"Researchers at MIT have shown that by surrounding the filament with a special crystal structure in the glass they can bounce back the energy which is usually lost in heat, while still allowing the light through.
They refer to the technique as ?recycling light? because the energy which would usually escape into the air is redirected back to the filament where it can create new light.
'It recycles the energy that would otherwise be wasted,' said Professor Marin Soljacic
The report I heard on the radio classed their current prototype as 3x as efficient as incandescents, which would only make them 1/3 as efficient as LEDs on the shelf today (Are the 200-300 lm/W LEDs any closer to the shelves yet?).
When I heard this on R4 a bit ago, I couldn't help but try to figure out whether the usual miserable suspects here will hate this because it's more efficient, hate it because it's new, or love it because it's incandescent.
or suspect that only half the story is here and that merely reflecting the infra red back in is hardly going to do the job on its own.
If you could coat the inside with a substance that stored energy and sent itt out in one direction as a higher frequency of light, ie visible then that would be a breakthrough, but getting quarts out of pint pots has not been successful to date, and therefore how is this going to actually work?
Brian
In theory if the bulb surface is a reflector to infra-red but transparent to the visible, the thing will simply not cool down so the filament will take a lot less power to stay hot.
Whether such a surface material can be made cheaply and effectively I do not know, but if it can, the claims are true.
Preventing thermal losses at wavelengths we can't see is half the battle
- less power is needed to keep the filament hot.
An ancient technology from the low pressure sodium lamp uses a clear InO coating on the inside of the outer vacuum envelope to prevent the escape of an otherwise strong IR spectral line and so improve thermodynamic efficiency. Same would potentially work in a filament lamp at least until tungsten vapour plates out on the envelope.
Making the filament a selective visible light emitter by making it a mirror finish and not a black body for near IR radiation would improve efficiency but I cannot see how they propose to coat it with something that will survive the thermal cycling stresses of a lamp filament!
It might make the filament more efficient than a basic tungsten one (at least until the coating falls off) but I doubt if it can rival the LED technology which is now closing in on the ultimate 200lumen/W target of the most efficient lamps available which are still low pressure sodium.
This is only a proof of concept prototype, they hope to get a lot better.
It could get even higher - this, from 2002, says "This could raise the efficiency of an incandescent electric bulb from 5 percent to greater than 60 percent"
(And the initial construction of an optical silicon lattice was 1998:
Or like me think (a) likely *realistic* time to market X, (b) my life expectancy Y, (c) heh ho.
A multi layer dichric reflector would do a similar job, they don't do it because it costs too much.
El Reg's take on it:
Just like i said earlier then and I hadn't seen the article.
Now can the tungsten migrate through the oxide layers or does it last forever??
'Could' reach 40% means very little. One of the problems is that IR would h it all the internal structures, glass, leads, ballotini fuse, the lot. So h ow much IR one can reflect back has its limits, and only a limited percenta ge would hit the filament. So in short, 40% efficiency is just drooling. As usual, it reads more as 'we want your money please.'
NT
300 layers to do it..
Not a cheap bulb. I guess interference type coatings could be used as reflect or transmit type filters
The other Sandia labs photonic lattice material is more interesting although how long it would maintain its structure at working temperature is still open to question. And real filaments rearrange their atoms when hot so the thing will rapidly degrade.
It is quite cute in a technology sense but I suspect that persuading the interference coating that it doesn't want to fall off when the filament warms up and expands more than the ceramic will be a problem.
The layers could be on the filament so the IR doesn't hit any of the other structures.
Basically you need a substrate you can deposite tungsten onto and then coat with 300 layers of oxide with different thicknesses. Then the IR would never leave the substrate. With enough layers you could get to the point where the only mechanism for the energy to leave would be as light but that isn't going to happen.
Next week: The CRT re-imagined with similarly reworked technology. (Just don't ask me to carry a 65" diagonal CRT - like the flat one I did carry the other day.)
The IR blocking layers would then be white hot & give off lots of IR.
Any substrate would need to be filament shaped to avoid massive heat loss & thus efficiency loss. Given filament behaviour, a W coated rod would simpl y deteriorate much worse & more rapidly than plain W. Anyway, what other ma terial is there to use? Very little survives white heat.
NT
Ceramic?
Next week is LED nanotechnology:-)
Of course it is all a load of bollocks.
ss & thus efficiency loss. Given filament behaviour, a W coated rod would s imply deteriorate much worse & more rapidly than plain W. Anyway, what othe r material is there to use? Very little survives white heat.
The element must be thin - ceramics are too fragile for that. A short fat e lement on ceramic is possible, but then you've got a very low voltage lamp that needs a transformer. Add to that the ceramic's positive tempco of cond uctance and you then need a ballast rather than a fixed voltage. But it wou ld light.
But... a short fat element suffers a lot of end losses. It also has enough thermal capacity to make it take a little time to light. And the faster det erioration issue is still there. And really one is no further forward.
Filament lamps have been worked on & experimented with for almost 2 centuri es, ever since the platinum filament lamp. Finding a practical way to impro ve them is thus extremely difficult.
NT
Have something to add? Share your thoughts — no account required.
Ask the community — no account required