Carbon-Based Nanomaterials and Their Ensembles for High Temperature Thermal Applications
Carbon nanomaterials, mainly including carbon nanotubes and graphene, have high potential for heat transfer applications at high temperatures because of their superb heat transport properties and good thermal stability. However, due to the small physical sizes of carbon nanomaterials, real-world applications often require an ensemble of them. The present study aims to characterizing the thermal properties of carbon nanomaterial ensembles and understanding the underlying mechanism with an emphasis on high temperature applications. A one-dimensional (1D) reference bar method is selected to perform thermal transport experiments on target materials. Despite its popularity for room temperature measurements, this method does not readily extend to the high-temperature regime, mainly due to oxidation and convective and radiative heat loss concerns. In this dissertation, a modified 1D reference bar test rig is presented that eliminates these problems. Oxidation and convection are avoided by vacuum. Radiation heat loss is accounted for by a data fitting algorithm. Monte Carlo simulation is used to quantify the uncertainty of this method. The system is also validated by testing a commercially available thermal interface material. One of the major drawbacks of the steady-state reference bar method is its slow test speed. Reaching thermal equilibrium takes a significant amount of time, from an hour up to days. Typical transient methods, which can perform tests much faster, require special instruments such as modulated heaters. In this dissertation, a new transient method is presented that can be used directly with existing 1D reference bar test rigs. Using the temperature response of the reference bars in time domain, thermal properties of the sample can be extracted. Uncertainty quantification shows that measurement accuracy is not lost compared to steady-state methods, but the fast test speed is shown to reduce the time needed to perform a test by as much as 40 times. Vertically oriented carbon nanotube (CNT) arrays hold high promise for thermal interface applications. However, such an ensemble of CNTs behaves much differently than a collection of isolated CNTs and suffers from various interface effects. After years of research, the thermal transport characteristics of CNT arrays are still not fully understood. Also, experimental data at elevated temperatures are lacking. Using the newly developed high temperature 1D reference bar test rig, thermal interface properties of CNT arrays are examined, and the results are presented in this dissertation. Thermal interface resistance of CNT arrays is found to consistently decrease at high temperatures for both thermomechanically matched and mismatched interfaces. The results also suggest that contact resistances between CNT tips and the opposing substrates are major contributions to the total interface resistances. A method of integrating CNT arrays to braze joints is also developed to improve CNT-based thermal interface materials. Braze alloys are found to infiltrate into CNT arrays and form strong chemical bonds. Thermal characterization results suggest very good thermal interface performance, which is further shown to be unaffected by thermomechanical stresses. Graphene aerogels are studied as another type of carbon nanomaterial ensemble. Their thermal conductivities are measured at varying volume fraction, temperature and compressive strain. Not surprisingly, increasing volume fraction and temperature are shown to increase the thermal conductivity. However, results imply that interfaces are critical to the material in terms of thermal transport. Thermal tests in compression and accompanying microscopy more vividly show the role of interfaces. The study demonstrates that with a combination of low density, defects and interface engineering, the thermal properties of graphene derivatives can be tuned across many orders of magnitude.
Fisher, Purdue University.
Off-Campus Purdue Users:
To access this dissertation, please log in to our