Ductile crystalline-amorphous nanolaminates

Abstract

It is known that the room-temperature plastic deformation of bulk metallic glasses is compromised by strain softening and shear localization, resulting in near-zero tensile ductility. The incorporation of metallic glasses into engineering materials, therefore, is often accompanied by complete brittleness or an apparent loss of useful tensile ductility. Here we report the observation of an exceptional tensile ductility in crystalline copper/copper-zirconium glass nanolaminates. These nanocrystalline-amorphous nanolaminates exhibit a high flow stress of 1.09 +/- 0.02 GPa, a nearly elastic-perfectly plastic behavior without necking, and a tensile elongation to failure of 13.8 +/- 1.7%, which is six to eight times higher than that typically observed in conventional crystalline-crystalline nanolaminates (<2%) and most other nanocrystalline materials. Transmission electron microscopy and atomistic simulations demonstrate that shear banding instability no longer afflicts the 5- to 10-nm-thick nanolaminate glassy layers during tensile deformation, which also act as high-capacity sinks for dislocations, enabling absorption of free volume and free energy transported by the dislocations; the amorphous-crystal interfaces exhibit unique inelastic shear (slip) transfer characteristics, fundamentally different from those of grain boundaries. Nanoscale metallic glass layers therefore may offer great benefits in engineering the plasticity of crystalline materials and opening new avenues for improving their strength and ductility.

Fig. 1.

Microstructures and tensile properties of Cu/Zr nanolaminates. (A and B) Cross-sectional (A) and plan-view (B) TEM images of the as-deposited 5/35 nanocrystalline Cu and amorphous Cu/Zr intermixing multilayer nanostructures. The average grain size in the nanocrystalline layers is approximately equal to the individual layer thickness. (C) Room-temperature tensile true stress–strain curves of the nanocrystalline–amorphous nanolaminate at the strain rate of 1 ×10⁻⁴ s⁻¹, in comparison with those of Cu/304 stainless steel crystalline multilayer with an individual layer thickness of 25 nm and pure nanocrystalline Cu with an average grain size of ≈30 nm. The curve for pure nanocrystalline Cu is an engineering stress–strain plot taken from ref. . The Cu/Zr nanolaminate has an average tensile elongation to failure of 13.8 ± 1.7% and a steady-state flow stress of 1,090 ± 20 MPa, in contrast with the low ductility (<2%) seen in conventional crystalline nanolaminates (Cu/304 SS) and pure nanocrystalline Cu. In addition, the nanoscale metallic-glass modulated nanolaminates exhibit a near-perfect plastic flow behavior without necking. (D) The top-view of the gauge section after fracture for 5/35 Cu/Zu nanolaminate.

Fig. 2.

Deformation microstructures. (A–D) Sequential TEM images of the 5/35 Cu/Zr nanolaminate at tensile strains of 0% (A), 2% (B), 7% (C), and 10% (D). The gradual reductions of individual nanocrystalline and nanoscale amorphous layers can be seen with increasing strains. Some deformation twins are discernable, but dislocation pileup is not observed at any strains. (E) A high-resolution TEM image of 5/35 Cu/Zr nanolaminate after fracture. Deformation twins can be seen inside several nanograins, two ends of which tend to terminate at the ACIs (green arrows) or at the GBs (red arrow), suggesting that the ACIs may have become the dislocation nucleation sources. Stacking faults are also seen at high-resolution TEM. (Scale bars: A–D, 20 nm; E, 50 nm.)

Fig. 3.

MD simulations of 5/10 Cu/Zr system (10 × 15 × 10 nm, 105,336 atoms) under periodic boundary conditions. (A) Dislocation nucleates at the bottom ACI. The central symmetry parameter, which characterizes the degree of inversion symmetry breaking around each atom, is used to visualize the stacking fault, bounded by a Shockley partial dislocation. (B) Dislocation is absorbed by the amorphous layer at the top ACI. Simultaneously, another stacking fault system is activated and intersects the first stacking fault in the middle, forming a sessile dislocation junction. (C) Activation of STZs when the partial dislocation hits the amorphous layer. Atoms with inelastic strain below the strain threshold are not shown. STZs in the amorphous layers are clearly visible as clusters of inelastically transformed atoms. One also sees STZs forming spontaneously inside the amorphous layer at the bottom of the image. (D) Distribution of inelastic shear strain after dislocation absorptions, in reference to a configuration before dislocation activities. ACI sliding (similar to GB sliding) shows up on the lower right and upper left corners of the bottom ACI.

Fig. 4.

Influence on dislocation structures by the presence of the amorphous phase, shown by T = 300 K MD simulations in the 5/35 Cu/Zr system (37 × 40 × 8 nm, 790,894 atoms). Atoms are color-coded by their coordination numbers [red, 11; blue, 13; green, 10; tan, 14; perfectly coordinated atoms (12) are not shown]. (A) A great burst of dislocation activities is induced initially by tension at the onset of yielding in the simulations, resulting in a dense sessile dislocation forest. (B) As simulation time progresses, the dislocation density decreases dramatically (after 11% additional strain with respect to A), indicating that the dislocation structures are drawn into the amorphous layers.