Dear LAMMPS users,
Hi,
Actually I'm a bit confused. In the following link which is manual of
LAMMPS it says that " If the simulation box is currently triclinic and
has non-zero tilt in xy, yz, or xz, then it cannot be converted to an
orthogonal box ".
LAMMPS Molecular Dynamics Simulator
<LAMMPS Molecular Dynamics Simulator the other hand, some
peopole on this forms suggested to use change_box command. What should I do
?! By the way, I have a triclinic simulation box and I should convert it
into a cubic before doing my simulations.
actually, what i recommended you to do is to build a supercell and
*before* even LAMMPS gets involved.
what you are asking for and the procedure is not very smart. LAMMPS will
not do the thinking for you.
there is no "magic" command (or sequence of commands) that will do what you
need to do without you actually thinking and figuring out the process of
what supercell you need.
/--------/
/ /
/ /
i
f you have a cell like shown above, then you need to replicate 3x in y
direction. so that the periodicity for an orthogonal box with the same area
matches that of the tilted box.
/--------/
/ /
/ /
/--------/
/ /
/ /
/--------/
/ /
/ /
/--------/
/
/
/ /
/--------/
/ /
/ /
/--------/
/ /
/ /
now you can wrap your coordinates back into an orthogonal box and the
structure will fit properly. NOTE: this is *wrapping* the coordinates, not
moving/deforming box and coordinates. you received these bad suggestions
because of the very bad and incomplete description of your problem.
--------|
|
>
--------|
|
>
--------|
|
>
if you insist on (near) cubic, you can replicate in x as well.
this is easy to work out with a pencil, a piece of paper, an some thinking.
it probably can also be done by using LAMMPS smartly, but unless you have
worked out all the steps i already explained before, you'll never succeed.
it is easier to do with some scripting outside of LAMMPS.
at any rate, this is basic geometry stuff and i feel a little embarrassed
to have to explain this at this level of detail.
axel.