Strong Lewis-acid coordinated PEO electrolyte achieves 4.8 V-class all-solid-state batteries over 580 Wh kg⁻¹

All-solid-state Li metal batteries using polymer electrolytes promise high energy density and improved safety, with PEO-based electrolytes favored for processability and interfacial contact. However, PEO suffers from poor oxidation stability at high voltages, especially against Ni-rich layered cathodes, limiting operation above ~4.3–4.5 V. The strong interfacial reactivity of ether oxygen chains with high-valence Ni at high states of charge leads to electrolyte decomposition and cathode surface degradation. The study aims to diagnose failure mechanisms in the 4.5–4.8 V range and to design a PEO-based electrolyte with enhanced oxidative stability and reduced interfacial interaction to enable 4.8 V-class solid-state batteries without sacrificing ionic conductivity.

Multiple strategies have been pursued to extend PEO-based electrolyte voltage windows: inorganic fillers, cathode surface coatings, and molecular grafting, enabling up to ~4.5 V-class operation. High-concentration salt electrolytes stabilize coordination structures but usually suffer from ultralow ionic conductivity, hindering Li⁺ transport. Metal-ion coordinated polymers have been proposed to enhance Li⁺ conduction by chelating polymer chains and widening ion channels, yet compatibility with high-voltage cathodes remains challenging. These prior approaches highlight the need to tune electrolyte coordination environments to decouple oxidative stability from ionic transport for ultra-high voltage (e.g., 4.8 V) operation.

Electrolyte design and preparation: A Lewis-acid coordinated PEO electrolyte was formulated by introducing Mg²⁺ and Al³⁺ as electron-withdrawing ligands to chelate EO chains and weaken their interaction with high-valence Ni. The optimized composition used EO:Li:Mg:Al = 18:1:0.25:0.25 with PEO (Mw = 2,000,000) and LiTFSI, using Mg(ClO₄)₂ and Al(ClO₄)₃ as Lewis-acid sources. Al(ClO₄)₃·9H₂O was dehydrated in Ar at 150 °C for 15 h (with Al debris present) to obtain anhydrous Al(ClO₄)₃; anhydrous Mg(ClO₄)₂ was also used. Electrolyte membranes (PEO-LiTFSI and PEO-Mg-Al-LiTFSI) were prepared by solution casting in anhydrous acetonitrile (ACN), drying under vacuum at 60 °C for 48 h, and hot pressing at 80 °C. Karl–Fisher titration confirmed low water content (~165 ppm) after drying. Thermal stability was assessed by TGA. Cathode preparation: Composite cathodes used either LiNi0.83Co0.12Mn0.05O2 (Ni83) or Li-rich Li1.14Ni0.136Co0.136Mn0.542O2, mixed with PEO-Mg-Al-LiTFSI electrolyte solution and carbon nanotubes (mass ratio 85:12:3) in NMP. Slurries were coated onto Al foils, dried at 65 °C (24 h, vacuum), then pressed at 1 MPa cm⁻² at 80 °C to densify. Cathode loadings exceeded 15 mg cm⁻²; pouch cells used 50 mg cm⁻² (double-sided coating). Cell assembly: CR2025 coin cells were assembled in Ar glovebox with Li metal anode and a double-layer solid electrolyte: a thin PEO-LiTFSI anode-side membrane (≤15 µm) and PEO-Mg-Al-LiTFSI cathode-side membrane, which also served as the ion conductor within the cathode. Pouch cells used a Li–Cu composite anode (copper mesh current collector), total anode thickness <20 µm, total electrolyte membrane thickness <40 µm, and a stack of four cathodes and five anodes; bare cell mass ~3.93 g, packaged mass ~4.31 g. Pouch cells were tested under 1 MPa pressure. Electrochemical testing: Galvanostatic tests were conducted at 60 °C on a Neware BTS system. Voltage windows: 2.8–4.8 V for Ni83 (1 C = 200 mA g⁻¹); 2.0–4.8 V for Li-rich cathodes. Linear sweep voltammetry (LSV) was performed on SS|electrolyte|Li cells at 60 °C, 1 mV s⁻¹. Ionic conductivity was determined via EIS on SS|electrolyte|SS cells using σ = L/(R S). Li⁺ transference number t⁺ was measured by DC polarization (10 mV) in Li|electrolyte|Li cells with pre/post EIS using t⁺ = (ΔV − I0R0)/(10ΔV − I1R1). Diffusion coefficients were estimated from MD simulations; experimental EIS tracked interfacial resistance evolution. Characterization: Operando synchrotron TXM and XAS at SSRF probed Ni-rich cathode particle chemistry and SOC distribution from 3.5–4.8 V; synchrotron nano-tomography assessed interphase composition and porosity after cycling. XPS analyzed surface chemistries (F 1s indicating LiF, MgF2, AlF3). Raman spectroscopy probed ion pairing/dissociation (TFSI-related bands). ICP-OES measured Ni concentration along charge. SEM imaged morphology and film uniformity. Computation: DFT (VASP, PBE-GGA, PAW, 500 eV cutoff, DFT-D dispersion) evaluated adsorption/reaction energies on LiNiO2 (104) slabs and HOMO/LUMO characteristics (Gaussian09, B3LYP/6-311++G(d,p)). MD simulations (OPLS-AA parameters; machine-learning force field) modeled coordination structures and radial distribution functions (RDFs), coordination numbers (CN), Li–O binding energies, and ion diffusion. ESP maps visualized charge distributions within solvation structures. Scale-up: Industrial continuous slurry-casting produced 50 m × 0.3 m rolls of PEO-Mg-Al-LiTFSI membranes. A 20 Ah all-solid-state polymer battery was assembled using the industrially produced membranes and evaluated for capacity, coulombic efficiency, and cycling.

• Failure mechanism in conventional PEO-LiTFSI at high voltage: Operando TXM/XAS from 4.3–4.8 V revealed formation of a Ni-deficient (“Ni-poor”) surface layer on Ni-rich cathodes uniquely in PEO cells, accompanied by reduced surface Ni valence, attributed to strong EO–cathode interactions. Calculated adsorption energy of PEO on LiNiO2 was strong (−0.76 eV) compared with EC (−0.42 eV), promoting Ni migration into electrolyte and interfacial degradation.
• Lewis-acid coordination design: Introducing Mg²⁺ and Al³⁺ chelates EO chains, lowering EO electron density and weakening EO–Ni interaction. MD shows anions (TFSI, ClO4−) enter inner solvation sheath, increasing inorganic-derived CEI formation propensity. RDF/CN analysis indicates stronger Mg/Al–O coordination than Li–O, weakening Li–EO and Li–TFSI binding and facilitating Li⁺ transport.
• Enhanced oxidative stability: ESP/HOMO analyses show the Lewis-acid coordinated structure with a low HOMO energy (−9.08 eV) and preferential anion decomposition sites. LSV: PEO-LiTFSI shows decomposition current rise >4.2 V vs Li⁺/Li; PEO-Mg-Al-LiTFSI remains stable up to ~5.0 V.
• Improved transport properties: Binding energy for Li dissociation from coordination decreases from 0.75 eV (PEO-LiTFSI) to 0.58 eV (PEO-Mg-Al-LiTFSI). Ionic conductivity of the dried PEO-Mg-Al-LiTFSI reaches 0.23 mS cm⁻¹ at room temperature (superior to PEO-LiTFSI below 60 °C). MD-derived Li⁺ diffusion coefficient increases from 1.21 × 10⁻7 cm² s⁻1 (PEO-LiTFSI) to 5.30 × 10⁻6 cm² s⁻1 (PEO-Mg-Al-LiTFSI). Li⁺ transference number t⁺ ≈ 0.67; Mg²⁺ and Al³⁺ are effectively immobile (1.62 × 10⁻10 and 7.05 × 10⁻10 cm² s⁻1). Li⁺ diffusion is 2.24× the sum of TFSI and ClO4− diffusion.
• Interfacial chemistry and morphology: XPS on cycled cathodes indicates inorganic-rich CEI (LiF, MgF2, AlF3) in PEO-Mg-Al-LiTFSI cells versus more organic-rich interphase for PEO-LiTFSI. Synchrotron tomography shows stronger X-ray absorption (inorganic) at the interface and lower post-cycle particle porosity for PEO-Mg-Al-LiTFSI, signifying fewer microcracks and better structural integrity.
• Cell performance at high voltage: Full cells with PEO-Mg-Al-LiTFSI (Ni83 cathode, 2.8–4.8 V) cycle for 300 cycles; with high loading (15.1 mg cm⁻2), capacity retention is 80.8% after 100 cycles. PEO-LiTFSI cells cannot operate at 4.8 V. EIS shows low interfacial resistance growth for PEO-Mg-Al-LiTFSI (stable Rinter after 100 cycles) versus rapid Rinter increase in PEO-LiTFSI (after 3 cycles).
• Pouch cell metrics: PEO-Mg-Al-LiTFSI pouch cells deliver 586 Wh kg⁻1 (bare cell 645 Wh kg⁻1) at 0.1 C with 80.6% retention after 50 cycles and 63.5% after 100 cycles at 60 °C under 1 MPa. With Li-rich cathodes, retention is 74.7% after 100 cycles (versus 61.8% for liquid electrolyte); PEO-LiTFSI shows only 28.1% retention after 100 cycles even at 2.8–4.2 V.
• Manufacturability: Industrial-scale continuous casting produced 50 m × 0.3 m electrolyte rolls; a 20 Ah all-solid-state polymer battery built from these membranes showed high capacity, high coulombic efficiency, and good cycling performance.

The study identifies strong EO–Ni interactions as a principal cause of interfacial degradation in high-voltage PEO-based batteries, leading to Ni dissolution at the cathode surface and chemically driven failure beyond 4.5 V. By coordinating EO chains with strong Lewis acids (Mg²⁺, Al³⁺), the EO electron density and solvating ability toward high-valence Ni are reduced, weakening adsorption and mitigating Ni surface depletion. The modified coordination structure shifts anions into the primary solvation sheath, favoring the formation of dense, inorganic-rich interphases (LiF, MgF2, AlF3), which block electron leakage and stabilize the interface. This simultaneously raises the oxidative stability window (>5 V) and, by weakening Li–EO/Li–TFSI binding and increasing Li⁺ transference number, maintains practical ionic conductivity and transport. Operando TXM/XAS confirms homogeneous SOC and the absence of a Ni-poor surface layer in the modified electrolyte, while tomography shows lower porosity and fewer microcracks, indicating improved structural stability. Collectively, these effects enable 4.8 V-class operation with extended cycling and high energy density in both coin and pouch formats, demonstrating a route to decouple oxidative stability from ionic transport in PEO-based solid electrolytes.

Introducing Mg²⁺ and Al³⁺ Lewis-acid coordination into PEO-LiTFSI effectively suppresses EO–cathode reactivity, elevates oxidative stability to ~5 V, and improves Li⁺ transport (higher diffusion coefficient and t⁺). The resulting inorganic-rich, robust CEI stabilizes high-voltage interfaces, enabling 4.8 V-class all-solid-state batteries with long cycle life and pouch cells exceeding 586 Wh kg⁻1. Industrial-scale manufacture of electrolyte membranes (50 m rolls) and assembly of a 20 Ah all-solid-state polymer battery underscore the engineering viability of this strategy. Future work may optimize anode interfacial chemistry to reduce impedance growth, explore alternative Lewis-acid species and concentrations for further performance gains, and evaluate long-term stability across broader temperature/pressure conditions and with diverse high-voltage cathode chemistries.

While the Lewis-acid coordination improves cathode-side stability, LUMO calculations and experiments indicate a slightly lowered LUMO (−2.11 eV vs −1.51 eV for PEO), leading to increased anode interfacial impedance and cyclic overpotential due to low-conductivity SEI components (e.g., MgF2, AlF3). Most cycling tests were conducted at 60 °C and under pressure (1 MPa for pouch cells), so room-temperature and low-pressure performance require further validation. The use of perchlorate salts necessitates careful dehydration and handling to ensure moisture-free conditions and safety; robustness under varying ambient conditions was only assessed in controlled, dry environments.