Odd blue glow after HID shutoff
Daniel J. Stern
I noticed something interesting while doing some experiments with Philips
D2R automotive HID capsules the other day. After the lamps are
de-energized, the glass outer envelope (the purpose of which is to block
UV emissions to permit HID headlamps with plastic components, because the
inner arc envelope is fused quartz and does not have UV-stop
properties) glows with a violet-blue color. The effect appears almost
immediately; a brief, sub-one-second burst of power applied and then
removed causes the blue glow in the glass shield in the immediate vicinity
of the inner arc capsule. A longer application of power, long enough to
vaporize the metal halides and get full output from the lamp, creates a
stronger and bigger area of blue glow, to the point that the entire outer
glass shield glows strongly "nuclear blue" after shutdown and for quite
some seconds afterwards.
Fascinating! Can anyone explain it? Some fluorescent mineral or oxide
present in the glass? I tried it with a Taiwanese-made D2S capsule, which
did NOT exhibit the effect at all. I do not have any Osram capsules to
compare to the Philips items.
DS
Ioannis
>
> I noticed something interesting while doing some experiments with Philips
> D2R automotive HID capsules the other day. After the lamps are
> de-energized, the glass outer envelope (the purpose of which is to block
> UV emissions to permit HID headlamps with plastic components, because the
> inner arc envelope is fused quartz and does not have UV-stop
> properties) glows with a violet-blue color.
Happens to regular quartz discharge tubes on high pressure mercury vapor
lamps, too.
[snip]
> Fascinating! Can anyone explain it? Some fluorescent mineral or oxide
> present in the glass? I tried it with a Taiwanese-made D2S capsule, which
> did NOT exhibit the effect at all. I do not have any Osram capsules to
> compare to the Philips items.
Excluding thermionic emissions from the quartz capsules, (i.e. excluding
the regular red glow from heated quartz when one turns a lamp off) when
the lamp shuts off, the temperature of the vapor inside the capsule is
still high enough to cause fluorescent and/or phosphorescent transitions
by indirect uv excitation on some of the residual mineral ions inside
the quartz material.
The glow disappears when the vapor cools off.
It can even be seen after a simple glow discharge on a mercury quartz
tube, which does not have high enough of a vapor temperature, but can
cause temporary phosphorescence on the quartz material, because of the
253.7 line.
> DS
--
Ioannis Galidakis
<http://users.forthnet.gr/ath/jgal/>
_______________________________________________________
"It was written that I should be loyal to the nightmare
of my choice." - Heart of Darkness Joseph Conrad
Daniel Stern Lighting
> > D2R automotive HID capsules the other day. After the lamps are
> > de-energized, the glass outer envelope (the purpose of which is to block
> > UV emissions to permit HID headlamps with plastic components, because the
> > inner arc envelope is fused quartz and does not have UV-stop
> > properties) glows with a violet-blue color.
> > Fascinating! Can anyone explain it? Some fluorescent mineral or oxide
> > present in the glass? I tried it with a Taiwanese-made D2S capsule, which
> > did NOT exhibit the effect at all. I do not have any Osram capsules to
> > compare to the Philips items.
>
> Excluding thermionic emissions from the quartz capsules, (i.e. excluding
> the regular red glow from heated quartz when one turns a lamp off) when
> the lamp shuts off, the temperature of the vapor inside the capsule is
> still high enough to cause fluorescent and/or phosphorescent transitions
> by indirect uv excitation on some of the residual mineral ions inside
> the quartz material.
I may misunderstand what you are saying, but the blue glow is very clearly
in the glass surrounding the quartz, not in the quartz itself. Do you
comments still apply? Presumably the glass that did not glow (in the
Taiwanese capsule) lacks these residual mineral ions.
DS
Ioannis
[snip]
> I may misunderstand what you are saying, but the blue glow is very clearly
> in the glass surrounding the quartz, not in the quartz itself. Do you
> comments still apply? Presumably the glass that did not glow (in the
> Taiwanese capsule) lacks these residual mineral ions.
Very likely. I would surmize that if it happens on Fused Silica tubes,
which are relatively free from residual foreign ions, the mechanism
would be all the more visible for glass types which are much less pure
than a quartz discharge FS tube.
I don't know what the outer glass for those bulbs you mention is made
of, but it is useful to keep in mind that many glasses and rock salts,
including many quartz varieties phosphoresce under uv.
The phosphorescence frequency (blue) indicates a quite low excitation
wavelength, which suggests the 253.7nm mercury lines.
For the case that you mention, there may be other phenomena at work here
as well, such as Raman or secondary resonance excitation, etc. but my
guess is that the glass phosphoresces because of the hot capsule vapor.
> DS
--
Ioannis Galidakis
<http://users.forthnet.gr/ath/jgal/>
________________________________________
"Nobody realizes that some people expend
tremendous energy merely to be normal."
-Albert Camus
Carl Tauber
shroud that are fluorescing. Some of the excited states can have
lifetimes of several tenths of a second. It's s smart way of
achieving the necessary UV block properties, since plain absorbers
convert the UV to heat. Fluorescent materials allow some of that
energy to be reradiated as visible light. Don or Ioannis will
probably know the exact numbers but I recall seeing estimates of the
UV flux from a metal halide lamp that were on the order of 25% of
the visible flux: IOW the potential heating effect of absorbing this
flux is large. That's why they don't put the absorbers directly into
the arc envelope.
The effect is probably not noticeable when the arc is lit since it
is so much brighter than the shroud. I wonder if this is what gives
certain HID headlights the characteristic blue/violet color when
viewed off-axis. Given your familiarity with the optics of these
headlights, can you estimate the beam distribution if the shroud
itself were the light source? Perhaps this was done deliberately to
enhance the look of these lamps compared to their competition.
Interestingly, I think that there is a relation to your post about
Neodymium lamps. I never understood the use of Nd because its
principal fluorescence is at 1064 nM which is not visible. Nd would
act purely as an absorber, subtracting light from the lamp output.
But other ions could be chosen that would fluoresce in the visible,
and perhaps enhance the efficacy of a metal halide lamp by
converting some of its UV output to visible. Perhaps even the CRI
could be enhanced if a material could be found with suitable red
emission. The effect would be much more significant for a metal
halide lamp than an incandescent.
Carl
-- Carl Tauber
carlt...@worldnet.att.com
Daniel Stern Lighting
> It sounds as though they have used some rare earth ions in the glass
> shroud that are fluorescing. Some of the excited states can have
> lifetimes of several tenths of a second. It's s smart way of achieving
> the necessary UV block properties, since plain absorbers convert the
> UV to heat. Fluorescent materials allow some of that energy to be
> reradiated as visible light.
This all makes perfect sense.
[lots of heat]
> That's why they don't put the absorbers directly into the arc
> envelope.
The original automotive HID capsule, the D1, and its successor, the D2,
had just the bare quartz arc envelope with no UV shield. Obviously, NO
plastic parts in headlamps for use with these capsules, and probably all
kinds of problems finding a UV-resistant topcoat for the reflector.
(Note, confusing as it is, the old D1 is long gone and N/A, but there is
now the D1R and D1S, which are D2R and D2S with ignitors integrated into
the base, allowing longer lengths of primary wire to a more compact
external ballast.)
> The effect is probably not noticeable when the arc is lit
I did notice it quite strongly in a specially-prepared optic and viewing
from a certain angle.
> so much brighter than the shroud. I wonder if this is what gives
> certain HID headlights the characteristic blue/violet color viewed
> off-axis [...] Perhaps this was done deliberately to enhance the look
> of these lamps compared to their competition.
You read my mind! :-{)}
> Given your familiarity with the optics of these headlights, can you
> estimate the beam distribution if the shroud itself were the light
> source?
BROAD and TALL. So if in fact the fluorescence of the shroud is as
extreme as it appears to be, it could potentially contribute materially to
the general blue/violet cast of the beam. However, the primary contributor
to the blue off-axis appearance is very clearly in the optics, and not in
the source. This is especially true of polyellipsoidal optics, which can,
as I pointed out in a previous post, be manipulated to create violet,
blue, yellow, red, or -- with some finesse -- probably whatever other
color one wants. It is interesting to note that Audi's recent Valeo-made
projector HID *and halogen* headlamps have an extremely strong dark
blue-violet appearance off axis in the fringe area of the cutoff. My
experimentation with the cutoff shield location relative to the condensing
lens shows that this blue fringe is very much a function of the cutoff
shield and the condensing lens.
It is almost certainly possible to eliminate or dramatically reduce this
chromatic aberration, but until the relevant colorimetry regulations
clearly demand uniform colorimetry throughout the beam, there's no legal
reason to do so, and "blue headlamps" are, very regrettably, all the
fashion rage.
It is also interesting to note that many paraboloid-reflector HID
headlamps do NOT exhibit a strongly-blue fringe, and in some cases the
beam itself isn't very strongly blue, but the blue cast it does have is
quite uniform. On the other hand, many of the "wanna be kewl" kids who
buy "HID retrofit kits" to replace the halogen bulbs in their headlamps
with HID capsules complain of excessive *yellow* coloration to the beam
that does not meet their expectations of blue. Certainly optics not
designed for HID light will not handle it properly, but I do wonder about
the contribution of the blue glow in some bulb brands' shroud.
> Interestingly, I think that there is a relation to your post about
> Neodymium lamps. I never understood the use of Nd because its
> principal fluorescence is at 1064 nM which is not visible. Nd would
> act purely as an absorber, subtracting light from the lamp output.
Nd changes the apparent color of the light. Some people find the
resultant light pleasant to work or live by. The subtraction of absolute
output makes such bulbs disastrous by which to drive, in my
experience. On the other hand, I have several Nd Oxide (sorry, don't know
the chemical formula notation) lamps in my living space, not because of
any magical benefits, but because the apparent light color works well with
what needs to be illuminated.
> other ions could be chosen that would fluoresce in the visible, and
> perhaps enhance the efficacy of a metal halide lamp by converting some
> of its UV output to visible. Perhaps even the CRI could be enhanced if
> a material could be found with suitable red emission.
You read my mind again! I don't know that red, per se, would be what to
shoot for, but certainly something in the orange-to-yellow-orange area
would be interesting to work with. It's demonstrated that the blue-heavy
output spectrum of an automotive HID beam is 46% more glaring for
observers on strictly colorimetric grounds (when intensity is corrected
for) than the output spectrum of an automotive halogen bulb.
Monkeying with the halides in the arc capsule to add some orange and
yellow back into the output would, as I understand it, reduce the efficacy
of the source (not to mention eliminating the visual "I spent a lot of
money on HID headlamps" product differentiation...). I wonder, though, if
color correction through shroud fluourescence like this could be made to
work usefully. One obstacle that I can think of right off the bat is that
smaller, more sharply-defined point sources are a lot easier to make into
well-controlled beam patterns than larger, fuzzily-defined point sources.
The "fuzzy" glowing plasma around the outside of the vaguely cylindrical
arc of an automotive HID capsule creates its own challenges in beam
control (and is why halogen headlamp optics that look for sharply-defined
filament coil edges don't properly handle fuzzily-defined edges of an
arc). If the shroud glow were to act more or less as a color filter, that
would be one thing. But if there were significant light emission from
shroud glow, it might make beam control difficult. Sounds like a fun
project to work with!
--Daniel
TO WRITE TO ME: Remove the headlamp from my return address.
.______DANIEL STERN LIGHTING______.
| http://lighting.mbz.org |
---
dastern "at" vrx "dot" net
Automotive Lighting and Signalling Services
NBCS b6f+wg++rp
Don Klipstein
>On Sun, 22 Jul 2001, Carl Tauber wrote:
>
>You read my mind again! I don't know that red, per se, would be what to
>shoot for, but certainly something in the orange-to-yellow-orange area
>would be interesting to work with.
I think red plus pure green like that from thallium would be great, by
illuminating red objects and green objects more brightly than yellow or
orange light would. Remember how red and green objects look under sodium
light?
>It's demonstrated that the blue-heavy
>output spectrum of an automotive HID beam is 46% more glaring for
>observers on strictly colorimetric grounds (when intensity is corrected
>for) than the output spectrum of an automotive halogen bulb.
>
>Monkeying with the halides in the arc capsule to add some orange and
>yellow back into the output would, as I understand it, reduce the efficacy
>of the source
Maybe not - more sodium could even improve the efficiency. But there
are some drawbacks:
1. Unless used to some extreme, its output (when used in a metal halide
lamp) is in a narrowish region of the spectrum from yellow to orange. I
expect its effects on color rendering to be negative.
2. Sodium produces a wider, more diffuse arc than other metal halide lamp
ingredients - especially scandium. Try using a lens to project an
image of a metal halide arc onto a wall - you will see an orange
fringe around the greenish-bluish-white core of the arc.
Making the arc wider can force lots of the optics designing work to be
redone.
3. Would, as Daniel says...
(not to mention eliminating the visual "I spent a lot of
>money on HID headlamps" product differentiation...). I wonder, though, if
>color correction through shroud fluourescence like this could be made to
>work usefully. One obstacle that I can think of right off the bat is that
>smaller, more sharply-defined point sources are a lot easier to make into
>well-controlled beam patterns than larger, fuzzily-defined point sources.
>The "fuzzy" glowing plasma around the outside of the vaguely cylindrical
>arc of an automotive HID capsule creates its own challenges in beam
>control (and is why halogen headlamp optics that look for sharply-defined
>filament coil edges don't properly handle fuzzily-defined edges of an
>arc). If the shroud glow were to act more or less as a color filter, that
>would be one thing. But if there were significant light emission from
>shroud glow, it might make beam control difficult. Sounds like a fun
>project to work with!
- Don Klipstein (d...@misty.com)
Don Klipstein
>It sounds as though they have used some rare earth ions in the glass
>shroud that are fluorescing. Some of the excited states can have
>lifetimes of several tenths of a second. It's s smart way of
>achieving the necessary UV block properties, since plain absorbers
>convert the UV to heat. Fluorescent materials allow some of that
>energy to be reradiated as visible light. Don or Ioannis will
>probably know the exact numbers but I recall seeing estimates of the
>UV flux from a metal halide lamp that were on the order of 25% of
>the visible flux: IOW the potential heating effect of absorbing this
>flux is large. That's why they don't put the absorbers directly into
>the arc envelope.
I think that 25 percent figure is very much on the high side.
A mercury lamp may have about 30-35 percent as much power in the 365-366
nm line cluster as it does in its blue, green and yellow lines.
Adding the halides increases radiation opportunity (mainly visible) and
this lowers the arc temperature enough to just about wipe out all UV
shorter than the 365-366 nm cluster, although enough 313 nm may remain to
cause problems. There is so little UV below 365-366 that the usual "DX"
phosphor is more of a diffusion coating than something that contributes
much to light output or even color correction. I have seen such MH lamps
and the lines from the "DX" phosphor are almost buried among scandium
lines and a weak orenge-red double line of sodium.
If the 365-366 maintained its strength, the over 50 percent increase in
visible light will knock the UV/visible ratio down to 20-25 percent. And
the 365-366 nm line does suffer. For a number out of my hat, I think more
like UV being 10-15 percent as much optical power as visible.
>The effect is probably not noticeable when the arc is lit since it
>is so much brighter than the shroud. I wonder if this is what gives
>certain HID headlights the characteristic blue/violet color when
>viewed off-axis. Given your familiarity with the optics of these
>headlights, can you estimate the beam distribution if the shroud
>itself were the light source? Perhaps this was done deliberately to
>enhance the look of these lamps compared to their competition.
I have seen some UV-absorbing 300 watt quartz halogen lamps for
torchiere lamps. The quartz tubing slightly fluoresces whitish blue.
Guess what - it glows under a regular blacklight, although it does not
convert longwave UV to visible light as efficiently as the optical
brightener in white cotton underwear does.
- Don Klipstein (d...@misty.com)
Daniel Stern Lighting
> >You read my mind again! I don't know that red, per se, would be what to
> >shoot for, but certainly something in the orange-to-yellow-orange area
> >would be interesting to work with.
>
> I think red plus pure green like that from thallium would be great, by
> illuminating red objects and green objects more brightly than yellow or
> orange light would. Remember how red and green objects look under sodium
> light?
<snip>
> Scandium
<snip>
Good points. And there are lots of reds and greens in the nighttime
traffic environment! My knowledge of the output from various elements is
quite limited without a textbook readily to hand. That said, the effect
of light color on various driving-performance tasks under mesopic
conditions (to say nothing of the various visual conditions caused by e.g.
direct glare, backdazzle in disturbed environments, etc.) has not been
extensively studied...yet. I could happily spend LONG periods of time in
a lab and a light tunnel if I had access to a ready means of creating
custom arc lamps.
Wonder what RPI's LRC has in that arena...
DS
Ioannis
[snip]
> I think that 25 percent figure is very much on the high side.
>
> A mercury lamp may have about 30-35 percent as much power in the 365-366
> nm line cluster as it does in its blue, green and yellow lines.
> Adding the halides increases radiation opportunity (mainly visible) and
> this lowers the arc temperature enough to just about wipe out all UV
> shorter than the 365-366 nm cluster, although enough 313 nm may remain to
> cause problems. There is so little UV below 365-366 that the usual "DX"
> phosphor is more of a diffusion coating than something that contributes
> much to light output or even color correction. I have seen such MH lamps
> and the lines from the "DX" phosphor are almost buried among scandium
> lines and a weak orenge-red double line of sodium.
Don's analysis is correct. However AFAIK the only commercial MH bulbs
that use a "DX" coating, are the Sylvania MetalArc/SuperMetalArc series.
Neither Philips nor (originally) OSRAM used phosphors on their MHs. The
only two types which used to use coatings were the OSRAM HQI-E 400W and
the Philips HPI 400W types. And in both of these cases they were just
diffusion coatings not phosphors.
I have also examined the spectra of both a coated MetalArc and a clear
MetalArc of the same Wattage and the difference in red phosphor
emissions is really minute. It's there alright, but it's buried under
scandium lines as Don says.
> If the 365-366 maintained its strength, the over 50 percent increase in
> visible light will knock the UV/visible ratio down to 20-25 percent. And
> the 365-366 nm line does suffer. For a number out of my hat, I think more
> like UV being 10-15 percent as much optical power as visible.
>
> >The effect is probably not noticeable when the arc is lit since it
> >is so much brighter than the shroud. I wonder if this is what gives
> >certain HID headlights the characteristic blue/violet color when
> >viewed off-axis. Given your familiarity with the optics of these
> >headlights, can you estimate the beam distribution if the shroud
> >itself were the light source? Perhaps this was done deliberately to
> >enhance the look of these lamps compared to their competition.
>
> I have seen some UV-absorbing 300 watt quartz halogen lamps for
> torchiere lamps. The quartz tubing slightly fluoresces whitish blue.
> Guess what - it glows under a regular blacklight, although it does not
> convert longwave UV to visible light as efficiently as the optical
> brightener in white cotton underwear does.
The most prominent example I can recall is a simple GE Germicidal tube.
Try the following, but exercise extreme caution with the light from this
lamp. Light it at night for 2-3 minutes and move away from it, then shut
it off and look at the tubing in total drakness. It glows really
brightly for a couple of minutes. The germicidal glass contains AFAIK
some rare earths, but don't know exactly which (If anybody knows its
exact composition, please share the info). And in this case the 253.7
cluster is not blocked, supressed or altered by anything so it acts
directly on the rare earth glass.
> - Don Klipstein (d...@misty.com)
--
Ioannis Galidakis
<http://users.forthnet.gr/ath/jgal/>
_________________________________________________________
"We've just replaced the Enterprise's Dilithium Crystals
with new Folgers Crystals. Let's see if they notice ..."
Carl Tauber
> I think that 25 percent figure is very much on the high side.
>
snip
>.... For a number out of my hat, I think more
>like UV being 10-15 percent as much optical power as visible.
Don,
I actually found the reference in my Random Access Filing System; it
refers to a 400 watt sodium-scandium metal halide lamp of the "plain
vanilla" type. So it's probably not directly applicable to the super
high pressure short-arc lamps like the automotive types.
Electrical input 400 watts
Discharge Radiation
UV 46 watts
Visible 136 watts
IR 38 watts
(Reference: Nelson, G.J., Gibson, R.G., and Jackson, A.D. : An
Efficacy Analysis of HID Lamps; JIES vol. 30, 1, winter 2001.
The authors are with Philips Lighting)
So the UV radiation is ~10% of electrical input and ~ 33% of the
visible output. I found this very surprising.
Carl
-- Carl Tauber
carlt...@worldnet.att.com
H Claus
They probably use an UV absorbing quartz, the same glass that is used for
the UV free halogen and HID (quartz) lamps. The material is fused silica
with small dopes of Cerium Oxide and Titanium Oxide. It absorbs almost 100%
below 350nm and cuts the 365nm radiation significantly. The fluorescence
that you see is caused by radiation in the 300 - 365nm range ... and there
is a little afterglow.
"Daniel J. Stern" <das...@engin.umich> wrote in message
news:Pine.HPX.4.21.010721...@topaz.engin.umich.edu...
Jon Connell + K. Brett Malak
ballasts for it) and I can't see the effect you mention. I would guess that
you are kicking some electrons up a few orbits - probably with uv, as this
lamp is much more blue at ignition - and they come crashing back down a few
seconds later losing the energy as light.
Ever tried peeling apart a Band aid wrapper in total darkness - the glue emits
orange / pink light as it tears apart...thats mechanical energy to light -
very strange... explain that one for me...
Jon Connell
"Daniel J. Stern" wrote:
--
Things That Swing Inc.
1661 11th Avenue #D6
Brooklyn, New York 11215
Telephone: 718.499.7832
Facsimile: 718.832.7717
http://www.thingsthatswing.com
in...@thingsthatswing.com
************************************
If this email is not intended for you, or you are not responsible for the
delivery of this message to the addressee, please note that this message may
contain Things That Swing Inc. Privileged/Proprietary Information. In such a
case, you may
not copy or deliver this message to anyone. You should destroy this message
and
kindly notify the sender by reply email. Information contained in this
message that does not relate to the business of Things That Swing Inc. is
neither endorsed by
nor attributable to Things That Swing Inc.
************************************
Daniel Stern Lighting
> I have the Phillips equivalent of that lamp here (we sell 120 / 230 volts
> ballasts for it)
For D2R/D2S...? Never heard of such a ballast. Interesting. I'm going
to write for more info.
> and I can't see the effect you mention.
Which specific lamp are you trying it with, and when were they produced?
I've pulled Philips D2Rs from a bunch of production runs going back a few
years, and they all do this. I have not tested it with a D2S, and wonder
if perhaps some surface treatment necessary for the shadow mask to adhere
to the UV screen might have some sort of fluorescent effect...But, that
said, we did have a very cogent explanation of the phenomenon a couple
weeks ago here.
> I would guess that you are kicking some electrons up a few orbits -
> probably with uv, as this lamp is much more blue at ignition
It's more blue at ignition because at ignition, it's a pure-Xenon arc.
Once the metal halides vaporize, it takes on a whiter appearance (and the
red-to-orange glow of hot quartz probably contributes to the apparent
color, as well.)
> Ever tried peeling apart a Band aid wrapper in total darkness - the
> glue emits orange / pink light as it tears apart...thats mechanical
> energy to light - very strange... explain that one for me.
I actually can: Wintergreen oil in the glue. Get any candy with
wintergreen oil in it (NOT artificially flavored), go in a dark bathroom,
let your eyes dark-adapt, dry your back teeth, grin and crunch the candy
in front of the mirror. Northern lights in your mouth! I don't remember
the name of the effect. Somethingluminescence... ;-{)}