Author Topic: Strip peak from spectra?  (Read 138 times)

Dan Ruscitto

  • Professor
  • ****
  • Posts: 72
    • GE Research Materials Characterization
Strip peak from spectra?
« on: July 25, 2022, 06:35:40 AM »
Hi all-
I'm new to DTSA-II but I'm trying to implement it in some tricky EDS quant we have here. Is there a straight-forward command to strip a peak from a spectrum? I have a Si3N4 standard with some carbon contamination and the software is not liking my choice of Si3N4 for a standard to quantify an unknown (i think it can't fit a proper background for N Ka) -- is there a straight-forward way i can strip out the carbon peak in the software (i have a graphite spectrum) and trick the program just for testing purposes before I go ahead and use a plasma asher to remove any carbon contamination present and re-aquire?
-Dan

NicholasRitchie

  • Moderator
  • Professor
  • *****
  • Posts: 107
    • NIST DTSA-II
Re: Strip peak from spectra?
« Reply #1 on: July 26, 2022, 06:00:16 AM »
There are a couple different ways to deal with this.
  • Brute force:  Select an ROI including the C peak on the spectrum display than use the Process => Trim menu to remove the C peak.  Make sure the composition of the resulting spectrum is listed as Si3N4 with no coating. You may have to select slightly less of the peak width than seems optimal since the algorithm uses background ROIs outside of the selected region.
  • More elegant (but more difficult and can fail if the spectrum fit isn't optimal):
    • Save the spectrum to disk.
    • Use the Tools menu => Calibration Alien menu item => "Calibrate using elemental reference" option
    • On the first page, select the saved spectrum and specify a material like "0.99*Si3N4+0.01*C" that includes C
    • Perform the fit and then on the "Fit results" page, check "Output elemental fits" and uncheck "Add to database", select "Finish"
    • From the command line, "> ls()" to list available spectra.  I identify the "s??" name of the original spectrum and the carbon elemental fit. Let's say "s8" and "s12" for the original and c fit.
    • From the command line "> display(s8-s12)" which will calculate the difference of the original spectrum and the carbon fit and display it.
    • Fix the "live time" and "standard composition" for the resulting spectrum using the "Spectrum Properties" dialog and the "Assign material" dialog.
"Do what you can, with what you have, where you are"
  - Teddy Roosevelt