Publication date: 10th April 2024
1. Introduction
Solid electrolytes with high ionic conductivities are essential for the development of all-solid-state batteries. Our prior research has identified perovskite–type AELiH3 (AE = Ca, Sr, Ba) as an emerging hydride–ion conductor.[1] Using the nudged elastic bandmethod revealed that the smaller A-site cation offers the lower migration energy. Experimentally, the synthesis of AE1–xNaxLiH3–x was attempted, and a similar trend in conductivity was observed. However, it was found that CaLiH3 could not be synthesized, possibly due to an unfavorable tolerance factor (t = 0.83) for the cubic perovskite structure.
In this study, we explored the compositional space in the SrLiH3–CaLiH3–NaLiH2 quasi-ternary diagram to investigate the formation range of the perovskite-type structure. Inspired by the NEB calculation results noted above, the A-site cations of the SrLiH3–NaLiH2 quasi-binary system were substituted by smaller Ca2+ ions, aiming to reduce the migration energy. Furthermore, the association energy between dopants and hydrogen vacancies, which is a constituent of the activation energy in addition to the migration energy, was calculated through DFT calculations. The relationship between the composition, the activation energy, and the ionic conductivity was investigated to establish material designing principle for perovskite-type hydride-ion conductors.
2. Methodology
The synthesis was carried out via the mechanochemical milling at 600 rpm for 12 h. The obtained samples were subjected to X-ray diffraction (XRD) measurement to identify the constituent phases. Synchrotron XRD (BL02, SPring-8, Japan), and neutron powder diffraction (NPD) data (BL09 SPICA, J-PARC, Japan) were obtained. The crystal structure was refined by Rietveld analysis using the Z-Rietveld software. The ionic conductivities of the uniaxially compressed powder samples were measured by alternating current impedance. Density functional theory calculations were performed using the projector-augmented wave method and PBEsol functional as implemented in the VASP code.
3. Results and discussion
As the first step, we examined the synthesis of Sr1–xCaxLiH3 solid solutions. In the range of 0≦x≦0.45, the single-phase of cubic perovskite-type structure were obtained. The activation energy decreased upon Ca-substitution, and reached minimum of 47.2 kJ mol–1 for x = 0.45. This result suggests the activation energy can be mitigated by employing a smaller A-site cation, in accordance with NEB calculation results.
Next, perovskite-type hydride-ion conductors within the SrLiH3–CaLiH3–NaLiH2 quasi-ternary system were explored. The composition of Sr0.75Ca0.1Na0.15LiH2.85 exhibited a maximum ionic conductivity of 1.2 × 10–4 S cm–1 at 100ºC, and Sr0.825Ca0.025Na0.15LiH2.85 exhibited a minimal activation energy of 41.8 kJ mol–1, superior over the previously reported hydride-ion conductor Sr0.925Na0.075LiH2.925 (6.4 × 10–5 S cm–1, 46.3 kJ mol–1) [1]. Structural refinement using NPD data of Sr0.7Ca0.1Na0.2LiH2.8 confirmed a larger atomic displacement parameter (0.0277) of hydrogen than that of Sr0.8Na0.2LiH2.8 (0.0231(3)), revealing that the A-site substitution by Ca offer more flatten energy landscape of hydrogen in terms of the average structure. In contrast, more substitution by Ca did not provide enhancement of the ionic conductivity. The association energy between the dopant (Na) and a hydrogen-vacancy in AE63Na1Li64H191 supercell (AE = Ca, Sr, Ba) was calculated to be 19, 14, and 22 kJ mol–1, respectively. The association energy in CaLiH3 was higher than that of SrLiH3, in other words, the local trapping of hydrogen-vacancy by Na is more prominent for CaLiH3. These results suggest the trade-off between the migration energy and the association energy in Sr1–xCaxNayLiH3–y. The optimal balance between the migration energy and association energy was achieved with small amounts of Ca (y = 0.025~0.1), leading to reduced activation energy and enhanced ionic conductivity compared to the previously reported perovskite-type hydride-ion conductors.
This study was supported by AGC Inc. and Japan Society for the Promotion of Science KAKENHI grants (Nos. 17H06145, 19H05793, 22K14472, and 23KJ0933). Synchrotron XRD experiments were carried out as projects approved by the Japan Synchrotron Radiation Research Institute (JASRI) (Proposal No. 2023A1647). The neutron diffraction experiment was approved by the Neutron Scattering Program Advisory Committee (Proposal Number 2019S06) of IMSS, KEK. The computing resources of the TSUBAME3.0 supercomputer at the Tokyo Institute of Technology were used.