Both Li-ion and Na-ion all solid state batteries are under development, where the absence of an liquid electrolyte potentially leads to improved safety and higher energy densities. Both the bulk ionic conductivity, as well as the conductivity over the electrolyte-electrolyte and electrolyte-electrode interfaces are prerequisites for efficient charge transport in solid state batteries. In many solid electrolytes the bulk mobility can be improved by introducing dopants that control the site occupancies by charge compensation. On the other hand the impact of solid-solid interfaces is anticipated to be one of the major bottlenecks towards high performance solid state batteries, making it of large importance to characterize the transport in detail.

The present study provides detailed insight in the Na-ion and Li-ion conductivity in Na₃PS₄¹ and Li₆PS₅Cl² sulfide solid electrolytes. The relative contribution of the bulk conductivity versus the conductivity over the solid-solid electrolyte and solid-solid electrolyte-electrode interfaces is revealed by comparing solid state NMR relaxation experiments with impedance spectroscopy. In addition X-ray and Neutron diffraction is performed to characterize the structures in detail. The structure and self diffusion measured by NMR is compared to DFT based Molecular Dynamics (MD) simulations that reveal the diffusion mechanism and the impact of halogen substitution in both cubic/tetragonal Na₃PS₄ (Na_3-yPS_4-yX_y, X=Cl,Br,…) and argyrodite Li₇PS₆ (Li₆PS₅X, X=Cl,Br,…) structures.

Solid state NMR relaxation experiments, shown in Figure 1c for Li₆PS₅Br, have the ability to directly probe the atomic scale Li-ion mobility in solid electrolytes and electrode materials^3-6offering a non-destructive method to probe the self-diffusion coefficients. Conversion to solid state conductivities allows complementary information to impedance spectroscopy, for instance giving insight in the role of grain boundaries on the overall conductivity, as illustrated in Figure 1d.

DFT MD-simulations, shown for Li₆PS₅Br in Figure 1a and 1b, allow to develop understanding of the diffusion mechanism of Li-ions and Na-ions in solid electrolytes and how these structures may be optimized to yield higher bulk conductivities. By systematically varying the halogen element to introduce Na-ion vacancies in both cubic/tetragonal Na₃PS₄ (Na_3-yPS_4-yX_y, X=Cl,Br,…) and in argyrodite Li₇PS₆ (Li₆PS₅X, X=Cl,Br,…) structures, the simulations reveal the impact of vacancies and the host structure on the conductivity. This gives insight why and what halogen dopant results in optimal structural conditions to improve the Na-ion and Li-ion conductivity.

By combining solid state NMR, Impedance, XRD and ND with DFT MD simulations, this study provides detailed insight in the Na-ion and Li-ion diffusion mechanism in in both cubic/tetragonal Na₃PS₄ (Na_3-yPS_4-yX_y, X=Cl,Br,…) and argyrodite Li₇PS₆ (Li₆PS₅X, X=Cl,Br,…) solid electrolytes and potential routes to improve its performance in all solid state batteries.

(1) Hayashi, A.; Noi, K.; Sakuda, A.; Tatsumisago, M. Nature Communications 2012, 3.

(2)  Deiseroth, H.-J.; Kong, S.-T.; Eckert, H.; Vannahme, J.; Reiner, C.; Zaiss, T.; Schlosser, M. Angewandte Chemie-International Edition 2008, 47, 755.

(3) Wagemaker, M.; van Eck, E. R. H.; Kentgens, A. P. M.; Mulder, F. M. Journal of Physical Chemistry B 2009, 113, 224.

(4) Wagemaker, M.; Kentgens, A. P. M.; Mulder, F. M. Nature 2002, 418, 397.

(5) Schmidt, W.; Bottke, P.; Sternad, M.; Gollob, P.; Hennige, V.; Wilkening, M. Chemistry of Materials 2015, 27, 1740.

(6) Epp, V.; Gun, O.; Deiseroth, H.-J.; Wilkening, M. Journal of Physical Chemistry Letters 2014, 4, 2118.

(7) Tanibata, N.; Noi, K.; Hayashi, A.; Kitamura, N.; Idemoto, Y.; Tatsumisago, M. Chemelectrochem 2014, 1, 1130.

Figure 1. (a) The Li-density and (b) Li-ion jumps for Li₆PS₅Br at 600 K from DFT molecular dynamics simulations. (c) ⁷Li NMR spin-lattice relaxation rates 1/T₁ (at 155.506MHz) of Li₆PS₅Br which allows to deduce the hopping rate and activation energy for Li-ion self-diffusion. (d) Temperature-dependence of Li jump rates of Li₆PS₅Br obtained from MD calculations and deduced from ⁷Li static SLR NMR measurements. The dash line in the figure shows an Arrhenius activation energy of 0.18 eV for bulk self-diffusion and s indicates the macroscopic conductivity measured with impedance spectroscopy. The lower value of s implies that diffusion over the grain boundary between different Li₆PS₅Br grains is limiting the conductivity of Li₆PS₅Br electrodes.