Multiscale modelling of sub-microchannel heat sinks for microprocessor cooling

The understanding of flow characteristics and heat transfer process through micro- and nanochannels has gained significant importance due to their effective thermal performance for MEMS-based applications. Due to the limitations of Computational Fluid Dynamics at higher Knudsen Number, Molecular Dynamics simulations act as a more suitable approach for capturing nanoscale flow. Therefore, this study is conducted to investigate heat dissipation from a sub-microchannel heat sink using a novel multi-scale approach. Seven scaled-down geometries are modelled, with the smallest channel length and height reaching 0.2 and 0.1 µm, respectively, which demonstrate how scaling down the channel influences dimensionless parameters; i.e. Reynolds Number, Knudsen Number and Nusselt Number. The characteristics of velocity and temperature profiles are also analysed, revealing vortex formation at sub-micro scale, which cannot be fully captured by conventional CFD due to the absence of interatomic interactions. The findings reveal a notable difference between MD and CFD outcomes, highlighting that flow and heat transfer properties are highly attributed to slip effects, temperature gradient, wall-fluid interactions and surface wettability. Thus, these findings contribute to a better understanding of flow and heat transfer mechanisms at sub-micron scale and support the development of coupled MD–CFD frameworks for accurate analysis.