<p> We use non-adiabatic <i>ab initio</i> molecular dynamics to study the influence of substituent side groups on the photoactive unit (Z)-hexa-1,3,5-triene (HT). The Time-Dependent Density Functional Theory Surface Hopping method (TDDFT-SH) is used to investigate the influence of substituted isopropyl and methyl groups on the excited state dynamics. The 1,4 and 2,5-substituted molecules are simulated: 2,5-dimethylhexa-1,3,5-triene (DMHT), 2-isopropyl-5-methyl-1,3,5-hexatriene (2,5-IMHT), 3,7-dimethylocta-1,3,5-triene (1,4-IMHT), and 2,5-diisopropyl-1,3,5-hexatriene (DIHT). We find that HT and 1,4-IMHT have the lowest ring-closing branching ratios of 5.3% and 1.0%, respectively. For the 2,5-substituted derivatives, the branching ratio increases with increasing size of the substituents, exhibiting yields of 9.78%, 19%, and 24% for DMHT, 2,5-IMHT, and DIHT, respectively. The reaction channels are shown to prefer certain conformation configurations at excitation, where the ring-closing reaction tends to originate from the gauche-Z-gauche (gZg) rotamer almost exclusively. In addition, there is a conformational dependency on absorption, gZg conformers have on average lower S₁ &larr; S₀ excitation energies that the other rotamers. Furthermore, we develop a method to calculate a predicted quantum yield that is in agreement with the wavelength-dependence observed in experiment for DMHT. In addition, the quantum yield method also predicts DIHT to have the highest CHD yield of 0.176 at 254 nm and 0.390 at 290 nm. </p><p> Additionally, we study the vitamin D derivative Tachysterol (Tachy) which exhibits similar photochemical properties as HT and its derivatives. We find the reaction channels of Tachy also have a conformation dependency, where the reactive products toxisterol-D1 (2.3%), previtamin D (1.4%) and cyclobutene toxisterol (0.7%) prefer cEc, cEt, and tEc configurations at excitation, leaving the tEt completely non-reactive. The rotamers similarly have a dependence on absorption as well, where the cEc configuration has the lowest energy S₁ &larr; S₀ excitation of the rotamers. The wavelength dependence of the rotamers should lead to selective properties of these molecules at excitation. An excitation to the red-shifted side of the maximum absorption peak will on average lead to excitations of the gZg rotamers more exclusively. </p><p>

<p> The search for robust, low-noise modelocked comb sources has attracted significant attention during the last two decades. Passively modelocked fiber lasers are among the most attractive comb sources. The most important design problems for a passively modelocked laser include: (1) finding a region in the laser&rsquo;s adjustable parameter space where it operates stably, (2) optimizing the pulse profile within that region, and (3) lowering the noise level. Adjustable parameters will typically include the cavity length, the pump power, and the amplifier gain, which may be a function of the pump power, the pump wavelength, and both the material and geometry of the gain medium. </p><p> There are two basic computational approaches for modeling passively modelocked laser systems: the evolutionary approach and the dynamical approach. In the evolutionary approach, which replicates the physical behavior of the laser, one launches light into the simulated laser and follows it for many round trips in the laser. If one obtains a stationary or periodically-stationary modelocked pulse, the laser is deemed stable and, if no such pulse is found, the laser is deemed unstable. The effect of noise can be studied by using a random number generator to add computational noise. In the dynamical approach, one first obtains a single modelocked pulse solution either analytically or by using the evolutionary approach. Next, one finds the pulse parameters as the laser parameters vary by solving a root-finding algorithm. One then linearizes the evolution equations about the steady-state solution and determines the eigenvalues of the linearized equation, which we refer to as the equation&rsquo;s dynamical spectrum. If any eigenvalue has a positive real part, then the modelocked pulse is unstable. The effect of noise can be determined by calculating the noise that enters each of the modes in the dynamical spectrum, whose amplitudes are described by either a Langevin process or a random walk process. </p><p> The evolutionary approach is intuitive and straightforward to program, and it is widely used. However, it is computationally time-consuming to determine the stable operating regions and can give ambiguous results near a stability boundary. When evaluating the noise levels, Monte Carlo simulations, which are based upon the evolutionary approach, are often prohibitively expensive computationally. By comparison, the dynamical approach is more difficult to program, but it is computationally rapid, yields unambiguous results for the stability, and avoids computationally expensive Monte Carlo simulations. The two approaches are complementary to each other. However, the dynamical approach can be a powerful tool for system design and optimization and has historically been undertilized. </p><p> In this dissertation, we discuss the dynamical approach that we have developed for design and optimization of passively modelocked laser systems. This approach provides deep insights into the instability mechanisms of the laser that impact or limit modelocking, and makes it possible to rapidly and unambiguously map out the regions of stable operation in a large parameter space. For a given system setup, we can calculate the noise level in the laser cavity within minutes on a desktop computer. </p><p> Compared to Monte Carlo simulations, we will show that the dynamical approach improves the computational efficiency by more than three orders of magnitude. We will apply the dynamical approach to a laser with a fast saturable absorber and to a laser with a slow saturable absorber. We apply our model of a laser with a slow saturable absorber to a fiber comb laser with a semiconductor absorbing mirror (SESAM) that was developed at National Institute of Standards and Technology (NIST), Boulder, CO. We optimize its parameters and show that it is possible to increase its output power and bandwidth while lowering the pump power that is needed.</p><p>

<p> In this study, several efficient and accurate mathematical models and numerical solutions to unconventional reservoir development problems are developed. The first is the three-dimensional embedded discrete fracture method (3D-EDFM), which is able to simulate fluid flow with multiple 3D hydraulic fractures with arbitrary strike and dip angles, shapes, curvatures, conductivities and connections. The second is a multi-porosity and multi-physics fluid flow model, which can capture gas flow behaviors in shales, which is complicated by highly heterogeneous and hierarchical rock structures (ranging from organic nanopores, inorganic nanopores, less permeable micro-fractures, more permeable macro-fractures to hydraulic fractures). The third is an iterative numerical approach combining the extended finite element method (X-FEM) and the embedded discrete fracture method (EDFM), which is developed for simulating the fluid-driven fracture propagation process in porous media. </p><p> Physical explanations and mathematical equations behind these mathematical models and numerical approaches are described in detail. Their advantages over alternative numerical methods are discussed. These numerical methods are incorporated into an in-house program. A series of synthetic but realistic cases are simulated. Simulated results reveal physical understandings qualitatively and match with available analytical solutions quantitatively. These novel mathematical models and computational solutions provide numerical approaches to understand complicated physical phenomena in developing unconventional reservoirs, thus they help in the better management of unconventional reservoirs. </p><p>

In recent years, continuum models of solvation have had exceptional success in materials simulations as well as condensed matter physics. They can easily capture the effects of disordered systems, such as neutral liquids or electrolytes solutions, on material interfaces without the need for expensive statistical sampling. The Environ library (www.quantum-environ.org) implements different continuum models and correction schemes, which is the focus of this presentation. Recently refactored into a stand-alone library, many changes have been introduced in Environ, making it more flexible and computationally efficient. Introduction of a double-cell formalism allows for faster ab initio DFT calculations while reparameterization of soft-sphere continuum model allows for smaller density cutoffs. Furthermore, Environ's periodic boundary conditions correction schemes have been expanded by including the AFC90 library, which allows for faster DFT calculations of partially periodic systems, such as slabs, wires, and isolated molecules. Finally, stand-alone Environ can now provide atomic and molecular descriptors, which can be used to characterize solvated interfaces, e.g. in machine learning applications. The specific details of the implementations are reviewed as well as their efficiency and some choice applications for different calculation setups and systems.

This thesis describes the computational tools that allow the simulation of polymers made up of an arbitrary number of three-level systems, the study of such systems and comparisons to experimental nonlinear optical spectra. The three-level system generator (3LSG), is designed to automatically generate the operators that describe the system, whether it is a closed system or an open quantum system (OQS) in contact with a bath, with just a few input parameters. A user is free to specify each 3LS's energy levels and transition dipoles between said levels, site couplings between the different units of the polymer and in the case of open systems, the rates and couplings describing the different relaxation processes taking place in an OQS, using the Redfield formalism. In either cases, the 3LSG is then capable of generating the Hamiltonian 𝐻₀ describing the closed system or the Liouvillian 𝓛₀ describing the open system from the various inputs. The Ultrafast Spectroscopy Suite (UFSS) is an open-source software suite used to perform the nonlinear optical spectroscopies simulations. It contains 4 main modules, one of which is the Hamiltonian/Liouvillian Generator (HLG), a module previously designed to model simpler two-level systems. The 3LSG is a sub-module of the HLG. The three-level system generator is used to replicate a theoretical model describing a copolymer model made of many identical pairs of squaraine monomers, where each monomer is a three-level system interacting with its neighbouring sites and a surrounding bath. The system automatically generated by the 3LSG is used, along with other spectroscopic calculation tools, to simulate high-order transient absorption (TA) spectroscopies and study the long-time behaviour of the 3rd-order to 13th-order excited state absorption (ESA) peaks in the TA signals. The 3LSG is used in conjunction with spectroscopic calculations tools as it was originally intented, though it may also be used by itself to study Hamiltonians and Liouvillians of electronic three-level systems.

Kohn-Sham Density Functional Theory (DFT) has been the most successful tool to probe the electronic structure, mainly the ground-state total energies and densities of many condensed matter systems has led to the development of various databases such as Materials Project (MP), Inorganic Crystal Structure Database (ICSD), and many others. These databases ignited the interest of the material science community towards Machine Learning (ML), leading to the development of a new sub-field in material science called material-informatics, which aims to uncover the interrelation between known features and material properties. ML techniques can handle and identify relationships in complex and arbitrarily high-dimensional spaces data, which are almost impossible for human reasoning. Unlike DFT, the ML approach uses data from past computations or experiments. In many cases, ML models have shown their superiority over DFT in terms of accuracy and efficiency in predicting various physical and chemical properties of materials. The incorporation of material property data obtained from atomistic simulations is crucial important to make continuous progress in data-driven methods. In this direction, we use DFT with Perdew-Burke-Ernzerhof (PBE), and Heyd–Scuseria–Ernzerhof (HSE) functionals, to introduce a family of mono-layer isostructural semiconducting tellurides M2N2Te8, with M = {Ti, Zr, Hf} and N = {Si, Ge}. These compounds have been identified to possess direct band gaps that are tunable from 1.0 eV to 1.3 eV, which are well suited for photonics and optoelectronics applications. Additionally, in-plane transport behavior is observed, and small electron and hole (0.11-015 Me) masses are identified along the dominant transport direction. High carrier mobility is found in these compounds, which shows great promise for applications in high-speed electronic devices. Detailed analysis of electronic structures reveals the presence of metal center bicapped trigonal prism as the structural building blocks in these compounds; a common feature in most of the group V chalcogenides helps to understand the atomic origins of promising properties of this unique class of 2D telluride materials. Atomistic simulations based on DFT theory played a vital role in the development of data-driven materials discovery process. However, the resource-based constraints have limited the high-throughput discovery process by using DFT. The main motivation of our work towards the application of machine learning in material science is to assist the discovery process using available material property data in various databases. Incorporation of physical principles in a network-based machine learning (ML) architecture is a fundamental step toward the continued development of artificial intelligence for materials science and condensed matter physics. In this work, as inspired by the Pauling’s rule, we propose that structure motifs (polyhedral formed by cations and surrounding anions) in inorganic crystals can serve as a central input to a machine learning framework for crystalline inorganic materials. We demonstrated that an unsupervised learning algorithm Motif2Vec is able to convert the presence of structural motifs and their connections in a large set of crystalline compounds into unique vectors. The connections among complex materials can be largely determined by the presence of different structural motifs, and their clustering information is identified by our Motif2Vec algorithm. To demonstrate the novel use of structure motif information, we show that a motif-centric learning framework can be effectively created by combining motif information with the recently developed atom-based graph neural networks to form an atom-motif hybrid graph network (AMDNet). Taking advantage of node and edge information on both atomic and motif level, the AMDNet is more accurate than a single graph network in predicting electronic structure related material properties such as band gaps. The work illustrates the route toward the fundamental design of graph neural network learning architecture for complex materials properties by incorporating beyond-atom physical principles. Due to the limitations in resources, it is not feasible to synthesize hundreds of thousands of materials listed in various databases by experiment or compute their detailed properties by using various electronic structure codes and state-of-the-art computational tools. Hence, the identification of an alternative route to screen such databases is very desirable. If identified, this route would be very helpful in reducing the material search space for any application. Categorizing materials based on their structural building blocks is very important to study the underlying physics and to understand the possible mechanisms for any application. Based on structure motifs, we purpose a novel way to categorize, analyze, and visualize the material space called a material network. The connection between any two nodes in this network is determined by using the calculated similarity value (Tanimoto-coeffecient) between each motif and its surrounding information, encoded in terms of a feature vector of length 64. By mapping a known compound, the network thus constructed can be used to screen compounds for the desired application. All the connections of the mapped compound are identified and extracted as a subgraph for further analysis. In our test screening for the transparent conducting oxides (TCO), the proposed network is successful in identifying compounds that are already listed as TCO in the literature. Thus, this indicates its usefulness in reducing the search space for the new TCO materials and various applications. This motif-based material network can serve as an alternate route for functional material discovery and design. / Physics

<p>Mount Erebus, Antarctica, is a volcano with frequent lava-lake eruptions known as strombolian eruptions. The larger of these eruptions create strong seismic waves and have a characteristic seismic signature that can be analyzed through three-component cross-correlation to distinguish smaller strombolian eruptions from the background noise of the volcano. The addition of an infrared camera on the rim of Mount Erebus allows for the confirmation of strombolian eruptions as opposed to unrelated seismic activity. This research finds that eruption events can also be detected categorizing the images using machine learning. Machine learning in seismology is now a commonly used technique, yet to date, no research using machine learning has ever been used in volcanology. Image categorization along with cross-correlation can improve automatic detection of strombolian eruptions.

<p> In this dissertation, we investigate the approach of pure SU(2) lattice gauge theory to its continuum limit using the deconfinement temperature, six gradient scales, and six cooling scales. We find that cooling scales exhibit similarly good scaling behavior as gradient scales, while being computationally more efficient. In addition, we estimate systematic error in continuum limit extrapolations of scale ratios by comparing standard scaling to asymptotic scaling. Finally we study topological observables in pure SU(2) using cooling to smooth the gauge fields, and investigate the sensitivity of cooling scales to topological charge. We find that large numbers of cooling sweeps lead to metastable charge sectors, without destroying physical instantons, provided the lattice spacing is fine enough and the volume is large enough. Continuum limit estimates of the topological susceptibility are obtained, of which we favor &chi;^1/4/<i>T_c</i> = 0.643(12). Differences between cooling scales in different topological sectors turn out to be too small to be detectable within our statistical error.</p><p>

abstract: In this dissertation, I investigate the electronic properties of two important silicon(Si)-based heterojunctions 1) hydrogenated amorphous silicon/crystalline silicon (a-Si:H/c-Si) which has already been commercialized in Heterojunction with Intrinsic Thin-layer (HIT) cells and 2) gallium phosphide/silicon (GaP/Si) which has been suggested to be a good candidate for replacing a-Si:H/c-Si in HIT cells in order to boost the HIT cell’s efficiency. In the first part, the defect states of amorphous silicon (a-Si) and a-Si:H material are studied using density functional theory (DFT). I first employ simulated annealing using molecular dynamics (MD) to create stable configurations of a-Si:H, and then analyze the atomic and electronic structure to investigate which structural defects interact with H, and how the electronic structure changes with H addition. I find that H atoms decrease the density of mid-gap states and increase the band gap of a-Si by binding to Si atoms with strained bonds. My results also indicate that Si atoms with strained bonds creates high-localized orbitals in the mobility gap of a-Si, and the binding of H atoms to them can dramatically decrease their degree of localization. In the second part, I explore the effect of the H binding configuration on the electronic properties of a-Si:H/c-Si heterostructure using density functional theory studies of models of the interface between a-Si:H and c-Si. The electronic properties from DFT show that depending on the energy difference between configurations, the electronic properties are sensitive to the H binding configurations. In the last part, I examine the electronic structure of GaP/Si(001) heterojunctions and the effect of hydrogen H passivation at the interface in comparison to interface mixing, through DFT calculations. My calculations show that due to the heterovalent mismatch nature of the GaP/Si interface, there is a high density of localized states at the abrupt GaP/Si interface due to the excess charge associated with heterovalent bonding, as reported elsewhere. I find that the addition of H leads to additional bonding at the interface which mitigates the charge imbalance, and greatly reduces the density of localized states, leading to a nearly ideal heterojunction. / Dissertation/Thesis / Doctoral Dissertation Electrical Engineering 2019

<p> Interactions between particles and their environment can alter the dynamics of biological systems. In crowded media like the cell, interactions with obstacles can introduce anomalous subdiffusion. Active matter systems, <i>e.g. </i>, bacterial swarms, are nonequilibrium fluids where interparticle interactions and activity cause collective motion and dynamical phases. In this thesis, I discuss my advances in the fields of crowded media and active matter. For crowded media, I studied the effects of soft obstacles and bound mobility on tracer diffusion using a lattice Monte Carlo model. I characterized how bound motion can minimize the effects of hindered anomalous diffusion and obstacle percolation, which has implications for protein movement and interactions in cells. I extended the analysis of binding and bound motion to study the effects of transport across biofilters like the nuclear pore complex (NPC). Using a minimal model, I made predictions on the selectivity of the NPC in terms of physical parameters. Finally, I looked at active matter systems. Using dynamical density functional theory, I studied the temporal evolution of a self-propelled needle system. I mapped out a dynamical phase diagram and discuss the connection between a banding instability and microscopic interactions.</p><p>