Page 1 of 1

desorption from 110 rutile MgH2

Posted: Wed Jan 20, 2010 12:32 am
by neb_user
Hey, I have been trying for a long time to reproduce some results I found in the literature for desorption of two H atoms from 110 rutile MgH2 to form a H2 molecule 5A above the surface but I can not find a reasonable MEP. At first my timestep was too big and the H atoms that desorbed got too close as a result and VASP errored out but I fixed that by following the advice I saw on some forum posts about using a conservative POTIM (like 0.01 until the forces were down to 0.5eV/Ang and the moving up to POTIM=0.1). This prevented the two Hs from getting to close and the run converged... but not to a resonable MEP. My lab partner suggested trying different SPRING values in INCAR... sure enough a SPRING value of -0.5 allowed the run to converge (even with the default POTIM of 0.5)... but, still, not to a reasonable MEP.

So I have to sort of unrelated questions:
1) assuming the 50% coverage (bridging Hs only on the 110 rutile MgH2 surface) I'm using to replicate the literature results found are in fact appropriate to use, what will it take for me to get the MEP?
2) assuming the 50% coverage is a "dangerous" local minima (by that I'm imagining other lower energy surface coverages that are effectively causing the mechanism to be multi-step: is there a tool that can take in the primitive lattice vectors and basis for a crystal and then also take in the miller index for the surface you want to expose to vacuum and then also take in the number of atomic layers and the coverage you want (surface 50% covered with atom X vs 100% covered, for example) and then you hit enter and program spits out the coordinates you would need for your POSCAR file (because if yes, I could quickly build different coverages and test them to see which is a more stable initial state).

Thanks for you time...
Jason

Re: desorption from 110 rutile MgH2

Posted: Sat Jan 23, 2010 5:35 am
by graeme
Changing the spring constant will probably not make any difference.

What do you mean by not finding a reasonable MEP? Is the barrier too high?

If there is a qualitatively different mechanism than the one you are finding starting from a linear NEB, see if you can help it get there. Look for intermediate minima along the way through which you can put the initial path.

We don't have any scripts for automatically generating those surfaces.