Author Topic: theoretical peak position  (Read 3820 times)

Ben Buse

  • Professor
  • ****
  • Posts: 488
theoretical peak position
« on: September 21, 2016, 06:00:13 AM »
Hi,

When you add an element and it loads the theoretical peak position say for Si of 77.674 - where does it get this from, is it a table or an equation? Is it possible to change it to the actual peak positions?

I notice in the manual there is a peak calibration centre in startwin, but I loaded startwin and this option was greyed out - is it only possible on certain instruments?

Thanks

Ben
« Last Edit: September 21, 2016, 08:25:08 AM by John Donovan »

Probeman

  • Emeritus
  • *****
  • Posts: 2823
  • Never sleeps...
    • John Donovan
Re: theoretical peak position
« Reply #1 on: September 21, 2016, 12:37:33 PM »
When you add an element and it loads the theoretical peak position say for Si of 77.674 - where does it get this from, is it a table or an equation? Is it possible to change it to the actual peak positions?

Hi Ben,
I must not be following you exactly because I know that you know, that you can enter any peak position simply by typing it in...  Yes, if you change an element/x-ray or spectrometer crystal, the program reloads the default on-peak to the value calculated from the XLINE.DAT table for this crystal. 

This emission line table is visible from the CalcZAF app and the values can be edited in CalcZAF if you want to change the default emission line values. Note that because the CalcZAF/PFE installer tags this file (and the XEDGE.DAT and XFLUR.DAT files) as NeverOverwrite="yes" and Permanent="yes", they will never get overwritten or removed by the installer.

Now of course since even the theoretical peak positions will vary from spectrometer to spectrometer due to the spectrometer mechanical calibration, I almost *never* utilize the default peak positions and instead simply load previously peaked emission lines from the SETUP.MDB element setup database by using the Load Element Setup button in the New Sample Setup button or even easier, by loading an entire sample setup using the Load File Setup from a previous run where all the elements are peaked and the backgrounds already checked. See here for more information:

http://probesoftware.com/smf/index.php?topic=262.msg1242#msg1242

I notice in the manual there is a peak calibration centre in startwin, but I loaded startwin and this option was greyed out - is it only possible on certain instruments?

Yes, the Startwin app has a feature to perform spectrometer calibrations over the range of each spectrometer (and crystal combination) using multiple elements on multiple standards. The description is in the PFE User Reference and I quote it here:

Quote
The Calibrate Peak Center option allows the user to specify a number of elements and corresponding standards for the program to perform multiple peak calibrations. After the peak centering has been performed the program will save the theoretical spectrometer positions along with the actual measured peak positions and a least squares fit to the data for the purpose of calculating a variable offset function which in turn is used to calculate the actual peak position of any element on that spectrometer/crystal combination.

There is a lot of details on this feature starting at page 105 or so.  See attached below for some calibration results from my SX51 at UC Berkeley (remember, you need to be logged in to see attachments).

Basically the Peak Calibration feature in StartWin is a method to calibrate one's spectrometers *without* a mechanical recalibration. Since JEOL spectrometers require more frequent mechanical recalibration than Cameca spectrometers, I'm not sure it's worth the trouble for JEOL instruments. But some may disagree.

The feature should be enabled and allows one to specify up to 5 element/x-ray combinations per spectrometer/crystal combination for spectrometer calibration.  If these calibrations are saved to the PFE peak calibration files (e.g., PROBEWIN-KA.CAL, PROBEWIN-KB.CAL, etc.), then they can be utilized when entering in a new element in the Elements/Cations dialog.

I haven't used this feature in ages, but Paul Carpenter was a big fan of this, at least at one time. We really should have someone write up a nice tutorial on it!  Interested?

OK, I checked and I see that you have to have the UseMultiplePeakCalibrationOffset flag in the Probewin.ini file [software] section set to a non-zero value to enable the Peak Calibration button.
john
« Last Edit: September 21, 2016, 01:09:27 PM by Probeman »
The only stupid question is the one not asked!

Ben Buse

  • Professor
  • ****
  • Posts: 488
Re: theoretical peak position
« Reply #2 on: September 22, 2016, 01:24:41 AM »
Hi John,

Thanks, I thought I might give calibrate peak centre a go - for when setting up a recipe/analysis from scratch since I have the peak search windows narrow - it often has to do a coarse scan, which is fine just takes a little bit longer. As you say one way is loading the latest peak up from an mdb file.

It be interesting to see how long the peak positions hold for.

Ben

Anette von der Handt

  • Global Moderator
  • Professor
  • *****
  • Posts: 351
    • UMN Probelab
Re: theoretical peak position
« Reply #3 on: September 22, 2016, 01:54:57 PM »
One really nice tool to investigate the short/long term fluctuations in peak positions (or standard intensities) is "Drift" under the PFE utilities. You can use it to mine data from your old mdb's (so if you are a long-time user of PfE, you have a nice data set). You can set some filters for the data to, for example only get Si ka data at 15kV on the Kakanui Hornblende and Drift creates a spreadsheet with peak position, intensities etc data. I run one standard (the Kakanui Hbl) for that reason more or less in all calibrations.

It is fun to cross-correlate the data then with service visits, room temperature data etc and it also provides a nice baseline for a machine in generell. Maybe a good way to explore how much drift you can expect.
Against the dark, a tall white fountain played.

Probeman

  • Emeritus
  • *****
  • Posts: 2823
  • Never sleeps...
    • John Donovan
Re: theoretical peak position
« Reply #4 on: September 22, 2016, 10:46:58 PM »
One really nice tool to investigate the short/long term fluctuations in peak positions (or standard intensities) is "Drift" under the PFE utilities. You can use it to mine data from your old mdb's (so if you are a long-time user of PfE, you have a nice data set). You can set some filters for the data to, for example only get Si ka data at 15kV on the Kakanui Hornblende and Drift creates a spreadsheet with peak position, intensities etc data. I run one standard (the Kakanui Hbl) for that reason more or less in all calibrations.

It is fun to cross-correlate the data then with service visits, room temperature data etc and it also provides a nice baseline for a machine in generell. Maybe a good way to explore how much drift you can expect.

Hi Anette,
You are totally correct.

I don't make as much use of the Drift application as much as I should.  It does provide excellent long term monitoring of standard, spectrometer and analyzing crystal "health".

A brief mention of the Drift app can be found here:

http://probesoftware.com/smf/index.php?topic=575.msg3565#msg3565

john
The only stupid question is the one not asked!

Ben Buse

  • Professor
  • ****
  • Posts: 488
Re: theoretical peak position
« Reply #5 on: October 12, 2016, 08:17:01 AM »
All spectrometers calibrated using StartWin and working well!

And thanks Anette I've just looked at Drift.exe and it looks like a good way to monitor the situation (and assess the stability of the spectrometers)

Ben
« Last Edit: October 12, 2016, 09:23:13 AM by John Donovan »

Probeman

  • Emeritus
  • *****
  • Posts: 2823
  • Never sleeps...
    • John Donovan
Re: theoretical peak position
« Reply #6 on: October 12, 2016, 11:09:52 AM »
All spectrometers calibrated using StartWin and working well!

Hi Ben,
Very good.

Remember, when you need to "touch up" your spectrometer calibrations, all the element and standard parameters should be saved so you can just run it again easily.

Feel free to share some plots like this with us:

http://probesoftware.com/smf/index.php?topic=797.msg5015#msg5015

It would be interesting to see how linear the JEOL specrometers are...
john
The only stupid question is the one not asked!

Probeman

  • Emeritus
  • *****
  • Posts: 2823
  • Never sleeps...
    • John Donovan
Re: theoretical peak position
« Reply #7 on: October 12, 2016, 11:43:08 AM »
And thanks Anette I've just looked at Drift.exe and it looks like a good way to monitor the situation (and assess the stability of the spectrometers)

Here is a post describing what the Drift app is useful for:

http://probesoftware.com/smf/index.php?topic=575.msg3565#msg3565

john
The only stupid question is the one not asked!

Ben Buse

  • Professor
  • ****
  • Posts: 488
Re: theoretical peak position
« Reply #8 on: October 17, 2016, 05:34:56 AM »
Hi

Some example plots for Ka.

« Last Edit: July 22, 2020, 05:19:32 AM by Ben Buse »