Rheology of the underfill flow process in a flip chip package

Abstract

Underfills that flow have traditionally shown superior reliability performance when compared to "No-flow" Underfills. This has ensured the necessity for "flowable" Underfills in the IC market particularly for microprocessors and other high performance chips. The speed of flow has been an issue in the industry. Faster flow would reduce cycle time, however, this should not be at the expense of reliability performance of the package. Schwiebert and Leong established the basic theory of the Underfill flow. Other authors have modified this with varying degrees of success, but in all these models the effect of the change in the Underfill viscosity during the flow process was not considered. This has rendered the theories incomplete and has been in pah the cause for errors in modelling the flow. In the first half of this study the Schwiebert and Leong model has been expanded to include the influence of viscosity change in the Underfill at the flow temperature. The resulting derivation is a more general form of the flow equation as compared to that of Schwiebert and Leong. The paper shows that the new equation can be reduced to the Schwiebert and Leong equation when the Underfill viscosity is considered to be constant. Experimental verification was done by first characterizing the viscosity change over time of three different Underfills. The flow distances versus time; the change in the rheology over the flow temperatures and time; and the change in 6he contact angle of the Underfill to a laminate substrate and polyimide die are also measured over time and temperature. The effect of the change in the contact angle over time at the Underfill flow temperatures and other parameters that influence Underfill flow are discussed These experimental results were then used to compare with the assumptions made in the theoretical prediction of the Underfill flow. Finally, the experimental flow results are compared with the theoretical model that was developed.