About a decade ago, the areal storage density on magnetic disk drives was advancing at a pace of about 25% per year. At that time semiconductor dynamic random access memory was advancing at a pace of about 40% per year, and many semiconductor proponents argued that the time would come when semiconductor devices offered a higher storage density and lower cost than magnetic disk storage. These proponents argued that with their faster access time, semiconductor devices would replace disk drives. In 1991, however, the magnetic disk drive industry responded by advancing the rate of progress in areal density to 60% per year, as illustrated in Fig. 2.3. With the recent announcement of recording at 10 Gb/in2 density by IBM, it is now believed that this rate of progress can be sustained at least into the next millenium. Hence, rather than converging toward the areal density of magnetic disk drives, the areal density of semiconductor memory devices is falling further behind that of magnetic disk drives, and it no longer appears likely that semiconductor memory devices will be a significant threat to magnetic storage.
Fig. 2.3. Areal density of magnetic hard disk drives and of dynamic random access memories as a function of the year of shipment (IBM).
In spite of the current rapid pace of advancement in disk drive technology, there are some obstacles to progress on the future horizon. Since 1957 when IBM introduced the first disk drive, the RAMAC, the areal density of magnetic disk recording has been increased over 2 million times by linear scaling of dimensions of the head, medium and head-medium spacing, while the sensitivity of read heads has been increased. While doing this, it has been necessary to reduce the size of the magnetic particles of which the medium is made in order to maintain the signal to noise ratio of the system. This is because the signal to noise ratio scales approximately with the number of magnetic particles contained within a bit. If this trend is continued, it is inevitable that the particle size will become so small that the magnetic energy of a particle, KUV, (where KU is the magnetic anisotropy energy holding the magnetization in its orientation and V is the volume of the particle) will decrease to values that approach thermal energy given by KBT (where KB is Boltzmann's constant and T is the absolute temperature). When this occurs, thermal energy alone may cause magnetic recordings to become unstable.
Although this thermal instability problem appears at first glance to be a problem in the magnetic recording medium, in reality part of the limitation comes from the magnetic recording head. Magnetic materials that have higher values of magnetic anisotropy energy KU exist; however, they require higher fields for recording than can be produced with existing recording heads. Before the field produced by the head is made large enough to write on the medium, the magnetic material from which the head is made saturates.
Modeling has indicated that if magnetic recording continues to be scaled linearly, this thermal instability limit will be reached at about 36 Gb/in2, or a factor of nine higher than the highest density disk drive made today. At the 60%/year growth rate, this density would be reached in about five years. However, modeling performed by a group of researchers working with the National Storage Industry Consortium (NSIC) has recently shown that, if the bit cell is not scaled linearly but decreased in track width more than in bit length, a density of 100 Gb/in2 may be achievable. Ultimately, however, thermal instabilities will become a problem, and means to circumvent this limitation must be found.
One possible means of extending magnetic recording density beyond what can be achieved using longitudinal magnetic recording is to use perpendicular magnetic recording. As illustrated in Fig. 2.4, in perpendicular magnetic recording, the medium is magnetized perpendicularly to the film plane, rather than in the plane. If a high permeability magnetic underlayer is placed under the perpendicularly magnetized thin film medium, then an image of the magnetic head pole is produced in the underlayer. Consequently, the perpendicular medium is effectively in the gap of the recording head, where the field is larger than the fringing field produced at a longitudinal medium by a longitudinal magnetic recording head. With this larger record field it is possible to record on media with higher KU and, consequently, smaller grain size V and smaller bit sizes. Advocates of perpendicular recording argue that, because of these higher record fields and the fact the media may be made thicker, perpendicular recording can achieve considerably higher densities without suffering from thermal instabilities than longitudinal recording can. Japanese companies and universities have long been proponents of perpendicular magnetic recording and have considerably more experience with it than we have in the United States.
Another possible means to circumvent the thermal stability problem of conventional longitudinal magnetic recording is to use some form of thermally assisted recording process. If the temperature of high KU media is raised, then a smaller field is adequate for recording. If the media are kept near room temperature for storage and readout, but raised in temperature during recording it may be possible to use media with intrinsically higher KU (and, consequently, improved thermal stability) while still being able to record on them. Dr. Miura of Fujitsu Laboratories, the Storage Research Consortium (SRC) and ASET suggested such an approach may have merit during this WTEC team's Japan trip.
Figure 2.4. A schematic diagram of a perpendicular magnetic recording medium with a soft underlayer and a single pole head.