Usually such new Li-ion conductive materials have been discovered through engineering manners, such as changing the composition of a known superionic compound, or adjusting the synthesis conditions. Recently, however, there have been other attempts to predict new SE materials by means of a computational approach using first-principles calculations. W. Richards et al. [4] predicted that Na₁₀SnP₂S₁₂, in which Li is substituted to Na in Li₁₀SnP₂S₁₂, would be a good Na ion conductor. They also synthesized Na₁₀SnP₂S₁₂ in practice and confirmed that the experimentally measured ionic conductivity and the activation energy of Na₁₀SnP₂S₁₂ agreed well with those of the predicted values.

In this study, we examined I^-4 type Li_1+2xZn_1-xPS₄ — a new Li ionic conductor computationally predicted by Richards et al. [5-6]. They predicted that the stoichiometric LiZnPS₄ is almost insulative, but non-stoichiometric compound of Li_1+2xZn_1-xPS₄ can show superionic conductivity with the predicted room temperature ionic conductivity (σ_25℃) as large as 50 mS/cm at x=0.5.

Li_1+2xZn_1-xPS₄ was synthesized by using solid state reaction. We ball-milled the starting materials and heated it at elevated temperature under vacuum. We measured the XRD patterns and the Raman spectra for Li_1+2xZn_1-xPS₄ at various x and confirmed that the samples with x ≤ 0.625 have I^-4 structure with little impurities.

Figure 1 shows the ionic conductivity at 25℃ (σ_25℃) and the activation energy (E_a) for Li_1+2xZn_1-xPS₄ at various x. Li_1+2xZn_1-xPS₄ shows negligible conductivity when x ≤ 0.2. At x = 0.375, σ_25℃ starts increasing and E_a starts decreasing. The highest σ_25℃ achieved was 5.7 x 10^-4 S/cm at x = 0.625. This tendency qualitatively agrees with the theoretical predictions [6], but there remains a considerable amount of discrepancy. When x ≥ 0.75, σ_25℃decreases, which likely corresponds to low phase purity and poor crystallinity of the material.

We also examined the feasibility of Li_2.25Zn_0.375PS₄(x=0.625) in a practical ASSB with NCM cathode and various anodes (graphite, indium, LTO), and found that this material was not stable with graphite and indium, but can be paired with NCM and LTO.

References

[1] N. Kamaya, K. Homma, Y. Yamakawa, M. Hirayama, R. Kanno, M. Yonemura, T. Kamiyama, Y. Kato, S. Hama, K. Kawamoto and A. Mitsui, Nature Materials, 10 (2011), 682.

[2] Y. Seino, T. Ota, K. Takada, A. Hayashi and M. Tatsumisago, Energy & Environmental Sciences, 7 (2014) 627.

[3] Y. Kato, S. Hori, T. Saito, K. Suzuki, M. Hirayama, A. Mitsui, M. Yonemura, H. Iba, R. Kanno, Nature Energy, 1 (2016) 16030.

[4] W. Richards, T. Tsujimura, L. Miara, Y. Wang, J. Kim, S. Ong, I. Uechi, N. Suzuki and G. Ceder, Nature Communications, 7 (2016) 11009.

[5] Y. Wang, W. D. Richards, S. P. Ong, L. J. Miara, J. C. Kim, Y. Mo and G. Ceder, Nature Materials, (2015), 14, 1026.

[6] W. Richards, Y. Wang, L. Miara, J. Kim and G. Ceder, Energy and Environmental Sciences, 9 (2016) 3272.

Figure 1. σ_25℃ and E_a for Li_1+2xZn_1-xPS₄.