Statics, dynamics, rheology and scission-recombination kinetics of self-assembling linear micelles are investigated at equlibrium state and under shear flow by computer simulations using a newly proposed mesoscopic model. We model the micelles as linear sequences of Brownian beads whose space-time evolution is governed by Langevin dynamics. A Monte Carlo algorithm controls the opening of a bond or the chain-end fusion. A kinetic parameter omega modelling the effect of a potential barrier along a kinetic path, is introduced in our model.<p>For equilibrium state we focus on the analysis of short and long time behaviors of the scission and recombination mechanisms. Our results show that at time scales larger than the life time of the average chain length, the kinetics is in agreement with the mean-field kinetics model of Cates. By studying macroscopic relaxation phenomena such as the average micelle length evolution after a T-jump, the monomer diffusion, and the zero shear relaxation function, we confirm that the effective kinetic constants found are indeed the relevant parameters when macroscopic relaxation is coupled to the kinetics of micelles.<p>For the non-equilibrium situation, we study the coupled effects of the shear flow and the scission-recombination kinetics, on the structural and rheological properties of this micellar system. Our study is performed in semi-dilute and dynamically unentangled regime conditions. The explored parameter omega range is chosen in order for the life time of the average size chain to remain shorter than its intrinsic (Rouse) longest relaxation time. Central to our analysis is the concept of dynamical unit of size Lambda, the chain fragment for which the life time tau_Lambda and the Rouse time are equal. Shear thinning, chain gyration tensor anisotropy, chain orientation and bond stretching are found to depend upon the reduced shear rate Beta_Lambda=gamma dot*tau_Lambda while the average micelle size is found to decrease with increasing shear rate, independently of the height of the barrier of the scission-recombination process. / Doctorat en sciences, Spécialisation physique / info:eu-repo/semantics/nonPublished

La diffraction des rayons X permet d’obtenir des informations sur la structure atomique et même sur la distribution de charges de composés sous forme cristalline, ce qui est d’une importance fondamentale pour la compréhension de leurs propriétés. Accéder expérimentalement à une description de la distribution de charges de systèmes macromoléculaires reste rarement possible malgré les améliorations techniques. Pour pallier cette limite, la transférabilité des paramètres de distributions de charges est un moyen fiable d’obtenir pour ces systèmes un modèle estimé et d’en déduire leurs propriétés électrostatiques. Les résultats présentés dans ce travail de thèse s’intègrent dans une dynamique visant à étendre les méthodes initialement réservées pour l’étude des petites molécules aux systèmes macromoléculaires. Il s’articule autour du développement de la suite de logiciels MoPro et de la banque de données ELMAM (Experimental Library of Multipolar Atom Model) pour l’étude des énergies des interactions électrostatiques au sein du site actif de complexes enzyme-inhibiteur. L’étude du fidarestat, un inhibiteur de l’holoenzyme aldose réductase, réalisée à partir de données obtenues à très haute résolution, est exposée dans ce manuscrit et a servi notamment à l’amélioration de la banque ELMAM en vue de l’étude des complexes holoenzymes aldo-keto réductase. A cette occasion, la légitimité du recours aux modèles transférés de distribution de charges a été discutée pour la première fois par une estimation statistique des incertitudes sur les énergies d’interaction électrostatique entre enzyme et inhibiteur / X-ray diffraction allows to obtain information about atomic structure and charge density distribution of crystal-state compounds, which is of main interest for the understanding of their properties. Reaching experimentally charge density distribution description of macromolecular systems is rarely possible despite technical improvements. To get around this limit, the transferability of charge density distribution parameters is a reliable way to obtain for these systems estimated model and to deduce their electrostatic properties. Works introduced in this PhD thesis manuscript take part in the will of extending methods initially for study of small molecules to macromolecular systems. It is centered on the development of the MoPro software suite and of ELMAM database (Experimental Library of Multipolar Atom Model) for the study of electrostatic interaction energies in enzyme-inhibitor complexes’ active site. The study of fidarestat, an inhibitor of aldose reductase holoenzyme, performed using ultra-high resolution data, is exposed in this manuscript and allowed to improve ELMAM database for the study of electrostatic interaction in aldo-keto reductase holoenzyme complexes. Moreover, the legitimacy of using transferred charge density distribution models was discussed for the first time, thanks to statistical estimation of uncertainties on electrostatic interaction energies between enzyme and inhibitor