Dear Rob, note that the doubly negative oxygen ion does not have bound states. The orbital energies are no longer negative for some of the occupied orbitals and the large r90% and r99% radii indicate that the orbitals are no longer localized on the atom. Also note the large number of warnings in the listing.
Likewise, the -0.8 state (test 10) has this problem, or is a border line case. The main difference of relativistic and non-relativistic calculations with atom is the treatment of electron spin, which you can also see from the output files. You may also want to do spin-polarized calcuations, which may be more comparable to the relativistic calculations. Regards, Ulrich. ----- Original Message ----- From: Koch, Robert J [mailto:[email protected]] To: [email protected] [mailto:[email protected]] Sent: Tue, 07 Jun 2011 22:24:55 +0200 Subject: [SIESTA-L] Oxygen All Electron Calculations > Hello all, > > I've been having some trouble with some ATOM all-electron calculations which > I'm trying to use to test my oxygen pseudopotential. > > I have attached two all electron input files for oxygen, identical except > that one is a relativistic calculation and the other is not. > > My first surprise was that for two of the reference states, the excitation > energy is largely different between a relativistic and non-relativistic > caclulation. In one case the difference is on the order of 20 Rydberg, and > in the other, as much as 70 Ry: > > > relativistic case: > &d total energy differences in series > > &d 1 2 3 4 5 6 7 > 8 9 > > &d total energy differences in series > > &d 10 11 12 > &d 1 0.0000 > &d 2 0.2193 0.0000 > &d 3 0.4386 0.2194 0.0000 > &d 4 0.6581 0.4389 0.2195 0.0000 > &d 5 0.8777 0.6584 0.4391 0.2196 0.0000 > &d 6 1.0974 0.8782 0.6588 0.4393 0.2197 0.0000 > &d 7 0.7591 0.5399 0.3205 0.1010 -0.1186 -0.3383 0.0000 > &d 8 0.4602 0.2409 0.0215 -0.1979 -0.4175 -0.6372 -0.2990 > 0.0000 > &d 9 0.1980 -0.0213 -0.2407 -0.4601 -0.6797 -0.8994 -0.5612 > -0.2622 0.0000 > &d 10 -0.4010 -0.6202 -0.8396 -1.0591 -1.2787 -1.4984 -1.1601 > -0.8611 -0.5989 0.0000 > &d 11 73.4604 73.2411 73.0217 72.8022 72.5826 72.3629 72.7012 > 73.0002 73.2624 73.8613 0.0000 > &d 12 1.0923 0.8730 0.6536 0.4342 0.2146 -0.0051 0.3331 > 0.6321 0.8943 1.4933 -72.3681 0.0000 > > > > > non-relativistic case: > > &d total energy differences in series > > &d 1 2 3 4 5 6 7 > 8 9 > > &d total energy differences in series > > &d 10 11 12 > &d 1 0.0000 > &d 2 0.2187 0.0000 > &d 3 0.4376 0.2189 0.0000 > &d 4 0.6566 0.4378 0.2190 0.0000 > &d 5 0.8756 0.6569 0.4380 0.2191 0.0000 > &d 6 1.0948 0.8761 0.6572 0.4383 0.2192 0.0000 > &d 7 0.7570 0.5383 0.3194 0.1004 -0.1186 -0.3378 0.0000 > &d 8 0.4585 0.2398 0.0209 -0.1981 -0.4172 -0.6363 -0.2985 > 0.0000 > &d 9 0.1967 -0.0220 -0.2409 -0.4598 -0.6789 -0.8981 -0.5603 > -0.2617 0.0000 > &d 10 51.9073 51.6886 51.4697 51.2508 51.0317 50.8125 51.1503 > 51.4489 51.7106 0.0000 > &d 11 53.9140 53.6953 53.4764 53.2575 53.0384 52.8192 53.1570 > 53.4555 53.7173 2.0067 0.0000 > &d 12 1.0928 0.8740 0.6552 0.4362 0.2171 -0.0020 0.3358 > 0.6343 0.8960 -50.8146 -52.8212 0.0000 > > > The Reference states are as follows: > ATM3 7-JUN-11 O Test1 -- GS 2s2 2p4 > &v&d > ATM3 7-JUN-11 O Test2 -- 2s1.8 2p4.2 > &v&d > ATM3 7-JUN-11 O Test3 -- 2s1.6 2p4.4 > &v&d > ATM3 7-JUN-11 O Test4 -- 2s1.4 2p4.6 > &v&d > ATM3 7-JUN-11 O Test5 -- 2s1.2 2p4.8 > &v&d > ATM3 7-JUN-11 O Test6 -- 2s1.0 2p5.0 > &v&d > ATM3 7-JUN-11 O Test7 -- 2s1.2 2p5.0 > &v&d > ATM3 7-JUN-11 O Test8 -- 2s1.4 2p5.0 > &v&d > ATM3 7-JUN-11 O Test9 -- 2s1.6 2p5.0 > &v&d > ATM3 7-JUN-11 O Test10 -- 2s1.8 2p5.0 > &v&d > ATM3 7-JUN-11 O Test11 -- 2s2.0 2p6.0 > &v&d > ATM3 7-JUN-11 O Test12 -- 2s2 2p3 3d1 > &v&d > > > > > Should I really see such a large difference between all electron > relativistic calculations and non-relativistic calculations, particularly in > an atom as light as oxygen? > > My second suprise was that in both the all-electron relativistic > calculations and non-relativistic calculations, the energy difference > between the oxygen ground state (2s2 2p4) and the fully ionized state (2s2 > 2p6) is extrememely large: over 50 Rydberg. > > Has anyone else expereinced this? I thought the point of these all electron > calculations provide accurate comparisons for electronic configurations > which were probable for the atom. > > Could someone on the list kindly run these input files through ATOM and > email me the output? My thoughts are that it is a possible compilation > issue. > > If anyone could offer any other advice I would also greatly appreciate it. > > Thanks, > Rob Koch > BS, Ceramic Engineering > Alfred University > --- Max-Planck-Institut für Eisenforschung GmbH Max-Planck-Straße 1 D-40237 Düsseldorf Handelsregister B 2533 Amtsgericht Düsseldorf Geschäftsführung Prof. Dr. Jörg Neugebauer Prof. Dr. Dierk Raabe Prof. Dr. Martin Stratmann Dipl.-Kfm. Herbert Wilk Ust.-Id.-Nr.: DE 11 93 58 514 Steuernummer: 105 5891 1000 -------------------------------------------------
