Information Storage Industry Consortium

Information Storage Industry Consortium THE HDD TECHNOLOGY ROADMAP PMR AND BEYOND Paul D. Frank December 7, 2006

WHO WE ARE INSIC the Information Storage Industry Consortium the collaborative research consortium for the worldwide information storage industry

WHO WE ARE What INSIC is: - An international storage technology research consortium What INSIC does: - Organizes & manages high-risk, pre-competitive competitive, collaborative research projects - Develops & publishes long-range storage technology and applications roadmaps - Coordinates & obtains funding for university research funding for university research in storage technology

International HDD Technology Workshop The INFORMATION STORAGE INDUSTRY CONSORTIUM and the STORAGE RESEARCH CONSORTIUM jointly present INTERNATIONAL HARD DISK DRIVE TECHNOLOGY ROADMAP WORKSHOP 2003

International HDD Technology Workshop Organizations Participating: Attendees: 87 (77% from industry, 23% from academia) Organizations Represented: 39 Countries Represented: 8 Agere Systems Akita Inst. Tech. Carnegie Mellon U. Data Storage Institute Ehime U. Fujitsu Georgia Tech Gifu U. Headway Technologies Hitachi GST Hitachi CRL Hutchinson Technology INSIC Komag Marvell Maxtor MMC Technology Read-Rite/WD SAE Magnetics Seagate Shinshu U. Showa Denko Samsung SISA Samsung SAIT San Jose State U. Stanford U. STMicroelectronics TarnoTek TDK Tohoku U. Tokyo Inst. Tech. Toshiba Toyota Tech. Inst. (TTI) TTI - Chicago U. Arizona U. Manchester U.C. Berkeley U.C. San Diego Western Digital

International HDD Technology Workshop September 11-12, 2003 South San Francisco, CA

Areal Density (Gbits per square inch) International HDD Technology Workshop 1000 100 10 Longitudinal demos Perpendicular demos Products Historical demos 40 %/yr 1999 demos 190 %/yr 1 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2000-2 demos 40 %/yr INSIC EHDR goal Products 1998-2002 100 %/yr workshop Year

International HDD Technology Workshop Some Conclusions of the Workshop: Agreement that areal density growth was slowing, but little consensus on the pace going forward. Areal density growth estimates ranged from less than 25% per year to nearly 100% per year. Little consensus on how far conventional perpendicular magnetic recording could be extended. Estimates ranged from less than 400 Gb/in 2 to more than 1 Tb/in 2, although it was generally conceded that even 1 Tb/in 2 would be difficult. Little consensus on what would be needed next beyond PMR: bit-patterned media, track-patterned media, thermally-assisted recording/hamr, self-organized magnetic arrays (SOMA), or combinations of some or all of the above.

HDD Areal Density Trends Today s View Areal Density (Gbits per square inch) 1000 100 10 Longitudinal demos Perpendicular demos Products 1991-9 demos 40 %/yr 2000-2 demos 40 %/yr 1999 demos 190 %/yr Recent demos 31 %/yr Products 1998-2000 100 %/yr INSIC EHDR goal Recent products 31 %/yr 1 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 Year

International HDD Technology Workshop 1 10 4 1 10 3 100 10 2003 2005 2007 2009 2011 2013 2015 Areal Density (Gb/in^2); 40% Growth Areal Density Longitudinal Areal Density Growth; Perpendicular Roy Gustafson

International HDD Technology Workshop Year 2009 (demo) 2009 2013 (demo) 2013 Areal Density (Gb/in 2 ) 477 881 971 1416 KTPI 291 348 547 644 KFCI 1642 2530 1776 2199 BAR 5.7 7.3 3.3 3.4 W w (nm) 44 44 23 23 W r (nm) 26.3 43 14.0 21.4 D grain (nm) 4.7 3.5 3.3 3.7 δ (nm) 15 20 15 15 σ_grain_area 0.15 0.15 0.1 0.1 H k (koe) 21.7 21.2 23.5 23.5 σh k 0.05 0.05 0.01 0.01 θ (degrees) 5 5 1 1 HMS (nm) 5 20 2 20 M s (emu/cm 3 ) 655 650 800 800 Jitter (%) 12 15 15 15 Relative Read Sensitivity 4.9? 6.4 6.1 4.4 θ (degrees) 45 45 Roy Gustafson

International HDD Technology Workshop Head to Medium Spacing Write Width 20 150 16 125 fly_h ( t ) (nm) 12 8 Ww( t ) (nm) 100 75 50 4 25 0 0 3 6 9 12 t (years) 0 0 3 6 9 12 t (years) 1 10 3 Track Density 50 Erase Width 40 ktpi Weras( t ) nm 30 20 10 Roy Gustafson 100 0 3 6 9 12 t (years) 0 0 3 6 9 12 t (years)

International HDD Technology Workshop Areal Density vs. Media Average Grain Diameter 1 0.75 0.5 0.25 0 3.5 4 4.5 5 5.5 Dgrain (nm) Areal Density (Tb/in 2 ) Roy Gustafson

International HDD Technology Workshop 3 Data Rate vs. Time 2.5 Data Rate (Gbits/second) 2 1.5 1 Crt = 1 0.5 2003 2005 2007 2009 2011 2013 Perp 3.5" 7200RPM Long 3.5" 7200RPM Perp 2.0" 15K RPM Long 2.0" 15K RPM Roy Gustafson

International HDD Technology Workshop Conclusions of the Evolutionary Architecture Subgroup: Writing technology appears to be at the top of the list of issues. The response to our projected reader sensitivity was that the required reader sensitivity was likely to be achievable. Projected jitter specs appeared to fit within the Channel group s power and cost budgets. Tribology (magnetic spacing) represents a challenge, as did the required media grain size. Data rate limits appear to exist; a limit of 3 Gb/sec appears achievable primarily limited by read process noise issues, but reductions in disk diameter should keep products within that limit. The projections in this report suggest that perpendicular recording technology will enable a factor of at least 10 improvement in the next decade beyond the 60 Gb/in 2 of 2003. Significant technology challenges exist. It looks like an interesting decade ahead! Roy Gustafson

Areal Density (Gbits per square inch) HDD Areal Density Trends vs. Roadmap 1000 100 10 Longitudinal demos Perpendicular demos Products 1991-9 demos 40 %/yr 2000-2 demos 40 %/yr 1 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 1999 demos 190 %/yr Recent demos 31 %/yr Products 1998-2000 100 %/yr INSIC EHDR goal Recent products Roadmap 31 Projections %/yr Year

HDD Areal Density Trends Demos & Products Areal Density (Gbits per square inch) 1000 100 10 Longitudinal demos Perpendicular demos Products 1991-9 demos 40 %/yr 2000-2 demos 40 %/yr 1999 demos 190 %/yr Recent demos 31 %/yr Products 1998-2000 100 %/yr INSIC EHDR goal ~3 years Recent products 31 %/yr 1 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 Year

Near-Future Technology Progress Technologies now being introduced but not yet fully developed: Perpendicular Recording (PMR) addresses thermal stability limit, allows higher areal densities expect continued advances in media (e.g., ECC media) and write-heads Tunnel-junction MR heads (TMR) CPP GMR heads? addresses readback sensitivity: improved head SNR higher Mb/s expect continued improvement in R/R, resistance, geometry Thermal Flying-Height Control (TFC) addresses spacing control: improved resolution higher BPI expect fuller utilization with experience + new hardware & algorithms Secondary-Actuator (milli- or micro-actuator) addresses track-following: higher loop bandwidth higher TPI some server & desktop drives; expect use in smaller form-factors Longer Blocks & Iterative Detection (not yet implemented in HDDs) addresses linear-density & format efficiency higher capacity & reliability 4kB data-blocks, soft decision, iterative detector/decoder (e.g. LDPC) R. Wood & P. Frank, presented at APMRC2006

Continued Advances in Media exchange coupled composite (ECC) media gradient media K u = 0.082 10 6 J/m 3 M s = 0.6 T K u = 0.675 10 6 J/m 3 M s = 0.19 T K u increasing K u = 0 M s = 0.94 T K u = 1.0 10 6 J/m 3 R.H. Victora & X. Shen (U. of Minnesota), IEEE Trans. Mag. 41, 2828 (2005) D. Suess (U. of Vienna), Appl. Phys. Lett. 89, 113105 (2006)

Jian-Ping Wang, U. of Minnesota Exchange Coupled Composite Media ECC Disk (MINT, UMN) Con. PMR Disk (MINT, UMN) Ref. PMR (A Media Company) Head Type Spin-Stand Testing Lub (1.0 nm)/carbon (3.5 nm) /Si(1.5nm) /CoCrPt- SiO 2 (8nm) /Pt(1.0nm)/CoCrPt- SiO 2 (20nm)/Ru(20nm)/Ta(2. 5nm)/SUL Lub (1.0 nm)/carbon (3.5 nm) /Si(1.5nm) /CoCrPt- SiO 2 (20nm) /Ru(20nm)/Ta(2.5nm)/SUL N.A. Targeting for 200 Gbit/in 2 Single Pole Head (A Head Company) Samsung Information System America SNR_Total (db) LF-TAA (mv) 18 16 14 12 10-10 -20 Saturation Curve ECC Ref. PMR Conv. PMR 8 0 10 20 30 40 50 60 Write Current (ma) 30 ECC Con. PMR 20 Ref. PMR 10 0 Roll-Curve Normalized Singal (%) Amplitude (mv/µa) 110 100 1.0 0.8 0.6 0.4 0.2 0.0 90 80 70 60 50 LF-Time Decay Linear density: 52.8 kfci ECC Con. PMR Ref. PMR 10 100 1000 Time (Second) Track Profile 1.4 ECC 1.2 Con. PMR Ref. PMR 100 200 300 400 500 600 700 800 900 Linear Density (kfci) -20-16 -12-8 -4 0 4 8 12 Write Offset(µin)

NSIC White Paper Large Block Size Initiative From 1999

Large Block Size Initiative Hard Disk Drive Organization Announces a New Sector Length Standard Sunnyvale, Calif. --- Mar. 21, 2006--- IDEMA, the International Disk Drive, Equipment, and Materials Association, announced the results of an industry committee assembled to identify a new and longer sector standard for future magnetic hard disk drives (HDDs). This Committee recommended replacing the 30 year-standard of 512 bytes with sectors having ability to store 4096 bytes. The IDEMA Long Data Block Committee was composed of members representing the major hard drive developers, as well as electronics and software companies. The Microsoft Corp. participated in this Committee and plans to include a 4K-byte sector capability in their upcoming operating system named Windows Vista. The Committee foresees the first hard drive products becoming available later this year or in 2007, and is asking the computer industry to recognize this new standard and prepare for its availability. Backward compatibility with existing 512-byte products, both in hardware and in software, will be defined and accommodated during the phase over period. It is projected that most disk drives will be eventually formatted for 4K-byte sectors.

HDD Areal Density Trends Demos & Products Areal Density (Gbits per square inch) 1000 100 10 Longitudinal demos Perpendicular demos Products 1991-9 demos 40 %/yr 2000-2 demos 40 %/yr 1999 demos 190 %/yr Recent demos 31 %/yr Products 1998-2000 100 %/yr INSIC EHDR goal 5~7 years Recent products 31 %/yr 1 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 Year

Beyond Conventional PMR The future is probably ALL perpendicular. Although modeling (and consensus) currently suggests that conventional PMR will not likely get us beyond 0.6 ~ 1.0 Tb/in 2, the leading candidate successor technologies, Thermally- Assisted or Heat-Assisted Magnetic Recording (HAMR) and Bit Patterned Media (BPM) are also likely to rely on perpendicular recording, and could perhaps be best thought of as extensions, rather than replacements, of perpendicular recording.

The Superparamagnetic Effect Magnetic stability requires that K u V/kT > 40 ~ 60 where K u = magnetic anisotropy energy V = magnetic grain volume k = Boltzmann s constant T = absolute temperature

Beyond Conventional Perpendicular Recording (two favorite technology options to extend thermal limit) Patterned Media (increased V, utilizing 1 large grain per bit) Thermal Assist (increased K u, utilizing very high anisotropy media) 1 bit = 1 island HAMR Read GMR Sensor laser high anisotropy medium sensitive to temperature heat spot write coils deposition islands Challenges: Disk Manufacture Lithography/Stamping Challenges: Head Integration New Media Development plus all the engineering challenges of scaling dimensions for >Terabit/in> 2! R. Wood & P. Frank, presented at APMRC2006

One Potential Solution: Bit Patterned Media From this: To this: ~ 1,0 00 grains per bit 20 nm 100-1000 grains per bit 1 grain per bit B. Terris, Hitachi Global Storage Technologies and G. Hughes, Center for Magnetic Recording Research, University of California San Diego

C. Munce, Hitachi GST Presented at DISKCON USA 2005 Source: www.idema.org

Patterned Media Challenges 1. Development of a cost-effective media fabrication process that is also extendible to higher areal densities. The process needs to be capable of creating a precise pattern of uniform magnetic islands of 15 nm diameter or less, with precise placement, over an entire disk surface, preferably with circumferential ordering. 2. New media development, which requires the magnetic properties to be tailored such that the individual magnetic islands are well-coupled, so as to take advantage of the effectively increased volume in order to maintain thermal stability.

Patterned Media Challenges 3. Resolution of the bit aspect ratio (BAR) conundrum: from the point of view of media fabrication, a 1:1 aspect ratio is probably optimal, since it give the largest minimum dimensions, while from head fabrication and servo system capability points of view, a higher BAR (perhaps on the order of 3:1 or 4:1) would be highly preferable. For example, at 1 Tbit/in 2 BAR KBPI KTPI Track Pitch Bit Length (nm) (nm) 1:1 1,000 1,000 25 25 4:1 2,000 500 50 12.5 7:1 2,650 378 70 10 10:1 3,160 316 80 8

Patterned Media Challenges 4. Write synchronization issues: Since the medium is no longer continuous, data writing must be synchronized with physical bit positions. 5. Head-medium interface stability issues: Recent studies indicate that maintaining the necessary head-medium spacing (with the necessary degree of stability) will likely require planarization of the patterned media, necessitating development of yet another process which needs to be cost-effective in volume production.

Another Potential Solution: HAMR M. Kryder, Seagate

Thermally-Assisted Recording Challenges 1. Development of a head which integrates not only an advanced read transducer and a write transducer, but an additional thermal transducer which is capable of delivering an extremely small spot of heat (50 nm in diameter or less) to the exact point on the medium which is being written (and does so in extremely close proximity to the transducer supplying the magnetic write field). A cost-effective manufacturing approach to such an integrated threetransducer head is also required, and the approach needs to be scalable to even higher areal densities.

M. Kryder, Seagate Presented at DISKCON USA 2006 Source: www.idema.org

Multilayer Optical Waveguide Reported at ISOM 2006 Multilayer Optical Head that Generates a Subwavelength Spot with High Optical Efficiency F. Tawa, S. Hasegawa, and W. Odajima, Fujitsu Laboratories Ltd.

Achieving Nanoscale Optical Confinement Reference: W. Challener, Seagate, presented at Microoptics 2005, Tokyo

Achieving Nanoscale Optical Confinement E 2 intensity in recording medium for beaked triangle antenna. The antenna is 100 nm long, has a 30 o apex angel, and is 50 nm thick. The beak height is 20 nm and the base of the beak is 20 nm x 20 nm. E 2 intensity in recording medium for canted bow-tie antenna Reference: INSIC International Optical Data Storage Roadmap, August 2006 (See also: Challener, W., et al., Optical Transducers for Near Field Recording, Japanese Journal of Applied Physics Vol. 45, No. 8B (2006), pp. 6632-6642)

Achieving Nanoscale Optical Confinement Reported at ISOM/ODS 2005 Writing 40-nm marks using a beaked metallic plate near-field optical probe T. Matsumoto, Y. Anzai, T. Shintani, K. Nakamura, and T. Nishida Storage Technology Research Center, Research & Development Group, Hitachi, Ltd.

Achieving Nanoscale Optical Confinement Reported at ISOM/ODS 2006 Paper ThA3: Writing 40-nm marks using a beaked metallic plate near-field optical probe T. Matsumoto, Y. Anzai, T. Shintani, K. Nakamura, and T. Nishida Storage Technology Research Center, Research & Development Group, Hitachi, Ltd.

Thermally-Assisted Recording Challenges 2. New media development, which requires the development of a finely-grained, veryhigh-anisotropy recording medium with the requisite magnetic and thermal properties and distributions.

Thermally-Assisted Recording Challenges Thermal management within: 3. the medium, so as to avoid heating the neighboring bits (either on adjacent tracks or on track) sufficiently to either erase them or thermally destabilize them. 4. the head-disk interface, so as to avoid degradation of the lubricant and overcoat layers. 5. the head, so as to avoid physical or magnetic property deterioration of any of the three transducers.

Beyond Conventional PMR The bottom line is that one or both of these approaches (HAMR or Patterned Media) will likely be needed within 2 to 4 years at the technology demonstration level and within 5 to 7 years for products, even at the current rate of areal density progress, which appears to be roughly 30%/year. Otherwise, areal density progress will slow even more

Beyond Conventional PMR What comes after HAMR and/or BPM? Possibly a combination of HAMR and BPM Ultimately, reading/writing of individual small magnetic grains, as might be possible with Self-Organized Magnetic Array (SOMA) media, which might also be combined with either BPM or HAMR or both, and might need to be written and read with something more like a probe transducer.

Beyond Lithography: Self Organized Arrays From this: Grain Diameter = 9.9±2.6 nm θ 50 = 6.8 σ D =0.26 Dc= 10.5 nm To this: Grain Diameter= 4.5±0.15nm σ D =0.03 D C =4.5 nm random Normalized Frequency 0.18 0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0 Histogram Lognormal Fit 1 3 5 7 9 11 13 15 17 19 Grain Size (nm) 20 nm Oriented CoPtCr Films Population % 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0 3.911 4.168 4.426 4.684 4.942 Size (nm) FePt Particles Deposited from Solvent D. Weller, Seagate

Beyond Lithography: Self Organized Arrays Arrays created by differential etching of phase-separated block copolymers. Self assembly is combined with lithographically patterned templates to achieve ordered arrays with controlled spatial arrangement. C. Ross, MIT

Beyond Lithography: Self Organized Arrays Toward single particle 1 2 per bit recording! 130 nm 3 9 Tbpsi Thermal Stability Limit 3 nm >40 Tbit/in 2 1 Conventional Granular Media 2 Bit Patterned Media 3 Single-Grain-Per-Bit Patterned Media Development time: ~ 13 years @ 60% CAGR and ~22 years @ 30% CAGR D. Weller, Seagate

The Future of Hard Disk Drive Technology? Possible HDD Areal Density Progression Areal Density (Gbits per square inch) 100000 10000 1000 100 10 1 Longitudinal demos Perpendicular demos Products EHDR Target ECC- type Media? Patterned Media? HAMR? If HDD products continued at ~ ~30%/yr 28 Perpendicular Magnetic Recording 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020 2022 2024 2026 2028 2030 SOMA Media? Year

INSIC s EHDR Research Program in Extremely High Density Recording EHDR Extended Goal Stable Recording for HDDs at 1 Terabit/in 2 and beyond Primary Areas of Investigation Perpendicular Magnetic Recording (primary focus to date) Exchange Coupled Composite Magnetic Media Recording Bit Patterned Media (likely to become a second front going forward, probably aimed at ~ 2 Terabits/in 2 ) Participants Drive Companies: Hitachi GST, Samsung, Seagate, Western Digital Industry Suppliers: Hutchinson, Magnecomp, MIPOX International National Research Institutes: Data Storage Institute (Singapore), NIST Boulder Labs (US), NIST Gaithersburg Labs (US) Leading Research Universities: 18 (on 3 continents) currently

INSIC s EHDR Research Program in Extremely High Density Recording EHDR Extended Goal Stable Recording for HDDs at 1 Terabit/in 2 and beyond Primary Areas of Investigation Perpendicular Magnetic Recording (primary focus to date) Exchange Coupled Composite Magnetic Media Recording Bit Patterned Media (likely become a second front going - Bob forward, Scranton, probably aimed at ~ 2 Terabits/in 2 ) Participants Implementing BPM and/or TAR/HAMR will be incredibly technically challenging and will require people working together worldwide. in today s opening address. Drive Companies: Hitachi GST, Samsung, Seagate, Western Digital Industry Suppliers: Hutchinson, Magnecomp, MIPOX International National Research Institutes: Data Storage Institute (Singapore), NIST Boulder Labs (US), NIST Gaithersburg Labs (US) Leading Research Universities: 18 (on 3 continents) currently

Acknowledgements The materials presented here are derived from a variety of sources, which I gratefully acknowledge, including in particular: - The draft reports of the Joint INSIC-SRC International Hard Disk Drive Technology Roadmap Workshop, held September 11-12, 2003, including those prepared by: Roy Gustafson (then at Seagate) and Gordon Hughes (UCSD). - Presentations from DISKCON USA 2005 and DISKCON USA 2006, sourced from the IDEMA web site. - A presentation by my colleague, Barry Schechtman (Executive Director Emeritus of INSIC), made at the University of Arizona, October 6, 2005. - A paper entitled A Perspective on the Future of Hard Disk Drive Technology, prepared jointly with Roger Wood, Hitachi GST, and presented at APMRC2006, December 1, 2006.