Thank you Nick for answering me.
 When they said in the paper we need convergence to the number of points
grid, which grid points they mean, is it the C or L contour.

Regards.

On Friday, September 20, 2013, Nick Papior Andersen wrote:

> They are not explained as they are not used anywhere in transiesta, maybe
> they where used in older versions.
>
> I have a very short answer about the convergence, I still think the paper
> is quite valuable in their explanation.
>
> As transiesta performs an integral with finite integration points you need
> to converge the number of integration points, for both the equilibrium
> contour and the non-equilibrium contour. Your biggest problem is the
> non-equilibrium contour as that cannot utilise the complex contour method
> and thus the integral has to be very fine close to the real axis. So here
> you need many points.
>
> Again, the paper provides details about each contour part. The manual
> links the different contour parts to transiesta variables.
>
> Kind regards Nick
>
>
>
> 2013/9/19 mahmoud ali <[email protected]>
>
>> Thank you Nick for your good answer.
>> I have found two parameters not explained in transiesta manual which I do
>> not know how they could affect the results, they are:
>> TS.RealContour.Emin
>> TS.RealContour.Emax
>>
>> Nick I have asked you before about the convergence tests in transiesta
>> and you have told me about the paper, but unfortunately I did not find what
>> I looking for.
>>
>> Regards
>>
>>
>> On Wed, Sep 18, 2013 at 2:10 AM, Nick Papior Andersen <
>> [email protected]> wrote:
>>
>>> Yes, that is transiesta shifts the Fermi levels of the systems to zero,
>>> and then it adjusts the fermi level by V/2 in both electrodes (with
>>> appropriate signs).
>>> This is a design choice of transiesta.
>>> The scattering regions Fermi level is shifted to be 0 so that the Fermi
>>> levels are "aligned" according to the bias.
>>>
>>> Kind regards Nick
>>>
>>>
>>> 2013/9/17 mahmoud ali <[email protected]>
>>>
>>>> Dear Transiesta Users,
>>>> As described in the manual, when there is a finite bias, the Fermi
>>>> energy of the left electrode is placed at V/2, and that of the right
>>>> electrode at -V/2.
>>>> So all systems will have the same position of fermi energy at V/2 for
>>>> LE and -V/2 RE??
>>>> Is this correct?
>>>>
>>>> Yours
>>>>
>>>>
>>>>
>>>
>>
>

Responder a