Journal Information
Vol. 8. Issue 4.
Pages 3603-3611 (July - August 2019)
Download PDF
More article options
Vol. 8. Issue 4.
Pages 3603-3611 (July - August 2019)
Original Article
DOI: 10.1016/j.jmrt.2019.05.023
Open Access
Bridging the local configurations and crystalline counterparts of bulk metallic glass by nanocalorimetry
Bingge Zhaoa,b, Bin Yangc, Javier Rodríguez-Viejod, Mannan Wua, Christoph Schickc,e, Qijie Zhaia, Yulai Gaoa,b,
Corresponding author

Corresponding author.
a Center for Advanced Solidification Technology (CAST), School of Materials Science and Engineering, Shanghai University, 99 Shangda Road, 200444 Shanghai, PR China
b Laboratory for Microstructures, Institute of Materials, Shanghai University, 99 Shangda Road, 200444 Shanghai, PR China
c Institute of Physics and Competence Centre CALOR, University of Rostock, Albert-Einstein-Street 23-24, Rostock 18051, Germany
d Physics Department, Universitat Autònoma de Barcelona, Bellaterra 08193, Spain
e Kazan Federal University, 18 Kremlyovskaya Street, Kazan 420008, Russian Federation
This item has received

Under a Creative Commons license
Article information
Full Text
Download PDF
Figures (6)
Show moreShow less
Additional material (1)

The structural understanding of crystallization in bulk metallic glasses (BMGs) has attracted much attention while rapid crystallization occurring under controllable conditions is less involved. In this study, a Ce68Al10Cu20Co2 (at.%) BMG was thermally devitrified by differential scanning calorimetry (DSC) and nanocalorimetry. At a heating rate of 10K/min by DSC, AlCe3 and Ce are the major crystalline phases after devitrification while Al13Co4 quasicrystals and Ce are the dominant phases in the crystallization products at a heating rate of 5000K/s by nanocalorimetry. Attributing to the covalent-like bond in Al–Co atom pairs, Al13Co4 quasicrystals precipitate in the primary crystallization and work as the precursors associating local atomic configurations in the glassy state with crystalline phases after crystallization. Attributing to the enhanced mobility of Cu atoms, compositional redistribution occurs in the as-cast sample. On nanocalorimetry heating, an unambiguous discrepancy in the nucleation and growth of the nano-sized Al13Co4 quasicrystals is thus triggered, contributing to an obvious difference in the crystal size. This research unveils the distinct crystallization behaviors of Ce-based BMG on rapid heating. The formation of quasicrystals demonstrates the multi-stage crystallization on rapid heating and bridges the structural gap between local atomic configurations of metallic glasses and crystalline phases.

Crystal growth
Metallic glass
Full Text

Since the first preparation of bulk metallic glasses (BMGs) by Duwez [1], tremendous attention from both the fundamental and practical disciplines has been paid to this area attributing to their intriguing structural and mechanical properties [2–5]. According to the classical nucleation theory (CNT), atom motions slow down and enthalpy decreases rapidly upon quenching the metallic melt below its melting temperature. With different undercooling, various nucleation and growth behaviors are expected, changing the final structure and mechanical properties [6–11]. If crystallization on cooling is totally suppressed, the undercooled liquid falls out of equilibrium, and therefore an amorphous state is preserved. Compared with crystalline phases, an excess Gibbs free energy is trapped in the metallic glass. Driven by such excess energy, BMG undergoes glass transition and crystallization at elevated temperatures. Up to now, the kinetics and structure evolution on crystallization in various BMGs have been synthetically demonstrated by thermal analysis and structural characterization [12,13]. However, most of these studies were performed under near-equilibrium conditions. For example, the heating rates of the conventional differential scanning calorimetry (DSC) are usually centered at tens of K/min [14]. In view of the pronounced relationship between solidification behavior and cooling rate [15], it is extremely attractive to comparatively figure out what is able to happen on rapid heating. However, due to the deficiency of conventional DSC, a novel calorimetry that can provide enhanced heating capability is vigorously needed. With the development of microelectromechanical systems (MEMS) technology, nanocalorimetry, which was initially designed by Allen and Hellman et al., provides a solution to this problem attributing to its ultrafast heating rate as high as 106K/s [16,17]. Furthermore, the sensitivity of nanocalorimetry is substantially improved, making it capable of capturing weak thermal signals and revealing phase transitions in nanoscale systems [18–20]. On rapid heating, it has been demonstrated that both crystallization kinetics and structure evolutions are possible to change. For example, a curved Kissinger plot depicting the crystallization kinetics of Ge2Sb2Te5 (at.%) was obtained by Greer, confirming the growth-controlled crystallization on heating and a decoupling between crystal growth and viscous flow in the deeply undercooled liquid [21]. This phenomenon is deemed to occur in many other fragile phase-change materials (PCMs) [22,23]. In terms of metallic glasses, Vlassak introduced the nanocalorimetry to the rapid crystallization of Cu50Zr50 (at.%) thin films, where the decoupling and crystallization asymmetry on heating and cooling were demonstrated as well [24]. Yang further identified the decoupling temperature in an Al-based metallic glass by heating at rates spanning over six orders of magnitude [25]. With the help of nanocalorimetry, our previous studies realized in situ preparation of metallic glass and made it possible to tune the microstructure of metallic glass, and the structural effect in determining the crystallization behaviors were revealed [26,27]. Another remarkable issue about the rapid crystallization by nanocalorimetry is the structure evolution. Some equilibrium phases can be suppressed while some other new metastable phases can form under extreme non-equilibrium condition [28–30]. If the heating rate is further increased, crystallization may be totally suppressed [9,31]. Based on those scientific outcomes, it is expected that nanocalorimetry opens new avenues to rapid crystallization under extreme conditions.

Differential fast scanning calorimetry (DFSC), a nanocalorimetric instrument designed by Schick [32], achieves in situ heating and quenching at rates up to 105K/s, and it has been widely employed to reveal the rapid phase transitions of metals [20,26,27,33–36]. Although high-temperature DFSC sensor became available recently [25], alloys with low crystallization temperature are still preferred to guarantee the full crystallization in studying the rapid crystallization of bulk metallic glasses. Ce-based BMG is deemed as an ideal candidate because of its high glass forming ability (GFA), low crystallization temperature, and good stability [37–39]. In this paper, Ce68Al10Cu20Co2 (at.%) BMG samples were crystallized by DSC and DFSC respectively, revealing that the crystallization products are evidently affected by heating rate. On DSC heating at 10K/min, Ce and AlCe3 with a small fraction of Al5Co2 were detected after crystallization. Whereas at the heating rate of 5000K/s by DFSC, Al13Co4 with the quasicrystalline structure, which precipitates in the early stage of crystallization, was captured. Based on the diffusion of Co under these two conditions, the whole crystallization trajectory was depicted. According to the atom pairs in this BMG and Gibbs free energy in Al–Co binary system, the role of Co element in the formation and evolution of Al–Co intermetallic compounds was determined. As a result, primary and secondary crystallization were proposed during the rapid devitrification of Ce68Al10Cu20Co2 BMG. Cu atoms, on the other hand, have higher mobility and can diffuse even at room temperature, leading to a composition fluctuation in the as-cast sample. Both the nucleation and growth of Al13Co4 quasicrystals on following heating are influenced by the diffusion of Cu, causing the discrepancy in quasicrystal sizes.


Ce68Al10Cu20Co2 (at.%) master alloy was prepared by arc melting pure Ce, Al, Cu and Co in high-purity argon atmosphere. The purity of Ce is about 99.5wt.% while other metals have purity at least 99.9wt.%. Then the ingot was melted and quenched by suction casting to get a cylindrical rod with a diameter of 3mm. The amorphous structure was examined by X-ray diffraction (XRD, D/Max-2200, Rigaku) with Cu Kα radiation (λ=0.154056nm) at a scanning rate of 4°/min. High-resolution transmission electron microscopy (HRTEM, JEM-2010F, JEOL) was used to further certify the amorphous structure. Using plastic abrasive paper, micro-sized samples were ground from the as-cast rod and then loaded onto a copper mesh. The wedge-shaped edge of these samples is sufficiently thin for HRTEM observation. Glass transition temperature (Tg), crystallization temperature (Tx) and melting temperature (Tm) were determined by DSC (Diamond, PerkinElmer Instruments) at heating rates of 5, 10, 20, 40, and 80K/min.

To identify crystallization products, one BMG sample was fully crystallized by heating it from 323K to 573K (before Tm, without melting) at 10K/min using DSC, which was immediately followed by free cooling. Crystallization products were then examined by XRD at a scanning rate of 4°/min.

As for DFSC measurement, XI 39395 sensor (Xensor Integration, Netherlands) with a measuring area of 60μm×70μm was used. A small amount of silicon oil was spread on the sensor to improve the thermal contact between sample and membrane. Micro-sized samples were cut from the as-cast rod. A sample having an estimated size of 20μm was picked up using a copper wire and then positioned in the center of the measuring area with the help of optical microscopy (Stemi 2000, Carl Zeiss). Then it was heated from 320K to 621K to get full crystallization without melting, which was followed by quenching at 5000K/s down to room temperature. After quenching, this crystallized sample was thinned by focused ion beam (FIB, Helios NanoLab 600i, FEI). Its morphology, structure and composition were then characterized by HRTEM equipped with energy dispersive spectroscopy (EDS, Oxford). For comparison, another as-cast sample was machined by FIB and characterized by HRTEM to show the structure after casting. In addition to the TEM bright-field images, high angle annular dark field-scanning transmission electron microscopy (HAADF-STEM) images were captured to distinguish the composition fluctuation both in the as-cast and as-crystallized sample.

3Results and discussion

The absence of sharp diffraction peaks in the XRD pattern (Fig. 1a) confirms the amorphous nature of the as-cast Ce68Al10Cu20Co2 sample. HRTEM image and selected area electron diffraction (SAED) pattern further verify the amorphous state. Fig. 1b represents DSC heating curves at different rates. Both glass transition and sharp crystallization peak are observed, confirming the XRD and HRTEM results in Fig. 1a. Tg, Tx and Tm are 358K, 430K and 641K respectively at the heating rate of 5K/min. They are considered as the reference temperatures for the following calorimetric measurements. Using Kissinger model, the activation energy of crystallization (Ea,c) is estimated as 239.4kJ/mol. This large value suggests that Ce68Al10Cu20Co2 has a higher thermal stability against crystallization in Ce-based metallic glasses, agreeing with its higher glass forming ability than those CeAlCu BMGs [38,40]. For Ce68Al10Cu20Co2 BMG, its Tg is close to room temperature, and enthalpy relaxation is possible to occur even just after suction casting. When the as-cast sample is heated, the relaxed enthalpy rejuvenates, and an endothermic peak rather than a step change is detected at the glass transition, as indicated in Fig. 1b.

Fig. 1.

Characterization on the as-cast sample. (a) XRD pattern of the as-cast Ce68Al10Cu20Co2 rod. The broad peak indicates the glassy nature of the sample. Insets are HRTEM image and corresponding SAED pattern, further indicating the amorphous state. (b) DSC curves of the as-cast samples. A dominating endothermic peak rather than the glass transition step is detected because of the enthalpy rejuvenation.


By DSC heating from 320K to 621K at 10K/min, the as-cast sample was fully crystallized. Both the DSC curve and corresponding XRD pattern are shown in Fig. 2. The dominant phases are Ce and AlCe3, which agrees well with the results in Ref. [41]. In addition, Al5Co2 is identified here after the crystallization by DSC. According to the Scherrer equation [42], the mean size of the crystals is roughly estimated to be 19nm.

Fig. 2.

XRD pattern of Ce68Al10Cu20Co2 sample after heating at 10K/min. AlCe3 and Ce are the principal crystals. The inset shows the DSC heating curve for full crystallization.


The lower heating rate by DSC guarantees the full crystallization while DFSC can expose the initial crystallization stage by the ultrafast heating rate. Fig. 3a shows the DFSC heating curve at 5000K/s. In contrast to the DSC traces in Fig. 1b, two endothermic peaks near Tg are detected in the DFSC heating curve. To exclude the effect of silicone oil, similar amount of silicone oil was spread on the sensor and was scanned using the same nanocalorimetric profile. No peaks corresponding to phase transitions of silicone oil are detected (Fig. S1), suggesting that the distinct behavior of glass transition and crystallization in DFSC measurement comes from Ce68Al10Cu20Co2. In addition, analogous results are observed in the nanocalorimetric measurement at 1000K/s and 10,000K/s (Fig. S2). A step scanning was performed on another sample to determine the dual endothermic peaks. The temperature profile is displayed in Fig. 3b: the sample is heated at 2000K/s from room temperature to 393K. After holding for 0.06s, it is quenched down to room temperature at 4000K/s. Then the cycle with temperature step of 20K is repeated until the final temperature reaches 553K. The results are shown in Fig. 3c. When the sample is cycled below Tg, there are no obvious peaks (Curves 1 and 2). But as the maximum temperature on heating is above Tg (Curves 3–7), an endothermic step rather than dual peaks is repeatedly observed, which is the reversible characteristic of glass transition [43]. It is noteworthy that the thermal history including structure relaxation can be erased when the metallic glass is heated to enter the undercooled liquid region. If the phase separation occurs in the as-cast sample, the double glass transition should always appear in the cycles without crystallization, which disagrees with the nanocalorimetry curves shown in Fig. 3b. This contradictory case implies that the dual peaks in glass transition are related to the structure heterogeneity formed during the room-temperature aging after the suction casting of the alloy, and its mechanism will be discussed later in this paper. The absence of the dual endothermic peaks in the DSC curves may be due to the limited sensitivity of this conventional calorimetric method. As the temperature increases further (Curve 8), an exothermic event associated with the crystallization of the undercooled liquid is detected. The next reheating curve (Curve 9) shows no traces of a glass transition, indicating the crystalline nature of the sample before the temperature scan. The overall crystallization process consists of two exothermic peaks, which will be demonstrated based on the structure characterization.

Fig. 3.

Nanocalorimetry curves indicating the glass transition and crystallization of the as-cast sample. (a) DFSC heating curve at 5000K/s. The glass transition and crystallization are patterned in green and blue respectively. (b) Temperature profile for step scanning. The temperature step between two cycles is 20K with heating and cooling rates of 2000K/s and 4000K/s, respectively. (c) Heating curves of the step scanning on one sample. When the sample is scanned below Tg (Curves 1 and 2), there are no visible phase transitions. One reversible endothermic step corresponding to glass transition is detected when the sample is scanned between amorphous solid and undercooled liquid. Since the thermal history can be erased at temperature higher than Tg, the double glass transition in the as-cast sample is associated with room-temperature aging and corresponding structural heterogeneity caused by atom rearrangement. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)


TEM bright-field images of the sample that was crystallized at 5000K/s are displayed in Fig. 4. Regions with different contrasts, marked as A and B respectively, are clearly distinguished (Fig. 4a), indicating variations in crystallographic orientations after rapid crystallization. In the high-resolution images, both regions contain a large number of nanocrystals of different sizes, as displayed in Fig. 4b. In area A, most of the crystals are less than 5nm while in area B, the crystals are several tens of nanometers. These nanocrystals suggest an extremely high nucleation rate but a slow growth velocity during crystallization, agreeing well with the growth-limited crystallization on heating [9,24]. As mentioned above, the mean size of the crystals after DSC heating is about 19nm, not far from those measured after fast heating. In contrast, Tkatch found a reduction of crystal size from 320nm to 48nm as the heating rate increased from 10K/min to 104K/s in FeB metallic glass [44]. While in the CuZrAl bulk metallic glass, the crystalline phase produced at 250K/s is about 1.5μm which is two times smaller than that obtained at 750K/s [45]. These two studies contradict current results, implying the crystallization of Ce68Al10Cu20Co2 is insensitive to the heating rate. Based on Farjas's research [46], crystal size on crystallization is essentially dependent on the ratio between activation energies of nucleation and growth rather than external experimental conditions, and that is why a comparison between nucleation rate and growth velocity is important [47]. To determine the relationship between crystal size and heating rate experimentally, further explorations are still needed. Apart from the ordered phases, a few amorphous phases were still retained after rapid crystallization (Fig. 4c). Theoretically, crystallization can be suppressed by ultrahigh heating rate, where amorphous structure can survive up to the melting point [48,49]. Experimentally, this has been verified in Zr41Ti14Cu12Ni10Be23 (at.%) and Au49Ag5.5Pd2.3Cu26.9Si16.3 (at.%) metallic glasses [6,9]. This scenario of the mixed crystalline-glassy structure highlights the crucial role of crystal growth in the formation of amorphous structure, as stated by Greer [5]. In current case, ordered clusters nucleate on previous cooling (suction casting). Then they grow on fast heating, where the growth rate is too sluggish to crystallize the whole sample, leaving the residual amorphous phase between different grains. This can be further demonstrated by the nano-sized crystals in Fig. 4.

Fig. 4.

Structure characterization on Ce68Al10Cu20Co2 after rapid crystallization by DFSC. (a) The bright-field TEM image showing the bright (A) and dark (B) contrasts in the sample. The contrast suggests the different crystallographic orientations. (b) HRTEM image showing the interface between area A and B. Apparent size variations are observed. (c) The residual amorphous phase after rapid crystallization. (d) HRTEM image showing Ce and Al13Co4. They are the main ordered phases after rapid crystallization. (e and f) The HRTEM images corresponding to region A and B respectively. Al13Co4 are confirmed in both regions with an obvious size variation. Insets are the FFT patterns corresponding to the HRTEM image.


By the fast Fourier transform (FFT) patterns, the ordered phases observed in Fig. 4d are consistent with Ce (ICDD-JCPDS card No. 65-3368) and Al13Co4 (ICDD-JCPDS card No. 65-1165) respectively. This result differs from the crystallization products at 10K/min (Fig. 2). Furthermore, it is interesting to note that Al13Co4 possesses an orthorhombic structure belonging to Pmn21 space group, and this phase has been confirmed as a quasicrystal [50,51]. Since Al13Co4 phase exhibits 10-fold symmetry in (100) plane [50], the typical diffraction pattern of a quasicrystal is not recognized in the FFT pattern (Fig. 4e and f). The formation of Al13Co4 quasicrystals can be illustrated as follows. In CeAlCuCo BMGs, Al-centered icosahedrons, which presumably form on quenching, are the predominant structural building blocks which constitute the essential units [52]. These local units are far below the critical size of a nucleus and cannot work as the heterogeneous sites like macroscopic impurities. As recognized, nucleation barrier arises from the difference between local structure of the undercooled liquid and the ordered crystals. Compared with a crystalline phase, there is a significant structural similarity between local icosahedron configurations and quasicrystalline phases. Although these local icosahedrons are very small, a decreased surface energy and nucleation barrier are generated [53], facilitating the formation of quasicrystals. This argument agrees with the consideration that the topographical short-range order dominates the nucleation of quasicrystals [54]. It has been demonstrated that the sp electrons of Al atoms can hybridize with the 3d electrons of late transition metals (such as Co, Ni, Fe) [38], which shortens the bond length and produces the covalent-like bond. According to Holland-Moritz's study, the bond length of Al–Co is only 2.535[54] which is smaller than Al–Cu (2.74Å) and Al–Ce bond (3.356Å) [55]. This situation suggests a high priority to form Al–Co compounds during the crystallization of Ce68Al10Cu20Co2. According to Al–Co binary phase diagram [56], there are several kinds of Al–Co intermetallic compounds. Based on Sudavtsova's study [57], Fig. 5 schematically plots the dependence of Gibbs free energy on Co content at the onset temperature of crystallization (500K). In the light of the thermodynamics, Al5Co2, a stable phase, should precipitate from the amorphous matrix because of its lower Gibbs free energy. However, in contrast to Al5Co2 crystals with hexagonal structure, Al13Co4 is a quasicrystalline phase that is structurally similar with the local atomic configurations units of the metallic glass. As mentioned above, it is the decreased nucleation barrier resulting from the structural similarity that contributes to the precipitation of Al13Co4 rather than Al5Co2. Following the nucleation, quasicrystal growth occurs on heating, in which the diffusion of Co is considered as the essential factor. As indicated by the free energy change in Fig. 5, metastable Al13Co4 quasicrystals evolve to stable crystalline Al5Co2 phases with a sufficient diffusion of Co at elevated temperatures. In summary, Al13Co4 quasicrystals can overcome the energy barrier and precipitate as the intermediate phase between local icosahedron structure and equilibrium crystals, which corresponds to the statement on evolution of quasicrystals.

Fig. 5.

Schematic depicting the evolution of Gibbs free energy versus reaction coordinate on crystallization. Al13Co4 quasicrystals preferentially precipitate attributing to their structural similarity with Al-centered icosahedron in BMG although Al5Co2 is more thermodynamically stable. At elevated temperatures, metastable quasicrystals react with undercooled liquid to form stable crystals. The free energy of Al13Co4 and Al5Co2 is calculated based on Ref. [57].


Based on the step scanning measurements, relaxation is confirmed in this metallic glass, which has an impact on the crystallization. As evidenced by a similar Ce68Al10Cu20Fe2 metallic glass [39], crystallization mechanism can be changed with relaxation, where multiple crystallization peaks are detected consequently. However, the crystallization products are the same in that study, implying the negligible effect of relaxation on final phase formation. In the present experiment, however, the crystallization products are significantly distinguished, suggesting that the split of the crystallization peak is caused by both the relaxation and the ultrahigh heating rate. Since the composition of quasicrystals is totally different from the amorphous matrix, a primary crystallization is deemed to occur, which is evidenced by the first crystallization peak on DFSC heating curve. Furthermore, several kinds of ordered phases including some Ce-based crystals are detected after crystallization, implying a complicated crystallization behavior in this multicomponent metallic glass. A secondary crystallization is deemed to occur in this alloy, corresponding to the second crystallization peak in DFSC heating curve. On DSC heating (inset of Fig. 2), several minor exothermal peaks after the main crystallization are detected while they are absent on DFSC heating (Fig. 3a). Consistent with this argument, Al5Co2 that is identified on DSC heating is not observed in DFSC measurements. Since the quasicrystalline phase, A113Co4, is thermodynamically metastable, it can evolve into stable Al5Co2 by reaction with undercooled liquid at elevated temperatures. On the ultrafast heating by DFSC, this transition may shift to temperature higher than 621K, which is beyond the crystallization temperature range. In other words, the evolution from Al13Co4 to Al5Co2 is limited, leaving some undercooled liquid, corresponding to the mixed glassy-crystalline structure shown in Fig. 4.

In addition to the phase variations on DSC and DFSC crystallization, Al13Co4 quasicrystals exhibit apparent size discrepancy after rapid crystallization, as demonstrated in Fig. 4b. To shed light on the mechanism behind this phenomenon, HAADF-STEM mode in addition to bright-field image was employed to determine the Z-contrast regions, as pictured in Fig. 6a. Using EDS, composition fluctuations of Cu and Al are detected while Ce and Co are uniformly distributed after rapid crystallization [58]. Associated with the TEM image shown in Fig. 4a, area A with refined nanocrystals is rich in Al element while Cu element is abundant in area B with coarse grains. As a comparison, Fig. 6b demonstrates the HAADF-STEM image of the as-cast BMG sample in which different contrasts are observed as well (more TEM images can be found in Fig. S3). In contrast to the as-crystallized sample, only a fluctuation of Cu element is observed in the fully amorphous regions. Ce68Al10Cu20Co2 BMG has a Tg close to room temperature, and it is therefore convincing that atoms begin to rearrange and diffuse just after casting [37], presumably resulting in the composition discrepancy. Consequently, both the nucleation and growth, which are central to the crystallization, are manifestly influenced on the following heating, and various crystallization behaviors are expected in different regions. For Ce68Al10Cu20Co2 system, Ce works as the matrix while Al and Cu are the main solute elements. Compared with Al, Cu atoms have a smaller radius, and a particular mobility is therefore achieved [59–61]. After casting and being stored at room temperature, Cu atoms can diffuse in the sample, inducing the composition heterogeneity in the amorphous matrix. For the atom pairs in Ce68Al10Cu20Co2 BMG, only Cu and Co displays a positive enthalpy of mixing (6kJ/mol [61]). According to Inoue's criteria [62,63], this positive mixing enthalpy can reduce the glass forming ability and deteriorate the stability of the undercooled melt. It is therefore easier for the nucleation to occur in the Cu-rich region. Although some external factors such as heating rate [9] and pressure [64] can affect the crystallization behavior, the nucleation and growth rates are the intrinsic factors controlling the crystallization of metallic glass. For the particular case of quasicrystals, chemical short-rang order is especially important in their growth [65]. In Cu-rich region, the repulsive force between Co and Cu is larger, and an additional energy for Al–Co pair is thus produced as compared with Cu-poor zones. In other words, the growth for Al–Co is accelerated by Cu element in the Cu-rich region. It is therefore proposed that both the facilitated nucleation and enhanced growth contribute to the coarse crystals in the Cu-rich area.

Fig. 6.

HAADF-STEM images of the (a) as-crystallized and (b) as-cast sample. Both Cu and Al segregate after crystallization, resulting in the contrasts in (a). In (b), only the segregation of Cu is detected while other elements are uniformly distributed. Insets in (b) are the high-resolution images for various areas, showing the amorphous structure.


Rapid crystallization of Ce68Al10Cu20Co2 BMG was realized by nanocalorimetry. When heated at 5000K/s, the sample shows an altered crystallization behavior as compared with that crystallized at 10K/min by conventional DSC, which digs out the initial stage of crystallization and depicts the trajectory of devitrification.

Multi-stage crystallization starting with the precipitation of Al13Co4 quasicrystals proceeds in Ce68Al10Cu20Co2 BMG. Working as the precursor on devitrification, this metastable phase bridges the structural gap between local atomic configurations in metallic glass and crystallization products. Co element plays the essential role in the nucleation of quasicrystals attributing to its covalent-like bond with Al.

Attributing to the particular mobility of Cu atoms, a composition redistribution is detected in the as-cast Ce68Al10Cu20Co2 BMG, which causes distinct crystallization behaviors in the different regions on the following heating. The enrichment of Cu destabilizes the undercooled melt and promotes the nucleation and growth of Al13Co4 quasicrystals on crystallization, resulting in the coarse crystals compared with Cu-poor regions.

The application of DFSC provides new insights into the devitrification of bulk metallic glasses, making it feasible to control and quantify phase transformations under extremely non-equilibrium conditions.

Conflicts of interest

The authors declare no conflicts of interest.


This work is supported by the National Natural Science Foundation of China (Grant Nos. 51671123, 51171105 and 50971086), the Program for Professor of Special Appointment (Eastern Scholar) at Shanghai Institutions of Higher Learning (Grant No. TP2014042). BGZ thanks the support by China Postdoctoral Science Foundation (Grant No. 2018M640376). CS acknowledge financial support from the Ministry of Education and Science of the Russian Federation (Grant No. 14.Y26.31.0019).

Appendix A
Supplementary data

The following are the supplementary data to this article:

W.K. Jun, R.H. Willens, P. Duwez.
Non-crystalline structure in solidified gold–silicon alloys.
Nature, 187 (1960), pp. 869-870
G. Kumar, A. Desai, J. Schroers.
Bulk metallic glass: the smaller the better.
Adv Mater, 23 (2011), pp. 461-476
T.B. Matias, G.H. Asato, B.T. Ramasco, W.J. Botta, C.S. Kiminami, C. Bolfarini.
Processing and characterization of amorphous magnesium based alloy for application in biomedical implants.
J Mater Res Technol, 3 (2014), pp. 203-209
S. Tantavisut, B. Lohwongwatana, A. Khamkongkaeo, A. Tanavalee, P. Tangpornprasert, P. Ittiravivong.
The novel toxic free titanium-based amorphous alloy for biomedical application.
J Mater Res Technol, 7 (2018), pp. 248-253
A.L. Greer.
New horizons for glass formation and stability.
Nat Mater, 14 (2015), pp. 542-546
J. Schroers, A. Masuhr, W.L. Johnson, R. Busch.
Pronounced asymmetry in the crystallization behavior during constant heating and cooling of a bulk metallic glass-forming liquid.
Phys Rev B, 60 (1999), pp. 11855-11858
Z. Chen, Y. Zhang, S. Wang, J.Y. Zhang, Q. Tao, P. Zhang.
Microstructure and mechanical properties of undercooled Fe80C5Si10B5 eutectic alloy.
J Alloys Compd, 747 (2018), pp. 846-853
J. Schroers, Y. Wu, R. Busch, W.L. Johnson.
Transition from nucleation controlled to growth controlled crystallization in Pd43Ni10Cu27P20 melts.
Acta Mater, 49 (2001), pp. 2773-2781
S. Pogatscher, P.J. Uggowitzer, J.F. Löffler.
In-situ probing of metallic glass formation and crystallization upon heating and cooling via fast differential scanning calorimetry.
Appl Phys Lett, 104 (2014), pp. 251908
S. Wang, Z. Chen, L.C. Feng, Y.Y. Liu, P. Zhang, Y.Z. He, et al.
Nano-phase formation accompanying phase separation in undercooled CoCrCuFeNi-3at.% Sn high entropy alloy.
Mater Charact, 144 (2018), pp. 516-521
C. Simon, J. Gao, Y. Mao, G. Wilde.
Fast scanning calorimetric study of nucleation rates and nucleation transitions of Au–Sn alloys.
Scr Mater, 139 (2017), pp. 13-16
A.A. Kündig, M. Ohnuma, T. Ohkubo, K. Hono.
Early crystallization stages in a Zr–Cu–Ni–Al–Ti metallic glass.
Acta Mater, 53 (2005), pp. 2091-2099
S. Pauly, S. Gorantla, G. Wang, U. Kühn, J. Eckert.
Transformation-mediated ductility in CuZr-based bulk metallic glasses.
Nat Mater, 9 (2010), pp. 473-477
C. Schick, D. Lexa, L. Leibowitz.
Differential scanning calorimetry and differential thermal analysis.
Characterization of materials, pp. 483-494
W. Kurz, B. Giovanola, R. Trivedi.
Theory of microstructural development during rapid solidification.
Acta Metall, 34 (1986), pp. 823-830
A.A. Minakov, C. Schick.
Ultrafast thermal processing and nanocalorimetry at heating and cooling rates up to 1MK/s.
Rev Sci Instrum, 78 (2007), pp. 073902
Y. Gao, B. Zhao, J.J. Vlassak, C. Schick.
Nanocalorimetry: door opened for in situ material characterization under extreme non-equilibrium conditions.
Prog Mater Sci, 104 (2019), pp. 53-137
S.L. Lai, G. Ramanath, L.H. Allen, P. Infante.
Heat capacity measurements of Sn nanostructures using a thin-film differential scanning calorimeter with 0.2nJ sensitivity.
Appl Phys Lett, 70 (1997), pp. 43-45
D. Lee, J.J. Vlassak.
Diffusion kinetics in binary CuZr and NiZr alloys in the super-cooled liquid and glass states studied by nanocalorimetry.
Scr Mater, 165 (2019), pp. 73-77
L. Li, B. Yang, B. Zhao, A.S. Abyzov, J.W.P. Schmelzer, C. Schick, et al.
Rapid solidification behavior of nano-sized Sn droplets embedded in the Al matrix by nanocalorimetry.
Mater Res Express, 1 (2014), pp. 045012
J. Orava, A.L. Greer, B. Gholipour, D.W. Hewak, C.E. Smith.
Characterization of supercooled liquid Ge2Sb2Te5 and its crystallization by ultrafast-heating calorimetry.
Nat Mater, 11 (2012), pp. 279-283
B. Chen, J. Momand, P.A. Vermeulen, B.J. Kooi.
Crystallization kinetics of supercooled liquid Ge–Sb based on ultrafast calorimetry.
Cryst Growth Des, 16 (2016), pp. 242-248
H. Weber, J. Orava, I. Kaban, J. Pries, A.L. Greer.
Correlating ultrafast calorimetry, viscosity, and structural measurements in liquid GeTe and Ge15Te85.
Phys Rev Mater, 2 (2018), pp. 093405
D. Lee, B. Zhao, E. Perim, H. Zhang, P. Gong, Y. Gao, et al.
Crystallization behavior upon heating and cooling in Cu50Zr50 metallic glass thin films.
Acta Mater, 121 (2016), pp. 68-77
B. Yang, J.W.P. Schmelzer, B. Zhao, Y. Gao, C. Schick.
Glass transition and primary crystallization of Al86Ni6Y4.5Co2La1.5 metallic glass at heating rates spanning over six orders of magnitude.
Scr Mater, 162 (2019), pp. 146-150
B. Zhao, B. Yang, A.S. Abyzov, J.W.P. Schmelzer, J. Rodríguez-Viejo, Q. Zhai, et al.
Beating homogeneous nucleation and tuning atomic ordering in glass-forming metals by nanocalorimetry.
Nano Lett, 17 (2017), pp. 7751-7760
B. Zhao, L. Li, Q. Zhai, Y. Gao.
Formation of amorphous structure in Sn3.5Ag droplet by in situ fast scanning calorimetry controllable quenching.
Appl Phys Lett, 103 (2013), pp. 131913
P. Swaminathan, M.D. Grapes, K. Woll, S.C. Barron, D.A. LaVan, T.P. Weihs.
Studying exothermic reactions in the Ni-Al system at rapid heating rates using a nanocalorimeter.
J Appl Phys, 113 (2013), pp. 143509
M. Molina-Ruiz, A.F. Lopeandía, M. González-Silveira, Y. Anahory, M. Guihard, G. Garcia, et al.
Formation of Pd2Si on single-crystalline Si (100) at ultrafast heating rates: an in-situ analysis by nanocalorimetry.
Appl Phys Lett, 102 (2013), pp. 143111
S. Pogatscher, D. Leutenegger, J.E.K. Schawe, P.J. Uggowitzer, J.F. Löffler.
Solid–solid phase transitions via melting in metals.
Nat Commun, 7 (2016), pp. 11113
W.L. Johnson, G. Kaltenboeck, M.D. Demetriou, J.P. Schramm, X. Liu, K. Samwer, et al.
Beating crystallization in glass-forming metals by millisecond heating and processing.
Science, 332 (2011), pp. 828-833
E. Zhuravlev, C. Schick.
Fast scanning power compensated differential scanning nano-calorimeter: 1. The device.
Thermochim Acta, 505 (2010), pp. 1-13
B. Zhao, L. Li, F. Lu, Q. Zhai, B. Yang, C. Schick, et al.
Phase transitions and nucleation mechanisms in metals studied by nanocalorimetry: a review.
Thermochim Acta, 603 (2015), pp. 2-23
Y. Gao, E. Zhuravlev, C. Zou, B. Yang, Q. Zhai, C. Schick.
Calorimetric measurements of undercooling in single micron sized SnAgCu particles in a wide range of cooling rates.
Thermochim Acta, 482 (2009), pp. 1-7
W. Wu, Q. Zhang, B. Zhao, L. Zhang, Q. Zhai, Y. Gao.
Solidification behavior of indium droplets embedded in aluminum by differential fast scanning calorimetry.
J Therm Anal Calorim, 135 (2018), pp. 2995-3003
B. Yang, J.H. Perepezko, J.W.P. Schmelzer, Y. Gao, C. Schick.
Dependence of crystal nucleation on prior liquid overheating by differential fast scanning calorimeter.
J Chem Phys, 140 (2014), pp. 104513
B. Zhang, D.Q. Zhao, M.X. Pan, W.H. Wang, A.L. Greer.
Amorphous metallic plastic.
Phys Rev Lett, 94 (2005), pp. 205502
B. Zhang, D.Q. Zhao, M.X. Pan, R.J. Wang, W.H. Wang.
Formation of cerium-based bulk metallic glasses.
Acta Mater, 54 (2006), pp. 3025-3032
Q. Luo, B. Zhang, D.Q. Zhao, R.J. Wang, M.X. Pan, W.H. Wang.
Aging and stability of cerium-based bulk metallic glass.
Appl Phys Lett, 88 (2006), pp. 151915
Z. Bian, A. Inoue.
Ultra-low glass transition temperatures in Ce-based bulk metallic glasses.
Mater Trans, 46 (2005), pp. 1857-1860
B. Zhang, R.J. Wang, D.Q. Zhao, M.X. Pan, W.H. Wang.
Superior glass-forming ability through microalloying in cerium-based alloys.
Phys Rev B, 73 (2006), pp. 092201
P. Scherrer.
Bestimmung der Grösse und der inneren Struktur von Kolloidteilchen mittels Röntgenstrahlen.
Nachr Ges Wiss, 26 (1918), pp. 98-100
G.P. Johari, A. Hallbrucker, E. Mayer.
The glass–liquid transition of hyperquenched water.
Nature, 330 (1987), pp. 552-553
V.I. Tkatch, A.I. Limanovskii, V.Y. Kameneva.
Studies of crystallization kinetics of Fe40Ni40P14B6 and Fe80B20 metallic glasses under non-isothermal conditions.
J Mater Sci, 32 (1997), pp. 5669-5677
I.V. Okulov, I.V. Soldatov, M.F. Sarmanova, I. Kaban, T. Gemming, K. Edström, et al.
Flash Joule heating for ductilization of metallic glasses.
Nat Commun, 6 (2015), pp. 7932
J. Farjas, P. Roura.
Solid-phase crystallization under continuous heating: kinetic and microstructure scaling laws.
J Mater Res, 23 (2008), pp. 418-426
J. Gerges, F. Affouard.
Predictive calculation of the crystallization tendency of model pharmaceuticals in the supercooled state from molecular dynamics simulations.
J Phys Chem B, 119 (2015), pp. 10768-10783
V.A. Shneidman, M.C. Weinberg.
Crystallization of rapidly heated amorphous solids.
J Non-Cryst Solids, 194 (1996), pp. 145-154
J. Orava, A.L. Greer.
Kissinger method applied to the crystallization of glass-forming liquids: regimes revealed by ultra-fast-heating calorimetry.
Thermochim Acta, 603 (2015), pp. 63-68
R. Addou, E. Gaudry, Deniozou Th, M. Heggen, M. Feuerbacher, P. Gille, et al.
Structure investigation of the (100) surface of the orthorhombic Al13Co4 crystal.
Phys Rev B, 80 (2009), pp. 014203
D. Kandaskalov, V. Fournée, J. Ledieu, É. Gaudry.
Adsorption properties of the o-Al13Co4 (100) surface toward molecules involved in the semihydrogenation of acetylene.
J Phys Chem C, 118 (2014), pp. 23032-23041
X. Xi, L.L. Li, B. Zhang, W. Wang, Y. Wu.
Correlation of atomic cluster symmetry and glass-forming ability of metallic glass.
Phys Rev Lett, 99 (2007), pp. 095501
Y.Q. Cheng, E. Ma.
Atomic-level structure and structure-property relationship in metallic glasses.
Prog Mater Sci, 56 (2011), pp. 379-473
D. Holland-Moritz.
Short-range order and solid–liquid interfaces in undercooled metallic melts.
Mater Sci Eng A, 304-306 (2001), pp. 108-113
S.W. Han, C.H. Booth, E.D. Bauer, P.H. Huang, Y.Y. Chen, J.M. Lawrence.
Lattice disorder and size-induced Kondo behavior in CeAl2 and CePt2+x.
Phys Rev Lett, 97 (2006), pp. 097204
P. Gille, B. Bauer.
Single crystal growth of Al13Co4 and Al13Fe4 from Al-rich solutions by the Czochralski method.
Cryst Res Technol, 43 (2008), pp. 1161-1167
M.A. Shevchenko, V.V. Berezutskii, M.I. Ivanov, V.G. Kudin, V.S. Sudavtsova.
Thermodynamic properties of alloys of the Al–Co and Al–Co–Sc systems.
Russ J Phys Chem, 88 (2014), pp. 729-734
B. Zhao, L. Li, Q. Zhai, Y. Gao.
Structure characterization of Sn-based and Ce-based alloys treated by ultrafast scanning.
Characterization of minerals, metals, and materials, pp. 227-234
C.A. Angell, K.L. Ngai, G.B. McKenna, P.F. McMillan, S.W. Martin.
Relaxation in glass forming liquids and amorphous solids.
J Appl Phys, 88 (2000), pp. 3113-3157
Z. Evenson, S.E. Naleway, S. Wei, O. Gross, J.J. Kruzic, I. Gallino, et al.
β relaxation and low-temperature aging in a Au-based bulk metallic glass: from elastic properties to atomic-scale structure.
Phys Rev B, 89 (2014), pp. 174204
A. Takeuchi, A. Inoue.
Classification of bulk metallic glasses by atomic size difference, heat of mixing and period of constituent elements and its application to characterization of the main alloying element.
Mater Trans, 46 (2005), pp. 2817-2829
A. Inoue.
Stabilization of supercooled liquid and opening-up of bulk glassy alloys.
Proc Jpn Acad Ser B, 73 (1997), pp. 19-24
A. Inoue.
Stabilization of metallic supercooled liquid and bulk amorphous alloys.
Acta Mater, 48 (2000), pp. 279-306
J. Zhang, Y.H. Wei, K.Q. Qiu, H.F. Zhang, M.X. Quan, Z.Q. Hua.
Crystallization kinetics and pressure effect on crystallization of Zr55Al10Ni5Cu30 bulk metallic glass.
Mater Sci Eng A, 357 (2003), pp. 386-391
J. Schroers, D. Holland-Moritz, D.M. Herlach, K. Urban.
Growth kinetics of quasicrystalline and polytetrahedral phases of Al–Pd–Mn, Al–Co, and Al–Fe from the undercooled melt.
Phys Rev B, 61 (2000), pp. 14500-14506
Copyright © 2019. The Authors
Journal of Materials Research and Technology

Subscribe to our newsletter

Article options
Supplemental materials
Cookies policy
To improve our services and products, we use cookies (own or third parties authorized) to show advertising related to client preferences through the analyses of navigation customer behavior. Continuing navigation will be considered as acceptance of this use. You can change the settings or obtain more information by clicking here.