The three-dimensional numerical simulation has been performed to study mold flow characteristics during injection molding process of stacked die packages. The modeling results revealed that flow front shape is highly non-uniform around the die stack-up units and is dependent on various design parameters including mold cap clearance, mold compound material properties, as well as the die stack-up geometric configurations. It was demonstrated that validated flow model can help guiding the design, material, and process optimization of 3D packaging development.
IntroductionAs the demand for speed and functionality increases for electronic handheld devices such as cell phones and PDAs, a stacked die package becomes one of the preferred options to provide the high performance and multi-functionality for a given electronic component. While the injection molding process is widely used for the microchip encapsulation, the various design, materials and process parameters will need to be further optimized for defect free process due to the increased level of complexity in package architecture.There have been studies of mold filling analysis in microchip encapsulation, however, focuses were given to the single die, not to 3D packaging such as staked-die packages [1][2][3][4]. Therefore, in this study, a three dimensional computational fluid dynamics (CFD) analysis has been performed on a stacked-die package, using a commercially available injection-molding software package, Moldflow, to establish the understanding of the mold flow characteristics in the complicated geometry during the encapsulation process on the effect of different design parameters.
Numerical Model DescriptionThe mass, momentum, and energy conservations for mold flow in the filling phase can be mathematically described by equations (1) through (3), respectively [2,5].-P+V(pv)0=0 (1) at (, +V(pvv)=-Vp+Vr+pg (2) OT (,cupat VT ,8'at i+y'+kVT+H at where iv is the velocity vector, p the density, 77 the viscosity, p the static pressure, r the stress tensor, and y the shear rate. Also the first term in the right hand side in equation (3) is the energy generation rate due to compression/expansion of the polymer melt followed by the energy generation rate due to viscous dissipation. The last term in equation (3) is the exothermic heat generation rate due to curing, where a is the degree of cure and H is the total heat of the cross linking reaction.In this simulation, the curing reaction is described with Kamal's autocatalytic model as shown in equation (4). (4) 3r= (K1 +K2ajXI-a)' K is the Arrhenius equation constant; m and n are the reaction orders. The temperature Arrhenius equation constant K is commonly described by equation (5).Ki Ai exp RT (5) Ai is the pre-exponential factor; Ei is the activation energy; R is the universal gas constant.For the non-Newtonian viscosity behavior of the polymer melt, the Cross-Exponential Macosko model [6] is used as described by following equation.