How to explain the difficulties in the coffinite synthesis from the study of uranothorite?

Article abstract of DOI:10.1021/ic2016758

The preparation of Th_1-xU_xSiO₄ uranothorite solid solutions was successfully undertaken under hydrothermal conditions (T = 250 °C). From XRD and EDS characterization, the formation of a complete solid solution between x = 0 (thorite) and x = 0.8 was evidenced. Nevertheless, additional (Th,U)O₂ dioxide and amorphous silica were systematically observed for the highest uranium mole loadings. The influence of kinetics parameters was then studied to avoid the formation of such side products. The variation of the synthesis duration allowed us to point out the initial formation of oxide phases then their evolution to a silicate phase through a dissolution/precipitation process close to that already described as coffinitization. Also, the uranium mole loading initially considered was found to significantly influence the kinetics of reaction, as this latter strongly slows down for x > 0.3. Under these conditions, the difficulties frequently reported in the literature for the synthesis of pure USiO₄ coffinite were assigned to a kinetic hindering associated with the coffinitization reaction.

Full text of DOI:10.1021/ic2016758

ARTICLE
pubs.acs.org/IC
How To Explain the Difficulties in the Coffinite Synthesis from the
Study of Uranothorite?
D. T. Costin,† A. Mesbah,† N. Clavier,*,† N. Dacheux,† C. Poinssot,‡ S. Szenknect,† and J. Ravaux†
†ICSM, UMR 5257 CEA/CNRS/UM2/ENSCM, Site de Marcoule ꢀ B^at. 426, BP 17171, 30207 Bagnols/Cꢀeze cedex, France
‡CEA, Nuclear Energy Division, DEN/DRCP/DIR, CEA Marcoule ꢀ B^at. 400, BP 17171, 30207 Bagnols/Cꢀeze cedex, France
S Supporting Information
b
ABSTRACT: The preparation of Th1ꢀxUxSiO4 uranothorite solid solutions
was successfully undertaken under hydrothermal conditions (T = 250 °C).
From XRD and EDS characterization, the formation of a complete solid
solution between x = 0 (thorite) and x = 0.8 was evidenced. Nevertheless,
additional (Th,U)O2 dioxide and amorphous silica were systematically
observed for the highest uranium mole loadings. The influence of kinetics
parameters was then studied to avoid the formation of such side products. The
variation of the synthesis duration allowed us to point out the initial formation
of oxide phases then their evolution to a silicate phase through a dissolution/
precipitation process close to that already described as coffinitization. Also, the
uranium mole loading initially considered was found to significantly influence
the kinetics of reaction, as this latter strongly slows down for x > 0.3. Under
these conditions, the difficulties frequently reported in the literature for the
synthesis of pure USiO4 coffinite were assigned to a kinetic hindering associated with the coffinitization reaction.
few reliable thermodynamic data related to coffinite formation
or solubility are available in the literature,10ꢀ13 probably due
to the persistent difficulties encountered in the preparation of
pure and single-phase USiO4 samples. Indeed, if a large amount
of the literature was dedicated to the synthesis of isostructural
zircon,14ꢀ16 hafnon,17,18 or thorite,19ꢀ22 only a few authors re-
ported the successful preparation of coffinite.
Fuchs and Hoekstra were the first to report the synthesis of
USiO4 by using a precipitation route under hydrothermal con-
ditions (T = 250 °C, t = 1 day).23 Nevertheless, UO2 or SiO2
were frequently observed as secondary phases, and the pH of the
initial mixture was pointed out as a key parameter. This protocol
was further adapted to reach single-phase compounds, especially
by controlling the surrounding atmosphere to ensure the tetra-
positive oxidation state of uranium.24,25 Nevertheless, all of the
methods proposed appeared hardly reproducible since different
authors failed to obtain coffinite under the same conditions26 or
did not clearly evidence its formation.27,28 Moreover, the pro-
ducts were never pure coffinite but rather a mixture of coffinite
and uraninite.
1. INTRODUCTION
High level radioactive waste storage or a nuclear spent fuel
repository is currently anticipated to be the standard option in
managing the final waste generated in the back-end of the nuclear
fuel cycle. Particularly, an underground repository was defined by
French law as the standard option for confining the high activity
and long-lived fission products and minor actinides elements.1,2
It is thus important to understand and simulate the behavior of
the various radionuclides considered during the evolution of such
a site, in particular after the long-term degradation of various
confinement barriers and subsequent radwaste leaching. Under
these conditions, the kinetics of dissolution as well as the nature
of the phases controlling the concentration of radioactive ele-
ments in solution through thermodynamics equilibria must be
considered carefully.
As many potential host rocks exhibit reducing conditions
associated with a silica-rich environment,3 the concentration of
uranium in the direct environment could be controlled by the
precipitation of uranium(IV) silicate, known as USiO4 coffinite,4
depending on the relative stability of coffinite and uraninite. Such
a reaction, called coffinitization, was already described in the
literature to occur in several natural sites like Oklo reactor
(Gabon)5 or Cigar Lake (Canada):6
On the other hand, several attempts were also made through
dry chemistry processes or solꢀgel chemistry. In the first case, a
mixture of UO2 and SiO2 was encapsulated in a platinum con-
tainer then heated between 250 and 300 °C under a pressure of
50ꢀ100 MPa and using a Si/SiO redox buffer.29 The second
UO2þx þ H4SiO4ðaqÞ a USiO4:2H2O þ 0:5xO2 v
ð1Þ
The natural presence of coffinite was also reported in various
Received: August 3, 2011
Published: September 29, 2011
uranium ore deposits all around the world.7ꢀ9 Nevertheless, very
r
2011 American Chemical Society
11117
dx.doi.org/10.1021/ic2016758 Inorg. Chem. 2011, 50, 11117–11126
|

Inorganic Chemistry
ARTICLE
method was based on the preparation of gelatinous mixtures of
Table 1. Precipitation Yields Determined for Th, U, and Si
from ICP-AES Measurements of the Supernatants
UO2(NO3)2 6H2O, TEOS, and NH4OH.30 These latter were
3
then heated in platinum capsules under hydrothermal conditions
(T = 395 °C, P = 500 bar). For both methods, the temperature
range was limited by the expected decomposition of coffinite into
UO2 and SiO2 above 500ꢀ600 °C.31 Under these conditions, no
reliable proof arguing for the formation of coffinite was provided
by the authors.
x (calcd)
thorium
uranium
silicon
0.3
0.5
0.8
>99.9
>99.9
>99.9
93.0 ( 2.7
95.2 ( 1.7
98.4 ( 1.8
97.8 ( 1.6
95.5 ( 1.7
94.4 ( 1.6
Since the direct determination of thermodynamic data requires
pure and single phase compounds, indirect methods involving
the solubility of Th1ꢀxUxSiO4 samples should be envisaged for
the evaluation of solubility constants of coffinite. Indeed, such
data could be reached by means of a solid solution approach
based on the study of Th1ꢀxUxSiO4 samples.1
The pH was then raised to close to 9 by adding 8 M NaOH, resulting in
the precipitation of a gelatinous phase, then finally buffered to 8.6 ( 0.1
by adding NaHCO3.
The resulting gel probably consisted of (Th,U)-silica colloids.45 It was
poured in a PTFE-lined Parr instruments autoclave (maximum volume
of 23 mL), then heated under hydrothermal conditions at 250 °C for
24 h. The precipitate obtained was separated from the supernatant by
centrifugation, washed three times with deionized water and ethanol,
then finally dried overnight at 60 °C in an oven.
This work was thus devoted to the preparation of synthetic
uranothorite solid solutions as a first step before the acquisition of
thermodynamic data. The formation of Th1ꢀxUxSiO4 uranothor-
ite occurring through direct substitution on the cation site was
often observed in natural samples.9,33ꢀ35 Nevertheless, this sub-
stitution was poorly described in terms of structural modifications
in the literature,36,37 either for synthetic or geological samples.
Associated XRD data remain scarce compared to other systems
involving actinides in zircon-type structures such as (Zr,Th)SiO4,
(Zr,U)SiO4, or (Zr,Pu)SiO4 solid solutions,38ꢀ41 probably due to
the metamict/amorphous state of the uranothorite samples.
A complete Th1ꢀxUxSiO4 series was then prepared through
hydrothermal synthesis then fully characterized from a structural
and microstructural point of view. The formation of a complete
solid solution between thorite and coffinite end members was
then discussed in light of the unit cell parameters determined
from synthetic samples. Moreover, particular attention was paid
to the influence of kinetics on the reaction mechanism in order to
supply possible explanations on the persistent difficulties en-
countered during attempts to prepare pure and single phase
coffinite samples.
ICP-AES analyses were then performed on the supernatants by the
means of a Spectro Arcos EOP device in order to evaluate the thorium,
uranium, and silicon recovery yields during the precipitation. The
apparatus was calibrated with SPEX standard solutions diluted to con-
centrations ranging from 0.1 to 5 mg Lꢀ1. The emission bands
3
considered were located at λ = 279.39, 385.96, and 409.01 nm for
uranium; at 274.72, 283.23, 283.73, and 401.91 nm for thorium; and at
152.67, 212.41, 251.61, and 288.16 nm for silicon.
Initial supernatant and washing solutions were poured together and
adjusted to a global volume of 50 mL by adding deionized water. An
aliquot of 0.5 mL was then removed and diluted to a final volume of
10 mL before ICP-AES measurement. From the results gathered in
Table 1, the precipitation of thorium appeared to be systematically
quantitative since its concentration in the supernatant was always found
below the detection limit of the apparatus (i.e., ∼0.05 ppm). On this
basis, less than 2.10ꢀ4 mmol remained in the supernatant, leading to a
precipitation yield over 99.9%.
On the other hand, uranium precipitation did not appear to be
quantitative, especially for the lower x values, as a probable consequence
of the presence of small amounts of uranium(VI) in the starting solution.
Nevertheless, uranium precipitation yield was found to increase with its
starting amount in solution and reached up to 99.6% for x = 0.8. An
opposite tendency was noted for silicon, which appeared to be under-
incorporated for high uranium amounts. As the cation/SiO4 mole ratio
was not significantly affected by the x value (and kept equal to 1.03), the
precipitation yield seemed to be controlled by the uranium concentra-
tion. Moreover, the lower precipitation yield measured for silicon when
considering the higher x values fitted well with the persistent difficulties
reported in the literature for preparing pure coffinite without any
simultaneous formation of UO2 dioxide as a secondary phase.
2.2. Characterization Techniques. Powder X-Ray Diffraction.
XRD diagrams were obtained by means of a Bruker D8 diffractometer
equipped with a Lynx-eye detector adopting the BraggꢀBrentano
geometry and using Cu Kα1,2 radiation (λ = 1.54184 Å). XRD patterns
were recorded at room temperature in the 5° e 2θ e 100° range, with a
step size of Δ(2θ) = 0.01° and a total counting time of 4 h. Pure silicon
was measured under the same conditions and used as standard to extract
the instrumental function. All powder patterns were refined with the
Rietveld method using the ThomsonꢀCoxꢀHastings Pseudo Voigt
convoluted with an axial divergence asymmetry function46 with the
Fullprof_Suite program.47 On the one hand, eight profile and three
structural parameters were refined for silicate phases, including a
broadening effect by an anisotropic size model, preferred orientation,
and thermal effect. On the other hand, only five profile and two
structural parameters were considered for oxide phases due to their
m3m Laue group.
2. EXPERIMENTAL SECTION
2.1. Chemicals and Synthesis Procedure. All of the reagents
used, including thorium nitrate hydrate, were of analytical grade and
supplied by Sigma-Aldrich, except uranium tetrachloride solutions. The
preparation of uranium tetrachloride was performed by dissolving
uranium metal in hydrochloric acid. The metal pieces were first washed
in 2 M HCl to eliminate eventual oxide traces present at the surface,
rinsed with water and ethanol, and finally dissolved in 6 M HCl. The
uranium concentration of the final solution was estimated to be 0.74 (
0.02 M using the titration method developed by Dacheux et al.42,43 The
thorium chloride concentrated solution was obtained by dissolving
thorium nitrate hydrate (Th(NO3)4 4ꢀ5H2O) in 8 M HCl.44 Several
3
cycles of alternate evaporation and dissolution in 8 M hydrochloric acid
were undertaken until all traces of nitrates were eliminated. The final
thorium concentration was estimated to be 0.55 ( 0.02 M.
Synthesis of Th1ꢀxUxSiO4 solid solutions was carried out by adapting
the protocol already described by Pointeau et al. for the preparation of
coffinite.25 In order to prevent any oxidation of U(IV) to U(VI), the
whole procedure was performed in an inert glovebox flushed by argon
(guaranteeing less than 1 ppm O2) instead of the reducing atmosphere
(N2 ꢀ 5% H2) used by Pointeau et al.25
Stoichiometric amounts of thorium and uranium chloride were first
diluted in 5 mL of deionized water in order to obtain 1 mmol of a starting
solution with the desired x stoichiometric ratio (0 e x e 0.8). This latter
was added dropwise to an aqueous solution containing 1.03 mmol of
Na2SiO3, resulting in a 3% molar excess in silicate compared to cations.
11118
dx.doi.org/10.1021/ic2016758 |Inorg. Chem. 2011, 50, 11117–11126

Inorganic Chemistry
ARTICLE
Figure 1. XRD patterns of uranothorite solid solutions with various
chemical compositions. The main XRD lines of additional Th1ꢀyUyO2
oxide phase are pointed out by asterisks.
Table 2. Results of X-EDS Analyses Obtained for
Th1ꢀxUxSiO4 Uranothorite Solid Solutions
x (calculated) Th (atom %) U (atom %) An(IV)/Si x (experimental)
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
15.1 ( 0.2
13.8 ( 0.2
13.0 ( 0.2
11.3 ( 0.3
9.8 ( 0.6
9.5 ( 0.7
6.4 ( 0.6
1.5 ( 0.1 0.99 ( 0.02
2.8 ( 0.2 1.00 ( 0.04
4.2 ( 0.2 1.07 ( 0.02
5.7 ( 0.5 1.04 ( 0.08
7.0 ( 0.7 1.02 ( 0.05
7.2 ( 0.7 1.01 ( 0.03
10.0 ( 0.7 0.98 ( 0.05
0.10 ( 0.01
0.17 ( 0.07
0.24 ( 0.01
0.33 ( 0.02
0.42 ( 0.03
0.43 ( 0.04
0.61 ( 0.04
Figure 2. Rietveld plot obtained for Th0.83U0.17SiO4 (a) and Th0.39
-
U
0.61SiO4 (b) samples. Solid line stands for calculated data.
No silicate formed after 24 hours
Scanning Electron Microscopy. SEM observations were directly
conducted on small powder samples without prior preparation such as
metallization, using a FEI Quanta 200 electronic microscope, equipped
either with an EverhartꢀThornley Detector (ETD) or a Back-Scaterred
Electron Detector (BSED), under high vacuum conditions with a very
low accelerating voltage (2ꢀ3.1 kV). These conditions were chosen in
order to create a beam deceleration effect that led to high resolution
images.
X-Ray Energy Dispersive Spectroscopy. X-EDS analyses were per-
formed using the Bruker AXS X-Flash 5010 detector coupled to the SEM
device. In order to determine quantitatively the elementary atomic per-
centages as well as the mole ratios, the powders were first embedded in an
epoxy resin. The surface of the sample was then polished to reach an
optical grade and metalized by carbon deposition. Experimental data were
finally collected from 30 different points by considering ThO2, UO2, and
albite (NaAlSi3O8) as standards for quantitative measurements.
Figure 3. Variation of the weight contents of Th1ꢀxUxSiO4 and
Th1ꢀyUyO2 phases versus expected uranium mole loading, xU.
four equal AnꢀO distances. The resulting AnO8 polyhedra share
an edge with the SiO4 tetrahedra to form chains along the c axis.
The second phase fitted well with the fluorite-type structure
(Fm3m space group) associated with thoriumꢀuranium(IV)
mixed dioxides (Th1ꢀyUyO2).51,52 The formation of such a
secondary phase was not surprising since the presence of oxide
phases associated with thoriumꢀuranium(IV) silicates was fre-
quently described in ore deposits53,54 and was one among the
causes of the late discovery of coffinite.8 Nevertheless, the
reflections associated with Th1ꢀyUyO2 appeared wide and/or
weak in our system, which suggested that the dioxide phase was
probably nanosized or poorly crystallized.
3. RESULTS AND DISCUSSION
3.1. Characterization of Uranothorite Solid Solutions. The
qualitative analysis of the XRD patterns (Figure 1) revealed
the formation of a two-phase system. The first phase corre-
sponded to the expected thoriumꢀuranium(IV) mixed silicate
(Th1ꢀxUxSiO4) and exhibited all of the characteristic XRD lines
of the tetragonal I41/amd zircon-type structure previously re-
ported for ThSiO4,48,49 USiO4,25 and other tetravalent actinide
(Pu, Np, Am) silicates.50 In this structural arrangement, the
actinide is surrounded by eight oxygen atoms with two sets of
11119
dx.doi.org/10.1021/ic2016758 |Inorg. Chem. 2011, 50, 11117–11126

Inorganic Chemistry
ARTICLE
Figure 4. SEM micrographs of the systems obtained during the preparation of Th1ꢀxUxSiO4 uranothorite solid solutions.
solutions appeared to efficiently maintain the tetrapositive state
of uranium in the reacting media during the synthesis since no
U(VI) is expected to be incorporated in the Th1ꢀxUxSiO4
phase.58
Finally, it is also important to note that no other silicate-based
phase was detected from the XRD experiments. On the basis of
the stoichiometry of the starting reagents and on the recovery
yields measured after precipitation, one should expect the
formation of a Si-enriched compound to reach the quantitative
precipitation of silicon. Under these conditions, the more likely
option lied in the precipitation of amorphous and/or gelatinous
SiO2 jointly with that of uranothorite and thoriumꢀuranium-
(IV) oxide solid solutions. Indeed, the formation of silica was
already observed previously in mineral assemblies,9 during the
synthesis of various zircon-type compounds, and in the Thꢀ
UꢀSiO4 system.39 The formation of such a phase was checked
during our study through μ-Raman spectroscopy where the
characteristic νS(SiO2) vibration mode59 was observed around
Figure 5. Refined unit cell volumes of Th1ꢀxUxSiO4 (9) and Th1ꢀy
UyO2 (b) phase versus expected x value.
-
480 cmꢀ1
.
All of the data collected (Table 2) were consistent with the
formation of a uranothorite solid solution with average (Th+U)/
SiO4 mole ratio values close to 1. The slight deviations observed
in several samples could be linked to the presence of the
additional oxide-based phase in the analyzed zone or to the small
Also, no XRD line related to uranyl-bearing phases such as
(UO2)2(SiO4) nH2O soddyite,55 [(UO2)8O2(OH)12] 12H2O
3
3
schoepite,56 or Na(UO2)(SiO3OH) 1.5H2O Na-boltwoodite57
3
was detected during our experiments. On this basis, the operat-
ing conditions set up for the preparation of uranothorite solid
11120
dx.doi.org/10.1021/ic2016758 |Inorg. Chem. 2011, 50, 11117–11126

Inorganic Chemistry
ARTICLE
Table 3. Refined Unit Cell Parameters of Uranothorite Solid Solutions and Oxide Side Productsa
Th1ꢀxUxSiO4 silicate
Th1ꢀyUyO2 dioxide
x (calcd)
x (exptl)
a (Å)
c (Å)
V (Å3)
a (Å)
V (Å3)
y (calcd)b
ref
0.0
7.1328(2)
6.3188(2)
321.48(2)
49
52
5.595(1)
175.2(9)
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
1.0
0
7.148(1)
7.119(1)
7.095(1)
7.085(2)
7.069(3)
7.056(1)
7.049(1)
7.028(1)
6.309(1)
6.317(1)
6.320(1)
6.310(1)
6.305(1)
6.299(1)
6.296(1)
6.286(1)
322.8(7)
320.2(6)
317.7(5)
316.8(7)
315.1(6)
313.6 (7)
312.8 (6)
310.5(6)
0.09(1)
0.17(1)
0.24(1)
0.33(2)
0.42(3)
0.43(4)
0.61(4)
5.498(2)
5.494(3)
5.475(4)
5.475(2)
5.471(6)
5.455(1)
166.2(7)
165.8(4)
164.2(8)
164.1(7)
163.8(8)
162.3(3)
0.65
0.68
0.79
0.80
0.82
0.92
6.995(1)
6.262(1)
304.1(6)
25
5.443(1)
161.3(8)
this work
a The rows in italics present the reference values used for oxide and silicate end members. b Interpolated from unit cell volume of ThO2 and UO2.
Figure 7. Evolution of Th1ꢀxUxSiO4 wt % versus heating time obtained
for x = 0.2 (9), 0.3 (b), 0.5 (2), and 0.8 (1) (T = 250 °C). The
maximum amount of uranothorite obtained for each composition is
indicated on the graph.
considering both phases, i.e., thoriumꢀuranium(IV) mixed
silicate and thoriumꢀuranium(IV) mixed dioxide. Under these
conditions, a good agreement was observed between experi-
mental and calculated data whatever the sample prepared:
indeed, both systems composed by one (such as Th0.83U0.17-
SiO4, Figure 2a) or two phases (e.g., Th0.39U0.61SiO4, Figure 2b)
led systematically to RBragg, RF, RP, and RWP reliability factors in
the 5ꢀ8% range (cited agreements were given by Fullprof47 after
the Rietveld refinement: RP and RWP are respectively profile and
weight profile factors; RBragg and RF depend on the intensities and
structure factor). The refinement of XRD patterns also led to
quantification of the amount of the different phases present in the
sample (Figure 3, precise data supplied as Supporting Information).
Three distinct domains were observed when studying the
variation of silicate and dioxide amounts versus the initial
chemical composition. For low uranium mole loadings (0 e
x e 0.3), only thoriumꢀuranium(IV) mixed silicate was de-
tected by XRD, meaning that less than 1 wt % of additional
dioxide phase was present in the sample. From x = 0.3 to 0.7, the
Figure 6. Evolution of the XRD pattern of the systems obtained for x =
0.5 versus the heating time.
size of the particle analyzed, which could lead to small bias in the
measurements.
Furthermore, uranium content was generally found to be
lower than that initially expected in the prepared samples for x >
0.2. This uranium depletion could arise partly from the small frac-
tion of uranyl present in the starting solution, which remained in
the supernatant all during the synthesis. Nevertheless, it mainly
came from the formation of the thoriumꢀuranium(IV) dioxide
secondary phase evidenced from XRD, which became more and
more important with increasing uranium mole loading x.
3.2. Influence of the Uranium Content on the Synthesis of
Uranothorite Solid Solutions. The Rietveld refinement of the
XRD patterns obtained from the sample series was performed by
11121
dx.doi.org/10.1021/ic2016758 |Inorg. Chem. 2011, 50, 11117–11126

Inorganic Chemistry
ARTICLE
synthesis process led to the coexistence of both oxide and silicate
phases. Nevertheless, the amount of thoriumꢀuranium(IV)
dioxide in the mixture appeared to continuously increase with
the uranium content in the starting solution, then became the
major phase for x = 0.7. For higher uranium mole loadings, this
tendency was exhausted, then led to the formation of thoriumꢀ
uranium(IV) dioxide as the only crystalline phase for x = 0.8. As
the recovery yield of uranium was a maximum for x = 0.8, the
uranium precipitation appeared to be more important in the
dioxide phase than in the uranothorite solid solution for our
operating conditions. Moreover, the absence of silicate phases for
higher uranium contents reflected the persisting difficulties in
easily obtaining a pure uranium(IV) silicate phase, as already
mentioned in the literature.
SEM micrographs (Figure 4) confirmed the presence of two
distinct phases in the samples prepared. On the one hand, mixed
silicates consisted of nanometric crystallites assembled into
bipyramidal aggregates, as already observed for natural samples
of uranium- and thorium-bearing zircon.60 This morphology
corresponds to the {111} bipyramid with no proof of the
presence of the {110} form. The size of such aggregates was
found to increase with the uranium content in the sample.
Indeed, statistic measurements obtained on at least 30 grains
led to average sizes increasing from 120 ( 20 nm (x = 0.1) to
330 ( 70 nm (x = 0.5) and then to 450 ( 60 nm (x = 0.7). On
the other hand, the size of the crystallites constituting the grains
(i.e., domain of coherence length), determined during the
Rietveld refinement of XRD data, did not follow the same trend
since it was systematically found between 10 and 30 nm, what-
ever the initial composition considered. Nevertheless, it allowed
for confirmation of the polycrystalline nature of the uranothorite
grains evidenced during these observations.
On the other hand, the secondary thoriumꢀuranium(IV)
dioxide phase appeared as spheres of about 2ꢀ5 nm scattered
onto the silicate crystals. As expected from the XRD data, the
occurrence of thoriumꢀuranium(IV) dioxide nanoparticles in-
creased significantly with the uranium mole content in the
sample to become the only phase obtained for x g 0.8 under
the operating conditions considered, i.e., T = 250 °C and t = 24 h.
Moreover, for this initial composition, nanostructured Th1ꢀyUyO2
particles seemed to be embedded in a gelatinous layer, which
could be associated with the precipitation of amorphous silica,
already suggested from the recovery yields determined through
ICP-AES measurements.
3.3. Variation of the Unit Cell Parameters versus the
Uranium Mole Loading. Correlatively to the quantification of
the phases formed during the hydrothermal process, the Rietveld
refinement of the XRD patterns led to the evaluation of the lattice
parameters of Th1ꢀxUxSiO4 uranothorite compounds (I41/amd
space group). The variation of the unit cell volume as a function
of the uranium mole loading in the silicate phase (Figure 5)
showed a regular decrease according to Vegard’s law, in good
agreement with the ioinic radii of Th4+ and U4+ in the 8-fold
coordination (1.05 Å and 1.00 Å, respectively).61 The linear
variation of the unit cell parameters (Table 3) then attested to the
existence of a complete solid solution between thorite and
coffinite end members. Also, it confirmed the homogeneity in
the chemical composition of the uranothorite samples pre-
pared and precluded the coexistence of multiple silicated
phases. On this basis, the preparation of uranothorite solid
solutions with x > 0.7 should be achievable from a thermodynamic
point of view. The results obtained under our operating conditions
Figure 8. Variation of the heating time required to reach the maximum
amount of Th1ꢀxUxSiO4 solid solution versus xcalcd (blue) and corre-
sponding xexptl/xcalcd ratio (red).
Table 4. Refined Unit Cell Volumes of Uranothorite Solid
Solutions versus Heating Time at 250°C
unit cell volume (Å3)
heating time (hours)
x = 0.2
x = 0.3
x = 0.5
x = 0.8
interpolateda
2
318.0
317.6(2)
316.3
316.9(2)
312.8
307.6
4
317.3(2)
316.8(1)
317.2(1)
317.7(1)
317.7(1)
317.5(1)
317.4(1)
316.9(1)
6
8
315.1(1)
315.4(1)
317.1(3)
313.2(2)
313.5(2)
313.6(1)
313.1(2)
312.6(1)
312.7(1)
16
24
32
48
64
72
168
336
308.7(2)
309.7(3)
309.9(1)
312.3(2)
312.9(1)
a Values interpolated from ThSiO449 and USiO425 unit cell volumes.
(i.e., formation of pure thoriumꢀuranium(IV) dioxide phases for
x g 0.8) then probably resulted from a kinetic hindering in the
precipitation mechanism.
The Rietveld refinement also allowed for a determination of
the unit cell parameter of the Th1ꢀyUyO2 phase (Table 3). As the
existence of a complete solid solution was previously described in
the literature in the ThO2ꢀUO2 binary system,62ꢀ64 the thorium
substitution rate in the dioxide phase (y) was estimated by
interpolation of the data obtained in this work with that of pure
ThO2 and intermediate Th1ꢀyUyO2 solid solutions.52 Since the
unit cell parameters of the pure UO2 end member were reported
to strongly depend on the U/O stoichiometry and on the
refinement method, a homemade reference was prepared by
firing U(C2O4)2 2H2O at T = 500 °C.65,66
3
As expected from the precipitation yields and the stoichiom-
etry of the silicate phase, Th1ꢀyUyO2 solid solution was found to
be strongly uranium-enriched (0.6 e y e 0.8). On this basis, the
quantitative precipitation of uranium for high uranium mole
contents could be linked to the formation of the U-enriched
11122
dx.doi.org/10.1021/ic2016758 |Inorg. Chem. 2011, 50, 11117–11126

Inorganic Chemistry
ARTICLE
Figure 9. SEM observations of the uranothorite samples obtained for x = 0.5 after 8, 16, 64, and 672 h of heating at 250 °C.
oxide phase. The preferential formation of Th1ꢀyUyO2 and SiO2
instead of Th1ꢀxUxSiO4 solid solutions under our operating
conditions for high uranium mole loadings fitted well with the
results obtained for coffinite.67 Also, it could be in favor of some
chemical processes based on the initial formation of Th1ꢀyUyO2
solid solutions then on a coffinitization-like reaction leading from
the Th1ꢀyUyO2 + SiO2 mixture to Th1ꢀxUxSiO4 solid solution as
already described in geological media.68,69
3.4. Understanding of the Uranothorite Kinetics of For-
mation. Since the operating conditions developed for the
synthesis of uranothorite solid solutions failed to prepare pure
and single phase samples for x > 0.3, the kinetics impact was
evaluated by varying the heating duration of the hydrothermal
treatment (T = 250 °C). As an example, the XRD pattern of the
sample prepared for x = 0.5 was recorded for various heating
times ranging from 8 to 672 h (Figure 6). From these data, the
intensity of the XRD lines associated with the oxide phase
(estimated from their height) was found to decrease significantly
when increasing the holding time for such hydrothermal condi-
tions which corresponded to an increase of the silicate amount in
the two-phase systems obtained.
15% after 8 h of heat treatment to about 95% after 672 h. Under
these conditions, the reaction mechanism leading to uranothorite
seems to be initiated by the nucleation of Th1ꢀyUyO2 nanopar-
ticles and of amorphous and/or colloidal silica. The formation of
Th1ꢀxUxSiO4 solid solutions is then obtained through a dissolu-
tion/precipitation process.
On this basis, the mixture of oxide phases initially present in
the reacting media appeared less stable than the Th1ꢀxUxSiO4
solid solution itself under our operating conditions. Such ob-
servation appeared in good agreement with the data calculated by
Fleche.13 It also fitted well with the geological observations which
reported the preferential precipitation of thorium and uranium as
Th1ꢀxUxSiO4 solid solutions in silicate-based media.7,70
Moreover, the variation of the uranothorite weight percent
was found to empirically follow a reverse exponential function of
time, already described in the literature for the precipitation of
silica from silicon oligomers and nanocolloids.71,72 Under these
conditions, the amount of silicate in the system strongly in-
creased during the first 50 h of heating, then reached a plateau. A
similar tendency was observed whatever the chemical composi-
tion of the starting mixture considered (Figure 7). Nevertheless,
the hydrothermal time required to obtain an optimal quantity of
silicate as well as its maximal amount was found to be strongly
dependent on the composition of the starting mixture. Indeed,
This phenomenon was confirmed through the phase quanti-
fication performed by Rietveld refinement (Figure 7): as an
example, for x = 0.5, the uranothorite weight percent varied from
11123
dx.doi.org/10.1021/ic2016758 |Inorg. Chem. 2011, 50, 11117–11126

Inorganic Chemistry
ARTICLE
while Th1ꢀxUxSiO4 solid solutions appeared to be single-phase
after a few hours for 0 e x e 0.3, the formation of the uranothorite
solid solution was only observed after 48 h of heating for x = 0.8,
whileits weight content onlyreached∼25wt%. Underthese latter
conditions, the rapid and preferential formation of the thoriumꢀ
uranium(IV) dioxide phase was also correlated with the decrease
of the silicon recovery yield at t = 24 h (Table 1).
These results suggest that, with the operating conditions
considered:
(i) The formation of USiO4 coffinite should only occur for
very long heating times, i.e., probably more than several
dozen days (Figure 8).
hydrothermal conditions, indicating that uranium(VI) present in
the supernatant did not precipitate under our operating condi-
tions. Nevertheless, the relative amount of uranium finally
incorporated into the uranothorite phase appeared to be barely
constant in the composition range considered, the xexptl/xcalcd
mole ratio being systematically close to 0.8 for x > 0.1 (Figure 8).
The decrease of the U-enriched dioxide phase amount was also
followed during the hydrothermal heat treatment at 250 °C by
SEM observations performed on the system prepared for x = 0.5
(Figure 9). After 8 h of heating, the system appeared mainly
composed of nanometric grains, assigned to the Th1ꢀyUyO2
phase, and to larger aggregates probably formed by amorphous
silica. The Th1ꢀxUxSiO4 grains were thus not clearly observed,
which agreed well with the low mass content determined under
such conditions (about 15 wt %). Conversely, bipyramidal grains
characteristic of zircon-type structures were systematically ob-
served for longer heating times. Moreover, the occurrence of the
Th1ꢀyUyO2 nanoparticles scattered onto the surface of ura-
nothorite grains was found to decrease progressively to become
almost invisible after 672 h. Also, it is worth noting that the grain
size of Th1ꢀxUxSiO4 was not significantly affected by the heating
duration. Indeed, whatever the heating time considered between
16 and 672 h, the bipyramidal grains were found between 100
and 300 nm in length. Simultaneously, no important increase of
the crystallite size was evidenced from Rietveld refinements of
XRD data collected for x = 0.2 (Table 5). Indeed, this latter was
found to increase moderately from about 100 to 170 Å between
2 and 48 h of heat treatment.
(ii) Only limited amounts of coffinite should be obtained,
with the additional major presence of uranium dioxide.
It is also important to note that the chemical composition of the
silicate phase obtained was almost constant whatever the heating
duration considered. Indeed, the variation of the unit cell volume
remained very weak (i.e., below 0.5%) when making the compar-
ison to the reference value determined for t = 24 h (Table 4).
Under these conditions, the total amount of uranium initially
precipitated in the Th1ꢀyUyO2 nanophase did not appear to be
incorporated in the final Th1ꢀxUxSiO4 solid solution. The deple-
tion in uranium should then occur during the dissolution step of
the dioxide through the partial oxidation of U(IV) as a uranyl
molecular ion which was not incorporated into the uranothorite
structure.58 This was correlated to a slight decrease of the uranium
precipitation yield when increasing the heating duration under
Table 5. Evolution of the Refined Crystallite Size versus
Heating Time (T = 250°C) for Th0.8U0.2SiO4
4. CONCLUSION
Th1ꢀxUxSiO4 uranothorite solid solutions were prepared
under hydrothermal conditions for 0 e x e 0.8 (T = 250 °C).
From XRD data, a complete solid solution was formed for x e
0.8 while additional U-enriched (Th,U)O2 dioxide and amor-
phous silica were systematically detected for the highest uranium
mole loadings. Particular attention was then paid to the influence
of kinetics parameters on the precipitation mechanism. Since the
heating
crystalite
size (Å)
anisotropy
(Å)
heating
crystalite
size (Å)
anisotropy
(Å)
time (h)
time (h)
2
4
6
8
94
128
142
146
(24
(25
(21
(26
16
32
48
134
168
172
(21
(28
(25
Figure 10. Schematic representation of the mechanism proposed for the formation of Th1ꢀxUxSiO4 uranothorite solid solutions.
11124
dx.doi.org/10.1021/ic2016758 |Inorg. Chem. 2011, 50, 11117–11126

Inorganic Chemistry
ARTICLE
amount of silicate compound in the polyphase system was found
to increase with heating time to reach a maximum value
depending on the x mole ratio, the precipitation of oxide phases
was proposed as a first step in the mechanism of formation of
Th1ꢀxUxSiO4 compounds (Figure 10). Their transformation into
uranothorite solid solutions is then expected through a dissolu-
tion/reprecipitation process close to that reported as coffinitiza-
tion. Nevertheless, this reaction was frequently associated with
a loss of about 20% of the initial uranium, probably due to the
partial oxidization of U(IV) into uranyl.
Also, the holding time required to obtain uranothorite strongly
increased with the uranium mole ratio. Consequently, the forma-
tionof coffinite, even if thermodynamically achievable, should only
occur for heating times exceeding several dozens of days. The
difficulties in synthesizing coffinite reported over the decades are
thus probably mainly linked to kinetics constraints associated with
the reaction of coffinitization. Such kinetic hindering then makes
the pure coffinite hardly obtainable on the laboratory time-scale
but allows its observation in natural ores.
(11) Langmuir, D. In Aqueous environmental geochemistry; Prentice
Hall, New Jersey, 1997.
(12) Guillaumont, R.; Fanghanel, T.; Fuger, J.; Grenthe, I.; Neck, V.;
Palmer, D. A.; Rand, M. H. In Update on the chemical thermodynamics of
uranium, neptunium, plutonium, americium and technicium; NEA, OECD:
Moulineaux, France, 2003.
(13) Fleche, J. L. Phys. Rev. B 2002, 65, 245116.
(14) Valꢁero, R.; Durand, B.; Guth, J. L.; Chopin, T. Micropor.
Mesopor. Mater. 1999, 29, 311–318.
(15) Alarcon, J. J. Eur. Ceram. Soc. 2000, 20, 1749–1758.
(16) Veytizou, C.; Quinson, J. F.; Jorand, Y. J. Eur. Ceram. Soc. 2002,
22, 2901–2909.
(17) Fuhrmann, J.; Pickardt, J. Z. Anorg. Allg. Chem. 1986, 532,
171–174.
(18) Kanno, Y. J. Mater. Sci. Lett. 1993, 12, 1807–1809.
(19) Frondel, C.; Collette, R. L. Am. Mineral. 1957, 42, 759–765.
(20) Fuchs, L. H. Am. Mineral. 1958, 43, 367–368.
(21) Sinha, D. P.; Prasad, R. J. Inorg. Nucl. Chem. 1973, 35,
2612–2614.
(22) Kamegashira, N. J. Mater. Sci. 1979, 14, 505–506.
(23) Fuchs, L. H.; Hoekstra, H. R. Am. Mineral. 1959, 44, 1057–1063.
(24) Mulak, J. J. Solid State Chem. 1977, 21, 117–126.
(25) Pointeau, V.; Deditius, A. P.; Miserque, F.; Renock, D.; Becker,
U.; Zhang, J.; Clavier, N.; Dacheux, N.; Poinssot, C.; Ewing, R. C. J. Nucl.
Mater. 2009, 393, 449–458.
(26) Pishva, F. In Synthesis of uranium silicate. Master Degree report,
East Texas State University, Commerce, TX, 1991.
(27) Uziemblo, N. H.; Thomas, L. E.; Schoenlein, L. H.; Mastel, B.;
Jenson, E. D. Mater. Res. Soc. Symp. Proc. 1987, 84, 161–171.
(28) Amme, M.; Renker, B.; Schmid, B.; Feth, M. P.; Bertagnolli, H.;
D€obelin, W. J. Nucl. Mater. 2002, 306, 202–212.
(29) Robit, V. In Etude des phases nꢁeoformꢁees lors de la dissolution du
combustible nuclꢁeaire en condition de stockage gꢁeologique : influence des ions
silicates, PhD Thesis, Universitꢁe Paris-Sud-11, France, 2005.
(30) Zimmer, P. In Etude expꢁerimentale ꢀa huate pression et haute
tempꢁearture du systꢀeme ternaire UO2-ThO2-SiO2 en prꢁesence d’une phase
fluide, CREGU report n 12; CREGU: Nancy, France, 1986.
(31) Lian, J.; Zhang, J. M.; Pointeau, V.; Zhang, F. X.; Lang, M.; Lu,
F. Y.; Poinssot, C.; Ewing, R. C. J. Nucl. Mater. 2009, 393, 481–486.
(32) Vitorge, P. In Law of mass action for co-precipitation, CEA report
no. CEA-R-6193; CEA: Gif-sur-yvette Cedex, France, 2008 (http://
www.vitorge.name/pierre/publis/08vit.pdf; accessed September 2011).
(33) Ifill, R. O.; Cooper, W. C.; Clark, A. H. CIM Bull. 1987, 80, 72–72.
(34) Lira, R.; Ripley, E. M. Contrib. Mineral. Petr. 1992, 110, 370–386.
(35) Sharma, G. S.; Purohit, R. K.; Roy, M.; Sengupta, B.; Singh, J.
J. Geol. Soc. Ind. 2000, 55, 189–196.
As the final aim of this work deals with the determination of
thermodynamic data, the preparation of pure Th1ꢀxUxSiO4 is
now under study. Multiparametric studies of the dissolution
kinetics of Th1ꢀxUxSiO4 solid solutions are also in progress.
’ ASSOCIATED CONTENT
S
Supporting Information. Refined weight amounts for
b
Th1ꢀxUxSiO4 and Th1ꢀyUyO2 phases in the obtained mixtures.
This material is available free of charge via the Internet at http://
pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
*Phone: + 33 4 66 33 92 08. Fax: + 33 4 66 79 76 11. E-mail:
nicolas.clavier@icsm.fr.
’ ACKNOWLEDGMENT
The authors are grateful to the GUTEC research group
(Geology of Uranium and Thorium: Extraction, Conversion) in-
cluded in the PACEN program of CNRS for its financial support.
(36) Pabst, A. Am. Mineral. 1951, 36, 557–562.
(37) Fuchs, L. H.; Gebert, E. Am. Mineral. 1958, 43, 243–248.
(38) Caruba, R.; Baumer, A.; Turco, G. Geochim. Cosmochim. Acta
1975, 39, 11–26.
’ REFERENCES
(39) Ushakov, S. V.; Gong, W.; Yagovkina, M. M.; Helean, K. B.;
Lutze, W.; Ewing, R. C. Ceram. Trans. 1999, 93, 357–363.
(40) Burakov, B. E.; Anderson, E. B.; Zamoryanskaya, M. V.;
Yagovkina, M. A.; Strykanova, E. E.; Nikolaeva, E. V. Mater. Res. Soc.
Symp. Proc. 2000, 663, 307–313.
(41) Geisler, T.; Burakov, B.; Yagovkina, M.; Garbuzov, V.;
Zamoryanskaya, M.; Zirlin, V.; Nikolaeva, L. J. Nucl. Mater. 2005,
336, 22–30.
(42) Dacheux, N.; Brandel, V.; Genet, M. New J. Chem. 1995, 19,
15–25.
(43) Dacheux, N.; Brandel, V.; Genet, M. New J. Chem. 1995, 19,
1029–1036.
(44) Dacheux, N.; Brandel, V.; Genet, M.; Bak, K.; Berthier, C. New
(1) French law no. 2006ꢀ739, 28 June 2006. http://www.assem-
blee-nationale.fr/12/dossiers/gestion_dechets_radioactifs_pro-
gramme.asp (accessed September 2011).
(2) Voizard, P.; Mayer, S.; Ouzounian, G. Proceedings of the 11th
International Conference on Environmental Remediation and Radioactive
Waste Management ICEM2007; ASME: New York, 2009, 125.
(3) Gaucher, E.; Blanc, P.; Bardot, F.; Braibant, G.; Buschaert, S.;
Crouzet, C.; Gautier, A.; Girard, J. P.; Jacquot, E.; Lassin, A.; Negrel, G.;
Tournassat, C.; Vinsot, A.; Altmann, S. C.R. Geosci. 2006, 338, 917–930.
(4) Stieff, L. R.; Stern, T. W.; Sherwood, A. M. Science 1955, 121,
608–609.
(5) Utsunomiya, S.; Deditius, A. P.; Pointeau, V.; Ewing, R. C.
Geochim. Cosmochim. Acta 2008, 72, 968–968.
(6) Janeczek, J.; Ewing, R. C. J. Nucl. Mater. 1992, 190, 157–173.
J. Chem. 1996, 20, 301–310.
(45) Dreissing, I.; Weiss, S.; Henning, C.; Bernhard, G.; Zanker, H.
Geochim. Cosmochim. Acta 2011, 75, 352–367.
(46) Finger, L. W.; Cox, D. E.; Jephcoat, A. P. J. App. Cryst. 1994,
27, 892–900.
(7) Cuney, M. Econ. Geol. 1978, 73, 1567–1610.
(8) Ludwig, K. R.; Grauch, R. I. Econ. Geol. 1980, 75, 296–302.
(9) F€orster, H. J. Lithos 2006, 88, 35–55.
(10) Langmuir, D. Geochim. Cosmochim. Acta 1978, 42, 547–569.
11125
dx.doi.org/10.1021/ic2016758 |Inorg. Chem. 2011, 50, 11117–11126

Inorganic Chemistry
ARTICLE
(47) Roisnel, T.; Rodriguez-Carvajal, J. In WinPLOTR: A windows
tool for powder diffraction pattern analysis, Materials Science Forum,
Trans. Tech. Publications: Switzerland, Vol. 378-381, 2001; pp 118-123.
(48) Taylor, M.; Ewing, R. C. Trans. Am. Geophys. Union 1975, 56,
1076–1076.
(49) Taylor, M.; Ewing, R. C. Acta Crystallogr., Sect. B 1978, 34,
1074–1079.
(50) Speer, J. A. Rev. Miner. 1980, 5, 113–135.
(51) Yang, J. H.; Kang, K. W.; Song, K. W.; Lee, C. B.; Jung, Y. H.
Nucl. Technol. 2004, 147, 113–119.
(52) Hubert, S.; Purans, J.; Heisbourg, G.; Moisy, P.; Dacheux, N.
Inorg. Chem. 2006, 45, 3887–3894.
(53) Fayek, M.; Janeczek, J.; Ewing, R. C. Appl. Geochem. 1997,
12, 549–565.
(54) Fayek, M.; Harrison, T. M.; Ewing, R. C.; Grove, M.; Coath,
C. D. Chem. Geol. 2002, 185, 205–225.
(55) Forbes, T. Z.; Horan, P.; Devine, T.; McInnis, D.; Burns, P. C.
Am. Mineral. 2011, 96, 202–206.
(56) Finch, R. J.; Hawthorne, F. C.; Ewing, R. C. Can. Miner. 1998,
36, 831–845.
(57) Burns, P. C. Can. Miner. 1998, 36, 1069–1075.
(58) Teterin, Y. A.; Utkin, I. O.; Melnikov, I. V.; Lebedev, A. M.;
Teterin, A. Y.; Ivanov, K. E.; Nikitin, A. S.; Vukchevich, L. J. Struct. Chem.
2000, 41, 965–971.
(59) Sidorov, T. A. Russ. J. Inorg. Chem. 2007, 52, 1586–1594.
(60) Reiners, P. W. Rev. Mineral. Geochem. 2005, 58, 151–179.
(61) Shannon, R. D. Acta Crystallogr., Sect. A 1976, 32, 751–767.
(62) Kanno, M.; Kokubo, S.; Furuya, H. J. Nucl. Sci. Technol. 1982,
19, 956–958.
(63) Dash, S.; Parida, S. C.; Singh, Z.; Sen, B. K.; Venugopal, V.
J. Nucl. Mater. 2009, 393, 267–281.
(64) Rousseau, G.; Fattahi, M.; Grambow, B.; Desgranges, L.;
Boucher, F.; Ouvrard, G.; Millot, N.; Niepce, J. C. J. Solid State Chem.
2009, 182, 2591–2597.
(65) Bharadwaj, D. S.; Vasudeva Murthy, A. R. Indian J. Chem. 1964,
12, 391–396.
(66) Duvieubourg-Garela, L.; Vigier, N.; Abraham, F.; Grandjean, S.
J. Solid State Chem. 2008, 181, 1899–1908.
(67) Robit-Pointeau, V.; Poinssot, C.; Vitorge, P.; Grambow, B.;
Cui, D.; Spahiu, K.; Catalette, H. Mater. Res. Soc. Symp. Proc. 2006,
932, 489–492.
(68) Janeczek, J.; Ewing, R. C. Mater. Res. Soc. Symp. Proc. 1992,
257, 497–504.
(69) Philippe, S.; Lancelot, J. R.; Clauer, N.; Pacquet, A. Can. J. Earth
Sci. 1993, 30, 720–730.
(70) Zacek, V.; Skoda, R.; Sulovsky, P. J. Geol. 2009, 54, 343–354.
(71) Icopini, G. A.; Brantley, S. L.; Heaney, P. J. Geochim. Cosmo-
chim. Acta 2005, 69, 293–303.
(72) Conrad, C. F.; Icopini, G. A.; Yasuhara, H.; Bandstra, J. Z.;
Brantley, S. L.; Heaney, P. J. Geochim. Cosmochim. Acta 2007, 71,
531–542.
11126
dx.doi.org/10.1021/ic2016758 |Inorg. Chem. 2011, 50, 11117–11126