[molpro-user] Saving orbitals to file and natural orbitals

[Seth Olsen]

Hi Steven,

Any method that produces a one-body reduced density matrix can be used 
to generate 'natural orbitals'.  That is really what natural orbital 
means - they diagonalize a 1-RDM.  Both MCSCF and MRCI could output 
these, and a variety of other methods also can.  If one wants to refer 
to a given set as the 'true' NO's, then the only clear choice is the set 
which diagonalizes the true 1-RDM of the system, contracted from the 
full N-body RDM.  It is the latter quantity that is hard to obtain.

As far as your orbital problems are concerned, I would look to the 
MATROP facility.  It allows one to print ASCII files with orbital 
coefficients, density matrix elements, etc.  I use it in ALL of my 
molpro jobs to save orbitals in non-machine-dependent form.

Cheers,

Seth

Stephen Harrison wrote:
> Dear Molpro User Group,
>
> I wonder if I could have some advice regarding some calculation inputs
> please? I'm still fairly new to the program and can't find much clear
> guidance regarding these issues in the documentation.
>
> Firstly, I understand that MOLPRO uses natural orbitals by default at
> MCSCF level, after discussing this with my supervisor we seem slightly
> perplexed by this. Are the MCSCF natural orbitals the same as normal MCSCF
> orbitals? because as far as my understanding goes I thought the 'true'
> natural orbitals only came out during the MRCI level calculation (i.e.
> after MCSCF has already run).
>
> Secondly, the documentation seems to make it clear that you can save
> orbitals to file using the orbital/save command. However I tried this and
> it seemed that MOLPRO didn't create these files I requested? (INPUT is
> below).
> I would like to take them and use them in an R-Matrix calculation so any
> format information you could give me would also be really helpful. I know
> it's possible to brute force it by copy and pasting orbitals from the
> output, but if I could get them saved in-situ, particularly in the correct
> format, it would be a great help.
>
> Here was the INPUT I tried to use, but after I couldn't find the files
> containing the orbitals.
>
> BASIS
>  default, cc-pVTZ
>  END
>
>  {RHF, ENERGY=1.e-8 ;
>        OCC, 9, 2, 2, 0;
>        ORBPRINT,5}
>
>  {MCSCF
>          THRESH, ENERGY=1.e-8;
>          OCC, 10,3,3,0 ; CLOSED, 7,0,0,0 ; FROZEN, 0,0,0,0 ;
>
>  WF, 25, 1, 1;
>  STATE, 1;
>  WEIGHT, 0.5;
>  ORBITAL, 2000.1
>
>  WF, 25, 2, 1;
>  STATE, 1;
>  WEIGHT, 0.25;
>  ORBITAL, 2000.2
>
>  WF, 25, 3, 1;
>  STATE, 1;
>  ORBITAL, 2000.3
>  WEIGHT, 0.25;
>
>          ORBPRINT,10}
>
>  {MRCI
>         THRESH, ENERGY=1.e-8;
>          OCC, 10,3,3,0 ; CLOSED, 7,0,0,0 ; CORE, 0,0,0,0 ;
>          STATE, 1;
>  }
>
>
> Thank you very much for any help, it would greatly be appreciated!
> Stephen Harrison, UCL
>
>
>
>