It's all in section 7.5 of the manual, but here are some answers: > could anybody please tell me whether SIESTA is able to relax > atomic and magnetic structures simultaneously and self-consistently.
Definitely so. > For example, if one wants to find out whether a non-collinear > ordering takes place in the magnetic system, should one just > specify NonCollinearSpin T and relax the system? It's the simplest possibility. I guess Siesta will do the calculation in the usual way: setting the maximum possible spin on each atom and then self-consistency does the job. Perhaps you should also set SpinPolarized to true; note that non-collinear spin could not work in parallel. I had problems when I tried to run a prosaic O2 molecule using non-colinear spin. Worked fine when I ran it serial. > Or one should also specify the initial guess for the orientation of spins > by specifying DM.InitSpin ? In the latter case do the specified > orientation (and magnitude) of the moments on atoms change during > relaxation or they remain fixed? This is another possibility, more sophisticated, but a bit more boring to prepare. As far as the manual says, in Siesta you can only fix the total spin of the system, not the spin of individual atoms. Perhaps it would be possible to fix the spin of individual atoms, but probably you'd have to hack into the code and implement it yourself, for the moment. Best regards, Marcos -- Dr. Marcos Verissimo Alves Post-Doctoral Fellow Unité de Physico-Chimie et de Physique des Matériaux (PCPM) Université Catholique de Louvain 1 Place Croix du Sud, B-1348 Louvain-la-Neuve Belgique ------ I have become so addicted to vi that I try to exit OpenOffice by typing :wq!
