Media Materials Are Key

While areal density improvements are enabled by read-write head improvements, media materials themselves provide the actual keys to realizing smaller bit sizes and true density improvements. Currently, bits are long, but narrow. To fit more bits on a disk, many manufacturers are looking for media that will support squarer bits. The problem engineers confront is that the smaller the grain size in the magnetic media coating, the less capability the grains have to hold a magnetic field or to resist heat.

Thermal stability is a key limiter of areal density improvement, most industry insiders concur. Heads can be made more sensitive, they can be flown closer to media to improve signal detection, and they can be improved in terms of tracking. Ultimately, however, the material properties of media ”how few grains can be used to obtain the best mix of coercivity (magnetic hardness) and resistance to temperature ”exert a limiting influence on density scaling.

Currently, a minimum of about 1,000 grains of magnetic material are required to store a bit of data. To make a bit smaller, one of three things must happen. Either 1) new materials must be found that will hold a detectible magnetic charge and resist SPE using fewer than 1,000 grains; 2) new materials or processes must be created that provide grains that are all uniform in size thereby creating a smaller domain; or 3) new signal detection technologies need to be discovered that will detect a magnetic charge using fewer grains.

In lieu of these improvements, experts say, the current bits per inch (BPI) limit for available media is right around 500,000 to 650,000 BPI. However, BPI is only one determinant of areal density; the other is tracks per inch (TPI).

In its quest for higher areal densities , the industry has looked for ways to write bits into a greater number of narrower tracks. The results of this approach are the drive geometries we see today, which offer TPI capabilities of about 20,000 TPI.

Expanding areal density by increasing TPI of a disk is not, however, a panacea for disk capacity growth. The number of tracks per inch that can be created on disk media is limited by several factors. These include the accuracy of the head position sensing system and resolution of the head recording elements. Significant improvements in the areas of head design, actuator control, and signal decoding will be required to pump up the TPI side of the areal density equation.

With current technology, tracks must be separated from each other by gaps of 90 to 100 nanometers in width to avoid the effects of fringe fields generated by write heads during the recording process. In the words of one expert, "Most write heads look like a horseshoe that extends across the width of a track. They write in a longitudinal (circumferential) direction, but they also generate fringe fields that extend radially. The fringe field causes side writing and side erasure, which is actually desirable and necessary to completely erase and overwrite old information. However, it also imposes a limit on how close tracks can be to one another since you do not want the fringe field generated while writing one track to accidentally erase data in an adjacent track."

More precise "head trimming" processes could lead to further TPI improvements. For example, experts say, a focused ion beam could be used to trim the write head and to narrow the width of the track that a writer writes . However, the read head, which is a sandwich of elements, poses a more difficult problem. The precision of the manufacturing process becomes much more important and much more difficult.

Getting to 100 Gb/in. ^[2] will require disk TPIs of 50,000 to 150,000. With tracks that are only 0.17 microns wide, not only will head design be an issue, so will actuator design. Secondary actuators will be required for heads to follow tracks accurately.

Finally, new technologies, such as turbocodes, will be required to separate the weak signals generated by smaller bits from background noise and to decode them accurately. Current methods for signal processing require a 20-decibel signal-to-noise ratio. With current media and channels, the industry is at least 6 decibels short of being able to work with the signal-to-noise ratio that would apply when dealing with the bit sizes entailed in disks with areal densities of 100 to 150 Gb/in. ^[2]

Design constraints are well understood , and industry insiders suggest that they are confident that materials and read-write technology improvements already in development will support the realization of drives with 100 to 150 Gb/in. ^[2] areal densities. However, this optimism is caveated with the concession that a materials breakthrough will be required to move conventional magnetic disk storage technology much beyond that point.