Ion pump "spiking" and beam stability issue

[Julien]

Hi everybody,

Three months ago we had our new JEOL-8230 installed. Whereas all tests passed with W-filament, we run into trouble when we hooked up and tested the ion pump, which is required for the LaB6 cathode. We have NOT yet mounted the LaB6 cathode, and all the following tests are done with a regular W-filament. Beam stability WITHOUT the ion pump is excellent and well beyond JEOL’s performances. The vacuum is acceptable according to JEOL:
- Sample chamber: 1E-4 Pa
- Ion pump: 5E-5 Pa

Symptom: The ion pump voltage is “spiking” (or burping as the engineer said), shifting from maybe 5.6-5.7 V at the high vacuum condition to ~5.8-5.9 V (with a rapid increase followed by a drop and a recovery period). In parallel the beam current is also unstable. These spikes occur once every hour, sometime once every 10 minutes or so. Not all the spikes in the ion pump appear to cause a disturbance in the beam current (??). Some time we see a large positive positive spike, sometime a small negative spike in the beam current, which could be due to our sampling period (1 second interval). It is not so clear, but it seems the spiking of the ion pump are occurring more and more commonly when it reach its steady-state (~1-2 hours after start-up). It is possible the beam is also having a strong positive spike followed by a small negative spike and a short recovery period, but we sample only one point over this process...

Here is a typical example of what is going on... (ion pump voltage in blue, beam current in red):



This is now 2 months the engineers are trying all their possible to solve this issue. It seems clear to us the source of the problem is the ion pump as we don’t see any problem when the pump is turned OFF. We have tried the following - without any success (spikes are always visible):
- Replacing the pump (this is now the 2nd pump showing the same behavior - JEOL wants now to try with another one but they are currently out-of-stock)
- Hi-potting the pump at 5 to 8 kV (if I remember the value used by the JEOL engineer). The tests after the first hi-potting  on the first pump appear to have improved the spiking behavior a little bit (from one spike every 5-10 min to one spike every 30 min). The 2nd and 3rd hi-potting improved it a little further (maybe down to “only” one spike every 2 hours), but the last and forth hi-potting have not solved the problem further… That’s when they decided to install the SECOND NEW ion pump. They also did some hi-potting on the 2nd pump - slight improvement, but the problem is not resolved.
- Turning OFF the HV (so no beam current) to see if there are any spikes in beam current showing up (grounding problem from the ion pump); this test show absolutely no current reading, it remained at 0 nA, but the ion pump was still showing spikes.
- The spikes in the beam current are visible at all scales, from 10 ^ -12 to 10 ^ -7 amp range. The spikes appear proportional to the current setting (i.e., spikes up to 10 ^ -11 when current is at -12 range, and spikes up to -6 range when setting is at -7).
- Switching the power supply for the ion pump using the secondary power supply that is normally used for the second ion pump (used on FEG only)
- Warming up the ion pump to ~160 °F (not too hot to avoid issue with O-rings / valves).
- Adding a direct ground from the ion pump to the main instrument ground.
- Cutting on the P-10 gas supply for over 24 hours. We were testing if the pump was actually spiking due to noble gas such as Ar leaking from the detector window and entering the ion pump. However, this made absolutely no differences. Also, the ion pump should be guaranteed to work even with some noble gases in the residual gas (I don’t know the exact “model” of pump provided by JEOL, but I've been told that the number printed on the back “20N“ should ensure it is OK to use it with some level of noble gas???).

The JEOL engineers (Ken S. and Chris O.) are kind of lost… Me too… Do you have any idea what could be a potential source for this problem?

My colleague has an RGA but we don’t know (yet) where we could hook it, maybe just above the main roughing pump valve near the ion pump. We do not have an external port available for the chamber (all occupied: 5 WDS, 1 EDS, 1 panchromatic CL, 1 video / OM). It might be interesting to see if this is a problem of contamination (other than P-10 gas). Would you recommend to do so, and if we hook it up, what kind of “problematic contaminants” that would cause this ion pump behavior should we be looking for? Noble gases? Other?

Many thanks for your considerations, and looking forward to see you () at M&M in Columbus OH next week!

Julien

[Beamsys]

While I don't specifically deal with electron probes, have seen the ion pump voltage "spiking" in SEM - most likely either ion pump, or controller, or both would need to be replaced. If you want to save money on buying new ion pump then have your original removed and send to Duniway (https://www.duniway.com/)  for full rebuild. Explain the issue while sending the pump, so they specifically test it for discharges. Once you get the rebuilt pump back - connect it to controller (it will be shipped sealed under UHV) and test before breaking the seal. For gas contamination - release seems to be too rapid and too periodic; I'd be guessing discharge/arcing as more likely cause.

Have no connection to Duniway other then being a very satisfied customer - they rebuild over 50 ion pumps for me over the years.
 
Valery

[Julien]

Dear EPMA lovers,

It turns out that after several months of erratic testing, replacement of ion pump, replacement and cleaning of several O-rings, our engineer found the “culprit”… The two O-rings on the condenser lenses were poorly manufactured and causing an air leak. Kudos to the patience and intense testing from our JEOL engineer, Ken Stern.

Julien