1
Optics in Information Technology and Telecommunications

The information industry, including the services it provides, is growing rapidly worldwide. Its annual revenue is estimated to exceed $1 trillion, which, at an average revenue of $200,000 per employee, translates into 5 million jobs.

The demand for new information services, including data, Internet, and broadband services, has combined with the supply of innovative information technology to move us rapidly into the information age (Figure 1.1). The information age has recently been called the "tera era" because its technology demands are terabit-per-second information transport, teraoperations-per-second computer processing power, and terabyte information storage (see Box 1.1). Because of the growing importance of image and video information, there is also demand for

FIGURE 1.1 The growth of the Internet and the World Wide Web is among the factors driving the growth of information technology. Although not all sources agree on the exact statistics of this growth, all agree that it is a spectacular phenomenon. Optics is an important enabler for information technology. (Courtesy of P. Shumate, Bellcore.)



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 29
Harnessing Light: Optical Science and Engineering for the 21st Century 1 Optics in Information Technology and Telecommunications The information industry, including the services it provides, is growing rapidly worldwide. Its annual revenue is estimated to exceed $1 trillion, which, at an average revenue of $200,000 per employee, translates into 5 million jobs. The demand for new information services, including data, Internet, and broadband services, has combined with the supply of innovative information technology to move us rapidly into the information age (Figure 1.1). The information age has recently been called the "tera era" because its technology demands are terabit-per-second information transport, teraoperations-per-second computer processing power, and terabyte information storage (see Box 1.1). Because of the growing importance of image and video information, there is also demand for FIGURE 1.1 The growth of the Internet and the World Wide Web is among the factors driving the growth of information technology. Although not all sources agree on the exact statistics of this growth, all agree that it is a spectacular phenomenon. Optics is an important enabler for information technology. (Courtesy of P. Shumate, Bellcore.)

OCR for page 29
Harnessing Light: Optical Science and Engineering for the 21st Century BOX 1.1 THE ''TERA ERA" VISION FOR INFORMATION TECHNOLOGY IN 10 TO 15 YEARS TRANSPORT Terabit-per-second backbone, long-haul networks · Access networks operating at hundreds of gigabits per second · Local area networks operating at tens of gigabits per second · 1 gigabit per second to the desktop PROCESSING Teraoperations-per-second computers · Terabit-per-second throughput switches · Multigigahertz clocks · Interconnections operating at hundreds of gigabytes per second STORAGE Terabyte data banks · Multiterabyte disk drives · Tens-of-gigabit memory chips The tera era is a 10- to 15-year vision for the needs of the information age, as articulated by Joel Birnbaum of Hewlett-Packard in October 1996. Projections for switching and details for storage have been added. This vision includes the need for cost-effective networks of virtually unlimited bandwidth. Note that the roadmaps of several key information technologies promise to meet the requirements of this vision: fiber transport capacity, computer processing power, and magnetic storage density are all advancing by a factor of 100 every 10 years. This implies that giga (109) performance will improve to tera (1012) performance within about 15 years. strong advances in display technology. Optics and electronics are partnering and complementing each other in meeting this demand for information technology and thus enabling the information age. There are five major technology segments in which optics plays a major role or has the chance to do so in the future. One of them is information transport over long distances, through large networks under the ocean, across continents, and in the local networks of the telephone and cable television systems. For this, optical fiber transmission is already the technology of choice, with a clear edge in cost and performance over competing technologies such as coaxial cable or satellite communications. Optical processing, including switching and networking, is another segment in which worldwide R&D hopes to open up new markets. A third segment is the storage of information, where technology has to meet rapidly increasing demands for more and more storage capacity. Optical techniques are a strongly growing complement to traditional magnetic storage. A fourth segment is the display of information, for which optics is the intrinsic, unavoidable link between the human eye and the electronics of a television or computer. The fifth important segment is the interface between electronic machinery and information recorded on paper, which includes printers, scanners, and copiers. This segment is not covered in this report since most of its

OCR for page 29
Harnessing Light: Optical Science and Engineering for the 21st Century R&D is conducted in industry, which—for competitive reasons—keeps these subjects quite proprietary; very little work is done in this area in university or government laboratories. The field of information technology provides many examples of the enabling role of optical technology. Often these enablers are small in size or cost but have an impact on a grand scale in large systems and applications. A tiny semiconductor laser, for example, enables the building of an optical transmitter, which enables a transmission system, which enables the construction of a telecommunications network, which enables the delivery of information age services such as multimedia or the Internet. Another example is the optical fiber, which enables the construction of an optical cable, which enables the construction of a network, and so on. A third example is the liquid crystal, which enables the flat-panel display, without which the laptop computer could not exist. These chains of enablers make it difficult to place firm dollar values on individual component technologies; components such as lasers or fibers are relatively inexpensive, but the service revenues of telecommunications networks are in the hundreds of billions of dollars. Box 1.2 gives a more detailed illustration of the enabling devices for long-haul information transport systems. As for optics in general, optical materials play an important enabling role. Chapter 6 gives further details of this topic (see Box 6.2, "Photonic Materials"). A few rather obvious but critical points about high-tech mass markets and low-cost manufacturing should be more widely understood and appreciated. They clearly apply to the mass markets for optical information technologies: BOX 1.2 ENABLING PHOTONIC DEVICES FOR PRESENT AND FUTURE LIGHTWAVE LONG-HAUL SYSTEMS Semiconductor lasers High-power pumps for fiber optical amplifiers Integrated laser-modulator transmitters Multiwavelength transmitters Tunable transmitters Soliton sources Semiconductor optical amplifiers Power amplifiers in integrated transmitters 1. 3-µm optical amplifiers Optical switches and switch arrays Wavelength converters and other nonlinear functions Photodetectors and OEIC receivers High-speed (>10 gigabit-per-second [Gb/s]) and high-sensitivity OEIC receivers High-speed (>10 Gb/s) avalanche photodiodes Planar integrated waveguide components Couplers Filters (fixed and tunable) Wavelength multiplexers and routers Modulators and switches Doped-waveguide devices Dispersion compensators Fiber-type components Doped-fiber devices (amplifiers and lasers) Couplers Grating filters Wavelength multiplexers Dispersion compensators Switches (mechanical) High-power lasers and brightness converters Isolators and circulators Nonlinear devices (Source: T. Li, AT&T Laboratories.)

OCR for page 29
Harnessing Light: Optical Science and Engineering for the 21st Century When a technology creates a mass market, its products become commodity items whose production requires a capability for low-cost manufacturing. Suppliers that have prepared and invested to develop competence and capability in low-cost manufacturing by the time the mass market emerges are likely to gain the largest market share. The revenues generated by a large market share provide a significant source of R&D funds. This fact results in a rapid learning curve for the technology and its manufacturing, which drives costs down further and creates barriers to market entry for other suppliers. Three established mass markets in optical information technology serve as excellent illustrations of the above points: (1) compact disk (CD)-based optical storage, (2) liquid crystal displays, and (3) cathode-ray tube displays. Each market has global revenues in excess of $20 billion per year, and in each market, offshore manufacturers planned early and were ready with timely low-cost manufacturing. As a result, more than 90% of the manufacturing of these high-tech, mass market products now occurs offshore. At least four emerging optical information technologies—fiber to the home, optical data links, small and miniature displays, and projection displays—have an excellent chance of creating global mass markets in the not too distant future that are similar in magnitude to the three listed above. We should learn from history and make sure that U.S. industry captures a good share of the emerging mass markets. The remainder of this chapter summarizes key trends in the four major technology segments described above, including the size of their markets and the technologies they compete with. It also outlines key challenges in this area for science, technology, education, and international competitiveness. The bulk of the chapter is based on the expert presentations and position papers listed in Appendix B. Information Transport Optics is now the preferred technology for the transmission of information over long distances. This technology, based on semiconductor junction lasers and optical fibers, is the technology of the information superhighways used to transmit voice, data, and video information. The major market segments include undersea transmission between continents, terrestrial long-distance telephone trunking between states and cities, local telephone exchange links between the central office and the home, and the networks of the cable television companies. Satellite-to-satellite links also offer exciting possibilities. Table 1.1 illustrates typical distance scales for these market segments.

OCR for page 29
Harnessing Light: Optical Science and Engineering for the 21st Century TABLE 1.1 Typical Distance Scales for Selected Information Transport Applications Application Distance (km) Satellite links 50,000 Undersea transmission 1,000-10,000 Terrestrial long-haul 20-1,000 Cable television links 10-20 Fiber in the local exchange 10-20 The optical fiber medium offers several advantages over earlier transmission media such as coaxial cables and copper wire pairs. The most significant among these are large transmission capacity (high bandwidth), large repeater spacings, small cable size, low cable weight, and immunity from electromagnetic interference. These advantages have led to large-scale installations of optical fiber all over the world. It is estimated that more than 100 million kilometers of fiber are now installed. Figure 1.2 provides more details on U.S. installations. Among the grand challenges that face the R&D community in this area are (1) the development of a cost-effective wavelength-division multiplexing technology for transmitting multiple signals on a single fiber, and FIGURE 1.2 Nearly 40 million kilometers of fiber have now been installed in the United States, and the number is still growing strongly. As the figure shows, long-haul installations (by both long-distance and local telephone companies) started the trend around 1982. Since 1990, fiber installations by cable television (CATV) companies have grown strongly. It is anticipated that fiber-to-the-curb or fiber-to-the-home installations by local telephone companies will follow. Note that a single fiber cable typically contains 20 to 40 optical fibers. (Courtesy of T Li, AT&T Laboratories.)

OCR for page 29
Harnessing Light: Optical Science and Engineering for the 21st Century (2) the development of a low-cost technology for large-scale deployment of fiber-to-the-home systems that can deliver broadband services. An issue common to all-optical information transport technologies is the lack of systems education at U.S. universities. They offer excellent programs on the required materials, device, and component aspects of the field, but they seem not to have found a way to integrate into their programs the worldwide trend toward a greater emphasis on software and systems. Four major application areas of optical information transport are discussed in this section: (1) long-distance transmission, (2) fiber to the home, (3) analog transmission, and (4) optical communications in space. Long-Distance Transmission Undersea systems and terrestrial long-haul systems have been early large-scale users of optical technology. Rapid advances in both photonics and electronics are producing a wealth of new technologies that continue to significantly increase the performance of these systems. Recent progress includes the development of optical fiber amplifiers, wavelength-division multiplexing technology, photonic integrated circuits, and video compression. The vast increase in single fiber transmission capacity illustrates the rapid pace of progress and serves as a key technology roadmap for the field (see Figure 1.3). It shows exponential growth, with capacity increasing by a factor of about 100 every 10 years. Transmission at 1 terabit-per-second (Tb/s) in the research laboratory was reported in 1996. Commercial systems appear to parallel the trend of laboratory demonstrations, with a lag of 3 to 7 years. To appreciate this staggering information capacity, recall that an optical fiber is just a thin strand of glass, about as thick as a hair. Contemplate one of your hairs and note that a terabit is a million megabits. This means that at the recently demonstrated capacity of 1 Tb/s, a hairlike fiber can transmit as many as 40 million data connections (at 28 kilobaud), 20 million conventional digital voice telephony channels, or half a million compressed digital television channels. Undersea Transmission Undersea fiber systems have become the major information highways between continents (see Figure 1.4). The first large system of this kind, linking the United States and Europe, was placed in service in 1988, as the successor to the transatlantic telegraph cable, which was installed in 1858, and the transatlantic telephone cable, which began operation in 1956. The old system used coaxial cables and carried information in analog form. It had a capacity of about 40 voice circuits. The new fiber systems use digital technology, which (as in computers and CDs) provides a considerable improvement in quality; at

OCR for page 29
Harnessing Light: Optical Science and Engineering for the 21st Century FIGURE 1.3 A technology roadmap for lightwave systems, showing the transmission per fiber achieved in leading long-distance optical fiber systems over the past 15 years. The graph also indicates the progression of technologies developed for this accomplishment. Note the logarithmic scale for capacity on the vertical axis, which means that the straight trend line implies an exponential growth in fiber capacity. The growth rate is approximately a factor of 100 every 10 years, about the same as the rate of increase in the processing power of computers. Trend lines are shown both for demonstrations of experimental systems in research laboratories and for the first introduction of commercial systems. The time lag between the two is 3 to 7 years. The major multiplexing systems used are marked ETDM, OTDM, and WDM: ETDM systems use only electronic timedivision multiplexing to achieve high capacity by interleaving pulses of a multiplicity of signals; OTDM systems use high-speed optical techniques to accomplish the same interleaving; WDM refers to wavelength-division multiplexing, in which different TDM signals are transmitted by light of different wavelengths. The asterisks mark a set of recent experiments that demonstrate terabit capacity using combinations of the above techniques. (Courtesy of CA. Murray, Lucent Technologies.)

OCR for page 29
Harnessing Light: Optical Science and Engineering for the 21st Century FIGURE 1.4 A map of the current global undersea optical fiber network. By the end of 1996 more than 300,000 km of optical fiber cable were installed in this network. Note that this map was totally empty before 1988, when the first large optical undersea system was deployed. Installations have been given names, such as "Columbus," or acronyms, such as TAT (for TransAtlantic Transmission), TPC (TransPacific Cable), and FLAG (Fiber Link Around the Globe). (Courtesy of P. Runge, Tyco Submarine Systems Laboratories.) the same time, their information capacity is at least 1,000 times larger. The rapid introduction of optical fiber into global networks is illustrated by the fact that in 1991, just three years after the first optical system was introduced, fiber already carried more international digital information traffic than satellites did. More than 300,000 km of undersea lightwave cable had been installed by the end of 1996. Typical system lengths are about 6,000 km in the Atlantic and 9,000 km in the Pacific. The rapid pace of progress is reflected in the fact that we can already distinguish three different generations of the technology, as described in Box 1.3. The dramatic increase of undersea cable capacity, enabled by optical fiber technology, has led to an equally dramatic decrease in the cost of a voice circuit across the oceans. The key reason for this is that the cost of constructing an undersea system, particularly of laying an intercontinental undersea cable, has remained almost the same, even though a single cable now carries at least 1,000 times more voice circuits than it did 40 years ago. An example of this dramatic cost decrease is given in Figure 1.5, which shows the cost of AT&T transatlantic systems and indicates a cost reduction by more than 1,000 times over the 40-year period. Terrestrial Systems Optical fiber systems for long-distance transmission on land (also called "trunking") provide the links between metropolitan telephone offices, between cities, and across the continent.

OCR for page 29
Harnessing Light: Optical Science and Engineering for the 21st Century BOX 1. 3 THREE GENERATIONS OF UNDERSEA LIGHTWAVE TECHNOLOGY The 1988 first-generation undersea technology uses conventional single-mode fiber operating at a wavelength of 1.3 μm. The system uses optoelectronic regenerators with semiconductor lasers and p-i-n photodetectors both made of InGaAsP. The bit rate of the optical signal stream in each fiber is 280 megabits per second (Mb/s), and the typical regenerator spacing is 70 km. A second generation of technology was used in the 1991 installations. It too employs conventional fibers, but with the operating wavelength moved to 1.55 μm to benefit from lower fiber losses. Single-frequency distributed-feedback lasers are used in optoelectronic regenerators, together with avalanche photodetectors. The bit rate per fiber is doubled to 560 Mb/s, and the typical repeater spacing is 150 km. The third-generation technology was prepared for installation in 1995. It uses optical repeaters with optical amplifiers. The systems operate at 1.55 μm and use dispersion-shifted fiber to provide near-zero dispersion in the operating range. The design bit rate per fiber is 5 gigabits per second (Gb/s). Year Technology Generation Voice Channels per Cable 1955 COAX 1 48 1963 COAX 2 140 1970 COAX 3 840 1976 COAX 4 4,200 1988 FIBER 1 8,000 1991 FIBER 2 16,000 1995 FIBER 3 122,880 The above table shows the information-carrying capacity of these three lightwave systems (FIBER 1-3) along with that of the preceding four generations of coaxial systems (COAX 1-4). Note that the data shown are for cable capacity without voice processing. Current undersea systems use digital voice processing to take advantage of silent periods in speech. This increases effective capacity by another factor of 5, enabling the third-generation lightwave system to carry more than 500,000 voice channels. Early metropolitan applications, such as those in the San Francisco Bay area, highlighted optical transmission's advantages. The Bay Area system, introduced in 1980, provided digital transmission at 45 megabits per second (Mb/s) with a repeater spacing of 7 km. This spacing was already sufficient to allow links between telephone offices to be direct, with no need for electronics outside the buildings. The ptical system also helped to save space in metropolitan cable ducts since a single fiber replaced the earlier T1 carrier's 28 pairs of copper wires, with a capacity of 1.5 Mb/s per pair.

OCR for page 29
Harnessing Light: Optical Science and Engineering for the 21st Century FIGURE 1.5 The falling cost per circuit of transatlantic telephone (TAT) systems. (Note the logarithmic vertical scale.) The first major lightwave installation in the United States was the Northeast Corridor system, which linked Washington with New York in 1983 and New York with Boston in 1984, a total length of 747 miles. Its technology carried 90 Mb/s per fiber, the equivalent of 1,344 digital voice channels. This system confirmed the inherently higher quality of digital technology, but it also brought out another point: the excellent technical and economic synergy between the digital lightwave system and the digital telephone switches along the route. No analog-to-digital conversion was needed to interconnect the fiber system to the 23 existing digital electronic switching systems along the route, which resulted in considerable cost savings. The Northeast Corridor was soon followed by more large-scale fiber installations, using rapidly advancing technology. More than 100 million kilometers of fiber had been installed around the world by the end of 1996, about one-third of the total being in the United States (see Figure 3.2). Terrestrial fiber systems, like undersea systems, have evolved through several technology generations. (See Figure 1.3 for more detail on this rapid progress.) Advanced systems now employ optical amplifiers and wavelength-division multiplexing (WDM). WDM is a multiwavelength technique that increases the number of digital information channels that can be sent over a fiber simultaneously by transmitting each channel on its own distinct optical wavelength, as described in the section "Optical Networking and Switching" below. Terrestrial long-haul networks will benefit significantly from amplified multiwavelength transmission systems designed to access the large inherent bandwidth in the installed fiber. Capacity will increase by a factor of 10 to 50, not only providing for ample and graceful growth, but also allowing flexibility in network architecture design and in the management of network restoration in case of a failure (e.g., a cable cut). The new networking flexibility will be exploited to enable novel routing of traffic. WDM system experiments involving more than 100 channels already have demonstrated ultrahigh-capacity transmission over long distances, and large-scale experimental projects are under way to explore the potential of WDM technology for high-capacity networking. Amplified 16-channel WDM systems at 2.5 gigabits per second (Gb/s) per channel

OCR for page 29
Harnessing Light: Optical Science and Engineering for the 21st Century TABLE 1.2 Estimated Global Markets for Long-Distance and Interoffice Lightwave Systems (U.S. dollars)   1995 1996 Undersea   $2 billion $2.5 billion   · Cable and deployment: ~70%       · Lightwave equipment: ~30%     Terrestrial   $2 billion $4 billion   · SONET (OC-48) and SDH (STM-16)       · Transmission equipment only       · Lightwave: ~20%       · Electronics: ~80%     NOTE: SONET = Synchronous Optical Network, SDH = Synchronous Digital Hierarchy. (Source: T. Li, AT&T Laboratories.) have been developed and are being manufactured for massive deployment in the embedded terrestrial network. These revolutionary system solutions will meet the demand of envisioned broadband services for many years to come and thus make lightwave communications the principal component of the global information services infrastructure. Table 1.2 gives estimates of the global markets for undersea and terrestrial lightwave systems. There are strong systems vendors in Europe, Japan, and North America. Because of the early large-scale deployment of WDM systems in the United States, North American vendors have gained an early market lead in WDM systems. Fiber to the Home Although most experts agree that in the future, fiber will be installed all the way from the telephone company central office to the home, opinions vary widely as to when this will happen. Developments in the United States and elsewhere are beginning to suggest that it may happen sooner than commonly thought. The technical benefits of fiber to the home (FTTH) include its well-known capacity for transmitting incredibly high bandwidths at relatively negligible losses. This "future-proofs" the resulting network against demands for rising bandwidths, which history shows will indeed be necessary. Since fiber lasts 30 years or more, labor-intensive installation of cable and passive components has to be done only once because no electronics are installed in the outside plant between office and home. Upgrades take place on the premises of the service provider and the customer. Nearer-term benefits of fiber include its small size and weight compared with metallic cable, especially coaxial cable; its total immunity to both inward and outward leakage of electromagnetic

OCR for page 29
Harnessing Light: Optical Science and Engineering for the 21st Century FIGURE 1.20 One future market expansion for displays is in flat, light, low-volume visual interfaces for personal digital assistants, Web browsers, augmented and virtual reality engines, and portable and wearable computers. The display shown here is a full-color VGA-sized device that carries nearly the information content of a personal computer monitor. (Courtesy of Planar America.) What is needed for the potential mass market is low-cost manufacturing, low power dissipation, high contrast ratio (an advantage of emissive displays), low-complexity imaging optics (for head-mounted and projection displays), low-voltage materials and devices, and hardware or software systems for three-dimensional imagery and interactive displays. In addition, there are specific needs for some technologies, such as faster switching for nematic liquid crystals with 8 bits of gray scale and electroluminescent materials with better white phosphor (especially in the blue). Projection Displays Projection displays can be considered a close relative of the slide projector. A key difference is that the "slide" is changeable in real-time, under the control of a computer or an electronic terminal. The system projects large images onto a screen or the wall, but the technology for "changing the slide" is closely related to that of miniature LCDs. For boardroom, classroom, or family room wall-sized displays, electronic projectors using LCD or micromirror technology are being commercialized by several U.S. companies. These systems use 1.3-inch active-matrix displays made on glass for transmission displays or on a silicon substrate when used in reflection. The latter presents an opportunity for the U.S. semiconductor industry. According to the Semiconductor Industry Association's National Technology Roadmap for Semiconductors, growth in this industry will be produced by the following: Achieving finer lithographic features; Improving interconnects through on-chip optics; and Creating new applications for integrated circuits (such as silicon-based displays). Electronic projector sales exceeded $1 billion in 1996 and are expected to double every year for the next few years. The potential market may exceed that of laptop computers by 2001. In addition to the technology needs described above for miniature displays, projection technology will require extremely bright illumination sources, with small spot sizes and long service lives. Very Large Displays Very large (nonprojection) displays are designed to deliver 20- to 40-inch images for such applications as high-end graphics workstations

OCR for page 29
Harnessing Light: Optical Science and Engineering for the 21st Century and future HDTV. Currently this is a niche market, but it has the potential to grow into a mass market. Conventional LCDs and CRTs are not expected to be scalable to these large sizes at a reasonable cost. Two flat-panel technologies now under development promise to reach 40-inch size, plasma displays and plasma-addressed LCDs. The United States has a technology edge in large plasma displays, and the development of a 21-inch full-color display with 1280 x 1024 resolution was announced in 1995. Plasma-addressed LCDs were invented in the United States (at Tektronix) but have already been licensed to Sony. As for other display technologies, a challenge for U.S. industry is to develop a high-volume, low-cost manufacturing capability before the mass market develops. Military and Avionics Displays Military and avionics displays are an important niche market for U.S. industry. With the increasing emphasis on graphics, images, and video information, the U.S. military finds displays increasingly critical. Typical applications are medium-sized displays for the digital battlefield and cockpit displays, as well as miniature devices for simulators and tank and gunsight displays. Although medium-sized displays are produced for the mass consumer market, it appears impractical to insert this technology directly into military applications, which have to work in a wide range of light levels and whose construction must be more rugged than that of civilian products. Government support of the flat-panel display industry has provided strong innovative impulses and has been a critical element in establishing a strong technology position in the current niche markets for small displays and very large displays. This could provide the catalyst and basis for U.S. reentry into the commercial and consumer markets, particularly if niche markets show fast growth and become mass markets. It should be stressed, however, that successful reentry will depend critically on the development of a high-volume, low-cost manufacturing capability. Note that the market for military displays is no more than 5% of the worldwide mass display markets. These mass markets can generate $3 billion to $4 billion annually in R&D resources that can be used to place their low-cost manufacturing on a fast learning curve. U.S. efforts toward reentry have to be measured against these vast resources. Educational and R&D Issues Display technology is a major outlet for the U.S. research community in optics and condensed-matter and materials physics. This community continues to create potential breakthrough technologies, but there is only relatively weak coupling of materials and device work to programs on display systems and applications. In fact, this study has not

OCR for page 29
Harnessing Light: Optical Science and Engineering for the 21st Century identified any course at a U.S. university in which display technology is taught in grand perspective, including materials, device, systems, and applications issues. In view of the growing importance of displays, there is a need to develop such cross-disciplinary educational curricula. Academic research on displays is limited to less than a dozen U.S. universities. The most notable academic program is at Kent State University, which has focused on building a center of excellence in liquid crystal technology. Other noteworthy university programs include Princeton University (organic luminescent displays), the University of Michigan (AMLCDs), the University of Colorado at Boulder (ferroelectric LCDs and liquid crystal on silicon), Georgia Tech (liquid crystal on silicon), and the Massachusetts Institute of Technology (liquid crystal on silicon). It is often claimed that U.S. universities find it difficult to obtain research contracts for display work. Most U.S. industrial display R&D is pursued by small and medium-sized companies, with a focus on military and avionics applications. These programs are concentrated on CRT, plasma, electroluminescent, field emission, and liquid crystal technologies. There is widespread consensus that the United States should try to reenter the mass display markets. However, there is no well-understood, broadly supported, well-coordinated strategy involving industry, universities, and the government that has the goal of accomplishing this task. Summary and Recommendations Information Transport Optical technology overwhelmingly dominates the long-distance transmission of information. More than 100 million kilometers of optical fiber have already been deployed worldwide, and deployment continues at the rapid rate of greater than 20 million kilometers per year (more than 2,000 km per hour, which is faster than Mach 2). Moreover, the transmission capacity of a single fiber has increased exponentially by about a factor of 100 per decade. A capacity of 1 Tb/s has been demonstrated in the research laboratory, and 20-Gb/s and faster systems are being deployed commercially. The introduction of optical amplifiers and wavelength-division multiplexing has been essential to recent advances, and these two key enablers have considerable potential for further innovation and application to WDM networking. Various optics-enabled systems are being evaluated worldwide for broadband applications in local exchange access. These include analog lightwave transmission, fiber to the curb, and fiber to the home. Over the next 20 years, the installation of a broadband local exchange access infrastructure will require an estimated $150 billion in capital

OCR for page 29
Harnessing Light: Optical Science and Engineering for the 21st Century investment in the United States alone and probably three to five times as much worldwide. This is a large-scale challenge for optics and associated technologies. The local access market is considerably larger than that for long-distance technology. Fiber to the home (FTTH) is generally considered the ultimately desired broadband technology. It offers a future-proof installation that can be adapted for most conceivable broadband services once demand for them develops. To make FTTH viable, cost reduction is needed in several key enabling technologies: low-cost lasers, low-cost packaging, low-cost manufacturing, and passive network topologies that allow cost sharing and reduce operational costs. FTTH will make it possible to avoid outside plant powering, which will result in significant savings in cost and operations. Lifeline telephony provided by FTTH will not function at home during power failures unless special provisions are made for powering. Increasing access bandwidth is a global challenge for the future. It will enable society to enjoy a variety of future information services. The rapid emergence of broadband telecommunications networks and broadband information services will have an enormously stimulating impact on commerce, industry, and defense worldwide. Congress should challenge industry and its regulatory agencies to ensure the rapid development and deployment of a cost-effective broadband fiber-to-the-home information infrastructure. Only by beginning this task now can the United States position itself to be world leader in both the broadband technology itself and its use in the service of society. The technology for analog lightwave transmission is improving rapidly and is enabling systems of lower cost and improved performance. This progress is broadening the range of applicability from cable television distribution to remote links for mobile radio cellular systems, as well as microwave and potentially millimeter-wave systems. Information transmission via laser beams promises lower cost and higher capacity for future links between satellites. The required know-how for this optical technology does exist at U.S. universities and federally funded laboratories; however, unlike their counterparts in Europe and Japan, U.S. industry and the U.S. government have not invested in the development of experimental optical payloads that would allow realistic economic assessment and timely market entry. Although U.S. universities are quite strong in the physics, materials, and component areas of information transport technology, they are relatively weak in applications and systems. There is a need for better linkage of university device research to systems research, in order to motivate and guide the former and speed the development of the latter. There are considerable research opportunities in component integration to make WDM technology more cost-effective and broaden its applicability. Low-cost optoelectronic packaging is essential for reducing system costs and should

OCR for page 29
Harnessing Light: Optical Science and Engineering for the 21st Century be considered an integral part of device research. Funding agencies should support better linkages between university researchers and industry and between research on devices and on multidisciplinary systems. Processing Data Links Low-cost optical data links are on the verge of becoming practical in local area networks that link together personal computers and workstations, file servers, printers, and data storage systems within computing clusters. To date, however, optical data links have achieved only modest market penetration, since copper is usually cheaper. The Optoelectronics Industry Development Association (OIDA) expects annual revenues for data communications equipment to reach $25 billion by 2003, with $7 billion of the total being for optoelectronic equipment. By 2013, OIDA expects the total market to double and the optoelectronic component to reach $30 billion. This market is global, and low cost is crucial. Further market penetration requires cost reductions, especially in manufacturing connectors and cables; packaging and alignment; improved molding or plastic packaging; and modeling and simulation tools for high-frequency, low-cost packages. Many optical data links employ parallel technology, using arrays to achieve economies of scale. The challenge is to enable research in the new R&D environment, in which low-cost, high-volume global markets mean low marginal profits. Strong university research must continue and should be focused on longer-term issues that are unlikely to be addressed by industry. Collaboration between industry and universities is vital. Optical Networking and Switching As telecommunications bandwidth increases, the cost of switching is becoming an ever-larger fraction of the cost per bit, and low-cost switching technology is becoming increasingly important. The increasing importance of low-cost switching will motivate the exploration of WDM and the use of optics in high-density logical routing switches. Switching system devices and concepts are in an early stage of development, so there are still major opportunities in materials and devices. There is strong competition from Europe and Japan in both WDM technology and high-density switching fabric technology. Experimental switching systems using WDM technology are being tested in Europe, Japan, and the United States. WDM systems use integrated optics crossbar switches to route different colors of light in different directions. These switches route the entire optical signal and thus are independent of the modulation format or the number of multiplexed signals. Network management and control systems are the major need

OCR for page 29
Harnessing Light: Optical Science and Engineering for the 21st Century for commercial acceptance of WDM optical networking and switching. Standards are necessary. Issues to be resolved include simplicity versus complexity, components, and architectures. Practical wavelength conversion devices are needed to allow full flexibility in WDM systems. High-density optical interconnect switching is a systems approach based on the current technology of electronic logical switches. Optics is introduced for routing within electrical logical switches to overcome the electrical limitations in high-density switches. The technology requires two-dimensional arrays of integrated receivers, transmitters, and smart pixels. The use of optics in high-density switching fabrics is still in the research stage. Component development and systems integration studies are under way. Significant deployment of optics in switching may start in about 5 years, be important in 10, and be crucial for successful systems in 15. Market size projections cannot yet be made, however, since the technologies are too immature. Purely optical switching has not yet become a reality because of difficulties in buffering, memory, and logic. In addition, ultrafast optical switching devices require too much power, are too simple logically, and are not yet practical for large-scale use. All-optical switching systems are unlikely to be practical unless there is a breakthrough in technology. Nonlinear materials, fibers, and devices that could lead to breakthrough technologies are excellent university research topics. Additional opportunities lie in exploiting ultrafast optical technology, such as short-pulse sources for WDM. Continued strong investment is recommended, with research carried out in both universities and industry. Optical Image Processing and Computing Optical logic has great difficulty in competing with electronics for computing since electrons perform logic much more easily than light. Breakthrough technologies will be required before general-purpose optical computing is likely to be important. Applications for optical processing are in correlators and niche applications for machine vision and pattern recognition. High-speed silicon provides strong competition for optical processing in many cases. Optical image recognition can be commercially viable, however, particularly in rare-event problems. The future of research on optical processing and computing lies in the application of technology to practical problems, with a premium on low cost. The development of practical hardware can come only by close interaction between device research and systems research. General University education and research must do a better job of interrelating devices and systems. Better interrelation of university and industry research is also necessary. Research should focus on cost reduction of

OCR for page 29
Harnessing Light: Optical Science and Engineering for the 21st Century optical and optoelectronic components, packaged subsystems, and full systems. It should strive for a seamless merger of optics and electronics via improved systems integration and device integration. Optoelectronic device development should be driven by systems needs, and system design by device realities. Designers of optical information processing systems should more fully exploit wavelength, space, and time. University research should be carried out with an understanding of industrial needs. Collaboration between universities and industry should be encouraged, and issues that stand in the way, such as intellectual property, should be addressed on a national level. It would be helpful for students and faculty to do part of their research on-site in industry. Device research and development should be carried out in a systems context. Although this usually occurs in industry, it must happen more in universities. Industry can assist universities by donating or lending components, enabling university researchers to carry out systems studies. Storage The requirement for digital information storage is growing at an enormous rate; it is expected to exceed 1020 bits in the year 2000, with an estimated annual market for data storage in excess of $100 billion. Of this market, optical storage will be about $12 billion. Each year, the number of bytes of shipped magnetic disk storage doubles; 125 petabytes shipped in 1996 will become 2 exabytes in the year 2000. These numbers do not include consumer-based electronics and entertainment storage (i.e., CD and DVD technology), which will add another $20 billion in drive sales alone with media sales exceeding $40 billion. Much of today's information is distributed and stored on paper, perhaps as much as 90% of it in the form of newspapers, magazines, and books—15 million tons of paper will be used annually by the year 2000. An increasing amount of the original information for these publications, however, is being captured electronically. Multimedia, parallel, and network-centric computing is increasing the need for larger, faster, cheaper storage. Magnetic disk storage density is currently at 2.6 Gb/in.2 and is growing at an annual rate of 60%. This translates into lower cost and higher performance. It is estimated that the cost of storage in the year 2000 will be 1 cent per megabyte and in the year 2005 less than 1/4 cent per megabyte. The limit for magnetic storage density is unknown at present but is estimated to be in excess of 100 Gb/in.². Optical storage densities have increased by a factor of seven, from 0.39 Gb/in.2 (CD-ROM) to 2.7 Gb/in.² with the recent introduction of DVD-video, and will continue to increase to 9.8 Gb/in.² when double-sided, double-layer DVD-ROM disks are available in a few years.

OCR for page 29
Harnessing Light: Optical Science and Engineering for the 21st Century CD-ROMs have already replaced floppy disks as the preferred medium for the distribution of programs, video games, and reference material. DVD technology with 4.7 to 17 GB capacity per disk is available to allow storage of a full-length movie on a single disk. Recordable CD technology is already available, with the DVD version not far behind. Extensions to CD and DVD technologies share many of the same pathways to higher density and have the added value of commonality with a large installed base. The inherent removability of optical storage media has made optical library systems an important part of the storage hierarchy, providing reasonably rapid on-line access to large databases. These systems are often used for archiving data. A large fraction of the data stored in optical libraries is on WORM media that use ablative, phase change, or dye polymer materials. Erasable libraries generally use magnetooptical materials. Read-only CD libraries are available that will transition to WORM as the use of CD-recordable technology become more widespread. DVD is expected to follow a similar product path. Erasable CD and DVD systems are on the horizon. The market for magnetic disk storage is dominated by U.S. industry, which has an 80% share of worldwide revenue. The market for CD-ROM drives, conventional optical drives, and CD libraries is dominated by foreign companies, with greater than 90% of the market. There is essentially no CD-audio player industry in the United States. It should be noted that even though U.S.-based companies dominate the magnetic disk market, all major manufacturers build their drives offshore. Three-dimensional storage, such as holographic, ETOM, and two-photon, has the potential to provide inexpensive, rapid access to large databases. The key to success is identifying and optimizing the proper materials systems. Storage and retrieval of information as large blocks (pages) of data can provide very high transfer rates in excess of 1 Gb/s. These storage technologies are ideally suited for the storage of digital images such as medical images, satellite images, geophysical data, and movies. The enabling technologies for information input (e.g., spatial light modulators) and information output (e.g., CCD arrays) are generally available. Most R&D on three-dimensional storage is performed in the United States; although many of the programs are exploratory and some have stopped short of fruition, two major industry-university consortia are in place to commercialize holographic recording. The unique requirement of specific market segments for systems with special performance characteristics is a driving force that will allow optical storage products to penetrate niche markets. Information storage of all formats, including electronic, optical, film, and paper, will continue to grow unabated, and storage products have enormous market potential. Magnetic recording will be the midto high-end, high-performance technology of choice for the foreseeable

OCR for page 29
Harnessing Light: Optical Science and Engineering for the 21st Century future. Extensions of the optical CD and DVD technology will capture a substantial share of the low- to midrange storage market. CD and DVD technologies will start to penetrate the high-end library storage market. Low drive cost and media standards that promote interchangeability and backward compatibility will be key to the evolving success of CD and DVD. U.S. industry is not playing a leadership role in the development of conventional, CD, and DVD optical storage systems. Although there is a window of opportunity between 2000 and 2005 for the deployment of three-dimensional storage techniques, they must compete on a cost-per-megabyte basis with the aggressive evolution of magnetic and optical (CD and DVD) technologies in the same time frame. The United States has a lead of about 2 to 3 years in the development of three-dimensional storage. To be successful, an optical storage product need not displace all other forms of storage but can coexist with them by providing cost-effective solutions to a subset of the market. Optics is a key enabling technology in the information storage marketplace. To push CD and DVD technologies to higher effective storage densities and performance levels, U.S. industry should develop multilayer storage media; low-cost optical systems for writing and reading data; and efficient, low-cost techniques for mass replication and assembly of multilayer disks. To retain the U.S. technological edge in three-dimensional recording, industry and universities should nurture and accelerate the development of advanced three-dimensional recording media, the design of low-cost optical systems, and the study of systems integration and architectures. It is imperative that these activities be coordinated among university and industrial researchers. DARPA should establish a program to seek new paradigms in optical storage that will reach toward the theoretical storage density limit of about 1.0 TB/cm3, with fast (>1 Gb/s) recording and retrieval. Displays Display technology is critical for the development of new information systems and services. Major change is transforming the display market and creating new opportunities for innovation. Once dominated by the CRT, the market is now split into two mass markets of about $20 billion each, plus a few niche markets. The mass market for large displays (greater than 15 inches, mostly televisions) is still dominated by the CRT, while the mass market for medium-sized displays (less than 15 inches, mostly computers) is dominated by LCDs. The U.S. military finds displays critical and has to maintain core competence in this field; however, military needs support only a small niche of the display market.

OCR for page 29
Harnessing Light: Optical Science and Engineering for the 21st Century No major innovations in CRT technology are expected, and scaling CRTs to very large sizes (greater than 40 inches) is expected to be difficult and costly. There are two approaches to overcoming this barrier for large workstation, simulator, and television applications: plasma displays and plasma-addressed LCDs. In medium-sized displays, there is a convergence of the computer and television formats. The industry standard is the active-matrix LCD, manufactured for 10.4-inch laptop computer applications. More than 10 million such displays were sold in 1996. The future trend for AMLCDs will be to increase their size to 12.4 inches and eventually to 14.1 inches. There is no clear winner yet in the race to develop a high-quality miniature display for head-mounted, handheld, and projection applications. Candidates include emissive technologies such as active-matrix electroluminescence, field emission, and liquid crystal on silicon; reflection-mode technologies include digital micromirrors and micromirror or grating displays. For head-mounted and handheld applications, R&D on small displays should focus on ergonomics. Before head-mounted displays are accepted by the consumer, they will have to be much lighter and more comfortable, and they should not cause nausea when used. Improved optical system designs have to be developed to make reflection displays as cost-effective and high in contrast as their transmissive and transmission-mode counterparts. Improvements are needed in many display technologies, including better white phosphors, faster-switching nematic liquid crystals, better lamps, and brighter blue LEDs. Although many currently used display technologies were invented in the United States, their development is for the most part carried on overseas, as is 95% of display manufacturing. The United States still has a lead in the development of very large and very small emissive and reflective display technology. Translating U.S. inventions into a share of the global market requires further R&D on display systems and applications. This competence is currently scattered and limited to a handful of U.S. universities and less than a dozen small and medium-sized companies. Because of the rapid learning curve and the large R&D resources (about $2 billion) generated by the mass market—heavy Japanese and Korean investment in low-cost manufacturing technology is expected to result in improved performance while lowering prices by at least 20% per year—it will be extremely difficult to displace liquid crystals from the mass market for medium-sized flat-panel displays. Major opportunities exist for new technologies to enter the niche markets for small displays, projection displays, and very large displays, and in the long term, the U.S. lead in these niche technologies, leveraged by investment in military displays, may establish a base for U.S. reentry into the mass consumer and commercial markets.

OCR for page 29
Harnessing Light: Optical Science and Engineering for the 21st Century The United States should devise and develop a broadly based and well-understood strategy to capture a significant share of the future global display market. This will require a well-coordinated effort that brings together partners from government, universities, and industry. Important components of such a strategy include timely development of a low-cost manufacturing capability, ensuring a core competence in displays for military needs, and developing multidisciplinary university courses and curricula on display systems and applications. Other key elements are building a broad consensus on the evaluation, roadmapping, and prioritization of promising display technologies and improving the coupling of the device, systems, and graphics software communities. The strategy should consider, among other options, U.S. mass market reentry via current market niches (e.g., very small or very large displays) in which the United States has a strong technical position.