Solar cells push efficiency beyond limits
- LONGi’s perovskite-on-silicon tandem solar cell reached a certified 34.85% efficiency, pushing past the single-junction silicon ceiling that has constrained mainstream solar for years. - The key threshold is 33.7% — the classic Shockley-Queisser limit for one junction — and tandem designs are now clearing it with lab-verified results. - That matters because the race is shifting from “can this beat silicon?” to “can this be built cheaply, durably, and at scale?”
Solar cells are running into a weirdly good problem. The old efficiency ceiling for plain silicon is no longer the main story. The new story is that stacked cells — especially perovskite-on-silicon tandems — are now posting certified numbers above the limit that shaped solar engineering for decades. LONGi’s 34.85% tandem result made that shift impossible to ignore, and newer triple-junction work is pushing the frontier even further. ### What limit are people talking about? The famous limit is the Shockley-Queisser limit — basically the maximum efficiency a single-junction solar cell can reach under normal sunlight. For a one-junction device, it sits around 33.7%, and practical silicon cells top out well below that. Commercial silicon panels are usually far lower still, which is why squeezing out another percentage point has gotten so hard. (pv-magazine.com) ### Why does silicon hit a ceiling? A single layer of semiconductor has to make one compromise. Tune it to catch high-energy photons well, and lower-energy light slips through. Tune it the other way, and high-energy light gets wasted as heat. That is the core tradeoff. Silicon has been optimized incredibly well, but one material trying to harvest the whole solar spectrum is like using one fishing net for every fish in the ocean — some of the catch always gets away. (pv-magazine.com) ### So what is the trick? Stack more than one cell. A tandem cell puts one absorber on top of another so each layer handles a different slice of sunlight. Perovskites are useful here because their bandgap can be tuned more easily than silicon’s, so a perovskite top cell can grab higher-energy light while the silicon bottom cell takes what passes through. That is how a tandem architecture can beat the single-junction ceiling without breaking physics — it changes the game being played. (nature.com) ### What actually changed in the lab? The clearest milestone is LONGi’s certified 34.85% two-terminal perovskite-silicon tandem cell, announced in April 2025 and validated by NREL. That result beat LONGi’s own earlier 34.6% mark and became the first certified double-junction tandem to clear the 33.7% single-junction limit. In plain English — this stopped being a theoretical talking point and became a measured device result. (nature.com) ### How did they get there? A lot of it comes down to interfaces — the places where one layer hands charges to the next. LONGi used a bilayer passivation strategy with lithium fluoride and EDAI molecules to reduce recombination losses and improve charge extraction. That sounds niche, but it is the whole battle in advanced solar cells: not just absorbing light, but getting the electrical charge out before it dies inside the material. (pv-magazine.com) ### Are tandems the only frontier? No — triple-junction cells are moving too. EPFL and CSEM reported a certified 30.02% perovskite-perovskite-silicon triple-junction cell in March 2026, using two perovskite layers plus silicon in one stack. That number is lower than LONGi’s best tandem, but the architecture has a higher long-run ceiling, with researchers arguing the path could go well above 40% if the materials and optics keep improving. (pv-magazine.com) ### If the numbers are so good, what is the catch? Durability and manufacturing. Perovskites can be efficient, but turning a record lab cell into a panel that survives heat, moisture, UV exposure, and years on a roof is the hard part. Reviews of the field keep landing on the same bottlenecks — stability, scalable fabrication, and integrating these stacks into large-area products without losing performance. ### Is this getting closer to the market? (pv-magazine.com) Yes, cautiously. U.S. companies Solx and Caelux said in April 2026 that they plan domestic tandem modules at 28% efficiency, with commercial volumes targeted for 2027. That is below the best lab records, but that gap is normal. The important shift is that companies are now choosing manufacturable architectures and talking in gigawatts, not just square centimeters. (nature.com) ### Bottom line The big change is not just that one solar cell got a flashy number. It is that the industry now has credible evidence that silicon’s old ceiling is no longer the ceiling that matters. The next fight is brutal but clearer — keep the efficiency, keep the stability, and make the stack cheap enough to win. (pv-magazine.com)
