The ABINIT Users Mailing List ( CLOSED )

[Nicola Marzari <marzari@mit.edu>]

Dear Max,


it seems all very good and reasonable, but if phonons is what you are
after, I'd go for a more stringent test on kpoints/smearing (as a side
note - if the quantities you are interested in were to depend only
on the Fermi surface, you'd need much tigher convergence tests).

Since phonons are what you are looking for, just displace one or
more atoms (by 5% on their interatomic distances, say), and
repeat the test, until forces are converged to your desired
tolerance (I'd say 2-3 cm-1 in the phonon frequencies).

                                nicola



Latévi Max LAWSON DAKU wrote:
> Dear Abinit users,
>
> I would like to study a family of related metallic systems.
> For the one with the simplest molecular formula (6 atoms
> in the formula unit), I tried to do a convergence study in
> order to determine the optimal value of tsmear (cold
> smearing; occopt=5).
>
> 1) Without smearing, I did full geometry optimisation with
>    nkpt = 10, 28, 60 and ecut values of 20, 30, 40, 50 Ha
>    (and an additional test optimisation with nkpt=60 and
>    ecut=80). For this FCC system, this gives "converged"
>    values of acell=7.260 angstrom and Etot=-167.260 Ha.
>    The "(tsmear=0)-convergence" in acell (> 0.001 A) is
>    obtained for ecut=40  Ha and nkpt=28.
>
>
> 2) With ecut set to 40 Ha and smearing turned on, I made
>    vary tsmear at fixed values of nkpt (10 pts -> circles;
>    28 pts -> squares; 60 pts -> diamonds). The attached
>    figures gives the dependences of acell and Etot on
>    tsmear. For tsmear <= 0.01 Ha, acell=2.589 A and
>    Etot=-167.266 Ha; independantly of the integration
>    grid used. The curves are actually similar in the studied
>    range.
>
> 3) Given these results, I would go with a tsmear value of
>    4 mHa and nkpt=10 for the determination of the bulk
>    modulus and its derivative, and  also of the phonons.


--
---------------------------------------------------------------------
Prof Nicola Marzari   Department of Materials Science and Engineering
13-5066   MIT   77 Massachusetts Avenue   Cambridge MA 02139-4307 USA
tel 617.4522758 fax 2586534 marzari@mit.edu http://quasiamore.mit.edu