Circularly polarized electroluminescence from chiral supramolecular semiconductor thin films

Here is the trade-off that has haunted chiral optics for years. Circularly polarized light — photons that spin either left-handed or right-handed — is genuinely useful for displays, sensing, and spintronic devices. But every practical attempt to generate it electrically from an organic light-emitting diode has forced a choice. Push for strong handedness and your device efficiency collapses. Build an efficient organic light-emitting diode and the chiral signal nearly vanishes. Chowdhury and colleagues at Cambridge, along with a large international team, just broke that trade-off. The answer turned out to be letting molecules organize themselves. The metric at the center of this story is the dissymmetry factor, g — specifically gEL for electroluminescence. It measures how strongly emitted light favors one handedness, running from zero for completely unpolarized light up to plus or minus two for perfectly circular polarization. Conventional host-guest organic light-emitting diodes disperse chiral emitter molecules in a molecular host to harvest both singlet and triplet excited states efficiently. But in those systems, the emitter is essentially isolated; it barely influences the electronic structure of the surrounding film. The result: dissymmetry factors below ten to the negative three. Essentially no chiral signal at all. The other extreme exists too. Cholesteric liquid crystals and similar supramolecular assemblies can reach dissymmetry values near plus or minus two — close to the theoretical maximum — because their chiral order is a large-scale photonic effect. But those systems haven't delivered the thin-film device performance needed for practical organic light-emitting diodes. Lanthanide emitters can be highly chiral but suffer from low electroluminescence quantum yield. The field was stuck: structure that creates chirality tends to wreck efficiency, and vice versa. Chowdhury and colleagues chose a different starting point entirely. Their molecule is a triazatruxene derivative called S-TAT, and what makes it special is that it doesn't stay isolated. When it crystallizes, it builds its own helix. Single-crystal X-ray analysis shows the molecules packing in a noncentrosymmetric space group, P6122, with six molecules completing one full turn of the helix. The helical pitch is 2.3 nanometres, with molecules spaced 0.38 nanometres apart inside dimers. Within each repeating unit, molecules rotate 40 degrees inside a dimer and 120 degrees between dimers — a precise, repeating twist that propagates through the entire crystal. That geometry does something remarkable to the electronic structure. First-principles calculations show that when isolated TAT molecules stack into this helix, their doubly degenerate frontier orbitals — the highest occupied and lowest unoccupied molecular orbitals — spread into dispersive electronic bands. Dimerization folds those bands and opens gaps. More importantly, the chiral stacking lifts the degeneracy between states carrying different orbital angular momentum: bands with magnetic quantum numbers of plus one and minus one separate in energy by up to about 30 millielectronvolts. Angular momentum is now baked into the band structure itself. The lowest-energy electronic transition carries right-handed circular polarization; the next transition up carries left-handed. The helix has turned a chemical preference into an optical one. The spectroscopic consequences are large. Crystallization shifts the optical gap from 3.2 electron volts for an isolated molecule down to 2.67 electron volts in the stack — a redshift of about half an electron volt — and experimentally the absorption edge moves from 380 nanometres into the green, around 510 nanometres. Photoluminescence quantum efficiency jumps from 14 percent in amorphous films to 58 percent in the crystal. The pure S-TAT crystal shows circularly polarized photoluminescence with a dissymmetry of negative 0.24. That's 24 percent — already far beyond what conventional host-guest organic light-emitting diodes achieve. But a crystal sitting in a vial isn't a device. The real challenge was translating that chiral order into a smooth, uniform thin film compatible with vacuum deposition — the same industrial process used to manufacture commercial organic light-emitting diodes. Getting chiral supramolecular structures to form reliably across device-scale areas, without cracking or delaminating, had not been demonstrated. The solution is elegant in its simplicity. Chowdhury and colleagues cosublime S-TAT as a low-concentration guest — just 10 weight percent — together with a structurally mismatched molecular host. They tested two hosts: CBP, a hole-transporting material, and DCzDCN, which is ambipolar, meaning it transports both holes and electrons. The word "mismatched" is doing real work here. The long alkyl chains on S-TAT reduce its interactions with the host, and S-TAT has a lower melting temperature — around 353 Kelvin — than either host. So when the film is first deposited, the mismatch suppresses premature crystallization. The film comes down amorphous and physically smooth, with the chiral molecules distributed evenly but not yet ordered. The host matrix holds everything together. Then comes the trigger. Gentle heating above S-TAT's melting temperature — or, in operating devices, the local Joule heating that can exceed 100 degrees Celsius — allows the S-TAT phase to melt rapidly and reorganize. It crystallizes in situ into helical, pi-stacked domains inside the slow-melting host scaffold. Polarized optical microscopy reveals spherulitic domains forming after annealing. Small-angle X-ray scattering confirms phase-segregated structures, and atomic force microscopy shows the spherulites are built from three-dimensional crystalline fibrils running parallel to the substrate. Confocal microscopy maps the result optically: green-emitting regions above 500 nanometres, associated with crystalline S-TAT domains, show photoluminescence lifetimes around 23 nanoseconds. Surrounding amorphous regions emit at shorter wavelengths with lifetimes below 5 nanoseconds. The chiral order is spatially localized, but it's there, and it survives device fabrication. The devices built from these films are where the two halves of the story come together. The best circularly polarized organic light-emitting diodes, using S-TAT in the DCzDCN host, reached an external quantum efficiency — the fraction of injected electrons that exit as useful photons — of approximately 16 percent. That's competitive with conventional, non-chiral organic light-emitting diodes. Peak current efficiency was 45 candela per ampere, turn-on voltage was just 2.2 volts, and maximum luminance hit 57,000 candela per square metre. These are not the numbers of a lab curiosity. They're device-level performance. And the chirality held. Electroluminescence dissymmetry factors reached the order of ten to the negative one — at least 10 percent — under electrical operation. That's the harder measurement: photoluminescence dissymmetry is observed under optical excitation, but electroluminescence dissymmetry requires the chiral order to survive the injection of charge carriers, electric fields, and heat. It did. Conventional host-guest organic light-emitting diodes sit below ten to the negative three. This work sits at ten to the negative one — two orders of magnitude better. Host choice mattered significantly. The DCzDCN host outperformed CBP on efficiency and roll-off, likely because its ambipolar character balances charge injection. By contrast, phenTAT dopants — where stronger host-guest interactions prevented nanophase segregation — showed external quantum efficiency around just 3 percent and no circularly polarized emission at all. Racemic n-TAT stacks gave similar efficiency but near-zero dissymmetry, exactly as expected from a material with no net handedness. The controls confirm the story: the chiral emission comes from the helical supramolecular order, not from any artifact of the device structure. The team also notes a thermally activated delayed fluorescence-like contribution to the high efficiency. Annealing brings the singlet excited state close to the triplet — the literature triplet energy for TAT is around 2.3 electron volts — and delayed emission accounts for up to about 20 percent of total photoluminescence. That triplet harvesting is part of why the numbers are so good. Device lifetime, even in unencapsulated devices measured in air, showed more than 100 hours of operation at 100 candela per square metre before brightness dropped to half. Not a finished product, but a viable starting point. The deeper implication is what the fabrication route enables. Vacuum cosublimation is already how the display industry deposits organic light-emitting diode films. The fact that chiral supramolecular order can be induced in situ, using standard equipment, with thermally triggered crystallization, means this isn't a laboratory technique requiring exotic processing. Other molecules could be designed with similar structural mismatches and low melting temperatures, self-assembling into different chiral geometries inside host scaffolds. Because the angular momentum in these systems is encoded in the band structure — not through spin-orbit coupling — the effects should extend beyond light emission to chiral-driven charge transport and spin-selective phenomena. Chowdhury and colleagues name spintronics and chiral-induced spin selectivity as directions this opens. The organic light-emitting diode is the proof of concept. The principle is a new fabrication strategy for chiral quantum materials, made with tools that already exist on the factory floor. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.