- 上一篇:underfill到底有何用？？？
- 下一篇:Underfill后导致手机耗电流增大
Adhesion and Debonding of Underfill to SiNx passivation
Motivation:
The overwhelming trend in the microelectronics industry today is to strive toward smaller feature dimensions on each chip. With the increased number of devices on one integrated circuit, there is a need for packaging that can accommodate the growing population of interconnects required to attach these devices to the circuit board. One effort to compact the chip package is to use a high-density solder ball grid array flip-chip configuration as shown in Figure 1.
Figure 1: Schematic diagram of high density flip-chip packaging.
With the chip lying face down on the substrate, solder balls provide direct interconnections between them. Thermal mismatch, one of the problems that is prevalent in IC processing, can cause large amounts of stress between the solder and the silicon. In fact, fatigue within the solder balls is often the first cause of failure in these packages. In order to aid in supporting the stresses that arise from constant mechanical, electrical and thermal cycling, a polymeric underfill is dispensed to surround the solder balls. To increase the strength and further equalize thermal expansion throughout the package, silica beads are dispersed throughout the underfill. During processing, these added beads tend to settle to the bottom interface of the underfill before the epoxy sets. This causes a gradation in mechanical properties within the underfill that can alter the failure mechanism in the structure. In addition, the polymer matrix is susceptible to moisture absorption, which can provide an additional driving force for underfill delamination from the passivation, the solder balls, or the silica beads.
Objective:
The purpose of this work is to determine the fundamental mechanisms causing adhesion and debonding at interfaces in microelectronic packaging. Samples are made by sandwiching underfill between two stiff silicon beams. These rigid substrates constrain the plasticity in the system to the layer thickness, allowing the use of linear elastic fracture mechanics to describe the deformation. Because of this constraint, the resulting adhesion values only reflect the energy required to cause debonding.
Two model underfills are of interest, one based on a bisphenol F epoxy and one based on an aliphatic epoxy. By using model systems with known compositions similar to those used in industry, the salient parameters that influence failure can be extracted. For this research, these epoxies are sandwiched between silicon beams passivated with silicon nitride. Examples of the sample cross-sections are shown in figure 2. These samples contain 20 wt% silica bead filler, and the extent of settling along the bottom interface can be seen.
a.)
b.)
Figure 2: SEM micrographs showing the cross-sections of two samples of underfill with 20wt% silica filler sandwiched between silicon beams with Si3N4 passivation: a) Aliphatic epoxy, b) Bisphenol F epoxy.
Adhesion to SiNx for each underfill is being studied with the amount of silica filler ranging from 0-70 wt%. The effect of environment is also being examined for both epoxy systems, specifically the effects of temperature and humidity.
Initial Results:
The bisphenol F system has provided the most surprising results in this research to date. Past work has shown that the effect of silica bead settling leads to lower adhesion values. In the bisphenol epoxy, however, the addition of silica filler appears to increase interface toughness. In addition, bisphenol F shows limited susceptibility to moisture-induced debonding, which has great implications in the development of moisture-resistant interfaces.
This research is funded by Semiconductor Research Corporation (SRC) Task 729.001.
