Dear Prof. Blaha, thank you for the reply.
> I don't know how exactly you got the laplacian, but my first guess > concerning the factor 2 would be that you did not consider the spin. I followed the "guide" you gave me a while back here on the mailing list: >You can use lapw0. In case.in0 specify as option 36 (instead if 13) >and put R2V instead of NR2V. > >(See vxclm2.f in SRC_lapw0) I did my (non-spinpolarized) SCF run as usual and then modified in0 to use option 36 and r2v instead of the defaults. > > In lapw0 everything is done spin-polarized, even in a non-spinpolarized > calculation and one has to add spin-up and dn laplacians to get the > total. I obtain one r2v file and in my working dir I do not see any files indicating an "up/down" output. So I followed your "old" advice >The file case.r2v contains the laplacian in the same functional form as >the density or the potential is usually stored. Thus you can use this file >instead of case.clmval and plot the Laplacian with x lapw5 >(you need to cp case.r2v case.clmval before running lapw5). and used the r2v file "as it is" to plot nabla**2. > I guess the rest is ok. The normalization of clmsum and clmval files are > different (only inside the spheres!), this explains the factor sqrt(4pi), > but the r2v file has the "clmval" normalization. Thank you for clarifying this. I will do some further tests and try to find out what is causing the deviations I observe - I do not see a mistake in my "procedure", yet. regards Georg Eickerling > > > Dr. Georg Eickerling schrieb: >> Dear WIEN users, >> >> I have a question concerning the use of LAPW5 to calculate the Laplacian >> of the Density on a grid. >> >> Using lapw0 [WIEN2k_08.1 (Release 14/12/2007)] I can obtain an r2v file >> containing the information >> about the Laplacian and copying the r2v to clmval I can use lapw5 to >> obtain grid files. >> >> I specify atomic units in the lapw5 input to keep WIEN from applying >> some conversion factors from >> e bohr-3 to e A-3 which would be useless in this case and in addition I >> specify the keyword VAL. >> >> This works all well and the plots in XCrysden look resonable but >> comparing the absolute values from WIEN >> to those obtained by experiment WIEN gives results which are too small >> by something close to a factor of 2 >> (which seems not to be "real" as the topological parameters obtained >> by WIEN's aim at the bond critical >> points are VERY close to the experimental values). >> >> In lapw5 I can specify "TOT" of "VAL" for the "normalization" which >> results in a division by a factor Sqrt(4Pi) in the >> case of using "TOT". My first (general) question is: What is the reason >> for this additional factor for the total density? >> >> Any then the second question: What would be the "proper" way to obtain >> the Laplacian in atomic units with LAPW5, >> should I use VAL or TOT or is there some other normalization/unit >> conversion I would have to do? >> >> Thanks in advance for any help! >> >> Georg Eickerling >> >> >> >> > -- ============================ Dr. Georg Eickerling Universitaet Augsburg Institut fuer Physik Lehrstuhl fuer Chemische Physik und Materialwissenschaften Universitaetsstr. 1 86159 Augsburg E-Mail: georg.eickerling at physik.uni-augsburg.de Phone: +49-821-598-3354 FAX: +49-821-598-3227 WWW: http://www.physik.uni-augsburg.de/cpm/ =====================================================