265 nm UV-C LEDs are getting efficient enough to matter — inside the wall-plug-efficiency race

We recently covered why 265 nm is the germicidal sweet spot for UV-C — it sits on DNA's absorption peak — but with a catch: the deepest UV-C wavelengths have always been the hardest and least efficient to manufacture, which is why much of the commercial UV-C LED market sits at the longer, easier 275–280 nm end. That gap is now closing. Stanley Electric reports a wall-plug efficiency (WPE) of 7.5% for its 265 nm UV-C LEDs — about three times its previous level — and a roadmap to 10% by the end of fiscal 2026.

What WPE is, and why it's the number that matters

Wall-plug efficiency is the share of electrical power going in that comes out as UV-C light. It is the honest, system-relevant figure — unlike external quantum efficiency (EQE), which can look impressive (~20% has been reported) while WPE stays in single digits, because electrical losses and getting the light out of the chip eat the difference. For replacing mercury lamps, WPE is what decides running cost and heat.

The numbers, in context

Stanley's figure (7.5% WPE, measured at 158 mW / 350 mA / 6 V) comes from a proprietary structure that pulls more light out of the aluminium-nitride substrate, plus a high-reflectivity electrode. For scale: deep-UV WPE has typically sat around 5%, so a 3× jump to 7.5% is real movement. And Stanley is not alone — ams OSRAM reported a 10.2% WPE at 265 nm in 2025, validated by Germany's national metrology institute (PTB), up from around 5.3%; academic work has pushed 275 nm to 9.1% with micro-mesh electrodes. The frontier is moving across the whole field, not at one vendor.

Why it matters

The case for UV-C LEDs over mercury was never germicidal power — it was mercury-free operation, instant on/off, compactness and lifetime. The weak point was efficiency at the biologically best wavelength. As 265 nm WPE climbs from ~5% toward 10%, the awkward trade-off we flagged earlier — "use 265 nm for the biology, accept worse efficiency" — softens. Water disinfection, where throughput scales directly with delivered UV power, is the first application to feel it.

The caveats

Three, to keep this honest. A sample-stage WPE figure (Stanley's samples ship from March 2026) is not a deployed-system efficiency — packaging, optics and driver losses still apply downstream. WPE is one axis: lifetime (L70 — the hours until output falls to 70% of initial), cost per milliwatt, and thermal management decide real deployments as much as efficiency does. And even at 7.5–10%, 265 nm still trails the longer UV-C wavelengths on raw efficiency — the gap is narrowing, not gone. But "narrowing" is the story: the wavelength that disinfects best is becoming one you can also afford to run.