[Wien] Electron density at the nucleus (Electron capture nuclear decay rate work)
I have been reading all the messages about the electron density at the Be nucleus under compression and would like to say a few things. My background is in experimental nuclear physics and I am very interested to undertsand quantitatively the results of electron capture experiments in compressed material. WIEN2K is probably the best availabale code at this time for this purpose. Given my background, please excuse me if I make any incorrect statements. I shall be grateful if you would kindly point out my mistakes. ? 1) Let me start with the Physics justification for thinking why Be 1s wave function should satisfy boundary conditions at the muffintin radius RMT(Be). As I understand, in this model, 1s electrons are seeing scf-potential of the crystal only within the Be sphere. Outside the Be sphere, it should see the potential of the interstitial region. Since there is an abrupt change of potential at the muffintin radius RMT(Be), so the wave function inside and outside the Be sphere should be different and there should be a matching boundary condition at RMT(Be). If we assume that outside the Be sphere, the 1s wave function should be that of a free Be ion, then it should be matched with the core wave function inside the Be sphere at RMT(Be). As a gross oversimplification, I suggested that the 1s wave function outside RMT(Be) might be taken as zero, because I thought that would be relatively easy to implement.(But I agree it was a?wrong boundary condition.)??However ?my main point is that the core wave function inside and outside the Be sphere should be different and there should be boundary conditions at RMT(Be). ? 2) I think whether compression would delocalize 1s wave function?should depend on the boundary condition applied. If the only boundary condition is that the core wave function would be zero at infinity, then of course, it will delocalize under compression. But probably there should be boundary conditions at RMT(Be). ? 3) I certainly agree that the tail of 1s wave function would experience more attractive potential when BeO is compressed. But I think that would affect the core wave function outside the Be sphere. It is not clear to me how that would affect the core wave function inside the Be sphere, particularly near the nucleus. The potential inside and outside the Be sphere is different and the wave functions should, in general, be different with a matching boundary condition at RMT(Be). ? 4) I certainly agree that the?contraction of 2s orbital would drive 1s orbital into expansion. But the reduction of 1s electron density at the nucleus is essentially independent of the muffintin radius used. I have done calculations of normal and compressed BeO cases keeping RMT(Be) the same in both the cases and have also done calculations by reducing RMT(Be) for the compressed case only. The change of 1s electron density at the nucleus remains the same always. The change of valence electrons in Be sphere is only 0.01 electrons and I can vary this number by adjusting RMT(Be). But that did not affect the change of 1s electron density at the nucleus. s-valence electrons in Be sphere can be made smaller?for the compressed case by adjusting RMT(Be), but still the result did not change. So I think that the effect of 2s orbital contraction on 1s electron density at the nucleus is probably very small. ? 5) I know about three experiments (done by different people) where the increase of electron capture rate by nuclei under compression?was seen and the effect is much more than expected from valence electrons. ? With best regards Amlan Ray Address Variable Energy Cyclotron Center 1/AF, Bidhan Nagar Kolkata - 700064 India -- next part -- An HTML attachment was scrubbed... URL: http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20100423/1bc4d143/attachment.htm
[Wien] Electron density at the nucleus (Electron capture nuclear decay rate work)
The construction of atomic spheres with a certain RMT is only a mathematical trick to obtain nicely represented wave functions and potentials in a convenient way. Of course there is a weak dependency of results on RMT, because series expansions converge better or worse with different RMTs, but there's no physics in it. RMT(Be). As I understand, in this model, 1s electrons are seeing scf-potential of the crystal only within the Be sphere. Outside the Be sphere, it should see the potential of the interstitial region. Since there is an abrupt change of potential at the muffintin radius RMT(Be), so the wave function inside and outside the Be sphere should be different and there should be a matching boundary condition at RMT(Be). No, the 1s electron sees the (spherical) potential not only inside RMT, but the potential is continued outside with a 1/r tail. (There is only ONE 1s wavefunction on a radial grid reaching to infinity.) Of course one can discuss this approximation, but as you have shown yourself, treating the 1s state as valence, where it sees the accurate non-spherical potential everywhere, does NOT change anything qualitatively (there is a limited basis set for the Be-s functions when you include 1s, but that does not matter for this purpose). However my main point is that the core wave function inside and outside the Be sphere should be different and there should be boundary conditions at RMT(Be). From the above it should be clear, that there is only ONE 1s function. For a core state, however, we make the approximation that the core-density outside the sphere is added as a constant smeared out over the whole interstitial. Also this is an approximation (and the code gives WARNINGS if the core leakage is too large), but again, your test with 1s as valence (where this is not done) proves that there is no real problem. PS: In the next release it will be possible to Fourieranalyze the leaking core density and get a correct charge distribution even with sizable core-leakage. -- - Peter Blaha Inst. Materials Chemistry, TU Vienna Getreidemarkt 9, A-1060 Vienna, Austria Tel: +43-1-5880115671 Fax: +43-1-5880115698 email: pblaha at theochem.tuwien.ac.at -
[Wien] LMO for LSMO
Why don't you create the struct file with w2web ??? Mn1NPT= 781 R0=0.0001 RMT=2.1000 Z: 25.0 O 1NPT= 781 R0=0.0001 RMT=1.5600 Z: 8.0 La1NPT= 781 R0=0.0005 RMT=3.6200 Z: 57.0 We have put lots of effort into it to select meaningful parameters. One cannot use arbitraryly chosen sphere radii, even the R0 values are not usable. Either you use the tools we provide, or you must carefully read the UG, the faq pages, S.Cotteniers book, the advises in the mailing list,... and follow them. Start a fresh case and create the struct file according to w2web. Lukasz Plucinski schrieb: Dear WIEN2k Experts, My goal is to calculate LSMO, so I try ideal LMO first. Initialization and first SCF cycle goes fine, but then I have error in SELECT in the second cycle. I reduced mixing to 0.01 and error still happens: FORTRAN STOP LAPW0 END FORTRAN STOP LAPW1 END FORTRAN STOP LAPW1 END FORTRAN STOP LAPW2 END FORTRAN STOP LAPW2 END FORTRAN STOP CORE END FORTRAN STOP CORE END FORTRAN STOP MIXER END in cycle 2ETEST: 0 CTEST: 0 FORTRAN STOP LAPW0 END FORTRAN STOP SELECT - Error stop error This is with standard PBE GGA, 500 k-points (7x7x7), and with spin-polarized calculations. I didn't try GGA+U yet. Error files and struct file attached. Same happens without spin-polarized but only in 6th SCF cycle -- error files for non-spin-polarized attached as ZIP. Could you please advise how to preceed ? Regards, Lukasz ___ Wien mailing list Wien at zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien -- - Peter Blaha Inst. Materials Chemistry, TU Vienna Getreidemarkt 9, A-1060 Vienna, Austria Tel: +43-1-5880115671 Fax: +43-1-5880115698 email: pblaha at theochem.tuwien.ac.at -
