Ultrasonic Spot Welding of Lightweight Alloys

The structural applications of lightweight alloys in the automotive and aerospace industries inevitably involve welding and joining of challenging dissimilar Mg-to-Al and Mg-to-steel while guaranteeing safety and structural integrity. Sound dissimilar lap joints were achieved via ultrasonic spot welding (USW) – an environment-friendly solid-state joining technique. The addition of Sn interlayer during USW effectively blocked the formation of brittle Al12Mg17 intermetallic compound in the Mg-to-Al dissimilar joints without interlayer, and led to the presence of a distinctive composite-like Sn and Mg2Sn eutectic structure in both Mg-to-Al and Mg-to-HSLA (high strength low alloy) steel joints. The lap shear strength of both types of dissimilar joints with a Sn interlayer was significantly higher than that of the corresponding dissimilar joints without interlayer. Failure during the tensile lap shear tests occurred mainly in the mode of cohesive failure in the Mg-to-Al dissimilar joints and in the mode of partial nugget pull-out in the Mg-to-HSLA steel dissimilar joints. In particular, the addition of Sn interlayer resulted in energy saving since the welding energy required to achieve the maximum strength decreased from 1250 J to 1000 J in the Mg-to-Al joints and from 1750 J to 1500 J in the Mg-to-HSLA steel joints.


INTRODUCTION AND MOTIVATION
Environmental pollution may be attributed to a large extent to the transportation industry, predominantly CO 2 emissions produced by automotive vehicles. These can have devastating effects on human society and the environment [1][2][3][4][5][6][7][8][9]. This problem can be resolved by considering such as use of alternate fuel sources, aerodynamic improvements, and powertrain enhancements. However, lightweighting of auto-body seems to be the best cost effective option for significantly decreasing CO 2 emissions and improving fuel efficiency [10]. It is projected a 22.5 kg reduction in mass would improve fuel efficiency by approximately 1% [1]. Thus, currently, the application and development of the ultralightweight magnesium (Mg) alloys have been considerably growing in the transportation industry due to its low density, superior damping capacity and high strength-to-weight ratio [1,[11][12][13][14][15][16][17]. Similarly, because of the lightweight and good mechanical strength properties of aluminum (Al) alloys, there usage is it is becoming progressively substantial in high-speed passenger rail bodies and heavyweight vehicles [18,19]. In addition, currently, steel is the most common metal used for the automotive vehicle industries. In order to achieve the desired properties of these multi-materials, welding and joining of these metals is essential.
This dissertation represents a contribution to novel ultrasonic spot welding (USW) technique to join similar and dissimilar lightweight alloys. Specifically, USWed Mg alloys' deformation mechanisms, dynamic recrystallization, its influence on crystallographic texture formation, and mechanical properties such as tensile and fatigue properties are studied.
Furthermore, characterization of challenging dissimilar USW of Mg-to-Al, Mg-to-steel and Al-to-steel, respectively were performed, where the formation of intermetallic compounds (IMCs) during dissimilar USW has been systematically studied. To improve the mechanical properties like lap shear tensile and fatigue, IMCs of dissimilar USWed joints were engineered with the help of a Sn interlayer and Zn coating. The remainder of this Chapter establishes the motivation and framework for welding and particular USW process.

Motivation for welding
Unfortunately, to this point, research has not yet reached to print automobile vehicles by 3D-Printing. Thus, one has to rely on the joining technologies, which are inevitably involved in the structural application of the transportation industry. Amongst all joining technologies, welding is the most cost effective, efficacious, fast and reliable method to permanently join materials. According to the document, Vision of Welding, by American Society of Welding [20], welding is a critical technique and typically the final step in the assembly, and plays a major role in ensuring structural performance. So far, no other technique is as widely used by manufacturers for joining metals. By considering only spot-weld industry, more than 5000 spot welds are required to manufacture a single car. Thus, the development of reliable spot weld method is essentially required. In spot welding industries, there are three major welding technologies available in the current market, which are resistance spot welding (RSW), friction stir spot welding (FSSW) and USW. Among them USW has several wide range of advantages compared to that of RSW and FSSW (discussed later in section 2.1.7).

Motivation for ultrasonic spot welding
Conventional fusion welding techniques produce large grains, porosity, voids and other defects in the weld zone that seriously degrades mechanical properties of joints. In the auto body manufacturing for the steel car bodies, RSW has been a predominant process [21,22].
Unfortunately, RSW is challenging and costly to apply to Mg and other alloys such as Al alloys owing to their tendency to degrade the electrodes, low strength at high temperature, and high conductivity [23]. In RSW, the high-energy requirement (50-100 kJ per weld) [24,25] and formation of large heat affected zone (HAZ) are main concerned. Recently, attention has been paid to two solid-state welding processes, namely FSSW and USW, because the liquid phase reaction in the fusion zone during RSW can be avoided. Although FSSW has the potential to produce effective welds between similar and dissimilar materials, a relatively long welding cycle (or time) would be a limiting factor for its widespread adoption in automotive manufacturing [22]. A different solid-state spot welding process that has not attracted much attention is ultrasonic spot welding (USW) [24][25][26][27] in the welding of dissimilar metals a rapid formation of brittle IMCs occurs, which can highly degrade the mechanical properties of welded joints [36]. Some limited work on the USW of dissimilar Mg-to-steel and Al-to-steel has recently been reported [37,38]. However, ways of enhancing the USW joint strength have not been systematically studied.
In this regard, microstructure and mechanical properties such as tensile and fatigue, of all welded joints used in structural applications must be evaluated to ensure the integrity and safety of the joints and structures. However, until now, the evaluation of differently welded joints in Mg and Al alloys using novel welding techniques in terms of mechanical properties is very limited.
The overall objective of this work is, therefore, to gain a better understanding of the principal factors of USW of similar and dissimilar lightweight alloys. The specific objectives include the following: • To evaluate the influence of USW in a Mg alloy on microstructure, crystallographic texture, lap shear tensile strength, fatigue resistance, and failure mode.
• To study the microstructure and mechanical properties of dissimilar USWed Mg-to-Al joints and evaluate the influence of introducing an interlayer on the mechanical properties.
• To characterize USWed joints of Mg-to-galvanized and ungalvanized steel with and without a interlayer.
• To study microstructure, tensile and fatigue properties of USWed joints of Al-to-steel alloy.

LITERATURE REVIEW
In this section, a literature review is presented to facilitate a concise overview of the USW of lightweight alloys, and to identify areas where further research is needed. The focus of this Chapter is to provide a review of the published research relevant to the topic of USW of lightweight alloys. This review is divided into two sections: the principles and fundamentals of USW system and prior work in the area of welding of Mg, Al and steel.

Principles and fundamentals of USW
While USW is extensively used in applications ranging from plastic industries, tube sealing, welding microscopic connections on microchips to, including electrical interconnections, the focus of this work is limited to the sheet metal welding, where the welding mechanism is thought to be different from that for thins foils, wires and plastics. In the following subsections, a brief outline of the USW process is presented.

History of USW
Some 60 years ago ultrasound was used to improve grain refinement in fusion welding and brazing [39]. Ultrasonic welding process was discovered in the late 1940s when, in research at the Aeroprojects Company of West Chester, Pennsylvania (the forerunner of the current Sonobond Corporation), ultrasonic vibrations were applied to conventional resistance welding equipment with the objective of decreasing surface resistance in spot welding of aluminium [39]. During this work, it was found that ultrasound alone was capable of producing a bonding of the metals. Similar work on welding by mechanical vibration was reported by Willrich in 1950, who obtained cold weld in the region of resistance welding by applying low frequency vibration to the tool that was designed for this purpose [40].
Ultrasonic metal welding equipment was first patented in 1960 [25].

USW equipment
In USW, there are numerous variations of the process available, such as those based on geometry, type of the weld and the metals to be joined. Among several different variations are, for example ring, torsion, seam, and line [41]. This study is focused on USW for sheet metals, which is most common to build automotive vehicle structure. This kind of weld typically used welding area of up to 50 mm 2 (in this study, 8 mm x 6 mm). Figure 2.1 shows two general configurations of USW system that can be used, namely a wedge-reed system and a lateral-drive system. Both systems work with similar vibrations mechanisms. However they are different in their application and shape [42]. The components of the both systems are briefly discussed in the following two sections.

Wedge-reed USW
In the wedge-reed system shown in   [41] ultrasonic vibration energy to the work piece that is fastened between the welding tip and the anvil. In this study, dual wedge-reed system was used (no anvil is present, rather it has wedge-reed configuration on both sides). The transducer produces a vibration from the piezoelectric disks. Clamping force is applied along the longitudinal axis of the reed. In this technique, the anvil is often a vibrating member, thus the overall motion across the interface increase [41]. Therefore, wedge-reed system is well known for joining higher-strength alloys.
The generation of amplitude in this system is in the range of 10 to 100 μm (peak to peak) [43]. In wedge-reed system, as the transducer is straightly brazed or welded to the reed and reed to the welding tip (via threaded screw), the transducer is only capable of driving the welding tip, and it has no control on weld parameters. Hence no resistance is directly received by it.

Lateral-drive USW
This system consists of a transducer, booster and horn with a welding tip, as seen in Figure   2-1(b). The combination of booster and horn sometimes known as 'welding stack,' which is connected together to the transducer. In a similar manner to the wedge-reed system, the transducer generates a vibration from the piezoelectric disks. When a voltage potential is applied across the piezoelectric crystals, a small mechanical displacement occurs. When the voltage is applied at high frequencies, the transducer produces a corresponding highfrequency mechanical displacement. Then, booster modifies mechanical amplitude produced by the transducer. The horn can further modify the amplitude of the vibration to meet the weld requirements [42]. The ultrasonic vibration of the welding horn is in a parallel direction to the specimen surface as shown in the Figure 2.1(b), creating a scrubbing motion at faying surfaces. This leads the friction, causing shear deformation and a flattening of the surface asperities and then subsequent weld formation. Unlike wedge-reed system, this system is suitable for joining thin gauges. This system can allow for measurement of process parameters, such as actual welding energy, mechanical amplitude and clamping force, more accurately than the wedge-reed system. These accurate values help to correlate the weld strength to the different welding conditions [41].

Principles of USW
The principles of USW and its range of applications remain not fully understood because most studies predominantly focus on ultrasonic plastic welding and its applications. For proper understanding of USW process, the schematic diagram of USW is shown in Figure   2.2. The welding process can be subdivided into four main phases: (i) A compressive static clamping force is applied by the sonotrode to the workpieces to be joined. The protuberances (bumps) of the two surfaces come in contact, but contaminants and oxides layer still avoid the workpieces from bonding.
(ii) Ones the power supply gives electrical energy, the transducer converts this electrical energy into mechanical energy with the same frequency and transmits mechanical vibration via the wedge and reed to the work pieces. Now, the protuberances in contact undergo a shear deformation (Figure 2.2). At first, these contact areas are localized in correspondence to very few irregularities, which grow in number and size during the further vibration cycles.
(iii) Now, this process breaks the oxide and contaminates layers and brings fresh metalto-metal contact and adhesion can occur (Figure 2.2). The plastic shear deformation produces heat due to the friction generation, which lowering the yield strength of the material while the weld area grows. Thus, the material becomes soft not only due to temperature but also to preferential absorption of acoustic energy at the material dislocations: the ultrasonic wave transferred into the metal lattice, and it absorbed the acoustic energy where the defects present (such as dislocation).
(iv) Finally, after the last welding cycle, area under the welding tips is completely Pressure plastically deformed and composed by recrystallized grains. At last, metallic bonding takes place all over the weld zone under the sonotrode. This description of the USW phenomenon provides a microstructural interpretation, which is drawn by DeVries [39].

Applications of USW
USW has extensive application in the industry such as automotive, electronics/electrical appliances, medical instruments and packaging. Conceivably the most common application of USW is in the automobile manufacturing industry, where it is used to weld the sheet metal to form a car. Most of the spot-welding, nowadays can be performed by completely automated industrial robots. Recently, the technique has been adopted for lightweight-softer sheet metal welding in aerospace fields. Another potential application of USW is to join Li-Iron battery cells. Generally, harder metals are more problematic to weld by technique, since more energy is required for achieving optimum weld, which can potentially damage the tool properties. These harder and thicker metals are welded by the so-called dual wedge-reed USW system, in which two transducers are installed to generate supplementary power (which is used in this study). Environments such as in water or a vacuum are also suitable for this process [41,44]. Currently, USW technique is used for alloys of copper, magnesium, aluminum, silver, titanium, nickel, and gold. Along with the spot welding, ultrasonic metal welding systems can be used successfully in different welding applications, such as ring welding, line welding, continuous seam welding [45,46]

Advantages of USW
In the spot welding industries, there are various types of joining techniques available in today's market such as RSW, FSSW, USW. and they all have some drawbacks and some advantages compared to USW. Several challenges are associated with the application of joining lightweight Al and Mg alloys in automotive body construction. In addition, because of the high heat input in RSW, it causes dimensional problems and a weakened HAZ. FSSW consumes much time (1 to 5 sec) compared to the USW (0.5 to 1 sec). Laser-welding equipment is costly and requires skilled operators. In today's aluminum vehicles, clinching, riveting, adhesive bonding find applications but they come with added complexity and cost or weight penalties [25]. Moreover, it can be seen from As discussed earlier in section 1.2 and as seen from Table 2.1, USW is more efficient than RSW since it uses only 0.6-1.5 kJ per weld compared to the RSW (50-100 kJ per weld) has the same advantages as FSSW, in that it is solid-sate friction joining process, but has a shorter welding time with good mechanical performance and no HAZ damage [35]. In comparison with FSSW, USW has been shown to have a shorter weld cycle (typically < 0.5 s) and produce high quality joints that are stronger than FSSW when compared on basis of the same nugget area [24,25]. Besides, the normal FSSW leaves an exit hole after welding [34].
From the point of view of energy consumption, USW is far more advantageous. For example, welding aluminum alloys using a USW process consumes only about 0.3 kWh per 1000 joints [35,22] compared to 20 kWh with RSW, and 2 kWh with FSSW [22]. It is worth to mention that the Ford Motor Company has recently (2005,2007) considered the feasibility of applying USW to Al body assemblies with promising results [24,25]. Some of the direct and relative costs of joining technologies such as RSW, SPR, GMAW, Adhesive bonding, USW, when applied to aluminum joining, are shown in Table 2.2. Among them, USW offers the possibility of joining aluminum automotive body structures in an environmentally friendly and cost-efficient manner.

USW parts and their functions
As shown in Figure 2.3, USW typically contains five key parts: (1) power supply, (2) piezoelectric transducer (3) wedge (4) reed, and (5) 46,49]. They also found that temperature gradient is much higher in a center of the welded joint along with the thickness of the welded samples. However, the actual temperature at the center of the welded joint during the USW process by experimental method is still elusive. Jahn et al. [24] and Bakavos and Prangnell [35] have studied the USWed Al alloy, where they studied the microstructure and mechanical properties of the joints. Bakavos and Prangnell [35] have found that USWed Al alloy can achieve as high as 3.5 kN lap shear fracture load within as short of 0.3 s. Jahn et al. [24] have also worked on the USWed Al alloys and investigated the influence of the anvil geometry and welding energy on microstructures of the welded joints. Since welding and joining of dissimilar metals are bound to involve in the structure of automotive vehicle body, Watanabe et al. [26], have investigated the feasibility of joining of Al-to-steel by USW process. However, so far, significant research is still absent for joining similar and dissimilar ultra lightweight Mg alloy. Thus, the following sections are briefly discussed on the prior work on the welding of Mg alloy by other welding technologies.

Welding of similar Mg Alloy
Because of specific physical and chemical properties of Mg, its welding requires less and precise power input. The solid-state reaction with oxygen forms an oxide layer immediately, when it is exposed to the open environment. This oxide layer is thermodynamically stable, which inhibits the joining process [50,51]. Mg has very low melting point, but the melting point of the oxide is very high. Therefore, prior to the welding of Mg, the oxide coating from the faying surfaces must be removed. Many of the joining technologies failed to remove the oxide layer while USW has potential to remove in first initial cycles due to higher shear force cause by the high frequency vibration. Further, Mg is a very active metal with high thermal conductivity, relatively high thermal expansion coefficient [50] and the rate of oxidation increases as the temperature is increased. Thus, welding of Mg is different from the welding of steels. In the following subsection, the feasibility of joining process of Mg alloy is briefly discussed.

Microstructure, grain size and recrystallization
The microstructure of a material can intensely influence the physical properties of materials such as, hardness, ductility, toughness, strength, wear resistance, corrosion resistance, high/low temperature behavior, and so on. Lang et al. [52]  The alteration from cellular dendritic to equiaxed dendritc structures in the nuggets is attributed to the changes of solidification conditions. It can be seen from    [54].
From these three studies and other similar works [55][56][57], it is seen that due to the higher heat generation in fusion welding, and high strain rate of solid state welding severely changes microstructure. Recently, Santella et al. [38] showed that during the USW of AZ31B-H24 Mg alloy-to-steel, the microstructure of AZ31B-H24 experienced dynamic recrystallization and grain growth. Dynamic recrystallization during USW of Al alloy was also reported by Bakavos and Prangnell [35]. However, it is still not known how the USW parameters influence on the microstructure of Mg alloy, whether it creates the distinct regions like several processes as stated earlier.
It is well known that the grain refinement is controlled by the strain rate and peak temperature. Generally, average grain size decreases with decreasing working temperature and increasing strain rate [58,[59][60][61]. In order to understand the simultaneous effect of both parameters, Zener and Hollomon [62] developed the Zener-Hollomon (Z) parameter which is expressed as, where ̇is the strain rate, R is the universal gas constant, T is the temperature, and Q is the related activation energy [58,[61][62]. This Z parameter appears to be useful in predicting the resulting grain size and has been widely used to combine the temperature and strain rate [59,60]. There are many studies indicating the dependence of product quality and internal structure performance on Z. Recently, Chang et al. [58] developed the relationship between grain size and Zener-Holloman parameter (Z) of friction stir processed (FSP) AZ31 Mg alloy, but information is missing in this respect for USW of Mg alloy.

Crystallographic texture of welded Mg alloy
In addition to the microstructure, mechanical properties are also closely related to crystallographic texture. Many chemical, mechanical and physical properties of crystals depend on their crystalline orientations, and wherever a crystallographic texture exists in polycrystalline materials, directionality or anisotropy of these properties are different [63].   significant interest over the years [64]. It is noted that, for hcp crystalline structure, such as Mg alloys, deformation is dominated by the active deformation modes of slip or twinning, and, hence, the development of texture [65]. The large difference in the critical resolved shear stress between the basal and non-basal slip result in significant anisotropic mechanical properties of Mg alloys [66], and hence the textures that develop during USW would have a strong influence on the mechanical properties. A number of researcher have studied the crystallographic texture (at different positions within the work piece, namely SZ, TMAZ, HAZ and BM) of FSP, FSW, RSW. and their influence on the mechanical properties such as compression and tensile behavior [67][68][69][70][71]. In addition, except for the work done by Zhu et al. [72,73] on the texture evolution of the USWed aluminum foil, most research in the area of USW have worked on the microstructure and the tensile properties of the USWed joints.
However, studies so far have not led to a fundamental understanding of crystallographic texture formation and their influence on mechanical properties of USW Mg alloys.

Fatigue behavior of welded Mg alloy
The process optimization for most of the welding techniques has been mainly determined via the lap shear tensile failure loads. However, the structural applications of these welded joints would inevitably involve dynamic loading in service. In view of the durability and safety of vehicles, the fatigue resistance of the welded joints is of vital importance. In the studies of

Welding of dissimilar alloys
There are many structure applications in which two or more metals are required. This brings about the need for joining dissimilar metals. Two factors are really important in the dissimilar welding of metals: intimate contact between two surfaces and IMCs of the welded samples. IMCs are extremely important, since it is closely related to the mechanical properties of the joints. In the following subsection, the feasibility of joining process of Mgto-Al, Mg-to-steel and Al-to-steel alloys are briefly discussed.

Welding of dissimilar Mg-to-Al alloys
There are a number of lightweight materials available for use in automobiles, but Al stands out for many applications due to its high corrosion resistance, recyclability and compatibility with existing manufacturing techniques. In order to achieve a good combination of the properties of Mg and Al alloys, owing to their lightweight nature, high strength-to-weight ratio, good castability and workability [84][85][86][87][88], the development of reliable joints between Mg and Al alloys is required. Fusion welding of Al and Mg alloys always produces large brittle IMCs and coarse grains, voids, porosities, distortions, large defects in a weld region.
This implies that conventional fusion welding process cannot be practically useful to join Mg and Al alloys. According to Rathod and Kutsuna [36] in situation of dissimilar welds such as Al-to-steel and Al-to-titanium, it is easy to understand solid/liquid state reaction at the faying surfaces of two metals, where only the metal with a lower melting temperature is melted.
However, it is problematic to apply this method to Mg-to-Al alloy joint, where difference between their melting points is minor (melting points of Al is ~660°C and Mg is ~650°C).
While joining Mg and Al alloys, the main disadvantage is the formation of brittle IMCs of Al 12 Mg 17 and Al 3 Mg 2 , and this degrade the mechanical properties of the joint [89]. These phases are brittle and lead to fracture, making it difficult to obtain a non-brittle and strong joint between Mg and Al alloys. Thus, many studies failed to achieve a sound joint between dissimilar welding of Mg and Al. A possible approach to improve the joint strength by reducing the tendency for the formation of brittle IMC would be to consider a third material as an interlayer which could interact with both Mg and Al from their respective binary diagrams [90,91] Several researchers have used third material (interlayer) between the two faying surface of the weld metals in lap joining [86,88,89,92,93] for increasing the lap shear strength of the weld joint.

Welding of dissimilar Mg-to-steel alloys
Joining Mg alloys to steel is challenging since the difference between the two is substantial  [85], and resistance spot welding [96] have been explored for dissimilar welding of Mg-to-steel. Liu and Zhao [95] studied the dissimilar lap joining of AZ31B Mg alloy-to-stainless steel 304 using laser-GTA and found poor mechanical properties of the joints due to the metallic oxides produced at the Mg-Fe interface. Wei et al. [97] studied the dissimilar lap joining of AZ31 Mg alloy-to-stainless steel SUS302 and found void and micro type defects at the interface. Chen and Nakata [85] reported that FSWed AZ31 Mg alloy-tozinc coated steel joints showed higher fracture loads than the FSWed AZ31 Mg alloy-tobrushed finish steel joints (without Zn coating), which demonstrates that the presence of a Zn coating considerably improved the weldability of Mg-to-steel. In most studies, Zn coated steel has been used for the welding of dissimilar Mg alloys to steel and the joints achieved higher lap shear tensile strength than the joints made with uncoated or bare steel. This study could be extended to some other interlayer, which could interact with both Mg and Fe elements. Liu et al. [92] in the hybrid laser-TIG welding of Mg-to-steel have used Sn as an interlayer and showed the improvement of the mechanical properties.

Welding of dissimilar Al-to-steel alloys
Many studies have shown that the major metallurgical problem of joining Al alloys with steel is the formation of brittle IMCs such as Fe 3

MATERIALS AND EXPERIMENTAL PROCEDURE
This Chapter provides details about the materials and experimental procedures used in this research. Specifically, details are presented for the equipment and methods that are used for the characterization of the USWed joints.

Experimental materials
The current research involves commercial 2 mm thick sheet AZ31B-H24 rolled Mg alloy with the composition listed in

Processing parameters of USW
The USW system employed was a dual wedge-reed, Sonobond-MH2016 machine. The  faying surfaces, the specimen were ground using 120 emery papers perpendicular to the ultrasonic vibration direction, and then washed using acetone and dried before welding.

Welding tip
Ultrasonic vibration is applied to the welding tip. When clamping force is applied to the tip, the texture of tool grips the weld specimens, which effectively isolates the relative motion to the joint interface. A sound joint could be formed with the proper welding tip's texture and geometry. Numerous welding tips available in the market, depending on the texture and geometry. The contact area of the flat tip was much greater than with the spherical tips, and therefore a lower applied power density was needed to apply. Thus, in this study, flat standard serrated (kurnl pattern) tip was used in all experiments as shown in the

Metallography
Cross-section samples for optical microscopy (OM) and JSM-6380LV Scanning electron microscope (SEM) were prepared from the welded lap joint coupons by sectioning through the weld region parallel to the vibration direction of USW. The mounted samples were manually ground with SiC papers up to a grit of #1200 with water as the lubricant and then polished using 6, 3 and 1 µm diamond paste followed by the 0.05 µm MASTERPREP solutions. The polishing lubricant for the diamond paste was a mixture of rust inhibiting solution and distilled water (10% solution by volume). Ethanol was used as a cleaning agent during the polishing stages. To reveal the microstructure of USWed Mg, etching was done using a solution of ethanol (10 ml), picric acid (5 g), acetic acid (5 ml) and water (10 ml).

Quantitative image analysis
Microscopic images were taken using an OM and SEM. Image analysis was performed using routine Clemex software to obtain the grain size. The Clemex image analysis system was composed of Clemex CMT software adaptable to ASTM standards, a Nikon optical microscope (10× eye piece, five different object lenses with magnifications of 5×, 10×, 20×, 40×, and 100×), a high-resolution digital camera, and a high performance computer to carry out the detailed analysis. Five images were taken at the center of the weld to calculate the average grain size of each welded sample. Point and line analysis was conducted on the welded joint using SEM, equipped with Oxford EDS.

X-ray diffraction for phase identification
For USWed Mg-to-Mg joints, XRD (machine manufactured by Panalytical )was carried out on top surfaces of BM, HAZ (5 mm away from the NZ) and NZ, fracture surface (FS). And for USWed dissimilar joints, XRD was performed on both matching fracture surfaces of welded joints after tensile shear tests, using CuKα radiation at 45 kV and 40 mA. The diffraction angle (2θ) of the incident X-ray beam varied from 20° to 100° with a step size of 0.05° and a dwell time of 2 s per step.

Temperature profile during USW
The temperature profile during USW was measured using a K-type thermocouple (Ni-Cr vs Ni-Al). In order to eliminate the effect on the welding process, a groove was made on the lower piece of each pair of weld samples. The thermocouple was laid into the groove and parallel to the vibration direction as shown in the Figure 3.3. The thermocouple was coupled with a data acquisition instrument for collecting the temperature data during the USW.

Microhardness test
A computerized Buehler microhardness testing machine was used for the micro indentation hardness tests at the center of the welded joint in USW of similar Mg alloy study, using a 1 1

K type Thermocouple
Groove load of 100 g for 15 s. The average of ten indentations was used for accuracy in microhardness test of USWed similar Mg alloys. In dissimilar welding study, as shown in

Lap shear tensile test
To evaluate the mechanical strength of the joints and establish the optimum welding conditions, tensile shear tests of the welds (samples dimensions is shown in Figure 3.1) were conducted to measure the lap-shear failure load using a fully computerized United testing machine with a constant crosshead speed of 1 mm min -1 in air at room temperature. As shown in the Figure 3.5, in the tensile lap shear testing, restraining shims or spacers were used to minimize rotation of the joints and maintain the shear loading as long as possible.

Welded sheets
Indentations Figure 3.5: A 3D view of the lap shear tensile and fatigue test specimen.

Fatigue Test
Fatigue tests were conducted on the welded joint (

Microstructure and grain size measurement
Microstructure characterization was focused on the nugget zone (NZ) at the center of the weld samples. Another reason is because, unlike Al, Mg has low stacking fault energy, in the range of 60-78 MJ/m 2 for pure Mg [103]. Additional possible factor is the high grain boundary diffusion rate of Mg compared to Al [58]. These three factors explain why dynamic recrystallization can take place in AZ31-H24 alloy during welding. Dynamic recrystallization during USW was also reported by Bakavos and Prangnell [35] and Allameh et al. [104]. A large shear strain rate is introduced by the high frequency vibration and high temperatures developed from friction and plastic deformation during USW; combination of both factors generate dynamic recrystallization and grain growth in the weld.
In addition, as energy inputs increase from 500 J to 3000 J, aspect ratio of the grains also increases from 1.54 to 1.63. During USW of sheet metal, normal and shear forces act on the parts to be welded and the weld interface. These normal and shear forces are the result of clamping pressure and the ultrasonic vibrations of the tool, respectively, which transferred onto the parts to be welded. It is believed that the reason behind the increasing aspect ratio is  Grains with large aspect ratio, dispersed along the flow lines, dominate the deformation microstructure. Similar microstructure was also observed by Jahn et al. [24].

Relationship between Zener-Hollomon parameter and grain size
As mention earlier in the literature review (section 2.2.1.1), Zener-Hollomon (Z) parameter appears to be useful in predicting the resulting grain size and has been widely used to combine the temperature and strain rate [59,60]. There are many studies indicating the dependence of product quality and internal structure performance on Z. The effect of grain refinement is controlled by the strain rate and the peak temperature; information is lacking in this respect for USW of AZ31B-H24 alloy. An objective of this chapter was to investigate the relationship between grain size and the Zener-Hollomon parameter. From Eqn. 2.1, in order to develop this relationship, shear strain rate and thermal profile of the USW of Mg alloy need to be identified. Thus, in following sub-sections, shear strain rate is calculated and thermal profile is measured.

Shear strain rate calculation
Using a frequency of 20 kHz in USW, very high strain rates and strains can develop during shearing of the small (micron-sized) asperities between the welding samples. Gunduz et al. showed that USW produces plastic strains in the materials being joined at very high strain rates (10 3 s -1 ) in a fraction of a second [105]. The strain rate estimated by Shriraman et al. in USW was 10 4 -10 5 s -1 [106]. In this study, shear strain rate was calculated as Bates et al. [107] estimated in vibration welding as, where, f is the frequency, is the thickness of the weld nugget between the plates and A is the amplitude of USW machine. The amplitude generated using USW ranges from 30-60 µm [26, 35,37,108,109] and the relationship between energy input and amplitude is given by [110]. As a first approximation, was used to calculate the amplitude (

Temperature measurement
During USW, heat is generated at the weld interface and the surrounding area as well as at the sonotrode (welding tip) surface owing to plastic deformation and friction. Plastic deformation in the specimen is caused by two different phenomena, namely, surface friction dissipation and volume ultrasonic softening effects. Temperature plays an important role in bond strength in the initial period [29]. There is limited published data on thermal measurements in USW because of the small size of the welds and rapid weld times [35].
Results from temperature measurements of the USW, using thermocouples that are placed at the center of the weld, are depicted in   kJ/mol during the dynamic recrystallization [58,60], the value of universal gas constant is 8.314 J/mol K and welding temperature was calculated as shown in  where D is the dynamically recrystallized grain size, Z is the Zener-Hollomon parameter, α and β are constants [58,59]. The previous studies on compression test, tensile test and FSP show that the average recrystallized grain size increases with increasing working temperature and decreasing strain rate [58,59]. However, in this investigation both temperature and strain rate increase dramatically as the energy input is increased from 500 J to 2500 J, resulting in an increase in grain size due to the effect of the temperature overwhelming the opposing effect of the strain rate. It should be noted that the dynamic recrystallized grains developed at higher Z values are much smaller than those generated at lower Z values ( Figure 4.3). It can be seen from

Microhardness profile
Another mechanical property that may be important to consider is hardness, which is a measure of a material's resistance to localized plastic deformation. For welds, microhardness testing is an easy way to measure a specimen's hardness. Hardness testing is performed on welds to evaluate the strength of the weld. In particular, the hardness at the center of the weld and around the HAZ is of interest, which can be helpful to estimate the brittleness of the weld. Here, to develop Hall-Petch type relationship between grain size and hardness, microrhardness test was only performed at the center of the nugget zone. The average of 10 indentations was used for accuracy.

Effects of USW on microhardness
Unlike RSW, USW displays no clearly discernible fusion zone or heat affected zone that can degrade the strength of the metals being joined [24,25]. It can be inferred that the HAZ is small, because of poor conductivity of the sonotrode material (steel) compared to the thermal conductivity of the samples. Hardness values over the entire area of USW fall within 10% of each other [25]. Similar variations were found in the base metal itself. The maximum hardness in the welded samples ranges from HV55 to HV72 and is lower than that found in base metal (HV73). It can be seen in Figure 4.4(a) that hardness decreases with increasing energy input due to increasing grain size.

Hall-Petch type relationship of microhardness of USWed Mg alloys
The well-known "Hall-Petch relationship" (relationship between yield strength and grain size) style was adopted to show the effect of grain size on hardness. Here, it is called as a "Hall-Petch type" relationship of hardness and grain size. As calculated, the linear relationship from the graph of Figure 4.5, the Hall-Petch type relationship of hardness and grain size is

Crystallographic texture of USWed Mg alloy
It is well know that the tensile properties of the weld depend on several microstructural factors, such as dislocation density and grain size, second phase particle [68]. However, besides these, crystallographic orientation also intensely influences plastic deformation during the tensile or compressive test because the plastic deformation arises from slip on the close-packed planes with the maximum critical resolved shear stresses [68].    In the BM, intensity of (0002) basal plane is higher than any of other planes, which is in complete contrast to the Mg powder data. This suggests that the BM itself exhibits a typical basal rolling texture, owing to the mechanical deformation during the cold rolling of Mg sheet. It can be seen from  ) on the FS, the intensity of (0002) plane was approximately two times higher than that in the BM, which could be the cause of lap shear tensile deformation. However, in NZ, the intensity of (10 1) plane was higher than the intensity of (0002) plane (similar like Mg powder data), which indicates nearly random grain orientation, reflecting its fully recrystallized coarse grain structure as shown in rpm [58]. It can be noticed that HAZ does not exhibit the texture changing, hence, it has similar texture profile as BM which is in agreement with some literatures reported that USW does not produce HAZ [24,25].       two components generates a pole figure of rolled Mg (BM of USW) that is almost a (0001) fiber texture. This (0001) fiber texture can be easily seen by prismatic (10 0) plane (circular ring type structure at 90° from the center of the pole figure). Moreover, the crystallographic fibering in the BM could be even better seen from the pyramidal planes (10 1 In NZ of 1000 J energy input, a strong basal texture with (0002) normal parallel to the ND was found (Figure 4.9(a), NZ-(0002) plane). The intensity of (0002) plane was much higher than that of the BM, suggesting that after the welding more grains followed the basal texture.

Crystallographic texture by pole figures profiles
Such a texture change occurring in the NZ was attributed to the intense localized shear plastic flow during USW. It can be seen from the (10 0

Lap shear tensile strength of joints
It is known that the structural application of any new materials' components requires proper welding and tensile resistance to guarantee their durability and safety. Thus, here, the lap shear strength test of the USWed Mg alloy was conducted and the result were plotted against welding energy inputs as shown in Figure 4.10. The maximum lap shear strength was found to be 89 MPa at 2000 J energy input (calculated as the maximum lap shear fracture load divided by the nugget area of (8×6 mm). As shown in Figure 4.10, with increasing energy input, lap shear strength increased in the beginning owing to high temperatures and strain rate, which accelerate diffusion between the Mg alloy sheets, but at very high energy inputs, it was decreased. This may be because, at low welding energies (1000 J), Mg has not been sufficiently deformed plastically to fill the valleys of the knurl patterns of the sonotrodes, as

Lap shear tensile fractography
In lap-shear tensile tests, the welds generally fracture along the weld interface for input energies of less than and equal to 2000 J, while fracture occurs at the periphery of the weld penetrate into the material more deeply with higher energy. In particular, due to this crack, the base materials were partly pulled out from the welded interface. Thus, further increase in energy input (2500 J) resulted in the whole nugget pull out as shown in Figure 4.11(e).
Similar nugget pull out phenomenon in the USW process was also reported in [22]. A magnified region for 1000 J energy input sample (    Figure 4.14 (c), and also observed in [117]. Finite element simulation showed that the normal tensile stress concentration at the periphery of the nugget could reach as high as more than five times the average stress under tensile-shear loading [117]. Thus, this micro level crack experienced higher stress concentration effect during cyclic loading which allowed the cracks to grow toward the outward Mg sheet. By closely observing    welding area corresponding to the area of the USW tip (8×6 mm 2 ). Likewise, for the FSSW, the hollow cylindrical area has been used (pin diameter was 5 mm and shoulder diameter was 13 mm) [34]; for the RSW the area of the electrode (electrode diameter 7.8 mm) has been used [75] and for the FSLW the product of pin diameter (6.35 mm) and sample width (20 mm) has been used [53]. It is seen that the USWed joints displayed a higher fatigue         here is based on the assumption that the majority of the samples failed from the kinked crack growth mode. This assumption is also supported by previous studies on FSSW [79,120,78,121].

Global stress intensity factor solutions for main cracks
To develop an engineering spot welding fatigue model, the three-dimensional spot weld problem can be idealized as a two-dimensional crack problem as reported by Newman and Dowling (ND) [78] and Lin et al. [79]. In order to estimate the fatigue life, it is first necessary to examine the global stress intensity factor solutions for the main notch tips treated as a crack in the lap shear specimens based on the studies of Pook [80], Swellam et al. [83], Zhang [81,82] and Lin et al. [79].
Pook [80] presented the global stress intensity factor solution for two circular plates with connection under uniform distributed loads along the clamped outer edges (here in this case point A and point B as shown in Figure 3.5) in the lap-shear specimens of thickness t and nugget radius r under the applied force P as, , (4.5) . (4.6) Pook [80] (Figures 4.14(a) and (b)), which is similar to a kinked crack. Thus, the local k I and k II solutions can be expressed as [78], , (4.13) , (4.14) where (k I ) 0 and (k II ) 0 represent the local k I and k II solutions for the kink length a approaching 0. In order to calculate the local stress intensity factors for the kinked crack, crack angle α = 45˚ has been used ( Figure 4.14(b)). Since the kinked crack growth is under local combined mode I and mode II loading conditions, an equivalent stress intensity factor, k eq, can be defined as [78], , (4.15) where β is an empirical constant to account for the sensitivity of materials to mode II loading conditions. In the absence of any further information in the literature for the value of β of Mg alloys, here in this study, β as 1 has been taken for the USWed Mg alloy joints (β = 1 fits much better than other value).

A fatigue crack growth model
A Paris crack growth relation as a function of could be expressed with an adjustment for the R-ratio effect as suggested by Walker [123], , (4.16) where N is the number of fatigue cycles, C and m are the material constants, and ∆k eq is the equivalent stress intensity range. As with Newman and Dowling [78], it is assumed that the local stress intensity factors remain almost constant throughout fatigue crack progress. Thus, the local stress intensity factor solutions considered in this study are assumed to be independent of the kink length a. Integrating Eqn. 4.16 gives, . (4.17) Since the fatigue crack will align and grow at an angle of α, t/sinα is substituted for the crack length a in Eqn. 4.17, where t is (t o -t I ) as discussed earlier.

Applicability of the fatigue model
To demonstrate the applicability of the proposed fatigue model for spot welds, the estimated fatigue lives based on the kinked fatigue crack growth model for the USWed AZ31B-H24 Mg alloy were compared with the experimental results of USWed Mg-to-Mg joints, as plotted in Figure 4.21. The predicted fatigue lives were obtained from the following  [81,82] and Pook [80].  [80] and Lin [79] provide more conservative curves compared with those of Swellam [83] and Zhang [81,82]. While some differences between the predicted fatigue lives and experimental data for the USWed Mg-to-Mg joints are still there, the present study in the fatigue life estimation seems to be somewhat better than the prediction using a structural stress model for the USWed Mg-tosteel samples [77]. Further studies in developing more accurate fatigue life prediction models are needed.

Summary
 USW of similar AZ31B-H24 Mg alloy sheets was successfully achieved.
 The influence of the input energy on the strain rate and temperature generated during USW was evaluated. After USW, the grain size was observed to increase with increasing welding energy.
 A higher welding energy input caused greater shear strain rate and higher peak temperatures. These both factors were mostly accountable for controlling the dynamically recrystallized grain size.  In this Chapter, it is seen that USW process has successfully achieved sound joints of similar Mg-to-Mg alloys, which shows the potential of USW process to join Mg alloy in spot welding industry. However, Al has already a wide variety of structural applications in the transportation industry due to their excellent properties, such as good ductility, formability and thermal conductivity. To achieve the combined properties of both alloys, recently, demand for dissimilar metal joints of Mg-to-Al alloy has risen. Thus, it is worth to develop some new potential welding process, since many other welding techniques failed to produce sound joints for this kind of challenging combination of materials. Chapter 5, will characterize the USW of Mg-to-Al joints and show how a Sn interlayer can increase the lap shear tensile strength of Mg-to-Al joints.

ULTRASONIC SPOT WELDING OF MAGNESIUM-TO-ALUMINUM ALLOYS †
The aim of this Chapter is to characterize the weld of Mg-to-Al joints, to study the cause of the failure in detail, and to improve the lap shear fracture strength by using a Sn interlayer inserted between the faying surfaces during USW. The selection of Sn in the present study was based on the Mg-Sn and Al-Sn binary phase diagrams [91,126], which show that Sn may interact with Mg and generate IMCs, while Sn is dissolved into Al to form a solid solution of Sn-Al. Furthermore, Sn was selected on the basis of the findings that it improved the wettability of Mg and Al during the welding process [127] and also refined the grain size in the Mg alloy [128].

Microstructure
Microstructural characterization was conducted across the welding line of the samples.  defects, such as crack or tunnel type defects. A sound joint was obtained under most of the welding conditions, but a crack was observed on the weld surface at energy inputs of 2500 J and above (similar crack as shown in 4.12 (b)). This crack was thought to be related to frictional heat generated between the welding tip (made of steel) and welded sheets and the restraint between them leading to tensile residual stresses in the sheet. It can be seen from heterogeneously. This may be because the temperature distribution is not homogenous across the welding line owing to the short initial welding time. In the 500 J energy input sample, discontinuous layers of IMC have been found while in the 3000 J energy input sample continuous layers of IMC have been observed owing to the equal distribution of the temperature at the higher-energy input ( Figure 5.1(b)). High local temperatures and strain rate and pressure promote chemical reaction and accelerate diffusion, leading to metallic bonding between the two sheets [129]. The diffusion and chemical reaction between the Mg and Al alloys result in the formation of IMC. During the USW, in first initial cycles, oxide films at the faying surfaces first breaks down locally at asperities on the contacting surfaces and bring fresh metal-to-metal contact. Now, by continuing applying high frequency vibratory energy for longer period of time permit inter-diffusion of Mg and Al to occur. If the kinetics is adequately rapid, it creates the simultaneous formation of IMCs throughout the faying surfaces. Several previous works on the dissimilar welds have suggested that the weld strength is seriously affected by the thickness of the IMC layer formed at the weld interface.
On examination of the cross sections of the Mg-to-Al welds, the IMC layer was evidently observable at the weld interface even after short welding times . Figure 5.1(c) shows that the IMC thickness increses with increasing welding energy input. At 500 J welding energy input, a discontinuous IMC layer with a maximum thickness of 2 µm was already present. With higher weld energy input (3000 J), the thickness of the IMC layer increased to 25 µm.

Energy-dispersive X-ray spectroscopy analysis
To identify the IMC phase in the interface layer formed between the Mg and Al alloy during the USW, EDS has been conducted at the center of the cross section and on the fracture surface of the weld on the 1500 J energy input samples. was also performed. However, IMCs of Al 12 Mg 17 were not found, which indicating that the failure predominantly occurred in-between the Al alloy and the intermetallic layer, which normally stayed at the Mg side or from the cracks of the IMCs in the reaction layer. The present dissimilar USW is likely exposed to peak temperatures above 460 ºC as confirmed by thermocouple measurements at the center of the weld of similar Mg alloy in earlier section 4.2.2. This peak temperature is sufficient for mutual diffusion between Mg and Al.
Even though USW is a solid-state joining process, in dissimilar USW it may be possible that constitutional liquation occurs due to the high frequency vibration and resulting heating.

X-ray diffraction analysis
For further separate verification of IMCs, the fracture surface of the Mg side of the welded sample was analyzed by XRD. Figure 5.3 shows large peaks of Mg and small peaks of IMCs

Microhardness
Unlike RSW, USW displays no clearly discernible fusion zone or heat affected zone that can degrade the strength of the metals being joined [24,25]. It can be inferred that in USW, the HAZ is small because of poor conductivity of the welding tip (steel) compared to the thermal conductivity of the samples. The hardness profile across the non-uniform IMC layer, which was measured along the diagonal line, is shown in Figure 5.4. It can be seen that hardness decreases with increasing energy input, owing to increasing grain size at higher temperature.
The non-uniform IMC layer in the weld center has hardness values between 200 and 300 HV.
This higher hardness is due to the brittle IMCs of Al 12 Mg 17 present in the center of the weld.
Similar results were also obtained by other researchers [88,131,132,93,134]. Therefore, the hardness test also verified the presence of the hard IMCs. Thus, USW of the Mg and Al alloys produces a brittle interfacial layer composed mainly of IMCs, which provide an easy fracture path.

Lap shear tensile strength of joints
The results of the lap shear tensile tests of the Mg and Al alloy weld joint are shown in  [135,136]. With increasing energy input, lap shear strength was increased in the beginning owing to high temperatures and strain rate, which accelerated diffusion in between Mg and Al alloy, but it decreased at very high-energy inputs. It seems that the decrease in lap shear strength is related to the thickness and presence of cracks in the reaction layer. Therefore, as thickness of the brittle IMC layer increases, the lap shear strength decreases. This was a consequence of a competition between the increasing diffusion bonding arising from higher temperatures and the deterioration effect of the intermetallic layer of increasing thicknesses. Figure 5. 5(b) shows the relationship between average value of the failure energy (as determined by integration of the lap shear fracture load curves) and welding energy input. It can be seen that at the beginning failure energy increases with the energy input and reaches a maximum around 1.14 kN.mm and then it decreases due to the larger brittle IMC layer. Based on these results ( Figure 5.1(c) and Figure 5.5(a)), it can be concluded that the thickness of IMC was closely related to the strength of Mg-to-Al USW joint. In Figure 5.2(a), the fracture surface appeared coarse and dark gray. The fracture surfaces on the Mg alloy consisted of protuberances and cleavage cracks, which were considered as proof of brittle fracture mode.

With tin (Sn) as an interlayer in-between the faying surface
Now, Sn interlayer was placed in between the faying surfaces of Mg and Al and SEM, EDS, micro hardness test were performed on the weld cross section. Lap shear tensile test was also performed to check the effect of Sn interlayer on the joint properties.

Microstructure
Microstructural characterization was conducted across the weld line of the samples.

Energy-dispersive X-ray spectroscopy analysis
The IMC layer displayed a composite-like eutectic structure, as shown in Figure 5  Mg 2 Sn would occur at a temperature of 203C [90].
In the USW the simultaneous application of localized high-frequency vibratory energy and moderate clamping force leads to relative motion and friction heat at the interfaces [24,15] between Al-Sn and Mg-Sn, which would cause the melting and coalescence of Sn. In the presence of the Sn interlayer in the USW, Al and Sn combine to form solid solutions, while Mg and Sn combine to form β-Sn and Mg 2 Sn IMCs. The Mg 2 Sn phase has an antifluoritetype (CaF 2 ) AB 2 crystal structure with a moderately high melting temperature of 770C and a lattice parameter of a = 0.676 nm [137]. It is apparent that the large Mg 2 Sn particles resulted from the eutectic reaction (L →β Sn+ Mg 2 Sn). The formation of the IMCs of Mg 2 Sn implies that the exothermic chemical reaction Mg+Sn→Mg 2 Sn+Q must occur in the interlayer, where Q is the mole heat release of Mg 2 Sn, which is −80.75 [138] kJ mol -1 , compared to the fusion heat of Al [139] and Mg [139] of 10 and 9 kJ mol -1 respectively.
Thus, it is possible that the large particles of Al could be remelted by the heat formation of Mg 2 Sn and resolidify as smaller particles or disappear under the condition of rapid cooling during the welding process. The addition of Sn to the lap joint was observed to refine the grain size in the fusion zone and the base Mg alloy [92,128] due to the presence of a eutectic structure, which restricts the growth of the Mg grains. Furthermore, it also improves the wetability of Mg and Al during the welding process [127]. Thus, the surface tension of the liquid is reduced so that more liquid spreads evenly over the surface of the base metal.

X-ray diffraction analysis
To

Microhardness
The hardness profile across the welded joint diagonally is shown in Figure 5.9. The hardness of the very thin IMC layer is the mean of measurements at three different locations along the

Lap shear tensile strength of joints
As shown in Figure 5.10(a), the maximum lap shear strength of Mg-to-Al USW joints with the addition of a Sn interlayer was improved (~41 MPa) when compared to that of a direct Mg-to-Al USW joints (~35MPa). In addition, the Sn interlayer also led to an energy saving since the optimal welding energy required to achieve the highest strength decreased from ~1250 J to ~1000 J. The failure energy shows a similar trend as shown in On the other hand, at higher energy inputs, the weld specimen was subjected to higher temperatures under larger vibration amplitude for a longer time, resulting in more Sn interlayer being squeezed out. As a result, the lap shear strength and failure energy of the Mg-to-Al USW joints with a Sn interlayer also increased initially with increasing welding energy, reached the maximum values, followed by a decrease with further increasing welding energy. Figure 5.10(c) shows a summary of the maximum lap shear strength for different types of joints made at a welding energy of 1000 J. It is seen that the USW Mg-toAl joint without a Sn interlayer was approximately 25% lower than the USW Al-to-Al joint and 60% lower than the USW Mg-to-Mg joint. However, the USW Mg-to-Al joint with a Sn interlayer was approximately 5% higher than the USW Al-to-Al joint and only 40% lower    (c) (d)    layer of Sn-Mg 2 Sn eutectic structure, which will be identified later.

Energy-dispersive X-ray spectroscopy (EDS) analysis
To identify the generation of phases during the USW, EDS line analysis has been performed across the center of the weld nugget of USWed Mg-to-bare steel, with galvanized steel and with Sn interlayer (placed in-between Mg and bare steel samples), respectively, and shown in Figure 6 To better understand the diffusion process of Mg and Zn, USW process is divided into three stages. In the first stage, the oxide film on the surface of Mg alloy was destroyed by the

X-ray diffraction analysis
To further verify the above microstructural observations, XRD data obtained on both matching fracture surfaces of USWed Mg-to-galvanized steel and Mg-to-bare steel with Sn interlayer joints after tensile shear tests are shown in Figure 6. as demonstrated by [145]. Failure in the USWed Mg-to-bare steel occurred as a combination of partial nugget pull-out mode and partial "cohesive failure" mode, giving rise to a higher tensile shear strength which will be seen in the following section.

Microhardness
The hardness profile diagonally across the welded joint is shown in Figure 6.4. It is seen that characteristic asymmetrical hardness profiles across the USWed Mg-to-galvanized steel and

Lap shear tensile strength of joints
As shown in interlayer led to an energy saving since the optimal welding energy required to achieve the highest strength decreased from ~1750 J to ~1500 J in the Mg-to-steel dissimilar joint.       [22,98,146,37,147,148], which can seriously degrade the mechanical properties of the joint. Therefore, it is worth to study the feasibility of USW to join Al-to-steel.

Energy dispersive spectroscopy analysis
To identify the generation of phases during the USW of Al and galvanized steel, EDS point and line scans have been performed across the center of the nugget zone and shown in Haddadi et al. [100] in USW and Ueda et al. [102] in RSW of Al-to-galvanized steel joint.
Thus, it can be concluded that in the USW of Al-to-galvanized steel, only Al-Zn eutectic film was present at the interface. However, by closely observing Figure 1(c), there is a very thin IMC layer present in-between the Al-Zn eutectic film and Fe matrix, called region 2 (shown by the arrow in Figures 7.1(c) and (d)). As can be seen in the EDS line scan- Figure   1(c), at around 15 µm distance, the   concentration of Al was starting to decrease and then suddenly spiked at a distance of approximately 21 µm and then again decreased. Further the scan clearly shows the higher amount of Fe element and a low count of 0 to 10 of Zn elements at the particular region 2, which evidently suggests that this layer is the IMCs of Al and Fe elements. According to the binary phase diagram of Al and Fe, these IMCs could be Fe 2 Al 5 and FeAl 3 phases, which were later confirmed by the XRD technique. Many studies have also reported the presence of these brittle phases during the welding of Al and steel, which is the main cause of fracture [22,98,146,37,147,148].   in-between the Al-Zn eutectic film and Fe matrix.

X-ray diffraction analysis
To further verify the above microstructural observations, XRD patterns obtained on both matching fracture surfaces of Al and galvanized steel sides after lap shear tensile test are shown in Figures 7.3(a) and (b). It is clear that apart from strong peaks of Al on the Al side  earlier, many studies have reported the presence of Fe 2 Al 5 and FeAl 3 brittle phases during the welding of Al-to-steel [22,37,98,146,147,148], even in the USW of Al-to-steel [22].
Haddadi et al. [100], have reported the formation of the brittle IMCs of Fe 5 Zn 21 and Fe 3 Zn 10 in the USW of Al-to-galvanized steel. However, in this present experiment, no peak was evident for the Fe-Zn IMCs. In the Al-Fe phase diagram [130], there are numerous Al-Fe IMCs, such as FeAl 3 , FeAl 2 , Fe 2 Al 5 , Fe 3 Al, and FeAl that are present in the binary phase diagram of Al-Fe. The first phase generated during metal-to-metal interaction is the phase with the most negative heat of formation at the concentration of the lowest eutectic of the binary system. Now, these phases react with each other to form another phase. The lowest eutectic point in Al-Fe binary phase diagram has at 0.02% Fe. At this particular composition, FeAl 3 has the lowest effective free energy of formation and therefore, generation of FeAl 3 phase is kinetically favored. Then the reaction between Fe and FeAl 3 phases will occur to form another phase with a composition between that of the interacting phases and closest to that of the lowest eutectic point, which is in the order of Fe 2 Al 5 , FeAl 2, FeAl, Fe 3 Al.

Lap shear tensile strength of joints
As shown in Figure 7.4, the maximum lap shear tensile strength of USWed Al-to-galvanized steel was ~76 MPa (~3.7 kN). The lap shear tensile load increased with increasing energy inputs and peaked 2000 J energy input and then decreased. At lower energy inputs, the temperature was not high enough to soften or diffuse the Zn interlayer into the Al. On the other hand, at higher energy inputs, the weld specimen was subjected to higher temperatures under larger vibration amplitude for a longer time, resulting in more Zn interlayer being squeezed out (Figure 7.1(b). As a result, more diffusion between Al and Fe occurred, which deteriorated strength. Thus, it can be concluded that this behavior was a consequence of the competition between the increasing diffusion bonding arising from higher temperatures at the higher energy inputs and the deteriorating effect of the brittle intermetallic layer of increasing thickness. Haddadi et al. [22] reported that the weld strength is limited by the development of an Al-Fe IMC layer, which increases in thickness with weld time. Thus, Watanabe et al. [37] obtained only a 0.6 kN lap shear tensile load for welding time of 2.5 s for the USW of A5052 Al alloy-to-SS400 mild steel (without any coating or interlayer).
Haddadi et al. [22] reported maximum lap shear strength of only 3.1 kN for USW of Al6111-  To further verify the fracture location and chemical composition presented in the Table 7 Therefore, it can be inferred that fracture could initially occur in-between the regions i and l with a small applied load and then with continuing higher applied load, the joints failed from the Al containing region. This is also verified by the magnified regions h and k (Figure 7.5(h) and (k)), where opposing shear-tear like fracture feature were found on Al and galvanized steel sides, suggesting that dimple-rupture failure mode occurred in-between the region h and k, while there was interfacial de-bonding-like failure in-between the region i and l.

Fatigue behavior and failure mode
Fatigue test results of the dissimilar USWed Al-to-galvanized steel joints obtained at room temperature, R = 0.2, and a frequency of 50 Hz are plotted in

Contribution
This study has demonstrated that USW can be used to produce satisfactory similar and dissimilar welds in difficult to weld light metals such as sheet Mg, Al and steel. It has also been shown that using an appropriate interlayer such as Sn in dissimilar USW it is possible to achieve sound welds with good mechanical properties through the formation of non-brittle IMCs.
The research results should effectively promote the applications of lightweight Mg and Al alloys in the automotive industry and contribute to the reduction of green house gas emissions.

Major conclusions
Joining of dissimilar materials Mg-to-Al, Mg-to-steel and Al-to-steel was successfully achieved using USW process.
1. USW joints exhibited acceptable lap shear tensile fracture strength and fatigue life with only ~5-10% hardness reduction at WZ.
2. In all joint combinations considered in this study, lap shear tensile strength of the joints increased with increasing energy inputs and then decreased.
3. Dynamic recrystallization was observed at the center of the Mg-to-Mg joints from ~1000 J welding energy and onwards due to large shear strain rate and higher peak temperature which led to a more random crystallographic texture. 4. In tensile and fatigue tests, all dissimilar joints failed from the interface while most of the similar joints failed at the edge of the NZ. USWed joints displayed higher lap shear tensile strength and longer fatigue life compared with other welding processes. 5. A life prediction model for spot welded lap joints based on Newman and Dowling [78] agreed well with the experimental results.
6. Dissimilar USW joints with interlayer exhibited higher lap shear tensile strength and consumed lower welding energy to achieve acceptable joints than dissimilar USW joints without interlayer.
7. The presence of suitable interlayer/coating (Sn and Zn) on the fracture surfaces reduced the tendency for brittle cleavage-like fracture and encouraged ductile type fracture behavior with rougher fracture surface.
8. The role of the interlayer in dissimilar USWed joints was to replace the formation of brittle IMCs in non-interlayer joints with solid solutions and non-brittle IMCs.
 In Mg-to-Al welds with Sn interlayer, the joints contained solid solution of Sn with Mg and Sn with Al as well as a Sn and Mg 2 Sn eutectic structure.
 In Mg-to-steel welds with Sn interlayer, the joints contained solid solution of Sn with Mg and Sn with Fe as well as a Sn and Mg 2 Sn eutectic structure.
 In Al-to-steel welds with Zn coating on steel surface, the joints contained Al-Zn eutectic and thin IMCs of FeAl 3 and Fe 2 Al 5.
9. The desirable welding energy input for USW of light weight alloys in this study with sample thickness of 1-3 mm was found to be between ~1000 J and ~2000 J.
Exceeding the 2000 J energy input led to the thinning of the welding sheet and failure at the edge of the NZ.

Recommendations for future work
A number of interesting extensions to the present work could be made in the future. These are listed below: 1 As seen earlier in the (