ICMCTF2013 Session TS1-1: Surface Engineering for Thermal Transport, Storage and Harvesting

The photoacoustic (PA) technique was used to measure the thermal conductivity of nanostructured materials without the use of a metal foil or transducer layers to absorb laser energy. Using the sample material as the absorption layer eliminates the need to bond the nanostructured material to a metal foil or deposit a metal film on the sample, which creates an unknown contact resistance between the sample layer and the metal. Because of the elimination of the additional contact resistance, the bare sample measurement has a much higher theoretical sensitivity to the thermal conductivity of the sample compared to the sensitivity when a metal foil or film transducer is used. The measurement technique is demonstrated on vertical forests of carbon nanotubes (CNTs), polymer nanotubes, graphite, and polymer films. Bare graphite and CNT samples were measured and compared directly with graphite and CNT samples with deposited metal film transducers and metal foil transducers. The accuracy of the thermal conductivity measurements was found to be significantly better without metal foils or deposited films atop the sample . Measurements of the thermal conductivity of graphite and polymer films without metal transducers were in good agreement with reference values. To compliment the experimental results, a theoretical analysis was conducted to show how to best increase the sensitivity of the measurement to the sample thermal conductivity while minimizing the error due to uncertainty in the optical absorption length. Some limitations to the direct absorption PA technique are also discussed.

A growing interest has developed in the past decade on the use of carbon nanotube (CNT) arrays as thermal interface materials (TIM). This interest on CNT-based thermal interfaces stems primarily from two factors – the high intrinsic thermal conductivity of individual CNTs and mechanical compliance of the CNT array. Innovative methods for the measurement of mechanical and thermal properties of CNT arrays have been developed by many others. However, modeling of CNT TIMs have mostly been limited to semi-empirical methods without detailed consideration of the CNT array microstructure, primarily due to the inherent randomness of the microstructure and the computational complexity involved in full atomistic modeling of CNTs. A compelling need exists for developing physics-based models in order to move from interpretation of experimental data to prediction and optimization of mechanical and thermal characteristics of CNT arrays. In this work, we report combined thermo-mechanical modeling of CNT arrays with coarse-grain methods. The coarse graining of CNTs allows us to model reasonably large numbers of CNTs at practical engineering scales within a reasonable computational time. Parametric studies include the effects of CNT height, diameter and volume fraction on important mechanical properties such as the Young’s modulus and buckling load of CNT arrays. A mesoscopic thermal network model couples with the coarse grain mechanics model and uses parameters derived from the various atomistic transport studies reported in prior literature. The thermal network model is used to estimate the diffusive and tip contact resistances of CNT arrays. We also report the pressure dependence of thermal resistance, tip contact area and compare our predictions with experimental values. Other useful information for thermal interface applications such as the effects of surface roughness and fillers such as wax are also predicted within the framework of the thermal network model.

Although suspended graphene has been reported to have very high in-plane thermal conductivity, most applications will require graphene to be encased within, or supported by, dielectric layers. To understand the thermal performance of graphene in this important configuration, I will describe two measurements of the thermal transport in graphene and ultrathin graphite encased within silicon dioxide. First, the thermal contact resistance between graphene and SiO2 was measured using a differential 3-omega method, revealing significantly better thermal contact than previous reports for carbon fibers and carbon nanotubes. Second, the in-plane thermal conductivity of encased graphene sheets was measured using a novel "heat spreader" method, which fits a measured temperature profile using a finite element method (FEM) with one free parameter. The results show that the constraints of the encasing layers reduce the in-plane thermal conductivity by at least a factor of 10 as compared to bulk graphite.

The thermal conductance for a series of metal – graphite interfaces has been experimentally measured with time–domain thermoreflectance (TDTR). For metals with Debye temperatures up to ~ 400 K, a linear relationship exists with the thermal conductance values. For metals with Debye temperatures in excess of ~ 400 K, the measured metal–graphite thermal conductance values remain constant near 60 MW m^-2 K^-1. Titanium showed slightly higher conductance than aluminum, despite the closeness of atomic mass and Debye temperature for the two metals. Surface analysis was used to identify the presence of titanium carbide at the interface in contrast to the aluminum and gold – carbon interfaces (with no detectable carbide phases). It was also observed that air–cleaved graphite surfaces in contact with metals yielded slightly higher thermal conductance than graphite surfaces cleaved in vacuo. Examination of samples with scanning electron microscopy revealed that the lack of absorbed molecules on the graphite surface resulted in differences in transducer film morphology, thereby altering the interface conductance. Classical molecular dynamic simulations of metal – carbon nanotube thermal conductance values were calculated and compared to the TDTR results. The upper limit of metal – graphite thermal conductance is attributed to the counteracting effects of decreased coupling and heat capacity for higher vibrational frequency modes of the lighter metals studied.

Engineering and testing thin films requires sensitive characterization tools. Optical techniques based on thermoreflectance have emerged as reliable way to measure thermal transport in nanoscale thin films and across their interfaces. We discuss the challenges and benefits of extending time-domain and frequency-domain thermoreflectance measurements into scanning microscopy tools that can create maps of thermal conductivity, thermal interface conductance, and several other properties. Calibration methods and resolution limitations are presented along with measurements of several thin-film systems.

Hydrogen storage materials are one of the important key materials for the futuristic clean energy sources. To achieve high hydrogen absorbing capacity and suitable absorption/desorption temperature, the study on nanostructured multilayer hydrogen absorbing materials is very important. A study on hydrogenation and dehydrogenation of Pd/Mg/Pd tri-layers, prepared by DC/RF magnetron sputtering has been conducted using XRD, FE-SEM and AFM. Hydrogenation of the films was carried out at (50-150°C) temperatures in a fix amount of hydrogen gas (1bar). Hydrogen contents in as-deposited and hydrogenated thin films have been estimated by Elastic Recoil Detection Analysis (ERDA) technique with 120 MeV ₁₀₇Ag⁺⁹ ions. The maximum hydrogen absorption (8.4×10¹⁷ H atoms/cm²) has been observed for 125°C among all samples studied. The hydrogen desorption kinetics has been enhanced by cooperative phenomena in Pd/Mg/Pd tri-layers system due to elastic interactions between nanostructured Mg and Pd interface. Low desorption temperature (80° C) has been observed for Pd/Mg/Pd tri-layer system.