Silicon photonics is emerging as a potential field to achieve optical interconnects towards the realization of ultra high bandwidth. The indirect band-gap property of silicon still remains as a big challenge to incorporate silicon photonic active device, for example, silicon-based laser. In the Laboratory of Semiconductor Materials at KTH, a monolithic integration platform based on nano-epitaxial lateral overgrowth (nano-ELOG) technique has been proposed to integrate III-V semiconductor materials with silicon for light source application. The integration process involves uneven surface morphology at different stages. The surfaces of the indium phosphide seed layer on silicon used for ELOG, the mask deposited on it (the silicon/silicon dioxide waveguide) and the ELOG indium phosphide layer grown on it prior to laser growth are often rough. In this thesis work, we have optimized chemical mechanical polishing (CMP) technique in order to achieve an even surface. The same procedure is also necessary to reach the optimal thickness of different layers to enable effective coupling of light from the laser source into the waveguide. CMP of indium phosphide to obtain an average surface roughness of < 1 nm has been optimized by a two-step polishing using different slurries; it results in a step height of ca 3 nm. Similarly the surface of silicon/silicon dioxide “waveguide” has also been optimized with the roughness of ~ 0.5 nm. In the latter case, a step height of 40 nm is retained and this increase with respect to InP is identified to be mainly due to limitations of the polishing machine which is different from that used for indium phosphide. The reduction in step heights with polishing time is analyzed and compared with an existing theoretical model. Our results are in good qualitative agreement with the model. The optimized surface morphology obtained in this work was tested for its suitability for integration. For this evaluation, InP was grown by ELOG in a hydride vapour phase epitaxy reactor with and without CMP of the involved surfaces. The surface after CMP yields layers of better surface morphology with fewer defects as revealed by atomic force microscopy, surface profilometer and cathodoluminescence analysis. The results indicate that the CMP process is useful for monolithic integration for silicon photonics.

Le diamant est le semi-conducteur par excellence pour les composants électroniques de puissance. Par conséquent, la technique de croissance du diamant et les dispositifs à base de diamant ont été largement étudiés dans le monde entier au cours des deux dernières décennies. A ce jour, les diodes Schottky à base de diamant sont les composants les plus avancés et les plus prometteurs. Cependant, pour remplacer Si et SiC dans cette filière technologique, des progrès importants dans la technologie du diamant sont nécessaires.L'amélioration de l’électronique de puissance dépend non seulement des caractéristiques intrinsèques du semi-conducteur mais aussi de l'architecture du dispositif. Les dispositifs verticaux et pseudo-verticaux offrent nombreux avantages comme la faible résistance spécifique, la haute tension de rupture et une taille réduite. De plus, la conception tridimensionnelle (3D) permet de réduire le champ électrique à l'intérieur du matériau et de tirer parti des qualités exceptionnelles du diamant si le mécanisme de croissance de ce matériau sur un substrat structuré était plus étudié et mieux compris.Ainsi, le but de ce manuscrit de thèse est de comprendre les mécanismes qui régissent la reprise de croissance en homoépitaxie sur un substrat structuré 3D en diamant. Pour atteindre cet objectif, les principaux mécanismes de la croissance CVD sur substrat gravé sont étudiés expérimentalement par microscopie électronique en transmission grâce à une approche stratigraphique originale qui permet de suivre la direction de croissance et la génération de défauts aux différentes étapes de la croissance. L’observation de divers secteurs de croissance, d’une tendance générale à la planarisation, et le rôle important joué par la concentration de méthane qui gouverne la vitesse de croissance, sont les principaux résultats de cette étude. Divers modèles d’interprétation allant de l’échelle atomistique à l’échelle macroscopique sont discutés. L'origine des dislocations, et des autres défauts étendus et superficiels a été déterminée. Les basses concentrations de méthane ont permis une croissance 3D de haute qualité. La reprise de croissance sur des micro-terrasses est présentée comme une alternative pour l’obtention de grandes surfaces sans défauts. Une nouvelle méthode basée sur le renforcement par solution solide est introduite comme alternative à la cathodoluminescence pour évaluer le dopage dans les régions riches en dislocations, où les défauts empêchent l’observation d’excitons. Tous les résultats obtenus ont été pris en compte pour réduire le nombre d’étapes technologiques nécessaires pour fabriquer des composants à architecture 3D en diamant (diode Schottky ou même MOSFET). Dans le cas des diodes Schottky, le procédé et l’architecture proposés ont les avantages suivants :- « Filtrer » les dislocations.- Permettre d’étendre la région de champ arbitrairement.- Eviter certaines étapes de photolithographie.- Améliorer le contact ohmique, par croissance d’une couche p+ sur une facette {111}. / Diamond is the ultimate semiconductor for power electronic devices. Consequently, diamond growth techniques and diamond-base devices have been intensively investigated over the last two decades all over the world. Among these power devices, diamond based Schottky diodes are the most advanced. However, for diamond to substitute present Si and SiC for power electronics, a substantial technological progress is still needed.The improvement of power devices depends not only on the intrinsic characteristic of the semiconductor, but also on the device architecture. Vertical and pseudo-vertical devices offer many advantages such as low-specific on-resistance, high breakdown voltage and a smaller size. Moreover, three-dimensional (3D) design allows to reduce the electric field inside the material and would make the most out of the superb material qualities of diamond, if only the diamond growth mechanism over patterned diamond substrates would be better studied and understood.Thus, the aim of this thesis manuscript is to understand the mechanisms governing the three-dimensional (3-D) shaped diamond substrates homoepitaxial overgrowth, in order to implement them in the design and fabrication of a Schottky device.To reach this goal, the main mechanisms of CVD growth over patterned substrates were experimentally investigated by transmission electronic microscopy using a novel stratigraphic approach that allowed to follow the growth direction and generation of defects at various stages. Evidence was provided for many sectors of growth, and a tendency to planarization, while the methane concentration and resulting growth rates were shown to be key parameters. Various interpretation models, from the atomistic to the macroscopic scale, were discussed. The origin of dislocations, and of other extended and superficial defects was determined. Low methane concentrations led to high quality 3-D overgrowth. The overgrowth of micro-terraces is proposed as a method to achieve large areas free of defects. A novel method based in solid solution strengthening was introduced as an alternative to cathodoluminescence to evaluate boron doping in dislocation-rich regions where extended defects usually hinder this approach. All the results obtained above have been taken into account to reduce the number of technological steps leading to a diamond based 3D device (Schottky diode or even MOSFET). In the case of Schottky diodes, a 3D design was proposed with the following advantages:- To “filter” dislocations.- To allow an arbitrary large field region.- To spare photolithography steps.- To improve ohmic contacts, as the p+ layer is grown on a {111} facet.

The purpose of this thesis was to study interactions between systemic fungal diseases and perennial plants. Using the systemic rust Puccinia minussensis on the host plant Lactuca sibirica, and the rust Puccinia pulsatillae on the host plant Pulsatilla pratensis, this thesis focused on: (i) the effects of systemic diseases on their hosts (ii) host and pathogen responses to abiotic factors, (iii) the importance of life history strategies for understanding host-pathogen interactions, and (iv) the evolutionary consequences of living in close associations. Results of greenhouse experiments showed that Lactuca sibirica had a high plasticity in growth, since it produced significantly more shoots in favourable than in unfavourable growth conditions. Both the disease levels and the number of healthy shoots (i.e. escape) were significantly higher under favourable conditions. Disease spread within the rhizome was found to be incomplete, and the risk of aecidial- infection decreased with distance from the parent. Furthermore, one isolate of the fungus had highest success and reduced the host plant biomass and shoot production more on the clone it was collected on compared to four other clones . In the field, disease levels were found to fluctuate more at localities subjected to disturbance, the host and pathogen abundances were found to be in phase and the pathogen showed no delayed response to increasing host densities. The rust Puccinia pulsatillae on Pulsatilla pratensis showed no fluctuations between years, low infection rates, and disease levels were higher in ungrazed compared to grazed sites. There was no escape from the disease in this system. A comparison of characteristics of different systemic fungi and hosts with different growth patterns indicated that the life history strategies of both host plants and pathogens need to be studied if the long-term consequences of host-pathogen interactions are to be predicted. / <p>Diss. (sammanfattning) Umeå : Umeå universitet, 1993, härtill 5 uppsatser.</p> / digitalisering@umu