Breaking symmetry in DFT singlet state calculation


  • Guest -
12:57:08 PM PDT - Tue, Jun 28th 2011
Hi, I am trying to compute the ground state of a singlet molecule which is supposed to be antiferromagnetic, but I can't manage to break the symmetry in order to observe the magnetic state even with the commands "ADAPT off" and "SYMM off" using ODFT (spin polarized computation). I guess I need to use an asymmetric atomic density as the initial guess. Is there a way to do so, such as the "Guess=Mix" command in Gaussian ? Thanks in advance.

  • Guest -
4:14:03 AM PDT - Wed, Jun 29th 2011
I am not familiar with ODFT calculations, but maybe the statement "symmetry c1" at the beginning of your geometry section would help.

  • Guest -
6:55:22 AM PDT - Wed, Jun 29th 2011
Thanks, but I am already using "symmetry c1" without success.

Forum Vet
5:13:12 PM PDT - Wed, Jul 6th 2011
There is something similar which needs to be documented:

   set mixang 45

Should allow you to do what you want. Just set this before "task"

Bert


Quote: Jun 28th 7:57 pm
Hi, I am trying to compute the ground state of a singlet molecule which is supposed to be antiferromagnetic, but I can't manage to break the symmetry in order to observe the magnetic state even with the commands "ADAPT off" and "SYMM off" using ODFT (spin polarized computation). I guess I need to use an asymmetric atomic density as the initial guess. Is there a way to do so, such as the "Guess=Mix" command in Gaussian ? Thanks in advance.

  • Guest -
9:52:55 AM PDT - Tue, Jul 19th 2011
Unfortunately, "set mixang 45" is not working either.

Forum Vet
1:32:04 PM PDT - Wed, Jul 20th 2011
Can you post the input deck, so I can take a look at it. Without knowing the system it's hard to give more suggestions and help.

Bert


Quote: Jul 19th 4:52 pm
Unfortunately, "set mixang 45" is not working either.

Forum Regular
7:43:53 AM PDT - Thu, Jul 21st 2011
Hi,

Have you tried the fragment approach ? There is an example on how to use fragments in the vectors section of the manual.

For example, if you are interested in opposite spins. i.e up on one atom and down on the other

Set mult 1 for one fragment

Set mult -1 for the other fragment

Converge them separately and then use the fragment guess approach to put them together.

Let us know if you need more help with this.

Best,
-Niri

  • Guest -
6:09:07 AM PDT - Tue, Jul 26th 2011
Yes, the fragment guess works well. Thanks


Forum >> NWChem's corner >> Running NWChem