Spin-Rotation constants

If you think you found a bug this is where you can report it
Post Reply
uachstno
Posts: 1
Joined: 25 May 2014, 13:18
First name(s): Stanislav
Last name(s): Komorovsky
Affiliation: CTCC, Uit The Arctic University of Norway
Country: Norway

Spin-Rotation constants

Post by uachstno » 25 May 2014, 13:30

When calculating spin-rotation constants with common gauge origin the off diagonal terms are incorrect.

Calculation with london orbitals:
34.084225 20.032869 0.000000
47.343131 30.898881 -0.000000
-0.000000 0.000000 32.643309

Calculation with common gauge in the center of mass:
34.255338 303.963380 0.000000
575.854478 30.607068 -0.000000
-0.000000 -0.000000 32.207030

I have used uncontracted cc-pvqz basis set to get converged results for calculation with common gauge.

My guess is that the off diagonal contribution from electric field integrals have wrong sing.

I have attached all imput files.

Stano
Attachments
SR-LN.dal
(115 Bytes) Downloaded 477 times
SR-CM.dal
(123 Bytes) Downloaded 470 times
MOL.mol
(344 Bytes) Downloaded 500 times

michaljz
Posts: 7
Joined: 15 Nov 2013, 11:46
First name(s): Michal
Last name(s): Jaszunski
Affiliation: ICHO PAN
Country: Poland

Re: Spin-Rotation constants

Post by michaljz » 04 Jun 2014, 12:05

hello,

it appears that the reported bug is related to a sign problem in the Diamagnetic contribution.

In a test H2O London-NoLondon HF calculation I get for the first H atom:

NoLondon
Diamagnetic contribution
------------------------

A -189.736232 139.393758 0.000000
B 274.103959 -201.376302 0.000000
C -0.000000 0.000000 -197.314587

with the signs of the off-diagonal terms reversed for the second H atom.
With the same *.mol file (thus, the same ordering of H atoms)

London calculation gives


A -189.550915 -146.733765 0.000000
B -259.481844 -201.027507 -0.000000
C 0.000000 0.000000 -196.861315

and this leads to very reasonable

Total spin rotation constants
-----------------------------

A 32.167417 19.594557 -0.000000
B 43.706389 29.313292 0.000000

whereas in the NoLondon calculation the total is close to that reported
earlier (see below) for a DFT common origin job (and the error is removed
reversing the sign, with the change of 2*139.393758)

regards,
Michal

On Sun, 25 May 2014 14:30:14 +0200, no-reply@daltonprogram.org wrote:

Posted by uachstno

When calculating spin-rotation constants with common gauge origin the off
diagonal terms are incorrect.

Calculation with london orbitals:
34.084225 20.032869 0.000000
47.343131 30.898881 -0.000000
-0.000000 0.000000 32.643309

Calculation with common gauge in the center of mass:
34.255338 303.963380 0.000000
575.854478 30.607068 -0.000000
-0.000000 -0.000000 32.207030

I have used uncontracted cc-pvqz basis set to get converged results for
calculation with common gauge.

My guess is that the off diagonal contribution from electric field
integrals have wrong sing.

I have attached all imput files.

Stano

----------

View the first unread post:

http://daltonprogram.org/forum/viewtopi ... ead#unread

Change your notifications settings:
http://daltonprogram.org/forum/ucp.php? ... boardwatch

taylor
Posts: 525
Joined: 15 Oct 2013, 05:37
First name(s): Peter
Middle name(s): Robert
Last name(s): Taylor
Affiliation: Tianjin University
Country: China

Re: Spin-Rotation constants

Post by taylor » 04 Jun 2014, 14:39

I realize that the original poster may want to be doing isotopically substituted systems like HOD (say), given that symmetry was explicitly turned off and the Isotope keyword was given for each atom in the .mol file. Out of interest (and I have to say it is "out of interest" because I do not quite see how it could lead to a problem and know very little about these types of calculations), is the sign error still there if symmetry is turned on (or allowed to be discovered automatically)? And how do HOD calculatiosn compare between no-London and London cases --- is there still an obvious sign problem in the former?

Best regards
Pete

michaljz
Posts: 7
Joined: 15 Nov 2013, 11:46
First name(s): Michal
Last name(s): Jaszunski
Affiliation: ICHO PAN
Country: Poland

Re: Spin-Rotation constants

Post by michaljz » 06 Jun 2014, 09:12

hello,

I repeated the previously described H2O calculation now allowing the
program to detect symmetry, and the results appear to be

all wrong.

Using another program I verified that correct results are obtained for

a) all the components without symmetry when London orbitals are used
b) diagonal components without symmetry in a .NOLONDON calculation
- as mentioned earlier, in this case there is most likely only a sign
errorin the calculation of the off-diagonal terms in the diamagnetic
contribution.

When symmetry is allowed, two H atoms are recognized as a pair, but
all the components of all the contributions (beginning with the nuclear
contribution to the spin-rotation constant) are apparently wrong,
and that means all the components (including the diagonal ones and the
trace).

My guess is that at some stage in the calculation wrong coordinates
are used for the relevant nuclei

regards,
Michal

taylor
Posts: 525
Joined: 15 Oct 2013, 05:37
First name(s): Peter
Middle name(s): Robert
Last name(s): Taylor
Affiliation: Tianjin University
Country: China

Re: Spin-Rotation constants

Post by taylor » 06 Jun 2014, 09:33

I guess I wish I hadn't asked what turning symmetry on would do!! Nevertheless, your (MJ's) observations and the earlier posting indicate that there's a problem (or maybe more than one) with calculating this tensor, and I guess one of the experts in the magnetic side of things (which I'm afraid does not include me) needs to look at this.

Best regards
Pete

Post Reply

Who is online

Users browsing this forum: No registered users and 2 guests