Li-In-S composite foil with built-in electric fields to stabilize Li/Li6PS5Cl interface for long-life all-solid-state batteries Authors: Chong Liu, Ruoyu Wang, Dewen Wang, Liyun Zhao, Taiguang Li, Tenghui Wang, Butian Chen, Tianran Zhang, Xiangfeng Liu Published: 2026 ABSTRACT AbstractSulfides-based all-solid-state lithium batteries show great potential due to their high safety and high energy density, yet severely suffer from sulfide electrolytes/Li interfacial instability and short cycle life. Here we propose a Li-In-S composite foil comprising Li2S, LixIn, and LiInS2 to stabilize the Li/Li6PS5Cl interface by constructing built-in electric fields, thereby enabling long-life all-solid-state lithium batteries. The work function difference between Li2S and LixIn generates built-in electric fields at the heterointerface, which traps the interfacial electrons and restricts their transfer to Li6PS5Cl, thereby suppressing the interfacial side reactions. Simultaneously, the built-in electric fields promote Li+ adsorption and diffusion inhibiting lithium dendrite growth. Li symmetrical cells display high critical current density over 4 mA cm−2 and Li plating/stripping stability over 2000 h at 1 mA cm−2. The assembled full cells with LiCoO2 and LiNi0.8Co0.1Mn0.1 (NCM811) demonstrate high capacity retention of 93% over 2000 cycles at 1 C and 87.7% over 1000 cycles at 1 C, respectively. Moreover, Li-In-S|Li6PS5Cl|NCM811 full cell shows rate capability up to 4 C. This work offers useful insights into the design of stable interfaces in all-solid-state batteries.Subject terms: Batteries, Batteries Introduction Lithium metal (Li) has emerged as a promising negative electrode material for high-energy batteries owing to its inherent advantages, including high theoretical specific capacity and low electrochemical potential1,2. Despite these advantages, the practical application of Li metal is far from succeeding since several critical challenges, such as uncontrolled lithium dendrite growth and safety issues associated with conventional liquid electrolyte, are not yet resolved3,4. In this context, all-solid-state lithium batteries (ASSLBs) have emerged as a promising alternative, in which liquid electrolytes are replaced with solid electrolytes to enhance both energy density and safety5,6. Among various solid electrolytes developed so far, sulfide electrolytes have garnered significant interest owing to their high ionic conductivity and formability7,8. However, the practical implementation of sulfide-based ASSLBs is hampered by interfacial incompatibility between Li metal and sulfide electrolyte, stemming from the significant difference in work functions between sulfide electrolytes and Li metal9,10. This interfacial incompatibility causes parasite reactions to produce undesirable decomposition products (Li2S, Li3P, etc.) that result in the elevated interfacial resistance and lithium dendrite growth, thereby restricting the electrochemical performances for sulfide-based ASSLBs11,12. To address these problems, reducing lithium activity has been proven to be an effective approach to stabilize the Li/sulfide electrolytes interface13,14. It has been demonstrated that Li alloys exhibit improved interfacial compatibility in sulfide-based ASSLBs15,16. Unfortunately, there is still a thermodynamic instability concern at the Li alloy/sulfide interface, along with the high cost, low Li+ diffusion rate, and reduced energy and volumetric densities of Li alloys17. Recently, composite Li with mixed conducting interphase (MCI), including a conductive Li alloy phase and other ion-conducting components18,19, is a better choice beyond Li-alloy. This architecture not only mitigates the chemical reactivity of lithium metal but also maintains high Li⁺ diffusion rates. The currently accepted interfacial stabilization mechanism is primarily attributed to the synergistic effects of the MCI in composite Li. However, it should be noted that the MCI may also facilitate electron transfer from lithium to the sulfide electrolyte, resulting in more severe interfacial side reactions, particularly when the MCI contains high-valence metal elements20–24. The reason for this is that the MCI eliminates the bottleneck of charge transfer at the traditional solid-solid interface and reduces the activation energy barrier10. Such controversy casts doubt on the effectiveness of the composite Li in suppressing the interfacial reactions. Therefore, rationally designing the component and in-depth understanding of the suppression factor of composite Li is highly desirable, yet remains a significant challenge. There are obvious differences in charge distribution, work function, and energy band structures among various components of MCI. These differences usually lead to the creation of built-in interfacial electric fields at the heterointerface between dissimilar phases. However, the fundamental role of built-in electric field (BIEF) in composite Li on suppressing interfacial reactions towards sulfide electrolyte has been largely overlooked. Herein, we design a composite Li-In-S foil comprising Li2S, LixIn, and LiInS2 with BIEFs and find that the BIEFs in the composite negative electrode play a critical role on suppressing undesired interfacial reactions between sulfide electrolyte and Li metal. The Li-In-S composite foil is fabricated by the dissolution-reprecipitation process for industrially relevant ASSLBs applications. The robust BIEF is created at the Li2S/LixIn heterointerface due to the distinct work functions between Li2S and LixIn. By comprehensive characterizations and theoretical study, we reveal that BIEF induces the electron trapping at the Li2S/LixIn heterointerface and inhibits the electron transfer to Li6PS5Cl electrolyte, thereby suppressing the interfacial reactions between Li metal and Li6PS5Cl. Furthermore, BIEF accelerates Li+ adsorption and diffusion at the Li2S/LixIn heterointerface, effectively suppressing the lithium dendrite growth. Benefiting from these merits, the symmetric cell with Li-In-S negative electrode enables a high critical current density (CCD) of up to 4 mA cm−2 and a high area capacity of 4 mAh cm−2 as well as a lifespan of more than 2000 h at 1 mA cm−2. Furthermore, the assembled ASSLBs demonstrate cycling performances with high capacity retention of 93% over 2000 cycles at 1 C for the Li-In-S|LiCoO2 (LCO) cell and 87.7% over 1000 cycles at 1 C for the Li-In-S|LiNi0.8Co0.1Mn0.1 (NCM811) cell, respectively. This work proposes a strategy to fundamentally address the critical interface issues in sulfide-based ASSLBs and should promote their development and applications. Results and discussion Fabrication and characterization of the Li-In-S foil A thin Li-In-S foil with Li2S/LixIn heterointerface and controlled thickness was fabricated by a facile spontaneous reaction between Li metal and In2S3 and mechanical kneading operations, as illustrated in Figs. S1–S3. Specifically, the In2S3 powder was first incorporated into molten Li to spontaneously generate the Li-In-S composite. The thin Li-In-S foil was then readily obtained by the mechanical rolling operations. The selection of In2S3 as the modified material for Li metal is deservedly deliberate. Specifically, incorporating In element into Li metal can decrease the lithium chemical potential and weaken the reduction activity, thereby suppressing the electrochemical decomposition of sulfide electrolytes15. Furthermore, the formed LixIn interphases can provide Li nucleation deposition sites for uniform Li+ deposition25 and accelerate the Li+ transfer at the electrode-electrolyte interface26. The choice of sulfur is due to its similar chemical properties to the LPSC electrolyte. Moreover, the formed lithiophilic Li2S phases with high Li+ adsorption energy and low electronic conductivity could modulate preferential Li+ deposition and prevent the electron transfer at the LPSC/negative electrode interface, thus avoiding the formation of lithium dendrite growth and unexpected interface reaction27,28. The distinct chemical properties of In and S elements induce the lattice order to lattice reconstruction when reacted with molten Li and promote the formation of immiscible solid phases (metallic LiₓIn and nonmetallic Li₂S), resulting in a naturally heterogeneous structure with abundant heterointerfaces (Fig. 1a). The fabrication of Li-In-S foil facilitates large-scale production for practical applications. As shown in Figs. 1b and S4, the controlled-thickness (50 μm) Li-In-S foil with a large size of 50 cm−2 can be readily prepared, which benefits from the strengthened machinability and flexibility, as verified by the winding test. We also screened the Li/In2S3 mass ratio to optimize the homogeneity of the Li-In-S composite, and the mass ratio of 5 wt% (marked as Li-In-S) is selected as the optimal material among other samples, as depicted in Figs. S5 and S6. Moreover, the Li-In-S composite negative electrode still exhibits a high specific capacity of 3194.8 mAh g−1 due to the limited amount of In2S3, which maintains the high energy density characteristic of the Li negative electrode (Fig. S7)19. The capacity of the Li-In-S composite Li negative electrode originates from the stripping of Li metal, consistent with the previous literature reports on composite Li negative electrode27. The analysis of the component characterization of the Li-In-S foil. a Schematic illustration of the reconstruction mechanism of the Li-In-S composite. b Digital photographs of the as-prepared Li-In-S foil. c XRD patterns of the Li-In-S foil. XPS spectra of d In 3d and e S 2p for the Li-In-S foil. f TOF-SIMS secondary ion 3D images of the Li-In-S foil. g Depth profile of LiInS−, LiS−, and Li3InO2− secondary ion fragments. h TEM and i HRTEM images of the Li-In-S foil and j the corresponding SAED image. To clearly analyze the chemical components of the Li-In-S foil, we first performed the X-ray diffraction (XRD) analysis (Fig. 1c). For the as-prepared Li-In-S foil, diffraction peaks well-indexed to Li2S and LixIn are clearly observed in addition to the two diffraction peaks of Li metal. In addition, some weak peaks in the range of 15°–25° are also displayed, which could be assigned to the LiInS2. The peaks of the raw In2S3 are not shown in the pattern, indicating that the In2S3 powder has been adequately consumed with molten Li. Moreover, X-ray photoelectron spectroscopy (XPS) further verifies the chemical component of the Li-In-S foil (Fig. 1d, e). In the In 3d XPS spectra, peaks at 449.9 and 442.7 eV could be assigned to LixIn29. The S 2p spectra also reveal XPS signals at 160.4 eV corresponding to Li2S30. Additionally, the In-S bond appears in both the In 3d and S 2p spectra, which may suggest the presence of LiInS2 phase in the Li-In-S foil. To further verify the component of Li-In-S foil, thermodynamic analysis based on first principles is conducted to reveal the reaction process between Li and In2S3. Based on the ternary phase diagram of the Li-In-S system (Fig. S3), all possible reaction equations with Li and In2S3 as reactants are listed in Table S1. These reactions are all spontaneous in terms of negative reaction energy, considering Li2S, LiInS2, and LixIn (abbreviated as LixInySz) as the reaction products. The reaction Li + In2S3 → Li2S + LiInS2 + LixIn is the most feasible one, depending on reaction energy changes. Both experimental and theoretical results indicate that the Li-In-S foil consists of a mixture of Li₂S and LiₓIn, along with a minor LiInS₂ phase. The top-view scanning electron microscopy (SEM) image and corresponding energy dispersive spectrometer (EDS) results of the as-synthesized Li-In-S foil demonstrate a homogeneous distribution of In and S elements without aggregation, indicating satisfactory compatibility of both In and S elements within molten Li (Fig. S8). The cross-sectional SEM images of Li-In-S foil with controllable thickness from 50 to 10 μm still display a complete morphology without structural failure (Fig. S9), suggesting the feasible processability of the Li-In-S foil. Furthermore, the morphology of Li and Li-In-S foils was measured by atomic force microscope (AFM) (Fig. S10). The roughness of Li-In-S foil increases slightly compared to that of pure Li foil due to the integration of In and S, in accordance with the SEM result observed above. To further identify the component depth distribution in the Li-In-S foil, the depth-profiling time-of-flight secondary-ion mass spectrometry (TOF-SIMS) was performed (Fig. 1f, g). The LiS−, Li3InO2−, and LiInS− secondary ion fragments represent the components of the Li-In-S foil. These signals exhibit a relatively uniform distribution even after 1200 s of sputtering, suggesting the as-prepared products are uniformly distributed rather than concentrated at the interface of the Li-In-S foil. The microstructure of the Li-In-S foil was further determined by high-resolution transmission electron microscopy (HRTEM). As depicted in Fig. 1h, tiny nanoparticles with an average radius of ~20 nm, in accordance with the top-view SEM observed surface morphology (Fig. S5). More importantly, the images in Fig. 1i present the lattice fringes with different interplanar distances of 0.331, 0.227, and 0.240 nm, well matching the Li2S (111), LiInS2 (112), and Li3InO3 (220) crystal planes, respectively, as further evidenced by the diffraction rings in the selected area electron diffraction (SAED) (Fig. 1j). These results directly demonstrate that the Li-In-S foil is composed of numerous LixInySz heterointerfaces. Electrochemical performance of the Li-In-S symmetric cells To evaluate the interfacial properties of the Li-In-S negative electrode toward the LPSC electrolyte, the symmetric Li-In-S|LPSC|Li-In-S cells were assembled and tested at stacking pressures of 10 MPa. As depicted in Fig. 2a, Tafel curves were first conducted to analyze the electrochemical kinetics at the Li electrode-electrolyte interface. Compared to the symmetric Li|LPSC|Li cell (0.409 mA cm−2), the cell with Li-In-S exhibits a higher exchange current density (0.752 mA cm−2). The increased exchange current density may be attributed to the formation of abundant LixInySz heterointerface, which indicates the faster charge transfer and improved electrochemical kinetics at the Li-In-S/LPSC interface31. Electrochemical performance of the Li and the Li-In-S symmetric cells at stacking pressures of 10 MPa. a Tafel curves of symmetric Li|LPSC|Li and Li-In-S|LPSC|Li-In-S cells. Critical current density (CCD) tests of b the Li symmetric cell and c the Li-In-S symmetric cell. Galvanostatic Li plating/stripping performance of the Li-In-S|LPSC|Li-In-S cell under d 0.2 mA cm−2 and e 0.5 mA cm−2. DRT transition result based on PEIS test under 0.1 mA cm−2 at different Li plating/stripping times of f Li|LPSC|Li and g Li-In-S|LPSC|Li-In-S symmetric cells at 25 °C. h The comparison of the Li-In-S symmetric cell performance with literature-reported results (Supplementary Table 2). Then, the CCD test was applied to evaluate the interfacial stability of the Li electrode-electrolyte32. As shown in Fig. 2b, the Li|LPSC|Li cell suffers from the short circuit under a current density of 0.5 mA cm−2 due to the severe interfacial reactions at the Li/LPSC interface, which is similar to the reported studies33. Note that the Li-In-S|LPSC|Li-In-S cell achieves a CCD value of 4 mA cm−2 (Fig. 2c), suggesting a better interfacial compatibility of LPSC electrolyte with Li-In-S negative electrode than pure Li. Although the LixInySz heterointerface can improve the interface stability between the negative electrode and LPSC electrolyte, the excessive amount (such as 10 wt.% and 15 wt.% of In2S3 for preparing) of the LixInySz heterointerface could also deteriorate the interfacial stability of the Li-In-S negative electrode with LPSC (Fig. S13) due to the limited Li+ transfer kinetics resulting from the excessive amount of Li2S, which possesses a high Li+ transfer barrier. Therefore, the Li-In-S negative electrode selected as the optimal amount of 5 wt.% enables satisfactory interfacial stability toward LPSC electrolyte under low stacking pressures of 10 MPa, competitive with these previously reported modified Li negative electrodes (Table S2). Furthermore, compared to the Li-In alloy, the Li-In-S negative electrode demonstrates significant advantages under the same test pressure (10 MPa) or higher pressure (150 MPa), and it also highlights the interface compatibility of the Li-In-S negative electrode with LPSC at low pressure (Fig. S14). Additionally, the long-term lithium plating/stripping tests of symmetric cells were further conducted to evaluate the electrochemical stability of the Li-In-S negative electrode against LPSC electrolyte. As shown in Fig. 2d, the Li-In-S|LPSC|Li-In-S cell maintains stability for over 2000 h at 0.2 mA cm−2 without a significant increase in overpotential. In sharp contrast, the symmetric cell with pure Li metal displays a significantly increased polarization and experiences a sudden decline in overpotential under the same current density, which implies the poor interfacial compatibility resulted from the severe interfacial side reactions and dendrite growth at the Li/LPSC interface (Figs. 2d and S15). Even under the higher current density of 1 mA cm−2 and areal capacity of 1 and 2 mAh cm−2, the symmetric cell with Li-In-S can still be cycled for more than 2000 h and 500 h (Figs. 2e and S16), further demonstrating the effectiveness of LixInySz heterointerface engineering in inhibiting interfacial reaction between LPSC and Li. Based on evaluating the long-term Li plating/stripping stability operating at high current density above, the Li-In-S|LPSC|Li-In-S cell rivals most previously reported sulfide-based symmetric cells (Figs. 2h and S17 and Table S2). To further reveal the relationship between interfacial impedance variations and interface evolution of the symmetric cell with Li and Li-In-S, in situ potentiostatic electrochemical impedance spectroscopy (PEIS) combined with the distribution of relaxation time (DRT) analysis was carried out at the current density of 0.1 mA cm−2 during continuous Li stripping. As illustrated in Fig. S18, theimpedance of the Li|LPSC|Li cell continuously decreases during Li stripping and then drops sharply after 10 h, indicating that the cell undergoes a short circuit. The irreversible change of interfacial impedance should be attributed to the lithium dendrite puncture and interfacial side reactions caused by continuous lithium deposition. Furthermore, the DRT technology was employed to determine the impedance evolution of each component in the Li stripping processes of the Li|LPSC|Li cell. DRT peaks at different relaxation time constants (τ) represent the characteristic response of each impedance component34. Specifically, the evolution of these peaks around 10−6 s and 10−4 s needs to be concerned because two peaks are assigned to the characteristics of the LPSC electrolyte and the solid electrolyte interphase (SEI), respectively. With the areal capacity of Li stripping in the Li|LPSC|Li cell increasing to 1 mAh cm−2, the intensity of these two peaks gradually disappears, which corresponds to the impedance evolution above. It can be explained that the interface components formed at the Li/LPSC interface cannot inhibit the puncture of lithium dendrites, resulting in the continuous degradation of the SEI and the LPSC electrolyte35. Different from the interface behavior of the Li symmetric cell with LPSC, the interfacial impedance and the corresponding SEI peaks of the cell with Li-In-S remain essentially unchanged, even reaching the areal capacity of 4.8 mAh cm−2 (Fig. 2f, g). The significantly enhanced areal capacity of Li deposition not only demonstrates interface compatibility of Li-In-S with LPSC electrolyte but also indicates the ability to accommodate deeper depths of discharge, which further highlights the advantages of the LixInySz heterointerface to effectively suppress interfacial side reactions of Li with LPSC electrolyte. Evolution analysis at the electrode/electrolyte interface To validate the effectiveness of the LixInySz heterointerface on the Li-In-S surface in suppressing the interface reaction, high-resolution XPS depth analysis of LPSC electrolytes from the Li|LPSC|Li and Li-In-S|LPSC|Li-In-S cells after Li plating/stripping was first performed at different etching times. In the S 2p XPS spectra of the Li/LPSC interface, a new peak located at 160.2 eV attributable to Li2S could be observed, obviously compared to that of the pristine LPSC electrolyte (Fig. 3a), which should result from the reduction decomposition of the LPSC electrolyte after contact with Li metal36. Synchronously, the existence of reduced P (130.2 and 131.2 eV) and Li3P (128.2 eV) in the P 2p XPS spectra of the Li/LPSC interface further indicates the poor interfacial compatibility of Li metal with LPSC electrolyte during the Li plating/stripping tests. The concentration of interfacial byproducts decreases gradually with increasing etching depth, suggesting the continuous decomposition reactions in the Li/LPSC interface. Comparatively, only a weak signal of Li2S in the S 2p XPS spectra for the Li-In-S/LPSC interface is observed after Li plating/stripping (Fig. 3b), which reveals the limited interfacial side reaction between the LPSC and Li-In-S surface. This is further supported by the P 2p XPS spectra of the Li-In-S/LPSC interface, where the reduced P and Li3P almost disappear. These results demonstrate that the Li-In-S foil is expectedly stable with LPSC electrolyte. In addition, the TOF-SIMS experiment was further conducted to quantitatively analyze the byproducts at the Li/LPSC interface and Li-In-S/LPSC interface before and after Li plating/stripping. Fig 3c exhibits the depth profiles of the proportion of various secondary ion species accounted for the total. The intensity of PS3− corresponding to the PS4 tetrahedron in the cycled Li/LPSC interface is reduced, which is lower than that of the cycled Li-In-S/LPSC interface. On the contrary, the characteristic signature of Li2S and reduced P (e.g., LiS− and LiP−) in the cycled Li/LPSC interface significantly increases and exceeds the variation of the cycled Li-In-S/LPSC interface. The distribution of reductive decomposition byproducts further confirms that the LixInySz heterointerface of Li-In-S can serve as the interfacial protection to mitigate interfacial reaction, in line with the above-analyzed XPS results. The Raman spectra measurements conducted in the different lithiation stages display the same evolution of Li/LPSC and Li-In-S/LPSC interfaces. For the Li/LPSC interface, in addition to the characteristic peaks of PS₄³⁻ in LPSC at 201, 270, 421, and 573 cm⁻¹, a new peak associated with Li₂Sₙ species containing -S–S- bonds was observed at 486 cm⁻¹ (Fig. 3d)37. This indicates the decomposition of PS43− tetrahedron of LPSC electrolyte along with the increase of lithiation areal capacity. In comparison, the peaks assigned to the decomposition products were not observed (Fig. 3e). suggesting that the stable LixInySz heterointerface suppresses the undesired reactions between LPSC and Li metal. Composition analysis and structure characterization at the negative electrode/electrolyte interface. High-resolution XPS spectra of the LPSC electrolyte interface from the symmetric cells after Li plating/stripping. The Li|LPSC|Li and Li-In-S|LPSC|Li-In-S symmetric cells were disassembled after Li plating/stripping at a current density of 0.2 mA cm−2/0.2 mAh cm−2 for 100 h under 25 °C, and the Li or Li-In-S electrodes were peeled off to expose the cycled LPSC electrolyte interface for XPS measurements. a S 2p and P 2p XPS spectra of the Li/LPSC interface and b S 2p and P 2p XPS spectra of the Li-In-S/LPSC interface after Li plating/stripping collected at different sputtering times. c TOF-SIMS depth profiling results of several representative fragments for the Li/LPSC and Li-In-S/LPSC interface before and after Li plating/stripping. The symmetric cells were disassembled under the same testing conditions mentioned above, and the Li or Li-In-S electrodes were peeled off to expose the electrode interface for TOF-SIMS measurements. Raman spectroscopy detection at the d Li/LPSC and e Li-In-S/LPSC interface in the Li symmetric cell after lithiation. The symmetric cells were disassembled under different lithiation states, and the specific capacity continued to increase by 0.2 mAh cm−2. The Li or Li-In-S electrodes were peeled off to expose the LPSC electrolyte interface for Raman measurements. Top-view SEM images of f the Li and g the Li-In-S electrodes exfoliated from the symmetric cells under the same testing conditions mentioned above. AFM images and corresponding Young’s modulus of h the Li and i the Li-In-S foil. Furthermore, the morphological evolution of the cycled Li was visually observed by SEM (Fig. 3f), where numerous agglomerates containing the decomposition byproducts and lithium dendrites were present on the surface of the cycled pure Li metal38. In contrast, the cycled Li-In-S exhibits a relatively flat and smooth surface without obvious agglomerates (Fig. 3g), demonstrating the efficacy of the heterointerface in suppressing lithium dendrite growth. In addition, the Young’s modulus of Li and Li-In-S foils was measured by AFM (Fig. 3h, i). The Young’s modulus of Li-In-S foil demonstrates a significant enhancement, averaging 7.2 GPa compared to pure Li at 2.4 GPa. It is well accepted that Young’s modulus of the interfacial materials over 6 GPa can effectively inhibit lithium dendrite growth39. These mechanical analyses indicate the improved strength of the Li-In-S foil with LixInySz heterointerface, which further confirms the potential of modified Li metal for stable electrode-electrolyte interfaces. Mechanism investigation at the electrode/electrolyte interface To comprehensively reveal the interfacial stabilization mechanism of Li-In-S with LPSC electrolyte, ab initio molecular dynamics (AIMD) simulations and first-principle calculations were employed to investigate the interface evolution at Li/LPSC and Li-In-S/LPSC interfaces (Supplementary Data 1 and 2). The construction of this structural model is derived from the above experimental characterization tests, and subsequent reasoning (Fig. 1i). For the Li/LPSC interfaces, severe decomposition of the PS43− tetrahedron is observed when the LPSC contacts with Li metal after AIMD, which is also similar to the first principle calculations (Figs. 4a, S19 and S24). The essence of the decomposition reaction is that Li metal easily transfers electrons to PS43− tetrahedron of the LPSC electrolyte, which induces the decomposition of the tetrahedron and forms the interface byproducts (Li2S and P2S5) (Fig. 4g). This decomposition mechanism indicates the interfacial incompatibility of LPSC with Li metal, which is evidenced by the structural variations observed in the Li/LPSC interface models based on the radial distribution function (RDF) profile (Figs. 4c and S23)40. Here, RDF describes the Li–S bond information of the specific area of the negative electrode/electrolyte contact interface and implies the formation of parasite LiSx phase at the interface. As shown in Fig. 4c, the strength of the Li–S bonds is relatively low at the initial contact of the Li/LPSC interface, which indicates that the interaction between Li metal and the S atoms of the LPSC electrolyte is relatively weak. As the contact time increases, the strength of the Li–S bonds gradually increases, and the length of the Li–S bonds also fluctuates. This indicates that the Li metal has completely combined with the S atoms of the LPSC electrolyte to form a large number of Li–S bonds, implying the continuous interface side reaction between Li metal and LPSC electrolyte. On the contrary, the RDF of the Li–S bonds at the initial state between Li-In-S (Li2S/LixIn) and LPSC is mainly attributed to the Li2S phase in the Li-In-S. The intensity of Li–S bonds remains almost unchanged during the simulation, indicating the relatively stable interfacial structure between Li-In-S and LPSC electrolyte (Figs. 4d and S22). This also implies that the interfacial reaction between Li-In-S and LPSC electrolyte is effectively inhibited. Mechanism investigation at the electrode/electrolyte interface. Snapshots of a Li/LPSC interface and b Li-In-S/LPSC interface after AIMD simulations. RDF evolutions of c Li–S interaction for Li/LPSC interface and d Li–S for Li-In-S/LPSC interface during the simulation. Electrostatic potential plots and the corresponding work functions of e LixIn and f Li2S models. Schematic diagram for the charge transfer at g Li/LPSC interface and h Li-In-S/LPSC interface. i The local electronic structure of the LixIn/Li2S interface. KPFM images and contact potential differences of j LixIn, k Li2S, and l Li-In-S. m Adsorption energies of Li2S, LixIn, and LixIn/Li2S for Li+. Interestingly, the Li2S/LPSC or LixIn/LPSC interface structure optimization modeling also displays the decomposition of PS4 tetrahedrons of LPSC (Figs. S20 and S21). These findings suggest that the suppressed interfacial reaction between Li-In-S and LPSC electrolyte is primarily attributed to the presence of LixInySz heterointerfaces. At present, previous studies have demonstrated that constructing composite Li with multiple components can effectively improve the interfacial stability with sulfide electrolyte. However, there is controversy on the role of composite Li with multiple components in interfacial reactions with sulfide electrolyte, and the contribution of the heterointerfaces in the composite negative electrode has been overlooked. To this end, we conducted further electronic structure analysis of LixInySz heterointerfaces in Li-In-S to reveal the stability mechanism of the Li-In-S/LPSC interface, since interfacial parasite reaction originates from the strong tendency of Li metal to donate electrons to the PS₄³⁻ tetrahedra in the LPSC electrolyte40. According to the band theory of solids, the charge transfer between different materials is directly affected by the work function (Wf)41. The work function of Li2S and LixIn in the LixInySz heterointerfaces of the Li-In-S composite negative electrode was therefore calculated to analyze the charge transfer behavior at the Li-In-S/LPSC interface. The work function is calculated by subtracting the Fermi level energy from the vacuum level energy. The Fermi level energy was obtained through density functional theory (DFT) calculations (Supplementary Data 2), and the vacuum energy level is usually taken as the average value or the maximum value of the stable region of the electrostatic potential in the vacuum layer. Here, the work function of Li2S and LixIn is calculated to be 3.78 and 3.48 eV, respectively (Fig. 4e, f). The work function difference could drive electron transfer from LixIn to Li2S, thus forming BIEF pointing toward Li2S at the LixIn/Li2S heterointerface (Fig. 4h). The BIEF traps the free electrons of the negative electrode at the LixIn/Li₂S heterointerface, which can be verified by the electron localization function (ELF) analysis (Figs. 4i and S25), thus restricting their subsequent migration to the LPSC electrolyte and suppressing the interfacial reaction between the Li-In-S and LPSC. To further confirm the existence of the BIEF, we have performed Kelvin probe force microscopy (KPFM) measurements to experimentally probe the contact potential difference (CPD), effective work function (Wf), and surface potential variation (ΔE) of LixIn, Li₂S, and Li-In-S. Based on the previous studies in the literature42,43, there is a positive correlation between the ΔE and the strength of the BIEF, and the higher ΔE values generally represent stronger the BIEFs properties. As shown in Fig. 4j–l, the LixIn and Li₂S exhibit markedly different CPD of approximately 1.92 V and 1.35 V, and different Wf of approximately 3.18 and 3.75 eV, respectively, corresponding to a substantial difference in their effective surface work functions. For Li-In-S, an intermediate CPD value of approximately 1.83 V and an intermediate Wf of approximately 3.27 eV is observed. This intermediate work function strongly suggests charge redistribution and Fermi-level equilibration at the LixIn/Li₂S heterointerface, which is a prerequisite for the formation of the BIEF. In addition to the shift in the average CPD, the Li-In-S shows a significantly enlarged surface potential variation (ΔE ≈ 405 mV, defined as max–min) compared with the LixIn and Li₂S, which further indicates the formation of the BIEFs at the Li2S/LixIn heterointerface. To sum up, the pronounced work function difference between LixIn and Li₂S provides the driving force for interfacial charge redistribution, leading to the formation of the BIEF at the Li2S/LixIn heterointerface, as evidenced by the intermediate Wf and the enhanced ΔE revealed by KPFM. In addition, the Li+ adsorption and diffusion enhancement of the BIEF is also confirmed in the calculated results and illustrated by the mechanism diagram in Figs. 4m and S26. This mechanism, which the BIEF can enhance the ions diffusion, has been widely discussed in previously published literature44. Compared to the Li2S (−2.04 eV) and LixIn (−2.17 eV), the adsorption energy of Li2S/LixIn heterointerface for Li+ is increased to −3.63 eV. With the BIEF, the S²⁻ anions in the Li₂S lattice have a strong Li+ electrostatic attraction and provide a strong Li+ adsorption site. Additionally, the work function difference could drive electron transfer from LixIn to Li2S, thus resulting in the accumulation of electrons around Li2S. Therefore, Li⁺ is more likely to preferentially adsorb near the Li₂S. With the Li⁺ adsorption, the positively charged Li⁺ accumulate, resulting in a change in the direction of the BIEF from Li2S to LixIn45. Under the drive of the BIEF, Li⁺ rapidly diffuses into the LixIn phase. Meanwhile, the concentration of Li⁺ near the interface of Li2S increases, and the high Li⁺ concentration gradient enhances the Li⁺ diffusion driving force46. Furthermore, the electric field undergoes dynamic evolution accompanied by charge neutralization after the Li+ adsorption, which promotes Li+ diffusion from the Li2S interface to the LixIn interface. These theoretical calculations above demonstrate that establishing BIEFs formed at the Li2S/LixIn heterointerface can manipulate the driving force for charge transfer, which not only effectively suppresses the interfacial reaction with LPSC but also promotes the uniform Li+ deposition and prevents the lithium dendrite growth. Performance of ASSLBs using the Li-In-S To investigate the influence of Li-In-S negative electrode containing LixInySz heterointerface on the electrochemical properties, the ASSLBs comprising the Li or Li-In-S negative electrode, LCO and NCM811 positive electrode, and LPSC electrolyte were assembled and tested at stacking pressures of 10 MPa. The long-term cycling performance of the all-solid-state batteries with LCO positive electrode was firstly evaluated. As shown in Fig. 5a, b, the cell with Li-In-S delivers an initial discharge capacity of 130 mAh g−1 at 0.1 C (1 C = 140 mA g−1) coupled with a high initial coulombic efficiency of 92.7% and demonstrates long-term cycling stability with a high capacity retention of 93% even over 2000 cycles. The cycling performance of the cell with Li-In-S should benefit from the suppression of the interfacial side reaction by the LixInySz heterointerface that enables the stable electrochemical behavior. By comparison, the cell with pure Li shows a significant decline in discharge capacity and poor cycle stability with low discharge capacity, which is caused by the serious interfacial reaction between LPSC and Li metal47. Furthermore, compared to the batteries assembled with Li-In alloy, the battery with Li-In-S demonstrates better cycle stability at both low (10 MPa) and high (150 MPa) test pressures (Fig. S28). This further highlights the advantages and potential of the Li-In-S under low test pressure conditions. The advantages of the LixInySz heterointerface were further verified by the rate performance of the cell. Compared with the Li|LPSC|LCO cell, the cell with Li-In-S achieves notable rate performance up to 3 C and delivers a high reversible capacity of 121.5 mAh g−1 when returning to 0.1 C (Fig. S27). Electrochemical performance of the all-solid-state cell with LCO and NCM811 positive electrodes at stacking pressures of 10 MPa. a Long-term cycling performances of the Li|LPSC|LCO and Li-In-S|LPSC|LCO cells at 1 C and b the corresponding charge−discharge curves of the Li-In-S|LPSC|LCO cell. c Rate performances of the Li|LPSC|NCM811 and Li-In-S|LPSC|NCM811 cells and d corresponding charge-discharge curves of the Li-In-S|LPSC|NCM811 cells at various rates. e Long-term cycling performances of the Li|LPSC|NCM811 and Li-In-S|LPSC|NCM811 cells at 1 C. f Cycling performance of the Li-In-S|LPSC|NCM811 cell under high positive electrode loading at 0.1 C. g The voltage profiles and h cycling performance of the Li-In-S|LPSC|NCM811 pouch cells. To further enhance the energy density of the battery, the cells using NCM811 as the positive electrode were assembled to evaluate the electrochemical performance of the Li-In-S negative electrode. Firstly, the rate capability of the cell with Li-In-S, as shown in Fig. 5c, is better than that of the cell with Li at various specific currents. Specifically, the cell with Li-In-S exhibits high discharge capacity of 193.7, 172.6, 152.6, 128.9, 110.2, 94.8, 81.5, and 67.1 mAh g−1 at 0.1, 0.2, 0.5, 1, 1.5, 2, 3, and 4 C, respectively, and the capacity can recuperate as the specific current returns to 0.1 C (1 C = 200 mA g−1) (Fig. 5c). In contrast, the discharge capacity of the cell with Li significantly decreases as the specific current increases. Furthermore, the charge-discharge curves of the cell with Li-In-S present a low polarization voltage at various current densities (Fig. 5d), suggesting enhanced electrochemical kinetics. This also further demonstrates that the interfacial stability of the electrolyte with the negative electrode can be effectively improved by the construction of the LixInySz heterointerface. Benefiting from the interfacial stability, the electrochemical performance of the cell with Li-In-S negative electrode demonstrates long cycling life (Figs. 5e and S29). The cell delivers a high reversible capacity of ~106.8 mAh g−1 and a high capacity retention of 87.7% over 1000 cycles at 1 C, suggesting high compatibility of LPSC with Li-In-S negative electrode, which is competitive with most previously reported sulfide-based lithium batteries in terms of cycling stability and rate performance (Table S3). Furthermore, the cell with a high positive electrode loading of 21 mg cm−2 can also exhibit a reversible capacity of 173.2 mAh g−1 and stable cycling performance for over 100 cycles (Fig. 5f). The charge and discharge capacities of the battery fluctuate simultaneously around 40 to 60 cycles, possibly due to slight temperature variations during electrochemical testing. Considering the practical application, the electrochemical performance of the all-solid-state pouch battery was also evaluated (Fig. 5g). Pairing the NCM811 composite electrode membrane with the Li-In-S negative electrode, the all-solid-state pouch battery delivers a reversible specific capacity of 152.4 mAh g−1 at 0.1 C with a high loading of 5.1 mAh cm−2 (Fig. 5h). The capacity retention is 93.4% after 20 cycles with a high capacity of 142.4 mAh g−1. Moreover, the high reversibility of the all-solid-state pouch battery can be further observed in Fig. 5h, demonstrating the interface stability of the Li-In-S negative electrode in the practical battery. In summary, we successfully construct BIEFs via designing a Li-In-S composite negative electrode to suppress the solid-solid interfacial reactions at the Li/sulfide electrolyte interface. The BIEF driven by the work function difference of Li2S and LixIn in Li-In-S triggers electrons trapping at the Li2S/LixIn heterointerface, which restricts the electrons transfer to LPSC electrolyte and thereby effectively suppresses interfacial reactions between LPSC electrolyte and Li metal. Additionally, the electron trapping effect at the Li2S/LixIn heterointerface significantly promotes the Li+ adsorption and diffusion, suppressing the lithium dendrite growth. Due to the above features, the symmetrical cell with Li-In-S negative electrode delivers a notable CCD of 4 mA cm−2 and a high areal capacity of 4 mAh cm−2, and a stable Li plating/stripping over 2000 h at 1 mA cm−2. The assembled full ASSLBs using Li-In-S negative electrode and LCO or NCM811 positive electrode demonstrate stable cycling performance (Li-In-S|LPSC|LCO: 93% over 2000 cycles at 1 C; Li-In-S|LPSC|NCM811: 87.7% over 1000 cycles at 1 C). Moreover, the Li-In-S|LPSC|NCM811 full cell achieves a rate capability of up to 4 C. The pouch cell with a high areal capacity of 5.1 mAh cm−2 also shows great application potential. This study highlights the crucial role of BIEFs in stabilizing the solid-solid interface and paves an alternative direction for designing practical metal negative electrodes for high-performance solid-state lithium batteries. Methods Preparation of LPSC electrolyte The Li6PS5Cl electrolyte (LPSC) was synthesized by the following processes. Firstly, the precursor powders with stoichiometric amounts of Li2S (Alfa Aesar, 99.9%), P2S5 (Aladdin, 99.9%), and LiCl (Aladdin, 99.9%) were placed into a ZrO2 ball-mill container (80 ml) with ZrO2 balls (d = 5 mm). The ratio of balls to materials is 40:1. The mechanical mixing process was carried out using a planetary ball-milling machine (Fritsch, Pulverisette 7) at 500 rpm for 16 h, with a milling/rest cycle of 10 min milling followed by 5 min rest. Then, the as-milled powders were pelletized at 350 MPa and annealed at 550 °C for 10 h with a heating rate of 5 °C min−1, followed by natural cooling to room temperature. The LPSC pellets were hand-ground in an agate mortar for use. All procedures were performed in an argon-filled glovebox. Fabrication of Li-In-S foil For the preparation of Li-In-S foil, the Li metal (China Energy Lithium Co., Ltd., 99.9%) was first placed in a stainless-steel crucible, followed by heating to ~300 °C to form molten Li. The In2S3 powder (Aladdin, 99.99%) with different amounts (5 wt.%, 10 wt.%, and 15 wt.%) was added into the molten Li and continuously stirred to form the homogeneous mixture. The Li-In-S foil with 5 wt% In2S3 was selected as the optimized sample. The Li-In-S composite material formed after cooling was scraped out and placed in a plastic film. The Li-In-S composite material wrapped in plastic film was placed on the roller press and repeatedly compressed to the desired thickness. The Li-In-S foil with a thickness of 50 μm was used for the electrochemical test. The fabrication processes above were conducted in an argon-filled glovebox. Characterizations The morphology of sulfide electrolyte, Li-In-S foil, and lithium negative electrode after Li plating/stripping was collected by the scanning electron microscope (SEM) and energy dispersive spectroscopy (EDS) by Hitachi SU8010. SEM images were collected at an accelerating voltage of 5 kV and a current of 10 μA, while EDS mapping was performed at an accelerating voltage of 15 kV and a current of 10 μA. Transmission electron microscopy (TEM) and SAED were applied to characterize the morphology of Li-In-S foil. XRD was employed to determine the phase composition of Li-In-S foil by the Smartlab with a Cu Kα radiation source, a step size of 0.01°, and a scan rate of 5° min−1. The AFM was carried out using the Bruker Dimension Icon in tapping mode to analyze the morphology and modulus of the samples. The scan size was set to 1 μm × 1 μm. TOF-SIMS was conducted on the ION-TOF (GmbH) to observe the chemical components on the Li-In-S foil. The sputtering with a Cs+ ion beam of 1 KV was applied for depth profiling analysis of Li-In-S foil on the selected analysis area of 200 × 200 μm. The TOF-SIMS spectra were normalized based on the total ion signal. XPS was conducted on a Thermo Scientific ESCALAB 250Xi equipped with a monochromated Al Kα X-ray source with a photon energy of 1486.6 eV and an X-ray power of 225 W to elucidate the chemical composition of the samples. The binding energy scale of the XPS spectra was calibrated using the hydrocarbon C 1s peak at 284.8 eV as the reference. The surface potentials of materials were measured by an Asylum Research MFP-3D Infinity AFM (Oxford Instruments) with a Kelvin probe. The scan size and scan rate were set to 5 μm × 5 μm and 2 Hz, respectively. To investigate the chemical composition and morphology of the electrode/electrolyte interface after Li plating/stripping, the cycled cells were disassembled, and the negative electrode was peeled off from the electrolyte to expose the interface after Li plating/stripping. The negative electrode or electrolyte samples were sealed in a vacuum transfer holder inside the glovebox and transferred for characterization to avoid air exposure. Electrochemical measurements The ionic conductivity of the LPSC electrolyte was measured by the electrochemical impedance spectroscopy (EIS) using the electrochemical workstation (PARSTAT MC) with the frequency ranging from 0.1 Hz to 1.0 MHz. The impedance spectra were collected with 71 data points. To assemble the blocking cell, 100 mg of LPSC electrolyte was first cold-pressed into a pellet under a pressure of 2 tons. Then, indium foils as current collectors served as blocking electrodes for EIS measurements. The Tafel curves of the Li|Li symmetric cells were recorded on the electrochemical workstation at a scan rate of 0.1 mV s−1 over a voltage range from −0.1 to 0.1 V. The operando PEIS measurements were performed under a constant current density of 0.1 mA cm−1. After each 1 h galvanostatic polarization, impedance spectra were recorded with a voltage amplitude of 10 mV over the frequency range from 1 MHz to 0.1 Hz. The DRT analyses corresponding to the PEIS spectra were performed using the MATLAB GUI toolbox developed by Prof. Francesco Ciucci48. To investigate the voltage profile and gravimetric capacity of Li-In-S composite negative electrode, the assembled Li-In-S/LPSC/Cu cells were subjected to the critical delithiation test at 0.5 mA cm−2 with a cut-off voltage of −0.4 V versus Li+/Li. The symmetric cell assembly For the fabrication of the symmetric cell, 100 mg of LPSC electrolyte was placed into a polyetheretherketone (PEEK) mold with a diameter of 10 mm and then cold-pressed into a pellet at 2 tons (LPSC pellet thickness, ~800 μm). Subsequently, two pieces of Li foil or Li-In-S foil were pasted on both sides of the pellet, and the copper foils as the collectors were attached to the Li foils or Li-In-S foils. Finally, the assembled symmetric cell was tested at a constant pressure of 10 MPa in a thermostat maintained at 25 °C. The variations of current densities for the CCD test in Fig. 2c are 0.1, 0.2, 0.4, 0.6, 0.8, 1, 1.2, 1.4, 1.6, 1.8, 2, 2.4, 3, 3.6, 4, 5, 6, 7, and 8 mA cm−2, respectively. The variations of current densities for the CCD test in Fig. 2b are 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, and 0.8 mA cm−2, respectively. The black lines in Fig. 2b, c represent the current density. All-solid-state battery assembly Typically, the assembly process of the ASSLBs is as follows: positive electrode (NCM811 or LCO, Canrd Technology Co., Ltd.) and LPSC in a mass ratio of 7:3 were first homogeneously mixed to obtain the composite electrode. For the assembly of Li-In-S|LPSC|NCM811 or Li-In-S|LPSC|LCO battery, 100 mg of LPSC electrolyte was pressed into a 10-mm-diameter pellet with a thickness of ~800 μm under a pressure of 240 MPa for 1 min using a mold cell consisting of stainless-steel columns and a PEEK holder. Then, the 10 mg of the composite electrode (mass loading: ~8.9 mg cm−2) was uniformly placed on the side of the LPSC pellet and pressed at 360 MPa for 1 min. Subsequently, a Li foil or Li-In-S foil (50 μm) was attached to the negative electrode side. Finally, the assembled cell was subjected to an external pressure of ~10 MPa during the process of the performance test. The external pressure (10 MPa) of the battery is controlled by the torque wrench. The performance measurements of the battery were evaluated using the Neware battery test system (CT-4008T-5 V 10 mA-164, Shenzhen, China) in a thermostat maintained at 25 °C. Before long-term cycling, the cells were activated by three charge-discharge cycles at 0.1 C, followed by testing at the targeted rate. The cut-off voltage ranges for the cells using LCO and NCM811 positive electrodes were set to 2.5–4.2 and 2.5–4.3 V, respectively. The pouch cell was assembled using three stacked layers comprising NCM811 composite electrode, LPSC, and Li-In-S foil, respectively. The NCM811 composite electrode film and LPSC separator film were prepared by a dry electrode process with 0.5 wt.% PTFE. Aluminum and nickel tabs were used for the positive electrode and negative electrode, respectively. The electrode, electrolyte and negative electrode films were stacked in an argon-filled glovebox to fabricate pouch cells, which was then vacuum-sealed in an aluminum bag and tested under a stack pressure of approximately 10 MPa. Theoretical calculation methods The interface structure evolution and work function were carried out with the Vienna Ab initio Simulation Package using periodic DFT49. The generalized gradient approximation with the Perdew–Burke–Ernzerhof functional was employed to describe the exchange–correlation interactions50. The convergence criteria for energy and forces were set to 1 × 10−4 eV and 0.01 eV Å−1, respectively. A k-point grid with a size of 3 × 3 × 3 was used to sample the Brillouin zone for optimizing the lattice parameters. These surface models, such as the Li/LPSC interface and Li-In-S/LPSC interface, were built and optimized to investigate the interfacial stability and interaction. AIMD simulations were conducted under the NVT ensemble employing a Nosé thermostat with a time step of 4 fs. RDF analysis was performed using the Open Visualization Tool (OVITO) for structural evolution. Supplementary information Supplementary Information Description of Additional Supplementary Files Supplementary Data 1 Supplementary Data 2 Transparent Peer Review file Source data Source data Author contributions The manuscript was written through the contributions of all authors. Xiangfeng Liu and Tianran Zhang provided supervision, funding acquisition, and manuscript validation. Chong Liu provided the lead experiment design, investigation, material synthesis, and characterizations. Ruoyu Wang performed the computational calculations. Dewen Wang, Liyun Zhao, Taiguang Li, Tenghui Wang, and Butian Chen conducted characterizations, data acquisition, and validation of this work. Peer review Peer review information Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. A peer review file is available. Funding X.L. discloses support for the research of this work from the National Natural Science Foundation of China (Grant No. 12375301, 12535018, 11975238, and 22005302), the Scientific Instrument Developing Project (Grant No. ZDKYYQ20180001), and the Strategic Priority Research Program (Grant No. GJ0090406) of the Chinese Academy of Sciences. C.L. discloses support for the research of this work from the China Postdoctoral Science Foundation (Grant Number 2025M770093). Data availability All the data generated in this study are provided in the Source data file. Source data are provided with this paper. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Tianran Zhang, Email: zhangtianran@ucas.ac.cn. Xiangfeng Liu, Email: liuxf@ucas.ac.cn. Supplementary information The online version contains supplementary material available at 10.1038/s41467-026-74049-1. References Associated Data Supplementary Materials Supplementary Information Description of Additional Supplementary Files Supplementary Data 1 Supplementary Data 2 Transparent Peer Review file Source data Data Availability Statement All the data generated in this study are provided in the Source data file. Source data are provided with this paper.