The ABINIT Users Mailing List ( CLOSED )

[Xavier Gonze <gonze@pcpm.ucl.ac.be>]

Dear Martin,

M.G.A.Tijssens wrote:
> Dear abiniters,
> I have been calculating elastic constants for NiTi. One deformation
> mode of the unit cell that I use is a simple shear. To accomplish this
> I include in the input file the lines:
>
>
>
> rprim: 1.0    0.0    0.0
>        0.0    1.0    0.0
>        0.0    0.0    1.0
>
> rprim+ 0.0    0.02   0.0
>        0.0    0.0    0.0
>        0.0    0.0    0.0
>
> As I understood, these give the directions of the vectors spanning
> the unit cell and they are multiplied by the dimensions given in
> acell (which is three times 5.718 in my case, corresponding to the
> minimum energy lattice parameter found for Troullier-Martins
> potentials).
>
> The results that I anticipated is that the energy will increase, and
> reach a maximum when (the first line of) rprim equals 1.0 0.5 0.0
> since at that moment the unit cell has been sheared by 45 degrees.
>
> However, the maximum is reached when the first line reads 1.0 1.0 0.0.
> Can anybody tell me if this is a mistake in abinit (doubtfull...)
> or my misinterpretation of rprim. If so, how then should I interpret
> rprim (and I did read the help file for this variable!).

To complement the answer given by Razvan, a long time ago ...
(if you are still interested ;-)

Let's say that your Ni atom is at the origin, and your Ti atom is
at 0.5 0.5 0.5 (or the reverse, no importance).
When the rprim reach 1.0 1.0 0.0, note that the
Ti atom located in 0.5 0.5 0.5 is not at all at the
BCC site of the simple cubic sublattice of Ni atoms , but at the center
of squares with four Ni atoms.
(Suggestion : look at the atomic coordinates xcart !)
In order to obtain the periodicity, you need to go to 1.0 2.0 0.0.
(Still, the periodicity will not be perfect, because the FFT
grid is not the same, but this is a 0.00001 Hartree effect, typically).

Xavier