Patent Application
20130175588
The subject matter of this patent application was developed without federal or state funding. Although both inventors named herein are the beneficiaries of support from various corporate organizations, and are employed by Universities at which they work on subject matter including the development and characterization of materials formed by the direct growth of graphene on various substrates, no University time, compensation or apparatus was employed, involved or relied on in the development of the subject matter disclosed and claimed herein.
1. Field of the Invention
The present invention relates to the development of a non-local spin valve, sometimes referred to as a spin field effect transistor, or spin-FET. Spin-FETS that can be operated at ambient temperatures offer the advantages of greater data storage in less space, with a reduced power consumption, and high sensitivity. Spin-FETS offer non-volatile solid state data storage the does not require the constant application of current to sustain it. Spin-FETS are, accordingly, an attractive technology to be used in the development of Magnetic Random Access Memory, or MRAMs. Spin-FETS also find application in logic devices, or combined logic and spin devices. Richter et al, Appl. Phys. Lett., 80 1291 (2002). Among the least consequential, but easily understood, effects of such advances is the ability to provide computing devices which need not consult a stored database before uploading, as most personal computers are arranged today. They offer an “instant on” capability: that is to say, nonvolatile memory and logic.
2. Background of the Technology
The spin-FET may have been first described as a desirable device to construct in 1990 (Datta et al, Appl. Phys. Lett., 56, 665 (1990)). These devices contemplate a non-magnetic layer which is used for transmitting and controlling the spin polarization of electrons from source to drain. A variety of field materials have been proposed over the years. Most of the concentrated effort in this field has looked at spin injection as a means of arriving at the capabilities offered by the theoretical spin-FET as described. Notwithstanding this effort, no room temperature spin-FET, that is scaleable, and reliable at low power consumption has yet been provided.
In one approach, the ability to deposit thin layers of cobalt on semi-conducting substrates such as GaAs was developed, in an effort to provide for a controlled magnetization perpendicular to the substrate. Bournel et al, Physics E, 10, 86-90 (2001). Ultimately, these researchers were unable to secure a stable, perpendicular magnetization for a ferromagnetic/semiconductor contact field, but research did demonstrate the ability to grow thin layers of oxidized Co on semiconductors such as gallium arsenide and silicon. As investigations into the provision of spin-FETs continue, the ability to adapt the resulting technology to existing materials will become increasingly important.
Our invention resides in a new type of non-local spin valve or spin field effect transistor (spin-FET), based upon graphene deposited directly onto a magnetically polarizeable electrically insulating or semiconducting substrate such as an oxide, and with the oxide dielectric, with net interface spin polarization, formed on top of a base ferromagnetic layer. In essence, the polarization of the substrate will result in polarization of the transition metal ions of the oxide dielectric at both the top and bottom interfaces, which will in turn lead to polarization of the graphene valence/conduction band electrons. Using appropriate magnetically polarizable source and drain electrodes, a low power, high on-off rate spin-FET can be constructed gated by the bottom base ferromagnetic layer, separated from what amounts to a narrow conduction channel by a dielectric oxide. An example of this is shown in
In developing this new or non-local spin-FET, although the applicants do not wish to be bound by this explanation, ferromagnetic polarization of the base layer, or gate, is effective to polarize the metal ions in the magnetic oxide at both the top and bottom interfaces. At the top interface, the one in intimate contact with the graphene conduction channel, these ions will undergo exchange interactions with the graphene valence/conduction band electrons. Such “proximity polarization” has been predicted for graphene [1]. Strong ion/conduction electron exchange interactions have been observed in impurity-doped oxide quantum dots, resulting in magnetic behavior in closed shell systems [2-4].
In elementary terms, in FET operation, when the polarization of the source and drain are in the same direction as that of the graphene, there will be a low non-local resistance (high current) between the source and drain. Switching the polarization of the base layer (or gate) will change the polarization of the graphene, and if it no longer matches the polarization of the source or drain, the non-local resistance will be much higher. This provides the basis for the non-local spin-valve, or coherent spin-FET.
In a “conventional” graphene-based spin valve (
In contrast, the inventive coherent spin-FET's properties derive from the fact that magnetic exchange interactions between oxide conduction electrons and oxide cations will result in the formation of magnetic polarons (
The accompanying drawings, which are incorporated herein and constitute part of this specification, illustrate exemplary embodiments of the invention, and, together with the general description given above and the detailed description given below, serve to explain the features of the invention.
Referring initially to the schematic illustration in
Graphene
this film must consist of (111)-oriented, sp2 carbon (graphene), as either a single layer or several layers, as desired to control potential oxide/graphene interactions such as charge transfer [10, 11]. This layer can be deposited by molecular beam epitaxy, or possibly by chemical or physical vapor deposition.
The deposition of graphene on a substrate has been described at some length in U.S. patent application Ser. No. 12/543,053 and U.S. patent application Ser. No. 12/980,763, both of which are included herein-by-reference. Additional advances in the control over few layer graphene deposition are provided in U.S. Provisional Patent Application Ser. Nos. 61/490,650 and 61/497,971, both of which are incorporated herein-by-reference. The controlled direct growth of graphene by MBE (layer-by-layer growth of macroscopically continuous graphene sheets on Co3O4(111) at 1000 K by carbon molecular beam epitaxy (MBE) from a graphite rod source) is described in detail in U.S. Provisional Patent Application Ser. No. 61/522,600. The disclosure of this pending application is incorporated herein-by-reference. Any of the methods described in the incorporated applications can be used to form the graphene field, with a preference for controlled molecular beam epitaxy.
Magnetic Oxide, Source and Drain:
This material electrically isolates the graphene from the ferromagnetic gate layer, and allows polarization of the graphene valence/conduction electrons via polarization of the cations in the magnetic oxide. Potential candidates include Co3O4(111), Fe3O4(111), NiO(111), and potentially spinels such as CoFe2O4(111), as well as Cr2O3(111), BaFe2O4. A critical feature is the polarization within the ion layer adjacent to the graphene (
Note, since NiO(111) has the same rocksalt structure as MgO(111), deposition of graphene may result in oxide reconstruction, destroying the chemical equivalence of graphene A sites and B sites [12], removing the HOMO/LUMO equivalence at the Dirac point and opening a band gap, as set forth in U.S. patent application Ser. No. 12/980,763. Development of bandgap potential in graphene bearing materials may provide important electronic advantages, in addition to preserving spintronic adaptability.
Ferromagnetic Gate:
This layer should have its axis of magnetic polarization that is easily switchable and ideally has a low coercive field for multifunction logic gates, and be ferromagnetic at room temperature. Appropriate materials include Co, Ni, or Fe.
In order to prepare a coherent spin-FET of the invention, a sapphire (aluminum oxide (0001)) substrate is prepared for deposition. An electron beam evaporator may be used to reduce movement between chambers and improve productivity, switching in various targets for deposition. Thus, a fifty angstrom layer of cobalt may be deposited under conventional conditions on the substrate at 750° K in UHV. This deposition is followed by an oxidation anneal at 1000° K which results in surface segregation of dissolved oxygen and the development of a thin layer of Co3O4 (111) (may be 2-5 ML thick). Graphene (2 or 3 ML) is deposited on the Co3O4 using molecular beam epitaxy at 1000° K, yielding a macroscopically continuous graphene film of approximately 3 ML thickness. Graphene may also be deposited via CVD and PVD processes, as disclosed, but MBE is preferred, not only because of the fine control and developed information for this method, but because it is compatible with the other process steps in the formation of the coherent spin-FET of the invention. This leads to high productivity.
In this example, the coherent spin-FET is finished with the application of Co source and drain, and connected in the device for use. On magnetization, it is stable and exhibits a very high on/off rate with low power consumption at temperatures substantially above room temperature. Thus, referring to
While the present invention has been disclosed both generically and with reference to specific embodiments and examples, these alternatives are not intended to be limiting unless reflected in the claims set forth below. The invention is limited only by the provisions of the claims, and their equivalents, as would be recognized by one of skill in the art to which this application is directed.
