perpendicular magnetic recording

PERPENDICULAR MAGNETICRECORDING

This page intentionally left blank

Perpendicular MagneticRecording

by

Sakhrat KhizroevCenter for Nanoscale Magnetic Devices,Department of Electrical and Computer Engineering,Florida International University,Miami, Florida, U.S.A.

and

Dmitri LitvinovUniversity of Houston,Houston, Texas, U.S.A.

Center for Applied Nanomagnetics,

KLUWER ACADEMIC PUBLISHERSNEW YORK, BOSTON, DORDRECHT, LONDON, MOSCOW

http://www.springerlink.com

eBook ISBN: 1-4020-2723-0Print ISBN: 1-4020-2662-5

©2005 Springer Science + Business Media, Inc.

Print ©2004 Kluwer Academic Publishers

All rights reserved

No part of this eBook may be reproduced or transmitted in any form or by any means, electronic,mechanical, recording, or otherwise, without written consent from the Publisher

Created in the United States of America

Visit Springer's eBookstore at: http://ebooks.springerlink.comand the Springer Global Website Online at: http://www.springeronline.com

Dordrecht

http://ebooks.springerlink.com

http://www.springeronline.com

TABLE OF CONTENTS

Page

Preface

Acknowledgements

CHAPTER 1

Fundamentals of Perpendicular Recording 1

1. A Historical Perspective 1

2. Superparamagnetic Limit 2

3. Dodging the Superparamagnetic Limit 5

3.1 Strong write field 5

3.2 Perfectly aligned media 9

3.3 Absence of the demagnetization field in bit transitions 10

4. Soft underlayer as a new system component 11

4.1 SUL as a major source of noise 12

4.2 SUL magnetic moment 13

4.3 SUL thickness 14

4.4 SUL influence on the resolution 15

5. Playback: New Signal Processing Schemes 16

6. Challenges of New Materials 17

6.1 Hard layer materials 18

6.2 High anisotropy SUL materials 21

7. How Far Will Perpendicular Recording Go 21

CHAPTER 2

Physics of Writing 23

1. Introduction 23

1.1 Chapter overview 24

2. Different Modes of Perpendicular Recording 24

ix x

2.1 Second perpendicular mode: a ring head and a perpendicular

medium without a soft underlayer 25

vi

2.1.1 Gap length dependence 28

2.1.2 Trailing pole thickness dependence 30

2.2 First perpendicular mode: a single pole head and a perpendicular

medium with a soft underlayer 31

2.2.1 Magnetic image model 31

2.2.2 Permanent magnet approximation 32

2.2.3 Recording by the field in the gap (perpendicular) versus

recording by the field fringing from the gap (longitudinal) 35

2.2.4 Is the increase of the recording field due to the use of a SUL

sufficient for adequate recording? 36

2.2.5 Quadruple ratio between saturation currents in

perpendicular and longitudinal recording 37

2.2.6 Focused-ion-beam trimmed single pole heads 38

2.2.7 Example 1: FIB trimming of a wide-track Censtor SPH into

a narrow-track SPH 38

2.2.8 Example 2: FIB trimming of a RH into a narrow-track SPH 39

2.2.9 Single pole head: design strategy 42

2.2.10 Definition of efficiency 43

2.2.11 Throat height dependence 44

2.2.12 Dependence on the pole trackwidth and thickness 50

2.2.13 Skew angle sensitivity of single pole head 52

2.2.14 Gap length dependence 56

Field modelling 58

SPH efficiency versus RH efficiency 63

Skew angle versus gap length 66

Single pole head of type III 68

Experiments to compare different types of SPH’s 68

2.2.15 Flying height limitation of single pole head design 70

2.2.16 Multiple magnetic image reflection 72

2.3 Modified first perpendicular mode: a shielded single pole head

and a perpendicular medium with a soft underlayer 79

2.3.1 Shielded single pole head 80

CHAPTER 3

Physics of Playback 85

1. Introduction 85


2. Playback in Perpendicular Recording 86

2.1 Analysis Methods 86

2.1.1 Direct calculation with Point-size Reader Approximation 86

2.1.2 Calculation based on the Reciprocity Principle 98

vii

2.1.3 Model of the magnetic image 101

Image “paradox” 102

2.1.4 Examples of reader designs 104

2.1.5 Basic reader comparison 106

2.1.6 Parallels between perpendicular and longitudinal recording 108

2.1.7 Influence of shields 111

Number of shields 111

Shield thickness 113

2.1.8 Soft underlayer versus no soft underlayer 113

2.1.9 Differential reader optimisation and single MR differential readers 114

2.1.10 Parallels between playback in perpendicular and

longitudinal magnetic recording: revisited 118

Overview of reader designs 118

Conformal mapping 120

Use of a soft underlayer 123

3-D BEM calculation 124

Conclusions on study of parallels between playback in

perpendicular and longitudinal recording 126

CHAPTER 4

Perpendicular Recording Media 127

1. Introduction 127


2. Perpendicular Recording (“Hard”) Layer 128

2.1 Types of Media 128

2.2 Continuous Media 131

2.3 Magnetic Field Calculation 131

2.4 Demagnetisation Field in Perpendicular Recording Layer 135

2.5 Stray Field from Perpendicular Magnetic Media 137

2.6 Well-defined Perpendicular Easy Axis: Thicker Recording Layer? 142

3. Soft Underlayer 144

3.1 Saturation Moment 145

3.2 Thickness of Soft Underlayer 146

3.3 SUL-to-ABS Separation 149

3.4 Anisotropy: Micromagnetics of SUL 150

3.5 Magnetic Biasing 153

3.6 Dynamics of Perpendicular Recording 155

3.6.1 Design of a high-data-rate perpendicular system 156

3.6.2 Kerr microscopy 159

References 162

Index 173

Preface

This book is intended to be of use to two groups of readers. The first group includes

engineers and scientists who have good familiarity with conventional magnetic recording

and intend to investigate perpendicular magnetic recording in more detail. The second

group comprises mostly graduate students and those who need at least a casual or,

perhaps, detailed knowledge of the applications of the physics of magnetism in ultra-high

density (with nanoscale bit and transducer dimensions) magnetic recording.

Today, many leading companies in the multi-billion-dollar data storage industry

demonstrate an unprecedented interest in perpendicular magnetic recording. It is

commonly believed that perpendicular magnetic recording is the most likely alternative

to the conventional magnetic technology – longitudinal magnetic recording.

Longitudinal magnetic recording that has been the core technology since the inception of

the magnetic data storage industry more than half a century ago is finally coming to terms

with reality. Reality screams that the areal density is limited by thermal instabilities in the

longitudinal magnetic media at areal densities not far beyond 100 Gbit/in2. As never

before, the industry is desperate to find an alternative technology that could maintain the

“usual” steady areal density growth rate to which it got used over the many years in the

past. This explains why recently perpendicular magnetic recording could suddenly

revitalize such a strong attention from the data storage community. Out of a number of

alternatives, perpendicular magnetic recording is technically the closest technology to

longitudinal recording. Therefore, with a switch to perpendicular recording it would cost

the least for the industry to change its current infrastructure. Perpendicular recording

promises to defer the superparamagnetic density limit in the magnetic media by at least a

factor of ten compared to longitudinal recording.

As companies in the ultra-competitive data storage industry promptly get in the global

race to develop perpendicular magnetic recording, many engineers find themselves not

adequately trained and experienced in the new technology. Although perpendicular

magnetic recording is similar to the conventional technology, it still has a number of

peculiarities and open issues that have never been encountered in longitudinal recording

and thus remain to be understood and resolved. The authors of this book believe that to

adequately address these issues, every engineer in the field should acquire sufficient

knowledge of the physics of perpendicular magnetic recording and be well aware of the

fundamental and sometimes barely perceptible differences between perpendicular and

longitudinal recording.

ix

The purpose of this book is to provide engineers and graduate students with the basic

knowledge in perpendicular magnetic recording. The book’s emphasis is on the basic

physics of perpendicular magnetic recording rather than a detailed description of one or

another particular technical implementation. The authors attempt to provide the most

basic guidelines to design a complete system to record and retrieve information from a

perpendicular magnetic medium with areal densities up to one terabit per square inch and

beyond. This defines also the structure of this book. The book is divided into four

chapters. Chapter 1 provides background on the most recent development in the field and

describes open questions and issues in perpendicular recording. Chapter 2 and Chapter 3

cover the recording (writing) and playback mechanisms, respectively. In these chapters,

not only the authors explain the physics of data recording and playback but also propose

technical solutions to the design optimization of recording and playback transducers.

Chapter 4 introduces the essentials of perpendicular magnetic recording media.

Throughout the book, the authors compare perpendicular and magnetic recording.

Currently, the authors of this book, Khizroev and Litvinov, are with Florida International

University and the University of Houston, respectively. The material presented in this

book has been collected as a result of several years of research of the authors together

with Florida International University, the University of Houston, Seagate Research, IBM

Almaden Research Center, and Carnegie Mellon University. Based on the

accomplishments of this research, Khizroev and Litvinov have co-authored over 12

patents that were granted to Seagate and IBM.

Acknowledgements

The authors would like to acknowledge the numerous insightful discussions on different

topics of perpendicular magnetic recording with Mark Kryder, Kent Howard, Roy

Chantrell, Dieter Weller, Song Xue, Erik Svedberg, Bin Lu, Alex Shukh, and many

others of Seagate Technology, David Thompson of IBM Research, Hal Rosen, Kurt

Rubin, Alex Taratorin, Pat Arnett, Yoshihiro Ikeda, Margaret Best, Neil Smith, Roger

Wood, Andy Moser, Wipul Jayasekara, Y. Sonobe, and many others of IBM Almaden

Research and currently of Hitachi Global Storage Division, Michael Mallary, Adam

Torabi, Anna Kostrov, and others of Maxtor Corporation, Stanley Charap, David

Lambeth, Robert White, Jim Bain, Albert Theile, Jimmy Zhu, David Laughlin, and others

of Carnegie Mellon University, Roman Chomko and Venkatesan Renugopalakrishnan of

Florida International University, Neil Bertram of the University of California at San

Diego, Heng Gong of Iomega Corporation, Jack Judy of the University of Minnesota,

Leon Abelmann and Cock Lodder of the University of Twente, Mark Re, Francis Liu,

Jing Zhang, Kroum Stoev, and others of ReadRite Corporation, Jim Miles of the

University of Manchester, Gerardo Bertero and David Wachenschwanz of Komag

Corporation, Yasushi Kanai of Niigata Institute of Technology, Kevin O’Grady and Jing

Wu of York University, Masaaki Futamoto of Hitachi Central Research Laboratory,

Hiroaki Muraoka of the University of Tohoku, Carolyn Ross of Massachusetts Institute

of Technology, Ben Hu of Headway Corporation, Paul Frank of Information Storage

x

Industry Consortium, Robert Doyle and Hideo Fujiwara of the University of Alabama,

and many others. The authors would like to express special gratitude to David

Thompson, retired IBM Fellow, for encouraging, challenging, mentoring, and supporting

this research since 1996. The authors would like to thank Shun-ichi Iwasaki who has

always been an inspiring symbol and supporter of the international efforts to develop

perpendicular magnetic recording. The authors would like to thank Simona Stefanescu

for invaluable help with reading the manuscript and the derivation and analysis of many

analytical expressions used throughout the book. Finally, the authors greatly appreciate

the dedicated efforts of Helga Melcherts of Kluwer Academic Publishers during the

preparation of the manuscript.

S. Khizroev, Miami

D. Litvinov, Houston

February 2004

xi

Chapter 1 Fundamentals of Perpendicular Recording

1

Chapter 1

Fundamentals of Perpendicular Recording

1. A Historical Perspective

It is believed that formally perpendicular magnetic recording [1,2,3,4] was proposed for

the first time by Professor Shun-ichi Iwasaki about two decades ago. However, sporadic

research efforts on the development of perpendicular recording started much earlier. It is

very likely that perpendicular recoding was mentioned for the first time in a new

computer design program at the University of California at Berkeley in the late 1940s [5].

This program was funded by the Office of Naval Research to pursue an intermediate

sized computer based on a magnetic drum memory. Among the first companies that have

demonstrated a serious interest in perpendicular magnetic recording are IBM Corporation

[1] and Ampex Corporation [3]. From 1955 until 1961, perpendicular recording was the

major disk-drive project at IBM. Ampex Corporation was another company that

considered perpendicular recording as a solution in the magnetic tape industry in the late

1950s. However, despite such a long history, because of the strong position of the

conventional technology (longitudinal magnetic recording), there have been only a few

products based on perpendicular recording. Among these products are a hard disk from

Censtor Corporation [6] and a floppy disk from Toshiba Corporation [7].

Today, as the conventional magnetic recording technology is finally facing its

fundamental limit due to thermal instabilities in the longitudinal magnetic media [8], the

strong interest in perpendicular recording as the prime alternative is coming back [9,10].

As envisioned by industry and academia leaders, perpendicular recording is the most

likely candidate for the technology to be implemented in the next generations of hard

disks. The most competitive virtue of this technology is the fact that while being

technically the closest alternative to longitudinal recording, it is capable of deferring the

(superparamagnetic) density limit beyond what is achievable with longitudinal recording.

Perpendicular Magnetic Recording 2

It is believed that perpendicular magnetic recording will enable to retain the current rapid

technology advancement for the next several generations of magnetic storage solutions.

The chapter will cover the basic principles underlying perpendicular recording as well as

the challenges associated with implementing the technology [11,12,13,14].

S N N S N SS N S NN S

Inductive

“Ring” Writer

MR Reader

Magnetizing

Coil

Write field Recording Media

Figure 1. A schematic diagram of a conventional longitudinal recording scheme employed in today’s hard

drives.

2. Superparamagnetic Limit

Data on a magnetic recording medium are stored by means of recording certain spatial

variations of the magnetization, where the variations represent the data. The relation

between the data and the magnetization pattern is defined through the special data

encoding. Figure 1 shows a simplified schematic diagram of a conventional longitudinal

recording system. The recording medium is engineered so that the preferred direction of

the magnetization, a so-called easy axis, lies in the plane of the recording layer. Using an

inductive “ring”-type writer, the magnetization of the grains (in the medium) is aligned

along the track in either positive or negative direction. The recorded data pattern is read

back using a magnetoresistive element. A change or no change in the magnetization

direction at the bit transitions corresponds to a 1 or to a 0, respectively. The lateral

dimensions of a bit, the smallest feature realized in a particular drive design, define the

areal bit density that the drive could support.

A conventional magnetic medium has a granular structure such that each bit consists of

several magnetic grains or magnetic clusters. The magnetic clusters/grains are usually

shaped irregularly and are randomly packed, as shown in Figure 2a. Consequently, the

Chapter 1 Fundamentals of Perpendicular Recording 3

recording bits and bit transitions are usually not perfect, which is illustrated in Figure 2b.

These imperfections lead to noise in the playback signal. The noise is kept below a

certain acceptable level by means of including a sufficiently large number of magnetic

grains into each bit. The resulting averaging reduces the level of noise. Therefore, to

satisfy the Scaling Law, as the areal density increases, the bit size and the average size of

the grains that constitutes each bit should be decreased. Independently, Figure 2b also

illustrates that the reduction of the average grain size (which is necessary to maintain the

signal-to-noise ratio with the density increase) results in the reduction of the statistically

defined bit transition. Typical grains in today’s media range from 5 to 15 nm.

Magnetic

grains

Bit transition(a) (b)

Figure 2. (a) A transmission electron micrograph (TEM) of a typical granular medium. (b) A schematic

diagram of a single bit transition in a granular medium.

One of the critical factors characterizing the reliability of a data storage device is the data

stability. Various parameters control the stability of the data against external factors.

Relative to the ambient temperature, which is manifested by thermal fluctuations in the

recording medium, the magnetic anisotropy energy stored in each magnetic grain is one

of the major determinants (assuming that the grains are magnetically independent). The

magnetic anisotropy energy defines the approximate amount of energy necessary to

reverse the direction of the magnetization of a grain. For a single grain, it is equal to

KUV, where KU is the magnetic anisotropy energy per unit volume and V is the volume

of the grain, as shown in a schematic diagram in Figure 3. According to the Statistical

Physics, even for a relatively large value of KU (compared to the characteristic thermal

energy density) there is a finite probability for the grain to reverse its magnetic moment

due to thermal fluctuations, as given by the following Expression.

)(exp0Tk

Eff

B, (1)

where subscripts “+” and “-“ stand for the “upward” and “downward” magnetization

directions, respectively, f0 is the characteristic frequency in the range from 109 to 1012 Hz

[58]. The exact value depends on the intrinsic medium properties linked to the quantum-


mechanical interactions within the grain, E is the energy barrier between the two energy

states. In the extreme case, E = KUV. (It should be reminded that the demagnetization

field reduces the effective energy barrier. The demagnetization field strongly depends on

the recording mode: it achieves its maximum in the bit transitions in longitudinal media,

while it achieves its minima in the bit transitions in perpendicular media.) For a medium

to be thermally stable, the above quantity KUV should be substantially greater (>~ 40

times) than the energy of a single quantum of thermal fluctuation, kBT, where kB is

Boltzman’s constant and T is the temperature [8]. As the ratio, KUV/ kBT , approaches

this “magical” number (~40), due to the above described exponential dependence, the

relaxation time, = 1/f, (which defines for how long can a particle remain in a stable

state) drastically changes. For example, Charap et el showed that when for a typical

longitudinal medium the ratio is reduced from 60 to 25, the relaxation time drops from

more than 3.5 x 106 years to 72 seconds, respectively [8]. From the above discussions,

one could derive an approximate “stability” expression to evaluate the minimum average

grain size below which the longitudinal recording medium would become thermally

unstable. (Conventionally, the boundary value for the ratio KUV/ kBT is assumed to be

40).

U

B

B

U

K

Tk

Tk

VK V 4040 , (2)

3 40min

1~U

B

KTk

tyArealDensiaa , (3)

As mentioned above, according to the Law of Scaling, higher areal densities require

smaller grain sizes. It follows that to sustain thermal stability with reduction of the

average grain size, the anisotropy energy density KU should be increased. There is a

choice of potential medium materials (e.g., Co/Pd multilayer) that would provide the

necessary increase in the anisotropy energy density. Unfortunately, as KU increases, so

does the write field necessary to efficiently write onto the medium. The problem lies right

here. It is believed that the saturation magnetization for “soft” magnetic materials (of

which recording heads are made) is fundamentally limited [15]. The current state-of-the-

art recording heads already use materials with magnetization close to the predicted

fundamental limit, ~ 26 kGauss.

In conventional longitudinal recording, the upper limit of the write field that a recording

head could generate is equal to 2 MS, where MS is the saturation magnetization of the

head material. The highest value of 4 MS for the materials available today is rapidly

approaching what is believed to be a fundamental limit of ~26 kGauss. This defines the

upper limit of the KU values that could be employed in a longitudinal medium and,

consequently, the maximum areal density achievable with conventional longitudinal

recording. It has been predicted that with the materials available today, the highest areal

density achievable with conventional longitudinal recording is ~100 Gbit/in2 [10,8].


Figure 3. A schematic diagram describing the anisotropy energy of a magnetic grain.

3. Dodging the Superparamagnetic Limit

Several aspects native to perpendicular recording make it superior to longitudinal

recording with respect to the superparamagnetic limit. Among the advantages are higher

write-field amplitude and sharper write-field gradients, thicker recording layers, absence

of the demagnetizing field in the bit transitions, higher playback amplitude, etc. The

specific nature of these advantages is discussed in detail below.

3.1. STRONG RECORDING FIELD

Figure 4 shows comparative schematic diagrams of conventional longitudinal and

perpendicular recording modes. While in longitudinal recording, the preferred direction

of the magnetization (in other words, the “easy” axis) lies in the plane of a recording

medium, in perpendicular recording, the “easy” axis is perpendicular to the plane of a

medium. In longitudinal recording, writing is performed by the fringing field emanating

from the gap region between the write-poles of a conventional “ring”-type recording

head. It is the geometry of the ring head that defines the upper limit of the write field of

2 MS, where MS is the saturation magnetization of the write-pole material. In

perpendicular recording, the write field is generated between the trailing pole of a single

pole head and a soft underlayer (SUL). SUL, a new component in a recording system, is

a soft magnetic material located below the recording layer. In such geometry, the upper

limit of the write field is equal to 4 MS, which is two times higher than the highest field


achievable with the longitudinal ring head. Higher write efficiency of the perpendicular

single-pole recording head (in combination with the SUL) will ba explained in greater

detail and is illustrated in Figure 5.

(a) SUL

Transition

Written

moment

in media

Coil

“Gap” field

Record.

layer

Yoke Trailing edge

(b)

CoilYoke

Fringing

fields

Recording

medium

TransitionWritten moment

in media

Figure 4. Schematic diagrams showing a side cross-section of (a) a typical perpendicular system including a

SPH and a double-layer medium with a SUL and (b) a longitudinal system, including a ring-head and a single-

layer recording medium.


(a)

“Gap” fields

Real head

Image head

Coil

SUL

boundary

Physical Gap Effective Gap

(b)

Figure 5. A schematic diagram and 3-d drawing of the magnetic imaging principle in perpendicular recording

using a medium with a soft underlayer.

It could be shown (the proof of this concept is beyond the scope of this book [13]) that to

evaluate the magnetic field above the SUL boundary, the SUL could be thought of as a perfect magnetic mirror such that the magnetic field above the SUL boundary is equal to


the net field generated by both the magnetic elements above the SUL boundary and their

images located below the SUL boundary. This concept is illustrated in Figure 5, where the SUL is replaced with an image of the recording head. From this picture it is clear that

in perpendicular recording the write process effectively occurs in the “gap” between the

magnetic poles of the real head and its image. This is in contrast to longitudinal recording, where writing is performed by the field fringing from the gap, as outlined

above. From simple superposition arguments, it is straighforward to show that the “in-

gap” field is equal to 4 MS while the highest value of the fringing field is equal to 2 MS

[11].

As shown above, the maximum write field achievable in perpendicular recording is

almost twice as large as the maximum write field achievable in longitudinal recording. The direct consequence is the ability to write onto a higher anisotropy medium (higher

KU). The use of higher anisotropy media materials allows higher areal densities without

compromising thermal stability of the recording data.

As illustrated in Figure 6, the spatial profile of the write field in perpendicular recording

is also more beneficial for achieving higher areal density (compared to longitudinal recording). The side gradients, i.e. the rate at which the field rolls off at the side edges of

a recording head, are usually substantially sharper than what one observes in longitudinal

recording. This property leads to better-defined tracks with a relatively narrow erase band. Along with better magnetic alignment of the media (see below), extremely narrow

tracks are possible to achieve.

0.5 um

19397

9698

Hx (Oe)

Along the track

6734

3373

Hx (Oe)

(a) (b)

Figure 6. Longitudinal head field contours and perpendicular head field contours from (a) a longitudinal head

with a 150 nm gap and (b) a perpendicular pole head with a pole thickness of 700 nm. The trackwidth is 50 nm

in both cases.

The single pole perpendicular write heads used to acquire the experimental data presented

in this chapter were fabricated via focused ion-beam (FIB) modification of conventional longitudinal writers [16]. It should be emphasized that the main difference in the design

of conventional perpendicular and longitudinal writers is the length of the gap between the magnetic write-poles. In terms of the write process, while in longitudinal recording

the writing occurs near the gap region, in perpendicular recording, the writing occurs near

the trailing edge of the trailing pole [17]. Figure 7 shows a state-of-the-art perpendicular recording head manufactured by FIB trimming of a conventional longitudinal write head


by increasing the gap length and trimming the trailing pole and the reader to specified

dimensions. Both the trailing pole and the reader are designed for a 60 nm track width.

FIBed Reader

FIBed Writer

Figure 7. A FIB image single pole perpendicular write head made by focused ion-beam etching of a

conventional longitudinal ring head. The trailing pole width is 60 nm.

3.2. PERFECTLY ALIGNED MEDIA

In conventional longitudinal recording, the easy axes of individual grains are randomly

oriented in the plane of a medium. (It should be recalled that the easy axis is the

energetically favorable axis/direction along which the magnetization of a grain is aligned in the absence of external magnetic fields.) Thus, in longitudinal recording, a large

fraction of the grains forming a bit has their easy axes severely misaligned with the bit

magnetization direction. Writing well-defined bit transitions on such randomly oriented media imposes stringent requirements onto the spatial profile of a write-field. If one

neglects the imperfections of a bit transition due to the granular nature of a medium, the

quality of the bit transition is defined mainly by the write-field profile. This is drastically different from perpendicular recording, in which the easy axis of each magnetic grain is

relatively well aligned in the direction perpendicular to the plain of the medium. Thus, in a perpendicular recording, the magnetization direction of a recorded bit always coincides

with the orientation of the easy axes of individual grains that form the bit. Well-defined

easy axis orientation relaxes the stringent requirements for the trailing and side write-field gradients necessary to achieve sharp transitions, thus enabling the use of thicker

media [14].

The intrinsically better alignment of perpendicular media helps to record narrow tracks

with well-defined transitions even into a relatively thick recording layer. A MFM image

of two adjacent tracks with a 65 nm trackpitch written into a 50 nm thick CoCr recording layer using a 60 nm wide single pole head is shown in Figure 8 [11]. This is equivalent to

a track density of ~400ktpi. It should be stressed that the state-of-the-art in longitudinal


recording for the track density is ~100 ktpi. The possibility of using thicker recording

layers further assists with improving thermal stability.

Figure 8. A MFM image of two tracks with a 65 nm trackpitch.

With respect to using well-aligned media, it should be remembered that previously it was shown that, although a well-aligned perpendicular medium might have a relatively small

average angle between the magnetization and the recording field, the torque created is

still sufficiently large to relatively rapidly switch the magnetization [18, 19].

S SNN

S

SN

N

longitudinal

perpendicular

More stable magnet

configuration

Figure 9. A schematics of the influence of demagnetizing field in longitudinal and perpendicular media.

3.3. ABSENCE OF THE DEMAGNETIZATION FIELD IN BIT TRANSITIONS

One of the major destabilizing factors in longitudinal recording medium is the strong demagnetizing field in the bit transitions. The destabilizing influence of the

demagnetizing field in the bit transitions is easy to see if one observes that two adjacent

bits with opposing magnetization directions repel in a similar way as two bar magnets


with the poles of the same polarity, such as north-north or south-south, facing each other.

The magnets would try to flip so that poles of opposite polarities are next to each other. This is illustrated in Figure 9.

The calculated demagnetizing field for the cases of longitudinal and perpendicular media for a single bit-transition is shown in Figure 10. In the longitudinal recording mode, high

demagnetizing field in the bit-transitions destabilize individual grains leading to a finite

transition width. This is opposite to perpendicular recording, in which the demagnetizing field reaches its minima in the bit-transitions, thus promoting ultra-narrow transitions

and, consequently, high-density recording.

It could also be noted that, unlike in longitudinal recording, the demagnetization field in

perpendicular recording decreases as the thickness increases, thus promoting a thicker recording layer, which in turn is beneficial for the thermal stability. With this respect, it is

common to note that although perpendicular recording promotes high densities, the

stronger influence of the demagnetization field at lower densities is a disadvantage of perpendicular recording.

-0.04 -0.02 0.00 0.02 0.04

-2000

-1000

0

1000

2000

Hz (

Oe)

Distance down the track (um)

T = 10 nm

T = 20 nm

-0.04 -0.02 0.00 0.02 0.04

-2000

-1000

0

1000

2000

Hx (O

e)

Distance along the track (um)

T = 10 nm

T = 20 nm

(a) (b)

Figure 10. The demagnetization field versus the distance down the track along the central planes of 10 nm and

20 nm thick recording layers for (a) perpendicular and (b) longitudinal recording media.

4. A Soft Underlayer as a New System Component

One of the key aspects of perpendicular recording that makes it superior to the

longitudinal recording with respect to superparamagnetic effects is utilization of media with a SUL. A single-pole head and a medium with a SUL perpendicular recording

system enables write fields in excess of 80% of 4 MS of the pole head/SUL material. This doubles the fields available in longitudinal recording, thus opening the possibility to

write on substantially higher anisotropy media and leading to better thermal stability. Acting as a magnetic mirror, SUL effectively doubles the recording layer thickness,

facilitating substantially stronger readout signals. Also, the effective thickness increase

due to the mirroring effects by a SUL leads to the reduction of the demagnetizing fields with a potential to further improve thermal stability.


Domain wall

(source of “magnetic charges”)

Fields from Wall (Source of Noise)

M M

Figure 11. A schematic of the stray fields generated by a SUL

While the utilization of perpendicular media with a SUL should make it possible to

postpone the superparamagnetic limit, the SUL introduces a number of technical challenges. Some of the issues related to the presence of the SUL are discussed below.

4.1. SUL AS A MAJOR SOURCE OF NOISE

Among the technical challenges introduced by the presence of a SUL is the fact that a not

properly optimized SUL material can introduce a significant amount of noise into the playback signal. The noise results from the stray field generated by the effective charges

resulting from domain walls in the SUL as illustrated in Figure 11.

Soft underlayerHard layer

-----

-----

Magnets

head

Figure 12. A schematic of experimental setup to magnetically bias SUL film.


Magnetic biasing of the SUL, i.e. forcing the SUL into a single magnetic domain state, allows to minimize the SUL noise. The biasing can be achieved either by application of

an external magnetic field or by engineering a SUL material with a built-in biasing field.

Figure 12 shows a schematic of the experimental setup to study the effect of magnetic biasing of the SUL on the noise. The magnetic biasing was achieved using two NdFeB

permanent magnets placed in the vicinity of the media. The placement of the magnets

was such that it allowed achieving complete saturation of the SUL underneath the reader. Special care was necessary to arrange the magnets sufficiently far from the recording

head ~2cm away in order not to affect the properties of the read element.

Figure 13 shows the playback signals from the two media with as deposited non-biased

(a) and magnetically biased (b) SUL’s. A substantial level of noise attributed to presence

of a large number of domain walls (confirmed by magnetic force microscopy) in the SUL

can be seen in Figure 13a. A drastic reduction of the noise (by at least 10dB) is clearly

observed in Figure 13b where the SUL is magnetically biased.

(a) (b)

Figure 13. Playback signal from two media with different SUL’s. (a) SUL with a large number of stripe

domains. The presence of stripe domains was confirmed using magnetic force microscopy. (b) Biased SUL with

domain walls swept out from the SUL material.

The magnetic biasing saturates SUL film forcing it into a pseudo-single domain state effectively sweeping the domain walls out of the SUL material. This results in

elimination of the SUL noise.

4.2. SUL MAGNETIC MOMENT

To properly design a perpendicular recording system that utilizes a medium with a SUL, it is critical to choose an appropriate SUL material. As illustrated in Figure 14, if the


magnetic moment of a SUL material is lower than the magnetic moment of the recording

pole tip, saturation of the SUL underneath the pole tip can occur.

Pole

tip

Soft underlayer

Saturated

region

H

Pole

tip

Soft underlayer

H

SUL 4 MS < Head 4 MS(saturated region under the pole tip

deteriorates gradients)

SUL 4 MS > Head 4 MS

(not saturated under the pole tip)

Figure 14. A schematic illustrating the saturation effect in the SUL is the magnetic moment of a SUL is lower

than the magnetic moment of the write pole tip.

The results of boundary element modeling for two different head/SUL combinations are

presented in Figure 15. It can be noticed that it is possible to generate strong recording

fields with the magnitude approaching 4 MS of the pole tip even if the SUL has a lower magnetic moment than the pole tip. However, saturation of the SUL will lead to a

substantial deterioration of the trailing field gradients. The trailing gradients in the case of the Permalloy based SUL are substantially worse than the trailing gradients in the case

when a FeAlN based SUL is used.

It follows that if high moment materials are used for write heads, e.g. CoFeB, FeAlN,

etc., the moment of the SUL material should match or exceed the moment of the pole tip material.

4.3. SUL THICKNESS

Another important issue related to the optimized design of a SUL is the SUL thickness.

Using simple considerations of magnetic flux conservation, the minimum thickness required for the SUL to function properly is given by Expression 4.

tippole

layersoft under S

tip poleS

layersoft under2

1w

M

Mt

, (4)

where the wpole tip is the width of the write pole tip, i.e. the dimension of the write pole tip

defining the track width. The evaluation of the above equation for the case of 100


Gbit/in2 areal density and 4:1 bit aspect ratio, i.e. a 160 nm wide pole tip, and the same

pole tip and SUL materials, gives the lower boundary on the SUL thickness of 80 nm. It should be stressed that this thickness is substantially smaller than the “required

minimum” (as often quoted in the literature, of “hundreds of nanometers to a micron).

-0.5 -0.4 -0.30

5

10

15

Hz (

kO

e)

Distance down the track ( m)

Permalloy

Isat

=100mA

FeAlN

Isat

=75mA

Figure 15. Trailing fields from a single pole perpendicular write head made out of FeAlN (4 MS =20kG) for

FeAlN and Permalloy (4 MS =10kG) SUL’s.

This important observation needs to be strongly emphasized. Due to material properties, the above mentioned problem of the SUL noise becomes increasingly aggravated with the

increasing thickness of the SUL.

4.4. SUL INFLUENCE ON THE RESOLUTION

An additional challenge that the presence of a SUL imposes is the potential deterioration

of the system resolution. During reading from a medium with a SUL, due to the magnetic

imaging of the SUL, the resolution could get distorted if the separation between the ABS and the SUL (sum of the recording layer thickness and the flying height) were

comparable to the reader thickness.

This phenomenon is clearly illustrated in the calculated [20] PW50 and the playback

signal versus the SUL-to-ABS distance, shown in Figure 16. PW50 is the physical width

of a single transition, the measure of the spatial resolution of a recording system. In these calculations, a fixed recording layer thickness of 10 nm was assumed, and the separation

between the bottom side of the recording layer and the SUL was varied from zero to a

finite value. For comparison, the dotted straight lines indicate the values for the case when there is no SUL used. It could be clearly observed that the resolution of the

modeled recording system substantially deteriorates at certain values of the ABS-to-SUL

separation. This suggests that a special care (of this separation) has to be taken to properly optimize the system’s resolution.


Although, in a properly designed system this resolution distortion could be almost

completely eliminated, it causes the resolution of a typical read head in a system with an underlayer to be at most as good as the resolution of an equivalent head in a system

without a SUL. It should be noted, however, the SUL definitely increases the playback

signal, which is desirable at high areal densities.

10 15 20 25 30

40

45

50

PW50 (with SUL)

PW50 (without SUL)

No

rmaliz

ed

Sig

na

l (a

rb.u

nits)

ABS to underlayer distance (nm)

PW

50 (

nm

)

0.6

0.7

0.8

0.9

1.0 Signal (with SUL)

Signal (without SUL)

Figure 16. PW50 and the normalized playback vs. the ABS to underlayer spacing. 30 nm GMR element and a

70 nm shield-to-shield spacing are assumed.

5. Playback: New Signal Processing Schemes

+

+

charges in the transition

+

+

+ + + + + + + + - - - - - - - - - -

- - - - - - - - - - + + + + + + +

Hstray

Hstray

M(a)

(b)

Figure 17. Diagrams showing the sources of stray fields in the case of (a) longitudinal recording, and (b)

perpendicular recording.

One of the drastic differences between perpendicular and longitudinal recording is the

difference in the playback signal. To help understand the basic difference in the playback process between longitudinal and perpendicular recording, schematic diagrams of the

stray field emanating from a longitudinal medium and perpendicular media without and

with a SUL are shown in Figures 17a and b, respectively. As could be noticed, in the


longitudinal case, the stray field emanates only from the transitions, with the fields near

the transitions oriented perpendicular to the disk plane. On the contrary, in the perpendicular cases, the stray field emanates from the effective magnetic “charge” at the

top and effective (due to the SUL) bottom surfaces of the recording layer, with the field

near the transitions oriented parallel to the disk plane.

As a result of the different magnetic “charge” distributions, the playback waveforms

differ drastically between longitudinal and perpendicular recording schemes. This

difference is illustrated in Figure 18, where typical low-density playback waveforms are shown for both perpendicular and longitudinal recording.

The above shown waveforms for perpendicular and longitudinal recording modes outline

a major difference between perpendicular and longitudinal recording. While in

longitudinal recording the signal is present only near the bit transitions, in perpendicular recording the signal is read not only near the bit transition but rather across the entire bit

area. It is possible to differentiate the perpendicular playback signal to make it similar to

the playback signal in longitudinal recording. However, it should be remembered that the differentiated perpendicular playback waveform is similar but not identical to the

longitudinal playback waveform. The difference arises in the absence of a transition

when there is no longitudinal playback signal while the differentiated perpendicular playback, although is relatively small in amplitude, is still finite (non-zero).

Longitudinal

PlaybackPerpendicular

Playback

Pla

yback S

ignal

Time

Pla

yba

ck S

igna

l

Time

Figure 18. Typical playback waveforms for perpendicular and longitudinal recording schemes.

It should be stressed that while not entirely suited to be processed by conventional longitudinal channels, perpendicular playback waveforms clearly contain more

information than typical longitudinal waveforms in which the signal is concentrated

mostly near the transitions. This property could be used to advantage in future channel designs.

6. Challenges of New Materials


While the requirements for the head materials used in perpendicular recording are similar

to those for the head materials used in longitudinal recording, the major differences exist with respect to the media materials. A typical perpendicular medium consists of two

magnetically active layers, a hard layer and a SUL, and a number of non-magnetic layers,

as shown in Figure 19. The hard layer has rather different magnetic properties compared to the hard layer utilized in conventional longitudinal recording. It should also be noted

that SUL has no analogy in longitudinal recording. The requirements for these two layers

are outlined below.

Figure 19. A schematics of a typical perpendicular medium.

6.1. HARD LAYER MATERIALS

The primary approach to the design of a perpendicular recording layer is in many ways similar to the design of a conventional longitudinal recording layer. All the media in use

today has granular structure, i.e. made of polycrystalline materials. Major goals inherent

to both longitudinal and perpendicular recording layer development are small grain size, small grain size distribution, texture control, optimization of the inter-granular exchange

de-coupling, etc. The large variety of today's perpendicular magnetic recording layer

types can be clearly divided into the two major categories: 1) Alloy based media, such as CoCr-alloys [21, 22], and 2) media based on magnetic multilayers, such as Co/Pt, Co/Pd

or other materials [23, 24]. Figure 20 contrasts the major difference between alloy and

multilayer media. In alloy media, the magnetic anisotropy is controlled by magnetic crystalline anisotropy. The alloy media are usually highly textured to insure well-defined

magnetic easy axis [25]. In magnetic multilayers, the magnetic anisotropy is controlled


through interfacial effects between a magnetic layer, such as Co, and a highly polarizable

spacer layer, such as Palladium or Platinum. In contrast to the alloy media, this set of materials as used in perpendicular media usually possesses a very weak texture.

Single crystal grains, arrows

represent the easy axes orientations

Co

Pd

Alloy Multilayer

Bi-layer

(a) (b)

Figure 20. A schematic representation of major microstructural differences between (a) an alloy medium and

(b) a multilayer medium.

Material-wise, perpendicular CoCr-based alloy recording layers are similar to

conventional longitudinal CoCr-based media, with the major difference being the

orientation of the magnetic easy axis. Therefore, a significant amount of information accumulated in the course of the longitudinal media development can be used to control

the critical parameters such as the grain size and the inter-granular exchange coupling. At

the same time, CoCr-based perpendicular media have some open issues. For example, it is not clear yet if it is possible to make a CoCr-based medium with sufficiently high

anisotropy to avoid superparamagnetic instabilities at ultra-high areal densities. It also has proven to be difficult to make CoCr-alloy based perpendicular recording layers with a

remanent squareness of 1. The remanent squareness is defined as a ratio between the

remanent magnetization, the value of magnetization on a M-H loop at H=0, and the saturation magnetization, the maximum value of magnetization. It is believed that a

remanent squareness of 1 is necessary for low-density bit pattern stability. Also, a

remanent squareness of less than 1 can lead to substantial amounts of DC noise. Various magnetic alloys such as L10 phases of FePt, CoPt, etc. are being studied as higher

anisotropy alternatives for the recording layer.

The magnetic multilayer based recording layers typically have significantly larger

anisotropy energies (Coercive fields of above 15 kOe have been reported.) and are thus

promising to be extendable to significantly higher recording densities. Another advantage of the magnetic multilayers is the fact that typically these materials have a remanent

squareness of 1.

To compare basic magnetic properties of CoCr-alloy and mutlilayer based recording

layers, typical M-H loops by a Kerr magnetometer for a 50 nm thick perpendicular CoCr

thin-film and a 52 nm thick Co/Pd structure (a stack of 40 sets of adjacent 3 and 10


Angstrom thick layers of Co and Pd, respectively) are shown in Figures 21a and b,

respectively. It can be noticed that in addition to the remanent squareness of 1, the Co/Pd structure exhibits nucleation fields in excess of 3kOe, a useful characteristic to avoid data

self-erasure due to stray fields. Meanwhile, the CoCr material shown in Figure 21a has a

squareness of 0.75. The CoCr and Co/Pd recording layers have coercive fields and magnetizations of approximately 3 kOe and 9 kOe and 300 emu/cc and 200 emu/cc,

respectively.

-6 -4 -2 0 2 4 6

-4

-2

0

2

4

Ke

rr S

ign

al (a

.u.)

Field (kOe)-10 -5 0 5 10

-5

0

5

Kerr

Sig

nal (a

.u.)

Field (kOe)

(a) (b)

Figure 21. An M-H loop of a 50nm thick (a) CoCr-alloy layer and (b) Co/Pd multilayer.

The direct consequence of remanent squareness less than 1 is shown in Figure 22, which

compares the spectral SNR distributions for the two media types. The CoCr medium

exhibits a significant amount of noise at lower linear densities. This is mainly due to the

fact that the dominant contribution to the noise at low linear density in the CoCr-based

medium comes from the DC noise which results from the relatively low value of

remanent squareness, as described below in more detail.

Figure 22. SNR versus the linear density for a CoCr-alloy (hollow circles) and a Co/Pd multilayer (hollow

squares).


6.2. HIGH ANISOTROPY SUL MATERIALS

Several design guidelines for SUL’s were discussed above including thickness

requirement and magnetic moment requirement. An additional parameter, which is critical to achieve optimized performance of a SUL in a perpendicular recording system,

is magnetic anisotropy of the SUL material. The dynamic properties [26, 27] and

influence of a SUL on system’s resolution [28] are affected by the value of the anisotropy field. The latter is illustrated in Figure 23, where the playback versus the linear density

(roll-off) curves are shown for identical perpendicular recording systems with different

SUL materials. The explanation of the quantum-mechanical nature of this effect is beyond the scope of this book. However, it should be mentioned that the deterioration of

the system’s resolution arises from the inability of lower anisotropy SUL materials to perfectly respond to rapid spatial variations of magnetization in the recording layer.

0 200 400 600 800 1000

-60

-50

-40

-30

-20

-10

0 FeAlN (Hk ~ 15 Oe)

Ni45

Fe55

(Hk ~ 50 Oe)

Permalloy (Hk ~ 5 Oe)

Pla

yb

ack (

dB

m)

Linear Density (kfci)

Figure 23. Playback roll-off curves for perpendicular recording media with identical recording layer but

different SUL’s.

The extent of the roll-off curves to higher linear densities for higher anisotropy SUL

indicates the advantage of using high anisotropy SUL materials.

7. How Far Will Perpendicular Recording Go?

It should be emphasized that perpendicular recording does not eliminate but rather defers

the superparamagnetic limit of longitudinal recording to higher areal densities. A number

of factors, including the availability of higher recording fields, the possibility of using

thicker and well-aligned media, and the absence of strong demagnetizing fields in the bit

transitions, contribute into deferring the superparamagnetic limit to substantially higher

areal densities. It has been shown that with all the factors taken into account, the

maximum areal density achievable with perpendicular recording configuration in the

development today is 500-1000 Gbit/in2 [10,29,30]. Once perpendicular magnetic


recording reaches its superparamagnetic limit, a new wave of technological innovations

will have to take place.

As mentioned in the beginning of this chapter, the foremost fundamental reason for the

existence of the superparamagnetic limit is the head materials constraint imposing the

limitation on the available head field that limits the utilization of higher anisotropy

media. Among the potential successors of perpendicular recording is heat-assisted

magnetic recording (HAMR) [31]. In HAMR, the anisotropy of a recording medium is

substantially reduced via local heating of the medium during the writing instance. To

accomplish the heating mission, a source of heat (envisioned as an ultra-small light

source) should be added in a recording system to locally increase the temperature of the

recording medium. The local increase of the medium temperature leads to the local

decrease of the medium coercivity enabling recording with relatively small magnetic

fields.

Additionally, patterned media can be utilized to further extend the limits of magnetic

recording [31]. In a patterned medium, the location and the size of the magnetic features

are pre-determined by the medium manufacturing process. Elimination of the element of

randomness characteristic to today’s polycrystalline recording media is a clear advantage

of the patterned medium approach. However, for such a medium to become a serious

contender to replace conventional alloy or multilayer media, an economically viable

manufacturing process will have to be developed [32,33].

It should be emphasized that due to the advantageous nature of perpendicular recording

in promoting extremely high areal bit densities (high write field amplitude, well aligned medium, sharp field gradients, absence of demagnetizing field at transitions, etc.), the

future technologies such as mentioned above HAMR and recording on a patterned

medium, are likely to be developed as extensions of perpendicular magnetic recording schemes [31] rather than to be based on conventional longitudinal recording.

Chapter 2 Physics of Writing

23

Chapter 2

Physics of Writing

1. Introduction

After fierce struggles to extend the life of longitudinal magnetic recording as the main

technology for another couple of years, the data storage industry is finally coming to

terms with reality. Reality tells that the areal density in cutting-edge laboratory demonstration systems is limited by thermal instabilities in the longitudinal magnetic

media [34]. Recent high areal density demonstrations of perpendicular recording clearly

demonstrate the strong interest of the data storage industry in this alternative technology today [35,36,37,38]. Compared to the conventional longitudinal recording mode, it is

believed that perpendicular recording is capable of deferring the superparamagnetic limit

to a substantially higher areal density due to the thicker recording layer and/or the use of a soft underlayer (SUL) [39]. Although perpendicular recording is certainly the closest

alternative to the conventional technology, its novelty also brings up new issues, not ever

encountered in longitudinal recording. These issues have to be well understood before the technology can be fully and most efficiently implemented. Major questions related to

perpendicular media and perpendicular playback and writing heads have been previously considered. However, relatively little attention has been given to the writing process at

areal densities beyond 100 Gbit/in2. For example, the role of soft magnetic shields in the

writing process is still an unresolved question: although the use of soft shields around the main pole of the writing head certainly increases the field gradient, its influence on the

magnitude of the recording field is still controversial. Another fundamental question is

the role of the soft underlayer in the writing process. These and many other questions associated with the writing process need to be considered altogether for the most efficient

design of the write head. Therefore, the intention of this Chapter is to analyze the writing

process in perpendicular recording from the global perspective of maximizing the achievable areal density.


1.1. CHAPTER OVERVIEW

In this Chapter, a detailed overview of the methodology to design a write transducer for

recording onto perpendicular media at areal densities beyond 1Tbit/in2 is presented. The

two basic modes of perpendicular recording, single-layer recording media in combination with a ring type head and double-layer recording media with a soft underlayer in

combination with a single pole head, are compared with each other theoretically and

experimentally. In addition, perpendicular recording is compared to longitudinal recording from the perspective of the writing process. The system efficiency is redefined

for perpendicular recording to take into account the critical role of the soft underlayer.

The effects of using “soft” magnetic shields around the trailing pole are analyzed. It is shown that at least a factor of two increase in the field can be obtained at areal densities

beyond 500 Gbit/in if shields are used. Such an open issue as the skew angle sensitivity in perpendicular recording is analyzed. It is shown that using “soft” magnetic shields

around the trailing pole substantially improves the skew angle sensitivity. Moreover,

using shields substantially improves the system efficiency and to some degree fulfils the role of the soft underlayer in perpendicular recording.

2. Different Modes of Perpendicular Recording

There are two basic modes of perpendicular recording [40]. The 1st mode utilizes a single pole head (SPH) for recording onto a double-layer perpendicular medium consisting of a

recording layer and a SUL, as shown in a diagram in Figure 1a [41]. As described below,

the use of the SUL is one of the most critical factors contributing to one of the best-known advantages of perpendicular recording, which is the ability to generate a recording

field of the order of 4 Ms, where Ms is the saturation magnetization for the recording head material [42-43]. For comparison, in conventional recording, the maximum

longitudinal recording field generated by a ring head (RH) is approximately 2 Ms [44]. The ability to generate a stronger field makes it feasible to record on a medium with

higher coercivity, which in turn further defers the superparamagnetic limit to a higher areal density [45]. The 2nd mode utilizes a regular RH for recording onto a single-layer

perpendicular medium, as shown in a diagram in Figure 1b. Although, the 1st mode is more widely exploited due to the advantages of the SUL, it is still reasonable to start with

the description of the 2nd mode, because the latter is fairly similar to the conventional

longitudinal mode and, therefore, is going to be a good transitional step towards development of a structured theory of perpendicular recording. Both the longitudinal

recording mode and the 2nd perpendicular recording mode rely on the utilization of a ring

head along with a medium without a soft underlayer. Through the comparison of these two recording modes, some of the critical features of perpendicular recording can be

made fairly apparent.

Besides the two basic modes, in some cases, some kind of an intermediate mode, e.g., a

RH and a medium with a SUL, or a SPH or a RH and a medium with a tilted

magnetization with or without a SUL can also be preferred, as discussed below. Moreover, it is shown that substantial modifications to basic head structures are required

Chapter 2 Physics of Writing 25

for the ability to record at densities beyond 1Tbit/in2. In the following Chapters,

advantages and issues associated with different recording modes are discussed in detail.

(a) SUL

Transition

Written

moment

in media

Coil

“Gap” field

Record.

layer

Yoke Trailing edge

(b)

CoilYoke

Fringing

fields

Recording

medium


in media

Trailing

pole

Figure 1. Diagram showing a side cross-section of a perpendicular system of (a) the 1st mode, including a SPH

and a double-layer medium with a SUL, and (b) the 2nd perpendicular mode, including a RH and a single-layer

recording medium.

2.1. SECOND PERPENDICULAR MODE: A RING HEAD AND A

PERPENDICULAR MEDIUM WITHOUT A SOFT UNDERLAYER

As mentioned above, the second mode of perpendicular recording, which uses a conventional longitudinal ring head and a medium without a SUL, still remains an arena

of exploration because of its resemblance to the conventional longitudinal mode and the

lack of the "not-yet-understood peculiarities” of the SUL in the first mode [42]. A


diagram showing a conventional longitudinal system is shown in Figure 2a. The only

structural difference between the second perpendicular mode and the conventional longitudinal mode is in the medium magnetization orientation: the magnetization is in the

plane and perpendicular to the disk plane for the longitudinal and perpendicular modes,

respectively. Also, in the perpendicular mode, the medium's "easy" axis is ideally aligned in one direction (in the direction perpendicular to the disk plane), while in the

longitudinal mode the "easy axes" are randomly oriented in the disk plane [46].

(a)

CoilYoke

Fringing

fields

Recording

medium


in media

(b)

Drive coil

W2

GT1 T2

P2P1TH

Figure 2. (a) Diagram showing a side cross-section of a typical longitudinal system, including a RH and a

recording medium. (b) 3D schematic diagram of a RH.


Because RH is a critical part of longitudinal recording, a more detailed diagram of the

conventional RH is shown in Figure 2b. Although, in most practical cases, the leading pole, P1, is typically substantially wider than the trailing pole, P2, in this Section, the

assumption that both poles, P1 and P2, have the same thickness, T1=T2, is used for

simplicity of explanation of the key issues. Because the actual recording takes place near the trailing edge of the gap length, the effective trackwidth is dominantly determined by

the width of the trailing pole, W2, and does not strongly depend on the width of the

leading pole [47]. Moreover, in the past, some recording head manufacturing companies, for example, ReadRite Corporation, indeed, utilized a ring head with identical leading

and trailing poles of the type shown in Figure 2b [48]. Using a specially developed

magnetic force microscopy (MFM) technique to separately measure individual components generated by such a RH, the perpendicular and longitudinal field profiles at

the ABS of such a RH were directly measured, as shown in Figure 3a [49]. The cross-sections of these field profiles along the central line in the track direction are shown in

Figure 3b.

(a)P1

P2

GAP

(b)

-10 -5 0 5 10-1.0

-0.5

0.0

0.5

1.0

Axis along the track, X ( m)

Hx a

nd H

z (

au)

Cr/CoCrPt

Ti/CoCrPt

Figure 3. (a) MFM images of the perpendicular and longitudinal field profiles taken at the ABS of a RH with a

200 nm gap length. (b) The cross-sections of these field profiles taken along the central line in the track

direction.


In general, the RH structure has been widely studied in its association to longitudinal

recording, and there is plenty of literature, which contains more detailed information about the RH design. In this work, the authors only discuss the aspects of the RH design,

which are of interest for perpendicular recording.

Before going into details of the head design analysis, it is worth reminding that

traditionally, the Karlqvist’s two-dimensional (2D) model has been utilized for describing

the magnetic properties [50]. However, today, as the areal density reaches the point, at which the trackwidth becomes fairly small, 2D calculations cannot give sufficient

accuracy. Therefore in this Chapter, results of 3D calculations made with boundary

element model (BEM)–based commercial field solver Amperes are shown [51].

2.1.1. Gap Length Dependence The 3D calculated along-the-track (X-) and perpendicular (Z-) field components for a RH

without a SUL at saturation for a set of 4 values of the gap length, 30, 70, 150, and 500

nm, are shown in Figures 4a-d, respectively. In these calculations, value for the flying height was 5 nm, and the trackwidth and the pole thickness were modeled to be 200 and

500 nm, respectively. Nevertheless, the efficiency depends on the gap length exactly as in

longitudinal recording [44,52]. The dependence of the system efficiency on the gap length is reflected in the saturation current dependence on the gap length, as shown in

Figure 5. The normalization factor, NF, necessary for determining the exact drive current

value depends on specific head parameters, including its dimensions and the location of the drive coil with respect to the ABS [52]. The saturation current is determined as the

current at which the recording field under the gap center at a 5 nm flying height starts to

saturate. Going back to the description of Figures 4a-d, with a gap length of 70 nm, the chosen parameters approximately correspond to areal density of 50 Gbit/in2.

-1000 -500 0 500 1000-1.0

-0.5

0.0

0.5

1.0

W2 = 200 nm

PT = 500 nm

G = 30 nm

Gap

region

Hz

Hx

Hx a

nd

Hz (

1/2

Ms)

Down the track (nm)-200 -100 0 100 200

-1.0

-0.5

0.0

0.5

1.0

W2 = 200 nm

PT = 500 nm

G = 70 nm

Gap

region

HzH

x

Hx a

nd H

z (

1/2

Ms)

Down the track (1/fly height)

(a) (b)


-1000 -500 0 500 1000-1.0

-0.5

0.0

0.5

1.0

W2 = 200 nm

PT = 500 nm

G = 150 nm

Gap

region

Hz

Hx

Hx a

nd

Hz (

1/2

Ms)

Down the track (nm)-1000 -500 0 500 1000

-1.0

-0.5

0.0

0.5

1.0

W2 = 200 nm

PT = 500 nm

G = 500 nm

Gap

region

Hz

Hx

Hx a

nd

Hz (

1/2

Ms)

Down the track (nm)

(c) (d)

Figure 4. Longitudinal and perpendicular field components versus the distance down the track for a RH with a

trackwidth of 200 nm, a pole thickness of 500 nm at four values of the gap length, (a) 30 nm, (b) 70 nm, (c) 150

nm, and (d) 500 nm.

Although, in practice, both field components, in-plane and perpendicular, simultaneously influence each recording event, ideally, the perpendicular and longitudinal field

components reflect the perpendicular and longitudinal recording modes, respectively.

From the plots, it can be seen that the longitudinal field component is fairly well localized in the gap region. In this case, the field near the trailing edge of the gap produces

recording. As a result, by having the gap length sufficiently small, a fairly sharp field profile and fairly large areal densities can be produced. However, there is a limit to

reducing the gap length. As the efficiency increases with the gap reduction, the less flux

leaks out through the gap region, thus resulting in the weaker recording field. Eventually, the recording field becomes too small for overcoming the medium coercivity field. For

example, in this particular case, this trade-off value of the gap, below which the

longitudinal field component starts to drop, is in the vicinity of the 70 nm value, as seen from Figures 4a-d. The tradeoff value is mostly determined by the flying height and the

trackwidth. The scenario is different for the perpendicular field component, for which a

fairly large value can be noted far beyond the gap region. As a result, in this case, recording is produced not by the field in the immediate vicinity of the gap region, but

rather by the field near the trailing edge of the trailing pole, as long as the field near the

trailing edge is larger than the coercivity field [53]. Also, it can be noted that the perpendicular field component at saturation even increases as the gap length is increased

in the considered range. This is caused by the reduction of the longitudinal field

contribution into the net flux as the gap increases and thus making the net field dominantly perpendicular. It can be noted that unlike in longitudinal recording, the

maximum field and the trailing field gradient are defined not only by the physical gap

length but also to substantial degree by the trailing pole tip geometry.

However, for the both systems, the efficiency fairly strongly depends on the gap length because of the use of a RH. For any recording mode, for which a RH is utilized with a

medium without a SUL, transitions are produced by the fields that fringe out from the gap

of the "closed" magnetic loop of the RH [54]. In other words, the gap region becomes a


part of the magnetic flux loop, and therefore the efficiency of the loop strongly depends

on the gap region. The dependence of the efficiency on the gap length is proportional to the dependence of the saturation current on the gap length. Assuming that the saturation

current is defined as the current at which the longitudinal field at the center of the gap

reaches 2 Ms, the calculated saturated current versus the gap length is shown in Figure 5.

0 200 400

50

100

150

200I m

ax (

au)

Gap (nm)

Figure 5. The maximum field current versus the gap length.

2.1.2. Trailing Pole Thickness Dependence

The calculated field components for a gap length of 70 nm and a trackwidth of 200 nm

are shown for a set of three values of the pole thickness, 100, 200, and 500 nm, in Figures 6a-c, respectively. It can be noted that although the longitudinal field component does

not vary substantially as the pole thickness is increased from 100 to 500 nm, the

perpendicular component increases more than by a factor of three. Moreover, while the perpendicular component is noticeably smaller than the longitudinal component at the

smallest value of the pole thickness, i.e. 100 nm, it becomes comparable to the

longitudinal component as the pole thickness in increased to 500 nm.

-1000 -500 0 500 1000-1.0

-0.5

0.0

0.5

1.0

G = 70 nm

W = 100 nm

PT = 100 nm

Gap

region

HzH

x

Hx a

nd H

z (

1/2

Ms)

Down the track (nm)-1000 -500 0 500 1000

-1.0

-0.5

0.0

0.5

1.0

Trailing

edge

G = 70 nm

W = 100 nm

PT = 200 nm

Gap

region

HzH

x

Hx a

nd

Hz (

1/2

Ms)

Down the track (nm)

(a) (b)


-1000 -500 0 500 1000-1.0

-0.5

0.0

0.5

1.0 Trailing

edge

G = 70 nm

W = 100 nm

PT=500 nm

Gap

region

HzH

x

Hx a

nd H

z (

1/2

Ms)

Down the track (nm)

(c)

Figure 6. Longitudinal and perpendicular field components versus the distance down the track for a RH with a

trackwidth of 100 nm, a gap length of 70 nm at three values of the pole thickness, (a) 100 nm, (b) 200 nm, and

(c) 100 nm.

In general, it could be noted that with respect to the recording field, the 2nd perpendicular

mode is quantitatively similar to the longitudinal mode. In both cases, the maximum field

never exceeds 2 Ms of the head material. Previously, the implementation of the RH writer in combination with perpendicular media with a SUL has also been published and

furthermore, the related material is going to be presented in this Chapter [55]. In SectionThe First Perpendicular Mode: A SPH and a Perpendicular Medium With a SUL, it is

shown that the perpendicular recording field can be increased by at least a factor of two,

i.e. can reach 4 Ms, if a medium with a SUL is utilized.

2.2. FIRST PERPENDICULAR MODE: A SINGLE POLE HEAD AND A

PERPENDICULAR MEDIUM WITH A SOFT UNDERLAYER

As shown in Figure 1a, besides the presence of a SUL, the first mode is different from the

second mode also in the type of the recording head: it is a SPH instead of a RH. Unlike

the RH, the SPH, utilized in combination with the SUL, has a physical gap that is substantially larger than the flying height. The purpose of the large gap is to force the

magnetic flux to flow through the SUL rather than through the gap region, thus to

enhance the perpendicular component of the magnetic field. Therefore, the SUL is an indispensable part of the recording head, as well as it is of the recording medium.

2.2.1. Magnetic Image Model

It is convenient to use the so-called "magnetic image" model for more transparent

description of recording processes in a system with a SUL [53]. According to this model, the SUL is replaced with a half-space, which contains a mirror image of the recording

head, as shown in a schematic diagram in Figure 7. Such replacement is theoretically

justified provided the SUL can be approximated to be ideal [56]. According to a theorem of differential equations, the Laplace’s Equation (a consequence of the Maxwell’s

Equations, convenient to use for the calculation of the magnetic field) has an


unambiguous solution if adequate boundary conditions are satisfied [57]. It appears that

in the above case with an ideal SUL the boundary conditions at the SUL top surface are the same as in the case with a mirror half-space provided that the magnetic "charge"

reverses its polarity when reflected into the mirror half-space. Together with the image

head, there are effectively two heads involved in each recording event, thus the net recording field becomes fairly large, as compared to the equivalent longitudinal case, as

discussed below.

2.2.2. Permanent Magnet Approximation

The fastest way to estimate a magnetic field generated by a SPH at saturation is probably

to use the permanent magnet approximation. In this approximation, the SPH is presented as an infinitely long vertical magnetic bar with finite cross-section dimensions, W

(trackwidth) and T (thickness), with its magnetization aligned (saturated) along the vertical axis. In such scenario, the magnetic field components can be directly calculated

using, for example, the equivalent “charge” model [58]. Thus derived formulas for a

saturated SPH without the presence of a soft underlayer are shown by Expressions 1a to c. Because of the problem symmetry, it is sufficient to calculate the field in one

coordinate quadrant, x > 0 and y > 0.

(a)

“Gap” fields

Real head

Image head

Coil

SUL

boundary

Physical Gap Effective Gap


(b)

Figure 7. (a) A cross-section diagram and (b) a 3-d drawing showing a mirror image representation of a

perpendicular system with an ideal soft underlayer.

22

2

2

22

22

2

2

22ln

22

2

2

22

22

2

2

22ln

2z

by

ax

by

zb

ya

xb

y

zb

ya

xb

y

zb

ya

xb

ysM

xH

(1a)

22

2

2

22

22

2

2

22ln

22

2

2

22

22

2

2

22ln

2z

by

ax

ax

zb

ya

xa

x

zbya

xa

x

zb

ya

xa

xsM

yH

(1b)


222

222

22

221tan

22

221tan4

zb

ya

xz

byax

zb

ya

xz

byaxs

M

zH

2

22

2

22

22

221tan

22

221tan4

zb

ya

xz

byax

zb

ya

xz

byaxs

M

. (1c)

The origin of the coordinate system is at the center of the pole tip air-bearing surface (ABS) with the vertical axis, Z, directed downward, as shown in Figure 8. Moreover, the

presence of the SUL can be simply taken into account through using the described above

"magnetic image model." In other words, the same expression can be utilized to calculate the extra recording field due to the image head located at the other side of the recording

layer. The spacing difference between the real and image heads is equal to the recording

layer thickness plus the separation between the recording layer and the SUL. The sum of the two fields gives the total recording field.

X

Y

W

Recording layer

Z

Main Pole Tip

Figure 8. A diagram showing the location of the origin of the coordinate system utilized in the calculations.


2.2.3. Recording By The Field In The Gap (Perpendicular) Versus Recording By The

Field Fringing From The Gap (Longitudinal)

When using the “magnetic mirror" image model, besides the physical gap length, also, the effective (magnetic) gap length can be introduced [59]. The effective (magnetic) gap,

defined as the spacing between the air bearing surfaces (ABS’s) of the real and image

heads, i.e. the two-fold separation between the ABS and the SUL, can be meaningfully compared to the physical gap of the RH [53]. It can be noticed that the SPH considered

together with its image resembles the RH rotated 90 degrees around the axis along the

cross-track direction, with the difference that the recording is produced in the “gap” itself [60]. In contrast, in the longitudinal case as well as in the case of the 2nd perpendicular

mode, the field fringing from the gap region produces recording, as shown in Figure 9. Any system exploiting a RH without a SUL is intrinsically built so that for the system to

be efficient, the gap length should be fairly small. It should be reminded, that the more

efficient a system is, the smaller amount of the magnetic flux leaks out on its way from the source (drive coil) to its destination (ABS). Consequently, substantial amount of the

magnetic flux just circulates in the magnetic ring yoke without being exploited for the

recording purpose itself, and, as noticed above, only the fringing field produces the actual recording. Typically, the maximum fringing field, which can be achieved in a recording

system of this type, is less than 2 Ms, where Ms is the saturation magnetization of the head material [54]. This limits the coercivity of a longitudinal medium, on which the

recording head can record [61]. On the contrary, it is due to the recording by the field in the “gap” region why the use of the SUL in the 1st perpendicular mode provides such a

drastic increase in the recording field at saturation. The calculated perpendicular and

longitudinal field components for a SPH with a gap length, G, of 1000 nm, a pole thickness, PT, of 500 nm, and a trackwidth, W, of 100 nm at saturation are shown in

Figure 10. It can be noticed that in this case the maximum perpendicular field is of the

order of 4 Ms. This allows writing on a medium with a higher anisotropy field. The anisotropy field defines the field, which needs to be applied for switching the

magnetization in the recording layer. In turn, the higher anisotropy medium means the higher density, to which the superparamagnetic limit can be deferred.

(a)Image SPH

Real SPH

Medium

Fields

in the Gap

(b)

Fields fringing from

the Gap

RH

Figure 9. Schematic diagrams showing (a) recording by the field in the "gap" in perpendicular recording and

(b) recording by the fringing field in longitudinal recording.


At this point, the SUL is assumed to be ideal. Also, the default modeling settings included a physical gap length of 1000 nm, a trackwidth of 100 nm, and a throat height of

500 nm, with a 20-nm separation between the ABS and the SUL. It can be noticed that

for the 1st perpendicular mode, the field profiles are qualitatively similar to the field profiles for the longitudinal mode, as shown in Figure 4, provided that the field

components are exchanged with each other according to the transformation Hx Hy and

Hy -Hx [60]. However, as previously mentioned, from the quantitative perspective, due to the use of the SUL the maximum perpendicular field in perpendicular recording is

approximately by a factor of two larger than the maximum longitudinal field in


0 750 1500 2250-1.0

-0.5

0.0

0.5

1.0

1.5

2.0Trailing

edge

W = 100 nm

PT = 500 nm

G = 1000 nm

Pole

region

Hz

Hx

Hx a

nd H

z (

1/2

Ms)

Down the track (nm)

Figure 10. Longitudinal and perpendicular field components versus the distance down the track (along the

central line) for a SPH with a gap length, G, of 1000 nm, a pole thickness, PT, of 500 nm, and a trackwidth, W,

of 100 nm.

2.2.4. Is the Increase of the Recording Field due to the Use of a SUL Sufficient for Adequate Recording?

Indeed, the use of a soft underlayer provides a two-fold increase of the recording field

component, as compared to the conventional longitudinal recording mode [42]. However, this comparison of the two recording modes is not equivalent. Above, it was shown that

in the longitudinal mode, in the gap region, besides the longitudinal field component,

there is also a substantial perpendicular component. For example, for a typical gap length of approximately 150 nm, as shown for the case in Figure 4c, both longitudinal and

perpendicular components reach approximately the same value, i.e. 2 Ms of the head material. On the contrary, in the perpendicular mode, the maximum longitudinal field

component is substantially smaller than the perpendicular field component. For example,


for the case shown in Figure 10, the perpendicular component almost reaches 4 Ms,

while the longitudinal component is substantially less than 1 Ms, i.e. the different is almost by a factor of four. As a result, the actual recording field is directed at angle

values of approximately 45 and 15 degree with respect to the medium magnetization for

the longitudinal and perpendicular modes, respectively. From the idealistic Stoner-Wohlfarth model, the switching field differs from the anisotropy field depending on the

angle between the recording field and the "easy" axis [62]. Moreover, the switching is

expected to be substantially easier for the 45-degree case, as compared to the 15-degree case. Another difference between these two recording modes results from the different

nature of the recording medium. For the perpendicular case, the magnetization is aligned

in one direction, i.e. the direction perpendicular to the disk plane, while, for the longitudinal case, the magnetization is directed randomly in the disk plane. Therefore,

although the realistic recording media might be substantially different from the idealistic Stoner-Wohlfarth model, for the fair comparison of the two recording modes, all the

described factors should be taken into account in more precise calculations. It is not done

in this Chapter, because the purpose is to describe the main concepts that help distinguish perpendicular recording. However, it could be noted that the second perpendicular mode,

i.e. the mode without a soft underlayer, is based on the use of the ring type head, similar

to the longitudinal mode, with all the advantages resulting from the larger torque angle between the recording field and the magnetization. This similarity to the longitudinal

mode makes the implementation of the second mode more straightforward, as compared

to the implementation of the first mode. Therefore, the second mode should not be totally ignored.

2.2.5. Quadruple Ratio between Saturation Currents in Perpendicular and Longitudinal Recording

Another advantage of perpendicular recording that can be noted from the mirror image

model is the fact that due to the SUL the effective number of the current sources is effectively doubled (see Figure 8). As a result, the perpendicular system needs

approximately only half as much current to generate the same magnetic field in the

effective gap, as compared to an equivalent longitudinal system. Below (in the following Section), it is shown that in the perpendicular case recording is produced in the effective

gap region. This is unlike the longitudinal mode, for which recording is produced by the field fringing from the gap. Because the field fringing from the gap is about only a half of

the field in the gap and the effective number of the drive current sources in the

perpendicular system is twice as many as in the longitudinal system, for the perpendicular system it takes approximately four times less drive current to generate the same recording

field as in the longitudinal system, with the other conditions equivalent. Although such a

fairly rough estimate does not take into account any non-linear effects that can take place in a recording system, it provides a good sense for the saturation currents in the two

systems. As an example, the maximum recording fields generated by RH and SPH versus

the drive current are shown in Figure 11.


-10 0 10 20 30 40 50 60 70

0.0

0.5

1.0

1.5

2.0

RH: Hx

SPH: Hz

Hx a

nd H

z (

1/2

Ms)

Drive Current (au)

Figure 11. The maximum recording fields for a RH and a SPH, each with a trackwidth of 500 nm, and a RH

with a throat height of 500 nm, a gap length of 70 nm, and a SPH with a gap of 1000 nm, an ABS to SUL

separation of 35 nm, a throat height of 250 nm.

In this calculation, each of the two heads was assumed to have the same trackwidth of

500 nm. The RH was modeled with a gap length of 70 nm and a throat height of 500 nm,

while the SPH was modeled with a pole thickness of 500 nm, a gap length of 1000 nm, an ABS to SUL separation of 35 nm, a throat height of 250 nm. It can be noted that the

linear region slope for the SPH is almost four times as large as the linear region slope for the RH. If recorded on media with the same coercivity field, the saturation current for a

perpendicular system should be expected to be four times as less as it is for an equivalent

longitudinal system.

2.2.6. Focused-ion-beam Trimmed Single Pole Heads

Using FIB trimming of regular relatively large RH’s or/and SPH’s, it is possible to fairly economically fabricate a set of individual recording SPH's, with a required set of

parameters, including the trackwidth, the pole thickness, the gap length, the throat height,

the shape of the leading (trailing) edge, and others [63]. Study of the FIB-fabricated devices could give a good insight into the operation of realistic magnetic devices.

2.2.7. Example 1:FIB Trimming of a Wide-track Censtor SPH into a Narrow-track SPH

By courtesy of Censtor Corporation, relatively wide SPH's (approximately, with a 1- mtrackwidth) were available for further modification via FIB trimming [64,65]. The

modification included the reduction of the trackwidth down to approximately 100 nm.

Scanning electron microscope (SEM) image of a FIB-fabricated 120-nm wide SPH is shown in Figure 12.


Two side

FIB-made trenches

Track direction, X

Probe ABS

Figure 12. SEM image of a FIB trimmed Censtor head (ABS view at a 20-degree tilt).

MFM images of the perpendicular and longitudinal components of the field generated at

the ABS of this head with a drive current in the over-saturated regime (above 1000

mA turn) are shown in Figures 13a and b, respectively. The central cross-sections of these field component profiles are shown in Figure 13c. Although, there is no SUL in this case, the symmetry of the measured field profiles look similar to the symmetry of the

modeled profiles with a SUL, as shown in Figure 10. As mentioned above, the SUL has

mostly a quantitative effect and thus does not substantially change the shape of the field profile.

2.2.8. Example 2:FIB Trimming of a RH into a Narrow-track SPH

By courtesy of IBM Corporation, relatively wide track RH's (approximately, with a 1- m

trackwidth) were available for further modification via FIB trimming. The modification included not only the reduction of the trackwidth down to approximately 60 nm, but also

the increase of the gap length from its original value of 150 nm to the required value of

approximately 1 m. Scanning electron microscopy (SEM) image of thus FIB-fabricated

60 nm wide SPH with a 1- m gap length is shown in Figure 14.


400 nm0 nm 0 nm 400 nm(a) (b)

-200 -100 0 100 200

-0.6

-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

Perpendicular

Longitudinal

MF

M s

ignal (a

u)

Distance down the track (nm)

(c)

Figure 13. MFM images of the (a) perpendicular and (b) longitudinal field components generated by a FIB

trimmed Censtor head with a trackwidth of 120 nm and a pole thickness of 200 nm. (c) Cross-section profiles of

the perpendicular and longitudinal components.


W= 60nm

G = 1000

Main pole tip

Figure 14. SEM image of a SPH FIB-made from a RH (ABS view at a 40-degree tilt).

A MFM image of two adjacent 65 nm wide tracks with periodic sets of transitions

recorded onto a CoCr-based perpendicular medium with a SUL, as shown in Figure 15,

clearly indicates the functionality of thus fabricated SPH despite its nanoscale size trackwidth. It should be reminded that there is a general concern that as the SPH pole tip

dimensions are reduced to sizes substantially less than the characteristic domain wall

width in the soft material of which the pole tip is made of, the magnetization not only might become fairly "hard" to switch but also might display substantially non-zero

remanence [66]. A more detailed analysis of this issue is presented below.

130 nm

Figure 15. MFM image of two adjacent 65 nm wide tracks recorded onto a CoCr-based perpendicular medium

with a SUL.


2.2.9. Single Pole Head: Design StrategyIn this Section, a more detailed description of the SPH structure is presented with the

purpose to explain the approach chosen to design the SPH geometry, as shown in Figure

16, and thus clarify the limitations of this head design and motivate an approach for future modifications. The limitations are fundamentally caused by the inability to

infinitely maintain the linear scaling of the system dimensions (for increasing the areal

density) below the value, at which the flying height reaches its smallest value that is physically feasible. It is believed that it is unlikely to be able to maintain a steady flying

height below approximately 5 nm because of the proximity to the size of the air

molecules critically participating in the recording head flying process. Therefore, a deviation from the straightforward scaling law is necessary for further increasing the

areal density. This deviation can be accomplished through the modification of the SPH design. Hence, the understanding of the principles utilized to design SPH geometry shall

make SPH modifications most efficient for satisfying the demand for the areal density

increase.

Drive

coil

I

G T

WTH

TL

ABS

Leading

poleMain

pole

Figure 16. A schematic diagram of a SPH.

Before going into details, it is worth reminding the major requirements towards a write

head in perpendicular recording:

1)the ability to generate a sufficiently strong field for recording onto a medium with adequate coercivity,

2)the ability to generate sufficiently large trailing and side field gradients for

recording sufficiently sharp transitions and narrow tracks, respectively, 3)the ability to localize the recording field in a fairly limited region along the track

so that the skew angle sensitivity is minimized (see Section Skew Angle Sensitivity),

4)the ability to maintain reasonable efficiency of a recording system.


Below, the analysis of the parameters, which influence the above-listed requirements, is

presented. Before going into a description of the design methodology, it is worth reminding that already today the flying height in every state-of-the-art recording system

is of the order of 5 nm, which is already close to the size of the air molecule. Therefore, it

is hard to see how the flying height can be further reduced, because the air flow process is a critical link in the ultimate operation of a magnetic hard-drive. This means that the

established over decades trend of increasing the recording density in a magnetic

recording system only through the direct application of the scaling law should be adjusted for creating next generation magnetic technologies. In other words, for obtaining the

maximum benefit and achieving the highest possible areal density, special attention

should be given to each component of the magnetic recording system.

2.2.10. Definition of EfficiencyBefore going into details of the design, such basic quality indicator as the efficiency of a

recording system should be redefined for perpendicular recording [44]. In longitudinal

recording, the efficiency is the ratio of the magnetic flux generated in the deep gap of the RH and the flux in the drive coil 52. As mentioned above, for perpendicular recording it

is not the physical gap but rather the effective (magnetic) gap, defined as the separation

between the SPH and its image, is equivalent to the physical gap of the RH. Therefore, it

makes sense to define the efficiency, , of a magnetic system of the 1st perpendicular

mode as the ratio of the magnetic flux in the magnetic gap (the flux under the pole tip ABS) and the flux in the drive coil, as shown in Figure 17,

= Bgap Agap / N I, (2)

where Agap is the deep gap cross-section area.

I

Bdrive

Bgap

SUL

boundary

IBdrive

Bgap

Image

SPH

RH

Longitudinal “Circuit” Perpendicular “Circuit”

Figure 17. Diagrams showing magnetic “circuits” in a longitudinal system with a RH and a perpendicular

system with a SPH and a SUL.


2.2.11. Throat Height Dependence

The throat, being the narrowest part of the magnetic flux loop (circuit), typically, is also the highest reluctance link of the magnetic loop [67]. Thus, by reducing the throat height,

the relative contribution of the throat region into the net reluctance of the magnetic circuit

is also reduced and therefore, the overall efficiency of the system is increased. Also, as described below, by reducing the throat height, the recording field at saturation is

increased. There are two competing factors contributing into the increase of the recording

field as a result of the throat height reduction. First, the field is increased because as a result of the throat height reduction the point inside the pole tip at which the saturation

starts to occur is shifted closer to the ABS. Calculated magnetization contours along the

central cross-track planes inside the main pole tip in two extreme cases, with fairly tall and sufficiently short throats, are shown in Figures 18a and b, respectively. The

magnetization profiles at saturation along the central vertical line inside the pole tip for these two cases are shown in Figures 18c. It can be observed that for the tall throat, the

saturation occurs near the top region of the throat, thus only a relatively small part of the

initial magnetic flux generated by the drive coil reaches the ABS. As the current is increased beyond the saturation value, the most of the flux is going to leak out from the

magnetic loop on its way from the drive coil to the ABS. In contrast, for the short throat,

the saturation starts to take place at the ABS, thus the maximum possible flux reaches the

ABS and therefore the maximum possible field (for a flat surface, of the order of 4 Ms)

can be generated. In other words, in the latter case, there is effectively more magnetic "charge" generated at the ABS. The "charge" at the ABS is the required source of the

recording field.

TH

10 kOe

2 kOe

20 kOe

Saturated

region

+ + +

“charge” at the ABS

TH

19.8 kOe

19.6 kOe

20 kOe

Saturated

region

+ + + + + + +

“charge” at the ABS

(a) (b)


0 200 400 600

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

1.1

500 nm

100 nm

ABS

Mz/M

s

Along the central line inside the Pole Tip (nm)

(c)

Figure 18. Calculated magnetization contours along the central cross-track planes inside the main pole for two

extreme cases at saturation: (a) a fairly tall throat and (b) a sufficiently short throat. (c) Magnetization profiles

along the central vertical line for the two cases at saturation.

Because the charge is located at the ABS, the ABS dimensions of the pole tip determine

the recorded bit sizes. Therefore, being local by its origin, this is a favorable effect of the

throat height reduction. Unfortunately, the throat height reduction results in another effect, which deteriorates the field gradients. This effect is due to the "charges" created

on the tilted walls above the throat height of the main pole, as shown in Figure 19. These

"charges" generate an extra field, which is not localized and therefore results in the deterioration of the field gradients, as shown below. As the throat height is reduced, the

"charge" at tilted walls is effectively moved closer to the ABS, and thus, the effective contribution of this unfavorable field increases.

It should be remembered that although a perpendicular medium is ideally symmetric with respect to any of the two in-plane directions, i.e. along and across the track, a typical

SPH, as shown in Figure 19, is not [68]. Because of fabrication process limitations,

typically, the throat top boundaries (the line at which walls starts to deviate from being vertical) are defined only at the two cross-track side walls of the main pole, and not at

any of the two along-track side walls, as shown in Figure 19. It should be reminded that

the magnetic "charge" is proportional to the change of the magnetization component normal to the boundary surface [69]. Therefore, in the particular case, the magnetic

"charge" is concentrated on the cross-track sides rather than on the leading and trailing

sides of the main pole. As a result, because of the different amount of the "charge" in these two cases, the throat height dependencies might be quantitatively different for the

field profiles along and across the track, respectively, as shown below.


TH

WPT

+ + +

+ + +

+ + +

+ + +

Magnetic

“charges”

on side walls

LS

M

Trailing

side wall

Cross-track

side wall

Figure 19. A diagram of a SPH pole tip showing location of side wall “charge."

0.0 0.2 0.4 0.6 0.8 1.00.0

0.5

1.0

1.5

2.0 TH = 100 nmTrailing

edge

50100

200

500 au

Hz (

10 x

kO

e)

Distance down the track ( m)0.0 0.2 0.4 0.6 0.8 1.0

0.0

0.5

1.0TH = 100 nm

Trailing

edge

50

100

200

500 au

Hz / H

z 0


(a)


0.0 0.2 0.4 0.6 0.8 1.00.0

0.5

1.0

1.5

2.0 TH = 500 nm

Trailing

edge

50100

200500 au

Hz (

10 x

kO

e)

Distance down the track ( m)0.0 0.2 0.4 0.6 0.8 1.0

0.0

0.5

1.0

TH = 500 nmTrailing

edge

50

100200

500 au

Hz /

Hz 0


(b)

Figure 20. Along-track profiles of the perpendicular field component and its normalized value for two values

of the throat height, (a) 100 and (b) 500 nm.

The along-track profiles of the perpendicular field component and its normalized value at a 5 nm flying height and a 20 nm separation between the ABS and the SUL at different

values of the drive current (in the arbitrary units) for two values of the throat height, 100

and 500 nm, are shown in Figures 20a-b, respectively. In this case, the side-wall tilt

angle, , as shown in Figure 19, was modeled to be 45 degrees. The perpendicular fields and their normalized values for the same set of parameters across the track are shown in

Figures 21a-b, respectively. As expected, it is observed that although it is easier to drive

more recording field in the case of the shorter throat, the undesirable off-track side field also increases due to the increased contribution from the "charge" at the tilted sidewalls.

To explicitly illustrate this effect, two cross-track perpendicular field profiles

corresponding to the two throat height values at saturation are put together in Figure 22a. The same profiles normalized to the corresponding values at the center of the track are

shown in Figure 22b. The normalized profiles directly illustrate the fact that the shape of

the field profile is substantially wider in the shorter throat height case.

0.0 0.1 0.2 0.3 0.40.0

0.5

1.0

1.5

2.0TH=100nm

50100

200

500 au

Hz (

1/2

Ms)

Distance across the track ( m)0.0 0.1 0.2 0.3 0.4

0.0

0.5

1.0 TH=100nm

50100

200

500 au

Hz /

Hz 0

Distance across the track ( m)

(a)


0.0 0.1 0.2 0.3 0.40.0

0.5

1.0

1.5

2.0TH=500nm

100200

500

1000 au

Hz (

1/2

Ms)

Distance across the track ( m)0.0 0.1 0.2 0.3 0.4

0.0

0.5

1.0

TH=500nm

100200

500

1000 au

Hz /

Hz 0


(b)

Figure 21. Cross-track profiles of the perpendicular field component and its normalized value for two values of

the throat height, (a) 100 and (b) 500 nm.

0.0 0.1 0.2 0.3

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

TH = 500 nm

TH = 100 nm

Hz (

10 x

kO

e)


(a)


0.0 0.1 0.2 0.3

0.2

0.4

0.6

0.8

1.0

1.2

TH = 500 nm

TH = 100 nm

Norm

aliz

ed H

z (

Hz/H

z o)


(b)

Figure 22. (a) The cross-track profiles at saturation for two values of the throat height, 100 and 500 nm. (b)

The normalized profiles at saturation.

The field 5 nm below the center of the main pole versus the drive current at three values of the throat height, 100, 200, and 500 nm, is shown in Figure 23a. The drive current is

given in arbitrary units because the exact value of the current depends on a number of

specific to each head design parameters, such as the exact location of the drive coil with respect to the ABS, the yoke geometry, etc. The saturation current can be defined at the

value, at which the first discontinuity (change of the slope) in the field dependence on the

current takes place. Thus derived saturation current (reflecting the system efficiency) versus the throat height is shown in Figure 23b. In summary, it can be noticed that

reduction of the throat height has two favorable effects, the increase of the recording field

and the reduction of the saturation current. However, the throat height can not be reduced entirely to zero, because the smaller the throat height is, the worse the side and trailing

field gradients are, as noted above (see Figure 22).

Here, it should be mentioned that for an ideally saturated state, the field due to the side

charge could be easily calculated according to the Coulomb Law for the “magnetic” charge on the sidewalls. It can be shown that at zero throat height and at a tilt angle of 45

degree, the extra field due to the side "charge" could substantially overcome 4 Ms (the maximum field assuming the pole tip with no side wall charge) provided the side wall is

sufficiently tall. Considering the side nature of the source of this field, the field due to the

side “charge” not only increases the field under the pole tip (on the track) but also creates an unfavorable field at the sides (off the track) and thus deteriorates the field gradient.

Therefore, for minimizing the contribution due to the side “charges”, it is preferable to


keep a sufficiently tall throat. In other words, there is a trade-off between the field

magnitude and the field gradient, and this trade-off can be controlled by the throat height.

0 200 400 600 800 1000

0.0

0.5

1.0

1.5

2.0

500 nm

200 nm

TH=100 nm

Hz 0 (

1/2

Ms)

Drive Current (au)100 200 300 400 500

70

75

80

85

90

Satu

ration C

urr

en

t (a

u)

Throat Height (nm)

(a) (b)

Figure 23. (a) The perpendicular field versus the drive current at three values of the throat heigth, 100, 200, and

500 nm. (b) The saturation current versus the throat height.

2.2.12. Dependence on the Pole Trackwidth and ThicknessAnother way to increase the recording field is to make each ABS cross-section dimension

of the SPH pole tip (pole thickness and trackwidth) as large as possible [70]. The

characteristic dimension, at which the field starts to substantially change, is determined by the doubled (due the “image” by the SUL) distance between the ABS and the SUL. The

trackwidth, W, of the SPH determines how narrow a track can be recorded. Therefore, the

trackwidth value is set by a required areal density value. For example, at an areal density beyond 100 Gbit/in2, the trackpitch (the trackwidth plus the guard band) should be smaller

than approximately 160 nm, assuming a 4:1 bit aspect ratio (BAR). Assuming that the

guard band occupies approximately a fifth part (20 percent) of the trackpitch, the SPH should have a trackwidth of approximately 120 nm for recording an approximately 130 nm

wide track. As to the pole thickness, as previously mentioned, in perpendicular recording, ideally, the actual recording takes place only near the trailing edge of the pole, therefore,

one can have the pole thickness as large as necessary for the maximum increase of the

recording field. The maximum recording at saturation versus the pole thickness for a given trackwidth of 120 nm is shown in Figure 24a. However, in practice, as explained in

Section below, the pole thickness cannot be made infinitely long because in a realistic

hard-drive, the skew angle is not always zero. The non-zero skew angle results in effectively recording a substantially wider track, as compared to the trackwidth of the pole

tip. As shown below, the pole thickness value of approximately 200 nm should reduce the

skew angle sensitivity to few percent of the trackwidth value, assuming approximately a 10 degree maximum skew angle and areal densities below approximately 400 Gbit/ in2. At

400 Gbit/ in2 areal density, assuming a 4:1 bit aspect ratio, the recorded trackpitch should


be 80 nm wide. Therefore, the trackwidth of the pole tip should be smaller than 80 nm.

The maximum saturation field versus the pole trackwidth at a fixed value of the pole thickness of 200 nm is shown in Figure 24b. At this point, it is worth reminding that the

image head is located effectively further away from the center of the recording layer as

compared to the real head, as shown in Figure 9, with the spacing difference being equal

to the recording layer thickness. Ideally, the net recording field of 4 Ms can be produced as a result of the contributions of the fields generated by both, the real and image heads,

respectively, with a 2 Ms field per each head. Assuming a 30-nm separation between the ABS and the SUL, at such high areal densities, the trackwidth (~< 80 nm) is of the same

order of magnitude as the doubled separation between the ABS and the SUL. Therefore, it

is not unnatural that the net recording field starts to substantially drop as the trackwidth is further reduced.

(a)

0 100 200 3000.0

0.5

1.0

1.5

W = 120 nmHz / 2

Ms

Pole Thickness (nm)

(b)

0 100 200 3000.0

0.5

1.0

1.5

2.0

T = 200 nmHz / 2

Ms

Trackwidth (nm)

Figure 24. The maximum field at saturation (a) versus the pole thickness for a SPH with a trackwidth of 200 nm,

and (b) versus the pole trackwidth at a fixed value of the thickness, 200 nm.


In summary, ideally, assuming a zero skew angle, the pole thickness can be made infinitely large because the recording is produced only near the trailing edge.

Nevertheless, the increase of the thickness results only in approximately 30 percent

increase if the trackwidth is kept as small as 120 nm. Moreover, in real conditions with a non-zero skew, the non-zero length of the pole thickness, T, results in substantial side

recording, as explained below in Section Skew Angle Sensitivity.

2.2.13. Skew Angle Sensitivity of Single Pole Head

One of the most serious issues during the future implementation of perpendicular

recording is believed to be the excessive sensitivity of a typical perpendicular recording system to the skew angle [41,71]. As mentioned above, unlike in longitudinal recording,

for which the recording is produced by the fringing field in the physical gap region of a RH, as shown in Figure 2, in perpendicular recording, the recording is produced in the

effective gap near the trailing edge of the main pole of a SPH, as shown in Figure 1. As a

result, one of the principal differences is the order of magnitude difference between typical sizes of the gap region of the RH and the trailing pole thickness of the SPH.

(a)

W2T2

Trailing edgeTrailing Pole Recorded Track

Skew angle

W

Side Band

(b)


(c)

Figure 25. (a) A schematic diagram illustrating the definition of the skew angle. (b) A diagram showing the

ABS of the trailing pole with a skew with respect to the track direction and transitions recorded with the skewed

trailing pole, thus creating a recorded track. (c) Guide diagram and MFM images of tracks recorded by a SPH

onto a CoCr-based perpendicular medium at zero skew and a 15 degree skew angle, respectively.

For the state-of-the-art recording RH and SPH suitable for areal densities of the order of

100 Gbit/in2, for example, the gap thickness and the trailing pole thickness are of the order of 50 nm and 1000 nm, respectively. Such a substantially larger thickness of the

SPH pole results in its extreme sensitivity to the skew angle. To help understand this

issue, the top view of the head assembly over the surface of a disk is shown in Figure 25a. This Figure also illustrates the definition of the skew angle, which is the angle

between the direction of the track and the head axis of symmetry. A diagram of a track

recorded by a SPH at a non-zero skew angle is shown in Figure 25b. It can be noted that at the condition of non-zero skew angle the recording is produced not only by the trailing

edge but also by one of the sides of the trailing pole. For comparison, MFM images of

two real tracks recorded with a SPH with a 500 nm thick pole at zero and a 15-degree skew angle onto a CoCr -based perpendicular medium are shown in Figure 25c.

Consequently, it is clear that the thicker the trailing pole is, the more sensitive the system

is to the skew angle. To a sufficient degree of approximation, the side written region is

proportional to the product T2 x sin , where T2 and are the pole thickness and the skew

angle, respectively, as shown in Figure 26. Assuming typical values for T2 and of approximately 1000 nm and 10 degrees, respectively, the side written region can be of the

order of 150 nm, which is unacceptable at areal densities beyond 100 Gbit/in2. It should

be reminded that the entire trackwidth is expected to be less than 150 nm at such high densities assuming a 4:1 bit aspect ratio (BAR).

P2

Track direction

Side written

region

Trailing edge

Figure 26. A diagram showing how the side recording is generated due to a non-zero skew angle.


One of the “perks” about the skew angle sensitivity in perpendicular recording is its

dependence on the linear density. MFM images of recording tracks written with a SPH at a 15-degree skew angle at five values of the linear density, 20, 40, 60, 80, and 100 kfci,

are shown in Figures 27a to e, respectively. These images clearly illustrate the

disappearance of the undesirable side written region with the increase of the linear density. The following describes another experiment illustrating the density dependence

of the skew angle sensitivity. Two sets of three tracks were recorded by a SPH at zero

and a 15-degree skew angle, respectively. The central track was recorded at a relatively low linear density of 25 kfci, while the two side tracks were recorded at a relatively high

linear density of 250 kfci. Then, a relatively narrow (80-nm wide) read head was used to

scan the tracks in the cross-track direction. Thus obtained track profiles are shown in Figure 28. Clearly, broadening of the effective trackwidth could be noticed for the

central track (which was recorded at the lower linear density value). Previously, the described linear density dependence was explained by the insufficient magnetic field

gradient in the side region [71].

Figure 27. MFM images of recording tracks written with a SPH at a 15 degree skew angle with linear densities

of a) 20 kfci, b) 40 kfci, c) 60 kfci, d) 80 kfci and e) 100 kfci.


-50 -25 0 25 50

1.5

2.0

2.5

Skew = 0 degrees

Pla

yback S

ignal (m

V)

Offset across the track ( in)

Skew = 15 degrees

250kfci 250kfci

25kfci

Figure 28. Track profiles for a set of three tracks (with the central and side tracks recorded at 25 and 250 kfci

linear densities, respectively) at two values of the skew angle, 0 and 15 degree, respectively.

The most straightforward "solution" for eliminating the skew sensitivity is the reduction

of the pole thickness. However, this solution is not adequate for ultra-high density

recording, because in this case the recording field, as shown earlier (See Section 2.2.12. Dependence on the Pole Trackwidth and Thickness), drastically drops and thus recording

on a sufficiently high coercivity medium becomes problematic [61]. For example, the

calculated perpendicular field at saturation versus the distance down the track near the trailing pole edge with a 120-nm trackwidth and with a 20-nm separation between the

ABS and the SUL at two different values of the pole thickness, 100 and 500 nm, is shown

in Figure 29. In this calculation, the material of which SPH and SUL were made was

modeled as a relatively high moment material with a 4 Ms of 2 T [72]. It can be seen that the reduction of the pole thickness to 100 nm reduces the field almost by a factor of two.

Another approach should be found to solve the fundamental issue of the skew angle

sensitivity. For example, it was shown that the use of a trapezoidal write pole could partially reduce the skew sensitivity [73,74]. Good understanding of the mechanisms

determining the recording field shall help find a more drastic solution.


0.9 1.0 1.1 1.20

4000

8000

12000

16000

PT = 0.1um

PT = 0.5um

Hz (

Oe

)


Figure 29. Modeled vertical fields near the trailing edge for two values of the pole thickness, PT, of 0.5 m and

0.1 m, with the same trackwidth of 0.1 m.

2.2.14. Gap Length Dependence

Above, in Section Second Perpendicular Mode: a Ring Head and a Perpendicular

Medium Without a Soft Underlayer, it was shown that as a direct consequence of the recording by the field fringing from the gap, properties of a system utilizing a RH without

a SUL, regardless of whether it is perpendicular or longitudinal recording, fairly strongly

depend on the physical gap length, as discussed above. This is in contrast with the case of the 1st perpendicular mode, for which no significant dependence on the physical gap

length can be expected as long as the gap length is substantially larger than the separation between the ABS and the SUL.

P1 P2tgap

P1

SUL

tP2-to-SUL

(a) (b)

Recording Layer

P2

tgap

MMFSource of

magnetic flux

(coils)

hP2wP2 tP2

Figure 30. A schematic diagram of: (a) the air bearing surface (ABS) view and (b) the side view of a single

pole head (SPH).


For the 2nd perpendicular mode, the physical gap is a part of the main path for the

magnetic flux in a recording system. As a result, in case of the 2nd mode, a stronger dependence on the gap length is expected. On the contrary, for the 1st perpendicular

mode, the main path for the magnetic flux does not go through the physical gap region, it

rather goes through the SUL, which explains the substantially weaker dependence on the physical gap length as long as the gap length is substantially larger than the separation

between the ABS and the SUL. However, if the gap length is reduced to the values

comparable with the separation between the head and the SUL, the dependence on the gap becomes essential for the 1st perpendicular mode as well. Because of the potential

practical benefits that could be envisioned based on the understanding of the dependence

on the gap length, a detailed analysis of the physical gap influence on the recording characteristic of a system with a SUL is given below [55].

-0.5 0.0 0.5

0

4

8

12

Ve

rtic

al F

ield

(kO

e)

Distance down the track ( m)0.0 0.5 1.0 1.5

-2

0

2

4

6

8

Vert

ical F

ield

(kO

e)


P1P2

Gap

Tra

ilin

g e

dg

e

of

the w

rite

fie

ld

P2P1

Tra

ilin

g e

dg

e

of

the

wri

te f

ield

Gap

Ring Head

(writing by “gap field”)Ring Head + SUL = Single Pole Head

(writing by the trailing edge of P2)

Figure 31. Vertical field profiles for a RH (medium without a SUL) and a SPH (medium with a SUL).

As mentioned above, commonly, it is believed that the gap, tgap, between the leading and

trailing poles in a SPH (See Figure 30) has to be sufficiently large (compared to the

separation between the air-bearing-surface (ABS) of the head and the SUL, tP2_to_SUL), i.e. tgap >> tP2_to_SUL, to achieve satisfactory recording performance.

In this Chapter, such SPHs will be referred to as type I. On the other hand, SPHs, for which the gap length is comparable or smaller than the ABS-to-SUL separation, will be

referred to as type II. In addition, there exists an additional qualititavely different type of

a SPH, in which the gap length is substantially (orders of magnitude) larger than any other magnetic dimension of the recording head. Such a head design will be referred to as

type III SPH. As shown below, the type III SPH is equivalent to a SPH in which the leading pole is not present at all. The recording performance of the type III SPH design is

also discussed below.


0.0 0.2 0.4 0.6 0.8 1.0

0.0

0.5

1.0

1.5

Leading

Edge

Trailing

Edge

Hz (

kO

e)


0nm

30nm

70nm

150nm

300nm

700nm

Figure 32. Modeled perpendicular write field for type I and type II SPH’s under a pole tip. The variation

parameter is the gap length.

One can note that physically the type II SPH design looks exactly as the RH design.

However, in the following analysis, it is chosen to differentiate between these two head

designs. For clarity, a SPH and a RH imply the presence and the absence of a SUL,

respectively. It should be reminded that recording with a RH onto a medium is

fundamentally different from recording with a SPH due to the different nature of the

write fields generated by these types of recording heads [53]. As mentioned earlier,

during recording with a SPH, the writing is accomplished by the trailing edge of the

trailing pole. This is in contrast to the recording with a RH, during which the writing is

accomplished near the trailing edge of the gap. If a conventional RH is used to write onto

a medium with a SUL, it acts as a SPH of type II. This is emphasized in Figure 31, where

the two cases of along-the-track vertical field profiles for the same writer are shown:

when the writer is used with a medium without a SUL (the writer acts as a RH) and when

the writer is used with a medium with a SUL (the writer acts as a SPH).

Field Modeling. The magnetic properties of SPH’s are investigated using the above

mentioned 3D-boundary element modeling software Amperes. In the calculations

presented below, an ideal SUL (with ideal magnetic imaging properties) was assumed.

For explanation clarity and without sacrificing the physical substance, the finite thickness

and permeability and the micromagnetic effects are ignored [12]. In these calculations,

the following values are chosen: the ABS-to-SUL separation = 30 nm; CoFe alloy with a

4 MS of 22 kG is chosen as the SUL and yoke material; the separation between the ABS

and the point at which the field values are measured is 10 nm; unless specified otherwise,

a 200 nm wide trailing pole is assumed.


0.0 0.2 0.4 0.6 0.8 1.0

0

5

10

15

20

Trailing

edge

Type II SPH

Hz (

kO

e)


1 mA turn

2 mA turn

5 mA turn

10 mA turn

20 mA turn

40 mA turn

60 mA turn

100 mA turn

160 mA turn

200 mA turn

Figure 33. Modeled perpendicular write field profile for a type II SPH (70nm gap) under the trailing pole tip at

different values of the magnetizing current.

The perpendicular field profiles for SPH’s with several different values of the gap

thickness are shown in Figure 32. According to the definition above, the 30-nm and 70-

nm gap SPH’s represent typical type II SPH’s (or RH’s if there is no SUL) while the 700-

nm gap SPH represents a type I SPH. A gapless SPH, i.e. a SPH with no gap at all (gap

thickness is equal to zero), is also a type II SPH.

It could be observed that regardless of the gap thickness, the magnetic properties at the

trailing edge are essentially identical including a maximum write field of ~18 kOe and a

maximum trailing gradient of ~ 350 Oe/nm, as shown in Figure 32. The maximum write

field here and later in the Chapter is defined as the value of the write field at a certain

maximum value of the magnetizing current, IMAX, above which the trailing gradient starts

to deteriorate. It should be stressed that a higher field is possible to achieve at the expense

of a deteriorated trailing gradient. As the gap becomes thinner, the field profile under P2

changes dramatically so that a substantial decrease of the magnitude of the vertical field

towards the leading edge of P2 is observed in the type II SPH.


0.2 0.4 0.6 0.8 1.0 1.2

0

5

10

15

20

Type I SPHTrailing

EdgeH

Z (

kO

e)


1 mA turn

2 mA turn

5 mA turn

10 mA turn

20 mA turn

40 mA turn

60 mA turn

100 mA turn

160 mA turn

Figure 34. Modeled perpendicular write field profile for a type I SPH (700nm gap) under the pole tip at

different values of the magnetizing current.

The longitudinal component of the write field under the P2 is negligible. It becomes

essential in the gap region and its maximum value, located near the trailing edge of the

P1, varies from ~3.5kOe to ~4.5kOe as the gap thickness is decreased from 700 nm to 30

nm. In this region, the vertical field is negligible and the net field amplitude is far below

the threshold value of the write field necessary to write onto media designed to be written

onto with a 18 kOe write field at the trailing edge of the P2.

The write field profiles for a type II SPH for several values of the magnetizing current are

shown in Figure 33. It can be noted that the above mentioned decrease of the vertical

component of the write field towards the leading edge of P2 in type II SPH’s is observed

only at larger values of the magnetizing current. For comparison, the perpendicular field

profiles for a type I SPH for several values of the magnetizing current are shown in

Figure 34.

It is instructive to plot the maximum value of the write field at the trailing edge and the

maximum value of the trailing gradient versus the current value in the magnetizing coils.

These dependencies are shown in Figures 35 and 36, respectively. Two distinct regimes

of the P2 magnetization process can be clearly identified: an initial steep increase of the

field and gradient values in a narrow range of the magnetizing current values is followed

by a weaker dependence of the field and gradient on the current. The second regime starts

to manifest itself as the top portion of the pole tip (throat) starts to saturate.


0 40 80 120 160 2000

5

10

15

20

HZ (

kO

e)

I (mA x turn)

30 nm gap

70 nm gap

150 nm gap

300 nm gap

700 nm gap

Figure 35. The amplitude of the write field at the trailing edge versus the current in the magnetizing coils for

different values of the gap thickness. The dotted line represents an 18 kOe level.

A marginal decrease of the trailing field gradient observed as the gap thickness is

decreased (see Figure 36), is only ~5% of the total trailing gradient magnitude.

0 50 100 150 200

50

100

150

200

250

300

Tra

iling G

radie

nt

(Oe/n

m)

Magnetizing Current (mA turn)

30 nm gap

70 nm gap

150 nm gap

300 nm gap

700 nm gap

Figure 36. The maximum trailing field gradient versus the current in the magnetizing coils for different values

of the gap thickness.

The effect of the decrease of the write field towards the leading edge of P2 is the result of

the specific magnetic configuration of the type II SPH. The difference between the

magnetic configurations of the type I and type II SPH’s is illustrated in Figure 37, where

the magnetic paths of the flux generated at P2 are compared for type I and type II SPH’s.

While in the type I SPH most of the magnetic flux flows directly into the SUL, in the


type II SPH the magnetic flux generated at P2 is distributed between the SUL and P1.

Consequently, in the type II SPH more flux reaches the trailing edge (rather than the

leading edge) of the ABS of P2. This can be clearly observed from Figure 38 where

contour plots of the vertical component of the magnetization in the middle plane of P2 for

both type I and type II SPH’s are shown.

P1

SUL

P2P1

SUL

P2

tP2tP2

(a) (b)

Figure 37. A schematic diagram illustrating the difference between the magnetic flux paths in (a) a type I SPH

(a) and (b) a type II SPH.

In the past, it has been shown that the trailing pole thickness is one of the limiting factors

controlling the maximum write field that a SPH can generate [4]. To maximize the write

field, the trailing pole has to be thicker than a certain critical thickness defined by the

overall magnetic system configuration. The SUL-to-ABS separation and the width of P2,

the parameter that defines the track width, are among the most critical parameters that

strongly affect the critical thickness of P2. However, it should be reminded that, for a

perpendicular system, the thicker the trailing pole is, the more sensitive the recording

system is to the skew angle [9].

The maximum write field at the trailing edge versus the trailing pole thickness for type I

and type II SPH’s with gap thickness values of 700 and 70 nm, respectively, is shown in

Figure 39. It could be observed that a thinner trailing pole could be utilized in a type II

SPH to achieve higher write fields.

The above phenomenon could be explained if one observes that in the type II SPH

configuration, a SUL could be thought of as being effectively wrapped around P2, as

illustrated in Figure 40. This effectively increases the total thickness of P2, thus enabling

a larger write field at a smaller value of P2 geometrical thickness.


Leading

edge

Trailing

edge

ABS

ABS

Type II SPH

Type I SPH

Figure 38. Vertical component of the magnetization in the middle plane of P2 for type I and type II SPH’s.

Darker shading corresponds to a higher value of MZ.

SPH Efficiency versus RH Efficiency. One of the concerns in using a type II SPH is the

efficiency of this type of head, i.e. the amount of current through the magnetizing coil

necessary to bring the head to the operating regime. It is commonly believed that the

relatively wide gap in a SPH is a prerequisite for the head to be efficient. However, as

shown below, the efficiency loss in the narrow-trackwidth type II SPH’s is only marginal.

0.2 0.4 0.6

1.0

1.2

1.4

1.6

1.8

HM

AX (

kO

e)

P2 Thickness ( m)

Type I SPH

Type II SPH

Figure 39. The maximum perpendicular write field for type I (700nm gap) and type II (70nm gap) SPH’s

under a pole tip.

The equivalent electrical circuit for the system utilizing an SPH is shown in Figure 41

[75]. In this approximation, it is assumed that the magnetic flux path through P2 is the

same for the magnetic flux flowing into the ABS of the P2 and into the gap region. If the

magnetizing coil is located near the ABS, the major contribution to the overall reluctance


of the magnetic circuit comes from the magnetic reluctance, RP2, of the throat region of

P2, the magnetic reluctance, RGAP, of the free space between P1 and P2, i.e. the gap

region, and the magnetic reluctance, RP2-to-SUL, of the free space between the P2 and SUL.

Then, neglecting RP1-to-SUL, RSUL, and RYOKE, for a given magneto-motive force, MMF,

generated by the magnetizing coil, the net magnetic flux, , is given by

.RR

RR R where,

RR

MMF

SULtoP2GAP

SULtoP2GAPAIR

AIRP2

(3)

P1

SUL

P2 P1

SUL

P2

tP2

tP2

Figure 40. A schematic drawing illustrating the effective increase of the P2 thickness in type II SPH.

It is straightforward to show that flux between the ABS of the P2 and the SUL, P2-to-SUL,

is given by

.

RR

R1R

MMF

SUL-to-P2

GAP

SUL-to-P2P2

SUL-to-P2 (4)

This equation can be rewritten in terms of the physical dimensions of the SPH. Then the

vertical magnetic field, B, under P2 is given by

,

ttt

ht1

h

MMFB

SUL-to-P2

P2GAP

P2SUL-to-P2P2

0 (5)

where 0 is permeability of air (free space), is the relative permeability of P2, hP2 is the

throat height, tP2 is the P2 thickness, wP2 is the P2 width (track width), tGAP is the gap

thickness, and tP2-to-SUL is the SUL-to-ABS separation, as illustrated in Figure 40.


~

RP2

RP2-to-SUL

RGAP

RYOKE

MMF (coil)

RP1-to-SUL

P2-to-SUL

RSUL

Figure 41. A schematic diagram showing an equivalent magnetic circuit for a narrow-gap SPH based

recording system.

To compare the efficiency of a typical type II SPH to the efficiency of a typical type I

SPH, the above equation is evaluated for the following two cases: type II SPH with tP2-to-

P1=tP2-to-SUL and hP2=tP2 and type I SPH with tGAP=infinity. This gives

.

t2h

MMFand B

th

MMFB

SUL-to-P2P2

0SPH II type

SUL-to-P2P2

0SPH I type

(6)

Because is usually a large number in excess of 100, the last two equations show that the

efficiency values for the two SPH types are similar.

It should be remembered that the above arguments are only an approximation based on a

number of assumptions such as zero flux leakage (fringing fields are neglected), the

absence of non-linearity in the yoke structure, uniform flux distribution, etc. However, as

shown below, based on the direct magnetic field modeling results, the above conclusion

that there should be no drastic difference between the efficiency values for the two SPH

types remains valid. The dependence of the saturation current, ISAT, on the gap thickness

is shown in Figure 42. ISAT is defined as the value of the current in the magnetizing coils

required to reach a certain operating write field. In this example, the operating write field

is chosen to be 18 kOe. It can been observed that although the saturation current increases

with the decrease of the gap length, a change of only ~30% is observed as the gap length

is decreased from 700nm to 70nm and a factor of two increase of the operating current is

observed for a 0-nm gap. It should be stressed that this is in contrary to a common belief

that the decrease of the gap thickness would lead to a dramatic, by orders of magnitude,

reduction of the SPH efficiency. It should also be stressed that the efficiency of a SPH is


strongly dependent on the head design and could be adjusted to match particular

recording system requirements.

0 200 400 600 800

125

150

175

200

225

250

Type I SPH

Type II SPH

Opera

ting C

urr

ent

(mA

turn

)

Gap Thickness (nm)

Figure 42. The dependence of the saturation current, ISAT, on the gap thickness.

Both of the above mentioned properties of type II SPH’s, i.e. the decrease of the write

field towards the gap and the ability to generate a stronger write field at a smaller value

of the trailing pole thickness, can be used to advantage to minimize the skew angle

sensitivity.

Skew Angle Versus Gap Length. It is expected that a type I SPH with a 500nm thick P2

at a 15-degree skew will write a 130 nm wide detrimental side band (illustrated in Figure

43) leading to necessity to reduce the track pitch by approximately 25%.

Recorded track

Side band due to

non-zero skew

Figure 43. Illustration of non-zero skew angle sensitivity.

This results in a substantial net loss in the areal bit density. If a type II SPH is used, e.g.

with a 30nm gap, the effective length of the P2 is substantially reduced. If a recording

medium with a 10% of switching field distribution is used, the effective P2 thickness of a

30-nm gap SPH reduces to ~150nm. Consequently, the detrimental side band width is

reduced to 40 nm. In this case, the track pitch needs to be reduced by only 8%, which is

clearly advantageous for maximizing the areal bit density.

The extreme case of the type II SPH design is a gapless SPH. According to the analysis

above, it is expected that the gapless SPH would have the best skew angle performance.

To confirm this expectation, the following experiment was performed. A gapless SPH


was manufactured according to the thin-film fabrication process. An SEM image of the

air bearing surface view of one of the manufactured gapless SPH is shown in Figure 44.

P1

P2

P1

P2

Figure 44. SEM image of the ABS view of a gapless SPH. (There is no gap between the trailing pole, P2, and

the leading pole, P1.)

To compare the skew sensitivity of a gapless SPH with an equivalent non-zero gap type

II SPH, two sets of tracks were recorded with a 1- m wide gapless SPH and a 1- m wide

type II SPH with a 1- m thick gap, respectively, at a skew angle varying from –15 to 15

degrees. At every skew angle, the effective trackwidth was measured by a relatively

narrow (120-nm wide) GMR read head. The measured skew angle dependences for the

two heads, respectively, are shown in Figure 45. The comparison clearly indicates a

substantially weaker dependence on the skew angle for the gapless head.

-15 -10 -5 0 5 10 15

1.2

1.4

1.6

1.8 1 m Gap

Gapless

Tra

ck w

idth

(m

)

Skew angle (degrees)

Figure 45. The effective readback trackwidth versus the skew angle for two SPH’s, with zero and a 1- m gap,

respectively.


Single Pole Head of Type III. As mentioned above, a type III SPH represents a distinct

type of SPH. In the type III SPH design, the role of P1 is built-in to be negligible, thus a

generic representative of this type of SPH is a SPH which does not have P1 pole at all.

The absence of P1 leads to the deteriorated efficiency of a type III SPH, as compared to

an equivalent type I SPH. However, for the dimensions considered in this book (to satisfy

areal densities beyond 100 Gbit/in2) at all other equivalent conditions, the saturation

current for the type III SPH is only ~30 % higher than the saturation current for the type I

SPH (See Figure 46).

0 50 100 150 200

0

5

10

15

20

HZ (

kO

e)

Magnetizing current (mA x turn)

Type I SPH (700nm gap)

Type II SPH (70nm gap)

Type III SPH

Figure 46. The amplitude of the trailing field versus magnetizing current for the three types of SPH’s.

It should be noted that although a type III SPH is not the most efficient type of SPH, the

absence of P1 helps to reduce the sensitivity of the SPH to a stray field [4].

Experiments to Compare Different Types of SPH’s. A type I SPH used in this

experiment was fabricated from one of two identical RH’s via focused ion-beam (FIB)

trimming to increase the gap thickness. The second RH was used as a type II SPH (see

discussion above). Additionally, both SPH’s were FIB trimmed to allow for a 300nm

trackwidth, as shown in Figure 47. The recording performances of the resultant type I

SPH with a gap of 1 m and of the type II SPH with a gap of 80 nm were compared. A

39 nm thick Co/Pd supperlattice based recording layer and a Ni45Fe55 (or FeAlN when

indicated) SUL at a 12 nm flying height were utilized.


Type I SPHType II SPH

Gap

Figure 47. Electron micrographs of a type II and type I SPH’s prepared from two identical RH’s via focus ion-

beam trimming. Both heads were trimmed to allow for a 300nm trackwidth. A white line on the type II SPH

micrograph is a guide for an eye to outline the location of the gap.

The spin-stand-measured saturation current (Isat) and PW50 versus the soft underlayer

thickness for the type I SPH and the type II SPH are shown in Figures 48 and 49,

respectively. Although for a sufficiently thick soft underlayer the Isat is larger for the type

II SPH, PW50 for the type II SPH is approximately as small as PW50 for the type I SPH.

When the SUL becomes too thin, the type I SPH, which is designed to operate with the

SUL cannot properly function, because the SUL saturates, while the type II SPH

gradually transforms into a RH when the SUL saturates and can still write on the media

with relatively low coercivity utilized in these experiments (Hc~2,500 Oe). Due to a

higher value of 4 MS of FeAlN, the thickness at which the FeAlN SUL begins to saturate

is smaller than the thickness at which the Permalloy SUL begins to saturate.

0.0 0.2 0.4 0.6 0.8

50

100

150

200

250

300

350

Satu

ration c

urr

ent (m

A turn

)

Underlayer thickness (um)

Type I and Permalloy

Type II and Permalloy

Type I FeAlN

Type II and FeAlN

Figure 48. Saturation current and (a 70 nm thick reader with a 100 nm shield to shield separation) versus the

SUL thickness.


From Figures 48 and 49 it can be observed that the thickness at which the SUL begins to

saturate is smaller for the type II SPH than for the type I SPH. The origin of this effect

becomes clear if one recalls that the minimum thickness of a SUL is defined by the

ability of the SUL to carry the magnetic flux emanating from the P2 (See Ref. [12]). In a

type II SPH a certain fraction of the flux gets channeled into the leading pole, P1, thus,

relaxing the minimum thickness requirement for a SUL. The possibility of using a thinner

SUL is an additional advantage of utilizing a type II SPH in a perpendicular recording

system.

0.0 0.2 0.4 0.6 0.8

130

140

150

160

170

180

PW

50 (

nm

)

Underlayer thickness (um)

Type I SPH

Type II SPH

Permalloy SUL

Figure 49. PW50 (a 70 nm thick reader with a 100 nm shield to shield separation) versus the SUL thickness.

The above shown experimental results are in agreement with the theoretical prediction of

only a marginal deterioration of the head efficiency in the type II SPH design versus the

type I SPH design. Although, many factors contribute to the measured value of PW50,

the invariance of the PW50 with respect to the type I or type II SPH design is indicative

that the trailing gradient is not much affected by the gap thickness.

2.2.15. Flying Height Limitation of Single Pole Head Design

As mentioned above, the fundamental density limitation of the regular SPH design is due

to the inability to scale the flying height as the areal density increase demands for the reduction of the flying height to values below physically impossible [61]. For example, it

is believed that the smallest achievable flying height is approximately 5 nm. It is hard to

see how one can make the flying height smaller considering that 5 nm is already of the order of the size of the air molecule. Therefore, assuming a constant flying height of

approximately 5 nm, as the trackwidth is reduced to satisfy the areal density increase, the

field generated at the location of the recording layer also decreases. Unfortunately, the field magnitude cannot be endlessly maintained via the reduction of the throat height. As

shown above, as the throat height becomes too small, the contribution to the recording


field from the magnetic “charges" on the tilted sidewalls increases. As a result, the cross-

track and trailing field gradients deteriorate [54]. Assuming the sidewall tilt angle is approximately 45 degree, the smallest value of the throat height at which the gradient

deterioration is less than 50 percent is approximately 100 nm. The recording field

generated at saturation under the center of a 300 nm thick trailing pole with a 100 nm throat height at a 5 nm flying height versus the distance across the track at 3 values of the

trackwidth, 25, 50, and 100 nm, is shown in Figure 50a. For example, at a 1Tbit/in2

density, the trackwidth is approximately 50 nm assuming a 4:1 BAR. The signal half-width defined as the distance along the track, at which the signal drops twice from its

maximum value, versus the trackwidth is shown in Figure 50b. It can be noted that as the

trackwidth becomes narrower than approximately 50 nm, the half-width ceases to strongly depend on the trackwidth. This is explained by the fact that as the trackwidth is

reduced below this critical value, the half-width is dominantly determined by the doubled separation between the ABS and the SUL along with the flying height, which, in this

case, are 20 and 5 nm, respectively. Also, as the trackwidth is reduced below

approximately 50 nm, the field magnitude drastically decreases for two reasons: 1) the field generated by an individual SPH drops as the trackwidth is reduced because the net

magnetic charge is reduced, and 2) the contribution of the field generated by the image

SPH drops faster with the trackwidth reduction because it is effectively further away from the center of the recording layer, as compared to the real SPH.

0 20 40 60 80 100

0.0

0.5

1.0

1.5

HW/2

10

T = 300 nm

2550 nm

W = 100 nm

Hz /

2M

s

Distance across the track (nm)

(a)


0 20 40 60 80 100

0

20

40

60

80

100

Vert

ical F

ield

Halfw

idth

(nm

)

Trackwidth (nm)

(b)

Figure 50. (a) The vertical field versus the distance across the track at saturation for a SPH with a 300 nm

thickness at 3 values of the trackwidth, 25, 50, and 100 nm. (b) The field half-width versus the trackwidth.

In summary, the main question regarding the write head at areal densities of the order of

1Tbit/in2 can be formulated: "How to maintain the recording field magnitude with the

reduction of the bit dimensions without deteriorating the field gradients?"

2.2.16. Multiple Magnetic Image Reflection

In a perpendicular system of the 2nd type, recording is produced by the trailing edge of

the trailing pole (TP) of a single pole head (SPH), as shown in Figure 51. The recording

field is controlled by the electrical current in a coil wrapped around the TP.

However, additional field sources contribute to the net recording field under the TP.

Previously, it was shown that the additional sources are due to the magnetic “charge” in a

recording medium, which, unlike in longitudinal recording, is concentrated not in the

transitions, but rather more uniformly distributed at the top and effective (due to the

presence of the SUL) bottom sides of the recording layer. Among these sources is the

field generated by tracks adjacent to the main track under the TP. From the field

superposition principle, the maximum field in this case is less than 2 Ms, where Ms is the

saturation moment of the recording layer. This effect exists in both longitudinal and

perpendicular recording leading to non-linear transition shift (NLTS).


Trailing

edge

Recording Layer

Leading Pole

(LP) TP

H

IWrite

coil

SUL

Figure 51. A schematic diagram showing how due to the multiple reflection a relatively small field under the

leading pole can be magnified into a relatively strong field under the trailing pole.

There is an additional effect inherent only to perpendicular recording with a SUL, which

contributes to the net field under the TP. The magnetic flux due to a bit-pattern in the

recording layer can be transferred from the leading pole to the trailing pole, as shown in

Figure 52. Although indirect, this effect is capable of generating a relatively large

additional field under the TP, as shown below. As described below, the process

underlying this effect can be explained in terms of the multiple (magnetic) image

reflection (MIR) of the surface magnetic charges in the recording layer, sandwiched

between two magnetic “mirrors”, the soft underlayer and the leading pole.

LP

SUL

TP

Recording Layer

Non-zero net

magnetization

Figure 52. A schematic diagram showing the origin of the MIR effect.


The intention of this Section is to utilize 3-D boundary element modeling (BEM)

supported by spin-stand and magnetic force microscopy (MFM) experiments to

investigate in detail the dependence of various parameters on the MIR effect.

In the absence of two magnetic mirrors above and below the recording layer, the stray

field emanating from a DC magnetized recording layer is negligibly small. The net stray

field is a sum of the oppositely directed fields generated by the top and bottom surface

“charge” of the recording layer. In other words, the magnetic field is “trapped” inside the

recording layer, as shown in Figure 53.

Next, the effect of the presence of the soft underlayer and the leading pole can be

analyzed. As earlier described, the soft underlayer acts as a mirror that creates an image

of the surface charge in the recording and thus increases the effective separation between

the effective bottom and top charge in the recording layer. The leading pole is the second

magnetic mirror added to the opposite side of the recording layer. This second mirror

creates another set of surface charge images and further increases the effective separation

between the effective bottom and top charge. The following analogy could illustrate the

rest in this process. Imagine yourself standing between two facing each other mirrors.

Ideally, due to the multiplicative reflection you should be able to see an infinite number

of images of yourself. Similarly, this multiplicative process leads to the effective

substantial separation of the surface charges from each other and thus, releases non-zero

magnetic flux (See Figure 53).

Ideally, assuming the head/medium system to be a 100% efficient magnetic flux guide,

i.e., with no flux leakage, according to the magnetic flux conservation, assuming a DC-

magnetized medium under the leading pole, the magnitude of the additional field,

Haddition, generated under the trailing pole is expected to be directly proportional to the

net magnetic moment of the recording layer, Ms,

Haddition ~ 4 Ms ALP/ ATP , (7)

where ALP and ATP are the ABS areas of the leading and trailing poles, respectively. The

linear dependence on the Ms and the ratio ALP/ ATP is valid as long as no saturation

occurs in the system. The linear dependence on the ratio ALP/ ATP becomes a crude

approximation when the trackwidth and, consequently, the area, ATP, is reduced down to

a size, at which the efficiency of the system starts to drop. Previous calculations indicate

that for a given perpendicular system configuration the efficiency drops to values less

than 60 percent as the trackwidth becomes narrower than approximately 300 nm. At

such narrow trackwidths, significant amount of the magnetic flux generated in the


region under the leading pole due to the MIR effect leaks out on its way to the trailing

pole.

+ + + + + + + +

- - - - - - - - - - -

+ =

+ + + + + + + +

- - - - - - - - - - - -

Soft Magnetic Material

SUL

Soft Magnetic Material

LP

Magnetic “charges” in the

recording layer

Magnetic “charges” moved to

due to multiple magnetic reflections

Magnetic “charges” moved to

due to multiple magnetic reflections

No magnetic flux gets out

DC Magnetized

Recording Layer

Magnetic Flux

Magnetic Flux

Figure 53. A schematic diagram helping understand the difference between two cases: a) magnetic flux trapped

inside the recording layer and b) the magnetic flux extracted from the recording layer region located between

SUL and LP due to the MIR effect.

Calculations indicate that in the presence of a SUL, a focused ion beam (FIB) trimmed

single pole head with a 300 nm trackwidth and a leading pole’s ABS cross-section of

2 m x 5 m is capable of generating an extra field under the trailing pole of up to

approximately 2000 Oe due to a DC magnetized recording medium with a Ms of 400

emu/cc located between the SUL and the LP. If a recording medium possesses relatively

low coercivity, i.e. less than ~ 2000 Oe, such a large additional field itself is sufficient to

entirely erase a previously recorded track. The following experiment was performed for

observing the effect. A double-layer CoCr based perpendicular medium with coercivity

of ~1800 Oe with a remanent squareness of 0.7 was preliminarily DC-erased in the

presence of an external vibrating sample magnetometer’s (VSM) field of approximately

3 T.

The SPH was run across the medium at the condition of zero current in the write coil. A

MFM image of a track recorded as a result of the described experiment is shown in

Figure 54a. Because the medium was preliminarily DC-erased and the write current was

zero, the only contribution to the net recording field could be due to the above described

effect of the extraction of the magnetic flux due to the MIR effect in the region under the


leading pole. For comparison, a MFM track recorded using the same head, now by

applying an alternating write current of 200 mA turn, is shown in Figure 54b.

(a)

(b)

Figure 54. MFM images of tracks recorded onto a CoCr based perpendicular medium with coercivity of

approximately 1800 Oe with a recording field (a) due to the effect of MIR to relocate the magnetic flux from the

region under the leading pole into region under the trailing pole and (b) by an alternating write current.

To study effects of MIR on recording, the following parameters were chosen: TP

trackwidth, W2, = 300 nm, TP thickness, T2, = 500 nm, throat height, TH, = 500 nm, LP

width, W1, = 2 , LP thickness, T1, = 2 m, gap length = 1 m, head magnetic moment,

Bs, = 2 T, recording layer thickness, t, = 20 nm, recording layer moment, Ms, = 200

emu/cc, and ABS to SUL separation = 30 nm.

Considering a DC-erased medium under the LP with a 2 x 2 cross-section, the

perpendicular component of the field generated under the TP at zero write current at 3

different values of the TP trackwidth, W2 = 100 nm, 300 nm and 500 nm, versus the

distance along the central line down the track at a 5 nm flying height is shown in Figure

55a. The reason why the field does not strongly depend on the trackwidth in this range is

the existence of the two essentially competing effects. On one hand, the field should

increase with the trackwidth reduction according to the Flux Conservation Law, as

described above. On the other hand, less flux reaches the TP as the trackwidth is


reduced, because the efficiency of the system drops with the trackwidth reduction in this

particular range. In any case, it can be seen that recording layer magnetization of only

200 emu/cc is sufficient to generate an extra field of more than 1000 Oe just due to the

MIR effect.

0.0 0.2 0.4 0.6 0.8

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

Hz (k

Oe)


W = 100 nm

W = 300 nm

W = 500 nm

0.0 0.2 0.4 0.6 0.8

0.00

0.25

0.50

0.75

1.00

1.25

1.50

1.75


1000 x 500

1000 x 1000

1400 x 1400

2000 x 2000

6000 x 2000

(a) (b)

Figure 55. Recording field under the TP due to the effect of MIR of the DC-erased recording layer between the

LP and SUL at different values of (a) TP widths, 100, 300 and 500 nm, and (b) at different cross-sections of the

LP.

The perpendicular recording field under the TP at a 5 nm flying height versus the

distance down the track is shown for a set of different values of the LP width and

thickness (W1xT1 nm2) in Figure 55b. It can be noted that the field depends on the net

ABS area of the LP, A = W1 x T1, rather than independently on the LP width and

thickness as long as the non-zero magnetization regions under the LP are relatively not

far from the location of the main track defined by the TP. Moreover, the dependence on

the LP area is rather significant in comparison with the dependence on the TP area. This

is explained by the relative proximity of the LP to the main source of the field, the region

at which the MIR effect occurs, along with relatively large cross-section dimensions of

the LP, as compared to the ABS-to-SUL separation. It can be noted that for fairly realistic

LP’s cross-section dimensions, e.g. such as 6 x 2 2, a recording field of as high as 1400

Oe could be generated under the TP, as shown in Figure 55b.

The important question to answer is: “What is more important: a particular bit pattern or

the net (average) magnetic moment under the leading pole?” At an initial condition of

zero net magnetic moment (ac-demagnetized medium) under the LP, a 500 nm wide DC-

erased track was modeled to be recorded along the central line under the LP along with

two 250 nm wide tracks located 250 nm away from the main track at each side with the

magnetization directed opposite to the direction of the magnetization in the main track, as


shown in Figure 56. As a result of this track pattern, the net magnetic moment under the

leading pole is still zero.

LP

SULAC-demagnetized

backgroundDC-magnetized

tracks

500 250 250

Figure 56. A front view cross-section diagram showing 3 DC-tracks with respect to the LP.

The magnetic field generated under the TP as a result of the described track pattern is

shown in Figure 57a. For comparison, the field generated at a condition of an entirely

DC-erased medium is also shown on the same plot. It could be observed that in the case

of zero net magnetic moment the generated field is negligibly small. Indirectly, this

result indicates that the dependence on the track location is not significant at least for the

considered off-center region of the order of 1 m, otherwise, the opposite polarity

magnetic flux currents under the LP would not have been able to cancel each other, and

therefore, the net field under the TP could not have been as small as it is. In another

modeled scenario, the location of a 500 nm wide track on the background of an AC-

erased medium was varied. The perpendicular field under the leading pole generated

when the track under the leading pole is located along the central line and 500 nm away

from the central line is shown in Figure 57b. It could be observed that the field for the

both cases is very similar. For the both cases, the track is entirely covered by the leading

pole. It should be noted that when the track was located outside the area covered by the

leading pole, the field under the leading pole was found to become negligibly small. For

reference, the dependence of the perpendicular field component versus the net

magnetization in the recording layer is shown in Figure 58.


0.0 0.2 0.4 0.6 0.8

0

2

4

6

8

10

12


DC-erased Track

Compensated Bit Pattern

0.00 0.25 0.50 0.75

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

HZ (

kO

e)


Centered Track

Off side Track

(a) (b)

Figure 57. Stray field versus areal density at a 200 ktpi at the cetner and at 5 percent off the edte of a bit at a 5

nm flying height.

0.0 0.5 1.00.0

0.5

1.0

1.5

Mave

(4 Ms)

Hz (

kO

e)

Figure 58. Vertical field under the TP due to the MIR effect for a 2 m x 6 m LP cross-section versus

normalized net magnetization in the recording layer.

In conclusion, the phenomenon of the MIR along with the physics of the magnetic flux

propagation explains why by making the leading pole sufficiently large it is possible to

generate an unacceptably large magnetic field under the trailing pole even at zero current

in the write coil.

2.3. MODIFIED FIRST PERPENDICULAR MODE: A SHIELDED SINGLE POLE

HEAD AND A PERPENDICULAR MEDIUM WITH A SOFT UNDERLAYER


2.3.1. Shielded Single Pole HeadOne of the previously proposed solutions is to build "soft" magnetic shields around the

main pole, as shown in Figure 59 [76, 77].

X

Y

G (Gap)SS

Main Pole

Shield

Figure 59. A diagram of the ABS view of a shielded single pole head (SSPH) pole tip configuration.

It can be noted that the shields are wrapped only around the trailing side and the two

cross-track sides of the main pole. Only these three sides are critical for recording,

because the two cross-track sides define the trackwidth and the trailing side defines the quality of each linear transition. No recording is supposed to take place at the leading

side, therefore, this side does not necessarily have to be covered with a shield. The direct

effect of the shielding is screening the unfavorable side field away from the recording medium, as shown in a cross track cross-section diagram in Figure 60. Consequently, the

constraints on the head structure, which were put on the regular SPH (without shielding)

for the purpose of reducing the effect of the side field, are substantially relaxed if shields are utilized. It should be reminded that for the regular SPH, the pole tip geometry is

chosen with a fairly large throat height with the purpose to reduce the side field. The cost of the fairly tall throat is the substantial reduction of the field magnitude and the system

efficiency, as shown above. On the contrary, for the case with shields, the throat height

can be substantially reduced for maintaining the fairly large field magnitude without loosing the field gradients. In other words, if shields are used, a substantially more

efficient pole structure can be implemented without loosing the field gradient. As an

example, the calculations were made to compare the recording field generated by a regular SPH with a throat height, TH, of 100 nm with the recording field generated by a

shielded SPH (SSPH) with a 50 nm throat height with a cross-track shield to shield

separation of 90 nm and a downtrack gap, G, between the write pole and the trailing


shield of 20 nm. The shield throat height, STH, was modeled to be 10 nm. In both cases,

the pole tip was modeled to be 50 nm wide and 300 nm thick. The central cross-track profiles for these two cases at saturation are shown in Figure 61. In practice, however,

there might be limitations due to the processing difficulties. For example, the shortest

possible today throat height is going to be dictated by the lapping process accuracy.

As a direct consequence of the ability to exploit a more efficient pole tip configuration,

the improved skew angle performance of SSPH can be mentioned. Due to the higher efficiency, as compared to SPH, a much thinner pole tip can be utilized to generate the

same recording field. Therefore, SSPH has substantially improved skew angle

performance, as compared to that of SPH, as discussed below.

SUL

Hrec

Hside

(a)

TH

SUL

Hrec

Shield ShieldRecording layer

(b)

STH

Figure 60. Diagrams showing the magnetic field propagation for the two cases of interest: (a) without and (b)

with shields.

For the SSPH design, the side cross-track and trailing field gradients are dominantly determined by the spacing between the main pole and the shields. This is in contrast with

the regular not-shielded SPH design, for which the gradients are determined not only by

the flying height and the separation between the ABS and the SUL but also to significant degree by the throat height. Evidently, the deadly limitation of a nonzero throat height in

the case of the regular SPH design is automatically removed in the case of the SSPH

design. In the latter case, even for a substantially shorter throat height, the undesired side cross-track and trailing fields are reduced due to the existence of a relatively low-

reluctance and well-defined return flux path via the shields.

As noted above, the reduction of the throat height to zero drastically increases the system

efficiency and allows a substantially larger amount of the magnetic flux generated by the

drive coil to reach the ABS. This automatically results in an improved skew angle performance of the SSPH design, as compared to the conventional SPH design, because

in this case a head with a substantially thinner pole tip can be utilized to generate a field

as strong as the field generated by an equivalent conventional SPH with a much thicker


main pole tip. As discussed above, the skew angle sensitivity is proportional to the pole

tip thickness.

0.0 0.1 0.2 0.3

0.0

0.5

1.0

1.5

W = 100 nmShielded

Regular

Hz (

10 x

kO

e)


Figure 61. Cross-track profiles for the two cases, a regular SPH and a shielded SPH (SSPH), at saturation.

The maximum trailing field at saturation versus the main pole thickness for the two cases,

a regular SPH with a 500 nm throat height and SSPH with a zero throat height, both with

a 120 nm trackwidth, is shown in Figure 62. To clearly illustrate the point, the field is shown normalized to its saturation value. It can be noted that indeed the field starts to

drop at a smaller value of the thickness for the case with shields.

0 200 400 600

0.0

0.2

0.4

0.6

0.8

1.0

SSPH

Regular SPH

Hz m

ax/H

z m

ax s

at

Thickness (nm)

Figure 62. The maximum trailing field at saturation (normalized to its saturation value) versus the pole tip

thickness for the two cases, a regular SPH with a 500 nm throat height and a shielded SPH with a zero throat

height. The trackwidth is 120 nm.

Another observation that can be made is the fact that if shields around the main pole are utilized, as described above, there is absolutely no need for the return pole separated with


a fairly large gap from the leading edge of the main pole, as shown in Figure 63a. The

shields wrapped around the main pole not only act as the gradient shapers but also perform the role of a return pole. As a result, a system with shields around the main pole

and without a return pole remains approximately as efficient as a regular system with a

return pole. As a consequence of the shields acting also as a flux return pole, the requirements on the use of a SUL are much less tight as compared to the regular SPH

case. It can be also observed that the shielded structure resembles the typical ring head

structure. The purpose of the separation between shields and the main pole is to avoid the side field and thus to distinctly define the recording transitions. Similarly, the purpose of

the gap between the two poles of the ring head structure is to define the recording

transitions. Moreover, similar to a system with a ring head, a system utilizing a shielded writer can be utilized without any SUL at all. As shown in Figure 63b, the fairly small

separation between the main pole and the shield provides sufficient efficiency. In most implementations, shields are coupled to the main pole through the back of the pole, as

shown in Figure 63c. In this case, the trailing and cross-track side field gradients are

determined by the flying height and the separation between the main pole and the shields rather than by the separation between the ABS and the SUL.

B

Return

pole

Main

pole

(a)

Main

poleShield

B

SUL

Shield

(b)

Main

pole

Shield

B

SUL

(c)

Figure 63. Diagrams showing the flux return paths in cases with a regular (a) SPH (along-track cross-section

view) and (b) a shielded SPH (SSPH) (front view cross-section), respectively. (c) Full-scale front view cross-

section of SSPH.


In fact, for the considered dimensions, the only noticeable difference between the two modified systems, with and without a SUL, is the fact that the system without a SUL

needs approximately 20 percent more current to saturate, as shown in the calculated

current dependencies in Figure 64. To clearly illustrate the main dependence, the field is shown normalized to its saturation value.

In summary, it can be concluded that the utilization of soft magnetic shields around the main pole results in the following advantages:

1)The recording field can be maintained to be fairly large (compared to the case without shields) with no substantial field gradient degradation at higher areal

densities.2)The field gradient can be controlled via varying the separation between the main

pole and each of the shields as well as by the pole tip and shield geometry.

3) If shields are used, the system efficiency could be increased through the reduction of the throat height. As a consequence, the skew angle sensitivity could be also

substantially reduced.

0 50 100 150 200

0.0

0.5

1.0

with NO SUL

with SUL

Hz m

ax/ H

z m

ax s

at

Drive current (au)

Figure 64. The maximum field versus the drive current for two configurations: SSPH with and without a SUL.

The field is shown normalized to its saturation value.

Chapter 3 Physics of Playback

85

Chapter 3

Physics of Playback

1. Introduction

Recent high areal density demonstrations of perpendicular recording clearly indicate an

increasing interest in this technology [78,79,80,81]. It is believed that, as compared to

conventional longitudinal recording, perpendicular recording is capable of deferring the superparamagnetic limit to a significantly higher areal density due to a thicker recording

layer and/or the use of a soft underlayer (SUL) [82]. Although perpendicular recording is

certainly the closest alternative to the conventional technology, its novelty also brings up new issues, not ever encountered in longitudinal recording. These issues have to be well

understood before the technology can be fully and most efficiently implemented

[83,84,85]. Major issues related to perpendicular media and perpendicular write heads have been previously considered [86,87,88,89]. However, relatively little attention has

been given to the playback process. For example, the role of the SUL in the playback

process is still an open question: although the SUL certainly increases the magnitude of the playback signal, its influence on the signal resolution is still controversial. Another

fundamental source of the difference between the playback processes in longitudinal and perpendicular recording is the difference in the magnetic “charge” configuration in

longitudinal and perpendicular media, respectively. Therefore, the intention of this

Chapter is to investigate the physics of the playback process in perpendicular recording.

1.1. CHAPTER OVERVIEW

In this Chapter, the physics of the playback process in perpendicular recording is

explored. It is shown that due to the existence of the two layers of the “magnetic charge”,

at the top and effective bottom surfaces of the recording layer, the stray field sensed by a reader rolls off with the areal density essentially differently than it does in longitudinal

recording. Unlike in longitudinal recording, in perpendicular recording, the recording

layer thickness is an extremely sensitive parameter, which provides extra flexibility in controlling the density roll-off. It is illustrated that for areal densities beyond


approximately 200 Gbit/in2, the slowest roll-off for both longitudinal and perpendicular

recording occurs at a bit aspect ratio of 1:1. A fundamental role of the soft underlayer in the playback process is investigated. It is illustrated that although at relatively low linear

and track densities the use of a soft underlayer increases the playback signal, the signal

does not depend on the use of a soft underlayer at high densities. It is shown that for both perpendicular modes, although at sufficiently low track densities (below ~ 50 ktpi), the

signal disappears at relatively low linear densities, there is a significant non-zero signal

even at zero linear density if the track density is sufficiently high (above ~ 300 ktpi). A magnetic image model is introduced to illustrate that the use of a soft underlayer could

not improve the resolution of a recording system. Moreover, it is shown that there is

range of the air-bearing-surface-to-soft-underlayer separation in which the playback resolution deteriorates. The guidelines are given on how to design a playback head to

avoid the operation in the region of the deteriorated resolution. Besides the conventional playback head design including a single read element surrounded by two soft shields

along the track, a number of other optimized for perpendicular recording designs are

explored. For example, it is illustrated that compared to conventional single-read-element shielded configurations, differential reader configurations display superior playback

properties in terms of both the playback amplitude and the spatial resolution.

2. Playback in Perpendicular Recording

2.1. ANALYSIS METHODS

In this Chapter, the playback process is analyzed as not just a detailed study of another read head design but rather an integral process, including a reader and a medium. Such an

integral consideration is especially critical for the perpendicular mode with a medium

with a soft underlayer (SUL). In this mode, the SUL is often viewed as an indispensable part of the recording head [85].

For broader and more insightful comprehension of the playback process in perpendicular recording, two analytic approaches, direct and reciprocity, are considered. The “direct

calculation” method addresses exclusively the fundamental contribution of a recording medium into the playback process. Thus, the fundamental issues related to the different

“charge” configuration in a perpendicular medium could be more explicitly studied. As

to the “reciprocity calculation,” it reflects the magnetic properties of the playback head.

2.1.1. Direct Calculation with Point-size Reader Approximation

To calculate the magnetic field emanating from a perpendicular recording medium with a periodically written bit pattern, an analytical 3D expression could be derived, as shown in

Equation 1 [90]. This expression takes into account the recording layer effective

thickness, (twice the physical thickness if a SUL is used), the hard layer saturation magnetization, Ms, and the bit length and width, a and b, respectively. The origin of the reference coordinate system is chosen to be located at a corner of a bit at the top surface

of the recording layer, as shown in Figure 1.

Chapter 3 Physics of Playback 87

Z

Y

X

(0,0,0)

b

a

Periodic bits

Figure 1. A diagram showing the location of the reference system with respect to a “effective” recording layer

with a periodicall written bit pattern.

To study the dependence on the bit aspect ratio (BAR) and the areal density (AD), the

transformation equations representing the bit length and width, a and b, through BAR and AD, could be used, as given by Equations 2 and 3, respectively.

0,1

1 1

sinsin32

2222

zz

oddn

oddk

yb

kx

a

n

kn

sMstrayzH

b

k

a

n

b

k

a

n

(1)

AD

BARa , (2)

ADBARb

1, (3)

In the “direct calculation” approximation, no head finite size effects are taken into account. Therefore, the “direct calculation” reflects only the contribution of a medium to

the playback process. Also, considering that the effect of the use of an “ideal” SUL on the


stray field is equivalent to a two-fold increase of the recording layer thickness, the same

expressions could be used also to model the case with a SUL just by replacing with 2 .However, it should be remembered, that the use of the SUL is not necessarily equivalent to the two-fold increase of the recording layer thickness in the sense of the energy, and,

therefore, this approximation could not be applied to predict phenomena associated with

the bit energy, e.g. the thermal instability effect [87].

+

+


+

+

+ + + + + + + + - - - - - - - - - -

- - - - - - - - - - + + + + + + +

Hstray

Hstray

+ + + + + + + + - - - - - - - - - -

- - - - - - - - - - + + + + + + +

Underlayer

boundaryMedium

image

M(a)

(b)

(c)

Figure 2. Diagrams showing the sources of stray fields in the case of (a) longitudinal recording, and

perpendicular recording (b) without and (c) with a SUL.

To help understand the basic difference in the playback process between longitudinal and perpendicular recording, schematic diagrams of the stray fields emanating from a single

transition in a longitudinal medium and a perpendicular media without and with a SUL

are shown in Figures 2a-c, respectively [91]. As can be noticed, in the longitudinal case, the stray fields emanate only from the transitions, with the fields near the transitions

oriented perpendicular to the disk plane. On the other hand, in the perpendicular cases,

the stray field emanates from the effective “magnetic charge” at the top and effective (due to the SUL) bottom surfaces of the recording layer, with the field right above the

transitions oriented parallel to the disk plane. The calculated stray fields for two values of

the recording layer thickness, 10 and 20 nm, with a Ms of 200 emu/cc, above an isolated magnetization transition at a 5 nm flying height are shown for the three cases in Figures

3a-c, respectively. In these calculations, the transition is assumed to be ideal. Also, in this example, for the description simplicity, infinitely wide tracks are assumed. Below in this


Chapter it will be shown that consideration of a finite trackwidth is critical to predict

realistic waveforms. To calculate the field in a longitudinal recording system, three-dimensional (3-D)-boundary element modeling (BEM) was exploited [92]. It should be

noted that, in general, the stray field in the perpendicular cases looks similar to the stray

field in the longitudinal case, provided that the perpendicular and in-plane components are interchanged. It is obvious that the effective distance away from the transitions, at

which most of the drop in the stray field occurs, is determined by the effective recording

layer thickness.

-0.10 -0.05 0.00 0.05 0.10

-500

0

500

1000

Hx and 10nm

Hx and 20nm

Hz and 10nm

Hz and 20nmLongitudinal

Hx a

nd H

z (

Oe)

Distance along the track (um)-0.10 -0.05 0.00 0.05 0.10

-500

0

500

1000H

x and 10nm

Hz and 10nm

Hx and 20nm

Hz and 20nm

Perpendicular

No Su

Hx a

nd H

z (

Oe)


(a) (b)

-0.10 -0.05 0.00 0.05 0.10

-500

0

500

1000

1500

Hx and 10nm

Hz and 20nm

Hz and 10nm

Hx and 20nmPerpendicular

With SU

Hx a

nd H

z (

Oe)


(c)

Figure 3. The along-the-track and perpendicular stray field components, Hx and Hz, versus the distance along

the track over a single transition in (a) a longitudinal medium and perpendicular media (b) without and (c) with

a SUL, with 10 and 20nm recording layer thickness values with a Ms of 200 emu/cc at a 5 nm flying height.

For example, because of the effective factor-of-two increase in the recording layer

thickness when the SUL is used, in the perpendicular case without a SUL the field drops more rapidly than it does in the case with a SUL. Below it is shown that the amount, by

which the field drops, is determined mostly by the trackwidth.


Before going into a more detailed study of the perpendicular stray field dependence on different parameters, such as the recording layer thickness, BAR, and others, it is helpful

to clearly understand the basic physics behind the origin of the stray field in

perpendicular recording. As previously mentioned (see Figures 2b and c), the net stray field in perpendicular recording consists of the oppositely directed fields generated by the

top and effective bottom “charges” of the recording layer. Therefore, because the net

stray field is the difference between the two fields, it is relatively strongly sensitive to the bit dimensions and the effective recording layer thickness. As an example, the field

profiles emanating from the top and bottom “charge,” as well as the net stray field at a 5

nm flying height for a 10 nm thick recording layer with a periodically written bit pattern with a 500 x 500 nm2 bit cell cross-section is shown in Figure 4a. The magnitude of the

stray field is normalized to 4 Ms, where Ms is the recording layer magnetization. It is obvious that for a relatively wide and long bit (as compared to the flying height) and if

the recording layer thickness is significantly smaller than each of the bit cell sizes, the magnitudes of the two fields over the center of a bit should be approximately equal to the

field from a uniformly “charged” plane (with a “charge” density of 4 Ms), i.e. to 2 Ms.Thus, in this case, being equal in magnitude and oppositely directed, the two fields

substantially cancel each other, so the net stray field should be relatively small at the center of a bit. On the contrary, for sufficiently small bits, as compared to the flying

height, although each of the fields, from the top and bottom “charges,” is less than 2 Ms,the fields are also more different from each other in their strengths, thus the net stray field

might not be negligible, as shown for a 50 x 50 nm2 bit cell cross-section in Figure 4b. By

changing the recording layer thickness from a relatively small to a sufficiently large, the contribution of the field from the effective bottom “charge” is changed from equal to

negligible compared to the contribution from the top “charge”. As a result, the net stray field at the center of a bit changes from zero to the field due to only the top “charge,” as

shown in Figure 4c for four different values of the square bit side, 25, 50, 200 and

500nm. It can be noted that the characteristic length, at which the field reaches its saturation value, is more sensitive to the bit length as the bit length becomes comparable

to the recording layer thickness. So far, the calculation has involved only the stray field

at the center of a bit. The field cancellation effect (FCE) due to the top and bottom “charge” is quantitatively different if the field is considered closer to one of the edges of

a bit. It is worth mentioning that, in general, the FCE is the cause of a typically observed

maximum value in perpendicular roll-off curves [84-86]. At sufficiently low densities, the net signal is small because of the FCE, while at sufficiently high densities, the net

signal is naturally small because the bit area of the top side of the recording layer

(containing the field generating “charge”) becomes fairly small. To illustrate how the FCE depends on the position with respect to the bit cell, the stray field versus the linear

density at a 200 ktpi track density at a 5 nm flying height at the center and five percent

(relative to the bit length) away from one of the edges of a bit is shown in Figure 5.


(a)

0 100 200 300 400 500-0.6

-0.4

-0.2

0.0

0.2

0.4

0.6

Net field

Top surface

Bottom surfaceH

z/4

Ms

Central line along the track (nm)

(b)

0 10 20 30 40 50-0.4

-0.2

0.0

0.2

0.4

Net field

Top surface

Bottom surface

Hz/4

Ms

Central line along the track (nm)

(c)

0 200 400 600

0.0

0.1

0.2

0.3

0.4

0.5

25 x 25

50 x 50

200 x 200

500 x 500

Hz/4

Ms

Recording layer thickness (nm)

Figure 4. The field profiles at a 5 nm flying height due the magnetic “charges” at the top and bottom surfaces

and the sum of the two fields from a 10 nm thick recording layer with a periodically written bit pattern with (a)

a 500x500nm2 and (b) 50x50nm2 bit cell cross-section. (c) The net stray field (at the bit center) at a 5 nm flying

height versus the thickness of the recording layer at different bit-cell cross-sections.


It can be noticed that the perpendicular field above the center peaks up at a significantly

higher density (~600kfci) than it does above the edge of a bit (~100 kfci). In the following Reciprocity Calculation Section, the existence of the maximum in a

perpendicular roll-off curve is explained with the help of a different model.

0 250 500 750 10000.0

0.1

0.2

0.3

At the Center

At the Edge

Norm

aliz

ed H

z


Figure 5. The stray field versus the areal density at a 200 ktpi at the cetner and at 5 percent off the edge of a

bit at a 5 nm flying height.

To explain the dependence of the signal on the trackwidth, the calculated normalized

isolated transition response at a 5 nm flying height for a 20nm thick recording layer without a SUL (equivalent to a 10nm thick layer with a SUL) at three values of the

recorded trackwidths, 80, 200, and 1000nm, provided the adjacent tracks are ac-

demagnetized (in other words, average magnetization is zero for the adjacent tracks), is shown in Fig. 6. Assuming an ideal transition, the perpendicular stray field could be

calculated via straightforward integration of the field produced by the point “charge” uniformly distributed within an individual track with a trackwidth of W and a recording

layer thickness of with the transition at X=0. It can be observed that the narrower the track is the smaller the amount of the field, which is lost away from the transition, is. This

is in agreement with the above-described field cancellation effect, according to which the

fields from the top and bottom “charges” essentially cancel each other only for a sufficiently wide track sufficiently far away from the transition.

Going back to the calculation of the field due to a periodic bit pattern (see Eqs. 1-3), the stray field emanating from the center of a bit in a periodically written bit pattern at a 5 nm

flying height, versus the linear density for media with three values of the recording layer


thickness, 10, 20 and 40nm, is shown for two values of the track density, 50 and 316ktpi,

in Figs. 7a and b, respectively. Considering that track densities of 316 and 50ktpi correspond to track pitches of approximately 80 and 500nm, respectively, it can be

noticed that for the narrower trackpitch, the stray field does not disappear at a low linear

density. In fact, agreeing with the symmetry of perpendicular recording with respect to the along and across the track directions, 316ktpi at a linear density of 50kfci corresponds

to 50ktpi at a linear density of 316kfci.

-60 -40 -20 0 20 40 60-0.15

-0.10

-0.05

0.00

0.05

0.10

0.15

200nm

1000nm

80nm

Hz/4

Ms

Distance down the track (nm)

Figure 6. The isolated transition response for a 20nm thick recording layer without a SUL (or 10nm thick

recording layer with a SUL) at 3 values of the trackwidth, 80, 200, and 1000nm at the condition of ac-

demagnetized adjacent tracks.

To stress the significance of this result, it is worth remembering that in perpendicular recording, typically (at relatively large trackpitch values), it is expected that the stray

field drastically would drop at low densities [93]. The current calculations indicate that this non-desirable effect of the field reduction at low linear densities does not exist at a

sufficiently narrow trackpitch. Another consequence of the FCE, which can be observed

from the graphs, is the fact that, although the net stray field, as expected, increases as the recording layer thickness increases, at sufficiently high linear densities the stray field

does not strongly (compared to the low density case) depend on the thickness.


(a)

0 750 1500 2250 30000.0

0.1

0.2

0.3

0.4

316kfc

i

316

50ktpiH

z/4

Ms

Linear density (kfci)

10 (5nm with SUL)

20 (10nm with SUL)

40 (20nm with SUL)

(b)

0 750 1500 2250 30000.0

0.1

0.2

0.3

0.4

50kfc

i

50

316ktpi

Hz/4

Ms

Linear density (kfci)

10 (5nm with SUL)

20 (10nm with SUL)

40 (20nm with SUL)

Figure 7. The stray field above the center of a bit in 10, 20 and 40 nm thick recording layers at a 5 nm flying

height versus the linear density for two values of the track density, (a) 50 ktpi and (b) 316 ktpi.

Also, it can be noticed that, considering that the use of a SUL is equivalent to a two-fold increase of the recording layer thickness, the use of a SUL can noticeably increase the net

signal only at relatively low linear and track densities and has a much weaker effect at

relatively high densities. For example, considering that a 20nm thick recording layer


without a SUL is equivalent to a 10nm thick recording layers with a SUL, it could be

noted that at linear densities below approximately 50 kfci at a 50 ktpi track density, the use of a SUL increases the net stray field almost by a factor of 2. The effect is weaker at a

316 ktpi track density: in this case, the use of a SUL increases the net stray field at low

densities by approximately a factor of 1.5.

0 50 100 150 2000.0

0.1

0.2

0.3

0.4

8:1

4:12:11:1

Hz/4

Ms

Recording layer thickness (nm)

Figure 8. The stray field over the center of a bit at a 5 nm flying height versus the recording layer thickness at a

200 Gbit/in2 at 4 values of BAR, 1:1, 2:1, 4:1 and 8:1.

It is obvious that the BAR value is also going to influence the degree of the field

cancellation effect. For example, aiming at the next year projected areal density of 200

Gbit/in2, the net stray field 5 nm away from the center of a bit in periodic 200 Gbit/in2

patterns for a set of values of BAR, 1:1 (57x57nm2), 2:1 (80x40nm2), 4:1 (112x28nm2),

and 8:1 (160x20nm2), versus the recording layer thickness is shown in Fig. 8.

It gives an additional insight into understanding the difference between perpendicular and

longitudinal recording in general, if one compares the playback signal for different

recording modes in the point-size sensor approximation [94]. Although, from the above noticed 90 degree symmetry of the stray fields between perpendicular and longitudinal

recording, it follows that a conventional longitudinal reader is not optimal in

perpendicular recording, today perpendicular recording still relies on the use of the longitudinal reader configuration. Therefore, because the conventional reader

configuration is designed to be most sensitive to the perpendicular stray field component, it makes sense to compare the perpendicular stray field components for the three

recording modes [95].


(a)

0 500 1000 1500 20000

200

400

600

Areal density (Gbit/in2)

No SUL

Str

ay F

ield

, H

z (

Oe

)

Areal Density (Gbit/in2)

BAR = 1:1

BAR = 2:1

BAR = 4:1

BAR = 8:1

(b)

0 500 1000 1500 20000

200

400

600

800


with SUL

Str

ay F

ield

, H

z (

Oe

)


BAR = 1:1

BAR = 2:1

BAR = 4:1

BAR = 8:1

(c)

0 500 1000 1500 2000

200

400

600

800

Hz (

Oe)


BAR = 1:4

BAR = 1:2

BAR = 1:1

BAR = 2:1

BAR = 4:1

BAR = 8:1

Figure 9. The perpendicular stray field component versus the areal density for the equivalent perpendicular

systems with a 10 nm thick recording layer (a) without and (b) with a SUL at 4 values of BAR, 1:1, 2:1, 4:1 and

8:1, and (c) a 10 nm thick longitudinal recording medium at 6 values of BAR, 1:4, 1:2, 1:1, 2:1, 4:1 and 8:1.


The bit averaged (averaged over the bit area) zero-to-peak perpendicular stray field at a 5 nm flying height over a periodic bit pattern recorded onto a 10 nm thick recording layer

versus the areal density at different values of BAR are shown for the two perpendicular

modes, without and with an ideal SUL, in Figs 9a and b, respectively. The track averaged (averaged over the track) zero-to-peak perpendicular stray field at a 5 nm flying height

over the ideal transition in a periodic bit pattern recorded onto a 10 nm thick longitudinal

recording medium versus the areal density at different values of BAR is shown in Fig. 9c. It should be remembered that perpendicular recording in the sense of the stray field is

symmetric with respect to the directions along and across the track because the

magnetization in the recording medium is directed perpendicular to the plane of the disk. Therefore, the results for BAR values of 2:1 and 4:1 are identical to the results for BAR

values of 1:2 and 1:4, respectively. On the contrary, longitudinal recording has a preferred orientation in the disk plane because the magnetization is recorded along the

track. Therefore, in the longitudinal case, it makes sense to consider BAR values

corresponding to the complete range of values, i.e. 1:4 is not the same as 4:1. From these graphs, it definitely can be seen that in the perpendicular and longitudinal cases, the stray

fields change very differently with the areal density. In the perpendicular cases, a density

cross-over region can be identified, at which the stray field increase with the BAR increase changes to the stray field decrease with the BAR increase. This cross-over

density region is around 500 and 200 Gbit/in2 for the cases without and with a SUL. In

other words, the larger the BAR value is, the larger the signal is if the areal density is less than approximately 500 and 200 Gbit/in2 for the two cases, respectively. On the contrary,

if the areal density is larger than approximately 500 and 200 Gbit/in2 for the two cases,

the smaller the BAR value is, the larger the signal is, i.e., in this case, the maximum signal is reached at the BAR value of 1:1. No cross-over region can be noticed in the

longitudinal case. Moreover, in the longitudinal case, regardless of the density, the signal

reaches the maximum when the BAR value is approximately 1:1. Although, it is not trivial to equivalently compare the playback capability in longitudinal and perpendicular

recording, to get a feel for how fast the signal drops with the areal density, the normalized

graphs corresponding to the best longitudinal case with a BAR value of 1:1 and the perpendicular cases with the same BAR values are shown in Fig. 10. From the above

conclusions it is clear that the perpendicular modes are relatively more sensitive to the recording layer thickness, and, for the current assumptions of the calculations it can be

concluded that for both perpendicular modes, the signal drops slower with the areal

density than it does in the best equivalent longitudinal case. Between the two perpendicular cases, in the case without a SUL the signal rolls off slower with the areal

density than it does in the case with a SUL. The only relevant difference between these

two cases is in the effective thickness of the recording layer, which is twice as large for the case with a SUL. In summary, the slower density roll-off in perpendicular recording

could be explained by a relatively strong dependence of the stray field on the ratio

between the BAR value and the recording layer thickness due to the field cancellation effect described above. In longitudinal recording, the density roll-off is mostly

determined by the reduction of the stray field with the reduction of the amount of the

effective “charge,” which generates the perpendicular stray field at the location separated from the transition by a flying height.


0 500 1000 1500 2000

0.2

0.4

0.6

0.8

1.0H

z n

orm

aliz

ed


Longitudinal

Perpendicular with no SUL

Perpendicular with SUL

Figure 10. The normalized perpendicular stray field component for the best longitudinal case (with a BAR

value of 1:1) and the two perpendicular cases without and with a SUL with a 10 nm thick recording layer at a 5

nm flying height versus the areal density.

The difference between the two perpendicular cases, with and without a SUL, is

described below with the help of the Reciprocity principle [96]-[22]. As previously described, the Reciprocity Principle is a convenient theoretical technique, which allows

one to take into account the finite playback head dimensions [85].

2.1.2. Calculation Based on the Reciprocity Principle

Previously, it was shown that a linear field response of the head magnetic material is a

sufficiently good approximation for the playback process even at trackwidths as small as 60 nm [97, 85]. Therefore, the Reciprocity Principle can be used for calculating the

playback signal [98,99,100]. According to the Reciprocity Principle, the playback signal

is proportional to the convolution of the magnetization distribution in the recording layer and the so-called sensitivity field of the head. The sensitivity field is the magnetic field

generated by the imaginary unit currents in imaginary coils wrapped around the read

element [85]. Because the sensitivity field depends on the head configuration, the Reciprocity Principle, as mentioned above, takes into account the contribution of the head

on the playback signal. Because the sensitivity field also depends on the presence of a SUL, it is clear that the SUL should be treated as a part of the playback head. The

Reciprocity Principle for the playback signal is applied via Equation 4,


rHMSI

ˆ~ 1 , (4)

where M is the magnetization in the recording layer and H is the sensitivity field

generated by the imaginary current, I. Considering that the magnetization in the recording layer is in the perpendicular orientation with a negligibly small transition

parameter, from Equation 4 it follows that for understanding the basic playback

phenomena in perpendicular recording, it is sufficient to calculate only the perpendicular component of the sensitivity field [85].

SSS

t

Shield 1 Shield 2

SUL

ABS to SUL

Interlayer

Recording layer

interlayer

Fly height

Read

element

Figure 11. A schematic diagram showing a conventional longitudinal reader design incorporated in

perpendicular recording.

At this stage of the analysis, non-zero dimensions of the playback head can be taken into account. For simplicity, a regular longitudinal read head with a shielded read element was

assumed in this analysis. A schematic diagram of such a head is shown in Fig. 11. The

sensitivity field was calculated as the field due to the imaginary coils around the read element with actual head/media parameters, thus, the interaction between the read head

and the SUL was taken into account [85]. To model an ideal SUL, the boundary

conditions were defined so that the top surface of the SUL had a constant scalar magnetic

potential, , of zero.


200 400 600 800 1000

0

5

10

ABS-to-SUL = 30nmFlying Height = 5 nm

ideal SUL No SUL


Pla

ybac

k (

a.u.)

Figure 12. Roll-off curves for the perpendicular modes without and with a SUL, with a ABS-to-SUL

separation of 30 nm.

It should be mentioned that although the playback characteristics in general are sensitive to the playback head dimensions, such as the trackwidth, the shield-to-shield separation

(SSS) and others, some of the fundamental conclusions could be made through a study of

one specific head configuration. In the calculation described below, a 100-nm wide and 30-nm thick read element with a linear MH-response in the direction of the dominant

magnetic flux propagation and with a 100-nm SSS separation was assumed. The calculated linear density roll-off curves at a 5-nm flying height with a 30-nm

ABS-to-SUL separation without and with a SUL are shown in Fig. 12. It can be noticed

that at the low-density limit the signal is smaller than at the intermediate densities for the both cases. This is in agreement with the above-described field cancellation effect.

Another observation, which can be made from these graphs is the fact that although, at

intermediate densities, in the case with a SUL the signal is significantly (by approximately 40 percent) larger than the signal in the case without a SUL, the signals for

the two cases converge at linear densities above approximately 450 kfci. This is in

agreement with the previous conclusions stating that at sufficiently high linear densities the use of a SUL does not contribute to the net signal. Also, it can be noticed that the

signal maxima occur at approximately 380 and 290 kfci for the cases without and with a

SUL, respectively. The smaller linear density, at which the maximum occurs, for the case with a SUL is explained by the fact that the use of a SUL effectively doubles the

recording layer thickness, thus, the characteristic bit size, at which the influence of the

effective bottom “charged” layer can be neglected, is larger for the case with a SUL. The reason why these characteristic densities do not also differ by a factor of two is explained

by the signal dependence on the BAR, which is another variable in these two cases.


Although the convergence of the two curves at high linear densities indicates that the use

of a SUL does not necessarily improve the resolution of a recording system, the final conclusions about the influence of the SUL on the resolution are still not trivial to make

from these simple calculations. Therefore, another model, which allows to make more

direct conclusions, is described below.

Real

head

Image

head

Recording layer

Buffer layers etc.

Center of the recording layer

Image head to the center of the recording layer

Real head to the center of the recording layer

The underlayer

boundary line

Figure 13. A diagram showing the image representation of a system with a SUL.

For better understanding of the following material, first, the so-called magnetic image

model, which is used to explain the dependence of the sensitivity field on the use of a SUL, is described [94].

2.1.3. Model of the Magnetic Image To sufficiently good approximation, during the playback process, the SUL can be

assumed to be ideally soft, i.e. it can be represented as a semi-infinitely thick layer with

infinite magnetic permeability and an infinite moment. Therefore, the magnetic mirror image model can be applied [14]. In this model, the ideally SUL is replaced by a half-

space, which contains a mirror image of the recording head, as shown in a schematic diagram in Fig. 13. According to a theorem of differential equations [101], Laplace’s

Equation (a consequence of the Maxwell’s Equations, convenient to use for the

calculation of the magnetic fields) has an unambiguous solution provided sufficient boundary conditions are satisfied. Using an ideal SUL automatically provides the same

boundary conditions as in the case with a half-space with a mirror image, providing the

following rule is applied to the currents determining the magnetic structure of the mirror image head: the vertical components of the magnetic fields generated by the real and

image heads add up, while the in-plane components subtract from each other. Below, the

magnetic model is used to describe a “paradoxical” phenomenon caused by the use of a SUL.


(a)

10 15 20 25 3035

40

45

50

PW

50

ABS to SUL distance (nm)

With SUL

No SUL

(b)

10 15 20 25 300.5

0.6

0.7

0.8

0.9

1.0

Norm

aliz

ed S

ignal


With SUL

No SUL

Figure 14. (a) PW50 and (b) the normalized playback signal versus the ABS-to-SUL separation. An equivalent

dependence for the case without a SUL is shown by the dotted curve. The oval surrounds the “bad” points, for

which the PW50 values are substantially larger than the PW50 value at the infinite separation between the ABS

and the SUL (equivalent to no presense of the SUL at all).

Image “Paradox.” A diagram showing the location of the playback head and its image

due to the SUL with respect to the recording layer is shown in Fig. 13. It can be noticed

that due to the presence of the SUL, the finite thickness of the recording layer adds asymmetry into this system. The center of the recording layer is closer to the real head

than to the image head by the thickness of the recording layer plus all the bottom

interlayer (or buffer layers). According to the Reciprocity Principle, the resolution of the


final system is determined by the sensitivity field in the region of the recording layer.

Therefore, the final resolution of the system with a SUL is determined by the sensitivity field, which, in this case, is the sum of the fields generated by the real head and the image

head. It can be noticed that the resolution by the image head is worse than the resolution

of the real head because of the effective spacing loss due to the finite recording layer thickness plus the interlayer thicknesses. Therefore, the resolution of a system with a

SUL intrinsically can not be better than the resolution of a system without a SUL.

However, the signal does go up due to a SUL, because, according to the image model, there is an extra contribution to the net signal due to the image head.

(a)

10 15 20 25 3015

20

25

30

35

40

45

50

t = 30nm

t = 10nm

t = 5nm

SSS = 50nm

PW

50 (

nm

)


(b)

10 15 20 25 3015

20

25

30

35

40

45

50

SSS = 100nm

t = 30nm

t = 10nm

t = 5nm

PW

50 (

nm

)

ABS to Underlayer (nm)

Figure 15. PW50 versus the ABS to underlayer separation at three different values of the read element

thickness, 5, 10, and 30 nm, for a shield-to-shield separation (SSS) of (a) 50 nm and (b) 100 nm.


The calculated PW50 and the normalized amplitude versus the distance between the air bearing surface (ABS) and the SUL are shown in Figs. 14a and b, respectively. PW50

was defined as the half-width of the peak-to-peak signal [17]. In these calculations, the

variable parameter, contributing to the net change in the ABS-to-SUL separation, was the net thickness of all the interlayers used, providing the flying height and the recording

layer thickness remained constant, 5 nm and 10 nm, respectively. In other words, a

variation in the ABS-to-SUL separation is produced via the change of the bottom net interlayer thickness. Naturally, at this condition no variation either in PW50 or in the

amplitude is observed, if no SUL is used, as shown by the horizontal dotted lines in Figs.

14a and b. As a reader, a conventional design with a 30 nm thick read element and a 100 nm shield-to-shield separation (SSS) was modeled. The feature to notice is the fact

(which is expected from the image model argument above) that there is a range of the ABS-to-SUL separation, in which PW50 for the case with a SUL peaks up by

approximately 30 percent relative to the PW50 value for the case without a SUL. This

characteristic region (in this particular case, approximately between 10 to 25 nm) of the deteriorated resolution is determined by the read head dimensions and to some degree by

the flying height and the recording layer thickness. Also, it can be noticed that the signal

with the SUL is always larger than the signal without the SUL, which agrees with the magnetic image model argument above.

The PW50 versus the ABS-to-SUL separation for 3 different values of the read element thickness, 5, 10, and 30 nm, is shown for two different values of the SSS, 50 and 100 nm,

in Figs. 15a and b, respectively. It can be seen that with the reader thickness reduction

from 30 to 10 nm, the maximum PW50 change decreases as well (it should be remembered that with the reader thickness reduction the playback amplitude decreases as

well). Another feature to observe is the fact that, although the amount of the maximum

PW50 change does not strongly depend on the SSS, the range over which it changes is significantly smaller for the smaller value of the SSS. For example, for the both values of

the SSS, 50 and 100 nm, the maximum PW50 change of above 30 % is observed for a 30

nm thick reader. At the same time, the range of the PW50 change is approximately 5 to 10 nm and 5 to 30 nm for SSS values of 50 and 100 nm, respectively.

2.1.4. Examples of Reader Designs

In this Section, the Reciprocity Principle is applied to analyze and compare four different

reader designs (See Figure 16): a) unshielded reader [102]; b) shielded [103,104]; c) differential reader [105,106,107,108]; d) shielded differential reader[109]. Variations of

shielded, differential, and shielded differential readers with the emphasis on various

aspect of recording performance and manufacturability will be considered as well. Playback from a perpendicular recording medium with a soft underlayer and a single

layer longitudinal recording medium will be investigated. For completeness, more exotic

configurations such a longitudinal recording medium with a soft underlayer (keeper layer) and perpendicular recording medium without a soft underlayer will be considered

as well.


MR

Sensor

Recording Medium Recording Medium

Recording MediumRecording Medium

MR

Sensor

MR

Sensor

MR

Sensor

MR

Sensor

MR

Sensor

(a) (b)

(c) (d)

shield shield

shield shield

Figure 16. Schematics of various reader designs: a) unshielded reader, b) shielded reader, c) differential

reader, d) shielded differential reader.

As mentioned earlier, the reciprocity principle is used to study playback performance of

various readers [98,110,111]. According to the reciprocity principle (See Equation 4), the

playback voltage of a linear playback head is equal to the convolution of the sensitivity

field (function) of the reader with the magnetization pattern written into the recording

layer. The sensitivity field is calculated as the field generated by the read-head, in which

the read sensor is substituted with an equivalent soft magnetic material with a current

carrying imaginary coil wrapped around [112].

ShieldShield

Recording Layer

tMR

dMR

tshield

hMR

dMR-shield

tRL

FH

hshield

Figure 17. Schematic of a shielded differential reader with the relevant dimensions outlined.


The reader design parameters used in calculations are similar to the ones suggested by M.

Mallary et. al. [113] for a 1Terabit/in2 perpendicular recording system design. Unless specified otherwise, the magnetic thickness of an MR sensor, tMR, is assumed to be 10nm.

The cross-track width of an MR sensor, wMR, is 40nm. The height of the MR sensor, hMR,

is 40nm. The separation between the MR sensor and the shields, dMR-shield, and the separation between the two MR sensors in a differential designs, dMR, are set to be 10nm

each. The flight-height, FH, is 5nm and the media thickness, tRL, is 10nm. The shields

thickness, tshield, is 100nm, the shield cross-track width, wshield, is 400nm, and the shield height, hshield, is 220nm. The dimensions mentioned above are shown in a schematic

drawing of a shielded differential reader in Figure 17. In this chapter, the presence of an ideal soft underlayer is modeled with symmetric boundary conditions on the top surface

of the soft underlayer. Magnetic field modeling based on boundary element approach is

utilized throughout the chapter [92].

2.1.5. Basic Reader Design Comparison

The presented calculations of the playback are based on the Reciprocity Principle. The Reciprocity Principle requires the knowledge of the sensitivity functions for the playback

heads [99]. Figures 18 and 19 show the z (vertical) and x (horizontal) components of the

sensitivity fields along the track for the four types of heads, respectively.

-200 -100 0 100 200

-6

-4

-2

0

2

4

6

Hz (

a.u

.)

Distance along the track (nm)

No shield (sul)

Shield (sul)

Diff (sul)

Shield Diff (sul)

Figure 18. Vertical component, Hz, of the sensitivity field for different reader types.


-200 -100 0 100 200

-2

-1

0

1

2

3

4

5

Hx (

a.u

.)


No shield (no sul)

Shield (no sul)

Diff (no sul)

Shield Diff (no sul)

Figure 19. Horizontal along-the-track component, Hx, of the sensitivity field for different reader types.

The playback signal off a perpendicular recording medium with a soft underlayer versus the linear density for the four reader designs is shown in Figure 20. While conventional

shielded reader provides improved performance over unshielded reader, both differential

reader configurations offer a major performance improvement in terms of higher playback amplitude and higher resolution. The unshielded differential reader offers

higher signal amplitude at lower linear densities. The shielded differential reader offers

the highest spatial resolution out of all the four designs.

500 1000 1500 2000 2500

0

5

10

15

20

25

30

35

Pla

yback (

a.u

)


No shield (sul)

Shield (sul)

Diff (sul)

Shield Diff (sul)

Figure 20. Playback off a perpendicular recording medium with a soft underlayer versus linear density for four

reader designs.


The playback signal off a single layer longitudinal medium versus the linear density for the four reader designs is shown in Figure 21. Similarly to the case of perpendicular

recording presented above, the conventionally used shielded reader provides improved

performance over the unshielded reader. Both differential reader configurations offer a major performance improvement in terms of higher playback amplitude and higher

resolution over their non-differential counterparts. The unshielded differential reader

offers higher signal amplitude at lower linear densities. The shielded differential reader offers the highest spatial resolution out of all the four designs.

0 500 1000 1500 2000 2500

0

5

10

15

Pla

yback (

a.u

.)


No shield (no sul)

Shield (no sul)

Diff (no sul)


Figure 21. Playback off a single layer longitudinal recording medium versus linear density for four reader

designs.

2.1.6. Parallels between perpendicular and longitudinal recording

It should be reminded that a conventional shielded reader when applied to longitudinal recording is equivalent to a differential reader when applied to perpendicular recording

[114]. This is illustrated in Figure 22 where the sensitivity fields of a (shielded) differential reader and a shielded reader are compared. It should be reminded that the

maximum value of the sensitivity field defines the maximum value of the playback

signal. Therefore, this graph illustrates that the differential reader, regardless of whether it is shielded or not, provides a larger playback signal than the equivalent regular shielded

reader.


-200 -100 0 100 200-6

-4

-2

0

2

4

6

Sensitiv

ity F

ield

(a.u

.)


Shielded Reader (Hx)

Diff Reader (Hz)

Shield Diff (Hz)

Figure 22. Sensitivity fields for shielded, differential and shielded differential readers.

The normalized sensitivity fields for the readers above are shown in Figure 23. It can be

observed that that normalized sensitivity functions of a shielded reader and a shielded

differential reader are almost identical while the sensitivity function of a not shielded differential reader has somewhat wider tails.

-200 -100 0 100 200

-1.0

-0.5

0.0

0.5

1.0

Sensitiv

ity F

ield

(a.u

.)


Shielded, Hx

Differential, Hz

Shielded Diff, Hz

Figure 23. Normalized sensitivity fields for a shielded, differential, and shielded differential readers.

For completeness of this description, it is instructive to compare the performance of the

above mentioned reader designs as applied to perpendicular and longitudinal recording.


0 500 1000 1500 2000 2500 3000

0

5

10

15

20

25

30

35

Pla

yback (

a.u

.)


No shield (perpendicular)

Shielded (perpendicular)

Differential (perpendicular)

Shield Diff (perpendicular)

No shield (longitudinal)

Shielded (longitudinal)

Differential (longitudinal)

Shield Diff (longitudinal)

Figure 24. Perpendicular and longitudinal systems playback for four reader designs.

Figure 24 compares the playback versus the linear density for a perpendicular medium with a soft underlayer and a longitudinal single layer medium with equivalent recording

layers. It can be observed that for all the considered reader designs, the playback

amplitude is higher for perpendicular recording (as compared to longitudinal recording), which is clearly an advantageous feature of perpendicular recording.

0 500 1000 1500 2000 2500

0

5

10

15

20

Pla

yback

(a.u

.)


Shielded (longitudinal)

Differential (longitudinal)

Shield Diff (longitudinal)

Shielded (perpendicular (no sul))

Differential (perpendicular (no sul))

Shield Diff (perpendicular (no sul))

Figure 25. Perpendicular without a soft underlayer and longitudinal systems playback for three reader designs.

For comparison, Figure 25 shows the playback signals off a perpendicular system with a medium without a soft underlayer and a single layer longitudinal medium. Figure 26


shows the playback signal off a perpendicular system with a medium with a soft

underlayer. Similarly, the same three reader designs as in Figure 25 are considered. The comparison of Figure 25 and Figure 26 indicates that the higher playback amplitude in

perpendicular recording is mostly due to the utilization of a medium with a soft

underlayer.

2.1.7. Influence of Shields

Number of Shields. The playback signal off a perpendicular recording medium with a soft underlayer and a single layer longitudinal recording medium versus the linear density

for three cases, 1) not shielded, 2) one-side shielded, and 3) two-side shielded single MR

sensor reader designs are shown in Figure 26 and Figure 27, respectively.

0 500 1000 1500 2000 2500 3000

0

5

10

15

20

25

30

Pla

yback (

a.u

.)


No shield z (sul)

One shield z (sul)

Two shield z (sul)

Figure 26. Playback off a perpendicular recording medium with a soft underlayer versus linear density for not

shielded, one-side shielded, and two-side shielded single MR sensor reader designs.

0 500 1000 1500 2000 2500

0

5

10

15

Pla

yback (

a.u

.)


No shield x (no sul)

One shield x (no sul)

Two shield x (no sul)

Figure 27. Playback off a single layer longitudinal medium versus linear density for not shielded, one side

shielded, and double shielded single MR sensor reader designs.


These graphs illustrate that the addition of shields improves the resolution of a reader at

higher linear densities for both perpendicular and longitudinal recording.

-200 -150 -100 -50 0 50 100 150 200

-2

0

2

4

6S

ensitiv

ity function (

a.u

.)


Thick shield, Hx (no sul)

Thick Shield, Hz (sul)

Thin Shield, Hx (no sul)

Thin Shield, Hz (sul)

Figure 28. Sensitivity functions for perpendicular and longitudinal shielded readers of two extreme values for

the shield thickness: thick shield – 100nm, thin shield – 10nm.

0 500 1000 1500 2000 2500

0

5

10

15

20

Pla

yback (

a.u

.)


Thick shield, Hx (no sul)

Thick Shield, Hz (sul)

Thin Shield, Hx (no sul)

Thin Shield, Hz (sul)

Figure 29. Playback versus linear density for perpendicular and longitudinal shielded readers of two extreme

values for the shield thickness: thick shield – 100nm, thin shield – 10nm.


Shield Thickness. Figures 28 and 29 show the sensitivity function and the roll-off curve,

respectively, for double-sided shielded readers for the cases of a single layer longitudinal medium and a perpendicular medium with a soft underlayer for 100nm and 10nm thick

shields.

Only a very weak dependence on the shield thickness could be observed for the cases

presented above.

0 500 1000 1500 2000 2500

0

5

10

15

20

25

30

35

Pla

yback (

a.u

.)


Shield (sul)

Shield (no sul)

Diff (sul)

Diff (no sul)

Shield Diff (sul)


Figure 30. Comparison of playbacks of three reader designs (shielded, differential, and shielded differential)

for the cases of perpendicular media with and without a soft underlayer.

2.1.8. Soft Underlayer Versus No Soft Underlayer

Figures 30 and 31 show the roll-off curves for perpendicular and longitudinal systems,

respectively, for three reader designs for media with and without a soft underlayer. It can be observed that while in the perpendicular system the addition of a soft underlayer

substantially increases the playback amplitude, in the longitudinal system the addition of

a soft underlayer (keeper layer) leads to a substantial drop in the playback signal. The physical explanation of the phenomenon is illustrated in Figure 32 where imaging

properties of a soft underlayer film are outlined for the cases of perpendicular and

longitudinal recording. In perpendicular recording, the addition of a soft underlayer effectively doubles the recording layer thickness and thus increases the amplitude of the

stray field. In longitudinal recording, the addition of a soft underlayer film creates an

effective layer underneath the recording layer with the magnetization oriented opposite to the magnetization written into the recording layer. As a result, the net stray field

decreases.


0 500 1000 1500 2000 2500

0

2

4

6

8

10

12

14

16

Pla

yback (

a.u

.)


Shield (no sul)

Diff (no sul)


Shield (sul)

Diff (sul)

Shield Diff (sul)

Figure 31. Comparison of playbacks of three reader designs (shielded, differential, and shielded differential)

for the cases of longitudinal media with and without a soft underlayer (keeper layer).

2.1.9. Differential Reader Optimization and Single MR Differential Readers

The performance of differential readers can be further optimized by adding a soft magnetic material bridge that magnetically couples the two sensors, as shown in Figures

33a and b. Additional configurations of a differential reader that are worth considering

are shown in Figure 33c and d, where one of the MR sensors is substituted with an equivalent soft magnetic material. The latter is simpler to manufacture as building

differential readers represents technological challenges associated with manufacturing of

double-MR elements with the outputs connected to form a differential circuit.

SUL

SUL

perpendicular

longitudinal

Figure 32. Illustration of the imaging properties of soft underlayer for the cases of perpendicular and

longitudinal recording schemes.



MR

Sensor

MR

Sensor

(c) (d)

shield shield


MR

Se

nsor

MR

Se

nsor

(a) (b)

shield shield

MR

Se

nsor

MR

Se

nsor

Figure 33. Schematics of bridged differential readers: a) not shielded, b) shielded, c) not shielded with one MR

element, d) shielded with one MR element.

Figures 34 and 35 show z and x components of the sensitivity function, respectively, for

three configurations of an unshielded differential reader. It can be observed that the

addition of the bridge connecting two MR elements substantially increases the magnitude of the sensitivity function. Also, it can be noted that all the readers have asymmetric

profiles along the track. Similarly, Figure 36 and 37 show z and x components of the

sensitivity functions, respectively, for the three configurations of a shielded differential reader. Again, it can be noted that the addition of the bridge connecting the two MR

elements substantially increases the magnitude of the sensitivity function. As in the case of the unshielded reader, single MR element based shielded readers have asymmetric

profiles along the track.

-200 -100 0 100 200

-10

-5

0

5

10

Hz (

a.u

.)


Bridged (sul)

Diff (sul)

Half Diff (sul)

Figure 34. Vertical component of the sensitivity function for three types of differential readers: diff –

conventional differential reader, bridged – differential reader with two MR elements connected by a soft

magnetic bridge, half diff – bridged with one of the MR elements substituted with soft magnetic material.


-200 -100 0 100 200

-2

0

2

4

6

8

10

Hx (

a.u

.)


Bridged (no sul)

Not bridged (no sul)

Half Diff (no sul)

Figure 35. Horizontal (along the track) component of the sensitivity function for three types of differential

readers: not bridged – conventional differential reader, bridged – differential reader with two MR elements

connected by a soft magnetic bridge, half diff – bridged with one of the MR elements substituted with soft

magnetic material.

-200 -100 0 100 200-10

-5

0

5

10

Hz (

a.u

.)


Bridged (sul)

Diff (sul)

Half Diff (sul)

Figure 36. Vertical component of the sensitivity function for 3 modifications of shielded differential readers:

diff – conventional differential reader, bridged – differential reader with two MR elements connected by a soft

magnetic bridge, half diff – bridged with one of the MR elements substituted with soft magnetic material.


-200 -100 0 100 200

-2

0

2

4

6

8

Hx (

a.u

.)


Bridged (no sul)

Diff (no sul)

Half Diff (no sul)

Figure 37. Horizontal component of the sensitivity function for 3 modifications of shielded differential

readers: diff – conventional differential reader, bridged – differential reader with two MR elements connected

by a soft magnetic bridge, half diff – bridged with one of the MR elements substituted with soft magnetic

material.

Figures 38 and 39 compare the roll-off curves for double MR element differential readers and modified single MR element differential readers as shown in Figure 33 for the cases

of perpendicular and longitudinal recording, respectively. It can be seen that in both

perpendicular and longitudinal recording, single MR sensor differential readers give better performance at lower linear densities than double MR sensor differential readers.

The performance at higher linear densities is approximately the same for all the

considered types of readers.

0 500 1000 1500 2000 2500 3000

0

10

20

30

40

50

Pla

yback (

a.u

.)


Diff (sul)

Shield Diff (sul)

Half Diff (sul)

Shield Half Diff (sul)

Figure 38. Playback off a perpendicular recording medium with a soft underlayer versus the linear density for

double-MR sensor and single-MR sensor differential readers.


0 500 1000 1500 2000 2500

0

5

10

15

20

Pla

yback

(a.u

.)


Diff (no sul)


Half Diff (no sul)

Half Shield Diff (no sul)

Figure 39. Playback off a single layer longitudinal medium versus linear density for double-MR sensor and

single-MR sensor differential readers.

2.1.10. Parallels Between Playback in Perpendicular and Longitudinal Magnetic Recording: Revisited

Although perpendicular recording is one of the closest alternatives to conventional

longitudinal recording, the implementation of this technology is subject to resolving a number of technical issues not encountered in longitudinal recording. As mentioned

above, relatively little attention has been given to the read process [115,85]. Currently,

the read heads used in perpendicular recording system prototypes remain largely unchanged from their original longitudinal versions [116]. It is not clear that such read

heads provide optimal playback performance. In addition, the application of longitudinal

readers in perpendicular recording leads to undesirable phenomena associated with adjacent track reading [117,118] and calls for modification of the existing read channels

[119,120, 121].

The subject of this Section is a read head design for a perpendicular recording system

equivalent in its playback characteristics to a conventional longitudinal reader used in


Overview of Reader Designs. A diagram of a conventional longitudinal read head (LRH) - shielded reader consisting of a read element (such as MR, GMR etc.) surrounded

by two magnetic shields - is shown in Figure 40 [122].


tMR

S1 S2

tSS

z

x

Figure 40. A side view diagram of a shielded reader.

Designed for the use in longitudinal recording, a typical LRH is configured so that it

predominantly senses the stray fields emanating from the bit transitions in a longitudinal

recording medium. In longitudinal media, stray fields near bit transitions are mostly of

perpendicular orientation and the magnitude of the perpendicular component of the stray

field, Hz, decays rapidly away from a bit transition. Therefore, a typical LRH is designed

to preferably sense the longitudinal field component, Hz. If designed to read the

longitudinal component of the stray field, Hx, the reader would sense stray fields

relatively far from a bit transition, because the rate of the decay of the Hx is substantially

slower, as shown in Figure 41a.

-0.10 -0.05 0.00 0.05 0.10

Hz

Hx

Distance along the track ( m)-0.10 -0.05 0.00 0.05 0.10

-500

0

500

1000

Hz

Hx

Hx a

nd H

z (a

rb.

units)


(a) (b)

Figure 41. The z- and x-components of the stray field at a 5nm flying height near the transition (X=0) in case

of (a) a longitudinal and (b) perpendicular medium.


In the case of a perpendicular medium, the vertical and longitudinal components of the

stray field are interchanged compared to the case of a longitudinal medium: the stray field near the bit transitions is predominantly longitudinal, while the stray field away from the

transitions is mostly perpendicular, as shown in Figure 41b. This is opposite to the

longitudinal medium for which the stray field near the transitions is predominately perpendicular while the stray field away from the transitions is mostly longitudinal, as

shown in Figure 41a. Therefore, in a magnetically equivalent perpendicular system, not

only should the medium orientation change from longitudinal to perpendicular, the read-head configuration should also be changed so that it preferably reads the longitudinal

component, Hx, instead of the perpendicular component, Hz. To develop such a

perpendicular read head (PRH) design, conformal mapping (CM) could be utilized [123]. It should be remembered that conformal mapping is a 2-D mathematical tool. However,

as shown below, this convenient conceptual instrument allows to avoid any unnecessary guess work in the design of an adequate perpendicular magnetic head. The concepts

developed for the 2-D case could be adequately extended in the 3-D case. Below, 3-D-

boundary element modeling (BEM) is utilized to apply the conceptual findings to design a perpendicular magnetic playback head which is magnetically equivalent to a magnetic

playback head used in longitudinal recording.

tMR

H

tSS

H

Figure 42. Transformation of a general geometry into a shielded reader at H .

Conformal Mapping. The LRH geometry, as shown in Figure 40, is a limiting (H )case of a more general geometry, as shown in Figure 42. The complementary object to

the general LRH geometry is a 180-degree rotated dual head (DH), as shown in Figure 43. Moreover, because the LRH and DH are complementary objects in a complex

variable plane, Z = X +iY, according to the Symmetry Principle, these objects could be

obtained as a result of correlated transformations of the positive and negative imaginary semi-planes of the complex variable plane, W = U + iV, by two complementary

functions, and)(wF )(wF , respectively. Function could be found according to the

Schwartz-Christoffel Transformation [124].

)(wF


SR

DH

ImW > 0

ImW < 0

Z=X+iY W=U+iV

Y=0)(wF

)(wF

Figure 43. The LRH and DH designs are the results of the transformation by complimentary functions,

and)(wF )(wF , of the top and bottom imaginary semi-planes.

In analogy with LRH, it can be noted that in the limiting case, H , the DH geometry transforms into geometry, consisting of two infinitely tall poles, as shown in Figure 44.

tGAP

H

tGAP + 2tPOLE

H

Figure 44. Infinite-throat-height limit (H ) of the DH geometry.

According to the Reciprocity Principle [125], the sensitivity field could be used as a

measure of the playback performance. According to this theory, the playback signal is a

convolution of the sensitivity field and the magnetization in the recording layer. It

should be remembered that the use of the Reciprocity Principle assumes that no

saturation processes occur in the system. [126]. The sensitivity field, as the field

generated by a unit electric current in imaginary coils wrapped around the

magnetoresistive element of the read head, is an intrinsic property of the read system.

The boundary conditions for the LRH are chosen such that the main pole and the shields

have two different values of the scalar potential, 1 and –1, respectively, as shown in

Figure 45a. The transformation of the boundary surface into the W-plane is a straight

line, ImW=0, with three intervals of specified scalar potentials, as shown in Figure 45a.

Similarly, the boundary conditions for the DH are chosen such that the two poles have

different scalar potential values, 1 and –1, as shown in Figure 45b. The transformation


of the boundary surface of the DH into the W-plane is the same straight line, W=0, with

two intervals of specified scalar potential, respectively. The intervals with specified

scalar potentials are used as the boundary conditions to solve the Laplace’s Equation in

the W-plane in the two cases, as shown in Equations 5 and 6, for the LRH and DH,

respectively.

(a)

ImW > 0Z=X+iY W=U+iV)(wF

A1

A2 A3

A4 A5 = A-1A-4

A-2A-3

A5 =

a1 a2 a3 a4a-1a-2a-3a-4Y = 0H

= -1 = 1 = -1

= -1 = 1

(b)

ImW < 0Z=X+iY W=U+iV)(wF

A1

A2 A3

A4 A5 = A-1A-4

A-2A-3

A5 =

a1 a2 a3 a4a-1a-2a-3a-4Y = 0H

= -1

= 1 = -1 = 1

Figure 45. Diagrams showing the “charge” configuration in (a) longitudinal and (b) perpendicular recording.

2

2 1

11

1

1Im

1

bnd

bnd bnd

Wt

dt

Wt

dtbnd

Wt

dt, (5)

2

1

1

2

Im1

bnd

bnd

bnd

bnd

Wt

dt

Wt

dt, (6)

As the last step to calculate the magnetic field components in the Z-space, it is fairly

illustrative to use the representation, as given by Expression 7 [127].


dWdZdW

d

xHiyH1

, (7)

where dZ/dW is the derivative of the above described transformation functions, and)(wF

)(wF , for the two cases of interest, respectively. Considering the complementary nature

of the two cases, it can be noted that Hx and Hy for the LRH are equivalent to –Hy and

Hx for the DH, respectively. Such a 90-degree rotation of the sensitivity field, as compared to the LRH, is exactly what is expected from the magnetically equivalent

PRH, as described above. Therefore, a pole tip configuration of the DH type satisfies the

requirement, providing the LRH and DH dimensions are related as tMR = tGAP, tSS = tGAP

+ 2 tPOLE, where tMR and tSS are the sensor thickness and the shield-to-shield separation

of the LRH, respectively, and tGAP and tPOLE are the gap and the thickness of each pole of

the DH, respectively (See Figure 42 and Figure 44).

Use of a Soft Underlayer. A typical perpendicular system includes a soft underlayer

(SUL), the presence of which is not directly addressed in the arguments presented in Section 0. In the presence of the SUL, the net sensitivity function of a reader consists of

the sensitivity function of the reader itself and of the sensitivity function of the reader’s

image (with respect to the SUL boundary). Due to the symmetry, the sensitivity function of the reader in the presence of the SUL is given by

t)(z(z)(z) (z) (z) RRIR HHHHH , (8)

where HR(z) and HI(z’) are the sensitivity functions of the reader and its image,

respectively, z and z’ are the distances from the ABS of the reader and its image, respectively, and tRL is the recording layer thickness. The playback is then given by

RLFH

FH

R

RLFH

FH

R

tt

tdtztz

I

tt

tdzz

IS 33 )()(1)()(1~ rHMrHM , (9)

where M(z) is the recording layer magnetization and tFH is the flight height. The

integration boundaries are shown for the z coordinate. The above equation can be

rewritten as

RLFH

FH

R

tt

t

dzzI

S

2

)()(1~ 3rHM , (10)


where M(z) for z> tFH +tRL is defined as a mirror image of the recording layer magnetization with respect to the SUL boundary. The latter equation describes the

playback in the absence of a SUL with the recording layer thickness doubled. Therefore,

a DH reader used in a perpendicular recording system with a SUL is magnetically equivalent to a LRH used in longitudinal recording, in which the recording layer

thickness is twice as thick as the recording layer thickness in the perpendicular system.

3D BEM Calculation. As a direct application of the developed concept, 3D-BEM

calculations were performed [92]. As an example of a PRH with the DH type pole

configuration, a yoke type MR or GMR head is considered, as shown in Figure 46.

GMR

PT

G

Yoke ABS

MR element

Figure 46. Side view diagram of a yoke type magnetoresistive head.

The sensitivity field components corresponding to the two head types, LRH and PRH, at a 5nm flying height are shown in Figure 47.

0.00 0.02 0.04 0.06 0.08 0.10

-1.0

-0.5

0.0

0.5

Longitudinal

Hx a

nd H

z (

norm

aliz

ed)

Down the track (um)

LR: Hx

LR: Hz

-1

0

1

Perpendicular

PR: Hx

PR: Hz

LRH: Hx

LRH: Hz

PRH: Hx

PRH: Hz

Figure 47. Sensitivity field components for the LRH and PRH at a 5nm flying height.


Assuming 3-D geometry, a different coordinate system was chosen: unlike in the system

implied in the 2D CM theory, the y-coordinate stands for the direction across the track, and the z-coordinate stands for the perpendicular direction. Both the standard LRH and

the yoke head of the PRH type were modeled to have the same trackwidth, W = 200 nm.

Also, the following ratio was kept between the LRH and PRH parameters: tMR = tGAP = 10nm, tSS = tGAP + 2tPOLE = 50 nm. In agreement with the theoretical conclusions above, it

can be noticed that Hx and Hz for LRH look similar to -Hz and Hx for the PRH. According

to the Reciprocity Principle, the playback signal is convolution between the sensitivity field and the magnetization in a recording medium. Therefore, it can be concluded the

two designs provide similar waveforms, if they are used with longitudinal and

perpendicular media, respectively. It should be remembered that the arguments presented in the previous section is an approximation. Consequently, the equivalency of a LRH and

a PRH is also an approximation. The normalized roll-off curves for a LRH and a PRH used with longitudinal and perpendicular media, respectively, are shown in Figure 48. It

can be noticed, that while both curves are similar, they are not identical.

0 500 1000 1500 2000 2500

0.0

0.2

0.4

0.6

0.8

1.0

Norm

aliz

ed P

layback A

mplit

ude


LRH: longitudinal medium

PRH: perpendicular medium

Figure 48. Normalized playback amplitude versus linear density for a LRH and a PRH used with longitudinal

and perpendicular media, respectively.

The implementation of the design above is subject to the fine control of the magnetic

domain noise within the yoke structure [128]. A design that addresses the issue of domain

noise control in a DH, is a dual element magnetoresistive reader [129,130], a schematic of which is shown in Figure 49. Because the sensitivity field is defined largely by the

yoke structure employed in the reader design, both the yoke type (G)MR sensor shown in Figure 46 and the dual (G)MR sensor shown in Figure 49 are magnetically equivalent.


MR1

MR2

PT

G

Yoke ABS

Figure 49. Side view diagram of a dual (G)MR element reader.

Conclusions on Study of Parallels between Playback in Perpendicular and Longitudinal

Recording. Conformal mapping theory was developed to show the playback equivalency between the conventional shielded GMR read-head configuration used in conjunction

with a longitudinal medium and the DH configuration used in conjunction with a

perpendicular medium. The approach chosen to design a PRH, which is magnetically equivalent to a LRH, was identifying the head configuration sensitive predominantly to

the longitudinal stray field, thus significantly reducing reading away from the transitions.

The requirements for maintaining the playback waveform equivalency between these two head configurations are: tMR=tGAP and tSS=tGAP+2tPOLE, where tMR and tSS are the thickness

and shield to shield separation of the LRH, respectively, and tGAP and tPOLE are the gap

and the thickness of each pole tip of the PRH, respectively. This concept was extended to 3D and supported by 3D BEM calculations of the sensitivity fields for the two types of

heads. As examples of the PRH’s in 3D case, a yoke type (G)MR reader and a dual

(G)MR reader were considered.

Chapter 4 Perpendicular Recording Media

127

Chapter 4

Perpendicular Recording Media

1. Introduction

As magnetic data storage industry is facing its fundamental limit due to thermal

instabilities in the longitudinal recording media [131], perpendicular magnetic recording

is becoming the center of attention in the industry [132,133].

The use of perpendicular magnetic recording media instead of conventional longitudinal

media is the main reason why perpendicular recording is considered to be the technology capable of deferring the superparamagnetic limit to areal densities much beyond 100

Gbit/in2. As described above, a typical perpendicular medium consists of two main

magnetic layers [134,135]: 1) the recording (magnetically “hard”) layer [136] and 2) the magnetic “soft” underlayer (SUL) [137,138, 139]. Such double-layer perpendicular

medium is usually used together with a single pole recording head, as described above in

Chapter Physics of Writing.

1.1.CHAPTER OVERVIEW

In this chapter, the results of theoretical and experimental study of some of the key issues

related to perpendicular magnetic media are presented. To stress the specific aspects of the recording physics native to perpendicular recording, a comparison between

longitudinal and perpendicular recording media is carried out throughout the entire

chapter. Specific attention is given to the role of the soft underlayer as a new component in the recording process. Among the discussed issues are the guidelines and the

underlying physics to choose the optimized parameters of the recording layer and the soft

underlayer and the integration of these two components. The noise due to the soft underlayer and means to minimize the noise are discussed. In addition, it is described

how Kerr microscopy could be utilized to study the dynamics of perpendicular recording

with a soft underlayer.


2. Perpendicular Recording (“Hard”) Layer

The primary approach to the design of a perpendicular recording layer is in many ways

similar to the design of a conventional longitudinal media. Major goals relevant to the development of both longitudinal and perpendicular recording layers are: achieving

sufficiently small grain size and grain size distribution, texture control, optimization of

the inter-granular quantum exchange de-coupling, etc. However, some aspects of the recording layer design are specific only to the perpendicular recording modes [140].

Understanding the fundamentals of the aspects inherent to perpendicular recording media

is the subject of this chapter.

2.1. TYPES OF MEDIA

Among the large variety of today’s perpendicular magnetic recording media, CoCr-based

alloys, Co/Pt-based mutlilayers, L10 phases of FePt, BaFe, and others, one could separate two large categories which have been most thoroughly explored for this purpose: (1)

CoCr-based alloy media and (2) media based on magnetic multilayers, such as Co/Pt,

Co/Pd or others [141,134,142,143,144,145].

Co

Pd

Figure 1. A schematic diagram of the cross-sectional view of a Co/Pd-multilayer-based recording layer.

Material-wise, perpendicular CoCr-based alloy [146,147,148] recording layers are similar

to the conventional longitudinal CoCr-based media, with the major difference being the orientation of the magnetic easy axis [149]. Therefore, a significant amount of

information accumulated in the course of the longitudinal media development could be used to control the critical parameters such as the grain size and the inter-granular

quantum exchange coupling in the perpendicular media. At the same time, the

development of CoCr-based perpendicular media has some unresolved issues not encountered in the development of the longitudinal media [150]. For example, it is not

yet clear whether it is feasible to engineer a CoCr-based medium with sufficiently high

anisotropy to avoid thermal instabilities at ultra-high areal densities. It has also proved not to be trivial to engineer CoCr-alloy-based perpendicular recording layers with a

remanent squareness of 1. It is believed that a remanent squareness of 1 is necessary for

the low-density bit pattern stability. Also, remanent squareness of < 1 can lead to substantial amounts of DC noise. Various magnetic alloys such as L10 phases of FePt,

CoPt, etc., are being studied as highest anisotropy alternatives for the recording layer

[151,152,153]. Selected material properties, such as the anisotropy density, Ku, the saturation magnetization, Ms, the anisotropy field, Hk, and the minimum stable grain size,

Chapter 4 Perpendicular Recording Media 129

a, as defined earlier in Chapter Fundamentals of Perpendicular Recording, for different

alloy systems, including Co-based alloys, L1 phases, and rare earth materials, are shown in Table 1.

Table 1Selected material properties (the anisotropy density, Ku, the saturation magnetization, Ms,

the anisotropy field, Hk, and the minimum stable grain size, a, as defined earlier in

Chapter Fundamentals of Perpendicular Recording) for different alloy systems: Co-based alloys, L1 phases, and rare earth materials.

Alloy System Material Anisotropy Saturation Magnetization Anisotropy Field Minimum stable grain size

Ku (107erg/cc) Ms (emu/cc) Hk (kOe) a (nm)

CoCrPtX 0.20 200-300 15-20 8-10

Co-alloy Co 0.45 1400 6.4 8.0

Co3Pt 2.00 1100 36 4.8

FePd 1.8 1100 33 5.0

L10-phase FePt 6.6-10 1140 116 2.8-3.3

CoPt 4.9 800 123 3.6

MnAl 1.7 560 69 5.1

Rare Earth Nd2Fe14B 4.6 1270 73 3.7

SmCo5 11-20 910 240-400 2.2-2.7

Figure 2. Top view TEM images of (a) a CoCrPtTa-alloy-based and (b) Co/Pd multilyaer –based recording

layers, respectively.

The multilayer-based recording layers (See Figure 1) typically have significantly larger

anisotropy (coercive fields of above 15 kOe have been reported) and thus promise to be extendable to significantly higher recording densities [154]. In these materials, the

magnetic anisotropy is controlled through the (surface) interfacial interaction between the

magnetic layer, (Cobalt) and a highly polarizable spacer layer (Palladium or Platinum). In


contrast to the alloy media, the multilayers typically display a very weak texture. Top

view TEM images of CoCrPtTa-alloy-based and Co/Pd-multilayer-based recording layers are shown in Figure 2a and b, respectively. The alloy-based and multilayer-based media

were sputter-deposited on Ti and ITO-based seed layers, respectively. A (thickness)

cross-sectional TEM image of the Co/Pd-based medium is shown in Figure 3. This image clearly indicates a columnar-type texture with an average column size of approximately

20 nm and with no ordered structure across the thickness. In other words, despite the

strong perpendicular anisotropy of the multilayer medium (due to the surface energy, as discussed earlier), typically no matching is detected between the sets of easy axes in the

adjacent Pd sub-layers.

Figure 3. A cross-sectional TEM image of a C/Pd-multilayer-based medium.

Another advantage of the magnetic multilayers is the fact that typically these materials have a remanent squareness of 1. The squareness of 1 indicates that the anisotropy field,

Hk, which keeps the magnetization in the perpendicular to the disk direction, is larger

than the maximum demagnetization field, 4 Ms. Consequently, because the demagnetization field reaches its maximum in the low-density limit, a medium with a

squareness of less than 1 tends to be unstable at low densities. In this case, the relatively strong demagnetization field substantially increases the chance of the magnetic moment

to be reversed as a result of thermal fluctuations [155,156].

To compare basic magnetic properties of CoCr-alloy and multiplayer based recording

layers, typical M-H loops [157] by a Kerr magnetometer [158,159] for a 50 nm thick perpendicular CoCr thin film and a 52 nm thick Co/Pd structure (a stack of 40 sets of

adjacent 3 and 10 A thick layers of Co and Pd, respectively) are shown in Figures 4a and

b, respectively. It can be noted that in addition to the remanent squareness of 1, the Co/Pd structure exhibits nucleation fields in excess of 3 kOe, a useful characteristic to avoid

data self-erasure due to stray fields. Meanwhile, the CoCr material shown in Figure 4a

has squareness of 0.75. The CoCr and Co/Pd recording layers have coercive field and magnetization of approximately 3 and 9 kOe and 300 and 200 emu/cm3, respectively.


Figure 4. A M-H loop of a 50 nm thick (a) CoCr-alloy layer and (b) Co/Pd multilayer.

The direct consequence of remanent squareness of < 1 is shown in Figure 5, which

compares the spectral SNR distributions for the two media types [160]. The CoCr

medium exhibits a significant amount of noise at lower linear densities. This is mainly due to the fact that the dominant contribution to the noise at low linear density in the

CoCr-base medium comes from the DC noise that results from the relatively low value of

remanent squareness, as described below in more details.

2.2. CONTINUOUS MEDIA

Also, it should be mentioned that there is another type of a magnetic recording medium,

which, similarly to a typical magneto-optical recording medium, due to relatively strong

exchange coupling between grains, acts as a magnetically continuous media [161]. In these so-called continuous magnetic materials, the bit separation is determined not by the

grain size, but rather by the domain wall width. The domain wall width (in these

relatively high anisotropy magnetic materials) could be as thin as few Angstroms. The coercivity field for these materials strongly depends on the mechanism and strength of

the pinning of the domain walls to naturally or artificially created defects. Today, because of many open questions continuous medium recording is not considered as the near-

future alternative to longitudinal recording, and research activities in this area are still

fairly rare. Therefore, the continuous materials are not covered in this chapter.

2.3. MAGNETIC FIELD CALCULATION

In this chapter, two approaches are used to calculate the magnetic fields. The analytical

solution of the Laplace’s Equation is used to calculate the stray and demagnetization field

for the cases of periodic bit patterns written into a perpendicular recording medium, as

shown in Equations 1 and 2, respectively, where and Ms are the hard layer thickness and saturation magnetization, and a and b are the bit length and width, respectively

[162,163].


Figure 5. SNR versus the linear density for a CoCr-alloy (hollow circles) and a Co/Pd multilayer (hollow

squares).

The origin of the reference coordinate system was chosen to be located at a corner of a bit at the top side of the recording layer. To study the dependencies on the BAR and the areal

density (AD) the transformation equations representing the bit length and width, a and b,

via BAR and AD, can be used, as shown by System of Equations 3. Note that the BAR is defined as the ratio of the track width to the bit length, b/a, and the effect on the fields of

the use of a SUL is equivalent to a two-fold increase of the recording layer thickness.

Therefore, the same expressions can be used to model an ideal SUL just via replacing

with 2 .

Figure 6. An illustration of the mirror imaging by an ideal SUL.


0,)

22

1(

22

1 1

sinsin32

zb

k

a

nz

b

k

a

n

oddn

oddk

yb

kx

a

n

kn

sM

strayzH

(1)

0,)(

1 1

sinsin32

2222

zzz

oddn

oddk

yb

kx

a

n

kn

sMdemagzH

b

k

a

n

b

k

a

n (2)

ADBARb

AD

BARa

1; (3)

Three-dimensional (3D) boundary element modeling (BEM) using a commercial field solver, Amperes, is used to calculate the magnetic field when bit patterns are written into

a longitudinal medium as well as to evaluate the magnetic field generated by magnetic

recording heads [164]. It should be noted that within the precision of the calculations, the BEM applied to periodic bit patterns in perpendicular media gives the results identical to

the results, which were calculated using the analytical solution [136].

The approximation of an “ideal” SUL is used in all calculations presented in this chapter

[165]. It should be reminded that the effect of the presence of an ideal SUL on the stray and demagnetizing fields generated by a recording layer is equivalent to the perfect

mirror-imaging of the recording layer with respect to the SUL boundary, as illustrated in

Figure 6. The fields above the SUL boundary are equal to the sum of the fields generated by the real recording layer and by its imaginary counterpart located below the SUL

boundary. If there is no separation between the recording layer and the SUL (in other

words, no buffer/exchange-decoupling layer is present), the use of the SUL is equivalent to a two-fold increase of the recording layer. Unless specified otherwise, it is assumed


that the thickness of the buffer layer is substantially smaller than the thickness of the

recording layer and, therefore, can be neglected.

Figure 7. The demagnetization field versus the distance down the track along the central planes of 10 and 20

nm thick recording layers for (a) longitudinal recording, (b) perpendicular recording without and (c) with a

SUL.

Figure 8. MFM images of tracks recorded into 30 nm thick CoCr alloys with a magnetization of (a) 200

emu/cm3 and (b) 400 emu/cm3.


It should be remembered that the use of a SUL is not equivalent to the effect of mirror-

imaging when net energy is to be evaluated. Therefore, the magnetic mirror-imaging should be used with caution when applied to the problems that deal with the bit energy.

For example, one cannot combine the energy of a bit with the “energy” of its magnetic

image to estimate the thermal stability of recorded information [166].

Figure 9. The maximum demagnetization field along the central line of 10, 20, and 40 nm recording layers

without a SUL at two values of linear density, 50 and 316 kfci.

2.4. DEMAGNETIZATION FIELD IN PERPENDICULAR RECORDING LAYER

The calculated normalized demagnetization field near a single ideal transition along the central plane of a recording layer is shown for longitudinal and perpendicular recording

layers with and without a SUL at two different values of the recording layer thickness, 10

and 20 nm, are shown in Figures 7a-c, respectively [162]. In these calculations, a relatively wide trackwidth is assumed. First, it can be noted that, unlike in longitudinal

recording, the demagnetization field in perpendicular recording decreases as the thickness

increases, thus promoting a larger thickness. If a perpendicular medium with a SUL is used, the SUL effectively further increases the recording layer thickness. Also, unlike in

the longitudinal medium, in both types of perpendicular media, the demagnetization field

reaches its minima at the transitions, thus promoting high-density recording. In this respect, it is common to notice that although perpendicular recording promotes high

densities, the stronger influence of the demagnetization field at lower densities is a

disadvantage of perpendicular recording.

One of the direct consequences of the strong demagnetization fields at low densities is a

relatively strong dc-noise from perpendicular media with a squareness of <1, as mentioned above. MFM images of wide (with respect to the media thickness) tracks

recorded into two 30 nm thick CoCr alloy films with a coercive field, Hc of


approximately 2 kOe and magnetization values of 100 and 400 emu/cm3 are shown in

Figures 8a and b, respectively. The presence of smoother and more uniform tracks in the case of the smaller magnetization indicates less dc-noise than in the other case.

Figure 10. The demagnetization field versus the areal density at the center of a bit (a) in the middle between

the top and bottom surfaces and (b) near the top surface of the recording layer for 10 nm thick recording layers

as a function of the areal density for different values of BAR.

The maximum demagnetization field at the center of a bit for three values of the

recording layer thickness with and without a SUL versus the track density for two values

of the linear density, 50 and 316 kfci, respectively, is shown in Figure 9. If considering only the magnetics of the recording process, perpendicular recording is symmetric with

respect to the directions along and across the track. For example, the demagnetization

field and stray field contours look similar (if rotated 900 in the plane) for the two sets of the following values for the linear and track densities: (1) 316 kfci and 50 ktpi, and (2) 50

kfci and 316 ktpi. Therefore, to avoid repetition, only one of the two data sets is

presented. In addition, it should be noted that under the assumption of zero spacing between the recording layer and the SUL, 10 and 20 nm cases with a SUL are equivalent

to 20 and 40 nm cases with no SUL, respectively.

As shown in Figure 9, the demagnetization field decreases with the increase of the track

density. The rate of the density roll-off is determined by the effective recording layer thickness and by the value of the linear density. (It should be reminded that the linear and

track densities are determined by the bit length and width, respectively.) It can be noted

that to avoid issues with thermal instability and dc-noise, a larger thickness and a higher track density are preferred. The SUL increases the effective recording layer thickness and

thus reduces the demagnetization field. However, it should be noted that the

demagnetization field cannot be reduced through the reduction of the trackwidth if the adjacent tracks mimic the main track. Therefore, a special encoding might be necessary

to avoid the unfavorable bit patterns at low densities. It is important to point out that at

present, the available encoding schemes (to accomplish the above-mentioned task) result in a substantial loss in the recording density.


+

+


+

+

+ + + + + + + + - - - - - - - - - -

- - - - - - - - - - + + + + + + +

Hstray

Hstray

+ + + + + + + + - - - - - - - - - -

- - - - - - - - - - + + + + + + +

Underlayer

boundaryMedium

image

M(a)

(b)

(c)

Charges on

two surfaces

Figure 11. Diagrams showing the "charge" distribution in (a) longitudinal media and perpendicular media (b)

without and (c) with a SUL.

An additional insight into the nature of the demagnetization field in the perpendicular

recording layer can be obtained if the demagnetizing field is plotted against the areal density for different values of the bit aspect ratio (BAR). The demagnetizing field at the

center of a bit in the middle between the top and bottom surfaces and near the top surface

of the recording layer is shown in Figures 10a and b, respectively. It can be seen that a higher value of BAR at a given areal density results in a lower demagnetizing field. This

is in contrast to longitudinal recording where the demagnetizing field increases as the value of BAR is increased.

2.5. STRAY FIELD FROM PERPENDICULAR RECORDING MEDIA

It should be mentioned that in perpendicular recording stray magnetic fields (the fields

sensed by a reader) emanate not from the magnetic “charge” in transitions, as in longitudinal recording, but from the “charge” at the top and effective (due to the use of

the SUL) bottom sides of the recording layer, as shown in Figures 11a-c [140]. Outside

the recording layer the fields from the top and effective bottom “charge” are of opposite directions; hence, they cancel each other if the recording layer thickness is significantly


smaller than the characteristic bit sizes. As shown in Figure 11c, the presence of the SUL

effectively doubles (in case of perfect imaging) the recording layer thickness.

Figure 12. The stray field above the center of a bit in 10, 20, and 40 nm thick recording layers without a SUL

at a 5 nm flying height versus the linear density for two values of the track density, 50 and 316 ktpi.

The net stray field emanating from the center of a bit in a periodically written bit patterns at two values of the track density, 50 and 316 ktpi, versus the linear density for media

with a recording thickness of 10, 20, and 40 nm with no SUL, are shown in Figure 12. It

should be stressed again that under the assumption of zero spacing between the recording layer and the SUL, 10 and 20 cases with SUL are equivalent to 20 and 40 nm cases with

no SUL, respectively. In these calculations, a flying height of 5 nm is assumed. It can be

noted that for the narrower trackwidth, the signal could be substantial even at relatively low linear bit densities because no total compensation of the fields generated by the top

and effective bottom charges occurs. It is not unnatural that at sufficiently high densities,

the stray field drops because the bit cell surface area containing the effective magnetic “charge” and, thus, also the net magnetic “charge” becomes relatively small. Regardless

of the effective recording layer thickness, the characteristic bit length at which this

happens cannot be smaller than approximately the flying height. At the same time, regardless of the linear and track density values, the stray field drops with the thickness

reduction because of the above-mentioned field cancellation effect due to the magnetic

“charge” at the top and effective bottom surfaces of the disk. (The latter effect is equivalent to the stray field reduction at some finite thickness as the bit dimensions

become substantially larger than the thickness.) In the intermediate case, the stray field generated by the effective “charge” at the top surface is sufficiently large to generate a


non-negligible field and at the same time cannot be compensated by the “charge” from

the effective bottom side. Therefore, the stray field’s dependence on the areal density has a maximum, as illustrated in Figure 12.

Figure 13. The stray field above the center of a bit in a 10 nm thick recording layer without a SUL at a 5 nm

flying height versus the linear density for different values of the spacing between a recording layer and a SUL.

From the above arguments, it is also clear that unless a special encoding is not performed

to limit the maximum physical bit dimensions, the stray field becomes negligibly small at sufficiently low densities.

Assuming that a special encoding is performed to avoid the low-density degradation of the stray field, the limiting lower boundary’s value for the recording layer thickness is

determined by the highest values of the required linear and track densities. For example,

aiming for 200 Gbit/in2 at a 2:1 bit aspect ratio (BAR), the track and linear densities would be 316 ktpi and 632 kfci, respectively. For these values of the track and linear

densities, the signal rapidly drops as the thickness becomes smaller than approximately

10 nm. By now, it was assumed that there is no separation between the recording layer and the SUL. The effect of a non-zero spacing between the recording layer and the SUL

is shown in Figure 13. It can be noted that the difference in the values of the stray field

between a 0 nm spacing and a 4 nm spacing can be as large as 20 %. Although this effect does not alter any of the conclusions presented in this chapter, it should be taken into

account when optimizing the design of a perpendicular recording system.


Figure 14. Perpendicular component of the stray field versus the areal density for the perpendicular mode with

and without a SUL.

Figure 15. Perpendicular component of the stray field versus the areal density for the longitudinal recording

mode.


It gives an additional insight into the difference between perpendicular and longitudinal recording if one compares the stray fields emanating from perpendicular and longitudinal

media as functions of the areal density at different values of bit aspect ratio (BAR). The

perpendicular component of the stray field at a 5 nm distance from the center of a bit in a periodically written pattern for a 10 nm thick hard layer at 3 values of BAR, 1:1, 4:1, and

8:1, is shown for perpendicular recording with and without a SUL in Figure 14. Since in

the case of longitudinal recording the directions along the track and across the track are not equivalent, one needs to consider the BAR values corresponding to both types of bits,

elongated along the across the track. The perpendicular component of the stray field at a

5 nm distance from the center of the transition in a periodically written pattern of a 10 nm thick longitudinal thick longitudinal hard layer at 6 values of BAR, 1:4, 1:2, 1:1, 2:1, 4:1,

and 8:1, is shown in Figure 15.

The pronounced peak in the perpendicular cases, as seen in Figure 14, can be explained

by the above-mentioned cancellation effect of the fields generated by the top and bottom magnetic “charge”. It can be noted that there exists a certain characteristic density of ~

100 and ~ 200 Gbit/in2 for the cases with and without a SUL, respectively, above and

below which the stray field depends differently on the areal bit density. This characteristic density increases with the reduction of the recording layer thickness.

Figure 16. The peak value of the perpendicular component of the stray field versus the recording layer

thickness for the perpendicular mode without a SUL.

The presence of the SUL is equivalent to the increase of the recording layer thickness by a factor of two. The observed essential sensitivity to the recording layer thickness can be

explained by the above-described nature of the magnetic “charge” in perpendicular

recording. As the areal density increases, the magnitude of the stray field becomes


independent of the recording layer thickness. As illustrated in Figure 16, the peak value

of the stray field strongly depends on the recording layer thickness and is independent of the BAR.

Roll-off curves show no peak in the case of longitudinal recording. Rather, they continuously drop for all values of BAR. The slowest roll-off is observed for a 1:1 value

of BAR.

Although the stray field roll-off curves are rather complex and strongly depend on the

value of BAR, it can be noted that in perpendicular recording, the stray field roll-offs

substantially slower than in longitudinal recording. This is due to the fact that the stray field is composed of the fields from the top and effective bottom “charge” of the

recording layer, respectively. The two fields are oppositely directed with respect to each other, thus the field generated by the bottom “charge” reduces the field generated by the

top “charge” and in overall reduces the net field. As the bit size becomes smaller, the

relative contribution of the bottom “charge” decreases thus slows the overall decrease of the field amplitude with the areal density increase.

2.6. WELL-DEFINED PERPENDICULAR EASY AXIS: THICKER RECORDING LAYER?

Figure 17. Schematic diagrams illustrating (a) a longitudinal medium with randomly oriented easy axes and

(b) a perpendicular medium with aligned easy axes, respectively.

In perpendicular recording, the easy axis is relatively well aligned in one direction

(perpendicular), unlike in a conventional longitudinal medium, in which the easy axes are

randomly oriented in the 2D plane of the disk, as shown in Figure 17. X-Ray “rocking” curve with a peak corresponding to the perpendicular easy crystalline axis (which

coincides with the perpendicular magnetic “easy” axis) for CoCrPtTa alloy medium

grown on a Ti seed layer is shown in Figure 18. The “rocking” curve indicates a texture


spread of less than 6.3 degree. Typically, the texture spread varies from 1 to 10 degrees,

depending on the deposition condition, seed layer(s), etc.

A well-defined easy axis potentially relaxes the stringent requirement to the trailing and

side writing field gradient necessary to achieve sharp transitions, thus enabling the use of thicker media. Consequently, this also explains the excellent overwrite in perpendicular

recording. Even, for sub-100-nm trackwidth recording, overwrite of the order of 40 dB

could be easily achieved [63].

0 1 0 2 0 3 0 4 00

1 0 0 0 0

2 0 0 0 0

3 0 0 0 0

4 0 0 0 0

5 0 0 0 0

6 0 0 0 0

7 0 0 0 0

8 0 0 0 0

0 1 0 2 0 3 0 4 00

1 0 0 0 0

2 0 0 0 0

3 0 0 0 0

4 0 0 0 0

5 0 0 0 0

6 0 0 0 0

7 0 0 0 0

8 0 0 0 0

FWHM=6.30

X-ray rocking curve

ideal

non-ideal

(a) (b)

Figure 18. (a) X-Ray rocking curve of a perpendicular CoCrPtTa alloy medium. (b) A schematic diagram

showing ideally and non-ideally aligned magnetic media.

The intrinsically better alignment of perpendicular media helps to record narrow tracks

with well-defined transitions even into a relatively thick recording layer, in contrast with longitudinal recording. A MFM image of two adjacent tracks with a 65 nm trackpitch

written into a 50 nm thick CoCr recording layer using a 50 nm wide single pole head is

shown in Figure 19.

Figure 19. A MFM image of two adjacent 65-nm wide tracks.


In this respect, it should be remembered that previously it was shown that, although well-

aligned perpendicular media might have a relatively small average angle between the magnetization and the perpendicular recording field, the torque between the

magnetization and the field is still sufficiently large to relatively rapidly switch the

magnetization [167]

3. Soft Underlayer

Figure 20. Mirror image representation of the SUL from the write head perspective.

One of the key features of perpendicular recording that makes it superior compared to longitudinal recording with respect to the superparamagnetic instabilities is the use of

media with a soft underlayer (SUL) [137,168,169]. Integration of a single pole head and

media with a SUL enables recording magnetic fields in excess of 80% of 4 Ms, where

4 Ms is the saturation magnetization of the head material. For comparison, state-of-the-

art longitudinal systems are only capable of recording fields of less than 2 Ms. Such a substantial increase in the recording field in perpendicular recording due to the use of a

SUL opens the possibility to record on media with substantially higher anisotropy and

thus leads to improved thermal stability. As mentioned earlier (See Figure 6), because of the un-ambiguity of the solutions of the Laplace’s Equation (used to calculate the

magnetic field) with respect to the adequate boundary conditions, ideally, the SUL can be

represented as a magnetic mirror and the effect of the SUL in the static sense is equivalent to the two-field increase of the thickness of the recording layer. The effective

increase of the recording layer leads to both increase of the stray magnetic field sensed by


a read head and reduction of the demagnetizing fields with a potential to further improve

thermal instability. The described view of the mirror imaging of the recording layer in the presence of the SUL is the perspective of the read head. From the perspective of the write

head (another mirror imaging perspective), the SUL can be equivalently replaced with a

mirror image head placed on the other side of the SUL symmetrically with respect to the SUL’s side adjacent to the recording layer, as shown in Figure 20.

However, while it is expected that the use of a SUL should make it possible to defer the superparamagnetic limit to areal densities beyond 1 Terabit/in2, the SUL also introduces a

number of technical challenges and issues. These issues should be resolved before

perpendicular recording can be fully implemented.

3.1. SATURATION MOMENT

The importance of the relation between the moment of the SUL’s material and the

moment of the recording pole tip’s material has been previously discussed [170]. It has been shown that saturation of the SUL can lead to a dramatic deterioration of the trailed

field gradients. According to the Maxwell’s laws, div B = 0 and H = 0, to avoid the SUL’s saturation under the pole tip, the following inequality has to hold,

4 MS SUL ASUL effective 4 MS Pole Tip AABS Pole Tip , (4)

where ASUL effective is the effective area of the SUL into which the magnetic flux emanating

from the pole tip enters and AABS Pole Tip is the area of the air bearing surface (ABS) of the

pole tip.

Figure 21. A schematic diagram of the recording head pole tip/SUL combination.

Figure 21 shows a schematic diagram of the combination of the recording pole tip and SUL. When the separation between the ABS of the pole tip and the SUL (d ~ 10-20 nm)


is substantially smaller than the lateral dimensions of the pole tip (lPole Tip ~ 300 – 1000

nm, wPole Tip ~ 50 – 150 nm), the effective area of the SUL, ASUL effective, is approximately equal to the ABS area of the pole tip, AABS Pole Tip. It follows that to avoid saturation in the

SUL, the magnetization of the SUL’s material should be equal or higher than the

magnetization of the recording tip’s material.

Figure 22. Trailing field from a single pole perpendicular write head made of FeAlN with a SUL made of two

different materials, FeAlN and Permalloy.

Although it is possible to generate strong recording fields (with magnitude approaching

4 Ms of the pole tip) even if the SUL has lower magnetization compared to the pole tip,

saturation of the SUL will lead to a substantial deterioration of the trailing field gradients. The results of the boundary element modeling (BEM) for two different head/SUL

combinations are presented in Figure 22. The two cases include the SUL made of 1- mthick Permalloy and FeAlN, respectively. Permalloy and FeAlN have saturation

magnetization of approximately 10 and 20 kGauss, respectively. As shown below, the thickness was chosen to be sufficiently large to avoid the geometrically caused saturation.

It can be observed that the trailing field gradient for the Permalloy-based SUL is

substantially lower than the trailing field gradient for the FelAlN-based SUL.

3.2. THICKNESS OF SOFT UNDERLAYER

The following approximate equation gives the value of the minimum thickness of the

SUL necessary to achieve the maximum recording field with the maximum trailing field

gradient, with assumption that the pole tip length is substantially larger than its width,

lPole Tip wPole Tip (See Figure 21),


TipPole

SULS

TipPoleS

MINSULw

M

Mt

21~ . (5)

.

The validity of the above equation could be proven using arguments similar to the arguments used in the previous section and is a consequence of the conservation of the

magnetic flux.

Figure 23. The amplitude and the gradient of the trailing field generated by a single pole head made of FeAlN

with a FeAlN-made SUL versus the SUL's thickness.

Figure 23 shows the results of the boundary element modeling of the magnetic field and

the field gradient for the recording field generated by a recording head with a 0.4 x 2 m2

pole tip cross-section with SULs of different thicknesses. Both the SUL and the pole tip

are assumed to be made of FeAlN with a 4 Ms of 20 kG. It can observed that as the

SUL’s thickness becomes smaller than approximately 0.2 m, the field gradient rapidly

drops and thus the performance of the recording system deteriorates.

Media with almost identical recording layers and different SULs were fabricated to test

the validity of the theoretical results outlines above. The pole tip was made of Ni45Fe55

alloy with a 4 Ms of 16 kG and the SUL was made of FelAlN with a 4 Ms of 20 kG. The


recording trackwidth (~ wPole Tip) was approximately 0.8 m. Then, according to Equation

5, tSUL MIN ~ 0.3 m.

Figure 24 shows the dependence of the saturation current on the thickness of the SUL. As

the SUL’s thickness is reduced below 0.3 m, the saturation current slowly increases, which can be explained by the partial saturation of the SUL. The substantially more rapid saturation with the thickness reduction below approximately 100 nm can be explained by

the total saturation of the SUL.

Figure 24. Saturation current versus the SUL's thickness. FeAlN-made SuL and 0.8- m wide Ni45Fe55 single

pole head were used.

Figure 25 presents roll-off curves for four different values of the SUL’s thickness, 01,

0.2, 0.3, and 0.4 m, obtained with a 0.8- m wide Ni45Fe55-made recording head. In agreement with the theoretical consideration and conferring with the result in Figure 24, the resolution of the recording system starts to deteriorate as the SUL’s thickness is

reduced below 0.3 m.

One of the consequences of the above discussion is the fact that to achieve a sufficiently

efficient recording process it is not necessary to make a SUL excessively thick. SUL with

thickness values of above 1 m have been routinely reported in literature. Often, it is not trivial to fabricate such a relatively thick SUL without loosing on the short-range

uniformity. This is not an acceptable feature, especially at high areal densities. To

illustrate an example of how the above-described concept can be utilized, a recording system for an areal density of 100 Gbit/in2 can be considered. Assuming a 4:1 bit aspect

ratio (BAR), at 100 Gbit/in2 density, a bit cell would have a 160 x 40 nm2 cross-section.

Assuming a SUL made of a high moment material such as CoFe composition with a


saturation magnetization of 24 kG, for a FeAlN-made recording head of the single pole

type, as analyzed above, with a saturation magnetization of 20 kG, the SUL can be as thin as ~ 70 nm. With allowing for such a thin SUL, it makes the deposition of the SUL to be

a fairly straightforward process even with conventional media deposition tools.

Figure 25. Roll-off curves for FeAlN-made SULs with different thickness values.

3.3. SUL-TO-ABS SEPARATION

Apart from the clear advantages (for a better recording field gradient and others) of

having the SUL as close to the air-bearing surface (ABS) of the recording head as

possible, certain consideration concerning the playback of the recording system should be made while designing a perpendicular recording medium with a SUL. Figure 26 shows

the dependence of the half-width, PW50, and the amplitude of the read sensitivity

function of a 30-nm thick reader as a function of the SUL-to-ABS separation.

Two effects could be identified from Figure 26. As expected, the amplitude of the read

sensitivity function decreases with the increase of the separation between the SUL and the ABS, i.e., the closer the SUL to the ABS is, the higher playback signal could be

expected. This trend suggests minimizing the SUL-to-ABS separation.

The other observation, which is crucial for optimizing the system’s resolution, is that

there exists a maximum in the value of the PW50 as the SUL-to-ABS distance is varied. From Figure 26, it is obvious that a special care should be taken to correctly choose SUL-

to-ABS distance. Shown below experimental data indeed indicate that this effect can

substantially deteriorate the playback performance of a perpendicular recording system with a SUL, thus making it, potentially, even worse than the playback performance of a

perpendicular recording system in which media without a SUL is used.


Figure 26. The dependence of the amplitude of the read sensitivity function and of the half-width PW50 on the

SUL-to-ABS separation.

Figure 27. Roll-off curves for media with and without a SUL.

Figure 27 compares roll-off curves for equivalent media with and without a SUL [171].

In both cases, the recording layer was made of CoCr-based alloy. The drastic effect of using perpendicular media with a SUL (the SUL-to-ABS distance is not optimized) is

clearly observed.

3.4. ANISOTROPY: MICROMAGNETICS OF SUL


This section addresses the micromagnetics of a SUL in the presence of stray fields generated by a recording layer [166,172,173,174]. It is confirmed that relatively high

anisotropy SUL materials need to be utilized to optimize the performance of the

recording system. As was elaborated in Reference [166], when the characteristic bit size

in the recording layer becomes comparable to the characteristic in-plane length, , in the SUL film (which is of the order of the domain wall thickness), the imaging ability of the

SUL deteriorates affecting the playback performance. This characteristic length is

defined via the following expression:

SKUMHAKA 2//~ , (6)

and its value along with other selected material properties for four popular SUL material

candidates is summarized in Table 2. The four materials are: 1) Ni81Fe19 (Permalloy), 2) FeAlN (one of the high moment nitrides), Ni45Fe55 (high moment composition with high

stress induces anisotropy), and Fe65Co35. Here, Hk and Ku are the anisotropy energy

density and the anisotropy field, respectively, 4 Ms is the saturation moment, A is the quantum exchange constant, and LD is the linear density.

Table 2Selected materials properties for four types of SUL: Ni81Fe19 (Permalloy), FeAlN,

Ni45Fe55 (with high stress induces anisotropy), and Fe65Co35. Hk is the anisotropy field,

4 Ms is the saturation moment, A is the quantum exchange constant, and LD is the linear density.

The inability of the SUL to image high frequency spatial variations of the magnetization

in the recording layer causes effective formation of a magnetically “dead” layer at the top of the SUL leading to effective increase of the SUL-to-ABS separation. As shown in

Figure 26, according to 3-D finite element calculations, the increase of the SUL-to-ABS

separation indeed leads to accelerated roll-offs.


The following magnetic force microscopy (MFM) experiment (courtesy of Isabel

Trindade) shows an effect of the existence of a “dead” layer in the SUL deposited on top of a demagnetized 20-nm thick Co/Pd perpendicular recording layer. To generate a

magnetic domain pattern, each Co/Pd thin film used in this experiment was demagnetized

before the SUL was deposited on top of it.

An MFM image of the surface of the demagnetized recording layer (before the SUL was

deposited on top of it) is shown in Figure 28a. Then, two SUL films made of Permalloy were deposited on two of the demagnetized thin films. The SUL films were different in

their thickness: they were 80 and 10 nm thick, respectively. MFM images of the two SUL

films are shown in Figures 28b and c, respectively. It could be noted that the 80-nm-thick SUL is sufficient to screen the demagnetization pattern in the recording layer while the

10-nm-thick SUL apparently fails to do so. This directly indicates that the 10-nm-thick SUL is magnetically “dead”. Other experimental data presented below demonstrate that

the micromagnetic behavior of the SUL leading to the accelerated roll-offs, indeed, takes

place and strongly depends on the anisotropy of the SUL’s material.

Co/Pd-multilayer-based media with close to identical recording layers and different SULs

were tested using spin-stand recording. The similarity of the recording layers was confirmed via measuring M-H-loops and microstructure of the films. All the measured

parameters for the recording layers, such as coercivity, saturation magnetization,

nucleation field, grain size, grain size distribution, etc., are within a 1-2 % error bar for all the recording layers.

(a) (b) (c)

Figure 28. MFM images of (a) a demagnetized Co/Pd perpendicular recording layer, and SUL thin films of

two thickness values, (b) 80 and (c) 10 nm, respectively, deposited on top of the recording layer.

Figure 29 shows the roll-off curves for three media with different SULs. It is clear that

the roll-off of the playback signal is substantially more dramatic if a SUL’s material with

a lower anisotropy field, Hk, is used.


It should be emphasized that the decrease in SUL’s relative permeability from 2000 for

Permalloy to 320 for high anisotropy Ni45Fe55 cannot explain the observed behavior. Boundary element based macroscopic calculations shown that lowering relatively

permeability of the SUL from infinity down to approximately 50 has virtually no effect

on the playback characteristic of the recording system.

Figure 29. Roll-off curves for media with identical recording layers and different SULs.

In summary, as outlined above and in Reference [10], the inability of a SUL with low magnitude of Hk to perfectly image the magnetic bits in the recording layer at a high

linear density may lead to distortions in the playback signal. The deteriorated imaging

results in the loss of resolution (increased PW50 and decreased amplitude of the read sensitivity function) and, subsequently, in the deterioration of the recording system’s

performance overall.

3.5. MAGNETIC BIASING

Among the main technical challenges introduced by the presence of a SUL is the fact that

the SUL contributes an additional source of noise [175]. A not properly optimized SUL

material can introduce a significant amount of noise into the playback signal. The noise is caused by the effective magnetic “charge” created in the numerous domain walls inside

the SUL. A MFM image of a 100-nm-thick SUL with visible domain walls in one of the

diagonal directions is shown in Figure 30. The magnetic charge in these walls is believed


to be the source of the noise. A solution to this problem is the magnetic biasing of the

SUL.

Figure 30. A MFM image of a 100-nm thick soft SUL with magnetic domain walls creating a wave ripple in

on of the diagonal directions.

Figure 31 shows a schematic diagram of the experimental setup to study the noise effects of the magnetic biasing of the SUL. The magnetic biasing can be achieved using two

NdFeB-based permanent magnets places in the vicinity of the recording media. The

placement of the magnets is chosen so that it allows to achieve complete saturation of the SUL underneath the reader. Special care should be taken to arrange the magnets

sufficiently far from the recording head not to affect properties of the read element.

Figure 31. A schematic diagram of the experimental to magnetically bias SUL films.


Figure 32 shows the playback signals from media with magnetically (a) non-biased and (b) biased SUL’s, respectively. As mentioned above, the substantial noise level in the

first case is attributed to the presence of a large number of domain walls. A drastic

reduction of the noise (by at least 10 dB) is clearly observed in the second case (when the SUL is biased). The explanation of this biasing effect is the following. Magnetic biasing

saturates the SUL and thus forces the entire disk into a pseudo-single domain state. In

other words, the magnetic biasing effectively sweeps away the domain walls, which in turn results in the elimination of the SUL’s noise.

Figure 32. Playback signal from two media with different SULs: (a) SUL with a large number of stripe

domains. (The presence of stripe domains was also confirmed with magnetic force microscopy.) (b) Biased

SUL with domain walls swept away from the SUL material.

3.6. DYNAMICS OF PERPENDICULAR RECORDING

Performance of perpendicular recording at high frequencies is a subject that has not been

sufficiently explored [176,177]. The demand for high data rate systems results from the

constantly increasing linear density [178]. Major open questions regarding the data rates in perpendicular recording are believed to be the magnetization switching in the

perpendicular recording layer (due to the relatively low “torque” because of the small angle between the magnetization and the field) and the influence of the soft underlayer

[167]. Previously, it was shown that the characteristic time for the magnetization

switching in the recording layer is the orders of magnitude smaller than the characteristic time for the magnetization switching in soft materials of a recording system [179]. In the

case of perpendicular recording, the soft materials are the head and soft underlayer

materials. Therefore, the head material and the soft underlayer are expected to dominate the (time) frequency dependent roll-off.

The write driver speed raises a serious concern at data rates at which the characteristic time response is of the order of one nanosecond or less. Nevertheless, the question of the

write driver speed limitation is not going to be a subject of this Chapter. Because this


question is as critical in perpendicular recording as it is in longitudinal recording, several

solutions have already been proposed previously [180]. The purpose of this Chapter is to study the intrinsic dynamic response of a perpendicular system with a soft underlayer and

compare it with the intrinsic dynamic response of a typical longitudinal system. Realizing

that the principal difference is caused by the presence of the soft underlayer in perpendicular recording, the study accentuates on the role of the soft underlayer.

Soft underlayer

Yoke

Magnetic flux

(a) (b)

Coil

Trailing

pole

tip

Closed path

definition

Figure 33. A diagram showing a closed magnetic flux path in (a) a longitudinal system and (b) a perpendicular

system with a soft underlayer

Three-dimensional (3D) finite element modeling, including micromagnetics, could be developed to define the design guidelines for high-speed perpendicular systems [181].

Kerr imaging microscopy could be utilized to directly study the magnetic flux dynamics

of a recording system with a soft underlayer [158,182,183].

3.6.1. Design of a High-Data-Rate Perpendicular System

Diagrams showing a typical longitudinal system and a perpendicular system with a soft underlayer are shown in Figures 33a and b, respectively. As mentioned earlier, assuming

that magnetically “hard” materials are typically orders of magnitude faster than “soft” materials in a system, the dominant contribution to the frequency roll-off is going to be

caused by the head inductance. The head inductance consists of two parts, the write coil

inductance and the inductance of the yoke plus the inductance of the soft underlayer in case of perpendicular recording. Effective inductance calculations, incorporating the eddy

current’s effect (resulting in the “skin” effect), could be carried out using 3D finite

element modeling (FEM). The narrow portions of the recording systems, such as the leading and the trailing pole tips in the longitudinal mode and the trailing pole tip and the

soft underlayer in the perpendicular mode, could be modeled micro-magnetically using

Landau-Lifschitz-Gilbert (LLG) Equation [184]. In the modeling described below, the head and the soft underlayer were assumed to be made of a high-moment FeAlN


compound with a saturation magnetization, Bs, of 20 kGauss and an anisotropy field, Hk,

of 10 Oe. A longitudinal ring head design with a trackwidth of 100 nm and a gap length of 60 nm was modeled. An equivalent perpendicular head modeled had a 100 nm wide

trailing pole with a 30 nm separation between the air bearing surface (ABS) and the soft

underlayer. The calculated switching time versus the closed magnetic path length of the perpendicular system with a set of different coil turns is shown in Figure 34. The closed

path is defined as the yoke length (measured at the inner yoke surface, as shown in Figure

33a) plus the length of the soft underlayer under the head, as shown in Figure 33b. The

soft underlayer was modeled to be 300 nm thick. The characteristic switching time, sw,was defined as the time necessary to generate the magnetic field sufficient to overcome

the anisotropy field of the recording layer, Hk. In this work, Hk. of 7 kOe was considered.

It is evident that the inductance of the system should decrease with the reduction of the number of coil turns. However, as the number of turns is reduced, the recording field

generated is also reduced. Fortunately, current state-of-the-art ultra-compact recording

systems, being intrinsically more efficient, require a smaller number of write coil turns [185]. This is especially true for a perpendicular system, which is even more efficient due

to the use of a soft underlayer [186]. For the particular system design, the minimum

number of turns necessary to generate a recording field of a value larger than the hard layer anisotropy field, Hk, was calculated to be two (Two turns generate approximately a

14-kOe-recording field). Although, the total circuit impedance could be optimized

externally, the number of turns around the yoke is an important factor to determine the impedance of the head itself. It could be observed that even a five-turn system is capable

of a switching time of less than one nsec if the magnetic path length is less than 20 m(see Figure 34).

15 20 25 30 35 40 45

1

10

2 turn

5 turn

10 turn

20 turn

Sw

itchin

g t

ime (

nsec)

Magnetic Closed Path Length (um)

Figure 34. The switching time versus the magnetic closed path length for perpendicular systems with 4

different numbers of coil turns: 2, 5, 10 and 20.


The switching time versus the soft underlayer thickness for a 5-turn coil system with a 20

m long flux path is shown in Figure 35a. To compare longitudinal recording and perpendicular recording, the switching time versus the magnetic closed path for an

equivalent longitudinal system with a 5-turn coil is shown in Figure 35b. The switching

time for the longitudinal system was determined as the characteristic time necessary to reverse the magnetization in the longitudinal medium with a coercivity field of 7 kOe. It

could be noted that for the perpendicular system, comparable switching times could be

achieved by adjusting the flux path length.

(a)

200 400 600 800 1000

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Sw

itchin

g t

ime

(n

sec)

Soft underlayer thickness (nm)

(b)

15 20 25 30 35 40 45

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

Sw

itchin

g t

ime (

nsec)

Magnetic closed path length (um)

Figure 35. (a) The switching time versus the soft underlayer’s thickness for a 5-turn coil system with a 20 m

flux path length. (b) The switching time versus the magnetic closed path length for a 5-turn coil longitudinal

system.


3.6.2. Kerr MicroscopyKerr microscopy was developed to directly study the dynamics of a recording system both

with and without a soft underlayer [158]. In this experimental setup, a Kerr image was

taken from a relatively small region under the recording pole tip. In the case of a perpendicular system, the image was taken from the region in the soft underlayer under the

pole tip, as shown in Figure 36.

Soft underlayer

Recording layer

Head

Kerr Imaging Location

Figure 36. A system to study the dynamics of a perpendicular system with a soft underlayer.

A focused ion beam (FIB) image of a focused-ion-beam (FIB)-trimmed perpendicular

head specially fabricated to study the dynamics of a perpendicular system is shown in Figure 37 [186,187].

Leading Pole, P1

P2

Trailing

edge

Gap trench

Figure 37. A FIB image (ABS view) of a FIB fabricated perpendicular head.


In this experiment, a square pulse with a rise time of less than 1 nsec was applied to study the head and soft underlayer response. A time sequence of the polar Kerr images taken

after the field step pulse was applied to a perpendicular system (with a soft underlayer) of

the above-described type is shown in Figure 38a. For comparison, a similar time sequence after the field step pulse was applied to an equivalent longitudinal system of the type

described above is shown in Figure 38b.

(a)

0 nsec 0.5 0.75

1 nsec 1.5 2 nsec

P2Gap

P1

(b)

0 nsec 0.5 0.75

1 1.5 2 nsec

P2

P1

Figure 38. A sequence of polar Kerr images taken at increasing time intervals after a field step pulse was

applied to (a) a perpendicular system with a soft underlayer and (b) an equivalent longitudinal system.

To summarize the results, Figure 39 shows the normalized write current pulse and the Kerr angle response at the point of the maximum field change for the studied longitudinal and

perpendicular systems. It can be observed that the switching times for the two

perpendicular and longitudinal systems are approximately in the same range, 0.75 to 1 nsec.


The above described Kerr imaging experiment clearly illustrates that a perpendicular

system, if optimally designed, should not demonstrate dramatically degraded dynamic performance compared to an equivalent longitudinal system.

0.0 0.5 1.0 1.5 2.0 2.5

-1.0

-0.5

0.0

0.5

1.0

Norm

aliz

ed S

ignal

(Curr

en

t an

d K

err

an

gle

)

Time (nsec)

Write Current

Perpendicular System

Longitudinal System

Figure 39. Normalized current of the write driver and the Kerr angle for the studied perpendicular and

longitudinal systems as a function of time.


REFERENCES

[1] A. S. Hoagland, “High-resolution magnetic recording structures,” IBM Jour., 91-104 (1958).

[2] S. Iwasaki and Y. Nakamura, ”An analysis for the magnetization mode for high density magnetic

recording,“ IEEE Trans. Magn.13, 1272 (1977).

[3] George J. Y. Fan, “Analysis of a practical perpendicular head for digital purposes,” J. Appl. Phys. 31

(5), 402S (1960).

[4] W. Cain, A. Payne, M. Baldwinson, R. Hempstead, “Challenges in the practical implementation of

perpendicular magnetic recording,” IEEE Trans. Magn. 32 (1), 97 (1996).

[5] P. Morton, “Report on the status of contract N70NR-295-XIV, NR-o43-066,” Intermediate Electronic

Computer Project, Univ. California, Berkeley, Sept. 1 – Dec. 1, 1949.

[6] H. Hamilton, “Status report: perpendicular contact recording on rigid media,” J. Magn. Soc. Jap. 18

(S1), 171-8 (1994).

[7] Y. Tanaka, A. Takeo, T. Hikosaka, “Dynamic read/write performance of perpendicular recording,”

IEEE Trans. Magn. 38 (1), 68 (2002).

[8] S.H. Charap, “Thermal stability of recorded information at high densities,” IEEE Trans. Magn. 33 (1),

978 (1997).

[9] D.A. Thompson, “The role of perpendicular recording in the future of hard disk storage,” J. Magn. Soc.

of Japan 21 (S2), 9 (1997).

[10] N. H. Bertram and M. Williams, “SNR and density limit estimates: a comparison of longitudinal and

perpendicular recording,” IEEE Trans. Magn. 36 (1), 4 (1999).

[11] S. Khizroev, M.H. Kryder, and D. Litvinov, “Next generation perpendicular system,” IEEE Trans.

Magn. 37 (4), 1922 (2001).

[12] D. Litvinov, M.H. Kryder, and S. Khizroev, “Recording physics of perpendicular media: soft

underlayers,” J. Magn. Magn. Mater. 232 (1-2), 84 (2001).

[13] S. Khizroev and Y.-K. Liu and K. Mountfield and M. H. Kryder and D. Litvinov, “Physics of

Perpendicular Recording: Write Process,” J. Magn. Magn. Mater. 246, 335 (2002).

[14] D. Litvinov, M.H. Kryder, and S. Khizroev, “Recording physics of perpendicular media: hard layers,”

J. Magn. Magn. Mater. 241 (2-3), 453 (2002).

[15] S. W. Wang, N.X. Sun, M. Yamaguchi, S. Yabukami, “Sandwich films – properties of a new soft

material,” Nature 407 (6801), 150-1 (2000).

[16] S.K. Khizroev and M.H. Kryder and Y. Ikeda and K. Rubin and P. Arnett and M. Best and D.A.

Thompson, “Recording heads with trackwidths suitable for 100Gbit/in2,” IEEE Trans. Magn., 35 (5),

2544 (1999).

[17] D. Litvinov, J. Wolfson, J. Bain, R. Gustafson, M.H. Kryder, and S. Khizroev, “The role of the gap in

perpendicular single pole heads,” IEEE Trans. Magn. 38 (4), 1658-63 (2002).

[18] A. Lyberatos, S. Khizroev, and D. Litvinov, “High speed coherent switching of fine grains,” IEEE

Trans. Magn. 37 (4), 1369 (2001).

[19] A. Lyberatos, D. Litvinov, and S. Khizroev, "Thermal effects in the high-speed switching of the

magnetization of fine grains," Japanese J. Appl. Phys. 42 (4A Part I), 1598-602 (2003).

References 163

[20] S. Khizroev, J. A. Bain, M. H. Kryder, “Considerations in the design of probe heads for 100 Gbit/in2

recording density,” IEEE Trans. Magn., 33 (5), pt.1, 2893-5 (1997).

[21] J.K. Howard, “Effect of nucleation layers on the growth and magnetic properties of CoCr and CoCr-X

films,” J. Vac. Sci. Techn. 4 (6), 2975 (1986).

[22] B. Lu, T. Klemmer, S. Khizroev, J.K. Howard, D. Litvinov, A.G. Roy, and D. Laughlin, “CoCrPtTa/Ti

perpendicular media deposited at high sputtering rate,” IEEE Trans. Magn. 37 (4), 1319 (2001).

[23] T.K. Hatwar and C.F. Brucker, “Coercivity enhancement of Co/Pt superlattices through underlayer

microstructure modification,” IEEE Trans. Magn. 31 (6), 3256 (1995).

[24] D. Litvinov, T. Roscamp, T. Klemmer, M. Wu, J.K. Howard, and S. Khizroev, “Co/Pd Multilayer

Based Recording Layers for Perpendicular Media,” MRS Proceedings T3.9, Vol. 674, (2001).

[25] D. Litvinov, H. Gong, D. Lambeth, J.K. Howard, and S. Khizroev, “Reflection high-energy electron

diffraction based texture determination: magnetic thin films for perpendicular media,” J. Appl. Phys. 87

(9), 5693 (2000).

[26] D. Litvinov, R. Chomko, J. Wolfson, E. Svedberg, J. Bain, R. White, R. Chantrell, S. Khizroev,

“Dynamics of Perpendicular Recording Heads,” IEEE Trans. Magn. 37 (4), 1376 (2001).

[27] J. Wolfson, J. Bain, D. Litvinov, and S. Khizroev J. Appl. Phys. 91 (10), 8665-9 (2002).

[28] D. Litvinov, R.M. Chomko, L. Abelmann, K. Ramstock, G. Chen, S. Khizroev, “Micromagnetics of a

soft underlayer,” IEEE Trans. Magn. 36 (5), 2483 (2000).

[29] R. Wood, J. Miles, and T. Olson, “Recording Technologies for Terabit per square inch Systems,” IEEE

Trans. Magn. 38 (4), 1711 (2002).

[30] M. Mallary, A. Torabi, and M. Benakli, “1 Tb/in2 Perpendicular Recording Conceptual Design,” IEEE

Trans. Magn. Magn. 38 (4), 1711 (2002).

[31] M.H. Kryder, “Future trends in magnetic storage technology,” Digest Book, 2nd (NA) PMRC, January

6-8, 2003, Monterey, California, 68 (2003).

[32] C. T. Rettner, S. Anders, T. Thomson, M. Albrecht, Y. Ikeda, M. E. Best, and B. D. Terris, “Magnetic

characterization and recording properties of patterned Co70Cr18Pt12 perpendicular media,” IEEE Trans.

Magn. Magn. 38 (4), 1725 (2002).

[33] J. Moritz, S. Landis, B. Dieny, A. Lebib, Y. Chen, B. Rodmacq, M. Belin, J. Fontaine, C. Donnet, and

J.P. Nozieres, “Patterned Media Using Pre-Etched Si Wafers Fabricated by Nano-Imprint and e-beam

Lithography,” IEEE Trans. Magn. Magn. 38 (4), 1731 (2002).

[34] S. H. Charap, P.-L. Lu, Y. He “Thermal Stability of Recorded Information at High Densities, “IEEE

Trans. Magn 33 (1), 978 (1997).

[35] MRS Bulletin, p. 387, May 2003.

[36] S. Iwasaki, “Lessons from research of perpendicular magnetic recording,” IEEE Trans. Magn. 39 (4),

1868 (2003).

[37] A. S. Hoagland, “History of magnetic disk storage based on perpendicular magnetic recording,” IEEE

Trans. Magn. 39 (4), 1871 (2003).

[38] F. Liu, K. Stoev, P. Luo, Y. Liu, Y. Chen, J. Chen, J. Wang, S.F.Gu, K.T. Kung, M. Lederman, M.

Krounbi, M. Re, A. Otsuki, S. Hong, “Perpendicular recording heads for extremely high-density

recording,” IEEE Trans. Magn. 39 (4), 1942 (2003).

[39] Private Communications with Mark Kryder, CTO of Seagate Technology (2003).


[40] S. Iwasaki and Y. Nakamura, “An analysis for the magnetization mode for high density magnetic

recording, “IEEE Trans. Magn. 13, 1272 (1977).

[41] D.A. Thompson, “The role of perpendicular recording in the future of hard disk storage,” J. Magn. Soc.

Of Japan 21, Supplement No. S2, 9-14 (1997).

[42] S. Khizroev, M. Kryder, and D. Litvinov, “Next generation perpendicular systems,” IEEE Trans.

Magn., 37 (4), 1922 (2001).

[43] D. Litvinov, M. Kryder, and S. Khizroev, “Recording physics of perpendicular media: soft

underlayers,” JMMM 232 (1-2), 84 (2001).

[44] C. Denis Mee and Eric D. Daniel, editors. Magnetic Recording, Volume III: Video, Audio, and

Instrumentation Recording. McGraw-Hill, New York, 1988.

[45] M. H. Kryder, W. Messner and L. R. Carley, “Approaches to 10 Gbit/in2 recording,” J. Appl. Phys., 79

(8), 4485 (1996).

[46] D. Litvinov, M. Kryder, and S. Khizroev, “Recording physics of perpendicular media: recording

layers,” JMMM 241(2-3), 453 (2002).

[47] M. L. Williams and R. L. Comstock, “Analytical model of the write process in digital magnetic

recording,” 17th Ann. AIP Conf. Proc., 1971.

[48] See www.readrite.com.

[49] D. Litvinov and S. Khizroev, "Orientation-sensitive magnetic force microscopy in future probe storage

applications," Appl. Phys. Lett. 81 (10), 1878 (2002); Editor's choice for the Virtual Journal of

Nanoscale Science & Technology, September 9th.

[50] O. Karlqvist, “Calculation of the magnetic field in the ferromagnetic layer of a magnetic drum,” Trans.

Roy. Inst. Technol., Stockholm, 86, 3-27 (1954).

[51] Amperes, Integrated Inc., Canada.

[52] R. E. Jones, Jr., “Analysis of the efficiency and inductance of multiturn thin film magnetic recording

heads,” IEEE Trans. Magn. 14 (4), 509-11 (1978).

[53] S. Khizroev, Y. Liu, K. Mountfield, M. Kryder, D. Litvinov, “Physics of perpendicular magnetic

recording: writing process,” JMMM 246 (1-2), 335-44 (2002).

[54] M. H. Kryder, W.-Y. Lai, “Modeling of narrow-track thin film write head fields,” IEEE Trans. Magn.

30 (6), 3873-5 (1994).

[55] D. Litvinov, J. Wolfson, J. A. Bain, R. W. Gustafson, M. H. Kryder, and S. Khizroev, "Narrow-gap

single pole heads," IEEE Trans. Magn. 38 (5), 2252-4 (2002).

[56] D. Litvinov, R. Chomko, G. Chen, L. Abelmann, K. Ramstock, S. Khizroev, "Micromagnetics of a Soft

Underlayer," IEEE Trans. Magn., 36 (5), 2483-5 (2001).

[57] M. V. Fedoryuk et al., “Asymptotic analysis: linear ordinary differential equations,” Berlin, Springer,

1993.

[58] Soshin Chikazumi, Physics of Ferromagnetism, Second Edition, Oxford, University Press, 1997.

[59] M. Williams, C. Rettner, K. Takano, and W. Weresin, “Perpendicular write process and head design,”

IEEE Trans. Magn. 38 (4), 1643 (2002).

[60] S. Khizroev and D. Litvinov, "Parallels between playback in perpendicular and longitudinal recording,"

JMMM 257 (1), 126-31 (2003).

[61] M. H. Kryder, “Data storage technologies for advanced computing,” Scientific American 117, 72

(1987).

References 165

[62] E. C. Stoner and E. P. Wohlfarth, "A mechanism of magnetic hysteresis in heterogenous alloys," Phil.

Trans. Royal Soc. A240, 599-642 (1948).

[63] S. Khizroev, M. H. Kryder, Y. Ikeda, K. Rubin, P. Arnett, M. Best, D. A. Thompson, “Recording heads

with trackwidths suitable for 100Gbit/in2 density,” IEEE Trans. Magn., Vol. 35 (5), p.2544, 1999.

[64] H. Hamilton, “Contact recording on perpendicular rigid media,” J. Mag. Soc. of Japan 15 (S2), 491

(1991).

[65] S. Khizroev, W. Jayasekara, J. A. Bain, R. E. Jones, Jr., M. H. Kryder, “MFM quantification of

magnetic fields generated by ultra-small single pole perpendicular heads,” IEEE Trans. Magn., 34 (4),

pt.1, 2030-2 (1998).

[66] Y. Nakamura, H. Muraoka, Y. Shimizu, and T. Inaguma, “Submicron trackwidth recording utilizing a

novel single pole head,” J. Magn. Soc. of Japan 18 (S1), 583 (1994).

[67] W. P. Jayasekara, S. Khizroev, M. H. Kryder, W. Weresin, P. Kasiraj, Fleming, “Inductive write heads

using high moment FeAlN pole,” IEEE Trans. Magn. , 35 (2) pt.1, 613-8 (1999).

[68] D. Litvinov, M.H. Kryder, and S. Khizroev, "Physics of Perpendicular Recording: Playback," Journal

of Applied Physics, 93(11), 9155-64 (2003).

[69] S. Khizroev, A. Lyberatos, M. H. Kryder, and D. Litvinov, "Physics of perpendicular recording: effects

of magnetic "charge" distribution," Japanese J. Appl. Phys., Part 2 Letters 41 (7A), L758-60 (2002).

[70] Private communications with M. Mallary, 1997.

[71] E. Svedberg, D. Litvinov, R. Gustafson, and S. Khizroev, "Magnetic force microscopy of skew angle

dependencies in perpendicular magnetic recording," J. Appl. Phys. 93 (3), 2828-33 (2003).

[72] R. E. Jones, Jr., “Fabrication of film heads with high moment materials,” Acta mater. 46 (11), 3805

(1998).

[73] M. Mallary, A Torabi, and M. Benakli, IEEE Trans. Magn. 38, 1719 (2002).

[74] Y. Kanai, R. Matsubara, H. Wanabe, H. Muraoka, and Y. Nakamura, IEEE Trans. Magn. 39, 1955

(2003).

[75] A.S. Hoagland and J.E. Monson, Digital Magnetic Recording, Wiley, New York, 1991.

[76] M. L. Mallary, S. C. Das, “Vertical magnetic recording arrangement,” U.S. patent No. Re.33949, Jun.

2, 1992.

[77] D. T. Wilton and D. J. Mapps, “An analysis of a shielded magnetic pole for perpendicular recording,”

IEEE Trans. Magn. 29 (6), 4182 (1993).


recording, “IEEE Trans. Magn. 13, 1272 (1977).

[79] George J. Y. Fan, “Analysis of a practical perpendicular head for digital purposes,” J. Appl. Phys. 31

(5), 402S (1960).

[80] H. Takano, Y. Nishida, T. Hamaguchi, F. Tomiyama, Y. Kawato, H. Sawaguchi, F. Tomiyama, Y.

Kawato, H. Sawaguchi, H. Aoi, N. Ohta, “A practical approach for realizing high-recording-density

HDDs by using perpendicular recording,” presented at the AIT-MINT 2001 Workshop on Magnetic

Information Storage Technology, March, 2001.

[81] F. Liu, K. Stoev, L. Leal, J. Wang, Y. Chen, S. Shi, H. Tong, M. Lederman, M. Re, “Perpendicular

high areal densities and high data rates,” presented at the AIT-MINT 2001 Workshop on Magnetic

Information Storage Technology, March, 2001.



Trans. Magn. 33 (1), 978 (1997).

[83] D. A. Thompson and J. S. Best, “The Extendibility of Magnetic Recording for Data Storage,” IBM

Executive Briefing, January 1998.

[84] W. Cain, A. Payne, M. Baldwinson, R. Hempstead, “Challenges in the practical implementation of

perpendicular magnetic recording,” IEEE Trans. Magn. 32 (1), 97 (1996).

[85] S. Khizroev, J. A. Bain, M. H. Kryder, “Considerations in the design of probe heads for 100 Gbit2

recording density,” IEEE Trans. Magn., 33 (5), pt.1, 2893 (1997).

[86] D. Litvinov, M. Kryder, and S. Khizroev, “Recording physics of perpendicular media: soft underlayer,”

JMMM 232 (1-2), 84-90 (2001).

[87] D. Litvinov, M. Kryder, S. Khizroev, “Recording physics of perpendicular media: recording layer,”

JMMM 232, 84 (2001).




with trackwidths suitable for 100 Gbit/in2 density, “ presented at Intermag’99, May 1999, IEEE Trans.

Magn. 35 (5), 2544 (1999).

[90] C. Kittel, Rev. Mod. Phys., 21, 541 (1949).


Magn. 37 (4), 1922 (2001).

[92] Amperes, the boundary element solver by Integrated Engineering Software, Winnipeg, Canada.

[93] K. Ouchi, “Recent advancements in perpendicular magnetic recording,” IEEE Trans. Magn. 37 (4),

1217 (2001).

[94] D. Litvinov, M. H. Kryder, S. Khizroev, “Recording physics of perpendicular media: recording layer,”

JMMM 241(2-3), 453-465 (2002).

[95] D. E. Heim, “The sensitivity function for shielded magnetoresistive heads by conformal mapping,”

“IEEE Trans. Magn. 19 (5), 1620 (1983).


magnetic fields generated by ultra-small single pole perpendicular heads,” IEEE Trans. Magn. 34 (4),

pt.1, 2038 (1998).

[97] S. A. S. Hoagland, “Digital Magnetic Recording”, Wiley and Sons, Inc., 1963.

[98] R. I. Potter, IEEE Trans. Magn. 10, 502 (1974).

[99] N. Smith, D. Wachenshwantz, “Magnetoresistive heads and the reciprocity principle,” IEEE Trans.

Magn. 23 (5), 2494 (1987).

[100] H. Muraoka, Y. Sugita, Y. Nakamura, “Simplified expression of shielded MR head response for

double-layer perpendicular medium,” IEEE Trans. Magn. 35 (5), 2235 (1999).


1993.

[102] R.P. Hunt, “A magneto-resistive readout transducer,” IEEE Trans. Magn. 7, 1 (1971).

[103] D.A. Thompson, “Magnetoresistive transducers in high-density magnetic recording,” AIP Conf. Proc.

24, 528 (1975).

References 167

[104] D.A. Thompson, L.T. Romankiw, A.F. Mayadas, “Thin film magnetoresistors in memory storage and

related applications,” IEEE Trans. Magn. 11 (4), 1039-50 (1975).

[105] R.S. Indeck, J.H. Judy, and S. Iwasaki, “A magnetoresistive Gradiometer,” IEEE Trans. Magn. 24 (6),

2617 (1988).

[106] H.S. Gill, V.W. Hesterman, G.J. Tarnopolsky, L.T. Tran, P.D. Frank, H. Hamilton, “A

magnetoresistive gradiometer for detection of perpendicularly recording magnetic transitions,” J. Appl.

Phys. 65 (1), 402 (1989).

[107] D.J. Mapps, “An inter-digitated magnetoresistive reply head,” U.K. Patent 2,143,071, October 28,

1987.

[108] R.E. Jones, Jr., M.H. Kryder, K.R. Mountield, J.I. Guzman, “Unshielded horizontal magnetoresistive

head and method of fabricating same,” US Patent 5,155,643, October 13, 1992.

[109] J.C. Mallinson, “Gradiometer Head Pulse Shapes,” IEEE Trans. Magn. 26 (2), 1123 (1990).

[110] G.J. Fan, “A study of the playback process of a magnetic ring head,” IBM J. of Res. & Dev. 5, 321-325

(1961).


Magn. 23 (5), 2494 (1987).

[112] S. Khizroev, J. A. Bain, M. H. Kryder, “Considerations in the design of probe heads for 100 Gbit2

recording density,” IEEE Trans. Magn. 33 (5), pt.1, 2893-5 (1997).

[113] M. Mallary, A. Torabi, M. Benakli, “1 Tb/in2 Perpendicular Recording Conceptual Design,” IEEE

Trans. Magn. 38, 1719 (2002).

[114] S. Khizroev and D. Litvinov, “Parallels between playback in perpendicular and longitudinal recording,”

JMMM 257 (1), 126-31 (2003).

[115] K. Klaassen, J. T. Contreras, J.C.L. van Peppen, “Read/write electronics front-end systems for hard

disk drives,” IEEE Trans. Magn. 40 (1), 263 (2004).

[116] A. M. Taratorin, M. Xiao, K. B. Klaassen, “Correlation of dc noise and transition noise in magnetic


[117] W. Eppler, A. Sunder, D. Karns, E. Kurtas, G. Ju, P. van der Heijden, Y. Kubota, H. Zhou, C.-H.

Chang, “Perpendicular recording near 100 Gbit/in2,” IEEE Trans. Magn. 39 (2), 663 (2003).

[118] P.A. Van der Heijden, D.W. Karns, T.W. Clinton, S.J. Heinrich, S. Batra, D.C. Karns, T.A. Roscamp,

E.D. Boerner, W.R. Eppler, “The effect of media background reading and writing in perpendicular

recording,” GB-09, MMM, Seattle, WA, 2001

[119] M. F. Erden, I. Ozgunes, E. M. Kurtas, W. Eppler, “General transform filters in perpendicular

recording architecture,” IEEE Trans. Magn. 38 (5), 2334 (2002).

[120] B.M. Lairson, K.Ho, and S. Sun, “PRML recording performance of multilayer Pt/Co alloy media,” J.

Magn. Soc. Japan 22 (S2), 135 (1997).

[121] I. Ozgunes, W. Eppler, “Synchronization-free dibit response extraction from PRBS waveforms,” IEEE

Trans. Magn. 39 (5), 2225 (2003).

[122] R. I. Potter, IEEE Trans. Magn. 10, 502 (1974).


1993.

[124] Schwarz and S. K. Decker, IEEE Trans. Magn. 15, 1622 (1979).



Magn. 23 (5), 2494 (1987).


magnetic fields generated by ultra-small single pole perpendicular heads,” IEEE Trans. Magn. 34 (4)

pt.1, 2030 (1998).

[127] D. E. Heim, “The sensitivity function for shielded magnetoresistive heads by conformal mapping,”

IEEE Trans. Magn. 19 (5), 1620 (1983).

[128] K. B. Klaassen and J. C. L. van Peppen, "Magnetic instability of thin-film recording heads," IEEE

Trans. Magn. 30, 375 (1994).

[129] D.J. Mapps, “An inter-digitated magnetoresistive reply head,” U.K. Patent 2,143,071, October 28,

1987.

[130] R.E. Jones, Jr., M.H. Kryder, K.R. Mountield, J.I. Guzman, “Unshielded horizontal magnetoresistive

head and method of fabricating same,” US Patent 5,155,643, October 13, 1992.


Trans. Magn., 33 (1), 978-83 (1997).


recording, “ IEEE Trans. Magn., 13, 1272 (1977).

[133] D. A. Thompson and J. S. Best, “The Extendibility of Magnetic Recording for Data Storage,” IBM

executive briefing, January 1998.

[134] K. Ouchi, Magnetic Materials for information storage devices, in: V.G. Bar’yakhtar et al. (Eds.),

Frontiers in Magnetism of Reduces Dimension Systems, 331-68 (1998).

[135] S. Iwasaki, Y. Nakamura & K. Ouchi, “Perpendicular magnetic recording with a composite anisotropy

film,” IEEE Trans. Magn., 15 (6), 1456 (1979).

[136] D. Litvinov, M. Kryder, and S. Khizroev, “Recording physics of perpendicular media: recording

layers,” JMMM 241(2-3), 453-465 (2002).

[137] D. Litvinov, M. Kryder, and S. Khizroev, “Recording physics of perpendicular media: soft

underlayers,” JMMM 232 (1-2), 84-90 (2001).

[138] Y. Tanaka, K. Shimomura, A. Takeo, “A study of perpendicular recording integration,” Technical

Report of the Institute of Electronics, Information and Communication Engineers (IEICE), 2002.

[139] K. Ouchi, N. Honda, “Overview of latest work on perpendicular recording media,” IEEE Trans. Magn.

2000, presented at TMRC’99.


Magn., 37 (4), 1922-4 (2001).

[141] J. Howard, J. Vac. Sci. Technol. A 4 (6), 2975 (1986).

[142] S. Nakagawa, S. Akiyama, M. Naoe, J. Magn. Soc. Jpn. 15 (Suppl. (S2)), 657 (1991).

[143] W. Peng, R. Victora, J. Judy, K. Gao, J. Sivertsen, J. Appl. Phys. 87 (9), 6358 (1991).

References 169

[144] L. Wu, S. Yanase, N. Honda, and K. Ouchi, “Magnetic properties of Co/Pd multilayers sputter-

deposited at high Ar gas pressure,” J. Magn. Soc. Jpn. 21, 301 (1997).

[145] T. Ando and T. Nishihara, “Experimental study of signal stability ontriple-layer perpendicular

recording media,” J. Magn. Soc. Japan 21 (S2), 522 (1997).

[146] Y. Hirayama, M. Futamoto, K. Kimoto, and K. Usami, IEEE Trans. Magn. 32, 3807 (1996).

[147] J. Araike, N. Honda, T. Keitoku, K. Ouchi, and S. Iwasaki, J. Magn. Magn. Mat. 155, 228 (1996).

[148] H. Hokkyo, S. Tsuboi, N. Ohshima, and K. Tagami, “Signal stability of double-layer perpendicular

magnetic recording media,” Proc. 1998 IEICE Ge. Conf., C7-2 (in Japanese), 1998.

[149] M. H. Kryder, “Data storage technologies for advanced computing,” Scientific American 117, 72

(1987).

[150] C. Denis Mee and Eric D. Daniel, “Magnetic recording,” published by McGraw-Hill, 1996.

[151] B. M. Lairson, M. R. Visokay, R. Sinclair, and B. M. Clemens, Appl. Phys. Lett. 62, 639 (1993).

[152] M. Watanabe and M. Homma, Jpn. J. Appl. Phys., Part 2 35, L1 264 (1996).

[153] J.-U. Thiele, L. Folks, M. F. Toney, and D. K. Weller, J. Appl. Phys. 84, 5686 (1998).

[154] L. Wu, N. Honda, K. Ouchi, IEEE Trans. Magn. 35 (5), 2775 (1999).

[155] T. Ando and T. Nishihara, “Triple-layer perpendicular recording media for high SN ratio and signal

stability,” IEEE Trans. Magn. 33, 2983 (1997).

[156] Y. Hirayama, K. Ito, Y. Honda, and M. Futamoto, “Magnetization decay in CoCr-alloy perpendicular

magnetic recording media,” J. Magn. Soc. Jpn. 21 (S2), 297 (1997).

[157] N. Honda, T. Kiya, and K. Ouchi, “Correspondence of medium noise and recording characteristics to

M-H loop in perpendicular recording media,” J. Magn. Soc. Jpn. 21 (S2), 505 (1997).

[158] S. Tamaru, J. Bain, R. van de Veerdonk, T. Crawford, M. Cavington, M. Kryder, “Measurement of

magnetostatic mode excitation and relaxation in Permalloy films using scanning Kerr microscopy,”

Intermag 2003, digest of technical papers, EA4 (2002).

[159] J. Wolfson, J. Bain, D. Litvinov, and S. Khizroev, “Dynamic Kerr imaging of soft underlayers (SUL’s)

for perpendicular magnetic recording applications,” J. Appl. Phys. 91(10), 8665-9 (2002).

[160] P. A. A. van der Heijden, D. W. Karns, T. W. Clinton, S. J. Heinrich, S. Batra, D. C. Karns, T. A.

Roscamp, E. D. Boerner, and W. R. Eppler, “The effect of media background on reading and writing in

perpendicular recording,” J. Appl. Phys. 91, 8372 (2002).

[161] X. Chen, M. H. Kryder, “Thermal stability of Co/Pt multilayers,” J. Appl. Phys. 85 (8), 5006 (1999).

[162] S. Khizroev, A. Lyberatos, M. H. Kryder, and D. Litvinov, "Physics of perpendicular recording: effects

of magnetic "charge" distribution," Japanese J. Appl. Phys., Part 2 Letters 41 (7A), L758-60 (2002).


[163] S. Chikazumi: Physics of ferromagnetizm, Oxford: Oxford (1997) 2nd edition.

[164] Amperes, Integrated Engineering Software, Inc., Winnipeg, Canada.

[165] D. Litvinov, M.H. Kryder, and S. Khizroev, "Physics of Perpendicular Recording: Playback," Journal

of Applied Physics, 93(11), 9155-64 (2003).

[166] D. Litvinov, R. Chomko, G. Chen, L. Abelmann, K. Ramstock, S. Khizroev, "Micromagnetics of a Soft

Underlayer," IEEE Trans. Magn., 36 (5), 2483-5 (2001).

[167] A. Lyberatos, D. Litvinov, and S. Khizroev, "Thermal effects in the high-speed switching of the

magnetization of fine grains," Japanese J. Appl. Phys. 42 (4A Part I), 1598-602 (2003).

[168] J. Ariake, N. Honda, K. Ouchi, and S. Iwasaki, “Preparation of double layer perpendicular media for

fine magnetic structure,” IEEE Trans. Magn. 36, (2000).

[169] Y. Kanai, R. Matsubara, IEEE Trans. Magn. 37 (4), 1357 (2001).



[171] S. Khizroev, J. A. Bain, M. H. Kryder, “Considerations in the design of probe heads for 100 Gbit/in2

recording density,” IEEE Trans. Magn., 33 (5), pt.1, 2893-5 (1997).

[172] J. Zhu, Y. Zheng, and X. Lin, “Micromagnetics of small size patterned exhange biased permalloy film

elements,” J. Appl. Phys, 81 (8), 4336 (1997).

[173] N. Bertram, X. Wang, H. Zhou, V. Safonov, IEEE Trans. Magn. 37 (4), 1521 (2001).

[174] A. Hubert, R. Schafer, Magnetic domains: the Analysis of Magnetic Microstructures, Springer, Berlin,

Heidelberg, New York, 1998, p. 125.

[175] W. Cain, et al, “Challenges in the practical implementation of perpendicular magnetic recording,” IEEE

Trans. Magn. 32, 97 (1996).

[176] K. Klaassen, J. T. Contreras, J.C.L. van Peppen, “Read/write electronics front-end systems for hard

disk drives,” IEEE Trans. Magn. 40 (1), 263 (2004).

[177] A. M. Taratorin, M. Xiao, K. B. Klaassen, “Correlation of dc noise and transition noise in magnetic


[178] W. Eppler, A. Sunder, D. Karns, E. Kurtas, G. Ju, P. van der Heijden, Y. Kubota, H. Zhou, C.-H.

Chang, “Perpendicular recording near 100 Gbit/in2,” IEEE Trans. Magn. 39 (2), 663 (2003).

[179] I. Tudosa, C. Stamm, A. B. Kashuba, F. King, H. C. Siegmann, J. Stohr, G. Ju, B. Lu, D. Weller, “The

ultimate speed of magnetic switching in granular recording media,” Nature 428, 831 (2004).

[180] K. Klaassen, J.C.L. van Peppen, “Magnetic media switching dynamics in thermal equilibrium,” IEEE

Trans. Magn. 37 (4), 1537 (2001).

References 171

[181] D. Litvinov, J. Wolfson, J. Bain, R. White, R. Chomko, R. Chantrell, and S. Khizroev, “Dynamics of

perpendicular recording,” IEEE Trans. Magn. 37 (4), 1376 ( 2001).

[182] D. Litvinov, A. Lyberatos, M. H. Kryder, J. Wolfson, J. A. Bain, and S. Khizroev, "Recording layer

influence on the dynamics of perpendicular recording," IEEE Trans. Magn 38 (5), 1994 (2002).

[183] D. Litvinov, J. Wolfson, J. A. Bain, R. W. Gustafson, M. H. Kryder, and S. Khizroev, "Narrow-gap

single pole heads," IEEE Trans. Magn. 38 (5), 2252 (2002).

[184] L. D. Landau and E. M. Lifschitz, Phys. Z. Sowjet., 8, 153 (1935).

[185] E. Valstyn, L. Shew, “Performance of single-turn film heads,” IEEE Trans. Magn. 9 (3), 317 (1973).


with trackwidths suitable for 100 Gbit/in2 density, “IEEE Trans. Magn., 35 (5), 2544-6 (1999).

[187] S. Khizroev, D. Litvinov, “Focused-ion-beam-based rapid prototyping of Nanoscale magnetic devices,”

Review in Nanotechnology 14, R7-15 (2004).

173

INDEX

Adjacent tracks, 9, 19, 41, 72, 92, 133

Air bearing surface (ABS), 35

AC demagnetized media, 77, 92

Areal density, 3

Barrium ferrite (BaFe), 128

Biasing of SUL noise, 13, 153

Bit (data) rate, 155

Bit transition, 3, 9

Boundary element modeling (BEM), 28, 74, 89,

120

Censtor heads, 1, 38

Channel (of data), 17

Closed (flux) loop (path), 29, 156, 157

CoCr- based media, 4, 9, 18

Co/Pt-multilayer based media, 4, 18

Coding (encoding), 2, 136

Coercivity, 22

Coercivity squareness, 75, 128

Conformal Mapping, 120

Data rate (see Bit Rate)

DC erase, 74

DC noise, 19.

Demagnetization field, 4, 11, 140

Demagnetized media (see AC demagnetized

media)

Detection (See Signal)

Differential equations, 31, 86, 101

Differential reader, 104

Domain, 13, 41, 152, 155

Domain noise, 125

Domain wall, 12, 153,

Domain width (thickness), 41, 131, 151

Dynamics (of recording), 127, 156, 159

Eddy current, 156

Efficiency, 6, 24, 28, 43, 49, 63,

Encoding (See Coding)

Erasure, 20, 130

Erase band, 8

FeAlN (high moment nitrides), 14, 68, 146, 151

Finite-element calculation, 156

Fly height (See Flying height)

Flying height, 15, 28, 42, 70

Focused ion beam (FIB), 8, 38, 40, 68

Frequency roll-off (See Roll-off)

Gap, 5, 8, 27

Giant magnetoresistive (GMR) head, 16, 67, 118,

124

Gradient (of the magnetic field), 5, 8, 14, 23, 29,

42

Grain size, 3, 18, 128, 131, 152

Heat-assisted magnetic recording (HAMR), 22

Head (see Recording head,

Single pole head (SPH),

Ring head (RH), Read head,

Magnetoresistive (MR) head,

GMR head, Write head,

Shielded head)

High-density recording, 1

Kerr (imaging), 19, 127, 130, 156, 159

L10 phases of FePt and others, 19, 128

Laplace’s Equation, 31, 101, 122, 131, 144

Linear density, 20, 54, 86, 100

Longitudinal field, 4, 5, 27,

Longitudinal recording, 1

Magnetic flux, 14, 30, 35, 43, 57, 74

Magnetic force microscopy (MFM), 9, 27, 39, 53,

143

Magnetic image, 31, 72, 101

Magnetic moment, 3

Magnetization, 2

Magnetometer, 19, 75

Magneto-optical media, 131

Magnetoresistive (See Magnetoresistive head)

Magnetoresistive head, 2, 121, 125, 171

Maxwell’s Equations, 31, 101, 145

MH-loop, 19

“Mirror” imaging, 7, 31, 73, 101

Multiplicative (magnetic reflection), 74

Noise (See DC noise, AC noise,

Signal-to-noise ratio)

Non-linear transition shift (NLTS), 72

174

Overwrite, 143

Particle (grain) size, 3, 18, 128

Patterned media, 22,

Perpendicular recording, 1

Playback head (See Read head)

Principle of Reciprocity, 86, 92, 98,

Read head, 15, 54, 67, 86, 99

Reciprocity Principle (See Principle of

Reciprocity)

Ring head, 6, 9, 24, 27, 56, 83

Roll-off curve, 21, 85, 92, 149

Saturation magnetization, 4, 19, 24, 35, 86, 128

Scalar (magnetic) potential, 99, 121,

Scaling law, 3, 42

Scanning electron microscopy (SEM), 38, 67

Schwartz-Christoffel Transformation, 120

Self-erasure (See Erasure)

Sensitivity field, 98,

Shielded head, 79, 99

Shields (magnetic), 23, 80

Side (track) reading, 54

Single domain wall (See Domain)

Signal (waveform, half-width PW50), 3, 11, 16,

71

Signal-to-noise ratio (SNR), 3, 20, 131

Single pole head (SPH), 5, 24, 31, 38, 42

“Skin” effect, 156

Soft (magnetic) underlayer (SUL), 5, 12, 15

Spinstand (measurements), 21

Spacing loss, 103

Squareness (See Coercivity squareness)

Stability (thermal) (See Thermal stability)

Superparamagnetic limit (See Thermal stability)

Surface (magnetic) charge, 17, 32, 70

Thermal stability, 1, 3, 8

Time sequence, 160

Track density, 9

Transmission electron microscopy (TEM), 3, 129

Trapezoidal (write pole), 55

Vibrating sample magnetometer (VSM), 75

Vertical recording (see Perpendicular recording)

Stoner-Wohlfarth mode, 37

X-Ray diffraction, 142

perpendicular magnetic recording

Documents