Fabrication of Metastable Crystalline Nanocomposites by Flash Annealing of Cu47.5Zr47.5Al5 Metallic Glass Using Joule Heating

Flash Joule-heating was applied to the Cu47.5Zr47.5Al5 metallic glass for designing fully crystalline metastable nanocomposites consisting of the metastable B2 CuZr and low-temperature equilibrium Cu10Zr7 phases. The onset of crystallization was in situ controlled by monitoring resistivity changes in the samples. The effect of heating rate and annealing time on the volume fraction of the crystalline phases and mechanical properties of the nanocomposites was studied in detail. Particularly, an increase of the heating rate and a decrease of the annealing time lead to a lower number of equilibrium Cu10Zr7 precipitates and an increase of tensile ductility. Tailoring of these non-equilibrium microstructures and mechanical properties may not be possible unless one starts with a fully glassy material that opens new perspectives for designing metastable nanomaterials with unique physical properties.

Metallic glasses are metallic solids with disordered, liquid-like atomic structures. Thermodynamically, metallic glasses are in a high-energy (metastable) state regardless of their

Materials and Methods
Samples were prepared under high purity argon atmosphere in two steps. First, Cu 47.5 Zr 47. 5 Al 5 (at. %) ingots were produced from Cu (99.99%), Zr (99.98%), and Al (99.99%), by arc-melting. In the second step, glassy ribbons were prepared from the ingots by melt-spinning. The metallic glass ribbons were annealed using an in-house designed set-up. The samples were characterized by X-ray diffraction (XRD, Stoe, STADI P with Mo-Kα1 radiation, Darmstadt, Germany) and scanning electron microscopy (SEM, Zeiss, Leo Gemini 1530, Oberkochen, Germany). Phase identification was done by means of the X'Pert High Score Plus (Malvern Panalytical, Malvern, UK) software, whereas SEM images were analysed using ImageJ (open source) software. Mechanical testing was performed with an Instron 8562 machine (Instron, Norwood, MA, United States) at a strain rate of 10 −4 s −1 at room temperature. The strain was measured by a laser extensometer (Fiedler Optoelektronik, Lützen, Germany). The gauge length of 5 mm was selected in the middle region of a ribbon sample.

Results and Discussion
Cu 47.5 Zr 47.5 Al 5 metallic glass can transform to a lower energy state through several devitrification paths. These are schematically illustrated in a continuous heating transformation (CHT) diagram in Figure 1.
Fast heating to temperatures above the crystallization temperature, T x , as depicted by line 1, leads to nucleation of the metastable B2 CuZr phase in the metallic glass matrix. Fast cooling is required to avoid decomposition of B2 CuZr into the thermodynamically favourable low temperature eutectic phases (LT-EPs) Cu 10 Zr 7 and CuZr 2 , and thus to stabilize B2 CuZr at room temperature. The ductile B2 CuZr crystals hinder localization of deformation in the glassy phase leading to tensile ductility of the glass-matrix composites [20]. Moving the cooling curve closer to the nose of the stability regime of the LT-EPs leads to a higher solid fraction of B2 CuZr, ultimately leading to complete suppression of the glassy phase. Heating along line 2 and cooling along lines a and b results in devitrification of the metallic glass into B2 CuZr with subsequent decomposition of B2 CuZr into the LT-EPs. As it has been shown in earlier works [20,33], the mechanical properties of the composites obtained depend strongly on the constituent phases.  [20].
Fast heating to temperatures above the crystallization temperature, Tx, as depicted by line 1, leads to nucleation of the metastable B2 CuZr phase in the metallic glass matrix. Fast cooling is required to avoid decomposition of B2 CuZr into the thermodynamically favourable low temperature eutectic phases (LT-EPs) Cu10Zr7 and CuZr2, and thus to stabilize B2 CuZr at room temperature. The ductile B2 CuZr crystals hinder localization of deformation in the glassy phase leading to tensile ductility of the glass-matrix composites [20]. Moving the cooling curve closer to the nose of the stability regime of the LT-EPs leads to a higher solid fraction of B2 CuZr, ultimately leading to complete suppression of the glassy phase. Heating along line 2 and cooling along lines a and b results in devitrification of the metallic glass into B2 CuZr with subsequent decomposition of B2 CuZr into the LT-EPs. As it has been shown in earlier works [20,33], the mechanical properties of the composites obtained depend strongly on the constituent phases.
In this work, we use the Joule heating for annealing of the Cu47.5Zr47.5Al5 metallic glass. Hereby, an electrical current is applied to a sample for a short time from several milliseconds to a few seconds, as described elsewhere [20]. The fast heating is favoured by the rather high resistivity of metallic glasses [36,37]. Since crystalline phases possess a higher conductivity compared to their glassy counterparts, devitrification of the glassy phase leads to a remarkable resistivity drop, which allows one to quite sensitively monitor crystallization processes [20]. The distribution of the applied current density versus the onset of the resistivity drop for the Cu47.5Zr47.5Al5 metallic glass can be found in our previous study [20]. The relatively short heating time during the Joule heating allows the adiabatic conditions to be nearly fulfilled and, therefore, the current density is proportional to the heating rate, as has been proven by measurement of heating rates using a thermocouple. The heating rates given in Table 1 have been estimated based on the crystallization temperature of the Cu47.5Zr47.5Al5 metallic glass (about 695 K at about 0.7 K s −1 heating rate) [38], and the measured time-to-crystallisation indicated by the resistivity drop. Particularly, the heating rate corresponding to the highest applied current density (namely, 59 ± 5 MA m −2 ) is ≥830 K s −1 . Figure 2 displays the effect of the heating rate (current density) on the microstructure of the flash-annealed Cu47.5Zr47.5Al5 metallic glass. X-ray diffraction analysis of the samples annealed at the lowest heating rate (≥150 K s −1 ) reveals the presence of two crystalline phases: the low-temperature equilibrium Cu10Zr7 and the metastable B2 CuZr phases ( Figure 2a). As the heating rate increases above 330 K s −1 , the intensity of Cu10Zr7 reflections decreases significantly, indicating a lower content of this phase in the sample (Figure 2b). Finally, at a heating rate of ≥830 K s −1 , there are no detectable peaks of the Cu10Zr7 phase and the sample mainly consists of the metastable B2 CuZr phase ( Figure  2c). In this work, we use the Joule heating for annealing of the Cu 47.5 Zr 47.5 Al 5 metallic glass. Hereby, an electrical current is applied to a sample for a short time from several milliseconds to a few seconds, as described elsewhere [20]. The fast heating is favoured by the rather high resistivity of metallic glasses [36,37]. Since crystalline phases possess a higher conductivity compared to their glassy counterparts, devitrification of the glassy phase leads to a remarkable resistivity drop, which allows one to quite sensitively monitor crystallization processes [20]. The distribution of the applied current density versus the onset of the resistivity drop for the Cu 47.5 Zr 47.5 Al 5 metallic glass can be found in our previous study [20]. The relatively short heating time during the Joule heating allows the adiabatic conditions to be nearly fulfilled and, therefore, the current density is proportional to the heating rate, as has been proven by measurement of heating rates using a thermocouple. The heating rates given in Table 1 have been estimated based on the crystallization temperature of the Cu 47.5 Zr 47.5 Al 5 metallic glass (about 695 K at about 0.7 K s −1 heating rate) [38], and the measured time-to-crystallisation indicated by the resistivity drop. Particularly, the heating rate corresponding to the highest applied current density (namely, 59 ± 5 MA m −2 ) is ≥830 K s −1 .  Figure 2 displays the effect of the heating rate (current density) on the microstructure of the flash-annealed Cu 47.5 Zr 47.5 Al 5 metallic glass. X-ray diffraction analysis of the samples annealed at the lowest heating rate (≥150 K s −1 ) reveals the presence of two crystalline phases: the low-temperature equilibrium Cu 10 Zr 7 and the metastable B2 CuZr phases ( Figure 2a). As the heating rate increases above 330 K s −1 , the intensity of Cu 10 Zr 7 reflections decreases significantly, indicating a lower content of this phase in the sample (Figure 2b). Finally, at a heating rate of ≥830 K s −1 , there are no detectable peaks of the Cu 10 Zr 7 phase and the sample mainly consists of the metastable B2 CuZr phase (Figure 2c). Figure 2d-f demonstrates the microstructure of the annealed Cu47.5Zr47.5Al5 metallic glass samples. The samples consist of dendritic Cu10Zr7 crystals that are homogeneously distributed in a B2 CuZr matrix. The volume fraction of the Cu10Zr7 dendrites increases with decreasing heating rate from about 2 ± 1 vol.% (≥830 K s −1 ) to 17 ± 3 vol.% (≥330 K s −1 ), finally reaching 41 ± 5 vol.% at the lowest heating rate (≥150 K s −1 ) ( Table 1). The size of the Cu10Zr7 dendrites also depends on the applied heating rate or current density. A larger dendrite size is achieved at a higher heating rate and vice versa. For example, the mean dendrite size in the sample subjected to the highest heating rate ≥830 K s −1 is 2.3 ± 0.2 µ m, while it drops to 1.1 ± 0.3 µ m for the samples obtained at the heating rate ≥150 K s −1 (Table 1). Along with the volume fraction and size of the Cu10Zr7 crystals, their number increases with higher heating rate from about 1.1 × 10 4 to 22.0 × 10 4 and reaches 61.2 × 10 4 particles per mm 2 .  Figure 2d-f demonstrates the microstructure of the annealed Cu 47.5 Zr 47.5 Al 5 metallic glass samples. The samples consist of dendritic Cu 10 Zr 7 crystals that are homogeneously distributed in a B2 CuZr matrix. The volume fraction of the Cu 10 Zr 7 dendrites increases with decreasing heating rate from about 2 ± 1 vol.% (≥830 K s −1 ) to 17 ± 3 vol.% (≥330 K s −1 ), finally reaching 41 ± 5 vol.% at the lowest heating rate (≥150 K s −1 ) ( Table 1). The size of the Cu 10 Zr 7 dendrites also depends on the applied heating rate or current density. A larger dendrite size is achieved at a higher heating rate and vice versa. For example, the mean dendrite size in the sample subjected to the highest heating rate ≥830 K s −1 is 2.3 ± 0.2 µm, while it drops to 1.1 ± 0.3 µm for the samples obtained at the heating rate ≥150 K s −1 ( Table 1). Along with the volume fraction and size of the Cu 10 Zr 7 crystals, their number increases with higher heating rate from about 1.1 × 10 4 to 22.0 × 10 4 and reaches 61.2 × 10 4 particles per mm 2 .
The volume fraction of Cu 10 Zr 7 dendrites in the B2 CuZr matrix can also be tuned by controlling the annealing time at a constant current density (heating rate), as shown in Figure 3. To reveal this effect, several samples annealed (i) until the resistivity drop, (ii) 1.6 s after the resistivity drop, and (iii) 2.2 s after the resistivity drop at a current density i 3 = 34 ± 5 MA m −2 (≥150 K s −1 ) were selected. X-ray analysis indicates the presence of two phases in these samples, namely, the low-temperature equilibrium Cu 10 Zr 7 phase and the metastable B2 CuZr phase. The intensity of the B2 CuZr phase peaks is highest for the samples annealed until the resistivity drop. A relatively small increase in annealing time (1.6-2.2 s) leads to a higher intensity of the Cu 10 Zr 7 peaks. This is in agreement with the findings from secondary electron micrographs (Figure 3d-e). The increase of annealing time at the constant heating rate (≥150 K s −1 ) leads to a higher volume fraction and a larger number of Cu 10 Zr 7 dendrites (Table 1). Particularly, the volume fraction of the Cu 10 Zr 7 dendrites increases from 41 ± 5 to Nanomaterials 2020, 10, 84 5 of 10 89 ± 3 vol.% for the annealing times until the resistivity drop and 2.2 s after the drop, respectively. In contrast to the effect of heating rate, the size of the Cu 10 Zr 7 dendrites varies insignificantly for different annealing times. These findings suggest that the volume fraction, number, and size of the Cu 10 Zr 7 crystals strongly depend on the specific heat treatment conditions and can be tuned by varying the current density being proportional to the heating rate.
equilibrium Cu10Zr7 phase and the metastable B2 CuZr phase. The intensity of the B2 CuZr phase peaks is highest for the samples annealed until the resistivity drop. A relatively small increase in annealing time (1.6-2.2 s) leads to a higher intensity of the Cu10Zr7 peaks. This is in agreement with the findings from secondary electron micrographs (Figure 3d-e). The increase of annealing time at the constant heating rate (≥150 K s −1 ) leads to a higher volume fraction and a larger number of Cu10Zr7 dendrites (Table 1). Particularly, the volume fraction of the Cu10Zr7 dendrites increases from 41 ± 5 to 89 ± 3 vol.% for the annealing times until the resistivity drop and 2.2 s after the drop, respectively. In contrast to the effect of heating rate, the size of the Cu10Zr7 dendrites varies insignificantly for different annealing times. These findings suggest that the volume fraction, number, and size of the Cu10Zr7 crystals strongly depend on the specific heat treatment conditions and can be tuned by varying the current density being proportional to the heating rate.  (Figure 3d is adopted from [20]).
In the current study, the flash Joule heating of the Cu47.5Zr47.5Al5 metallic glass leads to its devitrification into two phases: metastable B2 CuZr and equilibrium Cu10Zr7. This finding is in contrast to the equilibrium phase diagram [39] and some experimental as well as theoretical studies on the devitrification sequence of CuZr-based metallic glasses [33,35,38,40]. Based on theoretical considerations, Kaban et al. suggested that the Cu47.5Zr47.5Al5 metallic glass devitrifies following the sequence Cu10Zr7→CuZr2→B2 CuZr [40]. Experimental studies show that the Cu47.5Zr47.5Al5 metallic glass directly transforms into the Cu10Zr7 and CuZr2 equilibrium phases upon annealing at low heating rates of 10-40 K min −1 , while the CuZr2 phase precipitates after the Cu10Zr7 phase [38]. Recent reports demonstrated that rapid annealing of the Cu47.5Zr47.5Al5 metallic glass can suppress the formation of the low-temperature equilibrium phases completely [33] or partially [20], leading to its transformation into the non-equilibrium B2 CuZr phase. In the latter case, the Cu10Zr7 crystals are  (Figure 3d is adopted from [20]).
In the current study, the flash Joule heating of the Cu 47.5 Zr 47.5 Al 5 metallic glass leads to its devitrification into two phases: metastable B2 CuZr and equilibrium Cu 10 Zr 7 . This finding is in contrast to the equilibrium phase diagram [39] and some experimental as well as theoretical studies on the devitrification sequence of CuZr-based metallic glasses [33,35,38,40]. Based on theoretical considerations, Kaban et al. suggested that the Cu 47.5 Zr 47.5 Al 5 metallic glass devitrifies following the sequence Cu 10 Zr 7 →CuZr 2 →B2 CuZr [40]. Experimental studies show that the Cu 47.5 Zr 47.5 Al 5 metallic glass directly transforms into the Cu 10 Zr 7 and CuZr 2 equilibrium phases upon annealing at low heating rates of 10-40 K min −1 , while the CuZr 2 phase precipitates after the Cu 10 Zr 7 phase [38]. Recent reports demonstrated that rapid annealing of the Cu 47.5 Zr 47.5 Al 5 metallic glass can suppress the formation of the low-temperature equilibrium phases completely [33] or partially [20], leading to its transformation into the non-equilibrium B2 CuZr phase. In the latter case, the Cu 10 Zr 7 crystals are found within the B2 CuZr precipitates but the second equilibrium CuZr 2 phase is not observed. According to the devitrification experiments and theoretical studies [38,40], the CuZr 2 phase precipitates after Cu 10 Zr 7 , and, therefore, it can be assumed that in the current flash annealing case, the conditions for the precipitation of the CuZr 2 phase are not fulfilled. The CuZr 2 phase requires a specific stoichiometry for nucleation, which is probably not achieved during the short processing time.
The annealing conditions such as annealing temperature and time affect the nucleation and growth rate of precipitates. The growth rate is determined by the rate of diffusion and, therefore, it increases with increasing temperature [41]. The nucleation rate is also temperature-dependent and exhibits a maximum in an intermediate temperature range (Figure 4). According to the obtained results, the number of Cu 10 Zr 7 dendrites decreases while their average size increases at higher heating rate. This suggests that the B2-98 sample with the lowest number (1.1 × 10 4 ± 0.1 × mm −2 ) and the largest size (2.3 ± 0.3 µm) of Cu 10 Zr 7 dendrites was subjected to the highest transformation temperature. A decrease of the heating rate leads to a larger number of Cu 10 Zr 7 dendrites, which also become finer. This can be explained by a decrease of the average transformation temperature, which seems to be dependent on heating rate. The lowest applied heating rate (here, 150 K s −1 ) corresponds to the most optimum average transformation temperature for the highest nucleation rate. Therefore, increasing the annealing time in this case leads to a significant increase of the number of Cu 10 Zr 7 dendrites.

time.
The annealing conditions such as annealing temperature and time affect the nucleation and growth rate of precipitates. The growth rate is determined by the rate of diffusion and, therefore, it increases with increasing temperature [41]. The nucleation rate is also temperature-dependent and exhibits a maximum in an intermediate temperature range (Figure 4). According to the obtained results, the number of Cu10Zr7 dendrites decreases while their average size increases at higher heating rate. This suggests that the B2-98 sample with the lowest number (1.1 × 10 4 ± 0.1 × mm −2 ) and the largest size (2.3 ± 0.3 µ m) of Cu10Zr7 dendrites was subjected to the highest transformation temperature. A decrease of the heating rate leads to a larger number of Cu10Zr7 dendrites, which also become finer. This can be explained by a decrease of the average transformation temperature, which seems to be dependent on heating rate. The lowest applied heating rate (here, 150 K s −1 ) corresponds to the most optimum average transformation temperature for the highest nucleation rate. Therefore, increasing the annealing time in this case leads to a significant increase of the number of Cu10Zr7 dendrites. Nucleation and growth rate curves are drawn based on reference [41].
Tailored by flash annealing of the Cu47.5Zr47.5Al5 metallic glass, the composite microstructure of the current samples leads to notable tensile deformability and high strength ( Figure 5) comparable with that of metallic glass matrix composites [20,33]. Moreover, the crystalline samples exhibit pronounced strain-hardening behaviour. The yield strength of the samples increases from 700 ± 30 to 1440 ± 30 MPa with increasing volume fraction of Cu10Zr7 dendrites (Table 2). This strength increase with increasing volume fraction of Cu10Zr7 dendrites is at the cost of tensile deformability: the strainto-fracture decreases form 7.1 ± 0.5 to 1.8 ± 0.2% when the volume fraction of Cu10Zr7 dendrites increases from 2 ± 1 to 73 ± 4 vol.%. However, these fracture strain values are still in the range of interest for technological applications. For example, the B2-83 sample exhibits a tensile ductility of 7.5 ± 0.5%. Due to increasing strain hardening, the ultimate tensile strength increases from 1580 ± 50 Tailored by flash annealing of the Cu 47.5 Zr 47.5 Al 5 metallic glass, the composite microstructure of the current samples leads to notable tensile deformability and high strength ( Figure 5) comparable with that of metallic glass matrix composites [20,33]. Moreover, the crystalline samples exhibit pronounced strain-hardening behaviour. The yield strength of the samples increases from 700 ± 30 to 1440 ± 30 MPa with increasing volume fraction of Cu 10 Zr 7 dendrites (Table 2). This strength increase with increasing volume fraction of Cu 10 Zr 7 dendrites is at the cost of tensile deformability: the strain-to-fracture decreases form 7.1 ± 0.5 to 1.8 ± 0.2% when the volume fraction of Cu 10 Zr 7 dendrites increases from 2 ± 1 to 73 ± 4 vol.%. However, these fracture strain values are still in the range of interest for technological applications. For example, the B2-83 sample exhibits a tensile ductility of 7.5 ± 0.5%. Due to increasing strain hardening, the ultimate tensile strength increases from 1580 ± 50 to 1710 ± 50 MPa with higher volume fraction and number of Cu 10 Zr 7 dendrites in the B2 CuZr matrix (Table 2).

Conclusions
In summary, we have tailored different microstructures by devitrification of the Cu47.5Zr47.5Al5 metallic glass using flash Joule annealing. The size and volume fraction of the constituent phasesmetastable B2 CuZr and equilibrium Cu10Zr7-can be flexibly tuned by optimizing the heating rate and annealing time. The strength of the Cu10Zr7/B2 nanocomposites obtained through flash Joule heating and annealing exceeds that of the initial metallic glass and is comparable with that of metallic glass matrix composites. Hence, glassy materials provide a unique base for obtaining nonequilibrium microstructures by flash annealing with technologically attractive properties that cannot be achieved through conventional processing.

Conclusions
In summary, we have tailored different microstructures by devitrification of the Cu 47.5 Zr 47.5 Al 5 metallic glass using flash Joule annealing. The size and volume fraction of the constituent phases-metastable B2 CuZr and equilibrium Cu 10 Zr 7 -can be flexibly tuned by optimizing the heating rate and annealing time. The strength of the Cu 10 Zr 7 /B2 nanocomposites obtained through flash Joule heating and annealing exceeds that of the initial metallic glass and is comparable with that of metallic glass matrix composites. Hence, glassy materials provide a unique base for obtaining non-equilibrium microstructures by flash annealing with technologically attractive properties that cannot be achieved through conventional processing.