As I know, there are 3 reference energy to find or compare band level when we need to band alignment data
1. fermi energy set to zero
2. Vacuum energy (from slab calculation) set to zero
3. Align to Core orbital energy (which does not contribute to bonding)
I think third method is the most accurate, because core orbital of heavy atom will not participate bonding, then it will maintain its original atom orbital state.
Also I think first one is convenient, and easy to compare band level in same material, but hard to compare difference material's band level
and second one is same mechanism to experimental method, like UPS, but the DOS of slab and bulk is quite differ so I don't convince the result is right.
which one is the best method? and why?