Air exposure towards stable Li/Li₁₀GeP₂S₁₂ interface for all-solid-state lithium batteries

All-solid-state lithium batteries exhibit increased safety due to the employment of nonflammable inorganic solid electrolytes [1–3]. To date, various inorganic solid electrolytes, especially sulfide solid electrolytes, possessing high ionic conductivity have been developed [4–8], potentially achieving high-rate capability of all-solid-state lithium batteries. However, sulfide solid electrolytes exhibit extremely poor air stability and long-time air exposure leads to severe decomposition of sulfide electrolytes with destroyed structure and release of toxic H₂S gas [9, 10]. The accumulated decomposition products with poor ionic conductivity deteriorate the bulk and interface ion transfer, and finally resulting in rapid battery decay [10, 11].

Compared with commercial lithium-ion batteries, the employment of lithium metal anodes is an essential prerequisite to realize higher energy density all-solid-state batteries [2, 3, 12]. Among sulfide electrolytes, Li₁₀GeP₂S₁₂ possesses high ionic conductivity of 12 mS cm⁻¹, which exceeds to that of organic liquid electrolytes [4]. However, the sulfide solid electrolyte Li₁₀GeP₂S₁₂ strongly reacts with lithium metal and decomposes to Li₁₅Ge₄, Li₃P and Li₂S [13–15]. The large fraction of metallic Li₁₅Ge₄ at the Li/Li₁₀GeP₂S₁₂ interface can create electronically conducting pathways, which will continuously consume the bulk Li₁₀GeP₂S₁₂ and increase impedance until cell failure. Thus, a passivating layer without electronic conductivity is crucial to suppress the highly reactive Li/Li₁₀GeP₂S₁₂ interface. Modification of lithium metal through chemical reaction is an appealing strategy to stabilize the Li/Li₁₀GeP₂S₁₂ interface. Zhang et al [16] identified the in-situ formation of the LiH₂PO₄ by reacting the lithium metal with H₃PO₄. This LiH₂PO₄ layer is ionically conducting but electronically insulating, which can prevent the direct contact between Li₁₀GeP₂S₁₂ and lithium metal and passivate the interface reaction. A more effective approach is introducing an electronic insulating layer with high interface energy against lithium. Wan et al [17] constructed a bifunctional layer at Li/Li₁₀GeP₂S₁₂ interface by reacting the lithium meal with Mg(TFSI)₂-LiTFSI-DME liquid electrolyte. The sequential reduction of salts and solvent generates a gradient solid electrolyte interface Li_x Mg/LiF/polymer, resulting in a stabilized interface and demonstrating effective protection for Li₁₀GeP₂S₁₂. In addition, using bilayer composite electrolyte can also passivate the Li/Li₁₀GeP₂S₁₂ interface [18–20]. The cells employing Li-argyrodites Li_5.5PS_4.5Cl_1.5 or Li₁₀GeP₂S₁₂ exhibit distinct failure behavior toward lithium metal due to short circuit by lithium dendrite penetration for Li_5.5PS_4.5Cl_1.5 and increased overpotential by electrolyte decomposition for Li₁₀GeP₂S₁₂. The bilayer construction of Li/Li_5.5PS_4.5Cl_1.5/Li₁₀GeP₂S₁₂ shows excellent interface stability even at high current density, in which the Li_5.5PS_4.5Cl_1.5 layer as buffer layer is to isolate Li₁₀GeP₂S₁₂ from lithium metal and the Li₁₀GeP₂S₁₂ can prevent the lithium dendrite penetration [20]. Clearly, designing an artificial passivation interface layer by more convenient method is crucial to stabilize Li/Li₁₀GeP₂S₁₂ interface, while the reported approaches generally involve complicated reaction process or electrolyte multilayer structures.

Considering the severe reaction between sulfide electrolytes and moisture in air, interestingly, it is found that short-period air exposure of sulfide electrolytes with controlled humidity could provide effective passivation against lithium metal. The dramatically improved interface stability between Li₁₀GeP₂S₁₂ solid electrolyte and lithium metal is achieved by simply exposing the Li₁₀GeP₂S₁₂ pellet into air for dozens of seconds. This air-exposure treatment could rapidly generate a protective layer of Li₄P₂S₆, GeS₂ and Li₂HPO₃ coated on the surface of the Li₁₀GeP₂S₁₂ pellets. This protecting layer is ionically conductive but electronically insulation, which can not only physically isolate the contact between Li₁₀GeP₂S₁₂ and lithium metal but also effectively suppress the continuous decomposition of Li₁₀GeP₂S₁₂ reduced by lithium metal.

2.1. Preparation of Li₁₀GeP₂S₁₂ pellets with and without air-exposure treatment

2.2. Materials characterization

2.3. Electrochemical performance measurements

As shown in figure 1(a), in contrast with the nearly liner increasing Li plating/striping voltage for the symmetric Li/Li₁₀GeP₂S₁₂/Li cell without air-exposure treatment, the Li/air-exposed Li₁₀GeP₂S₁₂/Li cells show obviously sluggish increasement of the potential. The air-exposure durations were set at 10, 20, 30, 40, and 50 s. In addition, the suppressed increasement of impedance for the Li/air-exposed Li₁₀GeP₂S₁₂/Li cells was directly observed during the electrochemical cycling. Figure 1(b) presents the impedance of the Li/Li₁₀GeP₂S₁₂ or air-exposed Li₁₀GeP₂S₁₂/Li cells for the cycling tests. It can be clearly found that the continuously increasing impedance arising from strong reaction at Li/Li₁₀GeP₂S₁₂ interface was dramatically suppressed when the air-exposure treatment for Li₁₀GeP₂S₁₂ pellets was employed. Figure 1(c) presents the voltage of the symmetric Li/Li₁₀GeP₂S₁₂ or air-exposed Li₁₀GeP₂S₁₂/Li cells before and after cycling. The Li/Li₁₀GeP₂S₁₂/Li symmetric cell shows much higher polarization voltage than that of air-exposed Li₁₀GeP₂S₁₂ based symmetric cells before cycling, which is due to the decomposition reaction occurred once the Li₁₀GeP₂S₁₂ solid electrolytes contact with Li metal during assembling the symmetric cells. However, the obviously suppressed increasement of the voltage after cycling was detected after short-time exposure of 10 s. With increased air-exposure durations, such as 40 s and 50 s, the Li plating/striping voltages perform negligible increase even after 400 h cycling, which clearly demonstrates that the air-exposure treatment can effectively stabilize the Li/Li₁₀GeP₂S₁₂ interface. Whereas, the impedance of the Li/air-exposed Li₁₀GeP₂S₁₂/Li symmetric cells nearly linear increases with the air-exposure durations (figure 1(d)), which implies that the air-exposure treatment introduced a low ionically conducting layer at the interface. Moreover, the difference value of the impedance of the Li/air-exposed Li₁₀GeP₂S₁₂/Li cells for cycling tests (Δ R_t = R_{t, (400 h)} − R_{t, (0 h)}) were evaluated. As presented in figure 1(e), the symmetric cells using Li₁₀GeP₂S₁₂ with an optimal exposure duration of 40 s exhibit the smallest value of the impedance changes, showing the decreased impedance for the Li/40 s air-exposed LGPS/Li symmetric cell after cycling. Specifically, compared with the impedance of around 660 Ω before cycling, the impedance decreases to around 487 Ω after 400 h cycling (figure 1(f)). The decreased impedance could be attributed to the improved interfacial contact between the decomposition layer and Li metal due to the volume expansion of lithium metal during repeat plating/striping processes [18].

Zoom In Zoom Out Reset image size

Figure 2(a) shows the galvanostatic Li plating/striping of the Li/40 s air-exposed Li₁₀GeP₂S₁₂/Li and the Li/Li₁₀GeP₂S₁₂/Li cells. The Li/Li₁₀GeP₂S₁₂/Li cell shows a rapid increase in polarization voltage due to the continuous and deteriorative Li/Li₁₀GeP₂S₁₂ interface reaction. In contrast, the Li/40 s air-exposed Li₁₀GeP₂S₁₂/Li cell presents improved interface stability with small polarization voltage of 26 mV after 1000 h cycling. Moreover, the rate capability of the Li/40 s air-exposed Li₁₀GeP₂S₁₂/Li cell was evaluated in figure 2(b). When the current densities are set at 0.2 or 0.4 mA cm⁻², the polarization voltage is steady. However, gradual rise of polarization voltage was observed with increasing of current density to 0.6 and 0.8 mA cm⁻². Nevertheless, the Li/40 s air-exposed Li₁₀GeP₂S₁₂/Li cell can still stably cycle for 300 h at 0.2 mA cm⁻² under an areal capacity of 0.5 mAh cm⁻² (figure 2(c)).

Zoom In Zoom Out Reset image size

To understand the mechanism of the protective layer for improving the Li/Li₁₀GeP₂S₁₂ interface stability, the composition of the formed protective layer through air-exposure treatment was identified. Generally, most sulfide solid electrolytes are not stable to moisture because they easily react with H₂O by release of toxic H₂S gas [10, 22, 23]. Calpa et al [24] reported their DFT calculation results that the hydrolysis of Li₃PS₄ would decompose into Li₃PO₄ and H₂S. Ohtomo et al [9] experimentally detected the Li₃PO₄ phase after the 75Li₂S·25P₂S₅ electrolytes reacting with water. However, the reaction products of Li₁₀GeP₂S₁₂ and water are unclear. Through dissolving the Li₁₀GeP₂S₁₂ powder into water followed by vacuum drying at 150 °C for 3 h, the reaction products are determined to be Li₃PO₄ and Li₄GeO₄, as shown in figure S2. Noticeably, the Li₃PO₄ and Li₄GeO₄ were not observed on the surface of the air-exposed Li₁₀GeP₂S₁₂ pellet, indicating a different reaction process occur during the air-exposure treatment. As shown in figure 3(a), for the Li₁₀GeP₂S₁₂ pellet after 40 s air exposure, most of diffraction peaks can be indexed to Li₁₀GeP₂S₁₂ phase [25, 26]. Only three diffraction peaks at 12.58°, 14.60° and 17.56° belongs to Li₁₀GeP₂S₁₂ phase disappear and two new diffraction peaks at 16.07° and 16.99° were detected. After 30 min air exposure, large amounts of new diffraction peaks were observed, indicating new phases formed on the surface of Li₁₀GeP₂S₁₂ pellet. The diffraction peaks at 16.94°, 27.08°, 29.24°, 32.04°, 32.46°, 34.32°, 40.62° can be indexed to Li₄P₂S₆ (PDF#01-076-0992), while 15.48°, 16.94°, 26.42°, 28.82°, 31.38° belong to GeS₂ (PDF#00-030-0597) and 13.04°, 26.42°, 40.62° correspond to Li₂HPO₃ (PDF#00-035-0172). The XRD results suggest that a partial hydrolysis reaction occurs on the surface of the Li₁₀GeP₂S₁₂ pellet after air exposure with formation of the mixed phases of Li₄P₂S₆, GeS₂ and Li₂HPO₃, where the Li₄P₂S₆ is main phase as majority of diffraction peaks belong to it. The compound, Li₄P₂S₆, have been reported to exhibit considerable ionic conductivity and good interface stability towards lithium metal [27–30], making it desirable for interface protection for Li₁₀GeP₂S₁₂. Besides, both LiH₂PO₄ and Li₃PO₄ have been proven to be efficient protective layers for Li/solid electrolyte interface [16, 31–34], indicating phosphide, such as Li₂HPO₃, can stabilize the Li/Li₁₀GeP₂S₁₂ interface. Figures S3 and S4 presents the surface and cross-section morphology of Li₁₀GeP₂S₁₂ pellets before and after air exposure, showing the obvious decomposition layer coated on the surface of Li₁₀GeP₂S₁₂ pellet. Figure 3(b) presents schematic illustrations of the mechanism of the protective layer formed by air exposure for stabilizing the Li/Li₁₀GeP₂S₁₂ interface. For the Li/Li₁₀GeP₂S₁₂ interface, a mixed conductive interface is formed when Li₁₀GeP₂S₁₂ attaches with lithium metal, leading to continuously consume inner Li₁₀GeP₂S₁₂ and increase the cell impedance. After exposing Li₁₀GeP₂S₁₂ pellet in air, a passivating layer rapidly generated. On one hand, this layer can physically isolate the Li₁₀GeP₂S₁₂ from lithium metal. On the other hand, this layer comprising of the Li₄P₂S₆, GeS₂ and Li₂HPO₃ is lithium-ion permeable but electronic obstructed, which can effectively suppress the decomposition of Li₁₀GeP₂S₁₂ during the Li metal plating/striping.

Zoom In Zoom Out Reset image size

To further demonstrate the effect of the protective layer on stabilizing the Li/Li₁₀GeP₂S₁₂ interface, the LiCoO₂ based all-solid-state lithium metal batteries were fabricated by using both Li₁₀GeP₂S₁₂ and 40 s air-exposed Li₁₀GeP₂S₁₂ pellets as solid electrolyte. Figure 4(a) shows charge and discharge curves of the LiCoO₂/Li₁₀GeP₂S₁₂/Li battery, showing a rapid decay in specific capacity and large polarization after 5 cycles. In contrast, for the LiCoO₂/40 s air-exposed Li₁₀GeP₂S₁₂/Li battery, high reversible specific capacity and low degree of polarization are delivered (figure 4(b)). The LiCoO₂/40 s air-exposed Li₁₀GeP₂S₁₂/Li battery shows an initial charge specific capacity of 127 mA h g⁻¹ with high initial Columbic efficiency of 92% and delivered long cyclic stability for 100 cycles with capacity retention of 80%, as shown in figure 4(c). The good rate performances of the LiCoO₂/40 s air-exposed Li₁₀GeP₂S₁₂/Li battery were also presented (figures 4(d) and (e)), exhibiting the discharge capacities of 113, 87, 66, 46 mAh g⁻¹ at 0.1, 0.2, 0.5 and 1 C, respectively. The high reversible specific capacity and long cyclic stability strongly support the rapid generated protective layer by simple air-exposure treatment can effectively stabilize the Li/Li₁₀GeP₂S₁₂ interface.

Zoom In Zoom Out Reset image size

In summary, the strong reactive Li/Li₁₀GeP₂S₁₂ interface was effectively passivated via a rapid formed protective layer through simply exposing the Li₁₀GeP₂S₁₂ pellet to air for dozens of seconds. The protective layer coated on the surface of the Li₁₀GeP₂S₁₂ pellets is derived from the partial hydrolysis reaction of Li₁₀GeP₂S₁₂ in air, generating the decomposition products of Li₄P₂S₆, GeS₂ and Li₂HPO₃. This lithium-ion permeable but electronic obstructed layer can both separate the contact and effectively suppress the decomposition reaction between Li₁₀GeP₂S₁₂ and lithium metal during electrochemical cycling. After optimal air-exposure duration of 40 s, the Li/40 s air-exposed Li₁₀GeP₂S₁₂/Li symmetric cell presents long cyclic stability for 1000 h with small polarization voltage of 26 mV at 0.1 mA cm⁻². Compared with the LiCoO₂/Li₁₀GeP₂S₁₂/Li battery with rapid capacity decay after 10 cycles, the all-solid-state LiCoO₂/40 s air-exposed Li₁₀GeP₂S₁₂/Li battery shows long cyclic performances for 100 cycles with capacity retention of 80%, and good rate capabilities of discharge capacity of 113, 87, 66, 46 mAh g⁻¹ at 0.1, 0.2, 0.5 and 1 C, respectively.

The work was supported by the National Key R&D Program of China (Grant No. 2018YFB0905400), the National Natural Science Foundation of China (Grant Nos. U1964205, 51872303, 51902321 and 52172253), Zhejiang Provincial Key R&D Program of China (Grant No. 2022C01072), Ningbo S&T Innovation 2025 Major Special Programme (Grant Nos. 2019B10044 and 2021Z122) and Youth Innovation Promotion Association CAS (Y2021080).

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.

Wei Weng: investigation, methodology, formal analysis, data curation, visualization, writing–original draft. Dong Zhou: visualization, validation. Gaozhan Liu, Lin Shen: methodology, visualization. Mengqi Li, Xinshuang Chang: validation. Xiayin Yao: conceptualization, supervision, project administration, funding acquisition, resources, writing–review and editing.