Torygraph: MIT develops incandescent lamp that is more efficient than LED

[dennis]

I could be wrong but I doubt if much research into vacuum deposition of oxides onto filament lamps has been going on for anywhere near two centuries.

[tabbypurr]

No, but the basic physics aren't too mysterious.

NT

[The Natural Philosopher]

reflection is not absorption

[Alan Braggins]

The layers in the MIT demo are very close to the filament. Or at least the "thin sheet of polished tungsten, laser machined into a closely-packed, radiator-like structure that maximises the planar surface while allowing for resistive heating".

See also images on

TNP was suggesting you could do the same thing in principle putting the reflective layers on the bulb surface. Like existing dichroic coatings, but more so:

That has the problem of hitting the other internal structures.

[tabbypurr]

that's immaterial, with the 300 layers on the filament itself they would be white hot from conduction.

NT

[Andrew Gabriel]

It's been done in halogen lamps for 10-15 years. I think GE were the first with the 300/500W linear halogens (which were some thing like 225/400W with higher light output), but others followed suit.

I can't see anything in that newspaper article which is any different from the high efficiency halogens technique of 15 years ago.

That's been done by GE too. A pattern is applied to the filamant which generates a canceling interference pattern at IR wavelengths. IIRC, the problem was that the pattern didn't last, and they canceled the research when filament lamps were banned, rather than banning lamps of low efficiency.