CCL: DL_POLY error message "1-4 separation ..."
- From: "Bill Smith" <w.smith##dl.ac.uk>
- Subject: CCL: DL_POLY error message "1-4 separation
..."
- Date: Wed, 27 Sep 2006 06:49:28 -0400
Sent to CCL by: "Bill Smith" [w.smith---dl.ac.uk]
The error concerning the 1-4 distance exceeding the cutoff distance arises
because you have specified a dihedral potential in which the atoms in the 1 and
4 positions have separated to a distance larger than the van der waals cut-off
that is applied to pair forces.
This creates a problem for the routine that calculates the dihedral force, which
needs to calculate a 1-4 interaction, but there is no reference to the required
force in the bookkeeping arrays and no associated interpolation values in the
potential and force tables, which are constructed using the defined cut-off. It
is generally not safe simply to assume the 1-4 term is zero at this distance.
The problem arises in a simulation for two common reasons: the cut-off specified
is too short to allow for the 1-4 term in a `stretched' molecule; or the
parameters used to define the bonds, bond angles and dihedrals are incorrect or
possibly too `soft' and have allowed the molecule to assume an improbable
structure.
A third possibility is that the simulation is wildly unstable and is pushing the
structure to extremes. There should be other symptoms here - like high
temperatures and high configuration energies. The integration algorithm may also
show instability - large energy fluctuations for example.
If your cut-off is large (> 10A say) then you should check that the force
field has been correctly defined.