Carbon nanotubes (CNT) and graphene are considered as potential future candidates for many nano/microscale integrated devices due to their superior thermal properties. Both systems, however, exhibit ...significant anisotropy in their thermal conduction, limiting their performance as three-dimensional thermal transport materials. From thermal management perspective, one way to tailor this anisotropy is to consider designing alternative carbon-based architectures. This paper investigates the thermal transport in one such novel architecturea pillared-graphene (PG) network nanostructure which combines graphene sheets and carbon nanotubes to create a three-dimensional network. Nonequilibrium molecular dynamics simulations have been carried out using the AIREBO potential to calculate the thermal conductivity of pillared-graphene structures along parallel (in-plane) as well as perpendicular (out-of-plane) directions with respect to the graphene plane. The resulting thermal conductivity values for PG systems are discussed and compared with simulated values for pure CNT and graphite. Our results show that in these PG structures, the thermal transport is governed by the minimum interpillar distance and the CNT−pillar length. This is primarily attributed to scattering of phonons occurring at the CNT−graphene junctions in these nanostructures. We foresee that such architecture could potentially be used as a template for designing future structurally stable microscale systems with tailorable in-plane and out-of-plane thermal transport.
Molecular modeling of thermosetting polymers has been presented with special emphasis on building atomistic models. Different approaches to build highly cross-linked polymer networks are discussed. A ...multistep relaxation procedure for relaxing the molecular topology during cross-linking is proposed. This methodology is then applied to an epoxy-based thermoset (EPON-862/DETDA). Several materials properties such as density, glass transition temperature, thermal expansion coefficient, and volume shrinkage during curing are calculated and found to be in good agreement with experimental results. Along with the material’s properties, the simulations also highlight the distribution of molecular weight buildup and inception of gel point during the network formation.
Heat transfer of phase change material (PCM) in an open cell micro-foam structure was numerically studied. A high constant temperature was specified at the top surface of the structure. Each unit of ...the micro-foam is a body-centered-cubic (BCC) lattice embedded with spherical micro-pores. Two different simulation methodologies were applied. One is the high-fidelity direct numerical simulation (DNS), which allows for the effective thermo-physical parameters to be derived. The other methodology is a volume-averaged simulation based on one- and two-temperature models. Our results show that the volume-averaged simulation can accurately and efficiently capture the phase change process in PCM/micro-foam systems, with the effective thermal conductivity derived from direct simulations and expressed as a power law of porosity.
Molecular dynamics and molecular mechanics simulations have been used to study thermo-mechanical response of highly cross-linked polymers composed of epoxy resin DGEBA and hardener DETDA. The ...effective cross-linking approach used in this work allowed construction of a set of stress-free molecular models with high conversion degree containing up to 35000 atoms. The generated structures were used to investigate the influence of model size, length of epoxy strands, and degree of cure on thermo-mechanical properties. The calculated densities, coefficients of thermal expansion, and glass transition temperatures of the systems are found to be in good agreement with experimental data. The computationally efficient static deformation approach we used to calculate elastic constants of the systems successfully compensated for the large scattering of the mechanical properties data due to nanoscopically small volume of simulation cells and allowed comparison of properties of similar polymeric networks having minor differences in structure or chemistry. However, some of the elastic constants obtained using this approach were found to be higher than in real macroscopic samples. This can be attributed to both finite-size effect and to the limitations of the static deformation approach to account for dynamic effects. The observed dependence of properties on system size, in this work, can be used to estimate the contribution of large-scale defects and relaxation events into macroscopic properties of the thermosetting materials.
•The liquid octane accommodation coefficient αc was studied via molecular dynamics simulations.•The average lifetime of molecular pair-association in bulk liquid octane was estimated.•Octane ...condensation probabilities were estimated in the room temperature range.•The model of octane molecule was found to have a stronger effect than temperature on the computed αc.•The assumption that αc=1 was confirmed valid in the room temperature range.
The condensation coefficient, αc, also called mass accommodation coefficient, of liquid octane, is assumed to be 1.0 in fluid models of thin film evaporation even though it decreases with increasing liquid temperature. The temperature range where αc=1 is valid, is yet to be fully described. In this work, the condensation coefficient of liquid octane was estimated by employing large scale equilibrium classical molecular dynamics simulations and computing the condensation probability as the ratio of the number of octane molecules that condense in the liquid phase to the number of octane molecules incident at the interface from the gas phase. The effect of the model details on the computed value of the condensation coefficient was explored by using two different parameterizations of the octane molecule, the all-atom OPLS and the united-atom TraPPE force fields. The condensation probabilities were computed in the temperature range from 290K to 350K. Bulk properties of liquid octane were also studied via molecular dynamics simulations and the persistent residence times of nearest coordination of octane molecules were determined. At 300K, the temperature of interest in electronics cooling applications, the united-atom molecular model predicted αc=0.96 and the all-atom model predicted αc=0.93. No significant temperature dependence of the condensation coefficient was observed within the studied temperature range, therefore, supporting the assumption of αc=1.0 at room temperature. The two octane models predicted similar liquid structure and condensation probabilities. Compared to the all-atom molecular model, the united-atom model was found to predict shorter residence times in the first coordination shell of bulk octane molecules. This work extends the computational prediction of octane condensation coefficient into room temperature range.
We investigated molecular interactions involved in the selective binding of several short peptides derived from phage-display techniques (8−12 amino acids, excluding Cys) to surfaces of Au, Pd, and ...Pd−Au bimetal. The quantitative analysis of changes in energy and conformation upon adsorption on even {111} and {100} surfaces was carried out by molecular dynamics simulation using an efficient computational screening technique, including 1000 explicit water molecules and physically meaningful peptide concentrations at pH = 7. Changes in chain conformation from the solution to the adsorbed state over the course of multiple nanoseconds suggest that the peptides preferably interact with vacant sites of the face-centered cubic lattice above the metal surface. Residues that contribute to binding are in direct contact with the metal surfaces, and less-binding residues are separated from the surface by one or two water layers. The strength of adsorption ranges from 0 to −100 kcal/(mol peptide) and scales with the surface energy of the metal (Pd surfaces are more attractive than Au surfaces), the affinity of individual residues versus the affinity of water, and conformation aspects, as well as polarization and charge transfer at the metal interface (only qualitatively considered here). A hexagonal spacing of ∼1.6 Å between available lattice sites on the {111} surfaces accounts for the characteristic adsorption of aromatic side groups and various other residues (including Tyr, Phe, Asp, His, Arg, Asn, Ser), and a quadratic spacing of ∼2.8 Å between available lattice sites on the {100} surface accounts for a significantly lower affinity to all peptides in favor of mobile water molecules. The combination of these factors suggests a “soft epitaxy” mechanism of binding. On a bimetallic Pd−Au {111} surface, binding patterns are similar, and the polarity of the bimetal junction can modify the binding energy by ∼10 kcal/mol. The results are semiquantitatively supported by experimental measurements of the affinity of peptides and small molecules to metal surfaces as well as results from quantum-mechanical calculations on small peptide and surface fragments. Interfaces were modeled using the consistent valence force field extended for Lennard-Jones parameters for fcc metals which accurately reproduce surface and interface energies Heinz, H.; Vaia, R. A.; Farmer, B. L.; Naik, R. R. J. Phys. Chem. C 2008, 112, 17281−17290.
This paper employs fully atomistic molecular dynamics simulations to characterize relationships between structural and elastic properties of thermosetting polymers both in glassy and rubbery state. ...The polymer system investigated consists of epoxy resin DGEBA and hardener DETDA. An effective cross-linking procedure that enables generation of thermoset structures containing up to 35000 atoms with realistic structural characteristics was used. A dynamic deformation approach has been used that takes into consideration both potential energy and thermal motions in the structure. Small uniaxial, volumetric and shear deformations were applied to the systems to obtain elastic moduli. A method to independently determine Poisson's ratio was proposed that reduces statistical errors and circumvents the time scale limitations of molecular dynamics simulations. The influence of variables such as extent of curing and length of epoxy strands on elastic response at various temperatures was explored. Expected trends in the dependence of the elastic constants on these practical process parameters were shown. The relationship between the four independently calculated elastic constants was seen to comply with those predicted by the classical theory of linear elasticity in an isotropic medium, which provides confidence in the validity of these simulations. Moreover, the elastic properties obtained are also in good agreement with experimental data reported in the literature. Close agreements between predicted elastic constants and experimentally measured values underscore the ability of the approaches used in this study to provide realistic predictions of the mechanical response of thermosetting polymers, both in glassy and rubbery states. These results show significant improvement over earlier studies based on a static approach which takes into account the potential energy contribution to the elastic response but ignores temperature effect.
Display omitted
In this article, we have investigated the anisotropic nature of thermal transport in molybdenum disulphide using molecular dynamics simulations. At first, a force field has been validated with ...respect to crystal structure and experimental vibrational spectra of MoS
2. Thereafter, non-equilibrium MD simulations have been performed in two perpendicular directions (along as well as across the basal planes) to study thermal transport behavior. At room temperature, our results show an anisotropic factor of ∼4 in the values of thermal conductivity along two studied directions, which is in good agreement with recent experiments on MoS
2 thin films. However, the predicted values of thermal conductivity are about an order of magnitude higher with respect to experiments. The reasoning behind these differences has been discussed in terms of layer disorder and the large number of grain boundary interfaces in experimental thin films, which consisted of nano-crystalline MoS
2 grains with a predominant parallel or perpendicular basal plane orientation. Incorporation of phonon scattering via structure disorder and boundary interfaces were identified as further directions for the model refinement.
This article explores the transverse thermal conductance between two parallelbonded as well as nonbondedcarbon nanotubes, embedded in an epoxy matrix using nonequilibrium molecular dynamics ...simulations. Here, we study the effect of different organic linkersconnecting the two nanotubeson the thermal interface conductance and compare these results with those of nonbonded nanotubes. Our results suggest that incorporation of linker molecules significantly modifies overall interface conductance between nanotubes. Specifically, we find that the conductance increases with the number of linking functionality but shows an opposite trend with respect to the linker’s length, that is, the longer the linker is, the lower the conductance. We attribute this behavior to weakening of van der Waals interactions between carbon nanotubes in the case of longer linkers as well as possible scattering of thermal vibrations that occur along the linker molecules.
•We consider transient temperature effects on laser diode optical output.•Laser diode junction temperature affects both optical power level and laser light intensity.•The decrease in laser light ...intensity is of particular interest.•With high optical power, but lower laser intensity, a certain laser system will not be effective.•Our modeling yields a clear understanding of why diode thermal management is so important.
Laser diode optical output is studied and modeled. Four major diode parameters (threshold current, slope efficiency, central wavelength of output, and full-width half maximum of output), which are dependent on diode junction temperature, determine the optical output. The physics and equations representative thereof for each parameter are presented and incorporated into a multiphysics model of a high-power laser system (HPLS) to study the optical power/thermal interactions. Simulations are compared to show how optical power output of an HPLS changes when the temperature dependence of parameters are and are not accounted for in the model. The decrease in laser light intensity out of the HPLS as junction temperature changes is also studied. Intensity is sometimes a more important consideration than optical power because for most applications, laser light is only effective when the output power is focused over a very narrow wavelength range. The research provides higher fidelity diode modeling for effectively understanding optical/thermal interactions and the price to be paid for improper diode thermal management. The research supports our main goal of more accurately representing the thermal loads from the individual components of an HPLS.
