Spatial resolution of EDS versus WDS

[Michael Lance]

Hello all,

I recently collected two maps of an Ni-base superalloy using EDS (Pathfinder) and WDS. I used identical settings; 12 kV, 50 nA, 0.5 sec dwell time, 0.24 um pixel size and 256x192 pixels.

As expected, WDS was superior with the trace and light elements.

However, WDS also had better spatial resolution for the heavy elements (Cr, Fe, Ni) than the EDS map. Why might this be? Is it due to the better signal to noise ratio of WDS over EDS?

Thanks,

Michael

[Probeman]

Hello all,

I recently collected two maps of an Ni-base superalloy using EDS (Pathfinder) and WDS. I used identical settings; 12 kV, 50 nA, 0.5 sec dwell time, 0.24 um pixel size and 256x192 pixels.

As expected, WDS was superior with the trace and light elements.

However, WDS also had better spatial resolution for the heavy elements (Cr, Fe, Ni) than the EDS map. Why might this be? Is it due to the better signal to noise ratio of WDS over EDS?

Thanks,

Michael

Hi Michael,
This sounds quite interesting. Can you post an example of two maps so we can see the differences you are talking about?
john

[Michael Lance]

John,

I've attached an image showing both the EDS and WDS map of Chromium in this alloy.



-Michael

Edit by Donovan. Added image to image gallery and pasted link to post. TIF files can be problematic for "preview".

[BenjaminWade]

Possibly due to the different take-off angles of your WD to your EDS? "Shallower" take-off for EDS so longer path length/absorption compared to WDS manifest as more "out-of-focus" EDS map? No idea really, just a guess.

[Probe321]

The WDX counts X-rays at a specific wave length where the EDX counts X-rays over a energy range (energy window).  The Oxford AZTEC software can adjust the with of the energy range from the default setting. Adjusting the energy range to one channel may make the EDS map look similar to the WDX map. Keep in mind all the other variables that impact the analysis like location of detectors, detector distance from sample. pulse processor setting of the EDS detector etc.  Be interesting to see what the maps look like when reducing the EDS energy window size.

[Michael Lance]

I tried changing the energy range to one energy (I have Thermo Pathfinder) and it didn't make a difference for the spatial resolution.

I collected this EDS map using rate 5 which results in broader peaks than rate 1 so that may be a factor as you point out. I will have to try some more experimenting. I asked the question here because I thought maybe some of you have made a similar observation.

[Mike Matthews]

Is it definitely a different resolution and not a lower signal to noise on the ED map making it appear more fuzzy?

[Probeman]

It would help to see a cross section of both maps to look at this statistical noise vs. spatial resolution question.

[Philippe Pinard]

Besides what has already been suggested, I would check whether there is no image filter used in displaying the EDS map (e.g. smoothing or interpolation). Based on the image title, both maps are from quantified data, converted to at%? Are both using the Cr Ka?