Underpinning the use of indium as a neutron absorbing additive in zirconolite by X-ray absorption spectroscopy

Abstract

Indium (In) is a neutron absorbing additive that could feasibly be used to mitigate criticality in ceramic wasteforms containing Pu in the immobilised form, for which zirconolite (nominally CaZrTi₂O₇) is a candidate host phase. Herein, the solid solutions Ca_1-xZr_1-xIn_2xTi₂O₇ (0.10 ≤ x ≤ 1.00; air synthesis) and Ca_1-xU_xZrTi_2-2xIn_2xO₇ (x = 0.05, 0.10; air and argon synthesis) were investigated by conventional solid state sintering at a temperature of 1350 °C maintained for 20 h, with a view to characterise In³⁺ substitution behaviour in the zirconolite phase across the Ca²⁺, Zr⁴⁺ and Ti⁴⁺ sites. When targeting Ca_1-xZr_1-xIn_2xTi₂O₇, single phase zirconolite-2M was formed at In concentrations of 0.10 ≤ x ≤ 0.20; beyond x ≥ 0.20, a number of secondary In-containing phases were stabilised. Zirconolite-2M remained a constituent of the phase assemblage up to a concentration of x = 0.80, albeit at relatively low concentration beyond x ≥ 0.40. It was not possible to synthesise the In₂Ti₂O₇ end member compound using a solid state route. Analysis of the In K-edge XANES spectra in the single phase zirconolite-2M compounds confirmed that the In inventory was speciated as trivalent In³⁺, consistent with targeted oxidation state. However, fitting of the EXAFS region using the zirconolite-2M structural model was consistent with In³⁺ cations accommodated within the Ti⁴⁺ site, contrary to the targeted substitution scheme. When deploying U as a surrogate for immobilised Pu in the Ca_1-xU_xZrTi_2-2xIn_2xO₇ solid solution, it was demonstrated that, for both x = 0.05 and 0.10, In³⁺ was successfully able to stabilise zirconolite-2M when U was distributed predominantly as both U⁴⁺ and average U⁵⁺, when synthesised under argon and air, respectively, determined by U L₃-edge XANES analysis.

Introduction

Zirconolite (ideally monoclinic CaZrTi₂O₇; space group C2/c; Z = 8) is a candidate wasteform material for the immobilisation of actinides arising from reprocessing of spent nuclear fuel (SNF), such as U and Pu^1,2,3,4. The United Kingdom inventory of separated Pu is forecast to reach approximately 140 teHM (tonnes equivalent heavy metal) upon completion of the ongoing reprocessing campaign, which terminated in July 2022. At present, the UK policy favours reuse; material not suitable for this purpose should be immobilised to place beyond reach⁵. Should this policy not prove implementable, the technology to immobilize and dispose of the inventory would be required⁶. At present, the most technically feasible immobilisation pathway would see the conversion of the bulk inventory into solid ceramic monoliths, produced either by a campaign of conventional cold-press sintering (CPS) or hot isostatic pressing (HIP), prior to disposal in a geological disposal facility (GDF)^7,8.

It may be desirable for wasteforms containing a high fissile content to incorporate a suitable quantity of neutron absorbing additives to mitigate the potential for a criticality event to occur^9,10. The role of the neutron absorber within the wasteform is to ensure that the probability of a criticality event is mitigated by reducing the internal neutron flux via absorption. Accordingly, during wasteform development and composition scoping trails, during which potential solid solution regimes between the actinide portion and the host material are devised and characterised, it is desirable to ensure that sufficient concentrations of neutron absorbers can be co-accommodated. At present, a range of such additives and their respective incorporation mechanisms for zirconolite have been devised, mainly deploying Gd³⁺ and/or Hf^{4+2,11,12,13,14,15}. Whilst Gd³⁺ is typically accommodated over the Ca²⁺ and/or Zr⁴⁺ sites, requiring charge balance via a charge balancing species typically substituted for Ti⁴⁺ e.g. Ca_1-xZr_1-xGd_2xTi₂O₇ or Ca_1-xGd_xZrTi_2-xAl_xO₇, Hf⁴⁺ can be directly substituted for Zr⁴⁺ on the basis of near identical ionic radii (0.78 and 0.76 Å in sevenfold coordination, respectively), resulting in a very minor variation in unit cell volume¹⁶. However, there are a range of relatively uninvestigated additives that have been proposed as neutron poisons, such as Sm, Dy, Cd, B and In. At present, there is little to no reported data of In, Cd or B substitution within zirconolite and related titanate phases.

The incorporation of In³⁺ within zirconolite was reported by Begg et al., in which In³⁺ was deployed as a Ti³⁺ simulant¹⁷. A solid solution targeting CaZrTi_2-xIn_xO_7-2/x was fabricated by a cold press and sinter route, with a 1400 °C sintering temperature maintained for 20 h in air, with In³⁺ introduced as In₂O₃, targeting x = 0.25, 0.50 and 1.00. It was observed that zirconolite became unstable relative to perovskite and fluorite with increased substitution of In, with a zirconolite yield of 90%, 65% and 0% observed for x = 0.25, 0.50 and 1.00, respectively. It should be noted that In³⁺ substitution in the above solid solution was not charge balanced, with In³⁺ included to simulate the effects of Ti³⁺ ingrowth under reducing conditions rather than as a neutron poison. Another reported example of In-substitution in a zirconolite-like structure was KIn_0.33Ti_0.67Te₂O₇ fabricated by Lee et al., however this structure was shown to crystallise with orthorhombic unit cell symmetry in the space group Cmcm¹⁸. This material was reported to form a 3D framework in which each In³⁺ cation was bonded to 6 O atoms in octahedral coordination, again indicating that In³⁺ cations may preferentially occupy the Ti⁴⁺ sites in the CaZrTi₂O₇ structure. Herein, two novel zirconolite systems were investigated, the first of which was Ca_1-xZr_1-xIn_2xTi₂O₇, whereby In³⁺ was targeted across both Ca²⁺ and Zr⁴⁺ cation sites in equimolar quantity, resulting in a self-charge balancing solid solution. As will later be discussed, the first part of this study demonstrated that In³⁺ preferentially occupied the Ti⁴⁺ site in zirconolite; hence, the Ca_1-xU_xZrTi_2-2xIn_2xO₇ solid solution was also investigated, processed under both Ar and air, with the view to determine whether In³⁺ could successfully charge balance U⁴⁺ and U⁵⁺, respectively, at moderately low concentrations of U (x = 0.05 and 0.10).

Experimental methodology

Materials synthesis

Caution

Uranium is an alpha emitter. Manipulations, synthesis and characterisation was performed in a materials radiochemistry laboratory in a controlled area, using HEPA filtered fume hoods and a dedicated glovebox, following risk assessments and monitoring procedures.

Two solutions targeting Ca_1-xZr_1-xIn_2xTi₂O₇ (0.10 ≤ x ≤ 1.00, Δx = 0.10) and Ca_1-xU_xZrTi_2-2xIn_2xO₇ (x = 0.05, 0.10) were fabricated by a conventional mixed oxide synthesis route, whereby the precursors CaTiO₃ (Sigma Aldrich, 99.9% trace metals basis), ZrO₂ (Sigma Aldrich, 99.9% trace metals basis), TiO₂ (anatase, Sigma Aldrich, 99.9% trace metals basis), In₂O₃ (Sigma Aldrich, 99.9% trace metals basis) and UO₂ (ABSCO Ltd., depleted) were intimately mixed by planetary milling. Each reagent was dried at 800 °C prior to weighing, and added to a 45 mL ZrO₂-lined milling jar with isopropanol and ZrO₂ milling media. Each sample was homogenised at 400 rpm for 20 min, with the direction of milling reversed after 10 min intervals. The powder slurries were decanted and allowed to dry overnight at 80 °C to evaporate excess solvent; dried powders were further mixed by hand to break up agglomerates and pressed into the walls of a hardened steel die (ø = 10 mm) under 3 tonnes of uniaxial force to form green bodies. These pellets were then placed onto a ZrO₂ crucible and sintered at 1350 °C for 20 h (air for Ca_1-xZr_1-xIn_2xTi₂O₇; flowing Ar and air for Ca_1-xU_xZrTi_2-2xIn_2xO₇).

Materials characterisation

A portion of each sintered pellet was retained and finely ground for powder X-ray diffraction (XRD) using a Bruker D2 Phaser (Cu Kα source: λ = 1.5418 Å, Ni Filter) fitted with a Lynxeye Position Sensitive Detector, operating at 30 kV and 10 mA. Data were collected in the range 10° ≤ 2θ ≤ 70° with a stepsize of 0.02° s⁻¹. Phase identification and peak indexing was achieved using the PDF4+ database and Rietveld analysis was performed using the Bruker TOPAS package. Scanning electron microscopy (SEM) analysis was performed using a Hitachi TM3030 operating with a 15 kV accelerating voltage at a working distance of 8 mm, fitted with a Bruker Quantax 70 spectrometer for Energy Dispersive X-ray Spectrometry (EDS). Samples were prepared for SEM–EDS analysis by mounting in cold setting epoxy resin and curing for 24 h, prior to grinding using incremental grades of SiC paper, before polishing to a 1 μm optical finish using diamond suspension. In K-edge X-ray absorption spectroscopy (XAS) data were acquired at BL01B1, SPring-8 (Hyogo, Japan). The storage ring energy was operated at 8 GeV with a typical current of 100 mA. Measurements were obtained using a Si (311) double-crystal monochromator in transmission mode at room temperature. Data were collected alongside In₂O₃ and In foil reference compounds. U L₃-edge fluorescence and Zr K-edge transmission XAS data were collected at Diamond Light Source Beamline B18 (Oxford, UK), alongside a selection of reference compounds, containing U and Zr in a variety of coordination environments and U-oxidation states. A Si(111) double crystal monochromator was used to fine-tune incident synchrotron radiation, with the intensity of the incident and transmitted beam measured using ionization chambers, filled with a mixture of N₂ and He gas, operated in a stable region of the I/V curve. Data reduction and analysis was achieved using the Demeter software package¹⁹. Bond valence sums were calculated using the in-built function in the Artemis software package. Here, the bond valence sum parameters are derived from the work of Altermatt and Brown^20,21 and use the following equation:

$${s}_{ij}={e}^{\left[\frac{{R}_{ij}-{R}{^{\prime}}_{ij}}{b}\right]}$$

where s_ij is the bond valence, R_ij is the bond distance between atoms i and j and R′_ij and b are the empirically determined parameters by Altermatt and Brown.

Results and discussion

Phase evolution in the Ca_1-xZr_1-xIn_2xTi₂O₇ system

Powder XRD data for the formulations x = 0.10 and 0.20 were consistent with the formation of single phase zirconolite-2M. Unit cell dimensions were obtained by Rietveld analysis of powder diffraction data (Fig. 1, Table 1) corresponding to a decrease in the unit cell volume with increased In content. The unit cell parameters of pure CaZrTi₂O₇ have been previously reported as: a = 12.4462 Å, b = 7.2717 Å, c = 11.3813 Å, β = 100.5587° and V = 1012.60 ų²². There was therefore a consistent reduction in the lattice constants of the zirconolite-2M with respect to In substitution. No reflections representative of common secondary phases (ZrO₂, CaTiO₃) were evidenced in the diffraction patterns. Moreover, combined SEM–EDS analyses indicate the formation of high density microstructures comprised of a matrix of zirconolite-2M, with no phase separation visible by variation in backscattered electron contrast (Fig. 2). EDS analysis of the zirconolite-2M matrix formed for x = 0.10 and 0.20 confirmed the presence of In, indicated by the prominence of the In Lα and In Lβ emission lines present in the EDS spectrum (Fig. S1). The average composition of the zirconolite phase(s) was determined by semi-quantitative EDS analysis, normalised to seven oxygen atoms; the observed compositions are in good agreement with the targeted formulation (Table S1).

Figure 1

Rietveld refinement data for x = 0.10 (left) and x = 0.20 (right) compositions in the Ca_1-xZr_1-xIn_2xTi₂O₇ system fit to the zirconolite-2M structural model.

Table 1 Unit cell parameters for single phase In-doped zirconolite-2M compositions (x = 0.10, 0.20) in the Ca_1-xZr_1-xIn_2xTi₂O₇ system.

Figure 2

BSE images for x = 0.10 (left) and x = 0.20 (right) in the Ca_1-xZr_1-xIn_2xTi₂O₇ solid solution.

The In K-edge XANES spectra for single phase zirconolite-2M (x = 0.10 and 0.20) compositions are presented in Fig. 3. Data were collected alongside In₂O₃ and In foil reference compounds, representing In³⁺ and In⁰ oxidation states, respectively. For both In-containing zirconolite-2M compositions, the XANES spectra when overlaid were practically indistinguishable, indicative of identical speciation and coordination of In cations when targeting both x = 0.10 and 0.20. These spectra were comprised of a single intense absorption feature (E₀ = 27,945.5 eV) followed by a weak post-edge oscillation, with maxima at 28,000 eV. It was clear that position of the absorption edge of x = 0.10 and 0.20 compounds was identical to that of the In₂O₃ reference compound (space group Ia$\overline{3 }$) which contains trivalent In³⁺ distributed across two sites, both of which are sixfold coordinated to O²⁻, inferring uniform In³⁺ speciation within the zirconolite-2M phase. Moreover, linear combination fitting of the XANES spectra for x = 0.10 and 0.20 confirmed 100% In³⁺ speciation relative to In₂O₃ and In-foil reference compounds, with R-factors of 0.00469 and 0.00531, respectively.

Figure 3

In K-edge XANES spectra for In-doped zirconolite-2M at concentrations x = 0.10 and 0.20, alongside In₂O₃ and In foil reference compounds.

Fitting of the EXAFS spectra (Fig. 4, Table 2) for both the x = 0.10 and 0.20 compositions produced good fits and utilised very similar models that corresponded well to the expected zirconolite-2M structure. For x = 0.10, the best fit model (R-factor = 0.0078) included 6 O backscatterers at 2.16(1) Å, 2 Ti backscatterers at 3.33(1) Å, and 2 Ti backscatterers at 3.57(1) Å. For x = 0.20, the best fit model (R-factor = 0.0106) included a split first O shell, with 3 O backscatterers at 2.11(6) Å and 3 O backscatterers at 2.21(6) Å, 2 Ti backscatterers at 3.34(2) Å, and 2 Ti backscatterers at 3.58(2) Å. Both fits match well with the expected structure of zirconolite-2M according to Whittle et al.²³, with the exception of the second Ti shell at ~ 3.58 Å exhibiting an increase in the expected degeneracy (2 instead of the expected 1) since the Ca backscatterer at the same interatomic distance (R_eff = 3.57 Å) was not modelled to avoid over-parameterising the model. Given both the second Ti shell and the Ca backscatterers manifest at very similar areas of the EXAFS spectrum, it is likely that significant interference makes fitting the expected shells and their corresponding degeneracies (i.e. 1 Ti and 1 Ca as opposed to the best fit model 2 Ti) challenging. Similar challenges in fitting have been observed in other doped zirconolite systems²⁴ whereby the delineation of the exact type and number of backscatterers in the approximate range 3.5–3.6 Å was found to be not possible.

Figure 4

In K-edge XAS spectra for x = 0.10 and 0.20 compositions in the Ca_1-xZr_1-xIn_2xTi₂O₇ system. Left—k³-weighted EXAFS. Right—Fourier transform of the k³-weighted EXAFS, using a Hanning window function. Black lines are data and red lines are the best fit models for the data.

Table 2 Fitting parameters for In K-edge EXAFS data presented in Fig. 4.

The total degeneracy for the first O shell was found to be six in both samples (6 O at 2.16(1) Å for x = 0.10 and a split O shell of 3 O at 2.11(6) Å and 3 O at 2.21(6) Å for x = 0.20). The sixfold coordination of In³⁺, determined by these EXAFS analyses, implies that some compositional rearrangement is required to maintain charge balance across the single phase zirconolite-2M samples, for example (Ca_1-xTi_x)(Zr_1-xTi_x)Ti_2-2xIn_2xO₇, consistent with the known defect mechanisms in non-stoichiometric zirconolite²⁵. Indeed, the only previous study that explored In doping of zirconolite found that In readily expresses a preference for the fivefold Ti site¹⁷. The total degeneracy, and by extension the inferred mixture of In³⁺ coordination environments, correspond well to the bond valence sums (BVS) which sum to approximately 3 (2.973 for x = 0.10 and 3.014 for x = 0.20) for both samples. When trialing other models, with In³⁺ solely on the alternative Ti site (total O degeneracy of 5), the Ca site (total O degeneracy of 8) or Zr site (total O degeneracy of 7) the calculated BVS diverged markedly from the expected value of 3, and the overall model produced a qualitatively and quantitatively worse fit to the data.

As the targeted In³⁺ concentration was progressed beyond x ≥ 0.20, zirconolite-2M was no longer isolated as a single phase, with a number of ancillary In-bearing phases clearly distinguished by both XRD and combined SEM–EDS analyses. Nevertheless, zirconolite was detected by XRD up to x = 0.80, albeit at very low concentration (approximately 3 wt% at x = 0.80; see quantitative phase analysis in Table 3). Figure 5 shows the powder XRD data for compositions in the range 0.30 ≤ x ≤ 1.00. When targeting x = 0.30, a number of additional reflections corresponding to TiO₂ (P4₂/mnm) and InTi_0.75Ca_0.25O_3.25 (C2/m) were clearly observed, accounting for 2.3 ± 0.3 and 12.9 ± 0.3 wt% of the overall phase assemblage, respectively. SEM analysis (Fig. 6) confirmed the presence of all phases identified by XRD in the microstructure. In the compositional interval 0.40 ≤ x ≤ 0.60, an additional In-bearing cubic ZrO₂ phase was also observed (highlighted by arrows in Fig. 5), plateauing with a maximum fraction of ~ 24 wt% of the overall phase assemblage at x = 0.50. The relative concentrations of InTi_0.75Ca_0.25O_3.25 and TiO₂ steadily increased, with a corresponding decrease in the zirconolite-2M phase. It was interesting to note that zirconolite did not appear to undergo any structural transformations to the 4M or 3T polytypes, as is typically observed with excessive doping in the Ca²⁺ and/or Zr⁴⁺ sites¹. When targeting the end-member In₂Ti₂O₇ composition (i.e. x = 1.00), a two phase mixture of In₂TiO₅ (79.4 ± 0.34 wt%) and TiO₂ (20.6 ± 0.34 wt%) was identified by XRD and confirmed by SEM–EDS analyses, consistent with previous data²⁶. The crystal structure of the cubic pyrochlore phase is characterized as an anion-deficient fluorite superstructure, adopting the generic A₂B₂O₇ stoichiometry, resulting in two distinct cation sites; A cations are typically larger trivalent atoms (e.g. REE³⁺ = Dy, Y, Sm, Gd) whereas the B site is comprised of smaller, higher valence cations such as Ti⁴⁺ and Hf⁴⁺, although a number of variations do exist²⁷. The stability of the cubic pyrochlore phase (space group Fd $\overline{3 }$ m) is dictated by the size ratio of r_A/r_B whereby the A and B sites are eight and sixfold coordinated to O²⁻, respectively. Varying A and B site cations such that r_A/r_B > 1.78 results in the formation of a monoclinic structure, whereas r_A/r_B < 1.46 promotes the defect-fluorite structure type, in which oxygen vacancies are disordered across the sub-lattice. Considering the respective ionic radii of In³⁺ and Ti⁴⁺ in eight and sixfold coordination (0.92 Å and 0.605 Å) and the corresponding ratio (r_A/r_B = 1.52) the In₂Ti₂O₇ phase should form the cubic pyrochlore structure on the basis of ionic radius ratio, nevertheless, no yield of In₂Ti₂O₇ was observed at 1350 °C. Despite further attempts to form this phase with an increased sintering temperature of 1700 °C and extended dwell time of 24 h, we noted that these conditions were sufficient to completely volatilise the pellet during sintering and no yield was obtained.

Table 3 Quantitative phase analysis (QPA) for the Ca_1-xZr_1-xIn_2xTi₂O₇ solid solution calculated from Rietveld analysis of powder XRD data (* indicates phase purity).

Figure 5

Powder XRD data for 0.30 ≤ x ≤ 1.00 compositions in the Ca_1-xZr_1-xIn_2xTi₂O₇ system. TiO₂ reflections are labelled with closed circles. In₂TiO₅ reflections are labelled with stars. The prominent reflections contributing to In-doped c-ZrO₂ are indicated with arrows.

Figure 6

SEM analysis of x = 0.30, 0.50, 0.70 and 0.90 compositions in the Ca_1-xZr_1-xIn_2xTi₂O₇ solid solution.

U L₃ and Zr K-edge X-ray absorption near edge structure (XANES) analysis of the Ca_1-xU_xZrTi_2-2xIn_2xO₇ system

As In adopted the In³⁺ oxidation state when substituted in the Ca_1-xZr_1-xIn_2xTi₂O₇ solid solution, preferentially occupying the Ti site, it may be feasible for In³⁺ to act as a neutron poison and charge balancing species within the Ti⁴⁺ site. Moreover, to our knowledge, the only other reported instance of In³⁺ substitution in zirconolite was consistent with occupation in the Ti⁴⁺ site as a simulant for the reduced Ti³⁺ species¹⁷. Hence, two compositions in the Ca_1-xU_xZrTi_2-2xIn_2xO₇ system targeting x = 0.05 and 0.10 were also fabricated, with U deployed as a surrogate for Pu. Samples were first synthesised under argon, with a view to maintain the preferred U⁴⁺ valence configuration, as this is generally the target oxidation state for immobilised Pu in zirconolite. Powder XRD data for both compositions sintered under flowing Ar gas are shown in Fig. 7, and are consistent with the formation of near single phase zirconolite-2M (a small fraction of ZrO₂ was also observed) at both targeted concentrations of U, indicating the In³⁺ was successfully accommodated in solid solution. The unit cell parameters of the zirconolite-2M phase(s) were calculated by Rietveld analysis of powder XRD data and are listed in Table 3. A yield of near single phase zirconolite-2M (< 2 wt% free ZrO₂) was also obtained when compositions were synthesised in air, indicating that In³⁺ may capable of charge balancing U valence states greater than U⁴⁺.

Figure 7

Powder XRD data for Ca_0.95U_0.05ZrTi_1.90In_0.10O₇ (x = 0.05) and Ca_0.90U_0.10ZrTi_1.80In_0.20O₇ (x = 0.10) synthesised under both air and argon.

In order to determine the formal oxidation state of U, X-ray absorption near edge structure (XANES) spectra were collected at the U L₃-edge (~ 17,166 eV) in transmission mode (Fig. 8) alongside a series of reference compounds containing U with known oxidation state and coordination environment. Scanning across the U L₃ absorption edge probes the dipole transition from a core 2p_3/2 shell to the partially occupied 6d valence shell. The observed spectra for all measured compounds were comprised of a single intense absorption feature with additional resonance features present above the main absorption edge; the energy position of the absorption edge is correlated to the formal oxidation state of the absorbing atom in many actinide species.

Figure 8

Normalised U L₃-edge XANES for x = 0.05 (left) and x = 0.10 (right) compositions in the Ca_1-xU_xZrTi_2-2xIn_2xO₇ solid solution sintered in argon and air. Data are displayed alongside UO₂ (dashed) and CaUO₄ (dotted) reference compounds, representing U⁴⁺ and U⁶⁺, respectively.

The E₀ position of compositions sintered under flowing Ar was measured as the energy for the normalised absorbance μx = 0.5 (i.e. the energy at half the absorption edge), and was found to be 17,166.0(5) eV for both x = 0.05 and 0.10 compounds, indicative of uniform U speciation at both concentrations. These spectra displayed a clear resemblance to the UO₂ reference compound with respect to the normalised edge intensity, position and post-edge oscillation feature, consistent with U⁴⁺ speciation in the zirconolite-2M compounds. The E₀ position of the U L₃-edge in the air-synthesised compounds was shifted to higher energy relative to the UO₂ reference compound, measured to be 17,167.7(5) eV and 17,167.8(5) eV for x = 0.05 and 0.10, respectively, indicative of oxidised U. To establish the average oxidation states of U in both sets of compounds, a linear regression plot of the E₀ position (defined as μx = 0.5) relative to a large suite of reference compounds was constructed (Fig. 9). The oxidation states of U in the x = 0.05 and 0.10 were calculated to be 4.1 and 4.1 in argon; 5.2 and 5.3 in air. These data are strong evidence to support the use of zirconolite as a potential wasteform for U/Pu, as the zirconolite-2M structure is clearly capable of accommodating changes in oxidation state without phase separation. Moreover, it is clear that In³⁺ is capable of charge balancing both U⁴⁺ and U⁵⁺ within the zirconolite structure, in the case of the latter, implying a charge balancing mechanism consistent with Ca_1-xU⁵⁺_xZrTi_2-2xIn_2xO_7+x (Table 4).

Figure 9

Oxidation state of U in Ca_1-xU_xZrTi_2-2xIn_2xO₇ (x = 0.05, 0.10; air and argon) calculated by linear regression of E₀ position relative to reference compounds (U⁴⁺O₂, U⁴⁺Ti₂O₆, U⁴⁺SiO₄, U⁵⁺SbO₅, U⁵⁺MoO₄, LaU⁵⁺O₅, CaU⁶⁺O₄, Ca₃U⁶⁺O₆ and MgU⁶⁺O₄).

Table 4 Unit cell parameters for single phase U-doped zirconolite-2M compositions (x = 0.05, 0.10) in the Ca_1-xU_xZrTi_2-2xIn_2xO₇ system, synthesised in both air and argon.

XAS spectra were acquired at the Zr K-edge for x = 0.05 and 0.10 in the Ca_1-xU_xZrTi_2-2xIn_2xO₇ solid solution (for both air and argon) alongside a variety of reference compounds containing Zr in distinctive coordination environments (Figs. 10, 11). There were notable differences in the XANES spectra of Zr-containing compounds depending on the coordination of Zr atoms. However, the absorption features for each measured zirconolite compound were observed to be near-identical regardless of targeted U concentration or processing environment, and, in excellent agreement with those previously reported in the literature, including synthetic CaZrTi₂O₇^22,28 and annealed metamict (Ca,Th)ZrTi₂O₇ zirconolite^29,30. The XANES spectra are presented alongside a series of reference compounds, with variation in pre-edge intensity and main white-line absorption features consistent with Zr in a variety of coordination environments³¹. It was observed that the zirconolite and m-ZrO₂ compounds exhibited rounded features at the crest of the white line, which is typical of sevenfold coordinate Zr, whereas the ZrSiO₄, t-ZrO₂ and Li₂ZrO₃ compounds exhibited notable peak splitting which is typical of octahedral coordination, or, as in the case of ZrSiO₄, highly distorted eightfold coordination²². Qualitative analysis of the XANES region was consistent with the Zr atoms in zirconolite occupying the expected sevenfold coordination environment, as there are clear parallels with the m-^VIIZrO₂ reference compound. The Zr-K edge k³-weighted EXAFS modelling of the first oxygen coordination shell for each zirconolite specimen is presented in Fig. 10 (Table S2). The same model could be fit to all samples for both atmospheres (x = 0.05 and 0.10 for both argon and air) and was consistent with 7 O backscatterers present around the Zr central atom. In all fits, the first oxygen coordination shell was split into 4 O backscatterers at 2.10–2.11 Å and 3 O at 2.24–2.25 Å. These fits are in good agreement with the expected Zr coordination environment for zirconolite, in which a spread of 7 O backscatterers are present from 2.045 to 2.339 Å²³. A distinct yet low intensity pre-edge feature was observed for the zirconolite materials and Zr compounds, attributed to the formally forbidden 1s→4d transition; the relative prominence of this feature is increased for compounds for which Zr is located in a distorted, non-centrosymmetric coordination environment. The inset of Fig. 10 highlights the pre-edge region of zirconolite (represented by the x = 0.10 (argon) composition), consistent with sevenfold coordination, alongside ZrSiO₄, m-ZrO₂ and Li₂ZrO₃, representing eight, seven and sixfold coordinate Zr, respectively. There is a clear trend in the intensity of the pre-edge feature with reduced coordination, exhibited most clearly by the difference between Li₂^VIZrO₃ and ^VIIIZrSiO₄. The increased intensity of this feature is attributed to a lack of centrosymmetry around the absorbing Zr atom. Furthermore, the pre-edge shape and intensity of the zirconolite and m-ZrO₂ appears similar, further supporting the existence of sevenfold Zr in the U/In-doped zirconolite materials.

Figure 10

Zr K-edge XAS spectra for Ca_1-xU_xZrTi_2-2xIn_2xO₇ (x = 0.05, 0.10) processed under argon and air. Left - Zr K-edge XANES data. Right - k³-weighted EXAFS of the Zr K-edge using a Hanning function window. The fit shown is for the first coordination shell only; black lines are data and red lines are the best fit models for the data.

Figure 11

Zr K-edge XANES data for Ca_1-xU_xZrTi_2-2xIn_2xO₇ (x = 0.05, 0.10) processed under argon and air. Data are presented alongside a series of Zr reference compounds.

Conclusions

Herein, we have presented a novel examination of the solid solution behaviour of indium within the zirconolite phase, as a potential neutron absorbing additive. Two distinct substitution regimes were devised, targeting Ca_1-xZr_1-xIn_2xTi₂O₇ (targeting In³⁺ distributed equimolar across Ca²⁺ and Zr⁴⁺ sites) and Ca_1-xU_xZrTi_2-2xIn_2xO₇ (targeting In³⁺ across the Ti⁴⁺ site). In K-edge X-ray spectroscopy data confirmed In was uniformly present in the In³⁺ oxidation state, yet the prevailing In coordination environment was consistent with accommodation in the Ti⁴⁺ site in sixfold coordination, contrary to the targeted formulation. Nevertheless, single phase zirconolite-2M was able to form in the Ca_1-xZr_1-xIn_2xTi₂O₇ system for 0.10 ≤ x ≤ 0.20. Progressive substitution of In³⁺ beyond x ≥ 0.20 promoted the formation of several In-titanate phases, namely In₂TiO₅ and InTi_0.75Ca_0.25O_3.25, alongside TiO₂ and c-ZrO₂. In³⁺ was also utilised to successfully charge compensate U⁴⁺ and U⁵⁺ resulting in the formation of near single phase zirconolite-2M (accompanied by ~ 1–2 wt% ZrO₂) in the Ca_1-xU_xZrTi_2-2xIn_2xO₇ system at x = 0.05 and 0.10. The dominant oxidation states of U⁴⁺ (argon synthesis) and U⁵⁺ (air synthesis) were determined by U L₃-edge XANES analysis. Analysis of the XANES and EXAFS region of the Zr K-edge in these materials was consistent with Zr⁴⁺ occupying sevenfold coordinated sites in the zirconolite structure, as would be expected. These data form a useful contribution towards ongoing efforts to design, characterise and performance test titanate wasteform materials substituted with appropriate quantities of neutron poisoning additives.