Physical Origin of the Differential Voltage Minimum Associated with Lithium Plating in Li-Ion Batteries

Abstract

The main barrier to fast charging of Li-ion batteries at low temperatures is the risk of short-circuiting due to lithium plating. In-situ detection of Li plating is highly sought after in order to develop fast charging strategies that avoid plating. It is widely believed that Li plating after a single fast charge can be detected and quantified by using a minimum in the differential voltage (DV) signal during the subsequent discharge, which indicates how much lithium has been stripped. In this work, a pseudo-2D physics-based model is used to investigate the effect on Li plating and stripping of concentration-dependent diffusion coefficients in the active electrode materials. A new modelling protocol is also proposed, in order to distinguish the effects of fast charging, slow charging and Li plating/stripping. The model predicts that the DV minimum associated with Li stripping is in fact a shifted and more abrupt version of a minimum caused by the stage II-stage III transition in the graphite negative electrode. Therefore, the minimum cannot be used to quantify stripping. Using concentration-dependent diffusion coefficients yields qualitatively different results to previous work. This knowledge casts doubt on the utility of DV analysis for detecting Li plating.

A Review of Existing and Emerging Methods for Lithium Detection and Characterization in Li‐Ion and Li‐Metal Batteries

“…Modeling results presented by O'Kane et al highlight that changes in solid-state diffusion coefficient for graphite with concentration complicate whether dV/dQ analysis can be quantitative. [167] Campbell et al showed the absence of a differential voltage peak does www.advancedsciencenews.com not necessarily mean Li has not plated. [43] Further, most developments to date have focused on method development at low temperatures.…”

“…highlight that changes in solid‐state diffusion coefficient for graphite with concentration complicate whether dV/dQ analysis can be quantitative. [ 167 ] Campbell et al. showed the absence of a differential voltage peak does not necessarily mean Li has not plated.…”

“…[43] Further, most developments to date have focused on method development at low temperatures. [35,36,43,167] Thus, there is a need for increased synergy between experimental findings and modeling-based efforts to detect and characterize Li, to better understand and design against cell degradation.…”

“…[170] The exchange current density for Li plating/stripping has been reported to be between 0.001 and 100 A m −2 and cathode symmetry factors are reported between 0.5 and 0.7. Part of the large discrepancy within the P2D framework is that Li plating with a BV reaction is often treated as irreversible, [169,171] completely reversible, [167,170] or a has a set constant reversibility. [25,35] In reality, the reversibility is likely a strong function of many conditions and changes during operation.…”

“…[77,161,169] However, researchers have suggested that explicitly modeling the Li carbon species using either a volume-averaged Avrami formulation [173] or Cahn-Hillard formulation [12,174] improves accuracy. Further, the diffusion coefficient varies with lithiation amount [36,167] and may also change with local applied current. [175] Researchers have investigated using nucleation and growth principles, phase-field modeling, and SEI/particle morphology effects to accurately capture Li plating and stripping.…”