电子工程师手册（英文版五）：ch083

Morris, J.E., Martin, A., Weber, L.F. “Displays” The Electrical Engineering Handbook Ed. Richard C. Dorf Boca Raton: CRC Press LLC, 2000 83 Displays 83.1 Light-Emitting Diodes Semiconductor Device Principles?Semiconductor Materials?Device Efficiency?Interfacing 83.2 Liquid-Crystal Displays Principle of Operation?Interfacing 83.3 The Cathode Ray Tube Monochrome CRTs?Color CRTs?Contrast and Brightness?Measurements on CRTs?Projection Screen 83.4 Color Plasma Displays Introduction?Color Plasma Display Markets?Color Plasma Display Attributes?Gas Discharge Physics?Current Limiting for Plasma Displays?ac Plasma Displays?Color Plasma Display Devices?Gray Scale 83.1 Light-Emitting Diodes James E. Morris The light-emitting diode (LED) has found a multitude of roles as the field of optoelectronics has bloomed. Infrared devices are used in conjunction with spectrally matched phototransistors in optoisolation couplers, hand-held remote controllers, interruptive, reflective and fiber-optic sensing techniques, etc. Visible spectrum applications include simple status indicators and dynamic power level bar graphs on a stereo or tape deck. This section will concentrate on digital display applications of visible output devices. Semiconductor Device Principles The operation of an LED is based on the recombination of electrons and holes in a semiconductor. As an electron carrier in the conduction band recombines with a hole in the valence band, it loses energy DE equal to the bandgap E g with the emission of a photon of frequency u = c/l = DE/h (83.1) where l is the radiation wavelength and h is Planck’s constant. The incidence of recombination under equilibrium conditions is insufficient for practical applications but can be enhanced by increasing the minority carrier density. In an LED, this is accomplished by forward biasing the diode, the injected minority carriers recombining with the majority carriers within a few diffusion lengths of the junction edge. Figure 83.1 illustrates the process. The potential barrier eV o is reduced by forward bias eV, leading to net forward current and the minority carrier distributions shown on either side of the depletion layer. As the carriers diffuse away from the junction edges, these distributions decay exponentially because of recombination with the majority carriers. Each recombination event shown on either side of the junction gives off a photon. This process is called injection electroluminescence. James E. Morris State University of New York at Binghamton André Martin Hughes Display Products Larry F. Weber Plasmaco, subsidiary of Matsushita ? 2000 by CRC Press LLC Equation (83.1) implies that the radiation emitted will be monochromatic, but in practice DE > E g , and there is a spectral distribution corresponding to the energy distributions of the carriers in the conduction and valence bands. Semiconductor Materials Silicon is the most common material used in current semiconductor technologies, but it is not at all suitable for an LED. The reason is that silicon has an indirect bandgap, and a direct bandgap is required for process efficiency. Direct and indirect bandgaps are compared in Fig. 83.2, where carrier energy is plotted versus momentum for both cases. The photon momentum p = hl = hu/c (83.2) FIGURE 83.1 Light emission due to radiative recombination of injected carriers in a forward-biased pn junction. FIGURE 83.2 (a) Interband recombination in a direct-bandgap semiconductor; (b) recombination in an indirect-gap semiconductor also involves a momentum change. VB CB E E g eV n p p n x np n pO p nO E Fn E Fp Hole Injection Electron Injection Light Out Holes Depletion layer Minority- Carrier Density Light Out recombination e (V–V O ) p no e p no e -x/L e -x/L h recombination Light Output Indirect Transition Energy Change Momentum Change Injected Electron Direct Recombination with Hole VB CB E E Hole -h1/a h1/a0 (a) p VB Hole -h1/a h1/a0 (a) p ? 2000 by CRC Press LLC (where c is the velocity of light) is very small, and conservation of momentum can be readily accommodated by small deviations from the vertical transition shown in Fig. 83.2(a). For the indirect case illustrated in Fig. 83.2(b), the energy change DE defines the photon energy and momentum, again according to Eqs. (83.1) and (83.2), but conservation of momentum additionally requires that the much greater electron momentum on the order of h/2a be accounted for. For lattice dimensions, a, on the order of 10 –10 m and wavelengths, l, on the order of 10 –6 m, it is clearly not possible for both conservation criteria to be met without the participation of a third body, i.e., a phonon. The two consequences of this result are that the indirect transition is inefficient (in that it must transfer momentum and hence thermal energy to the lattice) and less likely to occur than the direct transition (because of the requirement for all three particles to simultaneously meet the energy and momentum conditions). Indirect bandgaps therefore lead to long diffusion lengths and recombination times, which produce good transistors but poor LEDs. The most common direct-bandgap semiconductor is GaAs, but the photon wavelength calculated for E g = E D = 1.43 eV as listed in Fig. 83.3(b) is in the infrared. Such a material may be ideal for communications and sensory optoelectronic applications but is unsuitable for display purposes. The bandgap may be adjusted, however, by the substitution of phosphorus for arsenic in the lattice as shown in Fig. 83.3(a). The color range listed corresponds to the range of LED colors commonly available: red, yellow, and green. The direct and indirect bandgaps, E D and E I , of GaAs 1–x P x vary with x as E D = 1.441 + 1.091x + 0.210x 2 (83.3) and E I = 1.977 + 0.144x + 0.211x 2 (83.4) [Wang, 1989], enabling one to design the material to produce the required LED color. FIGURE 83.3 (a) Plot of momentum versus bandgap energy, and (b) corresponding semiconductor parameters for various compounds of the GaAs/GaP system; (c) plot of momentum versus bandgap energy for indirect GaP materials showing special trapping levels. (Source: S. Gage et al., Optoelectronics/Fiber-Optics Applications Manual, 2nd ed., New York: Hewlett- Packard/McGraw-Hill, 1981, pp. 1.3–4. With permission.) ? 2000 by CRC Press LLC Note the continuous transition from the direct GaAs to the indirect GaP. The materials have an indirect bandgap for x > 0.4 and have the same problems as light emitters as silicon. The efficiency of an indirect-gap emitter can be greatly enhanced by the introduction of appropriate impurity recombination centers, as shown in Fig. 83.3(c). In the process shown, an injected minority carrier electron (in p-type material) is first trapped by the localized impurity (which is itself electrically neutral but which introduces a local potential to the lattice which attracts electrons). The center is then negatively charged and attracts a hole to complete the recombination process, which produces the photon. The recombination center solves the momentum transfer problem, because the trapped electron is localized to the impurity lattice site and has a momentum range according to the Heisenberg Uncertainty Principle of Dp ~ h/2pa (83.5) that is, sufficient to include the processes shown in the diagram at p ~ 0. In the cases used as examples, a nitrogen atom substitutes for a phosphorus, or a zinc–oxygen pair substitutes for adjacent gallium–phosphorus atoms in the GaAs 1–x P x lattice. The GaAs 1-x P x system is well established, but can only produce wavelengths defined by the range of energy gap widths, i.e., down to green. Blue LEDs require higher band-gap materials: (a) SiC technology is well developed for high temperature semiconductor applications, but it has an indirect band gap, so its emission efficiency is very poor [Pierret, 1996]. (b) GaN (and In/Al GaN alloys) is a direct band gap material system producing successful blue and blue- green devices [Jiles, 1994; Nakamura, 1995; Pierret, 1996]. (c) II-IV compounds such as ZnS and ZnSe possess direct band gaps in the 1.5–3.6eV range, offering the possibility of full spectrum LEDs within the single materials system [Jiles, 1994]. Device Efficiency In considering LED efficiencies, it is convenient to consider the emission process to consist of three distinct steps: (a) excitation, (b) recombination, and (c) extraction. These will be discussed with reference to Fig. 83.4. (a) Photons created by minority electron recombination on the p-type side of the junction are more likely to be successfully emitted from the surface of the device, for the structure shown in Fig. 83.4(a) and (b) if the p-type region is a thin surface layer. For a given total LED current, I, made up of electron, hole, and space- charge region recombination components, I n , I p , and I r , respectively, the electron injection efficiency (which provides the excitation) is g n = I n /(I n + I p + I r ) (83.6) In principle, all the physical processes described above apply equally to both electrons and holes. However, the electron mobility, m n , is greater than that of a hole, m p , and since I n /I p = N d m n /N a m p (83.7) (where N d , N a are n-type donor and p-type acceptor doping densities, respectively) greater g n is attainable for a given doping ratio than hole injection efficiency, g p . Consequently, LEDs are usually p-n + diodes constructed as in Fig. 83.4, with the p-layer at the surface. (b) Some of the recombinations undergone by the excess electron distribution, Dn, in the p-type region will lead to radiation of the photon desired, but others will not, because of the existence of doping and various impurity levels in the bandgap. The total recombination rate, R, can be written in terms of the radiative and nonradiative rates, R r and R nr , as R = R r + R nr (83.8) where R r = Dn/t r , R nr = Dn/t nr , R = Dn/t (83.9) ? 2000 by CRC Press LLC and where t r and t nr are the minority carrier lifetimes associated with the radiative and nonradiative recombi- nation processes, and t is the effective lifetime. The radiative efficiency is defined as h = R r /(R r + R nr ) = t/t r (83.10) and the internal quantum efficiency is h i = hg (83.11) (c) It is clear from Fig. 83.4 that many of the photons generated on either side of the junction will pass through sufficient bulk semiconductor to be reabsorbed. In fact the photon energy may be ideally suited to reabsorption if it exceeds the semiconductor direct bandgap. It is obvious, then, why GaAs is opaque and GaP transparent to photons from Ga(As:P) junctions. Clearly, a greater efficiency might be expected from the transparent substrate with reflecting contact [Fig. 83.4(b)]. The photon must strike the LED surface at an angle less than the critical angle for total internal reflection, q c , where sin q c = n ext /n LED = 1/n (83.12) and n ext , n LED are the external and internal refractive indices, respectively. For air, n ext = 1, but critical angle loss can be reduced by encapsulating the device in an epoxy lens cap [Fig. 83.4(c)] to increase both n ext > 1 and the angle of incidence at the air interface. Even within angles less than q c , there is Fresnel loss, with transmission ratio T = 4n/(1 + n) 2 (83.13) The total external quantum efficiency is then the fraction of photons emitted [Neamen, 1992], given by [Yang, 1988] h e = 1/(1 + av o /AT) (83.14) FIGURE 83.4Effect of (a) opaque substrate, (b) transparent substrate, and (c) encapsulation on photons emitted at the pn junction. A Absorbed Photons Graded Alloy GaAs 1-y P y (y=0 0.4) Graded Alloy GaAs 1-y P y Reflective Contact Emitted Photons Emitted Photons Predominantly Electron Injection Transparent Plastic Encapsulation Al Top Contact Light Output Insulating Layer Al Back Contact GaAs GaAs P GaAs 1-y P y Ga P B n p n n p (a) (b) (c) q c p ? 2000 by CRC Press LLC where a is the average absorption coefficient, v o is the LED volume, and A is the emitting area. In considering LED effectiveness for display purposes, one must also include radiation wavelength in relation to the spectral response of the human eye [Sze, 1985]. Although the GaP green LED is intrinsically less efficient than the GaAsP red LED, the eye compensates for the deficiency with a greater sensitivity to green. More recently developed heterojunction LEDs (Fig. 83.5) offer two mechanisms to improve LED efficiencies [Yang, 1988]. The electron injection efficiency can be enhanced, but, in addition, absorption losses through the wider 2.1-eV bandgap n-type layer are essentially eliminated for photons emitted by recombination in the lower 2.0-eV bandgap p-type region. Interfacing In circuit design applications, the LED may be treated much as a regular diode, but with a much greater forward voltage, V F . Since one usually seeks maximum brightness from the device, it is usually conducting heavily and V F approaches the contact potential. As one moves from GaAs to GaP [Fig. 83.3(a)], V F varies from about 1.5 to around 2.0 V. The variation in V F with temperature (at constant current) follows similar rules as apply to conventional diodes, but radiant power and wavelengths also change [Gage et al., 1981]. Single LEDs are commonly driven by logic gates, perhaps as status indicators, and some of the simplest interface circuits are shown in Fig. 83.6. In many cases, the gate output will not be able to source or sink sufficient current for visibility, and an amplifier will be required, as in Fig. 83.7. Bar graph displays are commonly FIGURE 83.5A GaAlAs heterojunction LED: (a) cross-sectional diagram; (b) energy-band diagram. FIGURE 83.6Digital logic can interface directly to LED lamps. (Source: S. Gage et al., Optoelectronics/Fiber-Optics Appli- cations Manual, 2nd ed., New York: Hewlett-Packard/McGraw-Hill, 1981, p. 2.20. With permission.) N GaAl 0.7 As 0.3 P GaAl 0.6 As 0.4 N GaAl 0.7 As P GaAl 0.6 As 0.4 p GaAs P GaAs Contact Contact 2.1 eV 2.0 eV 1.42 eV E ? ? 2000 by CRC Press LLC used to indicate signal level on audio equipment, with a modification of the position indicator seen in Fig. 83.8 to guide fine tuning. Matrix LED arrays can be used for flexible, high-density panel displays [Fig. 83.9(a)] and are conventionally controlled by row or column strobing [Fig. 83.9(b)] controlled by a microprocessor interface. Multiple LEDs are commonly packaged together in a single integrated device, organized in one of the standard display fonts [Fig. 83.10(a)], with decoding often included within the package [Fig. 83.10(b)]. The 7-segment FIGURE 83.7LED interfacing for (a) low-power transistor-transistor logic, (b) logic high drive, and (c) CMOS. (Source: M. Forbes and B.B. Brey, Digital Electronics, Indianapolis: Bobbs-Merrill, 1985, p. 242. With permission.) FIGURE 83.8Operational amplifiers or voltage comparators used to decode an analog signal into a bar graph or position indicator display. (Source: S. Gage et al., Optoelectronics/Fiber-Optics Applications Manual, 2nd ed., New York: Hewlett- Packard/McGraw-Hill, 1981, p. 23.3. With permission.) ? 2000 by CRC Press LLC display is adequate for hexadecimal applications, but the 16-segment is required for alphanumerics. To limit pin-out requirements, the LEDs of a single package are connected in either the common anode or common cathode configuration [Fig. 83.11(a)], with multiple display digits multiplexed as illustrated in Fig. 83.11(b). FIGURE 83.9Matrix displays. (a) One LED will be turned on by applying the proper signal to one x axis and one y axis. (b) Character generation using column strobe methods. (Source: S. Gage et al., Optoelectronics/Fiber-Optics Applications Manual, 2nd ed., New York: Hewlett-Packard/McGraw-Hill, 1981, pp. 2.25, 5.44. With permission.) FIGURE 83.10(a) Display fonts used in LED displays. (b) Construction features of a hermetic LED display. (Source: S. Gage et al., Optoelectronics/Fiber-Optics Applications Manual, 2nd ed., New York: Hewlett-Packard/McGraw-Hill, 1981, pp. 5.3, 5.6. With permission.) ? 2000 by CRC Press LLC Defining Terms External quantum efficiency: The proportion of the photons emitted from the pn junction that escape the device structure (but sometimes alternatively defined as h i h e ). Injection electroluminescence: Electroluminescence is the general term for optical emission resulting from the passage of electric current; injection electroluminescence refers to the case where the mechanism involves the injection of carriers across a pn junction. Internal quantum efficiency: The product of injection efficiency and radiative efficiency corresponds to the ratio of power radiated from the junction to electrical power supplied. Related Topic 22.1 Physical Properties References J. Allison, Electronic Engineering Semiconductors and Devices, 2nd ed., London: McGraw-Hill, 1990. M. Forbes and B. B. Brey, Digital Electronics, Indianapolis: Bobbs-Merrill, 1990. S. Gage, D. Evans, M. Hodapp, H. Sorensen, R. Jamison, and R. Krause, Optoelectronics/Fiber-Optics Applications Manual, 2nd ed., New York: Hewlett-Packard/McGraw-Hill, 1981. D. Jiles, Introduction to the Electronic Properties of Materials, London: Chapman & Hall, 1994. S. Nakamura, “A bright future for blue/green LEDs,” IEEE Circuits & Devices, 11(3), 19–23, 1995. D. A. Neamen, Semiconductor Physics and Devices: Basic Principles, Boston: Irwin, 1992. R. F. Pierret, Semiconductor Device Fundamentals, New York: Addison-Wesley, 1996. S. M. Sze, Semiconductor Devices: Physics and Technology, New York: Wiley, 1985. S. Wang, Fundamentals of Semiconductor Theory and Device Physics, Englewood Cliffs, N.J.: Prentice-Hall, 1989. E. S. Yang, Microelectronic Devices, New York: McGraw-Hill, 1988. FIGURE 83.11 (a) Generalized drive circuits for strobed operation. (b) Block diagram of a strobed (multiplexed) six-digit LED display. (Source: S. Gage et al., Optoelectronics/Fiber-Optics Applications Manual, 2nd ed., New York: Hewlett-Pack- ard/McGraw-Hill, 1981, pp. 5.25, 5.23. With permission.) ? 2000 by CRC Press LLC Further Information More extensive semiconductor device treatments of the LED are contained in Semiconductor Devices and Integrated Circuits by A. G. Milnes (Van Nostrand Reinhold, New York) and in Introduction to Optical Electronics by K. A. Jones (Harper and Row, New York). E. Uiga provides more interfacing and design detail for the LED as a circuit element in optoelectronics (Prentice-Hall, Englewood Cliffs, NJ, 1995). Wang [1989] considers second-order effects extensively. In Semiconductor Optoelectronics by J. Singh (McGraw-Hill, New York, 1996), the emphasis is on communications applications, but the temperature dependence and frequency response issues covered there are also relevant to displays. Chapter 2 of Gage et al. [1981] contains detailed information on the optical and thermal design constraints on the LED package and on LED back-lit display systems. Chapter 6 considers filtering and other techniques for the contrast enhancement required for direct sunlight viewing. Professional society magazines are good sources of up-to-date information at the non-specialist level, espe- cially the occasional special issues devoted to topic reviews. IEEE Spectrum is a good example, as is the IEEE Circuits & Devices magazine. 83.2 Liquid-Crystal Displays James E. Morris In a low-power CMOS digital system, the dissipation of a light-emitting diode (LED) or other comparable display technology can dominate the total system’s power requirements. In such circumstances the low-power dissipation advantage of CMOS technology can be completely lost. This is the situation in which liquid-crystal display (LCD) technology must be used. The LED (or other active system, such as a plasma or vacuum fluorescent display) emits optical power supplied (comparatively inefficiently) by the system battery or other source. The passive LCD is fundamentally different in that the optical power is supplied externally (by sunlight or room lighting typically) and the system source need supply only the relatively minute amount of power (microwatts per square centimeter) required to change the device’s reflective optical properties. Principle of Operation Materials classed as liquid crystals are typically liquid at high temperatures and solid at low temperatures, but in the intermediate temperature range they display characteristics of both. Although there are many different types of liquid crystals used, we will concentrate here on the use of nematic crystals in twisted nematic devices, the most common by far. The essential feature of a liquid crystal is the long rod-like molecule. In a nematic crystal, the molecules align as shown in Fig. 83.12. If the container surface is microscopically grooved, the interface molecules will be aligned by the grooves and intermolecular forces will maintain that orientation across the liquid crystal [Fig. 83.12(a)]. The molecules will align in an electric field, and beyond a critical value, the field may be sufficient to overcome the alignment with the grooves [Fig. 83.12(b)]. (In practice, the transition is not so abrupt, and groove alignment persists at the interface itself [Fig. 83.13].) The process of alignment in the electric field is the result of the anisotropic dielectric constant characteristic of liquid crystals. For the electric field parallel to the molecular alignment, e r = e || , and for a perpendicular field, e r = e ^ . In a “positive” liquid crystal, e || > e ^ , and the molecules align parallel to the field as described above in order to minimize the system’s potential energy. The principle of the twisted nematic cell is illustrated in Fig. 83.14. The confining plates, typically 10 mm apart, are grooved orthogonally, forcing the molecular orientation to spiral through 90 degrees [Fig. 83.14(a)]. In the LCD, two polarizers and a mirror are added as shown in Fig. 83.14(b). Incident ambient light is polarized and enters the liquid-crystal cell with the plane of polarization parallel to the molecular orientation. As the light traverses the cell, the plane of polarization is rotated by the twist in the liquid crystal, so that it reaches the opposite face with a polarization 90 degrees to the original direction, but parallel now to the direction of the second polarizer, through which it may therefore pass. The light is then reflected from the mirror and passes back through the cell, reversing the prior sequence. ? 2000 by CRC Press LLC When an electric field (greater than the critical field) is applied between the transparent electrodes, usually conductive indium–tin oxide (ITO) thin films, the 90-degree twist in the crystal is destroyed as the molecules align parallel to the field, so that the rotation of the light’s plane of polarization cannot be sustained. Conse- quently, the crossed polarizers effectively block reflection of the incident light from the backing mirror, and the surface appears to be dark, with excellent contrast to the light gray color of the device in the reflecting mode. The contrast ratio can be further enhanced by the use of the super twisted nematic crystal, where the molecular orientation is rotated through 270 degrees rather than 90 degrees. Transmission LCDs function very similarly to the devices just described, but without the mirror, which is replaced by a powered backlighting source. Obviously, the low-power advantage of the passive device is lost in FIGURE 83.12Liquid-crystal/grooved interface: (a) with no field applied, and (b) with an electric field e > a critical value. FIGURE 83.13Diagram of the orientation of the liquid-crystal axis in a cell (a) with no applied field, (b) with about twice the critical field, and (c) with several times the critical field. Note slight permanent tilt (a 0 ) and turn (b 0 ) at the surfaces. (Source: G. Baur, in The Physics and Chemistry of Liquid Crystal Devices, G.J. Sprokel, Ed., New York: Plenum, 1980, p. 62. With permission.) (a) (b) Liquid crystal molecules align parallel to grooves and to each other Solid container, surface microscopically grooved Liquid crystal Solid ? 2000 by CRC Press LLC this active alternative, but monochromatic backlighting does provide one means of constructing displays with varied background colors. Another form of color display is provided by cholesteric crystals. The three main types of liquid crystals, nematic, cholesteric, and smectic, are distinguished by the different types of molecular ordering they display. In the cholesteric crystal, the direction of molecular alignment rotates in each successive parallel plane (Fig. 83.15). The spatial period of the rotation, p, is called the pitch, and Bragg reflections occur when the wavelength of incident light meets the condition l = p/n (83.15) FIGURE 83.14(a) Twisted nematic cell, e = 0. (b) Liquid-crystal display element. (b) (a) Light Polarization Plane of Polarization Mirror Polarizer No. 2 Polarizer No. 1 Plate B , grooved in y direction, perpendicular to A Glass plate A, Microscopically grooved in x direction ITO Counter- Electrode Glass Plate, B Glass Plate, A Grooves Grooves Transparent (ITO) Top Electrode 10 μm Liquid Crystal Incident Light Plane of Polarization Twisted Liquid Crystal Light Polarization x y ? 2000 by CRC Press LLC where n is an integer. The liquid crystal can thus appear to be colored in incident white light. In practice, the color is strongly temperature dependent and the effect is more appropriate to temperature-sensing applications than to digital displays. In practice, the color pixels of a large area display will be organized in the traditional television RGB format, with the colors defined by external filters or internal dyes [Braithwaite and Weaver, 1995]. The dye molecules align with the LCD molecules, and absorb correctly polarized light. There is current interest in the development of liquid-crystal color switches where an electrical control signal would be able to change the device color, whether from white to monochromatic or continuously through the spectrum. Uiga [1995] discusses some LCD problems, in particular the slow response times and the variation of effective critical voltages with limited viewing angles, and the temperature dependences of both. Interfacing LCDs can be organized in all the ways available to competing technologies, e.g., LEDs (see Section 83.1), including seven-segment, alphanumeric, and dot matrix. The LCD differs from LED displays, where each pixel or segment must be a separate device, because the LCD segment or pixel areas are defined by transparent electrodes separated from a common overlapping backplane by a single liquid crystal [Fig. 83.16(a)]. In a large matrix array, it may take a significant period to scan all pixels, and the simple addressing scheme of Fig. 83.16(b) may lead to noticeable flicker. The high off resistance of the MOSFETs of Fig. 83.16(c) can reduce this problem by increasing the discharge time to hold the LCD on after the address pulse has gone. The MOSFETs in this active matrix technology are implemented in practice in the form of polysilicon or hydrogenated amorphous silicon (a-Si:H) thin film transistors (TFT) [Shur, 1990; Braithwaite and Weaver, 1990]. The interfacing requirements, which are otherwise similar in multiplexing techniques, etc., are complicated by the requirement for zero net dc bias across the cell in order to avoid electrochemical degradation of the material. LCDs require ac drive signals, and square waves of frequency between 25 Hz and 1 kHz are typically used [Wilson and Hawkes, 1989]. A square wave is applied to the backplane, with in-phase and antiphase signals to the counter electrode determining whether the given pixel or segment is on or off. In practice, the state is determined by the root-mean-square (rms) value of the differential voltage applied. Figure 83.18 illustrates the additional complexity that would be required by even a simple multiplexed addressing system. The backplane and segment drivers might correspond to rows and columns of a dot matrix, as implied in the diagram, or the backplanes may identify specific characters of an alphanumeric display. Calculating the rms values of the difference voltages shown gives 0.42 V tc for the on pixels and 0.24 V tc for off, from which V tc can be calculated for reliable operation if the critical voltage is known for the LCD to be used. FIGURE 83.15Cholesteric ordering: a large number of planes of nematic ordering are formed where the directors rotate as we move along a direction perpendicular to the planes. (Source: J. Wilson and J.F.B. Hawkes, Optoelectronics: An Intro- duction, London: Prentice-Hall, 1989, p. 145. With permission.) ? 2000 by CRC Press LLC Defining Terms Active matrix: Each pixel in a high density display matrix, such as for flat-screen television, requires its own active (switching element) driver (e.g., a TFT). Cholesteric: In the cholesteric liquid crystal, successive layers of aligned molecules are rotated naturally. Indium–tin oxide (ITO): A mixture of the semiconducting oxides SnO 2 and In 2 O 3 ; the most common transparent conductor. FIGURE 83.16 LCD addressing: (a) simple (seven-segment) addressing; (b) matrix addressing; and (c) matrix addressing with MOSFETs. FIGURE 83.17 Drive signals from a direct connect LCD driver. (Source: R. Lutz, Application Note 350, in Interface Databook, Santa Clara, Calif.: National Semiconductor Corporation, 1990, p. 4–109. With permission.) a 1 b c d d e f g 2 3 3 4 (b) (c) (a) 5 6 7 Top Contacts Top Contacts Common Bottom Contact Bottom Contacts S D G ? 2000 by CRC Press LLC Nematic: The type of liquid crystal in which the molecular chains align; such alignment can be controlled across the liquid crystal if it can be constrained at the boundaries. Twisted nematic: The alignments of the nematic planes are rotated through 90 degrees across the crystal by constraining alignments to be orthogonal at the boundaries. Related Topics 22.1 Physical Properties ? 83.1 Light-Emitting Diodes References J. Allison, Electronic Engineering Semiconductors and Devices, 2nd ed., London: McGraw-Hill, 1990. G. Baur, “Optical characteristics of liquid crystal displays,” in The Physics and Chemistry of Liquid Crystal Devices, G. J. Sprokel, Ed., New York: Plenum, 1980. N. Braithwaite and G. Weaver, Eds., Electronic Materials, Milton Keynes: The Open University/Butterworths, 1990. R. Lutz, “Designing an LCD dot matrix display interface, application note 350,” in Interface Databook, Santa Clara, Calif.: National Semiconductor Corporation, 1990. M. Shur, Physics of Semiconductor Devices, Englewood Cliffs, N.J.: Prentice-Hall, 1990. E. Uiga, Optoelectronics, Englewood Cliffs, N.J.: Prentice-Hall, 1995. J. Wilson and J. F. B. Hawkes, Optoelectronics: An Introduction, London: Prentice-Hall, 1989. Further Information Nematic liquid-crystal molecules typically incorporate two separated benzene rings in a complex chain molecule [Wilson and Hawkes, 1989]. The organic chemistry of liquid-crystal compounds will lie outside the interests of most readers but is briefly reviewed in “Liquid Crystal Materials for Display Devices,” by J. A. Castellano and K. J. Harrison in The Physics and Chemistry of Liquid Crystal Devices, edited by G. J. Sprokel [Plenum, 1980]. One technique used in liquid-crystal color switches requires the use of electrically controlled birefringence. This topic is covered at an elementary level by Wilson and Hawkes [1989]. FIGURE 83.18 Example of backplane and segment patterns. (Source: R. Lutz, Application Note 350, in Inter- face Databook, Santa Clara, Calif.: National Semiconductor Corporation, 1990, p. 4–109. With permission.) ? 2000 by CRC Press LLC An interesting historical perspective on the development of LCD technology is provided by the extensive reviews of 150 patents in the field contained in Liquid Crystal Devices, edited by T. Kallard (State of the Art Review, Vol. 7, Optosonic Press, New York, 1973). The book also contains a bibliography of more than 1100 entries. The various professional societies’ magazines are excellent sources of material for recent developments in this field (and others). These publications regularly devote a special issue to research developments in a single field, at a level intended for the non-specialist. A good example in the LCD area is provided by two articles on TFT silicon for active matrix displays contained in the Materials for Flat-Panel Displays issue of the MRS Bulletin, 21(3), March 1996 (Materials Research Society), which cover the transition from a-Si:H to polysilicon, to the prospects for single crystals. 83.3 The Cathode Ray Tube André Martin The cathode ray tube (CRT) is the element which, in a display, converts an electrical signal into visual information using an electron beam adequately intensity modulated and deflected to impinge on a cathodolu- minescent screen surface, in a glass envelope under vacuum. Because of the extensive growth of electronic communication since the Second World War, information is very often presented on CRTs, mainly when the information content exceeds 100,000 picture elements (pixels). Monochrome CRTs are widely used for computer terminals, radars, oscilloscopes, projection systems, etc., while high resolution color CRTs are preferred for imaging from computers, such as multimedia, CAD-CAM systems and digital image processors [Keller, 1991]. In 1991, the worldwide market for CRT monitors was 37 million units [Stanford Resources, 1992a], of which 20 million units were high-resolution color monitors. These 37 million do not take into account the oscilloscope, radar, projection, and other special-purpose tubes that would add a few hundred thousand to the figures. In 1996, high resolution color CRT monitors in current use are in excess of 200 million units and mroe than 50 million monitors are being produced. The importance of the CRT in the display world can be explained by two key factors: ?The CRT is using a single serial data input to generate a picture. ?The CRT is a very efficient light-emitting display. For example, a typical high-resolution 20-inch diagonal color monitor requires 70 W from the main ac 50/60-Hz power supply and generates a picture visible in any office environment with a luminous efficiency of 6 to 8 lm/W. These two factors, high luminous efficiency and convenience of addressing, make the CRT difficult to replace by any other type of display for large image contents. We will describe first the monochrome CRT and then discuss the color CRTs. Monochrome CRTs General The monochrome CRT [Martin, 1986] is composed of: ?A glass envelope with the necessary glass-to-metal seals for anode and electron gun connections. This envelope is under vacuum (about 10 –7 mm Hg). ?A cathodoluminescent screen deposited on the faceplate, usually aluminized to improve brightness and obtain good screen potential uniformity. ?An electron gun using a hot cathode to emit electrons that are accelerated toward the screen and deflected either by electrostatic plates or by an electromagnetic deflection coil. ?Various outside and inside conductive coatings to normalize the potentials. The Electron Gun A typical electron gun is composed of (Fig. 83.19): ? 2000 by CRC Press LLC ? 2000 by CRC Press LLC CATHODE RAY TUBE Vladimir K. Zworykin Patented December 6, 1938 #2,139,296 n excerpt from Zworykin’s patent application: I claim as my invention: 1. In combustion, a cathode ray image transmitting tube provided with an electrode in the form of a mosaic photosensitive target and with means for developing a ray of electrons and directing the ray at a surface of said target for scanning the same, a lens system for focusing an optical image of an object on the scanning portion of the surface of the target, such portion of said surface being concave toward the lens system with the curvature corresponding substantially to that of an imaginary spherical surface on which the lens system is able to focus the optical image sharply. A spherical (or possibly hyperbolic) photosensitive cathode helped correct for the distortion inherent in aiming the electron beam over a relatively wide dispersion to scan the whole “screen”. The basic principles behind Zworykin’s CRT, first demonstrated in 1929, have been used in hundreds of millions of radar screens, television tubes, and now computer monitors. (Copyright ? 1995, DewRay Products, Inc. Used with permission.) A ?A hot cathode that emits electrons when heated to ?800°C and a suitable potential is applied to the adjacent electrodes G 1 and G 2 . ?An apertured grid No. 1, also called G 1 , which is maintained negative with respect to the cathode and whose potential controls the flow of electrons from the cathode. ?An apertured grid No. 2 placed close to G 1 (usually a few thousandths of an inch) and set at a positive voltage of a few hundred volts with respect to the cathode. This G 2 attracts the electrons controlled by the G 1 aperture potential and shapes the beam. ?An anode composed of metal cylinders to accelerate the electron beam toward the focus electrode and a final anode to further accelerate the beam toward the screen, where it focuses into a spot. Figure 83.19 describes the unipotential lens focus structure, also called an EINZEL lens design. Another structure, widely used in modern CRTs, is the bipotential lens focus structure represented in Fig. 83.20. The bipotential structure is theoretically a better performer than the unipotential focus structure, because the lenses have less curvature of the line forces and less spherical aberration. When optimum resolution is required, electromagnetic focus is used instead of the electrostatic focus systems described previously. Because this magnetic focus lens just bends the electron trajectories without changing the electron’s speed, and because of its large diameter, the spherical aberration is reduced and the spot size is optimized. FIGURE 83.19A typical electron gun with unipotential lens structure. FIGURE 83.20A typical electron gun with bipotential lens structure. ? 2000 by CRC Press LLC The Cathode The cathode used on most CRTs is the oxide cathode, which consists typically of a heated nickel substrate coated with barium, strontium, and calcium oxides. This cathode works at a temperature of ?800°C and provides a dc emission density up to 0.2 amp/cm 2 and 2 amp/cm 2 peak current density. When higher current densities are required, a tungsten-impregnated cathode (porous tungsten matrix heated at ?1000°C and impregnated with barium and calcium aluminates) can be used. The impregnated cathode operates at higher temperature than the oxide cathode and requires sophisticated materials and techniques for its processing. The impregnated cathode, also known as a dispenser cathode, is commonly used on projection tubes and on other tubes where high beam currents are required, typically above 1.5 mA. The Electrostatic Deflection System An electrostatic deflection system (Fig. 83.21) consists of two sets of metal plates of length l symmetrically located with respect to the electron beam axis at a distance d of each other. At the anode outlet aperture (at potential V 0 ), the beam enters the deflection plates whose potentials are, respectively, V 0 + V D , V 0 – V D . The deflection angle a at the exit of plates is such that In order to increase the deflection sensitivity, plates are often flared to have an optimum contour. High-frequency deflection systems incorporate delay lines to match the electron beam speed in the deflection zone with the signal propagation speed in the delay line. The Electromagnetic Deflection System An electromagnetic deflection coil is composed of two perpendicular windings generating electromagnetic fields perpendicular to the trajectory of the electron beam in the vertical and horizontal planes. Figure 83.22 shows the principle of electromagnetic deflection where a field of length l is applied perpendicularly to the electron beam accelerated at V B . The beam, assuming the field intensity is uniform and of length l, is deflected onto a circular path of radius r. The corresponding angle of deflection is q such as: FIGURE 83.21Principle of electrostatic deflection. tana= 1 2 0 V V l d D sin . q= Nil DV B 268 ? 2000 by CRC Press LLC where Ni is the number of ampere turns generating the magnetic field, D the diameter of the cylindrical winding generating the field, l is its length, and V B the accelerating voltage expressed in volts. From the above, the deflection angle is conversely proportional to the coil diameter. As the coil diameter is limited by the tube neck diameter, it is preferable to use a small neck diameter to increase sensitivity to reduce deflection power. Because a small neck diameter cannot accommodate the large electrostatic focusing lenses required to reduce spherical aberration and spot size, a compromise must then be found between spot size and deflection power to achieve the best CRT performance when electrostatic focus is required. The Screen A cathodoluminescent material is characterized by several parameters: ?Color, usually expressed by its spectral distribution curve and also measured by color coordinates xy or u￠v￠ ?Temporal characteristics, such as decay time, usually measured at 10% of initial excitation These parameters, as well as the chemical composition of the luminescent materials, are listed in the E.I.A. Publication TEP 116 C [Publ. TEP]. Other parameters such as luminous efficiency (expressed in lumens/watt) or energy conversion efficiency (expressed in watts/watt) are also necessary and have to be required from manufacturers. Color CRTs The Shadow-Mask CRT Color CRTs are widely used in commercial television and for computer displays. The shadow-mask CRT [Morrell et al., 1974] uses three electron beams deflected by one deflection coil. The beams traverse a perforated metal mask before impinging on the selected luminescent screen material, which is usually made of stripes or dots of red, blue, and green phosphors. The arrangement of the electron optics and of the deflection system is such that the three electron beams converge on the screen after passing through the shadow-mask, each beam impinging on one color, red, blue, or green, only. FIGURE 83.22Principle of electromagnetic deflection. ? 2000 by CRC Press LLC Shadow-mask tubes use a mechanical selection of colors. The thin perforated steel or invar shadow-mask is welded onto a metallic frame suspended by supports in the tube glass faceplate. This structure is sensitive to shock and vibration, which may affect the position of the mask in the faceplate and the registration of the beam on the appropriate phosphor dots. These types of CRTs need specialized damping when ruggedization is required. Another type of shadow-mask CRT, the flat tension mask (FTM), is much more resistant to shock and vibration because of the thin shadow-mask foil tension sealed to the face plate. Suspended weight is minimal and the CRT can withstand high shock and vibration levels [Taki et al., 1996]. The shadow-mask tube is by far the most widespread tube for computer and high-resolution monitor displays. Two other types of color CRTs are practically immune to shock and vibration. These are the beam index tube and the penetration tube. The Beam Index CRT In the beam index tube, the RGB striped screen is intermixed with indexing stripes of a UV emitting luminescent material with very fast decay time. When excited by the electron beam, the index stripes emit a light pulse that is detected by photosensors located on the transparent bulb. The signals are then digitally processed, permitting video signals to be fired at the correct position of the electron beam on the screen. Beam index CRTs are basically used for avionic applications. A typical 150 ′ 150 mm 2 (6 ′ 6 in.) CRT can display images with a white brightness around 3400 cd/m 2 (1000 fL). The Penetration CRT With the penetration CRT, operation is based on the variation in depth of penetration of an electron beam in successive layers of different luminescent materials, typically red and green emitting. At low voltage, such as 9 kV, the first layer of luminescent material is excited and a red color is obtained. At high voltage, such as 18 kV, the electrons penetrate the red layer without losing much of their energy and excite the green layer to produce a slightly desaturated green. Typically, only four colors can be produced. These tubes can be built in any size and can also use a variety of short and long persistence luminescent materials to achieve variable persistence. This is convenient for radar or other specific limited color applications. The Beam Matrix Flat CRT [Tully, 1994] This flat panel CRT combines electron beams provided by a hot cathode, like a conventional CRT, and beam modulation and addressing thanks to a matrix of conductive metallic rows and columns. The electron beam generated impinges on a fluorescent screen identical to that of a standard CRT. Such tubes are, as of 1996, still in the development stage. Another type of flat panel using an electron beam is the field emission device (FED); these FEDs are in fact using electron beams generated in a very high vacuum by field emission, then these beams are modulated and addressed by sets of matrixed conductive electrodes, arranged in rows and columns, and impinged on a fluorescent screen [Kumar et al., 1994]. These FEDs are in the early stages of development and samples of 10-in. diagonal color panels have been demonstrated in 1996, during the SID conference. Although the FEDs are not exactly corresponding to the definition of a CRT, they are emissive devices using an electron beam to illuminate a fluorescent screen and hence can be assimilated to the family of CRTs. Contrast and Brightness Contrast ratio is defined as the ratio of the luminance L 1 of the picture element to the luminance L 2 of the background as follows: Brightness contrast, important to the observer, is defined as the ratio of the luminance of the picture element plus background L 1 + L 2 to the background luminance L 2 C L L R = 1 2 ? 2000 by CRC Press LLC In order to improve contrast, the usual technique is to use an absorbing faceplate or spectrally matched filter to absorb the light emitted by the luminescent screen once and the ambient incident light twice. In addition, antireflection coatings and antiglare treatments can be applied to the CRT face to reduce the reflections from the front surface. Depending upon the ambient lighting conditions, a compromise must usually be found between light output required and the contrast need. Measurements on CRTs Line Width and Modulation Transfer Function The emitted phosphor spot of a CRT usually presents a Gaussian energy distribution. The measurement is usually performed at the 50% height of the Gaussian curve and is called L 0.5 . It can be done also at the s point of the Gaussian curve and called L 0.6 , with L 0.5 = 1.175 L 0.6 = 1.175 Ls. Spot size is related to modulation transfer function (MTF) [E.I.A., 1986] by where MTF is fractional, L is the neperian logarithm, N is the number of cycles/millimeter, and L 0.5 is expressed in millimeters, and can be related to Ls by the formula L 0.5 = 1.175Ls The spot image is normally measured using a microscope with a suitable detector such as a fiber-optic probe coupled to a photomultiplier or a CCD array. Another method widely used is the shrinking raster technique. After a raster of n horizontal lines is scanned on the CRT screen, the vertical size of the screen is reduced until the line structure disappears and produces a uniform luminance to the observer. The number of raster lines is then divided into the raster height: Comparing L SR to Ls, authors find values of L SR between 1.17 and 1.23 Ls. Brightness Brightness is measured as area brightness (raster luminance) L R or peak line brightness L p . Peak line brightness is an inverse function of the writing speed, a direct function of the refresh rate of the displayed line and, of course, of the beam current. Raster luminance L R is related to the raster emitting surface (S), to the beam current (I), to the screen efficiency (e), transmission (T) of the faceplate, and to the screen voltage (V s ) by the formula where e is lumens/watt, I is amperes, S is in square meters, T is fractional, and p = 3.1416. There is no convenient relation between peak line brightness and raster brightness. C LL L C R = + =+ 12 2 1 L N L MTF N L MTF 05 2 1 356 1 053 1 . .() . == L n SR = compressed raster height L VI S T R S = p e ? 2000 by CRC Press LLC Illumination Illumination is usually measured by illuminance in lux where: 1 lux = 1 lumen/m 2 . The S.I. units are the candela/square meter or nit for luminance and the lux for illuminance. They are related to the commonly used footlambert (fL) and footcandle (fC) by the relations: The notion of illumination has to be differentiated from the notion of brightness. The footcandle is often used for illuminance and is produced approximately by a source having a luminance of 1 fL (when the reflectance of the surface is 1), which makes the contrast calculation easy. However, these units must not be confused. Projection Screen Very often a CRT is used to project an image onto a screen. For a lambertian screen, which rediffuses the incident light in all directions, we can write where B is the luminance off the projection screen, L is the luminance of the CRT, f is the focal length of the lens used, m is the magnification factor, d is lens diameter, n is aperture, f/d, and T is the transmission of the optics. The maximum flux F (lumens) from the CRT can be expressed by F = p LS where L is the CRT screen brightness (candelas/square meters) and S the emitting screen surface (square meters), p = 3.1416. Conclusion The worldwide market for 1996 CRTs is at a yearly level of 50 million units compared to 38 million units in 1991 [Stanford Resources, 1992]. High resolution color CRTs are the mainstream of this growth because there is progressive symbiosis between multimedia and high resolution color television with the upcoming of digital television and the development of the Internet. The growth of this market is presently foreseen by the electron tube community up to 100 million units yearly. It is now obvious that the competitive technologies have still to progress to be able to replace the CRT in many applications and that this progress is much slower than the sometimes optimistic predictions of the scientific and industrial community. It is also important to note that the Electronics Industries Association is monitoring several standard committees for CRTs and that proposals for standards are submitted by the industry and by laboratories such as the National Information Display Laboratory (see Further Information). These standards are following the evolution of measurement techniques. The adoption of these standards by the industry will make it possible for display users to choose their equipment according to criteria uniformly approved around the world. The CRT is widely used aboard commercial and military aircraft for cockpit displays, aboard surface ships and submarines, and aboard military and commercial vehicles. Because of its wide range of capabilities, versatility of use, brightness and contrast, image quality, and efficiency, the qualities of the CRT largely offset 1 1 342 11 10 fL cd/ft cd/m fC lm/ft 8 lux 1 lux 1lm/m 22 2 2 == == = p . . BL T nm = +41 22 () ? 2000 by CRC Press LLC its bulkiness and weight for most applications. CRTs are still dominating high-resolution applications ranging from 12-mm helmet-mounted displays to 1000-mm diagonal 2000 ′ 2000 pixels color monitors and, according to P. Brody [1980], will continue to dominate the market for a few decades. Defining Terms Cathode ray tube: A vacuum tube which uses cathode rays to generate a picture on a fluorescent screen. These cathode rays are in fact the electron beam deflected and modulated, which impinges on a phosphor screen to generate a picture according to a repetitive pattern refreshed at a frequency usually between 25 and 72 Hz. The term cathode rays stems from the discovery by Plücker [1858] and Hittorf [1869] of a blue glow present at some spots on a glass tube when studying high-voltage discharge in a low-pressure gas. Crookes [1879] showed that these cathode rays were deflected by a magnetic field and were in fact electrons emitted by the negative electrode. The more poetic term cathode rays has been kept instead of electron beam for the cathode ray tube. Contrast: For a cathode ray tube, contrast is the evaluation of the visibility of the picture presented on the phosphor screen in a given ambient lighting. Contrast is usually measured by the contrast ratio, which is the ratio of the luminance of the picture element under evaluation to the background luminance. Flux: Also called radiant flux, the radiant power emitted by a source. It can be expressed in watts for a radiometric source or in lumens when the source spectral energy distribution is between 400 and 700 nm. Illumination: The effect of a visible radiation flux received on a given surface. Illumination is measured by the illuminance, which is the luminous flux received by surface unit, usually expressed in lux. One lux equals 1 lumen/m 2 . Raster: Also called television raster; it is developed by a moving spot of light generated by an electron beam scanning a CRT phosphor screen in a predetermined and repetitive pattern. A picture is generated by modulating the beam intensity, hence the spot light output, when scanning the screen surface. Usually, horizontal lines are generated scanning in a left-to-right sequence and developed top-to-bottom of the image surface. Related Topic 83.4 Color Plasma Displays References T.P. Brody, “When—if ever—will the CRT be replaced by a flat display panel?” Microelectronics Journal, vol. 11, pp. 5–9, 1980. W. Crookes, “On the illumination of lines of molecular pressure and the trajectory of molecules,” Philos. Trans. R. Soc. London, vol. 170, pp. 135–164, 1879. E.I.A. J.T 20 Committee—Meeting #59 (1986)—Test Method—Measurement of M.T.F. for Monochrome CRTs by Fourier Transform. W. Hittorf, “über die Elektricit?tsleitung der Base,” Ann. Phys. (Leipzig) [2], vol. 136, pp. 1–31, 1869. P. Ke l l e r, The Cathode Ray Tube, Technology, History, and Applications, New York: Palisades Institute for Research Services, 1991. N. Kumar, H. Schmidt et al., “Development of nano-crystalline diamond based field emission displays”, SID 1994 Symposium Digest, Conference 6.1. A. Martin, Cathode Ray Tubes for Industrial and Military Applications, vol. 67, New York: Academic Press, 1986, pp. 183–328. A.M. Morrell, H.B. Law, E.G. Ramberg, and E.W. Herold, Color Television Picture Tubes, New York: Academic Press, 1974. J. Plücker, “über die Einwirkung der Magneten auf die elektrischen Entladungen in verdüunten Gasen,” Ann. Phys. (Leipzig) [2], vol. 103, pp. 88–106, 1858. Publication TEP 116C, Washington, D.C.: Electronic Industries Association. ? 2000 by CRC Press LLC Stanford Resources—Monitor Market Trends—1991, Menlo Park, Calif.: Information Associates, 1992a, 552 pp. Stanford Resources—Electronic Display World, San Jose, Calif.: Stanford Resources, Inc., vol. 12, no. 3, p. 4, 1992b. A. Taki, N. Arimoto, T. Okamoto, and Y. Ueda, “Development of 17-inch pure flat color monitor tube”, SID 1996 Symposium Digest, Conference 38.4. P. Tully, “Matsushita’s color flat panel”, Information Display, 6, 9–11, 1994. Further Information Electronic Industries Association (E.I.A.) 2001 I Street, N.W. Washington, D.C. 20006 E.I.A. prints a wide range of literature on electronics components, computers and industrial electronics, communications and services. E.I.A. also administers many committees for standardization, safety, etc., for cathode ray tubes and other components. E.I.A. offers a complete Electronics Technology Curriculum and technical training books, tapes, etc. National Information Display Laboratory David Sarnoff Research Center–CN 8619 Princeton, NJ 08543-8619 The NIDL’s strategic objective is to provide direct support to government users while promoting the devel- opment and commercialization of advanced soft copy technologies. The NIDL participates in the elaboration of standards for soft copy, for example, for high-resolution monitors, in close relation with E.I.A. 83.4 Color Plasma Displays Larry F. Weber Introduction The last few years have seen an explosive growth in manufacturing capacity and interest in full color plasma displays. This is fueled by the realization that plasma displays can fulfill the long sought after goal of consumer- affordable hang-on-the-wall flat-panel television displays with diagonals in the range of 20 to 60 in. Color plasma displays operate on the same physical principle as fluorescent lamps. A gas discharge generates ultraviolet light which excites a phosphor layer that fluoresces visible light. Differing phosphors are used for the red, green, and blue primaries and a full color moving image is obtained by modulating each primary color sub-pixel to one of typically 256 intensity levels at 60 times a second. One quantitative measurement of industrial activity is the list of major corporate efforts on the three distinct structures shown in Fig. 83.23. Each of these companies is now manufacturing or has demonstrated 40-in. or larger color plasma panels. Many of these companies are looking to plasma displays as the next display device opportunity that will follow the success model of the active matrix liquid crystal displays (AMLCDs). Color Plasma Display Markets The plasma display manufacturers have adopted the strategy of a strong attack on the greater than 40-in. diagonal NTSC television and high definition television (HDTV) markets. Display diagonals smaller than 20 in. are specifically avoided in this strategy. Plasma displays have found their proper place in the market by evading the fierce competition from the smaller diagonal LCDs and CRTs since both of these technologies have difficulty with 40- to 60-in. diagonals. In this diagonal range, plasma displays will compete primarily with projection displays. While projection systems have recently shown a very high level of achievement they are especially vulnerable to plasma because of their limited viewing angle and bulk. It is clear that projection systems have not found ? 2000 by CRC Press LLC much success in diagonal ranges where there is a wide viewing angle CRT alternative. Thus, a wide angle and thin plasma display alternative at a competitive price will easily dominate this large diagonal market. This market is quite large because of the expectation that many HDTV viewers will prefer the wide screen theater- like effect of the high resolution image. The challenge to the plasma displays will be to achieve prices competitive to those of projection systems. The potential market for large plasma displays is enormous. The most optimistic market projections show annual world-wide sales of about 10 million plasma displays in the year 2002. While this projection shows a very rapid growth, it is still only 5% of the slightly less than 200 million television sets that will be sold that year. More conservative projections still show $5 billion world-wide sales for plasma displays in 2002. Color Plasma Display Attributes Table 83.1 shows some of the attributes of color plasma displays which make them successful. The following reviews each attribute. 1.The electrical characteristics of the gas discharge allow plasma displays to be made with diagonals in the 20- to 60-in. range. Such large diagonals are facilitated by very strong non-linearity and inherent memory of the discharge, as discussed in Items 3 and 4 below, which present no practical limitations to the number of lines that can be multiplexed. Also, the high impedance characteristic (covered below in Item 11) coupled with the ability to use highly conductive opaque electrodes, greatly reduces electrode loss limitations to size. Monochrome plasma displays have been sold with sizes as great as 60-in. diagonal, FIGURE 83.23Major corporate efforts on fundamental structures. TABLE 83.1Color Plasma Display Attributes 1.Diagonals of 20 to 60 in. 2.Full 16 million colors 3.Very strong non-linearity 4.Inherent memory 5.Long lifetime 6.Very wide viewing angle 7.Instant update time 8.Good luminance and luminous efficiency 9.CRT-like manufacturing model 10.Tolerant to shock, vibration, and temperature extremes 11.Reasonable impedance characteristics 12.Precise digital grey scale 13.CRT-like color gamut ? 2000 by CRC Press LLC having over 4 million pixels. By 1996 full color plasma displays had been demonstrated with 46-in. diagonals and 4 million sub-pixels. 2.The all-digital gray scale technique used in color plasma displays allows each primary sub-pixel to display 256 or more intensity levels. This allows full 24-bit color or 16 million colors. These 256 intensity levels are not the limit of the plasma displays but rather a convenient design point for software and system compatibility. 3.The plasma display has a very large non-linearity due to the electrical characteristic of the gas discharge used in all plasma displays. This is an electrical non-linearity, meaning that below a certain threshold voltage, the gas discharge will emit no light. Of course, above that threshold voltage the gas discharge fires and emits a desired color. Very sharp non-linearity allows plasma displays to be multiplexed without limit which makes very large plasma displays practical. This is demonstrated by a number of recently developed 1280 ′ 1024 ′ 3 sub-pixel color plasma displays. This is a considerable advantage when compared to other display technologies such as the liquid crystal. The liquid crystal display does not have a very good non-linearity and, therefore, some other non-linear element, such as a thin film transistor, is sometimes added in series with each liquid crystal element to increase the display non- linearity. Of course, this greatly complicates and adds cost to this active matrix liquid crystal. 4.Most color plasma displays have inherent memory which is stored directly in the glass plasma panel. Memory is very desirable for flat panel displays because it allows the display to be very bright even for very large sizes. This is because a display with memory has a pixel duty cycle of one. Displays without memory have a pixel duty cycle of one divided by the number of scanned lines. Thus, as the non-memory displays get bigger and the number of scanned lines increases, the duty cycle and, therefore, the brightness of the display decrease. An additional value of memory is the elimination of flicker because the pixels are on all of the time. 5.The lifetime of color plasma displays can be long. Full color plasma display products have been delivered with specified lifetimes to half luminance of 10,000 to 30,000 h. While this is comparable to some CRT products, there is considerable effort to extend this lifetime further. The failure mode is usually a slow degradation in the phosphor that gradually decreases the display luminance. If all of the pixels are aged uniformly, with perhaps a randomly moving television image, the display will still be usable after the specified lifetime but at reduced luminance. However, displays used for computer images require a much tougher life specification because images such as icons may be left on the same screen location for long periods and burn in an image. These problems are very similar to the phosphor degradation observed on a CRT. 6.One of the major advantages of all plasma displays over liquid crystals is the very wide viewing angle. Plasma displays can even get brighter when viewed off axis and, therefore, have the widest viewing angle of any display technology. 7.Gas discharges switch in microseconds and so plasma displays can be updated instantly. Speed is especially important for the very popular mouse and cursor operations where the cursor would disappear when moving on a liquid crystal display. Full motion television images are not a challenge for plasma displays. 8.The luminance and luminous efficiency of color plasma displays are good. Displays having 450 candelas per meter squared at 1 lumen per watt have been demonstrated. Some other display technologies such as ac electroluminescence have a higher material luminous efficiency but this performance must be tempered by the fact that the plasma panel has 1000 times less electrical capacitance than the ac electroluminescent devices. The plasma panel frequently takes less power than the EL devices when the switching loss of the EL panel is considered. This favors plasma panels for larger numbers of scanned lines and favors EL panels for smaller numbers of scanned lines. The crossover point is somewhere in the region of a few hundred scanned lines. 9.Color plasma displays are manufactured in a plant that has considerable commonalty with CRT man- ufacturing plants. This contrasts sharply with the semiconductor-like manufacturing plant of AMLCDs. Therefore, the plasma display plant will cost much less than the AMLCD plant. 10.The structure can withstand very high levels of shock and vibration when properly mounted. Military plasma displays have been designed for in excess of 150 Gs of shock. Plasma displays can easily operate at both high and low temperature extremes. ac plasma displays have a temperature limit dependent ? 2000 by CRC Press LLC almost solely on the drive circuit characteristics. dc plasma displays, which use mercury, should not be operated for long periods at low temperatures without an external heater. All recently introduced color dc plasma displays do not have mercury and do not have this limitation. 11.Plasma displays have a high input impedance characteristic that makes them easy to drive. The dielectric constant of the gas is equal to one, which means that plasma displays have virtually the lowest possible electrode capacitance. This is 1000 times smaller than electroluminescent displays and about 100 times smaller than liquid crystal displays. This translates to lower current requirements and, therefore, smaller drive circuit silicon area for the plasma displays. While plasma displays do require 100-volt address drivers, it is frequently easier to design high voltage circuits than high current circuits. Also, 40-in. and larger displays can be designed with little power dissipation in panel electrodes. 12.The gray scale technique used in color plasma displays is 100% digital, which allows design of an all digital image system having reduced noise and increased stability in the color representations. This will become more important as signal sources with very high quality digital signals, such as those from digital video disks and HDTV, become widely available. 13.The color gamut of the available plasma display phosphors is very good. While the color coordinates do not yet exactly match those used in the CRT, future process adjustments are expected to produce the desired close match. Gas Discharge Physics A brief account of gas discharge physics will be covered below. A more detailed discussion of this material is presented in Weber [1985]. Figure 83.24 shows the important reactions that occur in a gas discharge for the monochrome gas mixture of neon and argon. The reactions in the gas volume include ionization (I), excitation (E), metastable generation (M), and Penning ionization (P). The three surface reactions that occur at the cathode cause ejection of electrons from the cathode by a bombarding neon ion, a neon metastable atom or by a high energy photon. The most important volume reaction is ionization (I), which can cause the generation of an avalanche in the gas volume as shown in Fig. 83.24. This avalanche is started by an electron near the cathode and as it grows toward the anode, it generates a large number of electron-ion pairs. The number of electron-ion pairs increases with increasing applied voltage across the gas. Ions, photons, or metastable atoms that are transported to the cathode can then eject electrons with a cathode surface-dependent probability and these ejected electrons will initiate further avalanches. These mechanisms act as a positive feedback system that becomes unstable when the loop gain is greater than 1. The onset of the unstable condition is defined as the gas firing voltage. Above this firing voltage the discharge current will continue to grow without bounds if the initial avalanche is primed with at least a single electron. Figure 83.25 shows the I-V characteristic of a typical gas discharge found in plasma displays. Note that the current is plotted on a log scale over nine orders of magnitude. The most striking feature is the very strong non-linearity at the firing voltage, which is a major attribute of gas discharges that allows matrix addressing. When the discharge current has sufficient magnitude, space charge distortion sets in and the characteristic achieves a negative resistance region. Most plasma displays operate near the junction of the normal and the abnormal glow regions of the characteristic. One critical aspect of gas discharges is the requirement for external priming as shown in the lower part of Fig. 83.25. The avalanche process shown in Fig. 83.24 needs at least one electron to start the discharge growth. Without this first electron, the discharge will not start at any voltage. Priming electrons can come from a number of different active particles created either by a prior discharge or by neighboring discharging pixels. Active particles include free electrons, free ions, metastable atoms, and ultraviolet photons. Figure 83.26 shows the characteristics of the glow discharge commonly found in operating plasma displays. The light comes from two luminous regions: the negative glow and the positive column. All plasma displays on the market today use light from the negative glow but a few research displays have used the light from the positive column. These regions are caused by the space charge distribution of the electrons and ions that distort the electric field and voltage distribution. ? 2000 by CRC Press LLC Current Limiting for Plasma Displays To avoid a catastrophic arc, the current in a gas discharge must be limited by some means. There are a number of ways of accomplishing this, but only two, shown in Fig. 83.27, have achieved commercial success. dc plasma displays use a resistor, a semiconductor current source, or a short applied voltage pulse to limit the current and have the electrodes in intimate contact with the gas discharge. ac plasma displays limit the current with an internal glass dielectric that couples the electrodes capacitively to the gas discharge. Most of the commercially successful monochrome dc displays have the resistors or current sources connected to a display electrode external to the panel which allows only one discharge to be ignited along that electrode at any one time. This works well for scanned displays. Multiple discharges and dc memory require placing the FIGURE 83.24Model of important gas discharge reactions. FIGURE 83.25The I-V characteristic of a gas discharge. ? 2000 by CRC Press LLC resistor internal to the panel in series with each pixel. Recent materials and process advances have allowed practical dc color displays with memory to be made having a resistor per sub-pixel. The ac displays can achieve memory and the necessary current limiting with a simple dielectric layer that forms a capacitor in series with each pixel. When a voltage pulse is applied to an ac panel, the discharge deposits FIGURE 83.26 Luminous regions of a gas discharge. FIGURE 83.27 The two current limiting techniques used in plasma display products. ? 2000 by CRC Press LLC a charge on the wall that reduces the voltage across the gas. After a short time, the discharge will extinguish and the light output will end until the applied voltage reverses polarity and a new discharge pulse occurs. This wall charge allows the ac plasma displays to operate in a memory mode, which greatly increases the brightness of large displays. ac Plasma Displays Figure 83.23 shows that currently, color plasma displays are dominated by the ac technology, so it is worth examining the ac monochrome structure shown in Fig. 83.28 [Criscimagna and Pleshko, 1980]. These panels are made by depositing thin film electrodes on the front and back substrates and then covering those electrodes with a thin dielectric glass. Recall from Fig. 83.27 that this dielectric glass makes a capacitor that is used to limit the discharge current. This dielectric also is used to store the charge that gives these panels inherent memory. The two substrates are then sealed together around the perimeter and filled with neon gas. The ac panels have a very simple structure that allows the pixels to be isolated simply by the action of electric fields. The ac plasma panels require the inner surface of the dielectric that is in contact with the gas to have a special coating of magnesium oxide. This MgO layer is necessary for the panel to have low operating voltages and long life. Being a refractory oxide, MgO sputters away at a very low rate and it is also well known for its high secondary electron emission. The largest plasma display product ever manufactured (prior to 1997) uses the Fig. 83.28 structure and has a diagonal of 60 in. with a 2048 ′ 2048 array of more than 4 million pixels [Wedding et. al., 1987]. This display operates at a very high update rate so that it will work with a standard NTSC video source. The memory feature allows this display to have the same luminance as the smaller page sized displays. ac displays require that an ac signal, called the sustain voltage, be applied during operation as shown on the right side of Fig. 83.27. A typical sustain frequency is 50 kHz. Figure 83.29 shows the details of this operation for a pixel in both the on and off states. When a pixel is discharging, charge collects on the dielectric glass walls and influences the voltage across the gas. The component of voltage due to this charge is called the wall voltage. When a pixel is on, the wall voltage changes for each polarity reversal of the sustain voltage. This change in wall voltage coincides with a pulse of light due to the gas discharge. When the pixel is off, there are no light pulses, and the wall voltage remains at a zero level. Pixel addressing is achieved through a partial discharge by introducing an address pulse timed between the sustain pulses. A write pulse causes the wall voltage to transit from zero volts to the final equilibrium wall voltage level. Likewise, an erase pulse causes the wall voltage to return to zero. FIGURE 83.28Monochrome ac plasma display structure. ? 2000 by CRC Press LLC Color Plasma Display Devices Color is achieved by placing phosphors in the plasma panel and then exciting those phosphors with the ultraviolet light of the gas discharge. This is the same principle as used in the fluorescent lamp. Xenon is the active UV generating gas which provides atomic resonance radiation at 147 nm and a molecular band centered at 173 nm. Neon or helium buffer gases are always mixed with the xenon. Figure 83.23 indicates the companies who are making a serious effort at developing the three dominant full color structures. The basic concepts of the two ac structures are shown in Fig. 83.30. The double substrate structure is very similar to the monochrome ac structure shown in Fig. 83.28. The single substrate structure separates the discharge cathode areas from the phosphor by applying the sustain voltage only to the lower electrodes while the phosphor is on the top. The single substrate approach promises longer phosphor life because it is not directly sputtered by the energetic ions that are directed toward the cathodes. The structure for the single substrate ac plasma displays is shown in Fig. 83.31 [Shinoda et al., 1993]. Note that the front and back substrates each have simple one dimensional features. Since the two substrate structures are positioned orthogonally, there is no critical alignment between the two substrates because the pixels will automatically occur wherever the orthogonal electrodes intersect. This allows for straightforward manufacture of large panels. The phosphors are placed on the rear substrate of the panel in Fig. 83.31 and are excited by the ultraviolet light generated by the electrodes on the top substrate. In this case, both sets of ac sustain electrodes are on the upper substrate. The ac voltage is applied to these electrodes in the normal way and the fringing fields from these electrodes reach into the gas and create a discharge. Note that the structure in Fig. 83.31 has glass barrier rib separators between each sub-pixel. This is necessary to reduce cross-talk between the different colors that will reduce the color purity. These barrier ribs do not transmit the 147-nm or 173-nm radiation generated by the xenon gas used in color plasma displays. The phosphors are placed on all walls of the sub-pixel channel except for the front plate which has the phosphor damaging sputtering activity at the cathodes. This nearly complete phosphor coverage of the walls maximizes luminance while minimizing sputtering damage. Other important features of the structure in Fig. 83.31 are the address electrodes buried beneath the phos- phors of the rear substrate. These are the column electrodes that are selectively pulsed depending on the input image data. While these address operations do create discharge activity that could potentially sputter damage FIGURE 83.29Sustain voltage, wall voltage, and light output for ac plasma pixels in the on and off states. ? 2000 by CRC Press LLC the phosphor, the address pulse frequency is orders of magnitude lower than the sustain frequency and so the amount of address damage is minimal. The sustain electrodes shown in Fig. 83.31 are made of a conductive transparent material such as tin oxide. Unfortunately, the resistance is orders of magnitude too high. To correct this problem, narrow bus electrodes of high conductivity materials, such as silver or chrome-copper-chrome, are placed over the tin oxide to reduce the electrode resistance to values on the order of 100 ohms. The ac double substrate device shown on the left of Fig. 83.30 has a structure very similar to the monochrome device of Fig. 83.28 [Doyeux and Deschamps, 1995]. The major differences are the introduction of glass barrier ribs to maintain color purity and the introduction of color phosphors. The phosphors are placed on one substrate and are carefully positioned to avoid the location directly over the electrode since the sputtering action over the electrode will cause significant phosphor degradation. Figure 83.32 shows the color dc plasma display structure [Koike et al., 1995]. The major difference between the dc and ac structures is the placement of resistors in series with each sub-pixel to limit the discharge current for the dc case as shown in Fig. 83.27. Another difference is the requirement for the dc device to have barrier ribs on all four edges of each sub-pixel. This is needed to prevent the dc discharge from spreading to neighboring sub-pixels in addition to the color purity issue discussed above for ac panels. Unlike the single substrate ac structure shown in Fig. 83.31, the alternate color phosphors are patterned along each row and each column in FIGURE 83.30Two major structural designs of color ac plasma displays. FIGURE 83.31Structure of single substrate color ac panel. ? 2000 by CRC Press LLC a two-dimensional structure. This requires careful alignment between the front and the rear substrates. Priming is achieved by means of the auxiliary cells and anodes placed between the sub-pixels that generate a discharge that is not visible to the viewer. Inherent memory is achieved for the dc plasma structure of Fig. 83.32 with the pulsed memory mode waveforms shown in Fig. 83.33 [Takano et al., 1994]. This operates on the principle that if there is metastable priming from the preceding discharge pulse, then the discharge will build up in a sufficiently short time to fully mature during the very short electrode voltage pulse. The electrode voltage pulse is adjusted to be sufficiently short to inhibit a sequence of discharges if there has not been an initiating higher amplitude address pulse. The pulse memory mode was invented over 25 years ago [Holz, 1972], and is the most widely studied technique for achieving memory in dc plasma displays. It was just recently introduced to commercial production. Gray Scale The ac or dc memory displays cannot use pulse intensity or pulse width modulation for gray scale because the pixels in memory mode are either on or off and such pulse perturbations would, in many cases, have the undesirable effect of changing the state of the pixel. Instead these memory displays achieve gray scale by modulating the percentage of time that the pixel is on in a given frame. This means that the pixels must be addressed multiple times per frame. In the sequence shown in Fig. 83.34 for 256 intensity levels, each frame is divided into eight sub-fields and each sub-field consists of an address period and a sustain period [Yoshikawa, 1992]. During a given address period, address pulses are applied to all pixels in the panel according to the sub- field image data. Each of the eight sub-fields has a sustain period with a different number of sustain cycles which emits an amount of light proportional to the number of sustain cycles. If each data bit of a given pixel FIGURE 83.32Structure of color dc panel. FIGURE 83.33Pulsed memory mode for dc plasma panels. ? 2000 by CRC Press LLC intensity word is allowed to control one of the sub-fields, then the total number of sustain cycles (and light) per frame will be proportional to the 8-bit intensity value. This results in a precise inherently digital gray scale. Defining Terms ac plasma displays: These employ an internal capacitive dielectric layer to limit the gas discharge current. dc plasma displays: These employ an internal or external resistor to limit the gas discharge current. Luminous efficiency: The measure of the display output light luminance for a given input power, usually measured in lumens per watt, which is equivalent to the nit. Memory: The property of a display pixel that allows it to remain stable in an initially established state of luminance. Memory gives a display high luminance and absence of flicker. Metastable atom: An atom in a temporary but long lived excited state in which photon emission is forbidden by electrodynamic theory. Metastables can give up their energy by ionizing other atoms or through wall collisions. Plasma: The fourth state of matter comprised of positive ions and negative electrons of equal and sufficiently high density to nearly cancel out any applied electric field. Not to be confused with blood plasma. Sputtering: The physical process whereby an ion with kinetic energy in the gas collides with a solid surface ejecting an atom from the solid into the gas. Related Topic 83.3 The Cathode Ray Tube References T. N. Criscimagna and P. Pleshko, “ac Plasma Display,” in Topics in Applied Physics, Vol. 40 Display Devices, Berlin: Springer-Verlag, 1980, pp. 91–150. H. Doyeux and J. Deschamps, “A high-resolution 19-in. 1024 ′ 768 Color ac PDP,” SID Intl. Symp., Orlando, pp. 811–814, 1995. FIGURE 83.34 Addressing sequence for 256 intensity gray scale in ac memory plasma displays. ? 2000 by CRC Press LLC G. E. Holz, “Pulsed gas discharge display with memory,” SID Intl. Symp., San Francisco, pp. 36–37, 1972. J. Koike et al., “Long-life, high luminance 40-in. color DC PDP for HDTV,” Intl. Display Res. Conf., Hamamatsu, Japan, pp. 943–944, 1995. T. Shinoda et al., “Development of technologies for large-area color ac plasma displays,” SID Intl. Symp., Seattle, pp. 161–164, 1993. Y. Takano et al., “A 40-in. dc-PDP with new pulse-memory drive scheme,” SID Intl. Symp., San Jose, pp. 731–734, 1994. L. F. Weber, “Color plasma displays,” SID Seminar Lecture Notes, Vol. 1, San Diego, pp. M-6/1–41, 1996. L. F. Weber, “Plasma displays,” in Flat-Panel Displays and CRTs, L. E. Tannas Jr., Ed., New York: Van Nostrand Reinhold, 1985, pp. 332–414. D. K. Wedding, P. S. Friedman, T. J. Soper, T. D. Holloway, and C. D. Reuter, “A 1.5 m diagonal ac gas discharge display,” SID Intl. Symp., New Orleans, pp. 96–99, 1987. K. Yoshikawa et al., “A full color ac plasma display with 256 gray scale,” Intl. Display Res. Conf., Hiroshima, pp. 605–608, 1992. Further Information A more detailed account of the material presented in this section and presentations of other display technologies are provided in Weber [1985] and Weber [1996]. The Society for Information Display (SID) annual International Symposium publishes a digest of technical papers which is the best source for new display developments. Tutorial material can be found in the annual SID Seminar Lecture Notes. More research-oriented papers can be found in the technical digest of the Inter- national Display Research Conference which rotates annually among Europe, Japan, and North America. In addition, SID publishes the quarterly Journal of the SID, which contains more detailed archival versions of selected papers from the conferences. These materials can be obtained from the SID at 1526 Brookhollow Drive, Suite 82, Santa Ana, CA 92705-5421, or see the web site http://www.display.org/sid. ? 2000 by CRC Press LLC