Solid-state lithium batteries (SSLBs) have become a key focus in next-generation energy storage technology development due to their inherent safety and high energy density. The performance of their core component - solid-state electrolytes (SSEs) - directly determines the overall battery performance^[1]. Among various solid-state electrolytes, polyethylene oxide (PEO)-based solid polymer electrolytes (SPEs) have been extensively studied for their excellent flexibility, good interfacial compatibility, and ease of processing^[4]. However, the high crystallinity of PEO results in low room-temperature ionic conductivity (typically 10^-6-10^-5 S cm^-1), which severely limits its practical applications at room temperature^[8]. To overcome this bottleneck, introducing inorganic fillers into the PEO matrix - has been proven to be one of the effective strategies^[11].

Among various inorganic fillers, garnet-type Lithium Lanthanum Zirconium Oxide (LLZO) has garnered significant attention due to its outstanding bulk lithium-ion conductivity and excellent electrochemical stability^[16]. There is much literature on the addition of LLZO nanoparticles and nanowires to polymer electrolytes to improve electrolyte performance^[22]. However, LLZO materials face a common practical challenge: their extreme sensitivity to moisture (H₂O) and carbon dioxide (CO₂) in the air, leading to frequent H⁺/Li⁺ exchange reactions on the surface, resulting in residual alkali layers such as LiOH and Li₂CO₃^[25,26]. Traditional research has typically regarded residual surface alkali as insulating impurity phases, believing they increase interfacial impedance and hinder ion transport^[27,28]. Consequently, extensive efforts have focused on removing residual alkali through post-processing techniques such as acid washing and heat treatment, aiming to achieve the filler’s intrinsic high conductivity^[29,30].

In this work, we unveil the positive role of residual surface alkali in optimizing electrolyte interface transport. Through precise control of the synthesis pathway, we prepared two LLZO fillers with distinct surface chemical states. Scheme 1A demonstrates the dry-LLZO preparation using a solvent-free method (dry method)^[31], where ball milling and calcination were conducted under fully sealed, air-free conditions, yielding a clean surface with minimal residual alkali. As shown in Scheme 1B, wet-LLZO was synthesized via solid-state synthesis and ground in isopropanol solvent and air (wet method)^[32], resulting in a surface rich in residual Li₂CO₃ alkali layer. Both fillers were incorporated into the PEO/LiTFSI system at equal proportions (15 wt%) to prepare the corresponding composite electrolytes CSPE-15 wet-LLZO and CSPE-15 dry-LLZO, enabling systematic investigation of the differential effects of surface residual alkali on electrolyte structure and performance. Comprehensive characterization reveals that residual alkali, such as Li₂CO₃ on the wet-LLZO surface, disrupts the regularity of PEO chain segments, thereby substantially decreasing their crystallinity. Benefiting from this, CSPE-15 wet-LLZO achieves a high ionic conductivity of 0.80 mS cm^-1 and a lithium-ion transference number of 0.36 at 60 °C. The assembled Li/LiFePO₄ all-solid-state battery maintains a discharge specific capacity of 142.5 mAh g^-1 and a coulombic efficiency exceeding 99% after 300 cycles at a 1.0 C rate. This study deepens the understanding of the surface characteristics of LLZO in SPEs and provides a novel, promising strategy for designing high-performance composite solid-state electrolytes.

Supplementary Figure 3 illustrates the preparation process of the PEO-based composite electrolytes containing LLZO fillers. The digital photographs in Supplementary Figure 4 show the 18 mm-diameter SPE-blank, CSPE-15 dry-LLZO, and CSPE-15 wet-LLZO electrolyte membranes, confirming that flexible and self-standing films were successfully obtained. The morphologies of dry-LLZO and wet-LLZO were examined by SEM, as shown in Figure 1A and 1B. The particle-size distribution histograms inserted in these two figures were obtained from SEM image analysis. The average particle size of dry-LLZO is 0.908 ± 0.346 µm, while that of wet-LLZO is 0.859 ± 0.422 µm, indicating that both fillers are mainly distributed in the submicrometer range. Although the two samples show similar average particle sizes, wet-LLZO presents a slightly broader size distribution, which may be associated with partial aggregation during wet processing and subsequent drying. In terms of morphology, dry-LLZO exhibits a more distinct granular structure and a relatively dry surface, whereas wet-LLZO shows less pronounced granular features. This difference is likely related to IPA-assisted mixing and the formation of surface residual alkali during air exposure. EDS elemental mapping further confirms the successful preparation of both dry-LLZO and wet-LLZO. As shown in Figure 1C-E, the electrolyte membrane of SPE-blank, CSPE-15 dry-LLZO, and CSPE-15 wet-LLZO exhibits a uniform thickness of approximately 80 µm. The cross-sections of SPE-blank and CSPE-15 wet-LLZO are relatively smooth and dense, whereas CSPE-15 dry-LLZO shows a considerably rougher cross-sectional morphology. This phenomenon can be ascribed to the inferior particle size uniformity of the dry-LLZO powder and its poor dispersibility in acetonitrile solvent without prior solvent infiltration.

The XRD patterns of dry-LLZO and wet-LLZO particles are shown in Figure 2A. Both wet- and dry-LLZO particles exhibited cubic phases after calcination at 850 °C for 6 h. Notably, the dry-LLZO displayed a broad full width at half maximum (FWHM) in its XRD pattern, accompanied by weak La₂O₃ impurity peaks (PDF#05-0602). This phenomenon is attributed to the high hardness of La₂O₃, which led to incomplete pulverization and uneven mixing during the dry mixing process. However, trace amounts of La₂O₃ are not expected to significantly affect the final sintered powder^[39]. The FTIR spectrum of wet-LLZO [Figure 2B] exhibits distinct characteristic peaks assigned to LiOH·H₂O and Li₂CO₃ (labeled i-viii). The peaks corresponding to LiOH·H₂O are identified as follows: the O-H stretching vibration at 3,566 cm^-1 (i), the H-O-H bending vibrations at 1,579 cm^-1 (ii) and 999 cm^-1 (iii), and the O-H-related vibrational modes at 678 cm^-1 and 632 cm^-1 (iv and v). Meanwhile, the characteristic peaks of Li₂CO₃ are observed at 1,504 cm^-1 and 1,438 cm^-1 (vi and vii, assigned to the C=O stretching of CO₃^2-) and at 863 cm^-1 (viii, assigned to the C=O bending of CO₃^2-). These results confirm the presence of substantial LiOH·H₂O and Li₂CO₃ on the surface of wet-LLZO particles. This surface residual alkali can be rationalized by the following mechanism: during ball milling, H⁺ ions derived from IPA and adsorbed moisture undergo ion exchange with Li⁺ ions in the LLZO lattice, generating CH₃CH(OLi)CH₃ and LiOH/LiOH·H₂O. During the subsequent drying process, these reaction products further react with atmospheric H₂O (ambient concentration: 10~16 g m^-3) and CO₂ (atmospheric concentration: ~0.8 g m^-3), ultimately leading to the formation of Li₂CO₃ on the surface of wet-LLZO^[40]. In contrast, the FTIR spectrum of dry-LLZO exhibits almost no characteristic peaks assignable to LiOH·H₂O (i-v), with only extremely weak residual signals corresponding to Li₂CO₃ (vi-viii), confirming the presence of only a negligible amount of surface residual alkali (LiOH and Li₂CO₃). Raman spectroscopy [Figure 2C] reveals that the characteristic peaks of both dry-LLZO and wet-LLZO are derived from the same LLZO matrix. Specifically, the Raman spectrum of wet-LLZO exhibited a distinct characteristic peak of Li₂CO₃, while dry-LLZO showed only a weak Li₂CO₃ peak. This result further confirms that the surface of dry-LLZO contains negligible residual alkali.

To further elucidate differences between LLZO particles synthesized by wet and dry methods, XPS was performed on both wet-LLZO and dry-LLZO samples. In the Li 1s spectrum [Figure 2D], Li₂CO₃ exhibits characteristic Li-O peaks at 54.83 eV (dry-LLZO) and 55.18 eV (wet-LLZO), while the peaks centered near 54.2 eV correspond to the Li-O bond within the LLZO lattice. Comparison of the two samples reveals a significantly stronger Li₂CO₃ peak for wet-LLZO relative to dry-LLZO. Analysis of the C 1s spectrum [Figure 2E] shows a peak at 289.85 eV, which, according to literature assignments, is associated with the C-O bond in Li₂CO₃^[41,42]. This carbonate-related peak is markedly more intense in wet-LLZO than in dry-LLZO, consistent with the observations in the Li 1s spectrum. In the O 1s spectrum [Figure 2F], peaks centered near 531.0 eV are attributed to Li₂CO₃ and LiOH species, while the peak at approximately 529 eV arises from lattice oxygen within LLZO. Wet-LLZO exhibits a significantly higher peak intensity ratio of residual alkali (Li₂CO₃/LiOH) to lattice oxygen than dry-LLZO, indicating more abundant surface alkali residues in the wet-synthesized material^[43]. Taken together, these XPS results demonstrate that wet- and dry-processed LLZO powders possess distinctly different surface alkali characteristics: wet-LLZO exhibits substantial residual Li₂CO₃ and LiOH, whereas dry-LLZO shows minimal surface alkali.

XRD was employed to evaluate the crystallinity of CSPE-15 wet-LLZO, CSPE-15 dry-LLZO, and SPE-blank [Figure 3A]. The SPE-blank sample displayed well-defined PEO crystalline peaks at the 2θ positions of 19° and 23°, corresponding to the (120) and (112) planes, respectively, as indexed in the PEO reference pattern (PDF card 45-0109). Incorporation of either wet-LLZO or dry-LLZO particles decreased the intensities of these PEO crystalline peaks, reflecting reduced polymer crystallinity. Notably, CSPE-15 wet-LLZO exhibited a more pronounced attenuation of PEO diffraction intensity compared to CSPE-15 dry-LLZO. DSC measurements [Figure 3B] further corroborated these findings. The glass transition temperature (T_g) of CSPE-15 wet-LLZO was -44.51 °C, lower than that of CSPE-15 dry-LLZO (-42.86 °C) and substantially lower than that of SPE-blank (-41.96 °C). This progressive decrease in T_g indicates enhanced disruption of PEO chain packing and reduced crystallinity upon LLZO incorporation, with wet-LLZO exerting a stronger effect than dry-LLZO. Collectively, XRD and DSC analyses consistently demonstrate that wet-LLZO incorporation is more effective than dry-LLZO in suppressing PEO crystallinity, which may be attributed to differences in particle-polymer interfacial interactions and dispersion behavior.

The thermal stability of SPE-blank, CSPE-15 wet-LLZO, and CSPE-15 dry-LLZO was evaluated and analyzed using TGA curves [Supplementary Figure 5]. Before 200 °C, all three samples exhibited approximately 2% to 3% weight loss due to water vapor evaporation. Between 240 and 700 °C, significant weight loss was observed, caused by the decomposition of PEO and LiTFSI. Further TGA analysis using Derivative Thermogravimetry (DTG) revealed that the thermal decomposition temperature of the SPE-blank electrolyte was 402 °C. After adding the fillers, the decomposition temperatures of CSPE-15 dry-LLZO and CSPE-15 wet-LLZO increased to 417 and 414 °C, respectively. Therefore, the addition of wet-LLZO and dry-LLZO fillers can enhance the thermal stability of CSPEs.

To investigate the electrochemical effects of residual alkali on the surface of LLZO on composite electrolytes, fundamental electrochemical performance tests were conducted on SPE-blank, CSPE-15 dry-LLZO, and CSPE-15 wet-LLZO. The Li⁺ conductivity curves with temperature variation are shown in Figure 3C. At 60 °C, the Li⁺ conductivity of CSPE-15 wet-LLZO was 0.8 mS cm^-1, while that of CSPE-15 dry-LLZO and SPE-blank were only 0.41 mS cm^-1 and 0.16 mS cm^-1, respectively. This is mainly attributed to the abundant residual alkali of Li₂CO₃ on the wet-LLZO surface, which further reduces the crystallinity of PEO on the basis of LLZO. As the PEO segment moves, the ionic conductivity is enhanced. The non-uniform particle size of dry-LLZO results in poor dispersion within acetonitrile, leading to suboptimal interfacial contact between the composite electrolyte film and the electrode. Furthermore, the presence of large particles impedes Li⁺ migration by obstructing efficient transport pathways. Consequently, the ionic conductivity of CSPE-15 dry-LLZO is lower than that of CSPE-15 wet-LLZO.

The electrochemical stability of the PEO-based solid electrolytes was evaluated using LSV, as shown in Figure 3D. CSPE-15 wet-LLZO maintained a stable voltage plateau up to approximately 5.04 V, indicating superior oxidative stability. In contrast, CSPE-15 dry-LLZO and SPE-blank reached instability at lower potentials of about 4.87 and 4.31 V, respectively, highlighting their limited voltage tolerance. Figure 3E-G present the Li⁺ transference number (t_Li⁺) for the three samples: CSPE-15 wet-LLZO exhibited a t_Li⁺ of 0.36, significantly higher than that of CSPE-15 dry-LLZO (0.28) and SPE-blank (0.20). This trend suggests that the incorporation of wet-LLZO not only enhances ionic conductivity but also improves the selective transport of Li⁺ ions, which is critical for the performance of solid-state battery systems.

The critical current density (CCD) is a key parameter for assessing a polymer electrolyte’s ability to suppress lithium dendrite penetration under high current conditions. Symmetric Li/SPEs/Li cells were assembled to determine CCD, with results illustrated in Figure 4A-C. The current density was increased stepwise at a rate of 0.05 mA cm^-2 per hour (0.5 h stripping followed by 0.5 h plating), and the corresponding cell voltage was monitored continuously. Lithium symmetric cells incorporating CSPE-15 dry-LLZO and SPE-blank exhibited pronounced polarization at current densities of 0.4 mA cm^-2 and 0.1 mA cm^-2, respectively, indicative of internal short circuits that precluded sustained cycling. In contrast, the cell employing CSPE-15 wet-LLZO maintained stable operation until reaching a short-circuit event at 0.65 mA cm^-2. This demonstrates that CSPE-15 wet-LLZO possesses markedly superior lithium dendrite suppression capability, enabling stable lithium stripping/plating at higher current densities compared with both dry-LLZO and blank polymer electrolyte systems.

EIS of the three lithium symmetric cells are shown in Figure 4D. The semicircular feature in the high- to mid-frequency range corresponds to the charge transfer resistance (R_ct)^[44]. Among the samples, the Li/CSPE-15 wet-LLZO/Li cell exhibited the lowest R_ct value, indicating that the presence of wet-LLZO effectively buffers interfacial potential fluctuations and thereby promotes interfacial stability during cycling. The long-term galvanostatic cycling performance of the lithium symmetric cells was evaluated at 60 °C under a current density of 0.1 mA cm^-2 with a fixed capacity cutoff of 0.1 mAh cm^-2, as shown in Figure 4E. The Li/CSPE-15 wet-LLZO/Li cell demonstrated exceptional electrochemical durability, sustaining stable cycling for over 1,000 h with a steady polarization voltage stabilized at approximately 30 mV. In contrast, the Li/SPE-blank/Li and Li/CSPE-15 dry-LLZO/Li cells maintained relatively high initial polarization voltages and experienced a rapid decline in polarization voltage to 0 V after 248 and 381 h, respectively. This abrupt voltage drop is attributed to irreversible degradation of the electrolyte-electrode interfacial layer induced by repeated lithium deposition and stripping. Consequently, these results underscore that the CSPE-15 wet-LLZO composite effectively preserves the structural and electrochemical integrity of the electrolyte/lithium metal interface. This interfacial robustness is pivotal to the superior long-term stability observed in the wet-LLZO-based system.

The EIS measurement of Li|LiFePO₄ full cells with different SPEs was tested at 60 °C [Figure 5A]. The obtained EIS spectra consist of ohmic impedance (first semicircle), charge transfer impedance (middle semicircle), and diffusion resistance (linear portion)^[45]. Compared to the SPE-blank control, the full cell incorporating CSPE-15 wet-LLZO exhibited a significant reduction in ohmic impedance from 21.4 Ω to 18.2 Ω and charge transfer impedance from 101.0 Ω to 28.5 Ω. These findings indicate that the addition of wet-LLZO effectively facilitates Li⁺ migration and lowers the overall impedance within the SPEs. Figure 5B presents the rate performance of Li/SPEs/LiFePO₄ full cells. At current rates of 0.2, 0.5, 1.0, and 3.0 C, the Li/CSPE-15 wet-LLZO/LiFePO₄ cell delivered discharge-specific capacities of 167.7, 166.8, 159.9, and 105.3 mAh g^-1, respectively. In comparison, the Li/CSPE-15 dry-LLZO/LiFePO₄ and Li/SPE-blank/LiFePO₄ cells exhibited notably lower capacities at the corresponding rates: 150.6, 147.3, 138.4, and 70.4 mAh g^-1 for the dry-LLZO composite, and 120.7, 115.8, 105.6, and 64.5 mAh g^-1 for the blank electrolyte. These results clearly indicate that the wet-LLZO-containing cell outperforms the other two electrolytes over all applied current rates. Moreover, when the current rate was returned from 3.0 C to 0.5 C, the specific capacity of the Li/CSPE-15 wet-LLZO/LiFePO₄ full cell fully recovered to its initial value, demonstrating excellent rate capability and electrochemical reversibility. This recovery underscores the excellent structural and interfacial stability imparted by the CSPE-15 wet-LLZO composite electrolyte during dynamic cycling under varied current loads.

The long-term cycling stability of solid-state lithium metal batteries was systematically investigated at 60 °C and 30 °C. The Li|LiFePO₄ battery using CSPE-15 wet-LLZO maintained a specific capacity of 156.8 mAh g^-1 and a capacity retention rate of 99.7% after 100 cycles at 0.2 C and 60 °C [Figure 5C]. By contrast, under identical test conditions, the discharge specific capacities of the cells based on CSPE-15 dry-LLZO and SPE-blank decreased to 128.3 mAh g^-1 and 116.8 mAh g^-1, respectively, after 100 cycles. Furthermore, the Li/CSPE-15 wet-LLZO/LiFePO₄ cell exhibited a more stable voltage plateau and lower polarization [Figure 5D]. Supplementary Figure 6 further compares the cycling performance of the three solid-state Li|LiFePO₄ cells at 0.5 C and 1.0 C under 60 °C with a LiFePO₄ loading of 1.5 mg cm^-2. The long-cycle results over 100 cycles at 0.5 C [Supplementary Figure 6A] demonstrate that the CSPE-15 wet-LLZO battery delivers the highest discharge-specific capacity among all samples within these 100 cycles. Its charge-discharge profiles recorded up to 150 cycles [Supplementary Figure 6B] further reveal a stable voltage plateau from the 10th to 150th cycle. At 1.0 C, the long-cycle performance after 300 cycles [Supplementary Figure 6C] indicates that CSPE-15 wet-LLZO maintains a discharge specific capacity of 142.5 mAh g^-1, with its coulombic efficiency exceeding 99%. In contrast, CSPE-15 dry-LLZO shows significant capacity degradation after only 200 cycles, while SPE-blank exhibits substantial capacity loss after merely 133 cycles, as evidenced by its charge-discharge curves [Supplementary Figure 6D], which confirm minimal polarization deterioration over 50-200 cycles. Collectively, the results from these subfigures demonstrate that the CSPE-15 wet-LLZO composite electrolyte achieves both high reversible capacity and excellent long-cycle stability. The long-term cycling stability of Li/SPEs/LiFePO₄ full cells was also evaluated at 30 °C via galvanostatic charge-discharge at 0.2 C. As shown in Figure 5E, the Li/CSPE-15 wet-LLZO/LiFePO₄ cell retains 107.9 mAh g^-1 after 130 cycles, while the Li/CSPE-15 dry-LLZO/LiFePO₄ and Li/SPE-blank/LiFePO₄ cells only deliver 98.6 mAh g^-1 and 72.4 mAh g^-1, respectively, under the same conditions. Notably, the wet-LLZO cell exhibits a flatter voltage profile with minimal polarization throughout cycling [Figure 5F], confirming superior interfacial stability and efficient charge transport. These results verify that the wet-LLZO composite electrolyte enables high-capacity, long-life operation of LiFePO₄ full cells even at 30 °C.

This study investigates the positive effect of residual alkali on the surface of LLZO on CSPEs. Combined with characterization results from DSC and XRD, it was found that residual alkali on the LLZO surface can reduce the crystallinity of PEO, thereby enhancing the lithium-ion conductivity of CSPEs. Consequently, CSPE with 15 wt%wet-LLZO exhibits favorable Li⁺ conductivity (0.80 mS cm^-1) and Li⁺ transference number (0.36) at 60 °C. The Li/CSPE-15 wet-LLZO/Li battery can achieve stable cycling for 1,000 h at 60 °C and 0.1 mA cm^-2 current density, with a critical current density of up to 0.65 mA cm^-2. Additionally, the Li|LiFePO₄ battery using CSPE-15 wet-LLZO demonstrates excellent rate performance (cycling specific capacity of 105.3 mAh g^-1 at 60 °C and 3.0 C) and good long-cycle performance (discharge specific capacity maintained at 156.8 mAh g^-1 after 100 cycles at 60 °C and 0.2 C). This study enhances the understanding of LLZO surface characteristics in solid polymer electrolytes and offers a new, effective approach for developing high-performance composite solid-state electrolytes.