27^th Symposium (International) on Combustion, Accepted for Publication

The Effects of Equivalence Ratio on the Formation of Polycyclic Aromatic Hydrocarbons and Soot in Premixed Methane Flames

Antonio M. Vincitore, Tyler R. Melton and Selim M. Senkan*, Department of Chemical Engineering, University of California

Los Angeles, CA 90095

* To whom correspondence should be addressed: Tel: (310) 206-4106, Fax: (310)206-4107, e.mail: senkan@seas.ucla.edu

ABSTRACT

The formation of polycyclic aromatic hydrocarbons (PAH) and soot has been investigated in atmospheric pressure, laminar, methane/oxygen/argon premixed flames as a function of mixture equivalence ratio. Mole fraction profiles of major products, trace aromatic, substituted aromatic and PAH were quantified by direct gas chromatography/mass spectrometry. In addition, soot particle diameters, number densities and volume fractions were determined using classical light scattering. The dependencies of flame species on equivalence ratio, using the expression, X_i^max = A_i ⁿⁱ were also determined; results reveal that the parameter (n) for stable aromatic precursors exhibit the following rank order: C₄H₂ (4.39) > C₃H₄ (3.96) > C₄H₆ (3.41) > c-C₅H₆ (3.30) > C₄H₄ (2.31) > C₂H₂ (1.35). For aromatic species, the values of n were in the following order: Phenylacetylene (10.88) > Benzene (10.21) > Indene (7.63) > Toluene (5.43). In comparison, PAH species were extremely sensitive to flame equivalence ratios, with the following n values: Fluoranthene (15.1) > Acenaphthylene (14.8) > Pyrene (13.8) > Anthracene (13.1) > Phenanthrene (12.9) > Naphthalene (11). These results indicate that PAH formation rates are not strongly influenced by the levels of acetylene present in the flame under the conditions investigated.

INTRODUCTION

Recently, we began a series of experimental studies of fuel-rich, premixed hydrocarbon flames to discern the relationships between combustion conditions and the formation of soot and polycyclic aromatic hydrocarbons (PAH) [1,2,3]. These results indicated that aromatic and PAH levels are strongly related to fuel structure and soot levels present within the flame. In earlier studies, PAH were advocated to be the precursors to soot formation [4], suggesting that an increase in soot production would be correlated to higher PAH levels. However, our recent experimental results indicate that this is not generally true [2], thus suggesting that PAH and soot formation are part of a complex chemical reaction network. Over the past years, numerous experimental studies involving premixed methane flames have been performed with a major focus on elucidating the physical and chemical processes associated with PAH and soot formation [5,6,7]. However, there still exists uncertainty with regard to reaction pathways leading to PAH production and their relationship to soot. To date, two main reaction routes have been proposed to describe the production of the first aromatic ring, i.e. benzene. The first route involves the reaction of C₄ and C₂ species yielding C₆s, followed by the cyclization and dehydrogenation of the adduct [4]. The formation of PAH were then accounted for by the activation of the aromatic ring followed by the addition of C₂H₂ and other carbon moieties [4]. The second route involves the recombination of resonantly stabilized C₃ species resulting in the direct formation of the aromatic ring [8]. Recently, new reaction pathways have also been proposed to account for the formation of PAH [9]. For example, the recombination of cyclopentadiene has been suggested to directly produce naphthalene, thus eliminating the need to form benzene first and expediting the PAH production process.

In this communication, we report on the detailed chemical structures of fuel-rich, laminar premixed methane flames as a function of mixture equivalence ratio in order to develop insights on the relation of flame chemistry with PAH and soot formation. The information acquired was also analyzed to establish trends such that better models describing PAH and soot formation can be developed. In an earlier study, Musick et al. [10] reported the C₁-C₄ chemical structures of low-pressure, premixed flames of CH₄/O₂/Ar over an equivalence ratio range of 0.92-1.94. However, no information on aromatic and PAH was provided.

EXPERIMENTAL

The experimental set-up has been described elsewhere [1], thus only a brief description will be given here. Atmospheric-pressure, premixed, laminar, flat-flames of CH₄/O₂/Ar were stabilized over a 50 mm diameter porous bronze burner. The burner was cooled by circulating an ethylene glycol solution at 104.3 ^oC. Flames were protected from the surrounding air by a concentric shield of Argon gas. Gas flows were regulated using high accuracy mass flow controllers (MKS, Burlington MA). The purities of the gases used in this experimental study were 99.99 %. The initial feed compositions and equivalence ratios for the flames studied are summarized in Table 1. In all the flames, the methane volumetric flow rate was maintained constant at 1.88 liters per minute, while the flows of argon and oxygen were varied. In addition, the argon mole fractions in all flames were maintained at 0.45. This allowed for a better control of the flame temperatures, thus a better comparison of the concentrations as a function of equivalence ratio. The chemical structures of five flames with equivalence ratios ranging from 2.0 to 2.6 were examined by withdrawing samples from within the flame using a heated quartz microbe with a 100 µm orifice at its tip. Samples were transported through a silica lined, stainless steel transfer line. Flames below an equivalence ratio of 2.0 were difficult to sample because of their close proximity to the burner surface. Flames with equivalence ratios above 2.6 were too unstable and sooting to sample. A quartz wool filter was placed inside the probe to trap soot particles. The entire sampling system was maintained at 300 ^oC to minimize the adsorption of large molecular weight PAH.

Experimental Conditions of Atmospheric, Premixed Methane/Oxygen/Argon Flames ^a.

^a V_o is defined as the cold gas velocity. T_o is defined as the burner surface temperature.

Samples were immediately analyzed by a computer controlled gas chromatograph/mass spectrometer (Hewlett-Packard 5890/5972) using both capillary (0.25 mm x 60 m HP-5) and packed bed (Hayesep T and Hayesep DB) columns. Major species were analyzed using the thermal conductivity detector (TCD). Some minor and all the trace species were analyzed using the mass spectrometer. Species concentrations were determined either by direct calibration standards (Matheson Gas, Sigma Aldrich), or by the use of mass spectral ionization cross section method [11]. The accuracy of latter has been reported to be within a factor of two [11], and was verified in previous studies [1]. The accuracy of the mole fractions for species determined by direct calibration is estimated to be ±15%. Both H₂ and H₂O concentrations were determined from H and O atomic balances.

Concentration profiles were generated by moving the entire burner assembly vertically up or down with respect to the fixed sampling probe. Positional accuracy associated with the concentration and temperature measurements is estimated to be ± 0.25 mm. Temperature measurements were obtained with a silica coated, 0.15 mm Pt-13%Rh/Pt thermocouple. Temperature measurements reported in this work were not corrected for radiation losses.

Soot sizes, number densities and volume fractions were determined using classical laser light scattering and extinction measurements [12]. For this, a collimated and focused (0.2 mm) 514.5 nm laser beam (Spectra Physics, 2037, 1.0 W) that was mechanically modulated at 1000 Hz, was passed through the flame. The light scattered by the soot particles was collected at 90^o from the incident beam with a lens and directed through an iris onto the cathode of a photomultiplier (PMT, Hamamatsu, R1463). The PMT signal was amplified using a lock-in amplifier (EG&G 5205). The transmitted light was measured with a photodiode and recorded with an oscilloscope. Polarized narrow band filters were placed in front of the photomultiplier and photodiode to minimize signal contamination from flame luminosity. The light scattering set-up was calibrated against a known, particle free, flow composed of argon. The calculation of the number densities and volume fractions were accomplished by assuming monodisperse particles having a complex refractive index of 1.54+0.58i [13].

DISCUSSION

In Fig.1, the temperature and soot profiles for the flames studied are presented. In this and subsequent figures, lines have been drawn through the data points to indicate trends. Due to crowding, Fig. 1 through Fig. 6 provide information for all flames excluding Flame 3. In Fig. 7 and 8, we report the PAH concentration profiles for all flames. As seen in Fig. 1, the flames based on their temperature profiles, were positioned at various distances from the burner surface, consistent with their equivalence ratios. For example, Flame 1 was positioned closest to the burner, because of its lowest equivalence ratio and thus highest burning velocity. In contrast, Flame 5 had the lowest burning velocity, and was positioned the farthest from the burner surface. The flame temperatures exhibited near linear profiles from the burner surface to the point of maximum temperature. It is also important to note that the maximum flame temperatures reached in all the flames were about 1300 ^oC, with a 25 ^oC difference in peak temperatures between the flames studied. This result is consistent with past experimental work [14]. At distances substantially away from the burner surface, soot deposition on the thermocouple bead was inevitable. This led to increased radiative heat losses from the thermocouple bead yielding in lower temperature readings. To minimize this effect, the thermocouple was kept in the sooting section of the flame for a minimal amount of time, and any accumulated soot was burned off with the aid of a small propane torch after each measurement [1].

FIG. 1. Temperature (T), soot particle diameter (d_p), number density (N) and volume fraction (f_v) measurements. Symbols used in soot measurements correspond to the following flames : (2.0), (2.2),(2.3), (2.4), (2.6).

In Fig. 1, the measured soot volume fraction, particle diameter and number density profiles are also reported. These properties were calculated by assuming monodisperse soot particles that were smaller than the wavelength of the light used (515.5 nm) in order to satisfy the Rayleigh scattering approximation. Although non-spherical and polydisperse soot occurs within flames to due the agglomeration of primary soot particles, the Rayleigh approximation has been demonstrated to hold to within a factor of three [15]. As evident from Fig. 1, the soot volume fraction steadily increases with increasing equivalence ratio and with distance above the burner surface. In contrast, the soot number density first increases with flame height and exhibits a maxima near the maximum flame temperature and then decreases with height above the burner. Within the flame, the decrease in soot number density can be attributed to the coagulation of elementary soot particles. As seen in Fig.1, soot particle diameters also increased with increasing equivalence ratio, closely tracking the soot volume fraction. These results were also seen in previous studies [14]. The measured maximum particle diameters ranged from 10 nm to 28 nm, for Flame 2 to Flame 5, respectively. We were not able to determine particle diameters for Flame 1.

In Fig. 2, the mole fraction profiles for the reactants CH₄ and O₂ and two of the major carbon containing products CO and CO₂ are presented. As evident from this figure, the profiles exhibit features consistent with the relative equivalence ratios. For example, both CH₄ and O₂ increasingly penetrate into the post flame zone with increasing equivalence ratio. As expected CO was the major combustion product in all cases.

FIG. 2. Concentration profiles of CH₄, O₂, CO and CO₂ for Flames 1-5. Symbols correspond to the following flames : (2.0), (2.2), (2.4), (2.6).

FIG. 3. Mole fraction profiles of H₂, H₂O, C₂H₂ and C₂H₄. Symbols correspond to the following flames : (2.0), (2.2), (2.4), (2.6).

The mole fraction profiles for H₂, H₂O, C₂H₂ and C₂H₄ presented in Fig. 3 also exhibit features consistent with the mixture equivalence ratio. However, the post flame H₂ and H₂O levels were not sensitive to equivalence ratio. It is also interesting to note that with the exception the lowest equivalence ratio case , i.e. =2.0, the post flame levels of acetylene also we re insensitive to the mixture equivalence ratio. This is an intriguing result, because of the significant influence equivalence ratio had on soot volume fraction and diameter (Fig. 1) and the generally believed notion that acetylene is an important soot growth species [16,17].

In Fig. 4, the mole fraction profiles for ethane, 1,3 propadiene/1-propyne, diacetylene and vinylacetylene are presented. As evident from this figure, the mole fraction profiles for ethane peaked early in the flame zone, indicating the role of this species in the pre-flame reaction zone. In contrast, 1,3 propadiene/1-propyne, diacetylene and vinylacetylene penetrated well into the post flame zone suggestive of their role in PAH and soot formation.

Also evident from Fig. 4, 1,3 propadiene/1-propyne, diacetylene and vinylacetylene appear to closely track the mixture equivalence ratio. That is, their formation is delayed and their peak mole fractions increased with increasing equivalence ratio.

FIG. 4. Concentration profiles of C₂H₆, C₃H₄, C₄H₂ and C₄H₄. Symbols correspond to the following flames : (2.0), (2.2), (2.4), (2.6).

FIG. 5. Profiles of 1,2 butadiene (a-C₄H₆), 1,3 butadiene (b-C₄H₆), 1,3 pentene-yne (C₅H₆) and cyclopentadiene (c-C₅H₆). Symbols correspond to the following flames : (2.0), (2.2), (2.4), (2.6).

In Fig. 5, the mole fraction profiles for 1,2 butadiene, 1,3 butadiene, 3-pentene-1-yne and cyclo-pentadiene are presented. With the exception of 3-pentene-1-yne, all of these species penetrated well into the post flame zone suggesting their possible role in PAH and soot formation chemistry. These species also exhibited trends consistent with the compositions of the flames studied, with increasing levels of production with increasing equivalence ratio.

In Fig. 6, the mole fraction profiles for the aromatic products, that include benzene, toluene, phenyl-acteylene and indene are presented. All of these species penetrated well into the post flame zone and exhibit trends consistent with the equivalence ratio.

FIG. 6. Profiles of benzene (C₆H₆), toluene (C₇H₆), phenylacetylene (C₈H₆) and indene (C₉H₈). Symbols correspond to the following flames : (2.0), (2.2), (2.4), (2.6).

The mole fraction profiles of six representative PAH measured in the flames are compared to one another in Fig.7 and Fig. 8. The structural designations of various PAH species are presented in Table 2. Although, there were many other PAH measured in the flames, space limitations prohibit the presentation of the entire set at this time. As evident from these figures, the levels of PAH produced in the flame steadily increased with flame height and generally leveled off at about 10 mm from the burner surface.

FIG. 7. Profiles of naphthalene (C₁₀H₈), acenaphthylene (C₁₂H₈) and anthracene (a-C₁₄H₁₀). Symbols correspond to the following flames : (2.0), (2.2),(2.3), (2.4), (2.6).

The dependencies of various chemical species on equivalence ratio were also analyzed using the empirical formula X_i^max = A_i ⁿⁱ [18]. In this expression, X_i^max is the maximum mole fraction of species i, A_i and n_i are the correlation parameters. The parameter n_i represents the sensitivity of species i on equivalence ratio, and has previously been used to explore the C₁-C₄ chemistry of low pressure CH₄/O₂/Ar flames at equivalence ratios range of 0.92-1.94 [10]. The values of A_i and n_i were determined from linear regression of our experimental data. The values of A_i and n_i, as well as the corresponding correlation coefficients are presented in Table 2. As evident from this table, the experimental results were well represented by the above empirical relationship with correlation coefficients higher than 0.9 for all species.

An inspection of Table 2 reveals that the mole fractions of aliphatic species produced in the flames were moderately sensitive to mixture equivalence ratio. The rank order of correlation parameter n_i were: C₄H₂ (4.39) > C₃H₄ (3.96) > C₄H₆ (3.41) > c-C₅H₆ (3.3) > b-C₄H₄ (2.31) > C₂H₂ (1.35). Although this rank order is consistent with similar determinations made in low pressure flames [10], the n_i values are considerably different. However, the latter is not surprising in view of the differences in equivalence ratios and pressures investigated.

FIG. 8. Profiles of phenanthrene (b-C₁₄H₁₀), fluoranthene (a-C₁₆H₁₀) and pyrene (b-C₁₆H₁₀). Symbols correspond to the following flames : (2.0), (2.2),(2.3), (2.4), (2.6).

In contrast to aliphatics, aromatics were significantly more sensitive to changes in equivalence ratio. The calculated n_i values for aromatic species were: Phenylacetylene (10.88) > Benzene (10.21) > Indene (7.63) > Toluene (5.43). Furthermore, PAH species were extremely sensitive to flame equivalence ratios, exhibiting the following rank order for n_i: Fluoranthene (15.1) > Acenaphthylene (14.8) > Pyrene (13.8) > Anthracene (13.1) > Phenanthrene (12.9) > Naphthalene (11).

In summary, the measurements of in-flame concentrations of aliphatics, aromatics and PAH exhibit different dependencies on the mixture equivalence ratio. PAH levels were most significantly affected by equivalence ratio followed by simple aromatics and aliphatics. The levels of acetylene were least sensitive to equivalence ratio. The results presented should be of considerable utility for the development and validation of models describing the formation of PAH and soot in flames.

Calculated values for the A_i factor and exponent n_i for minor, aromatic and PAH species.

ACKNOWLEDGMENTS

This research was supported, in part, by the US Environmental Protection Agency, the National Science Foundation, the UCLA Center for Clean Technology and the Petroleum Environmental Research Forum Project.

REFERENCES

1. Castaldi, M., Vincitore, A.M. and Senkan, S.M., Combust. Sci. Tech.107:1-19 (1995).

2. Senkan, S. and Castaldi, M.,Combust. Flame. 107: 141-150 (1996).

3. Castaldi, M. and Senkan, S.M., Combust. Sci. Tech. 116-117:167-181 (1996).

4. Frenklach, M. and Warnatz,J., Combust. Sci. Tech. 51: 265 (1987).

5. D'Alessio, A., Di Lorenzo, A., Sarofim A.F., Beretta, F., Masi, S. and Venitozzi, C., 15th Symposium (Int'l) on Combustion, The Combustion Institute, Pittsburgh PA, p.1427 , (1975).

6. Di Lorenzo, A., D'Alessio, A., Cincotti, V., Masi, S. and Venitozzi, C., 18th Symposium (Int'l) on Combustion, The Combustion Institute, Pittsburgh PA, p.485 , (1981).

7. Biordi, J. C., Lazzara, C. P. and Papp, J., 16th Symposium (Int'l) on Combustion, The Combustion Institute, Pittsburgh PA, p.1097, (1977).

8. Miller, J.A. and Melius, C.F., Combust. Flame, 91: 21-39 (1992).

9. Marinov, N.M., Pitz, W.J., Westbrook, C.K., Castaldi, M.J. and Senkan, S.M., Comb. Sci. and Tech., 116:211-287 (1996).

10. Musick, M. Vantiggelen, P. J., Vandooren, J., Combust. Flame. 105: 433 (1996).

11. Fitch, W.L. and Sauter, A.D., Anal. Chem. 55: 832-836 (1983).

12. Flower, W.L., SPIE, 644: 28 (1986).

13. Faeth, G.M. and Koylu, U.O., Combust. Sci. Tech. 108: 207 (1995).

14. Prado, G. and Lahaye, J., Particulate Carbon Formation During Combustion (Siegla, D.C. and Smith, G.W., Eds.), Plenum Press, NewYork-London, pp 143-175 (1981).

15. Koylu, U.O., Combust. Flame. 109: 488 (1997).

16. Frenklach, M., Clary, D., Gardiner, Jr., W.C. and Stein, S.E., 20th Symposium (Int'l) on Combustion, The Combustion Institute, Pittsburgh PA, p.887, (1984).

17. Weiner, A.M. and Harris, S.J., Combust. Flame. 77: 261 (1989).

18. Warnatz, J., in Soot in Combustion Systems and Its Toxic Properties (Prado, G. and Lahaye, J. Eds.), Plenum Press, NewYork-London, pp 128 (1983).