Author Topic: Question about : the command in ATK2010 IV calculation.  (Read 3076 times)

0 Members and 1 Guest are viewing this topic.

Offline fangyongxinxi

  • QuantumATK Guru
  • ****
  • Posts: 143
  • Reputation: 0
    • View Profile
Question about : the command in ATK2010 IV calculation.
« on: December 18, 2010, 02:55 »
Dear Sir,
I have a problem about the calculate IV loop in ATK2010 manual ,page 108, as the attachment.
1.   I know the “device_configuration.update()” means run scf. In this IV calculation, there is no “device_configuration.update()”, and I suppose there is no “ restore the old scf ” in this scipt , right ? ( ATK 2008, IV calculation, restore the old scf, saving a lot of time).
If we want to save the scf for each " voltage",  and restore the old scf for the next step calculation, just like ATK2008 did,  how to change the script ?

2.  what does the “device_configuration.set_calculator(…., initial state = device_configration)” means ?
It means “ scf process” or “just collect the parameters” ?

In one word, this script confused me, hope to get a detail discription ( compared with ATK 2008)

Offline Anders Blom

  • QuantumATK Staff
  • Supreme QuantumATK Wizard
  • *****
  • Posts: 5565
  • Country: dk
  • Reputation: 93
    • View Profile
    • QuantumATK at Synopsys
Re: Question about : the command in ATK2010 IV calculation.
« Reply #1 on: December 18, 2010, 10:11 »
If you want, you can put in the "update" line, it will work fine, but will not change anything :)

The reason is that when you request ATK to compute the transmission spectrum, it will check if the calculation is converged. If not, it will first automatically do the scf loop, then compute the transmission.

The "set_calculator" part will do exactly what the old "restore and initialize from" did, i.e. it will update the bias voltage and use the old calculation as a starting guess for the new (it's a loop, so in each subsequent step, device_configuration is the "old" bias point), which as you noted improves the convergence rate quite a lot for finite bias. This line does however not run the calculation, it only prepares it.

Also see http://quantumwise.com/publications/tutorials/mini-tutorials/98-i-v-curve-and-voltage-drop

Clear? ;)