Buschow K.H.J. (Ed.) Concise Encyclopedia of Magnetic and Superconducting Materials

Подождите немного. Документ загружается.

coating is lost during the simulated drying but the

media nevertheless exhibits a tendency for particle

alignment in both the plane of the web and along

the web direction. The correlation function for this

medium is shown in the Fig 9.

In contrast, the media made by drying the disper-

sion in a strong magnetic ﬁeld exhibits very strong

alignment in the ﬁeld direction, shown in Fig. 10. The

predicted alignment is expected to overestimate par-

ticle orientation as the model is not able to include the

disruptive effects arising from the high rate of solvent

loss. This strong alignment is necessary for the pro-

duction of good-quality magnetic storage media.

8. Magnetic Properties of Magnetic Media

Two hysteresis curves, one for an interacting and the

other for a noninteracting system, have been calcu-

lated from an initial saturated state for the zero-ﬁeld

media (Fig. 7(a)) and the ﬁeld-dried media (Fig. 7(b)).

The results for the interacting and noninteracting

Figure 8

Correlation function of the zero-ﬁeld media made from

nonsheared coating.

Figure 7

(a) View of a zero-ﬁeld media made from sheared

coating. (b) View of a ﬁeld-dried media made from

sheared coating. (c) Cross-sections of zero-ﬁeld media,

ﬁeld-dried media, and ﬁeld-dried media showing

channels.

Figure 9

Correlation function of the zero-ﬁeld media made from

sheared coating.

650

Magnetic Reco rding: Particulate Media, Micromagnetic Simulations

cases have been calculated using the same simulated

microstructure, therefore a direct comparison of

effects may be made.

The hysteresis curve for the zero-ﬁeld media

(Fig. 11) shows a relatively high squareness seen in

the noninteracting case and is indicative of the partial

alignment during the coating process. As expected,

the interparticle interactions give rise to a reduced

remanence and coercivity. Although the coercivity of

the medium is reasonable in comparison with exper-

iment, the remanence is rather low. The hysteresis

curve for the ﬁeld-dried media is shown in Fig. 12.

The coercivity is rather large, however the remanent

squareness is larger than the experimental result of

0.85–0.9. It would appear therefore that these two

cases represent extreme bounds with the properties of

real media lying somewhere between.

The d.c. demagnetization remanence curves have

also been calculated, from which the switching

ﬁeld distribution has be derived. For the well-aligned

ﬁeld-dried media, the magnetization was found to be

essentially perfectly irreversible with no significant

difference between the hysteresis loop and d.c.

demagnetization curve. The d.c. demagnetization

curves for the poorly aligned zero-ﬁeld media are

shown in Fig. 13.

The effect of interactions is to lower the remanence

and coercivity, which is to be expected and is nor-

mally observed experimentally. A second effect is also

clear, this being a dramatic increase of the initial

shape of the remanence curve in small ﬁelds. This is

accompanied by an increase in the width of the

switching ﬁeld distribution, (SFD), shown in Figs. 14

and 15, which presumably reﬂects a dispersion of the

local interaction ﬁeld, which seems reasonable.

In terms of recording properties, the increase in the

initial remanence shape due to interactions is signif-

icant. It might be expected, for example, to lead to

Figure 10

Correlation function of the ﬁeld-dried media made from

sheared coating.

Figure 11

Hysteresis curve with and without interactions for zero-

ﬁeld media.

Figure 12

Hysteresis curve with and without interactions for ﬁeld-

dried media.

Figure 13

D.c. Demagnetization curve with and without

interactions for the zero-ﬁeld media.

651

Magnetic Reco rding: Particulat e Media, Micromagnetic Simulations

transition broadening given that the increased re-

manence shape represents a reduced stability against

demagnetization. A further factor in this respect is

the relationship between the irreversible suscepti-

bility w

irr

and the time-dependent magnetization, i.e.,

¼ w

irr

, where S

is the coefﬁcient of magnetic

viscosity and H

is the ﬂuctuation ﬁeld. The differ-

entiation of the remanence curve (equivalent to the

SFD) is an experimental approximation to w

irr

Hence, an increase in the SFD ﬁelds of the order of

those occurring in magnetization transitions must be

expected to lead to an increase in the magnetic vis-

cosity and a consequent reduction in the thermal sta-

bility of stored information. This effect is crucial to

the potential degradation of the data during the long-

term storage of magnetic media.

9. Conclusion

‘‘First principles’’ calculations of the magnetic prop-

erties of the coated medium, i.e., one without ad hoc

assumptions relating to the physical microstructure,

have been made by employing a micromagnetic mod-

el of coated particulate media used in conjunction

with a molecular dynamic simulation of the medium

microstructure.

The calculations give good qualitative agreement

with experiment and demonstrate a strong depend-

ence of the magnetic properties on the detail of the

coating and orientation process, as is to be expected.

Detailed comparisons with experimental data are dif-

ﬁcult because of the proprietary nature of informa-

tion relating to the coating process. However, this

approach could form the basis of a systematic exper-

imental investigation which would be of great beneﬁt

in understanding both the magnetization processes of

coated media and the effects of the coating process on

magnetic and recording properties.

See also: Magnetic Recording Media: Tape Particles,

Magnetic Properties; Magnetic Recording Technolo-

gies: Overview

Bibliography

Allen M P, Tildesley D J 1990 Computer Simulation of Liquids.

Oxford Science Publications, Oxford

Bertram H N, Seberino C 1999 Numerical simulations of hys-

teresis in longitudinal magnetic tape. J. Magn. Magn. Mater.

193, 388–94

Brown W F Jr. 1979 Thermal ﬂuctuations of ﬁne ferromagnetic

particles. IEEE Trans. Magn. 15, 1196–208

Chantrell R W, Wongsam M, Hannay J D, Schreﬂ T, Richter H

J 1998 Computational approaches to thermally activated fast

relaxation. IEEE Trans. Magn. 34, 1839–45

Coverdale G N, Chantrell R W, Satoh A, Chang C -R 1998 A

molecular-dynamics model of the magnetic properties and

microstructure of strongly interacting colloidal dispersions.

IEEE Trans. Magn. 34, 378–80

Coverdale G N, Chantrell R W, Satoh A, Veich R J 1997

A molecular dynamics model of the magnetic properties

and microstructure of advanced metal particle dispersions.

J. Appl. Phys. 81, 3818–20

Deymier P A, Jung C, Raghaven S 1994 Molecular dynamics of

magnetic particulate dispersions. J. Appl. Phys. 75, 5571–3

Fincham D 1993 Leapfrog rotational algorithms for linear

molecules. Mol. Simul. 11, 79–89

Gilson R G 1993 Drying of particulate media coatings.

J. Magn. Magn. Mater. 120, 97–9

Gunal Y, Visscher P B 1996 Brownian dynamics simulation of

magnetic colloid aggregation. IEEE Trans. Magn. 32, 4049–

Newman J J, Yarbrough R B 1969 Theory of the motions of a

ﬁne magnetic particles in a Newtonian ﬂuid. IEEE Trans.

Magn. 5, 320–4

O’Grady K, Gilson R G, Hobby P C 1991 Magnetic pigment

dispersions (a tutorial review). J. Magn. Magn. Mater. 95 (3),

341–55

Okamoto K, Okazaki Y, Nagai N, Uedaira S 1996 Advanced

metal particles technologies for magnetic tapes. J. Magn.

Magn. Mater. 155 (1–3), 60–6

Richter H J, Veitch R J 1996 MP tape for high-density digital

recording. J. Magn. Magn. Mater. 155 (1–3), 80–2

Figure 14

Switching ﬁeld distribution with and without

interactions for zero-ﬁeld media.

Figure 15

Switching ﬁeld distribution with and without

interactions for ﬁeld-dried media.

652

Magnetic Reco rding: Particulate Media, Micromagnetic Simulations

Veitch R J 1996 Shear magnetometry for the analysis of mag-

netic tape dispersions. IEEE Trans. Magn. 32 (5), 4028–33

Visscher P B, Gunal Y 1997 Field-induced smectic ordering in

model magnetic inks. J. Appl. Phys. 18, 3827–9

Visscher P B, Gunal Y 1998 Self assembly in model magnetic

inks. IEEE Trans. Magn. 34, 1687–9

Voss M J, Booth R L, Zhu J G, Carlson L W 1993 Computed

hysteresis behavior and interaction effects in spheroidal par-

ticle ensembles. IEEE Trans. Magn. 29, 3652–7

G. N. Coverdale

University of Durham, County Durham, UK

Magnetic Recording: Patterned Media

The term ‘‘patterned media’’ refers to the storage of

data in arrays of discrete magnetic elements, such as

small, lithographically patterned pillars or bars. The

direction of magnetization in each element represents

a binary one or zero. This differs from data storage in

conventional hard disk or ﬂexible media, in which

regions of a magnetic ﬁlm are magnetized to repre-

sent the stored data. Patterned media systems are in

their infancy, but have excited great interest because

they should in principle be capable of storing data at

much higher density than is possible with conven-

tional media.

In conventional hard disks, data are stored as

magnetization patterns written onto a continuous

thin ﬁlm (Fig. 1(a)). The ﬁlm consists of small grains

of an alloy such as CoCrTaPt, 10–15 nm in diameter.

The grains are exchange-decoupled due to segrega-

tion of nonmagnetic elements at the grain bounda-

ries. Each bit written onto the medium covers an area

(cell) of hundreds of grains. As recording densities

increase, it is desirable to reduce the grain size in

order to maintain a large number of grains per bit cell

and thus to obtain a good signal-to-noise ratio.

However, as grain sizes are decreased, the magneti-

zation of the grains become thermally unstable, lead-

ing to the decay of recorded data patterns. This

occurs when thermal energy, kT, is sufﬁcient to cause

reversal of the magnetization within individual

grains. Here k is Boltzmann’s constant and T is the

temperature. This superparamagnetic behavior oc-

curs when KVo60 kT (Lu and Charap 1994), where

K is the magnetic anisotropy energy per unit volume

of the material and V is the grain volume. This

behavior sets a lower limit on the grain size of the

medium, and therefore limits the achievable signal-

to-noise ratio of the media. Increasing the anisotropy

K allows V to be decreased, but for high K it becomes

difﬁcult to write data patterns on the medium

because unreasonably large write ﬁelds would be re-

quired. An upper limit for recording density on con-

ventional media of 40 Gbin

2

was suggested (Charap

et al. 1997). However these predictions have proven

to be pessimistic, and improvements in longitudinal

and perpendicular media have continued. In 2005, for

example, Hitachi demonstrated 230 Gb/in

recording.

Any practical patterned media will need to support

densities well above those achievable with continuous

thin ﬁlm media in order to be economically viable.

Patterned media have been suggested as a means of

increasing storage density beyond that achievable

with thin ﬁlm media. A patterned medium (Fig. 1(b)

and 1(c)) consists of an array of nanoscale magnets,

each of which can store one bit of information de-

pending on its magnetization state. It is important

that each element behaves as a single magnetic do-

main, whether it is single crystal or polycrystalline.

The elements in patterned media are still subject to

thermal instability, but now the volume V can be as

large as the entire volume of the element instead of

equal to the grain volume. The elements could be as

small as 10 nm, potentially representing an increase in

density of 1000 times over thin ﬁlm media. The dif-

ﬁculties inherent in patterned media storage schemes

concern the fabrication of large-area, uniform arrays

of magnetic particles, and developing methods for

reading and writing the arrays. This article will dis-

cuss how arrays can be made, what magnetic prop-

erties are desirable in a patterned medium, and what

schemes have been proposed for incorporating pat-

terned media into practical storage devices.

1. Fabrication of Magnetic Media

The concept of patterning a hard disk using lithog-

raphy was ﬁrst proposed as a means of improving

tracking and signal-to-noise ratio (Lambert et al.

1991). However, as thermal instability limitations in

conventional media became more apparent, pat-

terned media appeared to offer a route to higher

density recording. Despite this promise, significant

development of patterned media did not occur until

the 1990s when high resolution lithography became

widely available for fabrication of arrays of magnetic

elements (Chou et al. 1994, Chou 1997, White et al.

1997).

The fabrication requirements for patterned media

are challenging. To make a structure consisting of 250

G particles in

2

, an array of elements with a period

below 50 nm is needed. The lithography method used

must therefore be capable of resolving 25 nm features

or better over an area several centimeters or greater in

diameter. This is well beyond the capabilities of the

optical lithography currently used in semiconductor

device fabrication, so higher resolution patterning

techniques are necessary. Early studies of small, pat-

terned structures used optical lithography to make

arrays with micron-scale features, but the most com-

mon method now used to make submicron arrays

with features of 10 nm and above is electron-beam

653

Magnetic Reco rding: Patterned Media

lithography (e.g., Chou et al. 1994, New et al. 1996,

O’Barr et al. 1997). However, this is a serial method in

which each element is exposed sequentially, so it is too

slow for economic patterning of large areas. Other

lithography methods have therefore been explored,

including x-ray nanolithography (Rousseaux et al.

1995), nano-imprint lithography (Wu et al. 1998), and

interference lithography (Fernandez et al. 1996, Haast

et al. 1998, Ross et al. 1999), all of which are capable

of patterning large areas. Self-assembly processes

have been proposed as an alternative patterning

route. In these methods, the self-assembly of, for in-

stance, pores in anodized alumina (Bao et al. 1996) or

domains in block copolymers (Zhu et al. 1997, Cheng

et al. 2001) are used as templates for the patterning of

magnetic arrays. Magnetic particles can also be made

by chemical synthesis and assembled onto a surface to

form a regular array (Sun and Murray 1999). Self-

assembly techniques can pattern large area structures

with good short-range order, but have the drawback

that long-range order is often poor. High resolution

lithography processes and self-assembly methods have

been successfully used to make large area magnetic

arrays on a laboratory scale for media prototypes,

and scaling up of many of these methods is under

development. A promising avenue for patterned

media formation is the combination of top-down

lithography with self-assembly (Naito et al. 2002,

Cheng et al. 2004).

The magnetic materials themselves are formed us-

ing a variety of methods. Commonly, a mask is de-

ﬁned by lithography or self-assembly processes,

consisting of an array of holes in a layer made of

photoresist or other material. Magnetic material is

then electrodeposited through the mask to form an

array of pillars, or evaporated over the mask fol-

lowed by a lift-off process to leave an array of cones

or pillars. Alternatively, a ﬁlm of magnetic material

can be initially deposited onto a substrate, then a

template or etch mask is deﬁned on top of the mag-

netic ﬁlm by lithography. The exposed magnetic ma-

terial is then etched away to leave an array of

magnetic dots. Thin ﬁlms can also be patterned using

a focussed ion beam or other direct writing technique,

or by imprint lithography (Lohau et al. 2001, Mc

Clelland et al. 2002). A further method relies on the

self-assembly of pre-formed magnetic particles on a

surface. Examples of some of these structures are

shown in Fig. 2. Each deposition method has its

advantages. Electrodeposition is ideal for making

Figure 1

Schematic of (a) conventional thin ﬁlm medium, consisting of single-domain grains. Bits are represented as transitions

between regions of opposite net magnetization. Each bit occupies an area of tens to hundreds of grains. (b) Patterned

medium with in-plane magnetization. Now the bits are deﬁned lithographically and have period p. The bits can be

polycrystalline (indicated by dotted lines) or single crystal, but magnetically they act as single domains. (c) Patterned

medium with out-of-plane magnetization. The period is p, height h and diameter d. Binary one and zero are indicated.

(Reprinted with permission J. Vac. Sci. Technol. B17, 3168; r American Institute of Physics 1999.)

654

Magnetic Reco rding: Patterned Media

Figure 2

Examples of arrays of patterned elements. (a) An array of electrodeposited nickel pillars, height 115 nm, diameter

57 nm, period 100 nm. (b) An array of evaporated Ni pyramids, 30 nm diameter, with 100 nm period. (c) An array of

etched bars made from a Co/Cu/Co multilayer, 170 nm long, 70 nm wide, 5 nm thick. (d) An array of Ni bars made by

evaporation and liftoff, 20 nm wide, 200 nm long and 35 nm thick. (e) Electron micrograph of a two-dimensional self-

assembled layer of 9 nm diameter Co particles. Inset: high-resolution image of a single particle. (Reprinted with

permission from: (a–c) J. Vac. Sci. Technol. B 17 3168; r American Institute of Physics 1999; (d) Proc. IEEE 85, 652;

r IEEE 1997; (e) J. Appl. Phys. 85, 4325; r American Institute of Physics 1999.)

655

Magnetic Reco rding: Patterned Media

features with a large aspect ratio, but it is restricted to

certain compositions. Evaporation and liftoff is con-

venient for ﬂat or conical shapes but is also restricted

in composition. Deposition of a ﬁlm or multilayer

structure by sputtering followed by etching offers the

greatest range of compositions and magnetic proper-

ties, but etching of the material into dots is often

difﬁcult and the thickness of the ﬁlms is limited. The

methods chosen will depend on the desired geometry

and composition of the array.

2. Magnetic Properties of Patterned Media

The optimum magnetic properties of patterned media

(the magnetic anisotropy, coercivity or switching

ﬁeld, and magnetic moment) are not as clearly un-

derstood as the properties of hard disk or ﬂexible

media, and will depend on the specific recording de-

vice in which the medium is incorporated. However,

some general considerations apply to all patterned

media. Most importantly, the elements need to be

single-domain particles with a number (usually two)

of well-deﬁned stable magnetization states to repre-

sent the data bits. For common ferromagnetic mate-

rials, single-domain behavior occurs in sub-100 nm

particles because it is energetically unfavorable for

domain walls to form in very small particles. Hence

for a practical patterned medium with sub-100 nm

period, the elements will automatically be single-

domain as depicted in Fig. 1(b) and Fig 1(c).

A single-domain particle can be characterized by

its total magnetic anisotropy K, its switching ﬁeld

, and its saturation magnetization M

. The an-

isotropy K controls the direction and number of the

preferred magnetization axes. A uniaxial anisotropy

deﬁnes a single preferred magnetization direction. K

includes two major contributions, one arising from

the particle shape, K

, and the other from its mag-

netocrystalline anisotropy, K

(Cullity 1972). Needle-

shaped particles have a uniaxial K

with magnitude

that increases with the particle aspect ratio and with

; the preferred magnetization direction is parallel

to the long axis of the particle. K

can therefore be

controlled by choice of particle shape. In contrast, the

magnetocrystalline anisotropy K

depends on the

material composition and crystal structure, and

quantiﬁes the preference for the magnetization to lie

parallel to certain crystallographic directions within

the crystal. Some materials such as cobalt have a

strong K

, which in cobalt causes the magnetization

to lie parallel to the [0001] direction. By choice of

materials, microstructure (amorphous, polycrystal-

line, or single crystal) and particle shapes, the total

anisotropy can therefore be controlled over a wide

range. Most patterned media structures rely on shape

anisotropy, though magnetocrystalline anisotropy

has been used to control magnetization in arrays of

iron rectangles (New et al. 1996).

The total anisotropy K is important because it

governs the switching ﬁeld H

, which is the magnetic

ﬁeld required to change the magnetization of an el-

ement from one direction to the other. This is the

most important magnetic parameter in terms of the

writing process because the recording system must

apply a ﬁeld of at least H

to write data on the

medium. H

is determined both by K and by the

mechanism of magnetization reversal. In the simplest

case, the magnetization of a single domain particle

reverses in a uniform manner known as coherent re-

versal (Stoner and Wohlfarth 1948) and the switching

ﬁeld is simply given by H

¼2K/M

when the ﬁeld is

applied parallel to the anisotropy axis. However,

measured switching ﬁelds are often smaller than this

due to nonuniform magnetization reversal processes

in which the magnetization reversal might occur by

curling (Frei et al. 1957), or it might start in one

region and propagate through the entire particle

(Fredkin et al. 1991, Gibson and Schultz 1993, Se-

berino and Bertram 1997). The switching ﬁeld for

nonuniform reversal depends on particle size and

shape, but as the particles become smaller, coherent

switching becomes the dominant reversal mode and

the switching ﬁelds of particles increase (e.g., Chou

1997). Clearly, a lower H

will make the medium

easier to write, but it will also make it more vulner-

able to thermal instability (superparamagnetism)

and to accidental erasure. Most patterned media

prototypes have switching ﬁelds of several hundred of

oersteds.

In terms of reading data back from the medium,

most recording devices are designed to detect the

magnetic ﬁeld surrounding each element. For single

domain particles, the ﬁeld approximates to a dipole

ﬁeld whose magnitude is proportional to M

and to

the particle volume (Cullity 1972). This ﬁeld decreas-

es as the inverse cube of the distance from the ele-

ment. A stronger ﬁeld will make detection of the

magnetization state of the element easier. However, if

the ﬁeld from an element is too strong, it may cause

inadvertent reversal (writing) of neighboring parti-

cles, which is undesirable. If the nearest-neighbor

interaction ﬁeld is H

, then in a square array of per-

pendicular magnetic dipoles, the maximum ﬁeld

experienced by one element due to all of its neigh-

bors is 9H

. Thus a criterion for arrays such as

Fig. 1(c) to support stable magnetization patterns is

that H

must exceed 9H

(Ross et al. 1999). This

limits how close together the elements can be placed.

Arrays in which the element separation is similar to

the element diameter are usually found to be stable.

3. Recording Systems

We have seen that a patterned medium can be de-

signed by choosing the particle size, shape, and ma-

terial to determine the switching ﬁeld of the elements

656

Magnetic Reco rding: Patterned Media

and by choosing the array geometry so that the

elements do not interact too strongly. A variety of

lithographic and self-assembly processes combined

with deposition and etching processes have been de-

veloped to make arrays with period of 100 nm and

smaller. However, the development of practical re-

cording systems has lagged behind the development

of the media.

Development of patterned recording systems has

followed two distinct paths as illustrated in Fig. 3.

The ﬁrst is a spinning-disk system, similar to a con-

ventional hard disk system but using a patterned in-

stead of a continuous hard disk medium. This has the

advantage that much of the existing hard drive tech-

nology is applicable, including the basic design of the

drive, actuators and other mechanical components,

and signal detection circuitry. The major challenges

for this device are to produce heads with extremely

narrow widths, to follow accurately the lithograph-

ically-deﬁned tracks, and to synchronize the writing

process with the physical location of the lithograph-

ically-deﬁned bits. In such a scheme the patterned

elements may be designed with in-plane (longitudi-

nal) magnetization directions such as Fig. 1(b), sim-

ilar to a conventional hard disk, or alternatively

the elements can have out-of-plane (perpendicular)

magnetization directions such as Fig. 1(c). Modeling

of the readback response of both types of media has

been carried out (Nair and New 1998, Boerner et al.

1999), and the reading and writing of patterned

media has been demonstrated (Todorovic et al. 1999,

Zhu et al. 2000, Albrecht et al. 2003) as shown in

Fig. 4, but a practical spinning-disk system has not

been made.

A more radical departure from conventional re-

cording is a probe-based system (Ohkubo et al. 1991,

Mamin et al. 1995). In this case, a scanning probe tip

similar to an atomic force microscope addresses an

array of elements. The elements are perpendicularly

magnetized as shown in Fig. 3(b). The probe tip is

magnetic and is mounted on the end of a ﬂexible

cantilever. The probe detects the magnetization state

of each element by responding to the force gradient

exerted by the magnetic ﬁeld emanating from the

element, which causes a change in the resonant fre-

quency of the cantilever. This is the same as the

operation of a magnetic force microscope. To write

data patterns on the medium, a magnetic ﬁeld less

than H

is applied to all the elements using an ex-

ternal coil. The magnetic probe tip is then moved

close to the target element, so that the sum of the ﬁeld

from the probe tip plus the external ﬁeld is sufﬁcient

to reverse that one element but not its neighbors

(Gibson and Schultz 1993, Kleiber et al. 1998). Var-

iations on this theme are also possible, for instance

using a separate tip for reading and writing (Chou

1997, Kong et al. 1997). An example of an array with

written features is shown in Fig. 5.

These probe-based systems are inherently slow be-

cause they rely on moving the probe tip to specific

Figure 3

Recording schemes using patterned media. (a) A spinning

disk. The disk is patterned with elements arranged in

circumferential tracks. The medium is written and read

using an ultra-narrow recording head. (b) A scanning

probe. A probe tip detects the magnetic ﬁeld from

perpendicular patterned elements. An external ﬁeld, plus

the ﬁeld from the probe tip, is used to write data.

Figure 4

A set of images taken using a magnetoresistive recording

head which was dragged across an array of nickel pillars

with 2-mm spacing. The white and dark images represent

up and down magnetizations of the pillars and the

linescans show the readback signal from the head.

Several different patterns have been written and read

back. (Reprinted with permission from Appl. Phys. Lett.

74, 2516; r American Institute of Physics 1999.)

657

Magnetic Reco rding: Patterned Media

locations in the patterned medium. To increase the

data rate, a large number of probe tips working in

parallel is necessary (Pohl 1995, Minne et al. 1996,

Binnig et al. 1999). In these devices, the challenges are

to make the device compact and mechanically robust,

and to design a reliable method for writing the un-

derlying patterned magnetic media.

One other storage scheme that should be men-

tioned in relation to patterned media is the magnetic

random access memory (MRAM) (e.g., Tehrani et al.

1999), which is described in more detail in the article

Random Access Memories: Magnetic . These devices

consist of an array of magnetoresistive elements,

connected together by conductor lines. Each element

stores one data bit, as in the patterned media systems

described above, but the elements in a MRAM con-

sist of more complex, multilayer ﬁlms such as spin-

valve or spin-tunneling structures such as the sample

shown in Fig. 2(c). Many of the issues of inter-ele-

ment interactions, thermal stability, and switching

behavior, as well as the need for high resolution lith-

ographic patterning, are common to both MRAM

elements and patterned media.

4. Summary

The onset of thermal instability (superparamagnet-

ism) in conventional hard-disk media is driving the

development of patterned media systems for areal

densities exceeding hundreds of Gbin

2

. Patterned

medium structures have already been fabricated using

a range of lithographic and deposition techniques,

and the switching ﬁeld and interactions between the

elements can be controlled by the choice of materials

and array geometry. The development of practical

storage schemes has lagged behind the development

of media, but prototype systems have been demon-

strated. Patterned media will most likely ﬁrst appear

in a spinning-disk system, but probe-based systems

have also been explored.

See also: Longitudinal and Perpendicular Recording:

Thermal Stability; Magnetic Recording Devices:

Future Technologies; Magnetic Recording Media:

Advanced; Magnetic Recording Technologies: Over-

view; Random Access Memories: Magnetic

Bibliography

Albrecht M, Rettner C T, Best M E, Terris B D 2003 Magnetic

coercivity patterns for magnetic recording on patterned

media. Appl. Phys. Lett. 83, 4363

Bao X H, Li F Y, Metzger R M 1996 Synthesis and magnetic

properties of electrodeposited metal particles on anodic alu-

mite ﬁlm. J. Appl. Phys. 79, 4866–8

Binnig G, Despont M, Drechsler U, Haberle W, Lutwyche M,

Vettinger P, Mamin H J, Chui B W, Kenny T W 1999

Ultrahigh density atomic force microscopy data storage with

erase capability. Appl. Phys. Lett. 74, 1329–31

Boerner E D, Bertram H N, Hughes G F 1999 Writing on

perpendicular patterned media at high density and data rate.

J. Appl. Phys. 85, 5318–20

Charap S H, Lu P L, He Y 1997 Thermal stability of recorded

information at high densities. IEEE Trans. Magn. 33, 978–83

Cheng J Y, Ross C A, Chan V Z H, Thomas E L, Lammertinck

R G H, Vancso G J 2001 Formation of a cobalt magnetic dot

array via block copolymer lithography. Advanced Materials

13, 1174

Cheng J Y, Mayes A M, Ross C A 2004 Nanostructure

engineering by templated self-assembly of block copolymers.

Nature Materials 3, 823

Chou S Y 1997 Patterned magnetic nanostructures and quan-

tized magnetic disks. Proc. IEEE. 85, 652–71

Chou S Y, Wei M, Krauss P R, Fisher P B 1994 Study of

nanoscale magnetic structures fabricated using electron-beam

lithography and quantum magnetic disk. J. Vac. Sci. Technol.

B12, 3695–8

Cullity B D 1972 Introduction to Magnetic Materials. Addison-

Wesley, Reading, MA

Fernandez A, Bedrossian P J, Baker S L, Vernon S P, Kania D

R 1996 Magnetic force microscopy of single-domain cobalt

dots patterned using interference lithography. IEEE Trans.

Magn. 32, 4472–4

Fredkin D R, Koehler T R, Smyth J F, Schultz S 1991 Mag-

netization reversal in permalloy particles: micromagnetic

computations. J. Appl. Phys. 69, 5276–8

Frei E H, Shtrickman S, Treves D 1957 Critical size and

nucleation ﬁeld of ideal ferromagnetic particles. Phys. Rev.

106, 446

Figure 5

(a) An array of nickel bars formed by evaporation and

liftoff, (b) magnetic image of sample after several bars

have been reversed to delineate the letter L, using a

magnetic probe tip. (Reprinted with permission from

IEEE Trans. Mag. 33, 3019; r IEEE 1997.)

658

Magnetic Reco rding: Patterned Media

Gibson G A, Schultz S 1993 Magnetic force microscope study

of the micromagnetics of submicrometer magnetic particles.

J. Appl. Phys. 73, 4516–21

Haast M A M, Schuirhuis S, Abelmann L, Lodder J C, Popma

Th J 1998 Reversal mechanism of submicron patterned

CoNi/Pt multilayers. IEEE Trans. Magn. 34, 1006–8

Kleiber M, Kummerlen F, Lohndorf M, Wadas A, Weiss D,

Wiesendanger R 1998 Magnetization switching of submi-

crometer Co dots inﬂuenced by a magnetic force microscope

tip. Phys. Rev. B 58, 5563–7

Kong L S, Zhuang L, Chou S Y 1997 Writing and reading on

7.5 Gb/in

longitudinal quantized magnetic disk using mag-

netic force microscope tips. IEEE Trans. Magn. 33, 3019–21

Lambert S E, Saunders I L, Patlach A M, Krounbi M T,

Hetzler S R 1991 Beyond discrete tracks: other aspects of

patterned media. J. Appl. Phys. 69, 4724–6

Lohau J, Moser A, Rettner C T, Best M E, Terris B D 2001

Writing and reading perpendicular magnetic recording media

patterned by a focused ion beam. Appl. Phys. Lett. 78, 990

Lu P L, Charap S H 1994 Magnetic viscosity in high density

recording 1994. J. Appl. Phys. 75, 5768–70

Mamin H J, Terris B D, Fan L S, Hoen S, Barrett R C, Rugar

D 1995 High density storage using proximal probe tech-

niques. IBM J. Res. Dev. 39 , 681–99

Mc Clelland G M, Hart M W, Rettner C T, Best M E, Carter

K R, Terris B D 2002 Nanoscale patterning of magnetic

islands by imprint lithography using a ﬂexible mold. Appl.

Phys. Lett. 81, 1483

Minne S C, Manalis S R, Atalar A, Quate C F 1996 Independ-

ent parallel lithography using the atomic force microscope.

J. Vac. Sci. Technol. B 14, 2456–61

Nair S K, New R M H 1998 Patterned media recording: noise

and channel equalization. IEEE Trans. Magn. 34, 1916–8

Naito K, Hieda H, Sakurai M, Kamata Y, Asakawa K 2002

2.5-inch disk patterned media prepared by an artiﬁcially

assisted self-assembling method. IEEE Trans. Magn. 38, 1949

New R M H, Pease R F W, White R L, Osgood R M, Babcock

K 1996 Magnetic force microscopy of single-domain single-

crystal iron particles with uniaxial surface anisotropy.

J. Appl. Phys. 79, 5851–3

O’Barr R, Yamamoto S, Schultz S, Xu W, Scherer A 1997

Fabrication and characterization of nanoscale arrays of nick-

el columns. J. Appl. Phys. 81, 4730–2

Ohkubo T, Kishigami J, Yanagisawa K, Kaneko R 1991 Sub-

micron magnetizing and its detection based on the point

magnetic recording concept. IEEE Trans. Magn. 27, 5286–8

Pohl D W 1995 Some thoughts about scanning probe microscopy,

micromechanics and storage. IBMJ.Res.Dev.39, 701–11

Ross C A, Smith H I, Savas T, Schattenburg M, Farhoud M,

Hwang M, Walsh M, Abraham M C, Ram R J 1999 Fab-

rication of patterned media for high-density magnetic stor-

age. J. Vac. Sci. Technol. B 17, 3168–76

Rousseaux F, Decanini D, Carcenac F, Cambril E, Ravet M F,

Chappert C, Bardou N, Bartenlian B, Veillet P 1995 Study of

large area high-density magnetic dot arrays fabricated using

synchrotron radiation-based x-ray lithography. J. Vac. Sci.

Technol. B 13, 2787–91

Seberino C, Bertram H N 1997 Micromagnetics of long ferro-

magnetic particles. IEEE Trans. Magn. 33, 3055–7

Stoner E C, Wohlfarth E P 1948 A mechanism of magnetic

hysteresis in heterogeneous alloys. Philos. Trans. R. Soc.

London A240, 599

Sun S H, Murray C B 1999 Synthesis of monodisperse cobalt

nanocrystals and their assembly into magnetic superlattices.

J. Appl. Phys. 85, 4325–30

Tehrani S, Chen E, Durlam M, DeHerrera M, Slaughter J M,

Shi J, Kerszykowski E 1999 High-density submicron magne-

toresistive random access memory. J. Appl. Phys. 85, 5822–7

Todorovic M, Schultz S, Wong J, Scherer A 1999 Writing and

reading of single magnetic domain per bit perpendicular pat-

terned media. Appl. Phys. Lett. 74, 2516–8

White R L, New R M H, Pease R F W 1997 Patterned media: a

viable route to 50 Gb/in

and up for magnetic recording?

IEEE Trans. Magn. 33, 990–5

Wu W, Cui B, Sun X, Zhang W, Zhuan L, Kong L, Chou S Y

1998 Large-area high-density quantized magnetic disks fab-

ricated using nanoimprintlithography. J. Vac. Sci. Technol.

B16, 3825–9

Zhu J G, Lin X, Guan L, Messner W 2000 Recording, noise

and servo characteristics of patterned thin ﬁlm media. IEEE

Trans. Magn. 36, 23–9

Zhu S, Gambino R J, Rafailovich M N, Sokolov J, Schwarz S

A, Gomez R D 1997 Microscopic magnetic characterization

of submicron cobalt islands prepared using self-assembled

polymer masking technique. IEEE Trans. Magn. 33, 3022–4

C. A. Ross

Massachusetts Institute of Technology, Cambridge

Massachusetts, USA

Magnetic Recording: Rigid Media,

Preparation

Hard disks have various formats with diameters var-

ying from 3.5 in (8.9 cm) for PCs to 2.5 in (6.4 cm) and

to 1.8 in (4.6 cm) for laptop and notebook computer

drives. The disk has a central hole for location on the

spindle of the disk drive motor. The rotating disk is

addressed by read/write head(s) ‘‘ﬂying’’ within tens

of nanometers of its surface. The disk substrate, on

which the magnetic ﬁlm is deposited, varies in thick-

ness from about 1.25 mm for 3.5 in disks to 0.5 mm for

the smaller format disks. Rotation speeds for the disk

are continually increasing to decrease access times,

and speeds approaching 10000 rpm are now used.

A cross-section of the structure on one side of a

conventional disk is shown in Fig. 1. The ﬁrst layer is

often a thick, mechanically hard and nonmagnetic

ﬁlm on which is deposited one or more underlayers or

seed layers that encourage a preferred microstructure

and crystallographic texture in the magnetic layer. It

can be seen that the magnetic ﬁlm is relatively thin.

The hard coating and the lubricating layer at the

head/disk interface protect the magnetic storage layer

in any contact stop–start phase of the head slider

ﬂight and also assist in that ﬂight.

1. The Substrate

The most common disk substrate material is an Al–

4% Mg alloy. The disk is prepared by stamping, and

659

Magnetic Reco rding: Rigid Media, Pr eparation