Author Topic: Transmission normalization ?  (Read 4476 times)

0 Members and 1 Guest are viewing this topic.

Offline ams_nanolab

  • Supreme QuantumATK Wizard
  • *****
  • Posts: 389
  • Country: in
  • Reputation: 11
    • View Profile
Transmission normalization ?
« on: November 4, 2015, 07:18 »
We have simulated the transport on unit cells of mos2, with periodic k-point settings in x and y directions (which i think is effectively considering infinite sheets), and not specifically sized supercells.

For such setting, will the absolute value depends on the width normal to the transport direction  (as in case of a supercell)??

Offline Umberto Martinez

  • Supreme QuantumATK Wizard
  • *****
  • Posts: 479
  • Country: dk
  • Reputation: 26
    • View Profile
Re: Transmission normalization ?
« Reply #1 on: November 4, 2015, 11:58 »
The absolute value of what??
of the current? yes, it depends on the size of your cell on the xy plane.
This is why sometime it is more convenient to plot the current densty.

Offline Anders Blom

  • QuantumATK Staff
  • Supreme QuantumATK Wizard
  • *****
  • Posts: 5565
  • Country: dk
  • Reputation: 93
    • View Profile
    • QuantumATK at Synopsys
Re: Transmission normalization ?
« Reply #2 on: November 4, 2015, 12:48 »
But this is also why you use periodic boundary conditions. Sure, making the width 2x larger gives 2x the current, but also takes longer to calculation. Just use k-points instead.

Offline ams_nanolab

  • Supreme QuantumATK Wizard
  • *****
  • Posts: 389
  • Country: in
  • Reputation: 11
    • View Profile
Re: Transmission normalization ?
« Reply #3 on: November 4, 2015, 17:19 »
Absolute value of transmission spectra I meant.

Offline ams_nanolab

  • Supreme QuantumATK Wizard
  • *****
  • Posts: 389
  • Country: in
  • Reputation: 11
    • View Profile
Re: Transmission normalization ?
« Reply #4 on: November 6, 2015, 11:57 »
one molecule with one level at resonance. Transmission is 2. Now consider two molecules in parallel, T=4. An infinite number corresponds to periodic boundary conditions and T=infinity. Therefore it should be safe to plot the transmission per unit cell width (Normalized transmission) for comparison purposes??

Or should it be the transmission per unit cell length (in the transmission direction)? What should be the correct normalization technique?

Offline Anders Blom

  • QuantumATK Staff
  • Supreme QuantumATK Wizard
  • *****
  • Posts: 5565
  • Country: dk
  • Reputation: 93
    • View Profile
    • QuantumATK at Synopsys
Re: Transmission normalization ?
« Reply #5 on: November 6, 2015, 16:41 »
An infinite system has infinite everything if  you talk about extrinsic quantities (volume, mass, number of atoms, etc). Therefore only intrinsic quantities make real sense (density, atoms per unit cell, transmission per molecule, etc)

Offline ams_nanolab

  • Supreme QuantumATK Wizard
  • *****
  • Posts: 389
  • Country: in
  • Reputation: 11
    • View Profile
Re: Transmission normalization ?
« Reply #6 on: November 7, 2015, 16:56 »
Okay, so how do I get the current density?

Offline Anders Blom

  • QuantumATK Staff
  • Supreme QuantumATK Wizard
  • *****
  • Posts: 5565
  • Country: dk
  • Reputation: 93
    • View Profile
    • QuantumATK at Synopsys
Re: Transmission normalization ?
« Reply #7 on: November 8, 2015, 22:35 »
The current divided by the cross section - if you have a 3D interface. Otherwise if, as you indicated, you have a single molecule in the central region, the relevant is probably current per molecule, which already what you calculate.