5975 filament deterioration symptoms?

[James_Ball]

One thing I have read on the SIS yttria filaments is that you need to lower the emission current a little because they produce more electrons at a given current. Not sure if that would cause the falloff with the first attempt or not.

I've not seen that. Their own SISweb Note 92: Yttria Coated Mass Spectrometer Filaments says to leave the Agilent default EIEnergy at 69.9 and that current and voltage will be correspondingly lower because it happens at about 500C lower temperatures while still producing 70eV. See figure #4.

I did have a source thermocouple fault on the second pump down (temperature was all over the place (target: 230C: 225,226,227,228,350, 450, 225, 228, 229, 280, etc...) I had the same fault during my May 2017 source cleaning and then it stabilized when I checked the mount. The instrument ran without problem right up till I burnt the filament. So I had bought a new thermocouple back in May and had it available and put it in right away when the fault happened again. I had wondered if temperature drift might be responsible but hard to see how it could vary enough without tripping a fault and I think I would have noticed changes while calibrating because I keep the source temperature visible on the Top App.

EI Energy will always be 69.9 but the Emission Current can be lowered because the filaments are more efficient at producing electrons. I think you have to manually lower the Emission Current as the autotune setting don't normally adjust that value. Best way to adjust the Emission Current is to run Repeller profiles and if the profile is not a nice bell curve shape, lower the current until it gives a better profile. If the profile from 219 and 69 just go up and up until the repeller maxes out, then the current is too high. I have had to drop emission current down as low as 20uA even on regular filaments, I believe default is 35uA.

[LALman]

...EI Energy will always be 69.9 but the Emission Current can be lowered because the filaments are more efficient at producing electrons. I think you have to manually lower the Emission Current as the autotune setting don't normally adjust that value. Best way to adjust the Emission Current is to run Repeller profiles and if the profile is not a nice bell curve shape, lower the current until it gives a better profile. If the profile from 219 and 69 just go up and up until the repeller maxes out, then the current is too high. I have had to drop emission current down as low as 20uA even on regular filaments, I believe default is 35uA.

With the standard filaments, It was always close or at the default maximum of 35 (as you say "up and up"). And, the peaks were barely rolling off when ramping repeller from 10 to 42. I let the Max Sensitivity Autotune set the Repeller. In this case Filament #1 is a standard filament and it came up with a repeller of 34.81 which just to the right the roll off at 36.

For the SisAlloy Filament #2, Autotune used the default Emission of 34.6 and EI Energy of 69.9; it chose a Repeller of 29.29 which is the center of the nice guassian peaks when ramping repeller from 10 to 42.

[tlahren]

Hi, I realize this is a 6 year old post but I just wanted to correct this for other people who stumble upon it.

The second filament is definitely no marketing gimmick -- it is actually critical to the function of the EI source. When you are using filament 1 the electrons emitted are reflected back by the 2nd filament. The electron can make several reflections back and forth in the ion volume before making a collision.

Thank you for the clarification. I guess I can understand the necessity for the reflection using the bottom filament or as I assumed previously a "collector plate" as found on instruments with only one filament. I was just told by an old Agilent Sales guy at one time that they decided to put two filaments in the final designs for the single quads because they could use it as a Marketing advantage by saying you can simply switch filaments without venting (I cannot personally verify this info). I was told that they never see the same performance out of filament #2. You're right, though, I was grossly inaccurate in referring to it as a "Gimmick." It is indeed more than that on these systems.

[aldehyde]

Hi, I realize this is a 6 year old post but I just wanted to correct this for other people who stumble upon it.

The second filament is definitely no marketing gimmick -- it is actually critical to the function of the EI source. When you are using filament 1 the electrons emitted are reflected back by the 2nd filament. The electron can make several reflections back and forth in the ion volume before making a collision.

Thank you for the clarification. I guess I can understand the necessity for the reflection using the bottom filament or as I assumed previously a "collector plate" as found on instruments with only one filament. I was just told by an old Agilent Sales guy at one time that they decided to put two filaments in the final designs for the single quads because they could use it as a Marketing advantage by saying you can simply switch filaments without venting (I cannot personally verify this info). I was told that they never see the same performance out of filament #2. You're right, though, I was grossly inaccurate in referring to it as a "Gimmick." It is indeed more than that on these systems.

I'm sure there is some degree of truth to the marketing bit.. When doing testing / troubleshooting / validation work I always tune both filaments. Generally the tune gain factor should be very similar on both filaments; tune gain factor is the detector amplification needed to get to 500k counts of PFTBA base peak. As long as the source and vacuum are OK, this value should be about 0.5-3.

If filament one breaks, it could definitely cause filament 2 to perform poorly. But if both filaments are functional they should work equally as well.

[gcmsguy123]

Around 2010, when Agilent was changing the EI filament part numbers multiple times, we had started to seen many problems with EI filaments. The problem was that at some point we would suddenly get a large number of low response data points for repeated standard measurements. At times the response would jump back up to where one would expect. It would randomly alternate like this, resulting in horrible precision, which could go one for weeks if not caught by the operator. Eventually the filament would die. In essence, instead of working well and then suddenly dying, the filaments would work well, then produce bad data for days to weeks, and then eventually die. Another problem was that the usable filament lifetime dropped by around 25%. All of my colleagues had similar problems.

It seems that SIS has an explanation for the problems:
https://www.sisweb.com/referenc/applnot ... arping.htm

We were advised to try Agilent CI filaments, which we did. Those worked great for us for a few years, even though we work in EI mode. Since CI filaments are straight and under tension, they have less of a tendency to sag/deform/warp than the curly EI filaments, which was apparently avoiding the problems that the EI filament had developed.

Unfortunately, around mid-2010, the performance of CI filaments also started to randomly deteriorated for us. Instead of regularly lasting up 6 or even 20 month, we suddenly had many CI filament failing in days or weeks or we could not properly tune the MSD with them, or they would just produce horrible precision from the start. Also, we noticed that we cannot predict anymore from one good (bad) filament whether other filaments from the same batch are good (bad). That used to work, not anymore. We also noticed that several CI filaments were not straight and those performed really bad.

We also tried the G7005-80000 filaments in 2021, but they produced artifacts for us. Since some of my colleagues recently reported good experience with them, so we may try them again. But they are even more expensive.

Overall, the experience is very frustrating and costly.

Has anyone found a solution?