QuantumATK Forum
QuantumATK => General Questions and Answers => Topic started by: Dong on August 31, 2010, 16:04
-
Dear everyone,
I can't calculate the magnetoresistance ratio (MR) for my two-probe system. I noticed that there several papers that calculate MR using atk, such as [PhysRevB.81.233403], [APPLIED PHYSICS LETTERS 96, 042104], [PhysRevB.78.224418] ... (you can search 'magnetoresistance' in the publication list of atk). These papers use the same method, calculating the currents/transmissions by seting the electrodes to be spin parallel (P) and antiparallel (AP). Also, there are similiar examples in the manuals/tutorials, like 'Mini-tutorial 2' in http://www.quantumwise.com/documents/tutorials/UpgradeGuide_10.8/XHTML/ .
My two-probe is similar to this 'Mini-tutorial 2'. I also use the atomic carbon linear chain as the electrode, but the bond distance is 1.27 angstrom (I get this value by minimizing the total energy and it also can be found in published papers), not 2.90 angstrom in the tutorial.
The problem is, I can only set the initial spin polarized states of electrodes to be P or AP. After scf steps, they (P and AP) both converged to a unpolarized state. SO I can't get the currents/transmissions for P and AP.
Are there any other methods? ???
Can anyone help me? :)
Thank you!
-
I am not perfectly sure, but let me try to give a guess anyway.
In short: Your are doing it right and the programs behaves correctly. If you make a carbon wire as it would be in nature,
it would be spin-unpolarized, hence if you make a carbon wire with the correct lattice constant as you do, you will
end up with the true ground state of the atomic chain.
If the atoms in the atomic chain is very long from one another, then the physics is sudden gone, since instead of atomic chain
you will have a string of very weakly, almost independent, atoms, and this system can be made spin-polarized,
and due to the very weak interaction between the atoms it will preserve the initial spin-configuration and can therefore have
a magnetoresistance
-
Is there any other method to calculate the MR for my system using atk?
-
Your system can not show any magnetoresistance, since it is not a magnetic system.
In order to be magnetic, it must be spin-polarized and atomic chain of carbon atoms is not spin-polarized.