Aged LiFePO₄ Batteries Show Greater Thermal Instability—But With Alarming Randomness
By Jacobin
In the rapidly evolving world of electric mobility, safety remains the silent linchpin of public trust. As automakers chase higher energy densities and faster charging times, one question lingers beneath the marketing gloss: What happens when batteries get old?
A recent study published in Battery Bimonthly delivers a sobering—and surprisingly unpredictable—answer. Researchers led by Lu Mi, professor at Xiamen University of Technology, conducted a controlled high-temperature stress test on both fresh and aged lithium iron phosphate (LiFePO₄) pouch cells, mimicking real-world thermal abuse conditions outlined in China’s national EV battery safety standard, GB 38031–2020. The results challenge conventional assumptions and underscore a critical blind spot in how battery safety is currently assessed: We’re testing new cells to certify old ones.
The stakes are high. Thermal runaway—the cascading, self-accelerating exothermic reaction that can lead to fire or explosion—is not just an engineering failure; it’s a reputational and regulatory liability. Yet nearly all regulatory certifications, third-party validations, and OEM validation protocols rely on brand-new cells to represent the entire lifecycle of a battery pack. That’s like approving a car’s crashworthiness based solely on its performance straight off the assembly line—ignoring how metal fatigue, corrosion, or worn suspension might alter outcomes after 100,000 miles.
What Lu Mi and his team discovered is that aging doesn’t just degrade capacity or increase internal resistance—it reshapes the physical geography of failure inside the cell. And far more troubling: the mode and severity of failure become less predictable, not more. In one aged cell, the aluminum-laminated pouch ruptured violently; in another—cycled more aggressively and degraded further—no rupture occurred at all. Neither caught fire, but the randomness itself is a red flag.
This isn’t a flaw in the cells tested per se. Rather, it’s a flaw in the testing paradigm.
The Hidden Geography of Degradation
To understand why aged cells behave differently under heat stress, you have to look beyond bulk metrics like state-of-health (SOH) or voltage curves. Inside a lithium-ion cell, degradation isn’t uniform—it’s topographic.
The team examined two aged 20 Ah LiFePO₄ pouch cells, both cycled between 2.50 V and 3.65 V:
- Cell #1: 1,622 cycles at 1C (20 A), retaining 90.3% of original capacity.
- Cell #2: 782 cycles with 2C charge / 1C discharge, retaining only 82.5%.
Higher charge rates accelerated degradation, as expected—but not always in intuitive ways. While Cell #2 had suffered greater capacity loss, it did not rupture during the 130°C/30-minute heating test. Cell #1—less degraded in terms of capacity—did rupture, with the inner polymer layer of the pouch splitting open laterally, exposing the wound electrode stack to ambient air.
Visually, both aged cells ballooned significantly, as did the fresh control cell (designated N). But only Cell #1 breached containment. Why?
The answer lies in where and how lithium is lost—and where it reappears.
Using post-test disassembly in an argon-filled glovebox, followed by region-specific electrochemical testing on recovered electrode patches, the researchers mapped local reactivity across the anode surface. In fresh cells, lithium distribution remained relatively even after heating. In aged cells? Not so much.
Electrode patches from Cell #2—especially from the central region (labeled a2)—showed markedly lower reversible lithium content after the heating test. Yet surprisingly, even in areas where metallic lithium had plated during cycling (a common consequence of fast charging and low-temperature operation), the underlying graphite lattice remained structurally intact, as confirmed by X-ray diffraction and Raman spectroscopy. The (002) graphite peak broadened, indicating increased disorder, but no phase collapse.
This suggests a nuanced failure mechanism: aging doesn’t necessarily “break” the graphite—it overloads parts of it. Lithium tends to plate preferentially at local hotspots: micro-regions with slightly higher current density, marginal electrode misalignment, or subtle variations in calendering pressure. Once lithium metal nucleates, it acts like a seed: subsequent cycles favor more plating at the same site. These lithium islands become electrochemically isolated “dead zones”—no longer contributing to capacity, but highly reactive when heated.
At 130°C, the solid-electrolyte interphase (SEI) begins to decompose. In fresh cells, this exposes a relatively stable, lithiated graphite surface. But in aged cells, it exposes metallic lithium—a substance that reacts exothermically with common carbonate-based electrolytes, producing gases like CO₂, C₂H₄, and CH₄. The more lithium plating, the more gas. And gas, trapped in a sealed pouch, means pressure—and eventual mechanical failure.
Yet here’s the paradox: Cell #2, with more severe aging and more expected plating, generated less catastrophic pressure buildup than Cell #1. How?
The researchers propose a counterintuitive explanation: heavy cycling may have pre-consumed reactive species. In aggressive cycling, side reactions happen during use—electrolyte is decomposed, lithium is irreversibly lost, gas may slowly vent through micro-leaks or accumulate in non-critical volumes. By the time the cell reaches end-of-life, less “fuel” remains for a violent thermal event.
Cell #1, cycled gently but many more times, may have retained more electrolyte—and thus more reactants—while still accumulating enough plated lithium to trigger intense, localized reactions during heating. The spatial clustering of reactivity led to concentrated gas evolution in one region, rupturing the pouch at its weakest seam.
It’s not just how much degradation—but what kind, where it sits, and what’s left behind—that dictates thermal fate.
The Separator: A Silent Sentinel Under Siege
Another critical finding centers on the separator—the thin polymer membrane that keeps anode and cathode apart. At elevated temperatures, conventional polyolefin separators (typically polyethylene or polypropylene blends) undergo “shutdown”: micropores melt and close, increasing resistance and (ideally) halting further current flow.
Scanning electron microscopy revealed that both aged and fresh cells experienced separator pore closure after the heating test—but the effect was more pronounced in Cell #2. Its separator showed extensive fusion and densification, correlating with higher post-test impedance measured by electrochemical impedance spectroscopy (EIS).
This might sound like a safety feature: more shutdown, less short-circuit risk, right? Not necessarily.
Excessive pore closure can create thermal gradients. Regions of the cell where the separator remains partially open may continue to conduct ions (and generate heat), while adjacent “shut down” zones stagnate. This mismatch amplifies local hotspots. Worse, if heating continues—say, in a real-world scenario where ambient temperature exceeds 130°C—the separator can shrink dramatically, pulling away from electrodes unevenly and enabling direct anode-cathode contact. Once internal shorting begins, thermal runaway can initiate in seconds.
The study notes that commercial cells increasingly use ceramic-coated separators to suppress shrinkage. While not tested here, the implication is clear: for aged cells, separator stability may be even more critical than for new ones. Yet no current safety standard mandates re-testing separators after aging.
The Randomness Problem—and Why It Matters
Perhaps the most unsettling conclusion from this work is not a specific failure mode, but the inconsistency of outcomes. Two cells, same chemistry, same form factor, same test protocol—yet one ruptures, the other doesn’t. Neither ignites, but under slightly harsher conditions (e.g., 140°C, or longer dwell time), the result could easily differ.
This randomness isn’t noise—it’s signal. It reflects microscale manufacturing variability (electrode coating thickness ±3 µm, tab welding resistance ±0.1 mΩ, pouch seal strength ±5 N) that becomes amplified by aging. In a new cell, such variations are negligible. In an aged cell, they determine whether a local hotspot becomes a runaway trigger or harmlessly dissipates.
For automakers and regulators, this is deeply problematic. Safety certifications assume repeatability. Crash tests, fire resistance, electrical isolation—all rely on deterministic outcomes. But if battery failure becomes inherently stochastic with age, then point-in-time certifications lose meaning.
Worse, field diagnostics can’t easily detect the risk. A battery management system (BMS) monitors pack voltage, temperature, and sometimes impedance—but it can’t map lithium plating distribution or separator microstructure. A cell with 85% SOH and perfectly normal telemetry could be sitting on a ticking time bomb, while its neighbor—same SOH, same usage history—is comparatively benign.
The implication? We need aging-inclusive safety protocols.
Toward a New Paradigm: Lifecycle-Aware Certification
The authors stop short of prescribing policy, but the logical next steps are clear.
First, mandatory post-aging safety validation for EV battery certification. Not just accelerated aging (e.g., 80% SOH via high-temperature storage), but dynamic aging—real cycling profiles that replicate urban stop-and-go, highway fast-charging, and seasonal temperature swings.
Second, development of non-invasive diagnostics for in-field plating detection. Techniques like incremental capacity analysis (ICA), differential voltage analysis (DVA), or even subtle acoustic signatures during rest periods may one day flag high-risk cells before they enter critical thermal environments—like summer parking in direct sun or prolonged high-power climbing.
Third, design for graceful degradation. Instead of optimizing solely for cycle life or energy density, pack engineers should prioritize failure mode consistency. Can venting be made more uniform? Can current collectors be structured to discourage localized plating? Can thermal interface materials be tuned to equalize temperature across the cell face—even as internal resistance rises?
One hopeful note: the study found that even in heavily plated regions, graphite structure remained recoverable. That opens the door for second-life applications where safety margins are higher (e.g., stationary storage) and for direct recycling of anodes without full reprocessing. Degraded ≠ disposable.
The Road Ahead
As the global EV fleet crosses 40 million units—and begins its inevitable transition from “new car smell” to “used battery anxiety”—the industry must confront an uncomfortable truth: Safety doesn’t age linearly. A cell that passes muster at Day 1 may behave unpredictably at Year 5, not because it’s defective, but because degradation is inherently heterogeneous.
What Lu Mi and his collaborators have shown is that thermal stability isn’t a fixed property—it’s a dynamic negotiation between materials, history, and microstructure. And like any negotiation, the outcome depends on who’s at the table, and how well they’ve prepared.
The next generation of battery standards must move beyond snapshot testing. They must simulate time. Because in the real world, batteries don’t fail when they’re fresh. They fail when they’re tired, uneven, and full of hidden stories—stories written in lithium, etched in graphite, sealed in aluminum, and waiting for heat to read them aloud.
Author Affiliations & Publication Info
Zufan Wang¹, Feihong Wang², Ning Ren³, Mi Lu¹∗
¹School of Materials Science and Engineering, Xiamen University of Technology, Xiamen, Fujian 361024, China
²Xiamen Products Quality Supervision & Inspection Institute, Xiamen, Fujian 361021, China
³Zhejiang Chilwee Chuangyuan Industry Co., Ltd., Huzhou, Zhejiang 313100, China
∗Corresponding author
Battery Bimonthly, Vol. 53, No. 2, Apr. 2023
DOI: 10.19535/j.1001-1579.2023.02.011