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 Progress and Achievements in Materials Research over the Past Decade This chapter highlights select examples of the advances that have been made since 2008 in the broad field of materials science and materials engineering. It is impos- sible to document all of the progress that has been made, so this chapter emphasizes a selection of achievements in fundamental understanding and enhancements of the properties of materials, and the acceleration of materials development and optimization of the properties through use of computational methods. What is not covered in this chapter are advances that have been made in computational methods, the developments in experimental tools and capabilities as well as in synthesis, and processing methods that are applied by the materials community. These enabling methodologies have undergone significant advances in the last decade sufficient to warrant a separate chapter. A significant development in this period was the announcement in 2011 by President Barack Obama of the Materials Genome Initiative (MGI). The ultimate goal of the MGI was âto discover, manufacture, and deploy advanced materials twice as fast, at a fraction of the cost.â1 The key to the initiative was the Materials Innovation Infrastructure, which realized that progress would be accelerated by working at the intersection of computational materials science and engineering, materials informatics, and synthesis and processing, as well as material character- ization and property assessment. This framework of combining the three areas was intended to operate synergistically across all seven components of the materials 1 Office of Science and Technology Policy, 2011, Materials Genome Initiative for Global Competitive- ness, Washington, DC, June. 26
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 27 development continuumâdiscovery, development, property optimization, systems design and integration, certification, manufacturing, and deployment (including sustainability and recovery). Reduction in the time from discovery to deployment and in the overall development cost was envisioned to be achieved by replacing the current linear approach to the materials development continuum with one in which each area was in constant communication with all other areas. These differ- ent design concepts are captured schematically in Figure 2.1. For the envisioned change to become effective, the different sectors engaged in materials engineering need to be educated about the efficiencies to be gained by this approach and become willing to use materials developed through this approach. An example, selected from many possibilities, of how this methodology has been applied in industry is given in Box 2.1. This initiative impacted the federal funding agencies, with the result that material databases have proliferated and teams of experimentalists and computational materials scientists and engineers have been formed to address scientific challenges. This has impacted the discipline, with more multidisciplinary teams forming to tackle key material challenges. 2.1 METALS 2.1.1 Accelerated Development Powered in part by widespread efforts such as the Integrated Computational Materials Engineering (ICME)2 approach to materials development, the National FIGURE 2.1â Linear versus continuous feedback materials continuum development. The traditional linear approach leads to longer and more expensive discovery to deployment cycles, whereas within the continuous feedback method each sector is in continual discussion with each and every other sector. The latter approach was seen as one in which the time from discovery to deployment of a new material would be reduced significantly from the typical 20 years for many industries. 2 National Research Council, 2008, Integrated Computational Materials Engineering: A Transforma- tional Discipline for Improved Competitiveness and National Security, The National Academies Press, Washington, D.C., https://doi.org/10.17226/12199.
28 F r o n t i e r s o f M at e r i a l s R e s e a r c h BOX 2.1 Accelerating Product Development Through Computational Modeling of Manufacturing Processes Advanced computational models are being used in manufacturing to accelerate the devel- opment and introduction of new materials into marketable products. The computational models capture manufacturing-history-sensitive materials information with manufacturing simulation and allow engineering product performance analysis to be conducted in silico. The impact is a significant acceleration of the product development cycle time. For example, Brunswick Cor- poration, and particularly its Mercury Marine division, used this approach for new aluminum and stainless-steel alloy development as well as castings to improve the performance of existing materials and products, and ultimately, to reduce production cost. Figure 2.1.1 compares the deformation behavior of two types of rivets for use on boats produced by the Lund boat division and Mercury Marine. By using this approach, the compa- nies compared different geometries and selected the one that met performance requirements in silico. This manufacturing deformation of each rivet was combined into an overall structural model of a boat and loaded with appropriate forces âin the computerâ to assess boat durabil- ity. Therefore, they avoided the need to make and test different geometries, shortened product development time by approximately 35 percent, reduced the risk associated with interaction between the manufacturing process and the loading on the boat in service. This improved the end product and reduced development costs. FI GURE 2.1 .1 âCalculation of the maximum principal stress distributions for two rivet con- figurations. Stress concentrations are evident at the upper corners of the left rivet as tiny red areas, suggesting that these might become failure sites. SOURCE: Courtesy of Mercury Marine.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 29 Nanotechnology Initiative, and the MGI,3 metals research during the past decade has achieved numerous advances. (Although this section focuses on metals, it is clear that the aforementioned programs cover all materials, and that progress has been made on all fronts.) For example, bulk structural alloys designed with self- organized nanoscale dispersoids with strengths exceeding 1.5 to 2 GPa while still retaining acceptable ductility and fracture toughness have been developed. This is a remarkable advance compared to historical practical strength limits of <1 GPa for 20th century structural materials, one that can enable significant lightweighting and cost reduction in key structural components for transportation and energy ap- plications. Similar improvements in high-strength structural alloys with adequate ductility or other targeted performance criteria have been discovered by efficient utilization of ICME approaches, leading to several examples of accelerated transi- tion from material invention/discovery to practical industrial implementation. An example of this acceleration and the impact on industries is provided in Box 2.2. High-entropy alloys, multiprincipal element alloys, or complex concentrated alloys are composed of nearly equimolar concentrations of five or more metals that form extended solid solutions.4 Since the first reported study on high-entropy alloys in 2004, research has dramatically expanded to approximately 1,000 journal publi- cations per year. These alloys can be fabricated using conventional metal fabrication techniques as single- or multiphase materials with a variety of crystal structures. The properties of these alloys are determined by a combination of the entropy (thermodynamics), sluggish diffusion (kinetics), structural distortion associated with the variation in the atomic size, and effects that are derived from the depen- dence of the properties on the composition. Several high-entropy alloys based on face-centered cubic and body-centered cubic crystal structures have been recently fabricated with mechanical properties superior to conventional alloys such as aus- tenitic stainless steel. As an example, Figure 2.2 compares the fatigue properties of different classes of metals in a plot of the fracture toughness, a measure of the ability of a material to tolerate an existing flaw such as a crack, Kc, against the yield strength of the material, Ïy, a measure of the stress at which a material will start to deform plastically.5 From this figure, it can be seen that high-entropy alloys exhibit some of the highest fracture toughness values of metals. The dashed lines in the figure represent the size of the crack tip plastic zone (=(1/Ï)(Kc/Ïy)2); this is a measure of the resistance of the material to the driving force behind propagation of a crack. A large plastic zone radius would indicate a higher resistance to crack propagation. 3 Office of Science and Technology Policy, 2011, Materials Genome Initiative for Global Competitive- ness, Washington, DC, June. 4 There is no universally agreed-upon definition of a high-entropy alloy. Some researchers describe them as having at least five elements, while others have described four-component alloys. 5 B. Gludovatz, A. Hohenwarter, D. Catoor, E.H. Chang, E.P. George, R.O. Ritchie, 2014, A fracture- resistant high-entropy alloy for cryogenic applications, Science 345:1153-1158.
30 F r o n t i e r s o f M at e r i a l s R e s e a r c h BOX 2.2 Accelerating the Process of Alloy Development Through Computation Ferrium S53 is an ultra-high-strength and corrosion-resistant steel that eliminates toxic cadmium plating, and Ferrium M54 is an upgrade from legacy alloys. Figure 2.2.1 demonstrates the time from development to deployment from 8 years for Ferrium S53 and 4 years for Ferrium M54 steels. Ferrium S53 is flying as safety-critical landing gear on U.S. Air Force A-10, T038, C-5, and KC-135 and on numerous SpaceX rocket flight-critical components. Ferrium M54 steel has been deployed on the U.S. Navyâs T-45 safety-critical hook shank component, offering more than twice the life of the incumbent steel while saving $3 million in overall program costs. Technology readiness levels (TRLs) are as follows: TRL 1 is basic principles observed; TRL 2 is technology concept formulated; TRL 3 is characteristic proof of concept; TRL 4 is validation in laboratory; TRL 5 is component validation in relevant environment; TRL 6 is system demonstra- tion in relevant environment; and TRL 7 is prototype demonstration in operational environment. FI GURE 2.2.1 âThe timeline (from 2000 to 2014) for the development of two alloys, Ferrium S53 steel (8 years) and Ferrium M54 steel (4 years), licensing them to a U.S. steel producer and deploying into demanding applications. The accelerated timeline in the improvement of these alloys and their deployment in applications was enabled by applying computational methods to optimize the composition and structure to achieve the desired properties. The left vertical axis indicates the different technology readiness levels (TRLs), and the right vertical axis indicates some of the important milestones. NOTE: MMPDS, metallic materials proper- ties development and standardization. SOURCE: W. Xiong and G. Olson, 2015, Integrated computational materials design for high-performance alloys, MRS Bulletin 40:1035-1044, doi:10.1557/mrs.2015.273, reproduced with permission.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 31 FIGURE 2.2â Comparison of the fracture toughness and yield strength of different materials show- ing the improvement in properties of high-entropy alloys. The dashed lines represent the radius of the crack tip plastic zone, which is equal to (1/Ï)(Kc/Ïy)2. This parameter gives a measure of the resistance of the material against the forces driving the advance of a crack. The larger the plastic zone size, the higher the resistance to crack advance. In other words, more energy will need to be provided to advance the crack, as some is dissipated by plastic deformation. SOURCE: From B. Gludovatz, A. Hohenwarter, D. Catoor, E.H. Chang, E.P. George, R.O. Ritchie, 2014, A fracture-resistant high-entropy alloy for cryogenic applications, Science 345:1153-1158, reprinted with permission from AAAS. 2.1.2 Bulk Metallic Glasses Similarly, evolutionary advances achieved with bulk metallic glasses during the past decade have fostered their progression from a scientific curiosity into a variety of specialized commercial products. Bulk metallic glasses offer near net shape6 formability, potential improvements in corrosion resistance, and useful strength and fracture toughness. The key scientific advance enabling the fabrica- tion of commercial bulk metallic glasses is associated with improved understand- ing of the atomic compositions that suppress crystallization during cooling (while also providing attractive mechanical properties). Since the maximum sample size 6 Near net shape is an industrial manufacturing technique. The name implies that the initial produc- tion of the item is very close to the final (net) shape, reducing the need for surface finishing.
32 F r o n t i e r s o f M at e r i a l s R e s e a r c h varies inversely with the critical cooling rate for crystallization, identification of atomic compositions with slow cooling rates for crystallization allow relatively large amorphous structures to be fabricated. Of the numerous metallic alloys that can be fabricated with an amorphous (glassy) structure by rapid quenching from the mol- ten state, more than 1,000 have been identified that can quench in the amorphous structure even for relatively slow cooling rates, thereby enabling bulk (>1 to 10 cm thick) amorphous structures to be fabricated from a variety of alloy compositions. 2.1.3 High-Performance Alloys A suite of specialized high-performance alloys with demonstrated significantly improved resistance to property degradation during exposure to extreme environ- ments such as nuclear reactor irradiation, high temperatures, high applied stresses, and high strain rates have emerged during the past decade. For example, improve- ments in the processing conditions of oxide-dispersion-strengthened Â erritic steels f (enabled by atomistic modeling and nanoscale structural and chemical charac- terization) has led to the creation of several families of new alloys with ultrahigh densities (~1024/m3) of nanoscale dispersoids. These new self-organized nano- structured alloys have room-temperature tensile strengths exceeding 1 to 2 GPa (greater than 50 percent stronger than conventional oxide-dispersion-strengthened or precipitation-strengthened steels), 10,000 hour thermal creep strengths exceed- ing 100 MPa for temperatures as high as 800Â°C (a nearly 200Â°C increase in upper use temperature compared to conventional ferritic steels), and unprecedented resistance to void swelling associated with neutron damage up to damage levels exceeding 500 displacements per atom (more than double the maximum allowable radiation damage level in conventional ferritic steels). These advances in high- performance, radiation-resistant structural materials provide the scientific basis for deploying new âGeneration IVâ nuclear reactors that offer significant potential improvements in fuel utilization, safety, electricity-generation economics, and reduced radioactive waste compared to current reactors. Controlling the size of the grain is a traditional approach to enhancing the properties of metallic alloys, with impressive improvements in room-temperature properties achieved for nanocrystalline grain sizes. Since most nanocrystalline grains are susceptible to dramatic coarsening during prolonged operation at room temperature or higher, research during the past decade has been instrumental in identifying the roles of composition and structure, such as grain boundary junc- tions, on modifying grain growth and accompanying mechanical properties. The complexity of the structure of the grain boundary was shown to be dependent on the presence of impurities, temperature, pressure, and chemical potential. The
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 33 interfacial phases formed have been termed complexions to distinguish them from bulk phases. It has been found that the impurities on the grain boundaries reside in specific sites, lower the interfacial energy, and, depending on the boundary character, can be periodically distributed along it. In the past decade, alloys designed with stabilized interfaces and married with processing paradigms to produce nanostructured material at scale were transi- tioned to successful commercialization applications. For example, electrodeposited nickel alloys stabilized with tungsten have been qualified for a variety of functional coating applications, including as a finish component in high-performance electri- cal connectors, where the alloy has been fielded on billions of components. Recent efforts at combining these nanostructure stabilization concepts in the domain of powder metallurgy have led to bulk sintered nanocrystalline components for use in construction tools and lightweighting applications. The shift in materials used in construction to lightweight alloys such as alumi- num and magnesium, development of high-strength low-alloy steels,7 implementa- tion of composite structural materials, and combinations thereof have introduced new concerns regarding corrosion. Environmental concerns over the use of toxic corrosion inhibitors, such as hexavalent chromium and cadmium, have necessitated the development of alternative coating materials and processes to prevent corrosion. Existing low-cost aqueous-based deposition processes such as immersion coatings and electrodeposition have led to new processes, such as zirconium-based conver- sion coatings and zinc-based alloys, to replace hexavalent chromium and cadmium, respectively. High- and low-temperature solid-particle spraying technologies have been developed and implemented that provide a means to create compounds and alloys with chemistry tailored to the specific corrosion environments. Development of rare-earth oxide corrosion inhibitors, such as cerium oxide or praseodymium oxide,8 for paints and primers has found application for military aircraft as envi- ronmentally benign replacements for carcinogenic compounds. Improvements in protective coatings for oil and gas pipelines, as well as bridges and buildings, have reduced the impact of corrosion under widely varying climatic conditions. 7 High-strength, low-alloy steel is a type of alloy steel that provides better mechanical properties or greater resistance to corrosion than carbon steel. 8 The rare-earth corrosion inhibitors are preferred over chromate-based inhibitors. The use of the latter is subject to regulation. Neither cerium nor praseodymium are listed as critical rare earths.
34 F r o n t i e r s o f M at e r i a l s R e s e a r c h 2.2 CERAMICS, GLASSES, COMPOSITES, AND HYBRID MATERIALS 2.2.1 Ceramics Ceramics and glasses have found high-value applications owing to their abili- ties to withstand harsh environments; as bulk, composite, and coating materials; and for their functionality in devices. Nonoxide ceramics, in particular, have prop- erties that are amenable to applications in harsh environments, ranging from high temperatures to tribological conditions to high impacts. These materials have been used as diesel filters, thermal barrier coatings, armor, turbine engine components, and refractories. Significant advances in the past decade have provided insights into the structure of these materials, improvements in processing, and ability to tailor the chemistries to achieve new properties. In addition, the past decade has seen the development of novel processes that achieve materials with increased durability by creating silica-rich layers that protect the material from oxidation. These ap- proaches allow the material to be used in new applications. This approach yielded a material with improved oxidation resistance. ZrB2/SiC could be considered as a matrix resin for composites, oxidation protection coatings, or possibly raw materi- als for additive manufacturing (AM) approaches Nonoxide ceramic materials also found application as oxidation reduction re- action cathodes in polymer electrolyte fuel cells or Li-air, Zn-air, and Al-air battery chemistries. Over the past decade, several materials have demonstrated improved fuel cell performance, including TiN, TiC, and TiB2. Further work to better under- stand how these materials achieve these improvements is ongoing, and the exact mechanism remains unclear. Extension of single-phase nonoxide materials to biphase materials has recently been pushed to the limit with the fabrication and characterization of high-entropy metal diborides such as (Hf0.2Zr0.2Ta0.2Nb0.2Ti0.2)B2. These materials offer promise as ultra-high-temperature ceramics, and the hardness and oxidation resistance of these materials appear to be greater than any of the single diborides. However, manufacturing of such five-component metal diborides may be difficult even though they consist of predominantly one solid-solution AlB2 type structure. Nonoxide ceramics have also found important uses as piezoelectrics for high- temperature sensors in automobiles, aircraft, and furnace and reactor monitoring systems, and as bulk acoustic resonators (necessary for filters in every cell phone). A typical high-temperature piezoelectric is aluminum nitride (AlN), which is both stiff and compatible with complementary metal-oxide semiconductor (CMOS) technology. However, its piezoelectric coefficient, d33, is exceedingly low, as is the low intrinsic electromechanical coupling factor. Scandium doping of AlN, motivated by the identification of metastable ScN structures from first-principle
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 35 calculations, demonstrates increases in d33 by a factor of 4,9 and coupling factors up to 12 percent.10 This significant advancement has now made its way to the corporate sector, to Murata, Taiyo Yuden, Qualcomm, Broadcom, and so on. 2.2.2 Glasses Advances in glasses over the past decade include a combination of expanded compositional options with next-generation wafer processing. Refractive indices and dispersion curves are customized by composition choices (e.g., germanium additions). Rare earths incorporated into chalcogenide glasses result in increased photoluminescence for optical amplifiers at 1300 and 1500 nm telecommunications windows.11 Fibers made from these materialsâfor example, Ga17Ge25As8.3S65 + 0.05% Pr3+ or Nd3+âhave resonance excitation from the rare earth, but also from the light absorption in the host glass that is then transferred to the rare earth. Wafer-scale processing of these materials to make devices is now possible. Binary volumetric diffractive optical elements are synthesized on a fused silica substrate with antireflective coatings below and above the Ge33As12Se55 (AMTIR-1) layer. Photopatterning has been improved to yield very fine feature control in beam shapers,12 beam splitters, or high-pass dichroic filters with adjustable parameters. Cut-off frequencies can be tuned from a few tens of picometers up to 10 or more nanometers. There is interest in fabricating photonic devices on stretchable substrates for applications such as epidermal sensing, strain-optical tuning, or aberration-free imaging. Depositing fairly low thermal expansion chalcogenide glasses on higher thermal expansion stretchable substrates, such as polydimethylsiloxane (PDMS) elastomers, generally produces high strain rectified by depositing the photonic device in a meandering or serpentine layout such that it can expand without frac- turing (Figure 2.3). Integration of graphene into wafer-scale device manufacturing is achieved by keeping the graphene-coated Si at room temperature during thermal deposition â9 A. Morito, T. Kamohara, K. Kano, A. Teshigahara, Y. Takeuchi, and N. Kawahara, 2009, Enhance- ment of piezoelectric response in scandium aluminum nitride alloy thin films prepared by dual reac- tive cosputtering, Advanced Materials 21(5):593-596. 10 M. Schneider, M. DeMiguel-Ramos, A.J. Flewitt, E. Iborra, and U. Schmid, 2017, Scandium aluminium nitride-based film bulk acoustic resonators, Multidisciplinary Digital Publishing Institute Proceedings 1(4):305. 11 ZBLAN glass fibers are also important for their transparency in the mid-infrared and potential uses in optical transport, doped-fiber amplifiers and lasers, and nonlinear optics. See Made in Space, âFiber Optics: Best in Class Fluoride-Based Fiber for Medical, Telecom, and Research,â http://made- inspace.us/mis-fiber/, accessed July 5, 2018. 12 From a few tens of picometers up to 10 or more nanometers.
36 F r o n t i e r s o f M at e r i a l s R e s e a r c h FIGURE 2.3â Fabrication process of the stretchable photonic devices. The steps are initial coating with polydimethylsiloxane (PDMS) and the epoxy-based photo resist (SU-8), fabrication of the glass device using serpentine layout, E-beam patterning, SU-8 patterning, removal of excess material, and delamination so that the stretchable PDMS layer becomes the substrate for the device. SOURCE: H. Lin, L. Li, Y. Huang, J. Li, C. Ramos, L. Vivien, J. Lonergan, K.A. Richardson, and J. Hu, 2016, âMonolithic High-Index-Contrast Stretchable Photonics,â Paper FF5F.2 in Frontiers in Optics 2016, OSA Technical Digest (online); used with permission. of the chalcogenide glass films.13 Patterning is achieved through fluorine plasma etching or liftoff. Microresonator devices in both near- and midwave infrared wave bands have been demonstrated without optical loss from the graphene layer. Integration of other 2D layers (WS2, MoTe2) has been demonstrated, potentially leading to photonic integration on 2D transparent conductor materials. Newer applications for chalcogenide glasses include natural gas sensing and superionic conduction for lithium- or silver-conducting solid electrolytes. 2.2.3 Composites and Hybrids Composite and hybrid materials have seen increased applications over the past decade, building on the foundation that investments in polymer-matrix compos- ite have achieved in broader understanding of these highly engineered materials. Monolithic materials can be easily shaped into engineering components via casting, forging, injection molding, machining, and a myriad of manufacturing processes, and historically the use of monolithic materials has far outpaced the use of hybrid materials in man-made engineering systems. That is changing with ever-Âincreasing demands for performance and design freedoms. Fiber- and particulate-based 13 U.S. Patent US20150206748A1, âGraphene Layer Formation at Low Substrate Temperature on a Metal and Carbon-Based Substrate,â Current Assignee University of Chicago Argonne, LLC.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 37 composites that offer a unique balance of properties have gained increased utility in the past half century, and the past decade in particular has seen maturation of composite technologies into the wholescale production of numerous products, in- cluding but not limited to bulletproof vests, massive wind turbines, and commercial airliners. Recent advances in the development and use of multicomponent hybrid material systems are well-documented. For example, during the past decade, the increase of carbon-fiber composites for structural components in aircraft increased from less than 25 percent in 2010 to around 50 percent at the time this report was prepared. The development and impact of ceramic matrix composites for applica- tion in aircraft engines and gas turbines is highlighted in Box 2.3. Much of this advance was the result of sustained support of fundamental and applied research in this area, along with a close cooperation between researchers and gas turbine manufacturers. Although the concept of composite and hybrid materials has a long legacy, the era of this class of materials took off with the investment in carbon-fiber poly- meric composites for aerospace applications. The ability to reduce weight through the materialâs high specific strength and modulus, the ability to form previously Â challenging complex shapes via the buildup of layers versus the hogging out of ingots, and the ability to place the structural properties where needed through the design of the built-up compositeâa crude form of AMâushered in significant investments that rapidly matured this class of materials from concept to military applications in recent decades. Today it is the basis of the Boeing 787, enabling a lower fuel burn and more comfortable passenger experience airplane, and the BMW i3 and i8, enabling battery-powered engines. Today, there is a strong industrial base that provides carbon fiber and carbon- fiber composites globally for a rapidly diversifying market. The commercial suc- cess of this specific class of composite materials has provided a foundation of new design, analysis, modeling, and manufacturing technology for the United States, and it has generated growth across the globe. Oil and gas and deep-sea structures see increasing use of composite pipes and structures; lightweight infrastructure and shelter applications have emerged; and transportation modes of all types are invest- ing in use of these tailorable materials. Technical advances over the past decade in processing science and manufacturing have enabled new fabrication approaches.14 Novel multidirectional weaving of reinforcements has been demonstrated, as have nanotailored composites. Improved fibers, polymers, and additives have been dem- onstrated in various fields of use. Processing science models that can simulate the creation of the material in its final component form have been developed in the past decade and show great promise in enabling the simulation of processing routes be- fore investing in laboratory or industrial fabrication. These advances must continue 14 R.M. Jones, 2014, Mechanics of Composite Materials, CRC Press.
38 F r o n t i e r s o f M at e r i a l s R e s e a r c h BOX 2.3 Development of Ceramic-Matrix Composites for Aircraft and Stationary Gas Turbines Increasing the operating temperatures by 100Â°C to 200Â°C and reducing the component weight results in an increase upward of 2 percent in efficiency, and consequently, fuel saving. The weight reduction arises because the density of ceramic-matrix composites (CMCs) is one- third that of the superalloys used in current turbine engines. Achieving these goals required development of CMCs through advances in both materials and their processing. The composite requires a high-strength, refractory fiber that can sustain thermal excursions. High-strength, creep-resistant, oxygen-free SiC fibers such as Hi-Nicalon Type S from NGS Advanced Fiber, provide these properties. Unlike the load transfer from matrix to fiber that is characteristic of polymer-matrix and metal-matrix composites, CMCs require a weak, low-toughness, fiber-matrix interface, which ensures that matrix cracks do not propagate through the fibers, thus ensuring their load-carrying function. This issue was addressed through the development of a multilayer, boron nitride-based coating that provides the adequate low toughness and, equally important, sustains the melt-infiltration (see Figure 2.3.1). Also equally important, the coating sustains the melt-infiltration composite-fabrication pro- cess without damage. Another essential development was the chemical-vapor deposition coating technique that was scaled to continuous fiber lengths. This ensured that the fiber coatings would remain independent. Last, the composite was produced by a melt-infiltration process, in which a coated fiber array is impregnated with a slurry of SiC and carbon and melt-infiltrated with silicon to produce a dense SiC matrix reinforced with SiC fibers. The development of a CMC suitable for aviation took more than 25 years of development, which highlights the challenge of translating materials development at the laboratory scale to product development. However, through federal agency support from the Department of Energy and other agencies, and companies such as General Electric, a new generation of jet engines employing CMCs was developed. The impact is that when CMCs are fully implemented in jet engines, fuel consumption will be reduced by 1.5 percent, which translates into a savings of more than $1 million per aircraft over a 10-year period.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 39 FIGURE 2.3 .1 âTop: Microstructure showing the ceramic matrix, the silicon carbide (SiC) fibers with a boron nitrideâbased coating. Bottom: CMC Stage 1 shrouds for the leading-edge avia- tion propulsion (LEAP) engine. There are 18 of these components per LEAP engine. SOURCE: Reprinted from G.S. Corman and K.L. Luthra, 2017, âDevelopment History of GEâs Prepreg Melt Infiltrated Ceramic Matrix Composites Material and Applications,â in Comprehensive Composite Materials II, Elsevier, New York, with permission from Elsevier.
40 F r o n t i e r s o f M at e r i a l s R e s e a r c h and be extended to capture a greater subset of the composites and hybrids that are being developed. In recognition of the critical need to be able to quickly model and downselect promising microstructures from among the multiple possible combi- nations of numerous constituents, multiscale modeling efforts have emerged, with some success in capturing the complex formations of composites and in particular the thermosetting composites that include chemical reactions. Early applications of the ICME in composites and hybrids have occurred in the past decade,15 with, for example, the development of a hybrid-disk concept for jet engine applications that would benefit from increased compressor exit temperature or increased time at elevated temperature for both the compressor and the high-pressure turbine.16 Unlike the composite strategies used for stiffening and strengthening in car- bon-fiber composites, advances in ceramic fiber-reinforced ceramics provide a solution to the traditional brittleness associated with ceramics coupled with greater refractoriness and reduced weight compared to superalloys. Ceramic-matrix com- posites (CMCs) were successfully employed, for the first time, in General Electricâs leading-edge aviation propulsion (LEAP) engine in 2017 (see Box 2.3). Driving the development and implementation for both aviation and stationary gas turbines is the scaling of engine efficiency with gas temperature in the engine. The major increase in material temperature afforded by CMCs compared to Ni-based superal- loys was achievable, through not only material selection, but also fundamentally new design and analysis approaches, as well as by innovative processing methods. Figure 2.4 shows the increase in the operating temperature with the development of Ni-based superalloys, and the increase in operating temperature achieved through the introduction of thermal barrier coatings (TBCs), without and with cooling.17 Over the period from 2010 to 2015, the increase in operating temperature achieved by alloy development was small and increased from 1120Â°C to 1140Â°C; with ther- mal barrier coatings, it increased from 1323Â°C to 1365Â°C; and with thermal bar- rier coatings with cooling, it increased from 1481Â°C to 1527Â°C. With thermal and environmental barriers with cooling, the temperature of CMCs will increase the operating temperature to more than 1620Â°C, increasing the operating temperature by about 100 degrees. Thermal protection systems are enabling space exploration. An example is the Solar Probe Plus satellite, which will venture closer to the Sun than any man-made object has ever gone (see Box 2.4). 15 B. Cowles, D. Backman, and R. Dutton, 2012, Verification and validation of ICME methods and models for aerospace applications, Integrating Materials and Manufacturing Innovation 1(1):2. 16 See B.A. Cowles and D. Backman, 2018, âAdvancement and Implementation of Integrated Com- putational Materials Engineering (ICME) for Aerospace Applications,â Preprint, http://www.dtic.mil/ dtic/tr/fulltext/u2/a529049.pdf, accessed August 8, 2018. 17 N.P. Padture, 2016, Advanced structural ceramics in aerospace propulsion, Nature Materials 15:804-809.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 41 FIGURE 2.4â The progression and projection of temperature capabilities of nickel (Ni)-based superal- loy (gray) with thermal barrier coatings (TBCs; green), ceramic-matrix composites (CMCs; blue), and maximum allowable gas temperatures with cooling (red). SOURCE: Reprinted by permission from Springer Nature: N.P. Padture, 2016, Advanced structural ceramics in aerospace propulsion, Nature Materials 15:804-809, Â© 2016. Advances in coating technologies have also engendered significant advances and increased reliability and use of multilayered wear, thermal, and environmental protection systems. Layered material systems are replacing advanced monolithic materials in a growing number of applications where the unique properties and functionality of each layer provides dramatically increased performance and life. TBCs for metallic Ni-base superalloys and environmental barrier coatings for CMCs are two examples of layered material systems that have been developed for extreme environments, consisting of high stresses and oxidizing and corrosive gases at temperatures on the order of 1500Â°C. These chemically, mechanically, and physi- cally diverse layers interact and evolve throughout the lifetime of the coating, and important phenomena affecting performance and durability occur in each layer and particularly at the interfaces between disparate materials. The successful implemen- tation of strain tolerance and low conductivity coatings have dramatically increased coating lifetimes. The use of layered materials extends across material class.
42 F r o n t i e r s o f M at e r i a l s R e s e a r c h BOX 2.4 Thermal Protection System The Parker Solar (Figure 2.4.1.) Probe will fly within 10 solar radii of the Sun, closer to the Sun than any other man-made spacecraft, in order to study the origins of the solar wind. The mission is enabled by development of a thermal protection system (TPS) that allows the spacecraft bus to function at about 30Â°C, while the shield takes the brunt of the Sun and withstands 1300Â°C on the front surface. The TPS is a multimaterial sandwich panel that opti- mizes the optical performance, mechanical integrity, and thermal insulation performance. The development of the TPS by the Johns Hopkins Applied Physics Laboratory required a success- ful teaming approach with Johns Hopkins University, NASA, government test facilities, and industry partners across the country. Consisting of a carbon-carbon and carbon-foam sandwich panel with a specially designed alumina coating, the 2 m diameter, 70 kg heat shield utilizes structurally integrated insulation to withstand launch loads and the extreme temperatures of the mission. This novel hybrid technology required development of new materials, utilization of high-temperature materials in new ways, and the elaboration of new testing approaches and methods. Owing to the novelty of the approach, extensive materials testing, manufacturing process development, structural analysis, thermal analysis, and testing were required to bring about a successful TPS in a little over 4 years. FIGURE 2.4 .1 âArtistâs concept of the Parker Solar Probe Plus with hybrid thermal protection system. SOURCE: NASA, âNASA Selects Science Investigations for Solar Probe Plus,â https:// www.nasa.gov/topics/solarsystem/sunearthsystem/main/solarprobeplus.html; courtesy of Johns Hopkins University Applied Physics Laboratory.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 43 The use of polymeric coatings for environmental protection has expanded in the past decade. For example, the development of high-surface-tension polymeric films with highly controlled chemistry and nanoscale features has fueled the use of tunable superhydrophobic and ice-phobic coatings that now protect microelectron- ics, solar cells, wind turbines, and airplane wings. 2.3 SEMICONDUCTORS AND OTHER ELECTRONIC MATERIALS Semiconductors are the workhorse materials for electronic and photonic device applicationsâmaking up integrated circuits, circuit boards, and light emitters, and incorporated into packaging materials, displays, and any number of controlling and monitoring devices.18 This section discusses some of the principal develop- ments in semiconductors and other electronic materials that have allowed conÂ tinued, significant advances in modern electronics and photonics. Like many of the mateÂ ials discussed in other sections of this chapter, the discovery paths followed r in electronic materials have, in many cases, been influenced or even been directed by the industrial environment in which these materials discoveries participate. This section will begin by describing some of those impacts, and then discuss how materials research (MR) has responded to these far-reaching factors. The final subsections of Section 2.3 will discuss the significant impacts that semiconductor research has made in optoelectronics and some of the developments in organic and flexible electronics. 2.3.1 Materials and Devices for Information Technology Integrated circuits have transformed information and computational tech- nologies. This revolution was enabled by five decades of improvements in the performance of silicon field-effect transistors (FETs) and has been characterized by exponentially compounded progress in the miniaturization of electronic devices and circuits. However, the past 20 years have encountered limits that have affected efforts at further miniaturization. The 1990s saw devices reach sizes where gate insulator thickness and operating voltage became less effective avenues for further device scaling. In the early 2000s, heat removal and power density considerations contributed to the plateauing of microprocessor clock frequencies. In the past decade, relevant feature sizes have reached the single digits of nanometers, where fundamental physics is imposing increasingly rigorous constraints. The responses by those in research and development (R&D) communities have been several-fold. Continued efforts to miniaturize have provided incremental 18 The committee thanks Thomas Theis, IBM (Emeritus), for his insightful thoughts and ideas in developing this portion of the report.
44 F r o n t i e r s o f M at e r i a l s R e s e a r c h gains. Parallel with those efforts have been concerted attempts to find alternative approaches to satisfy the needs of traditional information and computational tech- nology needs. These efforts have been seeded by a number of efforts jointly funded and directed by the public and private sectors (see Box 2.5). Last, another major push in electronic materials over the past 10 years, heavily entwined, at times, with the other efforts, has been to develop new materials that will be potentially useful for totally new computation and information capabilities and needs. BOX 2.5 Industrial/Government/Academic Joint Pursuits of New Solutions for Information Technologies Our ongoing information technology revolution has been enabled by six decades of exponentially compounding progress in the miniaturization of electronic devices and circuits. Of course, all exponential trends must eventually end, and there is mounting evidence that progress in miniaturization is slowing.1 Nevertheless, ongoing research supports an optimistic outlook for further advances in the hardware that enables information technology. The often- predicted, long-delayed, and gradual âend of Mooreâs lawâ is setting the stage for exciting new developments in information technology as the focus of research and development shifts from miniaturization of long-established devices to the coordinated introduction of new devices, new integration technologies, and new architectures for computing.2 The impetus for this shift in research investment goes back to the period 2003-2005, when microprocessor clock frequencies, constrained by considerations of heat removal and power density, suddenly plateaued.3 Aware that these constraints could not be fully addressed by continued incremental advancement of established technologies, representatives of several microelectronics manufacturers began to discuss a research program with longer-term goals. As a result, the Nanoelectronics Research Initiative (NRI) was chartered in 2005 by a group of Semiconductor Research Corporation (SRC) member companies to develop and administer a university-based research program aimed at entirely new devices and circuits for computing. In partnership with the National Science Foundation (NSF), NRI would fund university research to âdemonstrate novel computing devices capable of replacing the CMOS FET as a logic switch in the 2020 time frame.â In 2007, the National Institute of Standards and Technology joined the private-public partnership, resulting in the creation of four multiuniversity, multidisciplinary research centers. NRIâs bold and clearly articulated research goals caught the attention of fund- ing agencies in Europe and Asia and helped to spark new initiatives in those geographies. In 2013, the Defense Advanced Research Projects Agency (DARPA) joined with industry to fund STARnet, expanding the program and focusing U.S. university researchers on the exploration of post-CMOS devices. As the NRI and STARnet programs evolved, the interest of the industrial sponsors shifted from exploration of isolated devices toward co-development of new devices, circuits, and architectures. This was made explicit with new programs announced by SRC and NSF in 2016 and by SRC and DARPA in 2017. 1 T.N. Theis and H.-S.P. Wong, 2017, The end of Mooreâs law: a new beginning for informa- tion technology, Computing in Science and Engineering 19:41. 2 See https://eps.ieee.org/images/files/Roadmap/SIA-SRC-Vision-Report-3.30.17.pdf. 3 T.N. Theis and P.M. Solomon, 2010, Quest of the ânext switchâ: Prospects for greatly Â educed power dissipation in a successor to the silicon field-effect transistor, Proceedings of r the IEEE 98:2005. SOURCE: Courtesy of Tom Theis, IBM (Emeritus).
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 45 The importance of the interplay between device and MR exists in all of the examples discussed in this section. In some cases, MR is focused on material properties critical to the performance of an established device concept. In other cases, adventurous MR is inspiring new device concepts and making the impossible suddenly possible. In all cases, much more remains to be done. 2.3.2 Continued Miniaturization of Silicon-Based Field-Effect Devices Over the past decade, materials and process innovation made significant con- tributions to efforts to further miniaturize silicon-based field-effect devices. Se- lected materials research highlights include the following: â¢ New low-k dielectric films based on chemical vapor deposition (CVD, or similar) of âSiOCâ precursors. Initial advances were made by lower- ing the integrated dielectric of the film based on precursor design, then by introducing engineered porosity or air gaps into the integrated film. ultraviolet-Âcuring has been a critical ally to the successful integration of such fragile, low-k strategies. New capping or sealing layers protected the fragile dielectrics (especially films with porosity) from damage during subsequent processing steps or device operation. â¢ A wide array of new liners, barrier stacks, and capping agents have improved metallization, resistive-capacitive (RC) delay, and electromigration, espe- cially as feature sizes have been reduced. These include but are not limited to Cu, Ta, TaN, TiN, CuMn, MnN, Ru, W, and Co. â¢ A metal-insulator-metal capacitor, introduced to improve the voltage linear- ity properties across a chip. â¢ Industryâs migration to new gate dielectric materials with higher dielectric permittivity, primarily based on hafnium oxide, that greatly improved gate leakage at the <45 nm node. New metal gate electrodes for n- and p-type metal-on-semiconductor, atomic layer deposition (ALD), and chemical m Â echanical polishing (were all instrumental to realizing this transforma- tive advance. Last, in one of the most publicized innovations seen in the past decade, the semiconductor industry has moved to 3D FET called FinFETs (Figure 2.5). The 3D FinFET improved transistor performance by forming a conducting channel on all three sides of the vertical fin structure, improving power when compared to a traditional planar transistor at the same nominal dimension. Miniaturization performance gains over the past 10 years have also come through scaling driven by optical and extreme ultraviolet (EUV) lithography. Early in the past decade, traditional gains were provided by advances in optical
46 F r o n t i e r s o f M at e r i a l s R e s e a r c h FIGURE 2.5âTo sustain the silicon revolution, devices have pushed to 3D over the past decade. Devices such as IBMâs 5 nm node nanosheet transistor have achieved the highest layout efficiency. SOURCE: A. Tilley, 2017, âIBM Shows the World How to Build a Super Dense 5-Nanometer Chip,â Forbes, June 5, https://www.forbes.com/sites/aarontilley/2017/06/05/ibm-5nm-chips/#3ae8462d3c56; courtesy of IBM. lithography. For example, the industry moved from 193 nm dry to 193 nm immer- sion lithography by introducing water as an immersion fluid between the final lens element of the exposure tool and the photoresist.19 While providing a significant resolution enhancement (~40 percent), the immersion fluid caused leaching of photoresist components and defects arising from complex fluid dynamics of the high-speed exposure tool. In most cases, these problems were addressed by devel- oping materials for use as a spin-on topcoat to physically separate the photoresist and immersion media. As 193 nm lithography faced ultimate resolution issues (<90 nm pitch), indus- try turned to multipass patterning. Initially straightforward, âlitho-etch-litho-etchâ cycles generated the desired pattern by stitching together separate passes. The cost and complexity of this approach quickly exploded (mask design and count, feature variation, etc.), especially as it was applied to more complex situations. These in- cluded strategies such as self-aligned double patterning or other variants based on CVD20 or ALD, providing tight layer-to-layer alignment, and simplifying integrated fabrication flows. ALD also allowed for the integration of new high-k dielectrics and metal gates. New CVD and ALD precursors, thin films of various oxides and 19 Early this decade, a consensus formed that for 193 nm immersion lithography (193i), water would be the most likely candidate for 65 nm and 45 nm device nodes, bridging the gap between âdryâ or conventional 193 nm lithography and EUV lithography. See, for example, Optical Micro- lithography XVIII, 2005, edited by B.W. Smith, Proceedings of SPIE, 5754, SPIE, Bellingham, Wash., doi: 10.1117/12.600025. 20 CVD is chemical vapor deposition and is all types of depositions relying on chemical reactions. ALD is atomic layer deposition, a subset of CVD that used alternatively pulsed gas sources to prog- ress growth layer by layer. MBE is molecular beam epitaxy, which takes place in high vacuum with no carrier gas.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 47 nitrides, and processes that took a âbottom-up approachâ were critical for these efforts toward the continued miniaturization of electronics. Industryâs ultimate âtop-downâ 2D scaling strategy is extreme ultraviolet lithog- raphy, or EUVL,21 which was introduced in the past decade. With a wavelength of 13.5 nm (or 91.8 eV), this lithographic technique is a radical departure from previous optical approaches. As all matter absorbs radiation at this wavelength, EUVL exposures occur in a vacuum chamber using reflective rather than transmis- sive optics. Photoresist outgassing during EUVL exposure serves as an additional performance constraint that the patterning film must consider, as resist by-Âproducts will damage the costly lens elements. Second, unlike UV-based photoresists de- signed in an energy space well understood by chemists, EUV photons create a complex cascade of photoelectrons and secondary electrons that transact the ex- posed material in poorly understood ways. As such, stochastic EUVL simulations have been important for understanding the process margin and design rules for features patterned via this high-energy technique. All of these EUVL advances have benefited and will continue to benefit tremen- dously from the synchrotron research and light sources discussed in Chapter 4. For example, a Lawrence Berkeley National Laboratory EUV patterning tool was used to characterize and study more than 12,000 material systems, an EUV photomask microscope was used to prove the existence of defects visible to only EUV light, and an EUV scatterometer was instrumental in the development of sub-nm wavefronts needed for further development of novel photomask architectures and materials. 2.3.3 Alternative Solutions to Address the Bottlenecks in Field-Effect Transistors Low-Voltage, Low-Power Devices Ongoing research continues to provide possibilities for profound advances in the capability of information technologies, beyond continued miniaturization. One (of many) such examples is the negative capacitance field-effect transistor (NCFET), sometimes also called the ferroelectric field-effect transistor. The promis- ing 2008 theoretical proposal was directly motivated and funded by the SRCâs NRI program described in Box 2.5. It showed a new way to break the existing constraints on power and performance, but building the proposed device with known ferro- electric materials at the nm-scale dimensions required for cost-competitive digital circuits appeared impractical. The subsequent discovery of new materials exhibiting ferroelectric behavior in films as thin as a few nm ignited significant R&D activity. 21 T. Haga, 2018, The early days of R&D on EUV lithography and future expectations, Journal of Photopolymer Science and Technology 31:193.
48 F r o n t i e r s o f M at e r i a l s R e s e a r c h It is still too soon to predict when the NCFET will be commercialized and how important it will become for information technology, but a recent demonstration of a superior power-performance trade-off relative to closely comparable conventional FET circuits is very encouraging.22 The NCFET is just one example of an emerging class of transistor-like Â evices d that switch by physical principles that are fundamentally different from the op- erating principle of the conventional FET, and can thus transcend some of the FETâs fundamental limits.23 The tunneling FET is another exemplar device that has improved rapidly in recent years, enabled by materials advances such as the controlled growth of compositionally graded semiconductor nanowires. Other compelling low-voltage, low-power device concepts are less developed. None of these will advance without further advances in materials. New Materials Yielding New Classes of Memories The past decades have seen a rapid maturation of emerging memory tech- nologies based on novel materials and new cell architectures. These include spin-, phase change-, resistive- (or memristors), and ferroelectric-based memoriesâall now commercially available. Emerging random-access memory (RAM) devices include spin-transfer torque magnetic RAM (STT-MRAM), ferroelectric RAM, conductive bridge RAM, resistive RAM, and phase-change memory.24 These are very distinct devices, each based on a different class of materials, but all share some highly desirable attributes. In particular, they can all be fabricated at rela- tively low process temperatures, enabling integration of memory devices directly above blocks of logic. This fine-grained integration of memory and logic is seen as a key to the implementation of energy-efficient logic-in-memory and memory- in-logic architectures. At the same time, each of these devices offers advantages and disadvantages compared to the others. Each will advance only with advances in the properties of its requisite materials. The opportunity for STT-MRAM in the embedded memory market as an alternative or replacement to NOR flash is significant, as it is projected to be a good solution for harsh environments, such as the growing automotive market. 22 Z. Krivokapic, U. Rana, R. Galatage, A. Razavieh, A. Aziz, J. Liu, J. Shi, H.J. Kim, R. Sporer, C. Serrao, and A. Busquet, 2017, â14 nm Ferroelectric FinFET Technology with Steep Subthreshold Slope for Ultra-Low-Power Applications,â 2017 IEEE International Electron Devices Meeting, 15.1, http://www.proceedings.com/37997.html. 23 T.N. Theis and P.M. Solomon, 2010, In quest of the ânext switchâ: Prospects for greatly reduced power dissipation in a successor to the silicon field-effect transistor, Proceedings of the IEEE 98:2005. 24 T.N. Theis and H.-S.P. Wong, 2017, The end of Mooreâs law: A new beginning for information technology, Computing in Science and Engineering 19:41.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 49 Nanophotonics and Nonlinear Optical Materials Sophisticated nanophotonic communication networks based on passive wave- guides and electro-optic phase and amplitude modulators have been demonstrated in recent years and are approaching large-scale commercialization. Frontier re- search in nanophotonics is focused on demonstration and development of tiny nonlinear optical devices based on the large nonlinear optical coefficients obtain- able with 2D materials integrated in nm-scale optical resonators.25 This research relates to plasmonics, metamaterials, and coherent optics, and is enabling new technological opportunities in optical communications including energy-efficient all-optical gates and 100 THz all-optical modulators. Such devices may allow com- putational functions to be distributed in optical networks for smart routing and management of data flow. From Two to Three Dimensions The past decade has seen a number of efforts to expand two-dimensional circuitry into three dimensions. This includes the FinFET discussed above. More generally, 3D chips are becoming the de facto standard in the market, with most major suppliers creating 3D negative-AND gate (NAND) parts having 32 or more layers.26 Those manufacturers are pushing to 128 or more layers by the year 2020 and will soon sell more 3D NAND than traditional 2D NAND. While many of the basic fabrication materials are similar to those found in 2D NAND, the high aspect ratios (30:1 or more) in 3D NAND pose significant challenges for deposi- tion, patterning, high-aspect-ratio etching, and metallization. Overcoming these challenges required the introduction of paired deposition layers of customized oxide-nitride films, the use of an amorphous carbon hardmask, and novel plasma reactors that could yield smooth, vertical profiles. ALD techniques have become more prominent in 3D NAND, as ALD processes can deliver good uniformity and coverage, even for high-aspect-ratio features. While the price per 3D NAND wafer is higher than a 2D NAND wafer, the added performance per chip arising from 3D integration retains the promise of Mooreâs law economics, as the estimated cost per GB for the 3D NAND is lower than that for the 2D NAND part. The move to 3D design and layout plays an increasingly prominent part in the roadmap for semiconductors and electronics. Another example of materials research contributing to the development of 3D designs is in the area of radio frequency (RF) transformers. The ubiquitous wireless 25 A.Autere, H. Jussila, Y. Dai, Y. Wang, H. Lipsanen, and Z. Sun, 2018, Nonlinear optics with 2D layered materials, Advanced Materials 30:1705963. 26 3D NAND is driven by interconnect layers, which is a subset of the technological requirements for 3D logic.
50 F r o n t i e r s o f M at e r i a l s R e s e a r c h connectivity of devices has required the development of new on-chip components, and one such type of component, an RF transformer, has been developed using rolled-up membrane technology. These devices have large turn ratios, large cou- pling coefficients, and high maximum working frequencies. Such 3D structures, essentially nano- and microscale tubes, have been produced by fabricating semi- conductor scrolls through the application of a strain-induced self-rolling process. Control of the inner diameter of the tube is essential for application in some de- vices. By employing finite element modeling, researchers have studied the process by which the film is released, thereby guiding the engineering of the strained layer to obtain different rolled-up structures and predicting the diameters of the final product. Scrolls of Si/SiGe, strained SiNx, and GaAs/AlGaAs have been fabricated. The tubular SiNx structures, along with accompanying prepatterned metal layers, have been used to produce a novel on-chip tube inductor design platform to pro- duce inductors for application in radio frequency integrated circuits. Control of the size of two-layer metal InGaAs scrolls, in which the metal is used as the stressor, has been achieved through lithographic patterning, as seen in Figure 2.6. Combination Devices/MEMS The integration of mechanical and electronic components within a single mi- crofabricated package is a technology known as a microelectromechanical system (MEMS). Optimized and innovative technology, in combination with the eco- nomics of high volume/low unit cost production, has expanded the application of MEMS to include devices such as sensors, actuators, micropower generators, chemical reactors, and biomedical devices. Diversification of this portfolio will continue, but the extent will be predicated on the development of appropriate materials and fabrication technologies. The complexity of MEMS devices is increasing rapidly. Inkjet print heads, one of the original silicon-based MEMS devices, transformed document printing with a heating element and an orifice. Then silicon-based devices with moving parts enabled the development and commercialization of miniature pressure sensors, accelerometers, and gyros. Further miniaturization, bulk high-volume fabrication, and improved sensor design have enabled the integration of MEMS-based sensors with numerous consumer products, where they are primarily used for monitoring and control. 2.3.4 Semiconductors for Optoelectronics The past decade saw widespread commercial adaption of lighting using light- emitting diodes (LEDs) based on IIII-nitride semiconductors, developed through sustained materials and device research efforts over the previous few decades.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 51 FIGURE 2.6âTop: Simulation of a coaxial structure. Bottom: Coaxial structure rolled up by local stress. SOURCE: Reprinted with permission from W. Huang, S. Koric, X. Yu, K.J. Hsia, and X. Li, 2014, Precision structural engineering of self-rolled-up 3D nanomembranes guided by transient quasi-static FEM modeling, Nano Letters 14(11):6293-6297, doi:10.1021/nl5026369. Copyright 2014 American Chemical Society. Further research in this area over the past decade has led to continued improve- ments in the materials required for this important application. Gallium nitride on either sapphire or SiC is commonly used for LEDs. With efficiency gains reaching a maximum for LEDs grown on sapphire, the efficiency drop, the cost of sapphire wafers, and the challenges associated with large-scale fabrication have led to an in- terest in growing GaN on substrates such as GaN, Si, and metals. In the past decade, progress has been made in developing methods to produce bulk GaN substrates. The availability of such substrates has resulted in the production of GaN-on-GaN LEDs with an equivalent light output to conventional LEDs but on a much smaller
52 F r o n t i e r s o f M at e r i a l s R e s e a r c h chip. In addition, green lasers have been produced on semipolar GaN substrates. GaN-based LEDs have been fabricated on silicon substrates. This is a challenge because of the large difference in lattice parameter and thermal expansion coef- ficient. The issue of cracking owing to the thermal stresses has been addressed by growing multilayers of AlGaN/GaN on a prepatterned substrate. This technology offers the advantage that large-diameter silicon wafers are available at low cost and are processed routinely in the electronics industry. Semiconductor quantum dots have applications in lighting technologies, dis- play resolution, drug delivery, and imaging at the molecular level. In the past Âdecade, progress has been made in several areas including but not limited to tuning Â the emission spectrum in ultrasmall CdSe quantum dots through post-synthesis treatments; correlating the optical properties of the dots to the atomic structure on an individual quantum dot; ordering of InN quantum dots on Si substrate; and ordering of InAs-based quantum dots on GaAs. 2.3.5 Organic Semiconductors Conjugated semiconducting polymer/organic materials offer in some cases opportunities for a low-cost, additive, environmentally benign, printable elecÂ tronics manufacturing ecosystem built on materials designed to be lightweight, flexible, and solution processable over large areas. These materials offer oppor- tunities for functional electronic devices for applications including organic field- effect transistors, organic light-emitting diodes (OLEDs), organic photovoltaics, batteries, biomedical devices, and sensors,27 many of which form the case studies in Section 5.3. While it is fair to say that the synthesis of novel molecules has led to significant improvements in charge carrier mobility, which is the key metric that defines elec- trical performance, the solution processing and thin film deposition of conjugated polymers must also be properly controlled to obtain high-performance devices with the requisite crystallinity, intergrain connectivity, and alignment. Precise control of the electrical performance of a material depends critically on its solid-state thin film microstructure and on the intramolecular- to the device-length scales. The organic semiconductor technology that has become a marketplace reality in the past decade is the display technology based on OLEDs.28 Significant efforts went into the design of the materials and processes that enable todayâs displaysâ see, for example, Case 1 in Section 5.3. Lighter in weight and less likely to shatter 27 N.E. Persson, P.-H. Chu, M. McBride, M. Grover, and E. Reichmanis, 2017, Nucleation, growth, and alignment of poly(3-hexylthiophene) nanofibers for high-performance OFETs, Accounts of Chemical Research 50(4):932-942. 28 Chemical & Engineering News, 2016, The rise of OLED displays, July 11, Volume 94, Issue 28.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 53 than their LCD counterparts, vivid color OLED displays are now a mainstay of smartphones. And through efforts, TVs using OLED technology are becoming available. OLED display technology is also simpler than that of LCDs: the organic dis- plays are composed of pixels that individually emit red, green, and blue light to create an image. The conjugated organic molecules are positioned between two electrodes, and when current flows from the cathode to the anode, electrons and holes combine to emit light. Since black is simply created by leaving the requisite pixels off, the âblackâ is often called âtrue blackâ and is perhaps one reason that OLED displays exhibit much sharper and brighter images. OLED technology is also more energy efficientâto achieve the color black, no energy is consumed. While challenges surrounding materials and process costs continue, as does research to identify more stable materials for longer life displays, with the demon- stration of a viable technology, other active organic materials technologies are likely to follow. One particularly promising area for OLEDs is the use of data informatics approaches to help identify promising target molecules, coupled with the develop- ment of new, more sustainable approaches to materials synthesis. 2.3.6 Flexible Electronics In the past decade, flexible electronics29 have developed beyond earlier applications in curved and flexible displays and panels to new advances in foldable, stretchable, conformable devices. Increasingly, these are being widely applied as even softer, portable, wearable sensors, especially for continuous monitoring of physiological signals. Additional applications are in devices to enhance human physiology, as in controlling motion or haptic and tactile senses, and in internal devices, such as prostheses or interventional devices, as illustrated in Figure 2.7. There are different strategies for creating flexibility while retaining electronic properties, including strain minimization via nanoscale processing of established materials and synthesis of new functional nanomaterials. The nanoscale dimen- sions of the materials dramatically decrease flexural rigidity of devices while acting as electrodes, transport channels, and lightâemitting/photonâabsorption materials. Flexible designs with locally rigid materials have been created by shaping or pat- terning the material architecture to allow larger scale global deformability. This is a rapidly evolving area of materials research. 29 J.A. Rogers, M.G. Lagally, and R.G. Nuzzo, 2011, Synthesis, assembly, and applications of semi- conductor nanomembranes, Nature 477:45.
54 F r o n t i e r s o f M at e r i a l s R e s e a r c h FIGURE 2.7â Illustrations of uses of flexible electronics under development in physiological moni- toring and medicine. SOURCE: Reprinted with permission from Y. Liu, M. Pharr, and G.A. Salvatore, 2017, Lab-on-skin: A review of flexible and stretchable electronics for wearable health monitoring, ACS Nano 11(10):9614-9635, doi: 10.1021/acsnano.7b04898. Copyright 2017 American Chemical Society. 2.4 QUANTUM MATERIALS AND STRONGLY CORRELATED SYSTEMS 2.4.1 Superconductors and Strongly Correlated Electrons Superconductivity30 was serendipitously discovered in 1911, and the micro- scopic theory (Cooper pairing of electrons in Bardeen-Cooper-Schrieffer theory) was articulated in 1957. Applications for superconductors arise in their unique transport and quantum-mechanical properties: production of large magnetic fields (e.g., high-field research, magnetic resonance imaging, supercolliders); detection of small magnetic fields (e.g., superconducting quantum interference devices), high- frequency detection (e.g., radio astronomy); and energy transmission and produc- tion (e.g., power grid and turbines). Further, dozens of families of unconventional superconductors have been discovered; see Figure 2.8 for results, including the 30 J. Sarrao, W.-K. Kwok, I. Bozovic, I. Mazin, J.C. Seamus, L. Civale, D. Christen, et al., 2006, âBasic Research Needs in Superconductivity,â Basic Energy Sciences, U.S. Department of Energy, Office of Science, http://www.sc.doe.gov/bes/reports.lists.html.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 55 high critical-temperature (Tc) cuprates and iron-based superconductors. These disÂ overies have not only advanced the field of superconductivity but also fore- c shadowed quantum materials more broadly. In the past decade, superconductivity has remained a fertile field, with the discovery of both Fe-based superconductor families31 and hydrogen-rich ÂsuperÂconductors at extremes of pressure.32 In addition, the pursuit of novel superÂ conductors has enabled the development of new tools such as scanned probe microscopes. The enhanced integration of theory, experiment, and synthesis is accelerating progress toward understanding the origins of superconductivity in 8new materials discovered by the community. Nevertheless, there is still much 2.2. High-temperature superconductivity 200 HgBaCaCuO @ 30 GPa H2 S @ 155 GPa 150 TlBaCaCuO HgTlBaCaCuO BiSrCaCuO liq. CF4 HgBaCaCuO FeSe ï¬lm Critical temperature Tc [K] 100 YBaCuO SrFFeAs liq. N2 50 Cs3 C60 @ 1.4 GPa MgB2 40 SmOFFeAs LaSrCuO RbCsC60 LaBaCuO 30 Nb3 Ge BKBO YbPd2 B2 C LaOFFeAs Nb3 Sn PuCoGa5 20 liq. H2 K3 C60 Li @ 33 GPa CNT NbN V3 Si PuRhGa5 10 Pb Nb diamond UBe13 UPt3 UPd2 Al3 CeCoIn5 CeCu2 Si2 LaOFeP liq. He Hg CNT 0 1900 1940 1980 1985 1990 1995 2000 2005 2010 2015 Year Figure 2.4: 2.8â The observed superconducting transition temperature (Tc) of a variety of classes of su- FIGURE The history of some of the discovered superconducting compounds. Note in particular the BCS superconductors (green circles), the cuprates (bluediscoveries have increased the highest-observed perconductors is plotted as a function of time. Recent diamonds), and the iron-based superconductors (yellow squares). Note also the change in the axes at around year 1980 and Tc = 50 K. Partly based on Tc in a number of materials to unprecedented levels. BCS superconductors are displayed as green ï¬gures from  and [1, Fig. 2.1]. circles, cuprates as blue diamonds, carbon-based superconductors as purple and red triangles, heavy- fermion compounds as green stars, and iron-based superconductors as yellow squares. SOURCE: external magnetic ï¬eld decays inside âStructural Investigation in turn, how far into the supercon- Figure 2.4 from Pia Jensen Ray, 2015, a superconductor (or, of La(2-x)Sr(x)CuO(4+y)âFollowing ductor the screening currents occur), masterâs thesis, Niels Bohr Institute, Facultyhow the density Staging as a Function of Temperature,â while the coherence length determines of Science, Uni- versity of Copenhagen, doi:10.6084/m9.figshare.2075680.v2, Creative Commons Attribution-Share of electrons in the superconducting state can be excited and recover again without breaking Alike 4.0 International license. superconductivity. The ratio of Î» and Î¾ is the Ginzburg-Landau parameter, 31 C.Q. Choi, 2008, A new iron age: New class ofÎ»superconductor may help pin down mysterious Îº= , (2.2.1) physics, Scientific American, June 1. Î¾ 32 H. Wang, X. Li, G. Gao, Y. Li, and Y. Ma, 2018, Hydrogenârich superconductors at high pressures, and type-I and type-II superconductors are split up by the value of Îº by Wiley Interdisciplinary Reviews: Computational Molecular Science 8(1):e1330. 1 1 type-I: Îº < â , type-II: Îº > â . (2.2.2) 2 2 The HTSCs are furthermore classiï¬ed as extreme type-II superconductors, meaning that they have Îº â« 1, or Î¾ âª Î», and they are not well described by the BCS or Ginzburg-Landau theories.
56 F r o n t i e r s o f M at e r i a l s R e s e a r c h that researchers do not understand. A predictive theory of superconductivity in the cuprates remains elusive. Strongly Correlated Electrons Strong electron correlations play an important role in the physical proper- ties and quantum states of materials in two very different regimes: low-density semiconductors for which long-range Coulomb interactions remain unscreened, and high-carrier-density, narrow-band metallic systems, whose properties are in- fluenced by short-range (on-site) Coulomb interactions. During the past decade, as in prior decades, advancing the understanding of the wide range of unique phenomena that can be attributed to electron correlations has remained a major activity in materials physics. In the arena of Coulomb interactions in low-density systems, graphene has provided new ways of studying interactions in high magnetic fields, where unusual excitations, such as âcomposite Fermions,â can exist in the fractional quantum Hall regime. In strongly correlated metallic systems, which are often transition metal oxides, understanding the nature of unusual ordered states, such as spin-density waves, charge-density waves, nematic order, and superconductivity, has remained a major focus of investigation, as have associated phenomena, properties, and pathways between phases, such as metal-insulator transitions.33 A major focus over the past decade has been on quantum spin liquids.34 These can have an impact on data storage and memories in addition to further- ing the understanding of high-temperature superconductivity. In these materials, fluctuating spins do not order down to the lowest temperature but rather form a highly entangled state. Several different approaches to realizing these quantum spin liquids were pursued, exploiting avenues to obtain frustrated magnetic interac- tions. Significant experimental and theoretical work was motivated by the âKitaev modelâ of S = 1/2 Fermion on a honeycomb lattice. Experimentally, several pos- sible candidate systems were discovered, but a definitive proof that these materials are indeed quantum spin liquids is still outstanding. The search for quantum spin liquids was also one major motivation for extensive studies of the iridate family of materials over the past decade. These materials combine strong correlation physics and spin-orbit coupling. A second major motivation for the studies of iridates and related materials is the idea that exotic superconducting pairing symmetry might be realized in doped spin-orbit Mott insulators. In general, strong spin-orbit coupling 33 H. Yang, S.W. Kim, M. Chhowalla, and Y.H. Lee, 2017, Structural and quantum-state phase transitions in van der Waals layered materials, Nature Physics 13:931-937. 34 L. Savary and L. Balents, 2016, Quantum spin liquids: A review, Reports on Progress in Physics 80(1):016502.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 57 emerged as a major theme motivating new materials design and synthesis over the past decade, including in materials as diverse as correlated materials,35 topological insulators, and magnetic skyrmion systems. Thin Films The past decade also saw major progress in strongly correlated thin films and heterostructures. These studies are motivated by the ability to design novel quan- tum states using approaches such as quantum confinement or electric field gating that are not available in the corresponding bulk materials. For example, significant research activities over the past decade in rare-earth nickelate thin films were motivated by theoretical predictions of a cuprate-like Fermi surfaces in hetero- structures. Thin film studiesâin particular, strain and interface engineeringâled to new insights into the role of lattice coupling in strongly correlated phenomena such as metal-insulator transitions. Major advances were also made in thin film synthesisâin particular, molecular beam epitaxy (MBE)âof correlated materials. Examples include very high mobility complex oxide thin films grown by MBE and the demonstration of superconducting Sr2RuO4 films. Developments in the areas of computation, growth, and measurement, and their coordination36,37 have led to substantial advances in the field of quantum materials. The utilization of these techniques have led to the discovery of new superconductors, with Tc over 200 K in the hydrogen-sulfides at extremely high pressure, and to a deeper understanding of novel quantum phases owing to enor- mous advances in the measurement of electronic properties (e.g., tunneling, point contact, photoemission, and terahertz spectroscopies; some of these challenging measurements are done with surface tools in order to measure bulk properties), quantum oscillations, resonant inelastic X-ray scattering, and neutron scatteringâ all possible because of the substantial improvement in materials quality, crucial to this field. The Cooper pairing mechanism for the various families of uncon- ventional superconductors is also being clarifiedâphonons, spin-excitations, and orbital fluctuations all seem to have a role to play. Last, researchers have discovered, identified, and gained understanding and control of newer forms of quantum mat- ter, including topological insulators, and van der Waals semiconductors where the roles of strong electron correlations, broken symmetries, topology, and dimen- sionality are being explored. Researchers have learned to control and manipulate 35 J. Mannhart and D.G. Schlom, 2010, Oxide interfacesâan opportunity for electronics, Science 26:1607-1611. 36 C. Broholm, et al. âBasic Research Needs in Quantum Materials,â Basic Energy Sciences, U.S. Department of Energy, Office of Science, http://www.sc.doe.gov/bes/reports.lists.html. 37 D.N. Basov, R.D. Averitt, and D. Hsieh, 2017, Towards properties on demand in quantum materi- als, Nature Materials 16:1077-1088, doi:10.1038/nmat5017.
58 F r o n t i e r s o f M at e r i a l s R e s e a r c h their quantum phases with a variety of experimental techniques, including those at extreme conditions, such as high pressure, high photon flux, and high magnetic field. An example is the recent observation of the Higgs mode in a disordered su- perconductor38 and in NbN using time-resolved terahertz spectroscopy. 2.4.2 Magnetic Materials Magnetic materials principally are used in two major types of applicationsâ electromagnetic devices and magnetic information storage and logic. In elec- tromechanical applications, motors, and actuators, the quest for new materials focuses on hard magnetsâthat is, materials that develop strong external magnetic fields that, in turn, produce strong forces. These are particularly relevant today, as electric propulsion is becoming prevalent, and the performance of hard magnets enables the design of ever more powerful, compact, and lightweight motors; see, for example, Case 3 in Section 5.3. The second class of applications is in computers, and particularly magnetic memories and readouts. In the near future, distributed sensing and computing will require greatly reduced power consumption and the integration of multifunc- tionality into single-chip devices that will require new materials and architectures. Magnetism has an important role to play in this multidimensionality. Almost the entire industry of nonvolatile memories (those that retain the information when the power is switched off) are based on magnetic random-access memory (MRAM). So are magnetic recording devices such as hard drives. These devices are largely based on the physics of spin dynamics. Progress in this field has been driven by the discovery of the giant magnetoresistance, for which the 2007 Nobel Prize in physics was awarded jointly to Albert Fert and Peter GrÃ¼nberg. Over the course of the last three to four decades, progress has been made in developing new rare-earth-based hard magnets (neodymium-iron-boron or samarium-cobalt). These materials have enabled the development of compact high-power motors for electric transportation. The core of the work consisted in developing materials with a high remanent magnetization (the magnetic field that a magnet can produce by itself), a high coercive field (the amount of external magnetic field that a magnet can take without losing its own magnetization), and a high operating temperature. One drawback of the current materials is that they rely on rare-earth elements, which are costly to extract and purify. Nanotechnology has been of substantive help in this work, as nanometer-size grains of magnetic material tend to pin down the magnetic alignments in the grains and increase the 38 D.Sherman, U.S. Pracht, B. Gorshunov, S. Poran, J. Jesudasan, M. Chand, P. Raychaudhuri, et al., 2015, The Higgs mode in disordered superconductors close to a quantum phase transition, Nature Physics 11(2):188.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 59 coercivity. Progress in manufacturing of these materials includes the development of printable magnets (see Box 2.6). The past decade has seen considerable progress in spin dynamics and spin transport, typically in linear spin transport. Recent developments include the discovery of the spin-dependent Seebeck effects (thermoelectric effects of spin- polarized conduction electrons); the characterization of the out-of-equilibrium magnon system not only by its classical dispersion relation but also by the new theoretical concepts such as the magnon chemical potential. In magnonic systems, linear transport is described in terms of a spin conductance and a spin-Seebeck effect in either metallic ferromagnets like Permalloy, magnetic semiconductors like GaMnAs, or, most recently, in ferromagnetic insulators like yttrium-iron-garnet (YIG). Spin transport across interfaces enriches the number of possible effects because the interface breaks the symmetry present in uniform materials. In the past decade, the bilayer system consisting of a ferromagnet and a metal with strong spin-orbit interactions, most notably Pt/YIG, has become the paradigm for new discoveries and for new explanations developed for older discoveries, such as the anomalous Hall, spin-Hall, inverse spin-Hall, and spin-Nernst effects.39 These effects have led to a completely new way of measuring local magnetizations and spin currents: a new tool that will lead to future important discoveries in magnetic materials and widen the possibility of designing new spin-electronic devices, such as logic ele- ments, amplifiers, and oscillators. Topological insulator materials, which are also based on strong spin-orbit interactions, have been coupled in this way to ferro- magnetic solids, typically Bi2Se3/YIG. Spin-Torque Oscillators Spin-torque oscillators such as FeB/MgO/CoFeB trilayers in 100 nm diameter pillars predate the past decade. However, it was suggested in 2017 that they could provide an extraordinary hardware platform on which to perform neuromorphic computations. Neuromorphic computers are analog computers inspired by the logi- cal architecture of the brain. The architecture requires highly stable oscillators (the source of âbrain wavesâ) that give an output function that is nonlinear with respect to the input signal yet do not dissipate too much power. Spin-torque oscillators have all the desired characteristics, while being solid-state devices that operate at GHz frequencies. They have been demonstrated to provide a significant boost to speech-recognition algorithms (see Figure 2.9). 39 S. Meyer, Y.T. Chen, S. Wimmer, M. Althammer, T. Wimmer, R. Schlitz, S. GeprÃ¤gs, et al., 2017, Observation of the spin Nernst effect, Nature Materials 16(10):977-981, doi: 10.1038/nmat4964.
60 F r o n t i e r s o f M at e r i a l s R e s e a r c h BOX 2.6 Printable Magnets The Critical Materials Institute (CMI), which is part of the Department of Energyâs Ames Laboratory, has developed a process to print high-strength bonded magnets (see Figure 2.6.1) that outperform conventionally processed magnets in some applications. This allows for devices like generators and electric motors to be made with reduced demand for critical materials such as neodymium and dysprosium. The process was recognized with an R&D 100 Award in 2017 and offers electric machine designers some distinct advantages, as follows: â¢ Net shape or near-net shape production reduces machining losses of valuable magnet materials. â¢ The ability to print complex shapes allows for improved use of magnetic flux. â¢ Comparable motor performance can be achieved with less magnet material, or im- proved performance can be achieved with the same amount used in a design based on conventionally manufactured magnets. FI GURE 2.6 .1 âIsotropic, neodymium-iron-boron bonded permanent magnet, 3D-printed at the Department of Energyâs Manufacturing Demonstration Facility at Oak Ridge National Laboratory. NOTE: See L. Li, A. Tirado, I.C. Niebedim, O. Rios, B. Post, V. Kunc, R.R. L Â owden, et al., 2016, Big area additive manufacturing of high performance bonded NdFeB Âmagnets, ÂScientific Reports 6:36212, doi: 10.1038/srep36212. SOURCE: Courtesy of Oak Ridge Â ational Laboratory. N
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 61 a b c FIGURE 2.9â(a) FeB/MgO/CoFeB trilayer spin-torque oscillators improve speech recognition by 70 percent. Spoken numbers (1 to 10), input as an audio file, are preprocessed by a conventional computer and then fed to a highly stable but nonlinear spin-torque oscillator driven by a current. The output waveform in the current is analyzed for word recognition by a classical digital computer. (b) Trilayer spin-torque oscillators of a multilayer stack consisting of FeB/MgO/CoFeB as a current is passed through the stack. (c) The success rate during training with and without the spin-torque oscillator: adding the spin-torque oscillator increases the success rate of the word recognition by 70 percent. SOURCE: Panels (a) and (c) were committee generated. Panel (b) is from J. Torrejon, M. Riou, F.A. Araujo, P. Bortolotti, V. Cros, and J. Grollier, 2017, Neuromorphic computing with nanoscale spintronic oscillators, Nature 547:428-431. Antisymmetric exchange interactionsâfor example, the Dzyaloshinskii- Moriya interactionâfavor orthogonal spin alignment, which leads to a number of new skew effects such as a possible magnon Hall effect. This introduces chirality in the magnetizationâfor example, in skyrmions. Skyrmions are localized reversals of magnetization in an otherwise uniform ferromagnet (see Figure 2.10a). Their spatial extent is generally greater than a lattice spacing but still small (down to a few nanometers), and they move with little energy cost (several orders of magni- tude lower than that needed to move magnetic domain walls). These properties project the possibility that skyrmions can be used to create magnetic memory ele- ments with extremely high information density and low power consumption. The
62 F r o n t i e r s o f M at e r i a l s R e s e a r c h a b FIGURE 2.10â (a) The smallest possible perturbation of the otherwise uniform magnetization (up- arrows) of a ferromagnet is a skyrmion. Its core (down-arrow) is a single atom surrounded by a twisting of atomic magnetic moments that returns the spin texture to the background direction. (b) A triangular lattice of skyrmions has been observed in FeGe. SOURCE: (a) Reprinted with permission from Springer Nature: C. Pfleiderer, 2011, Magnetic order: surfaces get hairy, Nature Physics 7:673- 674, copyright 2011. (b) Reprinted with permission from Springer Nature: A. Fert, V. Cros, and J. Sampiao, 2013, Skyrmions on the track, Nature Nanotechnology 8:152-156, copyright 2013. integration of magnetic effects on single-chip devices will be further enhanced by the development of multiferroic materials, which combine magnetic effects with electrical ones. Multiferroic materials involve simultaneously two or more âferroicâ ordersâ for example, ferromagnetism and ferroelectricity. Thus, when ferroelectric solids (those that are spontaneously polarized with reversible electric polarization) also display ferromagnetic or ferrimagnetic order (spontaneous magnetic polarization), the materials are known as multiferroic. Two possibilities have come to the fore in the past 10 years. BiFeO3 is the sole single-phase multiferroic that demonstrates ambient temperature magnetoelectric coupling. Used in concert with CoO0.9Fe0.1 to amplify magnetization, BiFeO3 is the closest switching multiferroic to production. The second successful room-temperature multiferroic for switching is a lutetium- based superlattice consisting of hexagonal LuFeO3, a ferroelectric, with LuFe2O4, a ferrimagnet. In constructing the superlattice, a monolayer of LuFe2O4 is added once every 10 layers of the LuFeO3 through the addition of FeO during the molecular beam epitaxial growth to produce the multiferroic response. Switching is a signifi- cant hurdle in the development of devices. Optical control, to afford picosecond control, is likely required, so long as the switching is perfectly reversible down to the level of the magnetic domains. Multiferroic studies also have inspired new research and technology develop- ment in other material-related areas. For example, more general studies of possible mechanisms that promote spontaneous order in ferroelectrics are under way. Sig- nificant progress in the growth of oxide heterostructures has advanced to enable
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 63 joining of multiferroics with circuitry. Nonlinear laser spectroscopy has advanced to allow imaging of magnetic and ferroelectric domains and their interaction. 2.4.3 Two-Dimensional Quantum Materials Graphene The modern ascension of 2D materials began with the isolation and electrical measurement of single-atomic-layer graphite, or graphene, by Geim and Novoselov in 2004, which garnered a Nobel Prize in 2010.40 Although the band-structure of graphene had been calculated in 1947, and individual flakes had been imaged with electron microscopy, Geim and Novoselov demonstrated that one-atom-thick membranes could be simply created and easily fabricated into electronic devices. While the first graphene devices were created by mechanical exfoliation of graphite using adhesive tape, large-area growth can now be achieved via chemical vapor deposition, liquid-phase exfoliation, and synthesis on SiC.41 Graphene is a zero- gap semiconductor having properties such as high electron mobility (>15,000 cm2Vâ1sâ1), large thermal conductivity, mechanical strength and elasticity, optical transparency, impermeability, and high electrical sensitivity to adsorbates. Experiments on graphene have demonstrated novel and potentially useful electronic, photonic, mechanical, and thermal behavior. For example, it has been shown that charge carriers in graphene display relativistic effects such as tunneling through large barriers and can also display collective hydrodynamic flow, while the unique properties of grapheneâs optical absorption make it highly opaque. However, to enable graphene to be used in electronic and optoelectronic devices it is necessary to determine how to create a tunable bandgap. Different approaches have been developed including confinement, defect introduction, mismatch strain from a substrate, adsorption, and strain. Mechanically, graphene is similar to paper: it is hard to stretch but easy to bend. This property was predicted and later shown to be attributed to ripples in the graphene membrane. The high tensile strength of graphene means that it breaks at low strains. However, large deformation strains have been shown to be possible in graphene through the use of kiragami-like cuts, with reversible elongations of 240 percent having been achieved. A remarkable finding was that the electrical and thermal conductivity was insensitive to strainâa property achieved by the lack of deformation of the graphene itself. This use of strain engineering to graphene 40 See K.V. Academien, 2010, âGrapheneâThe Perfect Atomic Lattice,â https://www.nobelprize.org/ nobel_prizes/physics/laureates/2010/press.html. 41 E.O. Polat, O. Balci, N. Kakenov, H.B. Uzlu, C. Kocabas, and R. Dahiya, 2015, Synthesis of large area graphene for high performance in flexible optoelectronic devices, Scientific Reports 5:16744.
64 F r o n t i e r s o f M at e r i a l s R e s e a r c h sheets opens avenues to using them for stretchable electronics, hinges, springs, and thermal management. Further remarkable properties have been demonstrated in multilayer and encapsulated graphene. For example, the application of a gate voltage can induce a gap in bilayer graphene, leading to an ultrathin but mechanically robust semi- conducting material. Adding layers (âfew-layer grapheneâ) yields metallic material that maintains useful properties such as thinness, mechanical and thermal stabil- ity, transparency, and sensitivity to functionalization. It has been shown that the properties of graphene can be optimized by encapsulation in an inert material, particularly hexagonal boron nitride (h-BN), which is itself a monolayer insulator that is lattice-matched to graphene. There has been tremendous effort toward creating graphene-based applica- tions, with a few commercial products already at market, including security tags using graphene ink (Siren Technology), graphene-enhanced diaphragms for ear- phones (FiiO Electronics), graphene-composite tennis rackets (Head) and helmets (Catlike), graphene supercapacitors (Skeleton Technologies), and graphene-based molecular sensors (Nanomedical Diagnostics). Transition Metal Dichalcogenides Since the isolation of graphene, there has been a revolution in the discoveries of monolayer and few-layer materials that can be mechanically or chemically exfoli- ated. The observation in 2010 of the dramatic increase in the photoluminescence in a monolayer of molybdenum disulphide (MoS2) over that in the bulk material or even in bilayer MoS2 has led to an explosion in activities in the genre of mate- rials known as transition metal dichalcogenides, MX2, in which M is a transition metal and X is sulfur, selenium, or tellurium. These materials span the full range of electronic behavior from metallic to wide-bandgap insulators. These materials have diverse electronic, quantum, thermal, optical, chemical, and mechanical properties, ranging from MoS2 or phosphorene as n- and p-type semiconductors to NbSe2 as a superconductor and WTe2 as a Weyl semimetal (see Box 2.7 in Section 2.4.4 for an explanation of Weyl semimetals). They are also known as van der Waals materials because in the crystal the monolayers are held together by weak van der Waals interactions. The monolayer itself consists of an atomic layer of hexagonally packed metal atoms sandwiched between two layers of chalcogen atoms. This sandwich structure results in valence- satisfied atoms, making the basal plane inactive. As with graphene, monolayer transition metal dichalcogenides exhibit quantum confinement leading to enhance- ments in the electronic, optical, thermal, and mechanical properties that are sig- nificantly different from those of the bulk materialâfor example, the semimetallic to semiconducting transition observed in ultrathin TiS2, the metallic-to-insulating
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 65 BOX 2.7 Weyl Semimetals The topological insulator (TI) and Weyl semimetal (WSM) or Dirac semimetal (DSM) are shown in Figure 2.7.1. The topology of both a TI and a WSM/DSM originates from a similar inverted band structure. FI GURE 2.7.1 â(a) The spin-orbit coupling (SOC) opens a full gap after the band inversion in a TI, giving rise to metallic surface states on the surface. (b) In a WSM/DSM, the bulk bands are gapped by SOC in the 3D momentum space except at some isolating linearly cross- ing pointsânamely, Weyl points/Dirac pointsâas a 3D analog of graphene. Owing to the Â topology of the bulk bands, topological surface states appear on the surface and form exotic Fermi arcs. In a DSM all bands are doubly degenerate, while in a WSM the degeneracy is lifted owing to the breaking of the inversion symmetry or time-reversal symmetry, or both. The net Berry phase accuÂ ulated in the 2D k plane between a pair of Weyl points induces m a nonzero Chern number C = 1 with a quantized anomalous Hall effect, while the Berry Â phase is zero in other planes with C = 0. (c) The type-I WSM. The Fermi surface (FS) shrinks to zero at the Weyl points when the Fermi energy is sufficiently close to the Weyl points. (d) The type-II WSM. Owing to the strong tilting of the Weyl cone, the Weyl point acts as the touching point between electron and hole pockets in the FS. SOURCE: Reproduced with permission of Â nnual Reviews, Inc., from B. Yan and C. Felser, 2017, Topological materials: A Weyl semimetals, Annual Review of Condensed Matter Physics 8:1-19; permission conveyed Â through Copyright Clearance Center, Inc. Â
66 F r o n t i e r s o f M at e r i a l s R e s e a r c h transition of TaS2, and the indirect-to-direct bandgap transition together with the widening of the bandgap for molybdenum and tungsten dichalcogenides. Some of these 2D materials also exhibit interesting strongly correlated electron phenom- ena, such as charge density waves and superconductivity. Another very promising property of semiconducting monolayer transition metal dichalcogenides is the re- markably strong light-exciton interactions and greatly enhanced electron-electron interactions. Exciton binding energies in these materials can be hundreds of mil- lielectron volts, two orders of magnitude larger than what is seen in typical bulk semiconductors. Similar is the trend in the binding energies of trions (charged excitons), about tens of millielectron volts, much larger than that in ordinary semi- conductors, owing to reduced electron screening resulting from their 2D nature. Moreover, there is the possibility of full control of the valley and spin occupation by optical pumping with circularly polarized light. Monolayer MoS2 has already also been shown to yield field-effect transistors with very high current on-off ratios and to function within a vertical graphene-MoS2-graphene tunneling transistor structure. These properties, in combination with the promise of synthesis of large- area high-quality samples (notably by recently developed CVD methods), suggest interesting possibilities for applications of this and related materials in electronics and optoelectronics.42 Naturally, the structural and chemical terminations of the edge of the 2D transition metal dichalcogenides play a role in determining the local physical and chemical properties of these systems. Furthermore, vacancies in the basal plane and dopant can also change the local electronic structure. First principles calcula- tions have predicted that diverse edge passivation in metal dichalcogenides could produce different spin states that modify the edge electronic and magnetic proper- ties. In addition, high spin density could be localized surrounding metal vacan- cies. Calculations also predict that electronic structural changes brought about by vacancies could lead to chemical reactivity. While experimental exploration is still at an early stage, preliminary results already point to promising chemical activity of this otherwise inactive basal plane when pristine. Adsorbed metallic nanoparticles could also facilitate chemical reactions of 2D MoS2. Given the bandgap of about 1.8 eV, monolayer MoS2 is also a promising candidate for water splitting. Although monolayer MoS2 became popular as a possible opto-electronic material, its catalytic activities seem to have been realized at a faster pace. 42 S.Z. Butler, S.M. Hollen, L. Cao, Y. Cui, J.A. Gupta, H.R. GutiÃ©rrez, T.F. Heinz, et al., 2013, Progress, challenges, and opportunities in two-dimensional materials beyond graphene, ACS Nano 7(4):2898-2926.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 67 Looking Beyond Graphene and Transition Metal Dichalcogenides There are many layered materials that go beyond the transition metal dichal- cogenides (TMDs), including monochalcogenides (GaSe, etc.), monoelemental 2D semiconductors (silicene, phosphorene, germanene), black phosphorus, and MXenes (see Figure 2.11). The great ability to tune the bandgap, band offset, car- rier density, carrier polarity, and switching characteristics in 2D materials provides unparalleled control over device properties and possibly new physical phenomena. Devices based on atomically thin monolayers are the extreme scenario for the future lightweight, low-power consumption, and wearable electronics. Moreover, the true potential of these layered materials may emerge from the ability to stack them, layer by layer in any desired sequence, to create novel 3D architectures with entirely new functions. Elemental examples of 2D materials include the X-enes: graphene, phospho- rene, stanene, and germanene, to name a few. Two-dimensional allotropes and compounds are rapidly growing in number with 2D nitrides, such as h-BN, and TMDs, such as MoS2, receiving the greatest attention. Other TMDs having the form of MX2 (where M = Mo, W, Ti, Nb, etc., and X = S, Se, or Te) have started to gain significant interest. More complex compounds such as fluoro-X-enes, chloro-X- enes, X-anes, and MX-enes have also been theorized and demonstrated. Potential applications include energy harvesting and storage, sensing, pharmaceuticals, elec- tronics and photonics, and bioengineering. Another 2D material that was discovered in the wake of graphene is 2D h-BN, which was theoretically predicted to induce a bandgap in graphene when serving as its substrate. The flurry of activity in h-BN because of its role as a stable substrate, and as a gate dielectric or deep ultraviolet emitter, however, has promoted it as an FIGURE 2.11âThe chalcogenide families. SOURCE: Courtesy of Joshua Robinson, Penn State, https://news.psu.edu/story/466016/2017/05/01/research/stenciling-atoms-two-dimensional- materials-possible.
68 F r o n t i e r s o f M at e r i a l s R e s e a r c h interesting material in its own right. While being a wide-bandgap semiconductor, it has recently been shown to be an excellent hydrogenation catalyst, when laden with defects. As one of the few metal-free catalysts, it has emerged as a promising 2D material. In summary, there are now more than 600 different 2D materials recognized, most of which had 2D stability predicted only in the past decade, and some of which have not yet been synthesized. 2.4.4 Topological Materials It is rare that unexpected states of matter appear; however, in 2005-2006 it was predicted that a new quantum state, a topological insulator (TI), should exist in ma- terials where spin-orbit coupling causes electrons to develop an excitation spectrum with a nontrivial topology and wavefunctions characterized by an invariant (e.g., a Chern number) that remains unchanged under adiabatic deformations of the system. Unlike conventional atomic insulators, TIs exhibit gapless, conducting edge states in addition to a full insulating gap in the bulk so that the material behaves as an insulator in its interior, but the surface contains conducting states, meaning that electrons can move only along the surface of the material. The past decade saw the prediction and characterization of topological insula- tor properties in many materials, including InAs/GaSb heterojunctions, Bi1-xSbx alloys, Bi2Te3, GeBi2Te4, and SmB6. Additional progress in theoretical techniques include band structure analysis under spin-orbit interactions and systematic analy- sis of crystal symmetries, which have been developed to determine a multitude of topological properties and predict bulk and edge properties. In 2016, the efforts resulted in the Nobel Prize in physics being awarded âfor theoretical discoveries of topological phase transitions and topological phases of matter.â43 Beyond the basic recognition of topological properties, there has been significant progress in fabri- cating topological materials that clearly express the topological properties, which are needed for fundamental research or to enable device applications. Most materials, such as the Bi-based ones, were grown as bulk crystals. Thin films were directly exfoliated from crystals or grown by molecular-beam epitaxy, while nanowires of Bi2Se3 were grown by the vapor-liquid-solid mechanism. Dop- ing Bi2Se3 with Cu or Nb can induce superconductivity, while doping with Mn in- duces ferromagnetism. However, for many materials, it remains a challenge to tune the Fermi energy to within the bandgap, where topological states should dominate. For example, Bi2Se3 and related compounds suffer from strong n-doping owing to 43 The 2016 Nobel Prize in physics was awarded to David Thouless, Duncan Haldane, and Michael Kosterlitz. See, for example, E. Gibney and D. Castelvecchi, 2016, Physics of 2D exotic matter wins Nobel, Nature News 538(7623):18.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 69 Se vacancies; this was alleviated by adding bulk dopants such as Sb, or by utilizing undoped compounds such as SmB6. Soon after the initial predictions, transport measurements of 2D HgTe/CaTe quantum wells provided evidence of quantized edge states and angle-resolved photoemission spectroscopy (ARPES) of protected surfaces states with single Di- rac cones (see Figure 2.12). Since then, many other experiments have determined unusual properties of TIs: helical spin polarization was shown via spin-polarized ARPES on 3D TIs, while quantized edge states have been found in other 2D TIs such as ZrTe5 and InAs/GaSb. Scanning tunneling microscopy on 3D TIs near de- fects showed evidence of suppressed backscattering, while measurements of quan- tized polarization switching in light reflected off TIs is consistent with predicted ad- ditional cross-terms for magnetization and polarization in Maxwellâs equations (the quantized magneto-electric effect). Magnetically doped TIs exhibited a quantum anomalous Hall effect, where ferromagnetism and spin-orbit coupling combine to create a quantum Hall effect at zero external magnetic field. Other predicted be- havior in TIs, such as of fractional edge charges and bulk spin-charge separation, remain to be demonstrated. A topological superconductor is predicted to have a gapped bulk and gapless surface states, with excitations that are Majorana zero modes. Although many experiments to date have studied superconducting proximity effects in topologi- cal surface states, clear evidence of Majorana modes in these systems has not yet been demonstrated. Topological superconductivity was proposed to exist in semiconducting nanowires with strong spin-orbit coupling, proximity-coupled to s-wave superconductors, and placed in magnetic fields, so that spin-triplet FIGURE 2.12â High-resolution ARPES surface band dispersion through of Bi2-Î´CaÎ´Se3 (111). The arrows denote the topological spin polarization of the bands. The separation between surface and bulk states, as well as Dirac-like dispersion can be seen. SOURCE: Reprinted with permission from Springer Nature: D. Hsieh, Y. Xia, D. Qian, L. Wray, J.H. Dil, F. Meier, J. Osterwalder, et al., 2009, A tunable topological insulator in the spin helical Dirac transport regime, Nature 460:1101-1105, doi: 10.1038/nature08234, copyright 2009.
70 F r o n t i e r s o f M at e r i a l s R e s e a r c h superconductivity is induced. The predicted Majorana excitations have been dem- onstrated in InSb wires coupled to s-wave superconductors, as well as in a related system of atomic-scale iron nanowires on lead substrates. However, coherence, braiding, and other qubit properties remain to be demonstrated. The past decade has identified a wealth of materials, beyond 2D and 3D to- pological insulators, with interesting and potentially useful topological properties, opening a new materials area of âtopological quantum matter.â Among these are Weyl semimetals (see Box 2.7), conducting materials whose excitations are highly mobile fermions (a chiral half of a Dirac fermion); after theoretical predictions, Weyl properties were discovered in TaAs via ARPES. Weyl semimetals are a specific example of the broader category of âtopologi- cal metalsâ whose excitations are topologically protected fermionic quasiparticles. Topological crystalline insulators, such as SnTe, and Pb1-xSnxSe are topological phases of matter that are protected by crystal symmetries, including rotation and reflection; the Dirac cones and surface-state properties of these materials have been explored via ARPES and STM. New research has demonstrated that disorder and interactions may further tune the properties and phase transitions of topological materials. Last, a large variety of topological materials formed from 2D membranes has been proposed. The great advances in the topological physics of electron waves have naturally led to an explosion of research on topological properties of photons, phonons, plasmons, and other waves. Initial research in this arena focused on replicating electronic behavior in microwave photonics in easily constructed centimeter-scale systems. The earliest demonstration of such a system was a materials system con- sisting of a square array of gyromagnetic ferrite rods (providing coupling between electric and magnetic fields) in an external magnetic field (to break time-reversals invariance). This system exhibited the optical analog of the quantum Hall effect with a single topologically protected, unidirectional edge mode that propagated around arbitrary disorder without reflection. This and related systems enable novel device designs for a variety of optical applications. Optical systems involving a double-gyroid lattice in a magnetic field44 exhibit Weyl points and line nodes, and coupled infrared optical resonator waveguides with degenerate oppositely moving modes mimic the quantum spin Hall effect of topological insulators with edge states that have effective topological protection. Similarly inspired topological me- chanical and fluid mechanical models exhibiting protected edge modes have also 44 J.A. Dolan, B.D. Wilts, S. Vignolini, J.J. Baumberg, U. Steiner, and T.D. Wilkinson, 2014, Optical properties of gyroid structured materials: From photonic crystals to metamaterials, Advanced Optical Materials 3:12-32, https://doi.org/10.1002/adom.201400333.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 71 been constructed. For example, Vitelli and Irvine et al.45 have built a new type of mechanical metamaterial: a âgyroscopic metamaterialâ composed of rapidly spin- ning objects that are coupled to each other. At the edges of these materials, they find that sound waves are topologically protected so that they cannot be scattered back into the bulk. 2.4.5 QubitsâThe Building Blocks for Quantum Computers Quantum computers hold the promise of being able to solve some problems efficiently that are believed to be intractable for classical computers. They are chal- lenging to construct because one must have a way of specifying and controlling the quantum operations while avoiding decoherence arising from unintended interac- tions with the environment. A summary of the goals and comparative achievements of quantum computing is summarized in the report âTechnical Roadmap for Fault- Tolerant Quantum Computing,â46 Tremendous progress toward the development of quantum computers has been made in the past decade, and progress in materials has been critical to these achievements. The qubit types that have emerged in the past decade along with the pros and cons of each are summarized in Box 2.8. Of these qubit types, the two leading ones are superconducting qubits and ion-trap qubits. This section focuses on those qubits with materials challenges. The most commonly used superconducting qubit is the transmon, an anhar- monic oscillator where nonlinear inductance is provided by a Josephson junction shunted by a capacitance provided by large superconducting pads in proximity. The typical transmon utilizes the most mature and simply fabricated superconducting junction, Al-AlOx-Al, where the AlOx is grown inside the evaporation chamber without breaking vacuum. Nonetheless, amorphous AlOx has been measured with a two-level-system defect density of not less than 0.5 (mm2-GHz)â1. To put this in perspective, current transmons are tenths of a mm in size, and since the micro- waves used to communicate with the transmons are in the gigahertz range, even this minimized defect density will result in roughly one defect every 10 GHz per qubit. In order to advance this technology, a greater understanding and control of defects in these materials is necessary. Current promising directions toward remov- ing defect sources include growing an epitaxial insulating layer, such as epitaxial Al2O3 by creating Re-Al2O3-Re or Re-Al2O3-Al junctions, epitaxial MgO as in Re/ 45 L.M. Nash, D. Kleckner, A. Read, V. Vitelli, A.M. Turner, and W.T.M. Irvine, 2015, Topologi- cal mechanics of gyroscopic metamaterials, Proceedings of the National Academy of Sciences U.S.A. 112(47):14495-14500. 46 A. Fruchtman and I. Choi, 2016, Technical Roadmap for Fault-Tolerant Quantum Computing, Networked Quantum Information Technologies, Oxford, UK, October.
72 F r o n t i e r s o f M at e r i a l s R e s e a r c h BOX 2.8 Qubit Types In quantum computers, materials and physical processes are optimized to form qubits, a localized system that can take the values of not only one or zero but also any superposition of these. When two or more qubits interact, quantum entanglement can occur, enabling opera- tions on many qubits simultaneously, and inspiring visions of calculations that are beyond the reach of classical computers. Entangled states are still fragile, and rates of decoherence affect not only accuracy but also gate depth (number of consecutive operations). The leading qubit platforms are described in Figure 2.8.1. FIGURE 2.8.1 âThe leading qubit platforms. SOURCE: Figure adapted from G. Popkin, 2016, Quest for qubits, Science 354:1090-1093, reprinted with permission from AAAS. MgO/Al, epitaxial Nb-Al-Al2O3-Nb, or semiconducting nanowires sandwiched between aluminum leads. To protect transmons from the environment, an architecture called circuit quantum electrodynamics (cQED) is employed whereby the qubit is coupled to a microwave resonator, and that readout resonator is coupled to the external envi- ronment. For the cQED component, microwave resonators are usually patterned by a subtractive process from sputtered or expitaxial films, requiring materials that are simple to etch through a photomask (e.g., Al, Nb, Re, TiN). A low-power Q for these resonators is also sought. One style of transmon couples the qubit on a sap- phire substrate to the microwave field in a 3D cavity. Since qubits measured in this way have the highest lifetimes, this technique has been used for studies of surface participation on qubits, thereby informing qubit design. These cavities are typically made from high-purity aluminum, but other materials and coatings/processes have been considered. There have even been reports of high-Q 3D-printed cavities.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 73 Quantum annealing has been used to find the lowest energy state of a single challenging Hamiltonian, which is relevant for many optimization problems. In this computer, which uses a different type of transmon, switches are formed from Josephson junctions of two Nb layers separated by a thin layer of insulator such as AlOx. Variability in junction thickness is a challenge, ameliorated with additional circuitry and fields. The question of the amount of quantum speedup for this type of machine is an area of active scientific discussion. Topological qubits, which have emerged only in the past decade, have at- tracted much excitement because they can be inherently fault tolerant. These qubits are quasiparticle excitations of hybrid topological systems (e.g., superconducting proximity-coupled topological insulators, or proximity-coupled semiconducting nanowires in magnetic field). The excitations are characterized as nonabelian any- ons, which have superimposed quantum states that can be encoded in braided qua- siparticle trajectories. These braided paths are âtopologically protected,â meaning that they require less error correction than other types of qubits (e.g., transmons). Finding clear evidence of nonabelian anyonsâspecifically, Majorana excitationâ has been a challenge. Promising results have been achieved in InSb wires coupled to s-wave superconductors, as well as in a related system of atomic-scale iron nanow- ires on lead substrates. However, coherence, braiding, and other qubit properties remain to be demonstrated, and it remains unclear which systems, materials, and measurements will be optimal. Semiconductor quantum dots, which act as âartificial atomsâ with addressable spin and charge states, have the advantage of building on existing semiconductor technology infrastructure. Early quantum dots were made in heterostructures of GaAs and AlGaAs, but it is now much more common to use devices made in silicon or in Si/SiGe heterostructures, with the electrons confined and manipulated using voltages applied to lithographically defined metallic gates. Charge and nuclear spin noise are the dominant sources of decoherence and gate Â rrors. To counteract spin e noise, isotopically purified silicon is often used, while the effects of charge noise are mitigated by changes in qubit design including dopants. Recent work has dem- onstrated that other materials systems could be suitable for hosting quantum dot qubits. For example, an artificial double quantum dot molecule in a gated MoS2 van der Waals heterostructure has promising controllability, properties, as well as manufacturable reproducibility. The use of 2D chacolgenides gives promise for spin-valley qubits. Also, it has been proposed that donor atoms of âdeepâ chalcogen donors, such as sulfur, selenium, tellurium, and particularly 77Se+ is possible. The unique optical properties of 2D and heteroÂ tructured Â aterials open possibilities s m of accessing the strong coupling limit of cavity quantum electrodynamics using silicon photonic resonator technology and integrated silicon photonics. Qubits have been demonstrated by manipulation of optically active crystal defectsâfor example, nitrogen-vacancy (NV) centers in diamond. This has a negative charge, which creates an optically active, paramagnetic spin-1 complex.
74 F r o n t i e r s o f M at e r i a l s R e s e a r c h NV-centers are initialized by optical pumping and can be read out using optical (spin-dependent photoluminescence) or electronic methods. Entanglement and quantum logic operations have both been demonstrated. Another defect that acts like an artificial atomâwith a paramagnetic spin that has, like a nitrogen vacancy in diamond, relatively few decoherence interactionsâis the divacancy in silicon carbide. Scaling of spin systems to larger regular arrays is difficult because magnetic dipole interactions are detectable only up to around 30 nm, and placing vacancies to that density in diamond is a challenge. Another application of quantum information is to sensing and metrology, where entangled states are used as robust, sensitive, nanoscale sensors. Leading qubits for this application are nuclear and electronic spins in semiconductors, and NV in diamond. Quantum computing is an emerging technology that has appropriately at- tracted much attention. Given the current state of the art, MR can accelerate progress by focusing on challenges such as the role of defects and noise in device performance. 2.5 POLYMERS, BIOMATERIALS, AND OTHER SOFT MATTER 2.5.1 Polymers The past decade has seen a powerful acceleration of capabilities in the precision synthesis of polymers, consistent with a drive in every area of materials science to control, with fidelity and exactness, the placement and arrangement, not only of atoms and molecules, but also of defects, which often control material properties. This takes specific form in polymer synthesis in the control of polymer primary structure including programmed degrees of chain uniformity, monomer sequence, and placement of short and long branches, which has advanced greatly owing to the progress in controlled, living polymerization (CLP). CLP continues to expand capabilities beyond living anionic polymerization, especially in the realm of con- trolled free-radical polymerization (CFRP) including water-based ones, for which increasingly potent reagent and catalyst systems have been introduced. CFRP has been shown to be much less sensitive to environmental conditions than other CLP, which leads to increased versatility. Synthetic biology has also produced some intriguing advances in precision macromolecular synthesis of relevance to materials science. Biology, and all of its functionality, is controlled by the sequences formed from combinations of monomers inscribed in linear polymers. No such level of control is even remotely possible yet in chemically synthesized copolymers. Biological synthesis, however, can be commandeered to produce unnatural amino acid polymers, which can be further diversified by post-translational, or even post-polymerization, chemical
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 75 modification. Techniques that enable biological polymerization of a wide diversity of chemical types of monomers are currently very limited, but some progress has been made in polymerizing Î±-hydroxy acids and dipeptides. Work on reengineer- ing the ribosome and the development of cell-free synthetic biology are enabling these accomplishments. The functionality and properties of polymeric materials, of course, depend on structures at length scales between molecular and macroscopic. Areas of polymer materials science in which very important advances have been made in the past decade have been in self-assembly, yielding supramolecular, noncovalent structures and the solidification processes of crystallization and vitrification. Self-assembly, broadly construed here, is the set of mechanisms via which information-encoding structures at shorter length scales direct structure formation at larger length scales. The size of a polymer chain depends on the molecular weight and ranges from ~5 nm to ~50 nm for large molar mass chains. In semicrystalline polymers and in block copolymers, which undergo self-assembly by microphase separation, there is structure on the length scale of tens of nm, owing to the crystalline lamellar thickness in the former case, and the self-assembled nanostructures in the latter. There is mesoscale structure in the form of chain-folded lamellae in semicrystalline spherulites, in grains of ordered block copolymers, and in component domains in phase-separated polymer blends. Understanding of structure evolution on all of these length scales, and its impact on bulk properties, have seen enormous advances in the past decade. Polymerization-induced self-assembly and phase separation, hybrid covalent- noncovalent polymerizations, and the introduction of selective intrachain interac- tions have all been demonstrated effectively to generate tertiary and quaternary structures, leading to useful functional properties in synthetic polymers. Self- assembled building blocks have been shown to assemble further into higher-order structures to create new hierarchically structured materials. Self-assembly has come to be viewed as a chemical-reaction-like process, moving with some kinetics from reactants to products. Nucleic acid self-assembly, which brings programmability and addressability not possible with hydrophobic or van der Waals forces, is evolv- ing from initial concepts to practical possibilities in information and biomedical technologies. Likewise, peptide (and peptoid) nanotechnology has exploited self- assembly of amino acid polymers for practical applications in tissue engineering and biomedical nanoparticles. Directed self-assembly (DSA), through the use of chemical or topological templates, has advanced notably in the past decade from scientific exploration to a viable industrial technique for microprocessor and storage-device fabrication. Spatial control of patterns on the nanoscale in a scal- able, parallel, and robust fashion now seems within reach. This evolution of DSA from experimentation to concrete invention in nanolithography was steered and
76 F r o n t i e r s o f M at e r i a l s R e s e a r c h accelerated enormously by a closely connected program in computational materials science, an indication of much more to come. Progress in the past decade in the solidification processes of crystallization and vitrification has included: (1) computer simulations that enable visualization of the early stages of these processes; (2) quantitative determination of how structural re- laxation is slowed, and glass-transition temperatures suppressed, within 100 nm of free and solid surfaces; (3) recent engineering advances with ultra-high-molecular- weight polyethylene that have led to state-of-the-art artificial hip replacements; and (4) time-resolved structural characterizations using X-ray scattering at synchrotron sources in combination with rheology or processing. Perhaps the most striking accomplishment in the area of glassy polymers has been the development of vitrimers, a remarkable new class of plastics that exhibit properties similar to silica glasses. Permanently cross-linked materials have out- standing mechanical properties and solvent resistance, but they cannot be pro- cessed and reshaped once synthesized. Non-cross-linked polymers and those with reversible cross-links are processable, but they are soluble. Vitrimers are networks that can rearrange their topology by exchange reactions without depolymerization and remain insoluble and processable. They are self-healing and can be shaped, welded, and repaired by techniques that would be familiar to glassblowers. They derive from and have many of the same properties as thermosetting plastics widely used in manufacturing and especially in the automotive and aircraft industries. Their discovery and development relate directly to a fundamental understanding of the differences between the glass transition in polymers and in conventional inorganic glasses. The fluidity results from bond exchange. A new bond is formed only at the expense and at the position of an existing bond. Figure 2.13 illustrates the chemical process schematically and explains some of the characteristics of vitrimeric materials. In silica glass, the bond exchange is catalyzed by the presence of boron or other elemental impurities. Vitrimers mimic this basic bond exchange mechanism by a judicious choice of functional ligands and catalysts. A compelling demonstration of this mechanism is the dependence of the viscosity and melting temperature on catalyst concentration. Originally, the bonds were ester linkages in which exchange was catalyzed by zinc catalysts. Chemistries producing vitrimer-type exchange have now been expanded to include polybutadiene rubbers (linkages based on alkene metathesis) and catalyst-free polyurethane systems. The glass-like proper- ties mean that plastics can be blown, twisted, and shaped while cooling much as their inorganic cousins. Two separate vitrimer rods can be welded together, or a break healed simply by holding the pieces together and heating. Bond exchange at the interface welds the two pieces together. At low temperature, the rate of bond exchange slows considerably, making the material a solid with essentially no flow and thus recovering the attractive properties of the thermosetting plastic.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 77 FIGURE 2.13â Top: In a vitrimer, the number of interpolymer bonds are kept fixed. Bonds are formed at the expense of other bonds being opened using catalyzed bond-exchange reactions. Although the material is, at all times, a 3D, covalently cross-linked network, heating allows malleability through the bond exchange process illustrated. Therefore, unlike cross-linked networks without bond exchange, vitrimers can be processed similarly to non-cross-linked glassy polymers, can be welded, can enable self-healing, and can be recycled. Bottom: Over a large temperature range, the material behaves as a viscoelastic melt before becoming a rigid glass on cooling. This malleability allows processing/shap- ing of polymers in much the same way as conventional silica glass, or as non-cross-linked glassy polymers. The creation of the shapes shown here, as a means of illustrating the high level of fluidity that can be induced in vitrimers, is demonstrated. Once formed, these objects become rigid, glassy objects on cooling. SOURCE: Top: Adapted from W. Denissen, J.M. Winne, and F.E. Du Prez, 2016, Vitrimers: Permanent organic networks with glass-like fluidity, Chemical Science 7:30-38; published by The Royal Society of Chemistry. Bottom: D. Montarnal, M. Capelot, F. Tournilhac, and L. Leibler, 2011, Silica-like malleable materials from permanent organic networks, Science 334:965-968, Â© CNRS PhotothÃ¨que/ESPCI/Cyril FRÃSILLON.
78 F r o n t i e r s o f M at e r i a l s R e s e a r c h Vitrimers could be viewed as a particular class of self-healing materials, which repair and restore their functionality without external intervention after a damage event. Self-assembly implies a thermodynamic tendency to self-healing, although being thermodynamically spontaneous by no means implies kinetically instanta- neous. Self-assembled block copolymers and hydrogels have demonstrated self- healing. Closely related are olymeric materials with dynamic covalent bonds. This class of polymers combines intrinsic reversibility with the robustness of covalent bonds, thus enabling formation of mechanically stable, polymerâbased materials that are responsive to external stimuli. External stimuli can trigger self-healing including changes in pH, UV light, electrical potential, and mechanical strainâ for example, the central CâC bond in diarylbibenzofuranone, which is known to reversibly cleave and reform under mild conditions of physical stress. In the emerging field of mechanochemistry, a new concept in self-healing, mechanical deformation events trigger changes in optical and mechanical properties before a critical damage event occurs through mechanically triggered chemical events. Yet another category of self-healing materials consists of those that contain an encap- sulated healing agent, and in some cases a microvasculature to distribute it, creat- ing polymer self-healing somewhat analogous to wound healing in humans. An example of this latter type of self-healing is highlighted in Box 2.9 in the form of a coating that is currently transitioning to a commercial product. Today, self-healing has been demonstrated in thermoplastics and thermosets, electronics, and batteries, and a variety of self-healing materials are entering the marketplaceâfor example, self-healing technology based on dynamic bonds is currently being commercial- ized by the chemical company Arkema. Self-healing polymers have the potential to reduce plastic waste both by prolonging the useful life of plastic-based products and by facilitating their recyclability.47 Catalytic chemistry in polymer materials science has been progressing enor- mously in the past 10 years. New catalysts for carbon-fiber reinforced polymer, for metathesis polymerization, and for controlled olefin polymerization have led to many useful, new polymeric materials. Catalysts for new electroactive conjugated polymers, with applications from solar energy to nonlinear optics, have been in- troduced. New catalysts have also been central to a growing body of work aimed at green polymer science. Important recent work in this arena includes the catalytic conversion of sugars to lactic acid using zeolites; use of CO2 for polymer synthesis, attractive given its prevalence and the new developments in the catalytic conver- sion of CO2 to polymers with interesting and attractive properties; and the use of nontoxic metals for the catalytic conversion of monomers to polymers. Chemical 47 See C. Kirby and T. Abate, 2016, âA Super Stretchy, Self-Healing Material Could Lead to Artificial Muscle,â Stanford Engineering Magazine, April 18, https://engineering.stanford.edu/magazine/article/ super-stretchy-self-healing-material-could-lead-artificial-muscle.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 79 BOX 2.9 Basic Science to Commercial Product: Self-Healing Coatings The basic science behind the self-healing material shown in Figure 2.9.1 was conducted at the University of Illinois, Urbana-Champaign, and the researchers launched the company Autonomic Materials, with the first product appearing in 2017. FI GURE 2.9.1 âThe left figure shows schematically the concept of a self-healing polymer coat- ing on a metal substrate. The yellow spheres contain the catalyst, the blue spheres are the self-healing phase-separated agent droplets, and the orange is the polymer coating. A scratch on the coating breaks the yellow spheres, releasing the catalyst and the self-healing agent. The scratch fills with a polydimethylsiloxane filler, which protects the base metal from exposure to the environment. The right figure shows the protection of the self-healing polymer coating (left panel) over a conventional coating (right panel) on a steel plate following 5 days of expo- sure of the scratched plate in a 5 percent NaCl solution. The background images are optical micrographs showing the scratch, and the enlarged images in the center are scanning electron microscope images showing details of the scratch. The polydimethylsiloxane has filled the scratch. SOURCE: S.H. Cho, S.R. White, and P.V. Braun, 2009, Self-healing polymer coatings, Advanced Materials 21:645-649, Â© 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.
80 F r o n t i e r s o f M at e r i a l s R e s e a r c h recycling, catalytically activated, back to the original monomers or to new products that have value has emerged as an area of research with outstanding potential. All of materials science, polymer science included, is founded on the interrela- tionships among composition-structure-processing-properties. Bringing process- ing into the science considerations means understanding what is happening at high stresses, strains, and speeds, leading material scientists to give increasing attention in recent years to nonlinear and nonequilibrium phenomena. In this realm in poly- mer science, the past decade has seen transformational advances in the nonlinear mechanical response of polymer hydrogels. They are rooted in improved chemical strategies for engineering toughness into networks from a molecular level. These advances open the door to new life-altering advances in artificial tissues. In the un- derstanding of polymer melt rheology, especially its nonlinear aspects, the past 10 years have seen some convergence among bead-spring, tube, and slip-link models, both in their physical interpretations and in predictions of behavior. Progress has also been made on the theoretical and computational front through the develop- ment and use of coarse-grained and field-theoretic approaches, which rely on the well-established fact that many polymer properties do not depend on the detailed atomic, or even segmental, structures of the chains. As in other areas of materials science and engineering, the advances in computational power and computational methods has had a significant impact on polymer science. The introduction of deep learning48 methods to polymer science is in its infancy but is expected to have a significant impact in the coming decade. Advances in computational methods are discussed in detail in Chapter 4. In addition to mechanical and rheological properties, transport properties of ions, protons, electrons, and phonons in polymers have come to the fore in this decade, owing in part to increasing interest in applied technologies such as mem- branes for separations, energy-related technologies such as batteries, fuel cells, organic energy conversion devices, and sound-dampening materials. Molecular dynamics simulations have played a role in the design of new materials for poly- meric electrolytes in batteries. Related to ion transport in polymers is the broad topic of polyelectrolytes, which de Gennes once termed the âleast understood form of condensed matter.â Owing to progress in the past decade, that statement is probably no longer apt. Not only have many aspects of the solution structure of individual polyelectrolyte chains been clarified, again with the help of advanced theory and computation, but also interactions with other ions in solutionâparticularly, multivalent metal ions and 48 âDeep learning,â inspired by information processing and communications patterns in biologi- cal systems, is a machine learning method based on analysis of data representations, as opposed to task-specific algorithms.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 81 oppositely charged polyionsâhave come to the fore. Polyelectrolyte brushes49 have been shown to create extremely slick surfaces in low concentrations of monovalent salt but become much less slippery in the presence of multivalent ions. Twenty years ago, the technique of layer-by-layer (LBL) growth of alternating polyanions and polycations was introduced by Decher in a Science paper that has now been cited more than 10,000 times.50 Consistent with observations about self-assembly and rheology above, LBL leads to interesting and useful but inherently nonequilibrium structures. Not only has LBL been the subject of considerable follow-on work itself, but also it has led to the deeper exploration of polyelectrolyte complexes in the past decade. The long-standing, ad hoc Voorn-Overbeek theory of polyelectrolyte complexation has been shown to be conceptually inadequate and quantitatively inaccurate. Polyelectrolyte complexation is also a new driving force for block co- polymer self-assembly, with the need for caveats about possible nonequilibrium structures still applicable. Nanocomposite materials have come to the fore in the past 10 yearsâfor ex- ample, carbon nanotube and graphene composites with various polymers addedâ as they decrease small molecule transport and simultaneously increase electrical conductivity relative to the pure polymer. Cellulose nanocrystals can introduce hydrophilic channels in otherwise hydrophobic polymers. With respect to me- chanical properties, orders of magnitude in strength, with a concomitant increase in electrical conductivity, have been reported. Novel electric, magnetic, optical, and transport properties have also been demonstrated. Achievement of many of these gains have been enabled by enhanced ability to manipulate interfacial physics and chemistry in polymeric materials. Hybrid-bonding polymers are an important emerging class of macromolecu- lar soft materials in which the bonding among monomeric units is covalent or noncovalent in different nanoscale domains of their structure. These materials are therefore composed of covalent and supramolecular polymers and can be synthe- sized through various pathways, which includes the methodology in which covalent and supramolecular polymerizations occur simultaneously. This approach dem- onstrated the possibility to âcatalyzeâ covalent polymerization using noncovalent interactions of a forming supramolecular polymer with growing covalent chains. In this context, research in this area could reveal the principles behind ribosomal synthesis of proteins, which is definitely aided by noncovalent interactions in the environment of the growing chain. Earlier efforts had often focused on the covalent capture of ordered supramolecular polymers to create covalent polymers, or post- polymerization modifications via noncovalent bonding of small molecules without 49 Polyelectrolyte brushes are long polyelectrolyte chains densely affixed to a surface. 50G. Decher, 1997, Fuzzy nanoassemblies: Toward layered polymeric multicomposites, Science 277(5330):1232-1237, doi:10.1126/science.277.5330.1232.
82 F r o n t i e r s o f M at e r i a l s R e s e a r c h formation in either case of an actual supramolecular polymer within the structure. The dual nature of bonding among structural units introduces the potential for highly dynamic properties in polymeric soft matter that originate in the noncova- lent bonding of supramolecular polymers, integrated with the mechanical robust- ness of covalent polymers. Examples of dynamic properties include fast self-healing of defects, rapid response to external stimuli, enhanced rates of biodegradation, and new opportunities in recycling, toughening mechanisms, and electrical actuation. 2.5.2 Biomolecular and Bio-Inspired Materials Research on soft matter outside conventional covalent polymers over the past decade has been dominated by the area of self-assembling materials, most com- monly supramolecular polymers, organogels, DNA and peptide nanotechnology, supramolecular nanostructures, 2D materials, as well as metal-organic and covalent organic frameworks. Research on self-assembling materials has been motivated by great interest in bio-inspired systems as a rich source of ideas for novel functions in soft matter. This trend has come to define the field of âbiomolecular materials,â in which chemical structures found in biological systems are integrated into synthetic systems, or biological structures at all scales and their functions are emulated with abiotic chemistry. In this context, the broad field of biomolecular materials has been an important direction in soft materials research over the past 10 years. Advances in this field over the past decade have benefited enormously from novel synthetic strategies to create organic materials and characterization tools for soft matter, which include high-resolution microscopies, measurements on molecular dynamics, and scattering techniques. It has also become obvious that computational techniques, either coarse-grained or atomistic, are being established as an integral part of this research area. A common element in biomolecular ma- terials research over the past decade has been the search for the functional power of ordered structures obtained through self-assembly strategies. Another nascent objective has been the exploration of dynamic behavior in soft matter. This last goal connects to the objective of discovering materials with autonomous capacity for self-healing of defects, or the capacity to actuate or move in response to stimuli, thus emulating living organisms. Thermal energy and light have been used as common stimuli to develop this behavior in soft matter. Systems exhibiting dimen- sional and shape changes in response to an energy input are currently perceived as important for integration into smart systems useful in medicine, manufacturing, wearables, and robotics, among others.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 83 2.5.3 Biomaterials The global market of medical implants presently exceeds $100 billion, and the optimal performance of materials in this space has never been more critical given the demographic changes that have occurred worldwide. The most significant changes have been the great increase in aging populations, and the strong cultural aspiration to achieve the highest possible quality of life throughout life spans ap- proaching three-digit numbers. There is therefore great awareness that permanent implants used in orthopedic surgery, stents, and dental implants must perform optimally for longer periods of time. At another extreme, the field of regenerative medicine, seeking to rebuild tissues and organs lost to trauma, disease, aging, and congenital defects, has established itself over the past decade as an exciting global grand challenge and offers an avenue for addressing organ donor shortages. The 2014 book by D. Williams, Essential Biomaterials Science, reinforces these state- ments and nicely summarizes progress to that point. Advances in metallic, ceramic, and soft biomaterials over the past decade have significantly advanced efforts to address these needs. Metallic Biomaterials In the field of metallic biomaterials, there have been a number of active research areas that focus on improving implants. One direction has been the integration of secondary functions in metals beyond the load-bearing structural role of the parts. Examples include the modification of surfaces to render them antimicrobial, since preventing infection at tissue-implant interfaces remains an important challenge, and also improving methods to create porosity in order to promote bone ingrowth in orthopedic implants. For this particular purpose, development of effective meth- odologies to create metallic foams have been of interest recently. In the case of antimicrobial surfaces, for reasons that are not yet clear, silver nanostructures have been found effective at preventing bacterial infections. Attempts to understand the origin of antimicrobial properties of silver nanostructures have focused on the dissolution of silver ions from the nanostructures and the interactions of these metal ions with peptides, which are of course present in bacterial membranes. However, the specific mechanism behind this very important function of metallic nanostructures remains unknown. A different approach to promote bone ingrowth in metallic implants for fixation to the tissue has been the use of âosteoinductiveâ coatings on metals. The use of calcium phosphate ceramics such as apatites has been a common approach, but other efforts have contemplated the functionalization of metallic surfaces with biological macromolecules such as growth factors. In the case of cardiovascular biomaterials, there has been interest over the past decade
84 F r o n t i e r s o f M at e r i a l s R e s e a r c h and even earlier in improving methods to coat drug-eluting polymeric coatings on surfaces of metallic stents with the purpose of avoiding restenosis of blood vessels. A different area of interest in metallic biomaterials over the past decade has been âbiodegradableâ biometals. Examples of such metals include magnesium, iron, and zinc, since these metals are part of natural processes in mammalian biology. Zinc, for example, is found in hundreds of enzymes, and its presence is critical to their catalytic functions. There is also evidence that metals such as magnesium can have beneficial effects on the regeneration of bone. The use of iron alloys for stents have faced some challenges in animal models related to poor control of degradation rate, but at the same time their magnetic nature could help to stimulate cells positively through particulate displacement with external fields. The use of bioÂ ompatible biodegradable metals has remained an attractive objec- c tive over the past decade that will require moving forward a significant amount of basic science research. Access to fully biodegradable metallic parts would be particularly attractive in the context of bone regenerative medicine applications where the structural support of a metallic scaffold would be fully replaced by new tissue formation. Advances of the past decade have also shown that implantable electronic devices could in principle be developed using biodegradable metals. The area could have impact in other technologies outside biomaterialsâfor example, in strategies to create degradable and more easily recyclable electronic devices. Last, a major area of development in metallic biomaterials over the past decade is that of AM of implants. Techniques such as selective laser sintering of metallic powders, cryo-milling, and printing to create designed implants with tailored micro- and nanostructure are of great interest in personalized medicine. Ceramic Biomaterials Over the past decade, a major area of research in ceramic biomaterials involves the use of calcium-phosphate-based materials with applications in the area of bone implants. Many compositions have been explored and methodologies devel- oped to coat metallic implants. These ceramic materials have the capability to be osteoconductive, creating stable interfaces with living bone, and are also broadly biocompatible. Mirroring activity in metallic biomaterials, there has been great interest in developing strategies to create implants using AM techniques. As is the case for metals, AM offers the opportunity to customize implants to patients and, most importantly, to design macroscale and microscale architectures to optimize biological functionality.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 85 Soft Biomaterials Advances in polymer science of the past decade have impacted the field of soft biomaterials by providing novel approaches to improved physical properties and integration of function. In terms of physical properties, the field of soft biomaterials has benefited from progress in the area of hydrogel networks that are of interest in tissue engineering. Specifically, new chemical and processing approaches to the me- chanical reinforcement of hydrated macromolecular networks, via interpenetrating networks or the preparation of hybrid nanocomposites with internal reinforcement, offer the possibility to use them not only as transient scaffolds to template the growth of tissues but also as permanent artificial tissues in areas where regenera- tion is still distant. One example would be the replacement of intervertebral discs, which is a great societal need given the high incidence of degenerative disc disease. Related to this area, fundamental advances in the understanding of polyelectrolytes and coacervate phases offer the possibility to develop lubricious surfaces that may be useful in the development of artificial joints and catheters. In terms of function integration in soft biomaterials, the popular area of layer-by-layer materials has been used to integrate within individual layers different types of bioactivity as well as drug delivery functions. The past decade has seen an enormous advance in our ability to use and control self-assembly processes to create such new biomaterials. The most significant development in soft biomaterials over the past decade occurred in the area of âsupramolecular biomaterials.â Covalent polymers, nonbio- degradable or biodegradable, have been the materials of choice for soft biomateri- als (see Figure 2.14). An important research shift of the past decade in this field is the exploration of supramolecular polymers as biomaterials. In supramolecular polymers, bonding among monomeric units involves weak noncovalent interac- tions, thus creating many exciting new directions in soft materials. For example, the binding energies among monomers become tunable over a broad range, and the end result is a broad platform of soft materials in which the bond lifetimes can vary significantly. With bond lifetimes among monomers spanning microseconds to seconds, there is great potential for control of dynamic properties and new pro- cessing opportunities. Responsiveness to complex and interacting stimuli and re- configurability are now being enabled by some of the advances in polymer science. By far, the most popular chemistry in supramolecular biomaterials over the past decade has been the use of peptides and modified peptides such as peptide amphiphiles as the building blocks. Peptide supramolecular assemblies can com- pete with designed proteins to offer useful biological functions and a high diversity of structures. From a structural perspective, peptide assemblies can generate fila- ments, 2D materials, cross-linked networks, spheres, tubes, helical structures, and materials with higher structural complexity as researchers learn to master chemi- cal morphogenesis. Thus, the motivation for materials designed through peptide
86 F r o n t i e r s o f M at e r i a l s R e s e a r c h FIGURE 2.14â General structure of a peptide amphiphile (PA, center) surrounded by many of the supramolecular nanostructures that have been formed from these molecules to create supramolecular biomaterials. These were the first examples of supramolecular biomaterials exhibiting high levels of bioactivity in cell signaling for regenerative medicine, biocompatibility, and rapid biodegradation in vivo. SOURCE: M.P. Hendricks, K. Sato, L.C. Palmer, and S.I. Stupp, 2017, Supramolecular assembly of peptide amphiphiles, Accounts of Chemical Research, 50(10):2440-2448, https://pubs.acs.org/ doi/10.1021/acs.accounts.7b00297, Â© 2017 American Chemical Society. nanotechnology has been to create biomaterials that could be potentially bioactive, designed rationally to signal cells for tissue regeneration, and also with controlled structures for mechanics and delivery of therapeutic components. Peptide-based materials can also be programmed for self-assembly in water to create biomimetic extracellular matrices. The approach has been extremely success- ful in demonstrating the ability of these new bioactive and biodegradable materials to trigger regeneration of a wide variety of tissues. The exciting functional capacity of these supramolecular biomaterials seems to be partly based on their ability to form biomimetic nanofibers while accommodating a large diversity of signals that are displayed to cells. Furthermore, the supramolecular nature of this new type of biomaterial as opposed to conventional polymers has introduced the capability of rapid and full biodegradation after a âproductiveâ short-lived bioactive interaction with cells.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 87 Another important new area in soft biomaterials over the past decade has been the use of DNA to utilize the great fidelity of Watson-Crick pairing among nucleotides in order to generate designed structures across the scales. This is an emerging field, and many advances can be expected in the future with this approach to develop novel biomaterials. Soft biomaterials are also beginning to be impacted by the development of AM techniques in order to control macroscopic form and microscopic architecture. Printing also brings the possibility of creating constructs that are hybrids of soft biomaterials and cells, allowing the localization of cells in specific compartments of a 3D structure.51 It is worth mentioning that there has been a recent trend to the formation of Current Good Laboratory Practice (cGLP)52 and Current Good Manufacturing Practice (cGMP)53 facilities in several academic medical centers and universities. By and large, these facilities manufacture novel cellular, biomaterial, viral, and molecular materials for first-in-human use, general use (especially for cellular therapeutics), and treatments for rare diseases. As work progresses in the materials research area, and particularly in fields using DNA and similar biological materials, it will become increasingly important to be aware of the cGLP and cGMP efforts. 2.5.4 Soft Matter Colloids Soft matter54 is best described as matter with a small or vanishing elastic modulus that governs a materialâs ability to withstand deformation. In addition to many forms of polymers, it includes colloids, foams, emulsions, and granular and active matter composed of entities much larger than atoms but still much smaller than the sample size and arranged on a mesoscopic scale. Energy scales are often of the same order of magnitude as the thermal energy, but certainly no larger than those of atomic solids. Their large length scales thus require small elastic moduli, which scale as energy divided by volume. Soft matter is often strongly dissipative, 51 Another example is virus templating; see K.M. Bromley, A.J. Patil, A.W. Perriman, G. Stubbs, and S. Mann, 2008, Preparation of high quality nanowires by tobacco mosaic virus templating of gold nanoparticles, Journal of Materials Chemistry 18(40):4796-4801. 52 Food and Drug Administration, 1981, Guidance for Industry: Good Laboratory Practices, Ques- tions and Answers (revised December 1999 and July 2007), U.S. Department of Health and Hu- man Services, https://www.fda.gov/downloads/ICECI/EnforcementActions/BioresearchMonitoring/ UCM133748.pdf. 53 Food and Drug Administration, âFacts About the Current Good Manufacturing Practices (CGMPs),â updated June 25, 2018, https://www.fda.gov/drugs/developmentapprovalprocess/manu- facturing/ucm169105.htm. 54 S.R. Nagel, 2017, Experimental soft-matter science, Reviews of Modern Physics 89(2):025002.
88 F r o n t i e r s o f M at e r i a l s R e s e a r c h disordered, entropy dominated, far from equilibrium, and characterized by strong nonlinear and slow dynamical responses. Although these properties have been extensively studied over the past several decades, our understanding of them is still incomplete. This section will review representative advances in nonbiological soft matter (other than polymers). Colloidal science has undergone a revolution over the past decade or so. Whereas in the past, colloids were mostly restricted to micron-scale spherical particles suspended in an isotropic fluid and interacting through a few elementary forcesâsteric repulsion, electrostatic, attractive van der Waals, and depletionâit is now routine to create particles of essentially any shape (e.g., the letters of the al- phabet) with specific directional, reversible, and irreversible interactions dispersed in anisotropic liquid crystals in addition to isotropic fluids. Among these are Janus particles with inhomogeneous surface properties that control self-assembly of colloidal molecules and even open structures such as kagome and diamond-like lattices, lock-and-key particles, and rod-like biopolymers such as actin or micro- tubules, with controlled chirality, that mimic the idealized rods of the Onsager theory of liquid crystal order. The latter system forms large unilamellar membranes in the presence of depletants. These advances open new ways to probe and control phenomena like defect production in liquid crystals,55 self-assembly, and jamming. Clever synthesis and technologies adapted from other emerging areas such as DNA nanotechnology have contributed to the growth of colloidal science. For example, colloids with valence and directional bonding have been created56 by attaching single-stranded DNA to patches on patchy colloids. Hybridization with similar particles with complementary DNA single strands on their patches allows specific association and a new type of directional colloidal interaction mediated by the DNA. When densely packed finite clusters of spherical particles are swollen with an appropriate solvent and then polymerized, they form the aforementioned patch colloids, as shown in Figure 2.15. The specificity and design facility of DNA hybridization is a useful new tool even for spherical particles. Liquid Crystals The most common liquid crystals are formed by molecules with oily tails that avoid water. There are, however, hydrophilic molecules that form a liquid crystal phase when dissolved in water at sufficiently high concentration. Among these are 55 I.I. Smalyukh, 2018, Liquid crystal colloids, Annual Review of Condensed Matter Physics 9:207-226. 56 Y.F. Wang, Y. Wang, D.R. Breed, V.N. Manoharan, L. Feng, A.D. Hollingsworth, M. Weck, and D.J. Pine, 2012, Colloids with valence and specific directional bonding, Nature 491(7422):51-U61.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 89 FIGURE 2.15â Schematic of the formation of DNA-stranded patchy colloids. SOURCE: Reprinted by permission from Springer Nature: Y.F. Wang, Y. Wang, D.R. Breed, V.N. Manoharan, L. Feng, A.D. Hollingsworth, M. Weck, and D.J. Pine, 2012, Colloids with valence and specific directional bonding, Nature 491(7422):51-U61, Â© 2012. chromonic liquid crystals57 formed from flat molecules, which include an asthma drug, dies, and DNA nucleotides, with aromatic cores that aggregate in stacks that align to form nematic or columnar phases. Only recently have the viscosities and Frank elastic constants of chromonics been measured.58 Somewhat surprisingly, the twist constant is of order 6 to 11 times smaller than the bend constant, and it ap- pears the elusive saddle-splay constant, whose contribution to the energy integrates to the boundary, is of order 50 times larger than the twist constant.59 The result is that chromonics in confined geometries (as in a spherical micelle or a cylindrical capillary) develop unusual conformations.60 Oriented dried thin films of chromon- ics exhibit semiconducting properties. Applications involving chromonics include polarizing films, biosensors, and controlled self-assembly of nanorods. The ferronematic liquid crystal phase, the only known liquid that exhibits true long-range ferromagnetic order coexisting with nematic order, was discovered in the past decade.61 The original version of this phase was created by dispersing fer- romagnetic nanodiscs with magnetic moments normal to their flat planes into a standard nematic. The nanodiscs aligned with their magnetic axes parallel to each 57 H.S. Park, and O.D., Lavrentovich, 2012, âLyotropic Chromonic Liquid Crystals: Emerging Ap- plications,â pp. 449-484 in Liquid Crystals Beyond Displays: Chemistry, Physics, and Applications (Q. Li, ed.), Wiley & Sons, Hoboken, N.J. 58 S. Zhou, 2017, âElasticity, Viscosity, and Orientational Fluctuations of a Lyotropic Chromonic Nematic Liquid Crystal Disodium Cromoglycate,â pp. 51-75 in Lyotropic Chromonic Liquid Crystals, Springer, Cham. 59 Z.S. Davidson, L. Kang, J. Jeong, T. Still, P.J. Collings, T.C. Lubensky, and A.G. Yodh. 2015, Chi- ral structures and defects of lyotropic chromonic liquid crystals induced by saddle-splay elasticity, Physical Review E 91(5):050501. 60 J. Jeong, L. Kang, Z.S. Davidson, P.J. Collings, T.C. Lubensky, and A.G. Yodh, 2015, Chiral struc- tures from achiral liquid crystals in cylindrical capillaries, Proceedings of the National Academy of Sciences U.S.A. 112(15):E1837-E1844. 61 A. Mertelj, D. Lisjak, M. Drofenik, and M. ÄopiÄ, 2013, Ferromagnetism in suspensions of mag- netic platelets in liquid crystal, Nature 504(7479):237.
90 F r o n t i e r s o f M at e r i a l s R e s e a r c h other and parallel to the nematic anisotropy axis. An alternative version of the ferronematic was created by dispersing the magnetic nanoparticles in a nonpolar solvent that did not appreciably screen the Coulomb interaction between them.62 Last, a biaxial ferronematic63 in which a surface treatment of the nanomagnets forced the magnetic moments to align at an angle to the nematic director specifying the direction of uniaxial anisotropy. Because of the strong coupling to the direc- tor, their magnetic excitations are overdamped and do not exhibit the traditional dispersion with frequency proportional to wave-number squared. In the past decade, there have been advances in the use of liquid crystals for chemical and gas sensors. This is based on the disruption of the orientations of liquid crystal at either the liquid crystal/solid or the liquid crystal/air interface, caused by the analyte. Advances in the atomic/molecular scale design of chemo- responsive systems that resulted in the improvement in the sensor performance (selectivity and sensitivity to a given analyte and system response time) was shown to occur by tailoring the properties of the liquid crystal such that they were oriented parallel rather than perpendicular to that of the free surface while anchored in a perpendicular orientation at a surface decorated with metal cations. An example of this advance in sensor performance is shown in Figure 2.16.64 Also, modern computational chemistry methods have demonstrated great po- tential toward designing efficient chemoresponsive systems at the molecular scale and in a time frame much smaller than the traditional trial and error approach based on experiments alone. Granular Materials Granular materials are aggregates of individually solid grains that interact with their neighbors via contact forces. Their properties do not depend significantly on temperature, and they do not explore the many internal configurations that thermal equilibrium systems do (they are nonergodic). Granular materials are everywhereâ on sandy beaches, in loose dirt on mountain faces, in coal hoppers, in grain elevators, and in medicine bottles. They are important to industrial processing, and are found in new applications in fields such as robotics. Yet, in spite of their ubiquity and 62 M. Shuai, A. Klittnick, Y. Shen, G.P. Smith, M.R. Tuchband, C. Zhu, R.G. Petschek, et al., 2016, Spontaneous liquid crystal and ferromagnetic ordering of colloidal magnetic nanoplates, Nature Communications 7:10394. 63 Q. Liu, P.J. Ackerman, T.C. Lubensky, and I.I. Smalyukh, 2016, Biaxial ferromagnetic liquid crys- tal colloids, Proceedings of the National Academy of Sciences U.S.A.113(38):10479-10484. 64 K. Nayani, P. Rai, N. Bao, H. Yu, M. Mavrikakis, R.J. Twieg, and N.L. Abbott, 2018, Liquid crys- tals with interfacial ordering that enhances responsiveness to chemical targets, Advanced Materials 30(27):1706707.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 91 FIGURE 2.16âSchematic comparison of the response of 4-cyano-4â²-pentylbiphenyl (5CB) and a mixture of 4-(trans-4-pentylcyclohexyl)benzonitrile (PCH5) and the homologue 4-(trans-4-propylcy- clohexyl)benzonitrile (PCH3) to exposure to an analyte such as dimethyl methylphosphonate, a nerve agent stimulant. (A) Director profile of the liquid crystal, 5CB, with homeotropic anchoring at the free surface, prior to and after exposure to the analyte. (B) Director profile of PCH5/PCH3 with coordination interaction with metal-salt-decorated surface and a planar anchoring at the free interface, prior to and after exposure to the analyte. (C) The response time of the two sensors to different concentrations of the analyte. The PCH5/PCH3 liquid crystal sensor shows improved sensitivity and faster response time than 5CB. SOURCE: K. Nayani, P. Rai, N. Bao, H. Yu, M. Mavrikakis, R.J. Tweig, and N.L. Abbott, 2018, Liquid crystals with interfacial ordering that enhances responsiveness to chemical targets, Advanced Materials 30(27):1706707, Â© 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. importance, they continue to be poorly understood. Research over the last 10 years, however, has improved our ability to control and utilize granular materials. One of the most remarkable properties of granular materials is their ability to transform from a fluid to a solid state merely by the application of pressure, or equivalently by increasing the volume fraction of constituent particles. A bag of unjammed granular particles can easily spread around and engulf objects of the appropriate size. When air is sucked out of the bag, the density of particles increases until the jamming transition is reached. An object, now strongly clamped, can be displaced at will. The result is a universal gripper,65 a prototype material with tun- able adaptive compliance that can grab and move objects of any shape with obvious uses in robotics. Grains in granular materials can be of any size from the submicron (e.g., colloi- dal particles) to the meter scale and more (e.g., boulders). An application at the mi- cron scale is to force vesicles, which prefer to be spherical, to adopt nonequilibrium 65 E. Brown, N. Rodenberg, J. Amend, A. Mozeika, E. Steltz, M.R. Zakin, H. Lipson, and H.M. Jaeger, 2010, Universal robotic gripper based on the jamming of granular material, Proceedings of the National Academy of Sciences U.S.A. 107(44):18809-18814.
92 F r o n t i e r s o f M at e r i a l s R e s e a r c h shapes by absorbing particles on their surface at sufficient density that they jam, and thereby develop a nonvanishing shear modulus.66 Much of the work on granular materials has focused on jamming of spherical frictionless particles, but recent studies have emphasized the importance of fric- tion. The phase diagram of dry frictionless spheres as a function of external shear stress and granular volume fraction shows unjammed and isotropically jammed phases (Figure 2.17a), whereas that of particles with friction shows two other phases (Figure 2.17b): a fragile phase that resists forces along one dimension (in 2D) but not the other and a shear-jammed phase that resists stresses along both directions but with greater strength along one direction than the other.67 Work such as this on granular materials and jamming has had a significant impact on understanding of the mechanical and flow properties of dense suspensions. Packing of Hard Objects The structure of liquids, solids, and glasses is intimately related to the volume fraction of ordered and disordered (random) hard-sphere packings. Simulations have shown that hard tetrahedra form a columnar quasicrystal with 12-fold sym- metry with a packing fraction of 0.85 compared to 0.74 for spheres packed in a FIGURE 2.17â Phase diagrams for granular matter as a function of particle volume fractions Ï and shear stress, Ï. Left: Spherical frictionless particles showing jammed and unjammed regions and the zero-stress jamming point at ÏJ. Right: Spherical particles with friction showing the unjammed, jammed, fragile (F) and shear-jammed (SJ) regions. SOURCE: Reprinted by permission from Springer Nature: D. Bi, J. Zhang, B. Chakraborty, and R.P. Behringer, 2011, Jamming by shear, Nature 480:355- 358. Â© 2011. 66 M.M. Cui, T. Emrick, and T.P. Russell, 2013, Stabilizing liquid drops in nonequilibrium shapes by the interfacial jamming of nanoparticles, Science 342:460-463. 67 D. Bi, J. Zhang, B. Chakraborty, and R.P. Behringer, 2011, Jamming by shear, Nature 480(7377):355.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 93 face-centered cubic crystal structure (FCC) lattice. This tetrahedral packing frac- tion is noticeably higher than initial analytic estimates. A version of this lattice composed of nanoparticle square and triangular elements with targeted attractive interactions has been created.68 There is now an exhaustive âperiodic tableâ of the phases produced by different-shaped hard particles based on directional entropic forces. There are of course an infinite variety of shapes, and there has been some effort to identify the shape that produces the highest random-packing density. An exploration of packing of particles consisting of overlapping spheres yielded a triangular particle with a random packing fraction of 0.73,69 almost as high as that of spheres on an FCC lattice (see Figure 2.18). Particles with more exotic shapes, such as that of the letter Z (or Y, as discussed earlier), have interesting properties. They jam under their own weight, and as a result, they have entered the imagination of architects, who have built stable struc- tures such as an arch. Alternatively, string tethers can provide sufficient targeted compression to stabilize granular packing of architectural-size particles. This new field now goes under the heading of aleatory architecture. Properties of materials often result from the different forms of order, or broken symmetry, that they exhibit. Examples are the rigidity of crystals arising fromÂ their periodic structure and the optical properties of liquid crystals arisingÂ from FIGURE 2.18â Left: Configuration of tetrahedra in the 12-fold quasicrystal. Right: Local configura- tion of tetrahedral and the columnar structure they form. SOURCE: Reprinted by permission from Springer Nature: A. Haji-Akbari, M. Engel, A.S. Keys, X. Zheng, R.G. Petschek, P. Palffy-Muhoray, and S.C. Glotzer, 2009. Disordered, quasicrystalline and crystalline phases of densely packed tetrahedra, Nature 462:773-777, doi: 10.1038/nature08641, Â© 2009. 68 X. Ye, J. Chen, M.E. Irrgang, M. Engel, A. Dong, S.C. Glotzer, and C.B. Murray, 2017, Quasicrys- talline nanocrystal superlattice with partial matching rules, Nature Materials 16(2):214. 69 L.K. Roth and H.M. Jaeger, 2016, Optimizing packing fraction in granular media composed of overlapping spheres, Soft Matter 12:1107-1115.
94 F r o n t i e r s o f M at e r i a l s R e s e a r c h molecular alignment in a particular direction.70 A more subtle form of order, called hyperuniformity,71 is characterized by reduced density fluctuations. If particles are randomly distributed in a space of d-dimensions, the mean-square fluctuation in the number of particles in a sphere of radius R scales as the average number of points in a sphere or as N~ Rd. In hyperuniform systems, these fluctuations grow as Rs, where s lies between d and dâ1. Hyperuniformity can be achieved in disordered systems. Calculations and experiments show that hyperuniform structures have photonic bandgaps.72 While it is not yet clear whether hyperuniformity is required for the existence of bandgaps, it is now clear that complete bandgaps are possible without periodicity. This opens a fundamental question as to what properties are necessary for making bandgap materials. Hyperuniform disordered structures do not exist in equilibrium materials with short-range interactions. They have been produced out of equilibriumâfor example, in jammed systems at the jamming point and in driven systems such as periodically sheared-random organization dynamics, at their critical point. Hyperuniform patterns can be generated on a computer and additively manufactured. Soft 2D Materials Two-dimensional sheets are an interesting form of soft matter because of their ability to deform into the third dimension with very little energy cost. Atomically thick monolayer graphene is probably natureâs most perfect 2D material. It is best known for its intriguing electronic properties, but it also has very interesting elas- tic properties that were measured for the first time in the past decade.73 At zero temperature, these properties are controlled by the 2D Youngâs modulus measuring resistance to in-plane strain and the bending stiffness measuring resistance to out- of-plane bending; the former yielding a 3D Youngâs modulus for a stack of sheets three orders of magnitude greater than that of steel. Graphene is much like paper: it is hard to stretch but easy to bend. A quantitative measure of the relative strength of stretch to bend is the Foppl von-Karman number Î½K. A large Î½K implies easy folds. A 20 cm square piece of paper has Î½K~107, whereas the same square made of graphene at zero temperature has Î½K~1011. For graphene, unlike paper, thermal fluctuations are important. Thermally excited bending modes lead to substantial 70 P.Chaikin and T. Lubensky, 1995, Principles of Condensed Matter Physics, Cambridge University Press, Cambridge, U.K. 71 S. Torquato and F.H. Stillinger, 2003, Local density fluctuations, hyperuniformity, and order metrics, Physical Review E 68(4, Pt. 1):041113. 72 M. Florescu, P.J. Steinhardt, and S. Torquato, 2013, Optical cavities and waveguides in hyperuni- form disordered photonic solids, Physical Review B 87:165116.â¨ 73 M.K. Blees, A.W. Barnard, P.A. Rose, S.P. Roberts, K.L. McGill, P.Y. Huang, A.R. Ruyack, et al., 2015, Graphene kirigami, Nature 524(7564):204.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 95 length-scale-dependent renormalization of both the bending and Youngâs moduli. It is easy to see how these modes affect the bending modulus: corrugations in a flat sheet make it harder to bend the sheet. Measurements of the effective bending stiffness of a strip of graphene in a water shows a 4,000-fold enhancement relative to its bare zero-temperature value, in very good qualitative agreement with the ther- mal enhancement theory, but they do not rule out the possibility that quenched-in height fluctuations or other mechanisms might contribute to this enhancement. A second class of 2D sheets that deform into the third dimension are rubbery materials that have a much smaller Foppl-von-Karman number than graphene but can nonetheless fold easily. What is interesting about these sheets is that 2D patterning on them can produce controlled shape change upon change of external condition like temperature, chemical potential, or light intensity. In one version, a flat circular sheet of an N-isopropylacrylamide (NIPA) gel is prepared with a radial gradient of cross-link density. When heated, NIPA gels undergo a sharp reversible transition at a temperature Tc = 33Â°C, characterized by a volume reduction depen- dent on the density of cross-links. Thus, upon heating, an inhomogeneously cross- linked sheet will undergo differential shrinking that causes it to bend out of plane. The final 3D shape can be calculated with good accuracy from knowledge of the initial inhomogeneous cross-link density. An alternative version of this procedure is a halftone process (such as that used in newspaper photos) in which circular dots of higher rigidity are introduced on a lower-rigidity background sheet to produce inhomogeneous swelling factors controlled by the dot density. The uniaxial anisotropy of nematic elastomers and glasses (basically cross- linked nematic polymers) provide a different avenue for printing metric tensors on a flat sheet. Upon heating (cooling) these materials shrink (expand) in directions along the local director and expand (shrink) along the opposite direction. Heating thus causes initial flat sheets to deform into the third dimension imposed by their original imprinted director configuration. For example, a sheet with an implanted +1 disclination, in which the director follows contours that encircle the defect core, as shown in Figure 2.19, will stretch along its circumference but expand along its radii and deform to a cone. An extended version of the periodic structure shown in Figure 2.19 (c) can lift up to 100 times its own weight without buckling. As in the case of kirigami, algorithms,74 which have been experimentally verified, have been developed that specify two-dimensional director configurations that will yield any 3D shape, including, for example, a face.75 74 C.Modes and M. Warner, 2016, Shape-programmable materials, Physics Today 69(1):32. 75Aharoni et al., 2018, Universal inverse design of surfaces with thin nematic elastomer sheets, Proceedings of the National Academy of Sciences U.S.A. 115:7206.
96 F r o n t i e r s o f M at e r i a l s R e s e a r c h FIGURE 2.19â Kirigami example using nematic elastomers and glasses. (a) Patterns with director tangent to lines. (b) Closed patternâa configuration topologically equivalent to the +1 defect shown in (d). Upon heating, this pattern shrinks parallel to the director and expands in the perpendicular direction to form the pyramid in which sharp corners eventually relax. (c) A more complex director pattern, and (d) the resultant experimental 3D structure, which can support more than 100 times its own weight. (e) and (f) +1 and â1 topological defects, respectively. SOURCE: Panels (a-c, e-f) are reproduced with permission from C. Modes and M. Warner, 2016, Shape-programmable materials, Physics Today 69(1):32-38, Â© 2016 American Institute of Physics, https://doi.org/10.1063/PT.3.3051. Panel (d) is from T.H. Ware, M.E. McConney, J.J. Wie, V.P. Tondiglia, and T.J. White, 2015, Voxelated liquid crystal elastomers, Science 347(6225):982-984; reprinted with permission from AAAS. Active Matter Active-matter materials are composed of homogeneously distributed self- driven units capable of converting stored or ambient free energy to systematic movement. The interaction of active particles with each other, and with the me- dium they live in, gives rise to highly correlated collective motion and mechanical stress.76 These systems are fundamentally out of equilibrium, and there is no ther- modynamic maximum entropy principle to control their properties. As a result, even their steady-state behavior is often exotic. Biological cells are the quintessential form of active matter, but a variety of other forms, such as schools of fish, bacterial baths, shaken 2D granular gasses, various adenosine diphosphate-driven artificial systems, are currently being studied. Active matter is characterized by the symmetry of its active constituents (polar or apolar for filamentous particles or simply disc-shaped or spherical particles) and by the medium in which they move (e.g., on a dry substrate, or in a liquid host, 76See T.B. Liverpool, 2013, âBGERâVisits to Research Groups in the departments of Chemical Engineering, Evolutionary Biology and Mechanical Engineering at Princetonâ, https://research-infor- mation.bristol.ac.uk/en/activities/bger--visits-to-research-groups-in-the-departments-of-chemical- engineering-evolutionary-biology-and-mechanical-engineering-at-princeton%283ad890cb-55f1- 4664-8695-ff49e6f21246%29.html.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 97 possibly a viscoelastic, fluid). Hydrodynamic theories for polar and apolar media have been developed and applied to both model âin vitroâ systems and to living cells. They have predicted various properties of active matter such as spontaneous flow, rheological response, and the formation and annihilation of topological de- fects. More microscopic models with specific interactions between self-propelled particles or that include molecular motors like myosin as the source of active mo- tion have also proven to be very useful especially for numerical simulation (see Figure 2.20). Experiments have investigated various types of self-propelling units, including bacteria and particles on vibrating substrates. Of particular interest are those using molecular motors. In one version,77 a planar substrate is covered with immobilized molecular motors, which are fueled by adenosine triphosphate, that propel the actin filaments in the fluid above. Above a critical filament density, this system FIGURE 2.20â Experiments with bacterial colonies in biocompatible chromonic nematics on above patterned surfaces. (A) Externally imposed director pattern with a single â1 defect (diamond at the center surrounded by 4 +1 defects. (B) Polarizing optical microscopy texture of the pattern. (D) CounterÂ lockwise trajectories around the four +1 vortices. (E) Map of corresponding velocities. c SOURCE: From C. Peng, T. Turiv, Y. Guo, Q.-H. Wei, and O.D. Lavrentovich, 2016, Command of ac- tive matter by topological defects and patterns, Science 354:882-885, doi: 10.1126/science.aah6936, reprinted with permission from AAAS. 77 V. Schaller, C. Weber, C. Semmrich, E. Frey, and A.R. Bausch, 2010, Polar patterns of driven filaments, Nature 467(7311):73.
98 F r o n t i e r s o f M at e r i a l s R e s e a r c h self-organizes into coherently moving structures with persistent density modula- tions. In another version, bundled microtubules are set in motion by molecular motors.78 The result at sufficient density is a system with local nematic order with disclination defects that form, distort, and eventually annihilate in pairs. 2.6 ARCHITECTED MATERIALS In the past decade, the confluence of computational design of material prop- erties, fabrication processes with nanometer to micrometer control (origami- and kirigami-inspired fabrication, lithography, self-propagating photopolymers, AM etc.; see Chapter 4), and experimental characterization methods with equally fine resolution has made it possible to use material shape, or architecture, to create novel bulk materials with radically superior properties via architectural control at the appropriate scales. In much the same way that arches, columns, beams, and buttresses revolutionized the construction of buildings, towers, and bridges in past centuries, the materials community is now exploiting material architecture to ex- pand the material design space in multiple dimensions, independently manipulate material properties that are currently coupled, and develop materials with vastly superior properties than can be achieved with solid objects. Stochastic metallic foams have given way to metamaterials and microlattices with ever finer features and ever broader material classes and functionalities.79 Breakthrough improve- ments in specific strength and stiffness, energy absorptivity, negative Poisson ratio, zero thermal expansion, thermal conductivity, as well as electromagnetic sensing, filtering, cloaking, and communications have all been reported for laboratory quan- tities of materials in the past decade. Figure 2.21 shows how this approach can fill property gaps such as in the strength to density of materials. 2.7 MATERIALS FOR CATALYSIS Catalysts are materials that facilitate the chemical conversion of reactants to desired products. Their global importance can be gleaned from the simple fact that in 2018, they constitute a $20 billion industry. The majority of catalysts, ap- proximately 80 percent, are heterogeneous (made up of solid catalysts in a liquid reaction mixture); while some others (17 percent) are homogeneous (where the catalyst is in the same phaseâgas or liquidâas the reactants); and a small number 78 F.C. Keber, E. Loiseau, T. Sanchez, S.J. DeCamp, L. Giomi, M.J. Bowick, M.C. Marchet- ti, Z. Dogic, and A.R. Bausch, 2014, Topology and dynamics of active nematic vesicles, Science 345(6201):1135-1139. 79 G.V. Franks, C. Tallon, A.R. Studart, M.L. Sesso, and S. Leo, 2017, Colloidal processing: Enabling complex shaped ceramics with unique multiscale structures, Journal of the American Ceramics Society 100:458-490.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 99 FIGURE 2.21â Architected structures can be used to fill the void in producing lightweight materials with high yield strength. The figure shows the property space occupied by nickel-based and diamond- based lattices in comparison to conventional bulk materials. SOURCE: Republished with permission of from T.A. Schaedler and W.B. Carter, 2016, Architected cellular materials, Annual Review of Materials Research 46:187-210; permission conveyed through Copyright Clearance Center, Inc. (approximately 3 percent) are based on enzymes. Enzyme catalysts include a pro- tein with an active site that increases the rate of a chemical reaction. Continued advances in synthesis, characterization, and predictive modeling, as described in more detail in Chapter 4, have led to the birth of several new classes of materials for catalysis that may transcend the division among the three types men- tioned above. Much progress has been made through controlled synthesis routes, such as atomic layer deposition, and in situ techniques that work under reaction conditions. Characterization techniques include a number of spectroscopies, mi- croscopies, and chromatographies that are now capable of quantifying the adsorp- tion, desorption, diffusion, and reaction characteristics, as well as the structural changes in the catalyst, in real time. In tandem, high-throughput computational screening, armed with data-enabled methods, are helping to identify descriptors
100 F r o n t i e r s o f M at e r i a l s R e s e a r c h and material design principles that accelerate the discovery of cost-effective cata- lysts. As a result of these efforts, hybrid forms of catalysts have been developed that exhibit enhanced reactivity and product selectivity. The nature of the active site heterogeneous catalysis determines both its activ- ity and selectivity for specific chemical reactions. The ideal active site for a given reaction will turn over reactants to products at the highest possible rate and at the lowest reaction temperature, leading to substantial energy savings. Second, the ideal active site will produce only the desired set of products, with no by-products, which would save the energy required from subsequent costly product-separation processes. The third major consideration for the active site is its long-term stabil- ity in reactive environments. The ideal catalyst would not need regeneration or replacement for a period of several years. The past decade has seen a great deal of progress in terms of new materials, which could catalyze a range of important chemical reactions, but also the role of surface conditions such as in plasma-assisted hot electron catalysis.80 One class of such materials relies on shape-controlled wet synthesis of late transition metal nanoparticles. Figure 2.22 shows octahedral and cubic nanocages of platinum with a nanocage thickness of only a few atomic layers.81 These hollow structures with well- defined facets maximize atom efficiency by eliminating the need for core materials, FIGURE 2.22â Platinum hollow nanocages for stable and efficient low temperature catalysis. SOURCE: L. Zhang, L.T. Roling, X. Wang, M. Vara, M. Chi, J. Liu, S.-I. Choi, et al., 2015, Platinum-based nano- cages with subnanometer-thick walls and well-defined, controllable facets, Science 349:412-416, reprinted with permission from AAAS. 80 J.Y. Park, L.R. Baker, and G.A. Somorjai, 2015, Role of hot electrons and metal-oxide interfaces in surface chemistry and catalytic reactions, Chemical Reviews 115(8):2781-2817. 81 L. Zhang, L.T. Roling, X. Wang, M. Vara, M. Chi, J. Liu, S.-I. Choi, et al., 2015, Platinum- based nanocages with subnanometer-thick walls and well-defined, controllable facets, Science 349(6246):412-416.
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 101 meanwhile exposing both inner and outer faces of the shell to maximize the active catalytic surface area. First-principles theory has guided experiments toward the optimal thickness of each facet, as well as the most active type of facet exposed for structure-sensitive reactions. The hollow structures show oxygen reduction activity, a key reaction for low-temperature fuel cells, that has five times higher mass activity and substantially enhanced durability in comparison to commercial Pt/C. The ultimate size limit for nanoparticles is single-atom or subnanometer clus- ters. These atoms are typically supported on a substrate, and the catalytic activity can be associated with the metal atom as well as the surrounding support struc- ture. Molecular dynamics computer simulations have been used to understand the interaction and found that a single atom may be released from the nanocluster to become the active site for the reaction, the atom returning to the cluster on completion of the reaction. Advances in characterization capabilities, not only in the spatial, temporal, and energy resolution but also in the ability to conduct experiments in situ under operando conditions, have accelerated understanding of the dynamic structural and compositional changes that occur during the catalytic cycle.82 Multimode ex- perimental tools now allow the structural and compositional modifications to be probed in a single experiment in real time and operando conditions. Tomographic methods now provide three-dimensional insight of porous substrates and the dis- persion of the nanoparticles, and of oxides on and in substrates to identify active sites, and information on the elemental distribution. These in situ and operando capabilities have provided new insights as to the dynamic structural and compositional processes taking place during catalysis. For example, Ni-Fe nanoparticles realloy following reduction in H2 gas and dealloy after CO2 oxidation. This compositional change is highlighted in Figure 2.23, which shows energy dispersive X-ray spectroscopy maps of the change in the composi- tion of a Ni-Fe nanoparticle on oxidation. In the maps, iron is red and nickel is green.83 The redistribution of nickel and iron from the homogeneous distribution to a nonhomogeneous distribution on oxidation is evident. Iron, but not nickel, has been oxidized, and it is the iron oxide on the surface that reacts with carbon. This schematic of the dealloying and realloying, as shown in Figure 2.23(c),84 shows how the oxidation of iron and the elemental redistribution leads to a decrease in the accumulation of carbon, which is known to decrease the efficiency of catalysts in the dry reforming of methane. 82 F. Tao and P.A. Crozier, 2016, Atomic-scale observations of catalyst structures under reaction conditions and during catalysis, Chemical Reviews 116(6):3487-3539. 83 S.A. Theofanidis, V.V. Galvita, H. Poelman, and G.B. Marin, 2015, Enhanced carbon-resistant dry reforming Fe-Ni catalyst: role of Fe, ACS Catalysis 5(5):3028-3039. 84 Z. Bian, S. Das, M.H. Wai, P. Hongmanorom, and S. Kawi, 2017, A review on bimetallic nickelâ based catalysts for CO2 reforming of methane, ChemPhysChem 18(22):3117-3134.
102 F r o n t i e r s o f M at e r i a l s R e s e a r c h FIGURE 2.23â Energy dispersive X-ray spectroscopy maps of the change in composition of a Ni-Fe particle. (a) Composition after reduction in hydrogen gas. (b) Composition after oxidation in CO2. (c) Schematic of the dealloying and realloying reactions and the interaction of carbon with iron oxide. In the compositional maps, red corresponds to Fe and green to Ni. SOURCE: S.A. Theofanidis, V.V. Galvita, H. Poelman, and G.B. Marin, 2015, Enhanced carbon-resistant dry reforming Fe-Ni catalyst: role of Fe, ACS Catalysis 5(5):3028-3039, Â© 2015 American Chemical Society. Driven by the need to increase catalyst surface area, decrease the use of pre- cious metals, and improve the stability of nanoparticles, considerable progress has been made in confining nanoparticles within open structured systems, such as metal organic frameworks, novel zeolites, zeolitic imidazolate frameworks, porous organic polymers, materials, carbon nanotubes, and so on. A consequence of intro- ducing a confinement cage is that it creates a local environment that differs from that encountered in the bulk form. This has led to enhancements in the selectivity and reactivity as well as opened new reaction pathways with faster reaction rates. First-principles computational methods have been used extensively to interpret spectroscopic data, calculate binding energies and activation energy barriers, and provide electronic structure information as well as reaction energetics of elemen- tary reaction steps, leading to an unprecedented understanding of structure-activity and structure-selectivity correlations in catalysis. Further, ab initio thermodynam- ics has been shown to provide essential insights as to the nature of the catalytic phase under reaction conditions. For example, density functional theory calcula- tions coupled with ultrahigh vacuum surface science studies have revealed the dif- ferent forms particles can take on a surface as a function of pressure, temperature, and chemical potential of the environment to which the catalytic particles and their support are exposed to. Importantly, modern electronic structure theoretical methods, in conjunction with detailed microkinetic modeling, have enabled the prediction of catalytic materials with unique structure and composition, which pos- sess improved catalytic properties compared to materials found by trial and error. Another area of active research is the controlled deposition of oxophilic metal promoters (e.g., Fe, Mo, Re, Zr, Mn, Sn) onto the surface of reducible metal nanoparticles (e.g., Pt, Rh, Cu) to produce catalysts containing interfacial sites that show high activity for important reactions, such as water-gas shift, methane
P r o g r e s s a n d A c h i e v e m e n t s i n M at e r ia l s R e s e a r c h 103 conversion to ethylene, synthesis gas conversion to alcohols and alkenes, and hy- drogenolysis of cyclic ethers. Driven by the need to increase catalyst surface area and decrease the use of precious metals, attention has turned to novel geometries such as those afforded by metal organic frameworks (MOFs), novel zeolites, zeolitic imidazolate frame- works, mesoporous materials, 2D materials, and single-site catalysts. The challenge is to synthesize these materials with controlled, reproducible, and stable proper- ties. In this regard, carbon-based materials have attracted a great deal of attention because of their simple architecture, which facilitates heat- and mass-transport phenomena, thermodynamic stability, and the promise of a metal-free catalyst. Their high chemical inertness allows them to withstand operations in aggressive media. Surface functionalization or matrix doping with nitrogen and oxygen spe- cies allows further tuning of their surface reactivity. Application of carbon-based catalysts can be found in hydrogen oxidation, desulfurization, and liquid-phase transesterification reactions. 2.8 CONCLUSION The past decade has seen significant advances across all areas of materials science and engineering, much of this enabled by the teaming of experimentalists with computational materials scientists and the emerging importance of material informatics. Materials synthesis and processing to produce new materials as well as high-quality materials for fundamental investigations remains an ongoing chal- lenge. The pace of fundamental discovery accelerated and many new materials and processes, some unanticipated, moved into commerce and society during the past decade.