Dear Michael,

The essential difference is in a linear transit scan you are stepping
through vairous values of one or more variables, while optimising all other
degrees of freedom. They can be used to move a molecule into conformations
which are not favourable to investigate the energy required to perform a
particular geometry distortion, such as a rotation of a ligand, or a
stretching a bond.

IRC calculations start at a saddle point and descend the path of least
resistance down to the minima which the saddle points connect. In each step
you are allowing all degrees of freedom to vary, and so you should obtain
the reaction path which passes over the saddle point.

You are correct in your description of the difference between the two
calculations, however as your example is so simple it is perhaps difficult
to see how the two are so different. If you consider a more complex system
like the migratory insertion of Me in MeMn(CO)5. To determine whether it is
migration of Me, or insertion of CO you could do two linear transits. One to
move a methyl group to the CO carbon, and comparing this to moving a CO into
the Mn-Me bond. Comparing the two would give you an idea of the relative
barriers, and hopefully a good guess for you to optimise a transition state.

Once you have a transition state you then can do a IRC calculation to verify
it is indeed the correct TS, and to elucidate the timing of the process. In
this case it is much clearer to see that LT's are useful to explore a
potential energy surface, whereas IRC's are used to verify TS's, and
determine the exact reaction path (often an LT can step over an intermediate
you might have on the way to the TS, hopefully an IRC will stop here and not
at your reactant geometry).

I hope this clears up the confusion you may have had.
Ben

----- Original Message -----
From: "Michael Drummond" <mdrummon_at_chemistry.ohio-state.edu>
To: <adf-list_at_scm.com>
Sent: Sunday, April 18, 2004 7:30 PM
Subject: ADF-LIST: Linear Transit vs Intrinsic Reaction Coordinate

> I have a simple question: what is the difference between a Linear Transit
> (LT) calculation and an Intrinsic Reaction Coordinate (IRC) calculation?
>
> Let's examine a trivial example, the inversion of NH3 from its C3v ground
> state to a D3h transition state to another C3v ground state. If I were to
> examine the barrier for this process with an LT methodology, I'd probably
> define a dummy atom along the C3 axis "above" the nitrogen. I'd then
> define the H-N-Du angle as my LT parameter, vary it from whatever it is in
> the ground state (but certainly obtuse) to what it would be in the "other"
> ground state (now an acute angle). If I step the path in increments of,
> say, 5 degrees for this angle, I feel I'd have a pretty fair assessment of
> the barrier for this process.
>
> For the IRC calculation, my starting point would be the D3h transition
> state. I could tell the calculation to walk "forward" or "backward," and
> presumably it'd also give me a fair assessment of the barrier for NH3
> inversion.
>
> So am I correct in assuming that the major difference is that you must
> know where you're starting and which way you're going for an LT
> calculation, but you only need know a midpoint for an IRC calculation? Or
> is there a subtlety I'm missing due to the extreme trivial nature of my
> ammonia example above?
>
> Thanks,
> Michael Drummond
>
> Bursten Group
> Department of Chemistry
> The Ohio State University
> 614.292.5227
> mdrummon_at_chemistry.ohio-state.edu
>
>
Received on 2004-04-18 23:10:02