Should a home battery sit at 100 percent?

Most home batteries sold today use lithium iron phosphate (LFP) cells, and most of them offer some version of a backup reserve setting: hold the battery at or near full so it is ready when the grid drops. Owners reasonably ask whether that is slowly wearing the battery out. The honest answer is "somewhat, but the study people cite for it was asking a slightly different question."

What the Dahn lab actually did

The paper is "The Operation Window of Lithium Iron Phosphate/Graphite Cells Affects their Lifetime" by Eniko Zsoldos, Daphne Thompson, William Black, Saad Azam and Jeff Dahn, published open access in the Journal of The Electrochemical Society on 21 August 2024.

They used small 240 mAh LFP/graphite pouch cells (made dry by a manufacturer, then filled with the lab's own electrolytes) and charged and discharged them over five state-of-charge windows: 0–25%, 0–60%, 0–80%, 0–100% and 75–100%. The cycling ran at a C/3 rate for about 1,400 cycles, roughly 2,500 hours, with a slow full-range check-up cycle every 200 cycles to measure how much capacity was left. To speed up aging, the tests ran hot: 40 °C and 55 °C.

The ranking came out exactly in order of average charge level. Best to worst capacity retention: 0–25%, 0–60%, 0–80%, 0–100%, 75–100%. The conventional full 0–100% window landed in the middle.

The size of the effect is the striking part. For the cells built with the first of their two graphites, the authors write that temperature changed fade rates by roughly 15–50% and electrolyte salt by 5–30%, while the state-of-charge window changed them by 250–400%. In their words, average state of charge "was found to be the most critical factor" for fade, ahead of temperature, depth of discharge, salt and graphite choice, within these tests.

Why high charge hurts, per the paper

LFP cells mostly die by losing usable lithium (the authors call it lithium inventory loss, or shift loss), not by cathode damage. The paper identifies two routes that get worse at high charge. First, fully lithiated graphite reacts more with the electrolyte, and the authors measured that reactivity rising step by step with state of charge. Second, a chain of side reactions dissolves iron from the cathode and deposits it on the graphite, where it drives more electrolyte breakdown and more lithium loss.

One practical worry they also checked: does cycling at low charge cost you power, since cell voltage is lower there? Barely. LFP's flat voltage curve meant the average discharge voltage was 3.3 V in the 75–100% window and 3.15 V in the 0–25% window.

The part that rarely makes it into the advice

Everything above is about cycling: charging and discharging repeatedly while the cell sits in a given range. The same paper also ran a 1,000-hour voltage hold at 60 °C, with cells parked at 0% or 100% charge and not cycled. Those cells showed no observable iron deposition above background at either charge level. The authors conclude that the iron route "is independent of SOC during storage," that it happens because of the expansion and contraction of cycling, and that their results support the literature finding that "cycling is more detrimental to capacity fade than storage in LFP cells."

That does not mean storage at full charge is free. The hold experiment measured iron, not total capacity. The paper's own literature review notes that earlier work found storage at high charge does cause more lithium loss in both NMC and LFP cells, through slow growth of the graphite's surface film. And a widely cited 2012 calendar-aging study by Kassem and colleagues stored commercial LFP/graphite cells at 30, 45 and 60 °C and 30, 65 and 100% charge for up to eight months. Its abstract reports that capacity fade "strongly increases with storage temperature and to a lesser extent with the state of charge," and that cells stored at 30 °C showed no capacity fade over that period.

So the evidence lines up like this. Cycling up near full is the expensive case, and the Dahn result is strong there. Sitting full costs something too, but in the storage studies, heat matters more than charge level.

What this means for a battery at home

A backup-only battery that is held near full and rarely discharged is mostly a storage case. The Dahn cycling numbers overstate the penalty for it. The bigger lever is where the battery lives: a cool garage wall beats a sun-baked exterior wall. (Report 114 covers why LFP is especially temperature-sensitive.)

A self-consumption battery that charges from solar every day and discharges every evening is a cycling case. If it fills to 100% daily and spends evenings in the upper part of its range, it is doing something close to what faded fastest in the study. If your system lets you set the daily ceiling and you don't need the top 20% most days, this paper points toward a lower ceiling. The authors themselves suggest 0–80% (or 0–60%) as a reasonable trade, and say cycling only 0–25% would waste too much capacity to recommend.

A mixed setup is the trade every owner actually faces: reserve capacity available at the moment of an outage versus slower aging. There is no free answer. For a home in hurricane country, a full battery when the storm arrives may be worth more than a few percent of lifetime.

Where the study doesn't reach

These were lab-filled pouch cells, not commercial home-battery cells. Absolute lifetimes don't transfer, and since every dataset here shows cooler cells aging more slowly, a battery in a normal garage should outlast these test cells.

The tests ran hot on purpose. At 40 °C and 55 °C, the relative ranking of windows is the finding to carry over, not the hours.

"100%" on your app is not 100% in the paper. The paper's windows were defined by capacity on a lab cycler. The percentage a home battery shows is defined by its battery management system and its charging procedure (Report 150 covers this). A manufacturer may already hold back some of the top of the cell.

One study disagrees at high current. The authors flag work by Stroe and colleagues that found the opposite (lower average charge, more fade) when cycling at 4C and 42 °C, possibly because of lithium plating at that speed. Home batteries cycle far slower than that, closer to the Dahn conditions, but the disagreement is a reminder that the effect depends on how hard the cells are run.

I care about this because I help design the AI battery-cycling systems for a veteran-owned (HUBZone) energy-storage integrator, run by fellow veterans, and when to fill a battery and how far is exactly the kind of decision that logic makes. I do not own that company and earn nothing from this link. Full policy here. Nothing in this report involves them or any product they sell.