Unlocking Li superionic conductivity in face-centred cubic oxides via face-sharing configurations

Oxides with a face-centred cubic (fcc) anion sublattice are generally not considered as solid-state electrolytes as the structural framework is thought to be unfavourable for lithium (Li) superionic conduction. Here we demonstrate Li superionic conductivity in fcc-type oxides in which face-sharing Li configurations have been created through cation over-stoichiometry in rocksalt-type lattices via excess Li. We find that the face-sharing Li configurations create a novel spinel with unconventional stoichiometry and raise the energy of Li, thereby promoting fast Li-ion conduction. The over-stoichiometric Li–In–Sn–O compound exhibits a total Li superionic conductivity of 3.38 × 10−4 S cm−1 at room temperature with a low migration barrier of 255 meV. Our work unlocks the potential of designing Li superionic conductors in a prototypical structural framework with vast chemical flexibility, providing fertile ground for discovering new solid-state electrolytes.


Supplementary Note 1: Details of EIS fitting
A simple equivalent circuit model shown in the Figure S2 was used for the fitting of EIS spectrum, which has been frequently used for pure ionic conductors 1 .The constant phase element (CPE) is typically used to model the behavior of a double layer that is an imperfect capacitor.The parallel resistor and CPE (Resistor 2//CPE 1) elements were used to fit the semicircle in the Nyquist plot 2 .
Since only a single semicircle was observed in the Nyquist plot of o-LISO, the grain (bulk) and grain boundary contributions cannot be deconvoluted.The capacitance of CPE 1, 6.12×10 -10 F, implies the process stems from both grain and grain boundary contributions 3 .Thus, the resistance of Resistor 2 represents a sum of the bulk and grain boundary resistances, which was used to calculate the total ionic conductivity.The CPE 2 modeled the linear spike at low-frequency region for cells measured in an ion-blocking configuration, which is attributed to the accumulation of Liions at the interface.Li conductors (experimental reports) [4][5][6][7][8][9][10][11][12][13][14][15][16] .The red-shaded region refers to superionic conductors, which typically have room-temperature ionic conductivities higher than 10 -4 S cm -1 at room temperature and activation energies for ion conduction lower than 400 meV.However, the fine structure of Li3InO3-like phase cannot be refined because it only forms as a small impurity phase in the LISO sample with excessively over-stoichiometric Li.Since it doesn't have the desired face-sharing Li configuration, we think this phase is less of interest and didn't do a further study.Arrhenius plots of ionic conductivity values obtained from variable temperature EIS measurements for ns-LISO.

Supplementary Note 3: Li-ion conductivity of ns-LISO
The ns-LISO reported in the main text was synthesized by calcinate at 1050 °C for 6 h and then sintered at 1050 °C for 10 h.If we apply a longer calcination or sintering time, e.g.12h, that DRX phase is maintained, the ionic conductivity can be even lower and lower than the detection limit by EIS.This is likely due to the further Li-loss with longer calcination or sintering time.This phenomenon also suggest that the Li content affect the ionic conductivity a lot.

Supplementary Note 4: Rationale of using two-phase model in Rietveld refinement
The XRD pattern of o-LISO exhibits broad peaks characteristic of the s-phase and sharp peaks that are common to the o-DRX and s-phase.As the diffraction peaks of the o-DRX phase fully overlap with those of s-phase, it is not trivial to determine if the o-DRX and s-phase co-exist with different domain size, or whether a single s-phase with small domain size leads to selective peak broadening.
To differentiate these two scenarios, we tried refining the synchrotron XRD data using a single spinel-like phase with selective peak broadening.This refinement method has been reported in a previous study 17 .The result is shown in Fig. S9.Although the selective peak broadening can fit the broad peaks characteristic of s-phase, a large intensity disparity between calculated and observed curves is observed for those sharp peaks overlapping with o-DRX phase.This suggests that a single s-phase is insufficient to fit the diffraction pattern, confirming the presence of a DRX phase.Therefore, in the Rietveld refinement of synchrotron XRD and TOF-NPD data, we chose a two-phase model.Note: The fraction of site 1 may be underestimated because of the low MAS spinning rate and very short T2.Note that since the Li site occupancies in s-phase were obtained by high throughput grid search, their uncertainties are not provided.

FigureFigure S4 .
Figure S3.(a) The d.c.polarization curves of o-LISO using ion-blocking electrodes at different

Figure S7 .
Figure S7.The enlarged XRD patterns of LISO17 samples with the calcination time of 4 h

Figure S9 .
Figure S9.Rietveld refinement of synchrotron XRD of o-LISO using a single-phase model with

Figure S10 .FigureFigure S12 .
Figure S10.Rietveld refinement of time-of-flight neutron powder diffraction patterns of (a) Bank

Figure S13 .
Figure S13.TEM electron diffraction pattern collected on ns-LISO particle along the zone axis

Figure S14 .
Figure S14.TEM electron diffraction pattern collected on o-LISO particle along the zone axis of

Table S2 .
The measured metal atomic ratios of o-LISO and ns-LISO from elemental analysis.Note that the measured Li contents in both samples are lower than the initial composition because of the Li-loss during high-temperature heat treatment.But still o-LISO is over-stoichiometric and the Li content in o-LISO is much higher than ns-LISO.

Table S3 .
Crystallographic data of o-DRX and s-phase in o-LISO as obtained from Rietveld refinement of TOF-NPD data.