Search

Efficient thermal management of large vertically stacked integrated circuits (ICs) requires a material with high thermal conductivity inside the micrometer-sized gaps between the IC dies. Such an underfill material can be obtained by adding thermally conductive filler particles at high loadings to the adhesive matrix material. However, viscosity requirements for the state-of-the-art capillary-driven filling of particle-loaded adhesives limit the particle fill fraction to values lower than those necessary to reach true percolation. Accordingly, heat transfer through the composite material is dominated by conduction through the adhesive matrix material. We propose using an alternative, sequential filling method to achieve percolation with particles down to 1~\mu \textm in diameter. The percolating thermal underfill (PTU) can be achieved by a centrifuge-assisted filling of micrometer-sized particles, followed by a backfilling of a low-viscosity epoxy with long open time acting as matrix material, and a final thermal curing step. In this paper, we present and discuss the fabrication and relevant process parameters of the PTU composite material in chip stacks with critical dimensions below 30~\mu \textm . The wet dispensing of alumina particles with a diameter distribution of 1– 15~\mu \textm is proposed for overcoming the agglomeration observed for dry particle filling. Thermal conductivities of up to 3 W/( \ m\cdot \ K ) for the underfill system with a particle fill fraction of 61% were achieved, which is three times higher than those of commercially available capillary thermal underfills. Critical parameters in the formation of the percolating composite, such as the choice of filler material and its refinement, as well as the properties of the particle bed and the process parameters for the centrifugal filling, epoxy capillary backfilling, and composite curing are addressed. [ABSTRACT FROM PUBLISHER]

Heat dissipation from three-dimensional (3D) chip stacks can cause large thermal gradients due to the accumulation of dissipated heat and thermal interfaces from each integrated die. To reduce the overall thermal resistance and thereby the thermal gradients, this publication will provide an overview of several studies on the formation of sequential thermal underfills that result in percolation and quasi-areal thermal contacts between the filler particles in the composite material. The quasi-areal contacts are formed from nano-particles self-assembled by capillary bridging, so-called necks. Thermal conductivities of up to 2.5W/m K and 2.8W/m K were demonstrated experimentally for the percolating and the neck-based underfills, respectively. This is a substantial improvement with respect to a state-of-the-art capillary thermal underfill (0.7 W/m K). Critical parameters in the formation of sequential thermal underfills will be discussed, such as the material choice and refinement, as well as the characteristics and limitations of the individual process steps. Guidelines are provided on dry versus wet filling of filler particles, the optimal bimodal nanosuspension formulation and matrix material feed, and the over-pressure cure to mitigate voids in the underfill during backfilling. Finally, the sequential filling process is successfully applied on microprocessor demonstrator modules, without any detectable sign of degradation after 1500 thermal cycles, as well as to a two-die chip stack. The morphology and performance of the novel underfills are further discussed, ranging from particle arrangements in the filler particle bed, to cracks formed in the necks. The thermal and mechanical performance is benchmarked with respect to the capillary thermal and mechanical underfills. Finally, the thermal improvements within a chip stack are discussed. An 8- or 16-die chip stack can dissipate 46% and 65% more power with the optimized neck-based thermal underfill than with a state-of-the-artcapillary thermal underfill. [ABSTRACT FROM AUTHOR]

Efficient heat removal from integrated circuits arranged vertically in 3-D chip stacks requires thermally conductive underfill materials. The low-heat-transport performance of traditional capillary underfills can be improved by percolating the thermal conductive filler particles. We increased the thermal path by adding quasi-areal contacts using nanoparticle assemblies directed to the contact points of the percolating filler particles. We studied the formation and thermal effect of such nanoparticle neck assemblies in the filler-particle contact points using aqueous suspensions containing nanoparticles of different sizes, size distributions, and concentrations. An optimized binary mixture of small (28–43 nm) and large (200–300 nm) nanoparticles results in dense and defect-free neck assemblies. A neck-enhanced percolating thermal underfill (PTU) with a thermal conductivity as high as 2.4 W/mK was achieved using alumina filler and nanoparticles. Compared to a PTU, the addition of nanoparticle necks resulted in a more than twofold improvement in thermal conductivity. [ABSTRACT FROM PUBLISHER]

Silicon nanowires (SiNWs) are promising materials for the realization of highly-efficient and cost effective thermoelectric devices. Reduction of the thermal conductivity of such materials is a necessary and viable pathway to achieve sufficiently high thermoelectric efficiencies, which are inversely proportional to the thermal conductivity. In this article, vertically aligned forests of SiNW and germanium (Ge)-doped SiNW with diameters around 100 nm have been fabricated, and their thermal conductivity has been measured. The results show that discrete surface doping of Ge on SiNW arrays can lead to 23% reduction in thermal conductivity at room temperature compared to uncoated SiNWs. Such reduction can be further enhanced to 44% following a thermal annealing step. By analyzing the binding energy changes of Ge-3d and Si-2p using X-ray photoelectron spectroscopy, we demonstrate that surface doped Ge interacts strongly with Si, enhancing phonon scattering at the Si-Ge interface as has also been shown in non-equilibrium molecular dynamics studies of single nanowires. Overall, our results suggest a viable pathway to improve the energy conversion efficiency of nanowire-forest thermoelectric nanomaterials. [ABSTRACT FROM AUTHOR]