Interpolation Calculation
These variables control the interpolation calculations.
- Type:
STRING - Default:
ph_save - Description: Directory path containing the phonon save data to be read.
- Type:
STRING - Default:
ph_save_interpolation - Description: Directory path where the interpolation results will be written.
- Type:
LOGICAL - Default:
True - Description: If True, the code interpolates the deformation potential (\(\delta V_{scf}\)).
- Type:
LOGICAL - Default:
False - Description:
If
True, no symmetries are used; i.e., dynamical matrices and dvscf are printed on the full Brillouin zone.
- Type:
STRING - Default:
no - Description:
Specifies the type of Acoustic Sum Rule to apply for force constant matrix.
no: No acoustic sum rule applied.simple: Simple acoustic sum rule (corrects only the diagonal matrix elements).crystal: Translational acoustic sum rule imposed via orthogonal projection.all: Translational and rotational acoustic sum rules imposed via orthogonal projection.all_huang: (recommended) Translational and rotational sum rules, plus Huang invariances, imposed via orthogonal projection.
- Type:
STRING - Default:
no - Description:
Specifies the type of Acoustic Sum Rule to apply for Born charges.
no: No acoustic sum rule applied.simple: Simple acoustic sum rule (corrects only the diagonal matrix elements).crystal: Acoustic sum rule imposed via orthogonal projection.
- Type:
LOGICAL - Default:
False -
Description: If
True, applies non-analytic (LO-TO) splitting corrections to the dynamical matrices at \(\Gamma\) point.It should be noted that this flag does not completely switch off LO-TO treatment. It is only intended for switch on/off LO/TO treatment at the Gamma point. If you wish to completely disable LO-TO splitting, please remove or rename the
tensors.xmlfile in theph_savedirectory.If
qlist_fileis provided instead ofnq1,nq2, andnq3, thenlotoandloto_dirvariables are ignored. Seeqlist_filefor details.
- Type:
REAL (Array of 3) - Default:
0.0 0.0 0.0 -
Description: A vector (in cartesian units) specifying the direction approaching the \(\Gamma\) point (\(q \to 0\)) for LO-TO splitting. Internally, the code normalizes the vector, so the user does not need to provide a normalized vector.
If
loto = false, this flag is ignored.If the magnitude of the vector is smaller than \(10^{-5}\), it is ignored and treated as
loto = false.- Format: Three space-separated floats.
- Example:
loto_dir = 1.0 0.0 0.0(along x-axis)
If
qlist_fileis provided instead ofnq1,nq2, andnq3, thenlotoandloto_dirvariables are ignored. Seeqlist_filefor details.
- Type:
INTEGER - Default:
1 - Description: Dimensions of the fine q-point grid used for interpolation. Number of q points along reciprocal lattice vector \(b_1\).
- Type:
INTEGER - Default:
1 - Description: Dimensions of the fine q-point grid used for interpolation. Number of q points along reciprocal lattice vector \(b_2\).
- Type:
INTEGER - Default:
1 - Description: Dimensions of the fine q-point grid used for interpolation. Number of q points along reciprocal lattice vector \(b_3\).
- Type:
LOGICAL - Default:
False - Description:
If
True, the code will dump the change in the local part of the bare potential to disk. Note: in casewrite_dVbare = Trueandinterpolate_dvscf = False, dyn* files are not written.
- Type:
REAL - Default:
1.0 - Description:
Ewald parameter used for screened long-range interactions (dipoles, quadrupoles, phonon dynamical matrices).
- Must be a positive value.
- Type:
REAL - Default:
1.0 - Description:
Ewald summation parameter used specifically for the interpolation of dynamical matrices.
- Must be a positive value.
- Type:
STRING - Default:
"" -
- If a non-zero length string is provided, the code interpolates over this list instead of the grid defined by
nq1,nq2,nq3. - Format:
Description: Filename containing a specific list of q-points provided in reduced units to interpolate.
Note on LO-TO splitting
When
qlist_fileis provided, thelotoandloto_dirvariables are ignored. To obtain LO–TO splitting, include a small direction vector at the \(\Gamma\) point. Instead of sampling exactly at \(\Gamma\), use a displaced point \(\vec{q} = \Gamma + \vec{\epsilon}\), where \(\lvert \vec{\epsilon} \rvert\) is small and defines the direction of approach. When \(\vec{\epsilon}\) vector is converted to caresian units, its magnitude should be small; a recommended value is \(|\vec{\epsilon}_{cart}| = 10^{-4}\). - If a non-zero length string is provided, the code interpolates over this list instead of the grid defined by