a better incandescent light bulb

Neat. Sounds like the process would work well for a lot of things. Solar Collectors, stealth airplanes.

Dan

You can increase the intensity and shift the color of any incandescent bulb into the blue region while reducing relative power just by using a thinner filament. That process reduces bulb lifetime however. The following cites indicate that there is a downside to using a laser to thin a filament.

"Small variations in resistivity along the filament cause "hot spots" to form at points of higher resistivity; a variation of diameter of only 1% will cause a 25% reduction in service life. The hot spots evaporate faster than the rest of the filament, increasing resistance at that point?a positive feedback which ends in the familiar tiny gap in an otherwise healthy-looking filament."

"In flood lamps used for photographic lighting, the tradeoff is made in the other direction. Compared to general-service bulbs, for the same power, these bulbs produce far more light, and (more importantly) light at a higher color temperature, at the expense of greatly reduced life (which may be as short as

2 hours for a type P1 lamp)."

--Winston

I understood the article to say not that the filament is thinned, but that its surface is modified to increase emissivity.

I've got a prototype on my desk of some rebuildable lamps I designed and built a couple years ago for a customer who was doing similar research. They disassemble easily so new filaments can be mounted, and have cute plumbing and electrical feedthroughs for thorough purging. The most surprising part of the project, which I think I've mentioned here before, was obtaining a small quantity of getter. One of the ingredients in the getter for incandescent lamps is red phosphorous, which is apparently strictly controlled these days because it's used in cooking meth.

(...)

Chunlei Guo:

"We fired the laser beam right through the glass of the bulb and altered a small area on the filament. When we lit the bulb, we could actually see this one patch was clearly brighter than the rest of the filament."

Sounds to me as if the modification was just redistribution of filament metal. What would happen when you increase the resistance of a portion of a filament by thinning it? It dissipates more power because it's resistance is higher (P=I^2R) It glows more brightly than the rest of the filament but will cause the bulb to fail much earlier than it would have without the modification, I think. Would Occam be pleased with this guess? :)

That'll be the thing I learned today. :)

--Winston

I got the impression that the object was to "roughen" the surface. This would make the tungsten a better emitter. As they said, if you have good control of the size and spacing of the asperities, you will be able to selectively change the emissivity for different colours. If they merely wanted to selectively thin the filament, it would be orders of magnitude cheaper to do it when drawing the wire. Conventional bulb can (and are, for some types) be made more efficient by using a dichroic reflector so that much of the IR gets bounced back to the filament rather than heating the rest of the world.

Mark Rand RTFM

(...)

"Though Guo cannot yet make a simple bulb shine pure blue, for instance, he can change the overall radiated spectrum so that the tungsten, which normally radiates a yellowish light, could radiate a more purely white light."

You can do the same thing without a laser. Just push more power into the bulb and the color temperature will shift upward. Bulb life suffers, though.

OK, but how do I justify this nifty femtosecond pulsed laser then?

Gotta stay focused on the goal here.

Creation of that roughened surface means redistributing filament tungsten, yes? Subtract some here, add some there?

I'll bet you a dollar that we don't hear anything from Chunlei Guo or University of Rochester regarding bulb life tests. :)

--Winston

If the filament has a thin section the power density at that spot will indeed be higher, but the overall consumption of the lamp will be lower. (P=V^2/R) The article says, "... we could actually see this one patch was clearly brighter than the rest of the filament, but there was no change in the bulb's energy usage."

(...)

Yes, but bulb life is significantly reduced.

Can we really call it a "better light bulb" if we have to replace it ~twice as often? (I don't think so.)

"Small variations in resistivity along the filament cause "hot spots" to form at points of higher resistivity; a variation of diameter of only 1% will cause a 25% reduction in service life. The hot spots evaporate faster than the rest of the filament, increasing resistance at that point?a positive feedback which ends in the familiar tiny gap in an otherwise healthy-looking filament."

--Winston

As I said above, if the experiment created a hot spot by thinning the filament the resistance of the filament, and the power consumed, would change. The experimenters reported there was *no change* in the power consumption.

The relationship between power density, temperature, efficiency and life is well understood -- you can find it in old texts that date back to the early days of electric lighting. It's hard to believe that these guys, whose previous work involved fiddling with the emissivity of metals, are unaware of that relationship and have misinterpreted the results of their experiment.

In *net* consumption, no change was detected. I get that. I hypothesize that the local power consumption at the thinned area of the filament did increase, because it's increased resistance dropped more voltage across the thinned area for a given amount of current. The increased resistance of the filament as a whole would have decreased the filament current a tiny amount, cancelling the effect of the local power consumption increase.

Model it as two PTC resistors in series, one of which is about a percent of the value of the second one. Triple the value of the smaller resistance and it's power consumption triples. The net resistance of the network as a whole is decreased by a couple percent, causing the second resistor to decrease in value. As a result, current increases once again and a new equilbrium is established which is very close, if not identical to the pre-modification current. The bulb lifetime is significantly decreased but the study did not encompass bulb lifetime, just hotspots and spectral shift.

They didn't misinterpret, as far as I can see. They didn't say a word about the post-modification reliability of the bulb. You could argue that they committed an error of omission about that, if you were feeling very charitable. :)

--Winston

I respectfully disagree. The article was about changing the power distribution ratio and spectral shift in various parts of the filament. Our researchers cleverly implied that changing the thickness of a metal oxide layer (and the resulting emissivity changes) had something to do with the changes in the light bulb but the two effects are quite different and have nothing to do with each other, IMHO.

The filament is in a vacuum. Heating it does not cause it's surface to mix more readily with oxygen because there is no oxygen to speak of, in the vicinity of the filament.

Sure! Anyone who has held a torch to steel has seen how the surface changes the color of reflected light.

Theoretically, I agree. This isn't what our researchers were doing, however.

--Winston

=A0Our researchers cleverly implied that

I did not note any reference to " metal oxide layer "

Hmmm. I thought the whole article was about how the researchers were modifying the surface of the filament.

Dan

The article cleverly melded two *different* physical changes together, making it appear as if they were the same.

The article mentions the use of the laser to change the color of metal ala:

That is a heat process. The source doesn't need to be a femtosecond laser. You can use any sufficient source of high quality heat (and oxygen) to do that. We've all seen it as HAZ discoloration:

Yup. That was the clever part. You and the U.S. Air Force Office of Scientific Research were prompted to assume that the two different physical changes were directly related.

"In 2008, his team used a similar process to change the color of nearly any metal to blue, golden, and gray, in addition to the black he'd already accomplished."

[Please refer again to the HAZ photo link. What colors do you see? I see gray, blue, and golden starting from the most heated area to the least heated area of the HAZ.]

Here is a smoking gun: "Guo and Vorobeyv used that knowledge of how to control the size and shape of the nanostructures?and thus what colors of light those structures absorb and radiate?to change the amount of each wavelength of light the tungsten filament radiates."

What was this 'similar process'? I'll bet you a dollar that it was localized heating of a metal sample to grow a color-reflective oxide layer. How is that related to the process of vaporizing a tiny chunk out of a tungsten filament?

My point is that there is no relationship between these two concepts other than one can use heat in specific forms to accomplish them.

--Winston

Your arguments appear to be approaching blind denial of the claimed effect.

Have a look at a butterfly wing under an electron microscope and you'll get a better idea of what's going on.

Mark Rand RTFM

I prefer to think of it as 'informed denial of the claimed effect'. :)

What claim do I deny?

The claim, IIRC is that one can make a light bulb universally 'better' by exposing it's filament to laser light, because the resulting roughened surface somehow causes it to convert electrical power into visible - frequency photon emission more efficiently than does the smoother un-modified filament, without any change in the cross sectional area of the filament anywhere along it's length.

I don't deny that you can shift the average color temperature of the bulb upwards towards blue without increased bulb power by thinning the filament, but I do deny that the effect is produced by anything other than merely thinning the filament.

I also deny that laser beam exposure 'improves' the bulb because it will significantly decrease the amount of time that the lamp remains functional as compared to it's lifetime had it not been exposed.

A bulb design that fails significantly more often than average is not a better design. It is a worse design even if it is more efficient during it's short stay in the socket.

Iridescence?

Can you help me understand your point, please?

Are you saying that the iridescent surface of soap bubbles, butterfly wings and the inside of abalone shells make them more intense visible light sources than river rocks, blades of grass or concrete slabs for example?

I deny that, too. :)

--Winston

That's a pretty remarkable claim -- that, at least in the range where a normal lamp operates, the current flowing thru a filament is completely independent of the voltage applied to it. It's also very easy to test, so I did, and it just ain't so. Even using a 3-1/2 digit meter it was easy to observe the current thru a 75W bulb change with The bulb lifetime is significantly decreased but the study did not encompass

Then I'm not sure what you're claiming. That the researchers have convinced their peers at a major research university, the Air Force Office of Scientific Research, and the referees at Physical Review Letters, but not you, that blasting material off a filament with a laser is a significant achievement?

(...)

Your model shows a net reduction of voltage available to the bulb. In the experimenter's case, the amount of voltage to the bulb remained the same but it was distributed differently in the filament eg. a 'hot spot' area dissipating significantly more power than it had before it was thinned.

I said that the positive temperature coefficient of the filament would tend to limit the change of the current through the bulb as the power distribution in the bulb was changed by thinning some portion of the filament.

Your numbers show that the variation in current through your bulb was about half the variation in voltage across the bulb. A one volt change across the bulb caused a 2 milliamp change in current. This is the nonlinear positive temperature coefficient variation I was on about.

Had that PTC effect not been in place, we would expect to see a 5 milliamp change in current for a 1 volt change. A new current equilibrium was established that was within 0.35% of the pre-modification current, under a voltage change of 0.82%.

(...)

I'm not claiming much of anything.

Their claim, IIRC is that one can make a light bulb universally 'better' by exposing it's filament to laser light, because the resulting roughened surface somehow causes it to convert electrical power into visible - frequency photon emission more efficiently than does the smoother un-modified filament, without changing the cross sectional area of the filament anywhere along it's length.

I don't deny that you can shift the average color temperature of the bulb upwards towards blue without increased bulb power by thinning the filament, but I do deny that the effect is produced by anything other than merely thinning the filament.

I also deny that laser beam exposure 'improves' the bulb because it will significantly decrease the amount of time that the lamp remains functional as compared to it's lifetime had it not been exposed.

A bulb design that fails significantly more often than average is not a better design. It is a worse design even if it is more efficient during it's short stay in the socket.

Blasting material off a filament is a parlor trick, not a significant achievement.

I believe Mr. Guo and Mr. Vorobeyv are mistaken at best.

--Winston

Not my model. It's the one you proposed in your previous post ("Model it as two PTC resistors in series, one of which is about a percent of the value of the second one"), except I moved one resistance outside the lamp's envelope and turned a knob rather than tweaking it with a laser.

And in my experiment, the voltage across the two resistors was also constant.

What you said (in order to explain the experimenters' observation that no change in power accompanied the increase in light output) was that the PTC of the untreated section of the filament would cause the current to restabilize at a point "very close, if not identical to the pre-modification current." For that to happen, the current passing thru the untreated filament must be independent of the applied voltage, and that's clearly not the case.

Exactly. If treating a spot on the filament changed its resistance (as I changed the resistance in series with my filament), the resistance of the adjacent untreated filament would change in the opposite direction, but the magnitude of change would be smaller. In other words, there'd be a net change in the lamp's resistance, which would show up as a change in power.