Is a one-chip full monochrome color LED possible?

[EricB]

Hi All.
I have been fascinated by LED's as thge holy grail of lighting; an especially full color units, which are really becoming more common.
After reading on the doping process that determines the color output of c die chip; I was wondering if it were possible to have one chip that, perhaps with different current or something, could emit different monochrome colors,so you could get that rich 590nm amber, for instance, instead of the weaker yellows made by mixing red and green; and you could also get the 565nm yellow green, 600-20nm oranges, pure greens, cyan, etc.
Does anyone know if any comany is reasearging into suh a possibility?
Imagine the displays you could make with those!

[EricB]

My first post.^ No one seemed to know about it, apparently; but over the passing four years I would keep searching, and today I found these:

http://www.optotech.com/data_sheets/OTLA-0130_13.pdf
http://www.wipo.int/pctdb/en/wo.jsp?...1&DISPLAY=DESC
http://www.faqs.org/patents/app/20080207078

I wonder, just to be sure, if these were what I was talking about? It seems taking a chip of a particular wavelength, and adding a fluorescent material to get comparatively longer wavelengths, if I read correctly. and these patents are two years old already, and appear to be intended for lasers.

So has anyone seen anything like this? And is it possible to just have a semiconductor compond that can emit diferent wavelengths when different currents are applied, or something like that?

[Ken_McE]

I can't comment on the patents because I don't understand them. That first item is a white LED with a ring of smaller(?) colored LEDs around it. By lighting them selectively you get whatever tint you want. This is a variation on what I know as "RGB Mixing". RGB refers to Red, Green, Blue, the three primary colors that can be mixed to provide almost any color. RGB mixing has a lot of potential for high efficiency because each LED can be driven at its own optimum level. "White" LEDs are actually blue LEDs with a yellow phosphor on top. This gives you a whitish light but drops your efficiency. RGB mixing also allows you to shift easily from color to color, you are not locked into just one white. The classic problem with RGB mixing is that you get colored fringes around the edges of objects because the lights come from different angles.

So, RGB mixing gives results somewhat like what you are looking for, and it is here and easily doable right now.

You are also looking for a single light source that can change color. Incandescent filaments will change from orangey-brown to pure white if you under and fully drive them. If a steel bar is gradually heated it will glow in first red, then orange, then yellow to white as the heat goes up. I think you also get blue-white if you go up really hot. This gives you a controllable color changing light, but tends to run a little hot.

I am however not familiar at this time with any one light or emitter that you can easily color change by changing your electrical input. Sorry.

PS Mono = same or alike, chrome = color, thus monochrome means all the same color. You may have meant polychrome?

[2xTrinity]

Originally Posted by EricB

Hi All.
I have been fascinated by LED's as thge holy grail of lighting; an especially full color units, which are really becoming more common.
After reading on the doping process that determines the color output of c die chip; I was wondering if it were possible to have one chip that, perhaps with different current or something, could emit different monochrome colors,so you could get that rich 590nm amber, for instance, instead of the weaker yellows made by mixing red and green; and you could also get the 565nm yellow green, 600-20nm oranges, pure greens, cyan, etc.
Does anyone know if any comany is reasearging into suh a possibility?
Imagine the displays you could make with those!

I see no advantage in generating monochrome wavelengths for a display compared to using a mixture of primary colors. The mixture of primary colors will appear the same to human eyes, and it will be much easier to engineer such a system. There are certainly non-display applications, such as sensor applications, where having a particular wavelength may be important, however.

A better approach IMO would be to simply use more than three primary colors for special cases. For example, a red-amber-green-cyan-blue scheme selected at the right wavelength would be able to cover a wider gamut (or possible color combinations) than a simple RGB at much less expense than one with "infinite" primary colors.

Also, I am not sure what you are looking for is physically possible. The reason doping is adjusted is to change the energy band gap in the semiconduction junction. This is an intrinsic property of the matrial, not something that can be modulated using an electronics signal. In fact, there is even a hard limit to how much the wavelength/bandgap can be manipulated through doping for a particular material. For example, InGaN is the matrial commonly used to make LEDs in the near-violet, to green range. It happens to be most efficient at making light at 450nm (this is what white LEDs are based on). By adjusting doping concentrations, InGaN diodes can be made to produce green or near-uv, but at MUCH lower efficiency in both cases. There is no way you'd be able to dope a InGan to be able to make red light for example. And this is for adjusting the doping at the time of manufacturing, not in real time.

PS Mono = same or alike, chrome = color, thus monochrome means all the same color. You may have meant polychrome?

He meant monochrome as in, able to make arbitrary colors as a (variable) monochrome wavelength, rather than as a mixture of primary colors.

RGB mixing has a lot of potential for high efficiency because each LED can be driven at its own optimum level. "White" LEDs are actually blue LEDs with a yellow phosphor on top. This gives you a whitish light but drops your efficiency.

I think it's important that you distinguish between conversion efficiency (ie, making electricity into light) and luminous efficiency. A non-coated blue LED for example will appear dimmer than a white LED because our eyes are less sensitive to blue, than they are to the red/yellow/green wavelengtsh from the phosphor.

As it stands now, RGB mixing is way behind phosphor white LEDs for lumens/watt. The blue InGaN diodes used to drive the white LEDs are by far more efficient at converting electricity to light than any other LED, so even after the phosphor losses, it's still the most efficient option. The best blue LEDs convert about 60% of elcetricity to light. The best 530nm green LEDs are more like 5-10%. Yellow-green LEDs (at 550nm or so) are even less efficient.

The best RGB arrays that I know of are about 50 lumens/watt, compared to in excess of 120 lumens/watt for white LEDs under ideal conditions. Even neutral and warm white LEDs now are approaching 100lm/W @ 350mA.

You are also looking for a single light source that can change color. Incandescent filaments will change from orangey-brown to pure white if you under and fully drive them.

If true arbitrary monochrome color is necessary, the way to do it is use a continuous specrtum source like an incan, then use a prism to split that into component colors. Then use shutters to select a particualr color. For laboratory applications this is how it's done, though obviously with horrible efficiency.

If only color variation along the blackbody line (eg, orange -> white -> blue) is needed, the approach I'd use at this time is to mix cool white LED (6500k, ie Cree WC), warm white (2800k, ie Cree 8A) , cyan, and red. The cyan is used to fill the "gap" in the LED spectrum, and counter-balance the red so that the light doesn't look pinkish. With the right weighting of those colors, continuous variatino from ~2000-6500k should be possible. But you won't be able to make anything that isn't some shade of 'white'.

[saabluster]

I don't think anything will fill exactly what you said but this might come close.

[EricB]

Originally Posted by 2xTrinity

I see no advantage in generating monochrome wavelengths for a display compared to using a mixture of primary colors. The mixture of primary colors will appear the same to human eyes, and it will be much easier to engineer such a system.

That's actually not true, as an amber or orange made by mixing red and green (even if a single LED with rgb dies) looks nothing like those brilliant monochromatic 590 LED's, as well as the slightly longer orange and orange-red ones (600-619) which are common on signage as the most visible color. Monochromatic is much more saturated. And I don't think you can get the 565-570 lime green either in the same saturation. (The green in the windows logo comes close, but I've never seen that done with LED's, and you can still see the red in it).

There are certainly non-display applications, such as sensor applications, where having a particular wavelength may be important, however.

A better approach IMO would be to simply use more than three primary colors for special cases. For example, a red-amber-green-cyan-blue scheme selected at the right wavelength would be able to cover a wider gamut (or possible color combinations) than a simple RGB at much less expense than one with "infinite" primary colors.

Yeah, but that would make the unit too big; or you would just be using the separate LED's of those colors.

Also, I am not sure what you are looking for is physically possible. The reason doping is adjusted is to change the energy band gap in the semiconduction junction. This is an intrinsic property of the matrial, not something that can be modulated using an electronics signal. In fact, there is even a hard limit to how much the wavelength/bandgap can be manipulated through doping for a particular material. For example, InGaN is the matrial commonly used to make LEDs in the near-violet, to green range. It happens to be most efficient at making light at 450nm (this is what white LEDs are based on). By adjusting doping concentrations, InGaN diodes can be made to produce green or near-uv, but at MUCH lower efficiency in both cases. There is no way you'd be able to dope a InGan to be able to make red light for example. And this is for adjusting the doping at the time of manufacturing, not in real time.

OK, I realize that it would be hard to cross the InGan line, as the blues were nearly impossible without them (I didn't know if InGaN could make reds, and I know that old blue made with those compounds was too dim). So we could start with an InGaN chip that could cover the 400-540 range, and then 550 and up would be covered by the compound that makes those colors. Then, we could have a two-die LED that would cover the whole visible range. (then, we could add a third blue chip with phosphor for white; then you would have all the monochromatic colors, plus pure white, which you also can't really get with rgb mixing. Yet you could mix the other colors to get the rgb white also).

So it's the adjusting of doping concentations that makes the different colors? (I never knew what exactly differentiated the colors made by the same material).
So wouldn't there be a way to simulate different concentrations, then? Again, like applying a different current would simmulate a lighter concentration? (I'm not sure how that works, but to give an example). Is there a pattern, like the heavier the concentration, the higher the frequency, or something like that?

If true arbitrary monochrome color is necessary, the way to do it is use a continuous specrtum source like an incan, then use a prism to split that into component colors. Then use shutters to select a particualr color. For laboratory applications this is how it's done, though obviously with horrible efficiency.

Now, we don't want to be going back to no incandescent, do we? The whole point is to use an LED.

Originally Posted by saabluster

I don't think anything will fill exactly what you said but this might come close.

That looks like it might be. I'll have to take a closer look at it, to try to figure out what sme of those things mean. Like does it produce the whole "spaced apart" band at once, making a white LED with more wavelengths? (It mentions something about white LED's in the middle). Or can it turn on particular wavelengths at a time? That's what I was looking for. Or is this even talking about a single die/chip? I havn't seen it say that explicitly, yet. Even if it's the former, that would be a start, and all they would need to do is find a way to turn on one at a time.

Trinity and Ken, since you know alot about this stuff, what is your take of this item, given you said multiband was impossible?

[2xTrinity]

Originally Posted by EricB

That's actually not true, as an amber or orange made by mixing red and green (even if a single LED with rgb dies) looks nothing like those brilliant monochromatic 590 LED's, as well as the slightly longer orange and orange-red ones (600-619) which are common on signage as the most visible color.

You're right. However my proposed 5-primary scheme dramatically improves the situation. Amber will look very good (eg, that single LED only), and lime-green will be possible by mixing amber and green (with much better results than mixing red and green).
Likewise, mixing cyan LEDs and green LEDs will allow a lot more possible blue-green shades, than mixing a deep blue with a green LED.

Yeah, but that would make the unit too big; or you would just be using the separate LED's of those colors.

This wouldn't necessarily make the unit too big. For example, a red-amber-green-blue system could fit in a 2x2 grid. As LED dice are almost always inherently square, they are usually packed using "square packing" like this. Many RGB LEDs (in signs etc.) use one red, one blue, and two green dice aranged in a 2x2 grid already. Replacing one of the greens with a yellow/amber would improve the range of colors without changing the size of the system at all.

I personally would love to see Cree come out with a version of the MC-E loaded with separately-addressable red, amber, green and blue LEDs. It might take a fairly complicated set of diffusers and optics to mix the colors evenly, but it could sure make for some fun-to-use toys



The gray section is the range all possible colors, with saturated monochromatic colors along the rim. The highlighted "triangle" is the gamut, or possible range of colors that can be made with the displa (in this case a CRT, with red, green, and blue phoshpors)

The phosphors have fairly wide spectral line width (ie, they aren't very saturated), so that is why the primary colors are not along the edge. Filtered fluorescent (eg, your LCD monitor) and LED also can have fairly wide line width, leading to a limited gamut.

As you pointed out, it's not possible to make truly saturated colors in some ranges. Another scheme that has been proposed is to use laser light as the light source. Laser light is extremely monochromatic (ie, totally saturated). Look at that graph and imagine three points located at 650nm, 532nm, and 405nm RIGHT along the edge of the CIE curve. The triangle drawn between those three points covers a lot more area than the case of the CRT, meaning many more saturated colors would be possible with a laser light source.

Even an LED that could change its center-wavelength to any aribtrary value might have a hard time competing with a laser-based system, as even though the central wavelength might be moved, the LED will still have a significant spectral line width (read: even single color LEDs aren't perfectly saturated).


[quote]OK, I realize that it would be hard to cross the InGan line, as the blues were nearly impossible without them (I didn't know if InGaN could make reds, and I know that old blue made with those compounds was too dim). So we could start with an InGaN chip that could cover the 400-540 range, and then 550 and up would be covered by the compound that makes those colors. Then, we could have a two-die LED that would cover the whole visible range.

(then, we could add a third blue chip with phosphor for white; then you would have all the monochromatic colors, plus pure white, which you also can't really get with rgb mixing.

Interesting idea. mixing RGB to get white for a display system is analogous to the reason why black ink is used for CMYK systems -- as it's almost impossible to make true black by mixing cyan magenta and yellow.

For something like a display on a portable device, using a white LED would likely be more power efficient than making white with a mixture of RGB dice.

So it's the adjusting of doping concentations that makes the different colors? (I never knew what exactly differentiated the colors made by the same material).
So wouldn't there be a way to simulate different concentrations, then?

The only way I can think of would be to have a

Again, like applying a different current would simmulate a lighter concentration? (I'm not sure how that works, but to give an example). Is there a pattern, like the heavier the concentration, the higher the frequency, or something like that?

in general increasing the dopant concentrations will tend to result in a larger voltage drop (eg, produce shorter wavelengths) However, this change is proportional to the logarithm of the concentration, so it takes a very large change in doping to produce a small change in wavelength for an LED device.

If you make the concentration too high, then the highly concentrated impurities will start to significantly disrupt the structure of the crystal, and also result in very high unwanted resistance. This will set an upper limit on wavelength possible from a material.

If you make the concentrations too low (eg, to try to produce longer wavelengths), the device will simply become an insulator. What dopants do is add free electrons and holes that are capable of conducting current. With no doping/free carriers, there can be no current.

I can't think of any way to "simulate" different dopings in a material. I believe you are imagining a device behaving somewhat like a transistor -- where by applying a certain voltage or current you can produce light at a different color. Unfortunately, there is no way to modulate voltage-drop or effective of a semiconductor device that I know of. Amplifying a current flow is completely different than changing the fundamental properties of a material "on the fly".

That looks like it might be. I'll have to take a closer look at it, to try to figure out what sme of those things mean. Like does it produce the whole "spaced apart" band at once, making a white LED with more wavelengths? (It mentions something about white LED's in the middle). Or can it turn on particular wavelengths at a time? That's what I was looking for. Or is this even talking about a single die/chip? I havn't seen it say that explicitly, yet. Even if it's the former, that would be a start, and all they would need to do is find a way to turn on one at a time.

Trinity and Ken, since you know alot about this stuff, what is your take of this item, given you said multiband was impossible?

The actual device they are showing is somewhat interesting. They are in effect creating an LED that emits light at a whole bunch of different spectral lines by sandwiching many materials, with different dopings etc. each together. However, in the body of the paper, it is clear that this device is meant for generating some pre-selected spectrum of colors. Ie, a manufacturer could adjust the weights of the different color wells, to make a warm white, or neutral white, or blue-green, or whatever color LED without having to use phosphors.

However, there is no way for the end-user to change whcih color is being produced. There are only two sets of contacts -- all the user can determine is the brightness of the device by setting the current. The color produced will be dependent on the manufacturing process.

[EricB]

Originally Posted by 2xTrinity

You're right. However my proposed 5-primary scheme dramatically improves the situation. Amber will look very good (eg, that single LED only), and lime-green will be possible by mixing amber and green (with much better results than mixing red and green).

I've always wondered what green + amber would look like. With all the two-primary signs out now (RyG, RmB, ApB), they don't have ones with amber and the other colors.

Likewise, mixing cyan LEDs and green LEDs will allow a lot more possible blue-green shades, than mixing a deep blue with a green LED.

The B+G mixing makes a nice rich cyan (You can even see this on those 7-color LED's). It's much different than the aqua/turquoise 00FFFF of the screen. So I would think just adjusting the green to lower intensities after cyan would simulate mixing of it with cyan.

This wouldn't necessarily make the unit too big. For example, a red-amber-green-blue system could fit in a 2x2 grid. As LED dice are almost always inherently square, they are usually packed using "square packing" like this. Many RGB LEDs (in signs etc.) use one red, one blue, and two green dice aranged in a 2x2 grid already. Replacing one of the greens with a yellow/amber would improve the range of colors without changing the size of the system at all.

I knew about the RGB-IR units which Ledtronics had from the beginning. But you mentioned five colors. Would that still fit? An 8 or 10mm envelope, perhaps.

I personally would love to see Cree come out with a version of the MC-E loaded with separately-addressable red, amber, green and blue LEDs. It might take a fairly complicated set of diffusers and optics to mix the colors evenly, but it could sure make for some fun-to-use toys

All this time, I actually didn't know what MC-E was, so I just looked it up. So that's just a white with individually addressable dies? What's that for; brightness?

The gray section is the range all possible colors, with saturated monochromatic colors along the rim. The highlighted "triangle" is the gamut, or possible range of colors that can be made with the displa (in this case a CRT, with red, green, and blue phoshpors)

The phosphors have fairly wide spectral line width (ie, they aren't very saturated), so that is why the primary colors are not along the edge. Filtered fluorescent (eg, your LCD monitor) and LED also can have fairly wide line width, leading to a limited gamut.

As you pointed out, it's not possible to make truly saturated colors in some ranges. Another scheme that has been proposed is to use laser light as the light source. Laser light is extremely monochromatic (ie, totally saturated). Look at that graph and imagine three points located at 650nm, 532nm, and 405nm RIGHT along the edge of the CIE curve. The triangle drawn between those three points covers a lot more area than the case of the CRT, meaning many more saturated colors would be possible with a laser light source.

Even an LED that could change its center-wavelength to any aribtrary value might have a hard time competing with a laser-based system, as even though the central wavelength might be moved, the LED will still have a significant spectral line width (read: even single color LEDs aren't perfectly saturated).

I've heard about usinglasers as LED's. I don't know how that works. Lasers are generated by LED chips, arent they? How do you use one in place of an LED?

OK, I realize that it would be hard to cross the InGan line, as the blues were nearly impossible without them (I didn't know if InGaN could make reds, and I know that old blue made with those compounds was too dim). So we could start with an InGaN chip that could cover the 400-540 range, and then 550 and up would be covered by the compound that makes those colors. Then, we could have a two-die LED that would cover the whole visible range.

Interesting idea. mixing RGB to get white for a display system is analogous to the reason why black ink is used for CMYK systems -- as it's almost impossible to make true black by mixing cyan magenta and yellow.

For something like a display on a portable device, using a white LED would likely be more power efficient than making white with a mixture of RGB dice.

The only way I can think of would be to have a

Looks like what you were saying got cut off there.

in general increasing the dopant concentrations will tend to result in a larger voltage drop (eg, produce shorter wavelengths) However, this change is proportional to the logarithm of the concentration, so it takes a very large change in doping to produce a small change in wavelength for an LED device.

If you make the concentration too high, then the highly concentrated impurities will start to significantly disrupt the structure of the crystal, and also result in very high unwanted resistance. This will set an upper limit on wavelength possible from a material.

If you make the concentrations too low (eg, to try to produce longer wavelengths), the device will simply become an insulator. What dopants do is add free electrons and holes that are capable of conducting current. With no doping/free carriers, there can be no current.

I can't think of any way to "simulate" different dopings in a material. I believe you are imagining a device behaving somewhat like a transistor -- where by applying a certain voltage or current you can produce light at a different color. Unfortunately, there is no way to modulate voltage-drop or effective of a semiconductor device that I know of. Amplifying a current flow is completely different than changing the fundamental properties of a material "on the fly".


The actual device they are showing is somewhat interesting. They are in effect creating an LED that emits light at a whole bunch of different spectral lines by sandwiching many materials, with different dopings etc. each together. However, in the body of the paper, it is clear that this device is meant for generating some pre-selected spectrum of colors. Ie, a manufacturer could adjust the weights of the different color wells, to make a warm white, or neutral white, or blue-green, or whatever color LED without having to use phosphors.

However, there is no way for the end-user to change which color is being produced. There are only two sets of contacts -- all the user can determine is the brightness of the device by setting the current. The color produced will be dependent on the manufacturing process.

Well, wouldn't it be possible to hook up the contacts to the different layers so they can be turned on independently?

I also found this:
http://web.mit.edu/newsoffice/2002/dot.html

From the Wikipedia article on LED's that linked to this:

Quantum Dot LEDs
A new technique developed by Michael Bowers, a graduate student at Vanderbilt University in Nashville, involves coating a blue LED with quantum dots that glow white in response to the blue light from the LED. This technique produces a warm, yellowish-white light similar to that produced by incandescent bulbs.[30]

Quantum Dots are semiconductor nanocrystals that possess unique optical properties. Their emission color can be tuned from the visible throughout the infrared spectrum. This allows quantum dot LEDs to create almost any color on the CIE diagram. This provides more color options and better color rendering white LEDs. Quantum dot LEDs are available in the same package types as traditional phosphor based LEDs.

This says "tuned", but is that again only by the manufacturing process and not the end user?

[Light Sabre]

Kinda just thinking out loud here: check out of following links about multi-color LED products where a particular color can be selected:

ThinkGeek :: Multi-color LED Flashlight

ThinkGeek :: Multi-Color LED Lightbulb w/Remote

I have the NiteIze frisbee called the Disk-O that changes colors. Looks like it has a single led package in it with 3 separate dies Red, Green, Blue. Has a chip that cycles thru all the colors.

[saabluster]

Originally Posted by EricB

Quantum Dots are semiconductor nanocrystals that possess unique optical properties. Their emission color can be tuned from the visible throughout the infrared spectrum. This allows quantum dot LEDs to create almost any color on the CIE diagram. This provides more color options and better color rendering white LEDs. Quantum dot LEDs are available in the same package types as traditional phosphor based LEDs.

This says "tuned", but is that again only by the manufacturing process and not the end user?

They are not talking about something that can be tuned by an end user. The quantum dots just act in the same way as phosphors used in LEDs today do only they are more efficient and can create a broad spectrum of light output. The color of the LED is set on the day of manufacture.

[EricB]

Originally Posted by Light Sabre

Kinda just thinking out loud here: check out of following links about multi-color LED products where a particular color can be selected:

ThinkGeek :: Multi-color LED Flashlight

ThinkGeek :: Multi-Color LED Lightbulb w/Remote

I have the NiteIze frisbee called the Disk-O that changes colors. Looks like it has a single led package in it with 3 separate dies Red, Green, Blue. Has a chip that cycles thru all the colors.

Those use separate LED's; I'm talking about an LED with a single chip that can produce any visible wvelength without mixing colors.

That flashlight I want for Christmas, or whenever! That is the closest thing to what I had been asking about before. You can change the color, though you can't mix them completely. (I had mentioned perhaps slider dimmers for red green and blue so that you can make any color).

I have also finally seen that remote control bulb in a Chinatown bulb store. Nice colors. But not cost effective to me. If you could mix any color, or if it was cheaper, then I would get it.

[2xTrinity]

Originally Posted by EricB

All this time, I actually didn't know what MC-E was, so I just looked it up. So that's just a white with individually addressable dies? What's that for; brightness? I've heard about usinglasers as LED's. I don't know how that works. Lasers are generated by LED chips, arent they? How do you use one in place of an LED?

Well, the reasons there are for dice in one package is to increase lumen output. Other companies have quad-die white LEDs, such as the SSC-P7, but as far as I know the MC-E is the only one where the dice are separately addressable. The advantage is that it may be wired in series, parallel, or 2-series 2-parallel, making them more flexible to use with different power supplies. For example, it's a lot more efficient and cheaper for a wall wart transformer to generate 14V at ~700mA (series wiring) than than 3.5V at 2.8 A (parallel). In a flashlight application however running on say 3-4 NiMH batteries, parallel wiring may be easier than boosting the voltage up.

Well, wouldn't it be possible to hook up the contacts to the different layers so they can be turned on independently?

The layers create "quantum wells" that behave like phosphors -- they convert one form of light into another, except instead of emitting a broad spectrum of light, they re-emit light at specific discrete wavelengths. This gives manufacturers more options for fine-tuning the spectrum of their product than what is possible with phosphors. This will likely mean more possibilities for tint control for white LEDs. To produce white, they simpyl pack in tons of different quantum dots.

However, the only way for a user to create different colors would be to have spearate dice, each with a different quantum dot.

...This says "tuned", but is that again only by the manufacturing process and not the end user?

yes, unfortunately

That flashlight I want for Christmas, or whenever! That is the closest thing to what I had been asking about before. You can change the color, though you can't mix them completely. (I had mentioned perhaps slider dimmers for red green and blue so that you can make any color).

My guess is they are using RGB LEDs. The only problem with those when you use them in a flashlight, rather than using them as a "pixel" for a display, you end up projecting an image of the LED -- which means non-overlapping red, green, and blue regions.

A better approach to mix color, which is used in LED backlit projectors, is to use SEPARATE color LED packages, then mix the components using dichroic prisms (these are prisms that reflect some colors and transmit others). Projectors use this method to "split" a single white light source into RGB components.

I've heard about usinglasers as LED's. I don't know how that works. Lasers are generated by LED chips, arent they? How do you use one in place of an LED?

The lasers are being talked about more for projectors, or projection-based monitors rather than flat-panels. Basically three highly powerful lasers would replace the HID lamp in the projector. Through a series of lenses, the laser light would be expanded, then projected through LCDs or DLPs. The advantage is that the colors are FAR more saturated than what can be done with LEDs, or filtered white light. I suspect that this technology may remain vaporware however, as lasers powerful enough to act as a light source for a TV projector are not at all cheap.

[rushnrockt]

Originally Posted by 2xTrinity

Well, the reasons there are for dice in one package is to increase lumen output. Other companies have quad-die white LEDs, such as the SSC-P7, but as far as I know the MC-E is the only one where the dice are separately addressable. The advantage is that it may be wired in series, parallel, or 2-series 2-parallel, making them more flexible to use with different power supplies.

Just pitching in on that one, SSC-P7 is produced in either series, parallel or 2-series 2-parallel. How to obtain them outside of direct sampling, that I have not looked at.


As for the main topic, for one LED to be able to reproduce all colors, it will have to be able to (efficiently) knock out electrons at many different energy levels and that would require quite a complex chemistry of the semiconductor.

[EricB]

Originally Posted by 2xTrinity

My guess is they are using RGB LEDs.

If you mean a single LED with the three dies, rather than separate LED's (SRGB), then I really want that flashlight now!. (Come to think of it, the rgb bullb did look like it was producing solid colors, and not separate red, green and blue LED's

The only problem with those when you use them in a flashlight, rather than using them as a "pixel" for a display, you end up projecting an image of the LED -- which means non-overlapping red, green, and blue regions.

A better approach to mix color, which is used in LED backlit projectors, is to use SEPARATE color LED packages, then mix the components using dichroic prisms (these are prisms that reflect some colors and transmit others). Projectors use this method to "split" a single white light source into RGB components.

The separate images I can understand, but I still don't understand why separate LED's in a flashlight would reduce that problem. Would a single rgb behind a prism be even better?

The lasers are being talked about more for projectors, or projection-based monitors rather than flat-panels. Basically three highly powerful lasers would replace the HID lamp in the projector. Through a series of lenses, the laser light would be expanded, then projected through LCDs or DLPs. The advantage is that the colors are FAR more saturated than what can be done with LEDs, or filtered white light. I suspect that this technology may remain vaporware however, as lasers powerful enough to act as a light source for a TV projector are not at all cheap.

So is this basically sort of like cathode ray tube, with a beam producing the pixels on the screen? If so; I imagine it would be very good colorwise, perhaps even better than OLED (which still does not seem to have quite the saturation of the primaries as LED's). I wonder how it can be even more saturated than LED.
But then, it would seem you would bneed a big tube again. I guess it wouldn't have to be a vacuum, so like rear projection, you could just have it firing on a flat screen with nothing inbetween.

[2xTrinity]

Originally Posted by EricB

If you mean a single LED with the three dies, rather than separate LED's (SRGB), then I really want that flashlight now!. (Come to think of it, the rgb bullb did look like it was producing solid colors, and not separate red, green and blue LED's The separate images I can understand, but I still don't understand why separate LED's in a flashlight would reduce that problem. Would a single rgb behind a prism be even better?

Because the prism can make it "look like" the separate light sources are perfectly overlapping each other. In an RGB package, unless you diffuse the LED somehow, if you magnify it with a lens of reflector, the colors in the beam won't be perfectly overlapping -- there will be red, green, and blue "lobes" similar to the geometry of the LED surface.

So is this basically sort of like cathode ray tube, with a beam producing the pixels on the screen?
That could be one way of doing it. If so; I imagine it would be very good colorwise, perhaps even better than OLED (which still does not seem to have quite the saturation of the primaries as LED's). I wonder how it can be even more saturated than LED.

Lasers are extremely monochromatic -- they release only at one wavelength, whereas LEDs have maybe a +/- 15nm "band" of spectrum where they emit energy. CRTs are still based on phosphors, so they too have a fairly broad spectrum.

But then, it would seem you would bneed a big tube again. I guess it wouldn't have to be a vacuum, so like rear projection, you could just have it firing on a flat screen with nothing inbetween.

There's no need for a tube when you're projecting light (as opposed to electrons). You could even use lasers in a portable projector.

[EricB]

OK; here we go, now, finally; it looks like:
http://berkeley.edu/news/media/relea...graphene.shtml

(New article); this is being contrasted to current semiconductors in transistors and LED's which have "fixed" properties.

"Now, University of California, Berkeley, researchers have shown that a form of carbon called graphene has an electronic structure that can be controlled by an electrical field, an effect that can be exploited to make tunable electronic and photonic devices."

Still trying to understand some of this stuff about light waves and fields, but it looks like this is finally it!

Here also seems to be another kind, where a single die is made of micrometer sized "pixels" emitting different wavelengths:
http://www.itf.gov.hk/eng/Search/pro...?Prj_code=3016

Edit: on the first page, I thought it said aomething about the full visible spectrum. Now I see it says this particular mix can do the zero-250 meV bandgap which is "frequencies anywhere in the far- to mid-infrared range". (visible light is basically 1-3 eV). It does say "Ultimately, it could even be used for lasing materials generating light at frequencies from the terahertz to the infrared". Now, terahertz does encompass the visible spectrum, and I think part of the IR. So I hope that means it can someday expand to the visble range, even if it is just IR now.

[saabluster]

Originally Posted by energysmartin

hey,

True color sensors are based on one of the color models, most commonly the RGB model (red, ... SBCs are composed of a microprocessor, memory chip, and serial and parallel ... The most efficient RGB full color LED available in the industry.try here Energy Smart Industry (ESI)

reported

[DM51]

Originally Posted by saabluster

reported

Thanks - he's gone.