Nikolay
Gorshkov
^{a},
Egor
Baldin
^{b},
Dmitry
Stolbov
^{c},
Galina
Vorobieva
^{b},
Alexander
Shatov
^{b} and
Anna
Shlyakhtina
*^{b}
^{a}Yuri Gagarin State Technical University of Saratov, Saratov, Russia
^{b}N.N. Semenov Federal Research Center for Chemical Physics, Russian Academy of Sciences, Moscow, Russia. E-mail: annash@chph.ras.ru; annashl@inbox.ru
^{c}Department of Chemistry, Lomonosov Moscow State University, Moscow, Russia
First published on 6th August 2024
Solid solutions of rare earth titanates with high contents of rare earth oxides of up to 50–62% have been synthesized by the co-precipitation method and their structure, microstructure and conductivity in dry and wet air have been studied. Proton conductors have been found for the first time in solid solutions of rare earth titanates with a high content of Ln_{2}O_{3} (>50%) with a nominal formula composition of (Ln_{x}Ti_{1−x})_{4}O_{8−2x} (Ln = Yb, Er, Ho, 0.667 ≤ x ≤ 0.765). Among (Ln_{x}Ti_{1−x})_{4}O_{8−2x} (Ln = Yb, Er, Ho, x = 0.684), (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.684) showed the maximum conductivity in wet air. In this context, four additional compositions (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.718, 0.734, 0.75, and 0.765) were synthesized in the holmium series. An increase in the holmium content leads to an increase in the proton transfer coefficients; at the same time, a more complex nature of the dependence of the conductivity under dry and wet atmospheres is observed. For the fluorite-like solid solution (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (0.701 ≤ x ≤ 0.765), the proton transfer coefficients were found to be ∼0.9 in the range of 200–450 °C. As the temperature continues to rise, the proton conductivity decreases quite sharply and the transfer coefficient becomes as low as 0.3 at 700 °C. The increase in proton conductivity in the Yb–Er–Ho series is associated with an increase in the hydrophilic properties of rare earth cations. In the (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.667 ≤ x ≤ 0.765) series, the conductivity in wet air was ∼1 × 10^{−6} S cm^{−1} at 450 °C for most compositions. The conductivity of ceramics with x = 0.701 and 0.75 is about 2 times higher, which may be due to the optimal size of pyrochlore nanodomains in the fluorite matrix for x = 0.701 and the formation of pure fluorite for x = 0.75, respectively.
REE (rare-earth elements) titanates with a pyrochlore structure are known to be oxygen ion conductors, as are REE zirconates and hafnates, and have demonstrated high oxygen ion conductivity values. Doping with acceptor impurities significantly increased the conductivity of heavy rare earth titanates.^{8–11} In particular, for Yb_{2}Ti_{2}O_{7} and solid solutions based on it, the oxygen-ion conductivity reached a value of 0.02 S cm^{−1} at 740 °C.^{11} As for the fluorite-like “stuffed” pyrochlore Ln_{2}(Ti_{2−x}Ln_{x})O_{7−δ} (x = 0–0.68), in which some of the titanium positions are replaced by lanthanides, their magnetic properties,^{12–14} resistance to radiation effects^{15–18} and overall high temperature conductivity^{19–21} have been studied. Significant progress has been made in understanding the structure of titanates Ln_{2}TiO_{5}, in the context of their investigation as potential materials for nuclear waste disposal.^{13,22–24} It has been shown that the cubic and orthorhombic modifications of Ln_{2}TiO_{5} retain its crystallinity under irradiation, in contrast to Ln_{2}Ti_{2}O_{7} with a pyrochlore structure.^{25}
Polymorphism of Ln_{2}TiO_{5} compounds depends on the ionic radius of the lanthanide.^{26,27} It is known that middle rare earth titanates Ln_{2}TiO_{5} (Ln = Ho, Dy, Tb) exhibit the richest polymorphism and can crystallize in three polymorphic modifications: high-temperature cubic (fluorite Fmm), hexagonal (P6_{3}/mmc) and low-temperature orthorhombic (Pnma). Note that Tb_{2}TiO_{5} and Dy_{2}TiO_{5} are on the stability limit of the cubic phase. For example, Dy_{2}TiO_{5} fluorite was obtained in its pure form by rapid cooling at a rate of 30–50° min^{−1} from a temperature of 1630 °C.^{28} It is interesting to note that although Ln_{2}TiO_{5} titanates are fluorites (Fmm) in the long-range order (X-ray diffraction data), electron diffraction studies have shown that they contain nanosized domains of pyrochlore (Fmm) in the short-range order.^{14,16,17,29} It is therefore difficult to obtain pure Ln_{2}TiO_{5} (Ln = Ho, Dy, Tb) fluorites due to the rich polymorphism.
They tend to contain an admixture of the hexagonal phase and/or a nanosized phase of pyrochlore.^{26} Cubic fluorite phases are characteristics of most titanates from the end of the rare earth series Ln_{2}TiO_{5} (Ln = Er–Lu).^{26} For titanates of the large rare earths Ln_{2}TiO_{5} (Ln = La–Eu), orthorhombic modification occurs at low temperatures, but sometimes in a mixture with the cubic phase.^{19} No compounds with an orthorhombic structure have been obtained for Ln_{2}TiO_{5} (Ln = Ho–Lu).^{28} It has been found that the larger the lanthanide, the more obvious the presence of the pyrochlore phase in the fluorite matrix.^{28} The number of pyrochlore domains and their size also depend on the conditions of ceramic synthesis, in particular the cooling conditions. Slow cooling leads to the growth of pyrochlore or hexagonal phase domains in the fluorite matrix, which can be identified by XRD.^{17,30}
On studying the pyrochlore–fluorite disorder in the series of “stuffed” pyrochlore from Dy_{2}Ti_{2}O_{7} to Dy_{2}(Dy_{0.677}Ti_{1.33})O_{6.677} by ion-irradiation and transmission electron microscopy characterization methods, it was discovered that the initial pyrochlore Dy_{2}Ti_{2}O_{7} was also partially disordered, i.e. contained fluorite nanodomains.^{28} Thus, most compounds of the Ln_{2}TiO_{5} series (Ln = Tb–Lu) are typically multiphase and contain a mixture of polymorphic modifications in the form of nano-sized^{13–15} or microcrystalline phases.^{28}
Among the rare earth titanates studied, proton conductivity is only observed in acceptor-doped samarium titanates with a pyrochlore structure: Sm_{2}Ti_{1.92}Y_{0.08}O_{7−δ} and Sm_{1.92}Ca_{0.08}Ti_{2}O_{7−δ}.^{31} We emphasize that the least explored physical properties of solid solutions with Ln_{2}O_{3} concentrations above 50% in the literature, i.e. the interval for studying different properties is limited to the Ln_{2}(Ti_{2−x}Ln_{x})O_{7−δ} range (x = 0–0.68). Recently, in an impedance spectroscopy study of the series of Tm_{2}(Ti_{2−x}Tm_{x})O_{7−δ} (x = 0, 0.1, 0.18, 0.28) and (Tm_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.684) prepared by co-precipitation followed by high temperature firing, proton conductivity was discovered for a fluorite-like composition with a high degree of titanium substitution by thulium (Tm_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.684) (52% Tm_{2}O_{3}).^{32} The proton transfer coefficients were ∼0.65 in the temperature range of 25–450 °C and decreased to 0.1 as the temperature increased to 700 °C. Thus, like solid solutions based on the pyrochlore Sm_{2}Ti_{2}O_{7}, (Tm_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.684) has a significant proton contribution in the range of 300–400 °C. Note that “stuffed” pyrochlores Tm_{2}(Ti_{2−x}Tm_{x})O_{7−δ} (x = 0, 0.1, 0.18, 0.28) with a low degree of thulium substitution for titanium had no proton contribution. Therefore, there is a need to study the proton conductivity of the whole series of rare earth titanates with a fluorite-like structure. In this work, titanates of medium and heavy lanthanides (Ln_{x}Ti_{1−x})_{4}O_{8−2x} (Ln = Yb, Er, Ho; x = 0.684) are studied in dry and wet air. In the holmium system, a series of (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (0.667 ≤ x ≤ 0.765) solid solutions are synthesized and studied to identify the main trend in the change of the proton component with increasing holmium oxide content. All ceramics were obtained by the co-precipitation method, which allows the synthesis temperature of titanates to be significantly reduced compared to the solid phase synthesis method.^{32,33} In accordance with literature data, the purest fluorites are expected to be heavy lanthanide titanates (Ln_{x}Ti_{1−x})_{4}O_{8−2x} (Ln = Er, Yb; x = 0.667).^{28}
X-ray diffraction spectra of ceramics ground into powder were obtained at room temperature using the Rigaku Smartlab SE X-ray diffractometer (Cu Kα radiation, λ = 1.5418 Å, Bragg-reflection geometry, 40 kV, 50 mA; 2θ range was 10° to 70°, scan step 0.01°, scan rate 5° min^{−1}) in continuous mode. Rietveld refinement was carried out using the SmartLab Studio II software.
The local structure of ceramic materials under investigation was studied by Raman spectroscopy using a SENTERRA Raman microscope spectrometer (Bruker) with excitation at 785 nm.
The microstructure of the ceramic samples was examined using scanning electron microscopy (SEM) on a JEOL JSM-6390LA.
Thermogravimetric studies were carried out on a simultaneous thermal analysis device STA 449C (“NETZSCH”, Germany) at temperatures up to 1000 °C. The sample heating rate was 10 °C min^{−1}. Measurements were carried out under an oxygen atmosphere on a sample with a fluorite structure (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.75): before and after exposure to distilled water at room temperature for one day and two weeks.
Samples for impedance measurements were prepared in the form of discs with a diameter of 7.5 to 8.0 mm and a thickness of 0.7 to 1.5 mm, which were sintered at a temperature of 1600 °C for 4 hours. Pt paste (“ChemPur”, Germany) was used for the contacts. Impedance measurements under excitation of an AC electric field with an amplitude of 100 mV were performed using an impedance meter (Alpha-A Novocontrol, Novocontrol Technologies GmbH & Co. KG, Germany) in the AC frequency range of 10^{−1}–10^{6} Hz. Measurements were made in 50 °C increments over a range of 250–800 °C with isothermal exposure of the sample at each point for 2 hours to achieve thermal equilibrium. Impedance measurements under a dry atmosphere were accompanied by blowing the cell with dried air with a relative humidity of less than 1% at 25 °C, and in a wet atmosphere by blowing with humidified air with a relative humidity of 91% at 25 °C. The parameters of equivalent circuits were constructed using the EIS Spectrum Analyser program (Research Institute of Physical and Chemical Problems, Belarus) using the Powell algorithm.^{34}
Fig. 1 Diffraction patterns of fluorite-like solid solutions (Ln_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.684). |
Ln | a = 2a_{F}, Å | Phases | R _{wp}, % | χ ^{2} | Geometric density of the ceramic, (g cm^{−3}) | Relative density, % |
---|---|---|---|---|---|---|
Yb | 10.1983(8) | F | 10.77 | 1.59 | 7.28 | 91.8 |
Er | 10.267(7) | F + P + P2 + H | 15.32 | 2.798 | 6.45 | 85 |
Ho | 10.308(8) | F + P + H + P2 | 14.45 | 1.25 | 6.29 | — |
Fig. 2 shows the diffraction patterns of a series of holmium titanates (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (0.667 ≤ x ≤ 0.765). The (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.667) or Ho_{2}TiO_{5} compound is a mixture of two phases: (1) hexagonal phase β-Ho_{2}TiO_{5} with cell parameters a = 3.6135(4) Å, c = 11.91642 Å, ≈75 wt% and (2) cubic fluorite phase ≈25 wt%. Unfortunately, the presence of the pyrochlore P phase in this sample cannot be accurately assessed because its characteristic peaks (111) and (311) overlap with the hexagonal phase peaks and the next isolated peak (331) is only slightly above noise. In the holmium system, according to preliminary studies,^{26} it is impossible to obtain pure fluorite Ho_{2}TiO_{5} and it will always contain an admixture of the hexagonal phase and/or pyrochlore. The (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.684) sample was described in the previous paragraph. With increasing holmium content in the titanate, an increase in the unit cell parameter of the main fluorite phase is observed (Table 2). The weak lines of pyrochlore P in the diffraction pattern (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.701) are significantly broadened, indicating that this phase is nanosized. From the width at half maximum of the (311) peak, the size of the crystallites of the pyrochlore phase was estimated using the Scherrer formula and was 46.7 ± 5.5 nm. A further increase in the holmium content of the titanate leads to a decrease in the intensity of the diffraction lines of the nanopyrochlore. In the diffraction pattern of the (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.750) or Ho_{3}TiO_{6.5} sample, the nanopyrochlore lines are completely indistinguishable and it can be argued that it is pure fluorite in the long-range order. The sample (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.765) also contains no traces of nanopyrochlore, but Ho_{2}O_{3} impurities with a bixbyite structure were found in its composition (in Fig. 2 the corresponding peaks are marked with the symbol C). Thus, (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.750) is the only composition that crystallizes to pure fluorite from co-precipitated precursors after annealing at 1600 °C. Rietveld refinement of the XRD pattern of (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.684) and (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.75) (Fig. S1a and S1b†) is presented in the ESI.†
Fig. 2 Diffraction patterns of compounds in a series (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (0.667 ≤ x ≤ 0.765). Symbols: H – hexagonal phase; P – pyrochlore phase; C – bixbyite phase. |
x in (Ho_{x}Ti_{1−x})_{4}O_{8−2x} | Ln_{2}O_{3} content, mol % | a _{F}, Å | Additional phases | R _{wp}, % | χ ^{2} | Geometric density of the ceramics, (g cm^{−3}) | Relative density,^{b} % |
---|---|---|---|---|---|---|---|
a Hydrostatic weighing density. b The relative density was calculated by neglecting impurity phases and considering them either pure fluorite or pure hexagonal phase. | |||||||
0.667 | 50 | 5.1520(4) | H (≈75%) | 14.06 | 2.83 | 6.92 | 92.5 |
0.684 | 52 | 5.15440 | P (≈20%); H (≈9%); P2 (≈4%) | 14.45 | 1.25 | 6.29 | — |
0.701 | 54 | 5.16526 | Nano P | 11.66 | 4.23 | 6.64 | 88.1 |
0.718 | 56 | 5.185017(9) | 8.06 | 1.66 | 6.28 | 82.76 | |
0.734 | 58 | 5.173302(13) | 9.43 | 2.12 | 6.31 | 82.82 | |
0.750 | 60 | 5.19547(2) | — | 8.11 | 1.88 | 6.46/6.62^{a} | 86.5 |
0.765 | 62 | 5.20369(8) | Ho_{2}O_{3} (≈9%) | 10.21 | 2.28 | 6.66 | 86.47 |
All prepared holmium titanates (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (0.667 ≤ x ≤ 0.765) as well as the sample (Yb_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.684) were investigated by Raman spectroscopy (Fig. 3). As the holmium content of the titanates increases, the bands become wider and individual peaks are difficult to distinguish. The ordering of the Ho and Ti cations, characteristic of the pyrochlore structure, is lost and the structure becomes disordered fluorite, not only in the long-range order but also in the short-range order. At the same time, the spectra of (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.750) and (Yb_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.684), which are the purest fluorites according to the XRD data, practically duplicate each other. An additional narrow band can be seen in the spectrum of (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.765) at 372 cm^{−1}. This band is characteristic of oxides with a bixbyite structure and belongs to Ho_{2}O_{3}.^{37}
Unexpectedly, in the Raman spectra of the ceramics (Ho_{x}Ti_{1−x})_{4}O_{8−2x} with x = (0.677, 0.684, 0.701, 0.718, and 0.734), the pyrochlore bands were found to be weakly expressed, while the origin of most of the strong lines remained unclear.
The main characteristic Raman bands of the pyrochlore single crystal Ho_{2}Ti_{2}O_{7} are located at 311.09 cm^{−1} (vibration (F2g) associated with Ho–O stretching) and 522.16 cm^{−1} (vibration (A1g) associated with Ti–O stretching) according to ref. 38, where Raman spectra of single crystals of the whole series of pyrochlore rare earth titanates RE_{2}Ti_{2}O_{7} (RE = Sm–Lu, Y) were obtained. In Fig. 3 we observe in holmium ceramics a consistent extinction of bands close to the above-mentioned ∼315 and ∼500 cm^{−1} with increasing Ho_{2}O_{3} content in samples (Ho_{x}Ti_{1−x})_{4}O_{8−2x} with x = 0.667, 0.684, 0.701, and 0.718. According to ref. 39, where the Raman spectra of Y_{2}Ti_{2}O_{7} pyrochlore were studied (the ionic radii of Ho^{3+} and Y^{3+} in the octahedral environment are close: R Ho_{CN=8}^{3+} = 1.015 Å; R Y_{CN=8}^{3+} = 1.019 Å), a weaker band at ∼225 cm^{−1} appeared, also associated with Y–O stretching (F2g). In Fig. 3, for the (Ho_{x}Ti_{1−x})_{4}O_{8−2x} with x = 0.667, 0.684, 0.701, and 0.718 cm^{−1}, we note the consistent extinction of a similar band at ∼240 cm^{−1}. For ceramics with a higher Ho content, (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.734, 0.75, 0.765), all pyrochlore bands disappear completely.
An increase in the Ho content in (Ho_{x}Ti_{1−x})_{4}O_{8−2x} with x = (0.667–0.734) solid solutions leads to a slight broadening of the main Raman bands and a slight shift of all spectral lines, but the spectrum changes dramatically for (Ho_{x}Ti_{1−x})_{4}O_{8−2x} with x = (0.75), which is pure fluorite according to X-ray diffraction data (Fig. 2). It is clearly seen that the splitting into two bands at 716 and 759 cm^{−1} for the compositions (Ho_{x}Ti_{1−x})_{4}O_{8−2x} with x = (0.667–0.734) disappears completely during the formation of pure fluorite (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.75). In this case a broadband is formed with a maximum at 743 cm^{−1}. Similar behavior has been observed for fluorite-like compositions in the Y_{2}O_{3}–TiO_{2} system.^{40} With increasing Y_{2}O_{3} content, the characteristic splitting of the Raman bands of fluorite Y_{2}TiO_{5} into two components at 727 and 760 cm^{−1} disappears for (Y_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.75). One of the reasons for such splitting has been suggested in ref. 40 as residual internal stresses in the unit cells of related phases when they coexist in micro- and nanodomains. In addition, it should be noted that the strong bands at 292, 353, and 397 cm^{−1} characteristic of the compositions (Ho_{x}Ti_{1−x})_{4}O_{8−2x} with x = (0.667–0.734) also completely merge into one broadband for pure fluorite (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.75) (Fig. 3). Thus, the typical Raman spectrum of pure fluorites (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.75) and (Yb_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.684) consists of 2 broad bands in the region 240–459 cm^{−1} and 670–800 cm^{−1} with a pronounced maximum at 743 cm^{−1}. Strong Raman bands at 143, 166, 203, 292, 353, 397, and 743 cm^{−1} apparently belong to the hexagonal phase (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.667). This is in agreement with the X-ray diffraction data (Fig. 2). Probably, the local hexagonal phase is also preserved in Ho_{2}O_{3}-enriched compositions (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.684, 0.701, 0.718, 0.734).
SEM images illustrating the surface morphology of the ceramics discussed are shown in Fig. 4. The surface microstructure of erbium ceramics with the composition (Er_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.684) is similar to the microstructure of thulium ceramics with the same composition previously investigated.^{25} The grains are elongated in one direction and lie close together. Ytterbium ceramics are different from erbium and thulium ceramics, but similar motifs can be seen in them. All holmium ceramics, with the exception of (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.667), which is predominantly hexagonal according to XRD data, are characterised by the formation of large crystallites of the order of 20–40 μm. For two samples of the holmium series (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.701 and 0.75) with high total conductivity (Part 3.3.), extended defects are visible within large crystallites.
Fig. 4 SEM images of (Ln_{x}Ti_{1−x})_{4}O_{8−2x} (0.667 ≤ x ≤ 0.765) SEM image of the surface of the ceramic under investigation. |
On the surface of the (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.765) sample, which contains the highest amount of holmium among those examined, smaller grains of ≈2–5 μm can be observed between and inside large grains. It can be assumed that they belong to the Ho_{2}O_{3} detected in the XRD data. While in all the other samples examined the Ho and Ti cations are evenly distributed, for (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.765) the X-ray mapping shows places where there is no titanium but holmium (Fig. S2†). Notable features of all single-phase materials include low porosity.
Fig. 5 Impedance spectra for pure fluorite (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.684) in dry and wet air at 450 °C and 400 °C. B – grain bulk impedance, GB – grain boundary impedance. |
The temperature dependences of the total conductivity of the (Ln_{x}Ti_{1−x})_{4}O_{8−2x} (Ln = Yb, Er, Ho; x = 0.684) series of ceramics are shown in Fig. 6a. Clearly, conductivity increases with lanthanide radius in both dry and wet air. The proton conductivity is higher for the (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.684) solid solution and decreases successively for (Er_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.684) and (Yb_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.684). Fig. 6a demonstrates that the proton contribution remains up to 700 °C. A similar dependence has been observed for cubic molybdates with a bixbyite structure Ln_{6}MoO_{12−δ} (Ln = Er, Tm, Yb).^{41} The proton conductivity contribution exists in cubic Ln_{6}MoO_{12−δ} (Ln = Er, Tm, Yb) bixbyites up to 450–600 °C and decreases with the decrease of the lanthanide ionic radius.
Fig. 6 (a) Total conductivity as a function of temperature in dry and wet atmospheres and (b) proton transfer coefficients in (Ln_{x}Ti_{1−x})_{4}O_{8−2x} (Ln = Yb, Er, Ho; x = 0.684). |
When Ti^{+4} is replaced by Ln^{+3} (Yb, Er, Ho) cations in solid solutions of rare earth titanates with the formula (Ln_{x}Ti_{1−x})_{4}O_{8–2x}, oxygen vacancies appear:
(1) |
Here, titanate, as a complex oxide, undergoes classic quasi-chemical reactions with atmospheric oxygen (2) and water vapour (3):
(2) |
(3) |
Reactions (2) and (3) are not independent because the concentrations of holes, protons and oxygen vacancies are related by the condition of electrical neutrality, so the dissolution of water vapour, which reduces the concentration of vacancies, also reduces the concentration of holes. The total conductivity of the studied REE titanates is characterized by a mixed type and can be expressed in the general case as:
σ = σ_{ion} + σ_{h}, | (4) |
Under a dry air atmosphere (pO_{2} = 0.21 atm), the ionic conductivity σ^{dry}_{ion} is determined by the contribution of the oxygen conductivity, then the total conductivity takes the form:
σ^{dry} = σ^{2−}_{O} + σ_{h}, | (5) |
One can see that the conductivity values under a wet atmosphere are higher than those under a dry atmosphere. The increase in total conductivity under a wet atmosphere is due to the contribution of the proton conductivity to the ionic conductivity (σ^{wet}_{ion} = σ^{2−}_{O} + σ_{H}), where σ_{H} is the proton conductivity. At lower temperatures, the differences in conductivity between the dry and wet atmospheres become more pronounced. This indicates a hydration process leading to an increase in proton concentration. The analysis of the impedance data allowed us to obtain data σ^{dry} under a dry atmosphere and σ^{wet} under a wet atmosphere. Despite the mutual influence of reactions (2) and (3), the equation has been used to estimate proton transfer numbers, assuming that the difference in hole conductivity under dry and moist environments of atmospheric air can be neglected:
(6) |
A full description of the transport properties of the ceramics studied is given in Table 3. The activation energy of proton conduction in the temperature range of 450–200 °C decreases consistently with increasing ionic radius Ln.
Composition | (Ln_{x}Ti_{1−x})_{4}O_{8−2x} | t _{H} ^{eff} at 450 °C | Bulk conductivity at 450 °C/S cm^{−1} | Apparent activation energy (E_{a}) of bulk conductivity/eV | ||
---|---|---|---|---|---|---|
50.0% Ho_{2}O_{3} + 50.0% TiO_{2} | Ln = Ho; x = 0.667 | 0.75 | Dry air | 1.20 × 10^{−7} | 800–550 °C | 1.199 |
550–250 °C | 1.073 | |||||
Wet air | 4.83 × 10^{−7} | 800–600 °C | 1.200 | |||
600–250 °C | 0.944 | |||||
52.0% Ho_{2}O_{3} + 48.0% TiO_{2} | Ln = Ho; x = 0.684 | 0.74 | Dry air | 1.97 × 10^{−7} | 800–550 °C | 1.323 |
550–250 °C | 1.017 | |||||
Wet air | 7.47 × 10^{−7} | 800–600 °C | 1.150 | |||
600–250 °C | 0.878 | |||||
54.0% Ho_{2}O_{3} + 46.0% TiO_{2} | Ln = Ho; x = 0.701 | 0.80 | Dry air | 1.50 × 10^{−7} | 800–500 °C | 1.307 |
500–250 °C | 0.867 | |||||
Wet air | 7.75 × 10^{−7} | 800–600 °C | 1.158 | |||
600–250 °C | 0.745 | |||||
56.0% Ho_{2}O_{3} + 44.0% TiO_{2} | Ln = Ho; x = 0.718 | 0.81 | Dry air | 0.88 × 10^{−7} | 800–500 °C | 1.339 |
500–250 °C | 0.952 | |||||
Wet air | 4.68 × 10^{−7} | 800–600 °C | 1.201 | |||
600–250 °C | 0.761 | |||||
58.0% Ho_{2}O_{3} + 42.0% TiO_{2} | Ln = Ho; x = 0.734 | 0.82 | Dry air | 1.14 × 10^{−7} | 800–550 °C | 1.365 |
550–250 °C | 1.008 | |||||
Wet air | 6.26 × 10^{−7} | 800–600 °C | 1.148 | |||
600–250 °C | 0.778 | |||||
60.0% Ho_{2}O_{3} + 40.0% TiO_{2} | Ln = Ho; x = 0.750 | 0.83 | Dry air | 1.29 × 10^{−7} | 800–550 °C | 1.378 |
550–250 °C | 1.028 | |||||
Wet air | 7.78 × 10^{−7} | 800–600 °C | 1.150 | |||
600–250 °C | 0.721 | |||||
62% Ho_{2}O_{3} + 38.0% TiO_{2} | Ln = Ho; x = 0.765 | 0.87 | Dry air | 0.91 × 10^{−7} | 800–550 °C | 1.408 |
550–250 °C | 1.100 | |||||
Wet air | 7.06 × 10^{−7} | 800–600 °C | 1.189 | |||
600–250 °C | 0.864 | |||||
52.0% Er_{2}O_{3} + 48.0% TiO_{2} | Ln = Er; x = 0.74 | 0.73 | Dry air | 0.87 × 10^{−7} | 800–550 °C | 1.434 |
550–250 °C | 1.105 | |||||
Wet air | 3.26 × 10^{−7} | 800–600 °C | 1.240 | |||
600–250 °C | 0.934 | |||||
52.0% Yb_{2}O_{3} + 48.0% TiO_{2} | Ln = Yb; x = 0.74 | 0.67 | Dry air | 1.11 × 10^{−7} | 800–500 °C | 1.445 |
500–250 °C | 1.246 | |||||
Wet air | 3.37 × 10^{−7} | 800–600 °C | 1.393 | |||
600–250 °C | 1.010 |
Fig. 6b shows the temperature dependence of the proton transfer coefficient in these ceramics. At 300 °C the proton transfer coefficients are higher for titanates with a smaller ionic radius. However, the difference is insignificant and amounts to 0.1. At 400 °C they are almost the same for all compositions and amount to 0.8. At 600 °C, a drastic decrease in transfer coefficients is observed for all three compositions.
The reason for the increase in proton conductivity of fluorite-like REE titanates with increasing lanthanide ionic radius may be an increase in the basic properties of the REE cations, in the compounds in which proton transfer occurs much more easily. The holmium cation belongs to the middle series of rare earth elements and its compounds, such as Ho_{2}Ti_{2}O_{7} pyrochlore, are capable of retaining hydroxide water up to 1200 °C.^{42} For some systems, the composite effect is also known, where the conductivity of pure pyrochlore and fluorite is lower than that of the mixed two-phase region P + F or P1 + P2.^{33,43,44}
The total conductivity for the ceramics (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.667, 0.684, 0.701, 0.718, 0.734, 0.750, 0.765) measured under a dry air atmosphere (Fig. 7c) increases with the holmium content in the series 0.667, 0.684, and 0.701, with the greatest difference at temperatures from 600 °C to 250 °C. A further increase in holmium to x = 0.750 leads to a decrease in conductivity values, while the order of conductivity values at 800 °C for all (Ho_{x}Ti_{1−x})_{4}O_{8−2x} ceramics (x = 0.667, 0.684, 0.701, 0.718, 0.734, 0.750, 0.765) ∼1 × 10^{−4} S cm^{−1}.
For conductivity in humidified air, the nature of the dependence of the values for (Ho_{x}Ti_{1−x})_{4}O_{8−2x} ceramics (x = 0.667, 0.684, 0.701, 0.718, 0.734, 0.750, 0.765) on the holmium content is not preserved, so that the increase in conductivity for samples x = 0.667, 0.684, 0.701 is replaced by a decrease in values for sample x = 0.718, then an increase in conductivity is observed with increasing holmium content up to x = 0.750. The ceramic sample (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.765) has high conductivity values in the series depending on the holmium content at temperatures from 450 °C to 800 °C, but in the temperature range from 450 °C to 250 °C the conductivity shows a decrease compared to the sample x = 0.750 (Fig. 7d).
The observed dependence of proton transfer coefficients and conductivity under dry and wet atmospheres on the holmium content for the ceramic (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.667, 0.684, 0.701, 0.718, 0.734, 0.750, 0.765) can be explained by changes in the crystal structure of the samples (Fig. 7e). The conductivity in dry air from 600 °C to 250 °C is the maximum for ceramics that contain nano-sized pyrochlore in addition to the main fluorite phase, i.e. (Ho_{x}Ti_{1−x})_{4}O_{8−2x}x = 0.701. Conductivity in dry air decreases as the amount of nanopyrochlore decreases. In wet air, however, the opposite effect is observed. The conductivity in wet air compared to dry air increases more significantly the higher the parameter of the main fluorite phase and the less nanopyrochlore in the ceramics. The maximum transfer coefficients at 250 °C are obtained for the compound (Ho_{x}Ti_{1−x})_{4}O_{8−2x}x = 0.750 or Ho_{3}TiO_{6.5}, which we consider to be the purest fluorite. Although the parameter of the fluorite phase of the sample with x = 0.765 is higher than that with x = 0.750, its conductivity is lower under both atmospheres, which is explained by the presence of an impurity Ho_{2}O_{3} phase in it.
The time dependencies of the impedance spectra, which allow us to confirm the proton contribution to the conductivity for the best sample (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.75), are shown in Fig. S3a and Fig. S3b† at temperatures of 300 and 450 °C, respectively. The proton conduction will depend on the rate at which water is absorbed in the sample. Considering the rather low conductivity, diffusion coefficients will be very small. Hence, a slow change might indicate proton conduction.^{45}
At room temperature, proton-conducting complex oxides also hydrate naturally due to the humidity of the surrounding air. It therefore seems reasonable to investigate weight loss in these materials. However, previous studies of rare earth titanates have shown that, in addition to the dehydration process, there is a process of CO_{2} release and carbon combustion discovered in the synthesis from coprecipitated precursors during their heat treatment in a wide temperature range of 800–1600 °C.^{46–48} X-ray diffraction and Raman spectroscopy methods have shown that Ln_{2}Ti_{2}O_{7} (Ln = Gd, Tb, Dy) precursors, e.g. obtained by “wet” chemical methods, are either a mixture of amorphous Ln–Ti hydroxycarbonates^{47} or a mixture of rare earth hydroxycarbonates with titanium hydroxide.^{46} At the same time,^{47} a violation of stoichiometry in high temperature ceramics and the presence of up to 3 wt% graphite were noted.
For powdered ceramics with the highest conductivity (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.75), we studied the change in sample mass under an atmosphere of oxygen with two times heating. The ability to hydrate of (1) the original powder and after exposure (2) for 24 hours and (3) for two weeks was investigated.
Fig. 8 shows the behavior of fluorite (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.75) in dry oxygen when heated to 1000 °C in a STA cell. For clarity, the main figure shows the results of a first heating of the original powder, the same hydrated for 24 hours and the same hydrated for two weeks. Note that the weight loss of the original sample was 0.20%, while the sample hydrated for 24 hours showed a slightly greater weight loss of ∼0.24%. After 2 weeks in water at room temperature, the sample showed a slightly lower weight loss. It is known that during hydration at temperatures below 300 °C, hydroxide and hydroxycarbonate impurities may dissolve.^{48} This may be associated with a slightly lower weight loss after 2 weeks of exposure to water. Therefore, keeping the powder for 24 hours is sufficient to ensure maximum hydration, and a soaking time of at least 2 weeks is required to dissolve most of the impurity hydroxides and hydroxycarbonates in water.
Loss of pore water and weakly bound water has been observed in samples up to 530 °C.
However, it should be noted that in the same temperature range CO_{2} is released as a product of the decomposition of the remaining REE hydroxycarbonates;^{46} any carbonaceous impurities are burnt out, which is accompanied by an exo-effect at ∼320 °C (Fig. S4†).
The process of decomposition of basic carbonates and hydroxycarbonates of rare earths in air is a multi-stage one and takes place over a wide range of temperatures. For example, the decomposition of lanthanum hydroxycarbonate LaOHCO_{3}·nH_{2}O proceeds as follows: first to La_{2}O_{2}CO_{3} at a temperature of about 500 °C, and then to La_{2}O_{3} in the range 750–830 °C, according to the scheme:^{49}
La_{2}(CO_{3})_{3}·nH_{2}O → La_{2}(CO_{3})_{3} + nH_{2}O↑ | (7) |
La_{2}(CO_{3})_{3} ⇌ La_{2}O_{2}CO_{3} + 2CO_{2}↑ | (8) |
La_{2}O_{2}CO_{3} ⇌ La_{2}O_{3} + CO_{2}↑ | (9) |
In air, in the presence of CO_{2}, where ceramic synthesis takes place, this process is reversible. It is therefore not surprising that traces of carbonates can be preserved in ceramics based on lanthanide-containing complex oxides, synthesized by any method using nano-sized precursors, even after annealing at high temperatures (1200–1600 °C).^{46,50}
At temperatures T > 530 °C, mass loss occurs mainly due to the release of OH^{−} groups and strongly bound protons. In the inset to Fig. 8, the results of the second heating of the samples are shown: (1) initial powder, (2) after 24 h exposure in water, and (3) after two weeks exposure in water. A gradual decrease in mass loss during reheating was observed: 0.22, 0.18, and 0.16 wt% respectively.
Note, however, that the TG curve of the sample under investigation in oxygen does not reach a plateau even when heated to 1000 °C. It is most likely that the final decomposition of carbonaceous impurities occurs at a higher temperature T > 1000 °C, since in ref. 46 it is shown that titanates-pyrochlores of rare earth elements, for example, tend to retain a small amount of carbon-containing compounds even after high temperature annealing in the range of 1200–1600 °C due to relatively slow crystallization.
The structural study by X-ray diffraction and Raman spectroscopy methods showed that the maximum proton conductivity is exhibited by the solid solution (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.701), containing pyrochlore microdomains of optimal size in the fluorite matrix, or pure fluorite with the composition (Ho_{x}Ti_{1−x})_{4}O_{8−2x} (x = 0.75).
The authors confirm that the data supporting the findings of this study are available within the article [and/or] its ESI.†
Footnote |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4dt01493f |
This journal is © The Royal Society of Chemistry 2024 |