[QE-users] projwfc.x output for magnetic anisotropy calculations using QE 6.7.0

[Amar Kumar]

Dear QE users,

I am performing calculations for atomic and orbital resolved magnetic 
anisotropy energy (MAE). I used projwfc.x with spin-orbit coupling (SOC) 
and obtained the following output file format for a single magnetic 
moment orientation (e.g., in-plane).

Atom          1 Fe

... s_up, s_down

eband_l_m_atom   0   1   1    -0.8276049837E+02  -0.8361757118E+02

... s_up, s_down

eband_l_m_atom   0   1   1    -0.2608445926E+01  -0.2705318778E+01

... {p_up}, {p_down}

eband_l_m_atom   1   1   1    -0.4787349238E+02  -0.4864604079E+02

... (output continues for other orbitals)

eband_l_m_atom (eV) =        1           -0.5103815297E+03

I have a few questions regarding this output:

(i) Meaning of Last Two Columns: What do the last and second-to-last 
columns represent? Are they the total band energy (eband_tot) and 
projected total band energy (eband_proj_tot) in eV, respectively?

(ii) Fermi Energy Subtraction: Is it necessary to subtract the Fermi 
energy from all eigenvalues before calculating the MAE decomposition? Or 
is the Fermi energy already subtracted in the output?

(iii) s-orbital Contribution: When comparing orbital energies between 
two different magnetic moment orientations (e.g., phi=0 and phi=90) to 
obtain the MAE, the s-orbital also seems to contribute with a non-zero 
value. This is unexpected because the s-orbital shouldn't contribute to 
MAE, especially due to SOC. This behavior persists even after 
subtracting the Fermi energy.

Kindly help for the same, what am I doing wrong? Apologize for such 
basic questions.

Thank you for your time and assistance.


AMAR KUMAR
Ph.D. student
Indian Institute of Technology, Delhi