Home      Faculty      Research      Fast Dynamics Laboratory      Contact us

Project 1: Spin Clusters and Nanomagnetic Information Processing

1.1. Spin Clusters and Interactions

We have recently built an improved cluster-deposition system based on the early instrument of Haberland et al. [Surf. Rev. Lett. 3, 887, 1996 ]. A schematic view of the main components is shown below:

Schematic of universal cluster deposition system.

The system includes a sputtering gas-aggregation source in which clusters are formed in a 1-10 Torr inert gas, with diameters ranging from about 3 to 12 nm. The clusters impinge on a substrate alone or in combination with atoms from one or two other sources. The other source or sources are used to form a matrix for the magnetic clusters and/or underlayers or coating layers.

Schematic views of possible nanostructures that can be formed. (a) uncoated and (b) coated clusters embedded in a matrix.

A major advantage of this type of system is that the clusters have a much smaller size dispersion than grains in a typical vapor deposition system. Studies with a time-of-flight (TOF) mass spectrometer have shown that the observed log-normal diameter distribution has an rms standard deviation of sd/<d> » 0.1, and we have shown that applying the quadrupole mass selector to Co clusters reduces this figure to about 0.03. We have found interesting magnetic-interaction effects in Co clusters in Cu and SiO₂ matrices, as well as Co-moment reduction on the surface of Co clusters embedded in Cu. It is clear that research on the magnetic properties of cluster-assembled materials is in its infancy, establishing a fertile ground for fundamental studies and applications.
The possibility of fabricating nearly monodispersed clusters of a wide range of sizes and densities enables us to investigate many phenomena of fundamental importance to the magnetic behavior of materials. The individual cluster properties also are strongly affected by shape and surface effects, including surface spin disorder, uncompensated spins when clusters have an antiferromagnetic oxide layer, and surface (interface) anisotropies. Since many of these features can be controlled during the cluster-deposition process, it is possible to fabricate cluster-based materials with a wide range of characteristics.
Well-separated magnetic clusters embedded in a nonmagnetic metallic matrix interact via long-range magnetostatic interactions and through Ruderman-Kittel-Kasuya-Yoshida (RKKY) exchange mediated by the conduction electrons. As calculated by Skomski, both exchange and magnetostatic interactions scale as 1/R³, where R is the intercluster separation, but RKKY-type exchange interactions dominate when the elements are smaller than about 1 nm. It has also been investigated how superparamagnetic clusters compare to spin-glass-like systems. At low concentrations but with varying cluster sizes, one can investigate the magnetic properties of the individual clusters. Relatively few systematic studies have been done on these magnetic cluster assemblies since it has been difficult to vary the concentration independently of the cluster size utilizing the methods of previous researchers. The ability of our technique to independently control these parameters makes it ideal for systematic studies of magnetic clusters.
We produce cluster:matrix systems of the form TM:M where TM = Fe or Co, and M = Al₂O₃, Pt, and a-Ag_xSi_1-x with 0.1 < x < 0.2.
A second project involving spin clusters concerns high-anisotropy, cluster-assembled films for extremely high density magnetic recording. It is expected that present CoCrPtX longitudinal recording media and systems will reach their limits at an areal density of about 200 Gb/in². New systems are required, including perpendicular recording and altogether different media. Wood as considered the feasibility of recording at 1 Tb/in² and has determined the following set of media characteristics required for perpendicular recording at this density. They include coercivity H_c = 12 kOe, thickness t = 9 nm, grain diameter d = 8 ± 1 nm, perpendicular anisotropy, and minimal and/or controllable exchange coupling between grains.
We have shown that it is possible to prepare highly oriented L1₀ phase FePt:M nanocomposites with grain diameters in the 6-8 nm range, and H_c values of 5-12 kOe. We have studied M = Ag, C, SiO₂, and B₂O₃, as matrices, with excellent perpendicular anisotropy in all of these cases. We exploit the capabilities of the cluster-deposition system to produce FePd and MnPt particles in SiO₂, C, and other matrices with favorable structural and chemical stability. Composite targets (e.g., Fe and Pd or Mn and Pt) are used to create Fe_xPd_1-x or Mn_xPt_1-x alloy clusters with variable composition near x = 0.5. Heating the substrate to several hundred degrees Celsius and/or a rapid anneal provides the thermal energy to produce the ordered L1₀ phase without increasing the overall cluster size. Considerable magnetic and structural characterization (Sellmyer, Liou) will provide feedback on the fabrication parameters to produce controlled coercivity, grain size, perpendicular anisotropy, etc. Interface mixing and reactions will be studied through magnetization and photoemission measurements (Dowben). Theoretical modeling by Skomski and dynamics studies by Kirby are conducted provide information on the design of these spin-cluster materials as potential 1 Tb/in² recording media.