Ce travail porte sur l'étude des étapes initiales du dépôt de couches atomiques de ZnO (ALD) sur une surface (100) de In0,57Ga0,43As, par l'utilisation de techniques de caractérisation in situ (rayonnement synchrotron). En raison de la grande mobilité des électrons, le semi-conducteur III-V InGaAs est un matériau potentiel pour remplacer le canal de Silicium dans les transistors à effet de champ (MOSFET). Afin de diminuer la hauteur de la barrière Schottky et la résistance de contact, une couche ultra-mince (tunnel) de ZnO peut être insérée entre le métal et le semiconducteur InGaAs. Au cours de ces dernières années, la technique ALD, compatible avec les spécifications de l'industrie et basée sur des réactions chimiques de surface auto-limitantes, est utilisée pour la fabrication de films minces conformes et homogènes avec un contrôle sub-nanométrique de l’épaisseur. Cependant, le comportement au cours de la croissance diffère fortement en fonction de la surface du substrat. Ainsi, l'étude des premières étapes ALD est particulièrement intéressante afin d’améliorer la compréhension des mécanismes de croissance en vue de la création de films ultra-minces.Pour ce faire, nous avons développé et mis à niveau un réacteur thermique ALD (MOON) dédié. Il peut être installé sur des lignes de lumière synchrotron afin d’étudier la croissance des matériaux in situ avec des techniques telles que la fluorescence X, l’absorption X, la spectroscopie des rayons X ainsi que la diffraction X en incidence rasante. De plus, des techniques optiques de caractérisation in situ peuvent être utilisées en laboratoire ou couplées en milieu synchrotron. Les expériences au synchrotron ont été réalisées sur les lignes de lumière SIRIUS (SOLEIL, Saint-Aubin (France)) et ID3 (ESRF, Grenoble (France)).Nous montrons que dans sa phase initiale, la croissance ALD de ZnO est inhibée par le substrat (100) InGaAs, ce qui conduit à un régime transitoire avant le régime de croissance ALD stable. La première phase du régime transitoire conduit à la formation d’une couche d’oxyde de Zinc, ultra-mince (~1 nm d'épaisseur), fabriquée avec un taux de croissance très faible. L'absorption X et la diffusion X en incidence rasante montrent qu’à ce stade le matériau ZnO est désordonné (non cristallisé) et présente un ordre à courte distance caractérisé par une structure wurtzite embryonnaire. Ensuite, le régime transitoire entre dans une deuxième phase (croissance 3D), le taux de croissance par cycle (GPC) augmente, atteint un maximum puis diminue jusqu'à une valeur constante (croissance ALD stable). Afin de mieux comprendre le mode de croissance 3D nous avons développé un modèle géométrique qui schématise la croissance d’îlots hémisphériques par ALD. Ce modèle permet d'obtenir des paramètres quantitatifs de croissance.En modifiant le débit d’eau (H2O) utilisée comme réactif pendant le processus ALD, il est possible de contrôler le délai (ou le nombre de cycles) avant le début de la croissance 3D. Cet effet est très probablement lié à la variation de la densité des groupes hydroxyle à la surface de l'InGaAs. Par ailleurs, nous avons caractérisé la croissance ALD de ZnO pour différentes températures du substrat InGaAs (dans et hors fenêtre ALD). Les cartes de diffusion des RX réalisées en cours de dépôt, montrent l’apparition d’une phase cristallisée à longue distance en lien avec le démarrage de la croissance 3D. À température élevée, hors de la fenêtre ALD, nous observons une texturation de la couche ZnO lorsque son épaisseur augmente. Aucune relation d’épitaxie n’est observée.Enfin, nous rendons compte de l'utilisation de couches ZnO ultraminces sur InGaAs pour les contacts électriques. La résistance de contact des échantillons de métal/ZnO/InGaAs a été mesurée à l'aide de la méthode Transfert Length Method (TLM). Nous montrons que la résistivité de contact spécifique des tampons Al/p-InGaAs est réduite par l’insertion d’une couche tunnel ZnO entre l'Al et l'InGaAs dopé p. / This work focuses on the study of the initial stages of ZnO atomic layer deposition (ALD) on atomically flat (100) In0.57Ga0.43As surface, notably by using in situ synchrotron techniques. Due to high electron mobility, III-V InGaAs semiconductor has been recognized as a promising material to replace Silicon channel in the metal-oxide-semiconductor-field-effect transistors (MOSFET). Ultrathin ZnO layer on InGaAs can be used as a passivation layer at the interface with the gate transistor dielectric, as well as tunneling layer inserted in between metal/InGaAs contact to decrease the Schottky barrier height and the contact resistance. In the recent years, ALD technique based on self-limiting surface chemical reactions has received world-wide attention for manufacturing highly conformal and homogeneous thin films with sub-nanometer thickness control at low temperatures compatible with industry specifications. However, the growth behavior strongly differs depending on the substrate surfaces. Thus for the creation of few monolayers thick films, the study of ALD in the initial stages of growth is of particular interest for improving the understanding of the growth mechanisms.For that purpose, we have developed and upgraded a thermal ALD reactor (MOON:MOCVD/ALD growth of Oxide Nanostructures) dedicated to monitor the growth of materials by in situ characterization techniques. The MOON reactor can be moved to synchrotron centers for monitoring material growth in situ by using X-ray based techniques, notably X-ray fluorescence, X-ray absorption, XRR, and grazing incidence diffraction. Also, optical in situ techniques can be used in the laboratory. In this work, we show the results of experiments obtained at two synchrotron beamlines, i.e. SIRIUS (SOLEIL, Saint-Aubin (France)) and ID3 (ESRF, Grenoble (France)).We show that ZnO growth in the initial stages is inhibited by the (100) InGaAs substrate, leading to a transient regime prior to the steady ALD is achieved. We report a detailed investigation of this transient regime and find that an ultra-thin (~1-nm-thick) 2D layer is indeed fabricated but with a growth rate so low that one may believe that nothing has been deposited on the surface. We identify the structural and chemical properties of that ultra-thin layer. Only afterward does the substrate inhibited of type 2 growth mode begins: as the cycle number increases, the growth per cycle (GPC) increases, then reaches a maximum and level down to a constant value (steady growth). For a better understanding of the 3D growth mode by reproducing the experimental growth per cycle curves we have developed a geometric model that schematizes the growth of hemispheroid islands by ALD. We show that this model allows obtaining quantitative growth parameters.When water is used as a reactant, we showed that by changing the water flow during the ALD process, it is possible to control the time delay (or cycle number) prior to 3D growth begins. It is very likely that the water flow controls the density of hydroxyl groups on the InGaAs surface. We also demonstrated ZnO ALD for different InGaAs substrate temperatures. By combining in situ X-ray absorption and grazing incidence scattering techniques, we identified a short-range-order atomic structure of the ZnO material, with an embryonic ZnO wurtzite, prior to 3D growth, then a long-range-order structure is detected both by X-ray absorption and X-ray diffraction, together with the appearance of a microstructure. At higher growth temperature, outside of the ALD window, we observed the well-known ZnO texturing when the layer thickness increases.At last, we report on the use of ultrathin ZnO layers on InGaAs in the electrical contact structure. The contact resistance of metal/ZnO/InGaAs samples was measured using Transfer Length Method (TLM). We show that specific contact resistivity of Al/p-InGaAs pads is reduced by inserting a ZnO tunnel layer in between Al and p-doped InGaAs.

No description available.

<p>Thin films of cobalt oxide, iron oxide and niobium oxide, and nanostructured thin films of iron oxide, titanium oxide and multilayered iron oxide/titanium oxide have been deposited by Atomic Layer Deposition (ALD). The metal oxides were grown using the precursor combinations CoI2/O2, Fe(Cp)2/O2, NbI5/O2 and TiI4/H2O. The samples were analysed primarily with respect to phase content, morphology and growth characteristics.</p><p>Thin films deposited on Si (100) were found to be amorphous or polycrystalline, depending on deposition temperature and the oxide deposited; cobalt oxide was also deposited on MgO (100), where it was found to grow epitaxially with orientation (001)[100]Co3O4||(001)[100]MgO. As expected, the polycrystalline films were rougher than the amorphous or the epitaxial films. The deposition processes showed properties characteristic of self-limiting ALD growth; all processes were found to have a deposition temperature independent growth region. The deposited films contained zero or only small amounts of precursor residues.</p><p>The nanostructured films were grown using anodic aluminium oxide (AAO) or carbon nanosheets as templates. Nanotubes could be manufactured by depositing a thin film which covers the pore walls of the AAO template uniformly; free-standing nanotubes retaining the structure of the template could be fabricated by removing the template. Multilayered nanotubes could be obtained by depositing multiple layers of titanium dioxide and iron oxide in the pores of the AAO template. Carbon nanosheets were used to make titanium dioxide nanosheets with a conducting graphite backbone. The nucleation of the deposited titanium dioxide could be controlled by acid treatment of the carbon nanosheets.</p>

In-based III-V compound semiconductors have higher electron mobilities than either Si or Ge and direct band gaps. These properties could enable the fabrication of low power, high-speed n-channel metal oxide semiconductor field effect transistors (MOSFETs) and optoelectronics combining MOS technology with photonics. Since thermal and native oxides formed on III-V surfaces exhibit large current leakage and high densities of trap states, a key to incorporating these materials into advanced devices is the development of processing steps that form stable interfaces with dielectric layers. In this thesis, a processing flow consisting of native oxide removal using HF chemistries and deposition of high dielectric permittivity films using atomic layer deposition was investigated. Understanding the reaction mechanisms of these processes could provide the means of controlling composition and structure, yielding a desired electronic behavior. Quantitative X-ray photoelectron spectroscopy analysis of surfaces was coupled with electrical measurements on MOS capacitors of the interface quality in order to understand the nature of high-k/III-V interface defects and their repair. Ex situ liquid phase HF etching removed InSb, InAs, and InGaAs(100) native oxides and produced an Sb- or As-enriched surface, which oxidized when exposed to air. A 5 to 22 °A thick As- and Sb-rich residual oxide was left on the surface after etching and < 5 min of air exposure. The results showed that group V enrichment originated from the reduction of group V oxides by protons in the solution and the preferential reaction of HF with the group III atom of the substrate. A sub-atmospheric in situ gas phase HF/H2O process removed native oxide from InSb, InAs, and InGaAs(100) surfaces, producing an In or Ga fluoride-rich sacrificial layer. A 50 to 90% oxide removal was achieved and a 10 to 25 °A-thick overlayer consisting of mainly In and Ga fluorides was produced. The composition and morphology of the sacrificial layer were controlled by the partial pressure of H2O as well as the ratio of HF to H2O used. Water played a critical role in the process by directly participating in the etching reaction and promoting the desorption of fluoride etching products. Accumulation of thick fluoride layer at high HF to water partial pressure ratios prevented adsorption and diffusion of etchant to the buried residual oxide. When oxide was removed, HF preferentially reacted with In or Ga atoms from the substrate, enriching the surface with group III fluorides and producing approximately one monolayer of elemental group V atoms at the interface. Interface reactions occurred during atomic layer deposition of Al2O3, in which trimethylaluminum (TMA) removed surface oxides and fluorides. Chemically sharp InSb/Al2O3 and InGaAs/Al2O3 interfaces were achieved for gas phase HF-etched InSb and liquid phase HF-etched InGaAs. A ligand transfer mechanism promotes nucleation of Al2O3 and removal of III-V atoms from the sacrificial oxide and fluoride layers as volatile trimethyl indium, gallium, arsenic, and antimony. These reactions have been explained by the relative bond strength of surface and precursor metal atoms with O and F. Interaction of a InSb(100) surface with TiCl4 as a model for metal halide ALD precursors showed that similar ligand transfer reactions occured. Adsorbed chlorine from the dissociative adsorption of TiCl4 on the InSb surface at elevated temperature, however, preferentially etched In atoms from the substrate and produced a roughened surface. The quality of InGaAs/Al2O3 interfaces prepared by solvent cleaning and liquid phase HF were assesed electrically using capacitance-voltage and conductance measurements. Surface recombination velocity (SRV) values were extracted from the measurements to represent the net effect of interface defects, which includes defect density and capture cross section. The InGaAs/Al2O3 interface prepared by solvent cleaning consisted of interfacial native oxides while that etched in liquid phase HF consisted of submonolayer arsenic oxide. The two chemically contrasting interfaces, however, gave similar SRV values of 34.4±3.7 and 28.9±13.4 cm/s for native oxide and liquid phase HF prepared samples, respectively. This suggests that the presence or absence of oxides was not the only determining factor. Post Al2O3 deposition annealing in forming gas and NH3 ambient significantly improved the electrical quality for both surfaces, as shown by SRV values between 1 to 4 cm/s which is comparable to that of an ideal H-terminated Si surface. XPS analysis showed that the contribution from elemental As and Ga2O3 at the interface of both surfaces increased after annealing in forming gas and NH3, likely due to thermal or hydrogen-induced reaction between interfacial As oxide and Ga atoms in the substrate. There was no correlation between the atomic coverages of interfacial elemental As and oxides to the SRV values. High activity defects at III-V/Al2O3 interfaces are associated with interfacial dangling bonds which were passivated thermally and chemically by annealing in forming gas and NH3.

An intracranial aneurysm is a local dilation of an artery in the cerebral circulation. While the etiology of intracranial aneurysms is unknown, they likely result from a combination of factors including the weakening and degeneration of the collagen fibers and the internal elastic lamina comprising the arterial wall, as well as hemodynamic-associated stress resulting from blood pulsation inside the aneurysm sac. Intracranial aneurysm rupture leads to a devastating sequela, as 50% of patients die. In the U.S. alone there are approximately 30,000 cases of subarachnoid hemorrhage annually, a prevalence which has pushed practitioners to aggressively treat the aneurysm disease. Traditionally, intracranial aneurysms were managed with open craniotomy and microsurgical clipping; however, these treatment modalities carry relatively high morbidity and mortality depending upon the aneurysm location and surgical experience. In 2002 the International Subarachnoid Hemorrhage Aneurysm Trial established the superiority of the endovascular coiling of intracranial aneurysms compared to microsurgical clipping. This trial led to a paradigm shift in treating intracranial aneurysms with marked use of intracranial stenting, including devices used to assist endovascular coiling and stand-alone flow diverting devices. However, the placement of intracranial devices in the cerebral circulation mandates the adjunctive application of dual anti-platelet pharmaceuticals to minimize thromboembolic events, despite being associated with increased patient risk. This dissertation proposes a novel multilayer, nanometer-scale coating technology suitable for commercially available intracranial stents and flow diverting devices to minimize the use of dual anti-platelet therapy in the elective setting and expand the use of intracranial devices in the acute setting of ruptured intracranial aneurysms. A combination of qualitative and quantitative chemical characterization techniques was used to assess the composition, uniformity, and thickness of each coating layer on commercially available flow diverting devices; overall the coating was found to be relatively uniform and conformal to the device wires. Furthermore, in-vitro and in-vivo testing on commercially available intracranial devices suggest some hemocompatible and antithrombotic properties. Finally, the proposed coating technology can be modified for use as a platform for the attachment of FDA-approved molecules. With further optimization and testing this technology has the potential to minimize the adjunctive use of dual-antiplatelet therapy in the endovascular treatment of intracranial aneurysms.

In the last 50 years, the semiconductor industry has been scaling the silicon transistor to achieve faster devices, lower power consumption, and improve device performance. Transistor gate dimensions have become so small that short channel effects and gate leakage have become a significant problem. To address these issues, performance enhancement techniques such as strained silicon are used to improve mobility, while new high-k gate dielectric materials replace silicon oxide to reduce gate leakage. At some point the fundamental limit of silicon will be reached and the semiconductor industry will need to find an alternate solution. The advent of graphene led to the discovery of other layered materials such as the transition metal dichalcogenides. These materials have a layered structure similar to graphene and therefore possess some of the same qualities, but unlike graphene, these materials possess sizeable bandgaps between 1-2 eV making them useful for digital electronic applications. Since initially discovered, most of the research on these films has been from mechanically exfoliated flakes, which are easily produced due to the weak van der Waals force binding the layers together. For these materials to be considered for use in mainstream semiconductor technology, methods need to be explored to grow these films uniformly over a large area. In this research, atomic layer deposition (ALD) was employed as the growth technique used to produce large area uniform thin films of several different transition metal dichalcogenides. By optimizing the ALD growth parameters, it is possible to grow high quality films a few to several monolayers thick over a large area with good uniformity. This has been demonstrated and verified using several physical analytical tests such as Raman spectroscopy, photoluminescence, x-ray photoelectron spectroscopy, x-ray diffraction, transmission electron spectroscopy, and scanning electron microscopy, which show that these films possess the same qualities as those of the mechanically exfoliated films. Back-gated field effect transistors were created and electrical characterization was performed to determine if ALD grown films possess the same electronic properties as films produced from other methods. The tests revealed that the ALD grown films have high field effect mobility and high current on/off ratios. The WSe2 films also exhibited ambipolar electrical behavior making them a possible candidate for complementary metal-oxide semiconductor (CMOS) technology. Ab-initio density functional theory calculations were performed and compared to experimental properties of MoS2 and WSe2 films, which show that the ALD films grown in this research match theoretical predictions. The transconductance measurements from the WSe2 devices used, matched very well with the theoretical calculations, bridging the gap between experimental data and theoretical predictions. Based upon this research, ALD growth of TMD films proves to be a viable alternative for silicon based digital electronics.

Les MEMS RF sont des composants clés pour le développement de nombreuses fonctions de systèmes hyperfréquences plus efficaces et plus compactes (déphaseurs, module transmission/réception, réseau d’antennes à déphasage, circuits reconfigurables, réseau d’adaptation …). Pour le développement des prochaines générations de systèmes RADAR, Thales s’intéresse notamment à l’intégration de MEMS RF capacitifs pour développer des fonctions reconfigurables pouvant supporter des puissances hyperfréquences de l’ordre de 30 W. Les travaux exposés dans ce manuscrit se sont concentrés sur l’étude de matériaux diélectriques et de techniques de dépôts pour identifier, intégrer et démontrer la viabilité de diélectriques prometteurs pour les MEMS RF capacitifs de puissance. Les aspects relatifs à la fabrication de ces composants ont également été étudiés, particulièrement l’impact de la maitrise des états de surface sur les performances, la tenue en puissance et la défaillance des dispositifs. En outre, ces travaux ont montrés qu’avec l’introduction des matériaux déposés par ALD, la tenue en puissance des MEMS RF capacitifs n’est plus limitée par le diélectrique. En intégrant ces matériaux ALD, l’architecture des dispositifs devient le facteur limitant la tenue en puissance, particulièrement l’épaisseur de la membrane et la configuration du commutateur. En perspectives, différentes architectures ont donc été développées et étudiées pour adresser ces limitations de tenue en puissance. / RF MEMS are key components to improve the efficiency and size of numerous functions of microwave systems (Phase shifter, transmission/reception module, antennas array, reconfigurable systems, impedance matching…). To develop the next generation of RADAR systems, Thales takes special interest in the integration of capacitive RF MEMS devices to demonstrate reconfigurable functions with power handling capabilities up to 30 W. The work reported in this thesis did focus on the study of dielectric materials and deposition techniques to identify, integrate and demonstrate the advantages of promising dielectrics for capacitive RF MEMS power handling. The components fabrication aspects have also been studied, especially the impact of surface state quality on performances, power handling and devices failure mechanisms. Furthermore, this work did point out that with the integration of ALD material, power handling of capacitive RF MEMS is no longer limited by the capacitance dielectric. Furthermore, with the integration of ALD material the components design become the limiting factor for power handling, particularly the membrane thickness and the switch configuration. To open new prospects, several designs have been developed and studied to address these power handling limitations.

This thesis investigates the ability of transparent surface coatings to reduce xenon diffusion into plastic scintillators. The motivation for the work is improved radioxenon monitoring equipment, used with in the framework of the verification regime of the Comprehensive Nuclear-Test-Ban Treaty. A large part of the equipment used in this context incorporates plastic scintillators which are in direct contact with the radioactive gas to be detected. One problem with such setup is that radioxenon diffuses into the plastic scintillator material during the measurement, resulting in an unwanted memory effect consisting of residual activity left in the detector. In this work coatings of Al2O3 and SiO2, with thicknesses between 20 and 400 nm have been deposited onto flat plastic scintillator samples, and tested with respect to their Xe diffusion barrier capabilities. All tested coatings were found to reduce the memory effect, and 425 nm of Al2O3 showed the most promise. This coating was deposited onto a complete detector. Compared to uncoated detectors, the coated one presented a memory effect reduction of a factor of 1000. Simulations and measurements of the expected light collection efficiency of a coated detector were also performed, since it is important that this property is not degraded by the coating. It was shown that a smooth coating, with a similar refractive index as the one of the plastic, should not significantly affect the light collection and resolution. The resolution of the complete coated detector was also measured, showing a resolution comparable to uncoated detectors. The work conducted in this thesis proved that this coating approach is a viable solution to the memory effect problem, given that the results are reproducible, and that the quality of the coating is maintained over time.

Thin films of cobalt oxide, iron oxide and niobium oxide, and nanostructured thin films of iron oxide, titanium oxide and multilayered iron oxide/titanium oxide have been deposited by Atomic Layer Deposition (ALD). The metal oxides were grown using the precursor combinations CoI2/O2, Fe(Cp)2/O2, NbI5/O2 and TiI4/H2O. The samples were analysed primarily with respect to phase content, morphology and growth characteristics. Thin films deposited on Si (100) were found to be amorphous or polycrystalline, depending on deposition temperature and the oxide deposited; cobalt oxide was also deposited on MgO (100), where it was found to grow epitaxially with orientation (001)[100]Co3O4||(001)[100]MgO. As expected, the polycrystalline films were rougher than the amorphous or the epitaxial films. The deposition processes showed properties characteristic of self-limiting ALD growth; all processes were found to have a deposition temperature independent growth region. The deposited films contained zero or only small amounts of precursor residues. The nanostructured films were grown using anodic aluminium oxide (AAO) or carbon nanosheets as templates. Nanotubes could be manufactured by depositing a thin film which covers the pore walls of the AAO template uniformly; free-standing nanotubes retaining the structure of the template could be fabricated by removing the template. Multilayered nanotubes could be obtained by depositing multiple layers of titanium dioxide and iron oxide in the pores of the AAO template. Carbon nanosheets were used to make titanium dioxide nanosheets with a conducting graphite backbone. The nucleation of the deposited titanium dioxide could be controlled by acid treatment of the carbon nanosheets.

In the industry of metal cutting tools the conditions are extreme; the temperature can vary thousand degrees rapidly and the pressure can be tremendously high. To survive this kind of stress the cutting tool must be both hard and tough. In order to obtain these properties different coatings are used on a base of cemented carbide, WC-Co. Common coatings are hard ceramics like titanium nitride and titanium carbon-nitride with an outer layer of aluminium oxide. In this thesis the possibility of using titanium dioxide as an interlayer between titanium carbon-nitride and aluminium oxide to control the morphology and phase of aluminium oxide is investigated. Of the different aluminium oxide phases only the alpha-Al2O3 is stable. The titanium carbon-nitride coatings are made by CVD (chemical vapour deposition); also the alumina is deposited by CVD. The titanium dioxide was deposited by atomic layer deposition (ALD) which is a sequential CVD technique that allows a lower deposition temperature and better control of the film growth than CVD. The obtained thin films were analyzed using XRD, Raman spectroscopy, ESCA and SEM. To test the adhesion of the coatings the samples were sand blasted. A thin interlayer of titanium dioxide causes the aluminium oxide to grow as alpha-Al2O3, thinner TiO2 gave better adhesion.