-
Overcoming the Trilemma Issues of Ultrahigh DensityPerpendicular
Magnetic Recording Media
by L10-Fe(Co)Pt Materials
Fang WangKey Laboratory of Magnetic Molecules and Magnetic
Information
Material of Ministry of EducationSchool of Chemistry and
Materials Science of Shanxi Normal University
Linfen 041004, P. R. [email protected]
Hui XingDepartment of Physics, University at Bualo
SUNY, Bualo, NY 14260, [email protected]
Xiaohong Xu*
Key Laboratory of Magnetic Molecules and Magnetic
InformationMaterial of Ministry of Education
School of Chemistry and Materials Science of Shanxi Normal
UniversityLinfen 041004, P. R. China
[email protected]
Received 20 January 2015Accepted 6 April 2015Published 22 April
2015
L10-ordered FePt and CoPt (collectively called L10-Fe(Co)Pt in
this review) have become po-tential materials for future ultrahigh
density perpendicular magnetic recording (PMR) media dueto their
high magnetocrystalline anisotropy, rendering small grains with
high thermal stability.However, PMR media using such high
anisotropy faces the well-known trilemma issues amongthermal
stability, signal-to-noise ratio (SNR), and writability. This paper
will provide an over-view of the impact of L10-Fe(Co)Pt on
overcoming the superparamagnetic limit and balancingthe trilemma
issues for ultrahigh density PMR media. Here the research and
development ofL10-Fe(Co)Pt materials will be presented, from the
perspectives of enhancing thermal stability,SNR and writability.
Furthermore, we will provide some combined approaches to tackle
thechallenges in balancing the trilemma issues, focusing on
materials engineering.
Keywords: Perpendicular magnetic recording media; L10-Fe(Co)Pt;
trilemma; thermal stability;signal-to-noise ratio; writability.
*Corresponding author.
SPINVol. 5, No. 1 (2015) 1530002 (26 pages) World Scientic
Publishing CompanyDOI: 10.1142/S2010324715300029
1530002-1
SPIN
201
5.05
. Dow
nloa
ded
from
ww
w.w
orld
scie
ntifi
c.com
by N
ATI
ON
AL
UN
IVER
SITY
OF
SIN
GA
PORE
on
04/2
9/15
. For
per
sona
l use
onl
y.
-
1. Introduction
The demand for data storage devices has never beenhigher due to
the exponential growth of new infor-mation. Up to now, hard disk
drive (HDD) is stillthe main storage device among various
storagetechnologies for its cost per gigabyte of data al-though
solid state memories have gained momen-tum in personal computers
and electronic devicesmarket.1 On the other hand, the fast
developingcloud storage market provides new demand forHDD. Over the
past several decades, there wastremendous growth in the areal
density of HDDs,especially since the introduction of the giant
mag-netoresistance (GMR)2,3 head technology in 1996.However, the
rate of increase started to drop since2003, partly due to the
superparamagnetic behaviorof the longitudinal recording media,
which made itdicult to break 100Gb/in.2.4 Thus
CoCr-basedperpendicular magnetic recording (PMR) with themagnetic
moment of the bit oriented along the lmnormal was introduced to
overcome the super-paramagnetic eect.5 Moreover, both
signal-to-noiseratio (SNR) and areal density can be improved bythis
recording structure. The rst CoCrPt perpen-dicular recording HDD
with areal density of133Gbit/in.2 was rst demonstrated by Toshiba
in2004. The transition from longitudinal to perpen-dicular
recording has resulted in rapid growth inareal density for nearly
another decade. However,with the areal density approaching
1Tbit/in.2
today, the CoCrPt perpendicular recording media isstill facing
the superparamagnetic limit. In order toovercome this limit on
magnetic recording, the ex-ploitation of new technology or new
materials isnecessary.
It is well known that increasing the anisotropyof the media can
compensate for the super-paramagnetic eect caused by the reduction
in grainsize because the anisotropy barrier of
magnetizationreversal is proportional toKuV , whereKu and V arethe
uniaxial magnetocrystalline anisotropy and grainvolume,
respectively.6 Chemically orderedL10-alloyswith a
face-centered-tetragonal (fct) structure, suchas L10-Fe(Co)Pt, have
become one of the potentialmaterials due to their high
magnetocrystalline an-isotropy.7 Thus, L10-Fe(Co)Pt grains with
smallsizes can provide high SNR, which is determined bythe number
of grains in each bit (SNR 10log(N),where N is the number of grains
in a bit.).810 How-ever, the high anisotropy results in an increase
in
coercivity (Hc) proportional to Ku/Ms, whereMs isthe saturation
magnetization. Thus the writing eldwill have to be increased, which
may exceed the strayeld that can be supplied by the write head.
There-fore, thermal stability, SNR and writability areintertwined
in such a way that the improvement ofone parameter may lead to
deterioration of the other.These mutually restraining factors are
commonlycalled the trilemma issues of the magneticrecording media,
and the relations are shown inFig. 1. At present, it has become the
major roadblockto the ever increasing areal density growth
ofL10-Fe(Co)Pt perpendicular recording media.
11,12 Inorder to overcome and balance these issues,
severaltechnologies have been proposed, including bit pat-terned
media (BPM), heat assisted magnetic re-cording (HAMR), and
microwave assisted magneticrecording (MAMR). In this paper, we only
provide areview on the contributions of L10-Fe(Co)Pt onovercoming
and balancing the trilemma issues forultrahigh density magnetic
recording media. There-fore, we will focus on summarizing the
structure andmagnetic manipulation of L10-Fe(Co)Pt materialson
realizing the high thermal stability, high SNR, andhigh writability
of perpendicular recording media.
In the past decade, several new schemes wereproposed to overcome
the trilemma issues. In thefollowing sections, we review recent
progress, fo-cusing on material engineering in L10-Fe(Co)Ptalloys.
Section 2 discusses approaches to realizingphase transformation of
Fe(Co)Pt alloy to enhancethe thermal stability. Section 3 discusses
approachesto improving SNR, such as granular perpendicularmedia
(GPM), percolated perpendicular media(PPM) and BPM. Section 4
discusses approaches topromoting writability, including
texture-tilting-assisted media and domain-wall-assisted media.
Fig. 1. Trilemma issues of the PMR media.
F. Wang, H. Xing & X. Xu
1530002-2
SPIN
201
5.05
. Dow
nloa
ded
from
ww
w.w
orld
scie
ntifi
c.com
by N
ATI
ON
AL
UN
IVER
SITY
OF
SIN
GA
PORE
on
04/2
9/15
. For
per
sona
l use
onl
y.
-
However, any single approach mentioned above canonly mitigate
but not solve the trilemma problems.In the last section, we provide
an overview of thepredicted combining approaches to tackle the
tri-lemma challenges.
2. Approaches to EnhancingThermal Stability
The thermal stability depends on anisotropy barrier,which is
proportional to the Ku value of the mate-rial. It is known that Ku
of L10-ordered FePt andCoPt are 6:6 107 erg/cm3 and 4:9 107
erg/cm3,respectively, which are about 20 times higher thanthat of
CoCrPt used widely in the commercial re-cording media.6,7 It is
estimated that L10-FePt orL10-CoPt is thermally stable even for
grain size assmall as 34 nm. If one can obtain such smallgrains and
write the information on them, L10-Fe(Co)Pt-based media with an
ultrahigh arealdensity can be readily achieved. Usually, FePt
orCoPt alloy lms deposited at room temperatureare a disordered
face-centered-cubic (fcc) phasethat is magnetically soft. A high
temperaturetreatment, such as in situ substrate heating
orpost-deposition annealing at temperatures as highas 550750C is
necessary to obtain an orderedL10-Fe(Co)Pt phase.
1315 However, it is dicult toobtain (001) texture with
perpendicular easy axisorientation because the (111) plane is the
closestpacked plane with the lowest surface energy. Inaddition,
such a high annealing temperature is notsuitable for practical
applications. Therefore, manyattempts have been made to induce
perfect fct(001) texture and reduce the ordering temperature.Here,
two typical strategies are introduced to re-alize this goal, one is
stress-assisted phase trans-formation, and the other is
metal-doping-promotedphase transformation.
2.1. Driving L10-Fe(Co)Pt phasetransformation by
stress-assistedgrowth
Epitaxial growth is a common means to induce (001)texture in
L10-Fe(Co)Pt materials. For FePt orCoPt L10-alloys, people usually
use single crystalsubstrates, such as MgO or introduce an
additionalunderlayer, such as Ag and Cr, to induce the
phasetransformation from fcc to fct and to obtain (001)orientation
with the help of a small lattice mismatch
between FePt lms and substrates/underlayers.1629
Farrow et al. rst realized the control of chemicalordering and
magnetic anisotropy in epitaxial FePtlms prepared on Pt/MgO (001)
substrates.16
Figure 2 shows the specular X-ray diraction (XRD)patterns for
FePt lms grown at dierent tempera-tures. It is found that the
long-range order param-eter increases from near zero for lms grown
at100C to a maximum of 0.93 in lms grown at500C. Over this range,
the magnetic easy axischanges from in-plane to perpendicular.
Further-more, Shima et al. prepared ordered L10-FePt lmswith large
magnetic anisotropy by alternating Feand Pt monatomic layers on MgO
(001) substratesat low temperatures.17
Among these underlayers, Ag is one of the mostpopular one.
Figure 3 shows the lattice constants ofFePt and fcc-Ag crystal
structures. Lattice a of thedisordered fcc-FePt is 3.82 and c=a 1,
whilec (3.71) is shorter than a (3.86) for orderedL10-FePt and c=a
0:961. As shown in Fig. 3, Ag(001) plane has a slightly larger
lattice than fct-FePt(001) plane, and the stress caused by a small
mis-match (5.6%) between Ag (001) and FePt (001) canresult in the
shrinkage of the FePt lattice along thelm normal direction and thus
induce the orderingof FePt lms at a lower ordering temperature.
Fig. 2. XRD patterns for FePt/Pt(001)/MgO(001) lms atsubstrate
temperatures of 100C, 200C and 550C.16
Overcoming the Trilemma Issues of Ultrahigh Density PMR
1530002-3
SPIN
201
5.05
. Dow
nloa
ded
from
ww
w.w
orld
scie
ntifi
c.com
by N
ATI
ON
AL
UN
IVER
SITY
OF
SIN
GA
PORE
on
04/2
9/15
. For
per
sona
l use
onl
y.
-
Therefore, the stress-induced phase transformationshould be an
eective method to control the per-pendicular orientation and reduce
the orderingtemperature of L10-Fe(Co)Pt lms.
Several groups have prepared L10-FePt orL10-CoPt lms at low
temperatures by using dif-ferent underlayers, and also investigated
the de-pendence of the lattice mismatch on the ordering ofFe(Co)Pt
lm and the corresponding magnetic an-isotropy. Hsu et al. reported
that the epitaxialgrowth of FePt induced by the Ag underlayeris
clearly improved with increasing substrate tem-perature from 75C to
300C.18,19 Xu et al. dem-onstrated the importance of Ag underlayer
bycomparing the magnetic properties and structures ofFePt, FePt/Ag
and FePt/Cu thin lms depositedby magnetron sputtering.20,21 Figure
4 shows thehysteresis loops of the FePt thin lms annealed at350C
and 550C, respectively.21 It is found thatFePt/Ag thin lm has a
high coercivity of about6.2 kOe at a relatively low temperature of
350C,
while the coercivities of FePt and FePt/Cu lmsare only 1.2 kOe
and 0.1 kOe due to the presenceof fcc-FePt disordered phase. The
coercivity ofFePt/Cu lm only reaches about 3 kOe afterannealing at
550C. This is because the latticeparameters of Cu underlayer are
too small comparedwith that of FePt to induce the phase
transforma-tion of FePt lms. Furthermore, the dependenceof Ag
underlayer thickness on the orientation ofCoPt lms was
investigated. The (001) texture ofL10-CoPt/Ag lms deposited by
magnetron sput-tering can be improved signicantly by introducinga
thicker Ag underlayer.22,23
Besides Ag underlayer, Suzuki et al. preparedordered FePt thin
lms with perpendicular mag-netic anisotropy on Cr(100)
underlayer/MgO seedlayer at low temperature of 450C due to
theirproper mismatch.24 Based on the success of Cr (100)underlayer,
Wang et al. prepared the ordered FePtthin lms with fct (001)
texture on Cr100xRuxcomposite underlayer.25 Addition of Ru in
Cr
Fig. 3. Lattice constants of FePt and fcc-Ag crystal
structures.
(a) (b)
Fig. 4. In-plane hysteresis loops of the thin lms: (a) 350C and
(b) 550C.21
F. Wang, H. Xing & X. Xu
1530002-4
SPIN
201
5.05
. Dow
nloa
ded
from
ww
w.w
orld
scie
ntifi
c.com
by N
ATI
ON
AL
UN
IVER
SITY
OF
SIN
GA
PORE
on
04/2
9/15
. For
per
sona
l use
onl
y.
-
underlayer results in the formation of the FePt or-dered phase
with c-axis orientation perpendicular tothe lm plane at a lower
substrate temperature of350C. Figure 5 shows the out-of-plane and
in-planehysteresis loops of Cr91Ru9/Pt/FePt lm depositedat 400C.25
A thin Pt intermediate layer betweenthe FePt layer and the CrRu
underlayer is intro-duced to eectively resist the Cr diusion from
theCrRu underlayer into the FePt layer. The out-of-plane loop shows
a coercivity of 3.7 kOe with rem-anent magnetization squareness of
0.97, while thein-plane coercivity is only 190Oe. It indicates
thatthe fct (001) texture has been achieved at a lowerdeposition
temperature. They also found that acritical lattice mismatch near
6.3% to be the mostsuitable for improving the chemical ordering of
theFePt lms. Recently (001) textured FePt lm wasobtained on
MoC/CrRu/glass at 380C by usingmagnetron sputtering, in which the
MoC conductiveintermediate layer was used to resist the Cr
diusionat high deposition temperatures and promote theepitaxial
growth of the (001) texture FePt lm.26 Inaddition, the FePt grains
can be further separatedby excess carbon from MoC intermediate
layer,resulting in small intergrain exchange interaction.27
Other than these approaches, Bi,28 PtMn,29 andAuCu30 underlayers
were all used to induce thephase transformation and reduce the
ordering tem-perature of L10-Fe(Co)Pt alloys. However, Hottaet al.
recently conrmed that there is no markeddierence in the thickness
dependence of Ku inL10-FePt (001) single-crystal lms grown
epitaxi-ally on dierent substrates, although the latticemismatch
between FePt and the substrates is
markedly dierent (from 1.4% to 9.1%).31 However,Ku decreased
gradually as the lm thickness de-creased. It is likely that the
lattice mismatchbetween FePt and these substrates was relaxed inthe
rst 1 or 2 layers of FePt (001) lattices. There-fore, the lattice
mismatch may not be the mostcritical factor to obtain high Ku, the
lm thicknessalso plays an important role.
However, the usage of single crystal substrate orunderlayer is
restricted in actual applications be-cause a soft underlayer is
required under the re-cording layer. Zeng et al. obtained nearly
perfect(001)-oriented CoPt and FePt lms with none-pitaxial growth
by directly depositing lms on glassor thermally oxidized Si
substrates and rapidthermal annealing.32,33 Figure 6 shows the
typicalhysteresis loops for CoPt and FePt lms fabricatedby rapid
thermal annealing.32 It is seen that theeasy axis is in the
perpendicular direction for bothlms, and the perpendicular loops
show large co-ercivity and high remanence ratio. The
textureevolution mechanism was proposed to be the mis-match between
the thermal expansion coecientsof the metallic Fe(Co)Pt and glass
or SiO2, whichleads to large strain. This proposal was later
con-rmed by works of other groups.34,35 Dang et al.found that FePt
(001) texture is obtained onthermally oxidized Si substrates when
lms thick-ness is less than 10 nm, whereas (111) orientationappears
in the lms with the thickness larger than10 nm. This is similar to
Hotta's results.31 Theminimizing surface energy is proposed to
explainthe texture variation in the lms based on a theo-retical
mode.35
Fig. 5. Hysteresis loops of Cr91Ru9/Pt/FePt lm deposited
at400C.25
Fig. 6. Typical hysteresis loops for (a) CoPt annealed at750C
for 300 s and (b) FePt annealed at 550C for 5 s.32
Overcoming the Trilemma Issues of Ultrahigh Density PMR
1530002-5
SPIN
201
5.05
. Dow
nloa
ded
from
ww
w.w
orld
scie
ntifi
c.com
by N
ATI
ON
AL
UN
IVER
SITY
OF
SIN
GA
PORE
on
04/2
9/15
. For
per
sona
l use
onl
y.
-
2.2. Driving L10-Fe(Co)Pt phasetransformation by
metal-doping
Another eective way to promote the ordering ofL10-Fe(Co)Pt and
reduce the phase transformationtemperature is metal-doping.
However, dierentdriving mechanisms were proposed for
dierentadditives in L10-Fe(Co)Pt lms. Maeda et al. foundthat the
addition of Cu into FePt alloy lm is veryeective in reducing the
ordering temperature ofL10-FePt.
36,37 Figure 7 shows the coercivity of FePtand (FePt)8Cu15 lms
as a function of the annealingtemperature.36 The coercivity of
FePtCu lm isaround 5 kOe after annealing at 300C, whereasthat of
FePt shows several hundred Oe. The for-mation of the ternary FePtCu
alloy is considered tobe the main reason in reducing the ordering
tem-perature. The lattice parameters of the L10-orderedphase
suggest that Fe atoms are substituted by Cuatoms. The grain size
increases by the addition ofCu, suggesting that the decrease of the
annealingtemperature for ordering is due to the enhancedkinetics of
ordering during the alloying process.38
Similarly, Wang et al. prepared CoPtCu/Ag lmsby magnetron
sputtering and subsequent annealing,in which Ag underlayer plays a
dominant role ininducing the (001) texture, while Cu dopant is
usedto form CoPtCu ternary alloy. The CoPtCu/Aglms with
perpendicular orientation start to order ata lower annealing
temperature of 450C, which islower by 150C than the pure CoPt/Ag
lms.39
The inuences of Cu, Ag and Au additives on theL10 ordering,
texture and grain size of FePt thinlms are reported.4043 It is
suggested that Au andAg additives tend to segregate at the FePt
grainboundaries to inhibit FePt grain growth. However,Cu
substitution in FePt increases the average grain
size and lm roughness. FePt lms with Au or Agadditive show 12
kOe higher coercivity comparedto that of pure FePt lm after
annealing at 450C.The driving force of phase transformation
comesfrom the vacancy defects during the diusion out ofthe FePt
grains, resulting in an enhancement of L10ordering kinetics and
reduction of the orderingtemperature.
Besides, Kitakami et al. studied the eects ofadditional elements
Sn, Pb, Sb and Bi on the or-dering of L10-CoPt lms.
44 All these additives aredemonstrated to be very eective to
promote theordering of the samples annealed at 400C, which is200C
lower than that of pure CoPt. That is becausethese additives are
easy to diuse and segregate ontothe lm surfaces by post-annealing
due to their verylow surface free energy and extremely low
solubilityin CoPt, leading to a lot of defects to drive
phasetransformation. The results are similar to that of Auor Ag
additives. In order to investigate this point,Fig. 8 shows the
Auger electron spectroscopy depthproles of the CoPtSb lms before
and afterannealing at 650C.44 Clearly, Sb tends to diusetowards the
lm surfaces. Such surface segregationis caused by low surface free
energy and limitssolubility of Sb in CoPt. Lee et al. found that
Zr-doped FePt alloy lms could accelerate ordering
Fig. 7. In-plane coercivity Hc of FePt and (FePt)8Cu15 lmsas a
function of the annealing temperature.36
(a)
(b)
Fig. 8. Compositional depth proles for (a) as-made and
(b)annealed CoPt-Pb lms.44
F. Wang, H. Xing & X. Xu
1530002-6
SPIN
201
5.05
. Dow
nloa
ded
from
ww
w.w
orld
scie
ntifi
c.com
by N
ATI
ON
AL
UN
IVER
SITY
OF
SIN
GA
PORE
on
04/2
9/15
. For
per
sona
l use
onl
y.
-
transformation kinetics while keeping small grains.45
It is thought that the point defects and lattice
strainintroduced by Zr-doping activate the nucleation ofthe ordered
phase. Therefore, various metal-dopantswith dierent mechanism
contribute in inducing theordering of L10-Fe(Co)Pt thin lms.
Summarizing this section, many studies havebeen done in
obtaining L10-Fe(Co)Pt with high Kuvalues, including
stress-assisted growth and metal-doping. However, L10-Fe(Co)Pt with
high Kuvalues is just one of the basic requirements forultrahigh
density recording media. In the following,we will review strategies
to improve the SNR andwritability.
3. Approaches to Improving SNR
Both stress-assisted growth and metal-doping canhelp to obtain
L10-Fe(Co)Pt with large Ku valuesand perpendicular orientation at
relatively low tem-peratures. However, there exists a large
transitionnoise for any continuous medium. In order to reducethe
transition noise and achieve high SNR,46,47 sev-eral common
approaches have been adopted forL10-Fe(Co)Pt, such as GPM, PPM and
BPM.
3.1. Granular perpendicular media
In GPM, nonmetal additions have been used todecouple magnetic
grains. Several methods such ascosputtering or laminating FePt or
CoPt withnonmagnetic materials have been attempted to de-crease the
intergrain exchange interaction. For ex-ample, highly anisotropic
fct-Fe(Co)Pt nano-particles have been prepared and embedded in
Cmatrix by cosputtering deposition.48 Figure 9 is theTEM images of
annealed FePt/C lms with dier-ent carbon thickness.48 It is shown
that FePt par-ticles are embedded in C matrix and the particle
sizevaries from below 3 nm to about 8 nm with de-creasing the
carbon contents. This is because thenonmagnetic carbon atoms are
easy to diuse intothe grain boundaries to isolate the FePt
magneticgrains during the annealing process, resulting ina weak
intergrain exchange coupling. Also theincrease of C contents plays
a role in restrain-ing grain growth. Xu et al. also found that
thegrain size and intergrain interaction of the FePt/Cmultilayer
lms decreases with increasing C con-tent.49 The coercivity not only
depends on the Ccontent, but also on the structure of the
FePt/Cmutilayer. Moreover, L10-FePt particles with high
Fig. 9. TEM images of annealed FePt/C lms with dierent carbon
thickness.48
Overcoming the Trilemma Issues of Ultrahigh Density PMR
1530002-7
SPIN
201
5.05
. Dow
nloa
ded
from
ww
w.w
orld
scie
ntifi
c.com
by N
ATI
ON
AL
UN
IVER
SITY
OF
SIN
GA
PORE
on
04/2
9/15
. For
per
sona
l use
onl
y.
-
magnetocrystalline anisotropy and small particlesizes between 3
nm and 15 nm were fabricated byannealing FePt/BN multilayers at
high tempera-tures.50 The BN layers are used to control the
in-terparticle interactions of FePt grains.
Oxides are usually used as segregating materialsto achieve the
required L10-Fe(Co)Pt microstruc-ture.51,52 Nanocomposite FePt:SiO2
lms have beenfabricated by annealing the as-deposited
FePt/SiO2multilayers at high temperatures.53 These lmsconsist of
high-anisotropy L10-FePt particles em-bedded in a SiO2 matrix. The
coercivity in therange from 2 to 8 kOe and grain size of 10 nm or
lessare highly dependent on the annealing temperatureand SiO2
concentration. The nanostructured FePt:B2O3 lms with average grain
sizes from 4nm to17 nm were obtained by similar methods.54,55
Thec-axis of the FePt grains with a nearly perfect (001)orientation
can be obtained, resulting in perpen-dicularly oriented
nanocomposite lm with a highanisotropy constant of 3:5 107 erg/cm3.
Figure 10shows the XRD patterns of [Fe/Pt/B2O3]n lmsannealed at
550C.55 One can see that the relativeintensity of the (111) peak
decreases when theinitial B2O3 layer increases. The (111) peak
nearlydisappears and the (00n) peaks become dominant at
x 8, indicating the alignment of the c-axis ofFePt grains along
the normal direction. Furtherexperimental studies and ab initio
calculations alsoindicate that the B2O3 matrix results in strain
onFePt grains that changes the c=a ratio and thusmagnetic
properties such as Curie temperature.56 Inaddition, Al2O3 and ZrO2
were used to controlthe grain size and intergrain interaction of
L10-Fe(Co)Pt lms.57,58 Strong perpendicular anisotro-py, adjustable
coercivity, and ne grain size suggestthat oxide addition can play a
signicant role inreducing the exchange coupling interaction
betweenmagnetic grains.
In previous studies, Ag underlayer can induce(001)-oriented
L10-Fe(Co)Pt lms and nonmagneticC-doping can adjust particle size
and intergranularexchange coupling. It is well known that the
idealPMR media should have the perfect perpendicularorientation,
small and isolated grains, suitable co-ercivity, and low media
noise. Based on the aboveideas, we proposed a novel triple material
system of[CoPt/C]n/Ag/glass lms prepared by magnetronsputtering and
post-annealing.59 It is found that theoriented growth of L10-CoPt
lms is stronglyinuenced by both Ag underlayer thickness and
Ccontent. A nearly perfect (001) texture and a highperpendicular
magnetic anisotropy can be obtainedin the [CoPt/C]5/Ag lms. Xu et
al. also prepared(001)-oriented [C/CoPt/Ag]5 lms
60 and furtherAg/[CoPt/C]5/Ag lms with Ag as the underlayerand
top layer.61 Figure 11 shows the hysteresis loopsof the samples
with dierent structures.61 Sample Cof Ag(5 nm)/[CoPt(3 nm)/C(3
nm)]5/Ag(5 nm) lmhas a very large perpendicular coercivity of856
kAm1 (10.7 kOe) and a very low parallel co-ercivity of 63 kAm1 (0.8
kOe), which is consistentwith the XRD result showing the high
intensity(001) and (002) peaks of L10-CoPt. The strain en-ergy
caused by the Ag underlayer together with thediusion of Ag and C
atoms results in the en-hancement of the degree of chemical
ordering andthe development of the (001) texture for L10-CoPtlms.
Similarly, Chen et al. also prepared FePt-Clms with high
coercivity, (001) texture, and smallgrain size on MgO/CrRu/glass
substrate bycosputtering FePt and carbon at 350C, in whichMgO
underlayer is used to induce the ordering ofFePt at low
temperatures.62 It is clear that non-magnetic isolation combined
with underlayer is oneof the eective methods to obtain
L10-Fe(Co)Ptwith (001) texture.
Fig. 10. XRD patterns of [Fe(3)/Pt(4)/B2O3(x)]n lmsannealed at
550C for 30min. (a) x 2, n 10; (b) x 4,n 9; (c) x 8, n 7; and (d) x
12, n 6. Insets areXRD patterns of the corresponding as-deposited
lms.55
F. Wang, H. Xing & X. Xu
1530002-8
SPIN
201
5.05
. Dow
nloa
ded
from
ww
w.w
orld
scie
ntifi
c.com
by N
ATI
ON
AL
UN
IVER
SITY
OF
SIN
GA
PORE
on
04/2
9/15
. For
per
sona
l use
onl
y.
-
3.2. Percolated perpendicular media
Dierent from the GPM, a so-called PPM consistingof fully
exchanged coupled grains with densely dis-tributed nonmagnetic
pinning sites, is proposedtheoretically by Zhu et al. to reduce the
transitionnoise.6365 Figure 12 shows the illustration of
tran-sition boundaries for the GPM and PPM.64 Obvi-ously,
transition jitter noise is dominated by thegrain size and size
distribution for GPM. If the pin-ning sites are distributed more
densely than that ofthe grains in GPM, the resulting recorded
transitionboundaries would be signicantly less irregular, asshown
in the right picture. Consequently, transition
jitter noise would be signicantly reduced in PPM,while the
ferromagnetic exchange coupling betweenthe grains ensures sucient
thermal stability. Acomprehensive micromagnetic simulation study
hasshown that the medium transition noise can be op-timizedwith
amoderate exchange coupling constant.The switching time of the
percolated medium issmaller than 1 ns even with low switching eld
andsmall damping constant.66 Therefore, PPMmay oerbetter recording
properties over the present GPM inimproving the SNR of the
perpendicular media.
Based on this model, the rst PPM has beenreported in hcp
CoPt-SiO2 thin lms by the alter-nate sputtering of CoPt and SiO2
targets. Desired
Fig. 12. Illustration of transition boundaries for the present
GPM and PPM. The white dots in the PPM medium indicatesnonmagnetic
columnar grains which act as domain wall pinning sites.64
Fig. 11. Hysteresis loops of samples with dierent structures
before (insets) and after annealing at 600C. Here represents
out-of-plane hysteresis loops and represents in-plane ones.61
Overcoming the Trilemma Issues of Ultrahigh Density PMR
1530002-9
SPIN
201
5.05
. Dow
nloa
ded
from
ww
w.w
orld
scie
ntifi
c.com
by N
ATI
ON
AL
UN
IVER
SITY
OF
SIN
GA
PORE
on
04/2
9/15
. For
per
sona
l use
onl
y.
-
CoPt-SiO2 PPM microstructure was obtained uponpost-deposition
annealing.65,67 Figure 13 showsthe TEM images of CoPt-SiO2 and
(FePt)(MgO)/Pt/Cr thin lms.67,68 The plan-view image of
as-deposited CoPt-SiO2 lm [Fig. 13(a)] has a micro-structure
similar to conventional GPM. The plan-view microstructure of the
annealed sample is shownin Fig. 13(b). The magnetic grains are
magneticallyinterconnected, while the oxide forms
sphericalparticles in the grain boundaries of the lm. Theoxide
phase pins the magnetic domain walls, hin-dering their motion and
hence producing increasedcoercivity. Moreover, Sun et al. prepared
thepercolated perpendicular FePtMgO lms by con-ventional magnetron
sputtering.68 Magnetic mea-surements demonstrate that the
coercivity of themagnetic lm drastically increases from 2.1 to3.6
kOe as the MgO content is increased from 0 vol.% to 0.15 vol.%.
Here MgO is present as crystallinedots that are uniformly
precipitated in the FePtmatrix, which can be conrmed by
comparingFigs. 13(c) and 13(d). The MgO dots serve as pin-ning
sites of the domain wall and enhance perpen-dicular coercivity.
In order to form the PPM with evenly distributedpinning centers,
Lai et al. proposed a simple route tofabricate PPM, where the Co/Pt
multilayers weredeposited onto anodized alumina oxide
(AAO)substrates utilizing pores as pinning sites.6971
Coercivity, domain size, and switching eld canbe engineered by
controlling pore density. Figure 14shows the MFM/SEM images and
correspondinghysteresis loops of (Co/Pt)/Si lm and (Co/Pt)/AAO/Si
lms with dierent pore density.71 It isfound that the perpendicular
coercivity increaseslinearly with increasing pore density due to
thepinning eect imposed by the pores, which is con-sistent with
theoretical calculation for PPM. Abetter tolerance to switching-eld
distributions canthus be expected, which may help to achieve a
highSNR. In addition, Schulze et al. obtained a similarresult in
the Co/Pt multilayers deposited ontonanoperforated ZrO2
membrane.
72
Taken together, these domain wall pinning sitescan be either
nonmagnetic oxides or physical voidsor both of them. All previous
theoretical studies onPPM were done on exchange coupled magneticlms
with holes acting as pinning sites. However,
Fig. 13. Plan-view TEM image of CoPtSiO2 and FePtMgO thin lms.
(a) As-deposited CoPtSiO2 lm; (b) annealed CoPtSiO2lm; (c)
(FePt)(MgO)/Pt/Cr lm and (d) FePt/Pt/Cr trilayer lm.67,68
F. Wang, H. Xing & X. Xu
1530002-10
SPIN
201
5.05
. Dow
nloa
ded
from
ww
w.w
orld
scie
ntifi
c.com
by N
ATI
ON
AL
UN
IVER
SITY
OF
SIN
GA
PORE
on
04/2
9/15
. For
per
sona
l use
onl
y.
-
magnetic nanodots may be present in the holes, andthus the
magnetic interactions in such systems aremore complicated.
Therefore, Schulze et al. carriedout a series of micromagnetic
simulations to under-stand the modication of the pinning strength
ofdomain walls due to the magnetic interaction be-tween nanodots
and the surrounding lm.72 Acomparison of magnetic domain wall
pinning inPPM systems with and without nanodots is given inFig. 15.
The simulation data shows that the domainwall pinning behavior
strongly depends on the ex-change coecient Aint. When Aint 0 (no
exchange
coupling between lm and nanodots), the depinningeld (Hd is the
same as the system without nano-dots. The magnetic state of the
nanodots is not af-fected by the domain wall motion in the
surroundinglm. As Aint is increased from 0 to 1 106 ergs/cm,a
larger Hd is required as the nanodots stabilizes thepinned domain
wall. The dierence in morphologyprohibits a domain wall motion, as
shown in theright column of Fig. 15(b). The pinning strength
isdetermined not only by the geometrical properties ofthe template
but is also aected by the exchangecoupling between the lm and the
nanodots in the
(a) (b) (c) (d)
Fig. 14. MFM and SEM images of (Co/Pt)/Si (a), (Co/Pt)/AAO/Si
lms with pore density of 3:3 101 cm2 (b), and11:6 1010 cm2 (c), (d)
is the corresponding out-of-plane and in-plane hysteresis loops of
(a) and (c).71
Fig. 15. (a) Initial magnetization states for two types of PPM,
(b) domain wall displacement in an increasing eld H.
Themagnetization states in (b) correspond to the same eld step H H
Hd.72
Overcoming the Trilemma Issues of Ultrahigh Density PMR
1530002-11
SPIN
201
5.05
. Dow
nloa
ded
from
ww
w.w
orld
scie
ntifi
c.com
by N
ATI
ON
AL
UN
IVER
SITY
OF
SIN
GA
PORE
on
04/2
9/15
. For
per
sona
l use
onl
y.
-
system. However, the possibility of controlled pin-ning of
magnetic domain walls at even smallerlength scales, as required for
storage densities be-yond 10Tbit/in.2, remains an open
question.
3.3. Bit patterned media
To achieve ultrahigh areal density, BPM have beenproposed by
Chou and Krauss73,74 to reduce oreliminate transition noise without
a loss of thermalstability. Figure 16 shows the comparison
betweenconventional media and BPM.75 Compared to theconventional
granular media, BPM consist of peri-odic magnetic nanodot arrays,
where each dot canbe regarded as a separately recorded magnetic
bit,which can eectively enhance the SNR of perpen-dicular recording
media.76,77 However, one of themain challenges is to obtain the
smallest possiblefeature size for ultrahigh areal density. To
date,there are several methods to realize BPM. The mostcommon
approach is top-down physical fabricationtechnologies,78,79 in
which the magnetic nanoarraysare obtained by lithographic
patterning, ion milling,or focused ion beam (FIB). However, there
are somelimitations of such techniques, including thehigh
production cost and low throughput, as well asthe maximum
achievable areal density.
In view of the drawbacks of top-down approa-ches, some bottom-up
chemical template methodswere used to prepare BPM,80,81 for which
the bitsare formed by electrochemical deposition or sput-tering
using the self-organized templates. A pio-neering work was done by
Kim et al. to push thelimit of the areal density of the magnetic
nanodotarray.82 Ordered FePt nanodot arrays with a per-fect
perpendicular easy axis were deposited bymagnetron sputtering into
AAO templates fol-lowed by a rapid thermal annealing. Figure
17shows the SEM images, XRD pattern, and hys-teresis loops of
annealed FePt nanodots array.82
FePt nanodots with diameter of 18 nm and peri-odicity of 25 nm
have been fabricated, resultingin an areal density exceeding
1Tbit/in.2. Rapidthermal annealing converts the disordered fcc
to(001)-oriented L10-FePt nanodot arrays with per-pendicular
anisotropy and large coercivity. L10-Fe(Co)Pt nanowires and
nanorods were also pre-pared in AAO templates by
electrochemistryfollowed by annealing.83,84 These studies show
thatself-organized templates are low-cost with highuniformity and
easily controllable structural para-meters. In particular, for AAO
templates, not onlythe diameter and density can be adjusted, but
alsothe template can withstand high temperature of
Fig. 16. Comparison between conventional media and BPM.75
F. Wang, H. Xing & X. Xu
1530002-12
SPIN
201
5.05
. Dow
nloa
ded
from
ww
w.w
orld
scie
ntifi
c.com
by N
ATI
ON
AL
UN
IVER
SITY
OF
SIN
GA
PORE
on
04/2
9/15
. For
per
sona
l use
onl
y.
-
650C,85 which is favorable in fabricating L10-Fe(Co)Pt patterned
media. However, it is hard toobtain templates with small pore sizes
less than10 nm and thin pore wall less than 5 nm. It is
thusunfavorable to realize ultrahigh density L10-FePt-based
BPM.
In addition, the block copolymer template is usedto prepare BPM.
Naito et al. combined a diblockcopolymer template to fabricate a
long-range or-dered CoPt patterned media with 40 nm sized
dots,where single magnetic domains with an almostperpendicular
orientation were obtained in each
magnetic dot.81,86 Recently, Zhu et al. obtained thenanoscale
CoPtSiO2 magnetic media with high-coercivity using self-assembling
block copolymers asan etch-mask.87 Figure 18 is the TEM image
oftemplate CoPtSiO2 magnetic media and EDS in-tegrated intensity
maps of Co, Pt and Ru for onegrain. Clearly, the CoPt magnetic
grains are sur-rounded by the lighter-appearing amorphous SiO2.The
average grain size of the CoPt grains is found tobe 16.2 nm with a
standard deviation of 11%. Asseen from EDS mapping, the CoPt grain
is alsoclearly outlined, and it has grown on top of the
Fig. 17. SEM images, XRD pattern, and hysteresis loops of
annealed FePt nanodots array with dierent templates.82
Fig. 18. TEM image of template CoPtSiO2 magnetic media and EDS
integrated intensity maps of Co, Pt and Ru for one grain.87
Overcoming the Trilemma Issues of Ultrahigh Density PMR
1530002-13
SPIN
201
5.05
. Dow
nloa
ded
from
ww
w.w
orld
scie
ntifi
c.com
by N
ATI
ON
AL
UN
IVER
SITY
OF
SIN
GA
PORE
on
04/2
9/15
. For
per
sona
l use
onl
y.
-
dome. The SiO2 is in-between the CoPt grains andin the trenches
dened by the Pt domes.
Recently, Wang et al. proposed a nanopatterningprocess named as
the embedded mask patterning(EMP) to prepare FePt recording media,
as shownin Fig. 19.88 The mean size of FePt grain with 4.6 nmcan be
obtained, and the center to center distance isonly 6.3 nm. The FePt
grain size and packing den-sity can be adjusted and optimized by
changing thesputtering conditions of the embedded mask layer.This
EMP process is compatible with today's mediamanufacturing line
because each step could becompleted in vacuum without taking the
disc out ofthe chamber. Therefore, it provides an eectivemethod to
fabricate FePt BPM with low cost.
Comparing the advantages and disadvantagesamong the
above-mentioned GPM, PPM and BPM,it is clear that BPM is a better
choice to obtain highSNR due to the weak transition noise than
GPMand PPM. However, there are some diculties andchallenges to
realize BPM, in particular the di-culty in fabricating ordered
magnetic nanoarrayswith ultrahigh density at a reasonable cost.
Thisshould be the future direction if BPM is to be used inthe next
generation recording media.
4. Approaches to Promoting Writability
Even if L10-Fe(Co)Pt media could be tailored toachieve the
desired grain size, distribution, and
exchange decoupling, writing information on L10-Fe(Co)Pt media
is still a challenge due to its largeHc and therefore requires a
very high writing eld.To overcome this problem, researchers are
focusingon the development of new recording paradigm.A number of
advanced approaches such as texture-tilting-assisted media,
domain-wall-assisted media,and energy-assisted magnetic recording
were pro-posed to reduce writing eld of L10-Fe(Co)Pt-basedPMR
media. Here we will briey review the progressof
texture-tilting-assisted magnetic recording mediaand
domain-wall-assisted magnetic recording mediaby manipulating the
structures and magneticproperties of L10-Fe(Co)Pt materials.
4.1. Texture-tilting-assisted magneticrecording
Texture-tilting-assisted magnetic recording is a re-cording
scheme, in which the magnetic recording isaccomplished by tilting
the easy magnetization axisthat depends on the crystallographic
orientation ofthe magnetic material comprising the recordingmedium.
Tilted magnetic media with easy axis tilt-ing of 45 was rst
proposed by Gao et al.89,90 andrealized by experiments by Wanget
al.91,92 It allowsreducing the switching eld of high Ku
media,thereby leading to an improvement in the writ-ability. Figure
20 shows the schematic of the tiltedmagnetic media with a
single-pole writing head, anda plot of the normalized switching eld
versus thetilting angle.93 Obviously, three advantages can
beobtained by this design. First, the minimumswitching eld (Hs)
resulted from titling the easy-axis direction is only a half of
what is required inperpendicular media with a tilting angle of
zero.Second, a much better tolerance of switching-elddistribution
can be achieved. Third, a much fastermagnetization switching speed
can be realized whencompared to the untilted perpendicular
media.94
The areal density of tilted media could be more than62% higher
than that of traditional perpendicularmedia if Ku 7 106 erg/cm3.
Therefore, tiltingthe easy magnetization axis in magnetic
recordingmedia is becoming one of the eective methods toimprove the
writability without compromising on itsthermal stability.
Up to now, two main methods have been pro-posed to fabricate the
tilted media: one is the arti-cial-tilted easy magnetization axis
obtained byoblique surface fabrication method; another is
Fig. 19. The schematic drawing of EMP process: (1)
L10-FePtcontinuous layer was deposited on the substrate, (2) mask
layerRuSiO2 with a ne granular structure was deposited on theFePt
layer, (3) the SiO2 of mask layer was removed using areactive ion
etching process (RIE), (4) the Ru dots array pat-tern was
transferred to FePt layer using a RIE process.88
F. Wang, H. Xing & X. Xu
1530002-14
SPIN
201
5.05
. Dow
nloa
ded
from
ww
w.w
orld
scie
ntifi
c.com
by N
ATI
ON
AL
UN
IVER
SITY
OF
SIN
GA
PORE
on
04/2
9/15
. For
per
sona
l use
onl
y.
-
natural-tilted easy magnetization axis achieved bycontrolling
the relative textures. The rst articial-tilted CoCrPt recording
media was fabricated bycombining the oblique sputtering and
collimatedsputtering by Wang's group.91 Similarly,
L10-FePtnanoparticles with articial-tilted c-axis wereassembled
onto MgO (110) substrates with self-organized grooves, where the
c-axis of L10-FePtnanoparticle is tilted an angle of 45 from the
sub-strate normal.95 L10-FePt lms with dierent easy-axis
orientations also can be deposited onto thepyramid-type Si
substrates.96 Therefore, the use ofpreprocessed substrates and
oblique sputteringmethods are favorable for the fabrication of
arti-cial-tilted recording media. However, the obliqueincidence
approach may cause a large angular dis-persion around the tilted
preferred orientation.Albrecht et al. fabricated Co/Pd magnetic
multi-layers with articial-tilted easy-axis through thecurved
surface of spherical nanoparticles.97 Theminimum Hc appears at
45
between the appliedeld and lm normal, which is similar to the
simu-lated results of tilted media.89 The nanostructuresfabricated
by this method are monodisperse, singledomain, and uniform magnetic
anisotropy, whichare expected to provide higher density,
higherthermal stability, and faster switching when com-pared to
conventional PMR media.
To overcome the large angular dispersion causedby the
preparation process of articial-tilted media,
natural-tilted textures with tilted easy magnetiza-tion axis,
such as (101) or (111) textures ofL10-FePt, can be used as tilted
media, whereinthe easy magnetization axis is directly oriented
by3645 with respect to the medium surface normal.Natural-tilted
media has the similar advantageswith the articial-tilted
media.98,99 High-anisotropyL10-CoPt or L10-FePt lms having
natural-tilted(111) texture have been reported.100,101
Room-tem-perature angular remanence measurements (ARM)of L10-CoPt
were carried out in order to determinethe geometrical arrangement
of the easy axis in thelm, as shown in Fig. 21.100 The ARM curves
pro-vided the evidence of four out-of-plane maxima at 36 and 144 ,
within both the (110)and the (110) planes. The maximum
coercivities( 4.8 kOe) were observed, when the eld was ap-plied
along each maxima direction of the ARMcurves [Fig. 21(c)]. The
observed behavior is con-sistent with the presence of four easy
axes withmutually orthogonal in-plane projections, symmet-rically
tilted at an angle of 36 with respect to thelm plane. Such methods
can result in approxi-mately 75% reduction of the writing eld
without aloss of thermal stability.
Therefore, both the articial and naturally-tiltedmedia play a
certain role in improving the writ-ability of PMR without
decreasing its thermalstability. However, the articial-tilted media
cannegatively inuence the magnetic properties due to
(a) (b)
Fig. 20. (a) Schematic illustration of the tilted magnetic media
with a single-pole writing head, (b) normalized switching
eld(Hs=Hk; Hk, magnetic anisotropy eld) versus the angle () between
the external eld and easy axis of a grain exhibiting
uniformmagnetization reversal.93
Overcoming the Trilemma Issues of Ultrahigh Density PMR
1530002-15
SPIN
201
5.05
. Dow
nloa
ded
from
ww
w.w
orld
scie
ntifi
c.com
by N
ATI
ON
AL
UN
IVER
SITY
OF
SIN
GA
PORE
on
04/2
9/15
. For
per
sona
l use
onl
y.
-
the increased roughness, stress and shape anisotropyeects, and
the naturally-tilted media has an axialdistribution of the easy
magnetization directions. Itis therefore a great challenge to
design a head ca-pable of generating a uniform tilted eld,
whichlimits their practical applications in magnetic re-cording
media. Therefore, some new technologiesare still required to
realize the high writability ofL10-FePt-based recording media.
4.2. Domain-wall-assisted recording
Another way to reduce the switching eld is domain-wall-assisted
recording, including exchange-spring(ES), exchange coupled
composite (ECC), and ex-change coupled graded (ECG) recording
media. Allof them are composed of hard magnetic layer and
soft magnetic layer, which are coupled by sharp orgraded
interfaces, as shown in Fig. 22.102 The hardlayer with high Ku
provides a high energy barrier tomaintain high thermal stability,
while the soft layerwith high saturation magnetization switches at
lowapplied eld. A domain wall at the hard/soft inter-face provides
an additional exchange eld on thehard magnetic layer, which helps
to reduce theswitching eld of the recording media. Thus,
thesoft/hard composite media can improve the writ-ability, while
still maintaining a high thermal sta-bility. This idea is very
attractive because themethod for preparing the media is relatively
easywhen compared to that of the
above-mentionedtexture-tilting-assisted media.
The exchange spring scheme was introduced inPMR media in 2005 by
Suess.103,104 The soft
(a) (b)
(c)
Fig. 21. ARM curves in the planes of the MgO substrate: (a)
(110) and (b) (001), (c) schematic illustration of the four tilted
easyaxis model.100
F. Wang, H. Xing & X. Xu
1530002-16
SPIN
201
5.05
. Dow
nloa
ded
from
ww
w.w
orld
scie
ntifi
c.com
by N
ATI
ON
AL
UN
IVER
SITY
OF
SIN
GA
PORE
on
04/2
9/15
. For
per
sona
l use
onl
y.
-
magnetic layer is directly coupled with the hardmagnetic layer
by the soft/hard interface in ex-change-spring media. The simulated
magnetic re-versal process and hysteresis loop are shown inFig.
23.105 It can be seen that the soft layer can helpthe hard layer to
reverse its magnetization directionby exerting an additional
demagnetization eld. Theswitching eld of the soft/hard bilayer is
determinedby the pinning eld at the soft/hard interface,
Hp 1
4 2Khard Ksoft
Jhard;
where Khard and Ksoft are the anisotropies of hardlayer and soft
layer, respectively. Js is the saturationmagnetization of the hard
layer. The switching eldcan be reduced to one quarter of that of
the hardlayer whenKsoft is taken to be zero. When
soft-layeranisotropy is about one fth of the hard-layer
an-isotropy, the smallest switching eld, which is aboutone fth of
that of the hard layer, can beobtained.106 The inuences of soft
layer thickness onnucleation eld, coercivity and the
magnetizationreversal mechanism of exchange spring media
wereanalyzed by micromagnetic simulations.107109 It isan eective
method to improve the writability ofperpendicular recording
media.
Based on the above-mentioned theoretical stud-ies, L10-FePt/Fe
exchange spring lms were suc-cessfully prepared by Casoli et
al.110,111 In order toobtain an ordered L10-FePt hard layer, FePt
layerwas rst deposited onto MgO (001) substrate athigh temperature,
and then Fe soft layer was de-posited onto the L10-FePt hard layer
at room tem-perature to directly form L10-FePt/Fe exchangespring
lms. High anisotropy perpendicular systemswith moderate coercivity
can be easily obtained bycontrolling the thickness of the Fe soft
layer.112
Moreover, the control of the interface morphologycan adjust the
magnetic regime from rigid magnet toexchange-spring magnet because
of the hard/softinterlayer coupling.
Victora et al. theoretically proposed the ECCmedia for the
perpendicular recording during the
Fig. 22. Schematic illustration of the
domain-wall-assistedrecording.102
Fig. 23. Magnetic reverse process and hysteresis loop of the
exchange spring media.105
Overcoming the Trilemma Issues of Ultrahigh Density PMR
1530002-17
SPIN
201
5.05
. Dow
nloa
ded
from
ww
w.w
orld
scie
ntifi
c.com
by N
ATI
ON
AL
UN
IVER
SITY
OF
SIN
GA
PORE
on
04/2
9/15
. For
per
sona
l use
onl
y.
-
same period ES media were investigated.113,114 ECCmedia is based
on a structure with nonmagneticinterlayer between the soft layer
and the hard layer,by which the exchange interaction between
thelayers is coupled indirectly. The switching eld isone half of
that of the hard layer according to theratio between the thermal
barrier of the media andits switching eld 2E=Hs Ms V ,114 inwhich E
and Hs are the thermal barrier E 12Kh V is the switching eld,
respectively. Wanget al. rst carried out the experimental work on
suchmedia with CoPd or CoCrPt as hard layers.115,116
Figure 24 shows the exchange coupling dependenceof remnant
coercivity and thermal stability of the[Co/PdSiO]16/PdSi/FeSiO lms
on dierent PdSiinterlayer thickness.115 Adding nonmagnetic
PdSilayer between hard and soft layers is very helpful forfurther
reducing the coercivity. However, if it is toothick, the coercivity
would increase again because ofthe absence of exchange coupling
between soft andhard layers. They also found that the thermal
sta-bility factor KuV =KBT remains nearly unchangedwith changing
the thickness of the intermediatelayer. Clearly, the coercivity can
be tuned by con-trolling the interlayer coupling strength, while
theirthermal stability is still maintained. Furthermore,similar
results were obtained in hcp-CoCrPt ECCmedia with nonmagnetic Pt
and Pd layers asinterlayers.117120 In addition, Tang et al.
employedmagnetic interlayers to adjust the exchange cou-pling
strength between the soft and hard layers.Besides thickness, the
saturation magnetization ofthe interlayer can be used to control
the exchange
coupling strengthen between the two magnetic lay-ers, leading to
the reduction in coercivity of the ECCmedia.121,122 In a word, ECC
media can furtherimprove the writability compared to the
corre-sponding ES media.
On the basis of the studies of ES and ECC media,researchers
proposed theoretically ECG media witha multilayer structure,123
where the anisotropyvaries layer by layer from the hard to soft
layer.Thermal stability of graded media is dependent onthe
anisotropy of the hardest layer. If the number oflayers in exchange
spring media is increased fromtwo layers to N layers to form the
multilayerstructure of ECG media, the anisotropy dierence
ofadjacent layers will be decreased, resulting in asmaller pinning
eld at the interface,
Hp 1
4
2Kn1 KnJs
14N 1
2KhardJs
;
where Kn1, Kn and Khard are the anisotropies ofthe two adjacent
layers and hardest layer, respec-tively. Js is the saturation
magnetization of thehardest layer. The switching eld is decreased
withincreasing the numbers of layers. As the multilayersare
extended to graded media with continuouslyvarying anisotropy, the
pinning eld can be rewrit-ten as: Hp 2=Js
AKhard=tG
p, in which A and tG
are the exchange coecient and gradient thickness,respectively.
Obviously, the switching eld can bedecreased to an arbitrarily
small value if the gradi-ent thickness tG is large enough.
The rst epitaxial L10-FePt/Fe graded mediawere fabricated by
depositing a part of the Fe layerat elevated temperature.124,125 A
graded interface isformed between L10-FePt phase and Fe phase dueto
the interdiusion between layers at high tem-perature, which in turn
results in a continuouschange in magnetocrystalline anisotropy.
This is themain reason for coercivity reduction compared withthe
corresponding one with sharp interface. More-over, L10-CoPtTa2O5
and FePtTiO2 exchangecoupled multilayer media with well isolated
mag-netic grains were fabricated by adjusting Ta2O5 orTiO2 content
layer-by-layer.
126,127 In order tocontrol the ordering degree and anisotropy
gradi-ent, the FePtC graded lms can be fabricatedby varying the
substrate temperature layer bylayer.128,129
It is well known that higher substrate tempera-ture provides
kinetic energy for the interdiusionbetween the soft and hard
layers.130 Therefore, we
Fig. 24. Exchange coupling dependence of remnant coercivityand
thermal stability of [Co/PdSiO]16/PdSi/FeSiO lms withdierent PdSi
interlayer thickness.115
F. Wang, H. Xing & X. Xu
1530002-18
SPIN
201
5.05
. Dow
nloa
ded
from
ww
w.w
orld
scie
ntifi
c.com
by N
ATI
ON
AL
UN
IVER
SITY
OF
SIN
GA
PORE
on
04/2
9/15
. For
per
sona
l use
onl
y.
-
proposed using post-annealing method to formexchange coupled
graded media, which can be a veryeective method for realizing
interdiusion betweendierent layers.131 Graded lms should be
easilyobtained by changing the annealing parameters,such as
temperature, heating rate, and time.132,133
Our group combined sputtering and post-annealingtreatment to
fabricate a series of L10-FePt gradedlms with continuously varying
anisotropy on glasssubstrates. Figure 25 shows the schematic of
threestructures from bilayer to multilayers before andafter
annealing.102 The graded thickness can betuned by the structure
design and annealing tem-perature of the multilayers, in which the
nonmag-netic layers are used to adjust the
anisotropydistribution.
Figures 26(a) and 26(b) show the compositionaldepth proles of
FeAu/FePt lms before and afterannealing at 550C. Clearly, the
graded thicknesscan be increased due to the interdiusion when
post-annealing was carried out.134 In comparison to thelms with
sharp interfaces, the graded interface ismore favorable for
coercivity reduction.135 More-over, the graded L10-FePt:C/Fe lms
with a con-tinuous variation in anisotropy were
realizedexperimentally by varying the C concentration inthe FePt
hard layer. Nonmagnetic C layer playsan important role in tailoring
the gradient ofanisotropy and weakening the intergranular ex-change
interaction.136,137 We also simulated themagnetization reversal
process of the graded lm bythe object oriented micromagnetic
framework(OOMMF) software.138 Figures 26(c) and 26(d)show the
simulated hysteresis loops and magneti-zation reversal process of
the multilayer gradedlms. The coercivity of the multilayer graded
lmdecreases gradually with the decrease of the Ku
dierence between Fe soft layer and FePt hardlayer. The thermal
stability of the graded media isonly determined by the anisotropy
of the hardestpart of the media. The magnetization reversal
pro-cess agrees with the domain-wall-assisted reversalmechanism.
Thereafter, L10-FePt(hard)/CoPt(soft)graded lms were fabricated by
post-annealingbased on the dierent ordering temperatures be-tween
L10-FePt and L10-CoPt.
139 Zha et al. pre-pared a series of FePtCu graded lm by
adoptingpost-annealing method.140,141 Lee et al. also dem-onstrated
that the pinning eld is proportional tothe Ku dierence of the hard
and soft layers,
142
which is in agreement with our results.138 Further-more, the
pinning eld can be eciently decreasedafter an additional annealing
step, which is due to aheat-induced phase transformation of iron
oxidepresent at the interface between the hard and softlayers.
Therefore, post-annealing treatment is con-sidered to be a feasible
approach to prepare ECGmedia with continuously varying anisotropy.
Suesset al. predicted that the graded media should becapable of
ultrahigh-density recording of up to510Tbit/in.2, if the grain size
is assumed to be3.2 nm.7 In short, graded media should be a
moreeective approach than ECC and ES to improvewritability of
L10-FePt recording media.
Compared with texture-tilting-assisted media,the
domain-wall-assisted media can control the co-ercivity more easily
by optimizing the structures ofthe lms. Also the anisotropy
gradient and inter-grain interactions can be eectively adjusted
bynonmagnetic additives to improve the SNR. Giventhese
considerations, the domain-wall-assistedmedia, especially ECG
media, is more eective forpromoting the writability without the
loss of ther-mal stability.
(a) Bilayer lm (b) Sandwich-like lm (c) Multilayer lm
Fig. 25. Schematic of the three structures before and after
annealing.102
Overcoming the Trilemma Issues of Ultrahigh Density PMR
1530002-19
SPIN
201
5.05
. Dow
nloa
ded
from
ww
w.w
orld
scie
ntifi
c.com
by N
ATI
ON
AL
UN
IVER
SITY
OF
SIN
GA
PORE
on
04/2
9/15
. For
per
sona
l use
onl
y.
-
5. Balance of the Trilemma Issues forL10-Fe(Co)Pt
PerpendicularRecording Media
To achieve ultrahigh areal density for L10-Fe(Co)Ptmedia,
several proposals have been addressed so farto solve the trilemma
problems, including the above-mentioned PPM, BPM, tilted media,
domain-wall-assisted media. However, either of them alone maynot be
enough to address completely the trilemmaissues of L10-Fe(Co)Pt.
Fortunately, the abovereviewed approaches do not necessarily have
to bemutually exclusive. In order to write information onsuch high
anisotropy L10-Fe(Co)Pt, some form ofassisted recording that can be
switched at a su-ciently low applied eld have been used on BPM
tobalance the trilemma issues. If L10-Fe(Co)Pt lmswith ECC
structure is combined with BPM to formECC/BPM system, L10-Fe(Co)Pt
hard section andsoft section in ECC structure can ensure the
high
thermal stability and low write eld, while the bitpatterning can
provide the high signal-to-noiseratio.
Several groups have carried out relevant works onECC/BPM.
McCallum et al. prepared L10-FePt-based ECC/BPM with 180 nm pillars
on a 300 nmpitch representing a density of about 8Gbit/in.2
bye-beam lithography into a hard mask and subse-quent ion
milling.143 A 2.5-fold Hc reduction isobtained in this combination
system. Moreover, theswitching eld distribution (SFD) is
signicantlyreduced in ECC/BPM structures compared to thatof the
FePt single layer BPM system. The reductionin Hc and SFD observed
experimentally in thesestructures is consistent with micromagnetic
simu-lations that conrm a vertically incoherent pillarreversal from
the top to the bottom. The magneti-zation reversal of an areal
density of 1.5Tbit/in.2
ECC/BPM was simulated.144 The magnetic an-isotropy distribution
of ECC/BPM has a direct
(a) (b)
(c) (d)
Fig. 26. (a) and (b) Compositional depth proles of FeAu/FePt lms
before and after annealing at 550C, (c) and (d) simulatedhysteresis
loops and magnetization reversal process of the multilayer graded
lms.134,138
F. Wang, H. Xing & X. Xu
1530002-20
SPIN
201
5.05
. Dow
nloa
ded
from
ww
w.w
orld
scie
ntifi
c.com
by N
ATI
ON
AL
UN
IVER
SITY
OF
SIN
GA
PORE
on
04/2
9/15
. For
per
sona
l use
onl
y.
-
inuence on the switching eld distribution.L10-FePt ECC/BPM
nanopillar media were alsofabricated by continuously varying the
substratetemperature and then followed by electron-beamlithography
and ion milling.145 Moreover, FePt-based ECC/BPM with 31 nm bit
size and 37 nmpitch size were fabricated using diblock
copolymerlithography on 3 inch wafer by Wang et al.146,147
L10-CoPt/Ni composite nanowires with the diame-ter of 25 nm and
the length of 80 nm were fabricatedsuccessfully on AAO templates by
electrochemicaldeposition and post-annealing.148L10-CoPt/Ni
soft/hard composite nanowires exhibit an intermediatecoercivity of
1.96 kOe between those of CoPt array(10.97 kOe) and Ni array
(242Oe). Such a largereduction in coercivity leads to easier data
writing,showing a potential application of AAO templatesin
self-assembled media. Similarly, well-coupled L10-FePt/Fe and
L10-CoPt/FeCo composite nanotubeshave been prepared in AAO
templates.149,150
Furthermore, Goll et al. reported the large-areacomposite
L10-FePt/A1-FePt patterns fabricatedby ultraviolet nanoimprint
lithography in combi-nation with ICP reactive Ar-ion etching
ap-proach.142,151 It is clear that combination of ECCand BPM is an
attractive way to balance thetrilemma of PMR.
By now the exchange coupled graded media isvery eective in
reducing the write eld among do-main-wall-assisted media. Moreover,
L10-Fe(Co)Pt-based ECC/BPM combinations have made some
progress in recent years. Due to the advantages interms of ECG
and ECC/BPM, introducing ECGstructure into BPM to form ECG/BPM
structurewould be a promising media to pursue ultrahighareal
density in the future. This can provide opti-mum balance for the
trilemma of L10-based mag-netic recording media. In fact, Krone et
al. havesimulated that the switching eld of graded patternscan be
successively decreased with increasing num-ber of layers in the ECC
stack.152 A route for nar-rowing the switching eld distribution of
the bitarray is provided as well, which is vital for the
ap-plicability of the BPM concept in magnetic datastorage. Skomski
et al. have theoretically investi-gated how the magnetization
reversal processes ingraded recording media with columnar
structureaect the write eld and the areal density.153 Byusing
longer pillars, the write eld can be madearbitrarily small.
However, there is an optimumlength, beyond which writing becomes
dicultagain.
Up to now, there is little attempt to make ECG/BPM in
experiments, due to the great challengeassociated with its
fabrication. Combining ECGwith self-organized media might be a new
routeto overcome such challenges. Recently, Goll andBublat provided
a review on L10-FePt-based ECC/BPM, and the development of the
recording densityin conventional and advanced magnetic hard
diskdrives are shown in Fig. 27.154,155 It is unlikely thatthe
trilemma issue can be addressed by the material
Fig. 27. Development of the recording density in conventional
and advanced magnetic hard disk drives.155
Overcoming the Trilemma Issues of Ultrahigh Density PMR
1530002-21
SPIN
201
5.05
. Dow
nloa
ded
from
ww
w.w
orld
scie
ntifi
c.com
by N
ATI
ON
AL
UN
IVER
SITY
OF
SIN
GA
PORE
on
04/2
9/15
. For
per
sona
l use
onl
y.
-
engineering of the media alone. Besides ECC, energyassisted
recording such as HAMR and MAMR156158
is one of the promising schemes to combine withBPM in order to
overcome the writing eld limitof L10-Fe(Co)Pt-based media. These
combinedapproaches may very well be what are ultimatelyneeded to
push the areal density towards 10Tbit/in2.However, there are still
some signicant engineeringchallenges that need to be
resolved.159,160 Althoughmany attempts have been made to balance
the tri-lemma of L10-Fe(Co)Pt, there is still some
technicalproblems to restrict their industrial application asthe
next generation of recording media, especiallythe grain size. In
short, coordinated eorts fromboth the materials engineering and
technologies areneeded to balance the trilemma issues for
futurePMR.
6. Summary
In summary, many proposals have been made onmedia materials
engineering for L10-Fe(Co)Pt inorder to balance the trilemma of
perpendicular re-cording media. For the thermal stability,
stress-assisted growth and metal-doping methods are usedto reduce
the ordering temperature of L10-Fe(Co)Pt lm and obtain the perfect
fct (001) texture thatensure high thermal stability of the media.
For theSNR, GPM, PPM and BPM were designed to en-hance it from
dierent levels. Among them, BPM isconsidered to be the most
promising scheme to re-alize high SNR without a loss of thermal
stability.For the writability, both texture-tilting-assistedmedia
and domain-wall-assisted media can realizeits improvement on
materials engineering. In con-trast, the domain-wall-assisted
media, especially forECG media, is thought to be a more
eectiveapproach. However, it is necessary to combinesome
alternatives to balance the trilemma forL10-Fe(Co)Pt perpendicular
recording media dueto the shortages of single technology. Based on
theprogress of ECG and BPM, it is predicted thatL10-Fe(Co)Pt based
ECG/BPM should be one ofthe most eective paths to balance the
trilemmafrom the materials design, which would open upa new avenue
to realize an areal density of510Tbit/in.2 in the coming years.
Certainly, thereexists still a great challenge in production
technol-ogy, needing a synergic advance on the key tech-nologies of
media and heads.
Acknowledgments
The authors would like to acknowledge the usefuldiscussions with
Prof. Hao Zeng of University atBualo-SUNY and Prof. Dan Wei of
TsinghuaUniversity. The work was supported by the Na-tional Natural
Science Foundation of China (GrantNos. 51025101, 51101095,
11274214, 61434002), the863 Program (Grant No. 2014AA032904),
Founda-tions from the Ministry of Education of China(Grant Nos.
IRT1156, 20121404130001), ShanxiProvince Foundations (Grant Nos.
[2012]12, [2012]10, [2013]9).
References
1. B. Marchon, T. Pitchford, Y. T. Hsia and S.Gangopadhyay,
Advances in Tribology 2013,521086 (2013).
2. M. N. Baibich, J. M. Broto, A. Fert, F. Nguyen VanDau, F.
Petro, P. Etienne, G. Creuzet, A. Frie-derich and J. Chazelas,
Phys. Rev. Lett. 61, 2472(1988).
3. G. Binasch, P. Grnberg, F. Saurenbach and W.Zinn, Phys. Rev.
B 39, 4828 (1989).
4. Z. Y. Zhang, Y. C. Feng, T. Clinton, G. Badran,N. H. Yeh, G.
Tarnopolsky, E. Girt, M. Munteanu,S. Harkness, H. Richter, T.
Nolan, R. Ranjan,S. Hwang, G. Rauch, M. Ghaly, D. Larson,
E.Singleton, V. Vas'ko, J. Ho, F. Stageberg, V. Kong,K. Duxstad and
S. Slade, IEEE Trans. Magn. 38,1861 (2002).
5. S. Iwasaki and K. Takemura, IEEE Trans. Magn.11, 1173
(1975).
6. D. Weller, A. Moser, L. Folks, M. E. Best, L. Wen,M. F.
Toney, M. Schwickert, J. U. Thiele and M. F.Doerner, IEEE Trans.
Magn. 36, 10 (2000).
7. H. J. Richter, J. Magn. Magn. Mater. 321, 467(2009).
8. R. L. Comstock, J. Mater. Sci.: Mater. Electron.13, 509
(2002).
9. D. J. Sellmyer, M. Yu and R. D. Kirby, Nanostruct.Mater. 12,
1021 (1999).
10. H. J. Richter, J. Phys. D: Appl. Phys. 32, R147(1999).
11. O. Heinonen and K. Z. Gao, J. Magn. Magn. Mater.320, 2885
(2008).
12. H. J. Richter and A. Y. Dobin, J. Magn. Magn.Mater. 287, 41
(2005).
13. M. R. Visokay and R. Sinclair, Appl. Phys. Lett. 66,1692
(1995).
14. S. Stavroyiannis, I. Panagiotopoulos, D. Niarchos,J. A.
Chistodoulides, Y. Yang and G. C. Hadjipa-nayis, Appl. Phys. Lett.
73, 3453 (1999).
F. Wang, H. Xing & X. Xu
1530002-22
SPIN
201
5.05
. Dow
nloa
ded
from
ww
w.w
orld
scie
ntifi
c.com
by N
ATI
ON
AL
UN
IVER
SITY
OF
SIN
GA
PORE
on
04/2
9/15
. For
per
sona
l use
onl
y.
-
15. H. Zeng, M. L. Yan, N. Powers and D. J. Sellmyer,Appl. Phys.
Lett. 80, 2350 (2002).
16. R. F. C. Farrow, D. Weller, R. F. Marks, M. F.Toney, S. Hom,
G. R. Harp and A. Cebollada,Appl. Phys. Lett. 69, 1166 (1996).
17. T. Shima, T. Moriguchi, S. Mitani and K. Taka-nashi, Appl.
Phys. Lett. 80, 288 (2002).
18. Y. N. Hsu, S. Jeong, D. E. Laughlin and D. N.Lambeth, J.
Appl. Phys. 89, 7068 (2001).
19. Y. N. Hsu, S. Jeong, D. E. Laughlin and D. N.Lambeth, J.
Magn. Magn. Mater. 260, 282 (2003).
20. X. H. Xu, H. S. Wu, F. Wang and X. L. Li, Appl.Surf. Sci.
233, 1 (2004).
21. X. H. Xu, H. S. Wu, F. Wang and X. L. Li, ThinSolid Films
472, 222 (2005).
22. E. Manios, V. Karanasos, D. Niarchos and I.Panagiotopoulos,
J. Magn. Magn. Mater. 272276, 2169 (2004).
23. X. H. Xu, Z. G. Yang and H. S. Wu, J. Magn.Magn. Mater. 295,
106 (2005).
24. T. Suzuki, K. Harada, N. Honda and K. Ouchi, J.Magn. Magn.
Mater. 193, 85 (1999).
25. Y. F. Xu, J. S. Chen and J. P. Wang, Appl. Phys.Lett. 80,
3325 (2002).
26. J. L. Tsai, Q. S. Luo, P. R. Chen and Y. H. Chen, J.Appl.
Phys. 115, 17B715 (2014).
27. J. L. Tsai, Y. T. Tseng, C. R. Li and S. C. Fu, J.Appl.
Phys. 117, 17C502 (2015).
28. C. Feng, B. H. Li, G. Han, J. Teng, Y. Jiang, Q. L.Liu and
G. H. Yu, Appl. Phys. Lett. 88, 232109(2006).
29. C. C. Chiang, C. H. Lai and Y. C. Wu, Appl. Phys.Lett. 88,
152508 (2006).
30. Y. Zhu and J. W. Cai, Appl. Phys. Lett. 87,
032504(2005).
31. A. Hotta, T. Ono, M. Hatayama, K. Tsumura, N.Kikuchi, S.
Okamoto, O. Kitakami and T. Shi-matsu, J. Appl. Phys. 115, 17B712
(2014).
32. H. Zeng, M. L. Yan, N. Powers and D. J. Sellmyer,Appl. Phys.
Lett. 80, 2350 (2002).
33. M. L. Yan, N. Powers and D. J. Sellmyer, J. Appl.Phys. 93,
8292 (2003).
34. J. S. Kim, Y. M. Koo, B. J. Lee and S. R. Lee, J.Appl. Phys.
99, 053906 (2006).
35. H. G. Dang, L. W. Liu, L. Hao, T. L. Jin, M. F. Liu,J. W.
Cao, J. M. Bai, Y. Wang and F. L. Wei, J.Appl. Phys. 115, 17B117
(2014).
36. T. Maeda, T. Kai, A. Kikitsu, T. Nagase and J.Akiyama, Appl.
Phys. Lett. 80, 2147 (2002).
37. T. Maeda, A. Kikitsu, T. Kai, T. Nagase, H.Aikawa and J.
Akiyama, IEEE Trans. Magn. 38,2796 (2002).
38. Y. K. Takahashi, M. Ohnuma and K. Hono, J.Magn. Magn. Mater.
246, 259 (2002).
39. H. Wang, S. X. Xue, F. J. Yang, H. B. Wang, X.Cao, J. A.
Wang, Y. Gao, Z. B. Huang, C. P. Yang,W. Y. Cheung, S. P. Wong, Q.
Li and Z. Y. Li, ThinSolid Films 505, 77 (2006).
40. C. L. Platt, K. W. Wierman, E. B. Svedberg, R.Veerdonk, J.
K. Howard, A. G. Roy and D. E.Laughlin, J. Appl. Phys. 92, 6104
(2002).
41. K. Barmak, J. Kim, D. C. Berry, K. W. Wierman,E. B. Svedberg
and J. K. Howard, J. Appl. Phys.95, 7486 (2004).
42. Y. K. Takahashi and K. Hono, Scripta Mater. 53,403
(2005).
43. C. Y. You, Y. K. Takahashi and K. Hono, J. Appl.Phys. 100,
056105 (2006).
44. O. Kitakami, Y. Shimada, K. Oikawa, H. Daimonand K.
Fukamichi, Appl. Phys. Lett. 78, 1104(2001).
45. S. R. Lee, S. Yang, Y. K. Kim and J. G. Na, Appl.Phys. Lett.
78, 4001 (2001).
46. H. J. Richter, J. Phys. D 32D, 147 (1999).47. O. Chubykalo,
B. Lengseld, J. Kaufman and B.
Jones, J. Appl. Phys. 91, 3129 (2002).48. J. A. Christodoulides,
Y. Huang, Y. Zhang, G. C.
Hadjipanayis, I. Panagiotopoulos and D. Niarchos,J. Appl. Phys.
87, 6938 (2000).
49. X. H. Xu, H. S. Wu, X. L. Li and F. Wang, Mater.Chem. Phys.
90, 95 (2005).
50. M. Daniil, P. A. Farber, H. Okumura, G. C. Had-jipanayis and
D. Weller, J. Magn. Magn. Mater.246, 297 (2002).
51. T. P. Nolan, J. D. Risner, S. D. Harkness, I. E. Girt,S. Z.
Wu, G. Ju and R. Sinclair, IEEE Trans.Magn. 43, 639 (2007).
52. G. A. Bertero, D. Wachenschwanz, S. Malhotra, S.Velu, B.
Bian,D. Staord,Y.Wu,T.Yamashita andS. X. Wang, IEEE Trans. Magn.
38, 1627 (2002).
53. C. P. Luo and D. J. Sellmyer, Appl. Phys. Lett. 75,3162
(1999).
54. C. P. Luo, S. H. Liou, L. Gao, Y. Li and D. J.Sellmyer,
Appl. Phys. Lett. 77, 2225 (2000).
55. M. L. Yan, H. Zeng, N. Powers and D. J. Sellmyer,J. Appl.
Phys. 91, 8471 (2002).
56. H. Zeng, R. Sabirianov, O. Mryasov, M. L. Yan, K.Cho and D.
J. Sellmyer, Phys. Rev. B 66, 184425(2002).
57. M. Watanabe, T. Masumoto, D. H. Ping and K.Hono, Appl. Phys.
Lett. 76, 3971 (2000).
58. X. H. Xu, X. L. Li, F. Wang, F. X. Jiang and H. S.Wu, Appl.
Surf. Sci. 252, 2910 (2006).
59. X. H. Xu, T. Jin, X. L. Li, F. Wang, F. X. Jiang,Z. G. Yang
and H. S. Wu, Mater. Chem. Phys. 98,447 (2006).
60. X. H. Xu, T. Jin, X. L. Li, F. Wang, F. X. Jiang andH. S.
Wu, Thin Solid Films 515, 3936 (2007).
Overcoming the Trilemma Issues of Ultrahigh Density PMR
1530002-23
SPIN
201
5.05
. Dow
nloa
ded
from
ww
w.w
orld
scie
ntifi
c.com
by N
ATI
ON
AL
UN
IVER
SITY
OF
SIN
GA
PORE
on
04/2
9/15
. For
per
sona
l use
onl
y.
-
61. X. H. Xu, T. Jin, H. S. Wu, F. Wang, X. L. Li andF. X.
Jiang, Thin Solid Films 515, 5471 (2007).
62. J. S. Chen, B. C. Lim, J. F. Hu, B. Liu, G. M. Chowand G.
Ju, Appl. Phys. Lett. 91, 132506 (2007).
63. J. G. Zhu and Y. H. Tang, J. Appl. Phys. 99,08Q903
(2006).
64. J. G. Zhu and Y. H. Tang, IEEE Trans. Magn. 43,687
(2007).
65. D. E. Laughlin, Y. G. Peng, Y. L. Qin, M. Lin andJ. G. Zhu,
IEEE Trans. Magn. 43, 693 (2007).
66. Y. H. Tang and J. G. Zhu, IEEE Trans. Magn. 42,2360
(2006).
67. Y. L. Qin, D. E. Laughlin, Y. G. Peng and J. G.Zhu, IEEE
Trans. Magn. 43, 2136 (2007).
68. A. C. Sun, J. H. Hsu, P. C. Kuo and H. L. Huang,IEEE Trans.
Magn. 43, 2130 (2007).
69. M. T. Rahman, N. N. Shams, C. H. Lai, J. Fidlerand D. Suess,
Phys. Rev. B 81, 014418 (2010).
70. M. T. Rahman, C. H. Lai, D. Vokoun and N. N.Shams, IEEE
Trans. Magn. 43, 2133 (2007).
71. M. T. Rahman, N. N. Shams, Y. C. Wu, C. H. Laiand D. Suess,
Appl. Phys. Lett. 91, 132505 (2007).
72. C. Schulze, M. Faustini, J. Lee, H. Schletter, M. U.Lutz, P.
Krone, M. Gass, K. Sader, A. L. Bleloch,M. Hietschold, M. Fuger, D.
Suess, J. Fidler, U.Wol, V. Neu, D. Grosso, D. Makarov and
M.Albrecht, Nanotechnology 21, 495701 (2010).
73. S. Y. Chou and P. R. Krauss, J. Magn. Magn.Mater. 155, 151
(1996).
74. S. Y. Chou, M. S. Wei, P. R. Krauss and P. B.Fischer, J.
Appl. Phys. 76, 6673 (1994).
75. E. A. Dobisz, Z. Z. Bandic, T. W. Wu andT. Albrecht, Proc.
IEEE 96, 1836 (2008).
76. B. Terris and T. Thomson, J. Phys. D 38, R199(2005).
77. R. Sbiaa and S. N. Piramanayagam, RecentPatents Nanotechnol.
1, 29 (2007).
78. J. I. Martin, J. Nogues, K. Liu, J. L. Vicent andI. K.
Schuller, J. Magn. Magn. Mater. 256, 449(2003).
79. A. Kikitsu, Y. Kamata, M. Sakurai and K. Naito,IEEE Trans.
Magn. 43, 3685 (2007).
80. T. Thurn-Albrecht, J. Schotter, C. A. Kastle, N.Emley, T.
Shibauchi, L. Krusin-Elbaum, K. Guar-ini, C. T. Black, M. T.
Tuominen and T. P. Russell,Science 290, 2126 (2000).
81. H. Masuda, H. Yamada, M. Satoh, H. Asoh, M.Nakao and T.
Tamamura, Appl. Phys. Lett. 71,2770 (1997).
82. C. Kim, T. Loedding, S. Jang, H. Zeng, Y. Sui andD. J.
Sellmyer, Appl. Phys. Lett. 91, 172508 (2007).
83. A. I. Gapin, X. R. Ye, J. F. Aubuchon, L. H. Chen,Y. J. Tang
and S. Jin, J. Appl. Phys. 99, 08G902(2006).
84. J. Fu, S. Cherevko and C. H. Chung, Electrochem.Commun. 10,
514 (2008).
85. W. Lee, H. Han, A. Lotnyk, M. A. Schubert, S.Senz, M. Alexe,
D. Hesse, S. Baik and U. Gosele,Nat. Nanotechnol. 3, 402
(2008).
86. K. Naito, H. Hieda, M. Sakurai, Y. Kamata and K.Asakawa,
IEEE Trans. Magn. 38, 1949 (2002).
87. V. Sundar, J. X. Zhu, D. E. Laughlin and J. G. Zhu,Nano
Lett. 14, 1609 (2014).
88. H. Wang, H. B. Zhao, P. Quarterman and J. P.Wang, Appl.
Phys. Lett. 103, 052406 (2013).
89. K. Z. Gao, X. Wang and H. N. Bertram, J. Appl.Phys. 93, 7840
(2003).
90. K. Z. Gao and H. N. Bertram, IEEE Trans. Magn.38, 3675
(2002).
91. Y. F. Zheng, J. P. Wang and V. Ng, J. Appl. Phys.92, 8007
(2002).
92. J. P. Wang, Y. Y. Zou, C. H. Hee, T. C. Chong andY. F.
Zheng, IEEE Trans. Magn. 39, 1930 (2003).
93. J. P. Wang, Nat. Mater. 4, 191 (2005).94. Y. Y. Zou, J. P.
Wang, C. H. Hee and T. C. Chong,
Appl. Phys. Lett. 82, 2473 (2003).95. S. Okamoto, O. Kitakami,
T. Miyazaki, Y. Shi-
mada, Y. K. Takahashi and K. Hono, J. Appl.Phys. 96, 5217
(2004).
96. C. L. Zha, B. Ma, Z. Z. Zhang, T. R. Gao, F. X. Ganand Q. Y.
Jin, Appl. Phys. Lett. 89, 022506 (2006).
97. M. Albrecht, G. H. Hu, I. L. Guhr, T. C. Ulbrich,
J.Boneberg, P. Leiderer and G. Schatz, Nat. Mater.4, 203
(2005).
98. B. Liu and D. Weller, US patent, US7468214B2(2008).
99. L. J. Guan and J. G. Zhu, J. Appl. Phys. 93, 7735(2003).
100. G. Varvaro, E. Agostinelli, S. Laureti, A. M. Testa,A.
Generosi, B. Paci and V. R. Albertini, IEEETrans. Magn. 44, 643
(2008).
101. Y. C. Wang, P. Sharma and A. Makino, J. Phys.:Condens.
Matter. 24, 076004 (2012).
102. F. Wang and X. H. Xu, Chin. Phys. B 23, 036802(2014).
103. D. Suess, T. Schre, S. Fhler, M. Kirschner,G. Hrkac, F.
Dorfbauer and J. Fidler, Appl. Phys.Lett. 87, 012504 (2005).
104. G. Asti, M. Ghidini, R. Pellicelli, C. Pernechele,M. Solzi,
F. Albertini, F. Casoli, S. Fabbrici andL. Pareti, Phys. Rev. B 73,
094406 (2006).
105. D. Suess, J. Lee, J. Fidler and T. Schre, J. Magn.Magn.
Mater. 321, 545 (2009).
106. F. B. Hagedorn, J. Appl. Phys. 41, 2491 (1970).107. A. Y.
Dobin and H. J. Richter, J. Appl. Phys. 101,
09K108 (2007).108. D. Goll, S. Macke and H. Kronmuller, Physica
B
403, 338 (2008).
F. Wang, H. Xing & X. Xu
1530002-24
SPIN
201
5.05
. Dow
nloa
ded
from
ww
w.w
orld
scie
ntifi
c.com
by N
ATI
ON
AL
UN
IVER
SITY
OF
SIN
GA
PORE
on
04/2
9/15
. For
per
sona
l use
onl
y.
-
109. H. Kronmuller and D. Goll, Physica B 403, 237(2008).
110. F. Casoli, F. Albertini, S. Fabbrici, C. Bocchi, L.Nasi, R.
Cabassi and L. Pareti, IEEE Trans. Magn.41, 3877 (2005).
111. F. Casoli, F. Albertini, L. Nasi, S. Fabbrici, R.Cabassi,
F. Bolzoni and C. Bocchi, Appl. Phys.Lett. 92, 142506 (2008).
112. F. Casoli, F. Albertini, L. Nasi, S. Fabbrici, R.Cabassi,
F. Bolzoni, C. Bocchi and P. Luches, ActaMater. 58, 3594
(2010).
113. R. H. Victora and X. Shen, IEEE Trans. Magn. 41,537
(2005).
114. R. H. Victora and X. Shen, IEEE Trans. Magn. 41,2828
(2005).
115. J. P. Wang, W. K. Shen and J. M. Bai, IEEETrans. Magn. 41,
3181 (2005).
116. J. P. Wang, W. K. Shen and S. Y. Hong, IEEETrans. Magn. 43,
682 (2007).
117. K. C. Schuermann, J. D. Dutson, S. Z. Wu, S. D.Harkness, B.
Valcu, H. J. Richter, R. W. Chantrelland K. O'Grady, J. Appl. Phys.
99, 08Q904 (2006).
118. A. Berger, N. Supper, Y. Ikeda, B. Lengseld, A.Moser and E.
E. Fullerton, Appl. Phys. Lett. 93,122502 (2008).
119. T. Hauet, S. Florez, D. Margulies, Y. Ikeda, B.Lengseld, N.
Supper, K. Takano, O. Hellwig andB. D. Terris, Appl. Phys. Lett.
95, 222507 (2009).
120. T. Hauet, E. Dobisz, S. Florez, J. Park, B. Lengs-eld, B.
D. Terris and O. Hellwig, Appl. Phys. Lett.95, 262504 (2009).
121. G. Choe, Y. Ikeda, K. Zhang, K. Tang and M.Mirzamaani, IEEE
Trans. Magn. 45, 2694 (2009).
122. K. Tang, X. Bian, G. Choe, K. Takano, M. Mir-zamaani, G.
Wang, J. Zhang, Q. F. Xiao, Y. Ikeda,J. Risner-Jamtgaard and X. Xu,
IEEE Trans.Magn. 45, 786 (2009).
123. D. Suess, Appl. Phys. Lett. 89, 113105 (2006).124. D. Goll,
A. Breitling, L. Gu, P. A. Aken and W.
Sigle, J. Appl. Phys. 104, 083903 (2008).125. D. Goll, A.
Breitling and S. Macke, IEEE Trans.
Magn. 44, 3472 (2008).126. K. K. M. Pandey, J. S. Chen, G. M.
Chow and J. F.
Hu, Appl. Phys. Lett. 94, 232502 (2009).127. C. J. Jiang, J. S.
Chen, J. F. Hu and G. M. Chow,
J. Appl. Phys. 107, 123915 (2010).128. J. S. Chen, L. S. Huang,
J. F. Hu, G. Ju and G. M.
Chow, J. Phys. D: Appl. Phys. 43, 185001 (2010).129. L. S.
Huang, J. F. Hu, G. M. Chow and J. S. Chen,
J. Appl. Phys. 109, 063910 (2011).130. Y. H. Lin, J. H. Hsu, F.
T. Yuan, P. C. Kuo and
J. K. Mei, IEEE Trans. Magn. 49, 3679 (2013).131. X.H.Xu,B.H.
Lv,F.Wang,X.L. Li, F.X. Jiang and
H. S. Wu, J. Magn. Magn. Mater. 310, 2653 (2006).
132. L. Manna, E. C. Scher, L. S. Li and A. P. Alivisatos,J. Am.
Chem. Soc. 124, 7136 (2002).
133. K. H. Cai, C. Li, Y. Zhang, J. F. Xu, H. K. Lai andS. Y.
Chen, Appl. Surf. Sci. 254, 5363 (2008).
134. F. Wang, X. H. Xu, Y. Liang, J. Zhang and H. S.Wu, Appl.
Phys. Lett. 95, 022516 (2009).
135. F. Wang, X. H. Xu, Y. Liang, J. Zhang andJ. Zhang, Mater.
Chem. Phys. 126, 843 (2011).
136. F. Wang, J. Zhang, J. Zhang and X. H. Xu, J. Appl.Phys.
109, 07B731 (2011).
137. F. Wang, J. Zhang, J. Zhang, C. L. Wang, Z. F.Wang, H.
Zeng, M. G. Zhang and X. H. Xu, Appl.Surf. Sci. 271, 390
(2013).
138. J. Zhang, Y. Liu, F. Wang, J. Zhang, R. Q. Zhang,Z. F. Wang
and X. H. Xu, J. Appl. Phys. 111,073910 (2012).
139. V. Alexandrakis, D. Niarchos, K. Mergia, J. Lee, J.Fidler
and I. Panagiotopoulos, J. Appl. Phys. 107,013903 (2010).
140. C. L. Zha, R. K. Dumas, Y. Y. Fang, V. Bonanni, J.Nogues
and J. kerman, Appl. Phys. Lett. 97,182504 (2010).
141. Y. Y. Fang, R. K. Dumas, C. L. Zha and J. Aker-man, IEEE
Magn. Lett. 2, 5500104 (2011).
142. J. Lee, D. Makarov, C. Brombacher, B. Dymerska,D. Suess, M.
Albrecht and J. Fidler, Nanotechnol-ogy 25, 045604 (2014).
143. A. T. McCallum, P. Krone, F. Springer, C. Brom-bacher, M.
Albrecht, E. Dobisz, M. Grobis, D.Weller and O. Hellwig, Appl.
Phys. Lett. 98, 242503(2011).
144. P. Krone, D. Makarov, T. Schre and M. Albrecht,J. Appl.
Phys. 109, 103901 (2011).
145. J. Zhang, Z. Sun, J. Sun, S. Kang, S. Yu, G. Han,S. Yan, L.
Mei and D. Li, Appl. Phys. Lett. 102,152407 (2013).
146. H. Wang, W. M. Li, M. T. Rahman, H. B. Zhao,J. Ding, Y. J.
Chen and J. P. Wang, J. Appl. Phys.111, 07B914 (2012).
147. H. Wang, M. T. Rahman, H. B. Zhao, Y. Isowaki,Y. Kamata, A.
Kikitsu and J. P. Wang, J. Appl.Phys. 109, 07B754 (2011).
148. A. I. Gapin, X. R. Ye, L. H. Chen, D. Hong andS. Jin, IEEE
Trans. Magn. 43, 2451 (2007).
149. D. Zhou, M. G. Zhu, M. G. Zhou, Z. H. Guo and W.Li, J.
Appl. Phys. 109, 07B720 (2011).
150. H. L. Su, S. L. Tang, N. J. Tang, R. L. Wang, M. Luand Y.
W. Du, Nanotechnology 16, 2124 (2005).
151. T. Bublat and D. Goll, Nanotechnology 22, 315301(2011).
152. P. Krone, D. Makarov, T. Schre and M. Albrecht,Appl. Phys.
Lett. 97, 082501 (2010).
153. R. Skomski, T. A. George and D. J. Sellmyer, J.Appl. Phys.
103, 07F531 (2008).
Overcoming the Trilemma Issues of Ultrahigh Density PMR
1530002-25
SPIN
201
5.05
. Dow
nloa
ded
from
ww
w.w
orld
scie
ntifi
c.com
by N
ATI
ON
AL
UN
IVER
SITY
OF
SIN
GA
PORE
on
04/2
9/15
. For
per
sona
l use
onl
y.
-
154. D. Goll and T. Bublat, Phys. Status Solidi A 210,1261
(2013).
155. H. Kronmller and D. Goll, Phys. Status Solidi B248, 2361
(2011).
156. T.W. McDaniel, J. Appl. Phys. 112, 093920 (2012).157. T. J.
Fal and R. E. Camley, Appl. Phys. Lett. 97,
122506 (2010).158. M.V. Lubarda, S. Li, B. Livshitz, E.E.
Fullerton and
V. Lomakin, IEEE Trans. Magn. 47, 18 (2011).
159. T. Rausch, J. D. Trantham, A. S. Chu, H. Dak-roub, J. W.
Riddering, C. P. Henry, J. D. Kiely,E. C. Gage and J. W. Dykes,
IEEE Trans. Magn.49, 730 (2013).
160. X. B. Wang, K. Z. Gao, H. Zhou, A. Itagi, M.Seigler and E.
Gage, IEEE Trans. Magn. 49, 686(2013).
F. Wang, H. Xing & X. Xu
1530002-26
SPIN
201
5.05
. Dow
nloa
ded
from
ww
w.w
orld
scie
ntifi
c.com
by N
ATI
ON
AL
UN
IVER
SITY
OF
SIN
GA
PORE
on
04/2
9/15
. For
per
sona
l use
onl
y.
Overcoming the Trilemma Issues of Ultrahigh Density
Perpendicular Magnetic Recording Media by L10-Fe(Co)Pt Materials1.
Introduction2. Approaches to Enhancing Thermal Stability2.1.
Driving L10-Fe(Co)Pt phase transformation by stress-assisted
growth2.2. Driving L10-Fe(Co)Pt phase transformation by
metal-doping
3. Approaches to Improving SNR3.1. Granular perpendicular
media3.2. Percolated perpendicular media3.3. Bit patterned
media
4. Approaches to Promoting Writability4.1.
Texture-tilting-assisted magnetic recording4.2.
Domain-wall-assisted recording
5. Balance of the Trilemma Issues for L10-Fe(Co)Pt Perpendicular
Recording Media6. SummaryAcknowledgmentsReferences