The ABINIT Users Mailing List ( CLOSED )

[Erik Ylvisaker <eylvisaker@physics.ucdavis.edu>]

It looks like you're missing nband?  You should have one band for every
two electrons you're treating, plus a one or two extra (plus nbdbuf) so
the code can resolve states that are close in energy.  I usually feel
best about my calculation if I look at occ in the output file and see
that at least the highest few bands are unoccupied.

~Erik Ylvisaker

On Tue, 2005-07-12 at 07:32, silvia@cii.fc.ul.pt wrote:
> Dear ABINIT users,
> 
> I'm trying to optimize one phenol molecule using the Perdew-Burke-Ernzerhof 
> GGA functional (ixc=11) and Fritz-Haber-Institute (FHI) pseudopotentials 
> within the GGA aproximation.
> I'm considering a unit cell with dimensions 20.0 20.0 20.0. I'm using the 
> data set mode to vary the value of ecut while maintaining the cell 
> dimensions fixed. The problem is that I'm not getting convergence on the 
> force gradients during the Broyden minimization (ionmov=3). I've used 35.0 
> and 40.0 hartree as ecut values and 100 optimization steps (ntime=100). I'm 
> not achieving convergence on the SCF cycles either.
> 
> This is the input file I'm using:
> 
> ndtset 4
> acell 20.0 20.0 20.0
> ecut: 35.0 ecut+ 5
> 
> ntypat 3            
> znucl 8 6 1         
> 
> natom 13
> typat 1 1 1 1 1 1 2 2 2 2 2 3 2   
> 
> xangst
> 
>             0.939974   -0.024981    0.000001
>             0.221909   -1.222557    0.000001
>            -1.169193   -1.188258    0.000000
>            -1.854307    0.027925   -0.000001
>            -1.128953    1.217603    0.000000
>             0.265046    1.197679    0.000000
>             0.767230   -2.166033    0.000004
>            -1.725398   -2.126674    0.000001
>            -2.943885    0.047328    0.000000
>            -1.649437    2.176230   -0.000001
>             0.828963    2.133524   -0.000001
>             2.300265   -0.110628   -0.000003
>             2.673550    0.776176    0.000014
> 
> ixc  11
> 
> nkpt 1
> ntime 100
> ionmov 3
> tolmxf 5.0d-5
> 
> iscf 3            # Algorithm for SCF
> nstep 50          # Maximal number of SCF cycles
> toldff 5.0d-6
> diemac 1.0        
> diemix 0.5        
> 
>                  
> Can someone give me some clue on what I'm doing wrong or not doing at all? 
> Any help would be greatly appreciated...
> 
> Best regards,
> Slvia Estcio
> 
> 
>