Wide-bandgap perovskite solar cells lose up to 60 percent of their performance in the first few minutes of thermal cycling — a figure researchers at the Technical University of Munich (TUM) have now pinned down precisely, and think they can fix with a molecular anchor. The work, published in two peer-reviewed papers in Nature Communications (January 2026) and ACS Energy Letters (February 2026), and covered by Phys.org, Scientific Frontline, and Mirage News, is incremental. That's the point.
Perovskites have been five years away from commercial viability for roughly a decade. The efficiency numbers are real — certified single-junction cells hit 27.3 percent, and tandem perovskite-silicon cells have crossed 34.85 percent in the lab. The problem is always durability. Outdoor solar panels see temperature swings of 60°C or more between night and midday. Freeze-thaw cycles in colder climates add mechanical stress on top of that. The material that looks exceptional under controlled lab conditions tends to degrade fast under real-world ones.
The TUM team, working with collaborators at the Karlsruhe Institute of Technology (KIT), the Deutsches Elektronen-Synchrotron (DESY), and the KTH Royal Institute of Technology in Stockholm, set out to understand exactly why. Using high-resolution X-ray imaging at DESY's synchrotron, they watched wide-bandgap perovskite cells breathe in real-time as temperatures cycled from roughly 5°C to 80°C at about 10°C per minute. What they observed: the perovskite crystal lattice expands and contracts on each cycle, and the mismatch between the perovskite layer and the glass substrate beneath it — the material has a thermal expansion coefficient roughly 16 times higher than glass — generates mechanical strain that accumulates. Phase transitions between crystal structures (cubic to tetragonal) compound the problem, and non-radiative recombination centers build up, hammering the fill factor and open-circuit voltage. The result is a burn-in phase where the cell loses up to 60 percent of its relative performance before settling into slower degradation.
"The microscopic tug-of-war triggers this loss," said Dr. Kun Sun, lead author and researcher at the TUM Chair of Functional Materials. "Tensions arise inside the material and its structure changes — this costs power."
The second paper proposes a fix. The team tested a class of organic spacer molecules designed to act as a molecular scaffold inside the crystal structure, holding it together mechanically during thermal cycling. Common spacers led to structural breakdown. A bulkier molecule called PDMA (pyridine-2,6-dicarboxamide) performed better as a molecular anchor, more effectively constraining lattice strain. Control devices under the same cycling protocol retained just 34 percent of their initial power conversion efficiency after 300 minutes; DP devices — those treated with the PDMA-based dual passivation strategy — retained 46 percent. Not a complete solution, but a meaningful improvement. Tandem cells using the treated wide-bandgap top layer retained 94 percent of their original performance after over 200 minutes of thermal cycling.
The champion wide-bandgap single-junction device achieved 24.31 percent power conversion efficiency with dual passivation using 3-fluorophenethylamine iodide and ethylenediamine diiodide on a 0.05 cm² cell. The full tandem cell achieved 28.64 percent efficiency at room temperature, with an open-circuit voltage of 1.908 volts and a short-circuit current density of 19.8 mA/cm².
"Perovskite solar cells are widely seen as the next big leap in photovoltaics," reads the TUM press release — which is true, and also an understatement of how long the leap has been in progress. The new work doesn't change the five-year horizon so much as narrow down which problems need solving and which approaches have merit. The burn-in mechanism is now quantified. The anchor strategy is demonstrated in the lab.
Whether it scales to module-sized cells and multi-year outdoor exposure is a different question. The thermal cycling tests here run on the order of hundreds of minutes, not the tens of thousands of hours that represent a realistic outdoor lifetime. Several commercial ventures — Oxford PV, Saule Technologies, Tandem PV — are working on similar problems with different approaches. Oxford PV announced its first commercial perovskite-silicon tandem shipment in September 2024 — real deployment, but not proof of winter durability. Multi-year outdoor validation at scale remains sparse, and a year-long Joint Research Centre outdoor test published in early 2026 found divergent stability results across modules, underscoring that the operational stability of perovskite modules under real-world conditions remains a critical unsolved challenge.
Prof. Peter Müller-Buschbaum, group leader at the TUM Chair of Functional Materials and a co-author on both papers, put the work in context: "The future of photovoltaics is tandem. By understanding these microscopic mechanics, we are paving the way for a new generation of solar modules that are both highly efficient and durable enough for decades of outdoor use."
That's a defensible framing. Tandem architectures — stacking a perovskite wide-bandgap cell over silicon or another bottom cell — are where perovskite efficiency gains make the most sense commercially, because the wide-bandgap top cell handles the high-energy photons that silicon wastes. The thermal cycling problem the TUM team identified and partially addressed is one of the specific reasons wide-bandgap cells have been harder to stabilize than their mid-bandgap counterparts. Progress, not a product.
