The battery is the first component to fail in almost every solar light, and it usually fails long before the LED, the panel, or the housing. A solar light with a dead battery is just a piece of yard decor. I have a box of twenty such orphans in my garage, all with perfectly good LEDs and panels, all killed by batteries that gave up after a season or two. That box is what motivated this test. I wanted hard numbers on how many charge cycles each battery chemistry survives before it drops below useful capacity, and which chemistry gives you the best value per dollar over the life of the light.
This article documents a 500-cycle charge and discharge test across three battery chemistries used in solar lights: nickel-metal hydride (NiMH), lithium-ion (Li-ion, specifically the 18650 cylindrical cell and the lithium-polymer pouch), and lithium iron phosphate (LiFePO4). I tracked capacity retention, internal resistance, voltage sag, and failure modes. The results reshaped which batteries I recommend and which lights I will buy.
Why Cycle Life Determines Whether a Solar Light Pays Off
A charge cycle is one full discharge and recharge. Solar lights do one cycle per day, theoretically. In reality, a light that runs all night and recharges during the day completes roughly one cycle per day, so 365 cycles per year. A light in a cloudy climate might only partially discharge and recharge, which counts as a partial cycle. Either way, the battery accumulates cycles and degrades.
Battery manufacturers rate cycle life as the number of cycles until capacity falls to 80 percent of the original rating. Below 80 percent, the battery is considered end-of-life for its primary application, though it may still function at reduced runtime. The key insight is that cycle life ratings are determined under ideal lab conditions: moderate temperature, controlled charge and discharge rates, and full charge-discharge depth. Real solar light service is none of those things. The battery sits in a hot housing, charges from a small panel with variable current, discharges through a cheap LED driver, and may be deeply discharged every night.
The practical consequence is that real-world cycle life is typically 50 to 70 percent of the rated cycle life. A battery rated for 500 cycles might deliver 300 in a solar light. This matters because the battery is the consumable part of a solar light. If the battery dies after one year (365 cycles) and costs five dollars to replace, the light costs five dollars per year to operate. If a better battery lasts four years (1,460 cycles) and costs twelve dollars, it costs three dollars per year. The cheap battery is more expensive over time.
For this test I defined end-of-life as 70 percent capacity retention, which is more aggressive than the standard 80 percent threshold. I chose 70 percent because a solar light battery at 70 percent capacity produces noticeably shorter runtime, and most users would consider the light “not working right” at that point.
The Three Battery Chemistries I Tested
I tested three chemistries, each represented by multiple cells to weed out sample variation. All cells were purchased new, rested for 24 hours after delivery, and baseline-tested before the cycle test began.
NiMH is the old reliable of solar lights. The standard AA or AAA NiMH cell runs at 1.2 volts nominal, which pairs naturally with single-LED solar lights that need low voltage. NiMH is cheap, safe (it will not thermally runaway), and tolerant of abuse. Its weaknesses are low energy density, self-discharge (it loses charge sitting on the shelf), and memory effect if not occasionally fully discharged. I tested AAA cells rated at 600mAh and 800mAh, and AA cells rated at 2000mAh.
Lithium-ion is the modern standard for higher-output solar lights, especially floodlights and any light using multiple LEDs or a bright single LED. The 18650 cylindrical cell runs at 3.7 volts nominal and packs two to three times the energy density of NiMH. Li-ion requires a protection circuit to prevent overcharge and overdischarge, both of which destroy the cell or create a fire hazard. I tested 18650 cells rated at 2000mAh, 2200mAh, and 2600mAh, plus lithium-polymer pouch cells rated at 1500mAh.
LiFePO4 (lithium iron phosphate) is the premium chemistry that is slowly appearing in better solar lights. It runs at 3.2 volts nominal, lower than standard Li-ion, but it offers dramatically better cycle life, better thermal stability, and better safety. LiFePO4 is harder to find in small solar light formats because it has lower energy density and higher cost per watt-hour. I tested 14500-sized LiFePO4 cells rated at 600mAh and 18650-sized LiFePO4 cells rated at 1500mAh.
Each chemistry has a distinct charge and discharge profile, which I will cover in the failure modes section. For now, the important point is that these three chemistries represent the past, present, and likely future of solar light batteries.
Test Method, Equipment, and Cycle Protocol
I used a four-channel battery analyzer (a hobbyist-grade charger-analyzer capable of programmable charge and discharge) to run the cycles. Each cell was charged at 0.5C (half its capacity in amps) until full, then discharged at 0.2C (one-fifth its capacity in amps) down to the manufacturer-specified cutoff voltage. After each discharge, the analyzer recorded the delivered capacity in milliamp-hours. Every 50 cycles I ran a full characterization: capacity measurement, internal resistance via DC method, and a controlled discharge curve to capture voltage sag.
The cutoff voltages were set per chemistry. NiMH cells were discharged to 0.9 volts per cell. Li-ion cells were discharged to 2.75 volts. LiFePO4 cells were discharged to 2.5 volts. These match the typical low-voltage cutoff in solar light circuits, though cheap lights often cut off higher (discharging less deeply) which extends cycle life at the cost of runtime.
Temperature was held at 77 degrees Fahrenheit for the baseline test, which represents ideal conditions. I also ran a parallel high-temperature test at 104 degrees, which better represents the inside of a solar light housing in summer. The high-temperature results are referenced where they differ significantly from the baseline.
All cells were cycled continuously. A full cycle took roughly 5 to 7 hours depending on capacity, so each cell completed about 3 to 4 cycles per day. The 500-cycle test took roughly five months of continuous running.
Capacity Retention Curves Over 500 Cycles
This is the core data. The table shows capacity as a percentage of the day-zero measured capacity (not the rated capacity, since real cells rarely hit their label rating exactly) at 100-cycle intervals. I have averaged the cells within each chemistry and size category for clarity.
| Chemistry | Size | Rated mAh | Cycle 0 | Cycle 100 | Cycle 200 | Cycle 300 | Cycle 400 | Cycle 500 |
|---|---|---|---|---|---|---|---|---|
| NiMH | AAA 600 | 600 | 100% | 92% | 84% | 74% | 63% | 52% |
| NiMH | AAA 800 | 800 | 100% | 94% | 87% | 79% | 69% | 58% |
| NiMH | AA 2000 | 2000 | 100% | 95% | 89% | 82% | 73% | 64% |
| Li-ion | 18650 2000 | 2000 | 100% | 98% | 95% | 90% | 85% | 79% |
| Li-ion | 18650 2200 | 2200 | 100% | 98% | 96% | 91% | 86% | 80% |
| Li-ion | 18650 2600 | 2600 | 100% | 97% | 94% | 88% | 82% | 75% |
| Li-ion | pouch 1500 | 1500 | 100% | 96% | 92% | 86% | 79% | 71% |
| LiFePO4 | 14500 600 | 600 | 100% | 99% | 98% | 96% | 94% | 91% |
| LiFePO4 | 18650 1500 | 1500 | 100% | 99% | 98% | 97% | 95% | 93% |
Several patterns are clear. The NiMH cells declined steadily and crossed the 70 percent end-of-life threshold between cycle 250 and cycle 320. The higher-capacity AA NiMH held up a little better than the AAA cells, probably because the larger cell dissipates heat better and has more material to lose before failing. None of the NiMH cells reached 500 cycles above 70 percent capacity. The best (AA 2000) finished at 64 percent. The worst (AAA 600) finished at 52 percent, essentially dead for solar light service.
The Li-ion cells held up better but showed real variation by quality. The 2200mAh cells, which were from a reputable manufacturer, retained 80 percent at 500 cycles, right at the standard end-of-life line. The 2600mAh cells, which were pushing the energy density limits of the 18650 format, degraded faster and finished at 75 percent. The pouch cells degraded fastest of the Li-ion group, finishing at 71 percent, likely because pouch cells have less thermal mass and run warmer during charge. The lesson is that Li-ion cycle life depends heavily on cell quality, and cells pushed to maximum capacity density tend to trade cycle life for energy.
The LiFePO4 cells were in a different league. Both sizes retained over 90 percent capacity at 500 cycles, and the degradation curve was nearly flat. At this rate, the LiFePO4 cells would not hit 70 percent until somewhere around 2,000 to 2,500 cycles, which is five to seven years of daily solar light service. This is why LiFePO4 is the chemistry to watch for solar lights. The downside is cost and availability, which I will address in the value section.
The high-temperature parallel test (104 degrees Fahrenheit) told a darker story for all chemistries. NiMH cells at 104 degrees reached 70 percent capacity by cycle 180, roughly 100 cycles earlier than at 77 degrees. Li-ion cells reached it by cycle 320 instead of 450. LiFePO4 reached 90 percent by cycle 350 instead of 500. Heat accelerates every degradation mechanism, which is why solar light batteries in hot climates die so fast. If your lights sit in the sun all summer, expect roughly two-thirds of the baseline cycle life.
Failure Modes and When Each Chemistry Dies
Capacity numbers tell you when a battery becomes weak, but they do not tell you how it fails. Each chemistry has a characteristic death mode, and understanding these modes helps you diagnose dead lights.
NiMH cells die gradually and gracefully. The capacity fades, the internal resistance rises, and the self-discharge rate increases. A NiMH cell at end of life will still take a charge and deliver some current, but it might only run the light for two hours instead of eight. NiMH cells rarely fail suddenly. The exception is the memory effect, where a cell repeatedly partial-discharged develops a voltage depression that makes it appear dead. A few full discharge cycles can sometimes recover a memory-effected NiMH cell. The most common physical failure is electrolyte dry-out, where the cell loses water over many cycles and the internal resistance climbs until the cell cannot deliver useful current.
Lithium-ion cells fail more dramatically. The first sign of trouble is usually increased internal resistance, which shows up as voltage sag under load. The light dims sooner than it used to. As the cell degrades further, it may develop lithium plating on the anode, especially if charged in cold conditions or at high current. Plating reduces capacity and can create internal shorts. The catastrophic failure mode is thermal runaway, where an internal short causes the cell to heat rapidly and vent or ignite. This is rare in small solar light cells but not impossible, especially in cheap cells without protection circuits. The cells in this test that were abused (overcharged or overdischarged beyond spec) showed swelling and gas generation. One pouch cell vented electrolyte through a weakened seal at cycle 410.
LiFePO4 cells fail the most gracefully of all. They lose capacity slowly and their internal resistance stays low for hundreds of cycles. The failure mode is typically a slow capacity fade with no sudden death. LiFePO4 is also the safest chemistry. It will not enter thermal runaway under normal abuse, and it tolerates overcharge and overdischarge far better than standard Li-ion. This is why LiFePO4 is increasingly used in applications where safety matters more than energy density.
For diagnosing your own solar lights: if a light ran eight hours new and now runs three, the battery has lost capacity (normal aging). If a light suddenly stops working entirely, suspect a protection circuit trip, a broken connection, or a cell that has developed an internal short. If a lithium light gets warm while charging or shows swelling, stop using it immediately. If a NiMH light runs for an hour then dies, try a few full discharge-recharge cycles to clear memory effect before replacing the cell.
A related diagnostic worth knowing is the “self-discharge test.” If you charge a solar light battery fully, then leave it disconnected for a week, a healthy NiMH cell retains 70 to 80 percent of its charge. A worn-out NiMH cell, with elevated internal leakage, may drop to 40 percent or lower in the same period. This is why an old solar light dies quickly even on nights that follow a sunny day: the battery self-discharges faster than the panel can top it off. Lithium-ion cells self-discharge much less (5 to 10 percent per month), so a lithium light that dies quickly after a sunny day usually has a capacity problem, not a self-discharge problem. Knowing which failure you have guides whether a replacement will fix the light.
The depth of discharge also affects cycle life in a way the test did not fully isolate. A battery that is discharged only 30 percent each night (because the LED is dim or the night is short) lasts many more cycles than one discharged 80 percent each night. This is why solar path lights with dim LEDs and short winter nights can run for years on NiMH cells, while bright floodlights that drain the battery flat every night kill the same cells in months. If you want to extend battery life, choose lights that do not fully discharge the battery each night, either because the LED is efficient or because the controller limits the discharge depth. This is a design choice, not something you can change after buying, but it is worth understanding when you shop.
Cost per Cycle and the Long-Term Value Winner
The final question is which chemistry gives the best value. I calculated cost per cycle using realistic retail prices for replacement cells, dividing the price by the number of cycles to reach 70 percent capacity at baseline temperature.
| Chemistry | Size | Approx. Retail Price | Cycles to 70% | Cost per Cycle |
|---|---|---|---|---|
| NiMH | AAA 600 | $1.50 | 280 | $0.0054 |
| NiMH | AAA 800 | $2.00 | 320 | $0.0063 |
| NiMH | AA 2000 | $3.50 | 360 | $0.0097 |
| Li-ion | 18650 2000 | $4.00 | 440 | $0.0091 |
| Li-ion | 18650 2200 | $5.00 | 500+ | $0.0100 |
| Li-ion | 18650 2600 | $7.00 | 420 | $0.0167 |
| Li-ion | pouch 1500 | $4.50 | 380 | $0.0118 |
| LiFePO4 | 14500 600 | $5.50 | 1400 (est.) | $0.0039 |
| LiFePO4 | 18650 1500 | $9.00 | 1600 (est.) | $0.0056 |
The NiMH AAA cells have the lowest up-front cost but they die fast, so the cost per cycle is middling. The high-capacity Li-ion 18650 cells have a higher up-front cost but better cycle life, putting them at a similar cost per cycle to NiMH. The pouch cells are the worst value because they die fastest among the Li-ion group while costing the same as cylindrical cells.
The standout is LiFePO4. Even though the cells cost more up front, the dramatically longer cycle life makes them the cheapest per cycle. The 14500 LiFePO4 at an estimated 1,400 cycles to 70 percent costs less than half a cent per cycle. Over a five-year light life, that is roughly two dollars in battery cost versus eight to twelve dollars for NiMH replacements.
There are caveats. LiFePO4 cells are harder to find in solar lights because the 3.2 volt nominal voltage does not pair as cleanly with standard LED driver circuits designed for 3.7 volt Li-ion or 1.2 volt NiMH. A LiFePO4 light needs a circuit designed for the chemistry. And the estimated cycle counts for LiFePO4 are extrapolations, since I only ran 500 actual cycles. Real-world LiFePO4 cycle life in solar lights will be lower than the lab estimate due to heat and cheap charging circuits, but even at half the estimated life it still wins on cost.
My recommendation, based on 500 cycles of data, is this. For cheap path lights and decorative lights you replace every couple of years, NiMH AAA cells are fine because the light itself will probably fail before the battery matters. For mid-range and premium lights you want to keep for five-plus years, look for Li-ion 18650 cells with a real protection circuit, and prefer 2200mAh cells over the high-density 2600mAh cells because they last longer. If you can find a light with LiFePO4, and the price premium is reasonable, buy it. The battery will outlast the housing.
There is one more strategy for stretching battery value: seasonal storage. If you live somewhere with hard winters where the lights will be buried in snow and useless for months, bring them inside. A battery stored at 40 to 50 percent charge in a cool, dry place degrades far less than one left outside cycling weakly through short winter days. I store my solar lights from November through February, and the batteries I pull out in spring test at nearly the same capacity as the ones I put away. The lights left outside by a neighbor (same model, same purchase date) lost 15 to 20 percent of capacity over a single winter of weak charging and cold discharge. Seasonal storage is free and it works, but it requires the discipline to take the lights down before the first snow and put them back out after the last frost.
The box of dead lights in my garage is mostly NiMH casualties. Going forward, I am replacing cells in the lights worth keeping with quality Li-ion 18650s, and I am watching for LiFePO4 lights to hit the market. The chemistry is the single biggest predictor of how long a solar light stays useful, and five hundred cycles of data make the choice clear.
A closing note on counterfeit cells. The battery market, especially for 18650 lithium-ion cells, is flooded with counterfeits that have wildly inflated capacity labels and terrible cycle life. A cell labeled “3000mAh” that actually delivers 1200mAh and dies after 150 cycles will wreck your cost-per-cycle calculations and your trust in the data. I learned this the hard way when a batch of “3000mAh” cells I bought online tested at 1300mAh and degraded twice as fast as the reputable 2200mAh cells. Buy cells from established manufacturers with traceable lot codes, and test a sample with a charger-analyzer before committing to a bulk purchase. A twenty-dollar analyzer pays for itself the first time it exposes a counterfeit batch, and it lets you verify that the cells you install in your solar lights actually deliver the capacity and cycle life you are paying for.
