Page 93

2
New Materials and Structures

Our ability to make new materials and structures—both in bulk and in reduced dimensions or length scales—is inextricably linked to the advancement of our understanding of fundamental phenomena in condensed-matter and materials physics. This chapter describes some of the past decade' s advances in inorganic materials and structures. Some of the advances and promising new areas in organic materials are discussed in Chapter 5. As described in Box 2.1, an astonishing array of new materials with unexpected properties has come over the horizon. Improvements in synthesis and processing have led to dramatic improvements in the properties of established materials and our ability to exploit these properties. As a result, we can now fabricate new combinations of materials, features of reduced dimensions, and other characteristics that differ in significant ways from previous possibilities. Some of these developments have provided fertile ground for condensed-matter and materials physicists to explore novel fundamental phenomena; others show promise for finding applications quickly; some have the potential to change our lives.

New materials underlie the science and technology described throughout this report. Beyond condensed-matter and materials physics, they enable both science and future technologies. In some cases, entirely new and unexpected phenomena appear in a class of new materials. Layered cuprate high-temperature superconductors are a new class of materials that has kept experimentalists and theorists alike searching to understand the physical basis of high-temperature superconductivity. New materials sometimes allow entirely new device concepts to be realized or lead to a dramatic change in their scale, such as single-molecule wires made of carbon nanotubes; and new forms of already known materials can



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 93
Page 93 2 New Materials and Structures Our ability to make new materials and structures—both in bulk and in reduced dimensions or length scales—is inextricably linked to the advancement of our understanding of fundamental phenomena in condensed-matter and materials physics. This chapter describes some of the past decade' s advances in inorganic materials and structures. Some of the advances and promising new areas in organic materials are discussed in Chapter 5. As described in Box 2.1, an astonishing array of new materials with unexpected properties has come over the horizon. Improvements in synthesis and processing have led to dramatic improvements in the properties of established materials and our ability to exploit these properties. As a result, we can now fabricate new combinations of materials, features of reduced dimensions, and other characteristics that differ in significant ways from previous possibilities. Some of these developments have provided fertile ground for condensed-matter and materials physicists to explore novel fundamental phenomena; others show promise for finding applications quickly; some have the potential to change our lives. New materials underlie the science and technology described throughout this report. Beyond condensed-matter and materials physics, they enable both science and future technologies. In some cases, entirely new and unexpected phenomena appear in a class of new materials. Layered cuprate high-temperature superconductors are a new class of materials that has kept experimentalists and theorists alike searching to understand the physical basis of high-temperature superconductivity. New materials sometimes allow entirely new device concepts to be realized or lead to a dramatic change in their scale, such as single-molecule wires made of carbon nanotubes; and new forms of already known materials can

OCR for page 93
Page 94 possess different properties. Semiconductor nanoclusters, which emit light whose wavelength depends on cluster size, offer the possibility of tailoring material properties to suit a particular need. Even mature techniques, such as those for bulk crystal growth, demand continuous improvements in process control to produce the size or quality of material required for either technological applications or fundamental studies. Better understanding of the mechanisms at play in materials that have been known for decades can lead to new approaches that alleviate detrimental properties. An excellent example is the introduction of metallic oxide electrodes in ferroelectric devices, which reduces aging effects dramatically. Better understanding of the details of materials preparation can give rise to improvements in processing. Improved insight into the kinetics of epitaxial growth can dramati- BOX 2.1 Additions to the Zoo: New Materials and Structures of the Past Fifteen Years There have been far too many new developments in the past 15 years or so to document them all in detail, but all these developments have been made possible by advances in two intertwined areas: complexity and processing. Many of the new materials and structures are dramatically more complex, compositionally or structurally, than have been studied previously. In general, this trend has required advances in processing to allow control of the increased complexity. In other cases, the final product may not be much more complex than other well-known materials or structures, but the processing itself may need to be altered to achieve more control over the growth process in order to obtain the new material. Advances giving rise to new materials and structures fall into three categories. Some involve the synthesis of an entirely new compound or material. The advance may have been revolutionary, meaning that the properties of the new material (or in some cases its existence) could not have been predicted. In other cases, advances in processing have allowed fabrication of new or modified materials or structures whose properties were suspected before the material was actually made. This may allow a well-known compound to be remade in a new form with different properties. Third, well-known materials are sometimes found to exhibit new (in some cases unexpected) properties that appear when the ability to process them is improved. The new property may be found in a known material simply by looking at it in a new light, which shines on it as a result of insight gained from another materials system. The materials advances listed in Table 2.1.1 were driven by different motivations. Many addressed a technological need, such as the need to transfer or store information. Others were driven by scientific curiosity. Although the driver can be clearly identified in each case, the two sets are not mutually exclusive. Many discoveries that result from pure scientific curiosity ultimately find their way into products. For example, low-temperature superconductors are now used in magnets for magnetic resonance imaging. Other discoveries, though originally motivated by a technological need, give rise to very beautiful and fundamental insights. (Text box continued on next page)

OCR for page 93
Page 95 (Text box continued from previous page) For example, the fractional quantum Hall effect was first observed in high-mobility semiconductor structures now used in high-frequency applications. TABLE 2.1.1 Some New Inorganic Materials of the Past Fifteen Years Advance Driver Nature of Advance New compounds/materials High-temperature superconductors Science Revolutionary Organic superconductors Science Revolutionary Rare-earth optical amplifier Technology Evolutionary Intermetallic materials Technology Evolutionary High-field magnets Technology Evolutionary Organic electronic materials Technology Evolutionary Magnetooptical recording materials Technology Evolutionary Bulk amorphous metals Technology Evolutionary New structures of known materials Quasicrystals Science Revolutionary Buckyballs and related structures Science Revolutionary Nanoclusters Science Evolutionary Metallic hydrogen Science Evolutionary Bose-Einstein condensates Science Evolutionary Giant magnetoresistance materials Technology Revolutionary Porous silicon Technology Evolutionary Diamond films Technology Evolutionary Quantum dots Technology Evolutionary Foams/gels Technology Evolutionary New properties of known materials Gallium nitride Technology Revolutionary Silicon-germanium Technology Evolutionary   cally lower the growth temperature in semiconductor processing. Understanding and exploiting fundamental growth mechanisms can lead to previously unattainable structures as in the use of strain to induce the self-assembly of quantum dots. Many advances in condensed-matter and materials physics are the direct result of the availability of materials and structures of a quality not previously attainable. These materials and structures in turn exist because of improvements in the technology used to make, study, measure, and see them. The impetus for improvements in the materials is often a technological need, not the search for new knowledge. The new knowledge generated, however, in some cases itself becomes the enabler of revolutionary technology. This interplay is explored in Box 2.2.

OCR for page 93
Page 96 BOX 2.2 The Science—Technology Circle Tremendous advances in compound semiconductor devices were enabled by dramatic improvements in the growth of thin films that began in about 1970 with the invention of molecular beam epitaxy (MBE, see Figure 2.2.1). The direct antecedents of MBE were developments in vacuum technology beginning in the 1960s and continuing into the 1970s, driven by accelerator development and space physics. As the attainable vacuum improved, it became possible to keep a surface atomically clean for long enough to study it. Surface probes such as Auger spectroscopy and electron diffraction techniques allowed the clean surfaces to be studied. MBE enabled the controlled, layer-by-layer growth of compound semiconductors. The composition of the film could be changed abruptly. Extremely high mobility was achieved in GaAs-GaAlAs heterostructures through ''modulation doping.'' Research into these structures was pursued because of their utility in high electron-mobility transistors (HEMTs, see Figure 2.2.2) which are used today in high-speed electronics. Study of these layers at low temperatures in extremely high magnetic fields led to the discovery of the quantum Hall effect, which takes place in a two-dimensional electron "gas" produced in a transistor-like device. Under these conditions, electron correlations dominate, leading to precise quantization of the Hall conductance. As the quality of the layers was improved further, the mobility also improved, and the fractionally quantized Hall effect (FQHE) was discovered, in which the quantum number describing the system is a fraction rather than an integer (see Figure 2.2.3). The FQHE has subsequently been used for unprecedentedly accurate measurements of the fundamental quantity h/e2 (Planck's constant divided by the square of the charge of the electron).   Figure 2.2.1 Molecular beam epitaxy was invented at Bell Laboratories in about 1970 as an outgrowth of advances in vacuum technology and surface science techniques. (Courtesy of Bell Laboratories, Lucent Technologies.) (Box continued on next page)

OCR for page 93
Page 97 (Box continued from previous page)   Figure 2.2.2 A high electron-mobility transistor (HEMT) such as those used in cellular telephones. The round bonding pads are 100 µm in diameter, roughly the size of a human hair. The gate of the transistor, just 0.05 µm across, appears as the two narrow lines in the center of the scanning electron micrograph. (Courtesy of Sandia National Laboratories.)   Figure 2.2.3 A pictorial representation of the many-particle state that underlies the fractional quantum Hall effect. The height of the landscape represents the amplitude of the quantum wave of one electron as it travels among its companions (shown as balls). The arrows indicate the vortices induced by the magnetic field, which attach themselves to the electrons to form composite particles. (Courtesy of Bell Laboratories, Lucent Technologies.)

OCR for page 93
Page 98 The remainder of this chapter examines a few of the past decade's most impressive advances in materials and structures. The selections emphasize a number of themes that have emerged in materials research. Some of the discoveries have been completely unexpected. Others were predicted, although the experimentalists did not always know of these predictions when they did their work. Our thinking about new materials has changed fundamentally; we now consider dramatically more complex possibilities in our search for new materials than we did a decade ago. In some classes of materials that have been studied for many decades, we have achieved a much deeper understanding of physical and chemical mechanisms that govern their properties. This understanding in turn has led to improvements in the properties of the materials, either through elimination of problems inherent in existing materials by improved processing or by the introduction of new materials. Even in a material as thoroughly studied as carbon, a myriad of new forms has been discovered, exhibiting a wide range of properties. Shrinking the dimensions of well-known materials such as semiconductors has led to properties dramatically different from those of the bulk. New concepts in thin-film growth have led to improved film properties by changing the growth and processing windows. Finally, there has been a change in the attitude toward strain in heteroepitaxial systems that allows strain to be used to tailor the morphology as well as the electrical properties of the layers. The culmination of this effort is in the use of strain to induce self-assembly of quantum dots. Complex Oxides Surely one of the most surprising developments since the publication of the Brinkman report1 has been the discovery of high-temperature superconductivity in complex oxide materials, beginning in 1986 with the observation by Bednorz and Müller of superconductivity near 30 K in La2-xBaxCuO4. This discovery was rewarded with the 1987 Nobel Prize in Physics (see Table O.1). The field exploded with the discovery of superconductivity at temperatures in excess of the boiling point of liquid nitrogen (77 K). The family of known high-temperature superconducting materials now numbers near 100, with the highest superconducting transition temperature (Tc) above 130 K. High-temperature superconductivity has significantly altered the direction of condensed-matter and materials physics in several ways. The excitement generated by this totally unexpected discovery attracted researchers from throughout the field of condensed-matter and materials physics and beyond to the study of these fascinating materials. More recently, the principles that have been successful in the study of these materials have proven valuable in the study of other areas of condensed-matter and materials physics, most notably other sorts of oxides. 1 National Research Council [W F. Brinkman, study chair], Physics Through the 1990s, National Academy Press, Washington, D.C. (1986).

OCR for page 93
Page 99 High-temperature superconductors are much more complex than many of the materials that have occupied the attention of condensed-matter and materials physicists for many decades (see Figure 2.1). This complexity, however, is a two-edged sword, giving rise to a richness in the possible structures and properties of the materials, but also making the materials extremely challenging to produce, control, and understand. The crystal structures of these materials are dramatically more complicated and have lower symmetry than those of low-temperature superconductors or semiconductors. The physical properties are similarly anisotropic. This makes the control of crystallographic orientation extremely critical. The unit cell is large. A large unit cell and low symmetry offer many opportunities for the formation of defects during materials preparation, either in individual atomic sites or in long-range crystallographic perfection. The superconducting coherence length is of the order of the interatomic spacing in some crystallographic directions. This makes the materials exquisitely sensitive to defects—from atomic-scale defects, such as vacancies, interstitials, and substitutional atoms, to grain boundaries and other larger-scale imperfections. Separating the intrinsic properties of such materials from artifacts caused by defects is critical to gaining full understanding of the high-Tc phenomenon. It places extreme emphasis on materials preparation and serves as an example of the true collaboration that must exist between those who seek to understand and control the growth of the materials and those who probe their underlying physics, as illustrated in Box 2.3. Conversely, the carefully controlled introduction of Figure 2.1 Historical development of inorganic superconductors,  with Nobel prizes indicated by stars. Increasing  superconducting transition temperature correlates with  increased chemical complexity and more constituent  elements, as shown by the numbers 1 to 5. (BCS stands  for the Bardeen-Cooper-Schrieffer theory of classical  superconductors.) [MRS Bulletin 19, 26 (1994).]

OCR for page 93
Page 100 BOX 2.3 Vortex Matter: A Novel Window on Materials Physics In the past 5 years, a new area of condensed-matter investigation has emerged, based on the remarkable behavior of vortices in superconductors. For decades, type-II superconductors in a magnetic field have been understood as arrays of quantized tubes of magnetic flux, each surrounded by a circulating "vortex" of supercurrent that defines its interaction with its neighbors and the outside world. In traditional superconductors, thermal energy is limited to about 20 K by the superconducting transition temperature, and the vortex tubes form an elastic solid. High-temperature superconductors offer a new possibility: thermal energies up to about 100 K may melt the vortex solid, creating a novel liquid state with dramatically different properties that arise from the relative motion of vortices. As early as 1988, motion of vortices below Tc was found to create undesired dissipation in high-Tc cuprates. The thermodynamic nature of the melting phase transition and of the resulting vortex liquid has been vigorously debated. Theorists soon realized that vortex phases and phase transitions embody many fundamental features of condensed-matter physics, including reduced dimensionality, entanglement of flexible line objects, and the role of disorder on elastic media. Studies of vortex matter provide new insight into these basic materials physics issues in other condensed-matter environments. The diversity of equilibrium vortex phases is illustrated in the phase diagram of Figure 2.3.1. Although theoretical analysis of vortex liquids and solids abounded, experimentalists were frustrated by the quality of the available high-temperature superconducting materials. Real materials contain defects like impurity clusters, dislocations, twin boundaries, and rough surfaces. In superconductors, these defects generate pinning sites that immobilize vortices and remove them from participation in equilibrium behavior. The experimental observation of vortex phase transitions had to await more perfect materials with dramatically reduced defects and vortex pinning. In 1992, the first indications of vortex lattice melting were observed in electrical transport experiments. These measurements accurately located the melting line in the H-T plane and gave tantalizing but indirect evidence that the transition was first-order in clean materials. Further transport experiments suggested that first-order melting was destroyed by controlled pinning disorder and suggested the existence of a critical point in the melting line. These and other experimental observations created new interest and activity in the field. Experimentalists then sought the next level of fundamental information—thermodynamic characterization of the order and entropy changes on vortex melting—with magnetization and specific heat experiments. Such experiments require an even higher level of sample perfection to ensure thermodynamic equilibrium in the solid phase, where pinning effectiveness is significantly enhanced by shear elasticity. Sample size was a second serious problem: the most perfect crystals are also the smallest, making it extremely difficult to resolve the tiny magnetic and thermal signatures of melting from the much larger background. Nevertheless, sample preparation techniques continued to improve with better understanding of the roles of composition, growth rates, and annealing procedures. Improved materials enabled several landmark thermodynamic experiments, which have now settled the question of the thermodynamic order of the transition and raised new questions about critical points, vortex entanglement, and the dimensionality of the liquid. (Box continued on next page)

OCR for page 93
Page 101 (Box continued from previous page)   Figure 2.3.1 A suggested phase diagram of vortex matter in the magnetic field-temperature plane. Several vortex liquid and solid phases are illustrated, including a liquid of entangled vortex lines, a perfect hexagonal lattice, a polymer glass of entangled lines, and solid phases disordered by point pinning defects (vortex glass) or by line pinning defects (Bose glass). The melting transition is first-order from a lattice and proposed to be second-order or continuous from a glass. A critical point may occur on the melting line, where the first-order character disappears. The normal and vortex liquid states are separated by a fluctuation dominated by crossover rather than by a true phase transition. (Courtesy of Argonne National Laboratory.) Vortex matter has emerged as a vital field, with its own developing issues and international community of researchers. It extends traditional studies of atomic matter in several ways. For example, vortex density is linear in the applied magnetic field, so it can easily be changed by an order of magnitude with the twist of a dial. Experimental access to such a large density range is unheard of in atomic matter. The interactions among vortices are well-known Lorentz forces, which can be treated analytically or in simulation with no uncontrolled approximations. Advanced materials development has produced clean crystals with few pinning defects, revealing intrinsic thermodynamic behavior and its evolution under controlled disorder induced by electron or heavy ion irradiation. Finally, vortices can be set in motion by the Lorentz force from an externally applied transport current, enabling studies of driven phases, steady-state motion, and the new area of dynamic phase transitions. This remarkably rich microcosm of condensed-matter physics owes its existence to two materials developments: the landmark discovery of high-temperature superconductors, which introduced large thermal energies into the vortex phase diagram, and dramatic improvements in materials perfection, which enabled experimental studies of the delicate thermodynamics of collective vortex behavior.

OCR for page 93
Page 102 defects of a particular type into the material by, for example, ion irradiation or judicious atomic substitution allows the properties to be adjusted. The superconducting oxides with Tc above 77 K all contain at least four elements, two of which are copper and oxygen. Oxygen moves readily in these materials, during both sample preparation and subsequent processing. Changing the oxygen content by just a few percent can determine whether a material is a superconductor or an insulator. It can also govern the symmetry and crystal structure of the material, resulting in phase transformations during specimen preparation that, to date, have been unavoidable. Precise control of the stoichiometry of the metal constituents is also required to optimize the superconducting properties, although the consequences of deviations from ideal stoichiometry are not nearly as critical for the metals as for oxygen. Current interest in the high-temperature superconducting materials centers around two general areas: superconducting electronics and the carrying of large currents. The electronics applications can be further subdivided into logic and high-frequency applications. Electronics applications require thin films, generally in combination with films of other materials. The fabrication of reproducible tunnel junctions with useful properties for logic applications has been very challenging because of the incompatibility of high-temperature superconducting materials with most nonoxide barrier materials and the extremely short coherence length of the superconductor. Quite a few metallic oxides with compatible crystal structures have been identified and studied as a result of considerable research into suitable barrier materials. A promising area of application is in components for communications, particularly in the gigahertz frequency domain. The major issues are the surface resistance of the material and electrical nonlinearities at high frequencies. Though there has been considerable progress in improving surface resistance in the past few years, detailed understanding of the relationships between this and other relevant properties and the structure of the materials is still emerging. Technological applications demand large-area films that can be deposited fast enough to be economically viable. There has been dramatic progress, with high-quality films of YBa2Cu3O77-x (see Figure 2.2) now available on substrates several hundred square centimeters in area. Current-carrying applications require bulk material or thick films. Grain boundaries, especially those with significant misorientation between grains, are extremely detrimental to high critical currents because of both the extreme anisotropy of the materials properties and the properties of the grain boundaries themselves. The most successful approach for bulk materials with properties of potential technological interest has been the use of drawn, multifilament wires, especially in the bismuth system. The drawing induces alignment of the grains in the filaments and increases the critical-current density. More recently, biaxial orientation has been achieved in thick YBa2Cu3O77-x films deposited on metal substrates, either coated with an aligned buffer layer fabricated by ion beam-assisted deposition or with strong crystallographic alignment induced in the substrate by rolling.

OCR for page 93
Page 103 Figure 2.2 The orthorhombic crystal structure of  superconducting YBa2Cu3O7-x. The superconducting  transition temperature of this material is above 90 K.  Note the presence of four elements in the compound  and the low symmetry of the structure. These  characteristics make materials synthesis challenging  and give rise to dramatic anisotropy in the physical  properties of the material. (Courtesy of Princeton  University.) It has proven very fruitful to apply the principles discovered and techniques developed for high-temperature superconductivity to other classes of complex oxides. In some cases, this research has been driven by the need for materials with specific electronic or magnetic properties that are chemically and structurally compatible with high-temperature superconductors. These materials are typically needed as buffer or barrier layers. Compatible materials with other properties could be needed in the future if high-temperature superconducting devices are to be successfully integrated with devices having other functionality, such as memory and optical devices. Perhaps the most impressive demonstration of the application of lessons from high-temperature superconductivity has been the recent interest in colossal magnetoresistance in LaMnO3-derived materials (see Box 2.4).

OCR for page 93
Page 126 Artificially Structured Materials Artificially structured materials—materials whose structure or composition differs in some intentional way from materials available in nature—have enabled many of the advances in condensed-matter and materials physics over the past decade that are described in this report. Such materials are frequently dominated by interfaces, a feature that often leads to properties very different from those of bulk materials. Because artificially structured materials are frequently prepared far from thermodynamic equilibrium, they can exhibit phases or properties that are not otherwise achievable. Their multilayer structural length may be on the order of the length-scale characteristic of nonlocal physical phenomena in solids, making such materials ripe for fundamental investigations. The physical limits to their fabrication have been pushed to the greatest extent for semiconducting materials. The coming decade will undoubtedly see the same limits pushed for other classes of materials as well: complex oxides, polymers, biological materials, and composites are a few of the most exciting. Many if not most artificially structured materials involve heteroepitaxy, the crystallographically oriented growth of one material on a dissimilar one. In nearly all cases, heteroepitaxy involves a lattice mismatch between the different materials, which produces strain in the initial epitaxial layer. The strain is relieved as the epitaxial film thickness increases, by a roughening of the surface of the epitaxial layer or by the introduction of defects such as dislocations into the epitaxial layer or both. Controlling when and how strain relief occurs is a key issue in heteroepitaxy. The strained films discussed here are grown near the thermodynamic limit. Sputter-deposited films, such as those discussed in Box 2.8 and in the section on magnetic multilayers in Chapter 1, are deposited in the kinetic limit, which is required for alternating layers of extremely disparate materials. One distinct trend has been toward the use of more highly strained heteroepitaxial combinations, such as InGaAs/GaAs (see Figure 2.12) and SiGe/Si. Such systems must be approached with great care in order to achieve the optimum structural (and consequently electrical or optical) quality. Morphology-related strain relief is not a new phenomenon. Mounding in a film may partially relieve elastic stress of the epitaxial material within each mound. Even though there is additional compression of lattice planes at the grooves between the mounds, the roughened morphology is energetically favorable because the volume of material subjected to additional stress is much less than the volume experiencing partial stress relief. Another very important factor is the surface free energy. Roughening generally increases this energy, so that roughening is suppressed until the free-energy reduction in the system by stress relief is greater than the free-energy increase caused by surface area increase and step formation. Strain-induced roughening can be problematic in the fabrication of coherently strained device structures, for which it is important to understand the early

OCR for page 93
Page 127 BOX 2.8 Multilayers for X-Ray and Extreme Ultraviolet Optics Multilayers are artificially structured materials that are periodic in one dimension in composition or both composition and structure. These layered materials are, if perfect, equivalent to single crystals in one dimension. Thus the multilayer acts as a superlattice, diffracting longer-wavelength radiation in a manner directly analogous to the diffraction of x-rays by crystals. This application of multilayer structures as dispersion elements for soft x-rays and extreme ultraviolet radiation was the impetus for the first attempts to synthesize multilayer materials. Many factors determine the character of the multilayer response to an incident spectrum. The important parameters are the substrate quality (roughness and figure), the uniformity and thicknesses of the component layers, the x-ray optical constants of the component elements, the number of layers in the structure, the interfacial width between layers (i.e., interfacial abruptness in atomic position and composition), and roughness at layer interfaces. Many of these factors depend in turn on the synthesis process and the materials. Therefore, understanding of multilayer performance depends on a knowledge of the relationships among synthesis process, resultant microstructure, and properties for these engineered microstructure materials. The individual layers of the optics have a specific set of properties related to bulk forms of the materials. Primary issues include the compositions and structures of the layers, the x-ray optical properties of the layers, and uniformity of the areal density of atoms in the layers (see Figure 2.8.1). Specific synthesis questions relate to the film nucleation and growth behavior because deposition of material A onto a substrate or layer B may differ substantially from deposition of material B onto a substrate or layer A. Interfaces within the multilayer must also be controlled to an excruciating degree. They must be compositionally abrupt, smooth, clean, and flat. Recent work has shown that precise control of sputtering parameters during multilayer deposition allows control of individual layer thicknesses to an accuracy of better than ~0.01 nm, which greatly enhances reflectivity for both nickel-carbon and tungsten-carbon multilayers. Sputter deposition of multilayers typically produces higher quality structures than thermal source techniques. This has been attributed to ion bombardment by the sputter plasma resulting in smoother interfaces and higher reflectivities. Results of ion beam-assisted deposition support this proposal. Thermal-evaporation-source synthesized rhodium-carbon multilayers with and without argon ion bombardment (300 eV) at an incidence angle of 10 degrees show the effect. A gain of more than a factor of two in reflectivity was found for the samples "polished" by the incident ion beam. This increased reflectivity is attributed to smoothing of the interfaces between the carbon and rhodium by a factor of 30 percent by the ion bombardment. Combining these two improvements in control are likely to facilitate fabrication of higher quality multilayer structures, particularly of smaller periods. Multilayer structures may be optimized and engineered for specific spectral ranges by an analysis for optimum materials on the basis of their x-ray constants and an assessment of their suitability for multilayer microstructure synthesis. As an example, there are difficult spectral regions in which the lowest absorption materials useful as spacer layers are either toxic, such as beryllium, or unstable, such as lithium. Candidate materials such as magnesium are difficult to deposit as (Box continued on next page)

OCR for page 93
Page 128 BOX 2.8 Continued a uniform thin film because of its low melting point and high vapor pressure. Mg2Si has been identified as a possible new material for this application, and W/Mg2Si multilayers have reflectivities that are the highest reported in the 800 to 1300 eV range. Multilayer x-ray optics and instrumentation are now mature enough to be both an enabling technology and an area of scientific investigation in their own right. The promise that was held for soft x-ray and extreme ultraviolet multilayer optics is now coming to fruition, and many of the advanced optical systems envisioned in the late 1970s are becoming reality. Such x-ray optics will likely form the critical element of vacuum ultraviolet optics for the next generation of lithography in the semiconductor industry.   Figure 2.8.1 Transmission electron micrograph of a 6.9 nm period Mo2C/Si multilayer x-ray mirror (top) and the experimental and calculated reflectivity as a function of x-ray wavelength (bottom). The experimental reflectivity is 93.5 percent of the calculated values. [Reprinted with permission from T.W. Barbee, Jr., and M.A. Wall, ''Interface reaction characterization and interfacial effects in multi-layers,'' Proceedings of the SPIE 3113-20, 204 (1997). Copyright © 1997 SPIE.]

OCR for page 93
Page 129 Figure 2.12 Atomic force micrograph of self-assembled InAs islands deposited by  molecular beam epitaxy on a patterned GaAs(001) surface. The valleys  and hills, which have been defined by optical lithography and etching,  have a period of ~240 nm. The InAs islands, which are formed by  deposition of about 1.5 monolayers of InAs on the corrugated GaAs  surface, are preferentially located in the valleys of the surface. The  height and diameter of the InAs islands are 10 nm and ~20 nm,  respectively. These islands will be transformed into quantum dots  by in situ overgrowth of a GaAs cladding layer.  (Courtesy of the University of California at Santa Barbara.) stages of the transition in order to avoid or suppress three-dimensional growth. On the other hand, the strain-driven transition is beneficial for the self-assembly of quantum dots, in which it is necessary to control the size distribution and self-organizing behavior of the islands. It is critical to understand the kinetic pathways to island formation. Recent theoretical investigation suggests that systems with tensile stress could be more resistant to roughening than those with compressive stress. Recent work has supported this prediction in, for example, the Si-Ge system. Molecular dynamics modeling has attributed this observation to an increase in the energy of certain types of surface steps under tensile strain, which makes it energetically favorable for the surface to remain planar. One of the major contributions of scanning-probe microscopies to heteroepitaxy has been improved understanding of morphological evolution in heteroepitaxy. A general trend in all experiments is the decreasing size of typical morphological features with increasing misfit stress. In many highly mismatched systems, surface ripples exhibit a strong tendency to facet along inclined planes,

OCR for page 93
Page 130 Figure 2.13 Atomic force microscope image of an initially planar 2-nm thick  Si0.5Ge0.5 alloy layer on Si(001) after annealing to produce hut- shaped islands caused by strain in the layer. [MRS Bulletin 21, 31 (1996).] as in the epitaxial germanium "hut" clusters in Figure 2.13. The presence of {501} facets appears to be a general feature of strained-layer growth in the Si-Ge system. It is not understood, however, why such facets are stable and what role they play in the growth of coherently strained islands. The picture is emerging that {501} facets are the natural result of the desire to release as much elastic energy as possible without unduly creating energetically costly surface-step configurations. During growth, strain relaxation in heteroepitaxial systems can lead to changes in island shapes. In semiconductor systems, coherent islands are often faceted and characterized by large aspect ratios [up to a 10:1 base-to-height ratio in the case of the germanium "huts" and up to 50:1 in silver on Si(100)]. Recent calculations have shown that strained islands are likely to undergo a shape transformation during growth. Below a critical island size, the energy balance favors compact, symmetric islands; for large islands, elongated shapes with high aspect ratios are preferred. This suggests one approach to the challenge of producing quasi-one-dimensional quantum wire structures. Turning now to the other major strain-relief mechanism—defects such as

OCR for page 93
Page 131 dislocations—we enter an area of research with much longer antecedents. Traditionally, misfit dislocations have been avoided simply by keeping the film thickness at less than the "equilibrium critical thickness." More recently, other strategies have been introduced, such as reduced temperature growth or substrate patterning. When substrate patterning is used to isolate regions of the sample, relaxation is greatly retarded because most isolated regions contain no heterogeneous nucleation sites. These regions can be remarkably stable during thermal anneal, which shows that controlling nucleation may be the key to controlling, or even suppressing, relaxation. One of the most successful methods to date of fabricating quantum dots uses self-assembly that results from growth kinetics controlled by strained-layer epitaxy. A strain-induced transition from two- to three-dimensional growth results in the formation of coherently strained islands on the surface of the semiconductor. Using these principles, islands <<20 nm in diameter can be fabricated with size distributions within ± 10 percent. The resulting islands are pseudomorphically strained and dislocation free. The random distribution of islands can be modified by appropriate control of their nucleation and growth, kinetics. It was readily apparent early on that preferential nucleation of islands takes place at surface steps. This effect could be used to order islands. Recent progress in making extremely perfect kinkless steps over micron distances on Si(111) offers hope of ultimately achieving ordering of island assemblies by this approach. Another strategy for ordering the islands is based on the very sharp transition from two- to three-dimensional growth. A corrugated substrate with concave and convex areas will tend to flatten and minimize its surface energy during epitaxy through faster growth of the convex areas. Thus, when depositing a strained layer over such a surface, the critical thickness will be reached sooner in these areas, and quantum dots will nucleate preferentially in these regions. Another promising approach based on strain-induced nucleation allows for regulating the size and ordering of the islands in the growth direction. If two or more layers of quantum dots are grown sufficiently close to each other (closer than 10 nm for InAs/GaAs), it is possible to obtain self-alignment of the islands in the growth direction. Future Directions And Research Priorities The examples of new materials and structures presented in this chapter point out the major themes in the search for new and improved materials and properties that have characterized the past decade. The themes include the discovery of new and unexpected materials with novel properties and the use of new tools to provide improved understanding and control in well-known materials. These developments foreshadow many of the advances that are likely in the coming years. It is critical to emphasize, however, that many of the most exciting devel-

OCR for page 93
Page 132 opments have been complete surprises; this will almost certainly be true for the foreseeable future as well. Materials Properties by Design: Complexity The ability to tailor materials and structures to obtain a desired set of properties is in its infancy (see Box 2.9). Band-gap engineering, which has been achieved by judicious use of strain, alloying, and quantum-size effects, is one area that has had considerable impact. Control of the microstructure of a material, through processing and judicious choice of geometry or neighboring materials, has been used to control the physical properties of some materials. Artificially layered materials have been given unusual dielectric properties and have contributed to the search for new high-temperature superconductors. Future progress will build on these recent accomplishments. Eventually, we can expect to be able to tailor materials at the molecular level, building up materials molecule by molecule in three dimensions. This capability will enable truly three-dimensional designs with as much or as little symmetry as needed. In the future, a design will be able to incorporate structure at multiple length-scales, enabling the optimization of multiple properties that involve phenomena operating at very different dimensions. Today, the most advanced artificially structured materials utilize individual material constituents of the same or very similar classes: III-V semiconductors, for example. We can look forward to being able to use a much more colorful palette, not limited to a single class of materials or even just to inorganic materials. Polymers, organic molecules, and even biological molecules are likely to become integral parts of increasingly complex structures as we learn more about how to manipulate molecules individually. A glimpse into the possibilities is described in Box 2.10. As the structures that we are able to build become more complex, we will need the tools to be able to see, characterize, and manipulate them. We will need to be able to work with these structures on all of the length scales that are relevant for the properties we desire. The scanning-probe microscopies are an important step in this direction, but more, equally revolutionary advances will be required to truly take advantage of this new regime of materials design. Finally, we will need to be able to make predictions about these new materials. Our ability to predict the existence of new materials with interesting properties is extremely limited. Witness the complete surprise presented by the discovery of high-temperature superconductors or the lack of theoretical guidance concerning other interesting avenues of inquiry in the search for other high-temperature superconducting families. Theoretical guidance regarding promising synthetic routes for fabricating new materials would be most useful to experimentalists trying to prepare them. Finally, improved communication between theorists and experimentalists might help to shorten the gap between prediction and fabrication—several decades in the case of C60, for example.

OCR for page 93
Page 133 BOX 2.9 Combinatorial Chemistry and the Search for New Materials Systematically varying the composition of a multicomponent system to optimize its properties is a time-honored empirical method in materials synthesis. For example, "phase spreads" of thin films proved powerful in the study of metalinsulator transitions almost 2 decades ago. Scaling up the approach to allow the fabrication and testing of tens or hundreds of compositional variants requires the ability to prepare small volumes of precisely controlled composition under known preparation conditions as well as the capability to test these miniature samples for desired properties. Known more recently as combinatorial chemistry, this approach is used to make a large number of chemical variants in parallel, to screen them quickly and reliably for chemical activity, and to build a library of information about the resultant chemical diversity. Until recently, combinatorial chemistry has been used primarily to transform the way new drugs are discovered, but in the coming decade, it may have an impact on the search for other classes of new material as well. The most notable forays of combinatorial chemistry into non-medical arenas are in the areas of superconducting compounds and phosphors (see Figure 2.9.1). The applicability of the technique to the search for catalysts is also being investigated. All these materials share the property of being very complex, containing many elements and eluding prediction of their properties or even existence using any currently available theories or models. As the entire field of condensed-matter and materials physics moves toward increasingly complex systems, techniques such as combinatorial chemistry are likely to make a home for themselves alongside more traditional techniques such as bulk crystal growth or physical and chemical vapor deposition. For this promise to be realized, however, new tools need to be developed that can analyze and sort the large number of samples that are produced by this powerful technique. Figure 2.9.1 An array of different combinations of phosphors being screened for brightness in ultraviolet light. [Reprinted with permission from E.D. Isaacs, M.A. Marcus, G. Aeppli, X.-D. Xiang, X.-D. Sun, P. Schultz, H.-K. Kao, G.S. Cargill Ill, and R. Haushalter, "Syncrotron x-ray microbeam diagnostics of combinatorial synthesis," Applied Physics Letters 73, 1820 (1998). Copyright © 1998 American Institute of Physics.]

OCR for page 93
Page 134 BOX 2.10 Polymers Enable Porous Inorganic Materials Synthesis to Order Block copolymers self-assemble into ordered nanostructures consisting of three-dimensional arrays of spheres, cylinders, lamellae, and even bicontinuous domains depending on the lengths of the blocks. Polymer physicists have achieved a remarkably detailed understanding of the interplay between chain architecture and thermodynamics that leads to these nanostructures. As useful as this insight has been for all-polymeric materials, it now promises to be equally important for the synthesis of inorganics. A long-standing need has been for porous ceramics that have well-defined but large pore sizes (>5 nm). These can now be achieved using block copolymers as a template. A copolymer with hydrophilic and hydrophobic blocks is ordered into a nanostructure in which the hydrophobic block forms aligned cylinders. A ceramic precursor is absorbed preferentially into the hydrophilic matrix surrounding the cylinders and allowed to condense into a robust inorganic oxide network. The block copolymer can be extracted, leaving a ceramic matrix surrounding ordered uniform cylinderical pores, as shown in Figure 2.10.1. Pore size can be controlled between 5 and 30 nm simply by changing the length of the copolymer, leaving the ratio of blocks the same.   Figure 2.10.1 Transmission-electron microscope micrographs of mesoporous silica with pores sizes 6 and 9 nm. [Reprinted with permission from D. Zhao, D. Feng, Q. Huo, N. Melosh, G.H. Fredrickson, B.F. Chmelka, and G.D. Stucky, "Triblock copolymer syntheses of mesoporous silica with periodic 50 to 300 angstrom pores," Science 279, 550 (1998). Copyright © 1998 American Association for the Advancement of Science.]

OCR for page 93
Page 135 Synthesis and Processing: Control To make the increasingly complex materials and structures of the coming decades, tremendous advances in processing will be required. One can look forward to the day when arbitrarily complex materials and material combinations can be made with the same level of control as is possible for semiconductors today. One of the first classes of material that is likely to see the benefits of improved processing is the complex oxide family, including high-temperature superconductors. The complexity of these materials, however, pales in comparison with some of the other structures involving vastly different classes of material such as biological and inorganic materials. Although demonstrations of increasingly complex structures designed on the molecular level may be made using scanning-probe techniques, fabricating structures that can be studied intensively will require faster techniques that can make multiple samples. This almost certainly calls for a dramatic increase in our understanding of and ability to use self-assembly and biomimetic techniques to produce and process materials. In the past decade there has been impressive progress in the understanding and control of defects—what they are, where they come from, and how to eliminate them or control their placement when they serve a defined purpose—in some materials systems, especially semiconductors. For other materials to reach the same level of perfection and processing control, the same level of understanding will be required. Nanoscale fabrication and processing, wherein molecular chemistry and condensed-matter physics merge, will be key to achieving the level of control that will be needed to realize many of the exciting possibilities posed in this report. Physics: Understanding The materials and structures on the horizon offer rich possibilities for condensed-matter and materials physicists. More perfect materials will enable us to move toward developing a full understanding of the relationship between the detailed structure of a material and its properties. The ability to control defects will enable them to be studied themselves—how they interact with the material they inhabit and even how judiciously assembled collections of them interact with one another and with different defects. Advances of the past decade in probing surfaces and interfaces on the atomic scale offer the possibility that a full understanding of the initial stages of growth in systems more complex than silicon may one day truly be possible. Control of the structure of materials on various length scales simultaneously offers the opportunity to look for effects that result from the interplay of structure on these different length scales. Technology: Relevance Advances in new materials and structures have dramatically improved our

OCR for page 93
Page 136 lives in the past, and there is every reason to expect that new advances will have comparably great impact in the years to come. For this to happen, sustained research will be needed over many years. This research will need to have a balance between fundamental investigations into the physical mechanisms at play and research and engineering aimed at investigating the numerous questions that must be answered before a material can enter the technological mainstream: What can the material be used for? Is there a potential market of sufficient size to pay for the needed research and development? Is the advance so revolutionary, with improvements in customer capability so great, that it can found a new industry? If the improvement is in an area already occupied by an existing technology with significant infrastructure, can the material be integrated with the existing technology? And if so, is the improvement worth the development cost? Just as revolutionary advances in new materials and processes enabled the transistor, the optical fiber, the solid-state laser, and many other technologies that have improved our lives and strengthened the economy, new developments in materials and structures hold out the promise of revolutionary breakthroughs in the twenty-first century. Outstanding Scientific Questions • Can we complement empiricism with predictability in our search for new materials and structures with desired properties? Can we predict the composition and structure of a new material, its properties, and how to synthesize it? • Can we develop a full understanding of the initial stages of growth? • Can we develop a full understanding of the relationship between the detailed structure of a material and its properties? Can we truly control defects? Research Priorities • Tailor materials at the molecular level. • Use more complex combinations of materials: polymers, organic molecules, biological molecules, etc. • Develop new tools to synthesize, visualize, characterize, and manipulate new materials and structures. • Make increasingly complex materials and combinations with as much control as is currently possible in the making of semiconductors. • Increase our understanding of and the ability to use self-assembly and biomimetic techniques to produce and process materials. • Merge molecular chemistry and condensed-matter and materials physics to understand and control fabrication and processing on multiple length-scales. • Integrate processing of new materials and structures with existing technologies.