Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
CNew Materials INTRODUCTION The creation of new materials, always a driving force toward advances in condensed-matter physics, has in the past decade become an even more important, perhaps dominant, factor as our understand- ing of general physical phenomena (e.g., conductivity, magnetism, and superconductivity) has progressed to the point where intellectual challenge is now provided by departures in specific materials from the general behavior, rather than in this behavior of the class of materials as a whole. In many cases these departures represent the limiting behavior that also makes the phenomenon of technological value, for example, the highest stored energy density or lowest hysteresis loss in magnetic materials, the highest mobility in semiconductors, or the highest critical temperatures, magnetic fields, and currents in super- conductors. In this appendix we first identify the materials that have had a major impact on condensed-matter physics in the past decade, second understand how they were discovered, and third recommend how the chances of such discoveries can be maximized in the future. Given the record of the past, such discoveries are essential if the vitality of condensed-matter physics is to be assured. We define as new materials those that have led to the observation of new physical phenomena. Thus, new processing techniques, unless 248
APPENDIX C 249 SCIENTIFIC/TECHNICAL I MOTIVA TION | SYNTHESIS | CHEMI CA L /S TRUC TURA L CHARA CTERIZA TION 1 1 PHYSICAL CHARACTERIZA TION / BASIC KNOWLEDGE/TECHNICAL APPLICATIONS FIGURE C.1 The synthesis loop. they have resulted in the synthesis of materials with unusual proper- ties, will at most be only briefly mentioned. An underlying theme of this section is the importance of the . synthesis loop (Figure (~.1). This 1nterrelatlonsulp between synlnesls, the initial characterization of materials, and the evaluation of these materials, from which may come (unpredictably) either basic knowl- edge or technical applications, must be closed by feedback of these results to the individuals carrying out novel and creative synthesis. The motivation for completing this synthesis loop can be scientific or technological. NEW MATERIALS IN THE LAST DECADE One- and Two-Dimensional Transition-Metal Chalcogenides This family of materials has created great activity as a result of the observation, in 1974, of charge-density waves in them (see Chapter 11. This has led to questions of how charge-density waves move and of their pinning and to the prediction and observation of boundaries between commensurate and incommensurate regions of the wave. The
250 APPENDIX C charge-density wave has been observed directly by atomic-resolution electron microscopy (see Figure 1.4~. Materials with Open-Crystal Structures Although intercalated graphite is an old material, there has been considerable interest in the staging of the intercalants, the steps whereby spaces between the two-dimensional carbon layers become filled. In solid electrolytes the crystal structure is open enough that ion transport can occur through the material. Activity in this field is high, driven by the need for massive batteries for load leveling and mobile power sources. Some of the chalcogenides are also fast ion conductors when a third element (such as Li) is added. These materials are both electronic and ionic conductors and could be used as electrodes (not electrolytes) in high-energy density batteries. Zeolites can be used to support extremely small metallic particles or can be filled with metal, which results in an interesting network of filaments whose diameters are ~10 A. The Magnetic Superconductors The history of this field is a fascinating one with respect to the question of whether superconductivity and magnetism could simulta- neously exist in the same material (see Chapter 8~. In 1967 and 1971, members of a new family of ternary materials, now known as the Chevrel phases, were shown to exhibit superconductivity. Another new ternary compound, ErRh4B4, was first shown to be supercon- ducting at 8.6 K but then to return to the normal state with the onset of ferromagnetism at ~ 1 K. Chevrel phases were shown to exhibit superconductivity, which was then destroyed by ferromagnetism at a lower temperature. There followed exceptionally productive examples of the synthesis loop, where an extremely close coupling between basic scientific evaluation, theoretical input, and synthesis led to rapid progress and understanding of how superconductivity, ferromagnetism, and antifer- romagnetism can coexist. Organic Conductors A major advance in the past decade was the discovery that organic materials exhibit not only high metallic conductivity but also
APPENDIX C 251 superconductivity (see Chapter 8~. The materials of interest fall into a few groups. The so-called Bechgaard salts have remarkable transport properties (e.g., an unusually strong decrease in resistivity with decreasing temperature, whose rate is strongly affected by an applied magnetic field). Superconductivity was observed initially in one of them at ambient pressure and also in the organic compound polyacetylene, but under pressure. Although technically of consider- able interest, the fascination of this material lies in its physical properties, especially the nature of the charge carriers (solitons; see discussion in Chapters 2 and 10) produced in it by doping. New Superconductors Just before his untimely death B. T. Matthias discovered Ba(PbxBi~_x)03, which has the highest transition temperature (Tc) for compounds not containing a transition metal. We now know that this Tc is due to extremely strong electron-phonon coupling (the well- known mechanism of superconductivity) in a material with a low density of electron states rather than to an unusual coupling mecha- nism. More recently heavy-fermion superconductors have been dis- covered, having relatively low TC but anomalously large values of the electron mass, examples of which are CeCu2Si2 and UBe~3. Glasses This broad field encompasses the structure of glasses, transport and the metal-insulator transition, thermal properties at low temperatures, spin glasses, amorphous semiconductors, solitons in high-purity opti- cal fibers, and the effects of disorder and localization on superconduc- tivity. The synthesis methods used in the creation of glasses include not only rapid quenching of bulk materials but also the quenching of films onto low-temperature substrates, first used to produce an amor- phous material in 1954. Artificially Structured Materials The most remarkable advance in the past decade was the creation of what are known as artificially structured materials (see Chapter 1~. The classic example of this synthesis technique is the growth of semicon- ductors by molecular-beam epitaxy, which defines not only their layers but also the specific location of their constituents, dopants, and charge carriers. In addition to the enormous technological importance of this
252 APPENDIX C synthesis technique, it has led directly to new physical phenomena, including the quantized Hall effect and the fractionally quantized Hall effect (details are given in Chapter 14. IMPACT OF NEW SYNTHESIS TECHNIQUES ON CONDENSED-MATTER PHYSICS Thin Metallic Films A rapidly expanding field is that devoted to the conductivity of materials at low temperatures, especially those that undergo a transi- tion from metallic to insulating behavior because of disorder or reduction in their size. Interest in this field was stimulated by specific predictions of the maximum resistivity that can be reached in metals and made possible by advances in low-temperature techniques (Chap- ter 8), by improvements in thin-film deposition methods, and by the development of the photolithographic and processing techniques that allow patterning of the deposited films to dimensions << 1 lam. Popular materials for these studies include the two-dimensional electron gas in MOSFETs, in which the carrier density is tuned by the gate voltage, and codeposited mixtures of metals and insulators, where the carrier density is varied by changing the insulator content. Such studies have given new insight into the nature of electronic transport, elastic- and inelastic-scattering processes, the localization of electrons in disor- dered metals and the interactions between them, and even the phase coherence of the electron wave function in nonsuperconducting met- als. Epitaxial Materials The remarkable success of epitaxial growth in the synthesis of semiconductor materials has created new interest in many other systems, for example, metal on semiconductor, metal on metal, semiconductor on insulator (or vice versa), and insulator on insulator. A heavily studied class of metal-on-semiconductor materials is the family of transition-metal silicides on silicon, which have technological importance as conductors in VLSI circuits. One can grow such silicide films as perfect single-crystal films by a novel technique that uses a template for epitaxial growth. The growth of an epitaxial insulator on semiconductors again has technological implications, especially in the case of semiconductors
APPENDIX C 253 that do not naturally grow an insulating oxide with the quality of silicon oxide. Interestingly, the degree of lattice match between insulator and semiconductor does not seem to determine the quality of epitaxial growth. An ultimate aim of this research is the ability to grow repeated insulator and semiconductor layers, which then might be processed into three-dimensional circuits. The properties of such thin semicon- ductor layers, especially when patterned to <<1 Em in the lateral dimension, will be of interest in determining the physical limits on the sizes of electronic components. The deposition of metals on insulators has begun to benefit from advances in vacuum and surface-science techniques, and the growth of such layers is now taking place in ultrahigh vacuum, with in situ analysis of the growing film. The ability to control substrate tempera- ture and to deposit simultaneously or sequentially from multiple sources has led to the development of extremely versatile materials synthesis systems and may, to some extent, replace the traditional synthesis of bulk samples. Control of the substrate temperature gives a new dimension to this synthesis method, allowing the easy growth of amorphous and metastable phases. As an illustration, the highest superconducting transition temperature, that for Nb3Ge (T`. = 23 K), was achieved only by vacuum or chemical vapor deposition of Nb and Ge onto a carefully temperature-controlled substrate. Control of metastability has also been achieved by metal-on-metal epitaxy. Studies of metal-metal interfaces using ion channeling indicate that interracial strain can induce lattice match between two metals (or semiconductors) over moderately long growth distances. This results in an alternative method for the synthesis of metastable Nb3Ge, namely by epitaxy with stable Nb3Ir, which has the same lattice constant. Nb3A1 was also stabilized beyond the bulk-phase boundary by self- epitaxial growth, namely by continuously changing the Nb/A1 ratio away from the stable composition during growth of the film. Modification of the physical properties of a surface, and thus indirectly of the bulk material, has also been achieved by metal-on- metal overlayers, although epitaxy is not essential. For example, the absorption of hydrogen by bulk niobium was drastically increased by the deposition of a few monolayers of palladium on its surface. The inactivity of Au for catalyzing the rate of cyclohexane dehydrogenation to benzene was increased above that of Pt by the presence of two monolayers of Pt on the Au surface. The activity of bulk Pt for the same reaction was increased fourfold by the addition of a monolayer of Au.
254 APPENDIX C Metallic Superlattices The logical extension of work on metal epitaxy is to repeat the process many times, thus forming a metallic superlattice. The high level of recent activity in this field primarily addresses questions of structure and how the behavior of these multilayer materials is largely determined by the interfaces, which become essentially the whole of the material as the superlattice period is reduced. In only one system, Nb/Ta with perfect lattice matching between Nb and Ta, do films have an electronic mean free path that is appreciably longer than the period. In others, the mean free path is limited by scattering at the interfaces. In Nb/Ta, the phonon distribution function, measured by tunneling, clearly shows the alloying at the interfaces, as does x-ray scattering. Two surprising observations have come from these studies to date. One is the anomalous change in the elastic properties of some superlattices as the period is reduced to about 30 A. The other is the ease with which it appears possible to induce new lattice structures by epitaxy and interracial strain. Thus Zr can be stabilized in the bee phase in Nb/Zr superlattices, as can Co in Co/Cr. Both Mo/V and Nb/A1 have been shown to grow as strained superlattices, in exactly the same way as the strained semiconductor superlattice described in Chapter 5. Superlattices of metal with insulator, for example, Nb/Ge, constitute ideal systems for the study of dimensional effects in metals, in this case the crossover from two-dimensional to three-dimensional supercon- ductivity as the Ge thickness is reduced. It is clear that such research will continue with the goal of achieving single-crystal metal/insulator (or semiconductor) superlattices. Materials Modification New techniques for the modification of materials have emerged that both challenge current theories of crystal growth and have technolog- ical value. For example, ion beam/solid interactions are being used to dope and amorphize semiconductors and to produce new metallic alloys by implantation and ion mixing, while the interaction of ion beams with organic films produces conducting layers and has applica- tions to lithography. Laser processing of semiconductors can result in single-crystal growth and in anomalously high levels of incorporation of dopants. Laser annealing of semiconductors promises to improve the quality of devices and save billions of dollars for the electronics industry. Laser
APPENDIX C 255 alloying is also possible, creating alloy layers of extended solutions that are otherwise difficult to make. Lasers can be used to modify the structure of surface layers or films in an important manner. Depending on the laser pulse width used in heating, laser annealing or quenching can yield either amorphous or crystalline films of various materials. For surface modification, laser-induced photochemical deposition, chemical etching, and electroplating have been demonstrated to pro- vide order-of-magnitude improvement in speed over conventional methods. The fabrication of regular arrays and random distributions of metal particles in the 100-1000 A size range by deposition techniques, laser processing, and microlithography has helped to elucidate the phenom- enon of surface-enhanced Raman scattering. This is important as a probe of surface structure and adsorption and has improved our understanding of optical processes at surfaces. Explosive techniques are beginning to be used for the preparation of novel materials. For example, the A15 superconductor Nb~Si has been prepared with its highest transition temperature by this method, which also induced anomalous properties in the semiconductor CdS. Filamentary Materials By drawing from an ingot containing two components (say wires of Pt embedded in a Cu cylinder) it is possible to produce a fine wire in which the Pt exists as filaments <1000 A in diameter. For example, drawing a Cu ingot containing a dispersion of Nb particles results in a multifilamentary wire that can be reacted to form superconducting Nb3Sn filaments. The interest to date in these multifilamentary mate- rials has centered largely on their mechanical properties, but it has also been shown that by etching away the Cu matrix, fine single filaments can be produced that have unusual physical properties: high electrical resistivity and high strength, for example. Given the interest, men- tioned earlier, in systems with dimensions <1 ,um, this alternative method of making free-standing material with no strain at a substrate should be explored further. Metal Clusters The production of small clusters of a material by its evaporation into an inert gas, so that the particles agglomerate before condensing on a substrate, has been used for some time to produce three-dimensional samples of very small size, say <100 A in diameter. Recent interest in
256 APPENDIX C the physical properties of these materials is beginning to address the question: How large is a particle before it assumes the bulk properties of the material? A new approach to studies of metallic clusters has been to measure their mass distributions by time-of-flight techniques. There is clearly indication that certain (magic) numbers of atoms in a cluster are most stable. Novel sources of such clusters have been developed, including one in which a pulsed laser is used to vaporize a metal in a . . supersonic expansion nozz e. MAJOR CONCERNS From the discussion above, it is clear that materials that have received the greatest attention in the last decade can be divided roughly equally into bulk materials and artificially structured materials. The major concerns of the 1980s that can be identified from what has been said in this appendix are different with respect to these two classes of materials. In the field of artificially structured materials the central and crucial role of synthesis to the success of any program of research has emerged in a natural way. Either one person is responsible for all as- pects of synthesis, characterization, and evaluation or a closely knit group has emerged that is built around state-of-the-art facilities for synthesis, say of semiconductors by molecular-beam epitaxy (MBE) or metallic films by codeposition, or for processing to submicrometer di- mensions. In the case of processing, state of the art is an overstatement except in a few laboratories: in the majority the inventiveness of the individual scientist has to overcome the limitations of antiquated equipment. Thus, many novel methods for producing submicrometer structures have emerged. The main concern in this area is the extremely high cost of initiating research. For example, an MBE system costs about $500,000, a high-vacuum metals codeposition system with no surface characterization tools in place costs more than $300,000, and an electron-beam pattern-writing machine costs $1 million (although electron microscopes can be modified for this purpose at lower cost). This natural role of synthesis should be contrasted with the situation in research into new bulk materials, where synthesis is fragmented if it exists at all. PROJECTIONS FOR THE FUTURE Judging from the directions taken by condensed-matter physicists in the field of new materials in the past decade, some projections can be made into the next decade.
APPENDIX C 257 First, driven by technology and scientific challenge, great progress will be made in artificially structured materials on length scales that will be reduced below 100 A. The epitaxial growth of metals, insulators, and semiconductors in all combinations and as superlattices, and of clusters, will lead to new phenomena and new devices. Second, in studies of bulk materials, physicists will continue to extend their interest away from perfect single crystals of simple materials to materials with complex structures and large unit cells, especially structures with internal clusters or internal channels where properties can be modified by intercalation or ion insertion. Third, the U.S. condensed-matter community in general will con- tinue to contribute to materials evaluation of bulk materials rather than to their synthesis. Fourth, the evidence of the past decade is that the discovery of new materials will lead to the discovery of new physical phenomena. In some way, funding of science in the United States must be structured to encourage this to happen.
