A tandem solar cell crosses twenty-eight per cent
A team at the Institute of Chemistry in Beijing has built a perovskite-organic tandem cell that reaches a certified steady-state efficiency of 28.04 per cent and keeps ninety per cent of it after 625 hours of light. Lightweight, printable photovoltaics just moved a step toward usefulness.
Silicon solar cells have crept past twenty-six per cent efficiency, but they are heavy, rigid and hard to print on a flexible sheet. The newer thin-film routes, perovskite and organic cells, are light and roll-to-roll printable, but each on its own hits an efficiency wall well below silicon. Stacking the two into a tandem, so that a wide-bandgap perovskite front cell and a narrow-bandgap organic rear cell split the spectrum, is the obvious way around the wall. A paper in Nature, published online on 13 July 2026, by a team led by Li Yongfang and Meng Lei, pushes these tandem cells to a certified steady-state efficiency of 28.04 per cent.
The core problem the team solved is phase separation. Wide-bandgap perovskites blend iodide and bromide to absorb the right part of the spectrum, but under light the mixed halides tend to segregate, the voltage drops and the device decays. The usual fixes work while making the film, not while the cell runs. The team’s answer is an additive molecule, TDB, that changes behaviour depending on the stage of the device’s life. During crystallization it mixes with the perovskite precursors and produces a more uniform iodide-bromide film. Under illumination it activates at the grain boundaries and converts into a second molecule, TAB, which anchors more tightly to the perovskite surface and suppresses the iodine-related defects that drive phase separation.
The result is what the team calls whole-life management of the halide mix. The same additive stabilizes the material as it forms and as it ages, turning a high-bromide perovskite that was light-shy into one that tolerates light. The device reaches a 28.80 per cent conversion efficiency at its peak, certified at 28.04 per cent steady-state, a record for this class of tandem. Stability, usually the letdown, is also strong: after 625 hours of continuous illumination the cell still retains ninety per cent of its initial efficiency.
Why this matters is where these cells can go that silicon cannot. Printed from solution, roll-to-roll, and slot-die coated, they are thin, light and flexible, which suits building-integrated photovoltaics, portable power, wearables and drones. The team also points to space, where high specific power, watts per kilogram, is the figure that matters most, and a lightweight flexible panel is worth a great deal. A record on a laboratory square centimetre does not by itself deliver that, but a record paired with demonstrated stability is the kind of result that lets a technology move from papers to pilot lines.
The honest gaps are the usual ones. The record device is small; area scaling, where perovskite modules lose more than silicon ones, is still unproven. The 625-hour stability test is encouraging but short against the twenty-five-year warranty a roof demands. And the additive, for all its cleverness, is another chemical step in a manufacturing sequence that has to stay cheap. Tandems also have to clear the test of surviving heat, moisture and the thermal cycling of a real roof before they threaten silicon’s installed base.
Still, the direction is unmistakable. A few years ago perovskite-organic tandems sat in the low twenties, and stability was a footnote. A certified 28.04 per cent with 90 per cent retention after 625 hours is both a better number and a more credible one. The tandem is no longer a curiosity on the way to silicon; it is quietly turning into the lightweight, printable option that silicon never will be.