The ABINIT Users Mailing List ( CLOSED )

["Deskins, Nathaniel A" <nathaniel.deskins@pnl.gov>]

I'm assuming the main criteria for convergence is the cutoff energy.
This is a large box, so a single k-point should suffice. Here's the
results of a cutoff energy test.

Cutoff(hartree)   Erxn(eV)   D-TiCl(angstroms)
10                5.11       2.39
15                5.12       2.39
20                3.80       2.40
25                3.79       2.40
30                3.79       2.39
35                4.64       2.35

These show some pretty big oscillations. I couldn't go to 40 eV because
of memory restraints (2 GB required for 40 hartree). 

I'm guessing I'm doing something wrong, which is why I posted my input
files so perhaps someone might notice an obvious error.

Thank you,
Aaron

-----Original Message-----
From: Anglade Pierre-Matthieu [mailto:anglade@gmail.com] 
Sent: Friday, March 21, 2008 3:50 AM
To: forum@abinit.org
Subject: Re: [abinit-forum] Problems with TiCl Reaction

Hi,

Did you try to make a converged calculation ? What is the "absolute"
precision of your simulation ?
For instance usualy light d metals require large cutoff energy.

regards

PMA

On Thu, Mar 20, 2008 at 8:14 PM,  <nathaniel.deskins@pnl.gov> wrote:
> Hello all,
>
>   I'm looking at the following gas-phase reaction. TiCl --> Ti + Cl.
>  The ground state for TiCl is a quintet (3 unpaired e-), Ti is a
pentet(?) (4 unpaired e-), while Cl is a doublet (1 unpaired e-). I've
modeled this reaction with other planewave and gaussian codes. The
energy for the reaction varies between 4.1-5.1 eV (depending on basis
set and potential) while the Ti-Cl distance is near 2.22-2.23 angstroms.
All this was with PBE XC functional.
>
>   When I've tried Abinit with the FHI GGA-PBE Troullier-Martins
pseudopotentials I get strange results. My calculated energy is 6.58 eV
while the Ti-Cl distance is 2.40 angstroms. This was for a cutoff of 15
hartree. For cutoff of 30 hartree the energy is 3.8 eV(?!) and the
distance is 2.39 angstroms.
>
>   Either there is an issue with the pseudopotentials, or I'm doing
something wrong.
>
>  Below are my input files. Any help is appreciated.
>
>  Thank you,
>  Aaron
>
>
>  ti.in
>  acell 25 25 25    # The keyword "acell" refers to the
>  ntypat 1          # There is only one type of atom
>  znucl 22           # The keyword "znucl" refers to the atomic number
of the
>  natom 1           # There are two atoms
>  typat 1          # They both are of type 1, that is, Hydrogen
>  xcart             # This keyword indicates that the location of the
atoms
>    0.0 0.0 0.0    # Triplet giving the cartesian coordinates of atom
1, in Bohr
>  ecut 15.0         # Maximal plane-wave kinetic energy cut-off, in
Hartree
>  nkpt 1            # Only one k point is needed for isolated system,
>  nsppol 2
>  occopt 2
>  nband 8 8
>  occ 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0
>  spinat 0.0 0.0 4.0  # Initialisation of spin
>  nstep 50          # Maximal number of SCF cycles
>  toldfe 1.0d-6     # Will stop when, twice in a row, the difference
>  diemac 2.0        # Although this is not mandatory, it is worth to
>
>
>  cl.in
>  acell 25 25 25    # The keyword "acell" refers to the
>  ntypat 1          # There is only one type of atom
>  znucl 17           # The keyword "znucl" refers to the atomic number
of the
>  natom 1           # There are two atoms
>  typat 1          # They both are of type 1, that is, Hydrogen
>  xcart             # This keyword indicates that the location of the
atoms
>    0.0 0.0 0.0    # Triplet giving the cartesian coordinates of atom
1, in Bohr
>  ecut 15.0         # Maximal plane-wave kinetic energy cut-off, in
Hartree
>  nkpt 1            # Only one k point is needed for isolated system,
>  nsppol 2
>  occopt 2
>  nband 8 8
>  occ 1 1 1 0 0 0 0 0 1 1 1 1 0 0 0 0
>  spinat 0.0 0.0 1.0  # Initialisation of spin
>  nstep 50          # Maximal number of SCF cycles
>  toldfe 1.0d-6     # Will stop when, twice in a row, the difference
>  diemac 2.0        # Although this is not mandatory, it is worth to
>
>
>  ticl.in
>  acell 25 25 25    # The keyword "acell" refers to the
>  ntypat 2          # There is only one type of atom
>  znucl 22 17          # The keyword "znucl" refers to the atomic
number of the
>  natom 2           # There are two atoms
>  typat 1 2         # They both are of type 1, that is, Hydrogen
>  xcart             # This keyword indicates that the location of the
atoms
>    0.0 0.0 0.0    # Triplet giving the cartesian coordinates of atom
1, in Bohr
>    4.3 0 0
>  ecut 30.0         # Maximal plane-wave kinetic energy cut-off, in
Hartree
>  nkpt 1            # Only one k point is needed for isolated system,
>  nsppol 2
>  occopt 2
>  nband 12 12
>  occ 1 1 1 1 1 1 1 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0 0 0  spinat 0.0 0.0 
> 3.0
>        0.0 0.0 1.0
>  ionmov 2
>  tolmxf 5.0d-4
>  ntime 20
>  nstep 50          # Maximal number of SCF cycles
>  toldfe 1.0d-6     # Will stop when, twice in a row, the difference
>  diemac 2.0        # Although this is not mandatory, it is worth to
>



--
Pierre-Matthieu Anglade