- Email: [email protected]
Glass-transition process in an Au-based metallic glass
Journal of Non-Crystalline Solids 419 (2015) 12–15
Contents lists available at ScienceDirect
Journal of Non-Crystalline Solids journal homepage: www.elsevier.com/ locate/ jnoncrysol
Glass-transition process in an Au-based metallic glass Dmitri V. Louzguine-Luzgin a,⁎, Ichiro Seki b, Sergey V. Ketov a, Larissa V. Louzguina-Luzgina a,c, Vladislav I. Polkin c, Na Chen a, Hans Fecht d, Alexander N. Vasiliev e,c, Hitoshi Kawaji f a
WPI Advanced Institute for Materials Research, Tohoku University, Katahira 2-1-1, Aoba-Ku, Sendai 980-8577, Japan Institute for Materials Research, Tohoku University, Katahira 2-1-1, Aoba-Ku, Sendai 980-8577, Japan National University of Science and Technology “MISiS”, Moscow 119049, Russia d Institute of Micro- and Nanomaterials, University of Ulm, 89081 Ulm, Germany e Low Temperature Physics Department, Faculty of Physics, Moscow State University, 119992 Moscow, Russia f Tokyo Institute of Technology, Interdisciplinary Graduate School of Science and Engineering, 4259 Nagatsuta-cho, Midori-ku, Yokohama 226-8502, Japan b c
a r t i c l e
i n f o
Article history: Received 22 January 2015 Received in revised form 9 March 2015 Accepted 16 March 2015 Available online xxxx Keywords: Bulk metallic glass; Amorphous alloy; Nanoscale; Devitrification; Phase transformation
a b s t r a c t In the present paper, the glass-transition phenomenon in an Au-based metallic glass has been studied using a step-scan calorimetry measurement. The existence of two distinct slopes within the glass-transition region one starting at low temperature (about 340 K) and the other at higher temperature (about 380 K) likely indicates two glass-transition processes. This phenomenon is rather related to different diffusion coefficients of the alloying elements in this complex alloy in solid state. Structural relaxation of the glassy phase before reaching the glass-transition region also shows a complex behavior. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Apart from the amorphous materials obtained by thin film deposition [1] the first metallic glassy alloy (or metallic glass) [2], namely Au–Si, was produced by rapid solidification of the melt in 1960 [3]. Ternary Au–Si–Ge [4] and Au–Pb–Sb [5] alloys developed later demonstrated a better glassforming ability (GFA). Later other noble metal-based metallic glasses were also produced including bulk metallic glasses [6]. Bulk metallic glasses (BMGs) also called bulk metallic glassy alloys defined as 3dimensional massive glassy objects with a size of not less than 1 mm in every spatial dimension have been produced during the last 20 years in the thickness range of 100–102 mm by using various casting processes [7–11]. These materials having a non-crystalline structure [12] exhibit high strength [13], good wear and corrosion resistance [11,14]. The transition from liquid to glassy state taking place at the glass transition temperature (Tg) [15,16] is observed universally in various types of liquids at sufficiently fast cooling when crystallization is suppressed [17,18]. The process of vitrification in Pd-based BMGs was recently studied in-situ and intensification of covalent bonding with P atoms was found close to the glass-transition [19]. Glass formation is a complex phenomenon. Although, the glass-transition phenomenon in metallic glasses has been studied extensively [20–22] there are still ⁎ Corresponding author. E-mail address: [email protected] (D.V. Louzguine-Luzgin).
http://dx.doi.org/10.1016/j.jnoncrysol.2015.03.018 0022-3093/© 2015 Elsevier B.V. All rights reserved.
considerable uncertainties in this field of materials science [23,24]. There are several gaps in obtaining a clear picture of glass transition, especially in metallic liquids. For example, a thermodynamic treatment of glass transition has been suggested by Kauzmann [25]. Several theories are used to describe glass transition [26–28] though none is really comprehensive so far. In order to further study the phenomenon of glass transition, i.e. the transformation of a liquid into the glassy state and vise versa, in the present work we investigated the glass transition phenomenon in a Au49Cu26.9Ag5.5Pd2.3Si16.3 glassy alloy developed earlier [29]. In order to minimize the kinetic effects associated with structural relaxation upon continuous heating, we measured the specific heat capacity as a material property, on annealing the glassy alloy in a step scan mode. Step scan measurements allow equilibrating the glassy phase before each subsequent measurement [30]. The studied alloy has a sufficiently high glass-forming ability and has good stability of a supercooled liquid for such an investigation. 2. Experimental procedure The ingot of a Au49Cu26.9Ag5.5Pd2.3Si16.3 alloy (the composition is given in nominal at.%) was prepared by arc-melting mixtures of high purity elemental metals having more than 99.9 mass% purity in an argon atmosphere. From this ingot, glassy ribbon samples of about 20 μm in thickness and 1 mm in width were prepared by rapid
D.V. Louzguine-Luzgin et al. / Journal of Non-Crystalline Solids 419 (2015) 12–15
13
solidification of the melt on a single copper roller at a tangential wheel velocity of 40 m/s. Bulk glassy samples contain a surface layer of crystalline phases [31] which colored their surface in gold, and were not used in the present work. No visible presence of such a surface layer (unless it is extremely thin) was observed in ribbon samples. Then the glass-transition phenomenon was studied on continuous heating in a Seiko DSC and Perkin-Elmer Diamond DSC in a step scan mode. The glassy samples were heated and cooled at a rate of 83 mK/s (5 K/min), while the annealing (waiting) time between the steps was maintained at 60 s (1 min). The specific heat capacity at constant pressure (Cp) was measured by using heat release (ΔH) at each heating/cooling step. 3. Results
3
HF 4
HF 2 1
Fig. 2. XRD patterns of the as-solidified and annealed samples. The inset is a close up of the first diffraction maximum.
difference between the values of Cp at 410 and 340 K) is close to that of the Au81.4Si18.6 alloy [35] and higher than that of some other glasses [36]. At step scan mode crystallization starts above 430 K. On crystallization Cp decreases to a reduced value close to that of the initial glassy state being extrapolated to a higher temperature (see triangle symbols in Fig. 3). On cooling there is a weak hysteresis in the supercooled liquid state (the cooling curve is located somewhat lower in Cp values than the heating one), but in general, both heating and cooling cycles follow nearly the same trend showing a double-stage glass-transition. After that the sample was heated once again in the step scan mode, this time above the crystallization temperature (the sample crystallized above 430 K) and Cp was measured once again in a crystalline state (Fig. 3). After that the sample was melted in the DSC cell by heating up to 773 K, and then the Cp of a liquid was measured once again in the step scan mode on cooling at 83 mK/s to room temperature (Fig. 4). These experiments produced Cp values of the Au49Cu26.9Ag5.5Pd2.3Si16.3 alloy in glassy, crystalline, liquid and supercooled liquid states. In order to calculate the difference in entropy (ΔSl − c) between liquid and crystalline phases the temperature dependence of the
45 Glass Heating Cp Glass Cooling Cp
40
Crystal Heating Cp
Cp (J/molK)
Heat flow diffrence (arbitrary unit)
Heat Flow (arb.unit), exothermic
Fig. 1 shows the heat flow variation of the studied alloy on continuous heating as a function of temperature at a heating rate of 0.67 K/s. An enlarged plot up to 424 K indicates a possible double-stage structural relaxation shielding the glass-transition region visible at the first DSC run and when compared to the second run. In order to rule out possible crystallization the Au49Cu26.9Ag5.5Pd2.3 Si16.3 glassy sample was annealed at 400 K close to the glass-transition region and no clear changes in the X-ray diffraction pattern were found. There is a slight decrease in the half width of the first diffraction maximum as a result of structural relaxation below Tg (Fig. 2). Specific heat capacity (Cp) was measured in a step-scan mode on heating of an initial glassy phase up to above Tg and subsequent cooling to room temperature at 5 K/min (83 mK/s) with an annealing time of 60 s before the subsequent heating or cooling cycle (Fig. 3). The stepscan measurement allowed to get rid of the effect of structural relaxation (Fig. 1) as much as possible and to study the behavior of the glassy phase itself. The Cp of the glass increases gradually with increasing temperature from about 24 J/mol·K at room temperature. In the glasstransition region the slope of the plot changes twice: the first time at about 340 K and the second time at about 380 K. It may indicate two distinctly different glass-transition processes. On further heating the specific heat reaches a maximum of 40.5 J/mol·K in the supercooled liquid state and remains nearly constant in this region before the sample starts to crystallize. This value is close to that obtained for liquid copper near its melting temperature [32] and Zr-based bulk glass-forming alloys [33] in the supercooled liquid state but lower than those reported for Au–Pb–Sb [34]. The excess heat capacity of the supercooled liquid of about 15 J/mol·K (as a
35
Sigmoidal 2
30
Sigmoidal 1
25
300
350
400
450
500
Temperature (K) 20 Fig. 1. DSC trace scanned at 0.67 K/s (heat flow — HF) and a magnified 10× difference in heat flow signals (ΔHF) between the 1st and 2nd DSC heating experiments to the glasstransition region up to 424 K. 1st measurement — initial sample, 2nd measurement — the sample pre-heated to 424 K and cooled down within the DSC furnace at the cooling rate in the order of 1 K/s. 3rd measurement is continuous DSC scan up to crystallization. The onset crystallization temperature is 467 K.
300
350
400
450
Temperature (K) Fig. 3. Cp of glassy and crystalline phases measured on heating and on cooling (crystalline only on heating) as indicated. The heating and cooling rates were 83 mK/s. The annealing time between each heating and cooling step was 60 s.
14
D.V. Louzguine-Luzgin et al. / Journal of Non-Crystalline Solids 419 (2015) 12–15
Fig. 4. Cp of glassy (Cpg), liquid (Cpl) and crystalline (Cpc) phases measured in a step scan mode on heating and on cooling at 83 mK/s as shown with symbols. Solid lines represent fitting curves of the heat capacity of liquid (Cpl) and crystalline (Cpc) phases measured on cooling.
Cp of a liquid (Clp ), including the point for a supercooled liquid of 40.5 J/mol·K at 423 K, and crystalline (Ccp) phases were fitted with the following functions [37]: −2
l
Cp ¼ 3R þ a1 T þ b1 T
;
ð1Þ
and 2
c
Cp ¼ 3R þ a2 T þ b2 T ;
ð2Þ
where R is the gas constant while a and b are fitting parameters. Eq. (1) was found to describe well the temperature dependence of the heat capacity of a liquid phase, but the fit with Eq. (2) was not successful for C cp while a good one was obtained if a variable c is used instead of 3R. The value of ΔSl − c was calculated using the following formula:
l−c
ΔS
ZTl ðTÞ ¼ ΔSm −
ΔCl−c p ðTÞ dT T
ð3Þ
T
while Sm, the entropy of melting of 8.34 J/mol·K, was determined from the enthalpy of melting (Hm) equal to 5.278 kJ/mol. By using a DSC scan Tl was estimated at 632 K. The resulting values of ΔSl − are also plotted in Fig. 4. Thus the Kauzmann temperature was estimated at 340 K.
found at about 340 K (Fig. 3). This temperature is close to the calculated Kauzmann temperature for the entropy crisis (Fig. 4). A very close value of the Kauzmann temperature was recently obtained by another group [39,40]. The Cp of the supercooled liquid was found to rise on cooling as found for many metallic liquids [41,42] leading to a quick reduction in ΔSl − c. The changes in slope may be explained on the basis of the different diffusivities of the constituent elements in the glass. At low temperatures, diffusion of atoms by co-operative shearing is several orders of magnitude slower than atomic hopping [43]. At higher temperatures, however, co-operative diffusion of atoms becomes comparable to the hopping time. There are no data for diffusion of Si in crystalline Au but in Fig. 5 one can see that Ge (an element having similar properties to Si) has a two-orders in magnitude higher diffusion coefficient compared to the self-diffusion of Au in a wide temperature range [44]. Tracer diffusion of Cu is also notably faster than that of Au. Therefore the time scale for the shear of Si and Cu atoms becomes comparable to the hopping time and the liquid-like motion is predominant for these atoms. pffiffiffiffiffiffi According to the formula (l ¼ 2 Dt) the diffusion length l is proportional to the diffusion coefficient D, while t is time. Although atomic diffusion in the glassy state is significantly faster than in crystals the relative difference between the diffusion coefficients is similar. In other words elements which diffuse faster in crystalline alloys act in the same way in glasses, in general. For example, Cu and Ni are quickly diffusing elements both in crystalline Zr [45] and in Zr-based metallic glass [46], while Al has an intermediate diffusion coefficient followed by the most slowly diffusing element Zr. The above-mentioned simple analysis may indicate that each element: Au, Si and Cu (in a local group of atoms) undergoes glass transition at a certain range below which it moves in a solid-like manner and the glass transition in a multicomponent system does not occur uniformly but gradually over a temperature range reflecting the inhomogeneity at the atomic level [47]. The result is qualitatively similar to that observed for Zr-based metallic glass [48]. Ag and Pd are noble metals as well as Au. They are soluble in crystalline Au, and have atomic sizes similar to that of Au. Thus, their effect on the glass-transition process is likely the same as that of the major component Au. Moreover, the high stability of the nanostructured glassy sputtered Au-based structure observed recently [49,50] may be owing to the fact that TK is close to room temperature. The high stability of metallic glasses sputtered close to TK was found in Ref. [51]. These results are consistent with that obtained in the present research work.
4. Discussion Fig. 1 indicates a possible double-stage structural relaxation shielding the glass-transition region. A structural relaxation process consists of two stages: one starting at about 310 K another at about 380 K (Fig. 1). The first exothermic peak starting at 310 K may be related to structural relaxation and densification of the sample owing to a decrease in specific volume. The second exothermic peak is very close to Tg and its nature is not clear. Similar structural complex relaxation processes have been recently observed in the Cu55Hf25Ti15Pd5 alloy [38] and a double-stage variation of a relaxation signal was observed. As shown in Fig. 3, the changes in the slope of Cp are clearly observed within the glass-transition region between the glassy phase and liquid phase during heating and cooling. Instead of a single one it is well fitted with two sigmoid functions. These two regions rather indicate that there are two glass-transition processes: one taking place at a low temperature and the other at a high temperature and from Fig. 3 this process is reversible: the same effect is detected on cooling. The initial visible increase in Cp owing to the beginning of glass-transition was
Fig. 5. Self diffusion of Au and tracer diffusion of Cu and Ge in Au (extrapolated to wider range).
D.V. Louzguine-Luzgin et al. / Journal of Non-Crystalline Solids 419 (2015) 12–15
5. Conclusions The glass-transition phenomenon in Au49Cu26.9Ag5.5Pd2.3Si16.3 metallic glass has been studied using the step-scan calorimetry. The clear existence of two different slopes within the glass-transition region seems to indicate the occurrence of two separate glass-transition processes. These two processes one starting on heating at low temperatures around 340 K and the other at high temperatures around 380 K, indicate two independent glass-transition processes. This phenomenon is likely related to the different diffusion coefficients of the alloying elements in this alloy. Acknowledgements The authors gratefully acknowledge the financial support of the World Premier International Research Center Initiative (WPI), Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan, and in part by the Ministry of Education and Science of the Russian Federation in the framework of the Program to Increase the Competitiveness of NUST “MISiS” (No. К2-2014-013) and the European Marie Curie program VitriMetTech. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14]
[15] [16]
W. Buckel, R. Hilsch, Z. Phys. 138 (1954) 109. A.L. Greer, Science 267 (1995) 1947. W. Klement, R.H. Willens, P. Duwez, Nature 187 (1960) 869. H.S. Chen, D. Turnbull, Appl. Phys. Lett. 10 (1967) 284. H.J. Fecht, J.H. Perepezko, M.C. Lee, W.L. Johnson, J. Appl. Phys. 68 (1990) 4494. H.S. Chen, Acta Metall. 22 (1974) 1505. H.W. Kui, A.L. Greer, D. Turnbull, Appl. Phys. Lett. 45 (1982) 716. A. Inoue, Mater. Trans. JIM 36 (1995) 866. W.L. Johnson, MRS Bull. 24 (1999) 42. A. Inoue, Acta Mater. 48 (2000) 279. C. Suryanarayana, A. Inoue, Bulk Metallic Glasses, CRC Press, Boca Raton, FL, 2010. D.V. Louzguine-Luzgin, A.I. Bazlov, A.Yu. Churyumov, K. Georgarakis, A.R. Yavari, Phys. Solid State 55 (2013) 1985. D.V. Louzguine-Luzgin, L.V. Louzguina-Luzgina, A.Yu. Churyumov, Metals 3 (2013) 1. D.V. Louzguine-Luzgin, A. Inoue, Bulk metallic glasses. Formation, structure, properties, and applications, in: K.H.J. Buschow (Ed.)Handbook of Magnetic Materials, 21, Elsevier, 2013, p. 131. D. Turnbull, Contemp. Phys. 10 (1969) 473. D.S. Sanditov, S.Sh. Sangadiev, D.N. Tsydenova, Russ. J. Phys. Chem. 80 (2006) 634.
15
[17] P.G. Debenedetti, F.H. Stillinger, Nature 410 (2001) 259. [18] S. D. Kaloshkin and I. A. Tomilin, 280-281, (1996) 303. [19] D.V. Louzguine-Luzgin, R. Belosludov, A.R. Yavari, K. Georgarakis, G. Vaughan, Y. Kawazoe, T. Egami, A. Inoue, J. Appl. Phys. 110 (2011) 043519. [20] C.A. Angell, J. Non-Cryst. Solids 102 (1988) 205. [21] T. Egami, J. Non-Cryst. Solids 317 (2003) 30. [22] M.I. Ojovan, J. Non-Cryst. Solids 382 (2013) 79. [23] P.W. Anderson, Science 267 (1995) 1615. [24] D.V. Louzguine-Luzgin, J. Alloys Compd. 586 (2014) S2. [25] W. Kauzmann, Chem. Rev. 43 (1948) 219. [26] C.A. Angell, The glass transition, Pergamon Encyclopedia of Materials: Science and Technology, vol. 4, 2001, p. 3365. [27] E. Leutheusser, Phys. Rev. A 29 (1984) 2765. [28] J.K. Kruger, Lect. Notes Phys. 716 (2007) 61. [29] J. Schroers, B. Lohwongwatana, W.L. Johnson, A. Peker, Appl. Phys. Lett. 87 (2005) 061912. [30] Y. Moriya, T. Yoshida, H. Kawaji, T. Atake, M. Fukuhara, H. Kimura, A. Inoue, Mater. Sci. Eng. B 148 (2008) 207. [31] S.V. Ketov, N. Chen, A. Caron, A. Inoue, D.V. Louzguine-Luzgin, Appl. Phys. Lett. 101 (2012) 241905. [32] V.Ya. Chekhovskoi, V.D. Tarasov, Yu.V. Gusev, High Temperature 38 (2000) 394. [33] J. Bünz, T. Brink, K. Tsuchiya, F. Meng, G. Wilde, K. Albe, Phys. Rev. Lett. 112 (2014) 135501. [34] R. Busch, JOM 52 (2000) 39. [35] G.D. Fontana, L. Battezzati, Acta Mater. 61 (2013) 2260. [36] H.B. Ke, P. Wen, W.H. Wang, AIP Adv. 2 (2012) 041404. [37] S.C. Glade, R. Busch, D.S. Lee, W.L. Johnson, R.K. Wunderlich, H.J. Fecht, J. Appl. Phys. 87 (2000) 7242. [38] D.V. Louzguine-Luzgin, I. Seki, T. Yamamoto, H. Kawaji, C. Suryanarayana, A. Inoue, Intermetallics 23 (2012) 177. [39] G. Fiore, I. Ichim, L. Battezzati, J. Non-Cryst. Solids 356 (2010) 2218. [40] G.D. Fontana, G.L. Fiore, L. Battezzati, J. Non-Cryst. Solids 382 (2013) 95. [41] A.L. Greer, J. Less-Common Met. 145 (1988) 131. [42] J.H. Perepezko, J.S. Paik, J. Non-Cryst. Solids 61–62 (1984) 113. [43] M. Dzugutov, Europhys. Lett. 26 (1994) 533. [44] W.F. Gale, T.C. Totemeier (Eds.), Smithells Metals Reference Book, 8th editionElsevier Butterworth-Heinemann Ltd., Oxford UK, 2004, pp. 13–17. [45] R.A. Perez, H. Nakajima, F. Dyment, Mater. Trans. 44 (2003) 2. [46] F. Faupel, W. Frank, M.-P. Macht, H. Mehrer, V. Naundorf, K. Rätzke, H.R. Schober, S.K. Sharma, H. Teichler, Rev. Mod. Phys. 75 (2003) 237. [47] T. Egami, Mater. Trans. 43 (2002) 510. [48] D.V. Louzguine-Luzgin, I. Seki, T. Yamamoto, H. Kawaji, C. Suryanarayana, A. Inoue, Phys. Rev. B 81 (2010) 144202. [49] N. Chen, R. Frank, N. Asao, D.V. Louzguine-Luzgin, P. Sharma, J.Q. Wang, G.Q. Xie, Y. Ishikawa, N. Hatakeyama, Y.C. Lin, M. Esashi, Y. Yamamoto, A. Inoue, Acta Mater. 59 (2011) 6433. [50] J.Q. Wang, N. Chen, P. Liu, Z. Wang, D.V. Louzguine-Luzgin, M.W. Chen, J.H. Perepezko, Acta Mater. 79 (2014) 30. [51] S. Singh, M.D. Ediger, J.J. Pablo, Nat. Mater. 12 (2013) 139.
Contents lists available at ScienceDirect
Journal of Non-Crystalline Solids journal homepage: www.elsevier.com/ locate/ jnoncrysol
Glass-transition process in an Au-based metallic glass Dmitri V. Louzguine-Luzgin a,⁎, Ichiro Seki b, Sergey V. Ketov a, Larissa V. Louzguina-Luzgina a,c, Vladislav I. Polkin c, Na Chen a, Hans Fecht d, Alexander N. Vasiliev e,c, Hitoshi Kawaji f a
WPI Advanced Institute for Materials Research, Tohoku University, Katahira 2-1-1, Aoba-Ku, Sendai 980-8577, Japan Institute for Materials Research, Tohoku University, Katahira 2-1-1, Aoba-Ku, Sendai 980-8577, Japan National University of Science and Technology “MISiS”, Moscow 119049, Russia d Institute of Micro- and Nanomaterials, University of Ulm, 89081 Ulm, Germany e Low Temperature Physics Department, Faculty of Physics, Moscow State University, 119992 Moscow, Russia f Tokyo Institute of Technology, Interdisciplinary Graduate School of Science and Engineering, 4259 Nagatsuta-cho, Midori-ku, Yokohama 226-8502, Japan b c
a r t i c l e
i n f o
Article history: Received 22 January 2015 Received in revised form 9 March 2015 Accepted 16 March 2015 Available online xxxx Keywords: Bulk metallic glass; Amorphous alloy; Nanoscale; Devitrification; Phase transformation
a b s t r a c t In the present paper, the glass-transition phenomenon in an Au-based metallic glass has been studied using a step-scan calorimetry measurement. The existence of two distinct slopes within the glass-transition region one starting at low temperature (about 340 K) and the other at higher temperature (about 380 K) likely indicates two glass-transition processes. This phenomenon is rather related to different diffusion coefficients of the alloying elements in this complex alloy in solid state. Structural relaxation of the glassy phase before reaching the glass-transition region also shows a complex behavior. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Apart from the amorphous materials obtained by thin film deposition [1] the first metallic glassy alloy (or metallic glass) [2], namely Au–Si, was produced by rapid solidification of the melt in 1960 [3]. Ternary Au–Si–Ge [4] and Au–Pb–Sb [5] alloys developed later demonstrated a better glassforming ability (GFA). Later other noble metal-based metallic glasses were also produced including bulk metallic glasses [6]. Bulk metallic glasses (BMGs) also called bulk metallic glassy alloys defined as 3dimensional massive glassy objects with a size of not less than 1 mm in every spatial dimension have been produced during the last 20 years in the thickness range of 100–102 mm by using various casting processes [7–11]. These materials having a non-crystalline structure [12] exhibit high strength [13], good wear and corrosion resistance [11,14]. The transition from liquid to glassy state taking place at the glass transition temperature (Tg) [15,16] is observed universally in various types of liquids at sufficiently fast cooling when crystallization is suppressed [17,18]. The process of vitrification in Pd-based BMGs was recently studied in-situ and intensification of covalent bonding with P atoms was found close to the glass-transition [19]. Glass formation is a complex phenomenon. Although, the glass-transition phenomenon in metallic glasses has been studied extensively [20–22] there are still ⁎ Corresponding author. E-mail address: [email protected] (D.V. Louzguine-Luzgin).
http://dx.doi.org/10.1016/j.jnoncrysol.2015.03.018 0022-3093/© 2015 Elsevier B.V. All rights reserved.
considerable uncertainties in this field of materials science [23,24]. There are several gaps in obtaining a clear picture of glass transition, especially in metallic liquids. For example, a thermodynamic treatment of glass transition has been suggested by Kauzmann [25]. Several theories are used to describe glass transition [26–28] though none is really comprehensive so far. In order to further study the phenomenon of glass transition, i.e. the transformation of a liquid into the glassy state and vise versa, in the present work we investigated the glass transition phenomenon in a Au49Cu26.9Ag5.5Pd2.3Si16.3 glassy alloy developed earlier [29]. In order to minimize the kinetic effects associated with structural relaxation upon continuous heating, we measured the specific heat capacity as a material property, on annealing the glassy alloy in a step scan mode. Step scan measurements allow equilibrating the glassy phase before each subsequent measurement [30]. The studied alloy has a sufficiently high glass-forming ability and has good stability of a supercooled liquid for such an investigation. 2. Experimental procedure The ingot of a Au49Cu26.9Ag5.5Pd2.3Si16.3 alloy (the composition is given in nominal at.%) was prepared by arc-melting mixtures of high purity elemental metals having more than 99.9 mass% purity in an argon atmosphere. From this ingot, glassy ribbon samples of about 20 μm in thickness and 1 mm in width were prepared by rapid
D.V. Louzguine-Luzgin et al. / Journal of Non-Crystalline Solids 419 (2015) 12–15
13
solidification of the melt on a single copper roller at a tangential wheel velocity of 40 m/s. Bulk glassy samples contain a surface layer of crystalline phases [31] which colored their surface in gold, and were not used in the present work. No visible presence of such a surface layer (unless it is extremely thin) was observed in ribbon samples. Then the glass-transition phenomenon was studied on continuous heating in a Seiko DSC and Perkin-Elmer Diamond DSC in a step scan mode. The glassy samples were heated and cooled at a rate of 83 mK/s (5 K/min), while the annealing (waiting) time between the steps was maintained at 60 s (1 min). The specific heat capacity at constant pressure (Cp) was measured by using heat release (ΔH) at each heating/cooling step. 3. Results
3
HF 4
HF 2 1
Fig. 2. XRD patterns of the as-solidified and annealed samples. The inset is a close up of the first diffraction maximum.
difference between the values of Cp at 410 and 340 K) is close to that of the Au81.4Si18.6 alloy [35] and higher than that of some other glasses [36]. At step scan mode crystallization starts above 430 K. On crystallization Cp decreases to a reduced value close to that of the initial glassy state being extrapolated to a higher temperature (see triangle symbols in Fig. 3). On cooling there is a weak hysteresis in the supercooled liquid state (the cooling curve is located somewhat lower in Cp values than the heating one), but in general, both heating and cooling cycles follow nearly the same trend showing a double-stage glass-transition. After that the sample was heated once again in the step scan mode, this time above the crystallization temperature (the sample crystallized above 430 K) and Cp was measured once again in a crystalline state (Fig. 3). After that the sample was melted in the DSC cell by heating up to 773 K, and then the Cp of a liquid was measured once again in the step scan mode on cooling at 83 mK/s to room temperature (Fig. 4). These experiments produced Cp values of the Au49Cu26.9Ag5.5Pd2.3Si16.3 alloy in glassy, crystalline, liquid and supercooled liquid states. In order to calculate the difference in entropy (ΔSl − c) between liquid and crystalline phases the temperature dependence of the
45 Glass Heating Cp Glass Cooling Cp
40
Crystal Heating Cp
Cp (J/molK)
Heat flow diffrence (arbitrary unit)
Heat Flow (arb.unit), exothermic
Fig. 1 shows the heat flow variation of the studied alloy on continuous heating as a function of temperature at a heating rate of 0.67 K/s. An enlarged plot up to 424 K indicates a possible double-stage structural relaxation shielding the glass-transition region visible at the first DSC run and when compared to the second run. In order to rule out possible crystallization the Au49Cu26.9Ag5.5Pd2.3 Si16.3 glassy sample was annealed at 400 K close to the glass-transition region and no clear changes in the X-ray diffraction pattern were found. There is a slight decrease in the half width of the first diffraction maximum as a result of structural relaxation below Tg (Fig. 2). Specific heat capacity (Cp) was measured in a step-scan mode on heating of an initial glassy phase up to above Tg and subsequent cooling to room temperature at 5 K/min (83 mK/s) with an annealing time of 60 s before the subsequent heating or cooling cycle (Fig. 3). The stepscan measurement allowed to get rid of the effect of structural relaxation (Fig. 1) as much as possible and to study the behavior of the glassy phase itself. The Cp of the glass increases gradually with increasing temperature from about 24 J/mol·K at room temperature. In the glasstransition region the slope of the plot changes twice: the first time at about 340 K and the second time at about 380 K. It may indicate two distinctly different glass-transition processes. On further heating the specific heat reaches a maximum of 40.5 J/mol·K in the supercooled liquid state and remains nearly constant in this region before the sample starts to crystallize. This value is close to that obtained for liquid copper near its melting temperature [32] and Zr-based bulk glass-forming alloys [33] in the supercooled liquid state but lower than those reported for Au–Pb–Sb [34]. The excess heat capacity of the supercooled liquid of about 15 J/mol·K (as a
35
Sigmoidal 2
30
Sigmoidal 1
25
300
350
400
450
500
Temperature (K) 20 Fig. 1. DSC trace scanned at 0.67 K/s (heat flow — HF) and a magnified 10× difference in heat flow signals (ΔHF) between the 1st and 2nd DSC heating experiments to the glasstransition region up to 424 K. 1st measurement — initial sample, 2nd measurement — the sample pre-heated to 424 K and cooled down within the DSC furnace at the cooling rate in the order of 1 K/s. 3rd measurement is continuous DSC scan up to crystallization. The onset crystallization temperature is 467 K.
300
350
400
450
Temperature (K) Fig. 3. Cp of glassy and crystalline phases measured on heating and on cooling (crystalline only on heating) as indicated. The heating and cooling rates were 83 mK/s. The annealing time between each heating and cooling step was 60 s.
14
D.V. Louzguine-Luzgin et al. / Journal of Non-Crystalline Solids 419 (2015) 12–15
Fig. 4. Cp of glassy (Cpg), liquid (Cpl) and crystalline (Cpc) phases measured in a step scan mode on heating and on cooling at 83 mK/s as shown with symbols. Solid lines represent fitting curves of the heat capacity of liquid (Cpl) and crystalline (Cpc) phases measured on cooling.
Cp of a liquid (Clp ), including the point for a supercooled liquid of 40.5 J/mol·K at 423 K, and crystalline (Ccp) phases were fitted with the following functions [37]: −2
l
Cp ¼ 3R þ a1 T þ b1 T
;
ð1Þ
and 2
c
Cp ¼ 3R þ a2 T þ b2 T ;
ð2Þ
where R is the gas constant while a and b are fitting parameters. Eq. (1) was found to describe well the temperature dependence of the heat capacity of a liquid phase, but the fit with Eq. (2) was not successful for C cp while a good one was obtained if a variable c is used instead of 3R. The value of ΔSl − c was calculated using the following formula:
l−c
ΔS
ZTl ðTÞ ¼ ΔSm −
ΔCl−c p ðTÞ dT T
ð3Þ
T
while Sm, the entropy of melting of 8.34 J/mol·K, was determined from the enthalpy of melting (Hm) equal to 5.278 kJ/mol. By using a DSC scan Tl was estimated at 632 K. The resulting values of ΔSl − are also plotted in Fig. 4. Thus the Kauzmann temperature was estimated at 340 K.
found at about 340 K (Fig. 3). This temperature is close to the calculated Kauzmann temperature for the entropy crisis (Fig. 4). A very close value of the Kauzmann temperature was recently obtained by another group [39,40]. The Cp of the supercooled liquid was found to rise on cooling as found for many metallic liquids [41,42] leading to a quick reduction in ΔSl − c. The changes in slope may be explained on the basis of the different diffusivities of the constituent elements in the glass. At low temperatures, diffusion of atoms by co-operative shearing is several orders of magnitude slower than atomic hopping [43]. At higher temperatures, however, co-operative diffusion of atoms becomes comparable to the hopping time. There are no data for diffusion of Si in crystalline Au but in Fig. 5 one can see that Ge (an element having similar properties to Si) has a two-orders in magnitude higher diffusion coefficient compared to the self-diffusion of Au in a wide temperature range [44]. Tracer diffusion of Cu is also notably faster than that of Au. Therefore the time scale for the shear of Si and Cu atoms becomes comparable to the hopping time and the liquid-like motion is predominant for these atoms. pffiffiffiffiffiffi According to the formula (l ¼ 2 Dt) the diffusion length l is proportional to the diffusion coefficient D, while t is time. Although atomic diffusion in the glassy state is significantly faster than in crystals the relative difference between the diffusion coefficients is similar. In other words elements which diffuse faster in crystalline alloys act in the same way in glasses, in general. For example, Cu and Ni are quickly diffusing elements both in crystalline Zr [45] and in Zr-based metallic glass [46], while Al has an intermediate diffusion coefficient followed by the most slowly diffusing element Zr. The above-mentioned simple analysis may indicate that each element: Au, Si and Cu (in a local group of atoms) undergoes glass transition at a certain range below which it moves in a solid-like manner and the glass transition in a multicomponent system does not occur uniformly but gradually over a temperature range reflecting the inhomogeneity at the atomic level [47]. The result is qualitatively similar to that observed for Zr-based metallic glass [48]. Ag and Pd are noble metals as well as Au. They are soluble in crystalline Au, and have atomic sizes similar to that of Au. Thus, their effect on the glass-transition process is likely the same as that of the major component Au. Moreover, the high stability of the nanostructured glassy sputtered Au-based structure observed recently [49,50] may be owing to the fact that TK is close to room temperature. The high stability of metallic glasses sputtered close to TK was found in Ref. [51]. These results are consistent with that obtained in the present research work.
4. Discussion Fig. 1 indicates a possible double-stage structural relaxation shielding the glass-transition region. A structural relaxation process consists of two stages: one starting at about 310 K another at about 380 K (Fig. 1). The first exothermic peak starting at 310 K may be related to structural relaxation and densification of the sample owing to a decrease in specific volume. The second exothermic peak is very close to Tg and its nature is not clear. Similar structural complex relaxation processes have been recently observed in the Cu55Hf25Ti15Pd5 alloy [38] and a double-stage variation of a relaxation signal was observed. As shown in Fig. 3, the changes in the slope of Cp are clearly observed within the glass-transition region between the glassy phase and liquid phase during heating and cooling. Instead of a single one it is well fitted with two sigmoid functions. These two regions rather indicate that there are two glass-transition processes: one taking place at a low temperature and the other at a high temperature and from Fig. 3 this process is reversible: the same effect is detected on cooling. The initial visible increase in Cp owing to the beginning of glass-transition was
Fig. 5. Self diffusion of Au and tracer diffusion of Cu and Ge in Au (extrapolated to wider range).
D.V. Louzguine-Luzgin et al. / Journal of Non-Crystalline Solids 419 (2015) 12–15
5. Conclusions The glass-transition phenomenon in Au49Cu26.9Ag5.5Pd2.3Si16.3 metallic glass has been studied using the step-scan calorimetry. The clear existence of two different slopes within the glass-transition region seems to indicate the occurrence of two separate glass-transition processes. These two processes one starting on heating at low temperatures around 340 K and the other at high temperatures around 380 K, indicate two independent glass-transition processes. This phenomenon is likely related to the different diffusion coefficients of the alloying elements in this alloy. Acknowledgements The authors gratefully acknowledge the financial support of the World Premier International Research Center Initiative (WPI), Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan, and in part by the Ministry of Education and Science of the Russian Federation in the framework of the Program to Increase the Competitiveness of NUST “MISiS” (No. К2-2014-013) and the European Marie Curie program VitriMetTech. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14]
[15] [16]
W. Buckel, R. Hilsch, Z. Phys. 138 (1954) 109. A.L. Greer, Science 267 (1995) 1947. W. Klement, R.H. Willens, P. Duwez, Nature 187 (1960) 869. H.S. Chen, D. Turnbull, Appl. Phys. Lett. 10 (1967) 284. H.J. Fecht, J.H. Perepezko, M.C. Lee, W.L. Johnson, J. Appl. Phys. 68 (1990) 4494. H.S. Chen, Acta Metall. 22 (1974) 1505. H.W. Kui, A.L. Greer, D. Turnbull, Appl. Phys. Lett. 45 (1982) 716. A. Inoue, Mater. Trans. JIM 36 (1995) 866. W.L. Johnson, MRS Bull. 24 (1999) 42. A. Inoue, Acta Mater. 48 (2000) 279. C. Suryanarayana, A. Inoue, Bulk Metallic Glasses, CRC Press, Boca Raton, FL, 2010. D.V. Louzguine-Luzgin, A.I. Bazlov, A.Yu. Churyumov, K. Georgarakis, A.R. Yavari, Phys. Solid State 55 (2013) 1985. D.V. Louzguine-Luzgin, L.V. Louzguina-Luzgina, A.Yu. Churyumov, Metals 3 (2013) 1. D.V. Louzguine-Luzgin, A. Inoue, Bulk metallic glasses. Formation, structure, properties, and applications, in: K.H.J. Buschow (Ed.)Handbook of Magnetic Materials, 21, Elsevier, 2013, p. 131. D. Turnbull, Contemp. Phys. 10 (1969) 473. D.S. Sanditov, S.Sh. Sangadiev, D.N. Tsydenova, Russ. J. Phys. Chem. 80 (2006) 634.
15
[17] P.G. Debenedetti, F.H. Stillinger, Nature 410 (2001) 259. [18] S. D. Kaloshkin and I. A. Tomilin, 280-281, (1996) 303. [19] D.V. Louzguine-Luzgin, R. Belosludov, A.R. Yavari, K. Georgarakis, G. Vaughan, Y. Kawazoe, T. Egami, A. Inoue, J. Appl. Phys. 110 (2011) 043519. [20] C.A. Angell, J. Non-Cryst. Solids 102 (1988) 205. [21] T. Egami, J. Non-Cryst. Solids 317 (2003) 30. [22] M.I. Ojovan, J. Non-Cryst. Solids 382 (2013) 79. [23] P.W. Anderson, Science 267 (1995) 1615. [24] D.V. Louzguine-Luzgin, J. Alloys Compd. 586 (2014) S2. [25] W. Kauzmann, Chem. Rev. 43 (1948) 219. [26] C.A. Angell, The glass transition, Pergamon Encyclopedia of Materials: Science and Technology, vol. 4, 2001, p. 3365. [27] E. Leutheusser, Phys. Rev. A 29 (1984) 2765. [28] J.K. Kruger, Lect. Notes Phys. 716 (2007) 61. [29] J. Schroers, B. Lohwongwatana, W.L. Johnson, A. Peker, Appl. Phys. Lett. 87 (2005) 061912. [30] Y. Moriya, T. Yoshida, H. Kawaji, T. Atake, M. Fukuhara, H. Kimura, A. Inoue, Mater. Sci. Eng. B 148 (2008) 207. [31] S.V. Ketov, N. Chen, A. Caron, A. Inoue, D.V. Louzguine-Luzgin, Appl. Phys. Lett. 101 (2012) 241905. [32] V.Ya. Chekhovskoi, V.D. Tarasov, Yu.V. Gusev, High Temperature 38 (2000) 394. [33] J. Bünz, T. Brink, K. Tsuchiya, F. Meng, G. Wilde, K. Albe, Phys. Rev. Lett. 112 (2014) 135501. [34] R. Busch, JOM 52 (2000) 39. [35] G.D. Fontana, L. Battezzati, Acta Mater. 61 (2013) 2260. [36] H.B. Ke, P. Wen, W.H. Wang, AIP Adv. 2 (2012) 041404. [37] S.C. Glade, R. Busch, D.S. Lee, W.L. Johnson, R.K. Wunderlich, H.J. Fecht, J. Appl. Phys. 87 (2000) 7242. [38] D.V. Louzguine-Luzgin, I. Seki, T. Yamamoto, H. Kawaji, C. Suryanarayana, A. Inoue, Intermetallics 23 (2012) 177. [39] G. Fiore, I. Ichim, L. Battezzati, J. Non-Cryst. Solids 356 (2010) 2218. [40] G.D. Fontana, G.L. Fiore, L. Battezzati, J. Non-Cryst. Solids 382 (2013) 95. [41] A.L. Greer, J. Less-Common Met. 145 (1988) 131. [42] J.H. Perepezko, J.S. Paik, J. Non-Cryst. Solids 61–62 (1984) 113. [43] M. Dzugutov, Europhys. Lett. 26 (1994) 533. [44] W.F. Gale, T.C. Totemeier (Eds.), Smithells Metals Reference Book, 8th editionElsevier Butterworth-Heinemann Ltd., Oxford UK, 2004, pp. 13–17. [45] R.A. Perez, H. Nakajima, F. Dyment, Mater. Trans. 44 (2003) 2. [46] F. Faupel, W. Frank, M.-P. Macht, H. Mehrer, V. Naundorf, K. Rätzke, H.R. Schober, S.K. Sharma, H. Teichler, Rev. Mod. Phys. 75 (2003) 237. [47] T. Egami, Mater. Trans. 43 (2002) 510. [48] D.V. Louzguine-Luzgin, I. Seki, T. Yamamoto, H. Kawaji, C. Suryanarayana, A. Inoue, Phys. Rev. B 81 (2010) 144202. [49] N. Chen, R. Frank, N. Asao, D.V. Louzguine-Luzgin, P. Sharma, J.Q. Wang, G.Q. Xie, Y. Ishikawa, N. Hatakeyama, Y.C. Lin, M. Esashi, Y. Yamamoto, A. Inoue, Acta Mater. 59 (2011) 6433. [50] J.Q. Wang, N. Chen, P. Liu, Z. Wang, D.V. Louzguine-Luzgin, M.W. Chen, J.H. Perepezko, Acta Mater. 79 (2014) 30. [51] S. Singh, M.D. Ediger, J.J. Pablo, Nat. Mater. 12 (2013) 139.