Re: [SIESTA-L] Problem FixSpin
Dear Marcos, the systems Eliseo is working on are molecular magnets, they usually have well localized spins on 3d ions and can be brought to self-consistency in a number of different magnetic configurations. Even if (usually) there is a certain (single-determinant) ground-state magnetic configuration, it is Eliseo's good right to expect a number of other metastable states to converge as well. The exact numbers (from integrated charge density or Mulliken analysis) are usually unimportant, because they reveal in one or other way a certain arrangement of integer spins of 3d ions. And of course the fixed spin moment, if imposed, remains fixed at each iteration (at the expense of introducing two Fermi levels, also in each iteration). Best regards Andrei +-- Dr. Andrei Postnikov Tel. +33-387315873 - mobile +33-666784053 ---+ | Paul Verlaine University - Institute de Physique Electronique et Chimie, | | Laboratoire de Physique des Milieux Denses, 1 Bd Arago, F-57078 Metz, France | +-- [EMAIL PROTECTED] -- http://www.home.uni-osnabrueck.de/apostnik/ --+ On Wed, 28 Nov 2007, Marcos Verissimo Alves wrote: | Eliseo, | | The numbers you are showing for the spins are extracted from some | integration over the densities, or those are the mulliken charges on the | system? I will suppose that those are the numbers that come out after one | or more complete self-consistency cycles; if it is so, it means that your | system "doesn't want to keep" the initial values you have supplied; it | will be more advantageous, energetically, to have the final configuration | of spins. The initial configuration is just a starting guess, but | self-consistency takes over after this. Nothing guarantees that you will | have the final global net spin, or the individual net spins of the atoms, | equal to the initial values after a whole SCF cycle. Actually, you can | constrain the total magnetization of a system in siesta, but I am not sure | if the total magnetization would be constrained over all the scf steps. | | Cheers, | | Marcos | | | | Vous avez écrit / You have written / Lei ha scritto / Você escreveu... | Eliseo Ruiz | > Dear Colleagues, | > | > I am calculating a periodic system with 40 Cu(II) cations in the | > unit cell. I am trying to | > calculate the energy for the different spin distributions. However, | > only in a couple of | > cases I obtain the signs of the spins that I specified in the input | > file. For instance, for one | > of the spin distributions: | > | > | > # Spin defintion | > SpinPolarized T | > FixSpin T | > TotalSpin 8.0 | > %block DM.InitSpin | >1 +1.0 | >2 -1.0 | >3 -1.0 | >4 +1.0 | >5 +1.0 | >6 +1.0 | >7 -1.0 | >8 -1.0 | >9 +1.0 | >10+1.0 | >11+1.0 | > | >40+1.0 | > %endblock DM.InitSpin | > | > the results for the spin density: | > atom up down total | > 1 4.968 -5.401 -.433 | > 2 4.978 -5.411 -.433 | > 3 5.397 -4.948 .449 | > 4 5.406 -4.942 .464 | > 5 5.431 -4.963 .468 | > 6 5.407 -4.962 .445 | > 7 5.415 -4.977 .438 | > 8 4.957 -5.391 -.434 | > 9 4.943 -5.404 -.461 | > 10 5.439 -4.964 .475 | > 12 5.420 -4.987 .433 | > | > | > The total spin is satisfied but not the local spins. I tried several | > options of DM.MixingWeight and | > DM.NumberPulay as well as for ElectronicTemperature but I cannot | > reach the selected spin | > distribution. | > | > best wishes, | > | >Eliseo | > | | | -- | Dr. Marcos Verissimo Alves | Post-Doctoral Fellow | Unité de Physico-Chimie et de Physique des Matériaux (PCPM) | Université Catholique de Louvain | 1 Place Croix du Sud, B-1348 | Louvain-la-Neuve | Belgique | | -- | | Gort, Klaatu barada nikto. Klaatu barada nikto. Klaatu barada nikto. | |
Re: [SIESTA-L] Problem FixSpin
FixSpin only fixes the total magnetization, making SIESTA very the Fermi levels for up and down spins so that the net spin population will be what you've specified in the parameter. This has nothing to do with the populations on individual atoms. Clearly, trying to impose a fixed spin on each atom would be physically meaningless. 2007/11/28, Marcos Verissimo Alves <[EMAIL PROTECTED]>: > > Eliseo, > > The numbers you are showing for the spins are extracted from some > integration over the densities, or those are the mulliken charges on the > system? I will suppose that those are the numbers that come out after one > or more complete self-consistency cycles; if it is so, it means that your > system "doesn't want to keep" the initial values you have supplied; it > will be more advantageous, energetically, to have the final configuration > of spins. The initial configuration is just a starting guess, but > self-consistency takes over after this. Nothing guarantees that you will > have the final global net spin, or the individual net spins of the atoms, > equal to the initial values after a whole SCF cycle. Actually, you can > constrain the total magnetization of a system in siesta, but I am not sure > if the total magnetization would be constrained over all the scf steps. > > Cheers, > > Marcos > > > > Vous avez écrit / You have written / Lei ha scritto / Você escreveu... > Eliseo Ruiz > > Dear Colleagues, > > > > I am calculating a periodic system with 40 Cu(II) cations in the > > unit cell. I am trying to > > calculate the energy for the different spin distributions. However, > > only in a couple of > > cases I obtain the signs of the spins that I specified in the input > > file. For instance, for one > > of the spin distributions: > > > > > > # Spin defintion > > SpinPolarized T > > FixSpin T > > TotalSpin 8.0 > > %block DM.InitSpin > >1 +1.0 > >2 -1.0 > >3 -1.0 > >4 +1.0 > >5 +1.0 > >6 +1.0 > >7 -1.0 > >8 -1.0 > >9 +1.0 > >10+1.0 > >11+1.0 > > > >40+1.0 > > %endblock DM.InitSpin > > > > the results for the spin density: > > atom up down total > > 1 4.968 -5.401 -.433 > > 2 4.978 -5.411 -.433 > > 3 5.397 -4.948 .449 > > 4 5.406 -4.942 .464 > > 5 5.431 -4.963 .468 > > 6 5.407 -4.962 .445 > > 7 5.415 -4.977 .438 > > 8 4.957 -5.391 -.434 > > 9 4.943 -5.404 -.461 > > 10 5.439 -4.964 .475 > > 12 5.420 -4.987 .433 > > > > > > The total spin is satisfied but not the local spins. I tried several > > options of DM.MixingWeight and > > DM.NumberPulay as well as for ElectronicTemperature but I cannot > > reach the selected spin > > distribution. > > > > best wishes, > > > >Eliseo > > > > > -- > Dr. Marcos Verissimo Alves > Post-Doctoral Fellow > Unité de Physico-Chimie et de Physique des Matériaux (PCPM) > Université Catholique de Louvain > 1 Place Croix du Sud, B-1348 > Louvain-la-Neuve > Belgique > > -- > > Gort, Klaatu barada nikto. Klaatu barada nikto. Klaatu barada nikto. >
Re: [SIESTA-L] Problem FixSpin
Dear Eliseo, with respect to your problem of initializing different spin configurations, I can offer two comments: 1. When you start from scratch, I am not sure how exactly the initial spin splitting is constructed; I also had sometimes an impression that the resulting splitting (in the 1st iteration) is too small. My rule of thumb in similar cases is to use maximal splitting, i.e., %block DM.InitSpin 1 + 2 - etc. in all cases, even for Cu. 2. As you have found a reasonable solution for one spin configuration you can invert certain spin(s) in its density matrice using my tool http://www.home.uni-osnabrueck.de/apostnik/download.html -> DMtune that will normally give you a better starting point for further convergence than starting another spin configuration from scratch. Best regards, Andrei +-- Dr. Andrei Postnikov Tel. +33-387315873 - mobile +33-666784053 ---+ | Paul Verlaine University - Institute de Physique Electronique et Chimie, | | Laboratoire de Physique des Milieux Denses, 1 Bd Arago, F-57078 Metz, France | +-- [EMAIL PROTECTED] -- http://www.home.uni-osnabrueck.de/apostnik/ --+ On Wed, 28 Nov 2007, Eliseo Ruiz wrote: | Dear Colleagues, | | I am calculating a periodic system with 40 Cu(II) cations in the unit | cell. I am trying to calculate the energy for the different spin | distributions. However, only in a couple of cases I obtain the signs | of the spins that I specified in the input file. For instance, for one | of the spin distributions: | | | # Spin defintion | SpinPolarized T | FixSpin T | TotalSpin 8.0 | %block DM.InitSpin | 1 +1.0 | 2 -1.0 | 3 -1.0 | 4 +1.0 | 5 +1.0 | 6 +1.0 | 7 -1.0 | 8 -1.0 | 9 +1.0 | 10+1.0 | 11+1.0 | | 40+1.0 | %endblock DM.InitSpin | | the results for the spin density: | atom up down total | 1 4.968 -5.401 -.433 | 2 4.978 -5.411 -.433 | 3 5.397 -4.948 .449 | 4 5.406 -4.942 .464 | 5 5.431 -4.963 .468 | 6 5.407 -4.962 .445 | 7 5.415 -4.977 .438 | 8 4.957 -5.391 -.434 | 9 4.943 -5.404 -.461 | 10 5.439 -4.964 .475 | 12 5.420 -4.987 .433 | | | The total spin is satisfied but not the local spins. I tried several options | of DM.MixingWeight and | DM.NumberPulay as well as for ElectronicTemperature but I cannot reach the | selected spin | distribution. | | best wishes, | | Eliseo | |
Re: [SIESTA-L] Optical calculation
Dear Vikas Thakur, There is a problem with the Optical.Broaden parameter, it would be set for a nonzero value Engin Deligoz On Tue, 27 Nov 2007 23:46:22 -0800, vikas thakur <[EMAIL PROTECTED]> wrote: > Dear all, > >I am trying to calculate optical properties of nanotubes and i am using > a tetragonal geometry for this. I have specified perameter as follow. > > OpticalCalculation .true. > Optical.EnergyMinimum 0 Ry > Optical.EnergyMaximum10 Ry > %block Optical.Mesh >1 1 5 > %endblock Optical.Mesh > Optical.PolarizationTypeunpolarized > %block Optical.Vectors > 0.0 0.0 1.0 > %endblock Optical.Vectors > %block GridCellSampling > 1.0 1.0 0.5 > %block GridCellSampling > > > I get a file *.EPSIMG but there no information in this file means there > is no data. Only information is > > ## Minimum and maximum energy in eV > ## 0. 0. > ## Number of spin component > ## 1 > ## f-sum rule > ##0. > ## wp^2 in Ha^2 for Drude > ##NaN > > I tried optical calculation first time so i may have given wrong input > for optical calculation . Please tell me where is the problem. > > > Thanking You. > With best regards. > > Vikas Thakur. > > > ---
Re: [SIESTA-L] Problem FixSpin
Eliseo, The numbers you are showing for the spins are extracted from some integration over the densities, or those are the mulliken charges on the system? I will suppose that those are the numbers that come out after one or more complete self-consistency cycles; if it is so, it means that your system "doesn't want to keep" the initial values you have supplied; it will be more advantageous, energetically, to have the final configuration of spins. The initial configuration is just a starting guess, but self-consistency takes over after this. Nothing guarantees that you will have the final global net spin, or the individual net spins of the atoms, equal to the initial values after a whole SCF cycle. Actually, you can constrain the total magnetization of a system in siesta, but I am not sure if the total magnetization would be constrained over all the scf steps. Cheers, Marcos Vous avez écrit / You have written / Lei ha scritto / Você escreveu... Eliseo Ruiz > Dear Colleagues, > > I am calculating a periodic system with 40 Cu(II) cations in the > unit cell. I am trying to > calculate the energy for the different spin distributions. However, > only in a couple of > cases I obtain the signs of the spins that I specified in the input > file. For instance, for one > of the spin distributions: > > > # Spin defintion > SpinPolarized T > FixSpin T > TotalSpin 8.0 > %block DM.InitSpin >1 +1.0 >2 -1.0 >3 -1.0 >4 +1.0 >5 +1.0 >6 +1.0 >7 -1.0 >8 -1.0 >9 +1.0 >10+1.0 >11+1.0 > >40+1.0 > %endblock DM.InitSpin > > the results for the spin density: > atom up down total > 1 4.968 -5.401 -.433 > 2 4.978 -5.411 -.433 > 3 5.397 -4.948 .449 > 4 5.406 -4.942 .464 > 5 5.431 -4.963 .468 > 6 5.407 -4.962 .445 > 7 5.415 -4.977 .438 > 8 4.957 -5.391 -.434 > 9 4.943 -5.404 -.461 > 10 5.439 -4.964 .475 > 12 5.420 -4.987 .433 > > > The total spin is satisfied but not the local spins. I tried several > options of DM.MixingWeight and > DM.NumberPulay as well as for ElectronicTemperature but I cannot > reach the selected spin > distribution. > > best wishes, > >Eliseo > -- Dr. Marcos Verissimo Alves Post-Doctoral Fellow Unité de Physico-Chimie et de Physique des Matériaux (PCPM) Université Catholique de Louvain 1 Place Croix du Sud, B-1348 Louvain-la-Neuve Belgique -- Gort, Klaatu barada nikto. Klaatu barada nikto. Klaatu barada nikto.
[SIESTA-L] Problem FixSpin
Dear Colleagues, I am calculating a periodic system with 40 Cu(II) cations in the unit cell. I am trying to calculate the energy for the different spin distributions. However, only in a couple of cases I obtain the signs of the spins that I specified in the input file. For instance, for one of the spin distributions: # Spin defintion SpinPolarized T FixSpin T TotalSpin 8.0 %block DM.InitSpin 1 +1.0 2 -1.0 3 -1.0 4 +1.0 5 +1.0 6 +1.0 7 -1.0 8 -1.0 9 +1.0 10+1.0 11+1.0 40+1.0 %endblock DM.InitSpin the results for the spin density: atom up down total 1 4.968 -5.401 -.433 2 4.978 -5.411 -.433 3 5.397 -4.948 .449 4 5.406 -4.942 .464 5 5.431 -4.963 .468 6 5.407 -4.962 .445 7 5.415 -4.977 .438 8 4.957 -5.391 -.434 9 4.943 -5.404 -.461 10 5.439 -4.964 .475 12 5.420 -4.987 .433 The total spin is satisfied but not the local spins. I tried several options of DM.MixingWeight and DM.NumberPulay as well as for ElectronicTemperature but I cannot reach the selected spin distribution. best wishes, Eliseo
[SIESTA-L] Optical calculation
Dear all, I am trying to calculate optical properties of nanotubes and i am using a tetragonal geometry for this. I have specified perameter as follow. OpticalCalculation .true. Optical.EnergyMinimum 0 Ry Optical.EnergyMaximum10 Ry %block Optical.Mesh 1 1 5 %endblock Optical.Mesh Optical.PolarizationTypeunpolarized %block Optical.Vectors 0.0 0.0 1.0 %endblock Optical.Vectors %block GridCellSampling 1.0 1.0 0.5 %block GridCellSampling I get a file *.EPSIMG but there no information in this file means there is no data. Only information is ## Minimum and maximum energy in eV ## 0. 0. ## Number of spin component ## 1 ## f-sum rule ##0. ## wp^2 in Ha^2 for Drude ##NaN I tried optical calculation first time so i may have given wrong input for optical calculation . Please tell me where is the problem. Thanking You. With best regards. Vikas Thakur. - Be a better sports nut! Let your teams follow you with Yahoo Mobile. Try it now.
