Input files#

Setting file#

A setting file contains phonopy settings which are summarized at Setting tags. This file is passed to phonopy as an argument, e.g.,

% phonopy phonopy.conf

where the filename is arbitrary.

phonopy.yaml and phonopy_disp.yaml#

These are output files after the calculation or creating the displacements. These files contain the crystal structure information, primitive cell and supercell sizes, and also the calculator interface. Therefore with this file, users will not need to specify those crystal sturcutre related tags. This file format can be used with CELL_FILENAME tag or -c option:

$ phonopy -c phonopy_disp.yaml

FORCE_SETS, BORN, and FORCE_CONSTANTS information can be also stored in phonopy.yaml as the output after running phonopy, e.g.,

$ phonopy -c phonopy_disp.yaml --include-all --nac

Structure file#

Crystal structure is described by a file with specific format for each calculator, though the default crystal structure is written in VASP POSCAR format. See the detail of the calculator interfaces at Interfaces to calculators.

VASP POSCAR like format#

In the following, the VASP POSCAR format that phonopy can parse is explained. The format is simple. The first line is for your comment, where you can write anything you want. The second line is the ratio for lattice parameters. You can multiply by this number. The third to fifth lines give the lattice parameters, a, b, and c for the respective lines. The sixth line contains the number of atoms for each atomic species, which have to correspond to the atomic positions in the order. The seventh line should be written as Direct. This means that the atomic positions are represented in fractional (reduced) coordinates. When you write chemical symbols in the first line, they are read and those defined by the ATOM_NAME tag are overwritten.

Example of rutile-type silicon oxide crystal structure (VASP 4 style)#

Si O
   1.00000000000000
     4.2266540199664249    0.0000000000000000    0.0000000000000000
     0.0000000000000000    4.2266540199664249    0.0000000000000000
     0.0000000000000000    0.0000000000000000    2.6888359272289208
 2   4
Direct
  0.0000000000000000  0.0000000000000000  0.0000000000000000
  0.5000000000000000  0.5000000000000000  0.5000000000000000
  0.3067891334429594  0.3067891334429594  0.0000000000000000
  0.6932108665570406  0.6932108665570406  0.0000000000000000
  0.1932108665570406  0.8067891334429594  0.5000000000000000
  0.8067891334429594  0.1932108665570406  0.5000000000000000

Example of rutile-type silicon oxide crystal structure (VASP 5 style)#

The VASP 5.x style is also supported. Chemical symbols are inserted just before the line of the numbers of atoms. The chemical symbols in this line overwrite those defined by the ATOM_NAME tag and those defined by the first line of POSCAR.

Stishovite
   1.00000000000000
     4.2266540199664249    0.0000000000000000    0.0000000000000000
     0.0000000000000000    4.2266540199664249    0.0000000000000000
     0.0000000000000000    0.0000000000000000    2.6888359272289208
Si   O
 2   4
Direct
  0.0000000000000000  0.0000000000000000  0.0000000000000000
  0.5000000000000000  0.5000000000000000  0.5000000000000000
  0.3067891334429594  0.3067891334429594  0.0000000000000000
  0.6932108665570406  0.6932108665570406  0.0000000000000000
  0.1932108665570406  0.8067891334429594  0.5000000000000000
  0.8067891334429594  0.1932108665570406  0.5000000000000000

Force file (FORCE_SETS)#

Two types of FORCE_SETS formats are supported.

Type 1#

This format is the default format of phonopy and force constants can be calculated by built-in force constants calculator of phonopy by finite difference method, though external force constants calculator can be also used to obtain force constants with this format by the fitting approach.

This file gives sets of forces in supercells with finite atomic displacements. Each supercell involves one displaced atom. The first line is the number of atoms in supercell. The second line gives number of calculated supercells with displacements. Below the lines, sets of forces with displacements are written. In each set, firstly the atom number in supercell is written. Secondary, the atomic displacement in Cartesian coordinates is written. Below the displacement line, atomic forces in Cartesian coordinates are successively written. This is repeated for the set of displacements. Blank likes are simply ignored.

In the following example, the third line is the displaced atom number that corresponds to the atom number in the supercell created by phonopy. The fourth line gives the displacements in Cartesian coordinates. The lines below, the atomic forces in Cartesian coordinates are written. Once all the forces for a supercell are written, the next set of forces are written. This routine is repeated until the forces of all the displacements have been written.

Example#

48
2

1
  0.0050650623043761   0.0000000000000000   0.0086223630086415
  -0.0347116200   -0.0000026500   -0.0679795200
   0.0050392400   -0.0015711700   -0.0079514600
   0.0027380900   -0.0017851900   -0.0069206400
... (continue until all the forces for this displacement have written)

25
  0.0050650623043761   0.0000000000000000   0.0086223630086415
  -0.0017134500   -0.0001539800    0.0017333400
   0.0013248100    0.0001984300   -0.0001203700
  -0.0001310200   -0.0007955600    0.0003889300
... (continue until all the forces for this displacement have written)

Type 2#

Equivalent to DFSET of ALM code.

Each line has exactly 6 elements. The first three and second three elements give displacement and force of an atom in a supercell, respectively. One set with the number of lines of supercell atoms corresponds to one supercell calculation and the number of supercell calculations are concatenated as many as the user likes. This file is parsed to finally get displacements and forces to have the array shapes of displacements.shape = (num_supercells, num_atoms, 3) and forces.shape = (num_supercells, num_atoms, 3).

Force constants can be calculated by the fitting approach and this force constants calculation requires external force constants calculator such as ALM (invoked by --alm option). All the data are used for calculating force constants in the fitting (usually least square fitting) by the force constants calculator.

Example#

 0.00834956     0.00506291     0.00215683    -0.01723508    -0.00418148    -0.00376513
-0.00494556     0.00866021    -0.00073630     0.00849148    -0.01091833    -0.00458456
-0.00403290    -0.00837741     0.00368169     0.00476247     0.00907379    -0.00210179
-0.00462319     0.00361350    -0.00809745     0.00996582    -0.00320343     0.01904460
 0.00496785    -0.00596540    -0.00630352    -0.01882121    -0.00100787     0.01681980
...

FORCE_CONSTANTS and force_constants.hdf5#

If the force constants of a supercell are known, it is not necessary to prepared FORCES. Phonopy has an interface to read and write FORCE_CONSTANTS or force_constants.hdf5. To read and write these files are controlled by force constants tags and FC_FORMAT, READFC_FORMAT, WRITEFC_FORMAT. VASP users can use VASP DFPT interface to create FORCE_CONSTANTS from vasprun.xml. Quantum ESPRESSO users can use q2r.x to create force constants file by followng the instraction shown at Using q2r.x to create phonopy force constants file

Force constants are stored in either array shape of

  • Compact format: (n_patom, n_satom, 3, 3)

  • Full format: (n_satom, n_satom, 3, 3)

where n_satom and n_patom are the numbers of atoms in supercell and primitive cell, respectively.

Format of FORCE_CONSTANTS#

First line contains the first two elements of the shape of the force constants array, i.e., for (n_satom, n_satom, 3, 3), the first and second numbers are the same and are the number of atoms in the supercell, and for (n_patom, n_satom, 3, 3), they are the numbers of atoms in the primitive cell and supercell. If the first line contains only one number, it is assumed same as that of the former case.

Below second line, force constants between atoms are written by every four lines. In first line of the four lines, anything can be written, i.e., just ignored. Second to fourth lines of the four lines are for the second rank tensor of force constant in Cartesian coordinates, i.e.:

xx xy xz
yx yy yz
zx zy zz

Example#

32  32
1   1
  4.635786969900131    -0.000000000000000    -0.000000000000000
 -0.000000000000000     4.635786969900130    -0.000000000000000
 -0.000000000000000    -0.000000000000000     4.635786969900130
1   2
 -0.246720998398056    -0.000000000000000    -0.000000000000000
 -0.000000000000000     0.018256999881458    -0.000000000000000
 -0.000000000000000    -0.000000000000000     0.018256999881458
...
1  32
  0.002646999982813     0.018011999883049    -0.000000000000000
  0.018011999883049     0.002646999982813    -0.000000000000000
 -0.000000000000000    -0.000000000000000     0.035303999770773
2   1
 -0.246720998398056     0.000000000000000     0.000000000000000
  0.000000000000000     0.018256999881458     0.000000000000000
  0.000000000000000     0.000000000000000     0.018256999881458
...
32  32
  4.635786969900131     0.000000000000000     0.000000000000000
  0.000000000000000     4.635786969900130     0.000000000000000
  0.000000000000000     0.000000000000000     4.635786969900130

Format of force_constants.hdf5#

This is an alternative of FORCE_CONSTANTS but the data is stored in HDF5 format. See the detail of how to obtain this file, FC_FORMAT, READFC_FORMAT, WRITEFC_FORMAT.

The data are stored as follows. p2s_map is introduced at version 1.12.6. Force constants data can be stored in the array shape of either (n_satom, n_satom, 3, 3) or (n_patom, n_satom, 3, 3). In the later case, p2s_map is necessary for the consistency check and this gives the indices of atoms in the primitive cell in supercell index system.

In [1]: import h5py

In [2]: f = h5py.File("force_constants.hdf5", 'r')

In [3]: list(f)
Out[3]: ['force_constants', 'p2s_map']

In [4]: f['force_constants'].shape
Out[4]: (2, 64, 3, 3)

In [5]: f['p2s_map'][:]
Out[5]: array([ 0, 32], dtype=int32)

QPOINTS (optional)#

Specific q-points are calculated using QPOINTS = .TRUE. tag and QPOINTS file. The file format of QPOINTS is as follows. The first line gives the number of q-points. Then the successive lines give q-points in reduced coordinate of reciprocal space of the input unit cell.

Example#

512
-0.437500000000000  -0.437500000000000  -0.437500000000000
-0.312500000000000  -0.437500000000000  -0.437500000000000
-0.187500000000000  -0.437500000000000  -0.437500000000000
...

BORN (optional)#

This file is used with the --nac option or NAC tag.

The formula implemented is refered to Non-analytical term correction.

Format#

In the first line, unit conversion factor is given. In versions 1.10.4 or later, the default value for each calculater can be used if characters than numerical number are given. The default values for the calculaters are found at Default unit conversion factor for non-analytical term correction.

In the second line, dielectric constant \(\epsilon\) is specifed in Cartesian coordinates. The nine values correspond to the tensor elements of xx, xy, xz, yx, yy, yz, zx, zy, and zz.

From the third line, Born effective charges \(Z\) for the independent atoms in the primitive cell have to be written in Cartesian coordinates. The independent atoms can be found using the -v option. As shown below in the Al2O3 example, the independent atoms are marked by * in front of atomic positions:

% phonopy --pa R -v
  _ __ | |__   ___  _ __   ___   _ __  _   _
 | '_ \| '_ \ / _ \| '_ \ / _ \ | '_ \| | | |
 | |_) | | | | (_) | | | | (_) || |_) | |_| |
 | .__/|_| |_|\___/|_| |_|\___(_) .__/ \__, |
 |_|                            |_|    |___/
                                      2.12.0

Python version 3.9.6
Spglib version 1.16.2

Crystal structure was read from "phonopy_disp.yaml".
Unit of length: angstrom
Settings:
  Supercell: [2 2 1]
  Primitive matrix:
    [0.6666666666666666, -0.3333333333333333, -0.3333333333333333]
    [0.3333333333333333, 0.3333333333333333, -0.6666666666666666]
    [0.3333333333333333, 0.3333333333333333, 0.3333333333333333]
Spacegroup: R-3c (167)
------------------------------ primitive cell ------------------------------
Lattice vectors:
  a    2.387113172649497    1.378200432815288    4.337119797109073
  b   -2.387113172649497    1.378200432815288    4.337119797109073
  c    0.000000000000000   -2.756400865630577    4.337119797109073
Atomic positions (fractional):
   *1 Al  0.35209370789580  0.35209370789579  0.35209370789580  26.982
    2 Al  0.64790629210421  0.64790629210420  0.64790629210421  26.982
    3 Al  0.14790629210421  0.14790629210421  0.14790629210421  26.982
    4 Al  0.85209370789579  0.85209370789580  0.85209370789579  26.982
   *5 O   0.55580865378801  0.94419134621199  0.25000000000000  15.999
    6 O   0.44419134621199  0.05580865378802  0.75000000000000  15.999
    7 O   0.25000000000000  0.55580865378801  0.94419134621199  15.999
    8 O   0.75000000000000  0.44419134621200  0.05580865378801  15.999
    9 O   0.94419134621199  0.25000000000000  0.55580865378801  15.999
   10 O   0.05580865378801  0.75000000000001  0.44419134621199  15.999
-------------------------------- unit cell ---------------------------------
Lattice vectors:
  a    4.774226345298994    0.000000000000000   -0.000000000000000
  b   -2.387113172649497    4.134601298445865    0.000000000000000
  c   -0.000000000000000    0.000000000000000   13.011359391327222
Atomic positions (fractional):
   *1 Al  0.33333333333334  0.66666666666666  0.01876037456246  26.982 > 1
    2 Al  0.33333333333334  0.66666666666666  0.31457295877087  26.982 > 2
    3 Al  0.00000000000000  0.00000000000000  0.14790629210421  26.982 > 3
    4 Al  0.66666666666666  0.33333333333334  0.18542704122913  26.982 > 4
    5 Al  0.00000000000000  0.00000000000000  0.35209370789579  26.982 > 1
    6 Al  0.00000000000000  0.00000000000000  0.64790629210421  26.982 > 2
    7 Al  0.66666666666666  0.33333333333334  0.48123962543754  26.982 > 3
    8 Al  0.33333333333334  0.66666666666666  0.51876037456246  26.982 > 4
    9 Al  0.66666666666666  0.33333333333334  0.68542704122913  26.982 > 1
   10 Al  0.66666666666666  0.33333333333334  0.98123962543754  26.982 > 2
   11 Al  0.33333333333334  0.66666666666666  0.81457295877087  26.982 > 3
   12 Al  0.00000000000000  0.00000000000000  0.85209370789579  26.982 > 4
  *13 O   0.30580865378801  0.00000000000000  0.25000000000000  15.999 > 5
   14 O   0.36085801287865  0.33333333333334  0.08333333333334  15.999 > 6
   15 O   0.00000000000000  0.30580865378801  0.25000000000000  15.999 > 7
   16 O   0.66666666666666  0.02752467954532  0.08333333333334  15.999 > 8
   17 O   0.69419134621199  0.69419134621199  0.25000000000000  15.999 > 9
   18 O   0.97247532045468  0.63914198712135  0.08333333333334  15.999 > 10
   19 O   0.97247532045468  0.33333333333334  0.58333333333334  15.999 > 5
   20 O   0.02752467954532  0.66666666666666  0.41666666666666  15.999 > 6
   21 O   0.66666666666666  0.63914198712135  0.58333333333334  15.999 > 7
   22 O   0.33333333333334  0.36085801287865  0.41666666666666  15.999 > 8
   23 O   0.36085801287865  0.02752467954532  0.58333333333334  15.999 > 9
   24 O   0.63914198712135  0.97247532045468  0.41666666666666  15.999 > 10
   25 O   0.63914198712135  0.66666666666666  0.91666666666666  15.999 > 5
   26 O   0.69419134621199  0.00000000000000  0.75000000000000  15.999 > 6
   27 O   0.33333333333334  0.97247532045468  0.91666666666666  15.999 > 7
   28 O  -0.00000000000000  0.69419134621199  0.75000000000000  15.999 > 8
   29 O   0.02752467954532  0.36085801287865  0.91666666666666  15.999 > 9
   30 O   0.30580865378801  0.30580865378801  0.75000000000000  15.999 > 10
-------------------------------- super cell --------------------------------
...

If VASP is used as the calculator for Born effective charge, and the hexagonal unit cell is used for the calculation, the Born effective charge tensors of atoms No. 1 and 13 have to be written in BORN file.

Example#

 14.400
 3.269  0.000  0.000  0.000  3.269  0.000  0.000  0.000  3.234
 2.981  0.000  0.000  0.000  2.981  0.000  0.000  0.000  2.952
-1.935  0.000  0.000  0.000 -2.036 -0.261  0.000 -0.261 -1.968

or using the default NAC unit conversion factor (version 1.10.4 or later),

default value
 3.269  0.000  0.000  0.000  3.269  0.000  0.000  0.000  3.234
 2.981  0.000  0.000  0.000  2.981  0.000  0.000  0.000  2.952
-1.935  0.000  0.000  0.000 -2.036 -0.261  0.000 -0.261 -1.968