Most computing technology relies on the presence or absence of charge to represent binary information during computation. Limitations in reducing the size and energy consumption motivate exploring other fundamental information carriers such as electron spin. Magnetism, which is a macroscopic consequence of spin, is proven as storage method in hard drives and their "read" sensors use spin to transduce magnetic fields to current (charge). Yet directly utilizing the electron spin for information processing (as opposed to information storage) remains unrealized. Spin information tends to be very fragile and materials defects and non-ideal interfaces easily interfere with the information state. In this project, we implement new materials into classical spin devices towards predictably generating, transmitting and detecting well-characterized spin.
Achieving superb but controllable isolation of an information carrier is a pre-requisite for any computing paradigm to be feasible. For electrons, this isolation is accomplished with high quality insulating films. Ultra-thin (nanometer scale) insulators can allow quantum tunneling between two contacts, and in this case the insulator can be selected to select preferred state properties of the electrons, like symmetry, energy or spin. For optimizing spin polarization, the electron spin state can "filtered" using ferromagnetic insulators (FMIs) that attenuate one spin state much faster than the other. In this project, we are developing FMIs to form nanometer scale tunnel junctions combined with mesoscopic electronic devices to control and manipulate spin. The most stable and tunable of ultrathin insulators is often oxides, so we are putting more effort into oxide based materials than others.
From a technical perspective, this project features two major research thrusts:
The first of these thrusts is most advanced, where we have developed high quality aluminum oxide by plasma oxidation and confined by cobalt on either side. These tunnel barrier devices exhibit <40% resistance-area (RA) product drift over a monitoring time >100,000,000 seconds and have magnetoresistance (MR) values close to the theoretical limits. We are studying how these properties can be used to improve Josephson junctions in superconducting systems and used to design recipes for use with ferromagnetic insulators.
Our experimental approach to these efforts has the following features. We use a custom, multi-chamber UHV vacuum system to maintain cleanliness through the whole fabrication. We utilize plasmas for in situ cleaning and production of atomic radicals critical to produce high quality oxides (Figure 1). We also use in situ shadow masking, which allows for rapid prototyping by producing measurement-ready wafer (Figure 2). Our UHV pressure deposition chamber features electron-gun deposition and sputter deposition to allow a broad range of multilayers possibilities without removing samples. Finally, we are adding a spatially resolved photo-emission spectroscopy system (XPS/UPS/AES) to measure and monitor surface and interface chemistry at various points in the deposition. The combination of these capabilities is unusual and provides a unique opportunity to explore new materials for spin and quantum physics.