Lithium-ion batteries are widely used in 3C electronic products, new energy vehicles, and energy storage power plants due to their high energy density, high power density, high charging and discharging efficiency, long cycle life, and no memory effect [1, 2]. The all-solid-state lithium-ion batteries (ASSLBs) have become a current research hotspot because of the application of a nonflammable solid electrolyte to replace the role of liquid electrolyte and separator, greatly improving the safety of the battery while achieving a higher energy density [3-5]. The main solid electrolytes are sulfide electrolytes [6], oxide electrolytes [7], halide electrolytes [8, 9] and polymer electrolytes [10]. Among them, sulfide electrolytes show high lithium-ion conductivity, low grain boundary impedance, and good mechanical properties, and are considered to be the most promising next-generation high-performance energy storage devices [11, 12].

Li₂S-P₂S₅ system is one of the most widely researched binary sulfide electrolytes, which has the advantages of high ionic conductivity (> 10⁻⁴ S/cm), good stability to lithium, wide electrochemical window and simple preparation process, and has a greater prospect of application in ASSLBs [13, 14]. It is worth noting that Li₂S-P₂S₅ glass-ceramics generate a crystalline phase with high lithium-ion conductivity after heat treatment, and the lithium-ion conductivity is significantly increased. 78Li₂S-22P₂S₅ (x = 78) glass-ceramics is the component of xLi₂S-(100-x) P₂S₅ with the highest ionic conductivity [15]. D. Shin et al. [16] improved the lithium-ion conductivity by controlling the crystallization kinetics of the material. Nan's group [17] used 78Li₂S-22P₂S₅ electrolyte as both active material and electrolyte, and the prepared all-solid-state battery showed excellent electrochemical performance. In contrast, more studies on ASSLBs based on 78Li₂S-22P₂S₅ electrolytes focus on room temperature, while their electrochemical performance in a wider temperature range is unclear.

In this work, we synthesized 78Li₂S-22P₂S₅ glass-ceramic solid electrolytes with high ionic conductivity by exploring the optimal heat treatment temperature during the synthesis process. The LiNi_0.6Co_0.2Mn_0.2O₂/78Li₂S-22P₂S₅/Li-In batteries were assembled, and the electrochemical performance of the battery in a wide temperature range was compared. The battery showed excellent electrochemical performance at room temperature and high temperature (60 ℃). The changes of interfacial impedance during cycling at room temperature and high temperature (60 ℃) were compared by in-situ EIS and relaxation time distribution (DRT). The low interface impedance at high temperatures was the reason for its high discharge capacity, which was also proved by cyclic voltammetry (CV) and other kinetic characterizations. However, stacking pressure experiments show that there was a large volume change at high temperatures, which leads to a decrease in cycle stability.

To investigate the effect of heat treatment on the phase of the final materials, the 78Li₂S-22P₂S₅ glass-ceramic obtained from mechanical milling was annealed at different temperatures. Powder XRD and Raman spectra were applied to study the phase and structure evaluations. Fig. 1a shows the XRD patterns of 78Li₂S-22P₂S₅ glasses before and after the heat treatment. The typical halo pattern appears in the range of 20°~30° with no other strong diffraction peaks, suggesting the successful synthesis of the Li₂S-P₂S₅ glass solid electrolytes [15]. In comparison, the tiny diffraction peaks belonging to the crystallization phase appear in the pattern of 78Li₂S-22P₂S₅ glass-ceramic after 200 ℃. With increasing sintering temperatures from 215 ℃ to 250 ℃, more and more strong diffraction peaks are observed, indicating that the obtained materials possess a better crystallinity. Based on the above XRD results, it can be concluded that the annealing temperature should be higher than 215 ℃ to obtain the 78Li₂S-22P₂S₅ glass-ceramic solid electrolyte. The XRD patterns can be indexed to the thio-LISICON II phase [18]. Furthermore, Raman spectra were also performed. As shown in Fig. 1b, the typical Raman peak corresponds to the PS₄³⁻ tetrahedral unit in the structure is observed at ~422 cm⁻¹ [19]. Those peaks show a clear shift of the sintered samples with increasing sintering temperatures, suggesting the transformation of 78Li₂S-22P₂S₅ from glass to glass-ceramic phase [20], which agrees well with previous XRD results.

Furthermore, AC impedance was performed on the prepared 78Li₂S-22P₂S₅ glass and glass-ceramics electrolytes to determine the optimum sintering temperature for the highly conductive solid electrolyte. As shown in Fig. 1c, all materials show similar impedance spectra consisting of an arc in the high frequencies and a trail in the low frequencies. The 78Li₂S-22P₂S₅ glass shows a total resistance of 270 Ω, while the annealed samples deliver smaller resistances. The total resistances are 120, 100, 90, 240 and 250 Ω for 78Li₂S-22P₂S₅ glass annealed at 200, 215, 225, 235 and 250 ℃, respectively. The corresponding room temperature Li-ion conductivities are 0.35 mS/cm for 78Li₂S-22P₂S₅ glass and 0.85 mS/cm, 0.96 mS/cm, 1.06 mS/cm, 0.41 mS/cm and 0.37 mS/cm for 78Li₂S-22P₂S₅ glass ceramics annealed at these above temperatures, respectively (Fig. 1d). The 78Li₂S-22P₂S₅ glass ceramics shows higher Li-ion conductivity than that of the 78Li₂S-22P₂S₅ glass. The obtained 78Li₂S-22P₂S₅ glass ceramics electrolytes exhibit a trend of increasing and then decreasing Li-ion conductivity with these chosen annealing temperatures from 200 ℃ to 250 ℃ and deliver the highest conductivity (1.06 mS/cm) after sintered at 225 ℃. The activation energy (E_a) of the 78Li₂S-22P₂S₅ glass and glass-ceramics electrolyte annealed at 225 ℃ were deduced based on the temperature-dependent Li-ion conductivities, as shown in Fig. 1e. The AC impedance spectrum of those materials measured at different temperatures is shown in Fig. S1 (Supporting information). The 78Li₂S-22P₂S₅ glass electrolyte delivers an E_a of 0.20 eV, while the 78Li₂S-22P₂S₅ glass ceramic shows a similar E_a value of 0.21 eV. Besides the ionic conductivity, the electronic conductivity of the above solid electrolytes was also characterized using the DC polarization method. The corresponding electronic conductivity of 78Li₂S-22P₂S₅ glass and glass ceramic are 2.75 × 10⁻⁹ S/cm and 0.88 × 10⁻⁹ S/cm, respectively (Fig. 1f). Because of the high Li-ion conductivity and low electronic conductivity of 78Li₂S-22P₂S₅ glass ceramic obtained by annealing at 225 ℃, it was chosen as a solid electrolyte in this work in the following section to assemble all-solid-state batteries. As shown in Fig. S2 (Supporting information), the glass ceramic electrolyte shows a particle size of ~10 µm with uniform distribution of P and S based on the SEM images and EDS mapping results.

The battery performance of the prepared 78Li₂S-22P₂S₅ glass-ceramic electrolyte was evaluated by constructing all-solid-state lithium batteries using LiNi_0.6Mn_0.2Co_0.2O₂ cathode and Li-In anode. As depicted in Fig. 2a, the assembled battery displays a charge voltage platform of ~3.25 V (vs. Li-In) and a discharge voltage platform of ~3.15 V (vs. Li-In) during the first cycle at 0.1 C. It delivers an initial capacity of 216.6 mAh/g and discharge capacity of 149.1 mAh/g with a coulombic efficiency of 69%. After 50 cycles, it maintains a discharge capacity of 130 mAh/g with a capacity retention of 87%, as shown in Fig. 2b. During the cycling process, it keeps high coulombic efficiencies close to 100%, suggesting an excellent lithium intercalation/deintercalation behavior. Moreover, the rate performance of the assembled battery was also evaluated using different charge/discharge rates under room temperature. Fig. 2c shows the discharge voltage profiles of the assembled battery at various C rates. As shown in the figure, the discharge capacity and discharge plateau decrease with increasing C rates from 0.05 C to 1 C at room temperature. The assembled LiNi_0.6Mn_0.2Co_0.2O₂/g-c-225 ℃/Li−In cell shows discharge capacities of 155 mAh/g at 0.05 C, 122 mAh/g at 0.1 C, 90 mAh/g at 0.2 C, and 40 mAh/g at 1 C, respectively. After cycling at different rates, the recoverable capacity is close to 100% when the C rate returns to 0.1 C (Fig. 2d). As a typical inorganic solid electrolyte, the Li-ion conductivity of the 78Li₂S-22P₂S₅ glass-ceramic electrolyte increases at elevated temperatures, providing a Li-ion conductivity of 2.2 mS/cm at 60 ℃. This high conductivity enables constructing all-solid-state lithium batteries which are workable under high operating temperatures. The assembled battery was cycled at 60 ℃ to evaluate the corresponding battery performances. As shown in Fig. 2e, it exhibits similar charge/discharge plateaus at elevated temperature as cycled under room temperature at the same rate (0.1 C), while delivering a much higher initial discharge capacity (206 mAh/g). The higher discharge capacity is related to the smaller internal resistance of the battery and the faster lithium ion migration at high temperatures [21]. After 50 cycles, it sustains 77% of its initial discharge capacity and shows a discharge capacity of 158 mAh/g (Fig. 2f). Although the battery delivers much higher discharge capacities at elevated operating temperatures, it suffers a faster capacity degradation. Additionally, the assembled battery delivers higher discharge capacities at the same charge/discharge rates at 60 ℃ than operating at room temperature. Finally, this battery was also cycled under 0 ℃ at 0.1 C to investigate the electrochemical performances, as shown in Fig. S3a (Supporting information). It shows an initial specific discharge capacity of only 106 mAh/g with a low initial coulomb efficiency of 57% and suffers a significant discharge capacity decay in the subsequent cycles (Fig. S3b in Supporting information). Large polarization is observed in the initial charge/discharge profiles due to the low Li-ion conductivity of the 78Li₂S-22P₂S₅ glass-ceramic electrolyte at this operating temperature.

To unravel the working mechanism of the assembled solid-state battery at different operating temperatures, multiple characterization methods were performed. Li-ion kinetics plays a key role in electrochemical performance. To study the Li⁺ kinetics behaviors of the LiNi_0.6Mn_0.2Co_0.2O₂/g-c-225 ℃/Li-In battery at different temperatures, in-situ EIS was conducted. As shown in Fig. S4 (Supporting information), to evaluate the obtained impedance spectra, the impedance spectra can be divided into four parts according to the equivalent circuit model, including the bulk phase impedance of the electrolyte (R_{SE, bulk}), the grain boundary impedance of the electrolyte (R_{SE, gb}), the positive electrode/electrolyte interface impedance (R_SE/cathode) and the negative electrode/electrolyte interface impedance (R_SE/anode) [11]. As shown in Fig. 3, the changes in impedance spectrum under different charging and discharging states are carefully studied. During the initial charging process at room temperature, the total impedance increases with increasing charging voltages. In comparison, the total resistance first decreases slowly from 3.5 V to 3.1 V and then increases from 3.0 V to 2.8 V during the subsequent discharging process. (Figs. 3a and b) Moreover, in-situ EIS tests were also carried out on the battery when operated at 60 ℃. The total resistance of the assembled battery shows a similar change trend during the initial charging/discharging processes with much smaller resistance values at this temperature (Figs. 3c and d). To further understand the reason for those impedance variations, the corresponding impedance contribution and variation trend of each part obtained by fitting those impedance spectrums are shown in Fig. S5 (Supporting information). The R_{SE, bulk} and R_{SE, gb} impedance corresponding to the high frequency part basically remains at different states of charge (SOC) at room temperature, which proves that the solid electrolyte has good stability at room temperature. In contrast, when cycled at 60 ℃, the high-frequency part mainly comes from the R_{SE, bulk} part, while R_{SE, gb} can be almost ignored, and the changing trend is consistent with that at room temperature, indicating that the electrolyte still has good stability at 60 ℃. The resistance of R_SE/cathode increases with the depth of charge/discharge due to the accumulation of resistance caused by the side reactions between solid electrolytes/cathode materials and the local solid-solid contact failure caused by the volume change of electrode materials during cycling [22]. In contrast, R_SE/anode increases sharply under high discharge depth because of using lithium indium as the dynamic barrier of anode [23]. At 60 ℃, R_SE/cathode and R_SE/anode show a similar changing trend with significantly lower values due to the faster Li-ion mitigation at this operating temperature.

Compared with the traditional equivalent circuit model method, the relaxation time distribution (DRT) method can provide information about the change of impedance spectra of each part more intuitively and realistically [24]. As shown in Fig. 3e, during the charging process, a fully overlapping peak appears between 10⁻⁷ s and 10⁻⁶ s, corresponding to the R_{SE, bulk}. Meanwhile, the same overlapping peak between 10⁻⁵ s and 10⁻⁴ s is mainly the contribution of the R_{SE, gb}. These two parts are related to the process of conduction, and the conduction-based process is much faster, which is related to rapid relaxation [25]. The good overlap shows that this part is stable, which is in good agreement with the previous results based on the equivalent circuit model. An obvious increase in the intensity is observed for the peak between 10⁻³ s and 10⁻² s, which corresponds to the change of R_SE/cathode. Whereas the increase of the peak between 10⁻² s and 10⁰ s due to the slight rise of R_SE/anode is also observed. Moreover, an additional peak appears between 10⁰ and 10¹ s, which is mainly the evolution of the feedback diffusion impedance [26]. The intensity distribution plots based on the DRT results are displayed in Fig. 3f as to reveal the contribution of impedance in each frequency band during the charging more visually. As shown in Figs. 3g and h, a similar peak appears during the discharge process. When discharging to a lower potential, R_SE/cathode, R_SE/anode and diffusion impedance increase, which agrees well with the previous impedance results. As shown in Figs. 3i-l, significant differences are observed in the DRT results when the battery was cycled at 60 ℃. Only one overlapping peak is detected in the time constant between 10⁻⁷ s and 10⁻⁵ s, which is attributed to the disappearance of grain boundary impedance at high temperatures. Moreover, the time constants of R_SE/cathode and R_SE/anode increase from 10⁻³–10⁰ s to 10⁻⁴–10⁰ s due to the formation of solid electrolyte interface (SEI) and cathode–electrolyte interface (CEI) films. The increase of ionic conductivity at high temperatures has a significant effect on accelerating the above process [26].

More kinetic characterization was performed to clarify the degradation of electrochemical performances under different operating temperatures. Firstly, the differential capacitance versus voltage (dQ/dV) curves of the assembled ASSLBs worked under different operating temperatures were investigated. As shown in Figs. 4a and b, clear oxidation and reduction peaks are observed at around 3.1 V (vs. Li-In) at room temperature. The oxidation peak shifts towards higher voltages in the subsequent cycles, while the reduction peak moves towards lower voltages. During 50 cycles, those assembled ASSLBs show much smaller potential differences between the oxidation and reduction dQ/dV peaks at room temperature than that at 60 ℃, indicating higher reversibility and smaller polarizations. The battery delivers a much larger polarization when the operating temperature lowers to 0 ℃, which agrees well with the previous electrochemical performances (Fig. S6 in Supporting information). In addition, cyclic voltammetry (CV) was also used to study the kinetics of lithium ions in the battery when cycled at different operating temperatures. The lithium-ion mobility was roughly estimated by CV tests at different scan rates. As shown in Fig. S7 (Supporting information), the battery delivers higher lithium-ion transference numbers at elevated temperatures, which explains the higher initial discharge specific capacity at 60 ℃ in comparison with room temperature. This result corresponds to the ionic conductivity of the electrolyte at different temperatures. (Fig. S8 in Supporting information) The solid-solid contact between the active material and solid electrolyte in the cathode mixture also plays a crucial role in the electrochemical performance of all-solid-state batteries. To further investigate the volume changes during the cycling of the assembled solid-state battery at different operating temperatures, in-situ stack pressure measurements were performed when the battery cycled at 0.5 C at the corresponding temperatures. As shown in Figs. 4c and d, the stacking stress of the assembled battery increases during the initial charging process and decreases during the subsequent discharging process due to the volume changes of the electrodes during cycling [27]. At higher operating temperatures (60 ℃), the battery shows larger stress variations than that at room temperature, suggesting it suffers more serve volume changes at elevated temperatures. The solid-solid interfacial contacts between solid electrolytes and active materials are significantly affected by the volume variation during cycling. Large volume changes due to the intercalation/deintercalation of Li-ions during cycling lead to the loss of efficient contact, yielding inferior cycling performance, which agrees well with the fast degradation of discharge capacity at 60 ℃.

In conclusion, 78Li₂S-22P₂S₅ glass-ceramic electrolytes with high room temperature Li-ion conductivity up to 1.06 × 10^–3 S/cm have been prepared by carefully tailoring the heat treatment temperatures during the annealing process. Due to the ultrafast conductivity of the targeted solid electrolyte, the corresponding ASSLBs combined NCM622 cathode and Li-In alloy anode have been constructed and exhibit excellent electrochemical performances. The ASSLBs deliver a high initial discharge capacity of 149.0 mAh/g at room temperature when cycled at 0.1 C under room temperature and sustain 87.0% of the discharge capacity after 50 cycles. When the operating temperature rises to 60 ℃, it delivers a much higher initial discharge capacity of 206 mAh/g and lower capacity retention (77.0%) after 50 cycles at the same condition. Li-ion kinetics results confirm that the higher discharge capacity is associated with fast Li-ion mobility in the cathode mixture at elevated temperatures. While the faster degradation of discharge capacity at 60 ℃ is due to the larger volume changes during cycling at high operating temperature based on the in-situ stacking stress test results. This work provides a crucial role in volume changes and an interfacial section for the electrochemical performances of all-solid-state batteries in a wide operating temperature range.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

This work was supported by the National Key Research and Development Program (No. 2021YFB2400300) and the National Natural Science Foundation of China (No. 52177214). We gratefully acknowledge the Analytical and Testing Center of HUST for the technical support.

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2022.107859.