An emerging application of magnetic nanostructures is in magnetic random access memory (MRAM), a method for storing data bits using magnetic charges instead of electrical charges. MRAM is a revolutionary memory technology that could potentially replace today's semiconductor memory technologies and combines the best attributes of the three major memories including the density of dynamic RAM (DRAM), the speed of static RAM, and the nonvolatility of flash onto a "single" chip.
In addition to these benefits, MRAM has the built-in ability to withstand radiation doses that would destroy conventional memory, which could be useful in space and military applications. Replacing DRAM with MRAM could prevent data loss and start computers instantly without waiting for software to boot up. To successfully integrate magnetic structures in practical applications, advancements in nanocharacterization techniques have to occur.
One of the main challenges in characterization of ferromagnetic nanostructures is the detection of the spin state at nanoscale. Ideally, we would lmeasure the magnetic properties of single nanomagnets; however, this is extremely difficult because the output signal is beyond the detection limit of most conventional magnetometers. Magnetization probe techniques, such as vibrating sample magnetometer or alternating field gradient magnetometer, require a large volume of magnetic materials. The fabrication of such large arrays of identical nanomagnets is slow in the case of electron beam lithography. In addition, statistical smearing of the results can occur due to the size of the ensemble.
While magnetic force microscopy, Lorentz microscopy, and electron holography are microscopy techniques capable of spatially mapping the magnetization distribution within a single magnet, they cannot determine a quantitative hysteresis loop. MOKE is a well-established noninvasive method for probing the magnetization reversal in thin magnetic films. When a beam of plane-polarized light illuminates a magnetized material surface, the reflected light will generally be elliptically polarized, even if the incident light is polarized parallel or perpendicular to the plane of incidence. The magnetization of the reflecting surface is related to the degree of ellipticity of the resulting polarization. The only challenge is in obtaining a magneto-optical signal from small structures.
Currently, we are limited by how much we can successfully reduce a laser spot size to precisely obtain the reversal process of some nanoelements. The second challenge is the rate at which the data points are captured. MOKE magnetometry is used to measure the dynamic properties of nanostructures. Therefore, a high data acquisition rate is required to capture the change in a magnetic signal with a varying external magnetic field.
We successfully integrated a MOKE setup with National Instruments data acquisition (DAQ) tools to investigate the dynamic properties of patterned ferromagnetic nanostructures. Using NI DAQ modules, we recorded a high data capture rate of the optical signal and used the data to study the magnetization reversal processes of ferromagnetic nanostructures, which have potential application in magnetic logic devices.
We fabricated periodic arrays of circular dots 600 nm in diameter in two different configurations over an area of 100 by 100 μm2 on silicon substrates using deep ultraviolet lithography at 248 nm wavelengths followed by the lift-off process. The edge-to-edge spacing (s) between the dots varied from 55 to 600 nm. The thickness of the Ni80Fe20 layer (t) also varied from 25 to 80 nm and was deposited using electron beam evaporation at a rate of 0.2 Å/s. Figure 1 shows the representative scanning electron micrographs (SEM) of the two- lattice geometries for closely spaced dots.
MOKE is one of the magneto-optic effects in which an electromagnetic wave propagates through a medium that has been altered by the presence of a quasistatic magnetic field. The light reflected from a magnetized surface can change in polarization and reflectivity. In the optical aspects, the MOKE method is the interaction between an electromagnetic wave polarized linearly and a magnetized medium, causing the incident wave to reflect with an elliptical polarization. However, the light is known to penetrate nearly 20 nm into the surface for most metals, which means MOKE is not surface sensitive.
While in a microscopic view, the interaction of the spin-polarized electrons, due to the magnetization field M with the incident polarized light, results in the left/right circularly polarized output. MOKE can be categorized by the direction of the magnetization vector with respect to the reflecting surface and the plane of incidence. In our setup, we used the longitudinal MOKE, in which the magnetization vector is parallel to the reflection surface and the plane of incidence. The longitudinal setup involves light reflected at an angle from the reflection surface and not normal to it. After reflection from the magnetized surface, the principal component amplitude is reduced by the isotropic Fresnel reflection coefficient. The light is collected by the photodetector, which can be plotted against the varying external magnetic field to obtain a hysteresis loop.
The schematic block diagram of our MOKE setup is shown in Figure 2. For this experiment, we used a continuous wave solid-state red laser, which has a wavelength of 650 nm. The incident light is passed through a polarizer, followed by a beam expander to uniformly increase the beam diameter. The purpose of the polarizer is to rotate the plane of incidence by +45 degrees. The beam of light is focused on the sample using a low-focal length focussing lens, which can reduce the diameter to 3 μm. The patterned magnetic sample is placed in between two electromagnets, which are powered by a KEPCO power supply. We achieve the variable magnetic field by connecting a function generator to the power supply so we can obtain a sinusoidal waveform of a magnetic field with the desired frequency and amplitude. Also, we place a guass probe in between the electromagnets to acquire the real-time changes in the magnetic field. This probe is directly connected to its controller, which subsequently converts the voltage signal measured from the probe to the magnetic field.
Furthermore, we monitored the exact location of the laser spot using a monochromatic camera that constantly captures the image of the patterned nanostructures. By using translation stages (x, y, z, and θ motion) on which the sample is mounted, we accurately positioned the laser spot. We collected the reflected light from the sample using a second focusing lens that directs the beam onto a photoelastic modulator (PEM), which is an instrument for introducing the modulation in the signal for a better signal-to-noise ratio. This is followed by an analyzer that rotates the plane of the reflected light by -45 degrees. The photodetector connected to a lock-in amplifier demodulates the signal and finally collects the beam of light. The actual experimental setup of MOKE is correspondingly shown in Figure 3.
Results and Discussion
We used a PXI-1411 image acquisition module to acquire the first signal, which is the graphic image of the patterned nanostructure on which the laser spot should focus. The second signal is the real-time magnetic field from the gaussmeter controller, and the third and most important signal is the magnetic signal from the lock-in amplifier. The analysis of the last two signals produces a magnetization loop.
To enhance the signal acquisition rate, we used a PXI-1042Q chassis instead of a normal CPU. Another advantage includes synchronized signal acquisition because all the hardware cards share the same clock frequency as the PXI chassis. We captured the last two signals using an NI PXI-5922 digitizer with a 15 MHz maximum acquisition rate of. By acquiring all three signals together, we characterized the dynamic properties of magnetic nanostructures, which would otherwise not have been possible.
In addition, we used NI LabVIEW software to acquire and analyze real-time data. Figure 4 shows the screen images of the programs used for the entire measurement of the hysteresis loops. Figure 4(a) shows the screen image of the program used for positioning the laser spot directly on the sample. The patterned nanostructures can be easily captured on-screen using this program. Before acquiring the real signal from the guassmeter controller and the lock-in amplifier, we conducted a rough measurement of the loop using an NI PXI-6251 M Series multifunction DAQ module using the program shown in Figure 4(b).
We used the LabVIEW program to check the quality of the signal before capturing the real data. The final two signals are captured using the programs shown in Figure 4(c). We acquired the data in binary format and needed to convert it into ASCII format using the second program Figure 4(c). Depending on the frequency of the external magnetic field, the desired number of samples are captured. The final step is the reconstruction of the hysteresis loop, which is carried out through offline analysis of the captured data using the program shown in Figure 4(d). The result is a magnetization loop, which is a detector signal plotted as a function of varying magnetic fields.
To fully visualize the importance of NI hardware in obtaining the magnetization loops of magnetic nanostructures, Figure 5 shows the result responses from a full measurement cycle as a function of the angle of the applied external magnetic field for two-dot and three-dot lattice geometries. The obtained loops are shown to be significantly sensitive on the geometry of nanostructures and their corresponding orientations.
We demonstrated an experimental setup of MOKE magnetometry to characterize the magnetic properties of patterned nanostructures. NI hardware offered ultrafast data acquisition rates. As an example, we characterized the magnetization reversal process of ferromagnetic dots arranged in different lattice geometries as a function of the orientation of applied field direction. Therefore, we can use NI hardware and software tools to characterize nanoscale devices for their applications in nanotechnology.
Department of Electrical and Computer Engineering, National University of Singapore