Author Topic: Problem with AOMix  (Read 7187 times)

mwoch

  • Newbie
  • *
  • Posts: 5
  • Karma: +0/-0
Problem with AOMix
« on: April 11, 2015, 07:04:49 PM »
I am experiencing problems with AOMix:

When running a calculation with fragment molecular orbitals (AOMix-FO) I receive the error message: "Number of basis functions does not match the number of canonical orbitals."
The AOMix manual states: "[...] your QC software removed near-linearly dependent functions from the orbital set [...] typically one can employ a keyword to turn off the projection of basis functions."

Where do I find this keyword in turbomole?

Thanks for your help!

uwe

  • Global Moderator
  • Sr. Member
  • *****
  • Posts: 491
  • Karma: +0/-0
Re: Problem with AOMix
« Reply #1 on: April 12, 2015, 09:42:03 PM »
Hello,

Turbomole does not remove any basis functions even in cases they become almost linearly dependent. There must be a different reason for this error message.

Regards,

Uwe

mwoch

  • Newbie
  • *
  • Posts: 5
  • Karma: +0/-0
Re: Problem with AOMix
« Reply #2 on: April 13, 2015, 07:30:34 PM »
Hello Uwe,

thank you for your quick reply.
The problem was indeed a different one. I could get it working using 6d/10f basis. But now I am facing a new problem:

I am running calculations on iron complexes. In define I use 'bp' to switch to 6d/10f. In the next menu I use 'eht' to provide MOs and define crashes:

total number of sao's is    664 which exceeds nsymbf/nbf =    663,   664
MODTRACE: no modules on stack
fatal error in dimsym.


This error does not occur if I remove the iron atom.

Do you have any idea, what is happening here?

Regards,
Markus

uwe

  • Global Moderator
  • Sr. Member
  • *****
  • Posts: 491
  • Karma: +0/-0
Re: Problem with AOMix
« Reply #3 on: April 14, 2015, 09:35:38 AM »
Hello,

the t2aomix tool does the conversion to 6d automatically, so you do not have to rerun your job with cartesian basis sets again.

Does the basis set for Fe include ECPs? Try a usual def2-SVP or def2-TZVP or def2-QZVP basis set for the iron, they do not include ECPs for Fe.

Regards,

Uwe

mwoch

  • Newbie
  • *
  • Posts: 5
  • Karma: +0/-0
Re: Problem with AOMix
« Reply #4 on: April 14, 2015, 03:04:43 PM »
Hello,

I was using def2-TZVP without any ECP but it did not work. I made some simple tests with NH3BH3. Thats how I found out that it works, using 6d/10f.

Example:
NH3BH3, bp, def2-TZVP, ri, 5d/7f
Output from t2aomix:
Quote
[...]
              +--------------------------------------------------+
              |               basis set information                   |
              +--------------------------------------------------+

              we will work with the 1s 3p 5d 7f 9g ... basis set
              ...i.e. with spherical basis functions...

   type   atoms  prim   cont   basis
   ---------------------------------------------------------------------------
    n        1     46     31   def2-TZVP   [5s3p2d1f|11s6p2d1f]
    h        6      8      6   def2-TZVP   [3s1p|5s1p]
    b        1     46     31   def2-TZVP   [5s3p2d1f|11s6p2d1f]
   ---------------------------------------------------------------------------
   total:    8    140     98
   ---------------------------------------------------------------------------

   total number of primitive shells          :   46
   total number of contracted shells         :   46
   total number of cartesian basis functions :  108
   total number of SCF-basis functions       :   98

[...]

 all orbitals will be included in the transformation

 number of non-frozen orbitals          :    98
 number of non-frozen occupied orbitals :     9

[...]

And the corresponding output from aomix:
Quote
[...]
  Running the FRAGMENT MO (FO) calculation...
  Spin-restricted calculation
  108 (   98 canonical) orbitals

[...]

  Number of basis functions does not match the number of canonical orbitals.

Using 6d/10f I get:
Quote
[...]
              +--------------------------------------------------+
              |               basis set information                   |
              +--------------------------------------------------+

              we will work with the 1s 3p 6d 10f 15g ... basis set
              ...i.e. with Cartesian basis functions...

   type   atoms  prim   cont   basis
   ---------------------------------------------------------------------------
    n        1     51     36   def2-TZVP   [5s3p2d1f|11s6p2d1f]
    h        6      8      6   def2-TZVP   [3s1p|5s1p]
    b        1     51     36   def2-TZVP   [5s3p2d1f|11s6p2d1f]
   ---------------------------------------------------------------------------
   total:    8    150    108
   ---------------------------------------------------------------------------

   total number of primitive shells          :   46
   total number of contracted shells         :   46
   total number of cartesian basis functions :  108
   total number of SCF-basis functions       :  108

[...]

 all orbitals will be included in the transformation

 number of non-frozen orbitals          :   108
 number of non-frozen occupied orbitals :     9

[...]

And
Quote
[...]
  Running the FRAGMENT MO (FO) calculation...
  Spin-restricted calculation
  108 (   108 canonical) orbitals

So the automatical conversion to 6d might work but then some MOs are missing an aomix stops working.

Or do I make any mistakes?

uwe

  • Global Moderator
  • Sr. Member
  • *****
  • Posts: 491
  • Karma: +0/-0
Re: Problem with AOMix
« Reply #5 on: April 14, 2015, 04:36:31 PM »
Hello,

you are right, the number of molecular orbitals is the same as the number of spherical atomic basis functions, while the number of basis functions per molecular orbital is the number of cartesian basis functions. The latter one is transformed to 6d/10f only.

I am not sure why the number of orbitals must be equal to the number of basis functions in AOMix-FO, but perhaps there is a possibility to get around this restriction...

Using cartesian basis sets should work, please contact the Turbomole support to get help in case it does not.

Regards,

Uwe

« Last Edit: April 14, 2015, 11:37:52 PM by uwe »

mwoch

  • Newbie
  • *
  • Posts: 5
  • Karma: +0/-0
Re: Problem with AOMix
« Reply #6 on: April 14, 2015, 11:02:47 PM »
Allright, I will do so.

Thanks a lot!

riteshruhela

  • Newbie
  • *
  • Posts: 1
  • Karma: +0/-0
Re: Problem with AOMix
« Reply #7 on: July 08, 2016, 11:24:08 AM »
Hi,

Can anyone help me how to use t2aomix script in control file.
I am not getting the relevant input files for aomix programme.
I am using Turbomole 6.4

Regards