Efficiency records for perovskite-silicon tandem solar cells have been updated so frequently in recent years that following the raw numbers alone makes it hard to grasp the real progress being made. Against this backdrop, Germany's Helmholtz-Zentrum Berlin (HZB) announced a figure on August 26, 2026, that stopped at 30.3% (certified value: 29.7%)—falling short not only of the 35.5% industry record set by LONGi in July of the same year, but even of HZB's own 32.5% record from 2022. On the surface, this looks like a step backward. But the real story lies in the manufacturing method: instead of the solution-based coating process that struggles with scaling to large areas, HZB used a vacuum deposition technique borrowed from the OLED display industry to form the perovskite absorber layer.
Why 30.3% Isn't a "World Record"
According to HZB's announcement, the achievement involves a perovskite-silicon tandem solar cell made using a solvent-free co-evaporation method combined with an ultrathin CsCl (cesium chloride) seed layer, reaching 30.3% efficiency (certified value: 29.7%). The research was led by Viktor Skorjanc, with team leader Marcel Ross also named in the announcement. The work has reportedly been published in the journal Joule, and the corresponding arXiv preprint was submitted on November 28, 2025.
However, calling this 30.3% figure a "world record" would be inaccurate. The industry-wide record stands at 35.5%, set by LONGi on July 14, 2026, and certified by the European third-party certification body ESTI. HZB itself had already recorded 32.5% back in December 2022. The newly certified value of 29.7% falls 2.8 percentage points short of even that earlier self-record. Indeed, the original announcement avoids the phrase "world record" altogether, instead positioning the result more narrowly as an outstanding figure among vapor-deposition-based methods.
Judged purely on efficiency numbers, this announcement might look like stagnation rather than progress. But shift the evaluation axis to manufacturing method, and the picture changes. What HZB has done this time is switch the deposition of the perovskite absorber layer from a solution-based process to a solvent-free vacuum deposition process.
How OLED Vacuum Deposition Technology Improves Mass-Production Efficiency
Most perovskite absorber layers to date have been made using solution processing, in which precursor materials are dissolved in a solvent, coated onto a substrate, and then the solvent is evaporated away. Solution-coating techniques aimed at mass production, such as slot-die and blade coating, have also been under development—a joint team from KAUST and HZB reported a certified efficiency of 31.2% using blade coating. Yet these methods still face the challenge of uniformly controlling solvent drying speed and residue, and coating unevenness becomes more likely as the substrate grows larger. Mass-producing solar panels requires maintaining consistent quality across substrates ranging from tens of centimeters to a full meter in size, and this is where solution processing tends to hit a wall.
The co-evaporation method HZB adopted vaporizes the perovskite precursor materials in a vacuum and deposits them directly onto the substrate—no solvent is used in this absorber-layer deposition step. This technique repurposes vacuum deposition technology already established in OLED display manufacturing for use in perovskite solar cells, and it aligns well with the mass-production know-how already held by vacuum deposition equipment makers such as Von Ardenne and MBraun/CreaPhys. Because no solvent-handling step is needed for the absorber layer, variations in film thickness and composition across large areas become easier to suppress. That said, the hole-transport layer, made of MeO-2PACz, still relies on solution processing via spin-coating in this work, meaning the entire cell has not become fully solvent-free.
The technical key to this result is the addition of an ultrathin CsCl seed layer on top of the MeO-2PACz layer. This seed layer helps uniformize the uptake of organic precursors and prevents the localized segregation of lead iodide (PbI2) that tends to occur at the interface. As a result, the perovskite crystal grains grow larger and more uniformly, reducing defects and suppressing performance variability. Grain boundaries act as defect sites where charge carriers are prone to recombine and be lost, so enlarging grain size to reduce the total boundary area is the basic principle behind boosting both conversion efficiency and reproducibility.
Ross described the work as "a practical step toward bridging record-breaking laboratory efficiencies with reliable, industrially scalable tandem technology." It appears that the research team's real target was not the efficiency figure itself, but the feasibility of this bridge to industrial-scale production.
Seventeen Years of Efficiency Records: LONGi and HZB's Race
Research into perovskite solar cells is generally traced back to 2009, when Tsutomu Miyasaka and colleagues reported a dye-sensitized-type efficiency of 3.8%. Over the 17 years since then, the efficiency race has climbed to 35.5%. Placing the newly certified value of 29.7% within that trajectory, it trails LONGi's world record of 35.5% by 5.8 percentage points and HZB's own past record of 32.5% by 2.8 percentage points. Comparing only third-party-certified figures, HZB's vapor-deposition-based achievement still lags the industry's top level by more than 5 points.
Looking at the pace of LONGi's record updates: 33.9% in November 2023, 34.6% in June 2024, followed by 34.85%, then 35.2%, and finally 35.5% in July 2026. That's a gain of 1.6 points over roughly two years and eight months, showing LONGi has steadily kept pushing efficiency higher.
HZB, meanwhile, built up its own record with 29.15% in 2020, 29.8% in November 2021, and 32.5% in December 2022, but has since 2023 been overtaken in efficiency by LONGi and other competitors. Using HZB's certified value of 29.7% as the baseline, the 5.8-point gap with LONGi corresponds to roughly a 20% difference in power output, while the 2.8-point gap with HZB's own past record corresponds to about a 9% difference. This trajectory since 2023 suggests a deliberate shift away from the efficiency race and toward establishing a manufacturing process capable of mass production.
In its own technical materials, LONGi points to the theoretical maximum efficiency of a single-junction solar cell (the Shockley-Queisser limit at an optimal bandgap, 33.7%) to argue that a tandem structure could theoretically reach around 43%. However, this 33.7% figure represents a general upper bound at an ideal bandgap (about 1.34 eV); the theoretical limit based on silicon's own bandgap (about 1.12 eV) is closer to about 29.4%. The record for actual efficiency of single-junction silicon cells has been creeping closer to that 29.4% ceiling year by year—in May 2026, JA Solar and Gold Stone Energy reportedly achieved 28.2% under TÜV Rheinland certification (earlier reports from various sources cited figures of 26.81%, 27.09%, and 27.3%).
The Wall Facing Mass Production—And Japan's Point of Connection
The biggest obstacle to mass production is durability. Perovskite is known to have lower resistance to heat, humidity, and UV light than silicon, causing it to degrade faster. For commercial viability, the period during which output retains 80% of its initial value (T80) is generally considered to need to exceed 25 years. This 25-year benchmark roughly matches the warranty period typically set for existing silicon panels, and is seen as the minimum bar the perovskite camp must clear to catch up with silicon.
The fact that these compositions contain lead is also a source of environmental and health concern. The RoHS Directive, which restricts the use of hazardous substances in electronic equipment, includes an exemption for fixed installation solar panels designed and installed by specialists—but this doesn't mean all solar panels are uniformly excluded. Concerns about the risk of lead leaching into the environment remain persistent, and management systems for disposal and recycling stages have been flagged as a challenge. These two issues—durability and lead content concerns—remain concrete conditions that will determine whether mass production becomes viable.
Comparing this to an example that has actually reached commercialization highlights the distance that still separates laboratory records from real products. According to Oxford PV, the company began shipping the world's first commercial tandem module in September 2024, though that module's efficiency was 24.5%. Oxford PV claims this delivers up to 20% more power generation compared to a standard silicon-only panel. Its facility in Brandenburg, Germany, has a production capacity of around 100 MW—still a substantial gap remains between laboratory record values and commercial product performance.
As for developments in Japan, Sekisui Chemical announced on March 27, 2026, its decision to commercialize a film-type perovskite product called "SOLAFIL," and has begun supply discussions for projects in Saitama City, Shiga Prefecture, Fukuoka Prefecture, and other sites selected under Ministry of the Environment programs, as well as for Tokyo's Air Solar early-adoption initiative. As a lightweight, film-type single-junction cell, this falls into a different category from the high-efficiency tandem cells on crystalline silicon substrates that HZB works with. While Japan's display industry has long operated vacuum deposition processes for mass-producing OLED panels, no concrete plan to repurpose that equipment or know-how for perovskite solar cell mass production has been confirmed in either Sekisui Chemical's or HZB's announcements so far. What Sekisui Chemical's efforts do show is that, regardless of differences in deposition method, Japanese companies are independently exploring mass-production technology for perovskite solar cells.
Beyond Efficiency: The Next Step as a Manufacturing Technology
Neither the original article nor the arXiv preprint touches on concrete cost estimates for mass production, durability data, or a commercialization timeline. Numbers that would determine practical viability—such as yield rates on production lines or the retention of efficiency when scaled up to large-area panels—are also not provided. This announcement simply marks a technical milestone in the deposition process; the distance to an actual product has yet to be quantified.
Even so, reading this 30.3% figure as a "step backward" would be a mistake. What HZB has chosen is progress along a different axis than efficiency: shifting the deposition of the perovskite absorber layer from a solution-based process to a mass-production-capable vapor deposition process. Efficiency records will likely continue to be broken by LONGi and other companies going forward, but there's no guarantee those records can be reproduced as-is on a mass-production line. Rather than getting caught up in efficiency numbers alone, watching which manufacturing method reaches mass production lines first offers a more accurate way to gauge where this technology truly stands.
Ross's words about "bridging toward reliable, industrially scalable tandem technology" point precisely to where this technology is headed next. As durability data is disclosed and reproducibility at large scale is demonstrated, the path toward mass production should become more concrete. The next number HZB needs to show isn't a percentage—it's multi-year durability test data.