One can use Cartesian coordinates to input geometry and still
using Z-matrix for optimization - the simplest way and probably
the best way is indeed to use DLCs as it was already suggested,
however, this is not required so one can manually specify any
Z-matrix in the $zmat group.
Another point is that one can create and use the customized
"mixed" basis set definition file and use it as an external
basis set library - see e.g. http://classic.chem.msu.su/gran/gamess/ecp_tutorial.html
On Wed Apr 21 '10 10:03am, Stefan Boresch wrote
>We plan to scan energy surfaces of N-acetyl-methylamide amino
>acids ("blocked" amino acids) to verify (or disprove) "features" found in various force fields. In the case of, e.g., Ser, this is a 23 atom / 86 electron system. At both ends of these systems, there are methyl groups that are almost certainly devoid of any interest; however, they cannot be replaced by a hydrogen. Since eventually we may need multiple geometry optimizations on the MP2 level, I would like to use a minimal basis set for these, and a reasonable basis for the interesting part of the molecule.
>Now, if I set COORD=UNIQUE, then this is easy and well documented and
>the PUNCH file is my friend... However, in the scans of the energy surface I need dihedral angles frozen. This seems to imply using
>COORD=ZMT (or ZMTMPC). Is there a way to supply explicitly a basis set
>for each atom when using/relying on coordinates in z-matrix format?
>Alternatively, is there a way to freeze certain degrees of freedom
>if cartesian (unique) coordinates are used?
>Apologies if I overlook something trivial in the documentation; as
>far as quantum chemical calculations are concerned, I am a complete
>Thanks for any hints,
[ This message was edited on Fri Apr 30 '10 at 8:23pm by the author ]