forum@abinit.org
Subject: The ABINIT Users Mailing List ( CLOSED )
List archive
- From: Damien Caliste <damien.caliste@cea.fr>
- To: forum@abinit.org
- Subject: Re: [abinit-forum] Problems with TiCl Reaction
- Date: Tue, 25 Mar 2008 14:16:07 +0100
- Organization: CEA - recherche fondamentale
Hello,
Le 20/03/2008, nathaniel.deskins@pnl.gov a écrit :
> 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.
I've tested your input files within a wavelet basis set (that have the
same property as plane waves to be as precise as possible) and I obtain
results similar to the one you have with gaussians basis sets.
Here are the results:
hgrid crmult bond-length binding-energy
0.5 4 2.130 -7.7589
0.45 4.5 2.208 -7.3101
0.4 5 2.218 -4.5927
0.35 5.5 2.222 -4.5429
0.3 6 2.222 -4.5218
TiCl total energy
-18.479037693
-18.488242138
-18.494410365
-18.496620706
-18.497434335
Ti total energy
-3.3109492793
-3.3210559183
-3.4213936690
-3.4232515518
-3.4240442088
Cl total energy
-14.882951551
-14.898541633
-14.904236391
-14.906420534
-14.907213698
Where hgrid is in bohr and gives the precision of the basis set and
crmult gives the spatial expansion (as a multiplying factor), see
Abinit 5.5 documentation for wvl_crmult.
I've also attached the input files I used. The results for low
precision hgrid (> 0.4 bohr) can be discussed especially because the
DIIS goes to an excited states and because the spatial expansion is
too small, but this requires further investigations, I only give quick
and direct results. For better hgrid values (the lower, the better),
results can be trusted and are good.
Damien.
PS: the ti file is given with an equivalent occupation for the
degenerated states but above results are for asymmetric occupation, to
have them, change occ to 1111 instead of 111 1/3 1/3 1/3.
PPS: the 5 datasets of complete geometry relaxation have taken 40
minutes on 4 CPU with less than 400Mo per proc for the finer mesh
(equivalent to Ec=220Ht in the finer zones).
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 4 4
occ 1 1 1 1
1 1 1 0
fixmom 1
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
usewvl 1
icoulomb 1
optforces 0
optstress 0
iscf 2
nsym 1
nwfshist 5
ndtset 5
wvl_hgrid: 0.5
wvl_hgrid+ -0.05
wvl_crmult: 4
wvl_crmult+ 0.5
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 6 6
occ 1 1 1 1/3 1/3 1/3 0 0 0 0 0 0
fixmom 4
spinat 0.0 0.0 4.0 # Initialisation of spin
nstep 50 # Maximal number of SCF cycles
tolwfr 1.0d-4 # Will stop when, twice in a row, the difference
diemac 2.0 # Although this is not mandatory, it is worth to
usewvl 1
icoulomb 1
optforces 0
optstress 0
iscf 2
nsym 1
nwfshist 5
ndtset 5
wvl_hgrid: 0.5
wvl_hgrid+ -0.05
wvl_crmult: 4
wvl_crmult+ 0.5
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 6 6
occ 1 1 1 1 1 1 1 1 1 1 1 0
fixmom 2
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
usewvl 1
icoulomb 1
optstress 0
iscf 2
nsym 1
nwfshist 5
ndtset 5
wvl_hgrid: 0.5
wvl_hgrid+ -0.05
wvl_crmult: 4
wvl_crmult+ 0.5
- Problems with TiCl Reaction, nathaniel . deskins, 03/20/2008
- Re: [abinit-forum] Problems with TiCl Reaction, Anglade Pierre-Matthieu, 03/21/2008
- RE: [abinit-forum] Problems with TiCl Reaction, Deskins, Nathaniel A, 03/21/2008
- Re: [abinit-forum] Problems with TiCl Reaction, Anglade Pierre-Matthieu, 03/22/2008
- Re: [abinit-forum] Problems with TiCl Reaction, Dr MJ Verstraete, 03/23/2008
- Re: [abinit-forum] Problems with TiCl Reaction, Anglade Pierre-Matthieu, 03/22/2008
- RE: [abinit-forum] Problems with TiCl Reaction, Deskins, Nathaniel A, 03/21/2008
- Re: [abinit-forum] Problems with TiCl Reaction, Damien Caliste, 03/25/2008
- Re: [abinit-forum] Problems with TiCl Reaction, Anglade Pierre-Matthieu, 03/21/2008
Archive powered by MHonArc 2.6.16.