Une démarche de type génie de la réaction chimique est appliquée à l'hydroformylation modèle d'oct-1-ène par des complexes lipophobes du rhodium préparés à partir de Rh(CO)2(acac) en phase liquide ionique ([Bmim][PF6]) ou en phase liquide ionique supportée sur silice. La réaction étant contrôlée par la concentration des réactifs dans la phase liquide ionique catalytique, une première étape a consisté à mesurer ces concentrations tant pour les deux gaz (H2 et CO) que pour l'oct-1-ène à différentes températures et pressions. Diverses méthodes de mesures sont utilisées pour la solubilité de l'oléfine : thermogravimétrie et chromatographie gazeuse après extraction multiple d’espace de tête, en présence de solvant (décane) et du produit de la réaction (nonanal). Le transfert gaz-liquide, qui peut conditionner la vitesse de réaction dans ces milieux visqueux, est également mesuré par une technique dynamique de variation de pression, en liquide ionique pur et en mélange biphasique liquide ionique-phase organique, dans un réacteur autoclave à autoaspiration de gaz par arbre creux. Une corrélation générale est proposée montrant une forte influence de la vitesse d'agitation.Une étude cinétique est réalisée en conditions de transferts non limitants en gaz-liquide organique-liquide ionique avec la TPPTS comme ligand. Les comportements habituels de l’hydroformylation  en  phase organique ou en phase aqueuse sont retrouvés : ordre voisin de 1 pour H2, inhibition par CO à forte concentration, énergie d'activation élevée. Si le turnover est convenable (70 h-1), le rapport n/iso est par contre très bas ce qui n'est pas en faveur de ce système catalytique. Quelques résultats permettent aussi une première analyse de la catalyse biphasique avec le ligand sulfoxantphos et de la catalyse en phase liquide ionique supportée sur silice avec la TPPTS. / A chemical reaction engineering approach is applied to the hydroformylation of 1-octene using lipophobic complexes of rhodium prepared from Rh(CO)2(acac) in ionic liquid phase ([Bmim] [PF6]) or in the ionic liquid phase supported on silica. As the reaction is controlled by the concentration of the reagents in the catalytic ionic liquid phase, the concentrations of both gases (H2 and CO) and also of 1-octene are measured at various temperatures and pressures as an initial step. Different methods are used for the measurement of the olefin solubility inside the ionic liquid: thermogravimetry and multiple headspace chromatography, in the presence of solvent (decane) and reaction product (nonanal). The gas-liquid mass transfer, which can be a rate controlling step in these viscous media, is also measured by a dynamic technique of pressure variation, both in case of pure ionic liquid and biphasic mixture of ionic liquid and organic phase, in an autoclave reactor with self induced stirrer. A general correlation is proposed showing the strong influence of the agitation speed. A kinetic study is realized in no gas–liquid nor organic–ionic liquid mass transfer limiting conditions (chemical regime) with TPPTS as ligand. The usual hydroformylation behaviour is observed, as already found in organic phase or in aqueous phase: order close to 1 for H2, inhibition by CO at large concentration, and high activation energy. If the turnover frequency is suitable (70 h-1), the n/iso ratio is very low which is not favourable to this catalytic system. Some experimental results also allow a first analysis of biphasic catalysis with sulfoxantphos ligand and of ionic liquid phase supported catalysis with TPPTS ligand.

This thesis focus on the analysis of spin-lattice NMRD relaxation profilesmeasured in various complex systems such as proteins, zeolites and ionicliquids. Proton, deuterium and fluoride T1-NMRD relaxation profiles wereobtained from a fast-field cycling (FFC) instrument. It is found that alsopossible to obtain NMRD profiles from the molecular dynamics (MD)simulation trajectories. NMRD Profiles were analyzed by using differentrelaxation models, such as the Solomon-Bloembergen-Morgan (SBM) theoryand the Stochastic Liouville (SL) theory. Paper I described the hydration of protein PrxV obtained from a MDsimulation, and compared with the picture emerges from an analysis byusing a generally accepted relaxation model [appendix C]. The result showsthat the information from NMRD analysis is an averaged picture of watermolecules with similar relaxation times; and the MD simulations containsinformation of all types of interested water molecules with differentresidence times. In paper II NMRD profiles have been used to characterize the hydration ofthe oxygen-evolving complex in state S1 of photosystem II. NMRDexperiments were performed on both intact protein samples and Mndepletedsamples, and characteristic dispersion difference were foundbetween 0.03 MHz to 1 MHz; approximately. Both the SBM theory and theSL theory have been used to explain this dispersion difference, and it isfound that this is due to a paramagnetic enhancement of 1-2 water moleculesnearby ~10 Å from the spin center of the Mn4CaO5 cluster. The result showsthe reorientation of the molecular cluster is in μs time interval. Whencompare these two theories, the SL theory presented a better interpretationbecause parameters obtained from the SBM theory shows they didn’t fulfilthe presupposed perturbation criterion (the Kubo term). Paper III deals with the water dynamics in the restricted/confined spaces inthe zeolite samples (H-ZSM-5 and NH4-ZSM-5) and obtained by proton anddeuterium spin-lattice NMRD profiles. The results show that the spin-latticeNMRD can be used to characterize various zeolites. The temperature has aweak effect on the relaxation rate R1, but the change of different counter ionsmay change the hydration and the translational diffusion pores and givedifferent R1. Proton and fluoride NMRD profiles and MD simulations were both used tostudy the dynamics of BMIM[PF6] in paper IV. Results indicate the reorientation of the molecules are in the ns time regime, and the effectivecorrelation time obtained from 1H and 19F are the same. From the MDsimulation it is found the reorientation of [PF6]- ions is much faster (in ps)compare with BMIM+ ion which moves in the ns time range. With previous results, the FFC NMRD profiles are indeed very informativetools to study the molecular dynamics of complex systems. The MDsimulation can be used as a complementary method to obtain detailedinformation. By combine these two methods, it provide a more colorfulpicture in the study of protein hydration and liquid molecular dynamics.