A binder that spoils the cell. New interface chemistry could unlock perovskites

Perovskite cells have been breaking efficiency records for over a decade. And yet they still haven't hit the rooftops en masse. A breakthrough published in Science (2026) shows that the key link in this puzzle - literally - lies in an interface that was supposed to help but actually hurts.

Record in the lab, disaster on the roof

The short history of perovskite photovoltaics reads like an unprecedented success story. In just over a decade, energy conversion efficiency has increased from a few percent to more than 26 percent for a single laboratory cell - a pace that silicon technology has not approached during the same period of its development. Records are broken every few months, laboratories on four continents compete in the synthesis of new structures, in compositional engineering and in the search for defect passivators.

The fitness race is fascinating, but it can be deceiving. Efficiency is a single number, photographically captured in a short test under ideal conditions. The photovoltaic industry evaluates modules by a different yardstick: after 25 years on a roof, will they retain at least 80 percent of their original efficiency? For silicon, the answer is yes, supported by thousands of installations operating since the 1990s. For perovskites, we still don't know.

The problem is not laboratory. It is field, and results from the very nature of perovskites as materials.

When the prototype cell leaves the controlled environment and goes on the roof, other forces come into play: variable temperatures, ultraviolet radiation, the humidity-dry cycle, frost, biological contamination. Perovskites based on lead halides - especially the most popular FAPbI₃ with formamidine ions - are sensitive to each of these factors. They degrade faster and in more complex ways than silicon, often in ways that are difficult to predict and reverse.

For years, researchers have focused on the absorber material itself: halide composition, crystal structure engineering, and dopants to improve durability. Less attention was paid to the thin intermediate layer - the several-nanometer molecular film between the absorber and the conductive electrode. It was a costly mistake that was only exposed by a paper published in Science in 2026.

In the standard inverted perovskite cell architecture, an indium tin oxide (ITO) electrodeindium tin oxide) is covered with a thin layer of phosphonic acids (PA)phosphonic acids). These molecules act as promoters of adhesion and functional layers: they bind the electrode to the organic hole-selective layer, improve charge transport and - as has been assumed over years of research - stabilize the entire material stack. It was assumed.

Fakt:A team led by Chengbin Fei from the Department of Applied Physical Sciences at the University of North Carolina at Chapel Hill (UNC), in cooperation with the University of Colorado Boulder and with funding from the National Science Foundation and the Office of Naval Research, published a paper in Science (January 2026; DOI: 10.1126/science.adz7969) in which they describe how commonly used phosphonic acids become the perpetrators of the degradation they were supposed to stop. The mechanism is three-stage and all stages lead to the same end - the progressive decomposition of perovskite.

When PA molecules bind weakly to the ITO surface, they can detach from it as a result of thermal and photochemical stresses and then migrate deep into the cell structure. There, in the ion-rich perovskite environment, the protic acid groups of PA catalyze a series of destructive chemical reactions.

Pierwsza:oxidation of iodide ions (I⁻). Disturbance of the charge balance by oxidized iodide initiates damage to the crystal lattice and accelerates the formation of carrier traps, which reduce the quantum efficiency of the cell.

Druga:decomposition of formamidine ions (FA⁺), which is a key component of the FAPbI₃ absorber. Formamidine is relatively thermally unstable, and the presence of acidic PA groups from the detached molecules catalyzes its decomposition much faster than it would occur in pure perovskite.

Trzecia:reduction of lead ions (Pb²⁺) to metallic lead, which precipitates inside the cell structure. Metallic lead absorbs part of the solar spectrum, permanently destroys the transparency of the layer and creates conductive paths that shorten the electrical circuit.

Fakt:All three degradation mechanisms increase rapidly at elevated temperatures (tests were conducted at 85°C - the standard temperature used in accelerated aging tests) and exposure to UV radiation. These are conditions that every photovoltaic module will routinely encounter in summer, especially in southern Europe, Mediterranean regions and tropical zones.

The team's answer was not a new absorber or abandoning the inverted architecture. The researchers only changed the chemistry of the binder itself. They synthesized phosphonic acid based on a bisphenyl skeleton with bis(diarylamino) groups, a molecule that binds to ITO much more strongly than standard PAs used so far. A stronger bond means fewer molecules breaking off, a less free catalyst in contact with the perovskite, and slower and less destructive degradation. The solution is elegant because it does not require rebuilding the entire material stack - just substituting one component in the existing process is enough.

What the work opens - and what it honestly doesn't close

The quantitative results are noteworthy but require careful contextual interpretation.

Fakt:Single cells with the new phosphonic acid reached T₉₀ - the time after which the efficiency drops to 90% of the initial value - at the level of approximately 3000 hours under accelerated degradation conditions: temperature 85°C, solar simulator with UV component, maximum power point tracking (MPP tracking). Minimodules with an active area above 20 square centimeters retained efficiency above 22% and achieved T₉₀ of 2,200 hours under the same test conditions.

What does this work open up? Above all, it provides industry with a way to safely use PA layer architectures without the risk of the support layer actively degrading the device from the inside. This is important because inverted architecture with a PA layer is common not only in unit cells, but primarily in research on silicon/perovskite tandem cells.

Silicon/perovskite tandems are one of the most promising directions in photovoltaics. They combine mature, low-cost silicon technology with the high-energy absorption of perovskite, achieving efficiencies exceeding the theoretical limit for a single junction (33% according to the Shockley–Queisser limit). Several laboratories have demonstrated monolithic tandems in the 33–34% efficiency range. If the PA–ITO interface is stable, one of the main uncertainties of the tandem material stack is addressed.

Spekulacja, uzasadniona:a more durable PA layer could shorten the distance between lab prototypes and industrial certification-ready tandems. Module certification to IEC 61215 and IEC 61730 requires documented stability under multiple stress modes. Each component of the stack that is less reactive brings the entire module closer to passing certification tests without the need for costly re-optimization.

An honest list of what the job isniesolves is as important as its achievements.

Po pierwsze— the work does not prove long-term field durability. 3000 hours at 85°C with UV is a demanding but artificial laboratory scenario. The IEC 61215 certification standard defines qualification procedures, but the conversion between accelerated test hours and years of field operation depends on climate, panel orientation and local conditions. No laboratory test can fully replace multi-year field exposure.

Po drugie— the problem of moisture sensitivity is not solved. Perovskites based on lead halide hydrolyze on contact with water vapor. The work of Fei and colleagues focuses on thermal-photochemical reactivity; hermetic encapsulation against water remains a separate and unsolved engineering challenge, requiring solutions on the side of sealing materials, not the PA layer.

Po trzecie— ion migration in the absorber volume is not stopped. The electrical voltage applied during cell operation causes the movement of halide ions inside the crystal structure, leading to spatial heterogeneity of composition and local micro-damages. The new PA stabilizes the external interface but does not directly affect the ionic dynamics in the volume of the perovskite layer.

Po czwarte— the work does not scale the result to commercial-sized modules (typically 1–2 m²). A 20 cm² aperture is a significant step, but module manufacturers must cope with uniformly depositing thin molecular layers over large surfaces while maintaining process control on an industrial scale.

Global race and Polish silence

The photovoltaics market context is crucial to assessing the significance of this work.

Spekulacja, uzasadniona:installed photovoltaic capacity is growing faster than ever before, driven by cheap silicon modules and growing decarbonization ambitions. In this environment, perovskites must find a strategic niche - either through lower manufacturing cost than silicon or through significantly higher efficiency. Silicon/perovskite tandems represent the latter route, which is why any discovery that improves interface stability has potential commercial and geopolitical implications.

The work of Fei et al. fits into the global trend of intensive interlayer engineering that dominated the perovskite literature in 2024–2026. Dozens of research groups are working on selective layers, defect passivation, and monolithic tandem integration. The discovery that a common "helper" - phosphonic acid, which has been used for years - is actually a silent aggressor, may prompt the scientific community to revise previous stability results. Some of the published data on cell life may have been underestimated due to this unrecognized effect.

The practical implications are direct. Perovskite module manufacturers and material suppliers using inverted architectures with a PA layer can potentially adopt the new binder or design derivatives with a similarly strong bond to ITO. This does not require changing the entire production process - just substituting an interface component. In a world where certification and scaling costs are the dominant barriers to entry, this type of compatible modification has real industrial value.

Fakt, z zastrzeżeniem:Quantitative data on the perovskite market - production volumes, unit prices, market penetration forecasts - are scattered and often proprietary to manufacturers or paid analytical services. No publicly available source with a verifiable DOI or cited institution has been identified for this article and will not be cited here.

Polish angle: no identified Polish angle - requires deepening.

Available public information does not indicate that national research groups or companies are involved in work on PA interfaces for perovskites in the direction described by Fei's team. Research is being carried out on perovskite materials in Poland - there are groups working on this class of materials at technical universities - however, the lack of public reports confirming direct convergence with the described discovery makes it impossible to formulate a specific statement supported by facts.Spekulacja, uzasadniona:any company or academic group using inverted architecture with a PA layer is a natural recipient of the results of this work - regardless of the country of headquarters.

The scientific conclusion goes beyond a specific material and a specific interface. Interface engineering - this thinnest, often underestimated boundary between components - is increasingly turning out to be the decisive factor in the durability of a device. Once record efficiency is technically achievable, the next battlefield is time: how long will the cell maintain its performance on a real roof, in a real climate, for real years. And it is there, in the few nanometers between ITO and perovskite, that the key battle of next-generation photovoltaics is now taking place.

Sources

1. Fei, C. et al. "Limiting phosphonic acid interlayer–perovskite reactivity to stabilize perovskite solar modules."Science 391 (6780), 2026. DOI: 10.1126/science.adz7969
2. Crossref API - Article bibliographic metadata:api.crossref.org/works/10.1126/science.adz7969[access: 2026-06-05]
3. NREL Best Research-Cell Efficiency Chart (updated continuously). National Renewable Energy Laboratory, USA. Available:www.nrel.gov/pv/cell-efficiency.html