The SEM images of the obtained electrolytes directly reveal the grain size and grain boundary distribution of as-prepared electrolyte (PAL-C) and rolled electrolyte (PAL-R) electrolyte. The PAL-C electrolyte exhibits a compact spherulite polycrystalline structure with the spherulite diameter of 50 µm (Fig. 1a and S1a). For PEO-based SPE, li-ions are thought to be mainly transported through grain boundaries and amorphous phase. Therefore, large-grained PAL-C SPE obtained by casting is unfavorable for li-ion transport and limit the conductivity of electrolytes. Rolling treatment can break the electrolyte grain, which can significantly reduce the crystallinity and increase li-ion transport pathway. After press rolling, the large spherulite disappeared and the electrolyte showed a relatively homogeneous structure for PAL-R (Fig. 1b and S1b). This uniform homogeneous structure is considered to have obvious advantages in improving the conductivity of SPEs.
In order to further analyze the change of crystallinity before and after rolled, XRD test was conducted and the results are shown in Fig. 1c. The diffraction peaks of PAL-C electrolyte at 19.0° and 23.2° are sharp and intense, indicating the highly crystalline nature.[22, 23] As a comparison, the diffraction pattern of PAL-R exhibits several broad and weak peaks, suggesting that the crystallinity of the PAL-R greatly reduced after press rolling. Besides, the main XRD peaks of PEO at 19.0° are also characterized by significant changes in the full width at half maximum (0.216 for PAL-C and 0.323 for PAL-R), implying that the amorphous phase in the electrolyte has increased. The decrease in crystallinity is believed to have a significant effect on the improvement of the conductivity.
The DSC profiles of PAL-C and PAL-R electrolytes were tested and shown in Fig. 1d, which reveals the glass transition temperature (Tg) differences between the two electrolytes. The results suggest that the Tg of PAL-R is -49.17 °C, which is lower than that of PAL-C (-46.78 °C). This result shows that in PAL-R electrolyte, the movement of polymer segments can occur at lower temperature, which leads to a higher ionic conductivity than that in PAL-C electrolyte.
Ion conductivity σ of PAL-C and PAL-R SPE is calculated with the following equation:
where S, L, and R represent the geometric area of stainless steel blocking electrodes, the thickness of electrolytes, and the bulk resistance of the sample obtained from the impedance plots, respectively. The impedance spectra of PAL-C and PAL-R solid polymer electrolytes at different temperatures are tested and shown in Fig. S2. Figure 1c shows the temperature dependence of the calculated ionic conductivity of the PAL-C and PAL-R electrolytes. The as-prepared PAL-R electrolyte reaches an ionic conductivity of 7.58 × 10− 5 S cm− 1 at 25 °C and 1.03 × 10− 3 S cm− 1 at 60 °C, which is two times higher than that of PAL-C electrolyte (3.58 × 10− 5 S cm− 1 at 25 °C and 7.43 × 10− 4 S cm− 1 at 60 °C) and better than that of PEO-based SPE prepared through other methods. [11, 24, 25] The enhancement in li-ion conductivity is attributed to the crystallinity reduction of the PEO-based SPE after the press rolling process and is expected to lead to good battery performance. The relationship between log σ and 1000/T of the PAL-C and PAL-R SPEs reveals that the temperature dependence of conductivity follows Vogel-Tammann-Fulcher (VTF) empirical equation:[7, 13, 26, 27]
where σ, Ea, σ0, T and R represent the ionic conductivity, activation energy, pre-exponential factor, a temperature factor and the ideal gas constant, respectively. The Ea of PAL-C and PAL-R were calculated using the VTF equation (Fig. 1e), and the results show the fitting value of Ea for PAL-R is 5.0 × 10− 2 eV, which is much smaller than that for PAL-C (5.8 × 10− 2 eV). The lower Ea demonstrates that the li-ion movement in PAL-R electrolyte needs less energy than that in PAL-C electrolytes, indicating a higher conductivity.
The mechanical property of SPE is directly related to its barrier effect to lithium dendrite. Figure 1f shows the stress-strain test results of PAL-C and PAL-R SPEs. The ductility of the PAL-R SPE reaches to 1990%, which is much higher than that of PAL-C SPE (1470%). This reinforced ductility of PAL-R SPE would significantly improve the tolerance to dendrite penetration and inhibit short circuit in batteries.
The inhibiting effect of the two SPE on lithium dendrite grows is tested with a Li/SPE/Li symmetric cell. Before the test proceeds, Electrochemical Impedance Spectroscopy (EIS) is conducted to analysis the Li-SPE interface properties of different cells and the results are shown in Fig. 2a. The EIS plots are fitted with a simple mode which consists of ohmic resistance, interface resistance (Rf), charge transfer resistance (Rct), constant phase elements (CPE1 and 2) and Warburg diffusion resistance (Wo)[26]. The simulated results of Rf and Rct in the battery using PAL-R electrolyte are calculated to be 5.72 Ω and 17.65 Ω, respectively, which are smaller than those using PLA-C electrolytes (5.99 Ω and 21.77 Ω). This result indicated that amorphous phase in PAL-R SPE could improve the li-ion transportation and Li-electrolyte interface connection. After the initial EIS test, a 10 mV DC voltage was applied to the Li/SPE/Li symmetric cells to investigate the li-ion transference number in different SPE. Based on the current-time curve (Fig. 2b), impedance before and after polarization (Fig. 2a and S3), the li-ion transference number for PAL-R SPE is calculated to be 0.24, which is higher than the value of PAL-C SPE (0.16). This improvement can be attributed to the reduction of crystallinity phase, which releases more li-ions for ion transportation. After the EIS test, the Li/PAL-C/Li and Li/PAL-R/Li symmetric cells were charged and discharged at 60 °C for 30 min under current densities of 0.1, 0.2, and 0.3 mA cm− 2, respectively (Fig. 2c). From this result, we can find that the voltage of Li/PAL-R/Li cell can be stabilized at 33 mV and 67 mV at current densities of 0.1 and 0.2 mA cm− 2, respectively, which are much smaller than that of Li/PAL-C/Li (56 and 126 mV). For higher current density (0.3 mA cm− 2), PAL-R SEP can stably cycle for 200 cycles, but dendrite penetration occurs after only a few cycles of PAL-C SPE under the same current density. The surface morphologies of lithium electrode with different SPEs after 200 cycles at 0.2 mA cm− 2 were tested and shown in Fig. 2c and 2d. There are massive irregular lithium dendrites with the PAL-C SPE but a relatively smooth lithium surface with PAL-R SPE could be found. This result can be attributed to the high ionic conductivity and uniform ion transportation pathway of PAL-R SPE, which will lead to the uniform lithium deposition to avoid the internal short circuit caused by lithium dendrite growth.
Galvanostatic charge-discharge performance of all-solid li-ion batteries containing LiFePO4 (LFP) cathode, Li anode with different SPEs are tested and the results are shown in Fig. 3. Before galvanostatic charge-discharge test, the impedance of each battery was tested and fitted with an equivalent circuit model (inset of Fig. 3a). In this model, RΩ corresponds to the ohmic resistance; Rct represents the charge transfer resistance for electrochemical reactions; CPE is the constant phase angle element related to the double-layer capacitance of porous cathode and Zw is the finite length Warburg contribution. It is found that RΩ decreases from 17.1 Ω to 14.4 Ω and Rct decreases from 47.5 Ω to 33.1 Ω for battery with PAL-C and PAL-R SPE, respectively, as shown in Fig. 3a. The decreased RΩ and Rct can be attributed to the lower crystallinity of PAL-R SPE, which can provide more li-ion transportation pathways to enhance the conductivity of the electrolyte and facilitate redox reaction in the LFP electrode simultaneously.
Figure 3b shows the charge-discharge capacities of all-solid li-ion batteries with different SPE at 60 °C under current density of 0.2 C. The battery with PAL-R electrolyte delivers a discharge capacity of 162.6 mAh g− 1 with the discharge-charge voltage gap of 60 mV, while the battery with PAL-C electrolyte delivers a discharge capacity of 156.7 mAh g− 1 with the discharge-charge voltage gap of 82 mV. The increased discharge capacity and decreased voltage gap can be attributed to the higher conductivity and lower resistances of the PAL-R electrolyte compared with PAL-C electrolyte. The rate performance of all-solid li-ion batteries with different SPE was conducted under current densities of 0.1 C, 0.2 C, 0.5 C, 1 C and 0.2 C (Fig. 3c, 3d and S4), respectively. The results indicate that the battery with PAL-R can deliver a capacity of 164.3, 162.6, 161.8, 157.8 and 161.2 mAh g− 1, respectively. This performance is much better than the battery with PAL-C electrolyte, which only delivers capacities of 161.5, 156.7, 148.7, 142.1 and 151.8 mAh g− 1, respectively. This result illustrated that the PAL-R electrolyte could afford the high rate operation due to the higher conductivity.
The cycle performance of the battery with different SPE was tasted under the current density of 0.5 C (Fig. 3e). For LFP/PAL-C/Li cell, the discharge capacity maintains 117.1 mAh g− 1 after 300 cycles with the capacity decay rate of 0.071% per cycle. For comparison, the LFP/PAL-R/Li cell maintains a discharge capacity of 136.8 mAh g− 1 with the capacity decay rate of 0.048% per cycle under the same condition. The significantly improved cycle performance can be attributed to the higher ionic conductivity and better inhibitory effect on lithium dendrites.
In this work, we applied a simple press rolling technology to improve the performance of PEO-based SPE for all-solid li-ion batteries. The rolled PEO-based SPE shows a decreased crystallinity and increased amorphous phase, which are expected to be benefit for li-ion transportation. After treatment, PEO-based SPE delivers a doubling room temperature conductivity and decreased activation energy. It is experimentally shown the LiFePO4/SPE/Li all-solid li-ion battery with the rolled PEO-based SPE exhibits a rechargeable specific capacity of 162.6 mAh g− 1 with a discharge-charge voltage gap of 60 mV at a current density of 0.2 C, which is much better than that of the cast PEO-based SPE (156.7 mAh g− 1 and 82 mV). Furthermore, the capacity decay rate was reduced to 0.0048% per cycle after 300 cycles at 0.5 C. All the results show that the grain reforming technology is a promising technology to improve the performance of PEO-based SPE.