Hello,
In XPS measurements, I am observing consistent shifts in spin-orbit splitting values for different oxides of the same base element. I'm being told by many that this is not possible, that the spin-orbit split is an intrinsic property, and is therefore an error with my data collection method or data fitting. I'm fairly confident in my peak positions, but less confident in charge neutralization (since I didn't collect all of this data myself I don't have a solid record). Regardless, I would assume that components of the same spectra (i.e. 2p(1/2) & 2p(3/2)) should shift linearly for a charged surface. As a side note, all pass energies and collection step sizes have been the same throughout these measurements.
Is it possible to have different spin-orbit splitting values for various compounds or stochiometric variations of a compound containing the same element? (For example A0 v. AX2 v. A2X3 )
One of my thoughts is that magnetic domains might have a localized Zeeman effect on spin-orbit split components. Some folks I work with have done XPS measurements on superconductors before and after Tc , and they didn't observe a shift in spin-orbit split values. That being said, they aren't convinced what we are seeing could be related to ordering or magnetism. It just makes sense to me that spin-orbit split components can be shifted just by their crystallography or magnetic structure, by how much I just don't know. The shifts I see range from 0.2 - 0.5 eV between compounds.
Would anyone agree with this guess, if not, could you provide a physical justification for what I'm seeing? I'm fine with any suggestions on the instrument setup or physicality of a materials system.
Thank you.
Dear Keith White
Yes, it is possible. While the spin-orbit coupling is to a large extent an intrinsic effect, in a solid it can be modified by the underlying stoichometry. This modification happens through the orbital angular momentum (L) of electrons participating in chemical bonding. Since the wave function of valence electrons is distributed both within and outside the core regions, L can in principle have both intrinsic (from the core region) and extrinsic (from valence region) contributions. In solid, usually due to the so-called angular momentum quenching the latter is negligibly small. As such, people tend to consider the spin-orbit coupling L.S as an intrinsic effect. However, in some cases where the electrons are from a particular orbital manifold such as d-t2g and the crystal has an appropriate symmetry (like C4), the extrinsic part of L can remain unquenched, resulting in a significant enhancement of the spin-orbit splitting. Transition metal oxide perovskites are a prime example in which this phenomenon occurs.
Dear Keith,
As the previous answers pointed out, this is possible, since the spin-orbit coupling is dependent on the effective nuclear charge Z, which is shielded by the nuclear
charge corrected by electron shielding. The shielding is dependent on the oxidation state, i.e. spin-orbit-coupling increases with the oxidation state due to the reduction of shielding and thus an increase of the effective nuclear charge.
Appended you will find some data on Manganese compounds, which is able to adopt a large variety of oxidation states. According to literature, a similar trend was observed for Cr compounds too.
All the best,
Thomas
- Badges
- Science method
- Similar topics
- Methodology
- Crystallography