Design and Synthesis of Cubic K3−2xBaxSbSe4 Solid Electrolytes for K–O2 Batteries

Developing K‐ion conducting solid‐state electrolytes (SSEs) plays a critical role in the safe implementation of potassium batteries. In this work, a chalcogenide‐based potassium ion SSE is reported, K3SbSe4, which adopts a trigonal structure at room temperature. Single‐crystal structural analysis reveals a trigonal‐to‐cubic phase transition at the low temperature of 50 °C, which is the lowest among similar compounds and thus provides easy access to the cubic phase. The substitution of barium for potassium in K3SbSe4 leads to the creation of potassium vacancies, expansion of lattice parameters, and a transformation from a trigonal phase to a cubic phase. As a result, the maximum conductivity of K3−2xBaxSbSe4 reaches around 0.1 mS cm−1 at 40 °C for K2.2Ba0.4SbSe4, which is over two orders of magnitude higher than that of undoped K3SbSe4. This novel SSE is successfully employed in a K–O2 battery operating at room temperature where a polymer‐laminated K2.2Ba0.4SbSe4 pellet serves as a separator between the oxygen cathode and the potassium metal anode. Effective protection of the K metal anode against corrosion caused by O2 is demonstrated.


Introduction
The rising demand for large-scale energy storage devices calls for the technological development of economical, safe, and high-energy-density rechargeable batteries beyond the current lithium-ion batteries (LIBs) on the market.Potassium batteries (KBs), including both K metal batteries (KMBs) and K-ion batteries (KIBs), are regarded as promising candidates considering their various benefits, particularly when employed in the DOI: 10.1002/adma.202306809stationary electrical energy storage systems (ESSs) [1][2][3][4] : 1) The low standard reduction potential of potassium (−2.93 V vs standard electrode potential (E 0 )) results in decreased cutoff potentials for negative electrodes, enabling the possibility of KBs to operate within a broad range. [1]2) KBs are also more cost-efficient than LIBs due to the rich abundance of potassium in Earth's crust (1.5 wt% vs 0.0017 wt% for Li) and the significantly cheaper price of the potassium salt used as the raw material for electrode fabrication, K 2 CO 3 , compared to Li 2 CO 3 . [4]Moreover, the substitution of copper foil in LIBs with aluminum foil as a current collector in KIBs offers further cost reduction and lighter current collector weight. [5]3) Compared to Li + and Na + ions, K + exhibits a weaker Lewis acidity resulting in the smallest Stokes' radius and the highest ion mobility among the three ions in both ester and ether electrolytes. [3,4]4) Among all alkali metals, potassium (K) is the lightest element that forms thermodynamically stable superoxide, making it capable of achieving highly reversible K-O 2 batteries without the need for oxygen reduction/evolution electrocatalysts, setting them apart from other metal-O 2 battery systems. [6]5) The room-temperature Na-K liquid alloy enables plating without dendrite formation and facilitates the development of innovative structures like flow batteries. [7]Moreover, semi-solid Na-K has exhibited a significant K + critical current density (CCD) that exceeds 15 mA cm −2 , whereas Li solid electrolytes experience metal filament penetration during deposition at a CCD ranging from 0.1 to 1 mA cm -2 . [8]) Graphitic carbon forms the stage-1 K-graphite intercalation compound known as KC 8 .This characteristic enables graphitic carbon to be used as a viable anode material, which is a unique advantage over sodium in terms of transferability from the wellestablished lithium-ion technology. [9,10]he development of K-ion conducting solid-state electrolytes (SSEs) plays a critical role in the safe implementation of potassium batteries, considering the highly reactive nature of K metal.Employing nonflammable solid-state electrolytes (SSE) with good mechanical properties rather than organic liquid solvents can improve safety while simultaneously suppressing dendrite formation.This enables the secure and long-lasting performance of KIBs, including K-O 2 batteries and K-S batteries. [11]oreover, because only cations act as the charge carriers migrating across the immobile anion framework, SSEs possess high ionic transference numbers (close to one), while traditional organic liquid electrolytes only show ionic transference numbers below 0.2. [12]he study of K-ion conducting SSEs is still in its infancy when compared with that of Na-ion conductors.Johrendt et al. summarized the scarcity of known solid-state K-ion conductors, which are primarily oxide-based materials that achieve conductivities suitable for practical use (above 10 −4 S cm −1 ) only at high temperatures (300-400 °C). [13]Notably, commercial K-''-alumina achieves ionic conductivity of 8 × 10 −4 S cm −1 at room temperature, but its production cost is high due to sintering temperatures exceeding 1000 °C, limiting practical applications. [14,15]Recent findings include Johrendt et al.'s report on T5 KSi 2 P 3 with ambient conductivity up to 2.6 × 10 −4 S cm −1 , and our group's discovery of Ba-doped K 3 OI antiperovskite demonstrating excellent stability against reactive K metal. [13,16]tudying the crystal structure of effective Na-ion conducting materials can provide valuable insights into the key factors that enable fast ion transport in solids, which facilitates the discovery and design of promising K-ions conductors.Among the known sodium superionic conductors, thiophosphate-based solid electrolytes (Na 3 PnCh 4 (Pn = P, Sb; Ch = S, Se), Na 11 Sn 2 PnCh 12 (Pn = P, Sb; Ch = S, Se)) have stood out owing to their high ionic conductivities and excellent mechanical properties.In 1992, Jasen group first synthesized and reported a mild temperature (50 °C) ionic conductivity of 4.17 × 10 −6 S cm −1 for tetragonal Na 3 PS 4 (t-Na 3 PS 4 ) through solid-state reaction, which was overlooked due to its poor conductivity. [17]However, extensive research has been conducted to enhance Na 3 PS 4 's ionic conductivity to above 1 mS cm −1 using a variety of methods since Hayashi et al. reported in 2012 that ball-milling stabilized the cubic phase Na 3 PS 4 (c-Na 3 PS 4 ) with a high ambient temperature Na + ion conductivity of 2 × 10 −4 S cm −1 . [18][20][21][22][23][24] Moreover, cation doping with larger Ca 2+ was reported to help stabilize c-Na 3 PS 4 as well as introduce Na vacancies.Aliovalent replacement of Na + in tetragonal phase Na 3 PS 4 with Ca 2+ leads to a cubic phase Na 3−2x Ca x PS 4 with Na vacancies, which significantly improves Na-ion transport. [25]a 2.88 Sb 0.88 W 0.12 S 4 and Na 2.9 Sb 0.9 W 0.1 S 4 have also been reported to possess a cubic unit cell and the activation energy of 0.18 eV with ionic conductivity reaching around 32 mS cm −1 , exceeding the performance of all benchmark electrolytes.[20] These thiophosphate-based electrolytes possess the crystal structure made up of mobile Na cations and an open bodycentered cubic (bcc) anion framework that is composed of isolated polyhedral.In general, the building units of the skeleton are tetrahedral polyanions with central atoms from 4A or 5A groups, and the ligands Ch are S or Se like PCh 4 3− , SbCh 4 3− as well as SnCh 4 4− .[26] These polyanions form isotropic3D percolating channels while mobile cations Na + are dispersed at the interstitial sites.[27][28][29] Thus, we speculated that the bcc anion framework might be favorable for fast K-ions migration.Recently, our group reported a sulfide-based K-ions conductor -K 3 SbS 4 with an orthorhombic crystal structure, where the SbS 4 polyanions form a distorted bcc anion framework.[30] Although -K 3 SbS 4 can transform to cubic phase K 3 SbS 4 (-K 3 SbS 4 ) at a high temperature (above 370 °C), we were unable to test the ionic conductivity of -K 3 SbS 4 due to the limitation of our home-made Swagelok electrochemical cell.
Herein, we report the preparation of a new type of K-ion conducting material, K 3 SbSe 4 , which exists in two different crystal structures.The reversible trigonal to cubic phase transition of K 3 SbSe 4 was observed at a mild temperature (50 °C), which is the lowest among the related compounds.Moreover, the crystal structure of the new phase, c-K 3 SbSe 4 , was reported and resolved for the first time via employing single crystal X-ray diffraction.Attempts were made to stabilize the cubic phase at room temperature and improve the conductivity via Ba doping, introducing K vacancies, enlarging the lattice parameter, and tuning the defects.The maximum conductivity of K 3−2x Ba x SbSe 4 reaches around 0.1 mS cm −1 at 40 °C and the activation energy of the Ba-doped compound with the composition of K 2.2 Ba 0.4 SbSe 4 (0.28 eV) is comparable with those of other superionic conductors (usually 0.2-0.3eV).A polymer laminated K 2.2 Ba 0.4 SbSe 4 pellet is applied to construct room temperature two-compartment K-O 2 batteries in which the O 2 cathode and the K metal anode are separated by the K 2.2 Ba 0.4 SbSe 4 pellet.This two-compartment K-O 2 battery prototype enables the protection of the K metal anode by forming a stable SEI layer and blocking the diffusion of O 2 into the anode side.

Synthesis, Characterization, and Conductivity of K 3 SbSe 4
The synthesis of K 3 SbSe 4 was conducted via a liquid phase method with ethylenediamine (EDA) as the solvent based on the reaction 3K+Sb+4Se → K 3 SbSe 4 . [31]In more detail, elemental K, Sb, and Se powders with a molar ratio of 3:1:4 were dissolved in EDA, stirred at room temperature for 12 h, and then heated at 100 °C for 12 h.The solvent was removed, and the resulting powder was dried under vacuum at 120 °C for 24 h to remove the residual EDA.Further experimental details on the liquid phase reactions can be found in the Supporting Information.
Figure 1a shows the powder X-ray diffraction (PXRD) pattern of the synthesized K 3 SbSe 4 and the Rietveld refinement with an R3c trigonal structure (a = b ≠ c;  =  = 90°,  = 120°; K at 18b site, Sb at 6a site, Se1 at 18b sites and Se2 at 6a sites). [31]No noticeable impurities are observed, and the obtained lattice parameters are a (= b) = 11.42146(11)Å, c = 13.64956(19)Å (Table S1, Supporting Information).The Raman spectrum of K 3 SbSe 4 shows three distinct peaks located at 202, 256, and 265 cm −1 .These three peaks are ascribed to the vibrations of Sb-Se bonds within the isolated SbSe 4 group, confirming the existence of tetrahedron SbSe 4 3-polyanions (Figure S1, Supporting Information). [28]Note that the crystallographic unit cell consists of two types of Se atoms, with Se1 located on 18b sites and Se2 located on 6a sites, while all Sb atoms are located on 6a sites.The Sb-Se1 bond distance is 2.475 Å which is longer than that of Sb-Se2 (2.473 Å).The Se1-Sb-Se1 angle is 109.601°whileSe1-Sb-Se2 angle is 109.341°,indicating a slightly distorted SbSe 4 tetrahedron with Sb atoms deviating from the center of tetrahedron along the c axis.
The differential scanning calorimetry (DSC) measurements were conducted to investigate the phase changes of the mate- rials upon heating (Figure 1b).It can be observed that there is a pair of reversible peaks appearing at 50 °C upon heating and 40 °C during the following cooling, indicating the existence of a solid-solid phase transition from the trigonal phase to an unreported phase.Temperature variable XRD measurements of K 3 SbSe 4 were further conducted to examine the hightemperature phase of K 3 SbSe 4 (Figure 1c).At 60 °C, a previ-ously unreported phase is observed, leading to the conclusion of the existence of a second polymorph of K 3 SbSe 4 labeled as c-K 3 SbSe 4 .
By employing single-crystal X-ray diffraction (XRD) at 60 °C, the crystal structure of the new phase, c-K 3 SbSe 4 , was unequivocally resolved for the first time as shown in the inset of Figure 1c (see Table S2, Supporting Information for crystallographic details, Supporting Information).The unit cell of c-K 3 SbSe 4 can be fully indexed to cubic space group I-43m with lattice parameters of a (= b = c) = 7.9915 (6) Å at 60 °C, which is isostructural with the reported sodium superionic conductors of c-Na 3 SbS 4 and c-Na 3 PS 4 . [32,33]Calculated atomic coordinates and structural parameters are summarized in Table S2 (Supporting Information).In the cubic cell, the tetrahedral SbSe 4 3− units construct a body-centered cubic anion framework and K ions occupy 25% of the 24f sites, which bridge all six-coordinated "interstitial" 6b sites (the edge and face centers).The presence of partial occupancy in 24f sites implies the diffusion of K + within the structure.This observation allows for the identification of diffusion pathways aligned with the lattice vectors a, b, and c (See the detailed discussion in the following section).In contrast to trigonal K 3 SbSe 4 where tetrahedral SbSe 4 3− units show slight distortion, one Se position situates simply at 8c sites with a nondistorted SbSe 4 tetrahedron in the cubic K 3 SbSe 4 .The Sb-Se distance in c-K 3 SbSe 4 is 2.475 Å, which is the same as that of Sb-Se1 bond in t-K 3 SbSe 4 , and the Se-Sb-Se angle is 109.471˚.
The trigonal to cubic phase transition observed from temperature variable XRD measurements matches the phase transition temperature obtained from DSC results.The existence of a trigonal to cubic phase transition was predicted in 2020 from a groupsubgroup analysis for K 3 SbSe 4 . [34]To the best of our knowledge, this work is the first to verify such a phase transition experimentally and to resolve the cubic phase via single-crystal XRD analysis.Moreover, the transition temperature is the lowest among the related compounds, including Na 3 PS 4 (261 °C), [33] Na 3 SbS 4 (110 °C) , [21] and K 3 SbS 4 (360 °C). [30]The cell parameters of c-K 3 SbSe 4 are much larger than those of cubic Na 3 PS 4 (a = b = c = 6.98481(3)Å), [35] cubic Na 3 SbS 4 (a = b = c = 7.1910 (7) Å) [36] as well as cubic K 3 SbS 4 (a = b = c = 7.84765(16) Å), [30] even though c-K 3 SbS 4 was obtained and tested at an ultrahigh temperature (360 °C).
The ionic conductivities of the hot-pressed K 3 SbSe 4 samples were measured inside a homemade high-temperature Swagelok electrochemical cell, using the AC impedance method in the temperature range from 75 °C to 20 °C.The Nyquist plot of K 3 SbSe 4 measured at room temperature was fitted with an equivalent circuit as depicted in Figure 1d, and the fitted parameters are listed in Table S3 (Supporting Information).The Arrhenius conductivity plot of K 3 SbSe 4 from 75 °C to 20 °C is depicted in Figure 1e, where the relationship between log  and 1/T follows the Arrhenius law.The activation energy (E a ) for K ion migration is then calculated based on the slope of the linear Arrhenius conductivity plot following the equation of  =  0 exp(-E a /k B T), where  0 denotes the pre-exponential parameter and k B represents the Boltzmann constant.The range of ionic conductivity is from 1.22 × 10 −7 S cm at 20 °C to 1.05 × 10 −6 S cm −1 at 75 °C.Note that there is a sudden ionic conductivity drop from 50 °C to 45 °C and the activation energy of the low-temperature phase changes to 0.35 from 0.28 eV of the high-temperature phase on the same time, which matches the phase transition temperature shown in DSC during the cooling process.The observed ionic conductivity drops with activation energy increase is attributed to the cubic to trigonal phase change that occurs when the temperature is below 45 °C, which is later supported by temperature variable XRD results as well as the bond valence site (BVSE) analysis of these two phases.

Bond Valence Site Energy Analysis of t-K 3 SbSe 4 and c-K 3 SbSe 4
To analyze how the trigonal to cubic phase change affects the energies of the K-ions migration pathways in the lattice, BVSE analysis is utilized to analyze the crystal structure of both trigonal and cubic phase K 3 SbSe 4 through SoftBV software developed by Adams et al. [37] The BVSE map of trigonal phase K 3 SbSe 4 (space group: R3c) is depicted in Figure 2a.The purple spheres represent the lowest energy point that are occupied by K-ions and the pink as well as gray spheres represent the saddle points, labeled as SP1 and SP2.According to the calculated migration map, Kions has an anisotropic 3D migration pathway with two types of saddle points labeled as SP1 (pink) and SP2 (gray).The obtained saddle point coordinates are summarized in Table S4 (Supporting Information).K-ions are connected by SP1 or SP2 along the plane through a K-SP1-K-SP2-K pathway while these planes are connected through SP1.On the other hand, the migration map of cubic phase K 3 SbSe 4 (space group: I-43m) is shown in Figure 2b.K-ions in cubic phase possess an isotropic 3D migration pathway with one type of saddle points located in the middle of two K ions, connecting all the edge and face center located K-ions in the lattice (Table S5, Supporting Information).
Figure S2 (Supporting Information) compares bond-valence energy landscapes (BVEL) of K-ion diffusion in the two phases of K 3 SbSe 4 : t-K 3 SbSe 4 and c-K 3 SbSe 4 .The cubic phase (0.646 eV) shows a lower migration barrier compared with the trigonal phase (0.861 eV) K 3 SbSe 4 .This result supports our initial temperature-dependent EIS tests where the cubic phase K 3 SbSe 4 possesses higher conductivity and the lower activation energy compared with those of trigonal phase K 3 SbSe 4 .Notably, the BVSE analysis is a static approach that ignores the columbic repulsion between K-ions as well as the mobility of the anion sublattice.Considering that the analysis cannot handle correlated ionic movements and the mobile anion framework, it can only provide us with a potential K migration map and a qualitative analysis of the energy barrier.

Ba-Doped K 3 SbSe 4 (K 3−2x Ba x SbSe 4 )
Given the key role played by the K vacancy in fast-ion diffusion of the material, we conceive a way to tune the defect and the ionic conduction of the material by doping a multivalence cation of Ba 2+ in K 3 SbSe 4 .Ba 2+ is chosen due to its similar size of 1.48 Å compared to that (1.52 Å) of K + , so that the dopant is expected to occupy a K site within the structure.The K 3−2x Ba x SbSe 4 (x = 0.1, 0.2, 0.3, 0.4, and 0.5) samples were prepared through the reaction mentioned above by replacing quantitative K with Ba followed by thermal annealing at 300 °C (Detailed procedure can be found in Supporting Information).The Raman spectra of K 3−2x Ba x SbSe 4 (x = 0.1, 0.2, 0.3, and 0.4) exhibit three Raman bands in the range of 150 to 400 cm −1 , which are associated with various modes of SbSe 4 3− tetrahedron, revealing that Ba doping would not break the Sb-Se bonds (Figure S1, Supporting Information).
Figure 2c and Figures S3 and S4 (Supporting Information) show the XRD patterns of K 3−2x Ba x SbSe 4 (x = 0, 0.1, 0.2, 0.3, 0.4, and 0.5).Without Ba doping (x = 0), doublet peaks appear which implies the trigonal phase.As the amount of Ba increases, however, the doublet peaks merge to become single peaks, reflecting the formation of the cubic phase.Moreover, there is a noticeable shift of all reflections towards lower two-theta angles as the Ba content increases, which is more evident in the zoomed-in area (Figure 2c).The observed peak downshift suggests a gradual expansion in the volume of the unit cell after introducing Ba to the system.As the phase starts to become less crystalline at x = 0.5, the doping range of Ba in K 3 SbSe 4 is determined to be 0 < x ≤ 0.4 (Figure S4, Supporting Information).At K 3−2x Ba x SbSe 4 (x = 0.1), these "single" peaks are asymmetric as depicted in Figure S3 (Supporting Information), implying the transition state between cubic and trigonal phases.The DSC result of K 3−2x Ba x SbSe 4 (x = 0.1) shows one endothermic peak at 29 °C upon heating, implying that the phase-transition temperature of K 3−2x Ba x SbSe 4 is decreased after Ba-doping (Figure S5, Supporting Information).
All K 3−2x Ba x SbSe 4 (x = 0.2, 0.3, and 0.4) patterns fit to the cubic phase with Ba partially occupying the K sites.Figure 2d presents the X-ray Rietveld refinement profile for K 2.2 Ba 0.4 SbSe 4 , where all the peaks could be indexed to the cubic structure (a = 8.0224(2) Å, Z = 2, I-43m, Table S6, Supporting Information).The Rietveld refinement and detailed results for other compositions like K 3−2x Ba x SbSe 4 (x = 0.2 and 0.3) are also depicted in Figures S6 and S7 and Tables S7 and S8 (Supporting Information).The lattice parameters of the K 3−2x Ba x SbSe 4 compounds are plotted in Figure 3a, demonstrating a linear cell volume increase upon different Ba doping concentrations in the range of 0 < x ≤ 0.4.The maximum lattice volume is achieved in K 2.2 Ba 0.4 SbSe 4 , which is around 2.5% larger than that of c-K 3 SbSe 4 at a mild temperature (50 °C).Moreover, the energy-dispersive X-ray spectroscopic (EDS) element mapping analysis of the series of Badoped compounds K 3−2x Ba x SbSe 4 (x = 0.2, 0.3 and 0.4) reveal the presence of Ba throughout the entire SEM area and there is no noticeable composition-dependent contrast, suggesting that all the elements, including Ba, are uniformly distributed in these compounds (Figures S8-S10, Supporting Information).EDS elemental analysis also predicts these compositions to be K 2.36 Ba 0.17 SbSe 4 , K 2.26 Ba 0.27 SbSe 4 , and K 2.13 Ba 0.36 SbSe 4 with the Ba content analysis error of 0.17 ± 0.019, 0.27 ± 0.015, and 0.36 ± 0.029, respectively.The composition determined experimentally through both EDS elemental analysis results and Rietveld refinement analysis are nearly identical to the nominal composition of K 3−2x Ba x SbSe 4 (x = 0.2, 0.3, and 0.4).The conductivity of the pelletized K 3−2x Ba x SbSe 4 (x = 0.2, 0.3 and 0.4) was evaluated using the AC impedance method.To prepare the pellets, the electrolyte powder was hot-pressed at 200 °C under 400 MPa for 1 h inside an argon-filled glovebox.The hotpressed K 2.2 Ba 0.4 SbSe 4 pellet possesses a high relative density of 98% with the density of 3.65 g cm -3 .The cross-section scanning electron microscopy (SEM) images of pelletized K 2.2 Ba 0.4 SbSe 4 clearly show a dense microstructure, revealing the intimate contact between the particles within the pellet and facilitating the fast ionic diffusion (Figure S11, Supporting Information).Due to the expanded lattice with higher Ba contents, among the series of K 3−2x Ba x SbSe 4 (x = 0.2, 0.3 and 0.4) compounds, K 2.2 Ba 0.4 SbSe 4 shows the highest room-temperature (20 °C) ionic conductivity of 4.45 × 10 −5 S cm −1 and the lowest activation energy of 0.28 eV for conduction that is comparable with those of other superionic conductors (usually 0.2-0.3eV) (Figure 3b, Figure S12, Supporting Information).

Conduction Characteristics and Electrochemical
The Nyquist plot at 20 °C for K 2.2 Ba 0.4 SbSe 4 pellet is depicted as an example in Figure 3c.The fitted parameters are listed in Table S9 (Supporting Information).The Nyquist plot exhibits a characteristic semicircle in the higher-frequency region, corresponding to the bulk and grain boundary resistance of the electrolyte.Additionally, a spike is observed in the lower-frequency region, which originates from the capacitive behavior of the interface between the blocking electrode and the solid-state electrolyte.Note that the room-temperature ionic conductivity of K 2.2 Ba 0.4 SbSe 4 is over two orders of magnitude higher than that of the K 3 SbSe 4 and its activation energy is the same as that of c-K 3 SbSe 4 obtained after phase transition.Through the partial replacement of K with Ba in K 3 SbSe 4 (denoted as K 3−2x Ba x SbSe 4 ), the generation of potassium vacancies is achieved, accompanied by a transition from a trigonal to a cubic phase as well as an enlargement of lattice dimensions.The enhancement in ionic conductivity primarily stems from the higher vacancy concentration, the formation of interconnected isotropic 3D channels, and the expansion of lattice parameters, which could shed light on new approaches for optimizing ionic diffusion in solid-state electrolytes.Moreover, K 3−2x Ba x SbSe 4 (x = 0.2, 0.3, and 0.4) show linear Arrhenius relation without any abrupt ionic conductivity change within the temperature range of 210 °C-20 °C, suggesting that there is no phase transition during this temperature range.The DSC profile of K 2.2 Ba 0.4 SbSe 4 also reveals that neither phase transition nor decomposition occurs in the range of 10 °C-400 °C (Figure S13, Supporting Information).
The electrochemical stability window of K 2.2 Ba 0.4 SbSe 4 was determined using linear sweep voltammetry (LSV) at a scan rate of 0.1 mV s −1 (Figure 3d).The measurements were conducted using the cell configuration of K metal/1 M potassium bis(fluoroslufonyl)imide (KFSI)-1,2-dimethoxyethane (DME) /K 2.2 Ba 0.4 SbSe 4 -carbon paper.Additional information regarding the fabrication of the K 2.2 Ba 0.4 SbSe 4 cathode and LSV measurements can be found in Figure S14 (Supporting Information).The LSV curve indicates that K 2.2 Ba 0.4 SbSe 4 remains electrochemically stable up to 4.0 V versus K + /K.To ensure the exclusion of any interference resulting from the decomposition of the liquid electrolyte, the electrochemical window of the bulk low-concentration electrolyte (LCE) was also estimated.No  significant anodic current is observed up to 5.0 V, while the cathodic current occurs below 1 V for LCE.Hence, it can be concluded that K 2.2 Ba 0.4 SbSe 4 possesses a superb electrochemical stability within the voltage range of 1.2 to 4.0 V versus K + /K.

Room-Temperature Two-Compartment K-O 2 Batteries
As mentioned in the introduction, potassium (K) is the lightest alkali metal that forms thermodynamically stable superoxide thereby enabling highly reversible K-O 2 batteries without the need for oxygen reduction/evolution electrocatalysts and setting them apart from other metal-O 2 battery systems. [6]K-O 2 batteries only produce KO 2 as the sole discharge product through the facile one-electron reduction of O 2 with high round-trip energy efficiency (>95%). [9,38,39]However, the cycle lifetime and the further application of K-O 2 batteries are impeded by the instability of K metal anode toward organic liquid electrolytes as well as O 2 crossover from the cathode side.With the conventional potassium hexafluorophosphate (KPF 6 )-DME electrolyte, a thick yellow byproduct layer was always found on the cycled K metal surface after battery failure, indicating that the unstable K/electrolyte interface and the continuous corrosion of metallic K are the limiting factors for the limited cycle life of K-O 2 batteries. [40,41]Although the KFSI-DME electrolyte was found to enable highly reversible K plating and stripping, FSI − anion is susceptible to the nucleophilic attack from superoxide formed on the carbon cath-ode because of weak SO 2 -F bond. [42,43]Therefore, a K-ion conducting solid-state electrolyte (SSE) with high ionic conductivity is highly desirable to enable a two-compartment cell that can preserve the K metal anode by forming a stable solid electrolyte interface (SEI) and blocking the diffusion of O 2 into the anode side.
In our study, a two-compartment K-O 2 single cell was designed (Figure S15, Supporting Information).This cell configuration effectively protects the K anode from the undesired O 2 crossover in liquid electrolyte by utilizing a piece of polymerlaminated K 2.2 Ba 0.4 SbSe 4 pellet as the separator. [44]Moreover, due to the solid electrolyte's impermeability to solvents, this design enables the utilization of distinct anolyte and catholyte to meet different requirements from anode and cathode.Thus, the KFSI/DME anolyte is adopted in the two-compartment K-O 2 cells, which is beneficial for improving the K plating and striping efficiencies at the anode side. [40]Moreover, although diethylene glycol dimethyl ether (DEGDME) is less stable with K anode compared with DME solvent, the less volatile DEGDME possesses a stronger chelating ability owing to the increased ether chain length.As a result, the utilization of the KPF 6 /DEGDME catholyte can not only minimize the influence of solvent loss upon purging but enhance the solubility of KO 2 as well, preventing the cathode from premature blockage and facilitating the cathode discharge/charge reaction through the solutionmediated route.The representative first cycle voltage profiles of the assembled K-O 2 battery are plotted with a coulombic efficiency of 80%.(Figure S16a, Supporting Information).The discharge product of the battery is identified as pure KO 2 through Raman spectroscopy analysis, which then fully decomposed during the subsequent charge process (Figure S16b, Supporting Information).Moreover, the ionic conductivity of K 2.2 Ba 0.4 SbSe 4 pellet calculated from the battery discharge-charge voltage gap is 3.32 × 10 −5 S cm −1 , which is very close to that measured from electrochemical impedance spectroscopy (4.45 × 10 −5 S cm −1 ).
It is further demonstrated that the two-compartment K-O 2 battery can sustain over 100 stable cycles with an average coulombic efficiency of around 94% (Figure 4a).As shown in the typical voltage profiles of cells incorporating solid-state electrolytes, there are noticeable discharge and charge plateaus throughout the cycling process despite a gradual decay in the discharge voltage plateau (Figure S17, Supporting Information).The surface characterization of the K anode with and without protection after cycling was also conducted to validate the effectiveness of SSEs in eliminating O 2 crossover effect.The Raman spectrum analysis of the K anode without SSEs protection indicates that the yellowish byproduct layer observed on the anode consists of KO 2 and K 2 CO 3 which can be attributed to the side reactions between K anode and O 2 crossover from the cathode.It has been revealed in our prior study that the accumulation of these byproducts on the anode is mainly responsible for the capacity decay and battery failure in conventional K-O 2 batteries. [45]In stark contrast, the K metal anode disassembled from the two-compartment K-O 2 battery shows no presence of a yellowish layer even after 100 cycles (Figure 4b).Moreover, Raman analysis confirms that no KO 2 or other byproducts are formed on the cycled K anode.
The SEI formed on the K metal surface through the decomposition of KFSI-DME electrolyte was investigated using X-ray photoelectron spectroscopy (XPS).Figure S18 (Supporting Information) shows the XPS spectra from the K 2p, C 1s, O 1s, S 2p, and F 1s regions.The S 2p region shows peaks that are attributed to ─SO 2 − fragment and KSO x species, arising from the S─F and further N-S bond cleavage.Moreover, the F 1s spectrum revealed the presence of two distinct fluorine species, namely S─F species (687.7 eV) and KF with a lower binding energy of 682.9 eV.In the C 1s region, the decomposition products of the DME solvent include both C─C and C═O species and the O 1s peaks depicted in Figure S18d (Supporting Information) only suggest the presence of S═O, C═O species.These results strongly suggest that the incorporation of SSEs as a separator is an effective approach to further boost the lifespan of K-O 2 batteries by effectively suppressing the formation and accumulation of KO 2 byproducts on the K anode.Furthermore, XPS was also applied to probe the surface composition of the SSEs pellet before and after cycling, where no noticeable changes were observed (Figure S19, Supporting Information).
In order to gain deeper insights into possible reasons for an average coulombic efficiency of only 94%, the potential chemical decomposition reaction between K 2.2 Ba 0.4 SbSe 4 with KO 2 was investigated by mixing fine K 2.2 Ba 0.4 SbSe 4 powder with commercial KO 2 powder in the DME solution, with the addition of crownether to facilitate the dissolution of superoxide anions (mole ratio of K 2.2 Ba 0.4 SbSe 4 / KO 2 /18C6 = 1/1/2).The mixture was stirred at room temperature for one week, dried under vacuum at room temperature to remove the remaining solvent and characterized by PXRD and XPS analysis.After reacting with KO 2 , the PXRD pattern of K 2.2 Ba 0.4 SbSe 4 (black line) changes after the reaction (red line), exhibiting the transition state between trigonal and cubic phase after reaction (Figure S20, Supporting Information).XPS spectra of K 2.2 Ba 0.4 SbSe 4 and the products formed by reacting with KO 2 are depicted in Figure S21 from K 2p, C 1s, Se 3p, O 1s, Sb 3d, and Ba 3d regions .In the Se 3p 3/2 region, two distinct species are detected: K 2.2 Ba 0.4 SbSe 4 (160.3eV) and SeO 2 (164.5 eV), indicating the oxidation of Se element via the reaction.Within the Sb 3d spectrum, the presence of distinct peaks at 529.5 and 538.9 eV could correspond to the Sb─Se bond in the 3d 5/2 and 3d 3/2 regions, respectively.A minor neighboring peak at 532.4 eV is attributed to the C─O bond from the crown ether.Eventually in the Ba 3d 5/2 region, the peak located at 780.1 eV can be assigned to K 2.2 Ba 0.4 SbSe 4 while the peak at 783.5 eV indicates the presence of the lost Ba that is coordinated by the crown ether.As a result, it is expected that despite the low level of reaction between KO 2 and K 2.2 Ba 0.4 SbSe 4 , the KO 2 can still partially react with SSEs, which may induce KO 2 loss as well as the decomposition of the SSEs.To eliminate the side reaction mentioned above, an intrinsically stable buffer layer (such as a K-″-alumina layer) may be incorporated to block the physical contact between the K 2.2 Ba 0.4 SbSe 4 layer and the catholyte.This is a potential solution to improve the initial coulombic efficiency of cell., which is subject to further research.

Conclusion
In summary, we reported a chalcogenide-based K ion conducting SSEs (K 3 SbSe 4 ) with a room temperature trigonal structure, which undergoes the phase transformation to a cubic phase at 50 °C.Partial substitution of K with Ba in K 3 SbSe 4 (K 3−2x Ba x SbSe 4 ) not only incorporates K vacancies to the system but leads to a transition from a trigonal phase to a cubic phase as well, where polyanions form isotropic 3D interconnected channels and the mobile cations K + are dispersed within the interstitial sites.Consequently, the maximum conductivity of K 3−2x Ba x SbSe 4 , particularly in the case of K 2.2 Ba 0.4 SbSe 4 , reaches approximately 0.1 mS cm −1 at 40 °C, which is over two orders of magnitude higher than that of the undoped K 3 SbSe 4 .Subsequently, a two-compartment K-O 2 cell was implemented, separating the O 2 cathode from the K metal anode utilizing K 2.2 Ba 0.4 SbSe 4 SSEs.This design effectively prevents O 2 crossover and enables the utilization of various anolyte and catholyte compositions at the electrodes, addressing the distinct requirements of the anode and cathode and thus contributing to the decent cycling performance of the twocompartment K-O 2 cell.
Synthesis of K 3 SbSe 4 and Ba-Doped K 3 SbSe 4 : The synthesis of the K3SbSe 4 or Ba-doped K3SbSe 4 was carried out in EDA.Typically, 58.5 mg K (1.5 mmol) was dissolved in 5 mL EDA with the formation of a deep blue solution.Elemental Se and Sb were added to the above deep blue solution in a stoichiometric proportion of K:Sb:Se = 3:1:4.After addition, the mixture was stirred at room temperature for 12 h and then heated at 100 °C for another 12 h.The entire reaction was conducted under an inert atmosphere.The solvent was then removed, and the obtained powder was dried under vacuum at 120 °C for 24 h.The Ba-doped K3SbSe 4 samples can be prepared in a similar procedure mentioned above just by partially replacing quantitative K with Ba.The synthesized powder after removing EDA solvent was ground for 10 min and cold-pressed into a pellet.The sample pellet was placed inside a quartz tube under an Ar flow, which was heated to 300 °C for 2 h, kept at 300 °C for 12 h, and cooled down naturally to room temperature.
Bond Valence Site Energy Calculations of K 3 SbSe 4 : K + ion migration pathways in the crystal structure of tetragonal phase K3SbSe 4 (t-K 3 SbSe 4 ) and cubic phase K3SbSe 4 (c-K 3 SbSe 4 ) were analyzed by bond valence site energy (E BVSE ) calculations using the softBV bond valence parameter set.In this approach, the bond valence site energy of the mobile ion (K-ion in this study) can be considered as a Morse-type interaction with the environmental anions and a Coulomb repulsion of K-ion and other immobile cations.The Morse-type interaction can be separated into attractive interaction and short-range Born repulsion of K-ion and anions.
Characterization: The phase purity of the powder sample was characterized using an X-ray diffractometer (Bruker D8 Advance, Cu K source, 40 kV, 40 mA).Temperature-variable XRD of K3SbSe 4 was characterized using an X-ray diffractometer (Bruker D8 Venture Diffractometer-Cu -TXS) with a dry nitrogen (N2) flow.The powder sample of K3SbSe 4 was sealed in a capillary tube to prevent sample from air exposure and the 2-theta scan range is from 10°to 80°.The XRD results were fitted by Rietveld refinement to get the lattice parameter information.The single crystal X-ray diffraction studies were carried out on a Bruker Kappa Photon III CPAD diffractometer equipped with Mo K  radiation ( = 0.71073 Å).A 0.059 × 0.033 × 0.025 mm piece of an orange block was mounted on a Cryoloop with clear enamel.Data were collected in a N2 gas stream at 333.15(2) K using ϕ and ϖ scans.The crystal-to-detector distance was 80 mm using variable exposure time (2-5 s) depending on  with a scan width of 0.75°.Data collection was 100% complete to 25.00°in  (0.83Å).A total of 5682 reflections were collected covering the indices, -11< = h< = 11, -11< = k< = 11, -11< = l< = 11.167 reflections were found to be symmetrical independent, with a R int of 0.0197.Indexing and unit cell refinement indicated a body-centered, cubic lattice.The space group was found to be I-43m.The data were integrated using the Bruker SAINT software program and scaled using the SADABS software program.Solution by dualspace method (SHELXT) produced a complete phasing model for refinement.All atoms were refined anisotropically by full-matrix least-squares (SHELXL-2014).Crystallographic data are summarized in Table S2 (Supporting Information).Differential scanning calorimetry (DSC) measurements were done under a dry N 2 flow.Around 5-10 mg of air-sensitive powder samples were sealed by using Al hermetic pans in an Ar-filled glovebox.The cooling and heating rates were 5 °C min −1 .Scanning electron microscopic (SEM) analysis was performed using an FEI/Philips Sirion Field Emission SEM.For cross-section morphology characterization, a custom air-free SEM sample holder was used to prevent exposure of samples to the ambient air.EDS was also performed to explore the elemental composition and distribution in specific samples.The powders of electrolyte samples were sealed in a Borokapillaren capillary tube, spectra were then collected using a Renishaw inVia Raman microscope under a 633 nm edge laser excitation at room temperature.X-ray photoelectron spectroscopy (XPS) spectra were collected on a Nexsa G2 XPS System.All the spectra were calibrated with reference to the C─C peak binding energy at 284.8 eV.The curves were fitted using the CasaXPS software with a combined Gaussian-Lorentzian profile after subtracting the background.
Fabrication of K 2.2 Ba 0.4 SbSe 4 Cathode and Super-P Cathode: The K 2.2 Ba 0.4 SbSe 4 cathode was fabricated by mixing the active material of K 2.2 Ba 0.4 SbSe 4 powder, super-P carbon powder (MTI Corporation), and poly(vinylidenefluoride) (PVDF) (Sigma-Aldrich) at a weight ratio of 30:60:10 in N-methylpyrrolidone (NMP) (Sigma-Aldrich) drops.The slurry was ground thoroughly and then pasted onto the carbon paper, followed by drying at 90 °C for 48 hours under a vacuum.The super-P cathode was fabricated by following the same process except only super-P carbon powder and PVDF were mixed at a weight ratio of 85:15 in the NMP solvent.
Electrochemical Measurements: For the ionic conductivity tests, the sample pellets were hot-pressed at 200 °C under 400 MPa for 1 h inside the glovebox.Pt was coated onto both sides of the pellet by using a sputter coater to ensure sufficient electrical contact.Due to the air sensitivity of the sample, the alternating current (AC) impedance measurement was done with a homemade high-temperature electrochemical double-sealing clamp cell using a Gamry Reference 600 potentiostat.An AC voltage with an amplitude of 20 mV was applied to the cell.The frequency range was from 1 MHz to 0.01 Hz.The temperature was controlled by a Thermo Scientific muffle furnace and was monitored by a digital thermocouple.The resulting data were fitted using the ZView software.R s is the inner resistance of the set-up (cell + wire); R t is the total resistance of the electrolyte, including both the bulk resistance and grain boundary resistance; CPEdl is the constant phase element of the blocking electrode-electrolyte interface; CPEd is the constant phase element of the bulk pellet sandwiched between the two blocking electrodes.The values of ionic conductivity were determined from the measured resistance by using the expression:  = L/(A*R t ), where L is the thickness of the solid electrolyte, A is the area of the electrolyte, and R t is the final total resistance fitted from the Nyquist plot using the Zview software.The details of the fitting are shown in Figure 1d and Table S3 (Supporting Information).The electrochemical window of K 2.2 Ba 0.4 SbSe 4 was measured by linear sweep voltammetry (LSV) at the scan rate of 0.1 mV s −1 .The measurement was performed on the two-electrode cell composed of a K metal reference/counter electrode, a K 2.2 Ba 0.4 SbSe 4 working electrode, and 1 m KFSI/DME as the electrolyte.The measurement was carried out within a voltage range of 0.5-5.5 V (vs K + /K) to avoid interference with K plating/stripping.Typically, the measurements were conducted by increasing the voltage from the open circuit potential (OCP) to 5.5 V and decreasing the voltage from OCP to 0.5 V in two parallel cells separately.The intrinsic electrolyte window of 1 m KFSI/DME was estimated by using the blank super-P cathode within a voltage range of 0.5-5.5 V (vs K + /K).
Assembly of Two-Compartment K-O 2 Batteries Based on the As-Prepared Solid-State Electrolyte: The two-compartment K-O 2 cell configuration was constructed based on the reported K-O 2 cell configuration, where the O 2 cathode and the K metal anode were separated through a piece of a polymer laminated K 2.2 Ba 0.4 SbSe 4 pellet (Figure S14, Supporting Information).The K 2.2 Ba 0.4 SbSe 4 disc with a diameter of 13 mm and a thickness of 1 mm was polymer-sealed to a stainless-steel ring in an Ar-filled glovebox (<0.5 ppm of H 2 O and <1.5 ppm of O 2 ) and the active area of the pellet is 0.7 cm 2 .The K-O 2 cells were assembled by stacking a K metal anode (99.5%, Sigma-Aldrich, diameter 12 mm), a 19 mm tri-layer (PP-PE-PP) membrane (25 μm thickness, Celgard), a 15 mm glass fiber separator (GF/A, Whatman) soaked with 100 μL 1 m KFSI/ DME anolyte, a 15 mm trilayer (PP-PE-PP) membrane, the polymer laminated K 2.2 Ba 0.4 SbSe 4 pellet, a 15 mm trilayer (PP-PE-PP) membrane, a 15 mm glass fiber separator (GF/A, Whatman) soaked with 100 μL 0.5 m KPF 6 /DEGDME catholyte, a 15 mm trilayer (PP-PE-PP) membrane, and a carbon cathode (effective diameter of 15 mm, 0.2 mm in thickness).Additionally, 500 mL DEGDME solvent was added to the O 2 tank to saturate the O 2 .The O 2 chamber was eventually purged with high-purity O 2 (99.993%,UHP, gauge pressure = 1.5 psi).Galvanostatic discharge/charge tests were carried out using the Neware battery analyzer (BST8-WA) with the cutoff voltages set at 2.0 and 3.0 V (vs K + /K).A constant current of 0.05 mA (3.1 mA g −1 ) was applied if not specifically mentioned.The values of ionic conductivity were determined from the calculated resistance R by using the expression:  = L/(A*R), where L is the thickness of the solid electrolyte, A is the active area of the electrolyte, and R is calculated using the voltage gap and the current.
Chemical reactivity of K 2.2 Ba 0.4 SbSe 4 SSEs with KO 2 : The stability of K 2.2 Ba 0.4 SbSe 4 SSEs with KO 2 was examined by the following experiments and characterized by XPS and PXRD.The initial step involved introducing KO 2 powder into DME solvent, followed by the addition of 18-Crown-6 to facilitate the dissolution of superoxide.Subsequently, K 2.2 Ba 0.4 SbSe 4 powder was incorporated into the mixture and stirred at room temperature for a duration of one week (mole ratio of K 2.2 Ba 0.4 SbSe 4 /KO 2 /18C6 = 1/1/2).After the reaction, the solvent was removed, and the remaining precipitate was dried under a vacuum at room temperature to remove the remaining solvent.
Accession Codes: CCDC 2269525 contains the supplementary crystallographic data for this paper.This data can be obtained free of charge from The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif .

Figure 1 .
Figure 1.a) Powder X-ray diffraction (PXRD) data and Rietveld refinement for trigonal K 3 SbSe 4 .Black dots and red lines denote the observed and calculated XRD patterns, respectively.The green vertical lines mark the position of the reflections for K 3 SbSe 4 .The difference between the observed and calculated patterns is indicated by the grey line.The inset shows the R3c trigonal structure of the material, with K in purple, Sb in orange, Se1 in yellow, and Se2 in green.b) DSC results of K 3 SbSe 4 .c) Temperature variable XRD results of K 3 SbSe 4 at 25 and 60 °C using Cu target as the X-ray source.The trigonal to cubic phase transition temperature matches with the phase transition temperature observed in DSC.d) Impedance spectroscopy Nyquist plot of K 3 SbSe 4 sample at room temperature, showing the experimental data (black) and the fitting result from the equivalent circuit model (red).e) The Arrhenius plot of the ionic conductivity for K 3 SbSe 4 from 75 to 20 °C.The ionic conductivity test was conducted by cooling from high temperature to low temperature.

Figure 2 .
Figure 2. K-ion migration map of a) the trigonal phase and b) the cubic phase of K 3 SbSe 4 .Se atoms are eliminated to enhance clarity.The different colors show different energy isosurfaces.Darker colors indicate lower isosurface energy while paler colors indicate higher isosurface energy.c) Stacked, normalized PXRD patterns of K 3−2x Ba x SbSe 4 (x = 0, 0.2, 0.3, and 0.4) upon different Ba contents.The set of vertical lines at bottom represents a simulation of the cubic undoped K 3 SbSe 4 structure obtained at 50°C.PXRD peak evolution of K 3−2x Ba x SbSe 4 (x = 0, 0.2, 0.3, and 0.4) in detail within 2 = 26.5°-28°(highlighted in blue) is shown on the right side.d) Rietveld refinement plot of PXRD data for cubic K 2.2 Ba 0.4 SbSe 4 .Black dots and red lines denote the observed and calculated XRD patterns, respectively.The green sticks mark the position of the reflections for c-K 3 SbSe 4 .The difference between the observed and calculated patterns is indicated by the grey line.

Figure 3 .
Figure 3. a) Refined lattice parameters of K 3−2x Ba x SbSe 4 , showing the structural changes during substitution.b) The Arrhenius plots of the K ion conductivity for K 3−2x Ba x SbSe 4 (x = 0.2, 0.3 and 0.4) electrolytes tested by cooling from 210 °C to 20 °C.The activation energies of K migration within different compounds are labeled in the plot, revealing lower activation energy with higher Ba contents.c) Impedance spectroscopy Nyquist plot of K 3−2x Ba x SbSe 4 (x = 0.4) sample at room temperature, showing the experimental data (black) and the fitting result from the equivalent circuit model (red).d) The LSV curve of the K/K 2.2 Ba 0.4 SbSe 4 -carbon paper cell filled with the electrolyte 1 m KFSI-DME.The cell was tested at 20 °C with a scan rate of 0.1 mV s −1 , and current densities below 10 μA cm −2 are highlighted in the light blue region.

Figure 4 .
Figure 4. a) Long-term cycling performance and coulombic efficiencies of the two-compartment K-O 2 battery at a current of 0.05 mA.b) Raman spectra obtained for the following samples: the pristine bare K anode (black), the K anode collected from the two−compartment K-O 2 battery after 100 cycles (blue), and the unprotected K anode that is completely covered by a yellowish surface layer as the control sample (red).Inset: optical photos of these samples.