Article pubs.acs.org/EF
Impact of the Molecular Structure on Olefin Pyrolysis Kun Wang,† Stephanie M. Villano, and Anthony M. Dean* Chemical and Biological Engineering Department, Colorado School of Mines, Golden, Colorado 80401, United States ABSTRACT: The objective of this study is to explore the impact of the molecular structure on the rate and product distribution during olefin pyrolysis. Particular focus is on characterizing the reaction pathways, leading to the formation of molecular weight growth species. Propene, the smallest olefin that can form allylic radicals, and the three butene isomers were selected as model olefins. The experimental data were taken from several earlier studies that were conducted in a tubular flow reactor at an absolute pressure of ∼0.82 atm over a temperature range of 535−810 °C with a residence time of ∼2.4 s. The variations among the four olefins in terms of the observed conversions, generation of light products, and formation of molecular weight growth species were compared to the predictions of a fundamentally based detailed kinetic model with generally very satisfactory results. It was found that addition reactions of resonantly stabilized radicals to unsaturated products that can form unusually stable adducts, especially 1,3-butadiene and allene, are major channels for the formation of molecular weight growth products.
1. INTRODUCTION Practical transportation fuels, such as gasoline, aviation, and diesel, whether derived from conventional or alternative sources, are composed of a variety of hydrocarbons with significantly different molecular structures. The fuel structures have been shown to substantially impact combustion properties, such as octane number, soot formation tendencies, flame speed, ignition delay, etc. Substantial research has been devoted to the understanding of how the fuel structure impacts the kinetics of saturated hydrocarbons. Examples include ignition delay measurements for heptane isomers,1−4 fuel dependence for the formation of benzene,5,6 and studies of C5 and C6 linear or cyclic alkanes.7,8 Less attention has been paid to understanding structural effects on unsaturated hydrocarbons, specifically alkenes. These unsaturated hydrocarbons form a large fraction of the primary initial products from pyrolysis of alkanes, alcohol, and ethers, and their subsequent reactions have a substantial impact on the final product distributions. In particular, alkenes can undergo radical addition reactions that then lead to the formation of molecular weight growth (MWG) species that are most responsible for soot or deposit formation. Therefore, an accurate understanding for the impact of the molecular structure on the conversion of alkenes is essential in the prediction of fuel reactivity and optimization of the energy conversion processes that employ these fuels. An important feature of alkene pyrolysis is that high concentrations of resonantly stabilized free radicals (RSFRs), in particular allylic radicals, are generated. As a result of the lower energies of the allylic C−C and/or C−H bonds in olefins, allylic radicals are the likely initial dissociation products as well as being more likely to be formed by hydrogen abstraction reactions. The thermochemistry of reactions involving RSFRs is significantly different from the alkyl analogues. For example, abstraction reactions by RSFRs from alkanes will generally have higher activation energies, reflecting the substantial endothermicity of such reactions. Addition reactions to olefins will have higher barriers and form less stable adducts than the analogous alkyl radical additions. Unimolecular β-scission reactions of RSFRs will have higher © XXXX American Chemical Society
barriers. As a consequence, RSFRs are likely to accumulate to higher concentrations than these of “normal” (i.e., nonresonantly stabilized) alkyl radicals by ∼1−2 orders of magnitude at temperatures near 1000 K. Thus, even though the rate constants for the reactions of resonantly stabilized radicals (RSRs) are smaller than the analogous reactions of alkyl radicals, the actual reaction rates may be comparable. A variety of reactions involving allylic radicals have been investigated,9−13 including H-atom shifts, cyclizations, additions, recombinations, and H-abstraction reactions. One important finding is that the radical addition and recombination reactions can provide low-energy routes to the formation of higher molecular weight species. In these cases, the initially formed adducts from the addition reactions may isomerize to form cyclic radicals that may then form stable cyclic species by β-scission of H atoms or small alkyl groups. Recombination adducts may follow a similar pathway after H abstraction from the initial adduct. These stable cyclic products are important precursors for the formation of cyclopentadiene and benzene, the building blocks for polyaromatic hydrocarbons (PAHs). Among the alkenes, propene and the three butene isomers have been the focus of many studies because they are convenient to study experimentally and their small size makes them amenable to high-level electronic structure calculations. However, only a few prior studies have directly compared the decomposition kinetics of these species. Norinaga et al.14,15 studied the pyrolysis of propene as well as acetylene and ethylene at pressures of 2−15 kPa and temperatures of 1073− 1373 K. Multiple MWG species were characterized using on-/ off-line gas chromatography (GC), and a kinetic model was developed and extensively validated for hydrogen, small hydrocarbons, and aromatics. Schenk et al.16 conducted a detailed mass spectrometric and modeling study of lowpressure premixed flames of three butene isomers under fuelrich conditions. Multiple C5 species were identified and Received: March 14, 2017 Revised: April 29, 2017 Published: May 1, 2017 A
DOI: 10.1021/acs.energyfuels.7b00742 Energy Fuels XXXX, XXX, XXX−XXX
Article
Energy & Fuels
isothermal batch reactor module was used to interrogate the mechanism. In several places throughout the text, tables, and figures, we employ a chemical notation that omits the hydrogen atoms. For C3 and larger species, the radical site is indicated by “•”, a double bond is indicated by “”, a triple bond is indicated by “”, and a cyclic species is indicated by “cy”.
measured; the model provided qualitative descriptions of these products. Zhang et al.17 investigated the pyrolysis of the butene isomers in a low-pressure (∼3−11 Torr) tube reactor over a temperature range of 900−1900 K using synchrotron vacuum ultraviolet (VUV) photoionization mass spectrometry. In a recent study, Li et al.18 compared the high-pressure (10−50 atm) ignition reactivity of three butene isomers at temperatures ranging from 670 to 1350 K, with equivalence ratios of 0.5−2.0. Al-Shoaibi19,20 studied the pyrolysis of the butene isomers in a flow reactor over the temperature range of ∼550−750 °C at a pressure of ∼0.8 atm. Recently, we have conducted a series of studies aimed at improving the understanding of the mechanism for the formation of MWG species during the pyrolysis of propene21 and the three butene isomers.22,23 A detailed kinetic model was developed by a systematic application of rate constant assignments guided in a large part by rate rules based on extensive electronic structure calculations.24−28 The pressure dependence for the branching ratios of chemically activated reactions was incorporated into the mechanism.29 For each of the investigated species, the unadjusted model accurately predicted the observed conversions, production of light products, and formation of several important MWG species. Both the experiment and simulation revealed differences in the pyrolysis products of these four olefins. Most notable is that isobutene pyrolysis leads to the formation of more MWG products than the pyrolysis of the linear butene isomers and propene (shown in the Table of Contents graphic). The objective of the present work is to use this model to provide a comprehensive understanding of how the molecular structures of these alkenes impact the conversion and formation of major products and minor MWG species.
3. RESULTS AND DISCUSSION 3.1. Comparisons of Experimental Results to Model Predictions. Figure 1 compares the conversion as a function of
Figure 1. Temperature dependence for pyrolysis of propene and the three butene isomers. Initial conditions: 50 mol % olefin in N2, p of ∼0.82 atm, and hot zone residence time of ∼2.4 s (open symbols, this work; filled symbols, study by Al-Shoaibi;19,20 and solid lines, model predictions).
the temperature for propene and the three butene isomers. It also includes data from a prior study by Al-Shoaibi19,20 that were collected using a simpler version of the flow reactor that is employed here. The two data sets are very similar, attesting to the reproducibility of the measurements. 1-Butene is the most reactive species, starting to react at least 50 °C lower in temperature than the other investigated olefins. The next most reactive is 2-butene. Propene conversion appears to be comparable to isobutene at lower temperatures (≤675 °C) but becomes less reactive at higher temperatures. The slope of the 2-butene conversion curve is noticeably steeper than those of the other olefins, while the slope of the propene conversion curve is less steep and the rise time is noticeably longer at lower temperatures. The model predictions, calculated using the measured temperature profiles in the reactor, are also shown in Figure 1. In all cases, the model accurately captures the observed temperature dependence of the olefin conversions. The initiation reactions of alkenes (except ethylene) usually involve the breaking of an allylic bond, either a C−C bond (∼76 kcal/mol) or a C−H bond (∼86−88 kcal/mol). Only 1butene has the weaker C−C bond (reaction R1), explaining why the pyrolysis of this species begins at lower temperatures than the others (cf. Figure 1). Of the olefins investigated, 2butene is the only olefin that has a low-energy (∼65 kcal/mol) molecular elimination channel31,32 (R2) that competes with the bond fission reaction (reaction R3), and this explains its differently shaped conversion profile (another pathway for 2butene is isomerization to 1-butene, followed by breaking the weaker allylic C−C bond; we have examined this pathway and found it to be unimportant under these conditions22). Initiation for propene and isobutene involve breaking the stronger allylic C−H bond (reactions R4 and R5), leading to a shift in the
2. EXPERIMENTAL AND THEORETICAL METHODS In the current work, the data from our previous pyrolysis studies of propene,21 1- and 2-butene,22 and isobutene23 are compared to one another. These experiments were conducted using an atmospheric (high-altitude) pressure tubular flow reactor. The raw data sets and details of the procedure have been provided in earlier work.21−23 In general, a known flow rate of olefin was mixed with N2 and introduced into a 6 mm inner diameter heated tubular quartz reactor that was housed in an electric furnace. The axial temperature profile of the reactor was measured using a K-type thermocouple.21 The reaction gas stream was analyzed using GC. Permanent gases and light hydrocarbons were detected using a thermal conductivity detector. Light and heavy hydrocarbons were detected using a flame ionization detector that was connected in parallel to a mass spectrometer. For each olefin, the input concentration, residence time, and temperature were varied. This analysis considers the data sets where the initial olefin concentration was 50 mol % in nitrogen, the pressure was ∼0.82 atm, and the total flow rate into the reactor was 30 standard cubic centimeters per minute (sccm), which corresponds to an approximate residence time of 2.4 s in the constant-temperature region of the reactor (the exact residence time depends upon the temperature profile and extent of conversion and can be determined by the model simulations). These data sets exhibited the highest levels of conversion and the highest concentrations of higher molecular weight species. The detailed kinetic mechanism that was used in this study was recently used to describe the pyrolysis of the butene isomers.22,23 The details of the mechanism development procedure and the kinetic and thermodynamic database are provided therein. An earlier version of this model was used to describe propene pyrolysis.21 Simulations were performed using CHEMKIN-PRO.30 This program was used to generate concentration−time/temperature profiles and rates of production plots and perform sensitivity analysis. The plug flow module was used to model the experimental data sets, and the B
DOI: 10.1021/acs.energyfuels.7b00742 Energy Fuels XXXX, XXX, XXX−XXX
Article
Energy & Fuels
Figure 2. Product formation as a function of olefin conversion. Initial conditions: 50% olefin in N2, p of ∼0.82 atm, and hot zone residence time of ∼2.4 s (open symbols, this work; filled symbols, study by Al-Shoaibi;19,20 and solid lines, model predictions).
Shoaibi19,20 are also shown, and again, the agreement between the two data sets is quite good. Note that the temperatures at which each olefin reach a given degree of conversion are different (cf. Figure 1). For both methane and hydrogen, the selectivity increases in the order of 1-butene, 2-butene, propene, and isobutene and the variations with conversion and species are nicely captured by the model, with the exception that H2 is underpredicted for propene and isobutene at high conversions. Propene pyrolysis produces the most C2 products, followed by 1-butene, 2-butene, and isobutene. Again, the model describes the experimental data quite well. The C3 products for all of the olefins are grouped closely together at low conversion but begin to deviate widely at higher conversions. In all cases, there is a noticeable decrease in the slope and a significant spread in the concentrations. These trends are captured by the model predictions. The C4 product distributions vary widely and follow the trend of decreasing slope at higher conversions observed with the C3 products;
onset of pyrolysis to higher temperatures. Propene is unusual in that the initiation process also involves bimolecular reactions, including disproportionation and allyl addition reactions [the unimolecular reaction (reaction R4) only provides a minor contribution]. A detailed discussion has been provided in a previous study21 and will be briefly reviewed later in the text. CCCC → CCC• + CH3
(R1)
CCCC → CCCC + H 2
(R2)
CCCC → CCCC• + H
(R3)
CCC → CCC• + H
(R4)
CCC2 → CCC2• + H
(R5)
Figure 2 groups the product profiles as a function of olefin conversion in terms of the carbon number and compares these data to the model predictions. The prior data from AlC
DOI: 10.1021/acs.energyfuels.7b00742 Energy Fuels XXXX, XXX, XXX−XXX
Article
Energy & Fuels
was chosen to maintain a comparable conversion among the various olefins. This allows us to look at the product profiles as a function of time rather than conversion. Figure 3 shows the rates for the major decay channels for each olefin. For 1-butene, the allylic C−C bond fission channel (reaction R1) is significant at early times; for all the olefins, the allylic C−H bond fission channels are minor and are not included in the plots. For 2-butene, the molecular channel (reaction R2) is important and there is a noticeable delay in the reactions that involve radicals. Isobutene decay is dominated by secondary reactions that follow the initial dissociation (reaction R5). For propene, two disproportionation reactions provide important contributions at early times. Later on, the allyl addition to propene becomes significant. The plots show the importance of the secondary reactions during the decomposition of the different olefins. As shown in all of the subplots in Figure 3, the methyl radical and H atom play critical roles: the most important reaction is the hydrogen abstraction of the parent by the methyl radical, followed by the hydrogen abstraction of and/or addition to the parent by the H atom. The H-atom addition reaction is of special importance, because it not only breaks down the parent olefin but also produces a methyl radical to continue the reaction chain. Another important group of reactions involves resonantly stabilized allylic radicals, including the approximately thermoneutral hydrogen abstraction from the parent as well as the addition to the parent olefin. Figure 4 displays the concentration−time profiles for several selected radicals. As in Figure 3, the temperatures are chosen so that the conversions are comparable; however, some of the differences for a given radical can be attributed to the remaining slight differences in conversion. This is especially noticeable for 2-butene pyrolysis, where the production of some radicals is delayed in comparison to the other two isomers. As previously mentioned, under these conditions, olefin pyrolysis leads to the formation of high concentrations of RSRs. The most important radicals are allyl for propene and methyl allyl for the butene isomers. The two methyl allyl radical isomers are plotted together in the figure: for 1- and 2-butene pyrolysis, the primary radical formed via hydrogen abstraction is the linear 1-methyl allyl (CCCC•), while for isobutene pyrolysis, it is the branched 2-methyl allyl (CCC2•). For 1-butene pyrolysis, the allyl concentration is only slightly lower than that of 1methyl allyl. Note that the peak mole fraction of 2-methyl allyl for isobutene is approximately 3 times larger than that of 1methyl allyl for 1- and 2-butene. For all of the olefins, the cyclopentadienyl profile is shifted to later times. Its concentration is slightly higher than the methyl allyl radicals and allyl radical for the butene isomers but lower than allyl for propene. Propargyl (CCC•) is significant only for isobutene; note that it peaks much earlier than allyl, and the peak concentration is comparable to allyl. There are also remarkable differences in the predicted profiles of methyl, hydrogen, ethyl, and vinyl radicals among all of the olefins studied. Predictions of the major products are shown in Figure 5. In all cases, it is notable that the acetylene mole fractions are much lower than olefin and diolefin products. This indicates that, unlike the situation at combustion temperatures where the addition to acetylene dominates,36,37 under these relatively lowtemperature conditions, much of the MWG production is due to radical addition to olefins and/or diolefins. The most abundant olefin produced in propene pyrolysis is ethylene, which is primarily formed by the H addition to form chemically
here, a turnover in the selectivities at higher conversions becomes evident. 2-Butene pyrolysis leads to more C4 products (mostly 1,3-butadiene from reaction R2), followed by 1-butene and propene pyrolysis, with very little from isobutene pyrolysis. The observed trends are accurately predicted by the model. Figure 2 also displays the MWG products ranging from C5 to C11. The dominant products are in the C5−C7 range. More C5 products are produced from 1-butene pyrolysis; the C5 product profiles formed in 1-butene and isobutene pyrolysis show clear maxima. The propene pyrolysis data exhibit a plateau, but the 2-butene pyrolysis data continue to increase with conversion. The model predictions show similar behavior, although the isobutene pyrolysis predictions are slightly higher than measured. The C6 predictions closely track the data, showing that isobutene and propene pyrolysis produce more C6 species than 1- and 2-butene pyrolysis. Benzene is the dominant C6 product for all four olefins. Toluene is the primary C7 product, and much more is formed from isobutene pyrolysis. The predictions for propene, 2-butene, and isobutene pyrolysis closely match the data. The data scatter for 1-butene pyrolysis makes it difficult to judge whether the predicted rapid increase at low conversions is overestimated, but the predicted lower slope at higher conversions is consistent with the data. The concentrations of C8−C10 species are lower by about an order of magnitude relative to C5−C7 species. In general, isobutene and propene pyrolysis produce more of these species than 1and 2-butene pyrolysis. For the C8 products, the yield from isobutene pyrolysis is in good agreement with the model but the yields from the other three species are underpredicted, although the ordering is correct. Propene and isobutene pyrolysis produces comparable amounts of C9 products, followed by 2-butene pyrolysis and then 1-butene pyrolysis. The predictions for both 2-butene and isobutene pyrolysis are significantly underpredicted. The C10 data are quite similar to C9, although the predictions are much improved relative to C9 data. The formation of C11 and above species generally drops by another order of magnitude. Although Figure 2 groups species by carbon number, it should be noted that the model comparisons to individual species, including most MWG products, were generally quite satisfactory.21−23 These include comparisons at different dilution levels and several different residence times over a wider range of conditions, under atmospheric pressure with variation of the residence time and temperature. The model was also applied to several published pyrolysis data sets that were collected under significantly different conditions, e.g., lower pressures,14,15,17 higher temperatures,17,33−35 and faster residence times.17,33,34 In general, the model is able to satisfactorily predict these data as well. 3.2. Isothermal Batch Reactor Simulations. The comparisons of the model predictions to the extensive data sets suggest that the kinetic model captures the essential pyrolysis kinetics of these olefins. Thus, we are now in a position to use the model to provide insight into some of the observed similarities and differences. In particular, for a given level of conversion, isobutene pyrolysis leads to the formation of higher amounts of MWG species than observed for 1- and 2butene pyrolysis (cf. Table of Contents graphic). For simplicity, the analysis presented here employs isothermal batch reactor simulations. The initial concentration was set to 50 mol % in N2 at 1 atm, and the simulations were run for a duration of 5 s. The temperature for each simulation (675 °C for 1-butene, 715 °C for 2-butene, 765 °C for isobutene, and 800 °C for propene) D
DOI: 10.1021/acs.energyfuels.7b00742 Energy Fuels XXXX, XXX, XXX−XXX
Article
Energy & Fuels
Figure 3. Rate of production (ROP) analysis for the fuel major decay channels for propene and the three butene isomers. Initial conditions: 50 mol % olefin in N2, isothermal at specified T, and p = 1 atm. The temperature was adjusted to achieve similar levels of conversion.
Figure 4. Generation of major radicals during the conversion of propylene and the three butene isomers. Initial conditions: 50 mol % olefin in N2, isothermal at specified T (so that the extent of conversion is comparable), and p = 1 atm. The top row shows the important non-resonant radicals, and the bottom row shows the important resonant radicals.
activated n-propyl that then undergo a β-scission reaction (reaction R7). C3H6 + H → C2H4 + CH3
The rapid dissociation of ethyl leads to the formation of ethylene and the H atom (reaction R10) that can repeat reaction R9. Thus, 1-butene pyrolysis produces more ethylene and propene than the pyrolysis of the other two butene isomers.
(R7)
The two most abundant olefins are ethylene and propene for the pyrolysis of the butene isomers. Generally, propene is formed earlier via the chemically activated H addition to the parent (reaction R8), and the subsequent reaction of propene leads to the formation of ethylene (reaction R7). For 1-butene pyrolysis, the H-addition reaction at the non-terminal position of the double bond can form ethylene + ethyl (reaction R9).
1‐ , 2‐ , iso‐C4 H8 + H → C3H6 + CH3
(R8)
CCCC + H → C2H4 + C2H5
(R9)
C2H5 → C2H4 + H E
(R10) DOI: 10.1021/acs.energyfuels.7b00742 Energy Fuels XXXX, XXX, XXX−XXX
Article
Energy & Fuels
Figure 5. Formation of major products during the conversion of propylene and the three butene isomers. Initial conditions: 50 mol % olefin in N2, isothermal at specified T (so that the extent of conversion is comparable), and p = 1 atm.
Significantly more allene and propyne are formed in isobutene pyrolysis. Allene is primarily formed by the βscission of 2-methyl allyl (Ea ∼ 56 kcal/mol; reaction R11), which is generated from the parent. The subsequent isomerization of allene forms the more stable propyne. Allene formation in propene and 1-butene pyrolysis is primarily from the dissociation of allyl (reaction R12). This is a less efficient pathway compared to the analogous reaction of 2-methyl allyl (reaction R11) as a result of its significantly higher barrier (∼67 kcal/mol). CCC2• → aC3H4 + CH3
(R11)
CCC• → aC3H4 + H
(R12)
The three primary MWG species are 1,3-cyclopentadiene, benzene, and toluene. The pyrolysis of the three linear alkenes lead to more 1,3-cyclopentadiene, while isobutene pyrolysis leads to more benzene and toluene. The measured differences in both major products (Figure 2) and predicted radical concentrations (Figure 4) suggest that the major pathways to the production of MWG species will be different for the olefins studied. Specifically, in propene pyrolysis, ethylene and propylene are two important unsaturated species and allyl is the dominant radical; thus, the addition reactions of allyl to these two products become the primary MWG pathways. A significant distinction between the linear butenes and isobutene is that 1,3-butadiene is a significant product from the linear species, while allene is dominant for isobutene. As discussed in more detail below, the addition of RSFRs to 1,3-butadiene and allene produces adducts that are more stable (as a result of resonance), thereby becoming favorable pathways to MWG. For propene pyrolysis, the most abundant radical at early times is allyl; as it decays, 1,3-cyclopentadienyl becomes dominant. The reactions most responsible for MWG include allyl addition to propene, ethylene, and 1,3-butadiene. At very early times, the addition of allyl to the parent propene is also of special importance. Through this reaction, relatively “inert” allyl is consumed and the active H atom is produced to abstract from propene, continuing the reaction chain. The key feature of this potential energy surface (PES) is that the cyclization reactions and the subsequent β-scission reactions of the cyclic alkyl radicals are all lower in energy than the entrance channel. The addition to ethylene becomes important as conversion increases, leading to the production of cyclopentene, which,
More propyne is formed from 2-butene than 1-butene primarily as a result of the following pathway: CCCC (+R•) → CCC•C (+RH) → CCC (+CH3). The formation of 1,3-butadiene is significant for 1- and 2-butene pyrolysis. More 1,3-butadiene is formed from 2-butene than 1-butene pyrolysis as a result of the H2 molecular elimination channel (reaction R2). The decay is more pronounced at later times. Dissociation of 1-methyl allyl is the major pathway for 1,3-butadiene formation in propene and 1-butene pyrolysis (reaction R13). 1,3-Butadiene is produced at much later times in isobutene pyrolysis and is mainly formed by the isomerization of 1-butyne and 1,2-butadiene, which are generated by the recombination of propargyl (which has two resonance structures) and methyl. CCCC• → CCCC + H
(R13) F
DOI: 10.1021/acs.energyfuels.7b00742 Energy Fuels XXXX, XXX, XXX−XXX
Article
Energy & Fuels
Scheme 1. Formation of Cyclopentadiene, Benzene, and Toluene from Allyl Addition to Ethylene and Propylene in Propylene Pyrolysis
Scheme 2. Formation of Benzene and Toluene from Allyl and 1-Methyl Allyl (C4H7) Addition to 1,3-Butadiene in 1- and 2Butene Pyrolysis
Scheme 3. Formation of C7 Cyclic Species and Their Subsequent Reactions To Form C5 and C6 Species from 2-Methyl Allyl (C4H7) + Propyne/Allene (C3H4) in Isobutene Pyrolysis
after dehydrogenation, forms 1,3-cyclopentadiene. Although the concentration of 1,3-butadiene is lower, the faster addition rate constant as a result of the formation of a more stable initial adduct makes this reaction important as well. Abstraction from
the very weak C−H bond 1,3-cyclopentadiene forms cyclopentadienyl. Addition reactions of this radical to olefins are quite slow, as discussed in detail in ref 21. The primary reaction G
DOI: 10.1021/acs.energyfuels.7b00742 Energy Fuels XXXX, XXX, XXX−XXX
Article
Energy & Fuels
described above. The enthalpies of the reactants, the radical addition transition state (TS), the initially formed adduct, and the adduct well depth relative to the reactants are listed. Next, the barriers for the initial addition reaction and that for the adduct to return to reactants are listed. These are followed by the isomerization TS and the corresponding barrier for the formation of cyclic products. The difference in barriers for the two reactions of the initially formed adduct, either back to reactants or to cyclic product formation, is also shown. Finally, the overall rate constant, at 1000 K and 1 atm, for the formation of cyclic species is shown. Only the most favorable addition site is considered (i.e., the terminal carbon atom in 1,3-butadiene and the central carbon atom in allene that lead to the formation of RSRs). The reactions in Table 1 are ordered in terms of the well depth of the initially formed adduct. The first two form resonantly stabilized adducts involving a high-energy reactant (either vinyl or allene), resulting in relatively deep wells. The second set involves the formation of a resonantly stabilized adduct from a RSR, while the third set involves the formation of a non-resonantly stabilized adduct from a RSR. The first two reactions have relatively deep wells, with the result that the barrier for isomerization of the adduct is much lower than that for reversion to reactants. These reactions have relatively large rate constants for the overall process of formation of ring compounds. The rate constant for vinyl addition to 1,3butadiene is especially large; this is related to the much higher rate constant for the formation of the initial adduct, as indicated by the much lower barrier. As a result, this reaction sequence is generally quite important, even though the concentration of the vinyl radical is much lower (by 2−3 orders of magnitude) than the concentration of the allylic radicals. The next group of reactions have significantly shallower wells and higher barriers for the initial addition, typical of most cases when a RSR forms a resonantly stabilized adduct. For these particular cases, the initially formed linear adduct first isomerizes to another linear adduct prior to cyclization (note that we have included two pathways for methyl allyl addition to 1,3-butadiene to reflect the two resonance forms of methyl allyl). The much shallower wells for these reactions result in the differences in barrier height in the initial adduct, reverting to reactants versus isomerization that leads to the formation of cyclics, being much smaller than predicted for the adducts with deeper wells. As a result, the overall rate constants for the formation of cyclics are lower. The final group of reactions include those where the linear adduct is not resonantly stabilized, resulting in wells that are less deep. Generally, the rate constants for the formation of cyclics are lowest for these reactions. In these instances, the lower rate constants for cyclic formation are due to the relatively high barriers for the formation of the initial adduct.
is self-recombination, ultimately leading to the formation of naphthalene.21 For pyrolysis of the linear butenes, the addition reactions of vinyl, allyl, and 1-methyl allyl to 1,3-butadiene are important routes. The adducts formed by terminal addition to 1,3butadiene are resonantly stabilized species, thus forming deep initial wells, which can effectively lower the barriers toward product formation relative to dissociation back to reactants. For the case of vinyl addition, there is an additional advantage, i.e., the high energy of the reactant; thus, the formation of cyclic products is even more favored. The addition reactions of 2methyl allyl to the C3H4 isomers (i.e., allene and propyne) are the most important routes to MWG formation during isobutene pyrolysis. The non-terminal addition to allene, leading to a RSR, is the most important pathway. More detailed descriptions of these pathways are shown in Schemes 1−3, and the corresponding PESs are provided in our earlier studies.21−23 Another important difference between the linear and branched alkenes is that branched isobutene pyrolysis leads to significantly more propargyl radical (C3H3, CCC•), as shown in Figure 4. Propargyl is formed from hydrogen abstraction from allene and propyne, both of which are formed in higher concentrations in isobutene pyrolysis (cf. Figure 5). As documented in several prior studies, the recombination reaction of two propargyl radicals leads to benzene through a series of relatively low-energy isomerization pathways.38 It is useful to mention here that, even though the allylic radical concentration is much higher than the propargyl radical concentration, the self- or cross-recombination reactions of allylic radicals are not as important for MWG formation. This is because, in these cases, the adduct does not have a direct lowenergy isomerization pathway to form cyclic species, as shown by both experimental and theoretical studies,39−41 and is more likely to dissociate back to the reactants. Instead, a subsequent hydrogen abstraction reaction from the adduct needs to occur to allow the resulting radical to isomerize and then β-scission to form MWG products. Schemes 4 and 5 show the pathways for the self-recombination reactions of propargyl and allyl that lead to benzene formation. There are some helpful generalizations that may be used to connect the energetics of the radical addition reactions to the likelihood for the formation of MWG products. Table 1 summarizes the energetics for some of the addition reactions Scheme 4. Pathways Leading to the Formation of Benzene after the Recombination of Allyl Radicals
4. SUMMARY The objective of this study of olefin pyrolysis was to show the significant impact of the molecular structure on the reactivity as well as product distributions, especially on the MWG kinetics. A comparison of experimental data for propene and the three butene isomers to a fundamentally based detailed kinetic model resulted in generally good agreement in terms of both overall reactivity and major and minor product distributions. In all cases, RSRs are readily produced and accumulate to relatively high concentrations. The addition of RSRs to unsaturated species, both the parent and unsaturated products, are very H
DOI: 10.1021/acs.energyfuels.7b00742 Energy Fuels XXXX, XXX, XXX−XXX
Article
Energy & Fuels Scheme 5. Formation of Benzene from the Recombination of Two Propargyl Radicals
Table 1. Comparison of Energy Barriers among the Various Addition Reactions
Enthalpies of formation (kcal/mol). bApparent rate constant (cm3 mol−1 s−1). cRequires isomerization to the second linear isomer before cyclization.
a
(4) Silke, E. J.; Curran, H. J.; Simmie, J. M. The influence of fuel structure on combustion as demonstrated by the isomers of heptane: A rapid compression machine study. Proc. Combust. Inst. 2005, 30 (2), 2639−2647. (5) Zhang, H. R.; Eddings, E. G.; Sarofim, A. F.; Westbrook, C. K. Fuel-dependence for the formation of benzene. Proc. Combust. Inst. 2009, 32 (1), 377−385. (6) Hansen, N.; Kasper, T.; Yang, B.; Cool, T. A.; Li, W.; Westmoreland, P. R.; Oßwald, P.; Kohse-Höinghaus, K. Fuel-structure dependence of benzene formation processes in premixed flames fueled by C6H12 isomers. Proc. Combust. Inst. 2011, 33 (1), 585−592. (7) Kiefer, J. H.; Gupte, K. S.; Harding, L. B.; Klippenstein, S. J. Shock tube and theory investigation of cyclohexane and 1-hexene decomposition. J. Phys. Chem. A 2009, 113 (48), 13570−13583. (8) Cheng, Y.; Hu, E. J.; Deng, F. Q.; Yang, F. Y.; Zhang, Y. J.; Tang, C. L.; Huang, Z. H. Experimental and kinetic comparative study on ignition characteristics of 1-pentene and n-pentane. Fuel 2016, 172, 263−272. (9) Sabbe, M. K.; Reyniers, M.-F.; Van Speybroeck, V.; Waroquier, M.; Marin, G. B. Carbon-centered radical addition and β-scission reactions: Modeling of activation energies and prexponential factors. ChemPhysChem 2008, 9 (1), 124−140. (10) Saeys, M.; Reyniers, M.-F.; Marin, G. B.; Van Speybroeck, V.; Waroquier, M. Ab initio group contribution method for activation energies for radical additions. AIChE J. 2004, 50 (2), 426−444. (11) Saeys, M.; Reyniers, M.-F.; Van Speybroeck, V.; Waroquier, M.; Marin, G. B. Ab initio group contribution method for activation energies of hydrogen abstraction reactions. ChemPhysChem 2006, 7 (1), 188−199. (12) Matheu, D. M.; Green, W. H.; Grenda, J. M. Capturing pressure-dependence in automated mechanism generation: Reactions
important pathways to MWG. The most important pathways are the addition of these radicals to 1,3-butadiene and allene; here, RSRs are formed, leading to a much deeper adduct well. This tilts the product distribution toward the formation of cyclic products, as opposed to redissociation of the adduct out the entrance channel.
■
AUTHOR INFORMATION
Corresponding Author
*Telephone: 303-273-3643. E-mail:
[email protected]. ORCID
Kun Wang: 0000-0002-1854-3316 Present Address †
Kun Wang: Department of Mechanical Engineering, Stanford University, Stanford, California 94305, United States. Notes
The authors declare no competing financial interest.
■
REFERENCES
(1) Westbrook, C. K.; Pitz, W. J.; Curran, H. C.; Boercker, J.; Kunrath, E. Chemical kinetic modeling study of shock tube ignition of heptane isomers. Int. J. Chem. Kinet. 2001, 33 (12), 868−877. (2) Westbrook, C. K.; Pitz, W. J.; Boercker, J. E.; Curran, H. J.; Griffiths, J. F.; Mohamed, C.; Ribaucour, M. Detailed chemical kinetic reaction mechanisms for autoignition of isomers of heptane under rapid compression. Proc. Combust. Inst. 2002, 29 (1), 1311−1318. (3) Smith, J. M.; Simmie, J. M.; Curran, H. J. Autoignition of heptanes: Experiments and modeling. Int. J. Chem. Kinet. 2005, 37 (12), 728−736. I
DOI: 10.1021/acs.energyfuels.7b00742 Energy Fuels XXXX, XXX, XXX−XXX
Article
Energy & Fuels through cycloalkyl intermediates. Int. J. Chem. Kinet. 2003, 35 (3), 95− 119. (13) Xu, C.; Al-Shoaibi, A. S.; Wang, C. G.; Carstensen, H.-H.; Dean, A. M. Kinetic modeling of ethane pyrolysis at high conversion. J. Phys. Chem. A 2011, 115 (38), 10470−10490. (14) Norinaga, K.; Deutschmann, O. Detailed kinetic modeling of gas-phase reactions in the chemical vapor deposition of carbon from light hydrocarbons. Ind. Eng. Chem. Res. 2007, 46 (11), 3547−3557. (15) Norinaga, K.; Janardhanan, V. M.; Deutschmann, O. Detailed chemical kinetic modeling of pyrolysis of ethylene, acetylene, and propylene at 1073−1373 K with a plug-flow reactor model. Int. J. Chem. Kinet. 2008, 40 (4), 199−208. (16) Schenk, M.; Leon, L.; Moshammer, K.; Oßwald, P.; Zeuch, T.; Seidel, L.; Mauss, F.; Kohse-Höinghaus, K. Detailed mass spectrometric and modeling study of isomeric butene flames. Combust. Flame 2013, 160 (3), 487−503. (17) Wang, Z.; Cheng, Z.; Yuan, W.; Cai, J.; Zhang, L.; Zhang, F.; Qi, F.; Wang, J. An experimental and kinetic modeling study of cyclohexane pyrolysis at low pressure. Combust. Flame 2012, 159 (7), 2243−2253. (18) Li, Y.; Zhou, C.-W.; Somers, K. P.; Zhang, K. W.; Curran, H. J. The oxidation of 2-butene: A high pressure ignition delay, kinetic modeling study and reactivity comparison with isobutene and 1butene. Proc. Combust. Inst. 2017, 36 (1), 403−411. (19) Al-Shoaibi, A. S. Experimental and modeling analysis of hydrocarbon pyrolysis. Ph.D. Thesis, Colorado School of Mines, Golden, CO, 2008. (20) Al-Shoaibi, A. S.; Dean, A. M. Kinetic analysis of C4 alkane and alkene pyrolysis: Implications for SOFC operation. J. Fuel Cell Sci. Technol. 2010, 7 (4), 041015. (21) Wang, K.; Villano, S. M.; Dean, A. M. Fundamentally-based kinetic model for propene pyrolysis. Combust. Flame 2015, 162, 4456− 4470. (22) Wang, K.; Villano, S. M.; Dean, A. M. Experimental and kinetic modeling study of butene isomer pyrolysis: Part I. 1- and 2-Butene. Combust. Flame 2016, 173, 347−369. (23) Wang, K.; Villano, S. M.; Dean, A. M. Experimental and kinetic modeling study of butene isomer pyrolysis: Part II. IsoButene. Combust. Flame 2017, 176, 23−37. (24) Carstensen, H.-H.; Dean, A. M. Rate constant rules for the automated generation of gas-phase reaction mechanisms. J. Phys. Chem. A 2009, 113 (2), 367−380. (25) Wang, K.; Villano, S. M.; Dean, A. M. Reactivity−structurebased rate estimation rules for alkyl radical H atom shift and alkenyl radical cycloaddition reactions. J. Phys. Chem. A 2015, 119 (28), 7205−7221. (26) Wang, K.; Villano, S. M.; Dean, A. M. Reactions of allylic radicals that impact molecular weight growth kinetics. Phys. Chem. Chem. Phys. 2015, 17 (9), 6255−6273. (27) Wang, K.; Villano, S. M.; Dean, A. M. The impact of resonance stabilization on the intramolecular hydrogen-atom shift reactions of hydrocarbon radicals. ChemPhysChem 2015, 16 (12), 2635−2645. (28) Wang, K.; Villano, S. M.; Dean, A. M. Ab initio study of the influence of resonance stabilization on intramolecular ring closure reactions of hydrocarbon radicals. Phys. Chem. Chem. Phys. 2016, 18 (12), 8437−8452. (29) Chang, A. Y.; Bozzelli, J. W.; Dean, A. M. Kinetic analysis of complex chemical activation and unimolecular dissociation reactions using QRRK theory and the modified strong collision approximation. Z. Phys. Chem. 2000, 214 (11), 1533−1568. (30) Reaction Design. CHEMKIN-PRO; Reaction Design: San Diego, CA, 2008. (31) Alfassi, Z. B.; Golden, D. M.; Benson, S. W. The very lowpressure dehydrogenation of cis-2-butene. The activation energy for 1,4-H2 elimination. Int. J. Chem. Kinet. 1973, 5 (6), 991−1000. (32) Masson, D.; Richard, C.; Martin, R. Thermal isomerization of cis- or trans-2-butene: The unimolecular elimination of hydrogen from cis-2-butene around 500 °C. Int. J. Chem. Kinet. 1976, 8 (1), 37−44.
(33) Davis, S. G.; Law, C. K.; Wang, H. Propene pyrolysis and oxidation kinetics in a flow reactor and laminar flames. Combust. Flame 1999, 119 (4), 375−399. (34) Yasunaga, K.; Kuraguchi, Y.; Ikeuchi, R.; Masaoka, H.; Takahashi, O.; Koike, T.; Hidaka, Y. Shock tube and modeling study of isobutene pyrolysis and oxidation. Proc. Combust. Inst. 2009, 32 (1), 453−460. (35) Held, T. The oxidation of methanol, isobutene and methyl tertiary-butyl ether. Ph.D. Thesis, Princeton University, Princeton, NJ, 1993. (36) Wang, H.; Frenklach, M. A detailed kinetic modeling study of aromatics formation in laminar premixed acetylene and ethylene flames. Combust. Flame 1997, 110 (1−2), 173−221. (37) Richter, H.; Howard, J. B. Formation of polycyclic aromatic hydrocarbons and their growth to soota review of chemical reaction pathways. Prog. Energy Combust. Sci. 2000, 26 (4−6), 565−608. (38) Miller, J. A.; Klippenstein, S. J. The Recombination of propargyl radicals and other reactions on a C6H6 potential. J. Phys. Chem. A 2003, 107 (39), 7783−7799. (39) Georgievskii, Y.; Miller, J. A.; Klippenstein, S. J. Association rate constants for reactions between resonance-stabilized radicals: C3H3 + C3H3, C3H3 + C3H5, and C3H5 + C3H5. Phys. Chem. Chem. Phys. 2007, 9, 4259−4268. (40) Selby, T. M.; Meloni, G.; Goulay, F.; Leone, S. R.; Fahr, A.; Taatjes, C. A.; Osborn, D. L. Synchrotron photoionization mass spectrometry measurements of kinetics and product formation in the allyl radical (H2CCHCH2) self-reaction. J. Phys. Chem. A 2008, 112 (39), 9366−9373. (41) Matsugi, A.; Suma, K.; Miyoshi, A. Kinetics and mechanisms of the allyl + allyl and allyl + propargyl recombination reactions. J. Phys. Chem. A 2011, 115 (26), 7610−7624.
J
DOI: 10.1021/acs.energyfuels.7b00742 Energy Fuels XXXX, XXX, XXX−XXX
