Calhoun: The NPS Institutional Archive

Faculty and Researcher Publications Faculty and Researcher Publications Collection

2005-03-16

Redistribution of alloying elements in

quasicrystallized Zr65Al7.5Ni10Cu7.5Ag10

bulk metallic glass

Chen, M.W.

Phys. Rev. B 71, 092202 (2005)

http://hdl.handle.net/10945/51348

Redistribution of alloying elements in quasicrystallized Zr65Al7.5Ni10Cu7.5Ag10 bulk metallic glass

M. W. Chen,1,2,* A. Inoue,1 T. Sakurai,1 E. S. K. Menon,3 R. Nagarajan,4 and I. Dutta41Institute for Materials Research, Tohoku University, Sendai 980-8577, Japan

2Department of Mechanical Engineering, Johns Hopkins University, Baltimore, Maryland 21218, USA3Materials and Processes Division, UES, Inc., Dayton, Ohio 45432, USA

4Department of Mechanical Engineering, Naval Postgraduate School, Monterey, California 93943, USAsReceived 16 July 2004; revised manuscript received 18 October 2004; published 16 March 2005d

We report experimental evidence on redistributions of alloying elements in a partially quasicrystallizedZr65Al7.5Ni10Cu7.5Ag10 bulk metallic glass. Energy-filtered elemental maps show that the partitioning of silver,zirconium, and aluminum occurs in the quasicrystallized glass. Silver is found to enrich into quasicrystals, butzirconium and aluminum are slightly depleted from the quasicrystals. These observations suggest that theamorphous-to-quasicrystal transformation in this alloy is not a simple polymorphous reaction and a long-rangediffusion is involved in the devitrification, probably prior to the quasicrystallization.

DOI: 10.1103/PhysRevB.71.092202 PACS numberssd: 64.70.Pf, 81.30.2t, 68.37.Lp

Icosahedron has been considered as a structure unit indisordered metallic glasses due to its lack of translationalperiodicity and difficulty for growth compared to its crystalcounterparts. Icosahedral quasicrystals have recently beenobserved in a number of annealed Zr-based bulk metallicglasses.1–7 The formation of icosahedral phases, while theglasses are annealed within the supercooling liquid regions,supports the structural similarity between metallic glassesand icosahedrons. Meanwhile, it also suggests that the glass-forming ability and thermal stability of bulk metallic glassesare not solely controlled by the nucleation of crystallinephases as considered in traditional models. The formation ofquasicrystalline phases may have to be counted. Therefore,the understanding of the amorphous-to-quasicrystal transfor-mation mechanism turns out to be essential for exploring thebulk metallic glass phenomenon.8,9

In previous papers, the majority of the icosahedral phasesformed from annealed metallic glasses with a binary or ter-nary composition were determined to be products of a simplepolymorphous transformation during which no chemicalcomposition redistributions occur between glassy parentphases and icosahedral quasicrystals.10–13 Thermodynamicand kinetic analyses suggest that the transformations are ac-complished by a nucleation and growth process. The struc-tural similarity between metallic glasses and icosahedral qua-sicrystals yields much lower glass-quasicrystal interfaceenergies13 and results in that the transformations are gov-erned by nucleation processes. Regarding multicomponentZr-based bulk metallic glasses, Zr-Al-Ni-Cu-Ag alloys areone of the most investigated systems because relativelystable icosahedral phases can be formed in the alloys duringannealing in their supercooling liquid regions.3,14,15 How-ever, the amorphous-to-quasicrystal transformation mecha-nism in the multicomponent alloys is not fully understood.The pressure dependence of the quasicrystallization tempera-tures has been observed in a Zr65Al7.5Ni10Cu7.5Ag10 bulk me-tallic glass during high-pressure annealing, indicating that along-range diffusion may be involved into the trans-formation.16 Kinetic analysis of the alloy displayed that thetransformation rate of the icosahedral phase degrades withannealing. This appears to be induced by the diffusion field

overlapping caused by solute partitioning.17 These observa-tions indicate that the transformations in the Zr-Al-Ni-Cu-Ag bulk metallic glasses may not be a simple poly-morphous reaction, and chemical partitioning may take placeduring the devitrification. Therefore, direct chemical compo-sition measurements are desired to classify the transforma-tion mechanism.

In this Brief Report, we outline a systematic energy-filtered transmission electron microscopesEF-TEMdstudy of composition redistributions in the annealedZr65Al7.5Ni10Cu7.5Ag10 bulk metallic glass composed of qua-sicrystals embedding in a glassy matrix.

The Zr65Al7.5Ni10Cu7.5Ag10 sat. %d alloy was prepared bymelting high-purity constituent elements under an argon at-mosphere in an arc-melting furnace. Glassy ribbons with across section of about 0.0533.0 mm were prepared by asingle roller melt-spinning apparatus with a wheel velocityof 4000 rpm. The endothermic and exothermic reactions as-sociated with the glass transition temperature, and the crys-tallization starting temperature of the ribbons were measuredin a continuously heating mode by a differential scanningcalorimetersDSCd at a heating rate of 40 K/min. Two exo-thermic peaks appear in the DSC trace, indicating that thedevitrification of the metallic glass takes place in two steps.The glassy ribbons encapsulated in quartz tubes with a pres-sure of about 10−5 Torr were isothermally annealed at700 K, within the supercooled liquid region, for 4 min, andthen were quenched into water with the quartz tubes. Themorphologies of the annealed samples were investigated by aPhilips CM-300 FEG transmission electron microscopesTEMd, and the crystal structures of precipitated nanophasesin the annealed samples were investigated by nanobeamelectron diffraction with a nominal converged beam size of,5 nm. The distributions of the alloying elements in thesamples were investigated by a Gatan imaging filter system.TEM samples were prepared by ion milling with a liquidnitrogen cooling stage.

High-resolution electron microscopesHREMd observa-tions of the as-prepared samples demonstrated that the alloyis fully disordered and crystallites have not been seen. Figure

PHYSICAL REVIEW B 71, 092202s2005d

1098-0121/2005/71s9d/092202s4d/$23.00 ©2005 The American Physical Society092202-1

1 shows an example in which only a number of maze struc-tures, which may correspond to medium- or short-range clus-ters, can be observed. A fast Fourier transformation of theglassy phase only shows halo patterns, further confirmingthat the as-prepared alloy is fully amorphous. Additionally,EF-TEM characterizations did not find any noticeable chemi-cal variations in the glassy alloy, and the distribution of thecomponent elements in the glassy alloy appears to be uni-form.

The microstructure of the alloy isothermally annealed at700 K for 4 min is shown in Fig. 2sad. The precipitatedphase with a size around 20 nm can be observed, and theparticle density measured by the TEM observations is ap-proximately 931020 m−3 fFig. 1sadg. The separated micro-beam diffraction pattern taken from the quasicrystallinephase is shown in Fig. 2sbd, corresponding to the fivefoldsymmetry of the simple icosahedral point group. Weak dif-fraction halos are also visible in the microbeam diffractionpatterns, indicating that the matrix phase is the residualamorphous.

To explore the compositional difference between the qua-sicrystalline phase and the residual amorphous, two-dimensional distributions of the constituent elements in thenanostructured alloy were obtained by energy-filtered jump-ratio imaging, which is largely insensitive to variations inspecimen thickness and diffraction conditions.18 A small re-gion containing an icosahedral particle, highlighted in thezero-loss imagefFig. 2sbdg, was selected for EF-TEM analy-sis. From the elemental maps and separate electron energy-loss spectrasEELSd, it was determined that the nanoparticlecontains all constituent elements present in the alloy. How-ever, the particle is visible only in the jump-ratio imagesrecorded at the EELS edges of Zr-M45, Al-L23, and Ag-M45fFigs. 3sad–3scdg. Zr and Al exhibit dark contrast within thenanoparticle compared with the glassy matrix, indicating that

Zr and Al are depleted from the quasicrystalline phase. Incontrast, Ag shows enhanced contrast within the nanopar-ticle, suggesting that Ag enriches in the nanophase. In thejump-ratio images of Cu-L23 and Ni-L23, no chemical con-trast can be clearly identifiedfFigs. 3sed and 3sfdg, suggestingthat the distribution of Cu and Ni is uniform in the quasi-crystallized alloy and long-range diffusions of Cu and Nimay not be involved into the devitrification. Impurity oxygenhas been considered to be important in the formation oficosahedral phases and the devitrification of Zr-based bulkmetallic glasses.5,19,20A separate electron energy-loss spec-trum shows the existence of oxygen in the quasicrystallinephase and the residual amorphous matrix. However, the dis-tribution of oxygen appears to be uniform from the energy-filtered oxygen mapfFig. 3sddg. This suggests that oxygenpartitioning is not involved in the formation of the quasic-rystalline phase and impurity oxygen does not play any spe-

FIG. 1. Energy-filtered HREM image of an as-preparedZr65Al7.5Ni10Cu7.5Ag10 bulk metallic glass.

FIG. 2. sad Bright field images of the Zr65Al7.5Ni10Cu7.5Ag10

metallic glass annealed at 700 K for 4 min, andsbd a nanoparticleand corresponding nanobeam electron diffraction pattern. This par-ticle was selected for EF-TEM characterization in Fig. 3.

BRIEF REPORTS PHYSICAL REVIEW B71, 092202s2005d

092202-2

cial role in the formation of the quasicrystalline phase in thisalloy.

Devitrification mechanisms of metallic glasses are gener-ally divided into four categories: polymorphous, eutec-tic, primary, and devitrification with phase separ-ation.21 Unambiguously, the reaction observed in theZr65Al7.5Ni10Cu7.5Ag10 alloy is not eutectic because only onequasicrystalline phase formed during annealing. Polymor-phous transformations have been mainly observed in solidreactions of pure elements where no chemical compositionchanges are concerned during evolution of crystallographicstructures.22 The transformation is mainly performed byatom rearrangements at reactant-product interfaces by ashort-range diffusion. This sort of phase transformation hasalso been widely observed in the devitrification of metallicglasses because metastable crystalline phases with wide tol-erant composition ranges are easily formed from nonequilib-rium glasses when a long-range diffusion is restricted by lowtransformation temperatures. In most cases, the icosahedralquasicrystalline phases are metastable, but the compositionranges of quasicrystalline phases are usually very narrow be-

cause the icosahedral structures have strict requirements forconstituent atoms from size to chemical characteristics.Slight composition variations can induce the loss of the qua-sicrystalline phases, replaced by their crystalline counter-parts. Thus, the formation of quasicrystalline phases ismainly by polymorphous reactions. However, the current EF-TEM observations of the elemental partitioning in the par-tially quasicrystallized metallic glass suggest that the devit-rification in Zr65Al7.5Ni10Cu7.5Ag10 may not be a simplepolymorphous transformation and is most likely accom-plished by a reaction associated with a long-range diffusion.

The formation of a single phase accompanying the com-position redistribution is generally categorized as primarycrystallizationsdevitrificationd. A primary phase first precipi-tates out and results in a composition change in a residualglass that later crystallizes separately.21 However, the case inthe Zr65Al7.5Ni10Cu7.5Ag10 alloy appears to be different. Al-though the two-step devitrification can be observed in theDSC curve, the first exothermic reaction corresponds to theamorphous-to-quasicrystalline transformation and the secondone results from the transformation from quasicrystals—but

FIG. 3. sColord EF-TEM characterization of the elemental distributions in the quasicrystallized metallic glass. The succeeding imagesshow jump-ratio maps recorded with 20-eV windows on either side ofsad Zr-M45, sbd Ag-M45, scd Al-L23, sdd O-K, sed Cu-L23, and sfdNi-L23. The accompanying profiles highlight the composition difference between the nanoparticle and the glassy matrix.

BRIEF REPORTS PHYSICAL REVIEW B71, 092202s2005d

092202-3

not residual amorphous—to crystals.17,23–25 Definitely, thequasicrystallization in this alloy cannot be classified as theprimary transformation.

Based on the EF-TEM observations and the aforemen-tioned discussion, we propose that prior to the quasicrystal-lization, there is a modulated reaction in which the glassitself separates into silver-rich and silver-poor phases in thesupercooling liquid region. The existence of this kind ofmodulated reaction has been theoretically predicted in closeto the quasicrystalline composition range26 and experimen-tally observed in the quasicrystalline Al-Fe-Cu system.27

This reaction associated with an uphill diffusion of silver inthe Zr-Al-Ni-Cu-Ag system can be driven by the strongnegative heat of mixing between zirconium and silver.28

Silver-rich regions have lower thermal stability and quasi-crystallization temperatures15 and first transform into the

quasicrystalline phase by a polymorphous reaction. Thetransformation is extended by a continuous nucleation pro-cess from silver-rich to silver-poor regions with increasingquasicrystallization temperatures or annealing time. TheAvrime exponent of around 3.0, measured by kineticanalysis,29 suggests that new nuclei continuously appear dur-ing annealing, and the transformation is accomplished by thecontinuous nucleation process, exactly consistent with thisassumption. Additionally, it has been found that a highercontent of silver in the Zr-Cu-Ni-Al-Ag alloys significantlyreduces the icosahedral grain sizes and promotes the forma-tion of the quasicrystals.15 These observations further sup-port the proposed transformation mechanism because the pe-riodic wavelength and composition variation of modulatedstructures strongly rely on the concentrations of solutes.

*Author to whom correspondence should be addressed. Email ad-dress: mwchen@imr.tohoku.ac.jp

1U. Köster, J. Meinhardt, S. Roos, and H. Liebertz, Appl. Phys.Lett. 69, 179 s1996d.

2L. Q. Xing, J. Eckert, W. Loser, and L. Schultz, Appl. Phys. Lett.73, 2110s1998d.

3M. W. Chen, T. Zhang, A. Inoue, A. Sakai, and T. Sakurai, Appl.Phys. Lett.75, 1697s1999d.

4A. Inoue, Mater. Trans., JIM40, 1181s1999d.5B. S. Murty, D. H. Ping, K. Hono, and A. Inoue, Acta Mater.48,

3985 s2000d.6U. Kuhn, J. Eckert, N. Mattern, and L. Schultz, Appl. Phys. Lett.

77, 3176s2000d.7N. Wanderka, M.-P. Macht, M. Seidel, and S. Mechler, Appl.

Phys. Lett.77, 3935s2000d.8A. Inoue, T. Zhang, and T. Masumoto, Mater. Trans., JIM31, 177

s1990d.9A. Peker and W. L. Johnson, Appl. Phys. Lett.63, 2342s1993d.

10D. A. Lilienfeld, M. Nastasi, H. H. Johnson, D. G. Ast, and J. W.Mayer, Phys. Rev. Lett.55, 1587s1985d.

11S. J. Poon, A. J. Drehman, and K. R. Lawless, Phys. Rev. Lett.55, 2324s1985d.

12A. P. Tsai, A. Inoue, Y. Bizen, and T. Masumoto, Acta Metall.37,1443 s1989d.

13J. C. Holzer and K. F. Kelton, Acta Metall. Mater.39, 1833s1991d.

14M. W. Chen, A. Inoue, T. Zhang, A. Sakai, and T. Sakurai, Philos.Mag. Lett. 80, 263 s2000d.

15M. W. Chen, A. Inoue, T. Zhang, A. Sakai, and T. Sakurai,Intermetallics8, 493 s2000d.

16J. Z. Jiang, Y. X. Zhuang, H. Rasmussen, J. Saida, and A. Inoue,Phys. Rev. B64, 094208s2001d.

17J. K. Lee, G. Choi, D. H. Kim, and W. T. Kim, Appl. Phys. Lett.77, 978 s2000d.

18R. F. Egerton,Electron Energy-Loss Spectroscopy in the ElectronMicroscopesPlenum, New York, 1996d.

19J. Eckert, N. Mattern, M. Zinkevitch, and M. Seidel, Mater.Trans., JIM 39, 623 s1998d.

20M. W. Chen, A. Inoue, T. Sakurai, D. H. Ping, and K. Hono,Appl. Phys. Lett.74, 812 s1999d.

21R. W. Cahn, inMaterials Science and Technology, edited by R.W. Cahn, P. Haasen, and E. J. KramersVCH, New York, 1991d,Vol. 9, p. 530.

22J. W. Christian,Transformations in Metals and AlloyssPergamon,Oxford, 1975d, p. 105.

23J. Saida, M. Matsushita, T. Zhang, A. Inoue, M. W. Chen, and T.Sakurai, Appl. Phys. Lett.75, 3497s1999d.

24L. Liu and K. C. Chan, J. Alloys Compd.364, 146 s2004d.25M. W. Chen, I. Dutta, T. Zhang, A. Inoue, and T. Sakuri, Appl.

Phys. Lett.79, 42 s2001d.26A. Janner and T. Janssen, Phys. Rev. B15, 643 s1977d.27M. Audier, Y. Brechet, M. De. Boissieu, and P. Guyot, Philos.

Mag. B 63, 1375s1991d.28K. F. Kelton, J. Non-Cryst. Solids334/335, 253 s2004d.29M. Matsushita, J. Saida, T. Zhang, A. Inoue, M. W. Chen, and T.

Sakurai, Philos. Mag. Lett.80, 78 s2000d.

BRIEF REPORTS PHYSICAL REVIEW B71, 092202s2005d

092202-4