Heidelberg, Universiẗat, Diss., 2001.

La compréhension de la structure des verres est actuellement à l’origine de nombreuses recherches scientifiques. L’une des preuves expérimentale d’un certain ordre dans ceux-ci est la séparation de phase. En effet, ce phénomène est lié à la présence dans ces matériaux amorphes d’au moins deux phases de composition chimique différentes. Ainsi, une étude des prémices de la séparation de phase de type nucléation-croissance dans des verres de silicate de sodium a été menée. La RMN c’est révélée la plus efficace pour cette étude. L’utilisation des expériences de corrélation 29Si-29Si a permis de sonder le réseau silicaté à des échelles plus grandes que celles communément considérées. Nous avons donc pu identifier ce phénomène, caractériser la composition des phases qui apparaissaient et enfin établi un lien avec la cristallisation observée pour de long temps de recuits. A l’inverse, la séparation de phase dans les aluminosilicates de calcium étant visible à l’échelle macroscopique, nous voulions donc suivre la diminution de la séparation de phase avec l’ajout d’alumine. Ainsi, en élaborant un protocole de synthèse spécifique, nous avons pu synthétiser ces verres à hautes températures. Ils ont été étudiés par le biais de la microscopie électronique et de la RMN. Comme ces verres sont composés de nano-domaines vitreux intégrés dans une matrice de composition différente, nous avons pu obtenir des matériaux nano-structurés de taille contrôlée. L’utilisation de la RMN nous a permis de montrer que l’aluminium s’insérait sous forme de « clusters » dans le réseau silicaté. / The understanding of the vitreous state is actually a center of great interest in inorganic chemistry. In fact, even if a glass is often described as a totally disordered material, presence of structures at the atomic scale has to be considered. One of the first experimental proof against the random network theory is the observation of phase separation. In fact, in these systems, at least two glasses with different compositions are observed. In this way, the first steps of phase separation in sodium silicate glasses have been studied. This was afforded with the NMR technique using 29Si-29Si correlations experiments. Thus, the silicon network was probed at higher distances compared with which was commonly observed in these glasses. Therefore, we have been able to identify this phenomenon at the atomic scale, to define the composition of each glassy phase and to do a link with crystallization which happened after long heat treatment of the glass. At the contrary phase separation in calcium aluminosilicate glasses can be observed at the macroscopic scale, observation of its decrease with increasing amount of alumina has been under the scope of this study. The main challenge of this study was to synthesized high temperature glasses. Since an efficient protocol has been implemented, we have been able to obtained nano-structured materials with a controlled size of nano-domains. This caracterisation of the material has been possible using electron microscopy. Combined with NMR, we have been able to propose a mechanism to explain insertion of alumina in the silicon network.

The signal model in Kalman filtering is usually described by a linear dynamical equation driven by generating noise. In this thesis, we set the generating noise to be a white binary sequence and feed to a dynamic system and then forward to a phase modulation system. This discrete generating noise makes the generating signal to be discrete. However, there still are too many states for this discrete signal. In this thesis, we apply the tree structure to define the states. The number of states is reduced by focusing on the most probable working states. To save the bandwidth, we interpolate the generating signal in two methods. One is before the phase modulation, the other is after phase modulation. The former one has a better error rate than GMSK for SNR between- 2dB~5dB and if given the bandwidth around 500k(Hz). The former one has a better error rate than GNSK for SNR between - 3dB~5dB and if given the bandwidth around 500k(Hz). Our error performance is studied by computer simulation. This error performance is also successfully analyzed by computing the distance spectrums

Marburg, Univ., Diss., 2003.

Hamburg, Univ., Diss., 2000. / Computerdatei im Fernzugriff.

Hamburg, Univ., Diss., 2000. / Computerdatei im Fernzugriff.

Stuttgart, Univ., Diss., 2002. / Computerdatei im Fernzugriff.

Paderborn, Universiẗat, Diss., 2000.

Paderborn, Universiẗat, Diss., 2000.

Marburg, Universiẗat, Diss., 2003. / Erscheinungsjahr an der Haupttitelstelle: 2002.