Search results

The purpose of this paper is to investigate the reliability of wire bonds with three varying ball bond diameters, which are ball bonded with three different sizes of gold wires in light-emitting diode (LED) package under high-temperature environment. In automotive applications, “lifted ball bond” issue is a potential critical point for LED device reliability, as the wire bonds are usually stressed under high operating temperature during their lifetime. Moreover, the reliability of wire bonds in recent LED production has fallen under scrutiny due to the practice of reducing wire diameters to cut down production costs.

Three gold wires with sizes of 2, 1.5 and 1 mm were ball bonded on the LED chip bond pad via thermosonic wire bonding method to produce three different ball bond diameters, that is, 140, 120 and 100 μm, respectively. The reliability of these wire bond samples was then studied by performing isothermal aging at 200°C for the time interval of 30, 100 and 500 hours. To validate hypotheses based on the experimental data, COMSOL Multiphysics simulation was also applied to study the thermal stress distribution of wire bond under an elevated temperature.

Experimental results show that the interfacial adhesion of wire bond degrades significantly after aging at 200°C for 500 hours, and the rate of interfacial degradation was found to be more rapid in the wire bond with smaller ball bond diameter. Experimental results also show that ball bonds randomly elongate along an axis and deforms into elliptical shapes after isothermal aging, and ball bonds with smaller diameters develop more obvious elongations. This observation has not been reported in any previous studies. Simulation results show that higher thermal stress is induced in the wire bond with the decrease of ball bond diameter.

The reliability study of this paper provides measurements and explanation on the effects of wire diameter downsizing in wire bonds for automotive application. This is applicable as a reliability reference for industries who intend to reduce their production costs. Other than that, the analysis method of thermal stresses using COMSOL Multiphysics simulations can be extended by other COMSOL Multiphysics users in the future.

To resolve “lifted ball bond” issue, optimization of the bond pad surface quality and the wire bond parameter has been studied and reported in many studies, but the influence of ball bond diameter on wire bond reliability is rarely focused. Moreover, the observation of ball bonds randomly elongate and deform more into elliptical shape, and ball bond with smaller diameter has the highest elongation after isothermal aging also still has not been reported in any previous studies.

The purpose of this paper is to investigate the efficiencies of argon (Ar), oxygen (O₂) and O₂ followed by Ar (O₂→Ar) plasma treatments in terms of contaminant removal and wire bond interfacial adhesion improvement. The aim of this study is to resolve the “lifted ball bond” issue, which is one of the critical reliability checkpoints for light emitting diodes (LEDs) in automotive applications.

Ar, O₂ and O₂→Ar plasma treatments were applied to LED chip bond pad prior to wire bonding process with different treatment durations. Various surface characterization methods and contact angle measurement were then used to characterize the surface properties of these chip bond pads. To validate the improvements of Ar, O₂ and O₂→Ar plasma treatments to the wire bond interfacial adhesion, the chip bond pads were wire bonded and examined with a ball shear test. Moreover, the contact resistance of the wire bond interfaces was also measured by using four-point probe electrical measurements to complement the interfacial adhesion validation.

Surface characterization results show that O₂→Ar plasma treatment was able to remove the contaminant while maintaining relatively low oxygen impurity content on the bond pad surface after the treatment and was more effective as compared with the O₂ and Ar plasma treatments. However, O₂→Ar plasma treatment also simultaneously reduced high-polarity bonds on the chip bond pad, leading to a lower surface free energy than that with the O₂ plasma treatment. Ball shear test and contact resistance results showed that wire bond interfacial adhesion improvement after the O₂→Ar plasma treatment is lower than that with the O₂ plasma treatment, although it has the highest efficiency in surface contaminant removal.

To resolve “lifted ball bond” issue, optimization of plasma gas composition ratios and parameters for respective Ar and O₂ plasma treatments has been widely reported in many literatures; however, the O₂→Ar plasma treatment is still rarely focused. Moreover, the observation that wire bond interfacial adhesion improvement after O₂→Ar plasma treatment is lower than that with the O₂ plasma treatment although it has the highest efficiency in surface contaminant removal also has not been reported on similar studies elsewhere.