Radical Chemistry in the Thermal Decomposition of Anisole and

J. Phys. Chem. A 2010, 114, 9043–9056

9043

Radical Chemistry in the Thermal Decomposition of Anisole and Deuterated Anisoles: An Investigation of Aromatic Growth Adam M. Scheer,†,‡ Calvin Mukarakate,† David J. Robichaud,† G. Barney Ellison,‡ and Mark R. Nimlos*,† National Renewable Energy Laboratory, 1617 Cole BouleVard, Golden, Colorado 80401-3393 and Department of Chemistry and Biochemistry, UniVersity of Colorado-Boulder, Boulder, Colorado 80309-0215 ReceiVed: March 6, 2010; ReVised Manuscript ReceiVed: June 25, 2010

The pyrolyses of anisole (C6H5OCH3), d3-anisole (C6H5OCD3), and d8-anisole (C6D5OCD3) have been studied using a hyperthermal tubular reactor and photoionization reflectron time-of-flight mass spectrometer. Gas exiting the reactor is subject to an immediate supersonic expansion after a residence time of approximately 65 µs. This allows the detection of highly reactive radical intermediates. Our results confirm that the first steps in the thermal decomposition of anisole are the loss of a methyl group to form phenoxy radical, followed by ejection of a CO to form cyclopentadienyl radical (c-C5H5); C6H5OCH3 f C6H5O + CH3; C6H5O f c-C5H5 + CO. At high temperatures (Twall ) 1200 °C - 1300 °C) the c-C5H5 decomposes to propargyl radical (CH2CCH) and acetylene; c-C5H5 f CH2CCH + C2H2. The formation of benzene and naphthalene is demonstrated with 1 + 1 resonance-enhanced multiphoton ionization. Propargyl radical recombination is a significant benzene formation channel. However, we show the majority of benzene is formed by a ring expansion reaction of methylcyclopentadiene (C5H5CH3) resulting from methyl radical addition to cyclopentadienyl radical; CH3 + c-C5H5 f C5H5CH3 f C6H6 + 2H. The naphthalene is generated from cyclopentadienyl radical recombination; 2c-C5H5 f C5H5-C5H5 f C10H8 + 2H. The respective intermediate amu 79 and 129 species associated with these reactions are detected, confirming the stepwise nature of the decompositions. These reactions are verified by pyrolysis studies of cyclopentadiene (C5H6) and C5H5CH3 obtained from rapid thermal dissociation of the respective dimer compounds, as well as pyrolysis studies of propargyl bromide (BrCH2CCH). Introduction The formation of polycyclic aromatic hydrocarbons (PAHs) has been a focus of numerous pyrolysis and combustion studies of both small aliphatic hydrocarbons1-3 and larger unsaturated compounds.4-6 Polycyclic aromatic hydrocarbons are precursors to airborne soot produced in common combustion processes.7 Such particulates are known to be carcinogenic and pose a general health threat.7,8 Tars or large PAHs have been shown to inhibit biofuels production from biomass gasification.9,10 Also, a variety of PAHs are formed in atmospheric and interstellar chemistry11-13 in which low-temperature, low-pressure conditions exist. Generation of the first aromatic ring is considered a key step in PAH formation14 and the reactive chemistry of resonantly stabilized radicals is important in these processes. Because of their stability, such radicals, like propargyl and cyclopentadienyl, have high rates of formation and low rates of consumption. This allows them to exist in high concentrations and to have a great influence on the chemistry in combustion systems.15 In particular, the combination of two propargyl radicals (CH2CCH) to form benzene has been widely researched because of its potential to initiate molecular weight growth.

This reaction has been shown to be important in the combustion of acetylene and other small aliphatic molecules.3,14,16 * To whom correspondence should be addressed. E-mail: mark.nimlos@ nrel.gov. † National Renewable Energy Laboratory. ‡ University of Colorado-Boulder.

Several intermediate steps are required for the initial C6H6 species to form benzene.17-19 Though propargyl radical recombination is a fundamental reaction critical in the pyrolysis and combustion of small molecules, its relative significance is unclear when larger compounds are thermally decomposed. Another radical recombination reaction important in molecular weight growth is the addition of methyl radical to cyclopentadienyl radical (CPDR) to form methylcyclopentadiene (MCPD). Methylcyclopentadiene then loses two hydrogen atoms to form benzene via a bicyclic intermediate and ring expansion (Reaction 2).20-22

The pyrolysis of anisole (C6H5OCH3) provides the opportunity to study a case in which numerous reactive intermediates are present in relatively high concentrations. By examining both the unimolecular and bimolecular chemistry that ensues in the anisole pyrolysis gas mixture, several key precursor reactions that lead to PAHs can be probed, including those involving methyl radical, propargyl radical, and CPDR. Anisole also serves as an important model compound for lignin, a major component of biomass.23 Lignin consists of a network of cross-linked aryl ethers with methoxy groups bound to phenol rings prevalent throughout the macromolecule.24 Deciphering the unimolecular fragmentation mechanisms opera-

10.1021/jp102046p  2010 American Chemical Society Published on Web 08/10/2010

9044

J. Phys. Chem. A, Vol. 114, No. 34, 2010

Figure 1. A schematic of the hyperthermal nozzle/TOFMS system. See text for experimental details.

tive in the pyrolysis of anisole is a fundamental step in understanding the thermochemical breakdown of biomass. Studying the subsequent bimolecular reactions in the pyrolysis gas is of great importance in mapping the formation processes of undesirable PAHs, a key issue facing clean synthesis gas (CO + H2) production from the gasification of biomass.10 With the use of a hyperthermal tubular reactor described below and a reflectron photoionization mass spectrometer, we have studied the pyrolysis of anisole using both single photon ionization (PIMS; 118.2 nm) and 1 + 1 resonance-enhanced multiphoton ionization (REMPI). To validate the observed decompositions and verify numerous important bimolecular reactions, the pyrolysis of two forms of deuterated anisole (C6H5OCD3 and C6D5OCD3) were studied as well as the pyrolysis of propargyl bromide (BrCH2CCH), cyclopentadiene (c-C5H6; CPD) dimer, and methylcyclopentadiene (c-C5H5CH3; MCPD) dimer, compounds that readily thermolize at low temperatures resulting in pure monomer sources. Cyclopentadiene and MCPD provide probes of both bimolecular and unimolecular chemistry important in the later steps of anisole pyrolysis. Our results show consistent, predictable fragmentation patterns that allow the mapping of stepwise thermo-cracking processes that will lay the foundation for studies of larger aryl ethers. The anisole and deuterated anisole spectra confirm earlier results23 and elucidate a bimolecular reaction that generates MCPD and eventually benzene.21,22 The results of the dimer compound studies will be shown to confirm these reactions as well as corroborate naphthalene production from the recombination of CPDR. Proof that both benzene and naphthalene are being generated in anisole pyrolysis is achieved with REMPI measurements. Experimental Section A. Hyperthermal Tubular Reactor. Our hyperthermal tubular reactor is depicted in Figure 1 and is based on the original design of Chen.25 It contains a spring-loaded solenoid pulsed valve that controls the back and forth motion of a poppet that opens and closes a 0.25 mm diameter circular aperture. A 2.5 cm long, 1 mm diameter, resistively heated silicon carbide

Scheer et al. tube is mounted immediately after the aperture. The SiC tube can be resistively heated to 1500 °C and the temperature is monitored by a type C thermocouple attached to the outer wall of the tube. The temperature can be maintained to (10 °C. Because of variations in quality of thermocouple contact with the tubular reactor upon replacement, we report absolute temperatures on the outside of the tubular reactor to (100 °C. The detailed temperature distribution inside the tube is not yet well understood, but is a topic of ongoing research. To prevent radiative heat loss, the SiC tube is shielded by a 1 cm diameter alumina cylinder. The aperture faceplate is heat-sinked to a water-cooled flange to avoid damaging the valve and other internal components. No excess hydrogen should exist on the interior of the SiC tube. Thus chemistry due to H-atoms is limited to those generated in the pyrolysis sample gas. This assumption is corroborated by the d8-anisole (C6D5OCD3) results in which no species are detected resulting from H-atom induced reactions. A more detailed description of the hyperthermal pulsed tubular reactor was given previously.26 If the sample has sufficient vapor pressure at room temperature, the pulsed valve is operated with a chosen partial pressure of sample mixed with 2000 Torr of He carrier gas. If the sample vapor pressure is too low, or if increasing the bimolecular chemistry is desired in a sample with moderate vapor pressure, the sample is loaded in a small quartz tube and inserted immediately before the solenoid valve where it is heated appropriately. In such cases, 2000 Torr of He still backs the tubular reactor and is introduced as the carrier gas. The valve is operated at 30 Hz, creating a pulsed molecular beam. Upon exiting the nozzle, the pyrolysis vapor undergoes a free expansion and rapidly cools, quenching the reactive chemistry. A skimmer is used to select the forward traveling component of the expansion, creating a well-collimated beam. Because of the short residence time (∼65 µs) of the sample in the pyrolysis chamber and the subsequent rotationally cooled molecular beam, radicals, and other reactive intermediates are observable.26 This hyperthermal tubular reactor setup has been used in multiple prior studies of important reactive species and has been coupled with several different detection methods.23,26-33 This work represents the first use of the tubular reactor as an extensive probe of complex bimolecular chemistry. B. Single Photon Ionization Mass Spectrometry. The molecular beam enters the ionization region of a reflectron timeof-flight mass spectrometer (Jordan). The ninth harmonic of a Nd:YAG laser (10.49 eV, 118.2 nm) is generated by taking the third harmonic of the Nd:YAG (Spectra Physics) through an Ar/Xe tripling cell34 and is used to ionize the beam. The 118.2 nm light has sufficient energy to ionize most molecules of interest but lies below the fragmentation threshold for the majority of compounds. Certain small molecules such as CO and CH4 have ionization potentials too high to be overcome by our ionization source and thus are not observed. Propylene and NO are used to calibrate the mass spectrometer. Full-width-athalf-maximum peak widths of 0.075 amu are typical in our spectra. C. Resonance-Enhanced Multiphoton Ionization Mass Spectrometry. The molecular beam again enters the ionization region of the mass spectrometer (Jordan). The third harmonic (355 nm) of an Nd:YAG laser (Spectra Physics) is used to drive an optical parametric oscillator (OPO; versascan/MB, GWU) with a range of 500-710 nm. The light emitted by the OPO then enters a frequency doubling stage (GWU) resulting in a final, tunable range of 250-355 nm (3.50-4.97 eV) with an instrumental line width of 6.5 cm-1. If a molecule has a

Thermal Decomposition of Anisole and Deuterated Anisoles

J. Phys. Chem. A, Vol. 114, No. 34, 2010 9045

TABLE 1: Kinetic Model reaction

C6H5OCH3 f C6H5O + CH3 C6H5O f cyc-C5H5 + CO cyc-C5H5 f C3H3 + C2H2 C3H3 + C3H3 f products CH3 + CH3 f C2H6 cyc-C5H5 + CH3 f C5H5CH3 C5H5CH3 f cyc-C5H5 + CH3 C5H5CH3 f product

A (s-1) 2 × 10

n

Ea (kcal/mol)

ref

15

64

36

7.4 × 1011

44

37-40

1.98 × 10

68

15.0

6.26 × 10-11

-0.75

7.42 × 10-11

-0.69

124

41

0.25

42

0.73

43

2.3 × 10-15

21.6

44

5.5 × 1011

49.8

44

2.7 × 10

53.9

44

10

resonance in this region it will be excited by one photon. A second photon of the same energy can subsequently ionize the molecule. Aromatic molecules are often particularly sensitive to REMPI ionization in this region because of strong electronic excitations that result in signature resonance structures. Because the 1 + 1 REMPI scheme involves two photons of the same energy, it is limited to the detection of ions via resonances at energies greater than half the ionization potential of the molecule of interest. D. Data Acquisition. Data acquisition and signal averaging are performed using Labview. The individual mass spectra shown in the figures are composite averages of 2000-10000 individual mass spectra. Each data point comprising the REMPI wavelength scans are composite averages of 150-200 individual mass spectra. Step sizes of 0.02 and 0.05 nm were used (check figure caption). The reported wavelength scans show the averaged signal from one particular value of m/z. E. Sample Preparation. The anisole, deuterated-anisole, propargyl bromide, cyclopentadiene dimer, and methylcyclopentadiene dimer samples were obtained from Sigma-Aldrich. The anisole sample is listed as 99.9% pure, while the d3-anisole (C6H5OCD3) and perdeuterio anisole (C6D5OCD3) were listed at 99 and 98% pure, respectively. Each anisole sample underwent multiple freeze-thaw processes under vacuum to remove impurities of high vapor pressure. The dimer compounds were obtained from Sigma-Aldrich with the CPD dimer listed at 98% pure and the MCPD dimer listed at 93% pure. No additional purification was performed for the dimer compounds. The propargyl bromide was obtained from Sigma-Aldrich and was an 80% mixture in toluene. Multiple freeze-thaw processes were performed in an attempt to increase the proportion of propargyl bromide to toluene, but no noticeable difference in signal was observed as a result. The benzene and d1-benzene used as REMPI standards were obtained from Sigma Aldrich at stated purities of 99.9 and 98%, respectively. The naphthalene used as a REMPI standard was obtained from JT Baker at a stated purity of 99.9%. Results and Discussion A. Unimolecular Decompositions in Anisole Pyrolysis. To aid in the analysis of the mass spectral data a simple 8-reaction kinetic model was developed using Dizzy35 (Table 1).36-44 Because of the uncertainties in the physical conditions of the experiment (temperature and pressure) and the numerous secondary reactions excluded, this model should not be considered rigorous. However, it does provide a feasibility check of our analysis. The results and comparison with experiment are presented in Supporting Information Figure S1. i. The Pyrolysis of Anisole. The anisole PIMS spectra from room temperature to 1300 °C are shown in Figure 2. Very small pyrolysis peaks are visible at 800 °C (not shown). The anisole

data confirm the previous results of Friderichsen et al.23 Peaks observed in the mass spectra are consistent with the following unimolecular decomposition mechanism.

Masses associated with molecules observable in our TOF system (IP < 10.49 eV) are included in the reaction schemes for comparison with the mass spectra. Ionization potentials important for this work are listed in Table 2.45-63 The thermochemistry of anisole (C6H5OCH3) can be established. The heats of formation of anisole and phenol are known.64 Recently, energy-resolved, competitive threshold collision-induced dissociation methods with guided ion beam tandem mass spectrometry were used to measure the gas-phase acidities of phenol relative to HCN, H2S, and the HOO radical. The gas-phase acidity of phenol was measured to be ∆acidH298(C6H5OH) ) 348 ( 1 kcal/mol.65 The O-H bond dissociation enthalpy was measured by Nix et al.66 to be DH298(C6H5O-H) ) 85.8 ( 0.1 kcal/mol. The heats of formation of the methyl radical67 as well as the phenyl and methoxy radicals68 are well-known. From these data, one can compute the bond energies: DH298(C6H5O-CH3) ) 61.7 ( 0.3 kcal/mol and DH298(C6H5-OCH3) ) 101.8 ( 0.8 kcal/mol. These values agree very well with those reported by Arends et al.36 Considering these bond strengths, it is not surprising that the first step in the thermal decomposition of anisole is the loss of methyl radical to form phenoxy radical23,69 as shown in Reaction 3. The subsequent fragmentation of phenoxy radical has been shown in several studies to proceed through the ejection of CO and ring closure to form CPDR.37,38,40,70-72 As given in Table 2, CO has an ionization potential of 14.01 eV,49 which is too

9046


Scheer et al.

TABLE 2: Formulas, Molecular Masses, and Ionization Potentials for Molecules Important in This Worka molecule hydrogen methyl radical methane acetylene carbon monoxide propargyl radical diacetylene vinyl acetylene cyclopentadienyl radical cyclopentadiene triacetylene benzene bromine 1-methylcyclopentadiene 2-methylcyclopentadiene 5-methylcyclopentadiene phenoxy radical phenol phenylacetylene anisole naphthalene a

formula H 2* CH3 CH4* C 2H 2* CO* CHCCH2 C 4H 2 CHCCHCH2 C 5H 5 C 5H 6 C 6H 2 C 6H 6 Br* C5H5CH3 C5H5CH3 C5H5CH3 C 6H 5O C6H5OH C6H5CCH C6H5OCH3 C10H8

mass (amu) 2.02 15.02 16.04 26.02 28.00 39.0 2 50.02 52.03 65.04 66.05 74.02 78.05 78.92; 80.92 80.06 80.06 80.06 93.04 94.04 102.05 108.06 128.06

IE (eV) 45

15.43 9.8446 12.6147 11.4048 14.0149 8.6850 10.1751 9.5852 8.4153 8.5754 9.5056 9.2457 11.8155 8.4056 8.4658 8.4556 8.5659 8.4960 8.8261 8.2062 8.1463

IE (nm) 80.4 126.1 98.4 108.8 88.6 143.0 122.0 129.5 147.5 144.8 130.6 134.3 105.1 147.7 146.7 146.8 145.0 146.2 140.7 151.3 152.4

The * indicates a molecule with an ionization potential above 10.487 eV. These molecules will not be observed in our PIMS experiment.

Figure 2. TOF mass spectra of anisole pyrolyis at varying temperatures. The temperature is measured on the outside of the SiC tube. Y-axis scaling is done so the most dominant peak in each spectrum extends the length of the plot.

high to be overcome by our laser source, and is thus not observed here. However, because of the previous direct observations of CO from phenoxy radical pyrolysis, we did not consider it necessary to monitor. Using phenol as a starting material and validating with standards prepared according to known decomposition reactions, Khachatryan et al.73 have recently demonstrated with lowtemperature matrix isolation electron paramagnetic resonance spectroscopy that CPDR is indeed a major decomposition product of phenoxy radical pyrolysis. Reaction 4 is consistent with these previous experiments. The observation of phenoxy radical at 900 °C in anisole pyrolysis, despite the much smaller activation energy of Reaction 4 compared to Reaction 3, is a reflection of the associated transition state structures. The loss

of methyl radical from anisole (Ea(3) ) 64 kcal/mol) has a loose transition state and therefore a large pre-exponential factor (A(3) ) 2 × 1015 s-1).36 However, though the ring closing mechanism of Reaction 4 has a small activation energy (Ea(4) ) 44 kcal/mol),37-39 it involves a tight, multicentered transition state with a small pre-exponential factor (A(4) ) 7.4 × 1011 s-1).40 Using rate constants determined from these values with our residence time (∼65 µs), we compute that at 900 °C the phenoxy radical concentration should be approximately 20% of the parent anisole in the unpyrolyzed sample, a result that is consistent with our experimental observations (Figure 2). Propargyl radical, seen at higher temperatures at m/z 39, results from the unimolecular decomposition of CPDR (Reaction 5).74 This peak has been observed previously in anisole pyrolysis experiments23 but was not assigned. The decomposition of CPDR into acetylene and propargyl radical has been investigated computationally by Moskaleva et al.41 They found the barrier for the process to be 62.4 kcal/mol at the G2M(rcc,MP2) level of theory and a pressure dependent pre-exponential factor of A ) 2 × 1068 × T-15.0 s-1 at 1 atm. The unimolecular pyrolytic steps outlined above are compared to the kinetic model in Supporting Information Figure S1. There is qualitative agreement between experimental signal heights and the model predictions. Our kinetic model overestimates CPDR production and underestimates propargyl radical concentration at higher temperatures. The CPDR overestimation is likely due to the neglect of various bimolecular reactions that would reduce the concentration of CPDR in our reactor. ii. d3-Anisole and d8-Anisole Pyrolysis. To verify the above reactions, we have studied the pyrolysis of two deuterated anisole isomers. The PIMS spectra from room temperature through 1300 °C are shown in Figure 3 for d3-anisole and in Supporting Information Figure S2 for d8-anisole. As expected from Reactions 3-5, Figure 3 shows that the peaks at 93, 65, and 39 in d3-anisole pyrolysis corresponding to nondeuterated phenoxy radical, CPDR, and propargyl radical did not move compared to the analogous peaks in nondeuterated anisole pyrolysis. The parent d3-anisole peak and d3-methyl radical peak shifted three mass units higher. Some isotope scrambling is observed at higher temperatures in the d3-anisole pyrolysis experiment. This is evident in the masses displayed between


J. Phys. Chem. A, Vol. 114, No. 34, 2010 9047

Figure 5. As in Figure 2 for cyclopentadiene dimer pyrolysis at high partial pressures and high temperatures. At 500 °C, the dimer is completely dissociated to a pure monomer source (see Figure 4 inset). Figure 3. As in Figure 2 for methyl-deuterated-anisole pyrolysis.

to form CPDR and begins near 1100 °C, much higher than the initiation of pyrolysis in anisole. As would be expected from the anisole pyrolysis results, the CPDR then decomposes into propargyl radical and acetylene (Reaction 5). Similarly, the MCPD dimer, discussed below, is completely converted to the monomer by 500 °C (Reaction 7; Figure 6).

Figure 4. As in Figure 2 for cyclopentadiene dimer pyrolysis at low partial pressure. At 600 °C the dimer is completely dissociated to a pure monomer source. Inset: at 500 °C the dimer is completely dissociated to a pure monomer source.

15-18 and between 65-67 in the 1200 and 1300 °C mass spectra corresponding to different degrees of deuteration of methyl radical and CPDR, respectively. The origin of scrambling is likely hydrogen exchange upon collisions between methyl radical and CPDR. In d8-anisole, each peak shifted as expected according to complete deuteration, as can be seen in Supporting Information Figure S2. iii. Cyclopentadiene (CPD) Pyrolysis. The pyrolysis of the CPD dimer is useful in isolating the chemistry due to CPDR decomposition. The dimer is completely converted to the monomer by 600 °C via the retro Diels-Alder reaction shown in Reaction 6. The PIMS spectra of CPD pyrolysis under low partial pressure conditions are shown in Figure 4 while Figure 5 shows CPD pyrolysis at high partial pressure. The first step in the thermocracking process is the loss of a hydrogen atom

B. Bimolecular Reactions in Anisole Pyrolysis. The information we provide in this section will establish methylcyclopentadiene (MCPD) decomposition as the primary source of benzene in our system from 1000-1200 °C. Propargyl radical recombination will also be established as a significant benzene formation pathway. Though our basic kinetic analysis of anisole decomposition helps validate our observations, accurately modeling benzene formation from Reactions 1 and 2 is extremely complicated. The potential energy surface associated with propargyl radical recombination consists of no fewer than 11 stable C6H6 isomers, each of which could rearrange to form benzene. Miller and Klippenstein report a lumped mechanism for propargyl radical recombination that still includes 10 reactions and advise against simplifying further.19 Accurately modeling MCPD decomposition also requires accounting for numerous isomers, pathways, and intermediates.21,22,44 Because of these difficulties and the unknown pressure and temperature distributions in the tubular reactor, we have not undertaken a detailed kinetic analysis of benzene formation. It should be noted however that benzene formation from proparyl radical requires

9048


Figure 6. As in Figure 2 for methylcyclopentadiene dimer pyrolysis at high partial pressures. Inset: at 500 °C the dimer is completely dissociated to a pure monomer source.

the recombination of two radicals that will be produced from two anisole molecules, while a methyl radical and a cyclpentadienyl radical will be produced from a single anisole molecule. In addition, our experimental observations and kinetic modeling presented in the Supporting Information show that at temperatures below 1200 °C, the concentration of CPDR is much higher than propargyl radical. Thus, if the reaction rates for benzene formation from propargyl radical and CPDR are similar, one would expect the CPDR mechanisms to dominate. The following discussion is centered on establishing and elucidating mechanistic pathways important to formation of benzene, naphthalene, and other important species. We comment on kinetics throughout the text where appropriate and helpful. i. Proof of Benzene Formation. Resonance-enhanced multiphoton ionization (1 + 1) was used to confirm the formation of benzene, phenol, and naphthalene in our anisole pyrolysis experiments. Figure 7 shows REMPI mass spectra obtained at different wavelengths for anisole pyrolysis at 1200 °C. At 259.1 nm, benzene has a strong resonance and thus dominates the mass spectrum. At 275.1 nm, phenol, likely generated by a hydrogen abstraction reaction involving phenoxy radical, shows resonant behavior, while at 279.1 nm, naphthalene is the major product observed. The m/z 102 shown in the 275.1 nm REMPI mass spectrum is likely phenylacetylene (C6H5CCH) and its generation is discussed briefly later. In Figure 8, the REMPI wavelength spectrum of the m/z 78 species observed in anisole pyrolysis at 1200 °C is compared to the REMPI wavelength spectrum of a prepared benzene standard from 258-268 nm. The excellent agreement between the two spectra is definitive proof of the formation of benzene in the anisole pyrolysis vapor. The peak at 259.1 nm is due to the well-known 1B2u (υ6 V ) 1) r 1A1g (υ6 V ) 0) transition.75-77 The spectra in Figure 8 match closely those observed by KohseHo¨inghaus et al.78 and assignments are based on the work of Atkinson and Parmenter.77 The REMPI measurements of m/z 94 compared to a prepared phenol standard also confirm the presence of a small concentration of phenol in the anisole pyrolysis vapor. ii. The Role of Propargyl Radical. The first traces of propargyl radical (mass 39 in anisole and d3-anisole, mass 42

Scheer et al.

Figure 7. Resonance-enhanced multiphoton ionization TOF mass spectra of anisole pyrolysis at 1200 °C at different wavelengths. The temperature is measured on the outside of the SiC tube. Y-axis scaling is done so the most dominant peak in each spectrum extends the length of the plot.

Figure 8. Upper curve: 1 + 1 resonance-enhanced multiphoton ionization scan of m/z 78 observed in anisole pyrolysis at 1200 °C recorded with a 0.02 nm step size. The temperature is measured on the outside of the SiC tube. Bottom curve: 1 + 1 REMPI scan of m/z 78 for the benzene standard recorded with a 0.05 nm step size. Assignments of the A˜ 1B2u rX˜ 1A1g vibronic components (only the three largest peaks are labeled) are based on the work of Atkinson and Parmenter.77

in d8-anisole) are observed in PIMS around 1000 °C and propargyl becomes the dominant peak by 1300 °C. Several studies have used a propargyl-propargyl radical recombination mechanism to explain the observed formation of benzene and other amu 78 species from the decomposition of C2 to C4 hydrocarbons.3,14,16,79,80 However, several experimental factors indicate that propargyl radical recombination is not the main source of benzene observed in our anisole pyrolysis experiments. First, in both the anisole and deuterated anisole pyrolysis

Thermal Decomposition of Anisole and Deuterated Anisoles TABLE 3: Signal Size for the Indicated Peaks in the Anisole (C6H5OCH3), d3-Anisole (C6H5OCD3), and d8-Anisole (C6D5OCD3) Pyrolysis Experimentsa

J. Phys. Chem. A, Vol. 114, No. 34, 2010 9049 TABLE 4: Signal Size for the Indicated Peaks in the Anisole (C6H5OCH3) and Propargyl Bromide (HCCCH2BR) Pyrolysis Experimentsa

signal size (mV)

signal size (mV)

1200 °C

m/z ) 39 (42)

m/z ) 78 (84)

ratio 39/78

anisole d3-anisole d8-anisole

13.7 28.5 29.0

3.0 1.0 3.0

4.5 28.5 9.7

a The last column is the ratio of the signals associated with amu 39 (propargyl radical) and 78 (benzene) for anisole and d3-anisole and the ratio of amu 42 (perdeutero-propargyl radical) and 84 (perdeutero-benzene). The larger ratio observed in d3-anisole indicates that propargyl radical recombination does not account for the majority of benzene formation (see text).

Figure 9. As in Figure 2 for 80% propargyl bromide in 20% toluene. At 700 °C, the bromine atom dissociations are nearly complete leaving a clean propargyl radical source.

experiments the first traces of benzene appear at temperatures below those at which propargyl radical is seen. Second, in d3anisole, the propargyl radical is nondeuterated and is seen at m/z 39. However, the m/z 78 (84) peak is much smaller relative to the size of the propargyl radical peak than in the nondeuterated (perdeutero) anisole compounds (Table 3). This indicates that another mechanism must be responsible for the majority of benzene formation and furthermore that the mechanism involves methyl radical, a molecule that is the only deuterated fragment in the unimoleuclar d3-anisole decomposition and whose additive chemistry would lead to benzene formation at m/z greater than 78. To isolate the bimolecular chemistry due to propargyl radical recombination, we studied the pyrolysis of propargyl radicals using a propargyl bromide precursor in the tubular reactor. The results are shown in Figure 9 and agree very well with a previous study by Jochnowitz et al.28 Bromine has an IP of 11.8 eV55 and is thus not observed in our PIMS experiments. One can see a strong toluene peak at m/z 92 in the room-temperature spectrum. Toluene is added as a stabilizer to the commercial sample. The figure inset shows that pure toluene remains largely stable up to 1100 °C (some m/z 91 is observed) in the tubular

1000 °C 1100 °C 1200 °C

anisole propargylBr anisole propargylBr anisole propargylBr

m/z ) 39

m/z ) 78

ratio 78/39

0.55 21.86 4.3 11.95 13.8 9.79

0.85 2.35 2 1.34 3.1 1.32

1.55 0.11 0.47 0.11 0.22 0.13

a The last column is the ratio of the signals associated with m/z 78 (benzene) and m/z 39 (propargyl radical). The larger ratio observed at each temperature for anisole pyrolysis indicates benzene formation is largely independent of propargyl radical recombination.

reactor and will thus not influence propargyl radical chemistry aside from a small concentration of hydrogen donor molecules. Nearly all the propargyl bromide has decomposed by 700 °C, resulting in a clean source of propargyl radicals. At 900 °C, some m/z 76 and 78 is observed. These products grow relative to the propargyl radical peak and are clearly visible at 1100 °C. The peak at m/z 76 is likely benzyne (see Reaction 14 below). Table 4 shows the ratio of the m/z 78 peak to the m/z 39 propargyl radical peak at several temperatures in anisole and propargyl bromide pyrolyses. Note that no propargyl radical is observed in anisole pyrolysis below 1000 °C. At each temperature there is much more m/z 78 relative to propargyl radical observed in anisole pyrolysis than in propargyl bromide pyrolysis. As the temperature increases, the ratio of the m/z 78 peak height to the m/z 39 peak height decreases in anisole pyrolysis and increases in pyrolysis of propargyl bromide. These different ratios and trends support the interpretation that the primary benzene source in anisole pyrolysis is not propargyl radical recombination. These data are even more persuasive considering that the partial pressure of propargyl bromide was at least 5 times greater than that used in the anisole experiments, allowing for more propargyl-propargyl collisions. At 1200 °C, toluene pyrolysis showed both propargyl radical and benzene formation. The benzene is likely formed by propargyl radical recombination. iii. MCPD Formation. The peak at m/z 80 in the anisole mass spectra first resolved at 1000 °C (Figure 2) is due to methyl radical recombination with CPDR to form MCPD (Reaction 8). Observations of m/z 83 in the d3-anisole spectra and m/z 88 in the d8-anisole spectra (see Reactions 9 and 10) are consistent with this assignment. Pecullan et al.69 claim that methyl radical recombination with phenoxy radical to form methylcyclohexadienone, followed by the ejection of a CO, is the primary MCPD formation pathway in anisole pyrolysis. However, the production of methylcyclohexadieneones in our system is likely very limited. Phenoxy radical is never observed in high abundance and has a concentration below detection limits by 1100 °C. Though CO ejection from phenoxy radical (Reaction 4) proceeds with a modest pre-exponential factor (A(4) ) 7.4 × 1011 s-1)40 the small activation energy (Ea(4) ) 44 kcal/mol)37-39 becomes paramount at high temperatures. Using these kinetic parameters and those listed in Table 1 provided by Arends et al.36 for phenoxy radical formation from anisole (Reaction 3), we compute that at 1100 °C and at the ∼65 µs residence time of our tubular reactor, phenoxy radical should be present at a concentration of 1% that of the parent anisole in the unpyrolyzed sample. At 1100 °C no m/z 108, corresponding to both the parent anisole and methylcyclohexadieneone, is detected in PIMS or

9050


Scheer et al.

TABLE 5: Signal Size for the Indicated Peaks in Anisole (C6H5OCH3) Pyrolysis Observed at 1100 °C for Partial Pressures of 1 and 2 Torr Anisole in 2000 Torr Hea anisole 1100 °C

signal size (mV)

ratio X/65

ratio X/65

ratio 2 Torr/1 Torr

m/z

1 Torr

2 Torr

(1 Torr)

(2 Torr)

relative to amu 65

65 78 79 80

12.3 0.15 0.20 0.16

40.3 1.40 2.05 2.20

1 0.012 0.016 0.013

1 0.035 0.051 0.055

1 2.85 3.13 4.20

a The m/z 65 corresponds to cycopentadienyl radical resulting from the unimolecular decomposition of anisole. Columns 4 and 5 are the ratios of the indicated signals relative to amu 65. The last column is equal to the corresponding value in column 5 divided by the value in column 4. These values show that the concentrations of amu 78, 79, and 80 in anisole pyrolysis decrease dramatically relative to the concentration of unimoleuclar decomposion products when the partial pressure of sample gas is lowered. This is characteristic of products resulting from bimolecular reactions.

REMPI. In contrast, at 1100 °C, a significant amount of MCPD and its decomposition products are observed. Methylcyclohexadieneone formation and decomposition may be of much greater importance at lower temperatures and significantly longer residence times.

To verify the bimolecular nature of the production of MCPD, a lower partial pressure of 1 Torr anisole in 2000 Torr He was used. At 1100 °C, the 2 Torr sample yields approximately 4 times the amount of m/z 80 signal relative to the m/z 65 signal. Since CPDR at m/z 65 originates mostly from unimolecular decomposition, it can be used to normalize signal intensities between spectra taken with different anisole partial pressures. Table 5 details these results. Similar ratios were found for the analogous species in the d3-anisole and d8-anisole pyrolysis experiments. These observations are consistent with bimolecular reactions, as higher partial pressures will result in increased collision frequency between reactant molecules. iW. MCPD Decomposition. Along with m/z 80, peaks are also clearly visible in anisole at 1100 °C at m/z 79 and 78, while corresponding peaks are observed in d8-anisole at m/z 88, 86, and 84 (Figure 2 and Supporting Information Figure S2). These masses are consistent with two subsequent hydrogen atom loses/ abstractions from MCPD. Such a decomposition was studied by Dubnikova et al.21 to investigate the formation of benzene

SCHEME 1

from the pyrolysis of MCPD, a reaction later observed by the same group in shock tube studies.22 The mechanism involves a ring expansion after the initial hydrogen atom ejection. Similar ring expansions have also been observed in tert-butyl-1,3cyclopentadiene.81 We are not aware of any previous observation of the intermediate amu 79 species in the MCPD decomposition. Since the MCPD originates from a bimolecular reaction, the concentration of its decomposition products would also be expected to diminish relative to species resulting from unimolecular decompositions. As shown in Table 5, at 1100 °C, when a partial pressure of 1 Torr of anisole is used, the peaks at m/z 79 and 78 are smaller by a factor of approximately 3 relative to the CPDR peak compared to when a partial pressure of 2 Torr is used. Similar ratios are observed for the analogous species in the d3-anisole and d8-anisole pyrolysis experiments. To test the conclusion that benzene formation in anisole pyrolysis largely originates from MCPD decomposition, we studied the pyrolysis of MCPD (Figure 6) generated from the MCPD dimer (Reaction 7). The three isomers of MCPD are shown in Scheme 1 and are all expected to be generated from the dimer decomposition. Energy barriers for the 1,5 sigmatropic H-atom shift in CPD have been calculated to be approximately 25 kcal/mol82 and the energy barrier for 5-MCPD to 1-MCPD transition has been found to be ∼24 kcal/mol experimentally.83 Activation energies for the 5-MCPD to 1-MCPD and the 1-MCPD to 2-MCPD shifts have been calculated to be 24.8 and 25.1 kcal/mol respectively. The computed modified preexponential factors are 2.8 × 1012 × T/1000 and 3.3 × 1013 × T/1000 respectively, where T is temperature in K.44 The rate constants of 2 × 108 and 2 × 109 s-1 at 1000 °C indicate that such shifts can readily occur at the temperatures and time scales employed here. One can see in Figure 6 that indeed m/z peaks 79 and 78 are prominent in the mass spectra at the same temperatures anticipated from the anisole results. As expected, as the temperature is increased further, the two hydrogen atom loss reactions tend to completion and only benzene remains. The MCPD results presented in Figure 6 also indicate that a significant competing decomposition channel is methyl radical loss resulting in CPDR. This was also observed in a previous study of MCPD by Ikeda et al.84 and is in agreement with previous kinetic studies discussed below.22,44 The resulting CPDR can then decompose to propargyl radical and acetylene at higher temperatures as was observed in CPD pyrolysis (Figure 4). In MCPD pyrolysis at 900 °C, the first trace of m/z 78 is observed before any propargyl radical is detected (Figure 6). At 1000 °C, benzene is the dominant peak in the spectrum while only minimal propargyl radical is being generated. In CPD pyrolysis experiments in which higher partial pressures were used (Figure 5), no m/z 78 was observed below 1100 °C. These data help confirm that in the anisole system, even at high temperatures, benzene production is more likely to occur through MCPD decomposition (Reaction 2) than propargyl radical recombination (Reaction 1). Figure 10 shows the progression of the reactions near m/z 78 in MCPD and anisole as the temperature of the tubular reactor


Figure 10. As in Figure 2 for (a) methylcyclopentadiene dimer (left column) and (b) anisole (right column) pyrolyses at varying temperatures in the range of m/z 78. Diagnostic tests on anisole showed that the new thermocouple used to obtain the methylcyclopentadiene dimer exhibited the same 100 °C temperature shift displayed here.

is increased. One can see that the anisole and MCPD spectra are very similar. The discrepancy of ∼100 °C is due to a new thermocouple installed for the dimer compound studies. When anisole was studied again for diagnostic purposes, the same temperature shift was observed. As no CPDR is produced in CPD pyrolysis below 1100 °C, no further pyrolysis products or bimolecular reactions are observed below this temperature either. The three isomers of MCPD differ in the position of the methyl group relative to the sp3 ring carbon (Scheme 1). These structures have been calculated to be within 3 kcal/mol of each other.21 Because of the nature of the methyl recombination reaction with CPDR in anisole pyrolysis, 5-MCPD is initially generated in our tubular reactor. As mentioned, Ikeda et al.84 reported methyl radical dissociation from MCPD as a significant decomposition channel, an observation confirmed here (Figure 6). Those authors point out that the methyl radical loss must occur from the 5-MCPD isomer (73 kcal/mol)85 as the dissociation form the 1- or 2-MCPD isomers have barriers of greater than 100 kcal/mol. W. MCPD Kinetic Arguments. A survey of previous kinetic research associated with MCPD reveals a few insights important to interpreting our results. First, as discussed above, the three isomers shown in Scheme 1 all likely contribute to benzene formation to varying degrees. The numerous decomposition channels of the three isomers are very different and are not surprisingly governed by very different kinetics. These thermal dissociation pathways have been investigated by Lifshitz et al.21,22 and Sharma and Green.44 After the first hydrogen atom loss, the most probable pathway to benzene for any MCPD isomer is through isomerization to 5-methylenecyclopentadiene followed by ring expansion to cyclohexadienyl radical. Subsequent conversion of cyclohexadienyl radical to benzene is very fast. Second, methyl radical loss from MCPD is by far the fastest unimolecular decomposition channel. However, hydrogen abstraction reactions initiated by methyl radical or hydrogen atoms can facilitate the ring-opening mechanism.

J. Phys. Chem. A, Vol. 114, No. 34, 2010 9051 In the absence of collisions with reactant molecules, a simple analysis using kinetic parameters provided by Lifshitz et al.21,22 illustrates likely decomposition channels of the various MCPD isomers. Computed rate constants at 1000 °C show 5-MCPD to be approximately 50 times more likely to decompose to methyl radical and CPDR than it is to lose a H atom and form methylcyclopentadienyl radical. This difference is more than 3 orders of magnitude when comparing to 5-MCPD decomposition to 5-methylenecyclopentadiene. However, as we discuss below, the d3-anisole pyrolysis experiments demonstrate that generation of both methylcyclopentadienyl radical and methylenecyclopentadiene are significant decomposition channels. Furthermore, Figure 6 shows that we observe roughly equal amounts of benzene and CPDR from MCPD decomposition in our tubular reactor. Collisions with H and CH3 radicals can have a great influence on the MCPD decomposition branching ratios. For example, energy barriers for hydrogen abstraction from MCPD by atomic hydrogen and methyl radical to form methylcyclopentadienyl radical or methylenecyclopentadiene are estimated to be only 10.0 and 11.5 kcal/mol respectively and have moderate pre-exponential factors (AH2 ) 1014 s-1; ACH4 ) 5 × 1012 s-1). We would not be sensitive to the products of such reactions because H2 has an IP of 15.43 eV45 and methane has an IP of 12.6 eV,47 both greater than our 10.49 eV light. Considering that bimolecular chemistry involving methyl radical has already been shown to be significant at the conditions in our reactor, these abstraction reactions are very likely important. Another decomposition channel possible for both methylcyclopentadienyl radical and methylenecyclopentadiene is C-C bond breaking to form an open C6H7 species. Again using kinetic parameters provided by Lifshitz et al.21,22 and taking 1-MCPD as an example, we calculate that at 1000 °C 1-methylenecyclopentadiene is more than twice as likely to decompose to cyclohexadienyl radical via 5-methylenecyclopentadiene than it is to form an open C6H7 isomer. However, the situation is reversed when considering 5-methylcyclopentadienyl radical. At 1000 °C, this intermediate is calculated to decompose primarily to open C6H7 species. Depending on the extent of radical-induced hydrogen abstraction, the parent 5-MCPD isomer is expected to produce potentially much more 5-methylcyclopentadienyl radical than 5-methylenecyclopentadiene. Yet Lifshitz et al.22 report significantly more benzene production than open C6H6 compounds in their shock tube study of MCPD decomposition. This may be due to 5-MCPD loss via its more facile decomposition to CH3 + CPDR than compared to the 1and 2-MCPD isomers, leaving a preponderance of the latter two isomers to decompose to benzene. It may also be due to the C6H7 intermediates reforming ring compounds or decomposing further to smaller aliphatic hydrocarbons. On the basis of the above discussion, with all three isomers in Scheme 1 present in our reactor, decaying to the four radicals 5-methylcyclopentadienyl and 1-, 2-, and 5-methylenecyclopentadienyl, we expect the majority of the C6H6 PIMS signal observed is due to benzene. The d3-anisole pyrolysis experiments presented below are consistent with this interpretation. References 21, 22, and 44 contain much more detailed kinetics discussion, figures, structures, and electronic distributions for the numerous intermediates and transition states associated with MCPD decomposition. Wi. MCPD Decomposition in d3-Anisole Pyrolysis. One possibility for MCPD decomposition to benzene, shown in Reaction 11 for the case of d3-anisole pyrolysis, involves hydrogen loss/abstraction from the methyl group of 5-MCPD to form 5-methylenecyclopentadiene radical followed by cy-

9052


Scheer et al.

clohexadienyl radical formation via a bicyclic intermediate and subsequent ring expansion.21 Cyclohexadienyl radical will then rapidly lose a hydrogen atom and isomerize to benzene.

Figure 11. As in Figure 2 for methyl-deuterated anisole pyrolysis at varying temperatures in the range of m/z 80.

A different pathway shown in Reaction 12 is initiated by hydrogen loss/abstraction from the ring sp3 carbon to form 5-methylcyclopentadienyl radical, followed by isomerization to 5-methylenecyclopentadiene that allows for formation of the bicyclic intermediate that facilitates ring-opening. Isomerization of 5-methylcyclopentadienyl radical to 5-methylenecyclopentadiene is shown in the second step of Reaction 12. Because of the different masses occurring throughout Reactions 11 and 12, the d3-anisole pyrolysis experiments can be used to probe them separately. Our observation of masses 82 and 81 in d3-anisole pyrolysis indicates that the initial H-atom ejection can originate from either the methyl group or the ring. However, two PIMS observations suggest that Reaction 12 is preferred. First, as can be seen in Figure 11, in d3-anisole pyrolysis, significantly more m/z 82 is observed than 81 (assuming similar photoionization cross sections), indicating that it is more probable for the initial hydrogen loss to originate from the ring. Second, in the higher temperatures, much more m/z 80 is observed than 79. This would be expected if the final stable species was dideutero benzene, a product of Reaction 12. As discussed above, variations of these reactions will also exist for both the 1- and 2-MCPD compounds via isomerizations of the intermediates. The REMPI results on d3-anisole are consistent with these observations. One can see in Figure 12, that m/z 80, 79, and 78 follow the same fingerprint structure of benzene. The m/z 80 and 79 are shifted slightly due to changes in the vibrational band structure associated with the deuterium atoms. Monodeuterated benzene has C2V point group symmetry and the doubly degenerate ν6 mode is split into two modes, ν6a and ν6b. The 6a01 and 6b01 transitions have been observed at 258.84 and 258.82 nm, respectively.86 We would not resolve such a small splitting due to the temperature of our beam (∼50 K), but do observe a single peak centered at 258.83 nm in the d3-anisole pyrolysis experiment. In REMPI experiments of monodeuterated benzene, we reproduced the middle spectrum of Figure 12 very

Figure 12. 1 + 1 Resonance-enhanced multiphoton ionization scans of d3-anisole pyrolysis. The temperature is measured on the outside of the SiC tube. The spectra are shown with the same Y-axis scaling. Upper curve, m/z 80; middle curve, m/z 79; lower curve, m/z 78. Each series of lines of the same color help guide the eye to see shifts in the analogous transitions. See Figure 8 for spectroscopic assignments.

closely. A further shift to higher energy would be expected for d2-benzene and is observed in the upper scan of Figure 12. Lines to guide the eye show the consistent shifts in energy of the analogous transitions in the three benzene isomers, a reflection of the excitation of the ring modes. The three spectra shown in Figure 12 were taken at the same time and are to scale. The bottom curve (m/z 78) shows the lesser amount of benzene being generated due primarily to propargyl radical recombination. A slight amount of scrambling is also expected at 1200 °C and likely accounts for part of the various benzene isotopes observed. However, a careful analysis of the d3-anisole PIMS data confirms that propargyl radical recombination accounts for a small portion of the benzene concentration. As mentioned, at


Figure 13. As in Figure 2 for anisole (lower) and perdeutero-anisole (upper) pyrolyses showing the formation of naphthalene. The m/z 127 peak seen in both spectra is due to a small amount of iodine contamination.

higher temperatures significant isotope scrambling is observed in the d3-anisole data as is evident in the methyl radical peaks (m/z ) 15-18) in Figure 3. At 1200 °C, the m/z 15 peak corresponding to nondeuterated methyl radical is extremely small, while the m/z 78 peak is approximately twice as large. On the other hand, in anisole at 1200 °C the methyl radical peak is twice as large as the m/z 78 peak. Thus in d3-anisole, not all of the m/z 78 observed can be attributed to nondeuterated MCPD decomposition. At this temperature, propargyl radical is the major product seen. Thus our REMPI results on d3-anisole (Figure 12, bottom curve) indeed confirm that a small concentration of benzene is present due to propargyl radical recombination. Another possibility that must be addressed is that 5-methylenecyclopentadiene or 5-methylcyclopentadienyl radical can lose a hydrogen atom without undergoing a ring expansion, thus forming fulvene. Fulvene can then isomerize to benzene via a bicyclic pathway that involves several transition states and intermediates.21 The fulvene to benzene conversion proceeds at a fairly slow rate.44 In the flow tube experiments of Pecullan et al.69 benzene production in anisole pyrolysis was attributed to MCPD decomposition. No evidence for fulvene or linear C6H6 moieties was mentioned. No evidence for amu 78 compounds other than benzene was seen in our anisole REMPI experiments. However, we are unaware of existing fulvene UV or REMPI data with which to compare and thus cannot state definitively that fulvene does not contribute to m/z 78 formation in our PIMS experiments. C. Naphthalene Production. Naphthalene is an important precursor in the formation of larger PAHs.7 The recombination of two CPDRs has long been thought to be a primary source of naphthalene in decomposition and combustion processes.23,80,85,87-90 The mechanism predicted by Melius et al.85 has gained wide acceptance. This reaction involves the formation of a dihydrofulvane (C10H10, amu 130) with a single bond connecting two cyclic C5H5 moieties. Two subsequent hydrogen atom ejections lead to an amu 129 species and finally, naphthalene (Reaction 13).85 Numerous intermediates and transition states exist along

J. Phys. Chem. A, Vol. 114, No. 34, 2010 9053 the potential energy surface. Using BAC-MP4 (Bond-AdditivityCorrected Moeller-Plesset fourth-Order Perturbation) calculations, they found the largest single step activation energy to be 77.4 kcal/mol. A competing reaction is the formation of fulvalene85 at m/z 128. In the anisole pyrolysis study of Friderichsen et al.,23 a large CPDR signal was seen, but no naphthalene was observed. Though several experimental studies have attributed naphthalene observation to CPDR recombination,20,69,84,91 there have been no previous observations of the amu 129 precursor species. Filley et al.91 studied the pyrolysis of triphenylmethylcyclopentadiene with gas chromatography-mass spectrometric analysis of liquid products. They used a cylindrical furnace with a characteristic residence time of approximately 2 s. The triphenylmethylclopentadiene served as an efficient source of CPDR due to the high degree of resonance stabilization of the resulting radicals. At 550 °C, two peaks in the gas chromatogram with molecular ions of m/z 130 were reported as well as a m/z 128 peak corresponding to naphthalene. These two m/z 130 peaks are consistent with generation of two dihydrofulvalene isomers. As the temperature was increased, the m/z 130 peaks diminished until they disappeared completely by 700 °C, consistent with their assignment as naphthalene precursors.

In each of the compounds investigated here, we searched for a naphthalene signal. Figure 13 shows that at 1150 °C in both anisole and d8-anisole the m/z peak associated with naphthalene (128 in anisole and 136 in d8-anisole) was observed. Furthermore, the intermediate m/z peaks at 130 and 129 were resolved in the anisole pyrolysis, while a peak at 138 (C10D9) was resolved in the d8-anisole experiment. Because of waveform restrictions in our oscilloscope, the d8-anisole data was obtained at a lower resolution than was used for anisole and the dimer compounds. This explains the greater difficulty in resolving the intermediate peaks. The m/z 127 peak seen in both spectra is due to a small amount of iodine contamination. At this temperature, isotope scrambling was somewhat problematic in d3-anisole and considering the very small naphthalene signal, it is not surprising that it was not seen. Figure 14 shows that in both of the dimer compounds, the 130, 129, 128 reaction sequence was observed, indicating the same naphthalene formation mechanism. D. REMPI Confirmation of Naphthalene Production. The upper curve of Figure 15 shows the UV spectrum of gas phase naphthalene reproduced from Ferguson et al. from 266 to 285 nm.92 The middle curve is the m/z 128 REMPI signal observed in anisole pyrolysis at 1200 °C in the same range. There is excellent agreement between the UV and REMPI signals as would be expected for processes that share an initial single photon absorption step. Ejection of the excited electron into the continuum by absorption of a second photon is not expected to have a highly featured cross section as a function of wavelength.

9054


Figure 14. As in Figure 2 for cyclopentadiene dimer (lower) and methylcyclopentadiene dimer (upper) showing the formation of naphthalene at varying temperatures.

Figure 15. Upper curve: gas phase UV spectrum of naphthalene reproduced from Ferguson et al.92 Middle curve: resonance-enhanced multiphoton ionization scan of m/z 128 observed in anisole pyrolysis at 1200 °C. The temperature is measured on the outside of the SiC tube. Bottom curves: REMPI scan of m/z 128 for anisole pyrolysis at 1200 °C and for the naphthalene standard. Both lower REMPI curves are normalized by dividing by the square of the laser power.

When no very sharp resonances are seen, laser power fluctuations can make it difficult to definitively assign products based on observations of standards. This was problematic in our naphthalene REMPI experiments. Accounting for the power fluctuations caused significant additional noise in the spectra. However, to verify the m/z 128 signal observed in anisole pyrolysis against a naphthalene standard, we normalized small portions of each in a region containing a characteristic REMPI peak. The lower two curves of Figure 15 show the m/z 128 REMPI signals from 1200 °C anisole pyrolysis and a pure naphthalene standard. Both curves are corrected for power

Scheer et al. fluctuations by dividing by the square of the power entering the ionization chamber (the two-photon REMPI process has a quadratic dependence on laser power). The poorer signal-tonoise ratio is a reflection of the noise in the power readings. One can see the two spectra agree very well. Along with the UV comparison, this provides definitive proof of naphthalene production in anisole pyrolysis. E. Other Species. Peaks at m/z 77 and 76 (seen clearly in Figures 5; CPD, 9; propargyl bromide and 11; d3-anisole) likely result from Reaction 14 in which a hydrogen is lost in the propargyl radical recombination. Losing another hydrogen atom can lead to benzyne. Benzyne has been shown to decompose to acetylene and diacetylene at m/z 50 via a retro-Diels-Alder process with a barrier of 88.0 ( 0.5 kcal/mol (Reaction 15).32 Indeed, peaks at m/z 50 are observed at high temperatures in all of the above cases (52 in d8-anisole; Supporting Information Figure S2). The m/z 74 peaks seen in the same spectra are likely triacetylene. The m/z 74 was observed in the study of benzyne decomposition and were attributed to secondary reactions of benzyne. At 1300 °C, a peak at m/z 52 is observed in anisole, while d8-anisole displays a peak at 56. These masses likely indicate formation of vinyl acetylene (C4H4; C4D4). Peaks are resolved at this temperature in d3-anisole at each integer value between 50-55. The wide range of values results from isotope scrambling. Peaks at m/z 50 and 52 are also seen in both dimer compounds.

At high partial pressures and high temperatures in CPD pyrolysis, a series of peaks at 104, 103, and 102 is observed. This is shown in Supporting Information Figure S3. These likely correspond to a C8H8 compound resulting from propargyl radical recombination with CPDR. The subsequent hydrogen loss reactions may be similar to the ring expansion described above for MCPD. In this case, such a reaction could result in phenylacetylene at m/z 102. The propargyl radical recombination with CPDR may be an important precursor reaction in the growth of PAHs. Further studies of this reaction and its products would be interesting. Conclusions Using our hyperthermal tubular reactor with ∼65 µs retention times, we studied the pyrolysis of anisole using both single photon ionization as well as 1 + 1 resonance-enhanced multiphoton time-of-flight mass spectrometry. The results of methyl-deuterated anisole, perdeutero anisole, cyclopentadiene dimer and methylcyclopentadiene dimer pyrolysis experiments elucidate both the unimolecular and bimolecular chemistry observed in anisole pyrolysis. Our results show that though propargyl radical is a major product of the thermal decomposition of anisole, it has a limited role in benzene production. Instead, the formation and subsequent decomposition of methylcyclopentadiene is the major source of benzene. This work includes the first observation of the amu 79 species, likely due

Thermal Decomposition of Anisole and Deuterated Anisoles to several intermediate radical species, in the decomposition of methylcyclopentadiene. The recombination of two cyclopentadienyl radicals is shown to produce naphthalene through a mechanism involving two independent hydrogen losses. This work also achieves the first observation of the associated C10H9 precursor species at amu 129. One significant limitation of our study is the lack of an accurate temperature and pressure profile inside the tubular reactor. This prevents a detailed kinetic analysis of the observed reactions. We are in the process of addressing this issue experimentally and computationally. The kinetic analysis of many of the reactions discussed here has been done previously, including for propargyl radical recombination,17,19 decomposition of cyclopentadienyl radical,41 methyl radical recombination with cyclopentadienyl radical,44 decomposition and ring expansion of methylcyclopentadiene,21,22,44 and cyclopentadienyl radical recombination to naphthalene.90 These studies provide in-depth analysis of the respective reaction pathways. However, we note that an updated, detailed kinetic mechanism of anisole pyrolysis that includes propargyl recombination paths as well as methylcyclopentadiene decomposition would be very useful. To establish greater detail in the mechanistic interpretations, the use of 1 + 1′ REMPI in which the second photon has a higher energy would enable observation of ions via resonances at energies less than half of the respective ionization energies. Such an experiment could be particularly effective in the detection of numerous radical and other intermediate species. Acknowledgment. The U.S. DOE Office of the Biomass Program provided funding for this work. We thank Donald David, Ken Smith, Mark Jarvis, and AnGayle Vasiliou for technical support and Dr. Anthony Dean and Dr. Hans-Heinrich Carstensen for useful discussions. Supporting Information Available: Additional mass spectra can be found in the Supporting Information. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Kamphus, M.; Braun-Unkhoff, M.; Kohse-Hoinghaus, K. Combust. Flame 2008, 152, 28. (2) Norinaga, K.; Deutschmann, O. Ind. Eng. Chem. Res. 2007, 46, 3547. (3) Hidaka, Y.; Nakamura, T.; Miyauchi, A.; Shiraishi, T.; Kawano, H. Int. J. Chem. Kinet. 1989, 21, 643. (4) Yoon, S. S.; Anh, D. H.; Chung, S. H. Combust. Flame 2008, 154, 368. (5) Shukla, B.; Susa, A.; Miyoshi, A.; Koshi, M. J. Phys. Chem. A 2007, 111, 8308. (6) Somers, M. L.; McClaine, J. W.; Wornat, M. J. Proc. Combust. Inst. 2007, 31, 501. (7) Richter, H.; Howard, J. B. Prog. Energy Combust. Sci. 2000, 26, 565. (8) Perera, F. P. Science 1997, 278, 1068. (9) Balat, M. Energ Sources Part A 2008, 30, 620. (10) Han, J.; Kim, H. Renewable Sustainable Energy ReV. 2008, 12, 397. (11) Snow, T. P.; Le Page, V.; Keheyan, Y.; Bierbaum, V. M. Nature 1998, 391, 259. (12) Cook, D. J.; Schlemmer, S.; Balucani, N.; Wagner, D. R.; Steiner, B.; Saykally, R. J. Nature 1996, 380, 227. (13) Mebel, A. M.; Kislov, V. V.; Kaiser, R. I. J. Am. Chem. Soc. 2008, 130, 13618. (14) Miller, J. A.; Melius, C. F. Combust. Flame 1992, 91, 21. (15) McEnally, C. S.; Pfefferle, L. D.; Burak, A.; Kohse-Hoinghaus, K. Prog. Energy Combust. Sci. 2006, 32, 247. (16) Dagaut, P.; Cathonnet, M. Combust. Flame 1998, 113, 620. (17) Tang, W. Y.; Tranter, R. S.; Brezinsky, K. J. Phys. Chem. A 2006, 110, 2165. (18) Tang, W. Y.; Tranter, R. S.; Brezinsky, K. J. Phys. Chem. A 2005, 109, 6056. (19) Miller, J. A.; Klippenstein, S. J. J. Phys. Chem. A 2003, 107, 7783.

J. Phys. Chem. A, Vol. 114, No. 34, 2010 9055 (20) Lamprecht, A.; Atakan, B.; Kohse-Hoinghaus, K. Proc. Combust. Inst. 2000, 28, 1817. (21) Dubnikova, F.; Lifshitz, A. J. Phys. Chem. A 2002, 106, 8173. (22) Lifshitz, A.; Tamburu, C.; Suslensky, A.; Dubnikova, F. Proc. Combust. Inst. 2005, 30, 1039. (23) Friderichsen, A. V.; Shin, E.-J.; Evans, R. J.; Nimlos, M. R.; Dayton, D. C.; Ellison, G. B. Fuel 2001, 80, 1747. (24) Reale, S.; Di Tullio, A.; Spreti, N.; De Angelis, F. Mass Spectrom. ReV. 2004, 23, 87. (25) Chen, P. Unimolecular and bimolecular reaction dynamics; Wiley: New York, 1994. (26) Zhang, X.; Friderichsen, A. V.; Nandi, S.; Ellison, G. B.; David, D. E.; McKinnon, J. T.; Lindeman, T. G.; Dayton, D. C.; Nimlos, M. R. ReV. Sci. Instrum. 2003, 74, 3077. (27) Friderichsen, A. V.; Radziszewski, J. G.; Nimlos, M. R.; Winter, P. R.; Dayton, D. C.; David, D. E.; Ellison, G. B. J. Am. Chem. Soc. 2001, 123, 1977. (28) Jochnowitz, E. B.; Zhang, X.; Nimlos, M. R.; Varner, M. E.; Stanton, J. F.; Ellison, G. B. J. Phys. Chem. A 2005, 109, 3812. (29) Nandi, S.; Arnold, P. A.; Carpenter, B. K.; Nimlos, M. R.; Dayton, D. C.; Ellison, G. B. J. Phys. Chem. A 2001, 105, 7514. (30) Nandi, S.; Blanksby, S. J.; Zhang, X.; Nimlos, M. R.; Dayton, D. C.; Ellison, G. B. J. Phys. Chem. A 2002, 106, 7547. (31) Winter, P. R.; Rowland, B.; Hess, W. P.; Radziszewski, J. G.; Nimlos, M. R.; Ellison, G. B. J. Phys. Chem. A 1998, 102, 3238. (32) Zhang, X.; Kato, S.; Bierbaum, V. M.; Nimlos, M. R.; Ellison, G. B. J. Phys. Chem. A 2004, 108, 9733. (33) Zhang, X.; Maccarone, A. T.; Nimlos, M. R.; Kato, S.; Bierbaum, V. M.; Ellison, G. B.; Ruscic, B.; Simmonett, A. C.; Allen, W. D.; Schaefer, H. F. J. Chem. Phys. 2007, 126, 044312. (34) Lockyer, N. P.; Vickerman, J. C. Laser Chem. 1997, 17, 139. (35) Ramsey, S.; Orrell, D.; Bolouri, H. J. Bioinform. Comput. Biol. 2005, 3, 415. (36) Arends, I. W. C. E.; Louw, R.; Mulder, P. J. Phys. Chem. 1993, 97, 7914. (37) Lin, C. Y.; Lin, M. C. J. Phys. Chem. 1986, 90, 425. (38) Colussi, A. J.; Zabel, F.; Benson, S. W. Int. J. Chem. Kinet. 1977, 9, 161. (39) Schmoltner, A. M.; Anex, D. S.; Lee, Y. T. J. Phys. Chem. 1992, 96, 1236. (40) Liu, R. F.; Morokuma, L.; Mebel, A. M.; Liu, M. C. J. Phys. Chem. 1996, 100, 9314. (41) Moskaleva, L. V.; Lin, M. C. J. Comput. Chem. 2000, 21, 415. (42) Shafir, E. V.; Slagle, I. R.; Knyazev, V. D. J. Phys. Chem. A 2003, 107, 8893. (43) Wang, B. S.; Hou, H.; Yoder, L. M.; Muckerman, J. T.; Fockenberg, C. J. Phys. Chem. A 2003, 107, 11414. (44) Sharma, S.; Green, W. H. J. Phys. Chem. A 2009, 113, 8871. (45) McCormack, E.; Gilligan, J. M.; Cornaggia, C.; Eyler, E. E. Phys. ReV. A 1989, 39, 2260. (46) Berkowitz, J.; Ellison, G. B.; Gutman, D. J. Phys. Chem. 1994, 98, 2744. (47) Berkowitz, J.; Greene, J. P.; Cho, H.; Ruscic, B. J. Chem. Phys. 1987, 86, 674. (48) Hayaishi, T.; Iwata, S.; Sasanuma, M.; Ishiguro, E.; Morioka, Y.; Iida, Y.; Nakamura, M. J. Phys. B: At., Mol. Opt. Phys. 1982, 15, 79. (49) Erman, P.; Karawajczyk, A.; Rachlewkallne, E.; Stromholm, C.; Larsson, J.; Persson, A.; Zerne, R. Chem. Phys. Lett. 1993, 215, 173. (50) Minsek, D. W.; Chen, P. J. Phys. Chem. 1990, 94, 8399. (51) Bieri, G.; Schmelzer, A.; Asbrink, L.; Jonsson, M. Chem. Phys. 1980, 49, 213. (52) Rosenstock, H. M.; Mcculloh, K. E.; Lossing, F. P. Int. J. Mass Spectrom. Ion Processes 1977, 25, 327. (53) Lossing, F. P.; Traeger, J. C. J. Am. Chem. Soc. 1975, 97, 1579. (54) Derrick, P. J.; Asbrink, L.; Edqvist, O.; Lindholm, E. Spectrochim. Acta, Part A 1971, A 27, 2525. (55) De Leeuw, D. M.; Mooyman, R.; Delange, C. A. Chem. Phys. Lett. 1978, 54, 231. (56) Bieri, G.; Burger, F.; Heilbronner, E.; Maier, J. P. HelV. Chim. Acta 1977, 60, 2213. (57) Nemeth, G. I.; Selzle, H. L.; Schlag, E. W. Chem. Phys. Lett. 1993, 215, 151. (58) Meyer, F.; Harrison, A. G. Can. J. Chem. 1964, 42, 2256. (59) Dewar, M. J. S.; David, D. E. J. Am. Chem. Soc. 1980, 102, 7387. (60) Fuke, K.; Yoshiuchi, H.; Kaya, K.; Achiba, Y.; Sato, K.; Kimura, K. Chem. Phys. Lett. 1984, 108, 179. (61) Lichtenberger, D. L.; Renshaw, S. K.; Bullock, R. M. J. Am. Chem. Soc. 1993, 115, 3276. (62) Maier, J. P.; Turner, D. W. J. Chem. Soc., Faraday Trans. 2 1973, 521. (63) Cockett, M. C. R.; Ozeki, H.; Okuyama, K.; Kimura, K. J. Chem. Phys. 1993, 98, 7763.

9056


(64) Pedley, J. B.; Naylor, R. D.; Kirby, S. P. Thermochemistry of Organic Compounds; Chapman and Hall: New York, 1986. (65) Angel, L. A.; Ervin, K. M. J. Phys. Chem. A 2006, 110, 10392. (66) Nix, M. G. D.; Devine, A. L.; Cronin, B.; Dixon, R. N.; Ashfold, M. N. R. J. Chem. Phys. 2006, 125, 133318. (67) Ruscic, B.; Litorja, M.; Asher, R. L. J. Phys. Chem. A 2000, 103, 8625. (68) Blanksby, S. J.; Ellison, G. B. Acc. Chem. Res. 2003, 36, 255. (69) Pecullan, M.; Brezinsky, K.; Glassman, I. J. Phys. Chem. A 1997, 101, 3305. (70) Lin, C. Y.; Lin, M. C. Int. J. Chem. Kinet. 1985, 17, 1025. (71) Lovell, A. B.; Brezinsky, K.; Glassman, I. Int. J. Chem. Kinet. 1989, 21, 547. (72) Brezinsky, K.; Pecullan, M.; Glassman, I. J. Phys. Chem. A 1998, 102, 8614. (73) Khachatryan, L.; Adounkpe, J.; Dellinger, B. J. Phys. Chem. A 2008, 112, 481. (74) Moskaleva, L. V.; Madden, L. K.; Lin, M. C. Phys. Chem. Chem. Phys. 1999, 1, 3967. (75) Knight, A. E. W.; Parmenter, C. S.; Schuyler, M. W. J. Am. Chem. Soc. 1975, 97, 1993. (76) Page, R. H.; Shen, Y. R.; Lee, Y. T. J. Chem. Phys. 1988, 88, 5362. (77) Atkinson, G. H.; Parmenter, C. S. J. Mol. Spectrosc. 1978, 73, 20. (78) Kohse-Hoinghaus, K.; Atakan, B.; Lamprecht, A.; Alatorre, G. G.; Kamphus, M.; Kasper, T.; Liu, N. N. Phys. Chem. Chem. Phys. 2002, 4, 2056.

Scheer et al. (79) Kern, R. D.; Singh, H. J.; Wu, C. H. Int. J. Chem. Kinet. 1988, 20, 731. (80) Marinov, N. M.; Castaldi, M.; Pitz, W. J.; Senkan, S. M.; Melius, C. F.; Westbrook, C. K. Abstr. Pap. Am. Chem. Soc. 1996, 211, 138. (81) McGivern, W. S.; Manion, J. A.; Tsang, W. J. Phys. Chem. A 2006, 110, 12822. (82) Bachrach, S. M. J. Org. Chem. 1993, 58. (83) McLean, S.; Haynes, P. Tetrahedron 1965, 21, 2329. (84) Ikeda, E.; Tranter, R. S.; Kiefer, J. H.; Kern, R. D.; Singh, H. J.; Zhang, Q. Proc. Combust. Inst. 2000, 28, 1725. (85) Melius, C. F.; Colvin, M. E.; Marinov, N. M.; Pitz, W. J.; Senkan, S. M. 26th Symposium (International) on Combustion 1996, 685. (86) Riedle, E.; Beil, A.; Luckhaus, D.; Quack, M. Mol. Phys. 1994, 81, 1. (87) Dean, A. M. J. Phys. Chem. 1990, 94, 1432. (88) Marinov, N. M.; Pitz, W. J.; Westbrook, C. K.; Castaldi, M. J.; Senkan, S. M. Combust. Sci. Technol. 1996, 116, 211. (89) Marinov, N. M.; Pitz, W. J.; Westbrook, C. K.; Vincitore, A. M.; Castaldi, M. J.; Senkan, S. M.; Melius, C. F. Combust. Flame 1998, 114, 192. (90) Mebel, A. M.; Kislov, V. V. J. Phys. Chem. A 2009, 113, 9825. (91) Filley, J.; McKinnon, J. T. Combust. Flame 2001, 124, 721. (92) Ferguson, J.; Reeves, L. W.; Schneider, W. G. Can. J. Chem. 1957, 35, 1117.

JP102046P

Radical Chemistry in the Thermal Decomposition of Anisole and

Recommend Documents