CCL:G: Transition State TS Berry Optimization

[Billy McCann <thebillywayne##gmail.com>]

In addition to John's suggestion, I would add that calculating the force constants periodically will help towards geometric convergence, though with increased computation time. ÂI never dug into Gaussian to find out if one could specify the number of iterations between successive force constant calculations. ÂI just threw an 'opt=calcall' at it, which will calculate the force constants after every iteration. Â

When finding TS's, I would typically use some sort of bond length constraint and then perform a geometry minimization. ÂOnce that is complete, I'd release the constraint. ÂA relaxed PES scan also would help. ÂWith that, you're looking for an area where the energy is high, but the gradient is low. ÂAlso, look into the QST methods. Â

If you're looking for a book to help you with Gaussian, you can purchase the official Gaussian book, Exploring Chemistry With Electronic Structure Methods: A Guide to Using Gaussian [1].Â

Also check [2] for a guide on how to use the 'opt' keyword. ÂI keep this google operatorÂ[3]ÂaroundÂfor searching the Guassian keywords and such.

Hope that helps! Â

[1]Âhttp://amzn.com/0963676938

[2]Âhttp://www.gaussian.com/g_tech/g_ur/k_opt.htm

[3]Âsite:http://www.gaussian.com/g_tech/g_ur">www.gaussian.com/g_tech/g_ur

Billy Wayne McCann, Ph.D.

Oak Ridge National Lab

Chemical Sciences Division

Chemical Separations Group

On Tue, Sep 3, 2013 at 11:20 AM, Brian Jay Levandowski levandbj|a|plu.edu <owner-chemistry---ccl.net> wrote:

Sent to CCL by: "Brian Jay Levandowski" [levandbj{}plu.edu]
Hello,

I am running a TS Berry optimization with B3LYP 6-31+G. When the optimization converges (after
many iterations) the output geometry is a mess.My input geometry is my best guess from
minimizing the reactant then elongated the bond to 2.1 angstroms. I then froze the bond at 2.1
angstroms and did another minimization optimization to get my input geometry for the TS search.

In looking at the TS search intermediate gemoetries, intitially it looks like it is headed to the TS and
possibly converging, but then it jumps to the reactant and the energy spikes way up. It then heads
downhill and converges to a TS involving a C-C bond breaking instead of the C-O bond I initially
elongated.

My questions are:
1. I took one of the intermediate geometries that looked like what I would expect the intermediate
to look like and ran an IRC on it. It lead to the reactant and an intermediate, but it was not a true TS
(second derivative) from the TS Berry search as it did not converge on that geometry. So is it not the
TS I am looking for?

2. Could I constrain/freeze the C-C bond that is being broken in my TS Berry optimization to
prevent the spike from happening where it enters weird geometries? Or do constraints make the
system to artificial.

3. I would also like to if it is possibly to modify how much the bond length increases and each step
in which a calculation is done. In my optimization the C-O bond starts at 2.1 then jumps to 2.35,
2.58, 2.83... hits 3.85 then goes back down eventually to 2.49 (this is where I think the TS is, but it
never converges) then continues to 1.44 which is the bond length in the reactant then is spikes and
goes askew as mentioned above. Should I try and scan the distance of the C-O bond from say 1.85
to around 3.0 in increments of 0.1 angstroms? If so should I also constrain all the other variables or
does a scan constrain everything but the variables being manipulated?

Any pointers/resources on finding TS with gaussian would also be helpful. I am an undergrad
interested in going to grad. school for computational chemistry. I've used Lewar's computational
chemistry but it would be great if there was a guide more specific to gaussview/gaussian.

Thank you,
Brian Levandowski
levandbj],[plu.edu



-= This is automatically added to each message by the mailing script =-
E-mail to subscribers: CHEMISTRY---ccl.net or use:
   http://www.ccl.net/cgi-bin/ccl/send_ccl_message

E-mail to administrators: CHEMISTRY-REQUEST---ccl.net or use
   http://www.ccl.net/cgi-bin/ccl/send_ccl_message

Subscribe/Unsubscribe:
   http://www.ccl.net/chemistry/sub_unsub.shtml

Before posting, check wait time at: http://www.ccl.net

Job: http://www.ccl.net/jobs
Conferences: http://server.ccl.net/chemistry/announcements/conferences/

Search Messages: http://www.ccl.net/chemistry/searchccl/index.shtml
   http://www.ccl.net/spammers.txt

RTFI: http://www.ccl.net/chemistry/aboutccl/instructions/