Prog. Polym. Sci. 33 (2008) 581–630 Polymers for flexible displays: From material selection to device applications Myeon-Cheon Choi a , Youngkyoo Kim b , Chang-Sik Ha a, a Department of Polymer Science and Engineering, Pusan National University, Busan 609-735, Korea b Department of Chemical Engineering, Kyungpook National University, Daegu 702-701, Korea Received 22 January 2007; received in revised form 8 November 2007; accepted 8 November 2007 Available online 4 February 2008 Abstract With digitalization, plenty of information is being exchanged through electronic media, and consumers are demanding high quality, convenient, and portable digital devices. Currently, flat panel displays, such as liquid crystal displays (LCDs) and plasma display panels (PDPs), satisfy them with regard to quality. Convenience and portability will be realized with flexible displays in the future. Polymers are very promising materials for flexible displays with many advantageous charateristics including transparency, light weight, flexibility, and robustness. They are also some of the least expensive materials and are suitable for mass production via roll-to-roll processes. In this review, we will discuss the kinds of polymers that are used, where and how polymer materials are used and the challenges to overcome in developing flexible displays. r 2008 Elsevier Ltd. All rights reserved. Keywords: Electro-optic material; Encapsulation; Flexible display; Flexible substrate; Transparent electrode Contents 1. Introduction ..................................................................... 582 2. Polymer substrates ................................................................. 584 2.1. Potential polymer candidates for flexible substrates ..................................... 584 2.2. Property requirements that apply to flexible substrates ................................... 585 2.2.1. Clarity ................................................................ 586 2.2.2. Thermal stability ........................................................ 586 2.2.3. Surface properties ........................................................ 588 2.2.4. Chemical resistance ....................................................... 588 2.2.5. Mechanical properties ..................................................... 588 3. Barrier coatings ................................................................... 589 3.1. Mechanisms of device failure ..................................................... 589 3.2. Theories of gas permeation ...................................................... 590 ARTICLE IN PRESS www.elsevier.com/locate/ppolysci 0079-6700/$ - see front matter r 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.progpolymsci.2007.11.004 Corresponding author. Tel.: +82 51 510 2407; fax: +82 51 514 4331. E-mail address: [email protected] (C.-S. Ha).
Transcript
ARTICLE IN PRESS
0079-6700/$ - se
doi:10.1016/j.pr
�CorrespondE-mail addr
Prog. Polym. Sci. 33 (2008) 581–630
www.elsevier.com/locate/ppolysci
Polymers for flexible displays: From material selectionto device applications
aDepartment of Polymer Science and Engineering, Pusan National University, Busan 609-735, KoreabDepartment of Chemical Engineering, Kyungpook National University, Daegu 702-701, Korea
Received 22 January 2007; received in revised form 8 November 2007; accepted 8 November 2007
Available online 4 February 2008
Abstract
With digitalization, plenty of information is being exchanged through electronic media, and consumers are demanding
high quality, convenient, and portable digital devices. Currently, flat panel displays, such as liquid crystal displays (LCDs)
and plasma display panels (PDPs), satisfy them with regard to quality. Convenience and portability will be realized with
flexible displays in the future. Polymers are very promising materials for flexible displays with many advantageous
charateristics including transparency, light weight, flexibility, and robustness. They are also some of the least expensive
materials and are suitable for mass production via roll-to-roll processes. In this review, we will discuss the kinds of
polymers that are used, where and how polymer materials are used and the challenges to overcome in developing flexible
The topic of flexible displays has prompted manypopular news stories. How do we define a flexibledisplay? Flexible displays can be classified accordingto degree of flexibility: flat displays are made ofplastic or another non-glass backplane, but only forthe benefit of lightness or ruggedness; formed
displays are bent once, such as a curved automobiledashboard, but do not flex further; flexible displaysmay be bent or flexed during use, but not over arange that includes folding or rolling; rollable
displays are as flexible as fabric [1]. Recently,the literature on flexible displays has been expand-ing. It now includes a book on flexible flat paneldisplays written by Crawford [2] and a specialedition of the Proceedings of the IEEE on flexibledisplays [3].
The prospects for flexible displays are promising,although the timing still depends on technical andmanufacturing developments [4,5]. Electrophoreticdisplays such as electronic papers using plasticsubstrates, which have a relatively simple structure,are just beginning to be produced in quantities
approaching high volume. Displays that are in-tended to flex or roll during use may reach themarket in several years, pending further develop-ments in backplane and fabrication processes. Thenear-term revenue in dynamic signage and mobilephones will lead to the development of larger andmore sophisticated displays with flexibility androllability. Fig. 1, which has been adapted fromthe data of the iSuppli Flexible Display Report,shows the market prospects for flexible displaysfrom 2007 to 2013 [1].
Polymers are very promising materials for flexibledisplays with many advantages. They are transpar-ent, light in weight, flexible, and robust. Polymersare a good alternative to the glass substrates thathave been actively used for flat panel displays suchas liquid crystal displays (LCDs) and plasmadischarge panels (PDPs). Glass is so rigid that it isvery difficult to use in a flexible display. Polymershave mechanical properties that vary from strongrigidity, such as in engineering plastics, to softness,such as in rubber or polyethylene films. They aresome of the least expensive materials and aresuitable for mass production via roll-to-roll (RTR)
processes. Therefore, polymers are being consideredas the key materials for flexible displays in variousapplication areas including transparent substrates,electrodes, active materials for organic light-
emitting devices (OLEDs), LCDs and organicthin-film transistors (OTFTs), dielectric materials,and coating materials. All polymer-based flexibledisplays are being investigated.
In this review, we will discuss what, where, andhow polymer materials are used and the challengesto overcome in the flexible display field.
2. Polymer substrates
In flexible displays, the flexibility depends on thesubstrate. Three kinds of substrates are consideredto be flexible: thin glass, metal foil, and plastic. Thinglass films are bendable and have the highlydesirable qualities of glass [6]. However, they arebrittle. This property limits their application asflexible substrates. Metal foils can also handle high-process temperatures and provide a good barrier tomoisture and oxygen, without the problems ofbreakability [7]. However, metal only works fornon-transmissive displays and cannot handle multi-ple bends. In addition, it is an expensive material touse in large displays. Therefore, metal foils are goodcandidates for small and early applications offlexible displays. Plastic is the key material ofchoice, as it allows reasonable tradeoffs in mechan-ical, optical, and chemical performance. It is aninexpensive and useful material for in-line produc-tion via RTR processes. Multilayer-engineeredsubstrates will be required for most practicalapplications.
2.1. Potential polymer candidates for flexible
substrates
There are several polymers that are candidates forflexible substrates as shown in Fig. 2, which listssubstrates in terms of glass transition temperature
(Tg) [2,8]. These polymers are divided into threetypes, including crystalline, amorphous, and solu-tion-castable amorphous.
Thermoplastic semi-crystalline polymers avail-able for flexible displays include polyethyleneterephthalate (PET), polyethylene naphthalate(PEN), and polyetheretherketone (PEEK) [9,10].PEEK, whose Tg and Tm are �140 and 340 1C,respectively, is known be at the upper limit of semi-crystalline polymers that can be melt-processedbecause polymers with a Tg of higher than 140 1Cwill be degraded significantly during a melt process.Heat-stabilized semi-crystalline polymers providegood dimensional stability above their Tg, whichexpands their upper operating temperature. Thesecond group, amorphous polymers, includes poly-carbonate (PC) and polyethersulphone (PES)[11,12]. These are non-crystalline thermoplasticsthat can be melt-extruded or solvent casted. Thelast group is the amorphous polymers that cannotbe melt-processed, such as modified PC, polyarylate(PAR), polycyclic olefin (PCO) or polynorbonene(PNB), and polyimide (PI) [13–15]. Fabric materials[16], ultra-thin polymer films [17], and glass-reinforced plastic [18] have also been used assubstrates. Fig. 3 summarizes the Tg’s of thecandidate polymers for flexible substrates.
Table 1 shows a comparison of the characteristicsof polymer materials suitable for flexible substrates.Polyesters such as PET and PEN have advantageswith regard to clarity, coefficient of thermal expan-sion (CTE), chemical resistance, moisture absorp-tion, and price while their upper operatingtemperature and surface roughness are not so good.
ARTICLE IN PRESS
Fig. 2. Potential candidate polymers for flexible display sustrates.
PES has good clarity and a high upper operatingtemperature but has poor solvent resistance andis expensive. On the other hand, PI has high thermalstability and good mechanical and chemicalproperties but is orange-colored and expensive.Recently, however, many technical advances havebeen aimed at producing colorless PI by incorpor-ating fluorine, sulfone, or non-aromatic groups[19–25].
2.2. Property requirements that apply to flexible
substrates
Currently, OLEDs, which require the moststringent conditions of the flexible display devices,are based on glass substrates because they areperfectly satisfactory for flat panel applications. Toreplace glass, a plastic substrate needs to mimic theproperties of glass, including its clarity, dimensional
ARTICLE IN PRESS
Fig. 3. Glass transition temperature (Tg) of commercially
available polymers.
Table 1
Comparison of polymer materials for flexible substrates
PET PEN PC PAR PES PI
Optical clarity J J J J �
Upper operating temp. n J n J
Dimensional stability J J n n n J
Surface roughness � � J J J J
Solvent resistance J J � � � J
Moisture absorption J J n n � �
Young’s modulus J J n n n n
: Excellent, J: good, n: fair, � : poor.
Table 2
Minimum property requirements of a polymer substrate for flexible dis
Property
Polymer substrates Total light transmittance over 400
Haze (%)
Upper operating temperature (1C)
Coefficient of thermal expansion (
Average surface roughness (nm)
Chemical resistance
Barrier coated substrates Water vapor transmission rate (g/
stability, thermal stability, barrier properties, sol-vent resistance, low CTE, and smooth surface, aswell as possess good optical properties [2,26]. Theminimum property requirements of a polymersubstrate to be applied in flexible displays areshown in Table 2, which summarizes variousreported literature values up to the early 2007.Substrate materials affect the following processes:barrier coating, electrode deposition, patterning,and thin-film transistor (TFT) fabrication. There-fore, a circumspect approach to selecting the propersubstrate materials is required.
2.2.1. Clarity
Clarity is the most important property for bottomemissive displays, and a total light transmissionof 485% over 400–800nm coupled with a haze ofless than 0.7% are typical of what is required [2].Table 3 shows the properties of commercial basefilms [18]. PC and cyclic olefin copolymers (COC)have a high optical transparency. Common PIs arenot generally good enough for transparent devices,but transparent flexible PI substrates whose averageoptical transmission is about 86% in the visible lightrange at 250mm thickness have been reported.This fluorine-containing aromatic PI is knownto reduce intra- or intermolecular charge transfercomplex formation and to possess colorless transpar-ency [19].
2.2.2. Thermal stability
Thermal stability is another important issuefor polymer substrates. Polymer substrates are
plays
Requirement
–800nm (%) 485
o0.7
4150
ppm/1C) o20
o5
Resistance to acid, alkali, and solvent
m2/day/ OLED o10�6
LCD o10�3
EPD o10�2
TFT o10�3
o20
480
Ability to bend over a 1 in diameter 1000 times
ARTICLE IN PRESS
1.5
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
-0.2
-0.4
-0.6
-0.8
-1.00 20 40 60 80 100 120 140 160 180 200 225
Temperature (°C)
Hei
ght %
(%
)
Multi-ramp heating and coolingheat stabilised PEN 125 micron
Multi-ramp heating and cooling
PET Film
Fig. 4. Thermal mechanical analysis of PET and heat-stabilized PEN film [33]. Reproduced from MacDonald, Rollins, Eveson, Rustin,
and Handa by permission of Society for Information Display, California, USA.
Table 3
Properties of base film for polymer substrates (reproduced from Ito, Oka, Goto, and Umeda by permission of Japanese Journal of Applied
Physics, Tokyo, Japan)
PET PEN PC COC PES PI
Thickness (mm) 0.1 0.1 0.1 0.1 0.1 0.1
Total light transmittance (%) 90.4 87.0 92.0 94.5 89.0 30–60
Retardation (nm) Large Large 20 7 o10 Large
Reflactive index 1.66 1.75 1.56 1.51 1.6 –
Glass transition temperature (1C) 80 150 145 164 223 300oCoefficient of thermal expansion (ppm/1C) 33 20 75 70 54 8–20
Water absorption ratio (%) 0.5 0.4 0.2 o0.2 1.4 2.0–3.0
exposed to high temperatures during manufacturingprocesses such as barrier coating [27], electrodedeposition [28], patterning [29], and silicon-basedTFT fabrication [13,30]. In particular, TFTs arevery sensitive to dimensional changes due tothermal or mechanical stress [31].
There are two factors that have to be consideredwhen selecting polymer materials with the properthermal stability: CTE and Tg, where the polymerchains start to move to relax the stress that is storedduring the processes. In general, inorganic or metallayers used for barrier or electrode layers have muchlower CTEs than polymer substrates, and thedimensions of the polymer change significantly at
Tg [32]. Polymers with high Tgs such as PI and PESare better choices in this regard. For example, Limet al. used aliphatic cyclic PI [20] and fluorinated PI[19] as flexible substrates. A low CTE is alsoadvantageous for making dimensionally stable de-signs for devices [15]. A mismatch of CTE betweenlayers gives rise to strain and cracking underthermal cycling during device fabrication. High-temperature processes, such as the low-temperaturepoly-silicon (LTPS) process, require plastic sub-strates that can withstand very high temperatureswith CTEs of around 10–20 ppm/K or less to enablefilm deposition and device annealing at tempera-tures of up to 350 1C.
ARTICLE IN PRESS
Fig. 5. Cross-sectional structure of flexible displays.
Polyesters such as PET [33] and PEN [34] haveactively been researched as flexible substrates,although their Tgs of 78 and 120 1C, respectively,are too low to apply to OLED. MacDonald et al.[35] reported heat stabilization processes to impartthe thermal and dimensional stability necessary tosupport OLED manufacture. Fig. 4 gives a thermalmechanical analysis of 125 mm PEN and heat-stabilized PEN films. Heat-stabilized films exhibitminimal shrinkage, of the order of o0.1% andtypically o0.05%, when exposed to temperatures ofup to 180 1C for 5min. In addition, once heatstabilized, the Tg effects are essentially negated andthe PEN film remains a dimensionally reproduciblesubstrate up to 200 1C. This is within the perfor-mance requirements of a flexible substrate for anOLED display.
Ito et al. developed a new idea to employ flexiblefibrous glass-reinforced plastic (FRP) substrates forflexible displays to offer a low CTE of 14 ppm/1Cand a high transparency of about 89% at 400 nm[18]. New polymer substrates with a high thermalstability, named OPSs [36] and AryLiteTM [37],have also been reported.
2.2.3. Surface properties
Surface qualities such as roughness and cleanli-ness are essential to ensure the integrity ofsubsequent layers including barrier and conductivelayers. Semi-crystalline polymer films do not possessgood surface properties compared to amorphouspolymer films. Surface defects that remain in thesubstrates are detrimental to the performance of theactive layer of OLEDs, whose thickness is around100 nm, while picks greater than 50 nm in height areobserved on common optical grade PEN films. Theywill create defects like pinholes on the thin films ofbarriers and electrodes, forming dark spots inOLEDs [38]. The defects will also lead to crackswhen displays are bent. To provide a defect-freesurface, scratch resistant or planarizing layers arecoated to smooth over all the underlying substratesurface defects [39]. The coated film, called a surfaceengineered film, ensures good integrity for subse-quent barrier layers and conductive coatings.
2.2.4. Chemical resistance
Polymer substrates have to be exposed to a widerange of solvents and chemicals as well as moistureduring manufacturing processes, including cleaning,coating, and patterning processes. In general, semi-crystalline polymers such as PET and PEN have
strong resistance to solvents, while amorphouspolymers have poor solvent resistance. A typicallist of the materials that the substrate must becompatible with includes methanol, isopropanol,acetone, tetrahydrofuran, n-methylpyrrolidone,ethyl acetate, sulfuric acid, glacial acetic acid,hydrogen peroxide, and sodium hydroxide. Coatedpolymer substrates with a variety of organic orinorganic layers or a hard coating layer are used toprevent the invasion of solvents and moisture [2].
2.2.5. Mechanical properties
For the RTR process as well as for displayperformance, mechanical properties have to be care-fully considered [40]. Flexible OLEDs (FOLEDs),which demand stringent requirements among thedisplay technologies, usually contain a polymersubstrate, organic–inorganic multi-barrier-layer,brittle transparent inorganic anode, light-emittinglayer, TFT layer, metal cathode, and encapsulatinglayers as shown in Fig. 5 [41]. When the devices arebent, the mechanical discrepancy at the interfacebetween the organic and inorganic materials gen-erates mechanical failure in devices. For thosedevices inorganic thin films may be the source ofthe failure due to their brittle properties [42]. In thissense, all-organic-based devices are the best choicefor rollable displays. However, certain devices suchas FOLEDs require inorganic films, and it isimportant to understand and improve the mechan-ical limits of these materials [26].
The failure of devices depends on the arrange-ment, thickness, and properties of all the layers.Adhesion at the interfaces between the differentlayers under thermal cycling and environmentaltesting, wetting characteristics, and the ability towithstand flex testing are critical to determiningtheir robustness in use [35]. Flexible displays have
ARTICLE IN PRESSM.-C. Choi et al. / Prog. Polym. Sci. 33 (2008) 581–630 589
the advantage of mass production through a RTRprocess. Therefore, it is essential to understandthe mechanical properties required of flexiblesubstrates.
3. Barrier coatings
Inorganic transparent oxide films (e.g. silicon andaluminum oxide) on polymer films have been widelyused as gas barrier materials for food and medicalpackaging. They provide at best only two to threeorders of magnitude improvement over the oxygentransmission rates (OTR) of polymer substrates,whether deposited by plasma-enhanced chemicalvapor deposition (PECVD), sputtering, or evapora-tion [42].
However, electronic devices, especially OLEDs,demand more stringent barrier confinement. For anOLED lifetime of 410,000 h, the requirementsinclude water vapor transmission rates (WVTR) of10�6 g/m2/day and OTR of 10�5mL/m2/day [26,43].Fig. 6 shows the relative barrier properties andrequirements for electro-optic devices.
3.1. Mechanisms of device failure
Flexible displays, especially OLEDs, are extre-mely sensitive to water vapor and oxygen, which
Fig. 6. Requirement of WVTR and OTR for electronic devices.
bring about their degradation. There have beenmany mechanisms proposed for the degradation ofOLEDs including cathode oxidation, detachment ofthe organic layer from the anode, diffusion of theemitting layer material into the hole transport layer,detachment of the organic layer from the cathode,electrical shorts, electrochemical reactions at theelectrodes, oxygen-activated photochemical da-mage, and oxidation of polymers by oxygenoriginating from ITO.
Dark spots resulting from exposure to ambientconditions were found to be due to cathodedelamination by Liew et al. [44]. When the darkspots on the cathode were peeled off and thecathode was newly deposited, the OLEDs gaveuniform emission, indicating that the origin of thedark spots was cathode delamination. They sug-gested that the nucleation of the dark spots takesplace at the organic/cathode interface during thedeposition of the cathode.
Nuesch et al. investigated the influence of oxygenand water [45]. The electrochemical reduction ofwater at the cathode/organic interface leads tohydrogen evolution, creating bubbles below thecathode surface. Dark spots grown in a pure oxygenatmosphere are different from those grown in awater atmosphere, showing perfectly circular beha-vior and obeying linear growth kinetics for devices.
Kim et al. [46] looked at the black spots in anelectroluminescent polyfluorene system with poly(styrenesulfonate)-doped poly(3,4-ethylenedioxy-thiophene) (PEDOT/PSS) on ITO as the anodeand Ca/Al as the cathode. The non-emissive diskssurrounding the pinhole defects are characterized bya localized electrochemical reaction with reductionof the normal doped PEDOT/PSS to the dedopedmaterial and oxidation of the active metal.
Wang et al. used micrometer-sized silica particlesto create uniformly sized pinholes on the protectivecover so that oxygen and moisture would give riseto artificial dark spots [47]. They monitored in situthe linear growth of the dark spots with respect toparticle diameter as well as time and suggested thatdust contamination may be the major cause of darkspot formation, showing the distribution of the darkspots is Gaussian.
Chua et al. showed that, due to the roughening ofthe polymer/electrode interface caused by metalmigration, the close proximity of metal protrusionsleads to an increase in local current that degradesthe polymer [48]. Subsequent electrochemical andphotochemical reactions result in the formation of
ARTICLE IN PRESSM.-C. Choi et al. / Prog. Polym. Sci. 33 (2008) 581–630590
volatile species and a large concentration ofcarbonyl groups that quench luminescence fromthe polymer.
The degradation problem can be dealt with bysealing the devices in an inert atmosphere. There-fore, encapsulation of the devices and isolation ofthe active materials from the atmosphere are veryimportant to prolong the lifetime of flexible devices.
3.2. Theories of gas permeation
Inorganic thin films without defects possess goodbarrier properties applicable to OLED devices. Forexample, thin glass foils of around 50 mm thicknessperfectly satisfy this condition but they are toobrittle to apply to flexible displays. Currently, themost effective way to block gas intrusion whileretaining reliable mechanical properties is to usemultilayer gas barrier films with an alternatingorganic and inorganic thin-film structure [39].
However, there exist several challenges to under-standing the mechanisms in multilayer, thin-filmbarrier coatings on polymeric substrates [49]. First,direct imaging and characterization of the sizes andspatial density of defects that are 100 nm in diameterare difficult. Second, it is impractical to measurematerial properties such as effective diffusivity andsolubility for each layer in a multilayer system insitu. Third, a simple analysis of single oxide layersdeposited on PET may not accurately represent theindividual layers existing within a multilayer stack.
Fig. 7. Calculated lag-time and steady-state flux for defect spacings o
Burrows by permission of American Institute of Physics, Maryland, U
Roberts et al. [50] proposed a model for gaspermeation where gas and water vapor trans-port through gas barrier films is comprised ofcontributions from three components: un-hinderedtransport through ‘macro-defects’ (41 nm) in theoxide layer, hindered transport through ‘nano-defects’ (0.3–1.0 nm), and hindered transportthrough the amorphous lattice of the oxide (inter-stice o0.3 nm). The presence of nano-defectsindicates that the oxide layer is more similar to anano-porous solid such as zeolite than silica glasswith respect to permeation properties as it demon-strates greater permeability and lower activationenergy of permeation than the values expected forpolymers coated with glass.
Peukert et al. reported the influence of defects andmorphology on barrier properties for the case ofvacuum web coating of inorganic layers on poly-meric films [51]. Simulations performed via anumerical model revealed a complex interactionmechanism of film thickness, defect area, and defectspacing affecting the permeability. They showedthat the critical thickness that limits the perme-ability decreases by increasing the substrate filmthickness. When analyzing defect structures with thesame total defect area, the total transmission rate ofsmall defects occurring with high frequency exceedsthat for large defects with low frequency, indicatingthe importance of evaluating small defects.
Graff et al. [49] suggested a mechanism ofvapor permeation through multilayer barrier films.
f 100 and 1000mm [49]. Reproduced from Graff, Williford, and
SA.
ARTICLE IN PRESSM.-C. Choi et al. / Prog. Polym. Sci. 33 (2008) 581–630 591
They used both transient and steady-state vaporpermeation measurements combined with classicalFickian diffusion models to determine the mechan-ism of vapor permeation through multi-barrierstructures. They showed that the results obtainedare limited not by equilibrium diffusion but by lag-time effects caused by the extremely long effectivepath length for the diffusing gas, as shown in Fig. 7.It takes a long time to reach a steady state and, oncethe lag time is exceeded, the water vapor flux willretain a steady-state value that is substantiallyhigher, 10–3–10–4 g/m2/day, than the WVTR calcu-lated during the transient period. There are severalways to lower the diffusivity and solubility of thepolymer inter-layers to improve barrier perfor-mance such as inducing hydrophobic moieties ororganic/inorganic copolymers, physical modifica-tions through ion bombardment or cross-linking, orchemical modification through reactive etch orplasma surface treatment. However, the range ofimprovement may be small relative to improve-ments in the inorganic layer since the effectivediffusivity of the inorganic layer is at least fourorders of magnitude lower than that of the polymerinter-layers.
Henry et al. proposed a model for the mechanismof water vapor permeation through PET/AlOxNy
gas barrier films [52], identifying that water vaporpermeates predominantly through tortuous, nanos-cale permeation pathways in the AlOxNy structure.The nature of this permeation process is the result ofthe strong chemical interaction of the watermolecules with the AlOxNy pore walls, possibly atN-rich sites. The key requirements to give an ultrabarrier condition for optoelectronic devices are toincrease the density of the coatings as much aspossible to reduce the size and number density oftortuous pathways while maintaining an optimumAl/N ratio.
3.3. Barrier coatings on polymer substrates
There have been three research directions in thefield of transparent barrier coatings for flexiblesubstrates. The first is focused on the kinds ofinorganic materials that will be used. They includesilicon oxide [53–56], silicon nitride [57,58], siliconeoxynitride [59], aluminum oxide [49], and mixedoxide [60–62]. A perfect layer of bulk oxides, such asSiO2 which makes up glass substrates, of only a fewnanometers thickness can reduce the diffusion ofwater and oxygen to acceptable levels.
The second direction is the structure and thenumber of barrier layers, which are often calleddyads, that are alternatively layered with organicand inorganic materials. A typical single barrierlayer provides only two to three orders of magni-tude improvement in the barrier properties com-pared to a bare polymer substrate [39]. To satisfythe requirements of OLEDs, polymer substratesmust be coated with multilayer organic–inorganicthin films.
The third is the coating method of thin films.Generally, polymer materials that are used asflexible substrates require low-temperature pro-cesses, where the quality of the inorganic films isgenerally not good compared with high-temperatureprocesses. There has been much research dedicatedto finding a way to obtain high-quality films at lowtemperatures by developing new deposition tech-nologies and improving the deposition processes,including sputter deposition [63], electrobeam de-position [64], and PECVD [65–67]. All thesedeposition technologies are also available for in-line production through a RTR process, known asweb coating [40].
Vitex System developed ‘‘Flexible Glass,’’ whichis a commercially available, PEN-based substratewith the company’s Barix barrier that utilizesdefect-decoupling layers of vacuum deposited poly-acrylate between multiple layers of gas barriermaterials such as Al2O3 [39]. This substrate issufficiently impermeable to moisture and oxygenfor application to OLEDs with a low WVTR below5� 10�5 g/m2/day.
General Electric developed a transparent plasticsubstrate that provides the high hermeticity that isexpected to be required for many organic electronicdevices [11]. A PC film with high-temperaturecapability and a novel single-graded layer ultra-high barrier (UHB) coating can effectively stopdefects from propagating through the coatingthickness. The graded UHB coating with a thicknessof less than 1 mm can significantly reduce theeffective edge diffusion of water through the organicportion of the barrier coating, yielding a lowWVTRof 5� 10�6–5� 10�5 g/m2/day.
For silicon oxide barrier layers, Erlat et al. [68]demonstrated that several PECVD deposition con-ditions of SiOx, including deposition time, power,pressure, and gas flow rate, can significantly affectthe barrier performance. Sufficient time is requiredfor plasma stabilization and the deposition ofuniform SiOx coatings. Coating properties such as
ARTICLE IN PRESSM.-C. Choi et al. / Prog. Polym. Sci. 33 (2008) 581–630592
thickness and density, as well as composite proper-ties, are improved with increasing deposition powerdue to the enhanced excitation and dissociation ofthe siloxane and O2 molecules in the plasma. Anincrease in the substrate temperature improves themobility of the reactive species and expedites cross-linking of the SiOx network, reducing the propensityfor porous defects. The system pressure must besufficiently low and the flow rate of siloxane shouldbe sufficiently high. On the other hand, Wuu et al.[53] increased the deposition temperature from 80 to170 1C. As the substrate temperature increased, thedeposition rate, adhesion, and roughness valueswere found to increase while the WVTR decreasedto a value of nearly 0.3 g/m2/day at 150 1C. Underoptimum conditions, the WVTR for PES can bereduced to a level of 0.1 g/m2/day with a double-sided barrier coating of 100 nm on each side at150 1C.
In the case of silicon nitride layers, Wuu et al. [67]reported SiNx barrier coatings on flexible PESsubstrates by PECVD for oxygen and waterresistant applications. The NH3/SiH4 flow ratiohas played an important role in SiNx film char-acteristics, such as bond configuration, transmit-tance, refractive index, deposition rate, adhesion,WVTR, and OTR. The durability of thin barriercoatings on polymer substrates relies on theircoating thickness, and parylene/SiNx/PC resultedin an improved permeation resistance of the coatingas well as an improved resistance to crack formation[57]. The WVTR and OTR of the optimized SiNx
and parylene multilayers composed of parylene-(600nm)/SiNx(100nm)/parylene(600nm)/SiNx(100nm)on the PC substrate can be maintained at levelsnear 0.01 g/m2/day and 0.1mL/m2/day, respectively,after 3000-times cyclic bending. The effects ofplasma pretreatment for SiNx barrier coatings onAr, N2, and O2 plasma-treated PC substrates werealso reported [58]. After Ar plasma treatment for60 s, the WVTR and OTR of a 50-nm-thick SiNx
barrier coating on PC substrate after bending for6000 times decreased to values near 0.01 g/m2/dayand 0.1mL/m2/day, respectively.
On the other hand, Iwamori et al. [59] developeda new transparent gas barrier material based onsilicon oxynitride (SiON) via reactive sputtering innitrogen plasma. SiOxNy films have a fine andamorphous structure without pinholes or cracks,increasing the density and showing lower OTR thanthat of the SiOx films. Baik et al. studied the waterbarrier properties of composite films consisting of
silicon oxide and tin oxide, deposited by thermalevaporation on PC substrates [60]. Water vaporpermeation through the composite films is signifi-cantly affected by the chemical interaction of watervapor with the composite oxide films and themicrostructure of the composite oxide films. Asthe tin oxide is added to the silicon oxide, thepolarizability of the composite oxide films increasesand the composite films have a more densely packedmicrostructure with decreasing grain size. Kim et al.[61,62] investigated a variety of inorganic materialsand inorganic composites as barrier materials onpolymer substrates using the electron beam eva-poration system. They showed that a MgO thin filmhas a lower WVTR value than any other inorganicthin film and that the WVTR of the inorganic filmcan be dramatically minimized by adopting theinorganic composite.
Weaver et al. [69] reported long-lived FOLEDsfabricated on plastic substrates whose compositebarrier consists of alternating layers of polyacrylatefilms and an inorganic oxide. The permeation rateof water vapor through a flexible substrate wasreduced to less than 2� 10�6 g/m2/day and resultedin a half-life of 3800 h from an initial luminance of425 cd/m2.
3.4. Permeation rate measurements
Devices such as OLEDs have undergone signifi-cant development, but there are no standards for themeasurement of barrier properties. Highly sensitivepermeation measurements are crucial for the char-acterization and development of polymeric sub-strates for flexible display applications, especiallyOLEDs, which are very sensitive to water andoxygen. The current specifications for the permea-tion rates of OLED packaging are on the order of10�5mL/m2/day at standard temperature and pres-sure for oxygen and 10�6 g/m2/day for water vapor[43]. However, there is no method available commer-cially for the measurement of ultralow moisturepermeation for OLED applications. Therefore, thedevelopment of a moisture permeation measurementsystem will be essential for investigating permeationmechanisms and understanding device lifetimeand degradation phenomena. The commercial equip-ment available from MOCON is limited to asensitivity of 5� 10�4 g/m2/day under varying tem-perature conditions [70].
Nisato et al. [71] developed a method based ondetecting the optical degradation of calcium to
ARTICLE IN PRESSM.-C. Choi et al. / Prog. Polym. Sci. 33 (2008) 581–630 593
measure the permeation rate of OLEDs. Thiscalcium test method can effectively increase thespeed of research and development on thin-filmencapsulants suitable for OLED devices by provid-ing quantitative results and sample configurationsinaccessible to commercial equipment, which isnow limited to 5� 10�3mL/m2/day for O2 and5� 10�3 g/m2/day for H2O for 4-in diameter sam-ples of coated plastic. Effective water vaportransmission rates down to 10�6 g/m2/day weredetected. Paetzold et al. [72] proposed the idea ofmeasuring the amount of oxidative degradation in athin Ca sensor by in situ resistance measurements.The resulting baseline contribution to the WVTR ofa glass reference is below 10�6 g/m2/day underaccelerated test conditions. Ranade et al. [73] usedtwo chambers, with a high pressure side and anultra-high vacuum side, separated by the flexiblesample to be analyzed, allowing measurement ofpermeation rates as low as 1� 10�6 g/m2/day forHe, 1� 10�6 g/m2/day for O2, and 5� 10�7 g/m2/dayfor Ar.
4. Transparent electrodes
Most electro-optic devices such as LCDs,OLEDs, and electronic papers (e-papers) requireelectrically conductive and transparent electrodes.Indium tin oxide (ITO) thin films deposited on glasssubstrates have been widely used as transparentconducting electrodes in many electro-optic devicesbecause they possess attractive properties withrespect to visible transparency and electrical con-ductivity [2]. However, ITO thin films have severaldrawbacks. They are very expensive because indiumis one of the rarest material. They require high-temperature processes to get good qualities on glasssubstrates. Most of all, they are so brittle that theyare not easy to apply to flexible displays.
Studies aimed at developing new transparentelectrodes include several directions: developingnew low-temperature processes using ITO such aspulsed-laser deposition; developing alternative ma-terials to ITO such as zinc oxide; finding new rigidmaterials available for flexible displays such astransparent polymer electrodes.
4.1. Transparent conducting oxides (TCOs)
There have been remarkable applications ofTCOs in the area of flat panel displays such astelevisions, computers, electrochromic windows,
photovoltaics, and hand-held devices. Coupled withthe increased importance of TCO materials, therehave been many improvements over several years inthe science of these materials [74].
Among the TCOs, ITO has been used mostwidely as a transparent electrode material. ITOfilms show high optical transmittance of 490% inthe visible light region and a low electrical resistivity(2� 10�4O cm) when deposited on glass substratesunder optimized conditions, which include highsubstrate temperatures of 250–300 1C or a post-annealing process. But, the good qualities of atransparent electrode are difficult to obtain whenITO is deposited on polymer substrates that possesslow thermal resistance and high thermal expansion.Generally, low-temperature deposition results in highelectrical resistance and poor optical transparency.
To overcome these problems, Kim et al. [75]induced pulsed-laser deposition on PET substrates.They obtained an electrical resistivity as low as�4� 10�4O cm and an average transmittance of90% in the visible range at the substrate tempera-ture of 100 1C. Chung et al. [76] deposited ITO filmsat room temperature and annealed them through anXeCl excimer laser. They found that excimer laserannealing (ELA) significantly improves the crystal,electrical, optical, and etching characteristics ofroom temperature deposited ITO. The sheet resis-tance of irradiated films decreased from 1.91� 10�3
to 2.5� 10�4O cm, while the optical transmissionincreased from 70% to 85%. Many other depositiontechniques, including radio-frequency (RF) reactivemagnetron sputtering [77], sputter-type negativemetal ion source [78], and oxygen ion beam assisteddeposition [79], have also been reported. Lim et al.[20] prepared FOLEDs using fluorine-containingcolorless PI substrate with ITO deposition by RFmagnetron sputtering. They found that the sheetresistance of the ITO films on the colorless PIsubstrate is approximately 20O/cm2 when thesputtering temperature is 4150 1C and is lowerthan any resistance values that have been reportedpreviously for ITO films coated onto polymersubstrates. Fortunato et al. used a zinc oxide bufferlayer between polymeric substrates and ITO films.They reported that the electrical properties of ITOthin films deposited at room temperature on PENsubstrates were improved by more than two ordersof magnitude by using a zinc oxide (ZnO) bufferlayer [80].
On the other hand, there have been many studiesreplacing ITO with ZnO because ZnO exhibits high
ARTICLE IN PRESS
Fig. 8. Change in resistance after repeated bending to r ¼ 6mm
as function of the number of cycles for ITO and ITO–Ag–ITO
multilayers with Ag thickness of 4 and 12 nm [89]. Reproduced
from Lewis, Grego, Chalamala, Vick, and Temple by permission
optical transmission and good electrical conduction.In addition, ZnO is more stable in activatedhydrogen environments than other TCOs and isnot expensive compared to ITO. Ha et al. depositedaluminum-doped zinc oxide (AZO) thin films at thetwo-substrate deposition temperatures of 100 and200 1C with a RF magnetron sputtering system onthe fluorinated PI substrate [81]. The sheet resis-tance of the AZO film deposited onto the fluori-nated PI substrate at 200 1C and the opticaltransmission of the AZO/PI film in the visible lightrange are about 1.94� 10�4O cm and 81.47%,respectively. ZnO films with different dopingsystems were also investigated, such as Zr-dopedZnO thin films [82], Ti-doped zinc oxide thin films[83] and gallium-doped zinc oxide films [84].
Other materials constituting transparent conduct-ing electrodes, such as antimony-doped tin oxidefilms [85,86] and multi-oxide combinations includ-ing In2O3–ZnO films [87] and Cd–In–Sb–O films[88], were also reported.
4.2. TCO– metal– TCO (TMT) multilayers
TCO films such as ITO have been commonly usedas electrode materials in flat panel displays, provid-ing good electrical conductivity and high transpar-ency in the visible region. However, they are toobrittle to use as flexible displays. Metal films, on theother hand, have good mechanical properties forflexible display applications but they are nottransparent. In this regard, TCO–metal–TCO multi-layers such as ITO–metal–ITO have been studied asflexible transparent inorganic electrodes.
Lewis et al. [89] used ITO–Ag–ITO as a trans-parent electrode for a FOLED instead of ITO. Inthis case, the ITO layers provided good energy-levelalignment for efficient incorporation of holes intothe organic layers while the addition of the ductilesilver layer gave improved robustness under me-chanical strain. ITO–Ag–ITO provided significantlyreduced sheet resistance compared to ITO, yieldingabout 30O/cm2 and 80% optical transmission at anAg layer thickness of 8 nm, and improved bendingproperties both as a function of radius and as afunction of cycling. The resistance data for samplesbent to a radius of 6mm as a function of the numberof cycles are shown in Fig. 8. After 10,000 cycles thevalue of R/R0 for the ITO sample was 7.4.
Fahland et al. [90] also examined different layerstacks of the design ITO–Ag–ITO. They found thatthe layer stack can be optimized to provide sheet
resistances below 16O/cm2 at a total light transmis-sion of over 80% at 550 nm, as shown in Fig. 9.Under these conditions, the ITO was deposited atambient temperatures because the electrical proper-ties of ITO are of minor importance.
Liu et al. [91] reported ZnS/Ag/ZnS nano-multi-layers with a sheet resistance of 3O/cm2 and aluminous transmittance of about 90%.
4.3. Conducting polymers
Polymers had long been known as electricalinsulators when Heeger et al. in 1976 discoveredthe extremely improved conductive polymer, poly-acetylene [2]. With this event, thousands of scientistsbegan to conduct research to develop polymers thatare stable in the conducting state and processable atlow cost, including polyacetylene, polyaniline,polypyrrol, polyphenylene, poly(p-phenylene viny-lene), and polythiophene. Recently, polyanilinedoped with camphor sulfonic acid (CSA), i.e.PANI-CSA, was found to exhibit metallic conduc-tivity (4103 S/cm) [92].
Among the conducting polymers as shown inFig. 10, polyaniline and polypyrrol have alreadybeen commercialized in applications for corrosionprotection, antistatic materials and fiber sensors.Polyphenylene and poly(p-phenylene vinylene) havebeen investigated as emitting materials of OLEDsby controlling the effective conjugation length,while thiophene derivatives have been investigated
as active materials for OTFTs or solar cells bycontrolling the optical absorption and transport viathe organization of the thiophene chains anddomains [93].
Conducting polymers have attracted a great dealof attention as potential replacements for ITO inmany electronic and optoelectronic devices andespecially for flexible displays due to their goodmechanical strength and their electrical and optical
stability during bending [94,95]. Polythiophene,more specifically poly(3,4-ethylenedioxythiophene)(PEDOT), has been widely studied as the mostpreferred candidate for transparent conducting-polymer anodes, as it exhibits a very high conducti-vity of about 300 S/cm and is almost transparent inthin, oxidized films and shows a very high stabilityin the oxidized state [96,97]. The solubility problemwas subsequently improved by using a water-soluble
ARTICLE IN PRESSM.-C. Choi et al. / Prog. Polym. Sci. 33 (2008) 581–630596
polyelectrolyte, poly(styrene sulfonic acid) (PSS), asthe charge-balancing dopant during polymerizationto yield PEDOT/PSS. This combination resulted ina water-soluble polyelectrolyte system with goodfilm-forming properties, high conductivity of about10 S/cm, high visible light transmission and excellentstability. Films of PEDT/PSS can be heated in air at100 1C for over 1000 h with only a minimal changein conductivity. Due to their relatively low con-ductivity, conducting polymers, especially PEDOTderivatives, have been primarily used as a bufferlayer between ITO and a hole-transporting layer,providing a good interface between the ITO anodeand the organic hole transport layer [98].
Kim et al. [99] used PEDOT/PSS dispersed in anaqueous solution with added glycerol as the anodefor OLEDs. The surface sheet resistance of 130-nm-thick films used in the study was 1850O/cm2. Theiroptical transparency was 90%. Their work functionswere in the range of 5.070.1 eV, comparable to thatof a pre-cleaned and oxygen plasma-treated ITO.Devices using this polymer as an anode showedgood external EL quantum efficiency with 0.73% at100A/m2, which compares well with an OLED usinga standard ITO anode. In addition, an inducedbuffer layer between the conducting polymer and thehole-transporting layer reduced micro-shorts, whichlead to a leakage current, with a slight increase in thework function. They also reported a simple methodof patterning the conducting-polymer electrode onvarious substrates including plastics [100].
Joo et al. reported the effects of organic solventson the charge transport properties of the PEDOT/PSS systems, observing the increase in roomtemperature DC conductivity of the systems from�0.8 to �80 S/cm with a change of solvent [101].The screening effect due to a polar solvent betweendopants and polymer main chains plays an im-portant role in the charge transport properties, suchas conductivity and its temperature dependence.
Louwet et al. [102] added high boiling solventssuch as N-methyl pyrrolidone (NMP) during thecoating of PEDOT/PSS on polyester substrates,which decreases the surface resistance by threeorders of magnitude, resulting in transparent con-ducting electrodes with 350O/cm2 and a visual lighttransparency of 80%.
4.4. Carbon nanotube (CNT) thin films
Nowadays there are a significant number ofstudies of CNTs, especially single-walled carbon
nanotubes (SWNTs), and their applications, includ-ing CNT composites, electrical devices, hydrogenstorage, field emission devices, nanometer-sizedelectronic devices, and sensors, due to their uniqueelectrical, mechanical, and optical properties [103].In the application of flexible electronics, flexible andtransparent CNTs are some of the most promisingmaterials because SWNTs have metallic propertiesas well as semiconductive properties and theirelectronic behavior, such as their conductivity andtheir work function, can be adjusted by designingthe ways of rolling the graphene sheet.
Martel et al. used transparent SWNT sheets thatgave a luminance efficiency of the CNT-basedOLED of 1.4 cd/A at the maximum achievedbrightness of 2800 cd/m2, which is comparable toan optimized ITO anode device made under thesame experimental conditions [104]. Zhou et al.[105] applied arc-discharge nanotubes to hole-injection electrodes for OLEDs providing a sheetresistance of �160O/cm2 at 87% transparency.Marks et al. made polymer-based OLEDs usingSWNT films on flexible PET substrates [106].Rowell et al. [107] fabricated flexible transparentconducting electrodes by printing films of SWNTnetworks on plastic for flexible solar cells. Gruneret al. fabricated transparent and flexible transistorswhere both the bottom gate and the conductingchannel are CNT networks of different densities andparylene N is the gate insulator [108].
On the other hand, Shan and Cho showed thatfor SWNTs with smaller diameters (o1 nm), thework function is very sensitive to chirality ordiameter while for those of larger diameters(41 nm), there is no significant chirality or diameterdependence as shown in Fig. 11 [109]. For class Itubes, nanotube work functions are very close to thework function of a graphene sheet, �4.66 eV. Forclass II tubes, work functions of (n, 0) tubes increasedramatically, while those of (n, n) tubes show adecrease in work function; (6,0), (5,0), and (4,0)tubes are metallic due to s*�p* hybridization.
There have been various attempts to overcomethe low solubility of CNTs, including using variouscommon organic solvents [110,111] or oleum [112],cutting long materials into short, using open-endedpipes with 100–300 nm lengths and forming a stablecolloidal suspension in water with the help ofsurfactants [113], and dissolving full-length CNTsin common organic solvents [114].
For electrode patterning, Wu et al. [115] devel-oped a simple process including vacuum filtering of
ARTICLE IN PRESS
Fig. 11. Work functions of (n, n) and (n, 0) nanotubes of different diameter [109]. Reproduced from Shan and Cho by permission of
a dilute, surfactant-based suspension of purifiednanotubes onto a solvent-removable filtratingmembrane for the fabrication of ultra-thin, trans-parent, optically homogeneous, electrically conduct-ing films of pure SWNTs, and the transfer of thosefilms to various substrates. For an equivalent sheetresistance of 70–90% for 50 nm thin films, theelectrical resistivity was measured to be 1.5�10�4O cm, comparable to commercial ITO. Othermethods of deposition or patterning of SWNT,including spin-coat [116], vacuum filtering [117],transfer with poly(dimethysiloxane) (PDMS)stamps [118], line-patterning [119], the Langmuir–Blodgett method [120] and controlling dropletsusing a gas flow [121], have also been reported.
5. Electro-optic materials
5.1. Liquid crystal displays (LCDs)
Flat panel LCDs have extended their boundariesfrom small size mobile phones to large-size televi-sions because they are thin and lightweight and havea low power consumption and excellent resolution.However, there are critical problems of stabilitywhen applying conventional LCD technologies toflexible LCDs [122]. Flexing the panel creates forces
that will cause the liquid crystal to flow, resulting incell-gap variations across the panel, visual distor-tions, and artifacts, which can easily becomeirreversible [5]. Therefore, it has become a mostimportant challenge to prevent liquid crystal flowunder the pressure caused by the deformation that isunavoidable in flexible applications [123].
To ameliorate these cell-gap problems, liquidcrystal materials have been separated out in theform of droplets or domains randomly distributedthroughout the polymer binders. Technical ap-proaches include polymer-dispersed liquid crystals(PDLCs), pixel-encapsulated liquid crystals, photo-enforced stratification, and encapsulating liquidcrystals into polymer capsules.
PDLC films can be made either by phaseseparation methods [124] or by emulsion methods[16]. Schneider et al. [125] encapsulated liquidcrystal by polymerization-induced phase separationin which a homogeneous mixture of photo-reactivemonomers and liquid crystals is subsequentlyexposed to ultraviolet light and the liquid crystalsare continuously separated and form dropwisedomains to apply to flexible LCDs. They controlledphotopolymerization chemically via the fraction ofthe crosslinker in the polymer/LC compositethrough the functionality of the pre-polymer to give
ARTICLE IN PRESS
Fig. 13. A schematic process of fabricating a pixel-encapsulated flexible LCD having a multifunctional elastomer substrate: (a) an
elastomer substrate fabricated by a replica molding technique, (b) the SEM image of the PDMS elastomer substrate duplicated from the
master, and (c) the flexible LC cell with a multifunctional elastomer substrate used as the top substrate [128]. Reproduced from Kim,
Hong, Yoon, and Lee by permission of American Institute of Physics, Maryland, USA.
Fig. 12. Schematic diagram of the flexible encapsulated cholesteric liquid crystal displays [125]. Reproduced from Schneider, Nicholson,
Khan, Doane, and Chien by permission of Society for Information Display, California, USA.
large pancake-like droplets, as shown in Fig. 12, of�10 mm in diameter and producing very little lightscattering. Link et al. reported a technique forproducing monodisperse liquid crystal dropletsformed into a precisely ordered two-dimensionalhexagonal-close-packed monolayer from liquidcrystal emulsions, which switch between diffractionand transparency [126]. Chari et al. [127] fabricatedsingle-substrate cholesteric LCDs by colloidal self-assembly. This emulsion-based close-packed PDLCfilm was fabricated on a moving-web coatingmachine.
Lee et al. reported a pixel-encapsulated flexibleLCD with a multifunctional elastomer substratefabricated by a replica molding [128]. Fig. 13 shows
a process of fabricating a pixel-encapsulated flexibleLCD having a multifunctional elastomer substrate,with the scanning electron microscopy (SEM) imageof the PDMS elastomer substrate duplicated fromthe master and the flexible LC cell with a multi-functional elastomer substrate used as the topsubstrate. Kim et al. fabricated flexible LCDs witha stamped polymer wall structure and a phaseseparated polymer layer. The stable LC structurecould be achieved by isolating LC molecules intothe pixel surrounded by the micro-patterned PDMSwall structures [129].
Penterman et al. introduced single-substrateLCDs made by paintable and photo-enforcedstratification in which a coated film is transformed
ARTICLE IN PRESS
Fig. 14. The two-step photopolymerization process: (A) the UV
irradiation through a mask results in the formation of polymer
walls; (B) the second UV irradiation step results in the formation
of the polymer sheet; (C) top view of the polymer covered LC
layer seen through a polarization microscope [131]. Reproduced
from Vogels, Klink, Penterman, Koning, Huitema, and Broer by
permission of Society for Information Display, California, USA.
into a polymer-covered liquid crystal layer[130,131]. As shown in Fig. 14, the paintable LCDsare manufactured by a two-step photopolymeriza-tion process. The first UV exposure, using wave-lengths beyond the absorption region of the reactivestilbene dye, is through a mask, and the localizedpolymerization forms an array of polymer walls.During the second UV exposure, using UV light ofshorter wavelength in the absorption band of thedye, the polymer sheet is formed. The resulting
On the other hand, Lee et al. proposed a high-speed flexible display based on a deformed helixferroelectric liquid crystal in a vertically alignedconfiguration with a periodic array of columnarspacers on the top sides of in-plane electrodes [132].The mechanical stability of the flexible display wasachieved using a periodic array of columnar spacersformed directly on the top side of the in-planeelectrodes by a photolithography technique to giveflexiblility, uniform alignment, fast response andgray scale capability. Sikharulidze reported anorientational effect in a suspension of nematicliquid crystal with solid nanoparticles, controlledby an optically hidden electrophoretic effect [133].Polymer wall-stabilized smectic A liquid crystals tobe used for bistable flexible displays [134] and liquidcrystal microcapsules with a perpendicular align-ment shell [135] were also reported.
Samsung Electronics introduced large area(5.0 in) full color transmissive a-Si TFT-LCDs witha resolution of 400� 3� 300 lines (100 ppi) [14]. Allthe processes of TFT, color filter, and LC werecarried out below 150 1C on PES films. The overallmodule, including a backlight unit, is bendable andhas a thickness of about 1.2mm and a weight of22.0 g, which are one-third of the thickness andweight of normal glass-based displays. Kent Dis-plays developed the first ever reflective cholestericLCDs on single textile substrates made with simplecoating processes by sequential coating of variousfunctional layers on fabric materials [16]. Encapsu-lation of the cholesteric liquid crystal droplets in apolymer matrix and the mechanical flexibility of theconducting polymers allow them to create durableand highly conformable textile displays.
5.2. Electronic papers (e-papers)
When a TV or computer prevailed in our life, wethought newspapers would disappear. However, weare still getting news from newspapers because oftheir readability. It is very inconvenient to read longnovels on a computer screen. Paper-like readabilityis the most important target of electronic paper.Omodani studied how to achieve readability in hisexperiments, presenting the result that it may bepossible to improve display readability simply byadopting a free-reading style [136]. Electronic papercan be a thin, high-contrast, reflective display thatcan be flexed, bent, rolled-up, and folded and, in
ARTICLE IN PRESSM.-C. Choi et al. / Prog. Polym. Sci. 33 (2008) 581–630600
particular, it can have a very low power consump-tion. A portable device that uses this technology willlook more like a pad of paper than a standardlaptop. Applications of electronic papers includenewspapers, magazines, greeting cards, and evencereal boxes, bumper stickers, and wallpaper[137,138].
Most electronic papers use a bistable property,in which the image is not refreshed until rewritten,and a reflective characteristic, in which the imagecan be seen through another light source, incomparison with emissive displays such as LCDsor OLEDs. Therefore, they do not require aconstant voltage to maintain an image, whichgreatly decreases power consumption. These elec-tronic paper technologies include cholesteric, ferro-electric, PDLC, electrochromic, electrophoretic,electrodeposition, and OLED techniques. In thissection, we will mainly discuss electroporetic tech-niques [139].
Electrophoresis means physical motion under anelectrical force, while electrophoretic displays switchimages by the movement of particles under voltage[140]. Electrophoretic displays are almost invariablyreflective and bistable. They have less stringentbarrier requirements and can be easily made in largesizes compared to LCDs and OLEDs. The mainlimitation is a slow switching speed, makingelectrophoretics a poor choice for video or evenrapid scrolling. This largely eliminates the cellphone, game, and parts of the auto market.Electronic paper and electronic books are consid-ered as key applications for electrophoretics becauseof their bistability [4,5].
E-ink Corporation uses a microencapsulatedelectrophoretic material that consists of millions of
Fig. 15. Schematic diagram of the cross section of electronic-ink micr
Kodaira, and Inoue by permission of Society for Information Display,
microcapsules containing charged pigment particlesin a clear fluid. A negative voltage applied to the topsurface causes the positive white particles to moveto the top of the capsule and the surface to appearwhite, whereas reversing the electric field causes thenegative black particles to appear at the top surfaceand create a dark spot, as shown in Fig. 15.[141,142]. They fabricated a display on a bendableactive-matrix-array sheet, which is less than 0.3mmthick, has a pixel density of 160 pixels� 240 pixelsand a resolution of 96 pixels/in, and can be bent to aradius of curvature of 1.5 cm without any degrada-tion in contrast. They also introduced flexibleelectrophoretic displays with LTPS TFTs [142]and with OTFTs [9].
Gyricon Media uses a twisting-ball display[143,144]. Their basic display structure consists ofa thin layer of elastomeric material embedded withbichromal. Each sphere sits inside its own cavity,which is filled with silicon oil, allowing the sphere torotate. The spheres are fabricated such that onehemisphere appears white and one hemisphereappears another color, with each hemisphere havinga different permanent charge. In a stable situation,the balls gravitate to one side of the cavity andadhere to the wall. When an electric field is appliedacross the sheet of elastomer, the ball is releasedfrom the side of the cavity so it can rotate to alignwith the field. After the ball has rotated, it settlesagain against the cavity wall and becomes stablyattached until a reverse electric field is applied. If animage is converted into a planar voltage patternapplied to the surface of the sheet, then the electricalfield induces rotation of the balls and the image istransferred to the balls. Light reflected off the ballsdisplays the image.
o-capsules [142]. Reproduced from Kawai, Miyasaka, Miyazaki,
California, USA.
ARTICLE IN PRESSM.-C. Choi et al. / Prog. Polym. Sci. 33 (2008) 581–630 601
Sipix Imaging Inc. developed the Microcups, inwhich a particle suspension is deposited into80–160 mm wide compartments embossed into resinsubstrates [145,146]. An electrophoretic fluid com-prised of charged pigment (TiO2)-containing micro-particles dispersed in a colored dielectric solvent isenclosed and seamlessly top-sealed in the Micro-cupss. Color rendition may be achieved either byusing a color filter or by sequentially filling andsealing red (R), green (G), and blue (B) electro-phoretic fluids in the Microcupss. Excellent colloi-dal stability has been observed even after theelectrophoretic composition was centrifuged at1000g for more than 30min. They also developednew RTR manufacturing processes based on thistechnique.
Bridgestone Co. presented a new electrophoreticdisplay named Quick Response Liquid PowderDisplay (QR-LPDs) [147,148]. This display usesan electronic powder, named liquid powder, whichis a high-fluidity material that combines the proper-ties of a powder and a liquid and is highly sensitiveto electricity. Fig. 16 shows the cross-sectionalstructure and the operational principle ofQR-LPDs. Two types of powders, includingnegatively charged white and positively chargedblack versions, are put into an area between twoITO-patterned glass plates. The rib forms a cell gapand prevents the powders from mixing. The rest ofthe space is filled with ordinary air instead of liquid.Using this new material, QR-LPDs has shownoutstanding clarity similar to that of paper alongwith excellent image stability, quick response, highresolution, clear threshold characteristics, and lowpower consumption.
Hayes et al. introduced a novel reflective-displayprinciple, in which the optical switch is driven by theso-called electro-wetting effect [149,150]. This prin-ciple has the potential for use as a fast, high-
Fig. 16. The QR-LPDs architecture and the operation principle [14
permission of Society for Information Display, California, USA.
brightness color display. Fig. 17 demonstrates theprinciple of the reflective electro-wetting display.Fig. 17(a) shows the optical stack, comprising awhite substrate, a hydrophobic insulator, a coloredoil layer, and water. At equilibrium the colored oilnaturally forms a continuous film between the waterand the hydrophobic insulator. However, when avoltage difference is applied across the hydrophobicinsulator, an electrostatic energy is added to theenergy balance and the stacked state is no longerenergetically favorable. The energy of the systemcan be lowered by moving the water so that it makescontact with the insulator, thereby displacing the oil(Fig. 17(b)) and exposing the underlying whitesurface. The balance between the electrostatic andcapillary forces determines how far the oil is movedto the side. In this way, the optical properties of thestack when viewed from above can be continuouslytuned between a colored off-state and a white on-state, provided the pixel is sufficiently small so thatthe eye averages the optical response.
Other ideas including electrodeposition [151],photochromic compounding [152], fine particlesdispersed in a nematic liquid crystal [153], andhollow fibers [154] have also been reported.
5.3. Polymer light-emitting diodes (PLEDs)
OLEDs have many excellent properties as dis-plays including Lambertian emission, good colors,no cell-gap problem, and the potential for usingmany promising new solution-processing techni-ques. In this regard, the OLED is a promisingcandidate for flexible displays in the future [4].
However, there exist many challenges yet to beovercome [5]. The biggest challenge is water andoxygen sensitivity, which is a factor on both sides ofthe display when a polymer substrate is used. Wehave already discussed this issue in Section 3.
7]. Reproduced from Hattori, Yamada, Masuda and Nihei by
ARTICLE IN PRESS
Fig. 17. Electrowetting display principle: (a) no voltage applied, therefore a colored homogeneous oil film is present; (b) dc voltage
applied, causing the oil film to contract. The top view photographs in (c) and (d) demonstrate the corresponding oil retraction obtained
with a homogeneous electrode [150]. Reproduced from Feenstra, Hayes, Camps, Hage, and Johnson by permission of Society for
Second is to develop new active materials that havehigh thermal and mechanical stability, large-areaprocessibility, and high efficiency. Third is todevelop new manufacturing processes for FOLEDs,which are not well established at this time [155]. Thelast challenge concerns active back plain technol-ogy, which will be dealt with as the following topic.
OLEDs are classified as small molecule OLEDs(SMOLEDs) and polymer OLEDs (PLEDs). SMO-LEDs are being used for the small-sized displays ofmobile phones. Small molecules are easily synthe-sized and purified compared to polymers [156].Therefore, a large number of high-purity conjugatedsmall molecular materials are commercially avail-able. They are also easy to fabricate in a multilayerstructure and more readily allow optimization of theturn-on voltage, luminescence, and efficiency. How-ever, they require a high-vacuum manufacturingsystem for deposition and they are sensitive totemperature. Therefore, they are difficult to use in alarge-area deposition system. On the other hand,conjugated polymers have good film properties and
good mechanical and thermal stability but it isdifficult to synthesize high-purity conjugated poly-mers and to fabricate multilayer structures [157].Most of all, conjugated polymers can be more easilyapplied to large-area manufacture through cost-effective solvent processes such as spin-coat, ink jetprinting, dipping, and spraying. Therefore, polymermaterials are an excellent choice for RTR manu-facturing as well as flexible displays. In this section,we discuss materials for PLEDs and their fabrica-tion techniques.
5.3.1. Electron injection/transport materials
To optimize electron injection and transportingperformance, there are two factors to be considered:how to lower the injection barrier between thecathode and the organic material and how to raisethe recombination probability of the carriers.Methods for lowering the injection barrier betweenthe cathode and the organic material include using acathode of low work function, introducing a bufferlayer between the cathode and the organic material
ARTICLE IN PRESSM.-C. Choi et al. / Prog. Polym. Sci. 33 (2008) 581–630 603
layer to tune the energy levels, or using organicmaterials with a higher lowest unoccupied molecu-lar orbital (LUMO) [158].
In the energy-level alignment, lower work func-tion metals such as Mg and Ca are known to be wellmatched with organic materials, but they are verysensitive to moisture or oxygen compared to higherwork function metals such as Ag and Al. Therefore,many attempts to find stable and low work functioncathode materials have been reported [156]. Thework function of a cathode is also controlled bychanging the deposition thickness of the nanolayerof the cathode [159].
As a buffer layer, Elliott et al. prepared a series oftervalent conducting polymers, which have a lowwork function of o3.6 eV and is tunable, via thethermal polymerization of transition metal diiminecomplexes [160]. Kim et al. induced polymer-insulating nanolayers between the emitting layerand the aluminum cathode in PLEDs in which ahole is the major charge carrier, showing that adevice with a nanolayer of lower dielectric constantprovides a higher luminescence quantum efficiency[161]. Shu et al. inserted a layer of non-ionic neutralpoly(ethylene glycol)-based surfactant between anelectroluminescent layer and a high work functionaluminum cathode, which showed comparable oreven better device performance than that of controldevices using calcium as the cathode [162]. Theyproposed that when both surfactant and aluminumare used as the cathode, the abundant hole injectionthrough a hole-transporting layer and hole pile-upat the inner side of the EL/surfactant interfacemight cause an effective electric field to induce therealignment of the dipole moment of those polarsurfactant molecules, thus lowering the barrier forelectron injection.
As organic materials with higher LUMO, metalchelates or metal complexes such as tris(8-hydro-xyquinoline) aluminum (Alq3) are the most well-known electron transport materials [163–165].Organic materials modified with electron withdraw-ing moieties reduce the barrier of electron injectionand block the hole injection. These electron-with-drawing groups include oxadiazole [166–168], azole[169,170], benzothiadiazole [171–173], cyano[174,175], quinoline [176–178], bonyl [179], silole[180–182], and perfluorin [183,184], as shown inFig. 18. These groups are used as the backbone orpendant group of a small molecule or polymer[185,186] and as the backbone or pendant of anemitting conjugated polymer [187,188].
5.3.2. Hole injection/transport materials
Similar to the case of electron injection/transport,there are several ways to lower the hole-injectionbarrier between the anode and the conjugatedpolymers, including using high work functionanodes, anode modification, fabricating the multi-layer of materials with various ranges of ionizationpotentials, and modifying conjugated polymers[156]. As discussed in the previous section, TCOs,conducting polymers, and CNTs are considered asflexible transparent anodes.
Surface treatment of the ITO anodes throughoxidization by O2 plasma, CF4/O2 plasma, UVozone treatment as well as acid and base treatmentincreases the work function nearly to 5 eV [189,190].
An additional hole-injection layer between theanode and a conjugation polymer layer decreasesthe hole-injection barrier to form cascade-shapedenergy levels. These hole-injection materials includea thin layer of platinum [191], metal phthalocyanine[192–194], PEDOT/PSS [96,97], tetra(fluoro)-tetra(cyano) quinodimethane (TF-TCNQ) [195], tris(4-bromophenyl) aluminum hexachloroantimonate(TBAHA) [196], and 4,40,400-tris(N,N-diphenylami-no) triphenylamine (TDATA) [197].
A self-assembled monolayer (SAM) has beeninvestigated to manipulate the energy-level offset atthe ITO-hole transport layer. Day et al. used SAMsof polar adsorbate molecules with the dipoleoriented outward from the surface an artificialdipolar layer where the work function is increased[198,199]. With this method the threshold voltagefor light emission can be reduced and the maximumluminance increased. Fujihira et al. enhanced holeinjection by fine-tuning of the work functionthrough surface molecular design of ITO [200].
Marks et al. used a SAM with hole transportmoieties [201,202]. They modified traditional holetransport molecules such as 1,4-bis(phenyl-m-toly-lamino) biphenyl (TPD) and N,N0-bis(l-naphthyl)-N,N0-diphenyl-1,10-biphenyl-4,40-diamine (NPB)with trichlorosilyl groups and spin-coated themonto the ITO surface, enhancing the ITO-holetransport layer contact via robust covalent bonding,and got dramatic OLED device performanceenhancement. Modification with doped p-conju-gated polymers [203], a polymer hole-injection layer[204], and hole-injecting conducting-polymer com-positions [205] have also been reported.
As hole transport materials, several triarylamineand cabazole derivatives shown in Fig. 19 have beenstudied. To improve the thermal stability of
ARTICLE IN PRESS
N
S
OZn
N
S
O
N N
O
NN
O
*
*
OC12H25
C12H25O
n
N
C8H17
S S**
NS
Nm n
*
CNNC
C6H13 C6H13
C6H13 C6H13
*
n
N N S
C8H17
S*
C8H17*
Si
Me Me
SBNCH3
H3C
H3CCH3
CH3
CH3CH3
H3C
H3C
FF
FF
FFF
F
FF
F
F
FF
FF
F
F
n
N N
N N
Fig. 18. Examples of electron transport materials containing (a) metal chelates [164], (b) oxadiazole [167], (c) benzothiadiazole [171],
conventional hole transport materials containingtriarylamines such as TPD and NPB, which have alow glass transition temperature (Tg) of below100 1C leading to crystallization at elevated tem-peratures, new high Tg hole-injection materialsincluding biphenyl diamine derivertives [206], star-shaped molecules [207], dendrimers with a cabazoledendron [208], and spiro-linked biphenyl diamines[209] were developed. Cross-linking or polymeriza-
tion also presents good thermal properties. Kimet al. [210,211] synthesized a diamine that hastriarylamine moieties and used this product to makea PI film for hole injecting and transport. They alsoprepared a hole-injecting-transporting layer by insitu mixing hole-injecting and hole-transportingmaterials by evaporation under vacuum to reducethe number of organic layers in OLEDs. OLEDswith hole-injecting-transporting layers with 25%
ARTICLE IN PRESS
NN
H3C CH3
H3C
H3C
CH3
CH3
NN
N
N
N
N
N
N
N
N
N
N
N
N
* N N N N *
O
O
O
O n
N
N CH
2 OC
O
O
FF
FN
CH2OC
OF
FF
O
N
CH 2OC
O
FF
F
O
N
S
N
SNSN
NN
H3C
CH3
NN
Fig. 19. Examples of hole-transport materials such as: (a) NPB [201], (b) TPD [202], (c) high Tg triarylamines [206], (d) star-shaped
molecules [207], (e) dendrimer types [208], (f) crosslinkable molecules [215], and (g) high Tg polymers [210].
hole transport material content displayed thehighest efficiency and the longest lifetime, whileOLEDs with hole transport materials only exhibitedvery poor device lifetime despite the higher effi-ciency [212–214]. Jen et al. thermally cross-linkedhole transport layers [215] and Fechet et al.synthesized bipolar transport materials that retainhole and electron transport moieties [216]. Molecu-larly doped polymeric network nanolayers with
TPD were also investigated as a hole-injectionlayer [217].
Other attempts to improve hole transport proper-ties include blending with hole transport polymers,electron transport polymers and light-emittingpolymers (LEPs) [218], copolymers containingemitting and hole-transporting moieties [219], andcopolymers containing emitting, hole-transporting,and electron-transporting moieties [220].
ARTICLE IN PRESSM.-C. Choi et al. / Prog. Polym. Sci. 33 (2008) 581–630606
5.3.3. Electroluminescent polymers
The photoluminescence wavelength of LEPs canbe controlled by increasing or decreasing the bandgap of the polymer by proper molecular designs[157]. These molecular designs include introducingsubstituents on the polymer backbone that changethe electronic structure of the materials, e.g. byintroducing torsion into the polymer backbone viasterically bulky side chains that reduce the effectiveconjugation length of the polymer or by breakingthe chain conjugation with non-conjugated mono-mers during the copolymerization process [157].
*co
*O
O
ON N
OO
yx
*
C8H1
N OO
S
*S
* C8H17
C8H17n
C8H17 C8H17
N Zn2+ N
NN
NN
** n
2PF6-
Br
Br n
OOIr
NS 2
OOIr
N
2
** m n p q
Fig. 20. Examples of light-emitting polymers such as: (a) PPV derivat
(g) self-assembled polymers [263], and (h) organic–inorganic hybrid ma
Fig. 20 shows several representative emittingmaterials for OLEDs.
Poly(p-phenylenevinylene) (PPV) is the mostpopular conjugated polymer of the LEPs. It hasgood thermal stability, good film qualities, andsuitable color tunability. However it has poorsolubility, low stability of oxidation and structuraldefects at the effective conjugating lengths. Toimprove solubility, a variety of side chains such asalkyl, alkoxy and the bulky cholestanoxy group inPPV have been introduced. Huang et al. introducedsilyl groups into a conjugated polymer to afford
ARTICLE IN PRESSM.-C. Choi et al. / Prog. Polym. Sci. 33 (2008) 581–630 607
good processability, amorphousness, and good film-forming ability [221,222]. Tsutsui et al. synthesizedsterically hindered fluorenyl substituted PPVs,which have excellent solubility and a high quantumefficiency [223,224]. Hsu et al. also introducedfluorenyl substituents to improve solubility andincrease steric hindrance in poly(2,3-diphenyl-1,4-phenylene) (DP-PPV) [225]. On the other hand,another problem of PPVs is that electron injection ismuch more difficult than hole injection, resulting inan imbalance in the rates for electrons and holes anda shift of the recombination zone toward the regionnear the interface of the polymer/cathode. Jin et al.[188] introduced oxadialzole moieties as electronwithdrawing side chains of PPVs, which increasesthe LUMO energy level and improves deviceperformance [226]. Fluoro groups [227] and cyanogroups [228] have been used as electron withdrawinggroups. Generally the colors of the PPVs give greento orange-red emissions by controlling the sidechains [229]. For blue emissions, Karasz et al.synthesized a series of PPV-related alternating blockcopolymers containing a conjugated block and non-conjugated block, which can be used directly asblue-green LEPs [230,231]. Shim et al. synthesizedPPV-related copolymers containing carbazole orfluorine groups [232]. There were also investigationsof polymer blending to give color tuning [233] orhigh efficiency [234].
PLEDs based on polyfluorenes (PF) are apromising candidate for the next generation ofFOLED displays because of their excellent proper-ties, such as good thermal and chemical stability,high quantum yield, good film-forming and hole-transporting properties and, especially, their bluelight emission. Polyfluorene homopolymers have alarge band gap and emit blue light. Significantefforts have been made to obtain polyfluorenederivatives with a variety of photoluminescencewavelengths. Studies of blue-emitting materialsinclude fluorine/carbazole copolymers for deep blueOLEDs [219,235], polyfluorenes containing bipolorpendant groups [220], quinoxaline-containing poly-fluorenes to give high electron affinity [236],copolymers based on fluorene and 2,5-di(2-hexylox-yphenyl) thiazolothiazole [237], and hyperbranchedalternating copolymers of tetrabromoarylmethane/silane and 9,9-dihexylfluorene-2,7-diboronic acid[238]. As green-emitting materials, fluorene- andbenzothiadiazole-based conjugated copolymers[172], fluorene copolymers containing a phenylenegroup, biphenylene group, or a thienylene group
[239], and CN-poly(dihexylfluorenevinylene) [175]and polydioctylfluorene [240] have been reported.To get red emissions, a narrow band gap comono-mer was introduced into the polyfluorene backbone.The low-band gap polymers include fluorine- andbenzothiadiazole-based conjugated copolymers[241], copolymers derived from fluorene and benzo-selenadiazole [242,243], polymers derived fromfluorene and naphthoselenadiazole [244], conju-gated polymers containing Eu3+ [245], and poly-fluorene copolymers containing electron-deficient2-pyran-4-ylidene-malononitrile moieties [246]. Caoet al. introduced conjugated polyelectrolytes basedon polyfluorene, which are soluble in polar solventssuch as alcohol due to the ionic side groups that areattached to the conjugated main chain [247,248].There have also been studies to improve electro-luminescent efficiency and stability and reduceintermolecular interaction by introducing cross-linked polyfluorene polymers [249], polyfluorene/PPV copolymers containing the bulky pendantbis(4-alkoxyphenyl) groups in the C-9 position[250], and poly(aryl ether)s containing ter- andpentafluorene pendants [251].
Polythiophenes have high thermal and mechan-ical stability and stable color tunability but theyhave poor solubility, short lifetimes, low colorpurity and are difficult to process [252]. Severalinvestigations of thiophene derivatives were made[253–255].
On the other hand, Kim et al. [256] studied thedoping effect of blue light-emitting electron trans-port molecules in blue organic light-emitting de-vices. They also fabricated efficient blue OLEDswith charge carrier confining nanostructures formedby wide band gap molecular doping [257]. Alter-nating terphenylene carbazylenevinylene copolymer[258] and poly(fluorenylenevinylene-terphenylenevi-nylene) containing phenyl pendant groups [259] alsowas reported as an emitting material.
Cross-linked polymers have good thermal andmechanical properties as well as good solventresistance. Hikmet and Thomassen [260] used anelectron beam to cross-link electroluminescentpolymers while Haarer et al. and O’Neill et al.applied sol–gel processing [261] and radical poly-merization [262], respectively. Self-assembled poly-mers, including polymers using Zn2+ to assembleorganic building blocks [263], layer-by-layer self-assembled films with water-soluble polymers [264],self-assembling multilayers using the Langmuir–Blodgett technique [265], and polymer-dielectric
ARTICLE IN PRESS
Fig. 21. Schematic of the LITI process [291]. Reproduced from
Lee, Kim, Suh, Kang, Choi, Park, Kwon, Chung, Baetzold,
Bellmann, Savvateev, Wolk, and Webster by permission of
nanolayer composites [266] have also been studied.Electroluminescent dendrimers are an exciting newclass of materials for OLEDs. Dendrimers havedistinct advantages including a modular approachto their synthesis, the ability to utilize a greaterrange of luminescent chromophores, and, mostimportantly, the ability to control their electronicand processing properties independently [267,268].There have also been studies using polyhedraloligomeric silsesquioxanes with organic–inorganichybrid structures as cores of starlike LEPs [269,270].
Considerable progress has been made withelectrophosphorescent OLEDs based on smallmolecules as the host materials since the utilizationof triplet emitters to improve OLED efficiency wasproposed [271]. Polymer-based electrophosphores-cent OLEDs are important for the solution process.Chen et al. [272] obtained efficient red emissionelectrophosphorescent PLEDs by simultaneous in-corporation of Ir complexes and charge transportmoieties into the side chains of polyfluorene whilePark et al. synthesized carbazole-based copolymerstethering blue-emitting iridium(III) bis[(4,6-difluor-ophenyl)-pyridinato-N,C20] picolinate (FIrpic) viacovalent bonding [273]. Cao et al. used electropho-sphorescent chelating copolymers based on linkageisomers of naphthylpyridine–iridium complexeswith fluorine [274]. Samuel et al. used iridium-coreddendrimers for electroluminescent green phosphor-escent polymers [275]. Blending host polymers withguest metal complexes was also investigated. In thiscase, the iridium complexes were modified withfluorines [276] and alkyl groups [277] to providebetter solubility with the host polymers.
Hybrid polymer–quantum-dot light-emitting de-vices are a new area drawing significant attentiondue to their easy processibility and the ruggednessof the polymers and the exotic optical properties ofthe quantum dots. A variety of quantum dots suchas CdSe/ZnS [278], CdSe/CdS [279], PbS [280] andCdTe [281] have been investigated for application tolight-emitting materials. Quantum dots that aredispersed in a hole transport material at differentconcentrations with different surface modifications[282], a quantum dot monolayer that is sandwichedbetween the hole-transporting layer and electron-transporting layer [283,284], and a quantum dotusing a thermally polymerized hole transport layer[279] were investigated. Ha et al. considered theperformance of OLEDs consisting of PEDOT/PSS-Ag nanocomposites with different concentra-tions of quantum-sized silver nanoparticles as a
hole-injection layer, showing that the turn-onvoltage significantly decreases when the concentra-tion of the Ag nanoparticles increases [285]. Bulovicet al. explored a new method for forming a large-area ordered monolayer of colloidal nanocrystalquantum dots [286].
5.3.4. Patterning technologies
There have been many studies on the fabricationand patterning of active layers for the developmentof a full-color application. Vapor deposition usingshadow masks is the most common way offabricating small molecule OLEDs (SMOLEDs)demonstrating high performance. However, thistechnique has difficulties in shadow mask alignmentfor large area applications.
For a large-area or a RTR application manyprinting processes are available, including screen,gravure, offset and inkjet. Among them, inkjetprinting is a promising technique for RTR proces-sing [287]. The initial impetuses for creating jet-printing technology for displays were the depositionof PLEDs, for which conventional photolithogra-phy is difficult because of material sensitivity, andthe reduction of the fabrication cost of color filtersfor LCDs [288]. Presently, jet-printed color filtersare the leading application of the technology inproduction. They are simple and economic pro-cesses scalable to large area patterning of OTFTs[289] or PLEDs [290]. The reliability of the inkjetheads is a major challenge in addition to developingsolution-processible high-performance materials.
ARTICLE IN PRESSM.-C. Choi et al. / Prog. Polym. Sci. 33 (2008) 581–630 609
Another currently available patterning techniqueis laser-induced thermal imaging (LITI) developedby 3M [291,292]. The LITI process utilizes a donorfilm, a highly accurate laser exposure system, a LEP,and a substrate, as shown in Fig. 21. The donor filmconsists of a transparent base film with severalcoated layers. The layer adjacent to the base film is alight-to-heat conversion (LTHC) layer that convertslaser energy to heat. The LTHC layer comprises amaterial that absorbs at the wavelength of irradia-tion and converts a portion of the incident radiationinto sufficient heat to enable transfer in the thermaltransfer layer from the donor to the receptor. Thistechnique is also available for large area fabrica-tions through a dry process and can apply to amulti-stacking structure.
Photolithography [293], hot micro-contact print-ing [294], and radiation-induced sublimation trans-fer [295] have also been reported.
6. Thin-film transistors (TFTs)
Low-temperature process technologies in TFTfabrication are the most crucial for flexible displays[296]. The TFT processes developed for flat paneldisplay using rigid glass substrates cannot readily beapplied for use with flexible plastic substrates due tothe limitations of process temperature, lack ofdimensional stability, and thermal stresses betweenthe TFT thin films and the substrate.
There are two main approaches for fabricatingTFTs on plastic substrates. One is to transfer high-performance poly-silicon or single-crystal siliconTFT devices processed at a high temperature on aglass, quartz, or silicon substrate onto a flexibleplastic substrate [297]. This approach providesoptimal TFT device performance with respect tomobility, leakage current, stability, and uniformityof the TFTs, whereas it is not a cost-effectiveprocess because of the wastage of the glass substrateand the additional cost of the transfer process.Another is to fabricate a TFT array directly on theflexible polymer plastic substrate whereby the TFT,such as amorphous silicon (a-Si), LTPS, or organicsemiconductor (OTFT), is fabricated at a tempera-ture less than 150 1C directly on the flexible plasticsubstrate [298].
6.1. Amorphous silicon TFTs
Hydrogenated amorphous silicon (a-Si:H) TFTsare widely used as switching devices in active-matrix
LCDs (AMLCDs) at the moment. These devices aregenerally made from hydrogenated amorphoussilicon and silicon nitride layers formed on glasssubstrates by PECVD with processing temperaturesexceeding 300 1C.
Long et al. [15] fabricated amorphous siliconTFTs at a high temperature of 250 1C on PIsubstrates. However, polymer substrates are nor-mally incompatible with typical high-temperatureprocesses because the maximum working tempera-ture of most polymer substrates is between 100 and150 1C. On the other hand, the low depositiontemperature leads to an excess of hydrogen incor-porated onto the growing amorphous silicon sur-face, which gives a high defect density and poorelectronic properties such as low mobility and highgate current-leakage [299,300]. The hydrogen con-tent of the amorphous silicon film deposited byPECVD was reported to be �20 at% at 1001C and�10 at% at 250 1C [301].
To get high-quality amorphous silicon with lowhydrogen content on plastic substrates at a lowtemperature below 150 1C, the concentration ofsilane gas was controlled [302–304] or new deposi-tion techniques such as very-high frequency (VHF)PECVD [305], hot-wire chemical vapor deposition[306] and electron cyclotron resonance plasma-enhanced chemical vapor deposition (ECR-PECVD) [307] were introduced.
Sazonov et al. [308] fabricated TFTs withamorphous hydrogenated silicon (a-Si:H), nano-crystalline silicon (nc-Si), and amorphous siliconnitride (a-SiNx) films using plasma deposition equi-pment at process temperatures as low as 75 and120 1C. The a-Si:H TFTs showed good performancecharacteristics including field-effect mobilitiesof 0.8 and 0.6 cm2/V s and threshold voltages of4.5 and 4V at 75 and 120 1C, respectively. Wonet al. [309] used organic materials that have similarmechanical properties to the plastic substrates asgate insulators, instead of inorganic materials suchas SiNx, on PES substrates. They got a field-effectmobility of 0.4 cm2/V s and a threshold voltage of0.7V, which were better than the case of a siliconnitride gate insulator where the field-effect mobilityand threshold voltage were 0.3 cm2/V s and 5V,respectively, when they introduced benzocyclobu-tene (BCB) gate insulators. They also got a field-effect mobility of 0.5 cm2/V s and a thresholdvoltage of 4.7V using a poly(4-vinyl phenol)(PVP) gate insulator [310]. Hong et al. [311] deve-loped a 5.0-in transmissive type plastic amorphous
ARTICLE IN PRESSM.-C. Choi et al. / Prog. Polym. Sci. 33 (2008) 581–630610
TFT-LCD with a resolution of 400� 3� 300 lines.All the processes of TFT, color filter, and LC werecarried out below 150 1C on PES films.
6.2. Low-temperature poly-silicon TFTs
Low-temperature polycrystalline silicon (LTPS)TFTs have good mobility, which is higher than thatof amorphous silicon TFTs by two orders ofmagnitude, and can reduce the channel size of theswitching TFTs to increase the emitting zones aswell as provide a more stable threshold voltage overthe life of the device [312,313]. Many depositionprocesses have been developed to make low defectand high-quality polycrystalline silicon thin films,including direct deposition [314], solid phase crystal-lization [315], laser crystallization [316], ELA [317],and sequential lateral solidification (SLS) [318].
But the current LTPS processes for AMLCDsemploy a typical process temperature in the range of450–600 1C to recrystallize the amorphous siliconfilms via excimer lasers [319]. Therefore, it is difficultto use conventional LTPS TFT processes forpolymer substrates that require a low processtemperature under 150 1C. To circumvent thislimitation, there have been two kinds of investiga-tions, including direct fabrication on plastic sub-strates and transfer from glass to plastic substrates.
With regard to direct fabrication on plasticsubstrates, direct deposition, ELA and SLS havebeen actively studied. Cheng and Wagner [314]monolithically integrated p-channel and n-channelTFTs of nanocrystalline silicon on plastic substratesby direct deposition at a substrate temperature of150 1C to obtain a high electron field-effect mobility
Fig. 22. A schematic flow of the transfer process, from (a) to (d), to tra
substrate [323]. Reproduced from Asano, Kinoshita, and Otani by per
of �30 cm2/V s and a usable hole mobility of�0.35 cm2/V s. Gosain and Usui [317] createdpoly-silicon TFTs on plastic substrate at a substratetemperature of 110 1C, except for the tempera-ture rise during excimer laser irradiation forcrystallization providing a field-effect mobility of250 cm2/V s and sub-threshold swing of 0.16V/decade. Young et al. [320] also used this tech-nique on a variety of polymer substrates includingPI, PAR, polynorbonene (PNB), and PES. Kim etal. [318] employed a two-shot sequential lateralsolidification (TS-SLS) technique to obtain highthroughput for mass production. In this case, thefield-effect mobility and threshold voltage were181 cm2/V s and 1.6V, respectively. A thin-beamcrystallization method [321] and ion beam deposi-tion followed by eximer laser crystallization at roomtemperature [322] have also been reported.
Among the transfer technologies investigated,Akihiko et al. fabricated a bottom gate TFT devicelayer using an ELA process on a glass substrate andtransferred this TFT layer onto a plastic substrate[323]. Fig. 22 shows a schematic flow of the transferprocess. First, the stopper layer is deposited againstHF etching on a conventional glass substrate. Thedevice layer is formed through the LTPS processand glued to the second substrate with a removable,non-water-soluble glue (Fig. 22(a)). The glasssubstrate is etched off in hydrofluoric acid (HF) atroom temperature (Fig. 22(b)). The etching stopperis also removed in another kind of etching solution.The transparent polymer substrate is stuck onto theback surface of the device layer with a permanentadhesive (Fig. 22(c)) and then the second substrateis detached (Fig. 22(d)). The TFT layer on the
nsfer a thin film device layer from a glass substrate onto a plastic
mission of Society for Information Display, California, USA.
ARTICLE IN PRESSM.-C. Choi et al. / Prog. Polym. Sci. 33 (2008) 581–630 611
plastic substrate did not do any significant damageto the device layer and the changes in LTPS TFTcharacteristics with substrate bending revealed amobility change of 16% under 0.75% compression.Inoue et al. [324] developed a new technology thatenables the transfer of thin-film devices from anoriginal substrate to another substrate using laserirradiation, which is called surface-free technologyby layer annealing (SUFTLA). A polycrystallinesilicon TFT backplane for LCDs with integrateddrivers was fabricated using a low-temperatureprocess below 425 1C and successfully transferredfrom a glass substrate to a plastic film using thistechnology. They fabricated an all-plastic substrateTFT-LCD having a display area of 0.7 in measureddiagonally and a pixel count of 428� 238. Takechiet al. [325] reported on a very high rate and uniformglass etching with HF/HCl spray for transferringTFT arrays from a glass substrate to a flexiblesubstrate. Using HF/HCl spray etching, theyachieved both high etch rates of over 20mm/minand satisfactory etch-rate uniformity over a 150mmarea with an approximately 5% variation.
6.3. Organic thin-film transistors
There has been a lot of interest in the develop-ment of OTFTs in which organic semiconductorsare used as active layers because the organicsemiconductors are deposited at a low temperatureor processed through low-cost solution processessuch as spin coating or ink jet printing. In addition,they are compatible with polymer flexible sub-strates, compared to amorphous silicon processesor low-temperature polycrystalline silicon processes[326,327].
6.3.1. Organic semiconductors
As active materials of OTFTs, pentacene, andthiophene oligomers deposited by vacuum evapora-tion have been used to elucidate the transportmechanism in the active layer of organic semicon-ductors. In this case, field-effect mobility wasknown to depend on the deposition condition ofthe active layer and its morphology [328]. Trapsin the bulk, interface, and grain boundary of theactive layer affect the current–voltage characteris-tics of single crystalline and polycrystalline OTFTs[329,330]. The mobility in polycrystalline oligothio-phene increases linearly with grain size, and thetemperature dependence of the mobility changesdrastically from small grains, where the mobility is
thermally activated, to large grains, where themobility is practically temperature independent[331]. Fig. 23 shows potential OTFT materials thathave been studied.
Among small molecular semiconductors, penta-cene demonstrates good performance. Under opti-mized conditions, pentacene TFTs had mobilitiesranging from 0.5 to �3 cm2/V s, which is close tosingle-crystal mobility [12]. When a cross-linkedpolyvinylphenol-based copolymer was spin-coatedas a dielectric layer, good carrier mobility as large as3 cm2/V s was reported [332]. OTFTs with a surface-modified alumina dielectric layer showed highmobility of 2 cm2/V s [333]. For alkyl substitutedoligothiophenes, the length of the a,a0-substitutedalkyl side chains had a notable influence on the TFTperformance while the best performance wasobtained for molecules with relatively short alkylchains (2–6 carbons), which had good carriermobilities as large as 1.1 cm2/V s [334]. Functiona-lized indolo[3,2-b]carbazole was reported to havemobilities up to 0.12 cm2/V s and a current on–offratio of 107 [335], and a mobility of �8 cm2/V s wasreported for rubrene [336].
Devices resulting from solution processes gener-ally tend to exhibit poorer performance than thosemade with thermally evaporated organic smallmolecular semiconductors, due to their lowercrystallinity. As a semiconductor based on smallmolecules, the conversion of pentacene precursorsthrough heat treatment at 200 1C afforded a highlyordered active layer with a good mobility of0.89 cm2/V s [337]. Anthradithiophenes that arefunctionalized with triethylsilyl groups adopt a two-dimensional p-stacking arrangement and exhibitmobilities as high as 1 cm2/V s [338]. Dithiophene-tetrathiafulvalene, which forms single crystals dri-ven by p–p stacking with S?S interactions, alsoshowed high mobility of 1.4 cm2/V s [339]. Twosymmetrical a,o-substituted sexithiophene deriva-tives containing thermally removable solubilizinggroups showed overall mobilities as high as0.07 cm2/V s with on/off ratios of 108 [340]. Forpolymer semiconductors, especially regioregularpolythiophene, a higher molecular weight as wellas more strongly p–p interacting building blocks[341] yielded higher mobilities compared withpolymers of lower molecular weight. This is in spiteof the higher degree of crystallinity of the latter andis due to the fact that charge carriers can travelfarther along longer chains before they haveto hop to another chain [342]. A polythiophene
ARTICLE IN PRESS
N
N
C8H17
C8H17
SS
S
S
SS
SS
SS
SS
O
OC6H13
C4H9
O
OC6H13
C4H9
S S
SS **
C8H17 C8H17
nS
S S
SFF
FF
F F
FF
NN
O
O
O
O
C8H17C8H17S
N
N
S SS
CF3F3C
S S
O
SSO
C6F13
O
C6F13
S SS SNC
NCCN
NC CN
CN
C6H13
C6H13
N
N
N
N**
O O
n
SN
O
O
N F
FF
F
N
F
F
FF
NF
FF
F
N
F
F
FF
Cu N
N
N
N
SS
**
Fig. 23. Examples of organic semiconducting materials for TFTs: (a) solution processible pentacene precursor [337], (b) indolo carbazoles
semiconductor incorporating thieno[2,3-b]thio-phene, which is stable under ambient conditions,was introduced as a solution processable polymer
having a mobility of 0.15 cm2/V s [343]. The carriermobility is higher in oriented films along therubbing direction than in isotropic films and can
ARTICLE IN PRESSM.-C. Choi et al. / Prog. Polym. Sci. 33 (2008) 581–630 613
be as high as 0.02 cm2/V s at room temperature[344]. Poly(3-hexylthiophene) transistors fabricatedby spin coating from high boiling point solvents tofacilitate the growth of a highly crystalline filmexhibited a mobility of 0.12 cm2/V s [345].
N-type organic semiconductors have recentlydrawn increased attention because they can be usedas n-channel materials of complementary circuits,which are profitable with low power consumption[346]. To get a higher performance from n-typeOTFTs, n-channel semiconductors need to havehigh electron affinities with electron withdrawinggroups and also high mobilities, similar to those ofp-type semiconductors, as well as stability againstatmospheric oxidants such as O2 and H2O [347].Metallophthalocyanines with strong electron with-drawing groups were used in an n-channel activelayer that exhibited a mobility of 0.03 cm2/V s [348].Electron-deficient perfluoroarene substitution forelectron-rich thiophene rings lowered LUMO en-ergies while preserving rod-like molecular architec-tures, achieving n- and p-type activities and ahigh n-type activity with a mobility of 0.08 cm2/V s[349]. Dianhydride or diimide moieties containingnaphthalene and perylene derivatives have beenactively studied because of their relatively largeelectron affinities and p-stack structure andp-orbital interactions. OTFTs based on N-alkylperylene diimides achieved very high saturationelectron mobilities as high as 1.7 cm2/V s by care-fully adjusting film growth conditions and linearmobilities of 0.3–0.6 cm2/V s [350]. Yamashita et al.developed a new p-electron system with trifluor-omethylphenyl groups showing n-type performancewith electron mobilities of 0.12–0.30 cm2/V s [351]and thiazole oligomers with trifluoromethylphenylgroups showing very high mobilities of 1.83 cm2/V s[352]. Quaterthiophenes functionalized with carbo-nyl groups as electron withdrawing groups [353] andorigothiophenes capped with tricyanovinyl groups[354] have also been reported as high mobilityn-channel semiconductors and ambipolar transportsemiconductors. Babel and Jenekhe introducedconjugated ladder polymers as n-channel activelayers through a solution spin-coating process [355].The field-effect mobility of the electrons yieldedmobilities as high as 0.1 cm2/V s, which is compar-able with p-type polymer semiconductors.
Semiconductors including n- and p-type materialshave seen big improvements in performance in thepast few years. The adoption of complementarystructures that incorporate both p- and n-type
transistors for OTFTs presents many advantages,including reduced power dissipation and improvednoise margins compared to p- or n-type transistors.Organic complementary transistors also have great-er robustness in addition to better processabilityand compatibility with flexible substrates comparedto inorganic complementary transistors. Organiccomplementary circuits using hexadecafluorocopperphthalocyanine (F-CuPc) as the n-type semiconduc-tor and a-sexithiophene (a-6T) as the p-typesemiconductor have been introduced [346]. Solu-tion-processed ambipolar organic field-effect tran-sistors, which transport both holes and electrons,were fabricated using polymers based on interpene-trating networks as well as narrow band gap organicsemiconductors [356].
6.3.2. Gate dielectric materials
Recently, there has been much interest in the useof new dielectric materials as gate insulators toboost OTFT performance because improvements inthe charge carrier mobility of organic semiconduc-tor materials have begun to plateau around themobility of amorphous silicon TFT [357]. Highdielectric constant gate insulators generally areknown to accumulate more carriers on the interfaceof insulators and semiconductors are known toexhibit a low operation voltage [358]. However,Veres et al. [359] reported when a low dielectricconstant gate insulator is used for a device,performance is significantly improved with in-creased mobility, reduced threshold, and lowerhysteresis because of reduced energetic disorderand carrier localization at the interface. The surfaceroughness of an inorganic gate insulator is animportant parameter affecting OTFT performance;a rougher gate surface hinders the movement ofcharges because of the roughness of the valleys andresults in smaller pentacene grain size and lowerhole mobilities [360]. The evolution of pentacenethin films as revealed by photoelectron emissionmicroscopy showed the growth mechanisms anddynamics of pentacene thin films on the inorganicgate layer [361]. Fig. 24 shows several examples ofgate dielectric materials for OTFTs.
Inorganic gate insulators that have high dielectricconstant values compared to organic insulatorswere deposited by anodization to create anodizedultra-thin metal oxides with extraordinarily lowleakage, high breakdown field strength, and low-operating voltage [362]. The procedure to produce apolymeric smoothing layer resulted in an order of
magnitude improvement in charge mobility relativeto the rough dielectric [363]. The chemical treatmentthat creates the silicon oxide surface through SAMsof organic trichlorosilanes led to a higher mobility,a 20-fold improvement over the mobility on baresilicon oxide, in a polyfluorene copolymer [364]. Theelectric dipoles of SAMs with fluorines and aminogroups affected the accumulation of holes andelectrons in the transistor channel [365]. Phospho-nate-linked SAMs have greatly improved electricalproperties showing a near-zero threshold voltage[366]. Dielectric interface chemistry was reportedto affect n-type performance via interface elec-tron trapping [367]. Modification of the dielectric
interfaces by surface termination of SiO2 using1,1,1,3,3,3-hexamethyl-disilazane (HMDS) [368] andpoly(methyl silsesquioxane) (PMSSQ) [369] also ledto improved performance.
In addition to their superior compatibility withflexible substrates, polymeric insulators can producesmoother surfaces and allow more freedom tofabricate both top and bottom gate devices andinexpensive large-area applications by solutionprocesses compared to inorganic insulators. Severalpolymer gate insulators including PI [370], BCB[371,372], poly(vinyl phenol) (PVP) [373], poly(vinylalcohol) (PVA) [374], silsesquioxane polymers [375],and parylene polymers [376] have been investigated
ARTICLE IN PRESSM.-C. Choi et al. / Prog. Polym. Sci. 33 (2008) 581–630 615
previously. The operating voltage and powerdissipation of organic devices were dramaticallyreduced by exploiting the self-assembly of silane-based molecular dielectrics with a thickness of2.5 nm, providing operating voltages of 2V or less[377]. Surface-initiated ring-opening methathesispolymerization was used to form thin-film poly-mer-dielectric layers with thicknesses ofo100 nm to42 mm [378]. Marks et al. reported that blending ofcommercially available polymers and organosilanecross-linking agents affords robust, smooth, adher-ent, pinhole-free, high-capacitance, low-leakageultra-thin gate dielectric materials [379]. Gatedielectric multilayers containing a UV-cured PVP-poly(methyl methacrylate) (PMMA) layer and aPMSSQ layer showed improvements in OTFTperformance with mobility improved by as muchas 50 times to 0.15 cm2/V s and an on/off ratio of 106
[380]. Maliakal et al. [381] used a novel core-shell-nanoparticle based gate dielectric with titaniumoxide as the core material and polystyrene as theflexible shell to achieve 3.6-fold enhancement in thedielectric constant and mobilities approaching0.2 cm2/V s.
The solution processabilities of electrodes, semi-conductors, and dielectrics for OTFTs have pro-pelled the current surge of research with theirprobabilities of low-cost and large-area manufac-turing coupled with their outstanding features ofbeing physically compact, lightweight, and flexible.Sirringhaus et al. [382] developed the direct inkjetprinting of all polymer transistor circuits, includingvia-hole interconnections based on solution-pro-cessed polymer conductors, insulators, and self-organizing semiconductors. They also fabricated apolymer-based transistor device with a channellength of 500 nm by surface-energy-assisted inkjetprinting [383]. Halik et al. [384] developed a processfor the fabrication of fully patterned all-organicpentacene TFTs on flexible polymer substratesusing PEDOT/PSS for the gate electrode and forthe source and drain contacts and cross-linked PVPas the gate dielectric layer. All jet-printed polymerTFT [385] and a self-aligned self-assembly process[386] have also been reported.
6.4. Others
Transparent oxide semiconductors have recentlybeen proposed as active materials for TFTs whiletransparent electronic devices formed on flexiblesubstrates are expected to meet emerging technolo-
gical demands where silicon-based electronics can-not provide a solution [387]. Nomura et al. [388]proposed a novel semiconducting material, atransparent amorphous oxide semiconductor fromthe In–Ga–Zn–O (IGZO) system, for the activechannel in transparent TFTs. The a-IGZO isdeposited on PET at room temperature and exhibitsHall effect mobilities exceeding 10 cm2/V s, whichare an order of magnitude larger than those ofhydrogenated amorphous silicon, a low leak currentof �10�10 A, and an on/off ratio of 103 even duringand after bending.
CNTs are being actively studied because of theirexcellent electronic properties as new promisingactive materials for transparent flexible TFTs. Javeyet al. [389] showed that contacting semiconductingSWNTs via palladium with a high work functionand good wetting interactions with the nanotubesgreatly reduces the barriers for transport throughthe valence band of the nanotubes. With Pdcontacts, the ‘ON’ states of semiconducting nano-tubes can behave like ohmically contacted ballisticmetallic tubes, exhibiting room-temperature con-ductance near the ballistic transport limit. Gruneret al. fabricated transparent and flexible transistorswhere both the bottom gate and the conductingchannel are CNT networks of different densities andparylene N is the gate insulator [108]. Devicemobilities of 1 cm2/V s and on/off ratios of 100were obtained. Takenobu et al. [390] reportedSWNT transparent TFTs based on the solutionprocess using transparent electrodes that exhibit amobility of 0.5 cm2/V s and an on/off current ratioof �104 and that are highly flexible and bendable toa radius of 7.5mm without a significant loss inperformance.
There are several studies on transfer single-crystal silicon TFTs on flexible polymer substrates.Rogers et al. reported a high-yield fabricationstrategy for producing printable single-crystal sili-con ribbons from a bulk silicon wafer and printingthem onto thin plastic substrates showing goodelectrical properties and mechanical flexibility [391].Effective device mobilities, as evaluated in thelinear regime, were as high as 360 cm2/V s, andon/off ratios were 4103, representing importantsteps toward a low-cost approach to large-area,high-performance, mechanically flexible electronicsystems. They also developed large area, selectivetransfer processes of microstructured siliconbased on the printing approach on flexible sub-strates [392].
ARTICLE IN PRESSM.-C. Choi et al. / Prog. Polym. Sci. 33 (2008) 581–630616
7. Encapsulation
In the previous section, a multilayer barrierstructure on plastic substrates was demonstrated,which protects devices, especially OLEDs, frommoisture or oxygen [39]. A single layer of aninorganic compound such as SiO2, Al2O3, SiNx, orMgO can be adequate as a barrier layer for LCDs.A very dense and amorphous single layer also canbe used as an encapsulant for an OLED on glass.But in a flexible display, it may be difficult tomaintain the integrity of a single-layer barrier,especially if it is composed of oxides or nitridesbecause they tend to crack readily under tensilestress. Multilayer barriers present the most robustand forgiving properties in this regard. A defect inone layer does not affect or does not add to a defectin another layer [11].
To completely prevent gas ingression into thedevices, two kinds of gas sealing structure areneeded on the top and bottom sides. This gassealing technology is called encapsulation or passi-vation. To date, most OLED devices have beenencapsulated by sealing the device in an inertatmosphere such as nitrogen or argon using a metalthat can be secured by a bead of UV-cured epoxyresin. And a getter such as calcium oxide or bariumoxide is incorporated into the package to removeresidual water incorporated in the package ordiffusing through the epoxy seal over time [27].However, this kind of encapsulation is impossible toapply to the flexible display system.
There are a couple of ideas considered as nextgeneration encapsulation technologies for flexibledisplays as substitutes for the metal can [27]. One isto use barrier-coated polymer films, which have atransparent or opaque organic/inorganic multilayerstructure, in place of the metal can. This approachhas the advantages of allowing more flexible processconditions as well as robust mechanical properties.Another is a multilayer thin-film encapsulationwhere multilayer thin films are directly depositedon the OLED structures, providing a thinner devicestructure compared to the polymer film encapsula-tion. We have already discussed barrier-coatedpolymer films in the previous section. In this section,thin-film encapsulation will be described. Therequired conditions for thin-film encapsulation ofOLED displays include high transparency fortop emitting OLEDs, low-stress materials, lowprocessing temperature, densely packed andhighly conformal coating, pinhole-free coating and
compatibility with the active components ofOLEDs.
The multilayer barrier structure consists oftransparent organic layers that planarize the sub-strate surface and decouple defects in the oxide aswell as allow for greater flexibility in an otherwiserigid coating plus transparent inorganic metal oxidelayers that act as a practical diffusion barrier.Chwang et al. [393] encapsulated passive matrix,FOLEDs on flexible plastic substrates using anorganic/inorganic multilayer barrier encapsulationtechnology. The displays, based on electropho-sphorescent OLED technology, were deposited onbarrier-coated plastic substrate and hermeticallysealed with an optically transmissive multilayerbarrier coating encapsulation with alternatingAl2O3 and polyacrylate layers. Preliminary lifetimeto half initial luminance of the order of 200 h wasachieved on the passive matrix driven encapsulated80 dpi displays, and a 2500 h lifetime was achievedon a dc tested encapsulated 5mm2 FOLED testpixel. Visser et al. fabricated similar alternatingstacks of polymer and low-temperature inorganicoxide layers with a plasma-protective buffer depos-ited between the OLED device and the barrierstructure via the Vitex system, which demonstratedwater permeation in the range of 1� 10�6 g/m2/day[394]. Van Assche et al. [395] investigated thea-SiNx:H deposition process with only 3 plasmadeposited silicon nitride layers separated by a thinorganic layer showing a water permeation rate ofbelow 10�5 g/m2 per day and OLED lifetimes ofover 500 h at 60 1C and 90% RH.
Lifka et al. [396,397] proposed a novel multilayerstack of silicon nitride–silicon oxide–silicon nitri-de–silicon oxide–silicon nitride (NONON) as thin-film encapsulation. They used a combination ofPECVD silicon nitride and PECVD silicon oxide. Inthis case, the silicon oxide changes the chemicalinterface of a defective area and therefore enablesthe growth of silicon nitride, reducing the number ofpinholes through stacking of these layers. The waterpermeability of a single NON stack deposited at85 1C was measured to be less than 1� 10�6 g/m2/day.Yoshida et al. [398] encapsulated a full-color OLEDdisplay using SiON as a moisture barrier film on thesubstrate and SiN as a passivation film on thedevice. The display was approximately 0.2mmthick, weighted approximately 3 g, and it wasbendable. The estimated half-luminance decay timewas more than 5000 h with an initial luminance of1000 cd/m2, which is equivalent to that of a device
ARTICLE IN PRESSM.-C. Choi et al. / Prog. Polym. Sci. 33 (2008) 581–630 617
on a glass substrate. Akedo et al. [399] developedplasma-CVD SiNx/plasma-polymerized CNx:Hmultilayer films to improve the longevity ofpassivated OLEDs for automobile applications.The films had highly effective barrier againstmoisture even at high temperatures because thethermal stress of the films was released by the softCNx:H layers and no cracks were produced.
Kim et al. [61,62] investigated a variety ofinorganic materials and inorganic composites aspassivation materials on polymer substrates usingthe electron beam evaporation system. They showedthat the MgO thin film had a lower WVTR valuethan any other inorganic thin film, and that theWVTR of inorganic films can be dramaticallyminimized by adopting an inorganic composite asa passivation material.
8. Roll-to-roll (RTR) processes
The current method of producing display panels,circuit boards, and other electronic devices is abatch process using conventional vacuum deposi-tion and lithography pattern technologies on silicon
Fig. 25. Technical challenge
wafers or glass substrates. On the other hand, theRTR process is currently a well-known technologyto the film manufacturer in diverse areas such asnewspapers, labels, etc. [40].
RTR processing offers a significant advantagecompared with the conventional batch process, as itincreases throughput by allowing for greater levelsof automation and by eliminating the overhead timeinvolved in loading and unloading panels intolithographic imaging tools and chemical processingstations. This leads to lower contamination levelsand, thus, higher yields due to the minimal humanhandling that is needed to process the substrates [5].However, there are many challenges includingdisplay backbone manufacturing, where electronicsare fabricated on a plastic substrate with precisionand high yield, and fabrication of TFTs for active-matrix backplanes.
Current research includes two kinds of ap-proaches: the transfer of current processes from glasssubstrates to new RTR processes or the developmentof new technologies available for RTR processes[400]. However, there are many challenges toachieving real RTR processes [401,402].
s for flexible displays.
ARTICLE IN PRESSM.-C. Choi et al. / Prog. Polym. Sci. 33 (2008) 581–630618
9. Conclusions
Flexible displays will be the ultimate choice in thefuture in the display industry because of their manyadvantages including convenience, portability, andlarge-size applications as well as low-cost produc-tion through RTR processes.
However, there are many challenges to surmount,as summarized in Fig. 25. The thermal stability,solvent resistance, thermal expansion coefficients,and gas absorption of polymer substrates need to beimproved so that they are as good as glasssubstrates. OLEDs require almost perfect encapsu-lation against water vapor. Electrodes such asanodes or cathodes should endure repeated stressand be available for web coating. Active materialsof the displays should demonstrate high perfor-mance and stability as well as solvent processibilityapplicable for inkjet processes. With regard toTFTs, low-temperature processes need to be devel-oped and their performance and stability should beimproved to apply for TFTs as well as OLEDs.
With good properties and processiblity, polymershave attracted the attention of many scientists aspotential flexible display materials, including trans-parent substrates, transparent electrodes, and activematerials for OLEDs, LCDs and OTFTs, dielectricmaterials, and coating materials. All polymer-basedflexible displays are even being investigated.
In this article, we have reviewed the recentprogress in polymer materials that are potentialcandidates for use in flexible displays. We believethis review will give insights to readers on what,where, and how polymer materials are used and thechallenges yet to be overcome in flexible display.
Acknowledgments
The work was supported by the Korea Scienceand Engineering Foundation (KOSEF) through theNational Research Laboratory Program funded bythe Ministry of Science and Technology (MOST)(No. M10300000369-06J0000-36910), the SRC/ERC of MOST/KOSEF program (Grant #R11-2000-070-080020), and the Brain Korea 21 Project.
References
[1] Hill P. iSuppli flexible displays report. Veritas et Visus, Flex
Substr 2006;2(5):21–5.
[2] Crawford GP. Flexible flat panel display technology. New
York: Wiley; 2005.
[3] Nathan A, Chalamala BR. Special issue on: flexible
electronics technology. Proc IEEE 2005;93(7–8).
[4] Gasman L. OLED and paper-like display markets. Veritas
et Visus, Flex Substr 2006;2(6):22–5.
[5] Allen KJ. Reel to real: prospects for flexible displays. Proc
IEEE 2005;93:1394–9.
[6] Weber A, Deutschbein S, Plichta A, Habeck A. Thin glass-
polymer systems as flexible substrates for displays. SID Dig
![Laurence W. McKeen, PhD - Pentasil Used in Medical Devices.pdf · of branched polymers include star polymers, comb polymers, brush polymers, dendronized polymers [1], ladders, and](https://static.documents.pub/doc/80x56/5fd30108783da00f76371237/laurence-w-mckeen-phd-pentasil-used-in-medical-devicespdf-of-branched-polymers.jpg)
