After 3 days the Nickel surface collected only 5.3 atom% of Oxygen. Keep this in mind as you study this Peak-fit Model.

Shirley type background is highly recommended, but is it the correct one to use for all peak-fits?

Two types of background baselines are commonly used.  The questions are:  Which one should we use and why?

The linear region (868-862 eV) between these two Ni signals suggests that it is reasonable to peak-fit the stronger (2p3) signal by itself.

There is no obvious reason why the peak-fit should include both the Ni (2p3) and the Ni (2p1) signals.

Based on this we thinking we focus on the Ni (2p3) signal by itself.

One Question:
Is the peak at ~858 eV due to some sort of oxidized Nickel?

The smaller range including only the Ni (2p3) signal was used for this study. This is the raw data.

Based on a simple look at the spectrum, we envision the need for 3 or more peaks.

This is an assumption.

The Linear background type is black, and the Shirley curved type background is dotted red.

One question is:  Which background type should we use to fit this spectrum?

In this example, the data-analyst tested peak-fits while using both types of baselines.

For this test, the data-analyst decided to use the Shirley type (S-shaped) background-baseline.

This resulting peak-fit has 3 peaks with various FWHM values and 2 different G:L peak-shape ratios.

The main peak has a FWHM of 1.34 eV and is fit with a G:L ratio of 65:35 and 20% asymmetry in the tail.

The synthetic peak at ~858 eV was chosen because there is an obvious peak at 858 eV.  The FWHM was chosen by using the half-height of the far left slope. The resulting peak has a FWHM of 2.87 eV.

The remaining peak area between the two obvious peaks was given a synthetic peak located at ~855 eV and a large FWHM, 2.52 eV.

The peak-fit iteration was started, and the final result is displayed here.  The resulting Chi-square is 5.7.

The question is:  Are there any peaks that belong to oxidized Nickel? Remember that the surface has 5.3 atom % Oxygen.

In this test fit, the data-analyst used 7 synthetic peaks to fit the raw spectrum.

This peak-fit result has 7 peaks with various FWHM values while all G:L peak-shape ratios are the same:  80:20

The main peak has a FWHM of 1.17 eV and is fit with a G:L ratio of 80:20 with no  asymmetry in the tail.

The data-analyst used the same 1.17 eV FWHM to add 5 more synthetic peaks. The last peak was allowed to be broader.

The peak-fit iteration was started, and the final result is displayed here.  The resulting Chi-square is also 5.7.

The question is:  Are there any peaks that belong to oxidized Nickel? Remember that the surface has 5.3 atom % Oxygen.

As we look at these two peak-fits, the question is:

Which peak(s) belong to oxidized Nickel? Remember that the surface has 5.3 atom % Oxygen.

If we have reference spectra from pure NiO, Ni(OH)2, NiCO3, they might help us, but we do not have those spectra. We must use logic and our knowledge of empirical formulae to guide our choice.

One way of thinking is that each extra peak is due to oxidized Nickel.  The left peak-fit would then have 2 types of oxidized Nickel. The peak-fit on the right would have 6 types of oxidized Nickel. If we apply our knowledge of chemistry, we can make a good choice.

Which peak-fit makes logical sense?

The peak at ~858 eV is actually an unusual phenomenon involving a 2 electron photoemission.

The low level of Oxygen might be physi-sorbed Oxygen, or perhaps a C-O species, such as C=O that is physi-sorbed.

If we had copies of the O (1s) and C (1s) spectra, then those spectra might help us to choose the better baseline.

With so little oxygen present, and the odds of having 6 types of oxidized Nickel based on the linear baseline, the more logical choice, based on chemistry, is the Shirley.  The extra wide FWHM of the 2 smaller peaks then requires an explanation, and more thinking.

Iron (Fe)

Background Type – Shirley or Linear – Iron (Fe)