Dear Elisa,
No optimizer can guarantee to find the global minimum. If there is an energy
barrier between the ones you find with different methods, then the optimization
will find the one closest (i.e. on the same side of the barrier) one to the
starting structure.
If they are very close and there is no barrier between them, it's either not
fully converged, or there is an error (in the input or the program). Note that
the default for MD.NumCGsteps (if you use the CG optimizer; I don't remember the
ones for other methods) is a bit low, so check first if it converged.
Cheers,
Herbert
Elisa wrote:
Dear Siesta users,
I have optimized a solid made by Aluminum nanoclusters. First I relaxed the
atomic coordinates (MD.VaribleCell . false) and, after that, I relaxed the
lattice together with the atomic coordinates. Then I obtained an optimized
solid.
The problem is that I did some single point calculations changing the lattice
constant in order to obtain the Energy vs Volume graphic. However, I realized
that some other structures with different lattice constants are lower in
energy than my optimized solid.
Do someone know what can be the reason? What I'm doing wrong?
I appreciate any comment in advance !!
Thanks,
Elisa Jimenez Izal
Kimika Fakultatea
Euskal Herriko Unibertsitatea - University of The Basque Country
P.K. 1072, 20080 Donostia
Euskal Herria
E-mail: [email protected]
Phone: +34-943-015341
***********************************************************************
--
Herbert Fruchtl
Senior Scientific Computing Officer
School of Chemistry, School of Mathematics and Statistics
University of St Andrews
--
The University of St Andrews is a charity registered in Scotland:
No SC013532
