[CPMD-list] Geometry Opt

[Ari P Seitsonen]

Dear Xudong,

  Even without looking more deeply in your input one can see that there's
no way an oxygen dimer would fit in such a small lateral surface cell
(1x1). Thus it's no surprise that the eigenvalues are very unphysical, and
that the relaxation could give anything. Did you check the surface
geometry after the program had started?

  Other issues:

 - I'm not sure that the keyword MULTIPLICITY works with fractional
  occupation numbers
 - The cutoff energy is definitely too low for the pseudo potentials
  given; it _might_ speed up the initial relaxation, however later it's
  insufficient (especially for oxygen)
 - You probably don't have enough vacuum to separate the two surfaces
 - I nowadays LBFGS over BFGS, it seems to be faster in my case
 - Now you have only three layers of platinium, however if you have more
  layers you might want to fix some of them (CONSTRAINTS ... END
  CONSTRAINTS)
 - You try to use d as the local component for the Pt pseudo, that's not
  going to work; please try 'LOC=S' or 'LOC=P'
 - What are the core radii in the pseudo for oxygen? If they are say above
  1.3 Bohr you ought use 'LMAX=D', otherwise if the core radii are smaller
  than 1.1 Bohr the 'LMAX=P' is sufficient

  Hopefully this helps.

    Greetings,

       apsi

-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-
 Ari Paavo Seitsonen / Ari.P.Seitsonen at iki.fi / http://www.iki.fi/~apsi/
 Tel +41 1 635 44 97 / Fax +41 1 635 68 38 / GSM +41 79 719 09 35      
 Anschrift: Physikalisch Chemisches Institut (PCI), Universität Zürich (UniZh)
 Indirizzo: Winterthurerstraße 190, CH-8057 Zürich
 Address:   Schweiz / Svizzera / Suisse / Svizra / Switzerland

On Tue, 25 May 2004, Xudong Tang wrote:

> hi, CPMD ppl,
> 
> I'm running a simulation of oxygen obsorption on Pt surface. I do the 
> geometry optimization, but it just cannot reach convergence. I don't know if 
> there is anything wrong with my input file. Anybody can help me here?
> 
> BTW: Attached is the screen output.
> 
> Thanks.
> 
> Xudong
> 
> &CPMD
>     OPTIMIZE GEOMETRY
>     LSD
>     UNIT HESSIAN
>     BFGS
>     FREE ENERGY FUNCTIONAL
>     LANCZOS DIAGONALISATION
>     LANCZOS PARAMETERS
>       1   6 10   1.D-18
>     TROTTER FACTOR
>       0.001
>     BOGOLIUBOV CORRECTION OFF
>     GRAM-SCHMIDT ORTHOGONALISATION
>     CONVERGENCE
>       5.D-4  1.D-3
>     MAXSTEP
>       500
>     BROYDEN MIXING
>       0.15 200   0.01  0   8
>     ALEXANDER MIXING
>       1.1
>     ELECTRON TEMPERATURE
>       1000.
> 
>     COMPRESS WRITE32
>     STRUCTURE BONDS
> &END
> 
> &SYSTEM
>    MULTIPLICITY
>     3
> 
>    POINT GROUP
>     AUTO
>    SYMMETRY
>     4
>    CELL
>     2.772    1.0    4.082     1.0  1.0  1.0    (a=3.92/sqrt(2), 
> c/a=5*sqrt(2/3))
>    CUTOFF
>     35.000
>    ANGSTROM
>    STATES
>     70
>    TESR
>      3
>    KPOINTS MONKHORST-PACK fULL
>     7 7 1
> &END
> 
> &ATOMS
> *78-Pt_tm_PBE.pp KLEINMAN-BYLANDER
>    LMAX=D
>    3
>        0.000000000000      0.000000000000      0.238977135311
>        0.000000000000      1.600414946194      2.263060800000
>        1.386000000000      0.800207473097      4.287144464689
> 
> *08-O_tm_PBE.pp KLEINMAN-BYLANDER
>    LMAX=P
>    2
>        0.735555110510      1.284584334185      6.574083695600
>        1.957044889490      1.469915665815      6.752716304400
> 
> &END
> 
> &BASIS
>      PSEUDO AO 2
>       0  1
> &END
> &DFT
>    NEWCODE
>    FUNCTIONAL PBE
>    GC-CUTOFF
>      5.0e-6
> &END
> 
> _________________________________________________________________
> Stop worrying about overloading your inbox - get MSN Hotmail Extra Storage! 
> http://join.msn.click-url.com/go/onm00200362ave/direct/01/
>