Stoichiometries and thermodynamic stabilities for aqueous sulfate complexes of U(VI)

Article abstract of DOI:10.1021/ic701379q

The formation constants of UO₂SO₄ (aq), UO ₂(SO₄)₂^2-, and UO ₂(SO₄)₃^4- were measured in aqueous solutions from 10 to 75°C by time-resolved laser-induced fluorescence spectroscopy (TRLFS). A constant enthalpy of reaction approach was satisfactorily used to fit the thermodynamic parameters of stepwise complex formation reactions in a 0.1 M Na⁺ ionic medium: log₁₀ K₁(25°C) = 2.45 ± 0.05, Δ_rH₁ = 29.1 ± 4.0 kJ·mol^-1, log₁₀ K ₂(25°C) = 1.03 ± 0.04, and Δ_rH ₂ = 16.6 ± 4.5 kJ·mol^-1. While the enthalpy of the UO₂(SO₄)₂^2- formation reaction is in good agreement with calorimetric data, that for UO ₂SO₄ (aq) is higher than other values by a few kilojoules per mole. Incomplete knowledge of the speciation may have led to an underestimation of Δ_rH₁ in previous calorimetric studies. In fact, one of the published calorimetric determinations of Δ_rH₁ is here supported by the TRLFS results only when reinterpreted with a more correct equilibrium constant value, which shifts the fitted Δ_rH₁ value up by 9 kJ·mol ^-1. UO₂(SO₄)₃^4- was evidenced in a 3 M Na⁺ ionic medium: log₁₀ K ₃(25°C) = 0.76 ± 0.20 and Δ_rH₃ = 11 ± 8 kJ·mol^-1 were obtained. The fluorescence features of the sulfate complexes were observed to depend on the ionic conditions. Changes in the coordination mode (mono- and bidentate) of the sulfate ligands may explain these observations, in line with recent structural data.

Full text of DOI:10.1021/ic701379q

Inorg. Chem. 2008, 47, 2180-2189
Stoichiometries and Thermodynamic Stabilities for Aqueous Sulfate
Complexes of U(VI)
Thomas Vercouter,* Pierre Vitorge, Badia Amekraz, and Christophe Moulin
CEA Saclay DEN-DANS/DPC/SECR/LSRM, 91191 Gif-sur-YVette cedex, France
Received July 11, 2007
The formation constants of UO2SO4 (aq), UO2(SO4)22-, and UO2(SO4)34- were measured in aqueous solutions
from 10 to 75 °C by time-resolved laser-induced fluorescence spectroscopy (TRLFS). A constant enthalpy of reaction
approach was satisfactorily used to fit the thermodynamic parameters of stepwise complex formation reactions in
a 0.1 M Na+ ionic medium: log10 K1(25 °C) ) 2.45 ( 0.05, ∆rH1 ) 29.1 ( 4.0 kJ · mol-1, log10 K2(25 °C) ) 1.03
( 0.04, and ∆rH2 ) 16.6 ( 4.5 kJ· mol-1. While the enthalpy of the UO2(SO4)22- formation reaction is in good
agreement with calorimetric data, that for UO2SO4 (aq) is higher than other values by a few kilojoules per mole.
Incomplete knowledge of the speciation may have led to an underestimation of ∆rH1 in previous calorimetric studies.
In fact, one of the published calorimetric determinations of ∆rH1 is here supported by the TRLFS results only when
reinterpreted with a more correct equilibrium constant value, which shifts the fitted ∆rH1 value up by 9 kJ · mol-1.
UO2(SO4)34- was evidenced in a 3 M Na+ ionic medium: log10 K3(25 °C) ) 0.76 ( 0.20 and ∆rH3 ) 11 ( 8
kJ · mol-1 were obtained. The fluorescence features of the sulfate complexes were observed to depend on the
ionic conditions. Changes in the coordination mode (mono- and bidentate) of the sulfate ligands may explain these
observations, in line with recent structural data.
1. Introduction
been exploited by leaching with concentrated sulfuric acid.3,4
Consequently, geochemical modeling of the transport of
uranium through natural aquifers must account for sulfate
complexation.
Binary uranyl sulfate complexes may be dominant species
under acidic conditions, while ternary uranyl hydroxo-sulfate
complexes are stable at low to near-neutral pH conditions.5
Despite a large number of investigations on the sulfate
complexation of U(VI), the stoichiometries of the complexes
are still debated, and discrepancies are observed in their
formation data.6,7 Moreover, most of the data have been
obtained at ambient temperatures, while various temperatures
may be expected under environmental conditions; especially
higher temperatures could be reached in the vicinity of a
Chemical speciation of heavy and radioactive metal ions
in ground or surface waters is an important issue for the
modeling of their transport from polluted soils or radioactive
waste storage sites into the environment.1 A considerable
effort has been spent to better understand the interactions of
complexing agents present in soils with metal ions, and
particularly with uranium. Sulfate anions can be found at
relatively high concentration in natural waters and can
participate in the dissemination of uranium. For instance,
sulfate concentrations higher than 10 mM were evaluated in
the pore water of clay-rich rocks that may host a deep
geological repository of high-level radioactive waste;2 an
environmental concern is also the uranium contamination of
drinking water reservoirs near sites where uranium ore has
(4) Zeissler, K. O.; Nindel, K.; Hertwig, T. In Uranium mining and
hydrogeology IV; Merkel, B. J., Hasche-Berger, A., Eds.; Technische
Universität Bergakademie Freiberg: Freiberg, Germany, 2005; p 713.
(5) Grenthe, I.; Lagerman, B. Radiochim. Acta 1993, 61, 169–176.
(6) Meinrath, G.; Lis, S.; Piskula, Z.; Glatty, Z. J. Chem. Thermodyn.
2006, 38, 1274–1284.
* Author to whom correspondence should be addressed. Phone: +33-
1-69-08-26-59. Fax: +33-1-69-08-54-11. E-mail: thomas.vercouter@
cea.fr.
(1) Bruno, J. Trace Element Modelling. In Modelling in Aquatic Chemistry;
Grenthe, I., Puigdomenech, I., Eds.; OECD: Paris, 1997; p 593.
(2) Leroy, P.; Revil, A.; Altmann, S.; Tournassat, C. Geochim. Cosmochim.
Acta 2007, 71, 1087–1097.
(7) Guillaumont, R.; Fanghänel, 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 B.V.: Amsterdam, 2003.
(3) Carvalho, I. G.; Cidu, R.; Fanfani, L.; Pitsch, H.; Beaucaire, C.; Zuddas,
P. EnViron. Sci. Technol. 2005, 39, 8646–8652.
(8) Ahrland, S.; Kullberg, L. Acta Chem. Scand. 1971, 25, 3677–3691.
2180 Inorganic Chemistry, Vol. 47, No. 6, 2008
10.1021/ic701379q CCC: $40.75
2008 American Chemical Society
Published on Web 02/16/2008

Aqueous Sulfate Complexes of U(VI)
natural uranium was prepared by the dissolution of U3O8 in a hot
perchloric acid solution. The uranium concentration was measured
by inductively coupled plasma mass spectrometry. The uranium
concentration in the test solutions was obtained by adequate dilution
of this stock solution. NaClO4, H2O, and Na2SO4 were purchased
from Merck (R.P. Normapur) and used without further purification.
Perchloric acid and sodium hydroxide were used for [H+] adjust-
ments.
radioactive waste repository due to the radioactive decay
energy. In this work, we have studied the formation of binary
uranyl sulfate complexes as a function of the temperature.
The enthalpies of complex formation or dissociation
2-
reactions for UO2SO4 (aq) and UO2(SO4)2 have been
determined by calorimetry8–10 and from the temperature
coefficients of complexation constants.11–13 The critical
compilation of thermochemical data by the Nuclear Energy
Agency (NEA) revealed a fair agreement between the data
obtained by the former method, while larger discrepancies
were observed for the data obtained by the latter one.7,14
Indeed, calorimetric measurements provide precise enthalpy
changes of reaction when the speciation is well defined. On
the contrary, uncertain speciation may affect the ∆rH
determinations. Calorimetric results will be discussed here
in more detail because possible misinterpretations are
suspected due to either the choice of the ꢀ1° value9 or the
possible influence of polynuclear uranyl species.10
2.2. [H+] Measurements and Sulfate Speciation. [H+] was
measured using combined glass microelectrodes (Radiometer
Analytical, XC161). The original solution of the reference compart-
ment was replaced with either a 0.1 or 3 M NaClO4 aqueous
solution containing 0.01 M NaCl. Calibrations were performed with
solutions of known [H+] in 0.1 or 3 M ionic media, as detailed
elsewhere.19 [H+] was measured at the temperature of the laboratory
(23 ( 1 °C), at the beginning and at the end of each titration
experiment, to ensure that heating did not alter the solution
-
compositions by evaporation. The HSO4 dissociation constant,
Ka(T,I), was calculated for each temperature and each ionic strength
using Ka°(T) values from Dickson et al.,20 and using the formula
of specific ion interaction theory (SIT).7 The parameters in the
Debye–Hückel term are calculated at each temperature as detailed
elsewhere.16 The ion interaction coefficients are taken from the
literature.7 They are assumed to be temperature-independent in the
range 10–75 °C. This approximation has little impact on our
calculations because all titrations were carried out in a H+ range
where SO4 predominates over HSO4 (-log10 [H+] > 2.7). At
each temperature, the concentrations of H+, HSO4-, and SO42- in
the solutions were calculated from the mass conservation and
electroneutrality relationships, and Ka(T,I).
The third complex, UO2(SO4)34-, has usually been ne-
glected in the interpretations of experimental data because
its existence has been difficult to prove. Indeed, this complex
should form at high ionic strengths and high sulfate
concentrations, while most of the studies were carried out
in more dilute solutions. More recently, the formation of
UO2(SO4)34- in concentrated Na2SO4 solutions was suggested
by time-resolved laser-induced fluorescence spectroscopy
(TRLFS), and its formation constant ꢀ3 was estimated at
ambient temperature.15
2-
-
TRLFS has already been successfully used to determine
the influence of the temperature on the carbonate complex-
ation of Cm(III) in concentrated solutions and to derive
enthalpy and entropy changes of reaction from the temper-
ature coefficient of the equilibrium constant.16 TRLFS
combines very low detection limits and a high sensitivity
toward U(VI) complex formation in aqueous solutions,17,18
which enables an avoidance of the formation of polycationic
complexes. In the present study, sulfate complexation of
uranyl was investigated at variable temperatures between 10
and 75 °C at low and high ionic strengths by TRLFS. We
report new determinations of thermodynamic parameters and
2.3. Time-Resolved Laser-Induced Fluorescence Spectros-
copy. A 1.5 mL solution of U(VI) was placed in a quartz cell and
titrated by the addition of adequate volumes of a sulfate solution
with the same U(VI) concentration. The temperature of the solution
was equilibrated for at least 15 min after each addition and
maintained at (0.3 °C using water circulation in the cell holder.
The laser excitation source is a Nd:YAG Laser (Minilite II,
Continuum, U.S.A.) delivering an energy of 4 mJ at 355 nm. The
repetition rate was 10 Hz, and the pulse duration was about 5 ns.
The fluorescence from the solution sample was focused on the
entrance slit of a monochromator spectrograph (Acton 300i, Roper
Scientific, U.S.A.) using a combination of mirrors and lenses. It
was detected by an intensified CCD camera (Andor, U.K.) that was
triggered by the delayed output of the laser pulse. The fluorescence
spectra had a resolution better than 0.2 nm. The error on the
measured fluorescence intensity was estimated to be less than 2%
from the standard deviation of the fluorescence intensity of a
reference Eu(III) solution that was regularly measured.
2+
discuss the coordination of the UO2 ion in the sulfate
complexes.
2. Experimental Section
2.1. Materials. Millipore deionized water (Alpha-Q, 18.2 MΩ
cm) was used throughout the preparations. A stock solution of
(9) Bailey, A. R.; Larson, J. W. J. Phys. Chem. 1971, 75, 2368–2372.
(10) Ullman, W. J.; Schreiner, F. Radiochim. Acta 1986, 40, 179–183.
(11) Wallace, R. M. J. Phys. Chem. 1967, 71, 1271–1276.
(12) Day, R. A.; Powers, R. M. J. Am. Chem. Soc. 1954, 76, 3895–3897.
(13) Lietzke, M. H.; Stoughton, R. W. J. Phys. Chem. 1960, 64, 816–820.
(14) Grenthe, I.; Fuger, J.; Konings, R. J. M.; Lemire, R. J.; Muller, A. B.;
Nguyen-Trung, C.; Wanner, H. Chemical Thermodynamics of Ura-
nium; Wanner, H., Forest, I., Eds.; North-Holland: Amsterdam, 1992.
(15) Geipel, G.; Brachmann, A.; Brendler, V.; Bernhard, G.; Nitsche, H.
Radiochim. Acta 1996, 75, 199–204.
Fluorescence lifetimes were derived from the decay of the
intensity F(λ,D,W) measured at given wavelengths λ as a function
of the gate delay D and for a given gate width W. The decay curves
were fitted using eq 1, which results from the time integration of
the expression of the fluorescence signal when prefilter and postfilter
effects are neglected:18
(16) Vercouter, T.; Vitorge, P.; Amekraz, B.; Giffaut, E.; Hubert, S.;
Moulin, C. Inorg. Chem. 2005, 44, 5833–5843.
(17) Moulin, C.; Beaucaire, C.; Decambox, P.; Mauchien, P. Anal. Chim.
Acta 1990, 238, 291–296.
(19) Vercouter, T.; Vitorge, P.; Trigoulet, N.; Giffaut, E.; Moulin, C. New
J. Chem. 2005, 29, 544–553.
(20) Dickson, A. G.; Wesolowski, D. J.; Palmer, D. A.; Mesmer, R. E. J.
Phys. Chem. 1990, 94, 7978–7985.
(18) Moulin, C.; Decambox, P.; Moulin, V.; Decaillon, J. G. Anal. Chem.
1995, 67, 348–353.
Inorganic Chemistry, Vol. 47, No. 6, 2008 2181

Vercouter et al.
N
UO (SO )2-2i
[
[
]
]
D
τs
2
4 i
F(λ,D,W) ) k ×
[s] × fs0(λ) × τs × exp -
×
log10
) log10 mSO2- + log10 Ki (7)
∑
{
(
)
4-2i
4
(
)
s ) 0
UO (SO )
2
4 i-1
W
1 - exp -
(1)
The left-side member was determined from the ratios of
the amplitude factors of the corresponding fluorescence
contributions. The values of log10 Ki were obtained by a least-
squares analysis on the 1/σ2-weighted data points (where σ
is the evaluated error on the ratios), and accounting for the
effect of the ionic medium change, as explained in the next
section. The error on log10 Ki was calculated from the
standard deviation multiplied by the appropriate value of the
Student parameter for a 95% confidence interval.
The dependence of Ki (i ) 1-3) on temperature was
modeled by integrating the van’t Hoff isochore, which
involves the enthalpy ∆rHi and the heat capacity ∆rCp,i of
the reaction:
(
) }
]
[
τs
where k is an apparatus factor, N is the number of fluorescing
0
species, and τS and fS (λ) are the fluorescence lifetime and molar
fluorescence intensity at D ) 0 of the species S, respectively. The
parameters τS and fS (λ) were simultaneously fitted on decay curves
at three to five different emission wavelengths corresponding to
maximum intensity peaks.
Fluorescence spectra were recorded with a gate delay D of 0.5
µs after the laser pulse and a gate width W of 400 µs in order to
collect the largest part of the fluorescence of U(VI) species. The
gate width was voluntarily large in order to detect the possible
presence of hydroxide complexes of U(VI), which may have high
lifetime values, compared to UO22+ and the sulfate complexes.15,18
Then, eq 1 becomes
0
N
∆rHi(T ° )
[s] × F0(λ)
(2)
1
1
T°
F(λ,0.5,400) ) k ×
{
}
∑
s
log10 Ki(T) ) log10 Ki(T ° ) -
-
+
(
)
R ln(10) T
s ) 0
0
∆rCp,i
where FS (λ) is the molar spectral contribution of the species S at
these acquisition parameters. The amplitude factors k × [s] of each
spectral contribution are then proportional to the species concentra-
T°
T
- 1 + ln
(8)
°
( )
[
]
R ln(10) T
T
0
where R is the molar gas constant and T° ) 298.15 K. ∆rCp,i
contributes to a second-order term in eq 8 and can sometimes
be neglected.
tions. FS (λ) was represented with a sum of five Gaussian–Lorent-
zian functions in the range 465–565 nm. The contributions of each
species were determined by nonlinear curve fitting using the
OriginPro 7.5 software (OriginLab Corp., U.S.A.). In a first step,
only the spectra which exhibited isobestic points were simulta-
neously fitted to determine the two spectral components. In a second
step, the remaining spectra of the experimental set were fitted by
adding a spectral component. The errors on the amplitude factors
were estimated from the reproducibility of the fluorescence
measurements, and from the standard deviations in the fit of the
experimental spectra with two or three spectral components.
3.2. Ionic Medium Corrections. The constants Ki are
dependent on the activity coefficients of the reactants and
products, which were kept almost constant by using a
supporting electrolyte (NaClO4). In fact, the titration experi-
ments were performed so as to maintain constant [Na+] by
appropriate mixing of the two NaClO4 and Na2SO4 salts.
When [SO42-] was not negligible as compared to [ClO4 ],
-
ionic medium effects had to be explicitly taken into account.
Thus, the dependence of Ki on the change of the ionic
medium was calculated using the SIT formula:
3. Thermodynamic Description
3.1. Equilibrium Constants. Formations of sulfate com-
2+
plexes of UO2 are described with the following stepwise
log10 Ki - 8(i - 2)DD-H ) log10 Ki ° -
reactions:
{[ꢀ(i,j) - ꢀ(i - 1, j)]m } + ꢀ(Na+, SO2-)m
(9)
∑
j
4
Na+
UO22+ + SO24- h UO2SO4 (aq)
UO2SO4 (aq) + SO24- h UO2(SO4)22-
UO2(SO4)22- + SO24- h UO2(SO4)34-
(3)
(4)
(5)
j
where mj is the molality of the ion j, and ꢀ(k,j) is the SIT
coefficient for the interaction between UO2(SO4)k2–2k and the
ion j.
and the conditional equilibrium constants (in kg ·mol-1) with
i ) 1-3:
The Debye–Hückel term, DD-H, was calculated at each
temperature, as in section 2.2; the ion interaction coefficients
for U(VI) ions and SO42- with the medium ions Na+, ClO4 ,
-
UO (SO )2-2i
and SO42- were taken from the work by Geipel et al.,15 while
ꢀ(UO22+, SO42-) ) 0.12 kg · mol-1 was arbitrarily fixed
according to Grenthe and Lagerman.5 Their temperature
dependence was neglected in a first approximation, which
is relevant for solutions with 0.1 M ionic strength due to the
weak influence of the SIT terms. The SIT formula was also
used in the same manner to extrapolate the log10 Ki values
to I ) 0. The factors of molar-to-molal conversion were
calculated for each solution at 25 °C by a weighed average
of the factors for the single salts.7
[
]
2
4 i
Ki )
(6)
UO (SO )4-2i
m
]
SO24-
[
2
4 i-1
where m is the molality (mol · kg-1) and [ ] are molarities
(mol ·L-1). K1 is also noted ꢀ1. The choice of the concentra-
tion units in eq 6 is convenient for our purpose because the
fluorescence of the uranyl species is proportional to their
molarities (mol· L-1), but ionic medium corrections on Ki
should be made in the molality unit (mol ·kg-1). The validity
of the speciation model was also graphically checked with
the rearranged expression of eq 6:
2182 Inorganic Chemistry, Vol. 47, No. 6, 2008

Aqueous Sulfate Complexes of U(VI)
Table 1. Main Fluorescence Features of Complexes of U(VI) at 20 °C
main emission
lifetime
wavelengths (nm)
medium
0.01 M HClO4
Na2SO4-NaClO4, I ) 0.1 M
Na2SO4-NaClO4, I ) 0.1 M
1.5 M Na2SO4
(µs)
2+
UO2
471-488-510-534-560
477-493-515-538-565
481-496-518-542-569
477-494-516-539-565
482-498-520-544-n.d.
2.3 ( 0.2
n.d.
UO2SO4 (aq)
2-
UO2(SO4)2
n.d.
4-
UO2(SO4)3
18.1 ( 0.8a
30 ( 3
UO2OH+
NaClO4, I ) 0.1 M, pH 2.7
a
Estimation from measurements in solutions in which the complex dominates; the value might not correspond to a true fluorescence decay process (see
text).
4. Results
4.1. Speciation of U(VI) in a 0.1 M Na+ Ionic Medium.
The first series of experiments was carried out by titrating a
0.1 M NaClO4 solution with a 0.05 M Na2SO4 solution, both
with [U(VI)] ) 4.2 µM. Despite the acidic conditions, the
initial spectrum of the uncomplexed U(VI) was found to be
very similar to the one of UO2OH+, with a smaller contribu-
tion from the spectrum of UO22+ (more than 10 times lower).
The fluorescence yield of UO2OH+ is usually much higher
than that of other species, and UO2OH+ can be detected at
very low concentrations.21 Although the hydrolysis constant
is increased by approximately 1 order of magnitude from
25 to 70 °C,22 [UO2OH+] still remains negligible under our
conditions compared with [UO22+], even at 75 °C. These
two spectral components were determined by an adequate
choice of the acquisition parameters (Table 1). Their
measured lifetimes lie in the range of published values.23,24
The important fluorescence signal due to UO2OH+ was
Figure 1. Fluorescence spectra of 4.2 µM U(VI) in Na2SO4/NaClO4
solutions ([Na+] ) 0.1 M, -log10 [H+] ) 2.7) measured at 10 °C as a
function of [SO42-] from 0 (spectrum a) to 4.33 mM (spectrum b).
actually used as a measure of the concentration of the
uncomplexed U(VI), because the ratio [UO2OH+]/[UO2
]
2+
is constant at constant pH. Therefore, the measured signal
was corrected to account for the small variations of [H+]
along the titration. Indeed, at the beginning of the titrations,
-log10 [H+] equaled 2.7 for each temperature investigated
and slightly increased with the additions of significant
volumes of the titrating solution: the final values of -log10
[H+] were 2.7, 2.9, 3.0, 3.0, and 3.2 for 10, 20, 37, 50, and
75 °C, respectively.
2-
With the addition of sulfate to the solution, the fluores-
Figure 2. Fluorescence spectra assigned to UO2SO4 (aq) and UO2(SO4)2
at different temperatures in solutions with [Na+] ) 0.1 M from spectral
cence peaks of UO2OH+ tend to decrease while other peaks
decomposition.
2-
appear. For log10 mSO < -2.4, several isobestic points
4
are observed, which suggests the formation of UO2SO4 (aq)
(Figure 1). For higher sulfate concentrations, spectral changes
because the fluorescence lifetime decreases according to the
Arrhenius law. The positions of the peak maxima and the
widths at midheight appear to be unaltered in the temperature
range studied, except for the emissions at about 477 and 481
nm for UO2SO4 (aq) and UO2(SO4)22-, respectively; these
emissions actually originate from a different excited state
of U(VI), with all the other emissions at higher wave-
lengths,25 which explains why the temperature effect has no
reason to be the same for them. The wavelengths of the main
fluorescence bands are given in Table 1. The spectrum
assigned to UO2SO4 (aq) at 75 °C shows less favorable
emissions at 515 and 538 nm compared to that at 493 nm.
Conversely, the relative intensities in the spectrum assigned
indicate that another U(VI) species exists in solution, as
2-
expected with the formation of UO2(SO4)2
.
The decomposed fluorescence spectra of the two com-
plexes are presented in Figure 2 for each temperature. The
total intensity decreases when the temperature is increased
(21) Billard, I.; Ansoborlo, E.; Apperson, K.; Arpigny, S.; Azenha, M. E.;
Birch, D.; Bros, P.; Burrows, H. D.; Choppin, G.; Couston, L.; Dubois,
V.; Fanghänel, T.; Geipel, G.; Hubert, S.; Kim, J. I.; Kimura, T.;
Klenze, R.; Kronenberg, A.; Kumke, M.; Lagarde, G.; Lamarque, G.;
Lis, S.; Madic, C.; Meinrath, G.; Moulin, C.; Nagaishi, R.; Parker,
D.; Plancque, G.; Scherbaum, F.; Simoni, E.; Sinkov, S.; Viallesou-
branne, C. Appl. Spectrosc. 2003, 57, 1027–1038.
(22) Zanonato, P.; Di Bernardo, P.; Bismondo, A.; Liu, G. K.; Chen, X. Y.;
Rao, L. F. J. Am. Chem. Soc. 2004, 126, 5515–5522.
2-
to the second complex UO2(SO4)2 remain the same
(23) Kirishima, A.; Kimura, T.; Tochiyama, O.; Yoshida, Z. Radiochim.
Acta 2004, 92, 889–896.
whatever the temperature. Temperature changes affect the
(25) Bell, J. T.; Biggers, R. E. J. Mol. Spectrosc. 1968, 25, 312–329.
Inorganic Chemistry, Vol. 47, No. 6, 2008 2183
(24) Brachmann, A.; Geipel, G.; Bernhard, G.; Nitsche, H. Radiochim. Acta
2002, 90, 147–153.

Vercouter et al.
2-
Figure 3. Validation of the speciation model at [Na+] ) 0.1 M with the formation of UO2SO4 (aq) and UO2(SO4)2 at different temperatures.
probability of the radiative transitions and usually induce a
decrease of the fluorescence yield. The decreased relative
intensities of the 515 and 538 nm peaks of UO2SO4 (aq) at
75 °C do not result from a less-reliable spectral decomposi-
tion, although the fluorescence signal was much lower at 75
°C than that at lower temperatures: indeed, when forcing
the spectrum of UO2SO4 (aq) at 75 °C to have the same
relative intensities as those at the lower temperatures, the fit
of the measured spectra was totally unsatisfactory. Despite
insufficient information to discuss this spectroscopic obser-
vation further, we believe that it is not an artifact of the fit,
but that it is rather related to slight changes in the surround-
the correction that we made to account for [H+] variations
induces larger uncertainties on [UO22+] at 75 °C than at lower
temperatures because a significant increase of pH from 2.7
to 3.2 was expected. Moreover, no significant deviation is
observed for i ) 2 for which the UO2OH+ fluorescence
contribution has little impact. Consequently, the data points
2-
at low log10 mSO and 75 °C were given more weight in
4
the fit.
The dependence of the resulting log10 Ki values on the
temperature is represented in Figure 4a. An increase of the
temperature from 10 to 75 °C enhances the stability
of the mono- and disulfate complexes of U(VI) by an order
of magnitude, at most, on the K values. Consequently, when
the temperature is increased, complexation occurs at lower
sulfate concentrations. The corresponding thermodynamic
parameters at 25 °C were fitted on the basis of eq 8 with
heat capacities of reaction, ∆rCp,1 and ∆rCp,2, held at zero:
log10 K1 ) 2.45 ( 0.05, ∆rH1 ) 29.1 ( 4.0 kJ · mol-1, log10
K2 ) 1.03 ( 0.04, and ∆rH2 ) 16.6 ( 4.5 kJ · mol-1.
According to the uncertainties assigned to the log10 Ki values,
the enthalpies of reaction were assumed to be temperature-
independent. Thus, ∆rCp,1 and ∆rCp,2 could not be accurately
determined with our data sets.
2+
ings of the UO2 ion in UO2SO4 (aq) at 75 °C.
The attribution of the decomposed spectra to the UO2SO4
2-
(aq) and UO2(SO4)2 complexes was confirmed by the
dependence of the ratios of the U(VI) species concentrations
on mSO 2- using log–log representations (Figure 3). For each
4
temperature, linear variations with slopes of 1 well described
the data in accordance with eq 7 for i ) 1 and 2. This slope
analysis is a validation of the speciation model that was
considered (eqs 3 and 4), which is also in agreement with
the expected stoichiometries at 20 °C for such sulfate
concentrations.7 For i ) 1, the fitted curves slightly deviate
from a straight line at the highest sulfate concentrations
because of eq 9, which predicts a decrease of log10 K1 (by
0.18 at most) due to the medium effect. No indication of
4.2. Speciation of U(VI) in a 3 M Na+ Ionic Medium. The
4-
formation of UO2(SO4)3 was investigated by additional
experiments at higher ionic strengths, which enables a higher
reach of sulfate concentrations and usually stabilizes highly
charged species. Similarly to the previous series, a 3 M
NaClO4 solution was titrated with a 1.5 M Na2SO4 solution,
both solutions containing 0.4 µM U(VI). The total U(VI)
concentration was lowered compared to that in the experi-
ments at low ionic strength in order to keep unchanged the
detector acquisition parameters since the fluorescence of
U(VI) in concentrated sulfate solutions is much more intense.
A comparison with a series with 2.1 µM U(VI) at 20 °C
4-
the presence of UO2(SO4)3 was found since the fit was
not improved by introducing eq 5 in the model. At 75 °C,
the data points for i ) 1 fall under the straight line by more
2-
than their error bars for log10 mSO > -1.6 (Figure 3e).
4
This deviation cannot be explained by a variation of the
specific ion interaction coefficients at 75 °C, because their
influence in a 0.1 M Na+ ionic medium is small. Since the
concentration of the free uranyl is determined from the
fluorescence contributions of UO2OH+, it is more likely that
2184 Inorganic Chemistry, Vol. 47, No. 6, 2008

Aqueous Sulfate Complexes of U(VI)
Figure 5. Fluorescence spectra of a 2.1 µM U(VI) solution with 1.5 M
Na2SO4 (-log10 [H+] ) 3.0) measured at different temperatures.
tration in the solutions varies with [SO42-]. When this
spectrum shape was attributed to UO2(SO4)22-, the fit of the
data by adjusting the unknown spectrum of UO2SO4 (aq),
and the proportion of each species, was unacceptable because
the derived complexation constants K1 and K2 were unreal-
istic compared to literature values.26 Thus, the spectra
measured in the 1.5 M Na2SO4 solution (Figure 5) were
assigned to UO2(SO4)34-. It should be noted that the
fluorescence peaks are slightly shifted to the higher wave-
lengths as the temperature increases. These spectroscopic
changes might be related to modifications in the structure
of UO2(SO4)34-, and probably to the coordination mode of
the three sulfate ligands. When the spectra and the propor-
tions of the other complexes were fitted together on the
experimental spectra, all of the parameters were again so
correlated that the resulting values were irrelevant. Conse-
quently, the individual spectra of UO2SO4 (aq) and
Figure 4. Experimental values of complexation constants as a function of
temperature (a) for a 0.1 M Na+ ionic medium and (b) for a 3 M Na+ ionic
medium, based on K1 and K2 estimations (see text).
showed that the fluorescence signals were proportional to
the total U(VI) concentrations, as expected. [H+] was
maintained almost constant (-log10 [H+] ) 2.9 or 3.8) along
the titrations. The fluorescence contribution of UO2OH+ was
undetectable. It appears that the high ionic strength does not
promote the fluorescence of UO2OH+, on the contrary to
what was observed at low ionic strength. Thus, the uncom-
UO2(SO4)2 in the 3 M Na+ ionic medium could not be
2-
plexed uranyl was here fully characterized by the fluores-
unambiguously determined. Neither K1 nor K2 was deter-
mined in the final interpretation of these data. The values of
K1 and K2 were rather held constant during the fit by using
mean values calculated from three independent experimental
determinations by Ciavatta et al. for a 3 M NaClO4 medium
at 25 °C.26 The values of K1 and K2 at each temperature of
interest were calculated by using eq 8 and ∆rH1 and ∆rH2,
which were obtained in the 0.1 M NaClO4 medium (Table
2). The differences of ∆rHi between the 0.1 and 3 M NaClO4
media were supposed to be smaller than the uncertainties,
which is usually verified for similar systems and is consistent
with their theoretical dependence on activity coefficients.27
The sensitivity of the model to the experimental data was
tested according to eqs 2 and 7 by plotting the fluorescence
intensities at 494 nm as a function of the sulfate concentration
(Figure 6). Since K1 and K2 were fixed, only K3 and the molar
2+
cence of UO2
.
The peak maxima were continuously shifted to higher
wavelengths as the sulfate concentration increased up to 1.5
M. Attempts to interpret the data as for the previous series
of experiments were unsuccessful because the successive
formations of the complexes occur for close values of
[SO42-]. It was therefore difficult to accurately determine
the spectrum of each complex, and the fitting induced very
important correlation between all of the parameters. As a
first step, the fitting strategy was based on the measurements
at 20 °C. The fluorescence spectra of UO2SO4 (aq) and
UO2(SO4)2 determined in the 0.1 M Na+ ionic medium
2-
could not correctly fit the experimental spectra in the 3 M
Na+ ionic medium, even for the lowest sulfate concentrations
where the speciation should be completely defined by these
two complexes. It can be concluded that the individual
spectra are affected by the ionic medium and are not
transferable from specific ionic conditions to others, as
discussed in section 5.1. Consequently, the spectra were
treated independently from the results in the 0.1 M ionic
medium. At the highest sulfate concentrations (0.3–1.5 M),
the total intensity of the measured spectrum varies with
[SO42-], but the shape of the spectra remains unchanged.
The most probable reason for that is a predominance of the
spectral contribution of one single complex whose concen-
0
fluorescence intensities FS (494 nm) of the three complexes
were adjusted, resulting in a good fit of the data. As a
comparison, a poorer fit is presented when omitting the third
4-
complex, UO2(SO4)3 (Figure 6a). The diagram in Figure
6b shows that none of the complexes strongly dominates the
speciation, except at the highest sulfate concentrations where
(26) Ciavatta, L.; De Tommaso, G.; Iuliano, M. Ann. Chim. 2003, 93, 269–
279.
(27) Giffaut, E.; Vitorge, P.; Capdevila, H. J. Alloys Compd. 1994, 213,
278–285.
Inorganic Chemistry, Vol. 47, No. 6, 2008 2185

Vercouter et al.
Table 2. Thermodynamic Parameters for the Stepwise Formation of Sulfate Complexes of U(VI) at 25 °C
log10 Ki°
log10 Ki
∆rHi (kJ·mol-1
)
∆rSia (J·mol-1 ·K-1
)
exptl. medium (mol ·L-1
)
method
ref.
UO22+ + SO42- h UO2SO4 (aq)
20.79 ( 0.29
26.99 ( 1.84b
29.96 ( 2.38c
18.23 ( 0.17
19.6 ( 0.7
29.1 ( 4.0
21 ( 2
122
dilute H2SO4
calorimetry
9
151 ( 6b
163 ( 8c
p.w.
p.w.
8
10
p.w.
11
1 NaClO4
calorimetry
calorimetry
TRLFS
Donnan membrane
review
127 ( 2
145 ( 13
130 ( 5
∼0.01 Na2SO4, pH 2–3
3.29 ( 0.10
3.14 ( 0.03
3.15 ( 0.02
2.45 ( 0.05
0.1 NaClO4, pH 3–4
0.01–0.15 NH4ClO4, pH ∼ 5
19.5 ( 1.6
125.7 ( 5.4
14
2-
UO2SO4 (aq) + SO42- h UO2(SO4)2
16.88 ( 0.36
18.2 ( 2.1
16.6 ( 4.5
8 ( 5
1 NaClO4
0.01 Na2SO4, pH 2–3
0.1 NaClO4, pH 3–4
0.01–0.15 NH4ClO4, pH ∼ 5
calorimetry
calorimetry
TRLFS
Donnan membrane
review
8
84 ( 2
77 ( 15
47 ( 12
71.3 ( 4.6
10
p.w.
11
14
1.04 ( 0.10
1.06 ( 0.13
0.99 ( 0.05
1.03 ( 0.04
0.76 ( 0.20
15.6 ( 1.3
UO2(SO4)22- + SO42- h UO2(SO4)3
4-
11 ( 8
25 ( 27
3 NaClO4, pH 3–4
TRLFS
p.w.
a
b
c
Calculated from ∆rG ) ∆rH - T∆rS and ∆rG ) -RT ln K. Value obtained from ref 9, using log10 K1 ) 3.15 (NEA). Value obtained from ref 9,
using log10 K1 ) 3.29 (p.w.).
Figure 7. Typical fluorescence decays at the main wavelengths of emission
of a 4.2 µM U(VI) solution with 0.05 M Na2SO4 (-log10 [H+] ) 3.1) and
at 10 °C. The data are fitted with linear functions and a single lifetime
value (16.9 ( 0.1 µs) for all of the wavelengths (—).
Figure 6. (a) Fluorescence intensities measured at 494 nm as a function
of the sulfate concentration for 0.4 µM U(VI) in Na2SO4/NaClO4 solutions
([Na+] ) 3 M, -log10 [H+] ) 3.8) at 20 °C. Theoretical curves are fitted
to the data using K1 and K2 values from Ciavatta et al.26 for a two-complex
model and for a three-complex model (fitting K3). (b) Speciation diagram
corresponding to the three-complex model.
eq 1 and a single τS value of 16.9 ( 0.1 µs was successful,
while one may have expected at most four decay components
(if all species would have been detectable). At a given
temperature, the measured fluorescence lifetimes of U(VI)
increased with increasing sulfate concentration. Moreover,
at a given sulfate concentration, ln(τS) decreased linearly with
1/T (not shown) while different U(VI) species distributions
are expected. According to the Arrhenius relationship, which
often describes the temperature dependence of reaction
kinetics,23 the activation energy (Ea) of temperature quench-
ing of U(VI) fluorescence was determined by linear regres-
sion analysis for different sulfate concentrations: Ea ) 49.6
( 0.3, 47.8 ( 0.3, and 48.0 ( 0.3 kJ · mol-1 for 0.05, 0.15,
and 1.5 M Na2SO4 solutions, respectively. These values are
of the same order of magnitude as those obtained for Na2SO4/
NaClO4 solutions,28 as well as for sulfuric acid solutions,29
and for phosphate and fluoride solutions.23 It should be noted
that, at 20 °C and low sulfate concentrations where the molar
4-
UO2(SO4)3 represents more than 80% of the total U(VI)
concentration. This explains why we failed at obtaining the
individual spectra of UO2SO4 (aq) and UO2(SO4)22- by fitting
all the parameters simultaneously. The same analysis was
carried out for the other temperatures. The temperature
dependence of the experimental values of K3 in the 3 M Na+
ionic medium is shown in Figure 4b. The thermodynamic
parameters were determined by a weighed fit of the data
using eq 8: log10 K3(25 °C) ) 0.76 ( 0.20 and ∆rH3 ) 11
( 8 kJ· mol-1. A constant enthalpy approach was found to
be relevant to explain this data set, so ∆rCp,3 was not
determined.
4.3. Fluorescence Lifetimes of U(VI) in Aqueous Sul-
fate Solutions. The decays of fluorescence could be fitted
with monoexponential functions for each sulfate concentra-
tion and at each temperature. As an example, the fluorescence
decays measured at four different wavelengths are shown
for a sulfate solution of U(VI) where the expected speciation
is 7% UO22+, 54% UO2SO4 (aq), 26% UO2(SO4)22-, and 13%
UO2(SO4)34- at 10 °C (Figure 7). Simultaneous fitting with
(28) Kimura, T.; Nagaishi, R.; Ozaki, T.; Arisaka, M.; Yoshida, Z. J. Nucl.
Sci. Technol. 2002, 3, 233–239.
(29) Lotnik, S. V.; Khamidullina, L. A.; Kazakov, V. P. Radiochemistry
2003, 45, 550–554.
2186 Inorganic Chemistry, Vol. 47, No. 6, 2008

Aqueous Sulfate Complexes of U(VI)
2+
fraction of UO2 is more important, a biexponential decay
compositions of the test solutions. Recently, Hennig et al.
reported EXAFS data for several H2SO4/(NH4)2SO4 solutions
of U(VI).36 The bidentate sulfate coordination was deduced
from the U-S distances of 3.07–3.12 Å in most of the test
solutions, and also monodentate sulfate coordination in two
of them with U–S distances of 3.56–3.57 Å. The average
is observed with one short-life component (2.3 µs) in consis-
tency with the UO22+ lifetime and a longer-life component that
could rather correspond to an apparent lifetime of the sulfate
complexes.
The monoexponential decay curves can be explained by
either photochemical processes or by the small differences
of fluorescence lifetimes between the sulfate complexes,
which would then be hardly discriminated. Monoexponential
decays were also observed in another study at sulfate
concentrations where at least two sulfate complexes of U(VI)
were expected to form,30 whereas formation constants were
determined from multiexponential decays of the fluorescence
of U(VI) in sulfate solutions.15 Such inconsistencies for the
uranyl-sulfate aqueous system have already been pointed
out, but no definite conclusion has been drawn.21,31,32 It was
proposed that photochemical processes may interfere in the
determination of equilibrium constants by using TRLFS
when the formation/dissociation of complexes in their excited
states is rapid as compared to the rate of relaxation of the
excited complexes by fluorescence emission;32 it was also
concludedthatthisshouldnotbethecasefortheuranyl-sulfate
system. Then, it is more likely that the lifetimes of the sulfate
complexes of U(VI) are very close, and accurate determi-
nation of them is difficult. Therefore, only the value 18.1 (
0.8 µs was assigned to the fluorescence lifetime of UO2(SO4)34-
at 20 °C because this complex is highly predominant in the 1.5
M Na2SO4 solution.
2+
numbers of S atoms in the first coordination sphere of UO2
were fitted to the data to assess the complexes’ stoichiom-
etries and were compared to the expected speciation of
U(VI). The high ionic strengths required the choice of a
model for the calculation of ion activity coefficients.37
Hennig et al. have considered that, although the simple
Davies equation is not valid in the present range of ionic
strength (the Davies equation usually applies well at ionic
strengths lower than 0.1 mol ·kg-1), it could estimate the ion
activity coefficient without introducing larger errors than the
errors obtained when applying either the SIT or Pitzer model
with undetermined parameters.36 Similarly to the ionic
medium correction performed in the present work using the
SIT formula, we re-evaluated the speciation in the test
solutions of Hennig et al., by using reported ion interaction
coefficients7,34 and values for ion interaction with NH4+ by
analogy to Na+ as a first approximation. The calculation of
the free sulfate concentration also accounted for the degree
+
of association between NH4 and SO42-, but the ionic
medium effect on the corresponding association constant may
be misestimated under these conditions. This speciation
evaluation required an iterative procedure because the
calculated ionic strength and the free sulfate concentration
were correlated parameters. The final results were obtained
within only a few steps. Except for one sample, the calculated
ionic strength was higher than 0.6 mol ·L-1 and was between
2.4 and 6.8 mol ·L-1 for most of the solutions. The speciation
was very different from that proposed by Hennig et al.36
5. Discussion
5.1. Coordination of Sulfate in the Complexes. The
fluorescence spectrum of the sulfate complex of U(VI) was
sensitive to the 0.1 and 3 M Na+ ionic media, suggesting a
change in the local geometry of UO2 in the complexes.
Extended X-ray absorbance fine structure (EXAFS) analysis
of uranyl in sulfate solutions showed bidentate coordination
of the sulfate ligand in UO2SO4 (aq) and UO2(SO4)2
according to the U-S distances at about 3.12 Å, which is a
characteristic distance for bidentate SO4 in solid-state
2+
4-
Particularly, the UO2(SO4)3 complex was found to be
predominant in most of the samples (namely, C, D, E, and
F36). We believe that the speciation calculation with the SIT
formula is more correct despite the propagation of errors due
to estimated parameters. From our estimations of the species
concentrations, average numbers of S atoms coordinated to
UO22+ are expected to range between 2.6 and 3.0. The values
derived from the EXAFS spectra were one unit lower while
the Debye–Waller factor was fixed during the EXAFS fit
procedure.36 This latter constraint and the possible effect of
double-electron excitation36 should contribute to an increase
in the reported error of (15% in coordination numbers. Then,
while monodentate and bidentate coordination was clearly
evidenced in these solutions, the final interpretation of the
stoichiometries based on the average S coordination numbers
is ambiguous.
2-
,
2-
structures.33 The concentrations of the uranyl ion and
complexes in the test solutions were estimated by Moll et
al.33 from thermodynamic speciation, which was revised by
Vallet and Grenthe due to incorrect estimations of ion activity
coefficients.34 Conversely, U-S distances at about 3.67 Å,
indicating monodentate coordination, were determined by
high-energy X-ray scattering on a UO2SO4 (aq) solution,35
for which an accurate calculation of the speciation is difficult
due to the high total uranium concentrations.34 Both experi-
mental techniques provide reliable distances, while the
observed difference is likely due to differences in the
From the structural studies, it appears that either mono-
or bidentate coordination can be stabilized depending on the
sulfate concentration, uranium concentration, pH, or ionic
(30) Kimura, T.; Nagaishi, R.; Ozaki, T.; Kitatsuji, Y. Abstr. Pap. Am.
Chem. Soc. 2002, 223, B164–B165.
(31) Moriyasu, M.; Yokoyama, Y.; Ikeda, S. J. Inorg. Nucl. Chem. 1977,
39, 2199–2203.
(32) Billard, I.; Lutzenkirchen, K. Radiochim. Acta 2003, 91, 285–294.
(33) Moll, H.; Reich, T.; Hennig, C.; Rossberg, A.; Szabo, Z.; Grenthe, I.
Radiochim. Acta 2000, 88, 559–566.
(36) Hennig, C.; Schmeide, K.; Brendler, V.; Moll, H.; Tsushima, S.;
Scheinost, A. C. Inorg. Chem. 2007, 46, 5882–5892.
(37) Grenthe, I.; Plyasunov, A.; Spahiu, K. Estimation of medium effects
on thermodynamic data; Grenthe, I., Puigdomenech, I., Eds.; OECD:
Paris, 1997.
(34) Vallet, V.; Grenthe, I. C. R. Chim. 2007, 10, 905–915.
(35) Neuefeind, J.; Skanthakumar, S.; Soderholm, L. Inorg. Chem. 2004,
43, 2422–2426.
Inorganic Chemistry, Vol. 47, No. 6, 2008 2187

Vercouter et al.
medium. The geometry of different isomers of UO2SO4 (aq)
and UO2(SO4)2 has been calculated by quantum chemis-
selections, although the parameters used for the calculations
of activity coefficients were different from those recom-
mended by the NEA. Hence, no value is proposed at I ) 0
from our determination in the 3 M Na+ medium because
such ionic medium correction may induce a large uncertainty.
The enthalpies and entropies of the stepwise formation
reactions were both found to be positive, which is usually
observed for the complex formation of actinide and lan-
thanide ions with inorganic ligands. In this case, the entropic
stabilization drives the complexation reaction. The stepwise
formation reactions are endothermic, which suggests that the
main contribution to the enthalpy is the dehydration energy
of the reactants, which is less and less unfavorable as the
number of sulfate ligands increases. While ∆rH2 ) 16.6 (
4.5 kJ · mol-1 is in agreement with calorimetric determina-
tions,8,10 ∆rH1 ) 29.1 ( 4.0 kJ·mol-1 is significantly higher
than published values by a few kilojoules per mole (Table
2). Since a single thermodynamic data set should be
consistent with all experimental data, a closer examination
of previous calorimetric results seemed necessary.
∆rH1° and ∆rS1° have been determined by Bailey and
Larson from measured heats of solution and calculated heats
of dissociation of salts (i.e., excess enthalpy changes due to
the nonideality of the solutions), and assuming the value log10
K1° ) 2.72.9 Although this is a very careful work, the chosen
value of K1° is too low. We have therefore recalculated ∆rH1°
using different K1° values. In order to verify that our
numerical treatment was consistent with Bailey and Larson’s
results, the calculation was first performed using the same
assumptions as in the original work: we obtained ∆rH1° )
21.00 ( 0.33 kJ · mol-1, which compares to ∆rH1° ) 20.84
( 0.42 kJ ·mol-1 as reported by Bailey and Larson. It should
be noticed that the calculated term due to nonideality is quite
small because of the low ionic strength of the solutions, and
neglecting it led to ∆rH1° ) 21.42 ( 0.33 kJ · mol-1. The
∆rH1° values were markedly increased when using the K1°
value either recommended by the NEA (log10 K1° ) 3.15)7
or obtained at 25 °C in the present TRLFS investigation
(log10 K1° ) 3.29), leading to 26.99 ( 1.84 and 29.96 (
2.38 kJ · mol-1, respectively. These values of the enthalpy
of reaction better compare to the values obtained in the
present work, ∆rH1° ) 29.1 ( 4.0 kJ · mol-1, while the
uncertainties, calculated as the standard deviation in the four
experimental sets, increased significantly. The authors have
also determined ∆rH1° by using a different experimental path,
which consisted of measuring the heat of dilution of UO2-
(NO3)2 · 6H2O in K2SO4 solutions. Whereas their results
seemed to confirm their previous determination, we did not
manage to arrive at the same result according to the given
equations and data, and we suspect an error in the reported
data or results. We rather calculated 31.55 ( 1.46, 23.51 (
0.67, and 22.22 ( 0.50 kJ · mol-1 for ∆rH1° using log10 K1°
) 2.72 (original value), log10 K1° ) 3.15 (NEA), and log10
K1° ) 3.29 (this work), respectively. The value of ∆rH1°
decreased with increasing log10 K1°, but it is unclear whether
these determinations are reliable.
2-
try.34,36 Several isomers would have nearly the same energy
and could exist in solution. Moreover, the equilibrium
between the bidentate and the monodentate isomers of
UO2SO4 (aq) mostly depends on the water activity because
the change of coordination is necessarily accompanied by a
removal of at least one water molecule in the first coordina-
tion sphere of UO22+. Thus, the isomers of a complex are at
equilibrium concentrations when the water activity is constant.
We have concluded from our data treatment that the
fluorescence spectra assigned to UO2SO4 (aq) and
UO2(SO4)2 in the 0.1 M Na+ medium (Figure 2) are
2-
necessarily different from those in the 3 M Na+ medium
2+
(not determined). Hence, the average environment of UO2
in each complex probably changes when the ionic composi-
tion of the solution changes. Furthermore, in the range 10–75
°C, the fluorescence spectra of UO2SO4 (aq) and UO2(SO4)22-
in the 0.1 M Na+ medium do not show significant changes,
except for the spectrum of UO2SO4 (aq) at 75 °C, which
has different relative peak intensities (Figure 2). Conversely,
a batochromic shift of the spectrum of UO2(SO4)34- measured
in the 3 M Na+ medium is observed as the temperature
increases (Figure 5). These spectral changes could indicate
a modification of the equilibrium between isomers for
4-
UO2SO4 (aq) and UO2(SO4)3 that would be originated in
changes of water solvating properties when increasing the
temperature and either stronger or weaker bonding for the
third ligand. No spectral change is detected at 75 °C for
UO2(SO4)22-, suggesting that one isomer would predominate
at all temperatures. The more probable isomers are five-
coordinated complexes at ambient temperature.34 Thus,
4-
UO2(SO4)3 would have at least one monodentate sulfate
group because three bidentate sulfate groups would lead to
a six coordination. However, there is no reported U-S
distance of about 3.6 Å in solutions where this complex
would predominate.36
5.2. Thermodynamic Data. The values of log10 K1 and
log10 K2 determined in the 0.1 M Na+ medium were
extrapolated to I ) 0 using the SIT formula. The values at
25 °C compare well to the values selected by the NEA14
despite a slightly higher value for log10 K1° (Table 2). For
the formation of UO2(SO4)34-, Geipel et al. proposed log10
ꢀ3 ) 3.20 ( 0.25 from their analysis of the changes of the
fluorescence lifetime of U(VI) in a 1 M ionic medium at 25
°C.15 This value was selected by the NEA as the only
available one.7 Extrapolation to I ) 0 with the SIT formula
was performed by the NEA, using ∆ꢀ ) -0.11 instead of
the -0.34 originally used by Geipel et al., leading to log10
ꢀ3° ) 3.02 ( 0.38. In the present work, log10 ꢀ3 ) 3.68 (
0.24 was determined in a 3 M Na+ medium at 25 °C. This
value cannot be extrapolated from such a high-ionic-strength
medium to I ) 0 by using the SIT formula with a single ∆ꢀ
parameter. Indeed, in their study performed in a 3 M NaClO4
medium, Ciavatta et al. rather calculated the activity coef-
ficients of the UO2SO4 (aq) and UO2(SO4)22- complexes by
using empirical rules.26 The resulting values of ꢀ1° and ꢀ2°
were considered by the NEA to be consistent with their
The calorimetric measurements by Ahrland and Kullberg
were carried out under acidic conditions and at I ) 1 M,
2188 Inorganic Chemistry, Vol. 47, No. 6, 2008

Aqueous Sulfate Complexes of U(VI)
with uranyl concentrations ranging between 0.01 and 0.03
M. They obtained ∆rH1 ) 18.23 ( 0.17 kJ· mol-1, which is
lower than the one obtained from our TRLFS results, but
we could not find a reason for such a discrepancy. The value
∆rH2 ) 16.88 ( 0.36 kJ · mol-1 is in much better agreement
with our determination.
coordinations can be expected in UO2SO4 (aq) and
2-
UO2(SO4)2 depending on the solution composition, as
-
discussed above. ∆rH3 is low for UO2(Ac)3 (∼-2
39
kJ· mol-1) and highly negative for UO2(CO3)3 (∼-55
kJ ·mol-1),41 because the dehydration energy is probably low
in the third complex and can be overcome by the bond
formations. Comparison of the ∆rH3 values suggests that the
sulfate anions could act as monodentate and bidentate ligands
4-
The microcalorimetric study by Ullman and Schreiner was
performed by titrating U(VI) solutions with Na2SO4 solutions
at pH 2.1 and 2.7.10 The ionic strength of the solutions was
not noticed, but we calculated that it should range between
0.2 and 1.2 M according to the amounts of UO2(NO3)2 and
Na2SO4 introduced in the test solutions. A calculation of the
expected speciation at I ) 0.25 M (average value of the
experimental ionic strengths) using data selected by the NEA
shows that the hydrolysis species (UO2)2OH3+ and
2+
toward UO2 in the third complex, as in the case of
-
UO2(Ac)3 .
6. Conclusion
Three uranyl sulfate complexes, UO2SO4 (aq), UO2-
(SO4)22-, and UO2(SO4)34-, were identified by TRLFS in
aqueous solutions in the temperature range 10–75 °C.
UO2(SO4)34- was only detected at high sulfate concentrations
in a 3 M Na+ ionic medium. A set of thermodynamic
formation data was determined. Though the value of ∆rH1
was significantly higher than those from calorimetric studies,
it compares well with reinterpreted results from one of the
calorimetric studies. The value of ∆rH2 was found to be in
good agreement with calorimetric results. A value of ∆rH3
was determined for the first time. The fluorescence spectra
measured in low- and high-ionic-strength media were
interpreted as being indicative of the presence of different
isomers of UO2SO4 (aq) and UO2(SO4)22-, in accordance
with structural data that shows the existence of both bidentate
and monodentate chelation of sulfate in UO2SO4 (aq) and
2+
(UO2)2(OH)2 may have significantly formed (more than
10%) in the initial uranyl solutions at pH 2.7, and to a minor
extent at pH 2.1, while only UO22+, UO2SO4 (aq), and
UO2(SO4)22- were accounted for in the interpretation of the
data. Even small amounts of hydrolysis species can signifi-
cantly alter the calorimetric results since, for instance, the
corresponding enthalpy of dissociation of (UO2)2(OH)22+ is
about -48 kJ · mol-1.22 Furthermore, ternary hydroxo-sul-
fate complexes were evidenced more recently,5,7,38 and
2-
(UO2)2(OH)2(SO4)2 could be stable at a significant con-
centration under the reaction conditions. Thus, a possible
influence of other formation or dissociation reactions on the
measured heats of solution cannot be ruled out.
2-
∆rH1 and ∆rH2 in the 0.1 M Na+ ionic medium and ∆rH3
in the 3 M Na+ ionic medium were found to be positive,
following the order ∆rH1 > ∆rH2 > ∆rH3. Such a trend was
also observed for the complexation of U(VI) by acetate,8,39
which is a moderate oxygen-donor ligand, but stronger than
sulfate. The single charged acetate anion (Ac-) can bind to
the uranium ion with two oxygen atoms; bidentate coordina-
tion in the UO2Ac+ and UO2(Ac)2 (aq) complexes was
evidenced by EXAFS39 and X-ray absorption40 spec-
troscopies, while monodentate acetate coordination in
UO2(SO4)2 depending on the solution composition.
Acknowledgment. This work was supported by DSOE/
Basic Research and by DDIN/DPRGD within the cooperative
research and development CT4 program CEA-ANDRA-EDF
and is part of the 6th framework European community
program “FUNMIG” (FP6-516514). The authors are grateful
to anonymous reviewers for their fruitful comments.
Supporting Information Available: Experimental values of the
stepwise formation constants of UO2SO4 (aq), UO2(SO4)22-, and
UO2(SO4)34-. Fluorescence spectra of UO2OH+ at variable tem-
peratures. Fluorescence spectra of U(VI) in Na2SO4/NaClO4 media
with [Na+] ) 3 M at 20 °C. This material is available free of charge
via the Internet at http://pubs.acs.org.
UO2(Ac)3 was supported by EXAFS measurements.39 In
-
the case of sulfate complexation, both mono- and bidentate
(38) Ciavatta, L.; De Tommaso, G.; Iuliano, M. Ann. Chim. 2003, 93, 281–
290.
(39) Jiang, J.; Rao, L. F.; Di Bernardo, P.; Zanonato, P.; Bismondo, A.
J. Chem. Soc., Dalton Trans. 2002, 1832–1838.
(40) Bailey, E. H.; Mosselmans, J. F. W.; Schofield, P. F. Geochim.
Cosmochim. Acta 2004, 68, 1711–1722.
IC701379Q
(41) Lemire, R. J.; Tremaine, P. R. J. Chem. Eng. Data 1980, 25, 361–
370.
Inorganic Chemistry, Vol. 47, No. 6, 2008 2189