Grantee Research Project Results
Final Report: Soot, Precursor Particle and Higher Hydrocarbon Production in Flames
EPA Grant Number: R828167Title: Soot, Precursor Particle and Higher Hydrocarbon Production in Flames
Investigators: Pfefferle, Lisa , McEnally, Charles
Institution: Yale University
EPA Project Officer: Hahn, Intaek
Project Period: August 1, 2000 through July 31, 2002 (Extended to July 31, 2003)
Project Amount: $224,170
RFA: Exploratory Research - Engineering, Chemistry, and Physics) (1999) RFA Text | Recipients Lists
Research Category: Safer Chemicals , Water , Land and Waste Management , Air
Objective:
Soot formation is a complex, multi-step process; however it is useful to group the processes into two overall steps. The first involves decomposition of the fuel and the formation of the next highest aromatic ring containing species (e.g., for alkanes, benzene) from the decomposition products. The second includes the growth of additional rings onto these small aromatics, their nucleation into particles (likely involving liquid-like intermediate droplets), and surface growth and aggregation of the particles. The reason for distinguishing these two steps is that this research has shown that the specific mechanisms responsible for the first step depend on fuel composition and the extent of fuel/air premixing. In contrast, the mechanisms of the second part were found in this research to be fairly universal. Although the second step actually forms the soot particles, both steps affect the overall rate of soot production, and any change in the rate of small aromatics formation will change the amount of soot formed.
With this picture in mind, the overall focus of our research was to better understand the initial growth processes for soot in flames. This included developing methodologies for characterization of precursor and mature soot, identifying the flame location/gaseous environment where precursor particles are born, and using dopant studies to probe different mechanisms for soot nucleation. In our studies, we dope well characterized methane or ethylene flames with particular additives and model the perturbation effect. We find that this is a particularly powerful method of probing flame mechanisms because of the large number of degrees of freedom in the system. The flames are characterized by photoionization mass spectrometry, GC-MS, laser-induced fluorescence (OH, HCO), laser light scattering, laser -induced incandescence, and a number of physical soot characterization techniques.
Approach:
The capability to rapidly measure a wide range of both gas phase species and particulate types is of key importance for this study where a large array hydrocarbon species, combustion products, and both clear and mature soot particles are present throughout the flame. A distinctive capability of our laboratory is the ability to combine rapid and concurrent gas phase and particulate phase measurements over a wide range of conditions in soot laden diffusion flames. We will use VUV-MS and REMPI to simultaneously measure both combustion intermediates and products in a broad mass range (0-2000 amu) and LIF to provide OH concentrations. Soot will be probed using laser-induced incandescence (LII) to extract the soot concentration and primary particle size distribution. These measurements will be complemented by absorption and fluorescence measurements and the thermophoretic sampled particle diagnostic with TEM analysis (TPD/TEM). Since these techniques are based on the measurement of different physical properties, comparison of profiles made for the same conditions reveal changes in the nature of particulates throughout the flame.
Summary/Accomplishments (Outputs/Outcomes):
Our results demonstrate the sensitivity of the small aromatic formation mechanisms to both fuel structure and fuel/air mixing, and demonstrate the importance of this sensitivity to the overall rate of soot formation. Several new growth pathways for soot precursors were identified as discussed below. An important contribution of this research is the identification of chemical effects due to fuel structure and partial premixing as discriminated from thermal, residence time, or physical/transport factors. Three important factors that affect the mechanisms for aromatic hydrocarbon production, fuel structure, mixture effects , and fuel/air mixing, are discussed. We have also been developing correlations based on structural features to predict the individual fuel reaction pathways/ rates and sooting tendencies of complex fuel mixtures.
Selected specific findings are outlined below:
- For all nonaromatic additives doped into methane flame, soot always correlated with the first aromatic species formed. (Our flames are nucleation controlled.) The sootability of the additive relates to its decomposition pathway, with species decomposing to large resonantly stabilized radical species forming the most soot. C6 dienes, regardless of structure, are highly sooting and form similar soot levels regardless of structure because of their stability. This work points out the importance of fuel structure effects.
- In our lightly sooting co-flow flames, precursor particles are born at the region where the naphthalene concentration, as measured by mass spectrometry, reaches its maximum on the centerline of the flame. The precursor particles are born towards the center of the flame but are present throughout pure methane flames. Gas phase naphthalene concentrations are coincident with broad band laser-induced fluorescence (BLIF), a marker for soot precursor particles. (Gas phase species contributions are small compared to that from the “particles.”) The precursor particles peak is just upstream of soot measured by other techniques (e.g., laser light scattering, which measures both mature soot and the larger precursor particles and laser- induced incandescence, which measures only mature soot), suggesting that they are true soot precursors. The number density of soot particles remains approximately constant as the particles carbonize, another indication of the role of the transparent “precursor” particles.
- We have completed a study on small air addition to benzene-doped methane flames and found that for the levels of benzene found in practical fuels, the growth to higher molecular weight hydrocarbons largely proceeds through the hydrogen abstraction/carbon addition (HACA) mechanism. This is in contrast to the results found in lower temperature (i.e., around 1100K), lean premixed reactors of pure benzene oxidation when C5 routes to aromatics were observed to be important. This confirms our earlier hypothesis that for most fuel mixtures containing alkanes, the methyl radical concentration is high enough to compete effectively for cyclopentadienyl radical. It also points out the danger in applying pure fuel results to real fuel mixtures.
- To further explore the role of C5 recombination as a function of fuel/air premixing, air was added to the fuel side of non-premixed benzene flames to an equivalence ratio of 3. This allowed exploration of the importance of oxidative pyrolysis reactions to open new channels, providing proposed aromatic growth species such as cyclopentadienyl radical (see paper list below). Although cyclopentadiene is increased by air premixing, recombination of cyclopentadienyl radical to naphthalene never dominated two ring polycyclic aromatic hydrocarbon (PAH) growth by other mechanisms (such as acetylene addition) at our flame conditions. This contrasts with reports from some other investigators regarding the importance of the cyclopentadienyl recombination pathway to naphthalene, but is consistent with our earlier results using flames doped with cyclopentene and 13C labeled cyclopentene. Our results indicate that cyclopentadienyl recombination is not likely to be important in any practical flame condition. In laboratory experiments of very lean, moderate temperature benzene oxidation, this pathway is important (as shown in our earlier benzene oxidation work and by Jack Howard). In practical flames, however, even a small amount of alkane present creates enough methyl radicals that the reaction CH3 + C5H5 products dominate for consumption of cyclopentadienyl.
- Our work has shown evidence for significant synergistic effects. For example, adding methane to an ethylene fueled diffusion flame can actually increase PAH, even though methane is less “sooting” than ethylene due to the importance of the C3 route to benzene (see Roes ler, et al., in publications list). Methyl radicals play an important role in many reactions including ring enlargement of C5 cyclics. Even small amounts of alkanes in fuel mixtures can provide enough methyl radicals to allow these reaction pathways to contribute measurably. This is not clearly appreciated in the literature, as the C5 recombination route to aromatics in practical flames is still widely accepted.
Conclusions:
We have shown in this work that “transparent liquid-like” particles are the likely direct precursors of soot. Methods for their analysis have been developed. More broadly, we have addressed issues relating to soot formation from complex fuels and developed strategies for analysis of mechanisms and effects of fuel structure. A key issue, when addressing the complex kinetics issues posed by fuels containing mixtures of large hydrocarbons, is which of the thousands of possible reactions are important to practical flames. Our basic methodology, implemented and validated in our previous work, is to study flames that are perturbed in systematic ways to address specific mechanistic issues related to formation of small aromatics. Our detailed measurement and model allows comparison of changes in species profiles induced by the perturbation, as well as changes in computed and measured rates of production to help deconvolute errors in the production of one species caused by misprediction of another. Our perturbation strategy also helps in this respect because instead of just comparing measured and predicted profiles, we compare how the model and experiment reacts to a perturbation. Since our primary experimental diagnostics are rapid and on-line, all of the diagnostics for a set of five or six systematically varied flames can be completed in approximately a week. Thus, we combine extensive variation of flame conditions with detailed concentration and related measurements for a broad range of species at each condition. This is an important advantage because it allows us to span a large parameter space with detailed measurements to identify parameter regimes for detailed study where the most interesting “effects” occur and where model discrimination is greatest. Thus, we can identify where the sensitivity of an “effect” to the perturbed parameter is highest, including evaluating how changing flame conditions affects this sensitivity. This region is not where the effect is largest, since at that point the slope of the effect versus the perturbed parameter is zero. The high sensitivity of the on-line MS is thus also important because the highest sensitivity of a target species to perturbed parameters often occurs at low concentrations. Having identified regions of interest where the effect is changing rapidly and where comparison with the model raises additional questions, new experiments can be appropriately designed. Listed below are some important generalizations gleaned from this work which are not yet in our published literature. We are currently preparing a requested review article which will include these points and cite EPA support.
- We have found for a wide range of additives that “sootability” scales with the ease of forming the next higher aromatic ring. There are exceptions because some additives have very fast routes to the next higher ring or beyond. Characterization of the side chain/s alkyl aromatics is a good example where I believe that structure “scalability” can be used productively. The overall order for sooting tendency is tertiary or primary carbon < secondary carbon < quaternary carbon. This correlation holds over a wide range of fuels. The order is somewhat different than has been reported previously from estimates because alkyl aromatics with a secondary carbon attached to the ring decompose to benzyl and form benzene relatively quickly. We have also looked at xylenes but have not yet fully analyzed the pathways due to the multiplicity. It does look, however, as if scaling relationships are approachable for predicting sootability for fuel mixtures.
- The sooting tendency of paraffins scales with the type and number of carbon branches. The order is as expected: only primary < only secondary < only tertiary < only quaternary. (Interestingly, the analogous alcohols and ethers fall on top of the paraffins if all other flame conditions are held constant). For the heptane series, all of the isomers with two tertiary carbons were identically sooty because the isomerization of the heptyl radicals is faster than their decomposition, again supportive of the idea that useful groups and correlations can be developed. The same groupings also order “ignitibility,” with the more branched consumed more quickly (assuming that the fuel decomposition is dominated by fission processes; of course, in the limit of H abstraction dominating, the order is roughly reversed).
- Basic experiments and scaling approaches for development of mechanisms for complex fuels depend on the objective function to be satisfied. We believe that for many high temperature practical combustion systems, for example, the fuel disappearance proceeds significantly through unimolecular decomposition steps, simple or complex fission. This is contrary to most conventional wisdom and may be important, as it dictates the type of experiment needed to verify the important reaction steps.
Broader Implications of Our Work
Our results can be used to aid in design of emissions regulations. We have shown particular fuel structures considerably increasing soot production as well as developed general rules for fuel structure determining sooting behavior. We have also shown how slight premixing of air can dramatically alter emissions profiles leading to increased sooting. On a more general note, we carried out a feasibility study on determining health effects of soot with Professor John Wise (formerly at Yale, now at the University of Maine) to use a model lung cell system to assess the toxicity of soot. Clean sub-micron carbon soot was found to be reasonably nontoxic when compared to the National Institute of Standards and Technology urban particulate standard. Toxicity increased when aromatic compounds were added back. Although these were feasibility tests only, they point to the possibility of testing particular synergistic effects on toxicity that might guide burner design and fuel composition. In addition, it is important to characterize the toxicity of precursor soot versus mature carbon soot. As we design “cleaner” engines, we may be increasing precursor soot emissions which may, in fact, be significantly more toxic than mature carbon soot.
Expected Results:
In the proposed work we obtain a more complete characterization of the soot inception process in diffusion flames. Simultaneous measurement of both soot and soot precursors along with a map of stable species and OH radical will be important in testing mechanisms in the literature and pointing up areas where information is particularly needed. Characterization of precursor particles will be important for both risk assessment and management as the precursor particle emissions are currently uncontrolled and may provide more health risk than mature soot.
Journal Articles on this Report : 5 Displayed | Download in RIS Format
Other project views: | All 8 publications | 5 publications in selected types | All 5 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
Bennett BAV, McEnally CS, Pfefferle LD, Smooke MD, Colket MB. Computational and experimental study of axisymmetric coflow partially premixed ethylene/air flames. Combustion and Flame 2001;127(1-2):2004-2022. |
R828167 (Final) |
Exit Exit Exit |
|
McEnally CS, Pfefferle LD. The effects of slight premixing on fuel decomposition and hydrocarbon growth in benzene-doped methane nonpremixed flames. Combustion and Flame 2002;129(3):305-323. |
R828167 (Final) |
not available |
|
McEnally CS, Pfefferle LD. Experimental study of fuel decomposition and hydrocarbon growth processes for cyclohexane and related compounds in nonpremixed flames. Combustion and Flame 2004;136(1-2):155-167. |
R828167 (Final) |
Exit Exit Exit |
|
McEnally CS, Pfefferle LD, Atakan B, Kohse-Höinghaus K. Studies of aromatic hydrocarbon formation mechanisms in flames: progress towards closing the fuel gap. Progress in Energy and Combustion Science 2006;32(3):247-294. |
R828167 (Final) |
Exit |
|
Roesler JF, Martinot S, McEnally CS, Pfefferle LD, Delfau J-L, Vovelle C. Investigating the role of methane on the growth of aromatic hydrocarbons and soot in fundamental combustion processes. Combustion and Flame 2003;134(3):249-260. |
R828167 (Final) |
Exit Exit Exit |
Supplemental Keywords:
polyaromatic hydrocarbon,, RFA, Scientific Discipline, Air, Toxics, Waste, particulate matter, Environmental Chemistry, HAPS, Engineering, Chemistry, & Physics, Incineration/Combustion, particle size, particulates, hydrocarbon, particle precursor, soot nucleation, gas flow rates, PAH, particles, combustion, laser induced incandescense, thermophoretic sampled particle diagnostic, particle surface interactions, doped flame method, surface growth, combustion productsProgress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.