Apatite-type lanthanum silicates (ATLS) and apatite-type lanthanum germanates (ATLG), as novel electrolytes for the intermediate temperature solid oxide fuel cells (IT-SOFCs),^[1–15] have been attracting widespread attention of the scientific community due to their clean energy conversions, superior ionic conductivities, excellent long-term stabilities and low processing cost. In contrast to the fact that traditional fluorite and perovskite-type oxide ion conductors have a vacancy conduction mechanism,^[16–18] apatite-type materials have an especial interstitial oxide ion conduction mechanism,^[19,20] which could be responsible for their high oxide-ion conductivities. It has been found that among the apatite-type conductors, ATLG has a higher conductivity than that of ATLS at similar temperatures. However, ATLS have been investigated more extensively than ATLG due to several problems related to Ge loss at high temperature.^[21–24]

It is well known that the conductivity of solid electrolytes highly depends on the levels of purity and densification. In general, dense apatite-type ceramics prepared by conventional SSRM require the repeated long-time and high-temperature (1300 °C–1700 °C) sintering. However, high sintering temperature (melting point of GeO₂) will induce the Ge volatilization loss, further bringing the secondary phase La₂GeO₅ into the samples, finally resulting in the reduction of conductivity.^[21–24] Therefore, in order to exhibit their best electrical properties using SSRM, dense ATLG ceramics with high purity are very difficult to obtain. In order to overcome this drawback, some synthesis methods were developed to optimize the precursor powders of ATLG, such as the ball–mill method,^[25] sol–gel auto-combustion process,^[26] hydrothermal synthesis method,^[27] MSSM,^[28] etc. Comparatively, the MSSM as a low-cost, low-temperature, few steps and versatile method can provide a melting-salt environment to form high-quality ATLG powders, further to lay a solid foundation for gaining the highly dense and pure pellets at a lower sintering temperature.^[28–30]

In this paper, we successfully synthesize apatite-type La_9.33Ge₆O₂₆ powders by MSSM through using LiCl and LiCl/NaCl mixture (mass ratio 1:1) as molten salt, and the corresponding synthesis temperature reduces about 350 °C (LiCl), 500 °C (LiCl/NaCl) compared with by using the conventional SSRM, respectively, which can effectively avoid Ge volatility, agglomeration and crystallinity problems encountered in the conventional preparation process of precursor powders. Furthermore, on the basis of high-quality precursor powders obtained by the MSSM, highly dense and pure ceramic pellets are obtained at a lower temperature of 1100 °C, which exhibit higher conductivities and lower activation energies than those obtained by the conventional SSRM. In addition, the optimal mass ratio between reactant and molten salt, the optimal synthesizing temperature and the associated mechanisms are investigated in detail.

Original reactant powders of La₂O₃ (purity ) and GeO₂ (purity ) were weighed in the stoichiometric composition of La_9.33Ge₆O₂₆. Before weighing, La₂O₃ powders were precalcined at 1000 °C for 2 h to remove any lanthanum hydroxide or carbonate. The LiCl and LiCl/NaCl mixture (mass ratio 1:1) were taken as eutectic salt and weighed. The mass ratios of reactants to eutectic salt were fixed to be 1:1, 1:2, and 1:3, respectively. Mixtures (reactants and eutectic salt) were further ball-milled, dried, ground, sifted, and calcined at 600 °C–750 °C for 8 h, sequentially. Then the as-synthesized powders were washed repeatedly with de-ionized water and filtered to remove eutectic salt until there was no white precipitate in the filtrate by adding the AgNO₃ solution. Finally, La_9.33Ge₆O₂₆ powders were obtained through drying the products at 110 °C for 12 h. Then the as-synthesized powders were pressed under 100MPa for 30 min and sintered at 1100 °C for 2 h to obtain the dense pellets. In addition, for comparison, the pellets synthesized by conventional SSRM were prepared (the precursor powders were synthesized at 1100 °C for 4 h, and the final pellets were sintered at 1300 °C for 2 h).

Phase identification of samples was carried out by x-ray diffraction (XRD; Shimadu XRD-6000). The data were collected over the 2θ range of 20°–60° in steps of 0.02° by using Cu–K radiation. Microstructures of samples were examined by scanning electron microscopy (SEM; JEOL JSM-6700F) and transmission electron microscopy (TEM; JEOL JEM-2200FS). Electrical characterizations of dense pellets were carried out by electrochemical impedance spectroscopy (EIS; Solartron 1260 frequency response analyzer). The Pt electrodes were coated on both surfaces of pellets and sintered at 800 °C for 1 h. The EIS measurements were performed at a 50-°C interval between 300 °C and 850 °C in a frequency range of 0.1 Hz–1 MHz, then the impedance spectrum data were analyzed by Z-View software.

To confirm the optimal mass ratio of reactant and molten salt, the mixtures of reactants and LiCl salt with the mass ratios of 1:1, 1:2, and 1:3 are calcined at 650 °C, 700 °C, and 750 °C according to the melting temperature of molten salt (see Fig. 1), respectively. The corresponding XRD patterns are shown in Fig. 2. It can be concluded that the apatite structure cannot be acquired when the mass ratios are 1:1 and 1:3 at different temperatures, which implies that these two mass ratios are unsuitable for synthesizing apatite-type La_9.33Ge₆O₂₆ powders. It is mainly because the original reactant powders (La₂O₃ and GeO₂) cannot completely dissolve in the melting-salt environment when the mass ratio is of 1:1, but too much melting-salt environment will inhibit the precipitation of La_9.33Ge₆O₂₆ powders as the mass ratio is 1:3. In contrast, the apatite structure is obviously observed as the middle mass ratio of 1:2, and a pure apatite phase without any impurity is obtained at a calcining temperature of 750 °C, which indicates that the middle mass ratio of 1:2 is optimal, because this proper melting-salt environment can simultaneously ensure the good dissolution of original reactant powders and the good precipitation of La_9.33Ge₆O₂₆ powders. Comparing with the SSRM, the calcining temperature is much lower than 350 °C, which can effectively inhibit Ge loss and reduce energy consumption when preparing the precursor powders. However, there is a problem as LiCl salt is difficult to remove. In order to overcome this drawback, the proper NaCl is incorporated into LiCl as eutectic salt. Experimental results indicate that the LiCl/NaCl mixed salts with the mass ratio of 1:1 can effectively solve the above problem. More excitingly, the pure apatite-type La_9.33Ge₆O₂₆ powders are successfully synthesized at a lower temperature of 600 °C (see Fig. 3) due to the lower melting temperature of mixed salts (see the inset of Fig. 1), which is more than 550 °C lower than that of the molten-salt synthesis method.

Fig. 1.

	Figure Option ViewDownloadNew Window
	Fig. 1. DTA curve of LiCl salt among 50 °C–800 °C in air, and the inset is DTA curve of LiCl/NaCl mixed salts (mass ratio 1:1) among 50 °C–800 °C in air.

Fig. 2.

	Figure Option ViewDownloadNew Window
	Fig. 2. (color online) XRD patterns of the powders synthesized by LiCl MSSM with different mass ratio of reactants and LiCl salt (1:1, 1:2, and 1:3) at 650 °C, 700 °C, and 750 °C for 8 h, respectively.

Fig. 3.

	Figure Option ViewDownloadNew Window
	Fig. 3. XRD pattern of the powders synthesized by LiCl/NaCl MSSM at 600 °C for 8 h.

The microstructure of La_9.33Ge₆O₂₆ powders synthesized by MSSM are displayed in Fig. 4. As shown in the figure, the as-synthesized powders are nano-sized, homogeneous and less agglomerated particles, which should be ascribed to the effect of the melting-salt environment. The melting-salt environment can promote the dispersion of reactants to make them realize an atomic/molecule-scale reaction, which, meanwhile, can exist around the growing particles to effectively inhibit the agglomeration and control the particle sizes. Compared with the SSRM, the MSSM can effectively solve the main problems such as Ge volatility, agglomeration and crystallinity encountered in the preparation of precursor powders, and the high-quality precursor powders are obtained, which will be more beneficial for gaining the dense and pure pellets. In addition, it can be seen that the La_9.33Ge₆O₂₆powders synthesized in LiCl/NaCl mixed melting-salt environment are more homogeneous than that synthesized in the LiCl individual melting-salt environment, and the corresponding particles have an average size of around 200 nm.

Fig. 4.

	Figure Option ViewDownloadNew Window
	Fig. 4. SEM (a) and TEM (b) images of La9.33Ge6O26 powders obtained by LiCl MSSM at 750 °C for 8 h; and SEM (c) and TEM (d) images of La9.33Ge6O26 powders obtained by LiCl/NaCl MSSM at 600 °C for 8 h.

On the basis of high-quality precursor powders, the proper pellets are sintered at 1100 °C (which is lower than the melting point of GeO₂) for 2 h. The low sintering temperature and short sintering time can effectively avoid the high-temperature Ge loss problem encountered in conventional SSRM, which can be confirmed by the XRD pattern of pellets in Fig. 5. As shown in the figure, the proper pellet has a pure hexagonal apatite-type structure with no impurity (La₂GeO₅). Moreover, the relative density of pellets reaches a high level of ∼95% measured by the Archimedean method, which can be illustrated by the SEM images of pellets (Fig. 6). In addition, the pellets obtained by the more homogeneous precursor powders that are synthesized in LiCl/NaCl melting-salt environment have fewer grain boundaries, which should be more useful for oxide-ion conduction.

Fig. 5.

	Figure Option ViewDownloadNew Window
	Fig. 5. (color online) XRD patterns of La9.33Ge6O26 pellets obtained by MSSM at 1100 °C for 2 h and the compared pellets sintered by SSRM at 1300 °C for 2 h.

Fig. 6.

	Figure Option ViewDownloadNew Window
	Fig. 6. SEM images of the pellets prepared by the powders obtained by LiCl (a) and LiCl/NaCl, (b) MSSM at 1100 °C for 2 h.

The electrochemical impedance of ceramic pellets are measured in air from 300 °C to 850 °C in steps of 50 °C, and the complex impedance plots measured at 300 °C, 550 °C, and 700 °C are displayed in Fig. 7. In order to distinguish the contributions of bulk, grain boundary and electrode, the corresponding contributions are analyzed with the Z-View software by modeling the impedance spectrum with the equivalent circuit including four serial R_i and (CPE)_i parallels (i refers to “bulk” for bulk, “gb” denotes the grain boundary and “elec” the electrode). As shown, the impedance spectrum measured at 300 °C exhibits two independent semicircular arcs and a straight line from high frequency to low frequency, which corresponds to the conduction across the bulk (R_bulk), the conduction across grain boundary (R_gb) and the impedance response corresponding to electrode (R_elec), respectively. With the increasing of temperature, the arc corresponds to the grain contributions gradually disappearing, and the arc or spike corresponds to the grain boundary and electrode shifted to higher frequencies as shown in the impedance spectrum measured at 500 °C. When the temperature is beyond 700 °C, the arcs corresponding to the grain and grain boundary contributions disappear, and only an arc ascribed to the contribution of the electrode is observed, which indicates that the oxide-ion could diffuse through the entire thickness of bulk and grain boundary. Furthermore, the dense pellets sintered by more homogeneous precursor powders that are synthesized in LiCl/NaCl mixed melting-salt environment exhibit a better electrical property, which should be attributed to the higher density and fewer grain boundaries.

Fig. 7.

	Figure Option ViewDownloadNew Window
	Fig. 7. (color online) Electrochemical impedance spectra of La9.33Ge6O26 pellets sintered by the powders obtained by LiCl (a) and LiCl/NaCl (b) MSSM measured at 300 °C, 550 °C, and 700 °C, and the corresponding schematic equivalent circuits.

The corresponding conductivity is converted by the resistance R using Eq. (1), and the activation energy is calculated by the Arrhenius equation (2):

The conductivities of pellets prepared by MSSM and SSRM at different temperatures are plotted in Arrhenius plots (Fig. 8). As shown, the pellets obtained by MSSM exhibit evidently higher conductivities than by SSRM. The La_9.33Ge₆O₂₆ pellets prepared by the MSSM exhibits excellent conductivities with values of (LiCl) and (LiCl/NaCl) at 850 °C, respectively, which are higher than those of the pellets with the same compositions reported in Refs. [24]–[26], [31], and [32]. Furthermore, the corresponding activation energy is calculated by the Arrhenius equation. The activation energies ( , ) of pellets prepared by MSSM are lower than those ( ) of pellets prepared by SSRM, which should be ascribed to the high density, no impurity and fewer grain boundaries. In addition, consistent with the fewer grain boundaries, the dense pellets obtained by LiCl/NaCl MSSM exhibit a higher conductivity and a lower activation energy than those obtained by LiCl MSSM, which is consistent with the case of fewer grain boundaries.

Fig. 8.

	Figure Option ViewDownloadNew Window
	Fig. 8. (color online) Arrhenius plots for the conductivity of La9.33Ge6O26 pellets sintered by MSSM at 1100 °C and the compared pellets sintered by SSRM at 1300 °C.

High-quality La_9.33Ge₆O₂₆ powders with apatite structure are successfully synthesized via a facile MSSM by using LiCl, LiCl/NaCl mixture (mass ratio 1:1) as molten salt, respectively. The optimal mass ratio of reactant and molten salt is 1:2, and LiCl/NaCl mixed molten-salt is more beneficial for synthesizing high-quality La_9.33Ge₆O₂₆ powders. Compared with the SSRM, the MSSM can evidently reduce the synthesis temperature to aviod Ge loss and obtain homogeneous, non-agglomerated and well crystallized powders. Furthermore, the dense and pure ceramic pellets of La_9.33Ge₆O₂₆ are obtained at 1100 °C for 2 h, which exhibit higher conductivity and lower activation energy than those obtained by the SSRM, due to their high density, no impurity and fewer grain boundaries. Additionally, the dense pellets obtained by LiCl/NaCl MSSM exhibit the best conductivity and lowest activation energy.