Ultrasonic frequency in wire bonding

The frequency at which the transducer (the ultrasonic transducer) is excited in the bonding process is generally referred to as the bonding frequency. First of all, it is important to know that this bonding frequency is not fixed. The ultrasonic generator constantly adjusts the frequency during the bonding process and thus reacts to effects caused, for example, by the friction processes in the bond interface, the increasing deformation of the wire material, the sinking of the bond tool into the wire or the resonance effects of the bond pads. Depending on the wire bonding process, several possible bonding frequencies are available to the user. In heavy wire bonding, for example, these are 60 and 80 kHz. Occasionally, however, there are also heavy wire bonding processes that are set up with 40, 90 or even 120 kHz. Why is this so?

Three effects can be exploited when changing the bond frequency:

• Avoidance of critical resonances
• Reduction of the energy loss due to vibration of the bond pads
• Better control of the deformation behavior of the wire material (US softening)

As a result of a higher bonding frequency, the vibration amplitude decreases. This impacts the energy input into the bonding wire and the bonding interface. This effect may have to be compensated by an adapted tooldesign. Basically, the higher number of oscillations per second means that the process is faster. At the same time, the process is much more sensitive to disturbances. The process window becomes smaller at higher frequency - at least that is the general opinion. Whether the change in bonding frequency has any noticeable effect at all depends in the end on numerous factors, such as wire diameter, wire material, tooldesign, ultrasonic generator performance, amount of interfering factors superimposed on the effect, and much more. It is quite an experimental challenge to separate these effects cleanly.

In the past, I have experienced the determination of a bonding frequency predominantly as a trial-and-error process. A frequency was selected based on a recommendation or specification from the manufacturer. This specification is made, for example, on the basis of the bonding tool to be used (length, material, design) or is based on an empirical value. Afterwards, it is checked whether the process runs stably. If the process appears to be unstable, the first step is to check if there are other contributing factors. These include contamination, bonding pad and assembly design, component clamping, bonding parameters and material selection. Only after these factors have been checked the bonding frequency is changed. The reason for this is very simple. The bonding frequency can only be changed by replacing the transducer. This requires a modification of the bonding head and then the optimization of the bonding parameters must be carried out again. This is a time-consuming process.

Unfortunately, I do not know of a way to make this approach effective. In some applications, it is possible to perform orienting simulations and modal analyses. But the exact confirmation can only be achieved by bonding tests. These tests are time-consuming and there are also a large number of other influencing factors. It is very difficult for a beginner in bonding technology to work on this topic. Support is available from the bonding device manufacturer or specialists with the appropriate measurement technology.

By the way, you can check and measure a transducer very easily with appropriate test systems. Changes in its vibration behavior can thus be detected very reliably. This is important, for example, if your bonding process is currently not stable and you want to be certain that there is no hardware defect. One system is offered by the company F&K Physiktechnik (LINK). In addition to the mere electrical properties of the transducer, the movement of the tool tip can also be monitored with the help of a vibration measurement system (LINK). Such a system is useful for e.g.:

• Regular (e.g. weekly) control of the transducer function
• Measurement of bond tools before installation in the wire bonder
• Quick check of the transducer in case of bonding problems

Technical papers & articles

So far, I am not aware of any article that deals exhaustively with the subject for all bonding methods. But there are some interesting individual publications from which you can put together a good picture. Some of the articles are freely available, others are unfortunately behind paywalls. I wish you exciting insights while reading.

• Which Frequency is Best for Wirebonding? (LINK) (PDF LINK)
• Effects of Frequency and Surface Cleanliness of Al–Si Electrode on Ultrasonic Bonding Characteristics of Thick Al Wire Bonding (LINK)
• Ultrasonic wire bonding using high frequency 330, 600 kHz and complex vibration 190 kHz welding systems (LINK)
• Effects of different working frequencies on the joint formation in copper wire bonding (LINK)
• How to deal with resonances in wire bonding (PDF LINK)
• Advances in Wire Bonding Technology to Overcome Resonance Conditions (LINK)

Successful wire bonding!

Stefan Schmitz

PS: I tend to prefer the following bond frequencies. In my experience, the contact deformation is leaner and more defined at higher frequencies.

• Wedge/Wedge AlSi1 140 kHz
• Ball/Wedge Au 120 kHz
• Wedge/Wedge heavy wire Al 80 kHz

#wirebond #wirebonding #drahtbonden