Solvent effects play diverse roles in myriads of chemical, physical and biological processes. The solvent interacts with the solute by: i) non-specific (Coulombic, van der Waals interactions) and ii) specific interactions (hydrogen bonding, etc.). These interactions are responsible for solvation of the solute and are collectively termed as “solvent polarity”. Differential solvation of ground and excited electronic states is manifested in the absorption spectrum as a change in the band position, intensity or shape, which is termed as “solvatochromism”. Intermolecular hydrogen bonding (IHB) is a kind of specific solute-solvent interaction, which plays a key role in molecular or supramolecular photochemistry, as well as in photobiology. Solvation and its influence on various physico-chemical and biological processes can be understood by i) top-down; and ii) bottom-up approaches. In the top-down approach, the macroscopic properties like dielectric constant, refractive index are used to understand the microscopic solvation. This approach fails when specific interactions like hydrogen bonding interactions come into play, and furthermore it can reproduce only the macroscopic polarization of the solvent but fails miserably at the cybotactic region of solvation. With the recent advancements in the computational field, the molecular level description of solvation has been within reach for chemical physicists and experimentalists to corroborate their experimental results and in turn to visualize processes of fundamental or technologically relevant problems. The energy levels of the nπ* and ππ* singlet and triplet excited states of aromatic ketones are close-lying and therefore their energy levels can be altered by the substituents. The solvent polarity can be used as a surrogate to tune their energy levels. In certain cases, the lowest triplet or singlet excited states can switch their electronic character with increasing solvent polarity known as “electronic state switching” and thus modulate their photochemical or photophysical properties. Therefore, aromatic ketones were used as solvatochromic probes in this work. Comprehensive analyses of the solvent effects on xanthone (XT), 9,10-phenanthrenequinone (PQ) and benzophenone (Bzp) were carried out using steady-state and nanosecond time-resolved absorption, and resonance Raman spectroscopy in conjunction with ad hoc and classical-molecular dynamics and simulations generated supermolecule-continuum solvent model quantum mechanical calculations to corroborate the experimental outcomes and in turn to visualize the solvation process at the molecular level. The present thesis is divided into eight chapters and the summary of each chapter is described below: Chapter 1 provides a brief literature review of solvation effects and their influence on various physico-chemical and biological processes. Furthermore, the importance of understanding solvation at the molecular level and key concepts are discussed, which forms the heart of this thesis. Chapter 2 discusses the experimental and computational approaches used to study the solvation processes at the molecular level. A detailed explanation of spectroscopic techniques like resonance Raman (RR) and nanosecond-time resolved resonance Raman (ns-TR3) spectroscopy and their experimental and theoretical aspects are discussed, followed by a discussion on the fundamental concepts of computational methods like ab initio calculations density functional theory (DFT), and classical molecular dynamics and simulations (c-MDS) utilized in this study. Chapter 3 focuses on microscopic understanding of solvatochromic shifts observed for 9,10-phenanthrenequinone in protic solvents using UV-Vis and RR spectroscopy in conjunction with an ad hoc explicit solvation model and time-dependent density functional theory (TDDFT) calculations. The hypsochromic shift and bathochromic shift of the singlet nπ* and ππ* electronic transitions in protic solvents are due to hydrogen bond weakening and strengthening in the excited state for the corresponding electronic transitions, respectively as indicated by TD-DFT calculations and Kamlet-Taft linear solvation energy relationships. The hydrogen bond strengthening in the singlet ππ* excited state is further confirmed by Raman excitation profile (REP) analysis of PQ in different solvents. Furthermore, with increasing solvent polarity the two lowest singlet excited states undergo different hydrogen bonding mechanisms, leading to a decreasing energy gap between them. Therefore, hyperchromism of the nπ* transition has been hypothesized to be due to an increasing vibronic coupling between the lowest singlet nπ* and ππ* excited states. In Chapter 4, a real time observation of the thermal equilibrium between the lowest triplet excited states of PQ in acetonitrile solvent was carried out using ns-TR3 spectroscopy and this can explain its high reactivity towards H-atom abstraction, despite the fact that the lowest triplet excited state has ππ* character. Furthermore, extending the concept of hydrogen bonding mechanisms of the lowest singlet to the triplet excited states, the different hydrogen bonding mechanisms exhibited by them can lead to alteration of the intersystem crossing mechanisms in PQ. Chapter 5 highlights the very different role of intermolecular hydrogen bonding in the reduced reactivity of the xanthone (XT) triplet towards H-atom abstraction in protic solvents. The different hydrogen bonding mechanisms exhibited by the two lowest triplet excited states in protic solvents are derived from an ad hoc explicit solvation model, TD-DFT calculations and ns-time resolved absorption (ns-TRA): they separate them further in energy and thereby the nearest T2(nπ*) triplet state to the T1(ππ*) excited state plays an insignificant role in the reactivity towards H-atom abstraction, in contrast to the PQ triplet discussed in Chapter 4. Chapter 6 discusses the structure of XT triplet states using TR3 spectroscopy in combination with TD-DFT studies. The TR3 spectrum of the XT in acetonitrile identified a vibronic coupling mode responsible for the reactivity of XT towards H-atom abstraction, despite the fact that the lowest triplet excited state (T1) has ππ* character. This vibronic active mode is absent in the TR3 spectra of XT in protic solvents (methanol and ethanol). Furthermore, the REP analysis suggests that the nanosecond triplet-triplet absorption spectrum of XT in acetonitrile involves two different species, while in methanol it involves only one species. This observation is in agreement with the previous chapter (Chapter 5) which proposes a different hydrogen bonding mechanisms for the two lowest triplet excited states and their influence on the reduced reactivity towards H-atom abstraction. Chapters 3-6 emphasize the need for a proper solvation model at the molecular level to describe the various photophysical and photochemical processes of aromatic ketones. Therefore, Chapter 7 includes discussions on the bottom-up solvation methodology applied to benzophenone (Bzp) to understand its vibrational and electronic solvatochromic behaviour at the molecular level. Raman and UV-Vis spectroscopic techniques were used in conjunction with a c-MDS-generated supermolecule continuum solvation model DFT calculation to corroborate and to visualize the experimental outcome. The carbonyl stretching frequency of Bzp in protic solvents has two bands, corresponding to free carbonyl and hydrogen bonded carbonyl. Despite the fact that the macroscopic polarity of acetonitrile and methanol solvents are similar, the free carbonyl stretching of Bzp in methanol solvent was blue-shifted by 4 cm-1 with respect to the carbonyl stretching in acetonitrile solvent. The Gutmann’s acceptor number plot for carbonyl stretching frequencies indicates that the free carbonyl group is neighboured by a hydrophobic environment. The c-MDS-generated supermolecule-continuum solvation model DFT calculations suggest that the extended hydrogen bonding network of methanol solvent is responsible for the hydrophobic solvation around the free carbonyl. Furthermore, a linear correlation was obtained for the vibrational and electronic solvatochromism of the carbonyl frequency and energy of the singlet nπ* transition, respectively, which indicates that a variation in excitation wavelength for the singlet nπ* transition can arise from different solvation states. This can have implications for ultrafast processes associated with electron transfer, charge-transfer and also the photophysical aspects of excited states.Finally, Chapter 8 contains overall conclusions of the thesis and future directions for the present research area.

Noncovalent interactions are of great importance in studies on crystal design and drug discovery. One such noncovalent interaction, halogen bonding, is present between a covalently bound halogen atom and a Lewis base. A halogen bond is a directional interaction caused by the anisotropic distribution of charge on a halogen atom X covalently bound to A, which in turn forms a positive region called σ-hole on the A–X axis. Utilization of halogen bonds in lead optimization have been rarely considered in drug discovery until recently and yet more than 50% of the drug candidates are halogenated. To date, the halogen bond has not been subjected to practical molecular mechanical-molecular dynamics (MM-MD) study, where this noncovalent interaction cannot be described by conventional force fields because they do not account for the anisotropic distribution of the charge density on the halogen atoms. This problem was solved by the author and, for the first time, an extra-point of positive charge was used to represent the σ-hole on the halogen atom. This approach is called positive extra-point (PEP) approach. Interestingly, it was found that the performance of the PEP approach in describing halogen bond was better than the semiempirical methods including the recent halogen-bond corrected PM6 (PM6-DH2X) method. The PEP approach also gave promising results in describing other noncovalent halogen interactions, such as C–X···H and C–X···π-systems. The PEP resulted in an improvement in the accuracy of the electrostatic-potential derived charges of halogen-containing molecules, giving in turn better dipole moments and solvation free energies compared to high-level quantum mechanical and experimental data.With the aid of our PEP approach, the first MM-molecular dynamics (MM-MD) study of inhibitors that form a halogen bond with a receptor was performed for tetrahalobenzotriazole inhibitors complexed to cyclin-dependent protein kinase (CDK2). When the PEP approach was used, the calculated MM-generalized Born surface area (MM-GBSA)//MM-MD binding energies for halobenzimidazole and halobenzotriazole inhibitors complexed with protein kinase CK2 were found to correlate well with the corresponding experimental data, with correlation coefficients R2 of greater than 0.90. The nature and strength of halogen bonding in halo molecule···Lewis base complexes were studied in terms of molecular mechanics using our PEP approach. The contributions of the σ-hole (i.e., positively charged extra-point) and the halogen atom to the strength of this noncovalent interaction were clarified using the atomic parameter contribution to the molecular interaction approach. The molecular mechanical results revealed that the halogen bond is electrostatic and van der Waals in nature. The strength of the halogen bond increases with increasing the magnitude of the extra-point charge. The van der Waals interaction’s contribution to the halogen bond strength is most favorable in chloro complexes, whereas the electrostatic interaction is dominant in iodo complexes.The failure of the PM6 semiempirical method in describing noncovalent halogen interactions —not only halogen bonds, but also hydrogen bonds involving halogen atoms— was reported and corrected by the introduction of a second and third generation of noncovalent halogen interactions correction. The developed correction yielded promising results for the four examined noncovalent halogen interactions, namely: C–X···O, C–X···N, C–X···π-system, and C–X···H interactions.

Cette thèse de doctorat a été réalisée en cotutelle entre les Universités de Rennes 1 en France et de Florence en Italie. L’objectif de ce travail est tout d’abord de rationaliser les propriétés inter- et intramoléculaires de molécules-aimants (Single Molecule Magnet – SMM) à base d’ions lanthanides (“partie moléculaire”) et puis leur évolution une fois absorbé sur surface (''partie sur surface''). Ces deux aspects ont été examinés dans un cadre théorique et computationnel, en utilisant différentes techniques multi-niveaux, de periodic Density Functional Theory (pDFT) en utilisant une approche post-Hartree-Fock, en fonction de la variable expérimentale d’intérêt. Les molécules-aimants sont d'un intérêt particulier pour le design de nouveaux matériaux magnétiques dans la science des surfaces (comme la spintronique), mais elles permettent également une connaissance des propriétés électroniques et magnétiques approfondie est également nécessaire. / The Ph.D. project was a joint agreement between two universities: Université de Rennes 1 in France and Università di Firenze in Italy. The project aimed to shed light on the rationalization of the inter- and intramolecular properties of novel lanthanide-based Single Molecule Magnets, SMMs, (“molecular part”) and their evolution once adsorbed on surface (“surface part”). Both aspects are examined within a theoretical and computational framework, with different multi-level techniques ranging from periodic Density Functional Theory (pDFT) to post-Hartree-Focks approaches, depending on the experimental observable of interest. SMMs are, indeed, at the cutting-edge in the design of novel magnetic materials in surface science (as spintronics or memory storage devices), but for their exploitation a deep understanding of their electronic and magnetic properties is needed.

abstract: The structural and electronic properties of compositionally complex semiconductors have long been of both theoretical interest and engineering importance. As a new class of materials with an intrinsic compositional complexity, medium entropy alloys (MEAs) are immensely studied mainly for their excellent mechanical properties. The electronic properties of MEAs, however, are less well investigated. In this thesis, various properties such as electronic, spin, and thermal properties of two three-dimensional (3D) and two two-dimensional (2D) compositionally complex semiconductors are demonstrated to have promising various applications in photovoltaic, thermoelectric, and spin quantum bits (qubits).3D semiconducting Si-Ge-Sn and C3BN alloys is firstly introduced. Density functional theory (DFT) calculations and Monte Carlo simulations show that the Si1/3Ge1/3Sn1/3 MEA exhibits a large local distortion effect yet no chemical short-range order. Single vacancies in this MEA can be stabilized by bond reformations while the alloy retains semiconducting. DFT and molecular dynamics calculations predict that increasing the compositional disorder in SiyGeySnx MEAs enhances their electrical conductivity while weakens the thermal conductivity at room temperature, making the SiyGeySnx MEAs promising functional materials for thermoelectric devices. Furthermore, the nitrogen-vacancy (NV) center analog in C3BN (NV-C3BN) is studied to explore its applications in quantum computers. This analog possesses similar properties to the NV center in diamond such as a highly localized spin density and strong hyperfine interactions, making C3BN suitable for hosting spin qubits. The analog also displays two zero-phonon-line energies corresponding to wavelengths close to the ideal telecommunication band width, useful for quantum communications. 2D semiconducting transition metal chalcogenides (TMCs) and PtPN are also investigated. The quaternary compositionally complex TMCs show tunable properties such as in-plane lattice constants, band gaps, and band alignment, using a high through-put workflow from DFT calculations in conjunction with the virtual crystal approximation. A novel 2D semiconductor PtPN of direct bandgap is also predicted, based on pentagonal tessellation. The work in the thesis offers guidance to the experimental realization of these novel semiconductors, which serve as valuable prototypes of other compositionally complex systems from other elements. / Dissertation/Thesis / Doctoral Dissertation Materials Science and Engineering 2020

This work discusses applications of computational simulations to enzymatic systems with a particular focus on the effects of various small perturbations on cancer and disease-related systems. First, we cover the development of carbohydrate-based PET imaging ligands for Galectin-3, which is a protein overexpressed in pancreatic cancer tumors. We uncover several structural features for the ligands that can be used to improve their binding and efficacy. Second, we discuss the AlkB family of enzymes. AlkB is the E. coli DNA repair protein for alkylation damage, and has human homologues with slightly different functions and substrates. Each has a conserved active site with a catalytic iron and a coordinating His...His...Asp triad. We have applied molecular dynamics (MD) to investigate the effect of a novel single nucleotide polymorphism for AlkBH7, which is correlated with prostate cancer and has an unknown function. We show that the mutation leads to active site distortion, which has been confirmed by experiments. Thirdly, we investigate the unfolding of hen egg white lysozyme in 90% ethanol solution and low pH, to show the initial steps of unfolding from a native-like state to the disease-associated beta-sheet structure. We compare to mass spectrometry experiments and also show differing pathways based on protonation state. Finally, we discuss three different DNA polymerase systems. DNA polymerases are the primary proteins that replicate DNA during cell division, and have various extra or specific functions. We look at a proofreading-deficient DNA polymerase III mutant, the effects of solvent on DNA polymerase IV's ability to bypass bulky DNA adducts, and a variety of mutations on DNA polymerase kappa.

Chemical science spans multiple scales, from a single proton to the collection of proteins that make up a proteome. Throughout my graduate research career, I have developed statistical and machine learning models to better understand chemistry at these different scales, including predicting molecular properties of molecules in analytical and synthetic chemistry to integrating experiments with chemo-proteomic based machine models for drug design. Starting with the proteome, I will discuss repurposing compounds for mental health indications and visualizing the relationships between these disorders. Moving to the cellular level, I will introduce the use of the negative binomial distribution to find biomarkers collected using MS/MS and machine learning models (ML) used to select potent, non-toxic, small molecules for the treatment of castration--resistant prostate cancer (CRPC). For the protein scale, I will introduce CANDOCK, a docking method to rapidly and accurately dock small molecules, an algorithm which was used to create the ML model for CRPC. Next, I will showcase a deep learning model to determine small-molecule functional groups using FTIR and MS spectra. This will be followed by a similar approach used to identify if a small molecule will undergo a diagnostic reaction using mass spectrometry using a chemically interpretable graph-based machine learning method. Finally, I will examine chemistry at the proton level and how quantum mechanics combined with machine learning can be used to understand chemical reactions. I believe that chemical machine learning models have the potential to accelerate several aspects of drug discovery including discovery, process, and analytical chemistry.

Mass spectrometry with electrospray ionization is an excellent method for structural analysis of coordination compounds with outstanding sensitivity and selectivity. However, it fails to detect some low-polar rhenium complexes. This master thesis describes derivatization method of non-ionizable rhenium complexes with 1,2-dihydroxybenzene and 2,3- dihydroxytoluenene. Fragmentation mechanisms and structure of prepared complexes was studied using high resolution mass spectrometry and collision-induced dissociation (CID). Furthermore, density functional theory (DFT) computational method was used for prediction of bond cleavage based on bond lengthening.

<div>Deep neural networks (DNNs) have gained tremendous attention over the recent years due to their outstanding performance in solving many problems in different fields of science and technology. Currently, this field is of interest to many researchers and growing rapidly. The ability of DNNs to learn new concepts with minimal instructions facilitates applying current DNN-based methods to new problems. Here in this dissertation, three methods based on DNNs are discussed, tackling different problems in the field of computer-aided drug design.</div><div><br></div><div>The first method described addresses the problem of prediction of hydration properties from 3D structures of proteins without requiring molecular dynamics simulations. Water plays a major role in protein-ligand interactions and identifying (de)solvation contributions of water molecules can assist drug design. Two different model architectures are presented for the prediction the hydration information of proteins. The performance of the methods are compared with other conventional methods and experimental data. In addition, their applications in ligand optimization and pose prediction is shown.</div><div><br></div><div>The design of de novo molecules has always been of interest in the field of drug discovery. The second method describes a generative model that learns to derive features from protein sequences to design de novo compounds. We show how the model can be used to generate molecules similar to the known for the targets the model have not seen before and compare with benchmark generative models.</div><div><br></div><div>Finally, it is demonstrated how DNNs can learn to predict secondary structure propensity values derived from NMR ensembles. Secondary structure propensities are important in identifying flexible regions in proteins. Protein flexibility has a major role in drug-protein binding, and identifying such regions can assist in development of methods for ligand binding prediction. The prediction performance of the method is shown for several proteins with two or more known secondary structure conformations.</div>

<div>Conventional soybean oil and high-oleic soybean oil offer opportunities as bio-solvents for sweetening sour natural gas, adding value to the soybean oil industry and the natural gas industry. The rise of fracking in the United States and changing economics in the energy industry have increased use of natural gas, which is often rendered sour by high concentrations of hydrogen sulfide (H2S), a toxic and corrosive impurity. The present work evaluates the viability of both conventional and high-oleic soybean oil to act as bio-solvents for removing gaseous H2S. Predictive in silico methods, experimental validation, and economic feasibility analysis are included to draw conclusions regarding the overall capability and feasibility of using soybean oils as bio-solvents for gas sweetening.</div><div><br></div><div>In silico predictive methods for sweetening were implemented to assess the relationship between fatty acid composition in the soybean oils and the ability to effectively partition H2S from methane or nitrogen gases. The Conductor-like Screening Model for Real Solvents (COSMO-RS) was used to predict the partition coefficient (K) of H2S in a bi-phasic liquid-vapor system made up of fatty acids in the liquid phase and methane or nitrogen gas in the vapor phase. The fatty acid mass fractions represented those found in soybean or high-oleic soybean oil. Methane represented gas and nitrogen was considered in order to compare to experimental conditions. This proof of concept work predicted K values for H2S below 0.0005 at temperatures from 10 to 100 °C at atmospheric pressure; K values near zero indicate near-complete removal of H2S from the gas phase.</div><div><br></div><div>Experimental validation included equilibrium extraction experiments as well as data collection for isotherm model development. Experimental equilibrium studies were carried out at residence times ranging from 0 – 60 minutes with mixing at ambient conditions. Experiments resulted in K values below 0.1 for H2S in soybean oil and high-oleic soybean oil at 25 °C with residence times less than 15 minutes and a 2:1 gas to oil ratio. More than 90% of the H2S was removed from the gas phase within 15 minutes. Isotherm models demonstrated the saturation limits of the soybean oils and compared them to saturation limits in water and heptane. </div><div><br></div><div>Economic feasibility experiments used graphical and algebraic methods to determine the number of equilibrium stages needed to remove 99.9% of H2S from feed gas with H2S concentrations ranging from 40 – 400 ppm. A gas flow rate equivalent to industrial levels was used to design an extraction column. Capital costs and operating costs were estimated, along with the revenues to be gained from selling methane and selling recovered elemental sulfur as a secondary product. Solvent regeneration would need to exceed 98% in order to keep the cost of treating a unit of natural gas equal to or less than existing industrial methods. Suggestions for cutting costs and improving process viability are made.</div><div><br></div>

The work described herein is related to gas-phase ion-molecule reactions studied by using mass spectrometry. Chapter 2 describes density functional theory, a method used in chapters 4 and 5 to propose reaction mechanisms for reactions previously observed by others by using mass spectrometry. Chapter 3 describes a study that demonstrates that the fragmentation of saturated hydrocarbons occurs due to proton-transfer reactions that occur between these species and protonated molecules generated from molecules present in air such as nitrogen and water. Saturated hydrocarbons are studied in a wide variety of fields, and better methods to analyze complex mixtures of these compounds would facilitate their analysis. Chapter 4 discusses mechanisms of reactions for previously studied ion-molecule reactions of pyridynes studied by others by using mass spectrometry. Reactions of pyridynes are important to study arynes have been previously used in organic synthesis, and pyridine moieties are particularly common in biological compounds. Chapter 5 discusses density functional theory calculations used to determine why some organic polyradical undergo hydride abstractions from cyclohexane while others do not. The study discusses reactions taking place between both singlet and triplets states of the 2,5-didehydropyridinium cation and cyclohexane as a model, which are compared to reactions of the 2-pyridyl cation and 2-dehydropyridinium cation with cyclohexane. These studies may help improve our understanding of the reactivity-controlling factors of organic polyradicals, which may help improve toxic drug candidates like cytostatic enediynes.