Through a direct application of two cognitive science theories, conceptual metaphor (Lakoff & Johnson, 1980, 1999) and structure mapping (Gentner, 1983, 1989; Gentner & Markman, 1995), this project defined an instructional design model for the design, development, and assessment of metaphor-enhanced, computer-mediated learning environments. It used the model to produce an instructional product with a metaphor-based interface. The project also built a parallel learning environment that employed a concept map interface. To test the metaphor-based product's effectiveness at enabling learners to build rich mental models of a complex, abstract concept, the project ran fifty-seven preservice teachers (55 female, 2 male; mean age of 21) through the instruction, randomly assigning half to the concept map interface environment and half to the metaphor-based interface environment. Participants completed four essay-type assessment questions. Trained raters, blind to participant assignment, isolated any of the 13 targeted concepts present within participants' protocols and, through consensus, constructed a concept map for each participant, representing that participant's mental model of the targeted domain. Map attributes were translated into four weighted subscores (nodes, branches, levels, and cross-links) and summed. Comparison across the two groups indicated no significant difference for richness of mental model, t(55)=-.72, p > .05, although the discussion suggests methods for increasing the power in subsequent experimental sessions. A significant interaction between Subscore and Achievement, F(3,51)=33.42, p < .01, suggests that concept map cross-links are much more sensitive to differences in domain integration and the general richness of a participant's mental model than the level and branch subscores. This result has implications for classroom application. Concept maps have taken a place as a learner's, a teacher's, and a researcher's tool. With cross-domain validation and domain-specific extensions, specification of the relative sensitivity of various subscales, that is, the structure of the concept map, will enable educators to justify weighting scales and identify learner achievement. Credible concept map weighting scales also enhance learners' self-reliant and impartial assessment of personal growth in domain-specific knowledge. Results suggest that learners who have difficulty integrating domain concepts require direct, explicit instruction to help them to make connections between disparate conceptual strands. / Ph. D.

Design is hard and needs to be supported by software. One of the ways software can support designers is by providing analogical reasoning. To make analogical reasoning work well, the software makers need to know how to create a knowledge representation that will facilitate the kind of analogies that the designers want. This thesis will inform software makers by experimenting with two kinds of knowledge representations, called device-centric (DC) and environment-centric (EC), and to try to determine the relative benefits of using either one of them for analogical matching. We performed computational experiments, using Structure Mapping Engine for matching, to determine the quantity and quality of analogical matches that are produced when the representation is varied. We conducted a limited human experiment, using questionnaires and repertory grids, to determine if any of the computational results were novel, and to determine if the human similarity ratings between devices correlated with the computer results. We show that design software should use DC representations to produce a few focused matches which have high average weight. It should use EC representations to produce many matches some of high weight and some of low weight. Based on our human experiment, design software can use either DC or EC representations to produce novel matches. Our experiments also show that human matches correlate most strongly with a combined DC and EC representation and that their similarity reasons are more EC than DC. This suggests that designers tend to think more in EC terms than in DC terms.

Cette thèse, effectuée en collaboration entre le CEMES-CNRS, le laboratoire Satie (ENS Cachan) et NXP Semiconductors est motivée par la compréhension des mécanismes de défaillance des dispositifs MOSFET pour les applications dans l'industrie automobile. Un facteur limitant de la fiabilité à long terme des modules de puissance basse tension est le vieillissement électrothermique et/ou thermo-mécanique des parties métalliques de la source: métallisation en aluminium (ou alliage) et fils de connexion. A cause de la différence de coefficient de dilatation thermique entre la métallisation les oxydes et le substrat semi-conducteur, la température atteinte pendant les cycles de fonctionnement (quelques centaines de degrés), induit une déformation plastique inévitable dans le métal, qui est le matériau le plus mou dans l'architecture complexe du MOSFET. Nous avons caractérisé la microstructure métallique avant et après les tests de vieillissement électrothermique accélérés, en utilisant des techniques spécifiques du domaine de la métallurgie physique: microscopie électronique et ionique, cartographie d'orientation de grains et de la composition chimique. Pour la première fois, la métallisation de la source a été caractérisée sous les fils de connexion, qui sont cent fois plus épais que la métallisation. Cet emplacement est critique pour la fiabilité du composant, car le processus de soudure par ultrasons induit une déformation plastique importante qui peut affaiblir la métallisation initiale avant le vieillissement. Ceci est peu étudié dans la littérature en raison de la difficulté à accéder à la métallisation sous les fils sans altérer leur interface, souvent endommagée et fragilisée dans les modules vieillis. Nous avons mis en place des méthodes de préparation d'échantillon, basées sur le polissage ionique, pour étudier cette interface, sans introduire d'artefacts de préparation. Le processus de soudure à froid induit une déformation plastique sévère et non uniforme dans la métallisation sous les fils sans parvenir à recréer un bon contact électrique: de petites cavités et des résidus d'oxyde natif, ont été systématiquement observés à l'interface Al / Al, dans tous les modules analysés, avant et après vieillissement. Le mécanisme principal de défaillance des modules est la génération et la propagation de fissures de fatigue dans l'aluminium, associée à une oxydation locale qui empêche la fermeture de ces fissures. Sous et en dehors des fils de connexion, ces fissures traversent la métallisation perpendiculairement à la surface jusqu'au substrat en silicium en suivant les joints de grains. Cette fissuration est due à la diffusion intergranulaire accélérée des atomes d'aluminium. Dans la zone de soudure, le phénomène de fissuration parallèle à l'interface est favorisé par les imperfections initiales (cavités, oxyde). Les expériences de tomographie ionique ont montré que ces fissures sont confinées à l'interface fil-métal et ne se propagent pas dans le fil malgré sa plus faible résistance mécanique (Al pur, structure à grains plus grands). La propagation de la fissure le long de l'interface Al/Al peut provoquer une diminution du contact entre le fil et la métallisation de la source et éventuellement son décollement. Les fissures dans le métal source peuvent expliquer l'augmentation locale de la résistance et de la température du module qui accélère le processus de vieillissement jusqu'à l'échec. Cette étude a établi de nouvelles techniques dédiées et des méthodes de quantification pour évaluer le vieillissement des parties métalliques des modules MOSFET. La caractérisation complète de l'interface soudée, intrinsèquement défectueuse et la dégradation de la métallisation pendant le vieillissement électrothermique ouvrent la voie à l'amélioration possible les technologies actuelles et au développement potentiel de nouveaux procédés. / This thesis, a collaboration between CEMES-CNRS, Satie laboratory (ENS Cachan) and NXP Semiconductors is motivated by the comprehension of the failure mechanisms of low voltage power MOSFET devices produced for ap- plications in the automotive industry. A limiting factor for the long-term reliability of power modules is the electro- thermal and/or thermo-mechanical aging of the metallic parts of the source: Al metallization and bonding wires. At the temperature reached during the on-off operating cycles (few hundred degrees), the difference in the coefficient of thermal expansion between the metallization and the oxide and semicon- ductor parts induces an inevitable plastic deformation in the metal, which is the softest material in the complex MOSFET architecture. We have characterized the metal microstructure before and after accelerated electro-thermal aging tests, by using specific techniques from the field of the physical metallurgy: electron and ion microscopy, grain orientation and chem- ical composition mapping. For the first time the source metallization has been characterized both away and under the bonding connections, which are one hundred times thicker than the metallization layer. The latter is a critical loca- tion for the reliability assessment because the ultrasonic bonding process may weaken the initial metallization microstructure by adding an important plas- tic deformation prior to aging. This is, however, poorly stated in the literature because of the difficulty to access the metallization under the wires without damaging their bonding, which is known to be particularly weak in case of aged modules. In order to investigate the wire-metallization interface, we have set up origi- nal sample preparations, based on ion polishing, that allowed us to disclose the metallization under the bonding wires without introducing preparation artifacts in the microstructure. The bonding process induces a severe and non- uniform plastic deformation in the metallization under the wires without re- creating a good electrical contact: small cavities and native oxide residues, have been systematically observed at the Al/Al interface, in all the analyzed mod- ules, before and after aging. The main mechanism behind the device failure is the generation and propa- gation of fatigue cracks in the aluminum metallization, associated to a local Al oxidation that prevents these crack from closing. Away and under the wire bonds, they run perpendicularly from the surface down to the silicon substrate following the grain boundaries, due to an enhanced intergranular diffusion of aluminum atoms. In the bonding area, the phenomenon of parallel cracking is favored by the initial imperfections in the wire-metallization bonding. Ion to- mography experiments have shown that these cracks are confined to the wire- metal interface and do not propagate in the wire despite its lower strength (pure Al, larger grain structure). Crack propagation along the Al/Al interface can cause a contact reduction between the wire and the source metallization and eventually its failure. Such discontinuities in the metal can explain the lo- cal increase in the device resistance and temperature that accelerates the aging process until failure. This study settled new, dedicated techniques and quantification methods to as- sess the aging of the metal parts of MOSFET devices. The full characterization of the intrinsically defective interface generated by the bonding process and the metallization degradation during electro-thermal aging indicated paths to possible improvements of current technologies and potential developments of new processes.

Les villes sont confrontées à de nombreuses problématiques environnementales. Leurs gestionnaires ont besoin d'outils et d'une bonne connaissance de leur territoire. Un objectif est de mieux comprendre comment s'articulent les trames grise et verte pour les analyser et les représenter. Il s'agit aussi de proposer une méthodologie pour cartographier la structure urbaine à l'échelle des tissus en tenant compte de ces trames. Les bases de données existantes ne cartographient pas la végétation de manière exhaustive. Ainsi la première étape est d'extraire la végétation arborée et herbacée à partir d'images satellites Pléiades par une analyse orientée-objet et une classification par apprentissage actif. Sur la base de ces classifications et de données multi-sources, la cartographie des tissus se base sur une démarche d'extraction de connaissances à partir d'indicateurs issus de l'urbanisme et de l'écologie du paysage. Cette méthodologie est construite sur Strasbourg puis appliquée à Rennes. / Climate change presents cities with significant environmental challenges. Urban planners need decision-making tools and a better knowledge of their territory. One objective is to better understand the link between the grey and the green infrastructures in order to analyse and represent them. The second objective is to propose a methodology to map the urban structure at urban fabric scale taking into account the grey and green infrastructures. In current databases, vegetation is not mapped in an exhaustive way. Therefore the first step is to extract tree and grass vegetation using Pléiades satellite images using an object-based image analysis and an active learning classification. Based on those classifications and multi-sources data, an approach based on knowledge discovery in databases is proposed. It is focused on set of indicators mostly coming from urbanism and landscape ecology. The methodology is built on Strasbourg and applied on Rennes to validate and check its reproducibility.