Teoretická část této diplomové práce popisuje vlastnosti a možnosti modifikace poly(3-hydroxybutyrátu) (PHB) a amorfní poly(mléčné kyseliny) (PLA) a jejich směsí. V experimentální části je studována reaktivita Joncrylu, Raschigu a fosfitových činidel trifenylfosfitu, tris(nonylfenyl) fosfitu a difenylisodecylfosfitu s čistým PLA a PHB. Raschig, oligomerní aditivum na bázi polykarbodiimidu, prokázal v množství 2hm. % zvýšení viskozity taveniny obou polymerů, a proto byl použit k přípravě směsí o pěti hmotnostních poměrech PHB/(PHB+PLA). Vzorky s Raschigem a odpovídajícími nereaktivními vzorky byly studovány pomocí reologie, gelové permeační chromatografie a modulované diferenční kompenzační kalorimetrie. Výsledky naznačily reakce Raschigu v PHB/PLA směsích vedoucí k rozvětveným strukturám. Rychlost reakce však není nedostatečná ke kompenzaci poklesu viskozity v důsledku degradace při zpracování. Následně zůstává nezreagované množství Raschigu v matrici. Na základě těchto zjištění se dospělo k závěru, že Raschig se chová spíše jako relativně účinný stabilizátor reologických vlastností, než jako činidlo pro záměrnou modifikaci struktury směsí PHB/PLA. Za účelem studia mechanických vlastností těchto směsí byly ve dvoušnekovém extrudéru připraveny vzorky plastifikovány acetyltributylcitrátem.

La théorie classique de la valorisation des produits dérivés se repose sur l'absence de coûts de transaction et une liquidité infinie. Ces hypothèses sont toutefois ne plus véridiques dans le marché réel, en particulier quand la transaction est grande et les actifs non-liquides. Dans ce marché imparfait, on parle du prix de sur-réplication puisque la couverture parfaite est devenue parfois infaisable.La première partie de cette thèse se concentre sur la proposition d’un modèle qui intègre à la fois le coût de transaction et l’impact sur le prix du sous-jacent. Nous commençons par déduire la dynamique de l’actif en temps continu en tant que la limite de la dynamique en temps discret. Sous la contrainte d’une position nulle sur l’actif au début et à la maturité, nous obtenons une équation quasi-linéaire pour le prix du dérivé, au sens de viscosité. Nous offrons la stratégie de couverture parfaite lorsque l’équation admet une solution régulière. Quant à la couverture d’une option européenne “covered” sous la contrainte gamma, le principe de programme dynamique utilisé précédemment n'est plus valide. En suivant les techniques du cible stochastique et de l’équation différentielle partielle, nous démontrons que le prix de la sur-réplication est devenue une solution de viscosité d’une équation non linéaire de type parabolique. Nous construisons également la stratégie ε-optimale, et proposons un schéma numérique.La deuxième partie de cette thèse est consacrée aux études sur un nouveau schéma numérique d'EDSR, basé sur le processus de branchement. Nous rapprochons tout d’abord le générateur Lipschitzien par une suite de polynômes locaux, puis appliquons l’itération de Picard. Chaque itération de Picard peut être représentée en termes de processus de branchement. Nous démontrons la convergence de notre schéma sur l’horizon temporel infini. Un exemple concret est discuté à la fin dans l’objectif d’illustrer la performance de notre algorithme. / Classical derivatives pricing theory assumes frictionless market and infinite liquidity. These assumptions are however easily violated in real market, especially for large trades and illiquid assets. In this imperfect market, one has to consider the super-replication price as perfect hedging becomes infeasible sometimes.The first part of this dissertation focuses on proposing a model incorporating both liquidity cost and price impact. We start by deriving continuous time trading dynamics as the limit of discrete rebalancing policies. Under the constraint of holding zero underlying stock at the inception and the maturity, we obtain a quasi-linear pricing equation in the viscosity sense. A perfect hedging strategy is provided as soons as the equation admits a smooth solution. When it comes to hedging a covered European option under gamma constraint, the dynamic programming principle employed previously is no longer valid. Using stochastic target and partial differential equation smoothing techniques, we prove the super-replication price now becomes the viscosity solution of a fully non-linear parabolic equation. We also show how ε-optimal strategies can be constructed, and propose a numerical resolution scheme.The second part is dedicated to the numerical resolution of the Backward Stochastic Differential Equation (BSDE). We propose a purely forward numerical scheme, which first approximates an arbitrary Lipschitz driver by local polynomials and then applies the Picard iteration to converge to the original solution. Each Picard iteration can be represented in terms of branching diffusion systems, thus avoiding the usual estimation of conditional expectation. We also prove the convergence on an unlimited time horizon. Numerical simulation is also provided to illustrate the performance of the algorithm.

La construction du cerveau et de ses connexions pendant le développement reste une question ouverte dans la communauté scientifique. Des efforts fructueux ont été faits pour élucider les mécanismes de la croissance axonale, tels que la guidance axonale et les molécules de guidage. Cependant, des preuves récentes suggèrent que d'autres acteurs seraient impliqués dans la croissance des neurones in vivo. Notamment, les axones se développent dans des environnements mécaniquement contraints. Ainsi, pour bien comprendre ce processus dynamique, il faut prendre en compte les mécanismes collectifs et les interactions mécaniques au sein des populations axonales. Néanmoins, les techniques pour mesurer directement cela à partir de cerveaux vivants sont aujourd'hui insuffisantes ou lourdes à mettre en œuvre. Cette thèse résulte d'une collaboration multidisciplinaire, pour faire la lumière sur le développement axonal in vivo et les morphologies complexes des axones adultes. Notre travail a été inspiré et validé à partir d'images d'axones y individuels chez la drosophile, de type sauvage et modifiés génétiquement, que nous avons segmentés et normalisés. Nous avons d'abord proposé un cadre mathématique pour l'étude morphologique et la classification des groupes axonaux. A partir de cette analyse, nous avons émis l'hypothèse que la croissance axonale dérive d'un processus stochastique et que la variabilité et la complexité des arbres axonaux résultent de sa nature intrinsèque, ainsi que des stratégies d'élongation développées pour surmonter les contraintes mécaniques du cerveau en développement. Nous avons conçu un modèle mathématique de la croissance d'un axone isolé fondé sur des chaînes de Markov gaussiennes avec deux paramètres, représentant la rigidité axonale et l'attraction du champ cible. Nous avons estimé les paramètres de ce modèle à partir de données réelles et simulé la croissance des axones à l'échelle de populations et avec des contraintes spatiales pour tester notre hypothèse. Nous avons abordé des thèmes de mathématiques appliquées ainsi que de la biologie, et dévoilé des effets inexplorés de la croissance collective sur le développement axonal in vivo. / How the brain wires up during development remains an open question in the scientific community across disciplines. Fruitful efforts have been made to elucidate the mechanisms of axonal growth, such as pathfinding and guiding molecules. However, recent evidence suggests other actors to be involved in neuron growth in vivo. Notably, axons develop in populations and embedded in mechanically constrained environments. Thus, to fully understand this dynamic process, one must take into account collective mechanisms and mechanical interactions within the axonal populations. However, techniques to directly measure this from living brains are today lacking or heavy to implement. This thesis emerges from a multidisciplinary collaboration, to shed light on axonal development in vivo and how adult complex axonal morphologies are attained. Our work is inspired and validated from images of single wild type and mutated Drosophila y axons, which we have segmented and normalized. We first proposed a mathematical framework for the morphological study and classification of axonal groups. From this analysis we hypothesized that axon growth derives from a stochastic process, and that the variability and complexity of axonal trees result from its intrinsic nature, as well as from elongation strategies developed to overcome the mechanical constraints of the developing brain. We designed a mathematical model of single axon growth based on Gaussian Markov Chains with two parameters, accounting for axon rigidity and attraction to the target field. We estimated the model parameters from data, and simulated the growing axons embedded in spatially constraint populations to test our hypothesis. We dealt with themes from applied mathematics as well as from biology, and unveiled unexplored effects of collective growth on axonal development in vivo.

Major progress has been made during the last twenty years towards a better understanding of the evolution of angiosperms. Early molecular-phylogenetic analyses revealed three major groups, with eudicots as well as monocots being monophyletic, arisen from a paraphyletic group of dicotyledonous angiosperms (= basal angiosperms). Consistently, numerous phylogenetic studies based on sequence data have recovered the eudicot-clade and increased confidence in its existence. Furthermore this clade, which contains about 75% of angiosperm species diversity, is characterized by the possession of tricolpate and tricolpate-derived pollen and has thus also been called the tricolpate clade. Based on molecular-phylogenetic investigations several lineages, such as Ranunculales, Proteales (= Proteaceae, Nelumbonaceae, Platanaceae), Sabiaceae, Buxaceae plus Didymelaceae, and Trochodendraceae plus Tetracentraceae were shown as belonging to a early-diverging grade (early-diverging or “basal” eudicots), while larger groups like asterids, Caryophyllales, rosids, Santalales, and Saxifragales were identified as being members of a highly supported core clade, the so called “core eudicots”. Nevertheless, phylogenetic relationships among several lineages of the eudicots remained difficult to resolve. This thesis is mainly concentrated on fully resolving the branching order among the different clades of the early-diverging eudicots as well as on clarifying phylogenetic and systematic conditions within several lineages, based on phylogenetic reconstructions using sequence data of rapidly-evolving and non-coding molecular regions, such as spacers and introns. Commonly, fast-evolving and non-coding DNA was used to infer relationships among species and genera, as practised in chapter 3, due to the assumption of being inapplicable caused by putative high levels of homoplasy through multiple substitutions and frequent microstructural changes resulting in non-alignability. However, during the last few years numerous molecular-phylogenetic studies were able to present well resolved angiosperm trees on the basis of rapidly-evolving and non-coding regions from the large single copy region of the chloroplast genome comparable to multi-gene analyses concerning topology and statistical support. Mutational dynamics in spacers and introns was revealed to follow complex patterns related to structural constraints like the introns secondary structure. Therefore extreme sequence variability was always confirmed to mutational hotspots that could be excluded from calculations. Moreover it became clear that combining these non-coding regions with the fast-evolving matK gene can lead to further resolved and statistical supported trees. Chapter 1 deals with the placement of Sabiales inside the early-diverging eudicot grade, while investigating mutational dynamics as well as the utility of different kinds of non-coding and rapidly-evolving DNA within deep-level phylogenetics. It was done by analyzing a combination of nine regions from the large single copy region of the chloroplast genome, including spacers, the sole group I intron, three group II introns and the coding matK for a sampling of 56 taxa. The presented topology is in mainly congruence with the hypothesis on phylogenetic relationships among early-branching eudicots that was gained through the application of a reduced set of five non-coding and fast-evolving molecular markers, including the plastid petD (petB-petD spacer, petD group II intron) plus the trnL-F (trnL group I intron, trnL-F spacer) region and the matK gene. It showed a grade of Ranunculales, Sabiales, Proteales, Trochodendrales and Buxales. The current study differs in showing Sabiales as sister to Proteales in all phylogenetic analyses, in contrast to a second-branching inside early-diverging eudicots and a Bayesian tree displaying Sabiales branching after Proteales. All three hypotheses were tested concerning their likelihood. None of them was shown as being significantly declinable. Thus, albeit the number of characters and informative sites was doubled in comparision to the five-region investigation, the exact position of the Sabiales remained to be resolved with confidence. However, the advanced analyses of the phylogenetic structure of the three different non-coding partitions in comparison to coding genes resulted in the recognition of a significantly higher mean phylogenetic signal per informative character within spacers and introns than in the frequently applied slowly-evolving rbcL gene. The fast-evolving and well performing matK gene is shown to be nested within the non-coding partitions in this respect. Interestingly, the least constrained spacers displayed considerably less phylogenetic structure than both, the group I intron and the group II introns. Molecular evolution is again shown to follow certain patterns in angiosperms, as indicated by the occurrence of mutational hotspots and their connection to structural and functional constraints. This is especially shown for the group II introns studied where highly dynamic sequence parts were rather found in loops than stems. The aim of chapter 2 was to present a comprehensive reconstruction of the phylogenetic relationships inside the order of Ranunculales, the first-branching clade of the early-diverging eudicots, with an emphasis on the evolution of growth forms within the group. Currently, the order comprises seven families (Ranunculaceae, Berberidaceae, Menispermaceae, Lardizabalaceae, Circaeasteraceae – not included due to lacking plant material, Eupteleaceae, Papaveraceae) containing predominantly herbaceous groups as well as trees and lianescent/shrubby forms. A surprising result that emerged due to the increased use of molecular data within systematics during the last twenty years is the inclusion of the woody Eupteleaceae into Ranunculales. Because of its adaptation to wind pollination it was previously placed next to Hamamelididea. Although phylogenetic hypotheses agreed in the exclusion of Eupteleaceae and the predominantly herbaceous Papaveraceae from a core clade the branching order within early-diverging Ranunculales remained a question to be answered. Thus phylogenetic reconstructions based on molecular data of 50 taxa (including outgroup), applying the well-performing non-coding petD and trnL-F as well as the trnK/matK-psbA region including the coding matK, were carried out. The comprehensive sampling resulted in fully resolved and highly supported phylogenies in both, maximum parsimony and model based approaches, with family relations within the core clade being identical and Euptelea appearing as first branching lineage. However, the relationships among the early-diverging Ranunculales could not be resolved with confidence, a result in line with the finding made in chapter 1. The topology was further resolved as Lardizabalaceae being sister to the remaining members of the order, followed by Menispermaceae, Berberidaceae and Ranunculaceae, the latter sharing a sistergroup relationship. Inside the mainly lianescent Lardizabalaceae the shrubby Decaisnea was clearly depicted as first-branching. The systematic controversial Glaucidium and Hydrastis are shown to be early-diverging members of the Ranunculaceae. A central goal of chapter 3 was to test phylogenetic relationships among the members of the ranunculaceous tribe Anemoneae. Currently it consists of the subtribes Anemoninae including Anemone, Hepatica, Pulsatilla and Knowltonia, and Clematidinae, consisting of Archiclematis, Clematis and Naravelia. Furthermore the position and taxonomic rank of several lineages inside the subtribe Anemoninae were examined. Since recent comprehensive molecular-phylogenetic investigations have been carried out for the members of Clematidinae or Anemoninae, 63 species representing all major lineages of the two subtribes were included into analyses. Calculations were carried out on the basis of molecular data of the nuclear ribosomal ITS1&2 and the plastid atpB-rbcL intergenic spacer region. Phylogenetic reconstructions resulted in the recognition of two distinct clades within the tribe, thus corroborating the formation of the two subtribes. Within the subtribe Anemoninae the traditional genera Knowltonia, Pulsatilla and Hepatica are confidently shown to be nested within the genus Anemone. The preliminary classification of the genus, currently consisting of the two subgenera Anemone and Anemonidium, is complemented by the subgenus Hepatica.

Spiking neural networks (SNN) are dynamic models of biological neurons, that communicates with event-based signals called spikes. SNN that reproduce observed properties of biological senses like vision are developed to better understand how such systems function, and to learn how more efficient sensor systems can be engineered. A branching parameter describes the average probability for spikes to propagate between two different neuron populations. The adaptation of branching parameters towards critical values is known to be important for maximizing the sensitivity and dynamic range of SNN. In this thesis, a recently proposed SNN model for visual feature learning and pattern recognition known as the E-I Net model is studied and extended with a critical branching mechanism. The resulting modified E-I Net model is studied with numerical experiments and two different types of sensory queues. The experiments show that the modified E-I Net model demonstrates critical branching and power-law scaling behavior, as expected from SNN near criticality, but the power-laws are broken and the stimuli reconstruction error is higher compared to the error of the original E-I Net model. Thus, on the basis of these experiments, it is not clear how to properly extend the E-I Net model properly with a critical branching mechanism. The E-I Net model has a particular structure where the inhibitory neurons (I) are tuned to decorrelate the excitatory neurons (E) so that the visual features learned matches the angular and frequency distributions of feature detectors in visual cortex V1 and different stimuli are represented by sparse subsets of the neurons. The broken power-laws correspond to different scaling behavior at low and high spike rates, which may be related to the efficacy of inhibition in the model.

Dans cette thèse nous avons étudié des collisions entre des projectiles CnN⁺ (n=0,1,2,3) et des atomes d’Hélium à vitesse intermédiaire (2.25 u.a). A cette vitesse, proche de la vitesse des électrons sur les couches de valence externe des atomes et molécules, de nombreux processus électroniques prennent place avec une forte probabilité : ionisation (simple et multiple), excitation électronique, capture d’électron (simple et double). Nous avons mesuré les sections efficaces absolues de tous ces processus. Un autre aspect intéressant de la collision concerne la fragmentation des molécules excitées, que nous avons également mesurée précisément grâce à un dispositif dédié. Les expériences ont été effectuées auprès de l’accélérateur Tandem d’Orsay avec des faisceaux de quelques MeV d’énergie cinétique. Le dispositif AGAT a permis de réaliser les collisions (en condition de collision unique) et de mesurer tout à la fois les sections efficaces des processus et la fragmentation associée. Parallèlement nous avons simulé ces collisions d’un point de vue théorique en utilisant le modèle à Atomes et Electrons Indépendants (IAE) couplé à des calculs CTMC (Classical trajectory Monte Carlo). Sur cette base, nous avons prédit les sections efficaces qui se sont trouvées être en bon accord avec les mesures, à l’exception de la double capture d’électrons. Par ailleurs les rapports de branchement de dissociation des CnN⁺ après excitation électronique sont bien reproduits en utilisant la distribution d’énergie interne des espèces calculées avec le même modèle IAE/CTMC. Ces expériences nous ont permis de construire des « Breakdown Curves » (BDC), véritables cartes d’identité des molécules qui permettent de prévoir, dans le cadre d’une fragmentation statistique comment va fragmenter un système dont on connait l’énergie interne. Avec ces BDC nous avons pu prédire et recommander des rapports de branchement pour des voies de sortie de processus physiques et chimiques d’intérêt astrochimique. Ces données seront insérées dans la base internationale d'astrochimie the Kinetic Data Base for Astrochemistry KIDA. Cette thèse a été réalisée dans le cadre de l’Ecole Doctorale Ondes et Matière (EDOM) à l’Institut des Sciences Moléculaires d’Orsay (ISMO), à l’Université Paris-Sud Paris Saclay. / This thesis studies the aftermath of collision between singly positively charged Nitrogenated carbon species CnN⁺ (n=0,1,2,3) and neutral Helium atom at a velocity of 2.25 au. At this velocity, close to the velocity of outer electrons in atoms and molecules, several electronic processes take place and are near their maximum of probability such as ionisation (single, double, triple …), electronic excitation and electron capture (single and double). We looked at their cross sections and how their evolution with the molecule size. Following the collision the molecule can fragment, which leads to another interesting aspect, the fragmentation branching ratios. Collision experiments were done using a Tandem accelerator at Orsay that produced the CnN⁺ projectiles and a dedicated set-up, AGAT, to capture the flying fragments/intact molecule after collision according to their charge to mass ratio. Knowing the number of particles that are shot and the fact that our set-up allows no loss of fragments/intact molecule, we could get the probabilities of various fragments formed. Using these probabilities and a knowledge of the Helium jet profile used, we could measure their cross sections. The probabilities alone are sufficient to obtain the fragmentation branching ratios.The next step was to use a theoretical model to simulate the collision. We used Independent Atom and Electron (IAE) model coupled with Classical Trajectory Monte Carlo (CTMC) method to calculate the desired cross sections. A general good agreement was obtained, with the exception of double electron capture. The model could also predict, through the calculation of the species internal energy, the fragmentation branching ratios of cations CnN⁺ after electronic excitation. Also, the branching ratios were used to construct semi-empirical Breakdown Curves (BDCs), which are internal energy dependent dissociation branching ratios specific to each molecule, type, size and charge. With those, we could recommend products branching ratios to be used for various processes of astrochemical interest. The products branching ratios will be made available for a wider network of researchers under the international Kinetic Database for Astrochemistry (KIDA).This thesis was realized under the doctoral programme of Ecole Doctorale Ondes et Matiere (EDOM) with Institut des Sciences Moléculaires d’Orsay (ISMO) where the author was given an office and Université Paris-Sud where the author is formally enrolled.

A common injury type associated with vessel groundings on coral reefs is generation of coral rubble. As a substrate rubble is dynamic and can be mobilized by normal wave agitation, causing secondary damage to surviving corals and preventing coral larvae and invertebrates from surviving. Typical rubble management options include offsite removal and onsite stabilization using artificial materials. Sponges show promise as a restoration tool for rubble stabilization, especially sponges that are known to have rapid regeneration rates, low mortality, and the capability to quickly form attachment points to natural and artificial substrates. As an effective restoration tool a sustainable source of sponge is required in order to minimize impact on natural populations. The purpose of this project is to evaluate the feasibility of sponge propagation for use as a restoration material following groundings on coral reefs in Biscayne National Park, FL. If small sponge fragments could be collected from donor sponges and propagated on artificial growout structures (i.e. nurseries) a sustainable source of sponge fragments could be maintained for future restoration activities. If feasible, this technique would create a continually growing sponge source that would require a single harvest, reducing pressure on the local sponge populations. Fragments of four species of native sponges, Aplysina fulva¸ Amphimedon compressa, Aplysina cauliformis, and Iotrochota birotulata, were harvested and attached to artificial growout structures composed of PVC rods and cement paver tiles. Fragment necrosis, attachment, sponge retention (persistence throughout the study), complexity, and growth rates for all four species were recorded monthly for one year. Two different PVC types were utilized to compare attachment material effects on growth. Donor sponges were monitored to assess recovery and growth after harvesting, and intact sponges in the natural population were also monitored as a control. Transplanted fragments had the highest retention of any treatment (96-100% of fragments remained), growth rates ranging from 105% to 424%, low necrosis rates, and all but one species increased in complexity. There was no significant difference in sponge growth on different PVC types or between sites. Within one month all donor wounds inflicted by harvest were completely healed, and by the end of the study donors had significantly higher growth rates than the control treatment, ranging from 400% to 781% of initial sponge volume. Donor sponges also had low necrosis rates, but were less complex than controls. This is a promising, practical and low cost method, and based on these results, is a feasible way of producing sponges for us in restoration, with little negative impact on the natural populations.

Les polymères issus de la biomasse génèrent depuis quelques années un engouement certain puisqu’ils apparaissent comme de potentiels substituts aux polymères issus de l’industrie pétrolière. Parmi ces monomères récemment développés, l’acide succinique bio-sourcé a reçu une attention particulière notamment pour des applications dans le domaine des polyesters tels que le poly(butylène succinate). La présente thèse décrit la synthèse de ce polymère par estérification directe de l’acide succinique et du 1,4-butanediol dans un réacteur pilote de 7.5 L. Les principaux paramètres du procédé tels que l’excès de diol, la température de trans-estérification ainsi que la pureté de l’acide succinique ont été étudiés. Une attention particulière a été portée sur le choix du catalyseur (son type, la quantité utilisée …) afin d’observer son influence sur le procédé ainsi que sur la stabilité du polymère final. Puis différentes stratégies de modification du poly(butylene succinate) ont été testées pour améliorer à la fois sa mise en forme par extrusion gonflage et les propriétés des films obtenus. Dans ce but l’introduction d’agents de branchements, de silices nanométriques ainsi que des comonomères rigides a été étudiée. / Polymers issued from biomass present a growing interest, since they seem to be a suitable alternative to conventional petrochemical polymers. Among the newly developed monomers, bio-based succinic acid received a particular attention for its application in the synthesis of aliphatic polyesters such as poly(butylene succinate). The present thesis reports the synthesis of this polymer via the direct esterification of succinic acid and 1,4-butanediol in a 7.5 L pilote scale reactor. Main process parameters such as the diol exces, the trans-esterification temperature as well as the purity of succinic acid have been studied. In addition a special attention was taken to highlight the influence of the catalyst (its type, quantity ...) on the synthesis and on the stability of the resulting polymer. Then several strategies of modification of poly(butylene succinate) have been studied to improve the processability of the polymer via film extrusion blowing and to enhance the properties of polymer films. To that end the introduction of branching agents, silica nanofillers as well as rigid comonomers have been studied.

This master thesis describes deterministic context-free L-systems and its context in procedural modeling, especially in fractal geometry, deals with rewriting technique and its usage for modeling structures similar to plants. Further it describes more complex types of L-systems, especially their context and parametric variations, and shows usage of L-systems in computer graphics and describes its usage for procedural modeling of architecture. At the end of this thesis there are described some other possibilities of usage procedural modeling with L-systems and introduced some extensions of rewriting rules, which will be subject of future research.

Stress corrosion, SC, in some cases gives rise to stress corrosion cracking, SCC, which differs from purely stress intensity driven cracks in many aspects. They initiate and grow under the influence of an aggressive environment in a stressed substrate. They grow at low load and may branch. The phenomenon of SCC is very complex, both the initiation phase and crack extension itself of SCC is seemingly associated with arbitrariness due to the many unknown factors controlling the process. Such factors could be concentration of species in the environment, stress, stress concentration, electrical conditions, mass transport, and so on.In the present thesis, chemically assisted crack initiation and growth is studied with special focus on the initiation and branching of cracks. Polycarbonate plates are used as substrates subjected to an acetone environment. Experimental procedures for examining initiation and branching in polycarbonate are presented. An optical microscope is employed to study the substrate.The attack at initiation is quantified from pits found on the surface, and pits that act as origin for cracks is identified and the distribution is analysed. A growth criterion for surface cracks is formulated from the observations, and it is used to numerically simulate crack growth. The cracks are seen to coalesce, and this phenomenon is studied in detail. Branching sites of cracks growing in the bulk of polycarbonate are inspected at the sample surface. It is found that the total width of the crack branches are approximately the same as the width of the original crack. Also, angles of the branches are studied. Further, for comparison the crack growth in the bulk is simulated using a moving boundary problem based algorithm and similar behaviour of crack branching is found. / <p>Both papers in thesis as manuscript, paper II with title "Branching cracks in a layered material - Dissolution driven crack growth in polycarbonate"</p>