CoRoT (Convection, Rotation and planetary Transits) é uma missão espacial liderada pela Agência Espacial Francesa (CNES) em associação com diversos parceiros internacionais, entre eles o Brasil. Seus objetivos principais são o estudo da sismologia estelar e a procura por planetas extra-solares (exoplanetas). Ambos os programas científicos baseiam-se em uma fotometria de altíssima precisão e requerem observações ininterruptas de longa duração, possíveis somente a partir do espaço. Uma revisão da literatura indica a existência de três técnicas principais para a realização de fotometria estelar a partir de imagens capturadas em CCDs: fotometria por máscara (por abertura), por ajuste de limiar (por threshold), e por ajuste da resposta impulsiva do sistema de aquisição de imagens. A fotometria por máscara, ou por abertura, apresenta maior precisão para o registro de estrelas brilhantes, em cenários de maior estabilidade de atitude do satélite (situação de baixo jitter), e é a solução adotada a bordo pelo satélite CoRoT, por ser um algoritmo determinístico. A fotometria por ajuste da resposta impulsiva, também chamada de função de espalhamento do ponto (point-spread function, PSF), por sua vez, por levar em conta a resposta do sistema a uma fonte pontual de luz, permite a restauração da imagem original através de processos de deconvolução; apresenta maior precisão para estrelas fracas, ou em um cenário degradado, com perturbações devidas a radiação externa (stray light), ou em que o jitter de atitude do satélite seja elevado. Tal robustez é uma característica desejada no processo de restauração de imagens. Já a fotometria por threshold é aplicada somente em casos de jitter muito elevado e pobre conhecimento da resposta impulsiva do sistema, de modo que não é utilizada no satélite CoRoT. Este trabalho visa consolidar e potencializar a participação brasileira no projeto CoRoT e contribuir com os esforços associados à redução de dados da missão, através de uma proposta inovadora de fotometria híbrida, que se utilizará dos conhecimentos da PSF modelada do instrumento e da alta relação sinal/ruído alcançada com a fotometria por máscara realizada a bordo, baseada e fundamentada em resultados preliminares que atestam o potencial desta metodologia. Tal algoritmo permite um ganho substancial de precisão fotométrica em relação à técnica de abertura, resultando em uma melhor exploração dos dados disponíveis, dada a sua robustez em caso de degradação dos dados de entrada. / CoRoT (Convection, Rotation and planetary Transits) is a high precision photometry experience dedicated to stellar seismology and the search for extrasolar planets. The mission is led by the French Space Agency (CNES) in association with several French laboratories and international partners in Austria, Belgium, Germany, Spain and Brazil. Both scientific programs require great instrumental stability and long, uninterrupted observation runs, which take place simultaneously on adjacent regions of the sky. An overview of the literature displays three main techniques to perform stellar photometry from space CCD images: aperture photometry, threshold photometry and PSF-fitting photometry. Aperture photometry defines a mask which represents the CCD pixels to be summed up in the computation of the collected photon flux for a given star. This method presents very high precision for isolated, bright stars and for stable satellite attitude (low jitter scenarios). It is the data reduction technique implemented on-board, due both to link capacity constraints (given the large number of targets simultaneously observed by the instrument) and to its deterministic algorithm. Fitting photometry allows restoration of degraded stellar images through deconvolution processes, using the point spread function (PSF) of the optical instrument itself. This technique presents better performance for crowded fields and for faint stars; it also presents robustness in the presence of disturbances such as stray light or high satellite attitude jitter. Finally, threshold photometry takes into account only those pixels whose values are above a given pre-computed level. This method is used only in scenarios with excessive satellite depointing due to attitude jitter, or in cases where the instrumental PSF is poorly known. Therefore, it is not used in the CoRoT mission. This work intends to consolidate and strengthen the Brazilian share in the CoRoT Project, contributing to the efforts associated to the ground-based reduction of scientific data, through an innovative, hybrid photometry technique, which will take advantage of a high-resolution instrumental PSF and of the high signal-to-noise-ratio obtained in the on-board aperture photometry. Studies here described show that this methodology, compared to the Aperture technique, achieves gains in photometric precision and in the operational duty cycle, enabling robust and accurate data exploitation.

CoRoT (Convection, Rotation and planetary Transits) é uma missão espacial liderada pela Agência Espacial Francesa (CNES) em associação com diversos parceiros internacionais, entre eles o Brasil. Seus objetivos principais são o estudo da sismologia estelar e a procura por planetas extra-solares (exoplanetas). Ambos os programas científicos baseiam-se em uma fotometria de altíssima precisão e requerem observações ininterruptas de longa duração, possíveis somente a partir do espaço. Uma revisão da literatura indica a existência de três técnicas principais para a realização de fotometria estelar a partir de imagens capturadas em CCDs: fotometria por máscara (por abertura), por ajuste de limiar (por threshold), e por ajuste da resposta impulsiva do sistema de aquisição de imagens. A fotometria por máscara, ou por abertura, apresenta maior precisão para o registro de estrelas brilhantes, em cenários de maior estabilidade de atitude do satélite (situação de baixo jitter), e é a solução adotada a bordo pelo satélite CoRoT, por ser um algoritmo determinístico. A fotometria por ajuste da resposta impulsiva, também chamada de função de espalhamento do ponto (point-spread function, PSF), por sua vez, por levar em conta a resposta do sistema a uma fonte pontual de luz, permite a restauração da imagem original através de processos de deconvolução; apresenta maior precisão para estrelas fracas, ou em um cenário degradado, com perturbações devidas a radiação externa (stray light), ou em que o jitter de atitude do satélite seja elevado. Tal robustez é uma característica desejada no processo de restauração de imagens. Já a fotometria por threshold é aplicada somente em casos de jitter muito elevado e pobre conhecimento da resposta impulsiva do sistema, de modo que não é utilizada no satélite CoRoT. Este trabalho visa consolidar e potencializar a participação brasileira no projeto CoRoT e contribuir com os esforços associados à redução de dados da missão, através de uma proposta inovadora de fotometria híbrida, que se utilizará dos conhecimentos da PSF modelada do instrumento e da alta relação sinal/ruído alcançada com a fotometria por máscara realizada a bordo, baseada e fundamentada em resultados preliminares que atestam o potencial desta metodologia. Tal algoritmo permite um ganho substancial de precisão fotométrica em relação à técnica de abertura, resultando em uma melhor exploração dos dados disponíveis, dada a sua robustez em caso de degradação dos dados de entrada. / CoRoT (Convection, Rotation and planetary Transits) is a high precision photometry experience dedicated to stellar seismology and the search for extrasolar planets. The mission is led by the French Space Agency (CNES) in association with several French laboratories and international partners in Austria, Belgium, Germany, Spain and Brazil. Both scientific programs require great instrumental stability and long, uninterrupted observation runs, which take place simultaneously on adjacent regions of the sky. An overview of the literature displays three main techniques to perform stellar photometry from space CCD images: aperture photometry, threshold photometry and PSF-fitting photometry. Aperture photometry defines a mask which represents the CCD pixels to be summed up in the computation of the collected photon flux for a given star. This method presents very high precision for isolated, bright stars and for stable satellite attitude (low jitter scenarios). It is the data reduction technique implemented on-board, due both to link capacity constraints (given the large number of targets simultaneously observed by the instrument) and to its deterministic algorithm. Fitting photometry allows restoration of degraded stellar images through deconvolution processes, using the point spread function (PSF) of the optical instrument itself. This technique presents better performance for crowded fields and for faint stars; it also presents robustness in the presence of disturbances such as stray light or high satellite attitude jitter. Finally, threshold photometry takes into account only those pixels whose values are above a given pre-computed level. This method is used only in scenarios with excessive satellite depointing due to attitude jitter, or in cases where the instrumental PSF is poorly known. Therefore, it is not used in the CoRoT mission. This work intends to consolidate and strengthen the Brazilian share in the CoRoT Project, contributing to the efforts associated to the ground-based reduction of scientific data, through an innovative, hybrid photometry technique, which will take advantage of a high-resolution instrumental PSF and of the high signal-to-noise-ratio obtained in the on-board aperture photometry. Studies here described show that this methodology, compared to the Aperture technique, achieves gains in photometric precision and in the operational duty cycle, enabling robust and accurate data exploitation.

Globular clusters are dense and massive star clusters that are an integral part of any major galaxy. Careful studies of their stars, a single cluster may contain several millions of them, have revealed that the ages of many globular clusters are comparable to the age of the Universe. These remarkable ages make them valuable probes for the exploration of structure formation in the early universe or the assembly of our own galaxy, the Milky Way. A topic of current research relates to the question whether globular clusters harbour massive black holes in their centres. These black holes would bridge the gap from stellar mass black holes, that represent the final stage in the evolution of massive stars, to supermassive ones that reside in the centres of galaxies. For this reason, they are referred to as intermediate-mass black holes. The most reliable method to detect and to weigh a black hole is to study the motion of stars inside its sphere of influence. The measurement of Doppler shifts via spectroscopy allows one to carry out such dynamical studies. However, spectroscopic observations in dense stellar fields such as Galactic globular clusters are challenging. As a consequence of diffraction processes in the atmosphere and the finite resolution of a telescope, observed stars have a finite width characterized by the point spread function (PSF), hence they appear blended in crowded stellar fields. Classical spectroscopy does not preserve any spatial information, therefore it is impossible to separate the spectra of blended stars and to measure their velocities. Yet methods have been developed to perform imaging spectroscopy. One of those methods is integral field spectroscopy. In the course of this work, the first systematic study on the potential of integral field spectroscopy in the analysis of dense stellar fields is carried out. To this aim, a method is developed to reconstruct the PSF from the observed data and to use this information to extract the stellar spectra. Based on dedicated simulations, predictions are made on the number of stellar spectra that can be extracted from a given data set and the quality of those spectra. Furthermore, the influence of uncertainties in the recovered PSF on the extracted spectra are quantified. The results clearly show that compared to traditional approaches, this method makes a significantly larger number of stars accessible to a spectroscopic analysis. This systematic study goes hand in hand with the development of a software package to automatize the individual steps of the data analysis. It is applied to data of three Galactic globular clusters, M3, M13, and M92. The data have been observed with the PMAS integral field spectrograph at the Calar Alto observatory with the aim to constrain the presence of intermediate-mass black holes in the centres of the clusters. The application of the new analysis method yields samples of about 80 stars per cluster. These are by far the largest spectroscopic samples that have so far been obtained in the centre of any of the three clusters. In the course of the further analysis, Jeans models are calculated for each cluster that predict the velocity dispersion based on an assumed mass distribution inside the cluster. The comparison to the observed velocities of the stars shows that in none of the three clusters, a massive black hole is required to explain the observed kinematics. Instead, the observations rule out any black hole in M13 with a mass higher than 13000 solar masses at the 99.7% level. For the other two clusters, this limit is at significantly lower masses, namely 2500 solar masses in M3 and 2000 solar masses in M92. In M92, it is possible to lower this limit even further by a combined analysis of the extracted stars and the unresolved stellar component. This component consists of the numerous stars in the cluster that appear unresolved in the integral field data. The final limit of 1300 solar masses is the lowest limit obtained so far for a massive globular cluster. / Kugelsternhaufen sind dichte, gravitativ gebundene Ansammlungen von teilweise mehreren Millionen Sternen, die ein fester Bestandteil jeder massiven Galaxie sind. Aus der Untersuchung der Kugelsternhaufen in der Milchstraße weiß man, dass das Alter von vielen dieser Objekte vergleichbar ist mit jenem des Universums. Dies macht sie zu wertvollen Forschungsobjekten, beispielsweise um die Entstehung der Milchstraße und die Strukturbildung im frühen Universum zu verstehen. Eine aktuelle wissenschaftliche Fragestellung befasst sich damit, ob Kugelsternhaufen massive schwarze Löcher beherbergen. Diese würden eine Brücke schlagen von den stellaren schwarzen Löchern, die durch den Kollaps massereicher Sterne entstehen, zu den supermassiven schwarzen Löchern, welche man in den Zentren massiver Galaxien beobachtet. Man bezeichnet sie daher auch als mittelschwere schwarze Löcher. Die sicherste Diagnostik, um schwarze Löcher zu detektieren und ihre Masse zu bestimmen ist, die Bewegung der Sterne innerhalb ihrer gravitativen Einflusssphäre zu vermessen. Spektroskopische Untersuchungen vermögen dies über die Dopplerverschiebung von Spektrallinien, sind jedoch in dichten stellaren Feldern wie Kugelsternhaufen schwierig. Aufgrund der Turbulenz in der Atmosphäre und dem endlichen Auflösungsvermögen des Teleskops erscheinen die Sterne in den Beobachtungen nicht punktförmig, sondern mit einer durch die Punktspreizfunktion (PSF) gegebenen Breite. In dichten stellaren Feldern führt dies dazu, dass die Sterne überlappen. Da klassische spektroskopische Verfahren nicht bildgebend sind, lassen sich die Beiträge der Einzelsterne zu einem beobachteten Spektrum nicht trennen und die Geschwindigkeiten der Sterne können nicht vermessen werden. Bildgebende spektroskopische Verfahren, wie etwa die Integralfeld-Spektroskopie, bieten jedoch die Möglichkeit, die PSF zu rekonstruieren und basierend darauf die Spektren überlappender Sterne zu trennen. Im Rahmen der vorgelegten Arbeit wird das Potential der Integralfeld-Spektroskopie in der Beobachtung dichter stellarer Felder zum ersten Mal systematisch analysiert. Hierzu wird eine Methodik entwickelt, die das Extrahieren von Einzelsternspektren über eine Rekonstruktion der PSF aus den vorhandenen Daten erlaubt. Anhand von Simulationen werden Voraussagen darüber gemacht, wie viele Sternspektren aus einem gegebenen Datensatz extrahiert werden können, welche Qualität diese Spektren haben und wie sich Ungenauigkeiten in der rekonstruierten PSF auf die Analyse auswirken. Es zeigt sich hierbei, dass die entwickelte Methodik die spektroskopische Analyse von deutlich mehr Sternen erlaubt als klassische Verfahren. Parallel zu dieser systematischen Studie erfolgt die Entwicklung einer dezidierten Analysesoftware, welche im zweiten Teil der Arbeit auf Daten von drei Kugelsternhaufen angewendet wird, die mit dem PMAS Integralfeld-Spektrographen am Calar Alto Observatorium aufgenommen wurden: M3, M13 und M92. Die Auswertung dieser Daten liefert Spektren für eine Stichprobe von ungefähr 80 Sternen pro Kugelsternhaufen, weit mehr als bisher im Zentrum eines der drei Haufen verfügbar waren. In der weiteren Analyse werden Jeans Modelle für jedes der drei Objekte gerechnet. Diese erlauben basierend auf einer angenommenen Massenverteilung innerhalb des Kugelsternhaufens eine Vorhersage der Geschwindigkeitsdispersion der Sterne. Der Vergleich mit den gemessenen Geschwindigkeiten zeigt, dass in keinem der drei Haufen ein schwarzes Loch benötigt wird, um die Dynamik der zentrumsnahen Sterne zu erklären. Im Gegenteil, die Beobachtungen können zu 99,7-prozentiger Sicherheit ausschließen, dass sich in M13 ein schwarzes Loch mit einer Masse größer 13000 Sonnenmassen befindet. In den anderen beiden Haufen liegt diese Grenze noch bei deutlich geringeren Massen, nämlich bei 2500 Sonnenmassen in M3 und 2000 Sonnenmassen in M92. In M92 ist es außerdem möglich, das Limit noch weiter herabzusetzen durch eine zusätzliche Analyse der unaufgelösten stellaren Komponente. Diese Komponente besteht aus dem integrierten Licht all jener Sterne, die zu schwach und zahlreich sind als dass sie aus den verfügbaren Daten einzeln extrahiert werden könnten. Das endgültige Limit von 1300 Sonnenmassen ist das geringste, welches bisher in einem massiven Kugelsternhaufen gemessen wurde.