Joining aluminum alloy sheets is increasingly important in manufacturing. Traditional welding techniques create a heat-affected zone (HAZ) around the joint; however, solid-state joining methods such as impact welding produce joints without significant heat. Here, electrically vaporized foil actuators (VFA) provided the high-pressure pulses needed for impact welding. 0.96 mm thick AA6061-T6 and 0.76 mm thick AA5052 were joined in lap and spotlike configurations, at a variety of impact velocities. The welds failed in coach-peel outside the joint interface. The 5052 hardened within 100 μm of the interface. The 6061-T6 may have softened slightly within 50 μm of the interface.
Aluminum alloys combine low density and corrosion resistance with good mechanical properties after metallurgical strengthening mechanisms are applied. The high strength-to-density ratio of aluminum alloys makes them popular in industries such as automotive and aerospace manufacturing [1,2]. Body panels and structural components, such as decklids and roof rails, are increasingly being designed out of aluminum instead of steel . For that reason, the production of strong joints between aluminum components has become increasingly important.
Traditional fusion welding (via shielded metal arc welding, gas metal arc welding, or gas tungsten arc welding) and resistance spot welding produce a HAZ around the weld area . Recent advanced techniques, such as cold metal transfer joining, reduce heat input, but do not eliminate the HAZ . Some degree of HAZ is even observed in friction stir welding, a solid-state welding technique . In the HAZ, which can span a few millimeters [2,6] or even centimeters [4,7] from the weld perimeter, the metal temperature is substantially increased during the welding process. In age-hardened Al alloys like 6061-T6, heat from welding can lead to overaging of the alloy, causing a loss of material strength . In work-hardened alloys like 5052, strength is reduced by annealing of the material. As a result, materials become more susceptible to failure in the HAZ around the weld.
Impact welding, which produces a solid-state joint with minimal melting at the interface, is a good candidate for producing HAZ-free joints . In this technique, which has been practiced with explosives since the 1950s, one metal sheet (called a flyer) is driven by a high-pressure pulse to a high velocity, in order to impact another sheet (called a target) at an angle. At the point of impact, hydrodynamic behavior of the metals' surfaces results in the ejection of their surfaces as a jet. The jetted material contains surface oxides and other impurities and leaves behind clean metallic surfaces in high-pressure contact with each other . This leads to the formation of a metallurgical bond between the two parent sheets as the gap between them collapses and the impact point progresses to the edge of the sheets. The weld interface morphology can range from flat to varying degrees of waviness, and may or may not contain pores or a layer of intermetallic compounds, all depending on the material properties and the impact angle and velocity [10–12]. There thus exists a “welding window” of parameters that produce a strong weld for a given material combination [13,14]. The mechanism of formation of the wavy morphology is debated [15,16]; however, the purpose of this paper is not to critique or validate any particular theory.
Explosive welding (EXW) is useful for joining centimeter-thick flyers to targets as thick or thicker; however, this method performs poorly at smaller scales . Magnetic pulse welding (MPW)—also known as electromagnetic (EM) impact welding—has proven useful for aluminum flyers up to 1 mm thick with targets of varied thicknesses . MPW joints can exceed the base metal strength, as shown by Shribman et al. with AA7075-T6  and Okagawa and Aizawa with AA1050-O . Kore et al. also EM welded AA1050-O to itself  as well as to heat treated Al–Li alloy 1441 . The flyer velocity can be controlled and predicted based on the discharge energy for a given coil and the distance traveled by the flyer [23,24]. Unfortunately, the life of EM actuators is strongly limited by the magnetic pressure and cycle frequency that they experience [25,26]. Therefore, EM actuators that can drive a flyer to impact welding velocity (>400 m/s) tend not to last long.
Researchers at The Ohio State University (OSU) have recently developed vaporizing foil actuator welding (VFAW), a technique useful for impact welding of millimeter-thick sheets . In this technique, a capacitor bank charged to a high voltage is discharged through a thin aluminum foil, referred to as a foil actuator. The resulting high-current pulse (on the order of 100–150 kA) deposits more energy into the foil than is necessary to sublimate it, causing it to rapidly vaporize and create a strong pressure pulse on its surroundings, analogously to the explosive in EXW. Because the vaporization depends on the current density in the foil , the foil actuator shape can be tailored to initiate the pressure pulse in a desired region under a workpiece .
Materials and Methods
The flyer sheets (those driven by the pressure pulse) were 76.2 × 101.6 mm sheets cut from 0.96 mm thick AA6061-T6. The target sheets (those impacted by the flyers) were of the same dimensions, cut from 0.76 mm thick AA5052. The aluminum foil actuators were cut from 0.0762 mm thick AA11xx foil into the two shapes shown in Fig. 1(a).
“Patch”- and “spot”-weld VFAW setups are illustrated schematically in Figs. 1(c) and 1(d). The foil actuator was insulated inside a heat-sealed polyester sleeve and placed on a steel backing block. In the case of the spot-weld setup, a cylindrical riser 15 mm in diameter and 13 mm high was centered under the narrow region of the foil. This was in order to keep the pressure impulse on the flyer sheet localized, to avoid deformation of the surrounding area of the flyer by providing increased free volume outside the weld area in which the pressure could dissipate. The flyer sheet was centered on top of the narrow part of the foil actuator. The face of the flyer sheet and of the backing block that faced the foil were insulated with polyimide tape to prevent the current through the foil actuator from arcing to one or the other once the expanding aluminum vapor breached its insulation sleeve.
Standoff sheets were placed on the flyer plate, and the target plate was set on top of them. This standoff distance was necessary in order to allow the flyer to accelerate, and to create an angled impact with the target (Fig. 1(b)). In the patch-weld setup, the standoff distance was created with 1.6 mm thick Garolite sheets spaced 31.75 mm apart from each other (Fig. 1(c)). In the spot-weld setup, the standoffs were 1.58 mm thick Viton washers with a 19 mm inner diameter, centered over the narrow region of the foil. Finally, another steel backing block was placed on top of the assembly and bolted firmly to the bottom backing block in order to provide inertial restraint for the target sheet.
A Maxwell-Magneform capacitor bank with a short-circuit current rise time of about 12 μs was used as the charge source in these experiments. Its specifications are listed in Table 1. Repeated experiments were performed with bank charging energies of 4, 6, and 8 kJ, excepting the spot welds, which were only performed at 8 kJ. The voltage–time and current–time histories from the circuit were measured using a 1000:1 voltage divider and a 50 kA:1 V Rogowski coil, respectively. The velocity–time history of a moving flyer sheet could be recorded simultaneously via the photonic Doppler velocimetry (PDV) method developed by Strand et al.  and adapted at OSU . In experiments where velocimetry was desired, the PDV probe was oriented normal to the surface of the moving flyer, observing the flyer through a 5 mm diameter vertical channel drilled through the center of the top backing block and target sheet.
A sample at each energy level was cross-sectioned orthogonal to the plane of the weld interface. The cross sections were mounted and polished for hardness testing and optical and scanning electron microscopy (SEM). Hardness traverses across the weld interface, similar to Ref. , were recorded in five to seven places along each cross section using a Leco microindentation testing system at a load of 100 gf and a dwell time of 16 s. SEM images were taken with an FEI Quanta 200.
Coach-peel type tests were conducted on a MTS 370 load frame, using a 100 kN capacity load cell. The weld samples were bent into an “H” shape and fixtured as illustrated in Fig. 2, then pulled at 1 mm/min until failure occurred. Lap-shear testing was not conducted, because previous work showed that VFA impact welds which failed in the parent material during lap-shear testing could still fail in the weld interface during peel testing . The peel test is therefore considered to be the more rigorous test.
Results and Discussion
The evolution of a flyer sheet's velocity over the distance it traveled was calculated by integrating the velocity–time histories produced by PDV. Flyer velocity versus travel traces for each of the parameter variations are overlaid in Fig. 3. From these, it can be seen that the velocity of the flyer upon impact with the target increased successively as the charging energy of the capacitor bank was increased. This is commonly understood to be due to the increase in current density in the foil at vaporization (known as burst current density), which results from the higher input energy into the circuit at higher charging energies .
Though the difference in the impact velocity between each charging energy level may appear similar for the patch-weld samples, it is not indicative of a linear trend between input energy and velocity. Previous work by Vivek et al. systematically varied VFA input parameters . That work showed that the both measured flyer velocity and the energy deposited into a given mass of foil (which are related by the square of the current) approach a maximum limit as the charging energy increases for a particular resistive, inductive, capacitive (RLC) system.
The impact velocity for the spot weld was ∼960 m/s, while the impact velocity for the 8 kJ patch weld was ∼730 m/s. The large difference in velocities, when the charging energy was the same, is mainly due to the difference in geometry between the two foil actuator designs. While the energy deposited in the spot welding foil before it vaporized was somewhat less than in the patch welding actuator (2.83 kJ versus 3.50 kJ), the volume of aluminum being primarily vaporized was around a quarter of that in the patch actuators. Therefore, the burst current density was greater for the spot actuator, and the resulting pressure pulse against the flyer sheet was stronger. The spot welding setup may also have contributed to more efficient use of the pressure pulse; because the standoff was a ring, rather than having open ends, the lateral vapor confinement was likely better in that setup than in the patch welding setup.
Every trial produced a successful weld, in spite of the ∼500 m/s spread of impact velocities among the different parameter variations. As a result, it can be presumed that this range of velocities falls more or less within the welding window for this material pair. Representative samples are displayed in Fig. 4.
The minimum peel strengths of the patch-weld samples are presented in Table 2. These are specifically noted as minimum peel strengths because in every case, the weld samples failed in the weaker parent material, 5052, rather than across the weld interface. It is likely that, if the true peel strengths of the welds were measured, a trend dependent on impact velocity would be observed, depending on where the observed range of velocities lies in the welding window for the material pair.
Notable variation is observed in the strengths shown in Table 2, even among welds created using the same input energy. No definite trend based on input energy (and, presumably, impact velocity) can be deduced from this data. One potential reason for the inconsistency may be the variability of this method of testing. When the samples are bent into the H shape, the bend is made near the interface by hand, but no fixture has been made that would guarantee that the bend happens the same way every time. As a result, the lever arm between the weld perimeter and the bend varies in length, varying the effective tensile load on the perimeter of the weld.
More fundamentally, unlike the standard application of the peel test in adhesives, the bending of the flyer and target sheets makes failure more likely to occur at the strained bent region. Because of this, the 5052 failed in tension outside the weld at around 100 MPa, far below its typical 228 MPa ultimate tensile strength. Figures 5(a) and 5(b) show an 8 kJ patch-type weld sample before and after peel testing. Peel testing results could perhaps be improved by creating a smaller weld region relative to the length of sheet bent to 90 deg.
The spot-weld samples failed more consistently (Table 3). Each sample failed by shearing in the 5052 sheet around the perimeter of the weld, leaving behind a 5052 weld nugget on the 6061-T6 sheet. Figures 5(c) and 5(d) show an 8 kJ spot-type weld sample before and after peel testing. This failure resembles traditional weld nugget pullout, the desired mode of failure . As such, it is more comparable to known weld testing methods: the strength of resistance spot welds can be determined by a similar method, known as cross tension . Though that method more evenly distributes stress around the perimeter of the weld (likely improving strength measurements), the result is a similar weld nugget shear pullout failure. It would be ideal to modify the testing procedure for the VFA welds to match this cross tension method.
The results of microhardness testing are in Fig. 6. In Fig. 6(a), a representative view of the hardness testing pattern is visible, with multiple traverses made across the weld interface of each sample. The compiled results for each parameter variation are shown in Figs. 6(b)–6(e). The hardness profiles are remarkably consistent among them; in all cases, it can be seen that the base material hardness extends very nearly to the interface.
Some work hardening due to the impact is observed in the 5052 within 100 μm of the interface. A small amount of softening may also have occurred in the 6061-T6 within 50 μm of the interface, possibly due to annealing. However, this apparent trend may be due to the irregularity and invisibility of the wavy interface, the exact location of which was difficult to visually determine. The lower hardness recorded near the interface on the 6061-T6 side may in fact be due to overlap of the indenter with regions of 5052. If, however, softening truly occurred in the 6061-T6, its small extension from the interface indicates that any heat produced due to the impact pressure was very localized, even at higher impact velocities.
Most significantly, nowhere does the hardness fall below the base hardness of the weaker material. That little or no HAZ is present in these VFA welds is significant—as mentioned earlier, the HAZ of aluminum alloys welded by traditional means can range from millimeters to centimeters. As a result of the minimized HAZ, the strength around the joint perimeter is maximized.
Further examination of Al–Al VFA weld interfaces is warranted, starting with nanoindentation to measure the hardness across the interface in greater detail. Additionally, TEM analysis could reveal grain size and dislocation density in the materials near the interface to help better understand the slight changes in hardness near the weld interface.
The trials performed in this work illustrate how VFAW, like other welding techniques, can be used to produce a diverse range of weld lengths. Depending on the parts desired to be joined, welds with a size of a few square millimeters or centimeters can be produced in one split-second event, simply by varying the geometry of the foil actuator.
In this work, it was demonstrated that VFAW can be used to reproducibly create solid-state welds of varying lengths. Patch welds, resembling conventional lap welds, and spot welds similar to traditional spot welds were produced. Impact velocities ranged from 500 to 960 m/s and a successful weld was created in each case. Upon mechanical testing, the resulting Al–Al welds universally failed in the weaker parent material 5052, instead of in the joint interface.
Hardness testing revealed some process-induced changes in mechanical properties within 100 μm on either side of the weld interface. Some hardening occurred on the 5052 side, while the 6061 side may have experienced some decrease in hardness. The limited or nonexistent HAZs are a vast improvement over fusion welding, resistance spot welding, and even friction stir welding, in which HAZs may extend millimeters from the interface at a minimum.
Future work related to better understanding the structure–property relationships, improved mechanical testing methods, and process automation and scaling should be conducted.
Thanks are due to research group members: Bert C. Liu, Alexander Koenig, Drew Demmerle, and Geoffrey A. Taber for assistance with experimental trials, sample testing, and data organization.
This work was funded by the U.S. Department of Energy (DOE) Award No. DE-PI0000012, as well as support from the ALCOA Foundation.
- EM =
- EXW =
- gf =
gram-force, a nonstandard gravitational metric unit commonly used for hardness testing loads
- HAZ =
- HV =
Vickers pyramid number, a unit of hardness
- MPW =
magnetic pulse welding
- PDV =
photonic Doppler velocimetry
- RLC =
resistive, inductive, capacitive, in reference to an electrical circuit
- VFA =
vaporizing foil actuator(s)
- VFAW =
vaporizing foil actuator welding