Finding true cyan LEDs

[JoakimFlorence]

Re: A new source

If you look at blue-green, like 505-515 traffic signal LED's, it looks almost like cyan.

(The colors begin slowly normalizing themselves).

I agree. Saturation effects can make it very challenging to accurately judge precise color tint. Those 490 nm LEDs look kind of bluish for the first instant you look at the light, but after a while the eye starts adjusting and they begin to look more and more greenish than anything else. And all the surroundings look rose-tinted when you look away.

It is incredibly difficult to discern where precisely green ends and blue truly begins. It depends on how you look at it (literally) and I suspect relative color contrast plays a large part in our perception. When I look at green for a few seconds and then look at the blue-green spectral line it looks kind of blue. But then when I look at blue light for a few seconds and then look back at the spectral line it looks like a turquoise green. Obviously the spectral line does not change but my precise perception of the color hue changes. It is nearly impossible to decide objectively whether the line is more green or more blue.

For this reason I would consider it truly cyan, since I cannot objectively differentiate whether it has a little bit more green in it or a little more blue. It just seems blue-green, and that's all that can be said about it.

I'd really be interested if they made a bluish cyan colored wavelength LED, but alas, those are impossible to find. Kind of like the same color as those old VFD tubes that were used in the 1980s.

 

[JoakimFlorence]

Re: A new source

I just noticed that the LED digital display on my kitchen microwave is true cyan.

Not very high light output though.

 

[BVH]

Re: A new source

I still have this beauty by Mac and it is my all-time favorite light. Hardly use it but it's always in view when I'm on the PC. It uses the Q2H bin LED. Looks very Cyan to me when in the higher current modes. I also had the same bin LED put in an old Fenix L1P

http://www.candlepowerforums.com/vb/showthread.php?146752-Quad-Cyan-*SOLD*

 

[archimedes]

I have a nice cyan light engine, that was included in a large group of assorted vintage McLux parts I got a while back, and now runs in a Mirage_Man host ....

Sorry I don't know the details on the emitter, however

 

[JoakimFlorence]

Just got stopped by a police car. But what I noticed was the blue and red lights. The blue LEDs appeared to be a very soft slightly cyan-tinted blue hue, it definitely could be 485-490nm.

I know what 470nm and 480-485nm looks like, and these were not that color. But I just have difficulty believing the police lights could actually be using 485-490nm because I have searched everywhere and never seen that wavelength range available (at least not in >1W LEDs), so I'm considering if there could be some other explanation. Maybe the LEDs could be "ice blue" emitters and they only appear like 485-490nm behind the blue color filter. Ice blue emitters can commonly be purchased in cheap automotive lighting products and consist of a 470nm emitter covered by a little bit of green phosphor. These normally appear like a whitish blue, but passing through the blue plastic, that could be altering their spectrum. However, most ice blue emitters I have seen appear to be a more bluish color hue than cyan, so I'm not sure how plausible that is. These would probably have had to have been much more greenish ice blue emitters (edit: I think these are called "aqua blue"). Theoretically, passing such a "cold green" light spectrum through a blue filter could lead to a spectrum that contains 470-500nm wavelengths, and this could superficially appear similar to 485-490nm. Maybe.

Apparently there are some Japanese companies manufacturing semi-high power LEDs in this wavelength region.

Epitex L490-06 5mm greenish blue, peak wavelength 490nm typ.

Murubeni SMC490 SMD LED greenish blue 490nm typ.
20mA 3.5V brightness 75mcd

The specifications show that these LEDs appear to have very wide spectral emissions about 30nm half width
480-500nm minimum and maximum

Unlike those "490nm cyan" emitters found on ebay, these 490nm Japanese emitters are clearly categorized as "blue".

 

[CuriousOne]

I've measured some "cyan" emitters from ebay, with spectrophotometer. They have quite wide emission range, so using specific filters, it is possible to get "real" cyan from them, but at very high brightness cost.

 

[CuriousOne]

Bought some "480nm Cyan" leds from ebay. They are actually a blue leds with 450nm peak, wtih 520nm phosphor applied

 

[JoakimFlorence]

Bought some "480nm Cyan" leds from ebay. They are actually a blue leds with 450nm peak, wtih 520nm phosphor applied

Yes I know. This is very common if you cheap LEDs from China.


Update on the police car lights: I happened to see a police car driving by and the exterior blue-color filter plastic was cracked and I could distinctly see an LED emitter lit up inside. It seemed like it was kind of dangling a little bit on a strip, so apparently there had been some damage. What was interesting was the color of light being given off by this emitter, it was a whitish green-turquoise. I could not make out the color exactly since it was daytime but it had to be either a 495nm emitter or a green phosphor-converted emitter. Very interesting. What was this LED doing behind a blue color police light?

My conclusion: there are three possibilities. The first is that they are using 495nm cyan-green mixed with 470nm to achieve the optimal blue tint. I don't find this particularly likely since it's very rare for commercial products to mix different wavelength LEDs just to achieve something this subtle. The next possibility is they are using 495nm behind a blue filter. The blue filter would shift the color tint to blue (or a bluish cyan). Could be possible, but very inefficient, would cut down at least 60% of the light output, but that's cheap chinese engineering for you. The last possibility is phosphor conversion, and I actually have some of these chinese emitters. They are very cheap, and the color is whitish-green, or sometimes bluish-green (depending on the bin color). The bluish part comes from blue light from the emitter shining through the phosphor. I find this last possibility most likely. I didn't get a good look at the emitter but it appeared to be SMD-type. These green phosphor LEDs contain a trough in the cyan part of the spectrum that's about 60% lower than the green or blue maximum. So the spectrum is all filled in but there's just a deficiency of wavelengths in between maybe around 510 to 475nm. You put a blue color filter in place and that could potentially make the wavelength graph about level. This isn't all that different from using 470nm blue because the underlying emitter in these greenish phosphor LEDs (the cheap chinese ones I'm familiar with) is 470nm (not the usual 450-455nm in white LEDs). From the type of color of light I saw (coming from the bare LED), these would give off some 470nm blue light but there wouldn't really be a blue spike in the spectrum because the blue is much lower than the green. So here you have a simple, albeit inefficient, way of achieving a mostly smooth spectral distribution with the right color tint. And I'm sure these lights are manufactured in China.

 

[CuriousOne]

I have a segment of blue led cluster, removed from police car (These small sections of blue and red that go besides front vent openings). It has 5mm blue leds and additional blue filter in front. Will measure spectrum tomorrow.

 

[CuriousOne]

Oh, forgot to say, I found a seller of 5mm "true" cyan 582nm leds on taobao, who claims that 482nm claim is 100% true. I've ordered some, and they are on the way, let's see how it will go.

 

[CuriousOne]

These LEDs arrived, and as measurements show, he was correct, these are true cyan, single die, no phosphor or double crystals, just pure cyan leds.

 

[anuragwap]

Cheap 3535/XPE styled 'epileds' 490-495nm cyans when mounted on DTP copper boards, driven at 1 amp and properly heatsinked, blue-shift noticeably more my bin1 luxeon rebels mounted on sinkpad mounted on oversized ohmite heatsink. I don't have a spectrometer, but I guess the color is very close to 488nm dominant.

These LEDs arrived, and as measurements show, he was correct, these are true cyan, single die, no phosphor or double crystals, just pure cyan leds.

Although 482nm might be too blue, still will be an uncommon color. Can you please share the source with the rest of us?

 

[BVH]

I just acquired a Sekonic C-7000 Spectrometer and measured the PhotonFanatic "To Die For" Cyan emitters in my Mac's Quad Cyan light. The Peak Wavelength is about 495-497 NM. When the light is on HIGH, it appears very Cyan to me with a small bit of green. Lower levels produce more green. Has anyone decided what NM wavelength is really Pure Cyan? Is the 490-495 the best that can be had these days? If so, a source would be appreciated.

 

[FRITZHID]

488nm? Seems to be the most cyan-y cyan to my eyes.

 

[anuragwap]

If so, a source would be appreciated.

If you want to go the luxeon rebel way, see TCSS (at the bottom, they say all their cyans are bin1), but do contact them beforehand to make sure if they are indeed bin1. If you hate their steep price (I do), order some cut tapes of 'lxml-pe01-0060' from mouser/digikey/future and there is a very good chance of them being bin1 cyan. (You have to reflow them onto starboards though)

If you can live with the low lumens and high V_f of epiled chips, get them from aliexpress or ebay by searching '490nm cyan led' and order the ones that look like cree xp-e. As noted, you need to desolder and reflow them onto dtp copper boards to unleash their full potential. Realizing this, some chinese sellers on Ali are also offering them on flashlight friendly 16mm / 20mm dtp stars.

 

[JoakimFlorence]

Has anyone decided what NM wavelength is really Pure Cyan?

In terms of a spectral line (like something from a laser) I believe exactly 490 is cyan, where most people would be unable to differentiate whether the color looks more greenish or more bluish.
However, this does NOT translate over to a 490nm LED (so don't make that assumption!). There are two reasons for that. First, virtually all of the Chinese sourced "490nm" emitters that you will find anywhere should more accurately be described as 490-495nm and probably peak somewhere at about 492nm (this is just what I believe, from my personal observations). The second reason is that LEDs do not actually truly give light out at a single discrete wavelength value (although it is relatively narrow). The bandwidth of emissions trails off further towards the longer wavelengths. Normally this does not lead to any color difference, but when we're talking about LEDs in the cyan territory it's a different story, since the transition from blue to green is very sudden. This means that, in practice, a cyan LED is going to appear somewhat more greenish than its peak wavelength value would be. A real 490nm LED would probably not appear exactly the same color hue as a 490nm laser.

488nm is a very beautiful light blue color with a tinge of cyan.​

I have shopped around quite a bit and all of the LEDs advertised as 490nm were, what I would consider, a bit more on the greenish side of cyan. One of them was even covered with a thin bluish colored piece of plastic film that appeared to be designed to try to shift the color a little bit more towards a truer cyan (although it was only half-way successful towards this end, in my opinion).

So, to conclude this post and provide a simple answer to your question, if you wanted a true cyan color it would have to be a 489 nanometer LED (or to be more broad, a 488-490nm distribution, although I'm sure a 485-490 would be quite a lovely blue-cyan as well).

 

[ssanasisredna]

You are going to always be fighting both the spectral content of the "cyan" LED which is usually quite wide, and the eyes drastic shift in efficiencies in that area.

Perhaps the best way to do this is with a filter? http://www.saikoled.com/scientific.html <-- They do it with very expensive filters, but there could be a lower cost solution out there? It all depends on how perfect you want your cyan. If it's for a flashlight, then maybe some of the Polymer optics mixing optics designed for Luxeon rebels and adding together blue and green to a cyan to get your best perception of cyan. It will be like looking at cyan, but not like illuminating with cyan.

 

[anuragwap]

Off topic, but just noticed that DTRLPF now has 55mW 490nm Sharp diode lasers in his website for $50, they do 160mW when pushed.

 

[WeLight]

We stock the Nichia Cyan which we sell to aircraft landing light company so they are as cyan as we have seen