Spintronics exploits spin currents, i.e. electron currents in which a majority of electrons have the same spin and a minority have a spin oriented in the opposite direction.  These spin currents are often produced or probed using ferromagnetic materials whose magnetization imposes the direction of the majority spin. However, just as a classical electric current can be absorbed in an undesirable way (via leakage currents), a spin current can see its spin majority decrease under the effect of a "spin absorption".

Scientists wanted to know if the spin absorption in the vicinity of a ferromagnetic element depends on the orientation of the spin with respect to the magnetization of the material.

To do this, they fabricated a device using electron lithography that allows them to separate charge current and spin current and to access very sensitive measurements of microscopic magnetic states.  In their "lateral spin valve", a spin current is created in a copper wire, thanks to a first ferromagnetic electrode (made of Co, CoFe or NiFe), placed perpendicularly to the wire (above it) and then, after having propagated over a few hundred nanometers, it is probed thanks to a second electrode, identical to the first.

This device allows to measure the well-known spin absorption, in the classical case where the majority spin is parallel to the direction of the magnetization of the ferromagnetic material which probes it.

In another identically fabricated device, the physicists have introduced, halfway between the electrodes, a ferromagnetic nanodisc whose magnetization is perpendicular to that of the electrodes (and thus of the majority spin of the spin current). This second device allows, this time, to measure the spin absorption in the unusual case where the majority spin is perpendicular to the magnetization of the ferromagnetic material that disturbs it.

For the first time, the researchers show that the spin absorption depends very much on the orientation of the majority spin with respect to the magnetization of the material and specify quantitatively the values for the two different geometries (parallel or perpendicular spin). They thus determine the "spin mixing conductance", a difficult to access but fundamental parameter of spin transport.

This work was carried out in collaboration with the CNRS/Thales Joint Physics Unit (Palaiseau).