Stochastic Modeling of the Formation of Aromatics in Combustion

Abstract

Understanding the formation of polycyclic aromatic compounds (PACs) in combustion not only bridges the knowledge gap between the small gas-phase species and incipient soot particles, but may also help address the global emission issues of both PACs and soot. In this thesis, I present a kinetic mechanism utilizing reactive sites (i.e., the chemical and physical neighbourhoods) to describe the PAC growth in combustion. This kinetic mechanism was implemented for a stochastic modeling code (i.e., SNapS2) recently developed by the Violi Group. To address new experimental and computational discoveries, chemical reactions were gathered and categorized from various literature, while the reaction rate constants came from either literature or my own calculations to ensure full reversibility and thermodynamic consistency. These reactions were then implemented into SNapS2 kinetic mechanism with precise reactive site definitions to eliminate the possibility of steric hindrance and unrealistic reactions. Compared with the previous version of SNapS2, thanks to this new kinetic mechanism, the computational performance increased by an order of magnitude, enabling the simulation of complex two-dimensional flames. Some missing reaction pathways, which were identified from experimental evidence and simulations but not available in the literature, were explored and calculated using quantum chemistry methods. These newly discovered reactions were included in the SNapS2 kinetic mechanism as well, and some of them were already proven to be important under specific conditions. The characteristics of the PACs predicted with the kinetic mechanism were compared against different experimental measurements: mass spectra measured in a counterflow diffusion flame, the oxygen-to-carbon ratios obtained at different locations of a coflow diffusion flame, and the molecular structures observed in a premixed laminar flame. These successful validations demonstrate that the SNapS2 kinetic mechanism provides a high-fidelity, and yet generic, description of the PAC formation under various combustion conditions, making SNapS2 the first-of-its-kind to have such extensive flexibility and wealth of information. It greatly contributes to reveal the underlying chemical pathways to the experimental observations. Furthermore, SNapS2 code and the kinetic mechanism have shown its capability to provide valuable insights on the formation of aromatics beyond the limitation of diagnostics. For one application, spatial dependence of the PAC growth in an ethylene counterflow diffusion flame was characterized by SNapS2 simulations, revealing distinct PAC growth pathways for the streamlines starting from fuel side and oxidizer side. Given the fidelity of the SNapS2 predictions, it was also used to examine conditions that are impossible to test experimentally, like completely decoupling the effects of flame temperature when studying the effects of ethanol doping on the formation of aromatics, highlighting the chemical pathways that result in soot reduction. Both applications show the uniqueness and great potential of the model to obtain insights of the PAC formation when measurements are hard to obtain or experiments are difficult to control. Altogether, this dissertation lays a solid foundation that not only helps explain the experimental observations for the formation of soot precursors, but also provides a powerful tool for exploring the gas-phase nanoparticle growth that could drive the development of novel combustion technologies or the design of new nanomaterials.