Mechanochemical Synthesis of Potassium–Ion Conductor K₃SbS₄

Abstract

Antimony-based sulfide solid electrolytes exhibit high conductivity for alkaline cations. In this study, we synthesized K₃SbS₄ potassium-ion conductors using the mechanochemical method for the nominal compositions with x mol% excess K₂S (x = 0, 5, 10, and 15) to compensate for the chemical impurities in the K₂S reagent. The mechanochemically prepared samples showed X-ray diffraction patterns similar to β-K₃SbS₄ in all the compositions. Raman bands attributed to the SbS₄³⁻ unit were observed in all the samples. The ionic conductivities at 25 °C showed a positive correlation with increasing x, reaching a maximum ionic conductivity of 3.6 × 10⁻⁶ S cm⁻¹ at 10 mol% excess K₂S. Subsequent heat-treatment further enhanced the ionic conductivity, achieving 1.2 × 10⁻⁵ S cm⁻¹ at 25 °C. This improvement is attributed to the nominal composition being close to that of K₃SbS₄ by adjusting the excess amount of K₂S and the increased crystallinity of β-K₃SbS₄.

1. Introduction

Potassium-ion batteries have attracted considerable attention as next-generation rechargeable batteries owing to their cost-effectiveness, which stems from the abundance of potassium in the Earth’s crust and their ability to operate at high voltages, attributed to the low redox potential of potassium (−2.93 V vs. standard hydrogen electrode).^1–4 The replacement of organic liquid electrolytes in rechargeable batteries by non-flammable inorganic solid electrolytes is expected to improve battery safety, extend battery life and increase energy density.⁵ Oxide-based materials such as K-β′′-alumina⁶ are known as inorganic potassium ion conductors. However, oxide materials have issues such as poor formability and typically require the fabrication of high-temperature sintered bodies to minimize grain boundary resistance.

Sulfide-based materials have been studied as solid electrolytes for all-solid-state Li or Na batteries. They offer high ionic conductivity and excellent formability, enabling densification via simple room-temperature pressing.⁷ Sulfide solid electrolytes using phosphorus or antimony as the central element have been reported as lithium-ion or sodium-ion conductors. Among thioantimonate systems, Li₃SbS₄ crystal⁸ and Na₃SbS₄ crystal^9,10 show ionic conductivities of 10⁻⁹ S cm⁻¹ and 10⁻³ S cm⁻¹ at 25 °C, respectively. Antimony-based sulfides are advantageous for conducting larger cations.^11,12 We previously synthesized W-doped Na₃SbS₄ via the mechanochemical method followed by heat treatment, achieving an ionic conductivity exceeding 10⁻² S cm⁻¹ by introducing Na vacancies and stabilizing metastable cubic phase.¹¹

Previously, β-K₃SbS₄ in an orthorhombic crystal system (Cmc2₁) prepared by the solid-phase method showed an ionic conductivity of 2.5 × 10⁻⁶ S cm⁻¹ at 20 °C. The conductivity was further increased to 7.7 × 10⁻⁵ S cm⁻¹ at 20 °C by W-doping.¹³ Although K₂S is used as a raw material for synthesizing potassium-ion-conducting sulfides, most commercial K₂S reagents contain impurities such as K₂S₃, necessitating compositional adjustments to accurately obtain K₃SbS₄. In this study, the enhancement of ionic conductivity was targeted through the adjustment of potassium content and crystallization. Elemental analysis was performed on potassium sulfide reagent used as a raw material to investigate its K/S ratio. K₃SbS₄ was synthesized using 0–15 mol% excess K₂S reagent via the mechanochemical method, followed by heat treatment. The obtained electrolytes were characterized using X-ray diffraction and Raman spectroscopy, and their ionic conductivities were determined using AC impedance spectroscopy.

2. Experimental

2.1 Analysis of K₂S reagent

K₂S (Potassium sulfide) is used as a raw material for K-ion-conducting sulfide solid electrolytes. As the K₂S reagent contains polysulfides such as K₂S₃, an excess amount of the K₂S reagent has to be added to adjust the concentration of K ions in the sample to the target compositions. Figure 1 shows the X-ray diffraction (XRD) pattern of the K₂S reagent (95 %; Strem Chem, Inc.) used in this study. The XRD pattern is mainly attributable to K₂S. However, minor peaks were observed for K₂S₃, K₂S₂O₃, and K₂SO₄. In addition, the detailed K/S ratio in the K₂S reagent was analyzed. We measured the K ratio using an ion analyzer (IA-300, DKK-TOA Corp.) and the S ratio using a CHNS analyzer (Vario EL cube, Elementar Inc.). Table 1 shows the weight percentages (wt%) of K and S in the K₂S reagent, which contained 66.0 % K and 30.1 % S. Note that the determined weight percentages include K and S derived from impurities of K₂S₃, K₂S₂O₃, and K₂SO₄. The K/S ratio of the reagent was calculated to be K_1.8S, with a low potassium content.

Figure 1.

XRD pattern of reagent K₂S (Strem Chemicals).

Table 1. K and S ratio of the reagent K₂S.

K/wt%	S/wt%
66.0	30.1

2.2 Synthesis of K₃SbS₄

All synthesis processes were performed in a dry Ar atmosphere. K₂S (95 %; Strem Chem., Inc.), Sb₂S₃ (99.8 %; Nihonseiko Co., Ltd.), and S (99.99 %; Kojundo Chem., Co., Ltd.) powders were used as the starting materials for K₃SbS₄. Table 2 lists the molar ratio of the starting materials. The nominal composition of K₃SbS₄ was defined in terms of molar ratios, assuming that all starting materials were stoichiometric, and Table S1 lists the experimentally weighed amounts of the starting materials. To adjust the stoichiometric composition and compensate for the potassium deficiency in the K₃SbS₄, we prepared K₃SbS₄ samples with 5–15 mol% excess K₂S. A mixture of the starting materials was placed in a ZrO₂ milling pot (45 mL) together with 50 g of ZrO₂ balls (5 mm diameter) and subjected to mechanochemical treatment at 510 rpm for 5 h using a planetary ball mill (Pulverisette 7; Fritsch GmbH) to obtain K₃SbS₄ (x mol% excess K₂S).

Table 2. Nominal compositions of starting materials for the synthesis of K₃SbS₄ samples with x mol% excess K₂S.

x	K2S	Sb2S3	S
0	3	1	2
5	3.15	1	2
10	3.30	1	2
15	3.45	1	2

We also prepared a heat-treated sample of K₃SbS₄ (10 mol% excess K₂S). The sample powder obtained by mechanochemical treatment was pelletized into a 10 mm diameter disc by pressing under 360 MPa at 25 °C for 5 min, followed by heat treatment at 190 °C for 3 h.

2.3 Characterization

The crystalline phases of the synthesized samples were characterized by XRD using an X-ray diffractometer (SmartLab, Rigaku Corp.) with CuKα radiation. The diffraction data were collected over a 2θ range of 10° to 80°, with a step size of 0.02° and a scanning speed of 10° min⁻¹. A Raman spectrometer (RAMANtouch; Nanophoton Corp.) equipped with a 532 nm laser was used to identify the structural units in the samples. Differential thermal analysis (DTA) was performed using a thermal analyzer (Thermo Plus TG_DTA 8120; Rigaku Corp.), and thermal changes such as crystallization were observed. Energy-dispersive X-ray spectroscopy (EDX) analysis was performed using an EDX spectroscopy system (EMAXEvolution X-MAX, Horiba Ltd.). The ionic conductivity was measured via electrochemical impedance spectroscopy (EIS) using an impedance analyzer (1260A; AMETEK Inc., Solartron Analytical). Measurements were conducted over a frequency range 10⁻¹ to 10⁷ Hz with an applied voltage of 50 mV. Pellets with a diameters 10 mm were obtained by uniaxially pressing the powder samples at 360 MPa for 5 min. Thin gold films were then coated on both sides of the pellets as current collectors under vacuum using a quick coater (SC-701 MkII ADVANCE, Sanyu Co., Ltd.) installed in a dry Ar filled glove box.

3. Results and Discussion

Figure 2 shows the XRD patterns of the K₃SbS₄ samples with x mol% excess K₂S reagent (x = 0, 5, 10, and 15) prepared using the mechanochemical method. The XRD patterns for all compositions were similar to that of β-K₃SbS₄ with moderately broad peaks, indicating that samples containing β-K₃SbS₄ with low crystallinity were obtained. The Raman spectra of the samples prepared using the mechanochemical method are shown in Fig. 3. The three Raman bands around 355, 370, and 390 cm⁻¹, which are attributed to SbS₄³⁻ units, were observed in all samples. No peaks corresponding to the starting materials are observed in the XRD patterns or Raman spectra. The impurities of K₂S₂O₃ and K₂SO₄ derived from the K₂S reagent may remain as amorphous components.

Figure 2.

XRD patterns of K₃SbS₄ with x mol% excess K₂S (x = 0, 5, 10, and 15) prepared by the mechanochemical method.

Figure 3.

Raman spectra of K₃SbS₄ with x mol% excess K₂S prepared by the mechanochemical method.

Figure 4a shows the temperature dependence of the ionic conductivities, and Fig. 4b presents the dependence of the ionic conductivities at 25 °C on the composition of K₃SbS₄ prepared with x mol% excess K₂S. The conductivities of all the samples obeyed the Arrhenius equation. The Nyquist plots at 29.7 °C for the sample with 10 mol% excess K₂S is shown in Fig. S1. The typical profile of an ionic conductor with a semicircle and a spike (ion-blocking electrode behavior) was observed. The bulk and grain boundary components were not separated in the Nyquist plots; hence, the total conductivity was calculated from the resistance of the semicircle. Table 3 shows the ionic conductivities at 25 °C and activation energies for the K₃SbS₄ samples with x mol% excess K₂S. The ionic conductivities at 25 °C increased with increasing x, and the maximum ionic conductivity was obtained at 10 mol% excess K₂S. The activation energies also tended to decrease with the addition of K₂S; the sample with x = 10 exhibited a lower activation energy than that without excess K₂S. Table 4 lists the calculated potassium (K) contained in K₂S from the analysis results in Table 1 and the presumed K/Sb ratio of K₃SbS₄ with x mol% excess K₂S. A K/Sb ratio close to 3, which is the stoichiometric composition, resulted in the highest conductivity obtained at x = 10.

Figure 4.

(a) Temperature dependence of ionic conductivities and (b) composition dependences of ionic conductivities at 25 °C of K₃SbS₄ with x mol% excess K₂S prepared by the mechanochemical method.

Table 3. Ionic conductivities at 25 °C, and activation energies of K₃SbS₄ with x mol% excess K₂S prepared by the mechanochemical method.

x	σ25 °C/S cm−1	Ea/kJ mol−1
0	8.9 × 10−7	56
5	2.5 × 10−6	49
10	3.6 × 10−6	50
15	9.7 × 10−7	55

Table 4. Experimentally weighed amounts of starting materials K₂S, moles of potassium (K) contained in K₂S from the analysis results in Table 1, moles of antimony (Sb) in Sb₂S₃, and presumed K/Sb ratio of K₃SbS₄ samples with x mol% excess K₂S.

x	K2S/g	K contained in K2S/g	K contained in K2S/mol	Sb in Sb2S3/mol	Presumed K/Sb ratio
0	0.2247	0.1483	3.80 × 10−3	1.36 × 10−3	K2.79Sb
5	0.2313	0.1527	3.90 × 10−3	1.33 × 10−3	K2.93Sb
10	0.2366	0.1561	3.99 × 10−3	1.30 × 10−3	K3.07Sb
15	0.2424	0.1600	4.10 × 10−3	1.28 × 10−3	K3.21Sb

To improve the crystallinity of β-K₃SbS₄, we heat-treated the x = 10 sample. The heat-treatment temperature was determined based on the temperature close to the crystallization peak observed in the DTA curve (Fig. S2). Figure 5a shows XRD patterns and Fig. 5b shows the Raman spectra of K₃SbS₄ (x = 10) before and after heat-treatment at 190 °C. The peaks of the XRD pattern became sharper after heat treatment. The change in the full-width at half-maximum (FWHM) of the XRD peaks in the x = 10 sample before and after heat treatment is shown in Fig. S3; the FWHM decreased with the heat treatment. In contrast, the Raman spectra hardly changed before and after the heat-treatment, and the Raman bands attributed to SbS₄³⁻ units were observed. In the EDX spectra of K₃SbS₄ with 10 mol% excess K₂S (Fig. S4), the fraction of oxygen remained largely unchanged before and after heat treatment. The synthesized K₃SbS₄ still contained trace amounts of oxygen derived from impurities in the K₂S reagent. The weight ratio of the crystalline phase of β-K₃SbS₄ and amorphous phase was evaluated using the reference intensity ratio (RIR) method. K₃SbS₄ with 10 mol% excess K₂S was mixed with Al₂O₃ in a weight ratio of 1 : 1, and the obtained XRD pattern (Fig. S5) was analyzed using the Rietveld refinement method. The amorphous components of the milled sample decreased by crystallization from 38 wt% to 24 wt% through heat treatment (Table S2). Figure 6 shows the temperature-dependences of ionic conductivities of the x = 10 samples before and after heat treatment at 190 °C. By heat treatment, the ionic conductivity was increased from 3.6 × 10⁻⁶ S cm⁻¹ to 1.2 × 10⁻⁵ S cm⁻¹ at 25 °C, and the activation energy was decreased from 50 kJ mol⁻¹ to 39 kJ mol⁻¹. The increase in conductivity by heat treatment is attributed to the increase in crystallinity of β-K₃SbS₄. The potassium-ion conductivity in β-K₃SbS₄ crystals is suggested to be higher than that in the amorphous domains. The prepared K₃SbS₄ (x = 10) showed one order of magnitude higher conductivity than the previously reported β-K₃SbS₄ (2.5 × 10⁻⁶ S cm⁻¹).¹³ This is attributed to the nominal composition being close to that of K₃SbS₄ by adjusting the amount of K₂S added, and the increased crystallinity of β-K₃SbS₄. The prepared K₃SbS₄ exhibited the conductivity intermediate between that of crystalline Li₃SbS₄ (4.8 × 10⁻⁹ S cm⁻¹)⁸ and Na₃SbS₄ (3 × 10⁻³ S cm⁻¹).⁹ This suggests that the conductivity of thioantimonate solid electrolytes is not simply determined by the size of the carrier ions, but that the crystal structure with appropriate ionic conduction pathways governs ionic conductivity.

Figure 5.

(a) XRD patterns and (b) Raman spectra of K₃SbS₄ with 10 mol% excess K₂S before and after heat treatment at 190 °C.

Figure 6.

Temperature dependence of ionic conductivities of K₃SbS₄ with 10 mol% excess K₂S before and after heat treatment at 190 °C.

4. Conclusions

In this study, we first performed elemental analysis of the K₂S reagent. Based on the analysis, the composition of the K₂S reagent used in this study was calculated to be K_1.8S. Subsequently, K₃SbS₄ samples with 0–15 mol% excess K₂S were synthesized using the mechanochemical method. The XRD patterns and Raman spectra indicate that the prepared samples comprise crystalline β-K₃SbS₄ with partially amorphous component. The ionic conductivity increased with increasing excess K₂S, content; the sample prepared with 10 mol% excess K₂S showed the highest ionic conductivity of 3.6 × 10⁻⁶ S cm⁻¹. Furthermore, the ionic conductivity of the K₃SbS₄ sample with 10 mol% excess K₂S was increased to 1.2 × 10⁻⁵ S cm⁻¹ and the activation energy was decreased to 39 kJ mol⁻¹ by increasing the crystallinity of β-K₃SbS₄ through heat treatment.

Acknowledgments

This work was partly supported by the MEXT Program Data Creation and Utilization-Type Material Research and Development Project Grant Number SPMXP1122712807, JST Adopting Sustainable Partnerships for Innovative Research Ecosystem Grant Number JPMJAP2313, and JSPS KAKENHI (Grant Numbers JP24H02204 and JP25H00904).

Data Availability Statement

The data that support the findings of this study are openly available under the terms of the designated Creative Commons License in J-STAGE Data at https://doi.org/10.50892/data.electrochemistry.29389592.

• 1) Antimony-based sulfide solid electrolytes exhibit high conductivity for alkaline cations. In this study, we synthesized K₃SbS₄ potassium-ion conductors using the mechanochemical method for the nominal compositions with x mol% excess K₂S (x = 0, 5, 10, and 15) to compensate for the chemical impurities in the K₂S reagent. The mechanochemically prepared samples showed X-ray diffraction patterns similar to β-K₃SbS₄ in all the compositions. Raman bands attributed to the SbS₄³⁻ unit were observed in all the samples. The ionic conductivities at 25 °C showed a positive correlation with increasing x, reaching a maximum ionic conductivity of 3.6 × 10⁻⁶ S cm⁻¹ at 10 mol% excess K₂S. Subsequent heat-treatment further enhanced the ionic conductivity, achieving 1.2 × 10⁻⁵ S cm⁻¹ at 25 °C. This improvement is attributed to the nominal composition being close to that of K₃SbS₄ by adjusting the excess amount of K₂S and the increased crystallinity of β-K₃SbS₄.

CRediT Authorship Contribution Statement

Takehiro Nakao: Data curation (Lead), Investigation (Lead), Resources (Lead), Writing – original draft (Lead)

Chihiro Okushima: Formal analysis (Supporting), Investigation (Supporting), Resources (Supporting)

Takuya Kimura: Formal analysis (Equal), Investigation (Equal)

Akira Nasu: Formal analysis (Supporting), Investigation (Supporting)

Kota Motohashi: Supervision (Supporting), Writing – review & editing (Supporting)

Atsushi Sakuda: Supervision (Supporting), Writing – review & editing (Supporting)

Akitoshi Hayashi: Conceptualization (Lead), Funding acquisition (Lead), Supervision (Lead), Writing – review & editing (Lead)

Conflict of Interest

The authors declare no conflict of interest in the manuscript.

Funding

MEXT: SPMXP1122712807

JST: JPMJAP2313

JSPS: JP24H02204 and JP25H00904

Footnotes

T. Nakao: ECSJ Student Member

A. Nasu, K. Motohashi, and A. Sakuda: ECSJ Active Members

A. Hayashi: ECSJ Fellow

References

• 1)  J. C. Pramudita, D. Sehrawat, D. Goonetilleke, and N. Sharma, Adv. Energy Mater., 7, 1602911 (2017).
• 2)  R. Rajagopalan, Y. Tang, X. Ji, C. Jia, and H. Wang, Adv. Funct. Mater., 30, 1909486 (2020).
• 3)  N. Yabuuchi, K. Kubota, M. Dahbi, and S. Komaba, Chem. Rev., 114, 11636 (2014).
• 4)  Y. Marcus, Pure Appl. Chem., 57, 1129 (1985).
• 5)  J. Janek and W. G. Zeier, Nat. Energy, 1, 16141 (2016).
• 6)  N. Baffier, J. C. Badot, and P. Colomban, Mater. Res. Bull., 16, 259 (1981).
• 7)  A. Hayashi, A. Sakuda, and M. Tatsumisago, Front. Energy Res., 4, 25 (2016).
• 8)  T. Kimura, A. Kato, C. Hotehama, A. Sakuda, A. Hayashi, and M. Tatsumisago, Solid State Ionics, 333, 45 (2019).
• 9)  L. Zhang, D. Zhang, K. Yang, X. Yan, L. Wang, J. Mi, B. Xu, and Y. Li, Adv. Sci., 3, 1600089 (2016).
• 10)  F. Tsuji, S. Yubuchi, A. Sakuda, M. Tatsumisago, and A. Hayashi, J. Ceram. Soc. Jpn., 128, 641 (2020).
• 11)  A. Hayashi, N. Masuzawa, S. Yubuchi, F. Tsuji, C. Hotehama, A. Sakuda, and M. Tatsumisago, Nat. Commun., 10, 5266 (2019).
• 12)  T. Fuchs, S. P. Culver, P. Till, and W. G. Zeier, ACS Energy Lett., 5, 146 (2020).
• 13)  J. Shao, J. Zheng, L. Qin, S. Zhang, Y. Ren, and Y. Wu, Angew. Chem. Int. Ed., 61, e202200606 (2022).