Chemical Bond Covalency in Superionic Halide Solid‐State Electrolytes

Abstract

Halide solid‐state electrolytes (SSEs) are promising superionic conductors with high oxidative stability and ionic conductivity, making them attractive for all‐solid‐state lithium‐ion batteries. However, most studies have focused on ion‐stacking structures, overlooking the role of bond characteristics in ionic transport. Here, we investigate bond dynamics and the superionic transition (SIT) in bromide electrolyte, Li₃InBr₆, using synchrotron X‐ray techniques and ab initio molecular dynamics (AIMD) simulations. We demonstrate that the SIT in halide SSEs is driven by a thermally induced transition in bonding character (ionic to covalent) rather than a change in crystal phase. AIMD simulations further reveal enhanced Li⁺ diffusion and collective anion motion at elevated temperatures. Expanding our study to Li₃LnBr₆ (Ln = Gd, Tb, Ho, Tm, and Lu), we confirm the widespread occurrence of SIT in this material class, with Li₃GdBr₆ exhibiting the highest ionic conductivity (5.2 mS cm⁻¹ at 298 K). More importantly, the ionic‐covalent transition is highly tunable through electrolyte modifications, such as cation/anion substitution and synthesis methods. Our findings provide a new perspective on ionic transport, highlighting the critical role of chemical bond characteristics in halide SSEs.

Achieving high conductivity in halide solid‐state electrolytes requires understanding key factors for ionic diffusion. This study provides direct evidence that metal‐halide covalent bonding controls superionic transitions and enhances Li⁺ mobility. Tailoring covalency through composition, demonstrated via in situ studies and simulations on bromide‐based electrolytes, offers a novel bond‐centric design strategy for high‐performance SSEs.

Introduction

Achieving high ionic conductivity in solid‐state electrolytes (SSEs) remains a critical challenge for the development of all‐solid‐state batteries (ASSBs).^{[ 1 ]} Halide SSEs, with the general formula Li–Me–X (where “Me” represents one or more metal elements and “X” refers to one or more halogen elements), constitute a promising class of superionic conductors. Following Panasonic's pioneering work on Li₃YX₆ (X = Cl, Br),^{[ 2 ]} diverse halide systems have demonstrated high electrochemical oxidative stability (>4 V vs. Li⁺/Li) and ionic conductivity (>1 mS cm⁻¹).^{[ 3} , ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ^{9 ]} Previous studies on halide SSEs extensively explored ion‐stacking structures, including LiX‐derived anionic close‐packed structures^{[ 10} , ^{11 ]} (hexagonal close‐packed or cubic close‐packed arrangements), UCl₃‐type non‐close‐packed structures,^{[ 6} , ¹² , ^{13 ]} and glassy halides.^{[ 14} , ¹⁵ , ^{16 ]} Li⁺‐conducting behavior was further tailored through compositional tuning,^{[ 17} , ^{18 ]} elemental substitutions,^{[ 19} , ²⁰ , ²¹ , ²² , ^{23 ]} and anion mixing.^{[ 24} , ^{25 ]} However, the focus on crystal structure and composition has resulted in less attention being paid to other potentially crucial factors influencing ion transport, particularly the influence of covalent bonding interactions within these halide frameworks.

The concept that covalent bonding influences ionic conductivity was initially proposed by Phillips et al.^{[ 26 ]} in the 1970s and further explored by Aniya et al.^{[ 27} , ²⁸ , ^{29 ]} in their studies of Ag⁺ and Cu⁺ superionic conductors. These early works suggested that thermally driven fluctuations in bond character could facilitate the migration of the mobile ions. While this perspective was rarely applied to early Li⁺ conductor research, recent findings in Li⁺ halide SSEs are renewing interest in the role of bonding, especially in bromide‐based SSEs. For instance, a significant conductivity enhancement by an order of magnitude was observed in Li₃InBr₆ near room temperature (RT), referring to a superionic transition (SIT).^{[ 30} , ³¹ , ³² , ³³ , ³⁴ , ³⁵ , ³⁶ , ³⁷ , ³⁸ , ³⁹ , ⁴⁰ , ^{41 ]} Supporting this, first‐principles molecular dynamics simulations by Adelstein et al.^{[ 42 ]} correlated covalent interactions with the SIT in Li₃InBr₆, proposing that the transition reflects an order–disorder transition in bonding characteristics. Experimentally, Liu et al.^{[ 25 ]} demonstrated the impact of SITs in Li₃YCl₆ by observing collective anion motion above the superionic transition temperature (T _c). By lowering T _c via anion mixing, they achieved a record‐high ionic conductivity of 11 mS cm⁻¹ among halide at RT. While computational modeling and experimental observations offer preliminary insights into halide SSEs, a fundamental understanding of how covalent bonding affects ion transport mechanisms is still missing. Crucially, direct evidence explicitly connecting specific covalent features to ionic conductivity remains scarce. Therefore, advancing the development of high‐performance halide SSEs requires targeted research to elucidate the precise influence of covalent bonding on ion mobility and to explore strategies for effectively tailoring these bonding characteristics.

In this study, we selected Li₃InBr₆, which exhibits a significant SIT at near RT, as a model system to investigate the interplay between structural evolution, bonding dynamics, and ion transport. Using in situ structural characterizations, we observed a contraction of [InBr₆]³⁻ octahedra upon heating, indicating a shortened In─Br bond distance consistent with increased covalency. Complementary ab initio molecular dynamics (AIMD) simulations provided dynamic insights consistent with these structural changes, revealing enhanced Li⁺ diffusion and collective anion motion at elevated temperatures (600 K); this suggests that increased covalency facilitates cooperative ion transport. We further validated the crucial role of metal‐halide (Me─X) bond covalency through Y/Ga doping experiments, establishing a clear link between framework electronegativity, Me─X covalency, and the transition temperature T _c. Expanding our investigation to a series of bromide‐based electrolytes, Li₃LnBr₆ (Ln = Gd, Tb, Ho, Tm, and Lu), we confirmed the prevalence of SITs in this class of materials, with Li₃GdBr₆ exhibiting the highest room‐temperature ionic conductivity (5.2 mS cm⁻¹ at 298 K). Additionally, we explored the influence of typical electrolyte modification methods, including cation/anion substitution and synthesis strategies, on Me─X bond characteristics. Our findings provide a comprehensive understanding of the role of covalent bonding in influencing the SIT and ionic conductivity of halides, offering a novel perspective on ionic transport through the lens of chemical bond characteristics.

Results and Discussion

SIT of Li₃InBr₆

The Li₃InBr₆ compound was synthesized via a co‐melting solid‐state reaction (see methods for details). A nearly phase‐pure polycrystalline powder with no discernible impurities was confirmed by synchrotron‐based XRD and time‐of‐flight neutron diffraction (Figures 1a,b and S1). Combined Rietveld refinement of both X‐ray and neutron diffraction data revealed a cubic structure with the Fd‐3m space group for Li₃InBr₆ at room temperature, rather than the previously reported C2/m symmetry.^{[ 38 ]} This structure is isostructural to other spinel materials, including Li₂MgCl₄, Li₂MnCl₄, and Li₂Sc_2/3Cl₄. Detailed crystallographic parameters are provided in Table S1. Structural analysis of the Li₃InBr₆ spinel indicates a disordered distribution of In³⁺ and Li⁺ within the edge‐sharing octahedra (Wyckoff position 16d). Each 16d site shares faces with two partially occupied tetrahedral 8b sites containing Li⁺, while each 8b site shares faces with four 16d sites (Figure 1c,d). This interconnected polyhedral network suggests a potential 3D Li⁺ diffusion pathway.

Figure 1.

The crystal structure and SIT of Li₃InBr₆. a) Synchrotron‐based XRD and b) neutron diffraction patterns with the Rietveld refinement results of Li₃InBr₆. c) The refined crystal structure of spinel Li₃InBr₆. d) The structure illustrating Wyckoff position 16d with (Li/In)Br₆ octahedra and Wyckoff position of 8b with LiBr₄ tetrahedra. e) Arrhenius plots of the Li₃InBr₆ from 253 to 348 K. f) The thermogravimetry‐DSC result of Li₃InBr₆. g) Temperature‐dependent synchrotron‐based in situ XRD patterns of Li₃InBr₆. h) Temperature dependence of the lattice parameters and (Li/In)Br₆ polyhedral volume in Li₃InBr₆ derived from Rietveld refinement of XRD results.

Ionic conductivity measurements of Li₃InBr₆ were conducted using electrochemical impedance spectroscopy (EIS) over a temperature range of 253 to 348 K. Arrhenius plots of the ionic conductivity data are presented in Figure 1e. A notable SIT is evident between 303 and 320 K, characterized by a sharp increase in ionic conductivity by one to two orders of magnitude, reaching 1.0 mS cm⁻¹ at 348 K. The SIT in Li₃InBr₆ is further corroborated by the impedance spectra in the dispersive regime (Figure S2), where the plateau corresponds to the ionic diffusion resistance. Differential scanning calorimetry (DSC) analysis (Figure 1f) revealed an endothermic peak at 320 K, coinciding with the observed ionic conductivity transition in Li₃InBr₆ and confirming the occurrence of the SIT. Interestingly, despite a clear endothermic peak, in situ synchrotron‐based XRD patterns show no change in the crystal structure across the transition. As highlighted in Figure 1g, all observed diffraction peaks, including the (222), (400), (440), and (840) reflections, exhibit a subtle shift to lower angles with increasing temperature, consistent with thermal‐induced lattice expansion. Analysis of the in situ XRD data indicates that the SIT in Li₃InBr₆ is not driven by a transformation of the crystal structure. The temperature‐dependent evolution of the lattice parameter and the average polyhedral volume, derived from these XRD data, is presented in Figure 1h and Table S2. While both parameters generally increase with temperature due to thermal expansion, a notable deviation from this trend is observed within the 313 to 333 K temperature interval. This specific temperature range correlates well with the transition temperature identified from our ionic conductivity measurements.

Local Chemical Bond Frustration

Given the subtle changes observed in the long‐range structure of Li₃InBr₆, we employed pair distribution function (PDF) analysis to investigate the local structural evolution across the SIT. Figure 2a,b presents the pair distribution functions, G(r), of Li₃InBr₆. Consistent with the in situ XRD data, the distance of the majority of atom–atom pairs, such as Br–Br pairs, gradually increases due to lattice expansion with increasing temperature. However, the peak representing the shortest distance at approximately 2.7 Å, attributed to In(Li)–Br pairs, shows a decrease as the temperature rises above the transition temperature T _c , contradicting the trend of lattice expansion.

Figure 2.

Local structural evolution of Li₃InBr₆ across the SIT. a) Temperature‐dependent in situ PDF plots of Li₃InBr₆. b) PDF in the low‐r region (2–5 Å) for Li₃InBr₆. c) Temperature‐dependent in situ In K‐edge XANES of Li₃InBr₆ collected in fluorescence yield (FLY). d) Corresponding In K‐edge EXAFS plot in R space, with a k ²‐weighing, of Li₃InBr₆ at different temperatures. e) The In K‐edge EXAFS region in k‐space of Li₃InBr₆ at different temperatures. f) Temperature dependence of the In–Br interatomic distance and Debye–Waller factor in Li₃InBr₆. The error bar is derived from the uncertainty of the EXAFS fitting process. g) Schematic illustration of the contraction of the In–Br octahedra during the ionic‐to‐covalent transition.

To confirm the abnormal evolution of the In─Br bonding in Li₃InBr₆, we conducted In K‐edge XAS to probe the electronic state and average local environment of the In atoms. Figure 2c displays the temperature‐dependent In K‐edge X‐ray absorption near‐edge structure (XANES) spectra of Li₃InBr₆. The absorption edge (E ₀) at an incident photon energy of 27 947 eV corresponds to electronic transitions from the In 1s to In 5p unoccupied states. The E ₀ and post‐edge features in the XANES spectra remain largely unchanged across the measured temperature range, indicating no significant variation in the In³⁺ oxidation state. Figure 2d shows the Fourier‐transformed (FT) k ²‐weighted χ(k) In K‐edge extended X‐ray absorption fine structure (EXAFS) spectra of Li₃InBr₆ collected in situ at various temperatures. The first coordination shell, located at R ≅ 2.64 Å (corrected for phase shift), is attributed to backscattering from the nearest‐neighbor Br atoms. The intensity of this peak decreases with increasing temperature from 298 to 333 K, reflecting an increase in the Debye–Waller factor (σ ²) due to enhanced thermal disorder, as outlined in Table S3. Notably, the In─Br bond length remains constant up to 313 K and then begins to decrease as Li₃InBr₆ is further heated to 333 K, as evidenced by the shift in the peak position in R‐space and the k‐space EXAFS spectra from 313 to 333 K (Figure 2e,f).

Therefore, the decrease of the first peak observed in the PDF data (Figure 2b) originates from the contraction of In─Br bonding. In contrast, the overall increase observed in the Br–Br distance in Figure 2b reflects that the thermal expansion effect on these other Br‐containing pairs outweighs the specific contraction of the In─Br units when considering the average of all Br–Br correlations captured by PDF. Comprehensive analysis of both long‐ and short‐range structural data reveals the abnormal contraction of [InBr₆]³⁻ octahedra in the local structure. This bond contraction likely reflects a change in In─Br bond character (e.g., increased covalency), which may alter the surrounding Li⁺ dynamic and diffusion channels to facilitate ionic transport, as illustrated in Figure 2g.

To understand how the observed temperature‐induced contraction of [InBr₆]³⁻ octahedra influences Li⁺ mobility, we performed AIMD simulations on Li₃InBr₆ at various temperatures. As shown in Figures 3a and S3a, the calculated radial distribution function (g(r)) profiles of In–Br and Br–Br pairs exhibit trends consistent with in situ PDF and EXAFS results: the primary In–Br peak gradually shifted from ∼2.73 to ∼2.68 Å with increasing temperature, while the shortest Br–Br peak shifted from ∼3.94 to ∼4.02 Å. Correspondingly, the Li–Br pair distribution broadened at higher temperatures (Figure 3b), indicating a more diffuse Li⁺ distribution. Figure 3c shows the mean square displacement (MSD) of Li⁺ in Li₃InBr₆ after a 50 ps relaxation period at various temperatures. The MSD values at 500 and 600 K are significantly higher than those at lower temperatures, indicating substantially enhanced Li⁺ mobility. This increased mobility is visualized in the Li⁺ trajectories at higher temperatures (Figure 3d). Notably, together with the faster Li⁺ diffusion, a collective motion of Br⁻ ions along the b–c plane is observed in Li₃InBr₆ at 600 K (Figure 3e). This anionic movement is directionally correlated with the observed Li⁺ diffusion pathways. Furthermore, the onset of significantly enhanced Li⁺ diffusivity coincides with a marked increase in Br⁻ displacement (Figure S3b), suggesting that collective anion motion likely facilitates superionic Li⁺ conduction. A similar collective anion motion has been recently observed in Li₃YCl₃Br₃, where the collective motion is attributed to the origin of the SIT.^{[ 25 ]} Collective anion motion is further substantiated by the refinement of Br displacement from experimental PDF results (Figure S4). This analysis shows that above the SIT, specific Br atoms within the unit cell undergo larger, anisotropic displacements aligned with the Li⁺ diffusion pathway.

Figure 3.

The AIMD simulation of Li₃InBr₆. The g(r) profiles of a) In–Br and b) Li–Br pairs in the Li₃InBr₆ after AIMD simulations at various temperatures. c) The Li⁺ MSD plot for Li₃InBr₆ at 100, 300, 400, 500, and 600 K from AIMD simulations. d) The Li⁺ probability distribution (yellow) and e) the Br⁻ probability distribution (brown) in Li₃InBr₆ at 100, 300, and 600 K from AIMD simulations.

We propose that the observed contraction of the In─Br bond and the collective anion movement upon heating are linked to a fundamental change in chemical bond character—specifically, a shift from predominantly ionic toward increased covalent bonding. From a lattice dynamics perspective, mobile ion migration inherently involves local electron cloud distortions. Increased directional covalent bonding can significantly influence the dynamics of local structural motifs, facilitating phenomena such as polyhedral rotation (akin to the paddle‐wheel effect) and collective anion motion. Furthermore, enhancing covalent interactions can alter lattice vibrational modes (Raman spectroscopy results in Figure S5) and influence Li⁺ mobility kinetics, thereby affecting both the activation energy and the pre‐exponential factor in the Arrhenius relationship for ion diffusion. Therefore, we conclude that the SIT observed in these halides is not driven by a conventional structural phase transition but originates from a thermally induced evolution of chemical bonding toward greater covalency.

To investigate the influence of covalency on the SIT of Li₃InBr₆, we introduced small amounts of yttrium (Y) or gallium (Ga) into the structure, forming Li₃In_0.9M_0.1Br₆ (M = Y, Ga). This substitution strategy was based on the electronegativity differences between these elements (1.22 for Y, 1.81 for Ga, and 1.78 for In in the Pauling scale), aiming to modulate the bond covalency. The successful synthesis of the target compounds was confirmed by the XRD analysis (Figure 4a). The electronic structure and bonding characteristics were investigated by In L ₃‐edge XANES, which directly probes the electronic structure of Br 4d and In 5s‐Br 4p bonding orbitals. The In L ₃‐edge XANES spectra of Li₃InBr₆, Li₃In_0.9M_0.1Br₆ (M = Y, Ga), and an InBr₃ reference are shown in Figure 4b. Three distinct features, labeled c', d', and e', observed in the XANES spectra are also evident in the corresponding differential plots (Figures 4c and S6). While feature c' shows minimal changes upon doping, features d' and e' exhibit noticeable variations. Molecular orbital (MO) analysis was conducted for a [InBr₆]³⁻ octahedral to interpret these spectral features.^{[ 43} , ^{44 ]} Figures 4d and S7 depict the MO diagram, based on the assumption that metal ion 5s and 5p atomic orbitals interact in the symmetry‐adapted linear combinations of the halide ligand group orbitals. According to the proposed MO interpretation, we assigned the three features as follows: c' corresponds to the In 2p → Br 4d transition, d' to the In 2p → non‐bonding 5s transition, and e' to the In 2p → antibonding σ* transition. The observed changes in features d' and e' upon doping with Ga³⁺ and Y³⁺ confirm the expected modulation of In–Br covalency. Specifically, Ga³⁺ doping increases the covalency, while Y³⁺ doping decreases it. This leads to the following order of decreasing degree of covalency (In–Br): InBr₃ > Li₃In_0.9Ga_0.1Br₆ > Li₃InBr₆ > Li₃In_0.9Y_0.1Br₆.

Figure 4.

Correlation between the In─Br bond covalency in Li₃InBr₆ and SIT. a) The synchrotron‐based XRD plots of Li₃In_0.9M_0.1Br₆ (M = Y, Ga) compared to pristine Li₃InBr₆. b) The In L₃‐edge XANES region of Li₃InBr₆, Li₃In_0.9M_0.1Br₆ (M = Y, Ga), and InBr₃ reference. c) The corresponding first‐order differential of XANES spectra in b). d) Molecular orbital diagrams for a [InBr₆]³⁻ octahedron. e) The plots of ionic conductivities verse temperature for Li₃InBr₆ and Li₃In_0.9M_0.1Br₆ (M = Y, Ga).

Ionic conductivity measurements reveal a consistent shift of the T _c upon doping (Figure 4e). Ga³⁺, with its higher electronegativity, increases the Me–Br covalency, resulting in a lower T _c and enhanced ionic conductivity. Conversely, Y³⁺, with its lower electronegativity, decreases the Me–Br covalency, leading to a higher T _c and reduced ionic conductivity. These findings underscore the crucial role of covalency and the Me in Li–Me–X compounds in governing the SIT and ionic conductivity in halide SSEs. By precisely tuning the covalency of the Me─X bond, the motion of local structural motifs can be modified, thereby influencing SIT behavior and ionic diffusion.

Generalization of SIT in Halide Electrolytes

The SIT phenomenon remains valid for other bromide‐based electrolytes. A series of lithium lanthanide bromides, Li₃LnBr₆ (Ln = Gd, Tb, Ho, Tm, and Lu), were studied via a co‐melting method using LiBr and LnBr₃ precursors. Synchrotron XRD analysis (Figure 5a) reveals that these compounds crystallize in monoclinic structures with the C2/m space group. A slight shift in XRD peaks toward lower angles is observed with increasing lanthanide ionic radius, indicating an expansion of the unit cell volume. The EIS measurements were conducted on cold‐pressed pellets of Li₃LnBr₆ to determine their ionic conductivities. The Arrhenius plots of conductivity versus temperature (Figure 5b) show a transition in activation energy for all Li₃LnBr₆ compounds within the measured temperature range (−248 to 348 K). Temperature‐dependent XRD measurements of Li₃GdBr₆ and Li₃LuBr₆ confirm the absence of significant structural changes across the transition, supporting the presence of SITs in these materials (Figure S8). Among these bromides, Li₃GdBr₆ exhibits the highest RT ionic conductivity (5.2 mS cm⁻¹ at 298 K) as shown in Figure 5c. Chronoamperometry measurements confirm a negligible electronic conductivity (4.1 × 10⁻⁹ S cm⁻¹) for Li₃GdBr₆ (Figure S9). The other Li₃LnBr₆ compounds show decreasing RT ionic conductivities with decreasing lanthanide ionic radius: Tb (4.6 mS cm⁻¹), Ho (1.8 mS cm⁻¹), Tm (0.37 mS cm⁻¹), and Lu (0.18 mS cm⁻¹).

Figure 5.

Generalization of SIT in halide electrolyte. a) Synchrotron XRD patterns of Li₃LnBr₆ (Ln = Lu, Tm, Ho, Tb, and Gd). Ionic conductivity as a function of temperature for b) Li₃LnBr₆ (Ln = Lu, Tm, Ho, Tb, and Gd). c) Summary of Li⁺ conductivity of the Li₃LnBr₆ SSEs at 25 and 75 °C. Synchrotron XRD patterns of d) Li₃LuCl _x Br_{6‐ x} (x = 0, 1, 2, 3), e) Li₃Tb_0.3Lu_0.7Br₆ and Li₃Tb_0.7Lu_0.3Br₆, and f) Li₃LuBr₆ synthesized by different methods. Arrhenius plots of Li⁺ conductivity of g) Li₃LuCl _x Br_{6‐ x} (x = 0, 1, 2, 3), h) Li₃Tb_0.3Lu_0.7Br₆ and Li₃Tb_0.7Lu_0.3Br₆, and i) Li₃LuBr₆ synthesized by different methods.

Building on our understanding of the SIT mechanism, we further investigated how typical electrolyte modification methods affect Me─X bond characteristics in halide materials. Substituting Br with Cl in Li₃LuCl _x Br_{6‐ x} (x = 0, 1, 2, 3) progressively suppresses the SIT within the measured temperature range (Figure 5d,g), because anion‐mixing influences the ionic‐covalent transition and, consequently, the SIT. We also investigated the impact of Me³⁺ cation mixing on the SIT. Mixed‐cation compounds Li₃Tb_0.3Lu_0.7Br₆ and Li₃Tb_0.3Lu_0.7Br₆ were synthesized and compared to Li₃TbBr₆ and Li₃LuBr₆, which share similar crystal structures (Figure 5e). The ionic conductivities of the mixed‐cation compounds (1.1 mS cm⁻¹ at 298 K) are nearly identical and appeared between those of Li₃TbBr₆ and Li₃LuBr₆. Notably, the SIT is significantly reduced or suppressed in the mixed‐cation compounds (Figure 5h), suggesting that Me³⁺ cation mixing hinders the ionic‐covalent transition. To further examine the role of local structural distortions in the SIT, we synthesized Li₃LuBr₆ via ball milling (BM). The BM‐synthesized sample exhibits a broadened XRD pattern compared to annealed and co‐melted samples, indicating reduced crystallinity and a high degree of local distortion (Figure 5f). While the room‐temperature ionic conductivities of all Li₃LuBr₆ samples are similar (0.18–0.26 mS cm⁻¹), the SIT is suppressed in the BM‐synthesized sample (Figure 5i). This suggests that mechanochemically induced stacking faults and local structural distortions can also influence the ionic‐covalent transition and the SIT. In addition, the conductivity enhancement dependent on covalency can be extended to chloride‐based SSEs. For instance, the introduction of dopants with higher electronegativity (e.g., Al³⁺, Fe³⁺), which increases the covalency of the metal‐halide bonds, has been shown to favorably impact the ionic conductivity of Li₂ZrCl₆.^{[ 21} , ⁴⁵ , ^{46 ]} This observation is consistent with the principles discussed for bromide‐based SSEs.

These findings demonstrate the generalizability of SITs in halide electrolytes and highlight the potential for tuning electrolyte properties via cation/anion substitution and synthesis methodology. A deeper understanding of the factors governing SITs enables the rational design of high‐performance SSEs tailored for specific applications. To demonstrate the practical viability of these bromide electrolytes, ASSBs were constructed using the highly conductive Li₃GdBr₆ SSE with either a LiCoO₂ (LCO) or a Se cathode and a Li–In alloy anode (Figure 6). Leveraging the high ionic conductivity of Li₃GdBr₆, the LCO‐based ASSBs exhibits promising performance at both RT and −10 °C. Cyclic voltammetry (CV) at 0.02 mV s⁻¹ (Figure 6b) revealed minimal discharge polarization, consistent with the high conductivity of Li₃GdBr₆. Consequently, the LCO cell delivers a specific capacity of 92.2 mAh g⁻¹ at a 1 C discharge rate even at −10 °C (Figure 6c,d). Furthermore, these ASSBs demonstrate excellent cycling stability, retaining 85% of their initial capacity after 500 cycles at 25 °C (Figure 6e) and 98% after 100 cycles at −10 °C (Figure 6f). Moreover, the Li₃GdBr₆ SSE is proved compatible with a high‐capacity Se cathode (configuration in Figure 6g). At 25 °C, the Se‐based ASSB exhibits a high initial discharge capacity of 675.8 mAh g⁻¹ (Figure 6h). It also demonstrates good rate capability, delivering specific capacities of 620.4, 592.6, 531.6, and 422.3 mAh g⁻¹ at C‐rates of 0.2, 0.4, 1, and 2 C, respectively (Figure 6i). Corresponding charge/discharge curves are provided in Figure S10. Owing to its lower T _c and high ionic conductivity, the bromide electrolyte facilitates the development of low‐temperature ASSBs and high‐capacity Se cathodes.

Figure 6.

Electrochemical performance of ASSBs using the Li₃GdBr₆ SSE. a) Schematic of the LCO‐based ASSB configuration. b) Comparison of CV curves of the LCO‐based ASSB at 25 versus −10 °C. c) The rate capability from 0.05 to 1 C at −10 °C with a fixed charging rate of 0.05 C and d) corresponding discharge curves. e), f) The cyclic performance of LCO‐ASSBs at e) 25 °C and f) −10 °C, respectively. g) Schematic of the Se‐based ASSB configuration. h) Initial galvanostatic charge–discharge profile of the Se‐based ASSB at 25 °C. i) Rate capability of Se‐based ASSB at 25 °C.

Conclusion

In conclusion, using Li₃InBr₆ as a model system, this study elucidates the critical correlation between Me─X bond covalency and SIT in halide SSEs, particularly bromides. Through a combination of in situ XRD, PDF, and XAS studies and AIMD simulations, we demonstrate that the SIT in Li₃InBr₆ is not driven by a conventional crystallographic phase transition but by a thermally induced shift in bonding character from predominantly ionic toward increased covalency. This enhanced covalency promotes collective anion motion and enables a distinct ion transport mechanism. Furthermore, we establish that SITs are a common feature in a series of lithium lanthanide bromides, Li₃LnBr₆ (Ln = Gd, Tb, Ho, Tm, and Lu), with the highest RT conductivity reaching 5.2 mS cm⁻¹. Our findings also indicate that cationic and anionic polarizability, along with local structural distortions, significantly modulate this dynamic ionic‐covalent transition. Leveraging its relatively low T _c and high ionic conductivity, the Li₃GdBr₆ SSE enables promising performance in ASSBs. Specifically, it exhibits high discharge rate capability at −10 °C in ASSBs utilizing LCO cathodes and supports high discharge capacity in ASSBs employing Se cathodes. These findings deepen our understanding of the role of covalent bonding in the SIT of halide SSEs, providing valuable insights and guiding principles for the future rational design and optimization of high‐performance SSEs.

Author Contributions

J.F. designed the experiments and carried out the sample synthesis and characterizations. H.S. and S.W. performed the computational simulations. J.L. helped with the data analysis, electrochemical measurements, and manuscript writing. W.L., J.T.K., S.Z., X.H., and Y.H. helped with the synchrotron‐related measurements. J.L. performed neutron data collection and analyses. C.W., J.L., X.L., F.Z., and J.P. helped with interpreting and organizing the data. J.F. and J.L. discussed and wrote the paper. T.K.S. and X.S. supervised the project. All the authors helped to revise the final manuscript.

Conflict of Interests

The authors declare no conflict of interest.

Supporting information

Supporting Information

Fu J., Su H., Luo J., Li X., Liang J., Wang C., Kim J. T., Hu Y., Zhao F., Zhang S., Duan H., Hao X., Li W., Peng J., Liu J., Wang S., Sham T.‐K., Sun X., Angew. Chem. Int. Ed. 2025, 64, e202508835. 10.1002/anie.202508835

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.