Douglas, Probably irrelevant in your case, but I ran into this once myself: some codes can use a shifted k-point grid by default unless you specify otherwise, for atoms this would result in higher energy, increasing your absolute cohesive energy. I'm not sure what is the default behavior in SIESTA (now I always make sure to set explicitly a nonshifted grid), but I recall there being some kind of bug in SIESTA 2.something where it would spontaneosly introduce a shift.
In summary, you should maybe take a closer look at your atom calculations, and, of course, check the convergence of your cohesive energy (basis, k-points, integration grid). If everything is OK with these, then it probably is the fault of the functional (by the way, are you using GGA or LDA?). 2010/5/27 <[email protected]> > Hello, my name is Douglas Martins and I am a doctoral student at > UFJF-Brazil. > I'm starting to work with the SIESTA to calculate the cohesive energy of > the alloy Cu3Au, > I did the calculation for the league and then to the Au and Cu atoms alone, > then > subtracts the energy of the alloy Cu3Au the energies of Au and three times > the energy Cu. > But I got a value > 4.86 ev, but experimentally this value is 3.64 ev. I wonder if anyone knows > how to make > that calculation of cohesion energy correctly or if there is better value > in this energy > arquivo.out output. If someone can get me well as they should be the input > parameters for > the league and the atoms apart, I would greatly appreciate it. > A big hug to everyone and I await the return. > > Graciously > > Douglas > > -- > Esta mensagem foi enviada pelo CEFET-Mail > www.cefetmg.br > > > > -- > Sistema de Antivirus do CEFET-MG: > Esta mensagem foi verificada pelo sistema de antivirus e > acredita-se estar livre de perigo. > >
