Magnetohydrodynamics (MHD) studies the dynamics of an electrically conducting fluid under the influence of a magnetic field. Many astrophysical phenomena are related to MHD, and computer simulations are used to model these dynamics. In this thesis, we conduct MHD simulations of non-radiative black hole accretion as well as fast magnetic reconnection. By performing large scale three dimensional parallel MHD simulations on supercomputers and using a deformed-mesh algorithm, we were able to conduct very high dynamical range simulations of black hole accretion of Sgr A* at the Galactic Center. We find a generic set of solutions, and make specific predictions for currently feasible observations of rotation measure (RM). The magnetized accretion flow is subsonic and lacks outward convection flux, making the accretion rate very small and having a density slope of around $-1$. There is no tendency for the flows to become rotationally supported, and the slow time variability of the RM is a key quantitative signature of this accretion flow. We also provide a constructive numerical example of fast magnetic reconnection in a three-dimensional periodic box. Reconnection is initiated by a strong, localized perturbation to the field lines and the solution is intrinsically three-dimensional. Approximately $30\%$ of the magnetic energy is released in an event which lasts about one Alfv\'en time, but only after a delay during which the field lines evolve into a critical configuration. In the co-moving frame of the reconnection regions, reconnection occurs through an X-like point, analogous to the Petschek reconnection. The dynamics appear to be driven by global flows rather than local processes. In addition to issues pertaining to physics, we present results on the acceleration of MHD simulations using heterogeneous computing systems \cite. We have implemented the MHD code on a variety of heterogeneous and multi-core architectures (multi-core x86, Cell, Nvidia and ATI GPU) using different languages (FORTRAN, C, Cell, CUDA and OpenCL). Initial performance results for these systems are presented, and we conclude that substantial gains in performance over traditional systems are possible. In particular, it is possible to extract a greater percentage of peak theoretical performance from some heterogeneous systems when compared to x86 architectures.

Magnetohydrodynamics (MHD) studies the dynamics of an electrically conducting fluid under the influence of a magnetic field. Many astrophysical phenomena are related to MHD, and computer simulations are used to model these dynamics. In this thesis, we conduct MHD simulations of non-radiative black hole accretion as well as fast magnetic reconnection. By performing large scale three dimensional parallel MHD simulations on supercomputers and using a deformed-mesh algorithm, we were able to conduct very high dynamical range simulations of black hole accretion of Sgr A* at the Galactic Center. We find a generic set of solutions, and make specific predictions for currently feasible observations of rotation measure (RM). The magnetized accretion flow is subsonic and lacks outward convection flux, making the accretion rate very small and having a density slope of around $-1$. There is no tendency for the flows to become rotationally supported, and the slow time variability of the RM is a key quantitative signature of this accretion flow. We also provide a constructive numerical example of fast magnetic reconnection in a three-dimensional periodic box. Reconnection is initiated by a strong, localized perturbation to the field lines and the solution is intrinsically three-dimensional. Approximately $30\%$ of the magnetic energy is released in an event which lasts about one Alfv\'en time, but only after a delay during which the field lines evolve into a critical configuration. In the co-moving frame of the reconnection regions, reconnection occurs through an X-like point, analogous to the Petschek reconnection. The dynamics appear to be driven by global flows rather than local processes. In addition to issues pertaining to physics, we present results on the acceleration of MHD simulations using heterogeneous computing systems \cite. We have implemented the MHD code on a variety of heterogeneous and multi-core architectures (multi-core x86, Cell, Nvidia and ATI GPU) using different languages (FORTRAN, C, Cell, CUDA and OpenCL). Initial performance results for these systems are presented, and we conclude that substantial gains in performance over traditional systems are possible. In particular, it is possible to extract a greater percentage of peak theoretical performance from some heterogeneous systems when compared to x86 architectures.

The International Thermonuclear Experimental Reactor (ITER) is currently designed to use graphite targets in the divertor for power handling and impurity control. Understanding and quantifying chemical sputtering is therefore key to the success of fusion as a clean energy source. The principal goal of this thesis is to design and carry out experiments, then analyze and interpret the results in order to elucidate the role of chemical sputtering in carbon sources in the DIII-D tokamak. A self-contained gas puff system has been designed, constructed, and employed for in-situ study of chemical erosion. The porous plug injector (PPI) releases methane through a porous graphite surface into the divertor plasma at a precisely calibrated rate, minimizing perturbation to local plasma while replicating the immediate environment of methane molecules released from a solid graphite surface more accurately than done previously. For the first time in a tokamak environment, the methane flow rate used in a puffing experiment was the same order of magnitude as that expected from laboratory experiments for intrinsic chemical sputtering. Effective photon efficiencies for injection are reported; results are found to have significant dependencies on surface conditions and the divertor operating regime. The contribution of sputtering processes to sources of C0 and C+ are assessed through measurement of background and incremental spectroscopic emissions of both physically and chemically-released sputtering products and by CI, 910 nm line profile fitting. Comparison of background and incremental emissions of chemically-released products demonstrate a dramatic drop in production of CH in cold and detached conditions. Finally, the chemical erosion yield is calculated in both attached and cold-divertor conditions and found to be much closer to that measured ex-situ in ion beam experiments than previously determined in DII-D. These observations represent a positive result for ITER which will operate at all times with a detached divertor, i.e., a low chemical sputtering yield. Results and analysis techniques presented here point the direction for future experiments with the PPI for study of chemical sputtering in the tokamak edge environment.

The International Thermonuclear Experimental Reactor (ITER) is currently designed to use graphite targets in the divertor for power handling and impurity control. Understanding and quantifying chemical sputtering is therefore key to the success of fusion as a clean energy source. The principal goal of this thesis is to design and carry out experiments, then analyze and interpret the results in order to elucidate the role of chemical sputtering in carbon sources in the DIII-D tokamak. A self-contained gas puff system has been designed, constructed, and employed for in-situ study of chemical erosion. The porous plug injector (PPI) releases methane through a porous graphite surface into the divertor plasma at a precisely calibrated rate, minimizing perturbation to local plasma while replicating the immediate environment of methane molecules released from a solid graphite surface more accurately than done previously. For the first time in a tokamak environment, the methane flow rate used in a puffing experiment was the same order of magnitude as that expected from laboratory experiments for intrinsic chemical sputtering. Effective photon efficiencies for injection are reported; results are found to have significant dependencies on surface conditions and the divertor operating regime. The contribution of sputtering processes to sources of C0 and C+ are assessed through measurement of background and incremental spectroscopic emissions of both physically and chemically-released sputtering products and by CI, 910 nm line profile fitting. Comparison of background and incremental emissions of chemically-released products demonstrate a dramatic drop in production of CH in cold and detached conditions. Finally, the chemical erosion yield is calculated in both attached and cold-divertor conditions and found to be much closer to that measured ex-situ in ion beam experiments than previously determined in DII-D. These observations represent a positive result for ITER which will operate at all times with a detached divertor, i.e., a low chemical sputtering yield. Results and analysis techniques presented here point the direction for future experiments with the PPI for study of chemical sputtering in the tokamak edge environment.

Plumes are formed when a continuous buoyant forcing is supplied at a localized source. Buoyancy can be created by either a heat flux, a compositional difference between the fluid coming from the source and its surroundings, or a combination of both. In this thesis, two types of laminar plumes with different forcing mechanisms were investigated: forced plumes and autocatalytic plumes. The forced plumes were compositionally buoyant and were injected with inertial forcing into a fluid filled tank. The autocatalytic plumes were produced without mechanical forcing by buoyancy that was entirely the consequence of a nonlinear chemical reaction -- the iodate-arsenous acid (IAA) reaction. This reaction propagates as a reacting front and produces buoyancy by its exothermicity, and by the compositional difference between the reactant and product. Both the forced and autocatalytic plumes were examined in starting and steady states. The starting, or transient, state of the plume occurs when it initially rises through a fluid and develops a plume head on top of a trailing conduit. The steady state emerges after the plume head has risen to the top of a fluid filled tank leaving only a persistent conduit. Plume behaviour was studied through experimentation, simulation, and by using simple theoretical analysis. We performed the first ever study of plumes as they crossed over the transition between buoyancy-driven to momentum-driven flow. Regardless of the driving mechanism, forced plumes were found to exhibit a single power law relationship that explains their ascent velocity. However, the morphology of the plume heads was found to depend on the dominating driving mechanism. Confined heads were produced by buoyancy-driven plumes, and dispersed heads by momentum-driven plumes. Autocatalytic plumes were found to have rich dynamics that are a consequence of the interplay between fluid flow and chemical reaction. These plumes produced accelerating heads that detached from the conduit, forming free vortex rings. A second-generation head would then develop at the point of detachment. The detachment process for plumes was sensitively dependent on small fluctuations in their initial formation. In some cases, head detachment could occur multiple times for a single experimental run, thereby producing several generations of autocatalytic vortex rings. Head detachment was reproduced and studied using autocatalytic plume simulations. Autocatalytic flame balls, a phenomenon closely related to autocatalytic plumes, were also simulated. Flame balls were found to have three dynamical regimes. Below a critical radius, the smallest flame balls experienced front death. Above this radius, they formed elongating, reacting tails. The largest flame balls formed filamentary tails unable to sustain a reaction.

Plumes are formed when a continuous buoyant forcing is supplied at a localized source. Buoyancy can be created by either a heat flux, a compositional difference between the fluid coming from the source and its surroundings, or a combination of both. In this thesis, two types of laminar plumes with different forcing mechanisms were investigated: forced plumes and autocatalytic plumes. The forced plumes were compositionally buoyant and were injected with inertial forcing into a fluid filled tank. The autocatalytic plumes were produced without mechanical forcing by buoyancy that was entirely the consequence of a nonlinear chemical reaction -- the iodate-arsenous acid (IAA) reaction. This reaction propagates as a reacting front and produces buoyancy by its exothermicity, and by the compositional difference between the reactant and product. Both the forced and autocatalytic plumes were examined in starting and steady states. The starting, or transient, state of the plume occurs when it initially rises through a fluid and develops a plume head on top of a trailing conduit. The steady state emerges after the plume head has risen to the top of a fluid filled tank leaving only a persistent conduit. Plume behaviour was studied through experimentation, simulation, and by using simple theoretical analysis. We performed the first ever study of plumes as they crossed over the transition between buoyancy-driven to momentum-driven flow. Regardless of the driving mechanism, forced plumes were found to exhibit a single power law relationship that explains their ascent velocity. However, the morphology of the plume heads was found to depend on the dominating driving mechanism. Confined heads were produced by buoyancy-driven plumes, and dispersed heads by momentum-driven plumes. Autocatalytic plumes were found to have rich dynamics that are a consequence of the interplay between fluid flow and chemical reaction. These plumes produced accelerating heads that detached from the conduit, forming free vortex rings. A second-generation head would then develop at the point of detachment. The detachment process for plumes was sensitively dependent on small fluctuations in their initial formation. In some cases, head detachment could occur multiple times for a single experimental run, thereby producing several generations of autocatalytic vortex rings. Head detachment was reproduced and studied using autocatalytic plume simulations. Autocatalytic flame balls, a phenomenon closely related to autocatalytic plumes, were also simulated. Flame balls were found to have three dynamical regimes. Below a critical radius, the smallest flame balls experienced front death. Above this radius, they formed elongating, reacting tails. The largest flame balls formed filamentary tails unable to sustain a reaction.

Doctor of Philosophy / Department of Mechanical and Nuclear Engineering / William L. Dunn / Dense plasma focus (DPF) devices are pulsed power devices capable of producing short-lived, hot and dense plasmas (~10[superscript]19 cm[superscript]-3) through a fast compression of plasma sheath. A DPF device provides intense bursts of electrons and ion beams, X-rays, and 2.5 MeV neutrons when operated with deuterium through the fusion reaction [superscript]2H(d,n)[superscript]3He. The Kansas State University DPF machine was designed and constructed in early 2010. The device was characterized to determine its performance as a neutron source. The device was shown to produce 5.0x10[superscript]7 neutrons/pulse using a tungsten-copper anode. Such machines have the advantages of being non-radioactive, movable, and producing short pulses (typically tens of nanoseconds), which allows rapid interrogation. The signature-based radiation-scanning (SBRS) method has been used to distinguish targets that contain explosives or explosive surrogates from targets that contain materials called “inert,” meaning they are not explosive-like. Different targets were placed in front of the DPF source at a distance of 45 cm. Four BC-418 plastic scintillators were used to measure the direct neutron yield and the neutrons scattered from various targets; the neutron source and the detectors were shielded with layers of lead, stainless steel, and borated polyethylene to shield against the X-rays and neutrons. One of the plastic scintillators was set at 70[supercript]o and two were set at 110[superscript]o from the line of the neutron beam; a bare [superscript]3He tube was used for detecting scattered thermal neutrons. Twelve metal cans of one-gallon each containing four explosive surrogates and eight inert materials were used as targets. Nine materials in five-gallon cans including three explosive surrogates were also used. The SBRS method indicated a capability to distinguish the explosive surrogates in both experiments, although the five gallon targets gave more accurate results. The MCNP code was used to validate the experimental work and to simulate real explosives. The simulations indicated the possibility to use the time of flight (TOF) technique in future experimental work, and were able to distinguish all the real explosives from the inert materials.

Ultrafast electron diffraction is a novel pump-probe technique which aims to determine transient structures during photoinduced chemical reactions and other structural transitions. This technique provides structural information at the atomic level of inspection by using an electron pulse as a diffractive probe. The atomic motions of interest happen on the 100 fs = 10^(−13) s time scale. To observe these atomic motions, a probe which matches this time scale is required. In this thesis, I describe the development of an electron diffractometer which is capable of 200 fs temporal resolution while maintaining high signal level per electron pulse. This was made possible by the construction of an ultra-compact photoactivated 60 keV femtosecond electron gun. Traditional electron pulse characterization methods are unsuitable for high number density femtosecond electron pulses such as the pulses produced by this electron gun. I developed two techniques based on the laser ponderomotive force to reliably determine the duration of femtosecond electron pulses into the sub-100 fs range. These techniques produce a direct cross-correlation trace between the electron pulse and a laser pulse. The results of these measurements confirmed the temporal resolution of the newly developed femtosecond electron diffractometer. This cross-correlation technique was also used to calibrate a method for the determination of the temporal overlap of electron and laser pulses. These techniques provide the pulse diagnostics necessary to utilize the temporal resolution provided by femtosecond electron pulses. Owing to their high charge-to-mass ratio, electrons are a sensitive probe for electric fields. I used femtosecond electron pulses in an electron deflectometry experiment to directly observe the transient charge distributions produced during femtosecond laser ablation of a silicon (100) surface. We found an electric field strength of 3.5 × 10^6 V/m produced by the emission of 5.3 × 10^11 electrons/cm^2 just 3 ps after an excitation pulse of 5.6 J/cm^2 . This observation allowed us to rule out Coulomb explosion as the mechanism for ablation under the conditions present in this experiment.

The normal injection of a liquid jet into a gaseous crossflow has many engineering applications. In this thesis, detailed numerical simulations based on the level set method are employed to understand the physical mechanism underlying the jet ``surface breakup''. The numerical observations reveal the existence of hydrodynamic instabilities on the jet periphery. The temporal growth of such azimuthal instabilities leads to the formation of interface corrugations, which are eventually sheared off of the jet surface as sheet-like structures. The sheets finally undergo disintegration into ligaments and drops during the surface breakup process. Temporal linear stability analyses are employed to understand the nature of these instabilities. To facilitate the analysis, analytical solutions for the flow fields of the jet and the crossflow are derived. We identify the ``shear instability'' as the primary destabilization mechanism in the flow. This inherently inviscid mechanism opposes the previously suggested mechanism of surface breakup (known as ``boundary layer stripping''), which is based on a viscous interpretation. The influence of the jet-to-crossflow density ratio on the flow stability are also studied. The findings show that a higher density jet leads to higher wavenumber instabilities on the jet surface and thereby subsequent smaller drops and ligaments. The stability characteristics of the most amplified modes (i.e., the wavenumber and corresponding growth rate) obtained from stability analyses and numerical simulations are in good agreement. The stability results of the jet also show that the density may have a non-monotonic stabilizing/destabilizing effect on the flow stability. To investigate such effect, the concept of wave resonance are employed to physically interpret the inviscid instability mechanism in two-phase flows with sharp interfaces and linear velocity profiles. We demonstrate that neutrally stable waves are formed due to the density jump in the flow, in addition to the well-known vorticity (Rayleigh) waves. Under certain conditions, such neutral waves are capable of resonating and generating unstable modes. The resonance of different pairs of neutral waves, therefore, results in either stabilizing or destabilizing effect of density variation. We predict similar reasoning behind the density behavior in the jet in crossflow configuration with smoothly varying velocity and density profiles.

Ultrafast electron diffraction is a novel pump-probe technique which aims to determine transient structures during photoinduced chemical reactions and other structural transitions. This technique provides structural information at the atomic level of inspection by using an electron pulse as a diffractive probe. The atomic motions of interest happen on the 100 fs = 10^(−13) s time scale. To observe these atomic motions, a probe which matches this time scale is required. In this thesis, I describe the development of an electron diffractometer which is capable of 200 fs temporal resolution while maintaining high signal level per electron pulse. This was made possible by the construction of an ultra-compact photoactivated 60 keV femtosecond electron gun. Traditional electron pulse characterization methods are unsuitable for high number density femtosecond electron pulses such as the pulses produced by this electron gun. I developed two techniques based on the laser ponderomotive force to reliably determine the duration of femtosecond electron pulses into the sub-100 fs range. These techniques produce a direct cross-correlation trace between the electron pulse and a laser pulse. The results of these measurements confirmed the temporal resolution of the newly developed femtosecond electron diffractometer. This cross-correlation technique was also used to calibrate a method for the determination of the temporal overlap of electron and laser pulses. These techniques provide the pulse diagnostics necessary to utilize the temporal resolution provided by femtosecond electron pulses. Owing to their high charge-to-mass ratio, electrons are a sensitive probe for electric fields. I used femtosecond electron pulses in an electron deflectometry experiment to directly observe the transient charge distributions produced during femtosecond laser ablation of a silicon (100) surface. We found an electric field strength of 3.5 × 10^6 V/m produced by the emission of 5.3 × 10^11 electrons/cm^2 just 3 ps after an excitation pulse of 5.6 J/cm^2 . This observation allowed us to rule out Coulomb explosion as the mechanism for ablation under the conditions present in this experiment.