ABSTRACT
Figure 6.1 AD progress in IBM hard disks. Courtesy of Grochowski, E., Halem, R. D. (2003). IBM Syst. J., 42, 338. In the 1970s and 1980s, AD underwent an annual compound growth at a rate of ~30%. It should be noted that the significant improvement came in 1991, with the introduction of thin-film media as well as the magnetoresistive head. This accelerated AD growth from 30% to 60% per annum. Magnetic recording with an AD up to 10 Gbit/in2 was demonstrated in 1997. In data storage devices there has to be a medium for storing information. In magnetic recording the medium is tape or disk and satisfies two basic principles. The first is magnetized grains with
north and south poles out of which the magnetic stray field stems and can be sensed by a conventional magnetic field sensor. The second prerequisite is the ability to change the polarity of magnetic grains by applying an external magnetic field, which is usually produced using an electromagnet. In longitudinal recording technology were needed many grains to store a bit, which are isolated with zigzag transition. The polarities of small multigrain magnets are parallel to the surface of the hard disk. When two identical poles are next to each other a strong magnetic field emerges from the medium, but no field will emerge when opposed poles are next to each other. Therefore when a giant magnetoresistance (GMR) magnetic field sensor flies over the pole-pole transitions a voltage pulse is produced and synchronized with a clock pulse. When during a clock pulse the GMR sensor produces a voltage peak, it is represented by 1, and the opposite case, for example, the absence of voltage, is represented by 0. The volume V of grains typically decreases with an increase of AD. To compensate for a decrease in V, higher-magnetocrystalline-anisotropy (Ku) materials are needed to maintain sufficient stability [4]. However, at the superparamagnetic limit [5], scaling of the grain size necessary to maintain a sufficient signal-to-noise ratio (SNR) can no longer be compensated by increasing Ku, due to the limited write fields achievable with today’s write heads. The predicted superparamagnetic limit for conventional longitudinal recording is an AD of 150 Gbit/in2 [6]. 6.2 Perpendicular Recording Media for 1 Tb/in2
and beyondIn perpendicular recording, which is used nowadays for recording information in HDDs, magnetization stands out of plane. In this case the origin of the stray magnetic field is the center of bit cells rather than transitions. In the mid-1970s, perpendicular magnetic recording technology was proposed as a way to overcome the problem of demagnetized fields from recording transitions [3]. In a ferromagnetic or ferrimagnetic system there exists a demagnetizing field with a direction opposed to that of magnetization. The demagnetized field
is Hd = –N × M, where N is the demagnetizing tensor and M is the magnetization vector. N depends on the shape and direction relative to the magnetic field of the magnet. In longitudinal recording as the linear density increases, the distance between the magnetic charges decreases. When the distance between the charges decreases an increased demagnetized field in the opposite magnetization direction is expected, at higher linear densities (Fig. 6.2). In 1975 Iwasaki and Takemura experimentally observed that magnetic vortices will be produced in longitudinal media because of a stronger demagnetizing field when thicker films are used [8]. This circular magnetization would not be able to produce high output voltages at higher densities.
