|
    
Fragment mode
In Fragment mode more input is required than in Create mode:
you have to specify at least: (1) the atomic positions and (2) how
the total system is built up from fragments. We recommended to specify also
(3) the point group symmetry.
Example of an input file for the C2H4 molecule:
ATOMS
C 0 0 .6685
C 0 0 -.6685
H .927 0 -1.203
H -.927 0 -1.203
H .927 0 1.203
H -.927 0 1.203
end
fragments
C TAPE21c.dzp
H TAPE21h.dzp
end
symmetry D(2h)
end input
Three keys are used: atoms, fragments
and symmetry.
The first two are block keys.
atoms
defines the atomic positions: each record in the data block
contains the chemical symbol of an atom followed by its Cartesian coordinates
in Angstroms.
Z-matrix type input of atomic positions is also possible. This will be
explained in a later section.
fragments
lists the fragment files each record contains a
fragment type followed by the
corresponding fragment file. In
the example the files are local
files. Files in other directories are addressed by giving the complete
file path.
Note: if a parallel calculation
is performed, be sure that each 'kid' finds the specified fragment files. This
will usually require that the files are not local to the job, but first be moved to some shared
volume, and that the references to the fragment files in the input contain the
full path. An alternative is to ensure that the (local) files in the parent
directory are copied first to the 'kid' directories before the parallel
calculation starts.
symmetry
specifies the point group symmetry by a Schönfliess type symbol.
Appendix 3 contains a complete list of all Schönfliess symbols that are
recognized by adf. If no symmetry
is specified adf will take the
true symmetry of the nuclear frame as the user-specified symmetry. If (electric) fields are used, see later,
the computed symmetry will take this into account. Note that the computed
symmetry may not occur in the list of allowed symmetries (see Appendix 3), in
which case you have to explicitly specify the (lower) point group symmetry you
wish to apply.
The atomic coordinates must conform to the point group
symmetry; the program will check this and abort if the atomic system does
not have the specified symmetry. It is allowed, however, to specify a
lower symmetry than what is actually
present in the set of atomic positions. The specified symmetry determines how results are analyzed
and how irreducible representations and subspecies are labeled. It also
determines various algorithmic aspects: the program runs more efficiently
with the highest possible symmetry.
The spatial orientation of the molecular coordinate system is not arbitrary. ADF
requires for each pointgroup symmetry a specific standard orientation. In axial
groups for instance, the main rotation axis must be the z-axis. This
implies a restriction on how you can define the atomic coordinates under atoms.
The orientation requirements for all point groups are listed in Appendix 3. If the specified symmetry equals
the true symmetry of the nuclear frame adf
will adjust the input orientation of the molecule to the requirements (if
necessary). If you have specified a subgroup of the true nuclear symmetry, no
such orientation adjustment is carried out and the user has to make sure
that his input data yield the correct orientation, lest an error will occur.
Restrictions apply to the symmetry (as specified) of the
molecule, related to the symmetries of the fragments as they were
stipulated in the preceding fragment calculations. All symmetry operators
of the molecule that internally rotate or reflect a fragment but leave it at
the same position in the molecule, must also be operators of the symmetry group
in which the fragment has been computed. Furthermore, two fragments must
not be symmetry-equivalent in the molecule only by an improper rotation. The
implied internal reflection of the fragment must be one of the symmetry
operators in the point group symmetry that is used in the fragment
calculation and the molecular symmetry
group must also contain a proper rotation that maps the two fragments onto each
other.
The example of Fig.1 implicitly assumes that all fragments
are single atom fragments. When the
fragments are larger the data records in the atoms key have to be extended: you
must specify which atoms belong together in one fragment.
SYMMETRY T(D)
Atoms
Ni 0 0 0
C -1.06 -1.06 1.06 f=CO/1
C -1.06 1.06 -1.06 f=CO/2
C -1.06 1.06 -1.06 f=CO/3
C 1.06 -1.06 -1.06 f=CO/4
O 1.71 1.71 1.71 f=CO/1
O -1.71 -1.71 1.71 f=CO/2
O -1.71 1.71 -1.71 f=CO/3
O 1.71 -1.71 -1.71 f=CO/4
End
Fragments
CO TAPE21co.yesterday
Ni t 21ni.dzp
End
End Input
Another sample input file; using a single atom ni fragment and four molecular CO fragments. The keys symmetry and fragments operate as before. Again we have two types
of fragments (here: Ni and CO);
for each of them, the fragment file is specified.
Under the key atoms the chemical symbols and the nuclear coordinates are
listed. Added is the f=...-part;
f stands here for fragment and tells
the program that the carbon and oxygen atoms belong to CO fragments. The last part /n
enumerates the individual CO
fragments: here you define which C
and O belong together in one CO fragment.
The record for Ni contains no f=
part, implying the default for this
atom: it is a fragment on its own. In the C2H4 example before the default applied to all
atoms.
There are more possibilities with the keys atoms and fragments. This is worked out later. The purpose of
this section was to provide a quick and easy start.
    
|