Pristine or Highly Defective? Understanding the Role of Graphene Structure for Stable Lithium Metal Plating

Abstract

We are the first to examine the role of graphene host structure/chemistry in plating-stripping in lithium metal anodes employed for lithium metal batteries (LMBs). Structural and chemical defects are bad since highly defective graphene promotes unstable solid electrolyte interphase (SEI) growth. This consumes the FEC additive in the carbonate electrolyte and is correlated with rapid decay in CE and formation of filament-like Li dendrites. A unique flow-aided sonication exfoliation method is employed to synthesize "defect-free" graphene (df-G), allowing for a direct performance comparison with conventional reduced graphene oxide (r-GO). At cycle 1, the r-GO is better electrochemically wetted by Li than df-G, indicating that initially it is more lithiophilic. With cycling, the nucleation overpotential with r-GO becomes higher than with df-G, indicating less facile plating reactions. The df-G yields state-of-the-art electrochemical performance; stable plating at 0.5 - 4 mA/cm2, areal capacity up to 2 mAh/cm2, cycle 1 CE at 98% and cycle 100 CE at 94%. With df-G the post cycled metal surface is relatively smooth and dendrite-free. Conversely, r-GO templates have CE rapidly degrade from the onset, with extensive dendrites after cycling. Extensive SEI growth and associated FEC depletion with r-GO are further confirmed by electrochemical impedance analysis and surface science methods (XPS). We therefore propose the following design rule for next-generation supports for LMBs: An ideal architecture will promote copious heterogenous nucleation of the plating metal, shielding it from the electrolyte. In addition, it is essential that the host is itself non-catalytic towards SEI formation.