Solar cells get wired together the way plenty of other electronics do. A machine runs a thin metal ribbon across the front of one cell and the back of the next, heat goes into the joint, the solder melts, and that is your connection. Every conventional panel you have ever seen on a roof was built like that.
The trouble is that the two cell types the industry is betting on next cannot take the heat. So a group at Forschungszentrum Jülich in Germany went looking for a way to make the same joint cold, and what they came back with is a roll of tape.
So what is wrong with heat?
Silicon heterojunction cells hold their performance thanks to hydrogen sitting in an extremely thin layer of amorphous silicon on the wafer surface. Warm them much past 392 degrees Fahrenheit (200 degrees Celsius) and the hydrogen starts leaving, taking the passivation with it. Perovskite is fussier still. Jülich keeps its perovskite processing under 230 degrees Fahrenheit (110 degrees Celsius), because the crystal comes apart above that.
Standard infrared soldering runs hotter than either ceiling. That is why the industry has spent the past decade on low-temperature bismuth solder, conductive adhesives, coated copper wires and anything else that joins metal without cooking the cell underneath.
The tape is something you could already buy
Conductive tape is not a new invention. It turns up in electromagnetic shielding and inside consumer electronics, where the job is to stick two metal surfaces together and let current cross the bond. Yanxin Liu and his colleagues took three commercial tapes off the shelf and ran them through the abuse a solar module has to survive.
Two of the three are anisotropic, and that word is doing a lot of work here. An isotropic tape conducts in every direction at once, which is fine for shielding and bad on a solar cell, where two conductors sitting a millimeter apart are not supposed to talk to each other. An anisotropic tape conducts only through its thickness, top to bottom. Sideways it behaves like an insulator. Metal particles inside the adhesive touch the cell on one face and the ribbon on the other, and the current crosses through the particles rather than through the glue.
The thinnest tape has the least glue to hold on with
The thickest of the three is 100 micrometers of woven fabric coated on both sides, and that is the isotropic one. The middle option is a 50-micrometer anisotropic film filled with silver particles. The third is a 10-micrometer anisotropic film filled with nickel, and 10 micrometers works out to four ten-thousandths of an inch, thinner than any hair on your head.
Those nickel particles are bigger than the layer carrying them. Liu’s paper puts them at 20 to 50 micrometers across in a 10-micrometer adhesive, so the glue cannot bury them and they stand proud of it on both faces. Electrically that is the entire point. Mechanically it leaves almost no adhesive doing the gripping, and the peel tests said so plainly: the two thicker tapes held on firmly, and the nickel one did not.
So which one survived the chamber?
Jülich built single-cell modules with each tape and put them through damp heat, the certification soak that parks a module at 185 degrees Fahrenheit (85 degrees Celsius) and 85 percent humidity for weeks on end. The isotropic tape went first, giving up more than 5 percent of its efficiency in roughly 500 hours as its contact resistance climbed.
The two anisotropic tapes stayed in the fight. Across the full run the nickel tape lost around 2.5 percent of its efficiency with a 1 percent drop in fill factor, and the silver tape around 4 percent with a 2 percent drop.
Then came thermal cycling, 200 round trips down to minus 40 and back. The silver tape’s contact resistance climbed past the limit the team had set for itself, and its module shed about 2 percent in the first 100 cycles before settling down. The nickel tape’s resistance crept up too and stayed comfortably under the line, and that module finished less than 1 percent down, with its open-circuit voltage and current unchanged.
That line is not an arbitrary one. Jülich simulated what a bad joint costs a finished module and drew it at 0.015 ohms per square millimeter, the resistance at which the contact eats a tenth of a percentage point of efficiency. All three tapes started underneath it. One of them was still there at the end of both tests.
So what does Jülich do with that?
The tape that behaves best electrically is the one you would least want holding a ribbon in place for 25 years, and the group put that in its own summary: the next job is combining ECT-3’s stability with ECT-2’s peel force. I like that they wrote it down instead of burying it in a supplementary file. Plenty of interconnection papers would have.
So you can have the contact or the grip, and apparently not both yet.
Why not take the silver tape and live with the resistance? Because that resistance stays in the joint for the life of the panel, turning current into heat, and a module that starts a couple of percent down and then drifts further is the kind of thing that shows up as a warranty claim in year twelve rather than a bad reading on day one.
Silver is why anybody is doing this
Liu’s group frames the whole exercise around supply, which is the boring half of solar that we tend to skip past on the way to efficiency records. In a multi-terawatt solar industry, its slides put photovoltaics at half the world’s silver supply and a quarter of its bismuth. Bismuth is on that list because low-temperature solder is the current workaround for heat-sensitive cells, so one shortage gets solved by walking straight into another.
The market has already started moving without them. Solar manufacturing consumed 186.6 million ounces of silver in 2025, down from 197.5 million the year before, according to the World Silver Survey that the Silver Institute and Metals Focus published on April 15, 2026, and that drop happened while cell production kept rising. Manufacturers are putting less silver in each cell, because silver averaged just over $40 an ounce last year, up 42 percent. Recovery is getting serious too, which is why Georgia hosts a plant that strips old panels back into silver, copper and glass. A tape filled with nickel is that same instinct taken a step further.
Nobody has certified a panel built this way
Jülich laminated single cells in a laboratory. Nobody has certified a full panel built this way, and the group is careful to say its tapes perform comparably to the solder-free methods already in use rather than beating them. What the tape does drop is the curing step, because there is no wet adhesive to bake afterward.
The tape itself goes on cold. The module still gets laminated after that, and lamination is a hot press, so none of this turns panel assembly into a room-temperature process from end to end. A tape whose particles are bigger than its own glue also strikes me as an interesting conversation to have with a production engineer running a stringer at full speed.
Certification is the other wall. Any perovskite-silicon module needs 1,000 hours of damp heat and 200 cold cycles before a bank will finance a field of them, which is exactly the exam a Qcells tandem panel sat in July, and this tape has not faced that at module scale. The American end of tandem manufacturing has the same gap, which is why the California line printing perovskite onto cover glass is still waiting on somebody outside the building to certify its numbers.
Solar RRL published the study on September 1, 2026, with Yanxin Liu as first author and Kaining Ding’s group at Jülich behind it, and Liu had already walked through the reliability data at the Metallization and Interconnection Workshop. Peel force on that nickel tape is the number that decides whether any of it leaves the lab.
