous HEXS study of uranyl solutions,18 there have been no previous reports of longer-distance U
O solution correlations. Chloride Ion Dependence. FTs of the scattering patterns are presented in Figure 2 as a function of chloride ion concentrations. The corresponding raw data (total scattering structure functions), together with their background and solvent
solvent subtracted counterparts, SΔ(Q)
1, are included as Supporting Information. The first peak in the FT data at 1.76(2) Å, assigned to the linear dioxo coordination of U, is invariant for all the measured solutions. The second peak, at 2.41(3) Å, has a position invariant with chloride concentration, although the peak intensity is seen to diminish significantly, from 34(1) to 14(1) electrons as the Cl
ion concentration increases from 2.5 to 6.8 m. This peak is assigned to the uranyl coordination with equatorial waters. Partially resolved from the bound water correlation is a new peak at 2.72(3) Å assigned to U
Cl correlations; it increases in total intensity from 17(2) electrons in 2.5 m chloride to 45(2) electrons in 6.8 m solutions. Similar to the water peak, the chloride ion peak does not change position as a function of solution conditions but instead remains at 2.72(3) Å across the series. A similar result was found for the U
Cl distance using EXAFS spectroscopy on a series of uranyl chloride solutions.7,24 A principal component analysis on those data determined that only two independent spectra were required to fit all the data that fall into the phase space of our samples.21,24 Since there is no evidence of changing bond distances, and since both EXAFS and HEXS experiments average solution species, these results alone are not able to distinguish whether there are multiple species present, corresponding to UO2(H2O)xCly (y = 0, 1, 2, 3), or whether there are simply two limiting species, one of which would be y = 0 and the other y = 3 or perhaps 4. Published single crystal structures are not informative in this regard. There are few known single-crystal structures of a monomeric uranyl ion coordinated only with chloride and/or water.4,30
32 As a representative X-ray single-crystal structure, Cs2UO2Cl4 reveals a uranyl ion with four equatorial Cl ions bound at a distance of 2.671(1) Å.33 No monomeric mixed Cl/H2O structures have been reported. As illustrated in Figure 3 the total average equatorial coordination number, which is slightly less than 5 waters in a chlorine free 4961

dx.doi.org/10.1021/jp111551t |J. Phys. Chem. A 2011, 115, 4959–4967

The Journal of Physical Chemistry A

Figure 3. Ligation changes in the first coordination sphere as a function of total chloride ion concentration. These results rule out a simple substitution reaction of chloride ions for waters.

solution, is slightly increased in the 2.5 m Cl solutions, after which it decreases with increasing Cl concentration until it levels off at about 4.4, corresponding to 1.7 O and 2.7 Cl ligands per U. This result supports a previously published optical spectroscopy study in which the number of coordinating chloride ions was seen to saturate at concentrations above 6 M Cl
.34 This result rules out the simple chloride substitution mechanism modeled for this reaction,35 at least under the conditions of our experiments, and is inconsistent with one favored model chosen to fit EXAFS data on a similar series of solutions24 but consistent with another.7 The peak occurring at 4.4(1) Å for the chloride-free solution persists as the chloride ion concentration is increased from 2.5 to 6.8 m, although it changes shape and appears to move to higher r. The intensity in this region is fit with two Gaussians, one centered at the same r as the chloride-free solution, the other at 4.9(1) Å, with 2σ2 values of 0.083 and 0.05(1) Å2, respectively. This fitting approach was successful over the entire series of concentrations, revealing the number of electrons determined for the 4.4 Å peak decreases with increasing chloride ion concentration from 75(8) to 55(8) electrons. Extrapolating to a chloridefree solution based on this trend would yield 9(1) waters correlated in the second coordination sphere, comparable to the 10 waters that we previously reported for uranyl in a chloridefree, perchloric acid solution and is consistent with two water molecules for each inner-sphere water. Thus nine waters correspond well to the 4.5 to 5 waters found by both experiment and calculation for the uranyl ion in aqueous solution.17,18,29 The new peak at 4.9(1) Å grows in intensity until a chloride concentration of about 4 m after which it remains constant at about 14 electrons. The other notable change in the FT of the scattering pattern shown in Figure 4 is the new small peak at 3.4(1) Å that increases in intensity with increasing chloride ion, at concentrations above 4 m, to a maximum of about 7
8 electrons. If the electron counts of the three peaks, at 3.4 Å, 4.4 Å, and 4.9 Å are summed, their cumulative numbers do not change significantly with chloride ion concentration but instead remain relatively constant at about 83 electrons, suggesting that the changes in the scattering data are the result of a rearrangement of the waters in the second coordination sphere in response to ligation changes in the first coordination sphere and not directly from chloride ion coordination in the outer sphere. This suggestion is verified by the analysis of HEXS data obtained at

ARTICLE

Figure 4. Changes in the second coordination sphere as a function of chloride ion concentration. The new peak that appears at 3.4 Å is attributed to U
H correlations arising from the reorientation of second coordination sphere waters as the result of H bonding to the coordinating Cl
ions. The shoulder on the high-r side of the 4.4 Å peak is attributed to the O ions associated with the H-bonded waters.

constant-ion strength, as discussed below. Taken together, the data show that chloride binds to uranyl solely as an inner-sphere complex. There is no evidence in the data for additional scattering that could be interpreted as multinuclear uranyl species, as previously suggested by Raman data.36 The metrical parameters obtained from the chloride ion concentration-dependent series are presented in Table 1. Our interpretation of the second coordination sphere scattering from uranyl chloride solutions is markedly different to our Er3þ chloride experiments, where both inner and outer sphere coordination were observed and determined to have the same free energy of complexation.2 In the Er3þ experiments the secondcoordination-sphere scattering profile changes shape, requiring an additional peak to represent its chloride-ion dependence, and the overall number of electrons associated with the outer sphere correlation change significantly. In contrast, although the second coordination shell scattering profile changed for the uranyl ion as a function of chloride concentration, the total number of electrons did not change, suggesting a rearrangement of waters without the addition of chloride. This interpretation is supported by the stability constant analysis presented below. Stability Constant Determination. Further insights and confirmation of our coordination model for UO22þ
Cl
complexation can be gained from a comparison of stability constants obtained from HEXS data based on our peak assignments with published values.4,37 Toward this end, HEXS data were measured on a series of solutions with a constant ionic strength, in which the uranyl concentration was held fixed at 0.5 m and the ionic strength (IS) was maintained at 5.3 m by replacing chloride with perchlorate ion over the entire concentration range. Perchlorate was chosen as the alternate anion because it is expected to be only weakly correlating,8 as evidenced by the very similar scattering patterns observed between our initial experiments described above and those involving perchlorate. The chloride ion peak at 2.72(3) Å in each scattering pattern is fit with a Gaussian, as shown in Figure 5, which is then integrated to provide the average number of chloride ions correlated to uranyl as a function of free chloride ion concentration, 4962

dx.doi.org/10.1021/jp111551t |J. Phys. Chem. A 2011, 115, 4959–4967

The Journal of Physical Chemistry A

ARTICLE

Table 1. Fitting Results for 0.5 Uranyl Solutions as a Function of Chloride Ion Concentration 2nd coord.

a

U
O(eq)

first coord sphere

[Cl], m

distance (Å)a

2.5 3.0

U
Cl(eq)

1st coord. sphere

total UO22þ coordination

electrons r = 3.4 Å

electrons r = 4.4 Å

electrons r = 4.9 Å

sphere total

watersb

distance (Å)a

chlorideb

(1st coord. sphere)b

peakc

peakc,d

peakc,d

electronsd,e

2.40

4.2

2.72

1.0

5.2

1.6

75

2.40

3.7

2.72

1.3

5.0

0.82

72

12

85

3.5

2.40

3.5

2.72

1.5

5.0

1.4

70

15

86

4.0

2.41

2.9

2.72

1.8

4.7

1.6

66

16

84

4.5 5.0

2.41 2.41

2.5 2.2

2.72 2.72

2.0 2.1

4.4 4.3

6.0 6.3

65 64

14 14

85 84

5.5

2.42

2.0

2.72

2.3

4.3

6.6

61

14

82

6.0

2.41

1.7

2.72

2.6

4.3

7.2

59

14

80

6.5

2.41

1.8

2.72

2.6

4.3

7.9

58

14

80

5.0

82

Error 3σ = 0.02. b Error 3σ = 0.1. c Error 3σ = 0.5. d Relative error only. e Error 3σ = 2.

Figure 5. Fits of the first coordination sphere X-ray scattering as a function of chloride-ion concentratation. The data are fit with two Gaussians, centered at 2.41(2) (blue) and 2.72(3) (red) Å with 2σ2 = 0.013 and 0.020 Å2. Integration of the fitted Gaussians is interpreted as the number of electrons corresponding to the peak intensity for oxygen (waters) and chloride ions coordinating to uranyl, in units of electrons. The black curve represents the experimental data. The gray curve is the fit representing the sum of the two Gaussians.

n = (A
a)/B, in which n (nbar) corresponds to the average number of Cl
bound to UO22þ, A and B are the total Cl
and UO22þ concentrations, and a represents the free ligand concentration ([free Cl]). The experimental values of n as a function of chloride ion concentrations are listed in Table 2. The determination of the average number of chloride ions correlated with uranyl as a function of the anion concentration provides metrical information that can be used to estimate stability constants βN for the successive addition of Cl
to the uranyl
aqua complex in

solution15,38 according to UO2 ClN ð2
NÞ
þ Cl
T UO2 ClN þ 1 ð1
NÞ
ðN ¼ 0
3Þ KN ¼

½UO2 ClN þ 1  ½UO2 ClN ½Cl

βN ¼ 4963

N Y

Ki

i¼0

dx.doi.org/10.1021/jp111551t |J. Phys. Chem. A 2011, 115, 4959–4967

The Journal of Physical Chemistry A

ARTICLE

Table 2. Average Number of Chloride Ions Coordinating to Uranyl as a Function of Free Chloride Ion Concentration in Solution [Cl], m

free [Cl], m

nbar

0.

0.

0.

0.5

0.32

0.35

1.0

0.59

0.83

1.5

0.91

1.2

2.0 2.0

1.3 1.3

1.4 1.5

2.5

1.7

1.6

2.5

1.7

1.6

3.0

2.0

1.9

3.5

2.5

2.0

4.0

2.9

2.2

4.5

3.4

2.3

4.8

3.7

2.3

Figure 6. The average number of chloride ions coordinated to uranyl (nbar) as a function of the free chloride ion in solution (black circles). The solid line represents a fit to the data as described in the text, corresponding to β1 = 1.5(10) m
1, β2 = 0.8(4) m
2, and β3 = 0.40(2) m
3.

for which KN are the equilibrium constants for the successive (N = 0, 1, 2, 3) Cl
complexation. Stability constants are obtained from the HEXS results by fitting according to N

n¼

∑0 nβnan N

∑0 βnan

The data, plotted as n versus a, are presented in Figure 6 together with a line representing the best fit to the data. The best fit parameters are β1 = 1.5(10) m
1, β2 = 0.8(4) m
2, and β3 = 0.40(2) m
3, values relevant to an IS of 5.3 m. The large errors in the fit arise from a strong correlation between β1 and β3 (correlation coefficient 0.9) when fitting our data, a problem that would not improve significantly with the addition of more data within our accessible concentrations. The results obtained here compare very favorably to those previously reported values obtained by spectroscopic methods under similar solution conditions, β1 = 3.1 M
1, β2 = 2.4 M
2, and β3 = 0.85 M
3.37 Note the difference in concentration units, which accounts for some of the variation in the two sets of values. Although the published stability constants were previously used to estimate corresponding free energies by extrapolation to infinite dilution using specific ion interaction theory, our solution conditions are outside those for which this theory is applicable.39,40 Thus, our comparison is only with the original separations results. However, the thermodynamic stability constants, β0n of 1.48(7) and 0.12(8) for n = 1 and 2, respectively,39 are extrapolations to infinite dilution, using specific ion interaction theory (SIT) applied to stability constants determined from separations data.37 These values indicate the U
Cl complexes are very weak, raising questions about the validity of the complexes themselves or whether they are artifacts of the high-ionic strength conditions necessary to observe them.41 Our observation of correlation peaks in the FT of the HEXS scattering data confirms the presence in solution of these weak uranyl chloride complexes and the thermodynamic interpretation of the earlier separations results. The speciation diagram depicted in Figure 7, constructed using the beta values determined from HEXS data, is indistinguishable

Figure 7. The speciation diagram determined from our stability constants, plotted as a function of total chloride ion concentration, in molal units.

from one determined using the previously published parameters from the spectrophotometric study.37 The stability constants for the formation of uranyl chloride complexes are very weak, corresponding to free energies less than 3/2 kT, the thermal energy in solution at room temperature. Despite the small stabilization energies for the formation of the first and second chloride complexes, it is clear from Figures 6 and 7 that a third stability constant is required to account for the average number of chloride ions that bind to uranyl at high concentrations. The addition of a third chloride results in the formation of an anionic complex, UO2(H2O)xCl3
, further complicating the relative roles of electrostatics versus bonding and solvation in the formation of these complexes.29 The role of covalent bonding in uranyl-complex stabilization remains an area of study.42
48 Previous calorimetric measurements point to the large entropic contribution to the chloro-complex free energy, suggested to arise from the release of outer-sphere bound waters as the charge on the complex is reduced.37 Our HEXS results do not confirm this suggestion but instead indicate a change in second coordination sphere water orientation consistent with changes in hydrogen 4964

dx.doi.org/10.1021/jp111551t |J. Phys. Chem. A 2011, 115, 4959–4967

The Journal of Physical Chemistry A bonding. Theory and modeling of [UO2Clx]2
x complexes are necessary to understand the relative role of each of these contributions toward the overall stability of uranyl chloride anionic complexes in aqueous solution, with a concentration on the relative roles of UO22þ
Cl
bonding, charge transfer, and hydrogen bonding, both inner and outer sphere. Exemplary and informative for comparison with our experimental results is a published density functional theory (DFT) study of uranyl chloride complexes in aqueous solution.47 A comparison of the energies derived for increasing chloro ligation indicated that the most stable monochloro complex is five coordinate, in agreement with the data represented in Figures 3 and 7. Applying the same reasoning, our experimental results strongly support a four coordinate trichloro complex. The dichloro complex is much less straightforward, and we are unable to directly distinguish between five and four coordinate uranyl ions (i.e., three or two coordinating waters). However, using the speciation diagram in Figure 7 and the average number of coordinating ligands from the constant ionic strength fits, it is possible to fit an average coordination number of 4.3 for the dichloro species, suggesting an equilibrium between two and three coordinating waters. This situation is similar to that previously seen for the uranyl(VI) ion in acidic aqueous solution, where an equilibrium was proposed between four and five ligating waters.18 Following the same reasoning used therein, it is possible to estimate a free energy for the loss of a water ½UO2 Cl2 ðH2 OÞ3  T ½UO2 Cl2 ðH2 OÞ2  þ H2 O The equilbrium constant obtained from an average total coordination of 4.3 corresponds to a reaction free energy of
0.5(1) kcal/mol for the reaction as written. These results are in remarkable agreement with previously calculated values.47 Of addition interest from the same theoretical study is the convolution of the DFT-calculated bond distances with the energetics determining coordination number. The results suggest that the trend toward decreasing coordination number, which would favor shortening bonds, and decreasing charge on the complex, which would favor lengthening bonds, may offset each other and thus explain the invariance in bond distance with chloride ion concentration observed both here and in the EXAFS experiments.7,24

’ CONCLUSIONS The close agreement of our stability constants with accepted values for our experimental conditions4,5 confirms our assignment of ion correlations to peaks in the PDFs obtained from HEXS data. Whereas the assignments in the first coordination sphere are unambiguous, the same is not true for changes in electron distribution within the second coordination sphere, notably the peaks at 3.4 Å, 4.4 Å, and 4.9 Å. Our interpretation of the changing intensity distribution in the 3
5 Å range of the FT involves the redistribution of outer-sphere waters to accommodate changing ligation in the first coordination sphere. With no chloride present, there is one fairly broad water peak with no differentiation of the atomic contributions of O and H. As chloride is added, there is a shift of some scattering to longer correlation length and also the appearance and growth of a new small peak at 3.4 Å. We assign these changes to the redistribution of waters in the second coordination sphere, driven by hydrogen bonding of the second coordination sphere water with the innersphere bound chloride. The agreement of our β values with

ARTICLE

accepted values5 supports our interpretation of the HEXS results. These changes do not involve chlorine ligation in the second coordination sphere, which if present would significantly increase the stability constants obtained by the analysis of HEXS data. Thus, the peak at 3.4 Å arises from a U
H correlation associated with the second coordination sphere water. The O associated with that water is then pushed further from the uranyl and is responsible for the peak at 4.9 Å. The intensity ratio of the 3.4
4.9 Å peaks is considerably more than the 1:4 ratio expected for water at the higher chloride ion concentration, indicating that the hydrogen bonding may be more extensive than our simple model suggests. The reorganization of second coordination sphere without the loss of coordinating waters raises questions about the source of a large entropic contribution to the free energy that was determined by previous temperature-dependent studies on this system.37 The number of ligands directly coordinating with uranyl is seen to decrease as chloride is added to the first coordination sphere, such that the total average coordination in the uranyl equatorial plane is about 4.4 as the chloride solution concentration approaches about 6.8 m. This coordination includes an average of 1.7 bound waters, even at high chloride concentration. On the one hand, the overall decrease in coordination number, specifically coordinating waters, is consistent with the observed homoleptic coordination of uranyl with chloride, which is seen from a single-crystal structural refinement to be four coordinate,33 in contrast to the essentially five-coordinate aqua coordination.17 On the other hand, our excellent fit in the absence of a β4 is somewhat unexpected based on the structural literature. No monomeric mixed Cl/H2O uranyl structures have been reported to date; instead the published monomeric uranyl chloride structures all include the [UO2Cl4]2
moiety.4,30
32 Thus the available solid-state structures do not appear to represent the solution speciation, even at high chloride concentration. A similar observation was made for a dimeric Th hydroxide solution containing chloride ion.14 Although there was no evidence of a Th
Cl correlation in solution, the crystals obtained as the solution was evaporated contained Th
Cl hydroxo-bridged dimers. Aging the solutions without evaporation resulted in a poorly crystalline ThO2-based phase, again with no Th
Cl correlations. Surprisingly, neither the U
O(water) nor the U
Cl bond distances change with increasing chloride complexation but remain constant at 2.41(3) Å and 2.72(3) Å, respectively. Recent DFT calculations on several uranyl/chloride/water complexes indicate a tendency to increase uranyl
ligand distances when water is replaced by chloride and to decrease these same distances with decreasing coordination number.47 Since both factors are occurring simultaneously in this case, the two contributions may be offsetting. Using Brown’s bond valence model,49 assuming an invariant uranium bond valence and the coordination numbers determined from the HEXS results presented herein, the U
Cl bond distance is calculated to change approximately 5% over the measured chloride coordination, a change that should be observable in our data but is not. This result could indicate that there are only two species present in solution, a UO2[H2O]52þ moiety and a UO2[Cl]x2
x moiety, where x = 3 or 4, which would have very important implications in terms of U
Cl bonding effects but the chloride coordination observed at constant ionic strength rules out this hypothesis. All four species UO2
(Cl)x (x = 0
3) are necessary to represent the data shown in Figure 6. This conclusion is supported by an 4965

dx.doi.org/10.1021/jp111551t |J. Phys. Chem. A 2011, 115, 4959–4967

The Journal of Physical Chemistry A NMR study of uranyl chloride in water
acetone mixtures in which the cation hydration number decreases by one unit for each mole of Cl ion added to solution50 and Raman spectra, which show progressive complex formation up to the trichloro complex.36 However, unlike the UV visible spectra, which were inconsistent with any combination of five uranyl chloride species,36 our data show no evidence for a multinuclear uranyl species. The invariance of the U
water and U
Cl bond distances with changing coordination is of interest and would benefit significantly from more solid-state structural work on uranyl hydrous chloride and uranyl chloride inorganic complexes.

’ ASSOCIATED CONTENT

bS

Supporting Information. Plots of total X-ray scattering data from a series of chloride ion concentrations before and after background subtraction. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected].

’ ACKNOWLEDGMENT We acknowledge the Actinide Facility for technical support during synchrotron experiments. This work is supported by the U.S. DOE, OBES, Chemical Sciences, Geosciences, and Biosciences Division under Contract DE- AC02-06CH11357. The APS is supported by OBES, Materials Sciences, under the same contract number. ’ REFERENCES (1) Cosentino, U.; Villa, A.; Pitea, d.; Moro, G.; Barone, V. Extension of computational chemistry to the study of lanthanide(III) ions in aqueous solution: Implemenation and validation of a continuum solvent approach. J. Phys. Chem. B 2000, 104, 8001–8007. (2) Soderholm, L.; Skanthakumar, S.; Wilson, R. E. Structures and energetics of erbium chloride complexes in aqueous solution. J. Phys. Chem. A 2009, 113, 6391–6397. (3) Fernandez-Ramirez, E.; Jimenez-Reyes, M.; Solache-Rios, J. Effects of ionic strength and charge density on the stability of chloride complexes of trivalent lanthanides. J. Chem. Eng. Data 2008, 53, 1756–1761. (4) Grenthe, I.; Drozdzynski, J.; Fujino, T.; Buck, E. C.; AlbrechtSchmitt, T. E.; Wolf, S. F., Uranium. In The Chemistry of the Actinide and Transactinide Elements; Morss, L. R., Edelstein, N. E., Fuger, J., Katz, J. J., Eds.; Springer: Dordrecht, 2006; Vol. 1, pp 253
698. (5) Guillaumont, R.; Fanghanel, T.; Fuger, J.; Grenthe, I.; Neck, V.; Palmer, D. A.; Rand, M. H. Update on the Chemical Thermodynamics of Uranium, Neptunium, Plutonium, Americium, and Technetium; Elsevier: Amsterdam, 2003; Vol. 5, p 970. (6) Bernhard, G.; Geipel, G.; Brendler, V.; Nitsche, H. Uranium speciation in waters of different uranium mining areas. J. Alloys Compd. 1998, 271
273, 201–205. (7) Allen, P. G.; Bucher, J. J.; Shuh, D. K.; Edelstein, N. M.; Reich, T. Investigation of aquo and chloro complexes of UO22þ, NpO2þ, Np4þ, and Pu3þ by X-ray absorption fine structure spectroscopy. Inorg. Chem. 1997, 36 (21), 4676–4683. (8) S
emon, L.; Boehme, C.; Billard, I.; Hennig, C.; Lutzenkirchen, K.; Reich, T.; Rossberg, A.; Rossini, I.; Wipff, G. Do perchlorate and triflate anions bind to the uranyl cation in an acidic aqueous medium? A combined EXAFS and quantum mechanical investigation. ChemPhysChem 2001, 2 (10), 591–598.

ARTICLE

(9) Skanthakumar, S.; Soderholm, L., Studying actinide correlations in solution using high energy X-ray scattering. Mater. Res. Soc. Symp. Proc. 2006, 893 (Actinides 2005--Basic Science, Applications and Technology), 411
416. (10) Magini, M.; Licheri, G.; Paschina, G.; Piccaluga, G.; Pinna, G., X-ray diffraction of ions in aqueous solutions: hydration and complex formation; CRC Press, Inc.: Boca Raton, FL, 1988; p 267. (11) Egami, T.; Billinge, S. J. L. Underneath the Bragg Peaks: Structural Analysis of Complex Materials; Pergammon: Amsterdam, 2003. (12) Skanthakumar, S.; Antonio, M. R.; Wilson, R. E.; Soderholm, L. The Curium Aqua Ion. Inorg. Chem. 2007, 46, 3485–3491. (13) Wilson, R. E.; Skanthakumar, S.; Burns, P. C.; Soderholm, L. Structure of the homoleptic thorium(IV) aqua ion [Th(H2O)10]Br4. Angew. Chem., Int. Ed. 2007, 46, 8043–8045. (14) Wilson, R. E.; Skanthakumar, S.; Sigmon, G.; Burns, P. C.; Soderholm, L. Structure of dimeric hydrolysis products of thorium. Inorg. Chem. 2007, 46, 2368–2372. (15) Skanthakumar, S.; Antonio, M. R.; Soderholm, L. A comparison of neptunyl(V) and neptunyl(VI) solution coordination: The stability of cation-cation interactions. Inorg. Chem. 2008, 47, 4591–4595. (16) Aberg, M.; Ferri, D.; Glaser, J.; Grenthe, I. Structure of the hydrated dioxouranium(VI) ion in aqueous solution. An X-ray diffraction and 1H NMR study. Inorg. Chem. 1983, 22, 3986–3989. (17) Neuefeind, J.; Soderholm, L.; Skanthakumar, S. Experimental coordination environment of uranyl (VI) in aqueous solution. J. Phys. Chem. A 2004, 108, 2733–2739. (18) Soderholm, L.; Skanthakumar, S.; Neuefeind, J. Determination of actinide speciation in solution using high-energy X-ray scattering. Anal. Bioanal. Chem. 2005, 383 (1), 48–55. (19) Antonio, M. R.; Soderholm, L., X-ray absorption spectroscopy of the actinides. In Chemistry of the Actinide and Transactinide Elements, 3rd ed.; Morss, L. R., Fuger, J., Edelstein, N., Eds.; Springer: Dordrecht, 2006; pp 3086
3198. (20) Wahlgren, U.; Moll, H.; Grenthe, I.; Schimmelpfennig, B.; Maron, L.; Vallet, V.; Gropen, O. Structure of uranium(VI) in strong alkaline solutions. A combined theoretical and experimental investigation. J. Phys. Chem. A 1999, 103 (41), 8257–8264. (21) Wasserman, S. R.; Allen, P. G.; Shuh, D. K.; Bucher, J. J.; Edelstein, N. M. EXAFS and principal component analysis: a new shell game. J. Synchrotron Radiat. 1999, 6, 284–286. (22) Vallet, V.; Wahlgren, U.; Schimmelpfennig, B.; Moll, H.; Szabo, Z.; Grenthe, I. Solvent effects on uranium(VI) fluoride and hydroxide complexes studied by EXAFS and quantum chemistry. Inorg. Chem. 2001, 40 (14), 3516–3525. (23) Thompson, H. A.; Brown, G. E. J.; Parks, G. A. XAFS spectroscopic study of uranyl coordination in solids and aqueous solution. Am. Mineral. 1997, 82, 483–496. (24) Hennig, C.; Tutschku, J.; Rossberg, A.; Bernhard, G.; Scheinost, A. C. Comparative EXAFS investigation of uranium(VI) and -(IV) aquo chloro complexes in solution using a newly developed spectrochemical cell. Inorg. Chem. 2005, 44, 6655–6661. (25) Antonio, M. R.; Soderholm, L.; Williams, C. W.; Blaudeau, J.-P.; Bursten, B. E. Neptunium redox speciation. Radiochim. Acta 2001, 89, 17–25. (26) Tsushima, S. Hydration and water-exchange mechanism of the UO22þ ion revisited: The validity of the “nþ1” model. J. Phys. Chem. A 2007, 111, 3613–3617. (27) Nichols, P.; Bylaska, E. J.; Schenter, G. K.; de Jong, W. Equatorial and apical solvent shells of the UO22þ ion. J. Chem. Phys. 2008, 128, 124507.  (28) Wahlin, P.; Schimmelpfennig, B.; Wahlgren, U.; Grenthe, I.; Vallet, V. On the combined use of discrete solvent models and continuum descriptions of solvent effects in ligand exchange reactions: a case study of the uranyl(VI) aqua ion. Theor. Chem. Acc. 2009, 124, 377–384. (29) Gutowski, K. E.; Dixon, D. A. Predicting the energy of the water exchange reaction and free energy of solvation for the uranyl ion in aqueous solution. J. Phys. Chem. A 2006, 110, 8840–8856. (30) Brown, D. Compounds of uranium with chlorine, bromine and iodine. In Gmelin Handbook for Inorganic Chemistry; Springer: Berlin, 1979; Vol. Uranium Suppl., C9, pp 1
186. 4966

dx.doi.org/10.1021/jp111551t |J. Phys. Chem. A 2011, 115, 4959–4967

The Journal of Physical Chemistry A

ARTICLE

(31) Deifel, N. P.; Cahill, C. L. The uranyl tetrachloride anion as a tecton in the assembly of U(VI) hybrid materials. CrystEngComm 2009, 11, 2739–2744. (32) Deifel, N. P.; Cahill, C. L. Supramolecular chemistry with uranyl tetrahalide ([UO2X4]2- anions. C. R. Chim. 2010, 13, 747–754. (33) Watkin, D. J.; Denning, R. G.; Prout, K. Structure of dicaesium tetrachlorodioxouranium(VI). Acta Crystallogr., Sect. C: Cryst. Struct. Commun. 1991, 47, 2517–2519. (34) Sun, X.; Samples, C.; Daia, K.; Meyers, M.; Bumgarner, M. Behaviour of the uranyl (UO22þ) ion in different strongly acidic media: characterization of UO22þ in common acids by electronic absorption spectroscopy. J. Chem. Res. 2009, June, 351–355. (35) Rotzinger, F. P. Mechanism for the substitution of an aqua ligand of UO2(OH2)5 by chloride. J. Chem. Theroy Comput. 2008, 4, 1654–1658. (36) Nguyen-Trung, C.; Begun, G. M.; Palmer, D. A. Aqueous uranium complexes. 2. Raman spectroscopic study of the complex formation of the dioxouranium(VI) ion with a variety of inorganic and organic ligands. Inorg. Chem. 1992, 31, 5280–5287. (37) Awasthi, S. P.; Sundaresan, M. Spectrophotometric & calorimetric study of uranyl cation/chloride anion system in aqueous solution. Indian J. Chem. 1981, 20A, 378–381. (38) Rossotti, F. J. C.; Rossotti, H. The Determination of Stability Constants and other Equilibrium Constants in Solution; McGraw-Hill Book Company, Inc.: London, 1961; p 425. (39) Grenthe, I.; Fuger, J.; Konings, R. J. M.; Lemire, R. J.; Muller, A. B.; Nguyen-trung, C.; Wanner, H. Chemical Thermodynamics of Uranium; Elsevier: Amsterdam, 1992; Vol. 1. (40) Grenthe, I.; Plyasunov, V.; Spahiu, K., Estimation of Medium Effects on Thermodynamic Data. In Modelling in Aquatic Chemistry; Grenthe, I., Puigdomenech, I., Eds.; Nuclear Energy Agency OECD: Washington, DC, 1997; pp 325
426. (41) Marcus, Y.; Kertes, A. S. Ion Exchange and Solvent Extraction of Metal Complexes; John Wiley & Sons Ltd: London, 1969; p 1046. (42) Denning, R. G.; Green, J. C.; Hutchings, T. E.; Dallera, C.; Tagliaferri, A.; Giarda, K.; Brookes, N. B.; Braicovich, L. Covalency in the uranyl ion: A polarized x-ray spectroscopy study. J. Chem. Phys. 2002, 117, 8008–8020. (43) Pierloot, K.; van Besien, E. Electronic structure and spectrum of UO22þ and UO2Cl42-. J. Chem. Phys. 2005, 123, 204309. (44) Garcia-Hernandez, M.; Willnauer, C.; Kruger, S.; Moskaleva, L. V.; Rosch, N. Systematic DFT study of gas phase and solvated uranyl and neptunyl complexes [AnO2X4]n (An = U, Np; X = F, Cl, OH, n =
2; X = H2O, n = þ2). Inorg. Chem. 2006, 45, 1356–1366. (45) Pierloot, K.; van Besien, E.; van Lenthe, E.; Baerends, E. J. Electronic spectrum of UO22þ and [UO2Cl4]2- calculated with timedependent density functional theory. J. Chem. Phys. 2007, 126, 194311. (46) Denning, R. G. Electronic structure and bonding in actinyl ions and their analogs. J. Phys. Chem. A 2007, 111, 4125–4143. (47) B€uhl, M.; Sieffert, N.; Volodymyr, G.; Wipff, G. Density functional theory study of uranium(VI) aquo chloro complexes in aqueous solution. J. Phys. Chem. A 2008, 112, 2428–2436. (48) Paez-Hernandez, D.; Ramirez-Tagle, R.; Codorniu-Hernandez, E.; Montero-Cabrera, L. A.; Arratia-Perez, R. Quantum relativistic investigation about the coordination and bonding effects of different ligands on uranyl complexes. Polyhedron 2010, 29, 975–984. (49) Brown, I. D. The Chemical Bond in Inorganic Chemistry: The Bond Valence Model; Oxford University Press: Oxford, 2002; p 278. (50) Fratiello, A.; Kubo, V.; Lee, R. E.; Schuster, R. E. Direct proton magnetic resonance cation hydration study of uranyl perchlorate, nitrate, chloride, and bromide in water-acetone mixtures. J. Phys. Chem. 1970, 74, 3726–3730.

4967

dx.doi.org/10.1021/jp111551t |J. Phys. Chem. A 2011, 115, 4959–4967