Sie sind auf Seite 1von 6

Scripta Materialia 54 (2006) 321326 www.actamat-journals.

com

Metallic glasses as structural materials


M.F. Ashby a, A.L. Greer
a b

b,*

Department of Engineering, University of Cambridge, Trumpington Street, Cambridge CB2 1PZ, UK Department of Materials Science and Metallurgy, University of Cambridge, Pembroke Street, Cambridge CB2 3QZ, UK Received 9 June 2005; received in revised form 13 August 2005; accepted 30 September 2005 Available online 26 October 2005

Abstract The potential of metallic glasses as structural materials is assessed. A wide-ranging comparison with conventional engineering materials shows metallic glasses to be restricted to niche applications, but with outstanding properties awaiting wider application, for example in micro electro-mechanical systems devices. 2005 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
Keywords: Elastic behaviour; Metallic glasses; Toughness; Yield phenomena

1. Introduction Metallic glasses have certain remarkable attributes, but like many materials born of scientic research (in contrast to those developed to ll a specic market need) they come into the world homelessmaterials in search of an application. A useful starting point [1] is to (i) identify the attributes of the new material that are better than those of existing materials, (ii) identify the attributes that are worse, and then (iii) explore applications that exploit (i) and are insensitive to (ii). Table 1 attempts the rst two steps. The dening characteristic of metallic glasses, compared to conventional metallic materials, is their lack of crystallinity, and associated lack of microstructural features such as grain and phase boundaries. The excellent soft-magnetic properties (low coercive force and high permeability) of iron-based compositions, have led to commercial applications such as transformer cores, magnetic read-heads and magnetic shielding [2]. The glasses can also crystallize into very ne, uniform
*

Corresponding author. Tel.: +44 1223 334308; fax: +44 1223 334567. E-mail address: alg13@cam.ac.uk (A.L. Greer).

microstructures, exploited for example for their hard magnetism [3]. While some properties, for example, thermal conductivity, expansion and specic heat, are not unusual, the corrosion resistance and temperature-insensitive electrical conductivity are potentially attractive. Early metallic glasses were typically made by rapid quenching of the melt, with critical cooling rates of 105 106 K s1, severely limiting sample thickness. Starting with Pd-based compositions [4] and now extending to a very wide range of multicomponent alloys [57], critical cooling rates have been dramatically lowered to permit casting in bulk (i.e. with minimum dimension >1 cm). With the advent of these bulk metallic glasses (BMGs), there has been a revival of early interest [8] in mechanical properties. While metallic glasses can be multifunctional (e.g. it is important for a magnetic read head to have good wear resistance [9]), and can be exploited as corrosion- and wear-resistant coatings, the ability to make bulk components has stimulated interest in larger-scale structural applications. As noted in Table 1, BMGs can be moulded with an excellent surface nish, and this has further increased interest. Possible applications of BMGs have been surveyed by many authors [57,10,11]. Under ambient conditions, metallic glasses show a distinctive localization of the plastic deformation into shear bands [8], giving unusual combinations of properties, for

1359-6462/$ - see front matter 2005 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.scriptamat.2005.09.051

322

Table 1 Properties of metallic glasses related to potential applications Attributes General Attractive attributes Absence of microstructural features such as grain and phase boundaries and of related composition variations (e.g. segregation). This allows components with features of near-atomic scale High hardness, H, giving good wear and abrasion resistance High yield strength, ry Fracture toughness Kc and toughness Gc can be very high =3 1=2 High specic strengtha, ry/q, r2 y =q and ry =q 2 High resilience per unit volume and massa, r2 y =E and ry =Eq Low mechanical damping Unattractive Present cost of components and processing Optimization of composition for glass-forming ability prevents easy optimization for other properties, including low density Severe localization of plastic ow (shear-banding), giving zero ductility in tension Fracture toughness Kc and toughness Gc can be very low Can be embrittled by annealing Small process-zone size (d < 1 mm) means that larger components may fail in a brittle manner Instability above Tc limits high temperature use M.F. Ashby, A.L. Greer / Scripta Materialia 54 (2006) 321326

Mechanical

Thermal

Tg < Tc for some metallic glasses, allowing processing as a supercooled liquid (Tgglass-transition temp. Tctemp. of crystallization onset) High magnetic permeability Resistivity is nearly independent of temperature Lack of grain structure and associated microstructural features (e.g. solute segregation) gives corrosion resistance Some compositions biocompatible Low solidication shrinkage and lack of grain structure give high precision and nish in castings The high viscosity and low strain-rate sensitivity of the supercooled liquid permit thermoplastic forming Lack of grain structure allows a very high polish High hardness and corrosion resistance gives durability Aesthetics, present novelty and rarity make metallic glasses attractive for high-end life-style products Properties and processing favour lm-to-mm scale structures

Electrical and magnetic Chemical Environmental Processing

Relatively high magneto-striction gives energy loss in oscillating eld

Not easily recycled once in a product (non-conventional compositions) Current need for vacuum die-casting gives relatively slow production rate

Aesthetic Potential markets

Current high cost of material and processing limits applications to those with high value-added

Material indices of this kind are discussed in Ref. [14].

M.F. Ashby, A.L. Greer / Scripta Materialia 54 (2006) 321326

323

example zero ductility but high fracture toughness, which impede simple comparison with conventional engineering materials. The mechanical properties of metallic glasses (Table 1) are examined next to assess the potential of metallic glasses for structural applications. 2. Comparison of mechanical properties Enough data are now available on metallic glasses, mostly BMGs, to permit quantitative comparison of their properties (as a class of material) with conventional engineering materials. Fig. 1 shows elastic limit ry and Youngs modulus E for over 1500 metals, alloys and metalmatrix composites. On materials-selection charts of this kind, the ellipses enclose the range of values associated with given materials and material groups. The data on the conventional metallic materials are from a standard database [12]; the metallic glasses are identied by their compositions (in at.%). The metallic glasses stand out, lying at the upper edge of the populated region. They have high strength (the highest known exceeds 5 GPa [13]), which shows a correlation with Youngs modulus. The shading shows one fundamental boundary of the accessible region of the plot, limited by the theoretical strength (ry = E/ 20), which the metallic glasses approach more closely than any other bulk metallic material. The contours on Fig. 1 show the property combinations or material indices ry/E and r2 y =E . (The use of such indices is described in Ref. [14].) The rst of these is the yield

strain; the second we call the resilience, a measure of the ability of the material to store (and return) elastic energy, important for example in springs. By these measures the metallic glasses excel, having a larger yield strain and storing more elastic energy per unit volume than any of the other 1500 materials on the plot. Although not shown in Fig. 1, which is restricted to monolithic materials, it is of interest to note that BMG-based foams can be made, with attractive properties [15,16]. Fig. 2 shows the resilience, r2 y =E and loss coecient, g, a measure of the mechanical damping or energy loss in an elastic load cycle (dened in Ref. [14]). High resilience correlates with low loss coecient, reecting the contribution of local plastic ow to energy loss. This correlation is supported and extended by the inclusion of limited data on metallic glasses. Again the metallic glasses are exceptional. Their combination of high resilience and low damping is attractive for vibrating-reed systems such as gyroscopes and fast-acting springs, or in applications requiring transmission of elastic waves. Fig. 3 shows fracture toughness Kc (a measure of loadbearing capacity before fracture) and modulus E for some 2000 metals, ceramics, glasses and polymers. The metallic glasses (data from Ref. [17]) are identied, as before, by their compositions. For many applications, failure is energy-limited rather than load-limited; the appropriate material index is then not Kc, but rather the toughness Gc K 2 c =E . (The distinction is illustrated by polymers: they have low Kc, but rather good Gc, which is why they

Fig. 1. Elastic limit ry plotted against modulus E for 1507 metals, alloys, metal matrix composites and metallic glasses. The contours show the yield strain ry/E and the resilience r2 y =E.

324

M.F. Ashby, A.L. Greer / Scripta Materialia 54 (2006) 321326

Fig. 2. Resilience r2 y =E and loss coecient g for the same materials as Fig. 1.

Fig. 3. Fracture toughness and modulus for metals, alloys, ceramic, glasses, polymers and metallic glasses. The contours show the toughness Gc in kJ m2.

can be used for impact-resistant parts, for example, on cars.) The metallic glasses remarkably show Gc ranging over four decades, nearly the full known range for all mate-

rial types. The toughest metallic glasses lie among the best of metals. On annealing, some metallic glasses undergo severe embrittlement (see, e.g. Ref. [18], associated with

M.F. Ashby, A.L. Greer / Scripta Materialia 54 (2006) 321326

325

loss of free volume and changes in elastic moduli [17]). Other metallic glasses appear to be intrinsically brittle in their as-cast state [17]. While conventional metallic materials can also be brittle or can be embrittled, the metallic glasses can have particularly poor properties: their Gc can be as low as ceramics but not as low as glass. For brittle materials, there is a fundamental limit: Gc cannot be less than 2c, where c is the surface energy. Surface energy scales with Youngs modulus (c % Er0/20, where r0 is the atomic radius). This limit is indicated by the shaded area at the bottom of Fig. 3. Fig. 4 is a chart of fracture toughness against elastic limit (or yield strength) ry. The diagonal contours show the process-zone size, d, where d K2 c . pr2 y 1

The high toughness of some metallic glasses appears to be at variance with their near-zero ductility (i.e. plasticity in tension), which arises from shear localization. However, their plasticity in compression is measurable and can be large in some cases [1921]. By making a composite dispersion of a ductile crystalline phase in a BMG matrix [22], overall ductility can be achieved.

3. Discussion Applications, actual and potential, of metallic glasses as structural materials can now be considered. Tooling, particularly knife edges: The high hardness suggests use in precision tooling. The lack of grain structure allows a blade to be sharpened to an exceptional edge because there is no length scale above the atomic to limit it, and this is of interest for scalpels. Springs and clubs: The high elastic energy storage per unit volume and mass, and the low damping, give metallic glasses potential as springs. Their successful use in golfclub heads and in tennis racket frames is thought to exploit this feature, and there are many possible applications of springs in devices such as high-speed relays. Information storage and reproduction: Lack of grain structure and high hardness suggest that features of near-atomic scale could be moulded or etched into a metallic-glass surface to make masters for reproducing ultrahigh-density digital data.

If the ultimate size, d, of the plastic zone at the tip of a sharp crack is much smaller than the sample size, fast fracture follows. If d is greater than the sample size, brittle failure is not expected. Shear banding in metallic glasses favours failure which appears macroscopically brittle rather than ductile; nevertheless d remains a useful guide to the sample thickness below which there is likely to be plastic ow without fast fracture. Although some metallic glasses have very high fracture toughness Kc, they have small zone sizes (d < 1 mm) because ry is so large. In terms of d, a brittle metallic glass (for example Fe40Ni40P14B6) is even worse than conventional silicate glasses.

Fig. 4. Fracture toughness and elastic limit for the same materials as Fig. 3. The contours show the process-zone size d in mm.

326

M.F. Ashby, A.L. Greer / Scripta Materialia 54 (2006) 321326

Fashion items: The ability to take high polish and resist abrasion and corrosion (and sheer novelty) suggest the use of metallic glasses for rings, spectacle frames, watch cases, pens, mobile-phone cases and the like, though good mechanical properties remain important. Furthermore, the extensive current work on nding and optimizing new glass-forming compositions is expected to expand the range of exploitable properties. The less attractive features must not be ignored. The limited process-zone size restricts the structural use of metallic glasses to small components. Also, the current expense of metallic glasses limits their structural use to high-end products in which performance or aesthetics play le than price. However, a steady reduction in a greater ro the cost of BMG components is expected as a result of current work on new compositions using less expensive metals, and on improved processing such as treatments to permit the use of lower-purity base metals. MEMS devices: Micro electro-mechanical systems (MEMS) oer outstanding opportunities for the exploitation of metallic glasses. High yield strain, low loss coecient and resilience (and implied good fatigue resistance) are desirable, while low process-zone size and low ductility become unimportant at small scale. Crucially for MEMS fabrication, metallic glasses can be deposited as thin lms and the glass-forming ability in deposition opens up a much wider range of compositions than is possible for BMGs. Possibilities for metallic glasses to replace silicon in MEMS have been discussed [23]. There is at least one case of commercial exploitation: glassy Al3Ti-based hinges are the basis for the rotation of the micro-mirrors in digital light processor (DLP) technology for data projectors [24]. DLP devices with more than 106 mirrors on a single chip are being manufactured, and the sputter-deposited metallic-glass hinges have been tested through 1012 cycles without fatigue failurea good example of an application where the distinctive properties of metallic glasses can be exploited to best advantage. Another is metallic glass foams which show plasticity [15,16], and in which the struts can be of sub-mm crosssections and therefore below the process-zone size. 4. Conclusions While, for example, the magnetic properties of metallic glasses have been exploited for some decades, their mechanical properties are now receiving renewed attention. The distinctive plastic ow in metallic glasses precludes simple comparison with conventional alloys. For structural applications, the comparatively high cost of metallic glasses at present restricts their use to high value-added products, where performance, aesthetics or simple novelty are impor-

tant. Also, with their small process-zone size, typically <1 mm, even bulk metallic glasses are best applied in comparatively small components. For niche applications, however, metallic glasses have some outstanding properties, including: high strength, approaching the theoretical limit; uniquely high capacity for elastic energy storage per unit volume or mass; low damping; and in some cases very high toughness. Such properties, combined with an ability to be deposited as thin lms, makes metallic glasses attractive for many MEMS applications, some of which are already on the market. Acknowledgements chet and J.J. Lewandowski for The authors thank Y. Bre useful discussions. ALG is grateful for support from the European Commission RTN Ductilisation of bulk metallic glasses. References
chet Y, Ashby MF, Greer AL. Adv Eng Mater [1] Salimon A, Bre 2004;6:249. [2] Smith CH. In: Liebermann HH, editor. Rapidly solidied alloys, processes, structures, properties, applications. New York: Marcel Dekker; 1993. p. 61763. [3] Croat JJ, Herbst JF, Lee RW, Pinkerton FE. J Appl Phys 1984;55:2078. [4] Kui HW, Greer AL, Turnbull D. Appl Phys Lett 1984;45:615. [5] Johnson WL. MRS Bulletin 1999;24(10):42. [6] Inoue A. Acta Mater 2000;48:279. [7] Wang WH, Dong C, Shek CH. Mater Sci Eng R 2004;44:45. [8] Davis LA. Metallic Glasses. Metals Park (OH): ASM; 1978. p. 191 223. [9] Kohmoto O, Ohya K, Ojima T. IEEE Trans Magn 1989;25:4490. [10] http://www.liquidmetal.com/applications/. chet Y, Greer AL. Mater Sci Eng A [11] Salimon A, Ashby MF, Bre 2004;375377:385. [12] Cambridge Materials Selector (software), Granta Design Ltd, Cambridge, UK. [13] Inoue A, Shen BL, Koshiba H, Kato H, Yavari AR. Nature Mater 2003;2:661. [14] Ashby MF. Materials selection in mechanical design. 3rd ed. Oxford: Butterworth-Heinemann; 2005. [15] Wada T, Inoue A. Mater Trans 2004;45:2761. [16] Brothers AH, Dunand DC. Adv Mater 2005;17:484. [17] Lewandowski JJ, Wang WH, Greer AL. Philos Mag Lett 2005;85: 77. [18] Lewandowski JJ. Mater Trans JIM 2001;42:633. [19] Xing LQ, Li Y, Ramesh KT, Li J, Hufnagel TC. Phys Rev B 2001;64:180201. [20] Schroers J, Johnson WL. Phys Rev Lett 2004;93:255506. [21] Das J, Tang MB, Kim KB, Theissmann R, Baier F, Wang WH, et al. Phys Rev Lett 2005;94:205501. [22] Hays CC, Kim CP, Johnson WL. Phys Rev Lett 2000;84:2901. [23] Wesseling P, Nouri AS, Lewandowski JJ. Abstract for TMS annual meeting 2005. JOM 2004;56(11):362. [24] Tregilgas JH. Adv Mater Proc 2004;162(10):40.

Das könnte Ihnen auch gefallen