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

Moist air is a great challenge for manufacturing sulfide-based all-solid-state lithium batteries as the water in air will lead to severe decomposition of sulfide electrolytes and release H2S gas. However, different with direct reaction with water, short-period air exposure of Li10GeP2S12 sulfide electrolyte with controlled humidity can greatly enhance the stability of Li10GeP2S12 against lithium metal, thus realizing stable Li10GeP2S12 based all-solid-state lithium metal batteries. During air exposure, partial hydrolysis reaction occurs on the surface of Li10GeP2S12 pellets, rapidly generating a protective decomposition layer of Li4P2S6, GeS2 and Li2HPO3 in dozens of seconds. This ionically conductive but electronically insulation protecting layer can effectively prevent the severe interface reaction between Li10GeP2S12 and lithium metal during electrochemical cycling. The Li/40s-air-exposed Li10GeP2S12/Li cell shows long cycling stability for 1000 h. And the LiCoO2/40s-air-exposed Li10GeP2S12/Li batteries present good rate capability and long cyclic performances, showing capacity retention of 80% after 100 cycles.


Introduction
All-solid-state lithium batteries exhibit increased safety due to the employment of nonflammable inorganic solid electrolytes [1][2][3]. To date, various inorganic solid electrolytes, especially sulfide solid electrolytes, possessing high ionic conductivity have been developed [4][5][6][7][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 H 2 S 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, Li 10 GeP 2 S 12 possesses high ionic conductivity of 12 mS cm −1 , which exceeds to that of organic liquid electrolytes [4]. However, the sulfide solid electrolyte Li 10 GeP 2 S 12 strongly reacts with lithium metal and decomposes to Li 15 Ge 4 , Li 3 P and Li 2 S [13][14][15]. The large fraction of metallic Li 15 Ge 4 at the Li/Li 10 GeP 2 S 12 interface can create electronically conducting pathways, which will continuously consume the bulk Li 10 GeP 2 S 12 and increase impedance until cell failure. Thus, a passivating layer without electronic conductivity is crucial to suppress the highly reactive Li/Li 10 GeP 2 S 12 interface. Modification of lithium metal through chemical reaction is an appealing strategy to stabilize the Li/Li 10 GeP 2 S 12 interface. Zhang et al [16] identified the in-situ formation of the LiH 2 PO 4 by reacting the lithium metal with H 3 PO 4 . This LiH 2 PO 4 layer is ionically conducting but electronically insulating, which can prevent the direct contact between Li 10 GeP 2 S 12 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/Li 10 GeP 2 S 12 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 Li 10 GeP 2 S 12 . In addition, using bilayer composite electrolyte can also passivate the Li/Li 10 GeP 2 S 12 interface [18][19][20]. The cells employing Li-argyrodites Li 5.5 PS 4.5 Cl 1.5 or Li 10 GeP 2 S 12 exhibit distinct failure behavior toward lithium metal due to short circuit by lithium dendrite penetration for Li 5.5 PS 4.5 Cl 1.5 and increased overpotential by electrolyte decomposition for Li 10 GeP 2 S 12 . The bilayer construction of Li/Li 5.5 PS 4.5 Cl 1.5 /Li 10 GeP 2 S 12 shows excellent interface stability even at high current density, in which the Li 5.5 PS 4.5 Cl 1.5 layer as buffer layer is to isolate Li 10 GeP 2 S 12 from lithium metal and the Li 10 GeP 2 S 12 can prevent the lithium dendrite penetration [20]. Clearly, designing an artificial passivation interface layer by more convenient method is crucial to stabilize Li/Li 10 GeP 2 S 12 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 shortperiod air exposure of sulfide electrolytes with controlled humidity could provide effective passivation against lithium metal. The dramatically improved interface stability between Li 10 GeP 2 S 12 solid electrolyte and lithium metal is achieved by simply exposing the Li 10 GeP 2 S 12 pellet into air for dozens of seconds. This air-exposure treatment could rapidly generate a protective layer of Li 4 P 2 S 6 , GeS 2 and Li 2 HPO 3 coated on the surface of the Li 10 GeP 2 S 12 pellets. This protecting layer is ionically conductive but electronically insulation, which can not only physically isolate the contact between Li 10 GeP 2 S 12 and lithium metal but also effectively suppress the continuous decomposition of Li 10 GeP 2 S 12 reduced by lithium metal.

Preparation of Li 10 GeP 2 S 12 pellets with and without air-exposure treatment
The synthesis of Li 10 GeP 2 S 12 solid electrolytes can be found elsewhere [21]. The room temperature ionic conductivity of 6.13 × 10 −3 S cm −1 and its X-ray diffraction (XRD) pattern is shown in figure S1 (available online at stacks.iop.org/MF/1/021001/mmedia). The Li 10 GeP 2 S 12 pellet (10 mm diameter, ∼1 mm thickness) was prepared by cold pressing ∼150 mg of Li 10 GeP 2 S 12 powder under 180 MPa. For the air-exposure treatment of Li 10 GeP 2 S 12 electrolytes, both side of the Li 10 GeP 2 S 12 pellet were separately exposed to air in a constant temperature of 30 • C and humidity chamber with 45% humidity for different durations. Before one side of the Li 10 GeP 2 S 12 pellet was exposed to air, the other side was sealed to avoid secondary exposure. Air-exposed Li 10 GeP 2 S 12 electrolytes with duration of 40 s is labeled as 40 s air-exposed Li 10 GeP 2 S 12 .

Materials characterization
To identify the composition of interfacial layer of the air-exposed Li 10 GeP 2 S 12 pellets, XRD measurements were performed on Bruker D8 Advance Diffractometer with Cu Kα radiation (λ = 1.54178 Å). The EIS measurements for the symmetric cells were tested using Solartron 1470E electrochemical workstation (Solartron Public Co., Ltd) from 1 MHz to 0.1 Hz under 10 mV at 25 • C. Surface and crosssection morphology of Li 10 GeP 2 S 12 pellet before and after air exposure were investigated by a scanning electron microscope (Regulus-8230, Hitachi).

Electrochemical performance measurements
The lithium metal foils with thickness of 80 µm were used as electrode to assemble the symmetric cells and solid-state lithium metal batteries. To prepare the Li metal symmetric cells, two pieces of metallic lithium foils were attached on both sides of the electrolyte pellet and vacuum sealed in a pouch bag. Then the cells were isostatically pressed under 50 MPa for 5 min. The stainless steel attached with nickel tag was used as current collector. For testing the impedance of the symmetric cells after cycling tests, the galvanostatic Li plating/stripping was performed at 0.1 mA cm −2 and 0.1 mAh cm −2 under 25 • C. To fabricate the all-solid-state lithium metal batteries, the composite cathode was prepared by mixing LiNbO 3coated LiCoO 2 and Li 10 GeP 2 S 12 powders with 70:30 weight ratio. The composite cathode (∼2 mg cm −2 ) powder was spread on one side of Li 10 GeP 2 S 12 pellet and pressed at 180 MPa to obtain the integrated cathode-electrolyte pellet. The lithium metal was attached on the other side of electrolyte pellet and sealed in pouch bag. Before assembling the airexposed Li 10 GeP 2 S 12 based solid-state batteries, the side of Li 10 GeP 2 S 12 pellet integrated with cathode powder was sealed to avoid exposing to air. Charge/discharge measurements were conducted between 3.0 and 4.2 V at 25 • C using a multichannel battery test system (LAND CT-2001A, Wuhan Rambo Testing Equipment Co., Ltd).

Results
As shown in figure 1(a), in contrast with the nearly liner increasing Li plating/striping voltage for the symmetric Li/Li 10 GeP 2 S 12 /Li cell without air-exposure treatment, the Li/air-exposed Li 10 GeP 2 S 12 /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 Li 10 GeP 2 S 12 /Li cells was directly observed during the electrochemical cycling. Figure 1(b) presents the impedance of the Li/Li 10 GeP 2 S 12 or air-exposed Li 10 GeP 2 S 12 /Li cells for the cycling tests. It can be clearly found that the continuously increasing impedance arising from strong reaction at Li/Li 10 GeP 2 S 12 interface was dramatically suppressed when the air-exposure treatment for Li 10 GeP 2 S 12 pellets was employed. Figure 1(c) presents the voltage of the symmetric Li/Li 10 GeP 2 S 12 or air-exposed Li 10 GeP 2 S 12 /Li cells before and after cycling. The Li/Li 10 GeP 2 S 12 /Li symmetric cell shows much higher polarization voltage than that of airexposed Li 10 GeP 2 S 12 based symmetric cells before cycling, which is due to the decomposition reaction occurred once the Li 10 GeP 2 S 12 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/Li 10 GeP 2 S 12 interface. Whereas, the impedance of the Li/air-exposed Li 10 GeP 2 S 12 /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 Li 10 GeP 2 S 12 /Li cells for cycling tests (∆ R t = R t, (400 h) − R t, (0 h) ) were evaluated. As presented in figure 1(e), the symmetric cells using Li 10 GeP 2 S 12 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]. Figure 2(a) shows the galvanostatic Li plating/striping of the Li/40 s air-exposed Li 10 GeP 2 S 12 /Li and the Li/Li 10 GeP 2 S 12 /Li cells. The Li/Li 10 GeP 2 S 12 /Li cell shows a rapid increase in polarization voltage due to the continuous and deteriorative Li/Li 10 GeP 2 S 12 interface reaction. In contrast, the Li/40 s air-exposed Li 10 GeP 2 S 12 /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 Li 10 GeP 2 S 12 /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 Li 10 GeP 2 S 12 /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)).
To understand the mechanism of the protective layer for improving the Li/Li 10 GeP 2 S 12 interface stability, the composition of the formed protective layer through air-exposure Figure 1. Evaluation of the Li/Li 10 GeP 2 S 12 interface stability through Li plating/striping experiments and evolution of impedance for the symmetric Li/Li 10 GeP 2 S 12 or air-exposed Li 10 GeP 2 S 12 /Li cells. (a) Cyclic performance of the Li/Li 10 GeP 2 S 12 or air-exposed Li 10 GeP 2 S 12 /Li cells at 0.1 mA cm −2 and 0.1 mAh cm −2 . (b) EIS plots of the symmetric Li/Li 10 GeP 2 S 12 or air-exposed Li 10 GeP 2 S 12 /Li cells before cycling and after different cycling time. (c) Li plating/striping voltages of the Li/Li 10 GeP 2 S 12 or air-exposed Li 10 GeP 2 S 12 /Li cells before and after cycling. (d) Total impedance (Rt) of the Li/air-exposed Li 10 GeP 2 S 12 /Li cells before cycling. (e) Difference value of the total impedance of the Li/air-exposed Li 10 GeP 2 S 12 /Li cells for cycling tests (∆ Rt = R t (400 h) − R t (0 h) ). (f) Evolution of the total impedance (Rt) of the Li/40 s air-exposed Li 10 GeP 2 S 12 /Li cells for cycling tests. treatment was identified. Generally, most sulfide solid electrolytes are not stable to moisture because they easily react with H 2 O by release of toxic H 2 S gas [10,22,23]. Calpa et al [24] reported their DFT calculation results that the hydrolysis of Li 3 PS 4 would decompose into Li 3 PO 4 and H 2 S. Ohtomo et al [9] experimentally detected the Li 3 PO 4 phase after the 75Li 2 S·25P 2 S 5 electrolytes reacting with water. However, the reaction products of Li 10 GeP 2 S 12 and water are unclear. Through dissolving the Li 10 GeP 2 S 12 powder into water followed by vacuum drying at 150 • C for 3 h, the reaction products are determined to be Li 3 PO 4 and Li 4 GeO 4 , as shown in figure S2. Noticeably, the Li 3 PO 4 and Li 4 GeO 4 were not observed on the surface of the air-exposed Li 10 GeP 2 S 12 pellet, indicating a different reaction process occur during the air-exposure treatment. As shown in figure 3(a), for the Li 10 GeP 2 S 12 pellet after 40 s air exposure, most of diffraction peaks can be indexed to Li 10 GeP 2 S 12 phase [25,26]. Only three diffraction peaks at 12.58 • , 14.60 • and 17.56 • belongs to Li 10 GeP 2 S 12 phase disappear and two new diffraction peaks at 16.07 • and 16.99 • were detected. After 30 min The XRD results suggest that a partial hydrolysis reaction occurs on the surface of the Li 10 GeP 2 S 12 pellet after air exposure with formation of the mixed phases of Li 4 P 2 S 6 , GeS 2 and Li 2 HPO 3 , where the Li 4 P 2 S 6 is main phase as majority of diffraction peaks belong to it. The compound, Li 4 P 2 S 6 , have been reported to exhibit considerable ionic conductivity and good interface stability towards lithium metal [27][28][29][30], making it desirable for interface protection for Li 10 GeP 2 S 12 . Besides, both LiH 2 PO 4 and Li 3 PO 4 have been proven to be efficient protective layers for Li/solid electrolyte interface [16,[31][32][33][34], indicating phosphide, such as Li 2 HPO 3 , can stabilize the Li/Li 10 GeP 2 S 12 interface. Figures  S3 and S4 presents the surface and cross-section morphology of Li 10 GeP 2 S 12 pellets before and after air exposure, showing the obvious decomposition layer coated on the surface of Li 10 GeP 2 S 12 pellet. Figure 3(b) presents schematic illustrations of the mechanism of the protective layer formed by air exposure for stabilizing the Li/Li 10 GeP 2 S 12 interface. For the Li/Li 10 GeP 2 S 12 interface, a mixed conductive interface is formed when Li 10 GeP 2 S 12 attaches with lithium metal, leading to continuously consume inner Li 10 GeP 2 S 12 and increase the cell impedance. After exposing Li 10 GeP 2 S 12 pellet in air, a passivating layer rapidly generated. On one hand, this layer can physically isolate the Li 10 GeP 2 S 12 from lithium metal. On the other hand, this layer comprising of the Li 4 P 2 S 6 , GeS 2 and Li 2 HPO 3 is lithium-ion permeable but electronic obstructed, which can effectively suppress the decomposition of Li 10 GeP 2 S 12 during the Li metal plating/striping.
To further demonstrate the effect of the protective layer on stabilizing the Li/Li 10 GeP 2 S 12 interface, the LiCoO 2 based allsolid-state lithium metal batteries were fabricated by using both Li 10 GeP 2 S 12 and 40 s air-exposed Li 10 GeP 2 S 12 pellets as solid electrolyte. Figure 4(a) shows charge and discharge curves of the LiCoO 2 /Li 10 GeP 2 S 12 /Li battery, showing a rapid decay in specific capacity and large polarization after 5 cycles. In contrast, for the LiCoO 2 /40 s air-exposed Li 10 GeP 2 S 12 /Li battery, high reversible specific capacity and low degree of polarization are delivered ( figure 4(b)). The LiCoO 2 /40 s air-exposed Li 10 GeP 2 S 12 /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 LiCoO 2 /40 s air-exposed Li 10 GeP 2 S 12 /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/Li 10 GeP 2 S 12 interface. Electrochemical performances of all-solid-state lithium batteries. Charge and discharge curves of (a) the LiCoO 2 /Li 10 GeP 2 S 12 /Li battery and (b) the LiCoO 2 /40 s air-exposed Li 10 GeP 2 S 12 /Li battery at 0.1 C (1 C = 120 mA g −1 ) under 25 • C. (c) Cyclic performances of the LiCoO 2 /Li 10 GeP 2 S 12 /Li and LiCoO 2 /40 s-air-exposed Li 10 GeP 2 S 12 /Li batteries at 0.1 C under 25 • C. (d) Charge and discharge curves and (e) cyclic performances of the LiCoO 2 /40 s-air-exposed Li 10 GeP 2 S 12 /Li battery at different rates.

Conclusion
In summary, the strong reactive Li/Li 10 GeP 2 S 12 interface was effectively passivated via a rapid formed protective layer through simply exposing the Li 10 GeP 2 S 12 pellet to air for dozens of seconds. The protective layer coated on the surface of the Li 10 GeP 2 S 12 pellets is derived from the partial hydrolysis reaction of Li 10 GeP 2 S 12 in air, generating the decomposition products of Li 4 P 2 S 6 , GeS 2 and Li 2 HPO 3 . This lithium-ion permeable but electronic obstructed layer can both separate the contact and effectively suppress the decomposition reaction between Li 10 GeP 2 S 12 and lithium metal during electrochemical cycling. After optimal air-exposure duration of 40 s, the Li/40 s air-exposed Li 10 GeP 2 S 12 /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 LiCoO 2 /Li 10 GeP 2 S 12 /Li battery with rapid capacity decay after 10 cycles, the all-solid-state LiCoO 2 /40 s air-exposed Li 10 GeP 2 S 12 /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.