A team of researchers at Helmholtz-Zentrum Berlin (HZB), led by physicist Steve Albrecht, says it has pushed solar cell efficiency to 34.5% in the lab by stacking two ultra-thin perovskite layers on top of a conventional silicon cell, a triple-junction design that squeezes more electricity from sunlight than any previously tested device of its kind. The result, published in Nature in early 2026, has been independently certified, though it has not yet appeared on the U.S. National Renewable Energy Laboratory’s widely referenced Best Research-Cell Efficiency Chart.
“We have shown that adding a second perovskite absorber on top of silicon can push well beyond what tandems alone achieve,” Albrecht said in a statement accompanying the Nature publication. “But we are the first to say that a record on a small cell is only the starting point.”
The number matters because standard rooftop silicon panels sold today typically convert between 22% and 24% of incoming sunlight into electricity. A jump to 34.5%, even in a small lab cell, suggests that layered perovskite-silicon designs could eventually deliver substantially more power per square meter. But the distance between a thumbnail-sized prototype built under pristine conditions and a weatherproof panel bolted to a roof remains enormous, and the research raises as many questions as it answers.
How the cell works and what has been verified
Each layer in the triple-junction device is tuned to absorb a different band of the solar spectrum. The top perovskite layer captures high-energy blue and ultraviolet photons, the middle perovskite layer handles a portion of the visible range, and the silicon base collects lower-energy red and infrared light. By splitting the work this way, the cell wastes less energy as heat compared to a single-junction panel that can only efficiently convert one slice of the spectrum.
According to the Nature paper, the HZB researchers used advanced techniques to manage how charge carriers and photons move between layers, reducing losses at the interfaces where one material meets another. The bandgaps of each sub-cell were carefully matched so that voltage and current are shared efficiently across the stack. The 34.5% efficiency figure was measured by an independent testing facility under standardized conditions, as described in the paper.
Two other peer-reviewed results published in the same timeframe help frame the achievement. A team at Nanjing University reported a certified 33.6% efficiency for a flexible perovskite-silicon tandem, a two-junction design that trades the third absorber layer for a lightweight, bendable substrate. Separately, researchers at the King Abdullah University of Science and Technology (KAUST) published a paper in Nature Communications documenting certified steady-state efficiencies of 31.6% and 31.14%, measured by Europe’s JRC-ESTI testing authority, for perovskite-silicon tandems scaled to 60 cm². That same KAUST study reported that efficiency dropped to 28.9% at the 60 cm² size, down from the 31.6% steady-state figure recorded on smaller samples. The larger-area result is notable because efficiency almost always falls as cell size increases, and 60 cm² is far closer to the dimensions needed for commercial panels than the tiny samples typically used in record-setting experiments.
All three studies were peer-reviewed, and all report certification by recognized independent labs. Those are meaningful credibility markers. NREL’s efficiency chart, which compiles verified records submitted by research groups worldwide under measurement standards defined by IEC 60904-3 and ASTM G173, remains the definitive global scoreboard. If the 34.5% result is added to that chart, it will have cleared an additional layer of scrutiny.
The durability problem
Efficiency records grab headlines, but longevity determines commercial viability. Perovskite materials are notoriously sensitive to moisture, oxygen, and heat. Conventional crystalline silicon modules are qualified through punishing accelerated aging tests, including damp heat exposure, thermal cycling, and UV bombardment, that approximate 25 years of outdoor service. Manufacturers typically guarantee at least 80% of original power output after that period.
No comparable degradation data have been published for this particular triple-junction cell. The Nature paper describes the device’s structure and its certified efficiency measurement but does not include results from long-term stability testing or accelerated aging protocols. Related perovskite research has shown steady improvement in durability as scientists refine compositions, passivation coatings, and encapsulation methods, but the specific material stack used here has not been publicly subjected to the kind of stress tests that silicon panels routinely pass.
“Stability is the elephant in the room for every perovskite result,” said Jenny Chase, head of solar analysis at BloombergNEF. “Until a group publishes transparent degradation data under standard protocols, the industry will treat these numbers as science, not engineering.”
A cell that starts at 34.5% but loses significant output within a year would have little practical value. Until the researchers or an independent group publishes transparent degradation data for this device, the efficiency figure should be understood as a snapshot taken under controlled conditions, not a guarantee of real-world performance.
The scaling gap
The 60 cm² tandem cells from the KAUST-led Nature Communications study offer a useful reality check. As noted above, efficiency at that larger area dropped to 28.9%, a meaningful decline from the 31.6% steady-state figure recorded on smaller samples in the same research. That pattern has been consistent across perovskite studies for years: as the surface area grows, defects, pinholes, and thickness variations become harder to control, and performance suffers.
Scaling a triple-junction cell, which requires depositing two uniform perovskite films rather than one, compounds the challenge. No manufacturer has publicly outlined a production pathway from this lab result to commercial modules, and no independent cost analysis has estimated what it would take to deposit those layers at factory throughput. Questions about yield loss, materials utilization, and equipment costs remain unanswered in the available literature.
For context, the solar industry spent roughly two decades driving single-junction silicon from lab curiosity to the dominant electricity source it is today, with manufacturing scale, supply chain maturity, and relentless cost reduction doing as much of the work as any single efficiency breakthrough. Perovskite-silicon devices are far earlier in that arc.
Comparing the three results carefully
It is tempting to line up the headline numbers (34.5%, 33.6%, and 31.6%) and declare a winner, but the three studies used fundamentally different designs. The triple-junction device adds an extra perovskite absorber to capture more of the spectrum. The flexible tandem sacrifices some structural rigidity for a lightweight form factor that could open applications on curved surfaces or portable equipment. The 60 cm² study prioritizes scaling, accepting lower peak efficiency in exchange for a cell size that begins to resemble what factories would actually produce.
NREL’s chart accounts for these differences by sorting records into technology subcategories (single-junction silicon, two-junction perovskite-silicon tandem, multijunction concentrator, and others) so that each design competes against its own class. Readers should apply the same discipline: these are three data points on related but distinct development paths, not a single horse race.
What evidence would move the needle
Several pieces of evidence, none yet available, would substantially strengthen or weaken the case for this technology reaching the market. Published degradation data showing the triple-junction cell retaining at least 80% of its output after 1,000 hours of damp heat or thermal cycling would be a significant confidence builder. A pilot manufacturing study demonstrating that two perovskite layers can be deposited uniformly over areas larger than 200 cm² without catastrophic efficiency loss would address the scaling question directly. And an independent levelized-cost-of-electricity analysis comparing triple-junction modules with today’s commercial silicon and emerging tandem products would tell the industry whether the added complexity is worth the efficiency gain.
Until that evidence accumulates, the 34.5% figure is best understood as a verified laboratory milestone, one that demonstrates the theoretical headroom in perovskite-silicon architectures but does not, on its own, predict when or whether that potential will translate into panels on rooftops. The research trajectory is promising. The finish line is not yet in sight.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
