A lithium cell is not a container that holds electricity. It is a pair of materials that can each host lithium ions in their crystal structure, separated by a medium that passes ions and blocks electrons. Charging moves lithium from one host to the other, discharging moves it back, and the electrons forced the long way round through the external circuit do the work.
Nothing is consumed by design. Everything that degrades the cell is a side reaction that was not part of the plan.
What is actually moving
The positive electrode is typically a layered metal oxide containing lithium, or an olivine structured iron phosphate. The negative electrode is usually graphite, whose layered sheets accommodate lithium between them. Between them sits an electrolyte and a porous separator that keeps the electrodes apart while allowing ions through.
On charge, lithium ions leave the oxide lattice, cross the electrolyte, and insert themselves into the graphite. On discharge it reverses. The host structures expand and contract slightly but are not otherwise supposed to change, and the term for this reversible guest and host arrangement is intercalation. The distinction matters because it defines what wear is.
A cell does not run out of charge. It runs out of lithium still free to make the journey, and out of structure capable of hosting it.
The voltage window is a choice
The cell operates within a voltage window, commonly around three volts empty to roughly four point two full. That upper limit is a choice, not a physical boundary. Raising it extracts more lithium and yields more capacity, and it accelerates nearly every degradation mechanism at once, so the number is a negotiated position between capacity and lifetime.
Why the last stretch takes so long
Charging runs in two phases. Current is held constant while the cell voltage rises, and once the voltage reaches its upper limit the charger switches to holding voltage constant instead, letting current decay as the cell approaches equilibrium. This is why the first portion of a charge is fast and the final portion is not, and it is not a limitation of the charger.
Pushing high current into a nearly full cell would drive the voltage past its safe limit, and the electrode already accepts ions reluctantly at that point. The taper is the chemistry setting the pace.
The first and largest loss
The electrolyte is not chemically stable at the potential graphite reaches when charged. On the first charge it reacts at the graphite surface and forms a thin passivating layer, and that layer is essential, because once formed it conducts lithium while blocking further reaction. Without it the electrolyte would decompose continuously until the cell was destroyed.
Forming it consumes lithium permanently, which is why a cell is measurably lower in capacity after its first cycle than its materials suggest. That is expected and accounted for. The problem is that the layer keeps growing, thickening slowly over the life of the cell and consuming a little more lithium each time.
Growth accelerates with temperature and with electrode potential, so it happens faster when the cell is hot and faster when it is held at a high state of charge. It also happens while the cell sits idle, which is the mechanism behind the slow capacity loss of a battery that is never used at all.
When lithium stops going where it should
Intercalation takes time. Ions must cross the electrolyte, pass through the surface layer, and diffuse into the graphite. If lithium arrives faster than it can be absorbed, it has nowhere to go and deposits on the surface as metal instead.
This is plating, and it is the most damaging thing that routinely happens to a cell. Some deposited metal loses electrical contact and becomes permanently unavailable, some reacts with the electrolyte and consumes both, and the remainder can grow into needle like structures that reach the separator and create an internal short, which is the failure mode behind thermal runaway.
Two conditions cause it. High charging current, because ions arrive faster than they can be absorbed, and low temperature, because diffusion slows sharply as the cell gets cold while the charger continues delivering at the same rate. Charging a cold cell at a normal rate is considerably more harmful than charging a warm cell at a high rate, which is why devices refuse to fast charge below a threshold temperature.
The cathode wears too
Degradation is not confined to the negative electrode. Layered oxide cathodes expand and contract as lithium leaves and returns, and repeated cycling cracks the particles, exposing fresh surface to react with the electrolyte and creating regions that lose electrical connection. At high states of charge the cathode is also more chemically aggressive, transition metals can dissolve into the electrolyte and migrate to the negative electrode where they accelerate the growth of the passivating layer, and the electrolyte itself begins to oxidise.
Each effect is modest per cycle and none reverses. The compound consequence is that a cell held at full charge in a warm place is degrading through several independent pathways at once, without doing any useful work.
The mechanisms, summarised
| Mechanism | Triggered by | Reversible |
|---|---|---|
| Passivating layer growth | Heat, high charge, time | No |
| Lithium plating | Cold, high charge current | Partly |
| Cathode particle cracking | Repeated cycling | No |
| Metal dissolution | High state of charge | No |
| Electrolyte oxidation | High voltage | No |
Two kinds of ageing
Capacity fade is the loss of usable lithium and active material, and it is what the reported battery health figure describes. Impedance rise is the separate effect of the electrode surfaces becoming less conductive as layers thicken and contact degrades.
Impedance rise is why an ageing phone shuts down while showing charge remaining. Under heavy load a cell with elevated internal resistance sags below the minimum voltage the device can operate at, even though the chemistry still holds energy, and cold makes this dramatically worse. The battery has not run out, it has become unable to deliver current fast enough at an acceptable voltage.
Temperature governs the rate of nearly every mechanism here, and a widely used approximation is that degradation roughly doubles for every ten degrees of increase. A device left on a dashboard in the sun is ageing at several times its normal rate for as long as it stays there.
Why the percentage is an estimate
No practical method measures the lithium remaining in a cell directly. The management system infers it, integrating current in and out to track charge moved and periodically correcting that estimate against the cell's resting voltage. Both inputs are imperfect. Current integration accumulates error, and voltage correction requires the cell to be at rest, which rarely happens in a device in use, and depends on a relationship that shifts as the cell ages.
This is why the reading occasionally jumps and why it should be read as a well informed estimate rather than a measurement. Chemistry affects it directly, since iron phosphate cells have a notably flat voltage curve that is excellent for steady power and makes estimation from voltage considerably harder.
What follows in practice
Everything above points the same way. Time at a high state of charge accelerates surface layer growth and cathode stress, heat accelerates every mechanism, and high current charging risks plating, especially while cold.
The practical consequences are unglamorous. Keeping a cell in the middle of its range rather than parked at full measurably extends its life, which is exactly what optimised charging features implement by filling partially and completing only when the device predicts the charge will be needed. Charging slowly when there is no reason to hurry produces less heat and less stress.
Letting a cold device warm before charging avoids the worst single mechanism. And storing a device for months at full charge somewhere warm is the most effective way to damage it while doing nothing at all.
Note: the graphite electrode changes volume by roughly a tenth as lithium enters and leaves. Silicon holds far more lithium and swells by several times its own volume doing so, which is why silicon anodes promise large capacity gains and why containing that expansion has taken decades.