Re: [SIESTA-L] WARNING: Qtot, Tr[D*S] differ
Dear Eduardo, I was using siesta1.3p with no patches, hence the error (I found a 2004 post by Pablo Ordejon mentioning this problem). I switched to siesta-2.0 and it is gone. Thank you for your assistance. -Salvador. On Thu, 10 Apr 2008, Eduardo Anglada wrote: Hi Salvador, Could you post your fdf and pseudo? (if you prefer you send them directly to me) Best, Eduardo On 09/04/2008, at 21:06, Salvador Barraza-Lopez wrote: Hello list, I am doing a calculation where a gold surface is covered with a magnetic molecule, via sulfur (thiol). I first performed a calculation with a gold pseudo and basis including the 5d electrons. My basis was built explicitly and had 2zetas for the 6s, and 1 zeta for 6p and 5s channels. The band structure I get is very similar to that from plane-wave codes. Then, I built a pseudo without the 5d channel. I built my basis explicitly once again. For reasons I do not completely understand I am getting the following warning: siesta:1 -208997.1679 -208235.0221 -208235.0221 1.2912 -5.5563 timer: Routine,Calls,Time,% = IterSCF1 107338.054 20.77 siesta: WARNING: Qtot, Tr[D*S] = 870.00 859.492781 There is something not quite right. Could you share your experience on the problem, and perhaps point out what could be wrong with my basis? Thank you, -Salvador.
[SIESTA-L] Pseudo list not working today
Hello, I wish to use the Co pseudo from the new pseudo database. Somehow it does not seem to be working today: WARNING: The requested object does not exist on this server. The link you followed is either outdated, inaccurate. Please, excuse the inconvenience. This is the url: http://www.uam.es/departamentos/ciencias/fismateriac/siesta/Pseudos_GGA_Abinit/Co_html/Co.html Would it be possible to know when the pseudos list be back up? Thank you in advance, -Salvador.
[SIESTA-L] Electron occupancies in SIESTA
Hello SIESTA users, I would like to know if there is a way to obtain the electronic occupancies for a given eigenstate in SIESTA. I know one can determine this by the location of the Fermi energy but still would like to know if there is another way to obtain occupancies for electronic temperatures greater than zero (or of the order of an energy gap). Your comments will be greatly appreciated. With seasonal greetings, -Salvador.
[SIESTA-L] PDOS
Dear Marcos, David, Thank you for pointing out the meaning of the z-flag on the PDOS file. Yet, have you actually plotted the PDOS for spin-polarized systems at some point? I created a small script to plot the PDOS for a very simple system: A Mn dimer. I know the eigenvalues are very close to those obtained from a trusted plane wave code. The occupations as obtained from the EIG file and the location of the fermi levels from standard output are also consistent with the results from that plane wave code. If things are as both of you indicate, I should get, for the first column of the data the up-spin states at particular energies, and the corresponding down spin entries from the second column. Instead, what I have is a second column filled with zeroes (not a single peak when plotted, see red line in attached figure), and a first column that actually resembles the *sum* of spins (blue line; I am attaching a figure for clarity, hope it is okay). I know there are issues with non-ortonomal basis sets and that it is valid to get negative dos from those but, really, it looks like what I am getting is not the spin-up component but rather the sum. I am looking into the pdosg.F file to try to fix it (line 206). Can not determine how to do it momentarily. Comments and help on this matter will be greatly appreciated. Best, Salvador.attachment: Mn2.tif
[SIESTA-L] PDOS file
Hello, I am working with a spin-polarized system and I would like to plot the Projected density of states (PDOS). Looking in the manual, not much info can be found. I have two questions: z=1 is for spin-up electrons and z=2 for spin-down electrons? Also, the actual PDOS data has two columns, but the second one appears to only have zeroes. Could somebody please discuss why this is so? What is the second column for? Here's one example: orbital index=7 atom_index=1 species=Mn position= -2.473701 0.00 0.00 n=4 l=1 m= 0 z=1 data 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.01270 0.0 . . . Thank you in advance for your comments and clarification. -Salvador.
[SIESTA-L] Principal number channels on atom utility
Hi, I am trying to build a pseudopotential file for Mn within GGA-PBE. I have seen Miguel Pruneda's input file which is defined to include 3s, 3p, 3d and 4f channels. The pseudo is polarized (by leaving out two electrons on the 4s channel, which the pseudo program can not generate if s channels are already defined (the 3s channel). Here is one such sample pseudopotential generation file: pe Manganese tm2 1.70 n=Mn c=pbr 0.0 0.0 0.0 0.0 0.0 0.0 34 30 1.00 1.00 31 4.00 2.00 32 5.00 0.00 43 0.00 0.00 1.00 1.00 1.00 1.00 0.00 0.3 Looking at the OUT file after running atm, I have found . . . Mn pseudopotential generation: Troullier-Martins nlseigenvalue rc cdrc delta Core radius ( 1.00) outside wfn extremum ( 0.74) 1s 0.5 -7.6068910.9935450.7088600.303294 Core radius ( 1.00) outside wfn extremum ( 0.75) 2p -0.5 -5.3807480.9935450.6548401.634114 Core radius ( 1.00) outside wfn extremum ( 0.76) 2p 0.5 -5.2879350.9935450.6453831.604942 Core radius ( 1.00) outside wfn extremum ( 0.76) 3d -0.5 -1.7936330.9935450.4868733.054698 Core radius ( 1.00) outside wfn extremum ( 0.76) 3d 0.5 -1.7845600.9935450.4840283.047217 4f -0.5 -0.2776930.9935450.35 -2.589683 4f 0.5 -0.2776900.9935450.35 -2.600246 . . . I wonder why the n channel is running from 1 to 4 in those lines. All the rest in the OUT file looks normal. Have anyone else seen this? I guess it is just an innofensive bug in the program when it writes these things to the OUT file, without repercusions for the actual generated pseudos, but would like to be certain that is the case. Thank you in advance, -Salvador
Re: [SIESTA-L] atom label ordering problem
Hello Chu, The issue here is when you change the ordering on the Chamicalspecies block, you are changing the atomic species also in your coordinate set. Did you leave the coordinates unchanged? For instance, if an atom with coordinates X Y Z 1 was assigned to species 1, then by renaming the atomicspecies block this same atom now belongs to the complimentary specie (originally, to species 2). So you really need to change the atomicspecies block, and the labels on your coordinate set, which assign the atomic species to an atom at position XYZ. Best, Salvador. On Tue, 21 Nov 2006, Chu Chun FU wrote: Dear All, We are studying the FeCu system with Siesta and we have found some problems related to the labels (1 or 2) asigned to each chemical species. For example when considering the system containing 1 Fe atom and N-1 atoms of Cu with BCC structure (e.g. N = 54), the results (total energy and magnetic moment of Fe) differ depending on which atom is labeled as atom '1': If we put %block ChemicalSpeciesLabel 1 29 Cu 2 26 Fe %endblock ChemicalSpeciesLabel we obtain reasonable values for total energy and magnetic moment for the Fe atom (about 3 mu_B). However when we change the order and put %block ChemicalSpeciesLabel 1 26 Fe 2 29 Cu %endblock ChemicalSpeciesLabel we obtain a different value of total energy (0.6 eV higher) and zero magnetic moment for Fe ! We wonder if anybody had a similar problem before ? or any idea on the origing of this problem ? Thank you in advance for your information and help Chu Chun FU
Re: [SIESTA-L] About unoccupied channels in the basis set
Hi Marcos, I read some time ago a post by Nichols Romero on this subject. From that post, and the responses on it, it appeared that there is not a further implementation of the 'pseudo generation code' for two channels with the same angular momentum. Regards, Salvador.
Re: [SIESTA-L] Erbium
Hi Simone, As previously mentioned, your rcut are too small. I bet you are seeing a *linear* decrease on the total energy against lattice constant with your pseudo, aren't you? I've seen this behaviour myself whenever the rcuts were too small. Try the suggestion by Oleksandr. How abot all rcuts being set to 2.0 or somethig like that? Another thing you might want to fix is your radius for core correction: It is beyond your rcuts. usually this radii has to be as close to zero as you can, without making the pseudo too hard. There is a PRB paper by Marc Pederson, at NRL, on this. I've tried this with Siesta and too-small rcuts were in ocasions responsible for ghost states. So proceed with caution in that respect. Best, Salvador On Sat, 14 Oct 2006, Simone Sanna wrote: Dear all, I'm trying to generate a pseudopotential for Erbium with the code atom. The Erbium electronic configuration is: [Xe]4f12 6s2 I'd like to have the 5s and 5p orbitals as semicore and the 4f, 5d, 6s and 6p as valence, just like in the Ba example in the atom manual. So after choosing the various r_c like suggested by Mr. Salvador Barraza-Lopez in a recent contribution in this mailing list, and adding core correction, my input for atom looks like: %define OLD_CC pe Er tm2 3.0 Er car 0.0 0.0 0.0 0.0 0.0 0.0 94 50 1.00 1.00 51 3.00 3.00 52 0.00 0.00 43 7.00 5.00 1.11 1.11 1.22 0.50 1.00 1.50 The tests for the pseudopotential seem to be OK (to get a qualitative idea of the pseudos, have a look at the attachment). As I have semicore electrons I have to specify a PAO Basis. I use for Erbium the basis suggested by Valeria Miguel for Lanthanium (see the following lines or the SIESTA website), which includes 5s and 5p as semicore (simple zeta). The 6s and 5d orbitals are described with double zeta, and a simple-zeta 6p is used as polarization. Er 5 2.23376 n=5 0 1 E75.06727 3.90515 4.47365 1.0 n=6 0 2 E71.19116 3.69937 5.35635 4.45480 1.0 1.0 n=5 1 1 E 9.59098 0.01349 4.68524 1.0 n=6 1 1 E87.28316 3.65766 5.49527 1.0 n=5 2 1 E14.20169 4.25960 5.28396 1.0 Now, with these values I do not get any minimum in the total energy curve for bulk erbium (HCP structure) for different values of the lattice constant!!! This is not due to the deficit of LDA in simulating strong correlated electrons, because at least for the structure LDA is good enough as shown in the past. Can someone suggest me a better choice for the parameters and/or explain me why does my pseudopotential fail to reproduce metallic erbium? Do I make some mistake in the choice of the r_c or is the basis inadequate to correctly reproduce the erbium orbitals? Tanky you in advance for your time, Simone Sanna P.S The HGH pseudos downloadable for ABINIT are O.K. Is it possible to convert them in a format radable by SIESTA?
Re: [SIESTA-L] input_phenol
Hi Cornil, May I inquire how you obtained your value for PAO.EnergyShift? The problem originates from the overlap matrix. Choleski diagonalization requires S to be strictly equal to its transpose. I've seen this problem happening to me a couple times. My advice would be for you to plot your total energy against PAO.EnergyShift (yes, there will be *some* values of EnergyShift for which the run will be successful I believe). There is an extra variable indicating a magnitude for the split between orbitals, I can't remeber it off hand. Try this other variable too. One extra observation. Have you made tests of total energy against Meshcutoff? Your value is on the low limit. Best, Salvador. On Fri, 6 Oct 2006, cornil david wrote: Dear Alberto, Here is the complete input of my run -- SystemName molecule phenol SystemLabel phenol NumberOfSpecies 3 NumberOfAtoms 13 %block ChemicalSpeciesLabel 16 C 21 H 38 O %endblock ChemicalSpeciesLabel AtomicCoordinatesFormat NotScaledCartesianAng ZM.ForceTolLen 0.04 eV/Ang ZM.ForceTolAng 0.0001 eV/deg ZM.MaxDisplLen 0.1 Ang ZM.MaxDisplAng 20.0 deg %block Zmatrix molecule 1 0 0 0 0.000.00 0.00 0 0 0 !C 1 1 0 0 1.3951600.00 0.00 1 0 0 !C 1 2 1 0 1.394712 120.008632 0.00 1 1 0 !C 1 3 2 1 1.395427 119.994165-0.056843 1 1 1 !C 1 4 3 2 1.394825 119.993992 0.034114 1 1 1 !C 1 1 2 3 1.394829 119.998457 0.032348 1 1 1 !C 2 1 2 3 1.099610 119.997223 -179.972926 1 1 1 !H 2 3 2 1 1.099680 120.012795 179.961852 1 1 1 !H 2 4 3 2 1.099680 119.981142 -179.996436 1 1 1 !H 2 5 4 3 1.099761 120.011343 -179.999514 1 1 1 !H 2 6 1 2 1.099604 120.007997 179.989175 1 1 1 !H 3 2 1 6 1.43 119.980770 179.953248 1 1 1 !O 2 12 2 1 0.96 109.50-0.079100 1 1 1 !H %endblock Zmatrix ZM.UnitsLength Ang ZM.UnitsAngle deg xc.functional GGA xc.authors PBE MD.TypeOfRun CG MD.NumCGsteps 200 MeshCutoff 120 Ry PAO.EnergyShift 0.01 Ry DM.NumberPulay 5 DM.MixingWeight 0.3 - Découvrez un nouveau moyen de poser toutes vos questions quel que soit le sujet ! Yahoo! Questions/Réponses pour partager vos connaissances, vos opinions et vos expériences. Cliquez ici.
Re: [SIESTA-L] pseudopotential problem
of the exchange-correlation functional. So that's why I hoped if I download that particular pseudopotential for Bismuth, the things would work. I guess I could try with LDA. The problem is not LDA or GGA, but then one word of caution: don't mix LDA and GGA pseudopotentials, which will not give you good results (I'd say good results would simply be a coincidence in this case). If you decide that your pseudos are GGA, then they should all be GGA, or else all LDA. Cheers, Marcos Thank you very much. Regards Bozo -- -- Salvador Barraza-Lopez Department of Physics Virginia Tech Blacksburg VA, 24060 [EMAIL PROTECTED] (540) 231 3308
Re: [SIESTA-L] Gold band structure
Dear Oleksandr, Adam, Lucas, Thank you very much for your input. I knew the things I needed to look into were simple. It turn out the problem had to do with my slighlty-off definition of the W point. Thank you Lucas for the fdf file. I have to say the distribution function proposed makes calculations really fast for metals. I'll keep using it from now on. I did not find big changes with the fineness of the mesh nor the number of k-points although I was more concerned at the moment with qualitative agreement and now I'll look more carefully into convergence studies. Thank you for your help, Salva. * Salvador Barraza-Lopez 3223 BI University of Illinois +1(217) 244 1964 (Office) https://netfiles.uiuc.edu/barrazal/www/ *
[SIESTA-L] Gold band structure
Dear SIESTA users, I have been working on creating a pseudopotential for gold. I started checking postings and found a couple dating back to 2003. The pseudo there is LDA. I have tried a good number of parameters. I've taken matching radii close to the rmax in the all-electron calculations; have tried neutral and ionized configurations, core corrections, relativistic corrections as well as pseudos without: I think I have explored in a consistent form the options I can get from the pseudo. The problem I have is the following: The band structure looks reasonable, but I can *not* reproduce a degeneracy at the W-point for the second and third bands. The bands at the Gamma-K direction also look different from other published work and abinit test calculations. I also see the lack of splitting in a recent siesta calculation (Fig 6 on J. Phys. Cond. Matter 18 (2006), 7999. I feel I have tried what I can think of, but haven't had success. Am I doing a simple procedural mistake? Atached is a copy of my siesta input file for an LDA, spin unpolarized, non-relativistc pseudo; the one present in the listings from 2003. I have a database of pseudos, but all present the aforementioned splitting for the second and third band at the W point. As I am running out of ideas, I am eager to hear your comments and suggestions. Best regards, Salvador Barraza - SystemName Bulk gold, simplex run SystemLabel gold4p18 #SpinPolarized True NumberofSpecies 1 NumberofAtoms 1 LatticeConstant 4.07 Ang #XC.functional GGA #XC.authorsPBE ElectronicTemperature 300 K SolutionMethod Diagon Diag.DivideAndConquer True #UseSaveData True %block ChemicalSpeciesLabel 1 79 Au %endblock ChemicalSpeciesLabel PAO.Energyshift 0.002 Ry PAO.splitnorm 0.5 #PAO.Basis basis.fdf %block LatticeVectors 0.0 0.5 0.5 0.5 0.0 0.5 0.5 0.5 0.0 %endblock LatticeVectors %block AtomicCoordinatesAndAtomicSpecies 0.0 0.0 0.0 1 %endblock AtomicCoordinatesAndAtomicSpecies %block kgrid_Monkhorst_Pack 4 0 0 0.0 0 4 0 0.0 0 0 4 0.0 %endblock kgrid_Monkhorst_Pack %block BandLines 1 0.000.000.00 #Gamma 1600.000.002.00 #X 120 0.0 0.9153048 2.00 #W 1201.001.001.00 #L 1202.002.002.00 #Gamma 1601.501.500.00 #K %endblock BandLines MeshCutoff 200 Ry MaxSCFIterations 100 DM.NumberPulay 3 DM.MixingWeight 0.2 * Salvador Barraza-Lopez 3223 BI University of Illinois +1(217) 244 1964 (Office) https://netfiles.uiuc.edu/barrazal/www/ *
[SIESTA-L] Charge 'density(?)' from atom program
Dear developers and users, I have a couple questions regarding the charge plot from an all-electron calculation (or pseudo, does not matter) in atom. The first one is about the units in this plot. The horizontal scale is in Bohr but what would be the units for the vertical plot? Perhaps related to the previous question is: How is one to interpret the charge plot? It is the radial part I suppose, so it is not a density. (I could see this as being the density only for the case of closed shell atoms.) Waiting for your comments on this matter, Salvador.
Re: [SIESTA-L] Not converged ?
Hi, The variables you want to modify are DM.MixingWeight and DM.NumberPulay. If you can't make the run to converge, then make use of DM.NumberKick and DM.KickMixingWeight. Increasing the temperature might also help. Please have a careful look at the manual for descriptions and default values of those variables. Best regards, Salvador. * Salvador Barraza-Lopez 3223 BI University of Illinois +1(217) 244 1964 (Office) https://netfiles.uiuc.edu/barrazal/www/ *
Re: [SIESTA-L] [SIESTA-] Nanotubes cordinates-broadening
The DOS and PDOS are calculated independently from the self-consistent runs. What you need to do is to increase the number of k-points in your Monkhorst-Pack mesh and concurrently to increase sampling points and to decrease the width of your Gaussan smearing. You can stop when you're happy with what you get; This a limiting procedure, there's nothing to it. Please also look at the archive; there is an extensive discussion on this point. Regards, * Salvador Barraza-Lopez 3223 BI University of Illinois +1(217) 244 1964 (Office) https://netfiles.uiuc.edu/barrazal/www/ * ---BeginMessage--- Hello Roberto Veiga Thnx for send the web but my problem is this one I observed one thing that how can we find the width of the Gaussian broadening for Nanotubes bcos wen I chnage Gaussian broadening (0.02 to 0.20 eV) the DOS shape is t chnaged, specially the DOS at fermi level. Any Would you like to let me know how can we choose this parameter for Metal, Insulator, or Super conductor. Thnx Roberto Veiga [EMAIL PROTECTED] wrote: Dear Michael: about generating coordinates of carbon nanotubes, maybe you like to see my Internet-based applet (http://k.1asphost.com/tubeasp/tubeasp.asp). It generates both XYZ coordinates and a sample SIESTA input for carbon nanotubes of any chirality and number of unit cells, and also radially deformed. Regards, Roberto Veiga 2006/6/26, Michael Shin [EMAIL PROTECTED]: Thank you much navaratnarajah kuganathan I solved that problem now, but I observed one thing that how can we find the width of the Gaussian broadening for Nanotubes bcos wen I chnage Gaussian broadening (0.02 to 0.20 eV) the DOS shape is totally chnaged, specially the DOS at fermi level. Any Would you like to let me know how can we choose this parameter for Metl, Insulator, or Super conductor. Thnx navaratnarajah kuganathan [EMAIL PROTECTED] wrote: Dear Michael, I recomend to use two programmes for generating coordinates for carbon nanotubes.They are tube gen and wrapping.These can be downloaded from onlone.Addresses are http://turin.nss.udel.edu/research/tubegenonline.html and www.photon.t.u-tokyo.ac.jp/~maruyama/wrapping3/wrapping.html once you have the xyz or pdb files for nanotube you can visualise the structure of nanotubes in Crystal maker or Chemcraft.Using Chemcraft you can add atoms or chains. regards Kugan Michael Shin [EMAIL PROTECTED] wrote: Hello Users I read many artcicle about tranition metal(TM) or other element filled Nanotubes i.e [EMAIL PROTECTED] I know there are some softwares which can generate the atomic cordinates of Nanotube. I dont know how we can find the atmic cordinates for Fe filled Nanotubes, bcos Fe may be in BCC,FCC,Nanowires or in Atomic form. Is there any one who can help me in finding the cordinates of Fe or any other TM filled Nanotubes. I will be very thnkful for your help. Thnx Mich. - Talk is cheap. Use Yahoo! Messenger to make PC-to-Phone calls. Great rates starting at 1?min. - Yahoo! Messenger with Voice. PC-to-Phone calls for ridiculously low rates. - Do you Yahoo!? Next-gen email? Have it all with the all-new Yahoo! Mail Beta. - How low will we go? Check out Yahoo! Messengers low PC-to-Phone call rates.---End Message---
Re: [SIESTA-L] Questions on OrderN procedures
Dear Professor Ordejon, I will be looking forward for the next release. Thank you very much for your comments. Salvador Barraza. * Salvador Barraza-Lopez University of Illinois-UC 244 1964 (Office) 367 6756 (Home) *
[SIESTA-L] Questions on OrderN procedures
Dear SIESTA-developers/users I have explored the mailing list trying to obtain info regarding orderN procedures. But there is not much. Or to be precise, there is not enough for me to make a sucessful run. I have unstabilities after the second scf cicle. The gradient goes out of bounds. The fdf file in this (test) case works well with serial procedures. I am planning to build more complex systems and so I am using it for the purpose of attempting a sucessful run with the orderN algoritms. Thank you in advance for your advice Salvador Barraza == a) From output (I killed the process after the NaN values): == ... cgwf: iter = 602 grad = -0.000223 Eb(Ry) = -53.047003 cgwf: CG tolerance reached denmat: qtot (before DM normalization) = 93.9687 ordern: qtot (after DM normalization) = 96. siesta: Mulliken populations before mixing siesta: Mulliken populations after mixing siesta: Program's energy decomposition (eV): siesta: Eions = 7587.227555 siesta: Ena = 2368.195644 siesta: Ekin= 2228.689450 siesta: Enl = 235.059723 siesta: DEna=55.124991 siesta: DUscf = 0.966519 siesta: DUext = 0.00 siesta: Exc = -1109.533849 siesta: eta*DQ = 0.00 siesta: Emadel = 0.00 siesta: Eharris = -3723.679942 siesta: Etot= -3808.725076 siesta: FreeEng = -3808.725076 siesta: iscf =1 Eharris(eV) = -3723.6799 E_KS(eV) = -3808.7251 dDmax = 1.1037 timer: Routine,Calls,Time,% = IterSCF1 126.450 95.53 ordern: enum = 96. cgwf: iter =1 grad = -3.963003 Eb(Ry) = -42.044669 cgwf: iter =2 grad = -7.651252 Eb(Ry) = -42.547216 cgwf: iter =3 grad = -6.410114 Eb(Ry) = -42.720717 cgwf: iter =4 grad = -6.130575 Eb(Ry) = -42.838365 cgwf: iter =5 grad = -12.796178 Eb(Ry) = -42.997075 cgwf: iter =6 grad = -31.618497 Eb(Ry) = -43.353862 cgwf: iter =7 grad =-486.199456 Eb(Ry) = -44.444660 cgwf: iter =8 grad = -3798.219004 Eb(Ry) = -45.550972 cgwf: iter =9 grad = -48542.759773 Eb(Ry) = -48.997373 cgwf: iter = 10 grad = -275198.725463 Eb(Ry) = -52.861871 cgwf: iter = 11 grad =-4052458.136844 Eb(Ry) = -71.158725 cgwf: iter = 12 grad = -19499084.321046 Eb(Ry) = -87.201877 cgwf: iter = 13 grad = -416979082.612115 Eb(Ry) =-204.200298 cgwf: iter = 14 grad = -534715421.381929 Eb(Ry) = 493.020553 cgwf: iter = 15 grad = 122.512375 Eb(Ry) = -42.522687 cgwf: iter = 16 grad = -60.740570 Eb(Ry) = -42.534267 cgwf: iter = 17 grad = -19.437825 Eb(Ry) = -43.240224 cgwf: iter = 18 grad = -15.937511 Eb(Ry) = -43.565640 cgwf: iter = 19 grad = -18.396973 Eb(Ry) = -43.888326 cgwf: iter = 20 grad = -29.662959 Eb(Ry) = -44.311314 cgwf: iter = 21 grad =-437.021080 Eb(Ry) = -41.159444 cgwf: iter = 22 grad = -88.720173 Eb(Ry) = -43.746057 cgwf: iter = 23 grad = -80.426508 Eb(Ry) = -42.606757 cgwf: iter = 24 grad = -90.914064 Eb(Ry) = -43.377340 cgwf: iter = 25 grad = -75.478465 Eb(Ry) = -42.693587 cgwf: iter = 26 grad = -81.858472 Eb(Ry) = -43.354430 cgwf: iter = 27 grad = -91.267118 Eb(Ry) = -42.669000 cgwf: iter = 28 grad = -64.573457 Eb(Ry) = -43.365626 cgwf: iter = 29 grad =-104.885818 Eb(Ry) = -42.649302 cgwf: iter = 30 grad = -58.856621 Eb(Ry) = -43.352277 cgwf: iter = 31 grad =-184.142282 Eb(Ry) = -43.508836 cgwf: iter = 32 grad =-116.921952 Eb(Ry) = -42.537545 cgwf: iter = 33 grad = -38.830883 Eb(Ry) = -43.440271 cgwf: iter = 34 grad = -60.524839 Eb(Ry) = -43.929710 cgwf: iter = 35 grad = -2693.914372 Eb(Ry) = -37.714792 cgwf: iter = 36 grad = -5723.611318 Eb(Ry) = -37.099561 cgwf: iter = 37 grad =-1432567.548338 Eb(Ry) = -73.863247 cgwf: iter = 38 grad =-5893823.284952 Eb(Ry) = -81.978386 cgwf: iter = 39 grad =-968.847406 Eb(Ry) = -42.252302 cgwf: iter = 40 grad =-453.056769 Eb(Ry) = -42.293956 cgwf: iter = 41 grad = -72.135214 Eb(Ry) = -41.775811 cgwf: iter = 42 grad = -23594.568210 Eb(Ry) = -46.605841 cgwf: iter = 43