Remote Phosphor Formulation?

[jashhash]

I saw this article here: http://wot.motortrend.com/bmw-shows...amic-lightspot-work-126103.html#axzz2HKdY9JZW
about BMW's new method of creating a laser excited white light source. From what I understand, the process involves shining a blue laser at a phosphor which turns the light into a point source white light emitter. I'm interested in possibly creating a flashlight (thrower) using the same methodology, however that involves formulating a remote phosphor. I'm wondering if anyone on this forum has had experience formulating a remote phosphor.

 

[SemiMan]

I saw this article here: http://wot.motortrend.com/bmw-shows...amic-lightspot-work-126103.html#axzz2HKdY9JZW
about BMW's new method of creating a laser excited white light source. From what I understand, the process involves shining a blue laser at a phosphor which turns the light into a point source white light emitter. I'm interested in possibly creating a flashlight (thrower) using the same methodology, however that involves formulating a remote phosphor. I'm wondering if anyone on this forum has had experience formulating a remote phosphor.

That's not really something you formulate unless you are an expert in chemistry, physical chemistry ..... etc. It is something you buy. Quick web search will bring up several companies who sell phosphors.

 

[bshanahan14rulz]

Jashhash: I thought about doing this a while back. I say, if you have the time, go for it, and please post your results!

Find an LED that has good thermal properties, and use that as your target. Heatsink it well, because you want to keep your phosphor from degrading or burning as well as possible!

Focus laser to a point on the phosphor, or slightly bigger if you want to sacrifice divergence for perhaps a more rounded beam, and use a separate, short focal length aspheric (or even a reflector!) focused on the irradiated phosphor target.

Heck, might be easy enough to just take a crap drop-in and point a laser at it for a proof of concept. If you can get the hotspot to be smaller than it is with just the LED illuminating itself, it would be a good place to start!

basically, why make your own phosphor, when other companies have spent time, money, and research and have already come out with what they deem to be the best formulation. You just want to be able to cool it fast enough, so you'll want an LED with low thermal resistance.

 

[Harold_B]

I'm not clear if the suggestion is to use a LED die with a phosphor layer as a target but if so I suspect that the output would be very disappointing. A good rule-of-thumb is that if a material is a good emitter then it is also a good absorber. The die will absorb approximately half of the light. We use a 30mW 450nm laser for testing phosphors and while it's pretty bright it doesn't compare with a good die. A typical royal blue pump is around 300-400mW each. In the article they mention that in the demo they interrupted the beam of one laser and started a card on fire. Consider they are using three lasers in one light for a bulb replacement (typically around 1100lm). It would be fun to make a bench top laser light but stuffing it into a flashlight and getting a proper phosphor blend would be a trick. If you decide to go for it you might take a look at some of the Intematix products, in particular the premolded PC pucks.

 

[Optical Inferno]

I'm not clear if the suggestion is to use a LED die with a phosphor layer as a target but if so I suspect that the output would be very disappointing. A good rule-of-thumb is that if a material is a good emitter then it is also a good absorber. The die will absorb approximately half of the light. We use a 30mW 450nm laser for testing phosphors and while it's pretty bright it doesn't compare with a good die. A typical royal blue pump is around 300-400mW each. In the article they mention that in the demo they interrupted the beam of one laser and started a card on fire. Consider they are using three lasers in one light for a bulb replacement (typically around 1100lm). It would be fun to make a bench top laser light but stuffing it into a flashlight and getting a proper phosphor blend would be a trick. If you decide to go for it you might take a look at some of the Intematix products, in particular the premolded PC pucks.

+1 on the Intematix parts. Very impressive product line that is readily available from Digikey. I wouldn't invest too much time in experimenting with phosphor formulation as the others have said. Expensive and there are quite a few patent issues out there regarding phosphor formulation and application.

 

[AnAppleSnail]

Note: Please be extra careful with your laser experiments as you progress. Discussions with one of the high-power modders here brought out some safety concerns:

Lasers of interesting power are frightful to the eye. He was unable to come up with a satisfactory physical design to guarantee eye safety. I am not sure how Audi and other similar laser-LED lights prevent blue laser from escaping the light unit, but you should consider doing so.


If you simply put a phosphor-aspheric filter on a blue laser pointer, the failure mode of the phosphor is to pass unchanged, high-collimation laser light at whatever you point at. I considered certain electrical systems to detect this failure mode as it begins, but this is chancy.

Most mirroring designs would also have this failure mode. I assume you'll use adequate precautions with in-lab prototypes, but be sure that your field tests also have similar safety. The last thing I ever want to see is "Hey CPF, check out this article: Mad modder blasts innocent eyes."

 

[bshanahan14rulz]

A measly 400mW pump, sure that might be disappointing. But ~450nm single emitter (although wide transverse mode) diodes are available that are capable of 2000mW+. And isn't the idea for the phosphor to absorb the pump? If the pump gets through the phosphor, who cares if the LED die absorbs the leftovers?

I do agree that one of the biggest concerns is safety, since it is very hard to guarantee that there won't be sufficient reflection off of various internal surfaces that may result in a beam exiting the aperture. Some ways to minimize this is to make the aperture just the right size, use filters and internal beam blocks in places, and it should go without saying, investing in a good pair of goggles with the appropriate protection rating for the powers and wavelengths you are working with.

 

[Marcturus]

Lasers of interesting power are frightful to the eye. He was unable to come up with a satisfactory physical design to guarantee eye safety. I am not sure how Audi and other similar laser-LED lights prevent blue laser from escaping the light unit, but you should consider doing so.

Absolutely, I think these lamps should best be left to impress visitors at trade shows, and to distract competitors. At least it seems as if BMW is planning to deliver them together with a bio-fueled avian safety component that is presently undergoing programming to disable any malfunctioning laser within seconds:
http://www.ovb-online.de/bilder/201...61243-maeusebussard-steckt-scheinwerfer.9.jpg

 

[bshanahan14rulz]

lol @ that last picture!

I don't see what everyone is worried about with the laser based headlights. These devices you cannot even see the beams, not to mention the beams most likely aren't focused to infinity, but rather probably focused to half a meter or so, i.e. on the phosphor target. The light projected on the road is simply an image of the phosphor target, manipulated by a concave mirror, another safety feature.

Why aren't you people scared to use your DVD burners? There is a gap around the drive tray that is big enough for laser light to escape too, does that make the DVD burner dangerous to use?

And give me some money, I'll make a laser pumped phosphor based light that won't let any dangerous amount of the source pump radiation through.

All that said, I'll be the first to admit that, after realizing that this is probably gonna be a passive system instead of an actively scanning one, there is no real reason to use lasers in those headlights other than to say it looks cool and is easily marketable (fricken laser headlights!). Now, as for the OP's use, a laser is the best, most cost effective way to get that much radiation of a specific wavelength into that small of a spot. LEDs don't get anywhere near the optical "flux" that a laser has, and that is why if the phosphor can handle it, it would hopefully lead to a much higher surface luminance which equals more throw.

 

[AnAppleSnail]

lNow, as for the OP's use, a laser is the best, most cost effective way to get that much radiation of a specific wavelength into that small of a spot.

Why bother with lasers? I agree. Using converging lenses and any number of high-output blue LEDs may be an easier design to scale up. Say, five or six blue LEDs in recoil reflectors pointed at a given spot of phosphor. They don't have the surface brightness on their surface, but they do respond well (And mostly safely) to optical manipulation.

 

[RoGuE_StreaK]

investing in a good pair of goggles with the appropriate glovelengths

Fixed.

 

[SemiMan]

Why bother with lasers? I agree. Using converging lenses and any number of high-output blue LEDs may be an easier design to scale up. Say, five or six blue LEDs in recoil reflectors pointed at a given spot of phosphor. They don't have the surface brightness on their surface, but they do respond well (And mostly safely) to optical manipulation.

Optically not the same thing.

 

[bshanahan14rulz]

Here is my thought process. Imagine a multimode laser diode, emission area of ~15umX1um, outputting 1W. Compare it to a typical high power LED, emission area of 1mmX1mm, outputting 0.5W.

In order to get the same irradiance on the target as the laser diode focused down to a "point," you would need at least 100 (~133) of those 0.5W blue LEDs, all focused on the same spot, focused down as small as possible.
([1W/(15umx1um)]=[0.5W/(1mmx1mm)]x, x gives how many superimposed LED radiances to equal the laser diode radiance. From there, assume that optics can only do so much, so all I do is a raw comparison of the radiance and use the results to make inferences on the irradiance at the target)


I think this project has potential to create some serious throw, and if someone doesn't do it by this weekend, I might have to break out some old LEDs and give this a shot myself.

I don't have a regulated or known laser source, since I do not have a power meter, but I should be able to compare it to my best performing P60 (Q2 or Q3 5B XR, don't remember exact bin), a DRY turbo, and a ThruNite Ti on duraloop.
For the laser setup, I've got a direct-drive early generation 1W 445nm emitter, a 2.5" frosty wavy aspheric (If you've seen the aspherics from most cars, you know how crappy this aspheric lens is), and a random XR thermally epoxied to a copper 4U xeon heatsink.

If this is a good 'nuff setup for proof of concept, worst that could happen is I cook some phosphor ;-)

Now I just gotta talk the woman into helping me...

 

[Slotguy]

Simple proof of concept, just took my 405nm laser pointer and shined it on an XPG, returned the beam quite brightly in white. Phospher is already there, de-doming it would be brighter. Don't forget the protective glasses, very bright.

 

[bshanahan14rulz]

^yep, if you take an existing LED and optic combination and compare the lux of the hotspot from A: powering the LED and B: pumping with external source, if there is any way to make the hotspot lux higher, then you have proved the concept.

If I can show how simple it is, perhaps *someone* will pick it up and run with it. And maybe throw me a free sample or a box of ramen or something

 

[2xTrinity]

Here is my thought process. Imagine a multimode laser diode, emission area of ~15umX1um, outputting 1W. Compare it to a typical high power LED, emission area of 1mmX1mm, outputting 0.5W.

In order to get the same irradiance on the target as the laser diode focused down to a "point," <b>you would need at least 100 (~133) of those 0.5W blue LEDs, all focused on the same spot, focused down as small as possible. </b>
([1W/(15umx1um)]=[0.5W/(1mmx1mm)]x, x gives how many superimposed LED radiances to equal the laser diode radiance. From there, assume that optics can only do so much, so all I do is a raw comparison of the radiance and use the results to make inferences on the irradiance at the target)

What you described--imaging multiple LEDs onto the same spot and adding their irradiances to one another--is physically impossible to do with passive optics (eg reflection/refraction). it violates a concept called Etenude Conservation.

To simplify things I will assume that LEDs are uniform lambertian squares (not an unreasonable approximation). Etendue is essentially the product of the source area, and the [projected] solid-angle the light is emitted into.

E = A*Ω*n^2

A = source area
n = refractive index
Ω = π*sin(θ)^2 for a circularly symmetric beam (like a flashlight beam)
θ = divergence half-angle

The problem here is that LED A lambertian LED will emit into a full hemisphere (theta = 90 degrees). When you re-image that LED your lens will have a limited acceptance angle. If you pick magnification of 1 for example, your area A stays the same but your solid angle Ω has decreased. this represents a bottleneck in throughput (light emitted outside acceptance angels of the lenses is lost)

If you were to try to overlap the images of 150 different LEDs, all at 1:1 magnification you'd have to somehow arrange 150lenses in a hemispherical array around your phosphor. This means that each lens by itself could only fill as much as 1/150th of a hemisphere's worth of solid-angle.

Bottom like the surface brightness of your image will always be limited by the surface brightness of your source With 150 LEDs you can illuminate 150x as much area. If you try to demagnify your LEDs or focus that light into a smaller area, you will run into acceptance angel limitations whihc will result in lots of light not being collected.

Simple proof of concept, just took my 405nm laser pointer and shined it on an XPG, returned the beam quite brightly in white. Phospher is already there, de-doming it would be brighter. Don't forget the protective glasses, very bright.

De-doming will not necessarily make your LED brighter. The reason domes are used in the first place is to allow light emitted by the phosphor to escape into the air / not totally internally reflect. The phosphor is embedded in a polymer matrix w/ index of about 1.5.

The net effect of a domed vs de-domed LED is that your die image is magnified by a factor of the index N. If you neglect the effect of light being reflected, re-absorbed, and re-emitted by the phosphor, the actually brightness (radiance) of the die is the same whether it has the dome of nor, but the dome emitted emits far more light overall because it has a larger effective area.

 

[saabluster]

I mentioned in the past that I was developing a laser based lighting system. I had concerns as to safety and how to properly address this issue in a way that has inherent safety. Well sufficed to say I have solved all the issues. I have two different permutations. And no I'm not going into the details. Sorry. You guys are smart enough anyways I am sure you will figure it out. The only thing that stands between me and bringing this to market are regulatory issues. I am not adept enough or have the resources enough to figure out that side of things. I can tell you this. Laser is where throw lights will be in the future. Not LEDs. You can bank on that.

 

[AnAppleSnail]

What you described--imaging multiple LEDs onto the same spot and adding their irradiances to one another--is physically impossible to do with passive optics

I see. In other words, no passive optic will allow the beam to exceed the brightness of the source's surface brightness, because that is what is being projected?

But lasers, of course, would add this way. So the limit is what won't cook the phosphor, and safe design and regulations. No wonder it's the future to Michael.

 

[jashhash]

I would theoretically want to build something much brighter than the 1000 lumen headlamp of the BMW. For me to be motivated to build a prototype I would need a solution to the following problems:
1. Thermal management:
a. Need a phosphor capable of handling the laser's intense heat.
b. The phosphor would need to be applied to something that can move heat quickly (like applied directly to a heat pipe).

2. The above article also says this "laser" method has potential to be more efficient than current high power white LED's. I'm questioning the validity of this statement. Any thoughts on efficiency?

3. A precision point light source would have to be combined with an extremely precise reflector. The reflector doesn't have to be large diameter, but the shape would have to be extremely accurate. The Laser bloom would also have to be precisely centered within the focal point of the reflector.

 

[2xTrinity]

I would theoretically want to build something much brighter than the 1000 lumen headlamp of the BMW. For me to be motivated to build a prototype I would need a solution to the following problems:
1. Thermal management:
a. Need a phosphor capable of handling the laser's intense heat.
b. The phosphor would need to be applied to something that can move heat quickly (like applied directly to a heat pipe).

2. The above article also says this "laser" method has potential to be more efficient than current high power white LED's. I'm questioning the validity of this statement. Any thoughts on efficiency?

I doubt a laser system could beat a white LED for gross lumens/watt at the emitter, but it very possibly could on the system scale. Because a laser pumped phosphor is more like a point-source than an extended LED source it's quite posssible they are able to get the beam distribution they need with less light loss due to baffling/shielding stray light. I'd also expect heat may to be easier to manage with a remote phosphor system. As you pointed out, the phosphor could be coated directly onto a heatpipe without having to propagate the heat through an LED die.

Note the article confusingly point out that the laser source is "1000x brighter on half the power", By brightness they mean radiance (ie the light is emitted from a smaller point-like source), but to a lay reader "brightness" = output in lumens and this would imply the power conversion is 2000x better than LEDs which is nonsense.