pressure calculation

Hello Lammps programmers, in the calculation of compute_pressure is given
a simetric pressure tensor but for systems involving three-body
interaction forces is not valid( i use meam potential for gallium and
Stillinger-Weber potentials for Silicon), in lammps is there a way to
calculate the remaining terms of pressure tensor. which method is used to
calculate the contribution for long range coulombic force is Ewald
sumation our PPPM.

Thank you very much look forward to your prompt response.

Samuel

Samuel,

A few things you should be aware... First, there's has been a lot of
recent work to address the question of whether the stress(pressure) tensor
should be symmetric or not. Admal and Tadmor have shown that the symmetric
form is correct regardless of the type of interatomic potential assumed
(3-body or whatnot). For a thorough (and I mean comprehensive) discussion
see:

NC Admal & EB Tadmor, J. Elasticity, v 100, p. 63-143 (2010)
NC Admal & EB Tadmor, J. Chem. Phys., v 134, p. 184106 (2011)

This should mean that LAMMPS is all set to store the stress tensor,
provided adequate routines are implemented in the MEAM and SW pair
classes. This may take some investigation and possibly implementation on
your part.

As for long-range Coulombic forces via Ewald/PPPM, you're correct in that
the inter-atomic contributions to any stress measure will be missing. This
is a long-standing problem, and I'm not aware that lammps has any
solutions available.

Good luck,

Jon Zimmerman

The MEAM and SW potentials already calculate a total and per-atom
stress tensor. And it is symmetric. It is simply F dot r summed
over the atoms in each interaction. Which is as valid a way as
any to compute local stress and assign it to atoms. See the papers
Jon listed, and also the paper by A Thompson on the LAMMPS citation
WWW page.

Steve