Streitz-Mintmire potential

LAMMPS LAMMPS Mailing List Mirror
1

Hi LAMMPS team and users,

I was browsing through the archive of the 2013 workshop, and found from Ray Shan’s talk that LAMMPS is about to release the original SM potential (any tentative release date?). This is an exciting development for many of us who want to model oxygen-metal interaction with dynamic charge transfer with well-established EAMs.

Can you let me know, from your experience, that how the original SM behaves, in case of pure metal+O, when attached to Zhou and Wadley EAMs? Especially for non-bulk simulations under compression? I do understand that modification by Zhou (PRB 2004) is necessary for multi-metal simulations.

Thanks for your patience.

Sanket

2

Sanket,

Implementation of the Streitz-Mintmire potential will be released
soon; thank you for your interest and please stay tuned.

I am not sure what you are asking. SM is only a variable charge long
range Coulomb that can be hybrid with any short range pair styles,
e.g. EAM, Tersoff, LJ, etc. How the hybrid potential behaves depends
on the parameterization. Only way of knowing this for sure is to look
through the Literature and investigate it yourself. The
implementation does not yet include the modification after Zhou 2004.

Cheers,
Ray

3

Ray,

Thanks for your reply. My concern is mostly this: is it okay to pursue the original SM potential despite its known problems as reported by Zhou et. al. in PRB 69, 035402 (2004), or should it be best to implement the modification suggested by Zhou? I must say that I have never worked on ionic systems, and I am mostly a metals guy.

Particularly, I am worried about one of their comments in PRB: “We have also found that the original CTIP model is always unstable and always causes calculation overflow if it is combined with other EAM potentials such as that reported in Ref. 17 to even simulate an uncompressed ZrO2 or AlO2 surface. In addition, we have discovered that the overflow encountered during efforts to use the pair potential to approximate En (Ref. 26) also resulted from this instability. The constraint that the original CTIP model imposes on the choice of the nonelectrostatic potential including both format and parameters is clearly nonphysical and prevents a merging of the CTIP with the existing metal alloy EAM potential database for a variety of metals.”

I am interested in the Ni-O system, and hope the parameters works for it.

Thanks again for all your contribution !

Sanket

4

The Zhou modification to the original Streitz-Mintmire potential is
merely a charge barrier function over the electrostatics.

The "problems reported by Zhou" was already pointed out by Streitz and
Mintmire in their 1994 original paper that the SM electrostatics can
cause instabilities (likely due to Coulomb catastrophe) when coupled
with a pair potential. Therefore, Streitz and Mintmire recommended
the use of their variable charge electrostatics with many-body style
potentials. EAM is a pair potential with a many-body effect from the
electron density, so the situation should be a little better. But
problems can still occur, as noted by Zhou. I argue that the problems
is due to both EAM and SM. Note that SM can be used with any
many-body pair styles, not just EAM.

Ray

5

Hi Ray,

I agree. So lets wait and see how it behaves with different pair-potentials (different repulsive cores).

Zhou (PRB, 2004) also reported another issue- non-adherence to zero-charge in a multimetal system (no ionic bonds). While mathematically it is likely, do you think this is going to be a problem? Probably, we will have to do a few tests to answer these questions once you release the code.

Thanks for your time.

Sanket