|
Research:
Molecular Dynamics Computational Simulations
Molecular dynamics (MD) is a deterministic technique based on Newton's equations of motion as applied to individual atoms or molecules. It does not require knowledge of additional parameters except a well-established interatomic potential, and MD is a viable method for providing insight into the nanoscale world, especially where thermal transport is concerned. In the nanoEngineering laboratory, MD was used to study the influence that interfaces have on nanoscale thermal transport in solid argon and argon-based variant materials. To accomplish this, the mass m and bond strength e of pure argon were systematically modified to create variations on two types of heterogeneous systems: a layered film and a nanocomposite film. A comparison of thermal conductivity was made between simulation results and kinetic theory as well as thermal resistance network models and the acoustic mismatch model to help provide a basic understanding for analysis of the heterogeneous systems. Application of harmonic-oscillation and Debye approximations with yielded thermal conductivities that were on the same order of magnitude as the MD results.
Generally speaking, kinetic theory produced trend curves and the simulation data followed each other fairly well, comparing better with lower parameter values yet falling away slightly for higher parameter values. The differences were most likely due to the assumptions made in approximation the mean free path, phonon velocity, and heat capacity. The use of an experimental mean free path in the theoretical calculations that was likely lower than the true mean free path of the simulated materials undoubtedly contributed to the disagreement. Moreover, the known overestimation of the phonon velocity corresponding to pure argon likely indicated that the variant velocity estimates were also high. Not only did this affect the calculation of thermal conductivity directly, but heat capacity was also affected since it depends on velocity. To truly appreciate the deviation from kinetic theory type behavior, additional coding to capture the simulated heat capacity and phonon velocities of each simulation is necessary, and will be the focus of future work. The introduction of a second material into an argon film generally decreased its overall thermal conductivity. However, an exception was observed where slightly increased e values produced a slight increase in thermal conductivity over pure argon. A comparison of the layered system with the nanocomposite demonstrated that the presence of a nanoparticle was less influential in reducing thermal conductivity than the thin layer. This behavior was expected since the phonons had less of an obstruction along the path of heat flow. Application of the thermal resistance network without the AMM illustrated the significance of the interfacial effects as it predicted thermal conductivity values whose offset increased with greater interfacial disparity. Comparison of calculated AMM values to the simulated interfacial resistances showed promise; however, the AMM did not fully capture the behavior observed in the simulations. This may be due to several factors including the aforementioned assumptions used to predict phonon velocities which significantly affect the calculation of interfacial resistances, as well as the techniques used to collect and analyze the simulation data.
In all, the results indicate that combining a TRN with the AMM to describe nanocomposite thin films provides a means for investigating interfacial thermal resistance effects. More importantly, the results show that the configuration studied can be used to study more complex structures with reasonable success. Continuation of this work will involve improvements to the MD simulator, as well as the investigation of new systems that may exhibit interesting interfacial thermal effects. Moreover, the simulator will be expanded to collect the data necessary for calculation of dispersion relations, phonon velocities and heat capacities.
|
