In this work, I systematically investigate the capacity fading mechanisms of graphite||LiFePO₄ pouch-type energy storage batteries under various temperature conditions (25 °C, 45 °C, 60 °C, 70 °C, and 80 °C). The primary objective is to understand how elevated temperatures accelerate degradation and to identify the critical temperature threshold beyond which the aging mechanism changes. Through combined cycling tests, differential capacity analysis, and post-mortem characterization using scanning electron microscopy (SEM), inductively coupled plasma optical emission spectrometry (ICP-OES), and X-ray diffraction (XRD), I reveal that the activation energy of the aging process shifts when the cycling temperature exceeds 60 °C. This transition leads to accelerated solid electrolyte interphase (SEI) growth, active material particle cracking, and transition metal dissolution, ultimately causing faster capacity loss. These findings provide essential guidance for accelerating life tests and building reliable lifetime prediction models for large-format energy storage batteries.
LiFePO₄-based energy storage batteries are widely adopted for grid-scale applications due to their excellent structural stability, safety, and low cost. However, one major challenge in developing long-life products is the lengthy cycle life testing, which can consume up to 80% of the research and development timeline. Accelerated aging tests using elevated temperature are commonly employed, following the Arrhenius model:
$$ \theta = A e^{\frac{E_a}{k_B T}} $$
where θ is the characteristic life, A is the pre-exponential factor, Ea is the activation energy, kB is Boltzmann’s constant (8.617×10⁻⁵ eV/K), and T is the absolute temperature. Typically, the standard temperature is 25 °C. By increasing the test temperature, the aging process is accelerated, compressing the testing duration. However, if the temperature exceeds a certain threshold, the underlying degradation mechanisms may deviate from the Arrhenius relationship, leading to inaccurate lifetime predictions. For graphite||LiFePO₄ energy storage batteries, the main capacity fading sources are loss of lithium inventory (LLI) caused by SEI growth and electrolyte decomposition, and loss of active material (LAM) due to structural degradation and transition metal dissolution. In this study, I aim to identify the critical temperature where the aging mechanism changes and to quantify the corresponding effects.
Experimental Methods
Cell Preparation
I used 2.5 Ah graphite||LiFePO₄ soft-pack cells fabricated by a winding process. The cathode was carbon-coated LiFePO₄, the anode was artificial graphite, the separator was polypropylene (PP), and the electrolyte was LiPF₆ in a mixture of ethylene carbonate (EC), ethyl methyl carbonate (EMC), and dimethyl carbonate (DMC) with 2 wt% vinylene carbonate (VC) additive.
Cycling Tests
All cells were first subjected to a formation procedure: one cycle at 0.1 C charge/discharge between 2.5 V and 3.65 V, followed by three cycles at 1 C to determine the initial capacity C₀. Then, cycling tests were performed at constant temperatures of 25, 45, 60, 70, and 80 °C in temperature chambers (MHWX-2000). The protocol was CC-CV charge at 1 C to 3.65 V with a cut-off current of 0.05 C, rest 30 min, then 1 C discharge to 2.5 V, rest 30 min. Capacity retention was recorded every 50 cycles. Tests were stopped at predefined states-of-health (SOH) for post-mortem analysis.
Results and Discussion
Cycling Performance at Different Temperatures
Figure 2 shows the capacity retention versus equivalent cycle number (Neq) for the five temperatures. The initial discharge capacity increased by about 5% from 25 °C to 80 °C due to improved kinetics. However, the capacity fade rate accelerated with temperature. After 1200 cycles, the capacity retention was 95% at 25 °C, 90% at 45 °C, 85% at 60 °C, 80% at 70 °C, and 75% at 80 °C. To normalize the capacity differences caused by temperature, I used the equivalent cycle number Neq, defined as the total charge throughput divided by 2C₀.
The capacity loss Qloss follows the Arrhenius relationship:
$$ Q_{\text{loss}}(T, N) = A e^{-E_a / (RT)} N^z $$
where N is the cycle number, z is an exponent factor. Taking logarithms:
$$ \ln(Q_{\text{loss}}) = \ln A – \frac{E_a}{RT} + z \ln(N) $$
By plotting ln(Qloss) versus ln(Neq) for the linear region, I extracted the intercept ln(Q₀) at each temperature. The values of ln(Q₀) are linearly related to 1/T over the range of 25–60 °C, indicating a constant activation energy. However, at 70 °C and 80 °C, the data deviates from this linear trend, as shown in Figure 3. This suggests that the aging mechanism changes when the temperature exceeds 60 °C. The critical temperature is defined as the degradation mechanism transition point.
| Temperature (°C) | Capacity retention after 1200 cycles (%) |
|---|---|
| 25 | 95 |
| 45 | 90 |
| 60 | 85 |
| 70 | 80 |
| 80 | 75 |
Differential Capacity (dQ/dV) Analysis
To quantify the contributions of LLI and LAM, I conducted dQ/dV analysis for cells cycled to 90% SOH at different temperatures. The dQ/dV curve of a fresh cell shows two characteristic peaks: Peak I (around 3.35 V) corresponds to the phase transformation of LiFePO₄, and Peak II (around 3.42 V) is related to the graphite staging. The area under Peak I is proportional to available lithium inventory, while the area under Peak II is proportional to active material content. Using the method described in the literature, I calculated:
$$ \text{LLI} = \frac{S_I – S’_I}{S_I} \times 100\% $$
$$ \text{LAM} = \frac{S_{II} – S’_{II}}{S_{II}} \times 100\% $$
where S represents the peak area before cycling and S’ after cycling. Results are summarized in Table 2. As temperature increases, the proportion of capacity loss due to LLI increases significantly. At 80 °C, LLI accounts for over 75% of the total capacity loss, while LAM remains relatively small (~25%). This indicates that the dominant degradation mechanism at elevated temperatures is the loss of active lithium due to SEI formation and electrolyte decomposition, rather than structural damage to the active materials. However, at temperatures above 60 °C, the LAM fraction also increases, suggesting that material degradation becomes more pronounced.
| Temperature (°C) | LLI (%) | LAM (%) |
|---|---|---|
| 25 | 58.3 | 41.7 |
| 45 | 65.1 | 34.9 |
| 60 | 68.9 | 31.1 |
| 70 | 72.4 | 27.6 |
| 80 | 76.8 | 23.2 |
Morphology Evolution of Electrodes
Post-mortem SEM analysis was performed on graphite anodes and LiFePO₄ cathodes after cycling. Figure 6 shows surface images of fresh graphite and cycled anodes. At 25 °C and 45 °C, a thin layer of decomposition products is observed on the graphite surface. At 70 °C and 80 °C, the deposit layer becomes much thicker, indicating accelerated side reactions. Cross-sectional images (Figure 7) reveal that at 70 °C and above, internal cracks appear within the graphite particles. This is attributed to the penetration of electrolyte into grain boundaries during repeated lithium intercalation/deintercalation, generating SEI inside the particles and causing stress-induced cracking. Such cracking further exposes fresh surfaces to electrolyte, accelerating lithium loss.
For the LiFePO₄ cathodes, surface SEM (Figure 8) shows that particles cycled at 25–60 °C remain relatively smooth. However, at 70 °C and 80 °C, some large particles exhibit microcracks on the surface. Cross-section images (Figure 9) indicate that internal cracking is present even at lower temperatures (25–60 °C), likely due to anisotropic volume changes during cycling. At higher temperatures, the cracks become wider and more extensive. This is consistent with the increased LAM observed in dQ/dV analysis at elevated temperatures.
ICP Elemental Analysis of Anodes
I performed ICP-OES on graphite samples collected from cycled cells to quantify the deposition of phosphorus (P) and iron (Fe). The results are listed in Table 3. The P content increases with temperature, reflecting more electrolyte decomposition products (e.g., LiₓPFyOz) on the anode. The Fe content also rises markedly, especially at 80 °C, reaching over 800 ppm at 90% SOH and 2288 ppm at 80% SOH. This Fe originates from dissolution of LiFePO₄ due to HF generated by LiPF₆ decomposition at high temperature:
$$ \text{LiPF}_6 \rightleftharpoons \text{LiF} + \text{PF}_5 $$
$$ \text{PF}_5 + \text{H}_2\text{O} \rightarrow \text{POF}_3 + 2\text{HF} $$
$$ 2\text{HF} + \text{LiFePO}_4 \rightarrow \text{Fe}^{2+} + \text{LiH}_2\text{PO}_4 $$
The dissolved Fe²⁺ ions migrate to the anode, where they are reduced and incorporated into the SEI, catalyzing further electrolyte decomposition and lithium loss. This mechanism is one of the key factors accelerating capacity fade above 60 °C.
| Condition | P content (ppm) | Fe content (ppm) |
|---|---|---|
| 25 °C – 95% SOH | 1599.82 | 98.57 |
| 45 °C – 90% SOH | 2610.12 | 137.06 |
| 60 °C – 90% SOH | 3251.66 | 459.57 |
| 70 °C – 90% SOH | 3736.96 | 682.08 |
| 80 °C – 90% SOH | 5670.21 | 802.40 |
| 80 °C – 80% SOH | 8237.20 | 2287.79 |
XRD Analysis of Electrode Materials
XRD patterns of graphite anodes (Figure 10) show that the (002) peak position remains unchanged, confirming that the graphite layering structure is preserved. However, the full width at half maximum (FWHM) of the (002) peak increases with temperature, indicating a reduction in crystallite size (Table 4). This grain refinement is consistent with the cracking observed in SEM: micro-cracks propagate along grain boundaries, breaking the crystallites into smaller domains.
| Condition | 2θ (002) (°) | FWHM (002) | Crystallite size (nm) |
|---|---|---|---|
| 25 °C – 95% SOH | 26.444 | 0.230 | 392 |
| 45 °C – 90% SOH | 26.444 | 0.234 | 385 |
| 60 °C – 90% SOH | 26.443 | 0.237 | 380 |
| 70 °C – 90% SOH | 26.442 | 0.240 | 374 |
| 80 °C – 90% SOH | 26.440 | 0.243 | 370 |
For LiFePO₄ cathodes (Figure 11), the diffraction peaks do not shift, indicating no major structural change. However, by Rietveld refinement of the fully discharged state, I quantified the relative amounts of LiFePO₄ and FePO₄ phases (Table 5). The FePO₄ fraction increases with cycling temperature, reflecting incomplete re-lithiation due to active lithium loss. Notably, the FePO₄ fraction jumps from 16.58% at 60 °C to 18.89% at 70 °C, confirming the transition in aging mechanism beyond 60 °C.
| Condition | LiFePO₄ (%) | FePO₄ (%) |
|---|---|---|
| 25 °C – 95% SOH | 87.33 | 12.67 |
| 45 °C – 90% SOH | 85.05 | 14.95 |
| 60 °C – 90% SOH | 83.42 | 16.58 |
| 70 °C – 90% SOH | 81.11 | 18.89 |
Critical Temperature and Acceleration Factor
From the combined analysis, I clearly identify that the aging mechanism of this energy storage battery remains consistent when cycling at 25–60 °C, characterized by SEI growth and lithium inventory loss with constant activation energy. However, at 70 °C and above, additional degradation processes become dominant: graphite particle cracking, Fe dissolution and deposition, and cathode micro-cracking. This change in mechanism invalidates the simple Arrhenius extrapolation for lifetime prediction. Therefore, for accelerated life tests of LiFePO₄ energy storage batteries, the temperature should be kept below 60 °C to avoid mechanism distortion. The effective acceleration factor between 45 °C and 60 °C can be calculated from the ratio of capacity fade rates, while the 25 °C baseline is too slow for practical testing.
Conclusion
I have systematically studied the temperature-dependent capacity fading behavior of graphite||LiFePO₄ energy storage batteries. Through cycling tests, dQ/dV analysis, and post-mortem characterization, I demonstrate that the degradation mechanism transitions at approximately 60 °C. Below this temperature, capacity loss is dominated by LLI due to SEI formation, with stable activation energy. Above 60 °C, additional factors—including anode cracking, Fe dissolution, and cathode particle damage—accelerate degradation and alter the activation energy. To ensure valid acceleration tests and accurate lifetime predictions for this type of energy storage battery, it is recommended to limit the accelerated aging temperature to no more than 60 °C. The findings provide crucial insights for designing accelerated aging protocols and developing robust life models for grid-scale energy storage applications.