(Invited) Modeling of Lithium Plating Induced By Heterogeneities at Varied Length-Scales in Lithium-Ion Batteries

Abstract

Lithium (Li) plating is an unsafe, damaging side reaction that can occur during fast and/or low temperature charging of Li-ion batteries. The bulk driving force for Li plating is well understood. The side reaction becomes thermodynamically favorable when the negative electrode solid phase potential becomes less than the electrolyte phase potential. For today’s energy-dense batteries with thick electrodes, electrolyte Li+ transport limitations at high charge rates can lead to plating. High charge rates cause electrolyte salt depletion in the back of the negative electrode, shutting down reaction current there and causing excessive reaction current and plating at the electrode front. Pseudo 2D (P2D) macro-homogeneous models mathematically quantify the bulk onset of plating at the front of the negative electrode where it interfaces the separator. In practice however, Li does not plate uniformly across the entire negative electrode front. Instead, there are preferential regions where plating first occurs due to heterogeneities, sometimes at electrode edges, sometimes at the cell center and other times in seemingly random locations. Heterogeneous plating occurs earlier than the homogeneous P2D model predicts thus making it important to understand heterogeneities in order to achieve faster charge rates. Here we review known heterogeneities in Li-ion batteries and mathematical models that quantify their contribution to Li plating. Heterogeneities arise at all length scales of the battery, ranging from graphite crystal grain orientation at the sub-micron scale to meter-scale temperature distributions across large format cells and packs. Some heterogeneities such as edge effects are easy to avoid in design. To compensate for 2D transport effects as well as inevitable slight misalignment in an electrode stack for example, the negative electrode must overhang the positive electrode, typically by 0.5 to 1.0 mm. And in cases where the electrode stack is wetted with excess liquid electrolyte at its edges, the separator must further overhang the electrodes to prevent a preferential ion transport path through the free liquid electrolyte that may short-circuit the normal path through the separator. Other heterogeneities are not entirely possible to avoid and must be understood to quantify conservatism needed when setting fast charging limits. Using optical cell experiments and various modeling techniques, Harris1 and Thomas-Alyea2 have both shown preferential nucleation and reaction occurring due to graphite crystal anisotropy, particle-to-particle contact resistance and particle size/morphology differences. Complementing their previous analysis, we show results from a 3D microstructure model capturing early plating onset due to size/morphology effects across a 40-μm field of view. At a slightly larger length scale, local porosity and tortuosity variations caused by an inhomogeneous electrode coating can also lead to early onset of Li plating. We show a mesoscale model quantifies early onset of Li-plating at an approximately 50-μm diameter region of a 300x300 μm2 electrode plate area. The micro and mesoscale model domains are based on electrode 3D geometries obtained with computed tomography experiments. In large format cells, temperature, pressure and current collector potential gradients cause heterogeneous cell utilization that also leads to early onset of Li plating. A multi-scale multi-domain (MSMD) cell model provides examples of these cell-scale heterogeneities and their impact. Comparing magnitude of heterogeneities across the varied length scales, we comment on design of Li-battery materials, electrodes and cells to best suppress heterogeneity-driven Li plating. References K.E. Thomas-Alyea, C. Jung, R.B. Smith, M.Z. Bazant, “In Situ Observation and Mathematical Modeling of Lithium Distribution within Graphite,” J. Echem. Soc. 164 (11) E3063-E3072 (2017).S.J. Harris, E.K. Rahani, V.B. Shenoy, “Direct In Situ Observation and Numerical Simulations of Non-Shrinking-Core Behavior in an MCMB Graphite Composite Electrode,” J. Echem. Soc., 159 (9) A1501-A1507 (2012).G.-H. Kim, K. Smith, J. Lawrence-Simon, C. Yang, “Efficient and Extensible Quasi-Explicit Modular Nonlinear Multiscale Battery Model: GH-MSMD,” J. Echem. Soc., A1076-88 (2017).