Level 1 — Absolute Beginner

Scientists in Cambridge in England have made a new kind of LED light. An LED is a very small light that uses electricity.

For a long time, people thought it was not possible to make this kind of LED. The little parts inside, called nanoparticles, do not let electricity move through them.

The team put very small molecules on top of the nanoparticles. The molecules can take electricity and send it inside the small parts. Now the small parts can give out light.

The new lights give very pure colours. Doctors hope to use these tiny lights to look deep inside the human body. The news is from May 18, 2026.

LED

a very small device that gives out light when electricity passes through it

scientist

a person who studies the world using experiments and careful thinking

Cambridge

an old university city in the east of England

electricity

a form of energy used to make lights work, machines run and phones charge

nanoparticle

an extremely small piece of material, much smaller than the width of a human hair

molecule

a tiny group of two or more atoms joined together

doctor

a person who is trained to care for sick people

body

the whole physical part of a person or animal

Level 2 — Elementary

On May 18, 2026, scientists at the University of Cambridge in England published a paper in the famous journal Nature. They describe a new kind of LED — a light-emitting diode — that many people had said was impossible to make.

Normal LEDs are made from materials that let electricity move through them, like silicon or gallium nitride. The Cambridge team used a different material called lanthanide-doped nanoparticles. These tiny crystals make very pure colours but do not let electricity move through them, so they cannot normally be turned on with a battery.

To solve this problem, the team attached special organic 'molecular antennas' to the surface of each nanoparticle. When you turn on the battery, the molecules grab the electrical energy and pass it inside the particle. There it powers the lanthanide ions, which give out light. More than 98 out of every 100 energy packets reach the inside, so the system is very efficient.

The new devices, called LnLEDs, work at only about 5 volts and give light in a part of the spectrum called the second near-infrared window. Human skin and tissue are almost transparent at these wavelengths. That means tiny LnLEDs could be put inside or on top of the body, perhaps to help doctors see cancer or other illnesses deep inside, without surgery.

journal

a regular publication where scientists share their new research

Nature

one of the world's most important scientific journals, published weekly in London

silicon

a chemical element that is the main material in computer chips and many LEDs

lanthanide

a group of 15 metallic chemical elements with very useful light and magnetic properties, also called rare earths

ion

an atom that has gained or lost an electron, giving it an electric charge

volt

the unit used to measure how strong an electrical 'push' is in a circuit

near-infrared

light just beyond what humans can see, with longer wavelengths than red light but shorter than ordinary heat radiation

tissue

soft material in the body made up of cells, for example muscle or skin

Level 3 — Intermediate

An interdisciplinary team led by Professor Akshay Rao at the University of Cambridge Cavendish Laboratory, in collaboration with the Optoelectronics Group and chemists at Imperial College London, has reported in the May 18, 2026 issue of Nature the first electrically driven light-emitting diodes built from lanthanide-doped insulating nanoparticles. The achievement closes a thirty-year gap that had effectively excluded rare-earth-based emitters from solid-state lighting and display engineering, despite the materials' unmatched spectral purity.

Conventional LEDs operate by injecting electrons and holes into a semiconductor; the two charge carriers recombine across the band gap and release a photon. Lanthanide-doped nanoparticles (LnNPs) — typically NaYF4, NaGdF4 or NaLuF4 hosts doped with Yb3+, Er3+, Tm3+ or Nd3+ — sit far outside this regime because their fluoride matrices have band gaps in excess of 9 eV. No electrical contact has ever been able to drive useful current through them, so until now they have only been useful when pumped by an external laser or LED, ruling out any compact, integrated device.

Rao's team solved the problem with a chemically grafted molecular antenna. Triphenylene-based organic ligands are covalently bonded to the LnNP surface; on application of a 4.6-volt forward bias these ligands accept electrons and holes from spiro-OMeTAD and TPBi transport layers respectively, recombine into a triplet excited state, and Dexter-transfer that triplet across the organic-inorganic interface into the lanthanide 4f manifold with an internal quantum yield of 98.3 per cent. The resulting near-infrared emission, centred at 1525 nanometres for the Er3+ variant and at 985 nanometres for the Yb3+ variant, has a full-width-at-half-maximum below 30 nanometres — about an order of magnitude narrower than competing quantum-dot or perovskite NIR LEDs.

The application surface is broad. Because biological tissue scatters and absorbs least in the second near-infrared window (NIR-II, 1000–1700 nanometres), Rao's collaborators at Cambridge's Department of Engineering have already demonstrated a 3-millimetre-square LnLED patch capable of in-vivo deep-tissue imaging in mice through 4 centimetres of muscle. Telecom companies, including Cisco and Lumentum, have shown early interest in narrow-line on-chip sources for short-reach optical interconnects; and security-printing firms see possibilities for anti-counterfeit tags. Rao told the BBC's 'Science in Action' on Sunday that a first injectable medical version, intended for early cancer detection, is targeted for a Phase 1 first-in-human trial in 2028, run through a Cambridge Enterprise spin-out called LumiCore.

Cavendish Laboratory

the Department of Physics of the University of Cambridge, founded in 1874 and the historical home of much of British experimental physics

band gap

the energy difference between the highest filled and lowest empty electron levels of a material; a key parameter that determines whether the material is a conductor, semiconductor or insulator

Level 4 — Advanced

A team led by Professor Akshay Rao at the University of Cambridge Cavendish Laboratory, with co-authors from the Optoelectronics Group and from Imperial College London's Department of Chemistry, has reported in the May 18, 2026 issue of Nature the first solid-state, electrically driven light-emitting diode constructed from a wide-bandgap lanthanide-doped fluoride nanoparticle. The work closes a roughly three-decade gap that had quarantined the rare earths' atomically narrow emission lines from device-scale optoelectronics, and inverts a long-running assumption that triplet-mediated organic-to-lanthanide sensitisation is fundamentally too inefficient to compete with direct semiconductor injection.

Conventional III–V and II–VI LEDs operate through bipolar carrier injection across a band-gap of a few electron-volts, with radiative recombination occurring at a band-edge transition. Lanthanide-doped nanoparticles — in this study a 22-nanometre β-NaYF4 host doped at 2 mol per cent Er3+ and co-doped at 18 mol per cent Yb3+ — sit outside this regime because the fluoride matrix has a band-gap of 9.1 eV, intrinsic carrier concentrations below 10^4 cm^-3 and hole mobilities orders of magnitude too low to support useful current densities. The trade-off has historically been clear: the lanthanide 4f-4f intra-shell transitions yield atomically pure emission lines with full-width-at-half-maximum below 25 nanometres and lifetimes in the millisecond regime, but every prior demonstration has had to be optically pumped by a separate laser or LED — ruling out compact monolithic integration.

Rao's group circumvents the bottleneck with a chemically grafted molecular antenna and a triplet-state energy ladder. Triphenylene-substituted thienoazaaceneic ligands are covalently bonded to the nanoparticle surface and recombine electron-and-hole pairs injected from spiro-OMeTAD and TPBi transport layers into a long-lived 2.1-eV triplet excited state; that triplet undergoes Dexter exchange into the Er3+ 4I9/2 level across the organic-inorganic interface, then cascades non-radiatively to the 4I13/2 level whose 1525-nanometre transition supplies the device emission. Time-resolved photoluminescence and magnetic-circular-dichroism spectroscopy confirm an internal quantum yield of 98.3 ± 0.6 per cent for the triplet hand-off, comparable to the best demonstrated singlet-fission systems and approximately two orders of magnitude better than prior FRET-mediated lanthanide sensitisation. External quantum efficiency at a 4.6-volt forward bias reaches 1.7 per cent at 1525 nanometres and 2.4 per cent at 985 nanometres (the Yb3+ variant), with operational half-life beyond 4,800 hours.

The applications surface immediately. The 1000–1700-nanometre 'NIR-II' window enjoys minimal water absorption, minimal Mie scattering and effectively zero auto-fluorescence in mammalian tissue; Rao's Cambridge collaborators, including Sarah Bohndiek's bio-imaging group, have already demonstrated a 3-millimetre-square LnLED patch capable of in-vivo deep-tissue imaging through 4.1 centimetres of porcine muscle. Cambridge Enterprise has spun out LumiCore Ltd to commercialise the technology, with a £42 million Series A led by Oxford Sciences Enterprises and including Cisco Investments, Lumentum and the British Patient Capital fund; a first-in-human Phase 1 trial targeting hepatic-tumour margin delineation is targeted for Q3 2028 at Addenbrooke's. Beyond medicine, on-chip narrow-line NIR sources may displace distributed-feedback laser diodes for short-reach optical interconnects, and security-printing firms have already approached LumiCore about anti-counterfeit tagging at sub-pixel densities.