No description available.

Thema dieser Arbeit ist die Synthese von GaAs Nanodrähten mittels Molekularstrahlepitaxie. Dabei wird das Wachstum mittels Au- und jenes mittels selbst-induziertem VLS-Mechanismus verglichen. Die Au-induzierte Methode ist als vielseitiger Ansatz für die Herstellung von Nanodrähten bekannt. Darüberhinaus wird seit Neuerem der selbst-induzierte Mechanismus untersucht, bei dem Galliumtropfen die Rolle des Goldes übernehmen, um eine etwaige Verunreinigung mit Au von vornherein auszuschliessen. Mit beiden Wachstumsmethoden erzielen wir GaAs Nanodrähte mit großem Aspektverhältnis und epitaktischer Beziehung zum Si(111) Substrat. Während des Au-induzierten Wachstums entsteht eine parasitäre Schicht zwischen den Drähten, die mittels des selbst-induzierten Mechanismus vermieden werden. Alle GaAs Drähte sind vollständig relaxiert. Die durch die Gitterfehlanpassung (4,1\% zwischen GaAs und Si) verursachte Verspannung wird durch Versetzungen an der Grenzfläche abgebaut. Selbst-induzierte Drähten zeigen ausschließlich unpolare Seitenfacetten, während verschiedene polare Facetten für Au-induzierte Nanodrähte beschrieben werden. Mittels VLS-Nukleationstheorie könnne wir den Einfluss des Tropfenmaterials auf die Stabilität der verschiedenen Seitenfacetten erklären. Optoelektronische Anwendungen benötigen lange Minoritätsladungsträgerlebensdauern bei Raumtemperatur. Daher wurden mit (Al,Ga)As Hüllen ummantelte GaAs Nanodrähte mittels zeitaufgelöster PL vermessen. Das Ergebnis sind 2,5 ns für die selbst-induzierten aber nur 9 ps für die Au-induzierten Nanodrähte. Durch temperaturabhängige PL Messungen kann eine charakteristische Aktivierungsenergie von 77 meV nachgewiesen werden, die nur in den Au-induzierten Nanodrähten vorliegt. Dies suggeriert, dass sich Au aus den Tröpfchen in die GaAs Nanodrähte einbaut und dort als tiefes, nichtstrahlendes Rekombinationszentrum fungiert. / In this work the synthesis of GaAs nanowires by molecular beam epitaxy (MBE) using the vapour-liquid-solid (VLS) mechanism is investigated. A comparison between Au- and self-assisted VLS growth is at the centre of this thesis. While the Au-assisted method is established as a versatile tool for nanowire growth, the recently developed self-assisted variation results from the exchange of Au by Ga droplets and thus eliminates any possibility of Au incorporation. By both methods, we achieve nanowires with epitaxial alignment to the Si(111) substrates. Caused by differences during nanowire nucleation, a parasitic planar layer grows between the nanowires by the Au-assisted method, but can be avoided by the self-assisted method. Au-assisted nanowires grow predominantly in the metastable wurtzite crystal structure, while their self-assisted counterparts have the zincblende structure. All GaAs nanowires are fully relaxed and the strain arising from the lattice mismatch between GaAs and Si of 4.1\% is accommodated by misfit dislocations at the interface. Self-assisted GaAs nanowires are generally found to have vertical and non-polar side facets, while tilted and polar nanofacets were described for Au-assisted GaAs nanowires. We employ VLS nucleation theory to understand the effect of the droplet material on the lateral facets. Optoelectronic applications require long minority carrier lifetimes at room temperature. We fabricate GaAs/(Al,Ga)As core-shell nanowires and analyse them by transient photoluminescence (PL) spectroscopy. The results are 2.5 ns for the self-assisted nanowires as well as 9 ps for the Au-assisted nanowires. By temperature-dependent PL measurements we find a characteristic activation energy of 77 meV that is present only in the Au-assisted nanowires. We conclude that most likely Au is incorporated from the droplets into the GaAs nanowires and acts as a deep, non-radiative recombination centre.

At present, focus of the SiC crystal growth development is on improving the crystalline quality without polytype inclusions, micropipes and the occurrence of extended defects. The purity of the grown material, as well as intentional doping must be well controlled and the processes understood. High-quality substrates will significantly improve device performance and yield. One of the aims of the thesis is further understanding of polytype inclusion formation as well as impurity control in SiC bulk crystals grown using PVT method also termed seeded sublimation method. Carbonization of the source was identified as a major reason behind the polytype inclusion occurrence during the growth. The aim of this work was further understanding of sublimation growth process of 4H-SiC bulk crystals in vacuum, in absence of an inert gas. For comparison growth in argon atmosphere (at 5 mbar) was performed. The effect of the ambient on the impurity incorporation was studied for different growth temperatures. For better control of the process in vacuum, tantalum as a carbon getter was utilized. The focus of the SiC part of the thesis was put on further understanding of the PVT epitaxy with an emphasis on the high growth rate and purity of grown layers. High resistivity 4H-SiC samples grown by sublimation with high growth rate were studied. The measurements show resistivity values up to high 104 cm. By correlation between the growth conditions and SIMS results, a model was applied in which it is proposed that an isolated carbon vacancy donor-like level is a possible candidate responsible for compensation of the shallow acceptors in p-type 4H-SiC. A relation between cathodoluminescence (CL) and DLTS data is taken into account to support the model. To meet the requirements for high voltage blocking devices such as high voltage Schottky diodes and MOSFETs, 4H-SiC epitaxial layers have to exhibit low doping concentration in order to block reverse voltages up to few keV and at the same time have a low on-state resistance (Ron). High Ron leads to enhanced power consumption in the operation mode of the devices. In growth of thick layers for high voltage blocking devices, the conditions to achieve good on-state characteristics become more challenging due to the low doping and pronounced thicknesses needed, preferably in short growth periods. In case of high-speed epitaxy such as the sublimation, the need to apply higher growth temperature to yield the high growth rate, results in an increased concentration of background impurities in the layers as well as an influence on the intrinsic defects. On-state resistance Ron estimated from current density-voltage characteristics of Schottky diodes on thick sublimation layers exhibits variations from tens of mΩ.cm2 to tens of Ω.cm2 for different doping levels. In order to understand the occurrence of high on-state resistance, Schottky barrier heights were first estimated for both forward and reverse bias with the application of thermionic emission theory and were in agreement with literature reported values. Decrease in mobility with increasing temperature was observed and its dependencies of T–1.3 and T–2.0 for moderately doped and low doped samples, respectively, were estimated. From deep level measurements by Minority Carrier Transient Spectroscopy (MCTS), an influence of shallow boron related levels and D-center on the on-state resistance was observed, being more pronounced in low doped samples. Similar tendency was observed in depth profiling of Ron. This suggests a major role of boron in a compensation mechanism. In the second part of the thesis growth and characterization of GaN is presented. Excellent electron transport properties with high electron saturate drift velocity make GaN an excellent candidate for electronic devices. Especially, AlGaN/GaN based high electron mobility transistors (HEMT) have received an increased attention in last years due to their attractive properties. The presence of strong spontaneous and piezoelectric polarization due to the lattice mismatch between AlGaN and GaN is responsible for high free electrons concentrations present in the vicinity of the interface. Due to the spatial separation of electrons and ionized donors or surface states, 2DEG electron gas formed near the interface of the heterostructure exhibits high sheet carrier density and high mobility of electrons. Al0.23Ga0.77N/GaN based HEMT structures with an AlN exclusion layer on 100 mm semiinsulating 4H-SiC substrates have been grown by hot-wall MOCVD. The electrical properties of the two-dimensional electron gas (2DEG) such as electron mobility, sheet carrier density and sheet resistance were obtained from Hall measurements, capacitance-voltage and contact-less eddy-current techniques. The effect of different scattering mechanisms on the mobility have been taken into account and compared to the experimental data. Hall measurements were performed in the range of 80 to 600 K. Hall electron mobility is equal to 17140 cm2(Vs)-1 at 80 K, 2310 cm2(Vs)-1 at room temperature, and as high as 800 cm2(Vs)-1 at 450 K, while the sheet carrier density is 1.04x1013 cm-2 at room temperature and does not vary very much with temperature. Estimation of different electron scattering mechanisms reveals that at temperatures higher than room temperature, experimental mobility data is mainly limited by optical phonon scattering. At relevant high power device temperature (450 K) there is still an increase of mobility due to the AlN exclusion layer. We have studied the behaviour of Ga-face GaN epilayers after in-situ thermal treatment in different gas mixtures in a hot-wall MOCVD reactor. Influence of N2, N2+NH3 and N2+NH3+H2 ambient on the morphology was investigated in this work. The most stable thermal treatment conditions were obtained in the case of N2+NH3 gas ambients. We have also studied the effect of the increased molar ratio of hydrogen in order to establish proper etching conditions for hot-wall MOCVD growth.

This work deals with method for characterization of photovoltaic solar cells based on evaluation of solar cell response to fast transients. The voltage of the cell in both forward and reverse polarisation was controlled by current pulse exciting. Real reverse breakdown voltage and exact value of serial resistance of the cell can be obtained easily by evaluation of the transient curves recorded by digital osciloscope. By negotiation of the time constants of the cell response to excitation in forward polarisation the lifetime of minority carriers in semiconductor bulk can be estimated.

The thesis is focused on high temperature processes in crystalline solar cells production. Main topic is diffusion of traditional dopants phosphorus and boron. Diffusion processes for creating solar cells are different from classical diffusion in semiconductor industrial. It is reason why the thesis describes crated layers in detail. Knowledge of diffusion processes is used for creating bifacial solar cells and development of a new phosphorus emitter for conventional solar cells. Bifacial cells are a new type of cells. Developed new emitter increases efficiency and decreases cost of solar cells production. Another part the thesis is devoted to the prediction of diffusion processes. New models of phosphorus and boron diffusion for photovoltaic industrial are created in software SILVACO. Models correspond with real results.

O trabalho realizado nesta tese esteve apoiado em dois objetivos principais. O primeiro centrado na otimização das etapas e processos de fabricação de células solares de silício de alto rendimento envolvendo redução de custos. O segundo objetivo foi direcionado na implementação de células solares eficientes e não dependentes do armadilhamento de impurezas através da difusão de alumínio. Para levar a cabo estes objetivos de forma planejada, o trabalho dividiu-se em otimizações teóricas e implementações experimentais. As otimizações teóricas foram realizadas utilizando dois programas: um programa desenvolvido (simulacell.pas) e implementado no próprio LME (versão 2), e o outro adquirido comercialmente, PC1D. De acordo com os resultados obtidos em estruturas completas n+p e n++n+p foi possível concluir que tanto as estruturas formadas através de emissores homogêneos como as obtidas utilizando emissores duplamente difundidos permitem alcançar eficiências elevadas, 25,5% a 26,0%, respectivamente, em um amplo intervalo de espessuras e concentrações superficiais de dopantes. No que tange aos desenvolvimentos experimentais, este trabalho se inicia com o desenvolvimento de um processo simplificado de baixo custo, em células solares de silício Cz de baixa resistividade com estrutura n+pp+, tipo \"mesa\". Este processo simplificado também está baseado na difusão de fósforo e alumínio (P/Al), utilizando gases industriais e reagentes químicos de grau \"para análise\", como uma transposição do processo de fabricação anteriormente desenvolvido no LME-EPUSP em substratos de silício FZ utilizando tecnologia planar. A célula solar mais representativa do processo implementado, A-16-1, permitiu atingir eficiências no entorno de 17%. As implementações experimentais visaram inicialmente o desenvolvimento de um procedimento visando à qualificação de materiais de partida (silício), utilizando a técnica de decaimento fotocondutivo (PCD) através de dois procedimentos de passivação de superfícies; oxidações térmicas e difusões suaves de fósforo. Posteriormente, utilizando o sistema PCD, novas otimizações dos emissores de tipo n+ homogêneos e regiões de tipo p foram realizadas, seguidos por oxidações térmicas passivadoras hidrogenadas, preservando-se o tempo de vida do volume em valores elevados (aproximadamente 1ms, após a realização de todas as etapas térmicas). Estes resultados qualificam o silício e os materiais de consumo utilizados, assim como, o novo processo de fabricação desenvolvido. Esta técnica também permitiu qualificar os emissores com perfil Gaussianos processados, atingindo valores da ordem de 45fA/cm2 para densidades de recombinação em estruturas n+pn+. Desenvolveram-se também estruturas n+p em materiais Cz de baixa resistividade 2-3W.cm de dois diferentes fabricantes, e silício FZ com 0,5W.cm. Pôde ser comprovada a qualidade das etapas que compõem o processo completo otimizado tendo-se obtido tensões de circuito aberto-implícitas de 652,4mV (Si-Cz fabricante 1) e 662,6mV (Si-Cz fabricante 2), e 670,8mV (FZ). De acordo com simulações realizadas utilizando parâmetros habituais de dispositivos do próprio LME, estas tensões, quando associadas a um conjunto óptico frontal típico das células solares de alto rendimento do LME (texturização química aleatória e filme de SiO2), permitirão atingir valores entre 19% - 20%. Entretanto, utilizando texturização e camada dupla torna-se plausível atingir o marco de 21% de rendimento, ultrapassando assim a barreira dos 17% (recorde nacional), e comprovando a potencialidade da infra-estrutura deste laboratório para o desenvolvimento de células solares não dependentes do efeito do armadilhamento de impurezas através da difusão de alumínio. / The work developed at this thesis has been based on two main objectives. First, it was focused on the optimization of the steps and processes for the fabrication of high efficiency solar cells, reducing production costs. The latter objective was directed to develop solar cells that were efficient and non-dependent on impurities gettering performed through the aluminum diffusion. In order to attend the planned objectives the work was divided into the theoretical objectives and experimental developments. The theoretical optimizations were performed using two different program codes: one was developed at LME (simulacell.pas), being upgraded afterwards (version 2); and the other was acquired commercially, the PC1D. According to the obtained results in complete structures n+p and n++n+p, it was possible to conclude that the homogeneous and double diffused emitter structures can provide high efficiencies, from 25,5% to 26,0%, respectively, for a wide range of thicknesses and surface doping levels. Concerning the experimental developments, this work starts with a low cost simplified process, using Cz silicon solar cells with low base resistivity and the structure n+pp+, \"mesa\" type. This simplified process was also based on the phosphorus/ aluminum diffusion (P/Al), using industrial gases and for analysis grade chemical reagents, as a fabrication process transposition of the process previously developed at LME-EPUSP using silicon substrates with planar technology. The most representative solar cells of the implemented process, A-16-1, provided about a 17% efficiency. The experimental implementations aimed the development of procedure for starting material (silicon) qualification, by using the photoconductive decay technique (PCD) with two surface passivation procedures: thermal oxidation and light phosphorus diffusion. Later, using PCD system, new optimizations of n+ homogeneous emitters and p-type region were performed, followed by passivating thermal oxidations with hydrogenation, maintaining the volume lifetime at high values (approximately 1ms, after each thermal step). These results qualified the used silicon and the consumer materials, as well the new fabrication process developed. This technique has also allowed qualifying the processed Gaussian profile emitters, providing values about 45fA/cm2 for the recombination current density in n+pp+ structures. N+p structures were also developed using Cz silicon with low resistivity 2- 3W.cm of two different manufacturers and FZ with 0.5W.cm. It could be proved the quality of the steps of a complete optimized process resulting implicit open circuit voltages of 652.4mV (Cz silicon - manufacturer type 1), 662.6mV (Cz silicon - manufacturer type 2), and 670.8mV (FZ silicon). According to the theoretical simulations performed using the usual parameters of devices processed at LME (random chemical texturization and SiO2 film), efficiencies between 19%-20% can be reached. However, using a random texturization and a double layer anti-reflection system, a 21% efficiency becomes possible, surpassing the 17% barrier (national record), and proving the potentiality of this laboratory facility for the development of solar cells non-dependent on impurity gettering through the aluminum diffusion.

O trabalho realizado nesta tese esteve apoiado em dois objetivos principais. O primeiro centrado na otimização das etapas e processos de fabricação de células solares de silício de alto rendimento envolvendo redução de custos. O segundo objetivo foi direcionado na implementação de células solares eficientes e não dependentes do armadilhamento de impurezas através da difusão de alumínio. Para levar a cabo estes objetivos de forma planejada, o trabalho dividiu-se em otimizações teóricas e implementações experimentais. As otimizações teóricas foram realizadas utilizando dois programas: um programa desenvolvido (simulacell.pas) e implementado no próprio LME (versão 2), e o outro adquirido comercialmente, PC1D. De acordo com os resultados obtidos em estruturas completas n+p e n++n+p foi possível concluir que tanto as estruturas formadas através de emissores homogêneos como as obtidas utilizando emissores duplamente difundidos permitem alcançar eficiências elevadas, 25,5% a 26,0%, respectivamente, em um amplo intervalo de espessuras e concentrações superficiais de dopantes. No que tange aos desenvolvimentos experimentais, este trabalho se inicia com o desenvolvimento de um processo simplificado de baixo custo, em células solares de silício Cz de baixa resistividade com estrutura n+pp+, tipo \"mesa\". Este processo simplificado também está baseado na difusão de fósforo e alumínio (P/Al), utilizando gases industriais e reagentes químicos de grau \"para análise\", como uma transposição do processo de fabricação anteriormente desenvolvido no LME-EPUSP em substratos de silício FZ utilizando tecnologia planar. A célula solar mais representativa do processo implementado, A-16-1, permitiu atingir eficiências no entorno de 17%. As implementações experimentais visaram inicialmente o desenvolvimento de um procedimento visando à qualificação de materiais de partida (silício), utilizando a técnica de decaimento fotocondutivo (PCD) através de dois procedimentos de passivação de superfícies; oxidações térmicas e difusões suaves de fósforo. Posteriormente, utilizando o sistema PCD, novas otimizações dos emissores de tipo n+ homogêneos e regiões de tipo p foram realizadas, seguidos por oxidações térmicas passivadoras hidrogenadas, preservando-se o tempo de vida do volume em valores elevados (aproximadamente 1ms, após a realização de todas as etapas térmicas). Estes resultados qualificam o silício e os materiais de consumo utilizados, assim como, o novo processo de fabricação desenvolvido. Esta técnica também permitiu qualificar os emissores com perfil Gaussianos processados, atingindo valores da ordem de 45fA/cm2 para densidades de recombinação em estruturas n+pn+. Desenvolveram-se também estruturas n+p em materiais Cz de baixa resistividade 2-3W.cm de dois diferentes fabricantes, e silício FZ com 0,5W.cm. Pôde ser comprovada a qualidade das etapas que compõem o processo completo otimizado tendo-se obtido tensões de circuito aberto-implícitas de 652,4mV (Si-Cz fabricante 1) e 662,6mV (Si-Cz fabricante 2), e 670,8mV (FZ). De acordo com simulações realizadas utilizando parâmetros habituais de dispositivos do próprio LME, estas tensões, quando associadas a um conjunto óptico frontal típico das células solares de alto rendimento do LME (texturização química aleatória e filme de SiO2), permitirão atingir valores entre 19% - 20%. Entretanto, utilizando texturização e camada dupla torna-se plausível atingir o marco de 21% de rendimento, ultrapassando assim a barreira dos 17% (recorde nacional), e comprovando a potencialidade da infra-estrutura deste laboratório para o desenvolvimento de células solares não dependentes do efeito do armadilhamento de impurezas através da difusão de alumínio. / The work developed at this thesis has been based on two main objectives. First, it was focused on the optimization of the steps and processes for the fabrication of high efficiency solar cells, reducing production costs. The latter objective was directed to develop solar cells that were efficient and non-dependent on impurities gettering performed through the aluminum diffusion. In order to attend the planned objectives the work was divided into the theoretical objectives and experimental developments. The theoretical optimizations were performed using two different program codes: one was developed at LME (simulacell.pas), being upgraded afterwards (version 2); and the other was acquired commercially, the PC1D. According to the obtained results in complete structures n+p and n++n+p, it was possible to conclude that the homogeneous and double diffused emitter structures can provide high efficiencies, from 25,5% to 26,0%, respectively, for a wide range of thicknesses and surface doping levels. Concerning the experimental developments, this work starts with a low cost simplified process, using Cz silicon solar cells with low base resistivity and the structure n+pp+, \"mesa\" type. This simplified process was also based on the phosphorus/ aluminum diffusion (P/Al), using industrial gases and for analysis grade chemical reagents, as a fabrication process transposition of the process previously developed at LME-EPUSP using silicon substrates with planar technology. The most representative solar cells of the implemented process, A-16-1, provided about a 17% efficiency. The experimental implementations aimed the development of procedure for starting material (silicon) qualification, by using the photoconductive decay technique (PCD) with two surface passivation procedures: thermal oxidation and light phosphorus diffusion. Later, using PCD system, new optimizations of n+ homogeneous emitters and p-type region were performed, followed by passivating thermal oxidations with hydrogenation, maintaining the volume lifetime at high values (approximately 1ms, after each thermal step). These results qualified the used silicon and the consumer materials, as well the new fabrication process developed. This technique has also allowed qualifying the processed Gaussian profile emitters, providing values about 45fA/cm2 for the recombination current density in n+pp+ structures. N+p structures were also developed using Cz silicon with low resistivity 2- 3W.cm of two different manufacturers and FZ with 0.5W.cm. It could be proved the quality of the steps of a complete optimized process resulting implicit open circuit voltages of 652.4mV (Cz silicon - manufacturer type 1), 662.6mV (Cz silicon - manufacturer type 2), and 670.8mV (FZ silicon). According to the theoretical simulations performed using the usual parameters of devices processed at LME (random chemical texturization and SiO2 film), efficiencies between 19%-20% can be reached. However, using a random texturization and a double layer anti-reflection system, a 21% efficiency becomes possible, surpassing the 17% barrier (national record), and proving the potentiality of this laboratory facility for the development of solar cells non-dependent on impurity gettering through the aluminum diffusion.

Eine wichtige Eigenschaft von Leistungsdioden ist ihre Stoßstromrobustheit, also die Fähigkeit, einem kurzeitigen hohen Strom standzuhalten. Bei Dioden aus SiC wird dabei häufig auf die MPS-Struktur zurückgegriffen. In dieser Arbeit wird das Stoßstromverhalten von neuen 1200V-SiC-MPS-Dioden von Infineon untersucht. Dabei wird gemessen, welchem Strom die Dioden bei einem Halbsinus-Puls von 10ms standhalten. Das Ergebnis wird mit der Robustheit herkömmlicher SiC-Schottky-Dioden verglichen. Die Stoßstromrobustheit bei Parallelschaltung wird untersucht. Mittels elektro-thermischer Simulationen werden Möglichkeiten zur Verbesserung der Stoßstromrobustheit der Dioden erörtert. Hierbei wird die Schichtdicke erhöht und Kupfer statt Aluminium als Anodenmetallisierung angenommen. Des Weiteren wird das Simulationsmodell hinsichtlich der Lotschicht und eines reduzierten SiC-Substrats variiert. Die MPS-Dioden weisen bei hohem Strom einen negativen differentiellen Widerstand auf, hervorgerufen durch die Injektion von Minoritätsladungsträgern aus den p-dotierten Gebieten. Die Aktivierung der Injektion von Minoritätsladungsträgern in Abhängigkeit der Größe der p-dotierten Gebiete wird ebenfalls mittels Simulationen untersucht. Ein Vergleich verschiedener Caughey-Thomas-Parameter zur Modellierung der Ladungsträgermobilität in FEM-Bauelementsimulatoren wird durchgeführt. Das Ausschaltverhalten der MPS-Dioden wird unter verschiedenen Bedingungen gemessen. Dazu zählen das Abschalten unter Anwendungsbedingungen, das Abschalten unter Überlast von bis zum fünfzehnfachen Nennstrom, das Abschalten mit hohen Strom- und Spannungssteilheiten und das Abschalten in Parallelschaltung und mit zusätzlicher parasitärer Induktivität. Die Untersuchungen zeigen eine hohe Robustheit der neuen 1200V-SiC-MPS-Dioden. / The surge current ruggedness is an import property of power diodes. In case of SiC diodes this is realized by a Merged-pin-Schottky (MPS) structure. In this thesis the surge current ruggedness of novel 1200V SiC MPS diodes from Infineon is investigated. The maximum current during a half sine surge current pulse of 10ms is determined in measurements for various diodes and compared to the surge current ruggedness of conventional Schottky diodes. Furthermore, the surge current ruggedness in parallel arrangement is measured. By the means of electro-thermal simulations options to improve the surge current ruggedness are investigated. The simulation model is varied concerning the diodes anode metalization layer thickness and material. The layer thickness is increased and the typical aluminum is replaced by copper. Additionally, the influence of the silicon carbide substrate thickness and the solder layer thickness and material on the surge current ruggedness is simulated. The MPS diodes possess a negative differential resistance at high currents caused by the injection of minority carriers by the p-doped regions. The injection of minority carriers in dependence of the size of the p-regions is also examined in simulations. Furthermore, different parameter sets of the the Caughey-Thomas formula to describe the carrier mobility are compared. The turn-off behavior of the diodes is measured under different conditions such as the turn-off from fifteen times the rated current, the turn-off with high current and voltage slopes and the turn-off in parallel arrangement and with additional inductance. The investigation show a high robustness of the novel 1200V SiC MPS diodes.