Energy level splitting in Projected DOS

[bharadway]

Hello,

Using the ATK-DFT (NEGF) solver of QuantumATK, I am observing a weird splitting of energy levels in the projected DOS of my device. I have simulated my NMOS nanowire down to 5 K and have plotted the energy levels that appear in the gate region of the projected DoS as a function of gate voltage with Vds = 1 mV (attached to this post).

As can be seen, the energy levels change with temperature, experiencing a large splitting at colder temperatures.

Is there a physical interpretation of this splitting? I know it can't be Coulomb Blockade as the NEGF formalism used in QuantumATK describes a time-independent electronic distribution in the steady state.

[bharadway]

Pls help!!

[Tue Gunst]

Hi,
What is the units on the axis and exactly what energy levels are you plotting? It is obtained from a peak in the pdos?
It seems that the data is more noisy at low temperature.
In general, the thermal broadening secures that less strict numerical sertings can be used. At least for k-point samplings which is however less relevant for a nanowire.
Other than that it is hard to say more without more information.
Best, Tue

[bharadway]

Thank you for your reply, Tue.

On the y-axis is energy in eV.

I've attached examples of energy levels I've observed in the DoS. Note how there are clear energy levels in the central region which coincides with the gate region. In the plot in my original post, I plotted the energies at which these levels appear.