QuantumATK Forum
QuantumATK => General Questions and Answers => Topic started by: Dipankar Saha on July 6, 2015, 11:51
-
For a "Spin: Unpolarized" calculation...while obtaining Tr (k, e) ..... I find it's only for "up-spin" component...!!! Whereas, all other calculations...e.g. n(r), del_n(r) etc. realted to that are done for "all spin"....!!! :o What about the Seebeck co-eff. value...that we get from the "thermo-electric coefficients" plug-in ?? Is it considering the up-spin only??
Moreover, how do we select....del_Ef (shift in Fermi level)....in the same plugin??
Regards_
Dipankar
-
What about the Seebeck co-eff. value...that we get from the "thermo-electric coefficients" plug-in ?? Is it considering the up-spin only??
Or, it's an Averagre ??!! (Obviously, for Spin : UnPolarized .... whether this is up-spin or Average...should not matter.... )
-
The thermoelectric coefficients plugins uses the transmission spectrum and for unpolarized calculations it uses spin sum (up+down) in the evaluation.
what do you mean with "how to select the energy shift"?
you can have a look at this tutorials:
http://quantumwise.com/publications/tutorials/item/515-isotope-impurity-effect-on-a-single-wall-carbon-nanotube
-
The thermoelectric coefficients plugins uses the transmission spectrum and for unpolarized calculations it uses spin sum (up+down) in the evaluation.
1) Meaning.....for unpolarized calculations : even if the Tr(k, e ) is showing .... only the "up-spin" component as the active curve....; but still the calculation of "G(e)" (as shown in the plugin) considers the sum.....!!! :)
Is it correct??
__________
2) In the same plugin .....while calculating the individual co-efficients....we have to set "T"...as well as "del_Ef" .../ What this shift in Fermi level is relating to..??!!
Many thanks Umberto...!!! :)
Regards_
Dipankar
-
Dr. Umberto Martinez,
Looking forward for your reply... :) / Can you please tell me...how this "shift in Fermi level" is set ...??!!
[Some thing that is mentioned as...0.04 eV.... in the tutorial, you referred above.../ What "band edge" it corresponds to??!]
Regards_
Dipankar
-
Well you can calculate your coefficients at various energies.
The energy of the band edge is measured from the transmission spectrum
See also http://quantumwise.com/documents/tutorials/latest/Phonon/index.html/chap.thermoelectric.html#sect1.thermoelectric.analyze
-
Thanks for your reply..... :)
Best_
Dipankar
-
Hello,
One more thing that I wanted to know is___
In a 2D plot.... say the y-axis is showing Tr(k, e)... while the x-axis is given in terms of Energy (eV)......
Is this energy scale itself represents (E-EF) ....??
Thanks & Regards_
Dipankar
-
Is this energy scale itself represents (E-EF) ....??
It will be of great help....if you can please reply... :)
_____
Besides, what happens exactly at 0 V bias (as well as, say there is no del_T) ....?? Does the chem. potential merge with the Fermi energy level?? But in ATK EF is always represented as the energy zero..!!!
Thanks & Regards_
Dipankar
-
Moreover___
Is there any dipole formation at zero bias (something like, getting a local current??) ??!!! :-\
-
Yes, we will usually plot E-EF. This convention makes it easy to read the plot.
-
Thank you Jess for the clarification...!!! :)
Regards_
Dipankar
-
Using VNL-ATK 2015...I f we obtain DDOS... for a spin unpolarized calculation...it shows a projected 2D plot....where the down-spin component is exactly a mirror image of the up-spin one.
The problem is that....... none of the objects of any specific *.nc file, obtained using 2015, is compatible with the ATK-VNL 2014.... Thereby, I'm not being able to check....that if I render that DDOS object to the earlier version...
whether I will get the 'Sum' (i.e. , spin-up+spin-down = 2 x spin-up .... for this case) in DDOS v/s. Energy curve... or, it will be showing only the up-spin component??
Regards_
Dipankar
-
Great question. In ATK, an unpolarized DDOS consists of a spin-up and a spin-down component that are exactly identical. You can check this with "Text Representation". Both channels are plotted in the 2D Plot widget, but one of them can be "spin-flipped". Therefore, the total DDOS is 2xspin-up contribution = 2xspin-down contribution.
-
Many thanks Jess... / That's exactly what I wanted to know..!!! :)
Regards_
Dipankar
-
Hello,
One more thing ...that I find a bit confusing is__
We may obtain the conductance (or, say thermal conductance due to the contributions of the charge carriers) with help the "plugin"..../ Whatever the value that we get...are they calculated under the linear aprrox. ??
Again, using the same transmission co-eff. we do calculate SCF-current also..., but I don't think there is any approximation of linear coherent transport...!!! Isn't it?!!
Best_
Dipankar
-
Yes, as explained in http://quantumwise.com/documents/tutorials/latest/Phonon/index.html/chap.thermoelectric.html (http://quantumwise.com/documents/tutorials/latest/Phonon/index.html/chap.thermoelectric.html), the VNL plugin "Thermoelectric Coefficients" calculates various electronic and vibrational transport coefficients in the linear response regime.
I presume that by "SCF-current" you mean the electrode-to-electrode current derived from a TransmissionSpectrum after SCF is converged. The last sections of http://www.quantumwise.com/documents/manuals/latest/ReferenceManual/index.html/chap.negf.html#sect1.negf.current (http://www.quantumwise.com/documents/manuals/latest/ReferenceManual/index.html/chap.negf.html#sect1.negf.current) indicate that the energy-dependent device transmission coefficient may be considered a sum of transmission amplitudes at each particular energy.
-
Thank you Jess...for the reply...!! :)
the electrode-to-electrode current derived from a TransmissionSpectrum after SCF is converged...
Not exactly that... I meant the calculations of Tr(e) for individual bias voltages..... / However...what that I was asking ....is somewhat related to the Fermi distribution func. of the electrodes. Meaning, if there is not any approximation of linear coherent transport (considering the case of I-V calculation).........then the fL,R(mu, TL,R) should not be approximated or, simplified...!!! Is it not??
Regards_
Dipankar
-
Besides, we know that the Tr_e is insensitive to the temp . change.... But G(E) shows a significant variation with temp. If it's a linear electronic conductance...in that case...the only way you can include temp. effect is through the L_0 ... (again, del_T is also zero)....!! Correct??
However, for the other calculations (under the same linear response approx.)...e.g, "S" or, "K_e"...a del_T value should be reqd. ......How do you incorporate the "del_T" ....given a particular temp. T, say, 100 K, 200 K or, anything ??
-
The temperature enters the formalism through the fermi function, so it is an electron temperature.
Forinstance, at T=0 the conductance is given by the transmission coefficient at the fermi level, while for a finite temperature the transmission is averaged around the fermi level using the fermi function.
-
Dear Dr. Kurt Stokbro,
Thank you very much for the reply.... :)
________
So, in Near-equilibrium..if we take Taylor’s series expansion....of the f(E, mu_l) -f (E, mu_R) ...then I have two terms... In one of the terms , (-df0/dE) needs to be multiplied by del_V .......and in case of the other one.... (-df0/dE) is multiplied by del_T
How should we incorporate this del_V....? Directly into the Landauer formalism?? Or else, right from the NEGF calculation of the Transmission function...??
Regards_
Dipankar