Atomistic modeling of LiF microstructure ionic conductivity and its influence on nucleation and plating

The formation and degradation of the solid electrolyte interphase (SEI) and its underlying transport properties play an essential role in the overall performance of lithium-ion batteries. This paper presents classical molecular dynamics studies on polycrystalline inorganic lithium fluoride (LiF) layers to model and predict the SEI transport properties. The ionic conductivity is obtained from the lithium-ion diffusivity in polycrystalline structures of LiF using the Nernst-Einstein relation. The predicted molecular dynamics data are used in a continuum scale phase-field model to evaluate the plating kinetics under fast charging conditions. The analysis emphasizes that the SEI ionic conductivity properties impact the plating dynamics, where SEI's low ion conductivity value is prone to large plating and subsequent capacity degradation. The combination of atomic and continuum scale studies shown herein lays a foundation to tune in SEI transport properties to decrease the amount of lithium plating and improve the performance of fast-charging batteries.

Figure 1

Initial configuration of LiF polycrystal containing 64 grains in a (10 ${\mathrm{nm})}^3$ volume and its grain size distribution.

Figure 2

A single-particle simulation setup for understanding the effect of SEI conductivity on the plating kinetics under fast charge conditions: (a) schematic of electrolyte region (red), SEI (green), and anode graphite particle (black), (b) corresponding initial lithium concentration in electrolyte, SEI and graphite regions. In color bar, 1 represents fully lithiated state of graphite and 0 is complete delithiated state of graphite. Highlighted dashed box shows a magnified region used in results (see Sec. 3b).

Figure 3

Li mean square displacement in a $15\times 15\times 15{\mathrm{nm}}^3$ LiF system with eight grains at 300 K from 8 to 20 ns. Running average every 100 steps shown as black line.

Figure 4

Local ordering and atomic displacement in (10 ${\mathrm{nm})}^3$ polycrystal samples. [(a) and (b)] Grain identification based on coordination for 8 and 64 grains respectively. Atoms identified as grain boundaries are colored in gray and bulk atoms are colored by grain. [(c) and (d)] Color map of atomic displacement magnitude taken at 20 ns of MD run for 8 and 64 grain samples, respectively. Displacement magnitude color bar has been adjusted for better visualization, where white represents almost zero ion mobility.

Figure 5

Correlation between Li ion displacement magnitude and atomic local ordering as obtained from snapshot at the end of the 20 ns simulation in (10 ${\mathrm{nm})}^3$ sample with 8 grains. Atomic displacements values were sorted before being groups into bins of 500. Data points plotted in this figure show the average order versus the average displacement for each bin. Local ordering ranges from 0—meaning disorder atomic structure—to 1—identified as simple cubic lattice.

Figure 6

Predicted average plating overpotential at graphite/SEI and SEI/electrolyte interface for 4C-Rate charging.

Figure 7

Simulated metallic lithium nucleation and plating kinetics under a charging condition for 4C rate. (a) no SEI influence, (b) SEI influence. Color scheme is the same as used in Fig. 2.

Figure 8

SOC of the graphite anode for the onset of lithium nucleation on graphite as function of C rate for three SEI conductivity values: $1.5\times {10}^{-3}$ (red line), 0.64 (black line), and 6.4 S/m (blue line).