[lammps-users] elastic constants?

Marcin,

LAMMPS currently does not have a way to calculate elastic constants and bulk moduli directly. The new triclinic geometry enables future fixes to do just that. In the upcoming package we will be adding a general NpT ensemble. Variations in shape stemming from application of this fix will allow you to deduce the compliance and from it the elastic moduli. You can add a general NpT ensemble through a fix. Shape measurements could be done either through a fix or the new compute functionality.

Pieter

Elastic constants are equal to the second derivatives of the energy
density with respect to strain. I was not going to run npt to deduce
it, but calculate (numerically) the second derivatives of the
potential energy.
So the triclinic PBC is necessary, but not npt fix. Although i'm not
an expert here.

Marcin

Marcin,

Indeed you can calculate it that way too. I did a paper that way for the stiffness matrix of semicrystalline polyethylene (macromolecules 2006, 39, 439) by applying a higher order fit to the energy, followed by the derivatives. Getting the compliance matrix is a bit tricky because of statistical noise, but doable. Young's moduli etc follow from the compliance matrix.

Pieter