Hi there, On Fri, Jan 16, 2009 at 11:29 PM, Lydia Nemec <[email protected]> wrote:
> Hello! > > Thanks for your help! > Maybe it is just Friday night and I am to tired or it isn't obvious. Could > you > please give me a more detailed advice? Thanks again. > How do I include the Born charges in my Calculation? I couldn't find > something like UseSave... > You just need to have the system_label.BC file in the same directory than the system_label.FC when you run vibrator. So, instead of calculatin all the Born-charges for the supercell, you calculate them for the unit cell, then duplicate the lines for the duplicated/equivalent atoms. It should work... > As far as I understand the output, it will calculate the polarization 3 > times > for a single atom. How can I tell siesta that it does not need to calculate > the Born Charge again after moving an Atom in a new direction? > With finite differences, it has to calculate the polarization "at least 3 times". The Born-charges are defined as the derivatives of the polarization with respect to atomic positions (notice that they are not scalars, but tensors)! And I mean "at least 3 times" (one in each coordinate direction) because a more accurate result would be obtained if instead of 2 points (zero, and Delta_R) to get the derivative you have more points... This is one of the reasons why getting the Born-charges is so slow... Cheers, -- Miguel Alonso Pruneda Theory and Simulation Group @ CIN2
