Les batteries lithium-ion sont largement utilisées comme source d'énergie pour les appareils électroniques portables. L'oxyde de fer (principalement α-Fe2O3), l'un des oxydes de métal de transition les plus important, a suscité l’intérêt scientifique depuis qu'il a été reporté comme matériau d'anode pour les batteries lithium-ion en raison de sa capacité théorique élevée (1007 mAh g-1), de son respect de l'environnement, de son abondance et de son faible coût. Dans cette thèse, une électrode modèle en couche mince d'oxyde de fer a été préparée par simple oxydation thermique à 300 °C dans l'air d'un substrat de fer métallique pur, utilisé aussi comme collecteur de courant. Une variété de techniques d'analyse, électrochimiques (CV, EIS et cyclage galvanostatique), spectroscopiques (XPS, ToF-SIMS) et microscopiques (MEB et AFM), ont été mises en oeuvre pour étudier les mécanismes réactionnels et la chimie de surface de l'oxyde de fer à différents stades de lithiation/délithiation et cyclage. / Lithium-ion batteries (LIBs) are widely used as power sources for portable electronic devices. Iron oxide (mainly α-Fe2O3), as one of the most important transition metal oxide, has attracted attention due to its high theoretical capacity (1007 mAh g-1), environmental friendliness, abundance and low cost since reported as anode material for LIBs. In this thesis, an iron oxide thin film model electrode was prepared by simple thermal oxidation of pure metallic iron substrate at 300 oC in air, also used as a current collector. Electrochemical methods (CV, EIS and galvanostatic cycling) were combined with surface (XPS, ToF-SIMS) and microscopic (SEM, AFM) analytical techniques to investigate the reaction mechanisms and the surface chemistry of the iron oxide thin film at different stages of lithiation/delithiation and upon cycling.

Security is a critical part of every software developed today and it will be even more important going forward when more devices are getting connected to the internet. By striving to improve the quality of the code, in particular the security aspects, there might be a reduction in the number of vulnerabilities and improvements of the software developed. By looking at issues from past problems and studying the code in question to see whether it follows the SEI CERT secure coding standards, it is possible to tell if compliance to this standard would be helpful to reduce future problems. In this thesis an analysis of vulnerabilities, written in C and C++, reported in Common Vulnerabilities and Exposures (CVE), will be done to verify whether applying the SEI CERT secure coding standard will help reduce vulnerabilities. This study also evaluates the SEI CERT rule coverage of three different static analysis tools, Rosecheckers, PVS-Studio and CodeChecker by executing them on these vulnerabilities. By using three different metrics, true positive, false negative and the run time. The results of the study are promising since it shows that compliance to the SEI CERT standard does indeed reduce vulnerabilities. Of the analyzed vulnerabilities it was found that about 60% of these could have been avoided, if the standard had been followed. The results of the tools were of great interest as well, it showed that the tools did not perform as well as the manual analysis, however, all of them found some SEI CERT rule violations in different areas. Conclusively, a combination of manual analysis and these three static analysis tools would have resulted in the highest number of vulnerabilities avoided.

When an anode electrode potential is larger than the lowest unoccupied molecular orbital (LUMO) of the electrolyte, Li-ions and electrolyte molecules will participate in reduction reactions on the anode surface and form a solid electrolyte interface (SEI) layer. Active Li-ion consumption in the formation reactions is the main source of capacity loss (>50) and ageing in Li-ion batteries (LIBs). Due to the fast-occurring and complex nature of the electrochemical processes, conventional experimental techniques are not a feasible approach for capturing and characterizing the SEI formation phenomenon. The lack of experimental data and consequently the absence of potential parameters for crystal structures in this layer makes molecular dynamics~(MD) simulations inapplicable to it. Also, due to the multi-component multi-layer structure of the SEI, the smallest system representing an SEI layer is too large for employing the principles of quantum mechanics~(QM), that traditionally work with much smaller system sizes. Addressing this, this thesis presents a novel computational framework for coupling QM and MD calculations to simulate a system with the size limits of MD simulations independent of the experimental data. The QM evaluates sub-atomic properties such as energy barriers against diffusion and employs seven new algorithms to estimate potential parameters as the input of the MD simulations. Then MD simulations forecast SEI's properties including density, Young's Modules, Poisson's Ratio, thermal conductivity, and diffusion coefficient mechanisms. The output of the QM and MD calculations are employed to develop two macro-scale mathematical models for predicting battery ageing and battery performance, incorporating the impact of the SEI layer in addition to the cathode, anode, and separator parts. Finally, the results obtained have been validated with respect to the experimental data in different operational conditions. / Thesis / Doctor of Philosophy (PhD) / The limited lifespan of expensive batteries is the main obstacle to electrification of the transport sector, despite its necessity for addressing the current environmental issues. Li+/electrolyte reduction on the electrode surface is responsible for more than 50% of capacity loss and the consequent ageing is a complex and fast-occurring phenomenon (few ns) that cannot be easily resolved using conventional experimental and computational techniques. This thesis presents the development of some computational frameworks and demonstrates their employment to investigate this phenomenon from a multi-scale perspective, i.e., from a few electrons to an entire battery length scale, with the operating cycles ranging from a few ps to several months, employing Quantum Mechanics, Molecular Dynamics, and Macro-Scale Modeling. The frameworks have been successfully validated with respect to experimental data from the literature and have been applied successfully to highlight the parameters that impact ageing in batteries. The findings presented in this thesis can be used as the base for further research on next-gen durable batteries with liquid and solid-state electrolytes.

Radio Frequency Machine Learning (RFML) is the application of ML techniques to solve problems in the RF domain as an alternative to traditional digital-signal processing (DSP) techniques. Notable among these are the tasks of specific emitter identification (SEI), determining source identity of a received RF signal, and automated modulation classification (AMC), determining the modulation scheme of a received RF transmission. Both tasks have a number of algorithms that are effective on simulated data, but struggle to generalize to data collected in the real-world, partially due to the lack of available datasets upon which to train models and understand their limitations. This thesis covers the practical considerations for systems that can create high-quality datasets for RFML tasks, how variances from real-world effects in these datasets affect RFML algorithm performance, and how well models developed from these datasets are able to generalize and adapt across different receiver hardware platforms. Moreover, this thesis presents a proof-of-concept system for large-scale and efficient data generation, proven through the design and implementation of a custom platform capable of coordinating transmissions from nearly a hundred Software-Defined Radios (SDRs). This platform was used to rapidly perform experiments in both RFML performance sensitivity analysis and successful transfer between SDRs of trained models for both SEI and AMC algorithms. / Master of Science / Radio Frequency Machine Learning (RFML) is the application of machine learning techniques to solve problems having to do with radio signals as an alternative to traditional signal processing techniques. Notable among these are the tasks of specific emitter identification (SEI), determining source identity of a received signal, and automated modulation classification (AMC), determining the data encoding format of a received RF transmission. Both tasks have practical limitations related to the real-world collection of RF training data. This thesis presents a proof-of-concept for large-scale, efficient data generation and management, as proven through the design and construction of a custom platform capable of coordinating transmissions from nearly a hundred radios. This platform was used to rapidly perform experiments in both RFML performance sensitivity analysis and successful cross-radio transfer of trained behaviors.

La miniaturisation des appareils électroniques et la multiplication de leurs fonctionnalités conduisent à développer des microsources d’énergie adaptées, parmi lesquelles figurent les microbatteries au lithium. Malgré leurs excellentes performances, ces systèmes de stockage électrochimique tout solide restent toutefois limités en termes de capacité surfacique. Cette caractéristique étant intrinsèquement liée aux matériaux d’électrodes, nous avons choisi de nous intéresser à des couches minces de CuO, dont la capacité volumique théorique (426 µAh .cm-2.µm-1) est sensiblement plus élevée que celle des matériaux d’intercalation utilisés jusqu’à présent. Ce matériau réagit avec le lithium selon un mécanisme particulier, dit de conversion, qui induit la formation d’un système multiphasé et nanostructuré d’une grande complexité. Dans le cadre de ce travail, la compréhension des mécanismes électrochimiques et chimiques mis en jeu au cours du cyclage de couches minces d’oxyde de cuivre (CuO) a été l’objectif majeur. Celui-ci a nécessité une caractérisation fine du matériau actif d’électrode et des interfaces générées (interfaces solide/solide et interface solide/électrolyte). Ces études ont été principalement menées à partir de la Spectroscopie Photoélectronique à Rayonnement X (XPS), de la Microscopie à Force Atomique (AFM) et d’une modélisation théorique exploitant les méthodes de la chimie quantique. Les propriétés chimiques et morphologiques des couches minces de CuO cyclées ont été corrélées à leur comportement électrochimique. Une forte influence de leur structure et de leur morphologie initiales a pu être ainsi mise en évidence / The miniaturization of electronic components and the increasing number of their functionalities lead to the development of suitable energy microsources, among which lithium microbatteries appear. Despite the excellent performances of these all-solid-state electrochemical power sources, one main limitation that remains is their surface capacity. Its value being intrinsically connected to the nature of electrode materials, we chose to focus on CuO thin films which are characterized by a theoretical volumetric capacity (426 µAh .cm-2.µm-1) in far larger than the one of conventional intercalation materials used today. Indeed, this material reacts with lithium according to a particular mechanism, referred as conversion reaction, inducing the formation of a multiphase nanostructured system with a high complexity. In the framework of this study, understanding of electrochemical and chemical mechanisms which take place during the cycling of copper oxide thin films (CuO) was the main objective. This one has required a fine characterization of the electrode active material and the generated interfaces (solid/solid interfaces and solid/electrolyte interface). These studies have been mainly carried out with X-ray Photoelectron Spectroscopy (XPS), Atomic Force Microscopy (AFM) and theoretical approaches based on quantum chemistry methods. The chemical and morphological properties of the cycled CuO thin films have been linked to their electrochemical behavior. An important influence of their initial structure and morphology was then evidenced.

Ce projet de thèse porte sur la recherche de nouveaux électrolytes et additifs dans le but d’améliorer la cyclabilité d’une électrode négative composite de formule Si0.32Ni0.14Sn0.17Al0.04C0.35 et d’obtenir une interface électrode|électrolyte stable. En effet, comme la plupart des matériaux à base de silicium, ce composite de grande capacité (plus de 600 mA.h.g-1) souffre actuellement d’une faible durée de vie provenant essentiellement des expansions volumiques qu’il subit lors de sa lithiation et de sa SEI défaillante. Deux types d'électrolytes ont été évalués : (i) un mélange de carbonates d’alkyles EC/PC/3DMC auquel a été ajouté un sel de lithium (LiPF6, LiTFSI, LiFSI ou LiDFOB) ainsi que des additifs aidant à la formation de la SEI tels que le carbonate de vinylène (VC) ou le carbonate de fluoroéthylène (FEC), (ii) des liquides ioniques (LI) contenant un cation ammonium quaternaire (N1114+), imidazolium (EMI+) ou pyrrolidinium (PYR+), associé à un anion à charge délocalisée comme le bis(trifluorométhanesulfonyl)amidure (TFSI-) ou le bis(fluorosulfonyl)amidure (FSI-). L’analyse du diagramme d’ionicité de Walden a permis de mettre en évidence la bonne dissociation de LiFSI et LiPF6 dans EC/PC/3DMC assurant ainsi des conductivités ioniques supérieures à 12 mS.cm-1. Bien que possédant des propriétés de transport a priori moins intéressantes dans ce mélange ternaire que les autres sels, LiDFOB forme en réduction une SEI permettant au composite de fournir les meilleures performances en cyclage sans additif avec 560 mA.h.g-1 pour un rendement coulombique de 98,4%. L’ajout d’additif est cependant nécessaire pour atteindre les objectifs fixés par le projet en termes de rendement coulombique (>99,5%). Dans ce cas, l’ajout de 2%VC+10%FEC au mélange ternaire est le plus intéressant avec LiPF6. Le matériau fourni ainsi des capacités de 550 mA.h.g-1 durant une centaine de cycles à un régime de C/5 avec un rendement coulombique de 99,8%. En milieu LI, les performances optimales sont atteintes avec le [EMI][FSI] et 1 mol.L-1 de LiFSI. Le composite atteint alors une capacité de 635 mA.h.g-1 durant 100 cycles à un régime de C/5 avec un rendement coulombique très proche de 100%, tout en s’affranchissant de l’ajout d’additifs. Malgré une viscosité bien plus élevée que celles des mélanges de carbonates d’alkyles, cette formulation permet de générer une SEI plus stable dont la nature, principalement minérale, est issue majoritairement des produits de réduction de FSI-. / This study focuses on new electrolytes and additives in order to improve the cyclability of a Si0.32Ni0.14Sn0.17Al0.04C0.35 negative composite electrode (Si-Sn) and to obtain a stable electrolyte|electrolyte interface. Indeed, like most silicon-based materials, this high-capacity Si-Sn composite (over 600 mA.hg-1) currently suffers from a short cycle life due to volume expansion during charge-discharge processes leading to the degradation of the SEI. To improve the quality of the interface, two kinds of electrolytes were evaluated: (i) mixtures of alkyl carbonates EC/PC/3DMC in which a lithium salt (LiPF6, LiTFSI, LiFSI or LiDFOB) and additives like SEI builder (vinylene carbonate (VC) or fluoroethylene carbonate (FEC)) were added, (ii) ionic liquids (IL) based on quaternary ammonium (N1114+), imidazolium (EMI+) or pyrrolidinium (PYR+) cation, associated with delocalized charge anions such as bis(trifluoromethanesulfonyl)imide (TFSI-) or bis(fluorosulfonyl)imide (FSI-). The Walden diagram confirms the efficient dissociation of LiFSI and LiPF6 in EC/PC/3DM ensuring ionic conductivities as high as 12 mS.cm-1. Although possessing limited transport properties in such a ternary mixture compared to other salts, LiDFOB forms, without additional additives, an high quality SEI allowing the composite to provide the best performances in half cells (560 mA.hg-1 and 98.4% coulombic efficiency). The use of additive is however necessary to reach the objectives fixed by the ANR research project in terms of coulombic efficiency (>99.5%). In this case, the addition of 2%VC+10%FEC to the ternary mixture is the most interesting composition with LiPF6 as lithium salt. So, the Si-Sn nanocomposite material reaches 550 mA.h.g-1 during 100 cycles at C/5 with 99.8% efficiency. In IL, the best performances are achieved in [EMI][FSI]/LiFSI (1 mol.L-1). The performances of the Si-Sn composite reaches 635 mA.h.g-1 for 100 cycles at C/5 with coulombic efficiency close to 100%, without additives. This electrolyte formulation generates a stable SEI which the mainly mineral composition, is predominantly derived from the reduction products of FSI-.

Le fonctionnement des batteries lithium-ion, qu’il soit normal ou dans des conditions abusives, est accompagné d’une génération de gaz en particulier lors des premiers cycles. Celle-ci est intrinsèque au dispositif et est soumise à de nombreux paramètres tels que les matériaux d’électrodes utilisés, l’électrolyte ou encore les conditions opératoires. Cette génération de gaz est délétère : elle conduit à l’augmentation de la pression interne des batteries et pose donc des problèmes de sécurité. Cette étude vise à quantifier les volumes de gaz générés et à comprendre les mécanismes liés à la surpression dans les batteries. A cet effet, le format de batterie « pouch cell » a été adopté tout au long de ce travail de thèse. L’électrolyte choisi est le mélange EC:PC:3DMC + 1 mol.L-1 LiPF6. La première partie de ce travail est dédiée à la mise au point d’un protocole expérimental basé sur (i) l’analyse des matériaux d’électrodes (NMC, LFP, Gr, et LTO), (ii) la solubilité de gaz (O2, H2) comparées à (CO2, CH4) par PVT, et (iii) la quantification des volumes de gaz générés durant le cyclage en pouch cell, corrélée aux performances électrochimiques. Une analyse préalable en demi-piles et en dispositifs complets Gr//NMC et LTO//LFP a également été réalisée afin d’anticiper les performances attendues en pouch cells. Une analyse critique des données (de la littérature et de nos mesures) a permis de définir une procédure optimisée pour obtenir des résultats reproductibles et comparables lors des mesures de volume en pouch cells. La seconde partie de cette thèse consiste en la quantification du volume de gaz produit au cours du cyclage des pouch cells Gr//NMC, Gr//LFP, LTO//LFP et LTO//NMC. Ainsi, les tensions de fin de charge, l’effet du sel et de la température ont été discutés pour dégager les paramètres déterminants dans la génération de gaz en particulier lors de la formation de la SEI. Enfin, une analyse de la composition du gaz récupéré a été effectué par GC-MS et FTIR. A partir de résultats obtenus, des mécanismes ont été proposés et discutés. / The functioning of lithium-ion batteries, may it be under normal use or under abusive conditions, is accompanied by gas generation, especially during the first cycles. This extent of gas generation is dependent on the choice of electrode materials, the electrolyte, and the operating conditions. This gas generation is detrimental: the build-up of pressure leads to the over-pressure in the battery, raising serious concerns. This study is aimed at understanding the fundamental mechanisms governing these reactions. To do so, the « pouch cell » configuration was adopted throughout this thesis. The electrolyte we worked on is the mixture EC:PC:3DMC + 1 mol.L-1 LiPF6. The first chapter of this work is dedicated to development of an experimental protocol based on (i) the analysis of the electrodes materials (NMC, LFP, Gr and LTO), (ii) the gas solubilities (O2, H2) compared to (CO2, CH4) by PVT method, and (iii) the quantification of the volume of generated gases during the cycling of pouch cells which was correlated to the electrochemical performances. A preliminary analysis of half-cells and full cells Gr//NMC and LTO//LFP were also conducted to foresee the performances of the pouch cells. A critical analysis of data taken from the literature and from our own experiments enabled the optimization of a proper procedure to get reproducible and comparable results. The second part of this thesis consists in the quantification of the volume of gases generated during the cycling of Gr//NMC, Gr//LFP, LTO//LFP and LTO//NMC pouch cells. In that respect, the voltages of the end of charge and the effect of salt and of temperature were discussed to figure out the essential parameters in the gas generation and in particular during the formation of SEI. Lastly, a compositional analysis of gases was performed using GC-MS and FTIR. Based on those results, a mechanism is proposed and discussed herein.

Le développement des énergies renouvelables, telles que le solaire photovoltaïque ou l’éolien, est fortement conditionné par la nature intermittente de ces sources d’énergie. Cette intermittence se traduit par un décalage entre pics de production et de consommation. Le stockage de l’énergie électrique revêt donc un caractère primordial dans la gestion de ce décalage. Pour accomplir cette tâche, la technologie lithium-ion est une bonne candidate parmi les technologies de stockage électrochimique de l’énergie. Mais les applications visées exigent des durées de vie bien supérieures à celles requises pour l’électronique portable ou pour les véhicules électriques. En effet les performances des batteries, notamment en termes de capacité, doivent être préservées pendant des durées de 15 à 20 ans. Cette thèse a alors pour but l’étude des mécanismes de vieillissement à long terme d’accumulateurs Li-ion composés d’oxydes lamellaires Li(NixMnyCo1 x y)O2 à l’électrode positive et de graphite à l’électrode négative, en se focalisant sur les interfaces électrode/électrolyte qui sont le lieu privilégié des mécanismes de vieillissement. Ce travail a été réalisé à l'aide de la spectroscopie photoélectronique à rayonnement X (XPS) et de la spectroscopie d’impédance électrochimique (EIS), deux techniques complémentaires particulièrement bien adaptées à l’étude des interfaces, l'une permettant de sonder les environnements chimiques en extrême surface, l'autre donnant la réponse d’un système à une sollicitation électrique sinusoïdale de fréquence variable. La contrainte importante induite par les durées de vie visées (20 ans) ont conduit à simuler le vieillissement à long terme des batteries en leur faisant subir des sollicitations électrochimiques beaucoup plus importantes que lors d’une utilisation normale Les caractérisations par XPS et EIS ont été systématiquement mises en relation avec l’évolution des performances électrochimiques des batteries considérées. Cette étude a permis d'apporter des améliorations aux batteries pour apporter une meilleure réponse à ces phénomènes de vieillissement en termes de maintien des performances: modification de la formulation des électrodes, des électrolytes, de la nature des matériaux actifs, etc. / Development of renewable energy sources such as photovoltaic or wind energy is limited by the intermittent nature of these energy sources. This intermittent nature results in the mismatch between production and consumption peaks. As a result, the storage of electrical energy plays an essential role to manage this mismatch. To this aim, lithium-ion technology appears as a good candidate among other ways of electrochemical storage of energy. However the targeted applications require much greater life span than those commonly admitted for portable electronics or electric vehicles. Battery performances, e.g. rechargeable capacity, should be preserved over 15 or 20 years. This PhD thesis aims at studying the long-term aging mechanisms of Li-ion batteries made up of lamellar oxides Li(NixMnyCo1 x y)O2 at the positive electrode and graphite at the negative electrode. We focused on the electrode/electrolyte interfaces which are the major place of aging processes. The work has been performed by X-ray photoelectron spectroscopy (XPS) and electrochemical impedance spectroscopy (EIS), two complementary techniques especially adapted to the study of interfaces, the former giving access to the chemical environments of atoms at the surface, the latter giving the answer of a system to a sinusoidal electric current with various frequencies. An important technical constraint was the difference between the targeted life span for the application (20 years) and the duration of the thesis (3 years). In order to simulate long-term aging the batteries were submitted to electrochemical stress in much harder conditions than in normal use. XPS and EIS characterizations were constantly related to evolution of electrochemical performances of batteries. This study allowed us during the duration of the project to bring improvements to batteries in order to obtain a better response to aging mechanisms regarding retention of electrochemical performances: e.g. change of electrodes or electrolyte formulation, change of active materials composition, etc.

Målet med projektet var att undersöka samt implementera en fysikbaserad DFN modell för att simulera kalender samt cyklingåldrande av litiumbatterier som används i elbilar. Den fysikbaserade modellen var konstruerad baserad på ett Python biblioteket vid namn PyBaMM, vilket till skillnad från datadrivna modeller ger essentiell information om de kemiska processerna inuti batteriet. Den första delen av projektet täcker konceptet av kalenderåldring, vilket inkluderar en jämförelse mellan tre olika tre olika hastighetsbegränsande SEI modeller. Parametrar som påverkar det erhållna resultatet från modellen är identifierade, estimerade, och till slut validerade för att säkerhetsställa att modellen och parametrarna är identifierbara gentemot experimentella data. Resultatet av jämförelsen gav att SEI tillväxt begränsad av litium interstitiell diffusion är den mest optimala modellen att applicera när kalenderåldring för litiumbatterier ska modelleras. Resultaten visade också att endast en parameter, inre SEI litium interstitiell diffusivitet ska justeras för att erhålla optimal anpassning mot experimentella data. Andra delen av projektet använde resultatet från den första delen och litium plätering implementerades som en andraåldringsmekanism som undersöktes under tre olika laddningsprotokoll. Modellen var optimerad och anpassad gentemot experimentella data, där parametervärdet för kinetisk hasighetskonstanten för plätering var estimerad. Den optimerade modellen användes därefter för att erhålla mer information om elektrokemiska variabler för att kunna analysera samt beskrivaåldringsprocessen utan att behöva genomföra praktiska laborationer. Resultaten visade att mängden pläterat litium på den negativa elektroden ökade för celler som var exponerade till högre ström under laddningsprocessen, samt när cellerna var laddade vid höga SoC nivåer. Sammanfattningsvis, visade modellen hög potential att representera och evaluera experimentella data, samt tillhandahålla en inblick i elektrokemiska processer och kapacitetsförluster länkade till SEI tillväxt och litium plätering. Däremot, för att erhålla en högre grad noggrannhet av elektrokemiskaåldringsmekanismer i litiumbatterier, fler ytterligare mekanismer måste implementeras såsom mekanisk stress av både negativ och positiv elektrod. / The aim of this study was to investigate and apply a physics-based DFN model to simulate the calendar and cycling ageing of lithium-ion batteries manufactured for EV applications. The physics-based cell ageing model was constructed based on the open-source software Python library PyBaMM, which in comparison to data-driven models provides more essential information about the chemical process within the battery cell. The first part of the project covers the concept of calendar ageing which includes comparisons between three different rate-limiting SEI growth models. Parameters that affect the output from the physics-based model are isolated, estimated with numerical methods, and lastly validated to ensure that the model and the parameters rep- resent the physics behind the experimental data. It was found that the SEI growth limited by lithium interstitial diffusion is the most optimal model to apply for a physics-based model when modeling calendar ageing. It was also found that the only parameter that should be tuned against experimental data is the inner SEI lithium interstitial diffusivity. The second part of the project utilizes the results from the first part and introduces lithium plating as a second cell ageing mechanism under three different charging protocols. The model was optimized and fitted against experimental data by sweeping the lithium plating kinetic rate constant parameter. The optimized model was thereafter used to generate outputs that more thoroughly can explain the degradation effects of the cell without constructing real-world experiments. Where increased rate of plated lithium could be observed for the cell subjected to higher charging C-rate, and when the cells were charged at high SoC levels. To summarize, the model showed great potential in representing and evaluating the experimental data and providing the project with insight into the electrochemical processes and cell capacity losses of SEI growth and lithium plating. However, in order to achieve a higher accuracy of cell ageing model in relation to the lithium-ion cells used in customer vehicles, several additional cell degradation mechanisms have to be introduced, such as mechanical degradation of the two electrodes.

Målet med projektet var att undersöka samt implementera en fysikbaserad DFN modell för att simulera kalender samt cyklingåldrande av litiumbatterier som används i elbilar. Den fysikbaserade modellen var konstruerad baserad på ett Python biblioteket vid namn PyBaMM, vilket till skillnad från datadrivna modeller ger essentiell information om de kemiska processerna inuti batteriet. Den första delen av projektet täcker konceptet av kalenderåldring, vilket inkluderar en jämförelse mellan tre olika tre olika hastighetsbegränsande SEI modeller. Parametrar som påverkar det erhållna resultatet från modellen är identifierade, estimerade, och till slut validerade för att säkerhetsställa att modellen och parametrarna är identifierbara gentemot experimentella data. Resultatet av jämförelsen gav att SEI tillväxt begränsad av litium interstitiell diffusion är den mest optimala modellen att applicera när kalenderåldring för litiumbatterier ska modelleras. Resultaten visade också att endast en parameter, inre SEI litium interstitiell diffusivitet ska justeras för att erhålla optimal anpassning mot experimentella data. Andra delen av projektet använde resultatet från den första delen och litium plätering implementerades som en andraåldringsmekanism som undersöktes under tre olika laddningsprotokoll. Modellen var optimerad och anpassad gentemot experimentella data, där parametervärdet för kinetisk hasighetskonstanten för plätering var estimerad. Den optimerade modellen användes därefter för att erhålla mer information om elektrokemiska variabler för att kunna analysera samt beskriva åldringsprocessen utan att behöva genomföra praktiska laborationer. Resultaten visade att mängden pläterat litium på den negativa elektroden ökade för celler som var exponerade till högre ström under laddningsprocessen, samt när cellerna var laddade vid höga SoC nivåer. Sammanfattningsvis, visade modellen hög potential att representera och evaluera experimentella data, samt tillhandahålla en inblick i elektrokemiska processer och kapacitetsförluster länkade till SEI tillväxt och litium plätering. Däremot, för att erhålla en högre grad noggrannhet av elektrokemiska åldringsmekanismer i litiumbatterier, fler ytterligare mekanismer måste implementeras såsom mekanisk stress av både negativ och positiv elektrod. / The aim of this study was to investigate and apply a physics-based DFN model to simulate the calendar and cycling ageing of lithium-ion batteries manufactured for EV applications. The physics-based cell ageing model was constructed based on the open-source software Python library PyBaMM, which in comparison to data-driven models provides more essential information about the chemical process within the battery cell. The first part of the project covers the concept of calendar ageing which includes comparisons between three different rate-limiting SEI growth models. Parameters that affect the output from the physics-based model are isolated, estimated with numerical methods, and lastly validated to ensure that the model and the parameters rep- resent the physics behind the experimental data. It was found that the SEI growth limited by lithium interstitial diffusion is the most optimal model to apply for a physics-based model when modeling calendar ageing. It was also found that the only parameter that should be tuned against experimental data is the inner SEI lithium interstitial diffusivity. The second part of the project utilizes the results from the first part and introduces lithium plating as a second cell ageing mechanism under three different charging protocols. The model was optimized and fitted against experimental data by sweeping the lithium plating kinetic rate constant parameter. The optimized model was thereafter used to generate outputs that more thoroughly can explain the degradation effects of the cell without constructing real-world experiments. Where increased rate of plated lithium could be observed for the cell subjected to higher charging C-rate, and when the cells were charged at high SoC levels. To summarize, the model showed great potential in representing and evaluating the experimental data and providing the project with insight into the electrochemical processes and cell capacity losses of SEI growth and lithium plating. However, in order to achieve a higher accuracy of cell ageing model in relation to the lithium-ion cells used in customer vehicles, several additional cell degradation mechanisms have to be introduced, such as mechanical degradation of the two electrodes.