Peak-fitting a Single O (1s) Peak – Easy !!!

 

 

Peak-fitting a Single O (1s) Peak should be Easy !?

 

 

 

 

Step 1.  Add a Background with Useful Endpoints – Sherwood Background tested. Same as Iterated Shirley BG.

 

 

Test 1.  Use 1 synthetic peak and test for optimum Gaussian-Lorentzian peak-shape (test 75-90% G)

Check Chi-Square.  Check for Gaps at Top or Baseline

 

 

Test 2.  Use 2 synthetic peaks.  Testing for optimum Gaussian-Lorentzian peak-shape.  Test for 3rd peak!

Check Chi-Square.  Check for Gaps at Top or Baseline

 

 

Test 3.  Use 3 synthetic peaks with optimum Gaussian-Lorentzian sum peak-shape (90% G : 10% L)

Check Chi-Square.  Check for Gaps at Top or Baseline

 

 

Test 4.  Use 3 synthetic peaks – Testing effect of larger peaks at ends producing smaller chemical shifts

Check Chi-Square.  Check for Gaps at Top or Baseline

 

 

Test 5.  Use 3 synthetic peaks – Testing effect of larger peaks at ends with very small chemical shifts

Check Chi-Square.  Check for Gaps at Top or Baseline

 

 

Test 6.  Use 3 synthetic peaks – Testing effect of Constraining all FWHM to be 1.20 eV or 1.00 eV

Check Chi-Square.  Check for Gaps at Top or Baseline

 

 

 

Test 7.  Use 3 synthetic peaks – Testing effect of Constraining all FWHM to be 1.00 eV and Use 30% Lorentzian!

and

Test 8.  Use 5 synthetic peaks – Testing effect of Constraining all FWHM to be 1.20 eV and Use 15% Lorentzian

Chi-square?

 

PEAK-FIT OPTIMIZATION TESTS FINISHED.  Easy?

Please read the list of Questions that follows this Summary Slide

 

 

 

 

Question Set #1:

Based on chemical states expected which peak-fit is best ?

Based on typical FWHM which peak-fit is best ?

Based on typical chemical shifts which peak-fit is best ?

Based on chi-square which peak-fit is best ?

Based on gaps in peak area which peak-fit is best ?

Based on Gaussian:Lorentzian Sum (GLS) ratio which peak-fit is best ?

Question Set #2:

What are the “optimum” FWHMs values to use?

What is the “optimum” Gaussian-Lorentzian ratio to use?

What chemical shifts are reasonable?

What chemical states do you think are present?

How many peaks are reasonable?

How many synthetic peaks will you use to fit this spectrum of O (1s)?