[lammps-users] liquid metal deposition simulation questions

LAMMPS LAMMPS Mailing List Mirror
1

Hello,

I have taken on the task of simulating the impact on Ni-Al liquid metal droplets on a titanium surface using MD. I have experience with other MD codes, but very little with LAMMPS. I have a number of questions that members of this list can perhaps help with.

1. It seems from reading relevant literature that the standard potential functions used in liquid metal deposition simulations are based on EAM or FS+ potential functions. Can anyone point me to appropriate potential data for the Ni-Al system? What 'mixing' strategies are used for the force fields when 3 or more species (e.g. Ni+Al, Ti surface) are present?

2. What happens in the case of non-stoichiometric alloys (e.g. Ni + 5%wt Al)? Is there a recommended procedure for generation of such 'alchemical' potentials if I am not interested in distinguishing between the elements?

3. What is the recommended equilibration protocol for a liquid metal droplet using LAMMPS?

4. Is there any support within LAMMPS for defining, for a droplet impacting on a surface, a 'crash' region within the substrate in which the full potential expression is used, surrounded by a larger substrate region in which the atoms are constrained to only small deviations from their regular lattice positions, the interface between the two regions being handled with some kind of damping term (the goal being to reduce the effects of reflected impact pressure waves) ?

5. Can such simulations be 'coarse-grained' within LAMMPS to larger particles of the 'granular' type?

I would be very glad to get advice on any of these questions, pointers to literature or contact email addresses to groups doing similar research studies.

Many thanks in advance,
Steve Kirk

2

These are all good questions - but I can only give
simple answers to get you started

1. It seems from reading relevant literature that the standard potential
functions used in liquid metal deposition simulations are based on EAM
or FS+ potential functions. Can anyone point me to appropriate potential
  data for the Ni-Al system? What 'mixing' strategies are used for the
force fields when 3 or more species (e.g. Ni+Al, Ti surface) are present?

The pair_style eam doc page in LAMMPS has ptrs to a few WWW pages
that have EAM potentials. For NiAl I think a few exist, though they
may depend on what alloy/phase you are interested in.

From the LAMMPS perspective, you can mix as many elements as you

want. It just needs interactions rules for each I-J pair. These can be
inferred from single-element files (see pair_style eam), or included
directly in your EAM file (see pair_style eam/alloy).

2. What happens in the case of non-stoichiometric alloys (e.g. Ni + 5%wt
Al)? Is there a recommended procedure for generation of such
'alchemical' potentials if I am not interested in distinguishing between
the elements?

LAMMPS doesn't care - it's just an I-J interaction between atoms
of two elements. It's up to you to parameterize your EAM functionals
differently in the EAM file if you think a 2% alloy is different than
a 10% alloy.

3. What is the recommended equilibration protocol for a liquid metal
droplet using LAMMPS?

No different than any other equilibration. E.g. run NVT with
free boundaries.

4. Is there any support within LAMMPS for defining, for a droplet
impacting on a surface, a 'crash' region within the substrate in which
the full potential expression is used, surrounded by a larger substrate
region in which the atoms are constrained to only small deviations from
their regular lattice positions, the interface between the two regions
being handled with some kind of damping term (the goal being to reduce
the effects of reflected impact pressure waves) ?

You can define geometric regions and groups of atoms therin. Then
you could apply temperature damping (viscous or rescaling) in that outer
region. The interface is trickier. If you know precisely what you want
to do, then you could write a new "fix" that implements the damping
at the interface you want. But there isn't anything automatic for that in the
code.

5. Can such simulations be 'coarse-grained' within LAMMPS to larger
particles of the 'granular' type?

You can mix and match potentials and atom types (granular and metal),
but I haven't heard of any models like that for granular with atomistic.
It's not so much a LAMMPS issue as an issue of the physical model
you want.

3
  1. We did a similar work for impacts of a large cluster (E~ 50 eV/atom) on a metal surface using EAM potentials some time ago.
    We combined atomistic MD for the “hot spot” in the collision area with a continuum mechanics for the less deformed and “cold” areas. Please see the details of our simulation in PRB 61 (2000) 8744.
    I don’t think lammps is supposed to do this.