An HKUST team has stretched room-temperature spin coherence in a perovskite to about five nanoseconds, roughly 300 times longer than a conventional single well. The devices built on it are proofs of concept, not products.
The trick is a staircase. Instead of making one chiral quantum well and trying to inject spin from it into an ordinary semiconductor, where most of the polarisation is lost at the interface, the Hong Kong team grew a stack of wells of gradually varying width, all of them chiral, so that an electron passing from one to the next falls a small step in energy each time and keeps its orientation. They call it a spin cascade, and they built it with a two-step seed-assisted crystallisation that produces oriented wells across two-dimensional, quasi-two-dimensional and three-dimensional phases of the same material.
The work comes from Haipeng Lu’s group in the Department of Chemistry at the Hong Kong University of Science and Technology, with Jian Qiu and Yuling Huang as joint first authors. It was published in Nature Photonics on 15 October, and HKUST issued its own announcement on 9 November.
Why spin lifetime is the whole problem
Electronics encodes information in charge. Spintronics would encode it in the magnetic orientation of the electron, which in principle means switching with far less energy. The obstacle has never been generating spin polarisation. It is holding on to it. In most semiconductors the orientation scrambles within picoseconds at room temperature, which is why the striking demonstrations in this field have usually come from cryostats.
Halide perovskites are interesting here for an awkward reason: the heavy atoms that make them absorb and emit light so well also create strong spin-orbit coupling, which is what destroys spin information. The team’s architecture works against two separate loss channels. The graded, continuously chiral environment reduces the excess energy that drives Rashba splitting at interfaces, and the structural coupling between wells suppresses dephasing in the bulk phases. The reported result is spin coherence lasting about 5 nanoseconds at room temperature. HKUST describes that as roughly 300 times longer than conventional chiral single quantum well systems.
What was actually built
Three measured outputs follow from the long-lived polarisation. Circularly polarised luminescence with a luminescence dissymmetry factor of 1.4 x 10 to the power minus 2, spin-dependent charge transport with a polarisation degree above 80%, and two demonstrations: X-ray-triggered dissymmetric photopolymerisation, which points at making chiral molecules with light, and circularly polarised light photodetectors.
Each of those deserves proportion. Five nanoseconds is a long time by the standards of room-temperature perovskite spin physics and a short time by the standards of a working memory or logic device, which needs transport over device distances, electrical injection rather than optical pumping, and a way to read the spin state out. A dissymmetry factor of 1.4 x 10 to the power minus 2 means the circular polarisation of the emitted light is a small percentage effect, which is respectable for this class of material and far from what a display or communications component would require. The 300-fold figure comes from the university announcement, comparing against conventional single wells rather than against every published system. The photodetector and photopolymerisation results are proof-of-concept devices in a research paper with no reported operating lifetime, and lead-halide perovskites carry the field’s standing problems of moisture and thermal instability and lead content, none of which are addressed by this architecture.
The commercial read
There is no company here, no patent disclosed in the materials reviewed and no industrial partner named. Funding is academic: Lu’s group records a Research Grants Council collaborative grant for control of chirality and spin in halide perovskites awarded in January 2026, and a National Natural Science Foundation young scientists award in September 2026. Lu was promoted to associate professor in July 2026. This is a laboratory result on a long path.
The near-term application is also the least glamorous one, and that is often a good sign. Circularly polarised light detection is a real instrument market today, used in pharmaceutical chirality analysis, polarimetry and some imaging, and it is currently served by bulky optical assemblies of filters and waveplates. A semiconductor that detects handedness directly, at room temperature, could compress that into a chip, and the barrier to entry for an instrument component is far lower than for a logic device. The photopolymerisation result hints at a second route, using polarised light to bias the handedness of a chemical product, which matters because the wrong enantiomer of a drug is at best useless.
Beneficiaries, if the materials mature, would be analytical instrument makers, specialist photodetector suppliers and perovskite materials firms already selling into solar and X-ray imaging, since the processing route is compatible with solution deposition. The incumbents with most to lose are makers of the discrete optical components that circular polarisation analysis currently requires. These are reasoned scenarios, not announced products. For an investor the honest summary is that a room-temperature spin lifetime of nanoseconds removes one of the field’s standing objections, and that removing an objection is not the same as having a device.