QuantumATK Forum

QuantumATK => General Questions and Answers => Topic started by: Yueyang on October 5, 2016, 16:03

Title: Spin transmission spectrum messed up
Post by: Yueyang on October 5, 2016, 16:03
Hi,
I got a fake spin transmission spectrum,
half of the spin up spectrum is regraded  as spin down spectrum by the software,
and half of the spin down spectrum is regarded as spin up spectrum.
The data messed up!

compare the two figures below, they are obtained by changing the iteration tolerance from 0.005 to 0.002.
Title: Re: Spin transmission spectrum messed up
Post by: Petr Khomyakov on October 5, 2016, 19:07
What do you mean by "fake"? You have an individual transmission spectrum for each of the two spin channels. 

If your transmission spectrum depends on the tolerance parameter, it means that your calculation of the transmission curves is not converged yet. You have to decrease the tolerance parameter to the values for which this dependence virtually vanishes.
Title: Re: Spin transmission spectrum messed up
Post by: Yueyang on October 6, 2016, 05:24
Thank you, I will try, But I don't think it is the problem of tolerance
Title: Re: Spin transmission spectrum messed up
Post by: Daniele Stradi on October 6, 2016, 11:06
Hello,

to have a clearer answer, you should provide additional details such as an input script with the system you are studying.

Daniele
Title: Re: Spin transmission spectrum messed up
Post by: Anders Blom on October 6, 2016, 17:26
half of the spin up spectrum is regraded  as spin down spectrum by the software,
and half of the spin down spectrum is regarded as spin up spectrum.

No, that's not the case. Probably your up/down states are very close in energy and changing the tolerance flips them with respect to each other, and that causes the change in the transmission curves.