The underlying physics behind quantization axes for the electrons spin has never really been clear to me. I saw the Pauli matrices in my quantum courses, but the definition of different principle axes seems arbitrary.
In the case of a magnetic material, or spins in the presence of a magnetic field, there is a clear axis along which to define the spin. But what about when there is no magnetic field?
For instance, in materials with strong spin-orbit coupling electron wavefunctions are mixed between up and down spin polarization. i.e. spin polarized states are not eigenstates. Furthermore, this paper about Berry curvature in SOC systems says, " it cannot be assumed that their local spin quantization axes are antiparallel, as the interactions with neighboring bands may affect the direction of quantization" (DOI: 10.1103/PhysRevB.72.085110)
But what does up and down mean in this context? Can quantization axes be defined differently depending on the properties of the band structure?