The Solar Efficiency Record That Does Not Matter Yet

Perovskite/silicon tandem solar cells have a certification problem and a deployment problem. The certification problem is solved. The deployment problem is not.

Researchers at the Ningbo Institute of Materials Technology and Engineering, part of the Chinese Academy of Sciences, published results in Matter on May 2100187-6) showing their perovskite/silicon tandem cell achieved a certified power conversion efficiency of 32.89 percent — 33.33 percent measured — using a technique they call peak-selective passivation. The method deposits a thin layer of aluminum oxide onto the pyramid-textured surface of industrial silicon wafers using polystyrene nanospheres as a template, blocking the electrical leakage pathways that form where perovskite meets the rough texturing that all real silicon solar cells have. After 1,000 hours of continuous operation, the device retained roughly 90 percent of its initial efficiency.

The paper is peer-reviewed and the efficiency number is real. It is also not the world record. LONGi Solar holds the confirmed NREL-certified record at 34.85 percent, set in April 2025. Unconfirmed reports place LONGi closer to 35 percent, certified by the European Solar Test Installation, but no peer-reviewed publication or independent verification has appeared. The NIMTE result matters not because it beats the record, but because of what the researchers are claiming about manufacturability.

Every other lab doing this work uses flat silicon substrates — the kind you find in research environments, not factories. The pyramid-textured surface that covers virtually all commercial silicon solar cells creates a deposition problem: it's hard to coat it uniformly with perovskite, and the non-uniformity causes electrical shorts. The NIMTE team's approach is to selectively passivate just the peaks, using cheap polystyrene nanospheres as a shadow mask during Al2O3 deposition. Their claim — published, peer-reviewed, but not yet independently verified at scale — is that this is compatible with existing industrial production lines.

That phrase is doing a lot of work and should be treated with caution until someone actually runs it on a production line.

The adoption puzzle is where this story gets interesting. Perovskite/silicon tandem cells at certified efficiencies above 30 percent have existed for years. Oxford PV shipped a commercial perovskite batch to a US solar farm in 2025 under a licensing deal with Trina Solar. Grand View Research projects the perovskite solar cell market will grow from $464 million in 2025 to $32.4 billion by 2033, a compound annual growth rate that only makes sense if you believe deployment is about to accelerate dramatically. The technology has been real for long enough that "it's coming soon" has started to sound like "it's always coming soon."

Commercial solar, meanwhile, runs on PERC and TOPCon cells — 24 to 26 percent efficient — because those are bankable. Lenders finance solar projects based on known degradation rates, known failure modes, and known manufacturing yields. A perovskite/silicon tandem cell at 33 percent is more efficient on paper. Whether it lasts 25 years in a desert, whether it degrades differently under UV exposure, whether the perovskite layer degrades in ways that aren't visible until they're catastrophic — these are questions the 1,000-hour lab stability test does not answer.

The real story is not whether perovskite/silicon tandems work in a lab. It is who gets to manufacture them at scale — and that question has a geographic answer that should concern anyone who thought this was a story about clean energy democracy.

Perovskite deposition chemistry is a specialized skill set built in a handful of places. Oxford PV, the UK-German company that licensed its perovskite IP to Trina Solar in July 2025, represents one pole: Western academic and industrial expertise that has spent a decade trying to commercialize. China represents the other. The Chinese Academy of Sciences, which produced the NIMTE result, is not a lone academic lab — it operates at state scale, with government-backed funding, manufacturing relationships, and a strategic interest in owning the next generation of solar intellectual property. The deposition expertise required for uniform perovskite films — especially on textured silicon — is not distributed globally. It concentrates in places with the research infrastructure and the industrial will to pursue it.

The mineral dependencies cut the same way. Perovskite solar cells require cesium, lead, and iodine in quantities that are not evenly distributed globally. China dominates roughly 80 percent of global critical mineral refinery production, according to Trivium China. The same nation that leads in silicon wafer manufacturing also leads in processing the minerals that next-generation cells require. This is not coincidence. It is the result of deliberate industrial policy pursued over a decade.

The silicon wafer expertise that Western and Japanese companies spent 30 years building is becoming less central to the equation. Perovskite deposition — not silicon crystal growth — is where the new value will concentrate. The companies and nations positioned to benefit most from a perovskite transition are largely the same ones that already own the silicon supply chain. This is not a story of a new technology displacing an incumbent. It is a story of an incumbent using a new technology to consolidate further.

Whether this NIMTE result changes anything depends on whether anyone runs the peak-selective passivation process on an actual production line and reports back. That has not happened yet.