A different semiconductor platform
Perovskite LEDs are an active materials frontier. Peng and colleagues’ 2025 Nature paper uses large-grained, weakly confined caesium lead bromide rather than strongly confined structures. The work addresses competing losses, migration of ions and thermal behavior. These are device-material questions; the source is not a test of a complete white general-lighting luminaire. It is also a lead-containing material, so a future product would need its own containment and end-of-use assessment. [1]
A striking result, with an explicit extrapolation
The 2025 authors report peak external quantum efficiency of 22.0% and a half-lifetime extrapolated to 185,600 hours at an initial luminance of 100 cd/m² at room temperature. Extrapolated is essential here: this is not a device observed running for that entire period. Nor does a low-luminance lifetime establish performance at a different brightness or temperature. The study supports a materials strategy and a conditional lifetime estimate, not an equivalent service-life promise for an installed fixture. [1]
The earlier paper makes the contrast visible
Ding and colleagues’ 2024 Nature Photonics abstract reports T50 of 18.67 hours at 12,000 cd/m² and a low-luminance equivalent above 50,317 hours at 100 cd/m². It reports EQE of 29.5%. Those results illustrate why the initial luminance and estimation method must accompany the headline lifetime. The full subscription article was not reviewed here; this account is limited to the publisher’s accessible abstract. [2]
What EQE leaves unanswered
Editorial measurement explanation: external quantum efficiency counts emitted photons relative to injected electrons. Lumens weight visible output for human vision; electrical lm/W also depends on energy input and the emitted spectrum. EQE therefore cannot be read as luminous efficacy or as a percentage of useful room illumination. Likewise, luminance in cd/m² and illuminance in lux describe different quantities. Preserve the measurement boundary when comparing an experimental pixel and a luminaire.
A product has more failure paths than an emitter
Our development checklist includes package sealing, stable white spectral mixing, thermal behavior, drivers, optical extraction, production yield and serviceability. A room light needs enough area and total output for its task. Device research can remove one bottleneck without resolving all the others. Compare measured output decay, color shift and failure rates under intended use, rather than awarding a finished-product ranking from a single laboratory metric.
Scenario: specialized emitters before a broad lighting transition
Our hypothesis is that material-specific advantages may create useful niches before a general-lighting replacement. Stable, manufacturable multi-color or converted-white products would make the case stronger. A confident timetable would need sustained independent device results and production evidence. The 2024 and 2025 papers show why this field deserves a future chapter, while also showing why long extrapolated lifetimes should be read carefully.