No / To fully harness the enormous potential offered by interfaces between graphitic nanostructures and biomolecules, detailed connections between adsorbed conformations and adsorption behaviour are needed. To elucidate these links, a key approach, in partnership with experimental techniques, is molecular simulation. For this, a force-field (FF) that can appropriately capture the relevant physics and chemistry of these complex bio-interfaces, while allowing extensive conformational sampling, and also supporting inter-operability with known biological FFs, is a pivotal requirement. Here, we present and apply such a force-field, GRAPPA, designed to work with the CHARMM FF. GRAPPA is an efficiently implemented polarisable force-field, informed by extensive plane-wave DFT calculations using the revPBE-vdW-DF functional. GRAPPA adequately recovers the spatial and orientational structuring of the aqueous interface of graphene and carbon nanotubes, compared with more sophisticated approaches. We apply GRAPPA to determine the free energy of adsorption for a range of amino acids, identifying Trp, Tyr and Arg to have the strongest binding affinity and Asp to be a weak binder. The GRAPPA FF can be readily incorporated into mainstream simulation packages, and will enable large-scale polarisable biointerfacial simulations at graphitic interfaces, that will aid the development of biomolecule-mediated, solution-based graphene processing and self-assembly strategies. / Veski

Using computational tools ranging from molecular simulations – including both Monte Carlo and molecular dynamics methods – to quantum mechanical (QM) calculations (primarily at density functional theory (DFT) level), this work focuses on addressing some of the challenges faced in molecular simulations of gas adsorption in metal–organic frameworks (MOFs). This work consists of two themes: one concerns gas adsorption in MOFs with coordinatively unsaturated metal sites (cus’s), and the other one deals with predicting and understanding the breathing behaviour of the flexible MOF MIL-53(Sc). It has been shown experimentally that incorporation of cus’s – also known as “open” metal sites or unsaturated metal centres – into MOFs significantly enhances the uptake of certain gases such as CO2 and CH4. As a result of the considerably enhanced, localized guest-molecule interactions with the cus’s, it, however, remains a challenge to predict correctly adsorption isotherms and/or mechanisms in MOFs with cus’s using grand-canonical Monte Carlo (GCMC) simulations based on generic classical force fields. To address this problem, two multi-scale modelling approaches – which combine GCMC simulations with QM calculations – have been proposed in this work. The first approach is based on the direct implementation of a fluid–framework potential energy surface, calculated by a hybrid DFT/ab initio method, in the GCMC simulations. The second approach involves parameterization of ab initio force fields for GCMC simulations of gas adsorption in MOFs with cus’s. This approach focuses on the generation of accurate ab initio reference data, selection of semiempirical model potentials, and force-field fitting through a multi-objective genetic algorithm approach. The multi-scale simulation strategy not only yields adsorption isotherms in very good agreement with experimental data but also correctly captures adsorption mechanisms, including the adsorption on the cus’s, observed experimentally but absent from GCMC simulations based on generic force fields. The second challenge that this work aims to address concerns the “breathing” phenomenon of MOFs, in which the framework structure adapts its pore opening to accommodate guest molecules, for example. The breathing effect gives rise to some exceptional properties of these MOFs and hence promising applications. However, framework flexibility often poses a challenge for computational studies of such MOFs, because suitable flexible force fields for frameworks are lacking and the effort involved in developing a new one is no less a challenge. Here, an alternative to the force-field-based approach is adopted. Ab initio molecular dynamics (AIMD) simulations – which combine classical molecular dynamics simulations with electronic-structure calculations “on the fly” – have been deployed to study structural changes of the breathing MOF MIL-53(Sc) in response to changes in temperature over the range 100–623 K and adsorption of CO2 at 0–0.9 bar at 196 K. AIMD simulations employing dispersion-corrected DFT accurately simulated the experimentally observed closure of MIL-53(Sc) upon solvent removal and the transition of the empty MOF from the closed-pore phase to the very-narrow-pore phase with increasing temperature. AIMD simulations were also used to mimic the CO2 adsorption of MIL-53(Sc) in silico by allowing the MIL-53(Sc) framework to evolve freely in response to CO2 loadings corresponding to the two steps in the experimental adsorption isotherm. The resulting structures enabled the structure determination of the two CO2-containing intermediate and large-pore phases observed by experimental synchrotron X-ray diffraction studies with increasing CO2 pressure; this would not have been possible for the intermediate structure via conventional methods because of diffraction peak broadening. Furthermore, the strong and anisotropic peak broadening observed for the intermediate structure could be explained in terms of fluctuations of the framework predicted by the AIMD simulations. Fundamental insights from the molecular-level interactions further revealed the origin of the breathing of MIL-53(Sc) upon temperature variation and CO2 adsorption. Both the multi-scale simulation strategy for gas adsorption in MOFs with cus’s and the AIMD study of the stimuli-responsive breathing behaviour of MIL-53(Sc) illustrate the power and promise of combining molecular simulations with quantum mechanical calculations for the prediction and understanding of MOFs.

A common challenge in materials science is the "inverse design problem," wherein one seeks to use theoretical models to discover the microscopic characteristics (e.g., interparticle interactions) of a system which, if fabricated or synthesized, would yield a targeted material property. Inverse design problems are commonly addressed by stochastic optimization strategies like simulated annealing. Such approaches have the advantage of being general and easy to apply, and they can be effective as long as material properties required for evaluating the objective function of the optimization are feasible to accurately compute for thousands to millions of different trial interactions. This requirement typically means that "exact" yet computationally intensive methods for property predictions (e.g., molecular simulations) are impractical for use within such calculations. Approximate theories with analytical or simple numerical solutions are attractive alternatives, provided that they can make sufficiently accurate predictions for a wide range of microscopic interaction types. We propose a new approach, based on the fine discretization (i.e., terracing) of continuous pair interactions, that allows first-order mean-spherical approximation theory to predict the equilibrium structure and thermodynamics of a wide class of complex fluid pair interactions. We use this approach to predict the radial distribution functions and potential energies for systems with screened electrostatic repulsions, solute-mediated depletion interactions, and ramp-shaped repulsions. We create a web applet for introductory statistical mechanics courses using this approach to quickly estimate the equilibrium structure and thermodynamics of a fluid from its pair interaction. We use the applet to illustrate two fundamental fluid phenomena: the transition from ideal gas-like behavior to correlated-liquid behavior with increasing density in a system of hard spheres, and the water-like tradeoff between dominant length scales with changing temperature in a system with ramp-shaped repulsions. Finally, we test the accuracy of our approach and several other integral equation theories by comparing their predictions to simulated data for a series of different pair interactions. We introduce a simple cumulative structural error metric to quantify the comparison to simulation, and find that according to this metric, the reference hypernetted chain closure with a semi-empirical bridge function is the most accurate of the tested approximations. / text

For two decades, single-wall carbon nanotubes (SWCNTs) have captured the attention of the research community, and become one of the flagships of nanotechnology. Due to their remarkable electronic and optical properties, SWCNTs are prime candidates for the creation of novel and revolutionary electronic, medical, and energy technologies. However, a major stumbling block in the exploitation of nanotube-based technologies is the lack of control of nanotube structure (chirality) during synthesis, which is intimately related to the metallic or semiconductor character of the nanotube. Incomplete understanding of the nanotube growth mechanism hinders a rationale and cost-efficient search of experimental conditions that give way to structural (chiral) control. Thus, computational techniques such as density functional theory (DFT), and reactive molecular dynamics (RMD) are valuable tools that provide the necessary theoretical framework to guide the design of experiments. The nanotube chirality is determined by the helicity of the nanotube and its diameter. DFT calculations show that once a small nanotube 'seed' is nucleated, growth proceeds faster if the seed corresponds to a high chiral angle nanotube. Thus, a strategy to gain control of the nanotube structure during chemical vapor deposition synthesis must focus on controlling the structure of the nucleated nanotube seeds. DFT and RMD simulations demonstrate the viability of using the structures of catalyst particles over which nanotube growth proceeds as templates guiding nanotube growth toward desired chiralities. This effect occurs through epitaxial effects between the nanocatalyst and the nanotube growing on it. The effectiveness of such effects has a non-monotonic relationship with the size of the nanocatalyst, and its interaction with the support, and requires fine-tuning reaction conditions for its exploitation. RMD simulations also demonstrate that carbon bulk-diffusion and nanoparticle supersaturation are not needed to promote nanotube growth, hence reaction conditions that increase nanoparticle stability, but reduce carbon solubility, may be explored to achieve nanotube templated growth of desired chiralities. The effect of carbon dissolution was further demonstrated through analyses of calculated diffusion coefficients. The metallic nanocatalyst was determined to be in viscous solid state throughout growth, but with a less solid character during the induction/nucleation stage.

We present predictions for the statistical error due to finite sampling in the presence of thermal fluctuations in molecular simulation algorithms. Expressions for the fluid velocity, density and temperature are derived using equilibrium statistical mechanics. The results show that the number of samples needed to adequately resolve the flow-field scales as the inverse square of the Mach number. The theoretical results are verified for a dilute gas using direct Monte Carlo simulations. The agreement between theory and simulation verifies that the use of equilibrium theory is justified. / Singapore-MIT Alliance (SMA)

Structural evolution and its effect on optical absorption/emission behavior of derivatives of PFs upon heat treatment at different temperatures were studies by means of a combination of x-ray diffraction, transmission electron microscopy and molecular simulation. The main physical characteristics from results of this study over a series of PFs with alkyl side-chains may be summarized as the following: (1) The crystal structure of poly (9,9-di-n-octyl- 2,7-fluorene, PF8) and poly(9,9-bis(2- ethylhexyl)- 2,7-fluorene, PF26) are determined via a combination of selected area electron diffraction and molecular simulation. In PF8 case, there are 8 chains in an orthorhombic unit cell with dimensions a = 2.56 nm, b = 2.34 nm, c (chain axis) = 3.32 nm, space group P212121, and calculated density of 1.041 gcm-3. On the other hand, in PF26 case, there are 3 chains in a trigonal unit cell with dimensions a = 1.67 nm, b = 1.67 nm, c (chain axis) = 4.04 nm, space group P3, and calculated density of 0.991 gcm-3. (2) All the simulation results indicate that branched side-chains in the case of PF26 tend to fill the space among backbones. In contrast, the linear side-chains in the case of PF8 appear to embrace the neighboring backbone, favoring formation of layered structure. (3) As a consequence, co-planrity of PF backbones is decreased by the attached alkyl side-chains. This in turn results in lowered conjugation length, and in favor of blue light emission.

Water is abundant on earth and in the atmosphere and the most crucial liquid for life as we know it. It has been subject to rather intense research since more than a century and still holds secrets about its molecular structure and dynamics, particularly in the supercooled state, i. e. the metastable liquid below its melting point.  This thesis is concerned with different aspects of water and is written from a theoretical perspective. Simulation techniques are used to study structures and processes on the molecular level and to interpret experimental results. The evaporation kinetics of tiny water droplets is investigated in simulations with focus on the cooling process associated with evaporation. The temperature evolution of nanometer-sized droplets evaporating in vacuum is well described by the Knudsen theory of evaporation. The principle of evaporative cooling is used in experiments to rapidly cool water droplets to extremely low temperatures where water transforms into a highly structured low-density liquid in a continuous and accelerated fashion. For water at ambient conditions, a structural standard is established in form of a high precision radial distribution function as a result of x-ray diffraction experiments and simulations. Recent data even reveal intermediate range molecular correlations to distances of up to 17 Å in the bulk liquid. The barium fluoride (111) crystal surface has been suggested to be a template for ice formation because its surface lattice parameter almost coincides with that of the basal plane of hexagonal ice. Instead, water at the interface shows structural signatures of a high-density liquid at ambient and even at supercooled conditions. Inelastic neutron scattering experiments have shown a feature in the vibrational spectra of supercooled confined and protein hydration water which is connected to the so-called Boson peak of amorphous materials. We find a similar feature in simulations of bulk supercooled water and its emergence is associated with the transformation into a low-density liquid upon cooling. / <p>At the time of the doctoral defense, the following papers were unpublished and had a status as follows: Paper 1: Manuscript. Paper 4: Manuscript.</p>

The aim of this thesis is to study the structural and thermodynamical properties of polymers at liquid/liquid interfaces by means of multiscale molecular dynamics simulations. This thesis is presented in alternative format, and the results, consisting of three journal articles, are divided into two main parts. The first part of the thesis looks at the structural and dynamical changes as well as the thermodynamic stability of polymers of varying topology (linear and star-shaped) at interfaces by performing molecular dynamics simulations on model systems. It was found that homopolymers are attracted to the interface in both good and poor solvent conditions making them a surface active molecule, despite not being amphiphilic. In most cases changing polymer topology had only a minor effect on the desorption free energy. A noticeable dependence on polymer topology is only seen for relatively high molecular weight polymers at the interface. Examining separately the enthalpic and entropic components of the desorption free energy suggests that its largest contribution is the decrease in the interfacial free energy caused by the adsorption of the polymer at the interface. Furthermore, we propose a simple method to qualitatively predict the trend of the interfacial free energy as a function of the polymer molecular weight. In terms of the dynamics of a linear polymer, the scaling behaviour of the polymer confined between two liquids did not follow that predicted for polymers adsorbed onsolid or soft surfaces such as lipid bilayers. Additionally, the results show that in the diffusive regime the polymer behaves like in bulk solution following the Zimm model and with the hydrodynamic interactions dominating its dynamics. Further simulations carried out when the liquid interface is sandwiched between two solid walls show that when the confinement is a few times larger than the blob size the Rouse dynamics is recovered. The second part of the thesis focuses on optimizing the MARTINI coarse-grained (CG) Model, which retains certain chemical properties of molecules, to reproduce solubility of polymers, in specific polyethylene oxide (PEO), in both polar and non-polar solvents. Performing molecular dynamics simulations using this CG model will then enable us to study the properties PEO in octanol/water and hexane/water systems with increased length and timescales not accessible by atomistic simulations. The MARTINI CG method (Marrink et al., J. Phys. Chem. B, 2007, 111, 7812) is based on developing the optimal Lennard-Jones parameters to reproduce the partition free energy between water (polar solvent) and octanol (apolar solvent). Here we test the MARTINI CG method when modelling the partitioning properties of PEO, with increasing molecular weight between solvents of different polarity by comparing the results with atomistic simulation. We show that using simply the free energy of transfer from water to octanol to obtain the force parameters does not guarantee the transferability of the model to other solvents. Instead one needs to match the solvation (or hydration) free energies to ensure that the polymer has the correct polarity. We propose a simple method to select the Lennard-Jones parameter to match the solvation free energies for different beads. We also show that, even when the partition coefficient of the monomer is correct, even for modestly high molecular weight of the polymer the predicted partitioning properties could be wrong.

Microporous metal organic frameworks (MOFs) are a novel class of materials formed through self-assembly of inorganic and organic structural building units (SBUs). They show great promise for many applications thanks to record-breaking internal surface areas, high porosity as well as a wide variety of possible chemical compositions. Molecular simulation has been instrumental in the study of MOFs to date, and this thesis work aims to validate and expand upon these efforts through two distinct computational MOF investigations. Current separation technologies used for CO2/N2 mixtures, found in the greenhouse gas-emitting flue gas generated by coal-burning power plants, could greatly benefit from the improved cost-effective separation MOF technology offers. MOFs have shown great potential for CO2 capture due to their low heat capacities and high, selective uptake of CO2. To ensure that simulation techniques effectively predict quantitative MOF gas uptakes and selectivities, it is important that the simulation parameters used, such as force fields, are adequate. We show that in all cases explored, the force field in current widespread use for N2 adsorption over-predicts uptake by at least 50% of the experimental uptake in MOFs. We propose a new N2 model, NIMF (Nitrogen in MoFs), that has been parameterized using experimental N2 uptake data in a diverse range of MOFs found in literature. The NIMF force field yields high accuracy N2 uptakes and will allow for accurate simulated uptakes and selectivities in existing and hypothetical MOF materials and will facilitate accurate identification of promising materials for CO2 capture and storage as well as air separation for oxy-fuel combustion. We also present the results of grand canonical and canonical Monte Carlo (GCMC and canonical MC), DFT and molecular dynamics (MD) simulations as well as charge density analyses, on both CO2 and N,N-dimethylformamide adsorbed in Ba2TMA(NO3) and MIL-68(In), two MOFs with non-equivalent inorganic structural building units. We demonstrate the excellent agreement found between our simulation results and the solid-state NMR (SSNMR) experiments carried out by Professor Yining Huang (Western University) on these two MOFs. Molecular simulation enables discoveries which complement SSNMR such as the number, distribution and dynamics of guest binding sites within a MOF. We show that the combination of SSNMR and molecular simulation forms a powerful analytical procedure for characterizing MOFs, and this novel set of microscopic characterization techniques allows for the optimization of new and existing MOFs.

Dans le cadre d'un projet de modélisation multi-échelle des propriétés mécaniques des polymères semi-cristallins, nous avons entrepris au cours de cette thèse une étude à l'échelle moléculaire. Les polymères semi-cristallins se caractérisent par la coexistence de phases cristalline et amorphe et la modélisation à l'échelle moléculaire de ces matériaux est un défi scientifique. En effet, l'observation expérimentale d'une interface entre le cristal et l'amorphe est encore impossible. Il est donc nécessaire de réaliser des hypothèses pour la construction de cette interface. Par ailleurs, les dimensions des molécules étudiées, les temps de relaxation associés à leur dynamique et la différence de structure entre les phases constituent des difficultés supplémentaires pour la représentation aux échelles de taille et de temps de la modélisation moléculaire. Ce travail a été structuré selon deux axes de recherche : la construction d'un modèle moléculaire de polymère semi-cristallin et la validation de différentes méthodes de calcul des constantes élastiques à l'aide de la simulation moléculaire. L'originalité de ce travail a été, d'une part, la réalisation d'une étude comparative des différentes méthodes de calcul utilisées, et d'autre part, la construction d'un modèle prenant explicitement en compte l'interface entre les phases cristalline et amorphe. / As part of a project aiming to predict mechanical properties of semicrystalline polymers using multi-scale models, we did a numerical study at the molecular level during this thesis. Semicrystalline polymers are special in that they contain two phases: one crystalline and one amorphous. This makes their molecular modelling an actual scientific challenge. The interface between the phases is still not directly observable through experiment and in order to build a model of this interface, many assumptions and hypotheses are to be done. Furthermore, the length of the molecules, relaxation times associated with their dynamics and the difference of internal structure between the phases are parameters that have to be taken into consideration because of the typical scales of time and space in molecular simulation. This work is built along two axes: the construction of a molecular model for semicrystalline polymer and a review of the methods that are proposed to compute mechanical properties at the molecular scale. The originality of this work is, on the one hand, the comparative benchmark of the different computation methods, and, on the other hand, the making of a molecular model which takes explicitly in account the interface between amorphous and crystalline phases.