When a car sits at an EV fast-charging station and the battery level climbs back up in just ten-odd minutes, intense chemical reactions and physical deformation are unfolding simultaneously inside the electrodes. The charger's display shows a steadily rising charge percentage. But even when the charging behavior looks smooth on the surface, serious damage can be progressing in the microscopic world inside the electrode.
A research team led by Professor Kang Taek Lee and Professor EunAe Cho at the Korea Advanced Institute of Science and Technology (KAIST) used a 3D digital twin built from the structure of a commercial graphite anode to run numerical simulations of the localized degradation processes that occur inside electrodes during fast charging. The paper, co-authored with Yejin Kang, Seungsoo Jang, and Jung Hyeon Moon, was selected for the back cover of the journal InfoMat (published July 2026, DOI: 10.1002/inf2.70141). The team's central finding is that macroscopic averages—such as total binder content or average porosity across the whole electrode—cannot fully capture how an electrode degrades during fast charging.
Three microscopic degradation mechanisms occurring simultaneously inside the battery
During fast charging of a lithium-ion battery, lithium ions that have moved from the cathode travel through fine electrolyte-filled pores and burrow into the graphite particles of the anode. A polymer material called a binder holds the graphite particles together and bonds them to the current collector. When charging speed is pushed to an extreme, three interlinked physicochemical phenomena occur inside the electrode, degrading battery performance.
The first is metallic lithium plating. When the rate at which lithium ions insert into the graphite particle surface falls behind the rate at which ions are supplied from the electrolyte, excess ions that cannot enter the graphite lattice instead deposit as metallic lithium on the particle surface. This plated lithium becomes inactive lithium that no longer contributes to charge and discharge, permanently robbing the battery of capacity—and it can also grow into dendrites that pierce the separator, risking an internal short circuit.
The second is uneven growth of the solid electrolyte interphase (SEI). At the interface where graphite meets the electrolyte, decomposition of the electrolyte forms a protective SEI layer. Under the steep potential gradients created by fast charging, this layer can become excessively thick or form unevenly in localized spots, blocking ion pathways and increasing internal resistance.
The third is the concentration of internal stress caused by volume changes in the particles. Graphite particles expand by up to about 10% in volume as they absorb lithium. As particles expand, they press against neighboring particles and the binder, generating tiny mechanical stresses. When this stress becomes locally concentrated, it can cause the electrode structure to collapse or particles to become isolated.
Conventional physical experiments and standard one-dimensional electrochemical models have struggled to isolate how these three phenomena interfere with one another at the microscopic scale. Commercial electrode evaluations typically rely on macroscopic metrics such as the binder's weight fraction relative to total electrode volume or average porosity. Even when these average values are identical, the spatial arrangement of the constituent materials can dramatically alter internal ion transport pathways and reaction distribution.
Reconstructing 3D structure and analyzing parameters based on measured data
Based on the microstructural specifications of an actual commercial graphite anode, the research team built a 3D digital twin reproducing the three phases: graphite particles, binder, and the electrolyte-filled pore network. The group had previously established, in earlier work (Kang et al., Chemical Engineering Journal, 2025, DOI: 10.1016/j.cej.2025.164524), a coupled electrical, chemical, and mechanical modeling framework integrating volume expansion and SEI formation. In the present paper, they extended that framework to analyze microstructural heterogeneity under fast-charging conditions.
The simulation model's validity was confirmed against measured charging profile data. However, it should be noted that the individual degradation behaviors output by the model—such as the amount of lithium plating, the local distribution of SEI thickness, and the internal stress distribution—are predicted values from computational numerical simulation.
In the analysis, the following parameters were systematically varied as electrode design variables:
- Electrode thickness (a standard 50 μm and a higher-capacity 83 μm)
- Porosity distribution within the electrode
- Spatial arrangement of the binder (a steep gradient concentrated toward the separator side, versus a uniform or gentle gradient)
Through the simulations, the flow of lithium ions along the electrode depth, the locations where overpotential arose, the local risk of lithium plating, and the stress distribution from graphite expansion were visualized down to fine local regions.
Binder placement effects become more pronounced as electrode thickness increases
The key insight from the computational results is that looking only at the electrode's overall charge capacity can cause localized internal degradation to go unnoticed.
In the standard 50 μm-thick electrode model, comparing a case where the binder was heavily concentrated toward the separator side against a case where it was uniformly distributed throughout the electrode, the difference in apparent overall charge capacity was less than 4% (the underlying absolute charge capacity values for this comparison were not disclosed in the published materials). At first glance, binder distribution appears to have almost no effect on charging behavior.
Looking at the microscopic regions inside the electrode tells a very different story. When binder was concentrated near the separator, the pore pathways that lithium ions need to penetrate deeper into the electrode narrowed. As a result, the simulation showed that the amount of lithium plating occurring near the current collector increased by more than 10% compared to the model with uniformly distributed binder (again, the absolute value of local plating amount was not disclosed in the published materials).
The impact of this internal heterogeneity became clearly evident in the model where the electrode was thickened to 83 μm to pursue higher energy density. In thick electrodes where ions must travel farther, increased transport resistance from pore blockage directly limits the capacity the whole electrode can deliver. Compared to the model with a steep binder concentration gradient, the model with uniformly distributed binder showed roughly 18% higher charge capacity (a ratio comparison; the corresponding absolute capacity values were not disclosed in the published materials).
| Evaluation Parameter | 50 μm-thick anode (steep binder concentration vs. uniform distribution) | 83 μm-thick anode (steep binder concentration vs. uniform distribution) |
|---|---|---|
| Difference in apparent total charge capacity | Slight difference of under 4% (internal damage is masked) *absolute values not disclosed | Uniform distribution maintains roughly 18% higher charge capacity *absolute values not disclosed |
| Lithium plating near the current collector | Increases by more than 10% with binder concentration *absolute values not disclosed | Plating risk increases over a wide area due to delayed ion transport |
| Role of the pore network | Source of localized ionic resistance | Stress-relief buffer against particle expansion |
| Main degradation constraint | Localized overpotential and uneven lithium deposition | Overlapping effects of delayed ion diffusion along electrode depth and stress concentration |
The spatial arrangement of pores is also directly involved in relieving mechanical stress. When graphite particles expand during charging, if adequate pore space exists around the particles, that space acts as a buffer zone absorbing the deformation. In regions where pores are constricted or filled with binder, there is nowhere for the particle's volume change to be released, and local mechanical stress increases. The study confirmed that a design with appropriately placed high porosity both reduces ion diffusion resistance and significantly relieves localized stress concentration.
Simulation opens the door to electrode structure design, and remaining challenges
The conclusion presented by this research is that battery design for fast charging should shift its perspective—from managing total quantities, or "what to mix and how much," toward microstructural control, or "how to spatially arrange the constituent materials."
In an official statement, KAIST's Professor Kang Taek Lee said the significance of this research lies in using a 3D digital twin to visualize internal battery problems that were difficult to detect from overall charging performance figures alone. Conventional battery development has mainly relied on building prototype cells, repeating charge-discharge cycle tests, and then disassembling and analyzing them after degradation. By using digital twin analysis, it becomes possible to evaluate an electrode's internal ion transport properties and stress concentration risks in a virtual space before prototyping, and to optimize the structure accordingly.
Some academic caveats are needed when interpreting these results. This research is a theoretical prediction based on computational simulation; it did not involve manufacturing actual cells with the proposed binder arrangement and subjecting them to long-term cycle life testing or physical verification of suppressed plating. The model studied is based on the structure of a specific commercial graphite anode, and it remains unverified whether the same numerical results would apply directly to different material systems, such as silicon-added graphite anodes or all-solid-state batteries.
On the electrode manufacturing floor, intentionally controlling the depth-wise distribution of binder to within tens of micrometers of precision during the coating and drying processes remains a significant manufacturing challenge. Future research will focus on verifying how the ideal spatial arrangement indicated by the simulation can be reproduced on actual electrode production lines, and whether such electrodes demonstrate the predicted long life and fast-charging tolerance in physical experiments.