Buying a new probe

[dawncruth]

Hi all,
Perhaps I've contacted several of you directly. It was suggested to me to post my question on the forum, so folks in a similar position might benefit from the wide sharing of our collective wisdom.

Now onto the wisdom.

In May, I'll take over as the probe manager at the USGS Menlo Park. We are currently looking into a new probe as our current probe is the venerable JEOL 8900 SuperProbe. She is showing her age. Anyway, I've been tasked with making a wish list for this initial search. I would ask for your collective advice about:
1. What is a must have add on?
2. What should we avoid?
3. What do you see coming down the pipeline?

We are not set on any brand yet. So if you are a die-hard Cameca/JEOL fan - here's an opportunity to wax poetic about what you like and dislike about your respective instruments.

Any insight you could provide would be most welcome. And if you are a US citizen - consider it an opportunity to provide feedback into how your taxes are used!

Thanks!

Dawn

[Probeman]

In May, I'll take over as the probe manager at the USGS Menlo Park. We are currently looking into a new probe as our current probe is the venerable JEOL 8900 SuperProbe. She is showing her age. Anyway, I've been tasked with making a wish list for this initial search. I would ask for your collective advice about:
1. What is a must have add on?

Hi Dawn,
Congratulations on your new position!

My personal (probably quite jaded!) opinion regarding add-ons, is that a full featured EDS system is a must. It used to be common in the past that Cameca and JEOL provided stripped down EDS systems on their EPMA instruments. These EDS systems did not have many of the cool features that SEM users had available to them, for example spectrum imaging and stage automation.  The good news is that this situation has improved much for the better in recent years.  But one still needs to do their research and make sure they are getting all the options and capabilities in an EDS system for their microprobe.

Now because a microprobe tends to be run at higher beam currents than an SEM, my opinion is that one doesn't really need a detector larger than 10 sq mm or so, and in fact, I've had to utilize my EDS aperture turret on my SX100 in order to be able to run the EDS while acquiring trace elements by WDS at high beam currents:

https://probesoftware.com/smf/index.php?topic=342.msg1789#msg1789

We have a Thermo SDD on our Cameca SX100, but whether you get a Thermo or Bruker SDD (both are excellent systems) you definitely should get a complete EDS system from that vendor. Remember one can put a Thermo or Bruker detector on both JEOL or Cameca instruments.  Don't limit yourself. Shop around!

This next item isn't an add-on, and maybe not be that useful for geology, but I have two 4 crystal spectrometers on my instrument which is great for acquiring several light elements at a time. I would also make sure to get large area crystals for all the spectrometers that one can afford. One can never have too many photons.  Especially now that we seem to be using WDS more and more for trace elements and acquiring the major elements using EDS:

https://probesoftware.com/smf/index.php?topic=79.0

If you're interested in seeing more about EDS WDS integration, here's an example in Probe for EPMA specifically looking at traces using WDS, and majors using EDS, even when there is a significant interference of a major element (Si) on a trace element (Rb):

https://probesoftware.com/smf/index.php?topic=79.msg7818#msg7818

I wrote it intended for the students here, but even an expert can appreciate the work flow that was documented.  The other thing which I really like as an add-on for EPMA is the cryo-pump system we have on our SX100, as seen here:

https://probesoftware.com/smf/index.php?topic=646.msg3823#msg3823

We have it because we had to buy a diffusion pumped system, but this cryo cooled baffle works so much better than the oxygen jet/cold finger for carbon contamination, and amazingly the cryo system is actually cheaper!  Make sure to get the temperature controller also so one can see the actual temperature.   Even if I was to get a "dry pumped" vacuum system using scroll pumps I would still get this cryo pumped system as most of the contamination seems to come from the sample itself.  And because the cryo pump system is air cooled, it just runs and runs (at 100 Kelvin), and you don't need LN2.

Of course you'll probably want a good CL system for geology. If I was going to buy a new instrument I would think a lot about just getting more than the standard vendor provided CL solution. Perhaps some kind of insertable parabolic mirror CL collector, or maybe even a fiber coupled system.  Depending on how much CL means to your clients maybe even a Gatan or xCLent system. But that is a lot more money.

I would also try to make sure that I had a (large) PET crystal on every WDS spectrometer.  This can really come in handy for a number of analytical scenarios. E.g., measuring trace elements on all spectrometers. I do think that the WDS microprobe will become more and more focused on trace elements and be using the (integrated) EDS for the major elements quantification.

There's an interesting discussion on WDS spectrometer configurations here:

https://probesoftware.com/smf/index.php?topic=360.0

2. What should we avoid?

Don't assume that you have to buy a "package" solution from the vendor. You will have to do a lot more research, and spend more time working out details but you'll have a much higher performance system if you customize it for your specific laboratory requirements.

Also on what to avoid: a poor lab environment.  Make sure the lab *exceeds* the vibration, and especially, the temperature stability requirements of the vendors. Not just meets them, but exceeds them.  That may be impossible if you are putting the instrument in an existing lab, but it's worth thinking about.  See this topic on lab performance (you have to be logged in as a member to see this board):

https://probesoftware.com/smf/index.php?board=24.0

3. What do you see coming down the pipeline?

The biggest thing I see, and something I have been pushing for some time now is what we call the "holy grail of microanalysis". And that is integration of WDS and EDS for quantitative mapping:

https://probesoftware.com/smf/index.php?topic=400.msg3702#msg3702

We can already perform quantitative integrated EDS and WDS for point analyses as I previously described, but the next big thing is integrated EDS-WDS quant mapping.  This is something that I know both JEOL and Cameca are working on, but not sure where exactly they are in this process.

The other next big thing is solid state WDS detectors, or maybe tandem gas flow and solid state detectors. The idea being to acquire low energy photons with the flow detector and high energy photons (which will pass through the gas flow detector) with a pin diode detector mounted on the back of the flow detector:

https://probesoftware.com/smf/index.php?topic=644.msg4918#msg4918

No need for an SDD detector as the Bragg crystal already provides all the energy filtering one needs. Again, not sure where the vendors are with this technology.

We are not set on any brand yet. So if you are a die-hard Cameca/JEOL fan - here's an opportunity to wax poetic about what you like and dislike about your respective instruments.

I really like the visible light optics of the Cameca instruments.  Exceptionally clear and easy to focus.  I really like the FEG performance of JEOL field emission instruments. Amazing high resolution performance and stability. I really like the Cameca stage and spectrometer reproducibility.  The engineering is amazing and robust, they just run and run for years and years. 

Now if only one could buy a "hybrid" instrument and get the best of both worlds!

Any insight you could provide would be most welcome. And if you are a US citizen - consider it an opportunity to provide feedback into how your taxes are used!

Ha!      You asked for it (I know you mean tax supported microprobe purchases but I just can't help it)!  My feedback is: let's return the US to the tax rates of the 1950s and 1960s where the top marginal tax rates were around 80 to 90% (remember, we still had plenty of millionaires then, but the middle class was growing not shrinking!), and to take some of our bloated military budget and have a "Manhattan Project" level national effort to convert the US to renewable energy and smart grid storage, to show the world how we all should be protecting the planet's climate and air and water... But I'm just one vote, as I can't afford to buy any politicians.   

But enough politics! You also asked by email about purchase and performance specifications and so I thought I would just say a little about that.  When I bought my two last instruments we really put the vendors through some tough specifications and I don't regret it. See here for the gory details on our last purchase:

https://epmalab.uoregon.edu/purchase.htm

But it's a lot of work and stress.  On the other hand, we really got the attention of the instrument vendors!  But in the end I would say that one really only needs a few critical performance specifications.  Stage and spectrometer reproducibility is important of course, but if we think of the microprobe as a k-ratio machine (which is really what it's all about), the the test for simultaneous k-ratios on all spectrometers is probably the single most important test one can place into the purchase order.  And any specifications you stipulate do need to be part of the binding instrument acceptance!

It's quite simple really.  If you tune all your spectrometers to the same (moderate energy) emission line, say Mg Ka or Si Ka or Al Ka, and then measure two materials (they don't even need to be standards), each with significant differences in absorption for that emission line, you will quickly find out if your instrument is doing what it is supposed to do.  See this topic for more on this:

https://probesoftware.com/smf/index.php?topic=369.msg1948#msg1948

I will also point out that there is an actual board (again you have to be logged in as a forum member to see it) that is for JEOL vs. Cameca discussion:

https://probesoftware.com/smf/index.php?board=17.0

It's only got 63 posts (I guess most people are a little shy about telling us how they really feel about their instruments), but it is worth reading the whole thing.

Again, this is just my own (very biased) thoughts on these excellent and important questions, I'm sure many others have valuable experiences and advice to share. But please feel free to ask me more about where you think I'm not making any sense. Which is probably most of what I said!   

Again congrats on your new position!

[jrminter]

Probeman nailed it. I have two points to add:

If you are going add an EDS system as well, make certain you factor in off-line licenses. Some of the vendors are very good about this and give you a site license and your users can process data off-line to their heart's content. Others have much more restrictive licences that are both expensive and node-locked. I had one of these before I retired and it was a pain!

One other point (I know Probeman agrees but don't recall seeing it in his post...) You will need to live with this instrument for a LONG time. The bane of our existence has been software updates to cope with the Windows computers. You really want to write software support for the anticipated life of the instrument into the purchase contract.

Congratulations on getting a new instrument!

[Probeman]

Hi John, thanks.

Yes, the ability to reprocess data off-line is often overlooked.  I agree it is a very important aspect for running a lab.  I think I neglected to mention it because I'm so used to being able to re-process my WDS/EDS data (and for customers to reprocess their data and students to reprocess their data) on any computer anywhere, that it just didn't pop into my head.   

Of course that is because we use Probe for EPMA and that software can be distributed to every user for installation on their own computer, and also we use Thermo NSS/Pathfinder, who also allows their software to be copied on to any number of computers for data reprocessing.

Just FYI, I think Bruker (which used to charge for off-line processing) now also allows users to install their Esprit software on off-line computers for data re-processing.

On the Windows update issue please don't remind me!     I'm currently going through update hell on my Cameca PeakSight computer:

https://probesoftware.com/smf/index.php?topic=1171.msg8073#msg8073

It seems like I've spent most of the last year just updating my lab computers from Windows XP to Windows 7 or 10!   

[dawncruth]

This is all great information and I think will be very helpful as I build a "probe of dreams" instrument. Thanks all for your input!!

I will compile my email responses and share them here as well.

Dawn

[BenjaminWade]

My two cents...cant have too many LLIF xtals as well.
I have LPET on 4 spectrometers, but only LLIF on two. I now wish I had a third LLIF, maybe in place of my LTAP. For geo orientated work with sulfides and REE minerals you can't have too many LLIF xtals...

Cheers

[Probe321]

Make sure the computers that control the system have adequate memory and solid state drives with at least 30 inch monitors.  I have had to many issues with inadequate memory errors as the vendors do not supply enough computer ram for the tasks.  As a Federal employee I cannot endorse any one vendor, but  have found it helpful to have both vendor and Probe for EPMA software available.  I can send you a copy of the specifications used by NETL (they are public record and where on line, but think they have aged off the website) if you like them for review.

Concerning EDX it is nice to have, but as one senior researcher put if so simply why use EDX when you have WDX detectors.  The researcher wanted his work done with the WDX detector which leads into get the maximum spectrometers that you can put on the column.

WiFi for the service engineers and quests in the lab.  (Our labs have faraday cages built into the walls for field (and cell phone) cancelation)

[Probeman]

Probeman (aka John Donovan) is back after 5 days without power or Internet. 20 inches of snow still at our house...

Concerning EDX it is nice to have, but as one senior researcher put if so simply why use EDX when you have WDX detectors.

As someone who is quite partial towards WDS, I would respond to your senior researcher thusly: why waste time using WDS spectrometers for major elements?  Since WDS's forte is sensitivity, why not use EDS for the major elements and WDS for the trace elements? Some examples here:

https://probesoftware.com/smf/index.php?topic=79.msg7818#msg7818

That said, EDS can really only compete with WDS accuracy when using standards. And of course the best thing about WDS is that one *must* use standards. But if one does utilize standards for EDS quantification (and please note that one *must* use standards for EDS quant in Probe for EPMA), why not tune up your trace elements on the WDS spectrometers and acquire EDS elements for the major (and maybe even minor) elements?

Now it's true, if those EDS elements have low energy emission lines, or have elements with pathological spectral interferences, it gets more problematic, but for many quantitative situations, integrated EDS and WDS are excellent, and can really save time especially when analyzing for lots of elements.

The researcher wanted his work done with the WDX detector which leads into get the maximum spectrometers that you can put on the column.

I agree with this sentiment no matter what ones analysis strategy is. In fact, I personally would love to have a 6 WDS spectrometer instrument, of course with a nice EDS system as well. 

[aburnham]

We used to have a Cameca SX-100, replaced with a JEOL FEG probe.

JEOL software has some maddening problems that make it a pain to use (most notably the inability to combine data for the same element from multiple spectrometers e.g. Ti in quartz). It's a lot easier to get money for software when you write the budget so plan ahead to buy PfE if you go with JEOL.

We have found the H-type spectrometer that we bought to be a bit mixed in terms of performance. We have LDE1 and TAP. On TAP it barely gives better counts than a regular spectrometer for Mg and Al (but better for Na). Its smaller size means it can't do Si, unlike a regular TAP spectrometer. It works nicely for LDE1 as far as I can see (not much experience with it so far).