(178c) Larger Unsaturated Radicals under Combustion Conditions

INTRODUCTION: There has been much recent interest in the simulation of real combustion devices with realistic fuels. A direct impetus is the increasing price of oil, the consequent interest in alternative fuels and the continuing need for satisfying environmental regulation. These needs have come at the time when modern Computational Fluid Dynamics (CFD) programs can handle increasing amounts of detailed chemistry. Thus the fundamental basis of combustion, fluid dynamics and chemical kinetics has reached a stage where it is now possible to consider simulating the behavior of combustion systems in terms of basic interactions. This presentation is a summary of current attempts at providing the chemical kinetics databases that are necessary to enable this technology. We outline some of the problems and possible solutions. We use the pyrolysis kinetics of larger unsaturated radicals as examples. These organic fragments are among the immediate intermediates in combustion processes and characteristic of the breakdown of the large molecular weight compounds found in real fuel mixtures.

SCOPE: A complete combustion kinetics database contains fundamental information dealing with the interaction of individual fuel molecules with the bath and the reactive fragments. It is most conveniently divided into four separate modules. These are (a) pyrolytic cracking, (b)oxidative cracking, (c) light hydrocarbon combustion and (d) PAH/SOOT formation The last two modules deal with processes formed from the products of the reactions from modules (a) and (b). These reduce the initial large fuel molecules to smaller radicals and unsaturated entities. Most of the existing databases dealing with combustion of fuels belong in module (b)[1]. They almost always deal with individual molecules, for example heptane and are designed to reproduce results of oxidative properties such as ignition delays. They represent heroic efforts in terms of the scale of the work. A heptane combustion kinetics database can have hundreds of species and thousands of reactions. Module(c), light hydrocarbon combustion centered about GRIMECH[2] is probably the most developed in terms of completeness.. It is the area where there is the most experimental data. The PAH/SOOT databases(d)[3] are less developed. They usually begin with unsaturated compounds and are able to fit certain sets of data. There are uncertainties regarding completeness.

PRESENT EFFORT:We have concentrated on the development of the pyrolysis module. This provides inputs for the PAH/SOOT models beginning with a fuel molecule. It will extend the range of oxidation databases to richer mixtures. We focus on unimolecular processes involving the fuel radicals formed during the course of the pyrolytic process. Unimolecular processes are responsible for the breaking down of the fuel radicals to small unsaturates that are the starting point for PAH/SOOT models. The complicating factor is isomerization. This can involve the entire molecule and thus eliminates the simplicity of having only beta bond cleavage responsible for the product distribution. The high temperatures of combustion processes leads to distortions of the molecular distribution function and leads to pressure dependences. We have developed a program to account for such problems [4]. Quantitative details of true bimolecular processes under combustion conditions are much well understood. There are satisfactory correlations and rate expressions can be found in many of the databases. In the following we will use examples from our analysis of the pyrolytic decomposition of radicals derived from 1-hexene and n-butylbenzene. The former represents a product from the decomposition of normal alkanes with more than seven carbon atom. Olefins are also found in fuel mixtures. Branched aromatics are found in aviation fuels. There are interesting correlations regarding particulate formation with aromatic content in the fuel. The breakdown of unsaturated radicals involve much richer chemistry than that for alkyl radicals.

RESULTS: As much as possible we base our recommendations on experimental measurements. There are however very few direct determinations. We have taken data on related reactions, for example from measurements on the reverse process, and derived rate constants on the basis of detailed balance. There are also results that can be derived from chemical activation experiments. In other cases we rely on essentially empirical correlations, for example the effect of methyl or vinyl substitution. Where appropriate we carry out experimental studies. This involves generating the radical of interest from a suitable precursor and determining branching ratios from single pulse shock tube experiments.

For unsaturated radicals the situation is complicated by the presence of the site of unsaturation. This leads to resonance energy effects. However the 50kJ/mol resonance energy is not fully manifested in the transition state. Another complication is the presence of neophyl and homo- allylic rearrangements with the electron at the 2 position. For 1-olefinyl radicals with more than five carbon atoms there is the possibility of cyclization to form the cycloalkyl radical. It is these isomerization processes that represent the complicating factor in the decomposition of larger unsaturated radicals. We have earlier analyzed the available data on the decomposition of the three 1-pentenyl radicals [5]. In the case of 1-hexenyl, similar data does not exist. However reasonable estimates can be made on the basis of the effect of changing methyl to ethyl groupings. The most complicated situation arises from the unimolecular reactions of 1-hexenyl-6. There are three isomerization processes leading to the formation cyclohexyl, 1-hexenyl-3 and cyclopentylmethyl radicals. There has been a number of studies on the decomposition reactions of cyclohexyl radicals. At lower temperatures cyclic structures are more stable than 1-hexenyl-6 radicals. From single pulse shock tube experiments we find decomposition products of cyclohexyl, cyclopentylmethyl and 1-hexenyl-6 radicals. Together with results with cyclohexyl radicals as the starting reactant in single pulse shock tube experiments, it should be possible to give a quantitative picture of the complex reaction processes. An important aspect of this analysis is the implications on the decomposition of cyclohexyl radicals. Cyclanes are major component of the hydrocarbons in tar sands.

This leads into the analysis of the unimolecular reactions of the phenylbutyl radical. Benzylic and allylic resonance energies are virtually equal to each other. There exists experimental data on the addition of ethyl radicals to styrene. From detailed balance this leads to the rate constants for the decomposition of 1-phenylbutyl-1. 1-Phenylbutyl-2 beta bond fission will lead to the formation of allylbenzene and a methyl radical. The reverse reaction involves methyl addition to the terminal position in allylbenzene. The rate constants for this process should be very similar to those for methyl addition to the 1-olefins since the aromatic structure should not interact with the double bond in the 3 position. This rate expression is very similar to that for the analogous reaction involving methyl ejection from an alkyl radical structure. This fission is accompanied by the neophyl rearrangement that converts the secondary into a primary radical. 1-Phenylbutyl-3 beta bond fission leads to the formation of benzyl and propene. The stability of the former leads to large rate constants. There are no information on the addition of benzyl radicals to an unsaturated system. We have chosen to use the rate expression assigned to the decomposition of 1-pentenyl-5. 1-Phenylbutyl-4 decomposes by beta C-C bond leading to the formation of ethylene and 1-phenylethyl-2. Rate expression for this process is taken as equal to that for the decomposition of the larger 1-alkyl radicals since resonance energy effects are expected to be at a minimum. We have ignored the possibility of cyclization to form tetralin. The resonance energy of the aromatic ring should favor beta bond scission at least at high temperatures.

The systematic approach outlined here should assure the internal consistency of the database. This is especially important when one considers mixtures of surrogate compounds. It can also serve as a basis for automatic database generation.

REFERENCES

1. Curran, H. J., Gaffuri, P., Pitz, W. J., and Westbrook, C. K.,"A Comprehensive Modeling Study of Heptane Combustion". Comb. and Flame., 114, 149, 1998 2. Smith, P., Golden, D., Frenklach, M., Moriarty, N., Eiteneer, B., Goldenberg, M. Bowman, C., Hanson, R., Song, S., Gardiner, W., Lissianski, V., and Qin, Z.(1999). http://www.me.berkeley.edu/gri mech/. 3. Richter H, and Howard J.B., ?Formation of polycyclic aromatic hydrocarbons and their growth to soot - a review of chemical reaction pathways ?, Prog. Energ. Combust. 2000, 4, 565-608 4. Tsang, W., "A Pre-processor for the Generation of Chemical Kinetics Data for Simulations" AAIA-2001-0359, 39th AIAA Aerospace Sciences Meeting and Exhibit, Jamuary 8-11, 2001 Reno 5. Tsang, W., ?Mechanism and Rate Constants for the Decomposition of 1-Pentenyl Radicals?, J. Phys. Chem. (in press)