About current definition in SOT-driven skyrmion simulation

[Sandhya Ravi]

Hello MuMax3 Community

I am running a skyrmion racetrack simulation in MuMax3 using spin–orbit torque (SOT) parameters. I defined the current in my script as:

J = vector(0,0,2e11)

and set the SOT parameters:

SOTxi := -2
AlphaH := 0.15
Pol := alphaH
Lambda := 1
Epsilonprime := alphaH*SOTxi/2
Fixedlayer := vector(0,-1,0)

I would like to clarify:

(i) In this setup, is J treated as a charge current in the heavy metal layer, which        generates SOT, or as a spin current directly?

(ii) Are the SOTxi and AlphaH parameters applied correctly for modeling spin–orbit torque in a multilayer system?

I would greatly appreciate any guidance on interpreting the current and SOT parameters in MuMax3.

Thank you .

[Josh Lauzier]

Hi,

In this use-case, J is not directly a spin current. It is a charge current that is used to get the proper spin current. However: the heavy-metal layer and the electric to spin conversion is not explicitly modeled in mumax. J is just being chosen to give the correct spin current. In addition, if you're reusing the built in slonczewski STT and not custom fields in order to model SOT, it only looks at J_z, so one needs to be careful with that.

SOTxi is a material parameter depending on your system, and is determined by the ratio of your adiabatic vs non-adiabatic torques (a_j and b_j terms). Similar, alphaH is a material parameter (the spin hall angle). 

To see the full definition of how these terms are used, check out the mumax workshop here, in particular section 4 starting around slide 39 or so. It shows the full derivation, so you can see why certain parameters as set this way for SOT. Also refer to the 2 papers referenced on slide 41, which explain in more depth what things like Xi mean physically. 

Best,
Josh L.