Fig. 1. Illustration of the vaporizing foil actuator spot welding impacting process. (a) the initial setup (b) right after the vaporization of the Al foil. (c) after the entire flyer impacts the target. The two highlighted regions are locations in which the two base materials are expected to bond.

Fig. 2. SEM micrograph showing the AZ31/DP590 impact weld by vaporizing foil actuator welding (VFAW) method. Cracks appeared in the Mg side very close to the interface. Some epoxy went into the crack as an artifact result of the polishing. Two regions are further examined. High magnification SEM-EDX is done in region A (Fig. 3); STEM-EDX and HRTEM are done in region B (Figs. 4-6).

Fig. 3. Higher magnification SEM examination of the region A in bonding zone (a, b) SEM image of the DP590-interlayer-AZ31 microstructure of the weld sample, from different sectioned quarters of the same sample. (c) SEM-EDX image correspond to (b) showing the Mg as the bright field (d) SEM-EDX image showing the Fe as the bright field.

Fig. 4. Higher magnification SEM examination of region B in Fig. 2.

Fig. 5. STEM-EDXS image of chemical components of the interlayer. The images of Mn and Al are normalized to show the gradient. A very small amount of Zn element is seen and is therefore not shown.

Fig. 6. HRTEM images at the bonding zone (marked region B of Fig. 2) of the AZ31-DP590 impact weld sample (a) at the interface between the interlayer and DP590, and (b) inside the interlayer.

Fig. 7. The STEM-EDXS scanning result along a line across the interface between the interlayer and DP590. The size of the scanning electron beam is around 0.3?nm, significantly smaller than the interlayer width. Hence, no deconvolution of the spot size was necessary.

Fig. 8. Model setup for the coupled thermal-mechanical Eulerian-based FE simulation.

Fig. 9. Eulerian-FE simulation results of the AZ31-DP590 impact welding process (a) at t=1.8μs, impact first occurs in the center of the target due to the distributed velocity. (b) oblique impact propagates laterally to the corner, and the jetting of AZ31 occurs from the collision at t=2.2μs (c) oblique impact continues to propagate with further jetting, and unloading starts from the center (d) at t=2.75μs, the material jet flow hit the corner of DP590 target and cause local deformation, the oblique collision start to slow down. (e) the AZ31 collide down on the tilted edge of DP590 target (f) the entire sample finished unloading.

Fig. 10. Details of Eulerian-FE simulation results at collision front at time t=2.4μs. (a) the contour of the velocity component V1 in the transverse direction, unit is m?s-1. The arrow marked the area that collision occurs (b) the contour of the velocity component V2 in the vertical direction. unit is m?s-1. (c) The temperature distribution, unit is kelvin (d) the equivalent element volume-averaged plastic strain.

Fig. 11. Measurement of (a) velocity component V1 in transverse direction (b) velocity component V2 in the vertical direction and (c) temperature at the moving collision front during the lateral propagation of oblique impact.

Fig. 12. Model setup for the MD simulation.

Fig. 13. MD simulation prediction of impact interface nanostructure evolution. The Mg flyer has a normal impact velocity of Vnormal=350m?s-1, and a transverse velocity Vtransverse=2850m?s-1. (a) 10 ps right after impact: “mixing” of atoms is observed at the right corner on the interface. (b) after holding for 0.4ns, (c) at the end of 5?ns holding, and (d) after reloading.

Fig. 14. MD simulation result of interfacial nanostructure evolution in the impact sample during cooling, after cooling from 9200?K to (a) 2500?K (b) 1800?K, (c) 1700?K, (d) 1000?K, and (e) 300?K.

Fig. 15. HRTEM image close to the interface of impact weld sample.

Fig. 16. MD prediction of the impact weld interfacial nanostructure after cooling from 9200?K to 300?K in (a) 0.2?ns (b) 0.5 nanosecond, and (c) 1.0 nanosecond. The cooling rate is constant within the cooling period. Fig. 16(c) is the same as Fig. 14(e).

Fig. 17. (a, d) setup of Eulerian-FE simulation with two different initial flyer velocity distribution. (b, e) the corresponding snapshot during the lateral propagation of the oblique impact, (c, f) simulation results at the end of the unloading stage.

Fig. 18. Eulerian-FE simulation result of impact process using the same setup as in Fig. 8 but with material properties of AZ31 replaced by titanium and DP590 replaced by copper. Mechanical interlocking is observed.