<p>The heteroepitaxial growth of InAs self-assembled quantum wires on vicinal substrates is investigated. InGaAlAs lattice-matched to InP was first deposited onto an InP(001) substrate with and without a 0.9 degree off-cut toward the (110) direction, followed by the deposition of a strained layer of InAs. Dense InAs quantum wires were successfully grown on both nominally flat and vicinal substrates in order to observe the effect of the presence of atomic steps. The off-cut angle was chosen based on the wire spacing on a flat substrate to serve as a template for their nucleation and improve their size distribution for use as 1.55 um wavelength lasers required by the telecommunications industry. The results have shown a modest but statistically significant improvement in the width of their size distribution. In addition, a kinetic Monte Carlo simulation including full strain calculations was developed to further understand the nucleation process. The model developed here disproves the idea that InAs quantum wires are aligned towards the (-110) direction due to diffusion anisotropy. The simulation of the formation of quantum wires similar to those observed experimentally has been achieved and the Stranski-Krastanow growth mode is demonstrated.</p> / Master of Materials Science and Engineering (MMatSE)

Atomlagenabscheidung (ALD, Atomic Layer Deposition) ist als präzise Technik zur Abscheidung dünner Schichten bekannt. Mittels wechselweisen Einleitens von Precursorgasen in einen Reaktor erzeugt der Prozess auch auf strukturierten Substraten gleichmäßige dünne Schichten. Durch die selbstsättigende Natur der zu Grunde liegenden Reaktionen sind sowohl die Wachstumsrate als auch die Zusammensetzung wohldefiniert, weshalb sich Atomlagenabscheidung beispielsweise zur Herstellung nanoskopischer Bauelemente im Bereich der Mikroelektronik eignet. Obwohl Aluminiumoxid vermehrt Aufmerksamkeit für seine hohe Bandlücke (~9 eV) sowie die relativ hohe Dielektrizitätskonstante (k ~ 9) geerntet hat, ist oftmals trotz vielseitiger Untersuchungen der anwendbaren Precursorpaare nur wenig über die strukturellen Eigenschaften sowie die Wachstumskriterien der resultierenden Schichten bekannt. In dieser Arbeit wurde eine kombinierte Simulationsmethode entwickelt, mit der sich Atomlagenabscheidung mittels elementarer Reaktionen auf beliebig strukturierten Substraten simulieren lässt. Molekulardynamische Berechnungen ermöglichen dabei atomare Genauigkeit, wohingegen die Ankunft der individuellen Precursoratome durch Kinetic Monte Carlo-Methoden dargestellt werden. Diese Aufteilung erlaubt die Kopplung der molekulardynamischen Präzision mit den Größenordnungen einer KMC-Simulation, welche prinzipiell die Betrachtung von Milliarden von Atomen zulässt. Durch asynchrone Parallelisierung mit bis zu tausenden Arbeiterprozessen wird zudem die Effizienz gegenüber einer herkömmlichen Molekulardynamiksimulation ausreichend erhöht, um binnen weniger Stunden mehrere Abscheidungszyklen nahezu unabhängig von der Größe des betrachteten Raumes, welche im Bereich von Quadratmikrometern liegen kann, zu simulieren. Zur abschließenden Validierung des Modells und seiner Implementierung werden einerseits Versuche einfacher Schichtwachstumsprozesse unternommen, andererseits wird die Atomlagenabscheidung des wohluntersuchten Precursorpaares Trimethylaluminium (TMA, Al(CH3)3) und Wasser simuliert und die resultierende Schicht auf Übereinstimmung mit bestehenden Daten geprüft.

Atomlagenabscheidung (ALD, Atomic Layer Deposition) ist als präzise Technik zur Abscheidung dünner Schichten bekannt. Mittels wechselweisen Einleitens von Precursorgasen in einen Reaktor erzeugt der Prozess auch auf strukturierten Substraten gleichmäßige dünne Schichten. Durch die selbstsättigende Natur der zu Grunde liegenden Reaktionen sind sowohl die Wachstumsrate als auch die Zusammensetzung wohldefiniert, weshalb sich Atomlagenabscheidung beispielsweise zur Herstellung nanoskopischer Bauelemente im Bereich der Mikroelektronik eignet. Obwohl Aluminiumoxid vermehrt Aufmerksamkeit für seine hohe Bandlücke (~9 eV) sowie die relativ hohe Dielektrizitätskonstante (k ~ 9) geerntet hat, ist oftmals trotz vielseitiger Untersuchungen der anwendbaren Precursorpaare nur wenig über die strukturellen Eigenschaften sowie die Wachstumskriterien der resultierenden Schichten bekannt. In dieser Arbeit wurde eine kombinierte Simulationsmethode entwickelt, mit der sich Atomlagenabscheidung mittels elementarer Reaktionen auf beliebig strukturierten Substraten simulieren lässt. Molekulardynamische Berechnungen ermöglichen dabei atomare Genauigkeit, wohingegen die Ankunft der individuellen Precursoratome durch Kinetic Monte Carlo-Methoden dargestellt werden. Diese Aufteilung erlaubt die Kopplung der molekulardynamischen Präzision mit den Größenordnungen einer KMC-Simulation, welche prinzipiell die Betrachtung von Milliarden von Atomen zulässt. Durch asynchrone Parallelisierung mit bis zu tausenden Arbeiterprozessen wird zudem die Effizienz gegenüber einer herkömmlichen Molekulardynamiksimulation ausreichend erhöht, um binnen weniger Stunden mehrere Abscheidungszyklen nahezu unabhängig von der Größe des betrachteten Raumes, welche im Bereich von Quadratmikrometern liegen kann, zu simulieren. Zur abschließenden Validierung des Modells und seiner Implementierung werden einerseits Versuche einfacher Schichtwachstumsprozesse unternommen, andererseits wird die Atomlagenabscheidung des wohluntersuchten Precursorpaares Trimethylaluminium (TMA, Al(CH3)3) und Wasser simuliert und die resultierende Schicht auf Übereinstimmung mit bestehenden Daten geprüft.:1 Einführung 1.1 Anwendungen von Atomlagenabscheidung 1.2 Aktueller Stand 1.2.1 Experimentelle Untersuchungen 1.2.2 Kinetic Monte Carlo-Simulationen von Dwivedi 1.2.3 Kinetic Monte Carlo-Simulationen von Mazaleyrat 1.2.4 Molekulardynamik-Simulationen 1.2.5 Dichtefunktionaltheoretische Rechnungen von Musgrave 1.3 Motivation 2 Grundlagen 2.1 Atomlagenabscheidung 2.1.1 Einführung zur Atomlagenabscheidung 2.1.2 ALD von Metalloxiden 2.1.3 ALD von Al2O3 2.2 Kinetic Monte Carlo Methoden 2.2.1 KMC-Formalismus 2.2.2 KMC-Algorithmen 2.3 Molekulardynamik 2.3.1 Grundlagen 2.3.2 Methoden zur Ensembledarstellung 2.3.3 Potentialarten 2.3.4 Numerische Optimierungen 3 Kombiniertes Modell 3.1 Verwendetes Kinetic Monte Carlo-Modell 3.2 Kombiniertes Modell 3.2.1 Abscheidungszyklus 3.2.2 Simulationsraum 3.2.3 Ereignisse 3.2.4 Parallelisierungsmethode 3.2.5 Abhängigkeitsgraph 4 Implementierung 4.1 Existierende Software 4.1.1 LAMMPS 4.1.2 SPPARKS 4.1.3 Sonstige Software 4.2 LibKMC 4.2.1 Modularisierung 4.2.2 Abhängigkeiten 4.3 Implementierung des kombinierten Modells 4.3.1 Vorstellung der Software 4.3.2 Einbindung von LibKMC 4.3.3 Einbindung von LAMMPS 4.3.4 Host-Worker-System 4.3.5 Substratgenerierung 5 Validierung 5.1 Validierung des kombinierten Modelles 5.1.1 Wachstumskriterium 5.1.2 Sättigungskriterium 5.1.3 Parallelisierungseffizienz 5.2 Untersuchungen von Al2O3 5.2.1 Potentialuntersuchungen 5.2.2 Schichtwachstumseigenschaften 5.2.3 Strukturanalyse 6 Zusammenfassung und Ausblick Literaturverzeichnis Danksagung

As the size of modern transistors is continuously scaled down, challenges rise in almost every component of a silicon device. Formation of ultra shallow junction (USJ) with high activation level is particularly important for suppressing short channel effects. However, the formation of low resistance USJ is made difficult by dopant Transient Enhanced Diffusion (TED) and clustering-induced deactivation. In this work, we proposed a novel point defect engineering solution to address the arsenic TED challenge. By overlapping arsenic doped region with silicon interstitials and vacancies, we observed enhanced and retarded arsenic diffusion upon anneal, respectively. We explain this phenomenon by arsenic interstitial diffusion mechanism. In addition, we implemented this interstitial-based mechanism into a kinetic Monte Carlo (kMC) simulator. The key role of interstitials in arsenic TED is confirmed. And we demonstrated that the simulator has an improved prediction capability for arsenic TED and deactivation. As a long time unsolved process challenge, arsenic segregation at SiO₂/Si interface was investigated using density functional theory (DFT) calculation. The segregation-induced arsenic dose loss not only increases resistance but also may induce interface states. We identified three arsenic complex configurations, [chemical formula] , [chemical formula] and [chemical formula], which are highly stabilized at SiO₂/Si interface due to the unique local bonding environments. Therefore, they could contribute to arsenic segregation as both initial stage precursors and dopant trapping sites. Our calculation indicates that arsenic atoms trapped in such interface complexes are electrically inactive. Finally, the formation and evolution dynamics of these interface arsenic-defect complexes are discussed and kMC models are constructed to describe the segregation effects. A potential problem for the p-type USJ formation is the recently found transient fast boron diffusion during solid phase epitaxial regrowth process. Using DFT calculations and molecular dynamics simulation, we identified an interstitial-based mechanism of fast boron diffusion in amorphous silicon. The activation energy for this diffusion mechanism is in good agreement with experimental results. In addition, this mechanism is consistent with the experimentally reported transient and concentration-dependent features of boron diffusion in amorphous silicon. / text

Inert gases in metals can occur due to ion implantation, from a plasma in a magnetron device or as a result of being by-products of nuclear reactions. Mainly because of the nuclear applications, the properties of the inert gases, helium, argon and xenon in the body centred cubic (bcc) iron crystal are examined theoretically using a combination of molecular dynamics, static energy minimisation and long time scale techniques using empirical potential functions. The same techniques are also used to investigate argon and xenon in aluminium. The primary interest of the work occurred because of He produced in nuclear fission and its effect on the structural materials of a fission reactor. This structure is modelled with perfectly crystalline bcc Fe. In bcc iron, helium is shown to diffuse rapidly forming small bubbles over picosecond time scales, which reach a certain optimum size. In the initial phase of He accumulation, Fe interstitials are ejected. This occurs instantaneously for bubbles containing 5 He atoms and as the more He accumulates, more Fe interstitials are ejected. The most energetically favourable He to vacancy ratios at 0 K, vary from 1 : 1 for 5 vacancies up to about 4 : 1 for larger numbers of vacancies. An existing He bubble can be enlarged by a nearby collision cascade through the ejection of Fe interstitials, allowing more He to be trapped. Ar and Xe in bcc Fe prefer to be substitutional rather than interstitial and there are large barriers to be overcome for the inert gas atoms to diffuse from a substitutional site. Bubbles that form can again be enlarged by the presence of a nearby collision cascade or at very high temperatures. In this case the most energetically favourable vacancy ratios in the bubbles is 1: 1 for Ar and from 0.6: 1 to 0.8: 1 for Xe. For Ar and Xe, bubble formation is more likely as a direct result of radiation or radiation enhanced diffusion rather than diffusion from a substitutional site. Ar in aluminium is also studied. Ar atoms in fcc Al prefer to be substitutional rather than interstitial and evolution into substitutional occurs over picosecond time scales at room temperature. Bubble formation can occur more easily than in bcc iron, mainly because the barriers for vacancy diffusion are much lower but the time scales for bubble accumulation are much longer than those for He. A vacancy assisted mechanism is found which allows Ar to diffuse through the lattice. Finally some preliminary results on the energetics of different geometrical structures of larger Xe bubbles in Al are investigated since experiment has indicated that these can become facetted.

La formation et la morphologie des boîtes quantiques est un sujet d'un grand intérêt, ces structures ayant de nombreuses application potentielles, en particulier en microélectronique et optoélectronique. Cette thèse porte sur l'étude théorique et numérique de la croissance et de la morphologie d'ilots par épitaxie par jet moléculaire. Un premier modèle de croissance est une étude non-linéaire de l'instabilité de type Asaro-Tiller-Grinfeld, il convient pour les systèmes à faibles désaccords de maille, et est plus spécifiquement appliquée au cas où le désaccord de maille est anisotrope (voir le cas du GaN sur AlGaN). Le calcul de l'instabilité que nous avons effectué prend en compte les effets élastiques causés par le désaccord de maille, les effets de mouillage et les effets d'évaporation. La résolution numérique de l'instabilité nous permet de constater une croissance plus rapide dans le cas anisotrope comparé au cas isotrope, ainsi que la croissance d'ilots fortement anisotropes.Le deuxième modèle est basé sur des simulations Monte Carlo cinétiques, qui permettent de décrire la nucléation d'ilots 3D. Ces simulations sont utilisées pour les systèmes à fort désaccord de maille, comme Ge sur Si. Nos simulations prennent en compte la diffusion des adatomes, les effets élastiques, et un terme simulant la présence de facettes (105). Des ilots pyramidaux se formaent, conformément aux expériences et subissent un mûrissement interrompu. Les résultats obtenus ont été comparés au cas de la nucléation 2D, et on retrouve en particulier une densité d'ilots en loi de puissance par rapport au rapport D/F du coefficient de diffusion et du flux de déposition. / The growth and morphology of quantum dots is currently a popular subject as these structures have numerous potential uses, specifically in microelectronics and optoelectronics. Control of the size, shape and distribution of these dots is of critical importance for the uses that are being considered. This thesis presents a theoretical and numerical study of the growth of islands during molecular beam epitaxy. In order to study these dots, we used two models : a nonlinear study of an Asaro-Tiller-Grinfeld like instability, and kinetic Monte Carlo simulations. The first model is appropriate for low misfit systems, and is detailed in the case where misfit is anisotropic (this is the case when depositing GaN on AlGaN). In this case we took into account elastic effects, wetting effects and evaporation. Numerical calculations show faster growth, compared to the isotropic misfit case, and the growth of strongly anisotropic islands.The second model is based on kinetic Monte Carlo simulations that can describe 3D island nucleation. We use these simulations to study systems with high misfits, specifically Ge on Si. Adatom diffusion on a surface is considered and takes into account elastic effects, and surface energy anisotropy, that allows us to stabilize (105) facets. Simulation results show the growth of pyramid-shaped 3D islands, as observed in experiments, and their ripening is interrupted. The results of these simulations are then compared to the case of 2D nucleation, and we find that several of the known 2D properties also apply to 3D islands. Specifically, island density depends on a power law of D/F, the diffusion coefficient divided by the deposition flux.

As propriedades eletrônicas e estruturais de defeitos em bulk e superfície de semicondutores são estudadas através de cálculos de primeiros princípios. Apresentamos um estudo detalhado para as relaxações e distorções para diferentes estados de carga da vacância em Ge. Nosso principal resultado é que a vacância em Ge não é um sistema de U-negativo, ao contrário do Si. Nós estudamos a superfície Si(00l) e comparamos imagens de STM para estados vazios geradas teoricamente com imagens de experiências recentes de STM de alta resolução. Apresentamos também uma análise das possíveis configurações dímero de Ge sobre Si(00l) e uma comparação de imagens de STH geradas teoricamente com imagens experimentais recentes de estágio inicial de crescimento de Ge sobre Si(00l) na temperatura ambiente. Nós estudamos de­ feitos complexos em GaAs. Em particular, nós mostramos que o par de antiestrutura AsGa + GaAs pode existir em duas configurações, nn e nnn. Nossos resultados indicam que apesar da primeira estrutura ser mais favorável energeticamente ela apresenta uma barreira de recombinação relativamente menor, e consequentemente sua concentração é muito menor que a da estrutura nnn em experiências de irradiação seguida por annealing. Por último, nós apresentamos resultados da implementação do método de Monte Carlo Cinético para o estudo da difusão de As em Si baseada em resultados de primeiros princípios para as barreiras de migração. O principal objetivo é relacionar processos microscópicos com propriedades de grande escala, isto é alcançado com um certo grau de sucesso. / The electronic and structural properties of defects in bulk and surface of semiconductor are studied through first principles method. We present a detailed study of the lattice and relaxation and distortions for different charge states for the vacancy in Ge. Our main result is that the vacancy is not a negative U system, contrary to Si. We studied the surface Si(001) and compared theoretically generated STM images for empty states with recent experimental high resolution STM images. We also presented an analysis of the possible configurations of Ge dimer on Si(001) and a comparison of the theoretically generated STM images with recent experimental ones for the initial stage of Ge growth on Si(OO1) at low temperature. We studied defect complexes in GaAs. In particular, we show that the anti-structure pair AsGa+GaAs can exist in two possible configurations, nn and nnn. Our results indicate that although the first. structure is more energetic favorable it presents a relatively low recombination barrier which implies that is concentration is much lower than the nnn anti-structure pair in irradiation experiments followed by annealing. Finally we present results for implementation of the Kinetic Monte Carlo method for studying the As diffusion in Si, based on first principles results for the migration barriers. The main purpose is to link microscopic processes to large scale properties, and this is achieved with a certain degree of success.

A presente tese trata da simulação computacional dos estágios iniciais da eletrodeposição de cobalto sobre carbono e de cobre sobre ouro, utilizando o método de Monte Carlo Cinético. Os fenômenos de adsorção, dessorção, deposição e difusão superficial, entre outros, ocorrem simultânea e independentemente durante a eletrodeposição, em função de certa probabilidade. Sorteios aleatórios dos eventos de acordo com essa distribuição de probabilidades simulam o comportamento real complexo de um sistema durante a eletrodeposição. No caso do cobalto, foi utilizada uma descrição teórica detalhada das interações entre os átomos no sistema, que permitiu que fossem observadas as formações dos primeiros núcleos, a predileção pela eletrodeposição nos defeitos superficiais e os distintos mecanismos dependentes do sobrepotencial elétrico aplicado. No caso do cobre, foi utilizada uma descrição empírica mais simples da interação entre os átomos, o que permitiu que a quantidade de espécies simuladas fosse maior e aspectos relacionados ao crescimento do filme metálico fossem obtidos, tais quais a formação de buracos durante o processo e a obtenção de parâmetros importantes, como tamanho, altura e densidade média dos núcleos. Alguns experimentos foram realizados para fundamentar e complementar o estudo teórico / This thesis reports the computer simulation of the first stages of the electrodeposition of cobalt onto carbon and copper onto gold, using the Kinetic Monte Carlo method. Adsorption, desorption, deposition and surface diffusion, among other events, occur simultaneously and independently during the electrodeposition, according to a given probability. Random choices of events as a result of a probability distribution can simulate the real evolution of a complex system during the electrodeposition. Regarding the cobalt system, a detailed theoretical interatomic description was used, which allowed the observance of the formation of the first nuclei, the preference of the electrodeposition on the surface defects and the deposition mechanisms differences depending on the applied overpotential. Regarding copper, a much simpler, empiric-based interatomic description was used, allowing the simulation of a bigger number of particles. Some features regarding the copper film growth (like the formation of voids during the process) and some important deposition parameters (like the average size, height and density of the nuclei) were obtained. Some experiments were held in order to base and complement the theoretical study

A presente tese trata da simulação computacional dos estágios iniciais da eletrodeposição de cobalto sobre carbono e de cobre sobre ouro, utilizando o método de Monte Carlo Cinético. Os fenômenos de adsorção, dessorção, deposição e difusão superficial, entre outros, ocorrem simultânea e independentemente durante a eletrodeposição, em função de certa probabilidade. Sorteios aleatórios dos eventos de acordo com essa distribuição de probabilidades simulam o comportamento real complexo de um sistema durante a eletrodeposição. No caso do cobalto, foi utilizada uma descrição teórica detalhada das interações entre os átomos no sistema, que permitiu que fossem observadas as formações dos primeiros núcleos, a predileção pela eletrodeposição nos defeitos superficiais e os distintos mecanismos dependentes do sobrepotencial elétrico aplicado. No caso do cobre, foi utilizada uma descrição empírica mais simples da interação entre os átomos, o que permitiu que a quantidade de espécies simuladas fosse maior e aspectos relacionados ao crescimento do filme metálico fossem obtidos, tais quais a formação de buracos durante o processo e a obtenção de parâmetros importantes, como tamanho, altura e densidade média dos núcleos. Alguns experimentos foram realizados para fundamentar e complementar o estudo teórico / This thesis reports the computer simulation of the first stages of the electrodeposition of cobalt onto carbon and copper onto gold, using the Kinetic Monte Carlo method. Adsorption, desorption, deposition and surface diffusion, among other events, occur simultaneously and independently during the electrodeposition, according to a given probability. Random choices of events as a result of a probability distribution can simulate the real evolution of a complex system during the electrodeposition. Regarding the cobalt system, a detailed theoretical interatomic description was used, which allowed the observance of the formation of the first nuclei, the preference of the electrodeposition on the surface defects and the deposition mechanisms differences depending on the applied overpotential. Regarding copper, a much simpler, empiric-based interatomic description was used, allowing the simulation of a bigger number of particles. Some features regarding the copper film growth (like the formation of voids during the process) and some important deposition parameters (like the average size, height and density of the nuclei) were obtained. Some experiments were held in order to base and complement the theoretical study

As propriedades eletrônicas e estruturais de defeitos em bulk e superfície de semicondutores são estudadas através de cálculos de primeiros princípios. Apresentamos um estudo detalhado para as relaxações e distorções para diferentes estados de carga da vacância em Ge. Nosso principal resultado é que a vacância em Ge não é um sistema de U-negativo, ao contrário do Si. Nós estudamos a superfície Si(00l) e comparamos imagens de STM para estados vazios geradas teoricamente com imagens de experiências recentes de STM de alta resolução. Apresentamos também uma análise das possíveis configurações dímero de Ge sobre Si(00l) e uma comparação de imagens de STH geradas teoricamente com imagens experimentais recentes de estágio inicial de crescimento de Ge sobre Si(00l) na temperatura ambiente. Nós estudamos de­ feitos complexos em GaAs. Em particular, nós mostramos que o par de antiestrutura AsGa + GaAs pode existir em duas configurações, nn e nnn. Nossos resultados indicam que apesar da primeira estrutura ser mais favorável energeticamente ela apresenta uma barreira de recombinação relativamente menor, e consequentemente sua concentração é muito menor que a da estrutura nnn em experiências de irradiação seguida por annealing. Por último, nós apresentamos resultados da implementação do método de Monte Carlo Cinético para o estudo da difusão de As em Si baseada em resultados de primeiros princípios para as barreiras de migração. O principal objetivo é relacionar processos microscópicos com propriedades de grande escala, isto é alcançado com um certo grau de sucesso. / The electronic and structural properties of defects in bulk and surface of semiconductor are studied through first principles method. We present a detailed study of the lattice and relaxation and distortions for different charge states for the vacancy in Ge. Our main result is that the vacancy is not a negative U system, contrary to Si. We studied the surface Si(001) and compared theoretically generated STM images for empty states with recent experimental high resolution STM images. We also presented an analysis of the possible configurations of Ge dimer on Si(001) and a comparison of the theoretically generated STM images with recent experimental ones for the initial stage of Ge growth on Si(OO1) at low temperature. We studied defect complexes in GaAs. In particular, we show that the anti-structure pair AsGa+GaAs can exist in two possible configurations, nn and nnn. Our results indicate that although the first. structure is more energetic favorable it presents a relatively low recombination barrier which implies that is concentration is much lower than the nnn anti-structure pair in irradiation experiments followed by annealing. Finally we present results for implementation of the Kinetic Monte Carlo method for studying the As diffusion in Si, based on first principles results for the migration barriers. The main purpose is to link microscopic processes to large scale properties, and this is achieved with a certain degree of success.