Magneto-optic recording

This is a continuation from the previous tutorial - magneto-optic modulators and sensors.

In magneto-optic recording, digitized information stored in a magnetic thin film is read using the magneto-optic Faraday or Kerr effect. There are certain similarities between the principle of magneto-optic recording and that of the magneto-optic spatial light modulator. Indeed, because of its nonvolatility, a magneto-optic spatial light modulator also has the ability to hold digitized information for later access. Reading of the recorded information is performed using the magneto-optic Faraday effect. However, while the application of a magneto-optic spatial light modulator is primarily dynamic information processing, the purpose of magneto-optic recording is data storage and retrieval. Therefore, there are many fundamental differences between them due to different practical considerations.

The media for magneto-optic recording are ferromagnetic or ferrimagnetic thin films supported by nonmagnetic substrates. The presence of a sufficiently large uniaxial magnetic anisotropy with a positive anisotropy constant in the direction normal to the film surface is required to ensure that the film has two clearly distinguishable, oppositely directed magnetization states, which represent the binary logical states. In amorphous magnetic films prepared by evaporation or sputtering, this condition can be achieved by properly choosing the deposition parameters to create an anisotropic atomic arrangement along the film normal.

The materials suitable for the application of magneto-optic recording include magnetic oxides, particularly the Bi-substituted garnets, metallic Pt-Co and Pd-Co multilayers, and magnetic alloys.

The most popular magneto-optic recording materials today are amorphous ferrimagnetic rare-earth transition-metal (RE-TM) alloys containing one or more of the rare earths Gd, Tb, and Dy in addition to one or more of the transition metals Fe and Co. The most prominent examples are GdTbFe and TbFeCo alloys.

The magnetizations of these rare-earth and transition-metal atoms vary differently with temperature and are antiferromagnetically coupled. The rare-earth magnetization is larger at low temperatures, whereas the transition-metal magnetization is larger at high temperatures.

Consequently, an alloy of a proper RE-TM composition is a ferrimagnetic material that has a compensation temperature, \(T_\text{comp}\), below its Curie temperature \(T_\text{c}\). At \(T_\text{comp}\), the the magnetization of the rare earth and the transition metal are equal and opposite, resulting in zero net magnetization.

The coercive field, \(H_\text{c}\), exhibits a singularity tending toward infinity at \(T_\text{comp}\). Above the compensation point, \(H_\text{c}\) decreases as the temperature is increased toward \(T_\text{c}\). The magnetization and coercivity of such an alloy as a function of temperature are shown in figure 7-18(a) and (b), respectively.

The compensation and Curie temperatures, as well as the temperature characteristics of \(H_\text{c}\), of an alloy can be controlled by properly choosing the composition of the alloy.

The information is stored in the recording medium by means of magnetic domains. Writing, erasing, and rewriting are achieved by switching the magnetization direction through a thermomagnetic process with optical heating by a focused laser beam.

The composition of the medium is chosen to have a compensation point close to room temperature and a Curie temperature between 400 and 600 K.

On the one hand, this medium provides a high coercivity at room temperature to stabilize the information stored in the magnetic domains and to allow for a high storage density. On the other hand, the coercivity of the heated spot can be significantly lowered in the write process with a laser beam of a moderate power to raise the temperature of the heated spot near or above the Curie temperature.

The thermomagnetic switching process is based on a simple principle that the magnetization in a locally heated volume of the film can be oriented to the direction of an applied magnetic field when the coercivity is lowered at a high temperature to be less than the applied magnetic field.

Writing is accomplished by focusing a laser beam of a moderately high power, typically in the range of 5-10 mW, to a diffraction-limited spot on the medium. The write process is performed either by modulating the laser power at a constant magnetic field or by modulating the magnetic field at a constant laser power during the pass of the laser beam over the medium along the data track.

The latter permits direct overwrite, but its switching frequency is limited to less than 10 MHz due to the operation margins of the coil generating the magnetic field.

Erasure of the written information is accomplished by heating the medium with a constant laser power at a constant magnetic field to revert the magnetization of the bit to be erased back to the preset direction.

The simplest overwrite scheme uses two passes of the laser beam over the medium: the first pass for erasing the old information and the second for writing the new data. More sophisticated schemes using one or two laser beams in a single pass for direct overwrite are also developed, Magneto-optic recording is both erasable and rewritable because the thermomagnetic switching process is reversible.

The optimum magneto-optic disk is based on a multilayer structure to achieve long-term stability, high read-out efficiency, and high switching sensitivity. As shown in figure 7-19(a), a typical disk is composed of a pregrooved, transparent polyvinyl chloride or glass substrate of about 1 mm thickness, a precoated antireflection dielectric layer of 80 nm thickness, a magneto-optic medium layer of 45 nm thickness, a thin space layer, a reflective metallic layer of 30 nm thickness, and, finally, a protective polymer layer of a few micrometers thickness.

Figure 7-19.  (a) Multilayer structure and (b) tracking pregrooves of a magneto-optic disk.

The laser light used for writing, erasing, or reading is incident on the disk from the substrate side, as is also illustrated in figure 7-19(a). The composition and thickness of each layer, as well as the number of different layers, vary in disks designed for different applications.

The pregrooved structure serves for tracking in the recording system. The depth of the pregrooves is in the range of 50-70 nm, and the spacing between neighboring tracks is designed to be twice the focused laser beam spot size. A typical example is shown in figure 7-19(b).

By monitoring light reflected from the pregrooved structure, tracking servomechanisms maintain the position of the recording head accurately along the track and keep the focal point of the laser beam on the surface of the magneto-optic medium layer.

The written domain size is governed more by the magnetic properties of the medium and the domain nucleation and growth processes than by the diameter of the focused laser beam. Domain sizes smaller than the Rayleigh resolution limit, \(d_\text{R}=1.22\lambda/\text{NA}\), are quite easily written in a good magneto-optic medium. The practical spot size is approximately given by

\[\tag{7-54}d_0=\frac{0.5\lambda}{\text{NA}}\]

where \(\lambda\) is the laser wavelength and \(\text{NA}\) is the numerical aperture of the objective lens.

Taking into account the spacing between the pregrooved tracks in a disk, this yields an areal bit density of

\[\tag{7-55}\text{ABD}=\frac{1}{2d_0^2}=\frac{2(\text{NA})^2}{\lambda^2}\]

Typically an objective lens with a numerical aperture of 0.5 or greater is used so that the spot has submicrometer dimensions.

Example 7-6

A magneto-optic recording system uses a diode laser emitting at \(\lambda\) = 800 nm and an objective lens of \(\text{NA}\) = 0.5. For this system, the laser spot has a submicrometer spot size of 800 nm, according to (7-54). Using (7-55), it is found that the areal bit density is approximately \(78\text{ MBit cm}^{-2}\) (or \(503\text{ Mbit in}^{-2})\). For this bit density, the mark area,  which is the area for each bit, is \(1.28\mu\text{m}^{-2}\).

To achieve a high recording density, one avenue is to reduce the wavelength of the laser light utilized. According to (7-55), a 40% increase in areal bit density can be realized by replacing the diode lasers at 800 nm with red diode lasers of wavelengths at around 670 nm. The areal bit density can be quadrupled by using an InGaN laser at 400 nm wavelength. Further increase of the bit density can be accomplished by using lenses of high numerical aperture. The bit density is also dependent upon the data coding scheme. Using efficient codes for the data, such as those employing the magnetic flux changes between domains rather than the domains themselves to represent the information, the bit density can be effectively doubled. An areal bit density as high as \(100\text{ Gbit in}^{-2}\), equivalent to \(15.5\text{ Gbit cm}^{-2}\), has been achieved.

For effective optical heating, high optical absorption in the magnetic film is required. Consequently, the polar Kerr effect, rather than the Faraday effect, is most commonly used for the read process.

From (7-38) [refer to the magneto-optic Kerr effect tutorial], it can be seen that the polar Kerr effect angle \(\theta_\text{K}\) obeys the relation \(\theta_\text{K}(\pmb{M_0})=-\theta_\text{K}(-\pmb{M}_0)\). Therefore, the two possible directions of magnetization correspond to a positive and a negative Kerr signal, respectively.

The magneto-optic recording system consists of a polarization-sensitive optical head, as is shown in figure 7-20(a).

During readout, the laser power is typically reduced to about one-tenth of that used for writing, which is far below the threshold to write or erase. Upon reflection from the magnetic medium, the plane of polarization of the light is rotated by the polar Kerr effect.

From (7-38) [refer to the magneto-optic Kerr effect tutorial], it can be seen that in the polar Kerr effect a Kerr rotation angle is always accompanied by a Kerr ellipticity. Therefore, the reflected light is passed through a wave plate to compensate for the ellipticity introduced by the polar Kerr effect.

It is then decomposed into two components by a polarization beam splitter that is set at \(45^\circ\) with respect to the polarization of the incident light, as shown in figure 7-20(b).

The two components are directed to a set of differential photodetectors. With no magneto-optic rotation of the plane of polarization, the intensity of light at each photodetector is the same, yielding no difference signal between the two photodetectors. With a rotation of the plane of polarization, one photodetector receives more light than the other, resulting in a difference signal.

As shown in figure 7-20(b), with a polarization rotation angle of \(\theta_\text{K}\), the optical fields of the reflected light passing through the polarizing beam splitter to reach the two photodetectors are \(E_1=E_r\cos(\pi/4+\theta_\text{K})\) and \(E_2=E_r\sin(\pi/4+\theta_\text{K})\), respectively. Consequently, the signal current produced by the differential photodetectors is given by

\[\tag{7-56}i_\text{s}=\mathcal{R}P_0R\left[\sin^2\left(\frac{\pi}{4}+\theta_\text{K}\right)-\cos^2\left(\frac{\pi}{4}+\theta_\text{K}\right)\right]=\mathcal{R}P_0R\sin2\theta_\text{K}\]

where \(\mathcal{R}\) is the responsivity of the photodetectors, \(P_0\) is the average laser power incident upon the medium surface, and \(R\) is the reflectivity of the medium surface. Clearly, the polarity of the difference signal indicates the direction of the magnetization because \(i_\text{s}\) changes sign with \(\theta_\text{K}\).

From (7-56), the difference-signal current can be increased by increasing the detector responsivity, the laser power, the medium reflectivity, or the Kerr rotation angle. However, the important parameter characterizing the medium and the recording system is the signal-to-noise ratio (SNR) rather than the difference signal alone.

The SNR obtainable from the magneto-optic recording system is fundamentally limited by shot noise in the differential photodetectors. The shot-noise, which is determined by the total detector current rather than by the difference signal current, is given by

\[\tag{7-57}\bar{i_\text{n}^2}=2eB\mathcal{R}P_0R\]

where \(e\) is the electronic charge and \(B\) is the bandwidth of the detectors. Because the maximum values of \(|\theta_\text{K}|\) for magneto-optic recording media are on the order of 10 mrad, the shot-noise-limited SNR is given by

\[\tag{7-58}\text{SNR}=10\log\frac{\bar{i_\text{s}^2}}{\bar{i_\text{n}^2}}\approx10\log\frac{2\mathcal{R}P_0R\theta_\text{K}^2}{eB}\]

where \(\theta_\text{K}\) is in radians.

The SNR can be increased by increasing the laser power or the value of \(R\theta_\text{K}^2\). The value of \(R\theta_\text{K}^2\) is purely determined by the recording medium and is regarded as the figure of merit of a magneto-optic recording medium.

Because the laser power in the read process is limited by the threshold for writing and erasing, this figure of merit is a very important consideration in the choice of a medium.

For RE-TM alloys, the maximum value of \(|\theta_\text{K}|\) are below \(0.5^\circ\), and \(R\approx0.4\) at a laser wavelength around 800 nm. Other media have slightly higher values of \(\theta_\text{K}\), but have other disadvantages. For example, MnBi has a Kerr rotation angle of \(\theta_\text{K}=0.7^\circ\), but it is polycrystalline, resulting in a noise figure high above the shot-noise limit and a poor SNR in spite of the large Kerr rotation.

Example 7-7

A representative set of parameters for a magneto-optic recording system is \(\mathcal{R}=0.4\text{ A W}^{-1}\), \(P_0=1\text{ mW}\), \(\theta_\text{K}=0.4^\circ=7\text{ mrad}\), and \(B=10\text{ MHz}\), yielding a shot-noise-limited SNR of 40 dB.

In a magneto-optic recording system, there is no direct contact between the recording head and the medium. The spacing between the optical head and the disk can be of the order of millimeters. Therefore, head crashes are not a concern as they are in magnetic recording systems.

Because the optical access is provided through a transparent substrate and the laser beam is highly focused on the surface of the magneto-optic film, small particles on the back side of the substrate are out of the focal plane of the laser beam and do not significantly affect recording or readback.

These characteristics of the magneto-optic recording technology make it possible to have a removable disk.

The next part continues with the guided-wave magneto-optic devices tutorial.