Get documents about "Status and Future of Magnetic Data Storage Technology Mark
Document Sample
Status and Future of Magnetic Data Storage Technology possible for each bit to be a single grain. Thus, instead of
Mark H. Kryder requiring tens of grains per bit to obtain adequate SNR as
Seagate Research do present granular media, it would be possible to use
1251 Waterfront Place, Pittsburgh, PA 15222 only one grain per bit, greatly extending the ultimate areal
density of recording. The problem with patterned media
The areal density of magnetic recording on disc drives has recording is that, if perpendicular recording is in fact able
been advancing at a pace in excess of 100% per year in to achieve approximately 1 Tbit/in2, then the islands and
recent years. This has been supported by a continual spaces between them in patterned media will have to have
decrease in media grain size and grain size dispersion and features smaller than 12.5 nm. This is far beyond the
by the use of anti-ferromagnetically coupled (AFC) resolution of present lithographic systems and far beyond
media. Although it is expected that further gains will be what is projected by the Semiconductor Industry
achieved, it is clear that achieving these gains is becoming Association Lithography Roadmap for the next decade.
more difficult, because we are rapidly approaching grain For this reason, it is considered more likely that, if
sizes at which the media are approaching thermal patterned media is utilized, it will be in some sort of self-
instability, even with AFC media. ordered structure. Self-ordered magnetic arrays of FePt
have been demonstrated, for instance, and L10 FePt media
It is our expectation that at densities somewhere beyond have such large anisotropy that they would be thermally
100 Gbit/in2, perpendicular recording will be utilized and stable even with a particle size less than 3 nm. Such a
enable the areal density growth to be further extended. medium could, in principle, support thermally stable
By using a single-pole head and a soft-underlayer under recorded information at a density of the order of 50
the recording layer, perpendicular recording effectively Tbit/in2. Of course, such L10 FePt media would be
places the recording layer in the gap of the head, thus extremely high coercivity and would still require some
making it possible to produce larger head fields than novel means for writing – perhaps HAMR.
achievable with a ring head and to use thicker media. This
theoretically makes it possible to use media with higher It is concluded that, although longitudinal recording may
anisotropy and smaller grain diameters than in be approaching densities at which a change in technology
longitudinal recording, thus making it possible to will be required, it appears that perpendicular recording
maintain writability on media with smaller grain sizes should be able to extend the areal density growth to
while maintaining media SNR. These advantages have perhaps as much as 1 Tbit/in2. For densities beyond what
prompted researchers to predict that perpendicular perpendicular recording can achieve, both HAMR and
recording could achieve 0.5-1.0 Tbit/in2 areal density .
1,2,3
patterned media provide possible solutions. In the limit,
by combining HAMR and patterned media, it is
Beyond the limits of perpendicular recording other conceivable that 50 Tbit/in2 could be achieved. However,
approaches are expected to be required. Two approaches although perpendicular recording appears nearly ready to
that appear promising are heat-assisted magnetic be moved into products, only limited recording
recording (HAMR) and patterned media recording. With demonstrations have been made of HAMR and patterned
HAMR, very high coercivity media are written using both media recording, and considerable work remains to be
a thermal source and a magnetic field source done on these technologies, before they can be made into
simultaneously. Because of the application of heat, the products.
high coercivity of the media is reduced during the write
process to levels at which it may be written by
conventional magnetic head materials. Although simple
in concept, modeling has shown that to gain significantly
from HAMR, it will be necessary that the heat spot be
localized to a single track and have thermal gradients
similar in extent to the bit length. Moreover, to achieve
the data rates required, the thermal rise and fall times of
the media must be in the sub-nanosecond range. Given
that at 1 Tbit/in2, a square bit cell would be 25 nm on a
side, some form of near-field technique must be used to
produce the thermal spot. Designing a head with a near-
field thermal source and magnetic field gradient that are
co-located and of these dimensions is challenging;
however structures utilizing surface plasmon resonances
show promise of producing such small spot sizes.
Assuming a head design is evolved, it will still be
necessary to develop a reliable head-disc interface (HDI)
that will enable the head-medium spacings required and
that will withstand the temperatures of the HAMR
process. Testing has shown that at high temperatures
common in the HAMR process, conventional media
1
lubricants are vaporized and can lead to failure of the R. Wood, “The Feasibility of Magnetic Recording at 1 Terabit per
HDI. Square Inch, “ IEEE Trans. Mag., Vol. 36, No. 1, pp.36-42, January
2000.
2
H.N. Bertram and M. Williams, “SNR and Density Limit Estimates:
Another approach to achieving higher densities is to
A Comparison of Longitudinal and Perpendicular Recording,” IEEE
utilize patterned media. In patterned media recording, Trans. Mag., Vol. 36, No. 1, pp. 4-9, January 2000.
each bit is a defined island of magnetic material in a non- 3
M. Mallary, A. Torabi and M. Benakli, “1 Tb/in2 Perpendicular
magnetic matrix. With such a defined bit cell, it is Recording Conceptual Design,” to be published IEEE Trans. Mag.
Related docs
