| OCR Text |
Show N O is added into the flame, sometimes in significant quantity (0.1% mole fraction) after dilution in nitrogen. The burner is translated vertically to produce profiles of the radical species as a function of height above the burner surface. The laser beam propagates parallel to the burner surface and the fluorescence is collected at right angles to the laser, these optical paths remain fixed during the burner movement. The excitation and detection wavelengths for each radical are different, chosen from previous spectroscopic studies in our laboratory. CH, NO, AND PROMPT-NO FORMATION CH is crucial to prompt-NO formation, the main NO formation mechanism at temperatures below 1800K.8 Its reaction with air N 2 forms N atoms and H C N radicals; the N atoms promptly form N O by reaction with O 2 and O H , and the H C N is oxidized in subsequent steps to form another N atom. The reaction is highly temperature dependent, and as w e shall see, much more C H radicals are formed in rich flame regions. The concentrations of C H in these flames, near 10 ppm, is too small for measurements using standard absorption measurements and absolute LIF measurements are required to test the flame chemistry. These findings constitute the basis for control strategies on prompt N O formation in practical systems. As noted above, previous experiments in our laboratory established the difficulties in early GRI-Mech predictions of the concentration of the C H radical. These measurements were made in near stoichiometric and slightly rich propane/air (c|) = 1.00 and 1.15) and methane/air (<{> = 1.07) flames,9 and the predictions made using GRI-Mech together with the Premix flame code. 10 The propane flame forms a useful test for the mechanism of C H chemistry in natural gas flames because the propane is long converted into other hydrocarbon radicals before the C H appears. W e found that the then current version of GRI-Mech, with new values of the C H + O 2 rate coefficient, gave excellent predictions in the near stoichiometric flames but predicted higher values in the rich flames (although just within error bars). Because more C H is present in rich flames, a more comprehensive study of trends was necessary. W e first used the model to calculate C H concentrations in a variety of low pressure methane/N2/02 flames, to choose mixing ratios that would provide a sensitive test of the mechanism. W e selected five flames: "Standard", <)) = 1.07; "Lean", <j) = 0.80; and "Rich", ty = 1.27; plus two more rich flames with the same $ but increased and decreased flows of nitrogen. CH is determined using the A-X electronic system near 431 nm. For CH measurements, the quantum yield <X> varies with position in the flame, because it is controlled by collisional quenching with major species flame gases, whose composition varies with burner height. Temperature, because it determines the density in our constant pressure flame, and because the quenching rate coefficient varies with collision energy, also is important. W e have made direct measurements of quenching using time-dependent LIF throughout these flames, H showing there exists inadequate knowledge of high temperature quenching of C H by water vapor. Despite this w e can directly use the empirical results in Eq. 1. The other quantities in Eq. 1 needed for absolute measurements are the factors (Q/4n)er\V. For this radical, w e determine this grouped set of factors using Rayleigh scattering in room temperature N 2 introduced into the burner chamber at a variety of pressures below 100 Torr. All C H and Rayleigh measurements are made over a series of laser pulse energies II, to ensure a linear response of signal to laser power. Plots of this response and further details of this procedure are given in Refs. 5 and 12. A detailed error analysis, using replicate measurements and evaluating uncertainties in all the quantities of Eq. 1, is given in Ref. 13 and indicates errors in absolute concentrations between 15 and 2 0 %. The results are shown in Fig. 1. The solid line shows measured values and the dashed line the predictions from the most recent version of GRI-Mech (2.11) and the Premix code. (The heavy line is a temperature profile included here for reference and whose determination is described below.) The peak of the C H concentration has a value of 10.5 p p m in the Standard flame which is in excellent agreement with the previous, totally independent determination of 4 |
