Author Topic: Vertical Resistance at Low Temperatures  (Read 1658 times)

0 Members and 1 Guest are viewing this topic.

Offline saligram

  • New QuantumATK user
  • *
  • Posts: 1
  • Country: us
  • Reputation: 0
    • View Profile
Vertical Resistance at Low Temperatures
« on: May 13, 2021, 23:00 »
Hi,

I am trying to study the effect of tempeature on vertical resistance. I followed the tutorial on the webinars page (based on this paper https://journals.aps.org/prapplied/abstract/10.1103/PhysRevApplied.13.044045). The one place where I had the option of changing the temperature was in the LCAO calculator (line 318). I changed it to 4K, but the transmission per spin at the fermi energy level, used to calcuate the conductance doesn't change ans is exactly same as that of 300K. Am I doing it right ot is there any other way to account for the temperature.

The following is the script generated (attached)

Kindly let me know.

Thanks


Offline Anders Blom

  • QuantumATK Staff
  • Supreme QuantumATK Wizard
  • *****
  • Posts: 5394
  • Country: dk
  • Reputation: 89
    • View Profile
    • QuantumATK at Synopsys
Re: Vertical Resistance at Low Temperatures
« Reply #1 on: May 14, 2021, 07:16 »
The temperature you change this way is only the electron temperature, which is primarily used to avoid having a step-function Fermi distribution, which makes convergence hard. So, this "temperature" is not physical, it's more a numerical "cheat", in a sense, although for some thermoelectrical calculations the electron temperature can be used to compute properties like the Seebeck coefficient.

What you need is instead to change the ion temperature, which is a bit harder. First of all, you need to consider "what is temperature" in this context, or more precisely what is the effect of temperature on \resistance. This is well known, of course - in short, the atoms vibrate, which means they no longer form a perfect crystalline lattice, which causes scattering, and that is where "resistance" basically comes from.

So the way to capture the influence of temperature in calculations like this is to perturb the atoms from the perfect lattice. This should be done in a systematic way, and fortunately QuantumATK has precisely such a method, called "Special Thermal Displacement" - see https://docs.quantumatk.com/casestudies/std_transport for an example!

The main complication when using this method is that you need to evaluate the phonon spectrum of the system. If you have a simple metal and looking at grain boundaries, say, this is not such a big problem, as there are usually classical forcefields that can be used. But for more general situations (even a "simple" combination of two materials, like Ru and Ti as you have), one may need to resort to DFT for the phonons, which is a bit time-consuming (understatement). Or, find a way to combine individual potentials, or fit a new potential which is also possible with QuantumATK, but that's a whole other discussion.

Btw, since your initial structure was basically a combination of perfect crystals, what you actually already have computed is in fact the low-temperature case. It's the high-temperature situation that you obtain via the thermal displacements.
« Last Edit: May 14, 2021, 07:19 by Anders Blom »