Does a battery degrade even if unused?

Yes. The film on the anode surface (SEI) slowly thickens by reacting with electrolyte, locking away lithium. Growth slows as the film thickens, so loss is well approximated as proportional to the square root of time. Here, at 25 °C, loss is 3.0% after year 1 and 4.24% after year 2 — the second year adds less than half of the first.

For EV batteries, does driving or parking cause more degradation?

It depends on use. For a commuter car (0.1 cycles per day), 8.5% of the 12.6% loss after 8 years is calendar ageing, more than the 4.1% from cycling. For a delivery van cycling daily, cycle loss reaches 14.4% and exceeds calendar loss.

How much does temperature affect life?

A lot. With a 50 kJ/mol activation energy, ageing runs 1.92 times faster at 35 °C than at 25 °C, and 3.55 times at 45 °C. In the commuter example, time to 20% loss falls from 19.7 to 5.6 years. Activation energy varies with cell chemistry and strongly drives the result, so it must be determined from tests.

Battery Degradation Simulation

Category: Coupled Analysis – Electrochemistry | Revised 2026-10-01

Theory: calendar and cycle ageing

Overview

🙋

Isn't battery life set by how many times it can be charged?

🎓

Not by count alone. Li-ion capacity falls even when the cell just sits. The anode carries a thin film (SEI) formed by reaction with electrolyte; as it slowly thickens it consumes lithium, reducing what is usable. This is calendar ageing. Meanwhile each charge and discharge swells and shrinks electrode particles, cracking them and forcing the film to re-form — cycle ageing. Both run faster at higher temperature. Life estimates separate the two and add them according to use and temperature.

Semi-empirical model

$$ Q_{cal} = k_{cal}\,a(T)\,\sqrt{t},\qquad Q_{cyc} = k_{cyc}\,a(T)\,N^{0.55},\qquad Q = Q_{cal} + Q_{cyc} $$

$$ a(T) = \exp\!\left[\frac{E_a}{R}\left(\frac{1}{T_{ref}} - \frac{1}{T}\right)\right] $$

t is in days and N is equivalent full cycles. Here, 1 year at 25 °C gives 3.0% (k_cal), 1,000 cycles at 25 °C give 8.0% (k_cyc), and activation energy E_a = 50 kJ/mol. Simply adding the two is an approximation; in reality they interact.

Coffee Break Trivia

The film that protects and consumes

Peled named the anode surface film the SEI in 1979. It blocks electrons but passes lithium ions, stopping further electrolyte breakdown; the Li-ion cells commercialised in 1991 depend on this film forming stably. But it never stops completely and keeps slowly thickening, so the very mechanism that protects the cell slowly consumes it.

Worked examples

Example 1: calendar ageing (25 °C)

Years	1	2	4	8
Capacity loss	3.00%	4.24%	6.00%	8.49%

Being proportional to the square root of time, loss after 4 years is only twice year 1 and after 8 years about 2.8 times. Relative to 25 °C, temperature speeds it 1.92 times at 35 °C and 3.55 times at 45 °C.

Example 2: loss after 8 years by use and temperature

Case	Cycles in 8 yr	Calendar	Cycle	Total	Years to 20%
Commuter 0.1 cycle/day, 25 °C	292	8.5%	4.1%	12.6%	19.7
Commuter 0.1 cycle/day, 35 °C	292	16.3%	7.8%	24.2%	5.6
Van 1 cycle/day, 25 °C	2,920	8.5%	14.4%	22.9%	6.2
Van 1 cycle/day, 35 °C	2,920	16.3%	27.8%	44.1%	1.8

Calendar ageing dominates for the commuter, cycle ageing for the van. In both, a 10 °C rise in average temperature cuts life sharply.

🙋

For the commuter, 10 °C changes life from 19.7 to 5.6 years? That seems too big a difference.

🎓

That instinct matters; two effects stack here. First, temperature makes ageing 1.92 times faster. Second, calendar ageing follows the square root of time, so the time to reach the same 20% shrinks with the square of the rate. A 1.92× rate thus becomes more than a 3× difference in life. But the result depends strongly on activation energy: 50 kJ/mol is just an example, values vary with materials and state of charge, and at high temperature other mechanisms (electrolyte breakdown, cathode changes) can join in. And real cell temperature is set by charge/discharge heating and cooling, not ambient air. So life prediction couples battery thermal and ageing models, feeding in temperature histories for each drive pattern and climate. Storage and cycling tests at several temperatures to fit the constants are essential.

Modelling workflow

Run storage tests at several temperatures and states of charge to fit calendar constants.
Fit cycle constants from tests varying depth of discharge and current.
Use a battery thermal model to get temperature histories for each use pattern.
Feed temperature and state-of-charge histories into the ageing model and accumulate over time.
Revise constants against capacity records from vehicles or equipment in service.

Coffee Break Trivia

“Batteries that faded only in hot regions”

One vehicle model showed capacity fade varying greatly by region: with similar mileage, cars in hot regions faded nearly twice as fast. Investigation found battery temperature rising while parked in the daytime, often at full charge for long periods; calendar ageing accelerates with high temperature combined with high state of charge. Recommending battery cooling while parked and a lower daily charge limit slowed subsequent fade.

Common mistakes

Mistakes and fixes

Mistake	Effect	Fix
Life from cycle count only	Life overestimated for light use	Include calendar ageing
Ageing at ambient temperature	Self-heating acceleration missed	Couple with thermal model
Linear in time	Early loss under, long-term over	Square-root model
Fitting tests at one temperature	Activation energy undetermined	Test at several temperatures
Ignoring state of charge	Full-charge storage ageing missed	Function of state of charge

🙋

I'd like to learn related topics.

🎓

Rate this article

Thank you for your feedback!

Helpful

More details

Report error