The fast progress in renewable energy sources, emerging electronics, and electrical automobiles stimulate the development of next-generation energy storage devices. Among them, supercapacitors, possessing high power and high energy density, caught rabid attention worldwide. Theoretical studies on the electrode compositions, geometries, and interfaces delivered useful information on various material properties, which was essential for approaching high-performance supercapacitors. Recent computational screening techniques suggest potential electrode materials for future experimental designs and device fabrications. So far, transition metal oxides (TMO), carbon-based materials, and the newly discovered 2D MXenes, etc. were regarded as the promising supercapacitor electrodes due to their high theoretical capacitance, larger surface areas, and multiple constituents. In the first part of this dissertation, density functional theory (DFT) was employed to analyze the structure, energetics, and transport of hydrogen in 3d/4d transition metal perovskites (ABO3). The B-O bonding contributions before and after H absorption were unveiled. While simple chemical descriptors, allowing comprehensive searches for candidate proton-conductor perovskites with little computational cost, were provided. Likewise, the interfacial properties for the heterostructures MXene (Ti3C2Tx and Ti2CTx with T=-O, -OH, a mixture of -F and -O)/(pure or N-doped) anatase-TiO2 (101) were evaluated via DFT. We found that surface functional groups of MXene dramatically impacted the interface electronic charge transfers and interfacial configurations. In addition, the N-doped surface-O of TiO2 changed the electronic and geometric properties of these hybrid composites. Besides basic understandings of various material properties, perceiving energy storage mechanisms behind different types of supercapacitors is also a welcoming topic. In the second part of this dissertation, ab-initio molecular dynamics (AIMD) were utilized to study the proton transport performances at the mono water layer confined by Graphene-Ti3C2O2 (dissimilar interface) and Ti3C2O2-Ti3C2O2 (similar interface) with different intercalated-proton concentrations. The results showed that interfacial properties, as well as proton diffusion behavior, played a significant role in the faster proton surface-redox and transport process. Ultimately, this dissertation explores the fundamental knowledge of the recent sprout popular metal-oxides (perovskites), MXene-based composite electrode materials for supercapacitor applications. Meanwhile, we offered a theoretical interpretation of energy storage mechanisms for proton intercalated MXene-based layered interfaces.

Supercapacitors, such as electric double-layer capacitors (EDLCs) and pseudocapacitors, are becoming increasingly important in the field of electrical energy storage. Theoretical study of energy storage in EDLCs focuses on solving for the electric double-layer structure in different electrode geometries and electrolyte components, which can be achieved by molecular simulations such as classical molecular dynamics (MD), classical density functional theory (classical DFT), and Monte-Carlo (MC) methods. In recent years, combining first-principles and classical simulation to investigate the carbon-based EDLCs has become increasingly popular due to the importance of quantum capacitance in graphene-like 2D systems. More recently, the development of joint density functional theory (JDFT) enables self-consistent electronic-structure calculation for an electrode being solvated by an electrolyte. In contrast with the large amount of theoretical and computational effort on EDLCs, theoretical understanding of pseudocapacitance is very limited. In this dissertation project, we first study the carbon-based EDLCs and focus on several important aspects of EDLCs including quantum capacitance, nitrogen doping, dielectric screening, and edge effect. Then, based on our physical understanding on EDLCs, we designed novel 2D boron supercapacitors, which exhibit promising capacitive performance comparing with conventional carbon electrodes. Finally, we also introduced how to study pseudocapactive mechanism through first principle calculation with solvation model technique. Two typical pseudocapacitor systems are investigated: RuO2(110) and Ti3C2Tx (T=O,OH) in H2SO4 electrolyte. Based on our study, the relation between EDL and redox behavior were revealed. We summarize and conclude with an outlook for the future of materials design and discovery for capacitive energy storage.

Hydrogen-catalyst interaction is the foundation of many technologies and processes. Herein we employ first principles density functional theory (DFT) techniques to investigate how hydrogen interacts with different catalytic systems, including cobalt phosphide (CoP), h-BN monolayer supported on transition metals (h-BN/metal), atomically precise gold nanoclusters, and titanium dioxide (TiO2). Our study has important implications for the use of the catalysts for hydrogen evolution reaction (HER), CO2 reduction, and H2 activation.

Electrochemical HER using renewable energy can provide a sustainable supply of fuel for future societies with hydrogen as a key energy carrier. CoP is one of the most active earth-abundant electrocatalysts for HER. We have studied the adsorption structures and energetics of atomic hydrogen on several low-Miller-index surfaces of CoP. We find that CoP(111) is the most promising facet for high and long-lived HER activity. Compared to HER, CO2 reduction is much harder. Since HER competes with CO2 reduction at all negative potentials, a key criterion for selective catalysts in CO2 reduction is the avoidance of HER. We have investigated the electrocatalytic activities of h-BN/Ni, h-BN/Co, and h-BN/Cu. We find that the competitive HER channel can be filtered out by the surface h-BN monolayer for h-BN/Ni and h-BN/Co, making them very good electrocatalysts for CO2 reduction.

Atomically precise gold nanoclusters are a new class of catalysts, and it is able to establish structure-activity correlations due to their well-defined structures. Thiolate-protected gold nanoclusters have received considerable research attention, and have been reported to exhibit high HER activity. We have studied the thiolate-gold interface and find that the interfacial structures of thiolates on gold are surface sensitive. To understand HER on the thiolate-protected gold nanoclusters, we have studied how hydrogen interacts with [Au25(SR)18]q clusters (q=-1, 0, +1) and monoatom-doped bimetallic [M1Au24(SR)18]q clusters (M=Pt, Pd, Ag, Cu, Hg, Cd, and q=-2, -1, 0). We conclude that PtAu24(SR)18, PdAu24(SR)18, and CuAu24(SR)18 clusters can be very good electrocatalysts for HER. Unlike the above thiolate-protected gold nanoclusters, where the surface gold atoms are fully coordinated by the thiolates, the Au22(L8)6 cluster (L = 1,8-bis(diphenylphosphino) octane) is unique in that it contains eight coordinatively unsaturated (cus) gold atoms. We find that the eight cus gold atoms in the Au22(L8)6 cluster can adsorb hydrogen stronger than Pt, thereby being a potentially promising catalyst for HER.

Hydrogen interaction with metal oxides is of growing interest recently. TiO2 is one of the most extensively investigated metal oxides. We have studied the mechanisms of H2 activation on the surfaces of different TiO2 polymorphs, including rutile TiO2(110), anatse TiO2(101), and brookite TiO2(210). We find that for all the three surfaces, although the homolytic dissociation of H2 is thermodynamically more favorable, the heterolytic pathway is kinetically more favorable.

The selective conversion of alkanes to high value products remains a monumental catalytic challenge. Significant gaps in our knowledge remain regarding the mechanistic details of alkane activation and the chemical nature of the catalytically active sites for this class of reactions. In this thesis, systematic and thorough first principles studies of alkane activation on several major classes of promising materials, ranging from metal oxide surfaces to metal clusters and single metal atom sites, have provided us with the insight to tackle these open questions. To further expand these findings towards a comprehensive search of catalytically active materials, we developed chemical and catalytic descriptors which can predict these properties at a fraction of the computational cost. We find on oxide surfaces, the homolytic C-H activation pathway is favored, and correlated with the reducibility of the lattice oxygen. We identify hydrogen adsorption energy and vacancy formation as strongly correlated energetic descriptors for C-H activation energy and CH3 adsorption energy, which we use to predict catalytic properties of doped Co3O4 and compositionally diverse perovskite surfaces. To provide a structure-activity link between the oxygen and alkane activation, we developed a generally applicable coordination number descriptor, ‘ACN,’ which can quickly predict C-H activation barriers on various oxides. Furthermore, we identify a bulk descriptor, the metal-oxygen crystal orbital Hamilton population, which is correlated to the adsorption energy of radical adsorbates such as hydrogen and methyl. Meanwhile on metal clusters, structural fluxionality plays a pivotal role in catalytic performance. In the model system of Pt10-13, we utilize various structural features to identify Pt10 as a magic number cluster with low fluxionality which can explain its experimental inertness to alkanes. On single atoms, we demonstrate how heterolytic C-H cleavage pathway can be tuned through the metal d-orbital occupation and the substrate electronic band gap. These findings have allowed us to identify promising single atom catalysts with the elusive low-temperature methane activation ability. The work in this thesis aims to provide valuable descriptors and design principles for alkane activation and theoretical predictions of promising catalysts.

Membrane-based materials are an important branch in the field of gas separation. There are increasing number of membrane materials which are synthesized in recent years. Among them, the novel ultrathin membranes with uniform pores like nanoporous graphene attracted attention of many researchers, because their excellent gas separation ability based on molecular sieving effect. Several factors might influence the performance of ultrathin membranes, including pore size, pore shape, pore density, etc. However, the theoretical understanding of these factors has not been clearly addressed. In this dissertation, we firstly investigated the optimal pore size by first-principles density functional theory and simulated the pore-size effect for post-combustion carbon capture by grand canonical Monte Carlo. Next, we used molecular dynamics simulations to study the effect of pore density on the nanoporous graphene membrane. Then, we proposed a bilayer design of nanoporous graphene membrane with continuously tunable effective pore size for gas separations, such as CO2/CH4, N2/CH4, O2/N2. Meanwhile, we studied the effect of entropic selectivity by tracking the trajectories of gas-permeance events. After that, we proposed other design of graphene/ionic-liquid composites with tunable slit pore size. Finally, to satisfy the requirement of exploring huge material database for gas separation, we applied machine learning to predict the selectivity of porous carbon materials. We further used convolutional neural networks to search the optimal porous carbon material from an ultimate input feature. The same approach was also used for prediction of hydrogen locations in copper clusters, which are potential materials for hydrogen storage. The works in this dissertation aims to find and design ultrathin nanoporous materials for gas separation by different computational approaches.

Uranium’s economical consumption is based on the industrial nuclear fuel cycle. Today there remains 4-6 million tons of terrestrial uranium in Earth’s ores. However, as the projected population is estimated to double by 2040, current uranium reserves are only estimated to provide nuclear fuel for the next century. Unconventional methods for acquiring uranium are, therefore, required to maintain longevity and sustainability for future economic nuclear energy demands.

One unconventional alternative for mining uranium is diving from land and into the open seas. Although uranium resides uniformly in the oceans at very low concentrations of 3.3 ppb, in a vast body of Earths water, there exists an untapped reserve of uranium at an astounding 4.5 billion ton. Uranium exists in aqueous environment as the stable oxo-cation, in the hexavalent (VI) state called uranyl, UO22+, and as the neutral ternary-uranyl-carbonato, Ca2UO2(CO3)3, in seawater.

After six generations of research and development, current extraction methods rely on a polymer grafter amidoxime-based (poly(AO)) material for uranium recovery from seawater. One chemical advancement in adsorption capacity is from the inclusion the hydrophilic group carboxylate (Ac). However, displacing coordinating carbonates with poly(amidoxime) from uranyl, which is both at low concentrations and existing in soup of various metal-ions in a seawater matrix, still remains a formidable challenge. One of the multifaceted efforts involved with increasing adsorption selectivity is by the theoretical insight into uranium’s coordination chemistry with interacting ligands in seawater-related environments.

Herein we utilize current state-of-the-art, advanced computing technology to investigate coordinating uranium species in a completely explicit model. First, we employ a pure quantum mechanical (QM) simulations to pry into the chemistry of surrounding water molecules’ influence on the dominant Ca2[UO2(CO3)3] specie. Furthermore, large scale classical molecular dynamics (CMD) simulations are assisted with umbrella sampling to develop computational techniques for investigating coordinating uranyl-AO from calculating binding free-energies (Gbind). Furthermore, we investigate the influence of capacity by inspecting the interactions of mix-ligand containing AO/Ac. Additionally, employing umbrella sampling method coupled with WHAM, we investigate the sodium ion solvation environment of one of the most competing vanadium-containing species, HVO42-.

Constituting the bulk of rare-earth elements, lanthanides (Ln) need to be separated to fully realize their potential as critical materials in many important technologies. Solvent extraction is the most extensively used process to separate lanthanides on an industrial scale. Recently, preorganized ligands such as bis-lactam-1,10-phenanthroline show unique selectivity trends across the lanthanide series, indicating the synergistic effects of both N and O donors in complexing with lanthanides. So, we examined mixed N, O-donor ligands containing pyridinic N and N-oxide groups and evaluated their relative aqueous La(III)/Ln(III) selectivity by computing free energy changes for the exchange reaction between the designed ligands and a reference ligand via density functional theory (DFT). Three novel ligands show promise as excellent extractant agents in selectively separating trivalent lanthanides. However, no matter our DFT investigations or other discovery of new ligands in rare-earth separations by solvent extraction, they are still largely based on trial and error, a low-throughput and inefficient approach. As a result, deep neural networks on the available experimental data of distribution coefficients measured for hundreds of ligands for 14 Ln(III) ions have been trained to accurately and quickly predict their distribution coefficients for a given ligand and the extraction conditions. Four new-synthesized ligands were found that their predicted distribution coefficients from our trained machine-learning model match well with their experimental values. Then, this trained model was applied to some large ligand databases automatically generated by molecule generation tools based on string-based representations. Several hexalkyl-nitrilotriacetamide ligands were screened out with high potential in selective rare-earth separations. Therefore, our machine-learning approach paves the way for accelerating the discovery of new ligands for rare-earth separations. In additional to rare-earth elements, some theoretical insights towards atomically precise ligand-protected nanoclusters by first principles were performed. All work in this dissertation aim at chemically understanding the interactions between metals and organic ligands by different computational approaches.

The development of Single Atom Catalysts (SACs) is a milestone in the catalysis industry. SACs usually have unique electronic properties and ultrahigh reactivity for chemical reactions. The Turnover frequency (TOF) values of SACs are generally much larger than that of regular nanoparticle catalyst. However, due to the high surface energy of the SACs, the aggregation effect restrains the researchers from further increasing the loading of the single atoms over the supports. The aggregation, not like the coking which could be easily solved by reoxidation of the catalyst, is irreversible and would also lead to the deactivation of the catalyst. This conundrum prevents the implementation of SACs in industrial usage. To solve this problem, scientists usually leverage the Strong Metal-Support interaction (SIMI) effect to anchor the single atoms by the defect sites on the support (e.g., vacancies, hydroxyl group). In this dissertation, the first part of our work is to conduct the rational design of the SACs on the 2-dimensional (2D) hexagonal boron nitride (h-BN) materials. We assessed the possibility of noble metals Ru, Pt, and Au loaded on h-BN as SACs from three aspects (i.e., stability, activity, and selectivity). Both ab initio Molecular Dynamics (AIMD) simulations and anchoring energy calculations suggested that Au is not stable over the defect sites of h-BN. From the reaction energy plot of the propane dehydrogenation (PDH), we found that the Pt over B vacancy and Ru over N vacancy are promising for PDH reactions with low dehydration barriers. We also found that there is a trade of for the C-H bond activation and C3H6 desorption. This trade-off relationship let the researchers make a compromise between the activity and selectivity when conducting the catalyst design. The second part of this thesis focuses on the activity of the defect sites of the h-BN, we selected a set of vacancy sites with different sizes. From the defect formation energy calculations, we found that the normalized formation energy decreases with the increase of defect size. Also, we probed the CH4 activation over these vacancies and we found that the B-B pair over the N vacancy and B-N pair over the BN2 vacancy have excellent reactivity for CH4 activation. The third part of our work is to examine the activity and selectivity of CH4 oxidation to acetic acid reaction over single atom Rh dispersed Porphyrin-based Metal Organic Framework (pMOF) catalyst. We found that the Rh in-plane model favors the CH3OH formation while the Rh out-of-plane model favors the CH3COOH formation. This thesis gives the interpretation of the reaction mechanism of single atom catalysts over h-BN and pMOF catalysts.

The dream device for electrochemical energy storage (EES) should have both high energy density (as in batteries) and high power density (as in supercapacitors). Pseudocapacitors are a promising type of EES technologies toward that dream. The power density of a pseudocapacitor is mainly determined by the ion transport at the interface and inside the electrode. Therefore, atomistic understanding of the ion intercalation and transport is needed for rational design of high-performance pseudocapacitors. In this dissertation, we have investigated the ion transport, intercalation mechanism, interfacial structure, and spectroscopic features of two typical pseudocapacitive systems: MXenes (2D metal carbides and nitrides) and tungsten oxide hydrates. In the first part of the dissertation, we employed density functional theory (DFT) and first-principles molecular dynamics (FPMD) to study the ion insertion and transport in MXene electrodes. We first studied the proton dynamics in MXene-confined water to understand the relation between proton diffusion and surface redox behavior under confinement. Then, we studied the Li+ insertion into MXene in water-in-salts electrolyte. Our DFT calculation agrees well with the experimental electrochemical quartz crystal microbalance (EQCM) and in-situ X-ray diffraction (XRD) results, suggesting the desolvation-free Li+ (de)intercalation mechanism. Next, we investigated MXene edge structure by integrating the DFT calculation and structural search methods. The predicted most stable structure in vacuum exhibits an interlayer connection at the edge, while the configurational search of hydrated edge suggests the spontaneous CH4 formation, which was observed in previous experiments. In the second part of the dissertation, we studied the protonation sites and proton diffusion in tungsten oxides hydrates. We first investigated proton diffusion in tunnel h-WO3 and reported the optimal linear water density of 4 water/nm for highest proton conductivity. Next, we compared the calculated vibrational densities of states (VDOS) with inelastic neutron scattering (INS) measurement, to pinpoint proton at the bridging oxygen site in monoclinic WO3·H2O. Finally, we studied the proton adsorption in the H2W2O7 and revealed low proton adsorption energies at terminal oxygen sites. In summary, our simulation work provided significant mechanistic understanding on the pseudocapacitive behavior of MXenes and tungsten oxides hydrates from both thermodynamic and kinetic perspective, which will facilitate the development of high-performance pseudocapacitors for EES.

Carbon capture and gas separation is of great importance in addressing climate change, as greenhouse gas emissions continue to rise with major contributions from transportation and electricity production. Some materials have shown great affinity for CO2 capture, like metal-organic frameworks, zeolites, and functionalized ionic liquids. Limitations in these materials, such as regeneration costs, make it advantageous to develop porous media that are liquids at room temperature. This work employs computational chemistry to study two liquids used for gas separations: an ionic liquid with a polyethylene glycol-linked dication, and a novel porous ionic liquid with a permanent pore large enough to house small gas molecules. Using classical molecular dynamics, we are able to understand the behavior of these liquids on the atomic level, and design new materials to effectively tune their properties. The dicationic ionic liquid, [DBU-PEG][Tf2N]2, was synthesized previously, though not studied computationally. We find that our model well-replicated experimentally determined density, viscosity, and powder x-ray diffraction spectra. Our work reveals that the geminal dication, connecting two bicyclic amidine groups with a polyethylene glycol linker, plays a unique role in CO2 absorption, resulting from both the dication’s shape and charge distribution. We also present an anionic covalent cage-based porous liquid—the cage featuring a 4.25 Å pore. We simulate the bulk liquid with CO2, CH4, and N2 molecules, and find a strong preference for CO2 capture from the gas-cage interactions. This is supported by analysis of a gas molecule’s time-of-stay inside a cage, diffusivity of the gases in the system, and predicted solubilities from grand canonical Monte Carlo methods. The liquid also presents energetic and thermodynamic favorability for CO2 over the other gases studied. This work investigates two unique liquids’ performances in capturing carbon, and encourages further development of tunable liquids for gas separations. The geminal design of the cation in [DBU-PEG][Tf2N]2 inspires a similar approach of using dianions in ionic liquids, while the modular synthesis of the porous liquid can be easily modified to feature other functional groups. Such changes may lead to considerable effects in gas uptake and selectivity, and should be studied further.