Abstract

Structures of metal on graphene play important roles in batteries and hydrogen storage, and discovery of new structures could improve these applications. Through computational methods, the formation energy of a proposed structure can be calculated. Enumeration is the process of creating lists of possible structures to search for low formation energy. We present an enumeration method that allows for efficient treatment of vacancies and apply it to the case of six sites per two carbon atoms. If the separation between two sites in the lattice is less than a specified distance, the pair is referred to as forbidden. The enumeration skips structures where at least one forbidden pair has both sites non-vacant, speeding up the process exponentially by volume. This allows for enumeration of structures of high volumes that would be impossible using standard methods. The search method is applied to the case of scandium, titanium, yttrium, and zirconium on graphene. We report results of several low-energy structures, many of which do not currently appear in the literature. These structures may be useful in future applications of batteries and hydrogen storage.

Abstract

Because of its status as an emerging energy carrier, methods of hydrogen storage are in high demand. Using computational methods, we explore a very large space of graphene structures with different adsorption configurations of hydrogen and calcium as possible mechanisms for hydrogen storage. We expand our search space further than the traditional search space by a factor of almost 20 using cluster expansion to enumerate structures in a variety of computational cell shapes and sizes. This allows for periodicities not studied previously. We also introduce vacancies in the adsorption configurations and study their effects. We conclude that the cell size and shape is an important parameter in the search for stable structures. We also find that vacancies allow for structures with more than 40% calcium concentration without exhibiting signs of calcium stacking. We present four stable structures that show promise for hydrogen storage applications.

Timothy Hecht (Senior Thesis, June 2013,
Advisor: Bret Hess
)

Abstract

Jorge Jimenez Blanco (Senior Thesis, April 2013,
Advisor: Bret Hess
)

Abstract

MRI is widely used in medical imaging. Unlike x-ray and Computed Tomography, MRI uses no ionizing (x-ray) radiation, and is thus less damaging. However, MRI suffers from compara- tively long scan times for some applications. MRI data is sampled in the Fourier domain, and the sampling required for a full image can be quite time consuming in comparison to other medical imaging techniques. The use of certain schemes that allow us to reconstruct full images from under-sampled MRI data may potentially solve this problem. This thesis first provides an in- troduction to the basic principles behind MRI to serve as a foundation for understanding image reconstruction. The mathematical theory of compressed sensing is then presented, which allows images to be reconstructed from randomly under-sampled Fourier domain data. A description of how compressed sensing can be used in MRI to greatly accelerate image acquisition is then pro- vided. Finally, a variety of different MRI acquisition strategies using compressed sensing were simulated, and results are presented. A maximum reduction of 66% of scan time was achieved by the strategies explored with very little loss of image quality.

Abstract

Artificial neural networks have been effective in reducing computation time while achieving remarkable accuracy for a variety of difficult physics and materials science problems. Neural networks are trained iteratively by adjusting the size and shape of sums of non-linear functions by varying the function parameters to fit results for complex non-linear systems. For smaller structures, ab initio simulation methods can be used to determine absorption spectra under field perturbations. However, these methods are impractical for larger structures. Designing and training an artificial neural network with simulated data from density functional theory may allow time-dependent perturbation effects to be calculated more efficiently. I investigate the design considerations of neural network implementations for calculating perturbation-coupled electron oscillations in small molecules. The neural network structure presented is eventually shown to be flawed because it mishandled the complex-valued inputs and outputs that it was trained on. As a result, important complex behavior, required for an accurate approximation of the time-evolution for the system, was ignored. Despite this, valid theory and design considerations are discussed in connection with a new complex-valued network structure that may be adequate to solve the problem.

Abstract

Inspired by recent experimental success in coating carbon nanotubes with silicon, I have computationally modeled carbon nanotubes embedded in silicon using density functional theory. I have characterized the binding energy and electronic structure of several silicon-nanotube systems with diff erent nanotube radii and orientations in the silicon lattice.

Scott Horton (Senior Thesis, April 2010,
Advisor: Bret Hess
)

Abstract

Abstract

A real-space time-domain calculation of the frequency-dependent dielectric constant of nonmetallic crystals is outlined and the integrals required for this calculation are computed. The outline is based on time dependent current density functional theory and is partially implemented in the ab initio density functional theory fireball program. The addition of a vector potential to the Hamiltonian of the system is discussed as well as the need to include the current density in addition to the particle density. The derivation of gradient integrals within a localized atomic-like orbital basis is presented for use in constructing the current density. Due to the generality of the derivation we also give the derivation of the kinetic energy, dipole, and overlap interactions.

Reid Kraniski (Senior Thesis, April 2010,
Advisor: Bret Hess
)

Abstract

Abstract

We calculate the adhesion force between a carbon nanotube and differently shaped surfaces of a silicon block. We use the empirical molecular dynamics program “GULP” to test this. We vary the length of the adhesion contact surface and measure the respective adhesive force. We also vary the curvature of the surface and measure the adhesive force. We find that for contact surfaces between 5nm and 35nm the adhesive force varies linearly with the contact length. We also find that for a (7, 0) carbon nanotube which is kept at approximately 1.9Å from a curved silicon surface, the adhesion force increases rapidly for small radii as the number of nearest neighbor Si-C bonds increases and the nanotube becomes fully embedded in silicon.

Abstract

This study examines the interaction of silicon with carbon nanotubes(CNTs) as well as graphene. In this process the diffusion properties of silicon on CNTs and graphene were modeled. Using the computer program Fireball the behaviors of these systems were studied. The diffusion barriers for silicon and carbon on the surface of graphene were shown to be 0.42 eV and 0.93 eV respectively at zero temperature. The diffusion barrier for silicon on a graphene sheet at higher temperatures is shown to be 0.36 eV. The model of diffusion of a silicon atom on a graphene sheet is shown for temperatures in the range of 500 K to 3500 K. Various rolled hexagonal-planar formations of silicon nanotubes were shown to be stable with and without a carbon nanotube inside them.

Abstract

Photodarkening tellurium modified TiO2 nanoparticles exhibit TiO2 phase change from anatase and brookite when annealed below 700C, to rutile when annealed at or above 700C. An increase in size accompanies the phase change and causes smaller nanocrystals to become visible which cling to the surface of the larger rutile crystals. The smaller crystals are composed of TeO2, as identified by diffraction, and disappear when exposed to the electron beam (observed by TEM) for brief amounts of time. Anatase, brookite and rutile forms of TiO2 are compared by diffraction. A possible cause of photodarkening in samples annealed at 600C is the occurrence of surface interactions between Te and anatase or brookite. These surface interactions may be energetically allowed by annealing samples at temperatures near the TiO2 phase change temperature.

Abstract

We have expanded the capabilities of the ab initio tight-binding molecular dynamics package FIREBALL to include calculations of optical properties. Basic zero order approximation is based on transitions between Kohn-Sham states. Corrections for electron-electron interactions are based on time dependant density functional theory (TDDFT). Consistent with the FIREBALL approach, we use pre-calculated integrals and approximations to make the program faster.

Abstract

We have synthesized titanium dioxide nanocrystals that have been modified by addition of tellurium. After annealing, these nanocrystals become photosensitive, changing color from white to dark red when exposed to light. This change is stable, but reversible upon annealing. Te also causes a change in structure from rutile to anatase. The properties of these nanocrystals are explored.

Abstract

Four-orbital integrals containing the Coulomb kernel are computed to describe the interaction between two electrons as needed in most excited-state calculations. We outline their use in both time-dependent density-functional theory and G¨orling and Levy perturbation theory to find the excited-state properties of many-electron systems. The complete derivations are included to show how these six-dimensional integrals are simplified so that only three- or fourdimensional numerical integrations are required per integral. All numerical integrations are performed via the adaptive Simpson’s rule. The integrals are computed and stored for use in the ab initio density-functional theory fireball program. The integrals for each of the three distinct types of orbital arrangements are plotted for interactions between carbon and hydrogen atoms as a function of separation distance.

Abstract

We improved on the surface-hopping method of John C. Tully (1990) by eliminatingthe rapidly oscillating phase factors from the method. This allows fora significantly larger time step in the surface-hopping method. We also used asecond-order taylor series approximation instead of a constant Hamiltonian forthe time steps.. We used time-dependent perturbation theory to symbolicallysolve for the propagation of the wavefunction. We compare the two methodsusing a simple four-state one-dimensional model. Accuracy was determinedby comparing switching probabilities and expansion coefficients for differenttime step sizes. Our comparisons show the accelerated method runs five toten times faster for the same level of accuracy.

Abstract

We studied the shift that occurs in the electronic states of quantum dot ar- rays as dot spacing is varied. CdSe nanocrystals were synthesized and put into lms with optical quality sucient for absorption and electroabsorption exper- iments. Chemical procedures were used to make and stabilize the nanocrys- tals, and several dierent methods of CdSe synthesis were explored. Electrodes were used to put an electric eld across the CdSe samples during the electroab- sorption experiments. These were made by etching an electrode pattern into chromium covered sapphire disks using photolithographic procedures. The spacing between CdSe nanocrystals was varied using the polymer PMMA, and the absorption results suggested that the CdSe arrays progressed from a smooth spread of electron energy transitions to sharp electron energy transi- tions as the nanocrystal spacing increased.

Christopher Braniff (Senior Thesis, August 2006,
Advisor: Bret Hess
)

Abstract

n/a

Abstract

We apply the method of sum generation upconversion to the photoluminescence to study the dynamics of type Cadmium Telluride nanocrystals. Our sum generation technique uses type1 phasematching with a nonlinear betabarium borate crystal. We maximize the efficiency and give our best values. Three measures are incorporated to reduce stray photons due to the second harmonic generation of the excitation laser created in the crystal: noncollinear geometry, short pass filters and a monochromator. Applying this upconversion method to the photoluminescence of nanocrystals shows the rapid rise in luminescence following the integrated excitation pulse and its slow decay rate due to the to the decrease in carrier population. This technique allows us to observe the excitation lifetimes and the relaxation rate of carriers at the high energies.

Eric Wald (Capstone, September 2005,
Advisor: Bret Hess
)

Abstract

n/a

Josh Holt (Senior Thesis, August 2003,
Advisor: Bret Hess
)

Abstract

Christopher Bass (Senior Thesis, August 1999,
Advisor: Bret Hess
)

Abstract

Mary Landon Martin (Honors Thesis, June 1999,
Advisor: Bret Hess
)

Abstract

Rodion Tikhoplav (Masters Thesis, December 1999,
Advisor: Bret Hess
)

Abstract

We have studied the photoluminescence (PL) and optical absorption in Poly [2-methoxy, 5-(2’-ethyl-hexyloxy) phenylene vinylene] or MEH-PPV as a function of pressure up to 50 kbar. The PL band and the optical absorption edge redshift with pressure and the PL efficiency decreases dramatically. These features are similar to results reported earlier for pressure effect in unsubstituted poly phenylene vinylene (PPV). However, the initial pressure redshift in MEH-PPV is about twice as large as that in unsubstituted PPV, and the PL decays faster. The vibrionic structure in the PL is maintained up to the highest pressures. It appears that pressure increases the number of non-emissive interchain excitons, which decreases the photoluminescence.

Steve Summers (Senior Thesis, April 1997,
Advisor: Bret Hess
)

Abstract