Spin-orbit doublet fitting in XPS
p, d and f core levels split into two peaks by spin-orbit coupling, and the splitting and area ratio are physical constants — not free parameters. Fitting them as independent free peaks is one of the most common XPS mistakes: the optimizer, blind to the doublet structure, parks broad “ghost” peaks wherever they reduce χ².
The constants to fix
| Orbital | Components | Area ratio |
|---|---|---|
| p | p3/2 : p1/2 | 2 : 1 |
| d | d5/2 : d3/2 | 3 : 2 |
| f | f7/2 : f5/2 | 4 : 3 |
The splitting (eV) is element- and orbital-specific (e.g. S 2p 1.18, Mo 3d 3.15, Pt 4f 3.33).
How to do it in PikaXPS
In the peak table, set the peak’s Type to Doublet. PikaXPS then creates the partner with:
- partner centre = main + splitting (fixed),
- partner area = main × ratio (fixed),
- partner FWHM = main FWHM (linked).
If PikaXPS recognises the region (name it like “Ir 4f”), the splitting and ratio fill automatically; otherwise you enter them once. The same-species FWHMs stay linked, so the optimizer only fits what physics leaves free.
When the ratio looks off
Some metals (e.g. Ti) show ratios slightly off the theoretical value because of Coster–Kronig effects — start from the theoretical value anyway, and only release a constraint with a clear physical reason. The fit auditor will tell you if a freed ratio has drifted too far.
Comments & questions