Fast charging remains one of the most significant challenges in the widespread adoption of electric vehicles (EVs). While it solves one bottleneck—refueling time—it introduces others, notably safety concerns and accelerated battery aging. Lithium plating, a degradation mode triggered under certain fast-charging conditions, is one of the most pressing. One potential avenue to assess the onset of plating is through modeling the anode potential. Although attractive in theory, applying this approach in practice is far from straightforward.
Why Anode Potential Is Important
Lithium plating occurs when lithium ions deposit as metallic lithium on the surface of the anode, rather than intercalating into it. This behavior is tied to the electrochemical potential of the anode. If the anode potential drops below the equilibrium potential of lithium metal, plating becomes thermodynamically favorable.
This makes the anode potential a theoretically useful signal. If we can model it accurately under different operating conditions—such as current, temperature, and state of charge—we can predict when a battery enters a high-risk regime for plating. However, practical implementation is significantly more complex.
The Accuracy Challenge
Anode potential models require detailed representations of the battery's internal electrochemical and physical processes. These models often involve dozens of parameters, including reaction kinetics, ionic and electronic transport properties, thermal behavior, and microstructural features.
Many of these parameters are either not directly measurable or vary significantly with battery aging and usage. Obtaining them typically requires post-mortem cell disassembly, highly controlled laboratory tests, or specialized diagnostics that are not feasible at scale. Even with a full initial parameter set, several limitations remain:
Parameters change over time. As batteries age, properties such as diffusivity and conductivity degrade, making models based on fixed values increasingly inaccurate. While models can be adapted to include degradation-dependent changes, doing so significantly increases complexity. Each new degradation mechanism introduces additional unknowns. Obtaining the necessary values for these parameters requires extensive aging tests and post-mortem analyses—procedures that are both time-consuming and expensive.
Most models assume uniformity. They typically average behavior across an entire electrode, overlooking local variations where lithium plating often starts. Some enhanced models segment the electrode into distinct segments, which can provide insight into how current distributes under load. However, these segments are still relatively large. The resolution remains far too coarse—nowhere near the sub-millimeter scale needed to capture meaningful local inhomogeneities.
Key variables are inaccessible during operation. Internal anode potential cannot be directly measured in a functioning cell, making real-time validation impossible. As a result, most validation is limited to lab-based experiments on relatively fresh cells. Rarely are models validated on aged batteries, despite the fact that these are precisely the cases where accurate estimates of internal potential are most needed.
Together, these limitations constrain the reliability of anode potential predictions. A model that appears robust in theory may still fall short under practical, real-world conditions.
A Role Within a Larger Framework
Despite its limitations, anode potential modeling can play a useful role when combined with other diagnostic tools. For example, changes in open-circuit voltage (OCV) can be monitored through incremental capacity analysis (ICA) or differential voltage analysis (DVA). These techniques provide insight into degradation mechanisms such as loss of lithium inventory, which is often associated with lithium plating.
In such cases, anode potential modeling can serve as a complementary source of insight. It helps generate hypotheses about which operating conditions might lead to plating, which can then be tested and refined using empirical methods.
In research and lab settings, detailed models can still be valuable, especially when calibrating reduced-order models or understanding degradation under controlled conditions. In commercial applications, however, more practical approaches are needed.
Moving Toward Realistic Use Cases
The future of predictive battery diagnostics lies in hybrid approaches. Rather than relying on a single model to predict lithium plating, it is more effective to combine physical insight with field data and statistical trends. Reduced-order models, trained or informed by high-fidelity simulations and validated against empirical data, offer a more scalable path forward.
In practice, this means accepting some trade-offs. Perfect precision in modeling anode potential may not be achievable. But meaningful, directional insights—especially when paired with robust data analytics—can still guide better decisions around charging control, safety margins, and battery design.
Conclusion
Anode potential modeling provides a conceptually direct path to understanding lithium plating, but its practical limitations restrict its standalone utility. The challenges in parameterization, aging effects, and measurement access make full physical-chemical models difficult to deploy outside controlled environments.
Still, as part of a broader diagnostic toolkit, anode potential modeling retains value. When interpreted alongside empirical signatures of degradation, it can enhance our understanding of how fast charging affects battery health.
Improving battery safety and performance in the fast-charging era requires embracing the complexity of the system. That means combining theoretical modeling, empirical observation, and real-world analytics—not relying on any one method to do it all.
