perpendicular magnetic recording
TRANSCRIPT
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
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
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
1.1 Chapter overview 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
1.1 Chapter overview 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
This page intentionally left blank
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
This page intentionally left blank
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-
Perpendicular Magnetic Recording 4
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].
Chapter 1 Fundamentals of Perpendicular Recording 5
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
Perpendicular Magnetic Recording 6
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.
Chapter 1 Fundamentals of Perpendicular Recording 7
(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
Perpendicular Magnetic Recording 8
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
Chapter 1 Fundamentals of Perpendicular Recording 9
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
Perpendicular Magnetic Recording 10
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
Chapter 1 Fundamentals of Perpendicular Recording 11
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.
Perpendicular Magnetic Recording 12
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.
Chapter 1 Fundamentals of Perpendicular Recording 13
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
Perpendicular Magnetic Recording 14
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
Chapter 1 Fundamentals of Perpendicular Recording 15
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.
Perpendicular Magnetic Recording 16
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
Chapter 1 Fundamentals of Perpendicular Recording 17
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
Perpendicular Magnetic Recording 18
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
Chapter 1 Fundamentals of Perpendicular Recording 19
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
Perpendicular Magnetic Recording 20
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).
Chapter 1 Fundamentals of Perpendicular Recording 21
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
Perpendicular Magnetic Recording 22
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.
Perpendicular Magnetic Recording 24
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
TransitionWritten moment
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
Perpendicular Magnetic Recording 26
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
TransitionWritten moment
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.
Chapter 2 Physics of Writing 27
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.
Perpendicular Magnetic Recording 28
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)
Chapter 2 Physics of Writing 29
-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
Perpendicular Magnetic Recording 30
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)
Chapter 2 Physics of Writing 31
-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
Perpendicular Magnetic Recording 32
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
Chapter 2 Physics of Writing 33
(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)
Perpendicular Magnetic Recording 34
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.
Chapter 2 Physics of Writing 35
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.
Perpendicular Magnetic Recording 36
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
longitudinal recording.
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,
Chapter 2 Physics of Writing 37
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.
Perpendicular Magnetic Recording 38
-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.
Chapter 2 Physics of Writing 39
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.
Perpendicular Magnetic Recording 40
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.
Chapter 2 Physics of Writing 41
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.
Perpendicular Magnetic Recording 42
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.
Chapter 2 Physics of Writing 43
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.
Perpendicular Magnetic Recording 44
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)
Chapter 2 Physics of Writing 45
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.
Perpendicular Magnetic Recording 46
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
Distance down the track ( m)
(a)
Chapter 2 Physics of Writing 47
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
Distance down the track ( m)
(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)
Perpendicular Magnetic Recording 48
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
Distance across the track ( m)
(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)
Distance across the track ( m)
(a)
Chapter 2 Physics of Writing 49
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)
Distance across the track ( m)
(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
Perpendicular Magnetic Recording 50
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
Chapter 2 Physics of Writing 51
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.
Perpendicular Magnetic Recording 52
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)
Chapter 2 Physics of Writing 53
(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.
Perpendicular Magnetic Recording 54
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.
Chapter 2 Physics of Writing 55
-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.
Perpendicular Magnetic Recording 56
0.9 1.0 1.1 1.20
4000
8000
12000
16000
PT = 0.1um
PT = 0.5um
Hz (
Oe
)
Distance along the track (um)
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).
Chapter 2 Physics of Writing 57
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)
Distance down the track ( m)
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.
Perpendicular Magnetic Recording 58
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)
Distance down the track ( m)
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.
Chapter 2 Physics of Writing 59
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)
Distance down the track ( m)
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.
Perpendicular Magnetic Recording 60
0.2 0.4 0.6 0.8 1.0 1.2
0
5
10
15
20
Type I SPHTrailing
EdgeH
Z (
kO
e)
Distance down the track ( m)
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.
Chapter 2 Physics of Writing 61
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
Perpendicular Magnetic Recording 62
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.
Chapter 2 Physics of Writing 63
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
Perpendicular Magnetic Recording 64
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.
Chapter 2 Physics of Writing 65
~
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
Perpendicular Magnetic Recording 66
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
Chapter 2 Physics of Writing 67
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.
Perpendicular Magnetic Recording 68
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.
Chapter 2 Physics of Writing 69
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.
Perpendicular Magnetic Recording 70
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
Chapter 2 Physics of Writing 71
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)
Perpendicular Magnetic Recording 72
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).
Chapter 2 Physics of Writing 73
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.
Perpendicular Magnetic Recording 74
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
Chapter 2 Physics of Writing 75
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
Perpendicular Magnetic Recording 76
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
Chapter 2 Physics of Writing 77
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)
Distance down the track (um)
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
Distance down the track (um)
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
Perpendicular Magnetic Recording 78
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.
Chapter 2 Physics of Writing 79
0.0 0.2 0.4 0.6 0.8
0
2
4
6
8
10
12
Distance down the track (um)
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)
Distance down the track (um)
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
Perpendicular Magnetic Recording 80
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
Chapter 2 Physics of Writing 81
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
Perpendicular Magnetic Recording 82
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)
Distance across the track ( m)
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
Chapter 2 Physics of Writing 83
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.
Perpendicular Magnetic Recording 84
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
Perpendicular Magnetic Recording 86
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
Perpendicular Magnetic Recording 88
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].
+
+
charges in the transition
+
+
+ + + + + + + + - - - - - - - - - -
- - - - - - - - - - + + + + + + +
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 3 Physics of Playback 89
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)
Distance along the track (um)
(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)
Distance along the track (um)
(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.
Perpendicular Magnetic Recording 90
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.
Chapter 3 Physics of Playback 91
(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.
Perpendicular Magnetic Recording 92
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
Linear Density (kfci)
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
Chapter 3 Physics of Playback 93
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.
Perpendicular Magnetic Recording 94
(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
Chapter 3 Physics of Playback 95
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].
Perpendicular Magnetic Recording 96
(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
Areal density (Gbit/in2)
with SUL
Str
ay F
ield
, H
z (
Oe
)
Areal Density (Gbit/in2)
BAR = 1:1
BAR = 2:1
BAR = 4:1
BAR = 8:1
(c)
0 500 1000 1500 2000
200
400
600
800
Hz (
Oe)
Areal Density (Gbit/in2)
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.
Chapter 3 Physics of Playback 97
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.
Perpendicular Magnetic Recording 98
0 500 1000 1500 2000
0.2
0.4
0.6
0.8
1.0H
z n
orm
aliz
ed
Areal density (Gbit/in2)
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,
Chapter 3 Physics of Playback 99
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.
Perpendicular Magnetic Recording 100
200 400 600 800 1000
0
5
10
ABS-to-SUL = 30nmFlying Height = 5 nm
ideal SUL No SUL
Linear Density (kfci)
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.
Chapter 3 Physics of Playback 101
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.
Perpendicular Magnetic Recording 102
(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
ABS to SUL distance (nm)
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
Chapter 3 Physics of Playback 103
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
)
ABS to SUL distance (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.
Perpendicular Magnetic Recording 104
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.
Chapter 3 Physics of Playback 105
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.
Perpendicular Magnetic Recording 106
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.
Chapter 3 Physics of Playback 107
-200 -100 0 100 200
-2
-1
0
1
2
3
4
5
Hx (
a.u
.)
Distance along the track (nm)
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
)
Linear Density (kfci)
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.
Perpendicular Magnetic Recording 108
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
.)
Linear Density (kfci)
No shield (no sul)
Shield (no sul)
Diff (no sul)
Shield 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.
Chapter 3 Physics of Playback 109
-200 -100 0 100 200-6
-4
-2
0
2
4
6
Sensitiv
ity F
ield
(a.u
.)
Distance along the track (nm)
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
.)
Distance along the track (nm)
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.
Perpendicular Magnetic Recording 110
0 500 1000 1500 2000 2500 3000
0
5
10
15
20
25
30
35
Pla
yback (
a.u
.)
Linear Density (kfci)
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
.)
Linear Density (kfci)
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
Chapter 3 Physics of Playback 111
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
.)
Linear Density (kfci)
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
.)
Linear Density (kfci)
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.
Perpendicular Magnetic Recording 112
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
.)
Distance along the track (nm)
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
.)
Linear Density (kfci)
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.
Chapter 3 Physics of Playback 113
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
.)
Linear Density (kfci)
Shield (sul)
Shield (no sul)
Diff (sul)
Diff (no sul)
Shield Diff (sul)
Shield Diff (no 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.
Perpendicular Magnetic Recording 114
0 500 1000 1500 2000 2500
0
2
4
6
8
10
12
14
16
Pla
yback (
a.u
.)
Linear Density (kfci)
Shield (no sul)
Diff (no sul)
Shield 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.
Chapter 3 Physics of Playback 115
Recording MediumRecording Medium
MR
Sensor
MR
Sensor
(c) (d)
shield shield
Recording MediumRecording Medium
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
.)
Distance along the track (nm)
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.
Perpendicular Magnetic Recording 116
-200 -100 0 100 200
-2
0
2
4
6
8
10
Hx (
a.u
.)
Distance along the track (nm)
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
.)
Distance along the track (nm)
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.
Chapter 3 Physics of Playback 117
-200 -100 0 100 200
-2
0
2
4
6
8
Hx (
a.u
.)
Distance along the track (nm)
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
.)
Linear Density (kfci)
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.
Perpendicular Magnetic Recording 118
0 500 1000 1500 2000 2500
0
5
10
15
20
Pla
yback
(a.u
.)
Linear Density (kfci)
Diff (no sul)
Shield 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
longitudinal recording.
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].
Chapter 3 Physics of Playback 119
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)
Distance down the track ( m)
(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.
Perpendicular Magnetic Recording 120
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
Chapter 3 Physics of Playback 121
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
Perpendicular Magnetic Recording 122
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].
Chapter 3 Physics of Playback 123
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)
Perpendicular Magnetic Recording 124
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.
Chapter 3 Physics of Playback 125
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
Linear Density (kfci)
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.
Perpendicular Magnetic Recording 126
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.
Perpendicular Magnetic Recording 128
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
Perpendicular Magnetic Recording 130
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.
Chapter 4 Perpendicular Recording Media 131
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].
Perpendicular Magnetic Recording 132
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.
Chapter 4 Perpendicular Recording Media 133
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
Perpendicular Magnetic Recording 134
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.
Chapter 4 Perpendicular Recording Media 135
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
Perpendicular Magnetic Recording 136
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.
Chapter 4 Perpendicular Recording Media 137
+
+
charges in the transition
+
+
+ + + + + + + + - - - - - - - - - -
- - - - - - - - - - + + + + + + +
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
Perpendicular Magnetic Recording 138
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
Chapter 4 Perpendicular Recording Media 139
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.
Perpendicular Magnetic Recording 140
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.
Chapter 4 Perpendicular Recording Media 141
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
Perpendicular Magnetic Recording 142
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
Chapter 4 Perpendicular Recording Media 143
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.
Perpendicular Magnetic Recording 144
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
Chapter 4 Perpendicular Recording Media 145
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)
Perpendicular Magnetic Recording 146
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),
Chapter 4 Perpendicular Recording Media 147
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
Perpendicular Magnetic Recording 148
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
Chapter 4 Perpendicular Recording Media 149
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.
Perpendicular Magnetic Recording 150
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
Chapter 4 Perpendicular Recording Media 151
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.
Perpendicular Magnetic Recording 152
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.
Chapter 4 Perpendicular Recording Media 153
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
Perpendicular Magnetic Recording 154
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.
Chapter 4 Perpendicular Recording Media 155
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
Perpendicular Magnetic Recording 156
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
Chapter 4 Perpendicular Recording Media 157
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.
Perpendicular Magnetic Recording 158
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.
Chapter 4 Perpendicular Recording Media 159
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.
Perpendicular Magnetic Recording 160
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.
Chapter 4 Perpendicular Recording Media 161
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.
Perpendicular Magnetic Recording 162
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).
Perpendicular Magnetic Recording 164
[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).
[78] S. Iwasaki and Y. Nakamura, “An analysis for the magnetization mode for high density magnetic
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.
Perpendicular Magnetic Recording 166
[82] S. H. Charap, P.-L. Lu, Y. He “Thermal Stability of Recorded Information at High Densities, “IEEE
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).
[88] S. Khizroev, Y. Liu, K. Mountfield, M. Kryder, D. Litvinov, “Physics of perpendicular magnetic
recording: writing process,” JMMM 246 (1-2), 335-44 (2002).
[89] S. Khizroev, M. H. Kryder, Y. Ikeda, K. Rubin, P. Arnett, M. Best, D. A. Thompson, “Recording heads
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).
[91] S. Khizroev, M. Kryder, and D. Litvinov, “Next generation perpendicular systems,” IEEE Trans.
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).
[96] 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, 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).
[101] M. V. Fedoryuk et al., “Asymptotic analysis: linear ordinary differential equations,” Berlin, Springer,
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).
[111] N. Smith, D. Wachenshwantz, “Magnetoresistive heads and the reciprocity principle,” IEEE Trans.
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
recording,” IEEE Trans. Magn. 40 (1), 129 (2004).
[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).
[123] M. V. Fedoryuk et al., “Asymptotic analysis: linear ordinary differential equations,” Berlin, Springer,
1993.
[124] Schwarz and S. K. Decker, IEEE Trans. Magn. 15, 1622 (1979).
Perpendicular Magnetic Recording 168
[125] N. Smith, D. Wachenshwantz, “Magnetoresistive heads and the reciprocity principle,” IEEE Trans.
Magn. 23 (5), 2494 (1987).
[126] 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 (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.
[131] S. H. Charap, P.-L. Lu, Y. He “Thermal Stability of Recorded Information at High Densities, “IEEE
Trans. Magn., 33 (1), 978-83 (1997).
[132] S. Iwasaki and Y. Nakamura, “An analysis for the magnetization mode for high density magnetic
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.
[140] S. Khizroev, M. Kryder, and D. Litvinov, “Next generation perpendicular systems,” IEEE Trans.
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).
Perpendicular Magnetic Recording 170
[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).
[170] S. Khizroev, Y. Liu, K. Mountfield, M. Kryder, D. Litvinov, “Physics of perpendicular magnetic
recording: writing process,” JMMM 246 (1-2), 335-44 (2002).
[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
recording,” IEEE Trans. Magn. 40 (1), 129 (2004).
[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).
[186] S. Khizroev, M. H. Kryder, Y. Ikeda, K. Rubin, P. Arnett, M. Best, D. A. Thompson, “Recording heads
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).
This page intentionally left blank
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