Laser-Generated Ultrasound for Thermosonic Bonding

The aim of this project is to explore the use of laser-generated ultrasound in thermosonic (TS) bonding. TS bonding is a joining technique which uses a combination of heat, pressure and ultrasonic energy to facilitate the formation of strong metal-metal bonds. It is used mainly for attaching bond wires to silicon chips inside their packages, where it offers a number of advantages over other joining methods. For example, it involves no additional materials (e.g. solders or adhesives), and it can be carried out at lower temperature and pressure than thermo-compression bonding and lower ultrasonic power than pure ultrasonic welding.

An important potential application for TS bonding is flip chip assembly, a technique used in advanced electronics manufacturing. Flip chip allows unpackaged integrated circuits to be attached to a circuit board or other substrate in a face-down configuration, with electrical connections between the contact pads on the chip and the substrate being provided by conducting "bumps". Flip chip assembly offers several advantages over other chip attachment methods, such as higher electrical performance, higher interconnect density (more electrical connections per unit area), smaller footprint and lower height.

Flip chip processes based on solder attachment have been established for many years. However, with the continual drive for miniaturization they are approaching their limits in terms of interconnect density. Alternative approaches based on adhesive bonding are scalable to finer interconnect pitches, but do not achieve the performance or reliability required for many applications. TS bonding could form the basis of a highly reliable, ultra-fine-pitch flip chip technology. However, up to now it has proved challenging to develop robust processes, mainly because it is highly sensitive to co-planarity errors and bump height variations which can lead to bond strength non-uniformity and even damage to the chip. These issues become more severe as the chip size increases, and consequently TS flip chip has been limited to a narrow range of applications involving small devices with low interconnect count.

We propose to develop a TS bonding process in which pulsed laser light is used to generate ultrasound locally at specific bonding sites, using confined ablation of a sacrificial carrier tape sandwiched between the workpiece and a transparent bond head. This approach will enable us to deliver the ultrasonic energy in a flexible manner, allowing for the possibility of compensating for co-planarity and bump height errors. With the proposed system it will also be possible to pre-heat the interface locally by laser, yielding a process with very low overall thermal loading.

If successful, the proposed research will ultimately lead to a next generation flip chip technology with wide ranging applications in electronics manufacturing. The new process should also find applications in other fields such as MEMS (microelectromechanical systems) and optoelectronics where joining of delicate components is required.

The proposed research, if successful, will significantly expand the range of potential applications for thermosonic bonding. The main beneficiaries will be laser manufacturers, laser system integrators, and companies involved in electronics manufacturing, in particular manufacturers of bonding tools, contract electronics manufacturers, and electronic product manufacturers. Companies in these sectors could begin to benefit from the proposed research in the next 5-7 years.

Manufacturers of OEM laser products and laser system integrators will benefit from the emergence of a new industrial application for high-power pulsed lasers, while companies in the electronics manufacturing sector will benefit from the development of a new flip chip technology that is both highly reliable and scalable to finer interconnect pitches than solder-based processes. Such a technology will enable the development of new electronic products that have greater functionality for a given size. This is particularly relevant to portable consumer products, such as mobile 'phones, PDAs and music players, where there is a constant drive to reduce the size and weight while at the same time increasing functionality. Miniaturization is also a key driver for medical devices because it reduces the trauma associated with surgical procedures to implant intra-corporeal devices, and makes wearable devices such as hearing aids and vital signs monitors more comfortable for the patient. With ongoing developments in low-power electronics, and the emergence of viable techniques for wireless power delivery and energy harvesting that can reduce the size of batteries (or eliminate batteries altogether), current surface mount assembly methods will before long become a limiting factor in miniaturization.

The proposed technology will be unique in its ability to deliver highly localized ultrasonic power to a bonding interface. Because of this it will be possible to keep the total ultrasonic power relatively low, making the process attractive for applications such as assembly of hybrid MEMS (microelectromechanical systems) where handling of delicate micromachined parts is required. Also, because of its low overall thermal load, the process is also likely to have impact on applications requiring integration of components onto polymer substrates (e.g. RFIDs, plastic displays).

In addition to having an impact on quality of life through the enhancement of electronic products, the proposed research will have a positive environmental impact. In recent years there has been a global effort to eliminate lead from electronics manufacturing, and this has been achieved mainly by development of lead-free solders. However, a disadvantage of these materials is that they require higher reflow temperatures and hence increase the thermal budget of assembly processes. This in turn increases power consumption and contributes to pollution. In contrast, thermosonic bonding processes have a lower thermal budget than even lead-based soldering.