7
How Will the Information Technology Revolution Be Extended?

Extrapolation of Moore’s law suggests that in the next 20 to 30 years, electronic circuit elements will shrink to the size of single atoms. Even before this fundamental limit is reached, electronic circuits will have to operate in a new regime in which quantum mechanics cannot be ignored. New approaches to communications and information processing will have to be invented, and condensed-matter and materials physics (CMMP) will work with other disciplines to enable this transition. Among the many avenues already being explored in CMMP are devices based on spin rather than charge, molecular-scale circuit elements fashioned from carbon nanotubes, and novel computational engines based on biomolecules such as deoxyribonucleic acid (DNA). Perhaps most exotically, quantum information science envisions computation and communication based not on the familiar laws of classical physics but instead on the often counterintuitive laws of quantum mechanics. The familiar binary “bits” of today may tomorrow be replaced by arrays of quantum bits, or “qubits,” capable of encoding vastly more information. CMMP, the science that helped launch the information age, will play a pivotal role in determining its future.

THE ROAD AHEAD

Information technology (IT) pervades modern life. It is worth trillions of dollars per year in the global economy.1 About half of the productivity growth in

1

Dale W. Jorgenson, Econometrics, Volume 3: Economic Growth in the Information Age, Boston, Mass.: MIT Press, 2002.



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 127
7 How Will the Information Technology Revolution Be Extended? Extrapolation of Moore’s law suggests that in the next 20 to 30 years, electronic circuit elements will shrink to the size of single atoms. Even before this fundamental limit is reached, electronic circuits will have to operate in a new regime in which quantum mechanics cannot be ignored. New approaches to communications and information processing will have to be invented, and condensed-matter and materials physics (CMMP) will work with other disciplines to enable this transition. Among the many avenues already being explored in CMMP are devices based on spin rather than charge, molecular-scale circuit elements fashioned from carbon nanotubes, and novel computational engines based on biomolecules such as deoxyribonucleic acid (DNA). Perhaps most exotically, quantum information science envisions computation and communication based not on the familiar laws of classical physics but instead on the often counterintuitive laws of quantum mechanics. The familiar binary “bits” of today may tomorrow be replaced by arrays of quantum bits, or “qubits,” capable of encoding vastly more information. CMMP, the science that helped launch the information age, will play a pivotal role in determining its future. THE ROAD AHEAD Information technology (IT) pervades modern life. It is worth trillions of dollars per year in the global economy.1 About half of the productivity growth in 1 Dale W. Jorgenson, Econometrics, Volume : Economic Growth in the Information Age, Boston, Mass.: MIT Press, 2002. 

OCR for page 127
c o n d e n s e d - m at t e r m at e r i a l s P h ys i c s  and the U.S. economy is now attributed to information technology. Seminal inventions such as the transistor, the hard disk drive, and the communications laser eventually enabled the rise of the Internet, a watershed event in modern history. Indeed, some have compared the impact of the Internet to that of the printing press in extending broad access to information that was previously available only to the privileged. The great force behind this modern industrial revolution has been miniaturiza- tion—the repeated shrinking of the devices that process, store, and communicate information. To extend the IT revolution, new devices will have to be invented. Smaller devices tend to be faster. More important, they tend to be cheaper because many more devices can be manufactured at the same time. More devices per dollar mean more function per dollar, and ever-more-affordable function has enabled the industry to expand its products from yesterday’s mainframe computers to today’s bewildering array of consumer products. Virtually every electronic product con- tains at least one microprocessor, and game machines now pack the computational power of supercomputers from just a few years ago. The devices of IT have not just “shrunk.” They have shrunk in an exponentially compounding fashion. In 1965, it was noted that the number of transistors that could be built on a single silicon chip was doubling every few years. This doubling trend has roughly persisted over the past 40 or so years, elevating the trend to the status of a “law”—Moore’s law. But Moore’s law is not unique in IT. Analogous exponential doubling trends have been noted for information storage capacities of hard disk drives, digital communication rates, and many other key performance indicators of IT. Over time, such powerful trends can change the world. For ex- ample, since the introduction of the hard disk drive about 50 years ago, the areal density of information storage has increased by a factor of roughly 100 million. The resultant dive in the cost of information storage has been a key enabler for the rise of the Internet and the explosive growth of electronic commerce. What is required to drive such powerful trends? It is the repeated reinvention of the devices that store, process, and communicate information. For instance, the introduction to hard disk drives in 1998 of read-head sensors based on the giant magnetoresistance (GMR) effect greatly accelerated the above-mentioned doubling trend in information storage capacity. GMR is a subtle collective effect in magnetic materials, unknown to physics before 1988. It is striking that GMR was put to practical use in almost all computers manufactured worldwide by the late 1990s. Harnessing the effect for information technology involved the develop- ment of practical methods for manufacturing multilayered structures consisting of thin ferromagnetic films separated by metallic spacers of precisely controlled nanometer-scale thickness. A magnetic field passing through such a structure deter- mines whether the magnetic moments of adjacent ferromagnetic films are aligned parallel or antiparallel. Through the phenomenon of spin scattering, this alignment gives rise to low and high states of electrical resistance, respectively, corresponding to the 0 and 1 states of binary digital information. Thus, this “spin valve sensor”

OCR for page 127
how will i n f o r m at i o n t e c h n o lo g y r evo lu t i o n B e e xt e n d e d ?  the can translate the changing reversals of magnetization stored on a spinning hard disk surface into the electrical signals needed for digital information processing. Technologists found that these sensors could be readily miniaturized, allowing ever- smaller regions of magnetization to represent information on a hard drive. These developments contributed substantially to the huge drop in the cost of information storage in the past 15 years and the rapid rise of the Internet. Now, sparked by new developments in condensed-matter and materials phys- ics, magnetic sensors for hard disk drives are being reinvented again. Figure 7.1 shows a schematic of a GMR spin valve structure as originally utilized in the late 1990s and a schematic of the new magnetic tunnel junction (tunnel valve) sensor that is enabling disk drive manufacturers to further reduce the cost and increase the speed of information storage. Replacing the metallic spacer of the GMR structure with an electrical insulator means that electrons can pass between the two ferro- magnetic layers only by quantum tunneling. Recent advances in materials physics now allow the tunneling rate to be made exquisitely sensitive to a magnetic field, providing a new generation of sensors for hard disk drives. This complex interplay between the discovery of new physics and progress in miniaturization is not limited to magnetic sensors for hard disk drives. Figure 7.2 shows the current transition to higher-storage-density perpendicular magnetic recording from the traditional longitudinal geometry. This transition is being aided by the above-described improvements in magnetic sensors, but it also depends on FIGURE 7.1 (Top) Giant magnetoresistance (GMR) spin valve with the current flowing in the plane of the film. Two ferromagnetic layers are separated by a nonferromagnetic metallic spacer layer. The magnetoresistance is higher when the two magnetic layers have their magnetization aligned in parallel (high state, or binary “1”) and lower when antiparallel (low state, or binary “0”). (Bottom) A tunnel magnetoresistance (TMR) spin valve is indicated where the spacer is a nonmagnetic insulator and the current flow is perpendicular to the plane of the film structure. SOURCE: Courtesy of Hitachi Global Storage Technologies, Inc.

OCR for page 127
c o n d e n s e d - m at t e r m at e r i a l s P h ys i c s 0 and FIGURE 7.2 (a) Schematic illustration of a longitudinal recording system with a read and write head flying above the recording medium: t is the medium thickness; W, the width of the recorded track; and B, the bit size. (b) Schematic of magnetic alloy longitudinal recording layer showing amorphous grain boundaries as white and an average grain size of 8.5 nm. (c) Schematic illustration of a perpendicular recording system with a read and write head flying above the recording medium. The magnetization of the media is perpendicular to the films, and a soft magnetic underlayer is below the medium to enhance the perpendicular write fields from the head. SOURCE: Image courtesy of E. Fullerton, Uni- versity of California at San Diego; based on G. Srajer, L.H. Lewis, S.D. Bader, A.J. Epstein, C.S. Fadley, E.E. Fullerton, A. Hoffmann, J.B. Kortright, K.M. Krishnan, S.A. Majetich, T.S. Rahman, C.A. Ross, M.B. Salamon, I.K. Schuller, T.C. Schulthess, and J.Z. Sun, “Advances in Nanomagnetism Via X-Ray Techniques,” J. Magn. Magn. Mater., 307, 1-31 (2006).

OCR for page 127
how will i n f o r m at i o n t e c h n o lo g y r evo lu t i o n B e e xt e n d e d ?  the other distinct and dramatic advances in the understanding of micromagnetic ma- terials and structures. Indeed, the development of magnetic storage, the transistor, the optical communications laser, and other key devices of IT are all marked by the repeated introduction of new materials and by improved device structures based on new physical principles. Over the long run, even the key devices have themselves been repeatedly sup- planted and replaced by new (smaller, faster, cheaper) devices. Before there was random access memory based on the silicon transistor, magnetic core memories were dominant. Today they are found only in museums. The mechanical mecha- nisms of the first commercial tabulators were replaced by electromechanical relays and then by vacuum tubes. Commercial computing began to take off when tubes were replaced by discrete bipolar transistors and then by integrated circuits based first on bipolar transistors and then on more efficient field-effect transistors (FETs). History thus suggests that the key devices of IT may be reinvented again. There are also good scientific reasons to believe that this can happen; in order to extend the growing benefits of the IT revolution for several more decades, researchers must strive to make it happen. This is because the “old” established devices are nearing their physical limits. According to the International Technology Roadmap for Semiconductors,2 hard limits to the further miniaturization of the transistor are fast approaching. In par- ticular, undesired power dissipation, such as that caused by quantum mechanical tunneling, is already limiting further advances in speed; many believe it will limit further advances in miniaturization within a decade. Similarly, the physics of mag- netism is beginning to limit advances in hard disk drive technology. In particular, the volume of the magnetized regions that store each bit (“0” or “1”) of digital information cannot be shrunk below what is known as the superparamagnetic limit. Smaller magnets cannot retain their magnetic order long enough to be use- ful for the permanent storage of information. Figure 7.3 shows a high-resolution magnetic force microscopy image of magnetic bits tightly arranged along tracks in a magnetic recording medium. To extend the IT revolution, new devices for processing, storing, and com- municating information will have to be invented. The new devices will very likely be based on new materials and exploit physical principles that condensed-matter physicists are just beginning to explore. Some devices may operate in a new regime in which quantum mechanics will be embraced. New molecular-scale structural elements such as carbon nanotubes and organic spin valves and semiconductor nanowires and nanocrystals are of great interest for the new devices. This is because low-atomic-number elements, such as carbon, should cause less undesirable spin- 2Available at http://www.itrs.net/Links/2006Update/FinalToPost/00_ExecSum2006Update.pdf; last accessed September 17, 2007.

OCR for page 127
c o n d e n s e d - m at t e r m at e r i a l s P h ys i c s  and FIGURE 7.3 Have you ever seen a magnetic bit? Shown here is a high-resolution magnetic force microscopy image of magnetic bits arranged along tracks, as in magnetic recording media. The upper track is a square-wave bit pattern written at 1024 kilo-flux reversals per inch (kfci), and the bottom track is at 240 kfci. The upper track corresponds to a transition-to-transition distance of only 24.8 nm, which matches the maximum transition density used in a 36 Gbit/cm2 areal density demonstration. For comparison, the micron marker is the same length that a magnetic bit was in 1989. The white lines in the box, generated by an analysis program, indicate the average transition shape. The arrow indicates the direction of the head motion during the write process. SOURCE: A. Moser, C. Bonhote, Q. Dai, H. Do, B. Knigge, Y. Ikeda, Q. Le, B. Lengsfield, S. MacDonald, J. Li, V. Nayak, R. Payne, M. Schabes, N. Smith, K. Takano, C. Tsang, P. van der Heijden, W. Weresin, M. Williams, and M. Xiao, “Perpendicular Magnetic Recording Technology at 230 Gbit/in2,” J. Magn. Magn. Mater., 303, 271-275 (2006). orbit scattering in spin-based device structures. Thus, the information contained in electron spins can be propagated over longer distances. Figure 7.4 shows one ex- ample of how this new physics is being explored for potential applications in IT. Novel computational strategies based on the self-organizing properties of biomolecules such as DNA have also been demonstrated, and approaches based on the manipulation of molecular configuration are being explored. Perhaps most exotically, quantum information science envisions computation and communica- tion based not on the familiar laws of classical physics but instead on the some- times counterintuitive laws of quantum mechanics. The familiar binary “bits” of today may tomorrow be replaced by arrays of quantum bits, or “qubits,” capable of encoding vastly more information. Condensed-matter and materials physics, the science that helped to launch the information age, will likely play a pivotal role in determining its future. Thus, the next decade is one of transition as this burning issue is confronted: How will we extend the IT revolution? What device will allow the storage of mas- sive amounts of information more compactly than can be done with the hard disk drive? What device will allow the processing of information faster and cheaper than can be done with the silicon FET? These ubiquitous devices are the fruits of basic research in condensed-matter and materials physics stretching back to the 1930s and 1940s. The current form of these devices is the result of many decades of R&D

OCR for page 127
how will i n f o r m at i o n t e c h n o lo g y r evo lu t i o n B e e xt e n d e d ?  the FIGURE 7.4 Spintronic device example involving the detection of the spin Hall effect utilizing a lateral spin transistor-type geometry. (a) An atomic force microscopy image of the device, which consists of a thin aluminum Hall cross that is oxidized and in contact with two ferromagnetic electrodes of differ- ent widths (FM1 and FM2). (b) The measurement, where a current I is injected out of FM1 into the Al film and away from the Hall cross. A spin Hall voltage (VSH) is measured between the two Hall probes. VSH is caused by the separation of up and down spins due to the spin-orbit interaction in combination with a pure spin current. The top panel of (c) shows the spatial dependence of the spin-up and -down electrochemical potentials μ, while the black line represents μ in the absence of spin injection. The bottom panel of (c) illustrates the associated spin current Js. The polarized spins are injected near x = 0 and diffuse in both Al branches in opposite directions. The sign change in Js reflects the flow direction. (d) A spin-transistor measurement for device characterization, where I is injected out of FM1 into the Al film and away from FM2. An output voltage V is measured between FM2 and the left side of the Al film, where (e) depicts the same quantities as in (c). SOURCE: S.O. Valenzuela and M. Tinkham, “Direct Electronic Measurement of the Spin Hall Effect,” Nature, 442, 176 (2006).

OCR for page 127
c o n d e n s e d - m at t e r m at e r i a l s P h ys i c s  and and manufacturing experience. Many engineers and scientists believe that further rapid improvements will be possible for another 10 years or more. Replacing them with devices that can be smaller, faster, cheaper, and more reliable is an awesome challenge, but there are good scientific reasons to believe that this challenge can be met (see Figure 1.6 in Chapter 1). Doing so in a timely fashion will require increased investment and societal focus in the coming decade. NEW DEvICES FOR MASS STORAgE OF INFORMATION Since the physics of magnetism sets limits on the further development of hard disk drive technology, future devices for the mass storage of information may store information differently. “Scanning-probe storage” holds the promise of storing in- formation at a bit density approaching the density of atoms on a surface—perhaps a thousand times the areal density that will be obtainable with magnetic storage. Since the Nobel Prize–winning invention of the scanning tunneling microscope in the 1980s, a plethora of scanning probe techniques for the atomic-scale imaging, characterization, and modification of surfaces have been developed by condensed- matter physicists. While laboratory instruments already demonstrate a very limited ability to write and read information with atomic resolution, commercial develop- ment sacrifices “ultimate” storage density in favor of speed and reliability. To overcome speed limits in reading and writing information with a single scanning probe tip, arrays of thousands of tips are being developed, along with the multiplexed electronics to read and write from many tips in parallel. Researchers have reported rapid and reliable operation of many parallel scanning probe tips at storage densities of 120 gigabits (Gb; 120 billion bits) per square centimeter— several times the storage density of today’s hard disk drives. Operation of single tips at storage densities of 0.6 terabits (Tb; 0.6 trillion bits) per square centimeter sug- gests that the technology can be extended. For such prototypes, the scanning probe tips store information in the form of tiny indentations in a thin polymer film. Various industrial and academic research groups are pursuing a variety of other classes of materials and storage mechanisms. Experimental storage densities of sev- eral hundred trillion bits per square inch have been demonstrated by some research groups. However, understanding the atomic-scale interactions between probe tips and a variety of storage media and storage methods is key to realizing the enormous promise of this nascent information storage technology. That challenge will keep members of the CMMP community busy for many years to come. NEW SOLID-STATE MEMORY DEvICES Solid-state memory devices are used for information storage when hard disk drives are too slow or bulky. Today there are several distinct varieties of solid-state

OCR for page 127
how will i n f o r m at i o n t e c h n o lo g y r evo lu t i o n B e e xt e n d e d ?  the memory in widespread use. Static random access memory (RAM) is used when speed is of paramount importance. A circuit consisting of six FETs stores one bit of digital information. Static RAM retains information only as long as power is supplied. Dynamic random access memory is the dominant memory in computers. Slower than static RAM but still far faster than a hard disk, dynamic RAM repre- sents information as electronic charge stored on a capacitor integrated with a single FET. Dynamic RAM loses its information rapidly through charge leakage and must be intermittently refreshed while in use. Like static RAM it loses its information as soon as the power is turned off. Nonvolatile memory represents information as electronic charge trapped in insulating films built into specially designed FETs. Charge-retention times are typically about 10 years. This is the memory that is used in the memory cards of commercial MP3 players, digital cameras, and many other consumer products. All of these memory devices are based on the silicon FET, and all represent (or store) digital information in the form of packets of electronic charge. All are thus limited in their potential for further miniaturization by the same basic physical principles. The emerging memory devices being explored by condensed-matter physicists and electrical engineers will not have the same physical limits because they represent the information in physically different ways. Ferroelectrics provide one example of a solid-state memory device. These materials possess ordered electric dipoles. Ferroelectric memory devices are based on complex oxides that can exhibit large polarizations under the influence of an electric field. A digital “0” or “1” is represented by a persistent polarization cor- responding to a structural change in the material in which a central ion in the building block of the crystal structure moves to an off-center location. Ferroelectric memories are already used in some consumer devices, such as smart cards. They are also being explored as a promising medium for scanning probe information storage. There are many scientific challenges in understanding and controlling ferroelectric effects in nanoscale structures. The issue of perfecting thin insulating barriers of high-dielectric-constant materials is a challenge shared by the magneto- electronics community in its quest to harness the magnetic tunnel junction effects (described below). More broadly, it encompasses the issue of exploring coherent interfaces and the physics of hybrid structures in general. Magnetic random access memory represents a digital “0” or “1” by the high or low electrical resistance of a magnetic tunnel junction. The basic concept of the magnetic tunnel junction goes back decades, but continuous and rapid advances in the physics of nanostructured magnetic materials are just beginning to make it practical as a memory device. The first commercial magnetic RAM memory chip was recently announced in the United States. Magnetic RAM is nonvolatile—in other words, it retains its information even when the power is turned off. Magnetic RAM easily exceeds silicon FET-based nonvolatile memories—the kind used in

OCR for page 127
c o n d e n s e d - m at t e r m at e r i a l s P h ys i c s  and cameras and MP3 players—in terms of write speed and durability. It may eventu- ally replace many of the established memory devices, but that will require further invention and discovery. In particular, the recently demonstrated spin-transfer switching effects may eventually endow magnetic RAM with an unrivaled combi- nation of speed, low power consumption, and low cost. But, while magnetic RAM might seem like an obvious path to the future, technology and the marketplace can be fickle; many competing technologies are vying for future acceptance. For example, phase-change memory is another nonvolatile memory in the late stages of commercial development. Information is represented as a change in the resistance of a material that can be rapidly switched between amorphous (high- resistance) and crystalline (low-resistance) phases. The phase change is driven by appropriately timed electrical heating pulses. With good prospects for shrinking devices to very small dimensions, phase-change memory is seen as the successor to silicon nonvolatile memory in the rapidly growing consumer electronics market. The chalcogenide materials that enable it are thus under intense exploration and development. Still, the physics of phase changes in structures with dimensions on the order of 10 nm and smaller is just beginning to be explored. While the commercialization of magnetic RAM and phase-change memory are most advanced, many other new memory devices are being explored in universi- ties and industrial laboratories. A partial list of commercial development efforts includes memories based on changes in the conductivity of complex metal oxides, electrochemical reactions in small molecules, and electromechanical deformations of molecular-scale structures. All of these approaches to future memory devices are nonvolatile, and none stores its information in the form of a packet of elec- tronic charge. None is wed to a silicon substrate, although for now, the circuits that read and write the information to each memory device must be based on silicon FETs. NEW DEvICES FOR PROCESSINg INFORMATION While some new memory devices are beginning to enter the market, no new device for processing information is at a remotely similar state of development. The coming decade will be one of exploring many pathways. While some approaches outperform others in certain aspects, no one knows the best overall solution at this time. But just as the transistor, more than any other invention, enabled the explosive growth of IT in recent decades, so would the invention of something “smaller, faster, and cheaper” enable further explosive growth in the capabilities and applications of IT. The search for new logic device concepts is driving research into new materials and novel physical systems. The terminology to describe these areas stems from the word electronics—as in spintronics, orbitronics, plasmonics, and so forth—names

OCR for page 127
how will i n f o r m at i o n t e c h n o lo g y r evo lu t i o n B e e xt e n d e d ?  the that evoke a sense of imminent technological applications. In reality, the tech- nological applications of these scientific developments are still quite uncertain. However, the research is drawing the interest of top scientists in a way that reminds some observers of the excitement in the early days of solid-state electronics. The following examples should convey a sense of some of the innovative paths that lie ahead. Spintronics, also known as magnetoelectronics, represents a break with silicon technology because it is based on magnetic materials, possibly including magnetic semiconductors. Processing information that is represented by spin orientation offers intriguing possibilities for nearly dissipationless computing. For example, under the right conditions, spin precession can rapidly switch the spins from one orientation to another, while dissipating only a small fraction of the energy stored in the system. This possibly can reduce or avoid wasteful heating that occurs in electrical circuitry. Understanding spin dynamics in nanoscale structures provides an enormous basic challenge, because geometric confinement and physical proxim- ity of these materials can endow them with extraordinary properties. Many basic questions need to be addressed. Can the spin-transfer effects being developed for magnetic memory devices also be employed in logic devices (see Figure 7.5)? Will further advances in materials allow these effects to be harnessed in low- power devices? Also, the extension of spin-transfer effects within antiferromagnetic FIGURE 7.5 Illustration of spin-transfer effects and associated spin dynamics. The magnetic field H is aligned along the z-direction, collinear with the magnetic easy axis which has a uniaxial anisotropy, where M denotes the magnetic moment. The red trajectory at the right shows how a spin-polarized current of sufficient magnitude can switch the magnetization orientation via the spin-transfer effect. SOURCE: J.Z. Sun, “Spin Angular Momentum Transfer in Current-Perpendicular Nanomagnetic Junc- tions,” IBM J. Res. Dev., 50, 81 (2006).

OCR for page 127
c o n d e n s e d - m at t e r m at e r i a l s P h ys i c s  and structures remains relatively unexplored. Voltage-induced switching of magnetic domains remains another avenue to explore. Indeed, obtaining a truly compre- hensive understanding of magnetism in nanoscale structures would be a major breakthrough and is a challenge that might result in one or more Nobel Prizes. Multiferroics are yet more-exotic materials, including nanoscale composites, that merge magnetic (ferromagnetic) and ferroelectric effects. This enables the switching of a magnetic bit via an electric field rather than a magnetic field, or the switching of a ferroelectric polarization via a magnetic rather than an electric field pulse. Multiferroics may thus help to realize useful spintronic devices that can be made very small. Magnetic fields are inherently nonlocal in nature, and hence as device density increases, stray magnetic fields can adversely affect neighboring bits. Multiferroics may offer a way to surmount the hurdle. At the same time, basic physical properties, such as the coupling of the order parameters, present intrigue for the theorist and experimentalist alike. A still more speculative and less well traveled field is the nascent study of or- bitronics, in which silicon retains its central role. Theoretical studies predict that electric fields can be used to create orbital currents in the absence of conventional charge currents. The orbital angular momentum is transmitted perpendicular to the direction of electron movement, making an orbital Hall effect. The excitement is that the orbital current can be dissipationless and robust against disorder. Ex- perimental tests and realizations remain for the future. Photonics and plasmonics offer the promise of new logic devices based on light. At present, photonic devices such as the semiconductor laser are used for digital communication over relatively long distances—from a few meters to thousands of kilometers. The smallest photonic devices are generally limited by light diffrac- tion effects to dimensions somewhat greater than the wavelength of the light that is used. Plasmonics seeks to go beyond the diffraction limit of light and to explore new device concepts in subwavelength structures. Particles in electromagnetic fields provided by light can develop resonant charge oscillations (plasmons) that can be propagated and manipulated. It is hoped that this approach can take optical devices into the nanometer realm. A major challenge is to integrate photonic and electronic circuits, intimately linking communications and information processing on a single chip, as envisioned in Figure 7.6. The emergence of metamaterials, or negative index-of-refraction materials, also opens new areas of optics and optical materials for physical exploration, as discussed in Chapter 6. The implications for applications are in the area of wireless communication. The quest is not only to explore such exotic phenomena but to shrink the structures so that they operate at shorter wavelengths. It has been said that negative refraction is a subject with constant capacity for surprise, with inno- cent assumptions sometimes leading to unexpected and profound consequences. While the roots of modern geometric optics can be traced back to the work of

OCR for page 127
how will i n f o r m at i o n t e c h n o lo g y r evo lu t i o n B e e xt e n d e d ?  the FIGURE 7.6 Schematic of a futuristic silicon chip with monolithically integrated nanophotonic and electronic circuits. The convergence of electronics and photonics is on the cutting edge of integrated circuit research. The goal is to perfect the materials and processes in order to integrate on a chip ultracompact nanophotonic circuits for manipulating light signals similar to the way that electrical signals are manipulated in present-day computer chips. The example shown is an N-channel multiple of signals from a germanium photodetector, where the performance is monitored via complementary metal oxide semiconductor (CMOS) very large scale integrated (VLSI) logic circuitry integrated onto the optoelectronic chip. SOURCE: Image courtesy of IBM Corporation. Willebrod Snell in the 17th century, the 21st century is ripe for new insights and breakthroughs. Molecular electronics, sometimes dubbed moletronics, are devices that are constructed at the ultimate size limit of single molecules and in which single electrons are transported, rather than currents of ensembles of electrons flowing in circuitry. In the future, the microscopic theory of electronic transport should move well beyond the knowledge presented in today’s most advanced textbooks. At the present time, the effects of Coulomb blockade and percolation through nanoscale particle arrays are being pursued. For example, densely packed arrays of organic-ligand-coated gold nanocrystals have been reported to belong to a new class of artificial solids with tunable electronic transport properties. Such proper- ties stem from single-electron charging and quantum confinement energies at the level of individual particles, mediated by couplings to neighboring particles. The particles are so small that a single electron can block the tunneling of another electron onto the same particle or require the next electron to be of opposite spin as dictated by the Pauli principle. Thus, the physics of Coulomb blockades and the Pauli exclusion principle are experiencing a rebirth of interest. As in this example, such developments in molecular electronics are also motivating fabrication based

OCR for page 127
c o n d e n s e d - m at t e r m at e r i a l s P h ys i c s 0 and a b c FIGURE 7.7 Logic components on a molecular level: Logic AND gate based on carbon monoxide (CO) cascade on a copper surface. The blue, red, and green circles represent, respectively, CO molecules that are going to hop, that are not going to hop, or that have already hopped to their final location. (Left) Initial state of CO molecules. (Middle) The result of a cascade that propagates after triggering one input. (Right) The final con- figuration, in which all of the CO molecules have hopped to their final locations. SOURCE: Images courtesy of IBM Corporation. Figure 7-7 on self-assembly rather than lithography, and thus they encompass major inte- grated challenges that span the physical, chemical, biological, and manufacturing sciences. A second example at the ultimate limit of miniaturization involves the molecu- lar cascade of CO on a copper surface to form a logic AND gate. The system is la- boriously configured by means of scanning tunneling microscope manipulation of the CO molecules to form the initial state, as shown schematically in Figure 7.7. qUANTUM COMPUTINg No matter what device concepts are pursued, smaller devices must eventually embrace a profound change in the way we think about computing. Present devices average the behavior of a great many quantum particles. Make any of these devices small enough and there will be only a few quantum particles in the system. In this limit, the laws of quantum mechanics manifest themselves most vividly, as has also been discussed in the National Research Council’s atomic, molecular, and optical physics decadal survey.3 Thus, researchers need to traverse the threshold and look beyond the limits of classical physics. Quantum computing should have broad applications across scientific disciplines. An application of immediate interest, however, involves the factoring of large numbers into primes, a capability that is key to breaking cur- rent cryptographic protocols. In principle, a quantum computer should be expo- 3 National Research Council, Controlling the Quantum World: The Science of Atoms, Molecules, and Photons, Washington, D.C.: The National Academies Press, 2007.

OCR for page 127
how will i n f o r m at i o n t e c h n o lo g y r evo lu t i o n B e e xt e n d e d ?  the nentially faster at this task than a conventional computer. To illustrate the point, consider that, while a classical computer composed of a three-bit register can store only one of eight possible numbers at a given moment in time, a quantum com- puter with three qubits can store all eight simultaneously, owing to the coherent superposition of states possible in quantum systems. Once the register is prepared, the quantum computer can perform many different calculations in parallel. Now imagine going from three qubits to ten to a hundred. The basis for enabling such operations is to assemble arrays of quantum coherent nanoscale objects to serve as the qubits and logic gates of the quantum computer. Solid-state approaches to building qubits include, for example, the confined electron spins in semiconduc- tors and the superconducting currents in Josephson junctions. For example, an alternating-current-voltage pulse applied to one of two Josephson junction qubits connected by a capacitor can cause the two qubits to becoming entangled and oscillate between two combined states, if the energy differences between the “0” and “1” states are equal in both qubits. The characteristic oscillation frequency is analogous to that of a quantum state of an atom; hence the circuit is referred to as containing “artificial atoms.” See Figure 7.8 for a representation of a qubit that could be used in such a circuit. Unlike in classical computers, in quantum computers the bits are prepared in an initial state, interact with one another to entangle, and must then be read out while still in a coherent state. Many novel pathways are being pursued, including the use of biological templates such as DNA strands, to serve as a backplane to secure the qubits in place. The challenges of addressing readout and quantum error cor- rection have stimulated theorists and experimentalists alike and will continue to do so well into the future. The effort to build practical quantum devices is driving unprecedented progress in the ability to measure and control physical systems. The new capability will impact science and society well beyond the currently perceived boundaries of IT. Quantum computing might disrupt the status quo by making present-day encryption methods obsolete because of its potential to rapidly break conventional codes. But simultaneously, it holds promise of providing quantum encryption schemes that are fundamentally unbreakable. CONCLUSIONS In summary, the rise of IT has powered an ongoing transformation of the global economy and modern society. It has changed manufacturing, finance, communica- tions, commerce, entertainment, education, and science itself. Today’s world has thus been shaped tremendously by “old” inventions such as the transistor and the hard disk drive. But the continued growth of IT will depend on new inventions. Advances in IT are intimately connected to advances in the rest of the CMMP grand challenge areas. The new materials for future IT applications will likely

OCR for page 127
c o n d e n s e d - m at t e r m at e r i a l s P h ys i c s  and FIGURE 7.8 Josephson junction devices might be used in future quantum computers. The optical micrograph shows an “artificial atom” about the size of a human hair made with a superconducting circuit. The red arrow points to the heart of the qubit—the Josephson junction device that might be used in a future quantum computer to represent a “1,” a “0,” or both values at once. SOURCE: Cour- tesy of R. Bialczak and J.M. Martinis, University of California at Santa Barbara. embrace the emergent behavior discussed in Chapter 2. Also, the exploration of the nanoworld, described in Chapter 6, will be critical, as will the availability of advanced measurement techniques and major characterization and computational facilities (Chapter 11). Foremost will be the new understanding captured by the conceptual theories that are in the process of being formulated to develop under- standing of new states of condensed matter. What is expected of the industrial, academic, and national laboratory sectors of the research community in the next decade to extend the information technology

OCR for page 127
how will i n f o r m at i o n t e c h n o lo g y r evo lu t i o n B e e xt e n d e d ?  the revolution? Industrial laboratories must aggressively pursue new partnerships with the academic and national laboratory communities to promote critical, pre-com- petitive research that will lay the basis for future IT technologies (see Chapter 9 and the concluding recommendation). The Nanoelectronics Research Initiative, a consortium of microelectronics companies and state and federal agencies funding university research, is one possible example for the future. Universities and national laboratories can lower barriers to the creation of innovative start-up companies. Such companies can offer uniquely flexible environments for the creative devel- opment of new products aimed at new IT markets. The national laboratories can expand their tradition of stewardship of major characterization facilities for the research community at large. The scientific sophistication and technical complexity of new projects requires staffing, planning, and investment at a higher level than in prior eras. Such powerhouse teams of researchers will be capable of flexibly config- uring into large interdisciplinary working groups to pursue scientific challenges in information technology, helping maintain U.S. leadership in this key industry. To scientists, it is exhilarating to explore new materials and physical processes with the hope of extending for future generations the benefits of the IT revolu- tion that scientists and the public alike have had the good fortune to experience in their lifetimes. The case for strong investments in CMMP information technology research is compelling.