We assess a Fourier-based thermal model used in frequency-domain thermoreflectance data analysis. The Boltzmann transport equation (BTE) is first used to simulate sub-continuum phonon transport in a semi-infinite solid. We then compare the BTE-predicted temperature profiles to those predicted by an analytical solution of the Fourier-based conduction equation. The two models agree well when ωτ < 1, where ω is the surface-temperature modulation-frequency and τ is the bulk phonon relaxation time, but diverge when ωτ > 1.

Thermal transport by phonons in water/carbon nanotube (CNT) composite systems is investigated using molecular dynamics (MD) simulation. We calculate the spectral energy density of empty and water-filled CNTs and use it to extract the mode-specific phonon group velocities, relaxation times, and thermal conductivities. The total thermal conductivity predicted from the spectral energy density is consistent with what we predict using a direct application of the Fourier law in an MD simulation. The number of atoms and simulation runtime required to predict the spectral energy density, however, are both at least one order-of-magnitude smaller.

We investigate thermal transport in water/carbon nanotube (CNT) composite systems using molecular dynamics simulations. Carbon-carbon interactions are modeled using the second-generation REBO potential, water-water interactions are modeled using the TIP4P potential, and carbon-water interactions are modeled using a Lennard-Jones potential. The thermal conductivities of empty and water-filled CNTs with diameters between 0.83 nm and 1.66 nm are predicted using molecular dynamics simulation and a direct application of the Fourier law. For empty CNTs, the thermal conductivity decreases with increasing CNT diameter. As the CNT length approaches 1 micron, a length-independent thermal conductivity is obtained, indicative of diffusive phonon transport. When the CNTs are filled with water, the thermal conductivity decreases compared to the empty CNTs and transitions to diffusive phonon transport at shorter lengths. To understand this behavior, we calculate the spectral energy density of the empty and water-filled CNTs and calculate the mode-specific group velocities, relaxation times, and thermal conductivity. For the empty 1.10 nm diameter CNT, we show that the acoustic phonon modes account for 65 percent of the total thermal conductivity. This behavior is attributed to their long mean-free paths. When the CNT is filled with water, interactions with the water molecules shorten the acoustic mode mean-free path and lower the overall CNT thermal conductivity.

Semiconductor thin films are pervasive in advanced technological devices. For example, the active region in a quantum cascade laser contains films of direct bandgap semiconductors with thicknesses of 1–10 nm [1]. Superlattices, periodic materials that contain films of alternating species with thicknesses as small as 1 nm, are being studied for their potential to increase the efficiency of thermoelectric energy conversion devices [2]. Because these film thicknesses are less than the mean free paths of the phonons in the corresponding bulk material, the thermal transport properties of semiconductor thin films are different than their corresponding bulk values [3]. To aid in the design of devices employing thin films, accurate models for the thermal transport properties of the film are required.

The effective thermal conductivity of water/carbon nanotube (CNT) composite systems is predicted using molecular dynamics simulation. Both empty and water-filled CNTs with diameters ranging from 0.83 nm to 1.26 nm are considered. Using a direct application of the Fourier law, we explore the transition to diffusive phonon transport with increasing CNT length and identify the correlation between CNT diameter and fully-diffusive thermal conductivity. Using Green-Kubo linear response theory, we explore how the thermal conductivity of water inside CNT varies with tube diameter. We predict the effective thermal conductivity of the composite systems and examine how the phonon modes in the CNT are affected by interactions with the water molecules.

Quantum corrections can be used to map the thermal conductivity predicted in a classical framework [e.g., a molecular dynamics (MD) simulation] to a corresponding value in a quantum system. This procedure is accomplished by equating the total energies and energy fluxes of the classical and quantum systems. The validity of these corrections is questionable because they are introduced in an ad hoc manner and are not derived from first principles. In this work, the validity of these quantum corrections is examined by comparing the thermal conductivity of Stillinger-Weber silicon calculated using a full quantum mechanical treatment to a quantum-corrected value predicted from a classical framework between temperatures of 10 K and 1000 K. The quantum and classical predictions are obtained using anharmonic lattice dynamics calculations. We find discrepancies between the quantum-corrected predictions and the quantum predictions obtained directly. We investigate the causes of these discrepancies and from our findings, conclude that quantum thermal conductivities cannot be predicted by applying simple corrections to the values obtained from a purely classical framework.

Phonon scattering at the interface between two materials results in a thermal resistance, R [1]. An ability to accurately predict the thermal resistance of semiconductor interfaces is important in devices where phonon interface scattering is a significant contributor to the overall thermal resistance (e.g., computer chips with high component density). This ability will also lead to improvements in the design of semiconductor superlattices with low thermal conductivity, desirable in thermoelectric energy conversion applications [2].

Pressure-driven water flow through carbon nanotubes (CNTs) with diameters ranging from 1.66 nm to 4.99 nm is examined using molecular dynamics simulation. The flow rate enhancement, defined as the ratio of the observed flow rate to that predicted from the no-slip Hagen-Poiseuille relation, is calculated for each CNT. The enhancement decreases with increasing CNT diameter and ranges from 433 to 47. By calculating the variation of water viscosity and slip length as a function of CNT diameter, it is found that the results can be fully explained in the context of continuum fluid mechanics. The enhancements are lower than previously reported experimental results, which range from 560 to 100000, suggesting a miscalculation of the available flow area and/or the presence of an uncontrolled external driving force (such as an electric field) in the experiments.

Phonon scattering at an interface between two materials results in a thermal boundary resistance R , given by [1] R = Δ T q , ( 1 ) where Δ T and q are the temperature drop and heat flux across the interface. Predicting the thermal boundary resistance of semiconductor/semiconductor interfaces is important in devices where phonon interface scattering is a significant contributor to the overall thermal resistance (e.g., computer chips with high component density). Such predictions will also lead to improvements in the design of nanocomposite materials (e.g., superlattices) with low thermal conductivity, desirable in thermoelectric energy conversion applications [2].

Lattice dynamics calculations are used to investigate thermal transport in argon thin films with thicknesses ranging between one and ten nanometers. Quasi-harmonic lattice dynamics calculations are used to find the frequencies and mode shapes of non-interacting phonons. This information is then used as input for anharmonic lattice dynamics calculations, whereby we compute the frequency shift and lifetime of each phonon mode due to interactions with other phonons. The phonon frequencies, group velocities, and lifetimes determined with the lattice dynamics techniques are then used to compute the in-plane thermal conductivity of the thin films as a function of film thickness. The thermal conductivities predicted by the lattice dynamics methods are compared to predictions from molecular dynamics simulations. Differences in the phonon characteristics in thin films compared to bulk crystals are examined by comparing the contribution to the thermal conductivity as a function of frequency.

Phonon transport in argon and silicon thin films is examined using harmonic lattice dynamics theory and the Lennard-Jones and Stillinger-Weber potentials. Film thicknesses ranging from 0.8 to 33.5 nm for argon and 0.4 to 8.6 nm for silicon are examined at a temperature of 0 K. Both reconstructed films and films built using the bulk unit cell are considered. Phonon dispersion curves for the in-plane direction and the density of states are computed from lattice dynamics and compared to predictions for a bulk system. The results from the lattice dynamics calculations are used to predict the thermal conductivities of the bulk and thin film structures.

Molecular dynamics simulations and the non-equilibrium direct method are used to predict the thermal conductivity of a Si/Ge superlattice modeled by the Stillinger-Weber potential at a temperature of 300 K. We focus on the methodology of making the thermal conductivity prediction (limited effort has been made to model Si/Ge nanocomposites in the literature) and find that proper selection of the size and composition of the thermal reservoirs is important.

Molecular dynamics simulations are used to investigate the subcritical and supercritical evaporation of a Lennard-Jones (LJ) argon nanodroplet in its own vapor. Using a new technique to control both the ambient temperature and pressure, a range of conditions are considered to define a transition line between subcritical and supercritical evaporation. The evaporation is considered to be supercritical if the surface temperature of the droplet reaches the LJ argon critical temperature during its lifetime. Between ambient temperatures of 300 K and 800 K, the transition from subcritical to supercritical evaporation is observed to occur at an ambient pressure 1.4 times greater than the LJ argon critical pressure. For subcritical conditions, the droplet lifetimes obtained from the simulations are compared to independently predicted lifetimes from the D 2 law.

Si/Si 1− x Ge x superlattices are promising candidates for thermoelectric energy conversion applications [1, 2], as the phonon transport through them can be inhibited while maintaining desirable electrical transport properties. No comprehensive experimental study has been performed to map the thermal conductivity design space accessible by Si/Ge nanocomposites. By using atomistic modeling tools, interesting areas of the design space can be identified and then further explored experimentally.

Lattice dynamics calculations are used to investigate thermal transport in the face-centered cubic Lennard-Jones (LJ) argon crystal between temperatures of 20 and 80 K. First, quasi-harmonic lattice dynamics calculations are used to find the frequencies and mode shapes of non-interacting phonons [1]. This information is then used as input for anharmonic lattice dynamics calculations. Anharmonic lattice dynamics is a means of computing the frequency shift and lifetime of each phonon mode due to interactions with other phonons [2]. The phonon frequencies, group velocities, and lifetimes, determined with the lattice dynamics methods, are then used to compute the thermal conductivity. The thermal conductivities predicted by the lattice dynamics methods are compared to predictions from molecular dynamics simulations. The two methods are found to agree well at low temperature but diverge at higher temperatures (i.e., near the melting point). The properties of individual phonon modes are used to identify the modes that dominate thermal transport.

Thermal transport in crystals is governed by dynamic phenomena that take place at the atomic scale, namely phonon dispersion and scattering. A growing understanding of these mechanisms, coupled with increasingly capable nanofabrication and characterization technologies, provide a not-too-distant opportunity for designing a new class of materials with tailored thermal characteristics such as thermal conductivity, among other physical characteristics. Focusing on layered nanocomposites, also known as superlattices, modeled using the Lennard-Jones potential as a starting platform, we examine the effects of layering topology on the bulk property of thermal conductivity. We use molecular dynamics simulations to examine the link between structure and property; and employ ideas from phononic crystal design to investigate the potential of realizing dielectric crystals with exceedingly low thermal conductivities. This work potentially targets a range of applications such as thermal insulators for space applications and thermoelectrics for energy harvesting.

Molecular dynamics simulations and the non-equilibrium direct method are used to predict the effective cross-plane thermal conductivity of Si/Si 1 −x Ge x superlattices modeled by the Stillinger-Weber potential. The experimentally observed thermal conductivity design space and the methodology of making the thermal conductivity prediction with the direct method are reviewed. Preliminary results for the thermal conductivity prediction of a Si/Si 0.7 Ge 0.3 at a temperature of 500 K are discussed.

The thermal conductivity of the Lennard-Jones argon crystal is predicted using the Boltzmann transport equation under the single mode relaxation time approximation. The temperature and frequency dependence of the phonon dispersion and phonon relaxation times are obtained from lattice dynamics calculations based on the results of molecular dynamics simulations. No fitting parameters are required. The thermal conductivity results are in agreement with predictions from the simulations using the Green-Kubo method. Assuming that the dispersion is linear or temperature independent, as is often done in analytical calculations, leads to large errors in the predictions. This result suggests that the fitting parameters required in such calculations offset significant errors introduced by these simplifying assumptions.

The phononic band structure and thermal conductivity of a family of a two-atom unit cell Lennard-Jones crystals are predicted using molecular dynamics simulations. The structure consists of alternating layers of atoms with different masses, leading to anisotropic thermal properties. An increase in the mass ratio results in an increase in the width of the band gap and a decrease in the value of its central frequency. The thermal conductivity decreases with an increase in the mass ratio, and in all cases is lower than that for a monatomic unit cell. The thermal conductivity increases with an increase in the ratio of the central gap frequency to its width. The rate of this increase decreases and becomes less temperature-dependent at lower mass ratios. The results could be utilized towards the development of guidelines for the design of materials with extreme thermal transport properties.

Molecular dynamics simulations and the Green-Kubo method are used to predict the thermal conductivity of binary Lennard-Jones superlattices and alloys. The superlattice thermal conductivity trends are in agreement with those obtained through the direct method, verifying that the Green-Kubo method can be used to examine thermal transport in heterostructures. The simulation temperature and the constituent species are fixed while the superlattice period structure is varied with the goals of (i) minimizing the cross-plane thermal conductivity and (ii) maximizing the ratio of in-plane to cross-plane thermal conductivities. The superlattice thermal conductivity in both the cross-plane and in-plane directions is found to be greater than the corresponding alloy value and less than the value predicted from continuum theory. The anisotropy of the thermal conductivity tensor is found to be at a maximum for a superlattice with a uniform layer thickness. Lattice dynamics calculations are used to investigate the role of optical phonons in the thermal transport.