Is violet really the best way to go?

Anders Hoveland

Enlightened
Joined
Sep 1, 2012
Messages
858
Is violet really the best way to go?

http://www.ledinside.com/news/2015/8/purple_leds_to_replace_blue_leds


"It is only a matter of time before white LEDs using blue LED chips will disappear from the market", said Shuji Nakamura at a forum on GaN technology in July organized by Nikkei Asian Review.

Shuji Nakamura is the one who invented the blue LED, and consequently the white LED as well, which catapulted the world into the modern age of LED lighting technology.

So why would Nakamura think blue-chip based white LEDs are out? Because white LED technology based on violet-chips can produce a fuller coverage over the spectrum of different wavelengths, providing a "more complete" white light that feels more natural and, in particular, provides more accurate color rendering to different blue color hues than the white LEDs that use blue emitters do.

Theoretically, converting violet to white light should result in less efficiency than the current method of using blue chip emitters. But Nakamura has compensated for this inherent efficiency loss by developing his violet emitters based on "GaN-on-GaN technology". This results in a better crystal quality, which both enables the new violet chips to handle more power, and makes them more efficient. Of course, there's no inherent reason why blue LEDs could not be made with GaN-on-GaN technology, but for the time being Nakamura's new violet chip white LEDs have a higher efficiency than most of the regular white LEDs on the market.

So what's not to love?

The issue is that prolonged exposure to shorter wavelengths (such as violet) may not be healthy for eyes and skin.

Philips developed "CrispWhite" technology LED lighting for retail, which basically just adds some 410nm violet emitters to the regular white LED design. This has the advantage of making white colored objects with optical brighteners appear whiter and brighter, an attribute that is lacking with normal white LED lighting. A few years ago there was an online article that interviewed a Philips executive about the new CrispWhite technology. One of his comments was that they were a little reluctant to use shorter violet wavelengths because it was not really good for people to be under, that they were more for sterilization. Those were his words, as best I can remember. These CrispWhite LED spotlights are intended primarily for retail clothing display, not for mainstream lighting.
(While violet wavelengths are much less damaging than ultraviolet, they not completely innocuous either)

Another company called Vital Vio is marketing violet-emitter white LED fixtures to hospitals, with the idea that the presence of violet wavelengths will help disinfect the environment. (These violet-emitter LED fixtures are not even high CRI either; the technical specifications listed in the company brochure states 82 CRI).

Some time ago, GE developed the VIO LED, which used a 405nm violet emitter in conjunction with a blue phosphor. The whole reason behind using a violet chip, which was a completely novel concept at the time, was primarily to allow dimming in LED lamps without any color shift, since if the blue light emanated from the phosphor it wouldn't matter if the wavelength of the emitter shifted a little bit when the voltage was turned down. However, the GE VIO LED had the unintended benefit that the quality of the light appeared surprisingly natural (despite the CRI only being 85). Although the VIO never caught on and was discontinued, one of the members in this forum tested a 4W GE Vio in 2009 and commented at the quality of light (you have to remember, there were not a lot of super high CRI LED options available at this time).

Violet wavelengths of light are completely natural—it is a component of natural sunlight. But being out under the sun for the entire day is not a good thing. In fact few people actually do it, or at least do it continuously. This was even true in ancient times, though I'm not going to delve into the details of all that here, that could make for a long discussion. You're likely to get very badly sunburned if you stay out in direct sunlight for even 5 hours a day without any covering.

While artificial lighting such as Fluorescent tubes and Metal Halide put off less damaging ultraviolet wavelengths than sunlight, they do have a substantially higher ratio of plain violet in their spectrum. (Because of Rayleigh scattering in the atmosphere, the sunlight that actually reaches earth's surface has a somewhat truncated spectral distribution from a true Black-body) The advent of these forms of artificial lighting are actually relatively recent if we look back historically. Light bulbs started becoming widespread 125 years ago, and a decade before that gas mantle lamps were already in use. Neither of these two put out much violet light (not surprising for a 2800-2900K Black-body spectrum). Fluorescent tube lighting didn't really start appearing everywhere until the 1960's.

Many people complain about aching eyes from being in the office under fluorescent light all day, and several studies have indicated that fluorescent lighting may raise an individual's lifetime risk of carcinoma and cataracts. While "hazardous" would probably be too strong a word, it's probably not the healthiest thing, or the most conducive to individual wellbeing.

One of the advantages, one could say, of the switchover from fluorescent to LED is that it's light doesn't contain damaging shorter wavelengths. Is it really such a great idea to go backtracking and add back in the violet? This LED lighting is going to be everywhere, in schools, in offices, even in people's homes.
:eek:
 
Last edited:

SemiMan

Banned
Joined
Jan 13, 2005
Messages
3,899
It's a bit intellectually dishonest not to post the full article including rebuttals or the Nakamura is financially tied to a purple die company.

Even Nakamura was not quite forthcoming in his presentation. "Due to purple LEDs high crystal quality and good light extraction efficiency, its wall-plug efficiency (WPE) can reach 84%. Average blue LEDs WPE in general is between 50% to 60%."

1) WPE efficiency for purple LED is currently no where near 84%.

2) WPE for blue LEDs already regularly exceeds 60% and gets better all the time. Theoretical limits for both are similar.

Of course, Nakamura has also not solved Stokes losses, so the use of a purple pump will, based on current knowledge, always incur an efficiency hit compared to blue when stimulating phosphors at longer wavelengths.

Semiman
 

JoakimFlorence

Newly Enlightened
Joined
Jun 4, 2016
Messages
147
One thing I have noticed about LED streetlamps is that the color tint appears to be different depending on the viewing angle. Looking directly under the lamp the light is very blue-tinted, but looking at the lamp from a distance the lamp appears very greenish-yellow tinted. I suppose this is not surprising; while the yellowish phosphor is emanating light in all directions, much of the blue light being emanated from the diode underneath passes straight through without being diffused or changing direction.

If the blue light is being generated from the phosphor rather than the diode, the color tint is going to be consistent at different angles.
 

JoakimFlorence

Newly Enlightened
Joined
Jun 4, 2016
Messages
147
The way I understand it, just adding violet wavelengths does not really improve CRI. The reason violet chips improve the CRI is because the phosphor is then able to convert it to a range of different blue wavelengths, not just a single wavelength peak.

Of course, there are other potential ways to raise the CRI without violet chips. For example, the white LED light could simply incorporate some indigo (430-440nm) and sky blue (485nm) emitters. The spectrum does not need any violet light to approach 99 to 100 CRI.
 

JoakimFlorence

Newly Enlightened
Joined
Jun 4, 2016
Messages
147
A 440nm royal blue emitter could be used with a cyan phosphor. A 450nm wavelength might be just a little too long to get a cyan peak out of the phosphor, but typically a peak phosphor wavelength can be achieved that is 50 nanometers less than the excitation wavelength. So with a 440nm emitter, it should be possible to achieve a true ~490nm centered emission that is going to contain plenty of longer cyan-blue wavelengths. Then the gap between the 440nm and the cyan would be filled by the 450nm blue peak from a regular white LED. This type of design arrangement would lead to higher efficiency, because most of the blue light would be generated directly by the emitters rather than the phosphor, and the Stokes loss would be lower because we would not need to be converting violet wavelengths into red.

This would essentially have the same efficiency as regular white LEDs, and only involve 2 types of emitters. But a thought did occur to me. Whenever you have phosphors generating blue wavelengths mixed in the same batch with the other phosphors, that's going to lead to an inherent theoretical loss in efficiency. Because a portion of the blue light being emitted is going to be reabsorbed by the phosphor and downconverted into other wavelengths. So you have violet being converted to blue, in turn being converted to orange. That's going to cause efficiency losses. I'm not sure what this inherent efficiency loss is, maybe around 5 to 10 percent I'm guessing.

The other option would be to separate the phosphors into 2 distinct LEDs, perhaps right next to each other on the same chip. This could potentially necessitate some sort of light mixing optics, if color consistency within the beam is important. (Not too difficult; just have the LED chip facing backwards into a parabolic reflector, if you've ever seen those spotlights over the supermarket produce section)
 
Last edited:

SemiMan

Banned
Joined
Jan 13, 2005
Messages
3,899
One thing I have noticed about LED streetlamps is that the color tint appears to be different depending on the viewing angle. Looking directly under the lamp the light is very blue-tinted, but looking at the lamp from a distance the lamp appears very greenish-yellow tinted. I suppose this is not surprising; while the yellowish phosphor is emanating light in all directions, much of the blue light being emanated from the diode underneath passes straight through without being diffused or changing direction.

If the blue light is being generated from the phosphor rather than the diode, the color tint is going to be consistent at different angles.

Anders Hoveland, why are you posting under 2 different names?
 
Top