Calculating Point Defect Energies in Iron Using Molecular Dynamics Simulations

[nuclear85]

Hi everyone... I hope this is the right place to post.

Anyway, I am a first year grad student working on molecular dynamics simulations (with very little background in materials, yay) of iron. To start out, I'm just trying to replicate some materials constants with a basic lattice to make sure things are going well...

I want to calculate the energies of formation of interstitial dumbbells in various directions, <100>, <110>, <111>... I set up a perfect iron lattice, 10x10x10, and find it's total energy (-8244.8702 eV). Then, I insert an interstitial in the lattice inside one of the cubes (it's bcc iron), near where it's position would be for a <110> dumbbell, and let it run for about 100 timesteps to relax into a stable position. This is all at 0 K, by the way, and an NVE constant simulation. Now looking at the total energy, I have (-8141.51). I had thought that the magnitude of the energy should be higher with an interstitial, but it's not... so does anyone have an explanation for that, or a formula to calculate energy of intersitial formation?

I thought it would be analogous to the energy of vacancy formation equation (which I calculated fine), but doesn't seem to be... thanks

 

[handsomecat]

Basically, the ground-state configuration should have the lowest energy (ie. most negative). Any configuration with defects (eg. surface, interstitials, vacancy etc etc) should have higher energies (ie. not as negative as the ground state energy).

When performing MD simulations, one should always check for these things. This is because the MD potential-fitting process does not guarantee that all defects will have higher energy.

Now in your case, what you should be comparing is Energy per atom, since the system with the interstitial has one more atom than the perfect crystal.

The energy per atom of the system with the interstitial should be less negative than the energy per atom of the perfect crystal. Its not the magnitude. From the values you provided, this seems to be the case.

If the other way round happens, then the conclusion to be drawn is that the MD interatomic potential that you have chosen is not suitable to simulate interstitials.

Sounds interesting. What software are you using?

 

[nuclear85]

Hi handsomecat-

Thanks for the help... I am new to the whole materials thing, but what you said definitely makes sense. I am using the LAMMPS software (http://lammps.sandia.gov/), freely available. Right now I'm just trying to get a good feel for the program by simulating iron using a potential from Mendelev & Ackland... eventually I will be coding in a potential for chromium and studying dislocation and defect movement in FeCr.

 

[handsomecat]

Hi handsomecat-
eventually I will be coding in a potential for chromium and studying dislocation and defect movement in FeCr.

A few things you have to bear in mind. Kinda sharing my experience.

Right now you have a Fe-Fe potential, and you want to code in a Cr-Cr potential. But you still need the Fe-Cr potential ...

If there is no one else who has previously studied FeCr with MD, you will certainly have to develop a Fe-Cr MD potential that can simulate the Fe-Cr alloys that you want to. So pray hard that there is!

You may also have to do Density Functional Theory calculations to see how they can complement your research too.