In 2014, Tae-Woo Lee's team had just produced an early perovskite light-emitting diode. Its efficiency was only about 2%, yet Lee first had to persuade the university to file an international patent for a technology that had not yet taken shape. Twelve years later, his forum talk was no longer about whether a PeLED could emit light. It was about how an ever-smaller display pixel could preserve color, brightness and lifetime at the same time.

Immersive displays demand purer colors and higher brightness while pushing pixels below two micrometers. Perovskites absorb light strongly and emit over a narrow spectral range, making a thin layer well suited to color conversion. During a green-light demonstration, Lee reminded the audience that a camera could not capture the color purity visible to the eye. For a moment, the green shown on the screen and the green seen by the audience were two different things.

Researchers later formed the emitting region from nanocrystals, nanoscale polycrystalline films and quasi-two-dimensional structures, helping electrons and holes meet more readily and driving PeLED efficiency upward. After the records rose, the material still had three more passages to negotiate: strong light, heat and humidity; reproducible scale-up of synthesis; and the vacuum processes familiar to display manufacturing.

Durability begins with the soft perovskite lattice. The team built a hierarchical shell of PbSO4, SiO2 and polymer around the nanocrystals, isolating them from the environment while locking a lattice prone to movement. Lee said at the forum that a color-conversion film using this class of material had accumulated more than 14,000 hours of testing on a blue-light panel. The team's Science paper, meanwhile, reported T90 values of 12,000 and 27,234 hours for nanocrystals of different compositions—the time required for emission to fall to 90% of its initial level. The panel test and the nanocrystal tests in the paper used neither the same sample nor the same conditions. Their common lesson is how the shell fixes the lattice, not which hour count is larger.

A material that survives must also be made reliably at scale. Conventional hot injection requires high temperature, an inert environment and rapid cooling. As the vessel grows, each becomes an equipment and batch-control problem. The team instead cooled the precursor solution below 4°C and used cold injection to control nanocrystal growth. A 20-liter batch reported in Nature retained a photoluminescence quantum yield close to 100%, placing large-batch synthesis and high emission quality in the same experiment. Display production will still ask about lot-to-lot consistency and yield, but the work no longer stops at a small vial.

A second route meets the display industry's existing manufacturing methods directly. An organic spacer designed by the team replaces some halide sites and bonds with lead ions, steering crystallization during vacuum deposition toward what the paper calls an X-type quasi-two-dimensional structure. Devices reported in Nature Nanotechnology reached an external quantum efficiency of 21.9% and operated for more than 1,500 minutes. Those minutes are not a product lifetime. They do, however, move a material previously discussed mostly in solution into the language of vacuum deposition and pixel arrays.

Since the 2014 patent that first required persuasion, PeLEDs have ceased to be a question of whether they can light up. Lee separated the research into materials under blue light, a 20-liter reactor and a vacuum-deposited pixel array. The next step will not be answered by another efficiency record; it will be whether those three routes can connect within one manufacturing process.

Sources and further reading

Talk and speaker information

Primary papers