Determination of successive complexation constants in an ionic liquid: complexation of uo22+ with no3- in c4-mintf2n studied by uv-v is spectroscopy

Article abstract of DOI:10.1021/jp9107624

The complexation of UO₂²⁺ with NO₃ ^- has been investigated in the ionic liquid l-butyl-3- methylimidazolium bis(trifluoromethylsulfonyl)imide by UV-vis spectroscopy at T = 18.5?°C. The complexation is evidenced through the appearance of four peaks at 425, 438, 453, and 467 nm. EXAFS data indicate that the trinitrato complex, UO₂(NO₃)₃^-, is dominating the speciation for a reagent ratio of [NO₃^-]/[UO ₂^-]>3. Assuming three successive complexation steps, the conditional stability constants are calculated, the individual absorption spectra are derived, and a speciation plot is presented. ? 2010 American Chemical Society.

Full text of DOI:10.1021/jp9107624

4276
J. Phys. Chem. B 2010, 114, 4276–4282
Determination of Successive Complexation Constants in an Ionic Liquid: Complexation of
2+
-
UO2 with NO3 in C4-mimTf2N Studied by UV-Vis Spectroscopy
Sylvia Georg,† Isabelle Billard,*,† Ali Ouadi,† Clotilde Gaillard,‡ Laetitia Petitjean,§
Michel Picquet,§ and Vitaly Solov’ev⊥
IPHC, UMR 7178 CNRS-IN2P3 et UniVersite´ de Strasbourg, 23 rue du Loess, 67037 Strasbourg cedex 2,
France, IPN Lyon, UniVersite´ Lyon I, 4 rue E. Fermi, 69622 Villeurbanne Cedex, France, Institut de Chimie
Mole´culaire de l’UniVersite´ de Bourgogne, UMR 5260 CNRS - UniVersite´ de Bourgogne, 9 aVenue A. SaVary,
BP 47870, 21078 Dijon, cedex, France, and Institute of Physical Chemistry and Electrochemistry of Russian
Academy of Sciences, Leninsky pr. 31a, 119991 Moscow, Russia
ReceiVed: NoVember 12, 2009; ReVised Manuscript ReceiVed: February 9, 2010
The complexation of UO22+ with NO3- has been investigated in the ionic liquid 1-butyl-3-methylimidazolium
bis(trifluoromethylsulfonyl)imide by UV-vis spectroscopy at T ) 18.5 °C. The complexation is evidenced
through the appearance of four peaks at 425, 438, 453, and 467 nm. EXAFS data indicate that the trinitrato
-
-
complex, UO2(NO3)3 , is dominating the speciation for a reagent ratio of [NO3 ]/[UO22+] > 3. Assuming
three successive complexation steps, the conditional stability constants are calculated, the individual absorption
spectra are derived, and a speciation plot is presented.
Introduction
imidazolium based cation). The uranyl cation, UO22+, has been
chosen as the reacting moiety, based on its well-defined UV-vis
Ionic liquids (ILs) are a class of solvents,1 composed of an
(in)organic cation (most frequently imidazolium, pyrrolidinium,
spectra and the amount of literature data on complexation in
various solvents by UV-vis spectroscopy, that can be used for
comparison purposes.19–22 The choice of the ligand relied on
various considerations. Usually (i.e., in water, acetonitrile, or
acetone), successive complexation occurs, eventually up to a
limiting complex such as UO2Cl42-. In the case of chloride,
four successive complexation reactions have to be considered,
rending the data analysis rather complex. We have already
shown that UO2SO4 and (UO2)3(PO4)2 are not soluble in
C4-mimTf2N,23,24 so that we finally selected the nitrate anion
-
tetraalkylammonium, etc.) and an inorganic anion (PF6-, BF4 ,
Cl-, CF3SO3-, (CF3SO2)2N-, etc.), with so many industrially
beneficial properties (in particular, ILs are nonflammable and
nonvolatile, thus being often called “green solvents”) that an
increasing number of publications is devoted to their study as
replacement solvents in all fields of chemistry (electrochemistry,
catalysis, synthesis, liquid/liquid extraction, etc.).2–6 Some very
fundamental or basic chemical aspects are also covered, such
as solvation, complexation, or extraction mechanism of various
ions or molecules in ILs.3,7 Besides the potential of ILs for
innovative chemistry, owing to the “onion shell” structure8–10
of the ionic solvation sphere in these media, ILs represent a
real challenge for theoreticians because of their intrinsic
(chemical) complexity. At the same time, ILs offer a unique
chemical media to test the various correction laws (either
empirical or analytical) for activity coefficients at very high ionic
strengths that have been proposed and studied for the aqueous
case.11–13 Surprisingly enough, however, to our knowledge, not
a single paper is concerned with the determination of equilibrium
chemical constants in ILs, which is a prerequisite to the study
of ionic strength effects and Coulombic interactions in ILs.
-
NO3 as the ligand, leading to a limiting complex of 1:3
stoichiometry only, which is of interest to studies related to the
nuclear fuel cycle and has already been partially studied in
C4-mimTf2N.24–26
In this paper, we present UV-vis data on the successive
2+
complexation of the uranyl cation UO2 with nitrate anions
-
NO3 in the ionic liquid C4-mimTf2N as the solvent. Both
reacting moieties have been introduced as salts, the counterion
being either C4-mim+ or Tf2N- (i.e., UO2(Tf2N)2 and C4-
mimNO3 were used as reagents) in order to suppress the possible
effect of “spectator ions” that cannot be avoided in usual neutral
solvents. Complexation constants have been derived from
UV-vis spectrophotometric data in the spectral range of
350-490 nm using either a dedicated approach or the
CHEMEQUI program as a general method for computations
of equilibrium constants from experimental results of physico-
chemical measurements.
Our group is interested in actinide and lanthanide chemistry
in ILs, both on fundamental and applied aspects.14–18 In this
respect, we intended to experimentally derive the equilibrium
constants of an actinide cation with a charged ligand in a
“classical” IL, 1-butyl-3-methylimidazolium bis(trifluorometh-
ylsulfonyl)imide (further denoted as C4-mimTf2N, Tf2N- stand-
ing for the (CF3SO2)2N- anion and C4-mim+ being the
Experimental Section
Uranyl Salt UO2(Tf2N)2. The synthesis procedure is partially
based on the work of Nockemann and co-workers.26 HTf2N
(95% Aldrich, white solid, highly hygroscopic) is used as
received. It is dissolved in ultrapure water (5 mL for 5 g) to
give an acidic solution. UO3 (∼350 mg) is dissolved in 2.5 mL
of this acidic solution, which is thereafter mechanically shaken
for 18 h until total dissolution of the solid. The resulting pale-
* To whom correspondence should be addressed. E-mail:
isabelle.billard@ires.in2p3.fr. Phone: +33388106401. Fax: +33388106431.
† UMR 7178 CNRS-IN2P3 et Universite´ de Strasbourg.
‡ Universite´ Lyon I.
§ UMR 5260 CNRS - Universite´ de Bourgogne.
⊥ Institute of Physical Chemistry and Electrochemistry of Russian
Academy of Sciences.
10.1021/jp9107624  2010 American Chemical Society
Published on Web 03/10/2010

2+
-
Complexation of UO2 with NO3 in C4-mimTf2N
J. Phys. Chem. B, Vol. 114, No. 12, 2010 4277
yellow solution is then dried under vacuum in a two-step
procedure; first, the pressure is at p ) 3.5 × 10-2 mbar until
solidification starts, and then, it is at p ) 5 × 10-7 mbar for
∼60 h. All of the drying process is achieved without heating,
but the glass container is embedded in a beaker of water at room
temperature to avoid ice formation on the outside. No further
purification is performed as the boiling point of HTf2N is Tb )
91 °C, allowing evaporation of the acid in excess. The solid
formed is pale-yellow and highly hygroscopic. Once the salt is
prepared, it is kept in a closed glass tube under vacuum and is
dried thoroughly prior to any use to ensure reproducibility of
the UV-vis data.
aliquots of the uranyl stock solution were mixed with required
quantities of one of the nitrate stock solutions to obtain the
desired R ratio. The NO3- concentrations of the stock solutions
were high enough to ensure a rather limited dilution of the uranyl
concentration in the final solution, whatever the R value. The
solution was then carefully degassed/dried for 2 h by gentle
heating (T ) 60 °C) under vacuum (p ) 3.5 × 10-2 mbar).
The heating was then stopped, and natural cooling of the solution
back to ambient temperature was allowed for one additional
hour, still under vacuum. The solution was then rapidly
transferred to the quartz cell and measured readily after. IR
measurements have confirmed that the water amount in the
degassed solution is very low, that is, below the Karl Fischer
detection limit of 50 ppm (less than 0.5 water molecules per
uranyl units), as was already shown.29
C4-mimNO3. C4-mimNO3 is obtained by adaptation of the
halogenide-free procedures described by Tkatchenko and co-
workers27 and Rogers and co-workers.28 HNO3 (68% solution
in water, Acros Organics) is used as received. Acetonitrile
(99.9%, Carlo Erba, HPLC grade) is distilled under argon over
CaH2 prior to use. Analyses are performed at the PACSMUB
(Universite´ de Bourgogne, France) on Bruker Avance 300 and
500 MHz NMR spectrometers, on a Bruker Vector 22 FT-IR
spectrometer equipped with a Golden-Gate ATR accessory, and
on a Fisons Instruments EA1108 for elemental analysis. 1-Butyl-
3-methylimidazolium-2-carboxylate27 (10.71 g, 58.8 mmol) is
dissolved in 100 mL of distilled water in a two-necked round-
bottom flask equipped with a reflux condenser under argon. To
this solution, a diluted solution of HNO3 (3.89 mL, 68%, 58.8
mmol, 1 equiv) in 100 mL of distilled water is added through
a dropping funnel over a 5 min period at room temperature.
The mixture is stirred 3 h at 100 °C and then allowed to cool
to room temperature, and water is evaporated under vacuum (p
) 2 × 10-2 mbar) at 35 °C. The resulting pale-yellow
C4-mimNO3 is further dried under vacuum at 60 °C overnight,
slowly turning into a deep-yellow-to-brown oil. Decolorization
is subsequently performed by solubilizing C4-mimNO3 in 15
mL of acetonitrile and stirring this solution for 1 h over a small
amount of activated charcoal. Filtration over a plug of Celite
and evaporation of the solvent at room temperature under
vacuum affords 11.292 g (95% yield) of C4-mimNO3 as a pale-
yellow oil, slowly crystallizing at room temperature. A repeated
drying/degassing procedure (see next section for description)
leads to a yellow oil that doe not crystallize any more. 1H NMR
(500.13 MHz, D2O): δ ) 8.68 (br s, 1 H, NCHN), 7.46 (t, 1 H,
J ) 2 Hz, N-CHdCH-N), 7.41 (t, 1 H, J ) 2 Hz,
N-CHdCH-N), 4.18 (t, 2 H, J ) 7 Hz, N-CH2), 3.87 (s, 3
H, N-CH3), 1.83 (m, 2 H, N-CH2-CH2), 1.29 (m, 2 H,
CH2-CH3), 0.90 (t, 3 H, J ) 7.5 Hz, CH2-CH3) ppm. 13C{1H}
NMR (75.47 MHz, CDCl3): δ ) 136.8 (NCHN), 123.4
(N-CHdCH-N), 122.3 (N-CHdCH-N), 49.1 (N-CH2),
35.6 (N-CH3), 31.6 (N-CH2-CH2), 18.9 (CH2-CH3), 12.9
(CH2-CH3) ppm. IR (ATR, neat): 3146 (w), 3095, 2960, 2875
(w), 1572, 1326 (vs), 1167 cm-1. Elemental analysis (calculated
for C8H15N3O3): C 47.34 (47.75), H 7.74 (7.51), N 20.87 (20.88)
%.
Various trials and repeated experiments have shown that this
sample preparation procedure leads to the best reproducibility
of the UV-vis spectra. We suggest that letting the water in the
C4-mimTf2N solution equilibrate between the uranyl moieties
and the bulk ensures an efficient degassing of the solution
afterward and therefore a very low amount of remaining water
that is known to eventually enter the uranyl coordination sphere.
Measurements. UV-vis spectra were recorded on a Cary
100 system (Varian). The temperature was controlled at 18.5
°C. The data were acquired in the range of 200-800 nm. The
reference sample was pure dry C4-mimTf2N.
All uranyl solutions were followed by UV-vis spectroscopy
for a few days after being prepared. While the samples were
stored in the dark between two UV-vis measurements, no
caution was taken to keep them away from light all of the time,
especially during the drying procedure that was performed
without any light protection. Although a slight change in the
absorbance was observed for the lowest R ) [NO3-]/[U(VI)]
values in the range of 200-400 nm during the first 24 h after
sample preparation, the spectra appeared to be very stable with
time (as checked up to 8 days) above 400 nm. In agreement
with a previous work,29 this demonstrates that the quartz cuvette
and the Teflon tap are waterproof. All data presented in this
work correspond to UV-vis spectra recorded 4 days or more
after sample preparation.
Data Fitting. First, an attempt was made to analyze the data
through the method of principal components30,31 in the range of
385-485 nm to determine the exact number of chemical species,
but the results were not conclusive. Second, the data were fitted
according to two different procedures, hereafter referred to as
“monowavelength” fitting and “multiwavelength” computations,
by use of the CHEMEQUI program in the latter case.
In the monowavelength case, a Fortran routine based on least-
squares adjustment (CERN library) has been written to fit the
variation of the absorbance as a function of R at a fixed
wavelength (see Figures 1 and 2) for the two chemical models
proposed (see the Modeling and Fitting Results section). The
absorbance variation is obtained from the Beer’s law, and the
analytical expression for the concentration of the species is
obtained, with the help of the Maple software, by solving the
system of mass action laws and mass-balance equations. The
unknown parameters are the various equilibrium constants ꢀi
and the individual molar absorption εi at the chosen wavelength.
In the multiwavelength computations, the computer program
CHEMEQUI, which is appropriated for the treatment of
experimental data obtained via different physicochemical meth-
ods,32 was used to calculate the stability constants and the molar
absorbance values from the spectrophotometric data. In the
calculations, the vector ꢀ of the unknown stability constants
Sample Preparation. The reagent concentrations of the metal
2+
-
ion UO2 and the ligand NO3 were varied from 0.0096 to
0.0101 M (UO22+) and from 0 to 0.0341 M (NO3-) so that the
molar ratio R ) [NO3-]/[UO22+] varied from 0 to 3.4.
C4-mimTf2N was purchased from Solvionic, with no further
purification, and contained ∼300 ppm of water. Two stock
solutions of C4-mimNO3 (1 and 0.2 M) and one stock solution
of UO2(Tf2N)2 (10-2 M) in C4-mimTf2N were prepared. This
latter solution was kept on the shelves in the dark for 4 days,
so that the remaining water, either from the C4-mimTf2N or
from the uranyl salt, could equilibrate in the bulk. Then, 3 mL

4278 J. Phys. Chem. B, Vol. 114, No. 12, 2010
Georg et al.
is the number of species involved in equilibrium, ελj is the molar
absorbance of the jth species, and Cj is the equilibrium
concentration of the jth species. The concentrations Cj as a
function of the formation constants are expressed in Brinkley’s
form33
m
Cj ) exp(ln ꢀj +
ν ln C )
(3)
∑
jk
k
k)1
where ꢀj is the overall stability constant of the jth product from
m basic species (minimal set of reagents; here, two reagents
UO22+ and NO3-) and νjk is the stoichiometric coefficient for a
given reaction. Concentrations satisfy the mass-balance equations
s
ν C ) C0
k ) 1, 2, ..., m
(4)
∑
ki
i
k
i)1
Figure 1. Absorption spectra for various values of R, the molar ratio
[NO3-]/[UO22+], in C4-mimTf2N (see text).
where νki is the stoichiometric coefficient of the kth basic species
2+
(UO2 and NO3-) in the reaction in which the ith species is
formed. As a criteria of reliability of the equilibrium model,
we used the Hamilton’s R factor34 and residuals analysis (Aexp,λ
- Aλ) versus Aλ.35
CHEMEQUI uses three algorithms of minimization, rapid
gradient method EQ with analytical presentation of derivatives
∂Φ/∂ꢀ and ∂Φ/∂ε36,37 and reliable simplex38 and Monte Carlo
(low sensitivity to correlated calculated parameters)39 approaches
allowing one to significantly improve the reliability of the
stability constant calculations. Details of the algorithms used
in CHEMEQUI are given elsewhere.32 In the studied reagent
concentration range, for estimations of highly correlated stability
constants and related values, fitting algorithms of CHEMEQUI
were extended by an option for cyclic fitting of the parameters
one by one.
Experimental Results and Literature Review
Considering the pale-yellow color of C4-mimNO3, in a first
step, the absorbance of C4-mimNO3 in C4-mimTf2N was
measured over the range of 200-800 nm for concentrations
between 0 and 1 M (data not shown). The nitrate ions present
an absorption band peaking at 309 nm, and the Beer’s law
appears to be perfectly verified at λ ) 309 nm. By contrast,
there is no significant absorption above 350 nm, up to 1 M.
Therefore, in the calculations to follow, the specific absorption
of the “free” nitrate ion (i.e., not complexed to uranyl) was not
considered in the monowavelength approach, while a prelimi-
nary CHEMEQUI calculation demonstrated that the absorption
of the nitrate ion is similar to the experimental error; therefore,
it was not taken into account in further calculations.
Figure 2. Variation of the absorbance as a function of R. O: λ ) 438
nm; × : λ ) 467 nm. Thick solid line: fit according to model I. Thin
solid line: fit according to model II (see text).
and the matrix ε of the molar absorbance were fitting parameters
for the minimization of the squares of the residuals
n
Φ(ꢀ,ε) )
(A
- Aλ,i)2 ) minimum (1)
exp,λ,i
∑ ∑
λ
i)1
where n is the number of experimental points and Aexp,λ and Aλ
are the experimental and calculated absorbances, respectively,
for a given wavelength. The sets of data for the series of the
wavelengths λ were merged together and treated simultaneously
to calculate the ꢀ and ε values. In the program, the calculated
absorbance Aλ is given by the Beer’s equation
Figure 1 displays selected spectra for R values varying
between 0 and 3.5. As can be seen, the addition of nitrate ions
to the UO22+/C4-mimTf2N solution leads to a tremendous
increase of the signal above 420 nm, with the appearance of
four well-defined bands at 425, 438, 453, and 467 nm. A small
shoulder is present at 488 nm. These changes clearly evidence
s
2+
the complexation of UO2 with NO3-. Figure 2 presents the
Aλ ) A0λ + l
ε C
(2)
∑
absorbance variation as a function of R at λ ) 438 nm, showing
a leveling off above ∼2.5. As noted in the Experimental Section,
according to the sample preparation, the uranyl concentration
is not constant from one sample to another but is slightly
decreasing while R is increasing, owing to the dilution effect.
λj
j
j)1
Here, for a given wavelength λ, A0λ is the background absorption
or systematic error, l is the optical path length (here, 1 cm), s

2+
-
Complexation of UO2 with NO3 in C4-mimTf2N
J. Phys. Chem. B, Vol. 114, No. 12, 2010 4279
Modeling and Fitting Results
Such tremendous changes in the absorption spectra as a
consequence of nitrate addition are well-documented in various
solvents.Theyhavebeenevidencedinwater,21,23,40,41 water-acetone
mixtures,20 acetone,25 acetonitrile,22,25 and various ionic liquids,
either dry or not, C4-mimTf2N,23–26 C4-mimPF6, and
Me3NBuTf2N24 or C4-mpyrTf2N.26 However, most of these
works focused on the qualitative features of the absorption
spectra, and besides the excellent paper of Ikeda and co-
workers22 concerning complexation in acetonitrile, publications
dealing with the determination of equilibrium constants and/or
the stoichiometry of the successive uranyl-nitrate complexes
and their individual spectra are scarce. In water, the reasons
for such a lack of information are easily understandable. Owing
to the strong affinity of water toward uranyl, even the first nitrate
complexation is weak. Although various values can be found
in the literature for all three complexation steps, with which a
speciation diagram can be calculated,40 the most reliable values
All together, data already found in the literature speak in favor
of two different chemical interpretations, between which it is
impossible to decide, (i) step-by-step complexation involving
the species UO22+, UO2NO3+, UO2(NO3)2, and UO2(NO3)3
-
formed in accordance with three equilibria (model I) or (ii)
starting from UO22+ (R ) 0), less than three equilibria leading
-
to the limiting complex UO2(NO3)3 for R g 3 (model II).
Chemically speaking, these two models can be written as follows
Model I:
2+
-
+
ꢀ1
ꢀ2
ꢀ3
UO2 + NO3 a UO2NO3
2+
-
UO2 + 2NO3 a UO2(NO3)2
2+
-
-
UO2 + 3NO3 a UO2(NO3)3
to date are K1° ) 1.995 M-1 and K2° ) 0.07 M-1 42
This in
.
Model II:
turns explains why in most papers, only the first two complexes
have been observed in water, by Raman43 or fluorescence,44
although a recent paper mentions the presence of the trinitrato
uranyl complex in water as observed by UV-vis and EXAFS,
but without any complexation constant values or speciation.41
It is possible that increasing temperature may favor complex-
ation, but to our knowledge, in water, only the first nitrato
complex has been investigated as a function of T.21 From the
various UV-vis spectra displayed in the literature for the water
case, it is clear that the two first complexes present the
“characteristic” peaks already mentioned at 425, 438, 453, and
467 nm, the difference between the species being just a question
of molar absorbance, not of shape or position. The same
phenomenon has been observed in acetonitrile as the individual
absorption spectra obtained by spectral decomposition22 show
very similar peak positions for all three complexes, the molar
absorbance of the 1:1 and 1:2 complexes being rather similar,
while the 1:3 complex presents an absorption coefficient a factor
of ∼2 larger than the complexes of lower stoichiometry in the
range of 420-480 nm. In ILs, information is even more limited.
Besides the paper by Bradley et al.45 displaying an almost
featureless UV-vis spectra of U(VI) in pure liquid C4-mimNO3
(which corresponds to chemical conditions very different from
ours), all available UV-vis data present the bands at 425, 438,
453, and 467 nm, but since these data have been obtained for
2+
-
-
UO2 + 3NO3 a UO2(NO3)3
ꢀ3
Other solutions are possible for model II, but we restrained
our study to the sole case above as it differs widely from model
I. Whatever the model chosen, it is deduced from EXAFS that
for R g 3, the major component in the solution is the trinitrato
-
complex, UO2(NO3)3 . This has some implications on the value
of ꢀ3.
It has to be noted that the way we write the various
chemical equilibria envisioned to occur in C4-mimTf2N is
most probably not an exact description of the real solvation
sphere of the reacting species as we do not mention the
solvent.10 MD calculations in dry ILs have repeatedly shown
that the first coordination sphere of cations is mostly
composed of the anions constituting the IL entity, with slight
variations in the number of anions depending on the cation
charge and the anion chemical structure.8,9,24 It should be
recalled that the global charge of the entity composed of the
cation and of its first solvation sphere in ILs is, most of the
time, highly negative. It has also been shown that the second
solvation sphere in ILs is also of importance, and the global
charge of this entity is, by contrast, highly positive. In water,
2+
the notations UO22+, UO2 aq, or UO22+ · 5H2O are equivalent,
as far as mathematics is concerned, although the way to
consider the solvation sphere is different from one to the
other. By contrast, in ILs, the situation is unclear because
the way the mass action law is expressed may change
dramatically depending on how many solvation spheres are
considered. The global charge of the reacting moieties varies
depending on the choice made, which has implications for
the mathematical expression of the activity coefficients.
Finally, the way to express the activity of the IL solvent itself
is not firmly assessed. As a consequence, and in order to
conduct some fitting operations on the spectroscopic data that
we collected, we decide to use “conditional” equilibrium
constants, written as
-
a single U(VI)/NO3 ratio of 4 (or values above R ) 4 as in
refs 26 and 23), no information can be deduced concerning the
individual UV-vis absorption spectra of the (possible) succes-
sive complexes in ILs.
Other indications can be derived from EXAFS. Actually,
EXAFS measurements performed for R ) 423,25 evidence that
three nitrates on average are bound to uranyl. As EXAFS only
gives access to averaged coordination numbers, it could be
envisionned that this could be the weighted combination of
various species, including the tetranitrato complex, but consider-
ing the steric hindrance of NO3-, this hypothesis can be ruled
out. Similarly, EXAFS data point to two nitrates bound on
average in the case of UO2(NO3)2 dissolved in C4-mimTf2N.24
However, contrary to the trinitrato complex, which is unambigu-
ously evidenced through EXAFS, as already explained, this does
not prove that UO2(NO3)2 is the only species present in solution
for R ) 2, as discussed previously.24,41 In our recent work,46
we have also obtained EXAFS data for R ) 1, indicating one
nitrate on average around uranyl, again with no definite proof
- i
ꢀi ) [UO2(NO3)i(2-i)]/[UO22+][NO3 ]
where [X] represents the concentration of species X. In
complex formation studies in aqueous solutions, it is very
important to control the ionic strength of the solution, which
is most of the time achieved by the addition of an inert salt
in large quantities as compared to the reagents. In the case
+
of UO2NO3 being alone in solution or in the presence of
complexes of higher stoichiometry.

4280 J. Phys. Chem. B, Vol. 114, No. 12, 2010
Georg et al.
2+
-
TABLE 1: Calculated Stability Constants for the Complexation of UO2 with NO3 and the Molar Absorbance Values of the
Species in C4-mimTf2N at 18.5 °Ca
experimental datab: wavelengths, nm
log ꢀ1
log ꢀ2
log ꢀ3
HRF, %c
2.56
439, 440, 452, 453, 454, 466, 468
4.81 ( 0.45
8.31( 0.65
12.17 ( 0.81
molar absorbanced, cm-1 M-1
no.
wavelength, nm
ε1
ε2
ε3
ε4
1
2
3
4
5
6
7
439
440
452
453
454
466
468
3.25 (0.36)
4.18 (0.34)
4.13 (0.33)
3.92 (0.32)
5.78 (0.35)
7.46 (0.33)
7.25 (0.32)
9.47 (0.63)
8.06 (0.72)
5.35 (0.41)
5.80 (0.42)
12.08 (0.65)
9.45 (0.42)
9.78 (0.42)
12.7 (1.5)
24.2 (6.0)
7.2 (1.3)
1.8 (3.4)
12.2 (1.6)
12.1 (1.4)
6.0 (3.8)
13.73 (0.19)
18.18 (0.23)
12.31 (0.22)
10.60 (0.27)
18.80 (0.26)
18.58 (0.26)
16.26 (0.30)
a Equilibria UO2 + NO3 ) UO2(NO3)+, log ꢀ1; UO2 + 2NO3 ) UO2(NO3)2, log ꢀ2; UO2 + 3NO3 ) UO2(NO3)3-, log ꢀ3.
2+
-
2+
-
2+
-
b Twenty-nine experimental data points for each wavelength. c Hamilton’s R factor
n
n
2 1/2
HRF ) 100[
(A
- A )2/
(A
) ]
exp,λ,i
∑ ∑
∑ ∑
exp,λ,i
λ,i
λ
i)1
λ
i)1
See text for notation. d Molar absorbance: ε1 (UO22+), ε2 [UO2(NO3)+], ε3 [UO2(NO3)2], and ε4 [UO2(NO3)3-]. Uncertainties are given as
standard deviations in parentheses.
of ILs, supposing these media are mainly dissociated, the
ionic strength should be in the range of 5-9 M (depending
of the density of the used IL), owing to the cationic and
anionic entities of which they are composed. Therefore, the
complex formation occurring in our work should have a
negligible impact on the total ionic strength as uranium is in
the range of 10-2 M. Under this frame, the mathematical
treatment of the models is feasible (see Data Fitting section).
In a first step, monowavelength fitting was achieved at λ
) 438 nm. Model II was tested first as it contains fewer
parameters than model I. As a rule, the data appeared to be
fitted with similar ꢁ2 values for various sets of parameter
values (ꢀ3, ε0, ε3), where εi is the molar absorption coefficient
of UO2(NO3)i(2-i). This is due to the inherent correlation
Figure 3. Calculated (solid line) versus experimental (symbols) values
of absorbance for the model of formation of three complexes ML+,
between these parameters in a monowavelength fitting. Owing
-
ML2, ML3 (M ) UO22+, L ) NO3-) in the IL C4-mimTf2N. The
to the average uncertainty in the coordination numbers as
CHEMEQUI calculations were performed on an experimental data set
which includes seven data blocks for wavelengths 439, 440, 452, 453,
454, 466, and 468 nm.
-
derived by EXAFS, the presence of UO2(NO3)3 as a major
species above R ) 3 was taken into account by considering
only fitting results for which the percentage of this species
was calculated above 90% for R g 3. Given this condition,
the fit leads to log ꢀ3 ) 10.8 and ε0 ) 6.11 and ε3 ) 23.9
cm-1 M-1 (see Figure 2). In the case of model I, it has also
been checked that mathematical solutions of the fit leading
to [UO2(NO3)3-] > 90% for R g 3 are not the only ones to
be obtained with very similar satisfying ꢁ2 values. Actually,
as the data are rather smooth and the number of (correlated)
parameters is important, this is not surprising. As an
illustration, Figure 2 displays a fit of model II, for which
times until a user-defined value of change of the Hamilton’s R
factor (HRF)
n
n
2 1/2
HRF ) 100[
(A
- A )2/
(A
) ]
∑ ∑
∑ ∑
exp,λ,i
λ,i
exp,λ,i
λ
i)1
λ
i)1
Taking into account the EXAFS results, the model of
-
formation of three complexes ML+, ML2, and ML3 (M )
-
UO2(NO3)3 is equal to 87.5% of the total uranium concen-
UO22+, L ) NO3-) was analyzed. The stability constant
calculations were performed on a reduced experimental data
set, which included seven data blocks for wavelengths at 439,
440, 452, 453, 454, 466, and 468 nm. For each wavelength, 29
experimental data points were used. The selected bands in the
absorption spectra correspond to the three sharp peaks which
are most dependent on the reagent concentrations (Figure 1)
and ensure minimal correlations between the estimated param-
eters. According to the Hamilton’s R factor (HRF ) 2.56%;
Table 1), we have found reasonable agreement between the
calculated and experimental values of absorbances (Figure 3)
for the model. The calculated stability constants (logꢀ1 ) 4.81,
logꢀ2 ) 8.31 and logꢀ3 ) 12.17; logK1 ) 4.81, logK2 ) 3.50
tration at R ) 3 (log ꢀ1 ) 4.87, log ꢀ2 ) 8.0, and log ꢀ3 )
11.8; log K1 ) 4.87, log K2 ) 3.13, and log K3 ) 3.8).
In a second step in a multiwavelength computation procedure,
the CHEMEQUI program32 was used. Similarly, the main
difficulty in the calculations of stability constants and molar
absorbance values derived from the measured spectral data is
the high correlation coefficients between the desired parameters,
which lead to a shift of estimated parameters. To reduce the
shifts of the parameters, a special module for their computations
was developed and complemented in the CHEMEQUI program.
In this procedure, only one parameter is fitted at each step. The
procedure has systematically produced a cyclic optimization of
fitting parameters one at a time. This procedure is repeated many

2+
-
Complexation of UO2 with NO3 in C4-mimTf2N
J. Phys. Chem. B, Vol. 114, No. 12, 2010 4281
values of the stability constants derived in this work are much
higher than the values obtained in water for the same system
(log ꢀ1 ) 0.30 and log ꢀ2 ) 0.3142), while they are closer to
those derived in acetonitrile22 (log ꢀ1 ) 7.9, log ꢀ2 ) 15.0, and
log ꢀ3 ) 20.0), but no direct correlation can be found with the
dielectric constants or polarization abilities of these solvents.47
Nitrate ions appear to be strong complexing agents of uranyl
in C4-mimTf2N, a situation wich strongly ressembles that of
chloride ions toward transition-metal complexes, as evidenced
by Dyson and co-workers.48 Obviously, ILs induce very specific
interactions favoring species with the highest possible stoichi-
ometry. The successive complexation occurring in the system
under study goes from a positively charged species, UO22+, to
a negatively charged species, UO2(NO3)3-, both embedded in
an undefined number of solvation spheres of opposite sign at
each rank. As we demonstrated for the first time that the
determination of stability constants in ILs is feasible, we are
now facing the challenging understanding of the (possibly
mainly electrostatic) force fields that allow for such an efficient
(i.e., with high stability constant values) complexation process
in highly viscous liquids.
Figure 4. Percentage of the uranyl species as a function of the molar
2+
-
ratio R of the reagents for the complexation of UO2 with NO3 in
C4-mimTf2N. Data calculated by CHEMEQUI: UO22+ (1), UO2(NO3)+
-
2+
(2), UO2(NO3)2 (3), and UO2(NO3)3 (4); [UO2
]
≈ 0.01 M.
total
Conclusion
We have performed a comprehensive UV-vis study of the
-
system UO22+/NO3 in the ionic liquid C4-mimTf2N. The
conditional stability constants of the three successive complexes
have been calculated and the individual absorption spectra
derived for the first time in ILs. The use of dedicated salts (C4-
mimNO3 and UO2(Tf2N)2) allowed us to work without spectator
ions besides those composing the IL solvent itself, a situation
unusual to molecular solvents. These results demonstrate that
chemistry in ILs has reached a new step. The question of residual
water content and impurities that were still problematic just a
few years ago are now easily controlled, and it is possible to
obtain reproducible and reliable data for rather intricate com-
plexation chemistry pathways. With these data in hand and those
for other systems to be studied, theoreticians will soon have
the possibility to explore fundamental aspects of mass action
law and activity coefficients in ILs.
Figure 5. Calculated absorption spectra of the uranyl-containing
2+
species for a constant concentration of 0.01 M in C4-mimTf2N: UO2
(1), UO2(NO3)+ (2), UO2(NO3)2 (3), and UO2(NO3)3 (4).
-
+
and logK3 ) 3.86) of formation of the species UO2NO3
,
UO2(NO3)2 and UO2(NO3)3- in C4-mimTf2N at 18.5 °C and the
molar absorbances are presented in Table 1. The values of the
stability constants as assessed by CHEMEQUI conform to the
monowavelength fitting results, and the speciation (in percent-
age) of the uranyl species as a function of the molar ratio R of
Acknowledgment. This work has been performed under the
financial support of “Groupement National de Recherche”
PARIS, the European Network of Excellence (NoE) ACTINET,
and the CNRS program “Chimie Pour un De´veloppement
Durable” (RdR3).
the reagents is in a good agreement with the EXAFS results
2+
(Figure 4). Thus, for a total concentration of [UO2
]total ) 0.01
M and a molar ratio of R ) 1, the reagents form mostly
UO2NO3+, whereas at R g 3, the formation of the species
-
UO2(NO3)3 becomes dominant (Figure 4).
The values of the stability constants (Table 1) were utilized
to calculate the molar absorbances of the uranyl-containing
species and to derive their individual absorption spectra, using
3200 experimental data points including data blocks for 100
wavelengths from 384 to 485 nm. The calculated absorption
References and Notes
(1) Wasserscheid, P.; Welton, T. Ionic liquids in synthesis; Wiley-VCH:
Weinheim, Germany, 2008; Vol. 1 and 2.
(2) Buzzeo, C. M.; Evans, G. R.; Compton, G. R. ChemPhysChem
2004, 5, 1106.
-
spectra of the species UO2(NO3)2 and UO2(NO3)3 have the
(3) Dietz, L. M. Sep. Sci. Technol. 2006, 41, 2047.
(4) Binnemans, K. Chem. ReV. 2007, 107, 2592.
(5) Hao, J.; Zemb, T. Curr. Opin. Colloid Interface Sci. 2007, 12, 129.
(6) Wilkes, S. J. J. Mol. Catal. A: Chem. 2004, 214, 11.
(7) Cocalia, A. V.; Gutowski, E. K.; Rogers, D. R. Coord. Chem. ReV.
2006, 250, 755.
(8) Chaumont, A.; Wipff, G. Phys. Chem. Chem. Phys. 2003, 5, 3481.
(9) Schurhammer, R.; Wipff, G. J. Phys. Chem. B 2007, 111, 4659.
(10) Billard, I.; Gaillard, C. Radiochim. Acta 2009, 97, 355.
(11) Simonin, P. J.; Billard, I.; Hendrawan, H.; Bernard, O.; Lu¨tzen-
kirchen, K.; Se´mon, L. Phys. Chem. Chem. Phys. 2003, 5, 520.
(12) Abovsky, V.; Liu, Y.; Watanasiri, S. Fluid Phase Equilib. 1998,
150-151, 277.
(13) Felmy, R. A.; Rai, D. J. Solution Chem. 1999, 28, 533.
(14) Billard, I.; Mekki, S.; Gaillard, C.; Hesemann, P.; Moutiers, G.;
Mariet, C.; Labet, A.; Bu¨nzli, G., J. C. Eur. J. Inorg. Chem. 2004, 1190.
sharpest maxima at 438, 452, and 467 nm, whereas the medial
size maxima of UO2(NO3)+ are shifted to the long-wave side,
up to 3-4 nm (Figure 5), which induces a shift and an
enlargement of the peaks as this species is not negligible in the
speciation.
Although these stability constants are conditional stability
constants, as explained above, it would be possible to derive
real thermodynamical stability constants provided the exact
mathematical expression of mass action law in ILs is derived.
Actually, the use of C4-mimNO3 and UO2(Tf2N)2 for the
introduction of the reagents suppresses the presence of spectator
ions, which is unavoidable in traditional molecular solvents. The

4282 J. Phys. Chem. B, Vol. 114, No. 12, 2010
Georg et al.
(15) Mekki, S.; Wai, M. C.; Billard, I.; Moutiers, G.; Burt, J.; Yen, H. C.;
Wang, S. J.; Gaillard, C.; Ouadi, A.; Hesemann, P. Chem.sEur. J. 2006,
12, 1760.
(16) Gaillard, C.; Billard, I.; Chaumont, A.; Mekki, S.; Ouadi, A.;
Denecke, M.; Moutiers, G.; Wipff, G. Inorg. Chem. 2005, 44, 8355.
(17) Ouadi, A.; Klimchuk, O.; Gaillard, C.; Billard, I. Green Chem. 2007,
9, 1160.
(31) Ouadi, A.; Gadenne, B.; Hesemann, P.; Moreau, E. J. J.; Billard,
I.; Gaillard, C.; Mekki, S.; Moutiers, G. Chem.sEur. J. 2006, 12, 3074.
(32) Solov’ev, V.; Baulin, E. V.; Strakhova, N. N.; Kazachenko, P. V.;
Belsky, K. V.; Varnek, A.; Volkova, A. T.; Wipff, G. J. Chem. Soc., Perkin
Trans. 1998, 1489.
(33) Brinkley, R. S. J. Chem. Phys. 1947, 15, 107.
(34) Hartley, R. F.; Burgess, C. Solution equilibria; Ellis Horwood
Limited: Chichester, U.K., 1980.
(35) Seber; A. G. Linear hypothesis: a general theory; Hafner: New
York, 1966.
(36) Novikov, P. V.; Ignateva, I. T.; Raevskii, A. O. Zh. Neorg. Khim.
1986, 31, 1474.
(37) Novikov, P. V.; Raevskii, A. O. IzV. Akad. Nauk. 1983, 1336.
(38) Himmelblau, D. M. Applied non linear programming; McGraw-
Hill Book Company: Austin, TX, 1972.
(39) Conley, W. Int. J. Math. Educ. Technol. 1981, 12, 609.
(40) Houwer, D. S.; Go¨rller-Walrand, C. J. Alloys Compd. 2001, 323/
324, 683.
(41) Ikeda-Ohno, A.; Hennig, C.; Tsushima, S.; Scheinost, C. A.;
Bernhard, G.; Yaita, T. Inorg. Chem. 2009, 48, 7201.
(42) Ruas, A.; Bernard, O.; Caniffi, B.; Simonin, P. J.; Turq, P.; Blum,
L.; Moisy, P. J. Phys. Chem. B 2006, 110, 3435.
(43) Nguyen-Trung, C.; Begum, M. G.; Palmer, A. D. Inorg. Chem.
1992, 31, 5280.
(18) Stumpf, S.; Billard, I.; Gaillard, C.; Panak, P.; Dardenne, K.
Radiochim. Acta 2008, 96, 1.
(19) Go¨rller-Walrand, C.; De Jaegere, S. J. Chim. Phys. 1972, 726.
(20) Azenha, E. M.; Burrows, D. H.; Formosinho, J. S.; Leitao, P. M. L.;
Miguel, G. M. J. Chem. Soc., Dalton Trans. 1988, 2893.
(21) Suleimenov, M. O.; Seward, M. T.; Hovey, K. J. J. Solution Chem.
2007, 36, 1093.
(22) Ikeda, A.; Hennig, C.; Rossberg, A.; Tsushima, S.; Scheinost, C. A.;
Bernhard, G. Anal. Chem. 2008, 80, 1102.
(23) Billard, I.; Gaillard, C.; Hennig, C. Dalton Trans. 2007, 4214.
(24) Gaillard, C.; Chaumont, A.; Billard, I.; Hennig, C.; Ouadi, A.;
Wipff, G. Inorg. Chem. 2007, 46, 4815.
(25) Servaes, K.; Hennig, C.; Billard, I.; Gaillard, C.; Binnemans, K.;
Van Deun, R.; Go¨rller-Walrand, C. Eur. J. Inorg. Chem. 2007, 5120.
(26) Nockemann, P.; Servaes, K.; Van Deun, R.; Van Hecke, K.; Van
Meervelt, L.; Binnemans, K.; Go¨rller-Walrand, C. Inorg. Chem. 2007, 46,
11335.
(27) Picquet, M.; Poinsot, D.; Stutzmann, S.; Tkatchenko, I.; Tommasi,
I.; Wasserscheid, P.; Zimmermann, J. Topics Catal. 2004, 29, 139.
(28) Smiglak, M.; Holbrey, D. J.; Griffin, T. S.; Reichert, M. W.;
Swatloski, P. R.; Katritsky, R. A.; Yang, H.; Zhang, D.; Kirichenko, K.;
Rogers, D. R. Green Chem. 2007, 9.
(44) Couston, L.; Pouyat, D.; Moulin, C.; Decambox, P. Appl. Spectrosc.
1995, 49, 349.
(45) Bradley, E. A.; Hardacre, C.; Nieuwenhuyzen, M.; Pitner, R. W.;
Sanders, D.; Seddon, R. K.; Thied, C. R. Inorg. Chem. 2004, 43, 2503.
(46) Gaillard, C.; Chaumont, A.; Billard, I.; Hennig, C.; Ouadi, A.;
Georg, S.; Wippf, G. 2009, submitted.
(47) Reichardt, C. Green Chem. 2005, 7, 339.
(48) Daguenet, C.; Dyson, J. P. Organometallics 2004, 23, 6080.
(29) Billard, I.; Georg, S. HelV. Chim. Acta 2009, 92, 2227.
(30) Adams, J. M. Chemometrics in Analytical Spectroscopy; The Royal
Society of Chemistry: Cambridge, U.K., 1995.
JP9107624