evolution of perpendicular magnetic recording...
TRANSCRIPT
®
EVOLUTION OF PERPENDICULAR MAGNETIC RECORDING MEDIA AND UPCOMING CHALLENGES
Gerardo Bertero
Collaborators: Mike Mallary, Ram Acharya, Kumar
Srinivasan, Mrugesh Desai, Alex Chernyshov, Hua Yuan,
David Treves, Mike Madison, Bogdan Valcu, Eric Champion
Western Digital Corporation, San Jose, California, USA
IEEE Santa Clara Valley Chapter Oct. 25, 2011
®
Outline
Introduction
PMR Media
Evolution by Product Generation
Limitations Going Forward
Recording Technology Options
Microwave Assisted Magnetic Recording
Heat Assisted Magnetic Recording
HAMR Media Challenges
Bit Pattern Media (not covered here)
ASTC Industry Consortium
Summary and Discussions
2
®
Introduction
Introduced commercially in 2005, conventional
perpendicular recording can now be considered a
mature technology in its 6th (or 7th) product generation.
Alternative technologies such as Energy Assisted
Recording and Bit Patterned Media Recording are not
yet ready for productization.
Here we will address the key technical challenges
regarding media extendibility for CMR and will also
discuss the main issues in the development of media
for the next likely technologies.
3
®
Cross-Section TEM of Typical PMR Media Structure
Adhesion
SUL 1
Seed
Ru
NL 2
NL 1
SUL 2
Overcoat
Graded CoCrPtX-Oxide
• Advanced DS mag structure design
• Graded anisotropy
• Varying degrees of inter-granular
exchange coupling through the
thickness of grains
4
®
PMR Media Design By Generation
From the simple 2-layer capping layer media to the latest 7-8 layer
magnetic structures, PMR media has increased significantly in
complexity since it was first adopted.
Bottom
Cap
Mag 1
Cap
Mag 2
Seed
Cap
Seed
Cap
1st 2nd 5th
, 6th
2005 2010
Hard Magnetic Layers Only
5
®
ECC and Exchange-Spring Media
6
Inspired by the “tilted media” concept
Incoherent switching through column of grain
Compared to “Stoner-Wahlfarth” media, in its
simplest rendition it provides twice the energy
barrier with same switching field.
R.H. Victora and X. Shen, IEEE Transactions on Magnetics, 41, NO. 2, 537-542, (2005).
D. Suess et al., J. Magn. Magn. Mat, 551, 290-291, (2005).
®
By Courtesy of D. Suess
Dynamic-Spring Media (Media Write Field Boost)
Dynamic-spring media is an intrinsic magnetic self-assisted switching stack
The high Ku portion of the media provides for the required thermal stability
The softer portion of the media is used to adjust (lower) the switching field
This technology is currently being introduced into high volume manufacturing
Effective field to reverse storage
layer is a combination of:
1) Head Field
2) Assist Layer Ms*t
3) Interface Exchange
Assist layer
high Ms/low Ku
Storage layer
low Ms/high Ku
7
®
Graded Anisotropy Media
®
Micro-magnetic simulations predict further gains in performance by
extending the single ECC or Exchange-Spring concept to a system
where the anisotropy is graded.
Caveat is that switching has to take place by incoherent rotation for
advantage to be realized.
®
Advanced ECC Media
Concept:
Figure of merit = 2
Figure of merit > 4 Double EBL
Practice:
Single EBL
Two-Spin Model
Three-Spin Model
hard
Hc
J
K
CH
21
hard
Hc
J
KH
2
4
1
High Ku
Low Ku
High Ku
Low Ku
Gra
de
d A
nis
otr
op
y
C>4
9
®
10
Hs and KuV/kT as function of EBL 1/2
6800
7000
7200
7400
7600
7800
8000
8200
8400
8600
8800
150 200 250 300 350 400
Hs (
Oe)
KuV/kT
Single-EBL
Dual-EBL
EBL 1&2 = 0 nm
EBL 1 = 0 nm
EBL 2 = 0.7 nm
EBL 1 = increasing
EBL 2 = 0.7 nm
Double ECC structures allow for higher degree of freedom when
optimizing Hs and KuV/kT (i.e., writability and thermal stability)
®
ECC PMR Media
Double exchange-break layer media is latest generation media
structure
It allows higher degree of incoherent switching through the grain
column.
Provides higher design and performance optimization flexibility
Allows higher anisotropy (Ku) media to be used yet maintaining
good writability
Can result in thinner overall media stacks
It should be a key enabler for smaller media grains which is
required for extending conventional PMR recording past the 1
Tb/in2 mark.
11
®
Magnetic Grains and Bit Size
~40 Grains/Bit
~20 Grains/Bit
~11 Grains/Bit
~90 Grains/Bit
15 nm
500 Gb/in2
250 Gb/in2
800 Gb/in2
100 Gb/in2
12
To maintain SNR, number of grains per bit should ideally be kept constant
12
®
Magnetic Grains and Bit Size
13
To maintain SNR, number of grains per bit should ideally be kept constant
13
6
7
8
9
10
11
12
13
14
15
16
1 10 100 1000
Gra
in S
ize
(Ce
nte
r-T
o-C
en
ter)
, (n
m)
Areal Density (Gb/in2)
Perpendicular Longitudinal
Presented at TMRC 2010
®
Simple Picture of Grain Size Scaling in Media
Sa
WBPWSNR
2
5031.0
S
W
a
B
B
PWSNR
2
2
5031.0
~ constant over generations of products Decreasing with product generation
W: Read Width (MRW)
B: Bit Length
S: Cross-Track Corr. Length ~ Grain Size
a: Transition Parameter
Bit Length
n
GSa 7.0
Grain size will limit ultimately transition width
® 6
7
8
9
10
11
12
13
14
15
16
1 10 100 1000
Gra
in S
ize
(Ce
nte
r-T
o-C
en
ter)
, (n
m)
Areal Density (Gb/in2)
Perpendicular Longitudinal
Presented at TMRC 2010
®
Simple Picture of Grain Size Scaling in Media
®
Grain Size Limited Transitions
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
Gra
in S
ize
Lim
ite
d B
/a
Areal Density, Tb/in2
10 nm GS
9 nm GS
8 nm GS
7 nm GS
®
Grain Size Progression vs. Areal Density
ECC media structure was expected to yield higher areal density capability by
enabling smaller grain size in media.
In practice, grain size has remained practically unchanged since PMR introduction.
Instead, the trade-off of choice was to use ECC media to improve writability,
maintain acceptable adjacent track erasure, and optimize transition parameter with
utilization of higher Ku alloys and optimization of lateral exchange coupling through
the various layers. ®
6
7
8
9
10
11
12
13
14
15
16
1 10 100 1000
Gra
in S
ize
(C
en
ter-
To
-Cen
ter)
, (
nm
)
Areal Density (Gb/in2)
Perpendicular Longitudinal
®
0
10
20
30
40
50
60
70
80
90
100
110
120
2.53.54.55.56.57.58.59.5H
k [K
Oe]
Core Grain Size [nm]
CoCrPt-O (Ms=700 emu/cc)
FePt L10 (Ms=1100 emu/cc)
Grain Size and Ku Scaling
120-500
Gb/in2
700-900 Gb/in2
17
CoCrPt-O based
FePt(L10) based
1-1.5 Tb/in2
®
HDD Industry Historical Areal Density Trend
100 110 Production Year
1E-3
1E-2
1E-1
1E+0
1E+1
1E+2
1E+3
1E+4
1E+5
1E+6
HDD Products
Industry Lab Demos
Areal Density Perspective
AMR Head
60 70 80 90 2000 10
60% CGR
100% CGR
Perpendicular Recording
10 5
10 4
10 3
10 2
10 6
10
1
10 -1
10 -2
10 -3
25% CGR
(First Hard Disk Drive)
Thin Film Head
200 Million X
Increase
Thin Film Media
PRML Channels
GMR Heads
40% CGR PR Projection
TuMR Heads
AFC Media
Are
al D
ensi
ty M
egab
its/
in2
Are we slowing
down ?
18
®
Current Technology Is Reaching Maturity
2 Disk Mobile Historical Announces With Volume Ship of > 5M/Qtr
100
1000
Dec-0
6
Aug-0
7
Apr-
08
Dec-0
8
Sep-0
9
May-1
0
Jan-1
1
Sep-1
1
May-1
2
Feb-1
3
Oct-
13
Month-Year
HD
D C
ap
ac
ity
(G
By
tes
)
Company A
Company B
Company C
More than 5M/Qtr
200
500
400
300
700
600
800
900
Colored boxes show the press
announcement dates
purple cross show when
industry shipped more
than 5M/Qtr
80% CAGR enabled by large head-media
SNR gains
System features enable growth
19
®
®
Technology Options
SMR/TDMR Energy Assisted Recording Bit Patterned
Reader Shingle Writer
Progressive
Scan
Head
Motion
Laser Reader Spin-Torque
Oscillator Reader Write Pole
1 Bit = 1 Island Near Field Write Field
(Illustrations courtesy of Y. Shiroishi, Intermag 2009)
• Continued scaling requires innovations in systems technologies, materials science and process engineering to advance areal density
20
®
Technology Options for > 1.5 Tb/in2
21
0.10
1.00
10.00
20062007
20082009
20102011
20122013
20142015
20162017
20182019
Date (Year)
Are
al D
en
sit
y (
Tb
/in
2)
Conventional PMR (CM)
PMR Extendibility (Conventional & Shingle)
EAMR+BPM
MAMR
HAMR
TDMR
®
MAMR: Spin-Torque Oscillator Physics
The spin-transfer torque opposes the damping torque When both are equal, precession becomes
stable (condition “d” above)
Damping is energy loss per cycle (into lattice waves, spin waves, electron excitations, etc.)
Reduction of 30- 40% of media switching field
Effective gradients of 1500-3000 Oe/nm
22
®
Perpendicular Bias
Hitachi/NEDO STO Demonstration
WD Confidential 23
®
MAMR can extend the use of conventional PMR media. Higher Ku in
CoPtX-Oxide alloys must be realized.
With the higher Ku materials, media grain size must be decreased
aggressively for this technology to show its promise.
If MAMR is extendible to >2Tb/in2, new high Ku material systems will
be required
…but could be simpler than HAMR media as Curie temperature is not part of the
design criteria.
Companies now assessing MAMR’s potential. Not much energy has
been devoted to it across industry.
MAMR Technology and Media
24
®
Grain Size Expectations
25
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5
Are
al
Den
sit
y (
Tb
/in
2)
Grain Size CTC (nm)
CoCrPt-Oxide High Ku BPM
HAMR (granular
media)
5
15
25
35
45
55
65
45678910
Core Grain Size [nm]
Hk
[K
Oe
]
Anisotropy Field [KOe]
9 nm hard mag layer Thickness
70 KuV/kT
Conventional
High Ku (e.g., L10)
Grain size must come down to enable higher areal densities. HAMR
has potential to switch extremely high Ku materials.
®
HAMR Recording Fields
Y. S
hir
ois
hi, In
term
ag
2009)
By knowing T(x), M(x), Ku(x), Hk(x) and Happ (x) we can make simple
observations of the recording process for grains of different size.
Pole
T(x)
Hc(x)
Happ(x)
Media direction of Travel
Transition Location
dxdx
dTdT
dtVdx
V
26
®
Cross-sectional TEM Image of Experimental EAMR Media
Protective cap for FIB prep
FePt-X
Soft Magnetic Underlayer
Seed
Seed
Heat-Sink Layer Structure
®
Plan-View TEM
®
Heat Assisted Media Challenges
Grain Size and Morphology
Grain Size Uniformity
HK Uniformity
Grain-to-grain L10 atomic ordering uniformity
Grain-to-grain compositional uniformity
Intergranular Magnetic Exchange Uniformity
E.g., grain boundary width uniformity
Roughness
Thermal Design
Heat-sink and thermal barriers
Tc tailoring
Overcoat/Lubricant
28
®
Magnetization and Anisotropy as Function of Temperature
29
0.E+00
5.E+06
1.E+07
2.E+07
2.E+07
3.E+07
3.E+07
4.E+07
0
200
400
600
800
1000
1200
0 100 200 300 400 500 600 700 800
Man
geto
cry
sta
llin
e E
nerg
y, K
u (
erg
/cc)
MS(e
mu
/cc)
Temperature (K)
Ms (emu/cc)
Ku (erg/cc)
Knowing Ms(T) and Ku(T) allow us to estimate many aspects of the
writing process
®
EAMR Recording: Grain Size Effect on Transition Location
Happ(x)
Media direction of Travel
Hc(x)
Big Grain
Hc(x)
Small Grain
T>TB
Transition delta for big vs. small grains
TB: Blocking Temperature
Direction of applied field (consistent with magnetization in bit to be written)
Opposite orientation
Same orientation
30
®
Energy Balance for Grains at Either Side of Transition
Applied Field
M M
Applied Field
Magnetocrystalline: Ku x Sin2() x Vol
Zeeman : M.H= M x H x (Cos ) x Vol
Demagnetizing: 2 x x N x M2 x Vol
Magnetocrystalline + Zeeman - Demag Magnetocrystalline - Zeeman - Demag
For a particle (grain) to freeze magnetically its Energy Barrier/KbT must
be larger than a certain value.
31
®
-20
0
20
40
60
80
100
120
140
160
180
0 10 20 30 40 50 60 70 80
Distance from NFT Center (nm)
Eb
/kT
Eb/kT for written pattern
Upper Bound E Barrier/kT
Lower Bound E Barrier/kT
Energy Barriers During Writing
1T
1T
2T
2.M
(T).
H
2.M
(T).
H
2.M
(T).
H
Media motion
32
®
What Energy Barrier to Freeze Grains in Their State?
Contrary to conventional recording in HAMR, grains freeze magnetically once
their energy barrier is as low as 1.5-4 times kBT. Media grain volume difference
effects are magnified by this fact. Thermal gradients must be steep enough to
guarantee the freezing of all grains within a short distance from the intended
transition location.
0.001
0.01
0.1
1
0 1 2 3 4 5 6 7
Pro
ba
bil
ity o
f S
wit
ch
ing
Energy Barrier, Eb/kT
dxev
fP
x
xTk
xEb
B
0
)(
)(
0
33
sf
11011
0
®
Grain Size Sensitivity to Magnetization Freezing
As grain mean size is reduced, grain size distribution sigma must be
reduced significantly for linear density capability. Smaller grain size
does not automatically result in better performance.
34
450
480
510
540
570
600
630
1 3 5 7 9 11 13 15
Te
mp
era
ture
at
Eb
/kT
=2
C
ros
sin
g, K
Grain Size (Diameter), nm
)(
)(
)(
TMsMs
TKuKu
VolumeEE BB
2 4 6 8 10 12 14
Range
(DD)
Grain size
®
0.00
5.00
10.00
15.00
20.00
25.00
30.00
500 550 600 650 700
Temperature, K
Co
he
ren
cy
Le
ng
th, d
l, n
m
EAMR writing occurs here
FePt-L10 Coherency Length as Function of Temperature
At high temperature the domain wall width is significantly larger than at RT.
Increasing thickness beyond 10 nm may not necessarily increase energy barrier
based on this estimate of coherency length for FePt.
)(K
)A(4)(
u T
TTl d
T. Bublat and D. Goll, J. Appl. Phys. 108, 113910 (2010)
)(.)0(
117.0
2
)()(
3/2
0
0
0 TMContMg
TkTMTA s
s
B
B
bS
35
®
HAMR Media Grain Size Observations
Grains of smaller size require lower temperatures to magnetically freeze.
Under typical HAMR recording conditions freezing takes place near EB/kBT~1-3.
Grain size effects are magnified under EAMR (high T) conditions.
For grains in a given distribution to freeze within a distance corresponding to a fraction of the bit length, they must meet very tight size sigma requirements.
Sigma requirements for grain sizes in the 2-4 Tb/in2 are very stringent if Eb/kBT>>2 . What is the correct energy barrier to use ? (strongly dependent on media design).
Physics of high speed switching near Tc have not been sufficiently studied.
36
®
Hk Fluctuations near Tc
The promise of heat assisted recording is rooted on the premise of extremely high thermal gradients to write transitions much sharper than possible by just magnetic means.
In order to capitalize on the high switching effective gradients and achieve high recording linear density, The magnetic anisotropy of the media must be as high as possible
The Curie point of the media must be as low as possible (but safely higher than RT for archivability)
FePt is an nearly ideal system for satisfying both conditions above.
However, we assume time averaged properties for anisotropy vs. temperature of the media.
37
There are some suspicions that temporal fluctuations in Hk near Tc can be large and relevant during the writing process.
Modeling by R.Victora, UMN, support this conjecture, with particle size being a big factor.
If experiments confirm this behavior, ways to minimize the negative effects through novel layer structure designs will be needed.
®
Ultrafast Measurements at LCLS
Experimentally, these fluctuations could be time resolved by new tools
such as the LCLS facility at Stanford with femto-second resolution.
38
®
New Industry Consortium
“Advanced Storage Technology Consortium”
New consortium created to expand and enhance the power of R&D
funding Acceleration of technology development by targeted collaborations between storage industry
participants, suppliers, universities, laboratories, and institutes
Mission: member-directed, scalable R&D organization to address
fundamental technology challenges
Supply chain involvement
HDD technology roadmap
®
Why Do We Need Something Different ?
Pace of technology transitions and scope of change required means business as usual approach won’t work
Need to collimate and focus entire industry R&D to be successful
Need coordinated transitions in supply base - components, equipment, and materials
®
Expected Outputs from ASTC
Focused, collaborative research projects that will enable better understanding of key scientific challenges
Shared, realistic roadmap for the Industry
Solutions – science to engineering to manufacturing options – that will shorten time from invention to productization
41
®
ASTC Funding Allocations
By Area
HAMR BPM MAMR TDMR ADV. PMR
29 projects preselected for 1st year of funding
Systems Media Heads Signal Proc. HDI Tools
By System
42
®
Conventional perpendicular media will continue to evolve to higher
Ku and smaller grain size as we extend CPR technology.
Magnetic head fields and Ku limits of CoPtX-Oxide media will limit
densities to 1-1.5 Tb/in2 range. Shingling may help extend that
range.
Media for heat assisted recording based on FePt has improved
significantly but big challenges remain.
Reducing grain size and size distributions while preserving columnar morphology remain
challenging given the high processing temperatures required for L10 ordering.
Optimum heat-sink design of media to enhance thermal gradients is limited somewhat by
other functional requirements such as low surface roughness.
Pace of development needs to increase in order to productize these
new technologies as per roadmaps.
New ASTC consortium provides a forum for more open
collaboration in addressing the most challenging industry problems.
Summary And Conclusions
43