Hi Bedamani,
The contents of the *.FC file is in cartesian coordinates,
eV/Ang^2 units, though strictly it isn't directly the force
constants matrix, but a step just before.
Let's say FC.fdf file built by fcbuild contains N atoms, then
each of them gets a block of 6N rows by 3 columns in the *.FC
file, representing the forces on every atom, whenever every
atom of your basic, reference, unit cell moves a small distance
(delta) along +x,-x,+y,-y,+z,-z. The numbers written are not
the forces themselves but forces/delta.
Thus, if your unit cell contains M atoms, the *.FC file will
comprise M*6N lines. Finding the force constants matrix amounts
to locate those M atoms within the *.FC file, identifying to
which neighbour cell any interacting atoms belongs in, and averaging
rows for the + and - displacements. Of course, vibra knows that
geometric information and does that for you.
The system you specified is a fully periodic 2-atoms one, so
that M=N, and you have 2 * 6*2 * 3 = 72 total "elements" in the
*.FC file. This means you'll only get the gamma point,
*** is this what you really want *** ?.
If not, you must specify non-default values for (fcbuild) variables
SuperCell_[1,2,3], such as 1,1,1, namely, large enough to account for
all interactions with the reference, unit cell.
HTH,
Roberto
On Thu, 11 Oct 2012, bedamani singh wrote:
Dear siesta users,
We are trying to use the force-constant output from SIESTA for a
2-atom
(2-dimensions) crystal (graphene) as input into our program
which requires a force-constant matrix in Cartesian
coordinates. The required full matrix would be 6x6 ([basicaly 3n*3n, for
a n atom system],
with elements at the head Atom1x, Atom1y, Atom1z,Atom2x, Atom2y, Atom2z.
For SIESTA output from graphene we are getting 72
total elements which do not seem to be arranged in this order.
Can you tell us what the elements in the .FC file are, and how we could
transform to the needed Cartesian matrix?"
The input files I haved used in siesta calculation is attached as well as
the
output force-constant *.FC file.
We have used the executables fcbuild, siesta and vibrator included in
the Util directory of
siesta-3.1 software package.First, we run
$ fcbuild < graphene.fdf
this builds a file FC.fdf which is a partial input for Siesta. Then, We
run
$ siesta < graphene-siesta.fdf
Siesta provides the file graphene.FC with the force constant matrix.
Please help me.
best regards
bedamani
