Xenon versus Argon detectors

[Joe Boesenberg]

Hi All

With a little bit of luck, my lab will be looking to purchase a new JEOL microprobe next year. At the moment, I have a quote from JEOL for spectrometers that use XENON detectors. I was reading last week on the forum that these detectors have to be replaced every 3-5 years. What is the opinion of those who have purchased and used such detectors, are they worth it? Does anyone know what the ballpark cost is to replace them (I will inquire from JEOL at some point in the future)? What is the improvement over using argon detectors? Are there any other disadvantages to having them?

Thanks. Joe 

[Anette von der Handt]

Hi Joe,

Xenon detectors perform better with higher energy X-rays since Xenon is ionized more efficiently than Ar by shorter wavelengths*. Hence, Xenon detectors are commonly used in spectrometers with LIF crystals. P10 detectors are usually paired with LDE-X and TAP crystals. Spectrometers with PET crystals can go either way and usually it depends what other crystal types they are installed with. I personally like having a spectrometer with PET+P10 and another with PET+Xenon for flexibility and performance optimization.

The difference in performance can be quite significant and usually increases with increasing X-ray energy. The following numbers are from using Virtual WDS software.

Ni Ka on LIF+Xenon: 559076 cps/uA; 262.43 p/b
Ni Ka on LIF+P10: 234970 cps/uA; 339.42 p/b

Yb La on LIF+Xenon: 198394 cps/uA; 58.47 p/b
Yb La on LIF+P10: 84507 cps/uA; 75.04 p/b

For comparison, a lower energy X-ray line
Ca Ka on LIF+Xenon: 114641 cps/uA; 969.19p/b
Ca Ka on LIF+P10: 104984 cps/uA; 976.26 p/b

While 3-5 years is given as the usual lifetime of the detectors, I find that they often perform well much longer. Most service contracts include the replacement of the detectors. Either way, opting for all P10 detectors to minimize running costs would result in an EPMA very severely limited in its capabilities.

*FYI, Cameca used high pressure P10 detectors instead of sealed Xenon detectors.

[sem-geologist]

....
Either way, opting for all P10 detectors to minimize running costs would result in an EPMA very severely limited in its capabilities.

*FYI, Cameca used high pressure P10 detectors instead of sealed Xenon detectors.

As Cameca was mentioned here, as user of Cameca SX100 and SXFiveFE with only P10 detectors I share here my toughs. TLDR; The advantage of high pressure P10 detector is that intensities are long-term and short-term stable, where Xe detector is only short-term stable, and low-P P10 is only long-term stable:

detector gas	[Low P P10]	[HighP P10]	[Xe]
long-term stable	yes	yes	no
short-term stable	no	yes	yes

Low Pressure P10 requires re-calibration after seasonal changes, also they are clearly very sensitive to atmospheric front passing (drop or raise in pressure). If compared with low pressure P10 det., then, yes, sealed Xe detector looks immune to atmospheric changes, and that could be kind of advantage. However, Xe detector and P10 low pressure are aimed to different energies, and Xe is not used for low energy X-rays.Furthermore, it looks less advantageous when compared with high pressure P10 detector, which intensities are both - long and short time stable. The intensities on high pressure P10 detectors keep stable for few/many years - which makes the instrument more efficient as it needs less frequent (re-)calibrations - there is more time available for analysis. In contrast, with constant Xe escaping from Xe detector - the calibrations probably needs to be (re-)done quite often for all higher energy X-ray lines. Thus I would chose high pressure P10 detector any moment instead of Xe detector for LiF and PET, and would consider a probe without such an option to be defective-by-design. Such defect could only be compensated with lifetime free-of-charge unlimited Xe detector replacement contract or/and a significant discount (should be same price as P10 option) to compensate the time required for repeated re-calibrations. The only advantage compared with low pressure P10 is that recalibrations of Xe detector can be planned periodically in advance, there with low pressure P10 detectors the (re-)calibrations will follow local weather events.

BTW, if outlet pressure of P10 gas from detectors could be regulated to absolute value instead relative to the atmosphere (or atmosphere pressure of lab is stabilized), low pressure P10 could have also improved short-term stability. Maybe probe developers/builders could offer rather some ready-built-in solution for that... that would be some value-added worth to pay more.

P.S. I scratched my eyes twice, as p/b given by that Virtual WDS looks at disadvantage for Xenon. That is indeed correct, as too much signal is not good too, as it will over-saturate counting system. Higher intensity comes with downsides as it gets easy to go over 200 kcps for major elements. In conjunction with large XTALs it is easy to get intensities which hitherto sold counting electronics were not designed to cope with. Xe detector makes sense for trace analysis, but for major elements that could be kind of hindrance.

[Probeman]

To Anette's and SG's comments I would add that not only do sealed Xe detectors (as mentioned) need replacement after a period of a roughly 3 to 5 years, but that this is due to degradation of the sealed detector through two mechanisms, firstly though contamination of the detector due to internal out gassing and secondly through pumping out of the detector through small leaks in the window or even the detector body. 

Contamination of the detector can occur in flow detectors but on much longer times scales, but small leaks are not usually a problem unless they are large enough to affect the vacuum. Cameca has an advantage here in that there is a poly or mylar window between the spectrometer and sample chamber vacuum systems. The only downside with this window that I know of is a ~15 to 20% decrease in N Ka intensities due to absorption by the carbon based window.  However on the plus side, this window also helps to prevent Ar poisoning of the ion pump in the electron gun by leaky detectors.

A good way to monitor this degradation is to perform yearly checks on the peak to background of a specific (high energy) line in a specific material. What you will observe over time is that the intensity of the line will decrease, but more worrying is that the peak to background will decrease also. In fact this is another downside to the Xe detector in my experience, is that it will generally have a lower P/B than a flow detector even when new, especially for low energy emission lines.

Someday (I hope) we will have solid state detectors on our WDS spectrometers that will perform well from 100 eV to 20 keV and no need to lug around gas cylinders...   

Here are some related threads, specifically this one on stability and back flow regulation of P-10 gas detectors:

https://probesoftware.com/smf/index.php?topic=1109.0

And this topic on "unexplained" changes in intensities from one day to another:

https://probesoftware.com/smf/index.php?topic=1257.0

[Joe Boesenberg]

Anette, SG and John,

Thanks for all the information. Greatly appreciated.

Being my lab predominantly does routine olivine, pyx, feldspar, metal, and glass, and only occasionally wanders into trace elements, I would prefer a very stable system over high count rates for heavier elements. I have been spoiled for the last 26 years as I have worked with two different SX-100s, both with two high pressure spectrometers with LLIF and LPET crystals. I can always extend count times for trace elements, not to mention we have an excellent LA-ICP-MS lab in house. My lab has historically been financially challenged by not generating sufficient funds to maintain service contracts that cost $50k+ per year. Only recently am I approaching that level of revenue. That's what happens when there are (at least) six university microprobe labs within 100 miles of us. 

Does anyone know if JEOL offers high pressure argon counters? Definitely something I have to bring up with JEOL.

Thanks again.

[Probeman]

I don't think it's a heavy vs light element thing so much as it is a high energy emission line vs a low energy emission line thing.  For example with a 2 atm (high pressure) Ar detector, about 50% of Zn Ka lines are transmitted through the Cameca detector and not detected.  But with the new large area Bragg crystals maybe this isn't such an issue...

Here is a post discussing some issues buying a new instrument:

https://probesoftware.com/smf/index.php?topic=1179.msg8086#msg8086