Pressure-Dependent Competition among Reaction Pathways from

Subscriber access provided by La Trobe University Library

Article

Pressure-Dependent Competition among Reaction Pathways from First- and Second-O Additions in the Low-Temperature Oxidation of Tetrahydrofuran 2

Ivan O. Antonov, Judit Zádor, Brandon Rotavera, Ewa Papajak, David L. Osborn, Craig Allen Taatjes, and Leonid Sheps J. Phys. Chem. A, Just Accepted Manuscript • DOI: 10.1021/acs.jpca.6b05411 • Publication Date (Web): 21 Jul 2016 Downloaded from http://pubs.acs.org on July 24, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

The Journal of Physical Chemistry A is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Pressure-Dependent Competition among Reaction Pathways from First- And Second-O2 Additions in the Low-Temperature Oxidation of Tetrahydrofuran Ivan O. Antonov, Judit Zádor, Brandon Rotavera, Ewa Papajak, David L. Osborn, Craig A. Taatjes, and Leonid Sheps* Combustion Research Facility, Sandia National Laboratories, Livermore, CA 94551, USA

Corresponding author: [email protected] Tel: (925) 294-2927

1 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 29

ABSTRACT We report a combined experimental and quantum chemistry study of the initial reactions in lowtemperature oxidation of tetrahydrofuran (THF). Using synchrotron-based time-resolved VUV photoionization mass spectrometry, we probe numerous transient intermediates and products at P = 10 – 2000 Torr and T = 400 – 700 K. A key reaction sequence, revealed by our experiments, is the conversion of THF-yl peroxy to hydroperoxy-THF-yl radicals (QOOH), followed by a second O2 addition and subsequent decomposition to dihydrofuranyl hydroperoxide + HO2 or to γ-butyrolactone hydroperoxide + OH. The competition between these two pathways affects the degree of radical chain-branching and is likely of central importance in modeling the autoignition of THF. We interpret our data with the aid of quantum chemical calculations of the THF-yl + O2 and QOOH + O2 potential energy surfaces. Based on our results, we propose a simplified THF oxidation mechanism below 700 K, which involves the competition among unimolecular decomposition and oxidation pathways of QOOH.


Page 3 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


INTRODUCTION The combustion chemistry of tetrahydrofuran (THF) is of particular interest due to its fundamental role as a prototypical cyclic ether and its importance as a structural building block for proposed biofuels. The reactivity of THF in the low-temperature autoignition regime is governed by the temperature- (T) and pressure- (P) dependent competition between oxidation and unimolecular decomposition or isomerization pathways. THF combustion offers a way to explore the effects of molecular structure on reactivity, especially by comparing to analogous hydrocarbon fuels. Its five-membered ring constrains some of the isomerization pathways that are available to acyclic species; on the other hand, the oxygen atom in the ring lowers the strength of C−O and α-C−H bonds in THF, leading to ring-opening and β-scission pathways that are not favored in simple alkanes. From a practical standpoint, 2-methyl THF and 2,5-dimethyl THF have been suggested as promising biofuel alternatives to petroleum-based energy sources, based on engine performance and the potential for large-scale production from non-edible feedstock. Recently, a number of possible routes were identified to form 5- and 6-membered cyclic ether compounds similar to THF from lignocellulosic biomass by pyrolysis or catalytic conversion.1-4 Engine tests of THF derivatives show good knock resistance and stability,5 but higher NOx emissions than in conventional fuels, highlighting a need to better understand their combustion chemistry. In contrast to first-generation biofuels (ethanol and long-chain esters),6 these proposed alternatives share a cyclic ether core structure, and detailed understanding of their reactivity will be crucial to their efficient use in engines. Cyclic ethers are also ubiquitous products of hydrocarbon oxidation and may occur in large quantities in combustion systems even when they are not part of the fuel blend.7 However, their reactions are frequently missing from chemical models.8 At temperatures below ~700 K the framework for cyclic ether oxidation is similar to that of hydrocarbon fuels,9 as shown in Scheme 1. Throughout the manuscript we use the letters R, Q, P as short-hand notation for consecutive H-abstractions from the core molecular structure of the fuel: R is the radical produced by H loss from the fuel; Q is formed by H loss from R; P is the result of H loss from Q.

Scheme 1. A simplified mechanism of low-T oxidation of THF. The atom numbering convention is shown on the far left, and two possible THF-yl radical structures (α-THF-yl and β-THF-yl) are also shown.

The first step in THF oxidation is H-atom abstraction (by OH or other species) to form two 3 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 29

different THF-yl radicals: α-R (THF-2-yl) and β-R (THF-3-yl). Similar to other hydrocarbon fuels, these radicals can react with O2 to produce peroxy radicals (ROO) that can in turn decompose, self-react, or isomerize via internal H-abstraction to carbon-centered hydroperoxyalkyl radicals (QOOH). A crucial step in low-T autoignition is the so-called “second O2 addition” to QOOH, followed by another internal H-shift and OH elimination to form a ketohydroperoxide (KHP) intermediate (HOOP=O). KHP can decompose to OH + oxy-radical (OP=O) in a chain-branching sequence that ultimately yields three reactive radicals for every fuel radical consumed. There are only two direct kinetic studies of the elusive QOOH radicals: one derived from tert-butyl-hydroperoxide10 and one from 1,3-cycloheptadiene.11 The latter study is currently the only report of direct detection of QOOH. In contrast, closed-shell KHP intermediates are less reactive than QOOH and have been experimentally observed in dimethyl ether12 and butane13-16 oxidation studies using steady-state reactors. Second- and third-O2 addition pathways are an active research area in combustion17 and in atmospheric chemistry;18 however, experimental data on KHP formation and decomposition are still scarce. Scheme 1 is a good starting point to describe THF combustion; however, some reaction pathways of THF differ from those of straight-chain or branched alkane fuels, owing to the heterocyclic THF structure. For example, bond scission may occur in THF at different stages of oxidation. THF-yl radicals can ring-open, and at high T this is in fact their dominant fate.19-20 Ring opening may also be possible in QOOH and in products of second O2 addition. Similar pathways have been proposed in cyclohexane oxidation,21 but ring opening in THF should be more favored than in cycloalkanes due to the O atom, which lowers the energy barrier for the C5–O bond scission in α-THF-yl by ~12 kcal/mol,19 compared to the C–C bond scission in cyclopentyl.22-23 In addition, the strain of the relatively rigid 5-membered ring of THF is expected to significantly affect the intramolecular H-transfer pathways by raising the energy of their transition states. Another consequence of this rigidity is that various intermediates may be formed as cis- or trans- isomers which will in turn affect their subsequent reaction pathways. Work on high-temperature THF combustion began with thermal decomposition studies more than 60 years ago.24-25 Later, THF pyrolysis in a shock tube was examined by Lifshitz et al.26-27 They found that THF decomposes by ring breaking to C2H4O + C2H4 or C3H6 + CH2O or by C– H bond scission to THF-yl radicals, which ultimately yield CH3, CO and C2H4. More recently, a high-T shock tube and jet-stirred reactor study by Dagaut et al.28 proposed a detailed mechanism and used kinetic modeling with sensitivity analysis to identify major reaction pathways. Kasper et al.29 studied THF oxidation in a low-P flame with photoionization mass spectrometry, utilizing tunable synchrotron vacuum-ultraviolet (VUV) radiation. Their technique was sensitive to isomeric composition and allowed identification of numerous unstable intermediates, including radical species. A 2015 study by Tran et al.20 combined experiments and modeling of flames up to P = 1 atm and high-P ignition delays to develop an updated high-temperature THF combustion model. Simmie19 applied electronic structure calculations to determine barrier heights for hydrogen 4 ACS Paragon Plus Environment

Page 5 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


abstraction by H and CH3 from THF and for β-scission of the resulting THF-yl radicals. Verdicchio et al.30 studied the unimolecular decomposition of THF at the CBS-QB3 level. They found that THF decomposition is initiated by reactions involving diradical and carbene intermediates, and built a pyrolysis model that reproduced the results of Lifshitz et al.26-27 The current consensus is that THF combustion above ~1000 K is dominated by ring-opening and C–H bond scission pathways. THF-yl radicals undergo β-scissions to ultimately form formaldehyde + allyl, ethene + vinoxy, or the unsaturated compounds 2,3- and 2,5-dihydrofuran (DHF). Below ~1000 K the main β-scission pathway is the formation of 1-oxobutyl radicals and their decomposition to CO + propyl; this pathway may compete with oxidation reactions at some conditions. Decomposition of THF itself occurs at higher temperatures (T > 1200 K) via carbene or diradical intermediates, which then decompose by β-scission and internal H-abstraction, primarily to formaldehyde + propene or to 1-butenal. In contrast, few low-temperature THF oxidation studies are available. Molera et al.31 investigated THF oxidation at 493 K in a static reactor and identified 27 primary and secondary oxidation products. A key finding from their work was that ~66% of THF, consumed in the early stages of oxidation, forms succinic acid, HOC(O)CH2CH2C(O)OH. A recent study by Vanhove et al.32 probed the low-T oxidation chemistry and ignition delays of THF in a jet-stirred reactor and a rapid compression machine, coupled to gas chromatography and mass spectrometry detection. The authors observed two-stage ignition, significant low-T reactivity of THF starting at 550 K, and deviation from Arrhenius behavior at 680 – 810 K. They detected a variety of C1C4 products including alkanes, aldehydes, heterocycles, and esters, and proposed a model to explain their formation. While not a definitive mechanism, this low-T oxidation scheme does suggest that reactions of QOOH radicals give rise to many of the products observed in their experiments. In this work we studied the primary oxidation reactions of THF at T = 400 – 700 K and P = 10 – 2000 Torr, with a particular focus on second O2 addition to QOOH. We used time-resolved multiplexed photoionization mass spectrometry (PIMS) to detect and identify numerous products and intermediates involved in the initial steps of THF oxidation. To support our experimental observations, we calculated the potential energy surfaces (PES) for α-THF-yl + O2, β-THF-yl + O2, and αα'-QOOH + O2 (see the Results section for notation). The combined experimental and theoretical results show that reactions of just one of the 6 possible QOOH isomers dominate at our experimental conditions, constraining the mechanism of THF autoignition. EXPERIMENTAL METHODS Our experiments were performed in a time-resolved PIMS apparatus, using primarily the tunable VUV photoionization radiation at beamline 9.0.2 of the Lawrence Berkeley Laboratory Advanced Light Source synchrotron. In some datasets, when tunable VUV radiation was not needed for species identification, we used the emission of an H2 discharge lamp with an MgF2 high-energy cutoff filter, providing broadband (8.5 – 10.2 eV) ionizing radiation. Separate flow 5 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 29

reactors were used in this study for low- and high-P operation. We initiated THF oxidation in both reactors by Cl radical attack, generating Cl atoms via 248-nm photolysis of oxalyl chloride, (COCl)2, by an excimer laser. Constant T was achieved in both reactors by resistive heating; constant P was maintained by downstream throttling valves. Metered flows of THF, oxygen, (COCl)2, and He buffer gas were mixed in situ prior to entering the reactors, and flow velocities were always sufficient to completely renew the gas mixture between the photolysis laser pulses. The low-P reactor was described in detail in previous publications.33-34 Briefly, experiments were carried out in a 62 cm-long, 1.05 cm-inner diameter quartz tube. The photolysis laser fired along the tube and created a uniform axial distribution of Cl atoms. This, together with rapid radial diffusion at P = 10 Torr, used in this study, ensured that the sampled gas mixture was practically homogeneous at all times. A small fraction of the reaction mixture was withdrawn through a 600-µm pinhole in the side of the reactor into the ionization region of the mass spectrometer. The high-P reactor, operating on a similar principle to the low-P version, is a heated Inconel chamber (72 mm long by 76 mm outer diameter), capable of withstanding pressures over 100 atm at T = 1000 K. The reaction volume is in the central bore of the chamber (4 cm long, 0.5 cm diameter), surrounded by a quartz insert in order to minimize unwanted reactions on heated metal surfaces. In contrast to the low-P operation, diffusion at elevated pressures is slow, and special care is needed to maintain a sample that is homogeneous and free from temperature gradients. To ensure complete mixing, a helical static mixer was positioned in the gas inlet line immediately before the reactor. The photolysis laser was admitted through a quartz window at one end of the reactor and fired along its axis. An interchangeable sampling pinhole was positioned at the opposite end of the reactor, in line with the sample flow, which minimized the possible effects of radial inhomogeneity. The present study used a 100 µm-diameter pinhole orifice, chosen to obtain the appropriate sample flow rate into the mass spectrometer at our conditions. After exiting either reactor, the sampled mixture was expanded into a vacuum chamber, skimmed, and crossed with the ionizing radiation beam. The resulting ions were collimated and detected with a pulsed orthogonal time-of-flight (TOF) mass spectrometer.33-34 The mass resolution, m/∆m ≈ 1500, was sufficient to determine the elemental composition of small molecules: e.g., mass peaks corresponding to ketene (C2H2O, m/z = 42.011) and propene (C3H6, m/z = 42.047) could be easily resolved. The mass spectrometer operated at 50 kHz, acquiring a complete mass spectrum every 20 µs during a time window –20 ms < t < 130 ms, relative to the arrival of the photolysis laser pulse (t = 0). The ionization photon energy was scanned in steps of 0.025 eV over the range 8.2 – 11.2 eV to create a 3-dimensional data set I(m/z, t, E). Ion signals were normalized to the VUV radiation flux with a NIST-calibrated photodiode. The PIMS ion signal is described by the formula35-36 / , = Λ ∙ / ∙ ∙ .


Page 7 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


The ion signal / , is a sum of contributions from all isomers at the mass-to-charge ratio m/z. The quantities and are the photoionization (PI) cross-section and timedependent concentration of isomer i; / is the isomer-independent mass discrimination factor, and Λ is the isomer-, mass- and energy-independent sensitivity factor. The value of / was

determined daily using a calibration gas mixture and parameterized as / = ∙ ⁄ for interpolation; Λ was also measured every day, using the ion signal of THF (m/z = 72), whose absolute concentration was known. This method allows quantification of all species with known ionization cross-sections. In this work we recorded reference PI spectra of THF, 2,3-DHF, 2,5DHF, γ-butyrolactone, tetrahydrofuran-3-one, and butanedial following the procedure described previously.37 Published cross-sections were used for methyl and formaldehyde.34, 38

The flows of He and O2 were 99.9999% pure; the liquid reactants THF and oxalyl chloride (99.9% pure) were used as gases diluted in He. The initial Cl atom concentration, [Cl]0, was estimated using the known (COCl)2 concentration, its UV absorption cross-section39 at 248 nm (σ248 = 3.1·10-19 cm2), an assumed quantum yield Φ = 2 for Cl atom production, and the measured laser fluence (~70 mJ·cm-2). To identify specific product molecular structures we performed several experiments using partially deuterated 2,2,5,5-THF-d4 (99% D atom content, CDN Isotopes). Experimental conditions of the main datasets are summarized in Table 1. Additional experiments were conducted when needed, changing [O2] or [Cl]0 to distinguish the primary ROO decomposition products from secondary radical-radical products. Table 1. Summary of experimental conditions. We used photoionization energy scans from 8.2 to 11.2 eV to identify products and single-energy experiments to measure product yield dependence on T, P, or [O2]. An “X” in the last column indicates THF-d4 experiments, performed at the same experimental conditions. T, K 550 600 650 700 550 600 650 400 − 650 575 575

P, Torr

Concentrations, cm-3 [THF] [O2]

Energy, eV

[Cl]0

THF-d4

10

3.5 × 1014

1.0 × 1017

7.0 × 1012

8.2-11.2

X

10

3.5 × 10

14

1.0 × 10

17

7.0 × 10

12

8.2-11.2

X

1.3 × 10

14

1.0 × 10

17

7.0 × 10

12

8.2-11.2

X

1.3 × 10

14

7.0 × 10

12

10.2

-

6.7 × 10

14

5.1 × 10

13

8.2-11.2

-

6.4 × 10

14

4.5 × 10

13

8.2-11.2

X

1.0 × 10

15

8.0 × 10

13

8.2-11.2

2.2 × 10

15

1.1 × 10

14

broadband

1.2 × 10

15

1.1 × 10

14

11.0

-

1.2 × 10

15

1.1 × 10

14

11.0

-

10 10 1500 1500 1500 1500 340 − 1950 1500

0.01 − 1 × 10 3.1 × 10

18

3.0 × 10

18

3.0 × 10

18

3.0 × 10

18

1.2 × 10

18

6 − 50 × 10

17

17

X a

-

a

Ionization by H2 discharge lamp with broadband 8.5 - 10.2 eV emission, rather than with synchrotron VUV radiation.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 29

COMPUTATIONAL METHODS Potential energy surfaces of α-THF-yl + O2, β-THF-yl + O2 and QOOH + O2 reactions were calculated using KinBot software,40-41 an expert system that automatically explores the PES of interest using a set of generic rules to search for possible reaction pathways. KinBot constructs initial guesses for the structures of stationary points and bimolecular products and optimizes them using Gaussian 0942 electronic structure calculations. KinBot systematically searches for interconnected species and attempts to find all possible pathways within a user-defined energy range and network size. In this work the energy was restricted to not exceed 40 kcal/mol above the initial ROO adduct, and the network was explored up to species separated from ROO by 2 saddle points. The geometries and energies of stationary points were initially obtained by KinBot using B3LYP with a 6-31+G* basis set and were further refined using the CBS-QB3 approach. Adiabatic and vertical ionization energies (IE and VIE, respectively) of key species for which literature values were not available were calculated with Gaussian 09 by the CBS-QB3 method. Adiabatic energies were calculated at optimized neutral and cation geometries. VIE values were calculated at the optimized geometry of neutral species, when cation optimization led to species structurally different from the neutral (e.g. ROO+ dissociates to R+ + O2 upon ionization). RESULTS AND DISCUSSION Calculated potential energy surfaces The calculated PES for the reactions α-R + O2 and β-R+O2, shown in Fig. 1, give a theoretical framework for understanding the initial steps of THF oxidation. The energies of all stationary points throughout this report are referenced to that of the reactants for the PES being discussed. Both α- and β-ROO are formed over deep potential wells of -36.8 and -34.2 kcal/mol, respectively. Both peroxy radicals can directly eliminate HO2, forming either exclusively 2,3dihydrofuran (2.3-DHF) from α-ROO (-8.0 kcal/mol barrier) or a mix of 2,3- and 2,5-DHF from β-ROO (-2.8 and -4.9 kcal/mol barriers, respectively). In principle, ROO can produce carbonyl compounds + OH, but these pathways seem unlikely due to high barriers: γ-butyrolactone (GBL) from α-ROO (3.2 kcal/mol barrier) and tetrahydrofuran-3-one (THF-3-one) from β-ROO (7.8 kcal/mol barrier). Finally, each ROO can isomerize to one of 3 possible QOOH. We denote the QOOH species by a pair of letters (α or β) to indicate the location first of the –OOH group and then the unpaired electron relative to the O atom in the ring; the prime symbol indicated locations on the opposite sides of the ring (see Fig. 1). The α-ROO radical can isomerize to one of three different QOOH: αα'-QOOH, αβ'-QOOH, or αβ-QOOH. The lowest-energy pathway is isomerization to αα'-QOOH over a barrier of -16.0 kcal/mol, followed by dissociation to OH + butanedial (-10.1 kcal/mol barrier). The second lowest barrier leads to αβ'-QOOH, which has no low-energy decomposition pathways and most likely converts back to ROO. The highest-energy isomerization is to αβ-QOOH that can further decompose to HO2 + 2,3-DHF or OH + 2,3-epoxy-THF. The difference in the barrier heights 8 ACS Paragon Plus Environment

Page 9 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


leading to QOOH is due in part to the heterocyclic O atom that weakens the neighboring α-C−H bonds, and in part to the 6-membered ring transition state (TS) in αα'- and αβ'-QOOH, as opposed to 5-membered ring TS in the case of αβ-QOOH.

Figure 1. Calculated PES for α-R + O2 (top) and β-R + O2 (bottom), limited to energies less than 40 kcal/mol above the ROO well. All energies are calculated using the CBS-QB3 method, ZPEcorrected at 0 K, and shown relative to the reactants (R + O2). The lowest-energy pathways are drawn in bold. For clarity, some product channels are shown as simple arrows without explicitly drawn saddle points; the energies for these transition states are superimposed on the arrows.

The lowest-energy isomerization of β-ROO leads to βα'-QOOH via internal abstraction of a weakly-bound α-H and a six-membered ring TS (-11.7 kcal/mol). This compound can ring-open by C−O bond scission or decompose to OH + 2,4-epoxy-THF over barriers of -3.3 and -2.8 kcal/mol, respectively. The second-lowest isomerization barrier (-8.7 kcal/mol) corresponds to a 5-membered ring TS to βα-QOOH. This species can produce OH + 2,3-epoxy-THF through a low-barrier TS (-21.1 kcal/mol), form an open-chain radical by C−O bond scission, or eliminate HO2 to produce 2,3-DHF. The highest-energy isomerization forms ββ'-QOOH, which can decompose to OH + 3,4-epoxy-THF or HO2 + 2,5-DHF. 9 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 29

All six of the QOOH isomers can in principle be stabilized by collisions and react with another oxygen molecule. Although a quantitative prediction of product distributions requires kinetic calculations, the PES in Fig. 1 suggest that the most favorable reaction pathways for ROO in the absence of second O2 addition are: α-ROO → 2,3-DHF + HO2 (R1) α-ROO → αα'-QOOH → butanedial + OH (R2) β-ROO → βα-QOOH → 2,3-epoxy-THF + OH (R3) β-ROO → βα-QOOH → 2,3-DHF + HO2 (R4) β-ROO → 2,5-DHF + HO2 (R5) At sufficiently high pressures and oxygen concentration, QOOH + O2 reactions should compete with unimolecular decomposition channels, and will most likely involve the dominant QOOH isomers: αα'-QOOH and βα-QOOH. Experimental product assignment A typical experimental dataset, taken at T = 600 K and P = 10 Torr, is represented in Fig. 2. The top panel is a time-resolved mass spectrum, integrated over VUV photon energies 8.2 – 11.2 eV, and the bottom is a mass-resolved PI spectrum, integrated over kinetic times t = 0 – 50 ms. Ion signals present before the photolysis pulse arrival at t = 0 were subtracted from both images. The two panels contain complementary information: the PI spectra required to identify the species give rise to each ion peak, and the time evolution of those peaks.

Figure 2. PIMS probing of Cl-initiated THF oxidation at T = 600 K and P = 10 Torr. The falsecolor scale indicates ion signal intensity (in arb. units) Panel A: time-resolved mass spectrum, integrated over photon energies 8.2-11.2 eV. Panel B: mass-resolved PI spectrum, integrated over t = 0 – 50 ms after photolysis.

Integration of our data over photon energy and kinetic time yields transient product mass spectra, shown in Fig. 3 for T = 600 K, P = 10 Torr and 1500 Torr. The negative peaks in the mass spectra correspond to depleted ion signals of the parent THF peak (m/z = 72), its daughter ion 10 ACS Paragon Plus Environment

Page 11 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(m/z = 71) and its 13C isotopolog (m/z=73). Additional minor depletion signals at m/z = 63 and 65 are due to the dissociative ionization (DI) of oxalyl chloride (CO35Cl+ and CO37Cl+). The most intense positive peaks in THF oxidation at P = 10 Torr occur at m/z = 43, 58, 70, and 86. At P = 1500 Torr several additional mass peaks emerge; the signals at m/z = 69, 85, 102, and 103 are highlighted because they represent key product species with strong pressure-dependence. For comparison, a product mass spectrum of 2,2,5,5-THF-d4 oxidation at P = 10 Torr is plotted in red in Fig. 3. The shift of the ion peaks to higher masses indicates the number of D atoms in each product and facilitates the identification of their molecular structure.

Figure 3. Mass spectra of THF (blue) and 2,2,5,5-THF-d4 (red) oxidation at T = 600 K, P = 10 Torr and 1500 Torr, integrated over photon energies 8.2-11.2 eV and kinetic times t = 0 – 50 ms. Depletion signals of the oxalyl chloride (COCl)2+ cations at m/z = 126, 128, and 130 are omitted for clarity.

The time evolution of the m/z = 43, 58, 70, 86 and 102 signals at T = 600 K and P = 10 Torr is shown in Fig. 4; the time traces of the m/z = 69, 85, and 103 peaks are similar to that of m/z = 102, and are omitted from Fig. 4 for clarity. The rapid rise of the m/z = 43, 58, and 70 signals establishes them as primary products of THF oxidation, and we assign these peaks to the coproducts of HO2 elimination (m/z = 70) and OH elimination (m/z = 43, 58). The m/z = 69, 85, 102, and 103 peaks rise more slowly, indicative of secondary reactions products, and we assign them to the products of QOOH + O2.We cannot definitively identify the m/z = 86 peak, but its long rise time suggests that it may be due to radical-radical reactions. These assignments are discussed in detail in the following sections. In addition to the main ion peaks, we quantified the following minor products of THF oxidation: ethene (m/z = 28), formaldehyde (m/z = 30), propene (m/z = 42), acrolein (m/z = 56), and furan (m/z = 68). In all, we observed 18 stable and transient compounds in the THF oxidation reaction; the mass peaks and their assignments are listed in Table 2. The yields of all products with known absolute PI cross-sections are listed as a function of T and P in Table 3 for THF-d0 experiments and in the Supplementary Information for THF-d4 oxidation. 11 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 29

Figure 4. Time evolution of products C2H3O (m/z = 43), C3H6O (m/z = 58), C4H6O (m/z = 70), C4H6O2 (m/z = 86) and C4H6O3 (m/z = 102) at T = 600 K and P = 10 Torr, integrated over photon energies 8.2-11.2 eV. The signals were scaled to match their intensities at t = 25 ms. Table 2. Chemical species identified in the oxidation of THF-d0 and THF-d4. The major products are divided qualitatively by their kinetic behavior into primary stable products (with rapid signal rise), secondary stable products (with slow rise), and transient intermediates (with rapid rise followed by signal decay). See further text for species naming convention. Species

m/z

THF oxidation formula

2,2,5,5,-d4-THF oxidation m/z formula

Primary products Butanedial 2,3-dihydrofuran 2,5-dihydrofuran Transient intermediates α-THF-yl α-THF-yl peroxy β-THF-yl β-THF-yl peroxy αα'-OOQOOH Secondary products GBL-OOH Unassigned 2,3-DHF-αOOH 2,3-DHF-βOOH Minor products Ethene Formaldehyde Propene Acrolein Furan γ-butyrolactone THF-3-one a

43a 58a 70 70

C2H3O C3H6O C4H6O C4H6O

44a 60a 73 74

C2H3DO C3H4D2O C4H3D3O C4H2D4O

74 74a 75 75a

C4H4D3O C4H4D3O3 C4H3D4O C4H3D4O3

103

C4H7O3

85a 86 69a 102 69a 102

C4H5O2 C4H6O2 C4H5O C4H6O3 C4H5O C4H6O3

86a 89 71a 105 72a 106

C4H4DO2 C4H3D3O2 C4H3D2O C4H3D3O3 C4H2D3O C4H2D4O3

28 30 42 56 68 86 86

C2H4 CH2O C3H6 C3H4O C4H4O C4H6O2 C4H6O2

28,30 30,32 45 57,58 70 88 90

C2H4, C2H2D2 CH2O, CD2O C3H3D3 C3H3DO, C3H2D2O C4H2D2O C4H4D2O2 C4H2D4O2

DI products, rather than parent ions


Page 13 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Table 3. Quantified reaction products and the overall C atom yield (relative to THF consumed), determined by averaging over kinetic times 20 < t < 30ms at 10 Torr and 30 < t < 80ms at 1500 Torr. Uncertainties are ±25% of the stated value.

550 K

10 Torr 600 K 650 K

700 K

ethene formaldehyde acrolein furan 2,3-DHF 2,5-DHF propene

0.02 0.06 0.01 0.002 0.17 0.01 0.01

0.03 0.06 0.01 0.004 0.20 0.01 0.02

0.05 0.08 0.015 0.012 0.23 0.05

C atom yield

0.21

0.25

0.34

550 K

1500 Torr 600 K

650 K

0.10 0.10 0.02 0.014 0.22 0.03 0.14

0.07 0.02 0.03 0.02 -

0.05 0.02 0.07 0.02 -

0.03 0.04 0.02 0.09 0.005 0.02

0.46

0.08

0.12

0.15

R and ROO formation The first step in low-temperature THF oxidation is H atom abstraction from either the α or β position. Our experiments using 2,2,5,5-THF-d4 enable the separation of α-R and β-R radicals by mass (m/z = 74 and 75) and reveal that Cl atom attack produces mainly α-R at our experimental conditions. Figure S7 in the Supporting Information shows the relevant parts of the transient mass spectra at P = 10 Torr and T = 600 and 650 K. These spectra were averaged over ionization energies 8.2 – 8.5 eV in order to minimize possible contributions from DI of α- and β-ROO (calculated VIE of ~10.2 and 10.3 eV, respectively). The ratio of (m/z = 74)/(m/z = 75) signal intensities is ~ 10:1 at 650 K and 26:1 at 600K,which implies a corresponding dominance of α-H atom abstraction, assuming similar ionization cross-sections for the two THF-yl radicals. The kinetic isotope effect is likely to favor the formation of β-R in THF-d4 oxidation, since it involves abstraction of H atoms, rather than D. Therefore, the observed preference for α-R is likely to be even greater in THF-d0. The calculated α-C−H bond dissociation energy in THF is 4.5 kcal/mol lower than that of β-C−H (93.7 vs. 98.2 kcal/mol).19 At T = 550 – 650 K Simmie calculated19 the H-abstraction from the α site to be 5.5-7.5 times faster than from the β site in the case of CH3 attack, but only ~5-6% faster for H atom attack. The abstraction of α-H by OH radicals can be estimated from structureactivity relationships (SAR) of Kwok et al.43 as ~2.8 times faster than for β-H at 550K. No measurements or SAR estimates exist for site-specific H abstraction by Cl atoms in THF or other cyclic ethers. However, Alwe et al.44 reported similar total rate coefficients for Cl + THF and Cl + tetrahydropyran (THP, the analogous 6-member ring cyclic ether), whereas Giri et al.45 found that Cl reacts ~ 30% faster with THF than with THP, despite THP having an extra CH2 group. These results are contrary to the trend in linear hydrocarbons, where longer chains lead to higher Cl reaction rate coefficients,46 and are consistent with H-abstraction mainly at the α-carbon of ethers. Ballesteros et al.47 calculated the activation energy for α-H abstraction from THP by Cl to 13 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 29

be ~7 kcal/mol less than for the β- or γ- sites. Giri et al.48 later suggested that reactions of Cl with ethers in general involve a pre-reactive complex between the Cl and O atoms (bound by ~5 kcal/mol in the case of Cl + 1,4-dioxane), which further decreases the abstraction barrier from the α site. Formation of a similar complex in the Cl + THF reaction may contribute to the observed preference for α-H abstraction in our experiments. HO2 elimination pathways The ion peak at m/z = 70 has the composition C4H6O and corresponds to the coproducts of HO2 elimination from ROO, a typical major pathway in low-T hydrocarbon oxidation. The analysis of its PI spectrum (Fig. 5A) shows that this peak can be fit to a weighted sum of absolute reference spectra of 2,3- and 2,5-DHF. The fit coefficients are the contributions from each DHF isomer to the PI spectrum; when compared to the spectrum of the parent THF peak, they directly translate to concentrations of 2,3- and 2,5-DHF. The 2,3-DHF yield, relative to THF consumed, is 17 – 23% at 10 Torr and 3 – 9% at 1500 Torr, whereas 2,5-DHF yield is 0 – 2% of THF depletion at all pressures (Table 3).

Figure 5. PI spectra in THF oxidation (open circles) at T = 600 K, P = 10 Torr, integrated over 0 – 50 ms. Panel A: m/z = 70 (C4H6O+) peak. The red line is a fit to a weighted sum of 2,3- and 2,5DHF reference spectra, shown as black dashed lines (see text). Panel B: m/z = 58 (C3H6O+) and 43 (C2H3O+) peaks. The solid lines are reference spectra of the parent (m/z = 86, C4H6O2+) and daughter ions of butanedial, scaled for comparison.

The product branching of the HO2 elimination channels is in line with expectations, based on the PES in Fig. 1. Decomposition of α-ROO produces only 2,3-DHF. Decomposition of β-ROO can 14 ACS Paragon Plus Environment

Page 15 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


in principle form both DHF isomers. The lowest-energy channel for 2,3-DHF production is from βα-QOOH, whereas for 2,5-DHF it is directly from β-ROO. Without chemical kinetics calculations it is not possible to predict the branching ratio into the 2,3- and 2,5-DHF product channels on the β-R + O2 PES; however, the main isomer detected in our experiment is 2,3-DHF due to the initial dominance of α-R over β-R. In partially deuterated THF experiments the m/z = 70 channel splits in two peaks, m/z = 73 and 74, corresponding to 2,3-DHF-d3 and 2,5-DHF-d4, respectively. The yield of 2,5-DHF-d4 increases to 4-6% at 10 Torr and 2-4% at 1500 Torr as a consequence of the kinetic isotope effect. H atoms at the β site of 2,2,5,5,-THF-d4 favor the production of β-R and their subsequent conversion to 2,5-DHF-d4, when compared to THF-d0. In analogy to alkanes,35 HO2 elimination has some of the lowest-energy barriers in THF oxidation and likely involves a combination of formally direct production (i.e. well-skipping, R6) and thermal decomposition of ROO (R7): R + O2 → DHF + HO2 R + O2 → ROO → DHF + HO2

(R6) (R7)

Qualitatively, the decrease of HO2 elimination yield at 1500 Torr relative to 10 Torr is most likely due to the collisional quenching of formally direct production from α-R + O2. Production of DHF via αβ-QOOH is probably not important at our conditions, because it involves two potential barriers with higher energies than direct formation from ROO. OH elimination The calculated PES for THF-yl + O2 show eight OH elimination pathways to six distinct coproducts with formula C4H6O2 and molecular mass of 86, with butanedial being the main expected co-product based on computed barrier heights. However, the m/z = 86 ion peak in our experiments is mostly due to secondary reaction products, as we show below. Instead, we assign our largest peaks at m/z = 58 and 43 to butanedial. Figure 5B shows the PI spectra of the m/z = 58 and 43 ions at P = 10 Torr and T = 600 K, along with the reference spectrum of butanedial at the same temperature, scaled for visual comparison. Photoionization of butanedial is dominated by dissociative ionization at m/z = 58 (C3H6O+) and 43 (C2H3O+), with almost no signal at the parent mass, m/z = 86 (C4H6O2+). The DI peak at m/z = 58 comes from CO loss following intramolecular H transfer in the parent cation, whereas the m/z = 43 peak is likely due to neutral acetyl elimination. The CBS-QB3 calculated appearance energies of the parent and daughter ions of butanedial agree with the measured signal onsets: C4H6O2+, calculated IE = 9.63 eV, signal onset ~9.6 eV; C3H6O+, calculated AE = 9.76 eV, onset ~9.8 eV; C2H3O+, calculated AE = 9.77 eV, onset ~9.9 eV (see SI). Butanedial is only available commercially in a protected diacetal form, which requires acidcatalyzed hydrolysis to remove the protecting groups. The yield of pure butanedial in this reaction is not known, and therefore we could not determine its absolute ionization cross-section. However, even though we cannot quantify its production in THF oxidation, Fig. 5B demonstrates that our largest observed ion peaks at m/z = 58 and 43 are due almost entirely to butanedial. A 15 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 29

portion of the m/z = 58 spectrum above 10.5 eV differs slightly from the reference spectrum of butanedial. The source of this discrepancy is not direct ionization of any C3H6O species, all of which have ionization energies35, 49-51 below 10.5 eV. The ion peak at m/z = 43 appears to have a similar deviation, which may arise from imperfect normalization to the VUV photon flux. The m/z = 86 (C4H6O2) peak corresponds to the same sum formula as the co-products of OH elimination, yet its slow rise suggests that it is due to secondary reactions. When using THF-d4, this peak splits into three: m/z = 88, 89, and 90. The mass of the largest peak, m/z = 89, is consistent with 2,3- or 2,4-epoxy-THF-d3. However, neither of these species are likely sources of this peak: the calculated barriers to 2,4-epoxy-THF + OH are relatively high, and the only lowenergy pathway to 2,3-epoxy-THF + OH is via β-ROO, a minor channel in our experiments. Furthermore, the calculated IEs of 2,3- and 2,4-epoxy-THF (8.36 and 10.01 eV, respectively) do not match the observed signal onset of ~9.1 eV. As a result, we cannot definitively identify this peak but because of its time behavior it does not appear to be a primary product at our conditions. On the other hand, the two minor peaks at m/z = 88 and 90 in THF-d4 oxidation have short rise times. The PI spectrum of the m/z = 88 peak is well reproduced by a sum of the reference spectra of the two possible isomers at this mass: butanedial-d2 and γ-butyrolactone-d2. The m/z = 90 peak also has two possible sources: tetrahydrofuran-3-one-d4 and 3,4-epoxy-THF-d4, but its PI spectrum can be fully fit with the reference spectrum of THF-3-one alone. The rapid rise of the carbonyl species GBL and THF-3-one is unexpected, since our calculations do not show any low-energy pathways to their formation from ROO. These channels are analogous to the prompt production of acetone and propanal in propane oxidation,35 despite similarly high calculated barriers from propyl peroxy. Further computational studies are needed to find the source of carbonyl compounds, evidently formed through primary low-T autoignition reactions. To summarize, we find that OH elimination in THF oxidation at T < 650 K produces almost exclusively butanedial, although we cannot quantify its yield without the knowledge of its absolute ionization cross-section. According to the PES in Fig. 1, its source is αα'-QOOH via reaction R2, the only low-energy OH elimination pathway from α-ROO. In THF-d4 experiments the main DI peak of butanedial is at m/z = 60 (C3H4D2O+ + CO), confirming its formation from αα'-QOOH. In the case of β-ROO, the most likely OH elimination channel (to 2,3-epoxy-THF via reaction R3) may be competitive with HO2 elimination, based on comparable barrier heights. However, on the whole, butanedial is the main observed OH co-product due to the overwhelming preference for α-ROO in our experiments. Our findings are consistent with previous work on THF combustion. Vanhove et al. recently detected butanedial in a rapid compression machine,32 and Molera et al. found a 66% yield of succinic acid in a static reactor at T = 493K, P = 156 Torr.31 Succinic acid is an oxidation product of butanedial (also called succinic aldehyde), and the results of Molera et al. imply that butanedial is the main product of low-T THF oxidation. In our experiments the sum of all quantified products accounts for ~20 – 50% of the overall C atoms, consumed by THF oxidation 16 ACS Paragon Plus Environment

Page 17 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


at 10 Torr, and for 8 – 15% at 1500 Torr, (see Table 3). Assuming that butanedial is the main product that we did not quantify, its yield can be up to 50 – 80% at 10 Torr and up to ~90% at 1500 Torr, depending on T. Our measurements highlight a key difference in the low-T reactivity of ethers as compared with simple hydrocarbons: the presence of an oxygen atom in the molecular structure facilitates the breaking of C–O bonds and results in fragmentation channels that are not available in alkanes. As an example, n-butane oxidation at T ~ 550 – 700 K is dominated by HO2 elimination,13 and the OH + cyclic ether channels account for no more than ~1/3 of the total product distribution; in the case of propane35 oxidation, OH elimination yield does not exceed ~4% at T < 670 K. The barriers leading to cyclic ether formation in those compounds lie between +5 and -5 kcal/mol relative to the reactants. In contrast, the analogous OH elimination from αα'-QOOH in THF oxidation proceeds by breaking a C–O bond in the ring instead of forming a strained bicyclic compound, 2,5-epoxy-THF. The resulting barrier energy is -10.1 kcal/mol, which is the likely reason for butanedial + OH being the dominant reaction pathway. In linear ethers such as dimethyl ether, similar C–O bond scission produces OH and two CH2O fragments, analogous to ring-opening in cyclic ethers, and accounts for nearly all of the reaction products.52 Second O2 addition We observe four mass peaks in THF oxidation (at m/z = 69, 85, 102, and 103) that exhibit remarkably strong P-dependence, and we assign these peaks to the products of second O2 addition to αα'-QOOH, which we expect to be the most abundant QOOH isomer in our experiments. We support our assignments using THF-d4 oxidation experiments, calculated IE values, and additional calculations of the stationary points in the lowest-energy reaction pathways of αα'-QOOH + O2, as shown in Fig. 6.

Figure 6. Calculated PES for αα'-QOOH + O2, showing the reactions of cis- and transOOQOOH. The ZPE-corrected energies at 0 K, relative to the reactants, are calculated using the CBS-QB3 method. The pathways leading to products detected in our experiments are highlighted in bold.

The second O2 addition creates two isomers with similar energies, cis- and trans-αα'-OOQOOH. The trans- isomer can produce HO2 + 2,3-dihydrofuranyl-α-hydroperoxide (2,3-DHF-αOOH) or undergo intramolecular α-H atom transfer (cis to the peroxy group) to ultimately yield OH + γ17 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 29

butyrolactone hydroperoxide (GBL-OOH). The latter pathway is analogous to KHP formation in alkane oxidation. The calculated barriers to these products are -4.1 and -4.5 kcal/mol, respectively, relative to QOOH + O2. On the other hand, the only energetically accessible pathway for cis-OOQOOH is to HO2 + 2,3-DHF-αOOH (-4.2 kcal/mol barrier). GBL-OOH is not produced by cis-OOQOOH due to the strain of intramolecular trans-α-H transfer, which is a consequence of the rigid ring structure of THF. The sum formula of the m/z = 103 peak (C4H7O3) corresponds to ROO or QOOH in THF oxidation, yet it is not likely to arise from direct ionization of either of these species. Our calculations show that neither α- nor β-THF-yl peroxy have stable parent cations and instead dissociate to R+ + O2, like most organic ROO species.53 Meanwhile, QOOH are typically very reactive and hence elusive; in fact, the only reported observation is of a doubly-resonance stabilized QOOH,11 whereas QOOH in THF oxidation would not be resonance stabilized. On the other hand, OOQOOH are peroxy radicals and should therefore be similar to ROO. They are relatively stable and unreactive, and may accumulate in substantial concentration in our experiments. As with ROO, our calculations indicate that the parent cation of αα'-OOQOOH is unbound. The appearance energy for the O2 + 1C4H7O3+ DI channel is 7.89 eV, but the VIE of αα'-OOQOOH+ is ~10.4 eV. The large difference of these values suggests a gradually increasing PI cross-section, consistent with the observed m/z = 103 spectrum. Although we do not probe this species in THF-d4 experiments because of interference from other products, we tentatively assign the m/z = 103 peak to DI of αα'-OOQOOH. Our results are similar to the only other direct observation of OOQOOH in tert-butyl hydroperoxide oxidation,10 where OOQOOH was also detected solely via the QOOH+ + O2 DI signal. We detect no ion signal at the mass of the parent cation of GBL-OOH (C4H6O4, m/z = 118). However, the m/z = 85 peak (C4H5O2), which shifts to m/z = 86 in THF-d4 experiments, is consistent with dissociative ionization of this product. Again, our assignment is supported by calculations, which indicate that the parent GBL-OOH cation is unbound. The calculated appearance energy of the lowest-energy DI channel, 1C4H5O2+ + HO2, is 10.26 eV, and the VIE of GBL-OOH is 10.64 eV. Both of these values are in good agreement with the observed PI spectrum of the m/z = 85 peak, shown in Fig 7A. The remaining ion peaks at m/z = 102 (C4H6O3) and 69 (C4H5O) have identical time evolution and we assign them to the parent cation of dihydrofuranyl hydroperoxide (DHF-OOH) and a daughter ion formed by neutral HO2 loss, a common DI pathway in hydroperoxides (see Scheme 2). In THF-d4 experiments these peaks shift primarily to m/z = 105 and 71 with a minor contribution at m/z = 106 and 72. Six isotopomers of DHF-OOH can potentially form in the oxidation of THF-d4, as shown in Scheme 2, but the main peaks at m/z = 105 and 71 can only arise from one of them: 2,3-DHF-αOOD (Scheme 2.D). The calculated IE of 2,3-DHF-αOOD is 8.57 eV, and the appearance energy of the C4H5O+ DI fragment is 9.14 eV. These values agree well with the ionization onsets of the m/z = 102 and 69 peaks (see Fig. 7B), further confirming our conclusion. 18 ACS Paragon Plus Environment

Page 19 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 7. PI spectra of second-O2 addition products in THF oxidation at T = 600 K, P = 10 Torr, integrated over 0 – 50 ms. Panel A: m/z = 85 (C4H5O2+) peak. The vertical arrow indicates the calculated appearance energy of the C4H5O2+ DI product of GBL-OOH. Panel B: m/z = 69 (C4H5O+) and 102 (C4H6O3+) peaks. Vertical arrows mark the calculated appearance energy of the parent and daughter ions of 2,3-DHF-αOOH.

Scheme 2. The parent cations and DI products of all possible DHF-OOH isotopomers in THF-d4 oxidation. The main species detected in our experiments (2,3-DHF-αOOD) is highlighted in red. 19 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 29

Hydroperoxides are typically not included as primary products in low-T autoignition mechanisms, but rather as secondary products of radical-radical chemistry: R + HO2 (+M) → ROOH (+M) ROO + HO2 → ROOH + O2

(R8) (R9)

In principle, DHF-OOH could form in THF oxidation following H abstraction from DHF (the HO2-elimination co-product of THF): X + DHF → HX +DHF-H DHF-H + HO2 (+M) → DHF-OOH (+M)

(R10a) (R10b)

where X is a radical such as OH, Cl, H, or O atom. However, in 2,2,5,5,-THF-d4 oxidation the main route to hydroperoxyl is α-ROO → 2,3-DHF-d3 + HO2, meaning that the HO2 isotopolog should dominate in the reacting mixture. Therefore, secondary reactions such as R10 cannot account for 2,3-DHF-αOOD in our experiments, since its hydroperoxide group contains a D atom. The most likely alternative source of 2,3-DHF-αOOD is HO2 elimination following second O2 addition to αα'-QOOH. The unambiguous structure determination of this isomer, enabled by THF-d4 experiments, also reveals the mechanism for its production, shown in Scheme 3, which involves the calculated lowest-energy isomerization of α-ROO to αα'-QOOH and subsequent HO2 elimination from αα'-OOQOOH. The minor peak at m/z = 106 must come from 2,3-DHF-βOOD (Scheme 2.F), since the other mass 106 isomer (2,5-DHF-αOOD) cannot be formed from THF-d4. Production of 2,3-DHFβOOD starts with the isomerization of β-ROO to βα'-QOOH, which has the lowest entrance barrier and the highest exit barriers of all the QOOH isomers on the β-R + O2 surface. Consequently, βα'-QOOH is likely to be the most abundant QOOH species formed by β-ROO, and the most likely to react with a second O2 molecule rather than decompose. Although it is favored by the kinetic isotope effect in THF-d4, it is a minor peak in our experiments, so the yield of the corresponding isotopolog in THF-d0 oxidation is likely negligible.

Scheme 3. Mechanism for the production of 2,3-DHF-αOOD in THF-d4 oxidation

In summary, we simultaneously observe three key species in the second-O2 addition reaction of QOOH in THF oxidation – the OOQOOH radical and its decomposition products, GBL-OOH and 2,3-DHF-αOOH: αα'-QOOH + O2 → αα'-OOQOOH → GBL-OOH + OH αα'-QOOH + O2 → αα'-OOQOOH → 2,3-DHF-αOOH + HO2

(R11) (R12)

R12 collectively refers to DHF-OOH formation of from both cis- and trans-αα'-OOQOOH, 20 ACS Paragon Plus Environment

Page 21 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


whereas R11 proceeds exclusively through trans-αα'-OOQOOH. The competition between R11 and R12 is likely to influence the degree of radical chain-branching vs. chain-propagation behavior in THF oxidation. As with other KHP pathways, R11 should ultimately form 2 OH + oxy-radical, whereas R12 will produce at most a single OH through thermal dissociation of 2,3DHF-αOOH. Quantum chemical calculations of the PES in Fig. 6 reveal that although R11 and R12 have similar energy barriers from αα'-OOQOOH, the TS in R12 are somewhat more loose than in R11. As a result, preliminary calculations of the branching between R11 and R12 yield a ratio DHF-OOH/GBL-OOH ~4:1 at 400 K and ~10:1 at 700 K (see SI). Quantification of these products is not feasible in the current study, but will be a subject of a future manuscript, along with a more detailed theoretical and modeling treatment. THF-yl ring-opening pathways In the THF-yl pyrolysis mechanism by Simmie19, the fastest ring-opening channel of THF-yl radicals at T = 550 – 650 K is the C–O bond scission of α-R with a rate coefficient of ~2.5·105 s-1 at 650 K, according to the model of Tran et al.20 Other pathways, including C–C bond fission, are several orders of magnitude slower. The resulting 4-oxobutyl radical is predicted to either decompose to ethene + vinoxy or to isomerize and then decompose very rapidly to CO + propyl, which will in turn be efficiently oxidized to propene35 at our conditions. Calculations predict that CO + propyl production is favored over ethene + vinoxy by a factor of 2–6 at 650 K19-20 and by even larger margins at lower temperatures. Lastly, oxidation of 4-oxobutyl may compete with unimolecular dissociation, but the products of this reaction are not known. Indeed, we observed small amounts of propene and ethene (up to ~5% of consumed THF at 650 K); ethene was also present at lower T, but in smaller yields. In THF-d4 experiments propene was formed solely as CH2CHCD3 and ethene partially as CH2CD2, both of which are consistent with being products of 4-oxobutyl decomposition. To further explore the influence of ring-opening pathways in our experiments, we acquired additional data at P = 10 Torr and T = 700 K, varying the O2 concentration to examine the competition between bond scission and oxidation reactions. At 700 K we found that when we lowered [O2] from 1·1017 to 1·1015 cm-3, propene yield grew by a factor of 3, likely as a result of less competition from ROO formation. Oxygen-dependence of ethene yields was not quantified because the photon energy used in those experiments (10.2 eV) is below its IE. A full consideration of the effects of THF-yl β-scission pathways, including 4-oxobutyl + O2 and subsequent reactions, is beyond the scope of the present paper. Nonetheless, our results suggest that propene and (to a lesser extent) ethene detection is a good indicator of active ring-opening channels in our experiments, although not sufficient to quantify the total ring-opening yield. The near absence of propene signal at T = 550 – 600 K shows that ring-opening of THF-yl is not important at those temperatures, whereas it evidently starts to contribute at 650 K. We also observed small yields of formaldehyde (6 –8%) at T ≤ 650 K. In principle, it could arise from decomposition of β-R to allyl + CH2O. However, the predicted rate coefficient20 for this 21 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 29

reaction reaches only ~120 s-1 at 650 K, which is much slower than the first-order rate coefficient of O2 addition to THF-yl and subsequent decomposition or isomerization reactions. In THF-d4 experiments we observe d0 as well as d2 isotopologs of both ethene and formaldehyde, whereas only d2 species are predicted by the high-T decomposition mechanism of Simmie. Therefore, it is likely that at our conditions formaldehyde and ethene are also produced by some other pathways, perhaps related to the decomposition of QOOH or by secondary chemistry. TEMPERATURE- AND PRESSURE-DEPENDENCE OF THF AUTOIGNITION A simplified kinetic mechanism for our THF oxidation experiments at T ≤ 650 K is shown in Fig. 8. The initial Cl attack generates mainly α-R and a minor amount of β-R, which rapidly react with O2 to produce α- and β-ROO. A fraction of α-ROO (5 – 25%, depending on the conditions) undergoes HO2 elimination to form 2,3-DHF. The rest primarily isomerize to αα'QOOH, which decompose to OH + butanedial or react further with O2, producing αα'OOQOOH. Stabilized αα'-OOQOOH decompose to either 2,3-DHF-αOOH + HO2 or to GBLOOH + OH. In the case of β-ROO, we observe 2,5-DHF, which indicates that HO2 elimination pathways are active. Based on the calculated barrier heights, we assume that 2,3-DHF is also formed via the reaction R4, β-ROO → βα-QOOH → 2,3-DHF + HO2. However, our experiments do not probe this product channel separately from 2,3-DHF formed by α-ROO; we indicate this by a dashed arrow in Fig. 8. The only product that can be definitively attributed to isomerization into QOOH on the β-R + O2 PES is a minor amount of 2,3-DHF-βOOH in THF-d4 oxidation.

Figure 8. Primary THF oxidation sub-mechanism under our experimental conditions: T = 550 – 650 K, P = 10 – 2000 Torr. The arrow width qualitatively represents the relative importance of each reaction channel. The competition among the main reaction pathways (HO2 + DHF formation, unimolecular QOOH decomposition, second O2 addition) depends on T, P, and [O2]. At T > 650 K or at very low [O2], THF-yl β-scission channels may also be important.

The mechanism summarized above qualitatively explains the observed dependence of THF 22 ACS Paragon Plus Environment

Page 23 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


depletion and of the main product signals on experimental conditions, as shown in Fig. 9. In all cases we maintained constant [Cl]0. Panel 9A shows the [O2]-dependence at P = 1500 Torr and T = 575 K. Even at the lowest value of [O2] = 0.6·1018 cm-3 ROO formation completely dominates competing reactions, and increasing [O2] further does not alter the ratio between DHF + HO2 and isomerization to QOOH. As a result, the m/z = 70 ion signals are relatively independent of O2. However, [O2] does affect the competition between QOOH decomposition (to butanedial + OH) and second O2 addition (to form OOQOOH). Consequently, butanedial signal at m/z = 58 decreases by a factor of 2, whereas the signals from OOQOOH and its decomposition products DHF-OOH and GBL-OOH grow with increasing [O2]. Notably, THF depletion is nearly constant at ~3% regardless of [O2], which suggests that a decrease in butanedial + OH is offset by eventual OH release from other sources. Two such likely OH source are formation of GBL-OOH + OH and the thermal decomposition of GBL-OOH and DHF-OOH themselves.

Figure 9. Dependence of THF consumption and product ion signals on experimental conditions. The m/z = 85, 58, 70, 102, 103 intensities are integrated over t = 0 – 80 ms and arbitrarily scaled for comparison. Panel A: [O2]-dependence at P = 1500 Torr, T = 575 K; Panel B: P-dependence at T = 575 K, [O2] = 1.2·1018 cm-3; Panel C: T-dependence at P = 1500 Torr, [O2] = 3·1018cm-3. 23 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 29

THF depletion (%) is shown in black. Ion signals are normalized to [Cl]0 (using Oxalyl Chloride ion depletion) in panel A and to THF consumption (using THF ion depletion) in B and C.

Panel 9B shows that THF depletion increases steadily as a function of P from ~2% at 340 Torr to 6% at 1950 Torr (at T = 575 K and [O2] = 1.2·1018 cm-3) due to radical recycling via OH elimination. The ion signals in Fig. 9B are normalized to total THF consumed rather than to [Cl]0 in order to emphasize the relative branching ratio of ROO decomposition and oxidation. HO2 elimination co-products (m/z = 70) decrease in intensity by ~30% as pressure is increased, probably due to collisional quenching of formally direct processes (R6). Butanedial yield per ROO stays constant, consistent with the R + O2 reaction being in its high-pressure limit even at P = 340 Torr, similar to alkane oxidation.54 On the other hand, m/z = 85, 102 and 103 signals grow by a factor of 10, a remarkable increase that most likely reflects the important role of collisional stabilization of OOQOOH, also as predicted in alkane combustion.55 Finally, panel 9C shows the T-dependence of THF depletion and of the product yields, normalized to THF consumption, at P = 1500 Torr and [O2] = 3·1018cm-3. At T ≤ 450 K the measured THF depletion is ~4.5%, consistent with the estimated ratio of [Cl]0 relative to THF, 1:20. The product signals are small, implying that nearly all ROO are stabilized and do not react on the timescale of the experiment. At higher T, however, THF depletion increases to ~12%, indicating significant OH recycling. The yield of m/z = 70 (HO2 loss) and 58 (OH loss) ions rises by a factor of 10 and 20, respectively, at T = 650 K. The intensities of DHF-OOH and GBLOOH signals initially grow even more steeply, peak at T = 575 K, and decrease at higher T due to the onset of thermal decomposition. The OOQOOH signal at m/z = 103 also grows very steeply, but peaks at somewhat lower temperatures of ~525 – 550 K, which is consistent with DHF-OOH and GBL-OOH being a product of OOQOOH decomposition. CONCLUSION We probed the initial reactions in THF oxidation by time-resolved PIMS in a constant T and P chemical reactor. The major products of this reaction are HO2 + 2,3-DHF and OH + butanedial. In addition, we observed the competition between two reaction pathways following second O2 addition to αα'-QOOH: the formation of GBL-OOH + OH and a previously unexplored production of 2,3-DHF-OOH + HO2. The exact molecular structure of all products was determined using reference photoionization spectra, THF-d4 oxidation experiments, and quantum chemical calculations. With the aid of calculated PES, we propose a sub-mechanism for lowtemperature THF oxidation, which involves mainly the reactions of two critical intermediates: αROO and αα'-QOOH. It is commonly accepted that fuel ignition behavior depends on the competition between radical chain termination or propagation (e.g. unimolecular decomposition of QOOH) and chainbranching reactions like QOOH oxidation, also called the “second O2 addition”. In particular, OH + ketohydroperoxide formation from QOOH + O2 and subsequent KHP decomposition was recently recognized as a key chain-branching step, because it ultimately produces two OH 24 ACS Paragon Plus Environment

Page 25 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


radicals and an unstable oxy-radical (OP=O) for each initial ROO. In the case of THF oxidation, however, we show that the chain-branching production of OH + GBL-OOH (which is analogous to KHP) faces competition from HO2 + DHF-OOH at temperatures at least up to 700 K. Our results suggest that we expand the current QOOH oxidation models to include HO2 elimination and perhaps other reaction pathways in search for a complete picture of low-T autoignition. SUPPORTING INFORMATION. TST calculations of the 2,3-DHF-αOOH/GBL-OOH branching ratio, reference PI spectra measured in this work, product yields in THF-d4 experiments, mass spectra of α-R and β-R in THF-d4 experiments, calculated IE values, calculated geometries of stationary points on THF-yl + O2 PES, experimental time traces at all conditions. ACKNOWLEDGEMENTS We thank Mr. Kendrew Au (Sandia) and the staff at the Chemical Dynamics Beamline at the ALS for excellent technical support of these experiments. This work was supported by the Division of Chemical Sciences, Geosciences, and Biosciences, the Office of Basic Energy Sciences, the U.S. Department of Energy. LS and IO and supported through the Argonne-Sandia Consortium on High-Pressure Combustion Chemistry. Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the National Nuclear Security Administration, under contract DE-AC04-94AL85000. This research used resources of the Advanced Light Source, which is a DOE Office of Science User Facility.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 29

References 1. Geilen, F.; Engendahl, B.; Harwardt, A.; Marquardt, W.; Klankermayer, J.; Leitner, W. Selective and flexible transformation of biomass-derived platform chemicals by a multifunctional catalytic system. Angew. Chem., Int. Ed. 2010, 49, 5510 - 5514. 2. Mettler, M. S.; Paulsen, A. D.; Vlachos, D. G.; Dauenhauer, P. J. Pyrolytic conversion of cellulose to fuels: levoglucosan deoxygenation via elimination and cyclization within molten biomass. Energy & Environmental Science 2012, 5, 7864-7868. 3. Roman-Leshkov, Y.; Barrett, C. J.; Liu, Z. Y.; Dumesic, J. A. Production of dimethylfuran for liquid fuels from biomass-derived carbohydrates. Nature 2007, 447, 982-U5. 4. Yang, W.; Sen, A. One-step catalytic transformation of carbohydrates and cellulosic biomass to 2,5-dimethyltetrahydrofuran for liquid fuels. ChemSusChem 2010, 3, 597-603. 5. Thewes, M.; Muether, M.; Pischinger, S.; Budde, M.; Brunn, A.; Sehr, A.; Adomeit, P.; Kankermayer, J. Analysis of the impact of 2-methylfuran on mixture formation and combustion in a direct-injection spark-ignition engine. Energy Fuels 2011, 25, 5549-5561. 6. Westbrook, C. K. Biofuels Combustion. Annu. Rev. Phys. Chem. 2013, 64, 201-219. 7. Herbinet, O.; Bax, S.; Glaude, P.-A.; Carré, V.; Battin-Leclerc, F. Mass spectra of cyclic ethers formed in the low-temperature oxidation of a series of n-alkanes. Fuel 2011, 90, 528-535. 8. Battin-Leclerc, F.; Blurock, E.; Bounaceur, R.; Fournet, R.; Glaude, P.-A.; Herbinet, O.; Sirjean, B.; Warth, V. Towards cleaner combustion engines through groundbreaking detailed chemical kinetic models. Chem. Soc. Rev. 2011, 40, 4762-4782. 9. Zádor, J.; Taatjes, C. A.; Fernandes, R. X. Kinetics of elementary reactions in low-temperature autoignition chemistry. Prog. Ener. Comb. Sci. 2011, 37, 371-421. 10. Zador, J.; Huang, H. F.; Welz, O.; Zetterberg, J.; Osborn, D. L.; Taatjes, C. A. Directly measuring reaction kinetics of QOOH - a crucial but elusive intermediate in hydrocarbon autoignition. Phys. Chem. Chem. Phys. 2013, 15, 10753-10760. 11. Savee, J. D.; Papajak, E.; Rotavera, B.; Huang, H. F.; Eskola, A. J.; Welz, O.; Sheps, L.; Taatjes, C. A.; Zador, J.; Osborn, D. L. Direct observation and kinetics of a hydroperoxyalkyl radical (QOOH). Science 2015, 347, 643-646. 12. Moshammer, K.; Jasper, A. W.; Popolan-Vaida, D. M.; Lucassen, A.; Diévart, P.; Selim, H.; Eskola, A. J.; Taatjes, C. A.; Leone, S. R.; Sarathy, S. M. Detection and identification of the ketohydroperoxide (HOOCH2OCHO) and other intermediates during low-temperature oxidation of dimethyl ether. J. Phys. Chem. A 2015, 119, 7361-7374. 13. Eskola, A. J.; Welz, O.; Savee, J. D.; Osborn, D. L.; Taatjes, C. A. 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. J. Phys. Chem. A 2013, 117, 12216-12235. 14. Eskola, A. J.; Welz, O.; Zador, J.; Antonov, I. O.; Sheps, L.; Savee, J. D.; Osborn, D. L.; Taatjes, C. A. Probing the low-temperature chain-branching mechanism of n-butane autoignition chemistry via time-resolved measurements of ketohydroperoxide formation in photolytically initiated n-C4H10 oxidation. Proc. Comb. Inst. 2015, 35, 291-298. 15. Battin-Leclerc, F.; Herbinet, O.; Glaude, P. A.; Fournet, R.; Zhou, Z.; Deng, L.; Guo, H.; Xie, M.; Qi, F. Experimental confirmation of the low-temperature oxidation scheme of alkanes. Angew. Chem., Int. Ed. 2010, 49, 3169-3172. 16. Battin-Leclerc, F.; Herbinet, O.; Glaude, P.-A.; Fournet, R.; Zhou, Z.; Deng, L.; Guo, H.; Xie, M.; Qi, F. New experimental evidences about the formation and consumption of ketohydroperoxides. Proc. Comb. Inst. 2011, 33, 325-331. 17. Wang, Z.; Zhang, L.; Moshammer, K.; Popolan-Vaida, D. M.; Shankar, V. S. B.; Lucassen, A.; Hemken, C.; Taatjes, C. A.; Leone, S. R.; Kohse-Höinghaus, K. Additional chain-branching pathways in the low-temperature oxidation of branched alkanes. Combust. Flame 2015. 26 ACS Paragon Plus Environment

Page 27 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


18. Crounse, J. D.; Nielsen, L. B.; Jørgensen, S.; Kjaergaard, H. G.; Wennberg, P. O. Autoxidation of organic compounds in the atmosphere. J. Phys. Chem. Lett. 2013, 4, 3513-3520. 19. Simmie, J. M. Kinetics and thermochemistry of 2,5-dimethyltetrahydrofuran and related oxolanes: Next next-generation biofuels. J. Phys. Chem. A 2012, 116, 4528-4538. 20. Tran, L. S.; Verdicchio, M.; Monge, F.; Martin, R. C.; Bounaceeur, R.; Sirjean, B.; Glaude, P. A.; Alzueta, M. U.; Battin-Leclerc, F. An experimental and modeling study of the combustion of tetrahydrofuran. Combust. Flame 2015, 162, 1899-1918. 21. Fernandes, R. X.; Zador, J.; Jusinski, L. E.; Miller, J. A.; Taatjes, C. A. Formally direct pathways and low-temperature chain branching in hydrocarbon autoignition: the cyclohexyl + O2 reaction at high pressure. Phys. Chem. Chem. Phys. 2009, 11, 1320-1327. 22. Awan, I. A.; Burgess, D. R.; Tsang, W.; Manion, J. A. Shock tube study of the decomposition of cyclopentyl radicals. Proc. Comb. Inst. 2011, 33, 341-349. 23. Wang, K.; Villano, S. M.; Dean, A. M. Reactions of allylic radicals that impact molecular weight growth kinetics. Phys. Chem. Chem. Phys. 2015, 17, 6255-6273. 24. Klute, C. H.; Walters, W. D. The thermal decomposition of tetrahydrofuran. J. Am. Chem. Soc. 1946, 68, 506-511. 25. McDonald, G.; Lodge, N. M.; Walters, W. D. The effect of added gases upon the thermal decomposition of tetrahydrofuran. J. Am. Chem. Soc. 1951, 73, 1757-1760. 26. Lifshitz, A.; Bidani, M. Thermal-reactions of cyclic ethers at high-temperatures .5. Pyrolysis of 2,3-dihydrofuran behind reflected shocks. J. Phys. Chem. 1989, 93, 1139-1144. 27. Lifshitz, A.; Bidani, M.; Bidani, S. Thermal-reactions of cyclic ethers at high-temperatures .2. Pyrolysis of tetrahydrofuran behind reflected shocks. J. Phys. Chem. 1986, 90, 3422-3429. 28. Dagaut, P.; McGuinness, M.; Simmie, J. M.; Cathonnet, M. The ignition and oxidation of tetrahydrofuran. Experiments and kinetic modeling. Combust. Sci. Technol. 1998, 135, 3-29. 29. Kasper, T.; Lucassen, A.; Jasper, A. W.; Li, W. J.; Westmoreland, P. R.; Kohse-Hoinghaus, K.; Yang, B.; Wang, J.; Cool, T. A.; Hansen, N. Identification of tetrahydrofuran reaction pathways in premixed flames. Z. Phys. Chem. 2011, 225, 1237-1270. 30. Verdicchio, M.; Sirjean, B.; Tran, L. S.; Glaude, P. A.; Battin-Leclerc, F. Unimolecular decomposition of tetrahydrofuran: Carbene vs. diradical pathways. Proc. Comb. Inst. 2015, 35, 533-541. 31. Molera, M. J.; Couto, A.; Garciadominguez, J. A. Gas-phase oxidation of tetrahydrofuran. Int. J. Chem. Kinet. 1988, 20, 673-685. 32. Vanhove, G.; Yu, Y.; Boumehdi, M. A.; Frottier, O.; Herbinet, O.; Glaude, P.-A.; Battin-Leclerc, F. Experimental study of tetrahydrofuran oxidation and ignition in low-temperature conditions. Energy Fuels 2015. 33. Osborn, D. L.; Zou, P.; Johnsen, H.; Hayden, C. C.; Taatjes, C. A.; Knyazev, V. D.; North, S. W.; Peterka, D. S.; Ahmed, M.; Leone, S. R. The multiplexed chemical kinetic photoionization mass spectrometer: A new approach for isomer resolved chemical kinetics. Rev. Sci. Instrum. 2008, 79, 104103. 34. Savee, J. D.; Soorkia, S.; Welz, O.; Selby, T. M.; Taatjes, C. A.; Osborn, D. L. Absolute photoionization cross-section of the propargyl radical. J. Chem. Phys. 2012, 136, 134307. 35. Welz, O.; Burke, M. P.; Antonov, I. O.; Goldsmith, C. F.; Savee, J. D.; Osborn, D. L.; Taatjes, C. A.; Klippenstein, S. J.; Sheps, L. New insights into low-temperature oxidation of propane from synchrotron photoionization mass spectrometry and multi-scale informatics modeling. J. Phys. Chem. A 2015, 119, 7116 - 7129. 36. Welz, O.; Zádor, J.; Savee, J. D.; Sheps, L.; Osborn, D. L.; Taatjes, C. A. Low-temperature combustion chemistry of n-butanol: Principal oxidation pathways of hydroxybutyl radicals. J. Phys. Chem. A 2013, 117, 11983-12001. 37. Welz, O.; Zádor, J.; Savee, J. D.; Ng, Y.-H.; Meloni, G.; Fernandes, R. X.; Sheps, L.; Simmons, B. A.; Lee, T. S.; Osborn, D. L., et al. Low-temperature combustion chemistry of biofuels: Pathways in 27 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 29

the initial low-temperature (550 K – 750 K) oxidation chemistry of isopentanol. Phys. Chem. Chem. Phys. 2012, 14, 3112-3127. 38. Dodson, L. G.; Shen, L.; Savee, J. D.; Eddingsaas, N. C.; Welz, O.; Taatjes, C. A.; Osborn, D. L.; Sander, S. P.; Okumura, M. VUV photoionization cross sections of HO2, H2O2, and H2CO. J. Phys. Chem. A 2015, 119, 1279-1291. 39. Ghosh, B.; Papanastasiou, D. K.; Burkholder, J. B. Oxalyl chloride, ClC(O)C(O)Cl: UV/vis spectrum and Cl atom photolysis quantum yields at 193, 248, and 351 nm. J. Chem. Phys. 2012, 137, 164315. 40. Zádor, J.; Miller, J. A. Adventures on the C3H5O potential energy surface: OH + propyne, OH + allene and related reactions. Proc. Comb. Inst. 2015, 35, 181-188. 41. Zádor, J.; Najm, H. N. Automated exploration of the mechanism of elementary reactions (SAND2012-8095); Sandia National Laboratories: Livermore, CA, 2012. 42. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A., et al. Gaussian 09, Revision A.02, Gaussian, Inc.: Wallingford CT, 2009. 43. Kwok, E. S. C.; Atkinson, R. Estimation of hydroxyl radical reaction-rate constants for gas-phase organic-compounds using a structure-reactivity relationship - an update. Atmos. Environ. 1995, 29, 16851695. 44. Alwe, H. D.; Walawalkar, M.; Sharma, A.; Pushpa, K. K.; Dhanya, S.; Naik, P. D. Rate coefficients for the gas-phase reactions of chlorine atoms with cyclic ethers at 298 k. Int. J. Chem. Kinet. 2013, 45, 295-305. 45. Giri, B. R.; Roscoe, J. M. Kinetics of the reactions of cl atoms with several ethers. J. Phys. Chem. A 2010, 114, 8369-8375. 46. Ezell, M. J.; Wang, W. H.; Ezell, A. A.; Soskin, G.; Finlayson-Pitts, B. J. Kinetics of reactions of chlorine atoms with a series of alkenes at 1 atm and 298 K: structure and reactivity. Phys. Chem. Chem. Phys. 2002, 4, 5813-5820. 47. Ballesteros, B.; Ceacero-Vega, A. A.; Garzon, A.; Jimenez, E.; Albaladejo, J. Kinetics and mechanism of the tropospheric reaction of tetrahydropyran with Cl atoms. J. Photochem. Photobiol. AChem. 2009, 208, 186-194. 48. Giri, B. R.; Roscoe, J. M.; Gonzalez-Garcia, N.; Olzmann, M.; Lo, J. M. H.; Marriott, R. A. Experimental and theoretical investigation of the kinetics of the reaction of atomic chlorine with 1,4dioxane. J. Phys. Chem. A 2011, 115, 5105-5111. 49. Cool, T. A.; Wang, J.; Nakajima, K.; Taatjes, C. A.; McIlroy, A. Photoionization cross sections for reaction intermediates in hydrocarbon combustion. Int. J. Mass Spectrom. 2005, 247, 18-27. 50. Wang, J.; Yang, B.; Cool, T. A.; Hansen, N.; Kasper, T. Near-threshold absolute photoionization cross-sections of some reaction intermediates in combustion. Int. J. Mass Spectrom. 2008, 269, 210-220. 51. Bin, Y.; Juan, W.; Cool, T. A.; Hansen, N.; Skeen, S.; Osborn, D. L. Absolute photoionization cross-sections of some combustion intermediates. Int. J. Mass Spectrom. 2012, 309, 118-128. 52. Suzaki, K.; Tsuchiya, K.; Koshi, M.; Tezaki, A. Analysis of HO2 and OH formation mechanisms using FM and UV spectroscopy in dimethyl ether oxidation. J. Phys. Chem. A 2007, 111, 3776-3788. 53. Meloni, G.; Selby, T. M.; Goulay, F.; Leone, S. R.; Osborn, D. L.; Taatjes, C. A. Photoionization of 1-alkenylperoxy and alkylperoxy radicals and a general rule for the stability of their cations. J. Am. Chem. Soc. 2007, 129, 14019-14025. 54. Burke, M. P.; Goldsmith, C. F.; Klippenstein, S. J.; Welz, O.; Huang, H.; Antonov, I. O.; Savee, J. D.; Osborn, D. L.; Zádor, J.; Taatjes, C. A. Multiscale informatics for low-temperature propane oxidation: Further complexities in studies of complex reactions. J. Phys. Chem. A 2015, 119, 7095-7115. 55. Goldsmith, C. F.; Green, W. H.; Klippenstein, S. J. Role of O2 + QOOH in low-temperature ignition of propane. 1. Temperature and pressure dependent rate coefficients. J. Phys. Chem. A 2012, 116, 3325-3346. 28 ACS Paragon Plus Environment

Page 29 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


TOC graphic


Pressure-Dependent Competition among Reaction Pathways from

Recommend Documents