Nanoscale Domain Wall Devices

Nanoscale Domain Wall Devices

The study of magnetic nanostructures has been a very active research field. One of the reasons for this is that applications in ultrahigh-density magnetic recording require the element dimensions to be reduced down to the nanometer scale. Also, interesting changes in the magnetic behavior of nanostructures are expected as the overall size approaches some critical length. On the other hand, half-metallic ferromagnets have been extensively investigated for realizing spin-dependent devices with high magnetoresistance. These materials have a band gap in the minority spin density of states near the Fermi level and therefore exhibit a 100% spin polarization. Among these materials, chromium dioxide (CrO2) is one of the few experimentally proven half metals and possesses the largest spin polarization so far reported. Hence, the study of the properties of nanosized half-metallic ferromagnets is a very promising field.

The fabrication of CrO2 structures at the nanometer scale is a challenging task. So far, there has been no single reliable recipe available. This is due to the metastable nature of CrO2, which can easily decompose into Cr2O3. Prior to this work, various methods have been attempted to pattern small CrO2 structures, including wet etching, reactive ion etching (RIE) and focused ion beam milling. In all these methods, CrO2 thin films are deposited and then etched to make small patterns. These processing methods inevitably cause degradation of the quality of CrO2 after the etching step. Another approach for creating patterned chromium dioxide elements is the selective-area growth technique. In this method, patterned SiO2 is first deposited onto a TiO2 substrate and then CrO2 is grown selectively on the areas not covered by SiO2 because CrO2 has zero sticking coefficient on SiO2 substrates. This avoids the need to pattern the CrO2 film after deposition and, hence, subsequent degradation of the film.

The experimental process is shown schematically in Fig.1. To summarize, (1) the TiO2 substrate was first covered by a layer of amorphous SiO2 ~100 nm using rf sputtering, and then spin coated with e-beam resist. (2) After e-beam writing and subsequent development, the patterned resist was used as an etching mask for reactive ion etching of the underlying SiO2 layer in a CHF3 atmosphere. (3) Finally, the sample was carefully cleaned in acetone and de-ionized water before loading into oxidation furnace to deposit CrO2. Both polycrystalline and epitaxial CrO2 nanowires with sub-100-nm width can be obtained with this method.

The SEM images of CrO2 nanodot arrays in hexagon and square lattices are shown in Figs. 2. As can be seen, all the dots are distinct and well separated. The epitaxial (100) TiO2 substrate ensures that the grown CrO2 nanodots also have (100) texture and can be regarded as nanosized extensions of the substrate. Furthermore, it was observed that, first, the CrO2 covers the bottom of the e-beam defined circular holes and then beyond a certain thickness, the nanodots start to become rectangular shaped (viewed from the top) with long side parallel to the c axis ([001] direction). This indicates that the lateral growth rate in CrO2 nanodots is anisotropic and a maximum at an intermediate angle between the [010] and [001] directions. More specifically, in our case, this is the diagonal direction of the rectangle. It is believed that the anisotropy in surface binding
energy leads to the rectangular shape but the detailed kinetic
process is not very clear.

Chromium dioxide nanostructures with nanowire and ring shapes have also been fabricated on the same substrate to ensure identical magnetic and structural properties. The sample was first saturated along the easy axis [001] direction by applying a magnetic field of 2 kOe and then imaged in the remanent state using magnetic force microscopy (MFM). The MFM images were obtained with a Digital Instruments nanoscope IIIa scanning probe microscope in tapping mode with the scanning tip magnetized along the z direction (perpendicular to the sample surface).

Figures 3 are MFM images of CrO2 nanowires of various widths aligned along the [001] and [010] directions. In Fig. 3a, all the domains are aligned in plane along the magnetic easy axis. It is clear that the magnetic stray field signals at the element ends are very strong even for the 100 nm width, with no out-of-plane field observed in the rest of the wire. This indicates a high quality single domain structure with a fully in-plane magnetization. For the wires in Fig. 3b, lamellar domain structures with fragmented walls, aligned mainly along the easy axis direction, are observed. Each wire is either magnetized parallel or antiparallel to the [001] direction, across its width, leading to an alternating contrast in the MFM image. A similar result has been reported recently for epitaxial CrO2 wire with 2 micron width. This domain configuration implies that a strong magnetocrystalline anisotropy exists in the [001] direction for these epitaxial chromium dioxide elements. The formation of multiple domains can be attributed to the competition between magnetostatic energy and domain wall energy.

The magnetic image of an epitaxial CrO2 ring in the remanent state is displayed in Fig. 3c. Unlike the vortex and onion states found in other ferromagnetic rings, the epitaxial CrO2 ring displays a much more complicated domain structure. As can be seen, two large parallel domains are formed in the upper and lower parts of the ring and separated by many relatively small antiparallel domains in between. This configuration is consistent, again, with the existence of a strong uniaxial anisotropy along the easy axis direction, as
found in Fig. 3b. The strong magnetocrystalline anisotropy is more obvious in the MFM image of the Y-shaped structure shown in Fig. 3d. The shape anisotropy induced in this structure is almost negligible in comparison to the uniaxial anisotropy, given that every domain is aligned along the magnetic easy axis. Epitaxial CrO2 nanodots with a diameter of 250 nm also show a single domain configuration along the 001 direction, as displayed in Fig. 3e. These dots exhibit a strong signal-to-noise ratio and a strong magnetocrystalline anisotropy which dominates the shape anisotropy contribution, which suggest that epitaxial CrO2 nanodots are a potential candidate for high-density state storage media.


Please send your comments and thoughts to
Professor Gang Xiao