CCL:G: IRC calculation

 Sent to CCL by: Justin Finnerty [justin.finnerty*_*uni-oldenburg.de]
 On Tue, 2009-03-03 at 09:48 -0500, Kavitha Velappan
 kavitha.velappan(a)gmail.com wrote:
 > Sent to CCL by: "Kavitha  Velappan"
 [kavitha.velappan_-_gmail.com]
 > Hi all,
 >
 > I tried to do IRC calculation to follow the reaction path for conformer
 interconversion (Dimethoxymethane G+G- to G-G- conformer - rotation around OCH3
 bond). The route section looks like this:
 >
 > #N RB3LYP/6-31++G** IRC=(CalcFC,maxpoints=12)
 I have had this issue whenever I do an IRC with DFT in gaussian.  The
 problems are as you see, the IRC only goes in one direction and doesn't
 reach maxpoints.
 I found that the best solution to this is to reduce the STEPSIZE. The
 default is 10 and I have had much success with STEPSIZE=4 or 3.  (You
 have to increase MAXPOINTS accordingly!)  Using this I have successfully
 used IRC to go from the TS to close to the ground state.  Using a small
 stepsize will increase the length of the calculation but not as much as
 it first appears because the previous step of the IRC is now much closer
 to the new step so each step converges in two or three cycles.
 The other issue is getting both the forward and reverse directions to
 work. I have found no reliable way to do this automatically in gaussian
 using DFT - gaussian always seems to go in one direction! My normal
 solution is to manually modify the TS structure in the direction
 gaussian didn't go and start the IRC from the modified structure.  How I
 modify the structure is of course based on the motion vectors of the
 imaginary frequency.
 PS Your input uses CALCFC, surely you could use READFC from the freq
 calculation for the TS?
 Cheers
 	Justin
 --
 Dr Justin Finnerty
 Rm W3-1-165         Ph 49 (441) 798 3726
 Carl von Ossietzky Universität Oldenburg