Surface Degeneration of Li₃PS₄–LiI Glass-Ceramic Electrolyte by Exposure to Humidity-Controlled Air and Its Recovery by Thermal Treatment

Abstract

Sulfide-based solid electrolytes are desirable for use in all-solid-state batteries owing to their high ionic conductivity and plasticity. However, they generally degrade upon exposure to water and can generate toxic hydrogen sulfide even in dry-room atmospheres. To prevent their degradation, surface stabilization is required and further research into the degradation mechanism is necessary. In the present study, the stability of Li₃PS₄–LiI glass ceramic (LPSI) has been examined under low-humidity conditions. In contrast to an argyrodite-type solid electrolyte, exposure of LPSI to dry air with a dew point of −20 °C resulted in low H₂S-gas generation and reduced ionic conductivity of LPSI. Since the conductivity mostly recovered after vacuum heating at 100 °C, the H₂S generation is not considered to be the major reason for the reduction in conductivity. On the contrary, it is suggested that water molecules are present on the LPSI powder particles after dry-air exposure, resulting in the formation of a degraded LPSI layer and low ionic conductivity, and that most of the water molecules are removed during vacuum heating, resulting in the recovery of conductivity. Furthermore, optimal vacuum-heating conditions were obtained from X-ray diffraction and temperature-programmed desorption-mass spectrometry measurements, indicating an optimal temperature and heating time of 100 °C and 2 h, respectively. Impedance measurements were used to probe the degradation of the surface layer. The condition of the surface layer was affected by the pellet-forming pressure, and it was easier to detect the degradation of the surface layer when the pellets were formed at low pressures. This paper contributes to the formulation of guidelines for the development of water-resistant solid electrolytes.

1. Introduction

As one of the post-lithium-ion batteries, sulfide all-solid-state batteries have attracted increasing attention.^1–3 All-solid-state sulfide batteries have high energy density and durability; nevertheless, research is actively being conducted to achieve further improvements in these parameters.^4–15 The solid electrolyte is one of the main components that contributes to the high performance of sulfide all-solid-state batteries. Solid sulfide-based electrolytes have high ionic conductivity and plasticity.^1,2 However, their low resistance against moisture is problematic, particularly for battery fabrication, and their exposure to moisture even in a dry room environment leads to the generation of toxic hydrogen sulfide (H₂S) gas. For example, argyrodite-type electrolytes such as Li₆PS₅Cl are known to exhibit high ionic conductivity at room temperature but generate H₂S in dry air at a dew point (dp) of −20 °C, leading to the formation of Li₃PO₄, Li₂CO₃, Li₂SO₄, and LiCl on the surface.^16–20 In addition, the reduction of ionic conductivity to 35 % of the original value was reported after dry-air contact for only 1 h. To stabilize sulfide-based solid electrolytes against moisture, some researchers have examined the replacement of P with Sn, As, and Sb based on the hard and soft acids and bases (HSAB) theory.^21–25

The group of Tatsumisago investigated the stability of Li₂S–P₂S₅ sulfide in the ambient atmosphere and concluded that 75Li₂S·25P₂S₅ (Li₃PS₄) glass and glass-ceramic are fairly stable against moisture:²⁶ They discussed H₂S generation in sulfide-based solid electrolytes in terms of their composition and the S-containing structural unit, that is, S²⁻, P₂S₇, or PS₄, and concluded that the PS₄ unit is more resistant to attack from H₂O than the P₂S₇ and S²⁻ units. From this perspective, we focused on the Li₃PS₄ glass-ceramic electrolyte, which has a relatively stable PS₄ unit in its structure. They also found that the addition of third components such as Li₂O, ZnO, and FeS to Li₃PS₄ glass considerably suppressed H₂S generation.^7,27–29 The effect of LiI addition to Li₃PS₄ glass on moisture resistance has not been fully investigated;^30–32 however, Calpa et al. reported that the Li₄PS₄I (Li₃PS₄·LiI) crystal is highly stable against moisture in comparison with the β-Li₃PS₄ crystal.³³ Hence, the Li₃PS₄–LiI glass-ceramic was selected in this study, and its stability was investigated. Although the degeneration of the electrolytes under dry-room conditions is an important factor in the fabrication of all-solid-state batteries, few papers have discussed this phenomenon under low-humidity conditions in detail.^16–20 In this study, we focused on the degeneration of the electrolyte including the ionic conductivity upon exposure to air with a dew point similar to that in a dry room. Some researchers reported the recovery of the conductivity by the subsequent vacuum heating after exposure to ambient air, but the details of this process are unknown.^16–21 Thus, we also examined the recovery of the conductivity by vacuum heating after exposure to dry air.

2. Experimental

A sulfide glass-ceramic electrolyte (LPSI), Li₃PS₄·xLiI (x ≈ 0.2), was provided by Idemitsu Kosan (Japan) in powder form (D₅₀ ≈ 3.5 µm; BET surface area, 5.2 m² g⁻¹). For comparison purposes, an argyrodite-type sulfide electrolyte powder, Li₆PS₅Cl (LPSCl, D₅₀ ≈ 10 µm), from MSE Supplies (USA) was also examined. The electrolyte samples were exposed to humidity-controlled air according to the procedures in previous studies,^16,19,20,34 as follows. The humidity of air was controlled at a dew point of −20 °C-dp with an accuracy of ±2 °C-dp in a dry-air glove box (Seibu Giken, Japan). Dry air was pumped out of the glove box and flowed onto the quartz-wool-wrapped powder sample at a rate of 0.8 dm³ min⁻¹. H₂S concentration was measured downstream of the sample with an H₂S sensor (3000RS, Advanced Micro Instruments, USA) every 1 s. Vacuum-heating of the sample was carried out in a glass-tube oven.

The electrochemical impedance (Z) of LPSI and LPSCl was measured at 22 °C in the frequency range from 7 MHz to 1 Hz with an amplitude of 10 mV using a potentiostat/galvanostat/impedance analyzer (SP-200, Advanced Model, BioLogic, France). The sample powder (usually 0.1 g) was sandwiched between stainless steel electrodes and pelletized at various pressures ranging from 73 to 436 mPa. The pelleting of the samples for conductivity measurements was done after the samples had been exposed to moisture in powder form or after they had been exposed to moisture and vacuum-heated in powder form. The normalized impedance (Z_norm) was calculated using the equation Z_norm = (Z × S)/d, where S is the cross-sectional area of the pellet (1 cm²) and d is the thickness (usually ∼0.5 mm).

X-ray diffraction (XRD) of the electrolyte powder was measured using a Rigaku (Japan) Smart Lab diffractometer (MoKα, 45 kV and 200 mA) equipped with a temperature-control unit. The sample was sealed in a quartz-glass capillary in a dry-air glove box with a dew point of less than −80 °C-dp. When heating the sample during XRD measurements, the furnace temperature was raised at a heating rate of 5 °C min⁻¹, and the patterns were recorded with an integration time of 5 min after waiting for 5 min every 10 °C from 50 to 300 °C.

Transmission electron microscopy (TEM) analysis was performed using a JEM-2100F field-emission TEM (JEOL, Japan) equipped with a complementary metal–oxide–semiconductor camera (OneView, Gatan) and a vacuum transfer holder (model 648, Gatan, USA) according to the methods in the literature.^18,35,36 The sample powder was dispersed on a carbon grid in an Ar-filled glove box. Hollow-cone dark-field (HCDF) images were obtained for the observation of the crystal size and morphology. To identify the sample in detail, the electron diffraction (ED) patterns were also obtained and analyzed with “ProcessDiffraction” software.³⁷

X-ray photoelectron spectra (XPS) were recorded with an ESCA-3400HSE (Shimadzu, Japan). The X-ray source was MgKα radiation and the binding energy was corrected with reference to the C1s line at 284.6 eV attributed to the hydrocarbon moiety. The O/P surface molar ratio was calculated from the peak areas using the atomic sensitivity factors of O 1s (0.63) and P 2p (0.25).³⁸ The sample kept in an Ar-filled glove box was transferred to the XPS chamber in a dry room at ca. −20 °C-dp within 1 min. To remove the surface adsorbates, Ar-ion sputtering was carried out at 1 kV and 20 mA under a pressure of 8 × 10⁻⁴ Pa, while the sputter ratio at that condition for SiO₂ was to be about 0.5–0.9 nm min⁻¹.

Temperature-programmed desorption/decomposition mass spectrometry (TPD-MS) was performed at a heating rate of 10 °C min⁻¹ for temperatures of up to 500 °C. The gas was analyzed with a GC/MS QP2020Plus(10) (Shimadzu). The sample in the Ar-filled glove box was transferred to the TPD chamber without air exposure. In this study, in addition to the usual TPD-MS, in which the temperature was gradually increased as mentioned above, isothermal TPD-MS was performed, in which the operating temperature reached a defined value and was then held constant. In one experiment, the sample was heated up to 100 °C and held for 5 h at 100 °C.

3. Results and Discussion

3.1 Generation of H₂S with moisture

Hydrogen sulfide was generated from LPSI through contact with dry air at −20 °C-dp, as shown in Fig. 1. The total amount of H₂S generated was evaluated as 0.01 mmol g⁻¹ (0.3 cm³ g⁻¹) for the first 1 h. No generation of H₂S was detected from LPSI in contact with dry air at −50 °C-dp, where the detection limit was 0.0001 mmol g⁻¹. However, the amount of H₂S from LPSCl was as large as 0.4 mmol g⁻¹ (8 cm³ g⁻¹) when in air at −20 °C-dp for the first 1 h, showing that the reactivity of LPSI to moisture is very low in comparison with the argyrodite type electrolyte. This difference is thought to be caused by the difference in the stability of the S atoms: stable PS₄ units in LPSI and unstable S²⁻ units in LPSCl. Gas generation may also be suppressed to some extent with the addition of LiI to LPS.

Figure 1.

Accumulated amount of H₂S generated by a sample exposed to air stream with −20 °C dew point (dp). (red) LPSI glass ceramic. (black) Argyrodite-type crystal.

The sulfur content of LPSI is calculated as 19 mmol g⁻¹ from the compositional formula of Li₃PS₄·0.2LiI. Hence, only 0.05 % of sulfur in the electrolyte is reacted with dry air at −20 °C-dp to form H₂S for 1 h, indicating that the reaction occurs only on the surface. In the case of LPSCl, the sulfur content is also 19 mmol g⁻¹, and the amount of sulfur that reacted with water is attributed to the 2 % of sulfur in the argyrodite-type electrolyte, where it is also thought that only the surface of the electrolyte reacted with dry air.¹⁶ Tsukasaki et al. showed the formation of Li₂CO₃, Li₂SO₄, Li₃PO₄, and LiCl on the surface of the argyrodite-type electrolyte exposed to dry air of −20 °C-dp for 24 h, and they discussed the reactions occurring on Li₆PS₅Cl with H₂O, CO₂, and O₂ in the dry air.¹⁸

3.2 Conductivity of LPSI

Figure 2 shows the impedances in a Nyquist plot of the pellets of the LPSI solid electrolyte powder before and after exposure to dry air at −20 °C dp for 1 h with several pellet compaction pressures. The Nyquist plots for LPSI show the typical ion-blocking behavior of the impedance. The increase in the conductivity with an increase in the applied pressure is probably caused by a larger contact area of the electrolyte particle in the pellet.³⁹ The conductivity at 436 MPa was 1.9 × 10⁻³ S cm⁻². Semicircular arc attributed to the bulk conduction was not observed in the plots because the time constant for the bulk conduction of the sample was higher than the highest frequency of the measurement (1 MHz).

Figure 2.

Electrochemical impedance of an LPSI solid electrolyte (a) before and (b) after exposure to dry air of −20 °C dp. The pelletization pressure was increased from 73 to 436 MPa. Effects of pelletization pressure on thickness and conductivity of the LPSI solid electrolyte (c) before and (d) after exposure to dry air of −20 °C dp for 1 h. Thickness was shown in its relative value to that pressed at 73 MPa.

The generated amount of H₂S was very small after exposure to dry air at −20 °C-dp for 1 h for LPSI compared to LPSCl.¹⁶ However the conductivity considerably decreased to 7.1 × 10⁻⁴ S cm⁻¹ at 436 MPa after exposure (Figs. 2b, 2d). A semi-circular arc appeared at the high-frequency side of the Nyquist plots after air exposure (Fig. 2b). When increasing the applied pressure, the arc became ambiguous. The appearance of the semicircular arc suggests a significant increase in the resistance at the particle surface due to the exposure to moisture. This semicircular arc decreases with increasing pressure, which may imply that the low-ionic-conductivity phase formed on the surface was being crushed by the press, but the details are unknown. In any case, this low-ionic-conductivity phase appears to be detectable only when the pellets are pressed at a low pressure, such as 73 MPa, and difficult to detect when they are molded at much higher pressures, such as 436 MPa.

Zhao et al. reported that the conductivity of Sn-substituted Li₆PS₅I after exposure to air can be mostly recovered by vacuum-heating at 180 °C.⁴⁰ We also confirmed that the conductivity of Li₆PS₅Cl exposed to dry air at −20 °C-dp for 1 h recovers to 52 % of the original conductivity by vacuum-heating at 170 °C for 12 h, as reported in a previous paper.¹⁶ Hence, we examined the effect of the vacuum-heating on the conductivity of LPSI. Since the structural change of LPSI was detected by heating at 110 °C or above, as discussed later, the degraded sample was vacuum-heated at 100 °C for 2 h. The time period for the vacuum heating process was determined by the TPD-MS results presented in the following section. Using this treatment, the reduced conductivity obtained after the air exposure was mostly recovered. Figure 3 shows the impedances of the pelletized powder before exposure, after exposure, and after vacuum heating at 73 MPa. Here, the pelletization pressure was selected to be 73 MPa because the semicircular arc in the high-frequency region, which is thought to originate from the degraded grain boundary (particle surface) resistance, is apparent at this pressure. The low-ionic-conductivity phase that appears after air exposure disappears after heat treatment. In addition, the bulk conductivity after the heat treatment appeared to be slightly lower than that of the pristine pellet, which may be due to the change in the bulk crystalline state caused by the heat treatment. Only slight changes in the crystalline state were observed in the XRD measurement (discussed later in reference to Figs. 4a, 4c). The initial conductivity was 1.2 mS cm⁻¹ (100 %) but decreased to 0.45 mS cm⁻¹ (37 %) after exposure to dry air, and then increased to 0.94 mS cm⁻¹ (79 %) after the vacuum heat treatment. In the present LPSI case, the conductivity retention after exposure was similar to that of Li₆PS₅Cl (argyrodite, 36 %)¹⁶ but the conductivity retention after vacuum heating was much higher than that of Li₆PS₅Cl.¹⁶ This high recovery ratio is noteworthy. It is important to note that even though only a very small amount of H₂S is generated, the conductivity clearly decreases, and a low-ionic-conductivity phase appears to be formed at the grain boundary. The large recovery observed by vacuum heating compared to the case of argyrodite SE may suggest that the low-ionic-conductivity phase reversibly incorporates H₂O. In fact, some of its hydrates have been reported in some solid electrolytes,^41,42 including some Na-based sulfides^22,23,43,44 and Li₄SnS₄,^45,46 so it is possible that a similar phase is formed. If such a phase is stable in an ultra-high vacuum, it can be analyzed using several tools, including TEM.

Figure 3.

Electrochemical impedance in Nyquist plots of the LPSI solid electrolyte (black) before and (blue) after exposure to dry air of −20 °C dp, and (red) after the subsequent vacuum heating at 100 °C for 2 h. The pelletization pressure was 73 MPa.

Figure 4.

XRD patterns for LPSI. (a) As received, (b) exposed to dry air at −20 °C-dp for 1 h, (c) after vacuum heating at 100 °C for 2 h following the dry-air exposure for 1 h, and (d) exposed to dry air at −20 °C dp for 24 h.

3.3 X-ray diffraction of LPSI

Figure 4a shows the XRD pattern of LPSI. No significant change in the pattern was observed after the exposure to dry air at −20 °C-dp for 1 h (Fig. 4b) or after the vacuum-heating at 100 °C for 2 h following the exposure (Fig. 4c), indicating that the bulk of the powder is not affected by the 1-h exposure and subsequent vacuum drying. However, there is a very slight change in the width of the peak at around 9°, which may correspond to the slight change in bulk conductivity discussed in section 3.2.

The XRD pattern for LPSI changed slightly after the heating in the shoulder peak at 12.8° at around 120 °C, and the change became obvious at approximately 150 °C (Fig. 5a). The pattern at 300 °C can be identified as that of Li₄PS₄I (ICSD #432169). The growth of the peak at 12.8° suggests crystallization of the Li₄PS₄I species originally present in LPSI (the crystallite size is 11 nm at 25 °C, as determined from the width of the peak at 9.3°, while that at 300 °C is 30 nm).

Figure 5.

Thermal changes in XRD patterns of LPSI before and after the exposure of dry air at −20 °C-dp for 1 h. The XRD patterns were measured at 25 °C, and every 10 °C from 50 to 300 °C in a sealed glass capillary. The temperature rise rate was 5 °C min⁻¹, and XRD scans were performed for 5 min of integration time after a 5-min thermal equilibration (waiting) time at each step.

By heating LPSI that was exposed to dry air for 1 h, the crystallization became significant at approximately 110 °C (Fig. 5b). Thus, when recovering a degraded sample by heating, the temperature of the recovery process was selected to be below 100 °C. Notably, the measurement was carried out in a closed system under different conditions from those of vacuum heating. Under vacuum heating conditions, water or some gas molecules may be removed from the system. Since water was included in the sample by the dry-air exposure (as discussed later), the presence of water should enhance earlier crystallization compared with the as-received sample. The peaks in the pattern for LPSI at 25 °C (see Fig. 4a) were diminished after the dry-air exposure for 24 h and new unknown peaks appeared (Fig. 4d), showing the decomposition of the Li₄PS₄I species. Calpa et al. discussed the decomposition of Li₄PS₄I crystallite with moisture to Li₃PS₄ and LiI hydrates.³³ The phenomena of easy hydration of LiI even in sulfide solid electrolytes were also reported by Matsuda et al. for 60Li₃SbS₄·40LiI.⁴⁷

3.4 Surface morphology of LPSI with analytical TEM

HCDF images for LPSI are shown in Fig. 6. Debye–Scherrer rings and diffraction spots were observed in the ED pattern for the image region (insets in Fig. 6). The ED intensity profile calculated from the ED pattern showed the presence of diffraction peaks attributed to crystalline LiI. The diffraction of LiI simulated from the XRD pattern (ICSD #01-071-4666) is also shown in the right half of Fig. 6a. The intensity profile corresponding to the XRD pattern of the sample was simulated from that in Fig. 4a. The broad diffraction peak at 14 nm⁻¹ corresponds to the XRD peak at 9.3° in Fig. 4a. The ED pattern also shows Debye-Scherrer rings, which are considered to originate from LPSI, indicating that the LPSI bulk is in an amorphous state. The location that causes the LiI diffraction is shown as bright spots in the HCDF image, and it was found that LiI nanoparticles are present on the surface of LPSI. However, the peak attributed to LiI (311) at 22.4° (35 nm⁻¹ in scattering vector) was undetectable in the XRD pattern of LPSI (see Fig. 4a), implying that the crystal LiI content in the sample was very low.

Figure 6.

TEM observation results for LPSI. (a) As received, (b) after the exposure to dry air at −20 °C-dp for 1 h, and (c) after vacuum-heating at 100 °C for 2 h following the dry-air exposure for 1 h. HCDF images are shown on the left, together with their corresponding ED patterns in the respective insets. The bright contrast area in the HCDF images indicates the presence of LiI crystallites. ED intensity profiles are shown on the right (a) with LiI pattern calculated from ICSD #01-071-4666 and XRD of LPSI glass ceramic converted into intensity profiles with respect to the scattering vector. The HCDF images were obtained from the Debye–Scherrer rings and spots of the ED patterns.

After exposure to dry air followed by vacuum heating, no significant change was evident in the presence or particle size of the LiI nanocrystal or the crystallographic state of the LPSI bulk. In the argyrodite-type LPSCl sulfide solid electrolytes investigated in a previous study,¹⁸ ED patterns originating from decomposition products such as LiCl, Li₂CO₃, Li₃PO₄, and Li₂SO₄ were observed after exposure to dry air at −20 °C dp, whereas no such decomposition-product-derived ED patterns were observed in the LPSI samples after exposure to dry air at −20 °C dp in this study.

The lack of evident changes due to exposure to dry air measured using the TEM may be due to the return of the material to its original state after the vacuum-evacuation in the TEM measurements. However, the fact that the samples after vacuum heating were similar to those before exposure indicates that the material may return to its original state by vacuum heating, even after exposure to moisture. The absence of decomposition products could mean that the moisture did not decompose the solid electrolyte, but rather that water was simply incorporated into the solid electrolyte. This is supported by the fact that water was detected in the post-exposure samples by TPD-MS. The fact that the PS₄ unit in LPSI is chemically stable enough not to produce H₂S when attacked by water suggests the possibility of recovery by vacuum heating.

3.5 Surface analysis of LPSI by XPS

The surface properties of LPSI were analyzed by XPS. After Ar-ion sputtering for 30 and 60 s, the peak intensity of P 2p was discernibly increased, but the binding energy was almost the same as that without the sputtering. In the case of LPSI exposed to dry air for 1 h, the peak intensity of P 2p was considerably lower than that for the as-received LPSI. The peak intensity was significantly increased after the sputtering.

The O 1s peak for LPSI exposed to dry air for 1 h was considerably greater than that for LPSI without the exposure and the peak decreased after the sputtering, suggesting accumulation of adsorbates such as water on the surface of LPSI during the dry-air exposure. This is consistent with the result that the intensity of the P 2p peak increased with the decrease in that of the O 1s peak, because the surface adsorbates will disturb photoelectron emission from the surface of LPSI particles.

The O/P atomic ratio for LPSI exposed to dry air for 1 h was significantly higher than that for LPSI without the exposure (Fig. 7). The ratio was considerably decreased by the weak sputtering for 30 s, showing that the high O/P ratio is caused by the presence of surface adsorbate accumulated on LPSI during the dry-air exposure for 1 h. The reduction of the high O/P ratio for LPSI without the exposure after sputtering for 30 s suggests either that the surface adsorbate also exists on the as-received sample or that the surface was slightly oxidized during sample transfer from the glovebox to the XPS chamber. The O/P ratio for dry-air-exposed LPSI after sputtering for 60 s was similar to that after sputtering for 30 s, suggesting that the removal of the surface adsorbate by the sputtering is mostly completed within 30 s. In the case of LPSI without the exposure, a similar tendency was observed.

Figure 7.

Atomic surface ratio of O/P for LPSI with (square) and without (circle) exposure to dry air at −20 °C-dp for 1 h.

3.6 Temperature-programmed desorption from LPSI

To analyze the surface adsorbates on LPSI, TPD-MS was carried out. Desorption of H₂O was detected from LPSI before exposure to dry air. The desorption started at approximately 50 °C and ceased at 210 °C with heating at a rate of 10 °C min⁻¹, and the desorbed amount was 0.21 mmol g⁻¹ (0.38 wt%) (Fig. 8a). A small desorption of H₂S was detected at 100–500 °C. The desorbed amount of H₂S was 0.12 mmol g⁻¹ (0.40 wt%).

Figure 8.

Temperature-programmed desorption from LPSI, where temperature increase to 500 °C at a rate of 10 °C min⁻¹. (a) H₂O desorption from LPSI with (red) and without (black) the exposure of dry air at −20 °C-dp for 1 h at the heating rate of 10 °C min⁻¹. The total amount of H₂O desorption was 0.84 wt% (0.47 mmol g⁻¹) and 0.38 wt% (0.21 mmol g⁻¹). (b) Isothermal desorption at 100 °C from LPSI exposed to dry air for 1 h (red). The temperature is also displayed (black). The total amount of H₂O desorption was 0.60 wt% (0.33 mmol g⁻¹).

The H₂O desorption from LPSI after the exposure to dry air for 1 h also occurred in the same temperature range and the desorbed amount was 0.47 mmol g⁻¹ (0.84 wt%). From the dry-air-exposed sample, 0.33 mmol g⁻¹ (0.60 wt%) of water was desorbed by isothermal heating at 100 °C within 2 h (Fig. 8b). Hence, 0.14 mmol g⁻¹ (0.24 wt%) of water must be present in the LPSI evacuated at 100 °C for 2 h after exposure to dry air. This amount of water is attributed to a two-layer water film on the surface on the basis of the cross-sectional area of water (0.105 nm²)⁴⁸ and the BET surface area of LPSI. In the case of the as-received sample, the film thickness is three layers (0.21 mmol g⁻¹), and in the case of LPSI exposed to dry air for 1 h, the film thickness is six layers (0.47 mmol g⁻¹). It can also be assumed that some type of hydrated state of the solid electrolyte is formed on the surface of LPSI particles, and that this state has a relatively low ionic conductivity compared with the original state. Hydrated states were also reported for the sodium system of a sulfide solid electrolyte (Na₃SbS₄).⁴⁴

Hence, the removal of water from LPSI by heat evacuation should recover the ionic conductivity of the sample, and in fact, after the heat evacuation the conductivity was partially recovered. However, a close examination shows that the conductivity is not fully recovered even though the water content of the sample after evacuation is slightly lower than that of the as-received sample (Figs. 2b and 3b). In the XPS measurements, the adsorbate detected after the sputtering is probably the surface adsorbate that strongly interacts with the LPSI surface. Since the O/P ratio for the dry-air-exposed LPSI after the sputtering is higher than that of the as-received LPSI without the sputtering, the amount of water strongly adsorbed at the surface of dry-air-exposed LPSI should be greater than that of the latter sample. Hence, the presence of water with strong surface interactions may lower the conductivity. The generation of H₂S with moisture leads to the surface oxidation and therefore lowers the conductivity. In this study, a very low amount of H₂S was detected, implying a low rate of oxidation. No significant change was observed by TEM after dry-air exposure (see Fig. 6). Thus, it is thought that the adsorption of water, and hence the hydrated state formation on the surface is the major reason for the conductivity reduction in LPSI.

3.7 Optimization of the conditions for vacuum heating recovery

To optimize the vacuum heat treatment conditions, we conducted several tests, which we believe can also be used in optimizing vacuum heating recovery conditions for other solid electrolytes. The upper temperature limit for the recovery of the exposed sample can be optimized based on the temperature-rise XRD results. To perform this optimization, we investigated the upper temperature limit at which the XRD pattern stops showing detectable changes. In this case, we found that 100 °C was the upper temperature limit. Based on the results of TPD-MS at a fixed temperature, the retention time at a given temperature can also be optimized. To perform this optimization, we examined the time required to achieve a near-zero rate of water release, which was determined to be 2 h. However, the ordinary TPD-MS results also demonstrate that water cannot be completely extracted when the heat treatment temperature limit is set to 100 °C. It has been found that heating the sample with absorbed water to high temperatures can lead to adverse changes. Heat treatment protocols that start at low temperatures and then gradually increase the temperature while releasing water may be more suitable.

Acknowledgments

This study was conducted under the project of “Development of Fundamental Technologies for All Solid State Battery applied to Electric Vehicles (SOLiD-EV, JPNP18003)” commissioned by the New Energy and Industrial Technology Development Organization (NEDO).

CRediT Authorship Contribution Statement

Hikaru Sano: Data curation (Lead), Formal analysis (Lead), Investigation (Lead), Methodology (Lead), Validation (Lead), Visualization (Lead), Writing – original draft (Lead), Writing – review & editing (Lead)

Yusuke Morino: Data curation (Equal), Formal analysis (Equal), Investigation (Equal), Methodology (Equal), Validation (Equal), Visualization (Equal), Writing – review & editing (Equal)

Yasuyuki Matsumura: Data curation (Equal), Formal analysis (Equal), Investigation (Equal), Writing – review & editing (Equal)

Koji Kawamoto: Conceptualization (Equal), Data curation (Equal), Funding acquisition (Equal), Supervision (Lead), Visualization (Equal)

Hiroyuki Higuchi: Data curation (Equal), Formal analysis (Equal)

Noriyuki Yamamoto: Data curation (Equal), Formal analysis (Equal)

Atsunori Matsuda: Data curation (Equal), Formal analysis (Equal), Investigation (Equal), Methodology (Equal), Validation (Equal), Visualization (Equal), Writing – review & editing (Equal)

Hirofumi Tsukasaki: Conceptualization (Equal), Data curation (Equal), Formal analysis (Equal), Investigation (Equal), Methodology (Equal), Visualization (Equal), Writing – original draft (Equal)

Shigeo Mori: Conceptualization (Equal), Data curation (Equal), Formal analysis (Equal), Investigation (Equal), Methodology (Equal), Visualization (Equal), Writing – original draft (Equal), Writing – review & editing (Equal)

Atsushi Sakuda: Conceptualization (Equal), Data curation (Equal), Formal analysis (Equal), Investigation (Equal), Methodology (Equal), Visualization (Equal)

Akitoshi Hayashi: Conceptualization (Equal), Data curation (Equal), Methodology (Equal)

Conflict of Interest

The authors declare no conflict of interest in the manuscript.

Funding

New Energy and Industrial Technology Development Organization: JPNP18003

Footnotes

H. Sano, Y. Morino, Y. Matsumura, K. Kawamoto, A. Matsuda, H. Tsukasaki, and A. Sakuda: ECSJ Active Members

A. Hayashi: ECSJ Fellow

References

• 1)  Y. Kato, S. Hori, T. Saito, K. Suzuki, M. Hirayama, A. Mitsui, M. Yonemura, H. Iba, and R. Kanno, Nat. Energy, 1, 16030 (2016).
• 2)  A. Sakuda, A. Hayashi, and M. Tatsumisago, Curr. Opin. Electrochem., 6, 108 (2017).
• 3)  K. Takada, J. Power Sources, 394, 74 (2018).
• 4)  M. Nagao, A. Hayashi, and M. Tatsumisago, Electrochemistry, 80, 734 (2012).
• 5)  N. Okuzawa, Y. Yamamura, N. Ishii, and H. Morimoto, Electrochemistry, 90, 027003 (2022).
• 6)  R. Harada, K. Aso, A. Hayashi, and M. Tatsumisago, Electrochemistry, 83, 898 (2015).
• 7)  T. Ohtomo, A. Hayashi, M. Tatsumisago, and K. Kawamoto, Electrochemistry, 81, 428 (2013).
• 8)  T. Hakari, M. Deguchi, K. Mitsuhara, T. Ohta, K. Saito, Y. Orikasa, Y. Uchimoto, Y. Kowada, A. Hayashi, and M. Tatsumisago, Chem. Mater., 29, 4768 (2017).
• 9)  Y. Sakka, H. Yamashige, A. Watanabe, A. Takeuchi, M. Uesugi, K. Uesugi, and Y. Orikasa, J. Mater. Chem. A, 10, 16602 (2022).
• 10)  A. Suzuki, H. Tanaka, H. Yamashige, Y. Orikasa, Y. Niida, T. Kimura, K. Tono, M. Yabashi, T. Ishikawa, Y. Bessho, Y. Joti, and Y. Nishino, Nano Lett., 22, 4603 (2022).
• 11)  Y. Morino and S. Kanada, J. Power Sources, 509, 230376 (2021).
• 12)  Y. Morino and S. Kanada, ACS Appl. Mater. Interfaces, 15, 2979 (2023).
• 13)  Y. Morino, H. Tsukasaki, and S. Mori, Electrochemistry, 90, 047003 (2022).
• 14)  Y. Morino, J. Power Sources, 541, 231672 (2022).
• 15)  Y. Morino, Electrochemistry, 90, 027001 (2022).
• 16)  H. Sano, Y. Morino, A. Yabuki, S. Sato, N. Itayama, Y. Matsumura, M. Iwasaki, M. Takehara, T. Abe, Y. Ishiguro, T. Takahashi, N. Miyashita, A. Sakuda, and A. Hayashi, Electrochemistry, 90, 037012 (2022).
• 17)  H. Tsukasaki, H. Sano, Y. Morino, and S. Mori, Microsc. Microanal., 28, 338 (2022).
• 18)  H. Tsukasaki, H. Sano, K. Igarashi, A. Wakui, T. Yaguchi, and S. Mori, J. Power Sources, 524, 231085 (2022).
• 19)  Y. Morino, H. Sano, K. Kawamoto, K. Fukui, M. Takeuchi, A. Sakuda, and A. Hayashi, Solid State Ionics, 392, 116162 (2023).
• 20)  Y. Morino, H. Sano, A. Shiota, K. Kawamoto, T. Takanashi, N. Miyashita, A. Sakuda, and A. Hayashi, Electrochemistry, 91, 037005 (2023).
• 21)  G. Sahu, Z. Lin, J. Li, Z. Liu, N. Dudney, and C. Liang, Energy Environ. Sci., 7, 1053 (2014).
• 22)  H. Wang, Y. Chen, Z. D. Hood, G. Sahu, A. S. Pandian, J. K. Keum, K. An, and C. Liang, Angew. Chem., Int. Ed., 55, 8551 (2016).
• 23)  T. W. Kim, K. H. Park, Y. E. Choi, J. Y. Lee, and Y. S. Jung, J. Mater. Chem. A, 6, 840 (2018).
• 24)  T. Kimura, A. Kato, C. Hotehama, A. Sakuda, A. Hayashi, and M. Tatsumisago, Solid State Ionics, 333, 45 (2019).
• 25)  K. Kanazawa, S. Yubuchi, C. Hotehama, M. Otoyama, S. Shimono, H. Ishibashi, Y. Kubota, A. Sakuda, A. Hayashi, and M. Tatsumisago, Inorg. Chem., 57, 9925 (2018).
• 26)  H. Muramatsu, A. Hayashi, T. Ohtomo, S. Hama, and M. Tatsumisago, Solid State Ionics, 182, 116 (2011).
• 27)  T. Ohtomo, A. Hayashi, M. Tatsumisago, and K. Kawamoto, J. Non-Cryst. Solids, 364, 57 (2013).
• 28)  A. Hayashi, H. Muramatsu, T. Ohtomo, S. Hama, and M. Tatsumisago, J. Alloys Compd., 591, 247 (2014).
• 29)  A. Hayashi, H. Muramatsu, T. Ohtomo, S. Hama, and M. Tatsumisago, J. Mater. Chem. A, 1, 6320 (2013).
• 30)  N. H. H. Phuc, E. Hirahara, K. Morikawa, H. Muto, and A. Matsuda, J. Power Sources, 365, 7 (2017).
• 31)  M. Suyama, A. Kato, A. Sakuda, A. Hayashi, and M. Tatsumisago, Electrochim. Acta, 286, 158 (2018).
• 32)  E. Rangasamy, Z. Liu, M. Gobet, K. Pilar, G. Sahu, W. Zhou, H. Wu, S. Greenbaum, and C. Liang, J. Am. Chem. Soc., 137, 1384 (2015).
• 33)  M. Calpa, N. C. Rosero-Navarro, A. Miura, R. Jalem, Y. Tateyama, and K. Tadanaga, Appl. Mater. Today, 22, 100918 (2021).
• 34)  J. Sato, K. Terai, and F. Utono, Japanese Pat., (published unexamined patent application) 075855 (2020).
• 35)  H. Tsukasaki, S. Mori, H. Morimoto, A. Hayashi, and M. Tatsumisago, Sci. Rep., 7, 4142 (2017).
• 36)  H. Tsukasaki, M. Otoyama, T. Kimura, S. Mori, A. Sakuda, A. Hayashi, and M. Tatsumisago, J. Power Sources, 479, 228827 (2020).
• 37)  J. L. Lábár, Ultramicroscopy, 103, 237 (2005).
• 38)  Handbook of X-ray Photoelectron Spectroscopy (Eds. C. D. Wagner, W. M. Riggs, L. E. Davis, J. F. Moulder, and G. E. Muilenberg), Perkin-Elmer Corp., Minnesota (1978).
• 39)  K. Otani, T. Muta, T. Furuta, T. Miyuki, T. Kaburagi, and G. Inoue, J. Energy Storage, 58, 106279 (2023).
• 40)  F. Zhao, J. Liang, C. Yu, Q. Sun, X. Li, K. Adair, C. Wang, Y. Zhao, S. Zhang, W. Li, S. Deng, R. Li, Y. Huang, H. Huang, L. Zhang, S. Zhao, S. Lu, and X. Sun, Adv. Energy Mater., 10, 1903422 (2020).
• 41)  X. Li, J. Liang, J. Luo, M. Norouzi Banis, C. Wang, W. Li, S. Deng, C. Yu, F. Zhao, Y. Hu, T.-K. Sham, L. Zhang, S. Zhao, S. Lu, H. Huang, R. Li, K. R. Adair, and X. Sun, Energy Environ. Sci., 12, 2665 (2019).
• 42)  X. Li, J. Liang, X. Yang, K. R. Adair, C. Wang, F. Zhao, and X. Sun, Energy Environ. Sci., 13, 1429 (2020).
• 43)  A. Banerjee, K. H. Park, J. W. Heo, Y. J. Nam, C. K. Moon, S. M. Oh, S. T. Hong, and Y. S. Jung, Angew. Chem., Int. Ed., 55, 9634 (2016).
• 44)  Y. Tian, Y. Sun, C. Hannah, W. Karena, S. Bo, Y. Tian, Y. Sun, D. C. Hannah, Y. Xiao, H. Liu, and K. W. Chapman, Joule, 3, 1037 (2019).
• 45)  T. Kimura, T. Nakano, A. Sakuda, M. Tatsumisago, and A. Hayashi, 22nd Kagaku Denchi Zair. Kenkyukai Meet., I–05 (2021).
• 46)  T. Kimura, T. Nakano, A. Sakuda, M. Tatsumisago, and A. Hayashi, J. Phys. Chem. C, 127, 1303 (2023).
• 47)  R. Matsuda, H. Muto, and A. Matsuda, ACS Appl. Mater. Interfaces, 14, 52440 (2022).
• 48)  K. Kodera and Y. Onishi, Bull. Chem. Soc. Jpn., 32, 356 (1959).