The Verification Venue · the folklore everyone repeats, measured
The Climb Past the Shoulder
You've been told to charge to 80%, not 100%. Almost no one tells you why, or what it costs. The usual story — "100% instantly damages the battery" — is wrong: nothing snaps at the top. What's really happening is a trade, and you can drag it.
A lithium-ion cell doesn't hold one voltage. It rests higher the fuller it is — and the last fifth of the charge is crammed into a narrow band of volts right at the ceiling. That's the shoulder. Sitting up there, at high voltage, speeds up the slow chemistry that ages the cell. So a higher charge ceiling doesn't break anything today; it just spends cycles faster. Move the ceiling and watch both numbers move at once.
Drag from full down toward 80% and watch the resting voltage drop off the steep shoulder while the projected cycle life climbs — and the usable capacity you trade away grows. The cycle-life numbers are a band from two studies, not a single promise.
Cell rests at
4.20 V
on the shoulder — highest aging stress
Cycles to ~80% health (band)
300–500
baseline · the most you'll get
Usable capacity
100%
full charge, no trade
Here's the thing the folk model misses. From 4.0 V to 4.2 V — a swing of just 0.2 V — the cell goes from 73% to 100% full. The last 27% of the capacity lives inside that thin slice of voltage at the very top. To buy it you have to push the cell up onto the steepest, highest-voltage part of the curve and park it there, which is exactly where it ages fastest. The 80% rule isn't superstition; it's stepping back from the shoulder.
Why does high voltage age the cell? At high state of charge the graphite anode sits at a very low potential, which drives the electrolyte to keep reacting at its surface — growing the solid-electrolyte interphase (SEI), the film that slowly eats your cyclable lithium. It's not a switch that flips at 100%; it's a rate that rises with voltage, and the damage is gradual, accumulating over weeks and cycles (it grows roughly as √t). Keil et al.'s figures show high-SOC loss on the order of ~2–3× the moderate-SOC loss (read off their SOC-vs-fade plots) — the direction is settled science; the exact multiplier is not.
The check — every number recomputed in front of you
The resting voltage comes from a measured charge-voltage→capacity curve (PowerStream), inverted by linear interpolation — not a fitted formula. For your chosen ceiling of 100%:
The cycle-life number is shown as a band, because the real magnitude is cell- and temperature-dependent. The green row is your ceiling; the two sources broadly agree on direction and roughly agree near the 80% rule, and they disagree on the exact size — which is the honest point:
| ceiling | rests at | BU-808 cycles | usable cap. | PowerStream × |
|---|
—
Cross-check: both datasets independently put 4.00 V at 73% capacity. Run it yourself — node research/why-the-last-twenty-percent-costs-most/verify-why-the-last-twenty-percent-costs-most.mjs.
What's measured, what's cited, and every free choice
The OCV curve is one cell's measurement. The resting-voltage curve is PowerStream's published charge-voltage→capacity experiment on a small Li-poly cell. Because it was taken while charging at a low current, the voltages sit a hair above true open-circuit voltage (a small overpotential). Graphite's staging plateaus make the curve lumpy, not a clean smooth shoulder — so the robust claim is the one the data shows plainly: the cell rests higher the fuller it is, and the top fifth of charge (80→100%) spans about 0.16 V, roughly twice the 0.08 V of the middle fifth (40→60%). Different NMC/graphite cells shift the exact numbers; the compressed top-of-charge band is general.
The cycle-life trade is a band, on purpose. Two sources: Battery University's BU-808 table gives cycle ranges per charge voltage (e.g. 4.20 V → 300–500 cycles; 4.00 V → 850–1500); PowerStream gives a relative factor Ef = 2^(10·(4.2−V)) (4.10 V doubles, 4.00 V quadruples the cycle life). For "charge to ~80% (4.0 V) instead of 100%" these imply a multiplier of roughly 1.7× to 5× (BU-808 band) with PowerStream landing at 4× — inside the band. The popular "3× lifespan" claim sits in there too, but it's a single point in a wide, conditions-specific spread, not a law. Neither source is peer-reviewed primary data; they're widely-cited industry measurements, and the page labels them as such.
The mechanism is cited, not fitted. That higher voltage / higher state of charge accelerates aging — via low anode potential → electrolyte reduction → SEI growth, with fade ∝ √t — is from Keil et al., J. Electrochem. Soc. 163(9), A1872 (2016). The paper states the direction (high SOC ages faster, with a step near ~60% SOC), the √t fade, and the SEI / anode-potential cause; the ~2–3× magnitude is our reading of their SOC-vs-fade figures, not a number they report. We assert the shape they measured (rises with SOC, steps up past mid-charge, square-root in time); we do not refit their dataset on the page.
What's gradual vs instantaneous. Nothing is destroyed at 100%. "Damage" here is slow accumulation; a cell occasionally topped to 100% is fine. The 80% rule matters most for a cell that lives near full and warm (a laptop on a desk, a phone on a charger all night). Heat and high charge ceilings compound — the worst case is hot and full.