Propagating elastic vibrations dominate thermal conduction in amorphous silicon

Abstract

The thermal atomic vibrations of amorphous solids can be distinguished by whether they propagate as elastic waves or do not propagate due to lack of atomic periodicity. In a-Si, prior works concluded that nonpropagating waves are the dominant contributors to heat transport, with propagating waves being restricted to frequencies less than a few THz and scattered by anharmonicity. Here, we present a lattice and molecular dynamics analysis of vibrations in a-Si that supports a qualitatively different picture in which propagating elastic waves dominate the thermal conduction and are scattered by local fluctuations of elastic modulus rather than anharmonicity. We explicitly demonstrate the propagating nature of waves up to around 10 THz, and further show that pseudoperiodic structures with homogeneous elastic properties exhibit a marked temperature dependence characteristic of anharmonic interactions. Our work suggests that most heat is carried by propagating elastic waves in a-Si and demonstrates that manipulating local elastic modulus variations is a promising route to realize amorphous materials with extreme thermal properties.

Thermal Conductivity of Amorphous Materials

“…Recently, Moon et al provided a direct picture to distinguish between propagating and nonpropagating vibrational modes by conducting a “tuning fork” computer experiment. In their MD simulation, an a‐Si system of 4.3 nm × 4.3 nm × 43 nm was created and divided into 80 slabs in the long dimension.…”

“…a) 3 THz, b) 8 THz, and c) 16 THz. Adapted with permission . Copyright 2018, American Physical Society.…”

“…These authors further calculated the product of the lifetime and vibrational frequency as a function of frequency . As shown in Figure , a critical horizontal line is defined at which the lifetime is equal to the period of the wave, also known as the IR crossover from propagons to diffusons.…”

“…Clearly, the vibrational dynamics in amorphous silicon is not governed by anharmonicity. Adapted with permission . Copyright 2018, American Physical Society.…”

A unified deep neural network potential capable of predicting thermal conductivity of silicon in different phases

“…We note that many MD and theoretical studies show various results ranging from 1 to 3 Wm -1 K -1 [52][53][54][55][56], depending on the different EIPs and system sizes used. Moreover, recent experimental works show large variations due to different material preparation and measurement techniques, and the measured κ of thicker a-Si films can be as high as 4 Wm -1 K -1 at 300 K [57,58], with no convergence as the film thickness increases.…”