Tungsten Filament Life Poll

[John Donovan]

I've been getting some questions from other lab managers asking what sort of filament life one should expect from a tungsten filament, so I've decided to start poll to answer the question a little more scientifically.

That said, the answer may not only depend on the state of one's gun vacuum, but also the brand of filament and maybe even whether one utilizes new or rebuilt filaments.  There could be other factors as well, for example: how often does one re-saturate one's filament.  I remember a Cameca engineer (Sam Pindrys) who complained that most of the labs he visited, adjusted the filament heat when a new filament was installed, but almost never re-saturated the filament after that (until it burned out anyway).  Obviously since the saturation points drops as the filament wire thins, this is an important routine that should be performed at least every few weeks in my book.

That said, when I went through our SX100 log book I don't see quite the life expectancy I had hoped to see.  But in the last few years we've had some significant vacuum issues that have only recently been resolved so I will revisit this topic later this year to see if our improved vacuum improves our filament life. Here are some recent data from our Cameca SX100 (we're still running on the one from last December):

12/02/2015    New filament    ~1.5 months     infant mortality!?
10/14/2015    New filament    ~3 months
07/30/2015    New filament    ~8 months
01/09/2015    New filament    ~2.5 months
10/28/2014    New filament    ~1 month        infant mortality!?
09/17/2014    New filament    ~4.5 months
05/06/2014    New filament    ~2.5 months
02/28/2014    New filament    ~4 months
11/05/2013    New filament   

I should also mention that we have generally bought rebuilt filament from Energy Beam Sciences, so I can't really say whether that is part of the issue or not.  I do seem to remember when I had a supply of Cameca built new filaments at UC Berkeley on my SX51, that the filaments lasted 6 to 8 months or more.  But I don't have that log book here at Oregon so that would be difficult to check since the probe tech there is gone.

So please feel free to vote in the above poll and even more importantly please provide details in a post on whether you routinely re-saturate your filaments and whether you buy OEM or 3rd party new/rebuilt filaments.

[AndrewLocock]

With regard to W-filament life, I only have data for JEOL filaments since mid-2013, and for EBS filaments for Cameca since early 2015.
 
   days        JEOL (OEM)         Cameca (EBS-VL)
Average                53                         25
Median                 42                          23
Std Dev                39                          15
Max                      168                        56
Min                       16                          8
# filaments            17                         13

I tend to turn down the filament every few days on the Cameca, and every few weeks on the JEOL.

Thanks,
Andrew

[Brian Joy]

This is the complete tungsten filament history so far for the JXA-8230 at Queen’s University in Ontario:



I order new filaments directly from JEOL (~$50CAD apiece).  Generally I check the saturation about twice a week.  It appears that I can count on a filament to last 6-9 months or 1800-2400 hours at saturation.  As noted above, I’ve removed filaments prematurely twice due to unstable beam current, which I attribute to defects in those filaments (possibly hairline fractures?).

Edit: In response to a question that I received, I set the gun bias at a relatively high value to produce relatively low emission current.  At accelerating potential = 15 kV, emission current from a new filament is generally around 16 microamps; it decreases during the lifetime of the filament to ~12 microamps just before it breaks.

[Probeman]

My only experience with JEOL filaments were on the JEOL 6400 SEM.

Are the EPMA filaments similar to that instrument? That is, a pre-centered filament in a ceramic base with a metal alignment ring on the outside?

[Brian Joy]

My only experience with JEOL filaments were on the JEOL 6400 SEM.

Are the EPMA filaments similar to that instrument? That is, a pre-centered filament in a ceramic base with a metal alignment ring on the outside?

Yes, the JEOL K-type filament used in the JXA-8230 is very nicely pre-centered in a ceramic base within a metal alignment ring on the outside.

[neko]

Kinda funny that this would come up, as I just started putting our filament statistics into a spreadsheet about 10 days ago.

I only did 2015-2016 so far, Cameca SX-100, with EBS AR-CAMO-R (the ones with hips) rebuild filaments, UC Davis lab.

F1 - late 2014 through April 29 2015 - 775 operational hours
F2 - May 1 - May 31 - 252 hours (not sure if this was a false-out from out failing +/-15v supply or if we blew the filament. fixed the supply in November)
F3 - June 1 - June 28 - 99 hours (probably same false-out - this was also the month we had a vacuum valve go out and spent 3 weeks getting Cameca re-instated as a vendor  )
F4 - June 29 - Aug 9 - 412 hours
F5 - Aug 10 - Feb 2016 - 844 hours and still strong. (Jan 2017 update - F5's beam stability graph has become noisy so I replaced it - final hours: 1535 - here's hoping the new one lasts until 2019!)

Some of our filaments never seemed to want to saturate well - the heat was cranked up to 250 and they still looked 'ugly' as far as the gun scan image. We tried adjusting the height & etc but they wouldn't cooperate. The F4 filament was one of those. Beam was stable on the tests though so we crossed our fingers. We had one old VL-type (the one without the hips?) that ran for about 2500 hours before we took it out. [Just checked, 2425 hours total, including two weeks where it ran 539 hours non-stop. I'll have to go through the logs to find out if it was just left on by accident or if that was that one time we did over 2000 points per week on MORBs]

I'll make a better list when I get around to adding up the previous years. I have statistics going back to the 2002 installation!

We check the saturation every few days, generally run a few heat values under full saturation. Our latest filament has been one of the best behaved yet, the alignment's very nice and hasn't moved much, saturates around 220 (measured 234) and, well, we'll see how long it lasts!

[Jeremy Wykes]

I have a question about Cameca rebuild filaments. The filaments we would get at ANU via Graham H would be a mixture of "cross-welded" where the tungsten wire would be welded to the flats on opposite sides of the post for each side, or "same-side" welded, where the wire was welded to the flats on the same side of the post on each post. The "same-side" welded filaments are a nightmare. There is frequently insufficient travel in the Cameca wehnelt to get the "same-side" filaments centered.

Do others suffer from this problem with rebuild filaments?

[Probeman]

I have a question about Cameca rebuild filaments. The filaments we would get at ANU via Graham H would be a mixture of "cross-welded" where the tungsten wire would be welded to the flats on opposite sides of the post for each side, or "same-side" welded, where the wire was welded to the flats on the same side of the post on each post. The "same-side" welded filaments are a nightmare. There is frequently insufficient travel in the Cameca wehnelt to get the "same-side" filaments centered.

Do others suffer from this problem with rebuild filaments?

Hi Jeremy,
Our EBS (Energy Beam Sciences) rebuilt filaments are welded on the "same-side" but we usually have no problem centering them.  That's EBS part# AR-CAMO-R.  One thing that might help is that the filament supports are ground on their sides so they are quite thin where the filament is welded.

The problem we do see is possibly a mechanical stress issue where they are aligned mechanically to begin with, but after a few warming/cooling cycles they sometimes "bend over" and are no longer mechanically aligned as well as they had been.  I suspect that is one reason for "cross welding" the filaments as you mentioned.

If this "bending over" happens (I'd guess it occurs about 20% of the time for us), we usually have to remove the Whenelt assembly and mechanically re-align the filament a second time. That is, the offset is usually too large for the electronic adjustment range.

I wonder if an "annealing step" would prevent this?   But you'd need a good vacuum chamber on your bench with a suitable heating circuit, so probably just as well to take one's chances in the instrument.
john

[knshugart]

We have a box of old/dead Cameca filaments that a previous staff member was collecting because he knew a company that refurbished them. However, he left a while back and never passed the name of the company on.

Does anyone know of a company that will refurbish old tips? I feel it has to be more cost effective than buying new ones from Cameca.

And to answer the poll question, our W tips on the Cameca SX-100 generally last about 200 hours of active use. (Hard to say in days/months as we have long periods of inactivity on that equipment during which the filament is turned off.)

[Probeman]

We have a box of old/dead Cameca filaments that a previous staff member was collecting because he knew a company that refurbished them. However, he left a while back and never passed the name of the company on.

Does anyone know of a company that will refurbish old tips? I feel it has to be more cost effective than buying new ones from Cameca.

And to answer the poll question, our W tips on the Cameca SX-100 generally last about 200 hours of active use. (Hard to say in days/months as we have long periods of inactivity on that equipment during which the filament is turned off.)

See Energy Beam Sciences :

http://www.ebsciences.com/

This post (above) mentions the part # we use for our SX100:

http://probesoftware.com/smf/index.php?topic=668.msg4308#msg4308

[gmorgan@ou.edu]

We've been using EBS rebuilts for 20+ years. They are same-side, and I never had trouble with alignment on our SX50 but the Wehnelt in our SX100 doesn't have enough adjustment to align them well if at all. However, I find that I can lightly bend the wire to center the tip in the Wehnelt aperture, and then the trick is to warm it up VERY slowly (up to 100 cA can be fairly quick, say 25 cA steps over an hour, but then maybe 5 cA every 10-20 minutes to the 200+ cA range) and they seem to anneal in place. At least I'm 2 for 2 since trying it.

On the SX100 I miss the ability (or instruction) to write a local task like I had on the old SX50.  I'd just use a task to warm a new filament up overnight (1 cA every 5 minutes  for17-18 hrs).
For that matter, we had a very nice task to saturate the filament as well (increased the heat in 1 cA steps and plotted the emission curve)....

[Julien]

I have some detailed info on my W-filament operational hours on the 8600, but to be brief, in general, I can easily pull at least 1500 hours on a filament, and we always did quite a good amount of work (maps) at high current and high voltage (20-25 keV not unusual). I have been using EBS refurbished filament and JEOL filament (maybe for 2 years each). Although my 4-year period is statistically short, I had more "unluck" with refurbish EBS filament: some were not perfectly aligned and needed some adjustment, whereas other were behaving abnormally and usually being out after just a couple 100 hours or less. I should say that these refurbished filaments were prepared quite a long time ago (> 6-8 years?) and despite being in a dessicator all the time, they might have suffer from this long shelf time?

Julien

P.S. About the poll: the right question to ask is the operational hours, not the number of months. Also, all this might depend on the vacuum quality, the selection of the optimum saturation / gun bias, and maybe the column setting used (high voltage / current?), how the filament is heated (slowly / quickly) and what its standby value is... Some lab works 7/24 whereas other only have there beam on ~40 hours per week...

[John Donovan]

P.S. About the poll: the right question to ask is the operational hours, not the number of months. Also, all this might depend on the vacuum quality, the selection of the optimum saturation / gun bias, and maybe the column setting used (high voltage / current?), how the filament is heated (slowly / quickly) and what its standby value is... Some lab works 7/24 whereas other only have there beam on ~40 hours per week...

Hi Julien,
You are absolutely right, I am making a big assumption that most labs run roughly the same number of hours per month...  probably not a great assumption.  But at least it's an easy to measure number, usually by just glancing at the log book!
john

[neko]

On the SX100 I miss the ability (or instruction) to write a local task like I had on the old SX50.  I'd just use a task to warm a new filament up overnight (1 cA every 5 minutes  for17-18 hrs).
For that matter, we had a very nice task to saturate the filament as well (increased the heat in 1 cA steps and plotted the emission curve)....

You can fake that with scripts in the interpretor - I attached two since we manually warm up our filaments the first time over several hours, might as well automate it if I can.

warm_up_filament4.txt - warms up in increments of 5 cA once every 5 minutes until it hits 200. (~4 hours)
warm_up_filament24.txt - warms up in increments of 1cA once every 5 minutes until it hits 200. (~16 hours)

to run the script, place the text files in the sxpc/inter directory, open an interpretor window and type @warm_up_filament4.txt and it will start running (at least in Peak Sight 4.0). Nothing so fancy as automated saturation based on emission curves however.