The search for additional heavy neutral Higgs bosons predicted in Minimal Supersymmetric Extensions of the Standard Model is presented, using the direct decay channel into two tau leptons which themselves decay hadronically. The study is based on proton-proton collisions recorded in 2011 at a center-of-mass energy of 7 TeV with the ATLAS detector at the Large Hadron Collider at CERN. With a sample size corresponding to an integrated luminosity of 4.5 fb−1, no significant excess above the expected Standard Model background prediction is observed and CLs exclusion limits at a 95% confidence level are evaluated for values of the CP-odd Higgs boson mass mA between 140 GeV to 800 GeV within the context of the mhmax and mhmod± benchmark scenarios. The results are combined with searches for neutral Higgs bosons performed using proton-proton collisions at a center-of-mass energy of 8 TeV recorded with the ATLAS detector in 2012, with a corresponding integrated luminosity of 19.5 fb−1. The combination allowed an improvement of the exclusion limit at the order of 1 to 3 units in tan β. Within the context of this study, the structure of additional interactions during a single proton-proton collision (the “underlying event”) in di-jet final states is analyzed using collision data at a center-of-mass energy of 7 TeV recorded with the ATLAS detector in 2010, with a corresponding integrated luminosity of 37 pb−1. The contribution of the underlying event is measured up to an energy scale of 800 GeV and compared to the predictions of various models. For several models, significant deviations compared to the measurements are found and the results are provided for the optimization of simulation algorithms.

This work presents a study of the discovery potential for the neutral supersymmetric Higgs bosons h/A/H decaying to tau pairs with the ATLAS experiment at the LHC. The study is based on Monte Carlo samples which are scaled to state-of-the-art cross sections. The analyses are designed assuming an integrated luminosity of 30 1/fb and a center-of-mass energy of sqrt(s) = 14 TeV. The results are interpreted in the mmax h benchmark scenario. Two final states are analyzed: The dileptonic channel where the two tau leptons decay to electrons or muons and the lepton-hadron channel where one tau decays to an electron or muon and the other tau decays to hadrons. The study of the dilepton channel is based completely on the detailed ATLAS simulation, the analysis of the lepton-hadron channel is based on the fast simulation. The collinear approximation is used to reconstruct the Higgs boson mass and its performance is studied. Cuts are optimized in order to discriminate the signal from background and to maximize the discovery potential given a certain Higgs boson mass hypothesis. In the lepton-hadron channel the selection is split into two analyses depending on the number of identified b-jets. Procedures to estimate the dominant backgrounds from data are studied. The shape and normalization of the Z to tautau background are estimated from Z to leptonlepton control regions. The ttbar contributions to the signal regions are estimated from ttbar control regions. The individual analyses are combined and sensitivity predictions are made depending on the Higgs boson mass mA and the coupling parameter tanβ. The light neutral MSSM Higgs bosons with mA = 150 GeV can be discovered when at least tanbeta = 11 is realized in nature. The heavy neutral MSSM Higgs bosons with mA = 800 GeV can be discovered for tanbeta ≥ 44. However, due to the large width of the reconstructed Higgs boson mass and the mass degeneration, only the sum of at least two of the three Higgs boson signals will be visible.

This thesis presents a search for neutral bosons, such as new Higgs and Z' bosons, predicted by theories extending the Standard Model of particle physics. The search is performed in the di-tau analysis channel, where both tau leptons decay hadronically. Promising candidates of such theories are the Minimal Supersymmetric Standard Model (MSSM) and the non-universal G(221) model, which predict large couplings to tau leptons in large regions of their parameter space. Proton–proton collisions produced by the Large Hadron Collider (LHC) at a center-of-mass energy of 13 TeV, recorded with the ATLAS detector in 2015 and 2016, are analyzed for this search. The data correspond to an integrated luminosity of 36.1 fb^-1. This search relies on sophisticated algorithms for the reconstruction of hadronic tau decays from their decay products. This thesis presents a novel approach employing multivariate techniques to significantly improve existing algorithms, which became the default for reconstruction of hadronic tau decays in ATLAS since 2017. Additionally, the new method can provide useful information for subsequent tau identification algorithms. The MSSM extends the Higgs sector of the Standard Model by four additional Higgs bosons. Of particular interest for this thesis are the neutral CP-even H and CP-odd A bosons. The search for these bosons is performed in the mass range of 0.2 TeV to 2.25 TeV in two orthogonal categories depending on the number of identified b-quarks, each preferring one of the two considered production modes via gluon–gluon fusion or b-associated production. The data are in good agreement with the Standard Model prediction. Upper limits are set on the cross-section times branching fraction using a confidence level (CL) of 95 % independently for both production modes. Most stringent observed limits are found for a resonance mass of 1.5 TeV to be 4.94 fb and 3.65 fb for gluon–gluon fusion and b-associated production, respectively. The results are further interpreted in the hMSSM, mhmax and mhmod scenarios. Observed upper limits in the hMSSM scenario on tanβ are found to be between 4.6 at mA = 0.25 TeV and 41.4 at mA = 1.5 TeV. The search for additional Z' bosons is performed independently on the number of identified b-quarks in the mass range of 0.2 TeV to 4 TeV. As for the search for additional Higgs bosons no significant hint for new physics has been observed. 95 % CL observed upper limits are set on the cross-section times branching fraction for Z' bosons in the Sequential Standard Model (SSM) benchmark scenario between 20.5 pb at mZ' = 0.2 TeV and 7.74 fb at mZ' = 1.75 TeV. The observed upper limit for the highest considered mass of mZ' = 4 TeV is found to be 16 fb. Z' bosons in the SSM and the non-universal G(221) model are excluded at 95 % CL for masses below 2353 GeV and 2232 GeV, respectively.

While holding a firm position in popular culture and science fiction, Dark Matter (DM) is nonetheless a highly relevant topic at the forefront of modern particle physics. We study the applicability of characterizing DM particle candidates SUSY neutralino and sneutrino using Deep Learning (DL) methods. We focus on the monojet and mono-Z signatures and the emergence of missing transverse energy as the result of the undetectable DM candidates. Based on kinematic distributions of outgoing particles as input, a DM candidate classifier is built for each signature, along with a DM mass regressor. The DM candidate classifier obtained near perfect accuracy of 0.995 for the monojet, and 0.978 for mono-Z signature. The monojet and mono-Z mass regressors achieved a Mean Absolute Percentage Error (MAPE) of 17.9 % and 8.5 % respectively. Furthermore, we discuss both the shortcomings and simplification that our choice of model implied, as well as an interpretation of the results. Finally, we debate the prospects of DL in the discovery of new physics and it's use in experiments.

<p>This thesis addresses two separate naturalness issues which generically come to bear on physical theories with large extra dimensions, and so a gravity scale much lower than the Planck scale. The first is related to the observed stability of the proton, wherein we determine the relevant constraints on an additional gauge boson which conserves baryon number. Although several such proposals have been previously considered, our analysis is distinctive in its interest in lighter gauge boson masses (which naturally arise in such models), and in its focus on the dependence of constraints due to kinetic mixing effects. The second is related to the main purpose of large extra dimensions---namely, to address the smallness of the observed vacuum energy---wherein we compute the leading-order quantum corrections to the four-dimensional (4D) vacuum energy resulting from loops of extra-dimensional fields. We compute the contributions from bulk scalars (spin 0), fermions (spin 1/2), and gauge fields (spin 1) in a flux-stabilized, spheroidal extra-dimensional geometry whose rugby-ball shape is due to two codimension-2 branes---one at each pole. (We also obtain the corresponding beta functions for both bulk and brane operators.) These results are then combined to obtain the net contribution from various multiplets in the context of a particular supersymmetric extra-dimensional model that has been shown to give a vanishing result for the 4D vacuum energy at the classical level. Surprisingly, we find that supersymmetry can be preserved dynamically at one loop in the case of identical branes, without arranging any particular relationship between the brane parameters. Perturbing away from the case of identical branes is shown to give a positive 1-loop contribution to the 4D vacuum energy whose size is set by the radius of the extra dimensions.</p> / Doctor of Philosophy (PhD)

Seit der Entdeckung des Higgs-Bosons am Large Hadron Collider ergeben sich auf Grundlage der beobachteten Masse weitreichende Konsequenzen für das Standardmodell der Teilchenphysik (SM) sowie für darüber hinausgehende physikalische Theorien. In supersymmetrischen Modellen ist es möglich, die Masse des SM-artigen Higgs-Bosons vorherzusagen. Dies lässt Rückschlüsse auf das Massenspektrum der bisher unentdeckten Teilchen zu. Die gegenwärtigen Ausschlussgrenzen für einige supersymmetrische Teilchen liegen oberhalb der TeV-Skala. In Szenarien mit großen Abständen im Massenspektrum wird die Präzision der Higgs-Massenvorhersage durch die Resummation logarithmischer Terme, die von den Massen der zusätzlichen Teilchen abhängen, erheblich verbessert. Die Methode FlexibleEFTHiggs ermöglicht eine Berechnung der Higgs-Masse, die sowohl große Logarithmen resummiert als auch alle weiteren Terme bis zu einer festen Ordnung in der Störungstheorie berücksichtigt. In vorhergehenden Arbeiten wurde die Methode FlexibleEFTHiggs als Modul in das Softwarepaket FlexibleSUSY integriert. Um die Genauigkeit der Higgs-Massenberechnung zu optimieren, werden in dieser Arbeit Erweiterungen des FlexibleEFTHiggs-Ansatzes beschrieben, wobei ein Schwerpunkt in der konkreten Anwendung auf das Minimale Supersymmetrische Standardmodell (MSSM) besteht. Dies erfordert eine Weiterentwicklung des Verfahrens zur Bestimmung der quartischen Higgs-Kopplung. In diesem Zusammenhang wird die Bedeutung der Parametrisierung in den verwendeten Relationen erläutert und die Reihenentwicklung in SM-Kopplungen konsequent auf ein-Schleifenniveau trunkiert. Die vorgestellte Modifikation ist modellunabhängig und ermöglicht eine Resummation der nächstführenden Logarithmen (NLL). In Bezug auf die Resummation von Logarithmen höherer Ordnung sind weitere Änderungen in der Gleichung zur Bestimmung der quaritschen Kopplung erforderlich. Hierbei erweist es sich als vorteilhaft, die Beziehung durch eine Reihenentwicklung unter Verwendung der Parameter des vollständigen Modells auszudrücken. Basierend auf dem bestehenden MSSM-Spektrumgenerator in FlexibleSUSY, welcher Higgs-Massenkorrekturen bis zum Dreischleifenniveau berücksichtigt, wird eine Higgs-Massenberechnung unter Verwendung der hier entwickelten Neugestaltung der FlexibleEFTHiggs-Methode realisiert. In Verbindung mit dem Lösen der Gleichungen der Renormierungsgruppen auf vier-Schleifenniveau, ergibt sich somit eine Resummation der Logarithmen von bis zu N3LL. Zudem ermöglicht die Parametrisierung mittels der Parameter des vollen Modells eine Resummation von Beiträgen mit höchster Potenz in der Squark-Higgs-Kopplung Xq. Um dies darzulegen, werden systematische Betrachtungen des Auftretens von Xq in MSSM Green-Funktionen und Schwellenkorrekturen in zwei Parametrisierungen vorgestellt. Anschließend werden Resummationsgleichungen angegeben und die verbesserten Konvergenzeigenschaften aufgrund der Xq-Resummation in der neuen FlexibleEFTHiggs-Berechnung demonstriert. Zudem werden weitere numerische Ergebnisse vorgestellt und eine ausführliche Analyse der theoretischen Unsicherheiten für Referenzszenarien diskutiert. / Since the discovery of the Higgs boson at the Large Hadron Collider, the observed mass led to far-reaching consequences for the Standard Model of particle physics (SM) and for theories beyond it. In supersymmetric models, the mass of the SM-like Higgs boson can be predicted from the full-model information. This allows insights into the mass spectrum of the extended particle content that has not yet been revealed by experiments. In fact, the current exclusion limits for some of the supersymmetric particles exceed the TeV-threshold. In scenarios with a large mass hierarchy, the accuracy of Higgs-mass predictions is significantly improved by resumming logarithmic terms, which depend on the mass of the exotic particles. A way to efficiently organize the Higgs-mass calculation with a resummation of large logarithms, while including contributions at a fixed order, is achieved by the method FlexibleEFTHiggs. In previous work, the method FlexibleEFTHiggs was implemented as a module in the public code FlexibleSUSY. To increase the precision of the Higgs-mass calculation, this thesis describes several improvements of the FlexibleEFTHiggs approach, with a focus on the application to the Minimal Supersymmetric Standard Model (MSSM). We discuss a refined high-scale matching procedure allowing a consistent resummation of next-to leading logarithms (NLL) independent of the SUSY model. In this process, we address the aspect of parametrizing the high-scale matching in terms of SM couplings in conjunction with a strict truncation at 1-loop. With respect to the resummation of higher-order logarithms, a novel strategy is proposed to organize the high-scale matching. It is shown that the ideal setting is given if the matching is expressed in a perturbative expansion of full-model parameters. Building upon the existing MSSM spectrum generator from FlexibleSUSY, where the state-of-the-art self-energies and tadpoles are available at 2-loop and 3-loop orders, we have constructed a Higgs-mass calculation using a redesign of the FlexibleEFTHiggs method. Combined with a renormalization group running up to 4-loop, the Higgs-mass prediction achieves log-precision of N3LL. Analyzing the difference between the parametrizations in terms of SM parameters and full-model parameters, we discovered the ability to resum terms of maximal power in the squark-Higgs parameter Xq. To illustrate this statement, we perform a systematic investigation on the appearance of Xq in MSSM Green functions and threshold corrections in both parametrizations. As a result, a plethora of all-order resummation relations is presented, and we demonstrate the improved convergence properties of our updated calculation induced by the Xq-resummation. We provide numerical results and a thorough analysis of theoretical uncertainties for standard benchmark scenarios.

Dans cette thèse, nous analysons les propriétés géométriques des surfaces obtenues des solutions classiques des modèles sigma bosoniques et supersymétriques en deux dimensions ayant pour espace cible des variétés grassmanniennes G(m,n). Plus particulièrement, nous considérons la métrique, les formes fondamentales et la courbure gaussienne induites par ces surfaces naturellement plongées dans l'algèbre de Lie su(n). Le premier chapitre présente des outils préliminaires pour comprendre les éléments des chapitres suivants. Nous y présentons les théories de jauge non-abéliennes et les modèles sigma grassmanniens bosoniques ainsi que supersymétriques. Nous nous intéressons aussi à la construction de surfaces dans l'algèbre de Lie su(n) à partir des solutions des modèles sigma bosoniques. Les trois prochains chapitres, formant cette thèse, présentent les contraintes devant être imposées sur les solutions de ces modèles afin d'obtenir des surfaces à courbure gaussienne constante. Ces contraintes permettent d'obtenir une classification des solutions en fonction des valeurs possibles de la courbure. Les chapitres 2 et 3 de cette thèse présentent une analyse de ces surfaces et de leurs solutions classiques pour les modèles sigma grassmanniens bosoniques. Le quatrième consiste en une analyse analogue pour une extension supersymétrique N=2 des modèles sigma bosoniques G(1,n)=CP^(n-1) incluant quelques résultats sur les modèles grassmanniens. Dans le deuxième chapitre, nous étudions les propriétés géométriques des surfaces associées aux solutions holomorphes des modèles sigma grassmanniens bosoniques. Nous donnons une classification complète de ces solutions à courbure gaussienne constante pour les modèles G(2,n) pour n=3,4,5. De plus, nous établissons deux conjectures sur les valeurs constantes possibles de la courbure gaussienne pour G(m,n). Nous donnons aussi des éléments de preuve de ces conjectures en nous appuyant sur les immersions et les coordonnées de Plücker ainsi que la séquence de Veronese. Ces résultats sont publiés dans la revue Journal of Geometry and Physics. Le troisième chapitre présente une analyse des surfaces à courbure gaussienne constante associées aux solutions non-holomorphes des modèles sigma grassmanniens bosoniques. Ce travail généralise les résultats du premier article et donne un algorithme systématique pour l'obtention de telles surfaces issues des solutions connues des modèles. Ces résultats sont publiés dans la revue Journal of Geometry and Physics. Dans le dernier chapitre, nous considérons une extension supersymétrique N=2 du modèle sigma bosonique ayant pour espace cible G(1,n)=CP^(n-1). Ce chapitre décrit la géométrie des surfaces obtenues des solutions du modèle et démontre, dans le cas holomorphe, qu'elles ont une courbure gaussienne constante si et seulement si la solution holomorphe consiste en une généralisation de la séquence de Veronese. De plus, en utilisant une version invariante de jauge du modèle en termes de projecteurs orthogonaux, nous obtenons des solutions non-holomorphes et étudions la géométrie des surfaces associées à ces nouvelles solutions. Ces résultats sont soumis dans la revue Communications in Mathematical Physics. / In this Ph. D. thesis, we analyze the geometric properties of surfaces obtained from the classical solutions of the two-dimensional bosonic and supersymmetric sigma models which has Grassmann manifolds G(m,n) as target space. In particular, we consider the metric, the fundamental forms and the gaussian curvature induced by these surfaces which naturally live in the su(n) Lie algebra. The first chapter presents some preliminary tools to understand the elements of the following chapters. We present non-abelian gauge theories and bosonic grassmannian sigma models as well as its supersymmetric counterpart. Another section presents a construction of surfaces in the Lie algebra su(n) from the solutions of the bosonic sigma models. The three last chapters contained in this thesis presents the constraints that have to be imposed on the solutions of the models in order to generate constant gaussian curvature surfaces. From these constraints, we can give a classification of the solutions depending on the possible values of the curvature. The first two papers presents an investigation of these surfaces and of their associated solutions for the bosonic grassmannian sigma models. In the third paper, we generalize our approach to a supersymmetric extension of the bosonic CP^(n-1)= G(1,n) sigma model including some results for the general Grassmann manifold G(m,n). In chapter 2, we study the geometric properties of surfaces associated to holomorphic solutions of the grassmannian sigma models. We give a complete classification of these constant curvature solutions for the particular models G(2,n) with n=3,4,5. Furthermore, we establish two conjectures on the possible values of the gaussian curvature. We also give some elements of proof for these conjectures in terms of Plücker coordinates and immersions as well as Veronese curves. These results are published in the Journal of Geometry and Physics. The third chapter presents a similar analysis as in the second chapter in the case of non-holomorphic solutions of the bosonic grassmannian sigma models. This work generalizes the results obtained in the first paper and give a systematic algorithm to obtain such surfaces from the known solutions of the models. These results are published in the Journal of Geometry and Physics. In the last chapter of this thesis, we consider a N=2 supersymmetric extension of the bosonic sigma model which has the CP^(n-1)=G(1,n) manifold as target space. This chapter presents a geometric description of the surfaces obtained from the solutions of the model and shows, in the holomorphic case, that they have constant gaussian curvature if and only if the solutions consists of a generalization of the Veronese curve. Furthermore, using a gauge invariant formulation of the model in terms of orthogonal projectors, we obtain explicit non-holomorphic solutions and study the geometry of their associated surfaces. These results are submitted to Communications in Mathematical Physics.

La théorie de la Supersymétrie est étudiée ici en tant que théorie complémentaire au Modèle Standard, sachant que celui-ci n'explique qu'environ 5% de l'univers et est incapable de répondre à plusieurs questions fondamentales en physique des particules. Ce mémoire contient les résultats d'une recherche de Supersymétrie effectuée avec le détecteur ATLAS et utilisant des états finaux contenant entre autres une paire de leptons de même charge électrique ou trois leptons. Les données proviennent de collisions protons-protons à 13 TeV d'énergie dans le centre-de-masse produites au Grand Collisionneur de Hadrons (LHC) en 2015. L'analyse n'a trouvé aucun excès significatif au-delà des attentes du Modèle Standard mais a permis tout de même de poser de nouvelles limites sur la masse de certaines particules supersymétriques. Ce mémoire contient aussi l'étude exhaustive d'un bruit de fond important pour cette analyse, soit le bruit de fond provenant des électrons dont la charge est mal identifiée. L'extraction du taux d'inversion de charge, nécessaire pour connaître combien d'événements seront attribuables à ce bruit de fond, a démontré que la probabilité pour que la charge d'un électron soit mal identifiée par ATLAS variait du dixième de pourcent à 8-9% selon l'impulsion transverse et la pseudorapidité des électrons. Puis, une étude fut effectuée concernant l'élimination de ce bruit de fond via l'identification et la discrimination des électrons dont la charge est mal identifiée. Une analyse multi-variée se servant d'une méthode d'apprentissage par arbres de décision, basée sur les caractéristiques distinctives de ces électrons, montra qu'il était possible de conserver un haut taux d'électrons bien identifiés (95%) tout en rejetant la grande majorité des électrons possédant une charge mal identifiée (90-93%). / Since the Standard Model only explains about 5% of our universe and leaves us with a lot of open questions in fundamental particle physics, a new theory called Supersymmetry is studied as a complementary model to the Standard Model. A search for Supersymmetry with the ATLAS detector and using final states with same-sign leptons or three leptons is presented in this master thesis. The data used for this analysis were produced in 2015 by the Large Hadron Collider (LHC) using proton-proton collisions at 13 TeV of center-of-mass energy. No excess was found above the Standard Model expectations but we were able to set new limits on the mass of some supersymmetric particles. This thesis describes in detail the topic of the electron charge-flip background, which arises when the electric charge of an electron is mis-measured by the ATLAS detector. This is an important background to take into account when searching for Supersymmetry with same-sign leptons. The extraction of charge-flip probabilities, which is needed to determine the number of charge-flip events among our same-sign selection, was performed and found to vary from less than a percent to 8-9% depending on the transverse momentum and the pseudorapidity of the electron. The last part of this thesis consists in a study for the potential of rejection of charge-flip electrons. It was performed by identifying and discriminating those electrons based on a multi-variate analysis with a boosted decision tree method using distinctive properties of charge-flip electrons. It was found that we can reject the wide majority of mis-measured electrons (90-93%) while keeping a very high level of efficiency for well-measured ones (95%).

In dieser Arbeit stellen wir verschiedene Tests der (M)atrixtheorie-Vermutung vor. Die (M)atrixtheorievermutung besagt, dass die Dynamik von M-Theorie, der Urtheorie, die alle bekannten Stringtheorien und auch elfdimensionale Supergravitation als bestimmte Grenzfaelle enthalten soll, durch ein quantenmechanisches Matrixmodell gegeben ist. Insbesondere untersuchen wir Streuprozesse sowohl aus Sicht des Matrixmodells, als auch aus Sicht der Supergravitation, und vergleichen die resultierenden S-Matrixelemente. Wir finden beeindruckende Uebereinstimmung zwischen den beiden Theorien, solange wir uns auf klassische Supergravitation beschraenken. Sobald wir auch Quanteneffekte auf der Supergravitationsseite einbeziehen, hat diese Uebereinstimmung keinen Bestand. Des weiteren untersuchen wir die Frage, ob Loesungen der klassischen Matrixmodell-Bewegungsgleichungen mit Impulsuebertrag existieren, und finden eine negative Antwort. / In this thesis, we present several tests of the M(atrix)-Model conjecture that asserts that the dynamics of M-Theory, the eleven-dimensional Ur-theory containing all known string theories and also eleven-dimensional supergravity in specific limits, is given by a quantum mechanical matrix model. In particular, scattering processes are analyzed both from the M(atrix)-Model and from the supergravity perspective and the corresponding S-matrix elements are compared. We find impressive agreement between these two theories as long as only classical supergravity is considered. If one includes also quantum effects on the supergravity side, the agreement does not persist. In addition to these calculations, the question of the existence of classical solutions to the M(atrix)-Model equations of motion with momentum transfer is addressed and answered negatively.

Cette thèse présente trois études différentes basées sur les données de CMS du Run 2. Les deux premières sont des mesures des propriétés des amas dans le trajectographe à pistes de silicium de CMS, liées respectivement aux particules hautement ionisantes (HIP) et au partage de charge entre les pistes voisines (également appelé diaphonie). Le dernier sujet abordé dans ce document est la recherche du partenaire supersymétrique du quark top, appelé stop. Une augmentation de l’inefficacité de reconstruction des hits dans le trajectographe à pistes de silicium de CMS a été observée au cours des années 2015 et 2016. Les particules hautement ionisantes ont été identifiées comme une cause possible de ces inefficacités. Cette thèse apporte des résultats qualitatifs et quantitatifs sur l’effet HIP et sa probabilité. Le HIP n’était pas la source la plus importante d’inefficacité et, une fois la source identifiée et corrigée, les nouvelles données révèlent qu’après cette correction, le HIP représente à présent la principale source d’inefficacité. La seconde étude présentée porte sur les conditions utilisées dans la simulation du trajectographe par CMS afin de fournir des résultats réalistes. Ces conditions changent avec les conditions de fonctionnement du trajectographe et évoluent avec le vieillissement du trajectographe résultant des dommages causés par le rayonnement. Nous avons constaté que les paramètres de diaphonie obsolètes avaient une grande incidence sur la forme de l’amas. Dans cette thèse, les paramètres ont été réévalués et il a été confirmé que les nouveaux paramètres améliorent grandement l’accord des amas entre données et simulation. La dernière partie décrit en profondeur la recherche de stop en utilisant les données collectées en 2016 (correspondant à ∫L = 35.9 fb−1) avec un lepton dans l’état final. Aucun excès n’a été observé par rapport aux prédictions attendues par le modèle standard et les résultats ont été interprétés en terme de limites d’exclusion sur des modèles simplifiés. / This thesis presents three different studies based on the CMS Run 2 data. The first two are measurements of the cluster properties in the CMS silicon strip tracker related respectively to the highly ionizing particles (HIP) and the charge sharing among neighboring strips (also known as cross talk). The last topic discussed in this document is the search for the supersymmetric partner of the top quark, called the stop. An increase in the hit inefficiency of the CMS silicon strip tracker was observed during the years 2015 and 2016. The highly ionizing particles were identified as a possible cause of these inefficiencies. This thesis brings qualitative and quantitative results on the HIP effect and its probability. The HIP was found not to be the largest source of inefficiency at that time and once the source was identified and fixed, the new data revealed that after this fix the HIP now represents the major source of the hit inefficiency. The second study presented in this thesis focuses on the conditions plugged in CMS tracker simulation in order to provide realistic results. These conditions change with the tracker operating conditions and also evolve with tracker ageing resulting from the radiation damage. We identified that the outdated cross talk parameters largely impact the cluster width and seed charge. In this thesis the parameters were remeasured and it was confirmed that the new parameters largely improve the agreement of clusters between data and simulation. The last part describes deeply the stop analysis using data recorded in 2016 (corresponding to ∫L =35.9 fb−1) with single lepton in the final state. No excess was observed in the full 2016 data (∫L = 35.9 fb−1) with respect to the standard model background predictions and therefore exclusion limits in terms of simplified model spectra were derived.