National Academies Press: OpenBook

Frontiers in Crystalline Matter: From Discovery to Technology (2009)

Chapter: 2 Science and Technology of Crystalline Systems

« Previous: 1 Introduction
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 33
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 34
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 35
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 36
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 37
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 38
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 39
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 40
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 41
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 42
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 43
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 44
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 45
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 46
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 47
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 48
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 49
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 50
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 51
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 52
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 53
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 54
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 55
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 56
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 57
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 58
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 59
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 60
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 61
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 62
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 63
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 64
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 65
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 66
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 67
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 68
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 69
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 70
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 71
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 72
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 73
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 74
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 75
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 76
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 77
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 78
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 79
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 80
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 81
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 82
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 83
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 84
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 85
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 86
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 87
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 88
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 89
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 90
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 91
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 92
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 93
Suggested Citation:"2 Science and Technology of Crystalline Systems." National Research Council. 2009. Frontiers in Crystalline Matter: From Discovery to Technology. Washington, DC: The National Academies Press. doi: 10.17226/12640.
×
Page 94

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.

2 Science and Technology of Crystalline Systems The importance of symmetry in nature has been recognized since the time of the Greek philosophers. Today, new applications of symmetry continue to influence scientific thought in many fields, ranging from biology to astrophysics. In particular, the description of how symmetry in condensed matter is lowered or “broken,” such as when a liquid becomes a crystal, forms a mathematical connection with other disparate phenomena, such as the handedness of an oligomeric protein to the evo- lution of matter after the big bang. Indeed, the remarkable properties of graphene (see Box 2.1) originate from the unique way in which a single sheet of carbon atoms breaks spatial symmetry. Symmetry is described mathematically through the theory of space groups, which are the set of spatial translations and rotations that leave a crystal structure unchanged. The structural symmetry is defined at high temperatures, when the crystal forms. At lower temperatures, interactions of lower energy than that of the interatomic bonds yield states such as superconductivity, ferromagnetism, and ferroelectricity, which then lower the symmetry of the interacting particles and fields even farther. Crystal lattices can also serve as a laboratory for studying broken symmetries that arise in a much different context: for example, the generation of elementary particles. Elementary particle theorists often impose the symmetry of a lattice to calculate the consequences of subnuclear interactions such as the Higgs mecha- nism, at present being sought at the European Organization for Nuclear Research’s (CERN’s) Large Hadron Collider. Such theoretical concepts can also be realized in crystalline solids. For instance, certain magnetic systems on chain-like crystal 33

34 Frontiers in C rys ta l l i n e M at t e r BOX 2.1 Graphene Elemental carbon adopts a variety of crystalline forms—three-dimensional diamond, l ­ayered graphite, carbon nanotubes, and C60 fullerite. In 2005, a new form of carbon, graphene, was discovered. Graphene is a single sheet of hexagonally ordered carbon atoms. The purely two-dimensional nature of graphene sheets gives rise to an astounding array of new phenomena, among which are the following: • Behavior that mimics the relativistic motion of particles in high-energy accelerators, • New states of matter in the quantum Hall regime (in Chapter 1 of this report, see the subsection entitled “Example in the Area of Thin Films: Gallium Arsenide-Based H ­ eterostructures”), • Electronic conductivity at zero electron density, and • Extremely fast “ballistic” motion of electrons and holes even at room temperature. The latter property suggests a new class of electronic devices with switching speeds much greater than those achievable in silicon complementary metal oxide semiconductors, perhaps reaching terahertz frequencies. These properties and more result from an unusual electronic momentum-energy relation­ ship. Electrons in the hexagonal crystal structure of graphene behave like massless ­ relativistic electrons in a world with only two dimensions. Many materials possess a quasi-two-­dimensional hexagonal structure in which the sheets interact slightly with neighboring sheets, such as in g ­ raphite itself, thus breaking the special momentum-energy relationship of electrons in ­graphene. What makes graphene special is that the sheets are only one atom or a few atoms thick. The technical breakthrough that led to an explosion of research into graphene (now more than 1,000 papers per year; see Figure 2.1.1) was the discovery that crystals with a thickness of only a nanometer can be seen under an optical microscope when the crystals are placed on a Si wafer coated with a layer of silicon dioxide (SiO2) (see Figure 2.1.2). The SiO2 layer thick- ness must be precisely engineered—300 nm will not work, whereas 315 nm will—to produce interference contrast with the graphene crystal. Thus, the discovery of graphene is an instance of the combining of a novel measure- ment approach with a prosaic “synthesis” technique. A large part of the future challenge for creating graphene-based devices will be that of replacing these techniques with a scalable manufacturing process that does not sacrifice the unique properties of this remarkable form of crystalline matter.

Science and Technology of Crystalline Systems 35 1200 Number of “Graphene” Papers 1000 800 600 400 200 0 2003 2004 2005 2006 2007 2008 Year FIGURE 2.1.1  The number of papers published on an annual basis from 2003 through 2008 figure 2-1-1.eps that relate to graphene, revealed by a citation search for “graphene” on the Web of Science. FIGURE 2.1.2  Photograph of an approximately 3-nanometer-thick graphene flake on top of an oxidized silicon wafer. SOURCE: From K.S. Novoselov, A.K. Geim, S.V. Morozov et al., “Electric Field Effect in Atomically Thin Carbon Films,” Science, 306, 666 (2004); reprinted with permission from the Americanfigure 2-1-2.eps Association for the Advancement of Science. bitmap

36 Frontiers in C rys ta l l i n e M at t e r lattices are described by the same mathematics used to model elementary particle mass generation. Thus, crystalline symmetry provides an intellectual connection spanning 15 decades in energy! Crystalline systems are exemplars of condensed matter. The usual image of a crystal is one of clarity, which originates from the perfect ordering of atoms. In the laboratory, symmetry is manifested in regular x-ray scattering patterns, a phenomenon discovered by Sir William Lawrence Bragg and his father, Sir W ­ illiam Henry Bragg, for which they shared the 1915 Nobel Prize in physics. The microscopic ­crystal images shown in various figures in this report are constructed mathematically from such x-ray diffraction patterns; the produced images have a type of beauty appreciated especially by the scientists studying them. An additional measure of perfection comes in the physical properties of crys- tals. Examples abound in which “fragile” states are able to form only in crystals possessing a very low density of defects. One example is the set of fractional quantum Hall states discussed in Chapter 1, which are not seen in samples with crystalline disorder. Advances in measurement technology enable the probing of ever-higher degrees of crystalline perfection, allowing scientists to better match theories with experi- ments on exotic properties of matter. Currently, no general theoretical prescription exists for many of the new states of matter that emerge with increasing crystalline perfection. Discovering and understanding such states demand continual and close collaboration among synthesis, experiment, and theory. Controlling and minimizing defects in crystalline materials also constitute an important path to device innovation. Many next-generation devices for applica- tions such as solar energy, solid-state lighting, and novel sensors require crystalline order among nontraditional atomic, molecular, or nanoscale building blocks. Con- trolling such ordering will be key to creating devices that will form the foundation of these future technologies, just as controlling covalently bonded semiconductors enabled today’s microelectronics industry. Innovation involving crystalline materials has historically benefited from a fruitful interplay between basic research and device development, as described in Chapter 1. This interplay was guided by large industrial laboratories that oper- ated significant basic research divisions whose scientific agenda was to address technology goals of the corporation to which they belonged. Research scientists and engineers outside these corporations thus had a window onto technology roadmaps through interactions with fellow basic researchers in industry. Each of these laboratories often employed 100 or more discovery and growth of crystal- line materials (DGCM) scientists, including many crystal and film growers. They provided generally stable and substantial operating funds for DGCM activities and had the ability to respond rapidly to new materials opportunities without the need to pursue new funding to do so.

Science and Technology of Crystalline Systems 37 In the current business climate, U.S. corporations no longer sustain such basic research divisions, with the result that an important mechanism for communicat- ing technology needs has been greatly diminished. The impact of this loss is evi- denced by the reduction in industrial publications in physics from 998 in 1988 to 312 in 2005 (see also the discussion in the section entitled “The Role of Industry in Crystal Growth” in Chapter 3). The impact of this loss on crystalline materials could also be high, considering that most modern crystal growth techniques as well as a significant amount of materials discovery in the past 50 years were achieved in the integrated corporate laboratory environment. Achieving the scientific vision expressed in this chapter while addressing an increased need to innovate will carry with it the challenge that future research activities be organized to emulate the modalities of interaction in former industrial DGCM research environments. The recommendations of the committee presented in Chapter 4 include suggestions for new approaches that go beyond the industrial research model to accommodate the changing needs of DGCM researchers in academia, national laboratories, and industry. In the major sections immediately following, the committee presents a vision for the future of both bulk and thin-film crystalline systems in the form of three grand challenges. This chapter then concludes by discussing the needs for applied crystal growth in technology development and the role that characterization will continue to have in new crystalline materials discovery. GRAND CHALLENGES IN THE SCIENCE AND TECHNOLOGY OF CRYSTALLINE MATERIALS Advances in the science and technology of condensed matter will continue both to challenge and to enable society’s understanding of the physical world for the foreseeable future. Crystalline matter is at the heart of many of the most exciting areas of research, as shown by three grand challenges: • Grand Challenge 1. The Development of Next-Generation Crystalline ­Materials— New States of Matter and New Materials—for Future Information and Com- munications Technologies The ability to tailor both the symmetry and dimensionality of a crystalline lattice allows for the creation of a delicate balance between dissimilar ground states, especially in crystals supporting strongly correlated or quantum-mechanical interactions among the electrons. Such appropriately tailored crystals can exhibit completely new behavior: quantum phenomena such as superconductivity or quantum coherence at room temperature, the ability to switch and store multiple types of signals (electric, magnetic, and structural), or the attributes of physical

38 Frontiers in C rys ta l l i n e M at t e r states not anticipated by present theory. With respect to this grand challenge, the ensuing physics will be so unexpected that scientists possess a priori, besides a few general design guides such as effective dimensionality, little understanding of how atoms should be assembled. Thus, this challenge is tantamount to creating a research ecosystem for discovery-based inquiry in crystalline matter. • Grand Challenge 2. The Creation of New Crystalline Materials for Energy Production and Conversion Future energy technologies will require dramatic advances in crystalline mate- rials, including thermoelectric materials for heat-to-electricity conversion, solar photovoltaic materials for sunlight-to-electricity conversion, novel materials for hydrogen production and storage, new electrode and membrane materials for fuel cells, and affordable catalysts for feedstock-to-fuel conversion. Many such energy applications will require crystalline phases with low parasitic energy loss, such as low-cost solar cells with 50 percent power efficiency. Each individual topic in the area of crystalline materials for energy production and conversion requires a stretch goal for materials performance involving an unprecedented degree of atom-level control. Collectively these topics constitute a grand challenge for increasing energy availability and self-sufficiency in the United States. • Grand Challenge 3. Evolution in the Capacity to Create Crystalline Materials by Design In the next 10 years, advances in theory and modeling, coupled with dramatic increases in computational power, will enable the creation of completely new materials incorporating either thin-film or molecular subunits. More than ever before, these new materials will possess specific device functionalities. Realizing a materials-by-design approach would enable industry to move beyond current device limitations to provide cheaper and more efficient solar cells, high-power electronics, and devices with functionality not yet imagined. Grand Challenge 1: DEVELOPMENT OF Next-Generation Crystalline Materials—New States of Matter and New MATERIALS—FOR Future Information and Communications Technologies In analogy to the traditional states of matter—solid, liquid, and gas— r ­ esearchers have discovered a multitude of fascinating states in crystalline ­materials. These states depend on the effective spatial relationships between atoms within the crystal. Of particular importance is effective dimensionality and connectivity of

Science and Technology of Crystalline Systems 39 sub­structures within the crystal; for example, materials based on lamellar sheets of atoms can be considered two-dimensional for electronic or magnetic excita- tions. These excitations interact among themselves to create patterns, or states of matter, that are highly dependent on the effective dimensionality and con­nectivity. Such states of matter include modified traditional states of ferromagnetism, ferro­ electricity, and superconductivity. Other, more exotic states such as vortex matter, spin ice, and quantum states with no classical interpretation can occur in such sublattices. It is the ability of crystalline solids to emulate spaces of different dimensionality and connectivity that allows the realization of exotic new states of matter with novel properties. A few examples of future directions of new crystalline materials with novel properties are presented below. Scaffold Structures The challenge of creating novel states is that of using unusual and innovative combinations of elements to create an electronic structure in which there is competi- tion among possible states or forces, and then of fine-tuning the balance with chem- istry. One approach uses an atomic sublattice to serve as the scaffolding for another sublattice. Examples are shown in Figures 2.1 and 2.2. In the first example, the high-transition-temperature (Tc) superconductor YBa2Cu3O7, the super­conducting phase is formed from the sublattice of copper-oxygen (Cu-O) pyramids, with the Cu-O chains and larger yttrium and barium atoms providing the necessary spac- ing to create a two-dimensional electronic structure (Figure 2.1). In addition, the Cu-O chains form a spatially distinct subsystem that acts as a charge reservoir for the two-dimensional subsystem. Another example of atomic scaffolding is in the skutterudites. Skutterudites have the general formula RM4X12, where R is a rare-earth ion; M is iron (Fe), r ­ uthenium (Ru), or osmium (Os); and X is phosphorus (P), arsenic (As), or anti- mony (Sb). A structure is shown in Figure 2.2. Characterized by an open structure and large numbers of synthetic combinations, the skutterudites exhibit a wide array of physical properties, including metal-insulator transitions, heavy-fermion super- conductivity, and large thermoelectric power figure of merit. The latter property is aided by large voids in the crystal structure that allow large thermal vibrations for the atoms that reside in them—the atoms rattle around. This rattling is respon- sible for unusually large phonon scattering, which reduces thermal shorting in a thermoelectric cooling application. Another example of scaffolding is found in the molecular compound [BEDT- TTF]Mn[Crox3] (BEDT-TTF is bis(ethylenedithiolo)tetrathiafulvalene and ox is oxalate). Here, the coexistence of magnetic ordering and metallic-like electrical conductivity occurs due to alternating electrically conducting layers composed of [BEDT-TTF]+ and ferromagnetically ordered layers composed of {Mn[Crox3]}–.

40 Frontiers in C rys ta l l i n e M at t e r FIGURE 2.1  Crystal structure of YBa2Cu3O7 showing the sublattice of the copper-oxygen pyramids (blue) and the larger yttrium (yellow) and barium (green) atoms providing the spacing for a two- dimensional electronic structure. SOURCE: Courtesy of M.A. Subramanian, Oregon State University. FIGURE 2.2  Filled skutterudite LaFe3CoSb12 showing the large open spaces where the lanthanum ions (yellow) rattle to scatter phonons. Iron (Fe) and cobalt (Co) (red) and antimony (Sb) (blue) are also shown. SOURCE: Courtesy of D. Mandrus, Oak Ridge National Laboratory.

Science and Technology of Crystalline Systems 41 The final example is the newly discovered magnetic material of Figure 2.3, in which the presence of the second interpenetrating lattice led to anomalous magnetic switching behavior not observed before. Low-Dimensional Structures As mentioned above, crystalline materials can possess internal scaffolding structures with effective dimensionality lower than three dimensions. Such crys- talline structures possessing either chains or planes of interacting atoms modify the flow of energy for magnetic and electronic excitations, creating platforms for useful devices. This lower effective dimensionality can also lead to entirely new ground states. For instance, when local degrees of freedom are continuously variable, an atomic spin can point in any direction, and fluctuations suppress a long-range ordered state in one or two dimensions. In some cases, excitations in such systems can be “topological,” like knots in a string. Solitons, which are localized waves whose shape is unaffected by usual dispersive effects, are an FIGURE 2.3  Interpenetrating [Ru2(O2CMe)4]3[Cr(CN)6] lattices are depicted in orange and purple. SOURCE: Courtesy of J.S. Miller, University of Utah.

42 Frontiers in C rys ta l l i n e M at t e r example of topological magnetic excitations in one dimension. In two dimen- sions a common topological excitation is the vortex, which resembles a swirling pattern of spins. Topological character implies insensitivity to localized defects and dispersive effects that commonly destroy the coherence of harmonic, or wave- like, excitations. Such features are important for future information technologies; for instance, topological excitations in a two-dimensional electron gas have been discussed as possible stable bits of information in a quantum computer. Magnetic topological excitations are also potential “qubit” candidates, but the understand- ing required to design a material supporting such excitations is in the very early stages of development. Low-Dimensional High-Tc Cuprates Several central themes come together in the cuprate materials that exhibit high-Tc superconductivity, with YBa2Cu3O7 (Figure 2.1) being the best-studied example. These materials have copper oxide sheets that are doped and effectively two-dimensional by means of a superlattice spatially distinct from the sheets. Absent a widely accepted microscopic theory, most contending theories feature the hybridized copper oxygen band as a key component of the superconducting as well as the normal state. The formation of patterns indicative of density varia- tions in a high-Tc compound is shown in the scanning tunneling microscopy data of Figure 2.4. Low-Dimensional High-Thermopower Cobaltates The concepts of low dimensionality, geometric frustration of spin order- ing, structural and orbital degrees of freedom, correlated electron physics, and quantum fluctuations converge to yield the unexpected physical properties of the layered oxide cobaltates. In a structural family based on hexagonal CoO2 layers, these metallically conducting compounds display thermoelectric coefficients two orders of magnitude larger than those observed in conventional metals. Further- more, superconductivity is observed in the hydrate NaxCoO2·yH2O. The origin of these phenomena is not understood but could be due to a combination of low dimensionality and strong correlations unique to the combination of cobalt and oxygen. Structures Leading to Strong Competition of Internal Forces Patterns among low-energy degrees of freedom in a solid (such as those shown in Figure 2.4) can be thought of as distinct phases of matter. New phases are often found when competing forces are finely balanced at the microscopic

Science and Technology of Crystalline Systems 43 FIGURE 2.4  An example of anisotropic superconductivity in a cuprate material is seen in this scan- ning electronic micrograph in which thefigure 2-4.eps dark and light regions represent differing electron density. bitmap SOURCE: Courtesy of J.C. Seamus Davis, Cornell University. level, either chemically by means of a novel combination of elements, or physi- cally, for example by means of magnetic fields. The following are examples of competing states: localized versus itinerant electrons, spin alignment versus spin anti­alignment, or classical versus quantum fluctuations. When systems with finely balanced forces pass from one state to another, their physical response can become very large. An example is colossal magnetoresistance in manganite perovskites. Another example is the very large thermal resistance effect in vanadium oxide, used as the sensor in commercial infrared imaging systems. Physicists describe such behavior as ­“emerging” from the collective behavior of the large number of atoms (~1022) that comprise a solid. The synthesis scientist approaches the chal- lenge of creating such new emergent behavior by using different paradigmatic approaches to crystal growth. Geometrically Frustrated Structures One route to optimizing competing interactions is through geometrical frus- tration of magnetic interactions. In geometrically frustrated materials, the spins that constitute magnetic matter interact antiferromagnetically: interactions favor antiparallel alignment of neighboring spins, as shown in Figure 2.5. Such inter- actions cannot be simultaneously satisfied when the spins occupy a triangular l ­attice, as also shown in Figure 2.5. The inability to achieve a state that minimizes the energy of all two-body interactions, and the resulting low-energy entropy, are called geometrical frustration. This simple paradigm has deep consequences. When spins in anisotropic pyrochlore magnets are allowed to point only up or down, in a manner similar to that shown in Figure 2.5 for a triangle, the spin system

44 Frontiers in C rys ta l l i n e M at t e r ? FIGURE 2.5  (Left) An example of a two-dimensional antiferromagnetic spin arrangement. (Right) The quandary posed for antiferromagnetically interacting spins on a triangular lattice. figure 2-5.eps is formally identical to the hydrogen atoms in water ice; hence the distinct state of magnetic matter in which long-range order is frustrated at low temperature is called spin ice. When the spins are allowed to point in a continuously variable set of directions, like the needle of a compass, other magnetic states, called spin liquids, can arise. In the spinel antiferromagnet ZnCr2O4, geometrical frustration produces novel low-energy degrees of freedom. The near cancellation of the strong interactions between neighboring spins produces new, weakly interacting low- energy excitations described by a theory of topological excitations with long-range i ­ nteractions—a loose analogue of liquid crystals, but one in which the excitations retain their quantum character. Experimental studies of quantum spin liquids in frustrated magnets are in their infancy. Exciting issues include the possibility of making new topological quantum spin states, the use of these states for quantum computation, and the novel couplings between lattice and orbital fluctuations. Especially exciting are the largely unexplored regimes of chlorides, fluorides, and chalcogenides; most exotic spin systems studied to date are oxides. Indeed, the recent discovery of a spin-½ kagome lattice in the hydroxychloride compound ZnCu3(OH6)Cl2 has excited physicists for its indications of a quantum spin liquid and illustrates the potential of searching beyond oxides. Because many of the qualitatively new phases anticipated theoretically are extremely sensitive to defects, experimental progress is often entirely controlled by progress in materials synthesis techniques and in identifying new materials that realize specific lattice symmetries and/or that can be produced in a particularly pure form.

Science and Technology of Crystalline Systems 45 Orbital Degeneracy and Frustration Magnetism in a solid results from both the spin and angular momenta of electrons orbiting the nucleus. When different orbital states have the same energy, quantum fluctuations are increased and classical long-range order is suppressed. This orbital degeneracy is similar to geometrical frustration-induced spin degen- eracy in its close connection to the crystal lattice symmetry, and recent work has provided evidence for both orbital liquid and orbital glass phases in compounds with the spinel crystal structure. Phase-Change Materials When two different structural states are similar in energy, they compete with each other for the right to exist at a given temperature. Such a situation exists in the so-called phase-change materials. These are typically germanium-antimony- tellurium (Ge-Sb-Te) alloys that have, at room temperature, two different stable states, crystalline and amorphous. Either state can be locked in by using one of two different heating and cooling procedures. These procedures use a laser or a cur- rent pulse to produce a confined area (a bit) with elevated temperature. To create an amorphous bit, the material is heated well above the melting temperature for a very short time. To create the crystalline phase, a lower temperature, just above the melting temperature, is induced for a longer period of time. The resulting state, crystalline with higher conductivity or amorphous with lower ­conductivity, is then read either optically or electrically. Phase-change materials are already used in commercial products such as Blu-ray discs and High-Definition/Density Digital Versatile Disc (HD-DVD) storage and are expected soon in products such as random access memory disks for personal computers (PC-RAM) and super- r ­ esolution near-field structure discs. As an information-storage medium, phase- change materials are faster than flash memory and are more readily scalable to smaller bit sizes and therefore higher information density. Despite their advanced state of application, important aspects of the phase-change process, such as the dynamics of melting and recrystallization, are not completely understood. A better understanding of such dynamic, nanoscale crystallization processes will be needed to realize the full benefit of scaling to smaller bit sizes. Heavy-Fermion Metals In certain intermetallic systems, the conducting electrons can interact so strongly with coexisting magnetic atoms that the magnetically ordered state is suppressed and the magnetic character of the atoms is “dragged” along, or hybrid- ized, with the conducting electrons. With respect to the magnetic atoms, there is

46 Frontiers in C rys ta l l i n e M at t e r strong competition between magnetic long-range order and full itineracy of the electrons on the magnetic atom. In the itinerant limit, these electrons become “heavy” in the sense that specific heat and resistivity behave as though the mass of an otherwise free electron was increased. In some materials this increase is three orders of magnitude over the mass of an electron in free space. Examples of such materials include uranium- and cerium-based intermetallics such as UPt3 and CeCu2Si2. Several of these systems undergo a transition to a superconducting state as the temperature is lowered. Since the electron is carrying the vector character of the magnetic atoms, unusual types of superconductivity can occur. For instance, the macroscopic quantum state that constitutes superconductivity can become imbued with an anisotropy that normally characterizes atomic orbitals—for example, the s-, p-, and d-wave (and perhaps higher angular momentum) states. Despite the large amount of phenomenology developed for heavy fermions, there are major gaps in the scientific understanding of these systems. In certain heavy-fermion materials the temperature-dependent properties do not behave like free electrons within the so-called Landau Fermi-liquid description. Strange fractional exponents characterize thermodynamic properties at low temperatures, indicating a critical state that is qualitatively different from that of the well-known Landau Fermi-liquid. This leads to important questions: What other conducting states are possible beyond the Landau Fermi-liquid? Are there distinct universal states of which the Landau Fermi-liquid is but one, or are some of the phases part of an extended critical phase with nonuniversal and continuously varying exponents? Answers to these questions would represent major progress in the understanding of strongly correlated quantum systems. Other important questions from the experiments include these: Why are heavy-electron states so sensitive to crystal quality? Why are there so few heavy-electron superconductors? How does a nascent heavy-electron material with very few magnetic atoms (a so-called Kondo system) develop into a heavy-electron material? Extreme Conditions Another route to controlling competing interactions and creating new states is through the introduction of extreme environments. In many strongly correlated materials, energetically similar ground states compete for space in the crystal, which can lead to the coexistence of multiple ground states, such as superconductivity and charge order. Applying pressure, high magnetic or electric fields, or extreme temperatures can suppress one of these states while enhancing the other. For basic science, extreme environments offer the possibility of continuously tuning proper- ties to access unique critical parts of phase space that may be inaccessible through materials synthesis without introducing disorder. A practical outcome could be dramatically enhanced properties, such as stronger structural alloys or a higher-

Science and Technology of Crystalline Systems 47 performance superconductor stabilized at ambient conditions following synthesis under multiple extreme conditions. Multifunctional Structures Materials are often grouped into classes according to their functionality (see Appendix E). Metals, ferromagnets, and nonlinear optical materials, for example, are all defined by their response to external fields and resulting functionality when incorporated into a device. New classes of multifunctionality can be created by combining two (or more) distinct properties in one material such that there is coupling between these states. Alternatively, for materials where such coupling is small, multifunctionality can be thought of as the superposition of two desired responses in a single material. While the physics of such systems might not be new, they offer the possibility of reducing weight, power, and cost for device applications. Materials with anomalously large coupling to multiple stimuli such as magnetic and electric fields might, for example, lead to new information-storage systems. Since the functionality of a material often derives from a specific atomic element, creating a multifunctional material often involves creating compounds with specific types of atoms situated in the crystal structure so as to maximize their interaction. The advent of molecule-based magnets has heightened opportunities for magnetic materials with multiple properties. For example, light control and electrochemical control of magnetic behavior have already been demonstrated. (For a discussion of the thin-film approach to multifunctionality, see the sub­ section on “Thin Films,” under Grand Challenge 3). Much of the recent research in multifunctional materials has focused on magneto­electric response, discussed below. There are, however, many other generic types of multi­functionality, such as electrochromic, magnetostrictive, magnetooptic, and electro­optic that are com- prised by this field. An example of a magnetoelectric multifunctional material is shown in Figure 2.6. The number of specific atoms known to induce a single functionality is usually small. For instance, ferromagnets usually possess cobalt, iron, nickel, or manganese. The number of combinations allowed between atoms from different classes is large, however, particularly when different crystal structures are taken into account. In the above example of magnetoelectric multifunctional materials, the search for new compounds also requires synthesis and characterization of single crystals, given the need for the intrinsic response of the dielectric as well as the need to access anisotropic structures. Other examples of multifunctional properties are magnetooptic, magnetostrictive, and magnetic semiconducting responses. A material displaying any of these responses, especially at or above room temperature, with adequate sensitivity is all but guaranteed to have a major technological impact.

48 Frontiers in C rys ta l l i n e M at t e r FIGURE 2.6  Crystal structure for TbMnO3 showing atomic positions at room temperature (upper panel [a]) and the spatial variation and local spin direction along a given direction (lower panel [a]). (b) At low temperatures the spins develop an oscillating component along the c-axis so as to form an inver- sion symmetry-breaking spiral. This state is accompanied by inversion symmetry-breaking atomic displacements, hence ferroelectricity. Panels (c) and (d) show the electric polarization of TbMnO 3 in two different crystallographic directions as a function of temperature for different values of applied magnetic field. SOURCE: Reprinted by permission from Macmillan Publishers Ltd: Nature, T. Kimura, T. Goto, H. Shintani et al., Nature, 426, 55 copyright (2003). Electric Field Control of Magnetism, Magnetoelectrics, and Multiferroics The previous subsection describes an approach to magnetoelectric multiferro- ics that relies on novel combinations of atoms in a bulk crystal. New approaches to control and alter magnetism with an electric field, or, conversely, altering a ferroelectric state with a magnetic field, suggest novel thin-film multiplayer struc- tures. In the schematic example shown in Figure 2.7, control of magnetism with an electric field requires a coupling mechanism at the interface that inherently breaks time reversal symmetry, which is required to switch the state of magnetization. In contrast, strain coupling can be used to modulate magnetism, for example, in heterostructures.

Science and Technology of Crystalline Systems 49 FIGURE 2.7  Schematic illustration of a ferromagnet-multiferroic (FM) (ferroelectric [FE]/ anti­ferro­ magnetic [AFM]) heterostructure that can be used to control and switch ferromagnetism with an figure 2-7.eps electric field (E). The arrows in both layers depict the direction of the magnetic moments. bitmap Electrically Tunable Exchange Bias in Magnetic Heterostructures Exchange bias coupling is well known in magnetism and is exploited exten- sively in magnetic sensing and storage applications. Exchange bias typically occurs between an antiferromagnet and a ferromagnet; strong exchange interactions at the interface “pins” the spins in the ferromagnetic layer. The antiferromagnet in conventional exchange bias structures is typically a metal (manganese, iridium manganese) or a semiconductor (nickel oxide). Ferroelectric antiferromagnets (such as the multiferroic perovskites) present a new and exciting opportunity to attempt to modulate or control the magnetic structure by the application of an electric field. Such effects do not necessarily require a multiferroic material; the presence of a large magnetoelectric effect in the crystal would be sufficient. The presence of ferroelectricity in the crystal provides the opportunity to control antiferromagnetism in a permanent (nonvolatile) way. The next step is to couple the antiferromagnet to a ferromagnet and attempt to switch the ferro­magnetism through the application of an electric field. Indeed, this area of research promises to be exciting and has the potential for new device ­technologies. One such emerging technology is spintronics, which seeks to use the spin of the electron instead of its charge as the fundamental bit of information, which would lower the power consumption in a microprocessor. The use of magneto- electric coupling in multiferroic spintronics would enable the manipulation of spin ­magnetism not by slow and weak-coupling magnetic fields but by fast and strong-coupling electric fields.

50 Frontiers in C rys ta l l i n e M at t e r Search for New Multiferroic Crystalline Compounds Will there be a ferromagnetic-ferroelectric in future information systems? To answer this question, researchers must explore the frontier of materials design at the atomic level to create materials that simultaneously exhibit ferromagnetic and ferroelectric behavior. In many oxide systems, exchange interactions are mediated by the oxygen ions and lead to antiferromagnetic coupling (a key exception is the manganite system that exhibits so-called colossal magneto­resistance). This conundrum appears to present an ideal opportunity for innovative materials and device design. Even if device designers cannot capitalize on an idealized ferromagnetic-ferroelectric at room temperature, it may be possible to ­ create a canted magnetic structure, intermediate between antiferromagnetism and f ­ erromagnetism, that breaks inversion symmetry. Is it possible to tune canted magnetism such that macroscopically useful magnetic moments can be realized in a ferroelectric canted magnet? New Behavior in Artificial Structures and Interfaces The interface between two dissimilar crystalline materials provides a power- ful platform for exploring novel properties and controlling device functionality. As Herbert Kroemer, recipient of the Nobel Prize in physics in 2000, noted in his Nobel Lecture, “The interface is the device.” Starting with the notion of interface states crucial to the operation of the original transistor, the ability to engineer and control interfaces has progressed to present-day deposition techniques whereby atom-level control is a reality. Indeed the ability to control atomic deposition allows materials scientists to create completely artificial materials that have no analogue in bulk crystals. The excitement surrounding interface physics stems from the nascent capability of pushing device design down to the atomic level, where engineering, physics, and computational modeling come together. Conventional Semiconductor Heterostructures Three decades of research and development yielded practical devices such as the field-effect transistor, the quantum-well laser, and the quantum-cascade laser. They also provided insights into fundamental science from unusual quantum states such as the quantum Hall effects and exciton Bose-Einstein condensates. Develop- ing these types of structures in other materials is a priority: new interface behavior could include, for example, the formation of new electronic phases leading to   See http://nobelprize.org/nobel_prizes/physics/laureates/2000/kroemer-lecture.pdf. Last accessed June 3, 2009.

Science and Technology of Crystalline Systems 51 phenomena such as superconductivity, spin-polarized correlated electron gases, spin-transistor devices, new optoelectronic devices, and novel sensors. The ability to create chemically and structurally perfect interfaces is a prerequisite for the study of these electronic phenomena. Advances in deposition tools and instrumentation, discussed below, are needed to enable the control of such interfacial phenomena. The power of heterostructure interface design and engineering is illustrated with a few examples, spanning a spectrum of physical phenomena. Phonons, Phonon Confinement, and the Phonon Laser Tailoring acoustic phonon properties is important for terahertz-frequency phonon devices, including the generation and amplification of coherent phonons. Recently, terahertz acoustic cavities have been demonstrated with enormously amplified acoustic phonon/photon interaction, leading to the possibility of modify- ing the lifetime of optical phonons through tailored anharmonic processes. ­Acoustic cavities could also provide the required feedback mechanism for a phonon laser (see Figure 2.8). Important developments in terahertz acoustics are based primarily on compound semiconductors using mature epitaxial growth techniques, such as molecular-beam epitaxy (MBE), that enable the construction of heterostructures with atomically flat interfaces by design. Heterostructures of oxide materials such as BaTiO3 and SrTiO3 with strong coupling between sound, charge, and light offer an almost completely unexplored, but rich, terrain of versatile compounds with superior acoustic properties. They provide a range of acoustic impedances that can exceed the acoustic impedance mismatches in semiconductor heterostructures. In addition, they can be strongly piezoelectric, providing additional mechanisms that can significantly enhance sound-and-light coupling and allow electrical tuning of acoustic cavity wavelengths. Coupled Phenomena at Interfaces: Interconversion and Coupling Between Electrons, Phonons, and Photons Multiferroics provide a good example of coupling between two types of order parameters. An alternative way to approach materials design and synthesis is to explore the possibility of converting and/or coupling/decoupling interactions between electronic and thermal properties of matter. An obvious example is in thermoelectric materials, in which it is desirable to decouple electronic and thermal transport. The example of the phonon laser described above would be a natural consequence of the ability to confine phonons. Another key area of materials devel- opment is in the conversion of light energy to free-electric charge, as in photovoltaic solar cells. Artificially engineered nanostructures and heterostructures are the most obvious approach for exploring ways to overcome the exciton mean free path, for

52 Frontiers in C rys ta l l i n e M at t e r FIGURE 2.8  Schematic illustration of a device incorporating a resonant cavity for acoustic phonons inside an optical cavity, thereby enhancing interaction between sound and light. SOURCE: Reprinted by permission from Macmillan Publishers Ltd: Nature, Adapted from J.M. Worlock and M. Roukes, “Applied Physics: Son et Lumière,” Nature, 421, 802 (2003). example. Considerable success has been achieved in developing photovoltaic sys- tems with a mixture of materials on the nanoscale (see Figure 2.9). One approach to the optimization of such materials would involve incorporating crystalline molecular structures to replace the present disordered structures. Polariton Bose Condensates Strong ­­coupling between photons and electron-hole pairs in semiconductors is the basis for the conventional semiconductor laser. Using a high-quality Bragg mirror to confine the light, it is now possible to develop a new state of matter in which the photon and the electron-hole pair are coherent (forming a new particle called a polariton). Macroscopic condensation of these excitations leads to a new Bose-Einstein condensate formed at relatively warm temperatures of tens of degrees kelvin (Figure 2.10). With improved device fabrication and more strongly bound

Science and Technology of Crystalline Systems 53 FIGURE 2.9  Composite structure of two polymer materials—­­­(a) APFO-3 and (b) P3HT—to provide electron and hole transport, combined in a nanoscale composite with zinc oxide (ZnO) nanoparticles; (c) transmission electron microscope image of ZnO nanoparticles; (d) highest occupied molecular orbit (HOMO) and lowest unoccupied molecular orbit (LUMO) levels (in electronvolts) of APFO-3, P3HT, and ZnO nanoparticles with respect to common metal electrodes. SOURCE: Reprinted, with permission, from H.M.P. Wong, P. Wang, A. Abrusci et al., “Donor and Acceptor Behaviour in a Polyfluorene for Photovoltaics,” Journal of Physical Chemistry C, 111, 5244-5249 (2007). Copyright American Chemical Society. electron-hole pairs, it might be possible to obtain Bose-Einstein condensation temperatures that approach room temperature. Bulk Crystalline Matter Discovery Challenges Crystalline matter discovery can be broadly divided into two categories: growth of bulk crystals and growth of crystalline thin films. Thin-film growth is most closely associated with a level of control not attainable with bulk-crystal growth techniques. Challenges in thin-film growth are addressed in the section below

54 Frontiers in C rys ta l l i n e M at t e r FIGURE 2.10  (Top) A schematic view of quantum-well (QW) excitons coupled to microcavity photons. (Bottom) The momentum distribution of exciton-polaritons in a cadmium tellurium (CdTe) microcavity, showing the emergence of a coherent Bose-Einstein condensation peak of low momentum states as the density is increased. ���������������������������������������������������������������� Nature, SOURCE: �������������������������������������������������������� Reprinted by permission from Macmillan Publishers Ltd., J. Kasprzak, M. Richard, S. Kundermann et al., “Bose-Einstein Condensation of Exciton Polaritons,”� ���������������������������������������������������� Nature, 443, 409 (2006). entitled “Grand Challenge 3: Evolution in the Capacity to Create Crystalline Mate- rials by Design.” In this subsection, challenges are discussed for the growth of bulk-crystalline material, both in the highly exploratory phase of new materials development and in the growth of large single crystals. New materials synthesis using solid-state chemistry primarily yields poly- crystalline specimens composed of single crystals with dimensions on the order of 100 to 1,000 nanometers (nm). These crystallites are large enough to sup- port phenomena such as ferromagnetism or superconductivity; however, many of the phenomena of current interest are either anisotropic or extremely sensitive to grain ­boundaries, which dictates the need for crystal growth even in the dis­ covery phase. After the initial discovery of a novel phenomenon, single crystals are always required, regardless of the nature of the effect, in order to probe direction- d ­ ependence or defect-sensitive properties. Combinatorial chemistry was developed in the 1990s to accelerate the rate of discovery of new functional compounds. This technique proved to be successful in the search for new pharmaceuticals. In the materials arena, combinatorial chemis-

Science and Technology of Crystalline Systems 55 try is typically based on polycrystalline films with an inherent chemical phase and compositional distribution. While combinatorial chemistry has been successful in a small number of materials searches, the overall impact on materials science has been limited. Flux growth remains an important route to the growth of single crystals with novel functionalities. Recent examples include ZrW2O8 (pictured in Figure 2.22), a cubic material that contracts on heating from 0 to 1050 K, and the 115 family of strongly correlated metals such as CeCoIn5. The importance of these discoveries is reflected by the fact that more than 400 papers referencing one or both of these compounds were published in the year 2007. These compounds were discovered as a result of the ingenuity of individual scientists using little more than a crucible, a high-temperature furnace, and high-purity starting materials. The physical tasks required to crystallize a novel material are time-consuming and have not benefited substantially from automation. A polycrystalline ingot is prepared by mixing component powders in the solid state and performing multiple sinterings for every desired composition. Zone refining is required to grow a crys- tal from such a polycrystalline ingot, which can take from hours to days. For the synthesis of oxides, the installation of several dozen optical float zone furnaces that employ focused high-intensity light as the heat source (see Appendix D) in Japan in the 1990s expanded capabilities for oxide crystal growth to larger crystals, lower defect densities, and more complex stoichiometries. However, as is the case for most other advances in crystal growth technology, this development did not increase the ability to leverage human resources as have advances in measurement techniques. Direct human involvement requiring trained and highly skilled scientists remains an essential aspect of growing new crystalline materials. Grand Challenge 2: creation of new Crystalline Materials for Energy Production and Conversion Crystalline materials will play an important role in future energy systems. Examples abound of the relevance of crystalline materials in a broad array of energy applications that impact all facets of the energy cycle, from energy generation and transportation to storage. For photodiodes and photovoltaic (PV) conversion, single crystals provide a means for reaching ultimate performance limits for energy generation and control. For thermal energy scavenging, thermoelectric devices use monoliths containing small single-crystalline grains. Rechargeable batteries such as those used in portable electronics are based on the mobility of lithium ions between two crystalline electrodes. Crystalline materials also appear in the energy   ISI Web of Knowledge search, http://apps.isiknowledge.com/—–“CeCoIn5” OR “ZrW2O8” in Topic AND “2007” in Year Published.

56 Frontiers in C rys ta l l i n e M at t e r cycle in less obvious manifestations, such as in the energy savings associated with the high-strength crystalline turbine blades discussed in Chapter 1. The impor- tance of energy in every aspect of people’s lives and the potentially large impact of crystalline materials for energy production, storage, and efficiency demand a focused effort. Not only are crystalline materials currently in widespread use, but they will also continue to play a critical role in developing solutions to energy needs in the future. Described below are three representative areas of energy research that rely critically on the ability to control the growth of crystalline materials and that exemplify great opportunities for energy generation and conservation. These areas are solar energy and solid-state lighting (luminaries); superconductivity for electricity delivery, motors, and generators; and catalysts for fuel conversion and hydrogen storage. Band-Gap Engineering for Solar Energy and Solid-State Lighting Solid-state lighting relies on light produced by de-excitation of free charges injected into a junction between dissimilar semiconductors. The converse of this process arises in PV devices (solar cells), in which light absorbed at a junction induces the production of free charges (Figure 2.11). Discovery and growth of crystalline materials activities play a central role in advancing capabilities in these areas. At the basic science level, these two functionalities benefit from many of the same advances in fundamental understanding of energy and charge transport FIGURE 2.11  Simple schematic of a solar cell. The energy from incident light leads to the creation of an electron-hole pair, which separates to create electrical energy. SOURCE: Reprinted from U.S. figure 2-11.eps Department of Energy, Office of Science, Report of the Basic Energy Sciences Workshop on Solar bitmap Energy Utilization, April 18-21, 2005, Basic Research Needs for Solar Energy Utilization, Figure 30, available at http://www.sc.doe.gov/bes/reports/files/SEU_rpt.pdf.

Science and Technology of Crystalline Systems 57 across interfaces in heterogeneous thin-film systems. At the applied level, much of the development in these areas centers on creating new organic and inorganic mate- rials with greater efficiencies but whose large-scale fabrication costs render their use economically viable. While both technologies at present have small markets, they are viewed by the Department of Energy (DOE) and an increasing number of private-sector companies as fertile areas for further product development. The vast majority of today’s PV production is Si-based, single band-gap mate- rial, based on either single-crystal or multicrystalline ingots or on multi­crystalline ribbons. The crystallinity of the material plays an important role in both the effi- ciency and the cost of the device. At present, Si PV devices cost less than $3 per watt of peak power, with the price decreasing roughly as the cube root of cumulative production. The record efficiency for single-crystal Si PV devices with a single p-n junction is currently about 25 percent, and further advances are expected through materials processing research. While the economics of Si PVs are subject to the demand for high-purity material for use in integrated circuits and energy costs for producing high-purity crystalline Si, much research is devoted to improving the trade-off between the efficiency and cost of growth and the purity and crystallinity of ingots grown by the traditional Czochralski process. Most of the innovation in crystal growth is in non-ingot-based technologies such as Si ribbon growth, which circumvents the cost of slicing ingots to produce wafers. Semiconductors such as II-VI systems CdS, CdSe, and CdTe have band gaps larger than Si and thus increase the PV efficiency through increasing the open-circuit voltage. These materials are readily processed using scalable thin-film techniques such as spraying, solution growth, or electrodeposition combined with postdeposition processes to promote crystallinity and purity. The best performance comes from the highest-quality single-crystal-based PV devices, which approach the Shockley-Queisser efficiency limit for single band-gap PVs of 32 percent (Figure 2.12). Grain boundary recombination and related losses associated with noncrystallinity reduce the quantum yield of light-to-electricity energy conversion. Consequently, amorphous Si cells have half the efficiency of the highest-quality single crystals. This gain in efficiency is accompanied, how- ever, by significant increases in energy usage and cost associated with producing high-­quality single-crystal material. The search for alternative, higher-efficiency m ­ aterials has focused on compound semiconductor heterojunction cells made from III-V compounds that circumvent the Shockley-Queisser limit by incorporat- ing multiple band gaps in a heterojunction device. Power efficiencies on the order of 40 percent are currently achieved in three-junction devices. However, the cost of their complex fabrication processes and high-purity raw materials makes them impractical to manufacture at present on a large scale. Further progress will rely on solving the complex materials challenges that arise in fabricating multilayer junctions with high crystallinity, yet at low cost.

58 Frontiers in C rys ta l l i n e M at t e r FIGURE 2.12  Improvements in solar cell efficiency, by system, from 1976 to 2008. NOTE: NREL, National Renewable Energy Laboratory; FhG-ISE, Fraunhofer Institute for Solar Energy Systems; UNSW, University of New South Wales; EPFL, Ecole Polytechnique Federale de Lausanne; AMETEK, Ametek, Inc.; ARCO, ARCO Solar, a division of ARCO; Euro-CIS, European Consortium in CuInSe 2. SOURCE: This figure was developed by the National Renewable Energy Laboratory for the U.S. Depart- ment of Energy. Organic-based materials are potentially attractive candidates for PV appli- cations because of their low formation energies and associated low fabrication cost. Typical fabrication processes involve evaporation or spin deposition from a solution of organic molecules; in contrast, most Si PVs are fabricated from zone- refined single crystals. Perhaps not surprisingly, the best efficiency reported for organic-based cells is approximately 6 percent. Improvements to this figure of merit will come from enhancing both charge injection and extraction and quantum photoconversion efficiency. The performance of organic-based PVs is at present dominated by the existence of large densities of charge-trapping impurities that impede free charge transport as exciton transport. Research is required to improve the understanding of these fundamental charge and energy transport mechanisms in bulk materials and at interfaces and to assess achievable gains through the devel- opment of improved materials. One approach is to develop a level of ­ materials control comparable to that of Si and other mature semi­conductor technologies. There is no known intrinsic reason that the efficiencies of either inorganic or

Science and Technology of Crystalline Systems 59 organic PV devices cannot be substantially improved to produce electricity at a competitive price. Indoor lighting accounts for 20 percent of total electricity use in the United States, and light-emitting diodes (LEDs) have the potential to reduce electricity use for indoor lighting by 50 percent, making this area a very important focus for energy reduction. Compound semiconductors that emit light are generally single- crystal materials grown as thin films at very high purity using layer-by-layer growth processes such as MBE or metal-organic chemical vapor deposition. For reasons discussed above for PV applications, organic light-emitting diodes (OLEDs) are viewed as a potential improvement in solid-state lighting technology. However, despite progress in recent years, the purity of organic materials relevant to solid- state lighting remains more than three orders of magnitude below that for com- mon inorganic semiconductors. Again, the goal is to develop a level of materials control comparable to that of mature semiconductor technologies such that the efficiencies of organic materials can be sufficiently improved to meet this need. Reviews of the science and technology of solar energy and solid-state lighting are available from DOE. Superconductivity for Electricity Delivery Electrification was identified as “the greatest engineering achievement of the 20th century” by the National Academy of Engineering. As the U.S. use of electrical energy increases, more efficient means for transporting electricity over the nation’s power grid are required (see Figure 2.13). A report by DOE’s Office of Basic Energy Sciences entitled Basic Research Needs in Superconductivity to Secure our Energy Future specifies the importance of discovery and growth of crystalline materials in the form of superconducting materials in impacting the grid. Through this grid, electricity provides clean energy from a wide variety of sources to a wide variety of end users. However, in the 21st century, the growing demand in the United States for energy is already challenging the existing grid (e.g., when brownouts occur). At present, the annual world electrical energy consumption of approximately 2 terawatts (TW) is predicted to double and triple by 2050 and 2100, respectively, surpassing present grid capability. Today, over 60 percent of transported energy is lost in production and delivery, with 8 to10 percent being lost in the grid. Transporting energy across time zones may also become more important   Seehttp://www.science.doe.gov/bes/reports/files/SEU_rpt.pdf and http://www1.eere.energy.gov/ buildings/ssl/. Last accessed March 3, 2009.    W.A. Wulf, “Great Achievements and Grand Challenges,” The Bridge—Report of the National Acad- emy of Engineering, 30, No. 3/4 (Fall/Winter 2000). Available at http://www.nae.edu/nae/bridgecom. nsf/weblinks/NAEW-4STLP8?OpenDocument. Last accessed March 3, 2009.

60 Frontiers in C rys ta l l i n e M at t e r FIGURE 2.13  North American electric grid as viewed from space at night. SOURCE: Courtesy of �������������������� Goddard Space Flight Center. Available at http://svs.gsfc.nasa.gov/goto?2276� ����������������������������������� . with increased reliance on solar energy. Superconducting cables providing ­lossless energy transport offer the ultimate solution. Superconducting wires can carry currents up to five times the currents that copper wires with the same cross sec- tion can carry. Replacing sections of the grid with superconductors, particularly in the regions where the voltage is stepped down and resistive loss becomes more significant, would have a dramatic effect on grid efficiency. All superconductors need to be cooled to become superconducting, and the high-temperature superconductors, discovered in 1986, become superconducting at temperatures that are significantly easier to reach than those of conventional superconductors (Figure 2.14). They do not require expensive cryoliquids, such as liquid helium, and can easily be cooled with liquid air or mechanical refrig- eration. At present, the best high-temperature superconducting cables—called 2G, for ­second-generation wires—contain the high-temperature superconductor YBa2Cu3O7. DOE programs have successfully used these 2G wires in two pilot

Science and Technology of Crystalline Systems 61 Temperature (kelvin) FIGURE 2.14  Time evolution of the superconducting transition temperature, Tc, of various super- conducting materials. Note that certain classes exhibit dramatic increases and, as discussed in the figure 2-14.eps text, these increases incorporated some degree of predictive discovery. SOURCE: Report of the Basic ���������������������������� �������������������� bitmap Energy Sciences Workshop on Superconductivity, May 8-11, 2006, Basic Research Needs for Super- conductivity, Figure 24, p. 86, available at http://www.science.doe.gov/bes/reports/SC_rpt.pdf. programs: power substations, with one near Columbus, Ohio, in operation since 2006, and transmission lines in New York State. The 2G transmission lines have limited benefit to the grid, however, as the materials are extremely anisotropic and brittle and have limited reliability. Most important, they are expensive to produce, as they rely on a silver base and com- plex multilayer technology. The current capacity, Jc, and homogeneity of the 2G wires can be improved by engineering defects in the superconducting wires. These defects are needed because superconductors can experience loss due to motion of vortices, and defects can “pin” the vortices. Research is needed to optimize these defects and reduce loss. Third-generation, 3G, wires, which are more reliable, isotropic, and have tran- sition temperatures and Jc at least as high as those of the present high-­temperature superconductors, are needed in order to have an impact on the grid. Over the past 30 years, research on strongly correlated electron materials, such as super­conducting

62 Frontiers in C rys ta l l i n e M at t e r materials, has reached the point that scientists need not rely only on serendipitous discovery to identify the next material. Scientists have learned that emergent phases arise from phase boundaries between competing phases (such as between ­metallic and insulating phases). The universal phase diagram in Figure 2.15 is general to strongly correlated electron materials. That understanding has guided research from serendipitous discovery toward predictive discovery of materials. Figure 2.14 shows the dramatic increase achieved in the ­superconducting transition tempera- ture, Tc, over the years. One of the more interesting systems is the “heavy-fermion 1-1-5” metals (CeIrIn5, PuCoGa5) in which Tc ranges from 0.4 K to almost 20 K. This extraordinary span of Tc in materials of the same crystal structure is thought to arise from an unexpected interplay between magnetism and superconductivity. Such unexpected trends were discussed in a recent article that noted a style of new materials development called “luck by design.” While it is not quite at the predica- tive discovery stage, scientists are approaching it with what they have learned from previous successes in materials design. As stated, “The more groups that search for interesting and potentially useful materials, the more diverse and viable the ‘idea gene pool’ becomes.” Since a superconducting cable must be cooled below its transition tempera- ture, there are dual challenges in applications: the efficiency of energy flow can be improved either by making the present materials more efficient or by discovering a new material that has a higher Tc, requiring less refrigeration. For both approaches, there are many materials hurdles, both basic and applied, to overcome, and almost all of them require single crystals. At present there is not a homogeneous high-temperature superconductor with a high Tc and a high Jc that can be economically produced. However, over the past few years, it has become clear that significant progress can be achieved by pursuing research in three directions. • The first of these is basic research aimed at the discovery of new materials. As stated above, this direction not only will lead to improved superconductors, but also to improved understanding of the mechanism of high-temperature superconductivity, which would in turn lead to improved materials in an exciting and fruitful cycle. Both the understanding of ultimate performance and the search for alternate materials require single crystals. For example, the highest-Tc materials contain light ions whose location and motion are difficult to ascertain using x-ray diffraction and require neutron scattering. Because neutrons interact weakly with matter, large crystals are necessary for accurate information on the structure and role of lattice vibrations and spin fluctuations in such systems.   P.C. Canfield, “Fishing the Fermi Sea,” Nature Physics, 4, 167-169 (2008).   Ibid.

Science and Technology of Crystalline Systems 63 FIGURE 2.15  Canonical phase diagram to guide scientists in the pursuit of new superconducting figure 2-15.eps materials. Note that new phases emerge from competing phases, and in this particular case, it is clear bitmap how high-temperature superconductivity emerges from competing phases of antiferromagnetism and “strange metal” (labeled NFL for “non Fermi liquid”). Such phase diagrams guide the researcher past serendipitous discovery toward predictive discovery and then toward “materials by design.” NOTE: AFI, antiferromagnetic insulator; SG, spin-glass; CO, charge-ordered state; d-SC, d x2y2 symmetry; fl-SC, fluctuating superconductivity; M, metal; PG, pseudo-gap; TN, Neel temperature; T*, pseudo-gap temperature divider. SOURCE: ������������������������������������������������������������������� ��������������������������������������������������������������������������� Report of the Basic Energy Sciences Workshop on Superconductivity, May 8-11, 2006, Basic Research Needs for Superconductivity, Figure 21, p. 72, available at http://www. science.doe.gov/bes/reports/SC_rpt.pdf. • The second research direction is to improve the current-carrying capability of the superconductors, typically with engineered defects, as stated above. The 2G wires are composed of anisotropic materials, and if the 3G wires remain anisotropic, at an applications level a power grid will require such a low level of grain boundaries and misorientation between grains that power lines will essentially contain single crystals by the mile. At an applied science level, since the optimum superconducting materials (layered cuprates) are highly anisotropic from a structural perspective, it is essential to understand the details of their crystal growth in order to optimize the manufacturing processes. • The third research area involves the development of reliable and eco- nomical cables. Only a few laboratories in the United States address the second research direction (e.g., DOE’s Argonne National Laboratory and the National Science Foundation’s National High Magnetic Field Labora- tory at Florida State University). Efforts to optimize the cables themselves,

64 Frontiers in C rys ta l l i n e M at t e r including their current-carrying capability, are mostly confined to a few wire-manufacturing companies such as American Superconductor Cor- poration and Super Power, Inc. A review of the science and technology of superconductivity for energy delivery was recently prepared by DOE. Catalysts for Fuel and Hydrogen Storage The process needed to convert raw feedstocks, such as oil and coal, into specific fuels, such as gasoline or diesel fuel, relies on catalysts. An important and contem- porary research area is that of developing improved catalysts for hydrocarbon con- version reactions. By controlling the shape and size of pores in zeolite minerals, as well as other mesoporous materials, specific hydrocarbon interconversion ­reactions can be selectively favored with high efficiency. In addition, many other organic transformations occur on or are catalyzed by surfaces or edges of single crystals. The ability to grow large crystals, particularly with specific faces, is expected to lead to rapid progress in this area. An additional application of crystalline materials is found in the separation and storage of fuel gases. Developing the capability to safely store, transport, and use hydrogen is critical for the success of a hydrogen-based economy. Porosity occurs for many crystalline materials because of the size and density of voids in the crystal structure, and many crystalline mesoporous materials can be used to store hydrogen and other hydrocarbon gases and to separate gases. Ongoing and future work in this area involves the development of “designer crystals” with porosity matched to a specific fuel gas. Needed Crystal Growth Capability for Energy Conversion and Storage Energy conversion and storage are representative areas in which both the development of new crystalline materials and the optimization of existing mate­ rials are needed in order to advance important technologies. U.S. research efforts have identified new opportunities, but in general these efforts are not sufficient in size or scope to change the technological landscape dramatically in the near term. Some areas of product development, such as large-area Si devices for solar energy, are progressing through private investment. In general, however, all progress in areas of technology development can be accelerated by focused efforts in the ­critical paths outlined below.   Report of the Basic Energy Sciences Workshop on Superconductivity, May 8-11, 2006, Basic Research Needs for Superconductivity, available at http://www.science.doe.gov/bes/reports/list.html. Last accessed March 3, 2009.

Science and Technology of Crystalline Systems 65 Capabilities for Solar Energy and Lighting Both inorganic and organic solar cells are being pursued as potential commer- cial systems for solar energy production (see Figure 2.16). As discussed above, the highest efficiency is obtained with heterogenous-layered semiconductor materials in crystalline form. Thus, PV fabrication research should focus on the specific crystal growth problems of the following: (1) large area and thin layers of highly absorb- ing semiconductors; (2) high-mobility layered heterogeneous systems; (3) layered systems that use inexpensive materials to allow large-scale production—Si being one example, but organic materials offer an alternative route; and (4) ­materials that show potential for integration into heterostructures. Research projects to address these issues simultaneously, moving toward the goal of ubiquitous solar energy, are large scale and should mimic industrial processes. Such crystal growth research is also relevant to solid-state lighting applications. Some of the requirements mentioned above for a path to large-scale production apply to organic-based semiconductors. The putative low fabrication costs make organic-based solar cells appealing, but efficiencies need to be enhanced from the present approximately 6 percent. Enhancement of the charge injection and extrac- tion, as well as photoconversion efficiency, is essential. The performance of organic- based PVs is at present dominated by defects that trap charge and control exciton recombination rates; it is essential to develop higher-purity materials. At present, commercially available organic compounds contain several percent of impurity molecules that are incorporated into the crystal structure upon crystallization. It is expected that the feed material for crystallization will become more pure as production volume increases; partnering with chemical suppliers will be critical for FIGURE 2.16  Examples of organic single crystals with possible use in solar cells. (Left) Anthracene seen in blue backlight, (middle) α-hexathiophene, and (right) fluorinated copper phthalocyanine show- ing growth steps. SOURCE: Courtesy of Christian Kloc, Nanyang Technological University.

66 Frontiers in C rys ta l l i n e M at t e r this effort. One approach to understanding the role of impurities in organic PVs is to conduct research on single crystals such as those shown in Figure 2.16. These crystals are grown from vapor transport of sublimed molecules, and crystalline layers greater than 1 cm2 are currently obtainable for several compounds. Capabilities for Catalysts for Fuel and Hydrogen Storage The discovery and development of new materials capable of efficiently and selectively converting raw oil, coal, and other feedstocks into fuels or petrochemicals, as well as the storage and separation of gases such as hydrogen, require the discovery and development of improved catalysts and porous materials. New and improved “designer crystals” are sought. Advances in computer modeling to identify the shape and size of the void spaces within the crystalline lattice are essential, as are ideas and paradigms for new families of materials. Grand Challenge 3: Evolution in the Capacity to create Crystalline Materials by Design This section presents Grand Challenge 3: that of developing the capability of designing new materials from first principles to meet specific technological needs. Long a dream of scientists and engineers, such “materials-by-design” approaches are becoming increasingly possible through rapid advances in theory and model- ing, coupled with continuous increases in computational power. More than ever before, future devices will possess specific functionalities resulting from the design of the materials incorporated into the device. Realizing such an approach will free industry from current materials and device limitations, many of which are based on research carried out in academic and industrial research laboratories a generation ago. By combining materials by design with device science and engineering, new classes of devices will be created to provide lower-cost solar power, novel informa- tion and communications devices, and sensitive and multifunctional sensors. The section begins with a discussion of representative theoretical and computa- tional approaches to the design of materials, including some of the computational hurdles that must be overcome to fully implement those approaches. Examples of how the materials-by-design approach has already led to significant developments in the areas of structural alloys and sensors are then presented. Materials-by- design approaches that enable materials hitherto thought impossible to prepare are described next. These include the multiferroic materials previously discussed, as well as thermoelectric materials that are good electronic conductors but poor thermal conductors. The section ends with a discussion of selected specific tech- nological challenges that must be addressed for further progress in the discipline of materials by design to be achieved.

Science and Technology of Crystalline Systems 67 Theoretical and Computational Approaches to Materials by Design One of the most dramatic scientific achievements of the 20th century was the identification and development of the quantum-mechanical laws of matter. The quantum theory of solids that is built on those laws underpins our understanding of the physical nature of materials. It has proven remarkably successful in describ- ing materials with electrons that can be represented as distinct particles engaging in generally weak interactions. However, such materials are few, and the limitations of this approach have been made obvious by its failure over the past two decades to lead to an understanding of many new classes of materials that display correlated electron behavior. Recently, though, with increases in computational speed and the development of new algorithms, some of these limitations are being overcome. The ability to simulate phenomena has now advanced to the point at which scientists are confident that, in the near future, simulation and modeling can be used routinely for materials discovery and development. The materials themselves, and the insight gained from them, will almost certainly contribute to next-generation technologies, including nanoelectronics, functional materials, and quantum information. The flow diagram in Figure 2.17 illustrates how a typical computational design process takes place. The essential steps involve transcribing current theoretical understanding into a proposed trial compound. The properties of the compound can then be computed and compared to an experimental synthesis. The experi­ mental and theoretical properties will agree in a successful design by computation. It may also be possible to use stochastic algorithms to predict new trial compounds. For example, evolutionary or genetic algorithms have recently been used to predict alloy properties and new crystal structures. The promise of materials by design clearly exists. However, to fully realize the potential offered through materials by design, further advances must be made in a number of computational areas. Two such areas—the development of more-refined methods for predicting crystal structure and the simulation of crystal growth—are addressed here. Methods for Predicting Crystal Structures One of the most significant advances in the theory of materials occurred in the mid-1980s. At that time it was demonstrated that two ingredients could be combined to calculate the structural properties of crystals, with the exception of highly corre- lated systems. The first ingredient is the pseudopotential. This approximation replaced the electronic potential of all the electrons with a potential that replicated only the properties of the chemically active electrons, that is, those occupying valence states. The potential arising from the electrons in core states and the positively charged nucleus were combined to form a chemically inert “ion core” pseudopotential. This

68 Frontiers in C rys ta l l i n e M at t e r functionalities identified understand chemistry driving functionalities propose trial compounds calculate structures synthesize compounds compute properties measure properties computed = NO measured? NO develop improved YES develop improved theoretical methods synthesis methods measured = desired? NO YES rational design of new functional material FIGURE 2.17  A schematic illustration of the methodology by which theoretical approaches can be figure 2-17.eps used to carry out a rational design of new materials. pseudopotential can be determined from an atomic structure calculation and does not change from one environment to another. The valence states now fix the energy and length scales of the electronic structure because the pseudopotential describes only these states. As a result, very simple basis functions, such as plane waves and Gaussian shapes, or simple grids can be used to describe the electronic states. This enormously simplifies the problem, especially for crystalline matter. The second ingredient for calculating the structural properties of crystals is density functional theory implemented with the local density approximation or the generalized gradient approximation. Density functional theory allows the “many- electron” problem to be mapped onto a “one-electron” problem, resulting in the Kohn-Sham equation. This ingredient, coupled with pseudopotentials, enormously simplifies the problem and allows systems with hundreds, if not thousands, of atoms to be examined.

Science and Technology of Crystalline Systems 69 Using this approach, accurate structural energies can be calculated and used to make accurate predictions about crystalline materials and about more complex systems as well. Workers have demonstrated that in some cases—for example, crystals under high pressure—these methods often can be used to predict certain properties of crystals better than those properties can be measured. In principle, this situation opens up the possibility for one to consider predict- ing accurately the properties of hypothetical materials on the computer. However, this endeavor is in its infancy, and notable issues must be addressed. Electronic structure methods usually are employed to compute thermodynamic properties as opposed to kinetic behavior. Thus, it might be possible to predict that a given crystal has a given property, but it might be a very different issue to know whether the structure can be synthesized. A standard example is graphite versus diamond. Electronic structure methods can predict that diamond is a stable form of carbon at high pressures, but these methods do not indicate the kinetics of the synthesis process such as the reaction rate or the optimal growth conditions. Simulating Crystal Growth The growth of bulk single crystals remains one of the most challenging and astonishing technical endeavors of materials processing. For example, electronic- grade Si grown by the Czochralski method is one of the purest and most per- fect materials ever humanly produced. Current production technology routinely achieves impurity levels of less than parts per billion in single-crystal ingots of up to 400 mm in diameter and over 250 kg in mass, and these crystals are completely free of dislocations. Indeed, it is even possible today to control the distribution of microdefects, such as interstitials, vacancies, and voids, in bulk Si crystals. Modeling has played an important role in clarifying the physical mechanisms governing crystal growth and the engineering principles that must be employed for successful crystal growth processes. In the classical processing-­structure-­properties triad of materials science, one goal of ab initio modeling is to define the structure- properties relationship for a crystalline material, while the goal of crystal growth process modeling is to clarify the processing-structure connection. In the engineer- ing of bulk-crystal growth processes, the greatest purpose that model­ing can serve is to directly connect processing conditions to the structure, and therefore the properties, of the grown crystal. In engineering measures, modeling is employed to enable the mantra of “faster, better, cheaper”—that is, the understanding enabled by modeling can increase the process yield, improve the product quality, and decrease the process cost. Modeling of crystal growth must account for a host of physical mechanisms, representing both thermodynamic and kinetic phenomena and ranging over dispa- rate length scales and timescales. The challenge is how to represent realistically the

70 Frontiers in C rys ta l l i n e M at t e r most pertinent of these phenomena, and modeling approaches have varied from the continuum to the atomistic. The extent to which modeling can realistically represent bulk-crystal growth is limited, owing mainly to these challenges and the limits of current computers and numerical algorithms. Nevertheless, advances in theory and modeling have led to and will continue to lead to important advances in crystal growth. Bulk-crystal growth encompasses a wide variety of physical phenomena that occur over a vast range of length scales, making it among the most challenging of industrial processes to model. On a macroscopic scale, ranging from millimeters to meters, the transport of heat, mass, and momentum is always important. At a mesoscopic scale, ranging from tens of nanometers to tens of microns, the crystal interface can exhibit hillocks, steps, and other structures, even though it may appear macroscopically smooth. Mesoscale structure also occurs within the crystal in the form of grain boundaries or extended defects, and even within the melt, where structured complexes of molecular species sometimes appear. At a microscopic scale of nanometers or less, the fundamental mechanisms by which atoms or mol- ecules are incorporated into the growing crystal and how they interact within the solid at high temperatures ultimately determine its final structure. Bulk-crystal growth also encompasses a vast range of timescales. The longest of these is the time of growth, which can range from hours to months, depending on the type of crystal and the system used to grow it. Heat, mass, and momentum transport each have a characteristic timescale, depending on the dominant mecha- nism of transport and the physical properties of the system. If diffusion dominates transport, these timescales typically range from minutes to hours. Within the solution or melt, however, convective transport usually dominates, and the trans- port timescale is typically reduced to seconds or less. Phenomena occurring at mesoscopic length scales, such as morphological instabilities, typically evolve over timescales that are comparable to transport timescales. Atomistic events important in bulk growth range from the rate of incorporation of atoms into a solidification of interface, occurring over timescales on the order of milliseconds or less, to much longer times associated with concerted action, such as nucleation events. A model that includes all these physical phenomena, spanning at least nine orders of magnitude of length scales and timescales, is far beyond the capability of today’s computers. Therefore, a variety of modeling approaches have been employed to clarify phenomena on different scales. At a microscopic scale of tens of nanometers or less, ab initio molecular dynamics methods can be employed to study atomic behavior. Unfortunately, while these methods are quite rigorous in their approach, even using today’s fastest computers they are too computationally expensive to be applied to systems of more than a few thousand atoms or to be used to describe timescales of greater than hundreds of picoseconds. Molecular dynamics methods

Science and Technology of Crystalline Systems 71 based on classical potentials can compute for much larger ensembles and longer timescales. To model atomistic behavior at even larger length scales and longer timescales, kinetic Monte Carlo methods have been developed. On a macroscopic scale, continuum methods are gainfully applied. Significantly, there are many important phenomena associated with crystal growth that occur on a “mesoscale,” comprising hundreds of nanometers to tens of microns and occurring over long timescales (e.g., microseconds or longer). These phenomena are difficult to model by atomistic methods owing to the long timescales involved and are challenging for continuum methods owing to the very small length scales. Such phenomena require innovative, “multiscale” approaches, which, despite much fanfare, are still at nascent stages of development. This challenge of scales motivates the need to formulate crystal growth ­models that include enough physics to make realistic, usable predictions that are simple enough to remain tractable with today’s computational capabilities. Future chal- lenges for modeling must lead to more realistic representation of the multiscale interactions important in crystal growth systems. Models must be capable of describing detailed system geometry and design (e.g., furnace heat transfer for melt growth systems), three-dimensional and transient continuum transport (flows, heat and mass transfer), phase-change phenomena (thermodynamics and kinetics), and atomistic events. Progress is being made on all of these fronts, but many challenges remain. The understanding gained from more realistic, multi- scale models for ­crystal growth will ultimately lead to the ability to link crystal structure and ­properties with growth conditions and the macroscopic factors that influence them. Areas of Success in Creating Materials by Design Having discussed some of the theoretical and computational challenges that must be addressed in achieving future success in creating materials by design, the discussion now turns to some of the areas where success has already been reached. Thin Films The computational design of materials with novel functionalities, or combina- tions of functionalities, is becoming particularly valuable in the arena of thin-film heterostructures. In particular, the ability to combine different parent phases, to create new metastable phases, and to modify properties using epitaxial strain offers additional degrees of freedom in the design process, as described in Figure 2.17. Accurate calculations of interfacial behavior are computationally demanding. While macroscopic properties based on bulk behaviors can often be described accurately within semiclassical theories, the atomic-level response of an interface

72 Frontiers in C rys ta l l i n e M at t e r is intimately related to the local details of chemistry, bonding, structure, and elec- tronic properties, which often strongly depart from those of the constituent ­parent compounds. Therefore, detailed electronic structure methods are required for accurately reproducing the subtle balance of these competing factors. This issue is extraordinarily difficult to treat theoretically, given the absence of any periodicity in lattice mismatches. A particularly important methodological development for modeling device behavior has been the rigorous framework for applying finite electric fields within density functional theory, and its extension to metal-insulator heterostructures and interfaces. These developments have enabled detailed theoretical studies of i ­ nterfacial dielectric behavior, allowing the experimentally observed and techno- logically highly relevant suppression of the dielectric response of nanoscale capaci- tors to be understood and mitigated. Defects in Materials While the creation of defect-free, well-characterized crystalline matter is cru- cial for scientific study, most technologically relevant materials possess controlled densities of defects. Thus a quantitative understanding of defects in crystalline materials is required for the prediction of properties. A classic example is the con- trolled introduction of impurities in electronic materials, for example, the doping of semiconductors. Doping can dramatically alter the electronic properties of semiconductors; the introduction of 1 boron atom per 1,000 silicon atoms increases the conductivity of pure Si by three orders of magnitude at room temperature. Such profound changes are responsible for the functionality of almost all elec- tronic devices, ranging from LEDs to integrated circuits. For electronic materials, the theoretical tools outlined above are capable of helping design materials with specific defect-induced properties. For other systems, the issues of theoretical design are more problematic. Con- sider the example of the manganites. The parent phase of LaMnO3, for example, is an uninteresting, antiferromagnetically ordered insulator. However, with the introduction of holes into this system through cationic substitution (for example, Ca, Sr, or Ba doping at the La site), the phase complexity of this system dramatically increases and the properties vary greatly: For example, the colossal magnetoresistive effect replaces the magnetic insulator. Furthermore, recent work has established the existence of an electronically inhomogeneous structure (the so-called electronically phase-separated structure) in which regions that are rich in electrons (conducting and ferromagnetic) are interspersed with regions that are deficient in electrons (insulating and antiferromagnetic), although the whole material appears to be chemically homogeneous. Another example is the case of relaxor ferroelectrics, which are described as an ensemble of nanoscale polar regions that are highly sensi-

Science and Technology of Crystalline Systems 73 tive to their local chemical and electrical environment, thus leading to large dielectric and piezoelectric susceptibilities. The length scale of such processes in many of these materials is still under debate but is likely on the order of a few nanometers. These examples set the stage for a rather interesting question: Is it possible to design the architecture of inhomogeneous systems such that the optimum in functional responses is obtained? For example, can one design the size of the polaron in such materials (one that is suspected to be the cause for the onset of f ­ erromagnetism) through an atomic-scale “defective materials by design” algo- rithm? The types of theoretical tools that will be required to accomplish such tasks for these highly correlated systems are more complex than for the systems discussed above and are not as well understood. Materials with High Strength and Toughness: The Next Generation of Steels An area in which the materials-by-design philosophy has already paid hand- some dividends is the field of structural alloys, particularly alloy steels. Several decades ago, DOE made a strategic commitment to the development of next- generation alloy steels having the desirable combination of ultrahigh strength and high fracture toughness, along with high corrosion resistance in chemically and thermally hostile environments. This was accomplished through a materials- by-design approach, involving careful studies of the atomic-scale microstructure of steels prepared under ultrapure conditions. At the next higher length scale, a considerable amount of effort went into “designing,” through a combination of thermomechanical treatments, a microstructure that consisted of a fine distribu- tion of lath martensite (in contrast to twinned martensite, which directly leads to brittle fracture behavior) at the boundaries of which is dispersed a thin layer (a few tens of nanometers) of ductile austenite (the original parent phase from which the martensite forms). Such a nanocomposite microstructure displays the desirable combination of tensile strength and toughness. New Materials and Crystals for Sensors and Detectors Another area in which the materials-by-design concept has been making inroads is that of sensors and detectors. Sensor networks are becoming common, and sensors with new capabilities are widely sought. New materials and function- alities are essential components of sensing and its converse functionality, actua- tion. In a broad sense, sensors can be categorized in four general classes, based on their primary mode of functionality: (1) chemical sensors, (2) field sensors (such as pressure, electric, magnetic), (3) radiation sensors, and (4) biological sensors. Markets for each of these classes are rapidly expanding, and new materials and functionalities are a major driver of this expansion. Many of the devices are built

74 Frontiers in C rys ta l l i n e M at t e r using thin-film-based approaches. The focus here is on radiation sensors, since the current worldwide political climate has necessitated the implementation of a wide range of radiation sensors for threat mitigation. Single crystals are likely to be of great value in this area. Radiation Sensors Illicit trade in nuclear materials on world markets poses a long-term inter- national security threat and has created an urgent need for instrumentation that can rapidly, reliably, and inexpensively detect the x-ray and gamma-ray spectra of such materials. The core of such instrumentation is the radiation detector—a device that produces distinctive signals in the presence of radiation from nuclear mate- rial. Semiconductor materials have properties that make them exceptionally well suited for this task: in a single conversion step, semiconductor radiation detectors generate electrical pulses that are directly proportional to the energy of the x-rays and gamma rays of interest. This excellent linearity is one of the great assets of semiconductor detectors used in spectroscopy. The fundamental physics of the interaction of energetic electromagnetic radia- tion (10 keV to 10 MeV) with matter provides clear guidance in the choice of materials parameters for optimized detectors. The linear absorption cross section for the photoelectric effect (the conversion of the energy of electromagnetic radia- tion into kinetic energy of an electron) is proportional to Zn, where Z = atomic number and n = 4-5. Thus, the higher the Z of a detector material, the greater the stopping power for radiation and the more efficient a detector is for a given volume of material. The thinner detector layers needed with very high Z materials result in less stringent requirements for charge transport and collection. For gamma-ray energies between 100 keV and 1 MeV, the photoelectric absorption cross section of lead is more than 50 times larger than that of germanium (Ge), which, in turn, is 40 times larger than that of Si. A second important requirement for inexpensive radiation detectors relates to the operating temperature. Room temperature is clearly desirable. This in turn requires a band-gap energy of 1.6 eV or higher because the thermal generation of charge carriers must be small compared to the charge carriers produced by the radiation to be measured. At the same time, too large a band gap is detrimental because the energy required to form an electron-hole pair is proportional to the band gap energy. This means that large-band-gap semiconductors produce small amounts of charge for a given radiation energy, a disadvantage for high energy resolution measurements. The optimal material requires a trade-off of competing factors, so band-gap tunability is a desired feature in future materials systems. Finally, a third key parameter is related to the motion of radiation-induced charges through the semiconductor material. The typical planar radiation detec-

Science and Technology of Crystalline Systems 75 tor has two contacts on opposite faces of a cylindrical slice of material. In order to obtain identical signals, independent of the position of the incident radiation, the photon-generated holes and electrons need to travel all the way to their respec- tive contacts. The larger the carrier mobility and carrier lifetime, the easier it is to fulfill this requirement. The mobility-lifetime product is perhaps the most critical quality measure for a semiconductor material used in radiation detectors. Si and Ge have products for hole and electrons exceeding 1 cm2/V at 300 K and at 77 K, respectively, a value that allows the fabrication of large volume radiation detectors with excellent properties. All other radiation detector materials have significantly smaller products, especially for holes. This is why none of these existing materials can compete with Si or Ge detectors when energy resolution is most important. The need for room-temperature operation precludes the use of Ge detectors (the 0.7 eV band gap is too small) and has led to a search for high-Z, wide-band-gap (WBG) materials. The most successful candidate to date has been ZnxCd1-xTe, with x in the few percentage range. Recent results with 1 cm3-size crystals and a coplanar interdigitated contact structure have reached an energy resolution of approximately 1 percent for 662 keV 137Cs radiation. These results were achieved using only the collection of radiation-generated electrons, since the mobility-lifetime product of holes is too small to produce a signal. The efforts spent discovering and improving CdTe-based detectors span the past 30 years and act as a reminder of the difficulties that lie ahead for identifying and developing new materials. A number of other materials have been investigated for suitability for room-temperature radiation detection. These include HgI2, AlSb, and other mate­rials. None of these materials has yet produced results comparable to those achieved with ZnCdTe. Requirements created by nuclear threats for very large numbers of radiation detectors with 1 per- cent energy resolution at 1 MeV operating at or above room temperature demand new approaches. Potentially Successful Approach A brief review of the development of ultrapure Ge for gamma-ray ­spectroscopy may offer some insight into a potentially successful approach. The discovery in the early 1960s of the lithium (Li) drifting process for the formation of wide-­depletion- region p-i-n SiLi and GeLi detectors led to a new class of radiation detectors, vastly superior to the commonly used NaI(Tl) scintillation detectors. That GeLi detectors required cooling to achieve sufficiently small diode leakage currents was accept- able in view of the vast improvement in resolution. A serious problem was the u ­ nacceptable deterioration of device quality, related to only brief uncooled periods, which resulted in a loss of perfect Li compensation. Around 1970, Hall at the GE Research Laboratories suggested that it should be possible to develop ultrapure Ge for stable room-temperature radiation detectors. Goulding and coworkers at the

76 Frontiers in C rys ta l l i n e M at t e r Lawrence Berkeley National Laboratory initiated a program for the development of the ultrapure Ge single crystals and detectors. The Czochralski melt-growth tech- nique was selected, but many of the traditional crystal growth approaches had to be sacrificed for purity. New crucible materials were developed and a gas “blanket” of 1 atmosphere of flowing hydrogen was used to prevent residual impurities from reaching the melt. Dislocation-free crystals, naively expected to be the most perfect, showed poor charge trapping when made into detectors, and it was discovered that dislocations could play a key role in removing native defects. A large number of new defect centers were discovered and identified. In parallel to the crystal growth effort, radiation detector development and fabrication of gamma-ray spectrometer systems, including the detector, the cryostat, and all the associated electronics, were actively pursued. The entire operation was colocated in adjacent laboratories. New crystals were characterized with Hall-effect measurements and photothermal ionization spectroscopy within hours of crystal growth completion, and working detectors were fabricated within a few days. This extremely tight coupling between materials purification, crystal growth, and detector fabrication formed the recipe for success. Decoupling Electron and Phonon Transport: The Search for High-Efficiency Thermoelectrics Since the discovery of electricity, research on charge transport in materials has pushed the extremes of electrical conductivity, which now spans more than 20 orders of magnitude. Five decades ago, the invention of the transistor dem- onstrated external tuning of electrical conductivity, and this method of tuning formed the foundation for information processing that created the digital age. The societal impact of research in electrical conduction is clearly enormous. In contrast, heat transport in materials has received much less attention. The thermal conductivity of current solid materials spans only four orders of magnitude at room temperature, and its external tunability is limited. Yet the history of thermal transport goes back to primitive humans who combined empirical observations with applications (specific heat of stone for warmth, straw for insulation, wooden tools for manipulating fire, and so on). In modern times, research in thermal transport has enormous implications for the ability to efficiently convert, use, and store energy. Thermal barrier coatings in jet engines allow higher fuel efficiency, while thermal insulation in refrigeration systems has led to energy conservation. At present, the most serious issue in the miniaturization of electronic circuits is thermal management. The power density of microprocessors is expected to exceed that of a nuclear reactor before 2010 and to reach that of the surface of the Sun by 2015. In addition, power dissipation in organic and molecular electronics could be a key limitation. It is believed by many that if thermal transport could be

Science and Technology of Crystalline Systems 77 externally tuned and its limits extended, the impact would be felt in areas whose diversity ranges from energy conversion (e.g., thermoelectricity) and storage to next-­generation information processors (inorganic microprocessors, organic electronics, and qubit systems). In contrast to magnetoelectrics, in which a large coupling between magnetic and electric order is desirable, in thermoelectrics one desires to decouple electron and phonon transport. The performance of a thermoelectric material is typically described by the figure of merit ZT, where Z = S2σ/κ, σ is the electrical conduc- tivity; S is the thermopower, or Seebeck, coefficient; and κ is the total thermal conductivity. S and σ are contraindicative, meaning that they change in opposite directions with the carrier concentration. Thus, much of the recent work has focused on nanostructuring as a means of reducing the phonon contribution to the thermal conductivity in order to increase ZT. However, to date the most attractive material (from a commercial perspective) has been Bi2Te3, which was discovered several decades ago. Despite a large amount of research devoted to surpassing Bi2Te3, there remains a pressing need for new materials with higher ZT values. There is the need to discover materials that can behave like a perfect thermal insulator (akin to an electrical insulator) as well as superthermal conductors (for heat dissipation applications). Materials-by-Design Challenges Notwithstanding the tremendous progress in creating novel materials (struc- tural, electronic, functional, biological, and so on) over the past century, there are still huge challenges ahead for truly “designing” materials. The challenges remain- ing are (1) morphological length scales, (2) chemical stoichiometry control, and (3) matching energy scales. Challenges at Morphological Length Scales Today, various processes (bulk-crystal growth, thin-film growth) have enabled new materials with unprecedented properties to be made. In some cases (for exam- ple, MBE of GaAs-AlGaAs, discussed in Chapter 1 in the section entitled “Example in the Area of Thin Films: Gallium Arsenide-Based Heterostructures”), the depo- sition processes have evolved to an extent that atomic-scale control of structure and composition has become commonplace. For more complicated materials, however, such as the complex, correlated oxides, the situation is far less mature. MBE-like techniques are just beginning to be developed for the preparation of such materials in the form of epitaxial heterostructures. These heterostructures contain the desired phases, layered in the desired sequence with reasonable film- thickness control; however, the atomic-scale control of the interfaces between these

78 Frontiers in C rys ta l l i n e M at t e r phases, and ­ stoichiometry control (described below) both within these phases and at their interfaces, present tremendous challenges to synthesis. Owing to their more complicated crystal structures and lower symmetries, far more options for different interfaces are possible among these complex systems than among III-V semiconductors. In the case of oxide growth in particular, tools to provide surface- specific chemical and structural information during heterostructure growth need to be developed. The problems are equally challenging when considering crystals. In many cases, the complex phase equilibria in these systems (for example, lack of a congruent melting composition) mean that high-quality crystal growth processes, such as the Czochralski method, are not applicable. A good example is SrTiO3, which is an important substrate material for oxide heterostructure growth. The lack of high- quality substrates (in terms of both chemical stoichiometry and low defect con- centration) has become a critical limitation to progress in device development. The challenge in such cases is to create large-area (4- to 6-inch) wafers that approach the quality of state-of-the-art Si substrates. Challenges also arise at mesoscopic length scales. Over the past 20 years, a fun- damental insight that has emerged from studies of cuprate superconductors, relaxor ferroelectrics, and colossal magnetoresistive manganites is the importance of elec- tronic charge inhomogeneity, as distinct from structural homogeneity. In the case of the cuprates and manganites, the parent compounds are both anti­ferromagnetic insulators. Doping them with holes (typically through cationic doping) leads to the onset of superconductivity in the cuprates and large magnetoresistance in the manganites. In both cases, these phenomena are correlated with an inhomogeneous electronic distribution, typically over a length scale of a few nanometers to 100 nm. This is generically similar to relaxor ferroelectrics, in which the large dielectric and piezoelectric susceptibilities have been correlated with nanopolar domains. Thus, key challenges on this front are (1) to determine the characteristic length scale over which such inhomogeneities should exist for optimum responses from the material, (2) to develop experimental methods of characterizing and controlling electronic inhomogeneity, and (3) to understand theoretically the role of electronic i ­ nhomogeneity from a structure-property perspective. Chemical Stoichiometry Control Controlling composition (both cationic and anionic) in complex oxides con- tinues to be a key challenge. This challenge is illustrated through the example of oxygen content in perovskite oxides, aspects of which are common to strongly correlated oxide materials. It is widely accepted that oxygen vacancies form facilely in perovskite oxides. Furthermore, these vacancies are thought to be ionized, thus donating electrons to the conduction band. A 0.1 percent vacancy concentration,

Science and Technology of Crystalline Systems 79 readily achieved, would likely lead to an electron concentration on the order of 1019‑1020 e/cm3, a significant carrier concentration. A specific implication of the presence of oxygen vacancies is the evolution of polarization fatigue and imprint in ferroelectric capacitors. A large body of evidence now directly points to the role of such defects in the fatigue process. With this background, the challenge becomes one of probing these defects in crystals or thin films with sufficient chemical resolution (1 part per million) and spatial resolution (1 nm). Of particular importance is the behavior of oxygen, since it is highly mobile and yet highly reactive and critical for electrical and magnetic behavior. Thus the electronic manifestation of the defect states that accompany the non-stoichiometry in single-phase samples can have huge effects on the properties of thin films and heterostructures of complex materials and can mask their intrinsic properties. A specific example of this could be the magnetoelectric perovskite heterostructures exemplified by the model system of BiFeO3/BiCrO3. In this hetero­structure, it would be extremely desirable to be able to control the iron (Fe) or chromium (Cr) com- position across the heterostructure with atomic precision. This, however, can be a daunting task owing to the chemical similarities of Fe and Cr and the inter­diffusion that ensues. How does one create such precision-tailored heterostructures? How does one prevent interdiffusion? Applied Crystal Growth for Technology Development Grand Challenge 2, the Creation of New Crystalline Materials for Energy P ­ roduction and Conversion, addresses a specific technology area. Crystalline mate- rials are, however, crucially important in many other technologies. As was the case for basic research, the loss of industrial laboratories has had a large impact on crystal growth research focused toward technology development. In the past, such research was funded by corporate research and development (R&D) funds. Today research in applied crystal growth is modestly supported by programs such as the Small Business Innovation Research program through the Department of Energy, the National Science Foundation, and the Department of Defense. Such limited-term and application-focused funding, while useful for addressing some problems of importance, seems insufficient to support activities such as crystal growth, which by its nature relies on large infrastructure, a highly trained staff, and long-term institutional stability. It is also not clear that such a technology focus is well matched to the mission of any existing federally funded laboratory. Such a centralized laboratory that is industry oriented would have similarities to the centralized semiconductor R&D facility SEMATECH. While research of industrial importance has diminished in the United States, in other countries such research remains vital. Germany, for example, shows how such applied research can form a critical link in the science-to-product chain. A

80 Frontiers in C rys ta l l i n e M at t e r mixture of activities, from the growth of novel superconductors to the growth of large boules of Si for microelectronics, is supported in several German institutes, the largest of which is the Institute for Crystal Growth in Berlin (see also the section entitled “International Activities” in Chapter 3). These institutes are funded by a variety of sources, including the German federal government and private corpora- tions, and the institutes fulfill several needs not easily accomplished within either academia or modern corporations. First, they maintain a knowledge base for what might be termed large-scale or “industrial” growth techniques. These techniques are often applied to different materials classes (for example, floating-zone refine- ment is used both for semiconductors and for metals). It is critical for a country with a large industrial base in information, energy, and security technologies to maintain this expertise. Second, such expertise can be accessed by different com- panies, thereby increasing efficiency and speed to market. Finally, such activities provide an intellectual link between basic research and manufacturing. As basic researchers set strategic goals, their research should be informed by the range of possibilities and limitations in current manufacturing processes. The loss of such applied research in the United States is seen in many different areas. Crystalline Materials for Next-Generation Technologies Crystalline materials lie at the heart of many modern technologies. Past exam- ples are silicon for microelectronics, quartz in clocks and the Global Positioning System (GPS), gallium arsenide and indium phosphide in cellular phones, and ruby (aluminum oxide with chromium impurities) in lasers. In all of these examples, a nearly perfect crystal structure is needed for high performance. While research on existing technologies will need continuing improvement of known materials, developing technologies will require new materials as well as synthesis of known materials in high-purity single-crystal form. Examples are (1) gallium nitride (GaN) for energy-efficient lighting that some day could supplant incandescent lightbulbs, (2) new materials for use in detectors of weapons of mass destruction, (3) new semi- conductors for infrared optics and space communications, and (4) new approaches to producing large-area semiconductors for solar energy applications. Several areas of opportunity for research in crystalline materials for next-generation technologies are reviewed in the following pages. Next-Generation Crystalline Materials for Future Information Technology The earlier section entitled “Grand Challenge 1: The Development of Next- Generation Crystalline Materials—New States of Matter and New Materials—for Future Information and Communications Technologies” discussed a wealth of phenomena with potential for integration into future information systems. While

Science and Technology of Crystalline Systems 81 basic research on these materials provides technology directions, critical factors for eventual application will be determined farther down the development path where materials compatibility, materials cost and processing complexity, and com- patibility with adjacent technologies are determining factors for the adoption of a particular technology. Part of the challenge of developing a new technology lies in the research that attempts to create prototype devices based on a new effect. The more focused projects necessary for this stage of development require dedication of specialized apparatuses for growth and fabrication. Many of the research activities already mentioned fall into this category, including the production of high-purity small-molecule organic semiconductor feedstock, the development of techniques for creating large crystals of GaN and other next-generation semiconductors, and the control of defects in oxides. Of particular interest is intensive recent research on graphene—single atomic sheets of graphite—that led to the discovery of remark- able and interesting electronic properties (see Box 2.1). Experiments performed on sheets one atomic layer in thickness peeled from single crystals of graphite suggest that graphene holds the possibility of becoming the electronic material of the future. This vision of a future featuring graphene-based electronics now drives the science and technology on very interactive parallel paths; a similar situation occurred in the study of carbon nanostructures. Next-Generation Optical Devices for Security Crystals are already used in a wide variety of optical device applications, including laser sources, frequency mixing through nonlinear response, and optical modulation. Laser materials include fluorides such as rare-earth doped YLiF4 and CaF2 and rare-earth doped oxides such as the garnets, GdVO4­, and (Y,Gd)Ca4B3O10. Nonlinear crystals are typically oxides (for example, LiIO3 in the 800 nm range, and LiNbO3 in the 1,319 nm range). Optical modulation applications include amplitude and phase modulation using LiTaO3 and LiNbO3 crystals and Faraday rotation using the oxide garnet Tb3Ga5O12. Of great future interest is the use of nonlinear optical crystals for infrared- to-terahertz parametric sources. Here new materials are needed to provide higher efficiency and output power in the 2,000 to 8,000 nm (2- to 8-µm) wavelength range. Such materials require a highly nonlinear response with low absorption losses. Figure 2.18 shows some candidate materials envisioned for such purposes. Among these systems, crystalline ZnGeP2 shows great promise for its low loss and large nonlinear coefficient, as demonstrated in the terahertz image shown in Figure 2.19. New crystalline systems are also needed for mid-infrared communica- tions applications.

82 Frontiers in C rys ta l l i n e M at t e r FIGURE 2.18  Nonlinear figure of merit for several crystalline materials used in infrared frequency conversion. SOURCE: Courtesy of Peter Schunemann, BAE Systems. figure 2-18.eps bitmap FIGURE 2.19  Image of a razor blade inside an envelope obtained using a crystalline ZnGeP 2 terahertz source. SOURCE: Courtesy of Peter Schunemann, BAE Systems.

Science and Technology of Crystalline Systems 83 Crystalline Materials for Electromechanical Actuation A modern automobile contains more than 100 actuators. Future automotive design will use every opportunity to reduce weight and power consumption, and piezoelectric crystals and ceramics have been developed for actuating applications where speed and force are needed for a small displacement function. Such applica- tions include fuel injection, positioning, and fast switching. Actuators take a variety of forms, including stacked piezoelectric elements, tubes, and disk actuators. For wider applicability, materials requirements such as high operating temperatures, durability, and internal electrode composition must be satisfied. Substrates for Next-Generation Electronics The growth of high-quality crystalline thin films requires crystalline sub- strates with atomic lattice sizes that match the width of the thin film. Thin films of interest are typically oxides for nonlinear optics, solid-state lighting, or strongly correlated electronic states; or a non-oxide for emitter/detector or high-power radio-­frequency applications. As an example: for high-mobility ZnO/MgxZn1-xO films exhibiting the quantum Hall effect and p-ZnO/n-ZnO heterostructures exhibit­ing violet electroluminescence, ScAlMgO4 was used as the substrate. For AlGaN/GaN hetero­structures exhibiting mobility exceeding 1.6 × 105 cm2/Vsec, sapphire-Al2O3 substrates were used. While WBG compounds such as zinc oxide (ZnO) and GaN have been grown on oxide substrates, there is much interest in growing oxides on WBG substrates to enhance the ability to tune charge density in a WBG compound through the gating of an oxide-induced dielectric or ferro- electric state. In addition to their use in substrates, single crystals have also been used as high-purity sources in epitaxial growth. New Growth Techniques The development of new crystal growth techniques requires a focused research effort, independent of efforts that use established techniques to vary chemical com- position. New techniques usually strive to produce crystals that are purer, larger, or of a type that cannot be accessed using standard methods (e.g., growth at high pressure). Following are a few examples of promising techniques for advancing the science and technology of crystal growth. • The floating-zone (FZ) technique has become common for the growth of large oxide and intermetallic single crystals (see Appendix D). The precise control of process conditions that characterizes FZ growth can be further enhanced by introducing laser heating in place of conventional halogen

84 Frontiers in C rys ta l l i n e M at t e r or xenon lamps. A high-pressure environment can also be combined with FZ growth, providing the unique opportunity to grow single crystals that are either unstable at ambient pressure or have very high oxidation states, such as those containing Cu3+, Ni3+, and/or Fe4+. It is expected that further variations on the basic FZ process will be explored in the coming years. • Charge carrier doping often plays a key role in producing novel electronic phases, including superconductivity. In many cases, carrier doping has been achieved by chemical substitution, which often suffers from structural disorder introduced by the mismatch of either atomic size or electronic structure of the substituent, or from the presence of a miscibility gap. Electro­static doping, using a field-effect transistor (FET) structure fabricated on the surface of the crystal, is a nonchemical approach to modifying the carrier concentration while avoiding structural disorder. To date, however, the maximum number of charge carriers injected by such “FET chemistry” is limited up to ~10–19cm-3, due to the finite breakdown voltage of typical gate insulator materials. Recently, however, carrier injection using an elec- trolyte cell situated on the crystal surface, instead of a conventional FET structure, was reported. This electrochemical technique can inject carriers up to ~10–20cm-3 and has succeeded in “doping” the normally insulating SrTiO3, to induce a superconducting phase. Thus, co-joined FET structures provide a novel route to altering the electronic properties of crystalline matter, and new FET methods will likely be developed in the future. • Related to the interest in FET structures, interfaces in crystalline hetero- structures are now seen as a well-established type of crystalline matter for studying new electronic states, not only in covalent semiconductors but also in ionic oxides and van der Waals bonded organics. The need for precise control of such interfaces requires characterization of local electronic states on the atomic level using state-of-the-art spectroscopies, including angle resolved photoemission (ARPES) and scanning tunneling microscope (STM). In order to conduct interface characterization in situ during layer-by-layer growth, the integration of the thin-film chamber (either MBE or pulsed laser deposition [PLD]) with ARPES or STM should be pursued. Indeed, efforts are now under way to perform PLD growth in a chamber connected directly to a synchrotron x-ray beam line for photo­ emission spectroscopy. • In order to promote the search for new materials, a technique named combinatorial chemistry (CC) has been developed. This technique was first employed in the pharmaceutical industry to reduce the time and cost associ- ated with producing effective and competitive new drugs. Efforts have been made to apply CC to oxide thin films in order to introduce dopants with a graded concentration (also called composition-spread films). This tech-

Science and Technology of Crystalline Systems 85 nique has seen limited application; one example is the successful fabrication of correlated transition metal oxides such as the colossal magnetoresistance manganites. With these materials, researchers have investigated the phase diagram of the solid solution using a single film. Probes to characterize local lattice constant, resistivity, and magnetization of composition-spread film have been developed. In the future, CC is expected to be applied to an increasing number of materials problems in which compositions to opti- mize a particular property are desired. Future directions in crystal growth techniques will depend strongly on the classes of materials that define the mainstream topics in condensed-matter science. New compounds are constantly being discovered, and novel approaches to produc- ing large single crystals of these compounds are often required. Role of Characterization for New Crystalline Materials discovery The discovery of materials with novel, scientifically or technologically useful properties involves a wide range of scientific expertise, equipment, and processes in various institutions. Developments in the United States over the past 50 years have created an imbalance in this multifaceted process of new materials discovery and development. While facilities for materials characterization have increased capacity and expanded capabilities, the shrinking level of industrial basic research has led to a reduction in synthesis capabilities, as documented elsewhere in this report. This section focuses on opportunities to leverage the greater capacity for materials characterization to advance the scientific understanding and application of new crystalline materials. Laboratory-Scale Materials Characterization Tools In contrast to crystal growth, there has been extraordinary progress in ­laboratory- scale materials characterization tools. Mass production, automation, and informa- tion technology have greatly reduced the costs and increased the efficiency of such tools. Examples of improved materials characterization equipment in this category include x-ray diffraction instrumentation, Raman and infrared spectrometers, and cryogenic systems for specific heat, susceptibility, and transport measurements. Consider, for example, a typical activity that would follow the development of a new material: a specific heat measurement. Twenty years ago, a measurement of the low-temperature heat capacity of a single crystal required expertise that could only be gained through dedication over a career, in conjunction with technical skills in cryogenics, vacuum technology, analogue temperature control, signal process-

86 Frontiers in C rys ta l l i n e M at t e r ing, and data analysis. Today a young researcher with a typical start-up package can purchase a commercial cryostat with fully automated cryogenics, vacuum, and electronics. Heat capacity is an affordable option, as are alternating-current susceptibility, magnetization, thermal conductivity, and electrical transport. In addition, data collection is fully automated and can be monitored remotely. All of these features greatly increase the individual investigator’s measurement capacity. There has also been dramatic progress in data analysis and presentation tools. Twenty years ago data analysis involved the use of unforgiving mainframe com­ puters, and plotting was done by a graphic designer. These tasks are now completed faster and better by commercial software packages on desktop computers. As a result, experimentalists have the capacity to acquire, analyze, and publish compre- hensive data for a much wider range of samples than was the case just 10 years ago. The capacity to probe bulk properties of new materials has increased, perhaps by as much as an order of magnitude per experimentalist. National Facilities for Materials Characterization While bulk experiments carried out in a single-investigator laboratory are typi- cally completed first, full understanding of new materials often requires the use of national facilities that probe matter on the atomic scale. Table 2.1 provides an overview of major federally funded facilities for materials research. Experimental probes include neutrons, x-rays, and microscopy with photons and electrons. The role of facilities in the development of new materials and the sample synthesis requirements for each type of facility are discussed here. TABLE 2.1  Overview of National User Facilities for Materials Science, with FY 2007 User Statistics and Number of Publications for CY 2006 No. of Users No. of Papers Facility Laboratory Radiation in FY 2007a in CY 2006b NIST Center for Neutron NIST Neutrons, reactor 858 406 Research High Flux Isotope Reactor ORNL Neutrons, reactor 72 81 Spallation Neutron Source ORNL Neutrons, pulsed 24 68 spallation Los Alamos Neutron LANL Neutrons, pulsed 272 170 Science Center spallation

Science and Technology of Crystalline Systems 87 TABLE 2.1  (Continued) No. of Users No. of Papers Facility Laboratory Radiation in FY 2007a in CY 2006b Intense Pulsed Neutron ANL Neutrons, pulsed 173 103 Source (IPNS)c spallation National Synchrotron Light BNL X-rays, 2,219 612 Source synchrotron Advanced Photon Source ANL X-rays, 3,420 1,106 synchrotron Advanced Light Source LBNL X-rays, 1,748 593 synchrotron Stanford Synchrotron Stanford University X-rays, 1,151 320 Radiation Laboratory synchrotron Center for Microanalysis University of Illinois Electrons, 600 200 of Materialsd at Urbana-Champaign microscopy (approx.) (approx.) Electron Microscopy ANL Electrons, 199 89 Center microscopy National Center for LBNL Electrons, 183 150 Electron Microscopy microscopy National High Magnetic NHMFL High magnetic 1,144 404 Field Laboratorye fields NOTE: FY, fiscal year; CY, calendar year; NIST, National Institute of Standards and Technology; ORNL, Oak Ridge National Laboratory; LANL, Los Alamos National Laboratory; ANL, Argonne National Laboratory; BNL, Brookhaven National Laboratory; LBNL, Lawrence Berkeley National Laboratory; NHMFL, National High Magnetic Field Laboratory. a Data from the Department of Energy, available at http://www.sc.doe.gov/bes/users.htm, except as other­ wise noted below in footnote e. Users are defined generally as researchers who propose and conduct peer- reviewed experiments at a scientific facility. They include remote users (researchers granted authority to remotely produce data) and offsite users (researchers to whom the facility provides custom-­manufactured materials, tools or devices). An individual is counted as one user no matter how often or how long the researcher conducts experiments at the facility during the fiscal year. User data for the NIST Center for Neutron Research were obtained from private correspondence with operators of the facility. b Committee-collected data. c IPNS closed in February 2008. Table shows the final number of users in FY 2007 and the number of publications for CY 2006. d The Center for Microanalysis of Materials at the University of Illinois at Urbana-Champaign has not been designated a Basic Energy Sciences user facility since FY 2005; hence no current statistics are available. e User information regarding the National High Magnetic Field Laboratory is from the 2007 Annual Report for NHMFL, available at http://www.magnet.fsu.edu/mediacenter/publications/reports/annualreport-2007.pdf.

88 Frontiers in C rys ta l l i n e M at t e r While a laboratory-based x-ray source is the traditional route to atomic-scale structural information, synchrotron x-ray sources provide unprecedented sen­si­ tivity, accuracy, and detail. The much higher source brightness makes it possible to determine the structure of minute crystals (with size on the order of cubic microns) typical of the early stages of materials discovery work. With the current generation of synchrotron x-ray sources, it has even become possible to probe dynamic properties of solids and liquids such as phonons in micron-sized samples. Such information is critical for understanding physical properties, from electronic transport to thermal conductivity. Thin films and nanosize crystalline materials will be incorporated into many future technologies, from quantum computers to medical therapies. The lack of long-range order in one or more dimensions and the very small quantities of materials involved preclude the use of conventional structural probes. Synchrotron x-ray sources provide unique capabilities for prob- ing structure and dynamics under such conditions. Electron microscopy can provide direct, real-space, structural, chemical, and electronic information, resolved at the atomic scale for crystals, buried interfaces, and point or line defects. Recent examples have included the imaging of oxygen vacancies near oxide grain boundaries and in artificially grown heterostructures. Lattice distortions and strain fields can be measured to a precision of a few p ­ icometers using the 0.1 nm or better resolution images acquired on ­aberration- corrected transmission electron microscopes. Scanning transmission electron microscopy (STEM) offers good chemical sensitivity at a comparable resolution and can be used to detect and image individual dopant atoms inside a crystal. Electron energy loss spectroscopy can be performed simultaneously with STEM, and it provides very similar information to that obtained with x-ray absorp- tion spectroscopy, probing the local electronic structure, partitioned by chemical species and site at the atomic scale. Electron holography can measure real-time changes in the electric and magnetic fields in the thinned sample, with the sensi- tivity of a single fluxon. Sample preparation techniques developed for semiconductor failure analysis have greatly improved the quality and speed of sample preparation, which had long been a bottleneck for rapid electron characterization. As this is a real-space imaging method, very small sample quantities are adequate, and the method is well suited for initial explorations of new materials systems. Secondary phases can be identified and studied or avoided as needed. This suggests that electron microscopy can be most effective when coupled closely to sample growth for timely feedback and insight. However, a modern electron microscope can be as sophisticated and complicated to operate as a synchrotron beam line. While most microscopists are well trained in general structural analysis for materials science, very few laborato- ries, whether in the United States or abroad, are producing users trained to quan- titatively analyze and interpret the sophisticated spectra and images generated by

Science and Technology of Crystalline Systems 89 modern instruments. As the complexity of the instrumentation increases, so must efforts to train graduate student and postdoctoral fellows in the use of the latest instrumentation for materials science. Neutron scattering offers unique sensitivity to light atoms and magnetism, momentum-resolved spectroscopy spanning 10 orders of magnitude in energy (0.1 nanoelectronvolt to 1 eV), and the ability to penetrate thick samples, sample environment, and materials processing systems for in situ studies. For fundamental reasons that are directly tied to the utility of the technique, neutron interactions with matter are much weaker than for photons and electrons. In addition, because neutrons are tightly bound in nuclei, neutron sources provide orders-of-magnitude less flux on sample than photon flux at synchrotron facilities. While these factors have limited the use of neutrons for materials science in the past, the Spallation Neutron Source at the Oak Ridge National Laboratory and improved moderators and neutron optics at existing reactor facilities are leading to several orders-of- magnitude improvement in sensitivity and capacity. As a result, a substantial expansion in the use and impact of the neutron-scattering technique in research is occurring, which parallels the transition from rotating anodes to synchrotron sources for x-rays. High magnetic fields continue to be an essential tool for discovering and understanding new materials functionality. An applied magnetic field will alter the course of an electron in a circular path called a cyclotron orbit. At fields greater than 1 tesla (T) (10,000 times Earth’s field), the cyclotron orbit approaches interatomic dimensions. Measurements of electronic and thermal properties versus applied fields greater than a tesla can thus yield information on the electronic structure of a crystalline material. The National High Magnetic Field Laboratory in ­Tallahassee, Florida, is the primary U.S. facility with this focus; it routinely provides static mag- netic fields in excess of 30 T and can reach 100 T in short (millisecond) pulses. As a rule of thumb, 1 T approximates 1 kelvin (K), and thus 100 T fields enable the probing of electronic interactions of 100 K in strength. The availability of such fields enables the study of novel electronic phenomena such as critical phases in the quantum regime, including new fractional quantum Hall states; characterization of high-temperature superconductors for both basic science and large-current- carrying applications; and novel phases in magnetic materials such as spin ice and low-effective-dimensionality materials. Science from this type of facility requires a steady stream of qualitatively new materials, preferably in single-crystalline form so that intrinsic anisotropies can be explored. The growing capacity for facility-based materials characterization is illustrated by recent statistics for single-crystal experiments at the National Institute of Stan- dards and Technology (NIST) Center for Neutron Research. Figure 2.20 shows the number of single-crystal neutron experiments over time, classified by instrument. While single-crystal experiments hold the potential for more detailed atomic-scale

90 Frontiers in C rys ta l l i n e M at t e r 300 Reflectometer SANS TOF 250 Cold-TAS Therm-TAS 200 NCNR Crystal Experiments 150 100 50 0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 Year FIGURE 2.20  Number of single-crystal neutron-scattering experiments per year from 1997 through figure 2-20.eps 2007 at the NIST Center for Neutron Research (NCNR). Color coding indicates the instrument types: reflectometry, small angle neutron scattering (SANS), time of flight (TOF) spectroscopy, and cold or thermal triple axis spectroscopy (TAS). SOURCE: Data provided by Peter Gehring, NIST Center for Neutron Research. information, they also demand more of the instrumentation. The fivefold increase in the number of such experiments over the past decade is a result of progress in instrumentation and user access. With further improvements in these areas at both the NIST Center for Neutron Research and at the Oak Ridge National Laboratory, the capacity for single-crystal experiments will continue to grow dramatically in the coming decade. While single-crystal neutron scattering is seldom the first experiment to be conducted on a new material, such experiments are often necessary in order to understand and control new materials properties. In the preceding decade the supply of novel crystalline materials apparently kept up with the demand, as indi- cated by the factor-of-two average instrumentation oversubscription at NIST. But

Science and Technology of Crystalline Systems 91 as capacity grows, will there continue to be an adequate supply of exciting new materials for a strong scientific program? To answer that question, the committee examined the origin of the materials discoveries underlying the rapidly expanding activity in single-crystal neutron scat- tering at NIST. Figure 2.21 shows crystal experiments over time, classified by the origin of the underlying materials discovery. Europe, the United States, and Japan are the dominant locations of origin. The strong European representation can largely be attributed to the continuing impact of Bednorz and Müller’s discovery of high-temperature superconductivity in La2-xBaxCuO4. The large impact of Japan is associated with discoveries of novel correlated electron systems and magnetism in manganese (Mn)-doped gallium arsenide (GaAs). Materials classes underlying the U.S. impact include superconducting Y2BaCuO3, heavy-fermion intermetallics such as CeCoIn5, and strongly correlated oxides such as Sr3Ru2O7. However, a large frac- tion of the materials attributed to the United States could be called legacy materials with continuing scientific impact, such as LaCoO3 and LaMnO3. This indicates that Figure 2.21 is a lagging indicator of materials synthesis, not yet having picked up the decline in U.S. materials synthesis documented elsewhere in this report. 90 Original Materials Discovery: 80 Europe NCNR User Experiments on Crystals USA 70 Japan Other 60 Unidentified 50 40 30 20 10 0 2003 2004 2005 2006 2007 Year figure 2-21.eps FIGURE 2.21  User experiments at the NIST Center for Neutron Research (NCNR) for the years 2003 through 2007, classified by the origin (country or region) of the materials discovery that initiated the research. SOURCE: Data provided by Peter Gehring, NIST Center for Neutron Research.

92 Frontiers in C rys ta l l i n e M at t e r Opportunities Through Crystalline Matter Discovery The expanded capabilities for materials characterization both in the laboratory and at national facilities create extraordinary opportunities for accelerated progress in materials science. However, investment in and coordination with crystalline matter discovery are essential for realizing this potential. For nanoscale-structured materials, action has recently been taken in rec- ognition of this need. By colocating nanophase synthesis centers with national facilities to probe materials, the Department of Energy has created extraordinary conditions for synergy between crystalline matter discovery and characterization for nanostructured materials. An example is the Center for Nanophase Materials Science colocated with the Spallation Neutron Source. Combining deuterium- labeled polymer synthesis capabilities with small angle neutron scattering and neutron reflectometry provides unprecedented capabilities for creating and prob- ing self-assembled nanoscale structures for fundamental science and applications. Also, by combining the ability to create a thousand to a million copies of a given nanoscale structure with three-orders-of-magnitude-greater neutron brightness, it will be possible to probe phonons and magnons confined to the nanoscale. There are corresponding opportunities for extraordinary insight from x-ray diffraction and spectroscopy with microfocused synchrotron beams at the Advanced Photon Source, National Synchrotron Light Source, and Advanced Light Source on nano- structures created respectively at the colocated Center for Nanoscale Materials, Center for Functional Nanomaterials, and the Molecular Foundry. The advances in materials characterization techniques also create new oppor- tunities for understanding and controlling homogeneous crystalline materials. However, the utility and significance of the science produced depend critically on the resources devoted to discovering new materials and on producing them in suf- ficient quality and quantity for advanced characterization. For neutron-­scattering experiments, the quality and depth of information produced are typically limited only by the sample’s size and quality. Examples of samples used for recent experi- ments are shown in Figure 2.22. Two specific examples of crystal growth as the rate-limiting factor for scientific progress are provided below. The first of these examples is that almost two decades ago, neutron scatter- ing experiments uncovered a spectacular spin resonance in the superconducting state of YBa2Cu3O6+δ. It is a phenomenon that continues to be at the forefront of research in high-temperature superconductivity. To correlate the resonance energy with the superconducting gap amplitude, neutron and angle-resolved photo­ emission spectroscopy or scanning tunneling spectroscopy must be carried out on the same material. It took more than a decade to produce single crystals of ­cleavable Bi2Sr2Ca2CuO8+δ that were large enough for neutron-scattering experiments to accomplish this goal. At the time of the writing of this report, the correspond-

Science and Technology of Crystalline Systems 93 FIGURE 2.22  Images of crystals and sample holders for inelastic neutron scattering experiments. (Top left) ZrW2O8 grown using a flux technique by Glen Kowach, City College of New York. (Top right) Co-aligned ZrW2O8 single crystals used at the ISIS Facility to probe phonons in this negative thermal expansion material. (Bottom left) Co-aligned copper pyrazine dinitrate crystals used to probe an extended critical phase in a quasi-one-dimensional spin-1/2 antiferromagnet as a function of applied magnetic field at the NIST Center for Neutron Research. The crystals were grown in the group of Mark Turnbull and Christopher Landee at Clarke University. (Bottom right) Co-aligned Y 2BaNiO5 single crys- tals grown by H. Takagi and used to probe the Haldane singlet phase in a quasi-one-dimensional spin-1 antiferromagnet. The chains extend approximately along the cylinder axis of the samples. SOURCES: Courtesy of (top left) Glen Kowach, City College of New York; (top right) Joost van Duijn, Universidad Complutense de Madrid, Spain; (bottom left and right) Collin Broholm, Johns Hopkins University. ing letter to Nature reporting a spin resonance in the superconducting state of Bi2Sr2Ca2CuO8+δ had been cited 190 times since its publication in 1999. A second example of research limited by crystal growth capabilities is the recent effort to understand charge and spin dynamics in NaxCoO2.yH2O. The fun-   H.F. Fong, P. Bourges, Y. Sidis et al., “Neutron Scattering from Magnetic Excitations in Bi2Sr2CaCu2O8+δ,” Nature, 398, 588 (1999). Number of citations obtained from the ISI Web of S ­ cience, http://apps.isiknowledge.com. Last accessed April 2, 2009.

94 Frontiers in C rys ta l l i n e M at t e r damental importance of this material and of the exploratory materials synthesis that produced it is indicated by the more than 150 citations annually of the 2003 discovery paper by K. Takada et al. While neutron scattering would provide key information for understanding electronic correlations in this material, adequate single crystals are at present not available. The central challenge for understanding NaxCoO2.yH2O thus arguably lies in single-crystal synthesis. In summary, there has been extraordinary progress in the ability to probe new materials through laboratory-scale instrumentation and national user facilities. The quality of the science produced is, however, critically dependent on (1) the dis- covery of new crystalline materials and (2) the production of high-quality samples with the appropriate morphology and dimensions for advanced characterization. Increased emphasis on the discovery and growth of novel crystalline materials is needed to realize the potential of facilities for new science and for materials-based applications in technologies ranging from information to energy.   K. Takada, H. Sakurai, E. Takayama-Muromachi et al., “Superconductivity in Two-Dimensional CoO2 Layers,” Nature, 422, 53 (2003). Number of citations obtained from the ISI Web of Science, http://apps.isiknowledge.com. Last accessed April 2, 2009.

Next: 3 The Status of Activities in the Discovery and Growth of Crystalline Materials »
Frontiers in Crystalline Matter: From Discovery to Technology Get This Book
×
Buy Paperback | $59.00 Buy Ebook | $47.99
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

For much of the past 60 years, the U.S. research community dominated the discovery of new crystalline materials and the growth of large single crystals, placing the country at the forefront of fundamental advances in condensed-matter sciences and fueling the development of many of the new technologies at the core of U.S. economic growth. The opportunities offered by future developments in this field remain as promising as the achievements of the past. However, the past 20 years have seen a substantial deterioration in the United States' capability to pursue those opportunities at a time when several European and Asian countries have significantly increased investments in developing their own capacities in these areas. This book seeks both to set out the challenges and opportunities facing those who discover new crystalline materials and grow large crystals and to chart a way for the United States to reinvigorate its efforts and thereby return to a position of leadership in this field.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!