Introduction

Bulk Metallic Glasses (BMGs) represent a relatively new class of multi-component alloys with amorphous atomic structure and remarkable properties. These include exceptionally high strength and elastic limit, excellent wear and corrosion resistance1,2,3,4,5,6,7,8,9, soft magnetism, and unique ability for thermoplastic forming10. BMGs have a distinct disordered microstructure with structural changes induced by a variation in their free volume content, in contrast to the long-range atomic order in crystalline counterparts5. The different structural states of a BMG on a potential energy landscape are shown schematically in Fig. 1. A BMG in its “rejuvenated” state has higher energy and more loosely packed atoms compared to its as-cast state. On the contrary, the BMG in its “relaxed” state has relatively lower potential energy and higher atomic packing density as shown in Fig. 1. Structural relaxation typically leads to embrittlement of metallic glasses with plasticity reduction and decrease in fracture toughness11,12. Rejuvenation, in contrast, has been shown to improve the plasticity of metallic glasses by introducing atomic-scale defects and preventing brittle failure13,14,15,16,17,18,19,20. BMGs often exhibit limited plasticity because of the lack of dislocation-based deformation mechanisms when deformed at temperatures below their glass transition. Enhancement of ductility in BMGs, significant for structural applications, continues to be a challenge. Several thermomechanical processing routes have been explored to this end, including shot-peening21, cold rolling22,23, high-pressure treatment24,25,26,27,28,29,30, and triaxial compression31. Annealing below the glass transition temperature (Tg) may lead to embrittlement in some cases due to structural relaxation11,32,33. However, annealing above Tg34,35,36 and cryogenic thermal cycling37,38,39,40 has been shown to enhance plasticity. In addition, micro-alloying approach has also been explored for improving the mechanical properties of BMGs, such as with the addition of a small fraction of Ag41, Nb42, Hf43, Ti44, and Fe45. Substitution of Ti with Ag improved the glass forming ability (GFA) and mechanical properties of Zr-based BMGs46. Ti addition was found to change the short-range order in Zr-Al-Ni-Cu glass-forming alloys47 and in some cases enhance GFA and plasticity44,48.

Fig. 1
figure 1

Potential energy landscape illustrating the various structural states in bulk metallic glasses. Rejuvenation leads to an increase in potential energy and higher free volume content while relaxation leads to decrease in potential energy and lower free volume content compared to its as-cast state.

Full size image

Here, we report on the deformation mechanisms of a model Zr-Cu-Al BMG after cryogenic temperature cycling (thermal rejuvenation) studied with and without the addition of 1 at. % titanium. The Zr-Cu-Al system was chosen because it is the basis for development of numerous alloys in the Zr-based bulk metallic glass family and is a canonical example of a good glass former with a large processing window (wide supercooled liquid region)49,50. Moreover, it is free from toxic Be and does not have expensive noble metals like Pt, Pd, or rare earth elements, making it appealing for widespread applications. A small addition of Ti was investigated for further improvement in glass forming ability (GFA) and plasticity. Mechanical behavior of the Zr-Cu-Al-(Ti) BMGs was characterized in terms of changes in hardness, modulus, uniaxial compression response, shear band formation, and time dependent deformation. There are no reports in literature on simultaneously looking at rejuvenation of BMGs from cryogenic thermal cycling and microalloying. In addition, the time-dependent deformation behavior of BMGs after cryogenic thermal cycling has not been studied. We evaluated the effect of loading rate and report on the strain rate sensitivity (SRS) and relate it to the tendency for flow localization in the Zr-Cu-Al-(Ti) BMGs. Using nanoindentation and micropillar compression, we relate several aspects of the deformation behavior to the changes in structural state after cryogenic thermal cycling. Finally, using molecular dynamics (MD) simulation, we demonstrate that cryogenic thermal cycling leads to decrease in dense packing clusters and increase in loosely packed short range order clusters.

Methods

BMG ingots with the nominal composition of Zr47Cu46Al7 and Zr47Cu45Al7Ti1 were prepared by arc-melting a mixture of pure elements with a purity of 99.9% under a titanium gettered argon atmosphere in water-cooled copper crucibles. Each ingot was melted four times to ensure uniformity in composition. Plate samples in the dimensions of 3 mm × 10 mm × 10 mm was fabricated by suction casting into a copper mold under a purified Ar atmosphere. For thermal rejuvenation, cryogenic cycling was carried out by immersing the plate samples into liquid nitrogen for 12 h followed by holding at room temperature for 12 h for a total of 30 cycles. A schematic for thermal rejuvenation is shown in Fig. 1. The as-cast and rejuvenated samples were mechanically polished to a mirror finish and then cleaned in acetone and distilled water ultrasonically for 15 min for nano-mechanical tests. Rigaku III Ultima X-ray diffractometer (XRD, Rigaku Corporation, Tokyo, Japan) with 1.54 Å wavelength Cu-Kα radiation was used to confirm the amorphous structure of the samples. Differential scanning calorimetry (DSC) (NETZSCH DSC 404C) was conducted at a constant heating rate of 20 K/min to determine the characteristic temperatures and heat flow for the as-cast (AC) and rejuvenated (Rej) BMGs. Poisson`s ratio was measured for all the BMGs using resonant ultrasonic transducers. Rectangular samples in the dimensions of 10 mm × 10 mm × 2 mm were placed between ultrasonic transducers and the longitudinal and shear velocities were measured for each alloy to determine its Poisson`s ratio.

Hardness measurements and creep tests were performed using TI-Premier Triboindenter (Bruker, Minneapolis, MN, USA) with a Diamond Berkovich tip having an included angle of 142.30° and a tip radius of 150 nm. The nanoindentation hardness of the alloys in various structural states were measured using Oliver and Pharr method51. Fused quartz was used as a standard reference sample for initial tip calibration. Static constant load hold (CLH) creep tests were performed at room temperature by ramping to a maximum load of 1000 mN at two different loading rates of 10 mN/s and 100 mN/s. The samples were held at the peak load for 600 s, followed by unloading. The high load was used to minimize surface effects and avoid the indentation size effect (ISE) on hardness. A fast loading-rate was chosen to ensure that creep primarily occurred during the dwell time. At least fifteen indents were performed in each test condition and the distance between two neighboring indents was kept greater than 100 μm to avoid the overlap of their plastic zones. Thermal drift rate was maintained below 0.05 nm/s for all the tests.

For micro-pillar compression, pillars with aspect ratio of ~ 2 (height of the pillar ~ 6 μm and diameter of the pillar ~ 2.8 μm) were milled using FEI Nova NanoLab 200 focused ion beam SEM (FIB-SEM) in several steps using Ga ion beam current ranging from 5 nA to 10 pA. The top and bottom diameter of the micro-pillars were measured. As the taper angle was small (< 2°), the top diameter was used to calculate the cross-section area for further analysis. Micro-pillar compression tests were done in PI88 SEM Pico-indenter (Bruker, Minneapolis, MN, USA) with a 5 µm diameter flat diamond punch in displacement control mode. The recorded load versus displacement was converted to engineering stress–strain curve. Three to five micro-pillars were made for each sample condition to determine the standard deviation.

MD simulations were performed using a large-scale atomic/molecular massively parallel simulator (LAMMPS) using the embedded atom model (EAM) potential for Zr-Cu-Al BMG52. The alloys were rapidly quenched at the rate of 1013 K/s to obtain amorphous structure in Zr-Cu-Al and Zr-Cu-Al-Ti BMGs. The thermodynamic processes were carried out at isothermal-isobaric (NPT) ensemble based on the Nose–Hoover thermostat and barostat. Post-processing analysis of the MD simulations was carried out using OVITO tool to analyze the type and fraction of coordination polyhedral clusters or short-range ordering.

Results

Structural and thermal characterization

Figure 2(a) shows the XRD patterns obtained for the Zr-Cu-Al and Zr-Cu-Al-Ti alloys in as-cast (AC) and rejuvenated (Rej) states. A broad peak was observed between 2θ = 30°-45° for each of the alloys supporting their fully amorphous structure. Thermal analysis of the samples was done using DSC and heat flow as a function of temperature for all the alloys are shown in Fig. 2(b). The glass transition temperature (Tg) and onset temperature of crystallization (Tx) are marked by arrows in the DSC curves. All the four samples exhibited a small and broad exothermic peak prior to the glass transition, which is attributed to the annihilation of free volume. This was followed by a wide super-cooled liquid region and an exothermic peak associated with crystallization. Figure 2(c) shows the zoomed-in view of specific heat curves obtained using the DSC plots in Fig. 2(b), with the area under the curve for each alloy representing its relaxation enthalpy. Figure 2(d) shows the enthalpy of relaxation measured for each of the four alloys indicating significant increase in free volume with 1 at. % Ti addition. The Tg, Tx, and liquidus temperature (Tl) for the Zr-Cu-Al and Zr-Cu-Al-Ti alloys in as-cast (AC) condition are listed in Table 1. The relaxation enthalpy was 24.6 J/mol for Zr-Cu-Al and increased to 48.6 J/mol after rejuvenation. Similarly, the relaxation enthalpy was 183.8 J/mol for Zr-Cu-Al-Ti and increased to 190.2 J/mol after rejuvenation. The increase in relaxation enthalpy is much lower after rejuvenation for the Zr-Cu-Al-Ti alloy. This may be attributed to the increase in disorder with the addition of the 4th element53. Therefore, further disorder introduced by thermal strain may be relatively small in comparison to that seen for the ternary Zr-Cu-Al alloy. Disorder is related to increased “confusion” resulting from an increased diversity of atomic species and their different sizes with the introduction of the 4th element, Ti. With the addition of 1 at. % Ti, the Tg decreased from 718.8 K to 695.3 K, Tx decreased from 753.5 K to 736.3 K, and Tl decreased slightly. The reduced glass transition temperature (Trg = Tg/Tl), which is sometimes used as an indicator of GFA54, remained about the same with a value of ~ 0.6 for both the alloys.

Fig. 2
figure 2

(a) XRD pattern of as-cast (AC) and rejuvenated (Rej) Zr-Cu-Al and Zr-Cu-Al-Ti BMGs showing fully amorphous structure; (b) DSC plots for the four alloys showing the glass transition temperature (Tg) and onset temperature of crystallization (Tx); (c) Zoomed-in view of the specific heat capacity (CP) curves around glass transition, with shaded area shown for the as-cast samples; (d) Relaxation enthalpy (ΔHRel) calculated from the area under the DSC curves prior to glass transition for the four alloys.

Full size image
Table 1 Characteristic temperatures for Zr-Cu-Al and Zr-Cu-Al-Ti BMGs in as-cast (AC) condition measured using a DSC, where Tg is the glass transition temperature, Tx is the onset temperature of crystallization, Tl is the liquidus temperature, and Trg is the reduced glass transition temperature.
Full size table

Mechanical behavior

The hardness and modulus for the two alloys in their different structural states are listed in Table 2. For Zr-Cu-Al, the hardness and stiffness decreased by 6.3% and 2.4%, respectively, after rejuvenation. However, Zr-Cu-Al-Ti showed smaller changes with a decrease of 2.5% in hardness and around 1% in stiffness after rejuvenation. The decrease in hardness and stiffness for the rejuvenated alloys may be attributed to local atomic rearrangement and increase in free volume as suggested by the increase in their relaxation enthalpy (Fig. 2(d)).

Table 2 Hardness, reduced modulus, and yield strength for the Zr-Cu-Al and Zr-Cu-Al-Ti BMGs in different structural states.
Full size table

Micropillar compression tests allow for in situ observation of deformation behavior and shear band formation, which is not possible in bulk tests. Also, micropillar compression tests minimize the influence of surface defects that can affect bulk sample measurements. We fabricated and tested multiple micropillars across different areas of the sample to account for any variability, which is captured in the standard deviation of the results. Representative engineering stress–strain curves for the Zr-Cu-Al and Zr-Cu-Al-Ti BMGs obtained from micro-pillar compression tests are shown in Fig. 3(a). The yield strength was calculated at the point where the first shear band was observed in situ during the experiment (corresponding to ‘*’ in the Fig. 3 (a)) and correlated with the first load drop in the load–displacement curve and the values are summarized in Table 2 for all the alloys. The average yield strength obtained from the micro-pillar compression tests for the as-cast Zr-Cu-Al was 1752 MPa, which decreased to 1722 MPa after cryogenic thermal cycling (thermal rejuvenation). The average yield strength further decreased for Zr-Cu-Al-Ti as-cast alloy to 1671 MPa and for the corresponding rejuvenated alloy to 1632 MPa. Post-yield serrations observed in the stress–strain curves during compressive deformation of amorphous alloys are related to nucleation and propagation of shear bands, while pre-yield serrations may be attributed to structural rearrangements during elastic strain55,56. Figure 3(b) shows the stress drop magnitude and frequency calculated from the engineering stress–strain plots for all the four alloys. The magnitude of stress drops decreased, and the stress drop frequency increased as the free volume in the alloys increased suggesting the enhancement in plasticity after cryogenic thermal cycling. Figures 3(c1-c16) show the in-situ SEM images during compression for the four alloys at different strain levels as indicated on the figures. The micropillar for as-cast Zr-Cu-Al deformed by accommodation of all the strain in a single major shear band as shown in Figs. 3(c1-c4). Figures 3(c5-c8) show the in-situ SEM images of the micro-pillar for rejuvenated Zr-Cu-Al indicating the formation of a few secondary shear bands in addition to the primary shear band. Similarly, plastic deformation for the as-cast Zr-Cu-Al-Ti alloy was characterized by few secondary shear bands before a major shear band formed (Figs. 3(c9-c12)). The micropillars for rejuvenated Zr-Cu-Al-Ti alloy showed multiple secondary shear bands intersecting with each other as well as with the primary shear bands (Figs. 3(c13-c16)). This indicates significantly better plasticity accommodation for rejuvenated Zr-Cu-Al-Ti as evident from its relatively homogeneous plastic deformation behavior (Fig. 3(a)), smallest magnitude of stress drops, and higher stress drop frequency (Fig. 3(b)) compared to the other cases.

Fig. 3
figure 3

(a) Engineering stress–strain curves for Zr-Cu-Al and Zr-Cu-Al-Ti alloys in as-cast (AC) and rejuvenated (Rej) states, with origin of the curves shifted by arbitrary units (a.u.) for clarity of representation, the ‘*’ corresponds to the yield point of the sample; (b) Stress drop magnitude and stress drop frequency measured from the engineering stress–strain plots for the two alloys in AC and Rej states; (c1-16) In-situ SEM images showing the change in deformation behavior for the four alloys, illustrating primary shear bands (Pri SB) and secondary shear bands (Sec SB).

Full size image

Time-dependent deformation

Time-dependent deformation behavior of the alloys was studied at room temperature using nano-indentation. To evaluate the effect of loading rate, the samples were loaded to 1000 mN at the rates of 10 mN/s and 100 mN/s. The creep mechanism was evaluated by determining the creep strain rate sensitivity (SRS), m, which is the reciprocal of stress exponent (n), defined in the conventional power-law creep as57:

$$\dot{\varepsilon } = A\sigma^{n} = A^{\prime}H^{n}$$
(1)

where \(\dot{\varepsilon }\) is the strain rate, \(\sigma\) is the applied stress, and A or \(A^{\prime}\) is a temperature-dependent material constant. The strain rate \(\mathop { \left( \varepsilon \right) }\limits\) was calculated as:

$$\mathop {\varepsilon }\limits = \frac{1}{h}\frac{dh}{{dt}}$$
(2)

where \(\frac{dh}{{dt}}\) is the first derivative of the instantaneous creep displacement (h). The displacement rate dh/dt was obtained by fitting the displacement-holding time curve during constant load holding as58:

$$h \left( t \right) = h_{0} + a\left( {t - t_{0} } \right)^{p} + kt$$
(3)

where \(h_{0}\) and \(t_{0}\) are the indentation depth and time before the holding segment, and a, p and k are fitting constants. The hardness (H) for the self-similar indentation probe like Berkovich was calculated as:

$$H = \frac{P}{{24.5 h_{c}^{2} }}$$
(4)

where P is the applied load, hc is contact depth given by hc = hmax − 0.75 P/S for Berkovich indenter and hmax and S are maximum penetration depth and material stiffness, respectively. Slope of log (H) versus log \(\left( {\dot{\varepsilon }} \right)\) was used to calculate the strain rate sensitivity, m (reciprocal of stress exponent, n), which gives a measure of the creep deformation mechanism.

The creep displacement curves obtained by fitting the experimental nanoindentation data, using Eq. 3, for the four alloys during the constant holding period are shown in Figs. 4(a) and 4(b) at the loading rate of 10 mN/s and 100 mN/s, respectively. The creep displacement initially increased sharply with time followed by a decrease in the rate of increase. Overall, the creep displacement was higher for the rejuvenated alloys compared to their as-cast counterparts. The maximum creep displacement increased with loading rate, which may be attributed to less time for relaxation process and more free volume creation at the higher loading rates59,60. Similar trend has been reported for Pd-based, Fe-based, and Ce-based BMGs59,61,62 during nanoindentation at room temperature, as well as in some high entropy alloys63. Figures 4 (a) and 4 (b) also show the corresponding creep strain rate on the secondary y-axis plotted using Eq. 2. The strain rate decreased sharply with in the first 40 s reaching values on the order of 10–4/s by the end of the test. Supplementary figures S1 and S2 show that the hardness did not change significantly with depth well before the end of the test, indicating steady state creep64. Figure 4(c) shows the dependence of creep SRS on loading rate, which is related to the tendency for flow localization in a BMG. Higher SRS indicates more resistance to localization during plastic deformation59,65. In crystalline alloys, the creep mechanism is explained in terms of dislocation glide, diffusion, and grain boundary sliding66, which are fundamentally different compared to the mechanisms in amorphous alloys59,67. At lower loading rate, the creep SRS was lower for the rejuvenated alloys compared to their as-cast counterparts, as shown in Fig. 4 (c). Relatively higher free volume content in the rejuvenated samples promotes larger number of shear bands due to easier rearrangement of atoms65. This results in localized deformation as suggested by lower creep SRS for the rejuvenated samples. Increase in loading rate increases the flow stress leading to increased deformation volume or plastic zone size around the indent. Additionally, the shorter loading time does not allow for atomic rearrangement68 and may lead to higher free volume generation59. This higher rate of free volume generation at higher loading rate leads to more homogeneous plastic deformation69 as suggested from the higher creep SRS in Fig. 4(c).

Fig. 4
figure 4

Creep displacement (primary y-axis) and creep strain rate (secondary y-axis) versus holding time curves for Zr-Cu-Al and Zr-Cu-Al-Ti alloys at the loading rate of (a) 10 mN/s and (b) 100 mN/s; higher free volume in rejuvenated alloys led to larger creep displacement at all loading rates; (c) creep strain rate sensitivity of the four alloys at the two different loading rates.

Full size image

Discussion

Linking the shear transformation zone volume to deformation behavior

Figure 5 shows the measured yield strength (YS) and shear transformation zone (STZ) volume calculated based on the Johnson-Samwer cooperative shear model (CSM) as a function of Poisson’s ratio70. According to CSM, the activation energy is given as:

$$W^{*} = 4R_{0} G_{0} \gamma_{c}^{2} \left( {1 - \frac{\tau }{{\tau_{c} }}} \right)^{3/2} \Omega \xi$$
(5)

where Ω is the volume of STZ, \(G_{0}\) and \(\tau_{c}\) are shear modulus and threshold shear resistance of the metallic glass at 0 K, respectively, R0 and ξ are constants with approximate values of 0.5 and 3, respectively71 and, \(\gamma_{c}\) is the elastic limit with a value of 0.027 for BMGs71,72. Therefore, STZ volume may be calculated in terms of hardness and strain-rate sensitivity (m) as71:

$$\Omega = \left( {\frac{kT\sqrt 3 }{{2G_{0} \gamma_{c}^{2} }}} \right)\frac{1}{{R_{0} \xi }}\frac{1}{{m \left( {\frac{H}{{\tau_{c} }}} \right)(1 - \frac{{\tau_{CT} }}{{\tau_{c} }}) ^{\frac{1}{2}} }}$$
(6)

where \(\tau_{CT}\) is threshold shear resistance at temperature T, \(\tau_{c}\) /\(G_{0}\)≈ 0.036. Furthermore,

$$\frac{{\tau_{CT} }}{{\tau_{c} }} = 1 - \left( {0.44} \right)(T/Tg) ^{\frac{2}{3}}$$
(7)
Fig. 5
figure 5

Yield strength (primary y-axis, in blue) and shear transformation zone (STZ) volume (secondary y-axis, in green) as a function of Poisson’s ratio for the studied alloys. The data points represent the as-cast (AC) and rejuvenated (Rej) states for Zr-Cu-Al and Zr-Cu-Al-Ti alloys. The figure illustrates that the yield strength decreases while the STZ volume increases with the increase in Poisson's ratio among the studied alloys.

Full size image

The STZ volume for Zr-Cu-Al BMG increased from ~ 0.6 nm3 in as-cast state to ~ 0.8 nm3 in thermally rejuvenated state. Similarly, the STZ volume for Zr-Cu-Al-Ti BMG increased from ~ 0.9 nm3 in as-cast state to ~ 1.1 nm3 after thermal rejuvenation. The rejuvenated alloys showed larger STZ volume compared to their as-cast counterparts and the STZ volume increased with the addition of 1 at. % Ti to Zr-Cu-Al BMG. Larger STZ volume correlates with enhanced plasticity, as fewer STZs are needed to reach the critical shear strain for shear band formation. This reduction in the number of necessary STZs for shear band nucleation facilitates higher internal stress concentration, activating new flow units and multiple shear bands, thereby enhancing plasticity.

The yield strength decreased while the STZ volume increased with an increase in Poisson’s ratio among the studied alloys. This behavior may be attributed to the large negative heat of mixing among the constituent elements in a bulk metallic glass (BMG), which promotes chemical short-range ordering54,73. For example, the heat of mixing values are 0 kJ/mol for Ti-Zr, − 9 kJ/mol for Ti-Cu, and − 30 kJ/mol for Ti–Al, whereas they are − 23 kJ/mol for Zr-Cu and − 44 kJ/mol for Zr-Al74. These values suggest that Ti atoms preferentially substitute for Zr sites, as shown in recent studies74. Furthermore, the significant atomic size difference between Ti (0.145 nm) and Zr (0.160 nm) can lead to atomic-scale distortions and loosely packed local configurations. This atomic-scale distortion may explain the observed softening and reduction in stiffness with the addition of 1 at. % Ti to the Zr-Cu-Al BMG, as seen in Table 2.

Local atomic ordering

Molecular dynamics (MD) simulations were carried out to explain the pronounced differences in the deformation behavior of Zr-Cu-Al and Zr-Cu-Al-Ti BMGs in their as-cast and rejuvenated states. Figure 6 shows the fraction change in Voronoi index, calculated using spatial tessellation techniques, for the MD-simulated BMGs of composition identical to those used in the experiments75. The Voronoi index distinguishes coordination polyhedral clusters or short-range ordering (SRO) in the BMGs76.

Fig. 6
figure 6

Fraction change of Voronoi polyhedral clusters for Zr-Cu-Al and Zr-Cu-Al-Ti BMGs after cryogenic thermal cycling compared to their as-cast state.

Full size image

Figure 6 illustrates the distribution of major Voronoi polyhedra with Al, Ti, and Cu as center atoms. Previous reports classified the polyhedral clusters into dense and loose packing types77,78. Dense packing clusters are void-free, while loose packing clusters contain voids. After cryogenic thermal cycling, the Al-centered Voronoi index of < 0,0,12,0 > decreased by 29% and 27.5% in Zr-Cu-Al and Zr-Cu-Al-Ti BMGs, respectively. The dominant coordination polyhedra, < 0,2,8,1 > and < 0,2,8,2 > , correspond to bicapped square antiprism (BSAP) or distorted BSAP55. In Zr-Cu-Al BMG, the fraction of < 0,2,8,1 > polyhedra decreased by 4.2% and < 0,2,8,2 > by 3.1% after thermal rejuvenation. Similarly, in Zr-Cu-Al-Ti BMG, < 0,2,8,1 > decreased by 3.3% and < 0,2,8,2 > by 3.7%. Conversely, the fraction of loose-packed atomic clusters in both alloys mostly increased after cryogenic thermal cycling. This overall decrease in dense packing clusters and increase in loose packing clusters likely contributed to an increase in atomic-scale voids or free volume, thus enhancing plasticity79,80. The decrease in the number of dominant polyhedra increases the disordered regions where shear transformation nucleation is more likely to occur. Thus, the experimental differences in deformation behavior may be attributed to the degree of structural heterogeneity due to specific polyhedra in different structural states of the BMGs.

Recent MD simulations also show that rejuvenated Zr-Cu-Al glasses have a heterogeneous atomic cluster distribution, marked by a reduced presence of icosahedral (icos) and defective icosahedral clusters (deficos) and an increased prevalence of body-centered cubic (bcc) and face-centered cubic (fcc) + hexagonal close-packed (hcp) clusters81. This atomic restructuring leads to enhanced plasticity in metallic glasses, attributed to the higher nucleation sites for facilitating shear flow events. The addition of 1% Ti in Zr-Cu-Al BMG increases microstructural heterogeneity due to increased diversity of atomic species and their different sizes. Zr-Cu-Al-Ti BMG showed a larger STZ volume of 0.9 nm3 leading to enhanced plasticity compared to 0.6 nm3 for Zr-Cu-Al BMG. Thermal rejuvenation further improved these properties due to atomic-level structural changes.

Conclusions

In summary, the effect of thermal rejuvenation (i.e., with cryogenic cycling) on the plastic deformation behavior of Zr-Cu-Al and Zr-Cu-Al-Ti bulk metallic glasses was studied. The relaxation enthalpy increased while the hardness and strength decreased significantly with the addition of 1 at% Ti to Zr-Cu-Al BMG and further with thermal rejuvenation. The magnitude of stress drops during uniaxial compression decreased while their frequency increased after micro-alloying with 1 at. % Ti and further due to thermal rejuvenation, indicating the accommodation of strain in larger number of shear bands. The enhanced plasticity after micro-alloying with Ti and thermal rejuvenation was explained by higher number of sites for shear flow initiation due to increased heterogeneity in the system along with the decrease in dense packing clusters and increase in loosely packed short range order clusters. Time dependent deformation was relatively more homogenous and creep SRS was higher at higher loading rate (100 mN/s) compared to the response seen at lower loading rate (10 mN/s). This was attributed to the generation of excess free volume during loading at higher rates, thereby making the alloy less sensitive to the initial free volume.