An approximate explanation, but first a caveat: dealing with small RMT's in Wien2k is not completely trivial.
In Wien2k you have two different basis sets, a radial grid + orbitals inside the RMT, plane waves outside. If you use different RMTs for the Fe in effect you make them different. For instance, try setting up a Fe cell with two different RMTs (i.e. a supercell), converge it then do aim. You will find that there has been a small charge transfer between the Fe's, which will go to zero as you increase RKMAX. For your problem, I would go beyond setrmt (which is OK, but not always optimal) and use a smaller RMT for the C (maybe 1.3) and increase that for the Fe. You will have to do this by hand. Be careful about going too small with the C as if you do you will lose core charge (although .lcore helps); you also need to be careful about losing core electrons from the Fe. In both cases look in case.outputm to see how far the core integrals are from integers. Scale RKMAX so RKMAX/min(RMT) is roughly the same (e.g. 3.5), although you need to add a bit when you decrease the RMTs. Probably something like 5-6 should be good for your problem. How to chose the "right" RMTs? I don't have any great wisdom, others might have. I tend to use RMTCheck (look in the mailing list, or I can ask Peter to add it to the contributed software) and check that the step in the gradients at the RMT for the different atoms are comparable. In most cases this gives results similar to setrmt. In principle you can also do calculations for different values with RKMAX/min(RMT) fixed and look for the minimum energy -- the few times I've done this the results are about the same. Another method would be to try and balance the charge lost in the core integrals, which gives a similar result. 2011/4/21 Maxim Rakitin <rms85 at physics.susu.ac.ru>: > Thank you Evgenia, > > I've already seen that link, but I'd like to get information why it is not > possible from physical point of view. > > Maxim > > 21.04.2011 16:53, ??????? ???????? ?????: > > Dear Maxim, > > identical elements should have identical RMT's. > This link might help you http://www.wien2k.at/reg_user/faq/rmt.html . > > 2011/4/21 Maxim Rakitin <rms85 at physics.susu.ac.ru> >> >> Dear WIEN2k community, >> >> Is it correct to perform calculations with different Rmt radii for atoms >> of one type, e.g. Fe? >> >> I need to calculate total energy of a systems containing one interstitial >> carbon atom in bcc iron matrix. Carbon has quite large MT-radius comparing >> with H atom, so StructGen sets Rmt(Fe) to 1.59 a.u. and Rmt(C) to 1.41 a.u. >> for lattice constant 5.394 a.u. In some papers (C. Domain et al., PRB 69, >> 144112 (2004)) Rmt(C) is even large than I got (~1.6 a.u.) and Rmt(Fe) is >> also much larger (~2.4 a.u.). But if I use the radii from the article, >> WIEN2k doesn't allow me to perform further procedures - nn gives me >> information about overlapping spheres. >> >> Can I use Rmt(Fe)=2.4 a.u. in all volume of my supercell except of the >> nearest iron atoms? >> >> Thanks in advance for your answers. >> >> -- >> Best regards, >> ? Maxim Rakitin >> ? South Ural State University >> ? Chelyabinsk, Russia >> ? email: rms85 at physics.susu.ac.ru >> ? web: http://www.susu.ac.ru >> >> _______________________________________________ >> Wien mailing list >> Wien at zeus.theochem.tuwien.ac.at >> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien > > > > -- > Evgeniya Kabliman. > Insitute of Materials Chemistry? ? ? ?? Vienna University of Technology > Getreidemarkt 9/165-TC, A-1060 Vienna, Austria > Tel: +43 1 58801 15674? ? ? ? Fax: +43 1 58801 15698 > evgeniya at theochem.tuwien.ac.at? ? ? ? ? ? http://www.imc.tuwien.ac.at > > _______________________________________________ > Wien mailing list > Wien at zeus.theochem.tuwien.ac.at > http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien > > _______________________________________________ > Wien mailing list > Wien at zeus.theochem.tuwien.ac.at > http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien > > -- Laurence Marks Department of Materials Science and Engineering MSE Rm 2036 Cook Hall 2220 N Campus Drive Northwestern University Evanston, IL 60208, USA Tel: (847) 491-3996 Fax: (847) 491-7820 email: L-marks at northwestern dot edu Web: www.numis.northwestern.edu Chair, Commission on Electron Crystallography of IUCR www.numis.northwestern.edu/ Research is to see what everybody else has seen, and to think what nobody else has thought Albert Szent-Gyorgi