This workflow is designed to help you get started, or improve your workflow using ModeMap to analyse potential energy surfaces of phonon modes.
- You must have calculated a converged FORCE_SETS from phonopy previously, see phonopy for help.
- You must have
ModeMap.py
,ExtractTotalEnergies.py
,ModeMap_PostProcess.py
, andModeMap_PolynomialFit.py
in your main directory, see ModeMap.
A typical workflow looks like this:
Note: for the --q-range
argument I meant start end step, not start end stop
Using the scripts it gives a five step worklow - each step is highlighted on the mind map:
structures.sh
runsModeMap.py
for all the MODEs of interest and untar MODEMAP.tar.gz, which contains all the MPOSCARs.singlepoint_norun.sh
andsinglepoint_run.sh
distributes all MPOSCARs into single directories with INCAR, KPOINTS, POTCAR, and job script files. I recommend starting withsinglepoint_norun.sh
and then test that calculations are well converged on equilibrium and extreme structures before usingrun.sh
, which is specifically designed to run single point calculations for all structures apart from the equilibrium and extreme ones.energies.sh
runsExtractTotalEnergies.py
for all MODEs and cleans irrelevant files from all sub-directories.postprocess.sh
runsModeMap_PostProcess.py
for all MODEs and creates the directory Results, which stores the potential energy surface for each mode.fitting.sh
runsModeMap_PolyFit.py
for all MODEs.
The relationship between Q and displacement depends on the atomic mass. Therefore, keep in mind that a small value for Q might push light atoms very far and give something unphysical, where it would be fine for heavier atoms.
For reference, I have also attached typical annotated INCAR and KPOINTS files used for a single point calculation. The most important aspect of the INCAR is NSW = 0
to avoid relaxing the structure, and LREAL = .FALSE.
for accurate forces.
- Phonopy can be used to further investigate the bottom well structure of a soft mode, by looking at its symmetry for instance. Simply use the corresponding MPOSCAR. This can shed light on a potential displacive phase transition.
- ISODISTORT can be used to link the top and bottm well structures. You need to convert the corresponding MPOSCARs to .cif files that contain the symmetry elements for each structure. ISODISTORT is a great tool in general to explore the connection between structures.
- Bilbao Crystallographic Server to find k-point coordinates of a given space group. Make sure to use the CDML coefficients, not the ITA coordinates.
- VESTA to investigate MPOSCARs and verify that structures make sense and that we have not pushed too far.
- For more on displacive phase transitions, see Martin T. Dove's Theory of Displacive Phase Transitions in Minerals (1997).