List of input keywords

'[Trial]' : These options are not tested well

&calculation

theory

character, default=''

Choice of a theory to be used in the calculation.
Options:
dft / ground state calculation based on DFT
dft_md / ab initio MD simulations based on DFT (electronic ground state)
tddft_response / linear response TDDFT calculation in real time
tddft_pulse / simulations under pulsed electric field based on TDDFT
single_scale_maxwell_tddft / single-scale simulation coupling Maxwell and TDDFT
multi_scale_maxwell_tddft / multiscale simulation coupling Maxwell and TDDFT
maxwell / electromagnetic analysis using finite difference time domain (FDTD) method
dft_k_expand / convert checkpoint data of dft with k-points calculation to that of larger supercell system with gamma-point

yn_md

[Trial] character, default='n'

Available for theory='dft' (ground-state MD) and theory='tddft_pulse' (Ehrenfest MD).
Switch for molecular dynamics calculation.
Options:
'y' / enable
'n' / disable

yn_opt

[Trial] character, default='n'

Available for theory='dft'.
Switch for geometry optimization.
Options:
'y' / enable
'n' / disable

&control

sysname

character, default='default'

Available for all options of theory.
A prefix of output files.

base_directory

character, default='./'

Available for all options of theory.
Name of a directory where major output files are stored.

yn_restart

character, default='n'

Available for the DFT/TDDFT based options of theory.
Whether to continue previous calculation (restart) or start a new calculation.
Options:
'y' / enable (restart)
'n' / disable (new calculation)

directory_read_data

character, default='restart/'

Available for yn_restart='y'.
Directory name to read data that are required in the present calculation (restart) and were generated in previous calculations. For TDDFT based options, it specifies the name of the directory containing ground state results that were stored in 'data_for_restart'. When restarting from a checkpoint, it specifies the name of the directory that contains the checkpoint data.

yn_self_checkpoint

character, default='n'

Available for the DFT/TDDFT based options of theory.
With this option, each process writes/reads the restart (and checkpoint) data independently (self data format) so that the restart cost is reduced for large systems. Note that the number of processes and their assignments must be unchanged in restarting. The data is written out into 'checkpoint_gs_XXXXXX/' (DFT) or 'checkpoint_rt_XXXXXX/' (TDDFT).
Options:
'y' / enable
'n' / disable

checkpoint_interval

integer, default=-1

Available for the DFT/TDDFT based options of theory.
Interval of time steps (iteration steps) to write down the checkpoint data during the time-propagation (SCF iteration). Checkpoint data will not be written if a negative value is set.

yn_reset_step_restart

character, default='n'

Available for yn_restart='y' in the DFT/TDDFT based options of theory.
With this option, the counter of the SCF iteration step (for DFT) or the counter of the time propagation step (for TDDFT) is reset to 0 at the restart. In the SCF iteration, the density data in the previous SCF iteration step are abondoned.

read_gs_restart_data

character, default='all'

Available for yn_restart='y' with theory='dft'.
Specify which data are read in the restart. Specified data that are generated in the previous calculation and are contained in the restart (or checkpoint) directory are used in restarting the SCF iteration of DFT. The default option 'all' indicates the complete restart. In other options, a part of restart data are used (other data are prepared in the same way as in the initial SCF step).
Options:
all / all of restart data are read
all:single / same as all option but the data is read in the single file format even though the self data format is specified with yn_self_checkpoint='y' (i.e., the restart data is read in the single file format while written out in the self format)
rho_inout / only electron densities including those of previous iteration steps are read (from rho_inout.bin file)
rho_inout:single / same as rho_inout option but the data is read in the single file format even though the self data format is specified with yn_self_checkpoint='y'
rho / only the latest electron density is read (from user-made data)
wfn / only orbital wavefunctions are read

write_gs_restart_data

character, default='all'

Available for theory='dft'.
Options
all / all of restart data are written out
rho_inout / only electron densities including those of previous iteration steps are written out
wfn / only orbital wavefunctions are written out
checkpoint_only / the restart data are outputted only in the self data format (separated data for each process) at the last step into 'checkpoint_gs_XXXXXX/' directory (yn_self_checkpoint='y' is required) without generating the restart data into 'data_for_restart/' directory in the single file format.
Output data files are written out in the restart (or checkpoint) directory.
The default option 'all' gives the complete set of restart data.

time_shutodown

[Trial] real(8), default=-1d0

Available for the DFT/TDDFT based options of theory.
Timer for automatic shutdown. The unit is second.
If a negative time is set, the automatic shutdown will not be performed.

method_wf_distributor

character, default='single'

Available for the DFT/TDDFT based options of theory.
A method to save/load orbital wavefunctions.
Options
single: all orbital wavefunctions are saved(loaded) to(from) a single file.
slice : each orbital wavefunction is saved(loaded) to(from) a file. This choice reduces the I/O costs, and increase the flexiblility in handling files for large systems.

nblock_wf_distribute

integer, default='16'

Available for method_wf_distributor='slice'.
In the 'slice' mode, nblock_wf_distribute files are saved in one directory.

&units

unit_system

character, default='au'

Unit system to be used in input variables and some of output files.
If unit_system = 'A_eV_fs' is chosen, Angstrom for length, eV for energy, and fs for time are adopted.
For isolated systems specified by yn_periodic = 'n' in &system, a unit of 1/eV is used for the output files of DOS and PDOS if unit_system = 'A_eV_fs' is specified, while atomic unit is used if not. For other output files, the Angstrom/eV/fs units are used irrespective of the input keyword. For periodic systems specified by yn_periodic = 'n' in &system, the unit system specified by this input keyword is used for most output files. To confirm the unit, see the first few lines of output files.
Options:
'au' or 'a.u.' / atomic unit system
'A_eV_fs' / Angstrom-eV-fs unit system

&parallel

nproc_k

nproc_ob

nproc_rgrid(3)

integer, default=0

Options:
nproc_k/ Number of MPI parallelization for k-points of electron orbitals.
nproc_ob/ Number of MPI parallelization for orbital index of electron orbitals.
nproc_rgrid(3)'/ Number of MPI parallelization for each direction of real-space grid that are used for electron orbitals and density.

Defaults are 0 for nproc_k/nproc_ob and (0,0,0) for nproc_rgrid. In the default choice, MPI assignment is achieved atomatically. Users can specify nproc_k, nproc_ob, and nproc_rgrid manually. In that case, there are several constraints that should be fulfilled:
nproc_k must be set to 1 for &system/yn_periodic='n'.
nproc_k and nproc_ob must be set to 1 for theory='maxwell'.
nproc_k * nproc_ob * nproc_rgrid(1) * nproc_rgrid(2) * nproc_rgrid(3) = total number of processes.

yn_ffte

character, default='n'

Available for &system/yn_periodic='y'
For periodic systems, SALMON uses Fourier transformation to solve a poisson equation.
This switch selects if FFTE library is used or not. If FFTE is not used, the Fourier transformation in a simple algorithm is carried out.
Options
'y' / enable
'n' / disable

To enable it, following relations must be satisfied.
mod(num_rgrid(1), nproc_rgrid(2)) == 0
mod(num_rgrid(2), nproc_rgrid(2)) == 0
mod(num_rgrid(2), nproc_rgrid(3)) == 0
mod(num_rgrid(3), nproc_rgrid(3)) == 0

yn_scalapack

character, default='n'

Available for &calculation/theory='dft' or 'dft_md'
To calculate large systems, an eigenvalue problem in the subspace diagonalization becomes a bottle-neck in the ground state calculation. In SALMON, ScaLAPACK library can be used to solve the eigenvalue problem.
To enable it, it is necessary to link ScaLAPACK library when you build SALMON.
Options:
'y' / enable
'n' / disable

yn_eigenexa

character, default='n'

Available for &calculation/theory='dft' or 'dft_md'
SALMON can use RIKEN R-CCS EigenExa library to solve eigenvalue problem in subspace diagonalization. It is more efficient than ScaLAPACK to diagonalize matrices of large dimension. To enable it, it is necessary to link both ScaLAPACK and EigenExa libraries when you build SALMON.
Options:
'y' / enable
'n' / disable

yn_diagonalization_red_mem

character, Default='n'

Available for &parallel/yn_scalapack='y' or &parallel/yn_eigenexa='y'
This option reduces memory consumption in using ScaLAPACK/EigenExa libraries.
Options:
'y' / enable
'n' / disable

process_allocation

character, default='grid_sequential'

This controlls the order of process allocation.
Options:
'grid_sequential' / real-space grid major ordering.
'orbital_sequential' / orbital-space major ordering.

Suggestion:
&calculation/theory='dft' or 'dft_md' / 'orbital_sequential'
&calculation/theory='tddft*' or '*maxwell_tddft' / 'grid_sequential'

&system

yn_periodic

character, default='n'

Available for all options of theory.
Specify boundary condition for electron orbitals.
Options:
'y' / periodic systems (crystalline solids)
'n' / isolated systems (molecules and nano-particles)

spin

character, default='unpolarized'

Available for the DFT/TDDFT based options of theory.
It specifies the spin state of the system, spin-unpolarized (closed shell) and spin-polarized (open shell).
Options
'unpolarized' / spin-unpolarized systems (default)
'polarized' / spin-polarized systems

al(3)

real(8), default=0d0

Available for the DFT/TDDFT based options of theory.
Spatial box size or lattice constants for cuboid cell (x, y, z).
For nonorthogonal cell, see al_vec1(3), al_vec2(3), al_vec3(3).

al_vec1(3)

al_vec2(3)

al_vec3(3)

real(8), default=0d0

Available for yn_periodic = 'y' in the DFT/TDDFT based options of theory.
Primitive lattice vectors for nonorthogonal cell. For cuboid cell, see al(3).

nstate

integer, default=0

Available for the DFT/TDDFT based options of theory.
of orbitals/bands to be calculated. In the time evolution calculation of dielectrics, only occupied orbitals are evolved even when more nstate is specified.

nelec

integer, default=0

Available for the DFT/TDDFT based options of theory.
Number of valence electrons in the system.

nelec_spin(2)

integer, Default=0

Available for the DFT/TDDFT based options of theory.
Number of up/down-spin electrons are specified by nelec_spin(1)/nelec_spin(2).
This option is incompatible with nelec. (If nelec_spin is specified, nelec is ignored.)

temperature

real(8), default=-1d0

Available for DFT-based options of theory.
It specifies the temperature for electrons. The value must be given using the unit of energy as specified in &units/unit_system.
The kelvin unit can also be used by the keyword temperature_k instead of temperature (see next).
Occupation numbers are updated in every SCF step in the following way.
temperature < 0 / the occupation numbers are fixed by nelec (appropriate for systems with energy gap).
temperature = 0 / redistribution of the occupation numbers by the step function (metallic system at zero temperature).
temperature > 0 / redistribution of the occupation numbers by the Fermi-Dirac distribution function.

temperature_k

[Trial] real(8), default=-1d0

Available for DFT-based options of theory.
The same as temperature but kelvin is used as the unit.

nelem

integer, default=0

Available for the DFT/TDDFT based options of theory.
Number of atomic elements in the system.

natom

integer, default=0

Available for the DFT/TDDFT based options of theory.
Number of atoms in the system.

file_atom_red_coor

[Trial] character, default='none'

Available for the DFT/TDDFT based options of theory.
Name of the file that contains atomic positions given in reduced coordinates. This option is incompatible with &system/file_atom_coor, &atomic_coor, and &atomic_red_coor.

file_atom_coor

[Trial] character, default='none'

Available for the DFT/TDDFT based options of theory.
Name of the file that contains atomic Cartesian coordinates (The unit is specified by &units/unit_system). This option is incompatible with &system/file_atom_coor, &atomic_coor, and &atomic_red_coor.

yn_spinorbit

character, default='n' (currently not available)

Options
'y' / enable
'n' / disable

yn_symmetry

character, default='n' (currently not available)

Options
'y' / enable
'n' / disable

absorbing_boundary

[Trial] character, default='none'

Available for the TDDFT based option of theory with orthogonal unit cell.
Absorbing boundary condition for electrons. (T. Nakatsukasa et al., J. Chem. Phys., 114, 2550 (2001))
Options:
'none' / disable (default)
'z' / absorbing boundary region is set in z direction for 'yn_periodic = 'y'

imaginary_potential_w0

real(8), default='0d0'

Available when absorbing_boundary options is not 'none'.
Strength of the absorbing (imaginary) potential.

imaginary_potential_dr

real(8), default='0d0'

Available when absorbing_boundary options is not 'none'.
Thickness of the absorbing (imaginary) potential. For absorbing_boundary='z', the absorbing region is 0 < z < imagnary_potential_dr and al(3)-imagnary_potential_dr < z < al(3)

&atomic_red_coor

Atomic coordinates in periodic systems ('yn_periodoc = 'y') are specified in reduced coordinates using the following format:

'Si' 0.00 0.00 0.00 1
'Si' 0.25 0.25 0.25 1
...

Here, the information of atoms is ordered in row, the first row for the first atom, etc. The number of rows must be equal to &system/natom. Atomic spicies are written in the first column although they are not used in the calculation. The second, third and fourth columns are reduced coordinates for the first, second and third directions, respectively. The fifth column is a serial number of the atom spieces, which is defined in &pseudo.
This option is incompatible with &system/file_atom_red_coor, &system/file_atom_coor, and &atomic_coor.

&atomic_coor

Atomic coordinates are specified in the same way as atomic_red_coor but with length dimension. The unit chosen by &units/unit_length is applied.
This option is incompatible with &system/file_atom_red_coor, &system/file_atom_coor, and &atomic_red_coor.

&pseudo

izatom(:)

integer, default=-1

Available for the DFT/TDDFT based options of theory.
Atomic number of the element. The size of array is equal to &system/nelem.

file_pseudo(:)

character, default='none'

Available for the DFT/TDDFT based options of theory.
File name of the pseudopotential file. The size of array is equal to &system/nelem.

lmax_ps(:)

integer, default=-1

Available for the DFT/TDDFT based options of theory.
Maximum angular momentum of pseudopotential projectors.
If not given, values specified in the pseudopotential file will be used. The size of array is equal to &system/nelem.

lloc_ps(:)

integer, default=-1

Available for the DFT/TDDFT based options of theory.
Angular momentum of the pseudopotential that will be treated as local. The size of array is equal to &system/nelem.

yn_psmask(:)

[Trial] character, default='n'

Available for the DFT/TDDFT based options of theory.
Fourier filtering for pseudopotentials. The size of array is equal to &system/nelem.
Options:
'y' / enable
'n' / disable

alpha_mask(:)

[Trial] real(8), default=0.8d0

Available for the DFT/TDDFT based options of theory.
Parameter for the Fourier filtering of the pseudopotential. The size of array is equal to &system/nelem.

gamma_mask(:)

[Trial] real(8), default=1.8d0)

Available for the DFT/TDDFT based options of theory.
Parameter for the Fourier filtering of the pseudopotential. The size of array is equal to &system/nelem.

eta_mask(:)

[Trial] real(8), default=15.0d0

Available for the DFT/TDDFT based options of theory.
Parameter for the Fourier filtering of the pseudopotential. The size of array is equal to &system/nelem.

&functional

xc

character, default='none'

Available for the DFT/TDDFT based options of theory.
Exchange-correlation functional to be used.
In the present version, functionals 'PZ', 'PZM' and 'TBmBJ' are available for both yn_periodic = 'y' and 'n' calculations in the adiabatic approximation.
Options:
'PZ': Perdew-Zunger LDA :Phys. Rev. B 23, 5048 (1981).
'PZM': Perdew-Zunger LDA with modification to improve sooth connection between high density form and low density one. :J. P. Perdew and Alex Zunger, Phys. Rev. B 23, 5048 (1981).
'TBmBJ': Tran-Blaha meta-GGA exchange with Perdew-Wang correlation. :Fabien Tran and Peter Blaha, Phys. Rev. Lett. 102, 226401 (2008). John P. Perdew and Yue Wang, Phys. Rev. B 45, 13244 (1992). This potential is known to provide a reasonable description for the bandgap of various insulators. For this choice, the additional mixing parameter 'cval' may be specified. See below.

cval

real(8), default=-1d0

Available for xc='TBmBJ'.
Mixing parameter in Tran-Blaha meta-GGA exchange potential. If cval is set to a minus value, the mixing-parameter is evaluated by the formula in the original paper [Phys. Rev. Lett. 102, 226401 (2008)], \left\langle |\nabla \rho(\mathbf{r};t)| / \rho(\mathbf{r};t) \right\rangle. However, note that the value may be different from that in all electron calculations.

cname

xname

character, default='none'

Available for theory='XXX'.
XXX

alibxc

alibx

alibc

character, default='none'

Available for the DFT/TDDFT based options of theory.
Since version 1.1.0, exchange-correlation functionals in Libxc library (http://www.tddft.org/programs/libxc/) have been usable in SALMON. At present, usable functionals are limited to LDA and GGA. For periodic systems, meta-GGA functionals are usable as well. To specify the exchange-correlation potentials of Libxc library, there are two ways. If the exchange and correlation potentials are given separately, you need to specify both alibx and alibc separately. If the exchange and correlation potentials are given as a combined set, you need to specify alibxc. We show below an example:
&functional
alibx = 'LDA_X'
alibc = 'LDA_C_PZ'
Note that, the hybrid functionals (hybrid gga/mgga) are not supported.

To use libxc libraries, --enable-libxc option must be added in excecuting configure. The available option of the exchange-correlation functionals are listed in the LibXC website. [See http://www.tddft.org/programs/libxc/functionals/]

&rgrid

dl(3)

real(8), default=0d0

Available for the DFT/TDDFT based options of theory.
Spacing of real-space grids.
This cannot be used together with &rgrid/num_rgrid.

num_rgrid(3)

integer, default=0

Available for the DFT/TDDFT based options of theory.
Number of real-space grids for each direction.
This cannot be used together with &rgrid/dl.

&kgrid

num_kgrid(3)

integer, default=1

Available for yn_periodic='y' in the DFT/TDDFT based options of theory.
Number of k-points (grid points of k-vector) for each direction discretizing the Brillouin zone.

file_kw

character, default='none'

Available for yn_periodic='y' in the DFT/TDDFT based options of theory.
File name for a file that includes user specified k-points. This file will be read if num_kgrid is equal to 0 or negative values. The file should be described in the following format :

8 #(number of k-points)
1 -0.50 -0.50 -0.50 0.1250 #(id, kx, ky, kz, weight)
2 -0.50 -0.50 0.00 0.1250
3 -0.50 0.00 -0.50 0.1250
4 -0.50 0.00 0.00 0.1250
5 0.00 -0.50 -0.50 0.1250
6 0.00 -0.50 0.00 0.1250
7 0.00 0.00 -0.50 0.1250
8 0.00 0.00 0.00 0.1250

&tgrid

nt

integer, Default=0

Available for 'dft_md' and TDDFT-based options of theory.
Number of total time steps for real-time propagation.

dt

real(8), Default=0d0

Available for 'dft_md' and TDDFT-based options of theory.
Time step size.

gram_schmidt_interval

integer, default=-1

Available for TDDFT-based options of theory.
Interval of a time step for the Gram-Schmidt orthonormalization of the orbitals during time evolution calculations. If this is set to a negative value, no Gram-Schmidt orthogonalization will be achieved. If this is set to zero, the Gram-Schumidt orthogonalization is carried out once at the initial step only. Usually this Gram-Schmidt orthogonalization is not necessary and should not be used.

&propagation

n_hamil

integer, default=4
Available for TDDFT-based options of theory.
Order of the Taylor expansion adopted for the propagation operator.

propagator

character, default=middlepoint

Available for TDDFT-based options of theory.
Choice of the propagator in the time evolution calculation.
Options:
middlepoint / Hamiltoinan at midpoint of two-times is used in the propagation if yn_predictor_corrector = 'y'. Hamiltoian at the time t is used if yn_predictor_corrector = 'n'.
aetrs / time-reversal symmetry propagator. [M.A.L. Marques, A. Castro, G.F. Bertsch, and A. Rubio, Comput. Phys. Commun., 151 60 (2003)].

yn_predictor_corrector

character, default='n'
Available for TDDFT-based options of theory.
Switch of the predictor-corrector method of TDDFT.
For meta-GGA functionals (xc='tbmbj'), the predictor corrector is automatically used even with yn_predictor_corrector='n'.
Options:
'y' / enable
'n' / disable

yn_fix_func

[Currently not available] character, default='n'
Available for 'dft_md' and TDDFT-based options of theory.
Switch not to update the Hamiltonian during the time evolution, i.e., ground state Hamiltonian is used during the propagation.
Options:
'y' / enable
'n' / disable

&scf

method_init_wf

character, default='gauss'

Available for 'dft' and 'dft_md' options of theory.
The generation method of the initial orbitals at the begening of the SCF iteration in DFT calculations. For a stable calculation of very large systems, multiple gaussian functions are preferred for a stable calculation.
Options:
gauss / single gauss function per orbital centered at a position determined by random numbers
gauss2 / two gauss functions per orbital centered at positions determined by random numbers
gauss3 / three gauss functions per orbital centered at positions determined by random numbers
gauss4 / four gauss functions per orbital centered at positions determined by random numbers
gauss5 / five gauss functions per orbital centered at positions determined by random numbers
gauss10 / ten gauss functions per orbital centered at positions determined by random numbers
random / a random number is assigned at each real-space grid point of orbitals

method_init_density

[Trial] character, default='wf'

Available for 'dft' and 'dft_md' options of theory.
Specifying how to generate the initial density to start the SCF iteration in the DFT calculation.
Supported for limited formats of pseudopotentials ('KY' and 'UPF').
Options:
wf / generate from the initial wavefunctions (cf. method_init_wf).
pp / generate from a superposition of the pseudo-atom densities.

iseed_number_change

integer, default=0

Available for 'dft' and 'dft_md' options of theory.
Change a seed of random numbers that are used to generate initial orbitals. The value specified by this parameter is added to the seed.

nscf

integer, Default=300

Available for 'dft' and 'dft_md' options of theory.
Number of maximum SCF iterations in the DFT calculation.

method_min

character, Default='cg'

Available for 'dft' and 'dft_md' options of theory.
Method for updating orbitals in the SCF iteration. At present only confjugate gradient method is implemented.
Options:
cg / Conjugate-Gradient(CG) method

ncg

integer, default=4

Available for 'dft' and 'dft_md' options of theory.
Number of interations of conjugate-gradient method in the SCF iteration.

ncg_init

integer, default=4

Available for 'dft' and 'dft_md' options of theory.
Number of interations of conjugate-gradient method for the first SCF step.

method_mixing

character, default='broyden'

Available for 'dft' and 'dft_md' options of theory.
Method to update density/potential in the scf iteration.
Options:
simple / Simple mixing method
broyden / modified Broyden method
pulay / Pulay method

mixrate

real(8), default=0.5d0

Available for method_mixing='simple' in 'dft' and 'dft_md' options of theory.
Mixing ratio for simple mixing.

nmemory_mb

integer, default=8

Available for method_mixing='broyden' in 'dft' and 'dft_md' options of theory.
Number of previous densities to be stored in the SCF iteration using the modified Broyden method. This must be less than 21.

alpha_mb

real(8), default=0.75d0

Available for method_mixing='broyden' in 'dft' and 'dft_md' options of theory.
A parameter of the modified Broyden method.

nmemory_p

integer, default=4

Available for method_mixing='pulay' in 'dft' and 'dft_md' options of theory.
Number of previous densities to be stored in the SCF iteration using the Pulay method.

beta_p

real(8), default=0.75d0

Available for method_mixing='pulay' in 'dft' and 'dft_md' options of theory.
A parameter of the mixing rate of the Pulay method.

yn_auto_mixing

character, default='n'

Available for 'dft' and 'dft_md' options of theory.
Switch to change the mixing rate automatically (i.e. automatic adjustments of mixrate/alpha_mb/beta_p)
Options:
'y' / enable
'n' / disable

update_mixing_ratio

real(8), default=3.0d0

Available for yn_auto_mixing='y' in 'dft' and 'dft_md' options of theory.
Threshold for the change of the mixing rate in yn_auto_mixing='y' option. The mixing-rate is reduced to half when the ratio of the density differences between the current and previous iteration steps is larger than update_mixing_ratio.

yn_subspace_diagonalization

character, default='y'

Available for 'dft' and 'dft_md' options of theory.
Switch for the subspace diagonalization during SCF iterations.
Options:
'y' / enable
'n' / disable

convergence

character, default='rho_dne'

Available for 'dft' and 'dft_md' options of theory.
Specify a quantity that is used for convergence check of the SCF iteration.
Options:
'rho_dne'/ Convergence is checked by sum_ix|rho(ix,iter)-rho(ix,iter-1)|dx/N. N is &system/nelec.
'norm_rho'/ Convergence is checked by the square of the norm of the density difference, ||rho_iter(ix)-rho_iter-1(ix)||2=sum_ix|rho(ix,iter)-rho(ix,iter-1)|2.
'norm_rho_dng'/ Convergence is checked by ||rho_iter(ix)-rho_iter-1(ix)||2/(number of grids). "dng" means "devided by number of grids".
'norm_pot'/ Convergence is checked by ||Vlocal_iter(ix)-Vlocal_iter-1(ix)||2, where Vlocal is Vh + Vxc + Vps_local.
'pot_dng'/ Convergence is checked by ||Vlocal_iter(ix)-Vlocal_iter-1(ix)||2/(number of grids).

threshold

real(8), default=1d-17 [a.u.] (for convergence='rho_dne') and -1 (for other options of convergence))

Available for 'dft' and 'dft_md' options of theory.
Threshold of convergence that is specified by convergence keyword.

nscf_init_redistribution

integer, default=10

Available for 'dft' and 'dft_md' options of theory.
Number of initial iterations during which a redistribution of the occupation number is suppressed in the finite temperature calculation.

nscf_init_no_diagonal

integer, default=10

Available for &scf/yn_subspace_diagonalization='y' in 'dft' option of theory.
Number of initial iterations during which the subspace diagonalization will not be carried out.

nscf_init_mix_zero

integer, default=-1

Available for 'dft' option of theory.
The density will not be mixed (i.e. fixed) during the given number of the SCF iteration, that is, orbitals are optimized without updating the density.

conv_gap_mix_zero

real(8), default=99999d0

Available if nscf_init_mix_zero is positive value in the 'dft' option of theory.
Specify a condition to quit the fixed density iteration forced by step_initial_mix_zero option. Mixing of the density will start after the band-gap energy exceeds this parameter for consecutive five SCF iteration steps.

&emfield

trans_longi

character, default='tr'

Available for yn_periodic='y' in 'maxwell' and TDDFT based options of theory.
Specify the treatment of the polarization in the time evolution calculation.
Options:
'tr' / Transverse
'lo' / longitudinal
'2d' / 2D maxwell-TDDFT method (for more details, see film_thickness of &maxwell)

ae_shape1

ae_shape2

character, Default='none'

Available for 'maxwell' and TDDFT based options of theory.
Envelope shape of the first/second pulse. 'Acos2' indicates a cosine square envelope for vector potential, and 'Ecos2' a cosine square envelope for electric field.
Options:
'impulse' / A weak impulsive field is applied at t=0. This will be used to explore linear response properties. The magnitude of the impulse can be specified by e_impulse.
'Acos2' / Envelope of cos2for a vector potential.
'Acos3' / Envelope of cos3for a vector potential.
'Acos4' / Envelope of cos4for a vector potential.
'Acos6' / Envelope of cos6for a vector potential.
'Acos8' / Envelope of cos8for a vector potential.
'Ecos2' / Envelope of cos2for an electric field.
'Asin2cos' [Trial] / Envelope of sin2with cosine type oscillation for a vector potential.
'Asin2_cw' [Trial] / Envelope of sin2at the beginning and continuous wave after that for a vector potential (for 'ae_shape1' only).
'input' [Trial] / read the vector potential as a numerical table with file_input1 option (for 'ae_shape1' only).
'none' / no incident field is applied.

If 'Ecos2' is adopted, 'phi_cep1' must be chosen either 0.75 or 0.25, since otherwise the time integral of the electric field (vector potential at the end of the pulse) does not vanishi. There is no such restriction for 'Acos2' pulses.

For yn_periodic='n', available choices are limited to 'impulse', 'Acos2', and 'Ecos2'.

file_input1

character, default=''

Available if ae_shape1='input' is specified and theory='tddft_pulse'.
Name of an input file that contains user-defined vector potential. The file must be a numerical table separated by blank, having four columns; the first column is time and second to fourth columns are Ax/c, Ay/c, Az/c, repsectively. All the quantities are written using the units specified by unit_system. '#' and '!' may be used for a comment line.
Note that a linear interpolation will be applied when the time step differs from that used in the calculation.

e_impulse

real(8), Default=1d-2 a.u.

Available for 'maxwell' and TDDFT based options of theory.
Magnitude of the impulse in the impulsive perturbation. This valiable has the dimention of momentum, energy*time/length.

E_amplitude1

E_amplitude2

real(8), default=0d0

Available for 'maxwell' and TDDFT based options of theory.
Maximum amplitude of electric field for the first/second pulse. This valiable has the dimension of electric field, energy/(length*charge). This cannot be set with &emfield/I_wcm2_1 (I_wcm2_2) simultaneously.

I_wcm2_1

I_wcm2_2

real(8), default=-1d0

Available for 'maxwell' and TDDFT based options of theory.
Maximum intensity (W/cm2) of the first/second pulse. This valiable cannot be set with &emfield/E_amplitude1 (E_amplitude2) simultaneously. For this quantity, a unit of W/cm2is adopted irrespective of &units\unit_system.

tw1

tw2

real(8), default=0d0

Available for 'maxwell' and TDDFT based options of theory.
Duration of the first/second pulse (edge-to-edge time length).
Note that this is not the FWHM duration.

omega1

omega2

real(8), default=0d0

Available for 'maxwell' and TDDFT based options of theory.
Mean photon energy (average frequency multiplied by the Planck constant) of the first/second pulse.

epdir_re1(3)

epdir_re2(3)

real(8), default=1d0, 0d0, 0d0

Available for 'maxwell' and TDDFT based options of theory.
Real part of the polarization unit vector for the first/second pulse.

epdir_im1(3)

epdir_im2(3)

real(8), default=0d0

Available for 'maxwell' and TDDFT based options of theory.
Imaginary part of the polarization unit vector for the first/second pulse. Using both real 'epdir_re1' and imaginary 'epdir_im1' parts of the polarization vector, circularly and general ellipsoidary polarized pulses may be described.

phi_cep1

phi_cep2

real(8), default=0d0

Available for 'maxwell' and TDDFT based options of theory.
Carrier envelope phase of the first/second pulse. It specifies the CEP in unit of 2\pi.

t1_t2

real(8), default=0d0
Available for 'maxwell' and TDDFT based options of theory.
Time-delay between the first and the second pulses.

t1_start

real(8), default=0d0

Available for 'maxwell' and TDDFT based options of theory.
Shift the starting time of the first pulse. (this is not available for multiscale option).

num_dipole_source

integer, default=0

Available for TDDFT based options of theory.
Number of radiation sources to mimic optical near fields. Maximum number is 2.

vec_dipole_source(3,num_dipole_source)

real(8), default=0d0

Available for TDDFT based options of theory.
Dipole vectors of the radiation sources mimicing optical near fields.

cood_dipole_source(3,num_dipole_source)

real(8), default=0d0

Available for TDDFT based options of theory.
Coordinates of the radiation sources mimicing optical near fields.

rad_dipole_diele

real(8), default=2d0 [a.u.]

Available for TDDFT based options of theory.
Radii of dielectric spheres of the radiation sources mimicing optical near fields.

&singlescale[Trial]

method_singlescale

character, default='3d'

Available for theory='single_scale_maxwell_tddft'.
Type of single-scale Maxwell-TDDFT method.
Options:
'3d' / 3-dimensional FDTD + TDDFT
'1d' / 1-dimensional FDTD (along the z axis) + TDDFT
'1d_fourier' / '1d' with 3D Fourier component of the vector potential

cutoff_G2_emfield

real(8), default=-1d0

Available for theory='single_scale_maxwell_tddft'.
Cutoff energy of Fourier component of the vector potential when method_singlescale='1d_fourier'.

yn_symmetrized_stencil

[Trial] character, default='n'

Available for theory='single_scale_maxwell_tddft'.
Switch to symmetrize the finite differences for the product of vector potential and orbitals, (\nabla A(r) \cdot \psi(r)). This option improves hermiticity of the Hamiltonian although computational cost increases.

yn_put_wall_z_boundary

[Trial] character, default='n'

Available for DFT/TDDFT based options of theory.
Option to put potential wall near the boundary planes at z=0 and z=``&system/al(3)``. This potential prevents electrons from crossing the z-boundary plane. In the single-scale Maxwell-TDDFT method, the electron density on the z-boundary plane harms the norm conservation of electrons due to the discontinuity of the vectorpotential. The wall is described using the square of cosine function.
Options:
'y' / put the potential wall
'n' / no potential wall

wall_height

real(8), default=100.0 [eV]

Available for yn_put_wall_z_boundary='y'.
The height of the potential wall.

wall_width

real(8), default=5.0 [Angstrom]

Available for yn_put_wall_z_boundary='y'.
The width of the potential wall defined by the length from the potential peak (z=0 and z=``&system/al(3)``) to the edge.

&multiscale

fdtddim

[Trial] character, default='1d'

Available for theory='multi_scale_maxwell_tddft' with yn_periodic='y'
Dimension of macroscopic scale system (Maxwell(FDTD) calculation) in multiscale Maxwell-TDDFT method.
Options:
'3d' / 3-dimensional FDTD for macroscopic electromagnetism [currently not available]
'1d' / 1-dimensional FDTD (along the x-axis) for macroscopic electromagnetism

nx_m

integer, default=1

Available for theory='multi_scale_maxwell_tddft' with yn_periodic='y'
Number of macroscopic grid points inside materials for x-direction.

ny_m

nz_m

[Trial] integer, default=1)

Available for theory='multi_scale_maxwell_tddft' with yn_periodic='y'
Number of macroscopic grid points inside materials for (y/z)-direction.

hx_m

real(8), default=0d0
Available for theory='multi_scale_maxwell_tddft' with yn_periodic='y'
Grid spacing of macroscopic coordinate for x-direction.
Variable hx_m is deprecated, and will be moved to &units/dl_em(1)

hy_m

hz_m

[Trial] real(8), default=0d0

Available for theory='multi_scale_maxwell_tddft' with yn_periodic='y'
Grid spacing of macroscopic coordinate for (y/z)-direction.
Variable hy_m and hz_m are deprecated, and will be moved to &units/dl_em(2:3)

nxvacl_m

nxvacr_m

integer, default=1/0

Available for theory='multi_scale_maxwell_tddft' with yn_periodic='y'
Number of macroscopic grid points for vacumm region.
nxvacl_m / nxvacr_m specifies the number for negative / positive x-direction in front of the material.

&maxwell

al_em(3)

real(8), default=0d0

Available for theory='maxwell'.
Size of simulation box in electromagnetic analysis.

dl_em(3)

real(8), default=0d0

Available for theory='maxwell' and theory='multi_scale_maxwell_tddft'.
Spacing of real-space grids in electromagnetic analysis.

dt_em

real(8), default=0

Available for theory='maxwell'.
Time step for electromagnetic analysis.

nt_em

integer, default=0

Available for theory='maxwell'.
Number of total time steps of time propagation in electromagnetic analysis.

boundary_em(3,2)

character, default='default'

Available for theory='maxwell' and theory='multi_scale_maxwell_tddft'.
Boundary condition in electromagnetic analysis. The first index(1-3 rows) corresponds to x, y, and z axes. The second index(1-2 columns) corresponds to bottom and top of the axes.
Options:
'abc' / absorbing boundary
'pec' / perfect electric conductor
'periodic' / periodic boundary

If &system/yn_periodic='n', 'default', 'abc', and 'pec' can be chosen, where 'default' automatically chooses 'abc'. If &system/yn_periodic='y', 'default', 'abc', and 'periodic' can be chosen, where 'default' automatically chooses 'periodic'. | When theory='maxwell', perfectly matched layer(PML) is used for 'abc'.

shape_file

character, default='none'

Available for theory='maxwell'.
Name of input shape file in electromagnetic analysis. The shape file can be generated by using FDTD_make_shape in SALMON utilities (https://salmon-tddft.jp/utilities.html).

media_num

integer, default=0

Available for theory='maxwell'.
Number of media in electromagnetic analysis.

media_type(:)

character, default='vacuum'

Available for theory='maxwell'.
media_type(n) spesifies type of n-th media in electromagnetic analysis.
Options:
'vacuum'
'constant media'
'pec'
'lorentz-drude'
If 'lorentz-drude' is chosen, linear response calculation is feasible by setting &emfield/ae_shape1 or ae_shape2='impulse'.

epsilon_em(:)

real(8), Default=1d0

Available for theory='maxwell' and for TDDFT based options of theory with trans_longi='2d'.
For theory='maxwell', epsilon_em(n) spesifies relative permittivity of n-th media in electromagnetic analysis.
For TDDFT based options of theory with trans_longi='2d', the relative permittivity of the transparent media on both sides of the film is specified by epsilon_em(1) and epsilon_em(2), respectively.

mu_em(:)

real(8), default=1d0

Available for theory='maxwell'.
mu_em(n) spesifies relative permeability of n-th media in electromagnetic analysis.

sigma_em(:)

real(8), default=0d0

Available for theory='maxwell'.
sigma_em(n) spesifies conductivity of n-th media in electromagnetic analysis.

pole_num_ld(:)

integer, default=1

Available for theory='maxwell'.
pole_num_ld(n) spesifies number of poles of n-th media, available for type_media(n)='lorentz-drude' in electromagnetic analysis.

omega_p_ld(:)

real(8), default=0d0

Available for theory='maxwell'.
omega_p_ld(n) spesifies plasma frequency of n-th media, available for type_media(n)='lorentz-drude' in electromagnetic analysis.

f_ld(:,:)

real(8), default=0d0

Available for theory='maxwell'.
f_ld(n,m) spesifies m-th oscillator strength of n-th media, available for type_media='lorentz-drude' in electromagnetic analysis. The first index is the media ID whose maximum value is given by media_num. The second index is the pole ID whose maximum value is given by pole_num_ld(n).

gamma_ld(:,:)

real(8), default=0d0

Available for theory='maxwell'.
gamma_ld(n,m) spesifies m-th collision frequency of n-th media, available for type_media(n)='lorentz-drude' in electromagnetic analysis. The first index is the media ID whose maximum value is given by media_num. The second index is the pole ID whose maximum value is given by pole_num_ld(n).

omega_ld(:,:)

real(8), default=0d0

Available for theory='maxwell'.
omega_ld(n,m) spesifies m-th oscillator frequency of n-th media, available for type_media(n)='lorentz-drude' in electromagnetic analysis. The first index is the media ID whose maximum value is given by media_num. The second index is the pole ID whose maximum value is given by pole_num_ld(n).

wave_input

character, default='none'

Available for theory='maxwell'.
If 'source', the incident pulse in electromagnetic analysis is generated by the incident current source.

ek_dir1(3)

ek_dir2(3)

real(8), default=0d0

Available for theory='maxwell'.
Propagation direction of the first/second pulse.

source_loc1(3)

source_loc2(3)

real(8), default=0d0

Available for theory='maxwell'.
Location of the incident current source of the first/second pulse. Note that the coordinate system ranges from -al_em/2 to al_em/2 for &system/yn_periodic='n' while ranges from 0 to al_em for &system/yn_periodic='y'.

obs_num_em

integer, default=0

Available for theory='maxwell'.
Number of observation points in electromagnetic analysis. From the obtained results, figure and animation files can be generated by using SALMON utilities (https://salmon-tddft.jp/utilities.html).

obs_samp_em

integer, default=1

Available for theory='maxwell'.
Sampling time-step of the observation in electromagnetic analysis.

obs_loc_em(:,3)

real(8), default=0d0

Available for theory='maxwell'.
obs_loc_em(n,1:3)=x,y,z spesifies location of the n-th observation point in electromagnetic analysis. Note that the coordinate system ranges from -al_em/2 to al_em/2 for &system/yn_periodic='n' while ranges from 0 to al_em for &system/yn_periodic='y'.

yn_obs_plane_em(:)

character, default='n'

Available for theory='maxwell'.
Spesify whether or not to generate output of the electrmagnetic fields on the planes (xy, yz, and xz planes) for n-th observation point. This option must be 'y' for generating animation files by using FDTD_make_figani in SALMON utilities (https://salmon-tddft.jp/utilities.html).
Options:
'y'
'n'

yn_obs_plane_integral_em(:)

character, default='n'

Available for theory='maxwell'.
Specify whether or not to generate output of the spatial integration of electromagnetic fields on the planes (xy, yz, and xz planes) for n-th observation point.
Options:
'y'
'n'

yn_wf_em

character, default='y'

Available for theory='maxwell'.
Switch of a window function for linear response calculation. Available for &calculation/theory=maxwell.
Options:
'y'
'n'

film_thickness

real(8), default=0d0

Available for TDDFT based options of theory with trans_longi='2d'.
Thickness of the film for the 2D maxwell-TDDFT method.
The relative permittivity of the transparent media on both sides of the film can be specified by epsilon_em(1) and epsilon_em(2), respectively.

media_id_pml(3:2)

integer, default=0

Available for theory='maxwell'.
Media ID used in PML. The first index(1-3 rows) corresponds to x, y, and z axes. The second index(1-2 columns) corresponds to bottom and top of the axes.

media_id_source1

media_id_source2

integer, default=0
Available for theory='maxwell'.
Media ID used in incident current source1/source2 to generate the first/second pulse.

&analysis

projection_option

character, default='no'

Available for TDDFT based options of theory.
Methods of projection to analyze the excited states (e.g. the number of excited electrons).
Options:
'no' / no projection.
'gs' / projection to eigenstates of ground-state Hamiltonian.
'rt' / projection to eigenstates of instantaneous Hamiltonian. [currently not available]

out_projection_step

integer, default=100

Available for TDDFT based options of theory.
Resuts of the projection analysis will be outputted everty out_projection_step step during the time-propagation.

nenergy

integer, default=1000

Number of energy grid points for frequency-domain analysis. This parameter is used, for examples, in theory='tddft_response' and theory='maxwell'.

de

real(8), Default=0.01d0 (eV)

Energy grid size for frequency-domain analysis.
This parameter is used, for examples, in theory='tddft_response' and theory='maxwell'.

out_rt_energy_step

integer, default=10

Available for the TDDFT based option of theory.
Total energy is calculated and printed every out_rt_energy_step time steps.

yn_out_psi

character, default='n'

Available for theory='dft'.
Switch for output of orbitals.
Options:
'y' / enable
'n' / disable
The format of the output is specified by &analysis/format_voxel_data.

yn_out_dos

character, default='n'

Available for theory='dft'.
Switch for output of density of states.
Options:
'y' / enable
'n' / disable

yn_out_pdos

character, default='n'

Available for theory='dft'.
Switch for output of projected density of states.
Options:
'y' / enable
'n' / disable

yn_out_dos_set_fe_origin

character, default='n'

Available when yn_out_dos='y' or yn_out_pdos='y'.
Switch to set the Fermi energy to zero.
Options:
'y' / enable
'n' / disable
This option is not available if the temperature is not set in the calculation.

out_dos_start

real(8), default=-1d10 (eV)

out_dos_end

real(8), default=1d10 (eV)

Available when yn_out_dos='y' or yn_out_pdos='y'.
Lower/Upper bound of the energy range for the density of states spectra.
If this value is lower/higher than a specific value near the lowest/highest energy level, this parameter is re-set to the value.

out_dos_nenergy

integer, default=601

Available when yn_out_dos='y' or yn_out_pdos='y'.
Number of energy points sampled in the density of states spectra.

out_dos_function

character, default='gaussian'

Available when yn_out_dos='y' or yn_out_pdos='y'.
Choice of the smearing function for the density of states spectra.
Options:
gaussian / Gaussian function
lorentzian / Lorentzian function

out_dos_width

real(8), default=0.1d0 [eV]

Available when yn_out_dos='y' or yn_out_pdos='y'.
Smearing width used in the density of states spectra.

yn_out_dns

character, default='n'

Available for theory='dft'.
Switch to output electron density distribution of the ground state.
Options:
'y' / enable
'n' / disable

yn_out_dns_rt

character, default='n'

Available when theory='dft_md' or 'theory=tddft_pulse'.
Switch to output electron density distribution during the time-propagation.
Options:
'y' / enable
'n' / disable

out_dns_rt_step

integer, default=50

Available when theory='dft_md' or 'theory=tddft_pulse'.
Density is outputted every out_dns_rt_step steps.

yn_out_dns_ac_je

character, default='n'

Available for theory='single_scale_maxwell_tddft'.
Switch to print the electron density, vector potential, electronic current, and ionic coordinates every outdns_dns_ac_je_step time steps.
Options:
'y' / enable
'n' / disable
The data written in binary format are divided into files corresponding to the space-grid parallelization number.

out_dns_ac_je_step

integer, default=50

Available for theory='single_scale_maxwell_tddft'.
Electron density, vector potential, electronic current, and ionic coordinates are outputted every outdns_dns_ac_je_step time steps.

yn_out_dns_trans

[currently not available] character default='n'

Available for theory='tddft_pulse'.
Switch to calculate transition density at specified frequency omega (specified by out_dns_trans_energy), drho(r,omega)=FT(rho(r,t)-rho_gs(r))/T.
Options:
'y' / enable
'n' / disable

out_dns_trans_energy

[currently not available] real(8), default=1.55d0 [eV]

Available for theory='tddft_pulse'.
A frequency to output drho(r,omega)=FT(rho(r,t)-rho_gs(r))/T.

yn_out_elf

character, default='n'

Available for theory='dft'.
Switch to output the electron localization function.
Options:
'y' / enable
'n' / disable

yn_out_elf_rt

character, default='n'

Available for theory='dft_md', 'tddft_pulse'.
Switch to output the electron localization function during the time propagation every out_elf_rt_step time steps.
Options:
'y' / enable
'n' / disable

out_elf_rt_step

integer, default=50

Available for theory='dft_md', 'tddft_pulse'.
Electron localization function during the time propagation is outputted every out_elf_rt_step time steps.

yn_out_estatic_rt

character, default='n'

Available for theory='tddft_pulse'.
Switch to print the static electric field during the time propagation every out_estatic_rt_step time steps.
Options:
'y' / enable
'n' / disable

out_estatic_rt_step

integer, default=50

Available for theory='tddft_pulse'.
The static electric field during the time propagation is outputed every out_estatic_rt_step time steps.

yn_out_rvf_rt

character, default='n'

Available for TDDFT based options and 'dft_md' option of theory.
Switch to print the coordinates[A], velocities[au], forces[au] of atoms during time-propagation in SYSname_trj.xyz every out_rvf_rt_step time steps.
Options:
'y' / enable
'n' / disable
If yn_md='y', this option is automatically turned on.

out_rvf_rt_step

integer, default=10

Available for TDDFT based options and 'dft_md' option of theory.
The coordinates[A], velocities[au], forces[au] of atoms during time-propagation are outputed in SYSname_trj.xyz every out_rvf_rt_step time steps.

yn_out_tm

[Trial] character, default='n'

Available for yn_periodic='y' with theory='dft'.
Switch to calculate and print the transition matrix elements between occupied and virtual orbitals to SYSname_tm.data after the ground state calculation.
Options:
'y' / enable
'n' / disable

out_ms_step

integer, default=100

Available for theory='multi_scale_maxwell_tddft'.
Some quantities are printed every out_ms_step time step in the Maxwell-TDDFT multiscale calculations.

format_voxel_data

character, default='cube'

Available for yn_out_psi='y', yn_out_dns(_rt)='y', yn_out_dns_ac_je='y', yn_out_elf(_rt)='y', yn_out_estatic_rt='y'.
Option of the file format for three-dimensional volumetric data.
'avs' / AVS format
'cube' / cube format
'vtk' / vtk format

nsplit_voxel_data

integer, default=1

Available for format_voxel_data='avs'.
Number of separated files for three dimensional data.

yn_lr_w0_correction

[Trial] character, default='n' (currently not available)

Available for yn_periodic='y' with theory='tddft_response'.
Apply correction around zero frequency of dielectric function to suppress numerical error.
Options:
'y' / enable
'n' / disable

yn_out_perflog

character, default='y'

Available for all theory
Switch to print the performance log of routines and modules.
Options:
'y' / enable
'n' / disable

format_perflog

character, default='stdout'

Available for yn_out_perflog = 'y'
The output format of performance log.
Options:
'stdout' / standard output unit
'text' / save as a text file
'csv' / save as a csv format file

&poisson

layout_multipole

character, Default=3

Available for yn_periodic='n' in DFT and TDDFT based options of theory.
This papameter specify how to achieve multipole expansioin in the Hartree potential calculation.
Options:
1/ A single pole at the center.
2/ Multipoles are set at each center of atoms.
3/ Multipoles are set at the center of mass of electrons in prepared cuboids in each process.

num_multipole_xyz(3)

integer, default=0

Available for yn_periodic='n' in DFT and TDDFT based options of theory.
Number of multipoles. When default is set, the number of multipoles is calculated automatically.

lmax_multipole

[Trial] integer, default=4

Available for yn_periodic='n' in DFT and TDDFT based options of theory.
A maximum order of the multipole expansion to prepare boundary condition of Poisson equation.

threshold_cg

real(8), default=1d-15 [a.u.]

Available for yn_periodic='n' in DFT and TDDFT based options of theory.
A threshold for the convergence of the Hartree-cg calculation. A quantity examined is given by ||tVh(i)-tVh(i-1)||^2/(number of grids).

&ewald

newald

integer, default=4

Available for yn_periodic='y' in DFT/TDDFT based options of theory.
Parameter of the Ewald method for the ion-ion Coulombic interaction. Short-range part of the Ewald sum is calculated within newald-th nearlist neighbor cells.

aewald

real(8), default=0.5d0 [a.u.]

Available for yn_periodic='y' in DFT/TDDFT based options of theory.
Square of range separation parameter for Ewald method (This parameter is given only in atomic unit).

cutoff_r

real(8), default=-1d0

Available for yn_periodic='y' in DFT/TDDFT based options of theory.
Cut-off length in real-space. The length is automatically determined if cutoff_r < 0.

cutoff_r_buff

real(8), default=2d0 [a.u.]

Available for yn_periodic='y' in yn_md='y' or in theory='dft_md'.
Buffer length in radius for book-keeping for real-space interaction.

cutoff_g

real(8), Default=-1d0

Available for yn_periodic='y' in DFT/TDDFT based options of thddeory.
Cut-off in G-space in the Ewald method. No cut-off in default.

&opt[Trial]

nopt

integer, default=100

Available for yn_opt='y' in theory='dft'.
The maximum step number of geometry optimization.

convrg_opt_fmax

real(8), default=1d-3 (a.u.)

Available for yn_opt='y' in theory='dft'.
Convergence threshold of geometry optimization is specified for the maximum force acting on atoms.

max_step_len_adjust

real(8), default=-1d0

Available for yn_opt='y' in theory='dft'.
Set maximum optimization step length (if positive number is given)

&md[Trial]

ensemble

character, default='NVE'

Available for yn_md='y' or theory='dft_md'.
Ensemble in MD option:
Options:
NVE/ NVE ensemble (constant energy and volume system)
NVT/ NVT ensemble (constant temperature and volume system)

thermostat

character, default='nose-hoover'

Available for yn_md='y' or theory='dft_md'.
Thermostat in "NVT" option:
Options:
nose-hoover/ Nose-Hoover thermostat

step_velocity_scaling

integer, default=-1

Available for yn_md='y' or theory='dft_md'.
Time step interval for velocity-scaling. Velocity-scaling is applied if this is set to positive.

step_update_ps

integer, default=10

Available for yn_md='y' or theory='dft_md'.
Time step interval for updating pseudopotential (Larger number reduces computational time but increases inaccuracy).

temperature0_ion_k

real(8), Default=298.15d0 [K]

Available for yn_md='y' or theory='dft_md'.
Setting ionic temperature in unit of [K] for NVT ensemble, velocity scaling and generating initial velocities.

yn_set_ini_velocity

character, Default='n'

Available for yn_md='y' or theory='dft_md'.
Switch to generate initial velocities.
Options:
y/ Generate initial velocity with Maxwell-Bortzman distribution
n/ disable

file_ini_velocity

[Trial] character, default='none'

Available for yn_md='y' or theory='dft_md'.
File name for reading initial velocities. This is read if the file name is given, then, the priority is higher than use of set_ini_velocity and restart data of velocities. The format is simply vx(iatom) vy(iatom) vz(iatom) in each line. The order of atoms must be the same as the given coordinates in the main input file. In case of using nose-hoover thermostat, a thermostat variable should be put at the last line (all atomic unit).

thermostat_tau

real(8), default=1d0 [fs]
Available for yn_md='y' or theory='dft_md'.
Parameter in Nose-Hoover method: controlling time constant for temperature.

yn_stop_system_mom

character, default='n'

Available for yn_md='y' or theory='dft_md'.
Center of mass is fixed every time step.
Options:
y/ enable
n/ disable

&jellium

yn_jm

character, default='n'

Available for the DFT/TDDFT based options of theory.
Switch to use jellium model.
Options:
y/ enable
n/ disable
When yn_jm='y', &functional/xc must be 'pz'.

yn_charge_neutral_jm

character, default='y'

Available for yn_jm='y' in the DFT/TDDFT based options of theory.
Option to enforce exact charge neutrality :
Options:
y/ enable. rs_bohr_jm is modified to fulfill exact charge neutrality.
n/ disable. rs_bohr_jm is not modified, and there may appears small charge-neutrality error.

yn_output_dns_jm

character, default='y'

Available for yn_jm='y' in the DFT/TDDFT based options of theory.
Switch to output positive background charge density.
Options:
y/ enable
n/ disable

shape_file_jm

character, default='none'

Available for yn_jm='y' in the DFT/TDDFT based options of theory.
Name of input shape file that contains positive background charge density to be used in the jellium model calculations. The shape file can be generated by using FDTD_make_shape in SALMON utilities (https://salmon-tddft.jp/utilities.html). When shape_file_jm='none', the shape of the positive background charge density is specified by sphere_nion_jm and sphere_loc_jm which generate spherical shapes.

num_jm

integer, Default=0

Available for yn_jm='y' in the DFT/TDDFT based options of theory.
When shape_file_jm is not 'none', num_jm specifies number of media used in the jellium model. When shape_file_jm='none', num_jm specifies number of spherical shapes.

rs_bohr_jm(:)

real(8), default=0d0

Available for yn_jm='y' in the DFT/TDDFT based options of theory.
When shape_file_jm is not 'none', rs_bohr_jm(n) spesifies the Wigner-Seitz radius of n-th media. When shape_file_jm='none', rs_bohr_jm(n) spesifies the Wigner-Seitz radius of n-th sphere.

sphere_nion_jm(:)

integer, default=0

Available for yn_jm='y' and shape_file_jm='none' in the DFT/TDDFT based options of theory. sphere_nion_jm(n) spesifies ion number for n-th sphere. At present, only neutral systems can be treated.

sphere_loc_jm(:,3)

real(8), default=0d0

Available for yn_jm='y' and shape_file_jm='none' in the DFT/TDDFT based options of theory. sphere_loc_jm(n,1:3)=x,y,z spesifies location of center of mass for n-th sphere. Note that the coordinate system ranges from -al/2 to al/2 for &system/yn_periodic='n' while ranges from 0 to al for &system/yn_periodic='y'.

&code

yn_want_stencil_hand_vectorization

character, default='y'

Switch to use hand-vectorized optimization code of stencil in the hamiltonian calculation.
SALMON checks if the calculation can use the hand-vectorized code. If it fails, SALMON will use a typical implementation.

yn_want_communication_overlapping

character, default='n'

Available for theory='tddft*' or '*maxwell_tddft'
Switch to use computation/communication overlap algorithm to improve the performance of stencil in the hamiltonian calculation.
SALMON checks if the calculation can use the overlap algorithm. If it fails, SALMON will uses a non-overlap algorithm.

stencil_openmp_mode

character, default='auto'

This option selects an OpenMP parallelization mode of stencil in the hamiltonian calculation.
Options:
auto / SALMON decides the parallelization target automatically.
orbital / OpenMP parallelization is applied to orbital (and k-point) loop.
rgrid / OpenMP parallelization is applied to real-space grid loop.

current_openmp_mode

character, default='auto'

This selects an OpenMP parallelization mode of the current calculation.
Options:
auto / SALMON decides the parallelization target automatically.
orbital / OpenMP parallelization is applied to orbital (and k-point) loop.
rgrid / OpenMP parallelization is applied to real-space grid loop.

force_openmp_mode

character, default='auto'

This selects an OpenMP parallelization mode of the force calculation.
Options:
auto / SALMON decides the parallelization target automatically.
orbital / OpenMP parallelization is applied to orbital (and k-point) loop.
rgrid / OpenMP parallelization is applied to real-space grid loop.