- Email: [email protected]
Effect of Minor Similar Elements Substitution on Glass Forming Ability and Fragility of Al-Ni-based Amorphous Alloys
Available online at www.sciencedirect.com
Procedia Engineering 16 (2011) 755 – 762
International Workshop on Automobile, Power and Energy Engineering
Effect of Minor Similar Elements Substitution on Glass Forming Ability and Fragility of Al-Ni-based Amorphous Alloys Junzhe Sun, Xiufang Biana*, Yanwen Bai Key Laboratory for Liquid–Solid Structural Evolution and Processing of Materials, Ministry of Education Shandong University Jinan 250061,China
Abstract Compared with the Al84Ni10Ce6 alloy, the glass forming ability (GFA) of Al84Ni10(Ce3La3) was improved and the melt fragility was decreased through the method of La addition (similar to Ce element), which is in contrary to the conventional rules of dissimilar elements (large atomic size differences) addition. This phenomenon can be understood from the thermodynamics of decrease in ΔGl-x. Furthermore, the effects of other similar RE elements (Pr, Nd, Gd) substitution on GFA and different fragility parameters were investigated in Al84Ni10 (Ce3RE3) alloys, and it was found that smaller M corresponds to a better GFA in these alloys. In addition, it can be also confirmed that m could be just used in an individual alloy system which may attribute to the different formation mechanism of metallic glasses.
© 2010 Published by Elsevier Ltd. Selection and/or peer-review under responsibility of Society for Automobile, Power and Energy Engineering Keywords: Similar elements; Glass forming ability; Melt fragility
1. Introduction In recent years, Al-based metallic glasses have attracted increasing attention due to their superior specific strength and other good mechanical properties [1]. Although the Al-based bulk metallic glasses (BMGs) were reported [2], unfortunately the size of the Al-based amorphous is still limited to 1 mm. In the course of discovering the effective methods to develop large size of the Al-based BMGs, people have focused on the constituent elements with large atomic size differences and large negative mixing enthalpy,
* Corresponding author. Tel.: +5-318-839-2748; fax: +5-318-839-5011. E-mail address: [email protected].
1877-7058 © 2011 Published by Elsevier Ltd. doi:10.1016/j.proeng.2011.08.1151
756
Junzhe Sun et al. / Procedia Engineering 16 (2011) 755 – 762
which is helpful to promote the GFA by making more efficient atomic packing, while the effect of similar elements (similar atomic size but various electronic structures) addition has been neglected. However, the existence of the similar elements pairs can also have a beneficial impact in glass forming systems, such as the similar elements pairs of Co and Fe in Co-Fe-B-Si-Nb system [3], Ni and Fe in Ni–Fe–B–Si–Nb system [4], Cu and Ni in La-Al-Cu-Ni system [5]. The remarkable one is that Zhang et al. found the critical diameter of the quinary (La0.5Ce0.5)65Al10(Co0.6Cu0.4)25 BMG is significantly higher than that of the ternary Ln–Al–TM alloys (Ln = La or Ce; TM = Co or Cu), and the similar pairs of La and Ce, Co and Cu have greatly improved the GFA in this quinary alloy [6]. Recently liquid fragility has been used as a key to investigate the glass community. According to the deviation of the viscosity’s temperature dependence from Arrhenius behavior approaching Tg (the glass transition temperature), Angell [7] proposed the concept of fragility of supercooled liquid (m), which was extensively used to classify the strong-fragile characteristics of glass-forming liquids. However, it is hard to obtain m of the alloys with low GFA. And then, melt fragility (M), namely fragility of superheated melt, which is defined as the viscosity variation rate of superheated liquids towards the liquidus, has been proposed by Bian as a new concept and it can reflect the GFA directly, especially in marginal alloys like Al-based alloys [8]. In this paper we chose Al84Ni10Ce6 as the matrix alloy, the maximum thickness of which is up to 200 μm [9]. The lanthanide elements of La, Pr, Nd and Gd, which are similar to Ce, were used to substitute a small quantity of Ce. The influence of similar elements substitution on GFA, m and M and their relations in Al-based alloys were investigated in detail. And the reason for the improvement on GFA will be explained by thermodynamics of ΔGl-x and dynamics of viscosity. 2. Experimental procedure The samples of Al84Ni10(Ce3RE3) (RE = La, Pr, Nd, Gd) alloys were prepared from pure Al (99.9% mass), pure Ni (99.9% mass), and industry pure Ce, La, Pr, Nd and Gd. The purity of RE is about 99.6% mass. Then the prealloyed ingots were remelted and rapidly solidified into ribbons at different circumferential speeds by a single roller melt spinning apparatus. The structure of the ribbons was examined by X-ray diffraction (XRD) with a D/Max-rB diffractometer with Cu Kα radiation. Thermal properties of the ribbons were evaluated by a Netzsch DSC404 calorimeter using pure indium (99.999wt %) and pure zinc (99.999wt %) standards. Besides the viscosity, thermal scanning is also thought to be a reliable way to obtain the fragility of the supercooled liquid [10]. The parameter of the fragility of the supercooled liquid, m, can be calculated by the following equation [11],
m = ΔH ∗ / RTg , 20 ,
(1)
*
where ΔH is the active enthalpy of the glass transition, R is the gas constant, Tg,20 is the glass transition temperature at a scanning rate of 20 K/min examined by the DSC. Meanwhile, ΔH* can be calculated according to the different glass transition temperatures dependent on the heating rates. The equation is as follows [12, 13],
ln(Tg2 / ) = ΔH ∗ / RTg + Const
(2)
where Φ is the scanning rates, the glass transition temperature (Tg) can be determined by the DSC curves of different heating rates. During the DSC measurement, the samples were heating under different rates of 10, 20, 40 and 50 K/min, respectively.
757
Junzhe Sun et al. / Procedia Engineering 16 (2011) 755 – 762
For viscosity measurement, the samples were measured by an oscillating viscometer in a high vacuum atmosphere. The samples were first overheated to about 250 K above their liquidus and held for 1 h and then cooled to the required temperature for the viscosity measurements. The kinetic viscosity was calculated according to the Shvidakovskii’s equation [14]. The viscosity of the superheated melts can be successfully fitted by the Arrhenius equation in a wide temperature range above liquidus, = 0 exp( E / RT ) (3) where η0 is a pre-exponential constant, which is associated with the nature of the liquid, E is the activation energy for viscous flow and R is the gas constant. According to the definition, melt fragility (M) is defined as [8]:
M=
∂η(T)/η(TL ) ∂T/TL
T = TL
=
∂ηr ∂Tr
(4)
Tr =1
where TL is the liquidus temperature, ηL is the viscosity at the liquidus. According to the equation (3) and (4), M can be calculated by
M = Ε/RTL
(5)
3. Results and discussion In order to compare with each other, all the fully amorphous ribbons of Al84Ni10(Ce3RE3) (RE = La, Pr, Nd, Gd) alloys were obtained at the same circumferential speed of 14.7 m/s. The amorphous structures were also confirmed by the XRD patterns shown in Fig.1. The broad diffraction peaks found in these alloys indicate no detectable crystalline phase in them. From Fig.1 we can find that with the different elements addition, the main peaks of the XRD patterns have a slight deviation according to the Gaussian fitted curves. Fig. 2 shows the DSC curves of the Al84Ni10(Ce3RE3) (RE = La, Pr, Nd, Gd) ribbons during the heating and cooling process at a rate of 20K/min. The characteristic temperatures and parameters, such as the reduced glass transition temperature Trg(Trg = Tg/TL) [15] and γ(γ = Tx/(TgTL)) [16] for all the glassy ribbons are shown in Table 1. In our work it is difficult to get amorphous ribbons below the lowest circumferential speed v0. The relevant parameters of the Al84Ni10Ce6 were also listed for comparison and its parameters are close to the data in Ref. [18].
Intensity(a.u.)
d: Al Ni Ce Gd 84 10 3 3 c: Al Ni Ce Nd 84 10 3 3 b: Al Ni Ce Pr 84 10 3 3 a: Al Ni Ce La 84 10 3 3
10
20
30
40
50
60
70
80
2 (degree) Fig. 1. XRD patterns of the Al84Ni10(Ce3RE3) (RE = La, Pr, Nd, Gd) alloys
758
Junzhe Sun et al. / Procedia Engineering 16 (2011) 755 – 762
Exothermic
Al84Ni10Ce3Gd3 Al84Ni10Ce3Nd3
Al84Ni10Ce3Pr3 Al84Ni10Ce3La3
400
600
800
1000
Temperatuer(K)
1200
Fig. 2. DSC curves of Al84Ni10(Ce3RE3) (RE = La, Pr, Nd, Gd) alloys during the heating and cooling processes
Table 1 Thermal parameters of Al84Ni10Ce6 and Al84Ni10(Ce3RE3) (RE = La, Pr, Nd, Gd) alloys (ΔHf is the heat of fusion, ΔGl-x is the Gibbs free energy difference, ΔHmix is the mixing enthalpy, and v0 is the lowest circumferential speed to obtain fully amorphous ribbons) Alloys
Tg (K)
Tx (K)
TL (K)
Trg
γ
v0 (m/s)
ΔHf kJ/mol
ΔGl-x kJ/mol
ΔHmix kJ/mol
Al84Ni10Ce6 Al84Ni10Ce3La3 Al84Ni10Ce3Pr3 Al84Ni10Ce3Nd3 Al84Ni10Ce3Gd3
553.7 557.2 549.0 560.0 554.7
559.2 565.6 554.9 567.0 561.0
1172.3 1094.9 1186.6 1209.9 1230.2
0.472 0.509 0.463 0.463 0.451
0.324 0.342 0.320 0.320 0.314
11 5.5[17] 14.7 14.7 14.7
6.96 6.24 7.68 7.10 8.31
3.21 2.88 3.55 3.31 3.88
-15.72 -15.71 -15.75 -15.75 -15.86
From Table 1, it can be seen that the Al84Ni10Ce3La3 has a lower value of v0 than that of Al84Ni10Ce6, indicating that the coexistence of similar elements pairs of La and Ce has improved the GFA, while the addition of Pr, Nd and Gd elements has decreased the GFA. In order to further compare with each other, we use γ to characterize the GFA of these alloys. Generally speaking, the higher γ corresponds to the better GFA [16]. In the Al84Ni10(Ce3RE3) alloy system, the γ also shows a negative correlation with the lowest circumferential speed v0, indicating that γ is reliable to characterize the GFA in our work. The minor addition of different similar elements causes the change of the number and the shape of the main exothermic peaks for the crystallization, indicating that the minor addition affects the GFA via changing the crystalline process. Then we focus on the improvement of the GFA by the similar elements substitution. As we know, the RE elements have similar atomic size, chemical and physical properties. Since the similar element pairs of La-Ce are neighbours in the Periodic Table of Elements, the difference in atomic size between them is very small and the mixing enthalpy (ΔHABmix,, here A and B represent two kinds of elements) [19] between them is 0 kJ/mol. Therefore, the traditional criterion of mismatches in atomic size and the negative heat of mixing cannot give a reasonable explanation for the improvement in GFA. Consequently we want to discuss the possible reason in thermodynamic of Gibbs free energy difference and dynamic of viscosity for this phenomenon. Thermodynamically, the Gibbs free energy difference between the liquid and the crystalline state is a very important parameter for the glass forming. As is investigated by Busch et al. [20], the lower Gibbs
Junzhe Sun et al. / Procedia Engineering 16 (2011) 755 – 762
free energy difference (ΔGl-x(T)) between the liquid and the crystalline states corresponds to the better GFA, indicating that the high driving force for crystallization decreases the GFA. The difference in the Gibbs free energy between the liquid and crystalline states is given by [21, 22]:
Δ
−
= ⎡Δ ⎢⎣
−∫ Δ m
−
84Ni10Ce6
⎤− ⎥⎦
⎡ ⎢Δ ⎢⎣
−∫
m
Δ
−
⎤ ⎥ ⎥⎦
(c) Al84Ni10Ce3Pr3 (d)Al84Ni10Ce3Nd3 (e)Al84Ni10Ce3Gd3
According to the method mentioned previously, we calculated the parameter of the fragility of the supercooled liquids (m). The values of activation enthalpy for glass transition ΔH* and m are listed in Table 2. The different glass transition temperatures at different heating rates are obtained from different heating rates of DSC curves. Fig. 4 shows the correlation between ln(Tg2/Φ) and 1000/Tg for the Al84Ni10(Ce3RE3) (RE = La, Pr, Nd, Gd) alloys. The effect of similar elements substitution on the GFA in different Al-based alloys may also attribute to the change of the liquid behavior. As shown in Table 2,
759
760
Junzhe Sun et al. / Procedia Engineering 16 (2011) 755 – 762
Al84Ni10Ce3La3 alloy with the best GFA has the lowest m, indicating it has a more stable and stronger supercooled liquid than the other alloys. However, among the alloys of Al84Ni10(Ce3RE3) (RE = Pr, Nd, Gd), the Al84Ni10Ce3Gd3 alloy has the smallest m of 85.63 corresponding to the smallest γ of 0.314. Hence, it can be verified that a larger m corresponds to a better GFA. Generally speaking, the strong liquid usually exhibits a good GFA with a small m. The possible reason for this unusual tendency in Al-based alloys is the heterogeneous nucleation in liquids resulting from the existence of local heterogeneous domains, which decreases the influence of the viscosity in the glass transition region and promotes the nucleation [26]. It can be inferred that the smaller m corresponds to a better GFA only works in an individual alloy system, not in the different systems of Al84Ni10(Ce3RE3) (RE = La, Pr, Nd, Gd) alloys. This finding accords with the result that was found in Ref. [27]. Table 2 Activation enthalpy for glass transition (ΔH*), the fragility of the supercooled liquids (m), the fitting parameters of E/R and η0, and calculated parameters of ηL and M of Al84Ni10(Ce3RE3) (RE = La, Pr, Nd, Gd) alloys Alloys
ΔH* (kJ/mol)
m
η0 (10-3Pa.s)
E/R (K)
TL (K)
ηL (10-3Pa.s)
M
Al84Ni10Ce3La3 Al84Ni10Ce3Pr3 Al84Ni10Ce3Nd3 Al84Ni10Ce3Gd3
267.18 397.61 420.28 394.89
56.82 87.11 90.27 85.63
1.830 1.400 1.079 1.976
1117 1242 1255 1291
1094.9 1186.6 1209.9 1230.2
5.075 3.985 3.044 5.647
1.020 1.046 1.037 1.050
10.4 10.2
ln(Tg2/)
10.0
(a) (b) (c) (d)
9.8 9.6 9.4 9.2 9.0 8.8 8.6 1.76 1.77 1.78 1.79 1.80 1.81 1.82 1.83 1.84
1000/Tg(K-1) Fig. 4. Correlation between ln(Tg2/Φ) and 1000/Tg of (a) Al84Ni10Ce3La3 (b)Al84Ni10Ce3Pr3 (c) Al84Ni10Ce3Nd3 (d)Al84Ni10Ce3Gd3 alloys
On the other hand, the viscosity of the liquid is an important dynamics parameter to measure the process of the glass forming and it has a close relationship with the microstructure of the liquid. Fig. 5 shows the experimental viscosity data of Al84Ni10(Ce3RE3) (RE = La, Pr, Nd, Gd) alloy melts at high temperature during the cooling process. The discrete viscosity values are fitted by the Arrhenius law. Thermodynamic properties, especially the important one of mixing heat, which reflect the interaction between the components in the melts are considered to be useful in discussing the behaviors of the viscosity [28]. It can be inferred a stronger atoms interactions in the Al84Ni10(Ce3Gd3) alloy with the maximal viscosity, which corresponds to the largest mixing enthalpy ΔHmix of it. The values of the fitting
Junzhe Sun et al. / Procedia Engineering 16 (2011) 755 – 762
parameters according to equation (3) are listed in Table 2. Generally speaking, the lower M corresponds to the higher GFA, suggesting that the viscosity variation rate of superheated liquids towards the liquidus is slower and the liquid is more stable. In our work, it can be seen that the M has a good negative relationship with the GFA among these alloys from Table 2. The coexistence of La and Ce hinders the precipitation of the intermetallic compound, thus in the view of this improved the GFA.
6.0
Viscosity(10-3Pa*s)
5.5 5.0
(a) (b) (c) (d)
4.5 4.0 3.5 3.0 2.5 1150 1200 1250 1300 1350 1400 1450 1500 1550
Temperature(K) Fig. 5. The experimental viscosity data and fitting curves of (a) Al84Ni10Ce3La3 (b)Al84Ni10Ce3Pr3 (c) Al84Ni10Ce3Nd3 (d)Al84Ni10Ce3Gd3 alloys
4. Conclusions In this paper, the coexistence of similar element pairs of La-Ce improved the GFA, while the Pr, Nd and Gd addition decreased the GFA in Al-Ni-Ce alloy system. Among the Al84Ni10(Ce3RE3) (RE = La, Pr, Nd, Gd) alloys, the Gibbs free energy difference between the liquid and the crystalline states (ΔGl-x) has a negative relationship with the GFA according to the lowest circumferential speed v0 and γ. The decrease in ΔGl-x and melt fragility can give a reasonable explanation for the improvement in GFA by the minor substitution of similar elements. Due to the different formation mechanism, it can be verified that the fragility parameter of m could be generally confined to an individual alloy system. Acknowledgements This work was supported by National Basic Research Program of China 973 Program, Grant No. 2007CB613901 and National Natural Science Foundation of China (Nos.50831003, 50871062) and the Natural Science Foundation of Shandong province (Grant No. Z2008F08). References [1] Inoue A, Kimura H. Fabrications and mechanical properties of bulk amorphous, nanocrystalline, nanoquasicrystalline alloys in aluminum-based system. J. Light Met 2001; 1: 31-41. [2] Yang BJ, Yao JH, Zhang J, Yang HW, Wang JQ, Ma E. Al-rich bulk metallic glasses with plasticity and ultrahigh specific strength. Scripta Mater. 2009; 61: 423-426. [3] Chang CT, Shen BL, Inoue A. Co–Fe–B–Si–Nb bulk glassy alloys with superhigh strength and extremely low magnetostriction. Appl. Phys. Lett. 2006; 88: 011901.
761
762
Junzhe Sun et al. / Procedia Engineering 16 (2011) 755 – 762
[4] Shen BL, Chang CT, Inoue A. Ni-based bulk glassy alloys with superhigh strength of 3800 MPa in Ni–Fe–B–Si–Nb system. Appl. Phys. Lett. 2006; 88: 201903. [5] Tan H, Zhang Y, Ma D, Feng YP, Li Y. Optimum glass formation at off-eutectic composition and its relation to skewed eutectic coupled zone in the La based La–Al–(Cu,Ni) pseudo ternary system. Acta Mater. 2003; 51: 4551. [6] Zhang T, Li R, Pang SJ. Effect of similar elements on improving glass-forming ability of La–Ce-based alloys. J. Alloys Compd. 2009; 483: 60. [7] Angell CA. Spectroscopy simulation and scattering, and the medium range order problem in glass. J. Non-Cryst. Solids 1985; 73: 1-17. [8] Bian XF, Sun BA, Hu LN, Jia YB. Fragility of superheated melts and glass-forming ability in Al-based alloys. Phys. Lett. A 2005; 335: 61-67. [9] Inoue A. Amorphous, nanoquasicrystalline and nanocrystalline alloys in Al-based systems. Prog. Mater. Sci. 1998; 43: 365-520. [10] Crowley KJ, Zografi G. The use of thermal methods for predicting glass-former fragility. Thermochim. Acta 2001; 380: 79-93. [11] Brüning R, Mark S. Fragility of glass-forming systems and the width of the glass transition. J. Non-Cryst. Solids 1996; 205-207: 480-484. [12] Moynihan CT. Nonexponential relaxation in strong and fragile glass formers. Am. Ceram. Soc. 1993; 76: 1081-1087. [13] Kasap SO, Yannacopoulos S. Kinetics of structural relaxations in the glassy semiconductor a-Se. Mater. Res. 1989; 4: 893-905. [14] Schvidkovskii EG, Gostekhteorizdat, USSR, Science Press, 1955. [15] Lu ZP, Tan H, Ng SC, Li Y. The correlation between reduced glass transition temperature and glass forming ability of bulk metallic glasses. Scripta Mater. 2000; 42: 667-673. [16] Lu ZP, Liu CT. Glass Formation Criterion for Various Glass-Forming Systems. Phys. Rev. Lett. 2003, 91, 115505-1. [17] Si PC, Bian XF, Zhang JY, Li H, Sun MH, Zhao Y. The fragility of Al–Ni-based glass-forming melts. J. Phys.: Condens. Matter 2003; 15: 5409. [18] Hu LN, Bian XF, Wang WM, Zhang JY, Jia YB. Liquid fragility and characteristic of the structure corresponding to the prepeak of AlNiCe amorphous alloys. Acta Mater. 2004; 52: 4773-4781. [19] Boer FR, Boom R, Mattens WCM, Miedema AR, Niessen AK. Cohesion in metals. Amsterdam, North-Holland, 1989. [20] Bakke, Eric, Dissertation, Viscosity measurement of bulk metallic glass forming alloys Dissertation. California Institute of Technology, Pasadena, CA, 1997. [21] Glade SC, Busch R, Lee DS, Johnson WL. Thermodynamics of Cu47Ti34Zr11Ni8, Zr52.5Cu17.9Ni14.6Al10Ti5 and Zr57Cu15.4Ni12.6Al10Nb5 bulk metallic glass forming alloys. J. Appl. Phys. 2000; 87: 7242. [22] Lu ZP, Hu X, Li Y. Thermodynamics of La based La-Al-Cu-Ni-Co alloys studied by temperature modulated DSC. Intermetallics 2000 ; 8: 477-480. [23] Thompson CV, Spaepen F. On the approximation of the free energy change on crystallization. Acta Metall. 1979; 27: 18551859. [24] Fecht HJ, Free energy functions of undercooled liquid glass-forming Au-Pb-Sb alloys. Mater.Sci. Eng. A 1991; 133: 443. [25] Zhang Y, Tan H, Kong HZ, Yao B, Li Y. Glass-forming ability of Pr-(Cu,Ni)-Al alloys in eutectic. J. Mater. Res. 2003; 18: 664. [26] Gerhard W. Thermodynamics, viscous flow and relaxation dynamics of bulk glass-forming Pd alloys. Ann. Chim. Sci. Mat. 2002; 27: 49. [27] Song KK, Bian XF, Lv XQ, Xie MT, Jia R. The correlations among the fragility of supercooled liquids, the fragility of superheated melts, and the glass-forming ability for marginal metallic glasses. J. Appl. Phys. 2009; 105: 024304. [28] Sato Y, Sugisawa K, Aoki D, Yamamura T. Viscosities of Fe–Ni, Fe–Co and Ni–Co binary melts. Meas. Sci. Technol. 2005; 16: 363.
Procedia Engineering 16 (2011) 755 – 762
International Workshop on Automobile, Power and Energy Engineering
Effect of Minor Similar Elements Substitution on Glass Forming Ability and Fragility of Al-Ni-based Amorphous Alloys Junzhe Sun, Xiufang Biana*, Yanwen Bai Key Laboratory for Liquid–Solid Structural Evolution and Processing of Materials, Ministry of Education Shandong University Jinan 250061,China
Abstract Compared with the Al84Ni10Ce6 alloy, the glass forming ability (GFA) of Al84Ni10(Ce3La3) was improved and the melt fragility was decreased through the method of La addition (similar to Ce element), which is in contrary to the conventional rules of dissimilar elements (large atomic size differences) addition. This phenomenon can be understood from the thermodynamics of decrease in ΔGl-x. Furthermore, the effects of other similar RE elements (Pr, Nd, Gd) substitution on GFA and different fragility parameters were investigated in Al84Ni10 (Ce3RE3) alloys, and it was found that smaller M corresponds to a better GFA in these alloys. In addition, it can be also confirmed that m could be just used in an individual alloy system which may attribute to the different formation mechanism of metallic glasses.
© 2010 Published by Elsevier Ltd. Selection and/or peer-review under responsibility of Society for Automobile, Power and Energy Engineering Keywords: Similar elements; Glass forming ability; Melt fragility
1. Introduction In recent years, Al-based metallic glasses have attracted increasing attention due to their superior specific strength and other good mechanical properties [1]. Although the Al-based bulk metallic glasses (BMGs) were reported [2], unfortunately the size of the Al-based amorphous is still limited to 1 mm. In the course of discovering the effective methods to develop large size of the Al-based BMGs, people have focused on the constituent elements with large atomic size differences and large negative mixing enthalpy,
* Corresponding author. Tel.: +5-318-839-2748; fax: +5-318-839-5011. E-mail address: [email protected].
1877-7058 © 2011 Published by Elsevier Ltd. doi:10.1016/j.proeng.2011.08.1151
756
Junzhe Sun et al. / Procedia Engineering 16 (2011) 755 – 762
which is helpful to promote the GFA by making more efficient atomic packing, while the effect of similar elements (similar atomic size but various electronic structures) addition has been neglected. However, the existence of the similar elements pairs can also have a beneficial impact in glass forming systems, such as the similar elements pairs of Co and Fe in Co-Fe-B-Si-Nb system [3], Ni and Fe in Ni–Fe–B–Si–Nb system [4], Cu and Ni in La-Al-Cu-Ni system [5]. The remarkable one is that Zhang et al. found the critical diameter of the quinary (La0.5Ce0.5)65Al10(Co0.6Cu0.4)25 BMG is significantly higher than that of the ternary Ln–Al–TM alloys (Ln = La or Ce; TM = Co or Cu), and the similar pairs of La and Ce, Co and Cu have greatly improved the GFA in this quinary alloy [6]. Recently liquid fragility has been used as a key to investigate the glass community. According to the deviation of the viscosity’s temperature dependence from Arrhenius behavior approaching Tg (the glass transition temperature), Angell [7] proposed the concept of fragility of supercooled liquid (m), which was extensively used to classify the strong-fragile characteristics of glass-forming liquids. However, it is hard to obtain m of the alloys with low GFA. And then, melt fragility (M), namely fragility of superheated melt, which is defined as the viscosity variation rate of superheated liquids towards the liquidus, has been proposed by Bian as a new concept and it can reflect the GFA directly, especially in marginal alloys like Al-based alloys [8]. In this paper we chose Al84Ni10Ce6 as the matrix alloy, the maximum thickness of which is up to 200 μm [9]. The lanthanide elements of La, Pr, Nd and Gd, which are similar to Ce, were used to substitute a small quantity of Ce. The influence of similar elements substitution on GFA, m and M and their relations in Al-based alloys were investigated in detail. And the reason for the improvement on GFA will be explained by thermodynamics of ΔGl-x and dynamics of viscosity. 2. Experimental procedure The samples of Al84Ni10(Ce3RE3) (RE = La, Pr, Nd, Gd) alloys were prepared from pure Al (99.9% mass), pure Ni (99.9% mass), and industry pure Ce, La, Pr, Nd and Gd. The purity of RE is about 99.6% mass. Then the prealloyed ingots were remelted and rapidly solidified into ribbons at different circumferential speeds by a single roller melt spinning apparatus. The structure of the ribbons was examined by X-ray diffraction (XRD) with a D/Max-rB diffractometer with Cu Kα radiation. Thermal properties of the ribbons were evaluated by a Netzsch DSC404 calorimeter using pure indium (99.999wt %) and pure zinc (99.999wt %) standards. Besides the viscosity, thermal scanning is also thought to be a reliable way to obtain the fragility of the supercooled liquid [10]. The parameter of the fragility of the supercooled liquid, m, can be calculated by the following equation [11],
m = ΔH ∗ / RTg , 20 ,
(1)
*
where ΔH is the active enthalpy of the glass transition, R is the gas constant, Tg,20 is the glass transition temperature at a scanning rate of 20 K/min examined by the DSC. Meanwhile, ΔH* can be calculated according to the different glass transition temperatures dependent on the heating rates. The equation is as follows [12, 13],
ln(Tg2 / ) = ΔH ∗ / RTg + Const
(2)
where Φ is the scanning rates, the glass transition temperature (Tg) can be determined by the DSC curves of different heating rates. During the DSC measurement, the samples were heating under different rates of 10, 20, 40 and 50 K/min, respectively.
757
Junzhe Sun et al. / Procedia Engineering 16 (2011) 755 – 762
For viscosity measurement, the samples were measured by an oscillating viscometer in a high vacuum atmosphere. The samples were first overheated to about 250 K above their liquidus and held for 1 h and then cooled to the required temperature for the viscosity measurements. The kinetic viscosity was calculated according to the Shvidakovskii’s equation [14]. The viscosity of the superheated melts can be successfully fitted by the Arrhenius equation in a wide temperature range above liquidus, = 0 exp( E / RT ) (3) where η0 is a pre-exponential constant, which is associated with the nature of the liquid, E is the activation energy for viscous flow and R is the gas constant. According to the definition, melt fragility (M) is defined as [8]:
M=
∂η(T)/η(TL ) ∂T/TL
T = TL
=
∂ηr ∂Tr
(4)
Tr =1
where TL is the liquidus temperature, ηL is the viscosity at the liquidus. According to the equation (3) and (4), M can be calculated by
M = Ε/RTL
(5)
3. Results and discussion In order to compare with each other, all the fully amorphous ribbons of Al84Ni10(Ce3RE3) (RE = La, Pr, Nd, Gd) alloys were obtained at the same circumferential speed of 14.7 m/s. The amorphous structures were also confirmed by the XRD patterns shown in Fig.1. The broad diffraction peaks found in these alloys indicate no detectable crystalline phase in them. From Fig.1 we can find that with the different elements addition, the main peaks of the XRD patterns have a slight deviation according to the Gaussian fitted curves. Fig. 2 shows the DSC curves of the Al84Ni10(Ce3RE3) (RE = La, Pr, Nd, Gd) ribbons during the heating and cooling process at a rate of 20K/min. The characteristic temperatures and parameters, such as the reduced glass transition temperature Trg(Trg = Tg/TL) [15] and γ(γ = Tx/(TgTL)) [16] for all the glassy ribbons are shown in Table 1. In our work it is difficult to get amorphous ribbons below the lowest circumferential speed v0. The relevant parameters of the Al84Ni10Ce6 were also listed for comparison and its parameters are close to the data in Ref. [18].
Intensity(a.u.)
d: Al Ni Ce Gd 84 10 3 3 c: Al Ni Ce Nd 84 10 3 3 b: Al Ni Ce Pr 84 10 3 3 a: Al Ni Ce La 84 10 3 3
10
20
30
40
50
60
70
80
2 (degree) Fig. 1. XRD patterns of the Al84Ni10(Ce3RE3) (RE = La, Pr, Nd, Gd) alloys
758
Junzhe Sun et al. / Procedia Engineering 16 (2011) 755 – 762
Exothermic
Al84Ni10Ce3Gd3 Al84Ni10Ce3Nd3
Al84Ni10Ce3Pr3 Al84Ni10Ce3La3
400
600
800
1000
Temperatuer(K)
1200
Fig. 2. DSC curves of Al84Ni10(Ce3RE3) (RE = La, Pr, Nd, Gd) alloys during the heating and cooling processes
Table 1 Thermal parameters of Al84Ni10Ce6 and Al84Ni10(Ce3RE3) (RE = La, Pr, Nd, Gd) alloys (ΔHf is the heat of fusion, ΔGl-x is the Gibbs free energy difference, ΔHmix is the mixing enthalpy, and v0 is the lowest circumferential speed to obtain fully amorphous ribbons) Alloys
Tg (K)
Tx (K)
TL (K)
Trg
γ
v0 (m/s)
ΔHf kJ/mol
ΔGl-x kJ/mol
ΔHmix kJ/mol
Al84Ni10Ce6 Al84Ni10Ce3La3 Al84Ni10Ce3Pr3 Al84Ni10Ce3Nd3 Al84Ni10Ce3Gd3
553.7 557.2 549.0 560.0 554.7
559.2 565.6 554.9 567.0 561.0
1172.3 1094.9 1186.6 1209.9 1230.2
0.472 0.509 0.463 0.463 0.451
0.324 0.342 0.320 0.320 0.314
11 5.5[17] 14.7 14.7 14.7
6.96 6.24 7.68 7.10 8.31
3.21 2.88 3.55 3.31 3.88
-15.72 -15.71 -15.75 -15.75 -15.86
From Table 1, it can be seen that the Al84Ni10Ce3La3 has a lower value of v0 than that of Al84Ni10Ce6, indicating that the coexistence of similar elements pairs of La and Ce has improved the GFA, while the addition of Pr, Nd and Gd elements has decreased the GFA. In order to further compare with each other, we use γ to characterize the GFA of these alloys. Generally speaking, the higher γ corresponds to the better GFA [16]. In the Al84Ni10(Ce3RE3) alloy system, the γ also shows a negative correlation with the lowest circumferential speed v0, indicating that γ is reliable to characterize the GFA in our work. The minor addition of different similar elements causes the change of the number and the shape of the main exothermic peaks for the crystallization, indicating that the minor addition affects the GFA via changing the crystalline process. Then we focus on the improvement of the GFA by the similar elements substitution. As we know, the RE elements have similar atomic size, chemical and physical properties. Since the similar element pairs of La-Ce are neighbours in the Periodic Table of Elements, the difference in atomic size between them is very small and the mixing enthalpy (ΔHABmix,, here A and B represent two kinds of elements) [19] between them is 0 kJ/mol. Therefore, the traditional criterion of mismatches in atomic size and the negative heat of mixing cannot give a reasonable explanation for the improvement in GFA. Consequently we want to discuss the possible reason in thermodynamic of Gibbs free energy difference and dynamic of viscosity for this phenomenon. Thermodynamically, the Gibbs free energy difference between the liquid and the crystalline state is a very important parameter for the glass forming. As is investigated by Busch et al. [20], the lower Gibbs
Junzhe Sun et al. / Procedia Engineering 16 (2011) 755 – 762
free energy difference (ΔGl-x(T)) between the liquid and the crystalline states corresponds to the better GFA, indicating that the high driving force for crystallization decreases the GFA. The difference in the Gibbs free energy between the liquid and crystalline states is given by [21, 22]:
Δ
−
= ⎡Δ ⎢⎣
−∫ Δ m
−
84Ni10Ce6
⎤− ⎥⎦
⎡ ⎢Δ ⎢⎣
−∫
m
Δ
−
⎤ ⎥ ⎥⎦
(c) Al84Ni10Ce3Pr3 (d)Al84Ni10Ce3Nd3 (e)Al84Ni10Ce3Gd3
According to the method mentioned previously, we calculated the parameter of the fragility of the supercooled liquids (m). The values of activation enthalpy for glass transition ΔH* and m are listed in Table 2. The different glass transition temperatures at different heating rates are obtained from different heating rates of DSC curves. Fig. 4 shows the correlation between ln(Tg2/Φ) and 1000/Tg for the Al84Ni10(Ce3RE3) (RE = La, Pr, Nd, Gd) alloys. The effect of similar elements substitution on the GFA in different Al-based alloys may also attribute to the change of the liquid behavior. As shown in Table 2,
759
760
Junzhe Sun et al. / Procedia Engineering 16 (2011) 755 – 762
Al84Ni10Ce3La3 alloy with the best GFA has the lowest m, indicating it has a more stable and stronger supercooled liquid than the other alloys. However, among the alloys of Al84Ni10(Ce3RE3) (RE = Pr, Nd, Gd), the Al84Ni10Ce3Gd3 alloy has the smallest m of 85.63 corresponding to the smallest γ of 0.314. Hence, it can be verified that a larger m corresponds to a better GFA. Generally speaking, the strong liquid usually exhibits a good GFA with a small m. The possible reason for this unusual tendency in Al-based alloys is the heterogeneous nucleation in liquids resulting from the existence of local heterogeneous domains, which decreases the influence of the viscosity in the glass transition region and promotes the nucleation [26]. It can be inferred that the smaller m corresponds to a better GFA only works in an individual alloy system, not in the different systems of Al84Ni10(Ce3RE3) (RE = La, Pr, Nd, Gd) alloys. This finding accords with the result that was found in Ref. [27]. Table 2 Activation enthalpy for glass transition (ΔH*), the fragility of the supercooled liquids (m), the fitting parameters of E/R and η0, and calculated parameters of ηL and M of Al84Ni10(Ce3RE3) (RE = La, Pr, Nd, Gd) alloys Alloys
ΔH* (kJ/mol)
m
η0 (10-3Pa.s)
E/R (K)
TL (K)
ηL (10-3Pa.s)
M
Al84Ni10Ce3La3 Al84Ni10Ce3Pr3 Al84Ni10Ce3Nd3 Al84Ni10Ce3Gd3
267.18 397.61 420.28 394.89
56.82 87.11 90.27 85.63
1.830 1.400 1.079 1.976
1117 1242 1255 1291
1094.9 1186.6 1209.9 1230.2
5.075 3.985 3.044 5.647
1.020 1.046 1.037 1.050
10.4 10.2
ln(Tg2/)
10.0
(a) (b) (c) (d)
9.8 9.6 9.4 9.2 9.0 8.8 8.6 1.76 1.77 1.78 1.79 1.80 1.81 1.82 1.83 1.84
1000/Tg(K-1) Fig. 4. Correlation between ln(Tg2/Φ) and 1000/Tg of (a) Al84Ni10Ce3La3 (b)Al84Ni10Ce3Pr3 (c) Al84Ni10Ce3Nd3 (d)Al84Ni10Ce3Gd3 alloys
On the other hand, the viscosity of the liquid is an important dynamics parameter to measure the process of the glass forming and it has a close relationship with the microstructure of the liquid. Fig. 5 shows the experimental viscosity data of Al84Ni10(Ce3RE3) (RE = La, Pr, Nd, Gd) alloy melts at high temperature during the cooling process. The discrete viscosity values are fitted by the Arrhenius law. Thermodynamic properties, especially the important one of mixing heat, which reflect the interaction between the components in the melts are considered to be useful in discussing the behaviors of the viscosity [28]. It can be inferred a stronger atoms interactions in the Al84Ni10(Ce3Gd3) alloy with the maximal viscosity, which corresponds to the largest mixing enthalpy ΔHmix of it. The values of the fitting
Junzhe Sun et al. / Procedia Engineering 16 (2011) 755 – 762
parameters according to equation (3) are listed in Table 2. Generally speaking, the lower M corresponds to the higher GFA, suggesting that the viscosity variation rate of superheated liquids towards the liquidus is slower and the liquid is more stable. In our work, it can be seen that the M has a good negative relationship with the GFA among these alloys from Table 2. The coexistence of La and Ce hinders the precipitation of the intermetallic compound, thus in the view of this improved the GFA.
6.0
Viscosity(10-3Pa*s)
5.5 5.0
(a) (b) (c) (d)
4.5 4.0 3.5 3.0 2.5 1150 1200 1250 1300 1350 1400 1450 1500 1550
Temperature(K) Fig. 5. The experimental viscosity data and fitting curves of (a) Al84Ni10Ce3La3 (b)Al84Ni10Ce3Pr3 (c) Al84Ni10Ce3Nd3 (d)Al84Ni10Ce3Gd3 alloys
4. Conclusions In this paper, the coexistence of similar element pairs of La-Ce improved the GFA, while the Pr, Nd and Gd addition decreased the GFA in Al-Ni-Ce alloy system. Among the Al84Ni10(Ce3RE3) (RE = La, Pr, Nd, Gd) alloys, the Gibbs free energy difference between the liquid and the crystalline states (ΔGl-x) has a negative relationship with the GFA according to the lowest circumferential speed v0 and γ. The decrease in ΔGl-x and melt fragility can give a reasonable explanation for the improvement in GFA by the minor substitution of similar elements. Due to the different formation mechanism, it can be verified that the fragility parameter of m could be generally confined to an individual alloy system. Acknowledgements This work was supported by National Basic Research Program of China 973 Program, Grant No. 2007CB613901 and National Natural Science Foundation of China (Nos.50831003, 50871062) and the Natural Science Foundation of Shandong province (Grant No. Z2008F08). References [1] Inoue A, Kimura H. Fabrications and mechanical properties of bulk amorphous, nanocrystalline, nanoquasicrystalline alloys in aluminum-based system. J. Light Met 2001; 1: 31-41. [2] Yang BJ, Yao JH, Zhang J, Yang HW, Wang JQ, Ma E. Al-rich bulk metallic glasses with plasticity and ultrahigh specific strength. Scripta Mater. 2009; 61: 423-426. [3] Chang CT, Shen BL, Inoue A. Co–Fe–B–Si–Nb bulk glassy alloys with superhigh strength and extremely low magnetostriction. Appl. Phys. Lett. 2006; 88: 011901.
761
762
Junzhe Sun et al. / Procedia Engineering 16 (2011) 755 – 762
[4] Shen BL, Chang CT, Inoue A. Ni-based bulk glassy alloys with superhigh strength of 3800 MPa in Ni–Fe–B–Si–Nb system. Appl. Phys. Lett. 2006; 88: 201903. [5] Tan H, Zhang Y, Ma D, Feng YP, Li Y. Optimum glass formation at off-eutectic composition and its relation to skewed eutectic coupled zone in the La based La–Al–(Cu,Ni) pseudo ternary system. Acta Mater. 2003; 51: 4551. [6] Zhang T, Li R, Pang SJ. Effect of similar elements on improving glass-forming ability of La–Ce-based alloys. J. Alloys Compd. 2009; 483: 60. [7] Angell CA. Spectroscopy simulation and scattering, and the medium range order problem in glass. J. Non-Cryst. Solids 1985; 73: 1-17. [8] Bian XF, Sun BA, Hu LN, Jia YB. Fragility of superheated melts and glass-forming ability in Al-based alloys. Phys. Lett. A 2005; 335: 61-67. [9] Inoue A. Amorphous, nanoquasicrystalline and nanocrystalline alloys in Al-based systems. Prog. Mater. Sci. 1998; 43: 365-520. [10] Crowley KJ, Zografi G. The use of thermal methods for predicting glass-former fragility. Thermochim. Acta 2001; 380: 79-93. [11] Brüning R, Mark S. Fragility of glass-forming systems and the width of the glass transition. J. Non-Cryst. Solids 1996; 205-207: 480-484. [12] Moynihan CT. Nonexponential relaxation in strong and fragile glass formers. Am. Ceram. Soc. 1993; 76: 1081-1087. [13] Kasap SO, Yannacopoulos S. Kinetics of structural relaxations in the glassy semiconductor a-Se. Mater. Res. 1989; 4: 893-905. [14] Schvidkovskii EG, Gostekhteorizdat, USSR, Science Press, 1955. [15] Lu ZP, Tan H, Ng SC, Li Y. The correlation between reduced glass transition temperature and glass forming ability of bulk metallic glasses. Scripta Mater. 2000; 42: 667-673. [16] Lu ZP, Liu CT. Glass Formation Criterion for Various Glass-Forming Systems. Phys. Rev. Lett. 2003, 91, 115505-1. [17] Si PC, Bian XF, Zhang JY, Li H, Sun MH, Zhao Y. The fragility of Al–Ni-based glass-forming melts. J. Phys.: Condens. Matter 2003; 15: 5409. [18] Hu LN, Bian XF, Wang WM, Zhang JY, Jia YB. Liquid fragility and characteristic of the structure corresponding to the prepeak of AlNiCe amorphous alloys. Acta Mater. 2004; 52: 4773-4781. [19] Boer FR, Boom R, Mattens WCM, Miedema AR, Niessen AK. Cohesion in metals. Amsterdam, North-Holland, 1989. [20] Bakke, Eric, Dissertation, Viscosity measurement of bulk metallic glass forming alloys Dissertation. California Institute of Technology, Pasadena, CA, 1997. [21] Glade SC, Busch R, Lee DS, Johnson WL. Thermodynamics of Cu47Ti34Zr11Ni8, Zr52.5Cu17.9Ni14.6Al10Ti5 and Zr57Cu15.4Ni12.6Al10Nb5 bulk metallic glass forming alloys. J. Appl. Phys. 2000; 87: 7242. [22] Lu ZP, Hu X, Li Y. Thermodynamics of La based La-Al-Cu-Ni-Co alloys studied by temperature modulated DSC. Intermetallics 2000 ; 8: 477-480. [23] Thompson CV, Spaepen F. On the approximation of the free energy change on crystallization. Acta Metall. 1979; 27: 18551859. [24] Fecht HJ, Free energy functions of undercooled liquid glass-forming Au-Pb-Sb alloys. Mater.Sci. Eng. A 1991; 133: 443. [25] Zhang Y, Tan H, Kong HZ, Yao B, Li Y. Glass-forming ability of Pr-(Cu,Ni)-Al alloys in eutectic. J. Mater. Res. 2003; 18: 664. [26] Gerhard W. Thermodynamics, viscous flow and relaxation dynamics of bulk glass-forming Pd alloys. Ann. Chim. Sci. Mat. 2002; 27: 49. [27] Song KK, Bian XF, Lv XQ, Xie MT, Jia R. The correlations among the fragility of supercooled liquids, the fragility of superheated melts, and the glass-forming ability for marginal metallic glasses. J. Appl. Phys. 2009; 105: 024304. [28] Sato Y, Sugisawa K, Aoki D, Yamamura T. Viscosities of Fe–Ni, Fe–Co and Ni–Co binary melts. Meas. Sci. Technol. 2005; 16: 363.