Nitric oxide concentration and fluorescence lifetime in non-premixed atmospheric pressure flames.

Nitric oxide (NO) is an important combustion pollutant, and understanding NO formation and destruction chemistry requires accurate concentration measurements in geometrically simple flames. The increasing importance of combustion in non-premixed and partially premixed modes for power generation and transportation places additional emphasis on experiments under non-premixed conditions. Nitric oxide concentration profiles along the stagnation streamline were measured with laser induced fluorescence (LIF) in non-premixed, atmospheric-pressure counterflow flames stabilized under a semi-cylindrical burner (Tsuji-type). Concentration profiles were measured in pure methane/air flames and flames with NO seeded into the fuel flow at 100- to 600-ppm levels. Several air flow-rates were examined to study the effect of flame stretch. A picosecond-pulse tunable dye laser and microchannel plate photomultiplier tube were used to provide time-resolved LIF signal under linear (non-saturated) excitation, and fluorescence lifetimes and quenching rates were determined by de-convolving the temporal fluorescence signals. Coherent anti-Stokes Raman spectroscopy was used to measure temperature profiles in the flames; this information was used to extract quantitative species concentration data from the LIF measurements and provide a boundary condition for kinetic modeling. Measured peak NO concentrations were between 84 +/- 12-ppm (low flame stretch) and 50 +/- 8-ppm (high stretch). To evaluate the kinetic mechanisms, the flames were modeled computationally and NO concentrations were calculated. Absolute concentration predictions in seeded and unseeded flames were accurate to within the error imposed by uncertainty in the prompt NO production pathway rate (CH + N2 → HCN + N). However, the models over-predicted the decline in NO concentration with increasing air flow velocity. Calculated flame species concentrations were coupled with model estimations for collisional quenching cross sections, and these predicted fluorescence lifetimes were compared to directly measured LIF lifetimes. The predictions were accurate within experimental error in the fuel-lean and high-temperature regions of the flame. However, the model error in the fuel-rich region of the flame exceeded uncertainty limits; erroneously low model quenching cross sections for H2O and CO2 could explain the discrepancy, but this possibility has not been fully explored.PhDApplied SciencesMechanical engineeringUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/131766/2/3057940.pd