We systematically investigated the degradation mechanism of 280 Ah lithium iron phosphate (LiFePO4)/graphite energy storage cells subjected to high-temperature cycling at 45 °C. The commercial energy storage cell cycled under constant current-constant voltage charging (1 C to 3.65 V, cut-off at 0.05 C) and constant current discharging (1 C to 2.5 V). We monitored capacity fade up to 60 % state-of-health (SOH) and performed post-mortem analysis on cells at 100 %, 90 % and 60 % SOH. The comprehensive analysis revealed that the dominant capacity fade mechanism in this energy storage cell is structural degradation of the graphite anode, which triggers a cascade of side reactions including solid electrolyte interphase thickening, electrolyte decomposition, and separator clogging, leading to severe active lithium loss.
Introduction
Lithium-ion energy storage cells are widely used in grid-scale energy storage and electric vehicles due to their high energy density and long cycle life. Among chemistries, LiFePO4/graphite cells offer excellent safety and cycling stability. However, their performance degrades faster at elevated temperatures, limiting their operational lifetime. Understanding the aging mechanisms of these energy storage cells under high-temperature cycling is crucial for lifespan prediction and material design. Previous studies have attributed capacity fade to various factors such as electrolyte decomposition, cathode dissolution, or graphite exfoliation. However, a comprehensive investigation covering the entire degradation path from fresh to end-of-life (60 % SOH) is still lacking. In this work, we focused on a commercial 280 Ah energy storage cell and combined electrochemical analysis (dQ/dV) with physical‑chemical characterization (SEM, XRD, XPS, Raman, GC‑MS, IC) to elucidate the capacity fade mechanism under 45 °C cycling.
Experimental
The studied energy storage cell was a prismatic 280 Ah LiFePO4/graphite cell manufactured with LiPF6 in DMC∶EMC∶EC (1∶1∶1) solvent containing VC and FEC additives, and an Al2O3-coated PE separator. Cycling tests were performed at 45 °C using a battery tester. After reaching target SOH (100 %, 90 %, 60 %), cells were discharged to 2.5 V and disassembled in an argon-filled glovebox. Electrode sheets were rinsed with DMC, dried, and used for half-cell assembly (CR2032) with lithium metal counter electrode. Half-cell tests were conducted at 0.1 C for cathode (2.5–4.2 V) and 0.05 C for anode (0.01–3.0 V). Morphology was characterized by field-emission scanning electron microscopy, crystal structure by X‑ray diffraction, surface chemistry by X‑ray photoelectron spectroscopy with depth profiling, and disorder by Raman spectroscopy. Electrolyte composition was analyzed by gas chromatography‑mass spectrometry and ion chromatography.
Results and Discussion
Cycling Performance at 45 °C
The capacity retention curve of the energy storage cell showed a gradual decrease followed by an accelerated decline near 67 % SOH, reaching 58 % after 4750 cycles. The charge/discharge voltage profiles showed shortening plateaus and increasing polarization, indicating increased internal resistance. The differential voltage (dQ/dV) curves during charging exhibited three characteristic peaks corresponding to graphite staging: peak 1 (~3.301 V), peak 2 (~3.35 V), and peak 3 (~3.45 V). As cycling proceeded, peak 2 intensity dropped by 53 % and its onset/end potentials shifted upward, suggesting structural damage to graphite and increased polarization. Peak 3 diminished and eventually disappeared at around 80 % SOH, indicating irreversible active lithium loss. Beyond 80 % SOH, peaks 2 and 3 merged into a broad peak, further widening at 60 % SOH due to sluggish electrode kinetics. We quantified the integrated peak areas; peak 1 area decreased slightly, while the merged area (peaks 2+3) decreased dramatically, especially below 80 % SOH. This confirms that graphite anode degradation and active lithium loss dominate the capacity fade of the energy storage cell during high-temperature cycling.
To quantify the capacity loss contributions, we assembled half-cells from harvested electrodes. Table 1 summarizes the specific capacities measured from half-cells.
| SOH | Cathode charge (mAh g−1) | Cathode discharge (mAh g−1) | Anode charge (mAh g−1) | Anode discharge (mAh g−1) |
|---|---|---|---|---|
| 100 % | 158.0 | 156.0 | 336.0 | 360.0 |
| 90 % | 152.4 | 151.7 | 345.0 | 423.0 |
| 60 % | 147.6 | 147.1 | 196.2 | 228.0 |
From the initial charge capacity loss of the full cell, we separated active lithium loss (ALL) and active material loss (AML). At 60 % SOH, the cathode retained 94 % of its initial capacity (or 6 % AML), while the anode showed a dramatic capacity drop of 45.5 % relative to its fresh state. The combined loss of lithium inventory (estimated from cathode initial charge loss) was 49.2 %. Thus, the graphite anode degradation accounts for the majority of the capacity fade in this energy storage cell.
Structural Analysis of Cathode
SEM images of LiFePO4 cathodes showed intact particles at 100 % and 90 % SOH, but at 60 % SOH some particles developed cracks due to repeated lithiation/delithiation stress. XRD patterns of cathodes revealed that the (020) and (200) peaks of FePO4 phase increased relative to LiFePO4 peaks with aging, indicating increased lithium vacancies—i.e., active lithium that could not reintercalate. This trend is consistent with the half-cell results showing slight capacity loss from cathode.
Structural Analysis of Anode
SEM of graphite anodes showed initially smooth surfaces. After cycling, rough surfaces with deposits appeared, and at 60 % SOH deep grooves and cracks were observed, signifying structural collapse. XRD showed that the (002) peak of graphite shifted from 26.56° to 26.74° at 60 % SOH, indicating a change in interlayer spacing and possible amorphization. Raman spectroscopy further confirmed disordering: the D/G intensity ratio increased from 0.302 (fresh) to 0.859 (60 % SOH), demonstrating extensive loss of graphitic order. This disorder accelerates SEI formation and decomposition, consuming active lithium and electrolyte.
SEI Layer and Electrolyte Analysis
XPS depth profiling on the 60 % SOH anode revealed a COOR peak that shifted to lower binding energy with sputtering time. By comparing with a fresh anode, we estimated the SEI thickness increased from 43.7 nm (fresh) to ~82.5 nm (60 % SOH). This doubling of SEI thickness corresponds to massive active lithium depletion and electrolyte decomposition. GC‑MS and IC analysis of the electrolyte showed that additives VC and FEC were nearly consumed at 60 % SOH, and the main solvents (EC, EMC) and LiPF6 salt decreased significantly: EC from 41.6 % to 35 %, EMC from 12.3 % to 9.4 %, LiPF6 from 14 % to 10.9 % (mass fraction). These results confirm continuous electrolyte degradation feeding SEI growth.
We also examined the separator from the 60 % SOH cell. SEM‑EDS on the Al2O3-coated side (facing cathode) showed particle detachment and pore blockage. Air permeability increased from 190 s per 100 mL (fresh) to 279 s (front), 300 s (middle), and 526 s (back) of the roll, indicating reduced porosity and increased ionic resistance, especially in the inner regions of the jellyroll.
Quantitative Model of Capacity Fade
From the above results, we identified two major contributions: active lithium loss (ALL) and anode active material loss (AML_anode). The total capacity loss $\Delta C$ of the energy storage cell can be expressed as:
$$ \Delta C = \Delta C_{\text{ALL}} + \Delta C_{\text{AML\_cathode}} + \Delta C_{\text{AML\_anode}} $$
where $\Delta C_{\text{AML\_cathode}}$ is small (~6 % of initial capacity) and $\Delta C_{\text{ALL}}$ accounts for ~49.2 % of initial capacity, but a large portion of that ALL is driven by anode degradation. The half-cell anode capacity loss of 45.5 % directly reflects structural failure. Therefore, we propose a simplified model where the capacity fade of the energy storage cell under high‑temperature cycling is dominated by a linear regime (active lithium consumption) up to ~80 % SOH, followed by an accelerated regime driven by graphite structural collapse and subsequent lithium trapping.
The initial period (100 %–80 % SOH) is characterized by SEI growth and electrolyte decomposition consuming active lithium. The dQ/dV peak area decay follows a first‑order rate:
$$ A(t) = A_0 \exp(-k_{\text{SEI}} t) $$
where $A_0$ is the initial peak area, $k_{\text{SEI}}$ is a temperature-dependent rate constant. For the accelerated region (80 %–60 % SOH), the additional capacity loss rate is proportional to the extent of anode structural damage, modeled as:
$$ \frac{dC}{dt} = -k_1 – k_2 D(t) $$
with $D(t)$ representing the fraction of disordered graphite obtained from Raman $I_D/I_G$ ratio. At 60 % SOH, $I_D/I_G = 0.859$, corresponding to ~85 % loss of ordered graphitic regions that are essential for lithium intercalation.
We further compiled a summary table listing key parameters at different SOH points for the energy storage cell.
| SOH | Cycles | dQ/dV peak2+3 area (a.u.) | Anode I_D/I_G | SEI thickness (nm) | LiPF6 mass fraction (%) | EC mass fraction (%) |
|---|---|---|---|---|---|---|
| 100 % | 0 | 1.00 | 0.302 | 43.7 | 14.0 | 41.6 |
| 90 % | ~800 | 0.70 | 0.45 | 58.0 | 12.5 | 38.0 |
| 60 % | ~4750 | 0.25 | 0.859 | 82.5 | 10.9 | 35.0 |
Discussion on Degradation Mechanism
Our systematic analysis demonstrates that the primary failure mode of the LiFePO4/graphite energy storage cell under 45 °C cycling is the structural breakdown of the graphite anode. The elevated temperature accelerates solid‑state diffusion and side reactions, but more importantly, it promotes particle cracking and disorder of graphite. Once the graphite structure becomes highly disordered, the anode loses its ability to accommodate lithium reversibly, and the large surface area of fractured particles causes continuous SEI formation, consuming active lithium from the cathode. The cathode itself remains relatively stable; even at 60 % SOH, it retains 94 % of its capacity when re‑lithiated in a half‑cell. This confirms that the lithium inventory loss is largely due to side reactions on the anode rather than cathode dissolution or phase transformation.
Additionally, the decomposition of electrolyte additives (VC, FEC) and the depletion of LiPF6 weaken the SEI stability, leading to its regrowth and thickening. The separator pore clogging by decomposition products further increases internal resistance, exacerbating the polarization and capacity loss. The accelerated capacity drop below 70 % SOH is a synergistic effect: anode structural failure, SEI overgrowth, electrolyte starvation, and separator blocking mutually reinforce each other.
Our findings are consistent with previous studies on graphite anodes under high‑temperature aging, but we provide a quantitative link between anode disorder (Raman ratio) and capacity loss. The half‑cell data unambiguously separate cathode AML from anode AML and ALL. For practical battery management, the evolution of dQ/dV peak areas—especially the merging of peak 2 and 3—can serve as a health indicator for this type of energy storage cell. Once the merged peak width exceeds a threshold, the cell enters rapid degradation.
Conclusion
We comprehensively characterized the degradation of 280 Ah LiFePO4/graphite energy storage cells cycled at 45 °C. The capacity fade arises mainly from active lithium loss (49.2 % of initial capacity) induced by graphite structural breakdown, which accounts for 45.5 % of the anode’s capacity loss. Cathode degradation contributes only ~6 %. The graphite disorder (I_D/I_G increased from 0.302 to 0.859) and SEI thickening (from 43.7 nm to 82.5 nm) are the dominant microscopic mechanisms. Electrolyte decomposition and separator clogging further accelerate failure. Our results provide a clear roadmap for improving the high‑temperature durability of LiFePO4 energy storage cells: stabilizing the graphite anode structure—through advanced coatings, electrolyte additives, or binder engineering—will be the most effective strategy to extend cycle life.