Available online at www.sciencedirect.com
Scripta Materialia 63 (2010) 883–886 www.elsevier.com/locate/scriptamat
“Soft” atoms in Zr70Pd30 metal–metal amorphous alloy Liang Yang,a,* Gu-Qing Guo,a Jian-Zhong Jiang,b Lian-Yi Chenb and Shi-Hao Weic,d,* a
College of Materials Science and Technology, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, People’s Republic of China b International Center for New-Structured Materials (ICNSM) and Laboratory of New-Structured Materials, Department of Materials Science and Engineering, Zhejiang University, Hangzhou 310027, People’s Republic of China c Department of Physics, Ningbo University, Ningbo, Zhejiang 315211, People’s Republic of China d Department of Physics, Fudan University, Shanghai 200433, People’s Republic of China Received 8 June 2010; revised 28 June 2010; accepted 29 June 2010 Available online 7 July 2010
The atomic structure of Zr70Ni30 xPdx (x = 0–30) amorphous alloys (AAs) was systematically investigated via synchrotron radiation-based experiments and simulations. A relatively short Zr–Pd bond was detected in Zr70Pd30, which is attributed to the strong interaction between Zr and Pd “soft” atoms. This phenomenon was further explained by performing electronic interaction calculations upon some Voronoi clusters in the ZrPd sample and clusters extracted from the corresponding crystal phases. We suggest that soft atoms may have a signiﬁcant impact on the internal structure in some metal–metal AAs. Ó 2010 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. Keywords: Metallic glasses; Extended X-ray absorption ﬁne structure; X-ray diﬀraction; Reverse Monte Carlo simulation
The atomic structures of amorphous alloys (AAs) have been intensively investigated because their glassforming mechanisms and extraordinary properties are closely correlated with their atomic structures [1–8]. Long ago, Bernal suggested a random dense packing model using hard spheres to describe the microstructure of monatomic liquids [9,10], and this was later extended to model the atomic structures of AAs. However, this model ignored the chemical short-range order (CSRO). Recently, an eﬃcient dense packing principle was adopted to construct the atomic structures of metal–metal-type AAs , in which CSRO was suggested. In addition, in Miracle’s model [12,13], optimal clusters with solute-center and solvent-shell atoms were deduced based on the atomic radii that were the best ﬁt observed in a large number of condensed solids. As an important CSRO factor, covalent bonding between metal and metalloid atoms has also been recognized as aﬀecting atomic structure in many metal– metalloid AAs. Recently, relatively short atomic bonds (compared with the sum of Goldschmidt radii for atomic pairs or those in corresponding crystalline compounds) have been suggested in some metal–metal AAs [14–18], which indicates that atoms may be regarded as “soft” balls rather than “hard” spheres and impact on the * Corresponding authors. Tel.: +86 25 52112903; e-mail addresses: [email protected]
; [email protected]
internal structure in these glasses. The understanding of “soft” atoms and their bonds will shed light on the internal atomic structure in some metal–metal AAs. In this work, the atomic structures of Zr70Ni30–xPdx (x = 0, 5, 10, 20 and 30 at.%) AAs were systematically studied by using state-of-the-art synchrotron radiation techniques as well as a series of simulation and calculation methods. Explicit evidence and an explanation of shortened bonds between soft atoms in Zr70Pd30 (an early transition metal– late transition metal) AA are provided. Zr70Ni30 xPdx binary and ternary ingots were prepared by arc melting high-purity metals (99.9% Zr, 99.9% Ni and 99.9% Pd). Amorphous ribbons with a cross-section of 0.04 2 mm2 were produced from these ingots via single-roller melt spinning at a wheel surface velocity of 40 m s 1 in a puriﬁed Ar atmosphere. First, X-ray diﬀraction (XRD, using Cu Ka radiation) and high-resolution electron microscopy measurements conﬁrmed the amorphous state of the as-prepared samples. Subsequently, other room-temperature XRD measurements were performed on all samples with a high-energy synchrotron radiation monochromatic beam (about 100 keV) on beamline BW5 in HASYLAB, Germany. Twodimension diﬀraction data were collected by a Mar345 image plate, which was integrated to Q-space after subtracting the corresponding background using the Fit2D program . The output data were normalized using the PDFgetX software to obtain the structure factor
1359-6462/$ - see front matter Ó 2010 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.scriptamat.2010.06.044
L. Yang et al. / Scripta Materialia 63 (2010) 883–886
Figure 1. (a) S(Q), (b and c) EXAFS data obtained from experiment (black solid line) and RMC simulation (red dotted line) in Zr70Pd30 and Zr70Ni30 AAs, respectively. (d) The normalized Zr K-edge X-ray absorption near edge spectra of Zr70Ni30 xPdx (x = 0, 5, 10, 20 and 30) AAs and Zr foil. (For interpretation of the references to color in this ﬁgure legend, the reader is referred to the web version of this paper.) Table 1. Comparison of the ﬁrst-shell Zr–Zr and Zr–M (M = Pd or Ni) distances obtained from RMC simulation, SGAR and corresponding crystalline compounds. Zr70Pd30
(1) From RMC
Atomic pairs ˚) Zr–Zr (A ˚) Zr–M (A
(2) From SGAR
˚) Zr–Zr (A ˚) Zr–M (A
˚) Zr–Zr (A ˚) Zr–M (A
Zr2Pd1 3.23 2.89
Zr2Ni1 3.24 2.76
˚) Zr–Zr (A ˚) Zr–M (A ˚) Zr–Zr (A ˚) Zr–M (A
0.06 0.18 0.09 0.10
0.05 0.06 0.09 0.03
(3) From crystalline compounds (2)–(1) (3)–(1)
The bond lengths in the crystalline compounds are averaged values.
S(Q) according to the Faber–Ziman equation . Furthermore, extended X-ray absorption ﬁne structure (EXAFS) measurements for Zr, Ni and Pd K-edge were carried out using the transmission mode at beamline A1 in HASYLAB. The measured spectra were analyzed via the standard procedures of data reduction using the program Winxas . Moreover, the reverse Monte Carlo (RMC) simulation technique [22,23] was utilized to simulate their S(Q) and EXAFS data synchronously. Cubic boxes containing 20,000 random packed atoms were built that matched the compositions of the Zr70Ni30 xPdx AAs. The simulated atomic structural models were further carefully analyzed by the Voronoi tessellation method . As shown in Figure 1a–c, the simulated S(Q) and Zr, Pd and Ni K-edge EXAFS spectra match the experimental data fairly well in Zr70Ni30 and Zr70Pd30 AAs (the spectra of the ZrNiPd AAs are not shown here). Their atomic structural information could be deduced based on the theoretical equations in previous work . The structural results of ZrPd and ZrNi binary AAs are listed in Table 1, including the ﬁrst-shell coordination numbers (CNs) and the atomic pair distances (the bond lengths). Because of the complexity of six types of bonds in ZrNiPd AAs, their values are not listed in the table. The bond
lengths are compared with the sum of the Goldschmidt atomic radii (SGAR) and those extracted from the corresponding crystalline compounds (tetragonal phases of Zr2Ni1 and Zr2Pd1, obtained by annealing treatments on Zr70Ni30 and Zr70Pd30 AAs, respectively ). In Zr70Ni30 AA, it is shown that the Zr–Ni bond length does not change much compared with the SGAR or crystal values. However, the Zr–Pd bond in Zr70Pd30 AA diﬀers greatly from the values of both the SGAR and the tetrag˚ . This indional phase, which are shortened by 0.1–0.2 A cates that a strong interaction occurs between the Zr and Pd atoms in the ZrPd binary glassy alloy, i.e. these atoms can be regarded as “soft” atoms. To understand more of this bond shortening, we need to analyze the normalized Zr K-edge X-ray absorption near edge spectra of Zr70Ni30 xPdx AAs and high-puriﬁed Zr foil, which are plotted in Figure 1d. It is observed that energy position of the absorption edge of Zr70Pd30 is about 1 eV higher than that of Zr foil. Since all the spectra were measured under the same experimental conditions and reduced via the same procedure, any error of the instruments or the normalization should be the same. Therefore, this increase in edge position should reﬂect the intrinsic diﬀerence of bond types between Zr70Pd30 and pure Zr. In particular, this phenomenon indicates that strong Zr–Pd interaction exists in the former, which may contribute to the shortened ZrPd bond. The detailed analysis is as follows: it is acknowledged that the state of valence electrons relates closely to the energy position of the K-edge for the same metal absorber . Since Pauling’s electronegativities of Zr and Pd elements are 1.33 [28–30] and 2.20 [29,30], respectively, the valence-shell electron-pair may be closer to Pd when Zr and Pd atoms are bonded tightly together. Therefore, inside each Zr atom, the inner-shell electrons experience a slightly smaller repulsive force from the valence electron and a larger attractive force from the nucleus to keep its orbital, hence more ionization energy is needed to excite them to ﬂee from the host atom, leading to the higher energy position of the absorption edge in Zr spectrum. However, such speciﬁc bonds do not exist in Zr70Ni30 AA because its absorption edge does not shift to a higher energy position compared with that of Zr foil. In addition, a tendency for the absorption edge to shift to a higher energy position is also detected in Figure 1d when increasing the Pd concentration in the ZrNiPd system, which is attributed to the fact that the Pd–Zr pair has much stronger interaction than its Ni–Zr counterpart. To further clearly probe the atomic structure of Zr70Pd30, Voronoi tessellation is carried out based on the RMC simulated model, resulting in various indexed Voronoi clusters (VCs). The distribution of major VCs centered with Zr or Pd atoms is plotted in Figure 2. It is found that icosahedral-like clusters  (<0, 2, 8, 2>, <0, 3, 6, 3>, <0, 0, 12, 0>, <0, 1, 10, 2> and <0, 2, 8, 1>, etc.) coexist with non-icosahedral VCs (<0, 4, 4, 3>, <0, 3, 6, 2> and <0, 3, 6, 1>, etc.), building the atomic structure in this sample. Compared with the data in Figure 3 in our previous work studying Zr70Ni30 AA , it is observed that icosahedral-like clusters are more abundant in Zr70Pd30. This may explain why a primary phase transition by forming nanoicosahedral quasicrystal occurs in Zr70Pd30 after annealing, which is absent in Zr70Ni30 .
L. Yang et al. / Scripta Materialia 63 (2010) 883–886
Figure 2. Distribution of Voronoi clusters centered with (a) Zr or (b) Pd atoms in Zr70Pd30 AA.
Based on the cluster-level atomic structure deduced above, we may further study this shortened Zr–Pd bond in Zr70Pd30 AA by analyzing its electronic basis. The DMol3 cluster method based on density functional theory  within the generalized gradient approximation  was performed by calculating electronic interaction in two representative Zr-centered VCs indexed as <0, 2, 8, 2> and <0, 3, 6, 3>, as well as two clusters extracted from the Zr2Pd1 tetragonal phase . The relativistic effects were included. The three-dimensional (3-D) images of these two VCs after a predetermined relaxation of the structure, the corresponding total density of state (DOS) and the partial DOS originating from representative atoms are shown in Figure 3. The clusters in the corresponding crystal phase are also analyzed, as shown in Figure 4. The partial DOS plots of <0, 2, 8, 2> (see Fig. 3a and b) and <0, 3, 6, 3> (see Fig. 3c and d) show that the total DOS located at around 4.2 eV is predominantly contributed by Zr:5s (Zr1:5s contributes about 18%, with the rest contributed by other Zr:5s and Pd:4d5s). In the energy range from 3.7 eV to the Fermi energy level (EF), DOS is constructed by both Zr:4d5s5p and Pd:4d5s5p. We also notice that for the spin-up of the partial DOS in Figure 3a the levels slightly below EF are primarily derived from Pd2:5s5p orbitals (about 35%), the rest being from the 4d orbitals of other Zr and Pd atoms. These results clearly indicate that a strong hybridization between Pd:4d5s5p and Zr:4d5s5p atoms occurs in these VCs. Furthermore, it is interesting to note that there is also some hybridization between Zr1:4d and Pd2:4d around 4.0 eV (see Fig. 4a) and 1.7 eV (see Fig. 4a and b), between Zr2:4d and Pd1:4d/Pd10:4d around 1.4 eV and between Zr2:4d and Pd:4d5s slightly below EF (see Fig. 4c and d). However, it is relatively weaker than that in the selected VCs in ZrPd AA. The above calculation and analysis of the electronic basis are consistent ˚ in the cryswith the Zr–Pd bond being shortened by 0.1 A
Figure 3. 3-D pictures of the <0, 2, 8, 2> and <0, 3, 6, 3> VCs after a predetermined relaxation of the structure. The corresponding total and partial DOS plots are presented in (a and b) for <0, 2, 8, 2> and (c and d) for <0, 3, 6, 3>, respectively. Here we select partial DOS originating from the representative atoms of Zr1, Zr7, Pd2 and Pd4 in <0, 2, 8, 2> and Zr1, Zr5, Pd2 and Pd3 in <0, 3, 6, 3>, respectively. The partial DOS are displayed in threefold magniﬁcation. The number behind the atoms denotes their serial number in these clusters. The Fermi energy level (EF) possesses a position of 0 eV.
Figure 4. 3-D pictures of the Zr10Pd4 and Zr8Pd5 clusters extracted from the Zr2Pd1 tetragonal phase. The corresponding total DOS and partial DOS plots: (a) and (b) for Zr10Pd4, (c and d) for Zr8Pd5, respectively. Here we select partial DOS originating from the representative atoms of Zr1, Zr6, Zr14, Pd2 in Zr10Pd4 and Zr2, Pd1, Pd10 in Zr8Pd5, respectively.
˚ in the amorphous phase, compared tal phase and by 0.2 A with that in SGAR (see Table 1). Therefore, Zr and Pd atoms could be regarded as “soft” spheres in ZrPd AA. To further quantitatively address the bonding nature between the atoms in both ZrPd and ZrNi binary AAs, the average Mulliken charge (AMC) values of <0, 2, 8, 2> and <0, 3, 6, 3> Zr-centered VCs taken from ZrPd are listed in Table 2. For comparison, the clusters by replacing all Pd by Ni atoms were also calculated, without changing their positions in the initial models. For the Zr– Pd system, it is clear that only the Zr atom in the center (Zr1) gains electrons, the other Zr atoms lose electrons while all Pd are acceptors. The same tendency could be
L. Yang et al. / Scripta Materialia 63 (2010) 883–886
Table 2. The average Mulliken charges of two typical Zr-centered VCs, based on which we may construct clusters by replacing all Pd with Ni atoms and perform the same calculation. Mulliken occupation <0, 2, 8, 2> VC Zr1 Pd2 Pd3 Pd4 Zr5 Zr6 Zr7 Zr8 Zr9 Zr10 Zr11 Zr12 Zr13
<0, 3, 6, 3> VC 0.05 0.37 0.28 0.30 0.07 0.12 0.18 0.15 0.05 0.15 0.13 0.15 0.01
Zr1 Ni2 Ni3 Ni4 Zr5 Zr6 Zr7 Zr8 Zr9 Zr10 Zr11 Zr12 Zr13
0.09 0.09 0.09 0.11 0.01 0.06 0.08 0.05 0.00 0.07 0.06 0.07 0.03
Zr1 Pd2 Pd3 Zr4 Zr5 Pd6 Zr7 Zr8 Zr9 Zr10 Zr11 Zr12 Zr13
0.50 0.28 0.22 0.07 0.15 0.25 0.23 0.23 0.12 0.15 0.08 0.17 0.05
Zr1 Ni2 Ni3 Zr4 Zr5 Ni6 Zr7 Zr8 Zr9 Zr10 Zr11 Zr12 Zr13
0.08 0.02 0.06 0.04 0.04 0.09 0.09 0.01 0.04 0.08 0.05 0.10 0.02
The number behind the atoms denotes their serial number in the clusters depicted in Figure 3. The negative value represents an accumulation of electron.
observed for the Zr–Ni system. However, on average, it seems that Pd atoms are apt to gain more electrons than their Ni counterparts (see Table 2; for example, for <0, 2, 8, 2>, the AMC of Pd is about 0.3, while that of Ni is only 0.06). This can be understood by taking into account the fact that Pauling’s electronegativity of Pd (2.20) is larger than that of Ni (1.91) [29,30]. On the other hand, the diﬀerence in the atomic energy level between Zr:4d5s5p and Pd:4d5s5p is smaller than that between Zr:4d5s5p and Ni:3d4s4p, resulting in a stronger hybridization between Pd and Zr pairs. In summary, the shortened bonds between “soft” atoms in Zr70Pd30 AA were detected by RMC simulation using synchrotron radiation experimental data, and further explained by calculation and analysis of its electronic basis. While “hard sphere” atoms have been extensively adopted to establish a atomic structural model in a number of developed AAs, this phenomenon strongly suggests that soft atoms may also have a signiﬁcant impact on the internal structure in some metal–metal AAs. This study could enhance our understanding of the glass-forming mechanism as well as the unique properties of these novel glassy materials. The authors would like to thank HASYLAB in Germany, BSRF in Beijing and NSRL in Hefei for the use of the advanced synchrotron radiation facilities. Financial support from the National Natural Science Foundation of China (Grants Nos. 10805027 and 10804058), the Natural Science Foundation of Jiangsu Province (Grant No. BK2008397), the NUAA Research Funding (Grant No. NS2010168) and the Natural Science Foundation of Zhejiang Province (Grant No. Y607546) is gratefully acknowledged.    
G.D. Scott, Nature (London) 188 (1960) 908. P.H. Gaskell, J. Non-Cryst. Solids 32 (1979) 207. T. Egami, Y. Waseda, J. Non-Cryst. Solids 64 (1984) 113. J. Saida, N. Matsushita, A. Inoue, Appl. Phys. Lett. 79 (2001) 412.  J. Saida, M. Kasai, E. Matsubara, A. Inoue, Ann. Chim. Sci. Mater. 27 (2002) 77 (Elsevier, Paris 2002).  K. Saksl, H. Franz, P. Jo´va´ri, K. Klementev, E. Welter, A. Ehnes, J. Saida, A. Inoue, J.Z. Jiang, Appl. Phys. Lett. 83 (2003) 3924.
 C. Fan, P.K. Liaw, T.W. Wilson, W. Dmowski, H. Choo, C.T. Liu, Appl. Phys. Lett. 89 (2006) 111905.  H.W. Sheng, W.K. Luo, F.M. Alamgir, J.M. Bai, E. Ma, Nature (London) 439 (2006) 419.  J.D. Bernal, Nature (London) 185 (1960) 68.  J.D. Bernal, Proc. R. Soc. Lond. A 280 (1964) 299.  D. Ma, A.D. Stoica, L. Yang, X.L. Wang, Z.P. Lu, J. Neuefeind, M.J. Kramer, J.W. Richardson, T. Proﬀen, Appl. Phys. Lett. 90 (2007) 211908.  D.B. Miracle, W.S. Sanders, O.N. Senkov, Philos. Mag. 83 (2003) 2409.  D.B. Miracle, Acta Mater. 54 (2006) 4317.  A. Inoue, Bulk Amorphous Alloys – Practical Characteristics and Applications, Trans. Tech. Publications, Zurich, 1999, p. 141.  K. Ahn, D. Louca, S.J. Poon, G.J. Shiﬂet, Phys. Rev. B 70 (2004) 224103.  F.Q. Guo, S.J. Poon, G.J. Shiﬂet, Appl. Phys. Lett. 84 (2004) 37.  K. Saksl, P. Jovari, H. Franz, J.Z. Jiang, J. Appl. Phys. 97 (2005) 113507.  H.W. Sheng, H.Z. Liu, Y.Q. Cheng, J. Wen, P.L. Lee, W.K. Lou, S.D. Shastri, E. Ma, Nat. Mater. 6 (2007) 192.  A.P. Hammersley, S.O. Svensson, M. Hanﬂand, A.N. Fitch, D. Ha¨usermann, High Pressure Res. 14 (1996) 235.  T.E. Faber, J.M. Ziman, Philos. Mag. 11 (1965) 153.  T.J. Ressler, J. Synchrotron Radiat. 5 (1998) 118.  R.L. McGreevy, L. Pusztai, Mol. Simul. 1 (1988) 359.  R.L. McGreevy, J. Phys.: Condens. Matter 13 (2001) R877.  J.L. Finney, Nature 266 (1977) 309.  G. Evrard, L. Pusztai, J. Phys.: Condens. Matter 17 (2005) S1.  J. Saida, M. Kasai, E. Matsubara, A. Inoue, Ann. Chim. Sci. Mater. 27 (2002) 77.  R.S. Liu, L.Y. Jang, H.H. Hung, J.L. Tallon, Phys. Rev. B 63 (2001) 212507.  X.L. Li, X.F. Bian, L.N. Hu, Y.Q. Wu, J. Guo, J.Y. Zhang, J. Appl. Phys. 101 (2007) 103540.  J.A. Dean, Lange’s Handbook of Chemistry, 15th ed., McGraw-Hill, New York, 1999.  http://en.wikipedia.org/wiki/Electronegativity.  S.Y. Wang, M.J. Kramer, M. Xu, S. Wu, S.G. Hao, D.J. Sordelet, K.M. Ho, C.Z. Wang, Phys. Rev. B 79 (2009) 144205.  L. Yang, S. Yin, X.D. Wang, Q.P. Cao, J.Z. Jiang, J. Appl. Phys. 102 (2007) 083512.  B. Delley, J. Chem. Phys. 92 (1990) 508.  J.P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 77 (1996) 3865.