neb of spontaneous reactions

Message

mhyman · #1 Post by **mhyman** » Sun Aug 14, 2005 2:26 am

Will a neb calculation of a spontaneous reaction pathway converge in VASP?

Thanks,
Matt

tjf · #2 Post by **tjf** » Thu Aug 18, 2005 2:39 pm

What exactly do you mean by "spontaneous reaction pathway"?

mhyman · #3 Post by **mhyman** » Thu Aug 18, 2005 7:34 pm

One in which there is no activation barrier. The starting and end points of the pathway were extremely close to each other. Since the first post I ran single point calculations of three geometries linearly interpolated between image 00 and 01, which reveal an energy barrier of 0.02 eV.

tjf · #4 Post by **tjf** » Fri Aug 26, 2005 8:59 am

Well, I was going to say that if there's no barrier how is the reactant defined? The only option is for a bimolecular reaction. If there's no reaction, there's no chemical species.

Have you simply tried a NEB calc on your system?

#5 Post by **admin** » Fri Sep 16, 2005 12:25 pm

If there is no reaction barrier, you will probably not be able to find
any minimum for the 'initial state', because the system relaxes to
the 'final state' immediately, as it is the ground state structure.
So you have to choose the starting geometry according to some other criterium.

The NEB method should work though, because the relaxation of the
geometries only is done on the hyper-planes normal to the reaction
path. There only has to be taken care that the entire hyper-plane
does not drift away, which can be done eg. by choosing a rather small the number of ionic steps (NSW)