OCR Text |
Show A direct means of identifying the regions of high N O formation rates would be a valuable tool in the combustor design and evaluation process for minimizing N 0 X emissions, and for optimization of individual burners or combustors. PLIF measurements of N O can identify regions in a combustor or flame having high concentrations of N O , though these measurements cannot directly indicate where the N O is being produced. O n the other hand, w e have identified a technique that provides a direct visualization of the regions in a combustor or a flame having high N O production rates, and does not require laser excitation. This technique is based on a chemiluminescent process that closely correlates with the rate-limiting step in the Zel'dovich N O formation mechanism. The familiar high intensity "green" chemiluminescent emission from boron combustion is produced by the reaction,1: O + BO + M - B02' + M. This chemiluminescence reaction is analogous, because of its reaction with the oxygen atom radical, to the rate-limiting step in the Zel'dovich mechanism for thermal N O production, N2 + O - NO + N. The chemiluminescence reaction thus is a tracer for both O atoms and, as the experimental and modeling results show, thermal N O formation. Although the chemiluminescence reaction is a three-body reaction with a low activation energy (Ea = 2000 cal/mole) while the O atom reaction in the Zel'dovich mechanism is a two-body reaction with a high activation energy (Ea = 76,500 cal/mole), the close correlation obtained indicates that the O atom being the c o m m o n reactant dominates these other differences. Experimental A turbulent jet diffusion flame apparatus, similar to that used by Driscoll34 in NOx studies of turbulent flames, was used to investigate the correlation between boron chemiluminescence intensity and N O production in hydrogen and methane turbulent jet flames. This apparatus, shown in Figure 1, included a coaxial air nozzle (0.87 c m diameter) within which the fuel tube was located. Coaxial air was used in selected tests to increase the mixing rate, duplicating the tests of Driscoll.4 The main difference between the Driscoll apparatus and our apparatus was that we used a square cross section enclosure for the flame to allow better viewing, while Driscoll used a circular cross section enclosure. The flow measurement and metering devices were Gilmont® Accucal™ flowmeters having an accuracy of 2 % of the reading or one scale division. The fuel and boron seed compound were measured and metered separately and then mixed. Hydrogen (Grade 4, 99.997c) and methane (Grade 2, 9 9 % ) were the fuels used in the tests. Three seed compounds were used in the tests. T w o diborane (B2H6) mixtures, 5% B2H6 in N 2 and 5% B2H6 in H2, were obtained from Voltaix, Inc. In addition, trimethyl borate, (CH?0)3B, was used. The diborane mixtures were gases and could be metered directly. Trimethyl borate, on the other hand, is a liquid (b.p.=68°C). It was added to the fuel flow by diverting a controlled portion of the fuel flow 2 |