Recent developments in laser-induced incandescence (LII) for soot diagnostics in high-pressure laminar flames and engine-like Diesel combustion 1 2* M. Hofmann , B. Kock, T. Dreier, C. Schulz 1 BASF AG 67056, Ludwigshafen, Germany 2 Siemens AG 45478, Mühlheim, Germany 3 Institut für Verbrennung und Gasdynamik University of Duisburg-Essen, Duisburg, Germany The analysis of soot formation and oxidation is essential for Diesel engine development to meet future pollutant emis- sion reduction requirements. In recent years laser-induced incandescence (LII) has developed to a powerful tool for soot diagnostics in flames, even in the challenging environment of Diesel combustion. This work presents some re- cent applications of LII for soot particle sizing in high pressure combustion environments including premixed laminar ethylene/air flames, diesel combustion in a constant volume spray combustion chamber and in-cylinder engine com- bustion, with some emphasis on challenges in modeling time-resolved LII signal transients in these environments. Introduction particle sizes have been deduced from modeling For a better understanding of soot formation in the complete temporal LII signal decay, using theo- high pressure combustion situations, such as gas retical models treating heat, mass and radiative turbine and Diesel combustion quantitative, in-situ transfer of the laser-heated particle [10, 11]. A soot particle sizing is a valuable experimental ap- numerical tool for simulating TiRe-LII signals (LII- proach. Measurements, either under well-defined Sim) is available online [12]. conditions (steady laminar flames) or for unsteady, turbulent conditions (spray injections, Diesel en- Experiments and Results gines) create an experimental data base and sti- LAMINAR FLAMES – The knowledge of the in- mulate further development and validation of mod- fluence of environmental gas pressure on the LII- eling codes for soot formation and oxidation in signal is important in order to enable the quantita- these combustion environments. In addition, two- tive allocation of soot volume fraction and particle dimensional optical soot visualization allowed the size from LII measurements [13], and can best be development of conceptual models for Diesel investigated in laminar flames. For this purpose, combustion based on the qualitative understanding the burner used in our studies was installed inside of the processes in the Diesel spray flame [1]. Lat- a high-pressure chamber equipped with four quartz est developments for direct-injection (DI) Diesel windows for optical access. The burner matrix engines tend towards higher injection pressure and consists of a stainless steel sinter plate with a di- smaller nozzle diameters [2]. ameter of 20 mm. For stabilization, the central In the present work we focus on recent devel- sooting ethylene/air flame was surrounded by a opment of time-resolved laser-induced incandes- non-sooting methane/air flame (diameter 56 mm). cence (TiRe-LII) for the determination of mean The two flames were surrounded by a coflow of air soot particle size and size distributions in high- both for further stabilization of the flame and for pressure combustion environments prevalent ei- keeping the windows clean of soot and water. ther in steady laminar flames, transient Diesel Measurements were carried out for pressures up to spray combustion in constant volume pressure 10 bar for an equivalence ratio of  = 2.1, corres- vessels, or in small-size Diesel engines. Since gas ponding to a C/O ratio of 0.7. Further details are phase and/or soot particle temperatures are es- given in [13, 14]. sential when evaluating TiRe-LII signal profiles for For the LII experiments the beam of a short- particle sizing, in the described experiments differ- pulsed Nd:YAG laser at 1064 nm was aligned ent techniques were applied in obtaining this pa- through the burner. A small portion (1.9 mm diame- rameter. ter) of the beam was cut out with an aperture and Laser-induced incandescence has been used relay imaged onto the center of the burner in order successfully in high-pressure combustion environ- to obtain a homogeneous energy distribution within ments for measuring soot volume fractions in gas the beam. The resulting energy density in the ob- flames [3], spray flames [4], engine combustion [5, served volume was tuned in the range of 0.09– 2 6], and engine exhaust gases [7] both for point- 0.62 J/cm . For simultaneous particle pyrometry wise measurements and for two-dimensional imag- applications time-resolved LII signals were de- ing. Recent reviews state the current situation in tected at right angle at two different wavelengths LII experiments and modeling [8, 9]. Generally, (550 and 694 nm) with fast photomultipliers. * Corresponding author: thomas.dreier@uni-due.de Towards Clean Diesel Engines, TCDE2009 Gas phase temperature measurements were TiRe-LII experiments were performed with a accomplished by seeding the fresh gases with similar excitation / detection setup as described 0.5% to 2% of NO and using NO-LIF thermometry above (with only a single-color detection channel [15], where the A-X(0,0) band is probed at 225 nm active), acquiring LII decay profiles during prede- with a H2-Raman-shifted KrF excimer laser [16]. termined temporal delays after start of injection Simulated spectra were fitted to the experimental (SOI). Soot temperatures prior to particle laser data with absolute temperature, broadband back- heat-up were determined from spectrally resolved ground and total signal intensity as free parame- soot pyrometry using a spectrometer / camera ters. detection channel. A more detailed description of 0 the complete experiment and measurement tech- P HAB Tp Tg CMD g niques can be found in [17, 18]. [bar] [mm] [K] [K] [nm] Measurements were performed for initial gas 1 10 3758 1600 44 1.49 pressures between 1 and 3 MPa, injection pres- 2 10 3854 1710 31 1.54 sures between 50 and 130 MPa, and laser probe 5 5 3568 1850 10 1.84 timings between 5 and 16 ms after SOI. It is shown, that evaluated count mean particle diame- Table 1: Results of the measured temperatures ters (CMD) and standard deviations g are only and fit parameters of the comparison of the LII sig- slightly biased by the choice of typically assumed nal decay with the model at 1-5 bar. size distribution widths and gas temperatures. For a fixed combustion phase mean particle diameters Measured gas temperatures in the rich ethylene/air are not much affected by gas pressure, however flames at total pressures of 1, 2 and 5 bar in the they become smaller at high fuel injection pres- center (region of 10 mm wide  2 mm high) of the sure. At a mean chamber pressure of 1.39 MPa flame are listed in Table 1 with a relative uncertain- evaluated mean particle diameters increased by a ty of 3%. Spectral broadening of the lines in the factor of two for probe delays between 5 and 16 excitation spectrum reduces signal intensities and ms after start of injection, irrespective of the choic- spectral structure at higher pressures. The particle es of first-guess fitting variables, indicating a cer- temperatures at the time of laser heat-up deduced tain robustness of the least-squares fitting algo- from the 2-color pyrometry measurements are also rithm applied for TiRe-LII profile analysis. listed. ENGINE SOOT DIAGNOSTICS – The engine The time-resolved LII measurements show that used for in-cylinder LII was a single-cylinder, two- the LII decay rate in the heat conduction regime is stroke Diesel engine with a displacement volume linearly proportional to pressure, whereas compari- 3 of 250 cm with optical access through a custom son with soot volume fraction measurements by designed temperature-controlled (80°C) cylinder extinction does not show significant pressure de- head. The laser beam axis passed the center of pendence. When using prompt detection, calibra- the combustion chamber through two silica glass tion of the LII signal at atmospheric pressure windows, while the already described 2-color de- should be feasible for high-pressure applications. tection system has access to the combustion However, the influence of varying flame conditions chamber by a third window at the top of the cylind- on LII must be further addressed. er head. To keep windows clean and the thermal CONSTANT VOLUME CELL – For the compar- load on the cylinder head low the engine was mo- ison with model simulations, and for the develop- tored by an electrical asynchronous motor at a ment of optical diagnostics techniques, measure- constant speed of 1500 rpm and was fired for ments in cells where the fuel is injected into air at some individual cycles only. Further details of the high pressure and temperature and with no moving experiment are provided in [19]. All experiments pistons, such as in engines, often are more helpful. were performed at an injection crank angle of For this purpose measurements at the high- 23°CA before TDC and an equivalence ratio of  = temperature, high-pressure spray combustion 0.26. Soot particles were heated with a laser beam chamber at PSI (Switzerland) were performed. The 2 fluence at 1064 nm of 0.10 J/cm . Finally, a ther- cell is equipped with pneumatically actuated inlet mophoretic particle sampler was located in the and outlet valves, four sapphire optical windows exhaust gas manifold to get particle probes for (40 mm clear aperture) and a water cooled elec- further analysis by transmission electron microsco- tromagnetically actuated single-hole injection noz- py (TEM). zle. To simulate conditions close to top-dead cen- For the evaluation of the particle radiation sig- ter during Diesel combustion the cell is heated with nals in terms of particle size, it is important to know four heating cartridges (2 kW power consumption the respective mean combustion chamber pres- each) and is loaded with preheated, pressurized sure pg and temperature Tg. For the present engine air prior to fuel injection. The cell can be operated conditions, the pressure varied from nearly 80 bar with pressures up to 80 bar and wall temperatures at 0°CA to close to 1 bar at 100°CA. Gas tempera- up to 800 K, respectively. tures changed in this region from 2000 K to 1500 K, as was deduced both from two-color pyrometry resolved laser-induced incandescence (TR-LII). without laser heat-up and by calculating an individ- Proc. Combust. Inst., 2002. 29: p. 2775-2781. ual combustion cycle. With these two parameters it 7. Schraml, S., S. Will, and A. Leipertz, Perfor- was possible to evaluate the TiRe-LII signals. mance characteristics of TIRE-LII soot diagnos- tics in exhaust gases of Diesel engines. SAE Results are shown in Fig. 1. The CMD is in the Technical Paper Series No. 2000-01-2002, range of 30 to 75 nm, increases up to a crank an- 2000. gle of about 10°CA and then decreases again to- 8. Schulz, C., et al., Laser-induced incandes- wards a value of about 30 nm at 100°CA after cence: recent trends and current questions. TDC.80This behavior can be explained by particle 2500 Appl. Phys. B, 2006. 83(3): p. 333-354. formation and subsequent particle oxidation. g is 70 9. Michelsen, H.A., et al., Modeling Laser-Induced temperature Tg /K pressure pg /bar 2000 Incandescence of Soot: A Summary and Com- constant 60 at a value of 1.1 up to a crank angle of 70°CA and then increases to about 1.32. The two parison of LII Models. Appl. Phys. B, 2007: p. 50 1500 DOI: 10.1007/s00340-007-2619-5. 40 values of CMD and g, which are shown at circled 10. Michelsen, H.A., Understanding and predicting the 30right edge of the diagram, are results 1000 of the the temporal response of laser-induced incan- TEM analysis of the exhaust gas primary particles. descence from carbonaceous particles. J. 20 It is observed that the TiRe-LII measured sizes 500 at Chem. Phys., 2003. 118: p. 7012-7045. 100°CA10 after TDC approach the TEM determined 11. Liu, F., et al., Influence of polydisperse distribu- primary particle tions of both primary particle and aggregate 0 20sizes 40 in the exhaust 60 80 gas 100 quite well. crank angle /° a. TDC size on soot temperature in low-fluence LII. geometric standard deviation Appl. Phys. B, 2006. 83: p. 383-395. count median diameter /nm 100 1.5 12. Hofmann, M., B. Kock, and C. Schulz. A web- 80 1.4 based interface for modeling laser-induced in- candescence (LIISim). in European Combus- tion Meeting. 2007. Chania, Greece. 60 1.3 13. Hofmann, M., et al., Laser-induced incandes- cence (LII) for soot diagnostics at high pres- 40 1.2 sure. Appl. Opt., 2003. 42(12): p. 2052-2062. 14. Hofmann, M., et al. Soot diagnostics at ele- 20 1.1 vated pressure: Laser-induced incandescence and multi-line NO-LIF thermometry. in Fourth 1 Australian Conf. Laser Diagnost. Fluid Mech. 0 20 40 60 80 100 crank angle /° a. TDC And Combust. 2005. Adelaide. 15. Bessler, W.G. and C. Schulz, Quantitative Fig. 1: CMD and g evaluated from the LII signals multi-line NO-LIF temperature imaging. Appl. taken in the engine combustion chamber; symbols on Phys. B, 2004. 78: p. 519-533. the right axis are the respective values from the ex- 16. Dreier, T., A. Dreizler, and J. Wolfrum, The haust gas TEM sampling application of a raman-shifted tunable KrF ex- cimer laser for laser-Induced fluorescence combustion diagnostics. Appl. Phys. B, 1992. 55: p. 381-387. 17. Bougie, B., et al., Optical diagnostics of diesel References spray injections and combustion in a high- pressure high-temperature cell. Appl. Phys. B, 1. Dec, J.E., A conceptual model of DI Diesel 2005. 80: p. 1039-1045. combustion based on laser-sheet imaging. SAE 18. Ryser, R., T. Gerber, and T. Dreier, Soot par- Technical Paper Series 970873, 1997. ticle sizing during high-pressure Diesel spray 2. Pickett, L.M. and D.L. Siebers, Soot in diesel combustion via time-resolved laser-induced in- fuel jets: effects of ambient temperature, am- candescence. Combust. Flame, 2008. 155: p. bient density, and injection pressure. Combust. 430-439. & Flame, 2004. 138: p. 114-135. 19. Kock, B.F., et al., Two-color time-resolved LII 3. Ni, T., et al., Two-dimensional imaging of soot applied to soot particle sizing in the cylinder of volume fraction by the use of laser-induced in- a Diesel engine. Combust. Flame, 2006. 147: p. candescence. Appl. Opt., 1995. 34: p. 7083- 79-92. 7091. 4. Vander Wal, R.L. and D.L. Dietrich, Laser- induced incandescence applied to droplet com- bustion. Appl. Opt., 1995. 34: p. 1103-1107. 5. Dec, J.E., Soot Distribution in a D.I. Diesel Engine Using 2-D Imaging of Laser-Induced In- candescence, Elastic Scattering, and Flame Luminosity. SAE Technical Paper Series No. 920115, 1992. 6. Kock, B.F., T. Eckhardt, and P. Roth, In- cylinder sizing of Diesel particles by time-