The widespread adoption of lithium-ion batteries in electric vehicles, energy storage systems, and portable electronics has placed a premium on their long-term reliability and stability. In many applications, exposure to elevated temperatures during storage or operation is unavoidable. This exposure accelerates detrimental processes within the cell, leading to significant performance degradation, including pronounced capacity fade, shortened lifespan, and heightened safety risks. Particularly for lithium-ion batteries stored at a high state-of-charge (SOC), the thermodynamically unstable internal state exacerbates chemical and electrochemical side reactions, severely impacting performance and safety. Therefore, a detailed investigation into the capacity fade mechanisms of lithium-ion batteries under high-temperature storage conditions is crucial for developing effective strategies to mitigate this degradation.
Current research indicates that the capacity fade of lithium-ion batteries during high-temperature storage originates from several interconnected mechanisms. A primary contributor is parasitic reactions at the electrode-electrolyte interfaces. At the anode, electrolyte reduction leads to the continuous growth and reformation of the Solid Electrolyte Interphase (SEI). This process irreversibly consumes active lithium ions and electrolyte components. Storage at high temperature and high SOC is particularly detrimental as the anode resides at a lower potential, facilitating these reduction reactions. Furthermore, oxidation reactions occur at the cathode, decomposing the electrolyte and increasing interfacial impedance. Another significant mechanism is the loss of active material in both electrodes. For the anode, the persistent growth of the SEI layer can physically isolate and deactivate graphite particles, rendering them inaccessible for lithiation. For the cathode, especially in layered oxides like NCM or NCA, structural degradation, transition metal dissolution, and particle cracking can occur at elevated temperatures. The dissolved transition metal ions (e.g., Mn2+, Co2+, Ni2+) can migrate to the anode, where they are reduced and deposited, further catalyzing electrolyte decomposition and SEI growth, creating a vicious cycle of degradation. While these mechanisms are qualitatively understood, quantitative analysis of their individual contributions remains essential for targeted cell design and management strategy optimization.
This study focuses on quantitatively elucidating the degradation mechanisms of a pouch-type lithium-ion battery with a Nickel-Cobalt-Manganese (NCM) cathode and a graphite anode during storage at 60°C and 100% SOC. We employ a multi-faceted electrochemical analysis toolkit. By utilizing a cell equipped with a reference electrode, we can deconvolute the voltage contributions of the individual electrodes during low-rate cycling. Key analysis techniques include Incremental Capacity (IC) analysis, Differential Voltage (DV) analysis, and a voltage reconstruction method. These techniques allow us to systematically track changes in electrode states and quantitatively separate the contributions from loss of active lithium (LLI), loss of cathode active material (LAMPE), and loss of anode active material (LAMNE) to the overall capacity fade.
1. Experimental Methodology
1.1 Cell Configuration and Preparation
The lithium-ion batteries under investigation were custom-designed pouch cells. The cathode active material was a LiNi0.65Co0.07Mn0.28O2 (NCM) layered oxide, and the anode consisted of synthetic graphite. A critical feature for this study was the integration of a micro-reference electrode. A thin copper wire (50 µm diameter) was carefully inserted between the cathode and anode stacks during cell assembly. This wire was later electrochemically lithiated in-situ to function as a stable lithium reference electrode, enabling simultaneous monitoring of the cathode and anode potentials versus Li/Li+ during full-cell operation.
For baseline characterization, coin-type half-cells were fabricated using electrode samples taken from the same batch as the pouch cell electrodes. Cathode and anode discs were punched out and assembled versus lithium metal counter/reference electrodes in 2032-type coin cells. All cell assembly procedures (pouch and coin) were conducted in an argon-filled glovebox (M.Braun Unilab Pro) with oxygen and moisture levels maintained below 0.1 ppm.
1.2 Electrochemical Testing Protocols
Half-cell Testing: The cathode and anode half-cells were cycled at a low rate of 0.05 C (where 1 C is the current required to charge/discharge the nominal capacity in one hour). The voltage windows were set at 2.8–4.5 V for the cathode and 0.005–2.0 V for the anode versus Li/Li+. After two formation cycles, the charge curve for the cathode half-cell and the discharge curve for the anode half-cell from the second cycle were extracted. These curves, acquired under quasi-equilibrium conditions, represent the open-circuit potential (OCP) as a function of lithium content (x in LixGraphite and y in LiyNCM) and serve as the foundational data for the voltage reconstruction model.
Pouch Cell High-Temperature Storage Test: The three-electrode pouch lithium-ion battery was first charged to 100% SOC (4.2 V) using a standard constant-current constant-voltage (CCCV) protocol. In this charged state, the cell was placed into a temperature-controlled chamber at 60°C. The storage test was conducted for a total of 48 days. At intervals of 12 days, the cell was removed from the chamber and allowed to equilibrate to room temperature for at least 6 hours.
Following each storage period, a diagnostic low-rate charge test was performed. Prior to this test, the copper wire reference electrode was activated by plating it with lithium. This was done by applying a small constant current (30 µA) between the cathode and the copper wire for 4 hours, effectively creating a LixCu reference. The diagnostic test consisted of a 0.05 C constant-current (CC) charge to 4.20 V, a 30-minute rest, a 0.05 C CC discharge to 2.8 V, and another 30-minute rest. Two complete cycles were performed, and data from the second charge segment were used for analysis to ensure stable electrode conditions. The exceptionally low current rate minimizes polarization effects, allowing the measured full-cell, cathode, and anode voltages to closely approximate their respective thermodynamic equilibrium potentials. The key parameters for the storage experiment are summarized in Table 1.
| Parameter | Specification |
|---|---|
| Cell Type | Pouch, NCM/Graphite, with reference electrode |
| Initial Capacity | 57.83 mAh |
| Storage SOC | 100% (4.20 V) |
| Storage Temperature | 60°C |
| Total Storage Duration | 48 days |
| Diagnostic Interval | Every 12 days |
| Diagnostic Rate | 0.05 C |
2. Results and Discussion
2.1 Capacity Fade During Storage
The evolution of discharge capacity and capacity retention over the 48-day storage period is plotted. The lithium-ion battery exhibited a continuous decline in capacity. The initial capacity of 57.83 mAh decreased to 49.96 mAh after 48 days, corresponding to a capacity retention of 86.40%. This significant fade of 13.6% underscores the aggressive nature of the 60°C, high-SOC storage condition. The following sections delve into the electrochemical signatures to unravel the root causes of this fade.
2.2 Incremental Capacity (IC) Analysis
The IC curve, or dQ/dV versus V, is a sensitive indicator of phase transitions and changes in electrode stoichiometry within a lithium-ion battery. It is obtained by differentiating the charge/discharge capacity with respect to voltage. The IC curves for the full cell, cathode, and anode during the 0.05 C charge at different storage times are analyzed.
Three distinct peaks are observed in the IC curves for all three components (full cell, cathode, anode). These peaks correspond to the major two-phase transitions during (de)lithiation of the electrodes. For the graphite anode, the stages of lithiation (LiC6 → LiC12 → LiC18 etc.) appear as peaks. For the NCM cathode, the peaks correspond to ordered phase transitions (e.g., hexagonal to monoclinic, etc.) during lithium de-intercalation.
Full-cell IC Analysis: With increased storage time, the three peaks in the full-cell IC curve show a slight shift towards higher voltages, indicating a small increase in internal impedance. More importantly, the intensity (height) of all peaks diminishes significantly. A uniform decrease in the height of all IC peaks typically suggests a loss of active material, as fewer host sites are available for the phase transitions to occur.
Cathode IC Analysis: The cathode IC peaks also shift slightly to higher potentials and, crucially, show a dramatic reduction in intensity. This is a strong signature of cathode active material loss (LAMPE). The material loss could be due to structural degradation, particle isolation caused by surface film growth, or irreversible phase changes that render parts of the cathode electrochemically inactive.
Anode IC Analysis: Similarly, the anode IC peaks show a marked reduction in intensity with storage time, pointing towards loss of anode active material (LAMNE). This is commonly associated with the progressive growth of a thick, resistive SEI layer that encapsulates and electrically isolates graphite particles. Minor fluctuations in peak intensity during storage might be attributed to high-temperature-induced redistribution of lithium within the electrode, temporarily altering the local state-of-charge distribution.
In summary, the IC analysis qualitatively indicates that the capacity fade in this lithium-ion battery during high-temperature storage is driven by substantial degradation of both electrode materials (LAMPE and LAMNE), accompanied by a loss of active lithium (LLI) which is intrinsically linked to the SEI growth on the anode.
2.3 Differential Voltage (DV) Analysis
The DV curve, or dV/dQ versus Q, provides complementary information, often making specific features more discernible. It is derived by differentiating the voltage with respect to capacity. The DV curves for the full cell, cathode, and anode are analyzed in tandem.
The major peaks in the full-cell DV curve can be mapped to corresponding features in the cathode and anode DV curves. This mapping reveals which electrode governs the overall cell voltage at a given capacity. For instance, one full-cell peak (F1) is primarily governed by an anode feature (A1), while another peak (F2) is influenced by overlapping features from both the cathode (C2) and the anode (A2).
Tracking these features over storage time reveals critical trends: The position of anode peak A2 (associated with the LiC12 stage) shifts to lower capacities. This shift directly signifies a reduction in the accessible capacity of the graphite anode, i.e., LAMNE. Similarly, the shift of cathode peak C2 indicates LAMPE. Furthermore, the distance between the final DV peak (F2) and the end-of-charge point progressively decreases with storage. This shrinking “trailing edge” is a classic signature of loss of active lithium (LLI), as the lithium inventory becomes depleted before the electrodes can reach their full compositional limits.
The DV analysis corroborates the findings from the IC analysis, confirming that the dominant aging modes for this lithium-ion battery under the test conditions are the concurrent loss of active material from both electrodes and the loss of cyclable lithium. The correlation between the major IC/DV features and the electrode processes is summarized in Table 2.
| Feature Location | Primary Governing Electrode | Associated Process | Degradation Indicator |
|---|---|---|---|
| Full-cell IC Peak 1 / DV Peak F1 | Anode | Graphite Stage Transition | Peak intensity loss → LAMNE |
| Full-cell IC Peak 2 / DV Peak F2 | Cathode & Anode | NCM Phase Transition & Graphite Stage | Peak shift & intensity loss → LAMPE & LAMNE |
| DV Curve “Trailing Edge” | Full Cell | End of Charge | Edge shortening → LLI |
2.4 Quantitative Degradation Analysis via Voltage Reconstruction
While IC and DV analyses offer qualitative and semi-quantitative insights, the voltage reconstruction method enables a more precise quantification of the degradation modes. This method simulates the full-cell charge curve based on the half-cell equilibrium potentials and a set of adjustable parameters representing the cell’s state.
2.4.1 Methodology
The foundation of the model is the half-cell equilibrium curves: $$V_{pe} = f_{pe}(y)$$ for the cathode and $$V_{ne} = f_{ne}(x)$$ for the anode, where \(y\) and \(x\) represent the lithium stoichiometry in the cathode and anode, respectively (e.g., \(y\) in LiyNCM, \(x\) in LixC6).
For a full cell starting a charge from a known state, the initial stoichiometries are \(y_0\) (cathode) and \(x_0\) (anode). As charging progresses with a current \(I\), the stoichiometries change as:
$$y = y_0 – \frac{\int I dt}{C_{pe}}$$
$$x = x_0 + \frac{\int I dt}{C_{ne}}$$
where \(C_{pe}\) and \(C_{ne}\) are the effective capacities (related to the mass of active material) of the cathode and anode, respectively. The total lithium inventory in the cell is given by:
$$C_{Li} = y_0 \cdot C_{pe} + x_0 \cdot C_{ne}$$
During charge, the electrode and full-cell voltages are:
$$V_{pe} = f_{pe}\left(y_0 – \frac{\int I dt}{C_{pe}}\right)$$
$$V_{ne} = f_{ne}\left(x_0 + \frac{\int I dt}{C_{ne}}\right)$$
$$V_{full} = V_{pe} – V_{ne}$$
To analyze a degraded cell, the model parameters \(\{y_0, x_0, C_{pe}, C_{ne}\}\) are adjusted until the simulated \(V_{full}(Q)\) curve best fits the experimentally measured low-rate charge curve of the aged lithium-ion battery. By comparing the fitted parameters before and after aging, the losses are quantified:
- Loss of Cathode Active Material (LAMPE): \(\Delta C_{pe} / C_{pe, initial}\)
- Loss of Anode Active Material (LAMNE): \(\Delta C_{ne} / C_{ne, initial}\)
- Loss of Lithium Inventory (LLI): \(\Delta C_{Li} / C_{Li, initial}\)
The LLI can be conceptually understood as the sum of lithium trapped in the thickening SEI and any lithium made inaccessible due to cathode material loss. Therefore, the “pure” active lithium loss is approximately the total lithium inventory loss minus the lithium associated with the lost cathode material.
2.4.2 Reconstruction Results and Quantitative Fade
The voltage reconstruction was performed for the cell at different storage intervals. The model demonstrated excellent accuracy, with the reconstructed charge curves showing less than 2% error compared to the experimental data for both fresh and aged states, validating the model’s robustness.
The evolution of the fitted parameters reveals the degradation trajectory. The initial stoichiometries (\(y_0\), \(x_0\)) show minimal change, confirming the cell was maintained at ~100% SOC. The most pronounced changes are in the electrode capacities \(C_{pe}\) and \(C_{ne}\), which decrease steadily.
The quantitative breakdown of capacity fade after 48 days of storage is calculated as follows:
- Loss of Cathode Active Material (LAMPE): 6.32%
- Loss of Anode Active Material (LAMNE): 11.04%
- Loss of Active Lithium (LLI): 7.02%
This quantitative analysis provides a clear picture: the anode suffers the most severe degradation in terms of active material loss, which is consistent with the expected accelerated SEI growth and possible graphite particle isolation at high temperature and voltage. The cathode also experiences significant material loss, likely due to surface reconstruction, transition metal dissolution, or microcracking. The loss of active lithium is a direct consequence of continuous electrolyte reduction at the anode. The synergy between these mechanisms—LAMNE exposing fresh graphite surfaces requiring more SEI formation (consuming more Li), and possible transition metal dissolution from the cathode (LAMPE) poisoning the anode surface—creates a compounded degradation effect in the lithium-ion battery. The evolution of these parameters over time is summarized in Table 3, and the final quantitative contributions are presented in Table 4.
| Storage Time (days) | \(C_{pe}\) (mAh) | \(C_{ne}\) (mAh) | \(C_{Li}\) (mAh) | \(y_0\) | \(x_0\) |
|---|---|---|---|---|---|
| 0 (Initial) | 62.15 | 68.21 | 135.92 | 0.32 | 0.08 |
| 12 | 60.87 | 65.44 | 132.78 | 0.31 | 0.08 |
| 24 | 59.41 | 63.02 | 129.58 | 0.31 | 0.09 |
| 36 | 58.76 | 61.88 | 127.85 | 0.32 | 0.09 |
| 48 | 58.22 | 60.68 | 126.38 | 0.32 | 0.09 |
| Degradation Mode | Symbol | Loss Magnitude | Primary Likely Cause |
|---|---|---|---|
| Loss of Cathode Active Material | LAMPE | 6.32% | Surface degradation, TM dissolution, particle cracking |
| Loss of Anode Active Material | LAMNE | 11.04% | SEI overgrowth isolating graphite particles |
| Loss of Active Lithium | LLI | 7.02% | Consumption via SEI formation and reformation |
3. Conclusion
This study presents a comprehensive investigation into the degradation mechanisms of a commercial-type NCM/Graphite lithium-ion battery subjected to high-temperature (60°C) storage at full state-of-charge. By integrating in-situ electrode potential monitoring via a reference electrode with advanced electrochemical analysis techniques—Incremental Capacity (IC), Differential Voltage (DV), and a physics-based voltage reconstruction method—we have transitioned from qualitative observation to quantitative diagnosis.
The key findings are:
- Substantial Capacity Fade: The lithium-ion battery lost 13.6% of its initial capacity after 48 days under the specified storage conditions.
- Multi-Mode Degradation: The fade is not attributable to a single mechanism but is the result of three concurrent processes: significant loss of anode active material (LAMNE = 11.04%), notable loss of cathode active material (LAMPE = 6.32%), and considerable loss of active lithium inventory (LLI = 7.02%).
- Anode as Primary Degradation Site: The largest single contribution comes from the loss of accessible graphite material, overwhelmingly driven by the continuous, accelerated growth and thickening of the SEI layer at high temperature and low anode potential. This process simultaneously causes LAMNE and is the primary consumer of lithium (LLI).
- Cathode Degradation: The cathode also undergoes degradation, likely involving surface layer formation, transition metal ion dissolution, and possibly mechanical grain cracking, leading to active material loss.
The methodology demonstrated here—characterizing a lithium-ion battery using low-rate diagnostic cycles with reference electrode data and applying differential analysis and voltage reconstruction—provides a powerful, non-destructive framework for quantifying aging modes. This approach can be universally applied to various lithium-ion battery chemistries to deconvolute the complex interplay of degradation mechanisms during storage, cycling, or under different stress conditions. The quantitative insights gained are invaluable for guiding the development of more robust electrode materials, electrolytes, and cell designs to enhance the longevity and reliability of lithium-ion batteries, especially for applications demanding high-temperature tolerance.