Polymer Electrolyte/Sulfur Double‐Shelled Anisotropic Reduced Graphene Oxide Lamellar Scaffold Enables Stable and High‐Loading Cathode for Quasi‐Solid‐State Lithium‐Sulfur Batteries

Abstract Lithium‐sulfur batteries (LSBs) can replace lithium‐ion batteries by delivering a higher specific capacity. However, the areal capacity of current LSBs is low because the intrinsic limitations of sulfur make achieving a high sulfur loading difficult. Herein, the authors report vertically aligned reduced graphene oxide (rGO) with sulfur and poly(ethylene oxide)‐based polymer electrolyte double‐shell layers (VRG@S@PPE) as a high‐loading sulfur cathode. The addition of vapor‐grown carbon fiber (VGCF) into rGO is the key to success, as it allows for gas evacuation from internal nano/micropores without structural collapse, enabling perfect double‐shell layer contact. Owing to the anisotropic rGO lamellar structure that enables straightforward ion/electron transport and provides numerous active sites, sulfur‐infiltrated rGO reinforced via VGCF (VRG@S) exhibits a high capacity of 998 mAh g−1 after 100 cycles at 0.1 C under high sulfur loading (6 mg cm−2). Interestingly, an additional polymer electrolyte layer further increases the cycle retention (1005 and 718 mAh g−1 after 100 cycles at 0.1 and 1 C, respectively), because intimate contact between the solid polymer electrolyte and sulfur could suppress the loss of sulfur due to lithium polysulfide (LPS) shuttling or volume change during lithiation/delithiation. Therefore, it is possible to realize safe and stable quasi‐solid‐state LSBs with high sulfur loading.

First, a suspension in which sulfur and a conductive carbon agent were dispersed in several solvents (ethanol, NMP, and CS2) was directly dropped onto the VRG foam. As shown in Figure   S1a-c, most of the active materials aggregated on the surface and hardly penetrated into the pores of the VRG foam. To achieve efficient infiltration of sulfur into the VRG foam, we examined other approaches with CS2 solvent, because the lithium sulfur battery (LSB) using the cathode prepared with CS2 showed a higher initial specific capacity (1000 mAh g -1 ) in the charge/discharge curve than those using cathodes with other solvents (Figure S1d-f). Although the sulfur/CS2 solution penetrated well into the VRG foam without conductive carbon agents using the simple drop method ( Figure. S2a), the initial specific capacity was very poor (300 mAh g -1 ) at a current density of 0.1 C ( Figure S2d). This result indicates that conductive carbon agents should be included in the suspension to address the poor electrical conductivity of elemental sulfur. To ensure penetration of the sulfur-carbon suspension (S-C/CS2) into the foam, a decompression method using a syringe was utilized. As a result, the sulfur and carbon particles penetrated the pores of the VRG foam well ( Figure S2b). The initial specific capacity of the LSB also increased to 800 mAh g -1 ( Figure S2e). This implies that the air trapped in the micropores must be removed. Thus, vacuum filtration was performed to achieve sufficient decompression. By sucking air into the bottom, the S-C/CS2 suspension fully penetrated the VRG foam ( Figure S2c). The specific capacities after the first and second cycles were much higher (1000 and 800 mAh g -1 , respectively) than those of the syringe method ( Figure S2f).
However, a significant amount of sulfur and carbon materials were located at the bottom. Therefore, sonication was performed to ensure uniform distribution of sulfur and carbon within the foam and to eliminate air in the pores (Figure 3a).

Supplementary data 3. Coupling type between VGCF and rGO in VRG foam
In order to verify type of coupling between VGCF and rGO, FT-IR analysis was conducted for the bare rGO, VGCF and VRG ( Figure S5). The spectra of three samples were observed between 4000 and 650 cm -1 . The related peaks were appeared around 850, 1150, 1400 and 3400 cm -1 corresponded to C-C, C-O, C-O-C and O-H stretching, respectively. [42][43] The VGCF exhibited only C-C stretching peak around 1400 cm -1 . In contrast, the bare rGO and VRG showed all 4 peaks above-mentioned, and intensity of peaks exhibited that VRG was lower than bare rGO. Generally, when carbon-nanotube and rGO is chemically bonded, O-H stretching around 3400 cm -1 reaches down to a lower-wavelength (3000~3100 cm -1 ), making a shoulder peak. [44] However, such shoulder peak hardly observed in spectra of VRG. It indicates that there is no hydrogen bonding between VGCF and rGO due to the absence of the C-O-H functional group on the surface of VGCF. [45] Therefore, it suggests that the type of coupling between VGCF and rGO is purely mechanical mixing without chemical binding.

Supplementary data 4. Effect of PEO-based polymer electrolyte (PPE) infiltration methods on the structure and electrochemical properties of VRG@S@PPE cathodes
PPE was coated onto the surface of the VRG@S foam. Inspired by the success of sulfur infiltration, various methods were evaluated to find the best way to infiltrate PPE into the VRG@S foam. Direct melting of the PPE film can penetrate a large amount of PPE into the foam without void spaces. However, the harsh environment, i.e., high temperature and vacuum conditions, required for the direct melting of PPE induced the sublimation of sulfur in the VRG@S foam. The foam structure also collapsed because of an excess of PPE. Thus, PPE was dissolved in acetonitrile and then coated onto the surface of the VRG@S foam ( Figure S8).
First, the PPE solution was dropped directly onto the VRG@S foam. As a result, PPE hardly penetrated into the foam and mostly accumulated on the surface, blocking almost all the pores ( Figure S8a). Accordingly, the capacity was as low as approximately 300 mAh g -1 at a rate of 0.1 C ( Figure S8d), because sulfur utilization was reduced by the thick PPE layer deposited on the electrode surface. As with the previous sulfur infiltration approach, decompression was required to penetrate PPE into the foam without accumulation on the surface. Thus, decompression using a syringe was carried out for PPE infiltration. The PPE solution was forcibly penetrated into the foam while removing air in the pores through decompression using a syringe. Although this method better than the simple drop method, a large amount of PPE still barely infiltrated the foam and was accumulated on the surface ( Figure S8b). Although the discharge capacity of the corresponding cathode was relatively at approximately 1000 mAh g -1 after the first cycle, the charging process was insecure, and subsequently, overcharging occurred during the second cycle ( Figure S8e). Therefore, to infiltrate the PPE more effectively, the VRG@S foam was soaked in PPE solution and exposed to a vacuum environment.
Consequently, the amount of PPE accumulated on the surface was significantly reduced, and the porous surface of the foam was maintained ( Figure S8c). The resulting cathode showed a considerably high capacity (1600 mAh g -1 ) during the first cycle of the LSB (Figure S8f), although poor capacity retention was observed owing to the thick PPE coating. The capacity retention was improved by reducing the PPE coating amount from 30 to 10 mg ( Figure S9).
The capacity of 10 mg PPE-coated VRG@S only reduced from 1200 to 1100 mAh g -1 during the first three cycles and remained at that value even after the next 10 cycles, which indicates that the initial capacity retention was better than that of bare VRG@S foam without PPE.   Here, the infiltrated PPE mass were fixed to 30 mg, to confirm the advantage of sonication method more clearly. When the soaking/vacuum filtration method was used, a large amount of PPE that had not penetrated was accumulated on the surface and blocked the pores. In addition, PPE was non-uniformly aggregated inside the structure. In contrast, not only PPE was hardly stacked but also pores were clearly maintained without blocking in surface of VRG via the sonication method, where PPE which infiltrated internal structure was also uniformly coated on rGO sheet without agglomeration.