From owner-chemistry@ccl.net Fri May 14 20:06:00 2010
From: "David Hose anthrax_brothers^hotmail.com" <owner-chemistry],[server.ccl.net>
To: CCL
Subject: CCL:G: Freq and Stability
Message-Id: <-41862-100514162900-30529-l0o1/zYzWwGvVfzY6L4t+Q],[server.ccl.net>
X-Original-From: "David  Hose" <anthrax_brothers#hotmail.com>
Date: Fri, 14 May 2010 16:28:58 -0400


Sent to CCL by: "David  Hose" [anthrax_brothers(_)hotmail.com]
Anthony,

You might want to read Chris Cramer's posting about this from 2007.

http://ftp.ccl.net/chemistry/resources/messages/2007/06/14.012-dir/index.html

You'll also find some other useful tips in the days prior to Chris' posting.

> From personal experience, using a finer integration grid (int=ultrafine in G03) solves the problem 
of a number of small imaginary (<50 cm-1) frequencies.  However, you'll have to rerun all of other 
calculations for your series of compounds with the same grid size if you want to compare energies.  
Also, you'll have to take a hit on computational speed.

Regards,

Dave.


----------------
Dear CCL,

I've been running geometry optimizations and found stationary points. I have also been using SCF = 
QC (also UB3LYP/6-311+G(d,p) NOSYMM).

How can I be sure that my final answer is a local minimum? When I perform a frequency job after 
the optimization, I sometimes obtain imaginary frequencies, suggesting an nth-order saddle point 
(n is the number of imaginary frequencies). 

Does this also have to do with the stability of my wavefunction? I have used the keyword 
Stable=Opt for verification. But I find that I can have both a stable wavefunction and imaginary 
frequencies. How do I then find true local minima?

Also, is it possible to run both a STABLE and geometry OPT job with the same route? Or do I have to 
use the Link1 keyword? If so, is it possible to copy my checkpoint file through the route section?

Thanks

Anthony