2.2. Future information technologies

2.2.4. Spintronics

The aims of spintronics are the manipulation of electronic spin for the storage, transmission, and processing of information. Spintronics is promising to revolutionize data storage and processing in a development which commences with smaller devices that consume less power than present charge-based ones and may culminate in the implementation of quantum computing. Spintronics demonstrated first success already in metallic multilayers, most notably the oscillatory long-range magnetic coupling and its applications, i.e., the giant magnetoresistive reading head and the magnetic random access memory. The central unit of these devices is a quantum cavity for electrons and spin. The boundary layers are ferromagnets which impose their spin dependence on the electronic states inside the cavity.

Second-generation spintronic devices shall directly use spin-polarized currents. To generate these spin currents, spin injection from ferromagnets into semiconductors is being studied intensively, but the direct interfaces between these materials suffer from severe conductance mismatch problems. It is principally possible to create a spin current without ferromagnetic materials or external magnetic fields altogether if the polarization is created by virtue of the spin-orbit interaction instead of the Pauli principle. The spin-orbit interaction induced by an electric field E at a two-dimensional interface, i.e., the Rashba effect, has been studied intensively in semiconductor heterostructures and at metal surfaces. The main concept for a spin-field-effect transistor, however, still requires ferromagnets as polarizer and analyzer, and only the active element requires the Rashba spin-orbit interaction to perform spin rotation. The stronger the Rashba effect, the shorter is the distance required for spin rotation which is important for miniaturization.

The Rashba spin-orbit interaction has in recent years increased strongly. Giant values of the Rashba parameter _R on the order of 10¹⁰ eV Å have been reported for a Bi-Ag surface alloy (3·10¹⁰ eVÅ), for an Ir(111) surface state (1.3·10¹⁰ eVÅ), and for the three-dimensional system BiTeI (3.8·10¹⁰ eVÅ). On graphene intercalated with Au, also a giant spin-orbit splitting is observed 4 orders of magnitude larger than the intrinsic values of grapheme (Figure 14).

Figure 14. Left: Spin-resolved photoemission from a 4-monolayer film of Au on W(110). With increasing emission angle, the Rashba splitting of a Au quantum-well state increases. Right: Graphene on Ni(111) intercalated with a monolayer Au shows a giant Rashba splitting of ~100 meV on the p state near the Fermi energy. Following its dispersion through the hybridization with Au d-states, the origin of the spin polarization can be traced to the d-states. (Courtesy O. Rader, HZB).

The next challenge is to transfer these to a semiconducting or insulating platform. This has been achieved for Pb and Bi on Si. Also topological insulators, which ideally are insulating in their volumen, form extra Rashba-split states at their surface.

Similarly as in section 2.1.3, the time structure of BESSY^VSR will allow for angle-resolved time-of-flight electron spectrometers to be employed (ArTOF type). The gain in efficiency will lallow for a more wider useage of spin detectors.