Flow Proportional Counter Backflow Gas Regulation

[DavidAdams]

Here at the USGS Denver Microbeam Lab we are in the process of having our new JEOL 8530F+ Microprobe installed (YAY!). Everything has been going smoothly up until a couple of days ago when the engineer was checking the performance of the WDS detectors. Without going into details, the GFP counters aren't really working the way they need to be.  Because of the altitude in Denver the atmospheric pressure is significantly different than at sea level and that has a very noticeable affect on X-ray detection using the GFP counters versus the sealed Xenon detectors. It screws with the gain and bias voltages and P/B ratios. Changes in barometric pressure here has a very noticeable effect on the PHAs on our system much more so than I have seen on other instruments at lower altitudes. This is something that we've noticed for years on our old 8900 and we've recognized it and just dealt with it, but with the new probe being installed I wanted to get people's opinions and thoughts on ways to eliminate or minimize the effect of altitude and barometric pressure on the GFP counters.

My questions are: has anyone with a JEOL probe attempted some form of back pressure regulation at the end of the P-10 gas line from the GFP counters? Is this even possible or feasible? Is there some why to safely do this without blowing the detector windows that are under vacuum?

Any input/ideas/help would be greatly appreciated!

Thanks,
Dave

[DavidAdams]

Hi Owen,

Thanks for the reply!

The dibuthyl phthalate in the bubbler is already something that I do on the the old 8900 probe and I also gave some to Ken for the new 8530 probe bubbler. Unfortunately, that does very little to curb the back pressure problems that I experience up here. I also experienced the atmospheric pressure effects when I was running the 8530F in Perth but the effect was less dramatic then what I'm seeing here in Denver.

Thanks again!
Dave

[Probeman]

Hi Dave,
In Denver you are at roughly 85% sea level pressure, so I would expect that you can increase the back pressure by 15% and the detector windows should handle it fine.
john

[DavidAdams]

Hi John,

Thanks! That's definitely something I will try when JEOL hands over the 8530 to us. Hopefully that will be a simply easy fix if it works!!

Dave

[Karsten Goemann]

We got a bubbler from JEOL for our 8530F Plus as well.

I'm wondering why dibutyl phthalate is the fluid of choice (same for the bubblers on our Cameca SX100)? At least in Australia it is a class 9 dangerous substance, and the specified vapour pressure at room temperature is not particularly low either.

Wouldn't it be possible to use something like diffusion pump oil or even rotary pump oil? Many are non-hazardous, much lower vapour pressures. Santovac 5 is terribly expensive of course, but for a single small JEOL bubbler one wouldn't need much.

With the "clean vacuum system" of the newer generation machines, oil-free pumps, on-airlock plasma cleaners, cryo anticontamination etc, what would be the best bubbler fluid? Normally we probably wouldn't expect much back streaming, but if a counter window leaks or breaks...

Cheers,
Karsten

[D.]

Sorry to hijack, but I've always wondered why the FE instruments come with bubblers and the 8230's don't. ? . I assume there is no technical reason?

D.

[DavidAdams]

Karsten,

This is a good question and I return to my original post. If one put a back pressure regulator at the end of the gas flow chain between the last spectrometer and the bubbler I would guess that that would help solve any atmospheric pressure influences as well as help prevent any back streaming of any fluid (or anything else) in the bubbler in the event of a window failure.

D.

The FE 8530F in Perth, Western Australia didn't come with a bubbler either. I had to scrounge one up. Thankfully there was an old one in a drawer from a previous instrument. I just think JEOL randomly decides whether or not to send it with an instrument.

Cheers,
Dave

[Karsten Goemann]

Thanks Dave, we probably got the bubbler because we asked for it. The information given to us was that it is optional, so only supplied on request.

[Probeman]

I'm wondering why dibutyl phthalate is the fluid of choice (same for the bubblers on our Cameca SX100)? At least in Australia it is a class 9 dangerous substance, and the specified vapour pressure at room temperature is not particularly low either.

Wouldn't it be possible to use something like diffusion pump oil or even rotary pump oil? Many are non-hazardous, much lower vapour pressures. Santovac 5 is terribly expensive of course, but for a single small JEOL bubbler one wouldn't need much.

With the "clean vacuum system" of the newer generation machines, oil-free pumps, on-airlock plasma cleaners, cryo anticontamination etc, what would be the best bubbler fluid? Normally we probably wouldn't expect much back streaming, but if a counter window leaks or breaks...

I would think just some cheap silicone based diffusion pump oil would be fine to use as a bubbler fluid, but I have no special knowledge in this area.  Except that at room temperature, silicon oil is not going to be out gassing anything!

[Probeman]

This is a good question and I return to my original post. If one put a back pressure regulator at the end of the gas flow chain between the last spectrometer and the bubbler I would guess that that would help solve any atmospheric pressure influences as well as help prevent any back streaming of any fluid (or anything else) in the bubbler in the event of a window failure.

Cameca instruments have a bubbler on the exit of each spectrometer. Do JEOL instruments only have a single bubbler? 

The only advantage I can think of for separate bubblers is that it's easy to see which spectrometer has a broken detector window. Not that it happens very often, especially (never) since we added these "soft start" valves to our roughing pump line:

http://www.vatvalve.com/business/valves/catalog/H/311_1_A

https://probesoftware.com/smf/index.php?topic=120.msg487#msg487

[Karsten Goemann]

John,

Yes, by default only one bubbler on the JEOL as normally all the P10 channels are in series. I assume one reason for having separate regulators and bubblers on the Cameca would be that they use different pressures ("low" and "high" pressure counters) for different x-ray energy ranges. On the JEOL all P10 counters are at the same pressure as sealed Xe is used for high x-ray energies.

But I've seen at least one JEOL where the P10 line is split upstream from the spectros to have an individual supply for each P10 counter (and I think also a separate bubbler).

I was curious and did some testing on our 8530F+ where the three P10 spectrometers are currently in series. I moved the last spectro in the line on a major element peak, and then did the same for the first and second spectro while monitoring the count rate on the last spectro in the line. There is no noticeable change even at very high count rates on the first two spectrometers. So "recycling" of the gas further down the line does not appear to be affected by what's happening on spectrometers upstream.

I've also heard that one may see differences in count rates when changing the order of the spectrometer in the P10 gas chain on a JEOL instrument, i.e. the spectrometers further downstream can have lower count rates. But I'd assume that may be due to P10 leakage somewhere in the chain, e.g. a leaky detector window, or a tube not attached properly...?


Dave,

I don't have any experience with these back-pressure regulators. I can see how it might help in your case at high altitude, but do you think it would  also make a difference for us, being pretty much at sea level? I.e. in addition to preventing potential backstreaming, could they even out fluctuations in atmospheric pressure? There seem to be a range of designs available. How much of a pressure difference to atmospheric pressure would one have to set for them to operate properly? Obviously too much pressure on the detector windows etc would be a concern.

Cheers,
Karsten

[DavidAdams]

Karsten,

It would be interesting to see if a back-pressure regulator would be useful for JEOL probes at lower altitudes. I think we're going to try one here to see if it will actually smooth out the pressure effects we see. Have you monitored your PHA bias voltages on your P10 counters over time particularly when there is a big change in the atmospheric pressure? In Denver the voltages on our flow counters can often be ≥±10V during large changes in the weather, which, as you can imaging, has a noticeable affect on the analyses being performed. I don't remember what I was seeing when I was in Perth, but I don't remember it being quite so dramatic. The bias voltage on the Xenon counters is almost always the same. The greatest variation on the xenon counters that I've seen here is ±2V and that could be easily attributed to non-pressure related changes. Usually, however, the bias voltage is exactly the same for weeks and months at a time. A back-pressure regulator should, in theory, stop any atmospheric pressure related fluctuations. There are some very low pressure regulator options available (0-10 psi) and from what I understand there doesn't need to be a large pressure differential in those regulators to operate properly. I'm sure that the JEOL engineers will know what the maximum pressure before window failure would be.

Cheers,
Dave

[Brian Joy]

Hi Dave,

In Kingston, at ~100 m AMSL (and no P-10 back-pressure regulator), I see variations in bias voltage required to keep pulse amplitude distributions centered at 4 V using the gas-flow counters (but not with the sealed Xe counters, at least over short periods of time).  These swings can be of similar magnitude to the ones that you report and appear to correlate with variation in atmospheric pressure -- but I need to keep a better record of this in my log file to be absolutely certain.  Even if centering the PHA distribution at 4 V for a given count rate doesn't solve the problem of variation in deadtime with X-ray energy, it sure does provide a useful reference for monitoring fluctuations in the distribution due to various causes.

Brian

[Brian Joy]

For the past ten days I’ve been collecting data on P-10 gas flow counter anode bias as a function of atmospheric pressure; I’ve used the pressure recorded hourly at the Kingston airport (93 m AMSL).  I’ve measured Si Kα on wollastonite while keeping the count rate at 5000 s^-1 and adjusting the bias until I get the PHA distribution centered at 4 V.  I’ve done this in the PC-EPMA “base level” window using a step of 0.1 V and dwell time of 1 s.



Barometric pressure (corrected/adjusted?) averages about 100.5 kPa and usually stays within about 1 kPa of this value.  However, during intense storms in the fall, winter, and spring, it can range as low as ~97.5 kPa.  On the opposite end, after passage of a cold front, the pressure can rise as high as ~103.5 kPa.

I’m going to keep adding data to this plot.  I’m curious to see how much change occurs when barometric pressure is at more extreme values.

[Mike Matthews]

Excellent plot Brian, mind if I use it in my WDS lecture?

This is one of the reasons why the SDD-WDS that Ken Moran and Ric Wuhrer have been developing in Australia is a good idea - no pressure sensitivity.