A question about transmission pathway

[wot19920302]

Dear Qunantumwise Staffs:
         In VNL, threshold in transmission pathway is described as "only include scaled tranmission above the given threshold. The transmission t are scaled as (t/tmax)**ρ, where ρ is the scaling power."  So scaled transmission t_scale= (t/t_max)**ρ. I notice that the default  ρ is 0.00. What does it mean? Let's take the example,  if  local bond transmission between atom A and atom B  t=0.4,  the t_max=1, when default  ρ is 0.00, the scaled transmission   t_scale= (t/t_max)**ρ=(0.4/1)**0=1.  That's to say, for all transmission pathway,   when  ρ is 0.00, t_scale always equals to 1 and the threshold doesn't make sense in that case. Should I set ρ=1.00?   In addition, from the definition of scaled transmission, the radius of arrow seems the relative value ( (t/t_max)). Does it mean that I can't compare conductance of two different devices? 

[wot19920302]

can anyone help me?

[Daniele Stradi]

Hi,

this scaling procedure is required because in the DFT-LCAO calculations QuantumATK uses a non-orthogonal basis set, which, naively speaking, means that all the basis set functions used to represent the atomic orbitals of the system interact with each other, if their overlap is non-zero (this is controlled by the finite-range of the basis set functions). As a consequence, some of the current paths, normally those with very small amplitude, are not really relevant for a physically meaningful analysis. The scaling parameter allows one to remove them and focus on those pathways which have larger amplitude, which are normally those with most physical meaning.

Since the transmission pathways are essentially a decomposition of the total transmission into local contributions, I would say that the amplitude of the pathways can be compared only if their transmission functions can be also compared.

Regards,
Daniele

[wot19920302]

         Thanks! I have understood  the first question  with your kind help . However, I am still confused by the second question.
          Fig. 1(a) and Fig. 1(b) are spin-resolved transmission spectra under zero bias voltage for device-A and device-B, respectivley.  From Fig.1(a) and Fig.1(b), it can be seen that transmission peak for device-B at E=0.25 eV  is larger than that for device-A.  Therefore, I want to use transmission pathway to explain why the value of transmission peak at E=0.25 eV for device-B is larger than that for device-A.   The calculated transmission pathway for device-A is shown in Fig.2(a) and for device-B is shown in Fig.2(b).  Plotting parameters for Fig.2(a) and Fig.2(b) are the same and  displayed in Fig.3.  Obviously, compared to Fig.2(b), Fig.2(a)  shows much more backscattering transmission (red arrows) in center molecule.  Could I say  such backscattering transmission leads to lower transmission peak for device-A than device-B at E=0.25 eV? 

[Daniele Stradi]

Hi,

it looks like that your system is formed by the same electrodes, and that also the central region is similar, apart from a central molecule having a different configuration in the two calculations. In this case, the two transmission calculations should be comparable. 

Regards,
Daniele

[wot19920302]

Dear Stradi,
           Thanks for your help. I have the last question: in transmission pathway, radius and cap radius  of arrow indicate size of scaled transmission. I wonder are such two radii proportional to the value of scaled transmission?  e.g. For one device, If scaled transmission=1, then radius of arrow=0.2. Does it mean that if  scaled transmission=0.5, the radius of arrow=0.1?
            Yours sincerely

[wot19920302]

anyone helps me?

[Ulrik G. Vej-Hansen]

Yes, I would say so. All properties that are scaled are scaled with the same scaling factor.

[wot19920302]

thanks for your help!