[PDF] Three Dimensional Modeling and Characterization for Die Attach

76206_3three_dimensional_modeling_and_characterization_for_die_attach_process.pdf 1 IRU'LH$WWDFK3URFHVV Lin Bu, Wai Leong Ching, Ho Siow Ling, Minwoo Rhee, Yong Puay Fen 1 Abstract: A new three dimensional model for the die attach (DA) process is established and validated in the present study. With this model, the fluid flow characteristics of the DA process can be predicted accurately. Dynamic mesh and interface tracking method were adopted in the modeling to study the compression motion and the front of DA. Force driven model was conducted for the parametric studies of different bonding force. The model for the DA process was validated by the four materials, AP1, CA1, CA4 and DM60 in the optimized condition. Bond line thickness (BLT) can be predicted by simulation with ~20% accuracy. The simulation results show that viscosity is one of the key properties, which has a significant effect on the required bonding force, bonding time and DA contamination on the die top. Complete filling and DA contamination on the die top are two important standards to evaluate the good bonding force range in fluid dynamic analysis. Stress analysis illuminates that fillet area is very critical and experiences highest stress during the reflow process. Index Terms die attach process, bond line thickness (BLT), bonding force

I. INTRODUCTION

ie attach provides the mechanical support between the silicon die and the substrate, i.e,. leadframe, plastic or ceramic substrate. The die attach is also critical to the thermal and, for some applications, the electrical performance of the device. Significant results have been achieved in previous studies, focusing on mechanical analysis for the DA process. Dynamic mechanical analysis (DMA) was employed to characterize the modulus behavior of silver filled glass material. The method used a simulated DA process to understand the behavior of the storage modulus and the complex viscosity [1]. DMA taking into account multi-step curing was utilized to determine gelation times and melt viscosity under a shear mode by Taweeplengsangsuke J. [2]. They found that the longer the period of time at the lower temperature step of the 2-step curing gave rise to lower cure stress. In addition, the stress during the cool down process was investigated. At the point of decreasing temperature, the stress dramatically increases. The higher temperature difference, the larger the residual stress. Xiaosong M. et al. [3] developed a finite element model to predict the interface delamination

1The authors are with the Institute of Microelectronics, Agency for

Science, Technology and Research, 117685, Singapore. Corresponding author is Min Woo Rhee. (e-mail: bul@ime.a- star.edu.sg; wailc@ime.a-star.edu.sg; hosl@ime.a-star.edu.sg; mw.daniel.lee@gmail.com, sherylyong@ntu.edu.sg). issues encountered in the DA process. It is found that temperature has a large effect on the interface toughness (Gc). Gc greatly decreases with increasing temperature. In addition, moisture has no effects on interface toughness of copper and silver filled DA in their samples. Khoo Ly Hoon [4] et al. aimed to establish a robust DA process by evaluating various responses on DA epoxy with various values of epoxy viscosity. The findings of their study reveal that the epoxy viscosity within the tested range does not significantly affect DA and wire bond performance. Nicolas Heuck et al.[5] investigated the impact of the CTE of die-attach layers on the thermal stress in chip and attach layer, along with strategies to reduce the CTE of conventional silver sintered die-attach structures by adding materials like SiC or h-BN. They found, that the implementation of SiC and especially of h-BN

ȝ-attach layers can

lead to a stress reduction up to 30%. However, rheological simulations that can investigate the flow of the liquid DA under an applied force during the DA process are rare. In order to enhance the evenness of epoxy distribution along the peripheral of the die, Mark Lee [6] built a 2D model to investigate various dispensing patterns and to study their evolvement patterns. The thin-film assumption is used in the simulations. The studies shows that a suitable epoxy pattern is the key to ensure that the epoxy dispensed on the substrate can evolve to the final shape of the chip after the initial squeezing during the DA process. Complex dispensing patterns, i.e. snow star pattern are more likely to trap voids than basic dispensing pattern, i.e. x dispensing pattern. However, 2D simulation cannot capture the 3D real process very well and many important properties like surface tension and contact angle are not incorporated into the model. Important information such as BLT and contaminations on the die top in the DA process could not be obtained due to 2D constraint. The prediction of final BLT and die top contaminations are as important as epoxy dispensing patterns. Final BLT would have a significant effect on the reliability of the whole system. Man Wai Chan [7] et al. invented a way to measure BLT with laser equipment. They can also control BLT in a desired range by adjusting the four parameters, 1) adhesive dispensing pressure of the dispenser, 2) bond level of the bonding tool, 3) bond force exerted by the bonding tool and 4) bond delay of the bonding tool. In the present study, rheological simulations were carried out using the three dimensional model. Surface tension model and wall adhesion model are enabled to take into account the effect of surface tension along the interface of two fluids and the contact angle that the fluid makes with the wall. Appropriate bonding force and BLT ranges for the optimization of DA process can also be predicted by the present model. D 2

II. PACKAGE DESCRIPTION AND CRITICAL ISSUES

DURING DA PROCESS

Institute of Microelectronics at A*STAR Singapore and its industry partners have developed a DA technology, based on a

5×5 mm2 top chip. Fig. 1 demonstrates a schematic plot of DA

process.

Fig. 1. Schematic plot of DA process

The DA process consists of three stages. First stage is to dispense an adhesive with a dispenser onto the substrate. Then, in the second stage, a semiconductor die is picked and placed on the adhesive which has been dispensed onto the substrate with a bonding tool with vacuum. Thereafter, in the third stage BLT between the bottom surface of the semiconductor die and top surface of substrate on the process platform using a measuring device. (a) Overflow (b) Incomplete fill Fig. 2. Schematic plot of DA issues. (a) Overflow. (b) Incomplete fill. During the second stage, void issue, bad fillet and contamination on the die top are the key issues. Epoxy climbing along the edge of the die will lead to the formation of DA fillet. Excessive DA fillet can lead to DA contamination of the die surface. Too little of it may lead to die lifting and die cracking. These two issues are demonstrated in Fig.2. Voids maybe trapped by using inappropriate dispensing patterns. Voids in the epoxy not only increase thermal and electrical resistance, but also trigger electrical breakdown at extreme conditions. Overcoming these three issues is a balance of controlling dispensing pattern, bonding force and bonding time.

III. GOVERNING EQUATIONS

The equation for the conservation of mass, or continuity equation, shared by the two phases, can be written as follows [8], ࣔ࣋