Synthesis, structure and electrical properties of Li+-doped pyrochlore Gd2Zr2O7

Irina A. Anokhina, Irina E. Animitsa, Anastasia F. Buzina, Vladimir I. Voronin, Vladimir B. Vykhodets, Tatyana Е. Kurennykh, Yuri P. Zaikov Ural Federal University, 19 Mira St., Yekaterinburg, 620002, Russia b Institute of High Temperature Electrochemistry of the Ural Branch of the Russian Academy of Sciences, 20 Akademicheskaya St., Yekaterinburg, 620990, Russia c Institute of Metal Physics of the Ural Branch of the Russian Academy of Sciences, 18 S. Kovalevskoy St., Yekaterinburg, 620108, Russia *e-mail: irina.animitsa@urfu.ru


Introduction
The use of molten salt electrolytes opens up new ways of implementing resource-efficient waste disposal processes that are safe for humans and the environment, including those for recycling radioactive substances and organic compounds [1][2][3]. Electrochemical methods have a minimal environmental hazard and allow the creation of controlled, closed process circuits. For instance, the development of oxygen sensors is required for the measurement of oxygen concentration during chemical processes or direct measurement of the oxygen content in the process gas [4]. However, this problem has not been solved yet. Currently used materials for an oxygen ion activity sensor are susceptible to corrosion in halide-based melts, therefore they are ineffective because they cannot provide precise control of technological process parameters [5,6]. To implement modern, environmentally friendly technological developments, new materials that work under extreme conditions (high temperatures, aggressive environments) are needed.
From the point of view of the possible use of the material as an O 2sensor for Li + -halide melts, the Gd 2 Zr 2 O 7 crystalline matrix with a pyrochlore structure is of interest for investigation. Gadolinium zirconate Gd 2 Zr 2 O 7 has attracted consider-able attention of scientists in the last few years because of its high thermochemical stability and structural flexibility [7,8]. In this work, we studied the structure and electrical properties of the Li + -doped compound Gd 1.55 Li 0.45 Zr 2 O 6.55 in comparison with the undoped Gd 2 Zr 2 O 7 . Lithium was chosen as a dopant for two reasons: to create the oxygen deficiency and to prevent the possibility of ion exchange in Li +containing melts.

Experimental
T h e ph a s e s G d 2 Z r 2 O 7 a n d Gd 1.55 Li 0.45 Zr 2 O 6.55 were obtained by the solid-state method and Li-doped sample was also obtained by the modified Pechini method using glycerol-nitrate technique. Starting reagents Gd 2 O 3 (99.998%, VEKTON, Russia), ZrO 2 (99.99%, REACHIM, Russia) were preliminarily calcined at 600 °C for 3h, Li 2 CO 3 (99.99%, VEKTON, Russia) at 400 °C for 2 h. The synthesis was carried out in the temperature range of 700-1300 °C with a step of 100 °C with intermediate grindings.
The phase composition was monitored by the X-ray analysis after each stage.
The sample with nominal stoichiometry Gd 1.55 Li 0.45 Zr 2 O 6.55 was synthesized by a citrate-nitrate technique based on the Pechini method. In addition, glycerol was used as a fuel and a ligand in complexation reactions. The preliminarily calcined gadolinium oxide, zirconium oxide and lithium carbonate were weighed according to the stoichiometry and then the powders were dissolved in an excess of concentrated nitric acid, moreover, microwave treatment was used to dissolve Gd 2 O 3 and ZrO 2 upon heating at 80−100 °C. Then, the required amount of distilled water was added in order to reach the cations' concentration of ~1 mol/L. Then citric acid and glycerol were added. The solutions were mixed and heated at the temperature of 200 °C until spontaneous combustion occurred. To remove residual carbon, the powder were further annealed at the temperature of 700 °C. The final product was obtained by further heat treatments in the temperature range of 800-1300 °C with intermediate grinding in an agate mortar in ethanol.
X-ray analysis was carried out to control the phase composition and to determine the structural parameters of the studied samples. X-ray diffraction patterns were obtained at room temperature on a Rigaku MiniFlex600 diffractometer (Rigaku, Japan) with Cu Kα radiation in the range of 2θ = 10-80° with a step of 0.01° and at a scanning rate of 0.5 °/min. Calculation of interatomic distances and determination of unit cell parameters were performed using FullProf software.
The elemental analysis for the samples Gd 2 Zr 2 O 7 and Gd 1.55 Li 0.45 Zr 2 O 6.55 was carried out by set of methods: -Chemical analysis for gadolinium and zirconium was carried out after each annealing temperature by inductively coupled plasma atomic emission spectroscopy on an Optima 4300 DV Perkin Elmer spectrometer, USA. The 0.1 g sample was dissolved in a mixture of concentrated sulfuric acid and ammonium sulfate in the ratio of 3:2 by weight.
-The concentration of lithium in the samples was measured by nuclear reaction analysis (NRA) technique using a 2 МV Van-de-Graff accelerator. The reaction 7 Li (p, α) 4 Нe was used. The particle energy in the primary beam of protons was 762 keV. An incident beam was 2 mm in diameter. For the experiments, the powders were pressed into indium plate. As a result, a layer at least 2 μm thick containing only oxide particles was formed near the outer surface of the sample. The flat surface of the samples was perpendicular to the axis of the primary beam. Products of nuclear reaction were registered by a silicon surface-barrier detector with the registration angle of 160°. The irradiation dose was determined using a secondary monitor with the statistic error of about 1% [9]. The lithium concentration was determined by comparing the spectra from the samples to be measured and the reference sample with a constant (in depth) impurity concentration. The sample Li 2 ZrO 3 was used as the reference sample, in which the lithium content was taken equal to 33.33 at.%. The concentration of light elements was calculated using the stopping power data for the examined samples [10]. The techniques of measurement and calculation of light element concentrations were described in more details in [11].
Conductivity measurements were performed using сylindrical disc-shaped specimens with a diameter of 8−10 mm and a thickness of ~ 2 mm, which were pressed at 2.8 MPa and then calcined at temperature of 1500 °C for 5 hours (Gd 2 Zr 2 O 7 ) and at temperature of 1100 °C (Gd 1.55 Li 0.45 Zr 2 O 6.55 ) for 5 hours in case of the microwave sintering. The density of ceramic samples was measured by hydrostatic weighing in kerosene. Gd 2 Zr 2 O 7 ceramics had a relative density of 93-96%, and the Gd 1.55 Li 0.45 Zr 2 O 6.55 sample had a relative density of 70%.
Electrical conductivity of the studied samples was measured by the method of electrochemical impedance in a two-electrode cell using an Impendancemeter Z-500PX (Elins, Chernogolovka) over a frequency range of 100 Hz -1 MHz at the temperature range of 300-1000 °C. The platinum paste was applied on both sides of the sintered pellets and calcined at 900 °C for 3 hours.
The resistance was found by extrapolating the impedance spectrum to the real axis using ZView2 software. The electrical conductivity was calculated by the well-known formula: σ = l / (SR), where R is the resistance of the sample, l is its thickness, S is the cross-sectional area. Conductivity measurements were carried out in air and argon.

Structural features and phase characterization Pyrochlore phase Gd 2 Zr 2 O 7
The X-ray diffraction pattern of the Gd 2 Zr 2 O 7 sample obtained after heat treatment at 1500 °C is shown in Fig. 1.
The X-ray diffraction pattern contains intense reflections that correspond to the structure of fluorite, and a set of superstructural lines with low intensity, characterizing the structure of pyrochlore. The intensity of these reflections is determined by the difference in the scattering abilities of the A-and B-cations of the A 2 B 2 O 7 compound; in addition, their intensities are also determined by the dis-placement of O1 atoms from the ideal position (3/8 1/8 1/8). The presence of superstructural reflections on the X-ray diffraction pattern confirms the formation of the pyrochlore structure. However, in the model of perfectly ordered pyrochlore, the treatment of X-ray diffraction patterns does not give good agreement with the experimental data. Therefore, the following assumptions were made: 1) The possibility of anti-structural defects in the cationic sublattice, i.e. part of the Gd atoms in A-sites replaces the same part of Zr atoms in B-sites and vice versa. In this case, the total number of cations in the lattice remains constant.
2) The existence of the defects in the oxygen sublattice; in addition, taking into account the results of investigations in [12], it was suggested that oxygen atoms could be redistributed between the O1 position and the O3 vacant position in the ideal lattice. The results of fitting are shown in Fig. 1, and the refined structural parameters are shown in Table 1. The crystal lattice is shown in Fig. 2.
It can be seen (Table 1), that 20% inversion of atoms at the A-and B-sublattices is observed, while 10% of oxygen atoms move from O1 sites to the initially vacant   O3 sites. The results obtained are in good agreement with published data. For example, it was indicated [13] that the maximum degree of disordering of cationic positions in Gd 2 Zr 2 O 7 can reach 40%. Such structural features of Gd 2 Zr 2 O 7 are determined by the following conditions. The ordered structure of pyrochlore is formed only at R A / R B ≥ 1.46 [13]. In the case of Gd 2 Zr 2 O 7 , the ratio of the cation radii is R Gd / R Zr = 1.46 (the radius of Gd 3+ in the 8-fold coordination is 1.053 Å, and the radius of Zr 4+ in the 6-fold coordination is 0.72 Å [14]). Thus, according to geometric criteria, the Gd 2 Zr 2 O 7 phase is located near the boundary of the pyrochlore−fluorite transition, which leads to the inversion of atoms in the sublattices and 100% ordering is not observed. Moreover, the degree of disordering in the cationic and oxygen sublattices depends on the regimes of heat treatments [15][16][17]. High-temperature studies of the structure are necessary, and neutron diffraction experiments are needed to detect the changes in the oxygen sublattice.

The phase Gd 1.55 Li 0.45 Zr 2 O 6.55
The evolution of XRD patterns for the sample Gd 1.55 Li 0.45 Zr 2 O 6.55 , prepared by solution and solid-state techniques, is shown in Fig. 3. As can be seen, the fluorite phase was obtained at 700 °C for the sample, prepared by solution method; the impurity phases Li 2 ZrO 3 and LiGdO 2 were also found in traces amounts, which disappeared upon annealing at 900 °C. A single-phase sample with the pyrochlore structure was obtained at 1100 °C. At the same time, the formation of a fluorite phase for the sample, prepared by solid-state method, was observed at the temperature of 1100 °C; the gadolinium and zirconium oxides were present up to the temperature of 1200 °C. A single-phase sample was obtained only at 1300 °C. Thus, the solution method makes it possible to reduce the synthesis temperature of a single-phase sample by 200 °C.
Chemical composition of the obtained phases was monitored at all stages of the heat treatments. It was found that at all stages of the synthesis and sintering the atomic ratios of gadolinium and zirconium did not change and remained in exact accordance with the theoretical composition. However, the lithium content varied depending on the annealing temperature.  Table 2 presents the data on the lithium content obtained by the method of nuclear reactions. As can be seen, lithium loss does not occur up to temperatures of 1100 °C, but at higher heat treatment temperatures, stoichiometry is shifted, and lithium is not detected in the samples at annealing temperatures of 1300 °C. These data made it possible to select the necessary temperature-time sequences for synthesis of Li + -containing phase with a given lithium stoichiometry. Thus, for the solidphase synthesis method, it is appropriate to calcine the powder at temperatures up to 1000-1100 °C, press it into a pellet and then prepare a single-phase sample by microwave sintering at 1000-1100 °C.
X-ray diffraction pattern refinement results of the sample Gd 1.55 Li 0.45 Zr 2 O 6.55 , obtained under microwave heat treatment at 1000 °C for 5 hours, are shown in Fig. 4.
The sample has a pyrochlore structure, but the superstructural lines (111, 311, 331, 511, 531, 444) are very weak, which indicates a disordering of the pyrochlore structure. The lattice parameter was smaller compared to the undoped phase Gd 2 Zr 2 O 7 .
Electrical measurements Fig. 5 shows the evolution of the impedance spectra of Gd 1.55 Li 0.45 Zr 2 O 6.55 for various temperature ranges. As can be seen, in the general case, several relaxation processes can be distinguished, and the shape of the impedance diagram changed with temperature. At high temperatures (T > 800 °C), a semicircle, which does not start from zero and can be attributed to the grain boundary response, is well distinguished, the capacitance for this process corresponds to a value of ~10 -8 F/cm. With decreasing temperature, the low frequency arc was not clearly visible, and the first semicircle, starting from zero, become more pronounced. This first semicircle can be attributed to the relaxation process due to the grain volume resistance with capacity of ~ 10 -11 F/cm. Thus, the impedance of the studied electrochemical cell consists of the grain volume resistance R g with a constant phase element CPE g parallel to it, the resistance of the grain boundaries R gb with the parallel constant phase element CPE gb and the electrode impedance Z el (the equivalent circuit is shown in Fig. 5).
Doping with lithium did not affect the general shape of the impedance spectra  in comparison with the undoped sample. However, it should be noted, that the impedance spectra of the Li + -containing sample, obtained by the solution method, was significantly different. The large semicircle was observed due to the resistance of the grain boundaries over the main region of the studied frequencies (Fig. 6). Therefore, the grain volume resistance could not be determined.
Since the ceramics of the studied phases Gd 2 Zr 2 O 7 and Gd 1.55 Li 0.45 Zr 2 O 6.55 had different porosities, it was necessary to normalize the values for the less dense Gd 1.55 Li 0.45 Zr 2 O 6.55 sample to 95−100% density. Conductivity was recalculated according to the known relation: σ measured =σ o •d 2 , where d is relative density, σ measured is the apparent conductivity of the porous material, σ o is the intrinsic conductivity of the dense material [18]. A comparison of the temperature dependences of the bulk conductivities of the Gd 1.55 Li 0.45 Zr 2 O 6.55 sample obtained by solid-state synthesis and the undoped Gd 2 Zr 2 O 7 is shown in Fig. 7.  In order to clarify the nature of the dominant type of charge carrier, the measurements of electrical conductivity were performed in argon (pO 2 = 10 -5 atm). As can be seen (Fig. 7), the electrical conductivity of the undoped sample of Gd 2 Zr 2 O 7 in argon was lower than in air. This suggests that in air the Gd 2 Zr 2 O 7 sample is characterized by mixed ion-hole type conductivity. The obtained results are in good agreement with the literature data [19], where it was shown that the conductivity in air is mixed with the ionic transport number t ion = 0.57 at 780 °C, and the contribution of electronic conductivity increases at high temperatures. Assuming that the conductivity in argon atmosphere at low temperatures is predominantly ionic, the estimation of the ion transport number in air gives the value t ion = 0.98 at 500 °C, which correlates well with literature data [20].
On the contrary, for Li + -doped sample, the conductivities measured in argon and in air are almost equal. This indicates predominantly ionic conduc-tivity. However, comparing the ionic conductivities, measured in argon, it can be seen that the ionic conductivity of the Li-doped sample is lower than that of gadolinium zirconate Gd 2 Zr 2 O 7 . This indicates lower mobility of oxygen vacancies in the Gd 1.55 Li 0.45 Zr 2 O 6.55 sample. The decrease in mobility is probably due to the interaction of oxygen vacancies O V •• with the acceptor defect Li″ Gd , which has the opposite charge. In this case, the formation of a neutral associate {Li″ Gd • V O •• } × occurs. So, trapping leads to a decrease in the mobility of oxygen vacancies and, accordingly, to a decrease in the conductivity.

Conclusions
The phase with the pyrochlore structure Gd 1.55 Li 0.45 Zr 2 O 6.55 was obtained. Monitoring of lithium content in the sample by the nuclear reaction method allowed the development of synthesis methods with a given lithium stoichiometry. The introduction of lithium into the Gdsublattice led to a decrease in the lattice parameter a = 10.4830 (8) Å in comparison with Gd 2 Zr 2 O 7 (a = 10.5346(2) Å). The Li + -doped sample was characterized by predominant oxygen-ion transport over a wide range of temperatures. Although doping did not increase the oxygen-ion conductivity, it suppressed hole conductivity.