[QE-developers] Possible Bug with starting_magnetization tag in QE.7.3.1

[Alexander Duong]

Dear all,

I am Alexander Duong, an undergraduate student at the University of Ottawa. I recently used the Wannier90 input files from Tutorial 18: Iron — Berry curvature, anomalous Hall conductivity and optical conductivity to try to reproduce the Wannier-interpolated band structure of Fe. I did one run using Quantum ESPRESSO 7.2 with Wannier90 3.1.0, and one run using Quantum ESPRESSO v.7.3.1 with Wannier90 3.1.0. The only difference in the runs were the Quantum ESPRESSO versions.

The run with QE.7.2 gave me the expected band structure, but the run in QE.7.3.1 gave me some very unexpected differences. After going through the output, I realized that the self-consistent field run in QE.7.2 was giving a ferromagnetic ground state (as expected for Fe), whereas the self-consistent field run in QE.7.3.1 was giving a non-magnetic ground state. I then tested the starting_magnetization tag. The original value for both runs was  starting_magnetization(1) = -1. I then tried changing it to starting_magnetization(1) = -3 for the run using QE.7.3.1. When I do this, it finds the correct ferromagnetic ground state. Reading the input data description for the starting_magnetization tag, I understand it as the following: if starting_magnetization(1) = -1, that QE.7.3.1 should interpret it the same way as QE.7.2. However, it does not, as shown in the self-consistent field run. Therefore, I think there is either a possible bug with how the code interprets the starting_magnetization tag, or I believe the documentation should be clearer on how QE.7.3.1 interprets the starting_magnetization tag values, since I do notice that the description says "an absolute value greater than or equal to 1 ... will be interpreted as the site's magnetic moment ... Alternatively, the values can range between -1 and 1 ... For QE-v7.2 and older versions, only the second option is allowed." However, clearly, starting_magnetization(1) = -1 is allowed in QE.7.2 even though the absolute value of -1 is equal to 1, and -1 is not strictly between -1 and 1.

Thank you for your time and I look forward to hearing from you soon.

Best,
Alexander Duong (University of Ottawa)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/developers/attachments/20250228/6700da6b/attachment.html>