From vaskarpour at yahoo.com Tue Jan 4 14:53:51 2011 From: vaskarpour at yahoo.com (Vahid Askarpour) Date: Tue, 4 Jan 2011 05:53:51 -0800 (PST) Subject: [Wannier] Wannier functions of Cmc21 phase of Hydrogen Message-ID: <677634.55373.qm@web39404.mail.mud.yahoo.com> Dear Wannier Users, In an attempt to generate the Wannier functions of the Cmc21 phase of hydrogen at 100GPa (already published in PRB 62, 15505, 2000), I relaxed the structure (with ABINIT) using LDA psp with a cut off energy of 60Ha and a 11x11x11 k-grid (convergence tests were done). The unit cell parameters are: acell (Bohr)? 3.5802517040E+00? 3.5802517040E+00? 5.6948406845E+00 rprim?? ?4.9609471480E-01? 8.6826841123E-01? 0.0000000000E+00 ????????? ?-4.9609471480E-01? 8.6826841123E-01? 0.0000000000E+00 ?????????? 0.0000000000E+00? 0.0000000000E+00? 1.0000000000E+00 xred (reduced coordinates) 2.5121514893E-01? 2.5121514893E-01? 5.7250684394E-01 8.5213973344E-02? 8.5213973344E-02? 4.2749315606E-01 7.4878485107E-01? 7.4878485107E-01? 7.2506843938E-02 9.1478602666E-01? 9.1478602666E-01? 9.2749315606E-01 The structure is insulating with Fermi level at 2.83eV. With 4 atoms in the primitive cell, two bands are fully occupied. My wannier90.win is: num_wann 2??????? !(since two bands are occupied) dis_win_max 3.5 !(only the two occupied states are in this window) num_iter 10000 dis_num_iter 10000 dis_conv_tol 1.0E-10 iprint 1 wannier_plot =true wannier_plot_format=xcrysden wannier_plot_supercell=3 Begin Projections ! (Using two of the atoms as starting point) random f= 8.5213973344E-02,? 8.5213973344E-02,? 4.2749315606E-01 : l=0 f= 7.4878485107E-01,? 7.4878485107E-01,? 7.2506843938E-02 : l=0 End Projections However,?after 10,000 iterations, I get the following in the output: ?Cycle:? 10000 ? WF centre and spread??? 1? ( -0.118897, -0.090041,? 0.151445 )??? 17.92856764 ? WF centre and spread??? 2? ( -0.023007, -0.110858,? 0.105187 )??? 19.60921540 ? Sum of centres and spreads ( -0.141904, -0.200898,? 0.256632 )??? 37.53778304 ? 10000???? 0.102E+00???? 4.0933874374?????? 37.5377830446???? 432.59? <-- CONV ? O_D=???? 21.7668345 O_OD=???? 14.1241376 O_TOT=???? 37.5377830 <-- SPRD ?Delta: O_D=? 0.1058381E+00 O_OD= -0.4324867E-02 O_TOT=? 0.1015132E+00 <-- DLTA ?------------------------------------------------------------------------------ ?Writing checkpoint file wannier90.chk... done ?Final State ? WF centre and spread??? 1? ( -0.118897, -0.090041,? 0.151445 )??? 17.92856764 ? WF centre and spread??? 2? ( -0.023007, -0.110858,? 0.105187 )??? 19.60921540 ? Sum of centres and spreads ( -0.141904, -0.200898,? 0.256632 )??? 37.53778304 ???????? Spreads (Ang^2)?????? Omega I????? =???? 1.646810960 ?===============?????? Omega D????? =??? 21.766834519 ???????????????????????????????????????? Omega OD???? =??? 14.124137566 ? Final Spread (Ang^2)?????? Omega Total? =??? 37.537783045 ????? Wannier Function Num:??? 1?????? Maximum Im/Re Ratio =?? 18.718258 ????? Wannier Function Num:??? 2?????? Maximum Im/Re Ratio =?? 51.325554 The two Wannier functions are not real and the spread is large.??Is this output reasonable or am I missing something? Is there a way to lower the spread and make the functions real? I'd appreciate your input. Regards, Vahid Askarpour Physics Department Dalhousie University Halifax, Nova Scotia Canada -------------- next part -------------- An HTML attachment was scrubbed... URL: From jonathan.yates at materials.ox.ac.uk Tue Jan 4 20:15:54 2011 From: jonathan.yates at materials.ox.ac.uk (Jonathan Yates) Date: Tue, 4 Jan 2011 19:15:54 +0000 Subject: [Wannier] Wannier functions of Cmc21 phase of Hydrogen In-Reply-To: <677634.55373.qm@web39404.mail.mud.yahoo.com> References: <677634.55373.qm@web39404.mail.mud.yahoo.com> Message-ID: On 4 Jan 2011, at 13:53, Vahid Askarpour wrote: > > > The two Wannier functions are not real and the spread is large. Is this output reasonable or am I missing something? Is there a way to lower the spread and make the functions real? Vahid, My first suggestion is to set guiding_centres=.true. in the win file. (look through the mailing list archives to see what this does - put simply it can help the minimisation) Another suggestion is to try different projections - even just two randomly located Gaussians should be ok here. Jonathan -- Department of Materials, University of Oxford, Parks Road, Oxford, OX1 3PH, UK tel: +44 (0)1865 612797 http://users.ox.ac.uk/~oums0549/ From thanusit at kku.ac.th Wed Jan 26 05:19:32 2011 From: thanusit at kku.ac.th (=?UTF-8?B?4LiY4LiZ4Li54Liq4Li04LiX4LiY4Li04LmMIOC4muC4uA==?= =?UTF-8?B?4Lij4Li04LiZ4LiX4Lij4LmM4Lib4Lij4Liw4LmC4LiE4LiZ?=) Date: Wed, 26 Jan 2011 11:19:32 +0700 (ICT) Subject: [Wannier] Band disentanglement Energy windows Message-ID: Dear Wannier90 users I'm trying to perform the band structure interpolation from Wannier functions, using Wannier90 interfaced to the Abinit package. I've found that the interpolated band structure is strongly dependent on the disentanglement energy windows (Wannier90's input parameters: dis_win_min, dis_win_max, dis_froz_min, and dis_froz_max). I'd like to ask, for a given number of Wannier functions and a numbers of bands entering the interpolation, how we theoretically select these energy windows? Kind regards, Thanusit Dr Thanusit Burinprakhon Physics Department, Faculty of Science Khon Kaen university Khon Kaen, Thailand 40002 tel 00 66 4320-2222 ext. 2248 mobile 00 66 86-8620510 From a.mostofi at imperial.ac.uk Wed Jan 26 14:56:09 2011 From: a.mostofi at imperial.ac.uk (Arash Mostofi) Date: Wed, 26 Jan 2011 13:56:09 +0000 Subject: [Wannier] Band disentanglement Energy windows In-Reply-To: References: Message-ID: <4D4027F9.6060606@imperial.ac.uk> An HTML attachment was scrubbed... URL: From cmsl10 at gmail.com Wed Jan 26 15:33:31 2011 From: cmsl10 at gmail.com (Samir Ellzardane) Date: Wed, 26 Jan 2011 15:33:31 +0100 Subject: [Wannier] K-points error for pw2wannier.x Message-ID: Dear wannier users, I'm triying to do wannier calculation of simple structurale geometry of FePO4 with four atoms in the unit cell using LDA+U formalisme. So following all the steps of wannier calculations are made without any problem until pw2wannier.x step with this error (see below) Program POST-PROC v.4.1.2 starts ... Today is 21Jan2011 at 11:29:51 Reading nscf_save data file O.pbe-rrkjus.UPF: wavefunction(s) 2S renormalized Spin CASE ( up ) Wannier mode is: standalone ----------------- *** Reading nnkp ----------------- Checking info from wannier.nnkp file - Real lattice is ok - Reciprocal lattice is ok Something wrong! numk= 60 iknum= 136 Could anyone help me, Thank you in advance Samir Ellzardane, PhD. student Physics Department, University of Guelma 24000, Algeria -------------- next part -------------- An HTML attachment was scrubbed... URL: From a.mostofi at imperial.ac.uk Wed Jan 26 15:47:49 2011 From: a.mostofi at imperial.ac.uk (Arash Mostofi) Date: Wed, 26 Jan 2011 14:47:49 +0000 Subject: [Wannier] K-points error for pw2wannier.x In-Reply-To: References: Message-ID: <4D403415.1020508@imperial.ac.uk> An HTML attachment was scrubbed... URL: From cmsl10 at gmail.com Wed Jan 26 16:06:22 2011 From: cmsl10 at gmail.com (Samir Ellzardane) Date: Wed, 26 Jan 2011 16:06:22 +0100 Subject: [Wannier] K-points error for pw2wannier.x In-Reply-To: References: <4D403415.1020508@imperial.ac.uk> Message-ID: Dear Mostofi, > > Thank you for your fast reply, > > What you say is totally right, I used the same k-point generated from the > script kmesh.pl with all calculations (scf, nscf, ...), but seems to be > not the good method to solve my problem. > > Thank you a gain, > > Samir > > > 2011/1/26 Arash Mostofi > >> Dear Samir, >> >> This error is telling you that the k-point list in the PWscf input file >> doesn't match that in the Wannier90 (.win) input file. By default, PWscf >> uses symmetry to reduce the number of k-points used in the calculation. >> Wannier90, on the other hand, knows nothing about these symmetry operations. >> For this reason, the full list of k-points should be specified in the >> 'k_points' block of the PWscf input file for the nscf calculation as well as >> in the Wannier90 input file. I suspect that this is the cause of your error. >> >> >> Because typing out k-point coordinates is tedious for dense grids, there >> is a perl script called kmesh.pl in the 'utility' directory of the >> Wannier90 distribution that outputs the k-point coordinates for a given >> Monkhorst-Pach mesh. These can then be copied to your input file(s) >> >> Best wishes, >> >> Arash >> >> -- >> Dr Arash A Mostofi >> Senior Lecturer and RCUK Fellow >> Depts of Materials & Physics >> Imperial College London >> London SW7 2AZ, United Kingdom >> >> T +44 (0)207 594 8154 >> F +44 (0)207 594 6757 >> E a.mostofi at imperial.ac.uk >> W http://www.cmth.ph.ic.ac.uk/people/a.mostofi >> >> >> On 26/01/2011 14:33, Samir Ellzardane wrote: >> >> Dear wannier users, >> >> I'm triying to do wannier calculation of simple structurale geometry of >> FePO4 with four atoms in the unit cell using LDA+U formalisme. So following >> all the steps of wannier calculations are made without any problem until >> pw2wannier.x step with this error (see below) >> >> Program POST-PROC v.4.1.2 starts ... >> Today is 21Jan2011 at 11:29:51 >> >> Reading nscf_save data >> file O.pbe-rrkjus.UPF: wavefunction(s) 2S renormalized >> >> Spin CASE ( up ) >> >> Wannier mode is: standalone >> >> ----------------- >> *** Reading nnkp >> ----------------- >> >> Checking info from wannier.nnkp file >> >> - Real lattice is ok >> - Reciprocal lattice is ok >> Something wrong! >> numk= 60 iknum= 136 >> >> Could anyone help me, >> >> Thank you in advance >> >> Samir Ellzardane, PhD. student >> Physics Department, University of Guelma 24000, Algeria >> >> >> >> _______________________________________________ >> Wannier mailing list >> Wannier at quantum-espresso.org >> http://www.democritos.it/mailman/listinfo/wannier >> >> > -------------- next part -------------- An HTML attachment was scrubbed... URL: From cmsl10 at gmail.com Wed Jan 26 18:22:05 2011 From: cmsl10 at gmail.com (Samir Ellzardane) Date: Wed, 26 Jan 2011 18:22:05 +0100 Subject: [Wannier] K-points error for pw2wannier.x In-Reply-To: <4D403415.1020508@imperial.ac.uk> References: <4D403415.1020508@imperial.ac.uk> Message-ID: Thank for this fast reply, What you say is totally right, I used the same k-point generated from the script kmesh.pl with all calculations (scf, nscf, ...), but seems to be not the good method to solve my problem. What can you suggest? Thank you again, Samir 2011/1/26 Arash Mostofi > Dear Samir, > > This error is telling you that the k-point list in the PWscf input file > doesn't match that in the Wannier90 (.win) input file. By default, PWscf > uses symmetry to reduce the number of k-points used in the calculation. > Wannier90, on the other hand, knows nothing about these symmetry operations. > For this reason, the full list of k-points should be specified in the > 'k_points' block of the PWscf input file for the nscf calculation as well as > in the Wannier90 input file. I suspect that this is the cause of your error. > > > Because typing out k-point coordinates is tedious for dense grids, there is > a perl script called kmesh.pl in the 'utility' directory of the Wannier90 > distribution that outputs the k-point coordinates for a given Monkhorst-Pach > mesh. These can then be copied to your input file(s) > > Best wishes, > > Arash > > -- > Dr Arash A Mostofi > Senior Lecturer and RCUK Fellow > Depts of Materials & Physics > Imperial College London > London SW7 2AZ, United Kingdom > > T +44 (0)207 594 8154 > F +44 (0)207 594 6757 > E a.mostofi at imperial.ac.uk > W http://www.cmth.ph.ic.ac.uk/people/a.mostofi > > > On 26/01/2011 14:33, Samir Ellzardane wrote: > > Dear wannier users, > > I'm triying to do wannier calculation of simple structurale geometry of > FePO4 with four atoms in the unit cell using LDA+U formalisme. So following > all the steps of wannier calculations are made without any problem until > pw2wannier.x step with this error (see below) > > Program POST-PROC v.4.1.2 starts ... > Today is 21Jan2011 at 11:29:51 > > Reading nscf_save data > file O.pbe-rrkjus.UPF: wavefunction(s) 2S renormalized > > Spin CASE ( up ) > > Wannier mode is: standalone > > ----------------- > *** Reading nnkp > ----------------- > > Checking info from wannier.nnkp file > > - Real lattice is ok > - Reciprocal lattice is ok > Something wrong! > numk= 60 iknum= 136 > > Could anyone help me, > > Thank you in advance > > Samir Ellzardane, PhD. student > Physics Department, University of Guelma 24000, Algeria > > > > _______________________________________________ > Wannier mailing list > Wannier at quantum-espresso.org > http://www.democritos.it/mailman/listinfo/wannier > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jonathan.yates at materials.ox.ac.uk Wed Jan 26 20:49:03 2011 From: jonathan.yates at materials.ox.ac.uk (Jonathan Yates) Date: Wed, 26 Jan 2011 19:49:03 +0000 Subject: [Wannier] K-points error for pw2wannier.x In-Reply-To: References: <4D403415.1020508@imperial.ac.uk> Message-ID: On 26 Jan 2011, at 17:22, Samir Ellzardane wrote: > Thank for this fast reply, > > What you say is totally right, I used the same k-point generated from the script kmesh.pl with all calculations (scf, nscf, ...), but seems to be not the good method to solve my problem. What can you suggest? Arash's main point is still valid. When you run the nscf calculation in pwscf you are getting 168 kpoints, while in the nnkp file from wannier90 there are 60. You just need to understand which of these number is correct, and why the wrong number appears in one case. My guess (and you haven't provided sufficient information for anyone to do more than guess) is that pwscf is generating extra kpoints - try setting nosym=.true. in the system namelist. Note that for the scf calculation you can employ symmetry and use a reduced set of kpoints. The regular grid of kpoints is only required for the nscf part. Jonathan -- Department of Materials, University of Oxford, Parks Road, Oxford, OX1 3PH, UK tel: +44 (0)1865 612797 http://users.ox.ac.uk/~oums0549/ From cmsl10 at gmail.com Thu Jan 27 09:53:26 2011 From: cmsl10 at gmail.com (Samir Ellzardane) Date: Thu, 27 Jan 2011 09:53:26 +0100 Subject: [Wannier] K-points error for pw2wannier.x In-Reply-To: References: <4D403415.1020508@imperial.ac.uk> Message-ID: Thank you very much Jonathan for this suggestion. Best regards 2011/1/26 Jonathan Yates > > On 26 Jan 2011, at 17:22, Samir Ellzardane wrote: > > > Thank for this fast reply, > > > > What you say is totally right, I used the same k-point generated from > the script kmesh.pl with all calculations (scf, nscf, ...), but seems to > be not the good method to solve my problem. What can you suggest? > > Arash's main point is still valid. When you run the nscf calculation in > pwscf you are getting 168 kpoints, while in the nnkp file from wannier90 > there are 60. > You just need to understand which of these number is correct, and why the > wrong number appears in one case. > > My guess (and you haven't provided sufficient information for anyone to do > more than guess) is that pwscf is generating extra kpoints - try setting > nosym=.true. in the system namelist. > > Note that for the scf calculation you can employ symmetry and use a > reduced set of kpoints. The regular grid of kpoints is only required for the > nscf part. > > > Jonathan > > > > -- > Department of Materials, University of Oxford, Parks Road, Oxford, OX1 3PH, > UK > tel: +44 (0)1865 612797 http://users.ox.ac.uk/~oums0549/ > > _______________________________________________ > Wannier mailing list > Wannier at quantum-espresso.org > http://www.democritos.it/mailman/listinfo/wannier > -------------- next part -------------- An HTML attachment was scrubbed... URL: