Study on thermal conductivity model of polymer composite material for next-generation semiconductor package

Abstract

Recently, with the advent of the 4th industrial revolution and the hyper technologies, there has been a growing interest in the next-generation semiconductor device that can reliably compute and process the explosively increasing data. One proposed solution is the three-dimensional semiconductor package, which provides higher performance than conventional devices by vertically stacking chips. However, as more chips are stacked, the temperature of the device rises, posing new challenges in thermal management. Considering the heat path, EMC the encapsulant is the essential component that heat must pass, with its low thermal conductivity leads to bottleneck issues during the heat dissipation process. Improving the thermal conductivity of EMC, nonlinear increase of the property according to the filler fraction has not been fully explained, limiting the pace of advancements in this field.

In this study, a new kinetic model based on the Hill-Langmuir equation, a biochemical model, to describe the polymer-filler network system was designed. EMCs with high thermal conductivity filler such as Al2O3 and AlN showed a combination of two sigmoid curves with different n values, n=1.3 & 6.8, n=1.5 & n=8.1, respectively. The percolation thresholds were represented as the intersection points of the two sigmoid, measuring 31.6vol% and 41.5vol%, respectively.

Furthermore, the model was analyzed to identify ways to enhance the thermal conductivity of EMC. To achieve improved thermal conductivity at the certain filler volume fraction, thermal diffusivity curve with a larger n value and a smaller percolation threshold value is required. While n is a property proportional to the material's number density, the percolation threshold involves both morphological and chemical reaction terms, suggesting that controlling chemical reactions could enhance thermal conductivity.