You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
One of Euphonic's main jobs is to efficiently convert force constants to phonon modes. This allows e.g. numerical powder averaging over many q-points to be performed given outputs of codes such as CASTEP or Phonopy that provide force constants.
However, the reverse transformation QpointPhononModes -> ForceConstants could also be very useful. A similar transformation from dynamical matrices to to force constants is made in e.g. CASTEP to convert from DFPT results on a Gamma-centered mesh to the force constants of an equivalent supercell. Converting from a non-commensurate lattice would be an interesting but more difficult (i.e. numerical fitting) problem, and is out-of-scope for this Issue.
This ability to convert a set of QpointPhononModes to ForceConstants would be especially useful with external codes like GULP or CRYSTAL which, as far as I can tell, do generate supercell force constants internally but do not expose them to the user in a useful format. With such a code one could specify a mesh (as though constructing a DOS...) and use the results to reconstruct the ForceConstants in Euphonic for further manipulation and interpolation.
One of Euphonic's main jobs is to efficiently convert force constants to phonon modes. This allows e.g. numerical powder averaging over many q-points to be performed given outputs of codes such as CASTEP or Phonopy that provide force constants.
However, the reverse transformation QpointPhononModes -> ForceConstants could also be very useful. A similar transformation from dynamical matrices to to force constants is made in e.g. CASTEP to convert from DFPT results on a Gamma-centered mesh to the force constants of an equivalent supercell. Converting from a non-commensurate lattice would be an interesting but more difficult (i.e. numerical fitting) problem, and is out-of-scope for this Issue.
This ability to convert a set of QpointPhononModes to ForceConstants would be especially useful with external codes like GULP or CRYSTAL which, as far as I can tell, do generate supercell force constants internally but do not expose them to the user in a useful format. With such a code one could specify a mesh (as though constructing a DOS...) and use the results to reconstruct the ForceConstants in Euphonic for further manipulation and interpolation.
This would benefit from input of @DBJochym and @krefson
The text was updated successfully, but these errors were encountered: