No abstract

Lithium metal batteries are one of the promising technologies for future energy storage. One open challenge is the generation of a stable and well performing Solid Electrolyte Interphase (SEI) between lithium metal and electrolyte. Understanding the complex interaction of reactions at the lithium surface and the resulting SEI is crucial for knowledge-driven improvement of the SEI. This study reveals the internal species distribution and geometrical aspects of the native SEI during formation by model-based analysis. To achieve this, a combination of molecular dynamics, density functional theory, and stand-alone 3D-kinetic Monte Carlo simulations is used. The kinetic Monte Carlo model determines the SEI growth features over a long time and length scale so that the SEI can be analyzed quantitatively. The simulation confirms the frequently postulated layered SEI structure arising from the decomposition of an ethylene carbonate/lithium hexafluorophosphate (2 M) electrolyte with lithium metal. These layers are not clearly separated, which is contrary to what is often reported. The gradient distribution of the species within the SEI therefore corresponds to a partly mosaic structured SEI at the borders of the layers. At the lithium surface, an inorganic layer of lithium fluoride and then lithium carbonate is observed, followed by an organic, more porous SEI layer consisting of lithium ethylene dicarbonate. Simulations further reveal the strong prevalence of corrosion processes of the metal, which provide more than 99% of the lithium for the SEI reaction processes. The salt contributes less than 1% to the SEI formation. Additionally, SEI formation below and above the initial interface was observable. The here presented novel modeling approach allows an unprecedented in-depth analysis of processes during native SEI formation that can be used to improve design for high battery performance and durability.

Nowadays, lithium metal anodes are often referred to as the ‘holy grail’ of next-generation battery technology. Compared to graphite anodes used in state-of-the-art lithium-ion batteries they promise several times higher energy densities [1]. However, Li metal shows a high reactivity with liquid electrolytes and uncontrolled dendrite growth which hinders a safe and efficient cyclability and therefore the commercial application in rechargeable batteries. Thus, stabilization of the solid-electrolyte interphase (SEI) between liquid electrolyte and the lithium surface is the key challenge to be solved before bringing this promising battery technology to market. In order to suppress unwanted effects such as dendrite formation, persistent loss of active material and high kinetic losses, this layer of electrolyte decomposition products needs to be electrically insulating and should prevent direct contact of metal with electrolyte while ensuring a good ionic conductivity and mechanical stability.Such a beneficial SEI design requires detailed understanding of the fundamental formation mechanism and the resulting chemical SEI composition and structure depending on electrolyte composition. In literature, a large range of experimental methods are applied to characterize the SEI under various conditions [2]. These approaches are mainly descriptive and struggle to identify and understand the underlying mechanisms. Theoretical approaches like density functional theory (DFT), ab-initio molecular dynamics (MD) or classical MD are well suited to determine reaction mechanisms and to model the very beginning of SEI formation in the ns-range. However, they are too computationally expensive to model technical relevant length- and time scales [3].Therefore, this work focuses on overcoming these restrictions by developing an efficient multiscale-model with input from ab-initio calculations. The acceleration originates from focusing on rare events such as reaction or diffusion processes in a 3-dimensional kinetic Monte Carlo instead of considering the vibrational motions of single atoms like in molecular dynamics. Thereby, every process is represented by a probability derived from ab-initio calculations to ensure physical correctness. An additional coupling with efficient ordinary differential equations as shown in [4] ensures electroneutrality in the simulation box.Implementation of the degradation reactions for a 1.2M LiPF6/EC electrolyte obtained from molecular simulations (see Figure 1, appendix), allows to unravel the initial SEI formation, the emerging structure and its composition within the first µs. The simulation suggests an almost instantaneous formation of SEI species on the pristine Li metal within the first nanoseconds after immersion. Due to this early passivation layer the formation process slows down continuously until electron tunneling is blocked by SEI species and a first passivation of the lithium surface is reached. Interestingly, the formed passivation layer mainly consists of inorganic species such as LiF close ...