Why are color LEDs so inefficient? 2

[geneiusxie]

I'm looking at the efficiencies of color LEDs and many of them have terrible efficiencies compared to the best white LEDs like the Cree XM-L which gets between 100-160 lumens per watt. I know that the eye is less sensitive to blue and red light, but even green LEDs only have a fraction of the lumens per watt as the Cree XM-L or other white LEDs.
I thought green LEDs are made just by taking a blue LED and adding a green phosphor, just like how white LEDs are made just by taking a blue LED and adding a white phosphor.

Red LEDs are (I think) naturally red without needing phosphors (phosphors are bad because they reduce efficiency)

But in that case, green LEDs should have at least the same or greater lumens per watt as white LEDs be the same number of photons are released, but the eye is more sensitive to green light than other wavelengths.

Also, are blue/green/red lasers more efficient than LEDs? I have read documents stating that VCSELs (a type of vertical chip laser) convert 66% of electrical energy to light - compared to about 33% for the most efficient white LEDs. The only problem is whenever I google actual VCSELs for sale, they're always designed for data transmission so they have really low output power and efficiency - more like 6% or so.

So where are all the super-efficient color LEDs and lasers people keep writing about?

 

[IRstuff]

http://www.cree.com/LED-Components-and-Modules/Products/XLamp/Arrays-Directional/XLamp-XML-Color

Other than blue, they seem plausible

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

 

[MagicSmoker]

geniusxie - that is actually a very good question. My guess is that most colored LEDs are used as indicators, not lighting sources, per se, and so don't need to be terribly bright or all that efficient.

Now where you might find colored LEDs that are very bright and highly efficient is in outdoor displays, particularly for use in active billboards. Mind you this is purely conjecture on my part - it's a bit outside of my normal field of interest - but it seems reasonable.

As for semiconductor diode lasers, they are surprisingly fragile beasts. Indeed, mistreatment of diode lasers often causes them to stop lasing and just turn into LEDs.

 

[VEBill]

"...green LEDs are made just by taking a blue LED and adding a green phosphor..."

Green LEDs were available many many years before blue LEDs became available. If my memory is correct, LEDs of various colours became available on the practical market starting with red (at one time, essentially all consumer product LEDs were red) and more or less worked their way up the spectrum towards blue. It made the tech news when blue LEDs were finally developed.

I believe that the trick of using phosphor in LEDs is a more recent development on the LED timescale. Reason being, a typical phosphor is excited by a higher frequency (shorter wavelength) and emits light at a lower frequency (longer wavelength). I don't know phosphor theory, but I don't recall every seeing an example that worked in the opposite direction (going up spectrum).

E.g. Blue in, yellow out. (Blue + yellow makes "white".)
E.g. Flourescent tube UV to visible

Thus phosphor LEDs had to wait until blue and UV LEDs were invented.

Disclaimer: all based on recollection. I might be incorrect on some if this history.

 

[MacGyverS2000]

Where do I begin...

Green LEDs are not blues with phosphor, they're a different base material, like Gallium Nitride or Gallium Phosphide (blues are typically Silicon Carbide or Indium Galium Nitride).

But even if they were made with blue LED bases, the number of photons would likely not be identical... that's a function of the phosphor's efficiency. A yellow phosphor and a green phosphor would likely not have identical efficiencies given the same drive wavelength. You could figure it out very easily if you knew what the electron-volt energies were for each color emission.

Let's also not forget that blue and green are primary colors... impossible to get one primary from another by mixing in other colors (primary, secondary, or tertiary), so a green LED from a blue one does not make sense.


Dan - Owner

http://www.Hi-TecDesigns.com

 

[VEBill]

In answer to your main question, Wiki hints that it might be fierce competition between Cree and Nichia in the illumination market (white light).

Worth noting: these amazingly bright and amazingly efficient "White" LEDs are typically actually Blue LEDs with yellow phosphor added.

 

[geneiusxie]

So couldn't Cree just take their white XM-L and just skip adding the phosphor to make a super-efficient and powerful blue LED?

Apparently the appropriate term for phosphor efficiency is "quantum yield" which is basically the number of emitted photons divided by the number of photons absorbed.
According to

http://apps1.eere.energy.gov/buildings/publications/pdfs/ssl/ns/tian_phosphors_sandiego2011.pdf

high efficiency phosphors have yields of 70-90% so phosphor losses aren't nearly enough to explain why color LEDs are less than half as efficient as white LEDs (even after accounting for the retina's varying sensitivity to different wavelengths.)


Here's a brand name color LED that's designed for lighting - not some wimpy indicator LED.

http://www.cree.com/LED-Components-and-Modules/Products/XLamp/Arrays-Directional/XLamp-XML-Color

Also, nobody seems to have answered my question of where I can buy high efficiency lasers - 66% efficiency or better?
All the high power laser diodes used in 1000mw+ laser "pointers" seem to have a calculated efficiency of 33% or lower, which is pretty good, but still half as efficient as what was claimed for VCSELs.
Laser projectors seem to have a 11 lumen per watt efficiency according to my calculations.

 

[VEBill]

...nobody seems to have answered my question of where I can buy high efficiency lasers...

You've posted here in the Circuit design forum, where many people are quite familiar with LEDs. Fewer of us would have gone very deep into laser topics (such as efficiency).

Perhaps your laser question would be better posted in the Laser engineering forum.

Good luck.

 

[MacGyverS2000]

high efficiency phosphors have yields of 70-90% so phosphor losses aren't nearly enough to explain why color LEDs are less than half as efficient as white LEDs (even after accounting for the retina's varying sensitivity to different wavelengths.)

Your equating two totally different concepts. Lumen measurements are based upon the eye's sensitivity to a specific wavelength. Colored LEDs do not use phosphors, their base materials emit at the desired wavelengths.

Our eyes have a (roughly) order of magnitude difference in sensitivity in perceiving yellow light versus blue, so there's a huge boost in luminous efficacy when a yellow phosphor is excited by a blue LED, despite the <100% quantum yield of the phosphor itself. Since luminous flux is measured at a single frequency, things get interesting when you have a dual-emitting source.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[IRstuff]

Dan is alluding to luminosity function and luminous efficacy

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

 

[VEBill]

See also: [URL unfurl="true"]http://www.handprint.com/LS/CVS/color.html[/url]

It's a gold-mine of good info.

 

[jimkirk]

VE1BLL,

Slightly off topic, but for what it's worth, there are phosphors that convert infrared to visible light. I have a couple on laminated cards, one from General Electric and one from Radio Shack. Handy things when testing IR remote controls. (Be very careful if using with IR lasers.)

Here are a couple links where they may be available.

http://www.roithner-laser.com/cards.html

http://www.bromba.com/indicard.htm

More on topic, measuring light is fraught with peril. Do you measure what is visible to the human eye or an unweighted scale (photometry versus radiometry)? Total flux emitted or flux per area? Lots of different units and lots of conversions.

Lumileds has green and cyan LED with efficiencies of 100 lumens per watt or so. Blues and deep reds are usually measured differently because they fall on the edge of human perception. Yellow LED are generally not very efficient because of solid state physics issues I don't pretend to understand, but phosphor convert yellow LEDs (similar to "white" LEDs but without the blue component) are very efficient.

I agree also, the research into incandescent & fluorescent lamp replacement has driven the efficiency increases in "white" LEDs beyond other colors.

Here's another useful reference on light:

http://www.colorkinetics.com/support/whitepapers/evaluating_light_output.pdf

 

[VEBill]

jimkirk, thanks for the counter-example. Very good.

 

[MacGyverS2000]

Ugh, Color Kinetics... the only company to get away with patenting PWMing LEDs. I despise them for that, and Philips taking them over was a shame. I liken them to the Monster Cable type of the LED world.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[geneiusxie]

What... how is there a patent on PWM for LED's? I have that in four of my flashlights, and two of them are made by some obscure chinese company and I doubt they pay patent fees for anything (though I could be wrong.) If there is a patent, it's probably expired or they just aren't enforcing it.

 

[IRstuff]

Their patent application was filed in 2001:

http://www.freepatentsonline.com/6788011.html

Given the list of references, it seems to have been a fairly robust patent

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

 

[MacGyverS2000]

And they ARE enforcing it... they've made quite the little nest egg for themselves over the years with that (and similar) patents. Company down in Orlando got the short end of the stick with that one...

Dan - Owner

http://www.Hi-TecDesigns.com

 

[John2025]

Off topic, but the patent system is thoroughly broken. They give out patents for anything and let the courts and the lawyers fight it out. Instead of spurring innovation, it's just stifling productivity and the economy is going to suffer for it.
Rant over,
John