Building electrode skins for ultra-stable potassium metal batteries

Metal anodes promise high energy density due to their low redox potentials and high specific capacities, but their practical use is hampered by unstable solid electrolyte interphases (SEI), severe dendrite growth, and large volume changes during plating/stripping that cause interfacial rupture, low Coulombic efficiency, and safety risks. Prior approaches, including 3D hosts, alloy anodes, electrolyte and solvation optimization, and additive strategies, can partially mitigate dendrites but often suffer from limited mechanical robustness of the resulting SEI, scalability issues, and cost barriers. Inspired by the human skin—where an epidermis provides a protective barrier and a dermis offers elasticity and ionic transport—the study proposes a dual-layer metal electrode skin (MES) that increases surface flatness to homogenize electric fields and promotes in situ formation of a fluorine-rich, mechanically strong SEI. The research question is whether such a biomimetic MES can stabilize K metal interfaces, suppress dendrites, and enable ultra-stable cycling in symmetric, asymmetric, and full potassium metal batteries.

The paper situates itself within efforts to stabilize metal anodes by engineering electrodes and electrolytes. Reported strategies include constructing 3D current collectors or hosts to relieve volume change and homogenize ion flux, designing liquid alloy anodes, and tuning electrolyte chemistry via solvent/solvation modification, additives, and concentration adjustments. Although these approaches can yield more stable initial interfaces and reduce dendrite formation, the naturally formed SEI typically remains mechanically weak and prone to fracture during cycling, leading to continuous side reactions, fresh surface exposure, and electrolyte consumption. A tribo-electrochemistry-induced artificial SEI with amorphous KF has been shown to improve K+ diffusion and inhibit dendrites, but faces challenges in cycle life and large-scale preparation. Overall, the literature highlights the critical role of robust, mechanically resilient SEI and uniform interfacial fields, yet scalable, cost-effective solutions with long lifetimes are still lacking.

Materials synthesis: Fluorinated graphene oxide (F-GO) was synthesized via a mild liquid-phase reaction of graphene oxide (GO) with diethylaminosulfur trifluoride (DAST). In brief, GO (0.15 g) was dispersed in o-dichlorobenzene (DCB, 150 mL) by overnight stirring. DAST (2.5 mL) was added dropwise at 0 °C over 10 min, followed by 6 h sonication and 3 days stirring at ambient temperature. The reaction was quenched with methanol (100 mL), and the product was filtered, extensively washed with ethanol, acetone, and DI water, and vacuum-dried at 60 °C for 12 h. Construction of Cu@F-GO (MES host on Cu): F-GO was dispersed in ethanol, sonicated for 60 min, and transferred to Cu foil via a Langmuir–Blodgett-like process by injecting the suspension onto water, self-assembling a film, and lifting immersed Cu foils to collect films; repeated lifts obtained desired thickness. Films were dried at 80 °C (hot plate 3 min) and in vacuum overnight. Cu@GO was prepared similarly. Transfer to K (K@F-GO, i.e., K@MES): Lump K was rolled between plastic films to obtain uniform K foil. Cu@F-GO and K foil were stacked between plastic films and roll-pressed (roll distance ~80% of total stack thickness, speed 0.1 cm s⁻¹); the Cu was peeled off to leave F-GO on K (K@F-GO). K@GO and bare K controls were prepared analogously. Cell assembly and electrolytes: Symmetric cells (K||K) and asymmetric cells (Cu variants || K) used coin cells (CR2032), glass fiber separators (Whatman GF/D), and electrolytes of either 3 M KFSI in DME or 0.8 M KPF6 in EC:DMC (1:1). PB cathodes were prepared (active:carbon:PVDF = 6:3:1 in NMP), coated on Al, dried (80 °C, 24 h), and punched (1.2 cm). K anode thickness ~200 µm; PB loading ~1.5 mg cm⁻². Pouch cells used ~3×4 cm electrodes, controlled electrolyte volume (~600 µL), and N/P ~4. Electrochemical testing: Galvanostatic cycling used Neware BTS; CV, Tafel, and EIS (10⁻²–10⁵ Hz) were measured on a CHI660e workstation (CV/Tafel in −0.2 to 0.2 V at 2 mV s⁻¹). Symmetric cells were tested at various current densities (e.g., 0.5–3.5 mA cm⁻²) and capacities (0.5–3.5 mAh cm⁻²). Asymmetric cells for Coulombic efficiency used 0.5 mA cm⁻² and 0.5 mAh cm⁻². Full cells were tested at 500–3000 mA g⁻¹, including long-term cycling at 1000 mA g⁻¹. Temperature was controlled at 28 °C. Characterization: Surface morphology and cross-sections were examined by FESEM; TEM/HRTEM evaluated SEI thickness and structure; EDS mapped elemental distributions. XPS (survey and high-resolution C 1s, O 1s, F 1s) and FTIR probed GO/F-GO chemistry and fluorine evolution before/after cycling; depth profiling examined F distribution in SEI. AFM assessed surface roughness/flatness. Contact angles measured wettability in both electrolyte systems. In situ optical microscopy observed K deposition dynamics. An infrared thermal imager recorded heat release when K electrodes were exposed to air. Stability against electrolyte penetration was evaluated by potential holding at 4 V vs. K+/K for Cu@MES||K and bare Cu||K. Computation and simulation: DFT modeled K interaction with fluorinated graphene (graphene fluoride and F-GO) to assess C–F bond cleavage energetics and resulting KF formation and graphene reconstruction (sp³ to sp²). 3D Comsol Multiphysics electric currents module simulated electric field distribution and K deposition near rough vs. MES-flattened surfaces; simulations included ion migration under bias and conversion to metal upon surface contact, with deposition continuing from randomized positions until a preset height was reached. Universality tests: The MES fabrication and testing protocol was applied to Li and Na metal symmetric and asymmetric cells using corresponding electrolytes (LiFSI/LiPF6; NaFSI/NaPF6) and polypropylene separators, demonstrating performance generality.

• MES concept: A dual protective system mimicking skin—an artificial F-GO film (epidermis) increases surface flatness and wettability, homogenizing electric fields, while in situ fluorine release during K plating forms a fluoride-rich, mechanically robust SEI (dermis) that passivates the metal and resists cracking. • Surface flatness and field homogenization: AFM showed that F-GO coverage greatly increased surface flatness vs. bare Cu/K; Comsol simulations revealed that MES reduces tip-enhanced fields and produces more uniform near-surface K+ distributions, promoting uniform deposition. • In situ SEI strengthening via fluorine release: DFT predicted spontaneous C–F bond cleavage upon K contact with energy release of ~3.87 eV per K, forming KF and converting F-GO toward graphene (sp³→sp²). XPS/FTIR confirmed fluorine loss from F-GO after cycling and fluoride incorporation into SEI; depth profiling showed relatively uniform F content through the SEI thickness on MES vs. lower F on GO. • Mechanical robustness of SEI: Nanoindentation-derived modulus of SEI on MES increased with depth to ~5.3, 8.9, and 14.2 GPa at 5, 6, and 7 nm, respectively, vs. ~2.8, 3.6, and 4.5 GPa on GO, indicating substantially enhanced mechanical strength. • Electrochemical kinetics and impedance: CV showed higher current response for K@MES symmetric cells, Tafel analysis indicated higher exchange current density, and EIS revealed smaller and more stable interfacial impedance during cycling compared with K@GO and bare K. • Symmetric cell longevity: At 0.5 mA cm⁻² and 0.5 mAh cm⁻², K@MES symmetric cells cycled for ~2300 h (3 M KFSI/DME). K@GO failed at ~670 h and bare K at ~160 h. In 0.8 M KPF6 (EC/DMC), K@MES also outperformed controls, demonstrating electrolyte universality. Rate tests up to 3.5 mA cm⁻² and 3.5 mAh cm⁻² showed lower overpotential and stable operation for K@MES. • Asymmetric cell stability: Cu@MES||K achieved >3200 h (>1600 cycles) at 0.5 mA cm⁻² and 0.5 mAh cm⁻² with stable plating/stripping and lower nucleation overpotential—the longest reported life among comparable PMBs. Cu@GO||K and Cu||K survived ~300 and ~60 cycles, respectively, with sharp CE decline thereafter. • Dendrite suppression and morphology: In situ optical microscopy showed rapid dendrite initiation on bare K (within 300 s) and fewer but growing dendrites on K@GO; K@MES remained dendrite-free even at 600 s. SEM cross-sections of 2–5 mAh cm⁻² K deposits on Cu illustrated dense, uniform, dendrite-free layers on Cu@MES vs. porous, dendritic deposits on Cu@GO/bare Cu. • Electrolyte barrier and safety: Potential-hold tests at 4 V vs. K+/K showed Cu@MES||K maintained 4 V at near-zero current (electrolyte did not penetrate F-GO), while bare Cu corroded (current surged). Infrared thermography in air showed MES slowed exothermic reactions of K vs. bare K, indicating improved handling safety. • Full-cell performance (K@MES||Prussian blue): At 500 mA g⁻¹, K@MES delivered >110 mAh g⁻¹ after 500 cycles with lower polarization; bare K decayed to ~58 mAh g⁻¹; K@GO showed declining CE after ~500 cycles. Long-term cycling at 1000 mA g⁻¹ exhibited >5000 stable cycles with CE >98%. Rate capability with K@MES achieved ~110, 93, 81, and 75 mAh g⁻¹ at 500, 1000, 2000, and 3000 mA g⁻¹, respectively. Pouch cells (N/P ~4, ~600 µL electrolyte) cycled stably for 100 cycles at 1000 mA g⁻¹ with stable voltage profiles. • Generality: Applying MES to Li and Na metal symmetric/asymmetric cells yielded markedly extended lifetimes and reduced polarization even with conventional low-concentration electrolytes, underscoring broad applicability. • Scalability: The Langmuir–Blodgett assembly and roll-transfer process are simple and scalable, indicating potential for practical MES fabrication.

The biomimetic MES directly addresses the central challenge of K metal anodes—interfacial instability that fosters dendrites and rapid failure—by combining two synergistic protections. First, the F-GO film smooths the electrode, homogenizing interfacial electric fields and K+ flux to prevent tip-enhanced growth and enable uniform nucleation. Second, during initial plating, K cleaves C–F bonds in F-GO, supplying fluorine to form a KF-rich, inorganic-dominant SEI with significantly higher mechanical modulus and uniform composition through the depth, which resists cracking during volume changes. Together, these effects minimize continual SEI reformation and electrolyte consumption, thereby lowering interfacial impedance, reducing nucleation overpotentials, and stabilizing long-term cycling across electrolytes. The dramatic gains—a symmetric-cell life of ~2300 h, record asymmetric-cell life (>1600 cycles), and full-cell life >5000 cycles with CE >98%—demonstrate that robust interfacial mechanics and field homogenization are critical design levers for practical PMBs. The successful translation to Li and Na metals suggests that MES is a general interfacial engineering platform. The straightforward, scalable fabrication further enhances the relevance to real-world battery manufacturing, including pouch cells. Collectively, the findings advance understanding of SEI mechanics and interfacial physics in metal batteries and propose a viable pathway to safer, longer-lived high-energy cells.

The study introduces a skin-inspired metal electrode skin (MES) that combines an artificial F-GO top layer with an in situ formed fluoride-rich SEI to stabilize K metal anodes. By increasing surface flatness and homogenizing electric fields while simultaneously generating a mechanically robust, inorganic SEI, MES suppresses dendrites, reduces interfacial impedance, and enables ultra-stable cycling: ~2300 h in symmetric K cells, >1600 cycles in Cu@MES||K asymmetric cells, and >5000 cycles with CE >98% in K@MES||PB full cells alongside strong rate performance and pouch-cell validation. The approach is simple, scalable, and generalizable to Li and Na metal systems. Future work may explore optimization of MES thickness/composition, long-term chemical stability across broader temperatures and electrolytes, integration with high-loading cathodes under lean-electrolyte conditions, and manufacturing scale-up for large-format cells.