Kinetics of glass transition of La65Al20Co15 metallic glass

Kinetics of glass transition of La65Al20Co15 metallic glass

Journal of Alloys and Compounds 629 (2015) 11–15 Contents lists available at ScienceDirect Journal of Alloys and Compounds journal homepage: www.els...

950KB Sizes 0 Downloads 51 Views

Journal of Alloys and Compounds 629 (2015) 11–15

Contents lists available at ScienceDirect

Journal of Alloys and Compounds journal homepage: www.elsevier.com/locate/jalcom

Kinetics of glass transition of La65Al20Co15 metallic glass Ningbo Guo a, Chengying Tang a,b,⇑, Jiang Wang a, Chaohao Hu a, Huaiying Zhou a a b

Guangxi Key Laboratory for Informational Materials, Guilin University of Electronic Technology, 1 Jinji Road, Guilin, Guangxi 541004, PR China School of Materials Science and Engineering, Guilin University of Electronic Technology, 1 Jinji Road, Guilin, Guangxi 541004, PR China

a r t i c l e

i n f o

Article history: Received 30 September 2014 Received in revised form 20 November 2014 Accepted 1 December 2014 Available online 26 December 2014 Keywords: Metallic glasses Rare earth alloys and compounds Activation energy X-ray diffraction Glass-forming ability

a b s t r a c t The glass transition kinetics of melt-spun La65Al20Co15 amorphous alloy ribbons has been investigated using differential scanning calorimetry (DSC) technique with continuous heating of the sample at different heating rates. The dependence of the glass transition temperature (Tg) on the heating rate (b) has been determined by Lasocka’s relationship. The activation energy (E) of the glass transition has been obtained by employing Moynihan and Kissinger methods, respectively. It is found that the glass transition process cannot be described in terms of single activation energy. The variation of E with the extent of conversion (a) is further analyzed by using two iso-conversional methods. It is observed that the E of glass transition demonstrates a significant increase with the extent of conversion from amorphous to supercooled liquid phase of La65Al20Co15 metallic glass. Ó 2014 Elsevier B.V. All rights reserved.

1. Introduction In recent years, rare earth-based (RE-based) bulk metallic glasses (BMGs) have attracted increasing interests due to their novel physical properties and potential applications in the future, such as high glass-forming ability (GFA) [1–3], special magnetic properties [4–6], superplasticity and thermoplasticity in supercooled liquid region [3,7,8]. La–Al–Co metallic glass is one of La-based bulk metallic glasses with superior GFA and lower glass transition temperature [9,10], it shows regularly increasing of thermal stability, Vicker’s hardness Hv and fracture strength rf with the increase of elastic constants and the strong liquid behavior. Since the structural relaxation changing the properties of BMGs will take place after the BMGs were annealed near glass transition temperature (Tg), understanding the kinetics of glass transition of the alloy is of importance to determine the thermal stability and the mechanism of glass transition. The purpose of the present work is to experimentally investigate the kinetics of glass transition of La65Al20Co15 metallic glass using DSC with continuous heating of the sample at various heating rates. The heating rate dependence of the glass transition temperatures (Tg) and the glass transition activation energy throughout the glass transition region have been investigated.

⇑ Corresponding author at: School of Materials Science and Engineering, Guilin University of Electronic Technology, 1 Jinji Road, Guilin, Guangxi 541004, PR China. Tel.: +86 773 2291434; fax: +86 773 2290129. E-mail addresses: [email protected], [email protected] (C. Tang). http://dx.doi.org/10.1016/j.jallcom.2014.12.121 0925-8388/Ó 2014 Elsevier B.V. All rights reserved.

2. Experimental Ingot with nominal compositions of La65Al20Co15 (at%) was first prepared by melting pure La(99.99 wt%), Al(99.99 wt%) and Co(99.99 wt%) in an arc furnace under the argon atmosphere. Alloy ingot was re-melted four times in order to homogenize the alloy compositions with a total mass loss of less than 0.5%. The amorphous ribbons with about 3–4 mm width and 30–40 lm thickness were made by a single-roller melt spinning technique at a tangential wheel speed of 20 m/s under a partial high purity argon atmosphere. The nature of the quenched ribbons was firstly characterized by X-ray diffraction (XRD) with monochromatic Cu Ka radiation (k = 1.5405 Å, Rigaku D/ max2550VB, Japan). One ribbon sample with amorphous diffraction in its XRD patterns was then prepared by ion milling using ion beam energy at 4 keV and a low milling angle about 3° (Gatan Model 691 PIPS) without cooling to perform transmission electronic microscopy (TEM) observation. The TEM observations were carried out in a FEI Tecnai G2 F20 microscope with a point-to-point resolution of 0.24 nm. About 10 mg of the ribbon samples in the form of cut tape were then taken into an aluminum crucible and subjected to DSC (Netzsch STA449 F3 JupiterÒ) measurements under a flowing pure argon atmosphere supplied at the rate 20 ml/min at different heating rates (5, 10, 20, 30, 40 K/min), respectively. The temperature precision of this equipment is ±0.1 K with an average standard error of 1 K in the measured values. The temperature ranger covered in DSC measurement is from room temperature to 873.15 K. The DSC equipment is calibrated prior to measurements, using high-purity standard In, Sn, Bi, Zn, Al. The results of temperature and enthalpy calibrations obtained for the standard materials were within 3% of the values given in the literature [11].

3. Results and discussion Fig. 1 shows the XRD patterns for as melt-spun ribbons of alloy La65Al20Co15. Only a typical broad diffraction peak at the position of about 2h = 31.2° corresponding to the amorphous phase is observed and no evidence of crystalline Bragg peaks are detected, indicating a fully amorphous phase in the present alloy.

12

N. Guo et al. / Journal of Alloys and Compounds 629 (2015) 11–15

La65Al20Co15

10

20

30

40

50

60

70

80

90

Fig. 1. XRD patterns of the La65Al20Co15 metallic glass.

Tg-onset

Tg-end o 229.8 C

Tg-onset o 181.8 C

30K/min

20K/min 140 160 180 200 220 240 o

Temperature/ C

20K/min 10K/min 5K/min Tg-mid

100

2θ (degree)

Tg-end Exo

Exothermic Heat flow (a.u.)

Intensity (a.u.)

o

Tg-mid 205.8 C

40K/min

150

200

250

300

350

Temperature/ ◦C Fig. 3. DSC curves of the La65Al20Co15 metallic glass at different heating rates of 5, 10, 20, 30 and 40 K/min, respectively. The inset shows three assignments of the glass transition temperature (Tg).

Table 1 The glass transition temperature Tg of La65Al20Co15 metallic glass at different assignments.

Fig. 2. Bright field TEM and HRTEM images of the La65Al20Co15 metallic glass ribbons prepared by ion-milling. The inset in right corner shows the corresponding SAD patterns.

Fig. 2 shows the bright field (BF) transmission electron microscopy (TEM) images, the corresponding selected area diffraction (SAD) patterns (insets), and HRTEM images obtained from ionmilled La65Al20Co15 alloy. The resulting BF TEM images presented in Fig. 2(a) shows uniform amorphous feature and there is no distinct boundary. Both the SAD patterns and HRTEM images confirm that the TEM-studied sample regions of the samples are indeed fully amorphous. This finding is in good agreement with the XRD experimental results. Fig. 3 presents the DSC curves for alloy La65Al20Co15 at five different heating rates of 5, 10, 20, 30, 40 K/min, respectively. All the DSC traces exhibit an endothermic reaction, which is characteristic of glass transition region followed by an exothermic peak at higher temperatures and is the characteristic of crystallization region. The onset of the endothermic change is commonly used to define the

Heating rate b (K/min)

Tg-onset (°C)

Tg-mid (°C)

Tg-end (°C)

5 10 20 30 40

171.8 175.2 181.4 184.8 187.2

196.2 200.3 205.8 211.2 214.1

220.5 225.5 230.8 237.6 241.0

glass transition temperature Tg. To check the application of free volume model for the Kissinger’s equation to the glass transition, Tg is determined using three different definitions [12–16], which are shown in the inset of Fig. 3. In the present work, the extrapolated onset temperature of the endothermic trace of glass transition, Tg-onset, is determined from the intersection of the extrapolated straight line of initial change in the baseline slope of the corresponding region, the offset of the glass transition temperature, Tg-offset, is defined as the temperature corresponding to the intersection point of the tangents to the portions adjoining the final transition elbow in the DSC trace, and the midpoint temperature, Tg-mid, is determined from the average value of Tg-onset and Tg-offset. The obtained glass transition temperatures from Fig. 3 at five various heating rates are given in Table 1. The glass transition temperature shifts to higher temperature with increasing heating rate, indicating the kinetic nature of the glass transition. The glass transition temperature reflects the strength or rigidity of the glassy structure of the alloy. It has been widely observed that the dependence of Tg on the heating rate b follows empirical Lasocka’s relationship in the form [17–22]:

T g ¼ Ag þ Bg lnðbÞ

ð1Þ

where Ag and Bg are the constants for a given glass composition. The value of Ag is equal to the glass transition temperature for the heating rate of 1 K/min and the value of Bg reflects the rate dependence of the configurational changes within the glass transition region of the supercooled liquids: the lower the cooling rate of the melt, the lower the value of Bg [23]. In order to check if Eq. (1) describes the heating rate dependence of Tg, the Tg is plotted against ln b as shown in Fig. 4. The figure shows the shift of Tg follows Eq. (1) for all Tg assignments, indicating the validity of this equation for Tg of the La65Al20Co15 glass. The values of the constants Ag and Bg obtained

13

N. Guo et al. / Journal of Alloys and Compounds 629 (2015) 11–15

4.0

240

3.5

Tg-onset Tg-mid Tg-end

3.0

ln (β)

o

Tg/ C

220

200

2.5

Tg-onset Tg-mid Tg-end

2.0 180

1.5 1.5

2.0

2.5

3.0

3.5

1.95

ln (β)

dðln bÞ E ¼ dð1=T g Þ R

ð2Þ

where R is the gas constant, b is the heating rate, respectively. The plot of ln b against 1/Tg yields a straight line, the slope of which gives the activation energy E of thermal relaxation. Fig. 5 shows the plots of ln b vs. 1000/Tg for Ce65Al20Co15 glass. The obtained values of E for three different Tg assignments are listed in Table 2. The Kissinger equation is another approach that has been widely used for evaluation of glass transition activation energy from the dependence of Tg on the heating rate [26–29]. Although this equation was originally derived to describe the kinetics of chemical reactions, it is widely used to determine the activation energy of crystallization. On the basis of the free volume model of glass transition, Ruitenberg [27] showed that the Kissinger method can also be used to determine the glass transition activation energy. According to the Kissinger method, the glass

Table 2 The values of Lasocka parameters A and B, the activation energy of glass transition E for Tg at different assignments.

Tg-onset Tg-mid Tg-end

Lasocka parameters

Activation energy (kJ/mol)

A (°C)

Moynihan Eq.

158.8 181.2 203.6

7.60 8.68 9.83

217.9(±1.5) 214.4(±2.0) 244.6(±1.4)

2.10

2.15

2.20

2.25

Kissinger Eq. 213.7(±1.6) 206.5(±1.9) 246.8(±1.7)

Fig. 5. Plots of ln (b) vs. 1/Tg for La65Al20Co15 glass. The straight lines are fit to Moynihan equation.

transition activation energy can be obtained using the following expression:

d lnðb=T 2g Þ E ¼ R dð1=T g Þ

ð3Þ

The plot of ln (b/Tg2) vs. 1/Tg gives a straight line, the slope of which gives E of glass transition. Fig. 6 shows such a plot for Tg of La65Al20Co15 glass and the obtained values of E for three different Tg assignments are also listed in Table 2. As can be seen from Figs. 5 and 6, the data can be well fitted to straight lines indicating the applicability of both methods. It is evident from Table 2 that the there is a significance changes in the obtained values of activation energies E for the three different Tg assignments, which means that the activation energy is not constant throughout the glass transformation region. It is also found that the calculated activation energy using Moynihan and Kissinger equations is in good agreement with each other and the difference is within the experimental errors, although these two models are based on different theoretical methods. In order to investigate the variation of E with the extent of conversion a throughout the glass transition region, two isoconversional methods have been used. The extent of transformation (a) used in the isoconversional analysis is determined using the partial

-8.5

-9.0

ln (β /Tg 2)

from the least squares fittings of the data on Fig. 4 are given in Table 2. The observed change of Bg from Tg-onset to Tg-offset reveals a variation in the configurational states throughout the glass transition region. Similar indication of the physical significance of Bg was reported in a variety of glasses [15,20,21]. The heating rate dependence of Tg of a glass has usually interpreted in terms of thermal relaxation phenomena. Since three different definitions of Tg considered in this work correspond to different degree of conversion, it is useful to compare the activation energy E as determined using the different definitions. This will demonstrate if E varies with the degree of conversion or remains constant throughout the glass transition. It is reported that glass transition activation energy E is responsible for the molecular motion and rearrangement of the atoms around Tg. Two approaches are used to evaluate the activation energy of glass transition. The most frequently used approach for the kinetics of the glass transition is based on the Moynihan method [24,25].

B

2.05

1000/Tg

Fig. 4. The dependence of glass transition temperature Tg with ln b of the La65Al20Co15 metallic glass. The solid lines represent the least squares fit to Eq. (1).

Tg assignments

2.00

-9.5

Tg-onset Tg-mid Tg-end

-10.0

-10.5

-11.0

1.95

2.00

2.05

2.10

2.15

2.20

2.25

1000/Tg Fig. 6. Plots of ln (bTg2) vs. 1000/Tg for La65Al20Co15 metallic glass. The straight lines are fit to Kissinger equation.

14

N. Guo et al. / Journal of Alloys and Compounds 629 (2015) 11–15

250

1.00

KAS Vyazovkin

240

0.50

230

E (kJ/mol)

5K 10K /min 20K /min /mi n 30K /mi n 40K /mi n

α

0.75

20K/min

AT Exo

0.25

220 210 200

A

Tg-onset o

Tg-end

181.8 C

190

o

α= A T/A

229.8 C

180

0.00

180

195

210

o

225

Temperature/ C

180

195

210

225

240

170 0.0

255

0.2

o

Temperature/ C Fig. 7. Degree of transformation, a, as a function of temperature at different heating rate. The inset shows how the a was calculated.

area method. In this method, the value of a at temperature T is defined as a = AT/A, where A is the area under the curve of the endotherm between the temperature Tg-onset and Tg-end. AT is the area between Tg-onset and T. The obtained a versus T at different heating rates has been shown in Fig. 7 and Table 3. Using these experimental data of Fig. 7, the Vyazovkin [30,31], and the Kissinger–Akahira–Sunose (KAS) [26,32] methods are used to investigate the variation of E through the glass transition region. According to the advanced Vyazovkin method, for a set of n experiments carried out at different heating rates, the effective activation energy can be determined at any particular value of a by finding the value of Ea for which the objective function X is minimized, where



n X n X IðEa ; T a;i Þbj i¼1 j–i

ð4Þ

IðEa ; T a;j Þbi

where the subscripts i and j represent ordinal numbers of two experiments performed under different heating programs. The integral I(E, T) is given by

IðE; TÞ ¼

Z

Ta

exp

T a Da



 Ea dT RT

ð5Þ

The integral I(E, T) was numerically evaluated, using trapezoidal method, and Ea is the value that minimizes X in Eq. (4). Minimization is repeated for each value of a to obtain a dependence of Ea on a. On the basis of KAS method, the Ea can be determined for each a using the following expression.

d lnðbi =T 2a;i Þ Ea ¼ R dð1=T a;i Þ

ð6Þ

0.4

α

0.6

0.8

1.0

Fig. 8. The activation energy (E) of the glass transition as a function of the extent of conversion (a) determined using different iso-conversional methods.

where Ta,i is the set of temperatures related to a particular degree of conversion a at different heating rates, bi. Fig. 8 shows the variation of E of glass transition with the extent of conversion a using these two isoconversional methods. As shown in this figure, the E of glass transition throughout the glass transition region increases almost linearly with increasing degree of conversion a. This implies that the glass transition cannot be described by a single step mechanism. Multiple-step reactions involving relaxation process with different activation energy and mechanisms were demonstrated. For La65Al20Co15 glass, when a is smaller, the value of activation energy is lower, suggesting the La65Al20Co15 glass goes easily into the glass transition region. When a increases, the value of activation energy is higher. This means that the La65Al20Co15 glass gets hardly across the glass transition into the crystallization region. This finding was in good agreement with the metallic glasses reported by Wang et al. [20], Jankovic et al. [33], Joraid [34], and recently by Zhang et al. [15]. It should be noted that the E obtained by Vyazovkin method and KAS method agree well with each other only for higher values of a. This is maybe attributed to the different models adopted. 4. Conclusions Differential scanning calorimetry has been used to investigate the kinetics of glass transition of La65Al20Co15 glass with continuous heating of the sample at different heating rates. Using three different definitions of glass transition temperature, the heating rate dependence of the glass transition temperature (Tg) has been determined. The activation energy of glass transition is obtained by employing Moynihan and Kissinger methods, respectively. It is found that the variation of E throughout the glass transition region when different assignments of Tg are used. The variation

Table 3 Extent of transformation, a, as a function of temperatures at different heating rates. 5 K/min

10 K/min

20 K/min

30 K/min

40 K/min

a

T (°C)

a

T (°C)

a

T (°C)

a

T (°C)

a

T (°C)

0 0.04944 0.18668 0.43200 0.68816 0.87075 1

171.1 179.9 188.0 196.1 204.2 212.3 220.5

0 0.01229 0.10617 0.39738 0.68615 0.91016 1

175.2 183.5 191.8 201.1 209.4 217.9 225.5

0 0.02817 0.17602 0.43905 0.70388 0.90497 1

181.4 189.5 197.6 205.7 213.8 221.9 229.8

0 0.04012 0.20745 0.45245 0.69776 0.89938 1

184.8 193.6 202.4 211.2 220.0 228.8 237.6

0 0.06166 0.24797 0.49210 0.71287 0.89717 1

187.2 196.1 205.0 213.9 222.8 231.7 241.0

N. Guo et al. / Journal of Alloys and Compounds 629 (2015) 11–15

of E with a was further analyzed using two different iso-conversion methods. The activation energy of glass transition throughout the glass transition region increases almost linearly with the increasing a. The transformation from the amorphous to a supercooled liquid phase of La65Al20Co15 glass is a multi-step processes. Acknowledgements The financial support from the National Natural Science Foundation of China (51261005), Sino-German Center for Promotion of Science (GZ1002), the Natural Science Foundation of Guangxi (2011GNSCFC018002, 2012GXNSFGA060002) and the Key Lab of Guangxi Informational Materials (1210908-04-Z) are greatly acknowledged. References [1] [2] [3] [4] [5] [6] [7] [8]

D. Ding, M.B. Tang, L. Xia, J. Alloys Comp. 581 (2013) 828–831. Y. Ji, S. Pang, T. Zhang, J. Alloys Comp. 505 (2–3) (2010) 497–500. T. Zhang, R. Li, S.J. Pang, J. Alloys Comp. 483 (2009) 60–63. A. Wisniewski, R. Puzniak, Z. S´niadecki, A. Musiał, M. Jarek, B. Idzikowski, J. Alloys Comp. 618 (2015) 258–262. Z. S´niadecki, J. Marcin, I. Škorvánek, N. Pierunek, B. Idzikowski, J. Alloys Comp. 584 (2014) 477–482. L. Liang, X. Hui, C.M. Zhang, G.L. Chen, J. Alloys Comp. 463 (2008) 30–33. Z.P. Lu, X. Hu, Y. Li, S.C. Ng, Mater. Sci. Eng. A 304–306 (2001) 679–682. Y. Zhang, H. Tan, Y. Li, Mater. Sci. Eng. A 375 (2004) 436–439.

15

[9] Y.F. Ji, S.J. Pang, T. Zhang, J. Alloys Comp. 505 (2010) 497–500. [10] Y. Wang, W. Zhao, G. Li, Y. Li, R. Liu, J. Alloys Comp. 511 (2013) 185–188. [11] Y.S. Touloukian, Specific Heat of Non-Metallic Elements, Compounds and Mixtures Thermophysical Properties of Matter, Plenum Press, New York, 1970. pp. 25. [12] M.H.R. Lankhorst, J. Non-Cryst. Solids 297 (2002) 210–219. [13] M.K. Rabinal, J. Non-Cryst. Solids 241 (1998) 121–127. [14] C. Dohare, N. Mehta, J. Alloys Comp. 587 (2014) 565–572. [15] B. Zhang, C. Tang, W. Xu, W. Pan, J. Wang, H. Zhou, Mater. Chem. Phys. 142 (2013) 707–711. [16] N. Mehta, A. Kumar, J. Non-Cryst. Solids 358 (2012) 2783–2787. [17] A.B. Abd El-Moiz, N. Afify, M.M. Hafiz, Phys. B 182 (1992) 33–41. [18] A.L. Omar, M.A.I. Mousa, K.A. Marouf, Phys. B 395 (2007) 69–75. [19] N. Mehta, R.K. Shukla, A. Kumar, Chalcogenide Lett. 1 (10) (2004) 131–137. [20] S. Wang, S. Quan, C. Dong, Thermochim. Acta 532 (2012) 92–95. [21] A.A. Elabbar, J. Alloys Comp. 476 (2009) 125–129. [22] H. Kumar, N. Mehta, Mater. Chem. Phys. 134 (2012) 834–838. [23] A.A. Abu-Sehly, M.A. El-Oyoun, A.A. Elabbar, Thermochim. Acta 472 (2008) 25– 30. [24] C.T. Moynihan, P.K. Gupta, J. Non-Cryst. Solids 29 (1978) 143–158. [25] C.T. Moynihan, J. Am. Ceram. Soc. 76 (1993) 1081–1087. [26] H.E. Kissinger, Anal. Chem. 29 (1957) 1702–1706. [27] G. Ruitenberg, Thermochim. Acta 404 (2003) 207–211. [28] A.A. Abu-sehly, Mater. Chem. Phys. 125 (3) (2011) 672–677. [29] Z.Z. Yuan, B.X. Wang, Y.D. Tong, X.D. Chen, J. Alloys Comp. 429 (2007) 104–110. [30] S. Vyazovkin, N. Sbirrazzuoli, I. Dranca, Macromol. Rapid Commun. 25 (2004) 1708–1713. [31] S. Vyazovkin, A.K. Burnham, J.M. Criado, L.A. Perez-Maqueda, C. Popescu, N. Sbirrazzuoli, Thermochim. Acta 520 (2011) 1–19. [32] T. Akahira, T. Sunnose, Res. Rep. Chiba Inst. Technol. 16 (1972) 22–31. [33] B. Jankovic, B. Adnadevic, J. Ovanovic, Thermochim. Acta 452 (2007) 100–115. [34] A.A. Joraid, Thermochim. Acta 456 (2007) 1–6.