that is, if the user is performing second-order eigenvalue calculations using response-functions.
The variable bdeigrf is the maximum number of bands for which the second-order eigenvalues must be calculated: the full number of bands is still used during the computation of these corrections.
If bdeigrf is set to -1, the code will automatically set bdeigrf equal to nband.
Controls the range of atoms for which displacements will be considered
in non-linear computations (using the 2n+1 theorem), for the 1st perturbation.
May take values from 1 to natom, with
d3e_pert1_atpol(1)<=d3e_pert1_atpol(2).
See rfatpol for additional details.
Gives the directions to be considered in non-linear computations
(using the 2n+1 theorem), for the 1st perturbation.
The three elements corresponds to the three primitive
vectors, either in real space (atomic displacement),
or in reciprocal space (electric field perturbation).
See rfdir for additional details.
Turns on electric field perturbation in non-linear computation, as 1st perturbation.
Actually, such calculations requires first the non-self-consistent calculation
of derivatives with respect to k, independently of the electric field perturbation itself.
See rfelfd for additional details.
Turns on atomic displacement perturbation in non-linear computation, as 1st perturbation.
See rfphon for additional details.
Controls the range of atoms for which displacements will be considered
in non-linear computations (using the 2n+1 theorem), for the 2nd perturbation.
May take values from 1 to natom, with
d3e_pert2_atpol(1)<=d3e_pert2_atpol(2).
See rfatpol for additional details.
Gives the directions to be considered in non-linear computations
(using the 2n+1 theorem), for the 2nd perturbation.
The three elements corresponds to the three primitive
vectors, either in real space (atomic displacement),
or in reciprocal space (electric field perturbation).
See rfdir for additional details.
Turns on electric field perturbation in non-linear computation, as 2nd perturbation.
Actually, such calculations requires first the non-self-consistent calculation
of derivatives with respect to k, independently of the electric field perturbation itself.
See rfelfd for additional details.
Turns on atomic displacement perturbation in non-linear computation, as 2nd perturbation.
See rfphon for additional details.
Controls the range of atoms for which displacements will be considered
in non-linear computations (using the 2n+1 theorem), for the 3rd perturbation.
May take values from 1 to natom, with
d3e_pert3_atpol(1)<=d3e_pert3_atpol(2).
See rfatpol for additional details.
Gives the directions to be considered in non-linear computations
(using the 2n+1 theorem), for the 3rd perturbation.
The three elements corresponds to the three primitive
vectors, either in real space (atomic displacement),
or in reciprocal space (electric field perturbation).
See rfdir for additional details.
Turns on electric field perturbation in non-linear computation, as 3rd perturbation.
Actually, such calculations requires first the non-self-consistent calculation
of derivatives with respect to k, independently of the electric field perturbation itself.
See rfelfd for additional details.
Turns on atomic displacement perturbation in non-linear computation, as 3rd perturbation.
See rfphon for additional details.
It is the value of the "scissors operator", the
shift of conduction band eigenvalues,
used in response function calculations.
Can be specified in Ha (the default), Ry, eV or Kelvin, since
ecut
has the
'ENERGY' characteristics.
(1 Ha=27.2113845 eV)
Typical use is for response to electric field (rfelfd=3),
but NOT for d/dk (rfelfd=2) and phonon responses.
Turns on effective mass tensor calculations.
Such calculations requires the non-self-consistent calculation
of derivatives with respect to k, in the same dataset.
It must therefore be used with rfelfd=2.
At the present time, both norm-conserving (NC) and PAW calculations are supported. Also, for PAW calculations only, nspinor==2 and pawspnorb==1 (i.e. spin-orbit (SO) calculations) is supported. NC SO calculations are NOT currently supported. Also, for both NC and PAW, nspden/=1 and nsppol/=1 are NOT supported.
This variable controls the range of bands for which the effective mass is to be calculated. If a band is degenerate, all other bands of the degenerate group will automatically be treated, even if they were not part of the user specified range.
Allows the user to calculate the scalar effective mass of all bands specified by efmas_bands along specific directions in reciprocal space. This is particularly useful when considering degenerate bands, which are usually warped, and thus cannot have their dispersion (hessian) and effective mass expressed as a tensor. This allows the user to see the more complex angular behavior of effective masses in these cases, for instance.
When efmas_calc_dirs==0, no directions are read from the input file (using efmas_dirs) and the effective masses along the 3 cartesian directions are output by default.
When efmas_calc_dirs==1, 2 or 3, efmas_n_dirs directions are read from efmas_dirs, assuming cartesian, reduced or angular (theta,phi) coordinates, respectively. In the case efmas_calc_dirs==3, 2 real values per directions are read, whereas 3 real values are read in the two other cases.
Activate (==1) or not (==0) the treatment of degenerate bands (within a criterion efmas_deg_tol) using the transport equivalent effective mass idea (see PRB 89 155131 (2014)).
Energy difference below which 2 bands are considered degenerate (and treated using the formalism activated with efmas_deg==1). efmas_deg_tol has the 'ENERGY' characteristics.
For 2D or 1D systems, the band dispersion goes to 0 perpendicular to the system, which causes the inverse effective mass to be singular, i.e. the effective mass to be NaN. This keyword circumvents the problem by eliminating the troublesome dimensions from the inverse effective mass.
In 2D, the Z axis is ignored and, in 1D, the Z and Y axis are ignored.
Also, note that in the 2D degenerate case, a subtlety arises: the 'transport equivalent' effective mass does not determine the scale of the transport tensors (conductivity and others). Therefore, for this specific case, the factor by which these transport tensors should be scaled once determined from the 'transport equivatlent' effective mass tensor is output separately on the line immediately after the effective mass.
List of efmas_n_dirs directions to be considered according to the value of efmas_calc_dirs. The directions are specified by 3 real values if efmas_calc_dirs==1 or 2 and by 2 real values if efmas_calc_dirs==3.
Number of directions in efmas_dirs, to be considered according to efmas_calc_dirs.
When a band is degenerate, the usual definition of effective mass becomes invalid. However, it is still possible to define a 'transport equivalent mass tensor' that reproduces the contribution of the band to the conductivity tensor. To obtain this tensor, an integration over the solid sphere is required. The default value gives a tensor accurate to the 4th decimal in Ge.
that is, if the user is performing performing second-order eigenvalue calculations using response-functions.
The variable elph2_imagden determines the imaginary shift of the denominator of the sum-over-states
in the perturbation denominator, (e_{nk}-e_{n'k'}+i elph2_imagden).
One should use a width comparable with the Debye frequency or the maximum phonon frequency.
Can be
specified in Ha (the default), Ry, eV or Kelvin, since
ecut
has the
'ENERGY' characteristics.
(1 Ha=27.2113845 eV)
When optdriver==7, select the task to be performed. The choice is among:
that is, if the user is performing simulations of the electronic lifetimes induced by the electron-phonon coupling.
The variable esmear determines the width of the functions approximating the delta function, \delta(e_{nk}-e_{n'k'}),
present in the expression of the lifetimes. One should use a width comparable with the Debye frequency or the maximum phonon frequency.
Can be specified in Ha (the default), Ry, eV or Kelvin, since
ecut
has the
'ENERGY' characteristics.
(1 Ha=27.2113845 eV)
Can be used to suppress artificially the first-order change of Fermi energy, in case of Response Function calculation for metals at Q=0. The input variable frzfermi, if set to 1, allows to suppress this contribution, but this is incorrect.
If ieig2rf is greater then 0, the code will produce a file, named with the trailing suffix _EIGR2D, containing the second-order electronic eigenvalues for the perturbation. These files are used in the calculation of the thermal correction to the electronic eigenvalues.
If ieig2rf is set to 1, the second-order electronic eigenvalues will be calculated from the DFPT method (Sternheimer).
If ieig2rf is set to 2, the second-order electronic eigenvalues will be calculated from the Allen-Cardona method. (sum over states)
If ieig2rf is set to 3, the second-order electronic eigenvalues will be calculated from the DFPT method (sum over states) but using a different part of the code. This is equivalent to ieig2rf = 1 [debuging]
If ieig2rf is set to 4, the second-order electronic eigenvalues will be calculated from the dynamical DFPT method (Sternheimer).
The code will generate _EIGR2D.nc files that contain the electron-phonon matrix element squared on the space orthogonal to the active space.
The code will also produce _FAN.nc files that contain the electron-phonon matrix elements squared.
Note that ieig2rf=4 can only be used if Abinit is compiled with NETCDF support.
If ieig2rf is set to 5, the second-order electronic eigenvalues will be calculated from the dynamical DFPT method (Sternheimer).
The code will generate _EIGR2D.nc files that contain the electron-phonon matrix element square on the space orthogonal to the active space.
The code will also produce _GKK.nc files that contain electron-phonon matrix elements.
This option is preferable for large system to ieig2rf=4 as the GKK files take less much less disk space and memory (but run a little bit slower).
Note that ieig2rf=5 can only be used if Abinit is compiled with NETCDF support.
Related variables : bdeigrf,elph2_imagden,getgam_eig2nkq,smdelta
Related variables : bdeigrf,elph2_imagden,getgam_eig2nkq,smdelta
This variable defines the q-mesh used to compute the phonon DOS and the Eliashberg function via Fourier interpolation. Related input variables: ph_qshift and ph_nqshift.
This array contains the list of special q-points used to construct the q-path for phonon band structures and phonon linewidths. See also ph_nqpath and [ph_ndivsm.
The computation of third-order derivatives from the 2n+1 theorem
requires the first-order wavefunctions and densities obtained from
a linear response calculation. The standard approach in a linear
response calculation is (i) to compute only the
irreducible perturbations, and (ii) to use symmetries to
reduce the number of k-points for the k-point integration.
This approach cannot be applied, presently (v4.1),
if the first-order wavefunctions are to be used to compute third-order derivatives.
First, for electric fields, the code needs the derivatives
along the three directions. Still, in case of phonons, only the
irreducible perturbations are required.
Second, for both electric fields and phonons, the wavefunctions
must be available in half the BZ (kptopt=2), or the full BZ (kptopt=3).
During the linear response calculation, in order to prepare a non-linear
calculation, one should put prepanl to 1 in order
to force ABINIT (i) to compute the electric field perturbation
along the three directions explicitly, and (ii) to keep the full number of k-points.
The calculation of electron-phonon coupling quantities requires the presence of all the perturbations (all atoms in all directions) for the chosen set of (irreducible) q-points. To impose this and prevent ABINIT from using symmetry to reduce the number of perturbations, set prepgkk to 1. Use in conjunction with prtgkk.
If prtbbb is 1, print the band-by-band decomposition of Born effective charges and localization tensor, in case they are computed. See Ph. Ghosez and X. Gonze, J. Phys.: Condens. Matter 12, 9179 (2000).
UNUSABLE (in development)
Activates computation of second derivatives of wavefunctions with respect to wavevectors. This is not strictly a response function but is a needed auxiliary quantity in the calculations of 3rd-order derivatives of the energy (non-linear response). The directions for the derivatives are determined by rfdir (TO BE CORRECTED!).
Control the evaluation of the acoustic sum rule in effective charges and dynamical matrix at Gamma within a response function calculation (not active at the level of producing the DDB, but at the level of the phonon eigenfrequencies output).
Control the range
of atoms for which displacements will be considered
in phonon calculations (atomic polarizations), using the 2n+1 theorem.
These values are only relevant to phonon response function
calculations.
May take values from 1 to natom, with rfatpol(1)<=rfatpol(2).
The atoms to be moved will be defined by the
do-loop variable iatpol :
do iatpol=rfatpol(1),rfatpol(2)
For the calculation of a full dynamical matrix, use
rfatpol(1)=1 and rfatpol(2)=natom, together with
rfdir 1 1 1 . For selected elements of the
dynamical matrix, use different values of rfatpol and/or
rfdir. The name 'iatpol' is used for the part of the
internal variable ipert when it runs from 1 to natom. The
internal variable ipert can also assume values larger
than natom,
denoting perturbations of electric field or stress type (see
the response function help file
).
Activates computation of derivatives of ground state wavefunctions with respect to wavevectors. This is not strictly a response function but is a needed auxiliary quantity in the electric field calculations (see rfelfd) The directions for the derivatives are determined by rfdir.
Gives the directions
to be considered for response function calculations
(also for the Berry phase computation of the polarization, see
the berryopt input variable).
The three elements corresponds to the three primitive
vectors, either in real space (phonon calculations),
or in reciprocal space (d/dk, homogeneous electric field, homogeneous magnetic field
calculations). So, they generate a basis
for the generation of the dynamical matrix or
the macroscopic dielectric tensor or magnetic susceptibility and magnetic
shielding, or the effective
charge tensors.
If equal to 1, response functions, as defined
by rfddk,
rfelfd, rfphon, rfdir
and rfatpol, are to be computed
for the corresponding direction. If 0, this direction
should not be considered.
Turns on electric field response function calculations. Actually, such calculations requires first the non-self-consistent calculation of derivatives with respect to k, independently of the electric field perturbation itself.
Selects method used in response function calculations. Presently, only 1 is allowed.
It must be equal to 1 to run phonon response function calculations.
Used to run strain response-function calculations (e.g. needed to get elastic constants). Define, with rfdir, the set of perturbations.
Available to the developpers, to activate
the use of ipert=natom+5 and ipert=natom+6, two sets of perturbations
that the developpers can define.
In order to define and use correctly the new perturbations, the developper might have to include code lines or additional routines at the level of the following routines : dfpt_cgwf.F90, dfpt_dyout.F90, dfpt_symph.F90, dfpt_dyout.F90, dfpt_etot.F90, littlegroup_pert.F90, dfpt_looppert.F90, dfpt_mkcor.F90, dfpt_nstdy.F90, dfpt_nstwf.F90, respfn.F90, dfpt_scfcv.F90, irreducible_set_pert.F90, dfpt_vloca.F90, dfpt_vtorho.F90, dfpt_vtowfk.F90. In these routines, the developper should pay a particular attention to the rfpert array, defined in the routine respfn.F90 , as well as to the ipert local variable.
When smdelta in non-zero, it will trigger the calculation of the imaginary part of the second-order electronic eigenvalues, which can be related to the electronic lifetimes. The delta function is evaluated using:
The Matrix to be diagonalized in the Casida framework
(see "Time-Dependent Density Functional Response Theory of Molecular
systems: Theory, Computational Methods, and Functionals", by M.E. Casida,
in Recent Developments and Applications of Modern Density Functional
Theory, edited by J.M. Seminario (Elsevier, Amsterdam, 1996).)
is a NxN matrix, where, by default, N is the product of
the number of occupied states by the number of unoccupied states.
The input variable td_maxene allows to diminish N : it selects
only the pairs of occupied and unoccupied states for which the
Kohn-Sham energy difference is less than td_maxene.
The default value 0.0 means that all pairs are taken into account.
See td_mexcit for an alternative
way to decrease N.
The Matrix to be diagonalized in the Casida framework
(see "Time-Dependent Density Functional Response Theory of Molecular
systems: Theory, Computational Methods, and Functionals", by M.E. Casida,
in Recent Developments and Applications of Modern Density Functional
Theory, edited by J.M. Seminario (Elsevier, Amsterdam, 1996).)
is a NxN matrix, where, by default, N is the product of
the number of occupied states by the number of unoccupied states.
The input variable td_mexcit allows to diminish N : it selects
the first td_mexcit pairs of occupied and unoccupied states, ordered
with respect to increasing Kohn-Sham energy difference.
However, when td_mexcit is zero, all pairs are allowed.
See td_maxene
for an alternative
way to decrease N.