QuantumATK Forum

QuantumATK => General Questions and Answers => Topic started by: Dipankar Saha on July 6, 2015, 11:51

Title: Tr (k, e)/shift in Fermi level
Post 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
Title: Re: Tr (k, e)/shift in Fermi level
Post by: Dipankar Saha on July 7, 2015, 16:38
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.... )
Title: Re: Tr (k, e)/shift in Fermi level
Post by: Umberto Martinez on July 8, 2015, 10:48
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
Title: Re: Tr (k, e)/shift in Fermi level
Post by: Dipankar Saha on July 8, 2015, 12:36
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
Title: Re: Tr (k, e)/shift in Fermi level
Post by: Dipankar Saha on July 13, 2015, 15:22
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
Title: Re: Tr (k, e)/shift in Fermi level
Post by: Umberto Martinez on July 14, 2015, 15:04
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
Title: Re: Tr (k, e)/shift in Fermi level
Post by: Dipankar Saha on July 14, 2015, 20:09
Thanks for your reply..... :)

Best_
Dipankar

Title: Re: Tr (k, e)/shift in Fermi level
Post by: Dipankar Saha on July 22, 2015, 22:30
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
Title: Re: Tr (k, e)/shift in Fermi level
Post by: Dipankar Saha on July 23, 2015, 21:33
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
Title: Re: Tr (k, e)/shift in Fermi level
Post by: Dipankar Saha on July 23, 2015, 21:40
Moreover___
Is there any dipole formation at zero bias (something like, getting a local current??)  ??!!! :-\
Title: Re: Tr (k, e)/shift in Fermi level
Post by: Jess Wellendorff on July 27, 2015, 09:00
Yes, we will usually plot E-EF. This convention makes it easy to read the plot.
Title: Re: Tr (k, e)/shift in Fermi level
Post by: Dipankar Saha on July 27, 2015, 09:38
Thank you Jess for the clarification...!!!  :)

Regards_
Dipankar
Title: Re: Tr (k, e)/shift in Fermi level
Post by: Dipankar Saha on August 8, 2015, 08:45
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
Title: Re: Tr (k, e)/shift in Fermi level
Post by: Jess Wellendorff on August 10, 2015, 09:15
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.
Title: Re: Tr (k, e)/shift in Fermi level
Post by: Dipankar Saha on August 10, 2015, 12:11
Many thanks Jess... / That's exactly what I wanted to know..!!! :)

Regards_
Dipankar
Title: Re: Tr (k, e)/shift in Fermi level
Post by: Dipankar Saha on August 12, 2015, 17:59
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
Title: Re: Tr (k, e)/shift in Fermi level
Post by: Jess Wellendorff on August 17, 2015, 22:29
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.
Title: Re: Tr (k, e)/shift in Fermi level
Post by: Dipankar Saha on August 19, 2015, 14:13
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

Title: Re: Tr (k, e)/shift in Fermi level
Post by: Dipankar Saha on August 19, 2015, 14:31
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 ??
Title: Re: Tr (k, e)/shift in Fermi level
Post by: kstokbro on August 21, 2015, 11:29
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.
Title: Re: Tr (k, e)/shift in Fermi level
Post by: Dipankar Saha on August 24, 2015, 15:44
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