Les investigations des effets non-linéaires (NL) qui apparaissent dans les cristaux ferroélectriques deviennent de plus en plus approfondies, et au temps présent, les cristaux optiques NL deviennent de plus en plus utilisés pour le développement de nouvelles sources de radiations cohérente visibles, conversion de fréquences, ainsi que la détection et diverses transformations de signaux et d images. Un cristal très représentatif de cette classe de matériaux est le niobate de lithium, LiNbO3 (LN), qui depuis son apparition n a cessé de surprendre en révélant de plus en plus ses propriétés. Dans ce travail nous montrons comment la microsonde Raman peut être utilisée pour la caractérisation des microstructures de domaines ferroélectriques dans du niobate de lithium périodiquement polarisé (PPLN). L intensité Raman de modes transverses et longitudinaux de phonons optiques a été enregistrée au travers des stries des domaines ferroélectriques à la surface et en volume d échantillons en z-cut , congruents ou dopés, dont l origine provient de différentes techniques de fabrications. Le changement des intensités intégrées à travers ces structures de domaines a été attribué à une influence des contraintes mécaniques et partiellement du champ de dépolarisation écranté. Nous montrons ainsi l importance de la spectroscopie Raman et la place réelle qu elle occupe comme technique de caractérisation. Ceci ouvre la voie à de nombreuses applications dans ce champ d études. / The investigations of the nonlinear (NL) effects that appear in ferroelectric crystals are becoming more and deeper. At the present time, the NL optical crystals are more and more used for the development of new coherent sources of visible radiations, frequency conversion, beside the detection and several signals and images transformations. A very representative crystal of this material class is lithium niobate, LiNbO3 (LN), that since its appearance never stop to surprise with revealing more and more its properties. In this work, we show how Raman microprobe can be used for the characterisation of the ferroelectric domain microstructures in periodically poled lithium niobate (PPLN). The Raman intensity of transverse and longitudinal modes of optical phonons was recorded across the stripe ferroelectric domains at the surface of a z-cut congruent PPLN sample. The change of integrated intensity was attributed to the influence of mechanical stresses and partially screened depolarization fields. So, we show the importance of Raman spectroscopy and the real place that it takes as a characterisation technique. This open the way for numerous applications in this field of studies.

Despite recent advances in statistics, artificial neural network theory, and machine learning, nonlinear function estimation in high-dimensional space remains a nontrivial problem. As the response surface becomes more complicated and the dimensions of the input data increase, the dreaded "curse of dimensionality" takes hold, rendering the best of function approximation methods ineffective. This thesis takes a novel approach to solving the high-dimensional function estimation problem. In this work, we propose and develop two distinct parametric projection pursuit learning networks with wide-ranging applicability. Included in this work is a discussion of the choice of basis functions used as well as a description of the optimization schemes utilized to find the parameters that enable each network to best approximate a response surface. The essence of these new modeling methodologies is to approximate functions via the superposition of a series of piecewise one-dimensional models that are fit to specific directions, called projection directions. The key to the effectiveness of each model lies in its ability to find efficient projections for reducing the dimensionality of the input space to best fit an underlying response surface. Moreover, each method is capable of effectively selecting appropriate projections from the input data in the presence of relatively high levels of noise. This is accomplished by rigorously examining the theoretical conditions for approximating each solution space and taking full advantage of the principles of optimization to construct a pair of algorithms, each capable of effectively modeling high-dimensional nonlinear response surfaces to a higher degree of accuracy than previously possible.