Dissertations / Theses on the topic 'Molecular modeling'

Thesis (M.Eng. and S.B.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2001.
Includes bibliographical references (p. 81-83).
by Delphine Marguerite Denise Dean.
M.Eng.and S.B.

Diss. (sammanfattning) Stockholm : Stockholms universitet, 2010.
At the time of the doctoral defense, the following paper was unpublished and had a status as follows: Paper 3: Submitted. Härtill 4 uppsatser.

Cette thèse est une étude, au niveau moléculaire, par des méthodes de simulation numérique, des propriétés des aérosols organiques, notamment des aérosols marins, et de leur interaction avec des espèces présentes dans l’atmosphère. L’organisation de la matière organique au sein de ces aérosols joue un rôle fondamental pour leurs propriétés optiques, chimiques et leur rôle comme noyau de condensation pour les nuages.Dans une première partie, on présente contexte atmosphérique et les méthodes de dynamique moléculaire classique et les méthodes mixtes quantique/classique utilisées pour simuler ces aérosols. Ensuite, on décrit l’application de ces méthodes à trois cas.Tout d’abord, on a étudié, par dynamique moléculaire classique (logiciel Gromacs), l’organisation, notamment l’orientation, de molécules d’acide palmitique absorbées sur une surface de sel (NaCl) en fonction du taux de couverture en acide gras et de la température. On présente aussi une étude détaillée de l’influence de l’humidité sur l’organisation de ce film organique à la surface du sel, mettant en évidence l’existence d’îlots d’acide gras monocouches structurés. Dans une seconde étude, la réactivité de NO2 avec cet aérosol marin est traitée par une approche mixte quantique/classique (logiciel CP2K), avec prise en compte de l’impact de l’humidité sur cette réactivité.Enfin, la dernière étude concerne une étude par dynamique moléculaire de phases condensées organiques (n-butanol) et de leur interaction avec des molécules d’eau. Cette étude théorique, complémentaire d’expériences de jets moléculaires, a pour but de mieux comprendre le rôle fondamental que jouent ces interactions pour les propriétés des aérosols et des nuages
In this thesis numerical methods are used to study the properties, described at the molecular level, of organic aerosols, especially marine aerosols, and their interaction with species in the atmosphere. The organisation of the organic matter in these aerosols plays a key role for their optical, chemical properties, and their ability to act as a cloud condensation nuclei.The first part reviews atmospheric context and the methods (classical molecular dynamics and hybrid quantum/classical approaches) used in this thesis. Then applications to three cases are detailed.Firstly, the organization, more particularly the orientation, of palmitic acid molecules adsorbed on a salt (NaCl) surface as a function of the fatty acid coverage and temperature has been studied using classical molecular dynamics (Gromacs package). The impact of the humidity on the structuration of this organic coating has been described in details, showing the existence of structured fatty acid island-like monolayers on NaCl surface.In a second study, the reactivity of NO2 with these heterogeneous marine aerosols has been investigated by a hybrid quantum/classical method (CP2K package), with taking into account the effect of the humidity.The last study is a classical molecular dynamics of n-butanol crystal, water accommodation at these surfaces and simulation of water jet collision with n-butanol surface. These simulations, complementary to experiments, were performed to better understand the fundamental role of the water-organic matter interaction on the properties of the aerosols and clouds

Thesis (Ph.D)--Chemical Engineering, Georgia Institute of Technology, 2009.
Committee Chair: Ludovice, Pete; Committee Member: Grover, Martha; Committee Member: Jones, Christopher; Committee Member: Realff, Matthew; Committee Member: Sherrill, David. Part of the SMARTech Electronic Thesis and Dissertation Collection.

Amongst the many complex processes taking place in living cells, transport of cargoes across the cytosceleton is fundamental to cell viability and activity. To move cargoes between the different cell parts, cells employ Molecular Motors. The motors operate by transporting cargoes along the so-called cellular micro-tubules, namely rope-like structures that connect, for instance, the cell-nucleus and outer membrane. We introduce a new Markov Chain, the killed Quasi-Random-Walk, for such transport molecules and derive properties like the maximal run length and time. Furthermore we introduce permuted balance, which is a more flexible extension of the ordinary reversibility and introduce the notion of Time Duality, which compares certain passage times pathwise. We give a number of sufficient conditions for Time Duality based on the geometry of the transition graph. Both notions are closely related to properties of the killed Quasi-Random-Walk.

Les outils de simulation numérique permettent désormais l'analyse des interfaces à l'échelle moléculaire, tant du point de vue de leur structure que de leur comportement dynamique. Ainsi, dans mon travail de thèse, j'ai développé le logiciel PYTIM qui comprend les procédures les plus populaires d’analyse inter faciale à l'échelle moléculaire, fournissant une base solide pour les travaux de recherche sur les surfaces et interfaces.En utilisant ces méthodes, j’ai étudié le comportement dynamique des molécules situées aux interfaces de différents systèmes d'intérêt biologiques et atmosphériques. Ce faisant, j’ai étudié la corrélation entre la dynamique des molécules à la surface et les interactions intermoléculaires correspondantes.De plus, j’ai travaillé sur le calcul des profils de pression dans les systèmes simulés. Dans ce cas, définir localement une quantité macroscopique, la pression, à l’échelle microscopique représente un obstacle considérable. Nous avons cependant montré que les profils de pression peuvent être calculés dans des systèmes comprenant des charges ponctuelles via le contour de Harasima avec la méthode de sommation d’Ewald (PME). Par ailleurs, j'ai montré comment les contraintes rigides souvent utilisées dans les simulations introduisent un couplage entre les degrés de liberté translationnels (positions) et les degrés de liberté de rotation. La conséquence de ce couplage est que le tenseur d’énergie cinétique n'est plus constant même dans des systèmes en équilibre, ce qui peut introduire une différence significative dans le calcul de la tension de surface.Les méthodes développées au cours de mon travail de thèse ont permis de calculer, pour différents systèmes, la distribution de la tension superficielle près de l’interface, la relation entre la pression spinodale et le minimum du profil de pression latérale. Elles ont également permis de mieux comprendre les liens entre pression et mécanisme d’action des molécules anesthésiques,fournissant ainsi des bases moléculaires à l'hypothèse de Cantor.Enfin, j’ai étudié également l'équilibre gaz/solide en caractérisant, à l’aide de simulation de Monte Carlo dans l'ensemble grand canonique, le piégeage de molécules d'ammoniac dans un clathrate, sous conditions de pression et de température caractéristiques d'environnements extraterrestres
The tools of numerical simulation enable the analysis of interfaces at themolecular scale, both in terms of their structure and their dynamic behavior.Thus, in my thesis work, I developed the PYTIM software that includes the mostpopular procedures for interfacial analysis at the molecular level, providing asolid foundation for research work on surfaces and interfaces.Using these methods, I investigated the dynamic behavior of molecules at theinterfaces of different biological and atmospheric systems of interest. Indoing so, I studied the correlation between the dynamics of molecules on thesurface and the corresponding intermolecular interactions. In addition, Iworked on the calculation of pressure profiles in simulated systems. Inparticular, the localization of an inherently non-local quantity, the pressure,represents a considerable technical difficulty. I have shown that the pressureprofiles can be calculated in systems containing point charges via the Harasimacontour with mesh Ewald methods (PME). Moreover, I showed how the rigidconstraints often used in simulations introduce a coupling between thetranslational degrees of freedom and the rotational degrees of freedom. Theconsequence of this coupling is that the kinetic energy tensor is no longerconstant, even in equilibrium systems, which -- if neglected -- can introducesignificant errors in the calculation of the surface tension.The methods developed during my thesis work provided means to study variousproblems, such as the distribution of the surface tension near the interface,the relation between the spinodal pressure and the minimum of the lateralpressure profile. They also enabled the investigation of the possible linksbetween the lateral pressure profiles and the mechanism of action of anestheticmolecules, thus providing a molecular basis for the hypothesis ofCantor.Finally, I also studied gas/solid equilibrium characterizing, by Monte Carlosimulation in the grand canonical ensemble, the trapping of ammonia moleculesin a clathrate under conditions of pressure and temperature representative ofextraterrestrial environments

Dirhodium complexes such as carboxylates and carboxylamidates are very efficient metal catalysts used in the synthesis of pharmaceuticals and agrochemicals. Recent experimental work has indicated that there are significant differences in the isomeric ratios obtained among the possible products when synthesizing these complexes. The relative stabilities of the Rh2(NPhCOCH3)4 tolunitrile complexes, Rh2(NPhCOCH3)4(NCC6H4CH3)2, were determined at the HF/LANL2DZ ECP, 6-31G and DFT/B3LYP/LANL2DZ ECP, 6-31G levels of theory using NWChem 6.3. The LANL2DZ ECP (effective core potential) basis set was used for the rhodium atoms and 6-31G basis set was used for all other atoms. Specifically, the o-tolunitrile, m-tolunitrile, and p-tolunitrile complexes of the 2,2-trans and the 4,0- isomers of Rh2(NPhCOCH3)4 were compared.

Molecular materials have raised much interest in the last decades in the quest for new multifunctional devices. Among the multiple properties that those materials may present, one of the most typical is magnetism, which arises from the presence of unpaired electrons in the molecules that constitute the three-dimensional crystal. Magnetism has a macroscopic observable, the magnetic susceptibility (Ji), which is usually rationalized in terms of a set of JAB magnetic interactions between pairs of molecules. However, any experimental technique allows for such direct correspondence and, thus, the experimental interpretation of the magnetic properties usually requires further analysis from the point of view of computational chemistry. Consequently, the present PhD thesis is a contribution to the computational modeling of molecule-based magnetic materials. Specifically, we describe how the tools of computational chemistry may be used in order to study those materials from different perspectives. With this aim in mind, we have applied computational chemistry techniques to rationalize the magnetic properties of several systems of interest, ranging from metal-organic compounds, based on Cu(II), to pure organic radicals based on the DTA and Benzotriazinyl building blocks, and including compounds based on the metal-radical synthetic approach, and also spin crossover materials based on Fe(II). Along the thesis we have demonstrated that computational chemistry is a helpful discipline, capable to aid in the interpretation of experimental results and in the prediction of interesting properties, especially when working in close collaboration with experimentalists. In particular, the First Principles Bottom-Up (FPBU) procedure, extensively developed in our group, is a useful tool to rationalize the magnetic properties of any molecular magnetic material. To this purpose, the magnetic topology (i.e. the network of JAB within the crystal) is the key element. Regarding the magnetic topology, we have also demonstrated that it can be more intricate and complex than expected, and that it cannot be directly inferred from the coordination pattern of the molecule-based material. Therefore, the experimental assignation of the magnetic topology, by means of a fitting procedure, must be taken with caution. About the JAB values, we have proved that they depend on temperature, and that this dependence may be especially important when working with organic radicals. On this class of materials, we have analyzed how the JAB values evolve with time, and seen that this evolution may involve huge fluctuations of their magnitude as a consequence of the thermal motion at finite temperatures. Interestingly, we demonstrate herein that, when the JAB values depend non-linearly with the thermal vibrations of a material, the standard static perspective of magnetism is not valid to fully understand their magnetic properties, and that it is then required to adopt a dynamic perspective. Regarding the computational modeling of JAB values, we have seen that the combination of UB3LYP and the Broken-Symmetry approach yields JAB values, when transformed into the macroscopic observables, are in good agreement with experiment. In fact, we have demonstrated that, in order to predict the strength of a given JAB value, small distortions in the crystal structure can induce large variations, which may be much more important than the intrinsic error associated with the theoretical method employed. We have also observed that the counterions and diamagnetic ligands may have an important effect in defining the magnetic properties of a system. Overall, we have demonstrated that the magnetic topology and, thus, the macroscopic magnetic properties of a given material, cannot be understood without the proper knowledge of their crystal structure.
Els materials moleculars han despertat molt d'interès en les últimes dècades degut a la seva possible aplicació en nous dispositius multifuncionals. Entre les diferents propietats que aquests materials poden presentar, una de les més típiques és el magnetisme, el qual sorgeix de la presència d’electrons desaparellats en les molècules que constitueixen el cristall tridimensional. El magnetisme té un observable macroscòpic, la susceptibilitat magnètica (Ji), que sol ser racionalitzada en termes microscòpics mitjançant el conjunt d'interaccions magnètiques JAB entre determinats parells de molècules. No obstant això, cap tècnica experimental permet aquesta correspondència directa i, per tant, la interpretació experimental de les propietats magnètiques sol requerir d’un posterior anàlisi des del punt de vista de la química computacional. La present tesi doctoral pretén doncs contribuir en el camp del magnetisme molecular i, més concretament, en com es poden utilitzar les eines de la química computacional per a modelitzar materials magnètics moleculars des de diferents perspectives. Amb aquest objectiu en ment, s’han racionalitzat les propietats magnètiques de diversos sistemes d'interès, que van des de compostos metal•lorgànics basats en ions de Cu(II) o de Co(II), radicals orgànics purs, compostos basats en l’estratègia sintètica de “metall-radical”, i finalment també materials de spin crossover basats en Fe(II). Al llarg de la tesi s'ha demostrat que la química computacional és una disciplina útil, capaç d'ajudar a la interpretació dels resultats experimentals i en la predicció de propietats interessants, especialment quan es treballa en estreta col•laboració amb els experimentadors. En particular, el procediment de primers principis Bottom-Up (FPBU, per les seves sigles en anglès), desenvolupat àmpliament en el nostre grup, és una eina útil per racionalitzar les propietats magnètiques de qualsevol material magnètic molecular. Per a aquest propòsit, la topologia magnètica (és a dir, la xarxa de JAB dins del cristall) és l'element clau. A més, hem analitzat diversos factors que afecten aquesta topologia magnètica, com els contraions, els radicals diamagnètics o l’efecte de la temperatura, mitjançant el la seva manifestació en les vibracions del cristall i en la contracció (expansió) que pateix al refredar-se (escalfar-se).

Free volume and free volume distribution have long been used to explain differences in the gas transport properties of polymeric materials. However, only a few experimental techniques allow a comprehensive evaluation of polymeric void space. Through the use of computer simulations, the free volume was characterized of two polyester systems used for beverage packaging and polynorbornene, a unique polymer with possible applications in both microelectronic fabrication and membrane separations. Delaunay Tessellation was used to calculate the fractional free volume (FFV) of both polyethylene terephthalate (PET) and polyethylene naphthalate (PEN) molecular models. It was hypothesized that differences in the FFV distributions could be used to explain the higher experimental O2 solubility in PEN relative to PET. The analysis showed that there was no statistical difference between the FFV distributions for O2 sized penetrants. Clustering analysis was performed based upon the tetrahedra formed by Delaunay Tessellation to examine the connectivity of free volume pockets. These results show that there is a statistically larger number of small (containing less than 10 tetrahedra/cluster and between 20-30 and #506;3 in volume) clusters in PEN. It is this difference in small clusters which provides for the 30% higher O2 solubility in PEN. The free volume of a representative high molecular weight amorphous model of Hexafluroalcohol Subsituted Polynorbornene (HFA-PNB) was also characterized in to examine the shape of the free volume cavities and to draw correlations with the mean lifetime of ortho-positronium (o-Ps) from Positron Annihilation Lifetime Spectroscopy (PALS). Delaunay Tessellation and clustering analysis indicated that the free volume clusters in high molecular weight HFA-PNB are slightly non-spherical. Correcting lifetimes for the somewhat non-spherical shape of these free volume clusters was insufficient to reproduce experimentally measured positron annihilation lifetimes because the clusters contained many tortuous connections within the clusters. Inclusion of this connectivity information does produce a more accurate estimate of the measured life times. This indicates that the o-Ps does sample many tetrahedra in these static clusters, but does not freely sample every section of these clusters.

In this PhD thesis computational methods have been employed in order to study different biologically relevant systems. In the first part of the thesis two Ebola virus proteins were studied, namely Viral Protein 24 (VP24) and Viral Protein 35 (VP35), responsible for the inhibition of the immune response . After a brief theoretical introduction to the main computational methods employed in the thesis, a study of VP35 in complex with small organic molecules is presented. These compounds are able to inhibit the interaction between VP35 and viral nucleoprotein. This study confirms the experimental findings highlighting new important key interactions between the protein the inhibitors. Moreover, an Essential Dynamics analysis points out an interesting collective motion of the apo-form that is hindered by the presence of the ligands. Afterwards, the protein-protein interaction VP24-Karyopherin (KPNA) is studied. An atomistic analysis of the interactions at the interface leads to the design of a nonapeptide with VP24 binding capability. The peptide is derived from a KPNA subsequence and could potentially inhibit the VP24-KPNA interaction. Subsequently an analysis on the pockets present on VP24 surface in different solvents is performed. Once the most promising pocket has been located, a virtual screening on a subset of ZINC database is carried out, leading to the identification of few classes of molecules potentially able to bind VP24. Finally the effect of the osmolytes on VP24 protein structure is studied, pointing out how osmoprotectants and urea have opposite effects on the protein, the former stabilizing the folded state and the latter shifting the equilibrium to the denatured state. In the second part of the manuscript the study of the interaction of an antimicrobial peptide with a lipid membrane is presented. This work was carried out in the University of Groningen under the supervision of Prof. Siewert Jan Marrink in order to deepen the Coarse Grain method.

Los surfactantes son moléculas anfifílicas, con una cabeza solvofílica y una cola solvofóbica. Cuando la concentración de surfactante en solución es suficientemente alta, las moléculas se agregan entre ellas para proteger las partes solvofóbicas del contacto con el medio. Tales agregados pueden tener forma y tamaño muy diferentes, dependiendo del surfactante y de las condiciones del sistema. La auto-organización de los surfactantes (self-assembly), debida a un compromiso energético y entrópico de su estructura molecular, es la clave que permite observar cristales líquidos muy ordenados. En presencia de un precursor inorgánico y dependiendo de las interacciones que este precursor establece con el surfactante, se puede observar la formación del material híbrido. Los materiales híbridos constituyen un paso intermedio fundamental para la síntesis de los materiales mesoporosos ordenados, los cuales se obtienen eliminando la matriz orgánica (surfactante) del substrato inorgánico.
El presente estudio tiene como principal objetivo estudiar bajo cuales condiciones los sistemas formados por un surfactante, un precursor inorgánico y un solvente, se auto-organizan para dar lugar a estructuras híbridas muy ordenadas. En particular nos proponemos individuar cuales son las características más importantes que los precursores inorgánicos deberían tener para poder observar la formación de materiales mesoporosos ordenados.
Simulaciones Monte Carlo en el colectivo canónico han sido utilizadas para analizar la agregación de los surfactantes en estructuras complejas, como micelas, cilindros organizados en forma hexagonal, o laminas, a partir de configuraciones totalmente desordenadas. Con particular interés hemos analizado el rango de condiciones que llevan a la formación de las estructuras cilíndricas, y estas mismas estructuras han sido comparadas en función de algunas importantes características morfológicas, como el tamaño de poro, el grosor de las paredes, la presencia y accesibilidad de los grupos funcionales en los poros. El modelo usado representa las moléculas de surfactante y de precursor inorgánico como cadenas de segmentos en una red tridimensional que discretiza el espacio en sitios de volumen unitario. Este modelo no entra en el detalle de las características físicas y químicas de las moléculas, pero permite reproducir su agregación en estructuras complejas en un tiempo de cálculo muy razonable. La separación de fase ha sido también evaluada recorriendo a una teoría de campo medio, la quasi-chemical theory, que, aunque no pueda predecir la formación de estructuras ordenadas, ha sido muy útil para confirmar los resultados de las simulaciones, sobretodo cuando no se observa formación de fases ordenadas.
El estudio de surfactantes distintos, uno modelado por una cadena lineal y otro con una cabeza ramificada, nos ha permitido evaluar algunas diferencias estructurales de los materiales obtenidos. La ramificación de la cabeza, que merecería un estudio más profundo del que hemos descrito en este trabajo, ha evidenciado unas interesantes consecuencias en el tamaño de los poros. Este mismo surfactante con cabeza ramificada ha sido elegido para la síntesis de agregados cilíndricos utilizados como templates en la formación, agregación, y condensación de una capa de sílica modelada a través de un modelo atomístico. En particular, hemos aislado uno de los cilindros presentes en los cristales líquidos de estructura hexagonal, y a su alrededor hemos simulado la formación de una capa de sílica utilizando un modelo atomístico. De esta forma, hemos obtenido un poro típico de una estructura mesoporosa más realista, sin necesidad de asumir una forma mas o menos cilíndrica del template, por ser este generado de la auto-agregación del surfactante.
Surfactants are amphiphilic molecules with a solvophilic head and a solvophobic tail. When the surfactant concentration in a given solution is high enough, the molecules aggregate between them to shield the solvophobic part from the contact with the solvent. Such aggregates can show very different sizes and shapes, according to the surfactant and the conditions of the system. The surfactants self-assembly, being due to an energetic and entropic compromise of their molecular structure, is fundamental to observe the formation of very ordered liquid crystals. In the presence of an inorganic precursor and depending on the interactions established between such a precursor and the surfactant, it is possible to synthesize a hybrid material. Hybrid materials are the key step for the formation of periodic ordered mesoporous materials, which can be obtained by eliminating the organic soft matter (the surfactants) from the inorganic framework. Periodic ordered mesoporous materials represent an important family of porous materials as they find a large number of applications in several industrial fields, such as separations, catalysis, sensors, etc. In the last decade, the range of potential applications has increased with the possibility of functionalizing the pore walls by incorporating organic groups during the synthesis, or with post-synthesis treatments.
In this work, we are interested in studying the formation of ordered materials when hybrid organic-inorganic precursors are used. Lattice Monte Carlo simulations in the NVT ensemble have been used to study the equilibrium phase behavior and the synthesis of self-assembling ordered mesoporous materials formed by an organic template with amphiphilic properties and an inorganic precursor in a model solvent. Three classes of inorganic precursors have been modeled: terminal (R-Si-(OEt)3) and bridging ((EtO)3-Si-R-Si-(OEt)3)) organosilica precursors (OSPs), along with pure silica precursors (Si-(OEt)4). Each class has been studied by analyzing its solubility in the solvent and the solvophobicity of the inorganic group.
At high surfactant concentrations, periodic ordered structures, such as hexagonally-ordered cylinders or lamellas, can be obtained depending on the OSPs used. In particular, ordered structures were obtained in a wider range of conditions when bridging hydrophilic OSPs have been used, because a higher surfactant concentration was reached in the phase where the material was formed. Terminal and bridging OSPs produced ordered structures only when the organic group is solvophilic. In this case, a partial solubility between the precursor and the solvent or a lower temperature favored the formation of ordered phases.
With particular interest, we have analyzed the range of conditions leaving to the formation of cylindrical structures, which have been evaluated according to the pore size distribution, the pore wall thickness, the distribution and the accessibility of the functional organic groups around the pores. The phase behavior has been also evaluated by applying the quasi-chemical theory, which cannot predict the formation of ordered structures, but was very useful to confirm the results of simulations, especially when no ordered structures were observed.
The study of the phase and aggregation behavior of two different surfactants, one modeled by a linear chain of head segments and the other modeled by a branched-head, permitted us to evaluate some structural differences of the materials obtained.

In this thesis, we report theoretical investigations ofnonlinear and nonequilibrium quantum electronic transport propertiesof molecular transport junctions from atomistic first principles.The aim is to seek not only qualitative but alsoquantitative understanding of the correspondingexperimental data. At present, the challenges to quantitativetheoretical work in molecular electronics include two most importantquestions: (i) what is the proper atomic model for the experimentaldevices? (ii) how to accurately determine quantum transportproperties without any phenomenological parameters? Our research iscentered on these questions. We have systematically calculatedatomic structures of the molecular transport junctions by performingtotal energy structural relaxation using density functional theory(DFT). Our quantum transport calculations were carried out byimplementing DFT within the framework of Keldysh non-equilibriumGreen's functions (NEGF). The calculated data are directly comparedwith the corresponding experimental measurements. Our generalconclusion is that quantitative comparison with experimental datacan be made if the device contacts are correctly determined.We calculated properties of nonequilibrium spin injection from Nicontacts to octane-thiolate films which form a molecular spintronicsystem. The first principles results allow us to establish a clearphysical picture of how spins are injected from the Ni contactsthrough the Ni-molecule linkage to the molecule, why tunnelmagnetoresistance is rapidly reduced by the applied bias in anasymmetric manner, and to what extent ab initio transporttheory can make quantitative comparisons to the correspondingexperimental data. We found that extremely careful sampling of thetwo-dimensional Brillouin zone of the Ni surface is crucial foraccurate results in such a spintronic system.We investigated the role of contact formation and its resultingstructures to quantum transport in several molecular wires and showthat interface contacts critically control charge conduction. It wasfound, for Au/BDT/Au junctions, the H atom in -SH groupsenergetically prefers to be non-dissociative after the contactformation, which was supported by comparison between computed andmeasured break-down forces and bonding energies. TheH-non-dissociated (HND) junctions give equilibrium conductances from0.054G0 (equilibrium structure) to 0.020G0 (stretchedstructure) which is within a factor of 2-5 of the measureddata. On the other hand, for all H-dissociated contact structures - whichwere the assumed structures in the literature, the conductance is atleast more than an order of magnitude larger that the experimentalvalue. The HND-model significantly narrows down thetheory/experiment discrepancy. Finally, a by-product of this work isa comprehensive pseudopotential and atomic orbital basis setdatabase that has been carefully calibrated and can be used by theDFT community at large.
Cette thèse présentera nos recherches théoriquessur les propriétés quantiques de transport électroniquedes jonctions de transport moléculaire. Cette analyse a été effectuéeà l'aide de méthodes ab initio atomiques qui sont validesdans les régimes non-linéaire et hors-équilibre.L'objectif est de rechercher non seulement une compréhensionqualitative des données expérimentales mais aussi quantitative. Les deux questions les plus importantesquant au travail théorique en électronique moléculaire sont:(i) quel est le bon modèle atomique pour simuler les dispositifsexpérimentaux? (ii) comment déterminer avec précisionles propriétés de transport quantique sans l'utilisation deparamètres phénoménologiques?Nos recherches sont centrées sur ces questions. Nous avonssystématiquement calculéles structures atomiques de jonctions moléculaires en effectuantla relaxation structurelle dans le cadre de la théorie de la fonctionnelle de la densité (DFT).Les calculs de transport quantique ont été reàlises encombinant la DFT avec les fonctions de Green hors-équilibre deKeldysh (NEGF).Les calculs sont directement comparésaux données expérimentales correspondantes. Notre conclusiongénérale est qu'un accord quantitatif entre les valeurs théoriqueset empiriques est possible si la structure atomique du contact estcorrectement déterminée.Nous avons calculé les propriétés hors-équilibred'injection de spin à travers un film d'octane-thiole en contactavec des électrodes en Ni, formant ainsi un systèmespintronique moléculaire.Les résultats obtenus par premiers principes nous fournissentune compréhension claire sur la façon dont les spins sontinjectés à partir des électrodes en Ni à la molécule par la liaisonNi-molécule. De plus, nous expliquonspourquoi la magnéto-résistance à effet tunnel décroîtrapidement avecune augmentation du potentiel électrique, et ce, de manière asymétrique.Finalement, nous démontrons que la théorie ab initiodu transport électronique est en mesure d'effectuer des comparaisonsquantitatives avec les données expérimentales.Nous avons constaté qu'un échantillonnage minutieux de la zonede Brillouin 2D de la surface du Ni est crucial afin d'obtenir desrésultats précis dans un tel système spintronique.Nous avons étudié le rôle de la formation du contact,ainsi que la structure atomique associée sur l'influence dutransport quantique dans le cas de plusieurs jonctions moléculaires.Nous démontrons que l'interface reliant les électrodes aux moléculescontrôle très sensiblement la conduction de charge.Il a été trouvé, pour les jonctions Au/BDT/Au, quel'atome d'hydrogène dans les groupes -SHpréfère énergétiquement la non-dissociation après la formationdu contact. En effet, ceci a été corroboré par la comparaison entreles donnéees calculées et mesurées des forces de rupture et desénergies de liaison.Les jonctions avec l'hydrogène non-dissocié (HND) donnentdes valeurs de conductances à l'équilibre de 0.054G$_0$ (structure d'équilibre) à0.020G$_0$ (structure étirée). Ces valeurs sontà l'intérieur d'un facteur de 2-5 aux données expérimentales actuelles.D'autre part, toutes les structures de contact H-dissociées --- quiont été les structures supposées dans la littérature --- résultenten des valeurs de conductancescalculées au moins un ordre de grandeur plus élevé que lesvaleurs empiriques.Le modèle HND réduit de manière significative l'écart entrela théorie et les expériences. Pour terminer, une conséquence de ce travail estle regroupement d'une base de données complète incluant des pseudo-potentielset des orbitales atomiques. Celle-ci a été soigneusement calibrée et est disponibleà toute la communauté DFT.

This thesis begins with a review on the topic of molecular electronics. The purpose of this review is to motivate the need for good theory to understand and predict molecular electronics behaviour. At present the most promising theoretical formalism for dealing with this problem is a combination of density functional theory and nonequilibrium Green's functions (NEGF-DFT). This formalism is especially attractive because it is an ab-initio technique, meaning that it is completely from first principles and does not require any empirical parameters. An implementation of this formalism has been developed by the research group of Hong Guo and is presented and explained here. A few other implementations which are similar but differ in some ways are also discussed briefly to highlight their various advantages and disadvantages.
One of the difficulties of ab-initio calculations is that they can be extremely costly in terms of the computing time and memory that they require. For this reason, in addition to using appropriate approximations, sophisticated numerical analysis tech niques need to be used. One of the bottlenecks in the NEGF-DFT method is solving the Poisson equation on a large real space grid. For studying systems incorporating a gate voltage it is required to be able to solve this problem with nonperiodic boundary conditions. In order to do this a technique called multigrid is used. This thesis examines the multigrid technique and develops an efficient implementation for the purpose of use in the NEGF-DFT formalism. For large systems, where it is necessary to use especially large real space grids, it is desirable to run simulations on parallel computing clusters to handle the memory requirements and make the code run faster. For this reason a parallel implementation of multigrid is developed and tested for performance. The multigrid tool is incorporated into the NEGF-DFT formalism and tested to ensure that it is properly implemented. A few calculations are made on a benzenedithiol system with gold leads to show the effect of an applied gate voltage.

Approved for public release; distribution in unlimited.
Polymers have been widely used in various engineering applications. For more than a quarter century, the materials have been utilized intensively for the binding materials for composites. The material properties of the binding materials called matrix materials play an important role for the composite material behaviors. As a result, the objective of this study was to understand the mechanical properties of polymers. In particular, the goal was to develop insights as to how a molecular level structure is connected to the bulk properties of materials assuming homogeneity. To this end, molecular dynamics was used to model and simulate the polymeric behaviors. Polymeric chains were modeled using the bead and spring model along with interacting potentials. The study examined the effects of different sizes, densities, and numbers of molecules per chain on the shear moduli of the polymers. Furthermore, some preliminary study was also conducted for metallic particle reinforced polymer composites.

Thesis (Ph. D.)--California Institute of Technology, 1994. UM #94-06,216.
Advisor names found in the Acknowledgements pages of the thesis. Title from home page. Viewed 01/15/2010. Includes bibliographical references.

Organic electronic materials, possessing conjugated π-systems, are extensively used as the active layers in organic electronic devices, where they are responsible for charge transport. In this dissertation, we employ a combination of quantum-mechanical and molecular- mechanics methods to provide insight into how molecular structure, orientation, packing, and local molecular environment influence the energetic landscape experienced by an excess charge in these organic electronic materials. We begin with an overview of charge transport in organic electronic materials with a focus on electronic polarization while discussing recent models, followed by a review of the computational methods employed throughout our investigations. We provide a bottom-up approach to the problem of describing electronic polarization by first laying the framework of our model and comparing calculated properties of bulk materials to available experimental data and previously proposed models. We then explore the effects of changing the electronic structure of our systems though perfluorination, and investigate the effects of modifying the crystalline packing through the addition of bulky functional groups while investigating how the non-bonded interactions between molecular neighbors change in different packing motifs. As interfaces are common in organic electronics and important processes such as charge transport and charge separation occur at these interfaces, we model organic-vacuum and organic-organic interfaces to determine the effect changing the environment from bulk to interface has on the electronic polarization. We first investigate the effects of removing polarizable medium adjacent to the charge carrier and then, by modeling a realistic organic- organic interface in a model solar cell, probe the environment of each molecular site at the interface to gain a more complete understanding of the complex energetic landscape. Finally, we conclude with a study of the non-bonded interactions in linear oligoacene dimers, model π-conjugated materials, to assess the impact of dimer configuration and acene length on the intermolecular interaction energy, and highlight the importance of dispersion and charge penetration to these systems.

Thesis (S.M.)--Massachusetts Institute of Technology, Dept. of Materials Science and Engineering, 2013.
Title as it appears in MIT degrees awarded booklet, September 2012: Molecular simulation of primary nucleation and growth from oriented melts in polyethylene. Cataloged from PDF version of thesis.
Includes bibliographical references (p. 59-63).
The enhancement of the primary flow-induced nucleation rate in short chain alkanes (C20 and C150) has been examined for different levels of orientation by atomistic molecular dynamics simulations. The nucleation rate has been found to change drastically by varying average molecular orientation and temperature. For example, it is possible to accelerate nucleation kinetics by three orders of magnitude at the same temperature, but varying the average level of orientation (P2 (cos [Theta])) . The size of the critical nucleus has been found to increase with the level of undercooling Tm - T decrease, consistent with the classical nucleation theory. Our atomnistic molecular dynamics simulation model is even tractable at the small levels of undercooling, thus clearly demonstrating the effects of orientation (melt anisotropy) on nucleation kinetics when thermal nucleation is expected to be negligible. Furthermore, we calculate the influence of melt anisotropy on the growth rate. As expected, the growth rate is also altered by melt anisotropy. Furthermore, the growth rate maximum always occurs at the temperature above the nucleation kinetics maximum.
by Predrag Đjurdjević.
S.M.

Thesis (Ph. D.)--Biology, Georgia Institute of Technology, 2008.
Committee Chair: Harvey, Stephen C; Committee Member: Hud, Nicholas V; Committee Member: McCarty, Nael A; Committee Member: Wartell, Roger M.

This thesis is part of a project aimed at providing a comprehensive study of different galaxy components in the core of radio-loud ETGs, and look for kinematical signatures of feeding/feedback loops that can be causally related to the presence of radio jets. For this purpose, a volume limited (z < 0.03) sample of 11 radio galaxies was selected. This sample is the target of VLT/VIMOS (warm gas and stellar components) and ALMA (molecular gas). This thesis inquires into the kinematics of molecular gas in the centre of one of the sources in the sample: NGC 3100, a FRI radio galaxy hosted by a S0 galaxy at redshift z = 0.0088. NGC 3100 was observed with APEX at 230 GHz and showed a CO(2-1) line profile (double-horned) consistent with the presence of a rotating disk. The inner region of NGC 3100 was then imaged with ALMA. As part of this thesis, ALMA data was reduced and a data cube was obtained with a beam of 1.01''x0.73'', where a CO(2-1) 230-GHz line was clearly detected. The ALMA radio continuum data, revealed the inner part of the radio jets, entirely consistent with those imaged at similar resolution with the VLA. A full analysis of the CO(2-1) line emission was made through the integrated intensity map and the integrated velocity map. The mass of the molecular gas resulted in M = 1.85 +/- 0.4 x 10^8 M_sun, consistent with what found with APEX. The CO map was compared with the distribution of dust in the inner region of the host galaxy. A nice overlap was found for the structures detected in both images. The molecular gas disk shows a complex kinematics, with some warps in the velocity field. Two programs were used to model the disk: TiRiFiC and 3D-Barolo. The comparison of their results was helpful to better understand the kinematics of the gas. The modelling confirmed initial guesses about the inclination and the position angle of the gas disk, and allowed us to derive purely rotational velocity fields as well as fields including non-circular motions.

Designing novel drugs is a complex process which requires finding molecules in a vast chemical space that bind to a specific biomolecular target and have favorable physio-chemical properties. Machine learning methods can leverage previous data and use it for new predictions helping the processes of selection of molecule candidate without relying exclusively on experiments. Particularly, deep learning can be applied to extract complex patterns from simple representations. In this work we leverage deep learning to extract patterns from three-dimensional representations of molecules. We apply classification and regression models to predict bioactivity and binding affinity, respectively. Furthermore, we show that it is possible to predict ligand properties for a particular protein pocket. Finally, we employ deep generative modeling for compound design. Given a ligand shape we show that we can generate similar compounds, and given a protein pocket we can generate potentially binding compounds.
El disseny de drogues novells es un procés complex que requereix trobar les molècules adequades, entre un gran ventall de possibilitats, que siguin capaces d’unir-se a la proteïna desitjada amb unes propietats fisicoquímiques favorables. Els mètodes d’aprenentatge automàtic ens serveixen per a aprofitar dades antigues sobre les molècules i utilitzar-les per a noves prediccions, ajudant en el procés de selecció de molècules potencials sense la necessitat exclusiva d’experiments. Particularment, l’aprenentatge profund pot sera plicat per a extreure patrons complexos a partir de representacions simples. En aquesta tesi utilitzem l’aprenentatge profund per a extreure patrons a partir de representacions tridimensionals de molècules. Apliquem models de classificació i regressió per a predir la bioactivitat i l’afinitat d’unió, respectivament. A més, demostrem que podem predir les propietats dels lligands per a una cavitat proteica determinada. Finalment, utilitzem un model generatiu profund per a disseny de compostos. Donada una forma d’un lligand demostrem que podem generar compostos similars i, donada una cavitat proteica, podem generar compostos que potencialment s’hi podràn unir.

Diffusive molecular communication (MC) is a promising strategy for the transfer of information in synthetic networks at the nanoscale. If such devices could communicate, then it would expand their cumulative capacity and potentially enable applications such as cooperative diagnostics in medicine, bottom-up fabrication in manufacturing, and sensitive environmental monitoring. Diffusion-based MC relies on the random motion of information molecules due to collisions with other molecules. This dissertation presents a novel system model for three-dimensional diffusive MC where molecules can also be carried by steady uniform flow or participate in chemical reactions. The expected channel impulse response due to a point source of molecules is derived and its statistics are studied. The mutual information between consecutive observations at the receiver is also derived. A simulation framework that accommodates the details of the system model is introduced. A joint estimation problem is formulated for the underlying system model parameters. The Cramer-Rao lower bound on the variance of estimation error is derived. Maximum likelihood estimation is considered and shown to be better than the Cramer-Rao lower bound when it is biased. Peak-based estimators are proposed for the low-complexity estimation of any single channel parameter. Optimal and suboptimal receiver design is considered for detecting the transmission of ON/OFF keying impulses. Optimal joint detection provides a bound on detector performance. The weighted sum detector is proposed as a suboptimal alternative that is more physically realizable. The performance of a weighted sum detector can become comparable to that of the optimal detector when the environment has a mechanism to reduce intersymbol interference. A model for noise sources that continuously release molecules is studied. The time-varying and asymptotic impact of such sources is derived. The model for asymptotic noise is used to approximate the impact of multiuser interference and also the impact of older bits of intersymbol interference.
Applied Science, Faculty of
Electrical and Computer Engineering, Department of
Graduate

We present realistic molecular models for microporous and templated mesoporous carbons that describe the pore morphology and topology in a realistic way. In our work on modeling microporous carbons, we have developed a Hybrid Reverse Monte Carlo method to build molecular models that match the structural data obtained from experiments and capture the correct chemisty of the carbon atoms at the local level. Our method is based on Reverse Monte Carlo with an additional energy penalty term. The presence of the energy penalty term reduces the probability of having unrealistic features such as 3 and 4 member rings, reported in many previous Reverse Monte Carlo studies of carbons, in the resultant models. Hydrogen atoms, which are normally ignored or implicitly assumed in most of the simulation studies of porous carbons, are explicitly taken into account in our method. We built models for 3 saccharose based carbons and using a ring connectivity method, that we developed, we found that the resultant models contain highly defective and convoluted graphene segments. We also simulated adsorption of argon using grand canonical Monte Carlo simulations. The simulated isosteric heats of adsorption are in good agreement with experimental results. Our results show that the inclusion of explicit hydrogen atoms in the model is necessary to reproduce the local structure and adsorption properties of real carbons. In the second part of our project, we have used pseudo mimetic methods to develop molecular models for 3 silica templated mesoporous carbons: CMK-1, CMK-3 and CMK-5. These models include pore surface roughness and morphological defects found in real materials. We performed GCMC simulation of argon at 77K and found that the shape of the experimental adsorption isotherms is reproduced well by our mesoporous carbon models.

Tremendous advances in nanoscience during the past decades have drawn a new horizon for the future of science. Many biological and structural elements such as DNA, bio-membranes, nanotubes, nanowires and thin films have been studied carefully in the past decades. In this work we target to speed up the computational methods by incorporating the structural symmetries that nanostructures have. In particular, we use the Objective Structures (OS) framework to speed up molecular dynamics (MD), lattice dynamics (phonon analysis) and multiscale methods. OS framework is a generalization of the standard idea for crystal lattices of assuming periodicity of atomic positions with a large supercell. OS not only considers the translational periodicity of the structure, but also other symmetries such as rotational and screw symmetries. In addition to the computational efficiency afforded by Objective Structures, OS provides us with more flexibility in the shape of the unit cell and the form of the external deformation and loading, comparing to using the translational periodicity. This is because the deformation and loading should be consistent in all cells and not all deformations keep the periodicity of the structures. For instance, bending and twisting cannot be modeled with methods using the structure's periodicity. Using OS framework we then carefully studied carbon nanotubes under non-equilibrium deformations. We also studied the failure mechanism of pristine and twisted nanotubes under tensile loading. We found a range of failure mechanisms, including the formation of Stone-Wales defects, the opening of voids, and the motion of atoms out of the cross-section. We also used the OS framework to make concrete analogies between crystalline phonons and normal modes of vibration in non-crystalline but highly symmetric nanostructures.

N-Nitrosodimethylamine (NDMA), which is a carcinogenic and toxic N-nitrosamine, can be found in water resources associated with a multitude of processes in various industrial facilities or merely as a by-product of water or wastewater treatment. Therefore, the removal of NDMA from drinking water represents an important human safety and public health concern. The present paper presents a density functional theory study of NDMA adsorption in all-silica MFI, Na-ZSM-5 and H-ZSM-5 zeolites. The stability of NDMA inside the zeolite pores was investigated by calculating the amount of energy released during adsorption. Various configurations of adsorbed NDMA to the zeolites were investigated, predominantly at the intersection of straight and sinusoidal channels. The strength of the adsorption energies followed the order H-ZSM5 > Na-ZSM-5 > all-silica MFI. NDMA has a dipole moment and the strongest binding of NDMA occurred through the interactions of the negatively charged O atom of the molecule to positive atoms of the zeolite. Similar calculations were performed for water adsorption in these three zeolites. The adsorption energy of water to these three structures followed the order Na-ZSM5 > H-ZSM-5 > all-silica MFI. We also incorporated van der Waals corrections in the simulations, which had the effect of stabilizing NDMA within the zeolite channels, but did not significantly change the relative stability of the different adsorption geometries. It was concluded that H-ZSM-5 is the best choice to remove NDMA because it is strong enough to adsorb NDMA and it is not too strong in adsorption of water molecules.

It is impossible to determine the amount of money that is spent every replacing products damaged from wear, but it is safe to assume that it is in the millions of dollars. With metallic materials, liquid lubricants are often used to prevent wear from materials rubbing against one another. However, with polymeric materials, liquid lubricants cause swelling, creating an increase in friction and therefore increasing the wear. Therefore, a different method or methods to mitigate wear in polymers should be developed. For better understanding of the phenomenon of wear, scratch resistance testing can be used. For this project, classic molecular dynamics is used to study the mechanics of nanometer scale scratching on amorphous polymeric materials. As a first approach, a model was created for polyethylene, considering intramolecular and intermolecular interactions as well as mass and volume of the CH2 monomers in a polymer chain. The obtained results include analysis of penetration depth and recovery percentage related to indenter force and size.

One of the major challenges faced by the pharmaceutical industry is to accelerate the product innovation process and reduce the time-to-market for new drug developments. This involves billions of dollars of investment due to the large amount of experimentation and validation processes involved. A computational modeling approach, which could explore the design space rapidly, reduce uncertainty and make better, faster and safer decisions, fits into the overall goal and complements the product development process. Our research focuses on the early preclinical stage of the drug development process involving lead selection, optimization and candidate identification steps. Our work helps in screening the most favorable candidates based on the biopharmaceutical and pharmacokinetic properties. This helps in precipitating early development failures in the early drug discovery and candidate selection processes and reduces the rate of late-stage failures, which is more expensive. In our research, we successfully integrated two well-known models, namely the drug release model (dissolution model) with a drug transport model (compartmental absorption and transit (CAT) model) to predict the release, distribution, absorption and elimination of an oral drug through the gastrointestinal (GI) tract of the human body. In the CAT model, the GI tract is envisioned as a series of compartments, where each compartment is assumed to be a continuous stirred tank reactor (CSTR). We coupled the drug release model in the form of partial differential equations (PDE's) with the CAT model in the form of ordinary differential equations (ODE's). The developed model can also be used to design the drug tablet for target pharmacokinetic characteristics. The advantage of the suggested approach is that it includes the mechanism of drug release and also the properties of the polymer carrier into the model. The model is flexible and can be adapted based on the requirements of the clients. Through this model, we were also able to avoid depending on commercially available software which are very expensive. In the drug discovery and development process, the tablet formulation (oral drug delivery) is an important step. The tablet consists of active pharmaceutical ingredient (API), excipients and polymer. A controlled release of drug from this tablet usually involves swelling of the polymer, forming a gel layer and diffusion of drug through the gel layer into the body. The polymer is mainly responsible for controlling the release rate (of the drug from the tablet), which would lead to a desired therapeutic effect on the body. In our research, we also developed a molecular design strategy for generating molecular structures of polymer candidates with desired properties. Structure-property relationships and group contributions are used to estimate the polymer properties based on the polymer molecular structure, along with a computer aided technique to generate molecular structures of polymers having desired properties. In greater detail, we utilized group contribution models to estimate several desired polymer properties such as grass transition temperature (Tg), density (ρ) and linear expansion coefficient (α). We subsequently solved an optimization model, which generated molecular structures of polymers with desired property values. Some examples of new polymer repeat units are - [CONHCH₂ - CH₂NHCO]n -, - [CHOH - COO]n -. These repeat-units could potentially lead to novel polymers with interesting characteristics; a polymer chemist could further investigate these. We recognize the need to develop group contribution models for other polymer properties such as porosity of the polymer and diffusion coefficients of water and drug in the polymer, which are not currently available in literature. The geometric characteristics and the make-up of the drug tablet have a large impact on the drug release profile in the GI tract. We are exploring the concept of tablet customization, namely designing the dosage form of the tablet based on a desired release profile. We proposed tablet configurations which could lead to desired release profiles such as constant or zero-order release, Gaussian release and pulsatile release. We expect our work to aid in the product innovation process.
Ph. D.

L’une des application les plus remarquables de l’effet tunnel est le microscope à effet tunnel (STM) qui permet de cartographier spatialement et énergétiquement la répartition des électrons à la surface des matériaux avec une résolution nanométrique. Des avancées récentes permettent en outre d’exploiter la pointe du STM comme une source d’excitation locale des matériaux. Le travail de thèse présenté dans ce manuscrit vise à décrire et à modéliser les phénomènes impliqués lors d’une telle excitation. Nous présentons une modélisation des spectres d’absorption de molécule de phthalocyanine reposant sur des surfaces dans le cadre de la théorie de la fonctionnelle de densité dépendante du temps (TD-DFT). Nous montrons que l’analyse spectroscopique des transitions entre l’état fondamental et les états excités de la molécule permet de caractériser son état de contrainte. Nous mettons également en évidence une variété de spectres d’excitation selon la localisation de l’excitation de la molécule. Nous discutons la possibilité d’exploiter ce phénomène pour caractériser les transports d’énergie inter-moléculaire
One of the most remarkable applications of the tunnel effect is the Scanning Tunneling Microscope (STM), allowing to get the spatially and energetically map distribution of electrons on the surface of materials with nanometric resolution. Recent advances make it possible to exploit the tip of the STM as a source of local excitation of materials. The work presented in this manuscript aims to describe and model the phenomena involved in such excitation process. We present a modeling of the absorption spectra of phthalocyanine molecules lying on surfaces within the framework of the time-dependent density functional theory (TD-DFT). We show that spectroscopic analysis of the transitions between the ground state and the excited states of the molecule allows to characterize the stress inside the molecule. We also highlight a variety of excitation spectra depending on the location of the excitation of the molecule. We discuss the possibility of exploiting this phenomenon to characterize inter-molecular energy transport

Mestrado em Química
Esta dissertação explora o mundo nanoscópico de pequenos agregados onde as pontes de hidrogénio têm um papel preponderantes usando métodos quânticos ab-initio. No capítulo introdutório, a área da química computacional é apresentada e algumas noções teóricas referentes aos métodos ab-initio, discutidas. No Capítulo 2, o desempenho de vários níveis de teoria é avaliado através do estudo de pequenos agregados de água. O capítulo 3 discute a influência dos critérios de optimização no resultado deste processo, alertando para erros comuns. No Capítulo 4, hidratos gasosos de ácido trifluoroacético (TFA), nas formas dissociada e não-dissociada, são apresentados. Um número mínimo de 4 moléculas de água é necessário para induzir a transferência do protão do TFA para a rede de moléculas de água adjacente . No entanto, 5 moléculas de água são necessárias para que o agregado dissociado se torne mais estável que o seu análogo não dissociado. O Capítulo 5 propõe um novo esquema para o cálculo ab-initio de valores de pKa. Este esquema serve-se de hidratos de ácido microsolvatado, nas formas dissociada e não dissociada, em modelo de solvatação contínuo, para calcular a energia livre de dissociação em solução. Para o conjunto de espécies testadas, incluindo 10 ácidos carboxílicos, 1 amina e 2 aminoácidos, o erro médio absoluto é 1.11, o declive experimental 1.2 e o coeficiente de correlacção 0.92, o que indica um nível de exactidão aceitável.
This dissertation concerns the study of small hydrogen bonded systems through the use of quantum mechanical ab-initio methods. In the introductory chapter, the field of computational chemistry is presented and some basic theoretical notions concerning ab-initio methods are discussed. In Chapter 2, the performance of various levels of theory is assessed through the study of small water clusters. Chapter 3 discusses the influence of optimization criteria in the outcome of the optimization procedure, warning against common pitfalls. In Chapter 4, gas-phase hydrates of trifluoroacetic acid (TFA), in both dissociated and undissociated forms, are presented. A minimum of 4 water molecules is necessary to induce proton transfer from TFA to the neighboring water molecule network. However, 5 water molecules are needed to render the dissociated hydrate more stable than its undissociated counterpart. Chapter 5 proposes a new scheme for the ab-initio calculation of pKa values. It uses microsolvated acid hydrates, in both dissociated and undissociated forms, within a continuum solvation model, to calculate the dissociation free energy in solution. For the data set used, including 10 carboxylic acids, 1 amine and 2 aminoacids, the mean usigned error (MUE) of calculated pKa values is 1.11, the experimental slope 1.2 and the correlation 0.92, which denotes a reasonable level of accuracy.

Thesis (Ph.D.)--Chemical Engineering, Georgia Institute of Technology, 2009.
Committee Chair: Peter Ludovice; Committee Member: Amyn Teja; Committee Member: Arthur Ragauskas; Committee Member: William Koros; Committee Member: Yulin Deng.

Thesis (M.S. in Chemistry)--S.M.U., 2005.
Title from PDF title page (viewed Nov. 19, 2009). Source: Masters Abstracts International, Volume: 46-06, page: 3277. Adviser: John D. Buynak. Includes bibliographical references.

This research investigated the adsorption mechanisms of hydrophobic chlorinated contaminants in mineral micropores and on iron metal surfaces. Activated adsorption and desorption of trichloroethylene (TCE) in mineral micropores was studied using experimental and molecular modeling techniques. Adsorption of TCE on a silica gel adsorbent was measured using a frontal analysis chromatography technique at atmospheric and elevated fluid pressures. The results showed that the increase in pressure was able to rapidly induce the formation of a desorption resistant fraction. Grand Canonical Monte Carlo (GCMC) modeling was used to elucidate the nature of water and TCE behavior within silica micropores. TCE adsorption was energetically most favorable in pores that were minimally large enough to accommodate one TCE molecule. A molecular level study of the interactions between hydrophobic chlorinated contaminants and sediments was performed. GCMC simulations were preformed to investigate water and TCE adsorption in slit micropores confined by charged and uncharged silica surfaces. Gas-phase single-sorbate simulations with water or TCE were performed as well as mixture simulations of bulk water containing TCE at 1% of its saturation concentration. Aqueous-phase TCE at a concentration equal to 1% of its saturation concentration was able to completely displace adsorbed water in uncharged pores. In highly hydrophilic pores, TCE at this concentration was able to displace up to 50% of the adsorbed water. Metallic iron filings are becoming increasingly utilized as reactive agents for reductive dechlorination of solvents in contaminated groundwaters. This research also used molecular modeling to study chemical adsorption of TCE and PCE to iron surfaces. Quantum mechanical calculations were performed to determine the thermodynamic favorability and resulting structures for chemical adsorption of TCE and PCE to iron surfaces. Molecular mechanics modeling was used to study the effects of atomic hydrogen on the thermodynamic favorability for chemically adsorbed TCE and PCE. Because TCE and PCE react with iron surfaces, their adsorption to iron cannot be investigated experimentally. This makes molecular modeling approaches a useful complement to experimental investigations of chemical reaction phenomena.

Après son séjour au sein d'un réacteur nucléaire, le combustible contient encore une quantité importante de matières valorisables qu’il est intéressant de récupérer, à savoir le plutonium et l'uranium.La récupération et la purification de ces actinides sont réalisées à l'aide d'un procédé hydrométallurgique appelé PUREX (Plutonium Uranium Recovering by Extraction), basé sur les techniques d'extraction liquide-liquide. Ce procédé nécessite l'utilisation d'une molécule spécifique pour extraire Pu et U, le phosphate de tri-n-butyle TBP. Les N, N-dialkylamides (monoamides) sont considérés comme une famille alternative d'agents d'extraction au TBP en raison de leur forte capacité d'extraction des éléments Pu(IV) et U(VI). De plus, ces molécules présentent des caractéristiques intéressantes, telles que la forte dépendance des propriétés d'extraction (coefficient de distribution et sélectivité) à la structure des ligands ainsi qu'aux conditions chimiques. Afin de proposer le meilleur design de molécule d'extraction pour les futures usines de retraitement de combustible, il est crucial de comprendre la relation entre la structure et la capacité d'extraction. Cependant, le caractère radioactif de ces éléments combinés à leur complexité chimique rendent les études expérimentales de ces phases complexes. Par conséquent, la modélisation moléculaire semble être la solution idéale pour obtenir de nouvelles informations à l'échelle moléculaire.Dans la première partie de cette thèse, une étude quantique relativiste scalaire utilisant la théorie fonctionnelle de la densité a été réalisée pour déterminer l'influence de la nature de la chaîne alkyle monoamides sur la stabilité des complexes Pu(IV). Il a été possible de mieux comprendre la forte influence de la structure amide sur l'extraction du plutonium. Pour les deux complexes d’amide-plutonium-nitrate étudiés (complexes de sphères interne et externe), il a été constaté que l'introduction d'un groupe alkyle volumineux du côté carbonyle a un impact majeur sur l'énergie de complexation. L'impact de la polarité de la solution a été également étudié et jugé significatif.Dans le but d'étudier des systèmes plus réalistes, contenant des monoamides avec des longues chaînes alkyles, des actinides et des contre-ions, et d'aller au-delà de l'image statique de géométries optimisées au niveau QM/DFT avec des simulations de dynamique moléculaire classique, nous avons développé des champs de force polarisable pour les molécules de solvant (alcanes et monoamides) ajustés uniquement sur des calculs de chimie quantique.L'approche ab initio retenue pour le paramétrage ainsi que le champ de force résultant et détaillés nous permettent d'obtenir des propriétés macroscopiques comparables aux données expérimentales (thermodynamiques et structurales). L'excellent accord nous permet d'avoir confiance quant à la précision des prédictions réalisées sur les systèmes pures de monoamides. Enfin, les résultats préliminaires de simulations des mélanges monoamides-dodécane (DEHiBA/dodécane et DEHBA/dodécane) sont présentés
The nuclear fuel after its dwell time in reactor still bears a substantial amount of recoverable U and Pu. The recovery and purification of these actinides is achieved using a hydro-metallurgical process known as PUREX (Plutonium Uranium Recovering by EXtraction). Based on Liquid-Liquid extraction techniques, this process requires the use of a specific molecule to extract Pu and U, the tri-n-butylphosphate TBP. N,N-dialkylamides (monoamides) are regarded as an alternative family of extractants to TBP, as they are well-known for their strong extraction ability of Pu(IV) and U(VI) elements. In addition to this, they show some interesting features, such as, the strong dependence of the extraction properties (distribution coefficient and selectivity) on the ligands structure as well as chemical conditions. In order to propose the best extracting molecule design for future fuel reprocessing plants, it is crucial to understand the relationship between the structure and the extraction ability. However, the radioactivity of these elements combined with their chemical complexity make the study of these phases experimentally a real challenge. Hence, molecular modeling appears to be the golden solution for getting new insights on this issue.In the first part of this thesis, a relativistic density functional theory study was performed to investigate the influence of the monoamides alkyl chain nature on the relative stability of Pu(IV) complexes. It was possible to reach a better understanding of the strong influence of amide structure on plutonium extraction. For both investigated amide-plutonium-nitrate complexes (inner and outer-sphere complexes), it was found that the introduction of a bulky alkyl group on the carbonyl side has a major impact on the complexation energy. The impact of the polarity of the solution was also investigated and found to be significant.In the second part, within the aim of studying more realistic systems, i.e systems containing long alkyl chains monoamides, heavy elements and other counter ions, and to go beyond the static picture of QM/DFT optimized geometries with molecular dynamics simulations, we have developed a consistent polarizable FF model for the solvent molecules (alkanes, monoamides) based solely on quantum chemical calculations. The chosen ab initio parameterization approach as well as the final force field are presented. Then, the results of molecular dynamics simulations were compared to available experimental macroscopic thermodynamics and structural properties, and show an excellent agreement, making the predictions of properties of pure monoamides reliable. Finally, preliminary MD simulations results for monoamides-dodecane mixtures (DEHiBA/dodecane and DEHBA/dodecane) are presented

In this work, epoxy resins were modeled using all atom representations in nanoscale simulation boxes. Tetrafunctional epoxy and corresponding multifunctional amine were chosen as model materials. Algorithms of constructing interconnected network structures were invented developed to properly account for the chemical structures and computational cost. Monomers were generated in diamond lattice and crosslinked to model complex epoxy multifunctional network. The initial configurations were relaxed and equilibrated using molecular dynamics and suitable force field. Physical, thermal and mechanical properties resulting from equilibrated simulation box are in good agreement with experimental results. Possible impact of chemical degradation was studied by adopting oxidation and hydrolysis algorithms. Mechanism of degradation was based on bonds reaction probability and chemical structures of epoxies. Both oxidation and hydrolysis were found to decrease materials performances by reducing number of crosslinking points. Elastic modulus of materials was directly related to crosslinking density. Interfaces between two types of epoxies were constructed to study interactions at interfaces. Covalent bonds linking two components play an important role in interfacial strength. Free volume calculation helps to identify and monitor nucleation of crazes and voids within materials. It was found voids and cracks prefer to initiate and grow at 2 interfaces and lead to failures. Additional compatibilizer layers can improve overall composite performances by preventing void growth at interfaces. Diffusion pattern of water in epoxy resins was studied by tracking displacement of single molecules during certain time intervals. The characteristic of water diffusion in epoxies was interpreted by free volume theory. Reactive force field was introduced to study thermal degradation behavior of epoxy resins. Number of molecules and variation of different types of covalent bonds during heating processes were tracked and analyzed to uncover the degradation mechanism of epoxy resins.

Computational chemistry lies at the intersection of chemistry, physics, mathematics, and computer science, and can be used to explain the behavior of atoms and molecules, as well as to augment experiment. In this work, computational chemistry methods are used to predict structural and energetic properties of small molecules, i.e. molecules with less than 60 atoms. Different aspects of computational chemistry are examined in this work. The importance of examining the converged orbitals obtained in an electronic structure calculation is explained. The ability to more completely describe the orbital space through the extrapolation of energies obtained at increasing quality of basis set is investigated with the use of the Sapporo-nZP-2012 family of basis set. The correlation consistent Composite Approach (ccCA) is utilized to compute the enthalpies of formation of a set of molecules and the accuracy is compared with the target method, CCSD(T,FC1)/aug-cc-pCV∞Z-DK. Both methodologies are able to produce computed enthalpies of formation that are typically within 1 kcal mol-1 of reliable experiment. This demonstrates that ccCA can be used instead of much more computationally intensive methods (in terms of memory, processors, and time required for a calculation) with the expectation of similar accuracy yet at a reduced computational cost. The enthalpies of formation for systems containing s-block elements have been computed using the multireference variant of ccCA (MR-ccCA), which is designed specifically for systems that require an explicit treatment of nondynamical correlation. Density functional theory (DFT) has been used for the prediction of the structural properties of a set of lanthanide trihalide molecules as well as the reaction energetics for the rearrangement of diphosphine ligands around a triosmium cluster.

The recent determination of several crystal structures of voltage-gated ion channels has catalyzed computational efforts of studying these remarkable molecular machines that are able to conduct ions across biological membranes at extremely high rates without compromising the ion selectivity. Starting from the open crystal structures, we have studied the gating mechanism of these channels by molecular modeling techniques. Firstly, by applying a membrane potential, initial stages of the closing of the channel were captured, manifested in a secondary-structure change in the voltage-sensor. In a follow-up study, we found that the energetic cost of translocating this 310-helix conformation was significantly lower than in the original conformation. Thirdly, collaborators of ours identified new molecular constraints for different states along the gating pathway. We used those to build new protein models that were evaluated by simulations. All these results point to a gating mechanism where the S4 helix undergoes a secondary structure transformation during gating. These simulations also provide information about how the protein interacts with the surrounding membrane. In particular, we found that lipid molecules close to the protein diffuse together with it, forming a large dynamic lipid-protein cluster. This has important consequences for the understanding of protein-membrane interactions and for the theories of lateral diffusion of membrane proteins. Further, simulations of the simple ion channel antiamoebin were performed where different molecular models of the channel were evaluated by calculating ion conduction rates, which were compared to experimentally measured values. One of the models had a conductance consistent with the experimental data and was proposed to represent the biological active state of the channel. Finally, the underlying methods for simulating molecular systems were probed by implementing the CHARMM force field into the GROMACS simulation package. The implementation was verified and specific GROMACS-features were combined with CHARMM and evaluated on long timescales. The CHARMM interaction potential was found to sample relevant protein conformations indifferently of the model of solvent used.
At the time of the doctoral defense, the following paper was unpublished and had a status as follows: Paper 3: Manuscript.

En aquesta tesi es presenta la Teoria del Gradient de Densitat (DGT) [van der Waals, 1894] combinada amb una Equació d'Estat (EoS) amb base molecular la soft-SAFT [Blas i Vega, 1997], per predir de forma simultània el comportament de l’equilibri de fases i les propietats interfacials dels fluids industrials més significatius, lluny i prop de la regió crítica. Com qualsevol mètode estadístic ha de ser validat amb les dades de simulació existents en la literatura, ja que se suposa que els mètodes més precisos per descriure el comportament de la matèria en si mateixa són els mètodes de simulació. Una vegada que el model ha estat validat aquest es pot aplicar per calcular simultàniament l’equilibri de fases i les propietats interfacials dels fluids industrials seleccionats. El model ha estat provat amb fluids moleculars com els Lennard-Jones i amb fluids purs: n-alcans, alcohols, líquids iònics, refrigerants, nitrils, aigua, diòxid de carboni i d’altres molècules inorgàniques. Un pas endavant s'ha fet en el càlcul de les propietats interfacials de les mescles binàries d'interès industrial de compostos associants i no associants també d'una manera predictiva, evitant la necessitat d'ajustos addicionals. Les prediccions obtingudes amb la DGT + soft-SAFT s'han comparat amb les dades de simulació molecular i les dades experimentals disponibles, a més s'han obtingut tres correlacions pel paràmetre influència en funció del nombre de carbonis dels compostos. Aquestes s'han proposat pels alcans lleugers, els alcohols i una família de líquids iònics. Quan aquest pas s'aconsegueix, els models estan disponibles mitjançant paràmetres transferibles per predir el comportament d'altres compostos de la mateixa família, sèrie o amb propietats físiques similars. Per obtenir més informació sobre els fenòmens interfacials dels líquids iònics, mitjançant l'acoblament de la soft-SAFT amb la DGT hem predit la temperatura, la densitat i la pressió crítica dels líquids iònics més comuns. Aquesta és la primera vegada que una EoS és acoblada a la DGT per calcular simultàniament la tensió interfacial a temperatures elevades, mentre que captura el comportament asimptòtic prop de la regió crítica. A més, les propietats de la superfície, com l'entropia i l'entalpia de superfície, s'han derivat a partir de la dependència de la tensió superficial amb la temperatura, els resultats trobats estan d'acord qualitativament amb els valors reportats a la literatura tant de dades de simulació com de les contribucions experimentals. Finalment un estudi dels diferents perfils de densitat, incloent fluids purs i diferents tipus de mescles binàries, s'ha dut a terme per completar la revisió dels fenòmens interfacials. Així mateix es presenten perfils de densitat amb fenòmens d'absorció i desorció en la interfase, els quals són de gran importància en els processos de transport i producció. Aquesta tesis ha demostrat que l'acoblament d'una Equació d’Estat amb base molecular, la soft-SAFT [Blas i Vega, 1997] amb la teoria del Gradient de Densitat [van der Waals, 1894], és un mètode elegant per predir les propietats interfacials de compostos associants i no associants, així com de les seves mescles, amb un esforç computacional molt minse.
En esta tesis doctoral se presenta la Teoría del Gradiente de Densidad (DGT) [van der Waals, 1894] combinada con una Ecuación de Estado (EoS) con base molecular la soft-SAFT [Blas y Vega, 1997], para predecir de forma simultánea el comportamiento del equilibrio de fases y las propiedades interfaciales de los fluidos industriales más representativos. Como todo método estadístico tiene que ser validado con los datos de simulación existentes en la literatura, ya que se supone que los métodos más precisos para describir el comportamiento de la materia en sí misma son los métodos de simulación. Una vez que el modelo ha sido validado, se ha aplicado para calcular simultáneamente los equilibrios de fases y las propiedades interfaciales de los fluidos industriales más representativos, tanto lejos como cerca de la región crítica. El modelo ha sido probado con fluidos moleculares como Lennard-Jones y con fluidos puros: n-alcanos, alcoholes, líquidos iónicos, refrigerantes, nitrilos, agua, dióxido de carbono y otras moléculas inorgánicas. Un paso adelante se ha hecho el cálculo de las propiedades interfaciales de las mezclas binarias de interés industrial de compuestos asociantes y no asociantes también de una manera predictiva, evitando la necesidad de ajustes adicionales. Las predicciones obtenidas con la DGT + soft-SAFT se han comparado con los datos de simulación molecular y los datos experimentales disponibles, además se han obtenido tres correlaciones del parámetro influencia como una función del número de carbono de los compuestos. Estas se han propuesto para los alcanos ligeros, los alcoholes de cadena corta y una familia de líquidos iónicos. Una vez que este paso se logra, los modelos se pueden utilizar con parámetros transferibles para predecir el comportamiento de otros compuestos de la misma familia, serie o con propiedades físicas similares. Para obtener más información sobre los fenómenos interfaciales de los líquidos iónicos, mediante el acoplamiento de la soft-SAFT con la DGT hemos predicho la temperatura, la densidad y la presión crítica de los líquidos iónicos más comunes. Esta es la primera vez que una EoS es acoplada a la DGT para calcular simultáneamente la tensión interfacial a temperaturas elevadas, mientras que captura el comportamiento asintótico cerca de la región crítica. Además, las propiedades de la superficie, como la entropía y la entalpía de superficie, se han derivado a partir de la dependencia de la tensión superficial con la temperatura, los resultados encontrados están de acuerdo cualitativamente con los valores reportados en la literatura tanto de datos de simulación como de las contribuciones experimentales. Finalmente un estudio de los distintos perfiles de densidad, incluyendo fluidos puros y diferentes tipos de mezclas binarias, se ha llevado a cabo para completar la revisión de los fenómenos interfaciales. Asimismo se presentan algunos perfiles de densidad con fenómenos de absorción y desorción en la interfase, dado que estos son de gran relevancia tanto para el control de procesos como para el transporte de gases y fluidos. En esta tesis se demuestra que el acoplamiento de una Ecuación de Estado con base molecular, la soft-SAFT [Blas y Vega, 1997] con la teoría del Gradiente de Densidad [van der Waals, 1894], se presenta como un método elegante para predecir las propiedades interfaciales de compuestos asociantes y no asociantes, así como de sus mezclas, con un mínimo esfuerzo computacional.
In this PhD thesis, the Density Gradient Theory (DGT) [van der Waals, 1894] combined with a molecular-based Equation of State (EoS); the soft-SAFT [Blas and Vega, 1997], was applied to simultaneously predict the phase behavior and the interfacial properties of industrial relevant fluids. As the equation is based in statistical mechanics, its approximations and assumptions were assessed against simulation data for the same underlying model. Once the model was validated, it was applied to simultaneously calculate the phase equilibria and the interfacial properties of some of the most representative industrial fluids, far from and close to the critical region. In particular, the model has been tested with molecular model fluids as Lennard-Jones chains, giving excellent agreement with simulation data, and then applied to different pure fluids, including: n-alkanes, light alkanols, ionic liquids, refrigerants, nitriles, water, carbon dioxide and ammonia, among others. A step forward has been done by calculating the interfacial properties of the binary mixtures of industrial interest, with associating and nonassociating compounds, in a predictive manner, avoiding the need of additional fitting, and providing information for systems for which there is not experimental data available. In addition, three correlations of the influence parameter as a function of the carbon number have been proposed for the light alkanes, light alkanols and one ionic liquid family, allowing for predictions of properties of compounds not included in the fitting procedure. A final novel contribution of this Thesis work is the prediction of the critical temperature, density and pressure of the most common used ionic liquids by using soft-SAFT coupled with the DGT. This is to our best knowledge the first time that an EoS is coupled to the DGT to calculate simultaneously the interfacial tension at elevated temperatures, while capturing the asymptotic behavior as the critical region is approached. Moreover, the surface properties, such as surface entropy and surface enthalpy, have been derived from the surface tension dependence on temperature, finding a very good agreement with the values reported in the literature from simulation and experimental contributions. Finally, a throughout study of the different density profiles, including single fluids and different binary mixtures, has been carried out to complete the description of the interfacial phenomena. Absorption and desorption density profiles are also presented given their importance in transport and process control. The work developed here demonstrates that coupling an accurate molecularbased equation of state for phase properties, the soft-SAFT equation, with a simple and accurate theory for interfacial properties, the Density Gradient Theory, is a reliable tool to simultaneously predict the phase and interfacial properties of nonassociating and associating compounds, as well as their mixtures with a very slight computational effort and great accuracy.