11.3. ANPHON: Output files#

The files are listed by MODE. PREFIX is the value of the PREFIX tag. Several files come in two formats: the HDF5 version is written by default, and the plain-text version with FILE_FORMAT = text. The layouts of the HDF5 files are described in HDF5 files.

MODE = phonons#

Files written for the points of the &kpoint field. KPMODE 0, 1, and 2 give a list of points, band paths, and a uniform mesh, respectively.

File

Written when

Contents

PREFIX.bands

KPMODE = 1

Phonon dispersion (cm-1) along the paths

PREFIX.connection

KPMODE = 1, BCONNECT = 2

How the branches connect along the paths

PREFIX.dos

KPMODE = 2 (DOS = 1, default)

Phonon DOS; atom-projected DOS with PDOS = 1

PREFIX.tdos

KPMODE = 2, TDOS = 1

Two-phonon DOS at the irreducible \(k\) points

PREFIX.longitudinal_dos

KPMODE = 2, LONGITUDINAL_DOS = 1

Longitudinal-projected DOS

PREFIX.thermo

KPMODE = 2

Heat capacity, entropy, internal energy, and free energy vs. temperature. With FE_BUBBLE = 1, the bubble correction to the free energy is added as the last column (not included in the harmonic free energy)

PREFIX.msd

KPMODE = 2, PRINTMSD = 1

Mean-square displacements of atoms vs. temperature

PREFIX.ucorr

KPMODE = 2, UCORR = 1

Displacement–displacement correlation function (SHIFT_UCORR)

PREFIX.sps, PREFIX.sps_Bose

KPMODE = 2, SPS = 1 or 2

Three-phonon scattering phase space; .sps_Bose includes the Bose factor

PREFIX.phvel, PREFIX.phvel_all

PRINTVEL = 1 with KPMODE = 1 or 2

Group velocities along the paths, or at every mesh point (magnitude and components)

PREFIX[.band | .mesh].eval.h5

PRINTEVAL = 1

Phonon eigenvalues. The file name follows KPMODE 0, 1, 2 (no suffix, .band, .mesh); plain text PREFIX[.band | .mesh].eval with FILE_FORMAT = text

PREFIX[.band | .mesh].evec.h5

PRINTEVEC = 1

Phonon eigenvalues and eigenvectors, named as above; plain text PREFIX[.band | .mesh].evec with FILE_FORMAT = text. Both can be given to displace.py --evec

PREFIX[.band | .mesh].pr, .apr

PRINTPR = 1

Participation ratio and atomic participation ratio of every mode, named as above

PREFIX[.band | .mesh].axsf

PRINTXSF = 1

Mode directions at the given \(k\) points, for XCrySDen, named as above

PREFIX.animeNNN.xyz, PREFIX.animeNNN.axsf

ANIME (with ANIME_CELLSIZE)

Animation of mode NNN; XYZ (default) or AXSF with ANIME_FORMAT

PREFIX.gru_kpoints, PREFIX.gruneisen, PREFIX.gru_all

GRUNEISEN >= 1 with KPMODE = 0, 1, or 2

Grüneisen parameters at the given points, along the paths, or on the mesh: volumetric (GRUNEISEN = 1) or for each strain component (2, 3)

PREFIX_+.h5, PREFIX_-.h5

NEWFCS = 1

Force constants of the crystal strained by \(\pm u\), usable as FCSFILE (.xml with FILE_FORMAT = text)

PREFIX.dielec

DIELEC = 1 (needs BORNINFO)

Phonon contribution to the dielectric function \(\epsilon(\omega)\)

PREFIX.zmode

ZMODE = 1 (needs BORNINFO)

Mode effective charges of the zone-center modes

PREFIX.irreps

IRREPS = 1

Irreducible representations and IR/Raman activities at \(\Gamma\)

PREFIX.fc2_ewald

NONANALYTIC = 3, FC2_EWALD = 1 (any MODE)

Harmonic force constants split into the dipole–dipole and short-range parts

MODE = kappa#

File

Written when

Contents

PREFIX.kappa.h5

Always (default format)

All results: frequencies, velocities, linewidths, and thermal conductivity. Updated during the run and used for restarting; read by analyzer.py. See Thermal conductivity (PREFIX.kappa.h5)

PREFIX.result, PREFIX.4ph.result

FILE_FORMAT = text

Legacy text versions of the three- and four-phonon linewidths

PREFIX.kl

SOLVER = RTA

Thermal conductivity tensor vs. temperature

PREFIX.kl3, PREFIX.kl4

SOLVER = RTA, INCLUDE_4PH = 1 (instead of .kl)

Thermal conductivity with three-phonon scattering only, and with three- and four-phonon scattering

PREFIX.kl_iter

SOLVER = IBTE, VBTE, or DBTE

Thermal conductivity from the solution of the Boltzmann equation beyond the RTA

PREFIX.kl_spec

SOLVER = RTA, KAPPA_SPEC = 1

Spectrum of the thermal conductivity (diagonal components)

PREFIX.kl_coherent

SOLVER = RTA, KAPPA_COHERENT >= 1

Coherent (off-diagonal) contribution to the thermal conductivity

PREFIX.kc_elem

SOLVER = RTA, KAPPA_COHERENT = 2

Mode-resolved contributions to the coherent term

PREFIX.self_isotope

SOLVER = RTA, ISOTOPE = 2

Linewidths due to isotope scattering

PREFIX.interpolated_gamma

INCLUDE_4PH = 1, WRITE_INTERPOL = 1

Four-phonon linewidths interpolated onto the three-phonon mesh

PREFIX.velmat_check

KAPPA_COHERENT >= 1

Diagnostic of the velocity matrix: largest deviation of its diagonal from the group velocities and largest deviation from hermiticity

MODE = selfenergy#

File

Written when

Contents

PREFIX.selfenergy.h5

Always (default format)

Results of all target modes. See Self-energy (PREFIX.selfenergy.h5)

PREFIX.Gamma.N

FILE_FORMAT = text, LINEWIDTH = 1 (default)

Linewidth \(2\Gamma\) vs. temperature of target N

PREFIX.Shift.N

FILE_FORMAT = text, SHIFT = 1

Frequency shifts vs. temperature

PREFIX.Self.N

FILE_FORMAT = text, SELF_W = 1

Self-energy as a function of frequency

PREFIX.fw.N

FILE_FORMAT = text, FSTATE_W = 1

Frequency-resolved final-state analysis of the linewidth

PREFIX.spectrum

FILE_FORMAT = text, INTERPOLATE = 1

Spectral function \(A(\boldsymbol{q}, \omega)\), total and per branch

PREFIX.V3.N, .Phi3.N, .V4.N, .Phi4.N

PRINTV3, PRINTV4 = 1 or 2 (always text)

Three- and four-phonon matrix elements involving target N

In the text files, N numbers the targets on the integration mesh first and the other targets after them; the header of each file gives the wave vector and the branch. The same per-target files are written in MODE = kappa with the KS_INPUT analysis (SELF_ENERGY = 1, REALPART = 1, …).

MODE = SCPH and MODE = QHA#

The file names start with PREFIX.scph for MODE = SCPH and PREFIX.qha for MODE = QHA. With BUBBLE = 1, 2, 3, SCPH additionally writes the bubble-corrected versions PREFIX.scph+bubble(0), +bubble(w), or +bubble(wQP) of the eval, bands, dos, msd, ucorr, and dfc2 files.

File

Written when

Contents

PREFIX.scph.h5, PREFIX.qha.h5

Always (default format)

Results at every temperature, used by later runs (FC2_TEMPERATURE, RELAXED_STRUCTURE) and for restarting. The FC2 corrections of each temperature are projected onto the space group of the structure of that temperature (the spglib group of the relaxed structure with RELAX_STR > 0, at the run’s TOLERANCE; the operations of the run otherwise). See SCPH and QHA results (PREFIX.scph.h5, PREFIX.qha.h5), and Force constants and corrections with NONANALYTIC > 0 for the force constants with NONANALYTIC > 0

PREFIX.scph_eval, .qha_eval

KPMODE = 0

Frequencies vs. temperature at the given \(k\) points

PREFIX.scph_bands, .qha_bands

KPMODE = 1

Dispersion vs. temperature. With RELAX_STR > 0, the \(k\) axis is that of the reference cell at every temperature (the frequencies are those of the relaxed structure); a run with RELAXED_STRUCTURE = 1 uses the axis of the relaxed cell

PREFIX.scph_dos, .qha_dos

KPMODE = 2 (DOS = 1, default)

DOS vs. temperature

PREFIX.scph_thermo, .qha_thermo

KPMODE = 2

Heat capacity, entropy, and free energy vs. temperature. With FE_BUBBLE = 1 (.scph_thermo only), the bubble correction to the free energy is written in a separate column and included in F_{total}, but not in Cv and S. For a temperature whose SCPH or structural iterations did not converge, Cv, F and S are written as NaN (Phi0 is kept), and so are the anharmonic Cv correction (also at the neighbouring temperatures, which use it in a central difference) and the rows of that temperature in .scph_dos / .qha_dos, the MSD and ucorr files, and their +bubble versions. The log names what the arrays of that temperature hold: the last SCP iterate, or after a failed structural optimization a copy of the last converged temperature or the harmonic data. ALLOW_UNCONVERGED = 1 prints the values computed from those data instead

PREFIX.scph_hessian

BUBBLE = 4 (MODE = SCPH, RELAX_STR > 0)

Curvature of the SCP free energy at each converged temperature: frequencies of the SCPH and of the free-energy curvature (static bubble + quartic ladder) along the \(\Gamma\) displacements; negative values are unstable directions. With a cell (RELAX_STR = 2, 4), # lines with the clamped- and relaxed-ion elastic curvature (Voigt, GPa)

PREFIX.scph_fe.h5

BUBBLE = 4

The SCPH state file again, with the correction of the free-energy curvature at every point of the mesh (\(\Gamma\) for KMESH_SCPH = 1 1 1) added to the renormalized FC2. Use it with DFC2FILE (or FC2FILE) and FC2_TEMPERATURE for dispersions and DOS interpolated like the SCPH ones; exact at the points of the mesh. Temperatures without a curvature are flagged unconverged. Like the SCPH correction in PREFIX.scph.h5, the sum is projected onto the space group of the relaxed structure of each temperature, so it can differ from .scph_hessian / .scph_hessian_q at the level of numerical noise

PREFIX.scph_hessian_q

BUBBLE = 4 on a k-mesh

Curvature of the SCP free energy at the irreducible \(\boldsymbol{Q}\)-points of the mesh other than \(\Gamma\): for each converged temperature and \(\boldsymbol{Q}\), the frequencies of the SCPH and of the free-energy curvature (all branches); negative values are unstable modulations

PREFIX.scph_hessian_displace

BUBBLE = 4, a negative curvature

For each unstable direction, the &displace (DISPMODE = 1, Cartesian, bohr) and &strain fields of the converged structure moved 0.05 bohr along it, with the space group of the displaced structure

PREFIX.scph_msd, .qha_msd

KPMODE = 2, PRINTMSD = 1

Mean-square displacements vs. temperature

PREFIX.scph_ucorr, .qha_ucorr

KPMODE = 2, UCORR = 1

Displacement correlation function vs. temperature

PREFIX.scph_dielec, .qha_dielec

DIELEC = 1

Dielectric function vs. temperature

PREFIX.scph_dfc2, .qha_dfc2

Always

Change of the harmonic force constants \(\Delta\Phi_2\) at each temperature (see the formalism of the SCPH calculation), relative to the full harmonic dynamical matrix of the run; the comment line # Header = -1 : NONANALYTIC = n ; DELTA_BASELINE = 1 gives the NONANALYTIC to use with it (files without it are of version 1.x, see Force constants and corrections with NONANALYTIC > 0)

PREFIX.atom_disp, PREFIX.normal_disp

RELAX_STR > 0

Atomic displacements vs. temperature, in Cartesian and normal coordinates

PREFIX.umn_tensor

RELAX_STR = 2 or 4 (QHA: 2, 3, or 4)

Displacement gradient tensor \(u_{\mu\nu}\) vs. temperature

step_q0.txt, step_u0.txt, step_u_tensor.txt

RELAX_STR = 1, 2, or 4 (step_u_tensor.txt: 2 or 4)

History of the structural optimization at every step (fixed names in the working directory)

PREFIX.polarization

RELAX_STR = 1, 2, or 4 with BORNINFO

Polarization of the relaxed structure vs. temperature

PREFIX.V0

RELAX_STR > 0

Potential energy \(U_0\) vs. temperature

PREFIX.scph_dymat, PREFIX.renorm_harm_dymat

FILE_FORMAT = text (.renorm_harm_dymat: RELAX_STR > 0)

Legacy restart files: the change of the dynamical matrix

The optimization files (atom_disp, step_*.txt, …) and the dfc2 files are written by a new calculation, not by a restart.

Note

A calculation restarts automatically when the restart files of an earlier run with the same PREFIX are found (PREFIX.kappa.h5 or .result, PREFIX.scph.h5 or .scph_dymat, …). Remove them, or change PREFIX, to start from scratch.