How Low Can You Go?

Part 5
Put on Those UV-defying Blues Brothers Shades.
New Readers - Don’t start here. Go back to Lesker Tech Volume 1 Issue 4 where this stuff started as a trilogy
(Ha! A 5-part trilogy!). Also, my editor says this one’s so far out in left field, it will only interest the academic few.
But. . . it’s about outgassing. Isn’t everyone interested?

Introduction
Wow! Did Lesker Tech Vol. 2 Iss. 2 cause a brouhaha!
High Speed
To paraphrase - and distort to suit my purposes - two read-
ers scolded me for claiming UV light was useless for out-
Combinatorial Thin-Film
gassing chamber walls. I didn’t say that! I would never Materials Research...
say that! It isn’t true! Short wavelength UV has enough
Speeds film design and synthesis, and vastly
energy to break chemical bonds.
increases deposition tool throughput and efficiency.
But the furor made me reflect and . . . what I wrote wasn’t
We manufacture and sell worldwide thin-film
crystal clear. Let me re-phrase that . . . I didn’t define the
deposition systems incorporating Symyx®
type of lamps I was talking about and what’s written is, combinatorial patents and Symyx® Library
therefore, subject to misinterpretation. Aw heck, let’s face Studio® software.
it–I screwed up!
Kurt J. Lesker Co. and Symyx Technologies recently
In consequence, I’ll take one final stab at describing light entered into a worldwide licensing agreement.
as an outgassing agent with emphasis on the lighters and
what makes them tick.
We make your science easier.
I’ll start with words about wavelength so we’re all nod-
ding our heads at the same frequency when UV is men-
tioned. But, since it’s easier to understand the concepts,
this section is mostly about visible and IR wavelengths.

Then I’ll do a mini-tutorial on lamps–the obvious

lighters–emphasizing UV light output or lack thereof.

Finally, and this is a first, I’ll show results illustrating

UV degassing. Let me hasten to add, these results are
from Tom von Alten, one of the two readers. Tom sent
data taken during routine daily operations of his system.
The data are, therefore, unadorned, untweaked, and
unoptimized. . . the best kind!

©Kurt J. Lesker Company 2003 For more information, or to purchase vacuum products on-line, visit www.lesker.com.
 "...vacuum science is our business."
Wavelength UV-assisted outgassing. Selecting appropriate lamp types
To talk sensibly about light or, more generally, electro- require much more investigation than given here.
magnetic radiation, we have to talk wavelengths. Wave- Lamps divide into two groups:
length is, of course, the parameter that distinguishes: 1. Filament lamps (also called incandescent)
yellow light from green light; blue light from x-rays; or,
2. Discharge lamps (also called arc)
putting it in more gut-wrenching terms, your microwave
oven from your toaster. And to talk wavelength we have to I won’t define ‘discharge’ but you should know it involves
talk units. I’ll use nanometers (nm). If you like angstroms, making positive and negative species that make the gas
multiply every number I give by 10. If microns are your electrically conducting.
bag, then divide by 1000. The two groups are further sub-divided:
Visible radiation is from ~400nm (blue) to ~700nm (red)
so anything a bit less than 400nm is UV radiation and any- 1. Filament lamps
thing slightly greater than 700nm is IR radiation. How-
ever, don’t cut the limits too fine. Radiation at, say, 390nm A) Vacuum
is in the UV range all right, but each photon just doesn’t B) Gas Filled
carry enough oomph to break interesting chemical bonds. C) Halogen
You probably need to get down to 320nm or so to start
cutting the chemical mustard and breaking C–C bonds.
2. Discharge lamps
Similarly, but differently, just over the red line, photons
carry the maximum IR energy. Trouble is, assembling lots A) Fluorescent
of 710nm photons in one place is a challenge. For exam- B) Neon
ple, a filament at 2200°C, puts out ~96% of its energy C) Xenon
between 700nm and 10,000nm. However, it only puts
D) Sodium
out 0.5% between 700nm and 720nm. Its peak energy
output (the wavelength at which most of the photons are E) Mercury
released) is around 1172nm. F) High Intensity Discharge
For the theorists out there, yes, I’m assuming a blackbody
filament and I’m using a wonderful little applet: Why do so many different types exist? Well, it’s all about
http://thermal.sdsu.edu/testcenter/javaapplets/planckRadiation/ getting more photons in the visible range and getting
[For the practiquists out there, first: how do you like your more photons per watt. A report prepared for the Energy
new collective name? and second: don’t worry about this Efficiency and Conservation Authority of New Zealand
blackbody stuff. It’s explained in the next issue when I shows an interesting comparison.
discuss sample heating.]
The applet lets me make other pronouncements about Lamp Type Efficacies
2200°C filaments. Roughly 3% of the energy released is (lumens/W)
between 400nm and 700nm. So, lamps with 2200°C fil- Incandescent <20
aments are much better heaters than lighters. And UV Mercury Vapor
content? Between 100nm and 400nm the fraction of total (High Pressure) 50
energy emitted is 2.8 x 10-4 . . . that is, their UV emission
Mercury Vapor
characteristics are pathetic.
(Low Pressure Fluorescent) 80
Metal Halide
Shining a Light on Lamps (Mercury Vapor) 90
This is a seriously limited discussion of lamp variations. Sodium
If I went into every niche lamp manufacturers claim (High Pressure) 120
make their lamps different from the other guy’s, this
Sodium
would be a 10-part trilogy. I’m not saying the lamps
(Low Pressure) 150
aren’t different but I’m only giving you an introduction to
©Kurt J. Lesker Company 2003 For more information, or to purchase vacuum products on-line, visit www.lesker.com. 2
 "...vacuum science is our business."
Filament Lamps apocryphal). Fluorine, chlorine, bromine, and iodine are
all halogens and, in their elemental state, are a highly
1A. Vacuum reactive, nasty bunch - the chemical Kray twins.
Ordinary domestic lamps have But, hold the phone! Add iodine to a tungsten filament
tungsten filaments raised to lamp and all sorts of wonders happen. First; iodine (I) and
moderately high temperatures tungsten (W) react at high temperature; second; the prod-
in a transparent (at visible uct is volatile; and third (and quite fortuitously); the reac-
wavelengths) vacuum envelope. tion is reversible. That is,
The temperature is a compro-
W + 2I2 ! WI4
mise. Tungsten’s vapor pres-
sure increases with temperature goes to the left at very high temperatures and to the right
which means the filament evap- at just high temperatures.
orates faster, coats the envelope So, iodine reacts with the tungsten film on the envelope
walls (reducing light output), Ordinary Lamp.
to form WI4 vapor. These molecules bounce around until
and eventually ‘blows’. Clearly, the consumer doesn’t they contact a super-hot point where they promptly break
want this, so the filament temperature must be low. But into component atoms and deposit tungsten. And where’s
low filament temperature gives a very yellow light and that super-hot spot? Yep, you’ve got it, right where the
consumers want white light. filament is thinnest.
Vacuum incandescent lamps, as suggested by the 2200°C In other words, WI4 vapor breaks down and deposits tung-
filament data above, are reasonably efficient thermal sten in just the right place to restore the filament’s thin-
sources. But they are not used for vacuum bakeout because ning diameter to its original value. These devices, known
even a 100W lamp takes up so much volume. These lamps as tungsten-halogen or quartz-iodine lamps, are small,
are only installed in a vacuum chamber to illuminate high powered, efficient generators of thermal energy and
something that moves . . . are you into watching dinner are used extensively in vacuum as bakeout heaters.
rotate in the microwave oven too? And just why have I taken all this space on a heater when
describing lighters? Well, in Lesker Tech Vol. 2 Iss. 2, I
1B. Gas Filled was trying to make the point that tungsten halogen lamps,
while brilliant white, are still filament lamps. If I assume
A high pressure inert gas, argon or xenon, added to a reg-
they operate at 2700°C (where tungsten’s vapor pressure
ular filament lamp allows the wire’s temperature to be
is mid 10-5 torr) the handy-dandy applet shows the frac-
raised. The light gets noticeably whiter but, again, this
tion of total energy between 100nm and 400nm is ~0.2%.
type attracts no attention from vacuum designers.
I regard that as decidedly unspectacular UV performance.
For UV degassing applications, halogen lamps are not
1C. Halogen. lighters at all.
It’s obvious–to improve the light’s
whiteness, we must raise the fila- Discharge Lamps
ment’s temperature. But how do we
avoid the envelope’s coating getting 2A. Fluorescent
thicker quicker and the lamp start- If you take a mixture of mercury vapor
ing to flicker quicker before rapidly plus low pressure argon and put a
going permanent bye-byes? reasonable high voltage across it, a
Halogen
discharge forms and electrons are gen-
erated. Under the voltage’s influence,
Let’s digress. I’ll bet everyone has an opinion about the electrons accelerate until they slam
fluorides added to drinking water and toothpaste. Those into mercury atoms. This promotes Fluorescent
who do laundry might know Clorox 1 uses chlorine as a the mercury into an unstable, electronically excited state
bleach and we all know common salt is a chloride. The which, on returning to ground state, releases excess energy
more medically inclined will recognize iodine as a ‘tinc- as UV photons. Phosphors coating the fluorescent tube
ture’ or a component in ‘Betadine’ skin disinfectant. And absorb the UV and re-radiate (fluoresce) in the visible at
the medicinal uses of a bromide are legion (and probably a color determined by their chemistry. The manufacturer
©Kurt J. Lesker Company 2003 For more information, or to purchase vacuum products on-line, visit www.lesker.com. 3
 "...vacuum science is our business."
tailors the fluorescent’s ‘warmth’ or ‘coolness’ by choos- hazard indicators on power transmission towers; germi-
ing the right phosphor mix. cidal ‘cold’ sterilization; and forcing plant growth. Obvi-
The old idea of a T-12 standard 4 ft long, 1½! diam. 40W ously, for the last two, some UV emission is needed.
fluorescent tube has been blown out of the water. Nowa- So, xenon discharge lamps can be used in vacuum tech-
days, there’s a huge variety of shapes, wattages, and sizes nology for surface degassing. The trick is selecting a type
fitting the ‘Edison Screw’ socket. (Do UK’s bayonet fit- that fits these constraints:
tings have the same variety?) My house has three such 1. When operating, it must be vacuum compatible,
devices. I just hope they’re energy efficient because they which means (a) it doesn’t outgas; (b) it doesn’t
surely have no other desirable properties: the regularly explode; and (c) it doesn’t overheat.
shaped bulb in a table lamp has a horrible yellow hue; and
2. A high percentage of the power must be emitted
two ‘floods’ in the kitchen take two minutes to reach full
at useful UV wavelengths and not absorbed by
illumination.
the envelope.
From messing with photo-sensitive chemicals eons ago, I
3. Its wattage must be sufficient to “do the job.”
noted that normal fluorescent tubes emit tiny amounts of
UV. In Sadie’s Sans-Clothes Sunning Saloons, the lamp
tubes are made of materials that ‘up’ the UV ante. But, as 2D. Sodium
you know from your last visit to Sadie’s, these lamps gen- Most European, New Zealand,
erate little in the way of heat. So, they are not heaters and and South African readers have
the relatively low UV output (and huge size) of a tanning seen the ghostly (ghastly?)
bed lamp array means they’re useless as lighters too. effects of sodium vapor lamps
used for street and parking lot Sodium
2B. Neon illumination. Low pressure sodium vapor discharge lamps
are the most efficient at converting watts into lumens. But
You don’t have to visit
the light is an intense yellow with no UV and little IR.
Las Vegas to know exactly
what neon lights are. To High pressure sodium lamps appear ‘whiter’ but are
many, neon spells their Neon less efficient. For vacuum technology, however, both ver-
favorite German words: Busch, Schlitz, Pabts, and Open. sions are spectacularly uninteresting. But here’s a tip. LP
If you do visit Vegas, ignore the neon siren’s lure and sodium lamps make the red/pink/brown/black/yellow of
book seats at Cirque du Soleil. What about neon and our skin tones, look sickly at best. Stand under a portable
vacuum degassing? Forget it! Go book those seats. sodium lamp when your boss walks by. He’ll immediate
suspect hepatitis, an excess of ‘bili-bob rubin’ in your
blood (I’m from the south), and tell you to go home for a
2C. Xenon week. And it’s off to the Cirque du Soleil.
This designation covers a heap
of different constructions, gas
2E. Mercury
pressures, continuous or flash-
ing modes, and applications. In one of its many guises,
Basically, a high voltage mercury lamps are like
between two electrodes strikes Xenon fluorescent tubes with-
a discharge in high purity xenon at pressures from a few out phosphors. A little
hundred torr to a few atmospheres. The pressure affects argon, a little mercury, a
little electricity, all in a Mercury
the lamp’s spectral output, or ‘color,’ but just about any
pressure gives high intensity emissions at 235nm and silica envelope and, bingo, lots of UV. But manufacturers
lower. This is real UV output. Or at least it might be, have played games with pressure, additional gases, dis-
except these lamps often have UV opaque containers to charge geometries, etc, and have produced lamps with
reduce the risk of fried retinas or skin melanomas. phenomenal light (and UV) outputs. Go to a night-time
game at your local Australian-rules football stadium and
Applications are as varied as: camera flash guns and flood
you’ll see it perhaps better than in daylight, courtesy of
lights; warning flashers on some school buses; aircraft
high pressure mercury discharge lamps.
©Kurt J. Lesker Company 2003 For more information, or to purchase vacuum products on-line, visit www.lesker.com. 4
 "...vacuum science is our business."
If these lamps are so great, why mess with xenon or lights that blind’? Personally, I don’t get it. When I drive
sodium? Well, mercury lamps contain, ahem, mercury – at night, my focal point is my curb, near and far. I never
a well-known toxic element. Cleaning up after a lamp look at on-coming headlights, especially if they’re ‘flash-
explosion or even regular disposal when the lamp quits ing’ - the result of HID’s sharp lateral beam cut-off and a
becomes, at best, an ecological pain in the derrière. Using bumpy road. Oh, yeah. . . I also drive a car equipped with
a mercury lamp as a flash is also a tad tedious – the time to HID headlights but I insist that macht nichts.
reach full output from a cold start may be 10-15 minutes. HID lamps have generated enough light, and political heat,
Worse yet, hot re-starts may never happen! that European law permits them only on cars equipped
The technological cutting edge combines mercury and with self-leveling headlights. Consumers Reports is rec-
xenon at operating pressures of several atmospheres with ommending the NHTSA adopt this measure in the USA.
short arc geometries. With a UV transparent envelop these Do HID lights have any use in vacuum technology? I’d
goodies put out high intensity UV. The numbers for one have to say. . . hmmm, I dunno. Yes, they’re small. Yes,
particular mercury-xenon lamp are (percentages of total they last a long time. Yes, they can have high UV con-
energy): tent. But mostly the metal halides are added to improve
360-370nm 8.4% the visible emission characteristics, not to jazz up the
310-320nm 5.7% UV. Anyway, most HID lamps have UV opaque enve-
290-300nm 4.2% lopes. Literally blinding someone with your car head-
260-290nm ~6% lights comes under the heading of poor social skills.
Ignoring anything above 320nm, that’s almost 16%
appearing in the degassing photon range. Outgassing with UV
Again, the vacuum compatibility selection criteria given I am indebted to Tom von Alten for these results and his
for xenon lamps must be observed. But for this lamp type, permission to publish them. To avoid ‘seeming endorse-
the power issue is pretty much duck soup. Versions range ment’ issues, I’m not saying where Tom works but it’s big,
from 700W to 5000W. multi-national, and operates many vacuum processes.
Tom bought the lamp and power supply as a kit for
vacuum outgassing. The vendor has undergone recent
2F. HID
organizational changes and Tom’s not sure the kit is still
It’s possible to call just about any modern discharge offered. Both Tom and I have tried tracing the lamp, but
lamp a high intensity discharge (HID) device. However, its model number no longer appears on the lamp manu-
the term is reserved for mercury/inert gas lamps to facturer’s website. Given the rest of the product range,
which metal halides (indium, scandium, or thallium however, it’s probably a low or medium pressure mercury
iodides) have been added. Why add halides? Well, vapor discharge lamp.
to improve efficiency (see the table) but also to “fill out”
The graph shows the chamber pressure vs time for a
the spectral emission in the visible. These lamps give
system requiring frequent pumping from atmosphere to
a whiter shade of pale. (Any Procol Harum fans
‘high’ UHV. To introduce the results Tom wrote:
out there? No? Me neither!)
“We haven’t gone out of our way to do A/B comparison, as
we’re interested in results (getting the pressure down to use
the chamber as intended), rather than proving the reason-
ably obvious to ourselves.
“Attached are pumpdown curves from a run just before we
installed the UV bulb, and the first run with it. We ran it for
only an hour, and got half a decade better pressure in the
equivalent pumping time. (There were other differences as
you can see from offset in the first hour, so again, this isn’t
anything like UV lamp test data.)
HID (on the right)
“Further experience led us to run the bulb for 8 hours as
These lamps combine high intensity, high efficiency, small standard procedure, to get the chamber down to high UHV
size and have started popping up as – yes, even car head- pressure overnight.”
lights. Have you read reports condemning ‘blue head-
See Chart on Page 6
©Kurt J. Lesker Company 2003 For more information, or to purchase vacuum products on-line, visit www.lesker.com. 5
 "...vacuum science is our business."

Turbo pumping only

Turbo pumping only

UV lamp off
Ion pump opened

Ion pump opened

Pressure without UV 3x10-7 torr

Pressure with UV 7x10-8 torr

Final Thoughts on
UV Degassing WATCH OUT 2.
When it comes to your safety, I’m a nervous Nelly. UV radiation at <240nm has enough energy to dissociate
WATCH OUT 1. oxygen and form ozone. For sterilization, that’s just dandy
and there are Ozonizer lamps to do that. But ozone reacts
Attempting UV degassing without proper knowledge,
with polymeric materials, such as mucous membrane,
preparation, and protection is crazy! One quick glimpse of
flesh, and rubber gaskets, breaking the long molecular
a strong UV source may be your last glimpse of anything.
chains into short bits – how many sci-fi nuts remember the
Your eyes will be fried quicker than your microwave zaps
basic premise of The Andromeda Strain? Sound familiar?
that frozen TV dinner your spouse left (since you weren’t
home at dinner time). Don’t trifle with UV lamps! Make So, don’t flow oxygen (or air) through the chamber,
sure any viewports (yes, even “opaque” glass ones) are through the (turbo) pumps and into room air, or through
screened so the lamp cannot be accidentally observed. the chamber and into the (cryo) pumps when you’re using
And Tom advises “Wear UV protective glasses.” UV light to degas.

©Kurt J. Lesker Company 2003 For more information, or to purchase vacuum products on-line, visit www.lesker.com. 6
 "...vacuum science is our business."
WATCH OUT 3.
UV lamps are intended for operation in air at atmospheric
Kurt J. Lesker Company
pressure. If you intend exploring UV degassing tech- Office Locations
niques, let me suggest a few tests:
North American Canada Office
(a) Buy, or steal, a cheap vacuum oven. With the test lamp Headquarters Ph: 905.770.8450
mounted and suitable temperature measurement equip- Ph: 412.233.4200 Fax: 905.770.3723
ment in place, evacuate the oven and power up the lamp. Sales (Canada):
Fax: 412.233.4275
Leave it running for two or three times longer than you 1.800.465.2476
Sales (US):
expect an outgassing cycle to last in the real system.
1.800.245.1656 [email protected]
[email protected]
(b) Mount an RGA on the oven and carefully compare the
gas spectra before and during lamp operation. This will
Europe
give some feel for the lamp’s outgassing properties. West Coast Headquarters
Office Ph: +44 1424-719101
Ph: 925.449.0104 Fax: +44 1424-421160
(c) Mount a microscope slide, half covered with alumi-
Fax: 925.449.5227 [email protected]
num foil, near the lamp’s connector. After the test, look at
the dividing line between covered and uncovered surface Sales (US):
with a high powered light microscope. You’re looking 1.800.245.1656 Hungary Office
for visible differences indicating coating. If you’re lucky [email protected] Ph: +36 1 383-5322
enough to know a tame SEM operator, bend the alumi- Fax: +36 1 383-4369
num back on itself (so you have ‘coated’ and ‘uncoated’
[email protected]
surfaces adjacent) and have an energy dispersive x-rays
analysis done to identify the elements that evaporated
from the connector.
Comments and Suggestions
[email protected]
Good luck and remember, no peeking!
Subscription Requests
www.lesker.com/LeskerTech/subscribe.cfm

Lesker on the Road

Date Event Location Our Booth No.

Aug 19-20 Canadian Semiconductor Tech Conf. Ottawa, ON, Canada 7

Sept. 18 Pacific NW AVS Portland, OR

Oct 6-7 Advd Light Source Users’ Mtg Berkeley, CA

Oct 7 So CA AVS Anaheim, CA

Oct 16-17 Nanomaterials Crossroads Montreal, Canada

Nov 4-6 International AVS Baltimore, MD 529-535

Nov 19-21 OLEDs 2003 San Diego, CA

©Kurt J. Lesker Company 2003 For more information, or to purchase vacuum products on-line, visit www.lesker.com. 7