Search

In this study, a computational investigation of multiphase flow within supersonic micronozzles has been performed in an effort to quantify the impact on thrust production and nozzle efficiency. Mot...

Micronozzles represent a unique flow regime defined by low Reynolds numbers (Re

The ability of some nanostructured materials to perform as effective heterogeneous catalysts is potentially hindered by the failure of the liquid reactant to effectively wet the solid catalyst surface. In this work, two different chemical reactions, each involving a change of phase from liquid to gas on a solid catalyst surface, are investigated. The first reaction is the catalyzed decomposition of a H2O2 monopropellant within a micro-chemical reactor chamber, decorated with RuO2 nanorods (NRs). The second reaction involves the electrolysis of dilute aqueous solutions of H2SO4 performed with the cathode electrode coated with different densities and sizes of RuO2 NRs. In the catalyzed H2O2 decomposition, the reaction rate is observed to decrease with increasing catalyst surface density because of a failure of the liquid to wet on the catalyst surface. In the electrolysis experiment, however, the reaction rate increased in proportion to the surface density of RuO2 NRs. In this case, the electrical bias applied to drive the electrolysis reaction also causes an electrostatic force of attraction between the fluid and the NR coated surface, and thus assures effective wetting.

This computational study investigates nonlinear dynamics of unstable convection in a 3 D toroidal shaped thermal convection loop (i.e., thermosyphon) with heat flux boundary conditions; results are compared to prior 2 D simulations. The lower half of the thermosyphon is subjected to a positive heat flux into the system while the upper half is cooled by an equal-but-opposite heat flux out of the system. Water is employed as the working fluid with fully temperature dependent thermophysical properties and the system of governing equations is solved using a finite volume method. Numerical simulations are performed for varying magnitudes of heat flux ( 1.0 W / m 2 ⩽ q ″ ⩽ 1.0 × 10 4 W / m 2 ) to yield Rayleigh numbers (i.e., buoyant forcing) ranging from 2.83 × 10 4 ⩽ Ra ⩽ 2.83 × 10 8 . Delineation of multiple convective flow regimes is achieved through evolution of the bulk-mass-flow time-series and the trajectory of the mass flow attractor. Simulation results demonstrate that multiple regimes are possible and include: (1) conduction, (2) damped, stable convection that asymptotes to steady-state, (3) unstable, Lorenz-like chaotic convection with flow reversals, and (4) high Rayleigh, aperiodic stable convection without flow reversals. For the Rayleigh numbers considered, it is observed that certain flow regimes are not accessible in toroidal simulations owing to the constraints of additional surface boundaries in a 3 D system. The RMS of mass flow rate, power spectra of oscillatory behavior, dominant oscillatory frequency, and residence time are also described as a function of the buoyant forcing in the system.

Converging flows of a gas and a liquid at a microchannel cross junction, under proper conditions, can result in the formation of periodic, dispersed microslugs. This microslug formation phenomenon has been proposed as the basis for a fuel injection system in a novel, \'discrete\' monopropellant microthruster designed for use in next-generation miniaturized satellites. Previous experimental studies demonstrated the ability to generate fuel slugs with characteristics commensurate with the intended application during steady-state operation. In this work, numerical and experimental techniques are used to study the effect of valve actuation on slug characteristics, and the results are used to compare with equivalent steady-state slugs. Computational simulations of a valve with a 1 ms valve-actuation cycle show that as the ratio of the response time of the valve to the fully open time is increased, transient effects can increase slug length by up to 17%. The simulations also demonstrate that the effect of the valve is largely independent of surface tension coefficient, which is the thermophysical parameter most responsible for slug formation characteristics. Flow visualization experiments performed using a miniature valve with a 20 ms response time showed less than a 1% change in the length of slugs formed during the actuation cycle. The results of this study indicate that impulse bit and thrust calculations can discount transient effects for slower valves, but as valve technology improves transient effects may become more significant.

In this study, numerical computations are performed that examine the thrust production and efficiency of supersonic micronozzles with bell-shaped expanders. The bell geometry is favored on the macroscale for its flow alignment. To date, concerns over microfabrication challenges of the contoured geometry have limited its consideration for microscale applications. Three different bell expander configurations are examined (100% full bell, 80%, and 60%) for two-dimensional and three-dimensional duct configurations of varying depths (25–200 μm), and a decomposed H2O2 monopropellant is used as the working fluid, and the associated throat Reynolds numbers range from 15 to 800. Owing to the inherently low Reynolds numbers on the microscale, substantial viscous subsonic layers develop on the walls of the nozzle expander, retard the bulk flow, and reduce the nozzle performance. The thrust production and specific impulse efficiency are computed for the various flow scenarios and nozzle geometries to delineate the i...

The next generation of miniaturized satellites (`nanosats') feature dramatically reduced thrust and impulse requirements for purposes of spacecraft attitude control and maneuvering. The present study is a joint computational and experimental design effort at developing a new MEMS-based microreactor configuration for incorporation into a monopropellant micropropulsion system. Numerical models of the gas phase catalytic decomposition in microchannel configurations are used to obtain critical sizing requirements for the reactor design. The computational results show that the length scales necessary for complete decomposition are compatible with MEMS-based designs; however, it is also found that the catalytic process is dominated by mass diffusion characteristics within the flow at this scale. Experimentally, a microscale catalytic reactor prototype has been designed and microfabricated using MEMS techniques. The reactor uses self-assembled ruthenium oxide nanorods grown on the wall surfaces as a catalyst. Experimental testing indicates that only partial decomposition of the hydrogen peroxide is achieved. Among the potential sources of the incomplete decomposition, a likely cause appears to be the inability of the H2O2 reactant stream to adequately wet the surface of the catalyst film composed of a high surface density of RuO2 nanorods.

This computational study investigates the nonlinear dynamics of unstable convection in a 2D thermal convection loop (i.e., thermosyphon) with heat flux boundary conditions. The lower half of the thermosyphon is subjected to a positive heat flux into the system while the upper half is cooled by an equal-but-opposite heat flux out of the system. Water is employed as the working fluid with fully temperature dependent thermophysical properties and the system of governing equations is solved using a finite volume method. Numerical simulations are performed for varying levels of heat flux and varying strengths of gravity to yield Rayleigh numbers ranging from 1.5 × 10 2 to 2.8 × 10 7 . Simulation results demonstrate that multiple regimes are possible and include: (1) conduction, (2) damped, stable convection that asymptotes to steady-state, (3) unstable, Lorenz-like chaotic convection with flow reversals, and (4) high Rayleigh, aperiodic stable convection without flow reversals. Delineation of the various flow regimes, as characterized by the temporal evolution of bulk mass flow rate, is obtained in terms of heat flux, gravity, and the Rayleigh number. Temporal frequencies of the oscillatory behavior and residence time in a circulatory direction are explored and described for the various thermal and gravitational forcing (Rayleigh number) applied to the system.

Anumericalmodel to characterize the influence of wall heat transfer on performance of amicroelectromechanical systems (MEMS)-based supersonic nozzle is reported. Owing to the large surface-area-to-volume ratio and inherently lowReynolds numbers of aMEMSdevice, wall phenomena, such as viscous forces and heat transfer, play critical roles in shaping performance characteristics of themicronozzle. Viscous subsonic layers inhibit flow and can grow sufficiently large on the nozzle expanderwalls, potentiallymerging to cause the flow to be subsonic at the nozzle exit, and result in reduced efficiency and performance. Heat flux from the flow into the surrounding substrate can mitigate subsonic layer growth and improve overall thrust production. In this study, subsonic layer growth is quantified to characterize the impact on performance of micronozzles with a flowfield that is subject to wall heat transfer. Both twoand three-dimensional (3-D) simulations are performed for varying expanderhalf-angles (15 deg, 30 deg, and 45 deg) and varying throat Reynolds numbers (30–800), whereas the depth of the 3-D nozzle is varied (25–300 m). Simulation results and nozzle efficiencies are compared with inviscid theory, previous adiabatic results, and existing numerical and experimental data. It is found that heat loss to the substratewill further accelerate the supersonic core flowviaRayleigh flow theory and can reduce subsonic layer growth. These effects can combine to alter the micronozzle expansion angle, which maximizes thrust production and specific impulse efficiency.

A numerical investigation of three-dimensional flow in a converging–diverging microelectromechanical-systemsbased supersonic nozzle design is reported. Based on microfabrication techniques, the nozzle geometry is a twodimensional pattern etched into a substrate to a specified depth, thus resulting in a three-dimensional ductlike geometry and flowfield. Owing to the low Reynolds numbers associated with the microscale, viscous effects result in substantial subsonic layers that develop along expander walls and reduce nozzle efficiency. In this study, the threedimensional flowfield is simulated with a continuummodel and decomposed hydrogen peroxide as the working fluid monopropellant. Numerical simulations are performed over a range of nozzle expander half-angles (15–45 ), for several nozzle depths (25–400 m), and for varying throat Reynolds numbers (15–800). Simulation results are analyzed to determine subsonic layer growth rates and to delineate their impact on micronozzle thrust and performance. Specific impulse efficiencies are compared with quasi-one-dimensional theory, previous twodimensional simulations, and data from the micronozzle literature, including direct simulation Monte Carlo and experimental studies. It is found that three-dimensional impulse efficiencies can be much less than corresponding two-dimensional model results owing to the presence of viscous, subsonic layers on the additional walls in a threedimensional device. Themicronozzle depth is a critical parameter that impacts thrust and efficiency. For sufficiently shallow nozzles, it is possible for the subsonic layers to merge and decelerate the flow to subsonic conditions, thus degrading nozzle performance.

Transient laminar natural convection regimes occurring in a thermal convection loop heated from below and cooled from above are investigated numerically for a wide range of Rayleigh numbers spanning the interval from 10 3 to 2.6 × 10 7 . In the model system, the lower half of the loop is heated and maintained at a constant high temperature, while the upper half is cooled and maintained at a constant low temperature. A three-dimensional numerical model based on the finite volume method is used to solve the system of governing flow equations. Simulations are performed using water as the working fluid (Pr = 5.83) and detailed numerical results are presented and discussed for conduction, steady convection, and unsteady flow regimes. Although this subject has attracted researchers for decades, there have been no detailed three-dimensional numerical simulations of the dynamics of flow in the thermal convection loop. The objective of the present study is to fill this gap by presenting the temporal evolution of the velocity and temperature fields at key locations within the system. Emphasis is given to the analysis of dynamical behavior of the flow during the unsteady regime. The complexity of flow in the loop, which is characterized by vertical structures and flow recirculation, is visualized for the first time by performing detailed 3-D numerical simulations.

A numerical investigation of transient supersonic flow through a two-dimensional linear micronozzle has been performed. The baseline model for the study is derived from the NASA Goddard Space Flight Center microelectromechanical-systems-based hydrogen peroxide prototype microthruster. A hyperbolic-tangent actuation profile is used to simulate the opening and closing of a microvalve with a maximum inlet stagnation pressure of 250 kPa, which generates a maximum throat Reynolds number of Re 800. The complete duty cycle occurs over 1.7ms.Numerical simulations have been conducted for expanderhalf-angles of 10–50 , andboth slip and no-slip wall boundary conditions have been examined. The propulsion scheme employs 85%-pure hydrogen peroxide as the monopropellant fuel. Simulation results have been analyzed, and thrust production as a function of time has been quantified, along with the total impulse delivered. Micronozzle impulse efficiency has also been determined based on a theoretical maximum impulse achieved by a quasi-1-D inviscid flow responding instantaneously to the actuation profile. It is found that both the flow and thrust exhibit a response lag to the timevarying inlet pressure profile. Simulations indicate that a maximum efficiency and impulse occur for an expander half-angle of 30 for the no-slip wall boundaries, and the slip simulations demonstrate a maximum plateau in the range of 20–30 ; these angles are significantly larger than with traditional conical nozzle designs.

This paper numerically investigates the nonlinear dynamics of the unstable convection regime of the thermal convection loop, an experimental analogue of the Lorenz model. The lower half of the toroidal loop is heated and maintained at a constant high temperature, while the upper half is cooled at a constant low temperature. Subject to the proper boundary conditions, the system of governing equations is solved using a finite volume method. The numerical simulations are performed for water corresponding to Pr = 5.83 and Rayleigh number varying from 1000 to 150,000. In the case of a loop heated from below and cooled from above, it has been demonstrated theoretically and experimentally in the literature that multiple flow regimes are possible. Numerical results in terms of streamlines, isotherms, and local heat flux distributions along the walls are presented for each flow regime. Although several studies have investigated the chaotic regime of convection loops, there have been no detailed numerical simulations of the dynamics of flow reversals. Fine-scale flow behavior during the transition from one flow direction to another is illustrated by the temporal evolution of temperature distribution, mass flow rate, and local heat flux at selected locations in the system. Issues related to the observed Kelvin–Helmholtz instabilities are discussed.

A comprehensive numerical investigation of a steady viscous flow through a two-dimensional supersonic linear micronozzle has been performed. The baseline model for the study is derived from the NASAGoddard Space Flight Center microelectromechanical systems-based hydrogen peroxide prototype microthruster. On the microscale, substantial viscous subsonic layers may form on the nozzle expander walls, which reduce thrust and efficiency. One approach to compensate for the presence of these layers has been to designmicronozzleswith expansion angles larger than traditional macroscale nozzles. Numerical simulations have been conducted for a range of Reynolds numbers (Re 15–800) and for expander half-angles of 10–50 deg. Twodifferentmonopropellant fuels have been considered: decomposed 85% pure hydrogen peroxide and decomposed hydrazine. It is found that an inherent tradeoff exists between the viscous losses and the losses resulting from the nonaxial exit flow at larger expansion angles. Our simulations indicate that themaximumnozzle efficiency occurs for bothmonopropellants at a nozzle expansion halfangle of approximately 30 deg, which is significantly larger than that of traditional conical nozzle designs.

Physiological implications for the clotting of an artery, hemostasis, can be found by examining the time-averaged local strain rates in the vicinity of the puncture. The time-averaged leakage of fluid leaving a puncture in a pressurized vessel immersed in a quiescent, miscible medium is studied under pulsatile flow conditions. The time-averaged local strain rates at the puncture site can then be investigated. The time-averaged leakage and local strain rates at the puncture are characterized as functions of various hydrodynamic and geometric conditions. Dimensional analysis shows that the time-averaged leakage rate, Q*, is a function of the mean Reynolds number, the Womersley number, the pressure ratio between the center of the tube and the external hydrostatic pressure, P*, and the hole-to-main tube diameter ratio, D*. A 3-D finite volume computational model is constructed for laminar flow of a Newtonian fluid under pulsatile conditions and validated with supporting experiments. The time-averaged ...

This paper addresses the potential augmentation of natural convection heat transfer in Rayleigh–Benard enclosures when filled with a certain type of binary gas mixture. To form the binary gas mixtures, helium (He) is the primary gas and the secondary gases are nitrogen (N2), oxygen (O2), carbon dioxide (CO2) and methane (CH4). Each of the thermo-physical properties participating in the binary gas mixtures viscosity ηm, thermal conductivity λm, density ρm, and heat capacity at constant pressure Cp,m depends on the molar gas composition, temperature and pressure. Results are presented in terms of the maximum allied heat transfer coefficient hm,max/B at the optimal mole gas composition wopt, in the w-domain [0, 1] for the entire range of laminar and turbulent conditions. In the conduction regime, He provides the best heat transfer regardless of temperature. In the convection regime at 300 K a He–CO2 mixture usually provides the maximum heat transfer, whereas at 1000 K pure methane CH4 is the optimum. In addition, a detailed thermo-fluidic structure of the thermal convection patterns in the Rayleigh–Benard enclosure was analyzed by performing 2-D numerical simulations.

We revisit the problem of the stability of pulsatile pipe flow for axisymmetric perturbations. In contrast to the earlier approach based on the Chebyshev expansion for the spatial discretization [J. Appl. Mech. ASME 53 (1986) 187], we use the set of the eigenfunctions derived from the longwave limit of the Orr–Sommerfeld equation. We show that the Orr–Sommerfeld basis gives greater accuracy than the Chebyshev basis if fewer terms are used in the Galerkin expansion. For the time evolution of the flow perturbation, instead of the usual Floquet analysis, a different representation for the solution of the periodic system of linear differential equations is employed. We found that the flow structures corresponding to the largest energy growth are toroidal vortex tubes. They are stretched by the shear stress of the mean flow so that a maximum energy growth occurs. The flow perturbation subsequently decays due to viscous effects. The maximum energy growth is then evaluated over a range of Reynolds and Womersley numbers. Asymptotic solutions provided for the longwave limit as well as the limit of large Womersley numbers agree well with the numerical results, confirming the known linear stability of the flow.

A parametric, three-dimensional, computational study examining steady state performance of a planar plug micronozzle is conducted. Nozzle depth and throat Reynolds number are varied for a single plug contour designed using an approximation for the Method of Characteristics. The Reynolds number is varied from 80 − 820, for nozzle depths of 90μm-360μm. Thrust and specific impulse are used as metrics to assess nozzle performance. Results from 3D simulation are compared to 2D simulation for the same planar geometry and 3D results for a 30◦ linear-walled nozzle of the same throat size. For all Reynolds numbers examined, 3D nozzles were found to under-perform the 2D case. The margin between the 3D and 2D cases decreases with increasing nozzle depth. The 3D plug nozzle outperforms the 3D 30◦ linear-walled nozzle in both 2D and 3D simulation for the range of Reynolds numbers examined.

The development of micropropulsion systems for small satellites is an active area of research within the aerospace community. One of the most promising technologies is the development of monopropellant-based systems, such as hydrogen peroxide. Hydrogen peroxide is desirable for its high specific energy, modest specific impulse and “green” characteristics. A challenge for the development of MEMS-based hydrogen peroxide microthrusters is achieving complete decomposition over suitable length scales. In this work the potential for enhancing homogeneous catalysis through the introduction of an inert gas is investigated. By controlling the flow rate of the inert gas, an array of periodic monodisperse slugs containing an aqueous catalyst and the monopropellant are formed. As they travel downstream, the internal recirculation of the slugs improves mixing and decomposition. This study is a numerical investigation of this enhanced homogeneous catalysis. Simulations show that operating in the slug flow regime can increase the temperature rise by 888% when compared with lamellar flow, which would allow for complete decomposition over the intended length scales.

An evolutionary computing approach, based on Differential Evolution (DE), is explored as a means for solving a class of multi-satellite formation trajectory problems having topological constraints defined for a region of interest. The orbital mission constraints of NASA’s Magnetospheric Multi-Scale Mission (MMS) is used as motivation and a reference case upon which model test cases are developed for the DE-based trajectory optimization. A satellite formation (N = 4) is centered on a highly eccentric initial reference orbit and a single impulse for each satellite in the formation – deployed at specified true anomaly and with a specified configuration – is sought such that the formation maintains a tetrahedral structure within ±20◦ of apogee. While impulses were found that allowed for quality to be maximized within a bounded formation size, none of the trials were able to meet the full ±20◦ anomaly window; this is likely due to the highly-constraining single impulse limitation within this study, compared to the actual MMS mission design. Numerical tests were then extended to determine the largest ROI, centered about apogee, at which the tetrahedral quality and side-length goals were met. A select set of results was extended into multiple orbits about Earth so as to demonstrate the ability to account for long-term perturbation effects arising from Earth oblateness and atmospheric drag. While these cases did not completely satisfy mission requirements, the extension of computational framework to handle both multiple revolutions – and a ROI window within a given revolution – was successfully demonstrated.

A parametric, two-dimensional, computational study examining steady-state plug micronozzle performance has been conducted. As part of the study, a new method for plug contour construction is proposed. The performance of several different nozzle geometries is compared to that of a traditional plug nozzle geometry designed using the Method of Characteristics (MOC). New nozzle designs are derived from the MOC based design and geometric transformations are used to produce plug nozzles of reduced length. Spike lengths corresponding to 60, 50, 40, and 27% of the MOC nozzle’s length are examined. The throat Reynolds number is varied from 80–820. Thrust is used a metric to assess nozzle performance. The geometry which maximizes performance is found to vary with Reynolds number. It is observed that reducing the plug length improves thrust production for the range of Reynolds number examined.Copyright © 2014 by ASME

In this paper we use Differential Evolution (DE), with best-evolved results refined using a Nelder-Mead optimization, to solve complex problems in orbital mechanics relevant to low Earth orbits (LEO) and within the Earth-Moon system. A class of Lambert problems is examined to evaluate the performance and robustness of this evolutionary approach to orbit optimization. We evolve impulsive initial velocity vectors giving rise to intercept trajectories that take a spacecraft from given initial positions to specified target positions. We seek to minimize final positional error subject to time-of-flight and/or energy (fuel) constraints. We first validate that the method can recover known analytical solutions obtainable with the assumption of Keplerian motion. We then apply the method to more complex and realistic non-Keplerian problems incorporating trajectory perturbations arising in LEO due to the Earth’s oblateness and rarefied atmospheric drag. Finally, a rendezvous trajectory from LEO to the L4 Lagrange point is computed. The viable trajectories obtained for these challenging problems suggest the robustness of our computational approach for real-world orbital trajectory design in LEO situations where no analytical solution exists.

A host of manufacturing methods and mechanical processes depend on e cient convective mass and/or heat transfer. The use of ‘bound vortex’ impingement is shown to provide intense, localized, and well controlled heat and mass transfer enhancement. The bound vortex concept has been previously investigated by Vachon and Hitt as a means for enhancing surface shear stress for particulate dust removal applications. This study investigates the influence of bound vortex flow on the mass and heat transfer from an impinged surface over a range of flow conditions. Naphthalene sublimations techniques are employed to capture local mass transfer distributions and the Colburn analogy is used to interpret the results in terms of the local Nusselt number. Experimental flow conditions are recreated computationally and compare favorably with the experimental data. Convective mass and heat transport characteristics show that the bound vortex flow structure is capable of providing intense, localized, and well controlled transport enhancement. When compared to traditional a swirling flow impingement, an increase of 66% in peak Nusselt number and a 355% gain in average Nusselt value was observed for the optimal bound vortex condition.

Laminar flow mixing remains an active area of research within the microfluidics community. Traditional mixing methods often rely upon turbulent flow, which is generally not present on the micro-scale and so alternative approaches must be sought. This work studies enhanced laminar mixing for use in a proposed monopropellant microthruster based upon homogeneous catalysis in a flow with Re < 10. The enhancement is realized through the introduction of an inert gas at a channel junction, which can lead to the formation of discrete liquid slugs. These slugs contain the monopropellant and the catalyst and have an internal recirculation that is found to enhance mixing. The focus of this study is on the numerical investigation of this process with the goal of minimizing the mixing length and characterizing the dependence of mixing on inlet conditions. The slug formation process is found to decrease the minimum mixing length by a factor of up to 7.2, with much of the benefit of the multiphase flow occurring shortly after slug formation. As minimizing the dimensions of the microthruster is a key design consideration, this reduction in mixing length demonstrates the value of the enhanced laminar mixing for the proposed micropropulsion application.

We describe a simple new method for computing two-dimensional velocity vector fields of blood flows in microvascular networks. This method, known as optical flow, requires a time sequence of two or more images obtained by a process such as conventional videomicroscopy. There is no need to distinguish individual cells provided clear variations in the light intensity levels are present in the images. The result of the computation reveals the velocity distribution in the microvascular network and the geometry of the blood vessels. Although various implementations of the optical flow technique exist, we have found that the spatial correlation algorithm of Anandan performs best for microcirculatory flows. This report is an introduction to the application of optical flow to the microcirculation and provides instructions on how to obtain and run the codes.

Dust mitigation in Martian and Lunar environments is an increasingly active area of research within the aerospace community. Martian and Lunar environments produce electrically charged particles, which easily adhere to exposed surfaces. The work presented investigates the in uence of pulsed ow on a dust mitigation design reported by Vachon and Hitt in recent numerical and experimental studies. A computational study is performed to characterize a bound vortex ow condition over a range of inlet ow frequencies. Transient distributions of shear stress on the impinged surface are used to evaluate the potential of particle removal. Peak and average surface shear stress are used to determine possible shear stress enhancement and isosurfaces of vorticity are used to visualize temporal uctuations of the vortex core. A set of operating conditions are recreated experimentally to visualize particle removal and support computational results.

Gas and liquid converging at a microchannel cross junction results in the formation of periodic, dispersed microslugs. This microslug formation phenomena has been proposed as the basis for a fuel injection system in a novel, discrete monopropellant microthruster design for use in next-generation miniaturized satellites. Experimental work by McCabe et al. 1 demonstrated the ability to generate fuel slugs with characteristics commensurate with the intended application. In this study, a computational model of the microchannel junction is used to investigate the e↵ects of channel depth on the formation process and characteristics of fuel slugs. The simulations show the relative slug volume and detachment location increase with channel depth. Over the range of depths studied, the relative slug volume increased by up to 20% and the detachment location increased by 10 channel widths. These results demonstrate the impact of 3D e↵ects on the ability of the system to throttle the fuel flow rate to a level necessary for low thrust applications, which will have ramifications on the design and manufacture of the microthruster system.

at developing a new MEMS-based micro-reactor conguration for incorporation into a monopropellant micropropulsion system. Two-dimensional numerical models of the gas phase catalytic decomposition in microchannel congurations have been developed and used to obtain critical sizing requirements for the reactor design. The computational results show that the length scales necessary for complete decomposition are compatible with MEMSbased designs; however, it is also found that the results are highly sensitive to the mass diusion characteristics within the ow at this scale. Experimentally, a micro-scale catalytic reactor has been designed and microfabricated using MEMS techniques. The reactor uses self-assembled ruthenium oxide nanorods grown on the wall surfaces as a catalyst. It is found during experimental testing that only partial decomposition of the hydrogen peroxide occurs. The combination of multiphase and thermal quenching eects in a micro-scale geometry are identied as likely obstacles to the process.

Ruthenium oxide nanorods have been grown on Si wafer substrates under a variety of pre-existing surface conditions by reactive radio frequency sputtering in an electron cyclotron resonant plasma process. Nanorod formation by this method is fast relative to that observed in other processes reported in the literature, with nucleation being the rate determining step. Growth in the axial direction is limited by the availability of ruthenium precursors which competes with their consumption in the lateral growth of the nanorods. The availability of Ru precursors at the top of the nanorods can be controlled by surface diffusion and therefore substrate temperature. The ultimate length of the nanorods is determined by the mole fraction of oxygen used in the reactor ambient through the production of mobile Ru hyperoxide precursors. The results of this investigation show the way to develop a process for producing a high density field of nanorods with a specified length.