Search

Heterogeneous catalysis is universally acknowledged as a paramount approach in a wide range of the chemical, biological, and energy industries. Moreover, heterogeneous catalysis has also been applied to solving the problem of many aspects in our daily life, such as environmental enhancement, energy shortage, efficient synthesis. Just as its name implies, heterogeneous catalysis is catalysis where the phase of catalyst differs from the reactants or products. Normally it usually happens at the gas-solid or liquid-solid interface. This PhD theses mainly addresses two types of heterogeneous reactions: the one happens at the interface between gas and solid (chapter 3 and chapter 4), and the other occurs between liquid and solid (chapter 5 and chapter 6). According to the reaction properties, we tackle these two kinds of reactions with different methods. The gas-solid reactions are solved by traditional Density Functional Theory (DFT) method; while liquid-solid reactions are carried out using ab initio Molecular Dynamics (MD) simulation. Chapter 1 focuses on the background and theoretical methods applied in the projects. In chapter 2, fundamental theories and techniques in this thesis are introduced in detail, as well as VASP software packages. Chapter 3 comprehensively introduces the mechanism of Fisher-Tropsch process on stepped Ru(0001) surface. In this part, two most possible mechanisms are both discussed: CO insertion and carbide mechanism. Additionally, the coverage effect is introduced to acquire more realistic results, based on which we reveal the selectivity, activity and ratedetermined step of Fisher-Tropsch process from syngas to C2 compounds. Chapter 4 presents the distinct catalytic performances on three different types of NiO surface. The methane combustion process is investigated as a probe on these surfaces. In this part, activation of methane is studied under different coordination environment of Ni2+, such as sheet NiO, cubic NiO and octahedral NiO represented by NiO(110), NiO(100) and NiO(111), respectively. The mechanism schemes of the pathway are all simulated to search the most favorable methane oxidation route. Chapter 5 presents the calculation on Pt(111)/H2O interface to research the preference for CO adsorption in liquid with an advanced Molecular Dynamic (MD) approach, namely umbrella sampling. Both the hollow site and the top site are simulated for CO adsorption and desorption with free energy chemisorption energy obtained. In chapter 6, the slow-growth (SG) approach and constrained molecular dynamics (CMD) are adopted and combined to address the oxygen evolution reaction (OER) on the interface between water and IrO2(110) surface under electrostatic potential environment. The surface is terminated by adsorbed oxygen atoms on the iridium atoms to account for the electrochemical potential in experiments. The different reaction pathways are discussed to figure out the most possible mechanism. In order to make the reaction environment closer to the realistic condition, a constant potential model is achieved by correcting constant potential charge via workfunction calculations.

Molecular simulation is applied to understanding the behaviour of alkane liquids with the eventual goal of being able to predict the viscosity of an arbitrary alkane mixture from first principles. Such prediction would have numerous scientific and industrial applications, as alkanes are the largest component of fuels, lubricants, and waxes; furthermore, they form the backbones of a myriad of organic compounds. This dissertation details the creation of a potential, a model for how the atoms and molecules in the simulation interact, based on a systematic approximation of the quantum mechanical potential energy surface using machine learning. This approximation has the advantage of producing forces and energies of nearly quantum mechanical accuracy at a tiny fraction of the usual cost. It enables accurate simulation of the large systems and long timescales required for accurate prediction of properties such as the density and viscosity. The approach is developed and tested on methane, the simplest alkane, and investigations are made into potentials for longer, more complex alkanes. The results show that the approach is promising and should be pursued further to create an accurate machine learning potential for the alkanes. It could even be extended to more complex molecular liquids in the future.

The presence of the ordered $\gamma^{\prime}$ phase ($\text{Ni}_{3}\text{Al}$) in Ni-base superalloys is fundamental to the performance of engineering components such as turbine disks and blades which operate at high temperatures and loads. Hence for these alloys it is important to optimize their microstructure and phase composition. This is typically done by varying their chemistry and heat treatment to achieve an appropriate balance between $\gamma^{\prime}$ content and other constituents such as carbides, borides, oxides and topologically close packed phases. In this work we have set out to investigate the onset of $\gamma^{\prime}$ ordering in Ni-Al single crystals and in Ni-Al bicrystals containing coincidence site lattice grain boundaries (GBs) and we do this at high temperatures, which are representative of typical heat treatment schedules including quenching and annealing. For this we use the atomistic simulation methods of molecular dynamics (MD) and density functional theory (DFT). In the first part of this work we develop robust Bayesian classifiers to identify the $\gamma^{\prime}$ phase in large scale simulation boxes at high temperatures around 1500 K. We observe significant \gamma^{\prime} ordering in the simulations in the form of clusters of $\gamma^{\prime}$-like ordered atoms embedded in a $\gamma$ host solid solution and this happens within 100 ns. Single crystals are found to exhibit the expected homogeneous ordering with slight indications of chemical composition change and a positive correlation between the Al concentration and the concentration of $\gamma^{\prime}$ phase. In general, the ordering is found to take place faster in systems with GBs and preferentially adjacent to the GBs. The sole exception to this is the $\Sigma3 \left(111\right)$ tilt GB, which is a coherent twin. An analysis of the ensemble and time lag average displacements of the GBs reveals mostly `anomalous diffusion' behaviour. Increasing the Al content from pure Ni to Ni 20 at.% Al was found to either consistently increase or decrease the mobility of the GB as seen from the changing slope of the time lag displacement average. The movement of the GB can then be characterized as either `super' or `sub-diffusive' and is interpreted in terms of diffusion induced grain boundary migration, which is posited as a possible precursor to the appearance of serrated edge grain boundaries. In the second part of this work we develop a method for the training of empirical interatomic potentials to capture more elements in the alloy system. We focus on the embedded atom method (EAM) and use the Ni-Al system as a test case. Recently, empirical potentials have been developed based on results from DFT which utilize energies and forces, but neglect the electron densities, which are also available. Noting the importance of electron densities, we propose a route to include them into the training of EAM-type potentials via Bayesian linear regression. Electron density models obtained for structures with a range of bonding types are shown to accurately reproduce the electron densities from DFT. Also, the resulting empirical potentials accurately reproduce DFT energies and forces of all the phases considered within the Ni-Al system. Properties not included in the training process, such as stacking fault energies, are sometimes not reproduced with the desired accuracy and the reasons for this are discussed. General regression issues, known to the machine learning community, are identified as the main difficulty facing further development of empirical potentials using this approach.

The warm dense matter regime (WDM), defined by temperatures of a few electron volts and densities comparable with solids, is a complex state of matter where multi-body particle correlations and quantum effects play an important role in determining the overall structure and equation of state. The study of WDM states represents the laboratory analogue of the astrophysical environments found in the cores of planets and in the crusts of old stars, but also has practical applications for controlled thermonuclear fusion. Time resolved X-ray diffraction is used to study the temporal evolution of a sample from solid state towards WDM, either after irradiation with an intense proton/electron beam, in carbon samples, or direct laser illumination, in thin gold nanofoils. The electron-ion equilibration time is extracted through the use of the two-temperature model and in highly excited carbon shown to be longer than previously thought, this is attributed to strong ion-ion coupling screening the interaction (coupled mode theory). Calculation of the dynamic ion-ion structure factor is performed using orbital-free density functional theory (OF-DFT) and shown to compare well with Kohn-Sham DFT in both the static and dynamic cases. Experimental verification of these results is vital and measurement of the microscopic dynamics of warm dense aluminium have been successfully demonstrated through inelastic X-ray scattering. Using the self-seeded beam at the linear coherent light source (LCLS) scattering at a small momentum exchange allowed the first direct measurement of ion acoustic waves in WDM. This data provides the basis for a direct experimental test of many dense plasma theories through direct comparison with the ion-ion dynamic structure factor.

An understanding of the fundamental mechanism behind the relief of shear stress in single-crystal silicon subject to loading by shock-waves has to this day remained elusive. What is known is that this material undergoes a first-order pressure-induced polymorphic phase transition from its ambient pressure cubic-diamond (cd) crystal structure to its first stable high-pressure phase, known as β-Sn, at a pressure of about 120 kbar under hydrostatic compression. By investigating the evolution of the transition parameters for this phase transition as a function of increasing uniaxial shear stress representative of the effects of shock-compression via ab-initio Density Functional Theory computational techniques, we predict a significant lowering of the stress at which the phase transition occurs. This raises the question as to whether the onset of plastic response at the material's Hugoniot Elastic Limit (HEL) reported in experiments corresponds in fact to the phase transition itself, a very plausible possibility which has never been considered before. Furthermore, we present molecular dynamics simulations using a Tersoff-like potential of shock-compressed single crystals of silicon. We find an elastic response up to a critical stress, above which the shear stress is relieved by an inelastic response associated with a partial transformation to a new high-pressure phase, where both the new phase (Imma) and the original cubic diamond phase are under close to hydrostatic conditions. We note that these simulations are also consistent with shear stress relief provided directly by the shock-induced phase transition itself, without an intermediate state of plastic deformation of the cubic diamond phase.

Abstract: The conversion of lignocellulosic biomass to produce chemicals has gained significant interest recently. A broad spectrum of specialty chemicals can be obtained from lignocellulose-derived molecules. One of the most promising reactions in this category is the dehydration of cellulose-derived sugars such as glucose or fructose to 5-hydroxyl methyl furfural (5-HMF). 5-HMF is a valuable chemical platform from which polymers and biofuels can be produced. However, acid-catalyzed dehydration of fructose or glucose to 5-HMF is accompanied by side reactions, forming soluble polymers and insoluble humins. Humins are carbonaceous, polymeric by-products formed during acid-catalyzed condensed phase transformation of biomass-derived moieties. They are responsible for significant carbon loss and catalyst deactivation, and little is known about the chemistry and composition of humins. Additionally, condensed phase catalytic transformation of biomass is highly susceptible to solvent composition, where addition of aprotic solvents has shown to alter conversion and selectivity and can also avoid the formation of unwanted dehydration and rehydration by-products like humins and smaller molecular weight acids. Thus, revealing the mechanism of humins formation and elucidating the effect of solvent medium on humins formation chemistry are crucial knowledge gaps to fill to make quantum leaps in the development of biomass to chemicals technology. In this thesis, for the first time, the reaction pathways and the activation and reaction-free energies of all the elementary reaction steps of the 5-HMF initiated acid-catalyzed humins formation are computed using density functional theory (DFT)-based calculations. The reaction pathway elucidated in this work explains the subsequent chemistry of further polymerization leading to humins. The computed mechanism and suggested step growth polymerization scheme are in excellent agreement with the experimental spectroscopic data in the literature. This work provides mechanistic details of humins chemistry for the subsequent higher-scale kinetic modelling of humins formation and optimizes the reaction conditions to minimize them. The solvent effects are studied at two levels: physical interactions of the solvent with reacting species and preferential solvation as well as their chemical effects in altering reaction kinetics and thermodynamics. A computational approach of molecular mechanics-based molecular dynamics and well-tempered metadynamics methods is implemented to investigate physical effects. The effect of aprotic solvent on the interaction of reacting species with the hydronium ions (acid catalyst) is studied for the fructose dehydration to 5-HMF as a simulation system model. Analogous to heterogeneous catalysis, the interaction between the reactant and the catalyst and the competition between reactant and solvent to interact with the catalyst surface is believed to govern the reactivity. We demonstrate a remarkable enhancement in the interaction between the hydronium ions and the fructose hydroxyl group as the DMSO concentration increased. Also, adding DMSO improves the stabilization of the hydronium ions in the first solvation shell of fructose compared to that in the bulk solvent. On the other hand, hydronium ions became less stable near 5-HMF as the concentration of DMSO increased. These observations help to better understand the improved conversion and selectivity during the acid dehydration of fructose to 5-HMF experimentally observed when adding aprotic polar solvent to the reacting system. Three solvation modelling approaches are systematically compared to study the chemical effects of the solvent environment, i.e., implicit solvent, explicit, non-polarizable, equilibrated solvent molecules without dynamics and explicit solvent molecules with dynamics with polarizability. The model reaction is the acid-catalyzed protonation of 5-HMF, the first charge transfer step, leading to humins formation. The effect of the solvent environment on activation-free energy barriers was only captured when solvent and reaction dynamics were considered and when the solvent molecules were polarizable. The charge redistribution in the reacting species was observed to lead to the dynamic solvent reorganization; as the solvent composition changes, the Marcus reorganization energy changes, altering the reaction-free energy barriers. Additionally, the changes in the timescale of solvent reorganization led to non-equilibrium solvation. This thesis investigates the molecular-level formation of humins from biomass and their derivatives, identifying key intermediates and modeling humin formation in condensed-phase reactions during acid-catalyzed biomass conversion. These findings guide the selection of catalysts, solvents, and operating conditions to enhance desired product selectivity and reduce humin formation in biomass-to-chemical transformations.

The immense synthetic design space and material versatility have driven the exploration and development of organic semiconductors (OSC) over several decades. While many OSC designs focus on the chemistries of the molecular or polymer building blocks, a priori, multiscale control over the solid-state morphology is required for effective application of the active layer in a given technology. However, molecular assembly during solid-state formation is a complex function interconnecting the building block chemistry and the processing environment. Insufficient knowledge as to how these aspects engage, especially at the atomistic and molecular scales, has so far limited the ability to predict OSC solid-state morphology, leaving Edisonian approaches as the stalwart methods. Therefore, through multiscale simulations combining atomistic quantum scale modeling and state-of-the-art molecular dynamics (MD) techniques, we aim to establish first principles understanding required to synthetically regulate solid-state morphology of organic semiconductors (OSC) as a function of molecular chemistry and processing. In turn, we try to understand the deceivingly simple yet complex mechanisms behind molecular aggregation and crystallization of OSC. Simultaneously, we develop semi-to-fully automated high-throughput schemes to automate the complex and labor-intensive analyses to generate data based on various crystal structures in different crystallization environments. Ultimately, we aim to bridge molecular-scale information revealed on solid-state physical organization, understood in the context of chromophore chemistry and the molecular environment, with the macro scale properties to uncover useful guidelines for rational design and morphology regulation of OSC systems.

The development of catalysts for faster, more selective, and more sustainable chemical reactions remains an outstanding goal in chemistry. While advancements in statistical models have greatly improved our predictive ability, physical models remain our most valuable tools for mechanistic understanding. Density functional theory (DFT) calculations provide structural models of reaction intermediates and transition states and can accurately predict thermodynamic quantities. For larger systems, molecular dynamics (MD) simulations using classical force fields model time-dependent movements of protein-ligand complexes. This Thesis combines both approaches to model several designed and engineered catalysts.In Chapter 2, we investigate dirhodium catalysts that were designed to catalyze an enantioselective Si–H insertion. Our experimental collaborators observed that the enantioselectivity of the reaction correlated strongly with the Hammet constant of the carbene’s aryl substituents but lacked a clear understanding as to why. DFT calculations using an Rh2(OAc)4 catalyst showed that the aryl rings of the carbene rotate in the transition state, resulting in one aryl conjugating with the vacant carbene p-orbital. For asymmetrically substituted carbenes, the transition state with the electron-rich aryl conjugated to the vacant carbene p-orbital was always lower in energy, and this energy difference increased with greater electronic differences between the rings. We then showed that because of this geometric constraint, the chiral environment of the designed catalysts favored one enantiomeric transition state over the other. Again, the energy difference between transition states correlated with the differences in Hammet substituent constants between the rings. In Chapter 3, we investigate a laboratory-evolved family of flavin-dependent halogenases with orthogonal selectivity for the chlorination of tryptamine. Few mutations separate the wild type from the most distant mutant, and crystal structures show that there are minimal structural differences between the enzymes. DFT calculations and MD simulations established that a catalytic lysine activates HOCl through general acid catalysis for chlorination of the substrate. Energetic differences in the transition states or Wheland intermediates calculated by DFT showed the intrinsic site selectivity for each substrate. Docking calculations and MD simulations of protein-substrate complexes showed how each enzyme binds the substrate to position a single site closest to the halogenating species, influencing the site selectivity of this reaction. Our physical model was then used to predict site selectivity for several nonnative substrates. Finally, Chapter 4 investigates a set of computationally designed enzymes with varying activity for a chemiluminescence reaction. Initial designs were optimized by site-saturation mutagenesis, and it was not clear why the final proteins were so much more active than the initial designs. Our DFT calculations confirmed the proposed reaction mechanism and showed that anion formation in the substrate was critical for the reaction. Docking calculations and MD simulations showed that LuxSit-i, the most active variant, was the best at stabilizing this anion in the transition states. Changing the substituent resulted in worse stabilization during the simulations, which is consistent with the observed substrate specificity. All three of these chapters explore notable examples of catalyst engineering and provide physical models for their mechanisms. The insight gained here will hopefully lead to future generations of engineered catalysts with continuously improved capabilities.

Ion hydration is a central topic of discussion in the scientific community, given the important role that solvated ions play in fields as biochemistry, electrochemistry, and environmental chemistry. Despite the large amount of experimental and theoretical studies that involve ions in aqueous solutions, a unified molecular description is still missing: existing models fail to accurately capture the intricate ion-water interactions, leading to inconsistencies between experimental and theoretical results. The surface propensity of halide ions at the air/water interface has been debated for over a century, and the recent introduction of new architectures for the description of molecular interactions allows for the development of accurate and efficient models to describe their behavior.Advanced molecular modelling frameworks based on deep neural networks (DNNs) and many-body expansion (MBE) of the energy are explored and their limitations analysed. Data-driven MBE-based models as MB-nrg are able to provide chemical accuracy at great computational efficiency when compared to classical force fields (FFs), DFT-based ab initio molecular dynamics (AIMD), and modern DNN potentials. After developing an efficient active learning (AL) framework for the generation of comprehensive training sets, we proceed with the development of MB-nrg models for the description of the chloride, bromide, and iodide ions; these provide great accuracy in the description of both gas-phase clusters and bulk systems, by closely reproducing coupled cluster (CC) interaction energies and the experimental x-ray absorption spectra. Lastly, we address the errors associated with the use of semi-local exchange-correlation functionals in modeling hydrated ions within density functional theory (DFT): the recent introduction of density-corrected (DC) SCAN functionals provides a solution to the overdelocalization issue typically encountered in semi-local density functional approximations (DFAs), leading to considerable improvements in the energetics and structural features of hydrated ions. This dissertation presents a significant advancement in the understanding of ion hydration, showcasing novel methods for more precise theoretical predictions, and providing a solid foundation for future research in this challenging field.

Chemical production accounts for a substantial fraction of global energy use, and the majority of industrial processes rely heavily on precious metal heterogeneous catalysts. Fundamental knowledge of the surface structure of these metals is an absolute prerequisite for engineering of energy-efficient and sustainable catalysis. In this regard, first-principles atomistic modeling based on density functional theory (DFT) plays a crucial role in the advancement of rational catalyst design. However, the large computational cost of DFT precludes its use for dynamical simulations of realistic systems. Moreover, accurate identification of the active site remains a significant challenge, as surface composition and morphology are inherently dynamic at reaction conditions of interest, especially for soft metals such as Au, Ag, Pt, and Pd. Such restructuring has been shown to play a significant role in catalytic activity of many nanoparticle systems, but mechanistic understanding at the atomic level has remained scarce and challenging. To investigate the dynamics of restructuring and reactivity in catalytic surfaces, we develop and apply novel simulation methods and analysis workflow to model systems of increasing complexity. We first demonstrate the utility and limitations of the simplest static method – coordinate dragging – using an example system with a chemically diverse interface: diffusion of a water molecule through periodically porous graphene derivatives. We then proceed to highlight the power of proper transition state modeling and microkinetic modeling, combined with catalysis and machine learning-enabled spectroscopic analysis, to uncover the dynamic changes in the active site for hydrogen-deuterium exchange reaction in dilute Pd-in-Au catalysts in response to different catalyst treatments. Dynamic problems require dynamic solutions. As such, we employ molecular dynamics (MD) to elucidate the mechanism and timescale of surface restructuring. First, we perform MD simulation with an empirical potential followed by automated analysis to establish seven main classes of restructuring events by which a single Pd atom becomes incorporated into the Ag surface. Next, we develop and apply Bayesian active learning to train a fast and accurate machine-learning force field, which enables large-scale and long-timescale simulation of the restructuring of Pd deposited on Ag. Finally, the method is extended to reactive simulation and mechanistic analysis of H2 turnover on Pt at chemical accuracy. We close the thesis with a proposal of an automated workflow Catadynamics for discovery and characterization of surface dynamics and reactions directly from reactive MD simulations. The work presented in this thesis attests to the promise of machine-learning MD for further empowering computational catalysis and surface science.

Simulating matter at the atomistic scale can accelerate drug design and materials discovery, but the most accurate atomistic simulation methods are prohibitively expensive. In the past decade, machine learning (ML) has emerged as a powerful tool for combining the computational efficiency of classical force fields with the accuracy of quantum-mechanical methods such as density functional theory (DFT). However, most modern ML force fields are difficult to train, requiring thousands of expensive DFT calculations and detailed prior knowledge of the material of interest. In this thesis, I present a closed-loop Bayesian inference method that automates the training of many-body ML force fields using structures drawn "on the fly" from molecular dynamics simulations. Our online active learning algorithm uses the internal uncertainty of a Gaussian process (GP) regression model to decide whether to accept the model prediction or to perform a DFT calculation that updates the training set of the model. To enable large-scale molecular dynamics simulations with the resulting force fields, mean predictions of the GP are mapped onto much faster spline-based and parametric models. I discuss applications to superionic diffusion in silver iodide, hydrogen chemisorption on platinum, and martensitic phase transitions in the shape-memory alloy nickel titanium. The method is made available in an open-source software package called FLARE (Fast Learning of Atomistic Rare Events).

Self-assembled monolayers (SAMs) of alkanethiol molecules have been widely studied over the last three decades because of their diverse applications in the biomedical, nanotechnology, surface science, and electronics. It is also regarded as the model system to study the binding of organic molecules on the metal surfaces via thiol functional group. The robustness of the SAM structure combined with the ease of preparation makes it an ideal candidate for both fundamental and applied research. The structure of the alkanethiol SAMs on the Au(111) substrate is determined by the interplay between the alkyl chain packing and the interaction at the Au-S interface. Although our understanding of the SAM structure has significantly advanced over the last 35 years, an unambiguous atomic description of the Au-S interface and its influence on the chain packing remains elusive. In order to have better control of the SAM structure for different applications, we require a better understanding of the alkanethiol monolayer. In this work, we use a reductionist approach to determine the preferred head group positions driven by the chain packing, and by the interaction at the Au-S interface, separately. We use molecular dynamics (MD) to study the chain packing of the dense phase saturation coverage decanethiol SAM, and density functional theory (DFT) to study the interaction at the Au-S interface using an isolated methanethiol adsorbate. Alkane chains prefer a close-packed structure for the efficient interlocking of the methylene groups that minimizes the energy of the system. The molecular plane adjusts its orientation (molecular twist) depending on the spacing and the symmetry of the head groups to achieve a close packing of the chains. We first constrain the head groups at the high symmetry (√3 × √3)R30° sites to study the preferred combination of the molecular twists. We use this result as our baseline to study the effect of chain packing on the head group offset from the (√3 × √3)R30° sites. The position of the head groups also depends on the interaction at the Au-S interface. The preferred sites are determined by the tetrahedral coordination and the sp3 hybridization of the sulfur head groups. Relaxation and reconstruction (involve adatoms and/or surface vacancies) of the substrate also has significant influence on the preferred adsorption sites. We begin by determining the preferred positions of the head group on the unrelaxed substrate driven by the interaction at the Au-S interface alone. We then use the unrelaxed substrate as our reference to study the effect of substrate relaxation on the head group positions. To simulate a realistic model of the technologically interesting long-chain dense-phase alkanethiol SAM, we need to combine the effect of the chain packing and the interaction at the Au-S interface. Currently, we do not have a site-dependent force field for the Au-S interface to simulate the SAM structure using MD. On the other hand, the size of the problem is computationally too large for the DFT method. We demonstrated an approach to bridge the computational gap by using the atomic structure at the Au-S interface predicted by the DFT to study the effect on the chain packing. Using our DFT results, we predicted the symmetry of the adsorption site dependent dihedral force fields that can be used in MD to improve the prediction of the SAM structure.

Polymeric facilitated transport membranes (FTMs) incorporate reactive carriers (e.g., amino groups) in the polymeric matrix and uniquely utilize the carrier–CO2 reaction to enhance the CO2 permeation. Consequently, FTMs offer both high CO2 permeance and large CO2/N2 selectivity. In this study, a computational methodology was developed to elucidate how carriers influence the intricate interplay of underlying physical and chemical processes that determine the separation performance of FTMs.Commonly adopted amine carriers, including the poly(N-vinylformamide-co-vinylamine) (PNVF-co-VAm) fixed-site carrier and piperazine glycinate (PZ-Gly) and 2-(1-piperazinyl)ethylamine sarcosinate (PZEA-Sar) mobile carriers, were first studied. By using density functional theory (DFT) calculations, molecular dynamics (MD) simulations, and Monte Carlo methods, the carriers were evaluated from the perspectives of reaction chemistry, guest molecule (i.e., mobile carrier, CO2, and N2) diffusion, and N2 sorption. The computational observations were found in good agreement with experimental results, validating the reliability of the computational methodology.In order to improve the CO2 loading capacity, sterically hindered amine carriers were designed and shown to offer improved reaction chemistry. A new parameter of steric hindrance, N exposure, was proposed and validated to facilitate future design of carriers. Another class of reactive carriers with strong basicity have also been evaluated and demonstrated desirable reaction chemistry. These carriers are therefore expected to offer higher CO2 loading capacities than the amine carriers.The diffusion and N2 sorption in promising mobile carriers were then investigated using MD and Monte Carlo simulations, respectively. Sterically hindered amine mobile carriers typically demonstrated lower mobile carrier and CO2 diffusivities. Other candidates were found to offer fast guest molecule diffusion but suffered from larger N2 permeabilities.The water uptake of the FTM was identified to strongly influence gas permeation. Therefore, the computational prediction of water sorption in FTMs was also explored. For the first time to our knowledge, the equilibrium water uptake values of FTMs with PNVF-co-VAm and PZEA-Sar mobile carrier were computationally predicted and found in good agreement with experimental results. This is achieved using the chemical potential of adsorbed water calculated with the continuous fractional component Monte Carlo (CFCMC) method. The higher PZEA-Sar content in the FTM was shown to increase the water uptake value, which promoted the diffusion of CO2 and suppressed N2 sorption. The water sorption behavior, guest diffusion, and N2 sorption in FTMs with sterically hindered amines mobile carriers were subsequently discussed.The proposed computation methodology paves the way toward the future screening studies and design of better-performing carriers in FTMs. The insights generated herein can help navigate future experimental efforts.

In the following work, we discuss the development and application of efficient techniques either aimed at extending the applicability of the currently available methods in the liquid phases to larger system sizes, or at eliminating the artifacts with the existing techniques for several open-shell systems encountered in condensed phases. First, we have investigated the local solvation structure of aqueous hydroxyl radical and absorption spectrum by employing both the mixed quantum mechanics\slash molecular mechanics (QM/MM) framework and periodic density functional theory (DFT) framework. Theoretically, the presence of a hemibond (a two-center, three-electron bond) in this system has been debated for a long time. This structural motif has been explained as either an artifact arising from the self-interaction error (SIE) in DFT or an artifact because of the finite-size effects of the simulation cell but shown to play an important role in the absorption spectrum based on some theoretical studies on smaller representative clusters. Our investigations based on simulations with various DFT simulations suggest that a pseudo-hemibonded motif still persists in this system. We have also demonstrated that the population of hemibonds is extremely sensitive to the amount of exact Hartree-Fock (HF) exchange employed in the simulation. However, we have concluded that these hemibonded motifs play an outsized role in the absorption spectrum, even when present as a rare configuration, due to an intense charge-transfer transition in the hemibonded structures. To eliminate the artifacts arising from this SIE, we have then implemented the density-corrected DFT (DC-DFT) formalism along its analytical gradient, which has been proven to be powerful for the systems with larger density-driven errors. Afterward, we have applied this technique for studying the electronic structure descriptions of the hole defects in Al-doped silica and electron-polarons in anatase titanium oxide where most of the standard DFT functionals fail. Using DC-DFT with some cluster models, we have been successful in describing the proper localized nature of the hole in Al-doped as well as the proper structural distortion around the defect centers. This method has also removed a lot of density-driven error in the case of anatase. Next, we have applied this technique to study the stability ordering of different polymorphs of several organic crystals where standard DFT functionals perform poorly, but the ordering gets improved with the density-corrected versions. Furthermore, we have also done some method development with the QM/MM-Ewald framework by using CM5 charges (charge-model 5) in an effort to push the existing framework in simulating systems with larger QM sizes and longer timescales.

This thesis reports on incorporating various computational methods and tools, including first-principles calculations, mesoscale modeling, stochastic modeling, and Artificial intelli-gence to study structure-property relations in wide bandgap semiconductors. In the first project, we present first-principles calculations that elucidate the growth mecha-nism of ZnO photocatalytic materials for water-splitting applications used in sustainable energy storage. Polar surfaces of ZnO are known to have higher photoelectrochemical ac-tivity increasing water splitting efficiency. However, they are known to be unstable and less probable to form under normal conditions. In this project, we demonstrate these high-energy surfaces could be stabilized under certain growth conditions. This approach sug-gests a general solution for controlling the growth morphology, and it can be applied to other compounds to tailor their structure and obtain materials that are not normally stable. Further, we present a comprehensive study of diamond-based systems for application in high-power and high-frequency electronic devices. This includes two major projects to find solutions for SD of diamond. the first project studies the possibility of using hexagonal Bo-ron Nitride (hBN) and graphene 2D materials as acceptor or interface layer in diamond-based heterostructures. The second project reports modeling and analysis of amorphous vanadium pentoxide slabs as candidates for SD of diamond. Chapter six of this thesis introduces a novel AI-based approach to accelerate computational studies of molecule-surface (and molecule-molecule) interactions. To this end, we present an interactive method that couples Gaussian Processes, Bayesian Inference, and molecular dynamics simulations to accelerate the search for the minimum energy structure in mole-cule-surface interactions. This method addresses the problem of dealing with multiple con-figurations with similar energies. It enables making accurate predictions from relatively small datasets and quantifying the uncertainty associated with each prediction. The last chapter of this thesis introduces a Monte Carlo-based raytracing model, named LightCapture, that simulates light absorption in a microfluidic water-treatment reactor. The LightCapture model was developed to predict geometry - light interaction correlations in reactors consisting of micropillars. This is a critical step in determining the reactor’s overall photocatalytic efficiency. To evaluate the performance, the model was applied to determine light capture efficiency in microreactors that use an array of TiO2 photocatalytic micropil-lars, which are being developed for treating recycled water on spacecraft during deep space missions, and the results were in great agreement with experimental tests. This thesis provides guiding principles for accelerated discovery of wide bandgap semi-conductors. Many of the presented methods and tools have been successfully utilized to provide guiding principles for experimental fabrication.

Point defects in wide-bandgap semiconductors can be used to implement qubits that exploit spin-dependent fluorescence from electronic transitions. These qubits are the quantum mechanical analog of classical bits in that they leverage quantum mechanical properties such as superposition and entanglement and a spin-dependent fluorescence implies the ability to read out the qubit state optically. Fluorescent defects in wide-bandgap semiconductors such as diamond have long been studied, but an electronic spin coherence time exceeding one second at room temperature for the NV−, a defect in diamond comprised of a substitutional nitrogen atom adjacent to a carbon vacancy with a single negative charge, has only relatively recently been demonstrated. The demonstration of long-lived spin memory coupled with the fact that the NV− exhibits spin-dependent fluorescence spurred research on other point defects in diamond and in related materials such as silicon carbide for applications in quantum computing and quantum information as well as in magnetometry. However, maximizing the yield of point defects from the various techniques for creating them such as ion implantation or chemical vapor deposition is still not perfectly understood. Using density functional theory and kinetic Monte Carlo and molecular dynamics simulations, we provide an understanding of the thermodynamics and kinetics of point defect formation and migration that can be used to predict the annealing regimes that should lead to high yields and we provide a general framework that can be used to explain relative yields for various Fermi level regimes.

While optoelectronic devices built from organic semiconductors (OSC) continue to find their way into the commercial landscape, there remain numerous challenges to overcome to supplant traditional semiconductors in many technologies. Chief among these are low performance metrics relative to devices with conventional semiconductors and device stability. In order to overcome these challenges, a wide range of new OSC and processing technologies have been developed. However, there remains limited fundamental understanding of the relationship between molecular structure, packing in the solid state, and the resulting materials properties. Here, we make use of multiscale molecular models and utilize classical all-atom molecular dynamic (MD) simulations and quantum-chemical density functional theory (DFT) to explore the molecular structure-function relationships. Our results have provided distinctive insights into how the chemistries and processing of the OSC impact the material characteristics.

Amorphous (non-crystalline) materials offer new capabilities and applications relative to their corresponding crystalline phases due to their novel and tunable material properties, including but not limited to larger surface area, wider substrate compatibility, stronger corrosion resistance, and lower temperature processing than their crystal counterparts. However, even the most advanced experimental techniques (e.g., in situ NMR and ``inverse" approaches such as Reverse Monte Carlo (RMC) modeling of diffraction data) cannot directly measure the bulk atomic structure of such complex, non-crystalline materials. This, coupled with the lack of a structural theory of amorphous materials, has limited our understanding about the structure-property relationships required for the optimization of realistic devices. In this thesis, we compare and combine several atomistic modeling techniques for generating accurate and general structures of amorphous systems, both solid and liquid. We develop several metrics for measuring efficiency of configuration space sampling to guide and standardize the computational study of amorphous materials, and use these to differentiate between different samples of amorphous structures. Lastly, we use these structures to help elucidate structure-property relationships in these systems. In amorphous metal oxides, we characterize the types of coordination defects in amorphous metal oxide systems and subsequently identify potential electron and hole traps in their electronic structures. In a liquid system, water-in-salt electrolytes, we evaluate how the hydrogen bonding network and ion coordination shells change with composition, and how these are directly correlated to battery performance.

Mesostructured porous materials, especially metal-organic framework systems (MOFs), have received widespread attention by researchers over the course of several decades, aimed at exploiting their suitability as sensors, catalysts, adsorption materials, etc. These materials may exhibit high thermal stability for a variety of physical and chemical reactions, benefiting from their unique structural features such as their crystalline order and high internal surface area. Recently, within this class of materials, Covalent Organic Frameworks (COFs) are of particular interest, as they are generally composed of light elements, and offer nanoporous crystalline structure. In this work, we offer an atomic-scale view of the recently synthesized 3D woven COFs, known formally as COF-505-Cu and COF-506-Cu, using molecular simulation techniques, including ab initio Density Functional Theory (DFT) and Molecular Dynamics (MD). First, we explored the reversible demetallation of COF-505-Cu. Serving as the “linkers” of the structure, the covalent bonding between the metal atom and the organic thread is essential to maintain the overall weave of the structure. We proposed two distinct pathways for the demetallation reaction that breaks the covalent bonding, and calculated their respective free energy barriers. We also tested the re-insertion of other, as yet experimentally untested transition metals (Zn and Fe), to the structure, and found that re-formation of the covalent bonding with the metal is strongly affected by the nature of the ionic complex being inserted. Second, we studied the second-generation, non-interpenetrated, COF-506-Cu that was first synthesized in 2018. We adopted the potential function parameters devised for its predecessor (COF-505) to model this material. This force field model was then used to characterize the nature and rate of guest molecule diffusion and uptake in the COF-506-Cu weave. We found that the presence of more guest-accessible space inside the weave greatly increased the diffusion of small molecules, compared to the limited diffusion observed by Acevedo in COF-505. But for other molecules, like the larger THF and methyl orange, movement inside the weave is still constrained through molecular interactions to be tortuous and relatively slow; this was suggested without direct observation in the experimental work. The computationally predicted uptake of THF vapor compared quantitatively well with the experimental data. Those results provided strong validation of the original Acevedo force field model. Looking to the future, the validation of this model could provide a way to predict the efficacy of COF-506, or other variants of this woven COF, ahead of experimental synthesis or applications such as molecule capture (adsorption), separation processes, or catalysis.

Aqueous-phase heterogeneous catalysis is an important chemical process in applications such as water remediation, fuel cells, and the production of fuels and chemicals, including from biomass sources. However, designing alternative, improved catalyst materials for these applications is difficult due to fluctuations in the solvation environment surrounding the catalytic species. In order to elucidate the thermodynamics and kinetics of the relevant reactions, it is imperative to gain a better understanding of the roles of liquid water molecules in these reactions. In this work, a method combining both classical and quantum simulations was developed to generate configurations of liquid water molecules over catalytic species adsorbed on a catalyst surface, which can provide valuable insight into the roles of the liquid water reaction environment on aqueous-phase heterogeneous catalysis. The method developed in this work entails combining force field molecular dynamics (FFMD) and density functional theory (DFT) simulations. This method leverages the strengths of each type of simulation to enable the calculation of catalytic energies under “realistic” liquid water configurations. FFMD simulations are used to generate trajectories of liquid water configurations that include thermal fluctuations, while DFT simulations are used to capture the energies associated with bond breaking and forming that are required for microkinetic modeling and catalyst design studies. This FFMD-DFT method was used to calculate the interaction energies between the liquid water environment and the reaction intermediate or transition state species. The trend in the calculated interaction energies was shown to correlate with the trend in hydrogen-bond formation between liquid water molecules and the catalytic species. This work also demonstrated that entropic effects due to the thermal fluctuations in the solvation environment are a significant contribution to the free energies calculated for aqueous-phase, heterogeneously-catalyzed systems. The FFMD-DFT method was also used to calculate reaction energies, activation barriers, and pre-exponential factors to study the kinetics of example O—H and C—H cleavage reactions on a platinum catalyst surface under an aqueous reaction environment. Using this method, it was found that O—H cleavage reactions prefer H2O-mediated pathways, while C—H cleavage reactions prefer non-H2O-mediated pathways. In summary, the FFMD-DFT method developed in this work has been shown to be a robust technique for generating realistic liquid water configurations over catalytic species on a platinum catalyst surface. Those liquid water configurations can be used to calculate catalytic properties that can provide insight into the roles of water molecules in these reactions and facilitate microkinetic modeling and catalyst design studies.

Materials with features at the nanoscale can provide unique mechanical properties and increased functionality when included as part of a nanocomposite. This dissertation utilizes computational methods at multiple scales, including molecular dynamics (MD) and density functional theory (DFT), and the coupled atomistic and discrete dislocation multiscale method (CADD), to predict the mechanical properties of nanocomposites possessing nanomaterials that are either 1-D (carbyne chains), 2-D (graphene sheets), or 3-D (Al/amorphous-Si core-shell nanorod). The MD method is used to model Ni-graphene nanocomposites. The strength of a Ni-graphene nanocomposite is found to improve by increasing the gap between the graphene sheet and a crack embedded in the Ni matrix. Ni-graphene nanocomposites also show substantially greater strength than pure Ni, depending on the loading direction and crack orientation relative to the graphene sheet. Moreover, polycrystalline graphene may serve as a better reinforce in Ni-graphene nanocomposites due to its improved interfacial shear stress with the Ni matrix compared to pristine graphene. This work develops a patchwork quilt method for generating polycrystalline graphene sheets for use in MD models. Carbyne-based nanocomposites are modeled from first principles using DFT. This research finds that carbyne can only serve as an effective reinforcement in Ni-based nanocomposites when it is dielectrically screened from the Ni matrix, otherwise the carbyne structure is lost. When graphene is used as a dielectric screen, the local stiffness of the nanocomposite improves with the number of carbyne chains present. Specific stiffness is introduced as an alternative to elastic stiffness for characterizing low-dimensional materials because it is not dependent on volume when derived using an energy vs. strain relation. A two-material formulation of CADD is developed to model Al/a-Si core-shell nanorods under indentation/retraction. The structural deformation behavior is found to be dependent on the geometry of both core and shell. When present, the a-Si shell protects the Al core by delocalizing forces produced by the indenter. It is also found that substrate deformation becomes important for core-shell structures with sufficiently small cores. This work can help guide experimental and computational work related to the discussed 1-D, 2-D and 3-D nanomaterials and aid in future nanocomposite design.

Titanium (Ti) and its alloys have a wide range of applications in biomedical, automotive and aerospace industries due to their excellent strength to weight ratio and corrosion resistance. Alpha phase Ti has hexagonal closed packed (hcp) structure that shows anisotropic plastic deformation; 〈 a 〉 type slip on prism planes is the easiest to activate but cannot accommodate deformation along the 〈 c 〉 axis. The low temperature ductility of Ti is linked to twinning. Therefore, understanding the mechanisms behind the twin nucleation and growth in Ti alloys is important from both theoretical and industrial application points of view. To that end, the present study seeks a better understanding of the atomic scale processes involved in twin nucleation mechanisms and the effect of alpha-stabilizing solutes such as interstitial oxygen, substitutional aluminum and rare earth elements on twinning. Systematic molecular dynamics (MD) simulations are used to identify the underlying mechanism of twin nucleation from dislocation/grain boundary interactions. Density functional theory (DFT) simulations are employed to examine the effect of oxygen interstitials on the twinning behavior of Ti. A systematic framework has been developed to predict the diffusion of interstitial elements near the twin boundaries in hcp alloys. Next, uncertainty that arises from first-principles calculations in predicting diffusion coefficients are quantified. Finally, solute segregation to the twin boundaries as a new mechanism for dynamic strain aging (DSA) is investigated in Ti and other hcp alloys.

Atom probe tomography (APT) produces 3D, atom-by-atom images of a sample microstructure that provide impressive detail of materials systems. With advancements in detector and laser technology in the last couple decades, APT has gained new life and has established itself as a premier technique in materials characterization. New laser technology has allowed for data collection rates of 100M atoms/hour or more while new detector technology has pushed detection rates up to and beyond 80%. The technique holds huge promise for providing a direct link between the actual atomic microstructure of a material and its properties. Being a destructive technique by nature, and ultimately relying on observing the surface of the sample, the atom probe suffers from artifacts in data that are sometimes not apparent. Reconstruction algorithms, too, have been hampered in adapting to these artifacts, as for many of them the mechanisms are simply not known. Development of physically complete, accurate forward simulations of the field evaporation process can provide both the mechanistic information and a testing framework necessary for future advances in reconstruction algorithms.The work presented in this thesis is focused on expanding the capabilities of current atom probe simulations though application of both DFT and classical atomistic modeling techniques. Chapters 1-3 describe the theoretical underpinnings along with the details of the DFT calculations and atom probe simulation codes used in this research. Three projects focused on creating a more physically detailed model for simulation of atom probe are presented.Chapters 4 and 5 present work done developing a model for the site specific evaporation field of pure FCC emitters. This site specific model is parametrized in terms of the neighbor counts of an evaporated surface atom in order to more closely match the real evaporation sequence. In chapter 4, it is shown that this modified evaporation sequence is the origin of the "flower-petal" features in the aluminum (100) desorption pattern. The limitations of this simple neighbor model are then shown in chapter 5 when it is applied to the evaporation of nickel emitters.Chapter 6 presents a study using a modified version of the atom probe simulation code, TAPSim, which allows for the performance of molecular dynamics on simulated emitters in between evaporation steps. TAPSim-MD simulations of an Al0.98Cu0.02 emitter reveal the origin of the clustering artifact seen in experimental observations of AlCu alloys. The drag effect, an athermal surface migration process that occurs partly due to a non-ideal evaporation sequence of the alloy emitter, a new mechanism for these types of atom probe artifacts is described.Finally, DFT calculations of the surface binding energy of the dilute limit for alloyed adatoms are presented in chapter 7. Fitting a simple exponential function to the per-bond energy of the studied evaporations yields better results than fitting to the ZBEF computed by the Image Hump model, since the desorption energy should be neighbor dependent but the work function is surface dependent. The results of these DFT calculations suggest that for some alloyed elements, namely Mg when alloyed in Al, reduction of the ZBEF of the solute may occur. The weaker binding energy also leads to longer surface bonds, meaning Mg will be naturally more exposed, leading to a higher probability of undergoing evaporation before reaching a terrace edge. Predicting this type of situation is important so that it is not convoluted with other artifacts, such as the drag effect.While large advancements have been made in the hardware necessary to run the atom probe, advances in the development of reconstruction algorithms have not been so quick. Much of this stems from a lack of information about how exactly samples behave while in the instrument and how field evaporation proceeds for a given sample composition. Development of forward simulations with more complete physical models has shown promise for identifying the mechanisms for atom probe artifacts, and will be crucial as part of a testing framework for new reconstruction protocols. As the atom probe becomes more popular, the need for verifying and improving the quality of the data reconstructions becomes more important.

redictive modeling of fluid phase and sorption equilibria for reacting systems presents one of the grand challenges in the field of molecular simulation. Difficulties in the study of such systems arise from the need (i) to accurately model both strong, short-ranged interactions leading to the formation of chemical bonds and weak interactions representing the environment, and (ii) to sample the range of time scales involving frequent molecular collisions, slow diffusion, and infrequent reactive events. This thesis showcases some of my efforts in developing and applying advanced simulation methods to a variety of important systems. Chapters 2 and 3 describe how a novel Monte Carlo method (reactive first principles Monte Carlo or RxFPMC) can be used to overcome some limitations of existing methods for simulation of reactive systems. Chapter 4 shows how advanced sampling techniques in combination with sophisticated interatomic potentials can be used to elucidate nucleation pathways. Chapters 5 and 6 manifest how first principles simulations can be leveraged to understand liquid structure of novel complex solvents as well as reactive processes in such solvents. Finally, the last chapter discusses the use of smart sampling algorithms to study chemisorption of mixed ligands on nanoparticles.

The first part of this dissertation describes computational studies of higher-order cycloadditions with a focus on elucidating the origins of periselectivity. These reactions involve ambimodal transition states (TS) and bifurcations of potential energy surface.Chapter 1 describes density functional theory (DFT) studies of a transannular [6+4] cycloaddition proposed in the biosynthesis of heronamide A, in which an unbridged 10-membered ring is formed. Ring and steric strains are found to be essential in controlling the product stability that makes this reaction feasible. Chapter 2 presents an Environment-Perturbed Transition State Sampling method to explore the mechanism of the enzyme SpnF-catalyzed Diels–Alder. A [6+4] cycloaddition is also involved and enzyme enhances the formation of [4+2] product, with respect to the counterparts in the gas phase and in water. Chapter 3 explores the mechanisms and selectivities of the cycloadditions of tropone to dimethylfulvene discovered in 1967 by Houk. The two proposed pathways by which a key intermediate are formed are united through the discovery of an ambimodal [6+4]/[4+6] TS using DFT calculations and molecular dynamics (MD) simulations.The second part of this dissertation describes computational studies of Diels–Alder reactions and 1,3-dipolar cycloadditions using DFT calculations, with a focus on the mechanisms and the origins of regio- and stereoselectivities.Chapter 4 reports the biochemical characterization of a Diels–Alderase found in the biosynthetic pathway of the cytotoxic myceliothermophin natural products. A theozyme model rationalizes both the substrate- and stereoselectivity of the enzyme. Chapter 5 discusses the concerted versus stepwise mechanisms of a series of dehydro-Diels–Alder (DDA) reactions. The reactivity of DDA reactions is controlled by the distortion energies required to achieve the TS geometries. Chapter 6 explores the mechanisms of Lewis acid-catalyzed Diels–Alder reactions of aryl allenes and acrylates. A stepwise mechanism involving short-lived zwitterion intermediates is established. The [2+2] cycloaddition is not observed experimentally because of the greater charge separation in the first step of the [2+2] cycloaddition. Chapter 7 focuses on the regioselectivity of 1,3-dipolar cycloadditions of benzo and mesitonitrile oxides with alkynyl pinacol and MIDA boronates. Calculated relative free energies of activation reproduce the experimentally observed product ratios. The electronic energies of activation are mainly controlled by distortion energies. Chapter 8 describes DFT studies of the first example of diazo esters as dienophiles in intramolecular Diels–Alder reactions with dienes. For comparison, the reactivities of diazo esters as 1,3-dipoles withdienes have also been explored. The usually observed 1,3-dipolar cycloaddition was not observed because of strong tether distortion in the (3+2) TS.

Heat transfer is ubiquitous in both naturally occurring and engineered materials. As technology progresses, the length- and time-scales of thermal transport decreases, becoming comparable with the mean free paths and relaxation times of the vibrations that drive it. Increasingly, an atomistic-level understanding of thermal transport is pivotal in predicting and controlling heat transport in materials and devices. Modeling approaches that permit an atomistic understanding of heat transport, and the implementation of complex approximations of the phonon Boltzmann transport formalism include classical molecular dynamics and density functional theory (DFT). Results are presented for equilibrium molecular dynamics (EMD) simulations of the thermal conductivity of a series of clustering and non-clustering point defects in graphite using the Green–Kubo method, aimed to advance our knowledge of the evolution of the microstructure of graphite while in service in a graphite-moderated nuclear reactor. The Green–Kubo method — commonly used for predicting transport properties by scientists and engineers across fields — relates the property of interest to the lifetime of fluctuations in its thermodynamic driving potential. The integral of the autocorrelation fluctuations requires a long averaging time to reduce remnant noise and is a principal source of error. A new approach is proposed to quantify — on-the-fly — the uncertainty on transport properties computed using the Green-Kubo formulation, based on recognizing that the integrated noise is a random walk. EMD is also used to explore thermal transport in breathing metal--organic frameworks (MOFs), coveted for numerous applications due to their large surface area and modularity. A simple geometric model of thermal conductivity is proposed as a heuristic for the quick evaluation of transport in flexible MOFs, and a quantum-based approach is undertaken to explore deviations from the heuristic, such as rattler modes and phonon-focusing. Phonon properties calculated with DFT for (1) uranium dioxide and (2) silicon, for fuel and spintronics applications respectively, are also briefly discussed.

Solvation and solvent effects play an important role in diverse chemical processes ranging from reaction kinetics to molecular recognition, solubility, solvato-chromism and phase separations. Despite enormous activities in this field, quantitative solvation calculations remain an enormous intellectual challenge. My thesis is focused on development and application of molecular density functional theory (MDFT) and molecular dynamics (MD) simulation to predicting solvation properties. Accomplishments include 1) improved the average unsigned error of MDFT predictions for the room-temperature solvation free energies (SFE) of 504 pharmaceutical molecules in water from 1.04 kcal/mol to 0.66 kcal/mol; 2) established a more reliable numerical procedure to calculate the direct correlation functions (DCF) of solvent from MD simulations; 3) extended MDFT prediction of SFE to different temperatures and calibrated the theoretical results with experimental data for the hydration free energies of 5 nitrotolunenes and a library of 197 solutes at 277 K, 298 K and 313 K. In addition, I investigated the 3-dimensional (3D) solvation structure of amine-grafted silica gel in liquid water by applying a spherical harmonics expansion method to the MD trajectories. The simulation results provide evidence on the strong influence of the silica surface on hydration structure, which is often ignored in the theoretical analysis of surface reactions. Furthermore, I developed a hybrid method for predicting the SFE of spherical ions by combining MDFT with MD simulations. The numerical analysis justifies the universality of the bridge functional that can be reasonably approximated by the modified fundamental measure theory (MFMT) for hard-sphere systems.Results from this thesis demonstrate that the DCFs are important in application of MDFT to SFE predictions. Based DCF from on integral-equation methods, MDFT can also capture the temperature effect on SFE in good agreement with experiment. In addition, the hybrid MDFT-MD method provides accurate predictions of hydration free energies for charged solutes and the numerical analysis sheds light on future theoretical development. The efficient sampling method for generating 3D density profiles from MD may open up opportunities for application of MDFT to more complex systems, for example, protein solvation and enzyme kinetics. By studying the solvation structure of amine-grafted silica shell, I found that the silica surface affects not only the distribution of surrounding water but also the hydrogen-bonding network. This surface effect is long-ranged and can be reduced with longer grafted amine chains.

Supercapacitors are electrochemical energy storage devices known for their large power densities and long lifetimes yet limited energy densities. A conventional understanding of supercapacitors relates the high power to fast ion accumulation at the polarized electrode interface, forming the so-called electric double layer (EDL), and the low energy to limited electrode surface area (SA). To improve the energy density, the capacitance may be enhanced by using high SA electrode materials such as carbon-based nanomaterials. While promising results have been experimentally reported, capacitances have also been noted to exhibit a highly non-linear relationship with SA. These interesting observations suggest that a gap exists in our fundamental understanding of charge storage mechanisms in the EDL of carbon nanomaterials. Given that EDLs are typically on the order of 1-3 nm thick, theoretical simulations can elucidate these unknown physical insights in order to identify new design principles for future electrode materials. In this dissertation, we explore two broad types of carbon-based nanomaterials, which are separated into two Parts, using a combined density functional theory and classical molecular dynamics computational approach. In Part I, we study the capacitance using various chemically and/or structurally modified graphene (or graphene-derived) materials which is motivated by previous accounts of the limited capacitance using pristine graphene. Our analysis demonstrates the viability of dramatically improving the capacitance using graphene-derived materials owing to enhancements in the quantum capacitance with marginal effects on the double layer capacitance. In Part II, we investigate the capacitance using nanoporous carbon materials which is motivated by experimental observations that relate capacitance to pore width rather than SA. Our findings confirm that promoting ion confinement through pore width control can enhance capacitance, but also identify pore shape dispersity as another important structural feature that facilitates fast ion dynamics during charging/discharging. The work in this dissertation presents an overview of new insights into charge storage mechanisms using low-dimensional carbon-based nanomaterials and future directions for materials development. Moreover, we anticipate that the established methodologies and analyses can be broadly applicable to the study of other applications utilizing electrified interfaces, including capacitive deionization and liquid-gated field effect transistors.

This thesis discusses a general method for constructing interatomic potentials for crystal lattices based on a truncated Taylor series expansion. Such an interatomic potential may be constructed analytically, as in the case of a Lennard-Jones solid, or using ab initio methods such as Density Functional Theory. Specifically, it addresses the scope of application of the method, and demonstrates its practical importance in capturing anharmonicity for a Lennard-Jones solid. In particular, the third-order terms in the truncated potential are shown to accurately approximate the thermal conductivity of the standard interaction Lennard-Jones potential. The thesis also describes an efficient algorithm for locating the equilibrium lattice site of an atom in a three-dimensional crystal lattice displaced from its equilibrium position.In addition, a procedure is outlined to compute the coefficients in the truncated Taylor series expansion using Density Functional Theory, which is an ab initio method. The procedure is applied to germanium, a material for which, an exact analytical interatomic potential is not known. The computed coefficients may then be used to construct an interatomic potential for germanium of ab initio accuracy. This interatomic potential is then used in MD simulations to estimate various thermophysical properties of a germanium crystal lattice within the range of applicability of this method.

Phthalocyanines (Pc) have gained intense research attention in many diverse application areas due to their highly tunable electronic and structural properties through modification of the molecular periphery and metal center. Throughout this work a series of novel perfluoro-isopropyl substituted MPc have been investigated through theoretical methods. First, the synthetic mechanisms of these Pcs will be explored to gain insight into the experimentally observed Pc product distribution. By examining the electronic structure and formation energies of the various Pc precursors, we explain the product distribution as well as propose the formation of additional Pcs, which were not currently believed to form. The effect of metal center and peripheral modification on the Pc structural and electronic properties is also determined through a systematic investigation of several Pcs with varying degree of peripheral modification as well as several different metal centers. Increased modification of the Pc periphery with strongly electron withdrawing groups lowers the energy of the molecular frontier orbitals; increasing the chemical stability of the Pc. Open d-shell metal centers also introduce several partially occupied states near the top of the Pc valence band, which have electron density localized on the metal center. The bulky groups on the periphery of the Pc also act to mitigate molecular aggregation. To access the degree of aggregation as a function of peripheral modification, a molecular dynamics forcefield within the CHARMM parameterization model was developed specific to these Pcs. This also allows for the simulation of bulk and thin film properties important to various application areas. Finally, we propose a completely solid state dye sensitized solar cell (DSSC) design in which these chemically robust modified Pcs are sandwiched between n-TiO2 and p-NiO, acting as both photosensitizer and electron shuttle. Through analysis of the electronic structure of the Pc|semiconductor systems, the free energy associated with hole injection into the valence band of NiO upon photoexcitation of the sensitizer and electron injection into the conduction band of TiO2 from the reduced form of the Pc are calculated. Significant molecular orbital coupling between the Pc and semiconductors results in estimated charge transfer lifetimes on the femtosecond time scale on both NiO and TiO2. Additionally, the calculated excited state lifetimes of the Pc is found to be on the nanoseconds time scale, allowing ample time for charge transfer prior to the spontaneous relaxation of the Pc excited state. In the absence of a liquid electrolyte solution, the Pc molecule will need to also act as electron shuttle in our cell design. The charge transfer properties within the Marcus-Hush electron transfer theoretical framework are calculated. Results indicate that intermediate modification of the Pc periphery leads to high hole and electron mobilities. This is a promising result for our proposed DSSC design, but also makes these Pcs a viable semiconducting material in other application areas, such as light emitting diodes (LEDs) or organic field effect transistors (OFETs).

Research into actinides is of high interest because of their potential applications as an energy source and for the environmental implications therein. Global concern has arisen since the development of the actinide concept in the 1940s led to the industrial scale use of the commercial nuclear energy cycle and nuclear weapons production. Large quantities of waste have been generated from these processes inspiring efforts to address fundamental questions in actinide science. In this regard, the objective of this work is to use theory to provide insight and predictions into actinide chemistry, where experimental work is extremely challenging because of the intrinsic difficulties of the experiments themselves and the safety issues associated with this type of chemistry. This thesis is a collection of theoretical studies of actinide chemistry falling into three categories: quantum chemical and matrix isolation studies of small molecules, the electronic structure of organoactinide systems, and uranyl peroxide nanoclusters and other solid state actinide compounds. The work herein not only spans a wide range of systems size but also investigates a range of chemical problems. Various quantum chemical approaches have been employed. Wave function-based methods have been used to study the electronic structure of actinide containing molecules of small to middle-size. Among these methods, the complete active space self consistent field (CASSCF) approach with corrections from second-order perturbation theory (CASPT2), the generalized active space SCF (GASSCF) approach, and Møller-Plesset second-order perturbation theory (MP2) have been employed. Likewise, density functional theory (DFT) has been used along with analysis tools like bond energy decomposition, bond orders, and Bader's Atoms in Molecules. From these quantum chemical results, comparison with experimentally obtained structures and spectra are made.

The main focus of the thesis is to have better understanding of the atomic and electronic structures, vibrational dynamics and thermodynamics of metallic surfaces and bi-metallic nanoparticles (NPs) via a multi-scale simulational approach. The research presented here involves the study of the physical and chemical properties of metallic surfaces and NPs that are useful to determine their functionality in building novel materials. The study follows the 'bottom-up' approach for which the knowledge gathered at the scale of atoms and NPs serves as a base to build, at the macroscopic scale, materials with desired physical and chemical properties. We use a variety of theoretical and computational tools with different degrees of accuracy to study problems in different time and length scales. Interactions between the atoms are derived using both Density Functional Theory (DFT) and Embedded Atom Method (EAM), depending on the scale of the problem at hand. For some cases, both methods are used for the purpose of comparison. For revealing the local contributions to the vibrational dynamics and thermodynamics for the systems possessing site-specific environments, a local approach in real-space is used, namely Real Space Green's Function method (RSGF). For simulating diffusion of atoms/clusters and growth on metal surfaces, Molecular Statics (MS) and Molecular Dynamics (MD) methods are employed.

In the first part of this thesis, we present results from extensive molecular dynamics simulations of the collapse transitions of hydrophobic polymers in explicit water. The focus is to understand the roles that curvature and interactions associated with the polymer-water “interface” have on collapse thermodynamics. We show that model hydrophobic polymers can have parabolic, protein-like, temperature-dependent free energies of unfolding. Analysis of the water structure shows that the polymer-water interface can be characterized as soft and weakly dewetted. We also show that an appropriately defined surface tension for the polymer-water interface is independent of the attractive polymer-water interactions. This helped us to develop a perturbation model for predicting the effect of attractions on polymer collapse thermodynamics. In the second part, we explore connections between structure, thermodynamics, and dynamics of inhomogeneous fluids. First, we use molecular dynamics simulations and classical density functional theory (DFT) to study the hard-sphere fluid at approximately 103 equilibrium state points, spanning different confining geometries and particle-boundary interactions. We provide strong empirical evidence that both excess entropy and a new generalized measure of available volume for inhomogeneous fluids correlate excellently with self-diffusivity, approximately independent of the degree of confinement. Next, we study via simulations how tuning particle-wall interactions to flatten or enhance the particle layering of a model confined fluid impacts its self-diffusivity, viscosity, and entropy. Interestingly, interactions that eliminate particle layering can significantly reduce confined fluid mobility, whereas those that enhance layering can have the opposite effect. Excess entropy helps to understand and predict these trends. Finally, we explore the relationships between the effective interparticle interactions, static structure, and tracer diffusivity of a solute in a mixture. We show that knowledge of these relationships can allow one to “tune” the effective interparticle interactions of the solute in a way that increases its tracer diffusivity. One interesting consequence is that the mobility of a hard-sphere solute can be increased by adding a soft-repulsion to its interaction, effectively making it bigger.

As the primary medium through which planetary differentiation occurs, silicate liquids are key in the study of the thermal and chemical evolution of Earth. First principles molecular dynamics simulations were used to study the liquid state physics of the MgO - SiO2 join at pressure and temperature conditions relevant to the deep interiors of Earth-like planets, with special focus on the variation of liquid state thermodynamics and structure with pressure, temperature and composition. We find the structure of liquids to change continuously upon compression, and to differ markedly from that of the respective isochemical crystalline polymorphs. Si-O coordination increases from four to six upon two-fold compression in all compositions considered, with the increase strongly delayed in pure silica. Changes in liquid structure is expressed in the liquid state thermodynamic properties. A density crossover along the forsterite melting curve is found within the stability field of the mineral, a feature which a melting curve computed through the Lindemann criterion from the mean squared atomic displacements in forsterite is unable to reproduce. Composition dependent structural differences within the liquid are expressed as a liquid immiscibility field at low pressure in high silica compositions. Using our first principles molecular dynamics results, we develop a self-consistent thermodynamic description of liquid state thermodynamics, which also accounts for the thermal electronic contribution to the free energy, relevant to silicate liquids over a large range of pressures and temperature. With liquid state thermodynamics thus self-consistently constrained, we investigate the high pressure melting of MgO periclase and MgSiO3 perovskite. By calculating theoretical solid and liquid Hugoniot loci, we predict the changes in density and sound velocity which would be expected during shock melting of periclase and enstatite. We also apply the thermodynamic description to the thermodynamics of mixing along the extent of the binary. At low pressure the enthalpy of mixing is notably pressure dependent, primarily due to the disappearance of a maximum at high silica compositions with an increase in pressure. The structural mechanism responsible for liquid immiscibility is identified, and found to be stable only at low pressure.

Numerous enzymatic reactions are controlled by the chemistry of metallic ions. This dissertation investigates the electronic properties of three transition metal (copper, chromium, and nickel) complexes and describes modeling studies performed on glutathione synthetase. (1) Copper nitrene complexes were computationally characterized, as these complexes have yet to be experimentally isolated. (2) Multireference calculations were carried out on a symmetric C2v chromium dimer derived from the crystal structure of the [(tBu3SiO)Cr(µ-OSitBu3)]2 complex. (3) The T-shaped geometry of a three-coordinate β-diketiminate nickel(I) complex with a CO ligand was compared and contrasted with isoelectronic and isosteric copper(II) complexes. (4) Glutathione synthetase (GS), an enzyme that belongs to the ATP-grasp superfamily, catalyzes the (Mg, ATP)-dependent biosynthesis of glutathione (GSH) from γ-glutamylcysteine and glycine. The free and reactant forms of human GS (wild-type and glycine mutants) were modeled computationally by employing molecular dynamics simulations, as these currently have not been structurally characterized.

Density functional theory calculations are used to unravel the mysterious third step of aromatase catalysis. The feasibility of mechanisms in which the reduced ferrous dioxygen intermediate mediates androgen aromatization are explored and determined to be unlikely. Mechanisms for the aromatization/deformylation sequence which are initiated by 1beta-hydrogen atom abstraction by P450 Compound I are considered. 1beta-Hydrogen atom abstraction from substrates in the presence of the 2,3-enol encounters strikingly low barriers,whereas barriers for this same process are higher in the keto tautomer. Transition states for 1beta-hydrogen atom abstraction from enolized substrates in the presence of the 19-gem-diol decayed directly to the experimentally observed products. If the C19 aldehyde remains unhydrated, aromatization occurs with concomitant decarbonylation, and therefore does not support dehydration of the C19 aldehyde prior to the final catalytic step. Ab initio molecular dynamics on the confirmed that the 1beta-hydrogen atom abstraction and deformylation or decarbonylation occur in a non-synchronous, coordinated manner. These calculations support a dehydrogenase behavior of aromatase in the final catalytic step, which can be summarized by 1beta-hydrogen atom abstraction followed by gem-diol deprotonation. Aromatase, a cytochrome P450 hemoprotein that is responsible for estrogen biosynthesis by conversion of androgens into estrogens, has been an attractive target in the treatment of hormone-dependent breast cancer. There are several classes of natural products that exert potent activities in aromatase inhibition, with the flavonoids being most prominent. Previous studies have exploited flavone and flavanone scaffolds for the development of new aromatase inhibitors. In this dissertation, we describe design, synthesis, and biological evaluation of a novel series of 2-(4-pyridylmethyl)thioisoflavones as the first example of synthetic isoflavone-based aromatase inhibitors. The biological evaluation of a series of 2-azole and 2-thioazole isoflavones as potential aromatase inhibitors are described. Differences in inhibitory activity of triazole and imidazole inhibitors are rationalized with density functional theory to expose a key difference in the electronic structure of these molecules. In addition, difference binding spectra of inhibitors to immunoaffinity-purified aromatase produces classical Type II spectra consistent with coordination of the nitrogen lone pair electrons to the P450 heme.