Hello Sau Nguyen, There is a tool developed by LEV KANTOROVICH that can generate a Fortran code in which you specified the symmetry constrains you would like to use inside SIESTA. Using this approach, you have to compile SIESTA for each of your symmetry constrains (take a look at the SIESTA manual).
The link to the Lev page is: http://www.cmmp.ucl.ac.uk/~lev/ The link to the tool to fixing the symmetries: http://www.cmmp.ucl.ac.uk/~lev/codes/lev00/index.html Also, there are other ways to obtain the crystal structure for your system and it is using other tools developed specifically to obtain the atoms positions by ab initio methods. All these tools have the built in capability to constrain you crystallographic system. Some of these tools are: Endeavour ($$): http://www.crystalimpact.com/endeavour/download.htm Fox (free): http://vincefn.net/Fox/ Expo2009 (free): http://wwwba.ic.cnr.it/content/expo2009 Polymorph Predictor ($$$$$),part of the Material Studio package: http://accelrys.com/products/materials-studio/ Best regards, Camps On Mon, Jan 2, 2012 at 12:35 AM, Nguyen Doan Sau <[email protected]> wrote: > Hi Siesta users, > I need some help on the following problem. > I made two new compounds, one I solved crystal structure and the other > not. However, based on its synthesis, XRD, indexing, IR, and several > physical characterization methods, I may confirm that they are > iso-structural. Of course, I can do Rietveld refinement to get the crystal > structure of the latter by using the initial crystal structure of the > first. The problem is the crystal structure I solve involved 204 atoms in > unit cell (or 68 atoms in a asymmetric unit) and most of them is atom > having atomic number less than 10. As you know, the lab XRD or even > synchrotron X-ray diffraction in this case is not much help. I could not > have chance to get neutron data as its schedule is full. > The Rietveld refinement diverged when I turn on the refinement of oxygen > atoms. One of the reasons is the initial structure is far from the true > one. I definitely need to somehow obtain the initial crystal stru cture for > the refinement which is as close as the real structure. > I intend to obtain it by using the crystal structure relaxation in quantum > espresso code or siesta code. As you know, I could use some constrain to > keep some of special positions of atoms fixed while allowing the others > move. However, for atoms related by the space group symmetry, I want to > utilize the advances. For example, atom A in Wyckoff position (0,y,z) > related > to its mirror atom B > (1/2,-y,z+1/2)<http://www.cryst.ehu.es/cgi-bin/cryst/programs/find_comp_op?ita=31&xc=1/2&yc=-y&zc=zpp1/2&orgpos=1/2,-y,zpp1/2&standard=1>, > whenever, the code move A I would expect the code also move B in the way > that keep this relation. Based on my experience, the quantum espresso and > siesta codes don't have this implementa tion. One of the codes I know is > lmto ultilize the space group symmetry but don't allow structure relaxation > So I would ask the users in the community some ideas or any code that > allow me to do the structural relaxation that keep the crystal space group. > I would be very appreciated to it. > Best wishes, > Sau Nguyen, > U of H > >
