Dissertations / Theses on the topic 'Molecular simulation techniques'

Molecular dynamics trajectories that sample from a Gibbs, or canonical, distribution can be generated by introducing a modified Hamiltonian with additional degrees of freedom as described by Nose [46]. Although this method has found widespread use in its time re-parameterized Nose-Hoover form, the lack of a Hamiltonian, and the need to 'tune' thermostatting parameters has limited, its use compared to stochastic methods. In addition, since the proof of the correct sampling is based on an ergodic assumption, thermostatting small of stiff systems often does not given the correct distributions unless the Nose-Hoover chains [43] method is used, which inherits the Nose-Hoover deficiencies noted above. More recently the introduction of the Hamiltonian Nose-Poincare method [11], where symplectic integrators can be used for improved long term stability, has renewed interest in the possibility of Hamiltonian methods which can improve dynamical sampling. This class of methods, although applicable to small systems, has applications in large scale systems with complex chemical structure, such as protein-bath and quantum-classical models.;For Nose dynamics, it is often stated that the system is driven to equilibrium through a resonant interaction between the self-oscillation frequency of the thermostat variable and a natural frequency of the underlying system. By the introduction of multiple thermostat Hamiltonian formulations, which are not restricted to chains, it has been possible to clarify this perspective, using harmonic models, and exhibit practical deficiencies of the standard Nose-chain approach. This has led to the introduction of two Hamiltonian schemes, the Nose-Poincare chains method and the Recursive Multiple Thermostat (RMT) method. The RMT method obtains canonical sampling without the stability problems encountered with chains with the advantage that the choice of Nose mass is independent of the underlying system.

Nanotechnology is a multidisciplinary branch of science and technology that involves a widerange of different fields such as chemistry, materials science, physics or chemical engineeringwhose goal is the production of new functional materials and devicesthrough the control of their organization at the atomic and molecular scale.
Nanotechnology has jumped from research laboratories to our daily life and today all theprogresses made in this field have been translated into direct applications in different fields being electronics and computer science and biomedicine, where the most striking advances have beendone.
What differences nanotechnology from traditional chemistry and physics can be summarized inthree points: (i) Analysis and control of the matterat the atomic and molecular level focusing in individual atoms; (ii) the appearance of novel physical properties because of the nanoscopicdimensions; (iii) the possibility of generating new complex functional systems with novelproperties.
Modeling and theory are becoming vital to designing and improving nanodevices. The intrinsicnature of nano and supramolecular scale that involves tens, hundreds and thousands of atomsmakes computational chemistry the perfect ally to design new devices and predict their properties. Computational chemistry provides the perfect tools to describe the electronic structureand the dynamic behavior, as well as the properties derived from them, through quantummechanics and classical mechanics formalisms.
The suitability of such techniques in the design and improvement of nanodevices as well as theprediction of their properties is clearly proven throughout the four blocks in which this thesis isdivided:
· Nanotubes based on natural peptide sequences
Nanotubes have gained extensive interest because of their applicability in different fieldsranging from medicine to electronics. Among nanotubes, those based on natural peptidesequences taken from some natural proteins with a tubular or fibrillar motif are gaining a
broad attention because of their high biocompatibility, the possibility of adding functionalitiesby tuning them and their potentiality to self-assemble. The enhancement of the ability to retain the tubular geometry of such structures can be achieved by substituting targeted amino acids located in the more flexible parts of the nanoconstruct by synthetic amino acids withlow conformational flexibility providing a larger rigidity to the overall structure.
· Dendronized polymers
Dendronized polymers are a specific kind of macromolecule structure that consists of a linearpolymeric backbone where dendritic units are attached regularly leading to a highly branchedthree-dimensional architecture. This fact provides dendronized polymers the peculiarity of the coexistence within the same macromolecule of three topological regions: (i) the internalbackbone; (ii) the dendron region around the backbone and (iii) the external surface. Thesemolecules have a wide range of applications in different fields such as biomedical engineering, host-guest chemistry or catalysis.
· Theoretical study of ð-conjugated systems
Conducting polymers are polymers bearing a characteristic polyconjugated nature which makethem electronic conductors. In particular thiophene-based conducting polymers have been widely studied because of their electric and nonlinear optical properties, excellent environmentalstability and relatively low cost of production. Due to the crucial role played by the electronicstructure of these systems in their relevant properties, a good knowledge of it is a key factor todesign and improve new conducting polymers. To achieve this goal QM calculations suitperfectly to get accurate estimates of such properties.
· Molecular actuators and sensors based on conducting polymers
Both experimental and computational research in nanoactuators and nanosensors are widelyreported in the literature. Among them, those based in conducting polymers are flourishingbecause of their great transport properties, electrical conductivity or rate of energy migrationwhich provide amplified sensitivity in nanosensors and a rapid response in nanoactuators. In thissense electron-rich thiophene-based oligomers and polymers combined with versatilecalix[4]arenes units are presented in the present thesis. Calix[4]arenes are synthetic macrocyclic molecules consisting of four phenol or anisole rings connected via methylene bridges that canhost different guest molecules leading to conformational rearrangement of the whole device making it useful to be employed as a sensor or actuator.

Dissertation (Ph.D.) -- Worcester Polytechnic Institute.
Keywords: Activated carbons; Hydrogen cyanide; Methyl ethyl ketone; Adsorption; Mercury; Monte-Carlo; Solvents; Molecular simulations; Zeolites; Water; Methanol; Nanopores. Includes bibliographical references (leaves 147-150).

Els materials porosos s'utilitzen àmpliament en moltes branques de la ciència i tecnologia modernes com la catàlisi, la separació de mescles, la purificació de fluids i la fabricació de membranes. Per a que els sòlids porosos puguin aplicar-se amb èxit cal disposar d'una caracterització precisa de la superfície i de les propietats estructurals, així com també una bona comprensió del comportament físico-químic dels fluids dins dels porus. Alguns materials, com les zeolites, tenen estructures poroses ben definides, però d'altres, com els òxids porosos, carbons i vidres de porus controlat, són bastant amorfs. Per això, un tema clau i, sovint, complicat, és la caracterització adequada d'
aquests tipus de materials. Durant molts anys, l'adsorció de gasos s'ha emprat per estudiar les propietats de sòlids porosos, degut a que és un mètode ràpid, simple i que proporciona prou informació. Es van desenvolupar molts mètodes per extraure dades sobre la porositat i les propietats de la superfície de materials a partir d'isotermes d'adsorció. En les dues últimes dècades, amb l'ajuda dels ordinadors, cada cop més i més ràpids, l'ús de les tècniques de modelat molecular ha anat guanyant rellevància. En aquest context, l'objectiu general d'aquest treball de tesi és desenvolupar eines a escala molecular emprant la mecànica estadística i aplicant-la a la caracterització
de materials adsorbents.
Després d'una breu introducció en el tema (capítol 1), en el capítol 2 presentem una revisió de la metodologia bàsica emprada en aquest treball. En el capítol 3 hem implementat la teoria funcional de la densitat de mesures fonamentals o FMT (de l'anglès, Fundamental-Measured density functional theory), publicada per Kierlik i Rosinberg, per descriure l'adsorció de molècules Lennard-Jones en porus cilíndrics. Pel que sabem, aquest és el primer cop que la teoria s'aplica a la geometria cilíndrica. L'exactitud de la teoria en predir isotermes d'adsorció i perfils de densitat de partícules es compara amb simulacions Monte Carlo en el col·lectiu gran canònic per un rang ample de mides de porus. Aquesta comparació mostra que la concordança és molt bona en tots els casos. Addicionalment, s'ha aplicat la teoria a l'adsorció en porus plans per estudiar la influència de la geometria del porus en aquest fenomen. Els resultats indiquen que el confinament de la geometria cilíndrica introdueix diferències significatives en la forma de les isotermes d'adsorció i els perfils de densitat. Aquestes diferències són rellevants a l'hora de caracteritzar materials porosos. Els resultats indiquen que té lloc un comportament per capes en el porus cilíndric més petit que s'ha considerat, mentre que l'adsorció en un porus pla de la mateixa grandària necessita un potencial químic molt més alt per aconseguir una adsorció significant. A mida que el diàmetre del porus augmenta, la influència de la geometria es fa cada cop menys important, encara que es pot observar una certa desviació en la transició de condensació capil·lar. Addicionalment, per porus més amples, obtenim una adsorció multicapa amb condensació capil·lar a potencials químics alts, amb el mateix comportament qualitatiu observat en ambdues geometries. Quan el diàmetre assoleix el límit on els efectes de curvatura ja no són rellevants, el comportament quantitatiu del porus cilíndric es redueix al mateix que el del porus pla. La formació d'una capa fina adsorbent en mides de porus intermèdies i grans sembla correspondre a una transició de fase termodinàmica de segon ordre, per al rang de paràmetres utilitzat i les condicions termodinàmiques estudiades. No obstant, els resultats semblen indicar una interrelació entre aquest comportament i la transició pre-mullada (de la paraula anglesa prewetting) que s'observa en geometries semi-infinites, especialment al voltant del punt final crític de la línia pre-mullada. L'efecte del confinament és molt important en aquest comportament crossover (de pas). De la comparació de càlculs FMT amb resultats de la teoria funcional de la densitat no local, concloem que la FMT és una eina excel·lent per a l'estudi del comportament de fluids en geometries cilíndriques.
En el capítol 4 s'explica com hem aplicat la FMT juntament amb un mètode de regularització per estimar la distribució de mides de porus o PSD (de l'anglès, Pore-Size Distribution) de vidres porosos model. Hem escollit aquest material perquè va ser desenvolupat mitjançant tècniques de modelat molecular, i es pot comparar directament amb la teoria utilitzada en aquest treball. Un avantatge addicional d'aquests materials model, enfront els experimentals, és que, en el primer cas, la mida i forma dels porus són ben conegudes, així com també la posició dels àtoms en la superfície, esdevenint així un material perfecte per comprovar l'exactitud dels mètodes de caracterització teòrica disponibles. Com que hi ha diferents solucions de l'equació integral d'adsorció compatibles amb la isoterma d'adsorció experimental, i diversos factors poden amagar els defectes del model molecular, hem realitzat la caracterització d'una forma sistemàtica: primer hem comprovat l'exactitud de la FMT i el model de porus independent per predir les isotermes d'adsorció "experimentals" utilitzant la PSD ja coneguda per als materials. Això s'ha efectuat amb porus individuals plans i cilíndrics. En segon lloc, un cop la isoterma d'adsorció va ser reconstruïda amb èxit, vam invertir la isoterma d'adsorció integral amb un procediment de regularització. L'exactitud del mètode d'inversió s'ha comprovat també abans d'estimar la PSD de materials diferents. En últim lloc, un cop demostrat que el mètode és correcte, l'hem utilitzat per estimar la PSD de quatre materials. També hem estudiat la influència d'escollir alguns valors particulars de paràmetres moleculars per les interaccions fluid-fluid i sòlid-fluid en el comportament adsorbent d'aquests sistemes. Hem obtingut que el model de porus independent és adequat per als quatre materials investigats en aquest treball. La geometria plana sembla representar millor que la geometria cilíndrica el comportament adsorbent global. Pel que fa a la PSD obtinguda amb el nostre procediment, s'observa que les distribucions obtingudes mitjançant la inversió de la integral estan en millor concordança amb les distribucions geomètriques que les calculades amb el mètode Barrett-Joyner-Halenda (BJH). El locus del pic està situat a la mateixa mida de por, i tots ells són unimodals, mentre que les distribucions BJH mostren un màxim localitzat sistemàticament a porus més petits, estimant per sota la PSD del material, i no són unimodals. En quan a la geometria dels porus individuals que formen el material podem dir que, encara que la PSD és més ampla que les geomètriques, l'adsorció que es prediu per un conjunt de porus plans individuals està en un acord quasi quantitatiu amb la isoterma d'adsorció experimental.
Finalment, en el capítol 5 exposem com hem caracteritzat tres mostres diferents de g­alúmina, una d'elles sense tractament i les altres dues calcinades en un forn durant unes hores a 823 i 1023K. Per fer-ho hem mesurat isotermes d'adsorció de nitrogen a 77.35K en un equip Micromeretics ASAP 2000. A més, hem aprofitat les PSD's proporcionades pel programari de l'equip emprant el mètode BJH. Hem calculat isotermes teòriques mitjançant l'aproximació FMT. Hem invertit les equacions integrals d'adsorció amb el mètode de regularització i, finalment, hem obtingut les PSD's per les tres mostres d'alúmina, i les corresponents isotermes d'adsorció pels tres materials. D'aquesta forma hem observat la influència de la calcinació de l'alúmina en la seva PSD. A més, hem comprovat l'exactitud del mètode FMT/de regularització de manera sistemàtica. Quan comparem les PSD's obtingudes amb les corresponents distribucions BJH, hem verificat que, en els dos primers casos (alúmina no tractada i alúmina calcinada a 823K), el mètode BJH estima per sota la mida dels porus, proporcionant una PSD desviada cap a mides més petites. En el cas de l'alúmina calcinada a 1,023K, en la que el procés de sinterització produeix que els porus més petits desapareguin, afavorint els més grans, les PSD's del mètode BJH i les PSD's de la FMT/regularització són molt semblants. Amb això es corrobora el fet conegut de que el mètode BJH és força acurat en la regió macroporosa. Finalment, hem predit la isoterma d'adsorció d'un fluid diferent (età) a una altra temperatura (333K), en un dels materials caracteritzats (alúmina no tractada), amb l'ànim d'establir la robustesa de la PSD obtinguda. La concordança obtinguda mostra que és possible utilitzar aquest mètode de caracterització i extrapolar els resultats a altres condicions, mentre s'empri un nombre suficient de mides de porus per calcular la isoterma desitjada, i els paràmetres d'interacció sòlid-fluid es triïn adequadament.
Los materiales porosos se utilizan en muchas ramas de la ciencia y la tecnología, por ejemplo, se usan como catalizadores, en la separación de mezclas, en la purificación de fluidos, y en la fabricación de membranas. Su aplicación adecuada requiere de la caracterización precisa de sus propiedades superficiales y estructurales, además del conocimiento del comportamiento fisicoquímico de los fluidos cuando se encuentran dentro de los poros. Algunos materiales, como las zeolitas, tienen estructuras porosas bien definidas, pero otros en cambio (óxidos porosos, carbones, vidrios porosos con tamaño controlado) son bastante amorfos. Por lo tanto, una caracterización correcta de los materiales porosos es un área de estudio muy importante, la cual en algunos casos es una tarea sencilla pero en la mayoría no. Durante muchos años la adsorción de gases ha sido empleada para estudiar las propiedades de sólidos porosos, dado que es bastante fácil, simple y se puede obtener mucha información. Se han desarrollado muchos métodos para interpretar los datos experimentales y determinar la porosidad, las propiedades superficiales y la distribución de los tamaños de los poros de los materiales a partir de las isotermas de adsorción. En las dos últimas décadas, con la ayuda de las computadoras cada vez más rápidas, se ha extendido mucho el uso las técnicas de la mecánica estadística para realizar esta tarea. En este contexto, el objetivo general de esta tesis consiste en desarrollar herramientas a escala molecular utilizando la mecánica estadística para la caracterización de materiales adsorbentes.
Después de una breve introducción en el tema (capítulo 1), el capítulo 2 está dedicado a hacer una revisión de la metodología básica empleada en este trabajo. En el capítulo 3 hemos implementado la teoría funcional de la densidad de medidas fundamentales (FMT, del inglés Fundamental-Measure density functional theory) de Kierlik y Rosinberg para describir la adsorción de moléculas Lennard-Jones dentro de poros cilíndricos. Hasta donde sabemos, ésta es la primera vez que esta teoría es aplicada a geometría cilíndrica. La exactitud de la teoría en predecir las isotermas de adsorción y los perfiles de la densidad es verificada por comparación con simulaciones Monte Carlo en el colectivo Gran Canónico para un amplio intervalo de tamaños de poros, observándose una buena concordancia en todos los casos. Adicionalmente, la teoría ha sido aplicada a la adsorción en poros planos para estudiar la influencia de los poros en esta propiedad. Los resultados indican que el confinamiento de la geometría cilíndrica introduce diferencias significativas en la forma de las isotermas de adsorción y de los perfiles de la densidad. Estas diferencias son relevantes para la caracterización de los materiales porosos. Nuestros resultados indican que un comportamiento de formación de capa tiene lugar en el poro cilíndrico, mientras que la adsorción en un poro plano del mismo tamaño necesita un potencial químico mucho más alto para alcanzar una adsorción significativa. Cuando el tamaño de poro se incrementa, la influencia de la geometría se vuelve menos importante, pero aún se observa un cierto desplazamiento del lugar en el cual se da la transición de la condensación capilar. Adicionalmente, para poros más anchos, tenemos formación de multicapas con condensación capilar a potenciales químicos altos, observándose el mismo comportamiento cualitativo en ambas geometrías. Cuando el diámetro alcanza el límite en donde los efectos de la curvatura ya no son relevantes, el comportamiento cuantitativo de los poros cilíndricos y de los planos es muy similar. La formación de una fina película adsorbida a tamaños de poro grandes e intermedios parece corresponder a una transición de fase termodinámica de segundo orden, para el intervalo de parámetros usado y a las condiciones termodinámicas estudiadas. Sin embargo, los resultados encontrados parecen indicar que existe una relación entre este comportamiento y el de una transición de pre-mojado observada en geometrías semi-infinitas, especialmente en la vecindad del punto final crítico de la línea de pre-mojado. El efecto del confinamiento es muy importante en este comportamiento de transición. A partir de la comparación de los cálculos hechos con FMT y los hechos con la teoría funcional de la densidad no-local, concluimos que la FMT es una excelente herramienta para el estudio del comportamiento de los fluidos en geometrías cilíndricas confinadas.
En el capítulo 4 hemos aplicado la FMT en combinación con un método de regularización para estimar la distribución de tamaños de poros (PSD, del inglés Pore-Size Distribution) de materiales modelo que imitan a los vidrios porosos. Hemos elegido este material en particular porque fue desarrollado con técnicas de modelado molecular, y se puede hacer una comparación directa con la teoría aquí usada. Una ventaja adicional de estos materiales modelo, con respecto a los materiales reales, es que en este caso la forma y tamaño de los poros se conoce exactamente, además de que se sabe la posición de los átomos en la superficie, convirtiéndolo en un material ideal para verificar la exactitud de los métodos de caracterización teóricos disponibles. Dado que existen varias soluciones de la ecuación integral de adsorción compatibles con la isoterma de adsorción experimental, y que varios factores pueden ocultar los defectos del modelo molecular, hemos hecho la caracterización de una manera sistemática: primero hemos probado la exactitud de la FMT y del modelo de poros independientes para predecir las isotermas de adsorción "experimentales" usando la PSD geométrica ya conocida para estos materiales. Esto ha sido hecho tanto con los poros cilíndricos como con los planos. En segundo lugar, una vez que la isoterma de adsorción fue reconstruida, invertimos la isoterma integral de adsorción con un procedimiento de regularización. La exactitud del método de inversión ha sido verificado antes de estimar la PSD de los diferentes materiales. Finalmente, una vez que se ha establecido que el método es correcto, lo usamos para estimar las PSD's de estos cuatro materiales. Hemos estudiado también la influencia de elegir diferentes valores de los parámetros moleculares para la interacción fluido-fluido y para la sólido-fluido en el comportamiento de adsorción en estos sistemas. Los resultados indican que el modelo de poros independientes es adecuado para los cuatro materiales aquí investigados. La geometría plana parece representar el comportamiento de adsorción global mejor que la cilíndrica. En cuanto a lo que las PSD's obtenidas con nuestro procedimiento se refiere, las distribuciones resultantes a través de la inversión de la integral presentan una mejor concordancia con las distribuciones geométricas que las calculadas con el método Barrett-Joyner-Halenda (BJH). El máximo del pico está localizado en el mismo tamaño de poro, y las distribuciones son unimodales, mientras que las BJH's muestran un máximo sistemáticamente localizado a poros más pequeños, subestimando las PSD's del material, y éstas no son unimodales. Respecto a la geometría de los poros individuales que conforman el material, se puede decir, a pesar de que las PSD's son más dispersas que las geométricas, que la adsorción predicha por una colección de poros planos individuales tiene una concordancia casi cuantitativa con la isoterma de adsorción experimental.
Finalmente, en el capítulo 5 hemos caracterizado tres muestras diferentes de g­alúmina, una de ellas sin ningún tratamiento, y las otras dos calcinadas en un horno durante varias horas a 823 y a 1,023K. Para ello hemos medido isotermas de adsorción de nitrógeno a 77.35K en un equipo Micromeritics ASAP 2000. Adicionalmente, hemos usado las PSD's calculadas con el método BJH que proporciona el software del mismo equipo experimental para comparar. Hemos calculado las isotermas teóricas utilizando la FMT. Hemos invertido las ecuaciones integrales de adsorción con el método de regularización y, finalmente, hemos obtenido las PSD's para las tres muestras de alúmina, y las correspondientes isotermas de adsorción. De esta manera hemos podido observar la influencia de la calcinación de la alúmina en su PSD. Más aún, hemos probado la exactitud del método combinado FMT/Regularización de una manera sistemática. Cuando hemos comparado las PSD's obtenidas con las correspondientes BJH's, hemos verificado que en los dos primeros casos (alúmina sin tratamiento y alúmina calcinada a 823K) el método BJH subestima el tamaño de los poros, dando PSD's desplazadas a tamaños de poros más pequeños. En el caso de la alúmina calcinada a 1,023K, en la cual el proceso de sinterización ha producido la desaparición de los poros más pequeños en beneficio de los grandes, las PSD's BJH y las PSD's FMT/Regularización son muy similares. Con esto corroboramos el hecho conocido de que el método BJH es bastante exacto en la región de los macroporos. Para terminar, hemos predicho la isoterma de adsorción de un fluido diferente (etano) a una temperatura también diferente (333K) en uno de los materiales caracterizados (alúmina sin tratar) con la idea de comprobar sí la PSD obtenida es transferible a otras condiciones o no. La concordancia observada muestra que es posible usar este método de caracterización y extrapolar los resultados a otras condiciones, procurando que se utilice un número suficiente de tamaños de poro diferentes para calcular la isoterma deseada, y se elijan bien los parámetros de interacción sólido-fluido.
Porous materials are widely used in many branches of modern science and technology, such as catalysis, separation of mixtures, purification of fluids and fabrication of membranes. A successful application of porous solids requires a precise characterization of their surface and structural properties, as well as a good understanding of the physical and chemical behavior of fluids inside the pores. Some materials, such as zeolites, have well defined porous structures, but others, such as porous oxides, carbons and controlled-porous glasses, are quite amorphous. Therefore, a proper characterization of this kind of materials is an important topic, and more often than not, a complicated one. For many years, gas adsorption has been used to study properties of porous solids, since it is fast, simple and informative. Many methods were developed to extract information about porosity and surface properties of materials from adsorption isotherm data. In the last two decades, with the aid of the increasingly faster computers, the use of molecular modeling techniques has been gaining relevance. In this context, the general objective of this thesis is to develop tools at the molecular level using statistical mechanics for the characterization of adsorbent materials.
After a brief introduction on the topic (chapter 1), chapter 2 is devoted to a review of the basic methodology employed in this work. In chapter 3 we have implemented the Fundamental-Measure density functional theory (FMT) due to Kierlik and Rosinberg to describe the adsorption of Lennard-Jones molecules in cylindrical pores. To our best knowledge, this is the first time that this theory is applied to a cylindrical geometry. The accuracy of the theory in predicting adsorption isotherms and density profiles is checked by comparison with Grand Canonical Monte Carlo simulations for a wide range of pore sizes, showing very good agreement in all cases. In addition, the theory has been applied to the adsorption in slit-like pores to study the influence of the pore geometry on this property. The results indicate that the confinement of the cylindrical geometry introduces significant differences in the shape of the adsorption isotherms and density profiles. These differences are relevant for the characterization of porous materials. Our results indicate that a layering behavior takes place in the smallest cylindrical pore considered, while the adsorption in a planar pore of the same size needs a much higher chemical potential to achieve a significant adsorption. As the pore size increases, the influence of the geometry becomes less important, although a certain shift in the capillary condensation transition can still be observed. Additionally, for wider pores, we obtain multilayer adsorption with capillary condensation at high chemical potentials, with the same qualitative behavior observed for both geometries. When the diameter size reaches the limit where the curvature effects are not of further relevance, the cylindrical pores reduce to the same quantitative behavior of the slit-like pores. The formation of a thin adsorbed layer at intermediate and large pore sizes seems to correspond to a second order thermodynamic phase transition, for the range of parameters used and the thermodynamic conditions studied. However, the results found seem to indicate some relationship between this behavior and the prewetting transition observed in semi-infinite geometries, especially in the vicinity of the critical end point of the prewetting line. The effect of the confinement is very important in this crossover behavior. From the comparison of Fundamental-Measure density functional theory calculations versus non-local density functional theory results, we conclude that the FMT is an excellent tool for the study of the behavior of fluids in confined
cylindrical geometries.
In chapter 4 we have applied the FMT in conjunction with a regularization method to estimate the pore-size distribution (PSD) of model porous glasses. We have chosen this particular material because it was developed with molecular modeling techniques, and a direct comparison can be made with the theory used here. An additional advantage of these model materials, versus experimental ones, is that in this case the size and shape of the pores is well known, as well as the position of the atoms in the surface, making it a perfect material to check the accuracy of the theoretical characterization methods available. Since there are several solutions of the adsorption integral equation compatible with the experimental adsorption isotherm, and several factors can hide defects of the molecular model, we have done the characterization in a systematic manner: we have first checked the accuracy of the FMT and the independent pore model for predicting the "experimental" adsorption isotherms using the geometrical PSD already known for the materials. This has been done with individual cylindrical and slit-like pores. Secondly, once the adsorption isotherm was successfully reconstructed, we inverted the integral adsorption isotherm with a regularization procedure. The accuracy of the inversion method has also been checked before estimating the PSD of the different materials. Finally, once the method has been proved to be correct, we used it to estimate the PSD of four materials. We have also studied the influence of choosing different values of molecular parameters for the fluid-fluid and the solid-fluid interaction on the adsorption behavior of these systems. We have obtained that the independent pore model is adequate for the four materials investigated here. The slit-like geometry seems to represent the overall adsorption behavior better than the cylindrical geometry. As far as the PSD obtained with our procedure is concerned, the distributions obtained by inversion of the integral are in better agreement with the geometrical distributions than the ones calculated with the Barrett-Joyner-Halenda (BJH) method. The locus of the peak is at the same pore size, and all of them are unimodal, while the BJH distributions show a maximum systematically located at smaller pores, underestimating the PSD of the material, and they are not unimodal. Regarding the geometry of the individual pores that form the material, we can say that, although the PSD is broader than the geometrical ones, the adsorption predicted by a collection of individual slit-like pores is in almost quantitative agreement with the "experimental" adsorption isotherm.
Finally, in chapter 5 we have characterized three different samples of g­alumina, one of them without treatment and the others two calcined in a furnace during several hours at 823 and 1,023K. For this we have measured adsorption isotherms of nitrogen at 77.35K in a Micromeritics ASAP 2000 apparatus. Additionally, we have used the PSD's provided by the software of the experimental equipment using the BJH method. We have calculated theoretical isotherms by the FMT approach. We have inverted the adsorption integral equations with the regularization method and, finally, we have obtained the PSD's for our three samples of alumina, and the corresponding adsorption isotherms. In this way we have observed the influence of the calcination of alumina on its PSD. Moreover, we have tested the accuracy of the FMT/Regularization method in a systematic way. When we compared the PSD's obtained with the corresponding BJH distributions, we verified that in the two first cases (untreated alumina and alumina calcined at 823K) the BJH method underestimated the size of the pores, giving PSD's shifted to smaller sizes. In the case of alumina calcined at 1,023K, in which the sintering process has produced the disappearance of the smallest pores, favoring the wider ones, the BJH PSD's and the FMT/regularization PSD's perform very similar. With this, we corroborated the known fact that the BJH method is quite accurate in the macroporous region. Finally, we have predicted an adsorption isotherm of a different fluid (ethane) at a different temperature (333K) in one of the characterized materials (untreated alumina) with the aim of establishing the robustness of the PSD obtained. The agreement obtained shows that it is possible to use this characterization method and extrapolate the results at other conditions, provided that a enough number of different pore sizes are used to calculate the desirable isotherm, and the solid-fluid interaction parameters are well chosen.

La drépanocytose est une maladie génétique qui se caractérise par des globules rouges en forme de faucille. Chez les personnes atteintes de drépanocytose, ces globules rouges (GR) adhèrent à l’endothélium vasculaire et provoquent ainsi une vaso-occlusion. Ce phénomène s’explique par la surexpression de la protéine Lutheran (Lu) à la surface des globules rouges falciformes qui se lie fortement à la Laminine (Ln) 511/521 exprimée par l’endothélium vasculaire enflammé. Le but de cette étude est d’identifier un inhibiteur d’interaction protéine-protéine (PPI) qui possède une forte probabilité de liaison à Lu afin d’inhiber l’interaction Lu-Ln 511/521. Un criblage virtuel de 1 295 678 composés ciblant la protéine Lu a été réalisé. La validation préalable d’un protocole de scoring a été envisagée sur la protéine CD80 qui présente un site de liaison avec des caractéristiquestopologiques et physico-chimiques similaires au site de liaison prédit sur Lu ainsi que plusieurs ligands avec des constantes d’affinité connues. Ce protocole contient différentes étapes de sélection basées sur les affinités calculées (scores), des simulations de dynamique moléculaire et les propriétés moléculaires. Un protocole de scoring fiable a été validé sur CD80 avec le programme de docking DOCK6 et les fonctions de scoring XSCORE et MM-PBSA ainsi qu’avec la méthode decalcul FMO. L’application de ce protocole sur Lu a permis d’obtenir deux ligands validés par des tests in vitro qui font l’objet d’un dépôt de brevet. La fonction de scoring XSCORE a permis d’identifier neuf autres ligands qui semblent aussi être des candidats prometteurs pour inhiber l’interaction Lu-Ln 511/521
Drepanocytosis is a genetic blood disorder characterized by red blood cells that assume an abnormal sickle shape. In the pathogenesis of vaso-occlusive crises of sickle cell disease, red blood cells bind to the vascular endothelium and promote vaso-occlusion. At the surface of these sickle red blood cells, the overexpressed protein Lutheran (Lu) strongly interacts with the Laminin (Ln) 511/521.The aim of this study was to identify a protein-protein interaction (PPI) inhibitor with a highprobability of binding to Lu for the inhibition of the Lu-Ln 511/521 interaction. A virtual screening was performed with 1 295 678 compounds that target Lu. Prior validation of a robust scoring protocol was considered on the protein CD80 because this protein has a binding site with similar topological and physico-chemical characteristics and it also has a series of ligands with known affinity constants. This protocol consisted of multiple filtering steps based on calculated affinities (scores), molecular dynamics simulations and molecular properties. A robust scoring protocol was validated on the protein CD80 with the docking program DOCK6 and the scoring functions XSCORE and MM-PBSA and also with the FMO method. This protocol was applied to the protein Lu and we found two compounds that were validated by in vitro studies. The protection of these ligands by a patent is under process. Nine other compounds were identified by the scoring functionXSCORE and seem to be promising candidates for inhibiting the Lu-Ln 511/521 interaction

Elemental analysis of substances made of heavy elements and detection of light elements in heavy matrices are difficult by means of photon transmission techniques. Neutrons have been used in this work, taking unique advantage of their absorption and scattering properties, to detect the structure of industrial and biological objects made of strongly-neutron scattering or absorbing materials, or to study objects combining of high and low neutron cross section materials. The most convenient matrices and impurities amenable to neutron inspection were searched by obtaining expressions for minimum detectable mass and length fraction of elements in an object. Formulae to calculate the minimum required number of neutrons to detect an impurity in a matrix have also been developed. The optimum sample thickness to be investigated with a minimum number of neutrons is likewise derived. Calculations have been carried out for the minimum detectable mass fraction of hydrogen in a number of sample matrices of industrial interest and of elements in a water matrix highlighting the differences with photon attenuation measurements. Results are presented for three neutron energies cold (0.001 eV), thermal (0.025 eV), and fast (14 MeV); concentrations in the parts per million range are demonstrated. Fast neutrons were used because of their high penetration ability, in order to study bulk industrial and biological samples and for their adequacy in detection of light elements such as H, C, N and O in large objects. An attempt to simulate fast neutron transmission tomographs of biological samples was made using the MORSE-CGA Monte-Carlo code. The code was used to calculate transmission of multienergetic U-235 fast fission neutron source in a complex geometry for industrial and biological applications. A fast neutron collimator for radiography, a collimator for brain tomography and a tomography chamber were simulated to design a technique to estimate the effect of scattered neutrons in practical tomography. The macroscopic cross section and mean free path of neutrons for the media of the heterogeneous matrices were also obtained by using microscopic cross sections of elements from the DLC-100G package. Using a multienergetic source provided an opportunity to determine the optimum neutron energy for examination of objects. The analysis required establishing a technique to calculate the fraction of neutrons in each energy group for the 100 group structure of the DLC-100G package. Finally the simulated neutron tomographic images were reconstructed by using the neutron transmission data for different angles of the object, and reconstructing them by the filtered back projection technique. In non-destructive evaluation of medical organs by fast neutron simulation tomography the simulated tomography of prototype biological objects were able to distinguish brain in skull, bone-marrow in bone and bone in soft tissue with good contrast up to 0.42. These results are valuable to identify developing cystic lesions and daughter cyst within the marrow vascular spaces, solid bony tumors, aberrant masses in the facial bone, tumor in spine or other bone marrow abnormalities. In studying component characterisation of industrial objects non-destructively by fast neutron tomography a 3mm diameter duct containing engine-oil was detected at 40 cm depth inside an aluminium combustion engine with a remarkable contrast of 0.35. The minimum detectable mass of oil in aluminium for an optimum neutron energy was 0.1mg/g with a similar result for iron.

El objetivo de esta tesis se centra en el estudio de la transferencia de electrones (ET), una de las reacciones más simples y cruciales en bioquímica. Para dichos procesos, obtener información directa de los factores que lo promueves, asi como del camino de transferencia electronica, no es una tarea trivial. Dicha información a un nivel de conocimiento detallado atómico y electrónico, sin embargo, es muy valiosa en términos de una mejor comprensión del ciclo enzimático, que podría conducir, por ejemplo, a un diseño más eficaz de inhibidores. El objetivo principal de esta tesis es el desarrollo de una metodología para el estudio cuantitativo de la ET en los sistemas biológicos. En este sentido, hemos desarrollado un nuevo método para obtener el camino de transferencia electrónico, llamado QM/MM e-­‐ Pathway, que se puede aplicar en sistemas complejos con ET de largo alcance. El método se basa en una búsqueda sucesiva de residuos importantes para la ET, utilizando la modificación de la región quantica en métodos mixtos QM/MM, y siguiendo la evolución de la densidad de espín dentro de la zona de transferencia. Hemos demostrado la utilidad y la aplicabilidad del algoritmo en el complejo P450cam/Pdx, identificando el papel clave de la Arg112 (en P450cam) y del Asp48 (en Pdx), ambos conocidos en la literatura. Además de obtener caminos de ET, hemos cuantificado su importancia en términos del acoplamiento electrónico entre el dador y aceptor para los diferentes caminos. En este sentido, se realizaron dos estudios de la influencia del solvente y de la temperatura en el acoplamiento electrónico para sistemas modelo oligopéptidos. Ambos estudios revelaron que los valores del acoplamiento electrónico fluctúan fuertemente a lo largo de las trayectorias de dinámica molecular obtenidas, y el mecanismo de transferencia de electrones se ve ampliamente afectado por el espacio conformacional del sistema. La combinación del QM/MM e-­‐pathway y de los cálculos de acoplamiento electronico fueron utilizados finalmente para investigar la ET en el complejo CCP/Cytc. Nuestros hallazgos indican el papel fundamental del Trp191 en localizar un estadio intermedio para la transferencia electronica, así como el camino ET principal que incluye Ala194, Ala193, Gly192 y Trp191. Ambos hallazgos fueron confirmados a través de la literatura. Los resultados obtenidos para el muestro de manios de ET, junto con su evaluación a través de cálculos de acoplamiento electrónico, sugieren un enfoque sencillo y prometedor para investigar ET de largo alcance en proteínas.
The focus of this PhD thesis lies on electron transfer (ET) processes, belonging to the simplest but most crucial reactions in biochemistry. Getting direct information of the forces driving the process and the actual electron pathway is not a trivial task. Such atomic and electronic detailed information, however, is very valuable in terms of a better understanding of the enzymatic cycle, which might lead, for example, to more efficient protein inhibitor design. The main objective of this thesis was the development of a methodology for the quantitative study of ET in biological systems. In this regard, we developed a novel approach to map long-­‐range electron transfer pathways, called QM/MM e-­‐Pathway. The method is based on a successive search for important ET residues in terms of modifying the QM region following the evolution of the spin density of the electron (hole) within a given transfer region. We proved the usefulness and applicability of the algorithm on the P450cam/Pdx complex, indicating the key role of Arg112 of P450cam and Asp48 of Pdx for its ET pathway, both being known to be important from the literature. Besides only identifying the ET pathways, we further quantified their importance in terms of electronic coupling of donor and acceptor incorporating the particular pathway residues. Within this regard, we performed two systematic evaluations of the underlying reasons for the influence of solvent and temperature onto electronic coupling in oligopeptide model systems. Both studies revealed that electronic coupling values strongly fluctuate throughout the molecular dynamics trajectories obtained, and the mechanism of electron transfer is affected by the conformational space the system is able to occupy. Combining both ET mapping and electronic coupling calculations, we finally investigated the electron transfer in the CcP/Cytc complex. Our findings indicate the key role of Trp191 being the bridge-­‐localized state of the ET as well as the main pathway consisting of Ala194, Ala193, Gly192 and Trp191 between CcP and Cytc. Both findings were confirmed through the literature. Moreover, our calculations on several snapshots state a nongated ET mechanism in this protein complex. The methodology developed along this thesis, mapping ET pathways together with their evaluation through electronic coupling calculations, suggests a straightforward and promising approach to investigate long-­‐range ET in proteins.

Doctor of Philosophy
Department of Biochemistry and Molecular Biophysics
Jianhan Chen
Intrinsically disordered proteins (IDPs) are functional proteins that lack stable tertiary structures under physiological conditions. IDPs are key components of regulatory networks that dictate various aspects of cellular decision-making, and are over-represented in major disease pathways. For example, about 30% of eukaryotic proteins contain intrinsic disordered regions, and over 70% of cancer-associated proteins have been identified as IDPs. The highly heterogeneous nature of IDPs has presented significant challenge for experimental characterization using NMR, X-ray crystallography, or FRET. These challenges represent a unique opportunity for molecular mod- eling to make critical contributions. In this study, computer simulations at multiple scales were utilized to characterize the structural properties of unbound IDPs as well as to obtain a mechanistic understanding of IDP interactions. These studies of IDPs also reveal significant limitations in the current simulation methodology. In particular, successful simulations of biomolecules not only require accurate molecular models, but also depend on the ability to sufficiently sample the com- plex conformational space. By designing a realistic yet computationally tractable coarse-grained protein model, we demonstrated that the popular temperature replica exchange enhanced sampling is ineffective in driving faster reversible folding transitions for proteins. The second original contribution of this dissertation is the development of novel simulation methods for enhanced sampling of protein conformations, specifically, replica exchange with guided-annealing (RE-GA) method and multiscale enhanced sampling (MSES) method. We expect these methods to be highly useful in generating converged conformational ensembles.

This thesis addresses the sampling problem in a high-dimensional space, i.e., the computation of averages with respect to a defined probability density that is a function of many variables. Such sampling problems arise in many application areas, including molecular dynamics, multiscale models, and Bayesian sampling techniques used in emerging machine learning applications. Of particular interest are thermostat techniques, in the setting of a stochastic-dynamical system, that preserve the canonical Gibbs ensemble defined by an exponentiated energy function. In this thesis we explore theory, algorithms, and numerous applications in this setting. We begin by comparing numerical methods for particle-based models. The class of methods considered includes dissipative particle dynamics (DPD) as well as a newly proposed stochastic pairwise Nosé-Hoover-Langevin (PNHL) method. Splitting methods are developed and studied in terms of their thermodynamic accuracy, two-point correlation functions, and convergence. When computational efficiency is measured by the ratio of thermodynamic accuracy to CPU time, we report significant advantages in simulation for the PNHL method compared to popular alternative schemes in the low-friction regime, without degradation of convergence rate. We propose a pairwise adaptive Langevin (PAdL) thermostat that fully captures the dynamics of DPD and thus can be directly applied in the setting of momentum-conserving simulation. These methods are potentially valuable for nonequilibrium simulation of physical systems. We again report substantial improvements in both equilibrium and nonequilibrium simulations compared to popular schemes in the literature. We also discuss the proper treatment of the Lees-Edwards boundary conditions, an essential part of modelling shear flow. We also study numerical methods for sampling probability measures in high dimension where the underlying model is only approximately identified with a gradient system. These methods are important in multiscale modelling and in the design of new machine learning algorithms for inference and parameterization for large datasets, challenges which are increasingly important in "big data" applications. In addition to providing a more comprehensive discussion of the foundations of these methods, we propose a new numerical method for the adaptive Langevin/stochastic gradient Nosé-Hoover thermostat that achieves a dramatic improvement in numerical efficiency over the most popular stochastic gradient methods reported in the literature. We demonstrate that the newly established method inherits a superconvergence property (fourth order convergence to the invariant measure for configurational quantities) recently demonstrated in the setting of Langevin dynamics. Furthermore, we propose a covariance-controlled adaptive Langevin (CCAdL) thermostat that can effectively dissipate parameter-dependent noise while maintaining a desired target distribution. The proposed method achieves a substantial speedup over popular alternative schemes for large-scale machine learning applications.

Molecular dynamics simulations are computer simulations of the physical movements of atoms and molecules, and the interactions between them. In the particular cases we focus on (pharmaceutical drug design and enzymatic catalysis), molecular dynamics simulations predict the binding mode and binding affinity of a small molecule (the drug) with a biomolecule. Providing new tools and techniques to visualize and to interact with these simulations is crucial in understanding them. Throughout the development of this thesis, we have proposed new techniques to improve both the rendering quality and speed of different molecular representation models. Moreover, we developed a novel system to analyze docking simulations and the underlying interactions between the atom groups.
La dinámica molecular es una técnica de simulación por ordenador la cual intenta reproducir el movimiento de átomos y moléculas, y su interacción. En el marco de esta tesis (diseño de fármacos y cinética enzimática), las simulaciones de dinámica molecular predicen el modo y afinidad del acoplamiento entre una pequeña molécula (el fármaco) con una biomolecula. Proporcionar nuevas herramientas y técnicas para visualizar e interactuar con estas simulaciones es crucial para llegar a entenderlas correctamente. A lo largo del desarrollo de esta tesis, hemos propuesto nuevas técnicas para mejorar tanto la calidad del pintado como la velocidad de diferentes modelos de representación comúnmente usados en la visualización molecular. Además, hemos desarrollado un nuevo sistema para analizar tanto las simulaciones de acoplamiento molecular, como las fuerzas de interacción entre grupos moleculares que dirigen el proceso de acoplamiento.

By exploiting the concept of emergent network properties and the hierarchical nature of networks, we have constructed several levels of models facilitating the investigations of issues in the area of respiratory neural control. The first of such models is an intracellular second messenger pathway model, which has been shown to be an important contributor to intracellular calcium metabolism and mediate responses to neuromodulators such as serotonin. At the next level, we have constructed new single neuron models of respiratory-related neurons (e.g. the pre-Btzinger complex neuron and the Hypoglossal motoneuron), where the electrical activities of the neurons are linked to intracellular mechanisms responsible for chemical homeostasis. Beyond the level of individual neurons, we have constructed models of neuron populations where the effects of different component neurons, varying strengths and types of inter-neuron couplings, as well as network topology are investigated. Our results from these simulation studies at different structural levels are in line with experiment observations. The small-world topology, as observed in previous anatomical studies, has been shown here to support rhythm generation along with a variety of other network-level phenomena. The interactions between different inter-neuron coupling types simultaneously manifesting at time-scales orders of magnitude apart suggest possible explanations for variations in the outputs measured from the XII rootlet in experiments. In addition, we have demonstrated the significance of pacemakers, along with the importance of considering neuromodulations and second-messenger pathways in an attempt to understand important physiological functions such as breathing activities.

We present a highly-parallel implementation of the Langevin simulation method for modeling ferrofluids on Graphical Processor Units (GPU). Our method is based on the Barnes-Hut algorithm. As a benchmark we use the straightforward 'all-to-all interaction' algorithm. The obtained results are in good agreement with known theoretical model. With the proposed method we were able to follow the evolution of a system of one million interacting particles over long time-scales, the task hitherto is out of reach with the standard, CPU-based numerical schemes. When you are citing the document, use the following link http://essuir.sumdu.edu.ua/handle/123456789/35341

Graphene was first isolated in the lab in 2003 and this work was first published in 2004 by a research team at The University of Manchester. Since that date, graphene research has exploded due to its special properties. Phonons and molecular dynamic simulation provide valuable tools to study the molecular systems under different structure forms. They are helpful to study graphene ribbons and defects. On the other hand, many machine learning techniques were extensively used to analyse the enormous amounts of data resulted from the molecular simulations. As such, this thesis aimed to use one of the machine learning techniques to study phonons of graphene with single vacancy defect and graphene armchair nanoribbons. PCA can be used to transform the atomic velocities into orthogonal eigenvectors such that each eigenvector represents one of the phonon modes of graphene. This is helpful to visualize the atomic motion of a specific phonon mode. To provide orthogonal eigenvectors, PCA needs the data to be of gaussian distribution. The atomic velocities resulted from the molecular simulations follow gaussian distribution at the equilibrium state. Hence, the assumption of gaussian distribution needed by PCA is achieved. However, only some of the phonon modes can be calculated from the atomic velocities in their real space. Most of the phonon modes are calculated after transforming the atomic velocities to a reciprocal space (k space) using spatial Fourier transform. The k space atomic velocities are not following gaussian distribution. This thesis introduced a novel method to use PCA to isolate and visualize the phonon modes extracted from the k space velocities. To prove the feasibility of using PCA to isolate k space phonons, we conducted classical molecular simulations of graphene with different structures. The effect of single vacancy defect on graphene phonons was studied in comparison to the perfect graphene. In addition, the effect of the armchair ribbon width on graphene phonon modes was investigated. The results of the conducted molecular simulations were used with PCA to visualize some of the phonon modes of pristine graphene and armchair nanoribbons of graphene. We used PCA to present the evolution of the atomic motion of specific k space phonon modes of armchair ribbons: the first overtone of TA phonon mode and the highest overtone of TO phonon mode. The presented motions showed that the breathing like mode is a transition state between two opposite atomic motions of TA mode. In the method we introduced using PCA, we used the eigenvectors with the lowest eigenvalues to study the Fourier transformed atomic velocities. This method rotated the k space atomic velocities into the eigenvectors with the lowest eigenvalues which helped to isolate and visualize the k space phonon modes.

Adsorption is a low-energy separation process especially advantageous when the components to be separated are similar in nature or have a low molar concentration. The choice of the adsorbent is the key factor for a successful separation, and among them periodic mesoporous silicas (PMS) are of importance because of their pore sizes, shapes and connectivity. Furthermore, they can be modified by post-synthesis functionalisation, which provides a tool for tailoring them to specific applications. SBA-2 and STAC-1 are two types of PMS characterised by a three-dimensional pore system of spherical cages interconnected by a network of channels whose formation process was until now obscure. In this work the kinetic Monte Carlo (kMC) technique has been extended to simulate the synthesis of these complex materials, presenting evidence that the interconnecting network originates from spherical micelles touching during their close-packing aggregation in the synthesis. Moreover, for the first time atomistic models for these materials were obtained with realistic pore-surface roughness and details of the possible location of its interaction sites. Grand Canonical Monte Carlo (GCMC) simulations of nitrogen, methane and ethane adsorption in the materials pore models show excellent agreement with experimental results. In addition, their potential as design tools is explored by introducing surface groups for enhancing CO2 capture; and finally, application examples are presented for carbon dioxide capture from flue gases and for natural gas purification, as well as in the separation of n-butane / iso-butane isomers.

Research in the field of titanium-based complexes as potential anticancer drugs has led to impressive results in vitro, but there is still scope for further improvements. Significant effort has been put into synthetic research and biological evaluation of titanocene derivatives, as well as the identification of their main biological targets. Surprisingly, the area of computer-aided drug design (CADD) has not been utilized up-to-date for the design of novel titanium-based compounds. The work within the thesis describes the computational approaches employed to study the mode of action of titanium-based anticancer agents, highlighting the structural features required for biological activity. Consequently, novel titanium-based derivatives were designed and synthesized. Furthermore, the synthetic attempts for amidosilyl-substituted titanocene and ferrocene derivatives with potentially enhanced activity are also reported. The first chapter provides a thorough introduction into the developments in the field of metal-based antitumour agents with a particular emphasis on titanium-based agents and assessment of their cytotoxic activity. Chapter 2 describes the first systematic receptor- based docking approach utilized to understand the binding mode of titanium-based agents against human serum albumin (RSA). The first 3D-QSAR study, which is reported in Chapter 3, was performed on a series of titanocene complexes to gain insight into the key structural features vital for their biological activity and to advance the design of potent titanocene anticancer agents. The successful synthesis of six novel benzyl-substituted titanocene derivatives, which have been designed based on the 3D-QSAR analysis is described in Chapter 4. Chapter 5 focuses on the attempts made for the synthesis of a range of amidosilyl-substituted titanium and iron-based metallocenes. Although the amidosilyl- substituted derivatives could not be obtained as pure compounds, six titanium and iron- based metallocenes were successfully synthesized. Chapter 6 summarizes the results obtained and provides suggestions for areas of future research.

Studies of three different, but interlinked, aspects of membrane properties are presented here. First an analysis of both the lyotropic and thermotropic phase transitions of a phospholipid bilayer using the ELBA1.0 forcefield. This is believed to be the first time such transitions (including the rippled gel phase) have been observed using a coarse grained (CG) forcefield not specifically parameterised to do so. Further, analysis via enhanced sampling methods of the relative free energies of the phases is presented. These analyses confirm the enormous effects comparatively small changes to a forcefield can have on the aggregate behaviour of lipids modelled with it. Molecular dynamics studies on the bending rigidity of bilayers are also presented, including a comparison of 4 different computational methods for calculating the bending rigidity of bilayers. The system size dependency of such methods is compared, as well as their ability to reproduce trends in bending rigidity across lipid species already measured experimentally. The methods are compared across two different atomistic and one CG method for the first time and, despite quite different theoretical bases, are shown to produce surprisingly consistent results both with each other and with previously published experiment. Finally, two new parameterisations of cholesterol using the ELBA forcefield are explored. Their ability to induce the ordering and structure seen in atomistic simulations is measured and compared with another widely using coarse grained forcefield. ELBA's unique (amongst CG forcefields) direct compatibility with atomistic forcefields also allowed dual-resolution simulations of binary bilayers to be analysed. Results from such dual-resolution simulations are consistent with those resulting from atomistic simulation.

Although increased genome sequencing e orts have increased our understanding of genomic variability within many bacterial species, there has been limited application of this knowledge towards assessing current molecular typing methods and developing novel molecular typing methods. This thesis reports a novel in silico comparative genomic framework where the performance of typing methods is assessed on the basis of the discriminatory power of the method as well as the concordance of the method with a whole-genome phylogeny. Using this framework, we designed a comparative genomic ngerprinting (CGF) assay for Listeria monocytogenes through optimized molecular marker selection. In silico validation and assessment of the CGF assay against two other molecular typing methods for L. monocytogenes (multilocus sequence typing (MLST) and multiple virulence locus sequence typing (MVLST)) revealed that the CGF assay had better performance than these typing methods. Hence, optimized molecular marker selection can be used to produce highly discriminatory assays with high concordance to whole-genome phylogenies. The framework described in this thesis can be used to assess current molecular typing methods against whole-genome phylogenies and design the next generation of high-performance molecular typing methods from whole-genome sequence data.
xiii, 100 leaves : ill. ; 29 cm

Ce travail de thèse aborde l’un des défis technologiques liés au développement de nouvelles générations de transistors (FinFET, FDSOI), pour lesquels la gravure de couches ultraminces révèle plusieurs problèmes. En particulier, la gravure des espaceurs nitrure (SiN) doit être réalisée avec une précision nanométrique sans endommager les couches sous-jacentes, étape qui ne peut plus être réalisée par des plasmas conventionnels continus. Afin de dépasser cette limitation, une approche innovante a été récemment développée (dite Smart-Etch), qui s’appuie sur l'implantation d’ions légers et se déroule en deux étapes. Premièrement, le matériau à graver est exposé à un plasma ICP ou CCP d’hydrogène (H2) ou d’hélium (He); dans une deuxième étape, la couche modifiée est retirée sélectivement par gravure humide ou exposition à des réactifs gazeux. Afin d’appréhender les mécanismes fondamentaux de la première étape et assister le développement de cette nouvelle technologie, des simulations de dynamique moléculaire (MD) ont été réalisées pour étudier l'interaction des plasmas H2/He avec des couches de Si/SiN. La MD a été utilisée pour examiner comment la modification de ces substrats est affectée par l’énergie ionique, la dose ionique, la composition ionique ou le rapport flux de radicaux/ flux d’ions (dans le cas d’un plasma H2). En accord avec les expériences, les simulations de bombardement ionique He+ ou Hx+ (x = 1-3) sur Si/SiN montrent que l’implantation ionique est auto-limitée, et que l’évolution de la surface se déroule en deux étapes : une rapide modification en volume (sans gravure) suivie d'une saturation lente et de la formation d'une couche implantée stable en régime permanent (état stationnaire). Les mécanismes d'endommagement induit par les ions (rupture des liaisons Si-Si ou Si-N, piégeage/désorption d’He ou H2, formation de groupes SiHx (x = 1-3) en profondeur), sont étudiés et permettent d’apporter de nouveaux éléments de compréhension aux technologies Smart-Cut et Smart-Etch. L’exposition de substrats Si/SiN à un plasma H2 (impacts d’ions Hx+ et de radicaux H) a également été étudiée pour différentes conditions plasma. Dans ce cas, une transformation auto-limitée est observée mais les couches modifiées/hydrogénées sont simultanément gravées pendant l'implantation ionique, à un taux 10 fois inférieur pour SiN par rapport à Si. Les simulations montrent que modifier des substrats Si/SiN avec une précision nanométrique nécessite un contrôle prudent de l’énergie et du flux des ions incidents. En particulier, les faibles doses ioniques doivent être évitées car l’évolution de la surface ne peut pas être contrôlée précisément en régime transitoire (modification rapide). Dans les plasmas H2, les énergies ioniques élevées induisent des couches modifiées plus épaisses mais des taux d'hydrogénation plus faibles et moins homogènes. La composition ionique et le rapport flux de radicaux/ flux ions (Γ) doivent également être controllés avec précaution, notamment car la vitesse de gravure du matériau augmente avec Γ, ce qui empêche entre-autre la possibilité du Smart-Etch pour le silicium. Les simulations MD réalisées dans cette thèse permettent de clarifier divers phénomènes inexpliqués observés dans le Smart-Etch expérimentalement, et de révéler quelques problèmes possibles dans ce nouveau procédé. Finalement, une gamme de paramètres plasma est proposée pour optimiser cette première étape de Smart-Etch et contrôler la modification de SiN avec une précision sous-nanométrique
This PhD thesis focuses on technological challenges related to the development of advanced transistors (FinFET, FDSOI), where the etching of thin films reveals several issues. In particular, the etching of silicon nitride spacers should be achieved with a nanoscale precision without damaging the underlayers, a step which cannot be addressed by conventional CW plasmas. To overpass this limitation, an innovative approach was recently developed (so-called Smart Etch), which is based on light ion implantation and composed of two steps. First, the material to be etched is modified by exposure to a hydrogen (H2) or helium (He) ICP or CCP plasma; in a second step, the modified layer is selectively removed using wet etching or gaseous reactants only. To support the fundamental understanding of the first step and assist the development of this new technology, molecular dynamics (MD) simulations were performed to study the interaction between silicon/silicon nitride films and hydrogen/helium plasmas. MD was used to investigate how the substrates modification is affected by the ion energy, the ion dose, the ion composition or the radical-to-ion flux ratio (in the case of a H2 plasma). In agreement with experiments, simulations of He+ or Hx+ (x=1-3) ion bombardment of Si/SiN show that a self-limited ion implantation takes place with a surface evolution composed of two stages: a rapid volume modification (with no etching) followed by a slow saturation and the formation of a stable He- or H- implanted layer at steady state. The mechanisms of ion-induced damage (Si-Si or Si-N bond breaking, He or H2 trapping/desorption, SiHx (x=1-3) complex creation) are investigated and allow to bring new insights to both the Smart Cut and Smart Etch technologies. Si/SiN exposure to various H2 plasma conditions (with both Hx+ ions and H radicals) was then studied. In this case, a self-limited transformation is observed but the H-modified layers are simultaneously etched during the ion implantation, at a rate ~10 times smaller for SiN compared to Si. Simulations show that to modify Si/SiN thin films with a nanoscale precision by H2 or He plasmas, both the ion energy and the ion flux have to be controlled very cautiously. In particular, low ion doses, where the substrate evolution is in rapid modification stage, must be avoided since the substrate evolution cannot be precisely controlled. In H2 plasmas, high ion energies induce thicker modified layers but smaller and less homogeneous hydrogenation rates. The ion composition and the radical-to-ion flux ratio Γ must be considered as well, since the etch rate increases with Γ, compromising even the possibility to achieve a Smart Etch of silicon. The MD simulations performed in this thesis enable to clarify various unexplained phenomena seen in the Smart-Etch experimentally, and reveal some possible issues in this new process. In the end, a range for plasma parameters is proposed to optimize this first step of the Smart Etch process and to control the modification of SiN with a sub-nanoscale precision

Wet granulation process requires the addition of a coating agent or binder, typically composed of surfactants, water, plasticizers and fillers. In dry granulation however, the coating agent is added to the system in the form of fine solid particles. Our goals are to investigate the particles behaviour and agglomeration mechanism in dry and aqueous systems at the micro and meso scales, and also, to develop predictive methodologies and theoretical tools of investigation allowing to choose the adequate binder and to formulate the right coating solution. In this study we chose materials widely used in food and pharmaceutical industries, including; coating agents such as Hydroxypropyl-methylcellulose (HPMC) and Ethyl cellulose (EC), binders such as Polyvinylpyrrolidone (PVP) and Microcrystalline cellulose (MCC), hydrophobic filler such as Stearic acid (SA) and plasticizer such as Polyethylene glycol (PEG). A successful granulation requires good affinity between host and guest particles. In this context, in the first part of this work, two approaches to predict the binder-substrate affinity in dry and in aqueous media were compared; one based on the work of adhesion and the other based on the ideal tensile strength. The concept of ideal tensile strength was extended to ternary systems and applied for granulation in aqueous media. The developed approaches were thereafter tested for various systems (composed of PVP, MCC, HPMC, SA, EC, PEG and water) and compared to experimental observations. Approaches yielded results in good agreement with the experimental observations, but the work of adhesion approach might give more accurate affinity predictions on the particles affinity than the ideal tensile strength approach. Both approaches predicted that HPMC is a good binder for MCC. Results also indicated that PEG has a good affinity with HPMC and SA. In a second part of our work, we used mesoscale simulations and experimental techniques to investigate the structure of agglomerates formed in aqueous colloidal formulations used in coating and granulation processes. For the simulations, dissipative particle dynamics (DPD) and a coarse-grained approach were used. In the DPD method, the compounds were described as a set of soft beads interacting according to the Flory-Huggins model. The repulsive interactions between the beads were evaluated using the solubility parameter (δ) as input, where, δ was calculated by all-atom molecular simulations. The mesoscale simulation results were compared to experimental results obtained by Cryogenic-SEM, particle size distribution analysis and DSC technique. According to the DPD simulations, HPMC polymer is a better stabilizing agent for SA than PVP and MCC. In addition, HPMC is able to cover the SA particle with a thick layer ant to adsorb in depth into its inner core, preventing SA agglomeration and crystal growth. But, for high amounts of SA (above 10% (w/w)), HPMC is unable to fully stabilize SA. We also found that PEG polymer diffuses inside HPMC chains thereby extending and softening the composite polymer. Experimental results presented similar trends; particle size distribution analysis showed that in the presence of HPMC, for low percentages of SA (below 10% (w/w)), the majority of SA particles are below 1 μm in diameter. SEM images revealed that HPMC surrounds SA crystals with a hatching textured film and anchors on their surface.

La thermodiffusion, également appelé effet Soret, décrit le couplage entre les gradients de température et les flux massiques qui en résultent. Ce phénomène intervient dans de nombreux processus naturels et applications industrielles. En particulier, les réservoirs pétroliers sont sujets à ce phénomène impliquant des fluides multi constituants confinés dans une matrice poreuse et soumis à un gradient de température. Néanmoins, malgré beaucoup des progrès, il existe relativement peu de mesures fiables de ce phénomène et sa modélisation reste largement un problème ouvert. L’objectif principal de cette thèse s’inscrit dans ce cadre, à savoir développer une approche expérimental permettant de fournir des données de références sur la thermodiffusion notamment dans l’optique de quantifier l’effet de la pression sur cette dernière. Ainsi, durant cette thèse, nous avons développé une cellule de thermodiffusion en milieu libre qui permet d’étudier par shadowgraphie les fluctuations de non équilibre induites par effet Soret. L’appareil de mesure a ensuite été utilisé pour étudier deux mélanges binaires représentatifs de fluides pétroliers, à savoir le mélange équimassique tétraline/dodécane (en phase liquide) et le mélange dioxyde de carbone/méthane (en phases gaz et supercritique). En complément, des simulations de dynamique moléculaire ont été réalisées sur le mélange dioxyde de carbone/méthane. Par analyse dynamique des images de shadowgraphie nous avons pu déterminer les coefficients de diffusion et Soret en fonction de la pression pour le mélange tétraline/dodécane. Aux incertitudes près, nous observons une décroissance linéaire avec la pression pour ces coefficients. De plus nous avons observé l’effet du confinement de la cellule sur les fluctuations en très bon accord avec la théorie et les simulations. Pour le mélange dioxyde de carbone/méthane l’analyse dynamique a montré une cinétique difficilement accessible de par les limites physiques et informatiques du dispositif expérimental utilisé. L’analyse statique montre, quant à lui, une croissance rapide de l’amplitude des fluctuations avec la pression jusqu’à un seuil au-delà duquel elle décroît. Sur ce mélange les simulations de dynamique moléculaire ont montré un bon accord avec les prédictions théoriques
Thermodiffusion, also called the Soret effect, describes the coupling between temperature gradient and resulting fluxes. This phenomenon is involved in a number of natural and industrial processes. In particular, multi components fluids in petroleum reservoirs are subjected to this phenomenon because of the geo-thermal gradient. Nevertheless, in spite of a lot of advances, there are few available data of this phenomenon and the establishment of a theoretical model, able to give a quantitative estimation of these transport coefficients whatever molecules in presence, is still an open question. The principal aim of this thesis is to develop an experimental approach allowing providing reference data on thermodiffusion as a function of the pressure. During this thesis, we developed a high pressure thermodiffusion cell in free medium, enabling us to study concentration non-equilibrium fluctuations induced by the Soret effect by means of shadowgraph optical technique. With this setup we investigated two binary mixtures representatives of petroleum fluids; namely the equimassic tetralin/dodecane mixture in liquid phase and the carbon dioxide/methane mixture in gaseous and super critical state. Furthermore, molecular dynamic simulations on the second mixture were performed. Using a dynamic image analysis, we have measured molecular diffusion and Soret coefficient for the tetralin/dodecane mixture. Within experimental uncertainties, we observed a linear decrease of these coefficients with the pressure. Furthermore, we were able to observe the effect of confinement (finite size effect induced by cell vertical boundary conditions) on fluctuation dynamics, in good agreement with calculations and simulations based on hydrodynamic fluctuation theory on similar solutal Rayleigh number. Concerning the carbon dioxide/methane mixture, the dynamic analysis revealed a kinetic too fast for our experimental apparatus. Conversely, static analysis revealed a rapid increase of the non-equilibrium fluctuation magnitude as a function of the pressure up to a threshold beyond which it decreases. On this mixture, performed molecular dynamic simulations provided results in good agreement with expected theoretical behaviour

L'évolution des réserves de pétrole implique l'utilisation en raffinerie de pétroles bruts non conventionnels, bien souvent plus lourds et donc difficiles à caractériser. Les produits pétroliers sont en effet des mélanges chimiques extrêmement complexes. La partie légère et volatile peut être analysée par chromatographie en phase gazeuse couplée à la spectrométrie de masse (GC/MS), permettant l'identification des composés par l'utilisation de mesures de masses précises et de modèles de fragmentation. Cependant ces techniques sont inadaptées à l'analyse des fractions lourdes. Dans la pratique, la caractérisation des mélanges les plus complexes implique l'utilisation de spectromètres de masse à ultra-haute résolution généralement par analyse directe sans séparation chromatographique. La technique de référence est aujourd’hui la spectrométrie de masse à transformée de Fourier par résonance cyclotronique des ions (FTICR). Grâce à une résolution supérieure à 106 et à une précision de mesure de masse inférieure à 0,1 ppm, cet instrument permet de séparer toutes les espèces présentes dans un produit pétrolier et d'attribuer à chaque valeur de m/z une composition élémentaire unique. Ceci permet d'obtenir très facilement des cartes moléculaires qui peuvent être présentées graphiquement en utilisant le diagramme de Kendrick, le diagramme de van Krevelen ou le nombre d'insaturations (DBE) en fonction du nombre de carbones. Ce travail de thèse a permis grâce à la caractérisation moléculaire de produits pétroliers (Vacuum Gas Oil, Pétroles Bruts, Matériel Interfacial, Asphaltènes et Bio-Oil…) d'aborder la complexité de leur traitement dans l’outil de raffinage. Des protocoles d'analyses des échantillons ont été développés, à l'aide de différentes sources d'ionisation à pression atmosphérique (ESI, APCI et APPI) ainsi que par désorption/ionisation laser (LDI) sur le spectromètre de masse FTICR 12T. Les informations sur le contenu isomérique des produits pétroliers ont ensuite été déterminées grâce à l'apport de la spectrométrie de mobilité ionique (IMS)
The evolution of oil reserves requires the use in refineries of unconventional crude oils, which are often heavier and therefore difficult to characterize. Petroleum products are in fact extremely complex chemical mixtures. The light and volatile part can be analysed by gas chromatography coupled with mass spectrometry (GC/MS), allowing the identification of compounds by using precise mass measurements and fragmentation models. However, these techniques are inappropriate for the analysis of heavy fractions. In practice, the characterization of the most complex mixtures involves the use of ultra-high-resolution mass spectrometers generally by direct analysis without chromatographic separation. The reference technique today is Fourier Transform Ion Cyclotron Resonance Mass Spectrometry (FTICR). With a resolution of more than 106 and a mass measurement accuracy of less than 0.1 ppm, this instrument can separate all the species present in a petroleum product and assign a unique elemental composition to each m/z value. This makes it very easy to obtain molecular maps that can be presented graphically using the Kendrick diagram, the van Krevelen diagram or the number of unsaturations (DBE) as a function of the number of carbons. This thesis work has allowed thanks to the molecular characterization of petroleum products (Vacuum Gas Oil, Crude Oil, Interfacial Material, Asphaltenes and Bio-Oil...) addressing the complexity of their treatment in the refining tool. Protocols for sample analysis have been developed, using different sources of ionization at atmospheric pressure (ESI, APCI and APPI) as well as laser desorption/ionization (LDI) on the FTICR 12T mass spectrometer. Information on the isomeric content of petroleum products was then determined using ion mobility spectrometry (IMS)

博士
國立清華大學
分子醫學研究所
103
Chapter I: Site of metabolism prediction for FMO enzymes via machine learning and condensed Fukui function The flavin-containing monooxygenase (FMO) catalyzes xenobiotics with soft nucleophiles and also plays an important role in drug metabolism in Phase I enzymes. The site of metabolism (SOM) refers to the place where the reaction of metabolism occurs in a molecule. Identification of SOMs of a compound is not usually a low-cost task in drug discovery. Thus, a silico method to predict site of metabolism (SOMs) of FMOs would provide medical chemists information of SOMs before experiments. In this work, we developed a machine learning model combining quantum features (condensed Fukui function) and circular fingerprints to predict potential SOMs in a molecule. The final model via SVM was easily interpreted with only five features. In the training set with 10 CV showed an area under curve (AUC) value of ROC curve, 0.889, and the value of MCC,0.767. In the external validation, AUC value of the model was 0.801 and the accuracy (MCC) was 0.611. These showed the predictive power of our model and we wish such a research to assist medical chemists in the assessment of FMO metabolism at the preclinical stage of drug discovery. Chapter II: Interaction between Trehalose and MTHase from Sulfolobus solfataricus studied by theoretical computation and site-directed mutagenesis Maltooligosyltrehalose trehalohydrolase (MTHase) catalyzes the release of trehalose by cleaving the α-1,4-glucosidic linkage next to the α-1,1-linked terminal disaccharide of maltooligosyltrehalose. Computer simulation using the hydrogen bond analysis, free energy decomposition, and computational alanine scanning were employed to investigate the interaction between maltooligosyltrehalose and the enzyme. The same residues that were chosen for theoretical investigation were also studied by site-directed mutagenesis and enzyme kinetic analysis. The importance of residues determined either experimentally or computed theoretically were in good accord with each other. It was found that residues Y155, D156, and W218 of subsites -2 and -3 of the enzyme might play an important role in interacting with the ligand. The theoretically constructed structure of the enzyme-ligand complex was further validated through an ab initio quantum chemical calculation using the Gaussian09 package. The activation energy computed from this latter study was very similar to those reported in literatures for the same type of hydrolysis reactions. Chapter III: A Theoretical Study on the Alkaline Hydrolysis of Methyl Thioacetate in Aqueous Solution A base catalyzed hydrolysis reaction of thiolester has been studied in both gas and solution phases using two ab initio quantum mechanics calculations such as Gaussian09 and CPMD. The free energy surface along the reaction path is also constructed using a configuration sampling technique namely the metadynamics method. While there are two different reaction paths obtained for the potential profile of the base-catalyzed hydrolysis reaction for thiolester in gas phase, a triple-well reaction path is computed for the reaction in solution phase by both two quantum mechanics calculations. Unlike a SN2 mechanism (a concerted mechanism) found for the gas-phase reaction, a nucleophilic attack from the hydroxide ion on the carbonyl carbon to yield a tetrahedral intermediate (a stepwise mechanism) is observed for the solution phase reaction. Moreover, the energy profiles computed by these two theoretical calculations are found to be well comparable with those determined experimentally.

碩士
國立成功大學
機械工程學系碩博士班
92
This study is .focused on the efficiency improvement of molecular dynamics simulation and the simulation enlargement with parallel computing technique. Using semi-empirical potiential derived from LCAO(Linear Combination of Atomic Orbital) to accomplish the Large Scale Molecular Dynamics Simulation which can’t be done with Periodic Boundary Condition on personal computer. Two parallel algorithms, Atom-Decomposition Method and Spatial-Decomposition Method, for classical short-range molecular dynamics simulation are used. The two algorithms are successfully tested on a standard tight-binding benchmark problem for system sizes ranging from 20,000 to 1,6000,000 atoms on distributed-memory parallel machine which allows for message-passing of data between independently executing processors, such as linux cluster. We discuss the efficiency difference of two parallel algorithms in the same simulation. The relation between efficiency and computing processors are also discussed. Finally, the bottlenecks of parallel molecular dynamics simulation will be carried out, and the state-of-the-art ways overcomings the bottlenecks are also shown to be our future works.

Yes
In this article we briefly review the Boxed Molecular Dynamics (BXD) method, which allows analysis of thermodynamics and kinetics in complicated molecular systems. BXD is a multiscale technique, in which thermodynamics and long-time dynamics are recovered from a set of short-time simulations. In this article, we review previous applications of BXD to peptide cyclization, diamond etching, solution-phase organic reaction dynamics, and desorption of ions from self-assembled monolayers (SAMs). We also report preliminary results of simulations of diamond etching mechanisms and protein unfolding in AFM experiments. The latter demonstrate a correlation between the protein’s structural motifs and its potential of mean force (PMF). Simulations of these processes by standard molecular dynamics (MD) is typically not possible, since the experimental timescales are very long. However, BXD yields well-converged and physically meaningful results. Compared to other methods of accelerated MD, our BXD approach is very simple; it is easy to implement, and it provides an integrated approach for simultaneously obtaining both thermodynamics and kinetics. It also provides a strategy for obtaining statistically meaningful dynamical results in regions of configuration space that standard MD approaches would visit only very rarely.
DRG is grateful for funding from a Royal Society Research Fellowship. JB and DVS acknowledge the support of EPSRC (Grant No EP/E009824/1). E.M.-N. and S.A.V. are grateful to the “Centro de Supercomputación de Galicia (CESGA)” for the use of its computational resources, as well as to “Ministerio de Economía y Competitividad” (Grant No. CTQ2009-12588) for financial support. DS and E.M.-N. acknowledge the Leverhulme Trust for funding the E.M.-N. visit to Leeds by the grant “Accelerated classical and quantum molecular dynamics and its applications” (Grant No. VP1-2012-013).

博士
國立成功大學
機械工程學系碩博士班
90
This study is focused on the investigation of the physical mechanism about the nano thin film deposition using molecular dynamics and the phenomena of the size effect in molecular dynamics using parallel computing. The process parameters in the simulation are including deposition rate, incident energy, incident angle, and the substrate temperature. The simulated PVD processes contain evaporation, sputtering, ionized cluster beam deposition, and Damascene process. From the simulated results about the evaporation and the sputtering, deposition rate has no influence on the film structure in the evaporation process, but there would be an optimal range of deposition rate to obtain high quality film structure in the sputtering process; higher incident energy and substrate temperature will promote the film to grow in a two-dimensional mode and the smoother film surface will be obtained; larger incident angle encourages the self-shadowing effect, and it is not beneficial for the smoothing of the film surface. In ionized cluster beam deposition process, the transforms of the force, kinetic energy, and atom migration induced by the impact between the cluster and the substrate are along the close packed direction of the structure. In Damascene process, the trench will be filled from the bottom to the opening at higher incident energy and the void existing in the trench could be free by the atom diffusion at higher substrate temperature. The size effect in the Damascene simulation will be investigated by parallel computing. The simulation model will be enlarged and the simulation time domain will be prolonged by parallel computing. Comparing the simulated results from the smaller and the larger model, the size effect will appear obviously at smaller incident energy and the elevation of the substrate temperature. Finally, the bottlenecks of the molecular dynamics simulation will be carried out, and the state-of-the-art ways overcoming these bottlenecks are also shown to be our future works.

碩士
國立成功大學
化學系碩博士班
97
Nuclear magnetic resonance techniques have been applied to investigate the influences of diluent type, diluent content and temperature on the chemical shift, diffusion coefficient and the relaxation time for 1-(2-hydroxyethyl)-3-methylimidazolium tetrafluoroborate (HEMIMBF4).The macroscopic properties, like viscosity, density and conductivity of the mixture were measured. To understand further the details of ionic diffusion and conductivity, molecular dynamics simulations were also conducted. It’s known from the experiment that the 1H chemical shift difference has the order：H10＞H2＞H5∼H4, which indicates that the hydroethyl hydrogen atom of HEMIM+ ion forms a stronger hydrogen bond with BF4－ ion and diluent molecules than any other hydrogen atoms. The result is in accord with the partial charge obstained from the simulations. The NOESY spectra shows that most of the ions in pure HEMIMBF4 aggregate to form large clusters, the size of which diminishes with the increase of diluent content, which is further confirmed by the simulations. In general, the diffusion coefficient of BF4－ ion is greater than that of HEMIM+ ion. The diffusion rate of HEMIM+ at a acetonitrile (AN) molar fraction (XAN) of 0.8 is far greater than that at 0.7, owing to the rapid increase of coordinating acetonitrile molecules around HEMIM+ ion. As the diluent content increases, the viscosity of solution decreases, but the relaxation time, diffusion rate and specific conductivity increases. These phenomena can be interpreted with the coordination number, free ion fraction and ionic cluster size that are obtainable from the simulations. At constant temperature and diluent concentration, benzonitrile molecule has stronger interaction with HEMIM+ than AN. For HEMIMBF4 / BN system, the Stokes-Einstein radius is smaller than that of AN containing system, which leads to the high diffusion rate and conductivity. The origin can be traced back to the differences in the coordination number, free ion fraction, diluent size and solution viscosity. Based on this work, the surroundings around the HEMIM+ ion is now clear, which is believed to be helpful for the microscopic explanation of experimental observations.

Indiana University-Purdue University Indianapolis (IUPUI)
The multi-protein complex of F0F1 ATP synthase has been of great interest in the fields of microbiology and biochemistry, due to the ubiquitous use of ATP as a biological energy source. Efforts to better understand this complex have been made through structural determination of segments based on NMR and crystallographic data. Some experiments have provided useful data, while others have brought up more questions, especially when structures and functions are compared between bacteria and species with chloroplasts or mitochondria. The epsilon subunit is thought to play a signi cant role in the regulation of ATP synthesis and hydrolysis, yet the exact pathway is unknown due to the experimental difficulty in obtaining data along the transition pathway. Given starting and end point protein crystal structures, the transition pathway of the epsilon subunit was examined through computer simulation.The purpose of this investigation is to determine the likelihood of one such proposed mechanism for the involvement of the epsilon subunit in ATP regulation in bacterial species such as E. coli.

碩士
國立臺灣大學
高分子科學與工程學研究所
106
In this thesis, we constructed a ellipsoid-based coarse-grained model to study the mesoscale morphology evolution of DPDCPB:C70 small molecule organic solar cells during vacuum co-deposition processes. We coarsed the DPDCPB small molecule donor molecules, and the C70 acceptor molecules into three bonded ellipsoids and one single ellipsoid, respectively. The interactions between ellipsoids are described by the anisotropic Gay-Berne force field. By employing the covariant matrix evolutionary strategy (CMAES), we successfully parameterized the Gay-Berne potential by reproducing the potential energy surfaces from respective all atom atomistic simulations. With the parameterized Gay-Berne potential, for the first time, we are able to carry out molecular simulations of the vacuum co-deposition processes with system size compatible with experiments owing to the significantly reduced system degrees of freedom from coarse-graining. We studied the effects of DPDCPB:C70 deposition ratios, as well as the DPDCPB capping layer residual strains (densities) on the resultant composite film growth modes. Our simulation results indicate that C70 can easily aggregate into clusters regardless of concentrations (C70 deposition ratios). As a result, C70 clusters form the backbone of the film and the growth mode of C70 clusters literally determines the DPDCPB:C70 film morphologies. The residual strain induced by C70 supporting layer beneath the capping layer promote the three-dimensional growth mode of C70, leading to DPDCPB:C70 films with high surface roughness. In contrast, by relaxing the stress in the lateral dimensions, the three-dimensional growth modes can be noticeably suppressed. Hence, we also demonstrates the possibilities of manipulating film morphologies by applying lateral strains in the substrates. This thesis not only provides insights into morphology evolution during vacuum co-deposition processes of small molecule organic solar cells but also pave the pathway toward investigating the mesoscale structures of organic electronic devices such as the organic light emitting diodes (OLEDs).

La flexibilité est une caractéristique intrinsèque des protéines qui doivent, dès le mo- ment de leur synthèse, passer d’un état de chaîne linéaire à un état de structure tridimen- sionnelle repliée et enzymatiquement active. Certaines protéines restent flexibles une fois repliées et subissent des changements de conformation de grande amplitude lors de leur cycle enzymatique. D’autres contiennent des segments si flexibles que leur structure ne peut être résolue par des méthodes expérimentales. Dans cette thèse, nous présentons notre application de méthodes in silico d’analyse de la flexibilité des protéines : • À l’aide des méthodes de dynamique moléculaire dirigée et d’échantillonnage pa- rapluie, nous avons caractérisé les trajectoires de liaison de l’inhibiteur Z-pro- prolinal à la protéine Prolyl oligopeptidase et identifié la trajectoire la plus pro- bable. Nos simulations ont aussi identifié un mode probable de recrutement des ligands utilisant une boucle flexible de 19 acides aminés à l’interface des deux domaines de la protéine. • En utilisant les méthodes de dynamique moléculaire traditionnelle et dirigée, nous avons examiné la stabilité de la protéine SAV1866 dans sa forme fermée insérée dans une membrane lipidique et étudié un des modes d’ouverture possibles par la séparation de ses domaines liant le nucléotide. • Nous avons adapté auproblème de la prédiction de la structure des longues boucles flexibles la méthode d’activation et de relaxation ART-nouveau précédemment uti- lisée dans l’étude du repliement et de l’agrégation de protéines. Appliqué au replie- ment de boucles de 8 à 20 acides aminés, la méthode démontre une dépendance quadratique du temps d’exécution sur la longueur des boucles, rendant possible l’étude de boucles encore plus longues.
Flexibility is an intrinsic characteristic of proteins who from the moment of synthesis into a linear chain of amino acids, have to adopt an enzymatically active tridimensionnel structure. Some proteins stay flexible once folded and display large amplitude confor- mational changes during their enzymatic cycles. Others contain parts that are so flexible that their structure can’t be resolved using experimental methods. In this thesis, we present our application of in silico methods to the study of protein flexibility. • Using steered molecular dynamics and umbrella sampling, we characterized the binding trajectories of the Z-pro-prolinal inhibiter to the Prolyl oligopeptidase pro- tein and we identified the most probable trajectory. Our simulations also found a possible ligand recrutement mechanism that involves a 19 amino acids flexible loop at the interface of the two domains of the protein. • Using traditional and steered molecular dynamics, we examined the stability of the SAV1866 protein in its closed conformation in a lipid membrane and we studied one of its proposed opening modes by separating its nucleotide binding domains. • We also adapted the activation-relaxation technique ART-nouveau which was pre- viously used to study protein folding and aggregation to the problem of structure prediction of large flexible loops. When tested on loops of 8 to 20 amino acids, the method demonstrate a quadratic execution time dependance on the loop length, which makes it possible to use the method on even larger loops.

博士
國立中央大學
物理學系
102
Part I--An ultrafast shape-recognition technique was used to analyze the phase transition of finite-size clusters, which, according to our research, has not yet been accomplished. The shape of clusters is the unique property that distinguishes clusters from bulk systems, and is comprehensive and natural for structural analysis. In this study, an isothermal molecular dynamics simulation was performed to generate a structural database for shape recognition of Ag-Cu metallic clusters using empirical many-body potential. The probability contour of the shape similarity exhibits the characteristics of both the specific heat and Lindemann index (bond length fluctuation) of clusters. Moreover, our implementation of the substructure to the probability of shapes provides a detailed observation of the atom/shell-resolved analysis, and the behaviors of the clusters were reconstructed based on the statistical information. The method is efficient, flexible, and applicable in any type of finite-size system, including polymers and nanostructures. Part II--Folded conformations of proteins in thermodynamically stable states have long lifetimes. Between such stable folded conformations the protein will generally stray from one random conformation to another leading thus to rapid fluctuations. Brief structural changes therefore occur before folding and unfolding events. These short-lived movements are easily overlooked in studies of folding/unfolding for they represent momentary excursions of the protein to explore conformations in the neighborhood of the stable conformation. The present study looks for precursory signatures of protein folding/unfolding within these rapid fluctuations through a combination of three techniques: (1) ultrafast shape recognition, (2) time series segmentation, and (3) time series clustering. The first procedure measures the differences between statistical distributions of atoms in different conformations by calculating shape similarity indices from molecular dynamics simulation trajectories. The second procedure is used to discover the times at which the protein makes transitions from one conformation to another. Finally, the third technique exploits spatial fingerprints of the stable conformations, since strongly correlated atoms in different conformations are different because the bond and steric constraints are different, to map out the sequences of changes preceding the actual folding and unfolding events. The aforementioned high-frequency fluctuations are therefore characterized by distinct correlational changes and structural changes associated with rate-limiting precursors translate into brief segments. Guided by these technical procedures, we identify not only the signatures of transitions between α helix and β hairpin for transthyretin fragment TTR(105-115) (the model system chosen in this work for illustration), but also the important role played by weaker correlations in such protein folding dynamics. Part III--Improved basis sets for the study of polymer dynamics by means of the diffusion theory, and tests on a melt of cis-1,4-polyisoprene decamers, and a toluene solution of a 71-mer syndiotactic trans-1,2-polypentadiene were presented recently [R. Gaspari and A. Rapallo, J. Chem. Phys. 128, 244109 (2008)]. The proposed hybrid basis approach (HBA) combined two techniques, the long time sorting procedure and the maximum correlation approximation. The HBA takes advantage of the strength of these two techniques, and its basis sets proved to be very effective and computationally convenient in describing both local and global dynamics in cases of flexible synthetic polymers where the repeating unit is a unique type of monomer. The question then arises if the same efficacy continues when the HBA is applied to polymers of different monomers, variable local stiffness along the chain and with longer persistence length, which have different local and global dynamical properties against the above-mentioned systems. Important examples of this kind of molecular chains are the proteins, so that a fragment of the protein transthyretin is chosen as the system of the present study. This peptide corresponds to a sequence that is structured in β-sheets of the protein, and is located on the surface of the channel with thyroxin. The protein transthyretin forms amyloid fibrils in vivo, whereas the peptide fragment has been shown [C.P. Jaroniec, C.E. MacPhee, N.S. Astrof, C.M. Dobson, and R.G. Griffin, PNAS 99, 16748 (2002)] to form amyloid fibrils in vitro in extended β-sheet conformations. For these reasons the latter is given considerable attention in the literature, and studied also as an isolated fragment in water solution where both experimental and theoretical efforts have indicated the propensity of the system to form β turns or α-helices, but is otherwise predominantly unstructured. Differing from previous computational studies that employed implicit solvent, we performed in this work the classical molecular dynamics simulation on a realistic model solution with the peptide embedded in an explicit water environment, and calculated its dynamic properties both as an outcome of the simulations, and by the diffusion theory in reduced statistical-mechanical approach within HBA on the premise that the mode-coupling approach to the diffusion theory can give both the long-range and local dynamics starting from equilibrium averages which were obtained from detailed atomistic simulations.