The Chemiluminescence and Structure Properties of Normal / Inverse Diffusion Flames

e �ame emission spectrometry was applied to detect the distribution of excited radicals in two types CH4/O2 co�ow jet diffusion �ames (normal and inverse diffusion �ames). Combining the image analysis along with the spectrometry, the chemiluminescence and structure characteristics of these diffusion �ames were investigated.e results show that the inverse diffusion �ame (IDF) with relatively high inlet oxygen velocity is composed of two regions: a bright base and a tower on top of the base, which is quite different from the normal diffusion �ame (NDF).e�ame is divided into two regions along the �ame axis based onmaximumOH position (Region I: initial reaction zone; Region II: further oxidation zone). e degree of the further oxidization taking place in Region II is obvious in accordance with OH distribution, which is the main difference in reaction zone between fuel-rich condition and fuel-lean condition for NDFs. For IDFs, the change of OH distribution with increasing equivalence O/C ratio ([O/C]ee) in Region II is not conspicuous. More OH and CH are generated in IDFs, due to the inner high-speed O2 �ow promoting the mixing of fuel and oxygen to a certain extent.


Introduction
Most of the practical combustion systems such as coal gasi�ers, gas turbine engines and industrial stoves.employ diffusion combustion because of its better �ame stability, safety, and wide operating range as compared to premixed combustion [1].According to the feeding pattern of the fuel and oxidizer, there are two types of diffusion �ames: normal diffusion �ame (NDF) and inverse diffusion �ame (IDF).e IDF is a special �ame with an inner oxidizer jet surrounded by an outer fuel jet, with less soot produced as compared to NDF [2], so that the application of IDF in industry is becoming more and more widespread.In some processes of coal gasi�cation, the combustion of inner oxygen and fuel from the annulus forms the IDF.In the coke oven gas autothermal reforming technology, the �ame type is a typical IDF [3].
In recent times, there has been a growing interest in researching of IDF and its difference with NDF.e �rst detailed investigation was performed by Wu [4] for laminar methane IDF stabilized in a simple coaxial burner.He identi�ed six different regimes of IDF and found that IDF and NDF had almost the same visible appearance in the con�ned space.e comparative study on hydrogen IDF and NDF performed by Takagi et al. [5] revealed the occurrence of higher �ame tip temperature in IDF than NDF.ey de�ned a parameter called H 2 ratio to explain the occurrence of higher �ame tip temperature in H 2 IDF and the excess enthalpy in the central region of IDF.Kaplan and Kailasanath [6] numerically analyzed the �ow �eld effects of air-fuel jets on soot formation in laminar methane IDF and NDF con�gurations and observed less peak soot volume fraction in IDF as compared to NDF.Mikofski et al. [7] measured the �ame height of laminar IDFs, and predicted it using Roper's analysis for circular port burners.Based on the fact that Roper's analysis could be well applied to IDFs, they suggested that IDFs were similar in structure to NDFs.Sze et al. [8] conducted experiments on LPG-air IDFs stabilized on two different burners (one with circumferentially arranged ports and the other with coaxial jets) and reported the �ame shape, visible �ame length, temperature contour, and centerline oxygen concentration of these two different IDFs.In general, previous investigations were concerned mainly with the differences of soot formation characteristics, �ame temperature, and �ame appearance between NDFs and IDFs.ere have been very few investigations on �ame chemiluminescence characteristic, which is a common approach for characterizing the �ame reaction zones, �ame structures, and process parameters, such as fuel type, equivalence ratio, and strain rate [9][10][11].e interest of this paper is mainly focused on the chemiluminescence distribution characteristics of excited radicals (OH * and CH * ) along the axis of NDFs and IDFs with different equivalence O/C ratios.Furthermore, the detailed differences of �ame reaction zones and �ame structures between NDFs and IDFs are investigated by analyzing the chemiluminescence property.

Experimental Setup
2.1.Combustion System.Figure 1 shows a schematic diagram of the experimental setup, consisting of two main parts, a stainless steel combustion chamber and an optical measurement system.e height of the chamber was 200 mm, with an i.d. of 50 mm.A jet burner was mounted inside the chamber at the bottom.e burner consisted of two coaxial tubular layers: the inside tube had an i.d. of 0.8 mm, and an o.d. of 1.8 mm� the outer tube had an i.d. of 3 mm.e �ow rate was measured with mass �owmeter (Sevenstar, Inc., D07-19B).e chamber was designed to have a side quartz window to permit the optical measurement of the entire �ame by a spectrometer and to record the visible �ame image by a high resolution CCD camera (JAI Inc., BB-500CL).
Combustion was carried out in the chamber at atmospheric pressure.Pure methane (>99.9%) was used as fuel, and the �ow rate of fuel was �ept at 0.50 L/min.Pure oxygen (>99.9%) was used as oxidizer.e equivalence O/C ratio ([O/C]  ) was changed in a range of 0.60-1.20,which is de�ned by the following equation: where  optical bench (Table 2), detecting a certain wavelength range synchronously to achieve a high spectral resolution.� �beroptics probe with 25 ∘ �eld of view (FOV) and 3 mm diameter was directed to the �ame to collect the luminescence signals.e distance from the probe to the �ame was �xed in all measurements to ensure the detection of a constant projected �ame area within the �eld of view of the �ber.e spectrum was an average of 10 spectra obtained repeatedly under the same condition.e dark background spectrum, obtained before ignition or without �ame luminescence emissions for each experiment, was subtracted from the raw spectrum.
e raw intensity data collected by the spectrometer were subject to some errors: (1) losses in the intensity when the light passed through the lens; (2) attenuation losses in the optical �ber; (3) losses caused by the grating efficiency and CCD sensitivity of the spectrometer.Using a standard light source (Deuterium-Halogen lamp, Ocean Optics, Inc.) with known color temperature allowed the calibration of the optical system and reduced the errors from the lens, �ber-optic cable and the diffraction grating.e calibration factors are determined by modifying  mi / pi , calculated by the following equation: where  mi is the measured intensity at a given wavelength   ,  pi is the intensity calculated by Planck's Law of Radiation at   . 1 = 2 2 ,  2 = /, and the constants h, k, and c are Planck's constant, Boltzmann's constant and the speed of light, respectively.Figure 2 is the calibration curve for the UV and VIS ranges.e actual intensity of the object will then be calibrated by dividing the measured intensity at   by the calibration factors from Figure 2. e detailed procedure refers to the publication by Keyvana et al. [12].

Results and Discussion
3.1.Typical Emissions of OH * and CH * .In hydrocarbon �ames, the major excited radical radiations come from OH * and CH * .For OH * , the primary emission occurs at approximately 283 nm, 306 nm, and 309 nm.CH * radiates at about 390 nm and 431 nm.�xcited radicals are formed in �ames by two paths, thermal excitation and chemical excitation.ermal excitation is related to �ame temperature and the number of ground state radicals.Chemical exitation results from chemical reactions as other reactions produced ground state radicals, from which the radiation is called chemiluminescence.e thermal excitation way of OH * becomes more dominant at temperatures above 2800 K [13], thus the probability of thermal excitation way effecting OH * formation is small, which can be excluded.
Using the optical measurement system, the typical OH * and CH * emissions in the ultraviolet and visible regions with different transitions were obtained (the spectrum ranges from 200-600 nm), as shown in Figure 3. e precise radiation data are shown in Table 3.

e Comparison of Appearance between NDFs and IDFs.
Figure 4 shows the visible �ame images recorded by a high resolution CCD camera, and the exposure time is 1/100 s. e IDF is composed of two regions: a relatively bright base and a tower on top of the base, which is different from the NDF.According to the study of Sze et al. [8], at low oxygen jet velocity, the �ame shape is similar to that of a normal diffusion �ame, and at high enough oxygen jet velocity (Re > 2500), the IDF consists of two parts: a base and a tower.In this experiment, the inlet oxygen velocity (Re > 2500) is higher than the fuel velocity, at least 16 times higher.Wu [4] proposed that the shape of IDF is affected by both the inlet air momentum and the inlet fuel momentum.With the increase of oxygen jet velocity, the difference in the momentums of the two jets increases, leading to entrainment of some fuel into the central oxygen �ow, which creates two-zone structure.
When the oxygen velocity of IDF is up to 30 m/s, the �ame is blown out due to the excessive jet velocity.e �ame height of IDF is smaller than NDF, indicating the �ame propagation speed is relatively faster.e carbon particles produced in IDF is close to the fuel side, where the temperature is relatively low, so that the luminescence of unburned carbon particles is weaker than NDF.

e Change of OH * Distribution with Increasing [𝑂𝑂𝑂𝑂𝑂𝑂 𝑒𝑒
in NDFs.OH * distribution can be an indication of �ame reaction zone [14], and where the maximum OH * emission is where the reaction is the most vigorous (�ame-front).From the nozzle exit to the OH * peak position, fuel and oxygen are mixed depending on the action of interdiffusion and reach the stoichiometric proportion at the OH * peak position, where is Region I called initial reaction zone (Figure 5).In an ideal situation, the unburned fuel will be burned out when they reach the �ame-front, so it can be considered that the width of the �ame-front is in�nitely thin.For practical �ame, there are always few unburned fuel passing the �amefront (the OH * peak position), and going on burning in the �ame downstream region, where is the further oxidation zone (Region II).According the OH * peak position, the entire �ame has two regions along the axis: nozzle exit to the OH * peak position (Region I, initial reaction zone), OH * peak to the end of the �ame propagation (Region II, further oxidation zone).With the increase of [O/C]  , the change of OH * distribution in Region II is much more signi�cant than Region I (Figure 6).OH * generated in Region I is basically the same, and only more fuel is consumed along the �ame propagation direction, causing the originally position of the stable �amefront move downstream.At lower [O/C]  , the unburned fuel and carbon particles decomposed by fuel cannot be further oxidized due to the absence of O 2 , which forms the obvious upper luminous zone (yellow) as shown in Figure 4.Under fuel-lean condition, the adequate supply of oxygen provides the opportunity for further oxidation, leading to the increase in OH * generation.e degree of the further oxidization taking place in Region II is conspicuous in accordance with OH * distribution, which is the main difference in reaction zone between fuel-rich and fuel-lean condition for NDF.fuel and oxygen will be increased with higher oxygen speed, causing a stronger shearing action between these two �ows, so that the inner high-speed O 2 �ow promotes the mixing of fuel and oxygen in IDF.Due to this promotion effect, most of the fuel is oxidized in Region I, which may be the reason that the change of OH

Conclusions
e chemiluminescence distribution characteristics of excited radicals (OH * and CH * ) along the axis of NDFs and IDFs with different equivalence O/C ratios have been measured.Furthermore, the detailed differences of reaction zones and structures between NDFs and IDFs are investigated by analyzing the chemiluminescence property combining along with the image analysis.e results can be summarized as follows.
(1) e IDF is composed of two regions: a relatively bright base and a tower on top of the base, on account of the entrainment of some fuel into the central oxygen �ow under relatively higher inlet oxygen velocity condition.
(2) e entire �ame has two regions along the axis separated by the maximum OH * emission.Region I is initial reaction zone, where fuel and oxygen are mixed partially depending on the action of interdiffusion.Region II is further oxidation zone, because the combustion products of Region I, residual oxygen and a small quantity of fuel are further oxidized in this region.According to OH * distribution, the main difference of reaction zone between fuel-rich condition and fuel-lean condition for NDF, is that whether the further oxidization takes place in the Region II. (

F 1 :
e schematic diagram of the experimental setup.

F 2 :
UV and VIS calibration curves.

Figure 7 9 F 7 : 9 [F 8 :
shows the OH * distributions for different [O/C]  of IDFs.e pro�les have changed little with the increase of [O/C]  .e position of maximum OH * is still the boundary point of IDFs, and the upstream region is initial reaction region, downstream region is further oxidization region.Region I corresponds to the �ame relatively bright base, which is only a quarter of the entire �ame.Unlike NDFs, the change of OH * distribution with increasing [O/C]  in Region II is not obvious, and the OH * peak position does not move, indicating the reaction zone of IDF has no change basically for different [O/C]  (Figure 8).At the same [O/C]  ([O/C]  = 0.70), the OH * generated in IDF is signi�cantly more than NDF, and the distributing area is wider, as a result of more fuel and oxygen involved in the reaction.CH * is concentrated near the nozzle exit (less than 30 mm above the nozzle), with relatively weaker radiation comparing with OH * .e production of CH * in IDF is also more than NDF.e comparative velocity of the µW•cm −2 •nm −1 ) Distance from the nozzle (mm) [O/C] e = 0.6 [O/C] e = 0.65 [O/C] e = 0.7 [O/C] e = 0.75 [O/C] e = 0.8 [O/C] e = 0.85 [O/C] e = 0.e change of OH * distribution with increasing [O/C]  for IDFs.intensity (µW•cm −2 •nm −1 ) Distance from the nozzle (mm) [O/C] e = 0.e comparison of OH * , CH * distributions between NDFs and IDFs.

Figure 9
Figure 9 is the variation of maximum OH * emission with different [O/C]  .OH * peak increases regularly in NDF as the [O/C]  increases, because of the enrichment of O and O 2 and increasing temperature.e peak intensities of OH * show a trend from ascent to descent for IDF.e inner high-speed O 2 �ow promotes the mixing of fuel and oxygen certainly under relatively low [O/C]  .e downtrend may be caused by the variation of �ame temperature.According to the research about the temperature of IDF[8], a trend from increasing to decreasing is appeared with increasing equivalence O/C ratio, and the generation of OH * is in�uenced by the temperature to a certain extent.
T 1: Flame conditions a used in different experiment numbers.
a the IDF is blown out when   2 is up to 30 m/s.
[O/C]  is the actual O/C molar ratio calculated from the amount of fuel and oxygen feed, and [O/C]  is the stoichiometric O/C molar ratio.e equivalence O/C ratio was adjusted by changing O 2 �ow rate.Table 1 lists the numbers of experiment and conditions.
2.2.Optical Measurement Method.e optical measurement was carried out using an ocean optics spectrometer, which has the equivalent of four HR2000+ devices combined (the optical bench is a 101.6 mm-focal length symmetrical crossed Czerny-Turner type), and each device has speci�c T 2: Optical bench and settings for each HR2000+.
T 3: Radiation data of the major excited radicals.
distribution with increasing [O/C]  in Region II is not obvious for IDF.Under the relatively high [O/C]  condition ([O/C]  = 0.90), the distributions of OH * and CH * in NDF become wider signi�cantly, indicating that more fuel is oxidized, and the upper yellow zone is reduced correspondingly.But for IDF, both the OH * and CH * radiations decrease relatively compared with lower [O/C]  , leading to the NDF exceeding IDF in the productions of OH * and CH * . * 3) e change of OH * distribution with increasing [O/C]  in Region II is not obvious for IDF.At lower [O/C]  , the OH * generated in IDF is signi�cantly more than that in NDF, and the distributing area is wider, because of the inner high-speed O 2 �ow promoting the mixing of fuel and oxygen.Under the relatively high [O/C]  condition, the NDF exceeds IDF in the productions of OH * and CH * , which may be caused by the variation of IDF temperature.