<div>Dear all,<BR>As francesco d'acapito said (Tue Oct 5 11:17:28 CEST 2004 )<BR>>Having a glance to the case.in1 files? I've realized that there is apparently<BR>>no regular choice made by lstart in the linearization energy of local orbitals <BR>>(LO). Running the TiO2 case in the examples I've found (file TIO2_in1):<BR>--------------------------TiO2.in1-------------------------------<BR>WFFIL (WFPRI, SUPWF)<BR> 7.00 10 4 (R-MT*K-MAX; MAX L IN WF, V-NMT<BR> 0.30 5 0 (GLOBAL E-PARAMETER WITH n OTHER CHOICES, global APW/LAPW)<BR> 0 0.30 0.000 CONT 1<BR> 0 -4.35 0.005 STOP 1<BR> 1 -2.58 0.010 CONT 1<BR> 1
0.30 0.000 CONT 1<BR> 2 0.30 0.010 CONT 1<BR> 0.30 3 0 (GLOBAL E-PARAMETER WITH n OTHER CHOICES, global APW/LAPW)<BR> 0 -1.55 0.010 CONT 1<BR> 0 0.30 0.000 CONT 1<BR> 1 0.30 0.000 CONT 1<BR>K-VECTORS FROM UNIT:4 -7.0 1.5 emin/emax window<BR>-------------------------------------------------------------------<BR>Dr. Blaha said that(Tue Oct 5 17:27:59 CEST 2004 )<BR>>One gets (nearly) identical results when one sets <BR>>the 3p as LO and the 4p as "APW", or the other way around.<BR>and <BR>>WIEN has some energy criterium and selects the "more important" state as APW,<BR>>and the "less imp!
ortant"
one as LO. In the TiO2 example the O-2s and Ti 3p are<BR>>considered "more important" than their 3s/4p partners, while Ti 4s is more <BR>>important than Ti 3s (which is almost a core state).<BR>That means El_2 maybe higher than El_2 for the same l. But when I did:<BR>run_lapw -in1new 1<BR>I found:<BR>--------------------------TiO2.in1-------------------------------<BR>WFFIL (WFPRI, SUPWF)<BR> 7.00 10 4 (R-MT*K-MAX; MAX L IN WF, V-NMT<BR> .08694 6 0 global e-param with N other choices, napw<BR> 0 0.098 0.000 CONT 1<BR> 0 -3.756 0.002 CONT 1<BR> 1 0.169 0.000 CONT 1<BR> 1 -2.028 0.000 CONT 1<BR> 2
0.135 0.000 CONT 1<BR> 2 -0.855 0.000 CONT 1<BR> .08694 3 0 global e-param with N other choices, napw<BR> 0 0.164 0.000 CONT 1<BR> 0 -0.871 0.000 CONT 1<BR> 1 0.185 0.000 CONT 1<BR>K-VECTORS FROM UNIT:4 -7.0 1.5 emin/emax window<BR>-------------------------------------------------------------------<BR>El_2 is lower than El_1 for each l.</div> <div>Does it exists different energy criterium to select El, which are <BR>different for the parameter "-in1new" and without "-in1new". If I want to<BR>and an LO(El_2) manually, should I put it up of down to the El_1?</div> <div>The manual points out that: APWs are usually used for the d/f electrons.<BR!
>But
there are only APWs(for s,p and d electrons) in the case. <BR>Dose the program think the APWs are so important in this case?</div> <div>Since "-in1new" is so good a method, why not to use "-in1new 1" each time?<BR>Does it waste much time during calculations?</div> <div>I also calculate Fe slab of 7 layers(use WIEN2k_06). I use <BR>"runsp_lapw -in1new 1 ...", <BR>and then do "mini ...". At last I get a converged result.<BR>the :FER is 0.10514 Ryd.<BR>The *.in1 shows:<BR>----------------------------*.in1-------------------------------<BR>WFFIL (WFPRI, SUPWF)<BR> 7.00 10 4 (R-MT*K-MAX; MAX L IN WF, V-NMT<BR> -.19139 5 0 global e-param with N other choices, napw<BR> 0 -0.161 0.000 CONT 1<BR> 0 -6.162 0.000 CONT 1<BR> 1
-0.059 0.000 CONT 1<BR> 1 -3.672 0.000 CONT 1<BR> 2 0.019 0.000 CONT 1<BR> -.19139 5 0 global e-param with N other choices, napw<BR> 0 -0.233 0.000 CONT 1<BR> 0 -6.166 0.000 CONT 1<BR> 1 -0.078 0.000 CONT 1<BR> 1 -3.677 0.000 CONT 1<BR> 2 -0.014 0.000 CONT 1<BR> -.19139 4 0 global e-param with N other choices, napw<BR> 0 -0.240 0.000 CONT 1<BR> 1 -0.092 0.000 CONT 1<BR> 1 -3.856 0.000 CONT 1<BR> 2
-0.037 0.000 CONT 1<BR> -.19139 4 0 global e-param with N other choices, napw<BR> 0 -0.243 0.000 CONT 1<BR> 1 -0.105 0.000 CONT 1<BR> 1 -3.848 0.000 CONT 1<BR> 2 -0.021 0.000 CONT 1<BR>K-VECTORS FROM UNIT:4 -8.0 1.5 emin/emax window<BR>---------------------------------------------------------------<BR>I think the energy "-.19139" is OK(a bit lower than fermi energy)!<BR>The first atom is surface atom and the last one is the bulk atom.<BR>Why didn't it find the energy of 3s(semi core) in the last two Fe atoms,<BR>though they had appeared in the first two atoms! should I add them manually?</div> <div>Take the last atom as an example, if I make the program !
to search
Es(-0.243)<BR>automatically, it will search from Es(-0.243). But I just want to know <BR>the energy bound, to which the program stop searching?</div> <div>Another question about the source of main.f in SRC_kgen.<BR>in the line 151(WIEN2K_06), </div> <div>---------------------------------------------------------------<BR> READ(20,1511) LATTIC,NAT,ipgr<BR>---------------------------------------------------------------</div> <div>and in the line 171,</div> <div>---------------------------------------------------------------<BR> 1511 FORMAT(/,A4,23X,I3,/,30x,a4)<BR>---------------------------------------------------------------</div> <div>"20" stands for the file "case.struct". e.g. : TiO2.struct</div>
<div>-------------------------TiO2.struct----------------------------<BR>TiO2 <BR>P LATTICE,NONEQUIV.ATOMS: 2136_P42/mnm <BR>MODE OF CALC=RELA unit=ang <BR> 8.681405 8.681405 5.591702 90.000000 90.000000 90.000000<BR> ..............................<BR>---------------------------------------------------------------</div> <div>I kown variable LATTIC="P ",<BR> NAT=" 2",<BR>but what!
is the
value of ipgr? Why does need so many spaces "30x"?<BR>(Sorry,silly questions,I have just been learning Fortran for several days)</div> <div>Excuse me that maybe most of the questions are so silly, and several <BR>of them had been discussed many times.<BR>Hope that it won't waste your time!<BR>Best regards,<BR>Samspon Niu</div><p>
<hr size=1> Yahoo! Mail<br>
<a href="http://pa.yahoo.com/*http://us.rd.yahoo.com/evt=38867/*http://photomail.mail.yahoo.com">Use Photomail</a> to share photos without annoying attachments.