Concerted Influence of H2O and CO2: Moisture Exposure of Sulfide Solid Electrolyte Li4SnS4

Although moisture-induced deterioration mechanisms in sulfide solid electrolytes to enhance atmospheric stability have been investigated, the additional impact of CO2 exposure remains unclear. This study investigated the generation of H2S from Li4SnS4 under H2O and CO2 exposure. Li4SnS4 was exposed to Ar gas at a dew point of 0 °C with and without 500 ppm of CO2, and its ion conductive properties were evaluated. Although the lithium-ion conductivity of Li4SnS4 decreased regardless of the presence of CO2, the amount of H2S generated with CO2 was five times higher. To elucidate the underlying mechanism, X-ray diffraction and Raman spectroscopy were used. Without CO2, hydrate Li4SnS4·4H2O formation markedly increased, whereas, with CO2, it increased a little. The difference revealed distinct deterioration mechanisms leading to a decrease in lithium-ion conductivity: without CO2, adsorbed H2O and Li4SnS4·4H2O contributed to the decrease, while with CO2, a weak acid dissociation reaction could reduce the thermodynamic stability of the moisture-exposed Li4SnS4 surface including Li4SnS4·4H2O and adsorbed H2O, promoting H2S release and carbonate formation. This was supported by the recovery of lithium-ion conductivity after vacuum heating. The concerted influence of H2O and CO2 provides valuable insights into the fundamental deterioration mechanisms in sulfide solid electrolytes that could be applied in battery manufacturing processes.


■ INTRODUCTION
All-solid-state batteries using inorganic solid electrolytes (SEs) are considered promising energy devices due to their higherrate charge/discharge capability, longer lifetime expectancy compared to that of present liquid-type lithium-ion batteries (LIBs), enhanced safety features, and wider operating temperature range.−20 Recently, Li 4 SnS 4 (LSS) has attracted attention as an SE material capable of exhibiting both good lithium ion conductivity and superior moisture durability.For example, it has been reported that hexagonal LSS synthesized via mechanical milling followed by low-temperature heat treat-ment generates an exceptionally small quantity of H 2 S gas (∼0.2 cm 3 g −1 ) during exposure to temperatures ranging from 20 to 22 °C and relative humidity (R.H.) of 70% for 40 min 19 Besides, various attempts have also been made to dope LSS with Li 3 PS 4 20−22 and other additives 23,24 in order to promote enhanced lithium ionic conductivity while maintaining moisture durability.Investigating beyond material development, Kimura et al. have analyzed the mechanisms underlying the moisture durability in both hexagonal and orthorhombic LSS when exposed to humidified inert gases such as N 2 and Ar.Interestingly, the formation of hydrate Li 4 SnS 4 •4H 2 O upon exposure to humidity followed by its recovery to an anhydrous state via heat treatment under vacuum conditions was observed. 25It is considered that the generation of H 2 S gas is significantly suppressed due to the thermodynamically stable hydrate Li 4 SnS 4 •4H 2 O resulting from S atoms shared between LiS 2 (H 2 O) 2 tetrahedra and SnS 4 4− tetrahedra. 250][11][12][13]29 As a matter of fact, during industrial processes such as battery manufacturing, material synthesis, and storage, the environmental atmosphere typically contains not only moisture but also a mixture of various gases. Threfore, in this study, we focused on investigating the influence of CO 2 on moisture exposure.The LSS powder samples were exposed to moisture-controlled Ar gas flow with and without CO 2 while monitoring the amount of H 2 S gas generation.Afterward, the samples were characterized by lithium ionic conductivity measurement using electrochemical impedance spectroscopy (EIS), X-ray diffraction (XRD), and Raman spectroscopy to elucidate the effects associated with CO 2 during moisture exposure and to understand the corresponding deterioration mechanisms.Finally, vacuum heating was performed to verify the deterioration mechanisms and to demonstrate a recovery treatment for the successful implementation of all-solid-state batteries.

Material and Preparation.
A sulfide SE of LSS powder was synthesized according to a previously reported procedure; 26,27 Li 2 S, Sn, and S were introduced into ultrapure H 2 O in a molar ratio of 2:1:2.The mixture was dissolved while stirring at 80 °C for more than 12 h and then dried at 120 °C in a vacuum for 3 h to obtain a hexagonal LSS powder with an average particle size of approximately 1 μm.The controlled moisture exposure and H 2 S monitoring system (Figure 1a) was constructed based on the report by Yamada et al. 30 A gas cylinder containing Ar (<0.1 ppm of CO 2 ) or Ar + 500 ppm of CO 2 with a dew point below −80 °C (<0.5 ppm of H 2 O) was purchased (Grade 1, Taiyo Nippon Sanso Co., Japan) and connected to the upstream end of the system.The gas line was split into two branches.In one branched line, by bubbling argon gas through water, we generated a gas that contained water vapor.The two gas lines were then united under control of each flow rate by using mass flow controllers in order to prepare moisture-controlled gas.The dew point of the moisture-controlled gas was confirmed using an in-line dew point meter as 0 °C (∼6000 ppm of H 2 O) (Figure 1b).Then, 200 mg of LSS powder was exposed to the controlled-moisture gas at a flow rate of 0.8 L min −1 for 1 h.A H 2 S sensor (Model RS3000, Advanced Micro Instruments, USA) was connected to the line after the exposed LSS sample to monitor the amount of H 2 S gas generated.The total amount of generated H 2 S was calculated per 1 g of SE by accumulating the concentration value.The LSS samples before exposure are labeled as "pristine", the sample after exposure to a dew point of 0 °C without CO 2 as "without CO 2 ", and the sample after exposure to a dew point of 0 °C with CO 2 as "with CO 2 ," thereafter.Additionally, a recovery treatment demonstration was conducted under the same conditions as the synthesis process at 120 °C in a vacuum for 3 h.
Lithium Ionic Conductivity.The LSS samples, 80 mg, were pelletized at 360 MPa and restrained at a pressure of 98 ppm in a zirconia cylinder with a diameter of 10 mm between two stainless-steel (SUS) electrodes.The lithium ionic conductivity was evaluated using EIS measurement with a voltage amplitude of 30 mV in a frequency range of 10 6 −10 1 Hz at a temperature of 22 °C.Nyquist and Bode plots were used to describe the normalized impedance (measured in kΩ cm), taking into account the pellet thickness and electrode area.
X-ray Diffraction.XRD measurements were performed using a reflective configuration system (Empyrean, Malvern PANalytical, UK) and an airtight sample holder with knife edge.The diffraction patterns were obtained in the 2θ range 10−80°with a step width of 0.1°, using Cu Kα as the X-ray source.The diffraction patterns were analyzed using the software package "The General Structure and Analysis Software II (GSAS-II)". 31aman Spectroscopy.Raman spectroscopy measurements were conducted using a system (RAMANforce, Nanophoton, Japan) equipped with an incident green laser at a wavelength of 523 nm, which was directed through a quartz glass window in an airtight sample holder.The measurements were performed under controlled conditions to ensure that the signal-to-noise ratio of the Raman signal was within acceptable limits: The laser output was reduced to 10 μW μm −2

■ RESULTS AND DISCUSSION
Figure 2 shows the concentration of H 2 S gas generation (i.e., the generation rate) and the total amount of H 2 S (cc) per gram of LSS (cc g −1 ).In the case without CO 2 , the generation rate of H 2 S gas was a constant value of ∼0.6 ppm, and the total amount after 1 h exposure was 0.17 cc g −1 .The order of the total amount of H 2 S gas was roughly the same as previously reported results. 19,20By contrast, in the case of CO 2 , the behavior of H 2 S gas generation was obviously different; the generation rate of H 2 S gas increased with exposure time, resulting in a total amount of 0.95 cc g −1 , which was more than 5 times higher than the case without CO 2 .Here, it should be noted that the amount of H 2 S per LSS is lower than that per typical Li 3 PS 4 and argyrodite-structured Li 6 PS 5 Cl. 11,12,19Each H 2 S amount of 0.17 and 0.95 cc g −1 corresponds to merely <0.4 and 2 Å from the particle surface, respectively, when calculated geometrically under the assumption that one H 2 S molecule is generated from one SnS 4 4− unit.The estimated deterioration thickness also indicates that LSS has a remarkably high moisture durability.
The EIS data shown in Figure 3a reveal an increase in impedance compared to pristine in either case of moisture exposure with and without CO 2 .In other words, lithium ionic conductivity decreased upon moisture exposure with and without CO 2 .Notably, the decrease in lithium ionic conductivity was smaller for the case without CO 2 , although the total amount of H 2 S gas was smaller than that for the case with CO 2 .This result indicates that the decrease in lithium ionic conductivity was not solely caused by the decomposition of the LSS structure with H 2 S desorption.The discrepancy between the total amount of generated H 2 S and the retention of lithium ionic conductivity suggests that the respective deterioration modes in the case with and without CO 2 can be different from each other.
Figure 3b shows the EIS data of vacuum-heated samples after both moisture exposures compared to the pristine sample.It reveals that the deteriorated lithium ionic conductivities have been significantly recovered and returned to almost the same value as the pristine sample.However, in more detail, there were differences in the spectral shape in higher-frequency  regions.A small semicircle remained in the EIS data after vacuum heating the sample exposed to moisture and CO 2 .This suggests that the surface chemical state after vacuum heating differs from the pristine state.The semicircle in the higherfrequency region is attributed to surface species.In fact, EIS studies 11,27,32 revealed that the impedance components of some sulfide SEs increase due to surface degradation species that result from moisture exposure.Therefore, the incomplete recovery in EIS for the sample exposed to moisture and CO 2 indicates that irreversible reactions occurred at the SE surface.Table 1 summarizes the total amount of H 2 S gas, lithium ionic conductivity, and the retention value of lithium ionic conductivity for each sample.To elucidate in detail both deterioration and recovery mechanisms upon moisture exposure with and without CO 2 , XRD and Raman spectroscopy analyses were conducted.
Figure 4 shows the diffraction patterns of samples before and after moisture exposure with and without CO 2 and after vacuum-heated samples.The pristine LSS has a hexagonal single phase as previously reported (Figure 4a). 19,20,28In the case of moisture exposure without CO 2 (Figure 4b), several new peaks appeared, representing peaks at approximately 14.8, 23.8, 31.9, and 34.1°, among others.The diffraction pattern agrees with that of the hydrate Li 4 SnS 4 •4H 2 O crystal reported by Kimura et al., indicating a change in crystal structure due to hydration upon moisture exposure. 25By contrast, despite being exposed to the same amount of H 2 O, the diffraction pattern in the case with CO 2 has much lower-intensity peaks for hydrate Li 4 SnS 4 •4H 2 O, and the hexagonal phase remained more pronounced (Figure 4c).The molecular composition ratios of Li 4 SnS 4 •4H 2 O determined from quantitative analysis for both exposure conditions with and without CO 2 are approximately 14.1 and 5.6 mol %, respectively.In either case, the molar ratios of the hydrates are small, suggesting that the hydration reaction occurs only on the surface.These values correspond to hydrate layer thicknesses of approximately 25 and 9.6 nm from the particle surface, which are very small compared with the average SE particle size of approximately 1 μm.Furthermore, in both vacuum-heated samples (Figure 4d,e), patterns of hydrate Li 4 SnS 4 •4H 2 O have completely disappeared and returned to the hexagonal phase.Besides, carbonates, such as Li 2 CO 3 , are known to be deteriorating components associated with CO 2 on the sulfide SE surface. 12,14,29In fact, a nanocoating method has been proposed  for sulfide SE Li 6 PS 5 Cl particles, utilizing a surface reaction with CO 2 gas. 33However, no distinct peak corresponding to Li 2 CO 3 34,35 is observed, possibly owing to the limitation of XRD as a bulk analysis technique.
Surface analysis is required in addition to bulk analysis by XRD, because the hydration reaction of SE upon moisture exposure is considered to proceed gradually from the SE surface, as mentioned above.To analyze the surface while remaining in its hydrate state after moisture exposure, Raman spectroscopy was adopted rather than general surface analyses such as X-ray photoelectron spectroscopy (XPS) and secondary ion mass spectroscopy (SIMS).Yamamoto et al. also conducted Raman spectroscopy as a surface-sensitive measurement used to evaluate the reaction progress of the liquid-phase synthesis of Li 3 PS 4 by combining with XRD. 36ome studies have reported surface analysis for sulfide SE before and after moisture exposure. 12,14−38 Raman spectra in the regions of the Sn−S, O−H, and C−O vibrations are shown in Figure 5a.In the region of the Sn−S vibration, the main peak at the 354 cm −1 unit and the small subpeak at 300 cm −1 , corresponding to the SnS 4 4− unit, are observed in the pristine sample.This spectral shape of the pristine sample is consistent with the LSS synthesized by mechanical milling, as previously reported. 19After moisture exposure, a new small peak appears at a wavenumber lower than that of the main peak in both spectra of the exposed samples.The new peak at 340 cm −1 is assigned to hydrate Li 4 SnS 4 •4H 2 O. 25 The intensity of the hydrate peak is stronger without CO 2 than that with CO 2 .The hydrate peaks of both moisture-exposed samples disappeared after vacuum heating.The changes in the O−H stretching vibration region naturally align with the abovementioned change in the Sn−S vibration.Focusing on the O−H stretching vibration mode of H 2 O molecules is powerful for analyzing the chemical and physical states of H 2 O. 12,13,39−45 Although the Raman scattering sensitivity of the stretching vibration is not very high, a small change is observed.The slight peaks around 2950 cm −1 may be attributable to hydrocarbon contamination (C x H y ) in the Raman spectroscopy measurement cell. 12In the O−H stretching vibration region of pristine samples, almost no Raman signals are detected, but a broad band around 3200 cm −1 and a small peak at 3570 cm −1 , assigned to LiOH•H 2 O and the adsorbed outer layer H 2 O by the hydrogen bond network, 12,40 are observed.These may be very slight residuals due to the aqueous synthesis method or an unavoidable product by storage in a glovebox.Upon moisture exposure without CO 2 , a new peak around 3070 cm −1 appears in addition to the increase in the intensity of the broad band around 3200 cm −1 .Several studies on hydrates have reported that the vibrational peak of the H 2 O fixed in some crystal structures appears around 3100 cm −1 . 45−49 Therefore, the new peak is assigned to the internal H 2 O molecule in the hydrate Li 4 SnS 4 •4H 2 O crystal structure in the form of LiS 2 (H 2 O) 2 . 25Interestingly, despite the same moisture content of the exposure gas, the intensity of both internal H 2 O in LiS 2 (H 2 O) 2 and outer layer H 2 O is lower in the case with CO 2 than in the case without CO 2 , which is consistent with the result on Li 4− tetrahedron. 25As the hydration progresses, more hydrate is detected via XRD.However, if the frequency of H 2 O attacks on the LSS surface exceeds the rate at which it penetrates into the internal SE bulk region, then thermodynamic stability is disrupted, resulting in a slight amount of H 2 S gas generation.Kaib   12,13,53 As a result, hydrate Li 4 SnS 4 •4H 2 O (∼10 −9 S cm −125 ) and outer layer H 2 O decrease lithium ionic conductivity by inhibiting the conduction on the SE surface.However, these hydrated species reversibly return to the LSS by dehydration and desorption with vacuum heat treatment.As mentioned in the earlier paragraph, the thickness calculated for the H 2 S-released layer was extremely thin (<0.4 and 2 Å), and the lithium ionic conductivity was recovered to almost the same level as the pristine state.
For the moisture exposure with CO 2 , we propose that the deterioration mechanism upon exposure to moisture and CO 2 involves a "concerted influence" (Figure 6b).−56 In other words, CO 2 is more acidic than H 2 S. Therefore, the presence of CO 2 in H 2 O is considered to facilitate the release of H 2 S through a weak-acid dissociation reaction, which is a fundamental chemical reaction.The coordination reaction of CO 2 is more dominant.As a result, the hydrated species decrease and the surface undergoes hydrolysis, leading to the release of H 2 S. Simultaneously, a small amount of carbonates is also formed on the SE surface.Although the hydrated species can reversibly return to LSS through vacuum heating, the hydrolyzed layer (a few nanometers) caused by H 2 S releasing and carbonate species formation slightly decreases lithium ionic conductivity, resulting in irreversible deterioration on the SE surface even after surface dehydration by vacuum heating occurs.Therefore, the "concerted influence" resulting from the coexistence of H 2 O and CO 2 based on a weak-acid dissociation reaction reduces the thermodynamic stability of Li 4 SnS 4 •4H 2 O, promoting H 2 S release and carbonate species formation on the surface, ultimately leading to a decrease in lithium ionic conductivity.Here, it should be emphasized that carbonates, such as Li 2 CO 3 , may not promote H 2 S generation but may be merely byproducts resulting from the addition of CO 2 , which should be elucidated in detail in the future.The concerted influence of H 2 O and CO 2 suggests a new metric that should be considered in the battery manufacturing process.Additionally, the impact of SE surface degradation on battery performance, which has been partially reported, 33,57−59 is one of the future research targets.

■ CONCLUSIONS
In this study, we investigated the surface hydrolysis deterioration when a sulfide SE, LSS, was exposed to Ar gas at a dew point of 0 °C, both with and without 500 ppm of CO 2 .The amount of H 2 S gas generation varied depending on the presence or absence of CO 2 , despite being exposed to the same amount of H 2 O.In the presence of CO 2 , H 2 S gas generation increased by more than five times.However, the lithium ion conductivities significantly decreased after moisture exposure, regardless of the presence or absence of CO 2 .XRD and Raman spectroscopy analyses indicated that the deterioration mechanisms differed noticeably between the two cases.Without CO 2 , the thermodynamically stable hydrate Li 4 SnS 4 •4H 2 O formed on the surface, resulting in minimal H 2 S gas generation and demonstrating excellent reversibility through dehydration with vacuum heating.By contrast, in the presence of CO 2 , a weak acid dissociation reaction promoted the generation of H 2 S and carbonate species on the surface, leading to a decrease in the lithium ionic conductivity.The hydrolyzed species also reduced the reversibility upon dehydration.Further research is needed to elucidate intermediate reactions that occur when H 2 O and CO 2 coexist; moreover, the relationship between atmospheric conditions, such as dew point and gas species, as well as the influence of SE species itself should be explored in more detail.This finding of the "concerted influence" of H 2 O and CO 2 provides valuable insights into material development and future implementation.

■ AUTHOR INFORMATION Corresponding Authors
, and the laser exposure time was set to 10 s to prevent damage such as desorption of adsorbed H 2 O and decomposition of SE.The Raman spectra were calibrated by fixing the Si wafer peak at 520 cm −1 and by normalizing the peak intensity of the Sn−S vibration at 354 cm −1 of the SnS 4 4− unit.

Figure 1 .
Figure 1.(a) System of controlled moisture exposure and H 2 S monitoring.(b) Dew point monitoring when humidity was controlled by setting the target to a dew point of 0 °C.

Figure 2 .
Figure 2. H 2 S gas generation at a dew point of 0 °C in Ar: (a) without CO 2 and (b) with 500 ppm of CO 2 ; H 2 S concentration in gas flow (red, solid line) and total amount (blue, dashed line).The total amount of H 2 S was converted to cc per 1 g of LSS.

Figure 3 .
Figure 3. Nyquist and Bode plots of EIS data: (a) after moisture exposure and (b) after vacuum heating.Pristine (black, square), without CO 2 (green, circle), and with CO 2 (purple, triangle).Plots describe normalized impedance by pellet thickness and electrode area (unit of kΩ cm).The inset of Figure 3b is a magnified view in a higher-frequency region.

Figure 4 .
Figure 4. Diffraction pattern analysis: (a) pristine, 27 (b) moisture-exposed without CO 2 27 and (c) with CO 2 , and (d) vacuum-heated after exposure to moisture without CO 2 and (e) with CO 2 .Experimental data (black, cross mark), calculated data (red, solid line), difference between experimental and calculated data (pink, dashed line), vertical bars at the bottom corresponding to hexagonal Li 4 SnS 4 (blue) and Li 4 SnS 4 •4H 2 O (light green) crystal structures, respectively.Each percentage value indicates the mol % ratio.

Figure 5 .
Figure 5. Raman spectra in each region of pristine (black, dotted line), without CO 2 (green, solid line), and with CO 2 (purple, dashed line): (a) after moisture exposure 27 and (b) after vacuum heating.

4
SnS 4 •4H 2 O generation in the SnS 4 region.Besides, the hydrate markers, such as the new peak at 340 cm −1 in the Sn−S region and the O−H stretching vibration bands, in both moisture-exposed samples disappear after vacuum heating (Figure 5b).The trend of Raman peak changes in Sn−S and H−O−H vibration related to hydrate Li 4 SnS 4 •4H 2 O completely agrees with the result of XRD.However, a slight peak of symmetric stretching vibration of C− O in carbonates at ∼1090 cm −134,50,51 is observed by Raman spectroscopy only in the sample with CO 2 and remained even after vacuum heating.These results of XRD and Raman spectroscopy measurements reveal that the introduction of CO 2 with moisture has resulted in the suppression of hydrate Li 4 SnS 4 •4H 2 O generation and the formation of a small amount of irreversible carbonate species on the SE surface.The contrasting surface states propose different deterioration and recovery mechanisms and support the different behaviors of lithium ion conductivity for the cases with and without CO 2 .The deterioration and recovery mechanism upon moisture exposure without CO 2 is simply explained as follows (Figure 6a): The hydrate Li 4 SnS 4 •4H 2 O is generated on the LSS surface owing to H 2 O attack.The hydrate is thermodynamically stable owing to S sharing between the LiS 2 (H 2 O) 2 tetrahedron and the SnS 4

Figure 6 .
Figure 6.Schematic illustrations of the Li 4 SnS 4 surface upon moisture exposure (a) without and (b) with CO 2 .The coexistence of CO 2 decreases the thermodynamic stability of Li 4 SnS 4 •4H 2 O to promote H 2 S release based on a weak acid dissociation reaction and carbonate species generation on the surface, leading to a decrease in lithium ionic conductivity.

Table 1 .
Total Amount of H 2 S upon Moisture Exposure to Ar Gas at a Dew Point of 0 °C for 1 h, Lithium Ionic Conductivity, and Retention Value of Lithium Ionic Conductivity for Each Sample 2.83 × 10 −5 /99% 2.44 × 10 −5 /85% et al. previously proposed the crystal structure of Li 4 SnS 4 •13H 2 O as a hydrate with a larger number of H 2 O molecules.It exhibits a NaCl-type crystal structure, consisting of SnS 4 4− tetrahedral anion unit and [Li 4 (H 2 O) 13 ] 4+ hydrate complex cation units, which are connected through hydrogen bonds. 52Although Li 4 SnS 4 •4H 2 O and Li 4 SnS 4 •13H 2 O are presumed to be thermodynamically stable, if an intermediate state between them as a transition state is formed, thermodynamical stability decreases for the Scontaining units such as Li 2 S and SnS 4