Hi Saikat,
Assuming your structure is stable, negative frequencies generally mean poor convergence of either, the optimized structure or the following FC run. I have no experience with CNT, though the parameters you're using seem pretty ok to me, but for one: it is "Meshcutoff" not "Messcutoff" as you type. In any case, 150 Ry might be on the low side (depending on box's size). It is a particularly critical parameter, check the manual for the "eggbox" effect. Also, you might want to check the *.FC file that Siesta outputs. Namely, the forces for + and - displacements should be about the same, particularly on the moving atom (file structure has been explained in a recent post to the list). Good luck, Roberto On Mon, 15 Oct 2012, [email protected] wrote:
Dear siesta user, I am doing phonon calculation of carbon nanotube using siesta. But I am getting negative frequencies for more than half of the calculated frequencies. I wanted to know reason for getting negative frequencies of my calculation. I have optimized the structure using MD.MaxForceTol 0.001 eV/Ang And condition for scf are as follows DM.Tolerance 0.5d-4 DM.Require.Energy.Convergence true DM.Energy.Tolerance .5d-4 eV My input file for type of run- FC ---- SystemName Phonon calculation # Descriptive name of the system SystemLabel Phonon_fc # Short name for naming files # Output options LongOutput false # Species and atoms %block ChemicalSpeciesLabel 1 6 C %endblock ChemicalSpeciesLabel NumberOfSpecies 1 WriteKpoints true PAO.Basis<basis.fdf LatticeConstant 1 Ang Messcutoff 150 Ry Kgridcutoff 10.0 Ang xc.functional GGA xc.authors PBE DM.UseSaveDM true MaxSCFIterations 500 DM.MixingWeight 0.25 DM.NumberPulay 4 DM.Tolerance 0.4d-4 DM.Require.Energy.Convergence true DM.Energy.Tolerance .5d-4 eV #Atomic coordinates %include PhononFC.fdf AtomicCoordinatesFormat Fractional Thanking You. __________________ Saikat Paik IIT Kanpur UP,India
