H binding energy

Message

apple · #1 Post by **apple** » Tue Oct 25, 2011 2:51 am

hi,
I am looking at H atom on different metal surfaces. In one case, H is 2.14 A from the nearest surface atoms, which to me appears to be a big distance. Is it bound to the surface in spite of this large bond length? The binding energy when referenced to 0.5 *H2 is -1.2eV, which would imply it is, but can someone comment though.
Thanks.

Dr Vitaly V. G. Chaban · #2 Post by **Dr Vitaly V. G. Chaban** » Tue Oct 25, 2011 4:38 am

What algorithm do you apply to optimize the system geometry?

apple · #3 Post by **apple** » Wed Oct 26, 2011 5:09 am

I am using IALGO = 38 , EDIFF= 1E-05, EDIFFG=-1.0E-03. Why should it matter?
Thanks for reply.

Dr Vitaly V. G. Chaban · #4 Post by **Dr Vitaly V. G. Chaban** » Wed Oct 26, 2011 5:31 pm

[quote="apple"] Why should it matter?
[/quote]

This matters. If there are parts of system (in its initial POSCAR) which interact relatively weakly, the geometry optimization algorthms give an irrelevant output. In such cases, simulated annealing is preferrable.

Regarding your energy, it seems more or less reasonable to me, but I'd suggest to look through the recent literature. Again, try to reoptimize the geometry using annealing.