two minima without transition state?

[gump_813]

Hi, is it possible that two local minima are not connected by any saddle point?

[dano]

I don't think so. If there is really no saddle try tightening the convergence criteria and see if both minima converge to a single minima.

[ross]

[quote="dano"]I don't think so. If there is really no saddle try tightening the convergence criteria and see if both minima converge to a single minima.[/quote]
Hi,dano,I also encounter such a problem, they are really minima according to the VASP, and can not converge to one. I use forces(EDIFFG<0) as the convergence criteria. Maybe things will be changed if I turn to EDIFFG>0?
best regards
Ross

[dano]

Are all atoms in the system free to move? Could be the same local minima with just a translation or rotation? This would lead to the appearance of two adjacent local minima without a transition state between the two.