EPMA Vacuum Technology Discussion

[Probeman]

This is a topic for discussion of general vacuum issues in EPMA, both JEOL and Cameca instruments.

I'll start this thread by mentioning an excellent technology that Cameca provided for my SX100 instrument instead of the usual "cold trap finger" and "oxygen leak jet" hardware for minimizing carbon contamination.

When I actually set out to purchase the instrument about 5 or 6 years ago, the Euro was very unfavorable for the US, and although I had budgeted for a dry pumped turbo system, I could no longer afford it. So instead I suggested to Cameca that they provide a diffusion pumped system, with a cryo-trap over the diffusion pump and without the cold finger and oxygen jet.

The results were excellent as seen in our acceptance testing and the air cooled cryo-unit runs at 100 degrees kelvin! So it not only traps oil, but even water quite efficiently.  About once a year we warm it up to boil off all the contaminants with the secondary valve closed.

The link to the unit is here:

http://www.brooks.com/products/cryopumps-cryochillers/cryochillers/pcc-compact-coolers

The only trouble we've had over the last 6 years is to replace one cooling fan on the compressor. And zero LN2 and oxygen consumption...



Not too bad for an inexpensive and reliable oil pumped vacuum system!

Edit by John: There's another method which JEOL attaches to their airlock (see attached pdf image from Aachen), but I would prefer an "in-situ" cleaning method, so only the area of interest is cleaned as described below in the request for a UV laser for cleaning concurrently with the x-ray acquisition.

[Gseward]

Our SX100 has a turbo pumped system backed with scroll pumps. This test was done when the instrument was being installed. I have no idea what it is like today!

Description : test made on pure copper at 10kV 100nA.



Rate = 0.033% and is a similar result to John Donovan's Aqua Trap system. Not sure where the 'Limit without device' spec is from  - I'd be interested to see experimental results from a diffusion pumped system without any decontamination devices. It would also be interesting to compare numbers from somewhere that does a lot of Carbon analysis (Aachen 8530??).
Have you done a retest lately John? I have a plasma cleaner that I've been meaning to bolt onto the probe, so perhaps it is time to do a new test 'before and after'.

Our vacuum system hardware has been very reliable for 3.5 years. All I've had to do is replace the tip-seals in the scroll pumps a few times and clean the high vac gauge.

gareth

[Probeman]

This is a crazy idea, but I'd like to try a hydrocarbon experiment in my EPMA vacuum system and need to borrow or buy a small UV laser that I can mount just outside the instrument.  The specs my engineer suggests are as follows:

266 nm UV pulse laser (frequency quadrupled YAG), diode pumped solid state (DPSS)

Typically these devices are some 10 to 20 cm long. If anyone has one of these guys that we can borrow or buy (cheaply!) I would be very interested to hear from you.

Thanks.

[John Donovan]

Warren posted this recently on the list server and he said it was OK to re-post it here:

I presume your setup has three different purposes for those pumps and that each one is connected to pump a different part of the system. If so, I don't know that you will be able to combine them to a single vacuum manifold. They could be pumping against different loads at different vacuums. That is the case with our FEI Quanta variable pressure SEM. It has two roughing pumps because it needs to have two. One backs up the turbo pump at the best vacuum possible with whatever pumping load it has. The second pumps on the chamber when it is operating in variable pressure mode (10-160 Pa). In that mode, it would probably be more difficult (and expensive) to use a single pump. The second pump is tied in to help on roughing the chamber when first pumping down, but it is isolated and turned off before the turbo seriously comes into play. So, two pumps was by good design.

Now, if you have three pumps from three different instruments that are working at maximum vacuum all (or most) of the time, you might be able to design a system that could reduce the number of pumps. (Since parts came from different vendors, there was probably not an option to combine vacuum systems.) However, you would probably need extra valves and plumbing and a ballast tank to maintain high vacuum while you borrow a pump for roughing purposes. That would not be a trivial (or cheap) engineering exercise. My experience with engineering projects is that the cost could easily outrun the cost of three brand new vacuum pumps.

FWIW, it might be time to consider other aspects. For example, we swapped out our old, oil-sealed rotary pumps with oil-free dry scroll pumps. That reduced the contamination issues within our SEM. We have also been able to take advantage of manufacturers' occasional sales to get pumps at pretty serious discounts.

Warren Straszheim, Ph.D., manager
Materials Analysis and Research Lab
23 Town Engineering
515-294-8187

[sem-geologist]

Few questions about turbo pumps maybe someone could answer it.

A) turbomolecular pump emergency braking.
Do anyone knows what happens with turbo pump at mains power loss? Agilent TP manuals tells this:

Emergency stop can be achieved by removing the power supply.

, which I think is equal to situation with mains power loss/blackout. More recent manuals also added disclaimer:

...The use of this stop mode
could lead to faults and/or damages at the pump. ...

. The first sub-question is if that "emergency stop" is not the same as engaging "Stop" command?, where "stop" simply stops powering the motor and leaves it in inertial rotation state. If no air/gas at pump's vent port is applied, it would spin with very slow spindown for about 20-30 minutes until would fully stop on its own. Maybe disclaimer has in mind  the danger would come from other parts of vacuum system (i.e. probability of introducing atmosphere pressure by inoperable backing pump, as cutting of power rather of whole system (as EMO buttons do) and not just selectively of TP???). Or maybe pump is indeed stopped to halt in very short time, and thus danger comes from enormous strain introduced to blades due to fast, near instantaneous de-acceleration? Could such thing even be achieved at all? One of possible way I can think about would be shorting engine coils with high resistance, so TP would work at such event as generator with very huge load and would stall (that is not good not only for blades, but for coils or more precise both rotator and stator, where both would highly overheat (and rotator is practically impossible to cool).

B) electrical braking of TP
(Just to not confuse with emergency braking, at this point B it means a smooth slowing down of pump)
I looked through different manuals of TP, and none major vendors mentions such ability. I came across some hobbyist projects which implements such thing. There are some TPs on ebay (and similar) often either missing controllers, or being sold with damaged controller (i.e. cut wires and other atrocities). So naturally some hobbyist came up with own controller designs and some had implemented electric braking function.
    Meanwhile, all TP vendors have a unified way to slow down TP with introduction of small regulated air (or clean gas) stream at middle point of TP. For that all TPs have such vent port (inlet). Most of controllers has reserved pins for connection with valve connected to TP vent port, and controller then controls its operation, when "stop" command is being engaged, or if power is lost (where existence of emergency brakes hypothesized in point A by me are redundant).
    However, It gets interesting in official video from Agilent "Turbo Pump Care and Use 101 - Part 5 - Venting and Shutdown", at 3:29 there is a section "electric braking of the rotor" (https://www.youtube.com/watch?v=g33_NQsTMa0&t=209s). It is stated there that newest design has that mode. Actually I then dig deeper and found out that "newest" also include over 10 year Varian pumps, which ,honestly speaking, is not so "newest". It just seems this feature was never properly marketed and actually still is not. Even the most recent manuals and Datasheets of Agilent TwissTorr 305 has exactly 0 mentioning about that feature. It is cryptically described in integrated (small) controller manual (at RS232 communication protocol) and a bit more described (but also in very limited way) in the manual of rack-mountable controller. The feature can be completely overlooked if not knowing what to look for (it is called in controller manuals "Active Stop").
    I think it is very useful feature for SEM and EPMA, as there are situations where TP would be wished to be brought to halt (for safety), but without venting a chamber. I believe our Zeiss SEM use that function, as TP slowing down can be hear even if vent gas bottle is being closed at "vent chamber" command. Zeiss SEM also has feature of partial-vent which again, stops the TP, and keeps residual vacuum in the chamber (kind of energy efficient stand-by mode). As for EPMA-FEG, such mode would be useful for some of maintenance procedures, in example: swapping/doing maintenance of primary pump, doing maintenance of air compressor (all valves automatically closed, TP would loose backing vacuum), power saving during blackout (so UPS could keep UHV at gun for much much longer)...

  Had anyone hear anything about electric brakes of other vendor TP?

[zorch]

Some thoughts about TMP's I've seen implemented in various instruments:

Some JEOL SEM's in the 90's used a Seiko mag-lev TMP. A bit sensitive to mechanical shock as the TMP's rotor could be knocked out of bounds and the TMP would crash. Crash, meaning the rotor magnetic suspension would turn off and the rotor would fall onto the emergency bearing. Very loud and screechy. Only a few such crashes could be tolerated before permanent damage was done. This TMP could handle total power losses, but only if the back up battery were charged and in good condition (not older than its reasonable useful lifetime). If you are bringing one of these back to life, you'll need to replace that battery (NiCd?). This TMP needed a battery for use in suspending the rotor during RPM rampdown after power loss. I'm not sure, but this TMP could have employed electromagnetic braking on power loss to reduce RPM rampdown time. The need for a battery in that era of Seiko TMP+ contro implies that they had not employed a rotor/motor power generation to power the suspension and centering magnets.

In our IonTOF instrument, a Leybold mag-lev TMP serves the main chamber that does employ rotor/motor power generation to power the suspension and centering magnets. So there is no battery to go bad. I've never caused this TMP to crash via mechanical shock, not for lack of unfortunate incidents! The mag-lev system has noiselessly brought the TMP to zero RPM after numerous power down events. The controller did go bad once, and Oerlikon-Leybold repaired it by replacing all of the aluminum capacitors. I should have thought of that!