In - Situ Calibration Of Wire Bonder Ultrasonic System Using Integrated Microsensor

Based on investigations with an ESEC thermosonic wire bonder, we report a novel ultrasound calibration method which consists of measuring in-situ the friction heat generated by the gold ball as it scrubs on the oxide of an integrated microsensor. This procedure can be carried out in-process using a dedicated testchip and a modiÞed clamping plate. The calibration dependence on the process parameters bond force, ultrasonic energy, and initial ball diameter is quantiÞed. The testchip-to-testchip standard variation of the calibration signal taken from a single wire bonder corresponds to a horn amplitude variation of 0.01 µm. Introduction In the process of improving the quality of ultrasonic wire bonding in electronic packaging, the ultrasound level of the wire bonder is one of the most signiÞcant machine parameters to be optimized. Subjected to an alternating voltage, a stack of piezo-ceramics (transducer) produces ultrasonic vibrations which are ampliÞed by the horn of the bond arm. A bonding tool (capillary) is clamped at the end of the horn. It transfers the ultrasonic energy to the interface between the deformed gold ball and the aluminum bonding pad. In such a way, a bond can be made in short time even at room temperature [1]. Transducer impedance and efÞciency vary from one device to another within certain speciÞcation limits due to mechanical differences or capillary change. Transducer and horn characteristics can drift with time [2, 3] and regular calibration is necessary. For a given transducer input, different uncalibrated wire bonders can show more than 30% variation of the horn tip excursion amplitude [4]. This in turn affects the ultrasonic power delivered to the bond and therefore causes variations of crucial process responses like ball geometry and shear strength. This can affect the process feasibility especially for Þne pitch or ball grid array (BGA) applications. Machine-to-machine variations can be handled in several ways. The best method is to adjust the bond process parameters for each machine separately, until the process responses are within the speciÞcations. This procedure requires off-line inspection with appropriate optical and other equipment and is unpractical when many machines have to be calibrated for the same process. Then, calibration of the ultrasonic unit to a reference output is worthwhile. A common calibration method consists of adjusting the vibration amplitude at the horn tip to a desired value by laser interferometry [4] or another optical method [5]. The measurement is usually performed at the horn tip, …