Influence of Trace Nitrogen Oxides on Natural Gas Oxidation: Flow Reactor Measurements and Kinetic Modeling

The reactivity-promoting effect of trace nitrogen oxides (NO x ) on post-induction oxidation of a synthetic natural gas (2% ethane in methane) has been experimentally studied in a high-pressure laminar flow reactor (HPLFR) at 10 ± 0.1 atm, nominal reaction temperature of 818 ± 5 K, and several equivalence ratios (φ ∼ 0.5, 1.0, and 2.0). Each set of experimental measurements was simulated using several literature C0–C2 + NO x kinetic models, both recent and legacy, using approaches shown to lead to robust interpretation of present experimental conditions. Coupling between the NO x and C0–C2 submodel components of these models varies significantly in both qualitative (mechanistic) and quantitative character. A comparison among the experimental measurements and modeling results serves to highlight important kinetic features particular to application-relevant natural gas oxidation in presence of trace (∼25 ppm) NO x . Additional insight is offered by a baseline experiment with no NO x perturbation, which shows that synthetic natural gas exhibits only incipient reactivity under the present φ ∼ 1.0 experimental condition. A comparison across experimental measurements and simulation results suggests that the reaction CH3 + NO2 ↔ CH3O + NO, often cited as among the most important for NO x –natural gas coupling, insufficiently describes the principal net flux of NO x species at the relatively high pressures and low temperatures examined by present experiments. Simulation results indicate that accurate kinetics related to CH3O2 are necessary to describe a portion of NO ↔ NO2 cycling driven by fuel fragment chemistry. Modeling suggests that the formation of nitromethane (CH3NO2) from the relatively large and long-lived CH3 pool removes NO x from the pool of reactive intermediates, thus altering the reactivity initially imparted by trace NO x addition and the total pool of N atoms available as free NO x (NO + NO2). Frequently used kinetic models that lack (accurate) CH3O2- and CH3NO2-related submodels predict trends in overall reactivity and NO x mole fractions that vary from quantitatively distorted to qualitatively incorrect. These disparities have significant implications for combustor design/evaluation computations that rely on several present literature kinetic models, particularly in a “single digit” parts per million of NO x regulatory environment.