L’hydrogénation sélective d’hydrocarbures polyinsaturés présente un fort intérêt pour l’industrie pétrolière. Cette réaction est catalysée par des particules métalliques ou des alliages et a lieu sous pression d’hydrogène. Dans ce travail nous étudions la réaction d’hydrogénation sélective du butadiène en 1-butene en présence des catalyseurs Pt(111) et Sn/Pt-Pt(111). Cette étude a été menée à l’aide du programme VASP (Vienna Abinitio Simulation Package) qui permet de simuler des systèmes solides et surfaciques dans le cadre de la théorie de la fonctionnelle de la densité (DFT). Le choix des catalyseurs a été dicté par les communautés expérimentale et industrielle qui utilisent le plus souvent le platine car très actif, ou des alliages de platine-étain, moins actifs mais beaucoup plus sélectifs. L’hydrogénation du butadiène a déjà été étudiée par le passé mais uniquement à T=0 K et sans prendre en compte les conditions réelles de pression des réactifs. Notre but étant de comprendre les pas élémentaires de cette réaction, nous avons choisi de l’étudier à des conditions de T et de P proches de celles utilisées expérimentalement, c’est-à-dire 300-400 K et 1-10 bar. A cette fin, nous avons mis en place un modèle thermodynamique pour évaluer, dans un premier temps, la composition de la surface lors de la réaction. Il est alors apparu que la configuration de surface la plus stable correspond à une couverture de une monocouche d’hydrogène ce qui suggère un mécanisme de type Eley-Rideal. Puis nous avons étudié l’aspect cinétique ce cette réaction et calculé les chemins réactionnels pour différents recouvrements en hydrogène. Nous avons conclu qu’il existe une forte compétitivité entre le mécanisme pressenti, appelé Langmuir-Hinshelwood et impliquant des espèces fortement adsorbées, et le mécanisme Eley-Rideal proposant une adsorption faible de l’un des réactifs. Les calculs ont été effectués avec les fonctionnelles PBE et optPBE pour apporter un point de comparaison des méthodes. / Petroleum industry has a strong interest in the selective hydrogenation of polyunsaturated hydrocarbons. This reaction is catalyzed by metallic particles or alloys and happens under pressure of hydrogen. In this work, we study the selective hydrogenation of butadiene into 1-butene, on two model catalysts which are Pt(111) and Sn/Pt-Pt(111). For this, we used the VASP code (Vienna Abinitio Simulation Package) that allows to perform periodic calculations in the framework of the Density Functional Theory (DFT). The choice of the catalysts was driven by the experimental and industrial communities who mostly use platinum because of its high activity, or alloys such as tin-platinum alloy, less active but more selective. Butadiene hydrogenation was already modeled in the past but only at T=0 K and without taking into account the real pressure conditions of the reactants. Our aim is to understand the impact of the reaction conditions which is why we ran this study at T and P conditions close to the one used experimentally, e.g. 300-400 K et 1-10 bar. For this, we setup a thermodynamic model to evaluate in a first step the surface composition when the reaction occurs. It came out that the most stable surface configuration corresponds to a coverage of 1 ML of hydrogen which suggests an Eley-Rideal type mechanism. Then we studied the kinetic aspect of this reaction and we calculated the hydrogenation pathways for different coverages of hydrogen. We concluded that there is a strong competition between the sensed mechanism, called Langmuir-Hinshelwood mechanism and implying strongly adsorbed species, and the Eley-Rideal mechanism, proposing a weak adsorption of one of the two reactants.

Bimetallic Rh-V system was studied by means of surface science experimental methods. Properties of ultra-thin Rh-V layers supported by γ-Al2O3 were compared with model systems prepared by vacuum V deposition on Rh(111), Rh(110) and polycrystalline rhodium. Formation of ordered V- Rh(111)-(2×2), V-Rh(110)-(2×1) and V-Rh(110)-(1×2) subsurface alloys and their electronic and atomic structure were investigated and models of the surface reconstructions were proposed. Influence of the subsurface alloy formation on interaction with CO and O2 molecules as well as the influence of the molecule adsorption on this alloy was investigated. The bond between CO molecules and Rh-V alloy surface was weakened due to pronounced changes in surface valence band structure. Oxygen which adsorbed on the alloy surface reacted with the subsurface vanadium at elevated temperature and blocked the interaction of the metal substrate with CO molecules.

Abstract In this thesis oxidation of a pure and alloyed Cu surface was studied using density functional theory based calculations. The mechanism and energetics behind experimentally observed missing row reconstruction on a pure Cu(100) surface were studied. Examination of the formation of the missing row was approached by studying the profitableness of vacancy formation to the surface at oxygen coverages below and equal to 0.5 ML. However, an ideal surface was in all cases more favourable in energy than vacancy included structures. Therefore an additional Cu atom was added to the simulation cell to promote vacancy formation. A consequence of this on-surface Cu addition was the identifying of a new energetically favoured Cu-O-Cu chain formation. This newly identified structure is a possible transition state in the path from an ideal surface to a reconstructed one. Copper(I)oxide formation through sub-surface adsorption was examined on a reconstructed Cu(100) surface. At low O coverage on-surface adsorption was found to be much more favoured than adsorption on sub-surface sites. However, when coverage increased to 0.75 ML structures with both on-surface and sub-surface adsorbates possessed the lowest energy. Furthermore, the diffusion barriers to the sub-surface adsorption sites were low. Both these facts support the conclusion that the transition from the reconstructed missing row structure to the copper(I)oxide takes place without any large energy effort. The adsorption induced segregation of copper on a Ag/Cu(100) surface was studied by calculating the surface energies of slabs where the silver layer was positioned at different depths with varying O coverage. In addition, the segregation energies were also calculated with two Ag compositions. It was discovered that without oxygen adsorbates silver atoms lying on top of the surface were favoured but the surface segregation probability of Cu increases with increasing O coverage. The same calculation procedure was expanded to study the segregation probability of five additional alloy metals with a Cu surface. Without oxygen present Mg, Al and Zn possess a weak tendency to segregate to the surface, however, V and Cr prefer to stay in the bulk. The addition of oxygen adsorbates on the surface causes the segregation of all the dopants to the top of the surface. With the Al dopant the effect of the surface orientation and the oxygen coverage were also studied. The bonds between adsorbates and the surface were strongest with the densest surface, (111), and weakest with the most opened surface, (110). Furthermore, the enrichment probability was largest in the case of the (111) surface and lowest with the (110) surface. Probability increased with increasing oxygen coverage. / Tiivistelmä Tässä väitöstyössä on tutkittu puhtaan ja seostetun kuparipinnan hapettumista tiheysfunktionaaliteoriaan perustuvilla laskuilla. Kokeellisesti Cu(100) pinnan on havaittu rekonstruoituvan hapen vaikutuksesta. Rekonstruoitumisen aikana ideaalisesta pinnasta irtoaa joka neljäs pintakuparirivi ja muodostuu niin kutsuttu missing row- rakenne. Tämän rakenteen muodostumismekanismia ja muodostumiseen kuluvaa energiaa tutkittiin vertaamalla pintavakansillisten rakenteiden pintaenergioita vastaaviin ideaalipinnan arvoihin vaihtelevalla happipeitolla. Koska ideaalipinta oli kaikilla peittoasteilla energeettisesti suotuisampi kuin vakanssirakenne, yksittäisiä kupariatomeja lisättiin pinnalle stabiloimaan vakanssirakenteita. Kuparilisäyksen seurauksena löydettiin uusi energeettisesti edullinen Cu-O-Cu ketjurakenne. Tämä rakenne on mahdollinen ideaalisen ja rekonstruoituneen pinnan välinen siirtymätila. Kupari(I)oksidin muodostumista pinnanalaisen happiadsorption kautta tutkittiin rekonstruoituneella Cu(100) pinnalla. Pienillä happi-peittoasteilla pinnanpäällisen adsorption havaittiin olevan huomattavasti suotuisampaa kuin pinnanalaisen adsorption. Kuitenkin happi peiton kasvaessa 0,75 ML:iin rakenteet, joissa oli adsorbaatteja sekä pinnan alla että päällä, tulivat muita suotuisimmiksi. Myös lasketut energiavallit hapen diffuusiolle pinnalta pinnan alle olivat hyvin pieniä. Kaikki tulokset viittaavat siihen, että rekonstruoitunut Cu(100) pinta hapettuu helposti kupari(I)oksidiksi. Happiadsorption aikaansaamaa kuparin rikastumista (segregaatio) hopeakupariseosmetallin pinnalle tutkittiin laskemalla vaihtelevilla happipeitoilla pintaenergiat kuparipinnoille, joissa ensimmäinen, toinen ja kolmas kuparikerros oli korvattu hopealla. Myös segregaatioenergiat laskettiin kahdella eri seoskoostumuksella. Saatujen tulosten mukaan ilman happiadsoptiota hopea segregoituu pinnalle, kun taas hapen peittoasteen kasvaessa kupari rikastuu pinnalle. Samaa tutkimustapaa käytettiin myös tarkasteltaessa viiden muun seosmetallin segregaatiotodennäköisyyttä kuparipinnalla. Ilman pintahappea magnesiumilla, alumiinilla ja sinkillä oli heikko taipumus rikastua pinnalle, kun taas vanadium ja kromi pysyivät syvemmällä metallissa. Sen sijaan happea lisättäessä kaikki tutkitut seosmetallit pyrkivät siirtymään kohti pintaa. Myös hapen peiton sekä pintaorientaation vaikutusta segregaatioon tutkittiin hopeakuparipinnalla. Sekä happiadsorbaattien ja pinnan välinen sidos että segregaatiotodennäköisyys oli sitä voimakkaampaa, mitä tiheämmästä pinnasta oli kyse. Segregaatiotodennäköisyys kasvoi pinta-adsorption lisääntyessä.