Heat transfer at the nanoscale has been one of the primary concerns in the design of nanoelectronics and nanostructured materials for applications such as thermal management and thermoelectric energy conversion. This thesis examines the thermal transport in nanoscale thin films and two-dimensional (2D) materials using an optical pump-probe technique based on frequency domain thermoreflectance (FDTR). The design and implementation of a continuous-wave laser based FDTR system is described in detail. The system is extended to an imaging microscope capable of producing micrometer scale maps of several thermophysical properties simultaneously. An analytical formula, which accounts for experimental noise and uncertainty in the controlled model parameters, is derived to calculate the precision of thermoreflectance measurements. The FDTR system is used to study the anisotropic heat conduction in periodic nanoscale Mo/Si superlattices and a 2D material, graphene. The measured in-plane thermal conductivity values of the superlattices are in good agreement with calculations taking into account both electron and phonon thermal transport, using a phonon mean free path which depends on the Mo layer thickness. The measurement procedure of graphene is described in detail, including the sample preparation, sensitivity analysis, and parameter fitting. Various graphene flakes supported on SiO2 surfaces and atomically flat Muscovite mica surfaces are measured. The results show that the thermal conductivity of single-layer graphene can be improved by ~3 times by using a mica substrate compared to commonly used SiO2 substrates. In addition, comparison with the reported values of suspended graphene suggest that the out-of-plane flexural phonon modes may contribute at least 70% to the thermal conductivity of graphene. Finally, the thermal model is modified to include volumetric heating for the measurement of materials without a transducer layer. An amorphous silicon film deposited on fused silica and silicon substrates is measured to validate the model.

Les composants HEMTs (High Electron Mobility Transistors) à base d’AlGaN/GaN sont à ce jour les candidats les plus prometteurs pour des applications hyperfréquences de puissance, dû essentiellement à leur forte densité de porteurs et des mobilités électroniques élevées. Cependant, la température générée en condition réelle est un paramètre capital à mesurer, afin d’estimer précisément la fiabilité des composants et leur durée de vie. Pour ces raisons, nous avons comparé les méthodes de caractérisation thermique par thermoréflectance et par spectroscopie Raman car elles sont non destructives et avec une résolution spatiale submicronique. Ces techniques ont déjà prouvé leur faisabilité pour la caractérisation thermique des transistors, en modes de fonctionnement continu et pulsé. Nous comparons dans cette étude leurs adaptabilité et performance dans le cadre de la réalisation d’un banc d’essai thermique dédié. Ces méthodes sont reconnues pour ne caractériser que certaines catégories de matériaux : les métaux pour la thermoréflectance et les semiconducteurs pour la spectroscopie Raman, ce qui nous a conduit à l’éventualité de les combiner. Nous avons confronté des résultats obtenus par thermoréflectance à partir des équipements de deux fabricants commercialisant cette méthode, nous permettant ainsi de mettre en évidence des résultats originaux sur des aspects et inconvénients qui ne sont pas relayés dans la littérature. Avec la spectroscopie Raman, nous avons identifié les paramètres de métrologie qui permettent de réaliser un protocole de mesure thermique le plus répétable possible, et nous présentons également une technique innovante pour sonder les matériaux en surface, à l'aide du même équipement, et notamment les métaux. / At the moment, AlGaN/GaN HEMTs (High Electron Mobility Transistors) are the most promising for high-power hyperfrequency applications, essentially due to their large carrier density and a high electronic mobility. However, the temperature generating during operational conditions is a crucial parameter to measure, in order to estimate the reliability and durability of components. For these reasons, we compared thermoreflectance and Raman spectroscopy, that are non-destructive and possessing a submicronic spatial resolution. These techniques have already proven their feasibility as thermal characterization methods in both continuous wave and pulsed operational modes. We compare here their adaptability and performance to the conception of a thermal test bench. These methods are known for characterizing specific types of material: metals for thermoreflectance and semiconductors for Raman spectroscopy, leading us to the eventuality to combine them. We compared several results measured by thermoreflectance method with equipment from two different manufacturers that commercialize this technology, so we could highlight some aspects and drawbacks that are note relayed in the literature. With Raman spectroscopy, we identified metrology parameters allowing to realize a thermal measurement setup as reproducible as possible, and we also present an innovative method to probe surface material, especially metals.

No description available.

Nanofabrication techniques continue to advance and are rapidly becoming the primary route to enhancement for the electrical, thermal, and optical properties of materials. The work presented in this dissertation details fabrication and characterization techniques of thin films and nanoparticles for these purposes. The four primary areas of research presented here are thermoelectric enhancement through nanostructured thin films, an alternative frequency-domain thermoreflectance method for thin film thermal conductivity measurement, thermal rectification in nanodendritic porous silicon, and plasmonic enhancement in silver nanospheroids as a reverse photolithography technique. Nanostructured thermoelectrics have been proposed to greatly increase thermopower efficiency and to bring thermoelectrics to mainstream power generation and cooling applications. In our work, thermoelectric thin films of SbTe, BiTe, and PbTe grown by atomic layer deposition and electrochemical atomic layer deposition were characterized for enhanced performance over corresponding bulk materials. Seebeck coefficient measurements were performed at temperatures ranging from 77 K to 380 K. Atomic composition was verified by energy-dispersive X-ray spectroscopy and structures were imaged by scanning electron microscopy. All thin films measured were ultimately found to have a comparable or smaller Seebeck coefficient to corresponding materials made by conventional techniques, likely due to issues with the growth process. Frequency-domain thermoreflectance offers a minimally invasive optical pump-probe technique for measuring thermal conductivity. Like time-domain thermoreflectance, the version of frequency-domain thermoreflectance demonstrated here relies on a non-zero thermo-optic coefficient in the sample, but uses moderate cost continuous wave lasers modulated at kHz or MHz frequencies rather than a more expensive ultrafast laser system. The longer timescales of these frequency ranges enables this technique to take measurements of films with thicknesses ranging from 100 nm to 10 um, complimentary to time-domain thermoreflectance. This method differentiates itself from other frequency-domain methods in that it is also capable of simultaneous independent measurements of both the in plane and out of plane values of the thermal conductivity in anisotropic samples through relative reflective magnitude, rather than phase, measurements. We validated this alternate technique by measuring the thermal conductivity of Al2O3 and soda-lime and found agreement both with literature values and with separate measurements obtained with a conventional time-domain thermoreflectance setup. Thermal rectification has the potential to enhance microcircuit performance, improve thermoelectric efficiency, and enable the creation of thermal logic circuits. Passive thermal rectification has been proposed to occur in geometrically asymmetric nanostructures when heat conduction is dominated by ballistic phonons. Here, nanodendritic structures with branch widths of ~ 10 nm and lengths of ~ 20 nm connected to ~ 50 um long trunks were electrochemically etched from <111> silicon wafers. Thermal rectification measurements were performed at temperatures ranging from 80 K to 250 K by symmetric thermal conductivity measurements. No thermal rectification was ultimately found in these samples within the margin of thermal conductivity measurement error 1%. This result is consistent with another study which found thermal rectification with greater conduction in the direction opposite to what ballistic phonon heat conduction theories predicted. Plasmonic resonance concentrates incident photon energy and enables channeling of that energy into sub-wavelength volumes where it can be used for nanoscale applications. We demonstrated that surface plasmon polaritons induced in silver nanosphereoid films by 532 nm light defunctionalize previously photocleaved ligands adsorbed onto the films, to yield a reverse photolithographic technique. In this method, gold nanosphere conjugation were conjugated to a photocleaved ligand, however conjugation could be inhibited by exposing the cleaved ligand to 532 nm light and consequently yield a reversal technique. This defunctionalizion effect did not occur on gold films or nanoparticles conjugated with the ligand in IR spectroscopy, and was observed to have a reduced effect in silver films relative to silver nanospheroid film. As silver nanospheroid films and gold nanospheres of the size used in this study are known to have plasmon resonance in the green wavelengths, while gold and silver continuous films do not, this defunctionalization likely results from plasmonic effects. / Ph. D.

Lors de deux dernières décennies, la nano-structuration a permis une augmentation conséquente des performances thermoélectriques. Bien qu’à l’ origine le silicium (Si) ait une faible efficacité thermoélectrique, son efficacité sous forme de nanostructure, et notamment de nanofils, a provoqué un regain d’intérêt envers la conduction thermique au sein de ces nanostructures de Si. Bien que la conductivité thermique y ait été réduite de deux ordres de grandeur, les mécanismes de conduction thermique y demeurent flous. Une meilleure compréhension de ces mécanismes permettrait non seulement d’augmenter l’efficacité thermoélectrique mais aussi d’ouvrir la voie à un contrôle des phonons thermiques, de manière similaire à ce qui se fait pour les photons. L’objectif de ce travail de thèse était donc de développer une plateforme de caractérisation, d’étudier le transport thermique au sein de différentes nanostructures de Si et enfin de mettre en exergue la contribution du transport cohérent de phonons à la conduction thermique. Dans un premier temps, nous avons développé un système de mesure allant de pair avec une procédure de fabrication en salle blanche. La fabrication se déroule sur le site de l’institut de Sciences Industrielles et combine des manipulations chimiques, de la lithographie électronique, de la gravure plasma et du dépôt métallique. Le système de mesure est base sur la thermoreflectance : un changement de réflectivité d’un métal a une longueur d’onde particulière traduit un changement de température proportionnel. Nous avons dans un premier temps étudié le transport thermique au sein de simples membranes suspendues, suivi par des nanofils, le tout étant en accord avec les valeurs obtenues dans la littérature. Le transport thermique au sein des nanofils est bien diffus, à l’exception de fils de moins de 4 μm de long a la température de 4 K ou un régime partiellement balistique apparait. Une étude similaire au sein de structures périodiques 1D a démontré l’impact de la géométrie et l’aspect partiellement spéculaire des réflexions de phonons a basse température. Une étude sur des cristaux phononiques (PnCs) 2D a ensuite montré que même si la conduction est dominée par le rapport surface sur vole (S/V), la distance inter-trous devient cruciale lorsqu’elle est suffisamment petite. Enfin, il nous a été possible d’observer dans des PnCs 2D un ajustement de la conductivité thermique base entièrement sur la nature ondulatoire des phonons, réalisant par-là l’objectif de ce travail. / In the last two decades, nano-structuration has allowed thermoelectric efficiency to rise dramatically. Silicon (Si), originally a poor thermoelectric material, when scaled down, to form nanowires for example, has seen its efficiency improve enough to be accompanied by a renewed interest towards thermal transport in Si nanostructures. Although it is already possible to reduce thermal conductivity in Si nanostructures by nearly two orders of magnitude, thermal transport mechanisms remain unclear. A better understanding of these mechanisms could not only help to improve thermoelectric efficiency but also open up the path towards high-frequency thermal phonon control in similar ways that have been achieved with photons. The objective of this work was thus to develop a characterization platform, study thermal transport in various Si nanostructures, and ultimately highlight the contribution of the coherent phonon transport to thermal conductivity. First, we developed an optical characterization system alongside the fabrication process. Fabrication of the structures is realized on-site in clean rooms, using a combination of wet processes, electron-beam lithography, plasma etching and metal deposition. The characterization system is based on the thermoreflectance principle: the change in reflectivity of a metal at a certain wavelength is linked to its change in temperature. Based on this, we built a system specifically designed to measure suspended nanostructures. Then we studied the thermal properties of various kinds of nanostructures. Suspended unpatterned thin films served as a reference and were shown to be in good agreement with the literature as well as Si nanowires, in which thermal transport has been confirmed to be diffusive. Only at very low temperature and for short nanowires does a partially ballistic transport regime appear. While studying 1D periodic fishbone nanostructures, it was found that thermal conductivity could be adjusted by varying the shape which in turn impacts surface scattering. Furthermore, low temperature measurements confirmed once more the specularity of phonon scattering at the surfaces. Shifting the study towards 2D phononic crystals (PnCs), it was found that although thermal conductivity is mostly dominated by the surface-to-volume (S/V) ratio for most structures, when the limiting dimension, i.e. the inter-hole spacing, becomes small enough, thermal conductivity depends solely on this parameter, being independent of the S/V ratio. Lastly, we were able to observe, at low temperature in 2D PnCs, i.e. arrays of holes, thermal conduction tuning based on the wave nature of phonons, thus achieving the objective of this work.

Laser nano-patterning with near-field optical microscope (NSOM) and the associated energy transport analysis are achieved in this study. Based on combined experimental/theoretical analyses, it is found that laser nano-patterning with a NSOM probes strongly depend on the laser conditions and material properties of the target: the energy transport from the NSOM probes to the targets changes from pure optical to a combination of thermal and optical transport when the pulse duration of laser is increased from femtosecond to nanosecond. As a result, the mechanisms of nano-pattern formation on targets changes from nano-ablation to nano-oxidation/ recrystallization when the laser pulse duration is increased from femtosecond to nanosecond. Also, with the laser nano-patterning experiments, thermal damage of NSOM probes is observed which can be attributed to the low transport efficiency (10-4 – 10-6) and associated heating of the metal cladding of NSOM probes. The heating of NSOM probes are studied with developed time harmonic and transient thermoreflectance (TR) imaging. From time harmonic TR when the NSOM probes are driven with continuous laser, it is found that the location of heating of NSOM probes is ~20-30µm away from the NSOM tip. The strength of the heating is determined by the laser power (linear dependence), wavelength of the laser (stronger with short A), and aperture size of NSOM probes (stronger when aperture size < A/2). From the transient TR imaging when the NSOM probes are driven with pulsed laser, it is found that the peak temperature of the NSOM probe shifts much closer to the tip. The possible reason for the change in the location of peak temperature when continuous laser is changed to pulsed laser can be attributed to the competition between the heat generation and dissipation rates at different location of the probe: the tip experiences highest temperature with pulsed heating as the entire heating processes is adiabatic. The tip also experiences highest heat dissipation rate due to its large surface-to-volume ratio which overcomes the heat generation at the tip under quasi-steady state resulting in shift of the hot spot. The knowledge obtained in this study can be important in the future design of more efficient NSOM probes and other nano-optic devices.

abstract: Mechanical behavior of metallic thin films at room temperature (RT) is relatively well characterized. However, measuring the high temperature mechanical properties of thin films poses several challenges. These include ensuring uniformity in sample temperature and minimizing temporal fluctuations due to ambient heat loss, in addition to difficulties involved in mechanical testing of microscale samples. To address these issues, we designed and analyzed a MEMS-based high temperature tensile testing stage made from single crystal silicon. The freestanding thin film specimens were co-fabricated with the stage to ensure uniaxial loading. Multi-physics simulations of Joule heating, incorporating both radiation and convection heat transfer, were carried out using COMSOL to map the temperature distribution across the stage and the specimen. The simulations were validated using temperature measurements from a thermoreflectance microscope. / Dissertation/Thesis / Masters Thesis Materials Science and Engineering 2016

In recent decades the development of technologies capable to offer highly localized and precise temperature control has received increasing attention due to their relevance and applicability in numerous engineering fields. Multiple scientific papers have been written that focus on the enhancement of the performance of thermoelectric materials and micro-devices. This Ph.D. thesis in the field of Mechanical Engineering pursues three main research goals regarding electrothermal phenomena: (1) To provide an optimization design strategy for micro-thermoelectric coolers by analyzing the interplay between electrical and thermal fluxes during device operation. (2) To analyze the suitability of a device, based on micro-thermoelectric coolers, for controlling the thermal environment in microbiological systems. (3) To develop an experimental technique, based on optical pump-probe thermal imaging, to extract the thermal conductivity anisotropy of thin films. For this purpose, numerical simulations and experiments were carried out. The results show, that the design of micro-thermoelectric devices must take into account the impact of parameters that are typically neglected in the construction of macro scale devices. Poorly designed parameters, such as the metallic contacts, the distance between thermoelectric elements and their interaction with the substrate, carry severe reductions of the performance of micro-thermoelectric devices. It is demonstrated that the optimal performance is achieved when the thermoelectric legs are properly dimensioned, so that a balance of the Fourier and Joule fluxes is reached. Numerical analyses prove that micro-thermoelectric coolers offer a feasible alternative to overcome the current spatial and temperature limitations of conventional technologies and therefore enable to investigate the thermal environment of biological systems at the micro-scale. Guidelines for the implementation of the experimental platform are provided. The evaluation of the numerical and experimental data proves that optical pump-probe thermal imaging is suitable to characterize both the in-plane and the through-plane thermal conductivity of thin films. The experimental conditions to extract the anisotropy of the sample under study are determined. The outcome of this work yields new insights into electrothermal phenomena at the micro-scale and thus creates new routes in the design, fabrication and characterization of micro- thermoelectric materials and devices.:Acknowledgements IV Erklärung der Urheberschaft VI Summary VII Zusammenfassung VIII Table of content IX List of figures XI List of tables XIV Abbreviations and symbols XV 1 Introduction 1 1.1 Motivation 1 1.2 Outline of the thesis 4 1.2.1 Chapter 2 - Fundamentals 4 1.2.2 Chapter 3 - Design guidelines of micro-thermoelectric coolers 4 1.2.3 Chapter 4 - Development of a platform for biological systems experimentation 4 1.2.4 Chapter 5 - Development of a technique for thermal transport characterization in thin films 5 1.2.5 Chapter 6 - Main conclusion and future research 5 1.3 Main research objectives 5 2 Fundamentals 7 2.1 Thermoelectric phenomena 7 2.2 Performance estimation of micro-thermoelectric coolers 10 2.3 Finite element modelling 12 2.3.1 Introduction to finite element modelling 12 2.3.2 Finite element modelling of thermoelectric phenomena 17 2.4 Thermoreflectance imaging microscopy 19 3 Design guidelines of micro-thermoelectric coolers 26 3.1 Introduction 26 3.2 Micro-thermoelectric coolers: an alternative for thermal management 28 3.3 Analysis approach 29 3.3.1 Input current optimization 31 3.3.2 Metallic contacts 32 3.3.3 Leg pair geometry 35 3.3.4 Fill factor 38 3.3.5 Experimental characterization of µTECs 41 3.4 Summary 44 4 Development of a platform for biological systems experimentation 46 4.1 Introduction 46 4.2 Thermal analysis on biological systems 48 4.3 Platform conceptual proposal 50 4.4 Analysis approach 52 4.4.1 Input current optimization 52 4.4.2 Fill material 54 4.4.3 Thermotaxis 55 4.4.4 Top material 56 4.4.5 Cold spot optimization 58 4.5 Experimental platform construction 59 4.6 Summary 62 5 Development of a technique for thermal transport characterization in thin films 64 5.1 Introduction 64 5.2 Thermal anisotropy characterization in thin films 65 5.3 Experimental apparatus 66 5.4 Experimental measurements 69 5.5 Analysis approach 72 5.5.1 Thermal conductivity anisotropy analysis 76 5.5.2 Effect of the laser power on the temperature distribution 79 5.5.3 Enhancement of the system sensitivity 80 5.6 Summary 83 6 Main conclusion and future research 85 6.1 Main conclusion 85 6.2 Outlook 88 7 References 89 8 Scientific output 97 8.1 Publications in peer review journals 97 8.2 Selected conference abstracts 98 9 Curriculum vitae 99

Many modern technologies are enabled by the use of thin films and/or nanostructured composite materials. For example, many thermoelectric devices, solar cells, power electronics, thermal barrier coatings, and hard disk drives contain nanostructured materials where the thermal conductivity of the material is a critical parameter for the device performance. At the nanoscale, the mean free path and wavelength of heat carriers may become comparable to or smaller than the size of a nanostructured material and/or device. For nanostructured materials made from semiconductors and insulators, the additional phonon scattering mechanisms associated with the high density of interfaces and boundaries introduces additional resistances that can significantly change the thermal conductivity of the material as compared to a macroscale counterpart. Thus, better understanding and control of nanoscale heat conduction in solids is important scientifically and for the engineering applications mentioned above. In this dissertation, I discuss my work in two areas dealing with nanoscale thermal transport: (1) I describe my development and advancement of important thermal characterization tools for measurements of thermal and thermoelectric properties of a variety of materials from thin films to nanostructured bulk systems, and (2) I discuss my measurements on several materials systems done with these characterization tools. First, I describe the development, assembly, and modification of a time-domain thermoreflectance (TDTR) system that we use to measure the thermal conductivity and the interface thermal conductance of a variety of samples including nanocrystalline alloys of Ni-Fe and Co-P, bulk metallic glasses, and other thin films. Next, a unique thermoelectric measurement system was designed and assembled for measurements of electrical resistivity and thermopower of thermoelectric materials in the temperature range of 20 to 350 °C. Finally, a commercial Anter Flashline 3000 thermal diffusivity measurement system is used to measure the thermal diffusivitiy and heat capacity of bulk materials at high temperatures. With regards to the specific experiments, I examine the thermal conductivity and interface thermal conductance of two different types of nanocrystalline metallic alloys of nickel-iron and cobalt-phosphorus. I find that the thermal conductivity of the nanocrystalline alloys is reduced by a factor of approximately two from the thermal conductivity measured on metallic alloys with larger grain sizes. With subsequent molecular dynamics simulations performed by a collaborator, and my own electrical conductivity measurements, we determine that this strong reduction in thermal conductivity is the result of increased electron scattering at the grain boundaries, and that the phonon component of the thermal conductivity is largely unchanged by the grain boundaries. We also examine four complex bulk metallic glass (BMG) materials with compositions of Zr₅₀Cu₄₀Al₁₀, Cu_46.25Zr_44.25Al_7.5Er₂, Fe₄₈Cr₁₅Mo₁₄C₁₅B₆Er₂, and Ti_41.5Zr_2.5Hf₅Cu_42.5Ni_7.5Si₁. From these measurements, I find that the addition of even a small percentage of heavy atoms (i.e. Hf and Er) into complex disordered BMG structures can create a significant reduction in the phonon thermal conductivity of these materials. This work also indicates that the addition of these heavy atoms does not disrupt electron transport to the degree with which thermal transport is reduced. / Ph. D.

Heat generated by electronic devices must be dissipated in order to ensure reliability and prevent device failure. In order to design devices properly, it is important to have precise knowledge of materials' thermal properties at the nano and micro scales. Here we present a series of experimental studies of heat transport for two different types of material: a two dimensional (2D) material such as MoS2 and micron scale particles. We used frequency domain thermoreflectance (FDTR) to conduct all thermal property measurements. This technique can measure thin film thermal properties as well as characterize the interface between two materials. Molybdenum disulfide (MoS2), a transition metal dichalcogenide, is a 2D material that has potential applications as a transistor in nanoelectronics due to its semiconductor properties. We studied cross plane thermal transport across exfoliated monolayer and few layer MoS2 deposited on two distinct substrates: SiO2 and Muscovite mica. The cross plane direction is critical in layer structure devices since the largest thermal resistances are found along this way. The results show enhanced thermal transport across monolayer MoS2 on both substrates indicating that monolayer MoS2 has superior thermal properties for its use in electronic devices. On the other hand, thermally conductive micro particles are used as fillers in composite materials in order to improve the thermal conductivity of the host or matrix material. They can be embedded in polymers for die attach applications as well as in metals to create more efficient heat sinks. We developed new FDTR based thermal models that apply to isolated particles as well as particles surrounded by another material. We tested the models with isolated diamond and silicon micron size particles and with diamond particles embedded in tin. We were able to obtain the thermal conductivity of individual particles, an effective particle volume and the thermal interface conductance between a particle and its surrounding matrix. This technique could have important applications in industry since it could be used to measure in situ the thermal interface conductance between particles and their matrix, often the highest thermal resistance in composite materials.