Atomic clusters are attractive candidates for building motifs for new nano-assembled materials with desirable properties. At this nano-regime of matter, the size, shape, and composition of clusters changes their electronic structure and hence their properties. Computational modeling must work hand in hand with experiment to provide robust descriptions of the geometries and energetics of atomic clusters and how they might behave in a nano-assembled material. To this end, we have investigated three distinct species as model systems: antimony oxides SbxOy (x = 1, 2; y = 0 - 3), metal ion-solvent complexes Mm(NH3)n (M = Bi, Pb; m = 1 - 2, n = 0 - 4), and quantum dots Z10H16 (Z = Si, Ge) and β-Sn12H24. Their geometries and electronic structures have been determined using gradient-corrected density functional theory. The relative stabilities for antimony oxides have been examined by the respective comparison of highest-occupied and lowest-unoccupied molecular orbital (HOMO-LUMO) gaps and atomization energies. The superior electronic stability of Sb2O3 is indicated by its closed shell structure, wide HOMO-LUMO gap calculated to be 3.11 eV, and high atomization energy of 4.21 eV. Spin-orbit corrections were necessary for accurate calculation of the metal-solvent energetics, closing the gap between experimental and theoretical values by 1.05 eV for the electron affinity of the Pb atom. Quantum dot modeling of the well-established Si and Ge as well as the less-investigated Sn illuminated the accuracy of the CEP basis sets and the B3LYP functional over other DFT computational routes for clusters containing elements beyond the third row. Throughout, the results correlate well with experiment and higher order ab initio methods where data is available. These comparisons validate the accuracy of the computational routes used. This document was prepared in the Linux Ubuntu Open Office Suite 2.4.1.

Thèse numérisée par la Direction des bibliothèques de l'Université de Montréal.

La famille des alliages d'Heusler regroupe plusieurs composés considérés comme étant des candidats de choix pour être intégrés en tant qu'électrode magnétique dans des dispositifs d'électronique de spin performants. Les plus intéressants de ces alliages présentent en effet des températures de Curie élevées, une demi-métallicité théorique ainsi qu'un faible coefficient d'amortissement de Gilbert. Expérimentalement, les résultats obtenus sont cependant généralement moins probants que ceux annoncés numériquement. La première partie de cette thèse est donc consacrée à l'étude ab initio d'hypothèses usuellement utilisées pour expliquer les différences entre mesures expérimentales et résultats théoriques. Des calculs basés sur la théorie de la fonctionnelle de la densité ont été effectués pour comprendre l'impact de défauts structuraux (désordres chimiques partiels, déformations tétragonales, lacunes) sur les propriétés électroniques et magnétiques statiques et dynamiques d'alliages full-Heusler massifs à base de Co (Co2MnSi, Co2MnSn, Co2MnAl et Co2FeAl). Dans la seconde partie de cette thèse nous proposons d'étudier, avec les mêmes outils numériques, les caractéristiques physiques d'hétérostructures "tout-Heusler" prometteuses pour l'électronique de spin et rares dans la littérature. Nous nous sommes concentrés sur les variations des propriétés électroniques aux voisinages des interfaces demi- métal/isolant (Fe2TiSi/Co2MnSi, CoTiAs/Co2MnSi) ou demi-métal/métal (Fe2VAl/Co2MnSi, RhNiSi/Co2MnSi) composant les multicouches étudiées. Les résultats obtenus démontrent l'intérêt certain de ces structures pour des applications en électronique de spin, telles que les vannes de spin ou les jonctions tunnel magnétiques. / The Heusler alloy family contains several compounds considered to be prime candidates to be integrated as magnetic electrode into high-?performance spintronic devices. Some of these alloys indeed exhibit high Curie temperatures, have been predicted theoretically to be half-?metallic, and display a low Gilbert damping parameter. Nevertheless, the experimental results are generally less convincing than those reported numerically. The first part of this thesis is devoted to the ab initio study of hypotheses that are usually used to explain the differences between experimental and theoretical results. Calculations, based on the density functional theory, are then used to understand the impact of structural defects (partial chemical disorders, tetragonal deformation, vacancies) on the static and dynamic electronic and magnetic properties of Co-?based bulk full-Heusler alloys (Co2MnSi, Co2MnSn, Co2MnAl et Co2FeAl). In the second part of this thesis we propose to study, with the same numerical tools, the physical characteristics of "all-Heusler" heterostructures promising for spintronics and rare in the literature. We have focused on the variations of electronic properties in the vicinity of the "half-metal/insulator" (Fe2TiSi/Co2MnSi, CoTiAs/Co2MnSi) or "half-metal/non-magnetic metal" (Fe2VAl/Co2MnSi, RhNiSi/Co2MnSi) interfaces composing the studied multilayers. The obtained results highlight the interest of these structures for spintronic devices such as spin valves or magnetic tunnel junctions.

Neste trabalho foram realizadas simulações computacionais para investigar as propriedades eletrônicas de nanossistemas baseados em nanotubos de carbono e grafeno por meio de cálculos de primeiros princípios. Um dos nanossistemas investigados é formado por um nanotubo de carbono acoplado a eletrodos de nanofios de paládio encapsulados. Foi mostrado que estados provenientes dos eletrodos interagem fortemente com os estados do nanotubo de carbono. Cálculos de transporte eletrônico foram realizados para investigar a potencialidade desse nanossistema em aplicações como transistor de efeito de campo. Foi mostrado que a intensidade da corrente elétrica desse nanossistema pode ser variada com o campo elétrico de gate. Outro trabalho desenvolvido no presente trabalho tem como base um nanossistema formado pelo grafeno depositado nos substratos SiO2 amorfo e h-BN. Foi determinada a energia de adsorção e a quantidade de carga transferida para investigar a influênicas desses substratos na adsorção da molécula de H2 pelo grafeno. Foi mostrado que a energia de adsorção da molécula de H2 adsorivda na interface grafeno/SiO2 amorfo é menor em comparação com o grafeno suspenso ou disposto sobre o substrato h-BN. Além disso, a adsorção do H2 nessa região resulta em uma transferência de carga de uma ordem de grandeza maior em comparação com a adsorção no grafeno suspenso, sendo observado um deslocamento do Cone de Dirac em relação ao nível de Fermi. Esse estudo poderá contribuir para a construção de futuros sensores de H2 à base de grafeno. / In this work, ab initio calculations were performed within DFT framework to analyse electronic properties of Carbon nanotubes and grapheme nano systems. In this work, computer simulations were performed to investigate the electronic properties of nanosystems based on carbon nanotubes and graphene within DFT framework. One of these systems studied is a Carbon nanotube semiconductor coupled to encapsulated leads of Pd nanowires. It has been shown that leads states interact strongly with the carbon nanotube states. Electronic transport calculations were performed to unfold new applications of this system, such as the field effect transistor. We noticed that charge current intensity can be tuned by electrical field. We also described the influence of amorphous SiO2 and h-BN, in H2 energy adsorption and charge transfer, where both materials are used as graphene substrates. It was shown that the latter adsorption energy in the graphene/Si02 is smaller than graphene/h-Bn and the graphene suspended itself. In fact this adsorption results in a charge transference one order greater than in the suspended graphene, which can be seen as a vertical shift of the Dirac Cone. This study may improve the construction of future H2 sensors based on graphene.

Neste trabalho estudamos o efeito hidrofóbico no sistema benzenoágua. Nossa estratégia é avaliar as propriedades das ligações de hidrogênio entre as moléculas de água nas proximidades da molécula e benzeno. Utilizamos as ferramentas da mecânica quântica e o método de simulação computacional para este estudo. A análise estrutural e energética detalhada dos clusters benzenoágua mostra que a ligação de hidrogênio é mais forte quando na presença do benzeno. A investigação pelo método de simulação Monte Carlo corrobora estas conclusões e ainda fornece os efeitos da variação de tempeatura. Verificamos que o aumento da temperatura afeta todas as moléculas aumentando a desordem líquida, no entanto, constatamos a manutenção de uma estrutura de ligações de hidrogênio mais fortes as proximidades do benzeno. A interação entre duas moléculas de benzeno também foi analisada, mostrando que a interação benzenobenzeno é bem mais forte na presença da água. / The hydrophobic effect is studied for the benzene-water system. The properties of the hydrogen bond between the water molecules around the benzene is evaluated using both classical and quantum mechanical methods. Hydrophobic hydration analysis shows that the hydrogen bond interaction is stronger in the presence of benzene. This is verified both by ab initio quantum mechanical methods and classical Monte Carlo simulation. Temperature dependence is investigated. Although increasing temperature increases the disorder the hydrogen bonds between the water molecules are still stronger for those in the proximities of the benzene. Hydrophobic interaction is also investigated. It is seen that the benzene-benzene interaction is stronger in the water environment.

A molécula de hidrazina (N2H4) se decompõe endotermicamente. Ela é geralmente utilizada como propelente em satélites ou na composição de bipropelentes em veículos espaciais. Nesse trabalho a decomposição da hidrazina foi estudada na presença dos átomos de nitrogênio e oxigênio. Para tal estudo foram utilizados métodos de química quântica para calcular as propriedades termoquímicas das reações elementares e métodos de cinética química para obtenção das constantes de velocidade dessas reações. Com o objetivo de verificar a metodologia com melhor razão entre o custo computacional e qualidade do resultado para o estudo das propriedades termoquímicas, foram utilizados métodos ab initio altamente correlacionados, e esses resultados foram comparados aos resultados obtidos com diferentes funcionais DFT de correlação e troca. Para obter as constantes de velocidade das reações, foram utilizados os métodos TST e CVT, e os efeitos não clássicos foram considerados pelas aproximações ZCT e SCT. Para o cálculo CVT, a superfície de potencial foi construída pela metodologia dual-level, utilizando um funcional DFT para obter as propriedades ao longo do caminho reacional, e esses resultados foram melhorados com resultados obtidos pelo método CCSD(T) nas geometrias de equilíbrio.

Estudamos a adsorção de Te em superfícies de GaAs(001) e InAs(001) com periodicidades 1x1, 1x2, 2x1 e 2x2, para as concentrações de telúrio Tc= 0, 1/4, 1/2, 3/4 e 1. Realizamos cálculos dentro do formalismo do funcional da densidade empregando pseudopotenciais de norma conservada. Para a relaxação das estruturas empregamos a dinâmica molecular de Car e Parrinello. Nossos resultados apontam para uma redução na estabilidade das superfícies na proporçã em que aumenta a concentração de Te na uperfície. A cobertura de As (Tc = 0) é energeticamente mais favorável que as recobertas com qualquer concentração de Te, tanto na superfície de GaAs(001) quanto na de InAs(001). Observouse ainda nas superfícies com Tc = 0 Tc = 1, que a dimerização dos átomos de As e da ordem de 30% mais intensa que dos átomos e Te. Comparando as dimerizações do Te nas superfícies com concentrações Tc 1/2, observamos que elas são maiores sobre o InAs(001) (célula terminada em In) que sobre o GaAs(001) terminada em Ga). Outra tendência verificada e a \"flutuação\" do Te sobre as superfícies. Para uma mesma concentração verificamos a \"preferência de adsorção\" em sítios fora da cadeia de dímeros indicando uma adsorção monoatômica. Para concentração Tc = 1 na superfície GaAs(001):Te2x2 observamos duas geometrias possíveis: uma com cadeias de dímeros seguindo o modelo \"dimerrowmissing\" e outra com dois tipos de dímeros em posições alternadas ao longo da dir~ao (110). Analisando as energias de adsorção nas duas superfícies verificamos que a adsorção sobre o InAs é mais favorável que sobre o GAs. Analisamos ainda a estrutura eletrônica das superfícies em todas as reconstruções e concentrações consideradas e verificamos anda o caráter semicondutor das superfícies com concentração Tc = 1/2. / We have carried out ab initio density functional calculations to investigate the adsorption of Te on GaAs(001) and lnAs(001) surfaces as a function of Te surface coverage: Tc = 0, 1/4, 1/2, 3/4, 1. In order to determine the equilibrium atomic positions, the geometry was relaxed according he calculated total energy and forces following the Car Parrinello approach for molecular dynamics. Our calculations indicate that a full monolayer of As (Tc = 0) is energetically more favourable than any of the studied coverage of Te (Tc = 1/4, 1/2, 3/4, 1), where the stability is educed with increasing Te coverage. The dimerization of surface As atoms is about 30% more intense than surface Te atoms. Comparing the Te dimerization on InAs(001), In terminated, and aAs(OOI), Ga terminated, we observed that the Te atoms dimerize more over lnAs than GaAs surface. Another observation is the tendency of the Te atoms to \'l1oat\" from the surface with increasing coverage. For the same concentrations of Te the atoms \"prefer\" to be adsorbed on the offchain sites indicating a monoatomic adsorption. The adsorption energy of Te on InAs(001) is more favorable than GaAs(001). We also determined the surface band structure for all reconstructions and Te concentrations, veryfying the semiconductor nature for Tc = 1/2.

As the foremost production of electricity in the United State comes from coal-fired plants, there is much more to learn on the topic of mercury which is a common component in coal. The speciation of mercury in the flue gas determines the best control technology for a given system. Because of the difficulty in measuring mercury at different stages of the process, it is practical to use mercury reaction kinetics to theoretically determine mercury speciation based upon coal composition, plant equipment and operating conditions. Elemental mercury cannot be captured in wet scrubbers; however, its oxidized forms can. Chlorine is a reasonable oxidizing agent and is naturally found in bituminous coal, but bromine is an even better oxidizing agent because of its larger size, it has stronger London dispersion force interactions with mercury. Bromine additive technologies have recently been implemented in several companies to enhance mercury oxidation. Because capture technologies are highly dependent upon the form of mercury that is present, investigations into their speciation are extremely important. Though there have been numerous efforts to study mercury compounds as relevant to atmospheric studies, there is little data currently available for mercury compounds found in combustion flue gases. It would be particularly beneficial to obtain kinetic rate constants at various high temperature and pressure conditions typical for a combustion system. Prevalent species of mercury containing bromine in coal combustion flue gases were studied using density functional theory (DFT) and a broad range of ab initio methods. Reaction enthalpies, equilibrium bond distances, and vibrational frequencies were all predicted using DFT as well as coupled cluster (CC) methods. All electronic calculations were carried out using the Gaussian03 or MOLPRO software programs. Kinetic predictions of three first-stage and three second-stage oxidation reactions involving the formation of oxidized mercury via bromine containing compounds are presented. Understanding the speciation of mercury in the flue gases of coal combustion is paramount in developing efficient technologies to ensure its capture.

Les composites à matrice céramique SiC/SiC sont des matériaux prometteurs pour le remplacement des superalliages dans les moteurs pour l’aviation civile et militaire. La matrice de SiC des composites est réalisée par dépôt chimique en phase vapeur (CVD). Le précurseur actuellement utilisé à l’échelle industrielle est le méthyltrichlorosilane (MTS, CH3SiCl3) dilué dans le dihydrogène (H2). Ce système chimique peut dans certains cas conduire à des écarts à la stoechiométrie des dépôts ainsi qu’à des gradients de densification au sein du composite final. Deux nouvelles molécules ont été envisagées dans le cadre de la thèse pour remédier aux inconvénients du MTS. Le vinyltrichlorosilane (VTS, C2H3SiCl3), composé d’un groupement vinyle (deux carbones doublement liés), susceptible de stabiliser la liaison Si-C de la molécule et favoriser une réactivité couplée de silicium et du carbone, propice au dépôt de SiC pur. Le méthylsilane (MS, CH3SiH3), qui ne possède pas d’atomes de chlore limitant la réactivité hétérogène. Son utilisation devrait en outre permettre de réduire le traitement des effluents en sortie de réacteur.La pertinence de ces deux précurseurs a été évaluée par une approche expérimentale mettant en évidence les régimes cinétiques associés aux deux molécules et en s’intéressant à la nature du solide déposé. Cette approche est couplée à une modélisation de la phase homogène et à des mesures IRTF des gaz froids sortant du réacteur, dans le but de comprendre les phénomènes chimiques mis en jeu pendant le dépôt.À basse température, la décomposition du VTS produit des hydrocarbures réactifs qui inhibent le dépôt de SiC et conduisent à un large excès de carbone libre. Le solide devient stoechiométrique à température moyenne, au-delà d’un domaine de bi-stabilité de la vitesse de dépôt (R). C2H4 et SiCl4 sont alors majoritairement formés. Les transferts de masse (TM) limitent la vitesse de dépôt à haute température et un léger excès de C réapparait dans le solide, associé à la formation C2H2.Le système MS/Ar réagit à plus basse température que VTS/H2 et le rendement en solide est beaucoup plus élevé (seules des traces de CH4 sont détectées à haute température en sortie du réacteur). Inversement à VTS/H2, le solide est excédentaire en Si à basse température, stoechiométrique à température moyenne et, dans certains cas, riche en C à haute température (R est alors limitée par les TM).La composition, la structure et la texture des dépôts de SiC issus des deux systèmes ont été corrélées avec les conditions expérimentales (température, pression, temps de séjour). Sur substrat dense, les transitions observées traduisent des variations de concentration des précurseurs effectifs de C et de Si, voire des changements du mécanisme de dépôt. L’aptitude à l’infiltration (CVI) a également été jugée à l’aide de substrats poreux modèles. Les profils d’épaisseurs et de composition le long d’un canal traduisent l’effet du temps de diffusion des gaz le long du pore, mais aussi l’appauvrissement local en réactifs. Comparativement au système MTS/H2 testé dans des conditions de référence, des conditions plus favorables à la CVI ont été établies pour les systèmes VTS/H2 et MS/Ar. / The aim of this thesis is to understand the gas-phase mechanisms that occured during the chemical vapor deposition of silicon carbide from two precursors with different chemistry (chlorinated precursor and non-chrloinated one). Thanks to this study, optimal conditions for infiltration with these two precursors should be found.

Hybrid trap experiments are set-ups that allow to study the interaction between ions and atoms in cold controlled environment. In such context, molecular anions present specific theoretical and experimental interests and challenges. In this work, we have used extensive \textit{ab initio} methods to investigate several collisional anionic systems: (1) M + OH$^{-}$ (where M are alkali or alkaline earth atoms), (2) Rb and H + OH(H$_{2}$O)$_{n}^{-}$ (with $n=0,1,2,3,4$) and (3) Rb and Li + C$_{2}^{-}$. Several molecular properties such as vertical detachment energies or electroaffinities, optimized structures, harmonic frequencies, potential energy curves or surfaces, etc have been calculated using high level quantum chemistry approaches. The results have been used to make predictions on the related reactivity in low energy regime. We emphasis on electronic detachment processes by carefully analysing the difference between the neutral and anionic potential energy surface. The Rb + OH$^{-}$ system is currently under experimental investigation. Therefore, a detailed study of its reactivity is carried out in the present work. We have analysed the different reactive channels arising from both collision involving the ground state and first electronic excited state of Rb. Using our calculated potentials and a capture model based dynamics, we have extracted cross sections and rate constants. Comparison with other alkali and earth alkaline atoms are made. Hydrated hydroxide cluster anions are planned by the experimental group as upcoming studied systems. We present here our preliminary results on the possible outcome when considering collisions with Rb and we discuss their implications for hybrid trap experiments. We make comparison with H as a colliding partner and consider our results in the context of astrochemistry. Finally we propose the C$_{2}^{-}$ molecular anion as an alternative to OH$^{-}$. Its interaction and reactivity with Rb and Li are investigated and the results are used to motivate our suggestion. Furthermore, for the Rb+OH$^{-}$ and Rb+C$_{2}^{-}$ system, we have also investigated the effect of a non-thermal collision energy distribution on the rate constants. At last, in light of the discussions related to each topic, general conclusions on the use of molecular anions in hybrid trap experiments are drawn. / Doctorat en Sciences / info:eu-repo/semantics/nonPublished