[molpro-user] Linear Molecules and MP2 Frequnecies

Peter Knowles KnowlesPJ at cardiff.ac.uk
Fri Mar 14 17:14:46 GMT 2014


Dear Alex,

The trouble comes because for the RMP2 method, analytic second derivatives are not implemented, and therefore finite coordinate displacements are used for numerical differentiation. Some of these lower the symmetry of the molecule, and this can cause chaos if the job includes particular directives that assume, or even try to enforce, the high symmetry. A simple work-around is to use no symmetry at all, ie as in the following.
Best regards,
Peter 

symmetry,nosym;
orient,mass
ANG;
geomtyp=xyz
geometry={
3

  H          0.0000000000        0.0000000000       -2.8437425492
  Be         0.0000000000        0.0000000000       -1.5482621834
  Ar         0.0000000000        0.0000000000        0.4210358303
}
 gprint,orbitals;
 basis;
sp,h,aug-cc-pVDZ;c
spd,be,aug-cc-pwCVDZ;c
spd,ar,aug-cc-pwCVDZ;c
 end;
 {RHF;
wf,22,1,0;}
 pop;
 {RMP2
core,6}
 pop;
frequencies

On 13 Mar 2014, at 12:24, Simperler, Alexandra <a.simperler at imperial.ac.uk> wrote:

> Dear fellow-users and Experts
>  
> I am working on a user-question I cannot resolve.
>  
> I tried converting into z-matrix, various Orient options etc.
> The molecule is H-Br-Ar and is linear and being C*V MOLPRO picks up
> correctly the largest Abelian subgroup, c2v all goes well for HF and MP2 single points but once it comes to Frequencies we get a complaint that x cannot be imposed, and noorient is on which I thought was switched of by Orient to start with.
> We are using Molpro2012.1 btw
>  
> So my user tried:
> ***, some title
> memory,400,m
> symmetry,x,y;
> orient,mass
> ANG;
> geomtyp=xyz
> geometry={
> 3
>   H          0.0000000000        0.0000000000       -2.8437425492
>   Be         0.0000000000        0.0000000000       -1.5482621834
>   Ar         0.0000000000        0.0000000000        0.4210358303
> }
>  gprint,orbitals;
>  basis;
> sp,h,aug-cc-pVDZ;c
> spd,be,aug-cc-pwCVDZ;c
> spd,ar,aug-cc-pwCVDZ;c
>  end;
>  {RHF;
> wf,22,1,0;}
>  pop;
>  {RMP2
> core,3,1,1,0}
>  pop;
> frequencies
>  I personally tried a Z-Matrix approach:
>  ***, H+BeAr aug-cc-pCVQZ CCSD(T) single point calculation at
>  QCISD/aVQZ geometry !
> memory,400,m
> symmetry,x,y;
> orient,mass,planexz
> geometry={
> H
> Be , H , rhbe
> Ar , Be , rbear , H , 180
> }
>  rhbe=1.29548 Ang
> rbear=1.96930 Ang
>  gprint,orbitals;
>  basis;
> sp,h,aug-cc-pVDZ;c
> spd,be,aug-cc-pwCVDZ;c
> spd,ar,aug-cc-pwCVDZ;c
>  end;
>  {RHF;
> wf,22,1,0;}
>  pop;
>  {RMP2
> core,3,1,1,0}
>  pop;
> frequencies
>  
>  and I ended up with:
>  Computing numerical hessian using default procedure for command RMP2
>  Symmetry elements for unique displacements: X  Y
>   Using no symmetry in wavefunction calculations
>   Running default procedure: RHF-SCF000  RMP2
>  
>   Numerically approximating hessian using central energy differences
>   Task list generated. Total number of displacements:     24
>   Attempt to include non-existent symmetry X  Because of noorient
>  option, program will  not align to your requested symmetry elements 
>  Try again with different symmetry request  This error can also happen
> in geometry optimizations that specify
>    symmetry-breaking internal coordinates in terms of active variables 
>  User-specified symmetry elements: X,Y  Symmetry elements: Y  ? Error 
>  ? Symmetry error  ? The problem occurs in zmatrix.f:zmat_evaluate
>  
>  I am very puzzled because I don't know where the noorient comes from as we specify Orient in our inputs.
> Is it something trivial we missed? Then I apologize kindly in advance.
>  
> Kind regards
> Alex
>  
>  
> Dr Alexandra Simperler FHEA
> Software Sustainability Institute Fellow 2014
>  
> Department of Chemistry
> Imperial College London
> South Kensington
> London SW7 2AZ
> UK
>  
> Tel: +44 (0)20 759 41220
> Email: a.simperler at imperial.ac.uk
>  
> NEW: Follow us on twitter:
> https://twitter.com/NSCCS_team
>  
> EPSRC UK National Service for Computational Chemistry Software (NSCCS)
> http://www.nsccs.ac.uk
>  
> Software Sustainability Institute
> http://www.software.ac.uk/about
>  
>  
>  
> _______________________________________________
> Molpro-user mailing list
> Molpro-user at molpro.net
> http://www.molpro.net/mailman/listinfo/molpro-user

--
Prof. Peter J. Knowles             
School of Chemistry, Cardiff University, Main Building, Park Place, Cardiff CF10 3AT, UK
Telephone +44 29 208 79182 Email KnowlesPJ at Cardiff.ac.uk 
WWW http://www.cardiff.ac.uk/chemy/contactsandpeople/academicstaff/knowles.html









More information about the Molpro-user mailing list