Detailed experimental and kinetic modeling study of cyclopentadiene

Feb 7, 2018 - A combined experimental and kinetic modeling study is presented to improve the understanding of the formation of polycyclic aromatic hyd...
1 downloads 11 Views 6MB Size
Article pubs.acs.org/EF

Cite This: Energy Fuels XXXX, XXX, XXX−XXX

Detailed Experimental and Kinetic Modeling Study of Cyclopentadiene Pyrolysis in the Presence of Ethene Alexander J. Vervust,† Marko R. Djokic,† Shamel S. Merchant,‡ Hans-Heinrich Carstensen,† Alan E. Long,§ Guy B. Marin,† William H. Green,§ and Kevin M. Van Geem*,† †

Laboratory for Chemical Technology, Ghent University, Technologiepark 914, B-9052 Gent, Belgium ExxonMobil Research and Engineering, 1545 Route US 22 East, Annandale, New Jersey 08801, United States § Department of Chemical Engineering, Massachusetts Institute of Technology, 77 Massachusetts Ave., Cambridge, Massachusetts 02139, United States ‡

S Supporting Information *

ABSTRACT: A combined experimental and kinetic modeling study is presented to improve the understanding of the formation of polycyclic aromatic hydrocarbons at pyrolysis conditions. The copyrolysis of cyclopentadiene (CPD) and ethene was studied in a continuous flow tubular reactor at a pressure of 0.17 MPa and a dilution of 1 mol CPD/1 mol ethene/10 mol N2. The temperature was varied from 873 to 1163 K, resulting in cyclopentadiene conversions between 1 and 92%. Using an automated reaction network generator, RMG, we present an elementary step kinetic model for CPD pyrolysis that accurately predicts the initial formation of aromatic products. The model is able to reproduce the product yields measured during the pyrolysis of pure cyclopentadiene and the copyrolysis of cyclopentadiene and ethene. The addition of ethene as coreactant increases the benzene and toluene selectivity. In the absence of ethene, benzene formation is initiated by addition of a cyclopentadienyl radical to cyclopentadiene, following a complicated series of isomerizations and loss of a butadienyl radical. In the presence of ethene, the main pathway for the formation of benzene + CH3 shifts to ethene + cyclopentadiene. Toluene formation is initiated by vinyl radical addition to cyclopentadiene. Without the addition of ethene, vinyl radicals are mainly formed by hydrogen radical addition to ethyne. When ethene is added as coreactant, vinyl radical production happens via hydrogen abstraction from ethene.

1. INTRODUCTION Polycyclic aromatic hydrocarbons (PAHs) are hydrocarbons consisting of multiple aromatic rings. They are mainly produced during combustion of fuels and pyrolysis processes in oil refining. PAHs are known to be soot precursors, are carcinogenic, and contribute to global warming.1−5 Furthermore, PAHs are known coke precursors, which lead to coke formation in the reactor coils of steam crackers.6 Steam cracking is a pyrolytic process during which feedstocks containing mainly hydrocarbons are exposed to temperatures from 900 to 1200 K for short residence times of 0.1 to 0.5 s. At these conditions the feed is mainly converted into light olefins (i.e., ethene and propene) and monoaromatics (i.e., benzene, toluene, and xylene). Coke formation on the inner wall of the reactor coils results in decreased thermal efficiency and an increased pressure drop, having a detrimental effect on the economics of the process. The need for periodic shut-down of the steam cracker unit to perform decoking operations further increases the costs associated with coke formation during steam cracking. Knowledge of PAHs formation is thus not only valuable for understanding the formation of soot and reduce the amount of pollutants in the atmosphere but also is an important step in understanding the coke formation process in steam crackers. At pyrolysis conditions, resonance-stabilized radicals such as allyl, propargyl, and cyclopentadienyl are believed to play an important role in the formation of aromatics.7−11 Compared to regular radicals, these radicals have long lifetimes and can © XXXX American Chemical Society

therefore accumulate to relatively high concentrations. Important consumption routes for these resonance-stabilized radicals are recombination reactions with other radicals, leading to molecular weight growth. For instance, the self-recombination reaction of propargyl radicals has been shown to lead to benzene.9,12,13 Recombination of an allyl and a propargyl radical and the self-recombination of allyl radicals also leads to benzene formation.7,11,12,14,15 Similarly, naphthalene can be formed by recombination of propargyl and benzyl radicals.16 The cyclopentadienyl radical has been related with the formation of high amounts of mono- and polycyclic aromatics.17,18 As cyclopentadiene (CPD) is easily converted to cyclopentadienyl (CPDyl), the pyrolysis of CPD can provide important information on PAH formation from cyclopentadienyl radicals. A number of researchers have investigated the pyrolysis of CPD and shown the importance of cyclopentadienic intermediates in the formation of the initial aromatics.19−25 Several recent continuous flow reactor experiments have investigated the product distribution obtained in the thermal decomposition of CPD. Butler and Glassman26 measured a wide range of products including expected intermediates such as dihydronaphthalenes and methylcyclopentadienyls in atmospheric pressure experiments between 1000 and 1200 K using Received: November 17, 2017 Revised: February 7, 2018 Published: February 7, 2018 A

DOI: 10.1021/acs.energyfuels.7b03560 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels highly diluted CPD/nitrogen mixtures. Kim et al.27 expanded the CPD decomposition database to the low-temperature region (550−950 K) and confirmed benzene, naphthalene, and indene as the major products at their conditions (0.7% molar CPD in N2). In the most recent study, Djokic et al.28 investigated CPD pyrolysis at higher fuel concentrations (4% and 16.7% molar CPD in N2) and demonstrated the added value of online GC × GC-FID/(TOF-MS) analytics, enabling the identification and accurate quantification of PAH products such as anthracene, phenanthrene, and fluorene. All currently available experimental data sets on CPD pyrolysis in continuous flow reactors use pure CPD as the feed. The current work extends the available experimental database by performing CPD/ethene copyrolysis experiments. Ethene, alongside propene, is one of the major products in steam cracking, and CPD is easily formed as intermediate species (e.g., through the reactions of allyl + ethene11,29,30 and allyl + ethyne),31 followed by subsequent H losses. High CPD formation rates are also to be expected if renewable feedstock sources containing large amounts of oxygenates32 are used, because molecules such as phenol,21,22,33 anisole,34 and 2,5dimethylfuran35,36 are well-known precursors of CPD or CPDyl radicals. The reactivity of the cyclopentadienyl radical has been the subject of many studies. Self-recombination of cyclopentadienyl radicals has been found to lead to naphthalene formation.37−41 Pathways to indene and benzene, initiated by the addition of cyclopentadienyl to cyclopentadiene, have been reported.42−44 An alternative pathway to indene via reaction of CPDyl with acetylene has also been documented.45 The pathway to benzene via CPDyl + CH3 has been thoroughly elucidated.37,46 Pathways via CPD + C2H4, passing through norbornene, have been suggested.47,48 However, no direct evidence of this pathway through kinetic studies or investigation of the potential energy surface is available. To this date, no study has been able to corroborate to what extent each of the proposed elementary steps are contributing to the formation of PAHs during CPD pyrolysis. Although up-to-date kinetic models generally contain some CPD chemistry that leads to the formation of aromatic products, most of these were neither specifically developed nor tested for CPD pyrolysis applications. The kinetic mechanism developed by Wang et al.,49 POLIMI_1412,50 and Herbinet et al.51 are exceptions as they contain a wider range of reactions related to aromatics formation from CPD. Wang et al.49 extended the ethane pyrolysis model by Xu et al.52 to properly describe propene pyrolysis experiments. It contains reactions for the formation of the primary aromatic products (i.e., benzene, toluene, styrene, indene, and naphthalene from CPD) and has been shown to perform well for cyclopentane pyrolysis53 and therefore might be able to predict the CPD pyrolysis chemistry as well. POLIMI_141250 is a modular kinetic model, which is continuously extended toward new molecules. It has quite successfully been applied to the experimental data of Djokic et al.28 and thus is expected to describe the current CPD/ethene copyrolysis well, too. A distinct difference from the Wang model is that the POLIMI_141250 model contains chemistry for the formation of larger PAHs beyond naphthalene up to C20 species. The Herbinet et al.51 model distinguishes itself from the models discussed above in the exclusive use of elementary step reactions to describe the formation of the primary aromatics from CPD. It was developed for cyclopentene pyrolysis and is

therefore also potentially suited to predict CPD pyrolysis. The Herbinet et al.51 model, however, does not include all reactions relevant for CPD/ethene copyrolysis. Developing a kinetic model accounting for all possible reactions involved in the formation of PAHs from the pyrolysis of CPD manually is time-consuming and error-prone. Automated kinetic model generators, such as the Reaction Mechanism Generator (RMG)54 or Genesys55 are helpful tools when studying these complex systems. These tools automate many steps in the kinetic model generation process. However, some expert user knowledge is still required to be able to generate an accurate kinetic model, and some challenges, such as data scarcity, still need to be overcome before automatic kinetic model generators will be useable by a broader public.56 The objective of the present study is 4-fold. First, the CPD/ ethene copyrolysis is studied in a plug-flow reactor at conditions that lead to CPD consumptions between 1 and 92%. Ethene was chosen as coreactant to assess its effect on CPD pyrolysis and improve the understanding of the chemistry that leads to PAH and coke formation in steam crackers. Second, literature mechanisms thought to be able to predict the experimental data of this work and also of the earlier work by Djokic et al.28 are tested. Third, a new detailed elementary-step kinetic model is developed with the RMG tool, and it is shown that this kinetic model gives improved predictions of the available experimental data. Finally, reaction path analyses based on the new kinetic model are used to highlight the most important CPD consumption pathways and the effect of ethene in the copyrolysis experiments.

2. METHODOLOGY 2.1. Experimental Procedure. The CPD/ethene copyrolysis experiments were conducted using a continuous flow reactor, which has been described in detail in previous papers.28,35 Therefore, only the specifics related to this work are mentioned here. A detailed description of the setup is given in the Supporting Information. The setup consists of three main sections: a feed section, reactor section, and analysis section. Prior to the pyrolysis, dicyclopentadiene (Sigma-Aldrich 99+% purity) is melted (307 K) and fed to an evaporator kept at 473 K. This is 20 K above the boiling point of dicyclopentadiene and sufficient to convert dicyclopentadiene completely to CPD as shown by Kim et al.27 The coreactant ethene is mixed with N2 and heated to the same temperature. Both the evaporator and subsequent mixer are electrically heated and filled with quartz beads, assuring a constant feed and uniform mixing. For the experiments presented in this work, a 1:1 molar ratio of CPD to ethene was used. The dilution was set to 1 mol CPD/10 mol N2, while the temperature was varied between 873 and 1163 K. Flow rates were chosen in order to obtain CPD conversions ranging from a few percent up to 92%, corresponding to residence times between 300 and 400 ms. A summary of the operating conditions is given in Table 1. Products were analyzed using two gas chromatographs: (a) a refinery gas analyzer (RGA) and (b) a GC × GC. The latter is equipped with both a flame ionization detector (FID) and timeof-flight mass-spectrometer (TOF-MS), allowing both qualitative and quantitative analysis of the entire product stream, from methane to PAHs. The RGA detects all permanent gases such as N2 and H2 but also helps with an additional analysis of the C1−C4 cut. Response factors of all permanent gases and light hydrocarbons (C1−C4) were determined by means of a B

DOI: 10.1021/acs.energyfuels.7b03560 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

estimated by RMG using a Benson type group contribution method.67,68 To estimate the thermodynamic parameters of species with several resonance forms by group additivity, the thermodynamic parameters are calculated for each resonance structure of the species using group additivity, and RMG then selects the thermodynamic parameters of the resonance structure with the most stable enthalpy to represent the thermodynamic properties of the species. The kinetic parameters of reactions that are not available in the RMG database are determined from a hierarchical tree of rate estimation rules, each associated with its reaction family. Rate rules are defined for sets of functional group-based reactive sites through temperature-dependent rate coefficients described by the modified Arrhenius expressions. To estimate the kinetic rate parameters for a reaction, the most specific functional groups, describing that reaction, are used to match the rate rules in the database. For reversible reactions the forward reaction rate coefficient is estimated using rate rules, while the reverse rate is determined as the quotient of the rate of the forward reaction and the equilibrium coefficient to ensure thermodynamic consistency. More details describing the model generation in RMG can be found in the work of Van Geem et al.,69 Harper et al.,70 and most recently in Gao et al.54 Parameters of important reactions in the kinetic model are refined using quantum chemistry calculations. The resulting more accurate parameters are then added to the RMG database to improve future model predictions for both the current molecule and other similar molecules. The thermodynamic properties of important species and transition states of crucial reactions were calculated at the CBS-QB3 level71,72 using Gaussian 03.73 Bond additivity corrections were added to the calculated enthalpies of formation in order to improve the agreement between the CBS-QB3 calculated results and experimental enthalpies. These empirical corrections were taken from the work of Sabbe et al.74 and amount to −0.89, 0.84, −1.78, and −3.97 kJ mol−1 for a C−H, C−C, CC, and CC bond, respectively. Rate coefficients were calculated at the high pressure limit according to transition state theory

Table 1. Summary of the Experimental Conditions for 0.17 MPa, a Dilution of 1 mol CPD/1 mol Ethene/10 mol N2 and F0,CPD = 2.06 × 10−4 mol/s conditions temperature (K) CPD conversion (mol %) ethene conversion (mol %)

873

923

973

1023

1073

1123

1163

1.2

1.8

5.0

15.6

37.8

71.6

92.4

0.3

0.6

1.7

3.1

2.9

11.3

18.3

gaseous calibration mixture (Air Liquide, Belgium). The response factors of all C5+ hydrocarbons were estimated using the effective carbon number concept, relative to methane.57 Replicate experiments were performed in order to calculate the standard error, which was less than 10%. The mass balances closed within 5%. All component weight fractions were normalized to sum up to 100 wt %, enabling the straightforward interpretations of the results, as well as comparison with the modeling results. The molar C to H ratio in the feed was wellconserved in the reactor effluent, with a maximum deviation of 1.6% at 1073 K. 2.2. Kinetic Model Generation. The reaction mechanism of the kinetic model for the pyrolysis of cyclopentadiene consists of two parts. The first part is a submechanism describing the reactions of the C1−C4 species. The second part of the reaction mechanism describes the cyclopentadiene consumption and aromatics formation. The kinetic model was generated without adjusting any parameter to improve the agreement with experimental data. The complete kinetic model (CHEMKIN58 format, 461 species, 6334 reactions) is provided in the Supporting Information. Important reactions included in the kinetic model are discussed later. This section explains how the kinetic model was constructed. The AramcoMech2.059−65 mechanism has been employed to describe the reactions of the C1−C4 species. Even though the AramcoMech2.0 was developed primarily for combustion, it is also suited to model pyrolysis since pyrolytic reactions constitute an important part of combustion chemistry. AramcoMech2.0 has been thoroughly validated against a large set of experimental data thus justifying its use for the description of the base C1−C4 chemistry. To simplify the final mechanism, the oxygenates have been removed from the AramcoMech2.0 for use in the current kinetic model. The elementary steps contributing to the cyclopentadiene consumption and aromatics formation have been determined using the automatic reaction mechanism generator (RMG).54 Given the initial conditions, RMG automatically generates all reactions of the feed molecules (and their associated species), based on a set of 45 reaction families and libraries of reactions. It then selects the important reactions to include in the model, judging importance by reaction rates.66 This process is iterated until no more reactions are found that have rates greater than a specified tolerance. The thermochemical parameters of the AramcoMech2.0 are preserved. For all the other species, the thermochemical parameters of RMG are used. These thermochemical data are obtained from the RMG database whenever possible. Thermodynamic parameters for most of the important species in the present kinetic model are present in the RMG database. Any missing values for thermodynamic parameters are

k∞(T ) = κ(T )

q‡ kBT e−ΔE0 / RT n h ∏i = 1 qreactant − i

(1)

with k∞ the high-pressure limit rate coefficient, kB the Boltzmann constant, T the temperature, h the Planck constant, q the partition function per volume, ΔE0 the zero-point energy corrected energy barrier, and R the universal gas constant. The tunneling coefficient κ(T) was calculated using the Eckart method.75 Internal rotations of key components were treated as 1D hindered rotors according to the formalism described in Pitzer et al.76 Barriers for the most sensitive reactions were calculated at the RCCSD(T)-F12a level of theory using the double-ζ basis set (cc-vdz) on geometries obtained with the B3LYP/6-311G(2d,d,p) functional method. The RCCSD(T)F12 calculations were performed using Molpro.77 The optimized geometries for all species studied here are readily available in PES exploratory studies as referenced. Unless stated otherwise, the calculated high pressure limit rate coefficients are used for reactions estimated by RMG. But many of the smallmolecule reactions in AramcoMech2.0 and in the RMG database include pressure-dependence, derived either from experimental data or from master equation calculations. C

DOI: 10.1021/acs.energyfuels.7b03560 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels Table 2. Overview of the Available Experimental Data on CPD Pyrolysis in Flow Reactors conditions

CPD Kim et al.27

CPD Butler and Glassman26

CPD Djokic et al.28

CPD + ethene present work

DCPD purity (%) temperature (K) pressure (MPa) residence time (ms)a FCPD (mg/s) FEthene (mg/s) FN2 (mg/s) dilution (mol N2/mol CPD) CPD conversion (mol %) ethene conversion (mol %)

99.9 823−1223 0.1 3000 0.125 0 7.5 142 0−100 −

95 1106−1202 0.1 0−125b 0.008−0.028 0 3.86 324−958 12.8−62.3 −

99+ 873−1123 0.17 300−400 27/6 0 53/62 5/24 2.0−83.8/2.0−65.6 −

99+ 873−1163 0.17 300−400 13.6 5.8 57.2 10 1.2−92.3 0.3−18.3

Defined as the ratio of the reactor volume and the total volumetric flow rate at inlet conditions. bButler and Glassman sampled the product stream at different positions along the test section.

a

Figure 1. GC × GC-FID chromatograms of the online sampled CPD-ethene copyrolysis effluent at 923 K (top) and 1073 K (bottom), 0.17 MPa, and a dilution of 1 mol ethene/1 mol CPD/10 mol N2.

3. RESULTS AND DISCUSSION

CPD/1 mol ethene/10 mol N2. The CPD dilution in the copyrolysis experiments is thus twice that of Djokic et al.28 GC × GC-FID chromatograms of the online sampled product streams are shown in Figure 1. The copyrolysis of CPD and ethene produces a complex product spectrum with strongly temperature-dependent compositions. At 923 K, only a small number of pyrolysis products are detected, such as indene, naphthalene, benzene, toluene, and cyclopentene. In contrast to the data obtained at 923 K, the reactor effluent at 1073 K is

3.1. Co-Pyrolysis of Cyclopentadiene and Ethene. Table 2 gives an overview of the main features of the experimental data available in literature and the new experiments performed as part of this work. Djokic et al.28 studied the pyrolysis of CPD in the same continuous flow reactor as used in this work. The same molar feed to N2 ratio as was used by Djokic et al.28 for the low dilution experiments is used for the copyrolysis of CPD and ethene, resulting in a dilution of 1 mol D

DOI: 10.1021/acs.energyfuels.7b03560 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

Figure 2. Experimental and simulated yields of the main reaction products during the copyrolysis of CPD and ethene (0.17 MPa, dilution 1 mol ethene/1 mol CPD/10 mol N2, F0,CPD = 13.6 mg/s). Symbols, experimental data; dashed lines, model predictions with the POLIMI_141250 model; dotted lines, model predictions with the Wang et al.49 model; dash-dotted lines, model predictions with the Herbinet et al.51 model. See Supporting Information for a detailed overview of the experimental data at all studied conditions.

significantly more complex, containing on the order of one hundred different compounds. While more than 37% of CPD is consumed, only 3% of the ethene has reacted at 1073 K at the specific conditions. At the highest temperature, 92% of CPD is converted, yielding almost 44 wt % of PAHs with a carbon number up to C19. At 1163 K, roughly 18% of the ethene fed to the reactor is consumed. Yields of the major product species as a function of the temperature are shown in Figure 2. Detailed compositions measured at all studied conditions are given in the Supporting Information. In Table 3, the product selectivity at 1073 and 1123 K for the low dilution CPD pyrolysis from Djokic et al.28 and for the copyrolysis of CPD and ethene from this work are compared. A comparison of CPD conversion and product selectivity of the major products for all investigated temperatures can be found in the Supporting Information. Both data sets are obtained for similar residence times. The CPD conversion at 1073 K is only 37.8% for the copyrolysis as opposed to 61.7% for the pure pyrolysis. At 1123 K, the CPD conversion is still lower in the ethene copyrolysis experiments (71.6%) compared to that in the pure pyrolysis case (83.8%). However, the difference in conversion is less pronounced than at 1073 K. At 1123 K, 11.3% of ethene has been converted as opposed to 2.9% at 1073 K. The difference in CPD conversion can be attributed to the higher CPD dilution of the copyrolysis experiments. Comparison of the Djokic et al.28 data sets at low (5 mol N2/ mol CPD) and high (24 mol N2/mol CPD) dilutions reveals the same effect of dilution on the CPD conversion. At every studied temperature, the CPD conversion is lower for the high

Table 3. Comparison of Experimental Results for the Pyrolysis of CPD with Those of the Copyrolysis of CPD and Ethene conditions temperature (K) CPD conversion (mol %) ethene conversion (mol %)

CPD (Djokic et al.28 low dilution) 1073 61.7 −

selectivity (g product/100 g reacted feeda) permanent gases H2 1.33 CH4 4.38 alkenes ethene 1.13 cyclopentene 0.96 monoaromatics benzene 6.61 toluene 1.12 styrene 1.17 polyaromatics indene 17.31 naphthalene 28.05 larger PAHs 31.00 total aromatics

85.25

CPD + ethene (present work)

1123 83.8 −

1073 37.8 2.9

1123 71.6 11.3

1.47 4.11

0.73 4.14

1.29 5.55

1.47 0.07

− 1.10

− 0.19

6.43 1.65 1.10

10.15 1.58 0.62

16.00 3.70 1.72

11.48 25.34 44.43

17.66 26.34 30.34

14.26 31.28 19.64

90.44

86.70

86.60

a The feed is CPD in the work of Djokic et al.28 and CPD + ethene in the present work.

E

DOI: 10.1021/acs.energyfuels.7b03560 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

Figure 3. Reaction path analysis based on the current kinetic model for the formation of the primary aromatic products during the pyrolysis of CPD, with a dilution of 1 mol CPD/5 mol N2 at 1123 K (red and italic), and the copyrolysis of CPD and ethene, with a dilution of 1 mol CPD/1 mol ethene/10 mol N2 at 1123 K (blue and underlined) and 1023 K (green and bold), at 0.17 MPa and a CPD conversion of 10%. Percentages on a reaction path represent the reaction rate relative to the total CPD consumption rate. All reactions are reversible, arrows represent the net flux. The primary aromatic products and CPD are indicated by a shaded background. See the Supporting Information for the complete kinetic model.

selectivity is obtained in the copyrolysis experiments than in the pure CPD experiments. The increase in small aromatic species is balanced by a substantial reduction of larger PAHs (44.4% → 19.6%) in the copyrolysis experiments. There are several reasons for the lower selectivity to larger PAHs for the copyrolysis at 1123 K. First there is the effect of ethene. The ratio of the selectivity of the primary aromatic products to the selectivity of the larger PAHs decreases from 54%/31% at 1073 K to 46%/44% at 1123 K for the CPD pyrolysis. The opposite trend is observed for the CPD/ethene copyrolysis. The ratio increases from 56%/30% at 1073 K to 67%/20% at 1123 K. At 1123 K, a larger amount of ethene has been converted than at 1073 K, 11.3% at 1123 K versus 2.9% at 1073 K. Ethene thus considerably decreases the selectivity of larger PAHs in favor of the primary aromatic products. Second, there is the dilution effect. Molecular weight growth to larger PAHs requires a series of bimolecular reactions, which are hindered at higher dilutions. The higher dilution for the copyrolysis also explains the higher selectivity to the primary aromatic products as their consumption to form larger PAHs is hampered. This is most apparent for the naphthalene (25% → 31%) and indene (11% → 14%) selectivity. Besides the presence of ethene and CPD dilution there is the effect of the molar H to C ratio. The molar H to C ratio is 1.2 for the CPD pyrolysis and 1.43 for the CPD/ethene copyrolysis. As the

CPD dilution experiments than for the low CPD dilution experiments. Since the main consumption channels of CPD are bimolecular in nature (i.e., addition/recombination reactions involving CPDyl radicals), the CPD conversion decreases with higher dilution. The difference in conversion is further reduced moving from 1073 to 1123 K since more ethene reacts with CPD at 1123 K. At 1073 K, the product selectivity for the pyrolysis of CPD and the copyrolysis of CPD and ethene are fairly similar, except for the selectivity of the monoaromatics. The benzene selectivity is significantly higher for the copyrolysis (10.15%) than for the pure CPD pyrolysis (6.61%). The toluene selectivity is also higher for the copyrolysis (1.58% vs 1.12%). The styrene selectivity of 0.62% for the copyrolysis case is lower than that for the pure CPD pyrolysis (1.17%). The product selectivity for the CPD pyrolysis and CPD/ ethene copyrolysis differ significantly at 1123 K. The methane selectivity and the selectivity toward the primary aromatic products benzene, toluene, styrene, indene, and naphthalene are higher for the copyrolysis of CPD and ethene. The largest relative increases in selectivity are observed for benzene (6.4% → 16%) and toluene (1.65% → 3.7%). The temperaturedependence of the styrene selectivity is interesting because its selectivity was similar for pure and copyrolysis at 1073 K, but at 50 K higher temperatures, a significantly higher styrene F

DOI: 10.1021/acs.energyfuels.7b03560 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels formation of PAHs requires the loss of H atoms, the larger molar H to C ratio for the copyrolysis also contributes to the lower selectivity to larger PAHs. The large increase in benzene and toluene selectivity resulting from the addition of ethene as coreactant, hint at the existence of reaction pathways to benzene + CH3 and toluene via reaction of CPD with ethene. Reaction of CPD with ethene to form benzene requires the loss of a methyl radical, explaining the higher methane selectivity for the copyrolysis than for the pure CPD pyrolysis at 1123 K. An accurate kinetic model of the CPD/ethene copyrolysis is required to fully understand the effects of ethene on CPD pyrolysis. A comparison of experimental results and model predictions by the Wang et al.,49 POLIMI_1412,50 and Herbinet et al.51 models for CPD/ethene copyrolysis are shown in Figure 2. The Wang et al.49 model performs best at predicting the copyrolysis experiments. It accurately predicts the CPD conversion and benzene and naphthalene yields. However, the indene yield is overpredicted above 1023 K and toluene and styrene yields are underpredicted. Therefore, a new kinetic model is developed in this study, which is able to describe the formation of the primary aromatic products of the CPD/ethene copyrolysis. 3.2. Kinetic Model Description. The developed kinetic model comprises 6334 reactions between 461 species. 417 reactions and 73 species originate from the AramcoMech2.0 and 5917 reactions and 388 species have been generated by RMG. The high reactions to species ratio is caused by 4805 hydrogen abstraction reactions, generated by RMG. The algorithm of RMG tries to ensure that the relevant species and reactions for CPD pyrolysis are included in the final kinetic model. Not all species and reactions included in the kinetic model will be required; however, generating a reduced kinetic model is beyond the scope of this study. In this section, the reactions included in the kinetic model which are responsible for the formation of the primary aromatic products of CPD pyrolysis, namely naphthalene, indene, benzene, toluene, and styrene, are discussed as well as the CPD decomposition reactions. A summary of the kinetic parameters of the most important reactions, shown in Figure 3, as well as the complete kinetic model in CHEMKIN58 format can be found in the Supporting Information. In the kinetic model, naphthalene formation is initiated by self-recombination of CPDyl radicals. The C10H10 and C10H9 potential energy surfaces (PES), relevant for the formation of naphthalene, have recently been re-evaluated by Long et al.41 to determine pressure-dependent rate expressions. This data is included in the current CPD pyrolysis mechanism. Figure 4 shows the reactions responsible for naphthalene formation that have been included in the kinetic model. Similar isomers, namely the azulanes, azulanyls, fulvalanes, and fulvalanyls, have been grouped in Figure 4. The isomers within each group can easily interconvert due to the relatively low energy barriers for the hydrogen shift reactions.41 Naphthalene is formed via two parallel pathways in the kinetic model, namely the fulvalanyl pathway (rightmost pathway) and the azulanyl pathway (leftmost pathway). Both pathways are initiated by the selfrecombination of CPDyl radicals, forming a highly excited fulvalane molecule that can either stabilize through bath gas collisions, isomerize to azulane, or directly decompose to fulvalanyl and a H atom. Collisionally stabilized fulvalane forms a fulvalanyl radical by loss of an H atom. The fulvalanyl radical either directly loses a H atom to form fulvalene or it isomerizes to the dihydro-naphtyl radical which then loses a H atom to

Figure 4. Reactions on C10H10 and C10H9 potential energy surfaces, responsible for naphthalene formation, included in the kinetic model. All reactions are reversible, arrows represent the net flux.

form naphthalene. Analogously the azulanyl radical is formed from azulane by H loss and it subsequently either forms azulene + H or isomerizes to the dihydro-naphtyl radical before forming naphthalene + H. Long et al.41 found that the C10H10 species can accumulate to high concentrations, enabling bimolecular hydrogen abstraction reactions to be competitive with decomposition of the C10H10 species. Pressure-dependent rate parameters for all unimolecular reactions are taken from Long et al.41 The rate expressions for the bimolecular hydrogen abstraction reactions have been determined with RMG’s rate rules. The reactions on the C10H11 potential energy surface, responsible for the formation of indene and benzene, analyzed by Cavallotti et al.44 have been recalculated in this work. The thermochemical parameters resulting from this recalculation are included in the RMG database and have thus been included in the kinetic model. Indene formation proceeds via β opening of one of the C5 rings of the CPDyl and CPD adduct, see Figure 3. Cavallotti et al.44 found that this pathway to indene is entropically and energetically favored over other pathways involving the formation of tricyclic intermediates. As opposed to the reactions on the C10H10 potential energy surface, high pressure limit rate coefficients are used for the reactions responsible for the formation of indene in the kinetic model. The entrance channel on the C10H11 potential energy surface is low in energy compared to the initial barriers for isomerization reactions, as shown by Cavallotti et al.44 Therefore, the use of high pressure limit rate coefficients is assumed to be sufficiently accurate for the conditions of this study. Two routes to benzene on the C10H11 potential energy surface are included in the kinetic model. One via β opening of one of the C5 rings and the other via a tricyclic intermediate. Alternative reaction pathways to benzene via CPDyl + CH3 and G

DOI: 10.1021/acs.energyfuels.7b03560 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

Figure 5. Experimental and simulated yields of the main reaction products during the pyrolysis of CPD (0.17 MPa, dilution of 1 mol CPD/5 mol N2, F0,CPD = 27 mg/s). Symbols, low dilution experimental data of Djokic et al.;28 full lines, current model predictions; dashed lines, model predictions with the POLIMI_141250 model; dotted lines, model predictions with the Wang et al.49 model; dash-dotted lines, model predictions with the Herbinet et al.51 model. The current kinetic model is available in the Supporting Information.

fulvene + H37 are also included in the model. Radical recombination of CPDyl and C2H5 or Diels−Alder or ene reactions of CPD and ethene also lead to the formation of benzene, see Figure 3. These reactions first form 5-ethyl-1,3CPD, with the exception of the Diels−Alder reaction of CPD and ethene, which first forms norbornene that in turn reacts to 5-ethyl-1,3-CPD. The reaction pathway to benzene proceeds via loss of a methyl radical by 5-ethyl-1,3-CPD and then merges with the well-known CPDyl + CH3 pathway;37 namely, it undergoes a ring expansion reaction and a β-scission reaction with loss of hydrogen to form benzene. Thermochemical parameters for all but the Diels−Alder reactions are recalculated with the CBS-QB3 method and are now part of the RMG database. The kinetic parameters for the Diels−Alder reaction of CPD and ethene have been determined using the corresponding RMG rate rule. Toluene formation via the reactions benzene + CH3, CPD + C2H3, and CPDyl + ethene are included in the kinetic model using thermochemical parameters calculated at the CBS-QB3 level of theory. For styrene formation, the kinetic model includes pathways via benzene + C2H3, CPD + C3H5, CPDyl + C3H6, and CPDyl + C3H5. The part of the C8H9 potential energy surface relevant for styrene formation was calculated by Sharma et al.46 using the CBS-QB3 method. Kinetic parameters for the reactions leading to the formation of C8H9 species after reaction of CPD + C3H5, CPDyl + C3H6, or CPDyl + C3H5 have not been reported. Values for these kinetic parameters have been determined by RMG’s rate rules.

Besides the aforementioned reactions, responsible for molecular weight growth, the kinetic model includes two decomposition routes for CPD. The first is initiated by hydrogen addition to CPD. The C5H7 adduct undergoes a ring opening, a hydrogen shift, and a β-scission reaction to form ethyne and an allyl radical. The second route proceeds via the CPDyl radical, which decomposes to a propargyl radical and ethyne after a hydrogen shift, a ring-opening, and a β-scission reaction. Thermochemical parameters for both decomposition pathways have been recalculated as part of this work using the CBS-QB3 method making use of prior work by Dean31 and Moskaleva et al.,78 respectively. 3.3. Model Predictions and Validation. The generated kinetic model is validated by comparison with experimental data. Additionally, the performance of the model at predicting the experimental data is compared with the performance of the Wang et al.,49 POLIMI_1412,50 and Herbinet et al.51 models from literature. The current model has been used to simulate the cyclopentadiene pyrolysis experiments of Kim et al.,27 Butler and Glassman,26 Djokic et al.,28 and the experimental results of the copyrolysis of CPD and ethene. Reactor simulations with the current and literature kinetic models are performed using the plug flow reactor module in CHEMKINPRO.58 Model predictions for the Kim et al.27 and Butler and Glassman26 data sets are reported in the Supporting Information. The current model accurately predicts the CPD conversion of the Kim et al.27 data. The trends for the product yields of the primary aromatic products are accurately H

DOI: 10.1021/acs.energyfuels.7b03560 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

Figure 6. Experimental and simulated yields of the main reaction products during the pyrolysis of CPD (0.17 MPa, dilution of 1 mol CPD/24 mol N2, F0,CPD = 6 mg/s). Symbols, high dilution experimental data of Djokic et al.;28 full lines, current model predictions; dashed lines, model predictions with the POLIMI_141250 model; dotted lines, model predictions with the Wang et al.49 model; dash-dotted lines, model predictions with the Herbinet et al.51 model. The current kinetic model is available in the Supporting Information.

reported by Kim et al.27 and Butler and Glassman.26 A similar observation can be made for styrene. Given the inconsistencies in the available experimental data, the quality of the different experiments is probably not the same. Kim et al.27 did not use adequate online analysis equipment. Instead the reactor effluent was condensed and collected in a dual ice-cooled dichloromethane trap and analyzed offline. Butler and Glassman26 performed the experiments in a setup in which hot nitrogen is used to heat the fuel up to the desired reaction temperature. Mixing and reaction zones are not clearly separate, which leads to some uncertainties in the effective reaction time and the temperature profile.79 The setup used by Djokic et al.28 and in this study overcomes both shortcomings. It uses a high quality online GC × GC-FID/(TOF-MS) analytic system and the reaction zone and temperature profiles are well-controlled. Additional CPD pyrolysis experiments at well-defined and controlled operating conditions are needed to verify the accuracy of the existing data. The following discussion will focus on the newly acquired experimental data for the copyrolysis of CPD and ethene and the experimental data sets of Djokic et al.28 In Figure 5, experimental and simulated product yields for the low dilution CPD pyrolysis experiments of Djokic et al.28 are shown. The yields of the main products of CPD pyrolysis, calculated by the current model, agree well with the experimental data. The CPD conversion and the yields of

described, but the absolute values of the product yields are off. The naphthalene yield is overpredicted while the yields of the other primary aromatic products are underpredicted. The CPD conversion as well as naphthalene product yield of the experimental data reported by Butler and Glassman26 are accurately predicted by the current model. Product yields for the other primary aromatic products are underpredicted. While quantitative predictions of the product yields for the Kim et al.27 and Butler and Glassman26 data sets by the current model are less accurate, the model performs very well at predicting the CPD conversion and product yields of the low and high dilution experiments of Djokic et al.28 and CPD/ethene copyrolysis experiments. As the different experiments are performed at comparable conditions (the only major difference is the CPD dilution), the kinetic model was expected to reproduce all data evenly well, provided that all experimental data is accurate. Comparing the POLIMI_141250 and the current kinetic model predictions reveals some inconsistencies between the experimental data sets. For example, the POLIMI_1412 model predicts indene yields quite well for the Kim et al.27 and Butler and Glassman26 data, while indene formation for the Djokic et al.28 and the new CPD/ethene copyrolysis data sets are overpredicted. The current kinetic model on the other hand performs well at predicting indene yields of the Djokic et al.28 and the new CPD/ethene copyrolysis data set, while underpredicting indene yields I

DOI: 10.1021/acs.energyfuels.7b03560 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

Figure 7. Experimental and simulated yields of the main reaction products during the copyrolysis of CPD and ethene (0.17 MPa, dilution 1 mol ethene/1 mol CPD/10 mol N2, F0,CPD = 13.6 mg/s). Symbols, experimental data; full lines, current model predictions; dashed lines, model predictions with the POLIMI_141250 model; dotted lines, model predictions with the POLIMI_141250 model after adjusting the kinetic parameters for the POLIMI_141250 global reaction C2H4 + CPD → C6H6 + CH3 + H. The current kinetic model and a detailed overview of the experimental data at all studied conditions are available in the Supporting Information.

was validated for CPD pyrolysis, performs better at predicting the low and high dilution CPD pyrolysis data of Djokic et al.28 than the Wang et al.49 and the Herbinet et al.51 models. The POLIMI_141250 model is therefore retained for comparison with the current model’s performance at predicting the product yields of the copyrolysis experiments. In Figure 7, the model predictions are compared to the experimental data obtained for the copyrolysis of CPD and ethene. As observed for the pyrolysis of CPD (see Figure 5 and Figure 6), CPD conversion begins at around 950 K, suggesting that the coreactant ethene does not notably influence the CPD conversion temperature. Ethene seems to remain unreactive at temperatures below 1023 K. The CPD conversion and benzene and indene yields are accurately predicted by the current model. The toluene yield is clearly underpredicted for temperatures above 1073 K. As opposed to the pure CPD pyrolysis, the overpredictions of the styrene and naphthalene yields are less severe. This is directly linked to the fact that fewer larger PAHs are formed during the copyrolysis of CPD with ethene. As the formation of larger PAHs is not included in the kinetic model, the model predictions for styrene and naphthalene are better when less of these molecules are consumed to form larger PAHs. POLIMI_1412,50 the model which reproduces the pyrolysis of pure CPD quite well, overpredicts CPD consumption and product formation in the CPD/ethene copyrolysis experiments. Sensitivity analysis for the POLIMI_141250 model shows that the reaction C2H4 + CPD → C6H6 + CH3 + H has a strong

benzene and indene are accurately reproduced. The toluene yield is underpredicted, though. The computed styrene and naphthalene yields continue to increase with temperature, but it is experimentally observed that those yields plateau, most likely contributing to the formation of larger PAHs, which are experimentally found in high yields. As the focus of this study is on the formation of the primary aromatic products, these larger PAHs are not included in the present kinetic model, so styrene and naphthalene are expected to accumulate as final products in the model. Some cyclopentene is already measured at 873 K, while the model does not predict cyclopentene formation at this temperature. Even though no cyclopentene impurity was detected in the unreacted feedstock (see Table S2), cyclopentene may well be a product of the impurities in the feedstock. Similarly, cyclopentene is also already measured at the lowest investigated temperature of the high dilution experiments of Djokic et al.28 (see Figure 6) and the CPD/ ethene copyrolysis experiments (see Figure 7). Experimental and simulated product yields for the high dilution CPD pyrolysis experiments of Djokic et al.28 are shown in Figure 6. As for the low dilution experiments, the current model accurately predicts the CPD conversion and yields of the main products for the high dilution experiments. Indene is accurately predicted, while the benzene profile is shifted toward higher temperatures leading to its underprediction. The predicted styrene and naphthalene yields again continue to increase with temperature, which is likely caused by missing PAH formation chemistry. The POLIMI_141250 model, which J

DOI: 10.1021/acs.energyfuels.7b03560 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

Figure 8. Normalized sensitivity coefficients of the primary aromatic products, determined using the current kinetic model, at 1123 K, 0.17 MPa, and a CPD conversion of 10% for the pyrolysis of CPD, dilution of 1 mol CPD/5 mol N2 (dark bars), and the copyrolysis of CPD and ethene, dilution of 1 mol CPD/1 mol ethene/10 mol N2 (light bars). See the Supporting Information for the complete kinetic model.

impact on the reactivity as it results in the creation of two radicals. The current kinetic model contains this reaction as a sequence of elementary-steps, and the C−C scission reaction of 2-ethyl-1,3-CPD to form a methyl and a cyclopentadienylmethyl radical, see Figure 8, is easily identified as the reaction step that has the most influence on the rate of benzene formation. Using the kinetic parameters of this reaction step for the global reaction C2H4 + CPD → C6H6 + CH3 + H in the POLIMI_141250 model results in significantly improved predictions as shown by the dotted lines in Figure 7. 3.4. Rate of Production and Sensitivity Analysis. The majority of CPD pyrolysis products detected at the conditions of this study are aromatic species, even at low conversions. The dominant reaction pathways for the formation of the primary aromatic products, according to the current model, are shown in Figure 3 for different temperatures at 10% CPD conversion. CPD mainly forms CPDyl radicals via hydrogen abstraction reactions. However, a significant amount of CPDyl radicals are

also formed by C−H bonds scission. The CPDyl radicals are predominantly consumed by self-recombination and addition to CPD. The reaction pathway initiated by the selfrecombination of CPDyl radicals leads to the formation of naphthalene. Figure 4 shows the CPDyl self-recombination reaction can also lead to the formation of fulvalene and azulene. However, according to model predictions, the fulvalene and azulene yields are one to two orders of magnitude lower than the yields of the major products of CPD pyrolysis. A comparison between the naphthalene, fulvalene, and azulene yields for various temperatures can be found in the Supporting Information. The reaction pathway initiated by addition of a CPDyl radical to CPD is responsible for the formation of indene. For the CPD/ethene copyrolysis at 1023 K the selfrecombination reaction is responsible for 36.9% of the total CPD consumption rate increasing to 44.3% at 1123 K. The rate of CPDyl consumption by addition to CPD decreases from 13.8% of the total CPD consumption rate at 1023 K to 8.2% at K

DOI: 10.1021/acs.energyfuels.7b03560 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

Figure 9, during the pyrolysis of pure CPD at 1123 K, vinyl radicals are formed via hydrogen addition to ethyne. Adding

1123 K. The sensitivity analysis, see Figure 8, reveals that naphthalene formation is most sensitive to reactions that form CPDyl radicals, while indene formation is much less sensitive to such reactions. Therefore, a higher rate of CPDyl radical formation will result in a decreased indene to naphthalene ratio. The fraction of consumed CPD that forms CPDyl via C−H bond scission increases from 15.4% at 1023 K to 32% at 1123 K. This increase in CPDyl formation is reflected in the lower indene to naphthalene ratio at 1123 K. Ethene is not involved in the reaction pathways responsible for the formation of naphthalene and indene. The contributions of the naphthalene formation pathways change with temperature. At 1023 K, about 60% of the naphthalene is formed via the fulvalanyl pathway. This amount decreases to 50% at 1123 K. The reason for this decrease is that at low temperature the chemically activated fulvalane formed by CPDyl self-recombinations will preferably stabilize through collisions with the bath gas.40,41 Since the rate coefficient for isomerization of fulvalane to azulane increases rapidly with temperature, naphthalene formation via the azulanyl pathway gains importance, hence the contribution of the fulvalanyl channel decreases. The azulanyl and fulvalanyl radicals are formed predominantly by bimolecular hydrogen abstraction reactions by H, methyl, CPDyl, and C5H7 radicals from azulane and fulvalane, respectively. At 1023 K 98% of the C10H9 species are formed via hydrogen abstraction reactions, decreasing slightly to 94% at 1123 K, confirming the conclusion of Long et al.41 that the C10H10 species are long-lived enough for bimolecular hydrogen abstraction reactions to be important. Long et al.41 only considered hydrogen abstraction reactions by CPDyl radicals and found that up to 1020 K at 0.1 MPa and 1230 K at 10 MPa hydrogen abstraction reactions were preferred over C−H bond scission reactions for the formation of C10H9 species. As expected by Long et al.,41 the dominance of hydrogen abstraction reactions over C−H bond scission reactions extends to higher temperatures (here 1123 K at 0.17 MPa) when the full system radical pool is taken into account. Upon CPDyl addition to CPD, the adduct either proceeds via a reaction pathway ending with the loss of a methyl radical to form indene or via a reaction pathway which ends with the formation of benzene and a butadienyl radical. Figure 3 shows that benzene formation by the latter pathway proceeds via a tricyclic intermediate. When considering all possible benzene formation routes in the kinetic model, the main reaction pathway for the formation of benzene depends on the presence of ethene. During the pyrolysis of pure CPD, benzene is formed almost exclusively via the pathway starting with the addition of CPDyl to CPD. When a large amount of ethene is present, the dominant reaction pathway for benzene formation proceeds via 5-ethyl-1,3-CPD. At 1023 K, 5-ethyl-1,3-CPD is mainly formed via Diels−Alder reaction of ethene and CPD, resulting in the formation of norbornene, which in turn reacts to 5-ethyl-1,3CPD. The considerably high amount of CPD that is consumed via this pathway (11.4%) explains why a significant amount of benzene is already observed at 1023 K. At 1123 K, both recombination of CPDyl with C2H5 and Diels−Alder reaction of CPD with ethene contribute about equally to the formation of 5-ethyl-1,3-CPD. The main reactions yielding toluene and styrene are shown on the right side of Figure 3. Toluene is formed via addition of a vinyl radical to CPD. The recombination of CPDyl and an allyl radical leads to styrene. The addition of ethene increases the fraction of CPD that reacts to form toluene. As shown in

Figure 9. Reaction path analysis based on the current kinetic model for the formation of the vinyl and allyl radical during the pyrolysis of CPD, with a dilution of 1 mol CPD/5 mol N2 at 1123 K (red and italic), and the copyrolysis of CPD and ethene, with a dilution of 1 mol CPD/1 mol ethene/10 mol N2 at 1123 K (blue and underlined) and 1023 K (green and bold), at 0.17 MPa and a CPD conversion of 10%. Percentages on a reaction path represent the reaction rate relative to the total CPD consumption rate. All reactions are reversible, arrows represent the net flux. See the Supporting Information for the complete kinetic model.

ethene as coreactant shifts the main channel for vinyl radical production to hydrogen abstraction from ethene. The main reaction pathway for the formation of allyl radicals is the dissociation of CPD,31,80 regardless of the addition of ethene. The higher CPD dilution in the copyrolysis of CPD and ethene increases the importance of the CPD dissociation pathway initiated by hydrogen addition to CPD as all the subsequent reactions along this pathway are unimolecular. As a result the fraction of CPD that is consumed to form allyl radicals increases and therefore also the importance of the styrene formation pathway. Sensitivity analyses were performed to assess the effect of the aforementioned reactions on the formation of the primary aromatic products. The results of these analyses for the pyrolysis of CPD and the copyrolysis of CPD with ethene are shown in Figure 8. A positive sensitivity coefficient indicates that increasing the pre-exponential factor of the associated reaction will lead to an increase in the mole fraction of the target molecule and vice versa. All primary aromatic products are sensitive to the rate of C−H bond scission of CPD, which is the main radical initiation reaction in this system. Toluene, styrene, and naphthalene are most sensitive to the CPD C−H bond scission reaction. These products rely greatly on the presence of sufficient radicals for hydrogen abstraction reactions, addition reactions, and/or recombination reactions, as opposed to benzene and indene which only require the addition of one CPDyl radical to CPD or alternatively for benzene no involvement of additional radicals if it is formed via ene or Diels−Alder reaction of ethene with CPD, as seen in Figure 3. In agreement with the results of the reaction path analysis, for the pyrolysis of pure CPD, the benzene formation is most sensitive to reactions along the pathway initiated by the addition of CPDyl to CPD, while for the copyrolysis of CPD with ethene, benzene is most sensitive to reactions along the pathway initiated by reaction of ethene and CPD. The sensitivity of toluene to reactions involving the vinyl radical is greatly reduced when ethene is added to the feed as coreactant. The hydrogen shift reaction from the primary 5-methyl-1,3cyclohexadienyl radical to the tertiary 5-methyl-1,3-cyclohexadienyl radical is the most important reaction for the L

DOI: 10.1021/acs.energyfuels.7b03560 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels formation of toluene. The normalized sensitivity coefficients for the formation of styrene, indene, and naphthalene do not differ significantly between the pure CPD pyrolysis and the copyrolysis of CPD and ethene.

4. CONCLUSIONS The copyrolysis of cyclopentadiene (CPD) and ethene was studied both experimentally and via kinetic modeling at a pressure of 0.17 MPa, a dilution of 1 mol CPD/1 mol ethene/ 10 mol N2, and temperatures ranging from 873 to 1163 K. Flow rates were chosen to obtain CPD conversions ranging from 1% to 92%, corresponding to residence times between 300 and 400 ms. Similar to the pure CPD case, CPD consumption in the copyrolysis experiments starts at around 950 K. The ethene concentration on the other hand remains unchanged up to 1023 K. Indene and naphthalene are the major products. Compared to the pyrolysis of pure CPD as feed, the addition of ethene tends to increase the selectivity of benzene and toluene while reducing the formation of larger PAHs. Existing literature models do not fully capture the CPD/ ethene copyrolysis chemistry, often relying on simplified globalstep reactions. Therefore, a new kinetic model was developed using RMG, an automated reaction network generator. The kinetic model was validated against CPD data from the literature and the CPD/ethene copyrolysis data of this work. The model accurately predicts the conversion of both CPD and ethene, and the formation of the primary aromatic products up to 1100 K. Since the model does not include C11 and heavier species, it does not capture the growth of polycyclic aromatic hydrocarbons (PAHs) beyond naphthalene. The omission of these reactions which consume smaller aromatics explains the overprediction of naphthalene and styrene yields at high conversions. On the basis of the developed kinetic model, the reaction pathways active during the pyrolysis of CPD have been corroborated and the effect of ethene as coreactant was assessed. During pure CPD pyrolysis, the main pathway for the formation of benzene starts with the addition of a CPDyl radical to CPD. When ethene is present, most of the benzene is formed via 5-ethyl-1,3-CPD, much of it via the CPD + ethene pathway leading to the formation of benzene, a hydrogen radical and a methyl radical.47,48 Toluene is formed mostly via addition of a vinyl radical to CPD. The main production channel for vinyl radicals changes from hydrogen addition to ethyne to hydrogen abstraction from ethene, when ethene is added as coreactant. Future work should focus on the formation of larger PAHs beyond naphthalene to further improve the agreement between model and experiment. Kinetic models containing larger PAHs, such as tri- and tetracyclic PAHs, will consist of a huge number of intermediates and reactions. Therefore, automated network generation will prove to be a valuable tool in the development of such kinetic models.





reactions in the kinetic model, comparison of the model predicted and experimentally determined yields of the major products for the experiments of Kim et al.27 and Butler and Glassman,26 model predicted naphthalene, fulvalene, and azulene yields (PDF) Kinetic model for the pyrolysis of cyclopentadiene (CHEMKIN58 format) (TXT)

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Hans-Heinrich Carstensen: 0000-0003-0501-7605 Guy B. Marin: 0000-0002-6733-1213 William H. Green: 0000-0003-2603-9694 Kevin M. Van Geem: 0000-0003-4191-4960 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS A.J.V. acknowledges financial support from the Research Board of Ghent University (BOF). The research at the LCT has received funding from the Long Term Structural Methusalem Funding by the Flemish Government. S.S.M. acknowledges financial support by the Combustion Energy Frontier Research Center, funded by the U.S. Department of Energy, Office of Basic Energy Sciences under award number DE-SC0001198. A.E.L. and W.H.G. acknowledge financial support from the DOE Gas Phase Chemical Physics program, Grant DESC0014901. This research used resources of the National Energy Research Scientific Computing Center under Contract DE-AC02-05CH11231.



REFERENCES

(1) Liu, G.; Niu, Z.; Van Niekerk, D.; Xue, J.; Zheng, L. Polycyclic Aromatic Hydrocarbons (PAHs) from Coal Combustion: Emissions, Analysis, and Toxicology. In Reviews of Environmental Contamination and Toxicology, Whitacre, D. M., Ed. Springer: New York, NY, 2008; pp 1−28. (2) Brookes, P. Mutagenicity of polycyclic aromatic hydrocarbons. Mutat. Res., Rev. Genet. Toxicol. 1977, 39 (3), 257−283. (3) Broyde, S.; Hingerty, B. Mutagenicity of Polycyclic Aromatic Hydrocarbons and Amines: A Conformational Hypothesisa. Ann. N. Y. Acad. Sci. 1984, 435 (1), 119−122. (4) Finlayson-Pitts, B. J.; Pitts, J. N., Jr. Tropospheric air pollution: ozone, airborne toxics, polycyclic aromatic hydrocarbons, and particles. Science 1997, 276 (5315), 1045−1051. (5) Chen, D.; Zainuddin, Z.; Yapp, E.; Akroyd, J.; Mosbach, S.; Kraft, M. A fully coupled simulation of PAH and soot growth with a population balance model. Proc. Combust. Inst. 2013, 34 (1), 1827− 1835. (6) Kopinke, F. D.; Zimmermann, G.; Reyniers, G. C.; Froment, G. F. Relative rates of coke formation from hydrocarbons in steam cracking of naphtha. 3. Aromatic hydrocarbons. Ind. Eng. Chem. Res. 1993, 32 (11), 2620−2625. (7) Hansen, N.; Li, W.; Law, M. E.; Kasper, T.; Westmoreland, P. R.; Yang, B.; Cool, T. A.; Lucassen, A. The importance of fuel dissociation and propargyl + allyl association for the formation of benzene in a fuelrich 1-hexene flame. Phys. Chem. Chem. Phys. 2010, 12 (38), 12112− 12122. (8) Miller, J. A.; Pilling, M. J.; Troe, J. Unravelling combustion mechanisms through a quantitative understanding of elementary reactions. Proc. Combust. Inst. 2005, 30 (1), 43−88.

ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.energyfuels.7b03560. Detailed description of the experimental setup, detailed compositions measured at all studied conditions, overview of the kinetic parameters for the most important M

DOI: 10.1021/acs.energyfuels.7b03560 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

(30) Lewis, D. K.; Sarr, M.; Keil, M. Cyclopentene decomposition in shock waves. J. Phys. Chem. 1974, 78 (4), 436−439. (31) Dean, A. M. Detailed kinetic modeling of autocatalysis in methane pyrolysis. J. Phys. Chem. 1990, 94 (4), 1432−1439. (32) Djokic, M. R.; Dijkmans, T.; Yildiz, G.; Prins, W.; Van Geem, K. M. Quantitative analysis of crude and stabilized bio-oils by comprehensive two-dimensional gas-chromatography. J. Chromatogr. A 2012, 1257, 131−140. (33) Alzueta, M. U.; Glarborg, P.; Dam-Johansen, K. Experimental and kinetic modeling study of the oxidation of benzene. Int. J. Chem. Kinet. 2000, 32 (8), 498−522. (34) Scheer, A. M.; Mukarakate, C.; Robichaud, D. J.; Ellison, G. B.; Nimlos, M. R. Radical Chemistry in the Thermal Decomposition of Anisole and Deuterated Anisoles: An Investigation of Aromatic Growth. J. Phys. Chem. A 2010, 114 (34), 9043−9056. (35) Djokic, M.; Carstensen, H.-H.; Van Geem, K. M.; Marin, G. B. The thermal decomposition of 2,5-dimethylfuran. Proc. Combust. Inst. 2013, 34 (1), 251−258. (36) Alexandrino, K.; Millera, Á .; Bilbao, R.; Alzueta, M. U. Novel aspects in the pyrolysis and oxidation of 2,5-dimethylfuran. Proc. Combust. Inst. 2015, 35 (2), 1717−1725. (37) Melius, C. F.; Colvin, M. E.; Marinov, N. M.; Pit, W. J.; Senkan, S. M. Reaction mechanisms in aromatic hydrocarbon formation involving the C5H5 cyclopentadienyl moiety. Symp. (Int.) Combust., [Proc.] 1996, 26 (1), 685−692. (38) Kislov, V. V.; Mebel, A. M. The Formation of Naphthalene, Azulene, and Fulvalene from Cyclic C5 Species in Combustion: An Ab Initio/RRKM Study of 9-H-Fulvalenyl (C5H5−C5H4) Radical Rearrangements. J. Phys. Chem. A 2007, 111 (38), 9532−9543. (39) Mebel, A. M.; Kislov, V. V. Can the C5H5 + C5H5 → C10H10 → C10H9 + H/C10H8 + H2 Reaction Produce Naphthalene? An Ab Initio/RRKM Study. J. Phys. Chem. A 2009, 113 (36), 9825−9833. (40) Cavallotti, C.; Polino, D. On the kinetics of the C5H5 + C5H5 reaction. Proc. Combust. Inst. 2013, 34 (1), 557−564. (41) Long, A. E.; Merchant, S. S.; Vandeputte, A. G.; Carstensen, H.H.; Vervust, A. J.; Marin, G. B.; Van Geem, K. M.; Green, W. H. Pressure dependent kinetic analysis of pathways to naphthalene from cyclopentadienyl recombination. Combust. Flame 2018, 187, 247−256. (42) Wang, D.; Violi, A.; Kim, D. H.; Mullholland, J. A. Formation of Naphthalene, Indene, and Benzene from Cyclopentadiene Pyrolysis: A DFT Study. J. Phys. Chem. A 2006, 110 (14), 4719−4725. (43) Kislov, V.; Mebel, A. An Ab Initio G3-Type/Statistical Theory Study of the Formation of Indene in Combustion Flames. II. The Pathways Originating from Reactions of Cyclic C5 Species Cyclopentadiene and Cyclopentadienyl Radicals. J. Phys. Chem. A 2008, 112 (4), 700−716. (44) Cavallotti, C.; Polino, D.; Frassoldati, A.; Ranzi, E. Analysis of Some Reaction Pathways Active during Cyclopentadiene Pyrolysis. J. Phys. Chem. A 2012, 116 (13), 3313−3324. (45) Fascella, S.; Cavallotti, C.; Rota, R.; Carrà, S. The peculiar kinetics of the reaction between acetylene and the cyclopentadienyl radical. J. Phys. Chem. A 2005, 109 (33), 7546−7557. (46) Sharma, S.; Harper, M. R.; Green, W. H. Modeling of 1,3hexadiene, 2,4-hexadiene and 1,4-hexadiene-doped methane flames: Flame modeling, benzene and styrene formation. Combust. Flame 2010, 157 (7), 1331−1345. (47) Faravelli, T.; Goldaniga, A.; Ranzi, E. The kinetic modeling of soot precursors in ethylene flames. Symp. (Int.) Combust., [Proc.] 1998, 27 (1), 1489−1495. (48) Dente, M.; Ranzi, E.; Goossens, A. G. Detailed prediction of olefin yields from hydrocarbon pyrolysis through a fundamental simulation model (SPYRO). Comput. Chem. Eng. 1979, 3 (1−4), 61− 75. (49) Wang, K.; Villano, S. M.; Dean, A. M. Fundamentally-based kinetic model for propene pyrolysis. Combust. Flame 2015, 162 (12), 4456−4470. (50) Ranzi, E.; Frassoldati, A.; Grana, R.; Cuoci, A.; Faravelli, T.; Kelley, A. P.; Law, C. K. Hierarchical and comparative kinetic

(9) Miller, J. A.; Melius, C. F. Kinetic and thermodynamic issues in the formation of aromatic compounds in flames of aliphatic fuels. Combust. Flame 1992, 91 (1), 21−39. (10) Mulholland, J. A.; Lu, M.; Kim, D.-H. Pyrolytic growth of polycyclic aromatic hydrocarbons by cyclopentadienyl moieties. Proc. Combust. Inst. 2000, 28 (2), 2593−2599. (11) 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. (12) 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 (31), 4259−4268. (13) Tranter, R. S.; Yang, X.; Kiefer, J. H. Dissociation of C3H3I and rates for C3H3 combination at high temperatures. Proc. Combust. Inst. 2011, 33 (1), 259−265. (14) Miller, J. A.; Klippenstein, S. J.; Georgievskii, Y.; Harding, L. B.; Allen, W. D.; Simmonett, A. C. Reactions between ResonanceStabilized Radicals: Propargyl + Allyl. J. Phys. Chem. A 2010, 114 (14), 4881−4890. (15) Vermeire, F. H.; De Bruycker, R.; Herbinet, O.; Carstensen, H.H.; Battin-Leclerc, F.; Marin, G. B.; Van Geem, K. M. Experimental and kinetic modeling study of the pyrolysis and oxidation of 1,5hexadiene: The reactivity of allylic radicals and their role in the formation of aromatics. Fuel 2017, 208, 779−790. (16) Matsugi, A.; Miyoshi, A. Modeling of two- and three-ring aromatics formation in the pyrolysis of toluene. Proc. Combust. Inst. 2013, 34 (1), 269−277. (17) Gomez, A.; Sidebotham, G.; Glassman, I. Sooting behavior in temperature-controlled laminar diffusion flames. Combust. Flame 1984, 58 (1), 45−57. (18) Melton, T. R.; Inal, F.; Senkan, S. M. The effects of equivalence ratio on the formation of polycyclic aromatic hydrocarbons and soot in premixed ethane flames. Combust. Flame 2000, 121 (4), 671−678. (19) Szwarc, M. The determination of bond dissociation energies by pyrolytic methods. Chem. Rev. 1950, 47 (1), 75−173. (20) Spielmann, R.; Cramers, C. Cyclopentadienic compounds as intermediates in the thermal degradation of phenols. Kinetics of the Thermal Decomposition of Cyclopentadiene. Chromatographia 1972, 5 (12), 295−300. (21) Cypres, R.; Bettens, B. Mecanismes de fragmentation pyrolytique du phenol et des cresols. Tetrahedron 1974, 30 (10), 1253−1260. (22) Cypres, R.; Bettens, B. La formation de la plupart des composes aromatiques produits lors de la pyrolyse du phenol, ne fait pas intervenir le carbone porteur de la fonction hydroxyle. Tetrahedron 1975, 31 (4), 359−365. (23) Manion, J. A.; Louw, R. Rates, products, and mechanisms in the gas-phase hydrogenolysis of phenol between 922 and 1175 K. J. Phys. Chem. 1989, 93 (9), 3563−3574. (24) Roy, K.; Frank, P.; Just, T. Shock Tube Study of HighTemperature Reactions of Cyclopentadiene. Isr. J. Chem. 1996, 36 (3), 275−278. (25) Roy, K.; Braun-Unkhoff, M.; Frank, P.; Just, T. Kinetics of the cyclopentadiene decay and the recombination of cyclopentadienyl radicals with H-atoms: Enthalpy of formation of the cyclopentadienyl radical. Int. J. Chem. Kinet. 2001, 33 (12), 821−833. (26) Butler, R. G.; Glassman, I. Cyclopentadiene combustion in a plug flow reactor near 1150K. Proc. Combust. Inst. 2009, 32 (1), 395− 402. (27) Kim, D. H.; Mulholland, J. A.; Wang, D.; Violi, A. Pyrolytic hydrocarbon growth from cyclopentadiene. J. Phys. Chem. A 2010, 114 (47), 12411−12416. (28) Djokic, M. R.; Van Geem, K. M.; Cavallotti, C.; Frassoldati, A.; Ranzi, E.; Marin, G. B. An experimental and kinetic modeling study of cyclopentadiene pyrolysis: First growth of polycyclic aromatic hydrocarbons. Combust. Flame 2014, 161 (11), 2739−2751. (29) Bryce, W. A.; Ruzicka, D. J. REACTIONS OF ALLYL RADICALS WITH OLEFINS. Can. J. Chem. 1960, 38 (6), 835−844. N

DOI: 10.1021/acs.energyfuels.7b03560 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels modeling of laminar flame speeds of hydrocarbon and oxygenated fuels. Prog. Energy Combust. Sci. 2012, 38 (4), 468−501. (51) Herbinet, O.; Rodriguez, A.; Husson, B.; Battin-Leclerc, F.; Wang, Z.; Cheng, Z.; Qi, F. Study of the Formation of the First Aromatic Rings in the Pyrolysis of Cyclopentene. J. Phys. Chem. A 2016, 120 (5), 668−682. (52) Xu, C.; Al Shoaibi, A. S.; Wang, C.; Carstensen, H.-H.; Dean, A. M. Kinetic Modeling of Ethane Pyrolysis at High Conversion. J. Phys. Chem. A 2011, 115 (38), 10470−10490. (53) Khandavilli, M. V.; Vermeire, F. H.; Van de Vijver, R.; Djokic, M.; Carstensen, H.-H.; Van Geem, K. M.; Marin, G. B. Group additive modeling of cyclopentane pyrolysis. J. Anal. Appl. Pyrolysis 2017, 128, 437−450. (54) Gao, C. W.; Allen, J. W.; Green, W. H.; West, R. H. Reaction Mechanism Generator: Automatic construction of chemical kinetic mechanisms. Comput. Phys. Commun. 2016, 203, 212−225. (55) Vandewiele, N. M.; Van Geem, K. M.; Reyniers, M.-F.; Marin, G. B. Genesys: Kinetic model construction using chemo-informatics. Chem. Eng. J. 2012, 207−208 (0), 526−538. (56) Van de Vijver, R.; Vandewiele, N. M.; Bhoorasingh, P. L.; Slakman, B. L.; Seyedzadeh Khanshan, F.; Carstensen, H.-H.; Reyniers, M.-F.; Marin, G. B.; West, R. H.; Van Geem, K. M. Int. J. Chem. Kinet. 2015, 47, 199−231. (57) Beens, J.; Boelens, H.; Tijssen, R.; Blomberg, J. Quantitative Aspects of Comprehensive Two-Dimensional Gas Chromatography (GC× GC). J. High Resolut. Chromatogr. 1998, 21 (1), 47−54. (58) CHEMKIN-PRO 15131; Reaction Design: San Diego, 2013. (59) Li, Y.; Zhou, C.-W.; Somers, K. P.; Zhang, K.; Curran, H. J. The oxidation of 2-butene: A high pressure ignition delay, kinetic modeling study and reactivity comparison with isobutene and 1-butene. Proc. Combust. Inst. 2017, 36 (1), 403−411. (60) Zhou, C.-W.; Li, Y.; O’Connor, E.; Somers, K. P.; Thion, S.; Keesee, C.; Mathieu, O.; Petersen, E. L.; DeVerter, T. A.; Oehlschlaeger, M. A.; Kukkadapu, G.; Sung, C.-J.; Alrefae, M.; Khaled, F.; Farooq, A.; Dirrenberger, P.; Glaude, P.-A.; Battin-Leclerc, F.; Santner, J.; Ju, Y.; Held, T.; Haas, F. M.; Dryer, F. L.; Curran, H. J. A comprehensive experimental and modeling study of isobutene oxidation. Combust. Flame 2016, 167, 353−379. (61) Burke, U.; Metcalfe, W. K.; Burke, S. M.; Heufer, K. A.; Dagaut, P.; Curran, H. J. A detailed chemical kinetic modeling, ignition delay time and jet-stirred reactor study of methanol oxidation. Combust. Flame 2016, 165, 125−136. (62) Burke, S. M.; Metcalfe, W.; Herbinet, O.; Battin-Leclerc, F.; Haas, F. M.; Santner, J.; Dryer, F. L.; Curran, H. J. An experimental and modeling study of propene oxidation. Part 1: Speciation measurements in jet-stirred and flow reactors. Combust. Flame 2014, 161 (11), 2765−2784. (63) Burke, S. M.; Burke, U.; Mc Donagh, R.; Mathieu, O.; Osorio, I.; Keesee, C.; Morones, A.; Petersen, E. L.; Wang, W.; DeVerter, T. A.; Oehlschlaeger, M. A.; Rhodes, B.; Hanson, R. K.; Davidson, D. F.; Weber, B. W.; Sung, C.-J.; Santner, J.; Ju, Y.; Haas, F. M.; Dryer, F. L.; Volkov, E. N.; Nilsson, E. J. K.; Konnov, A. A.; Alrefae, M.; Khaled, F.; Farooq, A.; Dirrenberger, P.; Glaude, P.-A.; Battin-Leclerc, F.; Curran, H. J. An experimental and modeling study of propene oxidation. Part 2: Ignition delay time and flame speed measurements. Combust. Flame 2015, 162 (2), 296−314. (64) Metcalfe, W. K.; Burke, S. M.; Ahmed, S. S.; Curran, H. J. A hierarchical and comparative kinetic modeling study of C1− C2 hydrocarbon and oxygenated fuels. Int. J. Chem. Kinet. 2013, 45 (10), 638−675. (65) Kéromnès, A.; Metcalfe, W. K.; Heufer, K. A.; Donohoe, N.; Das, A. K.; Sung, C.-J.; Herzler, J.; Naumann, C.; Griebel, P.; Mathieu, O.; Krejci, M. C.; Petersen, E. L.; Pitz, W. J.; Curran, H. J. An experimental and detailed chemical kinetic modeling study of hydrogen and syngas mixture oxidation at elevated pressures. Combust. Flame 2013, 160 (6), 995−1011. (66) Susnow, R. G.; Dean, A. M.; Green, W. H.; Peczak, P.; Broadbelt, L. J. Rate-Based Construction of Kinetic Models for Complex Systems. J. Phys. Chem. A 1997, 101 (20), 3731−3740.

(67) Benson, S. W.; Buss, J. H. Additivity Rules for the Estimation of Molecular Properties. Thermodynamic Properties. J. Chem. Phys. 1958, 29 (3), 546−572. (68) Benson, S. W. Thermochemical Kinetics: Methods for the Estimation of Thermochemical Data and Rate Parameters; Wiley: New York, 1976. (69) Van Geem, K. M.; Reyniers, M. F.; Marin, G. B.; Song, J.; Matheu, D. M.; Green, W. H. Automatic Reaction Network Generation using RMG for Steam Cracking of n-Hexane. AIChE J. 2006, 52 (2), 718−730. (70) Harper, M. R.; Van Geem, K. M.; Pyl, S. P.; Marin, G. B.; Green, W. H. Comprehensive reaction mechanism for n-butanol pyrolysis and combustion. Combust. Flame 2011, 158 (1), 16−41. (71) Montgomery, J. A., Jr; Frisch, M. J.; Ochterski, J. W.; Petersson, G. A. A complete basis set model chemistry. VI. Use of density functional geometries and frequencies. J. Chem. Phys. 1999, 110 (6), 2822−2827. (72) Montgomery, J. A., Jr; Frisch, M. J.; Ochterski, J. W.; Petersson, G. A. A complete basis set model chemistry. VII. Use of the minimum population localization method. J. Chem. Phys. 2000, 112 (15), 6532− 6542. (73) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian, Inc.: Wallingford, CT, 2004. (74) Sabbe, M. K.; Saeys, M.; Reyniers, M.-F.; Marin, G. B.; Van Speybroeck, V.; Waroquier, M. Group Additive Values for the Gas Phase Standard Enthalpy of Formation of Hydrocarbons and Hydrocarbon Radicals. J. Phys. Chem. A 2005, 109 (33), 7466−7480. (75) Eckart, C. The penetration of a potential barrier by electrons. Phys. Rev. 1930, 35 (11), 1303. (76) Pitzer, K. S.; Gwinn, W. D. Energy Levels and Thermodynamic Functions for Molecules with Internal Rotation I. Rigid Frame with Attached Tops. J. Chem. Phys. 1942, 10 (7), 428−440. (77) Werner, H. J.; Knowles, P. J.; Knizia, G.; Manby, F. R.; Schütz, M. Molpro: a general-purpose quantum chemistry program package. Wiley Interdisciplinary Reviews: Computational Molecular Science 2012, 2 (2), 242−253. (78) Moskaleva, L. V.; Lin, M.-C. Unimolecular isomerization/ decomposition of cyclopentadienyl and related bimolecular reverse process: ab initio MO/statistical theory study. J. Comput. Chem. 2000, 21 (6), 415−425. (79) Zhao, Z.; Chaos, M.; Kazakov, A.; Dryer, F. L. Thermal decomposition reaction and a comprehensive kinetic model of dimethyl ether. Int. J. Chem. Kinet. 2008, 40 (1), 1−18. (80) Bouwman, J.; Bodi, A.; Oomens, J.; Hemberger, P. On the formation of cyclopentadiene in the C3H5 + C2H2 reaction. Phys. Chem. Chem. Phys. 2015, 17 (32), 20508−20514.

O

DOI: 10.1021/acs.energyfuels.7b03560 Energy Fuels XXXX, XXX, XXX−XXX