I think it should be possible to just omit the lattice constant and vector specifiation, then SIESTA chooses the box size automatically so that you end up with an isolated system calculation (smallest box to exclude direct image interactions +10%, if I'm not mistaken).
Mind, however, that if your system has, e.g., lagre electrostatic moments, you should pay attention to the effect that the box size has on your results. 2008/2/7, [EMAIL PROTECTED] <[EMAIL PROTECTED]>: > > > I am trying to calculate the band structure for 20 or so Ni atoms. > > However, > > Siesta repeats the cluster I specify and I end up with calculations that > > has > > more than 500 atoms. Is there anyway, I can stop the periodic > repetition? > > Choose the unit cell size large enough, so that the basis functions > of outmost atoms won't overlap with those of a repeated cell. > The maximal extension is given (with a not very intuitevely clear title) > as something like > Vna: Cut-off radius for the neutral-atom potential: 7.514923 > (in Bohr units). > > If this done right, you'll then see the message > siesta: System type = molecule > in the output file. > > Good luck > > Andrei Postnikov >
