Performing Integrated WDS and EDS Acquisition in PFE

[bgarcia]

OK, so John and I did some testing and found that JEOL eds is not returning any net intensities to PfEPMA, or rather the data file is not in the place it's supposed to be on the JEOL PC. Either way, I'll be doing some more investigation into 1) making sure the JEOL eds is functioning properly to begin with and 2) looking to see (with JEOL engineer) where the intensity files are kept.

Stay tuned!

[John Donovan]

OK, so John and I did some testing and found that JEOL eds is not returning any net intensities to PfEPMA, or rather the data file is not in the place it's supposed to be on the JEOL PC. Either way, I'll be doing some more investigation into 1) making sure the JEOL eds is functioning properly to begin with and 2) looking to see (with JEOL engineer) where the intensity files are kept.

Stay tuned!

To clarify a bit, the JEOL EDS interface is returning the EDS spectra as expected, but subsequently returns zero value intensities for net intensity requests (and no errors are reported).  At least for oxygen k-alpha. Ben will next try to get net intensities for other elements to see if it's a general issue.

Any ideas on what might be causing this?

[bgarcia]

Update: I've gotten the WDS/EDS integration package to successfully return intensities in PFE. JEOL installed a software update (1.19.0.1) for PC_EPMA.

I'm currently working through setting up my first "real" file with EDS enabled.

[Dan R]

I just got the EDS-WDS integration with my JEOL OEM EDS. So for clarity, when I get EDS net intensities back, where are the backgrounds coming from? 

That is are the default net intensities from the JEOL software correct?

[John Donovan]

Both the EDS net intensities and EDS background intensities are returned by the JEOL get EDS net intensity function call.

As to how accurate these net and background intensities are, I would have to refer you to someone who has tested the JEOL EDS system.  I believe Changkun Park at Korea Polar Reseach Institute has done some testing on his JEOL EDS system. Also Glenn Poirier at the University of Ottawa.

[John Donovan]

This is a new feature that we just added which allows one to plot WDS and EDS spectra together when using the integrated EDS and WDS acquisition feature in Probe for EPMA.

First acquire some WDS wavescans with the EDS acquisition turned on (see Acquisition Options dialog from the Acquire! window).  Might as well set the EDS acquisition time to use the sample acquisition time for the EDS acquisition runs the full time of the wavscan.

Then in the Plot! window check the Plot EDS Spectrum checkbox and for a WDS element, select corresponding Kilovolts for that element as shown here:



Then click the output menu:



Still a new feature, so let us know if there's anything that doesn't look right.  To see the WDS peak better you can check the No Peak Markers checkbox.

[dawncruth]

I'm finally getting to integrating the Thermo EDS and WDS setup, partially because one of our spectrometers went down. What better time to do this?

Three questions -- for now.
1. My EDS count times are longer than the Use Sample Counting time option. For example: I set up an olivine run which runs 264 seconds (PFE estimate), and the estimate given for EDS is 207 seconds. When I run the sample Thermo runs it for almost 6.5 minutes, so considerably longer.  I looked through the forum and couldn't find anything to explain that.

Note it also did this for my standards, which should only take 20 seconds total.
Note, I'm using a 20% deadtime

2. I started an automated olivine run with Si and Mg on EDS and the rest of the elements on WDS. It started great and then Si tanked by about 10 wt%. I re-standardized and cps/nA remained the same from my previous standard run. Usually, I'd repeak Si, but am not sure how to troubleshoot this as I didn't peak. I'm re-running the points to see what happens. Any thoughts?

3. I can't seem to specify O for stoichiometry calculations, when it was measured with EDS. I only have the EDS option.

[John Donovan]

I'm finally getting to integrating the Thermo EDS and WDS setup, partially because one of our spectrometers went down. What better time to do this?

Three questions -- for now.
1. My EDS count times are longer than the Use Sample Counting time option. For example: I set up an olivine run which runs 264 seconds (PFE estimate), and the estimate given for EDS is 207 seconds. When I run the sample Thermo runs it for almost 6.5 minutes, so considerably longer.  I looked through the forum and couldn't find anything to explain that.

Note it also did this for my standards, which should only take 20 seconds total.
Note, I'm using a 20% deadtime

The acquisition time is just an estimate and when EDS is involved the clock time is completely dependent on the actual dead times that occur.

One very important note: always use a *fixed* pulse processing time for EDS quant (that is, non-zero). If you have the "auto" pulse processing time selected in Thermo or PFE it will auto adjust based on the current count rate regardless of whether the faraday cup is in or out.  So when the cup comes out the pulse processing shifts to match the count rate, which messes everything up for EDS quant.

In other words be sure to set a fixed pulse processing time in PFE, then go to a sample, unblank the beam and set what the dead time in Pathfinder is.  Then enter that value when it asks you for the estimated dead time from the Automate window.

In practice is almost impossible to estimate this acquisition time accurately for EDS because the dead time depends on the sample and even more so on the beam current.

2. I started an automated olivine run with Si and Mg on EDS and the rest of the elements on WDS. It started great and then Si tanked by about 10 wt%. I re-standardized and cps/nA remained the same from my previous standard run. Usually, I'd repeak Si, but am not sure how to troubleshoot this as I didn't peak. I'm re-running the points to see what happens. Any thoughts?

I have absolutely no idea.

3. I can't seem to specify O for stoichiometry calculations, when it was measured with EDS. I only have the EDS option.

Are you trying to measure oxygen or calculate it by stoichiometry?  You can't do both things. 

That said you might be able to disable quant for measured oxygen and then add it in as an unanalyzed element for calculating by stoichiometry.

[dawncruth]

The acquisition time is just an estimate and when EDS is involved the clock time is completely dependent on the actual dead times that occur.

One very important note: always use a *fixed* pulse processing time for EDS quant (that is, non-zero). If you have the "auto" pulse processing time selected in Thermo or PFE it will auto adjust based on the current count rate regardless of whether the faraday cup is in or out.  So when the cup comes out the pulse processing shifts to match the count rate, which messes everything up for EDS quant.

In other words be sure to set a fixed pulse processing time in PFE, then go to a sample, unblank the beam and set what the dead time in Pathfinder is.  Then enter that value when it asks you for the estimated dead time from the Automate window.

In practice is almost impossible to estimate this acquisition time accurately for EDS because the dead time depends on the sample and even more so on the beam current.

Ok, that makes sense. We have a Thermo and the Time Constant (what I think you mean by pulse processing time) was set to auto. I'll have to play with it to see how to manage that.

Thermo users -- if y'all have tips on integrating EDS/WDS, I'd be most grateful!

Are you trying to measure oxygen or calculate it by stoichiometry?  You can't do both things. 

That said you might be able to disable quant for measured oxygen and then add it in as an unanalyzed element for calculating by stoichiometry.

That's exactly what I tried to do, disable quant. But when you disable quant - it doesn't allow you to select specify in the E/C properties window.

[John Donovan]

Are you trying to measure oxygen or calculate it by stoichiometry?  You can't do both things. 

That said you might be able to disable quant for measured oxygen and then add it in as an unanalyzed element for calculating by stoichiometry.

That's exactly what I tried to do, disable quant. But when you disable quant - it doesn't allow you to select specify in the E/C properties window.

Did you click on an empty row in the Elements/Cations dialog? 

[SteveSeddio]

1. My EDS count times are longer than the Use Sample Counting time option. For example: I set up an olivine run which runs 264 seconds (PFE estimate), and the estimate given for EDS is 207 seconds. When I run the sample Thermo runs it for almost 6.5 minutes, so considerably longer.  I looked through the forum and couldn't find anything to explain that.

Note it also did this for my standards, which should only take 20 seconds total.
Note, I'm using a 20% deadtime

2. I started an automated olivine run with Si and Mg on EDS and the rest of the elements on WDS. It started great and then Si tanked by about 10 wt%. I re-standardized and cps/nA remained the same from my previous standard run. Usually, I'd repeak Si, but am not sure how to troubleshoot this as I didn't peak. I'm re-running the points to see what happens. Any thoughts?

Hi Dawn,
I am happy to look at the Pathfinder data for items 1 and 2, if you want to zip the project and send it to me.
I may be able to discern what is going on.
-Steve

[John Donovan]

1. My EDS count times are longer than the Use Sample Counting time option. For example: I set up an olivine run which runs 264 seconds (PFE estimate), and the estimate given for EDS is 207 seconds. When I run the sample Thermo runs it for almost 6.5 minutes, so considerably longer.  I looked through the forum and couldn't find anything to explain that.

Note it also did this for my standards, which should only take 20 seconds total.
Note, I'm using a 20% deadtime

2. I started an automated olivine run with Si and Mg on EDS and the rest of the elements on WDS. It started great and then Si tanked by about 10 wt%. I re-standardized and cps/nA remained the same from my previous standard run. Usually, I'd repeak Si, but am not sure how to troubleshoot this as I didn't peak. I'm re-running the points to see what happens. Any thoughts?

Hi Dawn,
I am happy to look at the Pathfinder data for items 1 and 2, if you want to zip the project and send it to me.
I may be able to discern what is going on.
-Steve

I think the issue she was seeing is because she was using the "auto" time constant, which causes problems when doing quant and the faraday cup is being inserted and removed repeatedly.  Basically, the time constant is always changing during the acquisition and therefore the quant is affected.  The solution is to pick a fixed time constant for Pathfinder from the Probe for EPMA Acquisition Options dialog.

[bgarcia]

Update: I've gotten the WDS/EDS integration package to successfully return intensities in PFE. JEOL installed a software update (1.19.0.1) for PC_EPMA.
 
I'm currently working through setting up my first "real" file with EDS enabled.

Update Update:
I know this thread is long dead but I thought I'd put this here for posterity.

I was mistaken in that I posted the earlier message (below) just a little too quickly.. I DID NOT get the EDS to work using PfEPMA on our 8530. I had just gotten PfEPMA to show the peak it was scanning over, however, it did not return any net intensities besides zeros. 

I never really got beyond this point. Everything on the JEOL side worked fine; there were backgrounds and intensities galore. Eventually our JEOL guy ran out of things to try (he did a lot!), even after talking to other SEs and The Factory. The Factory just never really devoted much time to pursuing this so it died with them.

We're unfortunately unable to pursue this now as our 8530 hasn't worked in about 2 years. JEOL service in our region underwent experienced-staffing changes and they have been unable to devote someone knowledgeable to working on this full time.    When we do get the probe up and running again, I hope a solution for the issue at hand will have been found!

[John Donovan]

Update: I've gotten the WDS/EDS integration package to successfully return intensities in PFE. JEOL installed a software update (1.19.0.1) for PC_EPMA.
 
I'm currently working through setting up my first "real" file with EDS enabled.

Update Update:
I know this thread is long dead but I thought I'd put this here for posterity.

I was mistaken in that I posted the earlier message (below) just a little too quickly.. I DID NOT get the EDS to work using PfEPMA on our 8530. I had just gotten PfEPMA to show the peak it was scanning over, however, it did not return any net intensities besides zeros. 

Thanks for confirming this.

We did not learn about this issue until this last month when the Univ Toronto JEOL 8230 had a hard drive failure on their JEOL computer.

When JEOL replaced the hard drive and re-installed their software, Yanan found that she could connect to the JEOL EDS using Probe for EPMA, and even acquire an EDS spectrum, but the EDS element net intensities always came back as zeros.

That this appears to be caused by the JEOL software being updated is confirmed by Glenn Poirier who wrote to me, that he used to be able to obtain EDS net intensities from the JEOL EDS system in Probe for EPMA, but after his JEOL software was updated he no longer could get actual EDS element net intensities for quantification.

However, Julien Allaz at ETH Zurich, who has not updated his JEOL software, can still obtain EDS net intensities, so apparently JEOL broke something at some point in their 8230/8530 EDS interface.

That this is caused on the JEOL server side can be seen by using the JEOL provided test application: TestEDSLibIS.exe and the associated DLL: LibbEDSIS.dll. Which can be found in the Probe for EPMA application folder (see below).

Glenn did notice an error message when using the JEOL test app as seen here:



One thought occurs to me for those affected by this issue. Please search your JEOL hard drive and see if you can find more recent versions of the JEOL TestLibEDSIS.exe and LibEDSIS.dll files.

The files we have are dated from 08/17/2020. If you find more recent versions on your JEOL hard drive, please copy both files to the Probe for EPMA application folder, usually C:\Programs Files (x86)\Probe Software\Probe for EPMA, and overwrite the exiting files.

Then try again to acquire an EDS spectrum, and obtain net intensities.

[JonF]

We're unfortunately unable to pursue this now as our 8530 hasn't worked in about 2 years. JEOL service in our region underwent experienced-staffing changes and they have been unable to devote someone knowledgeable to working on this full time.    When we do get the probe up and running again, I hope a solution for the issue at hand will have been found!

Your probe has been offline for two YEARS?!

For the other two instruments, has anything changed on the Probe for EPMA computers? I'm assuming these are the PfE computers that are running the TestLibEDSIS.exe and trying to communicate with the JEOL PCs over LAN (with the updated JEOL software). 

Reason I ask is that I was googling the MSCOMCTL.OCX error that you mention and an interesting post on Stack Overflow (Stack Overflow) suggests this is a problem with a component of the Visual Basic 6.0 IDE that is no longer routinely distributed and needs to be installed and registered manually.

I checked on my JEOL PfE PC which does return EDS net intensities and the MSCOMCTL.OCX file is where it should be.

You can get the MSCOMCTL.OCX file from the (still available) IDE directly from Microsoft here: Microsoft website

And extract the required .OCX file by unzipping the .msi file using 7zip.

You then place the .OCX file in C:/Windows/SystemWOW64/ and register the component via an administrative console with

Code Select Expand

regsvr32 C:\Windows\SysWOW64\MSCOMCTL.OCX 
It could of course also be something to do with one of the dependencies for this file (which might hopefully also be in the .msi file above), or something completely different!