News:

:) You cannot see "members only" boards if you are not a member, so please join the forum!

Main Menu

WDS+EDS mapping

Started by ericwgh, October 15, 2013, 06:07:25 PM

Previous topic - Next topic

ericwgh

Hi John,
I tried running WDS + EDS maps today, using PI. The WDS part worked fine (the "read" spectrometer configuration is somewhat buggy), but I couldn't specify the EDS part. My Noran does work well with PFE for the quant analyses, though. Has the EDS signal not been fully implemented yet, or is there a place in the setup where I can adjust things?

Many thanks
Eric

John Donovan

Hi Eric,
Make sure you have the latest PI download. Lots of improvements.

http://www.probesoftware.com/Update.htm

When you say Noran, do you mean Thermo NSS?
John J. Donovan, Pres. 
(541) 343-3400

"Not Absolutely Certain, Yet Reliable"

ericwgh

Hi John,
our PI version 9.9.1768 is even newer than the newest version you announce on the update page (.1767). ;)
Sorry I wasn't clear about our EDS system. We have Thermo NSS (v.3.0), and it does interface with PFE (also newest version installed) to run quantitative analyses that combine WDS and EDS. That is why I wondered whether that feature was available for Probe Image. The EDS-inputs tab is there, I can check "enable" for EDS1, but I am not sure what do next. Typing in "Si ka" in the textbox produces a black map.

I have a wishlist and other PI-related questions, but I'll post those separately.

Many thanks
Eric

John Donovan

#3
Ah, OK.  I couldn't remember if you had the JEOL EDS or Thermo.

So here's the story: Thermo does not support ROI output from their hardware so you cannot connect any EDS ROIs from Thermo NSS to the JEOL mapping inputs in the JEOL software or Probe Image. Some people keep their JEOL EDS simply to have the old ROI mapping signals and mount the Thermo in the other port for the fancy stuff like spectrum imaging.

Frankly EDS ROIs are pretty old school now that we can acquire full EDS spectrum images. I use the EDS spectrum imaging 99% of the time without WDS, but...

Thermo does now offer a 5 WDS spectrometer input package for their NSS software so you can map WDS channels along with the spectrum image channels. I was the beta tester for this new feature for Thermo and here is an example I acquired on a sample for John Mavrogenes at ANU on some quartz samples from Mark Reed (UofO):



Both element maps are of potassium ka and acquired at the same time. Note that the EDS map has a higher geometric efficiency as the number of photons per pixel is larger, but the WDS map shows the much better sensitivity of the WDS acquisition and can discern the trace element variation easily.

So if you want to map WDS and EDS together I suggest one of two things.

1. Acquire the EDS SI image and save it. Then without removing the sample from the instrument, also acquire WDS maps of the sample tuned for the trace elements of interest on *roughly* the same area, quant it using CalcImage and then correlate the EDS SI and WDS maps sets with my new Hyper-Imaging feature which I should be releasing next week in CalcImage.

2. Or buy the new Thermo 5 input WDS package. I have no idea what it goes for but I'm sure someone at Thermo can help with that.
John J. Donovan, Pres. 
(541) 343-3400

"Not Absolutely Certain, Yet Reliable"

ericwgh

Hi John,
thanks for the plug for the Thermo software. Those are neat images, and illustrate exactly what I am after. I need the WDS sensitivity to resolve fine minor/trace element zoning, like P in olivine, by combining counts from more than 1 spectrometer. The "crude" EDS signal I would use mainly for modal analysis if it's acquired at the same time. Crude, because our 2008 SDD is only 10mm2, bought with nothing other than simple phase-ID in mind.

I am puzzled, however, that the EDS ROIs from Thermo NSS work with the quantitative analysis in PFE, but not with mapping... why is that different? Can you please explain?

Many thanks
Eric


John Donovan

Quote from: ericwgh on October 17, 2013, 12:39:12 AM
thanks for the plug for the Thermo software. Those are neat images, and illustrate exactly what I am after. I need the WDS sensitivity to resolve fine minor/trace element zoning, like P in olivine, by combining counts from more than 1 spectrometer.

There's another solution that I forgot to mention: we sell a switch box and cable set that allows one to connect their WDS spectrometers to the single WDS input on your existing Thermo hardware. Using this switch one can select any single channel or any combination of WDS channels (to improve sensitivity if they are tuned to the same emission line) and these WDS counts are then displayed in your existing NSS spectrum imaging application. Attached below are pictures of the front and back of this box when it was connected to my SX100 (before I switched over to the new 5 channel input hardware from Thermo).

There is another advantage of this switch box: if your sample is not flat and you want to minimize topographical effects in your acquisition, you can select two spectrometers that are opposite each other on the instrument and combine these two signals.  Oh, and here's another advantage: if you want to reduce Bragg defocusing effects in the acquisition (e.g., at low mag), you can combine WDS signals from two WDS spectrometers that are 90 degrees (or so) from each other!

Dang, this is a good idea for the new 5 channel WDS input hardware in NSS. I should talk to Thermo so they can modify their NSS to allow different combinations of WDS inputs channels.  :)

Quote from: ericwgh
The "crude" EDS signal I would use mainly for modal analysis if it's acquired at the same time. Crude, because our 2008 SDD is only 10mm2, bought with nothing other than simple phase-ID in mind.

Well I wouldn't say "crude", just smaller geometric efficiency. I have the same SDD detector on my Sx100, but by using the switchbox described above, you can double, triple, etc your WDS mapping sensitivity!

Quote from: ericwgh
I am puzzled, however, that the EDS ROIs from Thermo NSS work with the quantitative analysis in PFE, but not with mapping... why is that different? Can you please explain?

Yes.

We acquire and store the full EDS spectrum for every WDS analysis (standards and unknowns). When you decide to quant one of more elements by EDS (even if you are off-line for data re-processing), you specify the ROI(s), that is to say, element(s) of interest and and PFE uses the Thermo NSS to extract the net intensities for the specified element(s) (and also the net intensities from the assigned EDS standards for full k-ratio calculations), which are then combined with the WDS element k-ratios (if any) and then they are all cranked through the matrix correction together.

The point being that these ROIs are "software" ROIs as opposed to "hardware" ROIs and so cannot be sent to the JEOL (or Cameca) mapping inputs. However as I mentioned above, you can map using your existing Thermo single WDS channel input (by manually connecting it with a BNC cable), or buy the additional 4 WDS input hardware from Thermo, or buy the 5 channel WDS switch box from Probe Software.

Seems like a reasonable number of choices!   ;)
John J. Donovan, Pres. 
(541) 343-3400

"Not Absolutely Certain, Yet Reliable"

John Donovan

Quote from: ericwgh on October 17, 2013, 12:39:12 AM
I am puzzled, however, that the EDS ROIs from Thermo NSS work with the quantitative analysis in PFE, but not with mapping... why is that different? Can you please explain?

Basically the Regions Of Interest (ROI)s that PFE uses from the stored Thermo full spectra are software ROIs, not hardware ROIs. That is, the user specifies what elements (ROIs) they want the net intensities for, as seen here:



So PFE can quantify EDS elements along with WDS elements for a self-consistent quantitative analysis, but Thermo cannot output TTL pulses for mapping with the JEOL or Cameca EDS beam or stage scan mapping inputs.
John J. Donovan, Pres. 
(541) 343-3400

"Not Absolutely Certain, Yet Reliable"

David Steele

I've just reread the exchange between Eric and John re WD+ED mapping using Probe Image.  Can someone confirm my understanding that combined ED+WD mapping cannot be done using PI with a Thermo EDS in a single pass (on a JEOL 8530F)??

If my understanding is correct (i.e. the answer is no it can't be done as a single pass) I suspect I'm getting into a three pass strategy on our instrument here:

Pass 1 - acquire Panchromatic and spectroscopic CL  with up to 5 WDS elements using the JEOL software (and PCL detector) and (CSIRO) xCLent software.  Optimise beam conditions for CL acquisition.
Pass 2 - acquire EDS elements using NSS+UltraDry SDD (optimise beam conditions for EDS?)
Pass 3 - acquire 5 additional WDS elements (minors/traces) using PI, and crank up the wick for traces....

What file formats can we import into CalcImage??

As long as the JEOL stage is reproducible (LOL!) I don't see too many issues here.....

Thanks and cheers,
David

Probeman

#8
Quote from: David Steele on March 17, 2015, 12:44:12 AM
I've just reread the exchange between Eric and John re WD+ED mapping using Probe Image.  Can someone confirm my understanding that combined ED+WD mapping cannot be done using PI with a Thermo EDS in a single pass (on a JEOL 8530F)??

If my understanding is correct (i.e. the answer is no it can't be done as a single pass) I suspect I'm getting into a three pass strategy on our instrument here:

Pass 1 - acquire Panchromatic and spectroscopic CL  with up to 5 WDS elements using the JEOL software (and PCL detector) and (CSIRO) xCLent software.  Optimise beam conditions for CL acquisition.
Pass 2 - acquire EDS elements using NSS+UltraDry SDD (optimise beam conditions for EDS?)
Pass 3 - acquire 5 additional WDS elements (minors/traces) using PI, and crank up the wick for traces....

Hi David,
Good questions.  There are many ways to approach a characterization problem and the above would be one approach.  Yes, Thermo no longer has EDS ROI outputs for synchronous mapping with EPMA, though Bruker still does.  However, all this is about to change for the better...

To fully integrate EDS spectrum imaging and WDS mapping I am working with Cameca and Thermo to implement what should be a wonderful capability once it is finished as described here:

http://smf.probesoftware.com/index.php?topic=400.msg2174#msg2174

Camea has provided the pixel, line and frame sync signal pinouts already and Thermo already has a hardware sync input for saving the pixel EDS SI data.  I've already written an API for Thermo (and Bruker and JEOL), so hopefully everyone will soon get access to this soon.  I know John Fournelle is working with Cameca and Thermo on this now for his new SXFive, so you could ask him how things are proceeding...

Also check out the hyper imaging feature in CalcImage to correlate separately acquired WDS maps and Thermo EDS SI images:

http://smf.probesoftware.com/index.php?topic=83.0

Quote from: David Steele on March 17, 2015, 12:44:12 AM
What file formats can we import into CalcImage??

Right now just GRD and PrbImg files (attached to .TIF files).  They are the only image files that have the configuration info we need for quant calculations.

And of course the Classify input .DAT files that contain all the quant data in ASCII format for phase classification and now RGB composite imaging.
The only stupid question is the one not asked!

David Steele

#9
Thanks John.

I look forward to the development of the combined WD-ED (full spectrum) mapping capability via PI on our JEOL-Thermo hardware.  Now just getting panchromatic (JEOL) and/or xCLent spectroscopic CL into the mix would be great.... just a thought ;)

Edit by John: If the JEOL panchromatic signal in on one of the analog channels then just select that channel in Probe Image in the Analog Inputs tab...

I had an interesting chat with a senior professor here yesterday which went along HIS (not mine!) "traditional thought" lines of the use the QUT probe for quality quant spot microanalysis and do x-ray mapping on an SEM.  That x-ray mapping is "still" in the qualitative relative intensity realm was implied by his comment.  I pointed out that I would like to see quantitative ED-WD+/-CL integrated mapping used on our EPMA.  "Traditional" inertia to be overcome methinks....

Edit by John:  I think we've all gotten too used to looking at raw intensity images. Show them some quant trace element maps and that should make an impression. For example:

http://smf.probesoftware.com/index.php?topic=393.msg2185#msg2185

Roll on with the intended integrated ED-WD development ASAP please!!  (Who needs sleep...?!?  ;) )

Edit by John: Sleep?  What is this "sleep" you speak of....     ;)

Cheers,
David