7-29-98

desirability of venting SEM chambers

with either air (passed through a filter and

desiccant column) or bottled nitrogen?

In the case of LEO SEMs with LaB6 filaments

it strikes me as being an unnecessary expense to

vent with bottled nitrogen as the column region is

seldom vented. Also venting with bottled nitrogen

adds complexity as precautions must be taken not

to over pressurize the chamber and damage the

EDS window.

I would appreciate comments for my guidance.

Trevor Sewell

sewell@uctvms.uct.ac.za

We have found that boil off from liquid nitrogen is by far the best

venting material. It is dry, it is clean, it is cheal.We hane the whole

lab' plumbed with a line going back to a 100litre Dewer which if I

remember lasts for 3-4 weeks. You obviosly have to be careful with

pressure. There is nothing like getting a 100lb SEM stage whislting out

of the microscope and landing on your lap. Really brtings tears to your

eyes.

Patrick Echlin

Multi-Imaging Centre

University of Cambridge.

pe13@cus.cam.ac.uk

maintenance. I would be inclined to use nitrogen; it's

drier and cleaner. The cost is mostly in the rental of the

cylinder and not the gas. I used a full N2 cylinder to run

some pneumatic EM valves and instrument backfilling. When

the high pressure cylinder ran low, it was switched to

serve for gas agitation during film development. The low

pressure cylinder was always available to exchange for a

full cylinder when required without interrupting EM

operations.

Over-pressurising is avoided by not relying on a timer to

backfill the column but the use of a demand valve (from a

diver shop) this type of valve delivers gas only while the

diver, or the microscope "sucks".

A "T" junction with a needle outlet near the cylinder could

also be useful to dust-off particles from microscope parts

etc.

Jim Darley

ProSciTech Microscopy

PLUS

PO Box 111, Thuringowa QLD 4817 Australia

Phone +61 7 4774 0370 Fax: +61 7 4789 2313

jim@proscitech.com.au

This is a great topic for discussion as there are quite a few factors that come into play.

Basically the main criteria for any vacuum system is to keep the atmosphere in it

constant and clean. By venting and opening the door of the chamber in a SEM to change

samples obviously throws that criteria out the window.

As is always the case with microscopy we have to make compromises. The best solution

would be to fit an airlock to the chamber. In this way you would maintain a better

vacuum, however then you are limited to sample size and it's expensive to buy an

airlock.

The second option is to back fill with dry nitrogen gas. Nitrogen being the gas that is

most prevalent in the air that we breath and that vacuum manufactures have therefore

designed their vacuum pumps and gauges around.

By back filling with dry nitrogen you are maintaining a fairly constant atmosphere to

pump after venting and changing the sample. Maintaining the positive pressure in the

chamber, as you open the door, keeps the surrounding atmosphere out. Yes, the negative

side to this is the danger of putting to much pressure on the window of the ED detector

and therefore blowing it. This can be prevented by simply removing the latch that hold

the door closed.

The surface area of the door is much larger than the area of the window, so the door will

be forced open far before the pressure on the window of the ED detector suffers. In fact

on normal W filament systems where you vent the entire column, I have seen the gun

popping up and down due to dry nitrogen backfilling, when the door latch is still on. No

damage to the ED detector.

As mentioned the vacuum components are also suited to Nitrogen. Pirrani and penning

gauges are calibrated on Nitrogen pressures, so by using nitrogen you have a better

chance of attaining the same readings each time you pump again. The Turbo molecular

pump, in particular, prefers to pump nitrogen and is thus more efficient. Water vapour is

particularly difficult for a turbo pump to pump. So if your EM is situated near the coast

or you have a lot of heavy breathers using the EM, the humidity in the lab will change

on an hourly, daily and seasonal basis. On a bad day, with high humidity, the pumping

speed of the chamber to reach VAC OK status, could decrease in terms of 5 to 10

minutes from a good dry day.

Thus in conclusion ( yes I eventually got there), to backfill with dry nitrogen improves

pump down speed, cleanliness of the system, stability and repeatability of the required

vacuum and prolonged life of the vacuum pumps. This will be very helpful if you

require high resolution work or very accurate ED analysis from your system on a regular

basis. There is also no need to purchase an air lock for your system.

On the down side, as there always is in microscopy, there is the cost of the nitrogen,

ensuring the door latch is not left on to prevent ED window damage ( however I am not

aware of anyone where this has happened. ), changing samples in a hurry so as not to

waste too much Nitrogen and the hassle of making sure you have a supply of Nitrogen

handy all the time.

In my opinion, a SEM that is totally vented each time you need to change samples, i.e.

vent the column , gun and chamber, and is situated near the coast or any other high

humidity area, should definitely use dry nitrogen backfilling.

If you are simply venting the chamber but wish to do high resolution work or stable ED

analysis, I would still look at dry nitrogen backfilling.

If you are analysing or looking at really foul samples at low mag any way, I would not

bother.

That's it from me. I hope it helps and good luck.

Luc Harmsen

Anaspec, South Africa

International technical support on microscopy.

Tel: +27 (0) 11 476 3455

Fax:+27 (0) 11 476 7290

anaspec@icon.co.za

speed pumpdown somewhat, likely by minimizing water entry into the

vacuum system. It is of no help, of course, when I get the

ocassional specimen which outgasses significantly.

I have bottled nitrogen in the lab, but instead use the gas

take-off provided on the LN2 cryo which is also in the lab.

Over pressure is not a problem so long as I don't hold the chamber

door shut when venting....

Woody White - McDermott Technology, Inc.

Woody.N.White@mcdermott.com

atmospheric pressure. I'm from the better-safe-than-sorry school and

prefer to keep my chamber as clean as possible. As far as

overpressurization, it hasn't been a problem with our Amray scopes

because the stage/door is not locked to the chamber. At atmospheric

pressure, there is sufficient leakage past the O-ring seal to prevent

damage to our thin window EDS. I can only recall one incident in the

past 11 years (more than 100 microscope-years) when N2 broke a window.

That SEM was used by many people, was not well maintained, had

undiagnosed problems with the N2 purge (more flow! decrease cycle time!)

and a door that latched to the chamber.

My 2 cents

Harold J. Crossman

OSRAM SYLVANIA INC.

Lighting Research Center

71 Cherry Hill Dr.

Beverly, MA 01915

(978) 750-1717

crossman@osi.sylvania.com

How often are you venting the chamber ? We have found on our LEO 440 that

venting with nitrogen takes a long time (up to 30 minutes !!) (Maybe we're

doing something wrong, Luc?). Although the pump down time decreases, you're

probably going to lose more time venting than you'll gain with the decreased

pump down time if you're changing samples often. Venting with nitrogen will

keep the SEM cleaner, though, if time is not a problem.

Charles Bushell

CharlesB@mintek.co.za

mentioned here is ballistic damage -- particulates getting blasted off

the sample and punching a pinhole in the window. We see this problem

probably more often than implosion failures. Be aware of the geometry

of the venting flow pattern relative to the detector if you're considering

a customized venting system and analyze powders or other particulates.

Regards,

Rick Mott

rick@pgt.com

We vent all columns with nitrogen that is syphoned off a liquid nitrogen storage tank.

This nitrogen gas is also run through a tube with a drying agent to make sure it doesn't

pick up moisture from the lines.

We have scuba valves on the lines that will close as soon as you reach atmospheric

pressure. This eliminates the problem of accidentally pressurizing the microscope

column.

Debby Sherman, Manager Phone: 765-494-6666

Microscopy Center in Agriculture FAX: 765-494-5896

Dept. of Botany & Plant Pathology E-mail: sherman@aux.btny.purdue.edu

Purdue University or: emcenter@btny.purdue.edu

1057 Whistler Building

West Lafayette, IN 47907-1057

down. It shouldn't take any longer to vent with N2 than with air.

Check the you still have gas, that you don't have any kinks in your

supply hose, and that your valve is open. As a check, take the hose off

at the inlet to your vacuum system. Moisten your lips and have the gas

flow over your lips. You should just feel a slight cooling sensation.

That is a good flow rate for venting a system.

Walck. Scott D

walck@ppg.com

used to prevent the incursion of water vapor into the chamber since

most vacuum systems have a hard time dealing with it.

The use of a desiccant column is well advised for any instrument. In

the long run however, there are two problems. The first is the

regeneration of the desiccant. You have to pay attention and bake

the water out of it at regular intervals. The second is a long term

accumulation of desiccant dust. Over time, the desiccants appear to

produce a very fine dust that will eventually clog the chamber vent

valve, resulting in a service call.

Nitrogen systems can produce their own problems. The need to have a

constant source of nitrogen can be eased by using the vapor from a

large storage dewar if liquid nitrogen is always available in the lab

due to EDS or other cryogenic systems. Generally, the nitrogen is

cheap but the demmurage charges for the dewar are not. If you use a

little more nitrogen per month, the cost is minimal if you already

have the need for the large dewar.

The savings is basically in time. The primary volume in an SEM

vacuum system is in the sample chamber. It is not system

cleanliness that is in question here, it is basically the time

involved in pumping the chamber to required levels. If one can

exclude water vapor from entering the sample chamber during sample

changes, there will be an appreciable improvement in pump down times.

A well airconditioned room is a start, desiccants next and a

nitrogen backfill is best. Airconditioning is a necessary first

step, since a warm and humid environment can also cause other

problems such as the condensation of water on the exterior of water

cooling lines used for diffusion pump and electronics cooling.

Most system contamination comes not from gases not pumped out, but

from air leaks and pump oils. By the time a system reaches

operational vacuum, the overwhelming majority of environmental gases,

including water vapor, will be gone.

Allen R. Sampson

Advanced Research Systems

317 North 4th. Street

St. Charles, IL 60174

PH 630.513.7093 FAX 630.513.7092 Email: ars@mcs.net

WWW: http://www.mcs.net/~ars

Analytical instrument maintenance services

I have been responsible for a number of demonstration laboratories in the

past and I have always insisted upon using a dry gas rather than dirty old

(English) wet air.

Why?

1. Pump down time is improved considerably to ~50% in some cases, but it

takes months for this to happen.

2. Column cleanlyness improves so if you work at low kV (and you all

should try it) you will get a longer life from your column and its

apertures between cleaning.

3. Contamination is the biggest killer of high resolution microscopy and

this is also dramatically reduced.

Regards

Steve Chapman

Senior Consultant E.M.

Protrain, 16 Hedgerley, Chinnor, Oxford OX9 4TN, England.

Tel & Fax 44 (0)1844 353161

Web Site - http://ourworld.compuserve.com/homepages/protrain

outside (where the plumbing is required) or the large upright (5 feet

high - 2.5 feet diameter). Both of these have valves to take off the

vapor above the liquid. This pressure can be quite high and they also

have pressure relief valves and safety blowout plugs.

You are correct about what I was saying. I used a small dewar that is

used for absorption pumps which are basically styrofoam. You can use

the styrofoam that is used to protect acid bottles when they are shipped

or your dewar that you use to fill the EDS detectors. (If your are in a

bind, put a tube into the EDS detector dewars.) If you have an open

container, just use Tygon. If you have a narrow neck, use a copper or

stainless steel tube to go into the tank so that it won't go hard if it

is bent and attach the Tygon tube onto the warm end of the tube. Now

you have to clear the line of the air in it just like any other line

that you attach to a vacuum system. Just after you put the tube into

the LN2, there will be blow out at the other end of the tube until the

temperature equilibrates. Position that end over the inlet to the

vacuum system to flush out the inlet also. While the gas is flowing,

put the tube on the inlet and that will give you a fairly clean line.

It doesn't matter what diameter(s) that you use, just length. About 8-10

feet is good enough. If you use a stainless steel tube to insert in the

LN2, it will blow gas a little longer than if you use copper because of

its lower thermal conductivity.

-Scott

From: Luc Harmsen

To: 'Walck. Scott D.'

Subject: RE: Venting SEM chambers with bottled nitrogen or air

Date: Thursday, July 30, 1998 1:39AM

Hi Scott,

Thanks for the info on the N2 boil off. However I do not have a clear

understanding of how people tap off the nitrogen from the dewars.

If I understand you correctly, you simply put a pipe from the vent valve

directly into the liquid nitrogen dewar standing a few meters away from

the SEM. This pipe is then long enough to ensure that as you vent the

liquid nitrogen will be sucked up into the pipe and will turn into gas

before it reaches the SEM due to the length of the pipe ?

Would you be putting a thin pipe into the liquid nitrogen and then a

larger diameter pipe after that to the vent valve to ensure the N2 does

turn into a gaseous form ?

Thanks

Luc Harmsen

Anaspec, South Africa

International technical support on microscopy.

Tel: +27 (0) 11 476 3455

Fax:+27 (0) 11 476 7290

anaspec@icon.co.za

system is water (90 - 97% on an RGA). If your lab gets humid and you

have a dry nitrogen backfill, you can partially pump several times to

remove most of the water and still be ahead of a single pump-down. A

dessicator on your vent line will behave in a simalar manner, but won't

work as well a dry nitrogen under 1 - 2 psi gauge pressure. Also, 1 - 2

psi should not hurt a light element window.

Ken Converse

owner

Quality Images

third party SEM service

Delta, PA

qualityimages@netrax.net

if you're putting it into LN2. It does not become brittle or crack. It

is available from Cole-Parmer (and almost certainly from other vendors as

well). I have no commercial interest in tubing of any sort.

Yours,

Bill Tivol

tivol@wadsworth.org

purpose of venting an EM may be most convenient for some people. A metal

tube would be inserted into the dewar's liq. N2 and a Tygon tube could

connect it to the EM's inlet valve.

I recommend stainless tubing, not copper. Copper would lose too much N2

from the dewar, particularly if it's forgotten in place. Not only is copper

a good thermal conductor but this property is retained down to liq N2

temperatures. Most SS is a mediocre thermal conductor and becomes much worse

(which for our purposes is better) at low temperatures.

The second point is that open (not pressurised) dewars absorb O2 and

this becomes liquid O2 and is mixed in with the liq. nitrogen. No problem

when the dewar has just been refilled and it's true, liquid air is very dry.

BUT a dewar that was filled 2 or three weeks ago will contain not only a

high percentage of oxygen, but also a lot of small, suspended ice particles.

Many readers would be familiar with the "boiling ice syndrome" which can

cause vibrations and "noise" in EDS systems. After a year of operation an

EDS dewar would contain more water per litre of gas (800 litres of gas in a

litre of liq N2) than is ever found in air including during a tropical

downpour. Moisture problem can occur when using an open dewar in some

instances.

Dry oxygen maybe a lesser consideration, however, when the filament chamber

is to be vented, it's even more important to wait until the filament has

cooled. Exposure of a hot filament or LaB6 cathode to N2 is not a good idea,

exposure to O2 is worse.

Cheers

Jim Darley

ProSciTech Microscopy PLUS

PO Box 111, Thuringowa QLD 4817 Australia

Phone +61 7 4774 0370 Fax: +61 7 4789 2313

jim@proscitech.com.au

scattering effect compared to N2). Should we be concerned about any reaction

between the hot filament and the He? I would presume none. How about thermal

shock effects? Anybody care to comment?

Warren Straszheim

<wesaia@iastate.edu

I'm not sure what 'scattering' effects you believe helium will

improve. If you are referring to electron scattering during

operation, helium is probably harder for normal vacuum systems to

pump out then nitrogen, so the gain you are seeking more than likely

isn't there. I know of no practical difference between venting with

any dry gas.

Conductive heat dissipation is always much faster than

convective. In the two or three seconds it takes you switch the

accelerating voltage off and to hit the vent button, the majority of

the heat built up in the small mass of a tungsten filament will

probably be conducted off through the filament posts and electrical

contacts in the gun. As the backfill gas works its way into the

gun, the remainder of the heat will be removed rather slowly since

the leak provided by a normal vent valve is rather small, taking 30

seconds or more to completely vent.

I have never seen or heard of anything other than very loose

anecdotal evidence for increases in cathode life dependent on venting

methods. At the most, I'll accept customers turning the filament

drive down before turning off the accelerating voltage and venting

the instrument. Anything more than that is simply unnecessary. It

is far more important to be cognizant of small vacumm leaks,

particularily in areas that could bleed into the gun area. Air leaks

in the gun during operation are extremely damaging.

> At 02:52 PM 7/31/98 +1000, you wrote:

> >Dry oxygen maybe a lesser consideration, however, when the filament

> >chamber is to be vented, it's even more important to wait until the

> >filament has cooled. Exposure of a hot filament or LaB6 cathode to

> >N2 is not a good idea, exposure to O2 is worse. Cheers Jim Darley

> >ProSciTech Microscopy PLUS

>

> Hmmm. We vent our Hitachi with W filament with He (because of less

> scattering effect compared to N2). Should we be concerned about any

> reaction between the hot filament and the He? I would presume none.

> How about thermal shock effects? Anybody care to comment?

Allen R. Sampson

Advanced Research Systems

317 North 4th. Street

St. Charles, IL 60174

PH 630.513.7093 FAX 630.513.7092 Email: ars@mcs.net

WWW: http://www.mcs.net/~ars

I should have mentioned that our Hitachi is a 2460N low vacuum SEM. During

operation we He instead of air or N2 in order to reduce scattering effects

at the 40 Pa of atmosphere that we maintain in order to cancel charging on

our specimens. There was no intent of trying to reduce scatter by residual

gas at 10-5 torr levels since He might have displaced heavier molecules.

I was intrigued that a side benefit of our He use might be increased

filament life since the hot filament would not be exposed to N2 with its

detrimental effects. A coworker likes to wait until the filament cools for a

bit before venting the chamber (the gun chamber always gets vented with the

specimen chamber). However, I had been supposing that thermal issues might

be insignificant as you appear to corroborate. Since we just had a W

filament go 441 hours even with sample changes about every 2 hours, I

suppose that the Helium might be helping us some. However the benefits are

probably not worth someone switching over from N2 to He unless they too have

a "leaky" SEM.

Warren Straszheim

wesaia@iastate.edu

I would be curious if anyone has any real data on filament life as affected

by a cooling off period before chamber venting. Like many, I was trained

to always cool a filament for several minutes before admitting air, but if

it's really not necessary I'd be happy to stop. Sure would speed specimen

turn-around times when we're running a bunch through.

Another question: do you really find it necessary to use 40 pa to eliminate

charging on your 2460? Using both a 2460 and a 3200, I have almost always

been able to eliminate charging with pressures of 1-5 pa, which should give

substantially less scattering than 40 pa. Just wondering.

Finally---filament life of more than 400 hours is amazing!! And under

variable pressure with frequent specimen changes! I'd love to know the

secret to that trick. Maybe He is the wonder gas we've all been waiting

for....

Regards,

Randy

Randy Tindall

Electron Microscope Laboratory

Box 3EML

New Mexico State University

Las Cruces, NM 88003

rtindell@nmsu (work)

nrtindall@zianet.com (home)

at elevated temperatures is well known. I believe most of us who don't

wait for a cool down period are venting with dry nitrogen or some other

inert gas.

cheerios, shAf

<>/\<\/>/\<\/>/\<\/>/\ cogito, ergo zZOooOM /\<\/>/\<\/>/\<\/>/\<>

Michael Shaffer, R.A. - ICQ 210524

Geological Science's Electron Probe Facility - University of Oregon

mshaf@darkwing.uoregon.edu - http://darkwing.uoregon.edu/~mshaf/

>I would be curious if anyone has any real data on filament life as affected

>by a cooling off period before chamber venting. Like many, I was trained

>to always cool a filament for several minutes before admitting air, but if

>it's really not necessary I'd be happy to stop. Sure would speed specimen

>turn-around times when we're running a bunch through.

One operator likes to cool the filament for a while to get longer life. But

this filament has had a number of quick cool downs anyway. If it is a

reaction with nitrogen, then maybe we can vent the He to it right away. Then

again, we ran the scope in automatic mode with a beam shut off at the end of

the run. It might have had plenty of time to cool for most exchanges.

>Another question: do you really find it necessary to use 40 pa to eliminate

>charging on your 2460? Using both a 2460 and a 3200, I have almost always

>been able to eliminate charging with pressures of 1-5 pa, which should give

>substantially less scattering than 40 pa. Just wondering.

We have tried reducing the pressure and have run into charging. But we run a

lot of beam current to our concrete samples, around 1-2 nA by my guess.

Maybe the lower pressures work for lower beam current.

>Finally---filament life of more than 400 hours is amazing!! And under

>variable pressure with frequent specimen changes! I'd love to know the

>secret to that trick. Maybe He is the wonder gas we've all been waiting

>for....

I just hope that we can replicate the performance. Our normal life has been

in the 100 hour range.

Warren

As you all know the cathode assembly gets up to a pretty high temperature

too! If we vent with "dirty" air we cause contamination to deposit on the

hot cathode. I agree the filament cools very quickly but we all know the

cathode does not?

Waiting a few minutes after getting the READY indication before switching

on the filament and a few minutes after switching off the filament, prior

to letting in AIR, the result a difference in filament life and cathode

contamination.

The first point is that the READY indication means the vacuum is only just

good enough to use. Wait a few more minutes and the improvement in vacuum

will help give a better filament life and keep the column clean longer.

Put a beam down a column with a poor vacuum and you put up the

contamination! From the other direction allowing the cathode to cool prior

to letting in air will keep this cleaner too!

Just the same when you clean a column, it makes sense to clean the column

on a Friday afternoon, pump and wait over the weekend before putting a

beam down the column. Clean a column and then pass a beam through it

means that most of the vapours from the cleaning media have just deposited

on your clean components!

Steve Chapman

Senior Consultant E.M.

Protrain for courses and consultancy in electron microscopy world wide

Tel & Fax 44 (0)1844 353161

WWW http://ourworld.compuserve.com/homepages/protrain

PROTRAIN@CompuServe.COM

of any inert gas.

In the 1950th (no, I was not in labs then), a paper was

published The life of filaments. My long-term memory recall

advises that short of physical violence, filament life is

determined by vacuum, if it's under 5x10-5 torr and by

emission if it's above 15microamps.

It is assumed that throughout it's life the filament is

correctly saturated. I recall that there was a curve (or

maybe just a number of data points) showing filament

"death" at various vacua and emissions. In modern

instruments, emission is most frequently the limiting

factor. Reaching a filament life of 500 hours is no trouble

in microprobes at say 10 to 15 microamps, it's a little

less for a TEM, which commonly run at 30 microamps. Tell me

when a SEM operating at 80 or more microamps reaches 500

hours filament time; then it will be time to rewrite that

1950ish paper.

Cheers

Jim Darley

ProSciTech Microscopy

PLUS

PO Box 111, Thuringowa QLD 4817 Australia

Phone +61 7 4774 0370 Fax: +61 7 4789 2313

different gas species goes as the molecular weight of the molecule. As

a result, the pump speed for He compared to N2 are much lower. I could

look it up, but I think that it is about an order of magnitude

different. Another point is that, because the sensitivity of ion gauges

is about 0.25 for He, the gauge pressure and the true pressure will be

different. With a sensitivity of 0.25 for an ion gauge, the true

pressure is four times the value displayed. I'm not sure what the

values are for Penning gauges, but they are probably in the same ball

park.

If your vacuum system has ion pumps, DO NOT backfill with an inert gas,

especially He! Even with the special ion pumps that bury the inert

gases into the plates, they still "burp" the inert gas. Helium pump

speeds are extremely low.

One interesting aspect about your use of He around a hot filament is

that it does have a higher thermal conductivity than N2, so I guess that

it would cool the filament faster.

-Scott

Scott D. Walck, Ph.D.

PPG Industries, Inc.

Guys Run Rd. (packages)

P.O. Box 11472 (letters)

Pittsburgh, PA 15238-0472

Walck@PPG.com

(412) 820-8651 (office)

(412) 820-8161 (fax)