Atsushi Mineshige

The two types of c-axis-aligned polycrystals of lanthanum silicate oxyapatite (LSO) doped with K2O and Al2O3 were prepared by the templated grain growth method with different template/matrix mass ratios of 11.1/88.9 and 5.9/94.1. The template particles, K2O- and F-doped plate-like LSO crystals with developed {001} faces, and the matrix powder, mainly Al2O3-doped LSO, were those used in a previous study with the template/matrix mass ratio of 20.0/80.0. Considering the ionic conductivity and orientation degree of the three types of textured polycrystals, the intermediate mixing ratio of 11.1/88.9 was found to be optimal among the three, with the texture fraction of {0 0 l}apatite being 0.74. Thus, the random grain oriented polycrystal was prepared with the same bulk chemical composition as the control sample. As the temperature increased from 773 to 823 K, the bulk oxide-ion conductivity (σb) of the textured polycrystal increased from 1.04 × 1013 to 1.71 × 1013 S cm11 and the activation energy of conduction (Ea) was 0.61 eV. The σb value of the random grain oriented polycrystal increased steadily from 3.80 × 1015 to 5.26 × 1014 S cm11 with increasing temperature from 773 to 973 K (Ea = 0.92 eV). Comparing the σb values at the same temperatures, the former was 27.5 (773 K) and 21.2 (823 K) times higher than the latter. The chemical formula of the doped LSO in the textured polycrystal was determined from the average composition to be (La9.59K0.09□0.32)(Si5.50Al0.38□0.12)O26, where □ denotes vacancies in La and/or Si sites. The major chemical composition of the coexisting interstitial material was estimated to be 29.8 mol % La2O3, 44.9 mol % SiO2, and 25.3 mol % Al2O3. The mole fractions were determined by the lever rule to be 0.9824 for doped LSO and 0.0176 for interstitial material.

The disk-shaped sintered compacts consisting mainly of c-axis-aligned lanthanum silicate oxyapatite (LSO) polycrystals doped with K2O and Al2O3 were prepared by templated grain growth method. The tabular single crystals of K2O- and F-doped LSO were used as template particles, and the Al2O3-doped LSO as matrix powder. The chemical formula derived from the average chemical composition of the constituent LSO crystal grains was (La9.60K0.16□0.24)Σ=10(Si5.48Al0.38□0.14)Σ=6O26, where □ denotes vacancies in La and/or Si sites. One of the doped LSO grains was examined by single-crystal X-ray diffraction to confirm that the crystal structure was isomorphous to those previously reported. Other coexisting phases were 0.16 mol% LaAlO3 and 0.34 mol% SiO2-rich interstitial material (probably in a liquid state at high temperature). With increasing temperature from 773 to 1073 K, the total oxide-ion conductivity (σtotal) along the grain-alignment direction steadily increased from 1.31 × 10−4 to 3.21 × 10−3 S cm−1, each value of which was, at the same temperature, more than 7.7 times larger than that of the randomly grain-oriented polycrystal with the same bulk chemical composition. The larger σtotal-value of the former polycrystal would be due to the significantly higher oxide-ion conductivity along the c-axis direction of the doped LSO. A solid oxide fuel cell (SOFC) using the textured polycrystal as the electrolyte was constructed and tested for power generation. The maximum power density was satisfactorily high at 873 K, indicating that the c-axis-aligned polycrystals of doped LSO would be potentially applicable as electrolytes for the SOFCs operating at medium to low temperatures.

Ba0.6La0.4F2.4 (BLF) is one of the promising candidates for solid-state electrolytes in fluoride-ion batteries owing to the wide electrochemical potential window. Compared with a BLF single crystal, the microstructure-modified BLF polycrystals by mechanochemical synthesis have improved the F− ion conductivity. However, the fundamental mechanism of the F− ion conductivity enhancement is still unclear because of the lack of structural and chemical analysis at the atomic level. Here, we investigate the microstructure of the mechanochemically synthesized BLF, using atomic-resolution scanning transmission electron microscopy combined with electron energy-loss spectroscopy. We find that the La-rich tysonite nano-precipitates are formed in the BLF fluorite matrix. The observed unique microstructure in BLF realized by mechanochemical synthesis would be the origin of the F− ion conductivity enhancement.

Among the three inequivalent fluorides in tysonite CeF3(F1, F2, and F3 in the ratio of 12:4:2 per unit cell), we show by 19F solid-state NMR that F1 is solely responsible for the ion conductivity in Ce1-xSrxF3-x(x = 0.001) at 0 °C. It is further shown that the observed conductivity can be explained quantitatively by using the Nernst-Einstein relation with the F1-F1 exchange rate (ca. 6 × 105s-1) estimated from lineshape analysis with the carrier-F1 concentration. As for an alternative method to obtain the hopping rate of a carrier, we adopt the AC impedance method, for which the identity of the "carrier" is rather elusive. The observed AC impedance gives the carrier hopping rate of 3.5 × 107s-1, which is ca. 60 times of the F1-F1 exchange rate determined by NMR. The slower F1-F1 hopping rate is ascribed to the result of the long-time average of the faster carrier hopping rate. For the AC impedance analysis, the concentration of the carrier to realize the observed conductivity is much larger than that of the fluoride ion vacancy introduced by Sr doping. For explanation, we postulate that what influences AC impedance includes not only the vacancy but also fast-exchanging fluoride ions around the vacancy.

There has been a long-standing controversy over ion conducting pathways in tysonite. To unravel the pathways in CeF3, high-resolution MAS NMR spectra of 19F in CeF3 at −40–240 °C are analyzed on the bases of the calculated dipolar coupling tensor between 19F and paramagnetic electron spins at Ce3+ to assign the three spinning-sideband manifolds for the three crystallographically inequivalent sites (F1, F2, and F3). Clearly resolved sideband manifolds for F1∼F3 at 240 °C indicate that ion exchange among the three sites is not fast enough to affect the 19F spectrum, whereas F1 motion is suggested by its broader linewidth. Further, we show that addition of only 0.1% Sr2+ in CeF3 (Ce1-xSrxF3-x (x = 0.001)) brings significant effects on F-ion mobility. While some of the F2 ions remain at the F2 site, most of the F3 ions undergo exchange motion between F1 at 240 °C. The preference of the exchange pathways in CeF3 is thus consistent with that postulated for LaF3, that is, F1–F1 exchange occurs first followed by F1–F3, and F1–F2 exchange is most restricted.