Supramolecular interactions between EO unites/anions and AlOCWe introduce a kind of metal-oxo molecular cluster, aluminum-oxo rings (denoted as AlOC), to modify the microstructural heterogeneities of PEO-based polymer electrolyte to fabricate superionic PEO conductor (PEO-LiTFSI-AlOC). In detail, AlOC is constructed with inorganic Al8(OH)8 as cores and two types of organic ligands (R1: 4-methylpyrazole and R2: benzoate) as organic shells (Fig. 1a)36. The carboxylic acid part of benzoic acid and the N atom of 4-methylpyrazole act as a bridge, linking metal ions to form a ring. Therefore, the aromatic nucleus is oriented outwards of the molecular ring. Moreover, six Al8 rings will occupy the six faces of the cube to make a perfect (Al8)6 nanocage under π − π stacking interactions between adjacent rings with both mesoporous cages (21.19 Å) and microporous apertures (7.8 Å). Interaction region indicator (IRI) calculations on the guest-cluster interaction indicate there are substantial host–guest interactions that exist between the ligands and LiTFSI (Fig. 1b). The center of the aluminum molecular ring surrounded by four organic ligands (4-methylpyrazole) (site I) and the microporous channels between the nanocages (site II) could capture TFSI−. This results in a torsion of the organic ligand (Supplementary Discussion 1). In site I, TFSI− is trapped in the middle of the molecular ring by twists in the four 4-methylpyrazole ligands. TFSI− will also be captured in the micropore channel, site II, close to the benzoate and 4-methylpyrazole ligands. In addition, compared with site II, there is a hydrogen bond between the hydroxyl group and TFSI− at site I, which is more conducive to the adsorption of the guest. Therefore, the adsorption energy of site I for TFSI− is − 2.9 eV due to multiple interaction forces, which is lower than that at site II (− 2.0 eV). So both sites can make the N–Li bond lengthened (site I: 2.0 Å, site II: 2.1 Å, init. 1.9 Å). The lengthened N–Li bond reduces the ionic bond strength of LiTFSI and induces the dissociation of lithium salt in the polymer electrolyte. Furthermore, the significant adsorption of anions by AlOC may reduce the strength of the Coulomb interactions between anions and cations.Fig. 1: The supramolecular interaction and effect of aluminum-oxo rings with of lithium bis(trifluoromethane sulfonyl) imide.a the nanocage assembly process of Al8(OH)8(R1)8(R2)8 (R1 = 4-methylpyrazole, R2 = benzoate) and the simulated adsorption sites of lithium bis(trifluoromethane sulfonyl) imide (LiTFSI) on the nanocage. b the adsorption simulation details for Site I of the aluminum-oxo ring and Site II of the assembled mesoporous cage. In order to highlight the structural interaction, the non-interaction parts are virtualized by space-filling mode, while the important interaction sites and guests are represented by standard ball-and-stick mode. More details can be seen in Note S1. c the schematic of the supramolecular interaction of aluminum-oxo rings (AlOC) with poly(ethylene) oxide (PEO) segment. d the stretching vibration of ether oxide groups (C–O–C) in the region of 1050–1170 cm−1 of PEO and PEO-AlOC.In addition, the supramolecular interactions between AlOC and PEO could break the crystallization tendency of PEO (Fig. 1c). We took the three-monomer molecular unit of PEO (CH2CH2OCH2CH2OCH2CH2O) as a model to perform theoretical calculations. As shown in Supplementary Fig. 1, there are mainly four types of sites on the nanocage involved in the interactions with PEO. Among them, Site I and Site II interact with PEO chains through a single aluminum molecular ring. The PEO chains are located in the pores formed by alternating up and down benzene rings and pyrazole rings. Site III and Site IV interact with PEO chains through the conjugated pore cavity formed by two vertically packed aluminum molecular rings. Moreover, all sites are generated through the interactions of C–H···π and C–H···O formed by the organic ligands on the aluminum molecular ring and PEO. It is worth noting that Site IV has the lowest adsorption energy of – 1.18 eV, which is the most likely interaction site. Attenuated total reflection Fourier transform infrared (ATR-FTIR) spectra provide evidence for such interactions. As shown in Supplementary Fig. 2, when AlOC is mixed with PEO, two characteristic peaks of C–H in the aromatic nucleus also shifted from 722 cm−1 and 685 cm−1 to 728 cm−1 and 687 cm−1, respectively. In addition, the stretching vibration of ether oxide groups (C–O–C) in PEO at 1092 cm−1 will shift to 1097 cm−1 (Fig. 1d). The ether oxide groups (C–O–C) in PEO are the important parts of the chain folding structure in the PEO which usually induce the formation of crystalline phase of PEO37. On the other hand, the changes of Li+ and TFSI− in micro- and mesoscopic scales also influence the crystallization tendency of PEO. Obviously, such interactions will undoubtedly prevent the dense chain folding structure from the close packing between the segments to form the crystalline region.Multi-scale characterization for Li-ions conduction networkSuch supramolecular interactions have a great impact on the ion conduction network in the solid-state electrolytes. It is well known that the migration of ions in solid electrolytes is a multiscale process composed of mechanisms that manifest at different length scales, from the atomic scale up to the macroscopic scale. Importantly, the ionic conductivity of solid electrolytes is a function of all these mechanisms, which is typically measured on macroscopic samples through impedance spectroscopy measurements. The total macroscopic ionic conductivity is convoluted with each component, such as amorphous phases and crystalline phases. Given the complexity of ion conduction networks in solid-state electrolytes, the capabilities and limitations of single characterization technology could not provide comprehensive insights. In order to provide an accurate and reliable interpretation, we study ion conduction networks from atomic and nanoscale to macroscopic scales by multi-scale complementary characterization method.As shown in Fig. 2a, b, high-resolution solid-state nuclear magnetic resonance is used to detect the local environments of lithium ions in atomic and nanoscale38. With the addition of AlOC, the peak has an obvious shift from 0.24 ppm to 0.37 ppm. It might suggest that in PEO-LiTFSI-AlOC, a less constrained coordination environment of Li-ions was built. Ionic migration, which is accompanied by a constant shift in the coordinating environment, benefits from such a phenomenon. In micro- and mesoscopic scales, the effect of such supramolecular interactions may be manifested in the variation of lithium-ion concentration. As shown in Fig. 2c and Fig. 2d, we employ Raman mapping technology to study the distribution of lithium ions in polymer electrolytes. When LiTFSI is dissolved into PEO polymer, there is a Raman shift of S–N–S in TFSI− anion at 740 cm−1. The linearity of the peak areas at this characteristic Raman shift can be used to determine the difference in the molar ratio of nLiTFSI: nEO across the whole PEO electrolyte (see Supplementary Discussion 2 for detailed information)39. Although the PEO polymer electrolyte was usually prepared with a specific molar ratio of nLiTFSI: nEO, for example, 1:15 in this work, the real Li-ion concentration distribution is not as expected uniformly among the whole electrolyte. The chemical potential difference between the crystalline and amorphous phases leads to a clear and distinct nLiTFSI: nEO heterogeneity among the polymer electrolytes. The statistical result based on the threshold segmentation method shows the average molar ratio of Li:EO in the crystalline phase and amorphous phase is calculated to be 1:16.4 and 1:12.2, respectively. This means that the lithium-ion concentration in the amorphous phase is about 1.3 times that in the crystalline phase. Raman mapping results reveal that the Li-ion concentration in both the amorphous phase and crystalline phase of PEO-LiTFSI-AlOC is changed obviously (Fig. 2d), the average nLiTFSI: nEO in the crystalline phase and amorphous phase is calculated to be 1:19.5 and 1:10.7, respectively (see Supplementary Discussion 2). In this regard, the lithium-ion concentration in the amorphous region is about 1.8 times that in the crystalline region. Compared with PEO-LiTFSI, there are greater concentrations of lithium ions in amorphous phases and lesser concentrations of lithium ions in crystalline phases for PEO-LiTFSI-AlOC. Considering the resistance of ion migration in amorphous phases and crystalline phases, this phenomenon is normally beneficial to the lithium-ion conductivity of polymer electrolytes.Fig. 2: The characterization of the Li-ion conducting network in polymer electrolyte.a, b the 7Li MAS NMR spectra of PEO-LiTFSI (a) and PEO-LiTFSI-AlOC (b), respectively. c, d the Raman mapping characterization of the Li-ion concentration distribution in PEO-LiTFSI (c) and PEO-LiTFSI-AlOC (d), respectively; the left picture shows the 300 μm × 300 μm region at 10 × objective which overlaps the corresponding 10 × magnification optical image, and the right picture shows the small 100 μm × 100 μm region at 50 × objective. The color maps represent the molar ratio of Li:EO (More details can be found in Supplementary Discussion 2). The Li-ion conducting network of polymer electrolyte: e, f the volume renderings with the size of 300 μm × 300 μm × 90 μm of PEO-LiTFSI (e) and PEO-LiTFSI-AlOC (f); g, i the tomographic cross-section of PEO-LiTFSI (g) and PEO-LiTFSI-AlOC (i); (h, j) the proportion of crystalline and amorphous phases on the statistic of 150 XY plane ortho slices of PEO-LiTFSI (h) and PEO-LiTFSI-AlOC (j).In addition, the information in the macroscopic scale is important to study the link between the material properties of polymer electrolyte and their ion conduction performance. We further used high spatial resolution and tunable propagation phase contrast XCT to visualize and quantify the microstructural heterogeneities of PEO-based polymer electrolytes at the macroscopic scale. The tunable propagation phase contrast enables the visualization of materials with low atomic numbers materials and the distinguishing of the poor absorption contrast polymers with different densities, which is different from the conventional absorption contrast-based X-rays absorption characterization technique (Supplementary Fig. 3)40. Thus, tiny changes in PEO polymer electrolyte produce actually considerable phase drifts. As shown in Fig. 2e, a 3D morphological distribution of crystallization and amorphous phases with a high sub-micro (600 nm) voxel resolution in a macroscopic size around 300 µm × 300 µm × 90 µm of PEO-LiTFSI polymer electrolyte can be obtained. The spatial proportion of the crystalline phase is ca. 66.7 % while the amorphous phase is ca. 33.3 %. As a result, the amorphous phase which is believed to be the more efficient Li-ion transportation channel only occupies a very small amount of area, which was mostly surrounded by the large crystalline phases (Fig. 2g). The large size of the crystalline phases will result in a very low proportion of amorphous phases (< 40 %) based on the statistics of all ortho slices cross sections (Fig. 2h). This is unable to form a complete and continuous 3D amorphous network structure at such a low proportion of amorphous phases (Supplementary Fig. 4).There are significant changes in the spatial proportion of the crystalline region and the amorphous region for the PEO-based Li-ion conduction network (Fig. 2f and Supplementary Fig. 5). The spatial proportion of the amorphous phase is around 64.6 % and it is 35.4 % for the crystalline phase in PEO-LiTFSI-AlOC. Compared with the above XCT results of PEO-LiTFSI (Fig. 2e), the spatial proportion of the amorphous phase in PEO-LiTFSI-AlOC (64.6 %) is much higher than the one of PEO-LiTFSI (33.3 %). The continuity of amorphous phases in PEO-LiTFSI-AlOC has been significantly improved. In the tomographic cross-section (Fig. 2i), it is difficult to find any isolated amorphous phase, and most of the crystalline phases are independent small-sized spherulites. The particle size of the crystalline phase is also reduced to around ~ 20–40 µm (Supplementary Figs. 7, 8), which is much smaller than the ones in the PEO-LiTFSI. Thus, due to a much lower spatial proportion and smaller particle size of the crystalline phase, the proportion of the amorphous phase retains a high proportion of > 55 % based on the statistics of all cross sections, and even some planes accounted for as high as 68.2 % (Fig. 2j).For a multiscale ion conduction process, the information from several distinct techniques is inevitably interrelated. The internal connections need further excavated to analyze the link between the material properties of polymer electrolyte and their ion conduction performance (Supplementary Table. 1). Correlating lithium ion concentration and macroscopic component, we could calculate that there is as low as 39.4 % of the lithium ions in total for Li-ion conducting and transporting within the amorphous phases. The mobility of lithium-ion in amorphous phases is higher than lithium-ion in crystalline phases. Moreover, the isolated amorphous region with poor space connectivity, as discussed above, will not contribute the effective Li-ion conduction. Besides, the strong interaction between EO unites and Li-ion in conventional PEO-LiTFSI electrolytes will undoubtedly increase the dissolving of Li-ion salts. However, such strong Li-ions coordination will also retard the motion of Li-ion among the EO units in the polymer electrolytes10. As a result, conventional PEO-LiTFSI electrolytes exhibit as low as 10−6 S cm−1 conductivity at near room temperature. In contrast, the introduction of AlOC reconstructed the ion conduction network from multiple scales. The interactions at the atomic and molecular levels bring about a significant influence on the macroscopic polymer electrolyte to form the continuous amorphous phase structure. The statistical result also shows there is around 76.7 % of the lithium ions in total for Li-ion conducting and transporting via the amorphous phase network (it exceeds the proportion of 39.4 % in PEO-LiTFSI). These results indicate an effective and continuous Li-ion-rich conducting network is successfully built in the bulk of PEO-LiTFSI-AlOC. In addition, the adsorption of anions and segments on AlOC may reduce the resistance of the transformation of Li-ions coordination, which is beneficial for ion conduction.The Li-ions transport dynamics and conductivityDue to the reconstructing in the 3D Li-ion transportation network, the Li-ions in PEO-LiTFSI and PEO-LiTFSI-AlOC deliver distinguish conduction ability when coupled with the segmental relaxation of polymers41. Ion conduction in the crystalline phase with frozen segmental relaxation should be more sluggish than in amorphous phases with rapid segmental relaxation. The amorphous phases with ca. 33.3 % were divided into several isolated amorphous phases by crystalline phase with ca. 66.7 % in PEO-LiTFSI. In this condition, the limitation of the crystalline phase to the diffusion of Li-ions will cause a change in Li-ion transport kinetics. It appears as two distinct Arrhenius plots (Fig. 3f), with an activation energy of 1.21 eV and 0.42 eV below and higher than the melting temperature point of 59.6 °Ϲ (Supplementary Fig. 9), respectively. Due to the high activation energy below Tm, the ionic conductivity of PEO-LiTFSI shows a cliff-like decline, which has a very low ionic conductivity of 10−6 S cm−1 at near room temperature (Supplementary Table. 2). Due to the ability to reconstruct and enhance the ionic conduction network by AlOC, an effective and continuous Li-ion-rich conducting network can be built with a much higher spatial proportion of amorphous regions (64.6 %) and much higher available Li-ions concentrations in the amorphous Li-ion conducting polymer network. This effectively reduces the diffusive lengthscales of Li-ions in the crystalline phase. Thus, the Li-ion conductivity σ follows an Arrhenius equation with a lower activation energy of 0.42 eV in the whole temperature range, which was equivalent to that of PEO-LiTFSI in high temperature. As a result, the PEO-LiTFSI-AlOC superionic PEO conductor exhibited a molten-like Li-ion conduction behavior among the whole temperature range and delivered an ionic conductivity approaching to 1.87 × 10−4 S cm−1 at 35 °Ϲ. The Bruce–Vincent method d.c. polarization measurements also show the change in the lithium-ion transference number. PEO-LiTFSI only shows a low lithium ion transference number of 0.0931, which is much lower than 0.42 of PEO-LiTFSI-AlOC (Supplementary Fig. 10 and Supplementary Table. 3). It indicates the ionic current is carried predominantly by the Li-ion in PEO-LiTFSI-AlOC rather than its counter-ion.Fig. 3: The ionic conductivity of polymer electrolyte.The ionic conductivity (σ) versus temperature (T) of PEO-LiTFSI and PEO-LiTFSI-AlOC, the error bars correspond to the standard deviation from the average \(\log (\sigma )\) obtained by measuring at least five individual experiments at each temperature point. The specific fitting equation of Arrhenius plots and the calculation of activation energy can be found in Methods.The Li plating/stripping behavior in Li symmetric cellsSuch a 3D Li-ion conducting network with high connectivity also brings about an obvious influence on the Li electrochemical plating and stripping behaviors in symmetric Li cells. Due to the poor Li-ion conduction behavior, the Li|PEO-LiTFSI|Li symmetric cell has a large overpotential of 414.8 mV in the preliminary stage at 50 μA cm−2 at 35 °Ϲ (Supplementary Fig. 11). The overpotential will gradually increase to 570 mV and finally sudden short circuit at 366 h due to the Li dendrite penetration. However, there is no obvious voltage oscillation of the Li|PEO-LiTFSI-AlOC|Li symmetric cell at a current density of 50 μA cm−2 for an ultra-long duration of 2000 h (Fig. 4a). In addition, the Li plating/stripping behavior is highly stable when the current density increases from 25 μA cm−2 to 100 μA cm−2 (Fig. 4b). When the current density returned to 25 μA cm−2, the overpotential is recovered to ~ 33 mV and the Li electrochemical plating and stripping can be further operated steadily over 1700 h without short circuit (Supplementary Fig. 12a). PEO-LiTFSI-AlOC superionic PEO conductor even deliver an enhanced rate performance of Li plating/stripping at 50 °Ϲ (still below the melting point). Long-term cycle test confirms that PEO-LiTFSI-AlOC exhibits extremely stable Li plating/stripping behavior at a current density of 50 μA cm−2 for 2000 h with a stable overpotential of ~39.9 mV (Fig. 4c). The symmetric Li cell also shows stable Li plating/stripping behavior for 600 h with the overpotential of ~215.7 mV even at a higher current density of 250 μA cm−2 (Supplementary Fig. 12b). With the current density increasing from 50 to 400 μA cm−2, the overpotential increases from 66.6 mV to 394.5 mV (Fig. 4d). In contrast, under the current density of 250 μA cm−2, the Li|PEO-LiTFSI|Li symmetric cell shows a significant increase in overpotential at initial stage and sudden short circuit after 9 h (Supplementary Fig. 12c). This is because the insufficient Li-ion conductivity of PEO-LiTFSI which will result in the severe Li dendrite in the symmetric cell.Fig. 4: The Li plating/stripping performances of Li | PEO-LiTFSI-AlOC | Li symmetric cells.The long-term cycling performance of Li|PEO-LiTFSI-AlOC|Li symmetric cells at the current density of 50 μA cm−2 at 35 °Ϲ (a) and 50 °Ϲ (c). The inset figures show the voltage curves of the cycle at 1st, 500th, and 1000th. The cycling performance of Li|PEO-LiTFSI-AlOC|Li symmetric cells under different current densities at 35 °Ϲ (b) and 50 °Ϲ (d).The performance of all-solid-state Li | LiFePO4 full batteriesUnlike most of the PEO-based, all solid-state rechargeable batteries require high temperatures (usually higher than 60 °Ϲ) to allow an acceptable battery performance, the Li|PEO-LiTFSI-AlOC|LiFePO4 full batteries exhibit a good performance below its melting point temperature of 56.0 °Ϲ. For example, it exhibits typical discharge plateaus at 3.35 V and charge plateaus at 3.50 V during the Galvanostatic charging–discharging under 35 °Ϲ. High capacities of 161.8, 155.7, 142.2, and 127.5 mAh g−1 can be obtained at various current densities of 25 μA cm−2, 50 μA cm−2, 75 μA cm−2, and 100 μA cm−2, respectively (Fig. 5a). And the capacity can be recovered to 158.0 mAh g−1 when the current density was returned to 25 μA cm−2 (Supplementary Fig. 13). Particularly, the overpotentials are only 153 mV at 25 μA cm−2 and 226 mV at 50 μA cm−2, respectively. These results indicate a superior capacity retention at various current densities can be retained. Moreover, the as-built ASSBs show a stable battery cycling performance at 35 °Ϲ (Fig. 5e). There is no obvious capacity fading over 200 cycles and the specific capacities can be retained around 156 mAh g−1 at 25 μA cm−2 with an average coulombic efficiency (CE) higher than 99.99 %. Whereas Li|PEO-LiTFSI |LiFePO4 only delivers a discharge-specific capacity of ~ 120 mAh g−1, and the CE dramatically decreases after 40 cycles (Supplementary Fig. 14a). When the operating temperature increases to 50 °Ϲ, high and reversible capacity can be further enhanced with much lower overpotential even at a much higher current density (Fig. 5b). The cell maintains very low overpotentials of 113.8 mV at 100 μA cm−2 and 179.8 mV at 200 μA cm−2, respectively. Even at 400 μA cm−2, it delivers a reversible specific capacity of 139.0 mAh g−1 with an overpotential of 299 mV. In the long-term cycling at 250 μA cm−2, the cell also maintains a high discharge capacity of ~ 147 mAh g−1 and sustains > 96 % capacity over 200 cycles (Fig. 5e). In contrast, Li|PEO-LiTFSI|LiFePO4 has poor performance even at 50 °Ϲ with a discharge capacity of ~ 70 mAh g−1 at 250 μA cm−2 (Supplementary Fig. 14b).Fig. 5: The performances of PEO-LiTFSI-AlOC based all solid-state half-cell batteries using Li metal as anode and LiFePO4 as cathode (Li | PEO-LiTFSI-AlOC | LiFePO4).a, b the galvanostatic charge-discharge curves of Li|PEO-LiTFSI-AlOC|LiFePO4 as a function of current density at 35 °Ϲ (a) and 50 °Ϲ (b), respectively. c, d the galvanostatic charge-discharge curves as a function of cathode mass loading under 10 μA cm−2 at 35 °Ϲ (c) and 25 μA cm−2 at 50 °Ϲ (d) respectively; (e) the cycling performances of Li|PEO-LiTFSI-AlOC|LiFePO4 under different areal current densities and operating temperatures. f the battery performances comparison of this work and the reported polymer-based all-solid-state Li batteries (25–40 °Ϲ: green, 41–59 °Ϲ: blue, 60–69 °Ϲ: dark blue and higher than 70 °Ϲ: black). See Supplementary Table. 4 for details. g the galvanostatic charge-discharge curves of the pouch cell with LFP loading of 7.15 mg cm−2 at the current density of 10 μA cm−2.In all solid-state polymer rechargeable batteries, lithium ions can only be transported across the network of the polymer electrolyte via solid-solid interfaces. Thus, a sufficient 3D Li-ion conduction network across the whole electrode is crucial to allow the feasible intercalation/deintercalation of Li-ion in LiFePO4 cathode at a high mass loading. With the increase of the cathode mass loading, the electrode thickness will be increased from 43 μm at 4.1 mg cm−2 to 84 μm at 16.1 mg cm−2 (detailed see Methods, Supplementary Fig. 15 and Supplementary Fig. 16). At the operating temperature of 35 °C, Li|PEO-LiTFSI-AlOC|LiFePO4 full rechargeable batteries can still display high specific capacities of 157.6, 146.2, and 143.4 mAh g−1 with mass loadings of 2.0, 5.2, and 11.0 mg cm−2, respectively (Fig. 5c). And it also delivers a high specific capacity of over 150 mAh g−1 with an overpotential lower than 130.0 mV even at a very high mass loading of 16.8 mg cm−2 at 50 °Ϲ (Fig. 5d). The reversible specific capacities follow an almost linear relationship with the areal capacities at various mass loadings (Supplementary Fig. 17). And a high areal capacity of 1.58 mAh cm−2 and 2.53 mAh cm−2 can be achieved at 35 °Ϲ and 50 °Ϲ, respectively. These performances are advanced in the state-of-art of present polymer-based ASSBs (Supplementary Table. 4)5,6,7,8,9,12,17,24,25,28,42,43,44,45,46,47,48,49. As shown in Fig. 5f, it is still a challenge to make the areal capacity more than 1 mAh cm−2 at the operating temperatures of 25–40 °Ϲ. In order to exceed the areal capacity higher than 2 mAh cm−2, an operating temperature higher than the melting point of polymer electrolyte is usually required. Furthermore, we assemble a single-electrode pouch cell based on the single-side-coated LFP composite cathode (active area of 56 × 43 mm2), and the mass loading of LFP is 7.15 mg cm−2. As shown in Fig. 5g, the pouch cell delivers a reversible capacity of 163.7 mAh g−1, meaning a high output capacity of 1.17 mAh cm−2. Such greatly improved battery rate performance at various mass loading in this work is due to the enhancement and reconstruction of the continuous amorphous ionic conduction network of PEO-LiTFSI-AlOC.In summary, we introduce a kind of aluminum-oxo rings to effectively improve the Li-ion conduction network of PEO-based electrolytes via the supramolecular interactions, including the interactions of C-H···π and C-H···O between AlOC and PEO chains, host–guest interactions between AlOC and LiTFSI. The multi-scale complementary characterization technologies are adapted for the detailed study of ion conduction networks in polymer electrolytes. Lithium ions with different local environments in atomic and nanoscale are quantified by solid-state nuclear magnetic resonance. Raman spectroscopy mapping techniques and high spatial resolution and tunable propagation phase contrast XCT are employed to visualize and quantify microstructural heterogeneities of PEO-based polymer electrolytes. According to multi-scale characterization analysis, we also reveal that the conventional PEO-based electrolyte has up to ~ 60.6 % of lithium ions shackled within the crystalline phase. The amorphous Li-ion conducting polymer network with only 39.4 % available Li-ions is disconnected by the crystalline phase of PEO. In contrast, the PEO-LiTFSI-AlOC superionic PEO conductor has an effective and continuous Li-ion-rich conducting network with a much higher spatial proportion of amorphous regions of 64.6 % than that of 33.3 % in conventional PEO electrolyte. The available Li-ions in the amorphous Li-ion conducting polymer network are also increased to 76.7 %. Such reconstruction of the 3D ion conduction network brings in molten-like Li-ion conduction behavior with an ionic conductivity of 1.87 × 10−4 S cm−1 at 35 °Ϲ and a stable Li electrochemical plating/stripping performance over 2000 h. The as-built Li|PEO-LiTFSI-AlOC|LiFePO4 full batteries show a high rate performance and long cycling performance with capacity retention of more than 90 % and an average CE higher than 99.9 % over 200 cyclings and further enable a high-loading LiFePO4 cathode of 16.8 mg cm−2 with a specific capacity of 150 mAh cm−2.
