CCL: Limitations of DFT

 Sent to CCL by: "Benjamin  Ellingson" [ben*|*eyesopen.com]
 >>>>>>>>>>>>>>>>>>>>>>>>>>
 The experimental results were not conclusive in differentiating beteen the
 mechanisms, so I  proposed in the manuscript that a purely computationally-based
 follow-up paper on the intermediates, transition states (with IRC calculations),
 etc.  at the DFT B3LYP/6-311+G(d,p) level of theory(a paper I am currently
 working on) would help to distiguish between the mechanisms.
 <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
 The single-reference and spin-restricted warnings already discussed are
 important points.  Another thing to keep in mind is that the B3LYP method just
 isn't very good for kinetics.  There are much better DFT methods for calculating
 transition states and minimum energy paths.  All due respect to the immortal
 B3LYP, as it still holds its own when calculating atomization energies, but it
 is very difficult to do everything well.
 I would recommend using MPW1K and perhaps a newer method as well.  A good paper
 to look at is "The Design of Density Functionals that are Broadly Accurate
 for Thermochemistry, Thermochemical Kinetics, and Nonbonded Interactions,"
 Y. Zhao and D. G.. Truhlar, Journal of Physical Chemistry A 109, 5656-5667
 (2005).  There are newer functionals that are able to beat MPW1K in database MUE
 tests, but the older MPW1K just keeps sticking around towards the top and has
 build a track record.  Like any functional, MPW1K sometimes gives bad results,
 but it appears that this is often caused by multi-reference issues, and not
 necessarily by a problem the functional itself.
 Ben Ellingson, Ph.D.
 Truhlar Group alum (<-disclosure)