Enabling Paper-Like Displays Roll-to-Roll Manufacturing of Display Backplanes. Hewlett-Packard Company, Palo Alto, CA. Phicot Inc, Ames, IA

Enabling Paper-Like Displays Roll-to-Roll Manufacturing of Display Backplanes Carl Taussig, Bob Cobene, Rich Elder, Warren Jackson, Mehrban Jam, Albert Jeans, Hao Luo, Ping Mei, Craig Perlov, Hewlett-Packard Company, Palo Alto, CA Frank Jeffrey, Marcia Almanza-Workman, Kelly Beacom, Steve Braymen, Bob Garcia, Jason Hauschildt, Han-Jun Kim, Ohseung Kwon, Don Larson, Phicot Inc, Ames, IA

Outline Introduction What are the key attributes of a paper-like display Why we need paper-like displays High Performance Reflective Color Self-Aligned Imprint Lithography (SAIL) Benefits and challenges of roll-to-roll (R2R) manufacturing SAIL basics green manufacturing for green products toolset for 1/3m wide line world s first R2R active matrix displays ZXO based TFTs for enhanced performance

What is a Paper-Like Display? Paper Paper-like display Conventional display viewability Reflective: good in bright light as well as indoors Reflective; like paper Typically emissive or transmissive, not good in bright light form factor Paper is light, flexible, and rugged Plastic is light, flexible, and rugged Glass is fragile and heavy cost $0.08/m 2 $300/m 2 $846/m 2 power no power 10W/m 2 400W/m 2 color Great color No/poor color Good color speed ~1S video

Driving the adoption of paper-like display Ubiquitous availability of information: emergence of the cloud Ultrathin client: radio+display Mobility Ruggedness Viewable in any light Light weight: long battery life Sustainability Minimize conventional print Green appliance manufacturing Energy efficient operation The Readius a cell phone with a roll out display product of struggling Philips spinout Polymer Vision

Sustainability 1 / 3 waste is paper of which 43% is print Newspapers 12 Packaging & other 47.7 Commercial print 7.35 Office 6.6 Standard mail 5.8 36m tons/yr Millions of tons/year US Paper & board disposal, 2005 US EPA Magazines 2.5 Books 1.15 Directories 0.65

Macro-Trends Are Helping to Drive Electronic Paper Mobile Internet Clean Technology Printed Electronics Digital Media *courtesy of Mike McCreary, E Ink

Electronic Publishing is a Multi-$B Addressable Market $100B- $300B annual publishing industry today, about the same as the whole current display industry* Mobile electronic books have not previously succeeded because they lacked the attributes of paper: low cost, outdoor readability, light weight (low power), &mechanical toughness. A library in your hands *courtesy of Mike McCreary, E Ink

Newspapers Need an Alternative to Paper Paper newspaper subscriptions are dropping sharply 12 hour delay in receiving news newest generations of people are on-line much more increased sensitivity to ecology issues Newspaper profitability is under pressure as a result increasing energy costs On-line subscriptions are growing but it takes 50-100 on-line subscribers to make up for one lost paper subscription* Digital distribution enables personalization Geographic localization enhanced Individually targeted content and advertising A paper-like reader appliance is needed Low cost, portable, daylight readable, mechanically tough *courtesy of Mike McCreary, E Ink

Paper Newspapers and the Environment OIL OIL OIL OIL OIL OIL OIL OIL OIL OIL OIL OIL 1/5 ton per subscription per year *courtesy of Mike McCreary, E Ink

Economic Reasons to Move From Paper Newspapers Printing $6.7M Building, G&A $27.6M Advertising $7.3M Newroom $9.9M Circulation $10.1M Ink and paper $10.4M Composite Newspaper Business Profile* 100,000 Circulation $83.9M Revenue $72.1M Total Cost ~10% Profit Newspapers could eliminate $27M (~38%) from its variable budget by moving away from printed newspapers But it will be critical to keep subscriptions and advertising rates high with electronic newspapers *Published by Bill Richards (former NY Times and Washington Post reporter) *courtesy of Mike McCreary, E Ink

Architectures to address full color gamut additive side-by-side subpixels stacked layers additive subtractive max.=33% R+G+B = grey (dim) 400 500 600 700 [nm] R+G+B = murky white 400 500 600 700 [nm] -R -G -B = brown (so add black)

Why is it hard? Uncontrolled ambient light in - direction, color, Perfectly diffuse light out 3 or 4 electrodes sets per pixel Millions of pixels All have to work Optical management -very low loss (materials, processes, architecture) x (optical, electrical, cost) = Systems approach is essential If 13 interfaces present with each 98% efficient then max roundtrip reflectivity ~60%

Use of color selective mirrors can increase reflectivity by 20% compared to basic stacked design Blue mirror Green mirror Red (All) mirror

Brightness /contrast progress ΔL* (white black) 95 90 85 80 75 70 65 60 55 50 45 40 35 30 25 20 15 Target region Newsprint/ SNAP 10 55 60 65 70 75 80 85 90 Low DR -ve, no inters High DR +ve, rough resist, optimised gaps High DR +ve, rough resist Low DR -ve, rough resist High DR -ve, rough resist High DR +ve, luminit High DR +ve, Luminit, static High DR +ve, rough resist, optimised gaps, narrow illumination E-ink measured + perfect RGBW filter model low DR -ve plastic, luminit low DR -ve plastic, luminit, narrow illumination L* white state

HP Labs Reflective Color Demonstrator

Color demonstrator video

Challenges & Benefits of R2R Electronics Fabrication Benefits Challenges Lower substrate cost Steady state processing: high-throughput, high-yield Lower cleanroom requirements Cheaper equipment - better scaling? Lower process temperature Defect repair patterning Limited equipment available no previous generation

Imprint Lithography is the Best Choice for R2R Patterning Photolithography Imprint lithography Inkjet Physical mask Laser ablation Throughput Moderate: limited by step & repeat / stitching High: > 5 meters/min Low Limited only by deposition Low Resolution Limited by substrate flatness ~10μ 100nm demonstrated >10μ 10μ -100μ ~10μ Alignment Limited by substrate flatness ~10μ Self alignment possible External sensor required poor ~10μ Issues Scaling to large areas costly New technology Materials must be jettable Cleaning, particles Thermal effects, selectivity

SAIL (Self-Aligned Imprint Lithography): Process flow is radically different from conventional flat-panel / integrated circuit fabrication deposition imprint etch Vacuum deposition of metals, dielectrics, & semiconductors Multiple mask levels imprinted as single 3D structure Patterning completed w/ wet & dry processes 5μ Conventional Photo-Lith SAIL deposit deposit imprint etch strip/clean spin resist etch develop align/expose etch mask

Phicot s parent PowerFilm Solar is the first and only company to manufacture a-si solar cells on plastic with a R2R process

Basic Imprint Lithography Process 1: coated substrate 2: coat with polymer 3: emboss 1μm Pixel speed depends linearly on mobility but inversely with the square of channel length t pixel μ ( V 2 L 2 G V T ) 4: cure with UV ~40nm lines on 50μ polyimide 5: release 6: etch 3 2 20 μm 4 levels in 0.5 μ step heights 1 0 Multilevel structures on flex at 5m/min

SAIL: Self-Aligned Imprint Lithography Photolithography Multiple masking and alignment steps required Different mask used to pattern each layer Process induced distortion of 200ppm results in 20μ misalignment over 10cm SAIL SAIL encodes multiple patterns and alignments into thickness modulations of a monolithic masking structure Single mask used to pattern all the layers multiple times No misalignment because mask distorts with substrate

SAIL backplane: patterning process flow Imprint Etch Disassemble Then Remove Next Etch top remove polymer exposed through mask the metal remaining polymer on array contact and a top stack TFT second one Finally remove gate all stack contact semiconductor polymer layer the time metal down way at consisting layer a and expose time the expose (optionally) to to define substrate of area channel expose the dielectric to expose gate Top TFT structure. gate completed semiconductor undercutting covering other layers channel metal dielectric layers covering Begin TFT backplane the channel by form bottom the lines that were isolated Contact removing metal crossovers gate lines layer the top thin metal by the undercut (optional) regions Semiconductor isolate the gate lines Dielectric and the TFTs Bottom metal Process produces complete backplane: TFT Pixel electrode Data line Crossover Gate line

SAIL solves alignment problem & saves money Multiple photoresist applications dominate photolithography process materials costs cost per ft 2 Backplane materials costs for R2R photolith & SAIL $18.00 $16.00 $14.00 $12.00 $10.00 $8.00 $6.00 $4.00 $2.00 $0.00 Photolithography Cost of Patterning SAIL R2R SAIL R2R photolith (AGI) Web cost Strip-off 2P RIE etch oxide Under-cut Al (1-3 um) RIE etch n+ Wet etch Cr Thin down 2P (clear gate-pad) RIE etch n+&si&sin Pre-Cr-etch Cleaning Thin down 2P (clear gate-pad) Plasma etch Al RIE etch oxide RIE etch n+&si&sin Wet etch Cr Imprint SAIL structure SD metal deposition (Cr) PECVD oxide/nitride/si/n+ deposition Gate metal deposition (Al) Condition web (de-hydro) Web cost Align and Expose Sputter Dep Interconnect Align and Expose Sputter Dep/ ITO Align and Expose Ultrasonic Clean Si RIE & Resist Strip Align and Expose SiN, a-si, N+ dep Align and Expose Sputter Gate 1 Metal Web preparation $0.0 $0.5 $1.0 $1.5 $2.0 $2.5 $0.0 $0.5 $1.0 $1.5 $2.0 $2.5 $3.0

Green manufacturing for a green product Less is better: 50μm thick plastic vs. 0.7mm thick glass Less process materials: removal of photolith reduces process consumables Energy costs: transients involved in batch consume energy; steady state is more efficient Reduced clean room requirements: Smaller equipment footprint

The web rolled on the core is its own clean room HEPA filter Ambient Process Vacuum Process

Equipment footprint comparison between R2R and flat panel 330mm imprint system Gen10 cluster tool

Patterning scaling: R2R imprinter compared to panel stepper comparison made at equal throughput equipment cost scaling comparison: panel stepper vs R2R imprinter 10 0 equipment cost [M$] / throughput [cm 2 / S] 10-1 10-2 100 mm R2R imprinter 330 mm R2R imprinter 10-3 2 3 4 5 6 7 8 9 10 generation Scaling similar for R2R and panel; cost much lower for R2R

PECVD Scaling: R2R photovoltaic compared to panel comparison made at equal throughput equipment cost [M$] / throughput [cm 2 / S] 10 0 10-1 equipment cost scaling comparison: panel CVD vs R2R CVD 330 mm R2R PECVD 1 m R2R PECVD 10-2 1 2 3 4 5 6 7 8 9 10 generation Again; scaling similar for R2R and panel; cost much lower for R2R

13 production solar cell deposition R2R Tool Development 13 drum PECVD 13 drum RIE 13 wet etcher 4 imprinter 13 drum sputter 13 RIE 10 drum PECVD 13 imprinter 2005 2006 2007 2008 2009 2010

Web based tools are a lot simpler to build than big panel tools

Performance of Full-SAIL a-si TFTs Full SAIL TFTs with thinner dielectrics have greatly improved performance on-off ratio > 10 7 100μA on-current mobility from linear portion of transfer curve as high as 0.8 cm 2 /V/S near linear scaling of I on vs 1/L to L~2μm 1.E 1 1.E-05 W=100um Vsd=10.1V Mobility [cm 2 /V/S] 1.E 0 1.E-1 1.E-2 1.E-3 1.E 0 q1 =11.3663 p1 =0.83471 R2=0.96302 1.E 1 1.E 2 Channel Length [μm] Isd/(W/L) (A) 1.E-06 1.E-07 1.E-08 1.E-09 1.E-10 1.E-11 1.E-12 1.E-13 W [μm] L [μm] 100.0 1.0 100.0 2.0 100.0 5.0 100.0 10.0 100.0 20.0 100.0 50.0 100.0 100.0 1.E-14-10 0 10 20 30 40 Vg(V)

Initial display demonstrators SAIL Backplane on flexible substrate World s first active matrix display made exclusively with R2R processes (including E Ink Front Plane)

Process flow Yield Improvements: Pareto process at work Process Step Thin film deposition Imprinting Etching Device test Shunt defects Stress control Particle generation Master defects Stamp defects Imprint process Endpoint control Process design Contact liquefaction Probing errors/damage Severity 1/mm 2 unstable low low med low med med low low A bubble defect, voids are formed by insufficient volume of photopolymer to fill mold pinhole defect in metal caused by etchant diffusing through pinhole in oxide A crack defect typically results from imbalanced deposition stress Tenting defect formed by particle between stamp and substrate at imprinting time or by void in stamp Nonuniform imprinting results in premature mask erosion and feature loss Bridging caused by breakage of imprint stamp in narrow (~2u) regions

SAIL ZXO TFTs: Motivation, the 3P s Process simplification Elimination of contact layer removes 2 etch steps and one deposition step Huge increase in process margin for critical channel definition step Performance increase Even with direct metal contacts mobility is ~10X a-si at the same process temperature Higher mobility important for emissive pixels and edge electronics Pelucidity (transparency) Enable see thru displays Increase aspect ratio for conventional displays

SAIL ZXO TFTs: Process Simplification ZXO stack has 1 less layer then a-si Superior performance to a-si with no contact layer between top metal and semiconductor: 1 less deposition step 2 less etch steps: since all depositions performed before any etch steps some layers are etched multiple times Back-channel etch much easier to control with ZXO ZXO provides excellent etch stop for top metal etch whereas there is no etch selectivity between n+ μc-si and intrinsic a-si Embossed mask 100nm S/D metal Al 40nm N+ uc-si 30nm I-a-Si 200nm SiNx / 100nm SiOxGate dielectric 100nmGate (Al) substrate a-si stack Embossed mask W S/D metal ZXO Thermal oxide gate dielectric Substrate 6 p+ doped Si wafer Gate (Al) ZXO stack

SAIL ZTO TFTs: device measurements mobility for 100u long channels vs annealing temperature Transfer for W=50u, L=50u, Ta=300C 1.E-04 25 1.E-05 1.E-06 sat incr mobility (cm^2/v/s) 20 15 10 5 Ids, Ig (A) 1.E-07 1.E-08 1.E-09 1.E-10 1.E-11 Id for Vds=10V Ig 1.E-12 0 200 250 300 350 400 temperature (C) 1.E-13-10 0 10 20 30 Vg (V) Mobility strongly dependent on annealing temperature Performance unrealistically high due to high quality thermal oxide gate dielectric

Next steps towards commercialization September 22, 2008 PowerFilm announced that it has taken a license to the SAIL technology October 6 th, 2008 PowerFilm announced it has won a $1.4M / year cooperative agreement from the U.S. Army for development of a 'self powered flexible display'. HP Labs and PowerFilm will collaborate on the contract. PowerFilm Solar has created Phicot as a subsidiary to commercialize the technology

Acknowledgements The authors gratefully acknowledge the support of their collaborators and sponsors -- FlexTech Alliance - contract RFP04-112F ARL contract W911NF-08-2-0063 -- E Ink Corporation ASU Flexible Display Center Dupont Teijin