Quantum espresso/EPW output

[ollehellman]

Added an output option to print the second order force constants in EPW/QE XML format. This is useful if you want to do electron-phonon calculations in EPW with TDEP phonons.

[Guodonglin-cqust]

Hello Professor, regarding the 2D materials, the TEDP output for the force constants in Quantum Espresso has an issue with incorrect q-points. For example, in the case of graphene, using a 10x10x1 supercell for calculations, the expected q-point output should be 12x12x1. However, the q-points for the force constants are 9x9x7. I don’t understand why the q-point in the z-direction is 7 instead of 1.

[ollehellman]

Hello Professor, regarding the 2D materials, the TEDP output for the force constants in Quantum Espresso has an issue with incorrect q-points. For example, in the case of graphene, using a 10x10x1 supercell for calculations, the expected q-point output should be 12x12x1. However, the q-points for the force constants are 9x9x7. I don’t understand why the q-point in the z-direction is 7 instead of 1.

Do you have some example input files? I guess some kind of --twodimensional flag needs to be added, but in principle it should work anyway as long as the Z-axis is longer than the 12x12 in-plane supercell. Is that the case?

[Guodonglin-cqust]

300.zip
This is input files of graphene .

[ollehellman]

300.zip This is input files of graphene .

Just to be clear, can you plot the dispersion from tdep and the dispersions calculated by the converted ifc from epw/qe so that I can see that they really are different?

[Guodonglin-cqust]

There is no difference, except that EPW cannot read it and reports an error due to incorrect q-points. Based on a 10x10x1 q-point grid, the dyn and force constants were calculated using QE-ph. If TDEP is used to generate the QE force constants, because the q-point grid is not 10x10x1 but 9x9x7, EPW encounters an error when reading the force constants.

[ollehellman]

There is no difference, except that EPW cannot read it and reports an error due to incorrect q-points. Based on a 10x10x1 q-point grid, the dyn and force constants were calculated using QE-ph. If TDEP is used to generate the QE force constants, because the q-point grid is not 10x10x1 but 9x9x7, EPW encounters an error when reading the force constants.

Is that when using lifc = .true. in the EPW input? Alternatively, if you manually change the q-grid in the xml file, what happens? (since the IFCs are in realspace the q-grid should be irrevant in theory, not sure about in practice)

[Guodonglin-cqust]

After manually editing the force constant file produced by TDEP, I encountered problems in the epw.out output, as many tags like s_s1_m1_m2_m3 were missing.

 Reading interatomic force constants

end of file reached, tag s_s1_m1_m2_m3.1.1.10.1.1 not found
end of file reached, tag s_s1_m1_m2_m3.1.1.1.2.1 not found
end of file reached, tag s_s1_m1_m2_m3.1.1.2.2.1 not found
end of file reached, closing tag not found
end of file reached, tag s_s1_m1_m2_m3.1.1.3.2.1 not found
end of file reached, closing tag not found
end of file reached, tag s_s1_m1_m2_m3.1.1.4.2.1 not found
end of file reached, closing tag not found
end of file reached, tag s_s1_m1_m2_m3.1.1.5.2.1 not found
end of file reached, closing tag not found
end of file reached, tag s_s1_m1_m2_m3.1.1.6.2.1 not found
output.zip

[ollehellman]

Printing ifcs in EPW format, structures in QE format