Layered Ionic Conductor Na2Zn2TeO6: Crystal Structure and Stacking Faults

Abstract

Na-ion batteries (NIBs) are considered the most attractive and cost-effective alternative to Li-ion batteries (LIBs), due to the widespread abundance of sodium and its chemical similarity with Li, facilitating the conversion of battery technologies. Meanwhile, replacing the liquid electrolyte in traditional batteries with a super-ionic conductor material in the solid-state presents additional advantages in terms of safety and stability. The research for solid-state electrolytes (SSE) is consequently a very active field. Layered tellurates with formula Na2 M 2TeO6 (M = Co, Mg, Zn, Ni) have been reported to exhibit high Na-ion conductivity, around 10-4 S/cm, and crystallize into two distinct P2-layered modifications, structurally related to Na-deficient Na xMO2 (x < 0.7; M = Co, Fe, Mn) materials. Na2Ni2TeO6 (NNTO) adapts an AA stacking of hexagonal Ni-Te layers (P63/mcm), while Na2 M’ 2TeO6 (M’ = Zn, Co) adapts an alternative AB stacking (P6322) [1]. However, the structure of the Na2 M’ 2TeO6 compounds is not yet fully resolved, because the refinements reported in literature do not account for the broadening of a low-angle peak in the X-ray diffraction pattern which could be ascribed to the presence of stacking faults [1,2]. In this study, Na2Zn2TeO6 (NZTO) was synthesized via conventional solid-state methods. Synchrotron X-ray diffraction (SXRD) experiments were performed to verify the presence and quantify the amount of stacking faults in the structure. The standard Rietveld refinement against SXRD data shows that NZTO crystallizes into the P6322 structure (a = 5.289 Å, c = 11.244 Å), consistent with previous reports. The refinement exhibits a poor fit of Bragg intensity and could not describe the significant broadening of the (0 1 1) reflection, as shown in Figure 1. We performed density functional theory (DFT) calculations to verify the nature of these faults. We modeled two different NZTO modifications, the standard one (space group P6322) and the one obtained by substituting Zn for Ni in the NNTO structure. The energy difference is smaller than 10 meV, indicating the possible coexistence of these phases at room temperature or, indeed, the presence of the associated stacking faults in the P6322 structure. The latter hypothesis was directly tested by simulating the effect of NNTO-type stacking faults in NZTO and subsequent refinement against SXRD data employing the FAULTS software [3]. The simulated XRD patterns reveal that only the (0 1 1) reflection is drastically broadened and its intensity decreases as the fraction of NNTO-type stacking faults increases. Notably, the other peaks remain largely unchanged. The FAULTS-refined profiles exhibit a better fitting of the Bragg intensity and peak broadening of the (0 1 1) reflection, revealing 3% NNTO-type stacking faults in NZTO. In conclusion, a more accurate structural model of the NZTO compound is proposed and validated. Figure 1: Rietveld refinement (without stacking faults) and FAULTS refinement (with stacking faults) against SXRD data (λ = 0.3149 Å) for NZTO.