Author Topic: Hamiltonian Derivatives  (Read 1201 times)

0 Members and 1 Guest are viewing this topic.

Offline gozde

  • Heavy QuantumATK user
  • ***
  • Posts: 29
  • Country: tr
  • Reputation: 0
    • View Profile
Hamiltonian Derivatives
« on: April 14, 2023, 10:59 »
Hello,
I succeed to obtain hamiltonian derivatives for structures which have unitcell composed of 2, 3 and 4 atoms. However, executation tasks do not progress even one step in unitcell containing 24 atoms. I heva used thight parameters to obtain phonon dispersion accurately and I achieved. I guess this problem can be originated from high accuracy of the parameters of the LCAO calculator because I use, analysis from file (optimized bulk structure) before the study objects. Unfortunately executation task stays on the same step 1/73 for hours and does not progress.
Do you have any suggestion to make progress?

Offline Ulrik G. Vej-Hansen

  • QuantumATK Staff
  • Supreme QuantumATK Wizard
  • *****
  • Posts: 425
  • Country: dk
  • Reputation: 8
    • View Profile
Re: Hamiltonian Derivatives
« Reply #1 on: April 18, 2023, 09:52 »
Note that Hamiltonian Derivatives repeats your input cell to create a supercell, so it is expected that it will take a long time, especially with tight calculator settings.

Offline gozde

  • Heavy QuantumATK user
  • ***
  • Posts: 29
  • Country: tr
  • Reputation: 0
    • View Profile
Re: Hamiltonian Derivatives
« Reply #2 on: May 6, 2023, 13:25 »
Actually the main problem arises when qatk performing the Fermi level calculation. This first task, unfortunately is not over. You can find the log files at the attachmenet.

Offline gozde

  • Heavy QuantumATK user
  • ***
  • Posts: 29
  • Country: tr
  • Reputation: 0
    • View Profile
Re: Hamiltonian Derivatives
« Reply #3 on: May 17, 2023, 15:10 »
Can you suggest any recipe to overcome this problem?

Offline gozde

  • Heavy QuantumATK user
  • ***
  • Posts: 29
  • Country: tr
  • Reputation: 0
    • View Profile
Re: Hamiltonian Derivatives
« Reply #4 on: June 20, 2023, 11:24 »
Hello again,
Hamiltonian derivatives calculation has been completed by increasing atomic displacement from 0.01 Ang to 0.1 Ang. I know that Hamiltonian derivatives is calculated via finite difference method, so increasing the default value of atomic displacement gives the wrong result?