Ozonolysis of a Series of Methylated Alkenes: Reaction Rate Coefficients and Gas-Phase Products

Article abstract of DOI:10.1002/kin.20934

The gas-phase ozonolysis of three methylated alkenes, i.e., trans-2,2-dimethyl-3-hexene (22dM3H), trans-2,5-dimethyl-3-hexene (25dM3H), and 4-methyl-1-pentene (4M1P), has been investigated in the presence of sufficient hydroxyl radical scavenger in a laminar flow reactor at ambient temperature (296 ± 2 K) and P = 1 atm of dry air (RH ≤ 5%). Ozone levels in the reactor were monitored by an automatic analyzer. Alkene and gas-phase product concentrations were determined via online sampling either on three-bed adsorbent cartridges followed by thermodesorption and GC/FID-MS analysis or on 2,4-dinitrophenylhydrazine (DNPH) cartridges for subsequent HPLC/UV analysis. Reaction rate coefficients of (3.38 ± 0.12) × 10^-17 for 22dM3H and (2.71 ± 0.26) × 10^-17 for 25dM3H, both in cm³ molecule^-1 s^-1 units, have been obtained under pseudo-first-order conditions. Primary carbonyl products have been identified for the three investigated alkenes, and branching ratios are reported. In the case of 4M1P ozonolysis, the yield of a Criegee intermediate was indirectly determined. Kinetics and product study results are compared to those of literature when available. This work represents the first investigation of reaction products in the ozonolysis of 22dM3H, 25dM3H, and 4M1P in a flow reactor.

Full text of DOI:10.1002/kin.20934

Ozonolysis of a Series
of Methylated Alkenes:
Reaction Rate Coefficients
and Gas-Phase Products
1
1
1
2
´
TRISTAN BRAURE, VERONIQUE RIFFAULT, ALEXANDRE TOMAS, ROMEO IULIAN OLARIU,
CECILIA ARSENE,2 YURI BEDJANIAN,3 PATRICE CODDEVILLE1
1Mines Douai, SAGE, F-59508, Douai, France
2“Al. I. Cuza” University of Iasi, Faculty of Chemistry, 700506, Iasi, Romania
3Institut de Combustion, Ae´rothermique, Re´activite´ et Environnement, CNRS, F-45071, Orle´ans, France
Received 2 April 2015; revised 15 June 2015; accepted 17 June 2015
DOI 10.1002/kin.20934
Published online 23 July 2015 in Wiley Online Library (wileyonlinelibrary.com).
ABSTRACT: The gas-phase ozonolysis of three methylated alkenes, i.e., trans-2,2-dimethyl-
3-hexene (22dM3H), trans-2,5-dimethyl-3-hexene (25dM3H), and 4-methyl-1-pentene (4M1P),
has been investigated in the presence of sufficient hydroxyl radical scavenger in a laminar flow
reactor at ambient temperature (296 2 K) and P = 1 atm of dry air (RH ࣘ 5%). Ozone levels in
the reactor were monitored by an automatic analyzer. Alkene and gas-phase product concentra-
tions were determined via online sampling either on three-bed adsorbent cartridges followed
by thermodesorption and GC/FID-MS analysis or on 2,4-dinitrophenylhydrazine (DNPH) car-
tridges for subsequent HPLC/UV analysis. Reaction rate coefficients of (3.38 0.12) × 10−17 for
22dM3H and (2.71 0.26) × 10−17 for 25dM3H, both in cm3 molecule−1 s−1 units, have been
obtained under pseudo–first-order conditions. Primary carbonyl products have been identified
for the three investigated alkenes, and branching ratios are reported. In the case of 4M1P
ozonolysis, the yield of a Criegee intermediate was indirectly determined. Kinetics and product
study results are compared to those of literature when available. This work represents the first
investigaꢀtCion of reaction products in the ozonolysis of 22dM3H, 25dM3H, and 4M1P in a flow
reactor.
2015 Wiley Periodicals, Inc. Int J Chem Kinet 47: 596–605, 2015
hydroxyl radicals and carbonyl compounds or car-
boxylic acids in the atmosphere [1]. Some alkenes may
also be responsible for the formation of secondary or-
ganic aerosols [2], implying effects on human health
and climate by modifying the Earth’s radiative bal-
ance and cloud properties [3]. The first steps in alkene
ozonolysis are quite well known: Ozone is added across
the alkene double bond to form an unstable species
called a primary ozonide, which quickly decomposes to
INTRODUCTION
Alkene ozonolysis reactions play a major role in tro-
pospheric chemistry by leading to the formation of
Correspondence to: Ve´ronique Riffault; e-mail: veronique.
riffault@mines-douai.fr.
Supporting Information is available in the online issue at
www.wileyonlinelibrary.com.
ꢀC
2015 Wiley Periodicals, Inc.

OZONOLYSIS OF A SERIES OF METHYLATED ALKENES
597
form a primary carbonyl and an excited species known
as a Criegee Intermediate (CI) [4]. The yields of pri-
mary carbonyls for many alkenes are quite well docu-
mented in the literature; however, the high reactivity of
the CI makes their detection and quantification quite
difficult. Unresolved questions also remain regarding
their atmospheric fate and the products formed by the
reactions of these energy-rich intermediates with other
species including unimolecular and bimolecular reac-
tions [5–7]. Unimolecular reactions involve isomeriza-
tion followed by dissociation, leading to radical frag-
ments among which OH radicals, or ring closure via
the formation of an excited dioxirane followed by iso-
merization to hot acid and fragmentation. Bimolecular
reactions between CI and other atmospheric species
(e.g., H2O) become possible after stabilization of the
excited CI through collisional quenching. One of the
most important bimolecular reactions of the CI in the
atmosphere involves H2O, which leads to the forma-
tion of organic acids and H2O2. Isomerization of the
CI is directly dependent on the presence of the alkyl
substituent on the same side as the terminal O atom in
the CI, i.e. the CI is in a syn conformation [6]. In a
recent work from our group, the formation of a nonox-
idized intermediate product (2-methyl-1-propene) in
the ozonolysis of 2,4,4-trimethyl-2-pentene has been
reported, suggesting a new way for the Criegee birad-
ical to evolve through isomerization and decomposi-
tion [8]. According to this pathway, a CI bearing a
suitable hydrogen atom could follow an electron rear-
rangement leading to the formation of OH and HCO
radicals, together with a nonoxidized alkene.
The present study focuses on the ozonolysis
of 4-methyl-1-pentene (4M1P), trans-2,2-dimethyl-3-
hexene (22dM3H), and trans-2,5-dimethyl-3-hexene
(25dM3H), under controlled conditions. Each selected
alkene has an available hydrogen atom for the rear-
rangement discussed above, thus enabling the investi-
gation of the relevance of this channel in the ozonolysis
of longer alkenes (Fig. 1).
Some of the alkenes investigated in this work (e.g.,
4M1P) have been identified in common processes
such as tailpipe emissions of volatile hydrocarbons
from gasoline-powered motor vehicles [9], residential
fireplace combustion of wood [10], or in natural fires
of foliage, litter, and herbaceous matter [11]. To the
best of our knowledge, only one research group re-
ported results on these alkenes obtained in a smog
chamber [12–15]. In the present work, rate coefficients
and products yields for the three reactions have been
determined in a laminar flow reactor at atmospheric
pressure of dry air (RH < 5%) and ambient tempera-
ture, under room light.
MATERIAL AND METHODS
Laminar Flow Reactor
Experiments were carried out in a laminar flow re-
actor (10 cm internal diameter) equipped with a mo-
bile injector, which enables the reactor volume to be
changed and, thus, the reaction time. The entire ex-
perimental setup (Fig. 2) has already been described
in detail elsewhere [8]. The alkene gas flow was pro-
duced using a pressurized canister prepared by inject-
ing a known amount of the reactant volatile organic
compound (VOC) with a microsyringe and filling up
the canister to a few bars with zero air. Ozone was
generated by an ozone generator (model 146i; Thermo
Scientific, Villebon-sur-Yvette, France). The OH rad-
ical scavenger gas flow was produced either directly
by a pressurized bottle (5000 ppm of CO in N2) or
by using a pressurized canister in the same way as the
alkenes (2-butanol). Reactants, diluted in purified air
using mass flow controllers (model 247C 4-channel
readout; MKS Instruments, Munchen, Germany), are
sent to the top of the reactor in separate tubings and
mixed in the injection head.
The central outlet of the flow reactor was used
as an exhaust, whereas the gas phase was sampled
through the lateral outlet of the reactor for online analy-
sis. Organic compound concentrations were measured
using an online sampling of the gas flow on three-
bed adsorbent cartridges (Carbopack B, Carbopack
C, and Carbosieve SIII in 1:2:1 packing ratio from
Supelco, Saint-Quentin-Fallavier, France) with sub-
sequent desorption and analysis by a gas chromato-
graphic system (GC) equipped with parallel flame
ionization detection (FID) and mass spectrometry
(MS) (model 5957B; Agilent Technologies, Les Ulis,
France). A Varian capillary column (CP-Sil 5 CB
50 m, 0.32 mm, 1.2 µm; Les Ulis, France) was used
with the following temperature program: 0°C, 15 min
isothermal, 3°C min−1 to 90°C, then 10°C min−1 to
250°C and hold for 5 min. Helium was used as the
carrier gas with a column flow of ꢀ4 mL min−1. In ad-
dition to the product analysis through GC-FID/MS, the
most volatile carbonyl products were analyzed using
Figure 1 Molecular structures of the three alkenes studied.
International Journal of Chemical Kinetics DOI 10.1002/kin.20934

598
BRAURE ET AL.
Figure 2 Schematic of the experimental and instrumentation (MFC: mass flow controlIer).
2,4-dinitrophenylhydrazine (DNPH) derivatization
and further elution of the hydrazones from the car-
tridges with 4–6 mL acetonitrile. Separation of the ex-
tracted analytes was achieved by high-pressure liquid
chromatography (HPLC; Waters Alliance 2695; col-
umn C18, 25 cm × 4.6 mm × 5 mm at 40°C; Guyan-
court, France) using a ternary elution mobile phase
at 1.5 mL min−1 of acetonitrile/water/ tetrahydrofu-
ran (linear gradient from 0.3/0.2/0.5 to 0.65/0.05/0.3
in 15 min then hold 12 min for a total elution time of 27
min). Carbonyl compounds were detected with a UV–
vis detector (Waters Dual λ 2487) at 365 nm. Quantita-
tive analysis involved the use of external standards pre-
pared with carbonyl-DNPH derivatives synthesized in
the laboratory. Calibration curves were used to quantify
the carbonyl concentrations in the collected sample.
Ozoneconcentrationsweremonitoredbyanozoneana-
lyzer (model 49C; Thermo Environmental Instruments,
Villebon-sur-Yvette, France). Experiments were car-
ried out under room light, at ambient pressure and a
controlled temperature of (296 2) K, with a reaction
time ranging from 40 to 300 s.
reaction rate coefficient, kvoc, has been determined
following the ozone loss at different reaction times,
t, which is described by the following relation:
ꢀ
ꢁ
[O3]t
[O3]0
ln
= − (kVOC × [VOC]t + kwall) × t
(1)
where kwall is the first-order loss of ozone at the reactor
wall (in s−1), measured in separate experiments in the
absence of VOC. In excess of VOC, when [VOC]t re-
mains constant and equal to [VOC]0, Eq. (1) becomes
ꢀ
ꢁ
ꢂ
ꢃ
[O3]t
[O3]0
ln
= − kVꢁ OC + kwall × t
(2)
where kVꢁ OC = kVOC × [VOC]0.
The reaction rate coefficient kVOC can be determined
by subtracting kwall from the obtained linear regression
coefficients and plotting the resulting kVꢁ OC values ver-
sus the initial VOC concentrations. The plot of kVꢁ OC
versus [VOC]0 results in a linear dependence whose
slope is equal to kVOC. The overall uncertainty on the
rate coefficient, σk, has been calculated using the prop-
agation of uncertainty approach by Eq. (3):
Kinetic Experiments
In the kinetic experiments, the defined reactant con-
centrations allowed to work under pseudo–first-order
conditions, where the alkene concentration was at least
10 times higher than the ozone concentration. The
ꢄ
1/2
ꢀ
ꢁ
2
ꢅ
ꢆ
2
σkstat
k
σVOC
σk = K ×
+
(3)
[VOC]
International Journal of Chemical Kinetics DOI 10.1002/kin.20934

OZONOLYSIS OF A SERIES OF METHYLATED ALKENES
599
Table I Experimental Conditions and Results of the Ozonolysis Study for the Two Investigated Alkenes
[VOC]0
(1013 molecule cm−3
[O3]0
(1012 molecule cm−3
[CO]
Number of
Experiments
kVꢁ OC
(10−3 s−1
)
Alkene
)
)
(1015 molecule cm−3
)
22dM3H
25dM3H
2.19–8.56
2.87–10.63
1.67–2.39
1.81–7.53
5.74–7.58
0–8.61
17
5
0.55–2.12
0.95–3.08
with σkstat the 95% confidence interval on the slope
of the weighted linear least-squares fit and σVOC the
uncertainty on the concentration of the compound in
excess, considered as one standard deviation of several
GC/FID analyses of a cartridge doped with identical
amounts of the compound.
DETERMINATION OF THE REACTION
RATE COEFFICIENTS
The experimental conditions are summarized in
Table I. Kinetic studies of 22dM3H and 25dM3H
ozonolysis were carried out in VOC excess using vari-
ous initial reactant concentrations. An •OH scavenger,
CO, has been added in all 22dM3H experiments but
only once in the case of 25dM3H as it proved to
be unnecessary given the large excess of VOC and
the absence of significant secondary chemistry. The
rate coefficient of 4M1P ozonolysis had already been
determined in a previous study from our group as
k4M1P = (8.23 0.50) × 10−18 cm3 molecule−1 s−1 at
296 K [8].
Product Experiments
For product experiments, excess CO or 2-butanol was
added to scavenge the OH radicals possibly formed in
the reactions. Identification of the products was carried
out either by GC-MS analysis and mass spectra library
search (NIST MS Search 2.0) or by HPLC analysis
and comparison of retention times with authentic stan-
dards. Product yields were determined by recording
the formation of the products, and the loss of the re-
actants over time and by plotting the mole fraction of
product formed versus the mole fraction of reactant
consumed (VOC consumption was used in 22dM3H
and 25dM3H studies, whereas ozone consumption was
used for 4M1P since the coelution of a reaction product
with 4M1P would underestimate 4M1P consumption).
Using an •OH scavenger prevents the alkenes from ad-
For each initial alkene concentration, the pseudo–
first-order rate coefficient kVꢁ OC was retrieved using
Eq. (2). Examples of O3 decays obtained with dif-
ferent concentrations of VOC are plotted in Fig. 3 for
•
ditional consumption by OH and the products from
degradation by OH radical reactions.
Reagents
All chemicals were used as received without fur-
ther purification. 22dM3H (97%), 25dM3H (97%),
4M1P (98%), 3-methylbutanal (97%), 2,2-dimethyl-
propanal (96%), 2-methyl-propanal (>99%), propanal
(97%), isobutyl formate (97%), and acetone (>99.8%)
were purchased from Sigma-Aldrich (Saint-Quentin-
Fallavier, France); 2-methyl-1-propene (10 ppm in N2)
from Messer-Griesheim (Du¨sseldorf, Germany); 2-
butanone (>99%) and 2-butanol (>99%) from Merck
(Darmstadt, Germany). Carbon monoxide (5000 ppm
in N2) was supplied by Praxair (Danbury, CT, USA).
Purified air was produced by an air purifier (AZ 2020;
Claind, Wasquehal, France) and was characterized by a
relative humidity of less than 5% and an organic carbon
content of maximum 0.1 ppbv.
Figure 3 Examples of ozone consumption versus reaction
time for [22dM3H]0 (1013 molecule cm–3) = 0 (diamonds);
2.22 (circles); 5.16 (triangles); 8.56 (squares). Straight lines
are linear least-square fits forced through zero. The plotted
uncertainties (1σ) are graphical representations of the un-
certainty on ln ([O3]t/[O3]0) calculated using Eq. (4) (see
text).
International Journal of Chemical Kinetics DOI 10.1002/kin.20934

600
BRAURE ET AL.
The rate coefficients obtained in the present work
are lower—by 16% for 22dM3H and 29% for
25dM3H—than those reported by the only previous
study [12], which were determined as the average
of three kVOC values calculated from the measured
pseudo–first-order rate coefficients. However, in the
case of 22dM3H, the agreement is still satisfactory
when taking into account the uncertainties reported in
both studies (see also Fig. 4). For 25dM3H, the un-
certainties alone cannot explain the disagreement. It
should be emphasized that in this previous study, the
alkene initial concentration was not directly measured
but calculated from the injected reactant amount and
the chamber volume, which could induce a bias in
[VOC]0 needed for determining kVOC, whereas in the
present work the initial VOC concentration was accu-
rately measured by GC.
As evidenced by Grosjean and Grosjean [12], 1,2-
disubstituted alkenes bearing bulky substituents, such
as 22dM3H and 25dM3H, display much lower reac-
tivity than what would be expected from substituent
inductive effects alone, suggesting the contribution of
steric effects. From the data obtained in the present
work, it can be observed that the rate for 22dM3H is
higher than that for 25dM3H, suggesting that the two
methyl substituents induce a lower steric effect when
positioned together on the same carbon atom.
Figure 4 Pseudo–first-order rate coefficients kVꢁ OC versus
initial 22dM3H concentration (open circles). The straight line
is the weighted linear least-square fit forced through zero,
whereas dotted lines represent the 95% confidence interval.
Uncertainties on kVꢁ OC correspond to the 1σ precision of
the fit (0.04 × 10–17 s–1 for 22dM3H) and uncertainties on
[22dM3H]0 correspond to the 1σ error (repeatability) on
concentrations (3.4% for 22dM3H). Filled triangles are the
experimental data from Grosjean and Grosjean [12], shown
for comparison.
INVESTIGATION OF THE REACTION
PRODUCTS
22dM3H. Similar plots were obtained for 25dM3H
(Fig. S1 in the Supporting Information). The error bars
(1σ) plotted in Fig. 3 represent the sum of the variation
Mechanistic studies of 22dM3H, 25dM3H, and 4M1P
ozonolysis were carried out in conditions close to sto-
ichiometry with a few runs performed in the presence
of CO or 2-butanol as OH radical scavengers. The
coefficients (Vx ࣕ σx/x
viation of a set of samples xi and x
̅
, where σx is the standard de-
is its mean) of the
̅
recorded series of values for [O3]0 and [O3]t as derived
from the equation:
ꢉ
ꢇ
ꢀ
ꢁꢈ
ꢀ
ꢁ
ꢀ
ꢁ
2
2
[O3]t
[O3]0
σ [O3]0
σ [O3]t
σ [O3]0
σ[O3]t
σ
ln
=
+
+ 2 ×
×
(4)
[O3]0
[O3]t
[O3]0
[O3]t
experimental conditions in terms of concentrations and
number of experiments are summarized in Table III.
Figure 4 shows the dependence between the
pseudo–first-order coefficients for 22dM3H and the
initial alkene concentrations to determine the reaction
rate coefficient. The corresponding plot for 25dM3H
is presented in Fig. S2 in the Supporting Information.
From data presented in Fig. 4 (Fig S2 in the Support-
Primary Carbonyl Products
In the case of the 22dM3H ozonolysis, propanal and
2,2-dimethylpropanal are the expected primary car-
bonyl products as described in the following reactions:
ing Information, respectively), values of k22dM3H+O3
=
(3.38
0.12) × 10−17 and k25dM3H+O3 = (2.71
0.26) × 10−17 cm3 molecule−1 s−1 have been de-
rived. Table II summarizes the kinetic results from
the present work and compares them with the only
previous literature data obtained in a smog chamber
[12].
(CH3)3CCH = CHCH2CH3 + O3
→ CH3CH2CHO + [(CH3)3CC• (H) OO•]∗ (R1a)
→ (CH3)3CCHO + [CH3CH2C•HOO•]∗
(R1b)
International Journal of Chemical Kinetics DOI 10.1002/kin.20934

OZONOLYSIS OF A SERIES OF METHYLATED ALKENES
601
Table II Summary of Kinetic Results and Comparison with Literature Data
kVOC
Experimental
Setupa
Alkene
T (K)
(1017 cm3 molecule−1 s−1
)
Reference
This work
Grosjean and Grosjean [12]
This work
22dM3H
296
295
296
291
3.38 0.12
4.03 0.67
2.71 0.26
3.83 0.50
FR; VOC excess; OA
SC; VOC excess; OA
FR; VOC excess; OA
SC; VOC excess; OA
25dM3H
Grosjean and Grosjean [12]
aFR: flow reactor; OA: ozone analyzer; SC: smog chamber.
Table III Experimental Conditions for the Mechanistic Studies of the Three Alkene Ozonolyses
[VOC]0
(1013 molecule cm−3
[O3]0
(1013 molecule cm−3
[Scavenger]
(1015 molecule cm−3
Number of
Experiments
Alkene
)
)
)
22dM3H
25dM3H
4M1P
3.43–35.50
0.82–2.75
3.31–21.74
1.51–5.16
0.45–1.76
1.52–3.26
7.58–6250a
8.68–9.06a
0.04–3.53b
50
15
10
aCO; b2-Butanol.
whereas for the symmetrical 25dM3H, 2-
methylpropanal is the unique expected primary
carbonyl compound:
(22dMP) formed in the O3 + 22dM3H reaction (R1b)
and2-methyl-propanal formed in the O3 + 25dM3H re-
action (R2) with values about 30% smaller in our case.
Similar observations can be made for HCHO formed in
the O3 + 4M1P reaction (R3a), whose yield is higher
by 20%. The sum of the yields of the primary products
measured in the present work is closer to 100%, which
adds confidence to our data. Results also show that the
bulkier carbonyl is preferentially formed in the case
of the nonsymmetrical 22dM3H and 4M1P (Table IV),
consistent with literature data, indicating competing in-
fluences of substituent inductive and steric effects [15].
(CH3)2CHCH = CHCH (CH3)2 + O3
→ (CH3)2 CHCHO + [ (CH3)2 CHC• (H) OO•]∗
(R2)
Finally, 3-methylbutanal and formaldehyde are
formed in the ozonolysis of 4M1P as primary carbonyl
compounds through the following reactions:
Fate of the Criegee Intermediates
(CH3)2CHCH2CH = CH2 + O3 → HCHO
In addition to the primary carbonyls, other products
supposed to originate from reactions of the CI were
observed. The corresponding product yields were de-
termined in a similar way as the primary carbonyl ones
and are also reported in Table IV.
22dM3H. Acetaldehyde was the only product
clearly observed from 22dM3H ozonolysis with a mo-
lar yield of about 13%, which is somewhat lower than
the yield determined by Grosjean and Grosjean [15]
(ꢀ19%). It is expected to arise from the rearrange-
ment of the syn-[CH3CH2C•HOO•]* biradical (R1b)
concomitantly released with 22dMP. The mechanism
proposed by Grosjean and Grosjean [15] implies the
formation of an excited hydroxycarbonyl via a five-
member ring isomerization.
+[ (CH3)2 CHCH2C• (H) OO•)]∗
(R3a)
→ (CH3)2CHCH2CHO + [C•H2OO•]∗ (R3b)
All the expected primary carbonyls were observed
and quantified, and their molar yields are reported in
Table IV as well as the sum of the primary carbonyl
yields. It is noteworthy that these sums are close to
100% for the three investigated alkenes, in agreement
with the general reaction scheme of alkene ozonoly-
sis leading to the formation of two primary carbonyl
products by two reaction pathways.
Only one literature article reports data (also pre-
sented in Table IV) on the primary carbonyl yields
for 22dM3H, 25dM3H, and 4M1P ozonolysis reac-
tions [15]. Overall, the comparison of the data from
the two studies shows a satisfactory agreement. A
poorer agreement is obtained for 2,2-dimethylpropanal
syn − [CH3CH2C•HOO•]∗ → CH3CH = CHOOH∗
(R4)
International Journal of Chemical Kinetics DOI 10.1002/kin.20934

602
BRAURE ET AL.
Table IV Primary and CI-Derived Reaction Product Yields Obtained in This Work and Comparison with Literature
Data When Available
Primary Carbonyl Yields (%)a
CI-Derived Reaction Product Yields (%)a
Grosjean and
Grosjean and
Grosjean [15]
Pathway
Product
This Work
Product
This Work
Grosjean [15]
(CH3)3CCH=CHCH2CH3 (22dM3H)
1a
1b
Propanal
22dMP
37.0
60.2
1.0
0.7
33.5
84.5
0.5
2.4
–
–
Acetaldehyde
Methylglyoxal
Formaldehyde
ࢣ
12.9
ࣘ2.1
0.2
0.2
18.7
7.7
3.8
1.7
1.1
0.1
2.0
Unexplained
4.6
ࢣ
97.2
1.2
118.0
2.5
ࣘ19.6
30.2
(CH3)2CHCH=CHCH(CH3)2 (25dM3H)
2
2-Methyl-propanal
97.6
2.7
139.8
8.5
Acetone
19.5
1.1
28.4
5.4
(CH3)2CHCH2CH=CH2 (4M1P)
3a
3b
Formaldehyde
3-Methyl-butanal
ࢣ
52.6
58.9
111.5
1.4
1.4
2.0
44.1
70.6
1.2
4.9
5.0
2-Methyl-propanal
[C°H2OO°]*b
ࢣ
6.8
ࣙ46.2
ࣙ53.0
3.2
7.2
7.2
0.4
0.4
–
114.7
aUncertainties correspond to statistical errors on the fits.
bIndirectly quantified after conversion to s-butyl formate by 2-butanol (see text).
CH3CH = CHOOH∗ → CH3CH (OH) CHO•∗
CH3CH (OH) CHO•∗ → CH3CHOH + HC•O
CH3CHOH + O2 → CH3CHO + HO2•
studies used sampling on DNPH cartridges followed
by HPLC-UV analysis, with the difference that for
the previous study [15], sampling began when almost
all ozone had been consumed ([O3] ࣘ 3 ppb). In our
work, ozone was present during sampling on DNPH
cartridges, which may induce a bias in the determina-
tion of the carbonyl concentrations. Besides it has been
shown previously that ozone can react during sampling
and lower the collection efficiency of the cartridge [16].
Additionally, differences in relative humidity (<5% in
our study, and ꢀ55% in the previous one [15]) can
also (i) affect the collection efficiency of the cartridge
and thus the product yields [16] and (ii) enhance the
competition between different reaction pathways for
the CI, like decomposition or reaction with water.
The formation of nonoxidized products as observed
by Duncianu et al. [8] could also be expected for
22dM3H ozonolysis. Ethene and 2-methyl-1-propene
could then be formed from syn-[CH3CH2C•HOO•]*
and syn-[(CH3)3CC•HOO•]* species, respectively.
Among these products, only 2-methyl-1-propene was
tentatively observed, but without any significant evo-
lution of its signal. A maximum yield of 3% can be
inferred, whereas ethene was too volatile to be seen
with our current analytical methods.
(R5)
(R6)
(R7)
The formation of OH radicals and methylglyoxal
was also suggested from the isomerization of the same
biradical [15]; yet, only small amounts of methylgly-
oxal were observed in the present study, leading to
an upper limit of 2.1% in contrast with Grosjean and
Grosjean [15] (7.7%). The possible fate of the anti-
CH3CH2C•HOO• and (CH3)3CC•HOO• (R1a) birad-
icals is not clear and may give rise to a number of
organic species as well as OH radicals via the de-
composition of an excited organic acid as suggested
by Johnson and marston [6]. In the present work,
formaldehyde was also detected and quantified with
a yield of 4.6% (close to the value of 3.8 reported
by Grosjean and Grosjean [15] and yet its formation
still remains unexplained. No other products were ob-
served either by the GC/FID-MS or HPLC-UV ana-
lytical methods employed in the present study. For the
two observed products from CI reactions, i.e., acetalde-
hyde and methylglyoxal (Table IV), lower yields than
those previously reported were measured. The mea-
surement of acetaldehyde and methylglyoxal in both
25dM3H. Acetone was the only product from CI re-
actions observed in 25dM3H ozonolysis experiments.
It was assumed to be formed through a similar re-
action mechanism as for acetaldehyde (see above)
via a rearrangement of the syn-[(CH3)2CHC•HOO•]*
International Journal of Chemical Kinetics DOI 10.1002/kin.20934

OZONOLYSIS OF A SERIES OF METHYLATED ALKENES
603
k12 = 1 × 10−12 cm3 molecule−1
[C•H2OO•]∗ + [C•H2OO•]∗
s
−1 [20]
biradical formed by (R2) [15]. Acetone may also be
formed from the excited vinyl hydroperoxide by re-
leasing an OH radical [6]:
→ products (including formaldehyde) (R13)
(CH3)2 CHC•HOO• → (CH3)2 C• − CHO + HO•
(R8)
k13 = (4 – 6) × 10−11 cm3 molecule−1 s−1 [21]
In the atmosphere, (R11) dominates largely over
(R12) and (R13). In the present study performed in dry
conditions (RH < 5%), (R12) and (R13) still remain
negligible. Therefore, [C•H2OO•]* can lead to an addi-
tional formationof formaldehyde andthe measuredpri-
mary formaldehyde yield may be overestimated. This
could also be an explanation for the excess of primary
carbonyl yield observed in the case of 4M1P (sum of
primary carbonyl yields: 111%), formaldehyde result-
ing in that case both from the primary and CI reactions.
In addition to the primary and CI-derived carbonyls,
the mechanism proposed by Duncianu et al. [8] in the
case of 4M1P would lead to the formation of 2-methyl-
1-propene from syn-[(CH3)2CHCH2C•(H)OO•]*. As
for 22dM3H experiments, it was tentatively observed,
but without any significant evolution of its signal and
a maximum yield of 3% could be inferred.
s-Butyl formate (sBF; C5H10O2) was also observed
in the 4M1P ozonolysis study. The formation of this
ester was supposed to originate from the reaction be-
tween 2-butanol (used as the OH scavenger in these
experiments) and the CI, [C•H2OO•]*, formed together
with 3-methylbutanal (R3b). This reaction would pro-
ceed via the formation of a hydroperoxide as the main
product, which could then undergo a thermal decom-
position and lead to sBF formation [22]:
(CH3)2 C• − CHO + O2 → (CH3)2 CO2•CHO (R9)
(CH3)2 CO•2CHO + RO2•
→ (CH3)2 CO•CHO → (CH3)2 CO + HC•O
(R10)
The lower amount of acetone measured (yield of
ꢀ20% compared to the ꢀ30% obtained by Grosjean
and Grosjean [14]) indicates the existence of other
competitive reaction pathways from the anti- and syn-
[(CH3)2CHC•HOO•]* biradicals yielding products that
are not observed.
1-Propene could also be formed from the syn-
[(CH3)2CHC•HOO•]* biradical according to the mech-
anism proposed by Duncianu et al. [8], but it is believed
that this compound was too volatile to be seen with our
current analytical methods.
4M1P. 2-Methylpropanal was observed as a car-
bonyl product from CI reactions in 4M1P ozonol-
ysis [13]. As for acetaldehyde, it is expected to
be formed following the isomerization of the syn-
[C3H7CH2C•HOO•]* biradical (R3a) into an excited
vinyl hydroperoxide. Its low formation yield (ꢀ7%
compared to ꢀ53% formaldehyde) suggests that ei-
ther the isomerization pathway is not dominant for the
CI or the vinyl hydroperoxide species formed by this
pathway will predominantly evolve toward the forma-
tion of OH radicals and other multioxygenated organic
compounds. The anti-[C3H7CH2C•HOO•]* biradical
is expected to form an excited organic acid via a dioxi-
rane [6].
[C•H2OO•]∗ + C2H5 − CH (CH3) − OH
→ C2H5 − CH (CH3) − O − C (H2) OOH
(R14)
C2H5 − CH (CH3) − O − C (H2) OOH
→ C2H5 − CH (CH3) − O − C (H2) O• + HO•
The unimolecular chemistry of [C•H2OO•]* (R3b)
is fairly well known and will generate smaller frag-
(R15)
•
C2H5 − CH (CH3) − O − C (H2) O• + O2
→ C2H5 − CH (CH3) − O − C (H) O + HO2•
ments such as OH, CO, and CO2. A fraction of
C•H2OO• may stabilize and react with H2O, ozone,
and C•H2OO• itself, to give formaldehyde and other
products:
(R16)
[C•H2OO•]∗ + H2O → HCHO + H2O2 (R11)
k11 = (0.2–4) × 10−15 cm3 molecule−1 s−1 [17–19]
[C•H2OO•] ∗ + O3 → HCHO + 2 O2 (R12)
To validate this hypothesis, the 2-butanol concentra-
tion was varied from 0 to 3.53 × 1015 molecule cm−3
whereas the concentration of sBF formed was recorded
at a fixed reaction time (164 2 s). All other experi-
mental conditions (especially [4M1P]0 and [O3]0) were
International Journal of Chemical Kinetics DOI 10.1002/kin.20934

604
BRAURE ET AL.
of indirect detection of a CI was recently reported in
the ozonolysis of 2,3-dimethyl-2-butene by detecting
H2SO4 formed in the CI + SO2 reaction [24]. The
same CI + SO2 reaction was also used to determine
the yield of a stabilized CI formed in the ozonolysis of
β-pinene by calculating the consumption of SO2 [25].
In a similar way, methanol was utilized as a stabilized
CI scavenger to suppress some reaction pathways and
put into evidence the key role of C•H2OO• in the for-
mation of oligomeric hydroperoxides and secondary
organic aerosol during ethylene ozonolysis [26].
CONCLUSION
The rate coefficients of ozonolysis of trans-2,2-
dimethyl-3-hexene (22dM3H) and trans-2,5-dimethyl-
3-hexene (25dM3H) were measured under pseudo–
first-order conditions, and the following values were
Figure 5 s-Butyl formate formation versus initial 2-butanol
concentration. 4M1P and O3 initial concentrations and reac-
tion time were kept constant. The solid line is a double ex-
ponential fit through the data. The dotted line is the highest
possible s-butyl formate concentration for these experimental
conditions if all the CI are converted to sBF (see text).
obtained (in cm3 molecule−1
10−17 for 22dM3H and (2.71
s
−1): (3.38
0.12) ×
0.26) × 10−17 for
25dM3H. This work represents the first investigation
of reaction products in the ozonolysis of 22dM3H,
25dM3H, and 4M1P in a laminar flow reactor.
From the performed mechanistic investigations,
primary carbonyl products from the ozonolysis of
22dM3H, 25dM3H, and 4M1P have been identified
and branching ratios are reported. Total yields of the
primary carbonyl products are very close to 100%,
which is in agreement with the well-known mecha-
nism of alkene ozonolysis.
In terms of products arising from CI reactions, the
sums of the yields are less than 20% for 22dM3H and
25dM3H ozonolysis, whereas reaching about 50% in
the 4M1P ozonolysis when using a CI scavenger. This
clearly shows an incomplete comprehension of the CI
fate, either due to CI losses on the reactor wall or due
to unknown reactions or unidentified products. It is in-
teresting to note, however, that the indirect detection
of the CI by the use of 2-butanol as a radical scav-
enger enabled to detect indirectly the stabilized CI and
give an upper limit of its yield in the 4M1P ozonol-
ysis. Although the use of such scavengers is a good
way to study the fate of a CI in laboratory studies,
these reactions do not play any significant role in the
atmosphere, where CI radicals react preferentially with
other species.
kept unchanged. The concentration of sBF formed was
then plotted versus [2-butanol]0 (Fig. 5).
An increase of sBF was observed with increasing
2-butanol concentrations from 0 to 4 × 1015 molecule
cm−3 whereas a double exponential fit suggests a
plateau could be reached for [2-butanol] > 4 ×
1015 molecule cm−3. The yield of sBF could be cal-
culated as the ratio of [sBF]formed to [O3]reacted. At the
plateau, a maximum yield of 46.2% was calculated
([sBF]formed = 1.59 × 1012 and [O3]reacted = 3.44 ×
1012 molecule cm−3). sBF being a coproduct of 3-
methyl-butanal (after a reaction of [C•H2OO•]* with
2-butanol), a maximum yield of ꢀ59% (equal to the 3-
methyl-butanal yield measured in this study) could be
expected providing that all CIs would be stabilized and
converted. However, it can be observed that the mea-
sured sBF yield is 13% lower than the maximum pos-
sible yield. Since this measured yield results from the
scavenging of the stabilized [C•H2OO•]* and further
decomposition of the hydroperoxide, we can suggest
(i) that part of the C•H2OO• is not stabilized and un-
dergoes other reactions in our experimental conditions
and/or (ii) that the transformation of the hydroperox-
ide into sBF is only partial. As reported by Hasson
et al. [23] and references therein, the indirect detection
of C•H2OO• through its capture by organic compounds
(acetaldehyde, formaldehyde, methanoic acid) has al-
ready been used in the literature to determine stabilized
Criegee intermediate (SCI) yields. A similar method
The role of the conformation in syn- or anti-position
remains an unresolved issue in the evolution of the
Criegee biradical. While syn-conformers enable the
formation of an unsaturated hydroperoxide through
1,5-isomerization, anti-conformers are supposed to
evolve to an excited carboxylic acid which can de-
compose following four different ways [6]:
International Journal of Chemical Kinetics DOI 10.1002/kin.20934

OZONOLYSIS OF A SERIES OF METHYLATED ALKENES
605
i. •OH + acyl radical,
ii. CO2 + alkane,
7. Taatjes, C. A.; Shallcross, D. E.; Percival, C. J. Phys
Chem Chem Phys 2014, 16, 1704–1718.
8. Duncianu, M.; Olariu, R. I.; Riffault, V.; Visez, N.;
Tomas, A.; Coddeville, P. J Phys Chem A 2012, 116,
6169–6179.
iii. H2O + alkenal, and
iv. CO + alcohol.
9. Schauer, J. J.; Kleeman, M. J.; Cass, G. R.; Simoneit,
B. R. T. Environ Sci Technol 1999, 33, 1578–1587.
10. Schauer, J. J.; Kleeman, M. J.; Cass, G. R.; Simoneit,
B. R. T. Environ Sci Technol 2001, 35, 1716–1728.
11. Hays, M. D.; Geron, C. D.; Linna, K. J.; Smith, N.
D.; Schauer, J. J. Environ Sci Technol 2002, 36, 2281–
2295.
The formation of nonoxidized products as observed
by Duncianu et al. [8] is only possible from syn-
conformers. Since these were not observed in the
present work, it suggests either the predominance
of anti-conformation for the large CI investigated or
the importance of other reaction pathways for syn-
conformers under our experimental conditions.
12. Grosjean, E.; Grosjean, D. Int J Chem Kinet 1996, 28,
461–466.
13. Grosjean, E.; Grosjean, D. Atmos Environ 1996, 30,
4107–4113.
14. Grosjean, E.; Grosjean, D. Environ Sci Technol 1996,
30, 2036–2044.
15. Grosjean, E.; Grosjean, D. Environ Sci Technol 1997,
31, 2421–2427.
16. Ho, S. S. H.; Ip, H. S. S.; Ho, K. F.; Dai, W. T.; Cao, J.;
Ting Ng, L. P. Aerosol Air Qual Res 2013, 13, 1151–
1160.
17. Kjaergaard, H. G.; Kurte´n, T.; Nielsen, L. B.; Jorgensen,
S.; Wennberg, P. O. J Phys Chem Lett 2013, 4, 2525–
2529.
18. Stone, D.; Blitz, M.; Daubney, L.; Howes, N. U. M.;
Seakins, P. Phys Chem Chem Phys 2014, 16, 1139–
1149.
Mines Douai participates in the Research Institute of Indus-
trial Environment (IRENI), which is financed by the Com-
munaute´ Urbaine de Dunkerque, the Nord-Pas de Calais
Regional Council, the French Ministry of Education and Re-
search, and European funds (FEDER). The CaPPA project
(Chemical and Physical Properties of the Atmosphere) is
funded by the French National Research Agency (ANR)
through the PIA (Programme dꢁInvestissement dꢁAvenir) un-
der contract “ANR-11-LABX-0005–01” and by the Regional
Council “Nord-Pas de Calais” and “European Funds for
Regional Economic Development” (FEDER). TB is grate-
ful for a Ph.D. grant from Armines. RIO, and CA thank
Mines Douai and the IRENI for Invited Researcher Fellow-
ships and also PN-II-PCE-2011–3–0471 Project, contract no.
200/05.10.2011.
19. Welz, O.; Eskola, A. J.; Sheps, L.; Rotavera, B.; Savee,
J. D.; Scheer, A. M.; Osborn, D. L.; Lowe, D.; Murray
Booth, A.; Xiao, P.; Khan, M. A. H.; Percival, C. J.;
Shallcross, D. E.; Taatjes, C. A. Angew Chemie, Int Ed
2014, 53, 4547–4550.
20. Vereecken, L.; Harder, H.; Novelli, A. Phys Chem Chem
Phys 2014, 16, 4039–4049.
21. Buras, Z. J.; Elsamra, R. M. I.; Green, W. H. J Phys
Chem Lett 2014, 2224–2228.
22. Wolff, S.; Boddenberg, A.; Thamm, J.; Turner, W. V.;
Ga¨b, S. Atmos Environ 1997, 31 (18), 2965–2969.
23. Hasson, A. S.; Orzechowska, G.; Paulson, S. E. J
Geophys Res-Atmos 2001, 106, 34131–34142.
24. Berndt, T.; Jokinen, T.; Sipila¨, M.; Mauldin III, R. L.;
Herrmann, H.; Stratmann, F.; Junninen, H.; Kulmala, M.
Atmos Environ 2014, 89, 603–612.
25. Ahrens, J.; Carlsson, P. T. M.; Hertl, N.; Olzmann, M.;
Pfeifle, M.; Wolf, J. L.; Zeuch, T. Angew Chem, Int Ed
2014, 53, 715–719.
BIBLIOGRAPHY
1. Atkinson, R.; Aschmann, S. M. Environ Sci Technol
1993, 27, 1357–1363.
2. Kroll, J. H.; Seinfeld, J. H. Atmos Environ 2008, 42
(16), 3593–3624.
3. Kanakidou, M.; Seinfeld, J. H.; Pandis, S. N.; Barnes,
I.; Dentener, F. J.; Facchini, M. C.; Van Dingenen,
R.; Ervens, B.; Nenes, A.; Nielsen, C. J.; Swietlicki,
E.; Putaud, J. P.; Balkanski, Y.; Fuzzi, S.; Hjorth, J.;
Moortgat, G. K.; Winterhalter, R.; Myhre, C. E. L.;
Tsigaridis, K.; Vignati, E.; Stephanou, E.G.; Wilson,
J. Atmos Chem Phys 2005, 5, 1053–1123.
4. Criegee, R. Angew Chem, Int Ed Engl 1975, 14, 745–
752.
5. Atkinson, R.; Arey, J. Atmos Environ 2003, 37, 197–
219.
26. Sakamoto, Y.; Inomata, S.; Hirokawa, J. J Phys Chem A
2013, 117, 12912–12921.
6. Johnson, D.; Marston, G. Chem Soc Rev 2008, 37, 699–
716.
International Journal of Chemical Kinetics DOI 10.1002/kin.20934