Further insights into the chemistry of the Bi–U–O system

Abstract

Cubic fluorite-type phases have been reported in the U^IVO₂–Bi₂O₃ system for the entire compositional range, but an unusual non-linear variation of the lattice parameter with uranium substitution has been observed. In the current extensive investigation of the uranium(IV) oxide–bismuth(III) oxide system, this behaviour of the lattice parameter evolution with composition has been confirmed and its origin identified. Even under inert atmosphere at 800 °C, U^IV oxidises to U^V/U^VI as a function of the substitution degree. Thus, using a combination of three methods (XRD, XANES and Raman) we have identified the formation of the BiU^VO₄ and Bi₂U^VIO₆ compounds, within this series. Moreover, we present here the Rietveld refinement of BiU^VO₄ at room temperature and we report the thermal expansion of both BiU^VO₄ and Bi₂U^VIO₆ compounds.

---

Introduction

The Bi–U–O ternary system came to our attention in the frame of the safety assessment of lead bismuth eutectic (LBE) cooled fast reactors. The LBE cooled fast reactors concept is one of the systems being developed in Europe within the European Sustainable Nuclear Industrial Initiative (ESNII) and is the coolant of choice for the MYRRHA reactor project in Belgium.¹ Thus, the phase relations in the Pb–Bi–U–Pu–O system should be studied in detail, in order to identify possible phases that could be formed during inadvertent fuel (MOX) – coolant (LBE) interactions.²

Very few compounds have been identified in the Bi–U–O ternary system to date. Bismuth(III) monouranate, Bi₂U^VIO₆, is the only compound well characterized from a crystallographic point of view and has been prepared by several high- and low-temperature procedures.^3–7 It was reported to exist in two closely related crystalline modifications. The room temperature form, α-Bi₂UO₆ (space group C2)⁸ undergoes a phase transition above 600 °C into the β-form (P3̄ space group).^8,9 For this compound, the heat capacity was measured and a number of thermodynamic functions have been reported recently.¹⁰

BiU^VO₄ has been prepared and reported to have a defect fluorite structure (Fm3̄m) at room temperature.^11,12 The pentavalent oxidation state of uranium was deduced from the coproportionation reaction between stoichiometric amounts of UO₂, U₃O₈, and Bi₂O₃. Further magnetic susceptibility measurements indicated a paramagnetic behaviour,¹³ consistent with the pentavalent oxidation state of uranium. However, no Rietveld refinement has been published for this compound to date.

Cubic fluorite-type phases have been reported in the system U^IVO₂–Bi₂O₃ for the entire compositional range by Hund.¹⁴ The non-linear variation of the lattice parameters observed in this study was explained based on the fact that the substitution increases the amount of anionic vacancies with increasing Bi₂O₃ content. However, this explanation assumes that the uranium remains in oxidation state IV in the synthesised powders. Thus, we initiated a comprehensive study on the oxidation states in the Bi–U–O system, focused on both uranium and bismuth. We present here the solid state synthesis of different compositions in the Bi–U–O system, the formation of the BiU^VO₄ and Bi₂U^VIO₆ compounds in the reaction conditions, the Rietveld refinement of the BiU^VO₄ and the thermal expansion of BiU^VO₄ and Bi₂U^VIO₆.

Materials and methods

Synthetic procedures

The uranium(IV) oxide–bismuth(III) oxide solid solutions were produced using the solid state reaction procedure previously reported (Hund).¹⁴ Thus, stoichiometric amounts of commercial α-bismuth(III) oxide (Sigma-Aldrich, 99.999% trace metal basis) and UO_2.10 (COGEMA powder, stoichiometry calculated form the value of the lattice parameter) were reacted for 48 h at 800 °C under an argon atmosphere (with an oxygen content of 7 ppm) in specifically designed hermetically closed gloveboxes for handling radioactive materials. Note that the nonstoichiometric uranium oxide was chosen in this experiment in order to reproduce the working conditions presented by Hund.¹⁴ However, the composition having the ratio Bi : U = 1 : 1 and 2 : 1 were also produced starting from fully stoichiometric UO_2.00 under the same reaction conditions.

The α-Bi₂UO₆ used for HT-XRD experiments was prepared by conventional solid state reaction (5 h of at 800 °C under air) of commercial α-bismuth(III) oxide with a stoichiometric amount of amorphous uranium(VI) oxide.

X-ray powder diffraction

α-Bi₂UO₆ was characterised at room temperature by X-ray powder diffraction (XRD) using a Bruker D8 diffractometer mounted in a Bragg-Brentano configuration with a curved Ge(111) monochromator, a ceramic copper tube (40 kV, 40 mA) equipped with a LinxEye position sensitive detector. The data were collected by step scanning in the angle range 10° ≤ 2θ ≤ 120° at a 2θ step size of 0.0092°. For the measurement, the powder was deposited on a silicon wafer to minimize the background and dispersed on the surface with several drops of isopropanol. The structural refinement was performed using the Fullprof2k suite. The shape of the peaks was described by a Pseudo-Voigt function and the background was fitted based on linear interpolation between a set of about 50 background points. The scattering factors of all elements were used.

The thermal expansion of BiUO₄ and α-Bi₂UO₆ was followed by high-temperature X-ray diffraction experiments. The data were collected on a second Bruker D8 X-ray diffractometer mounted with a curved Ge (111) monochromator, a copper ceramic X-ray tube (40 kV, 40 mA), a LinxEye position sensitive detector and equipped with an Anton Paar HTK 2000 chamber. Measurements were conducted up to 600–700 °C under argon, in the angle range 16° ≤ 2θ ≤ 90° with a 2θ step size of 0.017°.

HR-XANES spectroscopy

Uranium M_IV edge. The U M_IV (3728 eV) high energy resolution XANES (HR-XANES) spectra were recorded at the KIT INE-Beamline for actinide research at the ANKA synchrotron radiation facility (Karlsruhe, Germany) on 5 mg of powdered samples mixed with 20 mg of BN. The HR-XANES spectra have been collected using an Johann type multi-analyser X-ray emission spectrometer (MAC-Spectrometer).^15–17 The incident X-ray beam was monochromatized by a Si(111) double crystal monochromator (DCM) and focused to 500 × 500 μm onto the sample. The MAC-Spectrometer comprised five spherically bent Si(220) analyser crystals with 1 m bending radius and a single diode silicon drift detector (KETEK VIRUS SDD). These two components and the sample were arranged in a vertical Rowland circle geometry and placed in a glove box filled with helium. Continuous helium flow has maintained a constant 0.1% O₂ content in the box during the experiments. The DCM was calibrated by setting 3724.5 eV to the maximum of the most intense absorption resonance (white line, WL) of a HR-XANES spectrum pf a UO₂ reference compound. A short HR-XANES spectrum and a normal U M_β (∼3337 eV) emission lines of this reference were measured after each sample. Normal emission lines were also measured before the HR-XANES spectra for each sample in order to monitor the stability of the spectrometer. The U M_IV HR-XANES spectra were obtained by recording the intensity of the U M_β emission line as a function of energy of the incident X-ray beam. The HR-XANES method is also often designated as high energy resolution fluorescence detected XANES (HERFD-XANES) or partial fluorescence yield XANES (PFY-XANES). The presented spectra have been normalized according to the maximum of the white line using the ATHENA software.¹⁸ This method provides unique qualitative informations about the oxidation states present in the systems. The spectra of (U^IV,U^V)₄O₉ and (U^V,U^VI)₃O₈ were collected earlier on the ID26 beamline of the ESRF synchrotron.^19,20 The experimental energy resolution at the INE-Beamline and the ID26 beamline are about 1.3 eV and 0.9 eV, respectively. Bismuth and Uranium L_III edge.

XANES data were recorded at the European Synchrotron Radiation Facility (ESRF, France). 5 mg of powdered sample were mixed with boron nitride (BN) and pressed into pellets for XANES measurements. The storage ring operating conditions were 6.0 GeV and 160–200 mA. The XANES spectra were recorded at room temperature in transmission mode at the ROBL beamline dedicated to actinide elements.²¹ A double crystal monochromator mounted with a Si (111) crystal was used. The energy calibration was achieved using platinum (Pt L_II edge: 13 273 eV) and yttrium (Y K edge: 17 038 eV) foil inserted between the second and third ionization chambers. For each XANES measurement, the spectra of the reference foil were systematically collected at the same time. Background subtraction and normalization have been achieved with ATHENA. To determine the oxidation states of bismuth, XANES spectra edges were compared to those of reference samples, i.e. Bi^III₂O₃, Bi^III₂UO₆ and NaBi^VO₃·xH₂O, that were collected during the same experimental campaign. The oxidation state of uranium, was determined by comparing the absorption edges of the U L_III XANES spectra to those of spectra of reference samples, i.e. U^IVO₂, (U^IV,U^V)₄O₉, (U^V,U^VI)₃O₈ and U^VIO₃. Oxidation states were derived using a linear combination fitting (LCF) of these normalized reference spectra.

Raman spectroscopy

Raman spectra were measured with a Jobin–Yvon® T64000 spectrometer used in the single spectrograph configuration. The excitation source was a Kr⁺ Coherent® continuous wave laser radiating at 647 nm, with a controllable nominal power, usually set to 100 mW at the exit of the cavity. The power impinging the sample surface was lower by approximately a factor 5. Spectra were measured in a confocal microscope with a 50-fold magnification and long focal distance (1 cm). This feature permits a good spectral resolution (±2 cm⁻¹) independently of the surface shape, with a spatial resolution of 2 μm × 2 μm on the sample surface. The spectrograph angle was calibrated with the T_2g excitation of a silicon single crystal, set at 520.5 cm⁻¹.²² The instrument is calibrated on a daily basis prior to measurement. Typical integration times ranged from 20 to 45 seconds with 3 cycles.

Scanning electron microscopy

A Philips XL40 Scanning Electron Microscope equipped with Energy Dispersive X-ray Spectroscopy (EDS) was used. The sample grains were deposited on the usual carbon sticker and covered with Carbon to avoid charging. Each sample was observed at different magnifications and EDS analysis performed to confirm its homogeneity, comparing the elemental composition on the general view with those from smaller regions.

Results and discussion

Elucidation of the non-linear behaviour of the lattice parameters in the uranium oxide–bismuth(III) oxide solid solution

Compositional characterisation. The relative chemical composition of the obtained Bi–U–O samples, as determined by SEM-EDS, is: (1) Bi_0.15U_0.85O_2.00; (2) Bi_0.25U_0.75O_2.00; (3) Bi_0.32U_0.68O_2.00; (4) Bi_0.40U_0.60O_2.00; (5) Bi_0.50U_0.50O_2.00; (6) Bi_0.60U_0.40O_1.95; (7) Bi_0.67U_0.33O_2.00; (8) Bi_0.79U_0.21O_1.81; and (9) Bi_0.85U_0.15O_1.70. The oxygen content was established based on the relative ionic metal composition (as determined by SEM-EDS) and their oxidation states (determined by XANES) as will be explained below (for the samples (1), (3), (5), (6), (7), and (8) – direct measurements, while for the samples (2), (4), and (9) – extrapolated).

XRD study. As already mentioned, cubic fluorite-type phases have been reported in the U^IV system, UO₂–Bi₂O₃, over the entire compositional range by Hund.¹⁴ A non-linear variation of the lattice parameter has been observed, which was explained by the increase of anionic vacancies with increasing Bi₂O₃ content, assuming uranium remained at the oxidation state IV (Hund).¹⁴ In the current investigation of the Bi–U–O system, we have confirmed this specific variation of the lattice parameter with composition, as shown in Fig. 1 and Table 1, suggesting that similar phases as Hund have been obtained.¹⁴ It can be observed that the samples rich in uranium follow a quasilinear trend up to Bi : U = 1 : 1.

Fig. 1 Lattice parameter variation in the Bi–U–O system as a function of the substitution degree. For Bi_0.67U_0.33O_2.00 (●), see discussions. Inset: XRD patterns of the different mixed oxides analysed in this study.

Table 1 Lattice parameters for the Bi–U–O compositions

Chemical composition^a	a, Å	b, Å	c, Å	β, °	V, Å^3
a The O/M ratios have been derived from the U molar fractions and XANES measurements.
Bi0.15U0.85O2.00	5.463 (1)				163.04 (3)
Bi0.25U0.75O2.00	5.465 (1)				163.22 (3)
Bi0.32U0.68O2.00	5.467 (1)				163.40 (3)
Bi0.40U0.60O2.00	5.472 (1)				163.85 (3)
Bi0.50U0.50O2.00	5.478 (1)				164.35 (3)
Bi0.60U0.40O1.95	5.516 (1)				167.83 (3)
Bi0.67U0.33O2.00	6.889 (3)	4.007(4)	9.691 (5)	90.13 (3)	266.96 (5)
Bi0.79U0.21O1.81	5.638 (1)				179.21 (3)
Bi0.85U0.15O1.70	5.612 (1)				176.75 (3)

---

In almost all the compositions, syntheses lead to the formation of well crystalline Bi–U–O solid solutions with a fluorite structure (inset Fig. 1). As this structural description comes from diffraction data, it only describes an “average” of the atomic organisation on a long distance. Being a solid-solution of uranium and bismuth, there is a shared crystallographic position of the cations, i.e. a random distribution of the two elements in space.

Only Bi_0.67U_0.33O_2.00 crystallises in a specific structure with a monoclinic space group. As illustrated in Fig. 2, the former structure is strongly fluorite related. For this reason, despite the poor quality of the refinement in Fm3̄m space group, the deduced lattice parameter obtained with this approach has been introduced in Fig. 1 for comparison with other compositions. Nevertheless, the monoclinic structure obtained for this composition (S.G. C2, a = 6.889(3) Å, b = 4.007(4) Å, c = 9.691(5) Å, β = 90.13(3)°) strongly suggest the formation of the monoclinic form of the U^VI compound α-Bi₂UO₆.^8–10 This hypothesis is at first quite surprising, having in mind the synthesis method (reaction of stoichiometric amounts of commercial α-Bi₂O₃ oxide and UO_2.10 at 800 °C under argon atmosphere). However, the oxidation state VI of uranium has been confirmed with XANES, as it will be presented in the next part. The source of the uranium oxidation are likely small amounts of molecular oxygen contained in the argon atmosphere.

Fig. 2 Rietveld refinement of the sample Bi_0.67U_0.33O_2.00 using a monoclinic cell (S.G. C2) and a cubic one (S.G. Fm3̄m). Despite a better description of the structure in the monoclinic space group, this second refinement highlights that the structure is fluorite related.

XANES measurements. In order to investigate the oxidation states of uranium and bismuth, we performed an extensive Bi L_III and U L_III XANES and U M_IV HR-XANES analysis for seven selected compositions. Unfortunately, the sample with the highest bismuth content could not be measured during the U M_IV HR-XANES experimental campaign.
Valence state of uranium. The U M_IV HR-XANES spectra of different Bi–U–O compositions are presented in Fig. 3 with those of the reference samples, i.e. U^IVO₂, (U^IV,U^V)₄O₉, (U^V,U^VI)₃O₈ and U^VIO₃. These spectra provide unique qualitative informations on the oxidation state in the systems. The uranium oxidation states and derived molar fractions from L_III XANES are provided in the Table 2. These results, together with those obtained from SEM-EDS analysis, were used to calculate the oxygen content within the different Bi–U–O compositions used during this study.

Fig. 3 U M_IV XANES spectra of Bi–U–O compositions.

Table 2 Uranium valence and molar fraction in the Bi–U–O compositions

Chemical composition^a	U^IV	U^V	U^VI	R f
a The O/M ratios have been derived from the U molar fractions in respect of electroneutrality.
Bi0.15U0.85O2.00	82 (5)	18 (5)	0	0.005
Bi0.32U0.68O2.00	44 (5)	56 (5)	0	0.020
Bi0.50U0.50O2.00	0	100	0
Bi0.60U0.40O1.95	0	74 (5)	26 (5)	0.002
Bi0.67U0.33O2.00	0	0	100 (5)	0.006
Bi0.79U0.21O1.81	0	10 (5)	90 (5)	0.003

---

As previously reported^19,20 herein we demonstrate that that the uranium oxidation states can be characterized using the U M_IV edge HR-XANES technique. The HR-XANES spectrum of U^IVO₂ has a pronounced peak at 3724.5 (2) eV, due to the electronic transitions from the 3d_3/2 core level to the unoccupied 5f states. The (U^IV,U^V)₄O₉ HR-XANES spectrum depicts a peak at the same energy position characteristic of U^IV and also a peak at 3725.5 (2) eV which corresponds to U^V. The (U^V,U^VI)₃O₈ HR-XANES spectrum exhibits a white line at 3726.2 (2) eV attributed to U^VI and a shoulder at the lower energy side with energy position characteristic for U^V. The energy positions of the white line of the U^VIO₃ spectrum and the U^VI peak in (U^V,U^VI)₃O₈ coincide (3726.2 (2) eV) and it is characteristic for U^VI compounds.

Except for Bi_0.79U_0.21O_1.81, the white lines of the Bi–U–O compositions shift toward higher energy with increasing the bismuth content. This suggests an overall oxidation of uranium over the compositional range. The XANES spectra of Bi_0.15U_0.85O_2.00 and Bi_0.32U_0.68O_2.00, exhibit the two features characteristics of U^IV and U^V. However, it can be observed that the amplitude of the U^IV peak decreases with increasing the bismuth content.

Regarding the Bi_0.50U_0.50O_2.00 sample, its white line is a symmetric peak well aligned with the U^V peak of (U^V,U^VI)₃O₈. As would be expected from the electroneutrality rule, the least common factor (LCF) confirms that uranium is purely pentavalent for this composition. When increasing the bismuth content, the LCF shows that U^VI is progressively formed in Bi_0.60U_0.40O_1.95 and that a pure U^VI compound is obtained on reaching Bi_0.67U_0.33O_2.00. The formation of U^VI is linked to a strong increase of lattice parameter. However, the white line of Bi_0.79U_0.21O_1.81 shifts toward lower energy indicating the presence of U^V. This is in agreement with the lattice parameter decrease observed by XRD and Raman measurements.

Valence state of bismuth. XANES measurements at the Bi L_III have been performed to elucidate the bismuth valence state. The Fig. 4a and b present the Bi L_III XANES spectra of the Bi–U–O samples.

Fig. 4 (a) Bi L_III XANES spectra of Bi_0.15U_0.85O_2.00, Bi_0.32U_0.68O_2.00, Bi_0.50U_0.50O_2.00, and Bi_0.60U_0.40O_1.95. (b) Bi L_III XANES spectra of Bi_0.60U_0.40O_1.95, Bi_0.67U_0.33O_2.00, Bi_0.79U_0.21O_1.81, and Bi_0.85U_0.15O_1.70.

In the Bi L_III XANES spectra, there is a shift toward higher energies between the Bi^III₂O₃, Bi^III₂UO₆ and NaBi^VO₃·xH₂O white lines passing from Bi^III to Bi^V. Although bismuth is trivalent in both Bi^III₂O₃ and Bi^III₂UO₆, the XANES spectra are different and shifted from one another. This has already been reported²³ for other isovalent compounds and can be understood from the difference of local environment. The NaBi^VO₃·xH₂O spectrum exhibits a pre-edge describing 2p_3/2–6s transitions corresponding to pentavalent bismuth where the 6s levels are unoccupied. This feature is not observed in our compositions, indicating that Bi remains trivalent.

The white lines of the spectra of Bi_0.15U_0.85O_2.00, Bi_0.32U_0.68O_2.00, and Bi_0.50U_0.50O_2.00 compositions have similar energy positions of their absorption edges and overall shapes. This is an indicator that the local structure of these materials is very close. However, the white lines of the remaining samples are totally different and are shifted from one another, suggesting a different local environment for each composition.

Raman spectroscopy

The Raman spectra of the samples are shown in Fig. 5. Samples with high uranium content (up to a molar ratio Bi : U = 1 : 1) reflect very poorly,²⁴ which explains the poor quality of their spectra (dashed black lines in Fig. 5). These spectra are compared with a spectrum from a polished surface of a pure commercial UO₂ pellet, shown in the thick bottom line in Fig. 5. Despite the noise, it can be seen that all these spectra are dominated by a band roughly centred at 445 cm⁻¹, which is assigned to the T_2g anti-symmetric U^IV–O stretching vibration typical of eight-fold coordinated uranium in the fluorite structure. It has been observed many times in UO₂ and isomorphous compounds.^25–30

Fig. 5 Raman spectra of the Bi–U–O compositions.

Factor group theory³¹ indicates that it is the only Raman active mode in the Fm3̄m space group. As Bi^III is dissolved in the UO_2+x lattice, this band broadens and its peak position shifts towards slightly higher energy.

Nonetheless, the fact that the T_2g band dominates the spectrum indicates that a fluorite structure is maintained for Bi_0.15U_0.85O_2.00, Bi_0.25U_0.75O_2.00, Bi_0.32U_0.68O_2.00, and Bi_0.40U_0.60O_2.00 compositions. The broadening of the T_2g mode could arise from the local disorder caused by the Bi^III cations in the fluorite structure. Close to the T_2g band, a few weak bands are observed around 570 cm⁻¹ and 630 cm⁻¹ which likely are linked to oxygen defects in the fluorite structure. Their presence reflects a breakdown of the translational symmetry and a reduction of the point group symmetry. Interestingly, their intensity is independent of the Bi^III concentration, which could suggest that doping with Bi^III has little effect on the oxygen sublattice and that charge compensation is then maintained by uranium atoms for compositions below Bi_0.50U_0.50O_2.00. The band at 630 cm⁻¹ increases for Bi_0.32U_0.68O_2.00 and Bi_0.40U_0.60O_2.00, which is similar to observations for U₄O₉. All of these observations are in agreement with the presence of mixed valence U^IV/U^V in the material and are consistent with XANES data.

At higher bismuth content, a different spectral form is observed (presented in green in Fig. 5), reflecting a complete change of the original fluorite local symmetry, though long range order probed by XRD indicates that a fluorite organisation remains for Bi_0.60U_0.40O_1.95, Bi_0.79U_0.21O_1.81 and Bi_0.85U_0.15O_1.70. These Raman spectra, of much better quality, are largely dominated by a strong band centred around 717 cm⁻¹. The increase in the number of Raman bands indicates a decrease of local symmetry. The band at 445 cm⁻¹ (T_2g) broadens with increasing the bismuth content together with the appearance of weak features at low frequencies ∼210 cm⁻¹ and 280 cm⁻¹ (typically for Bi_0.60U_0.40O_1.95 and Bi_0.67U_0.33O_2.00). These modes (at 210 cm⁻¹ and 280 cm⁻¹) are also present in the Raman spectrum of Bi₂O₃, correspond to the Van Hove singularities at the gamma point of the Brillouin zone and can be assigned to Bi–O stretching vibrations. The broadening of the 445 cm⁻¹ band is a consequence of an increased frequency distribution of the U–O and Bi–O vibrations, very likely due to an increased local disorder in the short range. Therefore, the assignment of this band to pure oxygen motions in a monoclinic or cubic phase is not straightforward. The XRD results suggest that this local disorder is also present in the long range for Bi_0.67U_0.33O_2.00, a monoclinic phase.

Despite the clear distortion at local range and by analogy with the Raman spectra of U₃O₈, UO₃, and UO₂²⁺ ion^32,33 the peak around 717 cm⁻¹ can be attributed to a Raman active U–O stretching vibration of UO8 polyhedra with non-bonding oxygens, in close geometry to the UO₂²⁺ species. This assignment agrees well with the Raman spectrum of the hexavalent uranium compound CaUO₄,³⁴ where this vibration was clearly identified as the ν₁ stretching mode of the uranyl UO₂²⁺ ion.

Combining these observations with the XRD and XANES results, one can draw several conclusions. It can be observed, that the spectra of Bi_0.32U_0.68O_2.00 and Bi_0.50U_0.50O_2.00 displayed in Fig. 6, the T_2g band typical of the fluorite structure and oxygen defect bands between 570 and 630 cm⁻¹.³⁵ The latter lines are attributed to T_1u LO phonons Raman inactive in the absence of oxygen defects. They are not affected by Bi^III doping, suggesting that only the uranium(IV) atoms change to uranium(V) to compensate the electric charge on addition of Bi^III. The T_2g band clearly dominates the spectrum of the uranium-rich samples suggesting that the fluorite structure is maintained, whereas, in the bismuth rich region two observations can be made: (i) the band at 445 cm⁻¹ broadens and its intensity reduces significantly mainly due to the increase of the local disorder at the short range distances and (ii) the dominance of the U–O stretching mode at 717 cm⁻¹ suggesting formation of UO₈ species with non-bonding oxygens (similar to uranyl ions).

Fig. 6 Raman spectra of Bi_0.50U_0.50O_2.00, Bi_0.32U_0.68O_2.00, and Bi_0.67U_0.33O_2.00.

The spectrum of the pure U^VI phase Bi_0.67U_0.33O_2.00 is even more strongly dominated by this latter line in agreement with the layer organisation of α-Bi₂UO₆.

A special case of a U^V phase: BiUO₄

At 50 mol% of UO₂ and Bi₂O₃, a fluorite-type phase of general formula U_0.5Bi_0.5O₂ (or BiUO₄) forms. The cell parameter 5.478(1) Å, is larger than UO₂ (5.471 Å). This is consistent with the average ionic radii of Bi^III_CN8 (1.17 Å) and U^V_CN8 (0.93 Å), with the former being slightly larger than ionic radius of U^IV_CN8 = 1.00 Å. In U_0.5Bi_0.5O₂ the average distance M–O is 2.372 Å. The X-ray diffraction pattern with Rietveld refinement is shown in Fig. 7 and fitting results are provided in Tables 3 and 4. For oxygen atoms the displacement parameter B_iso is high, suggesting local disorder in agreement with the Raman observations.

Fig. 7 Rietveld refinement of BiUO₄. Comparison between the observed (Y_obs, in red) and calculated (Y_calc, in black) X-ray diffraction pattern of BiUO₄ phase. Y_obs–Y_calc (in blue) is the difference between the experimental and calculated intensities. The Bragg reflections are marked in black ticks.

Table 3 Unit cell parameters and results of the Rietveld refinement for BiUO₄

R P = ∑[yi(obs) − yi(calc)]/∑yi(obs); Rwp = {∑wi[yi(obs) − yi(calc)]^2/∑yi(obs)}^1/2. RB = ∑[Ihkl(obs) − Ihkl(calc)]/∑Ihkl(obs); GOF = Rwp/Rexp.
Formula unit	BiUO4
a	5.478 (1) Å
V	164.35 Å^3
Space Group	Fm3̄m (N° = 225)
Z	4
Refined parameters	40
R wp	7.76%
R P	5.58%
R B	2.08%
R exp	3.83%
GOF	2.03

---

Table 4 Atomic positions in the BiUO₄ compound

Atom	Oxidation state	x	y	z	B iso (Å^2)
Bi	3+	0.000	0.000	0.000	0.38
U	5+	0.000	0.000	0.000	0.38
O	2−	0.250	0.250	0.250	2.89

---

The evolution of the unit-cell volume of X^IIIUO₄ phases as a function of eight-fold coordinated cationic radius (Fig. 8) would confirm that BiUO₄ belongs to this family. Note that FeUO₄ and CrUO₄, not shown in this graph, are not cubic.³⁶

Fig. 8 Evolution of the unit-cell volume of X^IIIUO₄ phases as a function of eight-fold coordinated cationic radius. The unit-cell volume values are taken from the literature,^37–44 while for BiUO₄ the present result is shown.

For some of these X^IIIUO₄-type compositions (X^III = Bi, Sc, and Y), the pentavalent oxidation state of uranium was proven based on magnetic susceptibility measurements.^13,45 This might be an indication that all the other X^IIIUO₄ compositions shown in this graph are pure pentavalent uranium phases.

Thermal expansion of BiUO₄ and α-Bi₂UO₆

The relative variation of the cell parameters of BiUO₄ and α-Bi₂UO₆ with temperature is presented in Fig. 9. The transition temperature of α-Bi₂UO₆ (C2 space group) to β-Bi₂UO₆ (P3̄ space group) occurs between 600 and 800 °C, while the cubic BiUO₄ start to decompose into two phases above 700 °C. The expansion is more limited along the a and b directions compared to c, because of the edge-sharing of the UO₈ polyhedra (the thermal expansion along the c axis is more pronounced because of the layer structure of α-Bi₂UO₆). Trivalent bismuth, which is seven-fold coordinated, is poorly bonded to oxygen.

Fig. 9 Relative variation of the cell parameters of BiUO₄ and α-Bi₂UO₆ with temperature.

The linear thermal expansion (LTE) is defined as LTE_{25 °C}(T) = (x(T) − x(25 °C))/x(25 °C). The polynomial interpolation of the experimental data are presented in Fig. 9 and gives for the cubic BiUO₄: LTE_{25 °C}(T) = −3 × 10⁻⁴ + 1.47 × 10⁻⁵T(°C) + 4.1 × 10⁻⁹T²(°C). Similarly, the LTE for the different orientations in Bi₂UO₆ are:

a: LTE_{25 °C}(T) = −1 × 10⁻⁴ + 3.2 × 10⁻⁶T(°C) + 8.8 × 10⁻⁹T²(°C)

b: LTE_{25 °C}(T) = 1 × 10⁻⁴ − 2.0 × 10⁻⁷T(°C) + 8.2 × 10⁻⁹T²(°C)

c: LTE_{25 °C}(T) = −5 × 10⁻⁴ + 1.65 × 10⁻⁵T(°C)

Conclusions

The LBE cooled fast reactor concept is the basis of the MYRRHA reactor project in Belgium. As a part of the safety assessment for such reactors, detailed knowledge of potential lead and bismuth uranates/plutonates phases that could form in the event of a pin breach is needed. The present work completes our study performed in order to identify potential reaction products in the Bi–Pb–U–Pu–O system for the safety assessment for LBE cooled fast reactors.

In the current extensive investigation of the uranium(IV) oxide–bismuth(III) oxide system, the unusual behaviour of the lattice parameter variation with composition as observed by Hund¹⁴ has been confirmed and its origin identified in the oxidation state of uranium. Even under inert atmosphere at 800 °C, U^IV oxidises to U^V/U^VI as a function of the substitution degree. We have identified the formation of the BiU^VO₄ and Bi₂U^VIO₆ compounds, within this series. All the compositions crystallise in a fluorite-related structure, but the actual symmetry of Bi₂U^VIO₆ is monoclinic. Raman spectrometry indicates changes in the local symmetry from six-fold uranium coordination to an eight-fold coordination (deduced as a function of composition) which is not observable in the long range ordering (probed by XRD). EXAFS or neutron diffraction measurements are needed in order to give more details on the local structure in those compositions.

Acknowledgements

This work was performed in the frame of the “Safe exploitation related chemistry for HLM reactors (SEARCH)” Collaborative Project (Contract Number: 295736) in the Seventh Framework Programme of the European Commission. The authors acknowledge Daniel Bouexière and Giorgio Pagliosa for room- and high-temperature X-ray diffraction measurements. We authors acknowledge the ANKA and ESRF synchrotron facilities for provision of beam time. XAS measurements have been supported by the European FP7 TALISMAN project, under contract with the European Commission.

Notes and references

1. J. Engelen, H. Aït Abderrahim, P. Baeten, D. De Bruyn and P. Leysen, Int. J. Hydrogen Energy, 2015, 40, 15137–15147.
2. J. F. Vigier, K. Popa, V. Tyrpekl, S. Gardeur, D. Freis and J. Somers, J. Nucl. Mater., 2015, 467, 840–847.
3. J. G. De Jong and Ph. A. Batist, Requeil, 1971, 90, 749–754.
4. H. Collette, V. Deremince-Mathieu, J. J. Verbist, Z. Gabelica, J. B. Nagy and E. G. Derouane, J. Mol. Catal., 1987, 42, 15–28.
5. J. M. Amarilla, R. M. Rojas and M. P. Herrero, Chem. Mater., 1995, 7, 341–347.
6. D. O. Charkin, D. N. Lebedev, S. Yu. Stefanovich and S. M. Kazakov, Solid State Sci., 2010, 12, 2079–2085.
7. N. L. Misra, A. K. Yadav, S. Dhara, S. K. Mishra, R. Phatak, A. K. Poswal, S. N. Jha, A. K. Sinha and D. Bhattacharyya, Anal. Sci., 2013, 29, 579–584.
8. R. N. Vannier, O. Thery, C. Kinowski, M. Huve, G. van Tendeloo, E. Suard and F. Abraham, J. Mater. Chem., 1999, 9, 435–443.
9. A. S. Kostner, J. P. P. Renaud and G. D. Rieck, Acta Crystallogr., Sect. B: Struct. Crystallogr. Cryst. Chem., 1975, 31, 127–131.
10. K. Popa, O. Beneš, P. E. Raison, J. C. Griveau, P. Pöml, E. Colineau, R. J. M. Konings and J. Somers, J. Nucl. Mater., 2015, 465, 653–656.
11. W. Rüdorff, H. Erfurth and S. Kemmler-Sack, Z. Anorg. Allg. Chem., 1967, 354, 273–238.
12. M. P. Eastman, H. G. Hecht and W. Burton Lewis, J. Chem. Phys., 1971, 54, 4141–4146.
13. C. Miyake, O. Kawasaki, K. J. Gotoh and A. Nakatani, J. Alloys Compd., 1993, 200, 187–190.
14. F. Hund, Z. Annorg. Allg. Chem., 1964, 333, 248–255.
15. J. Rothe, S. Butorin, K. Dardenne, M. A. Denecke, B. Kienzler, M. Löble, V. Metz, A. Seibert, M. Steppert, T. Vitova, C. Walther and H. Geckeis, Rev. Sci. Instrum., 2012, 83, 043105.
16. T. Vitova, M. A. Denecke, J. Göttlicher, K. Jorissen, J. J. Kas, K. Kvashnina, T. Prüβmann, J. J. Rehr and J. Rothe, J. Phys.: Conf. Ser., 2013, 430, 012117.
17. E. Kleymenov, J. van Bokhoven, C. David, P. Glatzel, J. Markus, R. Alonso-Mori, M. Studer, M. Willimann, A. Bergamaschi, B. Henrich and M. Nachtegaal, Rev. Sci. Instrum., 2011, 82, 065107.
18. B. Ravel and M. Newville, J. Synchrotron Radiat., 2005, 12, 537–541.
19. K. O. Kvashnina, S. M. Butorin, P. Martin and P. Glatzel, Phys. Rev. Lett., 2013, 111, 253002.
20. K. O. Kvashnina and F. M. F. de Groot, J. Electron Spectrosc. Relat. Phenom., 2014, 194, 88–93.
21. W. Matz, N. Schell, G. Bernhard, F. Prokert, T. Reich, J. Clauβner, W. Oehme, R. Schlenk, S. Dienel, H. Funke, F. Eichhorn, M. Betyl, D. Pröhl, U. Strauch, G. Hüttig, H. Krug, W. Neumann, V. Brendler, P. Reichel, M. A. Denecke and H. Nitsche, J. Synchrotron Radiat., 1999, 6, 1076–1085.
22. J. H. Parker, Jr., D. W. Feldman and M. Ashkin, Phys. Rev., 1967, 155, 712–714.
23. S. Salem-Sugui Jr., E. E. Alp, S. M. Mini, M. Ramanathan, J. C. Campuzano, G. Jennings, M. Faiz, S. Pei, B. Dabrowski, Y. Zheng, D. R. Richards and D. G. Hinks, Phys. Rev. B: Condens. Matter, 1991, 43, 5511.
24. G. M. Begun, R. G. Haire, W. R. Wilmarth and J. R. Peterson, J. Less-Common Met., 1990, 162, 129–133.
25. V. G. Keramidas and W. B. White, J. Chem. Phys., 1973, 59, 1561–1562.
26. D. Manara and B. Renker, J. Nucl. Mater., 2003, 321, 233–237.
27. T. Livneh and E. Sterer, Phys. Rev. B: Condens. Matter, 2006, 73, 085118.
28. C. Jégou, R. Caraballo, S. Peuget, D. Roudil, L. Desgranges and M. Magnin, J. Nucl. Mater., 2010, 405, 235–243.
29. M. J. Sarsfield, R. J. Taylor, C. Puxley and H. M. Steele, J. Nucl. Mater., 2012, 427, 333–342.
30. R. Böhler, M. J. Welland, D. Prieur, P. Cakir, T. Vitova, T. Pruessmann, I. Pidchenko, C. Hennig, C. Guéneau, R. J. M. Konings and D. Manara, J. Nucl. Mater., 2014, 448, 330–339.
31. T. Shimanouchi, M. Tsuboi and T. Miyazawa, J. Chem. Phys., 1961, 35, 1597–1612.
32. M. L. Palacios and S. H. Taylor, Appl. Spectrosc., 2000, 54, 1372–1378.
33. F. Pointurier and O. Marie, Spectrochim. Acta, Part B, 2010, 65, 797–804.
34. M. Liegeois-Duyckaerts, Spectrochim. Acta, Part A, 1977, 33, 709–713.
35. L. Desgranges, G. Baldinozzi, P. Simon, G. Guimbretière and A. Canizares, J. Raman Spectrosc., 2012, 43, 455–458.
36. X. Guo, E. Tiferet, L. Qi, J. M. Solomon, A. Lanzirotti, M. Newville, M. H. Engelhard, R. K. Kukkadapu, D. Wu, E. S. Ilton, M. Asta, S. R. Sutton, H. Xu and A. Navrotsky, Dalton Trans., 2016, 45, 4622–4632.
37. F. Hund and U. Peetz, Z. Anorg. Allg. Chem., 1952, 267, 189–197.
38. F. Hund and U. Peetz, Z. Anorg. Allg. Chem., 1952, 271, 6–16.
39. F. Hund, U. Peetz and G. Kottenhahn, Z. Anorg. Allg. Chem., 1955, 278, 184–191.
40. W. Rüdorff, S. Kemmler and H. Leutner, Angew. Chem., 1962, 12, 429.
41. J. Selbin and J. D. Ortego, Chem. Rev., 1968, 69, 657–671.
42. H. Weitzel and C. Keller, J. Solid State Chem., 1973, 13, 136–141.
43. I. B. de Alleluia, M. Hoshi, W. G. Jocher and C. Keller, J. Inorg. Nucl. Chem., 1981, 43, 1831–1834.
44. Y. Hinatsu and T. Fujino, J. Less-Common Met., 1989, 149, 197–205.
45. C. Miyake, T. Isobe, Y. Yoneda and S. Imito, Inorg. Chem. Acta, 1987, 140, 137–140.

---