[en] The Josephson effect appears when a weak-link establishes phase coherence between two superconductors. A unifying theory of this effect emerged in the 90's within the framework of mesoscopic physics. Based on two cornerstone concepts, conduction channels and Andreev reflection, it predicts the current-phase relation for the most basic weak-link: a single conduction channel of arbitrary transmission. This thesis illustrates this mesoscopic point of view with experiments on superconducting atomic size contacts. In particular, we have focused on the supercurrent peak around zero voltage, put into evidence the ac Josephson currents in a contact under constant bias voltage (Shapiro resonances and photon assisted multiple Andreev reflections), and performed direct measurements of the current-phase relation. (author)

[en] The future of hard X-ray astronomy relies on the development of new instruments able to focus photons of a hundred keV. Indeed, focalization allows an important improvement in sensitivity and angular resolution. Achieved by grazing incidence reflections on Wolter-I mirrors, its use currently limited to tens of keV can be extended to higher energies thanks to a specific coating and a large focal length. As X-ray observations are only possible above the atmosphere, the size of the observatories, and hence their focal length, was limited by the launcher capacity. Over the past few years, different technologies like extendible masts or formation flight have been studied to go beyond this limit. To gain a better understanding of these telescopes, I detail the Wolter-I mirror geometry, their coating reflectivity, the detection in semi-conductor as well as the dynamic related to extendible masts and formation flight. These telescopes are complex optical systems, subject to deformations during observation and need a fine metrology system to measure these deformations for image correction. To study their performance, I developed a code reproducing the real functioning of such a telescope. Each photon is considered individually, its path and interactions depend on the behavior of the telescope structure along with time. Each component of the telescope is modeled, as well as the metrology needed for the restitution of its dynamic. The path of the photon is computed in a three dimensional vector space, using Monte-Carlo methods to reproduce the mirror defaults, their reflectivity and the interactions in the detector. The simulation produces images and energy spectra, from which we can infer the angular resolution, the field of view, the effective area and the detection efficiency. In 2006, the Simbol-X mission was selected in the framework of the formation flight studies. This concept allows a large focal length, the telescope being distributed on two independent spacecrafts. However, the particular dynamic of the formation flight has consequences on the telescope performance and must be controlled. In the framework of this mission, my simulation tool was used to study the consequence of the motion of each satellite on the telescope performance and the influence of the metrology errors on the image reconstruction. This study led to the attitude control specification of each spacecraft and to the metrology accuracy specification. Considering these results, I demonstrate the feasibility of this kind of telescope. Beyond the Simbol-X mission, I have investigated the performance optimization of a hard X-ray telescope. Using my simulation, I have studied the impact of each parameter on the telescope performance. These studies have led to the design of the PheniX project, a telescope operating in the 1-200 keV band, proposed by the Centre d'Etude Spatial des Rayonnements in response to the European Space Agency M3 call. Thanks to a new coating and a large focal length obtained with an extendible mast, this telescope offers a level of performance at 100 keV more than 100 times better than the actual instruments. I present this project and its expected performances in the last part of my thesis. (author)

[en] During the first 7 yr of the INTEGRAL mission (2003-2009), Cyg X-1 has essentially been detected in its hard state (HS), with some incursions in intermediate HSs. This long, spectrally stable period allowed in particular the measurement of the polarization of the high-energy component that has long been observed above 200 keV in this peculiar object. This result strongly suggests that here we see the contribution of the jet, known to emit a strong synchrotron radio emission. In 2010 June, Cyg X-1 underwent a completed transition toward a soft state (SS). It gave us the unique opportunity to study in detail the corona emission in this spectral state, and to investigate in particular the behavior of the jet contribution. Indeed, during the SS, the hard X-ray emission decreases drastically, with its maximum energy shifted toward lower energy and its flux divided by a factor of ∼5-10. Interestingly, the radio emission follows a similar drop, supporting the correlation between the jet emission and the hard component, even though the flux is too low to quantify the polarization characteristics.

[en] We have developed a simulation tool for a Wolter I telescope operating in formation flight. The aim is to understand and predict the behavior of the Simbol-X instrument. As the geometry is variable, formation flight introduces new challenges and complex implications. Our code, based on Monte Carlo ray tracing, computes the full photon trajectories up to the detector plane, along with the relative drifts of the two spacecrafts. It takes into account angle and energy dependent interactions of the photons with the mirrors and applies to any grazing incidence telescope. The resulting images of simulated sources from 0.1 keV to 100 keV allow us to optimize the configuration of the instrument and to assess the performance of the Simbol-X telescope.

[en] Simbol-X is an X-Ray telescope operating in formation flight. It means that its optical performances will strongly depend on the drift of the two spacecrafts and its ability to measure these drifts for image reconstruction. We built a dynamical ray tracing code to study the impact of these parameters on the optical performance of Simbol-X (see Chauvin et al., these proceedings). Using the simulation tool we have developed, we have conducted detailed analyses of the impact of different parameters on the imaging performance of the Simbol-X telescope.

No abstract available

[en] Operational since 2002 on board the INTEGRAL observatory, the SPI spectrometer can be used to perform polarization measurements in the hard X-ray/softγ-ray domain (∼130 keV-8 MeV). However, this phenomenon is complex to measure at high energy and requires high fluxes. Cyg X-1 appears to be the best candidate amongst the X-ray binaries since it is one of the brightest persistent sources in this energy domain. Furthermore, a polarized component has recently been reported above 400 keV from IBIS data. We have therefore dedicated our efforts to developing the required tools to study the polarization in the INTEGRAL SPI data and have first applied them to 2.6 Ms of Cyg X-1 observations, covering 6.5 years of the INTEGRAL mission. We have found that the high energy emission of Cyg X-1 is indeed polarized, with a mean polarization fraction of 76% ± 15% at a position angle estimated to be 42°± 3°, for energies above 230 keV. The polarization fraction clearly increases with energy. In the 130-230 keV band, the polarization fraction is lower than 20%, but exceeds 75% between 370 and 850 keV, with the (total) emission vanishing above this energy. This result strongly suggests that the emission originates from the jet structure known to emit in the radio domain. The same synchrotron process could be responsible for the emission from radio to MeV, implying the presence of high energy electrons. This illustrates why the polarization of the high energy emission in compact objects is an increasingly important observational objective.

[en] Dosimetry in Nuclear Medicine is based on a common formalism (MIRD) using pre-calculated reference S Values. For decades, S Values generated from mathematical models were the only available reference. The transition from mathematical to voxel-based and meshed-based models generates a need for new reference data. The OpenDose project brings together resources and expertise through an international collaboration to generate, verify and disseminate reference dosimetric data to the Nuclear Medicine community.

[en] We present recent improvements in polarization analysis with the INTEGRAL SPI data. The SPI detector plane consists of 19 independent Ge crystals and can operate as a polarimeter. The anisotropy characteristics of Compton diffusions can provide information on the polarization parameters of the incident flux. By including the physics of the polarized Compton process in the instrument simulation, we are able to determine the instrument response for a linearly polarized emission at any position angle. We compare the observed data with the simulation sets by a minimum χ² technique to determine the polarization parameters of the source (angle and fraction). We have tested our analysis procedure with Crab Nebula observations and find a position angle similar to those previously reported in the literature, with a comfortable significance. Since the instrument response depends on the incident angle, each exposure in the SPI data requires its own set of simulations, calculated for 18 polarization angles (from 0° to 170° in steps of 10°) and unpolarized emission. The analysis of a large number of observations for a given source, required to obtain statistically significant results, represents a large amount of computing time, but it is the only way to access this complementary information in the hard X-ray regime. Indeed, major scientific advances are expected from such studies since the observational results will help to discriminate between the different models proposed for the high energy emission of compact objects like X-ray binaries and active galactic nuclei or gamma-ray bursts.

No abstract available