Files in this item

FilesDescriptionFormat

application/pdf

application/pdfSWAMINATHAN-GOPALAN-DISSERTATION-2019.pdf (6MB)
(no description provided)PDF

Description

Title:Development of physical models for the mesoscopic simulation of gas-surface interactions
Author(s):Swaminathan-Gopalan, Krishnan
Director of Research:Stephani, Kelly
Doctoral Committee Chair(s):Stephani, Kelly
Doctoral Committee Member(s):Ertekin, Elif; Glumac, Nick; Flaherty, David; Mansour, Nagi
Department / Program:Mechanical Sci & Engineering
Discipline:Mechanical Engineering
Degree Granting Institution:University of Illinois at Urbana-Champaign
Degree:Ph.D.
Genre:Dissertation
Subject(s):Gas-Surface Interactions
Surface Chemistry
Carbon oxidation
Abstract:This work is focused on the development of physically consistent models for the mesoscopic and macroscopic simulation of gas-surface interactions relevant for hypersonics and high-temperature aerothermodynamic applications. Both nonreactive and reactive interactions are considered with a special focus on desorption. The aim of the work is to employ microscopic information in the form of detailed experiments, numerical simulations, and fundamental theories, as a basis to construct general, accurate and physically realistic models for the interaction of oxygen: atomic (reactive) and molecular (non-reactive) with carbon surfaces: flat (vitreous) and complex porous microstructure (FiberForm®) at high temperatures ranging from 500 K to 2000 K. These models may be employed directly in conventional computational fluid dynamics (CFD), kinetic simulation, and material response tools for the study of non-equilibrium gas-surface interactions. A detailed finite-rate surface chemistry model for the interaction of oxygen with vitreous carbon (VC) surface is developed from molecular beam experimental data using direct simulation Monte Carlo (DSMC). First, a generalized finite-rate surface chemistry framework incorporating a comprehensive list of reaction mechanisms is developed and implemented into the DSMC solver. The various mechanisms include adsorption, desorption, Eley-Rideal (ER), and several types of Langmuir-Hinshelwood (LH) mechanisms. Both gas-surface (e.g., adsorption, ER) and pure-surface (e.g., desorption) reaction mechanisms are incorporated, and the framework also includes catalytic or surface altering mechanisms involving the participation of the bulk-phase species (e.g., bulk carbon atoms). Expressions for the microscopic parameters of reaction probabilities (for gas-surface reactions) and frequencies (for pure-surface reactions) that are required for DSMC are derived from the surface properties and macroscopic parameters such as rate constants, sticking coefficients, etc. This framework is used to numerically simulate the hyperthermal pulsed beam surface scattering experiments. Next, a general methodology for constructing finite rate surface chemistry models using time-of flight (TOF) and angular distribution data obtained from pulsed hyperthermal beam experiments is presented. A detailed study is performed to analyze the TOF distributions corresponding to the various reaction mechanisms at diverse conditions using the DSMC surface chemistry framework. This information is used to identify and isolate the products formed through different reaction mechanisms from the molecular beam experimental data of oxygen on vitreous carbon. A general methodology to derive the reaction rate constants which takes into account the pulsed nature of the beam is described and used to derive the rates within the vitreous carbon oxidation model. The constructed finite rate surface chemistry model provides excellent agreement with the experimental TOF and angular distribution as well as the total product fluxes. As a next step, the derived vitreous carbon oxidation model is extended to FiberForm®, which is used as a precursor of NASA’s TPS material Phenolic Impregnated Carbon Ablator (PICA). The purpose of this study is to investigate the reactive interaction of fibrous carbon with atomic oxygen in a complex microstructure, which is the primary source of carbon removal at lower temperatures. The detailed microstructure of FiberForm® obtained from X-ray micro-tomography is used in the porous microstructure analysis (PuMA) simulations to capture the complexity of the porous and fibrous characteristic of FiberForm®. Comparison between the experimental and PuMA time-of-flight (TOF) distributions are presented for both the reactive interaction of the oxygen beam and the nonreactive interaction of the argon beam. It was also found that a significantly higher amount of CO (up to 30% of the total product flux) is generated when the beam interacted with FiberForm®, when compared with vitreous carbon. This is postulated to be primarily a result of multiple collisions of oxygen with the fibers, resulting in an higher effective rate of CO production. Multiple collisions are also found to thermalize the O atoms, in addition to the adsorption/desorption process. The effect of microstructure is concluded to be crucial in determining the final composition and energy distributions of the products. Thus, an effective model for the oxygen interaction with FiberForm®, fully accounting for the detailed microstructure, for use in Computational Fluid Dynamics (CFD) and material response codes, is presented. In order to construct the effective surface chemistry model for FiberForm®, the VC model was applied to the detailed microstructure of FiberForm® to obtain the product fluxes at various porosities. At higher porosities, higher mole fractions of CO and lower amounts of O (up to 10% of the total product flux) were observed. This is due to the greater penetration of the incoming beam atoms into the microstructure leading to more collisions with the surface, resulting in the higher mole fraction of CO. This effect is more pronounced at higher temperatures when the probability of CO formation during a single collision is smaller. The effective model reaction mechanisms are assumed to be the same as that of the VC model, as well as the desorption rate constant values. Simulations performed using the constructed effective rates with a flat plate provided excellent agreement with the experimental TOF and angular distributions, and with the analyzed experimental fluxes. This effective model also provides excellent agreement with the PuMA data for the entire porosity range of interest. In order to study the non-reactive (inelastic) scattering process, Molecular dynamics /Quasi-classical trajectory (MD-QCT) simulations and molecule-surface scattering (MSS) theory are used. The system of interest in this work is the gas-surface interactions of O2 molecules striking a carbon surface. MD-QCT technique uses quasi-classical methods to represent the internal energies of the systems within MD and thus can be used to model accurate post-reaction and post-collision molecular internal energy distributions. The MSS theory employs a theoretical framework which accounts for several mechanisms of energy transfer between the substrate and gas particles including multi-phonon excitations, and translational, rotational and vibrational energy transfer. This framework is quasi-classical and employs classical treatment of translational and rotational modes while the vibrational mode is considered quantum-mechanically. A range of initial translational energies of the molecule and the surface temperature is considered to elucidate the dependence of the scattered molecule properties on these parameters. The quantities of interest in this work are the final energy (translational, rotational and vibrational) and polar angular distributions. The values of the fitted MSS model parameters are presented along with their variation with the initial molecule translational energy and the surface temperature, along with the physical significance of their variation. Finally, the desorption of O/CO from graphitic carbon surfaces is investigated using a one-dimensional model describing the adsorbate interactions with the surface phonon bath. The kinetics of desorption are described through the solution of a master equation for the time-dependent population of the adsorbate in an oscillator state, which is modified through thermal fluctuations at the surface. The interaction of the adsorbate with the surface phonons is explicitly captured by using the computed phonon density of states (PDOS) of the surface. The coupling of the adsorbate with the phonon bath results in the transition of the adsorbate up and down a vibrational ladder. The adsorbate-surface interaction is represented in the model using a Morse potential, which allows for the desorption process to be directly modeled as a transition from bound to free (continuum) state. The PDOS is an important input within the phonon-induced desorption (PID) model, which is a property of the material and the lattice and is highly sensitive to the presence of defects. The effect of random surface defects, etch pits, and adsorbates on the PDOS is considered in the present work. The presence of defects causes a redshift and broadening of the PDOS, which in turn changes the phonon frequency modes available for adsorbate coupling at the surface. This PDOS including defects is used within the PID model to predict the desorption rate constant. Using the realistic PDOS distributions, the PID model was used to compute the transition and desorption rates for both pristine and defective systems. Mathissen’s rule is used to compute the phonon relaxation time for pristine and defective systems based on the phonon scattering times for each of the different scattering processes. First, the desorption rates of the pristine system is fitted against the experimental values to obtain the Morse potential parameters for each of the observed adatoms. These Morse potential parameters are used along with the defective PDOS and phonon relaxation time to compute the desorption rates for the defective system. The defective system rates (both transition and desorption) were consistently lower in comparison with the pristine system. The difference between the transition rates is more significant at lower initial states due to higher energy spacing between the levels. In the case of the desorption rates, the difference between the defective and pristine system is more significant at higher temperatures. The desorption rates for each of the system shows an order of magnitude decrease with the strongly bound systems exhibiting the greatest reduction in the desorption rates.
Issue Date:2019-09-25
Type:Text
URI:http://hdl.handle.net/2142/106158
Rights Information:Copyright 2019 Krishnan Swaminathan Gopalan
Date Available in IDEALS:2020-03-02
Date Deposited:2019-12


This item appears in the following Collection(s)

Item Statistics