Introduction
Now, perhaps more than at any previous time, there is widespread appreciation of the urgent need for new functional materials and the critical role they play in matters of great importance to energy, the environment, sustainability, and the economy. A number of recent panel reports in the United States, including those coordinated by the National Research Council (NRC),1 the Department of Energy,2 and the Office of Science and Technology Policy of the National Science and Technology Council,Reference Fleming and Ratner3 have emphasized the importance of new, enabling materials in the economy and the need to critically examine the current status and future prospects. With support and encouragement from the US National Science Foundation, the authors of this contribution organized a workshop in 2011 on the theme of “Materials by Design” (http://www.mbd.mrl.ucsb.edu). Broadly defined, “Materials by Design” has the goal of achieving desired properties in materials through appropriate tuning of composition, structure, and architecture, which enables a given application. The expertise represented at the workshop ranged from solid-state chemistry, to organic, hybrid, nano, and liquid crystalline materials, to crystal growth, condensed matter experiment, and included both pencil-and-paper and computational theory. The questions of most interest to the participants were: How does one go about increasing the efficiencies in the process of new materials development, and concurrently to achieve this end, how does one make better, more effective use of theory and computation?
While the workshop participants were primarily based in the United States, we believe the findings presented have global relevance. Indeed, the integrated approaches to “Materials by Design” lend themselves very well to international collaborations.
As a starting point, we consider a typical flowchart (Figure 1a) depiction of the process of envisioning/designing and preparing a new material for a particular application, also sometimes referred to as discovery synthesis. Theory and chemical intuition—often in conjunction with the known literature—can combine to suggest a target material, but in many cases, the material produced is not that intended, so the role of serendipity cannot be overlooked. The synthesis of the material, often as small crystals or in polycrystalline form, followed by property measurements, can result in a return to searching for other materials or to the next steps of developing the material. The development process, from small and potentially defective crystals to samples that actually display the desired property, can be long and arduous, but has a potentially huge payoff. In the characterization stage, the properties could be other than that originally envisioned, potentially leading to new applications. Some of the illustrative examples, described later, are representative of various aspects of such flowcharts. The steps involved in discovery synthesis shown in Figure 1a in turn lead to the generation of the broad class of materials that form the base of the materials development pyramid, displayed in Figure 1b. This scheme, first described by Brian Sales at Oak Ridge National Laboratory (this version is from John Mitchell, Argonne National Laboratory), reflects on the very large number of samples of functional materials that normally require screening before a select few are then chosen for refining in the form of better samples and high quality thin films. These then have to be appropriately incorporated into the final device in order for the material to find use.
Schematic flowchart depicting the process of (a) discovery synthesis of new materials, leading to the formation of the base of the (b) materials pyramid. Panel (a) courtesy of P.W. Woodward, Ohio State University, and panel (b) courtesy of John Mitchell, Argonne National Laboratory, modified from a scheme originally suggested by Brian Sales, Oak Ridge National Laboratory.
The two panels of Figure 1 suggest a linear progression of materials development, from conception to application. A key recommendation of the Materials Genome Initiative for Global Competitiveness 2 reflects some of the intrinsic drawbacks of such a linear progression and points to how the different stages in the progression, when addressed by engineering or scientific teams at different institutions, results in “opportunities for feedback between stages that could accelerate the full continuum.” The initiative calls for greater transparency in the way that data can be shared and potentially “mined” such that at every step of the development process, all players can access the knowledge that informs every step, including their own. Concerning the materials pyramid displayed in Figure 1b, the NRC report1 mentioned previously points to the weakened status of efforts in the United States that are focused on crystal growth, including the growth of high quality thin films, particularly following the closure of some large industrial labs that were able to focus on long-term materials development. The specific role of one of these—Bell Labs—in enabling many new technologies, themselves frequently the result of advances in materials research, has been recently described.Reference Gertner4 A question of continuing concern in the context of the United States is whether research in universities and National Laboratories can fill the lacuna left by the closure of large industrial labs that were active in the arena of materials improvement and development.
The purpose of this article is to use illustrative examples in the creation and deployment of functional materials, both organic and inorganic, to make a number of points regarding the creative lifespan of materials, from birth to utility. Examples presented here emphasize the kind of intellectual and experimental resources that are required at the different stages. Suggestions for the advancement of the different domains of materials research are made, when appropriate. The takeaway message of this article is that the advancement of new functional materials science follows many different paths, but all of these paths benefit from interdisciplinary collaborations and, oftentimes, an openness to interesting surprises.
The development of GaN lighting
It is interesting to contemplate the emergence of efficient GaN light-emitting devices that allowed efficient blue and green light sources to become a reality. GaN is not a new material. First prepared in 1932,Reference Johnson, Parsons and Crew5 the structure was solved by Juza and Hahn in 1938.Reference Juza and Hahn6 According to the authoritative history by Schubert,Reference Schubert7 researchers at the RCA labs recognized the potential for a GaN-based blue emitter in the 1960s when they made the first metal-insulator-semiconductor blue emitter. Efficient emission, however, had to await p-type GaN and the formation of p-n junctions, leading to the development of “candela” class blue emittersReference Nakamura, Mukai and Senoh8 and, eventually in 1996, the blue laser diode.Reference Nakamura, Senoh, Nagahama, Iwasa, Yamada, Matsushita, Kiyoku and Sugimoto9 It is interesting to note the breakthroughs did not necessarily come from large, well-funded labs, but instead were the result of a few passionate and resourceful people focusing their efforts on an important problem. The proliferation of white solid-state lighting based on nitride emitters, with its promise of clean, reliable, and efficient lighting, suggests that even materials that are potentially very impactful need not be the product of a vast and concerted effort. This is not to suggest that such a focused effort would not pay dividends.
Computational inspiration for new multiferroics: The example of strained EuTiO3
Magnetoelectric multiferroics are compounds that can display concurrent and coupled magnetic and polar order, and, in the best possible case, these orders are respectively ferromagnetism and ferroelectricity.Reference Spaldin and Fiebib10, Reference Kimura11 This is an area where density functional calculations have played a particularly important role in unraveling the different processes and mechanisms that allow for the observation of these multiply ordered ground states. In one such example using density functional modeling, a mechanism for stabilizing a multiferroic phase in a material that is otherwise antiferromagnetic and paraelectric in the bulk was identified.Reference Fennie and Rabe12 It was suggested that such control could be obtained through tuning the competition between the different magnetic and structural instabilities in the compound. The control parameter was biaxial strain, which when applied to the material, EuTiO3, was proposed to give rise to a multiferroic figure-of-merit that was about a thousand-fold greater than the best previously known material. The original theoretical study chose to focus only on the application of compressive strain. The critical strain value necessary to drive the system into the multiferroic phase was about 1.1%. It was then suggested from the experimental side that there are many more substrates that would result in tensile strain being induced when EuTiO3 is grown epitaxially on them. As a result of this suggestion, the full strain phase diagram of EuTiO3 was obtained computationally, and it was found that significantly less tensile strain, about 0.6%, is required to drive the compound into a multiferroic phase.Reference Lee, Fang, Vlahos, Ke, Jung, Fitting Kourkoutis, Kim, Ryan, Heeg, Roeckerath, Goian, Bernhagen, Uecker, Hammel, Rabe, Kamba, Schubert, Freeland, Muller, Fennie, Schiffer, Gopalan, Johnston-Halperin and Schlom13 The smaller strain required also allows thicker and higher quality films to be grown. As described in a recent letter,Reference Lee, Fang, Vlahos, Ke, Jung, Fitting Kourkoutis, Kim, Ryan, Heeg, Roeckerath, Goian, Bernhagen, Uecker, Hammel, Rabe, Kamba, Schubert, Freeland, Muller, Fennie, Schiffer, Gopalan, Johnston-Halperin and Schlom13 strained EuTiO3 was grown on a variety of substrates and characterized by a team of collaborators from around the world, confirming the suggestions from computation. Figure 2 displays the computed property phase diagram of strained EuTiO3 as a function of the strain and schemes of the resulting crystal and magnetic structures. It should be noted that in these ABO3 perovskites, magnetism arises on the A site and polar behavior from the displacement of the B ions.
Suggestions from first-principles electronic structure calculations that perovskite EuTiO3 can, under both tensile and compressive strains, switch its ground state from being paraelectric (PE) and antiferromagnetic (AFM) to being ferroelectric (FE) and ferromagnetic (FM) have now been experimentally realized. The upper panel (a) shows the computed dependence of the ground state on strain, and the lower panel (b) displays schemes of different magnetic and polar orderings in the perovskite structure corresponding to the regimes in (a). Orange atoms at the corners of the perovskite cell are Eu2+ and carry spin, indicated by arrows. TiO6 octahedra are indicated with Ti4+ shown as green spheres at the centers of the octahedra. Figure courtesy of Craig Fennie, Cornell University.
The key role of formal theory: Development of topological insulators
The discovery of three-dimensional topological insulators provides an example of the interplay between different kinds of theory and experiment and, in particular, emphasizes that formal theory can play a key role in predicting and understanding new phenomena. Materials-specific modeling, as embodied in density functional electronic structure theory, can then play a collaborating role in suggesting real systems where certain phenomena may be observed. Interestingly, at a time when much of condensed matter physics is focused on correlated electronic phenomena, the understanding of topological insulators largely requires simple, single-particle theories.Reference Moore14, Reference Hassan and Kane15 Topological insulators are materials that are insulating in bulk but have protected metallic surface states as a result of spin-orbit coupling. Formal theoretical work, motivated by seeking generalizations of the quantum Hall and quantum spin Hall effects in two-dimensional systems, showed that similar mathematical structures (topological invariants) could, in principle, be generated by spin-orbit coupling, leading to new phases, dubbed topological insulators. Several candidate materials were then proposed based on heuristic arguments and first-principles calculations, contributing to the experimental discovery of the characteristic massless surface state and spin-momentum locking in BiSb alloys and Bi2Se3. This is an area that has been recently reviewed, for example, in References Reference Moore14 and Reference Hassan and Kane15. Following the initial role of formal theory, more materials-specific approaches have come into play, and a number of suggestions for interesting materials classes have been put forth, such as the suggestion that the half-Heusler XYZ structure type, comprising three interpenetrating fcc sublattices of (usually) two metals, X and Y, and a main group element, Z, is a rich playing field for the area of tunable, multifunctional topological insulators.Reference Chadov, Qi, Kübler, Fecher, Felser and Zhang16, Reference Lin, Wray, Xia, Xu, Jia, Cava, Bansil and Hassan17
Better samples lead to new science: The example of GaAs heterostructures
The science of topological insulators, referred to in the previous section, is in fact closely inspired by studies on the fractional quantized Hall effect. Single-crystal GaAs films are a natural extension of the deep experience of the scientific and technological community with single-crystal Si. A search by physicists for a more perfect semiconductor-insulator interface than realizable in Si-SiO2 led to the development of the GaAs-AlAs system, in which both the conducting GaAs and the insulating AlAs are incorporated in the same single crystal, usually in an alternating fashion. Using advances made in techniques to epitaxially grow single-crystal films, near-perfect crystal interfaces can be produced. In these films, the absence of a disordered electronic interface gives the freedom to reduce the conducting channel thickness to about 10 nm, which is near the quantum size of conduction electrons in GaAs. The electrons can thus move in the plane of the GaAs channel but not out of the GaAs layer because of the AlAs insulating barriers. Thus, a material that is effectively two-dimensional can be created.
The unusual nature of two-dimensional electrons can be probed in Hall-effect experiments. For metal samples, such as the gold leaf foil used originally by Edwin Hall in 1879, the resistance grows smoothly and linearly with an increase in the magnetic field B. However, for high-purity semiconducting systems at low temperatures, the Hall resistance shows a series of discrete plateau steps that cannot be understood without considering the quantum properties of the conducting electrons. Even more striking is that the numerical values of the various resistance plateaus are given by exact ratios of fundamental physical constants, at values of h/(νe 2), where h is Planck’s constant, e is the charge on the electron, and ν is the filling factor, which in the Quantum Hall Effect takes on integer values.
The realization that increasing the crystalline quality in the two-dimensional region could lead to a more robust quantum Hall effect led to experiments using the potentially higher-quality GaAs-AlGaAs interface. In 1982, Stormer et al.Reference Tsui, Stormer and Gossard18 published a truly novel experimental result shown in the top panel of Figure 3. Not only did they see the integer effect, but they also saw a clear plateau at ν = 1/3, in other words, a fractional rather than an integer value. They had discovered the fractional quantum Hall effect. This discovery in GaAs launched an effort to further improve the quality of the GaAs crystal samples. The panels in Figure 3 show the increasing richness of observed fractional states as the sample quality (measured by the electron mobility μ) has improved over the years. For a sample of mobility 5 × 105 cm2 V−1s−1,19 one sees that the states at 1/3 and 2/3 are joined by other states, and in the bottom-most trace where the mobility is of the order of 107 cm2 V−1s−1,Reference Pan, Stormer, Tsui, Pfeiffer, Baldwin and West20 these states are joined by a hierarchy of states, all having odd denominators and all organized around an absent state at ν = 1/2. These experiments provide dramatic evidence for a series of new emergent particles called composite fermions.Reference Jain21
Better samples lead to new science: Longitudinal resistance, Rxx, for three different two-dimensional electron gas samples as a function of the magnetic field, as expressed in terms of filling factor ν = nh/eB, where n is the electron density, h is the Planck constant, e is the charge on an electron, and B is the magnetic field. The samples are formed at the interface of crystalline GaAs and AlGaAs grown by molecular beam epitaxy. The samples are distinguished by different values of the mobility, μ, that have increased systematically over time. The upper frame shows the first report, in 1982, of fractional filling at ν = 1/3 and T = 0.48 K.Reference Tsui, Stormer and Gossard18 The middle frame shows new states at ν = 1/3, 2/5, 2/3 achieved in higher mobility samples at T = 0.55 K.Reference Stormer, Chang, Tsui, Hwang, Gossard and Wiegmann19 The bottom frame shows data from a typical modern high mobility sample at T = 0.35 K.Reference Pan, Stormer, Tsui, Pfeiffer, Baldwin and West20 This sample exhibits self similarity between the ν = 0 and ν = 1/2 states that is the basis for the “composite fermion” description of electronic states.
Better samples lead to better understanding: The example of SrTiO3 and conducting polymers
One of the lessons learned from GaAs is the importance of sample quality. This has recently become of particular relevance in the examination of thin films of high quality oxides and their heterointerfaces that have so captured the imagination with the possibility of oxide electronics.Reference Hwang, Iwasa, Kawasaki, Keimer, Nagaosa and Tokura22 The question of whether the kinds of growth processes that are applicable in the preparation of high quality GaAs films can be transferred to oxide thin films has recently been addressed.Reference Jalan, Moetakef and Stemmer23 Molecular beam epitaxy of SrTiO3 uses a molecular beam source for Sr and titanium tetraisopropoxide (TTIP) for the Ti source. A key to growing high quality films is to work under conditions, also frequently referred to as growth windows, which naturally lead to the correct stoichiometry of the phase being formed. In SrTiO3, this is appropriately probed using the unit cell parameter. SrTiO3 is an optimally bonded band semiconductor, and any change from the “1-1-3” stoichiometry results in underbonding, as seen from an expansion of the lattice. The growth windows therefore correspond to regions where the lattice parameter displays a plateau as a function of varying the ratio of starting materials, and this plateau is a minimum. This is displayed in the left panel of Figure 4a.Reference Jalan, Moetakef and Stemmer23 The use of such a growth window has resulted in complex oxide films with unprecedented mobilities,Reference Son, Moetakef, Jalan, Bierwagen, Wright, Engel-Herbert and Stemmer24 as seen in the right panel of Figure 4a. Complex oxides are fascinating materials because of the range of physical phenomena that they can display and, in bringing semiconducting oxides into this high-mobility regime, what seems to be one of the last materials frontiers is being traversed.
Better samples lead to better understanding. (a) The left panel displays the existence of growth of windows (shaded regions) in the hybrid molecular beam epitaxy (MBE) growth stoichiometric SrTiO3 as monitored from the cubic unit cell parameter a.Reference Jalan, Moetakef and Stemmer23 The horizontal axis displays the beam equivalent pressure (BEP) ratio of titanium tetraisopropoxide (TTIP) to strontium. The right panel displays the record mobilities that become achievable from the films grown in these growth windows.Reference Son, Moetakef, Jalan, Bierwagen, Wright, Engel-Herbert and Stemmer24 Figure courtesy of Susanne Stemmer. (b) Improved crystallinity in semiconducting polymer films results in significantly increased mobilities, as measured using a field-effect transistor geometry. Figure courtesy of Michael Chabinyc. R, generic organic group; S, sulfur; N, nitrogen; n, an integer; F8T2, dioctylfluorene-bithiophene; P3HT, poly(3-hexylthiophene); PBTTT, poly(2,5-bis(3-alkylthiophen-2-yl)thieno[3,2-b]thiophene).
In the area of organic electronics, discussed in the next section, the extent of delocalization is important in designing the electronic properties of molecules. In addition, it is also critical to think of the morphology and, specifically, the extent of crystallinity in the case of evaporated films. In Figure 4b, the mobility, measured using field-effect transistor geometry, of some amorphous, liquid crystalline, and semicrystalline conjugated polymers are displayed. A clear trend is observed that with increasing crystallinity, the mobility increases by two orders of magnitude. The availability of organic polymer films displaying increasingly higher mobility and potentially band-like transport enables better understanding and even modeling of these highly complex systems. Comparing organics with the oxides, and the more classical main group semiconductors referred to previously, it becomes evident that although the observed phenomena can differ by many orders of magnitude, some of the goals (e.g., the fabrication of transistors) in the very divergent materials classes can be similar.
The organic advance of optoelectronics: Light-emitters, photovoltaics, and transistors
Organic materials hold a great deal of promise for applications in electronic and optoelectronic devices. Both molecular and polymeric materials have several properties that make them well suited to these applications. Both are readily tailored by chemical substitution to grossly and finely tune their electronic properties, including filled and vacant state energies as well as their absorption and excited state energies. They can have conductivities ranging from as low as those of the best wide gap insulators, through semiconducting to metallic. Their optical absorptivity can be significantly larger than direct bandgap semiconductors, with extinction coefficients as high as 106 cm−1. They can also be designed to emit with high efficiency, ranging in color from the UV to the near-infrared. Lastly, both molecular and polymeric materials can be peripherally substituted to affect the mechanical and morphological properties without significantly affecting the electronic properties of the materials. This last property of organic materials is key to their future applications, as we move toward organic materials by design, in which the microscopic and macroscopic properties of the compounds are tailored independently to achieve the optimal materials for each application.
The first organic optoelectronic device to go into full-scale production was the organic light-emitting diode (OLED), and significantly more manufacturing capacity is being planned and added at this time. Electroluminescence in organic materials has been known since the mid-1960s.Reference Pope, Kallmann and Magnante25, Reference Helfrich and Schneidere26 However, the demonstration of the first heterostructured OLEDs in the late 1980s set the stage for major advances in efficiencies and color tunability of these devices.Reference Helfrich and Schneidere27, Reference Tang and VanSlyke28 By carefully controlling charge injection and recombination as well as the site of light emission in OLEDs, internal efficiencies have improved from ca. 5% in 1987 to 100% in the early part of this century.Reference Tang and VanSlyke29–Reference Flamigni, Barbieri, Sabatini, Ventura and Barigelletti32 The rapid development of OLEDs, from a laboratory demonstration to a commercial product, involved the dedicated work of a broad range of scientists and engineers working together to both understand and improve electroluminescent devices. The solutions that were developed involved the use of thin amorphous films. Materials need to be amorphous to achieve acceptable film uniformity for the 10 nm to 40 nm films that are used in OLEDs. These film thicknesses are necessary to keep the resistivity of the device low enough to achieve low drive voltages. Manufacturing these devices would be easier, and the power efficiencies of the devices would be improved if we could increase film thickness and decrease film resistivity, respectively. While these two goals seem mutually exclusive, if crystalline thin films—some of which are shown in Figure 4b—could be used for OLED preparation, it could be possible to both increase film thickness and decrease resistivity. Beyond simply growing uniform crystalline films, it is important to be able to predetermine the orientation of the lattice to maintain maximum conductivity in the desired direction, parallel or perpendicular to the substrate. Two other existing devices could benefit from such a level of materials design and control: photovoltaics and transistors.
Organic solar cells have the potential to achieve high efficiency for conversion of sunlight into electricity at a cost well below existing solar technologies. Organic photovoltaics (OPVs) with efficiencies exceeding 10% have been certified by the National Renewable Energy Laboratory,33 although the structures have not been published. Efficiencies of >25% are possible for single junction cells.Reference Ulbricht, Beyer, Freibe, Winter and Schubert34 While the OPV is in many ways the microscopic inverse of the OLED, there are important differences between the two. The high resistivity of organic materials can be tolerated in an OLED, because very thin films could be used to minimize the bias needed to inject and transport charge in the devices. If such thin films are used in OPVs, the problem of resistivity is minimized, but the films are so thin that very little light is absorbed. Thus, in an OLED, it is possible to restrict all light emission to a very thin film, typically less than 30 nm, while in an OPV, the light absorption must be spread over a much larger volume to achieve good light collection. The most efficient OPVs are those that use thermal or solvent annealing to achieve a degree of crystallinity in the organic materials. This is often done with donor (D) and acceptor (A) materials in an interpenetrating network to efficiently collect excitons (i.e., the bulk heterojunction, BHJ), since simple layered devices give exciton diffusion lengths that are too short to give efficient collection at the D/A interface. This is again a technology that would be well served by being able to control film crystallinity and morphology over a large area. Crystalline materials will have lower resistivity and higher exciton diffusion lengths than the amorphous or polycrystalline materials that are used currently. If it were possible to grow highly crystalline thin films of organic materials, with the preferred orientation of the organic moieties to direct excitons to the D/A interface and charges out the electrodes, OPVs could be fabricated with simple layered structures, rather than BHJs, and achieve efficiencies closer to the 25% theoretical limit.Reference Giebink, Wiederrecht, Wasielewski and Forrest35
Organic field-effect transistors, OFETs, have advanced tremendously over the last decade.Reference Małachowski and Żmija36, Reference Wu, Liu and Zhu37 A key limitation of OFETs is the carrier mobilities of organic materials. While values higher than 1 cm2V−1s−1 have been reported for organic materials, these are still many orders of magnitude below those seen for inorganic semiconductors, including GaAs and SrTiO3 mentioned earlier. The highest mobilities, and thus the best transistor performance, are observed for single-crystalline materials. An approach to using such materials in OFETs involves thermal annealing to achieve crystalline grains that are larger than the channel length.
The key to improving the efficiencies of organic electronic and optoelectronic devices is to develop the methodology to grow organic materials in predetermined structures and orientation in large area thin films. To develop this capability, it will take a multidisciplinary and consistent effort. This will start with theorists. It is possible to calculate many molecular parameters and even the bulk properties of a material in a given structure, to a limited degree. It has not been possible, however, to reliably predict the morphology an organic material will adopt in the solid state. In order to create methodologies for crystalline organic film growth, it is important to couple researchers in theoretical modeling with experimentalists to develop models that effectively deal with both kinetic and thermodynamic parameters to predict the structures of organic thin films. The first step is to develop deposition tools that allow the researcher to accurately control the deposition and growth parameters for the organic thin film, from both solution and vapor sources. Next, theorists and experimentalists need to work together to design, synthesize, and study the deposition of molecules that are predicted to give preferred structures under a set of deposition conditions. Materials that fail to give the expected structure can be used to refine the model and develop the deposition techniques further, so there are no real failures in this approach. Working through this cycle many times should provide a set of design parameters that can be used to control the morphologies and properties of organic thin films.
New directions in nanoscience
The field of nanomaterials has progressed considerably since its emergence as a distinct discipline from cluster and colloid science some 30 years ago. Advances have been spurred by new chemistry knowledge as well as new physical techniques and/or greater access to appropriate physical techniques (e.g., electron microscopy, atomic force microscopy). At the present stage of development, syntheses exist for a range of nanomaterials, mostly single-element phases or binaries. This is in contrast to the better-developed solid-state chemistry discipline, where highly complex materials—ternary, quaternary materials and beyond, and defect or dopant controlled materials—are at the forefront of science. With the excellent foundation of the past three decades of research, the field of nanomaterials is now poised to tackle a range of technical challenges. This includes the synthesis of new, more complex materials, better methods for nanoparticle integration, more detailed physical characterization of existing materials (wherein new phenomena can be found), combined theoretical/experimental approaches to understand nanoscale phenomena, and the deployment of nanomaterials in device architectures.
The rich phase diagrams established for solid-state materials provide a great source of new material compositions to be targeted on the nanoscale, both as homogeneous (alloys or compounds) and heterogeneous (e.g., core-shell) architectures. The formation of homogeneous structures of ternary phases has been achieved by direct synthesis from multicomponent solutions or from chemical transformations.Reference Vasquez, Henkes, Bauer and Schaak38 Likewise, the ability to tune the electro-optical properties of quantum dots by overcoating with a semiconductor of a different bandgap, or the creation of multifunctional structures by combining mutually exclusive properties (e.g., magnetism and photoluminescence) into a single nanostructureReference Kim, Achermann, Balet, Hollingsworth and Klimov39, Reference Redl, Cho, Murray and O’Brien40 have proven to be powerful paradigms in nanomaterial design. Structure-property-size correlations in complex nanomaterials will be a fruitful area of exploration for some time to come, driven by improved synthetic methods that enable high quality nanocrystals and nanocomposites to be generated.
Doping or the introduction of defects into bulk solids is an established method for tuning the optoelectronic properties, but doping on the nanoscale has proven to be a complex problem, restricting the kinds of phenomena that can be accessed in such materials. This problem has been addressed by a combination of experimental and theoretical approaches,Reference Erwin, Zu, Haftel, Efros, Kennedy and Norris41 which revealed the key issue to be low binding energies for dopant ions on certain crystalline faces. If the residence time is not sufficient for host layer overgrowth (i.e., incorporation), low to zero incorporation occurs. A key advantage of this “trapped dopant” model is that binding energies for ions and competing surfactants can be calculated for distinct facets, enabling doping to be enhanced by appropriate solvent choices and/or crystallite morphologies. Exploitation of this insight enables the effect of coupling of size and doping on the nanoscale to be probed, for example, the effects of quantum confinement effects on magnetic semiconductors. Combining theory with experiment is also shedding light on physical phenomena, as demonstrated by the application of d-band theory to the prediction of catalyst performanceReference Nikolla, Schwank and Linic42 or the modeling of interface scattering to predict optimal grain sizes in thermoelectric nanocomposites.Reference Popescu, Datta, Nolas and Woods43 The synergy that arises when experiment and theory act in a feedback loop enables the rational construction of nanomaterials for targeted applications.
Employing nanoparticles as discrete components for medical applications based on optical and magnetic properties was an early focus of investigation and continues apace. The integration of nanomaterials into electronic devices, of particular relevance for energy applications (photovoltaics, thermoelectrics, and batteries), has progressed more slowly, as it is predicated on conductive particle interfaces. Methods that generate well-formed discrete nanoparticles typically produce materials with long-chain organics on the surface, preventing facile transport of electrons. Advances have been made by the use of conducting ligands,Reference Milliron, Alivisatos, Pitois, Edder and Fréchet44 ligands that are thermally labile (t-butylthiol),Reference Webber and Brutchey45 replacement of organic ligands with inorganic ligands such as molecular chalcogenide clusters,Reference Kovalenko, Scheele and Talapin46 sulfide,Reference Zhang, Hu, Sun, Hovden, Wise, Muller and Robinson47 or hydrazine,Reference Urban, Talapin, Shevchenko and Murray48 and use of ligand removal coupled to direct inter-particle bonding (sol-gel particle assembly).Reference Mohanan, Arachchige and Brock49 A consequence of such modifications is changes in surface defects that modify luminescence characteristics and the potential reduction in quantum confinement in semiconductors. Alternative approaches focus on building nanostructures directly from molecular building blocks, generating gels or templated solids and obviating the need for ligand modification strategies altogether.Reference Rolison and Nazar50
The role of serendipity: New inorganic pigments from the search for multiferroics
In Horace Walpole’s 1754 retelling of a Persian fairy tale, three well-raised princes from Serendip—modern-day Sri Lanka—are continually making discoveries of things they are not in quest of. The word that has emerged into common currency from this tale, serendipity, is used to suggest the act of finding something useful without really looking for it, or even perhaps without prior knowledge of its existence. Modern materials science is a chronicle of serendipitous discovery exemplified by the original discovery of superconductivity by Kamerlingh-Onnes, the discovery in 1938 of polytetrafluoroethylene (Teflon) by Roy Plunkett of Dupont Company, or the recent discovery of superconductivity in layered transition metal pnictides by the Hosono group at the Tokyo Institute of Technology, who were in fact searching for new transparent conductors.Reference Kamihara, Hiramatsu, Hirano, Kawamura, Yanagi, Kamiya and Hosono51 In this regard, we emphasize the importance of exploratory synthesis in materials research, whose purpose is to escape the intellectual confines of what we know in order to discover the new and the unexpected.
The recent discovery of highly durable and non-toxic inorganic color pigments in oxides with transition metal cations in trigonal bipyramidal (TBP) coordination can be considered as a textbook example of how a systematic approach to discover new functional materials in one area can lead to a totally unexpected discovery in another, unrelated area. Both YInO3 and YMnO3 are known in the common orthorhombic and centrosymmetric perovskite structure, but they are also readily prepared in an acentric hexagonal structure (Figure 5, left panel),Reference Waintal and Chenavas52, Reference Pistorius and Kruger53 consisting of layers of Y3+ ions separating layers of corner-shared MO5 trigonal bipyramids (M = In or Mn). This structure has been of considerable recent interest because it exhibits an unusual form of improper ferroelectricity (meaning the ferroelectric polarization is not the primary driving force for the structural change) as a consequence of the tilting of MO5 polyhedra.Reference Van Aken, Meetsma and Palstra54, Reference Van Aken, Palstra, Filippetti and Spaldin55 Such ferroelectricity is compatible with M-site magnetism and therefore allows multiferroic behavior. To explore such multiferroic behavior, the complete YIn1−xMnxO3 solid solution was prepared. Despite the large size mismatch between In3+ and Mn3+, these cations are randomly disordered in the TBP sites.Reference Smith, Mizoguchi, Delaney, Spaldin, Sleight and Subramanian56 Unfortunately, such a disorder suppresses the multiferroic order parameter, rendering the system unsuitable for applications as a magnetoelectric material.Reference Dixit, Smith, Subramanian and Lawes57 However, a surprisingly intense and bright blue pigment is obtained when Mn3+ is introduced into the trigonal bipyramidal sites in hexagonal YInO3, and over much of the YIn1−xMnxO3 solid solution range, in spite of the fact that YInO3 and YMnO3 are white and black, respectively. An intense blue color is also obtained when Mn3+ is introduced into trigonal bipyramidal sites in other oxides. These unexpected discoveries have led to the rational design of other earth-abundant, environmentally benign and highly durable inorganic pigments. Substitution of Fe3+ in hexagonal YInO3 resulted in the formation of intense orange compounds,Reference Jiang, Li, Sleight and Subramanian58 whereas the simultaneous substitution of Cu2+ and Ti4+ along with Al3+ in TBP sites resulted in the formation of oxides exhibiting intense green color.Reference Smith, Sleight and Subramanian59
(a) Crystal structure of polar, hexagonal YMnO3 displaying the B cation (Mn atoms here) in a trigonal bipyramidal environment, indicated by the blue polyhedra. Orange spheres represent O atoms, and the dark gray spheres represent Y atoms. (b) Examples of oxide pigments prepared by appropriate substitution of transition metal ions into trigonal bipyramidal coordination.
Serendipitous discoveries require intellectual effort and are not simply a question of luck, and, indeed, we hark back to the tale that tells us that the princes were actually well trained to observe and interpret whatever they happened upon. Unpredictable results and discoveries are as much a part of scientific endeavor, but serendipity manifests when the alert and trained mind responds to an unexpected situation. Basic exploratory research necessarily incorporates serendipity, and thus sustained support for exploratory research should not be neglected.
Conclusions
The systematic search for new functional materials specifically designed to address existing challenges of technology has seen a recent and well-deserved surge of attention. In particular, the need to fill the materials discovery gap engendered by the shrinking of once-prolific industrial laboratories1 and the need (and the necessary means) for approaches that can speed up the arc from discovery to technologyReference Fleming and Ratner3 have been emphasized. In this article, we have simultaneously pointed to the needs in some emerging areas as well as the requirements for better materials that can enable new science. The consensus that is emerging is that in areas where a targeted search for better materials is necessary, there is a need for strong collaboration between theory and experiment at every level. At the same time, history tells us that exploratory research and the ability to be surprised should continue to play an important role and should be supported appropriately.
Acknowledgment
We are very grateful to the National Science Foundation (DMR 1115294) for supporting the Materials by Design Workshop, and especially to Drs. Linda Sapochak and Daniele Finotello for their support and encouragement. We thank all the participants of the Workshop for their many useful and creative inputs, and acknowledge Michael Chabinyc, Craig Fennie, John Mitchell, Joel Moore, Susanne Stemmer, and Pat Woodward for providing us with figures and explanations.
Ram Seshadri is a professor in the Materials Department and in the Department of Chemistry and Biochemistry at the University of California, Santa Barbara (UCSB). He received his BS degree in chemistry from St. Stephens College, Delhi, in 1989, and his PhD degree in solid-state chemistry from the Indian Institute of Science, Bangalore, in 1995. After some years as a postdoctoral fellow in Caen, France, and Mainz, Germany, he started a faculty career as an assistant professor in Bangalore in 1999, before moving to UCSB in 2002. Seshadri’s research program addresses structure-composition-property relations in functional inorganic materials, focusing currently on magnetic and correlated materials, catalysts, and phosphors. Seshadri can be reached at the Materials Department and the Department of Chemistry and Biochemistry, Materials Research Laboratory, University of California, Santa Barbara, CA, USA; tel. 805-893-6129; and email seshadri@mrl.ucsb.edu.
Stephanie L. Brock is a professor in the Department of Chemistry at Wayne State University. She received her BS degree in chemistry from the University of Washington in 1990 and her PhD degree in chemistry from the University of California, Davis, in 1995. From 1995 to 1999, she was a postdoctoral associate at the University of Connecticut working on nanoscale manganese oxide materials. Brock’s research focuses on the synthesis, properties, and applications of metal pnictide and chalcogenide nanomaterials; the development of sol-gel nanoparticle assembly methods and their application to non-traditional materials; and establishment of novel hybrid organic/inorganic nanomaterials for optoelectronic and drug-delivery applications. Brock can be reached at the Department of Chemistry, Wayne State University, Detroit, MI, USA; tel. 313-577-3102; and email sbrock@chem.wayne.edu.
Arthur P. Ramirez joined the University of California, Santa Cruz, as dean of the Baskin School of Engineering in May 2009 from LGS, a subsidiary of Alcatel-Lucent. He earned his BS and PhD degrees in physics at Yale University. Ramirez spent much of his career at AT&T Bell Laboratories, where he served as the director of the device physics research department before becoming leader of composite materials and device development for LGS. His research interests include frustrated systems, multiferroic magnetic materials, and organic semiconductors. Ramirez can be reached at the Baskin School of Engineering, University of California, Santa Cruz, CA, USA; tel. 831-459-2158; and email apr@ucsc.edu.
M.A. Subramanian is the Milton Harris Professor of Materials Science at Oregon State University. He received his BS and MS degrees from the University of Madras, and his PhD degree from the Indian Institute of Technology, Madras, in 1981. Following postdoctoral research at Texas A&M University (1982–1984), Subramanian joined Central Research & Development at DuPont, where he remained until his recent appointment at Oregon State. His research interests include designing new inorganic solid-state functional materials for emerging applications in electronics, solid-state energy conversion, and other areas. Subramanian can be reached at the Department of Chemistry, Oregon State University, Corvallis, OR, USA; tel. 541-737-8235; and email mas.subramanian@oregonstate.edu.
Mark E. Thompson is a professor of chemistry and materials science at the University of Southern California (USC). He received his BS degree in 1980 from the University of California, Berkeley, and his PhD degree in 1985 from the California Institute of Technology—both in chemistry. After two years as a postdoctoral fellow at Oxford University, he accepted a position as an assistant professor in the Chemistry Department at Princeton University. Thompson joined USC in 1995. His research program involves the study of new materials and devices for electroluminescence, solar energy conversion, chemical/biological sensing, and catalysis. Thompson can be reached at the Department of Chemistry, University of Southern California, Los Angeles, CA 90089, USA; tel. 213-704-6402; and email met@usc.edu.
