`
`INNOLUX CORP. v. PATENT OF SEMICONDUCTOR ENERGY
`
`LABORATORY CO, LTD.
`
`|PR2013-00068
`
`
`
`201. W. L. Taylor,J. Geophys. Res. BB, 3,575—3,583 [1978).
`202. D. E. Proctor.
`it. Uytenhogaordt, and B. M. Meredith, J.
`Geophys. Res. 931), 12,633—12,727 (1988).
`203. W. D. MacClemont and R. C. Murty, J. Appl. Meteor-oi. 1'1,
`786—795 (19731.
`
`204. GHHC User Services, Lightning Detection and Ranging
`{LEAR} Dolores Summary, Global Hydrology and Climate
`Center, Huntsville, Alabama, 2000; hitp:i/ghrc.n1.sfnnoso_
`goo .’ usoh‘eddmefldur.html.
`205- L. M. Meier, E. P. Kridor, and M. W. Meier, Mon. Weather
`Rev. 112, 1,134—1,14U{1934l.
`206. D. J. lioccippio, S. Heelcman. and S. J. Goodman, Free. 11th
`Inter. Confi Ammo. Electra, National Aeronautics and Space
`Administration. CP—1999-209261. Gunteravills, AL, 1999,
`pp. 254—251
`207. C. J. Noumann, J. Applied Meir-20ml. 10, 921—936 (1971).
`208. P. Richard and G. Aulfi‘ny,Raciio Sci. 20, 171-192 (1985).
`209. C. D. Hoyenga. J. Geophys. Res. 89]). 1,403—1,410 (1984).
`210- V. Motor, P. R. Krehbiel, and X. —M. Shoo, J. Geophys. Res.
`100D, 25,731-2fi,753 (1995).
`211. X. -M. Qhao, P. R. Krehhiel, R. J. Thomas, and W. Risen, J.
`Geophys. Res. icon, 2349—2383 (1995).
`212. J. W. Warwick, G. D. Hayenga. and J. W. Brosnshan, J.
`Geophys. Res. MC, EAST-3,433 [1979).
`213. C. Rhodes and P. R. Krehbiel, Geophys. Res. Lett. 16,
`1.169—1,172(1989l.
`214. E. Defer, C. Then-y, P. Blanchet, and P. Lsroche, Proc. 11th
`Int. Con)“. Armor. Elli-cm, Netional Aeronautics and Space
`Administration. (JP-1999409261, Guntersville. AL, 1999,
`pp. 14-— 17.
`215. X. -M. Shoo, D. N. Holden, and G. T. Rhodes, Geoplrys. Res.
`Lett. 23. 1,917—1,920(1996l.
`216. A. Galvan, V. Cooray, T. Gotechl, and V. Books, Free. 11th.
`Int. Confi (limos. Electra, National Aeronautics and Space
`Administration, (31751993309261, Guntersville, AL, 1999,
`pp. 162—165.
`'1‘. .l. 'I‘oomi, Proc. Hih Int. Conf. Aimee. Eileen, National
`Aeronautics and Space Administration, CHESS-209261,
`Guntorsville, AL, 1999. pp. 196-199.
`213. S. Souls, G. Molinié, S. Cheney, and N. Simond, Free. 11th.
`Int. (loaf. Atoms. Electm National Aeronautics and Space
`Administration. CP—1599-209261, Gunter-aville, AL, 1999,
`pp. 3344437.
`
`219. Y. Yair, 0.Altoratr., and Z. Levin, Prue. Nah. Int. Confi
`Amos. Electra, National Aeronautics and Space Administra—
`tion, CPleQQ-ZDQEISL, Guntersvillc, AL, 1999, pp. 460—463.
`
`LIQUID CRYSTAL DISPLAY TECHNOLOGY
`
`Gaseous P. Coawrono
`MICHAEL J. Escort
`Brown University
`Providence, RI
`
`INTRODUCTION
`
`As the display in most imaging systems is the final medium
`through which an image is rendered for manipulation
`and verification, an understanding of display technologies
`is essential to the imaging process. Because individuals
`
`LIQUID CRYSTAL DISPLAY TECHNOLOGY
`
`955
`
`working in the field of imaging science and technology
`may spend more time looking at a display screen than at
`anything else in their office or laboratory. it is imperative
`that it be comfortable to use and appropriate for the
`particular context.
`Twenty years ago, system manufacturers often inte-
`grated the electronic display directly into the system
`to provide a complete package for a specified imaging
`application. Although this approach does not afford much
`flexibility, the display could be closely matched to a ape-
`cific application because user requirements were well
`defined. This custom design approach enabled optimiz-
`ing the graphics controller, system software, and user
`interface for the display, user, and application require-
`ments. Despite the positive attributes of this “black-box”
`approach, such as high performance and superior applica-
`tion specific image quality. closed architecture platforms
`tend to be more expensive and suffer from incompatibi-
`lity with peripheral add-one and software packages not
`supported by the system manufacturer.
`Today.
`the situation is dramatically different due
`to the continual evolution of the graphics controller
`interface. By mixing images with text and graphics,
`software developers require more from the display to
`support moving images without diminishing display per-
`formance for static images. The graphical capability of
`today's standard computer platforms has now made it
`unprofitable for vendors of imaging systems to develop
`their own displays for system-specific teal-:5. End some
`now typically 'purchase a computer platform, display.
`and a variety of other peripherals fi-om multiple ven-
`dors and integrate them with ease (in, a plagued—
`play philosophy). In such a marketplace, one must be
`well educated to match display tachnology to appli-
`cation needs. This article provides the reader with a
`fundamental knowledge of working principles of liquid
`crystal displays (LCDBl. their capabilities, and their limi-
`tations.
`
`ADDRESSING DISPLAYS
`
`is
`it
`Before We delve into the operation of a LCD,
`important to understand how those displays are addressed
`and their impact on resolution,
`refresh rates, and
`image fidelity. Many treatises begin with material and
`device configurations, but we will first develop a basic
`understanding of electrical addressing schemes
`that
`apply to all LCDs. Our hope is that the reader will
`he better prepared to recognize the capabilities and
`limitations of the various display configurations presented
`afterward.
`
`A LCD with high-information content (c.g., computer
`or television screen) consists of a two-dimensional array of
`pixels, where a pixel is defined as the smallest switching
`element of the array. If the two-dimensional array has
`a total of N rows and M columns (N KM pixels), then
`in principle, there can be N x M electrical connections to
`control each pixel independently. This is known as direct
`addressing and is practical only for very low-resolution
`displays. For medium and higher resolution displays,
`
`
`
`LIQUID CRYSTAL DISPLAY TECHNOLOGY
`
`addressing is accomplished through pussiue— and active
`matrix techniques. Both of' these approaches require only
`N + M electrical connections, thereby greatly simplify the
`electronics, and make higher resolution possible.
`Luminance —voltage plots for three hypothetical dis-
`plays are depicted in Fig. 1. This display characteristic
`ultimately dictates the type of addressing that can be
`used to create images by using a LCD. Luminance is the
`physical measure of brightness of a display or any surface
`and most commonly has units of candelas per squared
`meter (cd/rn2), nits, or footlambcrts (fL). The two mea-
`surable quantities from the luminance—voltage curve that
`have the greatest impact on display addressing are the
`threshold voltage VT" (the voltage at which the luminance
`begins to increase) and a parameter A (the additional volt-
`age beyond VT“ needed to cause the display to approach or
`reach its highest luminance). if a liquid crystal (LC) mate-
`rial does not start to respond to an electronic stimulus until
`it has reached a well-defined voltage, then it is said to have
`a threshold; otherwise, if the display material responds to
`all voltages, then it is said to be thresholdless (1).
`For simple direct addressing schemes, like the seven-
`segment digit electrodes shown in Fig. 2,
`the three-
`holdllcss) nature of the material is irrelevant because
`the segmented electrodes (or pixels) are independent of
`each other. The appropriate combinations of segments are
`addressed by dedicated logic circuitry (is, every pixel is
`independently driven by its own external voltage source),
`and the screen refresh rate is only as long as needed for a
`single pixel to switch. Direct addressing is practical only
`for low-resolution displays (:59 pixels).
`In a passive addressing scheme (also known as
`multiplexing), one substrate has row electrodes, and the
`other substrate has column electrodes, as shown in Fig. 3.
`
`
`
`Threshold
`
`Luminance
`
`Voltage
`Figure 1. Luminance-voltage graph depicting the difference
`between materials that have well-defined thresholds and those
`materials that have no threshold (thresholdless). VTH is defined
`as the voltage at which the luminance bsgina to increase, and
`a parameter A is defined as the additional voltage beyond VTJ-I
`needed to cause the display to approach or reach its highest
`luminance.
`
` Conducting
`
`
`segment
`Figure 2. An example of direct addressing, where the segments
`are independently driven to create low-resolution information.
`
`Every pixel is uniquely determined by the region oi'ovcrlap
`of a row and a column electrode. making it possible
`to access N x M pixels with only N +M connections. A
`display that 1133 a passive matrix is driven one line
`at a time; that is, one row is selected for addressing,
`and all of the columns are addressed using voltages
`associated with the image for that row. At some time
`interval later, the next row receives an appropriate voltage
`pulse, and the columns again are addressed using the
`information required for that row. The net result is
`that a pixel
`is influenced only sufficiently to produce
`an optical cfi'ect when the time-averaged voltage [called
`the root-mean-square (rmsll across the row and column
`electrodes is beyond the threshold (VON ;_. Var"). All of the
`rows not being updated are driven by row voltages that
`will not affect the image information already present.
`The catch is that there will always be a voltage on
`the non selected rows, and therefore that voltage must
`satisfy the relationship Von 5 va so as not to induce
`an optical change. Therefore, a well-defined threshold in
`the luminance —voltage characteristic of the LCD material
`is necessary to prevent
`the non selected rows from
`being addressed (i.e., cross talk). LC materials typically
`exhibit threshold behavior, which many displays to be
`multiplexed; display materials that are thresholdless, such
`as electrophorctics (2) and gyricon (3), cannot.
`An expression can be derived that completely specifies
`the maximum number of rows NM that can be addressed
`in terms of the voltage threshold and the parameter A, a
`measure of the nonlinearity oftho LC material:
`
`
`1
`a .1
`VTJI _ «Jill MAXI
`
`(1)
`
`To maximize the number of addressable rows for higher
`resolution, one usually uses a material that has a very
`non-linear and steep luminance—voltage response, small
`A, rather than using materials that have a large Vm
`(Which increases power consumption}.
`Another important equation relates the rms voltages
`of the DN—state VQN’, and the OFFastate, Vow,
`to the
`maximum number of rows that is given by (For Now; >> 1):
`
`
`
`(2)
`
`
`
`LIQUID CRYSTAL DISPLAY TECHNOLOGY
`
`957
`
`Passive mulliplexing: Amplitude modulation
`
`
`
`
`
`Column signals
`
`Pixel voltage
`(row—column)
`
`Figure 3. A simple example of multi-
`plexing on an N-row matrix showing
`the row addressing waveforms. column
`addressing waveforms, and the cone;
`spending pixel waveforms. This type
`of multiplexing, lmown as amplitude
`modulation, varies the amplitude ofthe
`pixel waveform to achieve the desired
`image. The normally black pixel
`is
`"ON" if the rms average pixel voltage
`exceeds VT“. See color insert.
`
`where Vow/Vow is referred to as the selection ratio.
`When Nmax is large, this ratio approaches unity, leading
`to poor contrast because the difference would decrease
`between the luminance in the ON- and OFF-states.
`Therefore, Ell- l2l dictates the optical contrast ratio (CR)
`of the display, which is defined as the ratio of the
`luminance in the (IN-state and the luminance in the
`OFF-state (CR =LoN/Low). Equations (1) and (2) were
`first derived by Alt and Pleshlro in 1974 and remain the
`fundamental expressions governing multiplexing (4). To
`increase resolution using passive addressing techniques,
`a technique known a: “dual” scan can be implemented.
`In the deflector approach, two column drivers are used
`to address the upper N/2 rows and lower N12 rows.
`Examination of Eq. (1) shows that the dual-scan approach
`improves the ratio A/V-m by a factor of Mi (5).
`Although some LC configurations that use the mul—
`tiplexing technique have reasonable contrast for large
`values of N, other limitations enter into consideration
`such as response time (the time taken for a pixel to be
`switched to the fully CIN-state plus the time needed to
`relax to a completely OFF—state), viewing angle (the maxi-
`mum polar angle for which a display maintains reasonable
`contrast and minimal color degradation), and gray—scale
`issues (those that involve the operation of a pixel at a
`luminance which is intermediate between the ON- and
`OFFustates). The frame rate in a passively addressed dis-
`play is severely limited by the time it takes for a single
`row to switch, as well as the number ofrows in the display,
`because the frame rate must be greater or equal to the
`product of the two. In only a few exceptions, displays that
`use passive addressing will not support full video frame
`rates. However. this problem can be essentially elimi-
`nated using an active-matrix addressing scheme, where a
`nonlinear element is uscd for pixel isolation.
`Active-matrix addressing is recognized by the display
`industry as the ultimate solution For high fidelity, high
`
`full color, and significant gray—
`information content,
`scale applications. Additionally, this addressing approach
`can be used with thresholdless and large a materials
`because a discrete nonlinear switch is integrated into
`each pixel structure. An active-matrix LCD incorporates
`a two-dimensional circuit array (or matrix] to provide the
`electrical addressing of individual pixels (6). This matrix
`incorporates an active device in each pixel, usually a thin-
`film-trsnsistor (TFTJ, positioned at one of the corners.
`Almost all LCD pixels are essentially dielectric capacitors
`whose leakage is minimal when a charge is placed on
`the electrodes through the transistor. And due to the
`electrical isolation afforded by the transistor, the voltage
`on one pixel remains constant while other pixel elements
`are subsequently addressed; therefore, the Alt—Pleslflto
`limitation expressed by Eq. (2) does not constrain the
`contrast ratio, as it does in passive—addressing schemes.
`A schematic diagram of an active-matrix circuit is
`shown in Fig. 4, where each pixel element is defined by the
`overlap of row and column bus lines. The circuit diagram
`shows that each pixel has one TFT and a LG capacitor
`formed between a top conducting surface, (typically indium
`tin oxide (ITOll. and the active-matrix substrate. This
`approach is significantly more complex than passive
`addressing, as can be seen from the cross section of
`an active-matrix TFT display also shown in Fig. 4. The
`intricate underpinnings of activewmatrix addressing and
`the complex processing required to create one are beyond
`our scope, but the basic operation can be understood as
`follows. The display is addressed one line at a time, as
`in passive addressing. When a row (called the scan or
`gate line) is addressed, a positive voltage pulse of duration
`T/N (where N is again the number of rows and T is
`the frame time) is applied to the line and turns on all
`transistors in the row. The transistors act as switches
`that transfer electrical charges to the LG cells from the
`columns (called data or source lines). When subsequent
`
`
`
`n' tar-SI
`
`SIN,
`
`
`ITO electrode i
`'-
`
`
`Passivalion
`
`layer
`
`Source
`
`
`
`Black matrix
`
`Color filter
`
`Planarization
`layer
`
`Polymer
`allgnment
`layer
`
`Polarlzer
`
`LIQUID CRYSTAL DISPLAY TECHNOLOGY
`
`
`
`
`
`Glass
`
`'i‘illl.‘||'|i-‘I|Il lllll'llll lilll'illll l
`
`|Il|
`
`lli
`
`ll‘l'lll llll llllllilll lliiili'llilllll‘ltllllll— Pularizar
`
`Column driver
`
`
`
`Gate
`Insulating
`layer
`
`Figure 4. The driving circuit Of an active—matrix thin—libs transistor next to the cross section
`of an active—matrix pixel. As is apparent from the pixel structure becomes substantially more
`complex when active—matrix schemes are used. See color insert.
`
`capacitor
`electrode
`
`rows are addressed, a negative voltage is applied to the
`gate line that turns off all of the transistors in the row
`and holds the electrical charges in the LC capacitors for
`one frame until the line is addressed again. Alternating
`row-select voltages are required for most LCD materials,
`and the polarity of the data voltage is usually switched
`in alternate frames. The refresh rate in this addressing
`scheme is not limited by the number of rows but only by
`the response time of the LC.
`A nonlinear pixel element can be implemented by
`a variety of‘ approaches, but we will mention only the
`two most prominent active-matrix types; amorphous
`silicon (at-Si) and polycrystalline silicon (poly-Si). Both
`of these involve complex fabrication of thin films of
`silicon structures (<300 not) on glass or quartz substrates.
`The details of active-matrix derelopment are sufficiently
`complex to be far beyond the scope of this review, but
`we include a simple introduction to familiarize the reader
`with the basics.
`Pioneering work on the active matrix began in the
`early 1970!! ('7), and the now conventional cr‘Si approach
`was first proposed in 19'i‘9(3), where the deposited
`thin film of silicon has a small grain structure and is
`randomly configured (9). In this approach.
`inexpensive
`glass substrates are typically processed at ~3OD“C,
`using well established processes,
`including sputtering,
`photollthography (typically 5—8 photomasks), plasma—
`enhanced chemical vapor deposition (PECVD), wet
`chemical etch, and reactive ion etching (RED. The
`resulting electron mobility is low (~05 cmgN/s), but
`is more than adequate to form a useful switch
`the custodial. However,
`there are two prominent
`limitations. First.
`the aperture ratio {the transmissive
`area of the pixel divided by the total pixel area) of the
`display is decreased to 50—75% due to the use of an
`absorbing light shield to overcome the photouonductivity
`of the tie-Si TFT. Second, because the supporting row-
`and column~drivers must be mounted on additional 10
`chips bonded to the display substrate, the connections
`
`(34, 000) increase fabrication complexity and decrease
`reliability. Nonetheless, oraSi actiVs-matrix displays have
`been made in all sizes and are most popular as laptop
`monitors and other large-area LCDs. Current research
`issues include reducing the number of photomasks (10),
`increasing display resolution (ll), and increasing the
`aperture ratio (12).
`In contrast, a poly—Si substrate can yield TFTs Whose
`electron mobility is much higher (~44D orniN/s), but must
`be processed at significantly higher temperatures (6]. Fab-
`rication typically involves the same (It-Si process described
`before on a more expensive quartz substrate. Additional
`processing at the higher temperatures leads to recrys-
`tallization in a furnace or by laser annealing (13,14).
`An intriguing approach uses a laser ablation/annealing
`approach where CMOS-TFI‘S are fabricated at high tem-
`peratures using a quartz substrate and are subsequently
`transferred to flexible plastic substrates without any
`noticeable deterioration in polyaSi performance (15). In
`all approaches, the silicon grain becomes larger and more
`uniform and allows electrons to flow much more freely. The
`greatest benefits of these poly—Si substrates are the ability
`to fabricate row- and column-drivers directly on the periph—
`cry ofthe glass substrates and the reduction ofthe TFT size
`to N5 x 5 pm. Additionally, the aperture ratio can he made
`substantially higher. Disadvantages include the high pro-
`cess temperature, increased fabrication complexity (higher
`accuracy required in photolithography and ion implanta-
`tion), and the higher off‘dealtage current. Integration of
`driver electronics onto the substrate, increased display
`brightness, lower power consumption, and the ability to
`form smaller pixels at higher densities (>200 dpi) make
`these poly-Si displays particularly useful for n'iicrodisplays
`and medium size displays.
`Display addressing directly impacts resolution and
`optical performance. Because most imaging applications
`require high resolution, active—matrix addressing is the
`most prominent addressing approach in the imaging field.
`Because of the complexity of the substrate, activeumatrix
`
`
`
`addressing always inwlvcs more cost, but it enables
`a high-resolution regime. A rule of thumb for LCD
`technology is that passive addressing can achieve NMM 5
`400, whereas active addressing can achieve NM“; 3 1, 000‘.
`
`PROPERTIES OF LIQUID CRYSTAL MATERIALS
`
`To understand the nuts and bolts of a LCD, it is worth
`while to review briefly the material properties that make
`LCs functional for display applications. The liquid crystal
`phase is a state of matter that is intermediate between a
`solid crystal and an isotropic liquid; in fact, it has props
`crticc of both. The geometry of a LG molecule is highly
`anisotropic; these molecules can be four to six times longer
`than they are wide and are often modeled as rigid rods (16).
`The nematic phase is the simplest LC phase and is the
`most used in commercial display applications. This phase
`possesses only orientational order along the long axes of
`the elongated molecules, but no positional or bond orien—
`tational order, as in a conVcntional crystal. We discuss the
`ferroslectric smcctic 0“ liquid crystal phase later, so we
`will limit our discussion here to the nematic phase,
`The elastic properties of LCs are their most character»
`istic feature. At the display level, elastic theory is used
`to predict stable configurations and electric field-induced
`elastic deformations of the material that are responsi-
`ble for the image. The elastic theory expression is often
`written in the following form:
`
`f = iIKmv .nr +Kcslfl - v x n)2
`
`+K53[n x v x n)“ — endeared}.
`
`(3)
`
`Here. f is the free energy density; n is the nematic
`director; E is the applied electric field; K11, K32, and
`Km; are known as the splay,
`twist, and bend elastic
`constants, respectively; and As is the dielectric anisotropy.
`The nematic director is denoted as n and represents the
`average orientational symmetry axis of an ensemble of LC
`molecules. The elastic constants are typically of the order
`of 10“1 N, the dielectric anisotropy As is typically ~5—15
`for most display matcrials, and E is typically {1 V/pm
`for conventional transmissive LCD technology. Display
`performance parameters can be predicted by minimizing
`Eq. (3) with respect to the boundary conditions imposed
`by the surfaces of the display substrates to obtain tho
`configuration that has the lowest free energy. A number
`of mathematical representations can be used to calculate
`the device properties of LCDs (17). Equation (3) is used
`to predict the details of the field-aligned configuration of
`the LC material and its threshold voltage. The switching
`dynamics can be determined by balancing the elastic and
`electric field torques derived from Eq. (3) and the viscous
`torque oftbc material—y; are felt), where y; is the rotational
`viscosity. The rotational viscoaity y] is typically from 1—2
`poise for display materials. The switching time is highly
`dependent on the choice of LC material and configuration
`but is typically in the 1—30 ms range. The reader is
`referred to the following references to explore bacic LC
`properties further l' 18 —30).
`
`LIQUID CRYSTAL DISPLAY TECHNOLOGY
`
`959
`
`The techniques involved in transforming the configua
`rations derived from elastic theory and into the optical
`properties of LCDs have been treated extensively and will
`only be summarized here. In addition to their geometrical
`anisotropy, nematic LCs used in displays also possess
`optical anisotropy, or birefringence, that is responsible for
`the optical changes in contrast when one configuration is
`electrically switched to another configuration. The birc-
`fringcncc on. for LCs is in the range ofD.05—0.3; on. w 0.1
`is often used for display applications because it leads to a
`high contrast design known as the first minimum or maxi-
`mum (see later). The 2 x 2 Jones matrix technique (31)
`can be used to calculate the transmission oflight through
`a LCD at normal incidence but neglects Fresnel diffrac~
`tion and multiple reflection effects in thin layers. The
`more sophisticated 4 x 4 (32) method represents a com-
`plete solution of Maxwell's equations and is used for the
`general case of light transmission through a LCD at any
`angle. Nematic LC configurations pertinent to displays are
`most often modeled by this latter method by the display
`community (33).
`
`LIQUID CRYSTAL DISPLAY CONFIGURATIONS
`
`Having presented the basics ofdisplay addressing and the
`basic properties of nematic LCs, we can now introduce the
`Various configurations relevant to displays. It would be
`impossible to describe, even briefly, every type of current
`LCD configuration, so we will discuss the operational
`principles of a few that are most relevant to imaging.
`Broadly speaking, LCDs can be classified as transmissive
`or reflective. A transmissive display is a light shutter
`that modulates a powerful baclrlight. Reflective displays
`take advantage of ambient lighting to reflect light back
`to the viewer effectively and therefore do not require a
`hacklight. Transficctiva displays operate in both modes.
`Removing the bacldight can result in significant power
`saving and therefore longer battery life. These are largely
`being developed for portable applications rather than
`high—resolution imaging. Therefore, we will
`focus on
`transmissive displays and end by presenting a promising
`low—power reflective display technology.
`
`Twislcd Ncmalic (TN) Liquid Crystal Displays
`
`The most commercially successful LCD configuration,
`known as the twisted nematic (TN), uses cross-polarizcrs
`and a molecular orientation of the molecules whose long
`axis twists through a 90° angle between the orienting
`substrates. A crossasection of a TN—LCD, presented in
`Fig. 6, includes two glass substrates that have polarizcrs
`laminated on the outer surfaces and. an ITO conduct-
`ing layer on the inner surfaces covered by a polymer
`(c.g., polyimide). The polyimidc layer is mechanically
`rubbed with a cloth to create microgrooves on the sur-
`face that uniformly align the long axis or the LC molecules
`at the surface (34). The alignment of the rub direction of
`the two substrates is placed parallel to the transmission
`axis of the respective laminated polarizer, but perpendic-
`ular to each other. The surface alignment mechanisms
`introduced by the polyimide then result in a LG layer that
`
`
`
`LIQUID CRYSTAL DISPLAY TECHNOLOGY
`
`
`
`Polarizer
`
`Color filior
`mosalo
`
`.fl/M’.
`gamer
`
`Conducting
`layer (ITO)
`
`
`
`
`
`meoluiion
`
`
`Polymer
`layer
`
`Polarizer
`
`Baokllghl
`unpolarlzed
`llghi
`
`substrale
`
`
`
`Active matrix
`[or hlgh
`
`Figure 5. The Operation of the twisted nematic (TN) configuration, which is currently the most
`common LCD mode. The center illustration shows the optical. stack of all of the components of the
`TN display cell. The adjacent insets show the field-on and field-on" states of the nonnslly white
`(NW) configuration. Note that the cross-section is not drawn to scale. See color insert.
`
`has a 90° twist sandwiched between crossed polariaers. A
`51m: cell gap is used for most TN displays. In the figure,
`the extraordinary eitis (long axial of the LC molecules at
`the surfaces is parallel to the transmission axes of the
`polarizers (called the e-rnode). However, the o-mocle, When
`the ordinary axis is parallel instead, is also used. These
`two configurations load to slightly difi'erent transmission
`and viewing-angle characteristics (35). Finally, placing a
`red, green, and blue color filter array between the top sub-
`strate and ITO layer creates a full color display, known as
`spatial color synthesis.
`The TN principle of operation is as follows. After the
`first poiarizer, linearly polarized light will follow the LC
`twist and waveguide to an orthogonal linear polarization
`state. This process, sometimes called adiabatic following,
`enables light to escape through the second polariaer. This
`is valid if the twist angle (in this case n12) is much
`smaller than the retardation of the nematic (Brood/l),
`known as the Mauguin condition. Using common LC
`values, birefringence Art 2 0.10, coll thickness d = 5 are,
`and wavelength 3. = 1510 run, this ratio is 1: ML If this
`limit is not satisfied, then the light exiting the LC layer
`is elliptically polarized, and the output. from the top
`poleriser will be reduced. This configuration is called
`normally white (NW), and describes the zero—voltage state
`as the transmissive one. This is most common. in today’s
`technology. The polarizors can also be arranged in a
`parallel configuration, so that the zero-voltage state is
`black and the voltage-on state is the transmissive one.
`
`The transmission ofa NW-mode TN-LCD can be derived
`by using the Jones formalism (35):
`
`.
`
`2
`
`If
`
`r—!-
`
`T=£_lw
`
`2
`
`2
`
`1+u5"
`
`'
`
`(4)
`
`Here. it = Edema/A, and it is assumed that the director in at
`the display substrates is parallel to the transmission axis
`ofthe polai'izors. The 1/2 term indicates that the maximum
`transmission, as shown in Fig. 6, can be ideally 50% of the
`input light due to the loss in the first polarizer. Note that
`the maximum transmission occurs when the argument
`
`Transmissioni'it}
`
`14334'1'55fi5‘7
`
`a
`
`11
`
`Figure 6. The transmission of the NW-TN cell in the lleld-ofi'
`state as a function of the Mauguin parameter,
`
`
`
`of sinl‘ll'. . .l is an integral multiple of x, corresponding to
`n = fl. «1’33. J55", etc. The first maximum that occurs at
`J13 corresponds to the maximum brightness and contrast
`ratio. Optimizing around other maximums is also done
`(at M175, the Second maximum, for example), but this is
`not nearly as common. Nematic domains can sometimes
`appearin the TN mode, which degrade the contrast but can
`be overcome by using a polyirnidc alignment layer which
`induces an out-of-plane pretilt (typically in the range of
`2— 10").
`When a voltage is applied to the pixel, an electric field
`is created perpendicular to the substrates. Because the
`dielectric anisotropy of the LC is positive (As 2:- 0), the
`molecules tend to align parallel to the field. Equation (5)
`can be used to predict the actual threshold voltage of the
`twisted nematic configuration, which is the point where
`the molecules just begin to realign. The threshold voltage
`for 3 TN display is given by the following expression:
`
`Vol = a
`
`'_
`
`K1]_ [1 4— (Kl—3%)]
`suns
`4Kn
`
`112
`
`(5)
`
`Using typical values Kn = 10‘” N, K23 = 5.4 x 10“2 N,
`K33 = 15.9 x M"12 N, and As = 10.5, then Vri-i = 1.1 V, a
`common threshold for most nematic mixtures. Note that
`this threshold is the voltage where the LC starts to align
`due to the applied voltage and does not say anything about.
`the director profile. Above VT“, the broken symmetry ofthe
`twist due to the out-of-plane reorientation of the nematie
`molecules align perpendicular to the substrates and light
`passes through teli LC layer without any change in
`polarization. In the NW configuration, the output polarizer
`then absorbs the light. Grayscale levels are attainable
`when intermediate voltages are used. In summary, the
`TN—LCD therefore simply modulates the intensity of a
`powerful backlight by acting on the polarization state of
`the incident light.
`Notice that the display shown in Fig. 5 is addressed
`by an active matr'iX, Which is common because the
`luminance—voltage curve for the twisted nematic LCD
`is not very steep and is not conducive to multiplexing
`schemes (36,37); therefore an active matrix is exclusively
`used to address the twisted nematic for high-end, high-
`rcsolution imaging. For example, if we consider Eq.(1)
`and substitute typical values for 9. TN material (A = 0.7
`and V'rn = 1.8 V), the maximum number of rows that can
`be addressed is approximately six; therefore only very
`low resolution displays of this mode are possible using
`multiplexing schemes.
`The switching time (typically in the range of 10-30 ms)
`of the TN configuration is proportional to the viscosity
`and the square of the cell gap and therefore is very
`sensitive to the cell gap. Although thinner cells enable
`faster switching,
`they often compromise the Mauguin
`condition, which can reduce brightness and the contrast
`ratio. The art of display design consists, in large part, of
`balancing these parameters for maximum benefits.
`
`Superlwisted Nematic {STN} Liquid Crystal Displays
`
`While the TN-TF’I‘ display configuration has become the
`standard in applications that require high pixel densities
`
`LIQUID CRYSTAL DlSFLAY TECHNOLOGY
`
`951
`
`the supertwisted nematie
`and gray~scalc resolution,
`(STN) configuration (38.39) has been very successful
`in the low to medium resolution realm. The reasons
`
`for this primarily surround the ability of this display
`mode to be multiplexed (passivewmatrix addressing).
`As discussed earlier, an N xiii-pixel display requires
`N +M electrical connections using this passive scheme
`(the some number as for active-matrix addressing) but
`does not
`require a TFT matrix. And even though
`there are some optical performance trade-oil's involving