Aluminium nanopillars reduce thermal conductivity of silicon nanobeams

Abstract

In search of efficient thermoelectric nanostructures, many theoretical works predicted that nanopillars, placed on the surface of silicon membranes, nanobeams, or nanowires, can reduce the thermal conductivity of these nanostructures. To verify these predictions, we experimentally investigate heat conduction in suspended silicon nanobeams with periodic arrays of aluminium nanopillars. Our room temperature time-domain thermoreflectance experiments show that the nanobeams with nanopillars have 20% lower thermal conductivity as compared to pristine nanobeams. We discuss possible explanations of these data, including coherent effects, and conclude that the thermal conductivity is reduced mainly by phonon scattering at the pillar/beam interface due to the intermixing of aluminium and silicon atoms, as supported by the transmission electron microscopy. As this intermixing does not only reduce thermal conductivity but may also increase electrical conductivity, these nanostructures are exceptionally promising for thermoelectric applications.

Phonon and heat transport control using pillar-based phononic crystals

“…Figure 5(b) shows that the pillars indeed reduce the thermal conductivity, and this reduction is stronger for the pillars of larger diameter. However, contrary to the theoretical predictions [30,63], the temperature dependent measurements did not reveal significant changes in the thermal conductivity reduction at low temperatures. Analyzing the interface between the pillars and the beams, they found that the deposition of aluminum pillars causes amorphization and roughening of the surface under the pillars, as shown in the inset of Figure 5(b).…”

“…Reprinted from [62], with the permission of AIP Publishing. (b) The thermal conductivity of the nanobeams with pillars of different diameters measured by Anufriev et al [63] at room temperature. The inset shows the fragment of the nanobeam and transmission electron microscopy image of the pillar/beam interface featuring surface roughness. …”

“…To test the theoretical predictions more directly, Anufriev et al [63] measured the thermal conductivity of thin silicon beams with aluminum pillars of different diameters. Figure 5(b) shows that the pillars indeed reduce the thermal conductivity, and this reduction is stronger for the pillars of larger diameter.…”

“…On one hand, a stronger reduction was measured in silicon with pores (>50%) [9,18,19], nanodots (>70%) [13,68], polycrystalline grains (>80%) [15,69,70], dopants (>50%) [71,72] or germanium atoms (>70%) [14,71,73]. On the other hand, the 20% reduction by the pillars [63] is comparable to the reduction by holes (20–25%) [21,74] or slits (20–30%) [75] covering the same relative area. This relative comparison shows that pillars could achieve the same reduction in the thermal conductivity without sacrificing the material volume or introducing scattering points inside the bulk of the material.…”

Thermal Conductivity Reduction in a Nanophononic Metamaterial versus a Nanophononic Crystal: A Review and Comparative Analysis

“…Nanophononic metamaterials have also been studied using finite‐element (FE) analysis as commonly done for macroscale acoustic or elastic metamaterials . However, while linear continuous modeling may provide some insight on the dynamics of local resonances, nonlinear atomic‐scale models are necessary for accurate qualitative and quantitative capturing of the effects of the three key NPM mechanisms responsible for the thermal conductivity reduction.…”