Fracture surface morphology of compressed bulk metallic glass-matrix-composites and bulk metallic glass

Fracture surface morphology of compressed bulk metallic glass-matrix-composites and bulk metallic glass

Intermetallics 14 (2006) 982–986 www.elsevier.com/locate/intermet Fracture surface morphology of compressed bulk metallic glass-matrix- composites an...

444KB Sizes 2 Downloads 248 Views

Intermetallics 14 (2006) 982–986 www.elsevier.com/locate/intermet

Fracture surface morphology of compressed bulk metallic glass-matrix- composites and bulk metallic glass M. Kusy a,d,*, U. Ku¨hn a, A. Concustell a,b, A. Gebert a, J. Das a,c, J. Eckert a,c, L. Schultz a, M.D. Baro b a Institute for Metallic Materials, IFW Dresden, Helmholtzstraße 20, D-01069 Dresden, Germany Department de Fisica, Facultat de Cie`ncies, Universitat Auto`noma Barcelona, Edifici Cc, 08193 Bellaterra, Barcelona, Spain c Physical Metallurgy Division, Department of Materials and Geo Sciences, Darmstadt University of Technology, Petersenstraße 23, D-64287 Darmstadt, Germany d Department of Materials Engineering, Faculty of Materials Science and Technology, Slovak University of Technology, Bottova 24, 917 01, Trnava, Slovak Republic b

Available online 9 March 2006

Abstract The fracture morphology of Zr-based bulk metallic glass-matrix-composites (BMGCs) and Cu-based bulk metallic glass (BMG) after compression testing has been studied. The quasi-static compression fracture surface displays a mixture of three different distinct patterns: veinlike, smooth featureless and river-like features. The last one corresponds to the morphology known from tensile tests of BMGs. Moreover, randomly distributed transversal steps on the fracture plane are also present. This is in contrast to previous studies where a characteristic vein-like pattern is considered a unique feature of the fracture of BMGs under quasi-static uniaxial compression. The presence of different fracture features indicates that the development of the fracture plane occurs in a stepwise mode. q 2006 Published by Elsevier Ltd. Keywords: B. Glasses, metallic; C. Rapid solidification processing; F. Mechanical testing

1. Introduction During the last decade, one of the most important progresses in metallic glass research has been the development of new bulk metallic glasses (BMGs) [1,2], which can be formed at relatively low critical cooling rate (beloww102 K/s). These BMGs have received great attention as a new class of structural materials because of their excellent properties, such as high strength and high elasticity [3,4]. However, one of the major drawbacks of BMGs is their limited plasticity. At room temperature, BMGs fails catastrophically without appreciable plastic deformation under tension [3] and only very limited plastic deformation is observed under compression or bending [3,4]. The limited macroscopic ductility at temperatures below the glass transition temperature Tg represents an important disadvantage of the BMGs and has so far hindered broad application of this class of materials [5]. The problem of limited plastic deformation is related to the inhomogeneous plastic flow of metallic glasses, at * Corresponding author. E-mail address: [email protected] (M. Kusy).

0966-9795/$ - see front matter q 2006 Published by Elsevier Ltd. doi:10.1016/j.intermet.2006.01.017

temperatures below Tg. Under loading the sample deforms elastically until strain localization is initiated in the form of a shear band when reaching a critical local stress or free volume excess [6–8]. When a shear band propagates through the material a drop in the loading force is observed [3,9]. The material localized in the shear band behaves similarly to a thinlayer of a substance with low viscosity between two solid plates and possesses features of a perfect plastic behavior without strain hardening [3]. Therefore, propagation of a single shear band causes catastrophic failure of the material in tension. Since, the permanent deformation is localized only in the planes of a limited number of shear bands, the overall ductility of BMGs is negligible. In contrast, upon compression the material surrounding the shear band recovers elastically and the further propagation of the shear band is ceased. This creates the conditions for the formation of a number of shear bands and limited plastic deformation can take place, as it is actually observed in compressed BMGs [3,9] and [10–12] BMGCs. So far several studies on uniaxial tension [13–16], compression [3,9,13,14,17–19], three point bending [20–22], fracture toughness [23,24], fatigue [25,26] and indentation experiments [9,27] were reported in order to reveal the mechanisms of strain localization, plastic deformation, crack

M. Kusy et al. / Intermetallics 14 (2006) 982–986

initiation and propagation. The present work aims to report on the different fracture features developed during compression of Zr-based BMGCs (with low crystalline volume fraction) and Cu-based BMG. 2. Experimental Zr- and Cu-based master alloys were prepared by arcmelting a mixture of elements with a purity of at least 99.9% under Ar atmosphere. In order to ensure the homogeneity of the master alloys, lumps of 50 g were repeatedly re-melted. The nominal compositions are given in Table 1. From the master alloys, the cylindrical bulk alloys with a length of about 60 mm and a diameter of 3 mm were obtained using injection and suction copper mould casting methods. Their structure was checked using X-ray diffraction (XRD) and transmission electron microscopy (TEM). The compression test specimens were cut with a diameter (d) to length (l) ratio of 1:2 were cut from the rods. Quasi-static compression tests were performed with an INSTRON 8562 electromechanical testing device in a crosshead displacement control mode at strain rate ranging from 5.10K5 to 10K4 sK1. The fracture surfaces of the tested samples were examined with a NIKON SMZ800 optical microscope and a Jeol JSM-6400 scanning electron microscope (SEM). 3. Results Structural analysis reveals that the Cu-based BMGs are fully amorphous and the Zr-based rods exhibit of an amorphous matrix with a minor volume fraction (5 vol %) of crystalline CuZr2 intermetallic compound, as also found by Gebert [28]. and Mattern [29]. The results for the different samples are summarized in Table 1. The samples for the fracture surface analysis were selected using light microscopy. Only those samples, which failed with a macroscopically single fracture plane, were studied further. The angles of the fracture plane inclination with respect to the compression axes of selected samples as well as mechanical properties are given in Table 1. These data agree with results published elsewhere [9,13,14,17,19,30,31]. The SEM image of the Zr61Ti4Nb4Cu14Ni9Al9 BMGC presented in Fig. 1a shows that the fracture surface is covered mostly with a vein-like pattern (see inset in Fig. 1a). A similar fracture morphology was observed for other metallic glasses

983

with different chemical compositions [3,13,14,32,33]. This type of pattern is generally considered to be typical for the compression fracture of BMGs. The uniform arrangement of the veins agrees with the flow direction in the shear plane, as already pointed out by Zhang et al.[13]. Nevertheless, other distinct fracture patterns can be revealed at higher magnification. A river-like morphology is found on the fracture plane of all studied alloys, as exemplified in Fig. 1b (inset, magnified). This morphology appears in the form of islands surrounded by vein-like patterns and intermittent smooth regions that sometimes contain fine striations (Fig. 1b frame ii). Similar fracture surface features were found for the Zr61Ti6Nb1Cu14Ni9Al9 BMGC (see Fig. 2a and b). An enlarged micrograph (Fig. 2b) reveals areas where transversal steps are emerging from the main fracture plane, presenting the localized formation of a river-like pattern (see inset of Fig. 2b). An analogous morphology has been recently found by Xi et al. [34]. on the fracture surface of a brittle Mg-based BMG. Note that, when following the direction opposite to the direction of the shear force, there is no visible distinct boundary for the transition from the river-like pattern to the intermittent smooth regions and to the vein-like pattern. However, there is a sharp discontinuity between the vein-like and the river-like patterns in form of a transversal step on the fracture surface, as exemplified in Fig. 3 for Zr61Ti4Nb4Cu14Ni9Al9. Transversal steps of different sizes ranging from a few micrometers up to hundreds of micrometers were found. The river-like pattern consists of tributaries-fine ridges, which start from the featureless cores (see the arrow in the inset of Fig. 3) localized on the transversal steps and point to the intersection of the main shear plane with the transversal step. It is noteworthy that the river-like fracture pattern formed on the transversal steps resembles the pattern observed on the fracture surface after tensile testing [3,13,14,35]. The fractography of the Cu60Zr30Ti10 BMG is shown in Fig. 4. The central feature of the micrograph is a transversal step with a surface pattern of similar morphology as it was found for the BMGCs (see Figs. 2 and 3). 4. Discussion The presence of different morphologies on the fracture surface raises basic questions about the conditions as well as the time sequence of their formation. During the last century

Table 1 Chemical composition, mechanical properties and fracture plane inclination angles for the different investigated glassy alloys, where 3e and 3p are elastic and plastic strain, respectively BMG type

Chemical composition

Phases *TEM detected

3eC3p at yield strength (%)

Ultimate strength (MPa)

Maximum strain (%)

Fracture plane inclination angle (8)

Zr-based

Zr61Ti6Nb1Cu14Ni9Al9 Zr60Ti4Nb4Cu14Ni9Al9 Cu60Zr30Ti10

AmorphousC CuZr2* AmorphousC CuZr2* Amorphous

2

1705

2.1

43.5

2

1632

3.4

42

1.8

1931

2.4

40

Zr-based Cu-based

984

M. Kusy et al. / Intermetallics 14 (2006) 982–986

Fig. 1. Fractography of Zr61Ti4Nb4Cu14Ni9Al9: (a) overview of the fracture surface with regions of different morphology, the inset shows a detail of the vein-like pattern; (b) enlargement of a region with mixed fracture morphology—vein-like pattern, river-like pattern (frame i) and intermittent smooth regions (frame ii).

the development of the fractography of crystalline alloys resulted in a powerful tool for the prediction or failure analysis. This technique distinguishes the differences of various fracture features and attributes them to crack initiation, crack growth and rapid crack propagation [36]. In discussing our results, we will follow an analogy with the failure evolution of crystalline materials. We will try to identify and define different fracture features from a morphological point of view, and will consider their time-sequence and the possible scenario of their formation. The fracture surface analysis of both the Zr- and Cu-based BMGCs and BMGs reveals three distinct morphologies: (a) vein-like pattern, (b) intermittent smooth regions and (c) riverlike pattern. According to the literature, the vein-like pattern is a region where a considerable high temperature is achieved (temperatures as high as 900 K) due to local heating [15,32,33,37,38]. This morphology covers most of the fracture surface area of the studied alloys. Besides, the dominant vein-like morphology, some intermittent smooth regions are also observed on the examined fracture surfaces. Conner et al. [37] suggested that these smooth regions form during fast crack propagation when the crack tip recovers from a trap of a crystalline particle. After the

particle is sheared due to the high energy accumulated in the elastic-plastic zone of the crack tip, the crack propagates through the glassy matrix forming a smooth region. Subhash et al. [32] used the proposed model of the intermittent smooth region formation to explain a similar feature observed on the fracture plane of a BMG deformed at high strain rate. Both studies confirmed the essential role of fast crack propagation. Since, intermittent smooth regions are also observed in the present work, we propose that conditions similar to high rate crack propagation might be locally established also in our specimens. The river-like pattern is the third distinct morphology found randomly scattered over the fracture surface and frequently covering areas near the edge of the fracture plane. This morphology is a region where fine tributaries merging into coarser rivers are present on the fracture surface. The orientation of the tributaries reflects the imposed local stress and their onset points correspond to the nucleation site of a crack. Merging sets of rivers originating from neighboring flaws are observed to form loops, as it has been already been reported [13,14,35] (see the inset in Fig. 4b). The nucleation centers of the river-like morphology are located either (i) randomly on a transversal plane, (ii) near the intersection of the

Fig. 2. Fractography of Zr61Ti6Nb1Cu14Ni9Al9: (a) overview of the fracture surface with regions of different morphology; (b) enlargement of a region with transversal step and mixed fracture morphology—vein-like pattern, river-like pattern and intermittent smooth regions. Inset shows enlargement of the fracture morphology of the transversal step. The arrow point to the center of the featureless core where tributaries of river-like pattern nucleates.

M. Kusy et al. / Intermetallics 14 (2006) 982–986

Fig. 3. Transversal steps observed on the fracture surface of Zr60Ti4Nb4Cu14Ni9Al9 alloy. The inset shows details of the step-edge overflow and the feature-less core of the river-like pattern nucleation.

shear plane with transversal steps or (iii) on the main fracture plane in the vicinity of the transversal steps. The formation of the latter ones is mainly controlled by the local shear force acting between the corresponding fracture planes and, as a consequence, protrusions in the direction inversed to the shearforce are observed. The morphology of the river-like pattern formed on the transversal steps is very similar to that observed for BMGs fractured in tension [13,14]. However, the presence of local tensile stresses acting on the main fracture plain is not obvious during compression. In order to understand the role of the shear-bands in the formation of transversal steps, the resolved components of the superimposed compression force are overlapped with an array of shear bands or a transversal step penetrating onto the surface of the compression test specimen, as shown in Fig. 5. According to our observations (see Fig. 5) the shear bands can be divided into three distinct groups; (i) shear-bands aligned parallel with the fracture plane (plane i)—primary shear bands, (ii) shear bands perpendicular to the compression axis (plane ii) and (iii) shear bands approximately perpendicular to the fracture plane (plane iii)—secondary shear-bands.

Fig. 4. Transversal step and fracture morphology of neighboring regions formed on fracture surface of Cu60Zr20Ti20 upon compression. Inset shows enlargement of the transversal step with the river-like morphology.

985

Fig. 5. Fractography of Zr60Ti4Nb4Cu14Ni9Al9 alloy shows the fracture feature. The array of shear-bands is overlapped with the resoved s-normal and t-shear (tangential) component of compression force FC. Arrows i, ii and iii indicate primary shear-bands, shear-bands perpendicular to the compression axis and secondary shear-bands, respectively.

It should be considered that the crack nucleation and subsequent crack growth at the shear bands of group iii (secondary shear bands) and group i (primary shear bands) leading to failure are competing, since they have exactly the same stress magnitude. The behavior of the two sites and their interactions needs to be further discussed in terms of the free volume theory [6–8] and latest observations on the excess of free volume in shear-bands [39–42]. Also the calculations of the free energy change upon free volume coalescence should be taken into consideration [43]. The shear bands consist of a thinlayer of low viscous material with chemically and atomically changed short-range order with excess of free volume [40]. However, under compression there is no tensile component in any imaginary plane in the specimen. Moreover, the micromechanics of the intersection point of two different sets of plane (plane group i (primary) and plane group iii (secondary)) consisting of two different sets of shear bands without any shear support should be considered. It was earlier reported [44] that these intersection regions are highly strained and are regions with high stress concentration. Moreover, it is well known that strain accumulation in a shear band in its parallel direction can be very high [3–9]. This strain accumulation can be directly measured as the size of the shear step on the outer surface, which can be more than 10 mm, as evaluated in the present investigation. Contrary, in order to nucleate a crack, the critical displacement in the direction perpendicular to a secondary shear plane is 1 mm for Zr-base metallic glass, as calculated according to Argon et al. [39]. So at the instability prior to failure, a strain in the direction parallel to a shear band can be accumulated easily, whereas the same amount of strain in the direction perpendicular to the secondary shear band (plane iii) can be quite high. This will stimulate physical separation perpendicular to the weaker region, which is parallel to the imaginary planes iii. In this case the system will choose the secondary shear bands (where free volume has already been accumulated previously due to some strain accumulation) in order to join a set of parallel shear bands

986

M. Kusy et al. / Intermetallics 14 (2006) 982–986

along the direction of plane i. Basically, the secondary shear bands act as bridging ligaments between a set of parallel primary shear bands. Such physical separation is very similar to the failure under pure normal stress. This, in turn, helps to develop a core structure and a river-like pattern close to the plane iii and forms transversal steps on the fracture plane, as it is clearly evidenced in Figs. 1–5. In contrast, final failure along a shear plane causes the formation of melting features on the fracture surface leading to a vein-like pattern. 5. Conclusions The fracture surfaces formed during quasi-static compression tests are analyzed for Zr-based BMGCs with low crystalline volume fraction and Cu-based BMG. The obtained results show that the fracture surfaces of the studied alloys present the same motives of fracture morphology: a vein-like pattern, intermittent smooth regions and a river-like pattern. The river-like patterns as well as the intermittent smooth regions cover isolated regions of the fracture surface surrounded by a dominating vein-like pattern. The river-like pattern formed on transversal steps resembles morphology similar to the one observed on the fracture planes of samples loaded in tension. This is explained by easy separation along the secondary shear bands at the instability prior to failure due to the significant difference between values of critical strain accommodation along perpendicular and parallel direction with respect to the shear band planes. Possibly sub-critical crack nucleation starts from sites covered with the river-like pattern. The crack propagation further continues forming the intermittent smooth regions. The veinlike pattern develops formed during the ultimate failure of the compression specimens in the regions of the softened shear bands. Due to the fact that the same features are observed on the fracture surfaces of BMGCs with low crystalline volume fraction and for BMG it can be concluded that the mechanism of fracture surface formation of the two different classes of material is similar. Acknowledgements The authors are grateful to M. Frey, H.-J. Klauss and S. Kuszinski for technical assistance. We acknowledge the financial support of the European Commission under the Research and Training Network ‘Ductile BMG Composites’ MRTN-CT-2003-504692. A.C. is grateful for the support provided by Spanish MEC. Partial finantial suport from DURSI (2001-SGR-00189) and MEC (MAT 2004-01679) projects are also acknowledged.

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44]

Zhang T, Inoue A, Masumoto T. Mater Trans JIM 1991;32:1005. Peker A, Johnson WL. Appl Phys Lett 1993;63:2342. Pampillo CA. J Mater Sci 1975;10:1194. Gilman JJ. J Appl Phys 1975;46:1625. Inoue A. Acta Mater 2000;48(1):279. Spaepen F. Acta Metall 1975;23:615. Spaepen F. Acta Metall 1977;25:407. Argon AS. Acta Metall 1979;27:47. Wright WJ, Saha R, Nix WD. Mater Trans JIM 2001;42:642. Eckert J, Ku¨hn U, Mattern N, He G, Gebert A. Intermetallics 2002;10: 1183. Ku¨hn U, Eckert J, Mattern N, Schultz L. Appl Phys Lett 2002;80: 2478. He G, Eckert J, Lo¨ser W, Schultz L. Nat Mater 2003;2:33. Zhang ZF, Eckert J, Schultz L. Acta Mater 2003;51:1167. Zhang ZF, He G, Eckert J, Schultz L. Phys Rev Lett 2003;91:045505. Liu CT, Heatherly L, Easton DS, Carmichael CA, Schneibel JH, Chen CH, et al. Metall Mater Trans A 1998;29A:1811. Mukai T, Nieh TG, Kawamura Y, Inoue A, Higashi K. Intermetallics 2002;(10):1071. Inoue A, Zhang T, Masumoto T. Mater Trans JIM 1995;36:391. Lowhaphandu P, Ludrosky LA, Mongomery SL, Lewandowski JJ. Intermetallics 2000;8:487. He G, Lu J, Bian Z, Chen D, Chen G, Tu G, et al. Mater Trans 2001;42: 356. Conner RD, Johnson WL, Paton NE, Nix WD. J Appl Phys 2003;94:904. Hufnagel TC, El-Deiry P, Vinci RP. Scripta Mater 2000;43:1071. Yokoyama Y, Yamano K, Fukaura K, Sunada H, Inoue A. Mater Trans 2001;42:623. Gilbert CJ, Ritchie R, Johnson WL. Appl Phys Lett 1997;71:476. Gilbert CJ, Schroeder V, Ritchie RO. Metall Mater Trans A 1999;30A: 1739. Zhang ZF, Eckert J, Schultz L. J Mater Res 2003;18:456. Zhang ZF, Eckert J, Schultz L. Metall Mater Trans A 2004;35A:3489. Antoniou A, Bastawros AF, Lo CCH, Biner SB. Mater Sci Eng A 2005; 394:96. Gebert A, Eckert J, Schultz L. Acta Mater 1998;46:5475. Mattern N, Ku¨hn U, Sakowski J, Neuefeind J, Eckert J. Mater Trans 2002; 43:1947. Donovan PE. Acta Metall 1989;37:445. Zhang ZF, Eckert J. Phys Rev Lett 2005;94:094301. Subhash G, Dowding JR, Keczkes LJ. Mater Sci Eng 2002;A334:33. Bruck HA, Rosakis AJ, Johnson WL. J Mater Res 1996;11:503. Xi XK, Zhao DQ, Pan MX, Wang WH, Wu Y, Lewandowski JJ. Phys Rev Lett 2005;94:125510. Chou PC-P, Spaepen F. Acta Metall 1975;23:609. Dieter EG. Mechanical metallurgy. New York: McGraw-Hill; 1986. Conner RD, Choi-Yim H, Johnson WL. J Mater Res 1999;14:3292. Leamy HJ, Chen HS, Wang TT. Metall Trans 1972;3:699. Argon AS, Salama M. Mater Sci Eng 1976;23:219. Li J, Spaepen F, Hufnagel TC. Philos Mag A 2002;82:2623. Hajlaoui K, Benameur T, Vaughan G, Yavari AR. Scripta Mater 2004;51: 843. Yavari AR, Le Moulec A, Inoue A, Nishiyama N, Lupu N, Matsubara E, et al. Acta Mater 2005;53:1611. Wright WJ, Hufnagel TC, Nix WD. J Appl Phys 2003;93:1432. Kim KB, Das J, Baier F, Eckert J. Appl Phys Lett 2005;86:201909.