Turbulent flows and Combustion: Structural Analysis of Turbulent Flames by Means of Formaldehyde
Structural Analysis of Turbulent Flames by Means of Formaldehyde LIF
Combustion of natural gas still plays an important role as an energy source in practical combustion processes, and it most likely will in the coming years. Natural gas is characterized by its high calorific value and a low pollutant output. The interest in this energy source initiated many efforts to further optimize common combustion techniques and to minimize the pollutant output. Therefore, various experiments were implemented with the aim to characterize the flames. The objective is to understand combustion processes in their microscopic details and their localization in the flame. In diagnostics of laminar and turbulent flames the measurements have so far been based on the simultaneous detection of temperature, nitric oxide (NO) and hydroxyl radicals (OH). OH-LIF has been used frequently for the location of flame fronts. OH is formed in the reactive zone, but is present in the hot burned gases as well. Therefore, the region with the steepest gradients in OH concentration is often used for flame front location. For highly resolved measurements however, this approach is questionable. Paul et al. [1] suggested the use of the formyl radical (HCO) as a flamefront marker instead. In simulations based on determined chemistry models, HCO has proved to be one of the species best correlated to the flame front. Due to its high dissociation rate HCO is only found in very low concentrations and thus is not applicable to turbulent flames. But as shown by Paul et al. [1] one can avoid this problem by the detection of HCO via the product of the two-dimensional LIF intensity distributions of CH2O and OH which are the direct precursors of HCO:
CH2O + OH → HCO + H2O
For the measurements with this procedure an experimental setup was installed which allowed the simulaneous detection of two-dimensional images of temperature, CH2O-LIF und OH-LIF distributions in turbulent flames. The non-intrusive detection method also allowed measurements avoiding disturbance of the flow conditions and the turbulence.
Figure 1: Emission spectrum of H2CO after excitation at 353.2 nm
In preliminary examinations in a bunsen burner flame a spectrum was measured to identify the desired transition for 2D measurements at 353.2 nm which can be excited by means of a tuneable XeF excimer laser (Fig. 1). See also the following paragraph: CH2O spectroscopy in detail.
Our measurements included simultaneous instantaneous measurements in a Bunsen flame as well as the corresponding measurements in a standardized TECFLAM swirl burner. Fig. 2 shows some results of the measurements in a Bunsen flame. Fig. 3 shows the area of detection in the flame.
 |
 |
| Figure 2: Simultaneous instantaneous measurements in a Bunsen burner |
Figure 3: Area of detection in the flame |
Temperature distribution (by Rayleigh thermometry), CH2O-LIF distribution, OH-LIF distribution and the product of both LIF signals (CH2O and OH).
The closely confined occurrence of CH2O at the border of the freshgas zone (center of image) caused by the very small lifetime of the CH2O molecules can be clearly seen. The significantly larger lifetime of the OH radical in contrast causes its well visible spatially broad distribution. In comparison with the temperature measurement it can be seen that CH2O is formed early in the combustion process (proximity to freshgas), whereas OH is primarily found in the hottest regions of the flame. The last image of Fig. 2 shows the overlapping region of the two species. In this area the formation of HCO is possible according to the above reaction equation. Consequently, this is the region in which the flame front and the peak heat release are localized.
Formaldehyde (CH2O) is one out of a few spectroscopically well-characterized molecules which appear in the combustion process [2]. The CH2O molecule has six normal vibrations, as shown below (Fig. 4):
Very convenient for LIF measurements is the 
vibrational band, which is spread from 352 to 357 nm. Its band head is located at 353.20 nm and the excitation can be achieved by a narrow-band XeF excimer laser. The emissions of the XeF excimer laser occurs in ranges at about 348.8 nm, 351.1 nm and 353.2 nm. The emission at 353.2 nm overlaps with the long wave end of the H2CO absorption spectrum, in which the vibrational band is located.
Due to small absorption cross sections, at 348.8 nm neither formaldehyde nor any other combustion relevant species are absorbed. Weak excitation of CH2O must be expected by the laser emission at 351.1 nm caused by ASE (amplified spontaneous emission) of the laser, which amounts up to 40% of the total laser emisson. A selected transition with an overlap in this wavelength range, e.g. is the vibrational transition 
at 351.67 nm.
In Fig. 1 (see above) an emission spectrum of CH2O after excitation of the transition at 353.2 nm is shown. Spectroscopic measurements of the 
transition have shown that the CH2O fluorescence intensity at high emittances increases linearily whereas the fluorescence of polycyclic aromatic hydrocarbons, also absorbing in this range, shows early saturation.
Further advantages in the determination are: There is no absorption of the CH2O fluorescence by the flame and no photolytical formation of formaldehyde is possible.
References:
[1] Paul, P.H., Najm, H.N., Planar Laser-induced Fluorescence Imaging of Flame Heat Release Rate, Proc. Combust. Inst. 27, 43-50 (1998).
[2] Harrington, J.E., Smyth, K.S., Laser-induced fluorescence measurements of formaldehyde in a methane/air diffusion flame, Chem. Phys. Lett. 202, 196-202 (1993).
[3] S. Böckle, J. Kazenwadel, T. Kunzelmann, D.-I. Shin, C. Schulz, and J. Wolfrum, "Simultaneous single-shot laser-based imaging of formaldehyde, OH and temperature in turbulent flames," Proc. Combust. Inst. 28, 279-286 (2000).
[4] S. Böckle, J. Kazenwadel, T. Kunzelmann, D.-I. Shin, and C. Schulz, "Single-shot laser-induced fluorescence imaging of formaldehyde with XeF excimer excitation," Appl. Phys. B 70, 733-735 (2000).
[5] K.-U. Schildmacher, A. Hoffmann, L. Selle, R. Koch, C. Schulz, H.-J. Bauer, and T. Poinsot, "Unsteady flame and flow-field interaction of a premixed model gas turbine burner," Proc. Combust. Inst. 31, in press (2006).