Re: Dichromatic White LEDs

My apologies for an earlier attempt to followup being sent as an
e-mail-to-author because I hit the wrong key...

In article <1139791333.180514.141750@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>,
PhilRimmer@xxxxxxxxx wrote in part:

Anybody else had experience of these phosphor-free "white" LEDs? They
use a single chip (!) to produce two distinct frequencies of light
(blue and yellow). The CRI must be terrible. Reds and greens are very
dark, but shone in your eye they look white....and very bright!

Basically, they're what Fred Schubert proposed a couple of years back,
and have the potential to produce "white" light at 400lm/W, or so, as a
maximum theoretical efficacy.

Anybody seen any efficacy claims for these LEDs yet?

I have not yet seen efficacy claims, but I have had experience with a
couple such LEDs!

The ones I tried were zinc selenide whites. These had a single LED chip
that had two distinct emission bands - a "greenish blue" one and a
"yellow-orange" one.

The greenish blue emission band was the narrowest I ever saw in a
non-laser, non-superluminescent LED. My "eyeball estimate" with a
diffraction grating was in the 5-10 nm range for FWHM bandwidth. The peak
wavelength I estimated to be around 480 or in the 480's of nm.

The yellow-orange band appeared to me more ordinary.

Now for more on my experience with these:

Overall luminous efficacy was less than that of better 2-year-old
more-usual white LEDs that I had then, a couple years ago. So I would
guess 10 lumens out per watt in at best and probably less.

The color rendering index was outright horrible. I would dare to guess
a color rendering index in the 20's.

One positive point was voltage drop lower than usual for a white LED -
around 2.7-2.8 volts if I remember correctly.

Other negatives were requirement for less current and power, shorter
life expectancy, and life expectancy degraded by overcurrent even at
currents that should not cause much temperature rise in the more-usual
phosphored InGaN white LEDs.

As for 400 lumens per watt: I would guesstimate that the yellow-orange
band alone had about 400 lumens per watt of output in that band - along
with probably around 100-125 for the greenish blue band. Makes me think
probably around 300 lumens per watt of output. With the most efficient
popular visible LED of any kind that I am aware of being roughly 25%
efficient, I don't give much hope for 400 or 300 lumens per watt of input.

Dichromatic "white light" with 400 lumens per watt of output is
achievable. This can be done with two narrow bands, one centered around
450 nm and the other in the mid-upper 570's nm. For now, I would
guesstimate 450 lumens per watt out for CCT in the upper 3,000's with
monochromatic blue light in the upper 440's or close to 450 nm and
monochromatic yellow light with wavelength in the upper 570's or close to
580 nm. I have seen a CIE chromaticity diagram with curves drawn for
maximum lumens per watt of visible light for dichromatic light, and I
somewhat remember the maximum for white or warmish white around 440-450 or
so.
But surely even 100 lumens of white light out of an LED lamp per watt in
is a big deal and I have yet to hear of this being achieved, even for a
laboratory prototype.

I have heard a few and several years ago of "white" LED lamps having
more than one chip, but using only two colors (blue and yellow, or
greenish blue and orange). Back then these would have had higher overall
luminous efficacy than was achieved by blue LEDs having a phosphor that
produces wideband yellow-centered output to mix with some "unconverted"
blue output.
More recently, the blue LED chips used in white LEDs with phosphor have
outpaced orange and yellow LEDs in advancement of efficiency, so nowadays
white LEDs with a blue chip and phosphor achieve higher overall luminous
efficacy than do LED lamps with 2 different color chips - as well as much
better color rendering!

- Don Klipstein (don@xxxxxxxxx)
.

Relevant Pages

  • Re: 683 lm/W, how was this calculated?
    ... luminous efficacy of that wavelength is 683 lumens per watt of light. ... Wavelength of band center, ...
    (sci.engr.lighting)
  • Striaght Dope on efficiency of light sources
    ... efficiency of light sources, usually measured in lumens per watt. ... Another site claims that the best white LEDs give 37 lumens per ...
    (sci.energy)
  • Most Efficient Light Source??
    ... I'd read that incandescent bulbs produce about 13 lumens per watt, ... Japanese researcher has claimed to produce LEDs with double the normal ... strained silicon or strained silicon-germanium can result in charge ...
    (sci.energy)
  • Re: Circuit Design Q. - battery charger
    ... With their angle of 15 deg this would calc to a bit under ... lumen or about 14.4 lumen per watt. ... Thanks for the info:) I've been having a bit of trouble with LEDs ... I've been having trouble translating "watts" into "mcd (and lumens since ...
    (sci.electronics.design)
  • Re: Ultra bright LEDs
    ... The efficiency of production LEDs to top out around 20 Lumens ... figures of luminous efficacy in terms of lumens per watt. ...
    (sci.electronics.misc)
Loading