Messages

1

General Questions and Answers / questions about modeling Ni-Graphene-Ni MTJ

« on: March 26, 2013, 05:33 »

Hi, I developed a model to calculate the TMR of NGN MTJ as attachments showed and got results as follows,

Conductance Parallel Spin (Siemens)
Up=1.98e-05, Down=4.09e-05
Total = 6.08e-05
Conductance Anti-Parallel Spin (Siemens)
Up=2.20e-05, Down=2.20e-05
Total = 4.39e-05
TMR (optimistic) =    38.35 percent
TMR (pessimistic) =    16.09 percent

when I add double gates into the model and the gate voltages are both set to 0V, I get results like this,

Conductance Parallel Spin (Siemens)
Up=3.29e-05, Down=3.82e-05
Total = 7.11e-05

Conductance Anti-Parallel Spin (Siemens)
Up=3.12e-05, Down=3.12e-05
Total = 6.23e-05

TMR (optimistic) =    14.10 percent
TMR (pessimistic) =     6.59 percent

Maybe the absence of periodic boundary of the model caused the decrese of TMR, so I increase the width of the graphene barrier to get a TMR that closer to 38%. However, I get a result like this,

Conductance Parallel Spin (Siemens)
Up=4.80e-05, Down=5.79e-05
Total = 1.06e-04

Conductance Anti-Parallel Spin (Siemens)
Up=5.34e-05, Down=5.34e-05
Total = 1.07e-04

TMR (optimistic) =    -0.86 percent
TMR (pessimistic) =    -0.43 percent

I just simply increased the barrier width, and got a negative TMR, what is the problem? Is the model improper? The model is in the attachment, need your help, thanks for your time.

2

General Questions and Answers / Is this a proper Ni-Graphene-Ni MTJ model?

« on: March 4, 2013, 15:47 »

Dear Blom,

Many thanks for your answer to my prior questions, now I set up an improved model for my study as the attachment. This configration includes two gate in A and remains periodic in B. Is this geometry expected to get a right results?

Additionally, with this configration can I use periodic boundary conditions for the Poisson solver? Or I have to use Neumann boundary for both A and B?

Thanks for your time, best regards.

3

General Questions and Answers / a new model for Ni-Graphene-Ni MTJ with Gate Voltage

« on: March 2, 2013, 14:23 »

Hi, I have set up a new model to study the question as decribed in
http://quantumwise.com/forum/index.php?topic=2201.0
In this model  DFT and Neumann boundary condition are used. And and I want to calculate the TMR vs different gate voltage.  Now it should be an interface structure, and there is a piece of Graphene in the middle Ni.  The model is periodicity out of the screen plane.
Should this model be correct? Thank you.

4

General Questions and Answers / Re: about Ni-Graphene-Ni MTJ device with a gate voltage

« on: March 1, 2013, 04:57 »

Dear Anders Blom,
Many thanks for your answer and now I get what you mean. According to the present caulculation results, the TMR of this struture depends on the gate voltage. But this is not the question I meant to study.
The Ni-Graphene-Ni MTJ is an interface structure, in this case, what kind of model should I develop to consider the effect of an external electrical field on the TMR?
Should I use the model showed as x.jpg in the attachment, then how to deal with the external electrical field?
Thanks for your time, best regards.

The main problem is that it's not really a interface structure anymore, it's more something like a Ni nanowire with a few carbon atoms in the middle. I can't see from the graph it there is periodicity out of the screen plane (I assume there is), so perhaps it's more like two thin Ni slabs with a nanoribbon in the middle.

The primary question is more perhaps if this corresponds to anything that has been (or could be) measured experimentally. If not it can of course still be an interesting exercise to compute it, but the results are perhaps less useful then.

5

General Questions and Answers / about Ni-Graphene-Ni MTJ device with a gate voltage

« on: February 28, 2013, 12:51 »

Hi, I have devoloped a Ni-Graphene-Ni MTJ device with a gate as the attachment. I try to study the spin-dependent properties of the Junction and then calculate the TMR vs gate voltage.  In the model, DFT and Neumann boundary condition are used. However, is the model reasonable? Can someone help me? Thank you.