Direct conversion of uranium dioxide UO2 to uranium tetrafluoride UF4 using the fluorinated ionic liquid [Bmim][PF6]

Article abstract of DOI:10.1039/c9dt04327f

The industrial fluorination of UO₂ to UF₄ is based on a complex process involving the manipulation of a large amount of HF, a very toxic and corrosive gas. We present here a safer way to accomplish this reaction utilizing ionic liqui

Full text of DOI:10.1039/c9dt04327f

Dalton
Transactions
View Article Online
View Journal | View Issue
COMMUNICATION
Direct conversion of uranium dioxide UO2 to
uranium tetrafluoride UF4 using the fluorinated
ionic liquid [Bmim][PF6]†
Cite this: Dalton Trans., 2020, 49,
274
Received 8th November 2019,
Accepted 26th November 2019
Florian Joly,a Pardis Simon,a Xavier Trivelli,b Mehdi Arab,c Bertrand Morel,c
Pier Lorenzo Solari,d Jean-Francois Paul,a Philippe Moisye and
DOI: 10.1039/c9dt04327f
a,f
Christophe Volkringer
*
rsc.li/dalton
The industrial fluorination of UO2 to UF4 is based on a complex lated clusters in which U(IV) is surrounded by fluorides. So far,
process involving the manipulation of a large amount of HF, a very we have just counted one monomer ([UF6]2− 3
and two
)
toxic and corrosive gas. We present here a safer way to accomplish different dimers ([U2F13]5− and [U2F14]6−).4,5 These numbers
this reaction utilizing ionic liquid [Bmim][PF6] as a unique reaction are very poor compared to those of the numerous molecular
medium and fluoride source.
structures involving An(IV) and O-donor ligands.6
Other alternative approaches have already been proposed in
the literature to synthesize new U-fluorides, leading mostly to
the production of UF4 or UF6. For example, nitrogen trifluoride
(NF3) is a non-corrosive and stable gas, which has demon-
strated good efficiency for the fluorination of uranium and
other actinides.7,8 Ammonium bifluoride (NH4HF2) is a stable
salt melting at 126 °C. Its dissociation above 239 °C produces
especially gaseous HF that can be exploited for the fluorina-
tion of actinides at a relatively low temperature.9,10
Uranium is the main component of nuclear fuel and is used
for the production of energy.1 After its extraction from natural
ores and purification, this element is transformed into stable
oxides UO3 or U3O8, before further operations. To produce
nuclear fuel, these oxides are usually reduced in order to generate
UO2. This compound is well adapted to the fluorination of
uranium during a process called conversion.2 During this essen-
tial step, uranium is first precipitated as uranium tetrafluoride
UF4 by hydrofluorination. This solid can be stored under anhy-
drous conditions or converted into gaseous UF6. This last species
is the main reagent for uranium enrichment and is another
important intermediate for the fabrication of nuclear fuel.
The industrial fluorination of uranium dioxide into UF4
involves the manipulation of hydrogen fluoride HF at high
temperature. The very high toxicity of this gas requires precau-
tions in response to environmental and safety concerns. In
addition to UF4, the literature mentions other examples of
structures involving multidimensional (1, 2 or 3D) U(IV)-fluor-
ide sub-networks. However, very few examples deal with iso-
Thanks to their good stability, very low vapor pressure and
non-volatile nature, ionic liquids (ILs) can be considered as a
very safe solvent. These room-temperature molten salts have
already proved their efficiency for potential applications in
actinide chemistry11 like metal extraction,12 crystal engineer-
ing13 and solubilization of metal oxides.14 In the last example,
the solubilization of uranium oxides is facilitated by ILs
bearing a complexing chemical function (e.g. –COOH),15 by the
addition of nitric acid16 or using HF-based ionic liquids.17 On
the other hand, ILs can also be used as a fluoride source and a
solvent for the precipitation of fluorides.18 In this case, the
−
−
decomposition of a fluorinated anion (e.g. BF4 or PF6
)
aUnité de Catalyse et Chimie du Solide (UCCS), UMR CNRS 8181, Université de
Lille, ENSC-Lille, Bat. C7, Avenue Mendeleïev, 59655 Villeneuve d’Ascq, France.
E-mail: christophe.volkringer@ensc-lille.fr; Tel: +33 320 434 973
bUniversité de Lille, CNRS, UMR 8576, UGSF, Unité de Glycobiologie Structurale et
Fonctionnelle, Lille, F-59000, France
liberates a fluoride anion that is able to react with a solubilized
metallic cation. So far, this strategy has been established from
the salts of alkali metals,19 alkaline earth metals,20 transition
metals,21,22 lanthanides,23,24 and actinides,25 which are quite
soluble in ILs. In the last example involving An(IV) hexachloro
complexes, the authors proposed that water could play a cata-
lytic role in the decomposition of a hexafluorophosphate anion.
In this study, a fluorinated IL is used in a one-step reaction
for the dissolution of uranium oxide and its conversion into a
pure crystallized anhydrous uranium tetrafluoride (Fig. 1).
This approach is validated under ionothermal reaction con-
ditions with the transformation of UO2 into UF4.
cHall de recherche de Pierrelatte, Site ORANO du Tricastin, Route du Site de
Tricastin, 26701 Pierrelatte, France
dSynchrotron SOLEIL, L’Orme des Merisiers Saint Aubin, BP 48, F-91192 Gif sur
Yvette Cedex, France
eCEA, CEA, DEN, DMRC, Univ Montpellier, Marcoule, France
fInstitut Universitaire de France (IUF), 1 rue Descartes, 756231 Paris Cedex 05,
France
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
C9DT04327F
274 | Dalton Trans., 2020, 49, 274–278
This journal is © The Royal Society of Chemistry 2020

View Article Online
Dalton Transactions
Communication
ment, since no traces of UO2 are detected in the final collected
solid by powder X-ray diffraction. The same synthesis leading
to pure UF4 can be accelerated with a reaction time of 2 h
under microwave heating (150 °C, 1200 W) (Fig. 2). The solid
that crystallizes as a green microcrystalline powder under both
conventional and microwave heating conditions is character-
ized by BET surface areas (Kr, 77 K) of 10 and 6 m2/ g−1
respectively.
,
Fig. 1 Scheme of the one-pot conversion of UO2 into UF4 under
ionothermal conditions using the [Bmim][PF6] ionic liquid both as a
solvent and fluoride source.
Furthermore, inductively coupled plasma optical emission
spectroscopy (ICP OES) analysis (Fig. S2†) reveals traces of
uranium in solution (up to 8.6 mmol L−1 after 96 h, conven-
tional heating), leading to the precipitation of UF4 in a yield of
89% (based on U). This very low value does not allow the
identification of the uranium oxidation state by standard UV-
vis spectrophotometry (Fig. S3†).
Based on these observations, we supposed that the precipi-
tation of UF4 occurs on the surface of UO2 crystallites. Indeed,
the amount of solid UF4 increases gradually (Fig. S1†) in the
medium (simultaneously with the gradual disappearance of
UO2) and is not preceded in solution by a peak of concen-
tration of solubilized uranium (Fig. S3†). However, parallel to
the crystallization of UF4, we note a linear increase in the
quantity of uranium dissolved in the IL. Such behaviour is
attributed to the slow dissolution of UF4 in the IL.
X-ray Photoelectron Spectroscopy (XPS) was used to charac-
terize the oxidation state of solubilized uranium in the IL. The
F 1s core level spectrum (Fig. 3(a)) confirms the presence of
uranium fluoride species, as two peaks are found in the spec-
trum. The intense peak with binding energy (BE) 686.7 eV
corresponds to the fluorophosphate contribution from the
IL,26 whereas the minor peak with BE 684.5 eV is consistent
with uranium fluoride species.27 The uranium U 4f core level
spectrum shows one doublet peak, corresponding to the U
4f7/2 and U 4f5/2 half orbitals with BEs 382.2 eV and 393.0 eV
respectively and a full width at half maximum (FWHM) of
1.5 eV (Fig. 3(b)). An intense peak localized at 402.1 eV is also
By analogy with the industrial conversion process, we
selected uranium dioxide as the metal source for the pro-
duction of anhydrous uranium fluoride. The dissociation of
uranium fluoride in aqueous media requires the use of a
hydrophobic IL in order to avoid contact between water and
the crystalline uranium-based phase.
Among various commercial fluorinated ILs, we selected
1-butyl-3-methylimidazolium hexafluorophosphate, abbre-
viated as [Bmim][PF6]. To the best of our knowledge,
[Bmim][PF6] has never been exploited for the solubilization or
the fluorination of metal oxides.
In this study, black powder of UO2 (40 mg, 0.14 mmol) is
mixed with [Bmim][PF6] (2 ml) and stored in a Teflon lined
reactor. Typically, when the preparation is heated below 150 °C
for 24 h, we did not observe any modification. The transform-
ation of UO2 occurs in relatively short reaction times (<24 h)
from 180 °C (Fig. 2). Under these conditions and for increasing
contact times (Fig. S1†), we note the gradual change of the
ionic liquid from colourless to brown. In parallel, the powder
changes from black to green, leading to very small crystallites
(<5 µm). Powder X-ray diffraction analysis indicates that this
transformation is related to the conversion of black UO2 into
anhydrous uranium tetrafluoride UF4 (Fig. 2). The solid crystal-
lizes as a green microcrystalline powder, whatever the source
of heating may be (Fig. S1†). This chemical reaction effectu-
ated in a classic oven is quantitative after 96 h of thermal treat-
Fig. 2 PXRD diagrams of pure UO2, UF4 and the solids obtained from a
mixture of UO2 and [Bmim][PF6], heated at 180 °C in a conventional
oven from 12 h to 96 h or under microwave heating for 2 hours at
150 °C.
Fig. 3 XPS F 1s core level spectra (a) and N 1s core level spectra (b) of
the UF4 synthesised supernatant (180 °C, 24 h).
This journal is © The Royal Society of Chemistry 2020
Dalton Trans., 2020, 49, 274–278 | 275

View Article Online
Communication
Dalton Transactions
present in the spectrum and corresponds to the N 1s orbital
from the imidazolium species in the IL. In addition, two
minor peaks with BEs 389.3 eV and 400.1 eV are found in the
U 4f spectrum. These peaks correspond to the U 4f7/2 and U
4f5/2 satellite orbitals respectively, and are due to the shake-up
or charge transfer processes.27,28 The satellite-primary U 4f7/2
peak binding energy separation is equal to 7.1 eV and the sat-
ellite/primary U 4f7/2 peak intensity ratio is equal to 0.11. The
high U 4f7/2 BE, alongside the satellite-primary BE separation
and intensity ratio, is consistent with the presence of uranium
fluoride species with a 4+ oxidation state.27,29 Besides, the
presence of the U5+ species can be ruled out, as the U 4f7/2
peak from UF5 presents a large FWHM and almost no satellite
peak.
Fig. 4 19F (at 7.0 T) and 31P (at 9.4 T) NMR spectra recorded at 303 K of
neat [Bmim][PF6] (a) and the supernatant collected after a reaction of
24 h at 180 °C in a conventional oven (b).
In order to investigate the modification that is associated
with the ionic liquid, the solution was also analysed by solu-
tion-state NMR spectroscopy. The 13C NMR signature of the
supernatant previously treated at 180 °C for 24 h is comparable
to that of the unreacted IL, proving that the organic cation is
not decomposed during the reaction (Fig. S6†). Only a minor
difference is observed in the ionic liquid containing uranium,
which exhibits a small peak broadening due to the presence of
technique is not adapted for the detection of U(IV) complexes
due to the paramagnetic character of this cation (Fig. 4).
The characterization of the supernatant isolated from the
precipitation of UF4 (180 °C, 24 h) was completed by X-ray
absorption Spectroscopy (XAS). XAS measurements of the
sample were performed at the U-L3 edge. The shape of the
XANES spectrum (Fig. S7†) is typical of the signature of tetrava-
lent uranium in solution. The XANES region has a maximum
(white line) at 17180.8 eV and does not show a shoulder on the
right side of the line, typical of the uranium(VI) samples.31
Therefore, the EXAFS and the corresponding Fourier transform
(FT), as well as the associated fits, are shown in Fig. 5. One can
see qualitatively that the signal is dominated by a single contri-
bution and this is reflected in the FT by a main peak with a
maximum at 1.65 Å. By fitting the signal in the R space with a
single fluorine contribution, it is possible to achieve a good fit
with a U–F distance of 2.06 Å and compatible with five neigh-
paramagnetic U(IV) in solution. On the other hand, 19F and 31
P
−
NMR results confirm that the hexafluorophosphate anion PF6
from [Bmim][PF6] is more fragile under the reaction con-
ditions. After the thermal treatment (24 h, 180 °C), the initial
heptet at δ(31P) = −144.3 ppm (1JPF = 711 Hz) is partially trans-
formed to a triplet δ(31P) = −17.6 ppm (1JPF = 953 Hz), whereas
the 19F signature presents a doublet in both cases centred at
δ(19F) = 71.7 ppm (1JPF = 710 Hz) which shifts to δ(19F) =
83.4 ppm (1JPF = 954 Hz) at the end of the reaction (Fig. 3).
These new signals are consistent with the existence of the
PO2F2 anion,30 a known hydrolysis product of PF6 according
−
−
−
to eqn (1). PO2F2 is accompanied by the production of hydro-
fluoric acid. The initiation of PF6 hydrolysis is guaranteed by
−
bouring atoms like in the UF5 anion. The results are given in
−
Table S1.† Moreover, these distances are in good agreement
the traces of water in the initial IL (1 mmol L−1) and self-pro-
pagated by the generation of water during UF4 formation
(eqn (2)). This mechanism, clearly proved by NMR analysis, is
different from the process usually supposed for the fluorina-
−
tion of cations and based on the decomposition of PF6 into
gaseous phosphorus pentafluoride (PF5) and
anion (F−).19
a
fluoride
PF6ꢀ þ 2H2O ! PO2F2ꢀ þ 4HF
UO2 þ 4HF ! UF4 þ 2H2O
ð1Þ
ð2Þ
Based on the peak intensity (Fig. S4†), the quantity of
−
PO2F2 remains in a minority in solution, compared to the
original PF6− anion. Whereas this reaction induces the appear-
ance of HF, the whole process remains very safe. Indeed, the
quantity of this species must remain very low due to its con-
summation in a reaction cycle requiring its disappearance
(eqn (1) and (2)).
The different NMR spectra (13C, 19F, and 31P) do not show
other peaks that could correspond to the other hydrolysis pro-
Fig. 5 U-L3 edge k3-weighted EXAFS (top) and the corresponding
Fourier transform (bottom) of the supernatant sample. The blue lines
ducts and solubilized U(VI) complexes. Therefore, the NMR represent the experimental spectra and the dots represent the fits.
276 | Dalton Trans., 2020, 49, 274–278
This journal is © The Royal Society of Chemistry 2020

View Article Online
Dalton Transactions
Communication
with those already reported for other U(IV) and U(V)
Notes and references
fluorides.3,32 The existence of UF5 seems to be counterintui-
−
tive due to the very large size of the U(IV) cation, which is used
to adopt coordination numbers greater than or equal to 6.
However, it has to be noted that penta-coordinated uranium
was already mentioned in different complexes.33,34 The stabi-
1 C. R. Edwards and A. J. Oliver, JOM, 2000, 52, 12–20.
2 B. Morel and B. Duperret, J. Fluorine Chem., 2009, 130, 7–10.
3 K. S. Pedersen, K. R. Meihaus, A. Rogalev, F. Wilhelm,
D. Aravena, M. Amoza, E. Ruiz, J. R. Long, J. Bendix and
R. Clerac, Angew. Chem., Int. Ed., 2019, 58, 15650–15654.
4 R. J. Francis, P. S. Halasyamani, J. S. Bee and D. O’Hare,
J. Am. Chem. Soc., 1999, 121, 1609–1610.
5 Y. L. Lai, R. K. Chiang, K. H. Lii and S. L. Wang, Chem.
Mater., 2008, 20, 523–530.
6 K. E. Knope and L. Soderholm, Chem. Rev., 2013, 113, 944–994.
7 A. M. Casella, R. D. Scheele and B. K. McNamara, AIP Adv.,
2015, 5, 127230.
8 B. McNamara, R. Scheele, A. Kozelisky and M. Edwards,
J. Nucl. Mater., 2009, 394, 166–173.
9 J. B. S. Neto, E. F. U. de Carvalho, R. H. L. Garcia,
A. M. Saliba-Silva, H. G. Riella and M. Durazzo, Nucl. Eng.
Technol., 2017, 49, 1711–1716.
10 B. Claux, O. Benes, E. Capelli, P. Soucek and R. Meier,
J. Fluorine Chem., 2016, 183, 10–13.
11 K. Takao, T. J. Bell and Y. Ikeda, Inorg. Chem., 2013, 52,
3459–3472.
−
lization of UF5 in [Bmim][PF6] can be explained by the very
low concentration of this species and H2O, limiting the hydro-
lysis and/or the condensation of the inorganic anion.
Furthermore, an effect from the IL cannot be ruled out.
As mentioned above, the appearance of UF5− is correlated to
the crystallization of UF4. This phenomena can be interpreted
as the slow dissolution of UF4 in the IL, according to eqn (3)
ꢀ
UF4 þ Fꢀ ! UF5
ð3Þ
In conclusion, we have demonstrated for the first time the
direct conversion of UO2 to UF4 under ionothermal reaction
conditions. The chemical mechanism is based on a one pot
reaction process in which a fluorinated ionic liquid plays the
dual role of solvent and fluoride source. 19F and 31P NMR ana-
lyses indicate that the catalytic mechanism of fluorination
−
starts from the hydrolysis of the PF6 anions and generates
HF, which is immediately consumed for the fluorination of
−
12 X. Q. Sun, H. M. Luo and S. Dai, Chem. Rev., 2012, 112,
2100–2128.
13 A.-V. Mudring and S. Tang, Eur. J. Inorg. Chem., 2010, 2569–
2581.
uranium. The decomposition of PF6 in the presence of water
and the metallic cation is also accompanied by the production
of PO2F2−. H2O appears as the driving force of the conversion
−
of UO2 to UF4, since the hydrolysis of PF6 requires water
14 K. Binnemans, Chem. Rev., 2007, 107, 2592–2614.
15 C. J. Rao, K. A. Venkatesan, K. Nagarajan and
T. G. Srinivasan, Radiochim. Acta, 2008, 96, 403–409.
16 I. Billard, C. Gaillard and C. Hennig, Dalton Trans., 2007,
4214–4221.
molecules produced by the reaction between HF and UO2. The
initial hydrolysis of PF6 is permitted by the traces of water in
the initial IL. EXAFS/XANES and XPS spectroscopy analyses
indicate that the precipitation of UF4 is followed by its slow
dissolution as the UF5− anion.
−
17 C. A. Zarzana, G. S. Groenewold, M. T. Benson,
J. E. Delmore, T. Tsuda and R. Hagiwara, J. Am. Soc. Mass
Spectrom., 2018, 29, 1963–1970.
18 A. B. Pereiro, J. M. M. Araujo, S. Martinho, F. Alves, S. Nunes,
A. Matias, C. M. M. Duarte, L. P. N. Rebelo and
I. M. Marrucho, ACS Sustainable Chem. Eng., 2013, 1, 427–439.
19 J. Olchowka, M. Suta and C. Wickleder, Chem. – Eur. J.,
2017, 23, 12092–12095.
This new approach for the synthesis of UF4 is considered as
an interesting and very safe alternative for the usual industrial
process requiring high temperature and a large amount of HF.
Furthermore, it opens up new methods for the stabilization and
the synthesis of molecular U(IV)-fluorides, which can serve as a
building unit for the production of new fluorinated materials.
20 J. S. Xu and Y. J. Zhu, CrystEngComm, 2012, 14, 2630–2634.
21 D. S. Jacob, L. Bitton, J. Grinblat, I. Felner, Y. Koltypin and
A. Gedanken, Chem. Mater., 2006, 18, 3162–3168.
22 C. L. Li, L. Gu, S. Tsukimoto, P. A. van Aken and J. Maier,
Adv. Mater., 2010, 22, 3650–3654.
Conflicts of interest
There are no conflicts to declare.
23 C. Zhang, J. Chen, Y. C. Zhou and D. Q. Li, J. Phys. Chem.
C, 2008, 112, 10083–10088.
Acknowledgements
The authors would like to thank Nora Djelal, Laurence Burylo 24 D. S. Zhang, T. T. Yan, H. R. Li and L. Y. Shi, Microporous
and Philippe Devaux for their technical assistance with the Mesoporous Mater., 2011, 141, 110–118.
SEM image and powder XRD analysis. The Institut 25 S. I. Nikitenko, C. Berthon and P. Moisy, C. R. Chim., 2007,
Universitaire de France (IUF), Chevreul Institute (FR2638), 10, 1122–1127.
Fonds Européen de Développement Régional(FEDER), CNRS, 26 I. J. Villar-Garcia, E. F. Smith, A. W. Taylor, F. L. Qiu,
Région Nord Pas-de-Calais and Ministère de l’Education
Nationale de l’Enseignement Supérieur et de la Recherche are
K. R. J. Lovelock, R. G. Jones and P. Licence, Phys. Chem.
Chem. Phys., 2011, 13, 2797–2808.
acknowledged for funding this work. Dr F. Joly also thanks 27 E. Thibaut, J. P. Boutique, J. J. Verbist, J. C. Levet and
University of Lille and ORANO for his PhD fellowship.
H. Noel, J. Am. Chem. Soc., 1982, 104, 5266–5273.
This journal is © The Royal Society of Chemistry 2020
Dalton Trans., 2020, 49, 274–278 | 277

View Article Online
Communication
Dalton Transactions
28 T. Gouder, R. Eloirdi and R. Caciuffo, Sci. Rep., 2018, 8, 32 C. G. Gianopoulos, V. V. Zhurov, S. G. Minasian,
1–7.
E. R. Batista, C. Jelsch and A. A. Pinkerton, Inorg. Chem.,
2017, 56, 1775–1778.
33 C. J. Burns, D. L. Clark, R. J. Donohoe, P. B. Duval,
B. L. Scott and C. D. Tait, Inorg. Chem., 2000, 39, 5464–
5468.
34 C. A. P. Goodwin, F. Tuna, E. J. L. McInnes, S. T. Liddle,
J. McMaster, I. J. Vitorica-Yrezabal and D. P. Mills, Chem. –
Eur. J., 2014, 20, 14579–14583.
29 E. S. Ilton and P. S. Bagus, Surf. Interface Anal., 2011, 43,
1549–1560.
30 E. N. Keyzer, P. D. Matthews, Z. G. Liu, A. D. Bond,
C. P. Grey and D. S. Wright, Chem. Commun., 2017, 53,
4573–4576.
31 C. Hennig, A. Ikeda-Ohno, F. Emmerling, W. Kraus and
G. Bernhard, Dalton Trans., 2010, 39, 3744–3750.
278 | Dalton Trans., 2020, 49, 274–278
This journal is © The Royal Society of Chemistry 2020