UNCLASSIFIED/ / POR OFFICIAL USE ONLY 14 Decem ber 2009 ICOD: 1 Decem ber 2009 DIA-08-09 11-012 Defense Intelligence Reference Docum ent A cquisition T hreat S upport M etallic Glasses: Status and Prospects for Aerospace Applications UNCLASSIFIED// FOR OFFICIAL UOC ONLY UNCLASSIFIED/ /TUR UFFTCTAL U3E ONLY M etallic Glasses: Status and Prospects for Aerospace Applications Prepared by: Acquisition Support Division (DWO-3) Defense Warning Office Directorate for Analysis Defense Intelligence Agency Author: AAP Person 63 Administrative Note COPYRIGHT WARNING: Further dissemination of the photographs in this publication is not authorized. This product is one in a series of advanced technology reports produced in FY 2009 under the Defense Intelligence Agency, Defense W arning Office's Advanced Aerospace W eapon S ystem Applications (AAW S A) Program . Com m ents or questions pertaining to this docum ent should be addressed to |AAP Person 1 [ AAW S A Program M anager, Defense Intelligence Agency, ATTN: CLAR /DW O-3, Bldg 6000, W ashington, DC 2034 0-5100. UNCLASSIFIED// FOR OFFICIAL USE UNL! UNCLASSIFIED/ /FOR OFFICIAL USE ONLY Contents Summary...........................................................................................................................v Metallic Glasses................................................................................................................1 Structure.......................................................................................................................1 Processing....................................................................................................................2 Glass-Forming Alloys..............................................................................................2 Casting and Molding...............................................................................................4 Joining....................................................... 5 Foams..................................... 5 Thin Films and Coatings.........................................................................................5 Mechanical Behavior Near Room Temperature....................................................5 Stiffness: Elastic Deformation...............................................................................6 Strength and Ductility: Plastic Deformation........................................... 6 Fracture Toughness.................................................................................................8 Fatigue.....................................................................................................................9 Wear Resistance.................................................... 10 Corrosion and Stress-Corrosion Cracking...........................................................10 Mechanical Behavior at Elevated Temperature..................................................11 Other Properties: Magnetic, Electrical, Optical, Thermal, and Acoustic........12 Metallic Glass Matrix Composites................................................................................13 Processing and Structure of Composites................................................................13 Ex Situ Composites....................................................................................................14 In Situ Composites....................................................................................................14 Mechanical Properties of Composites......................................................................15 Strength and Ductility: Plastic Deformation...........................................................16 Fracture and Fatigue...................................................................................... 16 Aerospace Applications of Metallic Glasses................................................................16 Structural Applications.............................................................................................16 Other Applications.....................................................................................................19 iii UNCLASSIFIED/ / FOR OrTICIAL USE ONLY UNCLASSIFIED/ / FOR OrriCIAL USE ONhY Current Challenges and Prospects for the Future.............................. 20 Alloy Design...............................................................................................................20 Thermophysical Properties and Thermoplastic Processing.................................20 Composites and the Quest for Ductility..................................................................21 Summary and Recommendations................................... 22 Figures 1. Amorphous Versus Crystalline Structure..................................................................1 2. Critical Cooling Rate.................................................... 2 3. Examples of Processing of Metallic Glasses..............................................................4 4. Shear Bands................................. 8 5. Fatigue Limit of Metallic-Glass-Matrix Composites................. 10 6. Deformation Map for a Metallic Glasses..................................................................11 7. Cast Metallic Glass Wedge........................................................................................13 8. Microstructure of In Situ Metallic Glass Matrix Composite....................................15 9. Materials Property Charts..................................................... 18 Tables 1. Selected Bulk Glass-Forming Alloys........................ 3 2. Comparison of Strengths of Amorphous and Crystalline Aluminum Alloys.........7 iv UNCLASSIFIED// FOR OFFICIAL UOC OM L¥ UNCLASSIFIED/ / FOR OFFICIAL USE ONCT M etallic Glasses: Status and Prospects for Aerospace Applications Sum m ary Metallic glasses combine some of the advantageous mechanical properties of metals—strength, stiffness, and in some cases toughness—with the processing flexibility usually associated with thermoplastic polymers. The absence of crystalline defects allows metallic glasses to be much stronger than conventional alloys but also means they have near-zero tensile ductility and poor fatigue resistance. In structural applications, therefore, metallic glasses are most likely to be useful in the form of composites consisting of ductile crystalline dendrites in a metallic glass matrix. These dendritic composites sacrifice some strength but can have exceptionally high fracture toughness, as well as good fatigue resistance, and could replace high-strength steels in certain load-limited structural components in aerospace vehicles where space is limited. Because they are true glasses, thermoplastic forming near the glass transition temperature affords metallic glasses tremendous flexibility in processing. For instance, metallic glass components can be formed in a single step (for example, by injection molding) in complex geometries that would be difficult or impossible to produce with conventional alloys. In addition, metallic glass foams can be made with relative ease, raising the possibility of making structural foams with high strength and stiffness. Finally, because they lack a crystalline grain structure, metallic glasses can be used to form nanoscale features with high fidelity. This may make metallic glasses useful in a variety of micro-electromechanical systems (MEMS) applications. Metallic glasses also have significant limitations for aerospace applications, however. Foremost among these is a lack of good glass-forming alloys; in particular, there are no good aluminum-rich glass-forming alloys, the known titanium-based alloys are either relatively dense (owing to high concentrations of alloying elements) or contain beryllium, and the known magnesium- and iron-based alloys are all quite brittle, with low fracture toughness. Although metallic glass matrix composites can have outstanding properties (particularly strength and fracture toughness), the number of good composite systems known at present is also quite limited. Therefore, in order for metallic glasses (and their composites) to be of broad utility in aerospace structural applications, progress in the following areas is required: • Development of new lightweight alloys and composite systems, preferably by computational and/or combinatorial approaches rather than by trial and error. • Understanding of mechanical behavior, especially: v UNCLASSIFIED/ / FOR DEHCIAL USE ON Hl* UNCLASSIFIED/ /FOR OFFICIAL USE ONLY - [ The effect of alloy composition and structure on plastic deformation. - Microstructural design of composites for optimal toughness. • Development of processing techniques, including thermophysical processing of complex and/or nanoscale features as well as production of metallic glass foams. It is highly likely that continued work over the next 20-50 years will result in significant advances in all these areas, and that metallic glasses and metallic glass matrix composites will see increasing acceptance as structural materials. Whether or not they achieve widespread use in aerospace applications, however, depends critically on the development of new, lightweight alloys. vi UNCLASSIFIED// FOP nEfiiriAL USE ONL Y UNCLASSIFIED/ / FOR OFFICIAL USE ONLY M etallic Glasses STRUCTURE The atom ic-scale structure of m ost m etals and alloys is crystalline; that is, the atom s are arranged in a highly ordered m anner on a lattice that is periodic in three dim ensions, as depicted in Figure 1(a). In contrast to this crystalline structure, m etallic glasses lack the long-range order of a lattice and are therefore said to be am orphous, as depicted in Figure 1(b). Although the w ord "am orphous" im plies a com plete lack of structural order, in fact the atom ic structure of m etallic glasses is not truly random . Constraints on atom ic packing provide strong short-range order; for instance, on average the atom s have a particular num ber of nearest atom ic neighbors at a w ell- defined distance. But this short-range order persists only over distances of a few atom s; there is no long-range order as there is in a crystalline alloy. In m any w ays, the atom ic-scale structure of m etallic glasses m ore closely resem bles the highly disordered structure of a liquid than the structure of a crystalline alloy. Crystalline Amorphous (glass) Figure 1. Am orphous Versus Crystalline Structure. Schem atic atom ic-scale structure of crystalline (a) and am orphous (b) m etals. In a crystalline structure, order persists over long distances (m any atom ic dim ensions). In a glass, there is short range order but no long-range order. A corollary of this difference in structure is that the nature of structural defects is quite different betw een crystalline and am orphous alloys. Crystalline alloys, for exam ple, have extended linear defects in the crystal structure, called dislocations, that are (in large part) responsible for determ ining m echanical behavior. The lack of crystalline order precludes the existence of dislocations in m etallic glasses, but other sorts of defects can be present and m ay influence properties and behavior. From an applications point of view , the am orphous structure of m etallic glasses has tw o principal im plications. First, the m echanical properties of am orphous alloys are significantly different from those of their crystalline counterparts; som e of these differences are advantageous, but others are not. S econd, because m etallic glasses are 1 UNCLASSIFIED//FOR OFFICIAL USE ONLY UNCLASSIFIED / / W K OFFICIAL UOE ONLY glasses in the true sense of the w ord, rather than m elting abruptly (as crystalline m etals do), they soften and flow over a range of tem peratures in a m anner akin to com m on (oxide) glasses. This creates opportunities for trem endous flexibility in the processing of m etallic glasses. PROCESSING Glass-Form ing Alloys The key to m aking a m etallic glass is to retain the disordered, liquid-like atom ic scale structure during cooling from the m elt. All m aterials have a tendency to crystallize upon cooling because the crystalline state is the m ost stable structure at any tem perature below the m elting point. But crystallization takes tim e, so if the cooling is fast enough, it is possible to bypass crystallization and form an am orphous structure at the glass transition tem perature (Figure 2(a)). G lass form ation and crystallization are therefore com petitive processes; w hich one w ill occur depends on the m aterial and the processing conditions. (b) cd ^ 94 IOW 10 0 Pure nickel '5 10 10 10 ur Convention a f’ metallic glasses ^ (max thickness < 1 mm) Bulk metallic glasses (max thickness > I mm) 10 Z rM .T i, t'U !, N i^B c 0,2 03 0.4 03 Reduced glass transition temperature (T/T^) Figure 2 . Critical Cooling Rate, (a) Effect of the cooling rate on glass form ation If the cooling rate is slow (path 1), then the m elt crystallizes before going through the glass transition. If the cooling rate is fast enough (path 2), then the m elt can form a glass. The critical cooling rate (path 3) is the slow est rate at w hich the m elt can be cooled and still form a glass, (b) Critical cooling rates for various m etallic alloys The horizontal axis Is the glass transition tem perature norm alized to the m elting (liquidus) tem perature.1 ^ .XK NiX For som e m aterials, such as silica (silicon dioxide) and m ost therm oplastic polym ers, the crystallization process is slow because the crystal structures are com plex and the basic structural units (for exam ple, segm ents of polym er chains) are slow to rearrange into a crystalline form . These m aterials can therefore be produced in glassy form even at very low cooling rates; in fact, it can be difficult to crystallize them at all. M etals and alloys are another m atter because the crystal structures are relatively sim ple and the basic structural units are individual atom s, w hich are highly m obile. M etallic crystals nucleate and grow quickly, m aking production of a m etallic glass m ore challenging. 2 U N C LASSI FI ED/ /FOR OFFICIAL U6E ONLY UNCLASSIFIED / / FOR OFFICIAL USE ONLY One w ay to quantify the ability of a m etallic alloy to be produced in glassy form is through the critical cooling rate— the slow est rate at w hich a m etallic liquid m ay be cooled and still produce a fully am orphous structure, as show n in Figure 2(a). The critical cooling rate for a variety of m etallic glass-form ing alloys is show n in Figure 2(b). Early m etallic glasses (discovered in the 19 60s and 19 70s) w ere binary alloys w ith critical cooling rates typically on the order of 104 to 107 K /s. Achieving such high cooling rates requires specialized techniques (such as m elt spinning) and lim its the m axim um thickness of the m etallic glass to < 100 pm because of the need to rapidly extract heat from the m elt. As a result, these early m etallic glasses could be produced in only a lim ited range of form s, including ribbons, foils, w ires, and pow ders. Extensive research efforts in alloy design over the past tw o decades have resulted in the developm ent of m ulti-com ponent alloys w ith m uch low er critical cooling rates (0.1 K /s or even low er). This has enabled the production of m etallic glass specim ens in larger sizes— in som e cases exceeding 1-cm section thickness. Com m on practice in the field is to refer to any alloy capable of being cast into a section at least 1-m m thick as a "bulk" m etallic glass. These alloys m ay be cast or m olded into form s suitable for structural applications. At present, it is not possible to predict a priori the glass-form ing ability of an alloy of arbitrary com position. A variety of em pirical rules for selecting alloying elem ents and com positions have been proposed, and techniques have been dem onstrated for efficient searching of com position space. But identification of alloys w ith good glass-form ing ability is still m ostly a m atter of trial and error. As a result, the num ber of truly outstanding glass-form ing alloys (loosely defined as being able to be cast as a glass to a thickness of at least 1 cm ) is quite lim ited (see Table 1). Table 1. Selected Bulk Glass-Forming Alloys. S elected alloys reported to have excellent glass form ing ability, quantified here as the m axim um thickness of a fully am orphous casting.2 34 $ $ 7 8 Composition Maximum Thickness (mm) Reference MgssCuisAgsPdsGdio 10 2 Zr4i.2Tii3.8Cui2.sNiioBe22.5 50 3 Pd40CU30Nil0P20 72 4 Cu4?Zr4sAg4Al4 10 5 Pt57tsCU14.7Ni5.3P22.5 16 6 Ti4oZr2sNi3Cui2Be2o 14 7 Fe48CrisMoi4Er2Ci5B6 12 8 M oving from the laboratory to industrial practice, it is im portant to note that factors besides alloy com position can affect glass-form ing ability. In particular, som e alloys are sensitive to the presence of im purities; for exam ple, the glass-form ing ability of som e zirconium -containing alloys is dram atically reduced by the presence of oxygen. Processing conditions also influence the ability to m ake a glass; these m ay include the m aterial and surface finish of the m old and the tem perature of the liquid prior to casting. Finally, glass-form ing ability can be quite sensitive to sm all variations in com position, w hich m ay be difficult to control in industrial practice. 3 UNCLASSIFIED// FOR OFFICIAL UOC ONLY UNCLASSIFIED / / FOR OFFICIAL UDE ONLY Casting and M olding Like other alloys, m etallic glasses can be cast into net-shape or near-net-shape geom etries. Die casting into a perm anent (m etal) m old— because it provides the rapid heat transfer needed to m eet the requirem ent for relatively rapid cooling— is the m ost com m on casting technique. In m ost cases, casting is done in either a vacuum or an inert atm osphere to prevent form ation of oxide particles that prom ote crystallization. Conventional casting, how ever, does not take advantage of the flexibility afforded by the glassy nature of these alloys. If a m etallic glass is heated to a tem perature above its glass transition tem perature, it becom es a supercooled liquid. In this state, the viscosity drops w ith increasing tem perature over a w ide range, m aking it possible to control the viscosity by controlling the tem perature.1 This ability to control the viscosity enables m any of the processing techniques com m only used in m olding therm oplastic polym ers to be applied to m etallic glasses (Figure 3). 1 A crystalline m etal, In contrast, m elts abruptly, going from a rigid solid to a low -viscosity fluid very quickly. Figure 3. Exam ples of Processing of M etallic Glasses, (a) M icrospring produced by lithography and (b) thin’ w alled bottle produced by blow m olding. Im ages are courtesy of Professor Jan S chroers (Yale University). There are tw o im portant lim itations on processing of m etallic glasses in the supercooled liquid region. First, supercooled liquids are m etastable and have a tendency to crystallize, so there is a lim ited w indow of tim e (typically on the order of m inutes) in w hich the processing m ust be com pleted if the glassy structure is to be m aintained. S econd, the viscosity of m any glass-form ing alloys near the glass transition tem perature is too high for convenient processing. The viscosity can be reduced by increasing the processing tem perature, but higher tem peratures prom ote crystallization 4 UNCLASSIFIED/ / »^»»«^Tfti iipfoniy UNCLASSIFIED/ / POU OFFICIAL USE ONLY and thus reduce the w indow of tim e available for m olding. In practice, therefore, successful m olding requires careful control of the processing conditions. Joining S tructural applications inevitably require joining of com ponents, for instance by m echanical fasteners or adhesives or by w elding, soldering, or brazing. The use of fasteners and adhesives is m uch the sam e for m etallic glasses as for any other m etal. Techniques such as w elding, soldering, and brazing are potentially problem atic because they involve heating the glassy alloy, running the risk of crystallization (w hich could m ake the joint m ore brittle). In w elding, for instance, the m etal to be joined is actually m elted and then resolidifies upon cooling. In the case of a m etallic glass, care m ust be taken to ensure the cooling rate is fast enough to avoid crystallization. There is also a risk that the glassy m aterial in the heat-affected zone (near to but not in the m olten region) m ight crystallize. Laboratory tests of a variety of w elding techniques have been perform ed on several glass-form ing alloys w ith m ixed results, and it is clear that m uch rem ains to be done in this area. Foam s One particularly prom ising recent developm ent is the ability to produce m etallic glass foam s. H ere, the relatively high viscosity of glass-form ing alloys is an advantage in producing a stable foam structure that can be solidified, leaving a high-porosity foam w ith m etallic glass ligam ents.9 These foam s have high specific strength (that is, strength norm alized to density) and specific stiffness and could have excellent dam age tolerance, although this has not been dem onstrated. Thin Film s and Coatings The discussion above focuses on the processing of free-standing m etallic glasses, w ith an em phasis on structural applications. H ow ever, it is also possible to produce am orphous alloys as thin film s or coatings using techniques such as physical vapor deposition or electrodeposition. Although the thicknesses of m aterial that can be produced in this w ay are lim ited, they are useful for m aking am orphous alloy coatings (for w ear and corrosion resistance) or for thin film s for m agnetic or m icro­ electrom echanical system (M EM S ) applications. A distinct advantage of the thin film techniques is that because the effective cooling rates during vapor deposition are extrem ely high, a m uch w ider range of alloys can be produced in am orphous form than is possible w ith casting. This allow s the alloy com position to be tailored for optim ization of functional properties, w ith less concern about glass-form ing ability. M echanical B ehavior Near Room Tem perature W hen a m aterial is subjected to a stress, it can experience both elastic and plastic deform ations. Elastic deform ation occurs at low er stresses and is recoverable w hen the applied stress is rem oved. The lim it of elastic deform ation is defined by the yield stress— the point at w hich plastic (nonrecoverable) deform ation begins. M uch of the current interest in m etallic glasses arises because their yield stresses (that is, their strengths) can be m uch higher than those of crystalline alloys of sim ilar com position; this difference is a direct result of the novel atom ic-scale structure of m etallic glasses. The fracture and fatigue characteristics of m etallic glasses are also different from those 5 UNCLASSIFIED//FQ R OFFICIAL USE ONLY UNCLASSIFIED//rOR OFFICIAL USE ONLY of conventional alloys. In this section, w e review the m echanical behavior of m etallic glasses, w ith particular attention to properties of interest for aerospace applications. W e consider actual properties in detail in the section below on applications, w here w e com pare the properties of m etallic glasses w ith those of other advanced structural m aterials. Stiffness: Elastic Deform ation S tiffness is the resistance of a m aterial to elastic deform ation and is quantified by either the elastic m odulus (for tensile or com pressive loads) or the shear m odulus (for shear loading). M etallic glasses tend to be som ew hat (20-30 percent) less stiff than crystalline alloys of sim ilar com position. The low er m odulus is a consequence of the am orphous structure, in w hich atom s are (on average) slightly farther apart than in a crystalline alloy, enabling certain atom ic relaxations that are not possible in a crystal. The low er m odulus of am orphous alloys is clearly a concern in applications w here stiffness is a prim ary criterion, but it does present som e advantages. For instance, som e applications (springs, for exam ple) require the ability to store elastic strain energy (resilience), and here m etallic glasses do quite w ell. R esilience is also a key figure of m erit for snap-fit assem bly of m aterials w ithout fasteners. Overall, how ever, for structural applications, the low stiffness of m etallic glasses is a disadvantage. Strength and Ductility: Plastic Deform ation The theoretical strength of perfect, defect-free crystalline m etals is several orders of m agnitude larger than strengths m easured in typical laboratory experim ents. The difference exists because m etallic crystals inevitably have crystalline defects (dislocations) that are able to m ove at relatively low stresses and cause plastic (nonrecoverable) deform ation. Because dislocations cannot exist in an am orphous structure, in principle the strength of am orphous alloys should approach theoretical lim its based on the inherent strength of the atom ic bonds. As show n in Table 2, the strength of alum inum -based m etallic glasses can be tw o or three tim es greater than those of conventional (crystalline) high-strength alum inum alloys. S im ilarly high strengths are seen for other am orphous alloys; for instance, the best iron-based alloys have a strength of approxim ately 4 G Pa— again, tw o or three tim es greater than those of conventional high-strength steels.10 S uch high strengths create great interest in potential structural applications of m etallic glasses. 6 UNCLASSIFIED/ / EQ & Q EfilCIAIUlfiUUUX UNCLASSIFIED/ / FOP OK THAL UGC ONL¥ Table 2. Comparison of Strengths of Amorphous and Crystalline Aluminum Alloys. Com pared w ith the theoretical m axim um strength (taken to be p/30, w here p is the shear m odulus of pure alum inum ). Yield Stress (M Pa) % of Theoretical Strength Theoretical S trength (Defect- 1,600 - Free Crystal) Typical H igh- S trength Alum inum Alloy (7xxx S eries)11 4 00-500 25-31% Best Crystalline Alum inum Alloy12 770 4 8% Alum inum -Based M etallic G lass13 1,280 80% Unfortunately, the lack of dislocations in am orphous alloys is also their Achilles' heel. In crystalline alloys, dislocations m ove and m ultiply in response to applied stresses, resulting in dislocation tangles that increase the resistance to further dislocation m otion. This process, called strain hardening, is of crucial im portance because it m akes plastic deform ation stable. If one region of a crystalline m aterial yields and begins to plastically deform , the deform ing region strain hardens, and so another region w ill deform instead. The result is that the plastic deform ation is not concentrated but rather spreads through a large volum e of m aterial. M etallic glasses, lacking dislocations, do not strain harden and in fact strain soften in response to plastic deform ation. This m eans that as soon as any one region yields, any further deform ation w ill occur in the sam e region. This process, know n as shear localization, leads to the form ation of shear bands (Figure 4 ). In any loading geom etry w here the m etallic glass experiences significant tensile loading, fracture occurs on a single dom inant shear band w ith essentially zero tensile ductility.2 M etallic glasses therefore fracture in an abrupt, apparently brittle m anner on the m acroscopic scale (even though there can be significant plasticity on a m icroscopic scale). This lack of ductility is of obvious concern to designers interested in structural applications. Furtherm ore, it lim its the ability to fabricate m etallic glasses into different shapes by deform ation processing (by rolling or forging, for instance) after casting. 2 This assum es there is no geom etrical constraint preventing fracture. S om e geom etries (such as sim ple bending) can involve tensile loading, but there can still be significant plastic deform ation because the geom etrical constraints inhibit propagating of shear bands across the specim en. 7 UNCLASSIFIED// FOR OrriCIAh USE ONLY UNCLASSIFIED/ / FOR OFFICIAL USE ONET Figure 4. Shear B ands. Produced by bending of a zirconium -based m etallic glass.14 Fracture Toughness Fracture toughness is a m easure of a m aterial's resistance to grow th of cracks, a critical property for structural m aterials subjected to tensile loading. In very tough m etals, the toughness usually results from plastic deform ation that occurs near the tip of the advancing crack; plastic deform ation requires energy, and the need to provide this energy translates into resistance to crack grow th.3 Despite their lack of tensile ductility, at least som e m etallic glasses are not brittle in the sam e sense that ceram ics are, for exam ple, because they can experience significant plastic deform ation around the crack tip during fracture. For instance, the fracture toughness (K ic) of zirconium -based m etallic glasses is about 20 M Pam 1/21S— som ew hat low er than the ~ 55 M Pam 1/2 typical of crystalline zirconium alloys16 but m uch greater than the fracture toughness of ceram ics (typically 1-5 M Pa m 1/2). The fact that m etallic glasses are reasonably tough despite their lack of tensile ductility suggests structural applications are not out of the question. 3 In other m aterials, notably polym er-m atrix com posites, other m echanism s of toughening can be m ore im portant. H ow ever, som e m etallic glasses appear to be intrinsically brittle in that they fracture w ith only lim ited plastic deform ation near the crack top and thus have very low values of fracture toughness. For this reason, som e alloys that w ould otherw ise be highly desirable, such as iron-based m etallic glasses (for their high strength and low cost) and m agnesium -based glasses (for their low density), fall into this category. The physical origins of the difference betw een intrinsically brittle m etallic glasses and those capable of lim ited plastic deform ation (and thus som e toughness) are not w ell understood. 8 UNCLASSI FI ED/ / FOR OFFICIAL USE ONL¥ UNCLASSIFIED/ / FOA OFFICIAL USE ONLY Even som e m etallic glasses w ith reasonable toughness m ay be em brittled by exposure to elevated tem peratures. This m ay occur in the heat-affected zone during w elding (as discussed above), or it m ay be a by-product of processing in the supercooled liquid region (as in injection m olding, for instance). The causes of em brittlem ent are also not w ell understood, and there is no know n w ay to reverse em brittlem ent once it occurs. Fatigue Fatigue is a process by w hich m aterials can experience increm ental crack grow th ow ing to cyclic loading, even at stresses w ell below the yield stress. If unabated, fatigue cracks can grow to a critical length at w hich abrupt catastrophic fracture occurs. Up to 9 0 percent of failures of structural com ponents in service are estim ated to be caused by fatigue, m aking fatigue resistance of obvious im portance to designers. The fatigue resistance of m etallic glasses is not very good. A com m on m easure of fatigue resistance is the fatigue lim it— the stress am plitude (range) below w hich no fatigue failure w ill occur, regardless of the num ber of loading cycles the m aterial experiences, The fatigue lim it for high-strength crystalline alloys is typically about 4 0 percent of the tensile strength, but for m etallic glasses, it is only about 5 percent of the tensile strength (Figure 5). The reason for this difference has to do w ith the structure of the m aterial. In a crystalline alloy, there are m icrostructural features (such as grain boundaries and precipitate particles) that can inhibit the grow th of fatigue cracks. In m etallic glasses, the m icrostructure is com pletely featureless, and there is nothing to prevent fatigue cracks from grow ing once they have been initiated. The poor fatigue resistance of m etallic glasses is a critical lim itation for structural applications in aerospace because it im plies a need to overdesign com ponents to keep the stresses far below the yield stress. Thus, m uch of the advantage of having a high- strength m aterial in the first place is lost. The desire to im prove m etallic glasses' fatigue perform ance has led to the developm ent of m etaliic-glass-m atrix com posites w ith outstanding properties, as discussed below . 9 UNCLASSIFIED// B Q R OHICIAk W EE ONLY UNCLASSIFIED / / FOR OFFICIAL USE ONLY Number of cycles to failure Figure 5. Fatigue Lim it of M etallic Glasses and M etallic-Glass-M atrix Com posites. Fatigue life data for single-phase zirconium -based m etallic glass (red) and a dendritic m etallic glass m atrix com posite (blue). R epresentative data for steel (300-M ) of sim ilar tensile strength are show n for com parison.17 18 19 W ear Resistance Because of their high yield strength, m etallic glasses also have very high hardness. This, in turn, im plies they m ight have good tribological behavior, w hich w ould be of particular interest w hen com bined w ith the good corrosion resistance of som e alloys (see below ), opening up potential applications such as coatings on dry bearings for space applications.20 H ow ever, the tendency of m etallic glasses to form shear bands and (in som e cases) partially crystallize ow ing to deform ation m eans their w ear resistance is perhaps not as good as their high hardness w ould suggest. Nevertheless, the w ear resistance of m etallic glasses can still be quite good, and in fact one of the principal current m arkets for am orphous alloys is as w ear- and corrosion-resistant coatings for tools such as drill bits. Corrosion and Stress-Corrosion Cracking It is frequently stated that m etallic glasses have excellent corrosion resistance, but this is not alw ays true. The lack of grain boundaries and second-phase particles m akes som e m etallic glasses extrem ely resistant to corrosion, but this is not true of all alloys (som e of w hich oxidize rapidly in air). Broadly speaking, the corrosion resistance of nickel- and iron-based m etallic glasses is better than that of alloys based on zirconium , titanium , and copper (particularly in environm ents containing chloride ions).21 S om e alloys are susceptible to localized pitting corrosion, probably facilitated by the presence of crystalline inclusions. 10 UNCLASSIFIED/ / M IUHIGULJJSUW L. UNCLASSIFIED/ / BQ n OFFICIAL USE ONEY The subject of stress-corrosion cracking of m etallic glasses, despite its obvious im portance for structural applications, has received scant attention in the literature. W hat little w ork that has been done has focused on zirconium -based glasses, w ith the observation that these alloys are very susceptible to stress-corrosion cracking in aqueous environm ents containing chloride ions, likely ow ing to the fact that they do not form protective oxide surface layers.22 M echanical B ehavior at Elevated Tem perature The discussion above relates to m echanical behavior at tem peratures w ell below the glass transition tem perature. At elevated tem peratures, the strength drops and plastic deform ation transitions to a hom ogeneous m ode, occurring throughout the specim en instead of being localized into shear bands (Figure 6). Above the glass transition tem perature, the alloy becom es a fluid, w ith a viscosity that drops exponentially w ith increasing tem perature. Because the strength of the m aterial is low , tem peratures either above or below the glass transition m ay be useful for processing, as discussed above. H ow ever, the decrease in strength and the tendency for crystallization at elevated tem peratures preclude use of m etallic glasses from structural applications at tem peratures approaching the glass transition tem perature. T /T g Figure 6. Deform ation M ap for M etallic Glasses. As a function of tem perature (norm alized to the glass transition tem perature) and applied shear stress t (norm alized to the shear m odulus, p). At high stresses, plastic deform ation occurs inhom ogeneousiy, being localized into shear bands. At high tem peratures, plastic deform ation becom es hom ogeneous. The dashed lines represent different strain rates. The absolute stresses given are representative of the w ell-studied bulk m etallic glass Z r4 1,2Ti 13.8Cu 12.5Ni 10Be22.5, but the general features of the m ap are expected to apply to all m etallic glasses.23 11 UNCLASSIFIED// BOH OFFICIAL UOC CNET UNCLASSIFIED / / FOR OFFICIAL USE ONLY Other Properties; M agnetic, Electrical, Optical, Therm al, and Acoustic Although m ost of the current interest in m etallic glasses centers on their m echanical properties, it is appropriate to consider other properties of potential utility. Of these, the m agnetic properties of ferrom agnetic m etallic glasses stand out.24 A variety of ferrom agnetic glass-form ing alloys exist, m ostly based on transition m etals (iron, nickel, and cobalt). The presence of alloying elem ents (necessary to m ake the m aterial glass-form ing) m eans the saturation m agnetization of m etallic glasses is not as large as that of the pure elem ents. H ow ever, som e am orphous alloys have very low coercivity (a m easure of how strong a m agnetic field m ust be to change the direction of m agnetization of the m aterial) ow ing to the lack of crystalline defects (such as grain boundaries) and m agnetocrystalline anisotropy. In addition, the relatively high electrical resistivity of am orphous alloys (see below ) m inim izes eddy current losses caused by high-frequency m agnetization/dem agnetization. S om e am orphous alloys also have strong m agnetoelastic effects (coupling betw een m agnetic properties such as susceptibility or m agnetization and elastic strain). Current and potential future applications of these m agnetic properties are discussed below . Like crystalline alloys, m etallic glasses have conduction electrons that m ake them both electrically and therm ally conductive,25 although their structural disorder and high alloy content m ake them poor conductors. In addition, in a behavior that is useful in som e applications, the conductivity of m etallic glasses is not very sensitive to tem perature; an exception is near absolute zero, w here som e am orphous alloys becom e superconducting, Another consequence of the am orphous structure of m etallic glasses is that they tend to have very low acoustic dam ping. This m ay be useful in applications such as vibrating­ structure gyroscopes for vehicle orientation.26 A com m on m isperception am ong those hearing about m etallic glasses for the first tim e is to think they are transparent. This is not the case; am orphous alloys are highly reflective, w ith a shiny luster sim ilar to that of other m etals (Figure 7), This is a result of the presence of the conduction electrons, w hich scatter and absorb incident light. 12 U N C LASSI FI ED/ / FOP AEHCIAh UGE ONLY UNCLASSIFIED/ / FOK OFFICIAL USE ONhY Figure 7. Cast M etallic Glass W edge. W edge of a zirconium -based bulk m etallic glass produced by casting. Note the shiny m etallic luster, typical of m etallic glasses.27 M etallic Glass M atrix Com posites As discussed above, the lack of crystalline defects gives m etallic glasses high strength but com prom ises their ductility and fracture toughness. In particular, the tendency for plastic deform ation to localize into shear bands prevents the m aterial from deform ing in a "graceful" m anner. S o it should not be surprising that there have been m any attem pts to control shear band initiation and propagation by m aking com posite m aterials consisting of particles or fibers of som e other m aterial (m ost com m only a ductile crystalline m etal) in a m etallic glass m atrix. The idea is to produce a m aterial w ith im proved ductility, fracture toughness, and fatigue properties w hile (hopefully) not sacrificing the qualities— especially strength and processing flexibility— that m ake m etallic glasses interesting in the first place. PROCESSING AND STRUCTURE OF COM POSITES Broadly speaking, there are tw o kinds of m etallic glass m atrix com posites: ex situ and in situ. In ex situ com posites, the m etallic glass and the crystalline phase (be it in the form of particles or fibers) are physically com bined, for instance by adding particles to the m elt before casting. In situ com posites are different in that the crystalline phase is produced directly from the m elt (by precipitation) during processing. This fundam ental difference in processing leads to significant differences in structure and therefore in properties. 13 UNCLASSIFIED// FOR OFFICIAL USE ONLY UNCLASSIFIED/ / EOB Q K IfTftl M SB OM I Y EX SITU COM POSITES There are tw o basic w ays of m aking ex situ com posites, in w hich the m etallic glass m atrix and the crystalline phase are com bined physically, w ithout a chem ical reaction: • Add crystalline particles to a m elt of a glass-form ing alloy and then cast under conditions that allow the m atrix to form a m etallic glass. • M ake a preform of a crystalline phase (by packing fibers into a m old, for instance) and then cast the glass-form ing alloy around the preform . Both approaches have lim itations. In the first, the addition of particles to the m elt increases the viscosity (w hich is already quite high relative to non-glass-form ing alloys) considerably, ultim ately to a point w here casting becom es im possible. This lim its the volum e fraction of particles that can be added, w hich in turn lim its the control one has over the m icrostructure and, in particular, the spacing of the particles. W ith a perform , the volum e fraction of the crystalline phase can be m uch higher (up to about 80 percent by volum e), but the problem then is how to infiltrate the high-viscosity m elt into the preform w ithout leaving voids and w hile still ensuring sufficiently rapid cooling to form a glassy m atrix. W ith both approaches, interfacial reactions betw een the crystalline phase and the m elt can cause partial or com plete crystallization of the m atrix, degrading the m echanical properties. IN SITU COM POSITES The difficulty of m aking satisfactory ex situ com posites has led to the developm ent of a new approach in w hich the crystalline phase is precipitated directly from the m elt, either during casting28 or in a separate step prior to casting.29 30 Precipitation during casting, although easier, is problem atic from a practical standpoint because variations in the cooling rate (from the surface to the center of a casting, for instance) lead to significant variations in structure and, hence, in properties. One of the m ost prom ising recent advances in the m etallic glass field is the developm ent of in situ com posites in w hich the crystalline phase is precipitated as dendrites, either during casting (Figure 8) or by holding the alloy at an elevated tem perature prior to casting.31 By suitably choosing alloy com position, holding tim e, and tem perature, the volum e fraction, size, and spacing of the dendritic phase can be controlled. This control provides great flexibility in determ ining the m echanical properties of the resulting m aterial. Because the crystalline phase is produced prior to casting, variation in the cooling rate across the casting is m uch less im portant, though the cooling rate m ust still be sufficiently high to ensure the m atrix form s a glass during cooling. Once the glassy m atrix is form ed, the com posite can be reheated above the glass transition tem perature, allow ing for therm oplastic form ing in a m anner sim ilar to single-phase m etallic glasses (as described above). Finally, the presence of the dendritic second phase allow s for deform ation processes (for exam ple, by cold rolling or forging), sim ilar to crystalline alloys.32 14 UNCLASSIFIED// FOW OFFICIAL USE ONLY UNCLASSIFIED/ / FOR OFFICIAL UOt ONL¥ Figure 8* M icrostructure of In Situ M etallic Glass M atrix Com posite, W ith ductile crystalline dendrites. (a) S canning electron m icrograph show ing the dendrites (light gray) in the glassy m atrix (dark gray). (b) Com posite after plastic deform ation; note the m ultiplicity of slip steps, indicating extensive interaction of shear bands w ith the dendrites?3 The key lim itation of these in situ com posites is that not every alloy system is capable of form ing them . W hile any alloy w ill form crystalline phases at elevated tem peratures, usually the crystalline phases that form are brittle interm etallics that degrade rather than enhance the m echanical properties. To be effective in controlling shear bands, the precipitated phase needs to be ductile, have a shear m odulus low er than that of the glassy m atrix, and (preferably) form as dendrites. To date, the only published reports of system s that satisfy these criteria concern alloys based on early transition m etals, notably zirconium and titanium . W hether in situ com posites can be developed in other alloy system s rem ains to be seen. M ECHANICAL PROPERTIES OF COM POSITES The ability to produce m ixed am orphous-crystalline m icrostructure provides the ability to control the form ation and propagation of shear bands. The resulting m aterials can have good fracture and fatigue resistance w hile retaining the high strength and processing flexibility associated w ith m etallic glasses. The origin of these effects is related to the developm ent of a region of plastic deform ation at the tip of an advancing crack. For a crack opening under tensile loading, the size of the plastic region is approxim ately given by: (Equation 1) w here K ic is the plane-strain fracture toughness (m entioned above) and o, is the yield strength. The size of the plastic zone varies from ~ 1 pm for "intrinsically brittle" m etallic glasses to ~ 1 m m for glasses capable of som e plastic deform ation.34 If the m aterial has structure on this length scale (or if the sam ple itself is of this size), then 15 UNCLASSIFIED// FORfiFFTCTM ikf^hh UNCLASSIFIED/ / CAR OFFICIAL USE UNET deform ation can proceed in a stable m anner by generation and subsequent arrest of shear bands. The key to com posite design is to produce a m icrostructure w ith the correct length scale to prevent propagating shear bands from becom ing catastrophic cracks. This turns out to be relatively difficult w ith ex situ com posites, for reasons of processing described above. As a result, the recently developed dendritic in situ com posites have the m ost prom ising properties, and w e focus the rem ainder of our discussion on them . STRENGTH AND DUCTILITY: PLASTIC DEFORM ATION As w ith other com posite m aterials, the yield strength of m etallic glass m atrix com posites can be approxim ated as a sim ple rule of m ixtures based on the volum e fraction of the tw o phases. Because the ductile crystalline phases useful for lim iting shear band propagation are w eaker than the am orphous m atrix, in producing a com posite, som e sacrifice in strength is inevitable. H ow ever, the gains in tensile ductility can be significant. For instance, m onolithic titanium -based m etallic glasses (like all m etallic glasses) have essentially zero tensile ductility, but in situ com posites based on titanium have been reported w ith tensile elongation as large as 12 percent.35 This is com parable to the ductility of Ti-6AI-4 V (the m ost com m on conventional titanium alloy), but in a m aterial w ith about 30 percent greater strength. The properties of m etallic glass m atrix com posites and m ore conventional m aterials are further com pared below . FRACTURE AND FATIGUE The developm ent of a stable plastic zone m eans additional energy is required for crack propagation, m aking in situ com posites m uch m ore resistant to fracture and fatigue than are single-phase glasses. For instance, the plane-strain fracture toughness of som e zirconium -based in situ com posites can exceed 170 M Pa m 1/2— 7 tim es greater than that of single-phase glasses and greater than that of virtually any other m etallic alloy.36 This resistance to crack propagation is also m anifested as im proved fatigue perform ance. The fatigue strength of the zirconium -based in situ com posites is 20-30 percent of the tensile strength; in com parison, m onolithic m etallic glasses have a fatigue strength of only ~ 5 percent of the tensile strength.37 The fatigue strength of the in situ com posites is thus com parable to that of conventional structural alloys. Aerospace Applications of M etallic Glasses STRUCTURAL APPLICATIONS The key properties of m aterials for structural applications in aerospace are: • S trength. • S tiffness (Young's m odulus). • Density (w eight). • Fracture toughness (dam age tolerance). U N F I A C C TF T Fn //F »P n F F T C lftl IIC C M IM I V 16 UNCLASSIFIED/ / FOR OFFICIAL UOE ONLY • Fatigue resistance (including resistance to both fatigue crack initiation and fatigue crack grow th). • Corrosion resistance (including stress-corrosion cracking). • Cost (including raw m aterials, shaping, and assem bly). Figure 9 illustrates the m echanical properties of m etallic glasses and m etallic glass m atrix com posites com pared w ith other structural m aterials. S ince w eight is a particular concern in aerospace applications, in Figure 9 (a) w e norm alize both yield strength (ay) and stiffness (E) to density (p); tw o m aterials w ith the sam e specific strength (ay /p) or specific stiffness (E/p) could be used to produce a com ponent w ith the sam e overall strength or stiffness, respectively, at the sam e w eight. M aterials in the upper-right corner of the plot have the best com bination of strength and stiffness for a given w eight. Notice that the m etallic glasses (and dendritic com posites) can be stronger than virtually all crystalline m etals, although the stiffness of m etallic glasses tends to be som ew hat sm aller than that of crystalline alloys of sim ilar com position. Figure 9 (b) illustrates the dam age tolerance of m etallic glasses com pared w ith other m aterials. By plotting the fracture toughness (K Ic) against m odulus (E), w e can also com pare the fracture energy (G Ic ~ (K Ic)2/E) of the m aterials; the dashed diagonal lines are lines of constant fracture energy. Figure 9 (b) reveals several interesting aspects of the dam age tolerance of m etallic glasses. First, although the fracture toughness of som e m etallic glasses is com parable to that of crystalline m etals, som e m etallic glasses— m ost notably those based on iron (Fe) and m agnesium (M g)— are as brittle as any ceram ic. S econd, both the fracture toughness and the fracture energy of the dendritic m etallic glass m atrix com posites can be superior to those of all but the m ost fracture-resistant m etals. These considerations suggest the dendritic m etallic glass m atrix com posites m ight indeed find applications as structural m aterials in aircraft and/or spacecraft. The m ost obvious applications w ould be to replace steel in certain com ponents w here strength is critical but space is lim ited. These m ight include pylon structures and landing gear,38 although it has yet to be dem onstrated that the com posites can be fabricated in the sizes necessary. Furtherm ore, the corrosion and stress-corrosion cracking resistance of these m aterials has not been fully evaluated. 17 UNCLASSIFIED// FOR OFFICIAL USE UNCT UNCLASSIFIED//FOR OFFICIAL U5E ONLY Specific stiffness. E /p (MPa m3 kg ') Stiffness. E (GPa) Figure 9. M aterials Property Charts, (a) S trength and stiffness (both norm alized to density) of m etallic glasses (yellow ) and dendritic m etallic glass m atrix com posites (red) com pared w ith other m aterials, (b) Dam age tolerance. On this plot, the dashed lines represent contours of equal fracture energy. In both plots, polym er com posites (CFR P and G FR P) are represented by isotropic averages; continuous fiber com posites can have greater strength and stiffness in a direction parallel to the fibers.39 18 UNCLASSIFIED// BOB OFFICIAL USE ONLY UNCLASSIFIED/ / tQ R OFFICIAL UCE ONfeY M etallic glass foam s (see above) also provide intriguing possibilities for structural applications. It has recently been show n that m etallic glass foam s w ith outstanding strength can be form ed by controlling the size of the ligam ents betw een pores.4 0 This is a new developm ent, and these foam s have not been fully characterized, but it seem s likely that optim ized foam s w ill have a specific stiffness (E/p) superior to that of polym er foam s, along w ith high strength and acoustic dam ping. S uch structural foam s could be useful in applications requiring strength and stiffness under com pressive loads, such as structural panels for extraterrestrial buildings. Conceivably, such structural foam s m ight even be produced on site (from raw feedstock), reducing the volum e of m aterial that needs to be launched. A final possibility is that m etallic glasses m ight be com bined w ith polym er com posites into m etal-fiber lam inate m aterials. S im ilar lam inates (w ith crystalline alum inum alloys) are being em ployed in large quantities on the new Airbus 380 and are likely to find increased application in the future.4 1 The use of m etallic glasses in these lam inates is appealing because of their high specific strength (although the specific stiffness is low er than that of alum inum ). Furtherm ore, the individual layers in the lam inate are sufficiently thin that a w ide range of glass-form ing alloys m ight be considered (in contrast to thicker structural sections, w hich w ill be lim ited by the glass-form ing ability of the alloy). OTHER APPLICATIONS M onolithic m etallic glasses are unique am ong m etallic m aterials in having no m icrostructure at length scales of m ore than a few atom ic spacings. In principle then, m etallic glasses should be capable of replicating features dow n to this scale. This possibility is facilitated by the ability of m etallic glasses to be form ed in the supercooled liquid tem perature range w ith controllable viscosity. Indeed, superplastic form ing of m etallic glass surfaces w ith features as sm all as 13 nanom eters has been dem onstrated.4 2 This ability could be exploited for direct em bossing of nanostructures in polym ers or other m aterials. S tructures on this length scale are also potentially useful as diffraction gratings for ultraviolet and soft x-ray radiation. In a related area, m etallic glasses have a variety of useful properties for application in m icro-electrom echanical system (M EM S) actuators, including large elastic strains and high resilience (elastic strain energy storage), good corrosion and w ear resistance, and an excellent surface finish.4 3 The scale of these devices is sm aller than the plastic zone size (Equation 1 above), m aking brittle fracture unlikely. Furtherm ore, a m uch w ider variety of am orphous alloys can be m ade in thin film form (by vapor deposition) than is possible by casting. Finally, the m agnetic properties of certain am orphous alloys have long been exploited. For instance, their low coercivity and high electrical resistivity m ake ferrom agnetic am orphous alloys attractive as high-efficiency electrical transform ers, particularly at high frequencies. S uch applications are likely to continue w ell into the future. 19 UNCLASSIFIED// TOR OFFICIAL USE ONLY UNCLASSIFIED / / FOR OFFICIAL UOE ONLY Current Challenges and Prospects for the Future ALLOY DESIGN A critical lim itation of existing m etallic glass technology (and related com posites) is the relative dearth of alloys w ith good glass-form ing ability. The best glass-form ing alloys are either based on expensive elem ents (for exam ple, palladium ) or contain toxic elem ents (for exam ple, beryllium in the best zirconium - and titanium -based alloys). For aerospace applications, the m ost glaring lack is that, despite significant alloy design efforts in the United S tates (through the DAR PA S tructural Am orphous M etals program ), Japan, China, and elsew here, there are no good glass-form ing alloys based on alum inum . Attem pts to m ake alum inum -based m etallic glass com ponents by consolidating am orphous pow ders have m et w ith lim ited success. S im ilarly, all of the good iron-based m etallic glasses contain considerable am ounts of nonm etallic elem ents (notably carbon, boron, silicon, and/or phosphorus), w hich are thought to contribute to the very low fracture toughness of these alloys (Figure 9 (b)). H ow ever, there is reason to expect that further progress is possible. R ecent experim ental results have show n that som e of the em pirical "rules" of glass-form ing ability4 4 are actually quite flexible, and that glass-form ing ability is m uch m ore sensitive to com position than had been previously appreciated.4 5 S o it is highly probable that som e excellent glass-form ing alloys com positions rem ain to be discovered, possibly including som e low -density glasses based on alum inum . Identifying these good glass-form ing alloys w ill be a challenge. M ost alloy developm ent to date has been done w ith a brute-force approach, but com binatorial techniques4 6re likely to enable m uch m ore rapid screening. One issue is identification of suitable m etrics for glass-form ing ability, since the com binatorial approaches use vapor- deposited thin film s, and it is not clear w hat characteristics of such a film correlate w ith glass-form ing ability in the bulk. S im ilarly, continued developm ent of ab initio m olecular dynam ics techniques should enable identification of candidate alloys from com puter sim ulations, particularly as com puters continue to increase in pow er. One area that has received insufficient attention is the influence of processing conditions on glass-form ing ability. For instance, application of electrom agnetic vibrations during cooling reportedly significantly enhances the glass-form ing ability of m agnesium -based m etallic glasses.4 7 This approach could, in principle, be applied to other alloys, possibly greatly extending the range of alloys and com positions that can be produced as bulk m etallic glasses. THERM OPHYSICAL PROPERTIES AND THERM OPLASTIC PROCESSING M ost of the practical interest in single-phase (m onolithic) m etallic glasses centers on the potential for therm oplastic processing near to or above the glass transition tem perature. H ow ever, the therm ophysical properties and behavior of m etallic glasses are not w ell understood. For instance, the viscosity of the m etallic glass m elt (or supercooled liquid) is of critical im portance, but w e do not know how and w hy alloy com position influences viscosity. From an engineering point of view , the practical aspects of m olding of m etallic glasses are just beginning to be explored. Certainly m any 20 UNCLASSIFIED/ / fcQ B OFFICIAL USE ONLY UNCLASSIFIED / / FOK OFFICIAL USE UNLI parallels can be draw n w ith therm oplastic form ing of polym ers, but there are certain to be m any differences as w ell. Continued developm ents in this area are highly likely to result in the ability to produce com plex net-shape parts in a single processing step. This ease of processing could offset the higher raw m aterials costs for m etallic glasses, m aking them com petitive in a m uch w ider range of applications. Furtherm ore, as noted above, the ability to replicate extrem ely sm all features (< 20 nanom eters) in m etallic glasses is likely to be exploited in the m anufacture of nanostructured devices.4 8 4 9 Finally, developm ent of m etallic glass foam s w ill continue and w ill be aided by im proved understanding of therm ophysical properties. It is highly likely that foam s w ill be produced in a w ide range of glass-form ing alloys, and that techniques w ill be developed for precise control of the porosity, pore size, ligam ent size, and connectivity. This w ill allow the properties of these foam s to be tailored to particular applications. COM POSITES AND THE Q UEST FOR DUCTILITY From the point of view of structural applications, localization of plastic deform ation into shear bands is the single biggest challenge because this tendency lim its the tensile ductility, fracture toughness, and fatigue crack resistance of m etallic glasses. There m ay w ell be no solution to this problem for m onolithic m etallic glasses, for the sim ple reason that they lack any m icrostructure to interact w ith shear bands. Progress is likely to occur on tw o fronts. First, it is now w ell established that som e alloys are inherently brittle, in the sense that they experience very little plastic deform ation around a crack tip, w hile other alloys show extensive plastic deform ation (albeit localized into shear bands). The precise reason for this difference is not understood at present, but it seem s likely that it w ill be resolved w ith continued w ork on fundam ental aspects of plastic deform ation and fracture. This is likely to lead to developm ent of new alloys w ith reasonable fracture toughness, although not to tensile ductility. H ow ever, even this w ill be an im portant step if such alloys can be used as m atrices for com posites. S econd, in order to achieve tensile ductility, it appears to be necessary to have som e m icrostructural features to interact w ith shear bands. Furtherm ore, the length scale of the m icrostructure is clearly a critical param eter in arresting shear band propagation. Again, the precise reasons for this are not know n, but continued research quite likely w ill lead to an im proved understanding of the interactions betw een second-phase particles and shear bands. At present, the m ost prom ising approach to producing com posite m aterials w ith the proper m icrostructural length scale is the form ation of dendritic com posites, as discussed above. A critical lim itation is that this process has been dem onstrated in only tw o, closely related alloys and does not appear to be a general phenom enon. Unfortunately, our understanding of therm odynam ics and phase form ation in com plex m ulticom ponent alloys is not such that w e can predict a priori w hich alloys are capable of producing ductile dendrites in a glass-form ing m atrix. Until that understanding is developed, discovery of new dendritic com posite m aterials w ill rem ain a m atter of trial 21 UNCLASSI FI Fn/ /F»° OFnrTfti hceqniy UNCLASSIFIED / / FOR OFFICIAL USE ONLY and error. The potential benefits are significant, how ever, because there exists the possibility of m aking m aterials w ith exceptionally high strength, fracture toughness, and fatigue resistance. Sum m ary and Recom m endations M etallic glasses com bine som e of the advantageous m echanical properties of m etals— strength, stiffness, and in som e cases toughness— w ith the processing flexibility usually associated w ith therm oplastic polym ers. The absence of crystalline defects allow s m etallic glasses to be m uch stronger than conventional alloys but also m eans they have near-zero tensile ductility and poor fatigue resistance. In structural applications, therefore, m etallic glasses are m ost likely to be useful in the form of com posites consisting of ductile crystalline dendrites in a m etallic glass m atrix. These dendritic com posites sacrifice som e strength but can have exceptionally high fracture toughness, as w ell as good fatigue resistance, and could replace high-strength steels in certain load-lim ited structural com ponents in aerospace vehicles w here space is lim ited. Because they are true glasses, therm oplastic form ing near the glass transition tem perature affords m etallic glasses trem endous flexibility in processing. For instance, m etallic glass com ponents can be form ed in a single step (for exam ple, by injection m olding) in com plex geom etries that w ould be difficult or im possible to produce w ith conventional alloys. In addition, m etallic glass foam s can be m ade w ith relative ease, raising the possibility of m aking structural foam s w ith high strength and stiffness. Finally, because they lack a crystalline grain structure, m etallic glasses can be used to form nanoscale features w ith high fidelity. This m ay m ake m etallic glasses useful in a variety of m icro-electrom echanical system s (M EM S ) applications. M etallic glasses also have significant lim itations for aerospace applications, how ever. Forem ost am ong these is a lack of good glass-form ing alloys; in particular, there are no good alum inum -rich glass-form ing alloys, the know n titanium -based alloys are either relatively dense (ow ing to high concentrations of alloying elem ents) or contain beryllium , and the know n m agnesium - and iron-based alloys are all quite brittle, w ith low fracture toughness. Although m etallic glass m atrix com posites can have outstanding properties (particularly strength and fracture toughness), the num ber of good com posite system s know n at present is also quite lim ited. For m etallic glasses (and their com posites) to be of broad utility in aerospace structural applications, progress in the follow ing areas is required: • Developm ent of new lightw eight alloys and com posite system s, preferably by com putational and/or com binatorial approaches rather than by trial and error. • Understanding of m echanical behavior, especially: - The effect of alloy com position and structure on plastic deform ation. - M icrostructural design of com posites for optim al toughness. • Developm ent of processing techniques, including therm ophysical processing of com plex and/or nanoscale features as w ell as production of m etallic glass foam s. 22 UNCLASSIFIED// FAR Q EHCUU iicc nM iJt UNCLASSIFIED// rOR OrriGIAL USE ONLY It is highly likely that continued w ork over the next 20-50 years w ill result in significant advances in all these areas, and that m etallic glasses and m etallic glass m atrix com posites w ill see increasing acceptance as structural m aterials. W hether or not they achieve w idespread use in aerospace applications, how ever, depends critically on the developm ent of new , lightw eight alloys. 1 Z. P. Lu, Y. Liu, and C. T. Liu, Chapter 4 in Bulk M etallic G lasses, M . M iller and P. K . Liaw , eds. (S pringer, 2009 ). 2 H . M en, W . T. K im , and D. H . K im , M ater. Trans. 4 4 , 214 2 (200 3). 3 A. Peker and W . L. Johnson, Appl. Phys. Lett. 63, 234 2 (19 9 3). 4 N . N ishiyam a and A . Inoue, M ater. Trans. JIM 37, 1531 (1996). 5 W . Zhang, Q. S . Zhang, and A. Inoue, M ater. Trans. 50, 679 (2009 ). 6 J. S chroers and W . L. Johnson, Appl. Phys. Lett. 84 , 3666 (2004 ). 7 F. G uo, H . W ang, S . J. Poon, and G . J. S hiflet, Appl. Phys. Lett. 86, 09 19 07 (2005). 8 V. Ponnam balam , S . J. Poon, and G . J. S hiflet, J. M ater. R es. 19 , 1320 (2004 ). 9 A. H . Brothers and D. C. Dunand, S cripta M ater. 54 , 513 (2006). 10 X . J. G u, S . J. Poon, and G . J. S hiflet, J. M ater. R es. 2 2, 34 4 (2 007). 11 Data from M atW eb, < > .w w w .m atw eb.com 12 Data from M atW eb, < > .w w w .m atw eb.com 13 Y. H e, G . M . Dougherty, G . J. S hiflet, and S . J. Poon, Acta M etall. M ater. 4 1, 337 (19 9 3). 14 G raphic reproduced from C. A. S chuh, T. C. H ufnagel, and U. R am m am urty, Acta M ater. 55, 4 067 (2007) and used w ith the perm ission of Elsevier, Ltd. Original m icrograph Is from R . D. Conner, W . L. Johnson, N. E. Paton, and W . D. Nix, J. Appl. Phys. 9 4 , 9 04 (2003). 15 P. Low haphandu and J. J. Lew andw oski, S cripta M ater. 38, 1811 (19 9 8). 16 Data from M atW eb, < > .w w w .m atw eb.com 17 C. J. G ilbert, V. S chroeder, and R . 0. R itchie, M etall. M ater. Trans. A 30, 1739 (19 9 9 ). 10 B. M enzel and R . H . Dauskardt, Acta M ater. 54 , 9 3 5 (2006). 19 M . E. Launey, D. C. H ofm ann, W . L. Johnson, and R . 0. R itchie, Proc. Nat. Acad. S ci. 106, 4 9 86 (200 9 ). 20 A. L. G reer, K . L, R utherford, and I. M . H utchings, Int. M ater. R ev. 4 7, 8 7 (2002). 21 B. A. G reen, P. K . Liaw , and R , A. Buchanan, Chapter 8 in Bulk M etallic G lasses, M . M iller and P. K .Liaw , eds. (Springer, 2009 ). 22 V. S chroeder and R . O. R itchie, Acta M etall. 54 , 1785 (2006). 23 G raphic reproduced from C. A, S chuh, T. C. H ufnagel, and U. R am m am urty, Acta M ater. 55, 4 067 (2007) and used w ith the perm ission of Elsevier, Ltd. 24 A. H ernando and M . Vazquez, Ch. 17 in R apidly S olidified Alloys, H . H . Lieberm ann, ed. (M arcel Dekker, 19 9 3). 25 T. R ichm ond and H J. G untherodt, Ch. 14 in R apidly S olidified Alloys, H . H . Lieberm ann, ed. (M arcel Dekker, 19 9 3). 26 C. H aon, D. Cam el, B. Drevet, and J. M . Pelletier, M etall. M ater. Trans. A 39 , 179 1 (2008). 27 Photograph by Todd H ufnagel. 28 C. C. H ays, C. P. K im , and w . L. Johnson, Phys. R ev. Lett. 84 , 29 01 (2000). 29 C. Fan, R . T. Ott, and T. C. H ufnagel, Appl. Phys. Lett. 81, 1020 (2002). 30 D. H ofm ann, J.-Y. S uh, A. W iest, G . Duan, M .L. Lind, M . D. Dem etriou, and W . L. Johnson, Nature 4 51, 1085 (2008). 31 D. H ofm ann, J.-Y. S uh, A. W iest, G . Duan, M .L. Lind, M . D. Dem etriou, and W . L. Johnson, Nature 4 51, 1085 (2008). 32 D. H ofm ann, J.-Y. S uh, A. W iest, G . Duan, M .L. Lind, M . D. Dem etriou, and W . L. Johnson, Nature 4 51, 1085 (2008). 33 G raphics reproduced from C. A. S chuh, T. C. H ufnagel, and U. R am m am urty, Acta M ater. 55, 4 067 (2007) and used w ith the perm ission of Elsevier, Ltd. Original artw ork provided by Charlie H ays, Caltech. 34 M . F. Ashby and A. L. G reer, S cripta M ater. 54 , 3 21 (2 006). 35 D. C. H ofm ann, J.Y. S uh, A. W iest, M .L. Lind, M . D. Dem etriou, and W . L. Johnson, Proc. Nat. Acad. S ci. 105, 20136 (2008). 36 D, H ofm ann, J.-Y. S uh, A. W iest, G . Duan, M .L. Lind, M . D. Dem etriou, and W . L. Johnson, Nature 4 51, 1085 (2008). 37 M . E. Launey, D. C. H ofm ann, W . L. Johnson, and R . 0. R itchie, Proc. Nat. Acad. S ci. 106, 4 9 86 (200 9 ). 38 N. Barrington and M . Black, Ch. 1 in Aerospace M aterials, B. Cantor, H . Assender, and P. G rant, eds. (Institute of Physics, 19 9 8). 39 Data for m etallic glasses are from X. J. G u, S . J. Poon, and G . J. S hiflet, J. M ater. R es. 2 2, 34 4 (20 07); D. H ofm ann, J.-Y. S uh, A. W iest, G . Duan, M .L. Lind, M . D. Dem etriou, and W . L. Johnson, Nature 4 51, 1085 (2008); M . F. Ashby and A. L. G reer, S cripta M ater. 54 , 321 (2006); D. C. H ofm ann, J.Y. S uh, A. W iest, M .L. Lind, M . D. Dem etriou, and W . L. Johnson, Proc. Nat. Acad. S ci. 105, 20136 (2008); J. J. Lew andow ski, W . H . W ang, and A. L. G reer, Phil. M ag. Lett. 85, 77 (2005). Data for other m aterials are from Cam bridge M aterials S elector, < > .http://w w w .grantadesign.com / 23 UNCLASSIFIED// FOR Om CIAh USE ONLY UNCLASSIFIED / / FOR OFFICIAL USE ONLY 4 0 M . D. Dem etriou, C. Veazey, J. S . H arm on, J. P, S chram m , and W . L. Johnson, Phys. R ev. Lett. 101, 14 5702 (2008). 4 1 C. Verm eeren, ed. Around G lare: A New Aircraft M aterial in Context (K luw er, 2002). 4 2 D. V. Louzguine-Luzgin, D. B. M iracle, and A. Inoue, Adv. Eng. M ater. 10, 1008 (2008). 4 3 A. L. G reer, M aterials Today 12(1-2), 14 (2009 ). 4 4 D. V. Louzguine-Luzgin, D. B. M iracle, and A. Inoue, A dv. Eng. M ater. 10, 1008 (2008). 4 5 Y. Li, Q. G uo, J. A. K alb, and C. V. Thom pson, S cience 322, 1816 (2008). 4 6 Y. Li, Q. G uo, J. A. K alb, and C. V. Thom pson, S cience 32 2, 1816 (2008). 4 7 T. Tam ura, K . Am iya, R . S . R achm at, Y. M izutani, and K . M iw a, Nature M ater. 4 , 289 (2005). 4 8 G . K um ar, H . X . Tang, and J. S chroers, Nature 4 57, 868 (2009 ). 4 9 A. L. G reer, M aterials Today 12(1-2), 14 (2009 ). 24 UNCLASSIFIED/ / FOR OFFICIAL USE ONLY