Correctness of SOC

[fanchem]

The authors of another software pointed out that It is important to keep in mind that the wavefunctions are spinors in non-collinear calcu- lations, so the symmetry operators D{S|t} should act both on the spatial coordinate and in spin space. Neglecting the rotation in spin space would lead to significant errors in the computed e-ph matrix elements, especially in calcu- lations with SOC. And they mentioned in EPW symmetry operators do not act on the spin space.

Is this the case in EPW? If so, can we still trust the calculations with SOC?

[Josh Leveillee]

The action of the symmetry operator on spin space in EPW has been fixed by including the proper SU2 transformation on the spinors. The version of EPW containing this fix is the one found on the Quantum Espresso Foundation (https://gitlab.com/QEF/q-e).

However, even without the fix for spin-space, the errors expected for individual matrix elements are small. In Si, the error expected is on the order of 10^-5 meV, and for Pb the error was on the order of 10^-2 meV. Calculated results, such as the Eliashberg function of Pb, seem to remain essentially the same with or without the spin-space rotation. Nonetheless, spin-space rotations are now included in the development version of EPW mentioned above. Cheers!

-Josh

----
Joshua Leveillee
joshua.leveillee@gmail.com
Postdoctoral Researcher
Giustino Research Group, UT Austin