All-solid-state Li-ion batteries (ASSLIB), which replace the liquid electrolyte with solid materials, are expected to bring the improved chemical and structural stability and higher volumetric energy density than conventional Li-ion batteries. An emergent task for the next generation ASSLIB is to develop novel solid electrolyte material that has competitive Li-ion conductivity. The glassy lithium-phosphorus oxynitride (LiPON) has attracted attentions as a good candidate for that, however its practical rate capability in a full cell has been measured to be unexpectedly low, far below the computational predictions, which was attributed to the interface between solid electrolyte and cathode.

Recently, one research team has reported that the rate capability of LiPON-ASSLIB could be highly improved by adding BaTiO3 (BTO) nanoparticles to the LiPON/electrode interface (Adv. Energy Mater. 4, 1301416). They hypothesized that that the space-charge layer formed in the interface area was the origin of low Li-ion conductivity and the strong dielectric behavior of BTO additive rectified it. However, the verification of the claimed hypothesis has not been made due to experimental difficulty in the measurements of the space-charge layer. Rigorous computational study at atomic scales will enable us to verify this hypothesis and understand the mechanism of the improved rate capability. In this work, we perform first-principles calculations to 1) model the LiPON/Ni-Mn spinel interface structure with and without BTO additive and 2) investigate the effects of BTO additive on the Li-ion conductivity.

First, a model for the bulk LiPON is developed. There are several chemical compositions of LiPON (LixPOyNz; x=2y+3z-5). We select one sample structure, Li4PO3N, among them, for efficient study. The bulk Li4PO3N is obtained by substituting N ions for oxygen and adding extra Li ions to Li3PO4 crystal. Several ionic configurations that have different N substitution and Li addition are examined and the most stable one is identified. Next, the obtained bulk Li4PO3N is cropped and assembled with Ni-Mn spinel slab to create a model of the interface between LiPON electrolyte and Ni-Mn spinel cathode. We examine several combinations of different contacting facets and terminations, assuming that the interface structure will tend to have similar oxygen arrangement over Li4PO3N and Ni-Mn spinel parts, and find the most thermodynamically favorable one. Finally, we add BTO in the interface model to investigate its effect on the Li-ion conductivity. The activation energy barrier of Li-ion diffusion will be calculated in each case: bulk Li4PO3N, Li4PO3N/Ni-Mn spinel, and Li4PO3N-BTO/Ni-Mn spinel, and compared to investigate the effect of BTO additive on the Li-ion conductivity.

From analysis on the ions distribution and electric charge distribution, we propose a structural distortion of the Li4PO3N part, which is ascribed to the lattice mismatch between the Li4PO3N and Ni-Mn spinel, can be more significant factor to determine the ionic conductivity than the space-charge layer. To support our suggestion, we calculate the activation energy barrier of Li-ion diffusion through the bulk Li4PO3N by applying different mechanical strains.