How big is your supercell ? How many atoms ?? I expect it is much too small ! Use a supercell as large as possible on your computer, but definitely it should be at least 64 - 256 atoms.
If Be has a very different volume compared to the host, it needs space and this requires that the Cu atoms can move away slightly over a long range. Am 07.12.2010 13:58, schrieb Amlan Ray: > Dear Prof. Blaha, > Thank you for your reply regarding the valence s-electron density plot. > Actually I am studying the cases of implantation of 7Be in aluminum and > copper lattices. We have formed > supercells (using both neighbors and 2nd neighbors and the force > minimization) with the 7Be at octahedral positions and done the calculations. > We found that the total electron > density at 7Be nucleus increases by about 0.5% when 7Be is implanted in the > smaller copper lattice. The result is about the same whether both 1s and 2s > orbitals of 7Be are treated > as valence electrons or 1s as core and 2s as valence. (The experimental > result is in the opposite direction. It shows the decay rate is faster in > aluminum by (0.18+-0.13)%). > I think the predicted increase of the electron density when 7Be is implanted > in copper is because of the substantially smaller spatial dimension of the > copper lattice. If I run > the code the using 1s orbital of 7Be as core and 2s as valence, then I find > that the 2s electron density at 7Be nucleus increases by about 26% when 7Be > is in copper. In order to > probe further, I plotted the total valence electron charge density of 7Be as > a function of the distance from 7Be nucleus (please see the attached .pdf > plot). Although in the > attached plot, I am showing the total valence charge density, the bump around > 0.25 BU appears to be the first bump of 2s electron orbital. The second > larger bump of 2s orbital > apparently got mixed up with other valence orbitals and I did not show it in > the attached plot. The dash-dot line is for 7Be in Cu and the solid line is > for 7Be in Al. So it looks > like that we are getting 26% higher 2s electron density when 7Be is in > copper, because there is more (about 26%) 2s electrons in 7Be when implanted > in copper. However the smaller > spatial dimension of copper has not pushed the 2s orbital closer to the > nucleus. The 2s orbital seems to be at the right place. I have also looked at > the density of states versus > energy plot for the valence s-electrons of 7Be. I find that for 7Be in Al, > the energy of the s-electrons lie from about (--3 eV to +8.4 eV(Fermi > energy)) and for 7Be in Cu, the > energy of the s-electrons lies from (-1.3 eV to +10.8 eV (Fermi energy)). The > total number of valence s-electrons of 7Be is more when 7Be is in Al. > I am somewhat puzzled to find that as the energy of valence s-electrons of > 7Be in Cu is increasing considerably (compared to implantation in Al) and the > total number of valence > s-electrons summed up to Fermi energy is decreasing, the number of electrons > in the characteristic first bump of 2s orbital is increasing, thus increasing > the 2s electron density > at the nucleus. I was expecting that the copper having higher electron > affinity would take away s-electrons from 7Be reducing the number of 2s > electrons and the corresponding > charge density at the nucleus as found experimentally. > With best regards > Amlan Ray > Address > Amlan Ray > Variable Energy Cyclotron Center > 1/AF, Bidhan Nagar > Kolkata - 700064 > India > > > > > _______________________________________________ > Wien mailing list > Wien at zeus.theochem.tuwien.ac.at > http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien -- P.Blaha -------------------------------------------------------------------------- Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna Phone: +43-1-58801-15671 FAX: +43-1-58801-15698 Email: blaha at theochem.tuwien.ac.at WWW: http://info.tuwien.ac.at/theochem/ --------------------------------------------------------------------------