Perovskite nanocrystal LEDs have a step in their fabrication that everyone in the field knows is necessary and nobody enjoys. The crystals are born wearing a coat of long-chain organic ligands — oleate and oleylamine — that keeps them stable in solution and electrically insulated in a device. To make a working LED, researchers have generally had to strip that coat off after synthesis and swap in shorter ligands, a post-synthetic ligand exchange that adds a processing step, risks damaging the crystals, and is hard to reproduce the same way twice. A team led by Kookmin University, with Sungkyunkwan University and Hanyang University, reports a synthesis that removes the step at the source: a heat-up method built on an SN2 displacement reaction that produces cesium lead bromide (CsPbBr3) nanocrystals carrying far less oleate, stabilized by bromide instead. The crystals go straight into the LED’s emissive layer with no exchange step, and the devices reach 11.4% external quantum efficiency, up from 2.7% for conventionally made crystals, with peak brightness rising more than fivefold, from 59 to 314 cd/m².

The paper itself was published September 17 in Advanced Functional Materials; Kookmin University announced the result on November 6. The two-month gap is worth stating plainly: this is a process result whose significance is in the shortcut, not in a record number. At 11.4% EQE, the devices sit well below the nanocrystal-LED state of the art — green perovskite LEDs have been reported at 24.5% EQE and above this month alone. What the Kookmin result offers is a simpler route to a working device: four times the efficiency of the conventional-crystal control, achieved by deleting a step rather than adding one.

The problem: the coat that keeps the crystal alive and the device dead

Perovskite nanocrystals are, as the Kookmin announcement puts it, tiny crystals that glow brightly in pure colors and can be handled in liquid form, much like ink — the combination that makes them a leading candidate material for next-generation display LEDs. The standard way of making them, hot-injection synthesis introduced by Kovalenko and coworkers in 2015, uses oleic acid and oleylamine as the ligand pair: oleate binds to the crystal surface and oleylamine with it, and together they control the crystal’s growth and keep it colloidally stable in solution.

The trouble is what those ligands do once the crystals are in a film. Oleate is a long-chain organic molecule, an eighteen-carbon chain with a carboxylate head. In an LED, charge has to hop from crystal to crystal to reach the place where it recombines and emits light, and a shell of long insulating chains is an obstacle course for that hopping. The field’s standard answer has been post-synthetic ligand exchange: wash the crystals, strip the long ligands, and replace them with shorter ones — didodecyldimethylammonium bromide (DDAB) is the most widely used — that bind tightly and conduct better. It works: a 2018 Nanoscale study showed DDAB treatment lifting the photoluminescence quantum yield of CsPbBr3 nanocrystals from 70% to 96% and quadrupling device EQE against the untreated control. But ligand exchange is a second process with its own failure modes. Strip too little and the device stays insulating; strip too much and the crystals lose their passivation, aggregate, or degrade — and the balance is hard to reproduce from batch to batch, exactly the property a manufacturing process cannot tolerate.

The fix: an SN2 heat-up that puts bromide on the surface from the start

The synthesis is a heat-up method built on an SN2 displacement reaction. In an SN2 reaction, one molecule displaces part of another in a single concerted step — a nucleophile attacks a carbon center and the leaving group departs as the new bond forms. The announcement does not name the specific precursors, so this article names none. The heat-up route produces CsPbBr3 nanocrystals carrying far less oleate than the conventional synthesis, with bromide ions taking over much of the job of stabilizing the surface.

That substitution is the chemical heart of the paper. Bromide is a halide anion, small and directly part of the perovskite’s own chemistry, rather than a long organic chain draped over the surface. The team’s abstract states the comparison that makes the substitution work: the oleylamine/bromide (OAM/Br−) ligand system binds more strongly to the CsPbBr3 surface than the conventional oleylamine/oleate (OAM/oleate) system. Through experiments and theoretical calculations, the researchers showed that pairing bromide with oleylamine formed more stable bonds than the conventional pairing — so the crystals keep their strong light emission even with far fewer organic molecules on them, because the bromide that replaces the oleate is itself passivating the surface rather than merely leaving it bare.

The film-level evidence follows the same direction. Films made from the new crystals showed less variation in energy levels and more orderly crystal packing than films of conventionally made crystals, according to the university announcement, and charge moved between crystals far more easily — the abstract reports higher charge carrier mobility in the low-ligand films. The causal chain is coherent: fewer insulating chains on the surface means less energetic disorder, tighter packing, and easier charge transport into the emissive layer.

The device numbers, handled in order

The headline figure is the external quantum efficiency: a maximum of 11.4%, against 2.7% for devices built from conventionally made crystals. EQE is the fraction of electrons sent through the device that emerge as photons escaping it, and a fourfold-plus improvement on the same device architecture, with the crystal synthesis as the only change, isolates the ligand chemistry as the cause — the 2.7% figure is the same stack built from oleate-capped crystals, not a straw man from another group’s older device.

Peak brightness rises more than fivefold, from 59 to 314 cd/m². The ratio is the story — 5.3 times — rather than the absolute number, because 314 cd/m² is dim by display standards. The university announcement says so itself: the peak brightness falls below what many smartphone screens reach, and it did not report how long the devices keep working. A thousand nits is roughly the brightness of a phone or television display running outdoors; 314 nits is a laboratory brightness, not a display brightness. Read the luminance figure as evidence that charge is getting into the emissive layer and recombining there, not as a display claim.

The third reported figure is the emission linewidth: the range of wavelengths in the emitted light narrowed from 21 nanometers to 17 nanometers, a sign of purer color. Narrower emission means a wider color gamut, and a 4-nanometer narrowing on a green emitter is a meaningful improvement in spectral purity. It also corroborates the film story: less energetic disorder in the film means the crystals emit more uniformly, and a narrower linewidth is what that uniformity looks like in the spectrum.

What is not reported matters too. No photoluminescence quantum yield, turn-on voltage, current efficiency, or device area is stated in the material consulted, and no lifetime figure of any kind. The paper’s claim is a synthesis and integration claim, and its device evidence is two figures plus a linewidth. That is enough to support the claim as stated; it is not enough to support anything larger.

The lane: a process win, not an efficiency record

Place 11.4% against the green perovskite LED lane as it stands this month, and the framing writes itself. The Zhejiang University green devices covered on this site on November 7 reach 24.5% EQE with a measured T50 of 1,350 hours at 1,000 nits. A 2024 green device reached 27.1% EQE. Pure-red devices published November 1 reached 31.3% EQE. The Kookmin devices, at 11.4% EQE and 314 cd/m² peak brightness with no lifetime reported, are not competing with those numbers, and the paper does not pretend they are. The comparison that matters is the internal one: 11.4% against 2.7% on the same stack, a 4.2-fold gain bought by removing a processing step.

That is a different kind of result from an efficiency record, and it should be judged on its own terms. An efficiency record is a claim about the best device anyone has built; a process simplification is a claim about the cheapest reliable route to a working device. The ligand-exchange step the Kookmin synthesis deletes is a wet-chemical treatment applied to every batch of crystals before they can be used, with its own solvents, yield losses, and batch-to-batch variation — and it sits directly in the path between a laboratory synthesis and a production line, where every added step is a cost, a failure mode, and a source of variability.

There is a useful contrast with the other direction the field has taken on the same problem. A 2026 Advanced Materials paper from Park and coworkers keeps the conventional oleic-acid synthesis and then performs a hydrolysis-assisted one-step ligand exchange, using methyl acetate to displace the native oleate while introducing a π-conjugated pyridine carboxamide ligand. That is a better ligand exchange. The Kookmin approach is a zero-step ligand exchange: no exchange at all, because the synthesis never puts the long ligand there in the first place. The trade is control: a post-synthetic exchange lets you choose the replacement ligand after the crystals exist, while a synthesis-level fix commits to whatever surface chemistry the synthesis produces.

What this is not: a display device, a lifetime result, or a general recipe

Several of the limits are stated by the university itself rather than inferred by this desk.

First, the brightness. At 314 cd/m² peak, the devices are dimmer than the displays they would serve by a factor of three or more, and the announcement does not say at what voltage or current density that peak is reached. A device four times more efficient than its control but still dim is a device whose injection has improved without yet reaching display-relevant operating conditions.

Second, no lifetime. The university did not report how long the devices keep working, and the material consulted contains no T50, T90, or operational stability figure of any kind. For a display material, lifetime is the test that decides whether an efficiency figure matters, and its absence means the durability of the bromide-stabilized surface is entirely unproven. A surface with fewer organic ligands could plausibly be more vulnerable to moisture and oxygen, not less — the ligands that insulate charge also seal the crystal — and nothing in the material consulted addresses that question.

Third, the scale and scope. The material consulted describes laboratory devices built from one composition, green-emitting CsPbBr3. Whether the SN2 heat-up route works for red iodide or blue chloride nanocrystals is untested here, as is whether it works at the batch sizes a coating line needs or whether the low-oleate crystals stay colloidally stable in ink over the shelf life a production process requires. The university’s own framing is appropriately modest: controlling ligand density and surface stabilization at the synthesis stage simplifies device fabrication and could widen the approach’s use to other devices that convert between light and electricity. “Could widen” is the right verb; the widening has not been shown.

Fourth, the mechanism’s detail sits in the full text. The announcement names the SN2 reaction and the outcome, but the specific precursors, the reaction conditions, and the quantitative ligand-density measurements are not in the material consulted — and the detail that would let another group reproduce the synthesis awaits the full text.

What to watch

Three things, in order of how quickly they could change the assessment.

First, the full text. The paper is published — Advanced Functional Materials, volume 36, issue 75, article e77751 — so the precursor identities, the ligand-density quantification, and the device stack should be readable now. The first figures this desk will check are the photoluminescence quantum yield, which the abstract’s “improved optical properties” implies but does not number in the material consulted, and the control-device comparison in full: turn-on voltage and the current efficiency behind the 11.4% EQE.

Second, a lifetime. Any operational stability measurement — even a short one — would change the assessment more than another efficiency point, because it would test the obvious worry: that a crystal with fewer protective ligands degrades faster. A T50 at any brightness, on an encapsulated device, is the number that decides whether the synthesis shortcut is compatible with a real device or only with a demonstration.

Third, the other colors. A red or blue demonstration of the same SN2 heat-up route would show whether bromide-stabilized, low-oleate synthesis is a general recipe or a CsPbBr3-specific one. Blue is the harder test: chloride perovskites suffer halide segregation and a wider bandgap, and the ligand chemistry that stabilizes a green bromide crystal may not transfer.

The bottom line

A Kookmin University team, with Sungkyunkwan University and Hanyang University, has reported a heat-up synthesis built on an SN2 displacement that produces CsPbBr3 nanocrystals carrying far less oleate, stabilized by bromide instead — and the crystals go straight into an LED’s emissive layer with no post-synthetic ligand exchange. The devices reach 11.4% external quantum efficiency against 2.7% for conventionally made crystals, peak brightness rises more than fivefold from 59 to 314 cd/m², and the emission linewidth narrows from 21 to 17 nanometers. The paper, “Direct Integration of Conductive, Emissive Perovskite Nanocrystals Into Efficient Light-Emitting Diodes Without Post-Synthetic Ligand Exchange,” was published September 17 in Advanced Functional Materials (volume 36, issue 75, article e77751; DOI: 10.1002/adfm.77751), with Lee Yu-bin, Kim Ji-geon, and Seok Gyeong-eun as co-first authors and Kim Young-hoon, Lee Bo-ram, and Lee Tae-kyung as co-corresponding authors; Kookmin announced it November 6.

It is a process result, not an efficiency record: 11.4% EQE is well below the 24.5% green devices reported this month, the 314 cd/m² peak is dim by display standards, and no lifetime has been reported. Frame it as the field’s cleanest answer yet to the ligand problem — delete the exchange step by never putting the long ligand there — and let the full text, a lifetime measurement, and a red or blue demonstration decide whether a simpler synthesis becomes the standard one.

No investment recommendation is made or implied anywhere in this article.

Sources and further reading

  • Kim, Y.-H., Lee, B.-r., Lee, T.-k., et al. “Direct Integration of Conductive, Emissive Perovskite Nanocrystals Into Efficient Light-Emitting Diodes Without Post-Synthetic Ligand Exchange.” Advanced Functional Materials 36, issue 75, article e77751 (published September 17, 2026). DOI: 10.1002/adfm.77751 — the primary source: the paper title, journal, volume, issue, article number, DOI, and publication date, and the abstract’s claims — CsPbBr3 nanocrystals stabilized with Br− instead of anionic oleate ligands, stronger binding of the OAM/Br− system than the OAM/oleate system, suppressed surface defect formation, improved optical properties, higher charge carrier mobility in the low-ligand films, and a maximum EQE of 11.4% against 2.7% for conventional nanocrystal devices, achieved without post-synthetic ligand exchange. The full text was not directly readable through the sources available for this article (the publisher’s page returned an access error), so device figures beyond the EQE comparison and mechanism details beyond the abstract are reported from the university announcement and trade coverage, as stated below.
  • Park Sae-jin, Aju Press (November 6, 2026), “Korean team skips step, brightens perovskite LEDs” — the primary source for the device figures and the announcement framing: the 11.4% versus 2.7% EQE comparison, peak brightness of 314 versus 59 cd/m², the emission linewidth narrowing from 21 to 17 nanometers, the heat-up synthesis built on an SN2 displacement reaction, the far-lower oleate coverage with bromide taking over surface stabilization, the bromide–oleylamine pairing forming more stable bonds than oleylamine–oleate, the films’ reduced energy-level variation and more orderly packing, the direct placement of the crystals into the emissive layer, the co-first authors (Lee Yu-bin, Kim Ji-geon, Seok Gyeong-eun) and co-corresponding authors (Kim Young-hoon, Lee Bo-ram, Lee Tae-kyung) and their institutions, and the stated limits — the 314 cd/m² peak falling below smartphone-screen brightness and no reported device lifetime. Also the source for the journal’s impact factor of 19.9 and top-5-percent field ranking.
  • The Korea Times (via syndication, November 2026), “Joint research team develops new method to produce high-efficiency perovskite LEDs” — corroboration of the announcement: the three universities, the three co-corresponding authors and their departments (Kim Young-hoon, Chemistry, Kookmin; Lee Bo-ram, Advanced Materials Science and Engineering, Sungkyunkwan; Lee Tae-kyung, Organic and Nano Engineering, Hanyang), the co-first authors, and the paper’s publication in Advanced Functional Materials under its full title.
  • Excitonic.com, “Green PeLEDs Finally Clear the 1,000-Hour, 1,000-Nit Bar” (November 7, 2026) — the lane context: the Zhejiang University green devices at 24.5% EQE with a measured T50 of 1,350 hours at 1,000 nits, the pure-red 31.3% EQE result of November 1, and the 27.1% green EQE figure from 2024, all as previously reported on this site.
  • Park, et al. “Ligand-Mediated Surface Carrier Modulation in Perovskite Nanocrystals for Charge-Symmetric LEDs.” Advanced Materials (2026). DOI: 10.1002/adma.202520499 — the contrasting approach to the same ligand problem: a hydrolysis-assisted one-step ligand exchange using methyl acetate to displace native oleate and introduce a π-conjugated pyridine carboxamide ligand, cited as the better-exchange alternative to the Kookmin zero-exchange route.
  • Sun, S., et al. “Surface ligand modification of cesium lead bromide nanocrystals for improved light-emitting performance.” Nanoscale 10, 4173–4178 (2018). DOI: 10.1039/c7nr09126e — the classic ligand-exchange comparator: DDAB treatment lifting CsPbBr3 nanocrystal photoluminescence quantum yield from 70% to 96% and quadrupling device EQE against the untreated control, with a maximum luminance of 11,990 cd/m² and 2.1% EQE; also the source for the hot-injection synthesis background (Protesescu et al., Kovalenko group) and the role of oleylamine in nanocrystal formation.

A note on sourcing: the full body text of the Kim paper was not directly readable through the sources available for this article; the publisher’s page returned an access error. The abstract — via the Korea University Pure record — and the Kookmin University announcement via Aju Press are the sources for every figure and mechanism claim quoted here. No photoluminescence quantum yield, turn-on voltage, current efficiency, device area, encapsulation method, precursor identity, or lifetime figure is stated, because none appears in the material consulted. No market-size figure is quoted, because none is needed for this device result.