Faculty: Dr. James Rantschler
The main thrust of this research is to investigate the damping of magnetic materials in exchange coupled magnetic multilayers. Damping affects the high frequency response of magnetic materials, and these effects are critical parameters for the magnetic recording and magnetic solid state device industries. These industries construct gigahertz-speed devices that pass current through stacks of metallic layers, both magnetic and nonmagnetic. Some of the magnetic layers are coupled to each other through the nonmagnetic layers, and this coupling changes how these layers react to the nanosecond-scale changes in the magnetic field. This investigation is designed investigate how coupling through different materials changes the damping in magnetic layers.
The current understanding of these issues is lacking in three areas: the breadth of materials that have been investigated, the properties of these structures at high-frequency, and an understanding of the mechanisms of the spin-flip transition. In this application, we perform a systematic investigation of these issues through the deposition of trilayers of permalloy (Ni80Fe20) separated by various normal metals, measurement of their morphology using atomic force microscopy (AFM) and their domain properties with magnetic force microscopy (MFM), and the quantification of their damping properties through ferromagnetic resonance (FMR) spectroscopy.
This course of investigation is designed to examine this coupling by comparing series of structures of the two types depicted in the figure below: one designed to measure just the damping induced by the nonmagnetic metal (NM/FM/NM) and another, the coupling between the two layers (FM/NM/FM). In this program we will look at transition metals using one column of the periodic table at a time. We will begin with column 6B, which includes chromium Cr, molybdenum Mo, and tungsten W. This is for two reasons: firstly, this row has three non-magnetic, non-radioactive elements, and secondly, the damping induced through doping in these materials increases by a factor of five. In the all samples, the base ferromagnetic layer will be NiFe, the usual standard soft magnetic material whose high permeability and magnetic moment produce large signals and whose lack of magnetostriction will prevent spurious results due to stress in crystal growth. In investigations where coupling is not important, strong spin sinks (materials with small spin diffusion lengths) such as Palladium Pd and Platinum Pt will be used as a capping layer—part of this research will be in finding less expensive spin sinks that can be used in industry. In the coupled samples, the second layer will be NiFe doped with a small amount (10 atomic percent) of copper; all ferromagnetic metals have small spin diffusion lengths. this will shift the resonance field of the permalloy by more than three times the linewidth of the sample in plane (and 30x out of plane) with only a marginal change in its damping. The coupling will be measured by the angular size of the region where the two layers oscillate in unison—the most dramatic physical manifestation of dynamic coupling (the others being perturbations of resonance lines).
To properly carry out this research, UHV sputtering is important. The current sputter coater used at Xavier has a best pressure of 10-6 Torr, which results in additional scattering at interfaces and even within the normal metal samples. In the current system sputters copper with a resistivity is 4.0 μΩ·cm rather than sub-2 μΩ·cm possible below 1⤬10-8 Torr sputtered in similar systems that we
The expected results are that a greater damping will be seen in the two-magnetic layer systems than would be predicted by just being connected to nonmagnetic metals and related to the coupling strength of the system, a novel result.
This work will elucidate a poorly understood phenomenon in the damping of thin films in multilayer stacks. This understanding will increase our knowledge of magnetic effects at interfaces.
Further significance comes from the characterization of these interfaces. Engineers in the magnetic recording and spintronics industries need this information to tailor their GMR and TMR stacks to product specifications. This information exists only partially, and what exists has been found by various different methods, not all of them suited to the extraction of the damping parameter from linewidth or ringdown data. The data produced by this investigation will be perfectly suited to meet this need, and will position the investigator to produce additional data of the same type.
By measuring the spin diffusion length and spin mixing conductance of a larger range of materials, alternative spin sinks will be found that allow for more leeway in the design of spintronic devices and will increase American competitiveness in advanced technology.