A lithium-polymer cell does not have a single "capacity fade" mechanism. It has at least four, and each one leaves a distinct fingerprint on the cycle curve. Identifying which mechanism is dominant tells you which BMS or application parameter to change — and which failure mode is irreversible.
Why cycle curves look the way they do
A healthy LiPo cell loses capacity slowly and approximately linearly for the first 70–80% of its rated cycle life, then accelerates into "the knee" — a rapid capacity drop. The shape of the curve before the knee, the location of the knee, and the rate of post-knee decline are all diagnostic. Understanding the four mechanisms below helps you read those signals.
Mechanism 1: SEI growth (inevitable, manageable)
The solid electrolyte interphase (SEI) is a passivation layer that forms on the graphite anode during the first charge. A stable SEI is what makes lithium cells practical — it prevents continuous electrolyte decomposition at the anode surface. But SEI growth does not stop after formation. Every cycle adds a small amount of additional SEI material, consuming lithium from the active inventory and increasing internal resistance.
SEI growth is the dominant fade mechanism in cells operated within their rated conditions. It produces slow, linear fade starting from cycle 1 — the normal slope of a well-behaved cycle curve. High temperature accelerates SEI growth substantially: a cell cycled at 45 °C loses capacity roughly 2× faster than the same cell at 25 °C. Low charge cutoff voltage (e.g., 4.18 V instead of 4.20 V) reduces the rate by approximately 30–40%. This is the only mechanism you can slow down without sacrificing functionality — the others are either catastrophic or triggered by abuse.
Mechanism 2: Lithium plating (irreversible, triggered by charging mistakes)
When lithium ions cannot intercalate into graphite fast enough — because the charge rate is too high, the temperature is too low, or the graphite is already heavily lithiated — they deposit as metallic lithium on the anode surface. This metallic lithium does not re-intercalate during discharge; it is lost from the active inventory permanently. Plated lithium also forms dendritic structures that can pierce the separator and cause internal short circuits.
Lithium plating shows up on the cycle curve as a sudden step-down in capacity rather than the gradual fade of SEI growth. It also produces a characteristic voltage plateau during discharge that experienced engineers recognise. The triggers: charging above 1C at temperatures below 10 °C, or charging at any rate below 0 °C. The BMS fix is a temperature-gated charge rate limit: charge at C/10 below 5 °C, at C/5 below 10 °C, and at rated C-rate only above 15 °C. Low-temperature-grade electrolytes extend the safe operating window by about 10 °C.
Mechanism 3: Cathode particle cracking (chemistry-dependent)
NMC cathode materials undergo volume changes during lithiation and delithiation — typically 2–4% per cycle for NMC 523, and up to 7% for NMC 811. Over thousands of cycles, this mechanical stress fractures the cathode particles. Fractured particles expose fresh surface area that reacts with the electrolyte, accelerating local electrolyte decomposition and increasing impedance. They also create electrically isolated fragments that no longer contribute to capacity.
Cathode cracking is less common in small-format LiPo cells (which typically use LCO, HV-LCO, or NMC 111 with lower volume change) than in large cylindrical NMC811 cells for EVs. In consumer wearables and IoT devices, the more relevant concern is the cathode's HV-LCO chemistry used above 4.35 V — which can crack at its surface under repeated high-voltage cycling. Keeping the maximum charge voltage below the rated ceiling is the main mitigation.
Mechanism 4: Electrolyte depletion (accelerated by heat and HV)
Electrolyte — the LiPF₆ salt dissolved in organic carbonate solvents — is consumed over time by reactions at both electrodes. At the cathode, high voltage drives oxidative decomposition. At the anode, continued SEI growth consumes electrolyte as a reactant. At elevated temperature (above 50 °C), the LiPF₆ salt itself decomposes into HF, which attacks both the cathode coating and the copper current collector.
Electrolyte depletion shows up late in cell life as a sharp increase in internal resistance, often accompanied by the capacity knee. It is the final gating step that kills most cells — not particle cracking or plating — because by the time electrolyte is significantly depleted, the cell has already been compromised by one of the other three mechanisms.
Mechanism comparison table
| Mechanism | Typical onset | Rate of fade | Reversible? | Primary BMS parameter to control |
|---|---|---|---|---|
| SEI growth | Cycle 1 onwards | Slow, linear | No | Charge voltage ceiling, temperature |
| Lithium plating | Any cold/fast charge event | Step-change | No | Charge rate vs. temperature table |
| Cathode cracking | Mid-to-late life (after 500+ cycles) | Accelerating | No | Max charge voltage, C-rate during high-SoC phase |
| Electrolyte depletion | Late life | Rapid (at knee) | No | Temperature during operation, DoD window |
The single most effective intervention
If you can only change one parameter, change the upper charge voltage cutoff. Reducing the charge ceiling by 50–80 mV reduces the rate of both SEI growth and electrolyte decomposition, and prevents cathode cracking at the high-voltage surface. The capacity cost is 2–4% depending on chemistry. For most applications that claim a 3-year battery life target, the trade is overwhelmingly worthwhile. Cycle-life improvement of 30–50% from this single change is repeatable across LCO, HV-LCO, and NMC 111 chemistries.