New developments in advanced welding

Ultrasonic metal welding

K. GRAFF, Edison Welding Institute, USA

9.1 Introduction

The application of ultrasonic energy to materials joining processes has been in use for a number of years. While it was used for grain refinement in molten metals in the 1930s, for soldering, enhancement of resistance welding and in conjunction with arc welding in the 1940s, and for joining plastics in the 1950s, the ultrasonic metal welding process was first demonstrated in the early 1950s. It was found that ultrasonic vibrations were capable of creating a weld in metal parts without the need for melting the base metals.

The process of ultrasonic metal welding is one in which ultrasonic vibrations create a friction-like relative motion between two surfaces that are held together under pressure. The motion deforms, shears and flattens local surface asperities, dispersing interface oxides and contaminants, to bring metal-to - metal contact and bonding between the surfaces. The process takes place in the solid state, occurring without melting or fusion of the base metals.

Applications of ultrasonic welding are extensive, finding use in the electrical/ electronic, automotive, aerospace, appliance and medical products industries, as examples. Although nearly all metals can be welded with ultrasonics, the widest current uses typically involve various alloys of copper, aluminum, magnesium and related softer metal alloys, including gold and silver. A number of dissimilar metal combinations (e. g. copper and nickel) are readily welded with ultrasonics. It is used to produce joints between metal plates, sheets, foils, wires, ribbons and opposing flat surfaces. Joining several stranded copper cables into a single junction, used in automotive wire ‘harnesses’ is one common use. Ultrasonic welding is finding increasing applications for structural components in the automotive and aerospace industries. It is very useful for encapsulating temperature sensitive chemical or pyrotechnic materials. The closely related process of ultrasonic microbonding is widely used in the semiconductor and microelectronics industries.

There are several variations of the welding process. The most common, ultrasonic spot welding, uses two different configurations of equipment,

known as lateral drive and wedge-reed welding systems. Ultrasonic seam welding, used to obtain continuous welds in thin gage materials, and ultrasonic torsion welding, used for circular closure welds and stud welds, are added variations of the ultrasonic process, as is the previously noted ultrasonic microbonding. A number of other laboratory or experimental configurations of ultrasonic welding systems have been developed, including a means of producing ultrasonic butt welds.

This overview of the ultrasonic metal welding process will:

• Outline the principles of ultrasonic metal welding;

• Describe the features of several types of ultrasonic welding equipment;

• Summarize information on the mechanics and metallurgy of the ultrasonic weld;

• Review a number of applications of the process;

• Summarize advantages and disadvantages of the process;

• Examine future trends;

• Provide references for further study of the process.

9.2 Principles of ultrasonic metal welding

It was noted that ultrasonic welding uses high frequency mechanical vibrations to create a friction-like relative motion between two surfaces, causing deformation and shearing of surface asperities that disperse oxides and contaminants. This process brings metal-to-metal contact and bonding between the surfaces. Its features may be illustrated by considering a typical welder, known as a lateral drive system, as shown in Fig. 9.1.

The key elements of the system are the transducer, booster and sonotrode series of components, which produce and transmit the ultrasonic vibrations to the workpieces clamped between the sonotrode welding tip and a rigid anvil. The static force between the anvil and sonotrode tip is produced by a coupling moment created within the system enclosure by an arrangement of leveraged forces and pivots (details not shown in this simplified schematic).

The high frequency (e. g. 20 kHz) vibrations are produced in the transducer via a high frequency input voltage from the electronic power supply.

Since ultrasonic vibrations are the basis of this welding process, a basic understanding of the underlying vibrational behavior of the transducer-booster- sonotrode-system is important. This requires starting with the most important component, the ultrasonic transducer, shown in Fig. 9.2. Thus, from Fig. 9.2, we see that:

• The main components of the transducer are a front driver (usually aluminum or titanium), several piezoelectric disks (in pairs of two, four or six - sometimes up to eight), and a rear driver (usually steel), with this assembly held together by a bolt that serves to precompress the ceramics. The piezoelectric properties of the disks result in conversion of the high frequency electrical signal from the power supply to mechanical vibrations of the transducer. Between the disks are thin steel, copper or nickel electrodes which connect the disks to the external power supply.

• The transducer assembly vibrates, in a longitudinal direction (i. e. along the axis of the transducer, as shown by the arrow) at a resonant frequency determined by the dimensions of the drivers and ceramics. The distribution of vibration is shown by the curve labeled ‘vibration half wavelength’, indicating that the maximum amplitude is at the front end of the transducer. There is a point of minimum vibration, known as the ‘node’ which is usually designed to be located at the flange, where the transducer enclosure case may be attached, as shown in Fig. 9.2.

• The dimensions and materials of the various transducer components are selected so that the device vibrates, or more specifically ‘resonates’ at a specific frequency. An operating frequency of 20 kHz is commonly used in many systems, but transducer frequencies as low as 15 kHz or as high as 60 kHz may be found in metal welding systems (and yet higher, e. g. 120 kHz to 300 kHz, in electronic microbonding systems). An important

9.2 Ultrasonic transducer: (a) transducer assembly; (b) resonance behavior.

characteristic of the transducer is that its operating frequency (e. g. 20 kHz) is very sharply tuned to that specific frequency, with its vibration amplitude dropping off extremely rapidly just a few hundred hertz from its operating point. This is shown in Fig. 9.2(b), where the amplitude of transducer vibration is plotted against frequency and shows how just a slight shift from its resonant frequency, fr, results in a dramatic reduction in amplitude.

• The amplitude of vibration of the front end of the transducer is quite small, typically in the range of 10-30 mm, peak-to-peak, an invisible amount to the unaided eye. This, combined with the fact that operating frequencies are typically above the audible limit, results in little visible or audible action during a welding cycle, thus being deprived of the spectacular pyrotechnics of arc, resistance and laser processes.

The three components of transducer, booster and sonotrode, are shown assembled in Fig. 9.3 (the assembly is by threaded fasteners at the component interfaces). While each differs in shape from the transducer, the booster and sonotrode are in longitudinal vibration, tuned to the same frequency as the transducer, and with a node in the middle region of each. Each is operating at a half acoustical wavelength with the result that the overall assembly is 3 x 1/2 = 1+1/2 (1.5) wavelengths long. The exact value of the acoustic wavelength depends on the frequency, the materials of the different parts and their geometric shapes. To an approximation, at 20 kHz, the acoustic wavelength in several materials used in practical systems (e. g. steel, aluminum, titanium) is 25 cm, so that the length of the system shown in Fig. 9.3 would be of the order of

1.5 x 25 cm = 38 cm. The booster and sonotrode are shaped to amplify the vibrations of the transducer, so that as one progresses down the length of the assembly, the amplitude is first increased by the booster and then again by the sonotrode. The result is that an amplitude at the transducer, possibly of 20 mm, may be increased to as high as 100 mm at the weld tip on the sonotrode. The welding tip of the sonotrode may be detachable, held in place in a threaded connection, as suggested by Fig. 9.3, or be machined as an integral part of a single piece sonotrode.

The electronic power supply, shown in Fig. 9.1, provides the driving power to the system, converting line frequency to the ultrasonic frequency required by the transducer. Depending on the power rating of the transducer, and the application, the supply may need to provide power levels of tens to thousands of watts. The power supply system also provides additional control and operation functions. The piezoelectric transducers typically have a high level reactive (of a capacitance nature) electrical impedance of up to several hundred ohms which would match poorly with modern electronic power supplies that are designed for low impedance (e. g. 50 W) resistive loads. For this reason, impedance matching circuitry is typically built into the power supply.

Of equal importance is the fact that the sharply tuned resonant frequency of the transducer-booster-sonotrode system will change during system operation, and from one welding application to another. As one example, continued operation of the system will cause heating of the various parts, resulting in a change in system frequency. The welding application serves to impose a mechanical impedance on the transducer system through the welding tip, with this also being capable of changing the system resonant frequency. These and other effects can result in shifts in the sharp resonance curve shown in Fig. 9.2(b) away from the drive frequency and could greatly reduce vibration amplitude. However, all modern welding power supplies use frequency tracking circuitry to compensate for such shifts, so that the power supply frequency will stay tuned on the changing system resonant frequency. In addition to this, some welding power supplies are able to control the vibration amplitude of the transducer during the weld cycle, with this typically being done using concepts based on equivalent circuit representation of the transducer.

A second type of widely used configuration for ultrasonic metal welding is known as a wedge-reed system. Although using different principles to impart ultrasonic vibrations to the workpieces, it ends up achieving the same effect of a frictional-like relative motion at the surfaces of the parts. A wedge-reed system is shown in Fig. 9.4(a). The key elements of this system are the transducer, wedge and reed series of components, which produce and transmit the ultrasonic vibrations to the workpieces clamped between the sonotrode welding tip and an anvil. A static force is applied between the mass at the upper end of the reed and the anvil. As with the lateral drive, the high frequency transducer vibrations are produced by a high frequency input voltage from the electronic power supply.

The different vibration behavior of the wedge-reed system is shown in Fig. 9.4(b). The transducer is of the same form as that in the lateral drive system. The wedge serves the same purpose as the booster in the lateral drive system, acting to increase transducer vibration amplitude. Thus, the transducer - wedge is in longitudinal resonant vibration, as shown by the dashed wavelength

pattern in the figure, and produces vibrations in the direction of the arrow at the tip of the wedge. The wedge is solidly attached to the vertical reed by welding or brazing. The different feature of this system is that the wedge vibration drives the reed in a bending, or flexural, vibration mode. The general nature of the vibration pattern along the reed is shown by the solid line wave pattern. Being in bending vibrations means that the motion of the reed at any point is in a left-right (or transverse) direction, much like the vibrations of a string. This results in a transverse vibration of the welding tip against the workpieces, as shown in Fig. 9.4. Thus, the wedge-reed produces the same vibration effect at the workpieces as in the lateral drive (see Fig. 9.1), but by a slightly different means. Another variation of the lateral drive system involves the anvil, which also is in vibratory bending motion, as shown in the figure, although in some cases, a rigid anvil design may be used. Using a ‘contra-resonant’ design, the anvil may vibrate out of phase with the reed, thus increasing the net transverse motion across the parts.

With this outline of the vibration principles that underlie ultrasonic welding, we now can examine more closely the weld itself, shown in Fig. 9.5(a). It has been noted that both of the preceding welding systems end up applying the same type of transverse vibration from the weld tip to the top surface of the workpiece, as shown in Fig. 9.5. The top part moves in unison with the welding tip. This ‘anchoring’ of the part to the tip is usually assisted by a roughened or knurled surface to the tip, to engage better with the surface. Similarly, the bottom part remains anchored to the anvil, also usually assisted by a patterned anvil surface. As a result, the motion of the welding tip produces a relative motion between the parts at the contact surfaces. This

Static force

Sonotrode weld -►tip

Sonotrode weld tip

(b)

9.5 Ultrasonic metal weld: (a) transverse vibration imparted to workpieces, and weld zone; (b) normal and shear forces acting at weld zone.

relative, transverse motion, between the two opposing surfaces creates the friction-like action. This action, in turn, causes shearing and plastic deformation between asperities of the opposing surfaces, bringing about increasing areas of metal-to-metal contact and solid state bonding between the parts. The actual bonding zone is suggested by the shaded region between the two parts. If the two parts are separated at the faying surfaces, as shown in Fig. 9.5(b), two types of force components will be seen acting at the interface. First is the static clamping force, acting at right angles to the surfaces, and the second is a transverse shearing force, brought about by the friction-like transverse motion of the parts. These forces will be explored in more detail in Section 9.4.

Two unique features of ultrasonic metal welding are worth re-emphasising:

1 The nature of the motion between the parts during the metal welding process is one of transverse oscillation, where the opposing surfaces move parallel to one another. This is in distinct contrast to the allied process of ultrasonic plastic welding, where the opposing surfaces move at right angles to one another.

2 Although there is a local plastically deformed weld zone created between the parts, no melting of the metals occurs in this zone. The bonding is via a solid state bond, versus a fusion bond such as occurs in the weld nugget of a resistance spot weld, or in other fusion-based processes such as arc or laser welding.

Additional features of the overall bonding mechanism of the ultrasonic metal weld will be discussed in Section 9.4.

From the illustrations of the lateral drive and wedge-reed welding systems and of the weld zone it is evident that a number of parameters can affect the welding process. The main ones can be summarized as follows:

• ultrasonic frequency

• vibration amplitude

• static force

• power

• energy

• time

• materials being welded

• part geometry

• tooling.

Not all of these are independent parameters. For example, the energy delivered to the welder is dependent on the power-time relationship, while the vibration amplitude may depend on the power level. These are all noted here separately because different welding systems may be set up to operate with differing sets of independent variables.

9.2.1 Ultrasonic frequency

It has been noted that ultrasonic welding transducers are designed and tuned to operate at a specific frequency, with these frequencies ranging, for different systems and applications, from 15 kHz to 300 kHz. Most metal welding systems will operate at 20, 30 or 40kHz, with the higher frequencies, e. g. 60 or 120 kHz, being used in microbonding work.

It is sometimes asked whether there might be ‘critical frequencies’ from a fundamental metallurgical physics point of view (e. g. exciting dislocation fields) at which the welding process might be optimal. There is no evidence of such material behavior, at least in the frequency regimes of ultrasonic welding. Instead, welding frequency is governed by such matters as welding power requirements, which are governed by part dimensions and materials being welded, and the overall design of transducers and coupling components.

While reference has been made to the ‘single, specific operating frequency’ of a transducer, and the nature of this sharply tuned resonant behavior illustrated in Fig. 9.2(b), in actual operation, there are several factors that act to shift the transducer resonant frequency. Small dimensional changes due to system heating during operation, varying static force, different tooling and changing tool conditions due to wear, and the changing effects of the welding load can all act to cause both long-term, as well as quite rapid, changes in the system resonant frequency.

However, while these shifting conditions of resonance can be complex, the key point from a practical user’s standpoint is that modern power supplies employ sophisticated feedback control circuitry that automatically compensates for these shifting conditions and tracks the driving frequency of the transducer to maintain the system on resonance. Thus the changing conditions of welding effectively become transparent to the machine user, while the system is kept at resonance.

9.2.2 Vibration amplitude

The vibration amplitude of the welding tip at the weld is one of the key parameters affecting welding. Thus, it directly ties in to the energy delivered to the weld zone. It is again pointed out that ultrasonic vibration amplitudes are quite small quantities, being of the order of 10-50 mm at the weld and seldom exceeding 100 mm as a maximum. (The diameter of a human hair is in the range of 75-100 mm.)

In some welding systems, the amplitude is a dependent variable, being related to the power applied to the system. In other systems, the amplitude is an independent variable, capable of being set and controlled at the power supply because of added features of the feedback control system. (Interestingly, the purely mechanical parameter of vibration amplitude is controlled by purely electronic means, using certain equivalent circuit concepts between mechanical and electrical systems.) The selection of weld vibration amplitude will depend on the conditions of welding as governed by materials and tooling.

9.2.3 Static force

The static force is also a key parameter of ultrasonic welding. The force that is exerted on the workpieces via the welding tip and anvil, in pressing the parts firmly together, creates intimate contact between the opposing surfaces as preparation for the ultrasonic vibrations in the weld zone. The magnitude of the force will be strongly dependent on the materials and thicknesses being welded, as well as on the size of weld being produced and may range from tens to thousands of newtons. For example, producing a weld of 40 mm2 in 6000 series aluminums may use forces of the order of 1500N, while 10mm2 welds in 0.5mm thick soft copper sheet may require only 400N. In adjusting system parameters, there is typically found to be an optimum range of static force, below which welds will be weak to non-existent and above which excessive deformation of the parts may occur.

9.2.4 Power, energy, time

While individually listed as separate weld parameters, these are most conveniently examined in a unified manner, since they are all tied closely together. When a weld is made, the voltage and current result in a time- varying electric power flow to the transducer over a period of the weld cycle. This power-time curve can take many forms, depending on material types,

(a) (b)

9.6 Weld power: (a) examples of weld electrical power curves; (b) representative power curve.

dimensions and surface finishes, ultrasonic parameters such as amplitude and static force, and the particular features of a given welding system. Figure 9.6(a) is simply representative of the various forms that power curves can take during welding. However, these many details may be ignored for present purposes, and the simple curve of Fig. 9.6(b) used to illustrate the basic power, energy and time relations that would apply to more complex shapes.

A representative power curve, Fig. 9.6(b), will have a peak power (PP) and weld time (tw). The area under the power curve is the weld energy (joules) or, more specifically, the electrical energy supplied to the transducer during the weld cycle. It is evident that not all three can be independent. Thus, in Fig. 9.6(b) one can set the peak power, with the welder running until that level is reached, with weld energy and time being dependent on reaching a certain level. Or, energy can be set and the weld would run until such as the set level is achieved, and so forth for other variations.

The power delivered to the transducer from the power supply is converted to ultrasonic power at the weld. However, between the electrical input and the weld several conversions and transmission steps intervene. The actual power delivered to the weld zone is dependent on several factors that can include: (a) the efficiency of electromechanical conversion of the electrical input to mechanical output by the piezoelectric materials; (b) losses in the bulk materials and at interfaces of the transducer-booster-sonotrode system; (c) power radiated from the weld into the workpieces and the anvil structure. The power setting may be indicated in terms of the high-frequency power input to the transducer, or the load power (the power dissipated by the transducer-sonotrode-workpiece assembly). Currently, estimates of actual power (and energy) to the weld itself are made by measuring input electrical power and subtracting off estimated system losses determined from no-load system measurements.

The weld time of ultrasonic metal welds, as noted, may be a dependent or independent variable based on the type of welding system being used. By whatever means, however, metal welding times are quite short, typically being well under 1s in duration - thus 0.25 to 0.5 s are common. Longer welding times usually suggest the need to examine and possibly modify system parameters.

9.2.5 Materials

The single category ‘materials’ in fact encompasses a wide range of issues and parameters relating to ultrasonic metal welding. First, of course, is the type of material. As will be noted in Section 9.4, claims have been made that nearly every metal can be ultrasonically welded in some fashion. The properties of the materials, including modulus, yield strength and hardness are a key consideration. Generally speaking, the softer alloy metals, such as aluminum, copper, nickel, magnesium and gold/silver/platinum are most easily welded ultrasonically. With increasing alloy hardness, ultrasonic welding increases in difficulty. The material surface characteristics come next, with these including finish, oxides, coatings and contaminants. Given that the ultrasonic welding process by its very nature involves the transverse vibration of opposing surfaces, held together under pressure, it is evident that the conditions of these surfaces will play an important role.

9.2.6 Part geometry

The shape of the parts to be welded is also important. The dominant feature here is that of part thicknesses. Simply put, the thinner the parts, of whatever material, the better the chances of achieving ultrasonic welds. Increasing part thickness, in particular that of the part in contact with the welding tip, requires larger welding tip areas, increased levels of static force and generally increasing weld powers. Maximum thicknesses that can be achieved will obviously depend on the material being welded and the power levels available from a welder. For example, welding 1-2 mm 5XXX, 6XXX aluminum alloys is achievable with welding systems in the 2.5-3.5kW ranges. Another part geometry factor, becoming increasingly appreciated in welding larger parts, is the overall lateral size (width, breadth) and shape of the parts. The very vibrations that create an ultrasonic weld can, when transmitted away from the weld and into the surrounding part, affect the making of the weld itself, as well as affect previously made welds. Issues of part resonance, which may play little role for small parts, can become a factor in larger parts where dimensions may become of the order of ultrasonic vibration wavelengths. Typically, issues in this area can be accounted for by modifying part dimensions and stiffness, as well as use of supplemental clamping points to dampen vibrations.

The tooling consists of a sonotrode/welding tip that contacts the top surface of the top part, and an anvil that contacts the bottom surface of the bottom part. The tooling serves as part support and to transmit ultrasonic energy and the static force to the parts being welded. In most cases, detachable tool tips and replaceable anvils are used on ultrasonic welders. In some cases, the tool tip is machined as an integral part of a solid sonotrode. While the weld tip and anvil contact surfaces are usually flat, in some cases the weld tip may be designed with a slight convex curvature in order to change the contact stress patterns. The tooling contact surfaces typically have machined knurled patterns of grooves and lands, or other surface roughening, to improve gripping of the workpieces. A wide range of tooling hardnesses, heat treatments and coatings are employed to deal both with the wide range of materials to be welded and the wear conditions that are encountered for varied applications.