Solar cells just got a fresh efficiency record, and it comes from a promising combination of two thin-film technologies. Researchers at the Helmholtz-Zentrum Berlin (HZB) and Humboldt-Universität zu Berlin have set a new mark of 25.5% power conversion efficiency for a tandem solar cell that pairs a perovskite top layer with a copper-indium-gallium-selenide (CIGS) bottom layer.
The result, announced in early July 2026, was independently certified by the European Solar Test Installation (ESTI) and entered into the definitive Solar Cell Efficiency Tables published by the University of New South Wales. It edges out the same team's previous record of 24.6% and marks another step in a long climb toward the theoretical ceiling of tandem architectures.
Why tandem cells matter
A single-junction silicon solar cell has a hard efficiency limit — it can only absorb a portion of the sun's spectrum well. Tandem cells stack two absorbers with different bandgaps so that each layer captures a different slice of sunlight. The top perovskite layer soaks up higher-energy visible light, while the CIGS layer beneath it grabs lower-energy near-infrared photons that perovskite lets slip through.
Unlike perovskite-silicon tandems, the perovskite-CIGS pairing is fully thin-film. That opens the door to lighter, flexible, and potentially cheaper panels — useful for building facades, vehicles, and other places where rigid silicon modules struggle to fit.
How they pushed past 24.6%
To squeeze out the extra percentage point, the SOLMATES project team ran through a careful stack of engineering tweaks:
- Two different CIGS bottom cells with bandgaps of 1.05 eV and 1.1 eV were paired with matched aluminum-doped zinc oxide layers of different thicknesses.
- Multiple combinations of nickel oxide (NiOx) and self-assembled monolayers (SAMs) were screened as hole transport layers to reduce interfacial recombination losses.
- The electron-selective contact was refined by carefully controlling the initial thermal evaporation rate of buckminsterfullerene (C60) onto an ultra-thin, one-nanometer lithium fluoride passivation layer.
Each change nudged more of the incoming light into usable electric current — and cut down on the invisible losses that plague every solar cell.
The certified cell itself is small — 1.081 square centimeters — but the team also fabricated a mini-module using the same materials over 2.25 cm², achieving about 19.7% efficiency. That gap between lab cell and module is normal, and closing it is the next challenge as researchers move from record-chasing to something you could actually put on a roof.
"25.5% is only a stepping stone"
HZB researcher Guillermo Farias Basulto struck a confident tone in commentary accompanying the announcement, noting that the underlying physics of the current architecture suggests more room to grow. In-house testing of similar stacks has already reached 27.5% efficiency — a figure that, once independently certified, would put perovskite-CIGS tandems firmly in the same league as more established perovskite-silicon tandems that have crested 34% at the research level.
The SOLMATES project, funded by the European Union, is built around exactly this kind of iteration — pairing perovskite and CIGS specifically because the combination could scale to lightweight, flexible modules manufactured with lower-temperature processes than silicon requires.
What this means for solar's future
Commercial silicon panels sold today typically deliver around 20–22% efficiency. Every efficiency point matters: it means fewer panels, less land or roof area, and lower balance-of-system costs to generate the same electricity. Tandem architectures like this one are the most credible path to breaking through silicon's single-junction limits without exotic materials or extreme temperatures.
There's still work to do on stability, area scaling, and manufacturing before perovskite-CIGS modules land on rooftops. But records like 25.5% — properly certified, methodically improved — are the milestones that turn "someday" solar technologies into real products. Given how quickly this team has been raising its own bar, the number to beat may not stand for long.


