Reduced glass transition temperature and glass forming ability of bulk glass forming alloys

Reduced glass transition temperature and glass forming ability of bulk glass forming alloys

Journal of Non-Crystalline Solids 270 (2000) 103±114 www.elsevier.com/locate/jnoncrysol Reduced glass transition temperature and glass forming abili...

280KB Sizes 0 Downloads 34 Views

Journal of Non-Crystalline Solids 270 (2000) 103±114

www.elsevier.com/locate/jnoncrysol

Reduced glass transition temperature and glass forming ability of bulk glass forming alloys Z.P. Lu a, Y. Li a,*, S.C. Ng b a

Department of Materials Science, Faculty of Science, National University of Singapore, BLK S1A, Lower Kent Ridge Road, 119260 Singapore b Department of Physics, National University of Singapore, Singapore Received 9 September 1999; received in revised form 5 November 1999

Abstract Onset temperature (solidus) Tm and o€set temperature (liquidus) Tl of melting of a series of bulk glass forming alloys based on Zr, La, Mg, Pd and rare-earth elements have been measured by studying systematically the melting behaviour of these alloys using DTA or DSC. Bulk metallic glass formation has been found to be most e€ective at or near their eutectic points and less e€ective for o€-eutectic alloys. Reduced glass transition temperature Trg given by Tg /Tl is found to show a stronger correlation with critical cooling rate or critical section thickness for glass formation than Trg given by Tg /Tm . Ó 2000 Elsevier Science B.V. All rights reserved.

1. Introduction Since the late 1980s, bulk glass formation has been reported in various new alloy systems [1]. For example, 9 and 7 mm diameter glassy rods were obtained in La55 Al25 Ni10 Cu5 Co5 and Mg65 Cu25 Y10 alloys, respectively, by high pressure die casting [2,3]. Glass formation in diameters up to 30 mm in diameter has been obtained by suction casting in a Zr55 Al10 Ni5 Cu30 alloy [4], while a Be containing Zr-based alloy was reported to form glass in rod at least 14 mm by water quenching [5]. The latest report was a glassy rod 72 mm in diameter in Pd40 Ni20 Cu10 P30 alloy by water quenching after ¯uxing [6]. Bulk glass formation in these alloys

* Corresponding author. Tel.: +65 7763 604; fax: +65 7763 604. E-mail address: [email protected] (Y. Li).

indicates that they have high glass forming ability (GFA). GFA can be represented by many parameters [7]. The reduced glass transition temperature, Trg de®ned below is one of the widely used indicators of GFA of alloys: Trg ˆ

Tg ; Tl

…1†

where Tg is the glass transition temperature and Tl is the liquidus temperature. As the alloy concentration increases, Tg generally has a weak dependence on composition and while Tl often decreases more strongly. Thus, the interval between Tl and Tg generally decreases and the value of Trg increases with increasing alloying concentration so that the probability of being able to cool through the interval between Tl and Tg without crystallization is enhanced, i.e., the GFA is increased [8]. The ratio Tg /Tl also arises from the requirement

0022-3093/00/$ - see front matter Ó 2000 Elsevier Science B.V. All rights reserved. PII: S 0 0 2 2 - 3 0 9 3 ( 0 0 ) 0 0 0 6 4 - 8

104

Z.P. Lu et al. / Journal of Non-Crystalline Solids 270 (2000) 103±114

that the viscosity must be large at temperatures between the melting point and the glass transition temperature, as noted by Uhlmann and Yinnon [9]. The viscosity at Tg being constant, the higher the ratio Tg /Tl , the higher will be the viscosity at the nose of the TTT or CCT curves and hence the smaller will be the critical cooling rate for glass formation Rc . Accordingly, alloy systems for which the GFA is higher, are those with a deep eutectic or low-lying Tl compared with the melting points of the host metals, thus leading to high Trg [8,10]. From this analysis it can be seen that it is the liquidus temperature Tl that is important, not

the onset melting point Tm or the eutectic point Te , since these latter usually do not vary very much near a eutectic composition as is also the case for Tg . There are many reported values of Trg in the literature [7], but unfortunately most of them were calculated using Tg /Tm with minimal report of Tg /Tl [11,12]. In our recent report [7], it was found that the values of Trg based on Tg /Tm do not re¯ect the di€erences in the GFA among some bulk metallic forming alloys which have a range of critical cooling rates for glass forming. Therefore, they do not correlate well with GFA. In contrast, we have

Table 1 The glass transition temperature (Tg ), crystallization temperature (Tx ) obtained by DSC and onset melting temperature (Tm ) and o€set melting temperature (Tl ) obtained by DTA at heating rate of 20 K/min in Mg, Zr, La, Pd and rare-earth-based amorphous alloys (tolerance: <1 K) Alloys

Tg (K)

Tx (K)

Tm (K)

Tl (K)

Structures

Mg80 Ni10 Nd10 Mg75 Ni15 Nd10 Mg70 Ni15 Nd15 Mg65 Ni20 Nd15 Mg77 Ni18 Nd5 Mg90 Ni5 Nd5 Mg65 Cu25 Y10 Zr66 Al8 Ni26 Zr66 Al8 Cu7 Ni19 Zr66 Al8 Cu12 Ni14 Zr66 Al9 Cu16 Ni9 Zr65 Al7:5 Cu17:5 Ni10 Zr57 Ti5 Al10 Cu20 Ni8 Ti34 Zr11 Cu47 Ni8 La55 Al25 Ni20 La55 Al25 Ni15 Cu5 La55 Al25 Ni10 Cu10 La55 Al25 Ni5 Cu15 La55 Al25 Cu20 La55 Al25 Ni5 Cu10 Co5 Pd40 Cu30 Ni10 P20 Pd81:5 Cu2 Si16:5 Pd79:5 Cu4 Si16:5 Pd77:5 Cu6 Si16:5 Pd77 Cu6 Si17 Pd73:5 Cu10 Si16:5 Pd71:5 Cu12 Si16:5 Pd64:5 Cu19 Si16:5 Pd56:5 Cu27 Si16:5 Nd60 Fe30 Al10 Pr60 Fe30 Al10 Sm60 Fe30 Al10 Y60 Fe30 Al10

454.2 450.0 467.1 459.3 429.4 426.2 424.5 672.0 662.3 655.1 657.2 656.5 676.7 698.4 490.8 473.6 467.4 459.1 455.9 465.2 576.9 633.0 635.0 637.0 642.4 645.0 652.0 640.0 ± 591.0 575.0 593.0 572.0

470.5 470.4 489.4 501.4 437.2 449.0 479.4 707.6 720.7 732.5 736.7 735.6 720.0 727.2 555.1 541.2 547.2 520.0 494.8 541.8 655.8 670.0 675.0 678.0 686.4 685.0 680.0 640.0 ± 722.0 729.2 733.8 586.5

725.8 717.0 742.5 743.0 723.4 725.9 727.9 1188.5 1117.3 1109.1 1110.9 1108.6 1095.3 1119.0 711.6 659.7 662.1 663.4 672.1 660.9 741.5 1008.8 1019.3 1019.4 1019.7 1019.3 1019.6 1167.1 1167.3 929.3 873.4 905.2 1075.9

878.0 789.8 844.3 804.9 886.9 918.8 770.9 1251.0 1200.8 1172.1 1170.6 1167.6 1145.2 1169.2 941.3 899.6 835.0 878.1 896.1 822.5 836.0 1097.3 1086.0 1058.1 1128.4 1135.9 1153.6 1234.3 1248.4 958.0 950.1 946.6 1225.4

O€-eutectic Near-eutectic O€-eutectic Near-eutectic O€-eutectic O€-eutectic Eutectic O€-eutectic O€-eutectic O€-eutectic O€-eutectic Near-eutectic Near-eutectic Near-eutectic O€-eutectic O€-eutectic O€-eutectic O€-eutectic O€-eutectic O€-eutectic Near-eutectic O€-eutectic O€-eutectic Near-eutectic Near-eutectic O€-eutectic O€-eutectic Near-eutectic O€-eutectic Eutectic O€-eutectic O€-eutectic O€-eutectic

[18,19] [19] [19] [19]

Z.P. Lu et al. / Journal of Non-Crystalline Solids 270 (2000) 103±114

Fig. 1. Melting curves of three binary Al±Cu alloys obtained by DTA under a constant heating rate of 20 K/min, the insert indicates the de®nition of o€set melting point.

Fig. 2. Melting curves of Mg-based glass forming alloys obtained by DTA under a constant heating rate of 20 K/min.

found that Trg based on Tg /Tl has a strong correlation with the GFA in bulk glass forming Labased La55 Al25 (CuNi)20 alloys [13]. In the present

105

Fig. 3. Melting curves of six La-based bulk glass forming alloys obtained by DTA under a constant heating rate of 20 K/min.

Fig. 4. Melting curves of Zr-based glass forming alloys obtained by DTA under a constant heating rate of 20 K/min.

work, we present our systematic determination of reduced glass transition temperature Trg of bulk metallic glasses based on Zr, La, Mg, Pd and rare-

106

Z.P. Lu et al. / Journal of Non-Crystalline Solids 270 (2000) 103±114

cooling rate for glass formation and GFA of the bulk glass forming alloys is discussed.

2. Experimental

earth elements through the study of the melting behaviour of these alloys using DTA or DSC. The signi®cance of Trg based on Tg /Tm or Tg /Tl and the correlation between resulting Trg and the critical

All the alloys were produced either by arcmelting or by induction melting mixtures of constituent pure elements under an argon atmosphere. Their compositions are indicated in Table 1. Pieces of resulting ingots were sealed in quartz tubings with external diameter of 3 mm to minimize the oxidation during the DTA melting process. These tubes were evacuated and then back-®lled with argon. The onset melting temperature (solidus) Tm and the o€set melting temperature Tl of these alloys were obtained through the melting behaviour study carried out on a DTA at a heating rate of 20 K/min. The o€set melting temperature (Tl ) is determined as the intersection of the two tangents of the last part of the melting peak, as shown in the insert of Fig. 1. The calibration of temperature was carried out using Zn and Ag. It was noticed that the Y-based alloy reacted strongly with quartz tubing so Al2 O3 tubing was used instead for this particular compo-

Fig. 6. Melting curves of rare-earth-based alloys obtained by DTA under a constant heating rate of 20 K/min.

Fig. 7. DSC curves for several typical bulk amorphous alloys, the insert shows the determination of glass transition temperature.

Fig. 5. Melting curves of Pd-based glass forming alloys obtained by DTA under a constant heating rate of 20 K/min.

Z.P. Lu et al. / Journal of Non-Crystalline Solids 270 (2000) 103±114

107

Table 2 Summary of Tg /Tm , Tg /Tl , critical section thickness (Zmax ) and critical cooling rate (Rc ) for glass formation for Mg, Zr, La, Pd and rare-earth-based amorphous alloys

a

Tl ) Tm (K)

Zmax (mm) [Refs.]

Rc (K/s) [Refs.]

Methodsa [Refs.]

16.3 20.4 22.3 42.1 7.8 22.8 54.9

152.2 72.8 101.8 61.9 163.5 192.9 43.0

1251.4 [24] 46.1 [24] 178.2 [24] 30.0 [25] 4.9 ´ 104 [24] 5.3 ´ 104 [24] 50 [15]

A [24] A [24] A [24] A [24] A [24] A [24] C [26] D [3]

0.537 0.552 0.559 0.561 0.566 0.591 0.597

35.6 58.4 77.4 79.5 79.1 43.3 28.8

62.5 83.6 88.0 59.7 34.9 49.9 50.2

0.6 [24] 2.8 [24] 1.5 [24] 3.5 [24] 6 0.1 [24] 6 0.1 [24] 4.0 [26] 7.0 [3]

0.521 0.526 0.560 0.523 0.509 0.566

64.3 67.6 79.8 60.9 38.9 76.6

229.7 239.9 172.9 214.6 223.9 161.7

0.76 [35]

0.690

78.9

94.5

0.627 0.623 0.625 0.630 0.633 0.639 0.548 ±

± ± 0.64 [8] ± ± ± ± ±

0.577 0.585 0.602 0.569 0.568 0.565 0.519 ±

37 40 41.0 44.0 40.0 28.0 0.0 ±

88.5 66.7 69.8 108.7 116.6 134.0 67.0 81.1

0.636 0.658 0.655 0.532

0.85 [40] 0.90 [41] ± ±

0.617 0.605 0.626 0.467

131.0 154.2 140.8 14.5

28.7 76.7 41.4 149.5

Alloys

Tg /Tm

Tg /Tm [Refs.]

Tg /Tl

Tx ) Tg (K)

Mg80 Ni10 Nd10 Mg75 Ni15 Nd10 Mg70 Ni15 Nd15 Mg65 Ni20 Nd15 Mg77 Ni18 Nd5 Mg90 Ni5 Nd5 Mg65 Cu25 Y10

0.626 0.628 0.629 0.618 0.594 0.587 0.583

0.64 0.63 0.65 0.63 0.64 0.62 0.58

[24] [24] [24] [24] [24] [24] [26]

0.517 0.570 0.553 0.571 0.484 0.464 0.551

Zr66 Al8 Ni26 Zr66 Al8 Cu7 Ni19 Zr66 Al8 Cu12 Ni14 Zr66 Al9 Cu16 Ni9 Zr65 Al7:5 Cu17:5 Ni10 Zr57 Ti5 Al10 Cu20 Ni8 Ti34 Zr11 Cu47 Ni8

0.565 0.593 0.591 0.592 0.583 0.618 0.624

0.56 0.58 0.58 0.59 0.58 0.63 0.61

[27] [27] [27] [27] [28] [31] [32]

La55 Al25 Ni20 La55 Al25 Ni15 Cu5 La55 Al25 Ni10 Cu10 La55 Al25 Ni5 Cu15 La55 Al25 Cu20 La55 Al25 Ni5 Cu10 Co5 Pd40 Cu30 Ni10 P20

0.690 0.718 0.706 0.692 0.678 0.704

0.71 ± 0.70 ± 0.68 0.70

[2]

0.778

Pd81:5 Cu2 Si16:5 Pd79:5 Cu4 Si16:5 Pd77:5 Cu6 Si16:5 Pd77 Cu6 Si17 Pd73:5 Cu10 Si16:5 Pd71:5 Cu12 Si16:5 Pd64:5 Cu19 Si16:5 Pd56:5 Cu27 Si16:5 Nd60 Fe30 Al10 Pr60 Fe30 Al10 Sm60 Fe30 Al10 Y60 Fe30 Al10

[2] [2] [2]

16 [17] 20.0 [30] 4±5 [22,32] 3 [2] 5 [2] 3 [2] >9 [2] 72.0 [36] 40.0 [37] 1±3 [38,39] 0.75 [20] 1.5 [21] 2 [20] 1±3 [38,39] 1±3 [38,39] <0.1 [38,39] <0.001 [38,39] 15.0 [40] 3.0 [41]

66.6 [27] 22.7 [27] 9.8 [27] 4.1 [6] 1.5 [17] 10.0 [31] 100.0 [22] <250 [32] 67.5 [13,33] 34.5 [13] 22.5 [13] 35.9 [13] 72.3 [13] 18.8 [13,34] 0.10 [36] 1.57 [37] 500 [20] 100 [10] 125.0 [29]

>106 12.0 [40]

B [17] C [30] C [22] C [2] C [2] C [2] C [2] B [36,37] B [38,39] B [38,39] B [38,39] B [38,39] B [38,39] E [38,39] D [40] D [41]

A ± wedge casting; B ± water quenching; C ± mould casting; D ± suction casting; E ± melt spinning.

sition. The melting behaviour of some low melting alloys was also determined using a conventional DSC at a heating rate of 20 K/min using Al pans. From two to ®ve runs were carried out to obtain an average melting temperature value. Glass formation was obtained by melt-spinning using a single roller melt-spinner. The corresponding glass transition temperature Tg and crystallization temperature Tx were measured with a DSC at the same

heating rate of 20 K/min used for the melting studies. The glass transition temperature is de®ned in our study as the in¯ection point of the glass transition temperature (as demonstrated by the insert in Fig. 7). The maximum tolerance in the temperature measurement is 1 K and the error in the value of Trg given by Tg /Tm and Tg /Tl is mainly due to the error in the measurement of temperature and is within the range of 0.003.

108

Z.P. Lu et al. / Journal of Non-Crystalline Solids 270 (2000) 103±114

3. Results Fig. 1 shows that the melting curve of eutectic Al±17.6 at.% Cu alloy obtained by DTA under a constant heating rate of 20 K/min, has only a single sharp melting peak, while the melting curves of the two o€-eutectic Al±5 at.% Cu and Al±25 at.% Cu alloys have two melting peaks. The onset melting points Tm for the three alloys are 819.7, 821.5 and 822.5 K, respectively, while the liquidus temperatures Tl for the two o€-eutectic Al±5 at.% Cu and 10 at.% Cu alloys are 924.7 and 879.9 K, respectively. Fig. 2 shows the melting curves for Mg-based alloys among which Mg65 Cu25 Y10 , Mg75 Ni15 Nd10 and Mg65 Ni20 Nd15 alloys have only one melting peak indicating that these three alloys are very near to their eutectic compositions, respectively, while other alloys show multiple melting peaks with large melting interval indicating that they are o€-eutectic. Fig. 3 shows the melting curves of six La-based alloys exhibiting a sharp initial melting, followed by a wide melting interval with multiple melting peaks. No melting curves with single melting peak were observed for these six alloys indicating that they are not at eutectic points. However, the near invariance of Tm for La55 Al25 Ni15 Cu5 , La55 Al25 Ni10 Cu10 and La55 Al25 Cu20 alloys shows that there could be a eutectic reaction at a temperature of about 665 K for the three alloys. Fig. 4 shows that the Zr65 Al7:5 Cu17:5 Ni10 , Zr57 Ti5 Al10 Cu20 Ni8 and Zr11 Ti34 Cu47 Ni8 alloys exhibit one strong single melting peak with a weak melting peak, while other Zr-based alloys included certainly are not at eutectic points and they have large melting intervals. Furthermore, the near invariance of the onset melting point for the three quaternary Zr65 Al7:5 Cu17:5 Ni10 , Zr66 Al8 Cu7 Ni19 and Zr66 Al8 Cu12 Ni14 alloys indicates that they may be very near the eutectic temperature at 1108 K. Fig. 5 shows the melting behaviours of a series of Pd-based Pd±Si±Cu alloys with Si content ®xed at around 16.5 at.%. When the copper content increased from 2 to 6 at.%, the intensity of the second melting peak decreased. While the copper content increased further from 6 to 12 at.%, the intensity of the second melting peak increased again. This indicates that the Pd77:5 Cu6 Si16:5 alloy

is very near to a eutectic point. The near invariance of the onset melting points of these alloys at around 1019 K further pointed out that the eutectic temperature could be at this temperature. It is also noted that the melting temperature for Pd40 Cu10 Ni10 P40 is the lowest among the Pd-based alloys studied. Fig. 6 shows the melting behaviours of a few rare-earth-based alloys indicating that Nd60 Fe30 Al10 alloy could be around the eutectic point and other alloys are o€-eutectic with Y60 Fe30 Al10 alloy having highest liquidus of 1225 K. Representative DSC curves for a few bulk metallic glasses are shown in Fig. 7. All the values of onset melting point, Tm , o€set melting point Tl and the type of the reaction for the alloys studied are summarised in Table 1 together with their corresponding glass transition temperature Tg de®ned by the in¯ection point, crystallization temperature Tx . The reduced glass transition temperature given by Tg /Tl or Tg /Tm for all the present alloys together with the values of Tg /Tm for the same alloy from the literature are given in Table 2.

4. Discussion 4.1. Temperature measurement Our melting results for Al±Cu alloys indicated that the onset of melting points are very close to the previously published eutectic temperature 821.2 K [14], while the liquidus temperatures of 924.7 and 879.9 K for the two o€-eutectic Al±5 at.% and Al±25 at.% alloys, respectively, however, are about 20 K higher than the previously published liquidus temperatures of about 905 and 861 K, respectively [14]. These results show that it is convenient to use DTA to determine the onset melting temperature (eutectic point) and to detect the events of the transformation such as the melting of the eutectic or the primary phases during melting. On the other hand, the liquidus temperature obtained can be up to 20 K higher than the true temperature value under the heating rate of 20 K/min.

Z.P. Lu et al. / Journal of Non-Crystalline Solids 270 (2000) 103±114

109

Fig. 9. Glass transition temperature, Tg , onset and o€set temperatures of melting Tm and Tl and Trg based on Tg /Tl as well as critical cooling rate Rc for glass formation as a function of Cu content in six La-based bulk amorphous alloys.

Fig. 8. Critical cooling rate Rc as a function of Tg /Tm or Tg /Tl for (a) La-based, (b) Mg-based and (c) Zr-based alloys.

Our result of a single sharp melting peak for Mg65 Cu25 Y10 alloy is consistent with the observation of a single melting peak for the same alloy reported by Busch et al. [15]. They also concluded that this alloy is at a ternary eutectic point. The onset melting point of 729.9 K for Mg65 Cu25 Y10 is very close to 730 K for the same alloy obtained using DSC at a heating rate of 20 K/min reported by Busch et al. [15], while the onset melting points of 717.0 and 743.0 K for Mg75 Ni15 Nd10 and Mg65 Ni20 Nd15 alloys are close to 719 and 738 K of the corresponding alloys, respectively, reported by Li et al. [16] using DTA. The onset melting temperature of 1108.6 K for the Zr65 Al7:5 Cu17:5 Ni10 alloy is in agreement with that of 1107 K reported previously [17]. From Table 2, it can be seen that many of Trg given by Tg /Tm in the present study are very close to the ones reported in the literature,

indicating that our measurement of Tg and Tm are fairly consistent with those reported in the literature. 4.2. Eutectic composition and bulk glass formation It seems, from the present study of the melting behaviour that the best bulk glass forming alloys are those having eutectic composition or being very near to a eutectic composition. For example, Mg65 Cu20 Y10 alloy has the largest glass forming diameter of 7 mm among six Mg90 ÿ x Cux Y10 (x ˆ 5 to 30) alloys [3]. Our result as well as that of Busch et al. [15] shows that this alloy is at a ternary eutectic point. Similarly, each of the two best glass forming alloys, Mg65 Ni20 Nd15 and Mg70 Ni20 Nd10 with critical section thickness of 3.5 and 2 mm and critical cooling rates for glass formation of 50 and 400 K/s, respectively, among eight Mg±Ni±Nd glass forming alloys [16], is around a eutectic point, as shown by our present melting study (Fig. 2). It

110

Z.P. Lu et al. / Journal of Non-Crystalline Solids 270 (2000) 103±114

Fig. 11. Critical section thickness for glass forming as a function of Tg /Tm or Tg /Tl showing that generally the critical size increases as the value of Tg /Tm or Tg /Tl increases, but the later showed a narrower scatting. (1) Mg80 Ni10 Nd10 , (2) Mg75 Ni15 Nd10 , (3) Mg70 Ni15 Nd15 , (4) Mg65 Ni20 Nd15 , (5) Mg65 Cu25 Y10 , (6) Zr65 Al7:5 Cu17:5 Ni10 , (7) Zr57 Ti5 Al10 Cu20 Ni8 , (8) Ti34 Zr11 Cu47 Ni8 , (9) La55 Al25 Ni20 , (10) La55 Al25 Ni10 Cu10 , (11) La55 Al25 Cu20 , (12) La55 Al25 Ni5 Cu10 Co5 , (13) Pd40 Cu30 Ni10 P20 , (14) Pd77:5 Cu6 Si16:5 , (15) Nd60 Fe30 Al10 , (16) Pr60 Fe30 Al10 , (17) Pd40 Ni40 P20 , (18) Nd60 Al15 Ni10 Cu10 Fe5 , (19) Nd61 Al11 Ni8 Co5 Cu15 , (20) Zr41:2 Ti13:8 Cu12:5 Ni10 Be22:5 , (21) SiO2 .

Fig. 10. Critical cooling rate as a function of Tm /Tm or Tg /Tl for the data collected in Tables 2 and 3 showing that the later has a narrower scattering than the former.

is to be noticed that the alloy Mg90 Ni5 Nd5 has the largest melting interval with the poorest GFA among these eight alloys. Similarly in the Pd-based alloys, bulk glass formation has been reported in Pd77 Cu4 Si18 and Pd77 Cu6 Si17 alloys with critical thickness of 0.75 and 2 mm and critical cooling rates for glass formation of 500 and 125 K/s, respectively [20]. The formation of amorphous spheres 1.5 mm in diameter has been reported for Pd77 .5 Cu6 Si16:5 alloy [21]. Both Pd77 Cu6 Si17 and Pd77 .5 Cu6 Si16:5 alloys

showed very small melting interval in our study with high Trg values, while Pd79:5 Cu4 Si16:5 showed relatively large melting interval with higher critical cooling rate for glass formation of 500 K/s. The rest of Pd-based alloys are o€-eutectic with large melting interval. Although our melting study shows that the Zrbased alloys may not be exactly on a eutectic point, they are close to one. Bulk metallic glass formation has been reported in Zr57 Ti5 Al10 Cu20 Ni8 and Zr11 Ti34 Cu47 Ni8 alloys [22] which show a small melting interval (Fig. 4). The Nd60 Fe30 Al10 alloy was reported to form bulk glass rods of 3 and 12 mm in diameter by chill casting and suction casting, respectively [23], while the Pr60 Fe30 Al10 alloy was only reported to form bulk

Z.P. Lu et al. / Journal of Non-Crystalline Solids 270 (2000) 103±114

111

Table 3 Summary of Tg , Tx , Tm and Tl for Ni and other metallic and inorganic glasses Alloys

Tg (K) [Refs.]

Tx (K) [Refs.]

Tm (K) [Refs.]

Tl (K) [Refs.]

Ni Fe83 B17 Fe91 B9 Zr65 Be35 Ti63 Be37 Pd95 Si5 Pd82 Si18 Pd75 Si25 Pd40 Ni40 P20 Au77:8 Si8:4 Ge13:8 Nd60 Al15 Ni10 Cu10 Fe5 Nd61 Al11 Ni8 Co5 Cu15 Zr41:2 Ti13:8 Cu12:5 Ni10 Be22:5 SiO2 BS2 (BaOá2SiO2 ) ZBLAN (Zr53 Ba20 La4 Al3 Na20 )

425 [43] ± 600 [8,43] 623 [44] 673 [44] 647 [47] 648 [47] 656 [47] 580.0 [36] 293 [43] 430 [49] 445 [49] 638.0 [50] 1448 [51] 1002 [53] 533 [55]

425 [43] ± 600 [8,43] ± ± 647 ± 656 [42] 643.0 [36] 293 [43] 475 [49] 469 [49] 692.0 [50] ± 1137 [53,54] 625 [55]

1725 [43] 1447 [42] 1447 [42] 1238 [44] 1303 [44] 1094 [48] 1071 [8,48] 1096 [48] 855.0 [12,36] 605 [11] 709 [49] 729 [49] 937.0 [50] 1996 [52] 1605 [53] 715 [55]

1725 [43] 1447 [42] 1628 [8,42,43] 1238 [45] 1353 [46] 1688 [48] 1071 [8,48] 1343 [48] 991.0 [12,36] 629 [43] 779 [49] 744 [49] 993.0 [50] 1996 1605 [53,54] 737.3 [55]

glass of 3 mm in diameter by chill casting. This is consistent with the result of our melting study that the Nd60 Fe30 Al10 alloy has one a single melting peak, while the Pr60 Fe30 Al10 alloy has two melting peaks. To date no bulk glass formation has been reported for Y60 Fe30 Al10 and Sm60 Fe30 Al10 alloys, which have a large melting interval indicating that they are o€-eutectic alloys. It is quite clear that though bulk glass formation has been reported in the six La-based alloys, none of them are near a eutectic composition according to our melting study. However, the two most e€ective glass forming La55 Al25 Ni10 Cu10 and La55 Al25 Ni5 Cu10 Co5 alloys with 5 and 9 mm diameter glassy rods, respectively, were found to be the closest to a eutectic point among these six alloys. Furthermore, this could be an indication that even larger thickness of bulk glass may be obtained in the nearby eutectic alloy, which has a Trg value of 0.71. All these results seem to point out that the best bulk glass forming alloys in the present study are indeed at or near a eutectic point. This is consistent with the fact an alloy with a eutectic composition will have the highest Trg among the alloys around the eutectic point, particularly if it is a deep eutectic point with steeply converging of liquidus temperatures.

Structures

Eutectic O€-eutectic Eutectic O€-eutectic O€-eutectic Eutectic O€-eutectic O€-eutectic Eutectic O€-eutectic Near-eutectic Near-eutectic Congruent melting Congruent melting

4.3. Correlation between GFA and Trg Fig. 8(a)±(c) shows the critical cooling rates for glass formation as a function of Trg based on Tg /Tl or Tg /Tm for La, Mg and Zr-based alloys, respectively. It is clear from these ®gures that Trg based on Tg /Tl increases continuously with continuous decrease of critical cooling rate for glass formation, while the Trg based on Tg /Tm does not show such a trend. For example, the value of Tg /Tl for Mg±Ni±Nd alloys increased from 0.464 to 0.571 continuously as the critical cooling rate for glass formation decreased continuously from 5:2  104 to 30 K/s, while the value of Tg /Tm increased continuously from 0.587 to 0.629 as the critical cooling rate for glass formation decreased from 5:2  104 to the minimum value of 30 K/s and then increased to 1.3 ´ 103 K/s. Similarly, the values of Tg /Tl for La±Al±Ni±Cu±(Co) and Zr±Al±Ni±Cubased alloys increased from 0.509 to 0.566 and from 0.537 to 0.566 continuously as the critical cooling rate for glass formation decreased continuously from 72.3 to 18.8 K/s and from 66.6 to 1.5 K/s, respectively, while the values of Tg /Tm increased continuously from 0.678 to 0.718 and from 0.565 to 0.593 as the critical cooling rate for glass forming decreased from 72.3 to the minimum value of 18.8 K/s and from 66.6 to the minimum

112

Z.P. Lu et al. / Journal of Non-Crystalline Solids 270 (2000) 103±114

Table 4 Summary of Tg /Tm , Tg /Tl , critical section thickness (Zmax ) and critical cooling rate for glass forming (Rc ) for Ni and other metallic and organic glasses Alloys

Tg /Tm [Refs.]

Tg /Tl [Refs.]

Ni Fe83 B17 Fe91 B9 Zr65 Be35 Ti63 Be37 Pd95 Si5 Pd82 Si18 Pd75 Si25 Pd40 Ni40 P20

0.246 0.51 [8,42] 0.415 0.503 0.517 0.591 0.605 0.599 0.67 [36] 0.66 [57] 0.483 0.606 0.610 0.681 [5,60] 0.725 0.624 0.745

0.246 0.51 [8,42] 0.369 0.503 0.497 0.383 0.605 0.488 0.585

Au77:8 Si8:4 Ge13:8 Nd60 Al15 Ni10 Cu10 Fe5 Nd61 Al11 Ni8 Co5 Cu15 Zr41:2 Ti13:8 Cu12:5 Ni10 Be22:5 SiO2 BS2 (BaOá 2SiO2 ) ZBLAN (Zr53 Ba20 La4 Al3 Na20 ) a

0.466 0.552 0.598 0.642 0.725 0.624 0.723

Tx ) Tg (K)

±

Tl ) Tm (K)

181

Zmax (mm)

±

594 62.0

136.0

0 45 24 54.0

24 70 15 56.0 0 0 22.3

35 92

7.0 [58] 25 [59] 4±6 6 50.0 [22] 400 [52]

Rc (K/s) [Refs.]

Methodsa [Refs.]

3 ´ 1010 [8,43] 1 ´ 106 [8] 2.6 ´ 107 [8,43] 1 ´ 107 [44] 6.3 ´ 106 [44] 5 ´ 107 [47] 1.8 ´ 103 [8,43] 1 ´ 106 [47] 1.57 [36] 120 [35] 3 ´ 106 [10]

F [56]

1.0 [60] 2 ´ 10ÿ4 [52] 3.0 [53] 0.07 [55]

C [58] C [49] C [49] C [22]

C ± mould casting; F ± splat quenching.

value of 1.5 K/s and then increased to 34.5 and 22.7 K/s, respectively. This kind of trend could also be expected for Pd-based Pd±Si±Cu alloys, though their critical cooling rate data are incomplete. Large bulk glass formation has been reported in Pd77:5 Si6 Cu16:5 alloy which has the highest Tg /Tl value of 0.602 while it does not have the highest value of Tg /Tm among the Pd-based Pd±Si±Cu alloys studied in Table 2. Fig. 9 shows the onset melting temperature Tm , o€set melting temperature Tl , glass transition temperature Tg , reduced glass transition temperature Trg based on Tg /Tl and critical cooling rates of the six La-based alloys as a function of alloy composition showing the critical cooling rates has a strong correlation with Trg . All these results indicate that Trg based on Tg /Tl has a stronger correlation with the critical cooling rate, than Tg /Tm . This can be easily understood as Tm and Tg are both less composition dependent while Tl is strongly dependent on the composition as shown in Fig. 9. Finally, the critical cooling rate and critical section thickness for glass formation alloys based on Pd, Zr, La, Mg, rare-earths and some inorganic glasses are plotted in Figs. 10 and 11, respectively, as a function of Trg based on Tg /Tm or Tg /Tl from

the data collected in Tables 2±4 including previously reported data [8,10]. These results also show that the correlation between the critical cooling rate and Tg /Tl is much stronger and the scatter is much smaller than that between critical cooling rate and Tg /Tm , particularly for the present bulk glass froming alloys (the solid circles in Fig. 10). Based on this, it is concluded that the Trg based on Tg /Tl has a stronger correlation with the critical cooling rate and critical section thickness for glass formation than Tg /Tm . 5. Conclusions It has been found that the best bulk metallic glass forming alloys are at or near-eutectic composition, a result which is consistent with the fact that reduced glass transition temperature Trg given by Tg /Tl is highest at the eutectic composition. There is a strong correlation between GFA of these bulk glass forming alloys and their reduced glass transition temperatures, while a less strong correlation between Trg based on Tg /Tm and GFA is observed. This indicates that instead of Tg /Tm , Tg /Tl should be used for Trg .

Z.P. Lu et al. / Journal of Non-Crystalline Solids 270 (2000) 103±114

Acknowledgements We would like to thank Professor R. L uck for his valuable suggestions concerning the melting behavior study and Professor H. Jones for valuable comments on the manuscript.

References [1] Y. Li, J. Mater. Sci. Technol. 15 (2) (1999) 97. [2] A. Inoue, T. Nakamura, T. Sugita, T. Zhang, T. Masumoto, Mater. Trans. JIM 34 (4) (1993) 351. [3] A. Inoue, T. Nakamura, N. Nishiyama, T. Masumoto, Mater. Trans. JIM 33 (10) (1992) 937. [4] A. Inoue, T. Zhang, Mater. Trans. JIM 37 (2) (1996) 185. [5] A. Peker, W.L. Johnson, Appl. Phys. Lett. 63 (17) (1993) 2342. [6] A. Inoue, N. Nishiyama, H. Kimura, Mater. Trans. JIM 38 (2) (1997) 179. [7] Y. Li, S.C. Ng, C.K. Ong, H.H. Hng, T.T. Goh, Scr. Mater. 36 (7) (1997) 783. [8] H.A. Davies, in: B. Cantor (Ed.), Rapidly Quenched Metals. Part III, Metals Society, London, vol. 1, 1978, p. 1. [9] D.R. Uhlmann, H. Yinnon, in: Glasses Science and Technology, vol. 1, Academic Press, New York, 1983, p. 1. [10] H.A. Davies, Phys. Chem. Glasses 17 (1976) 159. [11] C.V. Thompson, A.L. Greer, F. Spaepen, Acta Metall. 31 (11) (1983) 1883. [12] Y. He, T.D. Shen, R.B. Schwarz, Metal. Mater. Trans. A 29A (1998) 1795. [13] Z.P. Lu, T.T. Goh, Y. Li, S.C. Ng, Acta Mater. 47 (7) (1999) 2215. [14] C.J. Smithells, in: Metals Reference Book, 5th Ed., Butterworth, London, 1976, p. 400. [15] R. Busch, W. Liu, W.L. Johnson, J. Appl. Phys. 83 (8) (1998) 4134. [16] Y. Li, H. Jones, H.A. Davies, Scr. Metall. 26 (1992) 1371. [17] A. Inoue, T. Zhang, N. Nishiyama, K. Ohba, T. Masumoto, Mater. Trans. JIM 34 (12) (1993) 1234. [18] Y. Li, S.C. Ng, Z.P. Lu, Y.P. Feng, K. Lu, Philos. Mag. Lett. 78 (3) (1998) 213. [19] Z.P. Lu, Y. Li, S.C. Ng, Y.P. Feng, K. Lu, J. Non-Cryst. Solids 250±252 (1998) 689. [20] J.R. Matey, A.C. Anderson, J. Non-Cryst. Solids 23 (1977) 129. [21] J. Steinberg, A.E. Lord, L.L. Lacy, J. Johnson, Appl. Phys. Lett. 38 (3) (1981) 135. [22] W.L. Johnson, Mater. Sci. Forum 225±227 (1996) 35. [23] A. Inoue, T. Zhang, A. Takeuchi, IEEE Trans. Magn. 33 (5) (1997) 3814. [24] Y. Li, H.Y. Liu, H. Jones, J. Mater. Sci. 31 (1996) 1957.

113

[25] Y. Li, H.Y. Liu, H.A. Davis, H. Jones, Mater. Sci. Eng. A 179&180 (1994) 628. [26] A. Inoue, A. Kato, T. Zhang, S.G. Kim, T. Masumoto, Mater. Trans. JIM 32 (7) (1989) 609. [27] H.H. Hng, Y. Li, S.C. Ng, C.K. Ong, J. Non-Cryst. Solids 208 (1996) 124. [28] T. Zhang, A. Inoue, T. Masumoto, Mater. Trans. JIM 32 (11) (1992) 1005. [29] W.J. Boettinger, in: T. Masumoto, K., Suzuki (Eds.), Proceedings of the Fourth International Conference On Rapidly Quenched Metals, Japan Institute of Metals, Sendai, 1982, p. 99. [30] L.Q. Xing, P. Ochin, J. Mater. Sci. Lett. 16 (1991) 1277. [31] L.Q. Xing, P. Ochin, M. Harmelin, Mater. Sci. Eng. A 220 (1996) 155. [32] X.H. Lin, W.L. Johnson, J. Appl. Phys. 78 (11) (1995) 6514. [33] Y. Li, S.C. Ng, Z.P. Lu, Mater. Lett. 34 (1998) 318. [34] T.T. Goh, X. Hu, Z.P. Lu, Y. Li, Mater. Sci. Forum 312±314 (1999) 247. [35] A. Inoue, N. Nishiyama, T. Matsuda, Mater. Trans. JIM 37 (2) (1996) 181. [36] N. Nishiyama, A. Inoue, Mater. Trans. JIM 38 (5) (1997) 464. [37] N. Nishiyama, A. Inoue, Mater. Trans. JIM 37 (10) (1996) 1531. [38] H.S. Chen, Acta Metall. 22 (1974) 1505. [39] L.F. Chua, H.W. Kui, J. Appl. Phys. 84 (11) (1998) 5993. [40] A. Inoue, T. Zhang, Mater. Sci. Eng. A 226±228 (1997) 393. [41] A. Inoue, T. Takeuchi, T. Zhang, Metall. Mater. Trans. A 29 (1998) 1779. [42] P.K. Liao, K.E. Spear, in: T.B. Massalski (Ed.), Binary Alloy Phase Diagrams, 2nd Ed., American Society for Metals, Metals Park, OH, 1986, p. 481. [43] R. Elliott, Eutectic Solidi®cation Processing, Crystalline and Glassy alloys, Robert Hartnoll, Bodmin, Cornwall, 1983, p. 38. [44] L.E. Tanner, R. Ray, Acta Metall. 27 (1979) 1727. [45] H. Okamoto, L.E. Tanner, J.P. Abriata, in: T.B. Massalski (Ed.), Binary Alloy Phase Diagrams, 2nd Ed., American Society for Metals, Metals Park, OH, 1986, p. 716. [46] I.L. Murray, in: T.B. Massalski (Ed.), Binary Alloy Phase Diagrams, 2nd Ed., American Society for Metals, Metals Park, OH, 1986, p. 704. [47] A.G. Lewis, H.A. Davis, Mater. Sci. Eng. 23 (1976) 179. [48] M.S. Chandrasekharaiah, in: T.B. Massalski (Ed.), Binary Alloy Phase Diagrams, 2nd Ed., American Society for Metals, Metals Park, OH, 1986, p. 3045. [49] Y. He, C.E. Price, S.J. Poon, Philos. Mag. Lett. 70 (6) (1994) 371. [50] R. Bush, Y.J. Kim, W.L. Johnson, J. Appl. Phys. 77 (8) (1995) 4039. [51] M.J. Weber, in: R.W. Cahn, P. Haasen, E.J. Kraners (Eds.), Materials, Science and Technology, vol. 9, 1991, p. 645. [52] D.R. Uhlmann, J. Non-Cryst. Solids 7 (1992) 337.

114

Z.P. Lu et al. / Journal of Non-Crystalline Solids 270 (2000) 103±114

[53] A.A. Cabral Jr, C. Fredericci, E.D. Zanotto, J. Non-Cryst. Solids 219 (1997) 182. [54] P.F. James, J. Non-Cryst. Solids 73 (1985) 1985. [55] W.G. Jordan, A. Jha, J. Therm. Anal. 42 (1994) 759. [56] H.A. Davis, J.B. Hull, Mater. Sci. Eng. 23 (1976) 193. [57] A.J. Drehman, A.L. Greer, Acta Metall. 32 (3) (1984) 323.

[58] H.W. Kui, A.L. Greer, D. Turnbull, Appl. Phys. Lett. 45 (1984) 615. [59] Y. He, R.B. Schwarz, J.I. Archuleta, Appl. Phys. Lett. 69 (1996) 1861. [60] Y.J. Kim, R. Bush, W.L. Johnson, A.J. Rulison, W.K. Rhim, Appl. Phys. Lett. 65 (1994) 2136.