Journal of Non-Crystalline Solids 317 (2003) 158–163 www.elsevier.com/locate/jnoncrysol
Section 4. Properties
Structure and properties of Zr–Ta–Cu–Ni–Al bulk metallic glasses and metallic glass matrix composites R.T. Ott *, C. Fan, J. Li, T.C. Hufnagel Department of Materials Science and Engineering, Johns Hopkins University, 3400 North Charles Street, 102 Maryland Hall, Baltimore, MD 21218, USA
Abstract We have developed a series of alloys which, upon cooling from the melt, form two-phase microstructures consisting of crystalline Ta-rich solid solution particles embedded in a bulk metallic glass matrix. These alloys have the general composition (Zr70 Ni10 Cu20 )90x Tax Al10 where 0 6 x 6 12 (all compositions are in atomic percent). These alloys have mechanical properties consistent with metallic glasses, including high yield strength (2 GPa) and large elastic elongation (2%). However, they also show much larger plastic strain to failure (up to 16%) in uniaxial compression than monolithic metallic glasses. Ó 2003 Elsevier Science B.V. All rights reserved. PACS: 61.43.Dq; 81.05.Kf; 62.20.Fe; 81.20
1. Introduction A common limitation of metallic glasses is their tendency to experience plastic (permanent) deformation in narrow regions called shear bands. This causes metallic glasses to fail in an apparently brittle manner in any loading condition (such as tension) where the shear bands are unconstrained [1–3]. As a result, monolithic metallic glasses typically display limited plastic flow (0–2% under uniaxial compression) at room temperature. Recently, the bulk metallic glass-forming alloy Zr59 Ta5 Cu18 Ni8 Al10 has been shown to display an
*
Corresponding author. Tel.: +1-410 516 0816; fax: +1-410 516 5293. E-mail address:
[email protected] (R.T. Ott).
enhanced plastic strain of 4.5% under quasistatic uniaxial compression [4]. The increase in plastic strain to failure in this alloy may be due to an increase in medium-range order (structure on the 1–2 nm scale). Other researchers have found that plasticity in bulk metallic glasses can be increased by forming a two phase microstructure consisting of a metallic glass matrix reinforced with a crystalline second phase [5–7]. A particularly promising type of these composite structures involves precipitation of a dendritic ductile intermetallic phase in a metallic glass matrix [8]. Here we describe the structure and properties of a bulk metallic glass matrix composite prepared using a relatively simple in situ processing method. The composite microstructure consists of a bulk metallic glass matrix surrounding homogeneously
0022-3093/03/$ - see front matter Ó 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0022-3093(02)01996-8
R.T. Ott et al. / Journal of Non-Crystalline Solids 317 (2003) 158–163
dispersed, micron-scale crystalline particles. The composite retains the characteristic properties of a metallic glass, including a glass transition temperature, high yield strength (2 GPa), and large elastic elongation (2%). Like monolithic metallic glasses, at room temperature the composite deforms by localized plastic deformation in shear bands. However, the second-phase particles, along with increased order in the local structure of the amorphous matrix, lead to significantly enhanced plastic strain to failure in uniaxial compression.
2. Experimental procedure The starting materials used in preparing the alloys were high purity metals: Cu (99.999%), Al (99.999%), Ta (99.995%), Ni (99.995%), and a Zr crystal bar with <300 ppm oxygen. We prepared alloys of composition (Zr70 Ni10 Cu20 )90x Tax Al10 with x ¼ 0; 2; 4; 5; 6; 8; 10; 12 (x ¼ at.% Ta) by arc melting in a Ti-gettered argon atmosphere on a water cooled copper hearth. Master alloy ingots were prepared by a two-step process: The elemental zirconium and tantalum were melted together to produce a homogeneous ingot; then the nickel, aluminum, and copper were melted together with the zirconium–tantalum alloy ingot. For each step, the ingot was melted and flipped several times to promote homogeneity. The final ingot was then cast into a copper mold to produce rods 3 mm in diameter and 5 cm in length. The phases present in the as-cast samples were examined with X-ray diffraction (XRD) using a Rigaku TTRAXS h=h rotating anode diffractometer with Cu Ka radiation (k ¼ 0:154 nm). The microstructure was examined using a Philips CM300 field emission transmission electron microscope (TEM) operated at 300 kV, a JEOL 6700F scanning electron microscope and a JEOL 8600 microprobe. The thermal properties of the composite samples were measured in a Perkin– Elmer Pyris 1 differential scanning calorimeter (DSC). Quasistatic compression tests were performed at room temperature using a MTS servohydraulic machine operated under displacement control at a strain rate of 104 s1 . Samples for quasistatic compression were produced by center-
159
less grinding the rods to 2.8 mm diameter; samples were then cut from the rods and the ends polished to ensure parallelism.
3. Results The XRD patterns for the as-cast (Zr70 Ni10 Cu20 )90x Tax Al10 (x ¼ 6 and 12 at.%) alloys are shown in Fig. 1. The diffraction patterns show a broad scattering feature at 37° 2h along with sharp Bragg peaks corresponding to a crystalline phase. The broad scattering feature is characteristic of an amorphous phase, which is the matrix of the composite. The lattice parameter of the crystalline which corresponds closely to phase is a0 ¼ 3:310 A ). No other crystalline that of bcc Ta (a0 ¼ 3:306 A peaks can be seen in the diffraction patterns. The intensity of the Ta peaks increases with increasing Ta concentration, indicating that the volume fraction of the crystalline phase increases with increasing Ta concentration. The as-cast microstructure for a 10% Ta alloy is shown in Fig. 2. The microstructure consists of homogeneously dispersed particles (light phase) in an amorphous matrix (dark phase). The particles are oblong in shape, and do not appear to possess a dendritic structure. The average size of the particles is approximately 20–50 lm. The average chemical
Fig. 1. XRD pattern showing the phases present in a 6% Ta alloy and a 12% Ta alloy.
160
R.T. Ott et al. / Journal of Non-Crystalline Solids 317 (2003) 158–163
Fig. 2. Back scattered electron micrograph of as-cast sample of Zr56 Ni8 Cu16 Ta10 Al10 showing micron-scale particles dispersed in amorphous matrix.
composition of the crystalline particles (determined by electron microprobe analysis) is Ta93 Zr5 (Cu+Ni+Al)2 (all compositions are in atomic percent and are accurate to 1 at.%) and is independent of the overall composition of the alloy. We examined the structure of the matrix phase with high resolution transmission electron microscopy (HRTEM). Fig. 3 shows a HRTEM micrograph and electron diffraction pattern for the 5% Ta alloy. The absence of lattice fringes in the image yields no evidence of any crystalline phase. Therefore, we conclude that there are no crystallites larger than about 2 nm (the minimum observable crystallite size in HRTEM) in the matrix. The volume fraction of the crystalline phase was determined by quantitative analysis of the SEM images, and is shown in Fig. 4. The volume percent of the Ta-rich particles scales linearly with the Ta concentration of the alloy above x ¼ 4%. The chemical composition of the matrix is also approximately 4% Ta. Together, these results indicate that the solubility limit of Ta in the amorphous phases is approximately 4%. The presence of Ta also influences the crystallization behavior of the amorphous alloys. Constant rate DSC scans 0% Ta and 6% Ta alloys are shown in Fig. 5. The 0% Ta alloy shows one large exotherm at around 760 K and a smaller, broad exotherm at higher temperature. In contrast, the
Fig. 3. HRTEM image and electron diffraction pattern (inset) of matrix for a 5% Ta alloy. The amorphous structure of the matrix is confirmed by the absence of lattice fringes.
Fig. 4. Volume percent of Ta-rich particles and Ta concentration in the amorphous matrix as a function of overall Ta content in the alloy.
6% Ta alloy has four exotherms in its DSC trace. Fig. 6 shows diffraction patterns for 0% and 6% Ta containing alloys which have been annealed through their first exotherms. The 0% Ta alloy crystallizes to form the tetragonal forms of CuZr2 , Al2 Zr3 , and NiZr2 , which is consistent with pre-
R.T. Ott et al. / Journal of Non-Crystalline Solids 317 (2003) 158–163
161
Fig. 5. Constant rate (10 K/s) DSC traces from a 0% Ta alloy and 6% Ta alloy. Note the distinct glass transition temperature, confirming that the matrix is a metallic glass.
Fig. 6. XRD patterns for a 0% Ta and 6% Ta alloy annealed isothermally through their first exotherms. The peaks corresponding to the quasicrystalline phase are indexed according to BancelÕs [12] method.
vious results on alloys of similar composition [9]. The diffraction pattern for the 6% Ta alloy, however, shows the presence of an icosohedral quasicrystalline phase (along with a residual amorphous phase) upon annealing through its first exotherm. The presence of the quasicrystalline phase was confirmed with TEM. Fig. 7(a) shows a brightfield image of the 6% Ta alloy that has been annealed through the first exotherm. Several quasicrystals ranging from 20 to 150 nm can be seen.
Fig. 7(b) shows a selected-area electron diffraction pattern for one of the quasicrystals. The diffraction pattern displays the characteristic fivefold symmetry of an icosohedral quasicrystalline phase. The overall crystallization sequence of the 6% Ta alloy is shown in Fig. 8. The amorphous matrix forms quasicrystals upon annealing through the first exotherm. When the sample is annealed for longer times, the tetragonal forms of NiZr2 and Al2 Zr3 are nucleated, while the intensity of the
Fig. 7. (a) TEM bright-field image showing quasicrystals along with residual amorphous phase. (b) SAED pattern with beam parallel to fivefold symmetry axis.
162
R.T. Ott et al. / Journal of Non-Crystalline Solids 317 (2003) 158–163
strength and elastic limit, but less than 1% plastic strain prior to failure and no work hardening. The 8% Ta alloy shows very similar elastic behavior and yield strength, but exhibits 16.6% apparent plastic strain before failure. (The definition of strain in this case is based on platen displacement. The deformation at large strains becomes non-uniform, and buckling occurs.) Several samples of this alloy were tested in compression and yielded a mean apparent plastic strain to failure of 16%. The 8% Ta alloy also shows work hardening during plastic deformation with a yield strength of 1750 MPa and an ultimate strength of approximately 2100 MPa.
Fig. 8. XRD patterns for a 6% Ta alloy showing the phases present after isothermal annealing.
peaks corresponding to the icosohedral phase is significantly reduced, which indicates that the quasicrystalline phase is metastable. When the sample is annealed through the last exotherm, CuZr2 is formed. This is in contrast to the 0% Ta alloy in which CuZr2 forms at a much lower temperature. Fig. 9 shows the stress–strain curves measured under quasistatic uniaxial compression for the 0% and 8% Ta alloys. The 0% Ta alloy shows behavior typical of a bulk metallic glass––large yield
Fig. 9. Stress–strain curve for a 0% Ta and an 8% Ta alloy. Note the high flow strength, apparent work hardening, and the extended region of plastic deformation for the 8% Ta alloy.
4. Discussion The microstructure of the composites appears to be a result of the two-step processing technique outlined above. When Zr and Ta are melted together to form a homogeneous ingot, the resulting microstructure is a supersaturated solid solution of Ta in Zr, as deduced from XRD patterns (not shown). When this ingot is melted together with the other elements, some of the Ta apparently precipitates out to form 10 lm particles, while the rest of the Ta remains in solution. The Ta-rich particles have a very high melting point, so it is likely that they remain solid when the ingot is melted and cast to form the composite. During casting, these small particles apparently agglomerate to form the larger particles apparent in the micrographs. The electron microprobe results seem to indicate that there is a solubility limit of approximately 4% Ta in the amorphous alloy. The presence of Ta in the amorphous phase has a significant effect on its crystallization by promoting the formation of quasicrystals. This is consistent with previous studies, which show that in Zr–Cu– Ni–Al alloys, additions of early transition metals promotes the formation of a quasicrystalline phase [10]. The formation mechanism of the icosohedral quasicrystalline phase has been related to structural order in the amorphous alloy [11]. It is possible that the Ta additions lead to an increase in medium-range order in the amorphous matrix similar to what Xing and coworkers reported in a
R.T. Ott et al. / Journal of Non-Crystalline Solids 317 (2003) 158–163
monolithic Zr59 Ta5 Cu18 Ni8 Al10 alloy which also showed enhanced plastic strain. The mechanical properties of the composite are a result of its unique microstructure. In addition to the presence of the Ta particles, the matrix of the composite is very similar in chemical composition to the Ôenhanced strainÕ monolithic glass reported by Xing and coworkers [4]. They speculated that the structure of the glass promotes branching of propagating shear bands, which redistributes plastic strain among several bands and thereby suppresses crack initiation. In the present case, this matrix material is combined with a dispersion of secondphase particles. The particles may act as nucleation sites for shear bands, or inhibit shear band propagation. Either of these would tend to increase the macroscopic plastic strain to failure. The strain hardening behavior of the composite is most likely due to strain hardening of the Ta particles.
5. Conclusion In situ composites consisting of homogeneously dispersed Ta-rich particles in an amorphous matrix can be made for several different alloy compositions. The volume fraction of the particles in the composite microstructure is proportional to the Ta content of the alloy, above the solubility limit of 4% Ta in the matrix. The composites show large apparent plastic strains in uniaxial compression, along with large elastic elongation and high yield strength.
163
Acknowledgements We gratefully acknowledge Mingwei Chen for assistance with the transmission electron microscopy measurements, Ken Livi for assistance with the electron microprobe analysis measurements and Haito Zhang for assistance with the mechanical testing. This work was supported by the U.S. Department of Energy under grant DE-F00298ER45699 and the Army Research Laboratory under grant DAAD19-01-2-003.
References [1] C.A. Pampillo, J. Mater. Sci. 10 (1975) 1194. [2] L.Q. Xing, C. Bertrand, J.P. Dallas, M. Cornet, Mater. Sci. Eng. A 241 (1998) 216. [3] A. Inoue, Acta Mater. 48 (2000) 279. [4] L.Q. Xing, Y. Li, K.T. Ramesh, J. Li, T.C. Hufnagel, Phys. Rev. B 64 (2001) R180201. [5] H. Kato, A. Inoue, Mater. Trans., JIM 38 (1997) 793. [6] H. Choi-Yim, W.L. Johnson, Appl. Phys. Lett. 71 (1997) 3808. [7] H. Choi-Yim, U. Koster, R. Busch, W.L. Johnson, Acta Mater. 47 (1999) 2455. [8] C.C. Hays, C.P. Kim, W.L. Johnson, Phys. Rev. Lett. 84 (2000) 2901. [9] N. Mattern, J. Eckert, M. Seidel, U. K€ uhn, S. Doyle, I. B€acher, Mater. Sci. Eng. A 226–228 (1997) 468. [10] J. Saida, A. Inoue, J. Phys.: Condens. Matter 13 (2001) L73. [11] L.Q. Xing, T.C. Hufnagel, J. Eckert, W. L€ oser, L. Schultz, Appl. Phys. Lett. 77 (2000) 1970. [12] P.A. Bancel, P.A. Heiney, P.W. Stephens, A.I. Goldman, P.M. Horn, Phys. Rev. Lett. 54 (1985) 2422.