Why Perovskite Solar Cells Wear Out, and How Researchers Are Making Them Last Outdoors

In our look at perovskite-silicon tandems leaving the lab, the commercial story was already clear: efficiency charts are no longer the bottleneck. Factories, pilot modules, and early shipments are. What still decides whether those watts become bankable projects is outdoor life—how the thin perovskite layer holds up when heat, humidity, UV, and day–night cycling stop being laboratory settings and start being weather.

That is the quieter half of the 2025–26 bet. Silicon modules already carry decades of field data and familiar warranty language. Metal-halide perovskites do not. They can be astonishingly efficient in a glovebox and fragile the first time moisture or intense light finds a weak seal. This piece stays with that problem: why perovskites degrade, which outdoor and stress-test results are starting to look encouraging, and what manufacturers still have to prove before lenders treat a tandem stack like ordinary silicon.

What actually breaks a perovskite cell

Metal-halide perovskites—often abbreviated as compounds like MAPbI₃ or FAPbI₃, depending on which organic or inorganic cation sits in the crystal—are excellent light absorbers. They are also soft ionic crystals. That chemistry is why they can be deposited from solutions at relatively low temperatures, and why they can fail in ways crystalline silicon almost never does.

Moisture is the classic villain. Water can hydrate the perovskite lattice, and with enough exposure the film can irreversibly decompose into lead iodide and other leftovers that kill the junction. The yellow tint of a ruined film is that lead iodide. Modern stacks fight this with better barrier films, edge seals, and glass-glass laminates, but packaging quality still shows up quickly in damp-heat tests.

Heat accelerates chemistry. Elevated temperatures speed ion movement, phase changes in mixed-halide films, and aging of organic transport layers. A module that looks fine at room temperature can lose fill factor when the absorber runs hot under full sun. That is one reason light-and-elevated-temperature (LT) stresses matter more for perovskites than for silicon checklists written decades earlier.

Light, especially UV, and oxygen together can drive photochemical damage. Oxygen under illumination can form superoxide that attacks the absorber, and the ISOS stability consensus notes that UV assists perovskite decomposition and can hurt cells built on mesoporous TiO₂ contacts. A 2026 Nature Energy review of perovskite modules lists the same cluster of stressors (UV light, oxygen, temperature cycling, and reverse bias once modules are wired into strings and partly shaded) and argues that standard accelerated tests often fail to predict outdoor performance.

Ion migration is the subtler failure mode. Halide and metal ions can move under an electric field and under light, creating hysteresis, reversible overnight recovery, and slow irreversible contact corrosion. A KAUST field study in Saudi Arabia found exactly that pair: a mostly reversible ion-migration loss plus silver contacts corroding into silver iodide. A cell that “recovers” overnight can still be dying by inches over months.

Why “passed IEC” is necessary but not sufficient

Silicon’s commercial language runs through standards such as IEC 61215: UV preconditioning, thermal cycling, humidity-freeze, and 1,000 hours of damp heat. Lenders understand those screens. In May 2025, Hanwha Qcells reported tandem modules from its German R&D pilot line passing critical IEC 61215-2:2021 / UL 61215 sequences—UV15, TC200, HF10, and DH1000—with tandem-appropriate multi-junction power measurement under IEC TS 60904-1-1, independently confirmed by TÜV Rheinland. The EU-funded PEPPERONI project has likewise reported encapsulated perovskite/silicon tandems meeting those IEC criteria with under 5% relative power loss.

Those milestones matter. They show mass-production-feasible stacks can survive the same type-approval hammer lenders already trust. They do not, by themselves, equal a 25- or 30-year field warranty. National-lab reliability work has long warned that classic damp-heat mainly probes packaging against moisture, while light-at-elevated-temperature stresses that matter for metal-halide perovskites are under-represented in legacy silicon suites. A 2025 PRX Energy study led by NREL researchers tested six batches of modules from four manufacturers and found that modules with a robust package that kept over 80% of peak efficiency after 100 hours of light at 55 °C were more likely to keep over 80% during 10 weeks outdoors. The authors call it a step toward a validated test that could go into a qualification standard, which tells you such a standard does not exist yet.

The community’s parallel language is ISOS (International Summit on Organic and Hybrid Photovoltaic Stability): dark, light-soak, thermal, and outdoor protocols meant to make lab claims comparable. A 2026 Chemical Society Reviews survey counted more than 43,000 perovskite devices reported between 2015 and 2025 in the Perovskite Database. More than 7,000 came with stability data tied to ISOS protocols, and of those, only 5 reports used the outdoor ISOS-O protocol. Efficiency papers remain easier to write than multi-year weather datasets.

What outdoor fleets are actually saying

Field numbers are still sparse, but they are no longer anecdotes alone.

In subtropical eastern China, a 2025 Nanoscale Advances study followed a micro power station of twenty FA_0.9Cs_0.1PbI₃ sub-modules (each 30 cm × 40 cm) for three years of continuous outdoor operation. The authors report only a 2.83% decline in power-conversion efficiency over that period, and they show that a tailored UV-to-blue accelerated aging dose (about 60 kWh/m² at 65 °C) tracked roughly two years of outdoor loss—evidence that perovskite-specific UV protocols can be predictive, not just punitive. The same paper points toward emerging draft language such as IEC TS 63624-1 for UV-related perovskite reliability work.

On the Red Sea coast, KAUST researchers previously reported a perovskite/silicon tandem retaining about 80% of its initial efficiency after a full year of outdoor operation in a hot, humid, high-irradiance climate—exactly the kind of site that stresses absorbers and packages harder than a mild temperate rack. Fill-factor fade and current-matching under real spectra and soiling were part of the story, not just peak lab efficiency. In September 2025, KAUST and Fraunhofer ISE also described a fully textured tandem reaching 33.1% efficiency with passivation compatible with industrial silicon pyramids and “extended” outdoor stability testing on that same Red Sea coast—useful as a process signal, even when the press release is thinner than a multi-year dataset.

Not every outdoor series is that gentle. A 2024 ACS Energy Letters outdoor campaign on perovskite mini-modules found the most durable configuration retaining about 78% of initial efficiency after one year, with clear burn-in loss rates and diurnal fade that partly recovered overnight—classic ion-migration fingerprints. The spread between “three years with a few percent loss” and “one year with double-digit relative loss” is the industry’s real map: architecture, encapsulation, climate, and measurement discipline still dominate outcomes.

How labs and factories are fighting degradation

The fix list is no longer mysterious; it is hard engineering at scale.

• Composition and interfaces. Moving away from fragile methylammonium-rich films toward formamidinium/cesium mixtures, mixed-halide band-gap tuning for tandems, and better surface passivation (including toughened self-assembled monolayers and diamine treatments) cuts trap-driven loss and ion generation. The 2025 KAUST–Fraunhofer passivation result is one example of making those chemistry tricks work on rough industrial silicon textures, not only on flat lab wafers.
• Contacts that do not eat the absorber. In the lead-tin perovskites used for all-perovskite tandems, a 2025 Nature Communications study traced much of the loss under heat and light to an acidic PEDOT:PSS contact that generated iodine vacancies. Swapping it for a more inert material, PTAA, brought ion-migration losses down to the level of ordinary lead perovskites.
• Encapsulation as the product. Glass-glass laminates, edge seals, desiccants, UV blockers, and barrier films are why a perovskite that hates water can still live outside. Japan’s Sekisui Chemical has said its film-type perovskite cells show durability equivalent to roughly 10 years on accelerated tests, and it is aiming for 20 years to match silicon while it scales roll-to-roll manufacturing—another reminder that lifetime and factory width are being solved together.
• Warranty ladders instead of overnight miracles. Oxford PV’s Gen 2 modules carry a 10-year product and performance warranty. In a January 2026 interview, CEO David Ward laid out targets of 26% efficiency with a 15-year lifetime in 2026, 27% with 20 years in 2027, and 30% with 30 years by 2030. That cadence matches what project finance actually buys: transparent degradation claims that lengthen only as datasets thicken.

What still has to be earned outdoors

Three gaps separate “promising” from “commodity.”

First, climate diversity. A three-year subtropical Chinese array and a Red Sea tandem year are valuable; they are not yet a global fleet spanning deserts, tropics, mountains, and freeze–thaw rooftops. ISOS-O and multi-site utility pilots (including work already underway from U.S. startups partnering with national labs and European operators) have to fill that map.

Second, standards that match perovskite physics. Passing IEC 61215 is table stakes. Validated light-plus-heat screens, reverse-bias / partial-shade protocols, and perovskite-specific UV doses—some of which are still draft technical specifications—need field correlation before they become bankability glue.

Third, module-scale honesty. Champion cells of a few square millimeters do not prove interconnects, laser scribes, or edge seals. The commercial path described in the tandem overview—M10 wafers, glass-glass modules, pilot factories—only closes when outdoor fleets of those larger formats publish transparent degradation rates lenders can underwrite.

The bottom line

Perovskites do not fail outdoors for one exotic reason. They fail for ordinary materials reasons stacked together: water chemistry, heat-driven ion motion, light-triggered reactions, and packaging that is either excellent or quickly exposed. The encouraging 2025–26 news is that those failure modes are being engineered against in factories, not only described in reviews—IEC tandem stress passes, a three-year outdoor array that lost only a few percent of its efficiency, and commercial warranty ladders that admit lifetime is still a product generation, not a press-release constant.

Efficiency made perovskites famous. Durability will decide whether silicon-plus-perovskite stacks become the default upgrade on scarce rooftops and land, or stay a specialty product with a short warranty. For the commercial context of who is shipping tandems and why silicon is the bottom cell, start with Perovskite Tandems Leave the Lab — The 2025–26 Commercial Bet.

If you want a hands-on sense of why sunlight intensity and panel orientation matter long before any tandem warranty debate, a handheld solar irradiance meter is the same kind of instrument outdoor test sites use to know what the sun actually delivered.

Measure the sun, or build with it