Synchrotron Photoionization Mass Spectrometry Measurements of

Oct 14, 2013 - and identified isomeric species by their photoionization spectra. For stable .... experimental system, multiplexed synchrotron photoion...
2 downloads 0 Views 4MB Size
Download PDF
Article pubs.acs.org/JPCA

Synchrotron Photoionization Mass Spectrometry Measurements of Product Formation in Low-Temperature n‑Butane Oxidation: Toward a Fundamental Understanding of Autoignition Chemistry and n‑C4H9 + O2/s‑C4H9 + O2 Reactions Arkke J. Eskola, Oliver Welz, John D. Savee, David L. Osborn, and Craig A. Taatjes* Combustion Research Facility, Sandia National Laboratories, Mail Stop 9055, Livermore, California 94551-0969, United States S Supporting Information *

ABSTRACT: Product formation in the laser-initiated low-temperature (575−700 K) oxidation of n-butane was investigated by using tunable synchrotron photoionization time-of-flight mass spectrometry at low pressure (∼4 Torr). Oxidation was triggered either by 351 nm photolysis of Cl 2 and subsequent fast Cl + n-butane reaction or by 248 nm photolysis of 1-iodobutane or 2-iodobutane. Iodobutane photolysis allowed isomer-specific preparation of either n-C4H9 or s-C4H9 radicals. Experiments probed the time-resolved formation of products and identified isomeric species by their photoionization spectra. For stable primary products of butyl + O2 reactions (e.g., butene or oxygen heterocycles) bimodal time behavior is observed; the initial prompt formation, primarily due to chemical activation, is followed by a slower component arising from the dissociation of thermalized butylperoxy or hydroperoxybutyl radicals. In addition, time-resolved formation of C4-ketohydroperoxides (C4H8O3, m/z = 104) was observed in the n-C4H9 + O2 and Cl-initiated oxidation experiments but not in the s-C4H9 + O2 measurements, suggesting isomeric selectivity in the combined process of the “second” oxygen addition to hydroperoxybutyl radicals and subsequent internal H-abstraction/dissociation leading to ketohydroperoxide + OH. To further constrain product isomer identification, Cl-initiated oxidation experiments were also performed with partially deuterated n-butanes (CD3CH2CH2CD3 and CH3CD2CD2CH3). From these experiments, the relative yields of butene product isomers (cis-2-butene, trans-2-butene, and 1-butene) from C4H8 + HO2 reaction channels and oxygenated product isomers (2,3dimethyloxirane, 2-methyloxetane, tetrahydrofuran, ethyloxirane, butanal, and butanone) associated with OH formation were determined. The current measurements show substantially different isomeric selectivity for oxygenated products than do recent jet-stirred reactor studies but are in reasonable agreement with measurements from butane addition to reacting H2/O2 mixtures at 753 K.

1. INTRODUCTION Despite decades of intense scientific work, there is still significant, continued interest in understanding hydrocarbon oxidation.1 For transportation applications, this interest is motivated in part by the development of new engine technologies, such as homogeneous charge compression ignition (HCCI) engines, that promise to combine high efficiency with low emissions.2 In HCCI engines, the combustion phasing is controlled by autoignition of the compressed fuel−air mixture. Predicting autoignition depends on understanding the often complex yet important low-temperature oxidation chemistry of a fuel, the modeling of which in turn relies on data from diverse experiments that differ in the range of conditions covered, the level of details obtained, and their time resolution.3,4 Low-temperature heat release depends on oxidation chemistry that begins with the R + O2 reactions, in which fuel radicals R are added to oxygen. These reactions proceed through alkylperoxy (RO2) radical intermediates and exhibit a temperature- and pressure-dependent competition among product channels. Some of these channels are effectively chain-terminating and inhibit oxidation at low temperature (e.g., formation of HO2 + alkene), © 2013 American Chemical Society

some are chain-propagating and maintain oxidation at low temperature (e.g., formation of OH + an oxygen heterocycle), and some are chain-branching and amplify oxidation at low temperature (e.g., formation of two OH radicals from a single hydroperoxyalkyl + O2 (“second” oxygen) reaction).1 Hydroperoxyalkyl (QOOH) radicals are produced by isomerization of the alkylperoxy radicals and provide additional complexity to the low temperature oxidation chemistry; this is because, for QOOH, several potential decomposition channels are open, in addition to reaction with O2. Although they are critical species for lowtemperature chain branching, the hydroperoxyalkyl radicals have never been observed directly. The kinetics of the R + O2 reaction, and the branching among the different product channels, depends on pressure and temperature, often strongly.1 Chemical activation is an important source of pressure dependence and can have a significant effect on the autoignition chemistry of R + O2 reactions; in time-resolved experiments initial prompt product Received: August 23, 2013 Revised: October 13, 2013 Published: October 14, 2013 12216

dx.doi.org/10.1021/jp408467g | J. Phys. Chem. A 2013, 117, 12216−12235

The Journal of Physical Chemistry A

Article

from QOOH + O2 reactions that have so far eluded direct experimental study. DeSain et al.11 investigated the Cl-initiated oxidation of n-butane by measuring the time-resolved HO2 formation in the reaction of both n- and s-C4H9 isomers with O2 and rationalized their results in reference to ab initio calculations of the stationary points on the n-butyl + O2 and s-butyl + O2 potential energy surfaces (PESs). The current work extends this earlier study, again probing the initial steps of the laser-initiated n-butane oxidation under well-controlled conditions but now also detecting the time-resolved formation of relevant species simultaneously by synchrotron MPIMS. The measurement of the time behavior of the product formation after pulsed-laser initiation can distinguish primary from secondary products and permits a kinetic separation of the chemical activation and the sequential thermal mechanisms for the product formation. Since the work of DeSain et al., several thorough and systematic ab initio calculations on the stationary points of the alkyl + O2 and QOOH + O2 PESs have been performed by Miyoshi,20 Green and co-workers,21,22 and the Dean group.23,24 The theoretical investigations have considerably expanded the understanding of the molecular structure effects on the initial oxidation reactions of alkanes, and the present results are interpreted in the context of those calculations.

formation is primarily due to the chemical activation, which is followed by the slower component arising from the dissociation of thermalized RO2 or QOOH radicals.5 A strategy for developing a fundamental and rigorous model of low-temperature oxidation chemistry is to build from small systems, for which experiment and theory can be carried out in great detail. The reactions of ethyl and propyl radicals with O2 have been extensively investigated,1,5 and recent ab initio/master equation simulations are in good agreement with experimental observations at low pressure (10 Torr).6 The oxidation mechanisms of these C2 and C3 hydrocarbons are essential constituents of any autoignition model for larger fuels. However, these small alkyl radicals do not contain all the structural motifs that determine the low-temperature combustion properties of larger alkyl radicals, and therefore n-butyl is likely a more useful combustion archetype.7 Combining the pulsed laser initiation of a reaction with the capability to simultaneously measure multiple (isomer-specific) species, all in a time-resolved manner, is a very powerful approach to investigate the complex chemistry associated with lowtemperature combustion or autoignition. Recently a new experimental system, multiplexed synchrotron photoionization time-of-flight mass spectrometry (MPIMS),8,9 has been developed that allows such time-resolved simultaneous probing of multiple species. By scanning the energy of the ionizing photons, isomers can be identified and distinguished on the basis of their photoionization spectra. Eskola et al.10 recently measured the product formation from Cl-initiated isobutane oxidation, that is, probing the isobutyl + O2 and tert-butyl + O2 reactions, by employing MPIMS under low-temperature combustion conditions (550−700 K). The current investigation continues the study of C4 hydrocarbons and focuses on the low-temperature oxidation of n-butane (specifically, on n-butyl + O2 and s-butyl + O2 reactions). The oxidation of n-butane is a more complicated system than that of the isobutane oxidation, in which the tertbutyl + O2 reaction has only one important channel, the formation of isobutene + HO2.11 An important detailed product study of n-butane oxidation was performed by Walker and co-workers,12,13 who added small amounts of n-C4H10 to slowly reacting mixtures of H2 + O2 + N2 at 480 °C in an aged boric acid-coated vessel. Those experiments used gas chromatography to measure all the reactant and stable product concentrations. The jet-stirred reactor studies of Battin-Leclerc and her co-workers14−17 have provided a wealth of information on the temperature-dependent products of the overall oxidation. The recent rapid compression machine (RCM) measurements of the autoignition delay times of n-butane were found to be in good quantitative agreement with modeling results.18 Similarly, in another recent study in which the RCM results were combined with shock-tube measurements performed at a higher temperature, a good overall agreement between shock-tube and RCM measurements and also between the experiments and a kinetic model was found.19 Despite the agreement for these high-level targets, some details of the n- and s-butyl radical oxidation processes still require elucidation. The chemical speciation in recent low-temperature, atmospheric pressure n-butane oxidation experiments in a jetstirred reactor has provided new constraints on low-temperature butane oxidation models, but questions remain, particularly concerning the fate of the QOOH radicals.15 Jet-stirred reactor experiments using synchrotron PIMS14,16,17 have detected ketohydroperoxide products that almost certainly originated

2. EXPERIMENTAL SECTION The time-resolved MPIMS8,25 experiments were performed using the Sandia multiplexed chemical kinetics reactor at the Chemical Dynamics Beamline (9.0.2) of the Advanced Light Source (ALS). The apparatus essentially consists of a heatable slow-flow quartz reactor with an inner diameter of 1.05 cm coupled to a mass spectrometer. The excimer laser photolysis of a suitable precursor along the reactor tube produces a uniform radical profile in the reactor. The reacting mixture is continuously sampled through an ∼0.65 mm diameter orifice in the sidewall of the reactor, and a molecular beam is formed from the small gas sample escaping into a vacuum surrounding the tube. Downstream from the orifice, a skimmer selects a center part of the beam before it enters an ionization region where quasicontinuous tunable vacuum ultraviolet (VUV) synchrotron radiation crosses the molecular beam. Ions formed are separated by mass by using an orthogonal-acceleration time-of-flight (TOF) mass spectrometer and detected by using a time-sensitive microchannel plate detector.26 The measurements were repeated at 4 or 10 Hz in order to increase the signal-to-noise ratio. The flow velocity was set to completely replace the gas in the reactor between subsequent laser pulses. The total pressure in the current experiments was 4 Torr; gas flows were set using calibrated mass-flow controllers. The time-resolved mass spectra I(m/z, t) were acquired from a sequence of complete TOF mass spectra taken at a repetition rate of 50 kHz (i.e., in kinetic steps of 20 μs) over a total kinetic time window of 50 ms (10 Hz) or 150 ms (4 Hz). The data collection was started 10 ms (10 Hz) or 20 ms (4 Hz) before the photolysis pulse. Full mass-, time-, and energy-resolved data sets, I(m/z, t, and E, respectively), were obtained by recording time-resolved mass spectra I(m/z, t) as a function of the photon energy E. Typically, the photon energy was scanned from 8.9 to 10.5 eV in steps of 25 meV and normalized to the ALS photon flux that was measured by using a calibrated photodiode (SXUV100). In this work, the signal was background-corrected by subtracting the average of the prephotolysis signal. Some additional experiments were carried out at a single photon energy of 11.0 eV in 12217

dx.doi.org/10.1021/jp408467g | J. Phys. Chem. A 2013, 117, 12216−12235

The Journal of Physical Chemistry A

Article

order to probe the products (C2H4, CH2O) that have higher ionization energies. The tunability of the synchrotron radiation enables isomerresolved species identification and quantification based on the comparison of the photoionization spectra from a chemical reaction with the absolute photoionization spectra of individual isomeric species.27 The photoionization spectra are obtained by integrating the time- and energy-resolved mass spectra I(m/z, t, E) over a range of kinetic times t, in this work typically from 0 to 20 ms after photolysis. The signal S(E) at a given mass-to-charge ratio m/z and photon energy E can be expressed as a sum of contributions from the ionization of different isomeric species i: S(E)m / z = Λ ∑ αiσi(E)ci

Data sets for the chlorine-initiated oxidation were taken at 575, 650, and 700 K. Additional experiments were performed at 575 K by employing 248 nm photolysis of 1-iodobutane (n-C4H9I) or 2-iodobutane (s-C4H9I) to selectively generate either n-butyl (1-butyl) or s-butyl (2-butyl) radicals: n‐C4 H 9I + hν (248 nm) → n‐C4 H 9 + I →other products

s‐C4 H 9I + hν (248 nm) → s‐C4 H 9 + I →other products

where Λ is an experimental proportionality factor common to all species. The mass discrimination factor αi, which accounts for mass-dependent sampling efficiencies and is required for the determination of the branching ratios for species with different masses, can be expressed as a power of the ion mass mi: αi = miβ. During this work an average best value of β = 0.60 ± 0.03 was empirically determined.28 The absolute partial photoionization cross sections of the isomeric species i at a given mass, σi(E), in eq E1 are essentially basis functions whose coefficients ci are optimized to fit the normalized signal S(E)m/z. Generally, at a given m/z ratio, the relative values of ci directly reflect the branching ratios of the individual isomers. In certain cases, the dissociative ionization of a higher-mass compound on the m/z of a lower-mass product occurs and has to be taken into account. For daughter ions from the dissociative ionization, an effective mass discrimination factor is determined from calibration measurements that relate the daughter and the parent ion signals. In the fit, the coefficient ci for the daughter is then determined by the parent ion signal and can be held fixed when fitting the normalized signal S(E)m/z at the lower m/z. In the current experiments, the oxidation of n-butane, n-butane-d6 (CD3CH2CH2CD3), or n-butane-d4 (CH3CD2CD2CH3) was initiated by the excimer laser photolysis of molecular chlorine at 351 nm: Cl 2 + hν (351nm) → 2Cl

(6a) (6b)

3. RESULTS An example of the experimental data is given in Figure 1, which shows time- and energy-resolved mass spectra for the Cl-initiated

Figure 1. Experimental data for Cl-initiated n-butane oxidation at 575 K. Upper panel: energy-resolved mass spectrum obtained by integration over 0−20 ms of kinetic time. Lower panel: time-resolved mass spectrum obtained by integration over photon energies from 8.9 to 10.5 eV.

(1)

The typical initial concentrations were [Cl]0 ≈ 1 × 10 cm−3, [Cl2] ≈ 4 × 1014 cm−3, [n-butane] ≈ 7 × 1014 cm−3, and [O2] ≈ 1 × 1016 cm−3, with helium added to reach a total pressure of 4 Torr. Once formed, the chlorine atoms react rapidly with the n-butane isotopolog to produce n-butyl (2a, 3a, or 4a) and s-butyl radicals (2b, 3b, or 4b): 13

oxidation of n-butane at 575 K. The formation of products at m/z = 56 (C4H8) and m/z = 72 (C4H8O) are known primary channels of the butyl + O2 reactions:11−13 C4H8 is the coproduct of the channels producing HO2, and C4H8O is the coproduct of the OH-forming channels from the QOOH decomposition. Some other products originate from secondary or side reactions, for example, butylhydroperoxide (C4H9OOH) at m/z = 90 (see Supporting Information, Figure S1). On the other hand, the situation is more complex for the signals at m/z = 42 and m/z = 44, which contain contributions both from the direct ionization of propene (m/z = 42) and acetaldehyde (m/z = 44) and from the dissociative ionization of the higher-mass products. One notable observation in the current work is a time-resolved product signal at m/z = 104, which we attribute to ketohydroperoxides formed in the QOOH + O2 reaction.14 In the following subsections these observations are discussed in more detail. 3.1. Isomer-Resolved Primary Products of Butyl + O2 Reactions. 3.1.1. Cl-Initiated n-Butane Oxidation. Figure 2 shows the time behavior and the photoionization spectra of the

̇ 2CH 2CH 2CH3 + HCl Cl + CH3CH 2CH 2CH3 → CH (2a)

̇ Cl + CH3CH 2CH 2CH3 → CH3CHCH 2CH3 + HCl (2b)

̇ 2 CH 2CH 2CD3 + DCl Cl + CD3CH 2CH 2CD3 → CD (3a)

̇ Cl + CD3CH 2CH 2CD3 → CD3CHCH 2CD3 + HCl (3b)

̇ 2CD2 CD2 CH3 + HCl Cl + CH3CD2 CD2 CH3 → CH (4a)

̇ Cl + CH3CD2 CD2 CH3 → CH3CDCD 2 CH3 + DCl

(5b)

In the iodobutane experiments typical initial concentrations were [C4H9]0 ≈ 4 × 1012 cm−3, [C4H9I] ≈ 4 × 1013 cm−3, and [O2] ≈ 1 × 1016 cm−3, with helium added to reach a total pressure of 4 Torr.

(E1)

i

(5a)

(4b) 12218

dx.doi.org/10.1021/jp408467g | J. Phys. Chem. A 2013, 117, 12216−12235

The Journal of Physical Chemistry A

Article

Figure 2. Comparison of time behaviors and photoionization spectra of the important products C4H8 (m/z = 56), C4H8O (m/z = 72), and C4H8O3 (m/z = 104, ketohydroperoxide) characterizing Cl-initiated n-butane oxidation at 575 K. (a, c, e) Time traces obtained by integration over photon energies from 8.9 to 10.5 eV. (b, d, f) Photoionization spectra obtained by integration over 0−20 ms of kinetic time. The black line in panel f is a photoionization spectrum of the m/z = 104 product obtained by Battin-Leclerc et al.14 in their n-butane low-temperature atmospheric pressure oxidation experiments at 630 K.

product mass peaks can be revealed and quantified. Figures 2b,d show the measured photoionization spectra of the m/z = 56 and m/z = 72 products and the corresponding fits using the measured absolute photoionization cross-section spectra of the pure compounds. Also shown are the absolute photoionization cross-section spectra of the individual isomers, which have been scaled by their relative contributions as determined from the fit to the normalized signal S(E)m/z (see eq E1). The products included in the fits are those predicted from the ab initio calculations of the stationary points on the n-C4H9 + O2 and sC4H9 + O2 PESs,11,20,22−24 depicted in Figure 3. On the basis of these calculations, the following product channels (eqs 7) for the n-C4H9 + O2 reaction can be deduced, which are also shown in Scheme 1.

important products C4H8 (m/z = 56), C4H8O (m/z = 72), and C4H8O3 (m/z = 104, ketohydroperoxide) from the Cl-initiated n-butane oxidation at 575 K. These time traces and photoionization spectra are obtained by further integrating the timeand energy-resolved mass spectra shown in Figure 1 over a specific m/z range. It is clear from Figure 2a that the time behavior of m/z = 56 is composed of two components: the initial prompt formation due to “formally direct” or “well-skipping” channels, principally the chemical activation, and the slower, longer time scale formation arising from the unimolecular dissociation of thermalized butylperoxy or hydroperoxybutyl radicals. An identical time behavior is observed for m/z = 72, as shown in Figure 2c in which the signal m/z = 72 is superimposed with the signal m/z = 56. However, the signals at both m/z = 56 and m/z = 72 are composed of more than one isomer. By fitting the measured photoionization spectra of the m/z = 56 and m/z = 72 products to the calibration spectra of the pure compounds, which have been measured individually in separate experiments using the same apparatus, the isomeric composition of the

n‐C4 H 9 + O2 → n‐C4 H 9OO → HO2 + 1‐butene

(7a)

̇ n‐C4 H 9 + O2 → n‐C4 H 9OO → CH3CH 2CHCH 2OOH → HO2 + 1‐butene 12219

(7b1)

dx.doi.org/10.1021/jp408467g | J. Phys. Chem. A 2013, 117, 12216−12235

The Journal of Physical Chemistry A

Article

Figure 3. Stationary point energies and reaction paths for reactions of 1-butyl (n-butyl) and 2-butyl (s-butyl) radicals with O2, taken from the calculations of DeSain et al.11,21

̇ n‐C4 H 9 + O2 → n‐C4 H 9OO → CH3CH 2CHCH 2OOH → OH + ethyloxirane

̇ s‐C4 H 9 + O2 → s‐C4 H 9OO → CH3CHCH(OOH)CH 3 → HO2 + 2‐butene

(7b2)

̇ s‐C4 H 9 + O2 → s‐C4 H 9OO → CH3CHCH(OOH)CH 3

̇ n‐C4 H 9 + O2 → n‐C4 H 9OO → CH3CHCH 2CH 2OOH → OH + 2‐methyloxetane

→ OH + 2,3‐dimethyloxirane

(7c)

→ OH + 2‐methyloxetane

(7d)

(8a)

̇ 2 s‐C4 H 9 + O2 → s‐C4 H 9OO → CH3CH 2CH(OOH)CH → HO2 + 1‐butene

(8b1)

̇ 2 s‐C4 H 9 + O2 → s‐C4 H 9OO → CH3CH 2CH(OOH)CH → OH + ethyloxirane s‐C4 H 9 + O2 → s‐C4 H 9OO → HO2 + 2‐butene

(8e)

In a fit of the measured photoionization spectrum of m/z = 56, it is also possible to distinguish between cis- and trans-2-butene products, which were observed in addition to 1-butene, as is shown in Figure 2b. Although the ionization energies of cis(9.11 ± 0.01 eV)29 and trans-2-butene (9.10 ± 0.01 eV)29 are very similar, the shapes of their photoionization spectra are sufficiently different over the range of energies employed, and the signal-to-noise ratio of the m/z = 56 product spectrum from the Cl-initiated n-butane oxidation is high enough, that it is possible to quantitatively distinguish these two isomers. A fit that includes the C4H8O isomers 2-methyloxetane, tetrahydrofuran, 2,3-dimethyloxirane, and ethyloxirane reproduces the measured photoionization product spectrum of m/z = 72 from the Cl-initiated n-butane oxidation, as shown in Figure 2d. The close

Similarly for the s-C4H9 + O2 reaction from Figure 3 and Scheme 2 (eqs 8): s‐C4 H 9 + O2 → s‐C4 H 9OO → HO2 + 1‐butene

(8d2)

̇ 2CH 2CH(OOH)CH3 s‐C4 H 9 + O2 → s‐C4 H 9OO → CH

̇ 2CH 2CH 2CH 2OOH n‐C4 H 9 + O2 → n‐C4 H 9OO → CH → OH + tetrahydrofuran

(8d1)

(8b2) (8c) 12220

dx.doi.org/10.1021/jp408467g | J. Phys. Chem. A 2013, 117, 12216−12235

The Journal of Physical Chemistry A

Article

Scheme 1. Reaction Paths for the n-C4H9 + O2 Reaction According to the Current Work

Scheme 2. Reaction Paths for the s-C4H9 + O2 Reaction According to the Current Work

of the Supporting Information). In the case of the n-C 4H 9 + O 2 reaction, the only observed C 4H 8 product at m/z = 56 is 1-butene (see Figure S3 of the Supporting Information), as expected (see Figure 3). The situation is more complicated for the m/z = 72 product channel due to the presence of several isomers; in order to properly fit the data, 2-methyloxetane, tetrahydrofuran, butanal, and ethyloxirane were included in the fit (see Figure S4 of the Supporting Information). Although not included as a product channel in the PES shown in Figure 3, butanal could still be a direct product of the n-C4H9 + O2 reaction with an intervening internal hydrogen migration, for example, an isomerization of 4-hydroperoxy-1-butyl radical, as indicated in eq 9:

similarity of the cis- and trans-2,3-dimethyloxirane spectra and the presence of several other isomers prevent separate quantification of the two 2,3-dimethyloxirane isomers. Instead, the absolute photoionization cross-section spectrum that is employed for 2,3-dimethyloxirane is the average of the measured spectra of its cis and trans isomers, which were measured individually using the current experimental apparatus. 3.1.2. Oxidation of Isomeric Butyl Radicals. Figure 4 considers the isomer-specific n-C4H9 + O2 reaction, which was initiated by hv (248 nm) photolysis of n-C4H9I at 575 K. In Figure 4a, the time profiles of m/z = 56 and m/z = 72 are shown. Comparing Figures 2a and 4a it can be seen that the yield of m/z = 72 relative to that of m/z = 56 has increased compared to the relative yields from the Cl-initiated n-butane oxidation (where the s-C4H9 + O2 reaction also contributes to the product spectrum), yet the time profiles overlap precisely (see Figure S2

̇ 2CH 2CH 2CH 2OOH → CH3CH 2CH 2CHOOH ̇ CH → CH3CH 2CH 2CHO + OH 12221

(9)

dx.doi.org/10.1021/jp408467g | J. Phys. Chem. A 2013, 117, 12216−12235

The Journal of Physical Chemistry A

Article

Figure 4. Comparison of time behaviors (C4H8 (m/z = 56), C4H8O (m/z = 72), and C4H8O3 (m/z = 104, ketohydroperoxide)) and the photoionization spectrum (C3H6 (m/z = 42)) of products formed via photolytically initiated (n-C4H9I, hv (248 nm)) n-C4H9 radical oxidation at 575 K. Time traces are obtained by integrating over photon energies from 8.9 to 10.5 eV, and the photoionization spectrum at m/z = 42 is obtained by integrating over 0−20 ms of kinetic time. In panel b, 40 time bins (20 μs each) have been integrated (instead of 10 elsewhere) to improve the signal-to-noise ratio, corresponding to an effective dwell time of 800 μs.

Figure 5. Comparison of time traces and photoionization spectra of C4H8 (m/z = 56) and C4H8O (m/z = 72) products formed in photolytically initiated s-C4H9 oxidation at 575 K. (a) Time traces obtained by integration over photon energies from 8.9 to 10.5 eV. (b) Photoionization spectra obtained by integration over 0−20 ms of kinetic time.

Figure 5 displays the data from the isomer-specific s-C4H9 + O2 reaction, which was initiated by the 248 nm photolysis of s-C4H9I at 575 K. In Figure 5a, the time behavior of the signals at m/z = 56 and m/z = 72 is shown. Again, because of the contributions from the formally direct and the sequential, thermal dissociation product channels, the signal at m/z = 56 is composed of two components; this applies to the time behavior of the signal at m/z = 72 as well, shown in Figure S5 of the Supporting Information. Comparing Figure 5a with Figures 3a and 4a, it is obvious that the signal of m/z = 72 relative to that at m/z = 56 has its lowest value among these three data sets, suggesting that the chain-propagating OH + cyclic ether channels are less important for the s-C4H9 + O2 reaction than they are for the n-C4H9 + O2 reaction. Three C4H8

isomers were observed as products in the s-C4H9 + O2 reaction, namely, 1-butene, cis-2-butene, and trans-2-butene (see Figure 5b), in accordance with the ab initio calculations (see Figure 3). The product signal at m/z = 72 was too weak and noisy to extract meaningful branching fractions from a fit of all the potential isomeric product calibration spectra to the data. 3.1.3. Cl-Initiated Oxidation of Partially Deuterated n-Butanes. The previous section showed that the signals at m/z = 56 and m/z = 72 are composed of several product isomers. Measurements employing the partially deuterated n-butanes, CD3CH2CH2CD3 or CH3CD2CD2CH 3, instead of undeuterated n-butane, CH3CH2CH2CH3, can help to resolve the active mechanisms leading to the isomeric products in more detail. In the oxidation of 12222

dx.doi.org/10.1021/jp408467g | J. Phys. Chem. A 2013, 117, 12216−12235

The Journal of Physical Chemistry A

Article

Scheme 3. Reaction Paths for the n-C4H4D5 + O2 Reaction According to the Current Work

the partly deuterated butanes, the butene and the cyclic ether isomers are separated by mass, which simplifies the photoionization product spectra at these masses and therefore facilitates the interpretation of the spectra. 3.1.3.1. CD3CH2CH2CD3. In the oxidation of CD3CH2CH2CD3, the butene product isomers appear at m/z = 61 (1-butene-d5, CD3CH2CHCD2) and m/z = 62 (cis/trans-2-butene-d6, CD3CHCHCD3), and the C4H8O isomers are spread out among m/z = 76 (tetrahydrofuran-d 4), m/z = 77 (2methyloxetane-d5 and ethyloxirane-d5), and m/z = 78 (2,3dimethyloxirane-d6). From the ab initio calculations presented in Figure 3, the following reaction channels can be deduced for the oxidation of the n-butyl radical, n-C4H4D5, formed in the Cl + CD3CH2CH2CD3 reaction (see also Scheme 3):

s‐C4 H3D6 + O2 → s‐C4 H3D6 OO ̇ 2 → DO2 + 1‐butene‐d5 → CD3CH 2CH(OOD)CD (11b1)

s‐C4 H3D6 + O2 → s‐C4 H3D6 OO ̇ 2 → OD + ethyloxirane‐d → CD3CH 2CH(OOD)CD 5 (11b2)

s‐C4 H3D6 + O2 → s‐C4 H3D6 OO → HO2 + 2‐butene‐d6 (11c)

s‐C4 H3D6 + O2 → s‐C4 H3D6 OO ̇ → CD3CHCH(OOH)CD 3 → HO2 + 2‐butene‐d6

n‐C4 H4D5 + O2 → n‐C4 H4D5OO → HO2 + 1‐butene‐d5

(11d1)

(10a)

s‐C4 H3D6 + O2 → s‐C4 H3D6 OO

n‐C4 H4D5 + O2 → n‐C4 H4D5OO

̇ → CD3CHCH(OOH)CD 3

̇ → CD3CH 2CHCD 2 OOH → HO2 + 1‐butene‐d5

→ OH + 2,3‐dimethyloxirane‐d6

(10b1)

s‐C4 H3D6 + O2 → s‐C4 H3D6 OO

n‐C4 H4D5 + O2 → n‐C4 H4D5OO

̇ 2 CH 2CH(OOD)CD3 → CD

̇ → CD3CH 2CHCD 2 OOH → OH + ethyloxirane‐d5

→ OD + 2‐methyloxetane‐d5

(10b2)

(11e)

Figure 6 shows the time profiles and the photoionization spectra of the butene products at m/z = 61 (1-butene-d5, CD3CH 2CHCD2) and m/z = 62 (cis/trans-2-butene-d6 , CD3CHCHCD3) from the Cl-initiated oxidation of the partially deuterated n-butane-d6, CD3CH2CH2CD3, at 650 K. At this temperature the difference in time scales of the formally direct and the sequential thermal product formation is less pronounced than they are in the undeuterated n-butane experiments at 575 K, as can be observed from Figure 6a. From the figure one can also observe a small difference in the time behavior of the 1-butene-d5 (m/z = 61) and 2-butene-d6 (m/z = 62) signals. This difference is a convolution of differences in the kinetics of the radical reactions with O2 (k(CD3CH2Ċ HCD3 + O2) vs k(CD3CH2CH2Ċ D2 + O2)), and the dissociation kinetics of the partially deuterated

n‐C4 H4D5 + O2 → n‐C4 H4D5OO ̇ → CD3CHCH 2CD2 OOH → OH + 2‐methyloxetane‐d5

(11d2)

(10c)

n‐C4 H4D5 + O2 → n‐C4 H4D5OO ̇ 2 CH 2CH 2CD2 OOD → OD + tetrahydrofuran‐d → CD 4 (10d1)

Similarly for the s-butyl radical, s-C4H3D6 (see also Scheme 4): s‐C4 H3D6 + O2 → s‐C4 H3D6 OO → DO2 + 1‐butene‐d5 (11a) 12223

dx.doi.org/10.1021/jp408467g | J. Phys. Chem. A 2013, 117, 12216−12235

The Journal of Physical Chemistry A

Article

Scheme 4. Reaction Paths for the s-C4H3D6 + O2 Reaction According to the Current Work

Figure 6. Experimental time traces and photoionization spectra of 1-butene (m/z = 61, C4H3D5) and 2-butene (m/z = 62, C4H2D6) products formed in Cl-initiated n-butane-d6 (CD3CH2CH2CD3) oxidation at 650 K. (a) Time traces obtained by integration over photon energies from 8.9 to 10.5 eV. (b) Photoionization spectra obtained by integration over 0−20 ms of kinetic time.

peroxy radicals to bimolecular products. Kinetic isotope effects also contribute to the differences; interestingly, in the CH3CD2CD2CH3 oxidation the time behavior of 1-butene-d3 is nearly identical to that of 2-butene-d2 (see Figure S6 of the Supporting Information). Figure 7a−c shows the photoionization spectra of the oxygenated products at m/z = 76, m/z = 77, and m/z = 78. As expected, the main product observed at m/z = 76 is tetrahydrofuran-d4 (see Figure 7a). However, a fit of the m/z = 76 photoionization spectrum can be significantly improved by including the contributions from butanal (