Many thanks to Riccardo Rurali and Emilio Artacho for their enlightening remarks about calculations with net charge!
Now that this topic is under discussion, I wonder (if anybody has experience) whether we should check the convergency vs. cell size only when comparing DIFFERENTLY charged systems, or as well when we compare different cases regarding to the SAME net charge and number of atoms? Similarly: can we neglect the warning of not-SC, BCC, FCC and skip caring about the Madelung energy when comparing identically charged systems in identical cells? It seems that the Madelung correction from the background charge will be the same for these cases, independently on what we are doing with the molecule inside the cell, but the I may have overlooked something. Thank you ion advance, Andrei Postnikov +-- Dr. habil. Andrei Postnikov ----- Tel. +49-541-969.2377 -- Fax .2351 ---+ | Institut fuer Festkoerperforschung - FZ Juelich, D-52425 Juelich, Germany | +-- [EMAIL PROTECTED] --------- http://www.home.uni-osnabrueck.de/apostnik/ --+ On Sat, 7 May 2005, Riccardo Rurali wrote: | Dear all, | | just a few minor remarks about the on-going discussion on charged systems. | | As far as I know, the neutralizing background charge is implemented, | regardless if the system is in cluster or supercell geometry. | As you know, this artificial background (needed so that the Coulomb energy | does not diverge) provokes a spourios interaction that is normally | (partially) corrected with the Madelung term. | | Now, what is true is that if your system is not a cluster the Madelung | correction is not calculated, even though the simulation box is of one of the | allowed type (SC, BCC or FCC). However, the reason is simply that, in the | case of a solid, the Madelung term needs to be scaled with the dielectric | constant and Siesta does not know it. | What I did in the past was commenting the appropiate IFs in the code in such a | way that the Madelung term is always calculated, but not summed to the total | energy. When the calculations was over, I read in output the value of the | Madelung term, scaled manually with the (experimental) dielectric constant | and add it to the total energy. | | Even so, if I got your point, you still have a problem: if your cell is not | SC, BCC or FCC you don't have the required alpha Madelung constant... | | A good idea could be calculating a large supercell. The obvious drawback with | respect to the option of a large cluster is that in this case the number of | atoms grows a lot (probably too much). In the limit of very large cell, the | amount of the Madelung correction (that you cannot calculate) becomes | negligible. | | Best Regards, | Riccardo | | -- | Riccardo Rurali | Laboratoire Collisions, Agrégats, Réactivité, IRSAMC | Université Paul Sabatier | 118 route de Narbonne | 31062 Toulouse | France | | e-mail: [EMAIL PROTECTED] | tel. +33 (0)5 61556071 | fax. +33 (0)5 61558317 | |
