[Pw_forum] force constants file
giannozz at nest.sns.it
Wed Apr 13 20:14:34 CEST 2005
On Monday 11 April 2005 11:23, Bartek Wiendlocha wrote:
> Could Anyone help me and write a (detailed..) description of the
> force constants file
A description, maybe. A detailed one (or even an accurate one),
> 2 3 4 5.6370000 0.0000000 1.0870000 0.0000000 0.0000000
> 0.0000000 --> lattice
number of types, number of atoms in the cell, type of lattice,
> 1 'Al ' -> atoms 24588.6885119930 -> ??
type #, label, atomic mass in a.u.
> 2 'B ' -> atoms 9851.88001536857 -> ??
> 1 1 0.0000000 0.0000000 0.0000000 -> positions
> 2 2 0.5000000 0.2886751 0.5435000
> 3 2 0.0000000 0.5773503 0.5435000
atom #, type #, positions in the cell in a0 units
T if the file contains epsilon and Z*, F otherwise
> 8 8 8 -> grid size
n1 n2 n3, defining the grid of q-points in the Brillouin Zone and
the side of the R-space supercell used to calculate the Fourier
transform. The latter is n1*tau1, n2*tau2, n3*tau3, where
tau1, tau2, tau3 are the primitive translations of the lattice
> 1 1 1 1 --> ??
indices: i,j,na,nb (polarization1 , polarization 2, atom 1 , atom 2)
> 1 1 1 --> ?? 1.56778730300E-01 --> force constant, but for which
m1 m2 m3 force_constant(m1,m2,m3,i,j,na,nb)
where m1=1,...,n1, m2=1,...,n2, m3=1,...,n3, define a lattice vector R:
R = (m1-1)*tau1 + (m2-1)*tau2 + (m3-1)*tau3
(if m1-1 > n1/2 and so on, refold m1-1 => m1-1-n1 and so on)
So: force on atom na in direction i, when moving atom nb in direction j,
in a cell that is R far away
> are the force constants divided by mass ?
> How to find the distance in real space (or the number of nearest
> neighbours) up to which the FC are calculated?
see above: approx. n1/2, n2/2, n3/2 unit cells in the three directions
Paolo Giannozzi e-mail: giannozz at nest.sns.it
Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513
Piazza dei Cavalieri 7 I-56126 Pisa, Italy
More information about the Pw_forum