optimization of the unit cell of molecular nitrogen. Relax and vc-relax.

Posted In DFT (Quantum Espresso)
  • Ivan Dario Arellano Ramirez 1 month ago

    I want to optimize parameters (unit cell, ecutwfc, kpoints) of molecular nitrogen. I would like to know if I need to use spin polarization. In the case it is necessary, what is the difference using tot_magnetization and starting magnetization. Do I need to do after this procedure a geometrical relaxation? When to use relax and vc-relax.

    I appreciate any help.

    I copy the input script here:

    #!/bin/bash

    for i in 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0 4.2 4.4 4.6 4.8 5.0 5.2 5.4 5.6 5.8 6.0 6.4 6.8 7.2 7.6 8.0

    do

    cat >n-$i.in << !

    &CONTROL

    !       calculation = 'relax' ,

            calculation = 'scf' ,

            restart_mode = 'from_scratch' ,

            outdir = './tmp' ,

            pseudo_dir = './', 

            prefix = 'n' ,

            tprnfor = .TRUE.

            tstress = .TRUE.

            verbosity = 'high'

            nstep  = 200

            etot_conv_thr = 1.d-4

            forc_conv_thr = 1.d-3

    /

    &SYSTEM

      ibrav = 4,

      celldm(1) = 20,

      celldm(3) = 2.5,

      nat  = 2,

      ntyp = 2,

      ecutwfc = 60.0,

      ecutrho = 240.0,

      occupations='smearing',

      smearing='methfessel-paxton',

      degauss=0.02,

      nspin = 2,

      starting_magnetization(1) = -0.3 ,

      starting_magnetization(2) =  0.3 ,

    /

    &ELECTRONS

          electron_maxstep = 400,

          conv_thr = 1.0d-6

          mixing_mode      = 'local-TF'

          mixing_beta      = 0.3

          diagonalization  = 'cg'

          startingwfc      = 'atomic+random'

    /

    &IONS

        trust_radius_max = 0.1D0

        ion_dynamics      = 'bfgs'

        ion_positions     = 'default' 

        pot_extrapolation = 'atomic'

        wfc_extrapolation = 'none'

    /

    ATOMIC_SPECIES

    N1   14.007      N.pbe-n-rrkjus_psl.1.0.0.UPF

    N2   14.007      N.pbe-n-rrkjus_psl.1.0.0.UPF

    ATOMIC_POSITIONS (angstrom)

    N1    0.00   0.00   0.00   0  0  0

    N2    0.00   0.00   $i

    K_POINTS automatic

    6  6  6  0  0  0

    !

    echo "a=$i" ; 

    done

     

  • Materials Square 1 month ago

    Hello.

     

    It is better to use the spin-polarization calculation when calculating a single atom (having a single electron). But for the molecule, it is not mandatory because the nspin=2 makes the calculation amount double. I recommend comparing the results with/without spin-polarization to find the more appropriate results. 

     

    And the 'starting_magnetization' is the starting value. The magnetization value will converge during the self-consistency steps. The 'tot_magnetization' determines the magnetization value of the total system. The magnetization value of each component will converge according to this value.

    For further information, please read the following link.

    https://www.quantum-espresso.org/Doc/INPUT_PW.html#idm266

     

    Could you give me detail for the last question? The relax and vc-relax perform the geometry optimization. 

  • Ivan Dario Arellano Ramirez 1 week ago

    Thank you for your response. What I want to ask in the last question is, what are the criteria to use relax or vc-relax? As far as I know, in relax the parameters are fixed, and in vc-relax parameters and atomic positions can be varied.

    I still have doubts with the use of tot_magnetization and starting magnetization. Let's suppose I have an isolated atom of Ga, it has 2 electrons in 4s and 1 electron in 4p subshell, what is the best way to perform a spin-polarization calculation? What about for a nitrogen isolated atom? It has 2 electrons in 2s and 3 electrons in 2p subshell.

    Last question! I want to do a SCF calculation for thallium (Tl) it has a hexagonal structure (SG=194). I have used around 8 different pseudopotentials (from quantum espresso) and all the time I got the same result, the lattice parameter does not converge in minimum energy, I got just a straight line.

    Tl lattice parameter

Please login or register to leave a response.