Concurrent Optimization of Crescent Bond Pull Force and Tail Breaking Force in a Thermosonic Cu Wire Bonding Process

In conventional wire bonding process optimization, the crescent bond is tested by destructive pulling the loop and measure the pull force (PF) required to break the bond. While this method assures the final quality, it does not necessarily minimize production stoppages which can reduce throughput significantly. Many stoppages are caused by short-tail and tail-lift errors which in turn are caused by reduced tail bond strength. The tail breaking force (TBF) is a measure for tail bond strength. A consistent tail breaking operation is needed for robust Cu wire bond production with little operator assistance required to restart stopped machines. We report the concurrent (simultaneous) optimization of PF and TBF using standard pull testing and a method that directly measures the tail bond strength in-process, respectively. The example process uses standard 25-$\mu$m-diameter Cu wire on standard Ag-plated leadframe diepads and wire loops oriented perpendicular to the ultrasonic horn. The optimization consists of 1) finding the factors to obtain a symmetrical bond shape, 2) choosing fixed values for bond time (25 ms) and heater stage temperature (220°C), and 3) adding the new optimization step of maximizing the TBF by iteratively optimizing the impact force (IF), bonding force (BF), and ultrasound (US) parameters. Among these parameters, the US and BF are the most significant. A process window (PW) is defined as the set of US/BF parameter combinations that result in a response being inside a previously defined range. PWs for PF and TBF are determined and compared with each other. There is only a partial overlap of these PWs. To increase the process capability index (cpk) for TBF, it is recommended to first carry out conventional PF optimization followed by a minimization of the bonding force parameter to the lowest value still fulfilling the PF cpk requirement.