Two-dimensional (2D) materials have properties that are often different from their three-dimensional (3D) bulk form. Many of these materials are stable at ambient conditions, which allows them to be integrated with other 2D- or 3D-materials to form heterostructures. Integration of various dimensional materials attains unique electrical and optical properties that aid in developing novel electronic devices. The interface of the heterogeneous integration of these films can exhibit a weak van der Waals-like bonding. In this thesis, an advanced characterization (from atomic to millimeter resolution) of various dimensional materials with weakly bonded interfaces is developed and employed to understand their behavior at scale. First, a large-area single-crystal cadmium telluride thin film is grown incommensurately and strain-free to a sapphire substrate despite a significant 3.7% lattice mismatch. The film remarkably delaminates as a bulk single crystal film due to an atomically thin tellurium that spontaneously forms at the interface. Aberration-corrected electron microscopy and spectroscopy reveal both the van der Waals-like structure and bonding at the film/substrate interface. Second, a large-area atomically thin gallium is intercalated at the interface of epitaxial graphene. Correlative microscopy workflows are applied to understand the thickness uniformity and area coverage of the 2D–gallium over few millimeters of the sample. Utilizing multiple correlative methods, SEM image contrast is found to be directly related to the presence of the intercalated gallium. The origin of the SEM contrast is investigated as a function of the surface potential. Then, the heterostructure characterization is scaled up over a few square millimeter areas by segmenting SEM images, each acquired with nanometer-scale resolution. Additionally, transmission electron microscopy is applied to investigate the interface of gallium–SiC, the gallium air–stability, and the role of the substrate on the heteroepitaxial growth of 2D–gallium, which charts a path for further development of these materials. / Thesis / Doctor of Philosophy (PhD)

Le but de cette étude est de mieux appréhender les phénomènes régissant l’épitaxie du diamant élaboré par MPCVD assisté pas polarisation sur silicium. Elle s’articule en deux grands axes : une première approche qualitative sur la réactivité du substrat et une seconde approche plus quantitative relative à l’influence des étapes de prétraitement sur le dépôt. L’importance de la mise en œuvre d’un protocole rigoureux est mise en exergue dans ce travail. Le rôle des différentes étapes de préparation des substrats de silicium allant du nettoyage ex situ jusqu’à la croissance du diamant en passant par le nettoyage in situ par plasma H2 et la polarisation a été étudiée grâce à des techniques d’investigations telles que la microscopie électronique à balayage et à transmission, la spectroscopie de photoélectrons X, la microscopie à force atomique et la diffraction en faisceau rasant d’électrons haute énergie. Ces analyses ont mis en évidence l’apparition de structures de type « voids » associées à la formation de carbure de silicium dès l’étape de décapage hydrogène. La formation de carbone amorphe dans les premiers instants de la germination du diamant semble être corrélée à la quantité de diamant constatée après croissance, l’importance du flux d’ions pendant l’étape de polarisation sur les densités a été démontrée en parallèle. Des taux d’épitaxie jusqu'à près de 30% ont été atteints grâce à l’optimisation du couple temps/tension de polarisation ou à une carburation du substrat précédant la polarisation. La technique d’analyse de diffraction d’électrons en faisceau convergent sur des films de diamant s’est révélée adaptée à l’étude statistique de l’orientation des cristaux / The goal of this study is to better apprehend the phenomena governing bias assisted MPCVD diamond epitaxy on silicon. It is organized into two main parts: a first qualitative approach on the reactivity of the substrate and a second approach more quantitative related to the pretreatment stages influence on the deposit. The importance of a rigorous protocol is put forward in this work. The role of the various stages of silicon substrates preparation going from ex situ cleaning, in situ cleaning (plasma H2), bias until diamond growth was studied thanks to investigations techniques such as transmission and scanning electron microscopies, X-ray photoelectrons spectroscopy, atomic force microscopy, reflexion high energy electrons diffraction. These analyses highlighted the appearance of structures of the type “voids” associated with the silicon carbide formation from the hydrogen etching stage. The amorphous carbon formation in the first moments of diamond nucleation seems to be correlated with the diamond quantity measured after growth, the importance of ions flow during nucleation stage on the densities was shown in parallel. Epitaxy ratio up to 30% were reached thanks to the optimization of the couple time/biasing or carburation of the substrate preceding bias. The convergent beam electron diffraction analysis of diamond films appeared well adapted to the statistical study of crystals orientation

Semiconductors represent the enabling technology that underpins the many advances that define modern society. One semiconductor that shows considerable promise in the fabrication of new devices is zinc oxide (ZnO). A fundamental understanding of the properties of a material is required in order to exploit its properties. The behaviour of dopants and defects relevant to optoelectronic device fabrication is of particular interest. However, acceptor doping of ZnO is currently controversial, as successful and reproducible acceptor doping has not yet been achieved. Acceptor doping of ZnO using silver (Ag) is explored in this thesis to contribute towards the understanding of defect introduction in ZnO. In addition, there is also increasing interest in exploring materials with unconventional properties, commonly referred to as metamaterials, particularly for optical applications. The previously unexplored unique combination of Ag and ZnO may enable the fabrication of those devices. Several key factors that affect heteroepitaxy film quality, and ultimately its properties, are buffer layers and substrate temperature. A lattice match between sapphire and ZnO was provided by using buffer layers of 1 nm magnesium oxide (MgO) and 7.9 nm low temperature ZnO. The highest quality film was grown at the highest temperature (800°C), with rms roughness of 2.9 nm, carrier concentration of 3.6x10¹⁶ cm⁻³, and mobility of 105 cm²/Vs. In contrast, dopant (Ag) incorporation occurs more readily below 600°C, with dopant incorporation of up to 1020 cm⁻³ measured. Ag manifests as a deep acceptor (up to 94% substitutionally on Zn lattice sites), as evident from decreasing carrier concentration with increasing Ag flux, and DLTS measurements indicating an acceptor trap at 319 meV. This suggests that Ag is suitable for introducing compensation in ZnO, but Ag acceptors are not sufficiently shallow to result in p-type material. However, the unique combination of ZnO and Ag also enables the fabrication of a novel device, namely a superlens. Initial experimental results show the possibility of imaging a 100 nm line as 132 nm, compared with the diffraction-limited resolution of 332 nm for the same line feature.

Le diamant monocristallin est un candidat prometteur pour les applications en électronique de puissance et l'hétéroépitaxie est une alternative crédible à la synthèse de ce matériau. Lors de ce travail de thèse, chacune des étapes de la synthèse de films de diamant hétéroépitaxié sur des pseudo-substrats de Ir/SrTiO₃/Si(001) a été finement étudiée afin de progresser dans la reproductibilité et la qualité cristalline de ces films. Ainsi, un système de réflectométrie laser a été installé sur le bâti d'épitaxie d'iridium afin de caractériser in situ l'épaisseur des films réalisés. Une nouvelle méthode basée sur traitement plasma permettant la recristallisation de l'iridium à des températures comprises entre 800°C et 900°C a été mise au point et brevetée. Ensuite, une caractérisation de la surface de l'iridium après l'étape de nucléation du diamant (BEN) par ellipsométrie spectroscopique a été réalisée en bâtissant un modèle ellipsométrique à partir d'une étude séquentielle en MEB, AFM et XPS. Cette étude démontre que l’ellipsométrie est sensible à la formation des domaines qui contiennent les cristaux de diamant épitaxiés. L'étape de nucléation a été étendue à des pseudo-substrats Ir/SrTiO₃/Si(001) de 10x10 mm². Une stratégie d'épaississement des films de diamant reposant sur deux étapes a été adoptée. La structure cristalline des films épaissis à plusieurs centaines de microns a été caractérisée par DRX et Raman. Des diodes Schottky latérales ont été fabriquées sur l'un des substrats épaissis. Les mesures électriques réalisées démontrent l'homogénéité du substrat de diamant hétéroépitaxié. Afin de mieux contrôler les premiers stades de la croissance, une nouvelle méthode de nucléation sélective a été mise au point et brevetée. Son application permettrait dans l'avenir d'obtenir une croissance latérale (ELO) dès la coalescence des premiers cristaux de diamant. / Single crystal diamond is a promising candidate for power electronics applications and heteroepitaxy is a credible alternative for the synthesis of this material. During this PhD, each step from the synthesis of heteroepitaxial diamond films on Ir/SrTiO₃/Si(001) pseudo-substrates was studied in details to progress in the reproducibility and the crystalline quality of films. Thus, a laser reflectometry system was installed on the iridium epitaxy reactor to characterize in situ the thickness of these films. A new approach based on plasma treatment leading to the iridium recrystallization at temperatures between 800°C and 900°C was developed and patented. Then, a characterization of the iridium surface after diamond nucleation (BEN) by spectroscopic ellipsometry was done by building an ellipsometric model based on a sequential study by SEM, AFM and XPS. Results demonstrate that ellipsometry is sensitive to the formation of domains including epitaxial diamond crystals. The nucleation step was extented to Ir/SrTiO₃/Si(001) pseudo-substrates with a size of 10x10 mm². A strategy for the thickening of diamond films based on two steps was adopted. The crystalline structure of films, few hundreds of microns thick, was characterized by XRD and Raman. Lateral Schottky diodes were built on one of the thick substrates. Electrical measurements demonstrate the homogeneity of the heteroepitaxial diamond substrate. To better control the growth early stages, a new method of selective nucleation was developed and patented. Its application in the future would make possible a lateral growth (ELO) from the coalescence of the first diamond crystals.

High quality crystalline semiconductor films are a key component in the production of electronic and opto-electronic devices, however, the requirement of latticed-matched single crystal substrates for the epitaxy of a thin film limits the available material systems which can be developed commercially. This strict lattice matching requirement is relaxed for two-dimensional layered materials grown via van der Waals epitaxy. Unfortunately, the same low surface energies of these layered materials also suppress the growth of three-dimensional materials upon them, preventing direct large area single-crystal growth. The work presented in this thesis will demonstrate and investigate the spontaneous van der Waals epitaxy, driven by weak chemical interaction, of a three-dimensional material on a three-dimensional material system. Despite a 3.7% lattice mismatch, high quality CdTe can be heteroepitaxially deposited on α-Al2O3 but with an incommensurate interface which demonstrates weak adhesion between the film/substrate. This weak adhesion is exploited by developing a strain driven epitaxial thin film transfer and handling method, which causes deposited layers and structures to separate at the substrate interface for transfer to secondary carrier substrates without effecting the film properties and leaving the original substrate for subsequent use. Simple transferred thin film crystalline II-VI heterostructure devices on flexible substrates are demonstrated, without the need for selective chemical etch layers, ion-implantation or complex post-processes as required by conventional fabrication techniques. Following a growth study of GaAs on three oxide substrates, the phenomena of epitaxial registry with apparent weak interface adhesion is demonstrated for another system, GaAs/α-Al2O3, where a layer transferred heterostructure device on a flexible substrate is also demonstrated. / Thesis / Doctor of Philosophy (PhD)

No description available.

Nous étudions ici l’heteroepitaxie du silicium-germanium (SiGe), un système qui est couramment considéré comme le stéréotype de l’´épitaxie des semi-conducteurs. Bien que ce système ait déjà attiré une attention considérable en raison de ses applications pour l’ingénierie des bandes dans l’industrie microélectronique, le défi majeur du développement de nouveaux dispositifs à base de SiGe reste la croissance épitaxiale contrôlable des nanostructures auto-assemblées. Il est bien connu que SiGe suit un mode de croissance de Stranski-Krastanov, qui passe par la croissance de couches bidimensionnelles suivie par la croissance d’ılots tridimensionnels. Sous cette dénomination générique ”Stranski-Krastanov”, plusieurs comportements différents peuvent être identifiés. Une compréhension globale de tous ces comportements est encore partiellement manquante en raison de la complexité et de l’interaction de la cinétique et des forces motrices dynamiques, empêchant le d´développement de nouveaux dispositifs. Dans ce travail, nous nous concentrons sur l’auto- assemblage des nanostructures SiGe à la suite de la quête de l’émission de lumière pour les dispositifs photoniques, optoélectroniques et nanoélectroniques à base de Si. Par Même si l’innovation dans les dispositifs à base de Si a été stimulée récemment par le d´développement de silicium complétement épuisé sur les transistors isolants, une véritable percée serait la démonstration de l´émission de lumière et / ou l’absorption par les éléments du groupe IV, car il permet une intégration pratique dans les semi-conducteurs actuels. Dans ce travail, nous montrons d’abord les différents régimes de croissance des films contraints, c’est-à-dire l’instabilité par rapport aux régimes de nucléation. Nous d´développons un modèle qui résout la course de ces deux voies de croissance et d´dévoile les mécanismes des différents modes d’évolution morphologique entrainés par l’élasticité. Dans la seconde partie, nous examinons en détail l’auto-organisation naturelle des îles cohérentes. L’effet élastique direct induit la répulsion entre les îles cohérentes. Cependant, l’énergie de surface dépendant de la déformation qui a été négligée précédemment dans l’analyse de l’interaction île-île est révélée pour provoquer une attraction entre les iles. Il peut compenser la répulsion élastique directe au cours de l´état initial de la nucléation et conduire au regroupement d’îlots cohérents. Dans une troisième partie, nous étudions l’influence des échelons du substrat vicinal sur la formation et l’auto-organisation des îles. Nous démontrons que l’anisotropie de relaxation de la contrainte produite par les bords des gradins est à l’origine de l’allongement de l’instabilité perpendiculaire aux marches. Un accord quantitatif entre l’allongement de l’instabilité et l’anisotropie de relaxation de la souche est trouvé, ce qui approfondit les compréhensions de la croissance hétéroépitaxiale sur le substrat vicinal. Dans la quatrième partie, nous développons un nouveau procédé basé sur la condensation Ge lors de l’oxydation thermique du SiGe dilué. On étudie la cinétique du procédé de condensation SiGe et on fabrique la couche épandeuse de SiGe totalement contrainte par ce procédé de condensation particulier. / We investigate here the heteroepitaxy of silicon-germanium (SiGe), a system which is commonly regarded as the stereotype of semiconductor epitaxy. While this system has already attracted a tremendous amount of attention due to its applications for band-gap engineering in microelectronic industry, the major challenge facing the development of new SiGe-based devices remains the con- trollable epitaxial growth of self-assembled nanostructures. It is well-known that SiGe follows a Stranski-Krastanov growth mode, which proceeds via the growth of bi-dimensionnal layers followed by the growth of three-dimensional islands. Under this generic “Stranski-Krastanov” designation, several different behaviors can be identified. An overall understanding of all these behavior is still partially missing due to the complexity and the interplay of kinetics and energetic driving forces, preventing the development of new devices.In this work we focus on the self-assembly of SiGe nanostructures following the quest of light emission for integrated Si-based photonic, optoelectronic and nanoelectronic devices.Even if the innovation in Si-based devices has been boosted recently by the development of ultra-thin body fully depleted silicon on insulator transistors, a real breakthrough would be the demonstration of light emission and/or absorption by group IV elements since it allows the conve- nient integration into the nowadays semiconductors.In this work we first demonstrate the different growth regimes of strained films, i.e. instability versus nucleation regimes. We develop a model which resolves the race of these two growth pathways and unveil the mechanisms of different modes of morphological evolution driven by elasticity.In the second part, we examine in details the natural self-organisation of coherent islands. The direct elastic effect induces repulsion between coherent islands. However, the strain-dependent surface energy which has been overlooked previously in analysis of the island-island interaction is revealed to cause an attraction between islands. It may compensate the direct elastic repulsion during the initial state of nucleation and lead to the clustering of coherent islands.In a third part we study the influence of miscut steps of vicinal substrate on the formation and self-organisation of islands. We demonstrate that the strain relaxation anisotropy produced by the step edges, is at the origin of the instability elongation perpendicular to steps. Quantitative agreement between the instability elongation and the anisotropy of strain relaxation is found, which deepens the understandings of hetero-epitaxial growth on vicinal substrate.In the fourth part we develop a new process based on Ge condensation during thermal oxidation of dilute SiGe. The kinetics of SiGe condensation process is investigated and the fully strained SiGe epilayer is fabricated via this particular condensation process. This process can be applied in fabrication of SiGe core-shell nanostructures, for which the direct deposition and growth process is found to be cumbersome in terms of the control of morphology and composition.As a whole, we studied the nanostructures of SiGe driven by its hetero-epitaxial elasticity. We proposed a model to compare two pathways of morphological evolution of SK growth and unearthed the mechanisms of the race and transition. We studied kinetics of island nucleation under the impact of elastic filed produced by an existing island. The peculiar role of strain-dependent surface energy is highlighted. Then the elasticity anisotropy induced by miscut steps on vicinal substrate is studied theoretically and experimentally. This anisotropy effectively induces the elongation of islands in one direction to form nanowires in good alignment. Then the kinetics of condensation of SiGe is studied, which is found to be an effective method in fabricating strained SiGe nanostructures.

The heteroepitaxial growth of cubic silicon carbide (3C-SiC) on (111) silicon (Si) substrates, via a horizontal hot-wall chemical vapor deposition (CVD) reactor, has been achieved. Growth was conducted using a two step process: first the Si substrate surface is converted to SiC via a carbonization process and second the growth of 3C-SiC is performed on the initial carbonized layer. During carbonization, the surface of the Si is converted to 3C-SiC, which helps to minimize the stress in the growing crystal. Propane (C3H8) and silane (SiH4), diluted in hydrogen (H2), were used as the carbon and silicon source, respectively. A deposition rate of approximately 10 µm/h was established during the initial process at a temperature of ~1380 °C. The optimized process produced films with X-ray rocking curve full-width at half-maximum (FWHM) values of 219 arcsec, which is significantly better than any other published results in the literature. Once this process was developed a lower temperature process was developed at a slower growth rate of ~2 µm/h at 1225 °C. The crystal quality was inferior at the reduced temperature but this new process allows for the growth of 3C-SiC(111) films on oxide release layers for MEMS applications. In addition, for electronic device applications, a lower temperature process reduces the generation of defects caused by the nearly 8 % mismatch in the coefficient of thermal expansion (CTE) between 3C-SiC and Si. Finally a new process using a poly-Si seed layer deposited on an oxide-coated Si wafer was used to form 3C-SiC films for MEMS applications. The results indicated initially that the films may even be monocrystalline (based on X-ray evaluation) but later analysis performed using TEM indicated they were highly-ordered polycrystalline films. The grown 3C-SiC films were analyzed using a variety of characterization techniques. The thickness of the films was assessed through Fourier Transform infrared (FTIR) spectroscopy, and confirmed (in the case of growth on poly-Si seed layers) by cross-section scanning electron microscopy (SEM). The SEM cross-sections were also used to investigate the 3C-SiC/oxide interface. The surface morphology of the films was inspected via Nomarsky interference optical microscopy, atomic force microscopy (AFM), and SEM. The crystalline quality of the films was determined through X-ray diffraction (XRD).

The heteroepitaxial growth of cubic silicon carbide (3C-SiC) on silicon (Si) substrates at high growth rates, via a horizontal hot-wall chemical vapor deposition (CVD) reactor, has been achieved. The final growth process was developed in three stages; an initial "baseline" development stage, an optimization stage, and a large area growth stage. In all cases the growth was conducted using a two step, carbonization plus growth, process. During carbonization, the surface of the Si is converted to 3C-SiC, which helps to minimize the stress in the growing crystal. Propane (C3H8) and silane (SiH4), diluted in hydrogen (H2), were used as the carbon and silicon source, respectively. A deposition rate of approximately 10 um/h was established during the baseline process. Once the baseline process proved to be repeatable, optimization of the process began. Through variations in temperature, pressure, and the Si/C ratio, thick 3C-SiC films (up to 22 um thick) and high deposition rates (up to 30 um/h) were obtained. The optimized process was then applied to growth on 50 mm diameter Si(100) wafers. The grown 3C-SiC films were analyzed using a variety of characterization techniques. The thickness of the films was assessed through Fourier Transform infrared (FTIR) spectroscopy, and confirmed by cross-section scanning electron microscopy (SEM). The SEM cross-sections were also used to investigate the 3C-SiC/Si interface. The surface morphology of the films was inspected via Nomarsky interference optical microscopy, atomic force microscopy (AFM), and SEM. The crystalline quality of the films was determined through X-ray diffraction (XRD) and low-temperature photoluminescence (LTPL) analysis. A mercury probe was used to make non-contact CV/IV measurements and determine the film doping.

Ce travail est consacré à l'hétéroépitaxie de GaN sur substrat Si, étudié à l’échelle nanométrique par microscopie électronique en transmission. On étudie dans un premier temps la croissance de la couche de nucléation d’AlN, en analysant la structure de son interface avec Si, et les premières étapes de croissance. Ceci est effectué en comparant les procédés MBE et MOCVD, pour lesquels des différences sont observées, comme la formation d’une couche amorphe interfaciale pour la MOCVD à haute température. Ensuite, la nanodiffraction par précession (N-PED) permet de déterminer la distribution de la déformation sur l’ensemble de l’hétérostructure. Avec l’aide de l’imagerie par microscopie TEM conventionnelle, on corrèle ainsi le comportement des dislocations traversantes (boucles et dislocations inclinées) à la relaxation des couches de GaN. Une croissance hybride basée sur une combinaison de couches MBE et MOCVD se révèle efficace quant à la réduction de la densité de dislocations, et est à priori intéressante pour des applications type LED. L’utilisation de la cathodoluminescence permet de mettre en évidence la présence d’impuretés et leurs effets au cours des différentes étapes de l’épitaxie. Finalement, nous explorons la possibilité d’utiliser une couche de Si fine sur isolant (SOI) comme substrat de type compliant pour la croissance des couches de GaN. Cette dernière étude est menée principalement pour des couches MBE, et demande à être étendue à des structures MOCVD, pour lesquelles les effets de compliances observés seront potentiellement plus importants. / This work is dedicated to GaN on Si heteroepitaxy, at nanometer scale by transmission electron microscopy. First we study the AlN buffer layer growth, by analysing its interface structure with Si substrate, and the first growth stages. This is done by comparing MBE and MOCVD growth processes, for which differences are observed, as the formation of amorphous inter-layer for high temperature MOCVD. Then nanodiffraction with precession (N-PED) allows us to determine strain distribution among the entire heterostructure. With the help of conventional TEM imaging, we correlate threading dislocation behaviour (loop and inclined dislocations) to the strain relaxation in GaN. Hybrid growth based on a combination of MBE and MOCVD layers appears to be quite efficient concerning the reduction of dislocation density, and thus interesting for LED applications. Cathodoluminescence highlights presence of impurities and their effect for each epitaxy stages. Finally we explore the possibility to grow GaN epilayers on a thin Si layer on insulator (SOI), as a compliant substrate. This last study is mainly conducted for MBE layers, and requires to be extended to MOCVD structures, for which observed compliant effects may potentially be more important.