The inputcard contains all the information needed on the system and the calculation. In this paragraph we analyze all the input parameters and the construction of this file.
A few remarks on the format of this file: Comments can be included and are ignored from the program, only data after the codewords are read in. You can place the codewords anywhere in the inputcard. Codewords have to be in uppercase so you can keep comments in lowercase. If a codeword is directly followed by an equal sign (=) the number(s) in the same line following the codeword will be read in. If no equal sign is present the value which is exactly below the codeword will be read in. This is explained more in subroutine "ioinput.f".
The first line contains the so called
running options (lines 1-2) in the example. There is a list of options mostly
used to take out a specific quantity or impose a condition to
the calculation. The program for the RUNNING OPTIONS always reads
everything in strings of 8 characters, so be careful when you have more
than one option. The most important running options are
| COMPLEX | Complex band structure plots (1 iteration) |
| BAND-STR | Band structure plots (1 iteration) |
| symG(k) | |
| T_hcore | |
| rigid-ef | The Ef is fixed during iterations |
| full inv | doesn't use O(N) algorithm for matrix inversion |
| this is usually used for bulk calculations | |
| but it depends on the form of the Dyson equation | |
| for long supercells you can use the O(N) algorithm | |
| sparse | Sparse matrix inversion |
| (not possible with decimation yet | |
| deci-out | write out the t-matrix in 'decifile' |
| DECIMATE | use decimation technique |
| DECIMATEONEBULK | Speed up decimation, step information used for |
| the left bulk is just copied to the right one. | |
| cut Four | |
| DOS-EF | |
| DOS REF | |
| QDOS | q-dos calculation (1 iteration ) |
| CURRENT | current calculation (1 iteration) |
| CONDUCT | |
| GENPOT | generates the file fort.3 which contains the potential |
| in a special format. Refer to the part on the | |
| VORONOI for more details. |
LMAX : value of l-cut-off for wavefunctions, a value of 2*lmax is used for the charge moments. Normally 2 is good only for the spin moments, 3 for the energies and someone should use at least 4 for the forces NSPIN: number of spins (2: spin-polarized, 1: paramagnetic) NATYP: number of inequivalent sites in the system (must be the same as NAEZ) KMT=3 (not used anymore)
Description of the lattice
ALATBASIS: lattice parameters in atomic units (a, b/a, c/a)
BASISCALE: the atomic coordinates are divided by the BASISCALE numbers
LATTICE: Gives the name in the k-points file, it can be any integer positive number
BRAVAIS: Bravais vectors in terms of ALATBASIS, Each line corresponds to one vector.
If someone does surfaces, he has just a two dimensional periodicity, the third
Bravais lattice should be (0.0, 0.0, 0.0) and the 
component of the first two
Bravais vectors should be also 0.0
NAEZ: same as NATYP
NEMB: not used
NEMBZ: not used
KAOEZ: not used
CARTESIAN : f or t (false or true).
false: if the atomic coordinates and
the vectors are expressed in units of the Bravais vectors (Wyckoff coordinates).
true: if the atomic coordinates are Cartesian coordinates
independent of the Bravais vectors.
RBASIS: coordinates of the Basis atoms, in 2d geometry these are the
coordinates of the different atoms which have the same
2d geometry (i.e. belong in different planes, or in different
sub lattices in the same plane if we have more atoms in the same plane)
Important for the scaling is that in case of 2d geometry
i.e. INTERFACE=.true. the z-component of the BRAVAIS
is always zero. This means that in this case the z-component
of RBASIS is always in Cartesian coordinates.
SCALING : scaling (usually not used) Scaling done with BASISCALE
INTERFACE: true (T) or false (F).
T : 2 - D Geometry
F : 3 - D Geometry (Periodically repeated)
The following variables are used in the case of 2D geometry. In the case of
slab calculation the information is used to produce the TB-clusters which for
the
boundary atoms extend outside the
slab region. In the case of semi-infinite geometry (decimation) they have the
extra role of defining the lattice outside our active region and this
information is used to calculate the electrostatic potential of the
semi-infinite system which converges fast since you have to add charge neutral
blocks of layers. Look later in the manual
for a more detailed description of the role of these parameters.
NRIGHTHO : number of layers taken into account right of the system.
NLBASIS : number of basis vectors to construct the right part of the system
NLEFTHOS : same as NRIGHTHO, but left of the system
NRBASIS: same as NLBASIS for the right part.
LEFTBASIS (X, Y, Z, REFPOT) vectors to construct the next layer on the left side.
Gives the coordinate of the inequivalent sites on the layer relative to the origin.
RIGHBASIS: (X, Y, Z, REFPOT) same but on the right side.
ZPERIODL: vector to construct the next layer on the left side.
Gives the coordinate of the inequivalent sites on the layer
relative to the origin.
ZPERIODR: same on the right side
Parameters for the TB-structure constants cluster
RCLUSTZ,
RCLUSTXY : radius (scaled with alat) determining the size of
the TB-cluster (along z axis, and x y axes)
(if equal, spherical cluster else cylindrical cluster)
The size of the cluster is written out, and it also affects
the size of the principal layer
ATOMINFO : description of core states and atom depending properties
The next line is comment line and then NATYP lines are needed
Z : nuclear charge
LMXC: lmax core,
argon core: lmxc = 1
krypton: lmxc = 2
xenon: lmxc = 3
KFG : configuration of core electrons:
argon core: 3 3 0 0=3s,3p (3d series)
krypton core: 4 4 3 0=4s,4p,3d (4d series)
xenon core+4f: 5 5 4 4=5s,5p,4d,4f (5d series)
CLS: type of cluster, used in case of structures with more than one type of TB-cluster
This information is not used by the voronoi program but is needed in the kkr.
The voronoi provides this info in it's output and should be taken from there.
REFPOT: used to set reference potentials of different heights or
different MT radii for each basis atom. More potentials are needed in the
4Ryshift file in this case usually set to 1
NTC : indexes different shape functions in case of atoms with
different geometric environment this information is provided by voronoi also.
Look at the part on how to get the initial potential and the shape functions
for more details.
FAC: not used set to 1.00
IRNS : number of radial points where the potential is non spherical,
but counted from the maximum radius in. Usually set to 208 or 278
RMT: Muffin-tin sphere in the case of FP calculations. This is important and
sets up the so-called muffin-tinization of the potential, used only by voronoi
WEIGHT: weight of each atom when the voronoi polyhedra are constructed.
For both RMT and WEIGHT look at the part on how to get the initial potential
and the shape functions for more details.
Energy contour
EMIN: lower energy of contour
EMAX: Maximum energy (in self consistent calculations
The Fermi level is read from the potential card). Attention the energies
for the integration should be such that they do not include any core states and
do not leave out any valence states. You can see the energetic position
of the core states in the potential cards.
In case of DOS calculation EMAX is used.
For the energy integration a rectangular contour with npt2 points parallel
to the real axis is used.
TK: temperature in Kelvin (determines the distance from the real axis)
Look at Phys. Rev. B 52, 11502 (1995)
NPOL: number of Matsubara poles for complex integration with temperature
if NPOL=0, the DOS if written out with NPT2 energy points.
NPT1,2,3: number of energies points for complex integration
npt2 points parallel to real axis,
npt1 points used to go up from the real axis
npt3 points used to go down to the real axis
IRNUMX: obsolete set to 10!
ITCCOR: Iterations for determining the core state energy, 40 is enough.
IRPCORE: Print information about the core states
IFILE:
IPE: not used ???
KWS: 1 for MT, 2 for ASA
KSHAPE: shape function for exact treatment of the potential
0=ASA,
1=the potential is replaced by a constant at the RMT (this is useful sometimes)
2=correct Full Potential
IRM: number of points on the radial space mesh in the MT sphere.
Normally we use 353 points for the ASA calculations and 484 for the FP Calculations
INS: 0 for ASA, 1 for FP
ICST: number of iterations in the Born scheme (for FP calculation).
The Born scheme is used to derive the 
wavefunctions needed
for the FP calculation from the 
wavefunctions (solutions of the
radial Schrödinger eauqtion) using the Lippman-Schwinger. Normally 2-3
iterations are enough. The criterion is converged total energies and/or forces
INSREF: 0 MT screening, 1 FP screening (set it to 0 )
KCOR: core relaxation:
KCOR=0 no core relaxation (frozen core calculation)
KCOR=1 SRA core relaxation
KCOR=2 non-SRA core relaxation
KVREL: Scalar relativistic calculation
KVREL=0 non relativist calculation
KVREL=1 SRA calculation
KEXCOR: exchange-correlation (0=von Barth Hedin, 1=Morruzi, Janak Williams,
2=Vosko, Wilk Nusair, planed GGA!)
KHYPERF: calculation of hyperfine field (0=no, 1=yes)
KHFIELD: external magnetic field
KEFG: calculation of Electric field gradient
KTE: calculation of total energy (0=no, 1=yes)
KPRE : writing out of total energy (0=no, 1=yes)
KVMAD: obsolete (set to 0)
KSCOEFF : set to 0
IPOTOUT: if set to 1 write out the potential after each iteration
IGREEFUN: write out the Green function (0=no, 1=yes)
ICC: sets the center of the cluster for impurity Green's function output
set it to 1 in case of impurity GF calculation
ISHIFT: set to 0
HFIELD: external magnetic field
VCONST:
KFORCE: calculates and writes out the Forces on each atom (0=no, 1=yes)
BZDIVIDE: divides the reciprocal lattice vectors and defines
the number of k-points in the IBZ. In reality four sets of k-points are used each
one corresponding on a different part of the integration over the complex
Energies surface. The most dense one is used near the Fermi level where we
need bigger accuracy. First the directory /mesh is searched for existing
files which have the same LATTICE name (look LATTICE) and the same
number of k-points
NSTEPS: number of iterations
IMIX: type of mixing:
0: linear mixing
1: Tchebychef mixing after 19 iteration
3: Broyden's first method
4: Broyden's second method
5: generalized Anderson mixing. This is the most powerful and quick method
but should be used only when the starting potential has been converged
up to the E-2 or E-3.
STRMIX: mixing parameter for linear mixing (0.01 typical value)
FCM : The moments are mixed with a FCM*STRMIX factor
QBOUND: convergence quality required
BRYMIX: Broyden mixing factor, normally it should be 2-3 times the STRMIX parameter
ITDBRY : number of steps taken into account for Broyden mixing
Filenames
FILES: input files 40 char long 4Ryshift I12 potential I13 madelung I40 shape I19 scoef I25
