Re: conjugation in mol mech

[tripos!metis!matt - at - uunet.UU.NET (Matt Clark)]

 Neil Ostlund makes an important point about atom types in
 molecular mechanics, and how different bond types may occur
 between two atom types.
 The Tripos force field solves this problem by adding an extra
 variable to the force field, the bond type. This applies in
 particular to bond stretch and torsion terms.
 This allows more flexibility in handling conjugated systems. The
 single bond length between two sp2 carbons is about 1.47 Angstroms,
 the double bond length is 1.33 Angtroms. There are different
 torsion parameters for single and aromatic bonds used in the
 case of biphenyl torsions, allowing these to be parametrized more
 accurately. It isn't the same as a pi calculation, but significantly
 improves the accuracy for conjugated systems.
 I would encourage anyone developing new force fields to use this
 idea; it greatly expands the possibilities for future expansion of
 the force field. Organometallic modeling is much easier with
 this scheme - you can assign different bond lengths for bridging
 bonds, dative bonds etc., between the same two atom types.