CCL:G: Gaussian 03 - optimization (linear, non linear) - negative frequency

Jim,
 
You've received several answers by now.  The practical consequences of the theory you have learned are as follows:
 
1. You *do* have to guess whether the molecule / ion is linear or bent. 
 
2. However, if the molecule is bent, you don't have to know the angle.
 
3. If you guess the molecule is linear and it's actually bent, you will get those imaginary  (negative) frequencies.  If you animate the frequencies in a program such as Gaussview, you will see the molecule "trying" to bend. 
 
(It's *always* a good idea to calculate frequencies and check that there are no negative/imaginary frequencies!)
 
4. If you guess the molecule is bent and it's actually linear, Gaussian will optimize the angle to ~179.9 degrees, and then will (usually) crash with an error message.
 
Regards,
--David Shobe
  Süd-Chemie, Inc.
 
Attention to detial is essentail.

From: owner-chemistry() ccl.net [mailto:owner-chemistry() ccl.net]
Sent: Friday, January 26, 2007 12:40 AM
To: Shobe, David
Subject: CCL:G: Gaussian 03 - optimization (linear, non linear) - negative frequency

Dear all

    I have a problem concerning optimization of molecules of more than 2 atoms.
I am using Gaussian 03 to optimize and perform a frequency calculation of molecules such as for example NaCN. NaCN is a T shape molecule and has a non-linear geometry.

 If I input a linear initial geometry and perform an optimization, will the molecule (which is actually non-linear) converge to a non-linear geometry

I have perform an optimization of linear NaCN and the stationary point was found.The optimized geometry remains linear. But when I perform optimization of
non-linear NaCN with initial angle 120, the required geometry (T shaped) was obtained.

My question is that must one have a knowledge of the geometry of a molecule under study in its ground state to perform optimization. If so what is the purpose of optimizing. Can an initial linear structure input converged to an optimized structure which is actually non-linear
Can Gaussian 03 predict the structure of a molecule for which  the geometry is unknown. If so how?

I know that Gaussian 03 required an approximate initial geometry. Is there a way
from theory to determine whether the molecule in question is linear, bent or  even a T shape.

For example OCS molecule is linear. If one is going to calculate the adiabatic electron affinity, one has to optimise the anion OCS- which I did with a linear
geometry just like with OCS. The optimize anion OCS- remains linear. But I have found from many publications that anion OCS- is actually  a bent shape molecules.Again I perform an optimization of OCS- but this time with a bent shape, angle 120. This optimized geometry with angle 136.7 coincide with the publications. The angle has change  significantly from angle 120 to 136.7.

Why the initial linear geometry of anion OCS- did not converge to the optimised geometry with angle 136.7

note that the electron affinity computed with linear anion OCS- and the bent shaped anion OCS- are totally different. The one which coincide with the experimental value is the bent shaped as required.

I perform a frequency calculation with the optimized geometry of the linear anion
OCS-  and obtain negative frequency (saddle point which specifies a transition state). Does a  negative frequency mean the optimized geometry has not actually
been optimized. Does in general a negative frequency is an indication that the initial geometry input is not correct and hence must alter the initial geometry just like I did with the bent shape.

Thanking you all

 Best regards

Jim


Check out the all-new Yahoo! Mail beta - Fire up a more powerful email and get things done faster.
This e-mail message may contain confidential and / or privileged information. If you are not an addressee or otherwise authorized to receive this message, you should not use, copy, disclose or take any action based on this e-mail or any information contained in the message. If you have received this material in error, please advise the sender immediately by reply e-mail and delete this message.
Thank you.