Synchrotron Photoionization Study of the Diisopropyl Ether Oxidation

Article abstract of DOI:10.1002/cphc.201901134

Scientific evidence has shown oxygenates help to reduce dangerous pollutants arising from burning fossil fuel in the automotive sector. For this reason, their use as additives has spread widely. The aim of this work consists in providing a comprehensive identification of the main primary oxidation products of diisopropyl ether (DIPE), one of the most promising among etheric oxygenates. The Cl-initiated oxidation of DIPE is examinated by using a vacuum ultraviolet (VUV) synchrotron radiation at the Advanced Light Source (ALS) of the Lawrence Berkeley National Laboratory (LBNL). Products are identified on the basis of their mass-to-charge ratio, shape of photoionization spectra, adiabatic ionization energies, and chemical kinetic profiles, at three different temperatures (298, 550, and 650 K). Acetone, propanal, propene, and isopropyl acetate have been identified as major reaction products. Acetone is the main primary product. Theoretical calculations using the composite CBS-QB3 method provided useful tools to validate the postulated reaction mechanisms leading to experimentally observed species. The formation of other species is also discussed.

Full text of DOI:10.1002/cphc.201901134

Accepted Article
Title: Synchrotron Photoionization Study of the Diisopropyl Ether
Oxidation
Authors: Andrea Giustini and Giovanni Meloni
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemPhysChem 10.1002/cphc.201901134
Link to VoR: http://dx.doi.org/10.1002/cphc.201901134
A Journal of
www.chemphyschem.org

10.1002/cphc.201901134
ChemPhysChem
Synchrotron Photoionization Study of the Diisopropyl Ether
Oxidation
Andrea Giustini
Giovanni Meloni
February 11, 2020
1 Abstract
2 Introduction
Air pollution arising from automotive emissions is to
date a point of major concern in terms of health
hazards and environmental consequences. Indeed,
other than carbon dioxide (CO2) and water vapor
(H2O), the combustion process can lead to the for-
mation of many species, including carbon monox-
ide (CO), uncombusted hydrocarbons (HCx), nitro-
gen oxides (NOx), and volatile organic compounds
(VOCs). CO and HCx are obtained when internal
combustion engines run during a rich regime, hence
a lower amount of oxygen is available leading to in-
complete combustion. NOx formation is very rapid
at high temperature, especially when air-fuel equiva-
lence ratio goes rich.[1] VOCs are a mixture of hydro-
carbons and aromatic compounds, with the latter be-
ing carcinogenic and mutagenic according to scientific
evidences, thus with serious health effects on human
life.[2, 3] They also have been shown to be able to en-
hance the tropospheric ozone levels during summer-
time when the sunlight is the highest and the proba-
bility of photochemical-ozone-precursor production is
higher.[4] Great effort has been made to reduce pol-
lutants deriving from fossil fuel consumption. The
elimination of lead and aromatic molecules as anti-
knock compounds (octane-enhancer) from gasoline
has been the first step to meet this goal, but it drove
the need to find valid alternative options to replace
the lost octane number.[1]
Scientific evidence has shown oxygenates help to re-
duce dangerous pollutants arising from burning fossil
fuel in the automotive sector. For this reason, their
use as additives has spread widely. The aim of this
work consists in providing a comprehensive identi-
fication of the main primary oxidation products of
diisopropyl ether (DIPE), one of the most promis-
ing among etheric oxygenates. The Cl-initiated ox-
idation of DIPE is examinated by using a vacuum
ultraviolet (VUV) synchrotron radiation at the Ad-
vanced Light Source (ALS) of the Lawrence Berkeley
National Laboratory (LBNL). Products are identi-
fied on the basis of their mass-to-charge ratio, shape
of photoionization spectra, adiabatic ionization en-
ergies, and chemical kinetic profiles, at three differ-
ent temperatures (298, 550, and 650 K). Acetone,
propanal, propene, and isopropyl acetate have been
identified as major reaction products. Acetone is the
main primary product. Theoretical calculations us-
ing the composite CBS-QB3 method provided useful
tools to validate the postulated reaction mechanisms
leading to experimentally observed species. The for-
mation of other species is also discussed.
A. Giustini and Prof. G. Meloni, Department of Physi-
cal and Chemical Sciences, University of L’Aquila, L’Aquila
(Italy)
Prof. G. Meloni, Department of Chemistry, University of
San Francisco, San Francisco, California 94117 (United States)
corresponding author
Oxygenates matched these requirements and over
the last three decades they have started to be
added to gasoline in order to get blends provid-
ing for cleaner-burning fuel.[5] The presence of oxy-
1
This article is protected by copyright. All rights reserved.

10.1002/cphc.201901134
ChemPhysChem
gen, indeed, guarantees a complete combustion re- mechanisms leading to them.
sulting into lower CO, HCx, and toxic exhausts,
but slightly higher NOx release, besides improving
engine performance.[6] No particular benefits have
been recorded in terms of reducing greenhouse gas
emissions. The replacement of aromatic compo-
nents with oxygenates causes the fuel-burning pro-
cess to be more energy-efficient, resulting into CO2
increase.[5] Among oxygenates, diisopropyl ether
(hereafter: DIPE) has been recognized as one of the
best fuel additives due to its excellent capability to
blend with gasoline.[7] DIPE also exhibits a much
lower water solubility and volatility, as well as a lack
of toxicity with respect to methyl tert-butyl ether
(MTBE), the most used oxygen-containing additive
in the United States until 2000.[8–11] Such proper-
ties are essential to prevent the dispersion in the en-
vironment, which could cause groundwater contami-
nation with dangerous effect on both human health
and ecosystem in general. In order to fully assess
DIPE environmental impact, an in-depth evaluation
of its combustion features is needed before endorsing
its effective usage as alternative oxygenate.[12] DIPE
is initially obtained by hydration of propylene into
This study aims to provide information about the
oxidation of DIPE relevant for determining its envi-
ronmental impact. Characterization of primary prod-
ucts was performed by investigating the Cl-initiated
oxidation process at three different temperatures. In
addition, electronic structure calculations were car-
ried out to explore the feasible reaction pathways
leading to products formation, giving key insights on
the explanation of experimental data sets.
3 Experimental Section
Experiments were performed at the Chemical Dy-
namics beamline of the Advanced Light Source (ALS)
situated at the Lawrence Berkeley National Labora-
tory (California, USA). A tunable synchrotron ra-
diation (7.2 - 25 eV) is used coupled with a mul-
tiplexed (all species detected simultaneously) time-
and energy-resolved mass spectrometer to probe reac-
tion species. Further details about the experimental
apparatus are discussed in previous works.[16–19]
The vapors of DIPE (Sigma-Aldrich, 99.0%) are
isopropanol, followed by addition of another propy- collected in a gas cylinder to reach an overall con-
lene molecule to yield the desirable ether. A par- centration of about 1.0% in excess of helium. In this
allel competitive dehydrative reaction of isopropanol work, DIPE reacts with O2 in the presence of the
can lead to the formation of DIPE. Both the etheri- radical precursor Cl2 at room temperature, 550, and
fication and hydration reactions are exothermic and 650 K and a pressure of 4, 7, and 7 Torr, respectively.
reversible.[7, 13] In recent years, researchers investi- Gaseous reactants are directed into a 62 cm long slow-
gated the oxidation of DIPE by focusing on several flow quartz reaction tube and regulated by means of
reaction pathways. Collins and co-workers[14] stud- calibrated mass flow controllers. The temperature of
ied the OH radical-initiated oxidation in the presence the reactor, wrapped by a 18 µm thick nichrome heat-
and absence of NOx. They found various products, ing tape, can be adjusted according to achieve the
such as isopropyl acetate, formaldehyde, and ace- desired setup. Additionally, in order to avoid tem-
tone, which are formed through reactions based on perature fluctuations, the reactor is surrounded with
C-C and C-O bond fission, and isomerisation mecha- an insulating square-weave, yttria-stabilized zirconia
nisms. In Wang et al. work[15], quantum chemistry (ZYW-15, Zircar Zirconia, Inc.) cloth sorrounded
calculations are carried out to evaluate atmospheric by two-halves of a gold-plated copper sheath. Fur-
oxidation of three different ethers, including DIPE. ther, a feedback controlled throttle valve connects a
They focused onto hydrogen abstraction by OH radi- Roots pump to the reaction tube so that the reac-
cals with and without NOx. Similarities with Collins’ tor pressure is allowed to be changed and maintained
work[14] have been reported, confirming isopropyl ac- throughout the experiments. The gas mixture is pho-
etate as the major product in the presence of NOx tolyzed by a 4 Hz-pulsed unfocused 351 nm (XeF) ex-
and acetone when NOx are missing. Discrepancies cimer laser, propagating collinearly down the quartz
have been reported just for the proposed reaction tube. Chlorine atoms are formed according to the
2
This article is protected by copyright. All rights reserved.

10.1002/cphc.201901134
ChemPhysChem
following equation:
Cl2
gion. In order to extract valuable observables, data
sets are sorted out into a 3D data block in which
the dimensions are referred to as mass-to-charge ra-
tio, time, and photon energy variables. Data analysis
consists of slicing 3D data into 2D images and further
into 1D plots to yield photoionization spectra and
kinetic traces used for qualitative and quantitative
studies. PI curves, for instance, are used to identify
species and to distinguish isomers from each other,
since their onset, shape, and intensity are strictly
correlated to ionization thresholds, Franck-Condon
factors, as well as concentration of chemicals. Thus,
by comparing literature, measured, or calculated PI
curves with experimental data, isomers identification
can be achieved. Adiabatic ionization energy (here-
after: AIE) of species is established by linear extrap-
olation of the initial onset of the photoionization sig-
nal as explained elsewhere.[23] Taking into account
the photon energy stepsize used for data acquisition
and possible existence of hot bands, an overall er-
ror of 0.05 eV is estimated for reported AIEs values
determined in this work.
hν
2Cl
The IUPAC Subcommittee for Gas Kinetic Data
Evaluation recommended 1.00 as the quantum yield
of Cl2 photolysis.[20] On the basis of the absorption
cross section value at 351 nm that is 1.82 x 10−19 cm2
according to Maric et al.[21], the recommended quan-
tum yield, and the concentration of Cl2 molecules,
which have been calculated to be 1.29
x
1013
molecules cm−3 at room temperature, 2.41 x 1013
molecules cm−3 at 550 K, and 2.04 x 1013 molecules
cm−3 at 650 K, the number density for Cl atoms is
8.33 x 1011 to 1.32 x 1012 molecules cm−3. Number
densities for DIPE are 2.37 x 1013, 2.20 x 1013, and
2.83 x 1013 molecules cm−3 at 298, 550, and 650 K,
respectively; number densities for O2 are 1.29 x 1016,
1.20 x 1015, and 1.03 x 1016 molecules cm−3 at 298,
550, and 650 K, respectively. The reaction species
effuse through a 650 µm wide pinhole located on the
side of the reactor, then are skimmed into the dif-
ferentially vacuumed ionization region of the instru-
ment. Once ionized, they are accellerated via a 50
kHz pulsed time-of-flight mass spectrometer with a
current mass resolution of 1600 and detected.[22] The
photoionization spectrum (PI) of isopropyl acetate
(Sigma-Aldrich, 99%) is recorded as well for data
analysis. A calibration experiment is performed by
flowing known amounts of DIPE and a calibration gas
mixture composed by propene, 1-butene, and ethene,
without photolysis. Authentic spectra are taken to
check the photoionization behaviour of DIPE.
4 Computational methods
Electronic structure calculations are performed to
compute adiabatic ionization energies when un-
known. All calculations are carried out by using
Gaussian 09 program[24] and the CBS-QB3 compos-
ite method.[25, 26] This model has a reported en-
ergy mean average deviation of 4-5 kJ mol−1 given
by Montgomery et al.[26] AIEs are yielded by the
difference between the zero-point corrected total elec-
tronic energies (ZPE) of the ground electronic states
of the neutral and of the cationic molecules. ZPE
are also used to derive enthalpies of reaction (4rHo)
useful to demonstrate that a proposed reaction path
is thermodynamically feasible.
Multiplexed time- and energy-resolved mass spec-
trometry experiments are developed to scan simulta-
neously all the species at different photon energies.
A tunable synchrotron radiation is used in a quasi-
continous fashion as ions provider starting before the
trigger time, which corresponds to the time when the
photolysis laser is fired and the radical-generating
precursor is decomposed. The photon energy is var-
ied from 9 to 11 eV with a 0.025 eV stepsize. At each
photon energy, the ion signal is background (prepho-
tolysis signal) subtracted, and subsequently normal-
ized to the ALS photocurrent, which is measured by
In addition, the appearance energies (AE) of
daughter ions are expressed as follows, if a dissoci-
ation barrier is not present:
AE = AIE + BDE
(1)
where BDE is the bond dissociation energy, i.e.,
a calibrated photodiode located in the ionization re- the difference beetween the zero-point energy of the
3
This article is protected by copyright. All rights reserved.

10.1002/cphc.201901134
ChemPhysChem
fragments and the parent ion ground state. AEs species are not appearing in the performed experi-
are coupled with thermodynamic values to provide ments. The reaction products are investigated after
useful information regarding enthalpies of formation subtracting the contribution of prephotolysis signals
(4f Ho) of species involved in photodissociative re- (background).
actions. In fact, since AE is the enthalpy of reaction
at 0 K experimentally-derived, and if the other en-
thalpies of formation are well-known, the third one
can be determined as a result of the following equa-
tion:
298 K
4rH0K = 4f H0oK(A+) + 4f H0oK(B∗)
(2)
− 4f H0oK(AB)
where AB is the generic neutral molecule, A+
its cationic fragment, and B∗ its radical fragment.
Relaxed potential energy surface (PES) scans were
carried out using the B3LYP/6-311G∗ level of the-
30
40
50
60
70
80
90
100
110
120
130
m/z
ory in order to locate possible first order saddle
points (transition states) along the PES, and then
reoptimized with the CBS-QB3 composite method.
The transition states are verified through intrin-
sic reaction coordinate (IRC) calculations[27] to en-
sure that the forward and reverse mechanisms fall
down into PES minima. If the photoionization spec-
trum of a reaction species is not found in the lit-
erature, it can be measured or, where appropri-
ate, it can be simulated by integrating the photo-
electron spectrum generated by approximating the
Franck-Condon factors for the vibronic transition
from the neutral to the cationic vibrational states of
the molecule. These simulations are performed us-
ing Franck-Condon (FC)[28–31] and Franck-Condon-
Herzberg-Teller methods (FCHT).[28]
550 K
40
50
60
70
80
90
100
110
120
130
m/z
650 K
5 Results and Discussion
Atmospheric oxidation of oxygenates tipically begins
with a reaction involving OH-radicals, as a result of
photolysis reactions due to sunlight. In this work,
atomic chlorine is replacing the atmospheric radi-
cals to initiate the hydrogen-abstraction from DIPE.
Typically, chlorinated species are obtained from Cl-
addition to unsaturated primary products and they
are identified on the basis of isotopic ratio for 37Cl
and 35Cl. However, peaks referring to chlorinated
40
50
60
70
80
90
100
110
120
130
m/z
Figure 1: Background subtracted mass spectra for Cl-
initiated oxidation of DIPE at 298 K, 550 K, 650 K, over the
photon energy range of 9.0 to 11.0 eV and over reaction time
range 0 - 150 ms.
4
This article is protected by copyright. All rights reserved.

10.1002/cphc.201901134
ChemPhysChem
Mass spectra in Figure 1 reveal negative (reactants species i at the photon energy E, k is the instrumen-
signal) and positive (products) peaks. The prominent tal costant, Si(E) is the ion signal at the specified
negative-peak at m/z = 87 corresponds to the main photon energy, and δi is the mass discrimination fac-
fragment of the photodissociation process of DIPE. tor that stands for the mass-dependent response of
The presence of the molecular ion at m/z = 102 also the instrument, here approximately equivalent to the
reveals a distinction between DIPE and other ethers, mass of the species i to the power of 0.67.[35] From
such as MTBE and ETBE, which are characterized the ratio of the signal of the species (P) for which we
by the absence of parent ion signal.[32] Additional want to determine the photoionization cross-section
daughter ions are recorded at m/z = 69 and m/z = to the signal of the reference species (R), we can de-
45, but they have much lower intensity than m/z = rive the desired quantity, as follows:
87 (3 and 0.1 %, respectively). There are not any
published experimental photoelectron and authen-
SP (E) CR δR
tic photoionization spectra of DIPE for comparison.
Williams et al.[33] assigned an ionization energy of
9.16 eV for DIPE, whereas Watanabe et al.[34] gave
a value of 9.20 eV, which are in agreement with the
experimentally-derived value of 9.10 ± 0.05 eV from
this work (see Figure 2).
σP (E) =
σR(E)
SR(E) CP δP
ꢀ
(4)
ꢁ
0.67
SP (E) CR mR
SR(E) CP mP
=
σR(E)
Figure 2 shows the total absolute photoioniza-
tion spectrum of diisopropyl ether obtained by sum-
ming up DIPE and its most abundant fragment rel-
ative cross-sections, estimated by using the above-
mentioned method.
28
Authentic PI curve of DIPE
Authentic PI curve of m/z = 87
Total PI spectrum of DIPE
24
20
16
12
At m/z = 42 a signal grows as the temperature
is increased from 550 K up to 650 K (see Figure 3).
The experimental onset is measured to be 9.70 ± 0.05
eV, which is in line with the value of 9.73 ± 0.02
eV for propene as observed by Traeger.[36] In order
to characterize this mass, propene PI spectrum by
Cool[37] has been superimposed onto the experimen-
tal plot. The first part is in very good agreement with
it, whereas the second part starting from roughly
10.1 eV is not assigned and it may derive from disso-
ciatively photoionizing secondary products. Its time
trace is slower than acetone time trace, suggesting it
arises from secondary chemistry reactions (see Figure
3c).At m/z = 43, a signal is detected in the experi-
ments at 550 and 650 K. It was thought to arise from
photoionization of isopropyl radicals but the experi-
mental PI spectrum illustrated in Figure 4a shows an
onset at around 9.9 eV, which is quite far from liter-
ature value of 7.37 ± 0.02 eV[38], suggesting that it
is clearly a fragment resulting from a dissociatively
photoionizing species. In addition, m/z = 43 kinetic
trace in Figure 4b exhibits a fast formation followed
8
4
9.0
9.5
10.0
10.5
11.0
Photon Energy (eV)
Figure 2: Authentic experimental photoionization spectrum
of diisopropyl ether.
When photoionization cross-sections of a species
are unknown, we used the absolute photoionization
spectrum of a reference species to compare with
the PI experimental spectrum of the species with a
known concentration in order to estimate its σE. This
estimation comes from the equation defining the ion
signal of a species i at the photon energy E:
Si(E) = kσi(E)δiCi
(3)
where Ci is the concentration of the species i, σi(E) by a stable signal, implying the probable involvement
represents the photoionization cross-section of the of a primary chemistry mechanism.
5
This article is protected by copyright. All rights reserved.

10.1002/cphc.201901134
ChemPhysChem
m/z = 42 Experimental PI Spectrum
Authentic PIE of propene (Cool et al.)
9.0
9.5
10.0
10.5
11.0
9.0
9.5
10.0
10.5
11.0
Photon Energy (eV)
Photon Energy (eV)
(a)
(a)
m/z = 42 Experimental PI Spectrum
Authentic PIE of propene (Cool et al.)
9.0
9.5
10.0
10.5
11.0
0
10
20
30
Photon Energy (eV)
Reaction Time (ms)
(b)
(b)
Figure 4: (a) Experimental photoionization plot of m/z = 43
fragment recorded at 550 K; (b) Experimental kinetic trace of
m/z = 43 from 0 (trigger time) to 30 ms at 550 K as well.
m/z = 58
m/z = 42
At m/z = 58 a peak is recorded in all experiments,
as shown in Figure 1. Figure 5 shows the experi-
mental photoionization plots of m/z = 58 product at
different temperatures along with the PI curves taken
from the literature with the aim of characterizing an
overall isomeric composition. The experimental on-
set is 9.67 ± 0.05 eV, which is in agreement with
the literature AIE measurement of 9.694 ± 0.006 eV
for acetone.[39] In addition, acetone PI spectrum by
Cool et al.[37] perfectly matches the very first part
of the experimental PI curve, which confirms this
assignment. A second feature is found at 9.96 eV,
where the experimental signal raises deviating from
acetone PI curve. This can be ascribed to the pres-
-10
0
10
20
30
Reaction Time (ms)
(c)
Figure 3: PI curves of propene[37] (blue line) superim-
posed onto the experimental photoionization plot of m/z =
42 recorded at (a) 550 K and (b) 650 K.(c) Time trace of m/z
= 42 compared to the time trace of m/z = 58 at 650K.
6
This article is protected by copyright. All rights reserved.

10.1002/cphc.201901134
ChemPhysChem
ence of propanal. Its AIE literature determination is
m/z = 58 Experimental PI Spectrum
Authentic PI of acetone (Cool et al.)
Authentic PI of propanal (Wang et al.)
All m/z = 58 isomers
9.96 ± 0.01 eV from Linstrom[40], consistent with the
experimental observed spectral shoulder. The sum-
mation of the PI literature curves of acetone[37] and
propanal[41] is in accordance with the experimental
plot of m/z = 58 at all three temperatures. The
relative intensities of the two isomers vary at differ-
ent temperatures, therefore, revealing an effect of the
temperature on the observed product yields.
m/z = 61 is detected at room temperature, while
it almost disappears at 550 and 650 K. Since this is
an odd mass, there are only two feasible options to
be considered for establishing its identity: it can be
either attributed to a radical species, which ionizes to
give the m/z = 61 cation, or it can be ascribed to a
cationic fragment coming from dissociative photoion-
ization of a larger species. Since the kinetic trace of
this mass (see Figure 6b) does not show a usual radi-
cal behaviour, it can be solely assigned to a dissocia-
tive photoionization fragment.
Furthermore, m/z = 61 is not very common in mass
spectrometry and a limited number of chemical com-
pounds can be used to explain its formation. By tak-
ing into account the chemical structure of DIPE, iso-
propyl acetate (hereafter: IPAC) has been proposed.
Benoit et al.[42] found m/z = 61 to derive from frag-
mentation of isopropyl acetate (m/z = 102) molecular
ion, whose intensity is extremely low due to its cation
ground state instability. They also reported the ap-
pearance energy value of 9.96 ± 0.05 eV, consistent
with the experimental value of this work, which is
10.03 ± 0.05 eV. An absolute photoionization spec-
trum has been recorded as well confirming the pres-
ence of IPAC, as illustrated in Figure 7.
The occurence of IPAC has been found out in pre-
vious studies, where it is clearly shown as being the
major product of DIPE photoxidation.[14, 15, 43]
Neverthless, experimental conditions were substan-
tially different owing to high concentration of NO
and HO2 radicals. There is solely one way to yield
IPAC starting from DIPE in the absence of either NO
and HO2, which involves self-reaction of alkylperoxy
radicals (ROO∗) to form alkylalkoxy radicals (RO∗),
followed by an α-β-cleavage (C-C bond fission) result-
ing into IPAC together with a methyl group release.
Several studies[44–50] supported a concerted Russell
9.0
9.0
9.0
9.5
10.0
10.5
10.5
10.5
11.0
11.0
11.0
Photon Energy (eV)
(a)
m/z = 58 experimental PI Spectrum
Authentic PI of acetone (Cool et al.)
Authentic PI of propanal (Wang et al.)
All m/z = 58 isomers
9.5
10.0
Photon Energy (eV)
(b)
m/z = 58 Experimental PI Spectrum
Authentic PI of acetone (Cool et al.)
Authentic PI of propanal (Wang et al.)
All m/z = 58 isomers
9.5
10.0
Photon Energy (eV)
(c)
Figure 5: PI curves of acetone[37] (blue line) and propanal[41]
(green line) superimposed onto the experimental photoioniza-
tion plot of m/z = 58 product recorded at (a) 298 K, (b) 550
K , and (c) 650 K.
7
This article is protected by copyright. All rights reserved.

10.1002/cphc.201901134
ChemPhysChem
9.0
9.5
10.0
10.5
11.0
Photon Energy (eV)
(a)
Figure 7: Authentic experimental photoionization spectrum
of isopropyl acetate.
topic abundance of m/z = 58, which is expected to
be 3.3 % (13C). Plus, the kinetic trace of m/z = 59
appears totally different from mass 58 (see Figure
8), implying that this signal corresponds to another
species.
m/z = 58
m/z = 59
0
10
20
30
Reaction Time (ms)
(b)
Figure 6: (a) Experimental photoionization plot of m/z =
61 fragment recorded at room temperature; (b) Experimental
kinetic trace of m/z = 61 from 0 (trigger time) to 30 ms at
room temperature as well.
0
10
20
30
arrangement[51, 52] passing through a tetraoxide in-
termediate for secondary ROO∗ or a cage-mechanism
for tertiary ROO∗, where few RO∗ escape the cage
along with O2 and, then, they undergo C-C bond
fission.[53, 54] m/z = 61 kinetic trace, which is equal
to the parent ion kinetic trace, suggests a fast forma-
tion consistent with a Russell mechanism rather than
a cage mechanism. However, since the characteriza-
tion of secondary products lies beyond the purpose
of this study, we cannot neither confirm nor dismiss
any mechanisms leading to the formation of IPAC.
Reaction Time (ms)
Figure 8: Experimental kinetic traces of m/z = 58 (red line)
and 59 (blue line) from 0 to 30 ms at room temperature.
Calculations have been performed to establish if it
comes from ionization of a radical or from fragmen-
tation of a dissociatevely photoionizing compound.
It was thought to derive from C-O bond fission
of alkylalkoxy radicals (RO∗) yielding a 1-methyl
ethoxy radical along with acetone. The correspond-
A peak at m/z = 59 is recorded only at room tem- ing Gibbs activation energy for this path is calculated
perature. The intensity does not match the real iso- to be 87 kJ mol−1 at room temperature, while the
8
This article is protected by copyright. All rights reserved.

10.1002/cphc.201901134
ChemPhysChem
Gibbs reaction energy is -19 kJ mol−1. In addition, radicals (ROR∗) as illustrated in Scheme 1. The first
William and Hamill[33] found m/z = 59 ascribed to sites are the α-carbons on both sides of the molecule,
1-methyl ethoxy radical generated by passing DIPE which give isopropoxy-2-isopropyl radical (hereafter:
gas through a platinum wire ”cracker”. The relative α-IPOXI) along with HCl. The CBS-QB3-calculated
AIE was 9.20 ± 0.05 eV, although we measure an enthalpy change of this process is - 43 kJ mol−1
.
onset at 9.60 ± 0.05 eV. Despite the computational The second equivalent sites where chlorine can ab-
efforts, no satisfactory results have been achieved in stract a hydrogen are the four β-carbons, from which
this direction, thus we ascribe this species to a frag- isopropoxy-1-isopropyl radical (hereafter: β-IPOXI)
ment of RO∗ photodissociation resulting into m/z = along with HCl is generated with a calculated reac-
59 cation along with m/z = 58. Taking into account tion enthalpy of -10 kJ mol−1. α-IPOXI production
the instability of this radical, we can justify the m/z results to be the most exothermic because α-carbon
= 59 kinetic trace as coming from RO∗ fast formation is a secondary carbon stabilized by inductive effect
and depletion.
from methyl groups. Indeed, Wallington and co-
At room temperature a low intensity peak ap- workers[43] found that secondary carbons are more
pears at m/z = 100, gradually disappearing as the reactive than primary carbons (80 % secondary vs.
temperature increases. Unfortunately, there are not 20 % primary) at room temperature, with a reac-
available PI spectra from literature to character- tivity increase of primary carbon sites expected to
ize m/z = 100. According to postulated mecha- occur at higher temperatures. This is in agreement
nisms (see below), we performed PI spectra simula- with previous studies[55, 56], where the vicinity of
tions of feasible compounds having m/z = 100, i.e., the oxygen has shown to weaken the C-H bond mak-
2,5-dimethyltetrahydrofuran (hereafter: 25DMTHF) ing the abstraction at this site more favorable than at
and isopropyl isopropenyl ether (hereafter: IIE). The the primary-carbon site. Further, Scheme 1 shows O2
25DMTHF simulated PI spectrum matches the onset addition to the radical site yielding alkoxyalkylper-
of the experimental m/z = 100 spectrum, whereas the oxy radicals (ROR0 OO∗): α-ROO (A) and β-ROO
IIE simulated PI spectrum does not match any fea- (B). These species can undergo intramolecular hydro-
ture of the experimental data (see Figure 9).
gen abstraction with the consequent isomerization to
hydroperoxyalkoxyalkyl radicals (hereafter: QOOH).
Then, labile QOOH radicals can undergo unimolec-
ular dissociation to form several products, which
are not all experimentally-observed in this study, by
means of α-β cleavages and cyclization processes as
detailed in the following section.
Experimental PI spectrum of m/z = 100
2-5,DMTHF PI simulation
IIE PI simulation
α-IPOXI Reaction Pathway. α-IPOXI radical
reacts with O2 to give α-peroxy radical (α-ROO),
with a calculated reaction enthalpy equal to -158
kJ mol−1. α-ROO, then, undergoes a unimolecular
process, which consists of a hydrogen shift from the
α- or β-carbons of the isopropyl side to the oxygen
of the other side to yield the resulting α-QOOH
(A1) and β-QOOH (A2) as shown in Scheme 2A.
H-shift isomerization, as described above, has been
found to be important in the oxidation process under
low-NO and low-HO2 conditions[15], which are fairly
9.0
9.5
10.0
10.5
11.0
Photon Energy (eV)
Figure 9: PI spectra simulations of 25DMTHF (blue line)
and IIE (green line) superimposed onto the experimental pho-
toionization plot of m/z = 100.
Postulated Mechanisms. Two distinct possible replicated in this study because of the total absence
sites from DIPE are expected to undergo a hydrogen of NO. The corresponding activation enthalpies for
abstraction by atomic chlorine radicals to form alkyl these reactions are 68 and 95 kJ mol−1 , whereas the
9
This article is protected by copyright. All rights reserved.

10.1002/cphc.201901134
ChemPhysChem
O
O
O
O
+O2
Cl
Cl
-HCl
-HCl
a-IPOXI (-43 kJ/mol)
A (-158 kJ/mol)
O
O
O
DIPE
O
O
+O2
ß-IPOXI (-10 kJ/mol)
B (-145 kJ/mol)
Scheme 1: Schematics of reaction pathways referring to Cl-initiated H-abstraction and subsequent O2 addition to form
alkoxyalkylperoxy radicals (ROROO∗). The CBS-QB3 calculated enthalpy of H-abstraction reactions are reported in paren-
theses
reaction enthalpies are 48 and 92 kJ mol−1, which activation barrier relative to A, this substituted diox-
are endothermic using A as the zero energy level, olane is not observed in our experiment. A fourth
as well as far below the thermochemical threshold route to decomposition for α-ROO radical species
(red line in Schemes 2A and 2B) allocated to the is the hydrogen abstraction on a CH3-group of the
α-IPOXI + O2 energy level. Everything above same isopropyl side to obtain isopropyl isopropenyl
the thermochemical threshold is considered to be ether and HO2, which was shown to be energetically
thermodynamically unfavourable. α-QOOH (A1) favourable with an activation enthalpy of 122 kJ
may decompose through a C-O bond fission leading mol−1. It is also predicted to arise via a concerted
to the formation of two molecules of acetone along mechanism without involving any QOOHs isomer. A
with OH. The driving path from A1 to acetone is signal at m/z = 100 is detected in the experiments
calculated to be exothermic with an energy release at room temperature. Several studies[57–59] found
of -144 kJ mol−1. β-QOOH (A2) may decompose unsaturated products formation by direct HO2 elim-
through a C-C bond fission leading to acetone and ination from alkyl radicals reactions with O2. Kaiser
propene along with HO2 radical, which is calculated et al.[59], by investigating reactions of ethyl radicals
to stand above the thermochemical threshold, thus with O2, Cl2, and Cl, considered two different
not feasible with respect to energetics. However, mechanisms yielding ethene: the first one deriving
A2 can undergo a cyclization path, where the from ethyl radicals with O2 reactions, most problable
terminal methyl radical joins to the peroxy oxygen at low pressure, and low Cl concentrations; whereas
on the other side, yielding the five-membered ring the second one coming from secondary reactions
compound 2,2,4-trimethyl-1,3-dioxolane (m/z = 116, of long-lifetime Cl atoms with ethyl radicals, most
C6H12O2) together with a hydroxyl radical. The probable at higher Cl2 concentrations and higher
activation energy and the released reaction energy pressure. Unfortunately, the IIE PI spectrum simu-
for this step are calculated to be 41 kJ mol−1 and lation simulation did not allow us to reproduce the
-192 kJ mol−1, respectively. Because of the high m/z = 100 experimental plot, which is conversely in
10
This article is protected by copyright. All rights reserved.

10.1002/cphc.201901134
ChemPhysChem
agreement with 2,5-DMTHF PI spectrum simulation (B3) is formed. Here, the carbon binds to the per-
(see Figure 9). Since it was not possible to find oxydic oxygen giving an oxyrane: isopropoxy methy-
a thermodynamically-favourable pathway yielding loxirane (m/z = 116). Reaction energy is estimated
2,5-DMTHF, this mass is considered a product from to be -72 kJ mol−1 with respect to B3, with an ac-
secondary reactions. As the temperature increases, tivation energy of 12 kJ mol−1. No corresponding
m/z = 100 progressively disappears and the m/z = observed signal for this species is due to its unbound
42 signal, assigned to propene, comes up. Propene cation dissociating into the isopropyl cation (m/z =
formation passes through chain-terminating reac- 43) and neutral methylene acetate (m/z = 73). By
tions of isopropyl radicals with oxygen also yielding using Equation 2, therefore, without considering a re-
HO2[60] and by β-IPOXI radicals decomposition. verse barrier for dissociation (thermochemical limit),
Isopropyl radicals come from the decomposition of the calculated AE for m/z = 43 is 9.85 eV, which is
α-IPOXI radicals, which is increasingly expected in perfect agreement with the measured AE for m/z
to occur as the temperature grows, since decom- = 43 of 9.85 ± 0.05 eV. We want to point out that
positions are endothermic reactions. Below the this assignment is only tentative since we cannot dis-
above-mentioned reactions are illustrated:
card other species that may dissociatevely photoion-
ize yielding m/z = 43. Finally, in the case of hydrogen
removal from the β-carbon on the same side, a α-
QOOH-α (B4) is created. The cyclization path leads
to isopropoxy oxetane (m/z = 116), which beacause
of extremely low Franck-Condon factors due to the
opening of the ring in the cation and high activation
barrier cannot be observed. The reaction energy is
-62 kJ mol−1, whereas the activation energy is 78 kJ
mol−1, with respect to B4. The transition state for
(CH3)2CHOC(CH3)2
(CH3)2CO + i-C3H7
i-C3H7 + O2
C3H6 + HO2
(CH3)2CHOCHCH2CH3
(CH3)2CHO + O2
(CH3)2CHO + C3H6
(CH3)2CO + HO2
β-IPOXI Reaction Pathway. β-IPOXI radical this step stands marginally above the thermochem-
reacts with O2 to yield β-peroxy radical (β-ROO), ical threshold. β-IPOXI direct decomposition path-
with a calculated reaction enthalpy of -145 kJ mol−1
.
way is considered to be another source of propene
β-ROO, further, is expected to undergo similar uni- at higher temperature experiments, since tempera-
molecular processes as α-ROO, consisting of an in- ture enhances the reactivity of β-carbons promoting
tramolecular H-abstraction to yield several QOOH β-IPOXI radicals formation and subsequent decom-
species as illustrated in Scheme 2B. When the hydro- position.
gen is removed from an α-carbon of the other side, an
Branching Fractions. Photoionization cross-
α-QOOH-β (B1) is produced. This species can un- sections, besides being important in products char-
dergo the same cyclization as β-QOOH (A2) yielding acterization, are also useful in quantitative analysis.
the heterocyclic acetal, 2,2,4-trimethyl-1,3-dioxolane. Concentrations of primary products relative to the
If hydrogen on the β-carbon on the other side is ex- reactant, also known as branching fractions, are ob-
tracted, β-QOOH-β (B2) is generated. Even here, tained taking the ratio of the observed ion intensity
the attack of the methyl radical on the peroxydic oxy- (see Equation 3) of a product (p) to the ion intensity
gen gives a six-membered cyclic ether 2,6-dimethyl- of the reactant (r). The following equation is used:
1,4-dioxane ((m/z = 116, C6H12O2). Energy released
is equivalent to -176 kJ mol−1 with an activation
Sp
σpδp
ꢀ
ꢁ
0.67
energy of 47 kJ mol−1 with respect to B2. These
cyclic species are not observed in our experiments
most likely due to their high activation barriers rel-
ative to B. In the event that hydrogen is removed
Cp
Cr
Spσrδr
Srσpδp
Spσr mr
=
=
=
(5)
Sr
σrδr
Srσp mp
Since the photoionization cross-section and ion sig-
from the α-carbon on the same side, a β-QOOH-α nal depend on the used photon energy, PI spectra
11
This article is protected by copyright. All rights reserved.

10.1002/cphc.201901134
ChemPhysChem
TS6(72;32)
a-IPOXI + O
2
0
-50
TS4(41;233)
TS5(122;24)
TS1(68;20)
Isopropenyl
isopropyl ether
+ HO
2
TS3(58;204)
Acetone + Propene
TS2(95;5)
+ HO
2
H3C
O
CH3
A2
O
OH
O
CH3 CH2
H3C
O
CH3
H3C
H3C
CH3
CH2
-100
-150
-200
-250
CH2 CH3
A1
OH
O
H3C
O
CH3
CH3 CH3
A
O
O
O
H3C
CH3
H3C
CH3
O
O
CH3 CH3
O
H3C
CH3
CH3
2 Acetone + OH
2,2,4-trimethyl-1,3-dioxolane
+ OH
-300
ß-IPOXI + O
2
TS14(146;207)
TS13(79;63)
0
OH
TS9(83;165)
OH
TS10(47,223)
O
O
TS12(104;68)
TS11(104;37)
TS7(65;30)
H2C
O
CH2
H3C
O
CH2
-50
-100
-150
-200
-250
-300
CH3 CH2
CH3 CH3
TS8(95;28)
2,5-dimethyltetrahydrofuran
isopropoxy oxetane
ether + OH
+ HO
2
TS14(12;84)
B2
B4
O
CH3
H3C
O
OH
B3
OH
B1
H3C
O
O
O
H3C
H3C
O
CH2
H3C
O
CH2
O
CH3 CH3
CH3 CH3
B
isopropoxy
methyloxirane
ether + OH
O
H3C
O
CH2
CH3
CH3
2,2,4-trimethyl-1,3-dioxolane
+ OH
CH3 CH3
O
H3C
H3C
CH3
O
O
O
O
2,6-dimethyl-1,4-dioxane
+ OH
H3C
O
CH3
CH3
Scheme 2: Decomposition pathways for α-ROO (A) and β-ROO (B). The red line indicates the reference energy level.
The Y axis accounts for the relative energy expressed as kJ mol−1. Numbers in parentheses correspond to energy barriers for
the forward reaction (left value) and reverse reaction (right value) starting from the A or B species.
12
This article is protected by copyright. All rights reserved.

10.1002/cphc.201901134
ChemPhysChem
Chemical Name m/z 298K 550K 650K
pathways to form propanal and 2-5,DMTHF as pri-
mary products, we consider them as deriving from
secondary chemistry, therefore, their branching frac-
tions are not listed in Table 1 either. Even though
IPAC derives from secondary reactions, it is a main
product at room temperature. It is worthwhile to
comment that this result suggests that at room tem-
perature the oxidation reaction of DIPE proceeds via
secondary reactions about 40-50 % of the time, most
likely through the self-reactions of α-ROO radicals.
As the temperature increases this path becomes less
relevant. The importance of the peroxy radicals self-
reaction has been recently reported[51] and the mech-
anisms were investigated. No particular changes are
noted when the temperature rises from 550 K up
to 650 K. From m/z = 61 time trace, which shows
the same time behavior of its parent ion IPAC and
the measured branching fractions, we can obtain a
rough estimate of the bimolecular rate constant (k3)
at room temperature for the ROO self-reaction yield-
ing IPAC by making some bold assumptions. The re-
action is complete at 15 ms (t3) when approximately
45 % DIPE is consumed. Taking a concentration of α-
ROO radicals (A) of roughly 8 x 1011 molecule cm−3
(we are assuming that all Cl atoms, limiting reac-
tant, yield α-IPOXI radicals, which all react with O2
forming α-ROO), the reaction proceeds with a k3 of
acetone
58
45
83
80
±15% ±29% ±28%
isopropyl
acetate
102
41
±14%
-
-
Table 1: Branching fractions for reaction products at each
experimental temperature calculated by integrating the ion in-
tensities over a reaction time range from 20 to 50 milliseconds.
are compared at the same highest energy in these
experiments (11 eV) where signal fluctuations are
minimized. Absolute photoionization cross-section of
acetone[37] was taken from literature. The overall
uncertainty of each branching fraction is calculated
using the propagation of the errors of all the quan-
tities in Equation 5. The error of the ion signal is
approximately 5 %, the error of σDIP E(11 eV) is 28
%, and the error of the mass discrimination factor is
3 % from reference.[35] Branching fractions relative
to DIPE are listed in Table 1 with uncertainties.
The calculated fractions of acetone may be affected 1 x 10−10molecule−1cm3s−1. In addition, from m/z
by possible secondary reactions, although based on = 58 time trace and its measured branching fraction
the observed kinetic traces we cut off part of the sig- we can get a crude estimate of the unimolecular rate
nal related to high reaction time (signal integration constant k to form acetone from A. The reaction is
is in the 0-30 ms range), thus cleaning up the spectra complete after 5 ms (t1) when 45 % of DIPE is de-
from secondary reactions as much as possible. We pleted yielding a k of 160 s−1. If the rate determing
also want to point out the fact that one thermody- step is the formation of α-QOOH (A1), then k = k1
namically feasible product cannot be quantified be- ' 160 s−1. Because of the measured branching frac-
cause of its poor Franck-Condon factors (isopropoxy tions of the species coming directly from the ROO
methyloxirane). Acetone is the only detected primary radical, i.e., acetone and IPAC, we can also derive k1
product. As the temperature increases, propene ap- from the k3, that is:
pears at both 550 K and 650 K experiments, sug-
k1
[Acetone]
[IPAC]
gesting a change in product-leading routes from 298
K to higher temperatures. Despite β-IPOXI radicals
decomposition is considered to be a primary reac-
tion to yield propene, it is also a product from sec-
ondary reactions involving isopropyl radicals, thus
its branching fractions are not listed. Since it was
=
= 1
(6)
(7)
2k3[A]0
ꢀ
ꢁ
1
1
1
k1 = 2
−
' 160s−1
t3 0.45[A]0
[A]0
where 2 is due to the fact that 2A are needed to
not possibile to find thermodynamically-favourable generate IPAC. This value is consistent with the one
13
This article is protected by copyright. All rights reserved.

10.1002/cphc.201901134
ChemPhysChem
obtained from the acetone time trace.
rector, Office of Science, Office of Basic Energy Sci-
ences, of the U.S. Department of Energy under Con-
tract No. DE-AC02-05CH11231. The authors also
acknowledge Prof. Fabio Ramondo from the Univer-
sity of L’Aquila (Italy) and Dr. Leonid Sheps from
6 Conclusions
We studied the Cl-initiated oxidation reaction of the Combustion Research Facility of Sandia National
DIPE at 298, 550, and 650 K, under low pressure Laboratories, who supported the authors on their
conditions. Acetone has been detected as the main work.
primary product. Branching fractions showed differ-
ences between the experiment at room temperature
8 Keywords
and higher temperatures. Indeed, the presence of iso-
propyl acetate in the experiment at room tempera-
ture suggests that a bimolecular channel is impor-
Branching Fractions, Combustion, Diisopropyl Ether
Oxidation, Multiplexed Synchrotron Photoionization
tant despite the low pressure conditions. The iso-
Mass Spectrometry, Photoionization Cross-section,
merization of peroxy radicals leading to QOOHs fol-
lowed by decomposition, i.e., unimolecular channel,
Primary Reactions, Secondary Reactions
is considered to be the source of acetone at room
temperature. Since cyclic ethers are not observed
9 TOC
at higher temperature, it seems that the cyclization
pathway do not occur or are significantly reduced,
due to the thermal decomposition of α-IPOXI and
β-IPOXI radicals, by C-O bond fission. These path-
ways are the major sources of acetone at higher tem-
Multiplexed synchrotron photoionization mass spec-
trometry has been employed to study the Cl-initiated
oxidation of diisopropyl ether at 298, 550, and 650
K. Acetone, propanal, propene, and isopropyl acetate
peratures. This finding results in agreement with a
have been identified as major reaction products. Ace-
previous study[61], where thermal decomposition of
tone is the main primary product at all temperatures.
radicals deriving from ethers has been found to be
At 298 K almost 50% of the reaction proceeds via the
much faster than reaction with O2 at 753 K. Propene
self-reaction of the initial peroxy radical to yield iso-
time trace shows a slower formation typical of sec-
propyl acetate.
ondary chemistry processes. Most likely, propene for-
mation passes through chain-terminating reactions of
isopropyl radicals with O2 also yielding HO2 and by
References
β-IPOXI radicals decomposition. Finally, because we
did not find any thermodynamically-favorable path-
ways to explain propanal formation as primary prod-
uct, we consider it as deriving from secondary chem-
istry.
[1] Gopinath Dhamodaran, Ganapathy Sun-
daram Esakkimuthu, and Yashwanth Kutti
Pochareddy. Fuel, 173:37–44, 2016.
[2] Ronald G. Harvey. Polyciclic Aromatic Hydro-
carbons, 2011.
7 Acknowledgment
[3] G. Grimmer.
2017.
Environmental Carcinogenics,
This work is supported by the American Chemical
Society Petroleum Research Fund Grant 56067-UR6.
The authors express their gratitude to Drs. Taat-
jes and Osborn from Sandia National Laboratories
for the utilization of SPIMS experimental apparatus.
The Advanced Light Source is supported by the Di-
[4] B. J. Finlayson-Pitts. Science, 276(5315):1045–
1051, 1997.
[5] Clean Fuels Development Coalition. Oxygenates
Fact Book, www.cleanfuelsdc.org, 2014.
14
This article is protected by copyright. All rights reserved.

10.1002/cphc.201901134
ChemPhysChem
[6] J Yanowitz, E Christensen, and R L McCormick. [19] Martin Y. Ng, Brittany M. Bryan, Jordan Nel-
Office of Scientific and Technical Information
(OSTI), 2011.
son, and Giovanni Meloni. J. Phys. Chem. A,
119(32):8667–8682, 2015.
[7] V. D. Stytsenko, A. A. Lavarinenko, L. E. Nadra, [20] IUPAC, Subcommittee on Gas Kinetic Data
and V. A. Vinokurov. Chem. Technol. Fuels
Oils, 44(5):364–369, 2008.
Evaluation - http://iupac.pole-ether.fr.
[21] D. Maric, J.P. Burrows, R. Meller, and
G.K. Moortgat. J. Photochem. Photobiol., A,
70(3):205–214, 1993.
[8] Pamela M. Franklin, Catherine P. Koshland,
Donald Lucas, and Robert F. Sawyer. Environ.
Sci. Technol., 34(18):3857–3863, 2000.
[22] Yasmin Fathi, Chelsea Price, and Giovanni Mel-
[9] Environmental Protection Agency: State Ac-
tions Banning MTBE. 2007.
oni. J. Phys. Chem. A, 121(15):2936–2950, 2017.
[23] William A. Chupka and Joseph Berkowitz. J.
Chem. Phys., 47(8):2921–2933, 1967.
[10] J. Sutherland and C. Adams. Environ. Eng. Sci.,
24(8):998–1005, 2007.
[24] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E.
Scuseria, M. A. Robb, J. R. Cheeseman, G. Scal-
mani, V. Barone, B. Mennucci, G. A. Petersson
et al. Gaussian 09, 2009.
[11] Matthew Winfough, Rong Yao, Martin Ng,
Katherine Catani, and Giovanni Meloni. J.
Phys. Chem. A, 121(7):1460–1469, 2017.
[12] Tom Shih, Yue Rong, Thomas Harmon, and [25] J. A. Montgomery, M. J. Frisch, J. W. Ochter-
Mel Suffet. Environ. Sci. Technol., 38(1):42–48,
2004.
ski, and G. A. Petersson. J. Chem. Phys.,
110(6):2822–2827, 1999.
[13] Frank P. Heese, Mark E. Dry, and Klaus P. [26] J. A. Montgomery, M. J. Frisch, J. W. Ochter-
M¨oller. Catal. Today, 49(1-3):327–335, 1999.
ski, and G. A. Petersson. J. Chem. Phys.,
112(15):6532–6542, 2000.
[14] E.M. Collins, H.W. Sidebottom, J.C. Wenger,
S. Le Calv´e, A. Mellouki, G. LeBras, E. Vil- [27] Wolfgang Quapp. Theor. Chem. Acc., 121(5-
lenave, and K. Wirtz. J. Photochem. Photobiol.,
A, 176(1-3):86–97, 2005.
6):227–237, 2008.
[28] D. L. Baulch, C. T. Bowman, C. J. Cobos,
R. A. Cox, Th. Just, J. A. Kerr, M. J. Pilling,
D. Stocker, J. Troe, W. Tsang, R. W. Walker,
and J. Warnatz. J. Phys. Chem. Ref. Data,
34(3):757–1397, 2005.
[15] Sainan Wang and Liming Wang. Phys. Chem.
Chem. Phys., 18(11):7707–7714, 2016.
[16] David L. Osborn, Peng Zou, Howard Johnsen,
Carl C. Hayden, Craig A. Taatjes, Vadim D.
Knyazev, Simon W. North, Darcy S. Peterka, [29] Fabrizio Santoro, Roberto Improta, Alessandro
Musahid Ahmed, and Stephen R. Leone. Rev.
Sci. Instrum., 79(10):104103, 2008.
Lami, Julien Bloino, and Vincenzo Barone. J.
Chem. Phys., 126(8):084509, 2007.
[17] Craig A. Taatjes, Nils Hansen, David L. Osborn, [30] Terrill A. Cool, Andrew McIlroy, Fei Qi,
Katharina Kohse-H¨oinghaus, Terrill A. Cool,
and Phillip R. Westmoreland. Phys. Chem.
Chem. Phys., 10(1):20–34, 2008.
Phillip R. Westmoreland, Lionel Poisson,
Darcy S. Peterka, and Musahid Ahmed. Rev.
Sci. Instrum., 76(9):094102, 2005.
[18] Amelia W. Ray, Craig A. Taatjes, Oliver Welz, [31] Matthew A. Oehlschlaeger, David F. Davidson,
David L. Osborn, and Giovanni Meloni. J. Phys.
Chem. A, 116(25):6720–6730, 2012.
and Ronald K. Hanson. J. Phys. Chem. A,
110(32):9867–9873, 2006.
15
This article is protected by copyright. All rights reserved.

10.1002/cphc.201901134
ChemPhysChem
[32] Mingfeng Xie, Zhongyue Zhou, Zhandong Wang, [45] Richard R. Hiatt, Theodore Mill, Katherine C.
Dongna Chen, and Fei Qi. Int. J. Mass Spec-
trom., 293(1-3):28–33, 2010.
Irwin, and Jane K. Castleman. J. Org. Chem.,
33(4):1421–1428, 1968.
[33] James M. Williams and William H. Hamill. J.
Chem. Phys., 49(10):4467–4477, 1968.
[46] Richard R. Hiatt, Theodore Mill, and Frank R.
Mayo. J. Org. Chem., 33(4):1416–1420, 1968.
[34] K. Watanabe, Toshio Nakayama, and Joseph
Mottl. J. Quant. Spectrosc. Radiat. Transfer,
2(4):369–382, 1962.
[47] R. Hiatt and Ludek Zigmund. Can. J. Chem.,
48(24):3967–3969, 1970.
[48] J. A. Howard and K. U. Ingold. Can. J. Chem.,
46(16):2655–2660, 1968.
[35] Oliver Welz, Judit Za´dor, John D. Savee, Mar-
tin Y. Ng, Giovanni Meloni, Ravi X. Fernandes,
Leonid Sheps, Blake A. Simmons, Taek Soon
Lee, David L. Osborn, and Craig A. Taatjes.
Phys. Chem. Chem. Phys., 14(9):3112, 2012.
[49] J. A. Howard and J. E. Bennett. Can. J. Chem.,
50(14):2374–2377, 1972.
[50] K. Adamic, J. A. Howard, and K. U. Ingold.
Can. J. Chem., 47(20):3803–3808, 1969.
[36] John C. Traeger. Int. J. Mass Spectrom. Ion
Processes, 58:259–271, 1984.
[51] Audrey R. Smith, Simone Di Muzio, Fabio Ra-
mondo, and Giovanni Meloni. Phys. Chem.
Chem. Phys., 21(20):10228–10237, 2019.
[37] Terrill A. Cool, Juan Wang, Koichi Nakajima,
Craig A. Taatjes, and Andrew Mcllroy. Int. J.
Mass Spectrom., 247(1-3):18–27, 2005.
[52] Glen A. Russell.
79(14):3871–3877, 1957.
J. Am. Chem. Soc.,
[38] John Dyke, Andrew Ellis, Neville Jonathan, and
Alan Morris. J. Chem. Soc., Faraday Trans. 2,
81(10):1573–1586, 1985.
[53] D. Lindsay, J. A. Howard, E. C. Horswill,
L. Iton, K. U. Ingold, T. Cobbley, and A. Ll.
Can. J. Chem., 51(6):870–880, 1973.
[39] Wayne M. Trott, Normand C. Blais, and Ed-
ward A. Walters. J. Chem. Phys., 69(7):3150–
3158, 1978.
[54] Richmond Lee, Ganna Gryn ova, K. U. Ingold,
and Michelle L. Coote. Phys. Chem. Chem.
Phys., 18(34):23673–23679, 2016.
[40] Linstrom. http://webbook.nist.gov, 2005.
[55] Peter McLoughlin, Rosaleen Kane, and Imelda
Shanahan. Int. J. Chem. Kinet., 25(3):137–149,
1993.
[41] Juan Wang, Bin Yang, Terrill A. Cool, Nils
Hansen, and Tina Kasper. Int. J. Mass Spec-
trom., 269(3):210–220, 2008.
[56] Linda Nelson, Oliver Rattigan, Ruaidhri
Neavyn, Howard Sidebottom, Jack Treacy, and
[42] F. M. Benoit, A. G. Harrison, and F. P. Lossing.
Org. Mass Spectrom., 12(2):78–82, 1977.
Ole John Nielsen.
22(11):1111–1126, 1990.
Int. J. Chem. Kinet.,
[43] Timothy J. Wallington, Jean M. Andino,
Alan R. Potts, Sara J. Rudy, Walter O.
Siegl, Zhengyu Zhang, Michael J. Kurylo,
and Robert E. Huie. Environ. Sci. Technol.,
27(1):98–104, 1993.
[57] Craig A. Taatjes.
110(13):4299–4312, 2006.
J. Phys. Chem. A,
[58] Adam M. Scheer, Oliver Welz, Judit Z´ador,
David L. Osborn, and Craig A. Taatjes. Phys.
Chem. Chem. Phys., 16(26):13027–13040, 2014.
[44] Richard R. Hiatt and Katherine C. Irwin. J.
Org. Chem., 33(4):1436–1441, 1968.
16
This article is protected by copyright. All rights reserved.

10.1002/cphc.201901134
ChemPhysChem
[59] E. W. Kaiser, L. Rimai, and T. J. Wallington.
J. Phys. Chem., 93(10):4094–4098, 1989.
[60] Edgar G. Estupin˜´an, Stephen J. Klippenstein,
and Craig A. Taatjes. J. Phys. Chem. B,
109(17):8374–8387, 2005.
[61] R. S. Tranter and R. W. Walker. Phys. Chem.
Chem. Phys., 3(21):4722–4732, 2001.
17
This article is protected by copyright. All rights reserved.