Analysis of non-contact and contact probe-to-sample thermal exchange for quantitative measurements of thin film and nanostructure thermal conductivity by the scanning hot probe method

Abstract

The ability of the technique to differentiate thin films from the substrate is investigated, and the sensitivity of the technique to thin films and samples with anisotropic properties is explored. The models (both analytical and finite element) developed and reported in this dissertation lead to the ability to measure samples which, by the standard procedure before this work, were unable to be accurately measured. While other techniques failed to be able to successfully interrogate the film thermal conductivity of a full set of double-wall carbon nanotubes infused into polymers, the methods developed in this work allowed non-contact scanning hot probe measurements to be successfully performed to obtain the film thermal conductivity for each sample. Finite element simulations accounting for the anisotropy of these thin film on sample materials show similar trends with independently measured in-plane thermal conductivity for the only two (of five) samples in the set which were successfully able to be measured by the independent technique. Investigations in contact mode with high resolution Pd probes, whose probe-to-sample clearance is difficult to control in a repeatable fashion, show that surface roughness affects the thermal contact resistance. This can lead to values of reported sample thermal conductivity which are misleading, when using the standard calibrated thermal exchange parameters on samples with significantly different surface roughness than the calibration samples. This affect was taken into account to report sample thermal conductivity of Bi2Te3 nanoflakes.

The ability to measure thermal properties of thin films and nanostructured materials is an important aspect of many fields of academic study. A strategy especially well-suited for nanoscale investigations of these properties is the scanning hot probe technique, which is unique in its ability to non-destructively interrogate the thermal properties with high resolution, both laterally as well as through the thickness of the material. Strategies to quantitatively determine sample thermal conductivity depend on probe calibration. State of the art calibration strategies assume that the area of thermal ex-change between probe and sample does not vary with sample thermal conductivity. However, little investigation has gone into determining whether or not that assumption is valid. This dissertation provides a rigorous study into the probe-to-sample heat transfer through the air gap at diffusive distances for a variety of values of sample thermal conductivity. It is demonstrated that the thermal exchange radius and gap/contact thermal resistance varies with sample thermal conductivity as well as tip-to-sample clearance in non-contact mode. In contact mode, it is demonstrated that higher thermal conductivity samples lead to a reduction in thermal exchange radius for Wollaston probe tips.

Conversely, in non-contact mode and in contact mode for sharper probe tips where air contributes the most to probe-to-sample heat transfer, the opposite trend occurs. This may be attributed to the relatively strong solid-to-solid conduction occurring between probe and sample for the Wollaston probes. A three-dimensional finite element (3DFE) model was developed to investigate how the calibrated thermal exchange parameters vary with sample thermal conductivity when calibrating the probe via the intersection method in non-contact mode at diffusive distances. The 3DFE model was then used to explore the limits of sensitivity of the experiment for a range of simulated experimental conditions. It is determined that, when operating the scanning hot probe technique in air at standard temperature and pressure using Wollaston probes, the technique is capable of measuring, within 20% uncertainty, samples with values of thermal conductivity up to 10 Wm-1K-1 in contact mode and up to 2 Wm-1K-1 in non-contact mode. By increasing the thermal conductivity of the probe’s surroundings (i.e. changing air to a more conductive gas), sensitivity in non-contact mode to sample thermal conductivity is improved, which suggests potential for future investigations using non-contact scanning hot probe to measure thermal conductivity of higher thermal conductivity samples.