Dear Andrei Postnikov Sir, Can you please guide me in how to do the 'eggbox test' ? I have gone through the tutorials and examples but am still unclear on how to do that test for a system. Will be very much thankful if you let me know the steps on testing the eggbox effect and eliminating it. I have seen the term 'eggbox energy' in siesta output file after completion of a run, but confused what to decipher from it. Thanking you and awaiting a reply.
Warm regards and wishing you a happy new year in advance, Sunetra Das. On 25 December 2017 at 19:52, Andrei Postnikov < andrei.postni...@univ-lorraine.fr> wrote: > Dear Najmeh, > > you are worrying about z-components of forces > and try to help it by increasing the k(xy) mesh. It means that your > bands dispersion becomes ever better in the plane, > however, I don't think this would have an important effect on the z-force! > > Three observations: > > 1) > Do you really have a problem? > Your z-forces on atoms 1 and 2 do nicely sum up to zero within 0.001 - > that is fine, - > and the fact that they (individually) fluctuate with k simply reveals the > fact > that your structure is not at equilibrium. Try to get a better equilibrium > - see 2). > If you ultimate goal is phonons, why don't you calculate just Gamma > and check if you get the acoustic modes close enough to zero? > > 2) > In order to suppress fluctuations of forces / energy, > the MeshCutoff is probably much more important that k-mesh. > Apparently you tried 300 Ry and did not increase it any further. > It might be not sufficient, depending on pseudos / basis. > (However, it might well be OK in your case, see 1). > In order to understand what's going on, the eggbox test > (rigidly moving all the atoms... along Z, your "problematic" direction) > for different values of MeshCutoff would help. > With insufficient MeshCutoff, the increase of the k-mesh alone won't help > you. > (and... do you really need 20 Ang in the z-direction? You only need > so much that the basis functions won't overlap). > > 3) > The k-mesh is not a variational parameter, > so you cannot expect a systematic lowering of the total energy > (although it usually happens, albeit with fluctuations). > For the phonon calculations, the total energy as such (i.e., > its absolute value) is irrelevant anyway. > > Best regards > > Andrei Postnikov > > > ----- Najmeh Honari <najmeh.honari...@gmail.com> a écrit : > > > Dear > > Siesta users, > > > > > > > I’m > > trying to simulate 2D borophene δ*6*, in order to calculate phonon > dispersion of the material. I did convergence > > test for mesh-cut-off and k-points. I selected mesh cutoff=300Ry and > started > > k-grid from 12x12x1 and increased it to 40x40x1. The issue is that > neither “total > > energy” nor the “forces on atoms” are converged in other words > fluctuations of > > energies and forces are obsereved. Here are a table of some energy and > forces > > for different k-grid meshes of borophene > > structure: > > > > >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> > > > > > K-grid > > > > > > > Total > > Energy (eV) > > > > > > > Atomic > > forces (eV/A) > > > > > > > > 15x15 > > > > > > > -206.414196 > > > > > > > 1 -0.000001 -0.000005 > > -0.131499 > > > > > 2 -0.000002 -0.000004 0.131227 > > > > > > > > 17x17 > > > > > > > -206.410794 > > > > > > > 1 0.000000 -0.000003 > > -0.423281 > > > > > 2 0.000004 -0.000001 0.423056 > > > > > > > > 20x20 > > > > > > > -206.413284 > > > > > > > 1 0.000001 -0.000006 > > -0.287325 > > > > > 2 -0.000002 0.000002 0.287087 > > > > > > > > 23x23 > > > > > > > -206.414429 > > > > > > > 1 0.000000 -0.000000 > > -0.355986 > > > > > 2 -0.000002 -0.000005 0.355733 > > > > > > > > 25x25 > > > > > > > -206.423565 > > > > > > > 1 -0.000001 -0.000000 > > -0.416997 > > > > > 2 -0.000002 -0.000003 0.416772 > > > > > > > > 27x27 > > > > > > > -206.424037 > > > > > > > 1 -0.000002 -0.000006 > > -0.308230 > > > > > 2 -0.000001 0.000000 0.307990 > > > > > > > > 30x30 > > > > > > > -206.422290 > > > > > > > 1 -0.000001 -0.000004 > > -0.313348 > > > > > 2 -0.000000 -0.000002 0.313111 > > > > > > > > 33x33 > > > > > > > -206.423341 > > > > > > > 1 -0.000002 -0.000008 > > -0.368439 > > > > > 2 -0.000001 -0.000005 0.368198 > > > > > > > > 35x35 > > > > > > > -206.422301 > > > > > > > 1 0.000002 -0.000004 > > -0.275519 > > > > > 2 -0.000002 0.000001 0.275263 > > > > > > > > 38x38 > > > > > > > -206.420742 > > > > > > > 1 0.000002 0.000000 > > -0.288896 > > > > > 2 -0.000000 0.000001 0.288655 > > > > > > > > 40x40 > > > > > > > -206.420015 > > > > > > > 1 -0.000001 -0.000005 > > -0.357026 > > > > > 2 -0.000000 0.000000 0.356782 > > > > > > > > > > > > > > > > The > > input.fdf file that is used in simulations is as follow: > > > > > > > > > > > >>>>> > > > > > SystemName B > > > > > SystemLabel Borophene > > > > > NumberOfAtoms 2 > > > > > NumberOfSpecies 1 > > > > > ################## > > > > > #Basis# > > > > > ################## > > > > > PAO.BasisSize DZP > > > > > ################## > > > > > #ChemicalSpecies# > > > > > ################## > > > > > %block > > ChemicalSpeciesLabel > > > > > 1 > > 5 B > > > > > %endblock > > ChemicalSpeciesLabel > > > > > ################################## > > > > > LatticeConstant 1.00 Ang > > > > > > > > > > %block LatticeVectors > > > > > 1.60000 0.0000 > > 0.0000 > > > > > 0.00000 2.8200 > > 0.0000 > > > > > 0.00000 0.0000 > > 20.000 > > > > > %endblock > > LatticeVectors > > > > > > > > > > AtomicCoordinatesFormat > > Ang > > > > > > > > > > %block > > AtomicCoordinatesAndAtomicSpecies > > > > > 0.00000 0.00000 > 0.00000 1 > > > > > 0.80000 1.41000 > > 0.91000 1 > > > > > %endblock AtomicCoordinatesAndAtomicSpecies > > > > > > > > > > BandLinesScale > ReciprocalLatticeVectors > > > > > > > > > > %block BandLines > > > > > 1 0.0000000000 0.0000000000 0.0000000000 Gamma > > > > > 200 0.5000000000 0.0000000000 0.0000000000 X > > > > > 200 0.5000000000 0.5000000000 0.0000000000 S > > > > > 200 0.0000000000 0.5000000000 0.0000000000 Y > > > > > 200 0.0000000000 0.0000000000 0.0000000000 Gamma > > > > > %endblock BandLines > > > > > > > > > > %block kgrid_Monkhorst_Pack > > > > > 15 > > 0 0 0.0 > > > > > 0 > > 15 0 0.0 > > > > > 0 > > 0 1 0.0 > > > > > %endblock > > kgrid_Monkhorst_pack > > > > > ################ > > > > > #DFT, Grid, SCF# > > > > > ################ > > > > > SolutionMethod diagon > > > > > XC.Functional GGA > > > > > XC.authors PBE > > > > > SpinPolarized F > > > > > NonCollinearSpin F > > > > > > FixSpin F > > > > > TotalSpin 0.0 > > > > > > SingleExcitation F > > > > > MeshCutoff 300 Ry > > > > > MaxSCFIterations 100 > > > > > DM.MixingWeight 0.3 > > > > > DM.NumberPulay 6 > > > > > DM.NumberKick 5 > > > > > DM.KickMixingWeight 0.2 > > > > > DM.MixSCF1 F > > > > > DM.Tolerance 0.0001 > > > > > DM.InitSpinAF F > > > > > > EggboxScale 1 eV > > > > > ##################### > > > > > #Eigenvalue problems# > > > > > ##################### > > > > > MD.TypeOfRun CG > > > > > MD.VariableCell F > > > > > MD.NumCGsteps 0 > > > > > MD.MaxCGDispl 0.2 Bohr > > > > > MD.PreconditionVariableCell 5.0 Ang > > > > > MD.MaxForceTol 0.001 eV/Ang > > > > > ################## > > > > > #Parallel options# > > > > > ################## > > > > > BlockSize 8 > > > > > DiagMemory 2.00 # Default value > > > > > DiagScale 1.25 > > > > > TryMemoryIncrease T # Default value > > > > > ################ > > > > > > > > > > > > > > > > It’s > > worth to say that I simulated BlackP by same INPUT.fdf file, and there > were not > > any fluctuation in energy and forces. Here are a table of some energy > > and forces for different k-grid meshes of Black phosphorene structure. > > > > > > >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> > > > > > K-grid > > > > > > > Total > > Energy (eV) > > > > > > > Atomic > > forces (eV/A) > > > > > > > > 12x12 > > > > > > > -834.239365 > > > > > > > 1 1.028242 0.000000 0.123389 > > > > > 2 -1.168924 -0.000003 > > -0.124082 > > > > > 3 1.182128 0.000001 > > -0.130042 > > > > > 4 -1.021002 0.000000 0.123803 > > > > > > > > 15x15 > > > > > > > -834.239563 > > > > > > > 1 1.026970 -0.000001 0.121664 > > > > > 2 -1.167701 -0.000001 > > -0.122354 > > > > > 3 1.180904 0.000003 > > -0.128320 > > > > > 4 -1.019726 -0.000000 0.122082 > > > > > > > > 18x18 > > > > > > > -834.239611 > > > > > > > 1 1.026466 -0.000001 0.121037 > > > > > 2 -1.167212 -0.000000 > > -0.121726 > > > > > 3 1.180422 0.000002 > > -0.127694 > > > > > 4 -1.019227 -0.000001 0.121454 > > > > > > > > 21x21 > > > > > > > -834.239629 > > > > > > > 1 1.026270 -0.000002 0.120792 > > > > > 2 -1.167025 -0.000001 > > -0.121484 > > > > > 3 1.180231 0.000001 > > -0.127446 > > > > > 4 -1.019028 -0.000001 0.121212 > > > > > > > > 24x24 > > > > > > > -834.239634 > > > > > > > 1 1.026188 -0.000001 0.120697 > > > > > 2 -1.166949 0.000000 > > -0.121387 > > > > > 3 1.180158 0.000002 > > -0.127350 > > > > > 4 -1.018948 -0.000001 0.121115 > > > > > > > > 27x27 > > > > > > > -834.239638 > > > > > > > 1 1.026157 -0.000000 0.120657 > > > > > 2 -1.166919 -0.000001 > > -0.121349 > > > > > 3 1.180126 0.000002 > > -0.127309 > > > > > 4 -1.018913 0.000000 0.121076 > > > > > > > > 30x30 > > > > > > > -834.239638 > > > > > > > 1 1.026142 -0.000002 0.120640 > > > > > 2 -1.166907 -0.000000 > > -0.121333 > > > > > 3 1.180114 0.000002 > > -0.127294 > > > > > 4 -1.018900 -0.000001 0.121060 > > > > > > > > > > > > > > > > Could > > you let me know If I did something wrong with .fdf file in the first > case? > > > > > Regards, > Najmeh Honari > > >
