Wire bonding is the most prevalent and robust die interconnection method used today. However, even with its popularity and status as the preferred die/lead-frame interconnection, there can be significant differences in the performance of wire bonds. In this month’s Tech Tips, the differences between ball and wedge wire bonding techniques are examined (Figure 4-1).

Ball bonding is typically associated with thermocompression (T/C) and thermosonic (T/S) joining methods. While T/C utilizes pressure and temperature to create a bond, T/S bonding adds ultrasonic energy to the process. During both methods, the end of the bond wire is converted to the ball shape by the application of an electronic flame-off (EFO). The ball is then positioned just above the bond pad on a substrate or package and connected to the bond pad. An intermetallic bond is created by interdiffusion between the wire materials and the pad metallization. The atomic interdiffusion is caused by pressure, temperature, and in the case of thermosonic bonding, ultrasonic energy. In ball-wedge bonding, the most common method of wire bonding, the first bond (source bond) takes the shape of a ball, and the second bond (destination bond) takes the shape of a wedge.

Gold wire balls created from the ball bonding process can also be used to create symmetrical wafer bumps (stud bumps). Using existing wire bond equipment to perform stud bumping is an excellent and cost efficient process for low to medium volume flip chip bumping. The material properties of gold wire provide significant advantages in strength, fatigue resistance, and electrical resistivity when compared to traditional tin-lead and lead-free solder bumps

Wedge bonding utilizes ultrasonic energy and pressure to create a bond between the wire and the bond pad. Wedge bonding is a low-temperature process that uses frequencies between 20 and 60 kHz for standard applications and 120 kHz for fine pitch applications. This cold welding process deforms the wire into the flat elongated shape of a wedge. The most common method of wedge bonding is wedge-wedge bonding, where both the source bond and the destination bond are formed with the geometry of a wedge.

Speed
One of the most notable differences between the two processes is the speed at which the wires can be attached. Ball bonds are considered omni-directional and allow wire placement at any angle with only x-y motions. In ball bonding, the second bond can be formed at an angle about the arc of the ball bond. In wedge-wedge bonding the wire is fed beneath the flat bottom of the tool. The destination bond must therefore be in line with the first bond. The alignment of the wedge-wedge bond results in a reduction in process speeds, which for automated ball bonders are in the 18,000 per hour range. Wedge bonding processing can take up to 2 to 3 times as long as ball bonding of the same chip. Despite improved wedge bonding capability, which reduces processing time, such as rotary bond head which allows radial fanning out of wedge bonding wires, ball bonding remains the most popular and robust system because of its speed.

Pitch
Wedge bonds have a slight advantage in pad spacing capability. Both ball and wedge bonds are capable of meeting the requirements of fine-pitch applications. However, wedge bonding deformation results in a footprint only 25 to 30 percent greater than the original diameter of the wire, so smaller bond areas are required. By contrast, ball bonds require an increase of approximately 60 to 80 percent in original wire diameter. The smaller bond allows for finer pitch bonding pads. Pitches less than 45µm are achievable with wedge bonding, while ball bonders using thin bond wire (< 25µm diameter) can achieve 60 µm spacing. Smaller spacing can be achieved in ball bonds by staggering pads. In both the ball and wedge process, fine pitch ball bonding requires more control over equipment, including high-resolution vision and positioning capability.

Bond wire
The ultrasonic wedge bonding process is typically performed with aluminum alloy wire on either aluminum or gold pad metallizations. Pure aluminum wire is too soft to be drawn into fine wire, so it is often alloyed with small percentages of silicon or magnesium to increase strength. Aluminum wire is not typically used in ball bonding because the heat used during the ball formation process can create oxides that hinder the joining process. Gold and copper wire can also be wedge bonded. Copper wire, desirable because of its high strength and stiffness, can be bonded at room temperature like aluminum wire. Special ceramic wedges allow for gold wire bonding at elevated temperatures, typically 125-150ºC.

Gold wire is commonly used during ball bonding because it easily deforms with pressure at elevated temperatures. Copper wire can also be used, but because gold is resistant to forming oxides at elevated temperatures, it is not as problematic in making a bond. The use of nitrogen as a “blanket” or “forming gas”, prevents copper oxides from forming during the bonding process. Copper is less costly than gold, has superior electrical properties, and is more compatible with copper chip metallizations, which have demonstrated considerable (30-50%) power savings on a digital chip. Because copper is stiffer than gold, the resulting deformation of the bond pad may damage the surface of the chip; therefore, tight processing controls are recommended when bonding copper.

Conclusion
Despite the wedge bond’s distinct advantages in pitch and the ability to process temperature-sensitive substrates and high power applications, the popularity of the ball bonding process continues to exceed that of wedge bonding. With processing speed as one of the main drivers for chip manufacturers, ball bonding will continue to dominate the market for traditional chip applications. Additionally, the ball bonder’s ability to create wafer bumps has helped to secure its position as the principal bonding technology.