Thermal decomposition of t-amyl methyl ether (TAME) studied by flash pyrolysis/supersonic expansion/vacuum ultraviolet photoionization time-of-flight mass spectrometry

Article abstract of DOI:10.1016/j.ijms.2010.11.003

Thermal decomposition of the oxygenated fuel component tert-amyl methyl ether (TAME) has been studied by flash pyrolysis up to 1250 K in a 20-100 μs time scale. Pyrolysis was followed by supersonic expansion to isolate intermediates and products, which are monitored by vacuum ultraviolet single-photon ionization time-of-flight mass spectrometry (VUV-SPI-TOFMS). The species detected, such as CH₃, C₂H₄, C ₂H₅, C₄H₈, C₅H ₁₀, C₃H₆O, and C₄H₈O, show competition between molecular elimination and bond fission pathways. The alkenes 2-methyl-1-butene (1) and 2-methyl-2-butene (2), the primary molecular elimination products of TAME, were separately pyrolyzed to evaluate the extent of secondary decompositions, as were the ketones (acetone and 2-butanone) produced by losses of two alkyl radicals. While vicinal elimination of methanol from TAME to form 1 and 2 in an approximate 3:1 ratio begins around 600 K and continues to dominate at higher temperatures, homolysis of TAME to form radicals onsets >825 K, yielding more acetone than 2-butanone. Contributions from secondary dissociations of the ketone and alkene products are evaluated.

Full text of DOI:10.1016/j.ijms.2010.11.003

International Journal of Mass Spectrometry 306 (2011) 210–218
Contents lists available at ScienceDirect
International Journal of Mass Spectrometry
journal homepage: www.elsevier.com/locate/ijms
Thermal decomposition of t-amyl methyl ether (TAME) studied by flash
pyrolysis/supersonic expansion/vacuum ultraviolet photoionization
time-of-flight mass spectrometry
Thomas Hellman Morton∗, Kevin H. Weber, Jingsong Zhang∗∗
Department of Chemistry, University of California, Riverside, CA 92521-0403, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 26 August 2010
Received in revised form 14 October 2010
Accepted 1 November 2010
Available online 9 November 2010
Thermal decomposition of the oxygenated fuel component tert-amyl methyl ether (TAME) has been
studied by flash pyrolysis up to 1250 K in a 20–100 ␮s time scale. Pyrolysis was followed by supersonic
expansion to isolate intermediates and products, which are monitored by vacuum ultraviolet single-
photon ionization time-of-flight mass spectrometry (VUV-SPI-TOFMS). The species detected, such as
CH3, C2H4, C2H5, C4H8, C5H10, C3H6O, and C4H8O, show competition between molecular elimination
and bond fission pathways. The alkenes 2-methyl-1-butene (1) and 2-methyl-2-butene (2), the primary
molecular elimination products of TAME, were separately pyrolyzed to evaluate the extent of secondary
decompositions, as were the ketones (acetone and 2-butanone) produced by losses of two alkyl radicals.
While vicinal elimination of methanol from TAME to form 1 and 2 in an approximate 3:1 ratio begins
around 600 K and continues to dominate at higher temperatures, homolysis of TAME to form radicals
onsets >825 K, yielding more acetone than 2-butanone. Contributions from secondary dissociations of
the ketone and alkene products are evaluated.
Dedicated to Tino Gäumann on the
occasion of his 85th birthday.
Keywords:
Fuel oxygenate
Pyrolysis
Free radical
Bond homolysis
DFT
© 2010 Elsevier B.V. All rights reserved.
CCSD
1. Introduction
additive to gasolines. The widely used octane number indicates the
extent of anti-knock properties of a fuel for spark-ignition motors.
At the onset of World War I in the early 20th century it became
critical to augment the power of spark-initiated internal combus-
tion engines to make aviation possible. The fundamental limitation
to increasing the compression ratio (and thus the power) for these
engines was occurrence of “knock”. Originally believed to occur
during the compression stroke of the four-stroke engine, knocking
or “pinging” was shown to result from pre-flame chain branch-
ing reactions during the power stroke [1]. When a fuel mixture
decomposes in an uncontrolled manner ahead of the desired flame
front, a sharp spike in pressure results, offset from the power
stroke, which rapidly destroys the engine cylinder head and pis-
ton. Many molecules were found to prevent knocking (e.g., iodine).
Although aniline was found to have superior anti-knock proper-
ties, the organometallic compound tetraethyllead (TEL) proved to
be most practical [2] and was rapidly incorporated as an anti-knock
By contrast, pro-knock additives confer an advantage for diesel
engines.
Over the years, concerns about automobile emissions have
increased. As catalytic converters were introduced in the 1970s,
methyl tert-butyl ether (2-methoxy-2-methylpropane, MTBE)
became employed as an anti-knock additive, being an afford-
able and cleaner burning replacement for TEL, which was not
compatible with catalytic converters. MTBE, tert-amyl methyl
ether (2-methoxy-2-methylbutane, TAME), and ethyl tert-butyl
ether (2-ethoxy-2-methylpropane, ETBE) have since been used as
components of more modern gasolines due to their anti-knock
properties, reduced volatile organic compound emissions, lack of
toxicity (relative to TEL), and cost-effectiveness.
Unfortunately, MTBE is somewhat soluble in water [3] and,
because of its odor [4], can render ground water unpalatable at
low concentrations. Incidents of leakage from underground stor-
age tanks into water supplies have caused MTBE to be outlawed in
many locales. As a result, higher homologues such as TAME (being
approximately three times less water soluble than MTBE [5]) have
been employed. Compounds that are even less soluble such as tert-
octyl methyl ether (2-methoxy-2,4,4-trimethylpentane, TOME)
[6,7], tert-hexyl methyl ether (2-methoxy-2,3-dimethylbutane,
∗
Corresponding author. Tel.: +1 951 827 4735; fax: +1 951 827 4713.
Corresponding author at: Air Pollution Research Center, University of California,
∗∗
Riverside, CA 92521, United States. Fax: +1 951 827 4713.
E-mail addresses: morton@citrus.ucr.edu (T.H. Morton), jingsong.zhang@ucr.edu
(J. Zhang).
1387-3806/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.ijms.2010.11.003

T.H. Morton et al. / International Journal of Mass Spectrometry 306 (2011) 210–218
211
MDMB) [8], and other homologues [9–14] have also been seri-
ously considered as replacements for MTBE. Nevertheless, the
smaller tert-butyl ethers TAME, ETBE, and MTBE continue to be used
because of the facility of manufacture at the refinery and the ease of
mixing with gasoline on site (although questions remain regarding
their ecotoxicity [15]).
bond homolysis compensates for the higher activation energy. Eq.
(3) illustrates the molecular elimination in competition with bond
cleavages that generate methyl radicals. The uppermost pathway
represents elimination via a 4-center cyclic transition state, while
the lower two depict typical homolyses.
R
R
The anti-knock ability of tert-alkyl ethers has been ascribed
to a propensity for thermal decomposition into alkene plus alco-
hol via concerted vicinal elimination. Both products themselves
confer high octane. Eq. (1) depicts molecular elimination for the
anti-knock compound TAME. The alkenes produced, 1 and 2, can
prevent knock by intercepting reactive radicals (particularly OH)
in the cool-flame region, forming resonance-stabilized radicals,
which are less reactive [16,17]. Conversely, this model implies that
a fuel additive, which itself initiates free radical formation, will
have a pro-knock effect. The present study continues this labora-
tory’s explorations of temperature regimes where MTBE [18] and
its higher homologues [19] undergo bond homolysis to generate
methyl radicals.
+
CX3OH
R
&
R
- RCH2
- CX3
- CH3
O
OCX3
O
tert-alkoxy radical
3, R=CH3
acetone
R=H, X=D: MTBE-d3
R=CH3, X=H: TAME
R
R
- CX3
OCX3
O
methoxyalkyl radical
ketone
4, R=CH3, X=H
(3)
1
OCH
3
molecular
elimination
o
24,25
Δ f
H
= 99 kJ/mol
Below 950 K, diisopropyl ether (a secondary alkyl ether) decom-
poses mainly by molecular elimination [38], while the tertiary alkyl
ethers MTBE [39–48], TAME [44–46,48], and ETBE [43–46] have
been reported to react nearly exclusively via molecular elimination.
Measurable levels of bond homolysis in methyl tert-alkyl ethers
begin to emerge at around 950 K, becoming visible by the detection
of methyl radicals among the pyrolysis products, as observed by
photoionization mass spectrometry [18,19,47]. A peculiarity of this
pathway is the difficulty in seeing the concomitant radical species,
tert-alkoxy or methoxyalkyl radicals, illustrated in Eq. (3). In the
case of MTBE, the methoxymethyl cation contributes the base peak
in the mass spectrum of the parent neutral, so it becomes virtually
impossible to ascertain whether this ion arises from photoioniza-
tion of neutral MTBE or of radical intermediates.
1000K
4
2
1
3
- CH OH
3
o
20,21
22,23
Δ f
H
=
-116 kJ/mol
-42 kJ/mol
1000K
2
o
24,26
Δ f
H
= 89 kJ/mol
1000K
(1)
As Eq. (1) summarizes, molecular elimination from TAME is
endothermic by ꢀH = 60–70 kJ mol−1 at 1000 K, depending on the
hydrocarbon product [20–27]. The entropy change for splitting one
molecule into two has a positive value, giving a value on the order
of TꢀS1000 = −105 kJ mol−1 for Eq. (1). In other words, molecular
elimination at 1000 K has a negative ꢀG.
2. Background
As the uppermost pathway in Eq. (3) depicts, the hydrocarbon
from molecular elimination product of MTBE, isobutene, appears
first and persists in competition with bond homolysis up to tem-
peratures at which it, too, undergoes thermal decomposition. Eq.
(1) illustrates the corresponding products from TAME. On the one
hand, isomers 1 and 2 should form in a 3:1 ratio based on purely
statistical considerations of the number of vicinal hydrogen atoms.
On the other hand, thermochemistry favors 2: the 1 ꢀ 2 equilibrium
proportions in the gas phase have been measured to as 7.5:92.5 at
room temperature [20] and should have a value on the order of
35:65 at 1000 K [24–26].
Thermal decomposition of a variety of alkyl ethers has been
studied using static and flow reactors, as well as shock tubes.
Among primary alkyl ethers, dimethyl ether decomposes via bond
homolysis to give free-radicals [28–31], while methoxyethane [32]
and diethyl ether exhibit both homolysis and molecular elimina-
tion [33–37]. Shock tube studies monitor bond homolysis in the
presence of oxygen at high temperatures by looking at hydroxyl
radicals produced via the sequence shown in Eq. (2).
•
CH3OR−→ CH3O• −→ H•ad−de→d O HO•
(2)
–R
–CH
O
2
2
As the lower pair of pathways in Eq. (3) depict, tert-alkoxy or
methoxyalkyl radicals such as RCH2CMe2O• or CD3OCMeCH2R•
ought readily to expel a molecule of ketone to yield a methyl radical
(CH3• or CD3•, respectively). The observation of both m/z 15 and m/z
18 from photoionization of pyrolysis products of MTBE-d3 corrob-
orates that expectation [18,19]. Although acetone is reported in the
shock tube study of MTBE, this product is seen under flash pyroly-
sis conditions to a lesser extent than might have been anticipated
Eq. (2) depicts a strategy for indirectly measuring the pro-
duction of methoxy radicals and other intermediates that lose a
hydrogen atom, but it does not detect methyl radicals. Methoxy
radicals dissociate to formaldehyde plus hydrogen atoms, and
these atoms react with added oxygen to produce hydroxyl radicals,
which are monitored by laser-induced fluorescence. In the present
work, 118.2 nm photoionization of the pyrolysis products pro-
vides direct observation of methyl radicals, although this technique
cannot detect hydrogen atoms, formaldehyde, carbon monoxide,
or other species with ionization energies above 10.5 eV. At tem-
peratures ≥1000 K, the pyrolysis/supersonic jet/photoionization
method used here provides an approach complementary to shock
tube studies.
•
on the basis of the stoichiometry Me3COMe → Me2C O + 2CH3
[18,19,48]. A 3:1 ratio of m/z 15 to m/z 18 from Me3COCD3 sug-
gests that the acetone itself decomposes further to yield carbon
monoxide and a net total of 4 methyl radicals.
The question arises as to why the ketone yield is low, even at
the onset of bond homolysis. To that end, this study examines the
pyrolysis of TAME, for which one of the molecular elimination prod-
ucts, 2-methyl-1-butene, 1, ought to display less thermal stability
than do acetone or 2-butanone, formed by two successive losses of
free radicals. The 298 K C2–C3 bond dissociation energies of acetone
and 2-butanone (based on the ꢀfH values for methyl radical [49],
Competition between molecular elimination and bond homol-
ysis depends on temperature. Molecular eliminations dominate
thermal decomposition pathways at moderately elevated temper-
atures, because they have lower activation energies, but at higher
temperatures the more favorable activation entropy of a simple

212
T.H. Morton et al. / International Journal of Mass Spectrometry 306 (2011) 210–218
ethyl radical [50] and acetyl radical [51]) are ꢀH = 353 kJ mol−1 and
ꢀH = 350 kJ mol−1, while the dissociation of 1 to give methallyl [52]
3. Experimental
plus methyl radicals has an endothermicity of ꢀH = 308 kJ mol−1
.
The thermal decomposition experiments were performed with
an apparatus that has been previously described [18]. Briefly,
it makes use of a Wiley–MacLaren type linear time-of-flight
mass spectrometer to observe pyrolysis products by means of
118.2 nm photoionization. TAME (99+%) and 2-methyl-2-butene
(99+%) were obtained from Sigma–Aldrich. 2-Methyl-1-butene
(99+%) was obtained from Columbia Organic Chemicals. Samples
were used without further purification and introduced to the
pyrolytic apparatus diluted in helium, argon, carbon tetrafluoride,
or sulfur hexafluoride by bubbling the carrier gas through the liq-
uid. The backing pressure of the gas mixture was maintained at
approximately 1.5 atm for all experiments.
The pyrolysis micro-reactor is based on the design described
by Chen and co-workers [53]. Thermal decompositions are carried
out by expanding the gas mixture through a heated SiC nozzle (Car-
borundum, heated length 10 mm, 2 mm o.d., 1 mm i.d.) attached to
a machinable alumina piece that isolates the nozzle thermally and
electrically from the pulsed valve. The nozzle is heated resistively
by means of two press-fitted graphite electrodes with the current
controlled by a Variac transformer. The temperature of the nozzle
was measured with a type C (Omega) thermocouple attached to the
exterior of the reaction zone, which had previously been calibrated
to the internal temperature. The reliability of the temperature mea-
surement for the nozzle has been investigated previously [19] and
was determined to be accurate to 50 K. The speed of the sample
flow was shown to be approximately sonic with a residence time
of around 20–100 ␮s depending on the carrier gas utilized [53,54].
Upon exit from the heated reaction zone the nascent products,
intermediates, and any remaining reactants undergo supersonic
expansion, which significantly reduces their internal energy con-
tent [55]. A molecular beam extracted from this expansion is then
intercepted by 118.2 nm photons (10.49 eV) produced by frequency
tripling of the 355 nm third harmonic of a Nd:YAG laser in a Xe
cell (approximately 20 Torr). The 118.2 nm light is focused by a
MgF2 lens through a small aperture into the photoionization region
with concomitant divergence of the fundamental 355 nm source to
minimize multiphoton ionization. TOF spectra are collected using
a multichannel plate detector and a digital oscilloscope with 512
laser shots averaged and then converted to mass spectra.
One therefore expects to see the ketone products and C5H10
concurrently from pyrolysis of TAME at the onset of bond homol-
ysis. As will be described below, experimental evidence indicates
that 1 constitutes the major isomer from molecular elimination. Yet
TAME does not yield ketones to the extent expected on the basis of
stoichiometry under conditions where pyrolysis produces radicals.
This leads to a variety of plausible suppositions. Surface cataly-
sis represents one possibility, a consideration that applies to all gas
phase pyrolyses (with the exception of shock tube studies). Perhaps
the hot surface catalyzes bond homolysis, so that radical products
stick to it, and additional surface catalysis degrades the ketones
further, as Eq. (4) depicts. Even if ketone decomposition does not
undergo surface catalysis, adhesion of the precursor radicals would
increase contact time, giving those products a longer period in
which to dissociate. Further consideration of the role of surface
chemistry lies outside the scope of the present study (except to
note that the consequences of radical formation in an automobile
cylinder do not depend on the homogeneity of the reaction).
(4)
Alternatively, if bond homolysis does occur homogeneously in
the gas phase, ketones from the second steps of Eq. (3) might be
formed with internal vibrational excitation. Because (as mentioned
below in Section 5) shock tube studies confirm the formation of
methyl radicals via a homogeneous gas phase reaction, this inter-
pretation appears as the most reasonable explanation at the present
moment.
The objectives here include an experimental estimate of the
yield of neutral radicals and of ketones from pyrolysis of TAME,
as well as a theoretical evaluation of the vibrational excitation that
might accompany further radical loss from an intermediate radical
(the second step of the lower pair of pathways in Eq. (3)). While the
results in hand do not rule out any of the aforementioned options,
they do place constraints on the homogeneous gas phase pathway.
A relatively limited amount of published experimental data doc-
ument the pyrolysis of TAME. Oxidation of TAME has previously
been observed by static reactor [43,44], high pressure jet-stirred
reactor [45], and in non-premixed flame [46]. The dominant decom-
position reaction over intermediate temperatures (700–1000 K)
produces 2-methyl-1-butene (1) and 2-methyl-2-butene (2) with
concomitant formation of methanol via the molecular elimination.
Homolytic bond scissions become competitive at higher tempera-
tures. The molecular elimination product branching distribution,
treated thus far as completely statistical, is further investigated
here.
The experimental approach – flash pyrolysis coupled to super-
sonic expansion and vacuum ultraviolet (VUV) photoionization
mass spectrometry – offers short reaction times to examine the
initial steps of the thermal decomposition, supersonic cooling to
quench the subsequent reaction and to minimize recombination
of the initial products and intermediates, and “soft” VUV photoion-
ization to minimize ion fragmentation allowing for direct detection
of transient intermediates. Quantum chemical calculations of bond
homolyses assist in elucidating the most likely reaction pathways
and provide a prelude to the experimental results.
Energetics of selected species involved in the decomposition
of TAME were calculated using the Gaussian03 program suite
[56]. Geometries were optimized using the cc-pVTZ basis set by
means of the B3LYP hybrid density functional theory method
[57], with harmonic normal modes calculated at the same level
of theory. Single-point calculations of electronic energies using
CCSD/cc-pVTZ were carried out on the B3LYP/cc-pVTZ optimized
structures.
4. Results
Setting aside its CHs, TAME possesses five nonequivalent single
bonds, of which four are calculated to have comparable bond disso-
ciation energies (the fifth, the C3–C4 bond, is much stronger, and its
cleavage will not be further discussed). Table 1 summarizes compu-
tations at two levels, including CCSD single point calculations using
the B3LYP/cc-pVTZ geometries. The calculations here focus on the
relative energies of the four competing homolytic pathways and
the barriers for loss of methyl radicals from intermediates 3 and 4
to yield 2-butanone. As can be seen, the ordering of homolytic ther-
mochemistries differs substantially between the B3LYP and CCSD
levels. The CCSD calculations predict the O-methyl bond to have
the lowest dissociation energy, consistent with previously pub-
lished calculations on MTBE [58]. In other words, CCSD predicts

T.H. Morton et al. / International Journal of Mass Spectrometry 306 (2011) 210–218
213
Table 1
Electronic energies (in a.u.) of radicals from competing bond homolyses of TAME and of transition states for radical expulsions from 3 and 4, along with zero
point energies and relative 1000 K bond dissociation energies (in kJ mol−1), as well as 0 K activation barriers for the radical expulsions.
that formation of t-amyloxy radical, 3, represents the lowest energy
homolytic cleavage of TAME.
ton energies well above threshold, ionized TAME fragments with
preference for expulsion of the larger alkyl group, as also observed
experimentally with other tertiary alkyl methyl ethers [18,19].
In Fig. 1A, a trace amount of 2,3,3-trimethyl-1-butene used to
facilitate mass calibration is observed as a small M+• peak at m/z 98
and, as the nozzle temperature is increased, its M−CH3+ peak at m/z
83 is also discernible. A small peak at m/z 78 is also noted, residual
C6H6 produced from these pyrolysis experiments at higher temper-
atures. With a nozzle temperature of 575 K, the intensity of the m/z
73 peak from photoionization of the parent TAME decreases, while
the intensity at m/z 87 (also from TAME) remains approximately the
same. A new peak at m/z 70 with approximately 20% the intensity
of the m/z 73 peak is now detected, corresponding to hydrocarbons
1 and 2. These are the molecular elimination products of TAME
(see Eq. (1)). Methanol produced concomitantly by same process
has an IE (10.85 eV) [60] that exceeds the photoionization energy
employed and hence cannot be seen. The peaks associated with
the trace impurities diminish in intensity except for the m/z 83
fragment from 2,3,3-trimethyl-1-butene, which now becomes vis-
ible. When the nozzle is further heated to temperatures of 875 and
950 K the intensities of m/z 73 and m/z 87 steadily decrease as the
parent compound is consumed. The product peak at m/z 70 grows
to roughly one half, then nearly equal intensities with respect to
TAME’s base peak at m/z 73. Small peaks at m/z 55 and 56 become
discernible, as well as small peaks at m/z 15, corresponding to CH3
The transition state calculations show barriers for subsequent
radical loss. The bottom pathway in Eq. (4) (loss of methyl from 4)
has a much higher activation energy than do the dissociations of
3. Thermodynamically, expulsion of ethyl radical from 3 to give
acetone has a ꢀH◦
of 5 kJ mol−1 less endothermic than does
298
expulsion of methyl radical to give 2-butanone, but the barrier
heights differ by 9–14 kJ mol−1 (depending on the level of com-
putation). In the transition state calculated for methyl loss from
˚
3, the radical carbon lies 2.106 A away from C2, while the transi-
˚
tion state for ethyl loss has a shorter distance, 2.074 A. With regard
to energetics, the difference between those two transition states
implies a greater yield of acetone than of 2-butanone from dissoci-
ation of intermediate 3. The computational results parallel the data
seen experimentally for the TAME molecular ion, which exhibits a
reverse activation barrier for methyl loss greater than for ethyl loss,
such that the latter fragmentation has a lower appearance energy
(even though it is more endothermic) and a greater abundance at
ionizing energies above threshold [59].
4.1. Pyrolysis of tert-amyl methyl ether (TAME)
Fig. 1A and B presents stack plots of mass spectra for the low
temperature pyrolysis, up to 1000 K of TAME seeded in argon and
sulfur hexafluoride, respectively. In Fig. 1A, with the nozzle at
room temperature, a molecular ion is barely detectable (M+• = m/z
102 for TAME), which has not previously been observed by mass
+
radical, and product acetone at m/z 58. At 1000 K, the M−CH3 ion
peak of TAME at m/z 87 is barely detectable and m/z 70 has increased
+
to nearly double the intensity of M−CH2CH3 ion peak at m/z 73.
+
The m/z 56 peak has grown in relative intensity to m/z 55. Small
peaks at m/z 58 and m/z 72 are detectable, corresponding to the
oxygenated products acetone (C3H6O) and 2-butanone (C4H8O),
as well as at m/z 68 corresponding to isoprene, C5H8. A series of
small peaks over m/z 40–43 are observed and the m/z 15 CH3 peak
intensity increases.
spectrometry. A large M−CH2CH3 ion base peak at m/z 73 is
+
observed accompanied by a M−CH3 peak at m/z 87 having 8% of
the base peak intensity. As mentioned above, photoionization of
TAME shows that the thermodynamically more favorable dissocia-
tion to m/z 87 has a higher appearance energy (AE) than for m/z 73:
AE (m/z 73) = 9.25 eV and AE (m/z 87) = 9.36 eV [59]. Even at pho-

214
T.H. Morton et al. / International Journal of Mass Spectrometry 306 (2011) 210–218
A
m/z 73
0.10
0.08
0.06
0.04
0.02
0.00
m/z 15
m/z 28
m/z 29
m/z 40
m/z 58
m/z 70 (intensity divided by 10)
m/z 72
m/z 70
m/z 55
m/z 40
m/z 87
1000K
950K
m/z 15
875K
575K
295K
750
800
850
900
950
1000
Heater Temperature (K)
Fig. 2. Intensities of selected fragment ions (relative to m/z 73) from pyrolysis of
TAME in argon below 1000 K.
0
10 20 30 40 50 60 70 80 90 100
m/z
B
velocity at the throat of the expansion is approximately sonic, hav-
ing an estimated residence time of 20 ␮s with helium as a carrier
gas [53]. Since the speed should track inversely with square root of
mass, the residence times of argon, carbon tetrafluoride and sulfur
hexafluoride ought to be approximately three, five, and six times
longer, respectively. With helium as a carrier gas the pyrolysis of
TAME onsets at temperatures >600 K (data not shown here); how-
ever, with argon carrier gas pyrolysis onsets below 600 K (Fig. 1A).
At a nozzle temperature of 875 K the intensity of m/z 70 relative to
m/z 73 is approximately 25% in helium and 50% in argon. The cooling
efficiency for these gases are similar and, as their heat capacities are
virtually the same, the difference in the onset and extent of pyrol-
ysis must be due to the different residence times. At temperatures
>950 K the spectra using helium or argon are indistinguishable. For
the polyatomic halogenated gases the presence of internal degrees
of freedom increases heat capacities relative to the monoatomic
gases. Residence times are longer, and the onset and extent of pyrol-
ysis are significantly delayed compared to inert gas carriers. With
sulfur hexafluoride as a carrier gas the onset of pyrolysis does not
start until temperatures >600 K (Fig. 1B), and, at a nozzle temper-
ature of 1050 K, the m/z 70 product peak is only roughly one half
the m/z 73 parent ion intensity, while in helium and argon m/z
73 peak is greatly reduced. The [M−CH3]+:[M−CH2CH3]+ ion ratio
(m/z 87:m/z 73) for pyrolyses in helium, argon, carbon tetrafluo-
ride, and sulfur hexafluoride carrier gases were determined. Within
experimental error all carrier gases give the same ratio of photoion-
ization fragmentation, with the [M−CH3]+ ion having 13 4% the
intensity of the [M−CH2CH3]+ ion, consistent with previous mass
spectrometric studies [59]. While the onset temperature for pyrol-
ysis and the extent of reaction do differ (as indicated by comparing
Fig. 1A and B), argon was chosen as the carrier gas for subsequent
experiments to produce best quality spectra.
Fig. 2 summarizes the relative abundances of significant pyrol-
ysis products formed in argon at temperatures <1000 K relative to
the m/z 73 base peak from ionization of TAME. While molecular
elimination of methanol (giving m/z 70) clearly dominates, radicals
from homolysis (m/z 15 and m/z 29) start to be observed at 835 K,
as does acetone (m/z 58), which results from the successive loss of
methyl and ethyl groups from TAME (the middle pathway in Eq.
(3)). Because it lies between m/z 70 and m/z 73, the mass spectro-
metric peak corresponding to 2-butanone (m/z 72, corresponding
to loss of two methyls from TAME) emerges from background at
higher temperatures, but it has an intensity approximately half that
of acetone. As Section 5 will note, these ketones exhibit an abun-
dance less than one would expect based on the observed intensities
of methyl and ethyl radicals.
m/z 73
m/z 70
m/z 15
m/z 55
m/z 40
m/z 87
1050K
1000K
975K
875K
575K
0
10 20 30 40 50 60 70 80 90 100
m/z
Fig. 1. (A) Stack plot of mass spectra for pyrolysis of TAME in argon with internal
nozzle temperatures from near room temperature (295 K) to 1000 K. (B) Stack plot
of mass spectra for pyrolysis of TAME in sulfur hexafluoride with internal nozzle
temperatures from 575 K to 1050 K. The mass spectra are shifted for clarity.
Fig. 1B presents a stack plot of mass spectra for the pyrolysis of
TAME with sulfur hexafluoride as the carrier gas. Although not pic-
+
tured here, the room temperature spectrum produces a M−CH3
peak with 8% of the intensity of the M−CH2CH3+, the same rela-
tive intensity observed in the argon entrained sample. At a nozzle
temperature of 595 K the product peak at m/z 70 is barely dis-
cernible, while at this nozzle temperature in argon, the m/z 70
peak has already reached about 20% of the m/z 73 base peak. The
+
M−CH3 ion peak at m/z 87 is found to have approximately 10%
+
of the intensity of the m/z 73 M−CH2CH3 ion peak, similar to
the argon data in Fig. 1A. With the higher nozzle temperature of
875 K, the m/z 70 product peak grows to roughly 10% of the m/z
73 intensity while m/z 87 and m/z 73 peaks remain essentially
unchanged. Nozzle temperatures of 975, 1000, and 1050 K do not
significantly change the intensities of m/z 73 and m/z 87 while the
peaks from 1 and 2 at m/z 70 increases to greater than half of the
m/z 73 peak intensity. New peaks at m/z 15, 40–43, 55, 56, and 58
are now clearly detected and, at 1050 K, a small peak at m/z 68 is
observed.
The carrier gas was changed to examine the effect on the pyroly-
sis due to (1) different residence times in the microreactor, (2) heat
capacities of carrier gas, and (3) cooling efficiency of the gas. The

T.H. Morton et al. / International Journal of Mass Spectrometry 306 (2011) 210–218
215
A
m/z 40
m/z 40
m/z 68
m/z 68
m/z 55
m/z 55
m/z 78
m/z 15
m/z 78
m/z 15
m/z 28
m/z 28
1240K
1205K
1190K
1135K
1090K
1240K
1200K
1160K
1110K
10
20
30
40
50
60
70
80
90
0
10
20
30
m/z
40
50
60
70
m/z
Fig. 3. Stack plot of mass spectra for pyrolysis of TAME in argon with internal nozzle
temperatures from 1100 K to 1240 K.
B
m/z 78
Overlap with other symbols in Fig. 2 obscures the fact that
methyl and ethyl radicals onset at the same temperature. At a
slightly higher temperature (875 K) a m/z 28 appears, which cor-
responds to loss of a hydrogen atom from the ethyl radical. Above
900 K the m/z 28 peak displays an intensity roughly half that of
m/z 29. Because ethylene has an IE (10.514 eV) very close to the
energy of the 118.2 nm photons used in these experiments, it seems
probable that ethylene has an abundance substantially greater than
indicated by the intensity of its peak in the photoionization mass
spectrum.
m/z 68
m/z 55
m/z 40
m/z 15
m/z 28
1245K
1190K
1170K
1110K
0 10
Fig. 3 presents a stack plot for the high temperature pyrolysis
(1110–1250 K) of TAME seeded in argon. Above 1000 K, the m/z 28
peak starts to increase relative to m/z 29 peak, and above 1100 K
the m/z 28 intensity exceeds that of m/z 29. At 1100 K little m/z 73 is
detected, being approximately 10% of the m/z 70 product (C5H10
)
20
30
m/z
40
50
60
70
peak. The intensity of the peak at m/z 40, C3H4, increases drasti-
cally to nearly the same intensity as m/z 70, while the signals at
m/z 15 (methyl radical), m/z 54 (C4H6), m/z 56 (C4H8), and m/z 68
(C5H8), have all increased in intensity. The production of C4H8 and
a portion of the CH3 signal is due to pyrolysis of TAME, while the
other signals are known products from the pyrolysis of the C5H10
hydrocarbons (vide infra) [46]. As the heater is raised to the higher
temperatures of 1160, 1200, and 1240 K the m/z 70 peak steadily
decreases in intensity, and in the highest temperature trace, 1240 K,
m/z 78 is detected, corresponding to C6H6 (plausibly from the self
combination of m/z 39 C3H3 radicals). The peaks at m/z 15, 28, and
40 continue to become more intense. The isobutylene peak at m/z
56 begins to slightly decrease in intensity, while new peaks over the
m/z range of 52–54 are observed. At these higher temperatures the
m/z 68 peak remains relatively constant and the m/z 73 peak from
TAME is hardly detectable. Arrhenius behavior would predict the
half life for TAME [44,45] at a temperature of 1200 K to be on the
order of 70–150 ␮s with respect to molecular elimination, which
is comparable to the estimated residence time of 60 ␮s for argon.
Experimentally, TAME is found here to decompose completely at
this temperature: the M−CH3 ion at m/z 87 and M−CH2CH3 ion at
m/z 73 corresponding to the photoionization fragments of TAME
are not observed >1200 K.
Fig. 4. (A) Stack plot of mass spectra for pyrolysis of 2-methyl-1-butene (1) in
argon with internal nozzle temperatures from 1090 K to 1240 K. (B) Stack plot of
mass spectra for pyrolysis of 2-methyl-2-butene (2) in argon with internal nozzle
temperatures from 1110 K to 1245 K.
1 and 2 in argon carrier gas were conducted. Fig. 4A and B presents
stack plots of mass spectra for temperatures from 1100 to 1250 K.
When comparing the two isomers in Fig. 4 the mass spectrometric
patterns of 1 and 2 undergoing pyrolysis shows significant differ-
ences. Relative to the molecular ion signal at m/z 70, a large amount
of m/z 68 is produced in the pyrolysis of 2 compared to 1. This
is explained by the propensity of 2 to undergo molecular decom-
position to evolve hydrogen gas via 1,4-elimination and produce
2-methyl-1,3-butadiene (isoprene) in preference to radical decom-
position. The pyrolysis of 1, on the other hand, produces a greater
amount of signal at m/z 40, C3H4, via loss of the elements of ethane.
Cleavage of a methyl radical to form methallyl radical followed by
subsequent loss of a second methyl radical provides a reasonable
explanation for m/z 40, predicting that the isomer thus produced
has the structure of allene.
Although the two alkenes have distinguishably different pyrol-
ysis spectra, a pair of obstacles limits an accurate determination of
the ratio of the product C5H10 isomers in the pyrolysis of TAME: (1)
differences in residence time for products compared to authentic
samples and (2) interference from competing pyrolytic processes.
In an attempt to determine the product isomer branching ratio from
4.2. Pyrolyses of 2-methyl-1-butene (1) and 2-methyl-2-butene
(2)
In order to evaluate the contribution of product fragmentation
and investigate relative product yields, pyrolyses of hydrocarbons

216
T.H. Morton et al. / International Journal of Mass Spectrometry 306 (2011) 210–218
of the alkenes produced by TAME is accelerated. The discussion
below deals with conclusions based on this inference.
m/z 15
m/z 28
m/z 29
m/z 40
m/z 55
0.15
0.10
0.05
0.00
5. Discussion
The observation of m/z 15 and m/z 29 from the supersonic jet
expansion/photoionization of TAME pyrolysis products testifies as
to the formation of radicals from bond homolysis. A recent shock
tube study [47], which reports the formation of ethane from high
temperature pyrolysis of MTBE, confirms the formation of methyl
radicals from homogeneous gas phase decomposition of highly
branched methoxyalkanes, as first described by this laboratory
a decade ago [18]. Other products reported from gas chromato-
graphic analysis of the shock tube products from MTBE also agree
with those observed in the present study, especially allene and
propyne (here observed as m/z 40 peaks). The peaks at m/z 40 also
arise from pyrolysis of 1 and 2 via loss of the elements of ethane.
This pathway is much more prominent for 1 than for 2, consistent
with the structural difference between those two isomeric alkenes.
Mass spectra of the pyrolysis products of TAME exhibit the same
m/z 73: m/z 87 intensity ratio as seen from photoionization of TAME
itself. In principle, three bond cleavages could generate free radicals
having these masses: the loss of ethyl radical to yield (CH3)2COCH3
(6, 73 amu) and the two pathways for methyl loss shown in Eq.
(3) (to yield 87 amu radicals, 3 and 4). Because it seems implausi-
ble that the proportions of bond cleavage of neutral TAME should
so closely match the fragmentation pattern of ionized TAME, the
aforementioned radicals are inferred to be largely absent.
As Table 1 summarizes, four bond homolyses can compete in the
thermal dissociation of neutral TAME below 1000 K. DFT and CCSD
calculations predict different orderings of the relative bond disso-
ciation energies, but it appears that the pyrolyses to yield methyl
radicals have thresholds comparable to the losses of methoxyl radi-
cal or ethyl radical. As Eq. (2) indicates, further dissociation renders
methoxyl radical invisible under the experimental conditions used
here. The concomitant tert-amyl radical, 5, can dissociate via further
loss of CH3 to give isobutene (m/z 56), which (so far as can be ascer-
tained from Fig. 1) exhibits an onset not far from the temperature
at which m/z 15 and m/z 29 begin to emerge.
At elevated temperatures virtually all organic free radicals from
bond homolyses (with the exception of methyl radical, CnHn, and
other highly unsaturated analogues) dissociate via loss of a hydro-
gen atom or CH3. This implies, in turn, that ketones – acetone and
2-butanone – ought to form. While ions at the appropriate masses
(m/z 58 and m/z 72) do appear, they exhibit intensities lower than
that of the concomitant methyl radicals at all temperatures. It is
not easy to account for the discrepancy simply in terms of relative
ionization cross sections.
According to the CCSD calculations in Table 1, the lowest energy
pathway for TAME homolysis cleaves the C–O bond to produce 3,
which can then lose ethyl radical to give acetone (the middle path-
way in Eq. (3)). Since m/z 15, m/z 29, and m/z 58 onset at the same
temperature, the experimental results in Fig. 2 agree with this pre-
diction, although the data do not permit an assessment of whether
CH3 or C2H5 comes off first. Shock tube studies of the thermal
decomposition of acetone, if fit to the Arrhenius equation [62,63],
predict that it should have a half-life for homogeneous dissociation
at 1000 K on the order of 3–8 ms. While more recent studies [64] fit
the data to non-Arrhenius behavior, acetone ought to remain stable
much longer that the contact time in the present experiments. The
following discussion interprets the comparatively low yields of ace-
tone and of 2-butanone in terms of vibrational excitation remaining
after expulsion of radicals in the second steps of Eq. (3).
650
700
750
800
850
900
950 1000 1050 1100
Heater Temperature (K)
Fig. 5. Intensities of selected fragment ions relative to the molecular ion at m/z 70
(minus peak intensities at temperatures below 700 K) from pyrolysis of 2-methyl-
1-butene (1) in argon.
the TAME pyrolysis, the relative amounts of m/z 40 and m/z 68 in
sample mixtures composed of 0%, 25%, 50%, and 100% 1 in 2 were
pyrolyzed and compared with the spectra TAME at higher tem-
perature (Fig. 3). It was apparent that the majority of molecular
elimination product is 1, consistent with the previous study [44].
This comparison suggested a branching ratio value of 75 20% 1
from TAME. Another noticeable differentiation between the 1 and
2 is the amount of m/z 78 observed. In Fig. 4A, pyrolysis of 1, a signal
at m/z 78 is detected at temperatures ≥1190 K. In the pyrolysis of 2
presented in Fig. 4B, a signal at m/z 78 is only barely detected in the
hottest temperature trace at 1245 K, because 2 undergoes molec-
ular elimination of H2 in preference to C3H4 (m/z 40) production.
Conversely, 1 prefers to decompose to C3H4 and can then lose a H
atom to form propargyl radicals, which then combine to form C6H6
[61].
Comparing Fig. 4A and B to Fig. 3 it is apparent that the high-
temperature pyrolysis of TAME produces mass spectra similar to
those produced by pyrolysis of 1 and 2. The most noticeable dif-
ferences distinguishing the TAME data from the C5H10 pyrolysis
experiments, however, are greater abundances of m/z 15, 28, and 56
from TAME. Fig. 5 summarizes the behavior of 2-methyl-1-butene
(1) at the lower temperatures at which it begins to decompose.
Fission of the sp3–sp3 carbon–carbon bond of 1 to form methallyl
radical represents the least endothermic homolysis pathway, but
the m/z 55 ion also corresponds to the base peak in the photoion-
ization mass spectra of 1 and 2. The emergence of this ion between
835 and 890 K above that substantial background in Fig. 5 (which
corrects for the presence of that peak in the mass spectrum of 1) is
taken to signal the onset of dissociation of neutral 1. The formation
of methyl radicals (m/z 15, which contributes a negligible peak to
the photoionization mass spectrum) ought to accompany homol-
ysis of 1 but does not become visible above noise level until 930 K
and, as Fig. 4A shows, exhibits an intensity equal to that of m/z
55 only above 1200 K. This discrepancy could be attributed to the
difference in 118.2 nm photoionization cross sections of the two
radicals.
Loss of the elements of ethane to yield C3H4 (m/z 40, which also
corresponds to a negligible peak in the mass spectra of 1 and 2)
provides another gauge of thermal decomposition of neutral 1. This
dissociation, which can be attributed to further loss of methyl radi-
cal from the methallyl radical, also emerges between 835 and 890 K,
the same onset temperature (within experimental uncertainty) as
in the pyrolysis of TAME (as Fig. 2 summarizes). In the case of TAME,
however, molecular elimination has to occur prior to decomposi-
tion to C3H4, which implies that C5H10 has a much shorter residence
time under those conditions. This suggests that further dissociation
To begin with, the evidence suggests that elimination of
methanol from TAME produces vibrationally excited C5H10. As
Fig. 2 indicates, the onset of m/z 40 from TAME at the same tem-

T.H. Morton et al. / International Journal of Mass Spectrometry 306 (2011) 210–218
217
perature as in the pyrolysis of 1 summarized in Fig. 5, even though
6. Conclusion
the C5H10 residence time is substantially shorter in the former case.
Loss of two successive methyl radicals from C5H10 accounts for the
formation of C3H4. In the TAME pyrolysis, as in the pyrolysis of 1
summarized in Fig. 5, m/z 40 appears at the same temperature as
does m/z 55 (after subtracting the contribution to the intensity of
latter from photoionization of C5H10). Close to onset the m/z 40:m/z
55 ratio has a value close to 0.15 in both cases. Because molecu-
lar elimination from TAME has an activation barrier ≥230 kJ mol−1
[44], much higher than the thermodynamic thresholds in Eq. (1),
it comes as little surprise that the resulting alkenes can form with
vibrational excitation that leads to decomposition rates faster than
from thermal activation by itself. The question arises whether the
same effect operates for the ketones formed from TAME.
Loss of alkyl radicals from intermediates 3 or 4 have calculated
electronic energies close to zero, for which ꢀεvib more than com-
pensates at high temperatures (an effect attenuated somewhat by
inclusion of ꢀε from molecular rotations): e.g., at B3LYP/cc-pVTZ
the exothermicity of methyl loss from 4 has a calculated value of
ꢀH◦ = −37 kJ mol−1 (taking into account the changes in ZPE and
harmonic canonical energies). The computed activation energies
summarized in the last three entries of Table 1 show surprisingly
high barriers for further expulsion of radicals from radicals 3 and
4. The excess internal energy from surmounting these barriers as
well as the net exothermicities could enhance the rate of further
dissociation of the resulting ketones.
Photoionization mass spectrometry of pyrolysis products
reveals that thermal dissociation of neutral TAME in argon displays
a threshold below 600 K, somewhat lower than the onset tem-
perature for MTBE in argon (approximately 700 K) with the same
experimental setup. Over the temperature range typically applica-
ble to operation of an internal combustion engine (700–1000 K),
molecular elimination of methanol to form 2-methyl-1-butene (1)
and 2-methyl-2-butene (2) dominates, with a nearly a statistical
distribution of products (3:1 ratio of 1–2). At temperatures >825 K
the observation of methyl radicals indicates the onset of bond
homolysis, as previously reported for MTBE. Ethyl radicals appear
from TAME with the same threshold but tend to lose H• to give ethy-
lene at higher temperatures. Acetone, 2-butanone, and C4H8 also
form via successive losses of two radicals, as does C3H4 from the
C5H10 products of molecular elimination. Acetone is more abun-
dant than 2-butanone, implying that losses of ethyl and methyl
from TAME occur with greater facility than loss of two methyls.
That experimental result agrees with CCSD predictions of 39 and
48 kJ mol−1 activation barriers for expulsion of ethyl and methyl
radicals, respectively, from the intermediate tert-amyloxy radical
(3). Secondary decomposition of the ketones and of C5H10, which
are not as robust as the isobutene formed from MTBE, takes place
and appears to be accelerated as a consequence of the internal
energy remaining after their formation.
The yield of m/z 15 near onset exceeds the combined yields of
the C2 fragments m/z 29 plus m/z 28. Although competing homol-
yses can also yield methyl radicals, the C2 yield ought to equal that
of acetone, but such is not the case <975 K. Since 3 and 4 have
calculated activation barriers >30 kJ mol−1 for subsequent expul-
sion of alkyl radicals, their dissociation might partition a significant
fraction of the excess energy into internal degrees of freedom of
the ketones (as occurs for the molecular elimination leading to
alkenes), giving rise to further decomposition and reducing ketone
abundance below that expected on the basis of stoichiometry.
The transition state calculations in Table 1 predict that the
tert-amyloxy radical ought to expel ethyl radical (to give acetone)
more readily than methyl (to give 2-butanone). Expulsion of the O-
methyl from 2-methoxy-2-butyl radical 4 (bottom path in Eq. (3))
offers yet another route to 2-butanone, but (as Fig. 2 summarizes)
the yield of acetone from TAME pyrolysis is at least twice as great
as that of 2-butanone. While this outcome is consistent with the
computational prediction (and in agreement with the lower exper-
imental appearance energy for formation of m/z 73 relative to m/z
87 from ionized TAME [59]), the facility with which the ketones
might undergo subsequent decomposition under the experimental
conditions stands in the way of a definitive confirmation.
In addition to the aforementioned rationalization of the low
abundance the expected ketones, other products warrant further
comment. TAME produces butenes, which might conceivably par-
allel the production of propene in the shock tube study of MTBE
[47]. On the one hand, as noted above, loss of methoxyl from TAME
to form (CH3)2CCH2CH3 radical 5 opens an avenue not available
in the case of MTBE: simple loss of a methyl radical from 5 to
produce (CH3)2C CH2. On the other hand, butenes also form in
the pyrolysis of 1 and 2, although to a lesser extent. Formation of
C4H8 from C5H10 could arise by methane loss via “roaming rad-
icals”, a possibility recently proposed for alkane pyrolyses [65].
Alternatively, the butenes from 1 and 2 may come from bimolecu-
lar reactions. One possibility supposes that addition of a hydrogen
atom to 1 or 2 followed by loss of a methyl radical yields butenes
(or that addition of a methyl radical to 2 followed by loss of an
ethyl radical yields isobutene, with the ethyl radicals then los-
ing a hydrogen atom quickly). These hypotheses remain to be
tested.
Acknowledgments
This work was supported by the University of California Energy
Institute and by NSF grants CHE-0848643 and CHE-0848517.
References
[1] T. Midgley Jr., Ind. Eng. Chem. 31 (1939) 504.
[2] T. Midgley Jr., T.A. Boyd, Ind. Eng. Chem. 14 (1922).
[3] A. Fischer, M. Muller, J. Klasmeier, Chemosphere 54 (2004) 689.
[4] X.L. Wang, M.A. Deshusses, Biodegradation 18 (2007) 37–50.
[5] R.M. Stephenson, J. Chem. Eng. Data 37 (1992) 80.
[6] M.J. Papachristos, J. Swithenbank, G.H. Priestman, S. Stournas, P. Polysis, E. Lois,
J. Inst. Energy 64 (1991) 113.
[7] L.K. Rihko-Struckmann, R.S. Karinene, A.O.I. Krause, K. Jakobsson, J.R. Aittamaa,
Chem. Eng. Process. 43 (2004) 57.
[8] J. Snelling, M.O. Barnett, D. Zha, J.S. Arey, Environ. Toxicol. Chem. 26 (2007)
2253.
[9] L.K. Rihko, A.O.I. Krause, Ind. Eng. Chem. 35 (1996) 2500.
[10] J. Liu, S. Wang, J.A. Guin, Fuel Process. Technol. 69 (2001) 205.
[11] D.E. Hendrickson, U.S. Patent Application W09516763.
[12] L.A. Smith, H.M. Putman, H.J. Semerak, C.S. Crossland, U.S. Patent 6,583,325.
[13] T. Evans, K.R. Edlund, U.S. Patent 2,010,356.
[14] J. Ignatius, H. Jaervelin, P. Lindqvist, Hydrocarbon Process. 74 (1995) 51.
[15] J. Bartling, M. Schloter, B.-M. Wilke, Biol. Fertil. Soils 46 (2010) 299.
[16] C.K. Westbrook, W.J. Pitz, Energy Technol. Rev. Feb/Mar (1991) 1.
[17] C.K. Westbrook, Chem. Ind. (UK) (1992) 562.
[18] S.D. Chambreau, J.S. Zhang, J.C. Traeger, T.H. Morton, Int. J. Mass Spectrom. 199
(2000) 17.
[19] K.H. Weber, J.S. Zhang, D. Borchardt, T.H. Morton, Int. J. Mass Spectrom. 249–250
(2006) 303.
[20] A. Heintz, S. Kapteina, S.P. Verevkin, J. Phys. Chem. B 111 (2007) 10975.
[21] A.M. Rozhnov, V.V. Safronov, S.P. Verevkin, K.G. Sharonov, V.I. Alenin, J. Chem.
Thermodyn. 23 (1991) 629.
[22] J.H.S. Green, Chem. Ind. (UK) (1960) 1215–1216.
[23] S.S. Chen, R.C. Wilhoit, B.J. Zwolinski, J. Phys. Chem. Ref. Data 6 (1977) 105.
[24] K.B. Wiberg, S. Hao, J. Org. Chem. 56 (1991) 5108.
[25] J.R. Durig, G.A. Guirgis, D.A.C. Compton, J. Phys. Chem. 84 (1980) 3547.
[26] D.W. Scott, G. Waddington, J.C. Smith, H.M. Huffman, J. Am. Chem. Soc. 71 (1949)
2767.
[27] Heat capacity for TAME up to 1000 K estimated as described in S.W. Benson,
Thermochemical Kinetics: Methods for the Estimation of Thermochemical Data
and Rate Parameters, 2nd ed., Wiley-Interscience, 1976.
[28] K.H. Anderson, S.W. Benson, J. Chem. Phys. 36 (1962) 2320.
[29] D.J. McKenney, K.J. Laidler, Can. J. Chem. 41 (1963) (1984).
[30] Y. Hidaka, K. Sato, M. Yamane, Combust. Flame 123 (2000) 1.
[31] R.D. Cook, D.F. Davidson, R.K. Hanson, J. Phys. Chem. A 113 (2009) 9974.

218
T.H. Morton et al. / International Journal of Mass Spectrometry 306 (2011) 210–218
[32] K. Yasunaga, T. Koike, H. Hoshikawa, Y. Hidaka, Proc. Combust. Inst. 31 (2007)
313.
[33] G.R. Freeman, C.J. Danby, C.N. Hinshelwood, Proc. R. Soc. A245 (1958) 28.
[34] C.J. Danby, G.R. Freeman, Proc. R. Soc. A245 (1958) 41.
[35] G.R. Freeman, Proc. R. Soc. A245 (1958) 75.
[36] K.J. Laidler, D.J. McKenney, Proc. R. Soc. A278 (1964) 505.
[37] A.T. Blades, G.W. Murphy, J. Am. Chem. Soc. 78 (1952) 1039.
[38] N.J. Daly, V.R. Stimson, Aust. J. Chem. 19 (1966) 239.
[39] N.J. Daly, C. Wentrup, Aust. J. Chem. 21 (1968) 1535.
[40] N.J. Daly, C. Wentrup, Aust. J. Chem. 21 (1968) 2711.
[41] K.Y. Choo, D.M. Golden, S.W. Benson, Int. J. Chem. Kinet. 6 (1974) 631.
[42] J.C. Brocard, F. Baronnet, Oxid. Commun. 1 (1980) 321.
[43] B. El Kadi, F. Baronnet, J. Chim. Phys. 92 (1995) 706.
[44] H. Bohm, F. Baronnet, B. El, Kadi, Phys. Chem. Chem. Phys. 2 (2000) (1929).
[45] A. Goldaniga, T. Faravelli, E. Ranzi, P. Dagaut, M. Cathonnet, Twenty-seventh
Symposium (International) on Combustion/The Combustion Institute, 1998,
pp. 353–360.
[46] C.S. McEnally, L.D. Pfefferle, Int. J. Chem. Kinet. 36 (2004) 345.
[47] K. Yasunaga, Y. Hidaka, A. Akitomo, T. Koike, 26th International Symposium on
Shock Waves, vol. 1, 2009, p. 177.
[48] T.C. Zhang, J. Wang, T. Yuan, X. Hong, L.D. Zhang, F. Qi, J. Phys. Chem. A 112
(2008) 10487.
[49] D.J. Grant, D.A. Dixon, J. Phys. Chem. A 113 (2009) 3656.
[50] A. Bodi, J.P. Kercher, C. Bond, P. Meteesatien, B. Sztaray, T. Baer, J. Phys. Chem.
A 110 (2006) 13425.
[53] D.W. Kohn, H. Clauberg, P. Chen, Rev. Sci. Instrum. 63 (1992) 4003.
[54] A.V. Friderichsen, J.G. Radziszewski, M.R. Nimios, P.R. Winter, D.C. Dayton, D.E.
David, G.B. Ellison, J. Am. Chem. Soc. 123 (2001) (1977).
[55] H. Clauberg, D.W. Minsek, P. Chen, J. Am. Chem. Soc. 114 (1992) 99.
[56] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheese-
man, J.A. Montgomery, Jr., T. Vreven, K.N. Kudin, J.C. Burant, J.M. Millam, S.S.
Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G.A.
Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J Hasegawa, M.
Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J.E. Knox, H.P.
Hratchian, J.B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R.E. Strat-
mann, O. Yazyev, A.J. Austin, R.; Cammi, C. Pomelli, J.W. Ochterski, P.Y. Ayala, K.
Morokuma, G.A. Voth, P. Salvador, J.J. Dannenberg, V.G. Zakrzewski, S. Dapprich,
A.D. Daniels, M.C. Strain, O. Farkas, D.K. Malick, A.D. Rabuck, K. Raghavachari, J.B.
Foresman, J.V. Ortiz, Q. Cui, A.G. Baboul, S. Clifford, J. Cioslowski, B.B. Stefanov,
G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R.L. Martin, D.J. Fox, T. Keith, M.A.
Al-Laham, C.Y. Peng, A. Nanayakkara, M. Challacombe, P.M.W. Gill, B. Johnson,
W. Chen, M.W. Wong, C. Gonzalez, J.A. Pople, J.A. Gaussian 03, Revision C.02,
Gaussian, Inc., Wallingford CT, 2004.
[57] A.D. Becke, J. Chem. Phys. 109 (2001) 9287.
[58] T.C. Zhang, L.D. Zhang, J. Wang, T. Yuan, X. Hong, F. Qi, J. Phys. Chem. A 112
(2008) 10495.
[59] J.C. Traeger, T.H. Morton, J. Phys. Chem. A 109 (2005) 10467.
[60] W. Tao, R.B. Klemm, F.L. Nesbitt, J.L. Stief, J. Phys. Chem. 96 (1992) 104.
[61] A. Vasiliou, M.R. Nimlos, J.W. Daily, G.B. Ellison, J. Phys. Chem. A 113 (2009)
8540.
[51] E.A. Fogleman, H. Koizumi, J.P. Kercher, B. Sztaray, T. Baer, J. Phys. Chem. A 108
(2004) 5288.
[52] W.R. Roth, F. Bauer, A. Beitat, T. Ebbrecht, M. Wustefeld, Chem. Ber. 124 (1991)
1453.
[62] K. Sato, Y. Hidaka, Combust. Flame 122 (2000) 291.
[63] J. Ernst, K. Spindler, H.G. Wagner, Ber. Bunsen-Gesell. 80 (1976) 645–650.
[64] S. Saxena, J.H. Kiefer, S.J. Klippenstein, Proc. Combust. Inst. 32 (2009) 123.
[65] L.B. Harding, S.J. Klippenstein, J. Phys. Chem. Lett. 1 (2010) 3016.