Search

International audience

International audience; This paper gives an overview of a multi-scale experimental study of the secondary hydriding phenomena that can occur in nuclear fuel cladding materials exposed to steam at high temperature (HT) after having burst (loss-of-coolant accident conditions). By coupling information from several facilities, including neutron radiography/tomography, electron probe micro analysis, micro elastic recoil detection analysis and micro laser induced breakdown spectroscopy, it was possible to map quantitatively, at different scales, the distribution of oxygen and hydrogen within M5™1 clad segments having experienced ballooning and burst at HT followed by steam oxidation at 1100 and 1200 °C and final direct water quenching down to room temperature. The results were very reproducible and it was confirmed that internal oxidation and secondary hydriding at HT of a cladding after burst can lead to strong axial and azimuthal gradients of hydrogen and oxygen concentrations, reaching 3000–4000 wt ppm and 1.0–1.2 wt% respectively within the β phase layer for the investigated conditions. Consistent with thermodynamic and kinetics considerations, oxygen diffusion into the prior-β layer was enhanced in the regions highly enriched in hydrogen, where the α(O) phase layer is thinner and the prior-β layer thicker. Finally the induced post-quenching hardening of the prior-β layer was mainly related to the local oxygen enrichment. Hardening directly induced by hydrogen was much less significant.

International audience; Plasticity of hexagonal titanium-aluminum alloys depends on the solute concentration and the order state of Al atoms. Development of short-range order (SRO) modifies Al strengthening and impacts the competition between prismatic and basal slip modes. Using ab initio calculations, we study the interaction of a screw dislocation with isolated Al atoms and Al pairs in a hcp Ti existing in short-range ordered Ti-Al alloys. Calculated activation energies reveal pronounced hardening caused by Al addition for both slip systems, which become competitive. This hardening is enhanced for Al pairs, resulting in a reduced plastic anisotropy of Ti-Al alloys with SRO.

International audience; The combination of electronic and nuclear energy deposition may have a significant effect on the defect production and evolution. In nuclear reactor, the nuclear fuel is exposed to the irradiation of several particles such as the fission fragments (FF). Along their path, the ratio of electronic to nuclear energy loss evolves. To understand the impact of this coupling on the fuel microstructure, single and dual-beam ion irradiations of uranium dioxide (UO2) were carried out. The damage evolution was investigated by Raman spectroscopy analysis correlated with Transmission Electronic Microscopy observations. A significant effect of electronic energy dissipation on defect formation and evolution was found, depending on the electronic energy loss level. With the increase of electronic energy loss, the microstructure evolution is even more pronounced.

International audience; Fe/Cr/Fe trilayers and multilayers are prepared as model systems designed to furnish simple data comparable with calculation results for diffusion properties in nuclear materials. Their structure (epitaxy, residual strains and dislocations) is characterized in detail. The film structure (strain and stress) is shown to be different on MgO20nm/SrTiO3 and MgO substrates due to the residual strain in the MgO buffer layer on SrTiO3. Superlattices with high crystalline quality are prepared, with Fe and Cr in coherent epitaxy. In-plane residual strain in Fe is +0.45(13)% on MgO substrates and decreases from 1.70(9)% to 0.47(2)% when increasing the thickness of the trilayers on MgO/SrTiO3 substrates. These strains enhance the contrast between Fe and Cr, opening the way to future kinetics studies using x-ray diffraction in this system, which is far more efficient (non-destructive and rapid) than high resolution transmission electron microscopy with electron energy loss spectroscopy or atom probe tomography

International audience; To explain irradiation creep, several mechanisms have been proposed. Some are based on the effect of stress on either nucleation or growth of dislocation loops. To investigate these mechanisms in aluminum we combine in-situ transmission electron microscope irradiation under stress and two simulation approaches (object kinetic Monte-Carlo, molecular dynamics Frenkel pair accumulation). We observe the selectivity of Frank loop variants under electron irradiation and applied stress. When the stress is turned on after loop formation, there is no loop variant selectivity, suggesting the absence of preferential absorption on already formed loops. Object kinetic Monte-Carlo simulations, including the effect of stress on the diffusion of point defects, show no growth rate difference between loop variants. Frenkel pair accumulation simulations exhibit variant selectivity of nucleated loops. This shows that loop selectivity is due to preferential nucleation of well oriented loops under stress and not to differential growth of loops.

The present work studied the evolution of stresses and associated mechanisms, related to viscoplastic behavior during relaxation, in the zirconia layer formed during the oxidation of Zircaloy-4 under a He/O2 mixture, at temperatures of 700 °C, 800 °C and 900 °C. Measurements by X-ray diffraction are performed in-situ under synchrotron radiation during oxidation to determine the evolutions of phases and stresses in the oxide layer with time. The results show that the zirconia formed contains a mixture of monoclinic and tetragonal phases. The proportion of the tetragonal phase depends on the oxidation temperature and decreases during oxidation. In order to better understand the influence of this evolution and localize the tetragonal phase, measurements by Raman spectroscopy have also been performed. These two phases are subjected to compressive stresses in directions perpendicular to the oxide layer growth direction. These stresses depend on temperature and relax during oxidation. Two mechanical models either considering the zirconia phases independently or considering an equivalent homogeneous oxide are proposed to describe the evolution of these stresses, considering that it is due to oxide viscoplasticity. The model parameters are analysed to discuss the mechanisms of viscoplastic flow in the oxide. Numerical values for the viscoplastic parameters of the model as well as for the corresponding activation energy are therefore provided.

Small addition of oxygen is currently used to improve the mechanical properties of titanium. It has been thought, until now, that oxygen fully dissolves in the titanium matrix, thus leading to a solid solution. However, we show here, with transmission electronic microscopy and X-ray diffraction, that oxygen can also form ordered precipitates with a Ti6O-type structure, even for a concentration as low as 0.15 wt% O. The ordered precipitates found in a series of pure binary Ti–O alloys are also visible in commercially pure titanium, including grades 2 and 4, two compositions widely used for technological applications. Oxygen-ordered precipitates are experimentally evidenced in a series of Ti–O alloys and in CP titanium. These findings question the paradigm defining oxygen only as a solute atom in titanium alloys.

International audience; Concentration profiles on the different sites of tungsten screw dislocation are calculated as a function of the nominal carbon atom concentration, temperature and dislocation density. Rigid lattice Monte Carlo simulations and associated mean-field modelling are performed based on interaction energies from ab initio calculations. We show that the repulsive interactions induce ordering phenomena that drastically change the concentration in the core of the dislocation and leads to ordered structures among the octahedral sites. We detail the domains of short-range ordering (SRO) and long-range ordering (LRO).

International audience; Free energy calculations in materials science are routinely hindered by the need to provide re-action coordinates that can meaningfully partition atomic configuration space, a prerequisite formost enhanced sampling approaches. Recent studies on molecular systems have highlighted thepossibility of constructing appropriate collective variables directly from atomic motions throughdeep learning techniques. Here we extend this class of approaches to condensed matter problems,using autoencoder neural networks to construct a scalar collective variable for use with the adaptivebiasing force method. Particular attention is given to design choices required for application tocrystalline systems with defects, including the filtering of thermal motions which otherwise domi-nate the autoencoder input. The machine-learning workflow is tested on body-centered cubic ironand its common defects, such as small vacancy or self-interstitial clusters and screw dislocations.For localized defects, excellent collective variables as well as derivatives, necessary for free energysampling, are systematically obtained. However, the approach has a limited accuracy when dealingwith reaction coordinates that include atomic displacements of a magnitude comparable to thermalmotions, e.g. the ones produced by the long-range elastic field of dislocations. We then combinethe extraction of collective variables by autoencoders with an adaptive biasing force free energymethod based on Bayesian inference. Using a vacancy migration as an example, we demonstratethe performance of coupling these two approaches for simultaneous discovery of reaction coordinatesand free energy sampling in systems with localized defects.

International audience; Oxygen content has always been limited in commercial titanium and titanium alloys due to its propensity to induce a severe ductility loss. Yet, its effect on the macroscopical behavior has never been clearly understood and is still rather unclear considering the wide variability in the literature results. Here, we investigate the tensile properties of α-titanium with oxygen contents ranging from 0.15 to 0.80 weight percent (wt%). While the strain-hardening ability of oxygen is maintained, no ductility drop is observed up to 0.60 wt% of oxygen, thus allowing exceptional combinations of mechanical properties with an ultimate tensile strength (UTS) of 800 MPa and 29% of elongation at fracture for the Ti-0.6O alloy. Both high strength and ductility of these alloys result from the dislocations/precipitate's interactions. It is proposed that these interactions induce an important cross-slip ac­tivity responsible for a dislocation multiplication and a high work-hardening rate. With the addition of Zr, alloys exhibit an even more promising combination of mechanical properties, achieving 1,075 MPa of UTS and 28% of elongation at fracture for the Ti-4.5Zr-0.8O alloy. The mechanical properties of TiO and TiZrO alloys brought out in this study surpass those of Ti-6Al-4V alloy and open significant prospects for developing a new generation of oxygen-tolerant titanium alloys.

International audience; Advanced austenitic materials are foreseen as cladding for next-generation fast reactors. They can suffer from void swelling, which is related to radiation-induced dislocation that may be strongly affected by chemical composition. Yet, the underlying mechanism leading to this latter effect is still unclear. Here, we investigate pure Ni, Ni-0.4wt.%Cr and Ni-0.4/0.8/1.2wt.%Ti as face-centered-cubic model materials, to gain an insight into the fundamental mechanism of solute effects on radiation damage in the early stage of irradiation. The microstructural evolution is studied using transmission electron microscopy in-situ self-ion irradiation at 450 and 510°C for doses up to 0.18 dpa, and supported by modelling using kinetic rate theory.Results reveal that micro-alloying has a spectacular impact on the dislocation loop nature, Burgers vector, mobility, number density, size, growth rate, and the critical sample thickness for their formation. A tiny amount of Ti effectively reduces the loop mobility and growth rate, and stabilizes self-interstitial loops against vacancy loops, depending on Ti content and temperature. Based on kinetic rate theory including production bias, this is explained by the suppression of interstitial loss to the surfaces. In the alloys, the equal distribution of Burgers vector and the relative strong increase of loop density is ascribed to the lower loop mobility and stacking fault energy, while both experiment and modeling show that irradiation leads to a Ti depletion at defects. Based on DFT data and present experimental evidences, we attribute these effects to the thermal segregation of oversized Ti atoms at strained lattice sites surrounding the defects.

It is generally considered that the elementary building blocks of defects in face-centred cubic (fcc) metals, e.g., interstitial dumbbells, coalesce directly into ever larger 2D dislocation loops, implying a continuous coarsening process. Here, we reveal that, prior to the formation of dislocation loops, interstitial atoms in fcc metals cluster into compact 3D inclusions of A15 Frank-Kasper phase. After reaching the critical size, A15 nano-phase inclusions act as a source of prismatic or faulted dislocation loops, dependent on the energy landscape of the host material. Using cutting-edge atomistic simulations we demonstrate this scenario in Al, Cu, and Ni. Our results explain the enigmatic 3D cluster structures observed in experiments combining diffuse X-ray scattering and resistivity recovery. Formation of compact nano-phase inclusions in fcc structure, along with previous observations in bcc structure, suggests that the fundamental mechanisms of interstitial defect formation are more complex than historically assumed and require a general revision. Interstitial-mediated formation of compact 3D precipitates can be a generic phenomenon, which should be further explored in systems with different crystallographic lattices.

The fuel used in pressurised water reactors (PWRs) is uranium dioxide UO₂ in the form of pellets in zirconium alloy cladding. Uranium dioxide is known to be thermodynamically unstable at room temperature and normal pressure and to oxidise. As the oxygen stoichiometry increases, the uranium oxide changes phase. The known hyper-stoichiometric phases are, in order of increasing oxidation: UO₂ , U₄O₉ , U₃O₇ then U₃O₈. In order to avoid any exposure of the radioactive fuel, and thus to meet safety criteria, a thorough knowledge of the material, its structures and phase transitions is therefore required. In this thesis, we explore a part of UO₂₊ₓ phase diagram in an attempt to add to the knowledge of the material. We are only interested in the phases between the stoichiometries UO₂ and U₄O₉. Between these stoichiometries, at low temperatures, only the two phases UO₂ and U₄O₉ coexist. As the temperature increases, a phase transition to a single-phase domain UO₂₊ₓ takes place. Calculations at the atomic scale using empirical potentials will allow an in-depth study of the different phases and transitions involved. A molecular dynamics study with different empirical potentials was launched. It allows a study of the UO₂₊ₓ material at different oxygen concentrations, with x between 0 and 0.25 and for different temperatures between 100K and 2000K. A description of the different phases of U₄O₉ and an in-depth study of the aggregates formed was made possible. The aggregate in U₄O₉ phase is the cuboctahedron (COT), which consists of 13 oxygen atoms. Their arrangement in the structure and geometry are studied. This method has limitations for the study of the phase transition. It is to overcome these limitations and to obtain a better exploration of the material that the Path Sampling Monte Carlo code is developed. This code is based on sampling pathways allowing the progressive addition and removal of an oxygen atom in the UO₂ mesh. The progressive aspect is coupled with molecular dynamics allowing the oxygen particle to be slowly inserted or de-inserted while letting the other atoms in the lattice move. This dynamic method reaches a better convergence as it allows the system to relax and find the most favourable location for the particle. Following these insertion and de-insertion of an oxygen atom, the chemical potential is estimated for each concentration between UO₂ and U₄O₉. Conditioning methods are applied to obtain a converged value in a short calculation time. The evolutions of the chemical potential at all concentrations between UO₂ and U₄O₉ and for ten temperatures between 200K and 2000K are obtained. These evolutions made it possible to identify precisely each COT formation in the mesh. The temperature and concentration at which it is no longer favourable to form them was also retrieved as well as the temperature at which the phase transition from U₄O₉ to the UO₂₊ₓ domain occurs. An in-depth exploration of the structures was also carried out in parallel, confirming what the chemical potential suggested. In conclusion, the UO₂₊ₓ phase diagram was explored with two computational methods, molecular dynamics and the Path Sampling Monte Carlo code.; Le combustible utilisé dans une grande partie des réacteurs à eau pressurisée (REP) est le dioxyde d'uranium UO₂ présent sous forme de pastille dans des gaines en alliage de zirconium. Le dioxyde d'uranium UO₂ est connu pour être instable thermodynamiquement à température et pression ambiante et pour s'oxyder. Lorsque la stœchiométrie en oxygène augmente, l'oxyde d'uranium change de phase. Les phases sur-stœchiométriques connues sont, dans l'ordre d'oxydation croissante : UO₂ , U₄O₉ , U₃O₇ puis U₃O₈. Pour éviter toute exposition du combustible radioactif à l'air, et donc pour respecter des critères de sécurité, il faut une connaissance approfondie du matériau, de ses structures et ses changements de phases sont donc nécessaires. Dans cette thèse, nous explorons une partie du diagramme de phases UO₂₊ₓ pour tenter d'apporter des éléments supplémentaires à la connaissance de l'oxyde d'uranium. Nous nous intéressons uniquement aux phases comprises entre les stœchiométries UO₂ et U₄O₉. Entre ces stœchiométries, aux basses températures, seules les deux phases UO₂ et U₄O₉ coexistent. Lorsque la température augmente, une transition de phase vers un domaine monophasé UO₂₊ₓ s'opère. Des calculs à l'échelle atomique à l'aide de potentiels empiriques vont permettre une étude approfondie des différentes phases et transitions mises en jeu. Dans un premier temps une étude en dynamique moléculaire avec différents potentiels empiriques est lancée. Elle permet une étude du matériau UO₂₊ₓ à différentes concentrations en oxygène, avec x compris entre 0 et 0.25 et pour différentes températures entre 100K et 2000K. Une description des différentes phases de U₄O₉ et une étude approfondie des agrégats formés a été rendue possible. L'agrégat majoritaire dans la phase U₄O₉ est le cuboctaèdre (COT), un agrégat constitué de 13 atomes d'oxygène. Leur arrangement dans la maille et leur géométrie en fonction de la température sont étudiés. Cette méthode présente des limites pour l'étude de la transition de phase. C'est pour palier à ces limites et obtenir une meilleure exploration du matériau que le code Monte Carlo des Chemins est développé. Ce code se base sur une approche d'échantillonnage de l'espace des chemins permettant l'ajout et le retrait progressif d'un atome d'oxygène dans la maille d'UO₂. L'aspect progressif est couplé avec une dynamique moléculaire permettant d'insérer ou désinsérer lentement la particule d'oxygène tout en laissant les autres atomes de la maille bouger. Cette méthode dynamique permet une meilleure convergence puisque permet au système de relaxer et trouver l'emplacement le plus favorable pour la particule. A la suite de cet échantillonnage des chemins permettant l'insertion ou la désinsertion d'un atome d'oxygène, le potentiel chimique est estimé pour chaque concentration entre UO₂ et U₄O₉. Des méthodes de conditionnement sont appliquées pour obtenir une valeur convergée dans un temps de calcul minimal. Ces calculs permettent d'obtenir l'évolution du potentiel chimique à toutes les concentrations entre UO₂ et U₄O₉ et pour dix températures entre 200K et 2000K. Les évolutions du potentiel chimique ont permis d'identifier précisément chaque formation de COT dans la maille. Puis, la température et la concentration à partir desquelles il n'est plus favorable d'en former et la transition de phases du domaine U₄O₉ vers UO₂₊ₓ. Une exploration approfondie des structures a aussi été menée en parallèle permettant de confirmer ce que laissait supposer le potentiel chimique. En conclusion le diagramme de phases UO₂₊ₓ a été exploré avec deux méthodes de calculs, de la dynamique moléculaire et le code en Monte Carlo des Chemins.

The burning of fossil fuels is the major culprit of greenhouse gases emission contributing to the global climate crisis. Among many alternative energy solutions, the nuclear power stands out for its capacity in massive production. However, the intense nuclear reactions create a detrimental environment for reactor materials. To understand the material damages caused by radiations, experiments traditionally use neutron beams. These experimental settings are often expensive with samples difficult to manipulate. These disadvantages are the reasons why the use of ion beams becomes a popular alternative. One key parameter for the ion irradiation is the electronic stopping power, whose value is usually obtained from the SRIM code. This empirical method is based on a very complete database. However, it may lose its reliability when the irradiated systems have hardly any experimental records in the database. Calculations from first-principles thus provide supplementary references to experiments. Implemented inside the MOLGW code, our simulation method for the electronic stopping power is based on the time-dependent density functional theory (TDDFT). Using localized Gaussian-type orbital basis functions, our code can describe all the electrons, core and valence. The main goal of this thesis is to incorporate basis functions following the movement of the projectile and simulate the irradiation of ions containing electrons. Modifications to the time-dependent equations, notably the appearance of an operator D, are necessary due to the movement of projectile basis. This extra operator enforces the conservation of the wave-function orthonormality and has an amount of energy contribution whose numerical precision depends on the basis set completeness. The Crank-Nicolson propagator is chosen over the second-order Magnus propagator due to less approximations required in the non-orthogonal basis. A predictor-corrector scheme is applied to evaluate the propagated states. In order to reduce the computational costs of large matrix updates during the propagation, only the cross target-projectile elements are selected for recalculations. The sanity of our implementations is verified by tests in atomic systems before applying the code to solid targets. Since our approach works with finite systems, the infinite crystal structure is represented by a cluster. Two main groups of modeling parameters vital to the convergence are studied: the cluster geometry and the basis set quality. We advocate for the use of a long and thin cylindrical shape for the clusters to guarantee the stabilization of the energy transfer between the projectile and the target. High quality cc-pVXZ Dunning basis are needed for target atoms and projectiles with core electrons. For lighter projectiles the standard cc-pVXZ basis is sufficient. The channeling stopping powers of hydrogen and helium are studied in fcc [001] aluminum target and converge faster when the cluster is centered on its channel rather than on its central column of atoms. The random electronic stopping powers of hydrogen and helium in lithium and aluminum are presented, as well as the results for aluminum ion in aluminum. Our calculations, although not fully converged, provide accurate qualitative insights of the stopping power behaviors in good general agreement with the SRIM prediction, the experimental data and the results obtained with plane-wave basis method from another group. We also report the random stopping powers of proton and antiproton in LiF calculated without projectile basis from a side project. Our results describe correctly the proton stopping power observed by experiments like the threshold velocity and the Barkas effect where the stopping power of antiproton is inferior to that of proton. Additionally, we have observed an unprecedented negative Barkas effect at very low velocities due to the destabilization of the 2p orbitals of F⁻ ions in the presence of a close antiproton., La consommation de combustibles fossiles est un facteur principal des émissions de gaz à effet de serre. Face à une crise climatique planétaire, des solutions énergétiques alternatives s'imposent. L'énergie nucléaire présente l'avantage de la production de masse, mais expose les matériaux de structure à un environnement extrême dont les dommages doivent être étudiés. Les expériences traditionnelles utilisant des faisceaux de neutrons sont souvent coûteuses avec des échantillons difficiles à manipuler. Une meilleure alternative est l’irradiation aux ions. Une quantité importante mesurée lors de l’irradiation aux ions est le pouvoir d’arrêt, dont la référence la plus standard est le SRIM. Cette méthode empirique de calculs est ajustée sur une base de données expérimentales très complète. Mais pour les irradiations peu documentées, la fiabilité du SRIM n'est pas garantie et les calculs ab initio peuvent servir de références supplémentaires. Implémentée au sein du code MOLGW, notre méthode de simulation s'appuie sur la théorie de la fonctionnelle de la densité dépendante du temps (TDDFT). L'utilisation de fonctions de base localisées du type gaussienne permet de décrire tous les électrons. L'objectif de cette thèse est d'introduire des fonctions de bases localisées suivant le projectile pour prendre en compte de ses électrons. Puisque la base du projectile se déplace, les équations de Schrödinger dépendantes du temps doivent être modifiées. On constate notamment l’apparition d’un opérateur D indispensable à la conservation d’orthonormalité de fonction d’onde. Cet opérateur contribue à une quantité d'énergie dont la précision numérique dépend fortement de la qualité de base. Le propagateur de Crank-Nicolson est choisi à la place du propagateur de Magnus du second ordre, en raison du fait que ce premier nécessite moins d'approximations dans une base nonorthogonale. Une stratégie de prédiction-correction est appliquée à la propagation. La mise à jour de matrices de grande taille demande beaucoup d'efforts de calcul pendant la propagation. Afin de réduire le coût, seuls les termes croisés entre la cible et le projectile sont sélectionnés pour le recalcul. L'implémentation a été validée par des tests de collisions atomiques avant d'être appliquée aux cibles solides. Notre approche traite les systèmes finis et les structures cristallines sont décrits par les agrégats. Deux groupes de paramètres de modélisation sont vérifiés par rapport à la convergence : la géométrie de l’agrégat et la qualité de base. On conseille un agrégat de forme cylindrique fin et long pour assurer la stationnarité du transfert d'énergie. Quant à la base, les séries cc-pVXZ de Dunning sont les plus adaptées pour les atomes de cible ainsi que pour les projectiles possédant d’électrons de coeur. Pour les projectiles plus légers, les séries standard cc-pVXZ suffisent. Les pouvoirs d’arrêt canalisés le long des directions cristallines (channeling en anglais) d’hydrogène et d’hélium sont étudiés dans un agrégat d’aluminium fcc [001]. Les résultats du channeling convergent plus rapidement quand l'agrégat est centré sur le canal. Les pouvoirs d’arrêt moyennés (random en anglais) d’hydrogène et d’hélium dans le lithium et l’aluminium sont analysés, ainsi que celui d’aluminium dans l’aluminium. Malgré un léger manque de convergence, nos résultats sont qualitativement corrects et en bon accord avec le SRIM, les données expérimentales et les résultats obtenu en base d'ondes planes par un autre groupe. Enfin une étude en parallèle de pouvoirs d’arrêt de proton et d’antiproton dans le LiF calculés sans la base de projectile sont montrés. Nos simulations décrivent correctement les comportements du pouvoir d’arrêt de proton observés par les expériences. En outre, on prédit un effet de Barkas négatif sans précédent à très basse vitesse. On explique cet effet par la déstabilisation d’une orbitale 2p d’ion F⁻ par un antiproton proche.

International audience; Crystal strength and plastic flow are controlled by the motion and interaction of dislocations, the line defects carrying atomic shear increments. Whereas, in most crystals, deformation develops in the crystallographic planes in which the glide force acting on dislocations is maximum, plasticity in body-centred cubic metals is more complex. Slip systems in which the resolved shear stress is not the highest can dominate at low temperature, leading to anomalous slip. Using in situ tensile tests in a transmission electron microscope we show that anomalous slip arises from the high mobility of multi-junctions , that is, junctions between more than two dislocations, which glide at a velocity several orders of magnitude larger than single dislocations. These multi-junctions result from the interaction of a simple binary junction with a gliding dislocation. Although elasticity theory predicts that these binary junctions should be unstable in crystals with a weak elastic anisotropy such as tungsten, both experiments and atomistic simulations reveal that such junctions can be created under dynamic conditions, in agreement with the existence of anomalous slip in almost all body-centred cubic metals, including tungsten .

International audience; Sampling the Minimum Energy Path (MEP) between two minima of a system is often hindered by the presence of an energy barrier separating the two metastable states. As a consequence, direct sampling based on Molecular Dynamics or Markov Chain Monte Carlo methods becomes inefficient, the crossing of the energy barrier being associated to a rare event. Augmented sampling methods based on the definition of collective variables or reaction coordinates allow to circumvent this limitation at the price of an arbitrary choice of the dimensionality reduction algorithm.We couple the statistical sampling techniques, namely Metadynamics and Invertible Neural Networks, with autoencoders so as to gradually learn the MEP and the collective variable at the same time. Learning is achieved through a succession of two steps: statistical sampling of the most probable path between the two minima and re-definition of the collective variable from the updated data points. The prototypical Mueller potential with nearly orthogonal minima is employed to demonstrate the ability of such coupling to unravel a complex MEP.

The damage induced in GaAs crystals irradiated with dual-ion beam (low-energy I2+ and high-energy Fe9+), producing simultaneous nuclear ( Sn) and electronic ( Se) energy depositions, was investigated using several characterization techniques. Analysis of the damage buildup shows that Sn alone (single 900 keV ion beam) leads, in a two-step process, to full amorphization of the irradiated layer (at a fluence of 1.5 nm−2) and to the development of a high (2.2%) elastic strain. Conversely, only one step in the disordering process is observed upon dual-ion beam irradiation (i.e., 900 keV I2+ and 27 MeV Fe9+, Sn& Se); hence, amorphization is prevented and the elastic strain remains very weak (below 0.2%). These results provide a strong evidence that, in GaAs, the electronic energy deposition can induce an efficient dynamic annealing of the damage created in collision cascades formed during nuclear energy deposition.

International audience; The elastic interactions between dislocations and lattice point defects (PDs) control the rate of PD elimination at dislocations, and thereby the PD spatial distribution within the microstructure. By relying on the Onsager formalism, we introduce a new partition of the PD diffusion driving force in a strain field. We present DFT-based calculations of straight and loop-form edge dislocation absorption efficiences and the resulting biases of vacancy and self-interstitial elimination in pure Fe, Ni, and Al metals. We highlight the effects of the dislocation density, the temperature and the dislocation orientation on the elastodiffusion behavior and the resulting PD redistribution near dislocations. Based on a detailed study of the sensitivity of absorption efficiencies to the boundary conditions, i.e., the capture radius and the associated PD concentrations, we introduce the notion of asymptotic absorption efficiencies and bias.

International audience; The interaction of C atoms with a screw and an edge dislocation is modelled at an atomic scale using an empirical Fe-C interatomic potential based on the Embedded Atom Method (EAM) and molecular statics simulations. Results of atomic simulations are compared with predictions of elasticity theory. It is shown that a quantitative agreement can be obtained between both modelling techniques as long as anisotropic elastic calculations are performed and both the dilatation and the tetragonal distortion induced by the C interstitial are considered. Using isotropic elasticity allows to predict the main trends of the interaction and considering only the interstitial dilatation will lead to a wrong interaction.

The comprehension of the fundamental mechanisms controlling the evolution and mutual interactions of extended defects such as dislocation and cavities is crucial to understand the behaviour of materials under extreme conditions. Underlying phenomena are hard to capture, either by numerical simulations or by experimental approaches. Therefore, building a complete mathematical formalism which can be efficiently implemented is at stake. In this work we develop a phase-field variational model that couples vacancy diffusion, dislocation climb and pore evolution, with the consideration of elastic interactions. We first present the construction of the model as well as its numerical implementation, including an improved solver for the equation controlling the vacancy field evolution. This drastically decreases the computational time and allows us to tackle diffusion at mesoscopic scales. We then validate our model on microstructures whose evolution can be predicted analytically. We finally present 2D simulation results on the interactions between climbing dislocations and pore assemblies, revealing the role of elastic interactions on the microstructural evolution., La compréhension des mécanismes fondamentaux à l’origine de l’évolution des défauts cristallins tels que les lacunes, les dislocations ou encore les cavités, ainsi que de leurs interactions mutuelles, est essentielle pour prédire le comportement des matériaux métalliques en conditions extrêmes. Les couplages multi-physiques en jeu sont cependant difficiles à identifier, tant par des simulations qu'à l'aide de dispositifs expérimentaux. Nous proposons ainsi un modèle champ de phase pour décrire les interactions entres lacunes, cavités et dislocations. Nous présentons dans un premier temps sa construction ainsi que son implémentation numérique, pensée de manière à minimiser les coûts de calculs. Nous détaillons en particulier notre approche originale de la résolution de l’équilibre chimique qui nous permet de simuler des phénomènes mésoscopiques sur les temps longs de la diffusion. Nous validons dans un second temps notre modèle dans le cadre de l'évolution de microstructures qui peuvent être approchées de manière analytique. Enfin, dans une troisième partie, nous étudions le comportement d’une microstructure 2D comportant plusieurs cavités et des dislocations. Nous montrons notamment que la prise en compte des interactions élastiques entre les cavités et les dislocations, généralement négligées dans ce type d’étude, modifie significativement la cinétique d’évolution de la microstructure.

International audience; Atomic scale simulations at finite temperature are an ideal approach to study the thermodynamic properties of magnetic transition metals. However, the development of interatomic potentials explicitly taking into account magnetic variables is a delicate task. In this context, we present a tight-binding model for magnetic transition metals in the Stoner approximation. This potential is integrated into a Monte Carlo structural relaxations code where trials of atomic displacements as well as fluctuations of local magnetic moments are performed to determine the thermodynamic equilibrium state of the considered systems. As an example, the Curie temperature of cobalt is investigated while showing the important role of atomic relaxations. Furthermore, our model is generalized to other transition metals highlighting a local magnetic moment distribution that varies with the gradual filling of the d states. Consequently, the successful validation of the potential for different magnetic configurations indicates its great transferability and makes it a good choice for atomistic simulations sampling a large configuration space.

Atomic-scale simulations, and in particular molecular dynamics (MD), are key assets to model the behavior of the structure of materials under the action of external stimuli, say temperature, strain or stress, irradiation, etc. Despite the widespread use of MD in condensed matter science, some basic material characteristics remain difficult to determine. This is, for instance, the case for the long-range strain tensor, and its root-mean-squared fluctuations, in disordered materials. In this work, computational diffraction is introduced as a fast and reliable structural characterization tool of atomic-scale simulation cells in the case of irradiated single crystals. In contrast to direct-space methods, computational diffraction operates in the reciprocal space and is therefore highly sensitive to long-range spatial correlations. With the example of irradiated UO2 single crystals, it is demonstrated that the normal strains, shear strains and rotations, as well as their root-mean-squared fluctuations (microstrain) and the atomic disorder, are straightforwardly and unambiguously determined. The methodology presented here has been developed with efficiency in mind, in order to be able to provide simple and reliable characterizations either operating in real time, in parallel with other analysis tools, or operating on very large data sets.

International audience; Surface energy is a key quantity that controls many physical properties of materials, yet determining its value at the nanoscale remains challenging. By using N-body interatomic potentials and performing analytical calculations, we develop a robust approach to determine the surface energy of metallic nanoparticles as a function of particle size and temperature. A strong increase in the surface energy is obtained when the size of the nanoparticle decreases in both the solid and liquid states. However, we show how the use of the classical spherical approximation to characterize the surface area of a nanoparticle leads to an almost constant surface energy with size as usually done to characterize the thermodynamic and kinetic properties of NPs in many works. We then propose a correction of the spherical approximation that is particularly useful for small size nanoparticles to improve the different models developed in the literature so far.

International audience; To improve the safety of nuclear power plants, a Cr protective layer is deposited on zirconium alloys to enhance oxidation resistance of the nuclear fuel cladding during both in-service and hypothetical accidental transients at High Temperature (HT) in Light Water Reactors. The formation of the Cr2O3 film on the coating surface considerably helps in reducing the oxidation kinetics of the zirconium alloy, especially during hypothetic Loss of Coolant Accident (LOCA). However, if the Cr coating is successful to increase the oxidation resistance at HT of the zirconium substrate, for in-service conditions, under neutron irradiation, Cr desquamation has to be avoided to guarantee a safe use of the Cr-coated zirconium alloys. Therefore, the adhesion properties have to be maintained despite the structural defects created by sustained neutron irradiation in the reactor environment. This paper proposes to study the behavior of the Zircaloy-Cr interface of a first generation Cr-coated material during a specific in situ ion irradiation. As deposited, the Cr-coated sample presents a f.c.c. C15 Laves-type intermetallic phase at the interface with off-stoichiometric composition. After irradiation and for the specific conditions applied, this interfacial phase has significantly dissolved. Energy Dispersion Spectroscopy revealed that the dissolution was accompanied by a counterintuitive “sharpening” effect.

International audience; Directly computing linear mass transport coefficients in stochastic models entails integrating over time the equilibrium correlations between atomic displacements. Here, we show how to improve the accuracy of kinetic Monte Carlo simulations via correlation splitting and conditioning, which statistically amounts to estimating the mass transport coefficients through a law of total diffusion. We illustrate the approach with kinetic path sampling simulations of atomic diffusion in a random alloy model in which percolating solute clusters trap the mediating vacancy. There, Green functions serve to generate first-passage paths escaping the traps and to propagate the long-time dynamics. When they also serve to estimate mean-squared displacements via conditioning, colossal reductions of statistical errors are achieved.

International audience; We report the \textit{ab initio} prediction of a negative Barkas coefficient in LiF insulator at low velocity (v< 0.25 a.u. ~ 2 keV). The electronic stopping power of protons in lithium fluoride has been extensively studied both experimentally and theoretically because of a controversial threshold effect. While our time-dependent density-functional theory (TDDFT) simulations confirm the presence of a velocity threshold below which the proton stopping power vanishes, our calculations demonstrate that the antiprotons do not experience such a threshold.The combination of those two contrasting behaviors gives rise to an unprecedented negative Barkas effect: the stopping power of antiprotons is larger than that of protons.We identify that the slow antiproton at close encounter destabilizes a p orbital of the F- anion pointing towards the antiproton. This particular orbital becomes highly polarizable and hence contributes much to the stopping power.

International audience; In response to Werheit's Comment, we first discuss the following three points: (i) Attribution of the Raman bands with the help of calculations based on the density functional perturbation theory (DFPT), (ii) similarity of the Raman spectrum of boron carbide to the Raman spectrum of α rhombohedral boron, and (iii) dependence of the experimental Raman spectrum on the frequency of the excitation laser. Central to the point at stake, we also present a Raman spectrum computed for the first time on a 405-atom unit-cell of boron carbide, where the simulation cell accounts for the substitutional disorder of the polar carbon atom in the six atomic positions that are equivalent in the trigonal symmetry.We thus obtain a theoretical spectrum whose agreement with experimentis unprecedented in boron carbide, all of the features being present, with however negligible intensities for the doublet at 270/320 cm-1. We argue that the doublet intensity seen in experiments is not purely vibrational ---as computed at mechanical equilibrium--- and that FT-Raman spectroscopy taken at 1.06 μm lacks the accuracy that is necessary to study vibrational modes of boron carbides.

International audience; This paper develops a unified presentation of the thermodynamic parameters of eleven point-defect models in UO$_2$ (defect formation energies and entropies); four of them are fitted to experimental data, while the remaining seven are obtained through atomistic simulations. This allows us to compare all the models on the same basis both among themselves and with a large set of experimental data of various physical quantities. Combining the assessed defect thermodynamics and the TAF-ID functions for U and U$_4$O$_9$ phases makes it possible to compute the U - O phase diagram in the vicinity of stoichiometric UO$_2$. The defect formation energies and entropies are very different from one model to another. Concerning the ability to reproduce experimental data, the fitted models usually correctly reproduce the data sets according to which they were fitted. For one atomistic-based model, the measurements of the oxygen potential as a function of the temperature and the O/M ratio are reproduced in a satisfactory manner while the phase diagram is more approximate. No model, either fitted or atomistic-based, reproduces simultaneously the measured conductivity and the oxygen concentration as functions of theoxygen potential. The difficulties of the atomistic-based models in predicting the O/M ratio as a function of the oxygen potential are thought to partly arise from an erroneous calculation of the oxygen molecule energy derived from ab initio techniques and probably also from a poor evaluation of the electron-hole Gibbs energy of formation; more generally improving the technique for calculating reliable defect entropies of formation appears of great importance. The difficulty of obtaining reliable experimental data close to the stoichiometry might also contribute to the limited agreement between calculations and measurements, which is reason enough to reassess the behavior of the material in this stoichiometry region comprehensively, using all the possible characterization techniques on each material sample. In particular, since several independent studies predict an unexpected crucial role of the uranium vacancies on the evolution of O/M,experimental assessment of this feature should be sought.

The microstructural evolution of nuclear materials is driven by the agglomeration and recombination of the defects that are created under irradiation, such as vacancies and interstitial atoms. Predicting the ageing kinetics of these irradiation defects is essential to understand how the mechanical properties of the materials will evolve with time. Many physical models describing the ageing kinetics account for the elementary jump frequencies of point defects at the atomic scale and involve a master equation governing the time evolution of a state probability vector given an initial state. Transition state theory and elasticity theory are used to compute the jump frequencies and construct the transition rate matrix that is the crucial ingredient of the master equation. In this thesis, we develop non-stochastic numerical techniques to characterize the motion of individual defects migrating over long distances prior to recombining with another mobile defect or being absorbed by a sink, usually an immobile extended defect. Our approach is based on the theory of absorbing Markov chains in which the absorbing states correspond to the recombinations of two mobile defects or to the absorptions of a defect by an immobile sink. The defect motions are then entirely determined by their first-passage time distribution to distant absorbing locations, no-passage distribution, and the probability fluxes to the sinks. These quantities directly stem from the spectral properties of the transition rate matrix and define the probabilistic laws of non-local events that can then be simulated by a first-passage kinetic Monte Carlo algorithm. They also allow computing the sink strengths that are the crucial input parameters of rate-equation cluster dynamics simulations. Assuming that defects undergo reversible diffusion, we show that the absorbing transition rate matrix is diagonally similar to a symmetric definite positive matrix. This feature greatly facilitates the extraction of its spectral properties by using iterative sparse solvers. We demonstrate the efficiency of the approach with direct computations of elastodiffusion properties of a vacancy around a cavity in aluminum and Monte Carlo computations of cluster diffusivity in low alloyed manganese steels. Additionally, we also assess the efficiency of various mathematical schemes to characterize the evolution laws of a point defect near a sink. In particular, we develop a scheme combining Krylov subspace projection and eigenvalue deflation. For our model system describing the absorption of a vacancy by a cavity in aluminum, a small Krylov subspace deflated by the unique eigenmode corresponding to the quasi-stationary distribution makes it possible to capture the kinetics of the defect absorption faithfully. Finally, we discuss how a first-passage kinetic Monte Carlo algorithm performing non-local moves of small self-interstitial clusters can be used to compute sink strengths efficiently., L'évolution microstructurale des matériaux pour le nucléaire est pilotée par l'agglomération et la recombinaison des défauts créés sous irradiation, tels que les lacunes et les atomes interstitiels. Prédire les cinétiques des défauts d'irradiation est essentiel si l'on souhaite comprendre comment les propriétés mécaniques des matériaux vont évoluer dans le temps. De nombreux modèles physiques décrivant les cinétiques de vieillissement prennent en compte les fréquences des sauts élémentaires des défauts ponctuels à l'échelle atomique et font intervenir une équation maîtresse régissant l'évolution temporelle d'un vecteur de probabilité d'état, étant donné un état initial. La théorie des états de transition et la théorie de l'élasticité sont ici utilisées pour calculer les fréquences de saut et construire la matrice des taux de transition qui est l'ingrédient crucial de l'équation maîtresse. Dans cette thèse, nous développons des techniques numériques non-stochastiques pour caractériser le mouvement de défauts individuels migrant sur de longues distances avant de se recombiner avec un autre défaut mobile ou d'être absorbés par un puits, généralement un défaut étendu immobile. Notre approche est fondée sur la théorie des chaînes de Markov absorbantes dans laquelle les états absorbants correspondent à des recombinaisons de défauts ou des absorptions de défauts mobiles par des puits fixes. Les mouvements des défauts sont alors entièrement déterminés par la distribution des temps de premier passage vers des sites éloignés, la distribution de non-passage, et les flux de probabilité vers les puits. Ces quantités découlent directement des propriétés spectrales de la matrice des taux de transition et définissent les lois probabilistes des événements non locaux qui peuvent ensuite être simulés par un algorithme de Monte Carlo cinétique de premier passage. Elles permettent également de calculer les forces de puits qui sont les paramètres d'entrée cruciaux des équations de cinétique chimique considérées dans les simulations de dynamique d'amas. En supposant que les défauts migrent suivant un processus de diffusion réversible, nous montrons que la matrice des taux de transition absorbante est diagonalement similaire à une matrice symétrique définie positive. Cette particularité facilite grandement l'extraction de ses propriétés spectrales par des solveurs itératifs creux. Nous démontrons l'efficacité de l'approche en calculant directement les propriétés d'élastodiffusion d'une lacune autour d'une cavité dans l'aluminium et en mesurant la diffusivité de petits amas de solutés dans des aciers faiblement alliés en manganèse. En outre, nous évaluons également l'efficacité de divers schémas mathématiques pour caractériser les lois d'évolution d'un défaut ponctuel près d'un puits. En particulier, nous développons un schéma combinant projection sur des sous-espaces de Krylov et déflation de modes propres. Dans le cas du système modèle décrivant l'absorption d'une lacune par une cavité dans l'aluminium, un petit sous-espace de Krylov déflaté par le mode propre unique correspondant à la distribution quasi-stationnaire est capable de capturer fidèlement la cinétique d'absorption du défaut. Enfin, nous discutons de la manière dont un algorithme de Monte Carlo cinétique peut être mis en œuvre pour calculer efficacement les forces de puits de petits amas auto-interstitiels migrant rapidement le long d'une ligne de glissement et effectuant occasionnellement des rotations.

The object of this thesis is to study plasticity of chromium, considered as a coating layer of zirconium fuel claddings, and in these conditions needs to be able to accommodate the deformation which is imposed to it without cracking. Chromium is one of the seven pure body-centered cubic transition metals, however the only one with a magnetic structure close to antiferromagnetism at temperatures below ambient. At the atomic scale, plastic deformation of these metals at low temperature is mainly governed by the motion of screw dislocations with a ½ Burgers vector, which experience a strong friction with the crystal lattice. However, plasticity of chromium still raises numerous questions (types of dislocations, influence of magnetism), and has been the focus of a few studies aiming at characterizing its mechanical behavior. Comparing chromium to the other body-centered cubic metals, this work helped conclude on the similarities and differences it presents.Using ab initio calculations, properties of ½ screw dislocations were studied in the two non-magnetic and antiferromagnetic phases of chromium, in order to qualify the impact of magnetism on its plasticity at low temperature. This study revealed the only consequence of magnetism is the generation of magnetic faults when these dislocations shear the crystal, since their Burgers vector disrupts the magnetic order of chromium at low temperature. Thus, ½ dislocations are constrained to move while dragging a magnetic fault behind them. This fault was then studied at finite temperature using Monte Carlo simulations coupled with a Heisenberg model developed for this work, evidencing their disappearance at a temperature below the transition to magnetic disorder, close to ambient, letting dislocations free to move without dragging magnetic faults. Magnetism therefore has an impact on the plasticity of chromium at very low temperature only, where these faults exist.Accounting for the atomic properties of dislocations, as well as their thermally activated glide through nucleation of kink pairs using a line tension model, we then developed a plastic flow law predicting the yield stress as a function of the mechanical loading and the temperature, until the athermal temperature. Through a systematic study across all body-centered cubic metals, for which numerous experimental data are available, predictions of the model were both compared to experiments and between metals, in order to confirm the plastic behavior of chromium is comparable to the other metals. This systematic approach allowing at the same time for a better understanding of the plasticity of chromium and the other body-centered cubic metals was then extended to the study of a new mechanism explaining anomalous slip at low temperature observed experimentally, operating through a cooperative motion of multiple dislocations leading to their fast motion, reproduced by our atomistic simulations. We then looked at the properties of dislocations having a Burgers vector, with close energy compared to ½ and observed in all these metals, to explain under which conditions they can participate to the plastic deformation, our atomistic simulations showing a locking of the multiplication mechanisms of these dislocations., L'objet de cette thèse est d'étudier la plasticité du chrome, envisagé comme revêtement des gaines combustibles en zirconium, et qui dans ces conditions doit être capable de supporter la déformation qui lui est imposée sans se fracturer. Le chrome est l'un des sept métaux de transition purs à structure cubique centrée, étant cependant l'unique d'entre eux ayant une structure magnétique proche de l'antiferromagnétisme à température inférieure à l'ambiante. A l'échelle atomique, la déformation plastique de ces métaux à basse température s'opère principalement par le mouvement des dislocations vis de vecteur de Burgers ½, subissant une friction importante avec le réseau cristallin. Or, la plasticité du chrome soulève encore de nombreuses questions (types de dislocations, effet du magnétisme), et a été l'objet de peu d'étude visant à caractériser son comportement mécanique. En comparant le chrome aux autres métaux cubiques centrés, cette étude a permis de conclure sur les similitudes et différences que celui-ci présente.A l'aide de calculs ab initio, les propriétés des dislocations vis ½ ont été étudiées dans les deux phases non-magnétique et antiferromagnétique du chrome, de manière à caractériser l'influence du magnétisme sur sa plasticité à basse température. Cette étude a révélé que la seule conséquence du magnétisme est la génération de fautes magnétiques lorsque les dislocations cisaillent le cristal, puisque leur vecteur de Burgers ne respecte pas l'ordre magnétique du chrome à basse température. Ainsi, les dislocations ½ sont contraintes à se déplacer en traînant derrière elles une faute. Celle-ci a ensuite été étudiée à température finie par simulations Monte Carlo couplées à un modèle de Heisenberg développé pour cette étude, mettant en évidence leur disparition à une température inférieure à la transition vers le désordre magnétique, proche de l'ambiante, laissant les dislocations libres de se déplacer sans traîner de faute. Le magnétisme n'a donc un impact sur la plasticité du chrome qu'à très basse température, où ces fautes existent.Tenant compte des propriétés atomiques des dislocations, ainsi que de leur mouvement thermiquement activé par germination de doubles décrochements à l'aide d'un modèle de tension de ligne, nous avons ensuite développé une loi d'écoulement plastique donnant la limite d'élasticité en fonction du chargement mécanique et de la température, jusqu'à la température athermale. Par le biais d'une étude systématique sur l'ensemble des métaux cubiques centrés, pour lesquels une grande variété de données expérimentales est disponible, les prédictions du modèle ont pu à la fois être comparées à l'expérience ainsi qu'entre métaux, de manière à confirmer que le comportement plastique du chrome est semblable à l'ensemble de ces métaux. Cette approche systématique permettant à la fois de mieux comprendre la plasticité du chrome et des autres métaux cubiques centrés s'est ensuite étendue à l'étude d'un nouveau mécanisme expliquant le glissement anomal à basse température observé expérimentalement, s'opérant par un mouvement coopératif de plusieurs dislocations entraînant leur déplacement rapide, reproduit par nos simulations atomiques. Nous nous sommes enfin intéressés aux propriétés de dislocations ayant un vecteur de Burgers , d'énergie proche des dislocations ½ et observées dans l'ensemble de ces métaux, afin d'expliquer dans quelles conditions celles-ci pourraient participer à la déformation plastique, nos simulations montrant un blocage des mécanismes de multiplication de ces dislocations.

International audience; The radiation damage induced in cerium dioxide CeO$_2$ and cerium-gadolinium mixed oxides (Ce, Gd)O$_{2-x}$ was studied by micro-Raman spectroscopy and diffuse UV-visible reflectance spectroscopy. Sintered samples of undoped CeO$_2$ and (Ce, Gd)O$_{2-x}$ for 1 mol% and 5 mol% of Gd$_2$O$_3$ were irradiated at room temperature with 2.5-MeV Ar$^{2+}$ and 12-MeV Ar$^{4+}$ ions up to high fluences (i.e. 5x10$^{14}$ cm$^{-2}$ and 2.1x10$^{14}$ cm$^{-2}$ respectively). The Raman spectra show that no amorphization of those oxides occurs whatsoever, regardless of the Gd content. However, the damage cross section deduced from the decay of F$_{2g}$ peak intensity versus fluence seems to decrease with the Gd$^{3+}$ content. Moreover, the reflectivity spectra show a clear decrease of the band-gap energy and increase of the Urbach energy arising from point defect accumulation. Similar effects are found for 1 mol% and 5 mol% Gd$_2$O$_3$, whereas a larger effect on band tailing is seen for undoped ceria. The respective roles of nuclear collisions and electronic excitations are discussed for both ion energies knowing that the two kinds of measurements do not probe the same sample depth. The substitution of Gd$^{3+}$ ions for Ce$^{4+}$ ions is likely reducing the damage and the formation of band tails related to the formation of Ce$^{3+}$ ions by electronic excitation processes.

International audience; Polyethylene is a widely used polymer and a prototypical polyolefin. Understanding degradation mechanisms at the atomic scale leading to oxidation under the effect of temperature and/or irradiation is crucial for many long term applications. Here we focus on a subset of reactions belonging to the generally assumed reaction scheme and involving alkoxy radicals. We follow a theoretical approach based on density functional theory (DFT), in order to assess the role of these species, which are short lived and thus hard to detect experimentally. For every considered reaction we calculate the reaction enthalpy as well as the energy barrier and we evaluate the influence of the local atomic environment, taking advantage of a model surface of a crystalline lamella which mimicks the interface between crystalline and amorphous zones. Our results suggest that, in certain conditions, the kinetic pathway can bypass the formation of hydroperoxides. The concentrations of alkoxy and peroxy radicals during radio-oxidation, as well as their ratio, are important parameters controlling the predominance of chain-scission or crosslinking of the polymer. The data presented here constitute part of a database that can be used to set up kinetic simulations based on homogeneous chemical kinetics or Monte Carlo algorithms.

International audience; We investigate the influence of elastic properties of point defects on dislocation climb under stress and irradiation. For this purpose, elastic dipole tensors and diaelastic polarizabilities are evaluated in aluminum for vacancies and self-interstitial atoms in their stable and saddle configurations, using density functional theory calculations. These parameters are introduced in an object kinetic Monte-Carlo code and a continuous diffusion model to estimate the stress dependence of dislocation climb, using a dipole of straight dislocations. We show that both parameters have an influence on absorption of point defects under stress, in agreement with previous analytical models. However, the effect of dipole tensor is found only 5 times larger than polarizability, whereas models predict a factor up to 30. In addition, including polarizability reverses the stress angular dependence when a uniaxial stress is applied orthogonal to the dislocation line, so in general polarizability cannot be ignored for simulations under applied stress. Further comparison with analytical models shows that they give a good description of angular dependence, provided saddle point configuration of point defects is not too anisotropic. For vacancies, which are strongly anisotropic in their saddle configuration, models fail to reproduce quantitatively lattice effects on stress angular dependence observed in simulations. Calculations show that dislocation climb velocity under irradiation is expected to be the highest if the stress is approximately orthogonal to the dislocation line, especially along the Burgers vector, and the lowest if the stress is applied close to the direction with the largest projection on the dislocation line.

Polyethylene (PE) is one of the most abundantly manufactured polymers for various applications. Among the various them, PE has been widely used as insulation of electrical cables in nuclear power plants (NPPs). However, because cable insulators consisting of polymers are most vulnerable in cable structures under high temperature or irradiation conditions (e.g., γ-rays), understanding the kinetic mechanisms involved in the polymer degradation is essential to ensure their long operation lifetime. This thesis study, based on density functional theory (DFT), aims to elucidate the degradation mechanisms of polyethylene mainly resulting from oxidation under irradiation, the so-called radio-oxidation process, that crucially triggers the aging of cables. Besides, antioxidants are commonly used as additives in order to slow down the aging process. The role of antioxidants and their action mechanism are carefully studied. The insight that is expected might find application in various fields, from electrical cable insulation in nuclear power stations to plastics used in everyday life., Le polyéthylène (PE) est l'un des polymères les plus abondamment fabriqués pour diverses applications. Parmi ces applications, le PE a été largement utilisé comme isolant de câbles électriques dans les centrales nucléaires (NPP). Cependant, étant donné que les isolants en polymères constituent le maillot le plus vulnérables de la globalité du câble sous des conditions de haute température ou d'irradiation (par exemple, les rayons γ), la compréhension des mécanismes cinétiques de dégradation est essentielle pour assurer une durée de vie la plus longue possible de ce composant en conditions opérationnelles. L'étude menée dans cette thèse, basée sur la théorie fonctionnelle de la densité (DFT), vise à élucider les mécanismes de dégradation du polyéthylène résultant principalement de l'oxydation sous irradiation, ou radio-oxydation, qui joue un rôle crucial dans le déclenchement du vieillissement des câbles. Par ailleurs, les antioxydants sont couramment utilisés comme additifs afin de ralentir le processus de vieillissement. Le rôle des antioxydants et leur mécanisme d'action sont soigneusement étudiés. Les résultats attendus pourraient trouver une application dans divers domaines, de l'isolation des câbles électriques dans les centrales nucléaires aux plastiques utilisés dans la vie quotidienne.

International audience; Excess point defects created by irradiation in metallic alloys diffuse and annihilate at sinks available in the microstructure, such as grain boundaries, dislocations or point defect clusters. Fluxes of defects create fluxes of alloying elements, leading to local changes of composition near the sinks and to a modification of the properties of the materials. The direction and the amplitude of this radiation induced segregation, its tendency to produce an enrichment or a depletion of solute, depends on a set of transport coefficients which are very difficult to measure experimentally. The understanding of radiation induced segregation phenomena has however made significant progress in recent years, thanks to the modeling at different scales of the diffusion and segregation mechanisms. We review here these different advances, and try to identify the key scientific issues that limit the development of predictive models, applicable to real alloys. The review addresses three main issues: the calculation of the transport coefficients from \textit{ab initio} calculations, the modeling of segregation kinetics at static point defects sinks---mainly by kinetic Monte Carlo or diffusion-reaction models---and the more challenging task of modeling the dynamic interplay between radiation-induced segregation and the sink microstructure evolution, especially when this evolution results from the annihilation of point defects. From this overview of the current state-of-the-art in this field, we discuss still-open questions and guidelines for what constitutes, in our opinion, the desirable future works on this topic.

International audience; In this work, the Cu clustering in Fe under irradiation is investigated using experiments, cluster dynamics and atomistic kinetic Monte Carlo (AKMC) simulations. In experiments, cast iron and model FeCu alloy samples were irradiated with 2 MeV electrons for 143 h at 140 °C. The post-irradiation microstructure was characterized using atom probe tomography. Cluster dynamics and AKMC methods were used to simulate the Cu clustering under the same irradiation conditions. Both simulation methods show satisfactory agreement with experiments, lending strength to the validity of the models. Finally, the Cu clustering in spent-fuel repository conditions for 10$^5$ years at 100 °C was simulated using both methods. The results indicate that potential hardening by Cu clustering is insignificant over 10$^5$ years.

International audience; In nuclear fission reactors, the amount of helium produced in materials by transmutation reactions at the end of the lifetime may reach several thousands of atomic parts per million (appm). Such high levels of helium production can impact the evolution of microstructures, particularly by forming helium bubbles. To better understand the role of helium on the stability of bubbles, a "variable-gap model" was parametrized with molecular dynamics (MD) calculations performed in nickel. This model predicts binding energies in a good agreement with MD values, especially for large bubbles. For very small bubbles, the influence of magic number sizes and faceting is more complex than can be described with the model. For these cases, it is proposed to use MD values directly.

International audience; The interplay of screw dislocations with carbon atoms is investigated in tungsten at high temperature using in situ straining experiments in a transmission electron microscope (TEM) and through ab initio calculations. When the temperature is high enough to activate carbon diffusion, above 1373 K, carbon segregates in the core of screw dislocations and modifies their mobility, even for a carbon concentration as low as 1 appm. TEM observations reveal the reappearance of a Peierls mechanism at these high temperatures, with screw dislocations gliding viscously through nucleation and propagation of kink-pairs. The mobility of screw dislocations saturated with carbon atoms is then investigated with ab initio calculations to determine kink-pair formation, nucleation and migration energies. These energies are used in kinetic Monte-Carlo simulations and in an analytical model to obtain the velocity of screw dislocations as a function of the temperature, the applied stress and the dislocation length. The obtained mobility law parametrised on ab initio calculations compares well with experiments.

International audience; Plasticity in hexagonal close-packed zirconium is controlled by screw dislocations which easily glide in the prismatic planes where they are dissociated. At high enough temperatures, these dislocations can deviate out of the prism planes to also glide in the first order pyramidal and basal planes. To get a better understanding of these secondary slip systems, we have performed molecular dynamics (MD) simulations of a screw dislocation gliding in a basal plane. The gliding dislocation remains dissociated in the prism plane where it performs a random motion and occasionally cross-slips out of its habit plane by the nucleation and propagation of a kink-pair. Deviation planes are always pyramidal, with an equal probability to cross-slip in the two pyramidal planes on both sides of the basal plane, thus leading to basal slip on average. Basal slip appears therefore as a combination of prismatic and pyramidal slip in the high stress regime explored in MD simulations. This is confirmed by nudged elastic band (NEB) calculations. But NEB calculations also reveal a change of glide mechanism for a decreasing applied stress. At low stress, kinks do not lie anymore in the pyramidal planes. They are now spread in the basal planes, thus fully compatible with a motion of the screw dislocation confined to the basal plane as seen in experiments. Basal slip, which is in competition with pyramidal slip, appears therefore favoured at low stress in pure zirconium.

International audience; Raman scattering is applied to probe the radiation damage in swift heavy ion-irradiated ceramics, namely zirconium nitride (ZrN), ceria (CeO$_2$), and yttria-stabilized zirconia (ZrO$_2$: Y, or YSZ) for about the same high electronic stopping powers of heavy ions. Raman spectra show that those ceramics are radiation-resistant materials which are not amorphized by such irradiations even for large track overlap at high fluences. However, the increase of the TA/LA and TO/LO band intensities versus fluence is evidenced after 100-MeV Xe ion irradiation of ZrN up to a saturation for the fluence of 3x10$^{12}$ cm$^{-2}$. The band growths are ascribed to the increase of the concentration of Zr and N vacancies induced by electronic excitations inside tracks. For ceria, the decrease and broadening of the main F$_{2g}$ peak of the fluorite-like structure and the growth of a broad defect band is observed versus fluence up to 10$^{14}$ cm$^{-2}$ for 200-MeV Xe ion irradiation. For YSZ, Raman spectra mainly give evidence of the intrinsic lattice disorder arising from the native oxygen vacancies, even up to high fluences (3x10$^{13}$ cm$^{-2}$) for 200-MeV I and 200-MeV Au ion irradiation. Results are discussed on the basis of the interplay between the native structural disorder and the radiation-induced disorder by electronic excitation in those three materials.

International audience; Ferritic Martensitic and ODS steels are widely studied for applications in the next generation of nuclear reactor. Their good mechanical properties and corrosion resistance make them good candidates. Under irradiation flux, migration of point defects leads to different mechanisms including radiation-induced segregation (RIS). This type of segregation affects defect sinks such as grain boundaries (GB). Many parameters influence the amplitude of the RIS such as temperature, dose, chemical composition of the alloy… This paper presents the quantification of GB plane variation on RIS. Two different Σ3 (60° ) GB were investigate using EBSD and TKD for structural characterization and APT and STEM/EDS for chemical quantification. This work revealed W-shape profiles of Cr across GBs and a heterogeneous precipitation of Cr-C rich particles in GBs planes after irradiation. Our work shows that not only the amplitude but also the sign of the Gibbsian Excess may be affected by the GB plane.

International audience; Most of the phase transformations modifying the microstructure, thereby the materials properties, are controlled by the diffusion of atoms. The rate but also the selection of phase transformations, depend on the concentration of lattice point defects (PDs), because substitutional atoms exchange with PDs to diffuse. PDs are non-conservative species. During manufacturing or in use, whenever PD diffusion and creation/annihilation reactions at extended defects in the microstructure, are slower than the kinetics of the microstructure, these PDs may not have their equilibrium concentration. A departure of PDs from local equilibrium can be transient under thermal conditions, or permanent in materials driven out of equilibrium as under irradiation.Non-equilibrium PDs can have a dramatic effect on the evolution of the microstructure or even on the stationary microstructure in driven system. We present an atomic kinetic Monte Carlo (AKMC) method, that is able to tackle the atomic-scale couplings between PD diffusion, annihilation/creation reactions and the kinetics of decomposition of a solid solution into a two-phase microstructure. By introducing PD source-and-sinks (SAS) at specific lattice sites, we control the PD reactions and highlight the role of non-equilibrium quenched-in point defects on the evolution kinetics of short range order parameters and subsequent second phase precipitation. Then, we open the discussion on various kinetic phenomena that require taking into account the role of non-equilibrium PDs at different scales of time and space.

International audience; The interaction of hydrogen with 2/1 screw dislocations is investigated in body-centered cubic tungsten and iron using ab initio calculations. Different core reconstructions are evidenced, depending on the number of hydrogen atoms introduced inside the dislocation core. Corresponding interaction energies are highly attractive in both metals, with a significant contribution of zero point energy associated with H vibrations, particularly in Fe. The pinning by hydrogen of the dislocation in its reconstructed core remains efficient for a local atomic fraction of hydrogen as low as 1/5. Other octahedral interstitial sites close to the reconstructed core are also attractive, contrary to the perfect bcc crystal where these sites are unstable and where hydrogen lies in the tetrahedral sites. Hydrogen recovers its bulk behavior only beyond the eighteenth neighbor octahedral sites. A simple pair interaction model is parameterized on ab initio calculations and used with a mean-field approximation to predict the concentration profiles of hydrogen segregated in and around the dislocation core. This segregation model predicts that dislocation core-sites remain completely decorated by hydrogen atoms up to at least 750 K and 200 K in W and Fe, respectively.

Low alloys steels are mainly made of iron and carbon, with additional alloying elements in low quantities. In these materials, secondary phases enriched in carbon -namely carbides- can appear. The formation of carbides can drastically change mechanical properties of steels. It may for example promote the appearance of cracks. In particular, the presence of carbides in ferritic steels used in the nuclear industry can raise some safety concerns. Cementite (Fe₃C) and the Mo₂C molybdenum carbide are two very common carbides in steels. The goal of this thesis is to investigate properties of these two carbides via modelling and computer simulations. Various methods are useful to investigate atomic-scale properties that are difficult to access experimentally. This way, we studied carbon diffusion in cementite and compared with indirect experimental data. We also quantified effects of alloying elements on cementite and Mo₂C properties. Comparisons are made with experimental data available at this time. In addition, we developed a modelling approach in order to investigate the precipitation of cementite.; Les aciers faiblement alliés sont principalement constitués de fer et de carbone, avec quelques éléments d’alliage supplémentaires en faible quantité. Dans ces matériaux, des phases secondaires enrichies en carbone -des carbures peuvent apparaître. La formation de carbures peut modifier radicalement les propriétés mécaniques des aciers. Elle peut par exemple favoriser l’apparition de fissures. En particulier, la présence de certains carbures dans les aciers ferritiques utilisés dans l’industrie nucléaire pourrait conduire à leur fragilisation. La cémentite (Fe₃C) et le carbure de molybdène Mo₂C sont deux carbures fréquents dans les aciers. L’objectif de cette thèse est d’étudier les propriétés de ces deux carbures en utilisant la modélisation et des simulations numériques. Ces méthodes permettent d’aborder des données inaccessibles expérimentalement. Ainsi, nous avons étudié la diffusion du carbone dans la cémentite et comparé nos résultats avec des données expérimentales indirectes. Nous avons également quantifié les effets des éléments d’alliage sur les propriétés de la cémentite et du Mo₂C. De plus, nous avons développé une stratégie de modélisation afin d’étudier la précipitation de la cémentite.

This thesis is a theoretical study on the temperature-dependent thermodynamic, point-defect formation and diffusion properties in fcc Fe-Ni alloys over the whole range of composition, with a focus on the magnetochemical effects. The results are derived from density functional theory (DFT) calculations and Monte Carlo simulations using a DFT-parametrized effective interaction model (EIM) with explicit atomic and spin variables. The first part of this work is focused on thermodynamics. We compute via DFT energetic, magnetic and vibrational properties and the bcc-fcc phase diagram below the Curie points, revealing the relative importance between magnetic and vibrational entropies. Combining Monte Carlo simulations with the EIM, we obtain an fcc phase diagram across the Curie points and demonstrate a close interplay between magnetic and chemical orders. Finally, we also discuss the effects of Mn and Cr additions on phase stability. The second part of the work is dedicated to point-defect properties. We develop Monte Carlo schemes to compute vacancy formation free energy in alloys. We show that vacancy formation in fcc Fe and Ni exhibits features that are well distinct from those in bcc Fe, pointing out the relevance of longitudinal spin fluctuations. The results in fcc Fe-Ni alloys reveal that magnetic disorder tends to increase vacancy formation free energy, while chemical disorder shows an opposite effect. Finally, we study via DFT the magnetic effects on the properties of self-interstitial atoms in fcc Fe and Ni. The final part of the work is devoted to vacancy-mediated diffusion. We perform kinetic Monte Carlo simulations to obtain temperature-dependent diffusion properties over the whole concentration range. The results allow to probe into the effects of magnetic and chemical transitions on diffusion properties and provide details inaccessible from experiments. This work fully takes into account the impacts of transversal and longitudinal spin fluctuations and the interplay between magnetism and chemical order. It provides an accurate and consistent prediction of thermodynamic, defect formation and diffusion properties in the Fe-Ni system, and contributes to a better understanding of effects of magnetism in austenitic steels. The applied approach is also transferable to the investigation of other magnetic alloys., Cette thèse porte sur les propriétés thermodynamiques, de formation des défauts ponctuels et de diffusion dans les alliages Fe-Ni cfc sur toute la gamme de composition et en fonction de la température, avec un intérêt particulier pour les effets magnétochimique. Les résultats sont obtenus à partir de calculs de la théorie fonctionnelle de la densité (DFT) et de simulations de Monte Carlo atomique incluant un modèle d'interaction effective (EIM) ajusté sur les calculs DFT. Ce modèle traite de façon explicite les interactions chimiques et magnétiques, et leur couplage. La première partie de ce travail est centrée sur la thermodynamique. Nous calculons via la méthode DFT, les propriétés énergétiques, magnétiques et vibrationnelles, ainsi que le diagramme de phase cc-cfc en-dessous de la température de Curie. Nous mettons en évidence les contributions relatives des effets magnétiques et vibrationnels. A l'aide des simulations Monte Carlo, nous obtenons le diagramme de phase cfc en-dessous et au-dessus des températures de Curie. Nous mettons en évidence un fort couplage entre les ordres magnétiques et chimiques. Enfin, nous étudions les effets d'un ajout de Mn ou de Cr sur le magnétisme et la stabilité des phases de l'alliage Fe-Ni. La deuxième partie du travail est consacrée à l'étude des propriétés des défauts ponctuels. Nous développons des algorithmes Monte Carlo pour calculer l'énergie libre de formation de la lacune dans les alliages. Nous montrons que la formation de lacune dans Fe et Ni cfc, présente des caractéristiques bien distinctes de celles dans Fe cc, soulignant ainsi l'importance des fluctuations de spin longitudinales. Nous montrons que dans les alliages, le désordre magnétique a tendance à augmenter l'énergie libre de formation de lacunes, tandis que le désordre chimique a un effet opposé. Enfin, nous étudions par la méthode DFT, les effets magnétiques sur les propriétés des atomes auto-interstitiels dans Fe et Ni cfc. La dernière partie du travail est consacrée à la diffusion atomique. Nous calculons les propriétés de diffusion de traceur et des défauts ponctuels en fonction de la température, sur la totalité de l'intervalle en composition de la phase austénitique. Nous mettons en évidence les effets des transitions magnétiques et chimiques sur les propriétés de diffusion et nous fournissons des données inaccessibles aux expériences. Ce travail prend pleinement en compte l'impact des fluctuations de spin transversales et longitudinales, ainsi que le couplage entre le magnétisme et l'ordre chimique. Il fournit une modélisation précise et cohérente des propriétés thermodynamiques, de formation des défauts et de diffusion du système Fe-Ni. Il contribue ainsi à une meilleure compréhension des effets du magnétisme dans les aciers austénitiques.

Atomic scale simulations at finite temperature are an ideal approach to study the thermodynamic properties of magnetic transition metals. However, the development of interatomic potentials explicitly taking into account magnetic variables is a delicate task. In this context, we present a tight-binding model for magnetic transition metals in the Stoner approximation. This potential is integrated into a Monte Carlo structural relaxations code where trials of atomic displacements as well as fluctuations of local magnetic moments are performed to determine the thermodynamic equilibrium state of the considered systems. As an example, the Curie temperature of cobalt is investigated while showing the important role of atomic relaxations. Furthermore, our model is generalized to other transition metals highlighting a local magnetic moment distribution that varies with the gradual filling of the d states. Consequently, the successful validation of the potential for different magnetic configurations indicates its great transferability makes it a good choice for atomistic simulations sampling a large configuration space.

International audience; We develop a new family of electronic structuremethods for capturing at the same time thedynamic and non-dynamic correlation effects.We combine natural orbital functional theory(NOFT) and many-body perturbation theory(MBPT) through a canonicalization procedureapplied to the natural orbitals to gain access toany MBPT approximation. We study three dif-ferent scenarios: Corrections based on second-order Møller-Plesset (MP2), Random-PhaseApproximation (RPA), and coupled-cluster sin-gles doubles (CCSD). Several chemical prob-lems involving different types of electron cor-relation in singlet and multiplet spin stateshave been considered. Our numerical tests re-veal that RPA-based and CCSD-based correc-tions provide similar relative errors in molecu-lar dissociation energies (De) to the results ob-tained using a MP2 correction. With respectto the MP2 case, the CCSD-based correctionimproves the prediction, while the RPA-basedcorrection reduces the computational cost.

International audience; Additive manufacturing (AM) is a promising technology for the design of materials with complex geometries with reduced cost and material waste. In order to be used in the nuclear industry, the capability of AM materials, in term of radiation resistance must be compared with materials elaborated in conventional ways. In this work, the radiation resistance of 316L austenitic stainless steels (ASSs) elaborated by AM is compared to a solution-annealed 316L ASS after irradiation with 5 MeV Fe$^{5+}$ for 3 dpa at 550 °C (873 K). After irradiation, cavities are mainly located near grain boundaries for all studied alloys. Intra-granular cavities are only found in the AM material after heat treatment and are likely to be remaining porosity already present before irradiation. No cavities in intra-granular position are found in the conventional 316L ASS or in the AM material after hot isostatic pressing (HIP) at 1100 °C. It suggests that the void swelling ASSs starts by the formation of cavities at grain boundaries followed by a formation of cavities in intra-granular position, conventionally studied. Loops in the AM material with a hot treatment at 700°C are heterogeneously distributed due to a bimodal distribution of grains in terms of intra-granular misorientation. The intra-granular misorientation drastically reduces the loop density after irradiation. Frank and perfect loops are found to be only interstitial-type loops. The larger cavities' size and the more advanced dislocations network in the AM HIP sample suggests a slightly reduced radiation swelling resistance for the AM material but further investigations at higher irradiation dose have to be done.