Re: [SIESTA-L] band-structure of supercell
Dear Andrei & Ricardo, Thanks for your response. When using a say 2x2 supercell the only thing that suppose to change is the level of degeneracy. Take the following argument : if the band lines are changed for a double cell, than one needs to converge the number of unit cells to get the right band structure. this is of course untrue and a Si crystal, say, would still have the same bandstructure if we double the number of atoms in the unit cell by taking a 2x1 super-cell. Therefore, the bandstructure of 2x2 graphene cell should look exactly the same as a 1x1 cell, given that the high symmetry points are scaled to the size of the BZ of a supercell. Since \Gamma point is not subjected to any scaling issues, the eigenvalues at this point of BZ should remain the same and only the degeneracy should increase. Cheers, Doron. On 5/4/2010 5:09 PM, Doron Naveh wrote: Dear Ricardo, Thank you! Actually, I get exactly the same results so the problem lies in understanding these results. I checked only the difference between highest occupied eigenvalue and lowest unoccupied one. Thanks very much to you and to Andrei P. - and sorry for the sloppiness.. BTW, my parents didn't have the periodic table on their mined when they named me ... :-) Cheers, Doron. On 5/4/2010 4:51 PM, Ricardo Faccio wrote: Dear Boron Pure graphene has no gap, maybe you're referring to the difference between eigenvalues for particular points in the 1BZ. I'm sending some results for graphene 1x1 and supercell 2x2 (path:/gamma, M, K, /gamma). Take a look at /Gamma, where eigenvalues at -3.2 eV and -7.3 eV "keep" their energy position. The others correspond to the folding into gamma as was commented by Andrei P. Maybe there are some issues with you convergence. What parameters did you use? Regards Ricardo - Dr. Ricardo Faccio Prof. Adjunto de Física Mail: Cryssmat-Lab., Cátedra de Física, DETEMA Facultad de Química, Universidad de la República Av. Gral. Flores 2124, C.C. 1157 C.P. 11800, Montevideo, Uruguay. E-mail: rfac...@fq.edu.uy Phone: 598 2 924 98 59 598 2 929 06 48 Fax:598 2 9241906 Web: http://cryssmat.fq.edu.uy/ricardo/ricardo.htm - - Original Message - From: "Doron Naveh" To: Sent: Tuesday, May 04, 2010 5:07 PM Subject: Re: [SIESTA-L] band-structure of supercell Hi Ricardo, Thanks for the prompt response. I know about the BZ folding, this is the reason why I took the BZ center as an example to make my point. The band gap at \Gamma decreases dramatically (factor of ~ 4) when I use the 2x2 cell. Since the \Gamma point is not subjected to folding issues, is it possible that there is a bug that translate the BZ center? Thanks, -D. On 5/4/2010 3:47 PM, Ricardo Faccio wrote: Dear Doron Remember when you´re working with supercells the 1st Brillouin zone become altered, particularlly reduced. So the orignal band structure, based on the original 1st BZ, is then folding in the new one. So, what you are seeing is a mixture of bandlines. In particular for graphene 2x2 supercell diract point is still present at (1/3, 1/3, 0), but you can check how Dirac point moves into Gamma point when supercells are multiple of three (eg: 3x3 and, 6x6 supercells). Best regards Ricardo - Dr. Ricardo Faccio Prof. Adjunto de Física Mail: Cryssmat-Lab., Cátedra de Física, DETEMA Facultad de Química, Universidad de la República Av. Gral. Flores 2124, C.C. 1157 C.P. 11800, Montevideo, Uruguay. E-mail: rfac...@fq.edu.uy Phone: 598 2 924 98 59 598 2 929 06 48 Fax:598 2 9241906 Web: http://cryssmat.fq.edu.uy/ricardo/ricardo.htm - ----- Original Message - From: "Doron Naveh" To: Sent: Tuesday, May 04, 2010 4:34 PM Subject: [SIESTA-L] band-structure of supercell Hi there siesters, I ran into a very basic problem. I'm trying to calculate the bandstructure of a slab. For that matter I an illustrate my problem on 2x2 graphene unit cell. The \Gamma point of the 2x2 cell gives wrong eigenvalues compared to the 1x1 cell (a factor of ~4!!) . Is there a rigid shift of the BZ center ? Here are the lattice vectors/atom position blocks for the 2x2 cell. for the primitive cell atom position is 0,0,0.5; 1/3,1/3,0.5 and the lattice parameter reduced by half. Thanks, -D. LatticeConstant 4.9200 Ang %block LatticeVectors 0.5 0.866025404 0.0 -0.5 0.866025404 0.0 0.0 0.0 3.0 %endblock LatticeVectors AtomicCoordinatesFormat Fractional %b
Re: [SIESTA-L] band-structure of supercell
Dear Ricardo, Thank you! Actually, I get exactly the same results so the problem lies in understanding these results. I checked only the difference between highest occupied eigenvalue and lowest unoccupied one. Thanks very much to you and to Andrei P. - and sorry for the sloppiness.. BTW, my parents didn't have the periodic table on their mined when they named me ... :-) Cheers, Doron. On 5/4/2010 4:51 PM, Ricardo Faccio wrote: Dear Boron Pure graphene has no gap, maybe you're referring to the difference between eigenvalues for particular points in the 1BZ. I'm sending some results for graphene 1x1 and supercell 2x2 (path:/gamma, M, K, /gamma). Take a look at /Gamma, where eigenvalues at -3.2 eV and -7.3 eV "keep" their energy position. The others correspond to the folding into gamma as was commented by Andrei P. Maybe there are some issues with you convergence. What parameters did you use? Regards Ricardo - Dr. Ricardo Faccio Prof. Adjunto de Física Mail: Cryssmat-Lab., Cátedra de Física, DETEMA Facultad de Química, Universidad de la República Av. Gral. Flores 2124, C.C. 1157 C.P. 11800, Montevideo, Uruguay. E-mail: rfac...@fq.edu.uy Phone: 598 2 924 98 59 598 2 929 06 48 Fax:598 2 9241906 Web: http://cryssmat.fq.edu.uy/ricardo/ricardo.htm - - Original Message - From: "Doron Naveh" To: Sent: Tuesday, May 04, 2010 5:07 PM Subject: Re: [SIESTA-L] band-structure of supercell Hi Ricardo, Thanks for the prompt response. I know about the BZ folding, this is the reason why I took the BZ center as an example to make my point. The band gap at \Gamma decreases dramatically (factor of ~ 4) when I use the 2x2 cell. Since the \Gamma point is not subjected to folding issues, is it possible that there is a bug that translate the BZ center? Thanks, -D. On 5/4/2010 3:47 PM, Ricardo Faccio wrote: Dear Doron Remember when you´re working with supercells the 1st Brillouin zone become altered, particularlly reduced. So the orignal band structure, based on the original 1st BZ, is then folding in the new one. So, what you are seeing is a mixture of bandlines. In particular for graphene 2x2 supercell diract point is still present at (1/3, 1/3, 0), but you can check how Dirac point moves into Gamma point when supercells are multiple of three (eg: 3x3 and, 6x6 supercells). Best regards Ricardo - Dr. Ricardo Faccio Prof. Adjunto de Física Mail: Cryssmat-Lab., Cátedra de Física, DETEMA Facultad de Química, Universidad de la República Av. Gral. Flores 2124, C.C. 1157 C.P. 11800, Montevideo, Uruguay. E-mail: rfac...@fq.edu.uy Phone: 598 2 924 98 59 598 2 929 06 48 Fax:598 2 9241906 Web: http://cryssmat.fq.edu.uy/ricardo/ricardo.htm - - Original Message - From: "Doron Naveh" To: Sent: Tuesday, May 04, 2010 4:34 PM Subject: [SIESTA-L] band-structure of supercell Hi there siesters, I ran into a very basic problem. I'm trying to calculate the bandstructure of a slab. For that matter I an illustrate my problem on 2x2 graphene unit cell. The \Gamma point of the 2x2 cell gives wrong eigenvalues compared to the 1x1 cell (a factor of ~4!!) . Is there a rigid shift of the BZ center ? Here are the lattice vectors/atom position blocks for the 2x2 cell. for the primitive cell atom position is 0,0,0.5; 1/3,1/3,0.5 and the lattice parameter reduced by half. Thanks, -D. LatticeConstant 4.9200 Ang %block LatticeVectors 0.5 0.866025404 0.0 -0.5 0.866025404 0.0 0.0 0.0 3.0 %endblock LatticeVectors AtomicCoordinatesFormat Fractional %block AtomicCoordinatesAndAtomicSpecies 0.0 0.0 0.500 1 0.1 0.1 0.500 1 0.0 0.5 0.500 1 0.1 0.6 0.500 1 0.5 0.0 0.500 1 0.6 0.1 0.500 1 0.5 0.5 0.500 1 0.6 0.6 0.500 1 %endblock AtomicCoordinatesAndAtomicSpecies BandLinesScale ReciprocalLatticeVectors %block BandLines 1 0.000 0.000 0.00 \Gamma %endblock BandLines
Re: [SIESTA-L] band-structure of supercell
Hi Ricardo, Thanks for the prompt response. I know about the BZ folding, this is the reason why I took the BZ center as an example to make my point. The band gap at \Gamma decreases dramatically (factor of ~ 4) when I use the 2x2 cell. Since the \Gamma point is not subjected to folding issues, is it possible that there is a bug that translate the BZ center? Thanks, -D. On 5/4/2010 3:47 PM, Ricardo Faccio wrote: Dear Doron Remember when you´re working with supercells the 1st Brillouin zone become altered, particularlly reduced. So the orignal band structure, based on the original 1st BZ, is then folding in the new one. So, what you are seeing is a mixture of bandlines. In particular for graphene 2x2 supercell diract point is still present at (1/3, 1/3, 0), but you can check how Dirac point moves into Gamma point when supercells are multiple of three (eg: 3x3 and, 6x6 supercells). Best regards Ricardo - Dr. Ricardo Faccio Prof. Adjunto de Física Mail: Cryssmat-Lab., Cátedra de Física, DETEMA Facultad de Química, Universidad de la República Av. Gral. Flores 2124, C.C. 1157 C.P. 11800, Montevideo, Uruguay. E-mail: rfac...@fq.edu.uy Phone: 598 2 924 98 59 598 2 929 06 48 Fax:598 2 9241906 Web: http://cryssmat.fq.edu.uy/ricardo/ricardo.htm - - Original Message - From: "Doron Naveh" To: Sent: Tuesday, May 04, 2010 4:34 PM Subject: [SIESTA-L] band-structure of supercell Hi there siesters, I ran into a very basic problem. I'm trying to calculate the bandstructure of a slab. For that matter I an illustrate my problem on 2x2 graphene unit cell. The \Gamma point of the 2x2 cell gives wrong eigenvalues compared to the 1x1 cell (a factor of ~4!!) . Is there a rigid shift of the BZ center ? Here are the lattice vectors/atom position blocks for the 2x2 cell. for the primitive cell atom position is 0,0,0.5; 1/3,1/3,0.5 and the lattice parameter reduced by half. Thanks, -D. LatticeConstant 4.9200 Ang %block LatticeVectors 0.5 0.866025404 0.0 -0.5 0.866025404 0.0 0.0 0.0 3.0 %endblock LatticeVectors AtomicCoordinatesFormat Fractional %block AtomicCoordinatesAndAtomicSpecies 0.0 0.0 0.500 1 0.1 0.1 0.500 1 0.0 0.5 0.500 1 0.1 0.6 0.500 1 0.5 0.0 0.500 1 0.6 0.1 0.500 1 0.5 0.5 0.500 1 0.6 0.6 0.500 1 %endblock AtomicCoordinatesAndAtomicSpecies BandLinesScale ReciprocalLatticeVectors %block BandLines 1 0.000 0.000 0.00 \Gamma %endblock BandLines
[SIESTA-L] band-structure of supercell
Hi there siesters, I ran into a very basic problem. I'm trying to calculate the bandstructure of a slab. For that matter I an illustrate my problem on 2x2 graphene unit cell. The \Gamma point of the 2x2 cell gives wrong eigenvalues compared to the 1x1 cell (a factor of ~4!!) . Is there a rigid shift of the BZ center ? Here are the lattice vectors/atom position blocks for the 2x2 cell. for the primitive cell atom position is 0,0,0.5; 1/3,1/3,0.5 and the lattice parameter reduced by half. Thanks, -D. LatticeConstant 4.9200 Ang %block LatticeVectors 0.5 0.866025404 0.0 -0.5 0.866025404 0.0 0.0 0.0 3.0 %endblock LatticeVectors AtomicCoordinatesFormat Fractional %block AtomicCoordinatesAndAtomicSpecies 0.0 0.0 0.500 1 0.1 0.1 0.500 1 0.0 0.5 0.500 1 0.1 0.6 0.500 1 0.5 0.0 0.500 1 0.6 0.1 0.500 1 0.5 0.5 0.500 1 0.6 0.6 0.500 1 %endblock AtomicCoordinatesAndAtomicSpecies BandLinesScale ReciprocalLatticeVectors %block BandLines 1 0.000 0.000 0.00 \Gamma %endblock BandLines
[SIESTA-L] Posible SPAM: RANGER TACC
Hi, Does anyone have an experience with compiling siesta on RANGER TACC ? http://services.tacc.utexas.edu/index.php/ranger-user-guide -D.
Re: [SIESTA-L] basis set overlap matrix
e, Temp ! Total > electronic charge and Temperature if(iv>0)write(ilog,*) "import_hsx: > total_electronic_charge, Temp (Ry):", real(total_electronic_charge,4), > real(Temp,4) deallocate(int_buff, sp_buff); close(ifile); > > end subroutine !import_hsx > > !! > !! > !! > subroutine sparse2full(ndim, M_full, M_sparse, col2nnzero, col2displ, > sparse_ind2row) use modul_precision !! external > integer(4), intent(in) :: ndim real(dp), intent(out) :: M_full(ndim,ndim) > real(dp), intent(in) :: M_sparse(:) integer(4), intent(in) :: > col2nnzero(ndim), col2displ(ndim), sparse_ind2row(:) > > > !! internal > integer :: icol, i, irow, sparse_ind > > do icol=1,ndim do i=1,col2nnzero(icol); sparse_ind = col2displ(icol)+i; irow > = sparse_ind2row(sparse_ind); > M_full(irow, icol) = M_sparse(sparse_ind) > enddo enddo > > end subroutine !sparse2full > > > On Fri, Mar 12, 2010 at 7:41 PM, Doron Naveh wrote: > > >> Hi, >> I'm trying to obtain the overlap matrix of basis set functions, >> does anyone know how? Thanks, >> Doron. >> >> >> > > > -- > Dr. Peter Koval > email: koval.pe...@gmail.com > inet: http://sites.google.com/site/kovalpeter/ > >
[SIESTA-L] basis set overlap matrix
Hi, I'm trying to obtain the overlap matrix of basis set functions, does anyone know how? Thanks, Doron.
[SIESTA-L] system.HSX files
Hi A question related to system.HS/system.HSX files: If I would like to get the overlap matrix of some non-scf data, say I have one grid of k-points to sample BZ for convergence and another set of "speial" k-points for which I would like the overlap matrix to correspond. a. Is this possible to do in SIESTA? b. I assumed that if I write the file system.WFSX, the overlap matrix should correspond to the content of this file, is it true? Thanks, D.
[SIESTA-L] site-projection of KS WFs
Dear users, Let me first apologize if the following questions are too simple, I'm a new user of SIESTA and could not find answers to these questions. I would like to project a KS orbital on some atomic basis functions. 1. Is there a ready-to use utility for such application? 2. If the answer to (1) is "No", is there an easy access to the atomic basis set functions in terms of radial function ? 3. If the answer to (1) and (2) is "No", could someone advise me on where in the code (subroutine name) I would have access to both the KS orbitals and basis-set functions, such that I would not need to allocate large arrays? Thanks!