unstable structure at zero pressure

Dear lammps users,

I am trying to use buckingham+coulombic potential to build a zirconia structure. I find that if the atoms fill the whole simulation box, the structure is stable, but I’m trying to simulate nanoparticle so a cluster will just occupy a small region, then the structure always become unstable. Is this because the atoms at cluster boundaries no longer have forces from neighbor as a result of PBC?

Any suggestion is well appreciated!

Cheers,

YAO

“God made the bulk; surfaces were invented by the devil”. Introducing surfaces in to your system is certainly going to change its morphology. Yet, what do you mean by unstable? Do you mean a simple rearrangement of atoms or an explosion-type of event? Did you try minimizing the initial cluster geometry first?

Carlos

Hello Carlos,

It’s like, first I build a crystalline ceramic structure, it should have the lowest potential, then I use minimization. If the atoms occupy the whole space, the positions of atoms stay untouched, if there are large void areas surrounding a cluster of atoms in the box, the cluster become amorphous. Does that make sense to you?

Again, thanks for replying.

YAO

Sounds reasonable, especially if the initial cluster is not too big. The atoms have to change positions because you have changed the way they interact when going from bulk to cluster. However, you want to make sure your minimization is converging. Additionally, keep in mind that this minimization will take you to a local minimum and not to the global minimum that the cluster can adopt given its total number of atoms.

Carlos

Sounds reasonable, especially if the initial cluster is not too big. The
atoms have to change positions because you have changed the way they
interact when going from bulk to cluster. However, you want to make sure
your minimization is converging. Additionally, keep in mind that this
minimization will take you to a local minimum and not to the global minimum
that the cluster can adopt given its total number of atoms.

on top of that, you have to keep in mind that for oxides and other
structures based on strongly polarized or ionic atoms surfaces and
edges can become highly energetic, if you cut them the "wrong" way and
also that there will be a significant change in the electronic
structure. while a simple potential with fixed charges may be good for
the bulk, it can become quite problematic for surfaces, edges, defects
and any nanoscale objects. usually, you need some kind of charge
equilibration process to properly model those kind of objects. LAMMPS
has a few options in that regard, it is worth looking closer at those,
study the relevant literature and make some tests.

axel.

I suggest you have a look at the work of Dean Sayle
(http://www.researchgate.net/profile/Dean_Sayle/publications/)
who has done extensive work on building clusters of oxides.

Nigel