A Series of Zr-Based Bulk Metallic Glasses with Room Temperature Plasticity

A group of plastic Zr-Al-Ni-Cu bulk metallic glasses (BMGs) with low Zr content was developed and their thermal and mechanical properties were investigated. The results show that these Zr-based BMGs have a single crystallization event for all heating rates in the studied temperature region. The glass transition temperature Tg decreases with increasing Zr content for all heating rates. There are two melting procedures for the BMGs whose Zr content is less than 52 at %, while three melting procedures for the other Zr-based BMGs. The second melting procedure is split into two melting procedures for Zr52.5Al12.2Ni12.6Cu22.7 and Zr53Al11.6Ni11.7Cu23.7 BMGs, while the first melting procedure is split into two melting procedures for the other BMGs. The activation energy decreases with increasing sensitivity index β for the studied Zr-based BMGs. The plastic strain εp is in the region of 0.2%–19.1% for these Zr-based BMGs. Both yield strength σy and fracture strength σf are smallest for Zr55Al8.9Ni7.3Cu28.8 BMG whose εp is largest among all studied Zr-based BMGs and reaches up to 19.1%. In addition, the mechanism for the large difference of the plasticity among the studied Zr-based BMGs is also discussed.

Room-temperature plasticity is one of important requirements for the engineering applications of the BMGs. Majority of Zr-based BMGs are room-temperature brittle although good room-temperature plasticity could be achieved through compositional design [3][4][5][6][7][8][9][10], minor addition [13], cold-rolling [14,15] and sand blasting [16]. For example, the room-temperature compression plasticity of Zr 70 Ni 16 Cu 6 Al 8 and Zr 70 Al 8 Cu 13.5 Ni 8.5 BMGs can reach up to 40% [6] and 70% [4], respectively. More interestingly, Liu et al. [3] developed three Zr-Al-Ni-Cu BMGs with super plasticity by carefully adjusting the composition. In addition, critical diameter (d c ) for the formation of amorphous state is another crucial factor for engineering applications of the BMGs. However, increasing efforts have claimed that there is a discrepancy between the plasticity and the d c for the BMG [17]. For example, the d c of Zr 55 Al 10 Ni 5 Cu 30 glass forming alloy can reach up to 30 cm [2], but its room-temperature plasticity is nearly zero [18,19]. Although the super plasticity can be achieved in Zr 61.88 Cu 18 Ni 10.12 Al 10 , Zr 64. 13 Cu 15.75 Ni 10.12 Al 10 and Zr 62 Cu 15.5 Ni 12.5 Al 10 BMGs, their d c s are only 5 mm [3]. Thus, it is a significant and challenging work for how to design Zr-Al-Ni-Cu BMGs with the combination of large d c and good room-temperature plasticity.
In the present work, a group of Zr-Al-Ni-Cu BMGs with room-temperature plasticity was carefully designed. It is found that the critical dimension and maximum plastic strain of the studied Zr-based BMGs can reach up to 8 mm and 19.1%, respectively. The present design concept would be a useful strategy for developing the BMGs with high GFA and plasticity.

Compositional Design
A method, namely weak constraint condition proposed by Cai et al. [20], was used to design a group of Zr-Al-Ni-Cu BMGs whose d c could reach up to 7.5 mm [20]. In addition, Cai et al. [20][21][22] found that e/a = ř C iˆEi (C i and E i are atomic fraction and conduction electron concentration of i-th element, respectively), R a = ř C iˆRi (R i is atomic radius of i-th element) and ∆H m = ř C iˆ∆ H mi (∆H mi is melting heat of i-th element) were strongly related with glass forming ability (GFA) of glass forming alloys. Recently, Cai et al. [23] developed a group of Zr-based BMGs with network structure by using the weak constraint condition method. Thus, the studied Zr-Al-Ni-Cu BMGs were developed by using the conditions of e/a = 1.38, R a = 1.496 nm, ř C i = 1, and ∆H m « 19 kJ¨mol´1. In addition, artificial neural network models were used to further optimize the compositions of these Zr-based BMGs [24]. The resultant compositions (in at %) of the studied Zr-Al-Ni-Cu glass forming alloys were listed in Table 1.

Materials and Methods
Pre-alloyed Zr-Al-Ni-Cu ingots with nominal compositions (see Table 1) were prepared by arc melting mixtures of Zr (99.99 wt %), Ni (99.99 wt %), Cu (99.99 wt %), and Al (99.99 wt %) in the Ti-gettered high purity argon atmosphere. Φ2 and Φ8 mm rods were prepared by suction casting into a water-cooled copper mold. The structures of the as-cast and DSC-tested samples were characterized by X-ray diffraction (XRD) using an X' Pert Pro MPD diffractometer with Cu-Kα (Holland Philips Company, Eindhoven, The Netherlands). The thermal analysis was conducted by DSC-404C differential scanning calorimeter (NETZSCH-Gerätebau GmbH, Bavarian State, Germany) at heating rates of 5-80 K¨min´1, respectively. Room-temperature uniaxial compression tests were performed on Φ2ˆ4 mm BMGs with a gauge aspect ratio of 2:1 using an Instron 3369 testing machine (Instron Corporation, Norwood, GA, USA) at a strain rate of 1ˆ10´5 s´1. Two polished end surfaces of the samples for the compression tests were parallel each other and vertical to the symmetry axis. Fracture morphologies of Φ2 mm samples were examined by SIRION 200 scanning electron microscopy (SEM, Holland FEI Company, Eindhoven, The Netherlands). It should be noted that at least three samples for all studied BMGs were examined in order to obtain reliable results. Table 1. Glass transition temperature T g , crystallization temperature T x , supercooled liquid region ∆T x , solidus temperature T m , liquidus temperature T L , reduced glass transition temperature T rg = T g /T L , and ∆T = T L´Tg at a heating rate of 30 K¨min´1, crystallization enthalpy ∆H x , activation energy for T g (E g ), activation energy for T x (E x ), sensitivity of T g and T x to heating rate (β g and β x ), fragility index m, yield strength σ y , fracture strength σ f , and plastic strain ε p .

Metallic Glasses
T g (K)  Figure 1 presents XRD patterns for Φ8 mm Zr-based glass forming alloys, which indicates that these Φ8 mm Zr-based glass forming alloys are in glassy states because no apparent crystalline Bragg peaks are observed.

Results
Figure 1 presents XRD patterns for Φ8 mm Zr-based glass forming alloys, which indicates that these Φ8 mm Zr-based glass forming alloys are in glassy states because no apparent crystalline Bragg peaks are observed. It is obvious from the DSC curves, as shown in Figure 2, that both glass transition and sharp crystallization events can be clearly observed at heating rates of 5-80 K•min −1 . It further confirms the glassy nature of these Zr-based BMGs. In addition, there is a single crystallization event in the studied temperature region for all heating rates, indicating one step crystallization procedure for all Zr-based BMGs. It is obvious from the DSC curves, as shown in Figure 2, that both glass transition and sharp crystallization events can be clearly observed at heating rates of 5-80 K¨min´1. It further confirms the glassy nature of these Zr-based BMGs. In addition, there is a single crystallization event in the studied temperature region for all heating rates, indicating one step crystallization procedure for all Zr-based BMGs.

Results
Figure 1 presents XRD patterns for Φ8 mm Zr-based glass forming alloys, which indicates that these Φ8 mm Zr-based glass forming alloys are in glassy states because no apparent crystalline Bragg peaks are observed. It is obvious from the DSC curves, as shown in Figure 2, that both glass transition and sharp crystallization events can be clearly observed at heating rates of 5-80 K•min −1 . It further confirms the glassy nature of these Zr-based BMGs. In addition, there is a single crystallization event in the studied temperature region for all heating rates, indicating one step crystallization procedure for all Zr-based BMGs.    The characteristic temperatures including glass transition temperature (T g ), crystallization temperature (T x ), crystallization peak temperature (T p ) and supercooled liquid region (∆T x = T x´Tg ) were carefully determined according to the DSC results (see Figure 2). The dependences of the characteristic temperatures on the heating rate are shown in Figure 3. These characteristic temperatures all increase with increasing heating rate for the studied Zr-based BMGs. The T g decreases with increasing Zr content for all heating rates, as shown in Figure 3a. Both T x and T p firstly increase with increasing Zr content and then decrease when the Zr content exceeds 51.5 at % for all heating rates, as shown in Figure 3b,c. However, there are complex relationships between the ∆T x and the Zr content for the studied heating rates, as shown in Figure 3d. The ∆T x for Zr 51 Al 14.2 Ni 15.9 Cu 18.9 BMG is the smallest among the studied Zr-based BMGs when the heating rate is less than 30 K¨min´1. When the heating rate exceeds 30 K¨min´1, the ∆T x for Zr 53 Al 11.6 Ni 11.7 Cu 23.7 BMG is the smallest among the studied Zr-based BMGs. The ∆T x for Zr 55 Al 8.9 Ni 7.3 Cu 28.8 BMG is the largest among the studied Zr-based BMGs when the heating rate is less than 40 K¨min´1. Zr 54 Al 10.2 Ni 9.4 Cu 26.4 BMG has the largest ∆T x among the studied Zr-based BMGs at a heating rate of 80 K¨min´1. The characteristic temperatures including glass transition temperature (Tg), crystallization temperature (Tx), crystallization peak temperature (Tp) and supercooled liquid region (ΔTx = Tx − Tg) were carefully determined according to the DSC results (see Figure 2). The dependences of the characteristic temperatures on the heating rate are shown in Figure 3. These characteristic temperatures all increase with increasing heating rate for the studied Zr-based BMGs. The Tg decreases with increasing Zr content for all heating rates, as shown in Figure 3a. Both Tx and Tp firstly increase with increasing Zr content and then decrease when the Zr content exceeds 51.5 at % for all heating rates, as shown in Figure 3b,c. However, there are complex relationships between the ΔTx and the Zr content for the studied heating rates, as shown in Figure 3d. The ΔTx for Zr51Al14.2Ni15.9Cu18.9 BMG is the smallest among the studied Zr-based BMGs when the heating rate is less than 30 K•min −1 . When the heating rate exceeds 30 K•min −1 , the ΔTx for Zr53Al11.6Ni11.7Cu23.7 BMG is the smallest among the studied Zr-based BMGs. The ΔTx for Zr55Al8.9Ni7.3Cu28.8 BMG is the largest among the studied Zr-based BMGs when the heating rate is less than 40 K•min −1 . Zr54Al10.2Ni9.4Cu26.4 BMG has the largest ΔTx among the studied Zr-based BMGs at a heating rate of 80 K•min −1 .   Table 1. As shown in Figure 4, an endothermal event for the glass transition procedure can be clearly seen for all studied Zr-based BMGs. The Tg decreases with increasing Zr content. The Tx slightly increases and then decreases with increasing Zr content when the Zr content exceeds 51.5 at %. Table 1 lists the average crystallization enthalpy (ΔHx) for the studied Zr-based BMGs at heating rates of 5-80 K•min −1 . The ΔHx is smallest for Zr55Al8.9Ni7.3Cu28.8 BMG and largest for Zr52.5Al12.2Ni12.6Cu22.7 BMG among the studied Zr-based BMGs.   Table 1. As shown in Figure 4, an endothermal event for the glass transition procedure can be clearly seen for all studied Zr-based BMGs. The T g decreases with increasing Zr content. The T x slightly increases and then decreases with increasing Zr content when the Zr content exceeds 51.5 at %. Table 1 lists the average crystallization enthalpy (∆H x ) for the studied Zr-based BMGs at heating rates of 5-80 K¨min´1.
The ∆H x is smallest for Zr 55 Al 8.9 Ni 7.3 Cu 28.8 BMG and largest for Zr 52.5 Al 12.2 Ni 12.6 Cu 22.7 BMG among the studied Zr-based BMGs. In addition, the melting curves at a heating rate of 30 K•min −1 are shown in Figure 5a. There are two melting procedures when the Zr content is less than 52 at %, while three melting procedures for the other Zr-based BMGs. It is well known that the eutectic alloy characterizes in a single melting procedure [25]. Figure 5a indicates that the studied Zr-based alloys are all off-eutectic compositions. The distance of two melting procedures is larger for Zr51Al14.2Ni15.9Cu18.9 and Zr51.5Al13.6Ni14.9Cu20 BMGs than for Zr52Al12.9Ni13.8Cu21.3 BMG. The second melting procedure is split into two melting procedures for Zr52.5Al12.2Ni12.6Cu22.7 and Zr53Al11.6Ni11.7Cu23.7 BMGs, while the first melting procedure is split into two melting procedures for the other BMGs. The values of the solidus temperature (Tm) and liquidus temperature (TL) are listed in Table 1. The dependences of the Tm, TL and TL-Tm on the Zr content are also presented in Figure 5b. The Tm decreases with increasing Zr content when the Zr content is less than or exceeds 52 at %, while inversely for the TL-Tm. However, there is a complex dependence of the TL on the Zr content. The TL is largest for Zr55Al8.9Ni7.3Cu28.8 BMG and smallest for Zr51Al14.2Ni15.9Cu18.9 BMG among the studied Zr-based BMGs. The activation energies of Tg, Tx, and Tp (Eg, Ex, and Ep) are calculated by Kissinger equation [26] and the corresponding fitting curves are presented in Figure 6. It is obvious from Figure 6 that the experimental data can be better fitted by Kissinger equation. The estimated Eg and Ex are listed in Table 1. In addition, the melting curves at a heating rate of 30 K¨min´1 are shown in Figure 5a. There are two melting procedures when the Zr content is less than 52 at %, while three melting procedures for the other Zr-based BMGs. It is well known that the eutectic alloy characterizes in a single melting procedure [25]. Figure 5a indicates that the studied Zr-based alloys are all off-eutectic compositions. The distance of two melting procedures is larger for Zr 51 Al 14.2 Ni 15.9 Cu 18.9 and Zr 51.5 Al 13.6 Ni 14.9 Cu 20 BMGs than for Zr 52 Al 12.9 Ni 13.8 Cu 21.3 BMG. The second melting procedure is split into two melting procedures for Zr 52.5 Al 12.2 Ni 12.6 Cu 22.7 and Zr 53 Al 11.6 Ni 11.7 Cu 23.7 BMGs, while the first melting procedure is split into two melting procedures for the other BMGs. The values of the solidus temperature (T m ) and liquidus temperature (T L ) are listed in Table 1. The dependences of the T m , T L and T L -T m on the Zr content are also presented in Figure 5b. The T m decreases with increasing Zr content when the Zr content is less than or exceeds 52 at %, while inversely for the T L -T m . However, there is a complex dependence of the T L on the Zr content. The T L is largest for Zr 55 Al 8.9 Ni 7.3 Cu 28.8 BMG and smallest for Zr 51 Al 14.2 Ni 15.9 Cu 18.9 BMG among the studied Zr-based BMGs. In addition, the melting curves at a heating rate of 30 K•min −1 are shown in Figure 5a. There are two melting procedures when the Zr content is less than 52 at %, while three melting procedures for the other Zr-based BMGs. It is well known that the eutectic alloy characterizes in a single melting procedure [25]. Figure 5a indicates that the studied Zr-based alloys are all off-eutectic compositions. The distance of two melting procedures is larger for Zr51Al14.2Ni15.9Cu18.9 and Zr51.5Al13.6Ni14.9Cu20 BMGs than for Zr52Al12.9Ni13.8Cu21.3 BMG. The second melting procedure is split into two melting procedures for Zr52.5Al12.2Ni12.6Cu22.7 and Zr53Al11.6Ni11.7Cu23.7 BMGs, while the first melting procedure is split into two melting procedures for the other BMGs. The values of the solidus temperature (Tm) and liquidus temperature (TL) are listed in Table 1. The dependences of the Tm, TL and TL-Tm on the Zr content are also presented in Figure 5b. The Tm decreases with increasing Zr content when the Zr content is less than or exceeds 52 at %, while inversely for the TL-Tm. However, there is a complex dependence of the TL on the Zr content. The TL is largest for Zr55Al8.9Ni7.3Cu28.8 BMG and smallest for Zr51Al14.2Ni15.9Cu18.9 BMG among the studied Zr-based BMGs. The activation energies of Tg, Tx, and Tp (Eg, Ex, and Ep) are calculated by Kissinger equation [26] and the corresponding fitting curves are presented in Figure 6. It is obvious from Figure 6 that the experimental data can be better fitted by Kissinger equation. The estimated Eg and Ex are listed in Table 1. The activation energies of T g , T x , and T p (E g , E x , and E p ) are calculated by Kissinger equation [26] and the corresponding fitting curves are presented in Figure 6. It is obvious from Figure 6 that the experimental data can be better fitted by Kissinger equation. The estimated E g and E x are listed in Table 1. The dependences of the Eg, Ex, and Ep on the Zr content are shown in Figure 7. It is clearly observed in Figure 7 that the dependences of the Eg, Ex, and Ep on the Zr content are all complex. The Eg is largest for Zr51Al14.2Ni15.9Cu18.9 BMG, while smallest for Zr55Al8.9Ni7.3Cu28.8 BMG. The Ex is smallest for Zr51Al14.2Ni15.9Cu18.9 BMG, while largest for Zr51.5Al13.6Ni14.9Cu20 BMG. The Ep is largest for Zr51Al14.2Ni15.9Cu18.9 BMG, while smallest for Zr53.5Al10.9Ni10.6Cu25 BMG. The dependences of the E g , E x , and E p on the Zr content are shown in Figure 7. It is clearly observed in Figure 7 that the dependences of the E g , E x , and E p on the Zr content are all complex. The E g is largest for Zr 51 Al 14  In addition, the dependences of the characteristic temperatures (Tg, Tx and Tp) on the lnB (B is heating rate) were investigated, as shown in Figure 8. It was found from Figure 8 that the characteristic temperatures (Tg, Tx and Tp) and the lnB can be well fitted by the Lasocka's equation [27]: (T is characteristic temperature, α and β are constants). Obviously, β can reflect the sensitivity of the characteristic temperature to the B. The higher the β, the more is the sensitivity of the characteristic temperature to the B. The βg and βx values are listed in Table 1. In addition, the dependences of the characteristic temperatures (T g , T x and T p ) on the lnB (B is heating rate) were investigated, as shown in Figure 8. It was found from Figure 8 that the characteristic temperatures (T g , T x and T p ) and the lnB can be well fitted by the Lasocka's equation [27]: T " α`βlnB (T is characteristic temperature, α and β are constants). Obviously, β can reflect the sensitivity of the characteristic temperature to the B. The higher the β, the more is the sensitivity of the characteristic temperature to the B. The β g and β x values are listed in Table 1. In addition, the dependences of the characteristic temperatures (Tg, Tx and Tp) on the lnB (B is heating rate) were investigated, as shown in Figure 8. It was found from Figure 8    The dependences of the β values on the Zr content are presented in Figure 9. Both βx and βp are larger than the βg, indicating that the crystallization procedure is more sensitive to the heating rate than the glass transition procedure. The βg is smallest for Zr52Al12.9Ni13.8Cu21.3 BMG, while largest for Zr55Al8.9Ni7.3Cu28.8 BMG. It indicates that the glass transition to the heating rate for the latter has largest sensitivity, while smallest for the former. The βx is largest for Zr51Al14.2Ni15.9Cu18.9 BMG, while smallest for Zr52.5Al12.2Ni12.6Cu22.7 BMG. The βp is smallest for Zr55Al8.9Ni7.3Cu28.8 BMG, while largest for Zr53.5Al10.9Ni10.6Cu25 BMG. The relationships between the activation energy and the β were also investigated, as shown in Figure 10. Obviously, the activation energy decreases with increasing β for all studied Zr-based BMGs, which is similar to the results for Zr-Al-Ni-Cu BMGs [28] and Cu-based metallic glasses [29]. Moreover, the liquid behaviors of the alloys were studied by the fragility index m. The m values of the alloys were estimated by  The dependences of the β values on the Zr content are presented in Figure 9. Both β x and β p are larger than the β g , indicating that the crystallization procedure is more sensitive to the heating rate than the glass transition procedure. The β g is smallest for Zr 52 Al 12  The dependences of the β values on the Zr content are presented in Figure 9. Both βx and βp are larger than the βg, indicating that the crystallization procedure is more sensitive to the heating rate than the glass transition procedure. The βg is smallest for Zr52Al12.9Ni13.8Cu21.3 BMG, while largest for Zr55Al8.9Ni7.3Cu28.8 BMG. It indicates that the glass transition to the heating rate for the latter has largest sensitivity, while smallest for the former. The βx is largest for Zr51Al14.2Ni15.9Cu18.9 BMG, while smallest for Zr52.5Al12.2Ni12.6Cu22.7 BMG. The βp is smallest for Zr55Al8.9Ni7.3Cu28.8 BMG, while largest for Zr53.5Al10.9Ni10.6Cu25 BMG. The relationships between the activation energy and the β were also investigated, as shown in Figure 10. Obviously, the activation energy decreases with increasing β for all studied Zr-based BMGs, which is similar to the results for Zr-Al-Ni-Cu BMGs [28] and Cu-based metallic glasses [29]. Moreover, the liquid behaviors of the alloys were studied by the fragility index m. The m values of the alloys were estimated by  Table 1. Obviously, the m value is largest for Zr51Al14.2Ni15.9Cu18.9 BMG, while smallest for Zr55Al8.9Ni7.3Cu28.8 BMG. The relationships between the activation energy and the β were also investigated, as shown in Figure 10. Obviously, the activation energy decreases with increasing β for all studied Zr-based BMGs, which is similar to the results for Zr-Al-Ni-Cu BMGs [28] and Cu-based metallic glasses [29]. Moreover, the liquid behaviors of the alloys were studied by the fragility index m. The m values of the alloys were estimated by m " [30] (D is the strength parameter, T 0 is the VF temperature, and T g is glass transition temperature at a heating rate of 5 K¨min´1). The D and T 0 are obtained through the best fitting of the tested T g values at different heating rates according to the Vogel-Fulcher (VF) relation [30]: BpT g q " AexprDT 0 {pT 0´Tg qs (A is a constant). The estimated m values for these Zr-based BMGs are listed in Table 1. Obviously, the m value is largest for Zr 51 Al 14.2 Ni 15.9 Cu 18.9 BMG, while smallest for Zr 55 Al 8.9 Ni 7.3 Cu 28.8 BMG. Mechanical properties of the studied Zr-based alloys were investigated by room-temperature uniaxial compression tests at a primary strain rate of 1 × 10 −5 s −1 . Stress-strain curves are shown in Figure 11 and the corresponding mechanical properties are listed in Table 1. It is obvious from Figure 11 that the plasticity can be clearly observed for all Zr-based BMGs and depends on the composition of the BMG. As shown in Table 1, the plastic strain (εp) is in the region of 0.2%~19.1% for these Zr-based BMGs. The εp is largest for Zr55Al8.9Ni7.3Cu28.8 BMG, while smallest for Zr51Al14.2Ni15.9Cu18.9 BMG. The yield strength (σy) can reach up to 2000 MPa for Zr51Al14.2Ni15.9Cu18.9 and Zr52Al12.9Ni13.8Cu21.3 BMGs whose εp is less than 1.0%. The fracture strength (σf) is more than 2000 MPa for the BMGs with 51-53 at % Zr. Interestingly, both σy and σf are smallest for Zr55Al8.9Ni7.3Cu28.8 BMG whose εp is largest among the studied Zr-based BMGs. Moreover, fracture surfaces and side surfaces for the fractured BMGs were investigated by SEM. Typical side surfaces for these fractured Zr-based BMGs are shown in Figure 12a-c. The shear bands can be observed on the side surfaces for all Zr-based BMGs. The magnitude and the density of the shear band increase with increasing plasticity. As for the Zr-based BMGs whose εp is less than 2.0%, Mechanical properties of the studied Zr-based alloys were investigated by room-temperature uniaxial compression tests at a primary strain rate of 1ˆ10´5 s´1. Stress-strain curves are shown in Figure 11 and the corresponding mechanical properties are listed in Table 1. It is obvious from Figure 11 that the plasticity can be clearly observed for all Zr-based BMGs and depends on the composition of the BMG. As shown in Table 1, the plastic strain (ε p ) is in the region of 0.2%~19.1% for these Zr-based BMGs. The ε p is largest for Zr 55 Al 8 BMGs whose ε p is less than 1.0%. The fracture strength (σ f ) is more than 2000 MPa for the BMGs with 51-53 at % Zr. Interestingly, both σ y and σ f are smallest for Zr 55 Al 8.9 Ni 7.3 Cu 28.8 BMG whose ε p is largest among the studied Zr-based BMGs. Mechanical properties of the studied Zr-based alloys were investigated by room-temperature uniaxial compression tests at a primary strain rate of 1 × 10 −5 s −1 . Stress-strain curves are shown in Figure 11 and the corresponding mechanical properties are listed in Table 1. It is obvious from Figure 11 that the plasticity can be clearly observed for all Zr-based BMGs and depends on the composition of the BMG. As shown in Table 1, the plastic strain (εp) is in the region of 0.2%~19.1% for these Zr-based BMGs. The εp is largest for Zr55Al8.9Ni7.3Cu28.8 BMG, while smallest for Zr51Al14.2Ni15.9Cu18.9 BMG. The yield strength (σy) can reach up to 2000 MPa for Zr51Al14.2Ni15.9Cu18.9 and Zr52Al12.9Ni13.8Cu21.3 BMGs whose εp is less than 1.0%. The fracture strength (σf) is more than 2000 MPa for the BMGs with 51-53 at % Zr. Interestingly, both σy and σf are smallest for Zr55Al8.9Ni7.3Cu28.8 BMG whose εp is largest among the studied Zr-based BMGs. Moreover, fracture surfaces and side surfaces for the fractured BMGs were investigated by SEM. Typical side surfaces for these fractured Zr-based BMGs are shown in Figure 12a-c. The shear bands can be observed on the side surfaces for all Zr-based BMGs. The magnitude and the density of the shear band increase with increasing plasticity. As for the Zr-based BMGs whose εp is less than 2.0%, Figure 11. Room-temperature uniaxial compression stress-strain curves for the studied Zr-based BMGs.
Moreover, fracture surfaces and side surfaces for the fractured BMGs were investigated by SEM. Typical side surfaces for these fractured Zr-based BMGs are shown in Figure 12a-c. The shear bands can be observed on the side surfaces for all Zr-based BMGs. The magnitude and the density of the shear band increase with increasing plasticity. As for the Zr-based BMGs whose ε p is less than 2.0%, the shear bands are scarcely and propagate vertical to the loading direction (see Figure 12a). The intersected shear bands can be observed for the Zr-based BMGs whose ε p is more than 4.0%, as shown in Figure 12b,c. In addition, Figure 12d-f presents typical fracture surfaces for these Zr-based BMGs. The fracture surfaces include smooth regions and vein-like regions for all BMGs. Although the grooves can be clearly observed for all BMGs, the depth and the width of the groove increase with increasing plasticity. More interestingly, the grooves are nearly parallel with each other for the BMGs whose plastic strain is less than 2.0%, while interdigitate each other for the BMGs whose plastic strain exceeds 4.0%.
the shear bands are scarcely and propagate vertical to the loading direction (see Figure 12a). The intersected shear bands can be observed for the Zr-based BMGs whose εp is more than 4.0%, as shown in Figure 12b,c. In addition, Figure 12d-f presents typical fracture surfaces for these Zr-based BMGs. The fracture surfaces include smooth regions and vein-like regions for all BMGs. Although the grooves can be clearly observed for all BMGs, the depth and the width of the groove increase with increasing plasticity. More interestingly, the grooves are nearly parallel with each other for the BMGs whose plastic strain is less than 2.0%, while interdigitate each other for the BMGs whose plastic strain exceeds 4.0%.

Discussion
As shown in Figure 11 and Table 1, there is large different plasticity for the studied Zr-based BMGs. The reasons would be as follows. Firstly, high GFA is advantageous of the homogeneous distribution of the atoms in alloys [17], indicating that the structural heterogeneity in the atomic scale would be difficult for the formation in the BMG with high GFA. The growing results have claimed that the heterogeneity would generally result in good plasticity [3,17,31,32]. In addition, Lu and Liu [33] found that the reduced glass transition temperature Trg (=Tg/TL) is a good indicator for the GFA of the glass forming alloy. The larger the Trg, the higher is the GFA of the BMG. It is obvious from Figure 13 and Table 1 that the Trg nearly decreases with increasing Zr content. It indicates that the GFA decreases with increasing Zr content.
It is well known that the GFA of the glass forming alloy originates from the microstructural configuration of the atomic arrangement related with the atomic size, type and interaction among the atoms. The atom radii for Zr, Al, Ni and Cu are 0.162, 0.143, 0.125 and 0.128 nm, respectively. The mixing enthalpy is −23 kJ/mol for Cu-Zr, −49 kJ/mol for Ni-Zr, −44 kJ/mol for Zr-Al, −22 kJ/mol for Al-Ni, −1 kJ/mol for Al-Cu, and +4 kJ/mol for Cu-Ni, respectively [33]. The packing of the atoms in alloys would be changed by the different atom content. Since there is large difference in the mixing enthalpy among atom pairs for the studied Zr-based alloys, the magnitude of short range orders (SRO) would be also changed by varying the atom content in alloys. In addition, there is large negative mixing enthalpy for Zr-Ni, Zr-Al, Zr-Cu and Al-Ni, the SROs corresponding to these atom pairs would be easily formed. XRD analysis (see Figure 14) was performed on the studied BMGs subjected to DSC test at a heating rate of 80 K•min −1 in order to roughly estimate the category and magnitude

Discussion
As shown in Figure 11 and Table 1, there is large different plasticity for the studied Zr-based BMGs. The reasons would be as follows. Firstly, high GFA is advantageous of the homogeneous distribution of the atoms in alloys [17], indicating that the structural heterogeneity in the atomic scale would be difficult for the formation in the BMG with high GFA. The growing results have claimed that the heterogeneity would generally result in good plasticity [3,17,31,32]. In addition, Lu and Liu [33] found that the reduced glass transition temperature T rg (=T g /T L ) is a good indicator for the GFA of the glass forming alloy. The larger the T rg , the higher is the GFA of the BMG. It is obvious from Figure 13 and Table 1 that the T rg nearly decreases with increasing Zr content. It indicates that the GFA decreases with increasing Zr content.
It is well known that the GFA of the glass forming alloy originates from the microstructural configuration of the atomic arrangement related with the atomic size, type and interaction among the atoms. The atom radii for Zr, Al, Ni and Cu are 0.162, 0.143, 0.125 and 0.128 nm, respectively. The mixing enthalpy is´23 kJ/mol for Cu-Zr,´49 kJ/mol for Ni-Zr,´44 kJ/mol for Zr-Al, 22 kJ/mol for Al-Ni,´1 kJ/mol for Al-Cu, and +4 kJ/mol for Cu-Ni, respectively [33]. The packing of the atoms in alloys would be changed by the different atom content. Since there is large difference in the mixing enthalpy among atom pairs for the studied Zr-based alloys, the magnitude of short range orders (SRO) would be also changed by varying the atom content in alloys. In addition, there is large negative mixing enthalpy for Zr-Ni, Zr-Al, Zr-Cu and Al-Ni, the SROs corresponding to these atom pairs would be easily formed. XRD analysis (see Figure 14) was performed on the studied BMGs subjected to DSC test at a heating rate of 80 K¨min´1 in order to roughly estimate the category and magnitude of the SRO in the studied Zr-based BMGs. A single NiZr 2 phase can be clearly observed in Zr 51 Al 14.2 Ni 15.9 Cu 18.9 alloy. A new Al 2 Zr 3 phase precipitates from the BMG matrix and its magnitude increases with increasing Zr content when the Zr content is less than 52.5%. The magnitude of Al 2 Zr 3 phase decreases when the Zr content exceeds 52.5%, then increases with increasing Zr content. It would be noted that another new CuZr 2 phase precipitates from the BMG matrix when the Zr content reaches up to 54%. At the same time, the magnitude of CuZr 2 phase increases with increasing Zr content. Thus the structural heterogeneity corresponding to the SRO is difficultly formed in Zr 51 Al 14.2 Ni 15.9 Cu 18.9 BMG, while inversely in Zr 55 Al 8.9 Ni 7.3 Cu 28.8 BMG. It indicates that the plasticity would be smaller for the former than for the latter (see Table 1).
Materials 2016, 9,408 13 of 17 of the SRO in the studied Zr-based BMGs. A single NiZr2 phase can be clearly observed in Zr51Al14.2Ni15.9Cu18.9 alloy. A new Al2Zr3 phase precipitates from the BMG matrix and its magnitude increases with increasing Zr content when the Zr content is less than 52.5%. The magnitude of Al2Zr3 phase decreases when the Zr content exceeds 52.5%, then increases with increasing Zr content. It would be noted that another new CuZr2 phase precipitates from the BMG matrix when the Zr content reaches up to 54%. At the same time, the magnitude of CuZr2 phase increases with increasing Zr content. Thus the structural heterogeneity corresponding to the SRO is difficultly formed in Zr51Al14.2Ni15.9Cu18.9 BMG, while inversely in Zr55Al8.9Ni7.3Cu28.8 BMG. It indicates that the plasticity would be smaller for the former than for the latter (see Table 1).  leading to the increase of the probability for the formation of the structural heterogeneity. As shown in Figure 13 and Table 1, the T  is largest for Zr55Al8.9Ni7.3Cu28.8 BMG and smallest for Zr51Al14.2Ni15.9Cu18.9 BMG. Thus, the structural heterogeneity is difficultly formed in Figure 13. Dependences of the T rg and T L -T g on the Zr content.
Materials 2016, 9,408 13 of 17 of the SRO in the studied Zr-based BMGs. A single NiZr2 phase can be clearly observed in Zr51Al14.2Ni15.9Cu18.9 alloy. A new Al2Zr3 phase precipitates from the BMG matrix and its magnitude increases with increasing Zr content when the Zr content is less than 52.5%. The magnitude of Al2Zr3 phase decreases when the Zr content exceeds 52.5%, then increases with increasing Zr content. It would be noted that another new CuZr2 phase precipitates from the BMG matrix when the Zr content reaches up to 54%. At the same time, the magnitude of CuZr2 phase increases with increasing Zr content. Thus the structural heterogeneity corresponding to the SRO is difficultly formed in Zr51Al14.2Ni15.9Cu18.9 BMG, while inversely in Zr55Al8.9Ni7.3Cu28.8 BMG. It indicates that the plasticity would be smaller for the former than for the latter (see Table 1).  leading to the increase of the probability for the formation of the structural heterogeneity. As shown in Figure 13 and Table 1, the T  is largest for Zr55Al8.9Ni7.3Cu28.8 BMG and smallest for Zr51Al14.2Ni15.9Cu18.9 BMG. Thus, the structural heterogeneity is difficultly formed in Secondly, the dwell time in the liquid state can be characterized by ∆T " T L´Tg : the larger the ∆T, the longer the dwell time. It would result in the long time for the atomic diffusion, eventually leading to the increase of the probability for the formation of the structural heterogeneity. As shown in Figure 13 and Table 1, the ∆T is largest for Zr 55 Al 8.9 Ni 7.3 Cu 28.8 BMG and smallest for Zr 51 Al 14.2 Ni 15.9 Cu 18.9 BMG. Thus, the structural heterogeneity is difficultly formed in Zr 51 Al 14.2 Ni 15.9 Cu 18.9 BMG, while inversely in Zr 55 Al 8.9 Ni 7.3 Cu 28.8 BMG. It would lead to larger plasticity for the latter than for the former.
Thirdly, it is well known that the plastic deformation of the BMG can be regarded as a localized transition from glass to supercooled liquid induced by external stress or temperature increase [17,34]. The external stress can lead to the temperature increase in the shear band [17,[35][36][37][38]. The BMGs with low T g could be easily gotten into the supercooled liquid region by the external stress and/or temperature. As shown in Figure 15, local melting phenomenon can be clearly observed for Zr 55 Al 8.9 Ni 7.3 Cu 28.8 BMG and not for the other studied Zr-based BMGs, which is advantageous of the enhancement of the plasticity [36]. Zr51Al14.2Ni15.9Cu18.9 BMG, while inversely in Zr55Al8.9Ni7.3Cu28.8 BMG. It would lead to larger plasticity for the latter than for the former. Thirdly, it is well known that the plastic deformation of the BMG can be regarded as a localized transition from glass to supercooled liquid induced by external stress or temperature increase [17,34]. The external stress can lead to the temperature increase in the shear band [17,[35][36][37][38]. The BMGs with low Tg could be easily gotten into the supercooled liquid region by the external stress and/or temperature. As shown in Figure 15, local melting phenomenon can be clearly observed for Zr55Al8.9Ni7.3Cu28.8 BMG and not for the other studied Zr-based BMGs, which is advantageous of the enhancement of the plasticity [36]. In addition, the activation energy for the glass transition Eg is equal to that for plastic flow Ep [17]. Low Tg or Eg would easily lead to yielding or plastic flow of the BMG. As shown in Table 1, both Tg and Eg are largest for Zr51Al14.2Ni15.9Cu18.9 BMG and smallest for Zr55Al8.9Ni7.3Cu28.8 BMG. Interestingly, the εp increases with decreasing Eg for the studied Zr-based BMGs, as shown in Figure  16. At the same time, large ΔTx indicates good stability of supercooled liquid, resulting in large plastic flow for the BMG. It is clearly seen from Table 1 that the ΔTx is largest for Zr55Al8.9Ni7.3Cu28.8 BMG and smallest for Zr51Al14.2Ni15.9Cu18.9 BMG. Moreover, the m value can be considered as the activation energy of the flow for the supercooled liquid and the steepness of the change of the viscosity (η) with the temperature (T) at Tg according to The smaller is the m, the lower is the activation energy and the longer is the time for the flow of the supercooled liquid. That is to say, the small m is advantageous of the plastic flow of the metallic glass and would result in large plasticity. As shown in Table 1, the m is largest for Zr51Al14.2Ni15.9Cu18.9 BMG and smallest for Zr55Al8.9Ni7.3Cu28.8 BMG. In fact, it can be clearly seen from Figure 16 that the εp increases with decreasing m for the studied Zr-based BMGs. In a word, the plasticity is largest for Zr55Al8.9Ni7.3Cu28.8 BMG and smallest for Zr51Al14.2Ni15.9Cu18.9 BMG. In addition, the activation energy for the glass transition E g is equal to that for plastic flow E p [17]. Low T g or E g would easily lead to yielding or plastic flow of the BMG. As shown in Table 1, both T g and E g are largest for Zr 51 Al 14.2 Ni 15.9 Cu 18.9 BMG and smallest for Zr 55 Al 8.9 Ni 7.3 Cu 28.8 BMG. Interestingly, the ε p increases with decreasing E g for the studied Zr-based BMGs, as shown in Figure 16. At the same time, large ∆T x indicates good stability of supercooled liquid, resulting in large plastic flow for the BMG. It is clearly seen from Table 1 that the ∆T x is largest for Zr 55 Al 8.9 Ni 7.3 Cu 28.8 BMG and smallest for Zr 51 Al 14.2 Ni 15.9 Cu 18.9 BMG. Moreover, the m value can be considered as the activation energy of the flow for the supercooled liquid and the steepness of the change of the viscosity (η) with the temperature (T) at T g according to m " Blgη BpT g {TqˇT "T g [17]: The smaller is the m, the lower is the activation energy and the longer is the time for the flow of the supercooled liquid. That is to say, the small m is advantageous of the plastic flow of the metallic glass and would result in large plasticity. As shown in Table 1, the m is largest for Zr 51 Al 14.2 Ni 15.9 Cu 18.9 BMG and smallest for Zr 55 Al 8.9 Ni 7.3 Cu 28.8 BMG. In fact, it can be clearly seen from Figure 16 that the ε p increases with decreasing m for the studied Zr-based BMGs. In a word, the plasticity is largest for Zr 55 Al 8.9 Ni 7.3 Cu 28.8 BMG and smallest for Zr 51 Al 14.2 Ni 15.9 Cu 18.9 BMG.

Conclusions
The studied Zr-Al-Ni-Cu BMGs can be cast into at least Φ8 mm and has a single crystallization event for all heating rates in the studied temperature region. The Tg decreases with increasing Zr content for all heating rates. Both Tx and Tp firstly increase with increasing Zr content and then decrease when the Zr content exceeds 51.5 at % for all heating rates. The ΔHx is smallest for Zr55Al8.9Ni7.3Cu28.8 BMG, while largest for Zr52.5Al12.2Ni12.6Cu22.7 BMG among the studied Zr-based BMGs.
There are two melting procedures when the Zr content is less than 52 at %, while three melting procedures for the other Zr-based BMGs. The distance of two melting procedures is larger for Zr51Al14.2Ni15.9Cu18.9 and Zr51.5Al13.6Ni14.9Cu20 BMGs than for Zr52Al12.9Ni13.8Cu21.3 BMG. The second melting procedure is split into two melting procedures for Zr52.5Al12.2Ni12.6Cu22.7 and Zr53Al11.6Ni11.7Cu23.7 BMGs, while the first melting procedure is split into two melting procedures for the other BMGs. The Tm decreases with increasing Zr content when the Zr content is less than or exceeds 52 at %, while inversely for the TL-Tm. The TL is largest for Zr55Al8.9Ni7.3Cu28.8 BMG and smallest for Zr51Al14.2Ni15.9Cu18.9 BMG among the studied Zr-based BMGs. The Trg decreases with increasing Zr content when the Zr content is less than or exceeds 52 at %, while inversely for the TL-Tg.
The εp is in the region of 0.2-19.1% for these Zr-based BMGs. The εp is largest for Zr55Al8.9Ni7.3Cu28.8 BMG, while smallest for Zr51Al14.2Ni15.9Cu18.9 BMG. The σy can reach up to 2000 MPa for Zr51Al14.2Ni15.9Cu18.9 and Zr52Al12.9Ni13.8Cu21.3 BMGs whose εp is less than 1.0%. The σf is more than 2000 MPa for the BMGs with 51~53 at % Zr. Both σy and σf are smallest for Zr55Al8.9Ni7.3Cu28.8 BMG whose εp is largest among the studied Zr-based BMGs. The depth and the width of the groove increase with increasing plasticity. The grooves are nearly parallel with each other for the BMGs whose εp is less than 2.0%, while interdigitate each other for the BMGs whose εp exceeds 4.0%.