Dear Ulas, I've looked at your .out file and it seems to me that your method is mostly good, but I do have a suggestion and I hope it helps, regarding your basis set: with transition metal oxides, I have found that the inclusion of semicore states is at times absolutely essential to find the correct global energy minima. That could be your problem here. In your case I would include 4p semicore states for Y and the 4f and 5d states for Bi. For help on how to do this, take a look at this basis input for Bi with semicore states included (or other elements, or search SIESTA-L): http://www.icmab.es/leem/siesta/Databases/BasisSets/Bi/gga/ - you will need to explicitly specify the basis of all your atoms in the fdf file and also need to create a new pseudopotential files for Y and Bi, created with the atom program. For O, you can look for a basis set inside a metal oxide as well. You can try to approach a reasonable lattice constant first, then use the simplex method in siesta-xxx-/Util/Optimizer to fine tune the soft confinement parameters and cutoff radii of the basis with your half-optimized crystal structure, then relax the lattice vectors again using your optimized basis. There are ample tutorials on the simplex method, both included in SIESTA's documentation in the Optimizer directory and also elsewhere on SIESTA-L. I'm pretty sure the inclusion of semicore states with correct pseudopotentials and a well-defined optimized basis set will get you to where you want to be just fine.
Hearty cheers! On Tue, Aug 6, 2013 at 9:43 AM, Ulas Koroglu <[email protected]> wrote: > Dear Siesta Users and Developers, > > I try to optimize BiYO3 crystal that has the cubic structure. While > experimental value of lattice parameter is a=5.484 Angstrom, but I found > a=4.095 Angstrom during the geometry optimization in Siesta 3.1. > There is a big different between the two (experimental and theoretical) > values. I used many ways to correct the theoretical result ( a=4.095 > Angstrom) but, I didn't achieve. I did not understand what problem is. If > you can suggest me a solution, I would be very appreciated. > > I have attached the input and output files. > > Best regards. > > Ulas Koroglu. > > > -- *Abraham Hmiel* Katherine Belz Groves Fellow in Nanoscience Xue Group, College of Nanoscale Science and Engineering at SUNY Albany http://abehmiel.net/about
