The aim of this thesis is to develop a deeper understanding and the technology in the nascent field of solid-state organic-inorganic perovskite solar cells. In recent years, perovskite materials have emerged as a low-cost, thin-film technology with efficiencies exceeding 16% challenging the quasi-paradigm that high efficiency photovoltaics must come at high costs. This thesis investigates perovskite solar cells in more detail with a focus on incorporating plasmonic nanostructures and perovskite film formation. Chapter 1 motivates the present work further followed by Chapter 2 which offers a brief background for solar cell fabrication and characterisation, perovskites in general, perovskite solar cells in specific, and plasmonics. Chapter 3 presents the field of plasmonics including simulation methods for various core-shell nanostructures such as gold-silica and silver-titania nanoparticles. The following Chapters 4 and 5 analyze plasmonic core-shell metal-dielectric nanoparticles embedded in perovskite solar cells. It is shown that using gold@silica or silver@titania NPs results in enhanced photocurrent and thus increased efficiency. After photoluminescence studies, this effect was attributed to an unexpected phenomenon in solar cells in which a lowered exciton binding energy generates a higher fraction of free charge. Embedding thermally unstable silver NPs required a low-temperature fabrication method which would not melt the Ag NPs. This work offers a new general direction for temperature sensitive elements. In Chapters 6 and 7, perovskite film formation is studied. Chapter 6 shows the existence of a previously unknown crystalline precursor state and an improved surface coverage by introducing a ramped annealing procedure. Based on this, Chapter 7 investigates different perovskite annealing protocols. The main finding was that an additional 130°C flash annealing step changed the film crystallinity dramatically and yielded a higher orientation of the perovskite crystals. The according solar cells showed an increased photocurrent attributed to a decrease in charge carrier recombination at the grain boundaries. Chapter 8 presents on-going work showing noteworthy first results for silica scaffolds, and layered, 2D perovskite structures for application in solar cells.

We study correlated two-level quantum impurity models coupled to a metallic conduction band in the hope of gaining insight into the physics of nanoscale quantum dot systems. We focus on the possibility of formation of a spin-1 impurity local moment which, on coupling to the band, generates an underscreened (USC) singular Fermi liquid state. By employing physical arguments and the numerical renormalization group (NRG) technique, we analyse such systems in detail examining in particular both the thermodynamic and dynamic properties, including the differential conductance. The quantum phase transitions occurring between the USC phase and a more ordinary Fermi liquid (FL) phase are analysed in detail. They are generically found to be of Kosterlitz-Thouless type; exceptions occur along lines of high symmetry where first-order transitions are found. A `Friedel-Luttinger sum rule' is derived and, together with a generalization of Luttinger's theorem to the USC phase, is used to obtain general results for the $T=0$ zero-bias conductance --- it is expressed solely in terms of the number of electrons present on the impurity and applicable in both the USC and FL phases. Relatedly, dynamical signatures of the quantum phase transition show two broad classes of behaviour corresponding to the collapse of either a resonance or antiresonance in the single-particle density of states. Evidence of both of these behaviours is seen in experimental devices. We study also the effect of a local magnetic field on both single- and two-level quantum impurities. In the former case we attempt to resolve some points of contention that remain in the literature. Specifically we show that the position of the maximum in the spin resolved density of states (and related peaks in the differential conductance) is not linear in the applied field, showing a more complicated form than a simple `Zeeman splitting'. The analytic result for the low-field asymptote is recovered. For two-level impurities we illustrate the manner in which the USC state is destroyed: due to two cancelling effects an abrupt change in the zero-bias conductance does not occur as one might expect. Comparison with experiment is made in both cases and used to interpret experimental findings in a manner contrary to previous suggestions. We find that experiments are very rarely in the limit of strong impurity-host coupling. Further, features in the differential conductance as a function of bias voltage should not be simply interpreted in terms of isolated quantum dot states. The many-body nature of such systems is crucially important to their observed properties.

Electrically detected magnetic resonance (EDMR) is a sensitive spectroscopic technique, which can be used to readout few to single electron spins in semiconductor and carbon nanodevices for applications in solid state quantum information processing (QIP). Since only electrically active defects contribute to the EDMR signal, this technique can be used further to investigate defects and impurities in photovoltaic devices, in which they limit the sunlight-to-energy conversion efficiency significantly. Here, I employ X-band EDMR for semiconductor defect analysis and identify the most important recombination centres in Czochralski silicon with oxide precipitates, which can be intentionally grown to confine detrimental metallic impurities to inactive regions of the wafer in order to serve as a defect-free substrate for modern silicon photovoltaic devices. Those experiments show that oxide precipitation is accompanied by the formation of silicon dangling bonds. Furthermore, I describe a very promising route towards the fabrication and readout of few to single electron spins in carbon nanotube devices, which can be characterised structurally via transmission electron microscopy in order to relate their electrical and spin properties with their structure. Finally, I employ EDMR to read out electron spin states in donor-doped silicon field-effect transistors as a prerequisite for their application in QIP. I report on a novel cryogenic probe head for EDMR experiments in resonant microwave cavities operating at 0.35 T (9.7 GHz, X-band) and 3.34 T (94 GHz, W-band). This approach overcomes the inherent limitations of conventional X-band EDMR and permits the investigation of paramagnetic states with a higher spectroscopic resolution and signal intensity. Both advantages are demonstrated and discussed. I further report on a novel mechanism giving rise to the EDMR effect in donor-doped silicon field-effect transistors, which is capable of explaining why the EDMR signal intensities of the conduction electrons are enhanced by a factor of ∼100, while the donor resonance signals increase by a factor of ∼20 from X- to W-band only. The spin-relaxation and dephasing times are extracted from a series of pulsed-EDMR measurements and confirm this model. The author gratefully acknowledges funding from Trinity College Oxford, Department of Materials, EPSRC DTA, and Konrad-Adenauer-Stiftung e.V. (Begabtenförderung).

Amorphous carbon film with Diamond like properties is the subject of intense interest in the past one and half decade. The unusual properties of these diamond like carbon films arise from the preponderance of SP3 tetrahedral bonding of carbon in the film. Depending on the processing technique and the processing conditions used, the structure of the films can range from amorphous carbon to large grain polycrystalline diamond. These deposited amorphous carbon films, which are smooth, may find their use in optoelectronics, in dielectric films and in microelectronics. These films are found to be chemically inhomogeneous(containing SP3 hybridized carbon in a matrix of SP2 hybridized non-graphitic carbon). There is a possibility of using these films as substrates in microelectronics, provided the deposited films are structurally smooth, are chemically homogeneous and are dopable with both types of impurities. A host of other advantages of using diamond like carbon as a substrate material in microelectronics made it a topic of interest to many investigators. This prompted the author to take up investigations on diamond like carbon films from the point of examining the electrical properties of these films and on the possibility of conceiving devices based on these films. This investigation dealt with, sputter deposition of diamond like carbon films and their electrical characteristics in MCS device structures. In this, emphasis is given to the importance of processing parameters involved and the effect of each parameter on the electrical and structural properties of the film. Various substrate treatments were done prior to sputtering and found that the DLC nature of the film exists in all the films but differ from one another in electrical resistivity, in nucleation density and in their adherence to the substrate. Films deposited on substrates treated with low vapour pressure oil resulted in compressive strain in the film and lead to very poor adhesion. The nucleation density increased when the substrates are pretreated with ultrasonic agitation in hard SiC grit. The substrate temperature had a direct impact on the resistivity of the film: resistivity decreases with increase in substrate temperature. The constituents of the plasma modified the structural properties of the film, e.g. the Hydrogen content in the plasma has resulted in increasing the SP3 hybridization content of the film, by acting as SP2- SP2 network terminator. Ultra violet light focused onto the substrate, in general, enhanced the deposition rate. Inclusion of Nitrogen in the plasma substantially increased the conductivity of the material and this is used in doping of the DLC film. The carbon films deposited on silicon are used for electrical characterisation. Deposition of metal electrode on the carbon film lead to the basic (MCS) device structure. The I vs.V characteristics of the MCS structure resemble those of junction diodes. From the I vs.V characteristics at different temperatures, it has been found that the reverse current goes through a maximum, drops back to certain level and once again increases with gradual increase in temperature. This behaviour of the structure with A1 as well as Ag as top electrode materials is explained by the heterojunction formed at the C-pSi interface. The initial increase in the reverse current is dominated by the drift of minority carriers across the depletion width at the reverse biased junction. With increase in temperature, the depletion width reduces to a minimum above a certain temperature, where the diffusion of carriers controls the current across the device. From the constructed energy-band diagram of heterojunction, it is shown that the change in the transport phenomena from drift of minority carriers to diffusion of majority carriers at the junction, introduces a barrier at the critical temperature; This is responsible for the drop in current at the critical temperature. This explains the anomaly of drop in reverse current with increase in temperature. The C vs. v characteristics showed a bell shaped behaviour indicating the presence of two junctions connected back to back. This confirms the type of contact formed at the metal-carbon interface and the type of conductivity of the film, concluding that A1 makes a Schottky contact where as Ag makes an ohmic contact and the deposited film behaves like n-type material. The C vs. V behaviour with temperature is explained by the two types of contacts in the case of Al-GpSi, i.e. Schottky contact at Al-C; and heterojunction at C-pSi interface. These C vs. V and I vs.V changes with temperature are in tune with each other and the model proposed takes care of all the characteristics observed. In case of Ag-GpSi, C vs. V with temperature shows junction like behaviour at elevated temperatures and are explained by the presence of the interface at C-pSi. It has been observed that in some of the carbon films, when an electric field of the order of l06 V/cm is applied, the reflectance of the Aluminium metal dot is increased by 5 times, coupled with a 50 to 100 times increase in the associated capacitance of the MCS structure. The increase in reflectance is explained by considering the film to be inhomogeneous with a matrix of varying dielectric constants (SP3 hybridized carbon in a medium of SP2 bonded carbon). The transformed film, is homogeneous and enhances the reflectance of the Aluminium dot. This is termed as "homogeneity induced smoothness." The transformation of inhomogeneous material to homogeneous material is further confirmed by the Raman spectroscopy, in which the broad peak is converted to a sharp peak changing the FWHM from 93 cm-1 to 4 cm-1 ; denoting the structural order in the film. To the best of our knowledge, this is the first investigation reporting the crystalline nature of the DLC, with structural order and the corresponding FWHM of the Raman peak as low as 4 cm-1. The preparational conditions of the film to get this transformation and the influence of various process parameters are examined. Devices based on Metal-Carbon-Oxide- Silicon (MCOS) structure are realized by thermally grown oxide/sputter deposited oxide on silicon, prior to carbon deposition. These structures showed voltage controlled negative resistance(VCNR) characteristics. The applied voltage and its distribution across the reverse biased junction and across the oxide gives rise to a negative resistance region. With the number of V vs. I characteristics measured, it is observed that the negative resistance region also shifts. This is attributed to the trapped charges in the carbon changing the distribution of applied voltage. This is explained by modifying the energy-band diagram. A concept of the accumalated charges at the oxide barrier filling up the higher energy states in the carbon and silicon, to become hot carriers is used. As long a. more voltage is dropped across the oxide, these hot carriers can surmount the barrier at the reverse biased junction. The flow of these carriers is cut off when the additional voltage is dropped across the reverse biased junction leading to a drop in the current. A further increase in the applied voltage nominally increases the current due to increase in the leakage current. A new hybrid (electrical/optical) read only memory (ROM) element is conceived and the way in which the information can be written and read is discussed. A two terminal negative resistance device using MCOS structure is fabricated and tested for its VCNR property. An analog memory device is proposed using the MCOS structure as gate in an FET. The work reported in this thesis has been divided into nine chapters. The introductory remarks on the importance of the area of research and about the work reported in this thesis are given in chapter one. Chapter two deals with some of the basic concepts related to understand the reported work. In chapter three the research work done by other investigators covering different aspects of this work is reported and some of their investigations are reviewed. Chapter four dealt with the various preparative techniques to deposit films, their structural characterisation, and the experimental work carried out to electrically characterize these films. Chapter five presents the I vs.V & C vs. V analysis and a model to qualitatively explain them. In chapter six field induced transformation phenomena of some of these films and its impact on the reflectance of the metal dot is dealt. Chapter seven consists of the MCOS device structure, its I vs.V characteristics and a model to explain the behaviour. Chapter eight presents the application part of same of the phenomena observed in conceiving a new hybrid ROM element and a two terminal negative resistance device. The concluding ninth chapter itemizes the important results of the work and suggestions to carry forward this work which can open up new vistas in the diamond like carbon film based technology and its applications in microelectronics.

Die Dunkelfeld off-axis Elektronenholographie (DFH) im Transmissionselektronenmikroskop ist eine nanoskalige Interferometriemethode, die es erlaubt, eine ausgewählte Beugungswelle eines Kristalls aufzuzeichnen und anschließend als zweidimensionale Amplituden- und Phasenverteilung zu rekonstruieren. Da sich aus dem Gradientenfeld der Phasenverteilung geometrische Verzerrungen des Kristallgitters bestimmen lassen, ermöglicht die DFH, Deformationsfelder in Kristallen zu vermessen. Damit eröffnen sich der Halbleiterindustrie vielversprechende Analysemöglichkeiten von lokalen mechanischen Verspannungen in Halbleiterkristallen insbesondere im Kanalbereich von Transistoren. Dabei verspricht die DFH eine höhere Ortsauflösung als rasternde, auf Elektronenbeugung mit möglichst fein fokussierten Elektronensonden basierende Methoden wie Nanobeugung. Jedoch steht die DFH als Analysemethode für mechanische Verspannungen bisher noch nicht standardmäßig zur Verfügung. Forschungs- und Entwicklungsbedarf besteht insbesondere hinsichtlich der Anpassung der Methodik auf kompliziertere Halbleiterstrukturen. Am Beispiel des Elementargitters wird demonstriert, wie einerseits die Gitterverzerrung die Phase der Beugungswelle moduliert, und wie andererseits aus dem Gradient der Phase diese Deformation wieder rekonstruiert werden kann. Zusätzlich wird die Modulation der Beugungswelle mit Hilfe eines erst kürzlich veröffentlichten analytischen Modells für den Zweistrahlfall erläutert. Spezielle Anpassungen der DFH im TEM erlauben, die geometrische Phase entweder mit 3...5 nm Lateralauflösung bei 200 nm breitem Gesichtsfeld oder mit 8...10 nm Lateralauflösung bei 800 nm breitem Gesichtsfeld aufzuzeichnen. Da die Deformationskarte durch numerische Ableitung der geometrischen Phase bestimmt wird, hängt die Signalauflösung der Deformationsmessung direkt von der Signalqualität in der rekonstruierten geometrischen Phase ab. Da die Ableitung das Rauschen verstärkt, werden verschiedene Strategien zur Rauschminderung und Signalverbesserung untersucht, u.a. werden Methoden zur Rauschfilterung eines DF-Hologramms oder zur Glättung der Deformationskarte vorgestellt. Durch Rekonstruktion einer gemittelten geometrischen Phase aus einer Dunkelfeldhologrammserie lassen sich Deformationen E mit einer Messabweichung von lediglich Delta_E=+/-0,05% bestimmen. Bei Aufzeichnung und Rekonstruktion der geometrischen Phase treten eine Reihe von Artefakten auf, die durch Fresnelsche Beugungssäume, defekte Detektorpixel sowie Verzeichnungen durch Projektivlinsen und Detektoroptik hervorgerufen werden. Da sie die Bestimmung der Deformationskarte erschweren, werden geeignete Methoden zur Vermeidung oder Korrektur vorgestellt. Die Präparation von TEM-Lamellen mit fokussiertem Ionenstrahl (FIB) verursacht Schädigungen der Probenoberfläche. Durch Vergleiche von DFH-Messungen mit Finite-Elemente-Simulationen wird gezeigt, dass die auf Oberflächenrelaxation zurückzuführenden Abweichungen vom simulierten Deformationszustand bei 120...160 nm Lamellendicke bis zu 10% betragen können. Präparationsbedingte lokale Dickenvariationen (Curtaining) können zu ähnlich großen Abweichungen führen. Anwendbarkeit und Funktionalität der DFH werden an modernen Halbleiterstrukturen untersucht. Die Vermessung einer verspannten SiGe-Schicht auf Si-Substrat zeigt eine sehr gute Übereinstimmung mit einem analytischen Modell. Die Abweichung beträgt ca. 10% und kann durch Oberflächenrelaxation an der SiGe/Si-Grenzfläche erklärt werden. Mittels SiGe an Source und Drain verspannte Transistoren dienen als Testobjekte für einen Vergleich von DFH und Nanobeugung. Beide Methoden liefern identische Ergebnisse. Der Vorteil der DFH besteht jedoch darin, das Deformationsfeld vollständig in Form einer zweidimensionalen Karte abzubilden, anstatt wie die Nanobeugung lediglich einzelne Profilschnitte zu messen. Die Deformationsmessung an SOI-Strukturen wird durch die leicht unterschiedliche Kristallorientierung (Miscut) zwischen SOI und Si-Substrat, das als Referenzbereich dient, erschwert. Die Deformationswerte im SOI zeigen ein Offset von 0,2% Dehnung gegenüber dem Si-Substrat. Der Miscut zwischen SOI und Si-Substrat kann zu 0,3°bestimmt werden. Für Transistoren mit tensiler Deckschicht gelingt es, Dehnungen von +0,3% in perfekter Übereinstimmung mit FE-Simulationen zu messen. Bei Transistoren, bei denen gleichzeitig eine kompressive Deckschicht und SiGe an Source und Drain eingesetzt werden, gelingt es mittels DFH, Stauchungen von -(0,1+/-0,05)% im Transistorkanal 5 nm unterhalb des Gateoxids nachzuweisen. / Dark-field off-axis electron holography (DFH) in a transmission electron microscope is based on the interference of a diffracted wave emanating from adjacent strained and unstrained sample areas to form a dark-field hologram, from which the phase of the diffracted wave can be reconstructed. Since the gradient of the phase parallel to the diffraction vector yields the lattice strain in this direction, a two-dimensional strain map can be derived. Therefore, DFH is considered to be a promising technique for strain metrology by semiconductor industry, especially for local strain measurements in the transistor channel. In particular, DFH offers better lateral resolution than scanning TEM-techniques based on electron diffraction with small focused electron probe like nano-beam diffraction. However, DFH is not yet available as a standard technique for strain metrology. Research is still needed to apply the method to complex devices. Using the example of a strained cosine lattice the phase modulation due to lattice distortions is discussed. In addition, modulation of the diffracted wave is approximated in two-beam diffraction condition. Adjustments of DFH in the TEM provide strain measurements with 3...5 nm lateral resolution at 200 nm field of view or 8...10 nm lateral resolution at 800 nm field of view. During recording and reconstruction of dark-field holograms several artifacts appear, for instance Fresnel diffraction, defective detector pixels, distortions of projective lenses or detector optics. Since they limit strain evaluation, suitable methods to either avoid or correct these artifacts are discussed. Sample preparation with focused ion beam (FIB) causes surface damage. Comparing DFH results with finite-element simulations reveals a deviation of 10% between simulation and experiment at 120...160 nm sample thickness due to surface relaxation. FIB-induced thickness variations (curtaining) lead to comparable deviations. Applicability of DFH for strain metrology is analyzed on several modern device structures. Strain measurements of SiGe-layers on Si-substrate correspond quite well with an analytic model. A residual deviation of 10% can be explained by surface relaxation close to the SiGe/Si-interface. Transistors strained by SiGe-source/drain serve as test objects for a comparison of DFH with nano-beam diffraction. Though both techniques reveal identical results, DFH is able to map the complete two-dimensional strain field, whereas nano-beam diffraction can only provide single line-scans. Strain mapping in silicon-on-insulator (SOI) is limited by the different crystal orientation (miscut) between the SOI layer and the Si-substrate, which serves as reference. Strain values in the SOI show an off-set of 0.2% in comparison to the unstrained Si-substrate. The miscut between SOI and Si-substrate is estimated to 0.3°. In transistor devices with tensile stress overlayers DFH is able to measure +0.3% tensile strain in excellent agreement with finite-element simulations. In devices with compressive overlayers and SiGe-source/drain a strain value of only -(0.1+/-0.05)% can be determined in the transistor channel 5nm beneath the gate oxide.

Made available in DSpace on 2016-06-02T20:15:30Z (GMT). No. of bitstreams: 1 5738.pdf: 3104025 bytes, checksum: 27f8126e91dc4b23ddd37a2e733a23fa (MD5) Previous issue date: 2014-01-29 / Universidade Federal de Sao Carlos / In the present thesis, we realize a computational modeling of semiconductor structures based on multiple quantum wells with filter barriers and on quantum wells with semiconductor diluted magnetic layers. We numerically solve the time-dependent Schrödinger s equation within the effective mass approximation, using the Split Operator method. Through the time evolved wave functions we access the dynamics quantities as the light assisted couplings of the states, in which the light is described by the inclusion of an oscillating electric field in the Hamiltonian. Then we determine the probabilities of absorption, oscillator strengths of the intersubband transitions induced by the light. Moreover we analyze the transmission probabilities and, in special, the system s photocurrent. The eigenstates and the eigenfunctions of the stationary states are also obtained within the method by simply making an imaginary time evolution. In the first work, the photocurrent of a multiple quantum well structure with filter barriers modulating the continuum above the wells was analyzed as a function of the applied bias. We find out an interesting dependence of the photocurrent with the applied field, as a differential negative photoconductance controlled by the field. We attribute this negative conductance to the interaction between the localized and extended states in the continuum, expressed by anticrossings between these states and the enhancement of the photocurrent at the crossings by the Landau-Zener-Stückelberg-Majorama like transitions. In the second work, it was evaluated the spin polarized photocurrent arising from quantum well s structures of GaMnAs, under light, electric and magnetic fields of few teslas. The study shows the existence of spectral domains in the THz ranges for which the proposed structure is strongly spin selective. For such photon frequencies, the photocurrent is spin polarized and the application of the external electric field reverts the polarization s signal. This behavior suggests the possibility of conveniently simple switching mechanisms. The physics underlying these results is studied and understood in terms of the spin-dependent coupling strengths emerging from the particular potential profiles of the heterostructures. We present two additional works related to the main ones. In the first additional one, we evaluated the dark current of the multiple quantum well structure with and without filter barriers. For doing this, we add totally the transmission probability through the structure in the Levine s model for the dark current. We observe that dark current is considerably reduced for the structure with the filter barriers when compared to the structure without these barriers. In the second additional work, we calculate the photocurrent in a ZnMnSe structure. We observe the generation of a spin polarized photocurrent controlled by the external electric field, as in the case of the GaMnAs structures. / Na presente tese, realizamos a modelagem computacional de estruturas semicondutoras baseadas em poços quânticos múltiplos com barreiras de filtro e em poços quânticos com camadas de material semicondutor magnético diluído. Para tanto, resolvemos numericamente a equação de Schrödinger dependente do tempo na aproximação de massa efetiva, por meio da evolução temporal das funções de onda do sistema, utilizando o chamado método do Split- Operator. Com as funções de onda evoluídas no tempo temos acesso às variáveis dinâmicas do sistema, como os acoplamentos entre os estados pela presença de luz, descrita na forma de um campo elétrico oscilante. Determinamos assim as probabilidades de absorção, forças de oscilador das transições intersubbandas geradas por essa excitação com luz, as probabilidades de transmissão através da estrutura e, em especial, o espectro de fotocorrente proveniente desses sistemas semicondutores. As autofunções e as autoenergias dos estados estacionários dos sistemas são obtidas pelo mesmo método realizando a evolução em tempo imaginário. No primeiro trabalho, a fotocorrente da estrutura de poços quânticos múltiplos com barreiras de filtro foi analisada em função do campo elétrico aplicado à estrutura. Foi encontrada uma dependência da fotocorrente com o campo elétrico bastante interessante, na forma de uma fotocondutância negativa controlada pelo campo elétrico aplicado à heteroestrutura. Atribuímos essa condutância negativa à interação entre estados localizados e estendidos no continuo se manifestando na forma de anticrossings e o aumento da fotocorrente para os valores de campo elétrico nos quais ocorrem esses crossings foi associado a transições de dois níveis do tipo Landau-Zener-Stückelberg-Majorama. No segundo trabalho, foi calculada a fotocorrente polarizada em spin de estruturas de poços quânticos de GaMnAs, na presença de um campo elétrico varável e um campo magnético de poucos teslas. O estudo mostrou a existência de domínios espectrais na região de THz do espectro eletromagnético, para os quais as estruturas propostas são altamente seletivas em spin. Para tais frequências, encontramos que a fotocorrente é polarizada em spin e a aplicação do campo elétrico é capaz de reverter forma muito eficiente o sinal da polarização. O comportamento observado sugere a possibilidade de mecanismos simples de controle sobre a fotocorrente e a Física por trás de tais efeitos foi entendida em termos dos acoplamentos dependentes de spin dos estados da estrutura, emergentes do perfil de potencial particular das heteroestruturas. Apresentamos dois trabalhos adicionais diretamente relacionados aos trabalhos principais. No primeiro trabalho, calculamos a corrente de escuro proveniente da estrutura de poços quânticos múltiplos com e sem barreiras de filtro, adicionando de forma integral a probabilidade de transmissão através da estrutura no modelo de Levine que determina essa corrente. Observamos que a presença das barreiras de filtro diminui significativamente a corrente de escuro dessa estrutura no regime de altos valores de campo elétrico. No segundo trabalho adicional, foi calculada a fotocorrente de uma estrutura de PQ com camada DMS, composta por ZnMnSe. Observamos a possibilidade de controle da polarização de spin com o campo elétrico, assim como no caso da estrutura composta de GaMnAs.

Fundação de Amparo a Pesquisa do Estado de Minas Gerais / The search for solutions to the different power quality problems, in special the ones related to the voltage level compliance to the standard values has produced, up to now, an extensive range of commercial products using different techniques. Despite this recognition, the challenge for alternative strategies, presenting low cost and low maintenance properties, are very attractive and this subject is still motivating research worldwide. In this context, this dissertation is focused on the proposal of a voltage regulating device, here referred as CET (Electromagnetic Voltage Compensator), which presents an innovative design to obtain the process of voltage regulation. The main idea is based on the voltage injection, been as an additive or subtractive way, through a physical arrangement, presenting as a full electromagnetic power device. Aiming operative dynamic properties, the proposal in question, as far as their operational design concerns, is based on electronic switching and control device. The proposal outlined here will encounter emphasizes to the CET physical structure, the control unit composition and the evaluation of the electric complex performance under different voltage deviations occurrences. The studies are presented and then discussed, using the results provided by the computational simulation carried out in the time domain simulator ATP throughout its classical feature the MODELS language. / A busca por soluções para os distintos problemas da qualidade da energia elétrica, com destaque às variações das tensões de suprimento, conta, na atualidade, com uma extensa gama de produtos visando, sobretudo, a regulação dinâmica da tensão de suprimento. Não obstante tal reconhecimento, os desafios por estratégias alternativas por compensadores com propriedades operacionais e econômicas mais atrativas continuam motivando pesquisas em todo o mundo. Neste contexto, a presente dissertação encontra-se focada numa proposta de controle de um dispositivo regulador de tensão, aqui denominado por CET (Compensador Eletromagnético de Tensão), que se apresenta com uma concepção inovadora no que tange ao processo da regulação da tensão. A ideia central apoia-se na injeção série de reforços de tensão, aditiva ou subtrativa, através de um arranjo físico totalmente eletromagnético quanto às suas unidades de potência. Objetivando propriedades operativas dinâmicas, a proposta em pauta, no que tange a sua concepção operativa, fundamenta-se em chaveamentos eletrônicos controlados. À luz destes fatos, o trabalho vai de encontro a esta filosofia e ressalta a estrutura física do dispositivo, a composição da unidade de controle e avalia o desempenho do dispositivo sob condições associadas com a ocorrência de desvios de tensão. Quanto aos estudos feitos e discutidos, estes se encontram alicerçados em investigações computacionais conduzidas através da plataforma ATP e recursos oferecidos pela linguagem MODELS. / Mestre em Ciências

There is a compelling need to explore different material options as well as device structures to facilitate smooth transistor scaling for higher speed, higher density and lower power. The enormous potential of nanoelectronics, and nanotechnology in general, offers us the possibility of designing devices with added functionality. However, at the same time, the new materials come with their own challenges that need to be overcome. In this work, we have addressed some of these challenges in the context of quasi-2D Silicon, III-V semiconductor and graphene. Bulk Si is the most widely used semiconductor with an indirect bandgap of about 1.1 eV. However, when Si is thinned down to sub-10nm regime, the quasi-2D nature of the system changes the electronic properties of the material significantly due to the strong geometrical confinement. Using a tight-binding study, we show that in addition to the increase in bandgap due to quantization, it is possible to transform the original in direct bandgap to a direct one. The effective masses at different valleys are also shown to vary uniquely in an anisotropic way. This ultra-thin Si, when used as a channel in a double gate MOSFET structure, creates so called “volume in version” which is extensively investigated in this work. It has been found that the both the quantum confinement as well as the gating effect play a significant role in determining the spatial distribution of the charge, which in turn has an important role in the characteristics of transistor. Compound III-V semiconductors, like Inx Ga1-xAs, provide low effective mass and low density of states. This, when coupled with strong confinement in a nanowire channel transistor, leads to the “Ultimate Quantum Capacitance Limit” (UQCL) regime of operation, where only the lowest subband is occupied. In this regime, the channel capacitance is much smaller than the oxide capacitance and hence dominates in the total gate capacitance. It is found that the gate capacitance change qualitatively in the UQCL regime, allowing multi-peak, non-monotonic capacitance-voltage characteristics. It is also shown that in an ideal condition, UQCL provides improved current saturation, on-off ratio and energy-delay product, but a degraded intrinsic gate delay. UQCL shows better immunity towards series resistance effect due to increased channel resistance, but is more prone to interfacial traps. A careful design can provide a better on-off ratio at a given gate delay in UQCL compared to conventional MOSFET scenario. To achieve the full advantages of both FinFET and HEMT in III-V domain, a hybrid structure, called “HFinFET” is proposed which provides excellent on performance like HEMT with good gate control like FinFET. During on state, the carriers in the channel are provided using a delta-doped layer(like HEMT) from the top of a fin-like non-planar channel, and during off state, the gates along the side of the fin(like FinFET) help to pull-off the carriers from the channel. Using an effective mass based coupled Poisson-Schrodinger simulation, the proposed structure is found to outperform the state of the art planar and non-planar MOSFETs. By careful optimization of the gate to source-drain underlap, it is shown that the design window of the device can be increased to meet ITRS projections at similar gate length. In addition, the performance degradation of HFinFET in presence of interface traps has been found to be significantly mitigated by tuning the underlap parameter. Graphene is a popular 2D hexagonal carbon crystal with extraordinary electronic, mechani-cal and chemical properties. However, the zero band gap of grapheme has limited its application in digital electronics. One could create a bandgap in grapheme by making quasi-1D strips, called nanoribbon. However, the bandgap of these nanoribbons depends on the the type of the edge, depending on which, one can obtain either semiconducting or metallic nanoribbon. It has been shown that by the application of an external transverse field along the sides of a nanoribbon, one could not only modulate the magnitude of the bandgap, but also change it from direct to indirect. This could open up interesting possibilities for novel electronic and optoelectronic applications. The asymmetric potential distribution inside the nanoribbon is found to result in such direct to indirect bandgap transition. The corresponding carrier masses are also found to be modulated by the external field, following a transition from a“slow”electron to a“fast” electron and vice-versa. Experimentally, it is difficult to control the bandgap in nanoribbons as precise edge control at nanometer scale is nontrivial. One could also open a bandgap in a bilayer graphene, by the application of vertical electric field, which has raised a lot of interest for digital applications. Using a self-consistent tight binding theory, it is found that, inspite of this bandgap opening, the intrinsic bias dependent electronic structure and the screening effect limit the subthreshold slope of a metal source drain bilayer grapheme transistor at a relatively higher value-much above the Boltzmann limit. This in turn reduces the on-off ratio of the transistor significantly. To overcome this poor on-off ratio problem, a semiconductor source-drain structure has been proposed, where the minority carrier injection from the drain is largely switched off due to the bandgap of the drain. Using a self-consistent Non-Equilibrium Green’s Function(NEGF) approach, the proposed device is found to be extremely promising providing unipolar grapheme devices with large on-off ratio, improved subthreshold slope and better current saturation. At high drain bias, the transport properties of grapheme is extremely intriguing with a number of nontrivial effects. Optical phonons in monolayer grapheme couple with carriers in a much stronger way as compared to a bilayer due to selection rules. However, it is difficult to experimentally probe this through transport measurements in substrate supported grapheme as the surface polar phonons with typical low activation energy dominates the total scattering. However, at large drain field, the carriers obtain sufficient energy to interact with the optical phonons, and create so called ‘hot phonons’ which we have experimentally found to result in a negative differential conductance(NDC). The magnitude of this NDC is found to be much stronger in monolayer than in bilayer, which agrees with theoretical calculations. This NDC has also been shown to be compensated by extra minority carrier injection from drain at large bias resulting in an excellent current saturation through a fundamentally different mechanism as compared to velocity saturation. A transport model has been proposed based on the theory, and the experimental observations are found to be in agreement with the model.

Die Dunkelfeld off-axis Elektronenholographie (DFH) im Transmissionselektronenmikroskop ist eine nanoskalige Interferometriemethode, die es erlaubt, eine ausgewählte Beugungswelle eines Kristalls aufzuzeichnen und anschließend als zweidimensionale Amplituden- und Phasenverteilung zu rekonstruieren. Da sich aus dem Gradientenfeld der Phasenverteilung geometrische Verzerrungen des Kristallgitters bestimmen lassen, ermöglicht die DFH, Deformationsfelder in Kristallen zu vermessen. Damit eröffnen sich der Halbleiterindustrie vielversprechende Analysemöglichkeiten von lokalen mechanischen Verspannungen in Halbleiterkristallen insbesondere im Kanalbereich von Transistoren. Dabei verspricht die DFH eine höhere Ortsauflösung als rasternde, auf Elektronenbeugung mit möglichst fein fokussierten Elektronensonden basierende Methoden wie Nanobeugung. Jedoch steht die DFH als Analysemethode für mechanische Verspannungen bisher noch nicht standardmäßig zur Verfügung. Forschungs- und Entwicklungsbedarf besteht insbesondere hinsichtlich der Anpassung der Methodik auf kompliziertere Halbleiterstrukturen. Am Beispiel des Elementargitters wird demonstriert, wie einerseits die Gitterverzerrung die Phase der Beugungswelle moduliert, und wie andererseits aus dem Gradient der Phase diese Deformation wieder rekonstruiert werden kann. Zusätzlich wird die Modulation der Beugungswelle mit Hilfe eines erst kürzlich veröffentlichten analytischen Modells für den Zweistrahlfall erläutert. Spezielle Anpassungen der DFH im TEM erlauben, die geometrische Phase entweder mit 3...5 nm Lateralauflösung bei 200 nm breitem Gesichtsfeld oder mit 8...10 nm Lateralauflösung bei 800 nm breitem Gesichtsfeld aufzuzeichnen. Da die Deformationskarte durch numerische Ableitung der geometrischen Phase bestimmt wird, hängt die Signalauflösung der Deformationsmessung direkt von der Signalqualität in der rekonstruierten geometrischen Phase ab. Da die Ableitung das Rauschen verstärkt, werden verschiedene Strategien zur Rauschminderung und Signalverbesserung untersucht, u.a. werden Methoden zur Rauschfilterung eines DF-Hologramms oder zur Glättung der Deformationskarte vorgestellt. Durch Rekonstruktion einer gemittelten geometrischen Phase aus einer Dunkelfeldhologrammserie lassen sich Deformationen E mit einer Messabweichung von lediglich Delta_E=+/-0,05% bestimmen. Bei Aufzeichnung und Rekonstruktion der geometrischen Phase treten eine Reihe von Artefakten auf, die durch Fresnelsche Beugungssäume, defekte Detektorpixel sowie Verzeichnungen durch Projektivlinsen und Detektoroptik hervorgerufen werden. Da sie die Bestimmung der Deformationskarte erschweren, werden geeignete Methoden zur Vermeidung oder Korrektur vorgestellt. Die Präparation von TEM-Lamellen mit fokussiertem Ionenstrahl (FIB) verursacht Schädigungen der Probenoberfläche. Durch Vergleiche von DFH-Messungen mit Finite-Elemente-Simulationen wird gezeigt, dass die auf Oberflächenrelaxation zurückzuführenden Abweichungen vom simulierten Deformationszustand bei 120...160 nm Lamellendicke bis zu 10% betragen können. Präparationsbedingte lokale Dickenvariationen (Curtaining) können zu ähnlich großen Abweichungen führen. Anwendbarkeit und Funktionalität der DFH werden an modernen Halbleiterstrukturen untersucht. Die Vermessung einer verspannten SiGe-Schicht auf Si-Substrat zeigt eine sehr gute Übereinstimmung mit einem analytischen Modell. Die Abweichung beträgt ca. 10% und kann durch Oberflächenrelaxation an der SiGe/Si-Grenzfläche erklärt werden. Mittels SiGe an Source und Drain verspannte Transistoren dienen als Testobjekte für einen Vergleich von DFH und Nanobeugung. Beide Methoden liefern identische Ergebnisse. Der Vorteil der DFH besteht jedoch darin, das Deformationsfeld vollständig in Form einer zweidimensionalen Karte abzubilden, anstatt wie die Nanobeugung lediglich einzelne Profilschnitte zu messen. Die Deformationsmessung an SOI-Strukturen wird durch die leicht unterschiedliche Kristallorientierung (Miscut) zwischen SOI und Si-Substrat, das als Referenzbereich dient, erschwert. Die Deformationswerte im SOI zeigen ein Offset von 0,2% Dehnung gegenüber dem Si-Substrat. Der Miscut zwischen SOI und Si-Substrat kann zu 0,3°bestimmt werden. Für Transistoren mit tensiler Deckschicht gelingt es, Dehnungen von +0,3% in perfekter Übereinstimmung mit FE-Simulationen zu messen. Bei Transistoren, bei denen gleichzeitig eine kompressive Deckschicht und SiGe an Source und Drain eingesetzt werden, gelingt es mittels DFH, Stauchungen von -(0,1+/-0,05)% im Transistorkanal 5 nm unterhalb des Gateoxids nachzuweisen.:1 Einleitung 2 Grundlagen der Elastizitätstheorie 2.1 Der Verzerrungstensor 2.2 Der Spannungstensor 2.3 Das Hooke’sche Gesetz 2.4 Zusammenfassung 3 Mechanisch verspannte Transistoren 3.1 Der MOSFET 3.2 Techniken zur Spannungserzeugung 3.2.1 SiGe- und Si:C-Source/Drain-Gebiete 3.2.2 Verspannte Deckschichten 3.3 Mechanische Verspannung und Ladungsträgerbeweglichkeit 3.4 Zusammenfassung 4 Beugungswelle und geometrische Phase 4.1 Transmissionselektronenmikroskopie 4.1.1 Aufbau eines Transmissionselektronenmikroskops 4.1.2 Hellfeld- und Dunkelfeldabbildung 4.2 Beugung am Kristallgitter 4.2.1 Bragg- und Laue-Beugungsbedingung 4.2.2 Ewaldkugel 4.2.3 Beugungswelle 4.3 Geometrische Phase 4.3.1 Geometrische Phase in kinematischer Näherung 4.3.2 Veranschaulichung der geometrischen Phase am Elementargitter 4.3.3 Grenzen der geometrische Phase 4.3.4 Geometrische Phase bei dynamischer Streuung 4.3.4.1 Streuung im deformierten Kristall 4.3.4.2 Zweistrahlfall im deformierten Kristall 4.3.4.3 Analytische Lösung für z-unabhängige Verschiebung 4.3.4.4 Näherungslösung für z-abhängige Verschiebung 4.3.4.5 Konsequenzen für die Deformationsmessung 4.4 Zusammenfassung 5 Spezialverfahren der Dunkelfeld off-axis Elektronenholographie 5.1 Aufnahme von Dunkelfeldhologrammen 5.1.1 Voraussetzungen 5.1.2 Versuchsaufbau 5.1.3 Rekonstruktion der Beugungswelle 5.2 Bestimmung der Gitterdeformation 5.2.1 Gitterdeformation in g_ref-Richtung 5.2.2 Gitterdeformation in (x,y)-Ebene 5.3 Optimierung des Tecnai F20 Mikroskops für die Dunkelfeldholographie 5.3.1 Anforderungen 5.3.2 Limitierungen durch experimentellen Aufbau 5.3.3 Zusätzliche Freiheitsgrade mit Cs-Korrektor und Pseudo-Lorentz Linse 5.3.4 Verbleibende Limitierungen 5.3.4.1 Begrenzte Beleuchtungskippung 5.3.4.2 Defokussierte Blende in der hinteren Brennebene 5.4 Aufbereitung und Rekonstruktion von Dunkelfeldhologrammen 5.4.1 Beseitigen fehlerhafter Pixel 5.4.2 Entfernen der Fresnelschen Beugungssäume 5.4.3 Wahl der Rekonstruktionsmaske 5.4.4 Filterung der Hologrammintensität mit Wiener-Filter 5.5 Einfluss und Korrektur von Verzeichnungen 5.5.1 Verzeichnungskorrektur mittels Leerwelle 5.5.2 Verzeichnungskorrektur mittels Verzeichnungskarte 5.5.3 Vergleich der Korrekturmethoden 5.6 Vorzeichen der Beugungswelle 5.7 Numerische Ableitung der Phase und Rauschen 5.8 Kalibrierung von Phasen- und Deformationskarte 5.9 Glättung der Dehnungskarte 5.10 Aufzeichnung und Rekonstruktion einer Dunkelfeldhologrammserie 5.11 Maximierung der Intensität in der Beugungswelle 5.11.1 Zweistrahlfall und gekippte Dunkelfeldbeleuchtung 5.11.2 Optimale Probendicke 5.12 Einfluss der Objektkippung an Grenzflächen 5.13 Präparationseinflüsse 5.13.1 Curtaining 5.13.2 Relaxation in FIB-Lamellen 5.13.3 Amorphe Oberflächen 5.13.4 Verbiegung von FIB-Lamellen 5.14 Zusammenfassung 6 Verspannungsmessungen an aktuellen Halbleiterstrukturen 6.1 Gitterdeformation in SiGe-Schicht auf Si-Substrat 6.2 Mit SiGe verspannte Transistoren auf Bulk-Silizium 6.2.1 Transistorstrukturen mit SiGe-S-Source/Drain-Gebieten 6.2.2 Vergleich von Dunkelfeldholographie und Nanobeugung 6.3 Mit SiGe verspannte Transistoren auf Silicon-on-Insulator (SOI) 6.4 Transistorstrukturen mit verspannten Deckschichten 6.4.1 Erste Experimente 6.4.2 Mittels Wolframschicht verspannte Teststruktur 6.4.3 Mittels TPEN-Schicht verspannter n-MOSFET 6.4.4 Mittels CPEN-Schicht und SiGe verspannter p-MOSFET 6.5 Zusammenfassung 7 Zusammenfassung / Dark-field off-axis electron holography (DFH) in a transmission electron microscope is based on the interference of a diffracted wave emanating from adjacent strained and unstrained sample areas to form a dark-field hologram, from which the phase of the diffracted wave can be reconstructed. Since the gradient of the phase parallel to the diffraction vector yields the lattice strain in this direction, a two-dimensional strain map can be derived. Therefore, DFH is considered to be a promising technique for strain metrology by semiconductor industry, especially for local strain measurements in the transistor channel. In particular, DFH offers better lateral resolution than scanning TEM-techniques based on electron diffraction with small focused electron probe like nano-beam diffraction. However, DFH is not yet available as a standard technique for strain metrology. Research is still needed to apply the method to complex devices. Using the example of a strained cosine lattice the phase modulation due to lattice distortions is discussed. In addition, modulation of the diffracted wave is approximated in two-beam diffraction condition. Adjustments of DFH in the TEM provide strain measurements with 3...5 nm lateral resolution at 200 nm field of view or 8...10 nm lateral resolution at 800 nm field of view. During recording and reconstruction of dark-field holograms several artifacts appear, for instance Fresnel diffraction, defective detector pixels, distortions of projective lenses or detector optics. Since they limit strain evaluation, suitable methods to either avoid or correct these artifacts are discussed. Sample preparation with focused ion beam (FIB) causes surface damage. Comparing DFH results with finite-element simulations reveals a deviation of 10% between simulation and experiment at 120...160 nm sample thickness due to surface relaxation. FIB-induced thickness variations (curtaining) lead to comparable deviations. Applicability of DFH for strain metrology is analyzed on several modern device structures. Strain measurements of SiGe-layers on Si-substrate correspond quite well with an analytic model. A residual deviation of 10% can be explained by surface relaxation close to the SiGe/Si-interface. Transistors strained by SiGe-source/drain serve as test objects for a comparison of DFH with nano-beam diffraction. Though both techniques reveal identical results, DFH is able to map the complete two-dimensional strain field, whereas nano-beam diffraction can only provide single line-scans. Strain mapping in silicon-on-insulator (SOI) is limited by the different crystal orientation (miscut) between the SOI layer and the Si-substrate, which serves as reference. Strain values in the SOI show an off-set of 0.2% in comparison to the unstrained Si-substrate. The miscut between SOI and Si-substrate is estimated to 0.3°. In transistor devices with tensile stress overlayers DFH is able to measure +0.3% tensile strain in excellent agreement with finite-element simulations. In devices with compressive overlayers and SiGe-source/drain a strain value of only -(0.1+/-0.05)% can be determined in the transistor channel 5nm beneath the gate oxide.:1 Einleitung 2 Grundlagen der Elastizitätstheorie 2.1 Der Verzerrungstensor 2.2 Der Spannungstensor 2.3 Das Hooke’sche Gesetz 2.4 Zusammenfassung 3 Mechanisch verspannte Transistoren 3.1 Der MOSFET 3.2 Techniken zur Spannungserzeugung 3.2.1 SiGe- und Si:C-Source/Drain-Gebiete 3.2.2 Verspannte Deckschichten 3.3 Mechanische Verspannung und Ladungsträgerbeweglichkeit 3.4 Zusammenfassung 4 Beugungswelle und geometrische Phase 4.1 Transmissionselektronenmikroskopie 4.1.1 Aufbau eines Transmissionselektronenmikroskops 4.1.2 Hellfeld- und Dunkelfeldabbildung 4.2 Beugung am Kristallgitter 4.2.1 Bragg- und Laue-Beugungsbedingung 4.2.2 Ewaldkugel 4.2.3 Beugungswelle 4.3 Geometrische Phase 4.3.1 Geometrische Phase in kinematischer Näherung 4.3.2 Veranschaulichung der geometrischen Phase am Elementargitter 4.3.3 Grenzen der geometrische Phase 4.3.4 Geometrische Phase bei dynamischer Streuung 4.3.4.1 Streuung im deformierten Kristall 4.3.4.2 Zweistrahlfall im deformierten Kristall 4.3.4.3 Analytische Lösung für z-unabhängige Verschiebung 4.3.4.4 Näherungslösung für z-abhängige Verschiebung 4.3.4.5 Konsequenzen für die Deformationsmessung 4.4 Zusammenfassung 5 Spezialverfahren der Dunkelfeld off-axis Elektronenholographie 5.1 Aufnahme von Dunkelfeldhologrammen 5.1.1 Voraussetzungen 5.1.2 Versuchsaufbau 5.1.3 Rekonstruktion der Beugungswelle 5.2 Bestimmung der Gitterdeformation 5.2.1 Gitterdeformation in g_ref-Richtung 5.2.2 Gitterdeformation in (x,y)-Ebene 5.3 Optimierung des Tecnai F20 Mikroskops für die Dunkelfeldholographie 5.3.1 Anforderungen 5.3.2 Limitierungen durch experimentellen Aufbau 5.3.3 Zusätzliche Freiheitsgrade mit Cs-Korrektor und Pseudo-Lorentz Linse 5.3.4 Verbleibende Limitierungen 5.3.4.1 Begrenzte Beleuchtungskippung 5.3.4.2 Defokussierte Blende in der hinteren Brennebene 5.4 Aufbereitung und Rekonstruktion von Dunkelfeldhologrammen 5.4.1 Beseitigen fehlerhafter Pixel 5.4.2 Entfernen der Fresnelschen Beugungssäume 5.4.3 Wahl der Rekonstruktionsmaske 5.4.4 Filterung der Hologrammintensität mit Wiener-Filter 5.5 Einfluss und Korrektur von Verzeichnungen 5.5.1 Verzeichnungskorrektur mittels Leerwelle 5.5.2 Verzeichnungskorrektur mittels Verzeichnungskarte 5.5.3 Vergleich der Korrekturmethoden 5.6 Vorzeichen der Beugungswelle 5.7 Numerische Ableitung der Phase und Rauschen 5.8 Kalibrierung von Phasen- und Deformationskarte 5.9 Glättung der Dehnungskarte 5.10 Aufzeichnung und Rekonstruktion einer Dunkelfeldhologrammserie 5.11 Maximierung der Intensität in der Beugungswelle 5.11.1 Zweistrahlfall und gekippte Dunkelfeldbeleuchtung 5.11.2 Optimale Probendicke 5.12 Einfluss der Objektkippung an Grenzflächen 5.13 Präparationseinflüsse 5.13.1 Curtaining 5.13.2 Relaxation in FIB-Lamellen 5.13.3 Amorphe Oberflächen 5.13.4 Verbiegung von FIB-Lamellen 5.14 Zusammenfassung 6 Verspannungsmessungen an aktuellen Halbleiterstrukturen 6.1 Gitterdeformation in SiGe-Schicht auf Si-Substrat 6.2 Mit SiGe verspannte Transistoren auf Bulk-Silizium 6.2.1 Transistorstrukturen mit SiGe-S-Source/Drain-Gebieten 6.2.2 Vergleich von Dunkelfeldholographie und Nanobeugung 6.3 Mit SiGe verspannte Transistoren auf Silicon-on-Insulator (SOI) 6.4 Transistorstrukturen mit verspannten Deckschichten 6.4.1 Erste Experimente 6.4.2 Mittels Wolframschicht verspannte Teststruktur 6.4.3 Mittels TPEN-Schicht verspannter n-MOSFET 6.4.4 Mittels CPEN-Schicht und SiGe verspannter p-MOSFET 6.5 Zusammenfassung 7 Zusammenfassung

<div>Power semiconductor devices play an important role in many areas, including household</div><div>appliances, electric vehicles, high speed trains, electric power stations, and renewable energy</div><div>conversion. In the modern era, silicon based devices have dominated the semiconductor</div><div>market, including power electronics, because of their low cost and high performance. The</div><div>applications of devices rated 600 V - 6.5 kV are still dominated by silicon devices, but they</div><div>are nearly reaching fundamental material limits. New wide band gap materials such as silicon</div><div>carbide (SiC) offer significant performance improvements due to superior material properties</div><div>for such applications in and beyond this voltage range. 4H-SiC is a strong candidate</div><div>among other wide band gap materials because of its high critical electric field, high thermal</div><div>conductivity, compatibility with silicon processing techniques, and the availability of high</div><div>quality conductive substrates.</div><div>Vertical DMOSFETs and insulated gate bipolar transistors (IGBT) are key devices for</div><div>high voltage applications. High blocking voltages require thick drift regions with very light</div><div>doping, leading to specific on-resistance (R_ON,SP ) that increases with the square of blocking</div><div>voltage (V_BR). In theory, superjunction drift regions could provide a solution because of a</div><div>linear dependence of R_ON,SP on V_BR when charge balance between the pillars is achieved</div><div>through extremely tight process control. In this thesis, we have concluded that superjunction</div><div>devices inevitably have at least some level of charge imbalance which leads to a quadratic</div><div>relationship between V_BR and R_ON,SP . We then proposed an optimization methodology to</div><div>achieve improved performance in the presence of this inevitable imbalance.</div><div>On the other hand, an IGBT combines the benefits of a conductivity modulated drift</div><div>region for significantly reduced specific on-resistance with the voltage controlled input of a</div><div>MOSFET. Silicon carbide n-channel IGBTs would have lower conduction losses than equivalent</div><div>DMOSFETs beyond 6.5 kV, but traditionally have not been feasible below 15 kV. This</div><div>is due to the fact that the n+ substrate must be removed to access the p+ collector of the</div><div>IGBT, and devices below 15 kV have drift layers too thin to be mechanically self-supporting.</div><div>In this thesis, we have demonstrated the world’s first functional 10 kV class n-IGBT with</div><div>a waffle substrate through simulation, process development, fabrication and characterization.</div><div><div>The waffle substrate would provide the required mechanical support for this class of devices.</div><div>The fabricated IGBT has exhibited a differential R_ON,SP of 160 mohm</div><div>.cm², less than half of</div><div>what would be expected without conductivity modulation. An extensive fabrication process</div><div>development for integrating a waffle substrate into an active IGBT structure is described</div><div>in this thesis. This process enables an entirely new class of moderate voltage SiC IGBTs,</div><div>opening up new applications for SiC power devices.</div></div>