[molpro-user] How to define occ, closed and core n1, n2, n3, n4, n5, n6, n7, n8?

[Grant Hill]

Hi,

Well, without wanting to sound too rude, learning a little group theory is probably a good thing at this point. The values you require to define occ, closed, etc cards will depend on your molecule and its point group. Running a few small test calculations can help illustrate this, for example:

-----------

***,transplatin
memory,40,m

basis={
default,vdz,pt=vdz-pp
}

geomtyp=xyz
geom={
  11

Pt    0.0000000000    0.0000000000    0.0000000000
Cl    0.1452381549    0.0000000000    2.3420937397
N     2.0547889056    0.0000000000   -0.0920752634
Cl   -0.1452381549    0.0000000000   -2.3420937397
N    -2.0547889056    0.0000000000    0.0920752634
H     2.4403034093    0.0000000000    0.858871999
H     2.4053137827   -0.8249612003   -0.5873520537
H     2.4053137827    0.8249612003   -0.5873520537
H    -2.4053137827    0.8249612003    0.5873520537
H    -2.4403034093    0.0000000000   -0.858871999
H    -2.4053137827   -0.8249612003    0.5873520537
}

int

-----------

This will simply evaluate the integrals and then stop. In the output you will see that C2h symmetry has been assigned and a little further down in the output is the following:

 NUCLEAR CHARGE:                   72
 NUMBER OF PRIMITIVE AOS:         255
 NUMBER OF SYMMETRY AOS:          242
 NUMBER OF CONTRACTIONS:          132   (  45Ag  +  22Au  +  44Bu  +  21Bg  )
 NUMBER OF CORE ORBITALS:          16   (   6Ag  +   2Au  +   7Bu  +   1Bg  )
 NUMBER OF VALENCE ORBITALS:       28   (  12Ag  +   3Au  +   8Bu  +   5Bg  )

where the labels Ag, Au, Bu and Bg are the irreps of the C2h group. You can now see that if you used a core,6,2,7,1 card, you would have 6 Ag, 2 Au, 7 Bu and 1 Bg orbitals in the core (which is also the default in this case). In terms of reading around this, I would recommend Molecular Symmetry by Willock as a good starting point, then the books by Herzberg are very useful, especially for small molecules.

I hope this is enough to get you started, and that it provides a little insight into why multi-reference methods are not a black-box.

Grant



On 31 May 2011, at 08:27, Wenjun Li wrote:

> Dear Molpro Users,
> 
> I am a very new trial user of Molpro. I found that one key difference between Molpro and Gaussian is that, Molpro tries to solve the Schrodinger Equation as accurately as possible using all kinds of great methods to consider the electron correlation very well, especially using Multi-Reference Methods and lots of great and new methods developed from the authors of Molpro, this is great. 
> 
> However, to use Molpro is also becoming much more complicated than Gaussian, mainly because of two key reasons: (1) Molpro does not have a very great interface program for the input and output files editing and showing so far, certainly Molpro-View is great for this, but it is still not good enough at all comparing with GaussView, GaussView can almost let user to set up a calculation very quickly and easily, even though the user does not have a very good background on quantum chemistry. I heard of that, now Molpro developers are working on this kinds of interface program for Molpro, which will be definitely valuable for users. (2) Molpro requires the users to input lots of very details for solving the electronic structure calculation, including picking up different methods, options and directives, which really requires the users to have a very solid background on quantum chemistry, or else you will not really be able to use Molpro accurately and efficiently, just like me.
> 
> I am one of the users who's background is not in quantum chemistry, I know very little on molecular symmetry stuff,  and I have only a very weak quantum chemistry background, which makes me lots of troubles to use Molpro. One very key difficulty for me so far is that, I can never figure out how to define occ, closed and core n1,n2,n3,n4,n5,n6,n7,n8. For example for occ,n1,n2,n3,n4,n5,n6,n7,n8, ni is the number of occupied orbitals in the irreducible representation i. This is the only very short description for OCC in Molpro Manuals, but I can never figure out how to define occ, closed and core n1,n2,n3,n4,n5,n6,n7,n8. 
> 
> So far what I understood is like that, n1 is the # of sigma orbitals, n2 is the # of sigma* orbitals, n3 is the # of pi-x orbitals, n4 is the # of pi-x* orbitals, n5 is the # of pi-y orbitals, n6 is the # of pi-y* orbitals, then what is n7 and n8? Am I right? Most likely I am wrong, I am actually always confused about how to Define the number of occupied orbitals in each symmetry. I believe this is most likely due to my weak quantum chemistry background. So may I ask for some suggestions from the Molpro-users. How can I define occ, closed and core n1,n2,n3,n4,n5,n6,n7,n8.? Where can I find some more references or descriptions for this? Or maybe can some one suggest me to read some kinds of textbook, so that I can understand this background info.
> 
> Thanks a lot for all the help and suggestions in advance. Sorry for the long email, but maybe I did say something for very new users, who does not have a solid quantum chemistry background. Thanks a lot again.
> 
> 
> Best regards,
> 
> 
> Wenjun
> 
> -- 
> Wenjun LI 
> ---------------------------------------------------------------
> Chemical & Biomolecular Engineering,
> North Carolina State University,
> Engineering Building I, Box 7905,
> 911 Partners Way, Raleigh, NC 27695
> ---------------------------------------------------------------
> _______________________________________________
> Molpro-user mailing list
> Molpro-user at molpro.net
> http://www.molpro.net/mailman/listinfo/molpro-user