Is the IV curve correct?

[lknife]

Dear experts,

I got my IV curve using ATKSE:Extended Huckel method, seen in the attached figure. I have a question on the spectral current. Since the energy did not cover the current region (the blue area), I am not sure if the IV curve is correct. Can anybody here help me with this problem?

Thank you very much for your help!

[Petr Khomyakov]

There is no problem. The blue-colored area just shows you the difference between the chemical potentials in the left and right electrodes for a set of applied voltages, e.g., E_Fermi_left = E_Fermi_right (blue line becomes a point) for V_bias=0 Volt, but the spectral current is defined in the energy range from -infinity to +infinity, in principle. In practice, one always sets upper and lower limits for this energy range. You may take a look at our tutorials and manual on the transport calculations, see http://docs.quantumwise.com/tutorials/atk_transport_calculations/atk_transport_calculations.html,
http://docs.quantumwise.com/manuals/NEGF.html and many other examples at http://docs.quantumwise.com/.

[lknife]

OK, I see. Thank you very much for your kind reply. I was previously confused with this statement in the tutorial http://docs.quantumwise.com/tutorials/atk_transport_calculations/atk_transport_calculations.html that, "Pay special attention to the spectral current: All possible contributions to the total current must be included inside the specified energy window, in order to get accurate I(V) points. If this is not the case, you should consider redoing the analysis using a wider energy window."

From my spectral current plot, although the current was not distributed symmetrically to energy (maybe due to the characteristics of a p-n junction), it was fully covered by the energy range. Thus, the IV curve should be OK. Is it right?

[Petr Khomyakov]

The statement made in the tutorial is correct, in principle. As a rule of thumb, the energy range chosen for calculating the IV curve must be much larger than the chemical potential difference.  If you are in doubt, you need to do a convergence test by increasing the energy range to see how the IV curve changes. Otherwise, the conclusion on the IV curve correctness will be based on a belief, rather than on a calculation.

[lknife]

Thank you for your reply. It's really helpful.

I want to know if there is a easy criterion to help determining whether the energy range is suitable or not.  If one just use a large enough energy range for the calculation every time, it should be very time-consuming and actually it's no need to do like that sometimes. 

However, if one used a relative smaller energy range, it might be not enough. It's really a daunting and frustrated job if in any case I need to repeat the time-consuming calculation many times. Taking the spectral current plot in my question as an example, can I roughly tell that the energy range is suitable?

[Petr Khomyakov]

This depends on the voltage and temperature as both enter the Fermi-Dirac distribution in the equation for the current, see http://docs.quantumwise.com/manuals/NEGF.html.

For example, the energy range is to be exactly given by the chemical potential difference at T=0K, but it needs to be extended upon increasing the temperature to account for the tails of the Fermi-Dirac distribution, which decays exponentially.  So, the energy range does not need to be too huge, but I cannot tell you exact criterion because it also depends on the transmission spectrum.

I can guess that for your particular problem the energy range chosen looks good enough, assuming that the temperature is of 300 K -> 0.025 eV.

[lknife]

Thank you for your kind reply. I really got something from your help.