In the production of electronic devices based upon the principles of spintronics, that is, using the location and sign of the spin of the electron rather than its charge as the pre-eminent factor under control, it is possible to include in such devices elements termed ‘spin valves’. Spin valves conventionally function by controlling the ability of one part of the valve, which forms part of an electrical circuit, to pass a spin-polarised electrical current, or not. This control is effected by other parts of the valve, which typically create and change magnetic fields in such a way as to allow or impede the spin-polarised current in the conducting part.
Such devices are known in ferromagnetic metallic systems, and involve two metallic ferromagnetic layers, the one controlling the magnetic state and thus the current flow in the other: such devices are currently commercially available as ‘giant magneto-resistive’ (GMR) elements in e.g., read heads employed with magnetic recording media. Analogous devices are also known and have been described in ferromagnetic semiconductor systems, and devices have also been made employing two ferromagnetic layers, where the first ferromagnetic layer is a metallic system and the second ferromagnetic layer is a non-metallic system. The effect can also be used in TMR (tunneling magneto-resistive) devices in spintronics.
However all such systems hitherto described actually consist of three layers, being the two magnetic layers separated by a non-magnetic ‘barrier’ layer. This barrier layer is essential in all such conventional systems, and serves to magnetically separate the two magnetic layers so that the interaction between the two magnetic layers is controllable, and so they do not act magnetically as one single layer. This barrier layer is typically composed of copper or similar in metallic GMR samples, an insulator such as AlOx, in metallic TMR structures, or an undoped semiconductor in Semiconductor TMR devices.
In the present disclosure, a novel effect has been observed, wherein two ferromagnetic materials, one metallic and one semiconductor, e.g., permalloy (NiFe, abbreviated: Py) and GaMnAs, directly deposited the one on the other, can be switched independently. This is a very interesting effect, and is believed to arise from the fact that the carriers in each material (electrons for NiFe, holes for GaMnAs) are different, and so bringing the two layers in direct contact does not lead to the two layers acting magnetically as a single layer.
It is also a commercially useful effect, as the non-magnetic interlayer previously thought necessary for such devices can be discarded: a two-layer device could be cheaper, faster, have higher efficiency, and have better signal to noise characteristics.
The charge transport for magnetoresistance phenomena, which gives a different resistance for such GMR/TMR devices depending on the magnetic orientation of the layers, as in a traditional GMR/TMR device, is dependant on the nature of the interface: for devices where the transport through this interface has an ohmic character, it would yield a GMR-type structure, whereas if a Schottky or p-n barrier is present at the interface, the device would act as a TMR.
Beyond the above, it is possible to set up in GaMnAs and other systems states in which the magnetizations of the two layers are neither parallel nor antiparallel, but which have more complex geometrical relationships; the simplest of these involve the two magnetizations remaining in the plane of the material layers but being offset by a certain angle, e.g., 90°, and the more complex of which involving magnetizations not in the plane or planes of the materials (one or both). Such more complex geometrical cases lead to operational behaviors where the system has three or more stable states, in comparison to the two stable states of hitherto known devices. These three or more states can be used directly for more complex computations than the essentially binary devices hitherto described. In addition, if the lower level is made of a material which exhibits tunneling anisotropic magnetoresistance (TAMR), then a TAMR component may also be present, potentially increasing the number of operable states even further.
The invention will be illustrated in connection with a detailed description of embodiments shown in the drawings.
We now provide a detailed technical description of one embodiment of the device in question. The sample consist two active layers deposited on a standard GaAs substrate and buffer. The first layer grown on the buffer is a thin film of GaMnAs grown by MBE. This is followed by a ˜2 nm layer of Py deposited in-situ onto the GaMnAs (i.e. the sample is transferred from the MBE growth chamber to the Py sputtering chamber under UHV conditions). The Py is deposited by magnetron sputtering, creating a magnetic anisotropy in the layer. The Py layer can especially be chosen between 1 and 5 and preferred between 1.5 and 2.5 nanometer thickness.
The bulk material is first characterized by SQUID magnetometry to confirm that the magnetization direction in each layer can be independently modified. This is put into evidence in
Finally,
We now turn to a transport characterization of the sample, which is put into evidence in
It is also interesting to note that preliminary measurements suggest that the part of the MR which comes from the Py layer may survive past the Curie temperature of GaMnAs.
This survival of part of the effect above the Tc of the GaMnAs, suggest that it is related to the Py, either because of an intrinsic property of this layer, or by the action of the Py on the Mn atoms in the semiconductor and may represent a way of pushing TAMR above room temperature.
The legend on the right of
Reference numeral 110 in
Reference numeral 130 in
Filing Document | Filing Date | Country | Kind | 371c Date |
---|---|---|---|---|
PCT/CH06/00488 | 9/12/2006 | WO | 00 | 8/4/2008 |
Number | Date | Country | |
---|---|---|---|
60716075 | Sep 2005 | US |