Non-isothermal crystallization kinetics of Mg60Zn30Ti5Si5 amorphous alloy prepared by mechanical alloying

Accepted Manuscript Non-isothermal crystallization kinetics of Mg60Zn30Ti5Si5 amorphous alloy prepared by mechanical alloying K. Chen, H.M. Yang, J.S. Gao, X.L. Li, C.G. Yu, G.X. Ma, X. Yuan PII:

S0925-8388(16)31831-X

DOI:

10.1016/j.jallcom.2016.06.107

Reference:

JALCOM 37976

To appear in:

Journal of Alloys and Compounds

Received Date: 4 January 2016 Revised Date:

10 June 2016

Accepted Date: 11 June 2016

Please cite this article as: K. Chen, H.M. Yang, J.S. Gao, X.L. Li, C.G. Yu, G.X. Ma, X. Yuan, Nonisothermal crystallization kinetics of Mg60Zn30Ti5Si5 amorphous alloy prepared by mechanical alloying, Journal of Alloys and Compounds (2016), doi: 10.1016/j.jallcom.2016.06.107. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Non-isothermal crystallization kinetics of Mg60Zn30Ti5Si5 amorphous alloy prepared by mechanical alloying K. Chena*, H.M. Yanga, J.S. Gaob, X.L. Lia, C.G. Yua, G.X. Maa, X. Yuana

RI PT

a. College of Engineering, Nanjing Agricultural University, Nanjing 210031，China b. College of Materials Science and Engineering, Anhui University of Science and Technology, Huainan Anhui 232001, China

SC

Abstract: Mg60Zn30Ti5Si5 amorphous alloy was prepared via mechanical alloying

M AN U

(MA) and heat treated in different way. X-ray diffractometer (XRD) was used to detect the phases of alloy. The crystallization characteristics and kinetics of amorphous alloy were investigated by differential scanning calorimetry (DSC). Tests reveal that raw materials transform to amorphous alloy after 6 hours mechanical

TE D

alloying; during subsequent heating, MgZn2, Mg2Si and Zn were crystallized in turn. Apparent activation energies (Ea) and local activation energies (El) of non-isothermal crystallization were calculated according to Kissinger and Ozawa models based on

EP

DSC data. Results show that Ea3＞Ea2＞Ea1; El1 decreases while El2 increases with

AC C

crystallization; El3 is almost stable during heating. Avrami parameters (n) indicate that the first devitrification is interface controlled growth and later precipitations are diffusion controlled growth. Keywords: amorphous alloy; mechanical alloying; Mg60Zn30Ti5Si5; apparent activation energy of crystallization; local activation energy

* Corresponding author. Tel.: +86 25 58606580. E-mail address: [email protected] (K.Chen).

ACCEPTED MANUSCRIPT 1. Introduction As research focus, amorphous magnesium alloys and composites have advantages of low density, high rigidity, wear resistance and corrosion resistance [1-6]. During

RI PT

this decade, many studies on Zn-containing Mg alloys, such as Mg-Zn-Ca, Mg-Zn-Cu and Mg-Zn-Y, were reported and got much academic attention [2, 3, 5-10]. This is because Mg-Zn alloy has great potential in biomedical engineering. So, various

SC

Mg-Zn base amorphous alloys were developed. For example, Mg-Zn-Ca amorphous

M AN U

alloy has great potential as biomaterial, because of their well compatibility with bone cell and elastic moduli similar to bones. Besides above merits，they’re biodegradable and inexpensive [7,8,10,11]. Based on above background, this study focused on Mg-Zn-Ti-Si system, in which nontoxic Ti and Si were added to replace Ca for their

TE D

lower-price.

Mechanical alloying is one of the most effective methods for preparing amorphous materials with various compositions, including alloys which cannot be prepared by

EP

smelting. Furthermore, it’s attractive to acquire bulk amorphous composite or porous

AC C

materials through subsequent powder compact/sinter processing [12]. Therefore，in this research Mg60Zn30Ti5Si5 amorphous alloy was prepared by simple and convenient MA processing.

Crystallization characteristics and kinetics are important issues for metal glass. The

application and processing of metal glass greatly correlate with its stability and crystallization behavior. So far，there’re many articles which focus on crystallization of bulk metallic glasses (BMGs) prepared by rapid solidification (RS) [6, 13-18]. But

ACCEPTED MANUSCRIPT crystallization kinetics of Mg based metal glass prepared by MA was less reported. Considering great structure difference between BMGs and amorphous metal powder prepared by MA, it’s necessary to clarify crystallization process of the latter. In

RI PT

addition, the crystallization of quaternary system is complicate compared with that of binary or ternary alloys. So, precipitation procedure and DSC analyses of Mg60Zn30Ti5Si5 amorphous alloy were presented and discussed in this work. As

SC

emphasis, non-isothermal crystallization kinetics of Mg60Zn30Ti5Si5 alloy was detailed

M AN U

analyzed. The results could provide references for its utilization and possible following processing such as sintering.

2. Experimental

TE D

The raw materials is Mg (99.9 %), Zn (98 %), Ti (99.99 %), and Si (99.99 %) powders with atom ratio of 60︰30︰5︰5, respectively. Metal powder, stainless balls (Φ = 20 mm, 10 mm) and stearic acid, as a process controller, were added into ball

EP

milling tank. Samples were milled for 6 hours in a XQM-4L frequency planetary ball

AC C

mill at rotation speed of 420 rpm. After milling, amorphous powder was soaked in acetone to remove stearic acid,

then filtrated and dried in the air. In order to investigate the crystallization process, a part of Mg60Zn30Ti5Si5 powder was heated to 423 K, 473K, 523 K, 573 K, 673 K, 773 K and 873 K at heating rate of 20 K/min separately; then cooled to room temperature for subsequent phase detection. The phases of samples were analyzed using Panalytical X’Pert Powder X-ray diffractometer.

ACCEPTED MANUSCRIPT In another hand, amorphous Mg60Zn30Ti5Si5 alloy were non-isothermal crystallized by continuously elevating its temperature to 823 K at rates of 5 K/min, 10 K/min, 20 K/min and 30 K/min. The heating and cooling process were carried out by Netzsch

Ar. Meanwhile, DSC curves were recorded for later analysis.

SC

3. Results and discussion

RI PT

STA449F3 simultaneous thermal analysis instrument under protection atmosphere of

M AN U

3.1 Phase evolution of Mg60Zn30Ti5Si5 alloy prepared by MA

Fig. 1 is XRD pattern of Mg60Zn30Ti5Si5 alloy prepared by MA. It presents a typical broad hump of which 2theta value is between 35° and 45° and a few low peaks. That indicates metal powder generally transformed to amorphous alloy after 6 h

TE D

milling. There is a small hump at the position that 2θ ≈ 20°. The pre-peak implies short-range order in sample; it is an important characteristic of metal glass. The remaining peaks with small intensity correspond to tiny Si and Mg grains which were

EP

unalloyed during MA.

AC C

Mg60Zn30Ti5Si5 amorphous powders were heated to 423 K, 473K, 523 K, 573 K, 673 K, 773 K and 873 K, separately. Corresponding XRD patterns of heat-treated powder were shown in Fig. 2. Crystallization sequence starting from amorphous alloy was indicated from XRD patterns. First，intensity of Si diffraction peaks is enlarged. With temperature elevating, MgZn2 precipitated from amorphous alloy at ~473 K. This process lasted for a long time and the temperature range is ~100 K. Crystallization of Mg2Si started before temperature reach 573 K and ended before 673

ACCEPTED MANUSCRIPT K. At last, diffraction peaks of Zn raised before 773 K. According to XRD pattern of sample heated to 873 K, it is mainly comprised of α-Mg, MgZn2, Mg2Si, Zn and Si. 3.2 DSC curves

RI PT

To further clarify crystallization process of Mg60Zn30Ti5Si5 amorphous alloy, its continuous heating DSC curves at different heating rate are shown in Fig. 3. Each curve exhibits three exothermic peaks which are characteristics of crystallization.

SC

Take DSC curve measured at heating rate of 20 K/min for example, the first

M AN U

exothermic hump has a large temperature range of 450 ~ 580 K. It shows the heat release which caused by crystallization of MgZn2, according to above XRD analysis. The second exothermic peak is close to the first one and occurs during 580 ~ 640 K. Corresponding phase transition is devitrification of Mg2Si. Finally, precipitation of Zn

TE D

results in the last exothermic peak which temperature is between 720 and 800 K. Usually, there is only one main devitrification peak in DSC curve of bulk metal glass prepared by RS [6, 13~15]. But, in contrast to that of most Mg based metal glass,

EP

phase evolution of Mg60Zn30Ti5Si5 amorphous alloy is still sophisticated. Similar

AC C

results appeared with other Mg amorphous powders prepared by MA. For example, DSC curve of Mg60Zn35Ca5 amorphous powder contains two exothermic peaks with formation of Mg-Zn intermetallics [10]. In this research, except for 3 times crystallization, unalloyed tiny Si grains remains in Mg60Zn30Ti5Si5 alloy during heating. Comparing DSC curves obtained at diverse heating rate, exothermic peaks shifts to higher temperature with heating rate (β) increasing. Characteristic temperatures

ACCEPTED MANUSCRIPT derived from DSC curves are listed in Table 1. Because the first exothermic peak is overlap with the second one in each DSC curve, onset temperatures are inaccurate. So, Tp (crystallization peak temperature) value is used as characteristic temperature in

RI PT

following analysis. Table 1 shows that Tp1, Tp2 and Tp3 increase as β increasing. On the other hand, the areas of exothermic peaks also enlarge with β increasing. These crystallization

SC

phenomena of Mg60Zn30Ti5Si5 amorphous alloy indicate kinetic feature. The heating

M AN U

rate dependence of crystallization is caused by the fact that nucleation is a thermally activated process.

3.3 Activation energy of crystallization (E)

3.3.1 Apparent activation energy of crystallization (Ea)

TE D

The activation energy of crystallization (E) is average energy which is necessary for transition of amorphous structure to crystalline. It presents the difficulty of crystallization for an amorphous material. The larger activation energy, the more

EP

stable of amorphous alloy is.

AC C

Kissinger method is often used to calculate apparent activation energy (Ea) of crystallization [19，20]. The equation is: ln

=−

+C

(1)

where Ea is activation energy, β is heating rate during the experiment, T is the characteristic temperatures such as Tx (onset temperature of crystallization) and Tp (temperature of peak point), C1 is a constant.

Since the Tx of the second crystallization peak is not exact, Tp values were used as

ACCEPTED MANUSCRIPT T for calculation in this study. According to three theories, the relationships of ln(β/Tp2) -10000/Tp was derived from Tp and reflected in Fig. 4. In the figure, points were fitted to straight lines and slopes of fit lines are Eα values. Eα values of three precipitations

RI PT

were also listed in charts. Activation energy of first crystallization (Ea1) is ~91 kJ·mol-1; Ea2 is ~116 kJ·mol-1 and Ea3 is ~155 kJ·mol-1. Ea of first crystallization of Mg60Zn30Ti5Si5 alloy is not as

SC

high as that of some other amorphous magnesium alloy prepared by rapid

M AN U

solidification [14]. Similar low value occurs for amorphous Mg alloys when DSC curves contain 2 or more crystallization peaks [21, 22]. In this study, the sample was severely deformed after mechanical alloying and in high energy state. So, driving force for crystallization is very high while activation energy is low. The activation

TE D

energy of the third crystallization which occurs at higher temperature is larger than that of first and second crystallization. The reason is that diffusion channel decreased after former precipitations. Therefore, more energy is required for the third

EP

transformation.

AC C

The total Ea of three peaks is considerable high, i.e. large energy is needed for Mg60Zn30Ti5Si5 alloy to transform from amorphous to entirely crystal. Moreover, temperature difference between the third crystallization and the second one is large (about 150 K). This may be advantageous to preparing partly crystallized amorphous magnesium composite. The alloy powder could be sintered at temperature above Tx2 but lower than Tx3 to generate a nanocomposite alloy exhibiting a mixture of amorphous phase and nanocrystalline Mg alloy [10, 23].

ACCEPTED MANUSCRIPT 3.3.2 Local activation energy of crystallization (El) Limitation of Kissinger method is that conversion rates at Tp be regarded as consistent for different heating rates. But the fact may deviates from this requirement.

RI PT

So, local activation energy at different transaction stage should be concerned. According to Ozawa-Flynn-Wall formula for non-isothermal crystallization [24], crystallization activation energy can be evaluated: x

+C

SC

ln( ) = −1.052

(2)

M AN U

where Tx is the temperature corresponding to the certain crystallized volume fraction (x), C2 is a constant. While x could be obtained by the ratio of part area before Tx and total area for an exothermal peak.

Crystallized fraction x as functions of temperature Tx at different heating rates of

TE D

Mg60Zn30Ti5Si5 alloy are plotted in Fig. 6(a), (b) and (c), respectively. The curves with different temperature range show typical sigmoid shape. The middle parts of curves are nearly straight lines.

EP

Local activation energy (El) at certain crystallized fraction could be calculated from

AC C

plot values in Fig. 5 by above Ozawa- Flynn-Wall formula. The results are listed in Fig. 6. It’s found that values of El1 and El2 derived from Kissinger formula are close to local activation energy when x is about 0.4. Variation of local activation energy with crystallized fraction (x) exhibits the nucleation and growth behavior in the crystallization process. With x increasing, El1 increases while El2 and El3 decrease. Local activation energy of BMGs prepared by rapid solidification often rises at first and then decrease along with crystallize process. This phenomenon is not coincidence

ACCEPTED MANUSCRIPT with El of Mg60Zn30Ti5Si5 alloy. Because there are still lots of tiny area of short-range ordered atoms after MA, it’s easy to form nuclei in the alloy, i.e. the nucleation energy is much smaller than that of BMGs. In fact, the value of El1 (about 70 kJ·mol-1) is

RI PT

fairly low at beginning. With growth of the nucleation, it’s more difficult for atoms to diffusion. Therefore, El1 is increasing for the first crystallization of Mg60Zn30Ti5Si5 alloy. Until Zn atoms are consumed, the alloy doesn’t entirely change to crystal. For

SC

the second crystallization, El2 is big at starting and slowly decreases from about 150

M AN U

kJ·mol-1 to 100 kJ·mol-1 with x raises. This is because the diffusion distance of Mg, Si atoms shortens. El3 is higher and almost stable during third precipitation. It’s corresponds to the phase separation of unstable Mg-Zn compound. 3.4 Avrami parameter (n)

TE D

Avrami parameter, which is represented by n, is an important index of grain growth mold during isothermal crystallization. But, Augis and Bennett figure out a way to

EP

calculate n value of non-isothermal crystallization [25]. The formula is as following: =

.

∆

∙

(3)

AC C

where ∆T is the full width at half maximum of the crystallization peak, T corresponds to temperature of maximum of crystallization peak, i.e. Tp.

So, peak 1 and peak 2 in each DSC curve were fitted and separated by using

software. Then ∆T values of three exothermic peaks were measured (listed in Table 2). Substituting the ∆T and Ea derived by Kissinger formula, n values were calculated and shown in Table 2 as well. For peak 1, n values of four curves are all smaller than 1. So the first crystallization

ACCEPTED MANUSCRIPT is interface controlled growth with grain boundary nucleation after saturation [26]. This corresponds to MgZn2 crystallization based on above XRD pattern. In the alloy, ratio of Mg and Zn atoms is 2:1, which exceed their mutualsolubility. Inevitably,

RI PT

MgZn2 precipitates during heating. Since Mg and Zn are abundant and uniformly mixed, it’s easy for MgZn2 to nucleate without diffusion. That’s the reason of interface controlled growth occurs at first.

SC

n values of peak 2 and peak 3 are about 2. The corresponding transformations are

M AN U

diffusion controlled growth which nucleation rate is decreased. The result is consistent with above calculation of activation energy: along with crystallization and growth, more and more grains are form; diffusion channels of atoms are decreasing and crystallization becomes relatively difficult. More energy is needed for atoms to

4. Conclusions

TE D

move and form new phase. As results, nucleation rate is lower and Ea3﹥Ea2﹥Ea1.

EP

1. Mg60Zn30Ti5Si5 alloy changes to amorphous after 6 hours mechanical alloying；

AC C

MgZn2, Mg2Si and Zn form during non-isothermal heating sequentially. 2. Based on DSC curves, apparent activation energy of each crystallization peak was calculated according to Kissinger formula, and Ea3＞Ea2＞Ea1. Ea1, Ea2 are ~91

kJ·mol-1 and ~116 kJ·mol-1, separately. But, Ea3 is higher (~155 kJ·mol-1) and

temperature difference between the third peak and the second peak is large. So, Mg60Zn30Ti5Si5 amorphous alloy is hopeful to be made into partly crystallized amorphous composite.

ACCEPTED MANUSCRIPT 3. The calculated values of local crystallization activation energy are consistence with apparent activation energy values. Variation trends are different for three precipitations: El1 decreases while El2 increases with crystallization; El3 is almost

RI PT

stable during heating. 4. Avrami parameters (n) indicate that the first crystallization is interface controlled

SC

growth; the mechanism of later precipitations is diffusion controlled growth.

M AN U

Acknowledgements

This research is supported by the Fundamental Research Funds for the Central Universities (No.KYZ201657) and the Fundamental Research Funds for the Central

References

TE D

Universities (No.KJSY201518).

[1] L.A. Deibler, K.O. Findley. Characterization of an amorphous–crystalline Mg-based

EP

composite, J. Alloys Compd. 463, 1-2 (2008) 173–179.

AC C

[2] G.B. Liu, P. Gao, Z. Xue, S.Q. Yang, M.L. Zhang. Study on the formation of new Mg–Cu– Ti–Y quaternary bulk metallic glasses with high mechanical strength, J. Non-Cryst. Solids. 358 (2012) 3084–3088.

[3] K.Q. Qiu, N.N. Hu, H.B. Zhang, W.H. Jiang, Y.L. Ren, P.K. Liaw. Mechanical properties and fracture mechanism of as-cast Mg77TM12Zn5Y6(TM = Cu, Ni) bulk amorphous matrix composites, J. Alloys Compd. 478 (2009) 419–422. [4] E. Koç, M.B. Kannan, M. Ünal, E. Candan. Influence of zinc on the microstructure,

ACCEPTED MANUSCRIPT mechanical properties and in vitro corrosion behavior of magnesium–zinc binary alloys, J. Alloys Compd. 648 (2015) 291-296. [5] M.S. Dambatta, S. Izman, B. Yahaya, J.Y. Lim, D. Kurniawan. Mg-based bulk metallic

RI PT

glasses for biodegradable implant materials: A review on glass forming ability, mechanical properties, and biocompatibility, J. Non-Cryst. Solids. 426 (2015) 110–115.

[6] R. Babilas, K. Cesarz-Andraczke, R. Nowosielski, A. Burian. Structure, Properties, and

SC

Crystallization of Mg-Cu-Y-Zn Bulk Metallic Glasses, J. Mater. Eng. Perform. 23, 6 (2014)

M AN U

2241-2246.

[7] K. Yu, L.J. Chen , J. Zhao, S.J. Li, Y.L. Dai, Q. Huang, Z.M. Yu. In vitro corrosion behavior and in vivo biodegradation of biomedical β-Ca3(PO4)2/Mg–Zn composites, Acta Biomater. 8, 7 (2012) 2845-2855

TE D

[8] F.X. Qin, G.Q. Xie, Z.H. Dan, S.L. Zhu, I. Seki. Corrosion behavior and mechanical properties of Mg-Zn-Ca amorphous alloys, Intermetallics. 42 (2013) 9-13. [9] Y. Dai, J.H. Li, B.X. Liu. First-principles molecular dynamics simulations to study the

EP

crystal-to-amorphous transition in the Mg-Zn system, Intermetallics. 29, (2012) 75-79.

AC C

[10] M.K. Datta, D. Chou, D. Hong, P. Saha, S.J. Chung, B.Lee, A. Sirinterlikci, M. Ramanathan, A. Roy, P.N. Kumta. Structure and thermal stability of biodegradable Mg– Zn–Ca based amorphous alloys synthesized by mechanical alloying, Mater. Sci. Eng. B. 17, 620 (2011)1637-1643.

[11] J.F. Wang, S. Huang, Y.Y. Wei, S.F. Guo, F.S. Pan. Enhanced mechanical properties and corrosion resistance of a Mg–Zn–Ca bulk metallic glass composite by Fe particle addition, Mater. Lett. 91 (2013) 311–314.

ACCEPTED MANUSCRIPT [12] C. Yang, L.H. Liu, Y.G. Yao, Y.H. Li, Y.Y. Li. Intrinsic relationship between crystallization mechanism of metallic glass powder and microstructure of bulk alloys fabricated by powder consolidation and crystallization of amorphous phase, J. Alloys Compd. 586 (2014)

RI PT

542–548. [13] G.T. Bae, S. B. Lee, N.J. Kim. Effect of alloying elements on the crystallization kinetics of Mg–Cu–(Y, Gd) bulk amorphous alloys, Mater. Sci. Eng., A. 449–451 (2007) 489–492.

SC

[14] L.J. Huang, L. Li, G.Y. Liang, Y.L. Guo, D.C. Wu. Crystallization kinetics of

M AN U

Mg65Cu25Nd10 amorphous alloy, J. Non-Cryst. Solids. 354 (2008) 1048–1053. [15] Y.D. Sun, P. Shen, Z.Q. Li, J.S. Liu, M.Q. Cong, M. Jiang. Kinetics of crystallization process of Mg–Cu–Gd based bulk metallic glasses, J. Non-Cryst. Solids. 358 (2012) 1120–1127.

TE D

[16] B. Gun, K.J. Laws, M. Ferry. Static and dynamic crystallization in Mg–Cu–Y bulk metallic glass, J. Non-Cryst. Solids. 352 (2006) 3887–3895. [17] J. Mu, H.F. Zhang. Glass forming ability and crystallization kinetics of Al-Mg-Ni-La

EP

metallic glasses, Adv. Mater. Res. 960-961(2014)161-164.

AC C

[18] C. Peng, Z.H. Chen, X.Y. Zhao, A.L. Zhang, L.K. Zhang, D. Chen. . Crystallization kinetics of Zr60Cu25Fe5Al10 bulk metallic glass, J. Non-Cryst. Solids. 405, 11 (2014)7-11.

[19] H.E. Kissinger. Variation of peak temperature with heating rate in differential thermal analysis, J. Res. Nat. Bur. Stand. 57, 4 (1956) 217-221.

[20] H.E. Kissinger, A. Chem. Reaction kinetics in differential thermal analysis, Anal. Chem. 29, 11 (2002)1702-1706. [21] Á. Révész, Á. Kis-Tóth, L.K. Varga, J.L. Lábár, T. Spassov. High glass forming ability

ACCEPTED MANUSCRIPT correlated with microstructure and hydrogen storage properties of a Mg-Cu-Ag-Y glass. Int. J. Hydrogen Energy. 39, 17(2014) 9230-9240. [22] Andrey A. Tsarkov, Erzhena N. Zanaeva, Alexander Yu. Churyumov, Sergey V. Ketov,

RI PT

Dmitri V. Louzguine-Luzgin. Crystallization kinetics of Mg–Cu–Yb–Ca–Ag metallic glasses. Mater. Charact. 111 (2016) 75–80.

[23] M. Krasnowski, T. Kulik. Nanocrystalline or amorphous matrix Al60Fe15Ti15(Co/Mg/Zr)5

SC

-5%B composites produced by consolidation of mechanically alloyed powders -

M AN U

lightweight materials with high hardness, Intermetallics. 28 (2012) 120-127. [24] J. H. Flynn. The isoconversional method for determination of energy of activation at constant heating rates. J. Therm. Anal. 27, 1 (1983) 95-102.

[25] J.A. Augis，J.E. Bennett. Calculation of Avrami parameters for heterogeneous solid state

283-292.

TE D

reactions using a modification of the Kissinger method, J. Therm. Anal. 13, 2 (1978)

[26] J. Málek. The applicability of Johnson-Mehl-Avrami model in the thermal analysis of the

AC C

EP

crystallization kinetics of glasses, Thermochim. Acta. 267 (1995) 61-73.

ACCEPTED MANUSCRIPT Fig. 1. XRD pattern of Mg60Zn30Ti5Si5 alloy prepared by MA. Fig. 2. XRD pattern of Mg60Zn30Ti5Si5 alloys heat-treated at different temperatures. Fig. 3. Continuous heating DSC curves of Mg60Zn30Ti5Si5 amorphous alloy at different heating

RI PT

rates. Fig. 4. Plots and fitting lines of three peaks derived from Kissinger theory.

Fig. 5. Variation of Crystallized fraction (x) with temperature (a) peak 1, (b) peak 2, (c) peak 3.

AC C

EP

TE D

M AN U

SC

Fig. 6. Local activation energy (El) at different Crystallized fraction (x).

ACCEPTED MANUSCRIPT Table 1 Temperature values of crystallization peaks (Tp) obtained from DSC curves. Tp1 /K

Tp2 /K

Tp3 /K

5

510.8

566.2

716.4

10

527

582

732.4

20

544.2

597.8

752.7

30

552.2

606.6

764.8

AC C

EP

TE D

M AN U

SC

RI PT

HeatingRate / K·min-1

ACCEPTED MANUSCRIPT Table 2 values of ∆T and n for each exothermic peak. ∆T1

∆T2

∆T3

n1

n2

n3

5

95.1

34.5

35.2

0.625

1.656

1.955

10

101

20.9

35.3

0.626

2.889

2.038

20

94.1

26.9

39.2

0.717

2.368

1.938

30

71.7

35.7

37.2

0.968

1.837

2.109

AC C

EP

TE D

M AN U

SC

RI PT

Heating Rate / K·min-1

ACCEPTED MANUSCRIPT

20

40

60

2θ /

ο

RI PT

Cps

Mg Si

80

AC C

EP

TE D

M AN U

SC

Fig. 1. XRD pattern of Mg60Zn30Ti5Si5 alloy prepared by MA.

ACCEPTED MANUSCRIPT α-Mg

Si Zn

Mg2Si MgZn2

873 K

Cps

773 K

RI PT

673 K 573 K 523 K 473 K 423 K

40

60 2θ /

ο

80

SC

20

AC C

EP

TE D

M AN U

Fig. 2. XRD pattern of Mg60Zn30Ti5Si5 alloys heat-treated at different temperatures.

ACCEPTED MANUSCRIPT

Heat flow/ (µV/mg)

EXO

10 K/min 20 K/min 30 K/min

Tp3

500

Tp2 600

Temperature / K

700

800

SC

Tp1 400

RI PT

5 K/min

M AN U

Fig. 3. Continuous heating DSC curves of Mg60Zn30Ti5Si5 amorphous alloy at different heating

AC C

EP

TE D

rates.

ACCEPTED MANUSCRIPT

Fit lines Peak 1 Eα1=91.25KJ/mol

Peak 2 Eα2=116.56KJ/mol

-10

Peak 3 Eα3=154.95KJ/mol -11

-2.2

-2.0

-1.8 -1

M AN U

-10000/Tp (K )

-1.6

-1.4

SC

-2.4

RI PT

2

ln ( β /Tp )

-9

AC C

EP

TE D

Fig. 4. Plots and fitting lines of three peaks derived from Kissinger theory.

ACCEPTED MANUSCRIPT

100

80

RI PT

60

40

5 K/min 10 K/min 20 K/min 30 K/min

20

0 450

500

550

100

60

40

0 500

TE D

Crystallized fraction (α) / (%)

(b) 80

20

600

650

M AN U

Temperature /K

SC

Crystallized fraction (α) / (%)

(a)

550

600

650

700

800

850

EP

Temperature /K

(c)

AC C

Crystallized fraction (α) / (%)

100

80

60

40

20

0 650

700

750

Temperature /K

Fig. 5. Variation of Crystallized fraction (x) with temperature (a) peak 1, (b)peak 2, (c) peak 3.

ACCEPTED MANUSCRIPT

peak1 peak2 peak3

200

RI PT

El / (kJ/mol)

150

100

0 0

20

40

60

SC

50

80

100

M AN U

Conversion fraction / %

AC C

EP

TE D

Fig. 6. Local activation energy (El) at different Crystallized fraction (x).

ACCEPTED MANUSCRIPT This study focuses on crystallization kinetics of Mg-Zn-Ti-Si amorphous alloy which was rarely studied. MgZn2, Mg2Si and Zn were crystallized in turn during heating process.

Ea1.

RI PT

Apparent activation energies (Ea) of crystallizations were calculated and Ea3＞Ea2＞

Local activation energy for 1st crystallization (El1) decreases while El2 increases with crystallizing; El3 is almost stable during heating.

AC C

EP

TE D

M AN U

precipitations is diffusion controlled growth.

SC

The first crystallization is interface controlled growth; the mechanism of later