Synthesis and in situ ion irradiation of A-site de ﬁ cient zirconate perovskite ceramics † of Materials Chemistry A PAPER

The in ﬂ uence of A-site vacancies on zirconate perovskite structured ceramics of formula A 1 (cid:1) x La 2 x /3 ZrO 3 (A ¼ Ca, Sr, Ba) has been investigated using 400 keV Kr + in situ ion irradiation in a TEM with varied temperature from 100 K to 673 K. The solubility limit of La within the Ba 1 (cid:1) x La 2 x /3 ZrO 3 system was 5.33 ml% ( x ¼ 0.08) using a solid-state synthesis route, with space group symmetry of Pm (cid:1) 3 m for x # 0.08. The Ca 0.9 La 0.067 ZrO 3 and Sr 0.9 La 0.067 ZrO 3 samples were found to synthesise single-phase with the Pcmn and Pbnm symmetries of their respective end member structures. The addition of 6.67 ml% La ( x ¼ 0.1) in CaZrO 3 and SrZrO 3 was found to decrease the critical temperature of amorphisation, T c , in comparison with the undoped structures. A similar decrease was found for the 5.33 ml% La doped BaZrO 3 structure. Namely, values were observed for doped samples of z 404 K, z 520 K and z 621 K (A ¼ Sr, Ca, Ba respectively), compared to z 440 K, z 623 K and z 675 K for SrZrO 3 , CaZrO 3 and BaZrO 3 , respectively. This follows similar behaviour to that observed in titanate systems at low dopant levels, and these ﬁ ndings have been discussed in relation to both the presence of cation vacancies in the doped samples and the presence of the La cation. Void formation, growth and induced morphology changes have also been characterised, onset by ion irradiation.


Introduction
Perovskite structured materials have been widely investigated for various material applications, including as fuel cells, [1][2][3][4] dielectric resonators, 5-7 photovoltaics, 8,9 light emitting diodes, 10 lasers, 11,12 next generation nuclear fuels 13,14 and as hightemperature superconducting magnets. [15][16][17][18] Another potential application is in the immobilisation of nuclear waste streams as a host ceramic matrix. The namesake perovskite CaTiO 3 has been widely studied as a direct immobilisation matrix, including as a constituent of the SYNROC study, [19][20][21][22][23][24] while it is a commonly observed secondary actinide bearing phase when targeting zirconolite structured systems. [25][26][27][28][29][30] In the context of nuclear waste disposal, waste immobilisation matrices will be required to operate with a service lifetime on geological timescales (z10 6 years) due to the long half-lives of the isotopes in question. Over such timescales, doping a ceramic structure like perovskite with actinides, i.e. U, Pu, Am, will impart signicant radiation damage to the crystalline matrix from alpha-decay processes. Key to ceramic performance in this context will be the damage required to amorphise the crystalline structure, a consequence of which could be increased leaching rates in contact with ground water should a disposal canister fail. In this context, previous studies of radiation damage in perovskites have focused on the titanate series, [31][32][33][34][35][36][37][38][39][40][41][42][43][44][45][46][47] while niobates and tantalates have also received attention. 48 Zirconate systems, namely CaZrO 3 , SrZrO 3 and BaZrO 3 , have received little attention, although the response of CaZrO 3 to swi heavy ions has been previously investigated. 49 The focus of the present study, therefore, is investigation of radiation damage effects within these zirconate systems using in situ ion irradiation techniques. Such techniques are frequently employed to calculate amorphisation doses over accelerated timescales.
Perovskite structured materials take the base structural formula ABO 3 and an array of elemental combinations have been synthesised and characterised. The aristotype perovskite structure is exhibited by SrTiO 3 in a cubic Pm 3m arrangement, with Sr cations located at the unit cell corners, a body centred Ti cation and face-centred corner sharing O atoms, forming corner sharing TiO 6 octahedra throughout the structure. While SrTiO 3 exhibits the ideal ratio of A/B site cation radii (as governed by the Goldschmidt tolerance factor 50 ), distorted structures can form when this ratio changes through the tilting of BO 6 octahedra and, for example, anti-parallel A-site cation displacement. Within this study, the CaZrO 3 , SrZrO 3 and BaZrO 3 compositions have been investigated, and these structures are shown graphically in Fig. 1. CaZrO 3 presents at room temperature in the orthorhombic Pcmn (a À a À c + using Glazer notation 51 ), with CaZrO 3 reported to remain orthorhombic <2023 K prior to a transition to cubic symmetry. 52,53 SrZrO 3 takes the Pbnm (a À a À c + ) structure, with temperature induced phase transitions investigated by various groups. [54][55][56] Most recently Hasegawa et al. 57 revisited the SrZrO 3 system and reported transitions to Ibmm (a À a À c 0 )at1042K,I4/mcm (a 0 a 0 c À ) at 1119 K and to Pm 3m (a 0 a 0 a 0 ) at 1375 K. BaZrO 3 has the aristotype perovskite structure of Pm 3m (a 0 a 0 a 0 ), with no temperature related phase-transitions reported within the literature from 10 K to 1273 K. [58][59][60] Both space groups Pcmn and Pbnm are non-stand settings of Pnma (space group 62), however, within this work we refer to the most commonly used settings within the literature.
A further consequence of doping the perovskite ABO 3 structure is the likely induction of cation vacancy defects. Perovskite structures have been reported to exhibit a remarkable tolerance for such cation deciency, and this has been hypothesised in a disposal context through doping experiments with La and Gd. 61 When considering the multivalent nature of those isotopes suitable for ceramic immobilisation, understanding the effects of cation deciency on radiation effects is of key interest. Titanate compositions investigated in this regard include the A-site decient Sr 1Àx La 2x/3 TiO 3 (ref. 62) and Ca 1Àx -La 2x/3 TiO 3 (ref. 63 and 64) systems. A-site cation vacancies are produced within these systems whereby a cation vacancy is induced for every two dopant La cations. In both of these systems, in situ ion irradiation in a transmission electron microscope (TEM) has shown a reduction in the critical temperature of amorphisation, T c , in the 0.1 # x # 0.4 region. This value is dened as the temperature at which the damage rate equals the rate of recovery from damage, for a specic irradiation condition. A lower T c value can therefore be interpreted as a sample having recovery characteristics that require less thermal energy to suppress amorphisation under irradiation. While this could be linked to a possible higher displacement energy for La, for $ 0.5 T c was found to increase with respect to each end member, producing the opposite trend at higher La contents. It has been hypothesised that the increased A-site vacancy content in these systems increases recovery rates through increased defect migration and the increased likelihood for recombination events on the perovskite A-site. Therefore, as well as studying the end-member zirconate systems, this paper presents work to investigate whether such effects translate to zirconate perovskite systems in the context of nuclear waste disposal. To this purpose, samples of Ca 0.9 -La 0.067 ZrO 3 ,S r 0.9 La 0.067 ZrO 3 and Ba 0.92 La 0.053 ZrO 3 were also synthesised and investigated using in situ ion irradiation. Synthesis and irradiation of a Ba 0.9 La 0.067 ZrO 3 sample was intended, however, a study to nd the solubility limit of La within BaZrO 3 determined this sample would not synthesise, as is presented within this paper.

Sample synthesis
Compositions of A 1Àx La 2x/3 ZrO 3 (A ¼ Ca, Sr; x ¼ 0 and 0.1) and Ba 1Àx La 2x/3 ZrO 3 (x ¼ 0, 0.02, 0.04, 0.06, 0.08 and 0.1) were synthesised via a solid-state route. The additional Ba 1Àx La 2x/ 3 ZrO 3 samples were synthesised in order to nd the solubility limit of La in this system. Stoichiometric quantities of CaCO 3 , SrCO 3 , BaCO 3 , ZrO 2 (all 99.9%; Sigma-Aldrich), each dried at 180 C, and La 2 O 3 (99.9%; Sigma-Aldrich), dried at 800 C, were weighed to stoichiometric amounts and milled in isopropanol using a planetary ball mill with silicon nitride milling media for 15 minutes at 400 rpm. Aer milling to homogeneity, powders were dried at 80 C, sieved (212 mm; steel) and uniaxially pressed into 10 mm pellets to produce green bodies. Samples were then red in zirconia crucibles at 1450 C for 48 hours in air, reground using an agate pestle and mortar and re-sieved to produce ne, reacted powders.

X-ray diffraction
X-ray diffraction (XRD) was undertaken using a Panalytical X'pert 3 powder diffractometer, operating at 45 mV, 40 mA with Cu Ka (l ¼ 1.5418Å). The scan range was 10 to 90 with a step size of 0.02 and step time of 0.2 s. A 1/4 divergent slit was employed with 0.02 rad Soller slits. Powders were backloaded into sample holders with the rear surface of packed powders exposed to the X-ray beam, minimising preferential orientation. All data were calibrated using an external NIST 640e line position standard (ESI Fig. 1 †). Lattice parameter renement was Fig. 1 Structures of (a) CaZrO 3 (Pcmn; a À a À c + ), (b) SrZrO 3 (Pbnm; a À a À c + ) and (c) BaZrO 3 (Pm3m; a 0 a 0 a 0 ), as viewed slightly tilted off the [100] cubic direction. Ca ions are presented as red spheres, Sr ions as blue spheres, Ba ions as white spheres and ZrO 6 octahedra in pink. Figures were produced using the CrystalMaker software package using structural information from Levin et al. 7 Kennedy et al. 54 and Parida et al. 59 undertaken using the Bruker TOPAS soware package, with all errors reported as calculated by the renement process. Such errors were likely underestimated due to the use of a laboratorybased X-ray source and this should be noted when referring to the lattice parameters.

Transmission Electron microscopy
TEM analysis was undertaken using a JEOL F200, operating at 200 keV, with images and electron diffraction patterns captured using a Gatan OneView. Electron energy loss spectroscopy (EELS) was undertaken using a Gatan Image Filter (GIF) Quantum SE. Samples were prepared for TEM via the powdered "crushed grain" route, with powder ground in isopropanol and pipetted onto a carbon holey lined Cu TEM grid.

In situ ion irradiation
In situ ion irradiation was undertaken at the University of Hudderseld's MIAMI (Microscopes and Ion Accelerators for Materials Investigations) facility. The MIAMI-2 system was utilised, which comprises a Danfysik 921A ion source coupled with a 350 kV National Electrostatics Corporation ion accelerator, with the ion beam entering the TEM column of a 300 kV Hitachi H-9500 TEM at an angle of 18.7 to the electron beam. 65 Samples were prepared via the crushed grain route as described previously, with Au grids used for temperatures of >300 K and Cu for #300 K. Prior to ion irradiation, the presence of each intended element in the target composition of the specic grains monitored was conrmed via EELS spectroscopy using a Gatan Image Filter (GIF) Quantum SE.
The ion beam setup employed was 400 keV Kr + ions with a ux of 2 AE 0.2 Â 10 12 ions per cm 2 .K r + was employed for comparative purposes with the studies of Smith et al. 63 and Whittle et al. 64 Nuclear/electron stopping powers and ratios for each composition were estimated using the SRIM Monte Carlo code. 66 The "Ion Distribution and Quick Calculation of Damage" option was used with an average displacement energy of 50 eV set for each constituent element. Values calculated using SRIM are summarised in Table 1, including the determined theoretical densities of the powders used in the calculations and a ratio of electron to nuclear stopping powers (ENSP). Graphical SRIM curves are presented in ESI Fig. 2. † Under these conditions, stopping mechanisms were dominated by the nuclear component in all cases. Considering a TEM grain thickness of <100 nm, it was a requirement to induce displacement damage across the entire grain to observe full amorphisation. Peak displacement damage depth was estimated by SRIM to be 90 AE 10 nm in all cases, extending to a full damage range of z200 nm, ensuring the entire grain was subject to displacement damage. For each sample, the Kr retention to a thickness of <100 nm was estimated to be <5% by integrating the SRIM proles for implantation depth.
The electron beam valve was closed while ion irradiation was undertaken in order to avoid dual-beam synergistic effects, with the ion beam blocked off at intervals to allow for imaging and electron diffraction analysis. Several grains were monitored at each interval using bright-eld (BF) imaging and selective area electron diffraction (SAED). A Gatan 652 heating holder was used for irradiations at elevated temperatures, and a Gatan 636 liquid nitrogen cooling stage was used for temperatures below ambient. An average of the time/uence interval at which all Bragg diffraction peaks diminished and only amorphous diffuse scattering was observed within SAED patterns and the previous interval was taken as the critical amorphisation uence, F c , at a given temperature. This process for a SrZrO 3 grain is shown in Fig. 2. The error in F c was taken as the difference between the two intervals in which F c was observed to fall. An average of F c from all monitored grains gave the nal F c value for a particular sample and temperature. An additional 10% error in the ux was applied to account for uctuations in the ion beam. The ion beam ux was also measured before and aer irradiation to further monitor beam stability.
By measuring F c as a function of temperature, the critical temperature of amorphisation, T c , was determined. T c is dened as the temperature at which the rate of damage induced equals the rate of recovery from damage under specic irradiation conditions. As previously implemented by various groups, 37,62,63,67-76 values of T c were calculated using a nonlinear least squares renement of eqn (1) using several models that can be accessed elsewhere. 77,78 where F c0 is the critical amorphisation uence at 0 K, E a is the activation energy for recovery from radiation damage, k b is the Boltzmann constant (8.6173 Â 10 À5 eV K À1 ) and T is temperature. F c0 was extrapolated from the acquired data using eqn (1). a Density (g cm À3 ), energy loss to nuclear stopping (dE/dx) n (eV nm À1 ), energy loss to electronic stopping (dE/dx) e (eV nm À1 ) and electronic to nuclear stopping power ratio ENSP.
The frequent underestimation of E a has been reported using eqn (1) when compared to defect migration energies reported in oxides. Therefore, as for other studies, the relationship developed by Weber 77 was employed to determine both the thermal and irradiation assisted activation energies for recovery from damage, as outlined in eqn (2).
where f is the ion ux and n the effective jump frequency. Values for irradiation-assisted (10 12 s À1 ) and thermal (10 9 s À1 ) jump frequency were used based on previously reported estimations to allow for calculation of E a values for both thermal (E th ) and irradiation assisted (E irr ). 37 (110) in the ideal cubic Pm 3m. For the Ca 1Àx La 2x/3 ZrO 3 and Sr 1Àx La 2x/3 ZrO 3 systems, no additional peaks were identied other than those related to their end member counterparts, conrming the parent structure was maintained by the doped samples, i.e. Pcmn for CaZrO 3 and Ca 0.9 La 0.067 ZrO 3 ; Pbnm for SrZrO 3 and Sr 0.9 La 0.067 ZrO 3 . La incorporation into the structures was conrmed through lattice parameter contraction/expansion, as can be observed through the shiso fd i ffraction peaks to lower 2q values (higher d-spacing) for Ca 0.9 La 0.067 ZrO 3 ,a n d higher 2q values (lower d-spacing) for Sr 0.9 La 0.067 ZrO 3 .T h e s e shis are readily observable in the magnied regions in Fig. 3a and b, and lattice parameters for all samples are reported in Table 2 with tting errors. The expansion in lattice parameters for Ca 0.9 La 0.067 ZrO 3 was due to the smaller ionic radius of La (1.16 A) when replacing Ca (1.12Å) in 8-fold coordination, while Sr 0.9 La 0.067 ZrO 3 lattice parameters contract due to the larger Sr (1.26Å) that is replaced, again in 8-fold coordination.
While lattice contraction was observed in the Ba 0.9 La 0.067 -ZrO 3 sample as compared to BaZrO 3 , a secondary phase of the pyrochlore-structured La 2 Zr 2 O 7 was detected (see Fig. 3c). This suggested the solubility of La within the Ba 1Àx La 2x/3 ZrO 3 system was below x ¼ 0.1 for this synthesis route. This was likely due to the large difference in size between Ba (1.61Å) and La (1.36Å) in the A-site 12-fold coordination state in Pm 3m. The Goldschmidt tolerance factor, 50 t, has been extensively used to predict the chemical compositions that will form as perovskites, however,  to our knowledge no tolerance factor exists that can accurately predict the effects of A-site vacancies on the likelihood of perovskite formation. Considering an average cation radius based on either a fully occupied La site (i.e. Ba 0.9 La 0.1 ZrO 3 )o r the correct ratio (ignoring the presence of vacancies, as suggested by Ganguly et al. 79 ) reduces t below the ideal BaZrO 3 value of z1, but is still in the regime for perovskite formation. As Whittle et al. 80 report, vacancy defects can occupy more space than a fully-occupied cation lattice site, which is contrary to both these calculations. Without a tolerance factor that truly incorporates a specic vacancy-type defect, such calculations can never be considered accurate. Even so, the incorporation of a quantity of La within the target Ba 0.9 La 0.067 ZrO 3 formula was conrmed by the peak shi observed in Fig. 3c, conrming lattice expansion due to ionic substitution.
Samples of composition x ¼ 0.02, 0.04, 0.06 and 0.08 were produced to identify the La solubility limit in Ba 1Àx La 2x/3 ZrO 3 . XRD patterns of the Ba 1Àx La 2x/3 ZrO 3 system for x ¼ 0, 0.02, 0.04, 0.06, 0.08 and 0.1 are presented in Fig. 4a, with magnications of the 27 # 2q # 32 region in Fig. 4b and the (110) reection in Fig. 4c. A magnication of the 27 # 2q # 32 region in which the (222) pyrochlore La 2 Zr 2 O 7 reection was observed is presented in Fig. 4b. For x # 0.08, no such reection was observed, conrming the solubility limit of La in the region 0.08 # x < 0.1. For x # 0.08 the Pm 3m structure was present, conrming the single-phase nature of these samples. A shi to higher 2q values was observed incrementally with increasing La content (Fig. 4c) until x ¼ 0.1, with calculated lattice parameters and unit cell volumes reported in Table 2. A linear decrease in unit cell volume was observed for the Ba 1Àx La 2x/3 ZrO 3 system until the solubility limit was reached at x ¼ 0.1. The contraction and expansion of the unit cells of the Sr 1Àx La 2x/3 ZrO 3 and Ca 1Àx La 2x/ Pcmn (a À a À c + ) CaZrO 3 5.58781 (8)   contraction (6.04%, equivalent to 0.25Å), which was not observed in our analysis. It is unlikely that such contraction could be induced by cation substitution, and if the lattice parameter reduction is taken as reported for x ¼ 0 / 0.02 / 0.04 / 0.06 / 0.08 / 0.1, the associated reduction in the cubic parameter would be equivalent to 6.04% / 0.37% / 0.18% / 0.17% / 0.02%. Errors are not reported by the authors, and it is likely that the 0.02% reduction from x ¼ 0.08 / 0.1 may be within error, and a solubility limit of La in the solid-solution had also been reached in this case. Electron diffraction was further used to conrm the space group symmetry of each doped composition, with SAED patterns taken with the electron beam orientated down the [001] axis presented for the Ca 0.9 La 0.067 ZrO 3 ,S r 0.9 La 0.067 ZrO 3 and Ba 0.08 La 0.053 ZrO 3 compositions presented in Fig. 5a, b and c, respectively. In each instance, the patterns from these doped samples were found to index to the counterpart parent end member, namely Pcmn for CaZrO 3 , Pbnm for SrZrO 3 and Pm 3m for BaZrO 3 . The presence of La within powdered specimens of the Ca 0.9 La 0.067 ZrO 3 ,S r 0.9 La 0.067 ZrO 3 and Ba 0.92 La 0.053 ZrO 3 samples was further conrmed using TEM/EELS. BF micrographs and corresponding EELS spectra are presented within ESI Fig. 3, 4

In situ ion irradiation
Values of critical amorphisation uences for all samples are presented in Fig. 6a, with ts from the nonlinear least-squares renements using eqn (1) also presented. Included within Fig. 6a are values for SrZrO 3 at 473 K, and BaZrO 3 at 673 K, at which amorphisation was not achieved, presented as grey data points. These values lie above the calculated T c curves for the respective samples, suggesting these irradiations took place at temperatures that prevented amorphisation. Values for F c0 , T c and E a , calculated using eqn (1), and E th and E irr values, calculated using eqn (2), are reported in Table 3. Calculated values for T c are also shown graphically in Fig. 6b.
Firstly, the end member compositions have T c values with the trend of SrZrO 3 (440 K) < CaZrO 3 (623 K) < BaZrO 3 (675 K). This was similar to the titanate compositions of CaTiO 3 , SrTiO 3 and BaTiO 3 reported by Meldrum et al. 48 which showed a T c relationship of SrTiO 3 (425 K) < CaTiO 3 (440 K) < BaTiO 3 (550 K). In all instances, the T c values for the zirconate end members calculated in this study were appreciably higher than their A-site titanate equivalents. While Meldrum et al. 48 employed a 1 MeV Kr + ion beam as opposed to the 400 keV Kr + beam utilised in the present study, a signicant difference in T c under the two different conditions would not be expected considering the SRIM calculations discussed previously, although the temperature specic F c values may be altered. Most comparisons within the present study are with those using 800 keV to 1 MeV Kr + ion beams, and while a signicant difference in calculated T c values is not expected, the difference in the ion beam used should be noted. This includes the studies of A-site decient systems previously reported. [62][63][64] The relationship between titanates and zirconates differs from those in the pyrochlore system, in which zirconates such as La 2 Zr 2 O 7 and Gd 2 Zr 2 O 7 show a remarkable tolerance for radiation damage compared to their titanate counterparts. [82][83][84][85] These relationships are governed by anti-site defect formation energies and phase transitions to the defect uorite   (1) are also presented. The grey solid square refers to a SrZrO 3 fluence at 473 K and the grey solid circle to a BaZrO 3 fluence at 673 K; amorphisation was not achieved in either of these cases. (b) Values of the critical temperature of amorphisation, T c for each sample plotted with atomic number. Each T c value is located above the respective "Ca", "Sr" and "Ba" markers referring to pristine and La doped CaZrO 3 , SrZrO 3 and BaZrO 3 samples. Pristine samples are presented in black, and La doped in red. structure, 35,86 processes that are not replicated within perovskites. The calculated T c values suggest the specic A-site cation plays a larger role than structure within perovskite systems. For the titanate systems, T c values take the structural trend Pm 3m < Pbnm < I4/mcm, while for the zirconates the trend is Pbnm ¼ Pcmn < Pm 3m, showing no relationship with relative symmetry, tolerance factor 50 or A/B site radii ratio. A recent study by Meena et al. 87 suggests a relative increase in the ionic-nature of bonding for SrTiO 3 > BaTiO 3 > CaTiO 3 , which does not follow the T c trends observed. With regard to displacement energy, E d , a variety of values through both experimental and theoretical routes have been determined, but no conclusive values can be assigned to these systems. While values have been determined for CaZrO 3 , the authors noted the reported cation values were appreciably lower than previously measured oxide samples, 88 preventing comparison between titanate and zirconate systems. It is thus hypothesised that the exible nature of the Ti(III)/Ti(IV) B-site, in comparison with the highly-refractory Zr(IV), may promote defect mobility and an increased tolerance for oxygen displacements at higher temperatures. This could take the form of ATi 1Àx 4+ Ti x 3+ O 3Àd phases being induced during irradiation, whereas zirconium reduction is far less likely, and the resultant oxygen deciency is harder to accommodate. This may also be the case for the as produced pristine materials, as has been observed for (Ba, Sr, Ca)TiO 3 perovskites. 89 For the La-doped samples, similar trends as with doped titanate perovskites are observed. For instance, a decrease in T c between CaTiO 3 and Ca 0.9 La 0.067 TiO 3 is reported of z140 K, 63 while a drop of z100 K is reported here for the counterpart zirconates. Similarly, a reduction in T c of z 90 K is reported between SrTiO 3 and Sr 0.85 La 0.1 TiO 3 (ref. 62) and a decrease of z40 K within the zirconate counterparts is found in the present investigation (note that the titanate contains 3.33 mol% greater La content than within our study). In direct comparison, the titanates exhibit a lower T c for both Ca and Sr based end members, with Ca 0.9 La 0.067 TiO 3 (<300 K) > Ca 0.9 La 0.067 ZrO 3 (520 K) and Sr 0.9 La 0.067 TiO 3 (308 K) > Sr 0.9 La 0.067 ZrO 3 (402 K). No structural changes were induced between the parent and doped samples in any of these cases. It is possible that La has an intrinsically higher E d than each parent A-site constituent. Considering the observations reported regarding the Sr 1Àx La 2x/ 3 TiO 3 and Ca 1Àx La 2x/3 TiO 3 systems, in which La doping of x > 0.4 samples increased T c with regard to the parent end member, it seems unlikely that an increased A-site displacement energy is the sole cause of the trends observed in these zirconates. Values for La within La 2 Zr 2 O 7 are reported to be similar to Ba, Sr and Ca in perovskite systems, namely E d ¼ 65 AE 20 eV, 90-92 providing further evidence against such a conclusion. As is reported in Table 3, the activation energies for both thermal and irradiation assisted recovery are reduced for all doped samples with relation to their respective end member counterparts. This does suggest that increased defect migration and ionic diffusion may be the cause of the reduction in T c for doped samples, however direct measurement of such characteristics would be required to prove this hypothesis.
Several studies on the effects of doping and sintering characteristics on perovskite systems have been reported, and the resultant consequences on cation and oxygen deciency. [93][94][95][96] Within studies of Sr 1Àx La 2x/3 TiO 3 and Ca 1Àx La 2x/3 TiO 3 , 62,63 samples were sintered at 1573 K in air, while the bulk ion irradiation study showing increased room temperature resistance to amorphisation for 0.2 # x # 0.4 used pellets sintered at 1673 K. 64 Oxygen loss has been shown to rapidly increase with sintering temperature in the A-site decient Sr 0.85 La 0.1 TiO 3 system, even at temperatures of 1473 K in owing O 2 . 97 This has the consequence of causing signicant reductions in the activation energy, E a , for bulk ionic conductivity, which can be readily linked to the concentration and mobility of oxygen vacancies within the system. Akin et al. 97 report a reduction in E a from 0.97 eV for SrTiO 3 sintered at 1723 K in O 2 to 0.16 eV for Sr 0.85 La 0.1 TiO 3 sintered under the same conditions. This decreased energy requirement for oxygen vacancy mobility may lead to increased recovery rates, particularly at elevated temperatures, that lead to reductions in T c . This does not take into consideration sintering temperatures. E a values of 1.64 eV and 1.79 eV are reported in Sr 0.85 La 0.1 TiO 3 sintered at 1473 K under O 2 and N 2 , respectively, and values of 0.16 eV at 1723 K in O 2 . They report the same sample sintered at 1723 K in N 2 as being too conductive to measure even at 10 K, suggesting an even lower E a value for this sintering condition. It may be that such effects regarding sintering atmosphere and temperature translate to radiation damage kinetics, but it should be noted the effects of sintering temperature are not as dramatic in SrTiO 3 . While these are feasible explanations, such mechanisms have not been tested in the present study and it is not possible to transfer this theory to the zirconates studied here,  (6) a Critical amorphisation uence at 0 K F c0 (Â 10 14 ions per cm 2 ), critical temperature of amorphisation, T c (K), activation energy E a (eV), activation energy for thermally assisted recovery E th (eV) and activation energy for irradiation assisted recovery E irr (eV).
other than through empirical observations of the similarities in T c . Based on the T c observations made, it is proposed that a similar mechanism leads to the reductions in T c for doped samples, i.e. less oxygen deciency within a zirconate system as compared to a titanate. Studying bulk pellets of A-site decient samples sintered under various temperatures and atmospheres may elucidate this behaviour, and studies are now taking place in this regard. Fig. 7 presents under-focus (z500 nm) BFTEM micrographs of an irradiated grain of the Ca 0.9 La 0.067 ZrO 3 sample at 473 K ( Fig. 7a-d) and 100 K (Fig. 7e-h). At a uence of 0.3 of the critical uence for amorphisation (0.3F c ; Fig. 7b) at 473 K, a high density of light circular contrast regions was observed. With increased uence to 0.8F c (Fig. 7c), these circular regions grew in size whilst decreasing in areal density, suggesting a migration and agglomeration of these features. On rst observation, it could be argued this was the result of Kr bubble formation, with the elevated temperature promoting Kr migration and agglomeration into bubbles. Forming Kr bubbles in ceramic oxide samples normally requires the implantation of on the order of 10 15 to 10 16 ions per cm 2 , and considering the <5% Kr retention (as predicted by SRIM) under these conditions it is unlikely this contrast was the result of Kr bubbles. Furthermore, previously observed bubble sizes in ceramic oxide systems are on the order of z2nm, even at these elevated temperatures. [98][99][100][101] Therefore, it is likely these areas are void-type defects, i.e. a high concentration of irradiation induced vacancies, with the elevated temperatures allowing for void growth and agglomeration. It can be qualitatively observed that at 0.8F c (Fig. 7c), larger voids were present and in higher density than at the amorphisation uence, F c (Fig. 7d). This process was likely driven by void migration and growth toward the grain edges, with the voids annihilating at the surfaces that act as a defect sink. While these observations evidence the microstructural build-up of defects under irradiation, such defects do not appear to have signicantly affected the increased recovery rates onset by increased thermal energy, with an amorphisation uence of F c ¼ 7.3 AE 0.8 Â 10 15 ions per cm 2 for this sample at 473 K.
The morphology of the grain was also observed to change, with the grain edges smoothing with increasing uence compared to the irregularly shaped pristine grain. This may have been due to material sputtering, however grain swelling has been reported in various ceramics under ion irradiation, and it may be that a combination of irradiation and temperature induced migration induced a "ow" of material that induced such a change in morphology. In contrast, at 100 K, no void formation was observed at any uence (Fig. 7e-h), in a similar manner to void formation in steels that requires elevated temperatures. Indeed, it may be that the increased recovery rates at 473 K were required to induce a signicant density of vacancy defects to allow for the formation of observable voids, or that at 100 K there is insignicant defect mobility to allow for the required agglomeration. Similarly, changes in grain morphology at this temperature were severely reduced, however this may have resulted from either reduced temperature preventing "ow" or reduced sputtering at a lower uence. Such effects were not observed at room temperature or below in any samples, conrming this was a thermally driven process.

Conclusions
Solid-solutions of single-phase Ca 0.9 La 0.067 ZrO 3 and Sr 0.9 -La 0.067 ZrO 3 were produced, and the solubility limit of La in BaZrO 3 , through the compositional formula Ba 1Àx La 2x/3 ZrO 3 , was found at x ¼ 0.08. Each of these samples maintained their parent end member structure, namely Pcmn, Pbnm and Pm 3m for CaZrO 3 , SrZrO 3 and BaZrO 3 , respectively. Values of the critical temperature of amorphisation, T c , were calculated of z440 K, z623 K and z675 K for SrZrO 3 , CaZrO 3 and BaZrO 3 , respectively, compared to z404 K, z520 K and z621 K for the La doped counterparts (Ba 0.08 La 0.053 ZrO 3 for the Ba sample). The response of the materials to ion irradiation were rationalised by linking to studies regarding ionic conductivity within doped perovskite titanates, and in particular oxygen vacancy mobility within such systems. Void formation, migration and agglomeration at elevated temperatures was observed, including changes in grain morphology, mechanisms that were not observed below room temperature.

Conflicts of interest
There are no conicts of interest to declare.