Measuring Aluminum

[knshugart]

I'm interested in an accurate measurement of Aluminum in a Cobalt alloy. We've run an EPMA scan in our lab, had someone else run an EPMA scan in their lab, and run an EDS scan. All three come up sufficiently different that I can can't trust the value. I think the largest issue may be in the way I am measuring the standard for Aluminum. I was wondering if someone could give me step by step instructions on their best practices for Al. (As a note, I'm on a Cameca SX-100 using the OIM software.)

Thoughts? Suggestions?

[Ben Buse]

What do you mean by a EPMA scan. - are you quantifying using wavescans.

Normally you peak up on Aluminium metal standard, measure pk counts, and bkg counts. (Wavescan may be useful to identify bkg positions on standard and unknown). Then measure pk counts and bkg counts on unknown sample (AlCo alloy).

Things to check
Is you Aluminum metal standard heavily oxidised?

Also their may be a matrix correction problem (once you've done all the measures correctly) - do you have AlCo alloy standards.
Have a look at the ZAF values - are the matrix corrections large. If the matrix corrections are large you could try different phi-rho-z model or MAC.

Here's an example of matrix correction using calczaf for alloy with 50% wt.% Al and 50% Co



You can see Al has a absorption correction of 1.5 and Z correction of 1.4

Ben

[Probeman]

I'm interested in an accurate measurement of Aluminum in a Cobalt alloy. We've run an EPMA scan in our lab, had someone else run an EPMA scan in their lab, and run an EDS scan. All three come up sufficiently different that I can can't trust the value. I think the largest issue may be in the way I am measuring the standard for Aluminum. I was wondering if someone could give me step by step instructions on their best practices for Al. (As a note, I'm on a Cameca SX-100 using the OIM software.)

Thoughts? Suggestions?

Ben already anticipated most of my answer!    So this is a trace Al measured in a Co matrix?

As Ben said, if I were you I'd start by entering the nominal composition into CalcZAF (it's a free download), and see what the correction factors look like.  You can also enter the composition into the Standard database (comes with CalcZAF), and have it calculate the nominal spectral interferences.

I'd then perform a wavescan on all the measured elements and check the backgrounds for any weirdness.  Feel free to plot some wavescans here and we can look at them together.