Molecule - Introduction to the DFTB package

  • : preg_replace(): The /e modifier is deprecated, use preg_replace_callback instead in /home/icode/Documents/icode_web/modules/codefilter/codefilter.module on line 121.
  • : preg_replace(): The /e modifier is deprecated, use preg_replace_callback instead in /home/icode/Documents/icode_web/modules/codefilter/codefilter.module on line 122.
  • : preg_replace(): The /e modifier is deprecated, use preg_replace_callback instead in /home/icode/Documents/icode_web/modules/codefilter/codefilter.module on line 126.
  • : preg_replace(): The /e modifier is deprecated, use preg_replace_callback instead in /home/icode/Documents/icode_web/modules/codefilter/codefilter.module on line 127.

A short manual



Description of the DFTB package

The DFTB package is a density-functional based tight-binding scheme which can be used for clusters and supercells. It runs in the standard tight-binding as well as in the charge-self-consistent (scc) mode. Implementations in f77 are available for single-processor machines and for the T3E.
The program is called dftb. It has been developed by a variety of scientists at the Technical Universities in Chemnitz and Dresden. Here is an example of an input file for dftb. Replace the dummy variables by their appropriate values and call dftb < dftb.input

file dftb.input, for relaxation of a tiophene molecule containing S, C and H atoms

rel_mode force scc_on scctol read_charges
'structure'
electrons
moveatoms
'out'
3 2 1
'~/slko/ss_me.spl'
'~/slko/sc_me.spl'
'~/slko/sh_me.spl'
'~/slko/cs_me.spl'
'~/slko/cc_me.spl'
'~/slko/ch_me.spl'
'~/slko/sh_me.spl'
'~/slko/ch_me.spl'
'~/slko/hh_me.spl'
stepwidth atemp etemp vwscale nofsteps

rel_mode INTEGER from 1-7: specifies the relaxation mode:
1 ... MD with scaling of velocities according to temperature
2 ... MD without scaling of velocities according to temperature
3 ... Steepest descent (velocities are set to zero after each step)
4 ... Conjugate gradient relaxation
5 ... Constraint Conjugate gradient relaxation
6 ... Mulliken analysis and atomic energy calculation
7 ... Option number 6 + prints out the eigenvectors
force REAL: specifies the tolerance for the force ( in a.u.): If the total force acting on each atom in the structure is smaller than fmax, the conjugate gradient or steepest descent routine is converged and the program terminates.
scc_on BOOLEAN: switches the scc mode on/off.
scctol REAL: specifies the tolerance for the scc cycle.
read_charges BOOLEAN: read existing charges in file CHR.DAT if in scc mode.

structure STRING: name of input structure. Must be in gen format.
electrons INTEGER: number of electrons in structure.
moveatoms INTEGER: number of movable atoms in structure.
out STRING: name of output structure.

Maximal angular momentum (lmax + 1) of the outer shell of the atoms found in the input structure.

Names of the Slater-Koster files for the interaction of the atoms contained in the input structure in the order 1-1, 1-2, ..., 1-n, 2-1, ..., 2-n, ..., n-n. For symmetry reasons the interaction files including hydrogen x-h and h-x are identical.

stepwidth REAL: specifies the stepsize ( in a.u. = 2.4189e-17 s) for the relaxation (not relevant for conjugate gradient (rel_mode 4), where it is adjusted automatically).
atemp REAL: specifies the atomic temperature.
etemp REAL: specifies the electronic temperature (Fermi-distribution).
vwscale REAL: is the probability to rescale the velocities after each step according to the temperature
nofsteps INTEGER: number of steps in the relaxation.
Multiple input lines are only allowed in rel_mode 1, they do not make sense in any other case.

Output files:
ATM.DAT: atomic energies
CHR.DAT: Mulliken charges
SPE.DAT: spectrum of Kohn-Sham single particle eigenvalues
ENERGY.TMP: energies and forces during relaxation
FRC.DAT: forces

Description of the general (gen) format for clusters and supercells

The gen format can be used to describe clusters and supercells. It is more or less the xyz format used by the visualizer xmol extended to periodic structures. The format can be converted to xyz via gen2xyz. The reverse can be done by xyz2gen.

structure file for tiophene SC4H4

   9 C
S C H  Tiophene
1   3    -2.305709     0.000000     0.840199
2   2     0.711561     0.000000    -0.696378
3   2    -0.711561     0.000000    -0.696378
4   2    -1.243090     0.000000     0.585641
5   1     0.000000     0.000000     1.755602
6   2     1.243090     0.000000     0.585641
7   3     1.321862     0.000000    -1.607264
8   3    -1.321862     0.000000    -1.607264
9   3     2.305709     0.000000     0.840199

This example of a gen file describes a tiophene molecule.
The first line contains the number of atoms followed by C for Cluster or S for Supercell.
The second line contains the chemical symbols of the elements which may be followed by a Comment.
The following lines describe the atoms: The first number is the atom number (only a help, not read in by the programs). The second number is the atom type. Then follow the coordinates x, y, z in Å. A binding matrix may follow but is often redundant, since xmol called in xyz format creates it itself.

structure file for diamond

   8 S
C Diamond
1   1   0.891748   0.891748   0.891748   4   5  6  7  8
2   1   0.891748   2.675249   2.675249   4   5  6  7  8
3   1   2.675249   0.891748   2.675249   4   5  6  7  8
4   1   2.675249   2.675249   0.891748   4   5  6  7  8
5   1   1.783501   1.783501   3.566997   4   1  2  3  4
6   1   1.783501   3.566997   1.783501   4   1  2  3  4 
7   1   3.566997   1.783501   1.783501   4   1  2  3  4
8   1   3.566997   3.566997   3.566997   4   1  2  3  4
0.000000   0.000000   0.000000
3.567002   0.000000   0.000000
0.000000   3.567002   0.000000
0.000000   0.000000   3.567002

This example of a gen file describes a diamond structure.
The line behind the description of the atom positions contains the origin of the supercell in Å.
The last three lines are the supercell vectors in Å.