C2N – Spintronics without magnetism

Towards spintronics without magnetism

Charge Electronics is based on semiconductors, using Si and InP substrates for transistors. Spintronics oriented research is so far using magnetic metals, semimagnetic semicnductorsa or semiconductors with magnetic contacts .The spin effect is observed with rater involved optical experiment (two photon absorption, Kerr rotation in III-V alloys. In this project a significant edge spin polarization is measured at room temperature by conventional transport in a highly doped III-V semiconductor heterostructure, with the electron system in a 65 nm wide quantum well. This is measured as a transverse voltage V_y, in an optical lithography designed micrometric Hall bar, excited by a longitudinal current I_x. The spin momentum S_z is quantized perpendicularly to the sample surface. No external magnetic field is applied to get this Hall-like transverse voltage. I am sorry not to disclose the material stacking used in this successful study, because further MBE growth are prevented during many months by the lab move out to a new building followed by growth reactor cleaning and couples of weeks high temperature backing.

The figure on the left shows the calculated spin-split Fermi surfaces of our MBE grown sample The spin-orbit splitting is assigned to the Dresselhaus effect, resulting from the lack of spatial inversion symmetry in the structure materials. The Dresselhaus parameter A43 is obtained from k.p theory for the electron system host material. However the full Shrödinger-Pauli Hamiltonian with the spin-orbit coupling term exhibit time reversal symmetry. Therefore this energy contour must exhibit Kramers degeneracy i.e. E(k, up)=E(-k, down). Because the band are significantly split, this imply the occurrence of 2 Discontinuity Points for a given spin orientation, indicated by DP labels and black arrows on Fig.1. We believe that these points occur along the direction of applied current.

Fig.2 Measurements of the transverse Hall voltage as a function of applied current on the Hall bar. The load resistance in serial with the sample is equal to 5k ohm (on the red curve) for the largest current and is increased by this value at each lower current point up to 50k ohm. The points on the green curve are obtained exchanging the 2 current connections with the 2 voltage connections.

The transverse voltage is at least one order of magnitude larger than possible resistance components mixing. We believe that these data prove that the linear transverse spin current idecreases linearly in time being depolarized by spin rotation under the spin orbit coupling induced internal magnetic field. Exchanging current and voltage pads, the transverse voltage decays and the current has to be increased by a factor corresponding approximatively to the ratio L/l, L length of the Hall bar, l distance between opposite voltage contacts ! with an accuracy of 30%.

Collaborations : Ulf Gennser and Antonella Cavanna for the MBE system conditioning.
Hourari AbdouSouleïman for the optical lithography and optimization of the contact metallization to get both ohmicity and good condition for wire thermocompression ( heat+US).
A publication is in preparation : Phenomenology of the intrinsic Hall effect in parametric III-V Epilayers.