Nitin Kumarab,
K. Sandeep Raocd,
Anuab,
A. K. Sahue,
S. N. Achary
*cd and
S. K. Deshpandea
aUGC-DAE Consortium for Scientific Research, Mumbai Centre, Bhabha Atomic Research Centre, Mumbai 400085, India
bSavitribai Phule Pune University, Ganeshkhind, Pune 411007, India
cChemistry Division, Bhabha Atomic Research Centre, Mumbai 400085, India. E-mail: sachary@barc.gov.in
dHomi Bhabha National Institute, Anushakti Nagar, Mumbai 400094, India
eGlass and Advanced Materials Division, Bhabha Atomic Research Centre, Mumbai 400085, India
First published on 28th July 2025
A series of hexagonal perovskite derivative compounds Ba3WV1−xGexO8.5−x/2 (x = 0, 0.1, 0.2, and 0.3) were synthesized using a solid-state reaction route and characterized by powder XRD, SEM-EDX and dielectric spectroscopy. All the samples were found to have a rhombohedral lattice (space group: Rm) and showed a systematic increasing trend in unit cell parameters with increasing Ge4+ content. High-temperature XRD studies on the samples confirmed their stability up to 1023 K. Temperature- and frequency-dependent dielectric studies indicated high ionic conduction in all these samples and systematic enhancement in ionic conductivity with increasing Ge4+ incorporation. Among the studied samples, Ba3WV0.7Ge0.3O8.5 showed the highest dc conductivity (∼1.80 × 10−5 S cm−1 at 973 K), while Ba3WVO8.5 exhibited a value of 1.03 × 10−7 S cm−1 under the same conditions. This enhancement in conductivity upon incorporation of Ge4+ ions was attributed to the easy diffusion of oxide ions in the lattice due to increasing unit cell volume, oxygen ion vacancies and anion disorder in the BaO3 layer owing to preferential formation of tetrahedral GeO4. The temperature evolution of ionic conductivity of the samples showed Arrhenius behaviour, and activation energies were in the range of 0.9–1.6 eV. In addition, a changeover in conduction mechanism was observed at higher temperatures. The shape parameters in the electrical modulus spectra were determined using Havriliak–Negami (H–N) function fitting, and it was found that the distribution of relaxation times increased with the Ge4+ content due to a widely distributed spatial arrangement of carriers and energetics for carrier migration.
High ionic conductivities have been reported in compounds of only a few structural families, such as fluorite,7 apatite,8 melilite,9 and perovskite.10–12 A number of studies are focused on normal perovskite-type compounds due to their flexible structure, which can accommodate varieties of cation and anion vacancies, and among them, Na0.5Bi0.5TiO3 and LaGaO3 are considered good solid-state electrolytes for moderate-temperature applications.10–12 Such compositional and structural flexibilities allow convenient modes to tune carrier and transport properties in perovskite-type materials, and that makes them attractive for solar cell,13 optoelectronic and sensor applications.14–16 Recently, high oxide ion conduction at moderate temperatures in certain types of hexagonal perovskites like Ba3M′M′′O8.5 (M′ = W and Mo; M′′ = Nb and V) have been reported in the literature.17–20 The high chemical and thermal stability along with high ionic conductivity at moderate temperatures of such compounds attracted widespread research attention to understand and improve the ionic conductivity.21–24 Fop et al. have reported conductivity as high as 10−2 to 10−3 S cm−1 at around 873 K with high stability in an atmosphere with wider partial pressure of oxygen in hexagonal Ba3NbW/MoO8.5 systems.21,22 It may be mentioned here that the structure of the hexagonal perovskite derivative compounds is intermediate between that of perovskite 9R and the palmierite structure.25–28 The 9R perovskite has a 3D structure built by linking of trimers of face-shared MO6 octahedra through corner oxygen atoms, while the palmierite structure is made up of isolated MO4 tetrahedra.20,27–29 Indeed, the palmierite structure is a 9R perovskite with ordered arrangement of cations and oxygen vacancies, where one of the three closely packed [BaO3] layers is replaced by an oxygen-deficient [BaO2] layer.18,20,29 This arrangement separates the face-shared octahedral trimers of 9R to two isolated octahedra separated by a vacant octahedral space. In the 9R perovskites, there are three different types of oxygen sites (O1, O2 and O3), and two of them (O2, and O3) are underoccupied. These two underoccupied anion sites form disordered coordination polyhedra, which can be approximated by the disordering of octahedra and tetrahedra. Thus, Ba3WNbO8.5 has a more disordered structure than the 9R perovskite and palmierite, with partial occupancies of the metal site (M1 and M2) and oxygen site (O2 and O3) layers, while Ba3WVO8.5 has a completely vacant M2 site and the M1 site is fully occupied.20,30
The inherent oxygen vacancies and the disordering of coordination polyhedra have been shown as the origin for their high oxide ion conduction.25,26,31,32 The anion site disorder has been shown to be related to occupancy and site symmetry of the O3 ion.33–35 Additionally, it has been shown that the alteration of oxygen vacancies and ratio of tetrahedra to octahedra enhances the ionic conductivity,28,34,35 and hence, such materials have been the subject of many recent studies. The variation in conductivity with composition in Ba3W1+xV1−xO8.5+x/2,28,33 Ba3MoNb1−xTaxO8.5−δ,31 Ba3MoNb1−xGaxO8.5−δ32 and Ba3W1+xNb1−xO8.5+x/234 has been explained by these factors. Fop et al. have shown that substitution of a part of Nb5+ of Ba3WNbO8.5 by V5+ enhances the conductivity solely due to the increased disorder by increasing the fraction of tetrahedral MO4 units over octahedral MO6 units.21 The enhancement in conductivity of the Ba3Mo1+xNb1−2xGexO8.5 (x = 0, 0.1, and 0.2) system has been attributed to the increase in the concentration of tetrahedral units.35
Recently, the role of oxygen ion vacancy and stoichiometry of cations in the ionic conductivity of Ba3WVO8.5 has been studied by investigating the structure and conductivity of Ba3W1−xV1+xO8.5−x/2.28 The bulk conductivity of Ba3W1−xVxO8.5+x/2 has been tuned to about 1.2 × 10−4 S cm−1 by decreasing the vanadium content, while under similar conditions, Ba3WVO8.5 had a conductivity of about 2.3 × 10−6 S cm−1 (1023 K).28 Further, it was reported that the conductivity is enhanced by both increasing and decreasing the oxygen stoichiometry in the samples, which has been attributed to the easy migration of ions and the increase in the concentration of the mobile ions. Additionally, it was concluded that the fraction of tetrahedral units increases with the increase in V5+ concentration in the samples. Cheng et al.35 have reported conductivity of about 1.23 × 10−4 S cm−1 for Ba3Mo1.2Nb0.6Ge0.2O8.5 at 673 K, which is about two times higher than that of Ba3MoNbO8.5 (5.8 × 10−5 S cm−1) and Ba3MoNb0.8V0.2O8.5 (7.6 × 10−5 S cm−1) around the same temperature. This suggests that tuning of anion vacancies or introducing excess anions as well as disorders in anion sites can further enhance the ionic conductivity.
In order to substantiate the role of oxygen ion vacancies and the fraction of tetrahedral units, a systematic study was conducted on the structure and dielectric properties of Ba3WV1−xGexO8.5−x/2 (x = 0, 0.1, 0.2, 0.3). To obtain single-phase samples, Ge4+ can be considered as an ideal substituent for V5+ due to their comparable ionic radii. It is known that Ge4+ has a preference for tetrahedral coordination due to a smaller ionic radius, and hence, can change the ratio of M1–O4/M1–O6 in these compounds. Additionally, a systematic increase in oxygen ion vacancies can be introduced by the substitution of V5+ with Ge4+. Moreover, the preference of Ge4+ for forming tetrahedral coordination can also alter the anion disorder in the BaO3 layer of hexagonal perovskite system. By adopting this strategy, a conductivity of 1.80 × 10−5 S cm−1 at 973 K could be achieved at Ba3WV0.7Ge0.3O8.35, while that of Ba3WVO8.5 was 1.03 × 10−7 S cm−1 at a similar temperature. From the analyses of impedance data using equivalent circuit models, it could be revealed that the bulk conductivity of the grains systematically increases with the increase in Ge4+ concentration in the compositions, while the activation energies for conductivity remain at 0.9 to 1.6 eV. In order to understand the conductivity mechanisms, the modulus spectra were analysed, and it was concluded that the distribution of relaxation time is appreciably asymmetric with respect to the mean value of all the samples, which increases with the increase in Ge4+ concentration. This is a characteristic feature of the distribution of spatial arrangement of carriers and the energetics for carrier migration. From the in situ high-temperature XRD studies, structural and compositional stability of all the samples was concluded. The details of the structure and transport properties at high temperatures are explained herein. Such a study can provide further insights into the influence of such substitutions on the structure and conductivity, which is relevant to the possible applications of this class of materials as solid electrolytes in fuel cells.
In order to examine the thermal stability, structural characterization was carried out by in situ high-temperature X-ray diffraction studies. All the high-temperature XRD patterns were recorded using a PROTO AXRD theta–theta X-ray diffractometer with an Anton Paar HTK 1600 high-temperature attachment. Slurries of a finely powdered sample in acetone were smeared on a platinum strip used as a sample holder and heater. The samples were heated to the desired temperature and equilibrated for 5 minutes prior to recording the diffraction data. The XRD patterns were recorded at several temperatures from room temperature to 973 K. The XRD data were analyzed by the Rietveld refinement method using the Fullprof program.36
The measurement of dielectric properties was carried out using an Alpha-AN impedance analyzer (Novocontrol Technologies, Germany) equipped with a Novotherm HT1200 high-temperature attachment. Sintered cylindrical pellets of the samples were coated with platinum paste and annealed at 973 K to ensure proper electrical contact. The platinum-coated pellets were placed in a dielectric sample cell (Probostat, NorECS, Norway) mounted in a tubular furnace, in a parallel-plate capacitor configuration with platinum electrodes. The temperature was varied from 623 K to 973 K. The complex dielectric parameters were measured in the frequency range of 10 Hz–10 MHz. All the measurements were carried out while heating in air and with an applied rms ac voltage of 0.5 V. Data at every temperature were recorded with an error of 0.5 °C in temperature.
Further, the powder XRD data of the samples recorded at 323 K were analysed by the Rietveld refinement method using a structural model based on the reported structural details of Ba3WVO8.5.28,30,33,37 In this model, Ba atoms are located on two atomic sites Ba1 (3a) and Ba2 (6c), while W, V and Ge are on site M1 (6c) according to their stoichiometry. Another possible site for the octahedral cation, M2 (6c), was assumed as a fully vacant site as reported earlier for the Ba3WVO8.5 compound (30). The oxygen ions occupy three sites, namely, O1 (18h), O2 (9e) and O3 (6c). The site O1 is fully occupied while the sites O2 and O3 are under-occupied, hence retaining the intrinsic anion vacancy arising from the composition. For the refinement of data for Ba3WVO8.5, the M1 site was occupied by both W and V with an occupancy of 0.5 for each, while for the Ge-doped samples, the Ge ions are distributed in the M1 site keeping the occupancy of V and Ge totaling to 0.5. Attempts to fill the cations in the M2 site were not successful in the Rietveld refinement as their occupancies turn out to be negative on refinement. Hence, similar to the earlier reports,28,30,33,37 the M2 site is kept vacant while maintaining the cation stoichiometry as the nominal composition of the samples. For the refinement of anions, the O1 (18h) site is considered as a fully occupied site and the rest of the anions as per the stoichiometry are placed at the O2 (9e) site, i.e., occ. 0.833 and O3 site (6c) as 0.0. Subsequently, the O3 sites are occupied by liberating the O2 site anions to O3 sites with a constraint, and the total anions at O2 and O3 sites, i.e., gO2 + gO3 = 7.5 − x/2, are in agreement with the nominal composition. Since the anion stoichiometry decreases due to the substitution of V5+ with Ge4+, the refinement of occupancies of underoccupied oxygen sites was carried out by using a constraint of nominal anion stoichiometry. The occupation of a larger fraction of O2− ions at the 6c site (i.e., O3) compared to 9e (O2) is in accordance with the larger fraction of tetrahedral group compared to the octahedral group in Ba3WVO8.5 and Ba3WV1−xGexO8.5−x/2.
It may be noted here that several possible sites have been proposed for the O3 atoms, viz. 6c (0, 0, z; z ∼ 0.5), 3b (0, 0, 0. 5) or 36i (x,y,z, x ∼ 0, y ∼ 0, and z ∼ 0.5). This variation can be due to the distortion and orientational disorder of the M1O4 tetrahedra. In the present study, the refinement was carried out using O3 atoms at 6c sites, and appreciable displacement from the 3b (special position) along the c-axis is evident.
The typical structure of the hexagonal perovskite-type Ba3WVO8.5 (BWVO) compound, depicting different sites is shown in Fig. 1. It can be seen from the figure that O2 atoms make up the M1O6 octahedra, while the O3 atoms make up the M1O4 tetrahedra. Thus, the octahedral and tetrahedral units of M1 atoms are related to the occupancy of the O2 and O3 atoms. All the XRD patterns of Ba3WV1−xGexO8.5−x/2 (x = 0, 0.1, 0.2, and 0.3) samples could be fitted very well as single-phase rhombohedral (space group: R3m), which suggests the structural stability and formation of the single-phase sample on Ge substitutions. The refined unit cell parameters and residuals for the Rietveld refinements of the XRD data are summarized in Table 1. The Rietveld refinement plots for the XRD data of BWVO (x = 0) and BWVO3 (x = 0.3) samples recorded at 323 K are shown in Fig. 2, and those of samples with x = 0.1 and 0.2 are given in Fig. S3 of the SI. The refined position coordinates and other refinement parameters for BWVO at 323 K are given in Table S1.
Ba3WV1−xGexO8.5−x/2 | ||||
---|---|---|---|---|
BWVO x = 0 | BWVGO1 x = 0.1 | BWVGO2 x = 0.2 | BWVGO3 x = 0.3 | |
323 K | ||||
a (Å) | 5.8230(1) | 5.8292(1) | 5.8395(1) | 5.8529(1) |
c (Å) | 21.1211(4) | 21.1542(2) | 21.1654(2) | 21.1818(2) |
V (Å)3 | 620.22(2) | 622.50(1) | 625.05(1) | 628.40(1) |
χ2 | 1.87 | 1.41 | 1.30 | 1.31 |
Rp (%) | 4.81 | 4.29 | 4.30 | 4.29 |
Rwp (%) | 6.55 | 5.56 | 5.66 | 5.65 |
Rexp (%) | 4.80 | 4.69 | 4.97 | 4.94 |
1023 K | ||||
a (Å) | 5.9027(1) | 5.9083(1) | 5.9196(1) | 5.9332(1) |
c (Å) | 21.2852(4) | 21.3082(3) | 21.3198(2) | 21.3372(3) |
V (Å)3 | 642.26(2) | 644.17(1) | 646.98(1) | 650.51(1) |
χ2 | 1.57 | 1.43 | 1.22 | 1.35 |
Rp (%) | 4.69 | 4.55 | 4.26 | 4.60 |
Rwp (%) | 6.29 | 5.85 | 5.54 | 5.93 |
Rexp (%) | 5.02 | 4.90 | 5.02 | 5.10 |
The thermal and structural stabilities of the samples are crucial factors in determining their suitability as electrolyte materials in solid oxide fuel cells. In this regard, the structural stability and evolution of structural parameters with temperature were studied by analysing the recorded high-temperature XRD data. The XRD patterns recorded at higher temperatures were analysed by the Rietveld refinement method using the observed structural details of the corresponding sample at 323 K. In all the cases, the HT-XRD patterns recorded up to 1023 K could be successfully refined with the considered rhombohedral lattice. This suggests that all the studied materials do not have any phase transition or decomposition within the studied temperature range. The XRD patterns of the BWVGO3 sample show additional peaks due to platinum (Pt), which was used in the sample holder cum heater. Hence, the refinement of the XRD patterns of the BWVGO3 sample was carried out by considering Pt as an additional phase.
Typical Rietveld refinement plots for the XRD data of Ba3WVO8.5 (BWVO), and Ba3WV0.7Ge0.3O8.35 (BWVGO3) samples recorded at 1023 K are shown in Fig. 2. The Rietveld refinement plots of XRD data of samples with x = 0.1 and 0.2 are given in the SI (Fig. S3). The refined structural parameters for all the samples at 1023 K are given in Table S1. The unit cell parameters of Ba3WV1−xGexO8.5−x/2 (x = 0, 0.1, 0.2, and 0.3) samples at 1023 K are presented in Table 1. Appreciably, good residuals were achieved for the refinement of the XRD data recorded at all the temperatures, and the representative residuals for the two temperatures are also given in Table 1.
Fig. 3(a) shows the variation in unit cell parameters of the Ba3WV1−xGexO8.5−x/2 (x = 0, 0.1, 0.2, 0.3) samples with x, i.e. the Ge content. It can be seen that the unit cell parameters (a, c) and the cell volume V increase with the increase in Ge content in the series. This is in accordance with the larger ionic radius of Ge4+ compared to that of V5+ (ionic radii of Ge4+ and V5+ in tetrahedral coordination are 0.390 Å and 0.355 Å, respectively).38 Similar expansion of the unit cell was reported by Cheng et al.35 for the Ba3Mo1+xNb1−2xGexO8.5 series. Compared to the study by Cheng et al.,35 the present studied series showed an increase in unit cell parameters even though there is an increase in anion vacancies. This suggests that the expansion due to the larger ionic radii of Ge4+ ions dominates over the possible contraction due to the anion vacancies in the lattice. The variations in unit cell parameters (a, c and V) of Ba3WV1−xGexO8.5−x/2 (x = 0, 0.1, 0.2, and 0.3) with temperature are shown in Fig. 3(b)–(d). The plots show an almost linear expansion for all the samples with temperature. It was also observed that the temperature-induced lattice expansion is more or less similar for all the studied samples. Hence, the change in the anion vacancy with the increase in Ge concentration might be the governing factor for their ionic conductivity. It is reported in the literature that ionic conductivity in the perovskite-type structure increases as the unit cell volume increases.31,32 Thus, the expanded lattice without any change in the crystal structure of the present studied samples at higher temperatures is likely to facilitate the migration of ions.
![]() | ||
Fig. 3 Variation in the lattice parameters of Ba3WV1−xGexO8.5−x/2 (x = 0, 0.1, 0.2, and 0.3) with x (Ge4+ concentrations) at 323 K (a) and with temperature (b–d). |
All samples exhibit irregular grains of spherical to cubical shapes having sizes in the range of 1 to 5 microns. Moreover, in all these samples, the grains are well formed and interconnected. The elemental distribution and composition were analysed using the EDX spectra collected in the area-scan mode. The EDX spectra for the samples are displayed in the SI (Fig. S4), and they exhibit almost similar spectral characteristics. Additionally, it may be mentioned here that no elements other than those expected are observed in the studied samples. The elemental compositions were determined by EDX analysis, and the details of elemental compositions are presented in Table 2. Furthermore, the EDX analysis indicates a consistent trend in the variation of Germanium (Ge) and Vanadium (V) across the series, which adheres to the stoichiometry relevant to the samples. Further, to support the grain structures, we conducted a TEM study for a representative sample (BWVGO3), and the grains and their connectivity could be visualized with more clarity (Fig. S5 of the SI). The crystalline natures of the grains were also confirmed by EDX patterns.
Sample | Ba atomic % | W atomic % | V atomic % | Ge atomic % | ||||
---|---|---|---|---|---|---|---|---|
Exp. | Obs. | Exp. | Obs. | Exp. | Obs. | Exp. | Obs. | |
BWVO | 60 | 60.5 | 20 | 21.4 | 20 | 18.1 | 0 | 0 |
BWVGO1 | 60 | 62.9 | 20 | 17.4 | 18 | 17.9 | 2 | 1.8 |
BWVGO2 | 60 | 61.3 | 20 | 19.2 | 16 | 15.7 | 4 | 3.8 |
BWVGO3 | 60 | 61.4 | 20 | 19.0 | 14 | 13.8 | 6 | 5.9 |
![]() | ||
Fig. 5 Variation in the real part of complex permittivity (ε′) with frequency at different temperatures for the Ba3WV1−xGexO8.5−x/2 (x = 0, 0.1, 0.2, and 0.3) series. |
Furthermore, it was noticed that the permittivity of the compounds increases with the increase in Ge concentration. The values of ε′(f) at 114 Hz and 823 K were found to be 114, 572, 493 and 1554 for BWVO, BWVGO1, BWVGO2 and BWVGO3, respectively. The variation in permittivity in this series follows the same trend as in their conductivity (discussed later in this section). Thus, the permittivity in these samples is mainly due to polarization caused by the diffusion of charge carriers.
To understand the ion diffusion process, the imaginary part of the permittivity, ε′′(f) (dielectric loss), was extracted and analyzed in a similar manner, as reported in the literature.39,40,44,45 Fig. 6 shows the variation in ε′′(f) with frequency at different temperatures. At lower frequencies, the value of ε′′(f) increases with temperature as in the case of ε′(f). The loss spectra show two distinct features for BWVO and three features for the Ge-doped samples, similar to those observed in ε′(f) plots. The characteristic relaxation frequency of an underlying process is usually observed as a peak in the loss spectrum. However, in the present case, these features are smeared due to the ion hopping conductivity process, which is commonly overlapped and correlated.
σ*(ω) = σ′(ω) + iσ′′(ω) = iωε*(ω) | (1) |
The real part σ′(f) of ac conductivity of the Ba3WV1−xGexO8.5−x/2 (x = 0, 0.1, 0.2, and 0.3) samples is shown in Fig. 7. It can be seen from Fig. 7 that the variation in σ′(f) with frequency shows a plateau region at lower frequencies and a dispersion region at higher frequencies. Further, it can be seen that with the increase in temperature, the plateau region shifts upward and the onset of dispersion region shifts towards a higher frequency. At higher temperatures, a second plateau and a dispersion region, which can be attributed to interfacial polarization at grain boundaries, or to electrode effects, appear at a lower frequency. The high-frequency plateau and dispersion region were used to extract the dc conductivity (as only this region corresponds to the bulk contribution at higher temperature), using the Jonscher Power Law (JPL) given by eqn (2):28,47
σ′(f) = σdc + Afs | (2) |
The extracted bulk dc conductivity shows an increasing trend with the increasing temperature for all samples, and the value of dc conductivity at a particular temperature increases with the increase in Ge concentration in the Ba3WV1−xGexO8.5−x/2 (x = 0, 0.1, 0.2, and 0.3) series. The higher conductivity of the Ge-containing samples can be attributed to the increasing vacancy concentration as well as increasing unit cell volume. However, it is noticed that conductivity for BWVGO1 (x = 0.1) is larger than that of BWVO, while for BWVGO2 (x = 0.2), it shows a lower value than that of BWVGO1. It is again larger for the BWVGO3 sample (x = 0.3). The anomalous behaviour observed in conductivity data extracted from JPL is due to highly overlapping contributions from grain and grain boundaries of the samples. This anomalous behavior could be resolved by the analyses of impedance data using equivalent circuit models, and they are explained later. Strong overlapping of grain and grain-boundary contributions in the conductivity spectra is commonly observed in such hexagonal perovskite systems such as Ba3WNbO8.519 and Ba3W1.2Nb0.8O8.6.34 The values of dc conductivity obtained using JPL are in the order of BWVO < BWVGO2 < BWVGO1 < BWVGO3. The BWVGO3 sample shows the maximum value of σdc in this series, which is about 21 times that of the parent sample (BWVO) at 973 K.
The evolution of dc conductivity (σdc) of the Ba3WV1−xGexO8.5−x/2 samples with temperature is shown in Fig. 8. In all the cases, a discontinuity in the plot of σdc vs. 1000/T, where T is the temperature in Kelvin, is observed at higher temperatures.
![]() | ||
Fig. 8 Plots of σdc vs. 1000/T for the Ba3WV1−xGexO8.5−x/2 (x = 0, 0.1, 0.2, and 0.3) series. The lines are linear fits according to eqn (3). |
In all the samples, the variation in lnσdc shows discontinuity at a certain temperature, and for each case, the data were fitted to two straight lines, one temperature above the discontinuity (region-I, i.e., T > 823 K) and other below the discontinuity (region-II, i.e., T < 823 K). The activation energy (Ea) of all the samples in these two regions was obtained by fitting the linear regions of the plots of ln
σdc versus 1/T data using the Arrhenius relation:28,44,45
σdc = σ0exp(−Ea/kBT) | (3) |
The activation energies for the BWVO, BWVGO1, BWVGO2 and BWVGO3 samples in the region-II are 0.90, 1.07, 0.99 and 1.09 eV, respectively, while in region-I, the activation energies for the corresponding samples are 1.47, 1.14, 2.28 and 1.32 eV. Although there is no clear correlation between the activation energies, various samples were noticed, in each sample the activation energy in the high temperature region (region-I) is found to be higher than that in the lower temperature region (region-II). This may be due to two different types of processes or migrating species, one involving loosely bonded charge carriers that can easily migrate, while the other involving relatively strongly bonded species that get activated only at a higher temperature.
In order to elucidate the contributions of grain and grain-boundary to the total conductivity of Ba3WV1−xGexO8.5−x/2 samples, the frequency-dependent complex impedance Z*(f) = Z′(f) + iZ′′(f) was analyzed using the equivalent circuit method. Fig. 9 shows the complex impedance plots of the Ba3WV1−xGexO8.5−x/2 (x = 0, 0.1, 0.2, and 0.3) samples recorded at different temperatures. For BWVO, a clear arc-like feature could be observed in the spectra only above 823 K, while in the doped samples, noise-free data could be observed from 723 K onwards. These spectra were analyzed using WinFIT software (Novocontrol Technologies, Germany). For all the samples, the complex impedance spectra appeared as depressed semicircles with their centers lying below the X-axis. This indicates that all the samples behave like diffusing capacitors. In an ideal capacitive response, the phase shift in the electrical signal will be π/2 and this gives an ideal semicircle in the impedance plot. However, a phase shift of nπ/2, where n lies between 0 and 1, arises from a diffusive (non-ideal) capacitor. This leads to depressed semicircles, i.e., the center of the semicircle appears below the X-axis, in the complex impedance plot.48 This behavior is generally represented by a parallel combination of a resistance with a constant phase element (CPE), with the impedance given as , where the parameter Y0, expressed in Ohm−1 sn or F sn−1, encapsulates the information regarding capacitance,
, and n ranges from 0 to 1.44,45,48 Therefore, the impedance spectra of all the samples were analyzed using parallel combinations of a resistance R and CPE instead of a capacitor C. For BWVO, the impedance spectra could be well modeled by combining two RCPE (parallel R and CPE) circuits in series. However, to fit the spectra for other samples, three parallel R and CPE (constant phase element) circuits in series representing grain, grain boundary, and electrode polarization have been used. Typical fits of the impedance spectra are shown in Fig. 9. It may be noted that an identical circuit model was used to find out the grain, grain-boundary and electrode resistance by impedance fitting in the study on the Ba7Nb4MoO20 compound.26
Fig. 10 shows the variation in grain (Rg) and grain boundary resistance (Rgb) of all the samples in this series. Both Rg and Rgb decrease with the increase in temperature. It is clear from Fig. 9 that the resistance in the grains systematically decreases with the increase in Ge concentration. However, the resistance of the grain boundary (Rgb) follows the order as Rgb(BWVO) > Rgb(BWVGO2) > Rgb(BWVGO1) > Rgb(BWVGO3). The anomalous behavior observed in the grain boundary of BWVGO1 and BWVGO2 can be due to the segregation of secondary phase or deviation of composition in the sample at the grain boundary, while sintering the sample which could not be detected in both XRD and SEM studies. Such behavior was reported earlier by McCombie et al. for Ba3WNbO8.5.19,34 Thus, the anomalous behavior in the dc-conductivity data extracted by JPL (shown in Fig. 8) can be due to the overlap of grain and grain boundary contributions. A systematic variation of the conductivity of the grains with Ge concentration is in accordance with the increasing oxygen vacancy and the increasing unit cell volume. The dc conductivity of the grains at different temperatures was calculated from the grain resistance (Rg) for all the samples using eqn (4), and the values are shown in Fig. 11.
![]() | (4) |
![]() | ||
Fig. 10 Variation in Rg (grain resistance) and Rgb (grain boundary resistance) with temperature and composition. Lines are guides to the eyes. |
![]() | ||
Fig. 11 Bulk (grain) dc conductivity vs. 1000/T (K) for all the samples obtained from the complex impedance analysis. The lines are linear fits according to eqn (3). |
The values of grain dc conductivity for BWVO, BWVGO1, BWVGO2 and BWVGO3 at 973 K are 1.03 × 10−7, 4.83 × 10−6, 2.94 × 10−6 and 1.80 × 10−5 S cm−1, respectively. A systematically increasing trend in conductivity with the increase in Ge concentration is in accordance with the increasing anion vacancies and ease of hopping of ions from occupied to vacant sites. The temperature dependence of the grain conductivity could be fit very well with the Arrhenius relation (eqn (3)), and the estimated activation energy for the samples lies in the range of 1.4 eV to 1.6 eV. However, no change in slope was noticed in variation of σdc of grains conductivity as that observed in σdc obtained from JPL presented earlier. Thus, the anomalous behaviors of BWVGO1 and BWVGO2 explained earlier are due to the dominating contributions of grain boundaries.
![]() | (5) |
Physically, the complex electrical modulus is the frequency-domain representation of the evolution of the time-dependent electric field inside the material. The frequency- and temperature-dependent imaginary parts of the modulus spectra can be used to understand the conduction process in the sample. Usually, ionic conductors show broad peaks in the imaginary part (M′′(f)) of the modulus spectra due to a wider distribution of times of the relaxation processes in the material.49 The variations in M′′(f) with frequency and temperature for the samples studied are shown in Fig. 12.
![]() | ||
Fig. 12 Imaginary part of the complex electrical modulus for the samples at different temperatures. Red lines are fits to the H–N functional form according to eqn (6). |
In all the samples, the M′′(f) vs. f plots show broadened peaks that shift towards a higher frequency with the increase in temperature. This is indicative of the ion hopping process with a wide relaxation time distribution. Moreover, it can be seen in Fig. 12 that at a given temperature, the peaks in modulus spectra shift toward a higher frequency with the increase in the concentration of Ge in BWVO. This suggests that the BWVGO3 sample has the shortest mean relaxation time and exhibits the ion hopping process at the shortest characteristic time scale.
To further probe the relaxation in the samples, the M′′(f) spectra were fit with the Havriliak–Negami (H–N) relaxation functional form as given in eqn (6):47
![]() | (6) |
The temperature dependence of relaxation time τM obtained by fitting eqn (6) to the M′′(f) spectra is shown in Fig. 13, and it is found to follow the Arrhenius relation28 as given below:
τM = τ0exp(−EM/kBT) | (7) |
![]() | ||
Fig. 13 Temperature dependence of the relaxation time (τM) derived from the modulus spectra. The solid lines represent linear fits using eqn (7). |
Similar to the variation in σdc with temperature (Fig. 8), a discontinuity in the variation of τM was observed in all compounds except BWVO. The activation energies (EM) were extracted by fitting the conductivity relaxation time τM data shown in Fig. 13 with eqn (7) over the two different linear regions, namely region I (T > 823 K) and region II (T < 823 K). The extracted activation energy values for the BWVO, BWVGO1, BWVGO2 and BWVGO3 samples in region-I are 1.43, 1.36, 1.04 and 1.54 eV, respectively. In region-II, the activation energies for BWVGO1, BWVGO2 and BWVGO3 are 1.04, 1.14 and 1.11 eV, respectively. In the case of BWVO, the relaxation frequency in the low-temperature region (region-II) could not be followed due to its noisy modulus spectra. However, the overall trend of EM for the samples is similar to that seen for the σdc data extracted from JPL analyses and the differences may be due to errors in fitting.
The shape parameters (α and β) provide a significant understanding of the nature of relaxation. For an ideal Debye-type system, both α and β should be equal to 1, while for a non-Debye type relaxation, these parameters take values less than 1. The values α and β at every temperature, obtained by fitting the electrical modulus data of Fig. 12 with eqn (6), are shown in Fig. 14.
![]() | ||
Fig. 14 Variation in the shape parameters α and β obtained by fitting the electrical modulus spectra (Fig. 12) to the H–N function (eqn (6)) for all the samples. Solid lines are guides to the eye. |
It can be seen that the values of α for the samples do not show any appreciable variations with temperature, while a decreasing trend beyond certain temperature is noticed in BWVGO3. Moreover, the values of α for Ge-containing samples are lower than that of the undoped sample (BWVO). Since the shape parameter α is related to the symmetrical broadening of the M′′(f) peak, and it represents the spread of relaxation times, it can be inferred that the spread of relaxation times is greater in the Ge-substituted compounds than that in BWVO. The shape parameter β is related to the asymmetry of relaxation time distribution about a mean relaxation time. It can be seen from Fig. 14 that β remains nearly the same, except a slight increase for the BWVGO3 sample beyond 900 K, and is much smaller than 1 in all the samples over the entire temperature range studied. The degree of asymmetry of modulus spectra represented by shape parameter β (called asymmetry parameter) is associated with the ionic conduction mechanism. In all the studied samples, the values of β are appreciably lower then 1, which indicates that the distribution of relaxation time is appreciably asymmetric with respect to a mean value, which is the characteristic feature of the distribution of spatial arrangement of carriers and the energetics for carrier migration. The lower β values suggest that ionic hopping depends on not only the average energy barrier but also the range of energy barriers resulting from the differences in migration path, lattice or local distortion, and the binding strength of mobile ions. Moreover, the closer values of β at a lower temperature for all the samples indicate similar mechanisms of ion migration in all of them, which may be due to similar energy barriers and hopping distances in their lattices. In the present studied system, the replacement of V5+ with Ge4+ not only generates oxygen vacancies but also creates local distortion due to the preference for the formation of tetrahedra rather than octahedra by the Ge4+ ions. This may alter the local distortion, and hence, ion hopping mechanism as well as the barrier energy for hopping the lattice. However, at higher temperatures, the sample with x = 0.3, i.e., BWVGO3, shows a slight increasing trend, which can be due to excessive O2− ion vacancies, which may lead to the opening of more hopping paths than others.
From the studies on ion transport by JPS and impedance spectral analyses, it was evidenced that the ionic conductivity increases on substitution of Ge4+ in the lattice of Ba3WVO8.5. Earlier Fop et al.21 have shown that Ba3WVO8.5 can show a bulk conductivity as high as 2.0 × 10−3 S cm−1 at 1173 K, which is about one order lower than the stabilized zirconia around that temperature. Moreover, the authors have indicated that the increased ionic conductivity of Ba3Nb1−xVxO8.5 with the V5+ ion concentration even without alteration in anion stoichiometry is due to the preferential formation of tetrahedral VO4 units in the structure, which results in excessive disordered anion sites due to the displacement of anions. The conductivity of about 0.01 S cm−1 has been shown in the composition Ba3Nb0.9V0.1O8.5 at 873 K.21 Studies of Cheng et al.35 on Ba3Mo1+xNb1−2xGexO8.5 indicate the conductivity of about 1.23 × 10−4 S cm−1 in the composition Ba3Mo1.2Nb0.6Ge0.2O8.5 at 400 °C, which is about two times higher than that of Ba3MoNbO8.5 (5.8 × 10−5 S cm−1) and Ba3MoNb0.8V0.2O8.5 (7.6 × 10−5 S cm−1) at around a similar temperature. This has been attributed to the tetrahedral preference of Ge4+ ions, as concluded by Fop et al.18,21 In the studies on the Ba3W1+xV1−xO8.5+x/2 system, a bulk conductivity of about 1.2 × 10−4 S cm−1 at 750 °C has been shown in the composition Ba3W1.1V0.9O8.55, while under similar conditions, the pristine Ba3WVO8.5 shows a conductivity of about two-order lower value (2.3 × 10−6 S cm−1).28 Thus, it is apparent that the tuning of anion vacancies or introducing excess anions as well as disorder in anion sites can enhance the ionic conductivity. In this study, it was observed that the conductivity increased with the increase in Ge4+ concentration in Ba3WV1−xGexO8.5−x/2 compositions, and a conductivity of about 1.80 × 10−5 S cm−1 was observed for the BWVGO3 sample at 700 °C, while at a similar temperature, the observed conductivity of Ba3WVO8.5 was about 1.03 × 10−7 S cm−1. This can be due to the synergetic effects of introduction of anion vacancies, dilation of lattices and disordered anions in the BaO3 layer of the Ba3WVO8.5 lattice due to the preferential formation of GeO4 tetrahedra. Thus, it can be suggested that the ionic conductivity of such hexagonal perovskites can be enhanced by altering the mobile anion concentration, anion vacancies, disorder in anion and cation sites, lattice distortion, nature and stability of coordination polyhedra around the smaller cations, etc.
The improved oxide ion conductivity noted in Ge4+-substituted Ba3WVO8.5 materials, along with their structural and thermal stability up to 1023 K, positions these compounds as promising candidates for application as electrolytes in intermediate-temperature solid oxide fuel cells (IT-SOFCs). The existence of adjustable oxygen vacancies and modes to influence conduction pathways via aliovalent substitution (Ge4+ replacing V5+) may facilitate the development of customized ionic conductors. Furthermore, the details of composition, structure and electrical properties presented herein will be useful to design materials for intermediate-temperature solid oxide fuel cells, oxygen sensors, and ionic switching devices.
The powder XRD data of Ba3WVO8.5 (BWVO) (Cu Kα X-ray) indexed according to the JCPDS data, comparison of the powder XRD data of all the studied composition, Rietveld refinement plots for BWVGO1 and BWVGO2 at 323 K and 1023 K, refined structural parameters of all the BWVGO samples at 323 and 1023 K, EDX spectra BWVGO1, BWVGO2 and BWVGO3 samples and TEM image and ED pattern for BWVGO3. See DOI: https://doi.org/10.1039/d5tc01785h
This journal is © The Royal Society of Chemistry 2025 |