Development of a new flow reactor for kinetic studies. Application to the ozonolysis of a series of alkenes

Article abstract of DOI:10.1021/jp211480x

A new flow reactor has been developed to study ozonolysis reactions at ambient pressure and room temperature (297 ± 2 K). The reaction kinetics of O₃ with 4-methyl-1-pentene (4M1P), 2-methyl-2-pentene (2M2P), 2,4,4-trimethyl-1-pentene (tM1P), 2,4,4-trimethyl-2-pentene (tM2P) and α-pinene have been investigated under pseudo-first-order conditions. Absolute measurements of the rate coefficients have been carried out by recording O₃ consumption in excess of organic compound. Alkene concentrations have been determined by sampling adsorbent cartridges that were thermodesorbed and analyzed by gas-chromatography coupled to flame ionization detection. Complementary experimental data have been obtained using a 250 L Teflon smog chamber. The following ozonolysis rate coefficients can be proposed (in cm³ molecule^-1 s^-1): k_4M1P = (8.23 ± 0.50) × 10^-18, k_2M2P = (4.54 ± 0.96) × 10^-16, k_tM1P = (1.48 ± 0.11) × 10^-17, k_tM2P = (1.25 ± 0.10) × 10 ^-16, and k_α-pinene = (1.29 ± 0.16) × 10^-16, in very good agreement with literature values. The products of tM2P ozonolysis have been investigated, and branching ratios of (21.4 ± 2.8)% and (73.9 ± 7.3)% have been determined for acetone and 2,2-dimethyl-propanal, respectively. Additionally, a new nonoxidized intermediate, 2-methyl-1-propene, has been identified and quantified. A topological SAR analysis was also performed to strengthen the consistency of the kinetic data obtained with this new flow reactor.

Full text of DOI:10.1021/jp211480x

Article
pubs.acs.org/JPCA
Development of a New Flow Reactor for Kinetic Studies. Application
to the Ozonolysis of a Series of Alkenes
Marius Duncianu,†,‡ Romeo Iulian Olariu,§ Veronique Riffault,*,†,‡ Nicolas Visez,†,∥ Alexandre Tomas,†,‡
́
and Patrice Coddeville†,‡
†Universite
́
Lille Nord de France, F-59000, Lille, France
‡EMDouai, CE, F-59508 Douai, France
§Faculty of Chemistry, “Al. I. Cuza” University of Iasi, 11 Carol I, 700506, Iasi, Romania
∥Physico-Chimie des Processus de Combustion et de l’Atmospher
̀ ́
e, UMR CNRS 8522, Universite de Lille 1,
59655 Villeneuve d’Ascq, France
ABSTRACT: A new flow reactor has been developed to study ozonolysis
reactions at ambient pressure and room temperature (297 2 K). The reaction
kinetics of O3 with 4-methyl-1-pentene (4M1P), 2-methyl-2-pentene (2M2P),
2,4,4-trimethyl-1-pentene (tM1P), 2,4,4-trimethyl-2-pentene (tM2P) and
α-pinene have been investigated under pseudo-first-order conditions. Absolute
measurements of the rate coefficients have been carried out by recording O3
consumption in excess of organic compound. Alkene concentrations have been
determined by sampling adsorbent cartridges that were thermodesorbed and
analyzed by gas-chromatography coupled to flame ionization detection.
Complementary experimental data have been obtained using a 250 L Teflon
smog chamber. The following ozonolysis rate coefficients can be proposed
(in cm3 molecule−1 s−1): k4M1P = (8.23 0.50) × 10−18, k2M2P = (4.54
0.96) × 10−16, ktM1P = (1.48 0.11) × 10−17, ktM2P = (1.25 0.10) × 10−16
,
and kα‑pinene = (1.29 0.16) × 10−16, in very good agreement with literature values. The products of tM2P ozonolysis have been
investigated, and branching ratios of (21.4 2.8)% and (73.9 7.3)% have been determined for acetone and 2,2-dimethyl-propanal,
respectively. Additionally, a new nonoxidized intermediate, 2-methyl-1-propene, has been identified and quantified. A topological SAR
analysis was also performed to strengthen the consistency of the kinetic data obtained with this new flow reactor.
study confirms their versatility and the capacity of also using
them in kinetic studies.
INTRODUCTION
■
A significant part of the total primary volatile organic carbon
(VOC) emissions to the atmosphere is formed by alkenes, and
their ozonolysis continues to receive attention, due to its
important role in atmospheric chemistry. The alkene ozonolysis
rate coefficients are useful as input data for comprehensive
atmospheric chemical models describing air quality at urban or
regional scales, where ozonolysis may be the most important
sink for alkenes.
Some of the alkenes investigated in this work (2M2P and
4M1P) have been identified in common processes such as
tailpipe emissions of volatile hydrocarbons from gasoline-
powered motor vehicles,17 residential fireplace combustion of
wood,18 or in natural fires of foliage, litter, and herbaceous
matter.19 α-Pinene is a well-known biogenic compound that
forms secondary organic aerosols (SOA) in the atmosphere
following oxidation processes.
Under atmospheric conditions, the reactions of ozone with
alkenes provide free radicals and reactive intermediates. Gas-
phase ozonolysis reactions involving alkenes can be a significant
source of hydroxyl radicals in the atmosphere1−3 and precursors
of carbonyls or carboxylic acids.4−6
In this work, we have studied the ozonolysis kinetics of four
methylated pentenes, i.e., 4-methyl-1-pentene (4M1P),
2-methyl-2-pentene (2M2P), 2,4,4-trimethyl-1-pentene (tM1P),
2,4,4-trimethyl-2-pentene (tM2P), and α-pinene in a flow
reactor newly developed in our laboratory, aiming to investigate
the primary particle formation steps. The literature describes
a limited number of flow reactors used in the study of the
formation, evolution, or aging of aerosols,7−16 and the present
It should be pointed out that the four pentenes have been
selected for the validation step because their O3 reaction rate
coefficients are already known in the literature; and yet, while
O3 + alkene reaction kinetics and mechanisms are generally
investigated in environmental simulation chambers with large
reaction times, the present work on the pentenes is to the best
of our knowledge the first one carried out at much shorter
reaction times.
Special Issue: A. R. Ravishankara Festschrift
Received: November 29, 2011
Revised: January 23, 2012
Published: January 23, 2012
© 2012 American Chemical Society
6169
dx.doi.org/10.1021/jp211480x | J. Phys. Chem. A 2012, 116, 6169−6179

The Journal of Physical Chemistry A
Article
Moreover, current uncertainties on such alkene oxidation
mechanisms result from the incomplete description of the
excited Criegee intermediate decay channels, which determine
the extent to which each of these channels generate radicals.20
α-Pinene ozonolysis was also studied, since the reaction system
is more complex and involves heterogeneous processes due to
the formation of SOA.21,22 Besides the main homogeneous
process, the presence of particulate matter could induce supple-
mental reactions in the particulate phase or adsorptive processes.
To add more confidence to the data obtained with the flow
tube, similar kinetic studies have been performed on the pen-
tenes using a Teflon bag smog chamber. Finally, a structure−
activity relationship (SAR) analysis has been applied to prove
the consistency of the obtained results.
on two different lines and to ensure homogeneity before their
entrance into the reaction volume. The passage of the fluid into
the reactor is assured by a sintered glass wall welded to the
mixing chamber filled with glass beads (diameter: 2 mm), thus
providing optimal conditions to rapidly achieve a laminar flow
regime. The injection head has been designed to allow good mixing
of the reactive gases at transit times of seconds (1−2.5 s) and to
support the formation of a laminar flow (Reynolds number,
Re ∼ 20−50) in the reactor. The laminar flow in the reactor
ensures a stationary mode as a function of the reaction volume,
at atmospheric pressure and room temperature (T = 297 2 K).
At the bottom of the reactor, two exhaust lines consisting of
1/4-in. glass tubes are used to collect samples. The central one
reaches the laminar flow region roughly 15 cm above the bottom
of the reactor. The reactor height is defined by the distance
between the injector head and the central sampling point. The
volume difference between the central sampling line and the
lateral one was taken into account to correct the reaction time for
each reactor height.
1. EXPERIMENTAL METHODS
1.1. Flow Reactor Setup for Kinetic Measurements.
The new aerosol flow tube is schematically represented in
Figure 1. The cylindrical reactor (Pyrex tube with 1 m length
́
Both Reynolds (Re) and Peclet (Pe) dimensionless numbers
can be calculated in order to characterize transport phenomena
in fluid flows:
LU
D
Pe =
= Re × Sc
(I)
where L is the characteristic length, U is the average velocity, D
is the mass diffusion coefficient, and Sc is the Schmidt number
equal to the ratio between the kinematic viscosity and D. For all
experiments, the Pec
́
let numbers were inferior to 50, indicating
nonplug flows.23
A pulse tracer study similar to previous approaches24 was
performed using either O3 or CO to validate the calculation of
the contact time into the reactor (which is defined as the ratio
between the reactor height, hR, and the average velocity). The
velocity profile in laminar flow conditions presents a parabolic
distribution (zero at the wall, maximum at the center of the
reactor), implying that tracer molecules traveling at the center
of the cylinder will be detected first, while the average velocity
will correspond to the time of the first derivate (dO3/dt)
maximum. The concentration profiles of the tracer analytes
were monitored with 1 s time resolution measurements using
spectroscopic analyzers (Models 48C and 49C, Thermo
Environmental Instruments) for various heights of the reactor
and several characteristic flows. The example given in Figure 2a
Figure 1. Schematic of the experimental setup and instrumentation
(MFC: mass flow controller; hR: reactor height).
and 10 cm i.d.) has been designed to work with total flows of
about 1.5 to 5 L min−1, corresponding to reaction times ranging
between 30 s and ∼5 min. Its vertical position assures a
gravitational equilibrium of the flow, and a sliding injector
ending with a mixing chamber is used to introduce the reagents
Figure 2. (a) Concentration profiles of ozone used as a tracer analyte to characterize the reactor flow for three reactor heights. (b) Correction factor
between the calculated average residence time (tR_calc) and the measured residence time (tR_meas) for several heights of the reactor (hR) and several
characteristic bulk flows (in mL min−1) in the reactor in laminar flow conditions using two tracers. The broad line is the exponential regression fit of
all the data, while the dotted lines represent the 95% confidence interval. The gray area represents the Re/hR range of all the kinetic measurements.
6170
dx.doi.org/10.1021/jp211480x | J. Phys. Chem. A 2012, 116, 6169−6179

The Journal of Physical Chemistry A
Article
shows the ozone time profiles and their first derivatives
(Δ[O3]/Δt, which were fitted using Gaussian curves) for a
3150 mL min−1 flow. The average bulk velocity was corrected
using the ratio between the calculated residence time, tR_calc, as
derived from the laminar flow theory and the measured
residence time, tR_meas, spanning from 28 to 177 s for the tracer
study. The ratio was close to unity for the most part of the flow
range (described by Re, values ranging from 20 to 64 for the
tracer study) whatever the reactor heights (hR, expressed in cm),
as seen in Figure 2b. The gray area represents the Re/hR
range of all the kinetic measurements. The response time of the
tracer analyzer and the diffusion coefficient of the analyte
due to the concentration gradient were taken into account
to correct the estimation of the measured residence time. The
diffusion correction was under 5% for most experiments and
never exceeded 10%.
For alkene ozonolysis, a continuous gas flow of the selected
VOC is achieved using a pressurized canister, with a few
milliliters per minute being released through a mass flow
controller (MFC) in a premixing chamber. Purified air is used
as a carrier gas and ensures the dilution of the VOC at the
requested level of concentration. On the other hand, an ozone
generator (Model 146i, Thermo Scientific) provides a con-
trolled flow of oxidant which reaches the injection head via a
separate line. The controlled flows of ozone and volatile organic
compound enable predefined and stable concentrations at the
top of the reactor to be obtained.
Initial concentrations in the flow reactor were (0.81−8.1) ppmv
for the alkene (1 ppmv = 2.46 × 1013 molecules cm−3 at 298 K
and 1 atm) and (61−374) ppbv for ozone.
The outlet analytical chain includes an online O3 analyzer
(Model 49C, Thermo Environmental Instruments) based on
UV absorption spectroscopy. The absence of interferences
coming from the alkenes investigated in the present study were
checked by considering their UV absorption spectra, as well as
by measurement tests in the absence of ozone. As for possible
interferences from the ozonolysis products, alkene concen-
trations were increased progressively until the total con-
sumption of ozone in the reactor was reached, while the O3
analyzer presented a background value similar to the back-
ground value of purified air.
1.2. Smog Chamber Setup. In order to obtain additional
confirmation of the kinetic parameters obtained with the aerosol
flow reactor, complementary experiments were performed on
the four pentenes in a smog chamber that has been described
previously.26 Only the aspects specific to the current experi-
ments are presented below.
The reaction chamber consists of a Teflon bag with a volume
of about 250 L working at atmospheric pressure, (297 2) K,
and in the dark. Teflon tubes are used for the introduction of
the reactants and the sampling by the analytical instruments.
Ozone was produced by flowing purified air through a UV lamp
ozonizer (Model 165, Thermo Environmental Instrument),
which was directly injected in the Teflon chamber at a flow rate
of about 2.8 L min−1. The ozone concentration was determined
as a function of reaction time by using an online ozone analyzer
(Model 41M, Environnement S.A.) with a time resolution of
about 20 s. The analyzer sampling flow rate and precision were
0.90 L min−1 and 5 ppb, respectively. Initial O3 concentrations
in the reaction chamber were in the range 70−300 ppbv.
All the investigated alkenes being liquid reagents, they were
injected with microliter syringes into a small heated glass cell
placed on a secondary input line of clean air in order to ensure
a complete and rapid evaporation. Initial alkene concentrations −
calculated by taking into account the injected volume of
alkene and the total volume of the bag − were in the range
1.1−3.4 ppmv. Additional control experiments were carried out
to check the stability of the reactants in the Teflon bag,
showing that all the selected olefins and O3 do not suffer any
significant wall losses while in the chamber.
In a typical experiment, ozone was first added into the Teflon
bag. At time zero, the pentene was injected in the reactor using
a high flow of purified air (∼15 L min−1), thus ensuring a rapid
dilution of the organic reactant in the ozone-containing Teflon
chamber. For each investigated compound, one experiment was
performed using cyclohexane as an OH scavenger, and no
significant difference was observed in the ozone decay and,
therefore, in the ozonolysis rate coefficient.
1.3. Absolute Rate Measurements. Kinetic Determi-
nation Formalism. Rate coefficients for the investigated
ozone reactions were determined in both reactors by
monitoring the O3 decay rates in the presence of excess and
known concentrations of the alkene (ALK). As the ozone loss
caused by wall deposition was shown to be negligible, the
temporal profile of O3 is governed by a unique process:
The volume of the flow reactor can be modified by moving
up and down the sliding injection head, thus allowing one to
monitor the evolution of the reagents and products at various
reaction times and to estimate kinetic and mechanistic
parameters. Different reaction times may also be achieved by
changing the total flow rate for a given distance between the
movable injection head and the sampling outlet.
The side line was used to sample adsorption cartridges, to
determine the ozone concentration and evacuate the main flow,
while the central one will be used in further studies for aerosol
sampling in isokinetic conditions. The cartridges filled with
three different adsorbents (Carbopack C, Carbopack B, and
Carbosieve SIII in 1:2:1 packing ratio from Supelco) were
sampled at 50 mL min−1 before injecting ozone to determine
the initial concentration of the VOC in excess. A cartridge
bypass was installed to continuously purge the Teflon line with
the exhaust air of the reactor before and after sampling. The
reactant concentration was estimated using calibration curves
obtained with cartridges doped with known concentrations of
analytes. Six cartridges were sampled in the absence of ozone at
different heights in the reactor to check for the concentration of
the organic reactant, the potential wall losses (found negligible
for all compounds), as well as the reproducibility and accuracy
of the sampling. It should be noted that in the presence of
ozone, a KI-coated copper tube upstream from the sampling
cartridge was used as a denuder to avoid the loss of the alkenes
through O3 reaction on the adsorbent.25
For the kinetic experiments in the flow reactor, sampled
cartridges were then thermally desorbed (within 24 h) in a two-
stage thermodesorber (Markes Unity2) and analyzed by GC/
FID (Agilent 7890A). A Varian capillary column (CP-Sil 5 CB
50 m, 0.32 mm, 1.2 μm) was used with the following tempera-
ture program: 35 °C, 15 min isothermal, 4 °C min−1 to 125 °C,
then 20 °C min−1 to 250 °C and held for 6 min. Helium was
used as the carrier gas with a column flow of ∼3 mL min−1.
O + ALK → products
3
leading to the following rate equation:
−d ln[O ] /dt = k
3 t ALK
× [O ] × [ALK]
3 t
(II)
t
6171
dx.doi.org/10.1021/jp211480x | J. Phys. Chem. A 2012, 116, 6169−6179

The Journal of Physical Chemistry A
Article
where [ALK]t and [O3]t are the concentrations in alkene and
ozone at time t and kALK is the ozonolysis rate coefficient.
With the initial alkene concentration [ALK]0 being in excess
by at least a factor of 10 over the initial ozone concentration,
the alkene concentration remains essentially constant through-
out the reaction time. This approximation is only valid if there
are no other additional processes, such as competitive reaction,
wall loss, or/and OH radical reactions that could contribute to
the alkene decay except the ozonolysis.
On the basis of literature data, it is well-known that the
reaction of ozone with unsaturated species can produce OH
radicals,2,27−33 which may further react with the organic
reactant, resulting in a significant change in the alkene concen-
tration. Complementary experiments were thus performed with
cyclohexane (in smog chamber and flow tube experiments) or
CO (in flow tube experiments) added as OH scavenger. CO
was preferred because it is an effective scavenger of OH34,35
without causing interference in the product study. After data
evaluation for experiments with and without OH scavenger,
no significant difference was observed in the rate coefficients
given the uncertainties. It should also be noted that the reac-
tions were generally carried out in a large excess of the organic
reactant, so that the possible additional loss of the alkene by
reaction with OH would have a limited effect on the alkene
concentration.
air was added in the reactor flow during ozone level mea-
surements, and its concentration was corrected for dilution.
Gaseous samples were collected online (Thermo Desorption
System, Gerstel) on adsorbent tubes kept at 0 °C, trapping the
volatile organic compounds at a rate of 50 mL min−1 for a total
volume of 500 mL. Once collected, the analytes are thermo-
desorbed and transferred toward a cryogenic trapping (Cooled
Injection System, CIS) capillary tube of small diameter (2 mm)
filled with a few milligrams of adsorbent (Carbopack B, ∼ 8 mg)
to focus analytes before entering the chromatographic column.
The chromatographic and spectrometric analysis was further
performed by a GC/FID-MS (Agilent 6890N/5975B) instru-
ment equipped with a DB-5 ms (123−5563; 60 m × 0.32 mm,
1 μm) column. The separation was carried out at a 4 mL min−1,
22.1 psi, He carrier flow. The temperature program was started
in cryogenic conditions at 0 °C hold for 5 min, followed by a
3 °C min−1 ramp until 90 °C and a steep ramp of 20 °C min−1
up to 250 °C (5 min hold).
Reproducibility and breakthrough volume tests were equally
performed in the absence of ozone and in ozonolysis conditions
for different sampling volumes. The results showed an excellent
correlation between the sampling volume and the quantity
of analyte as measured by the flame ionization detector (FID)
(r2 > 0.998) with a slope value greater by more than 1 order of
magnitude than the value of the intercept, suggesting the
absence of breakthrough and adsorbent rinsing effects.
On the basis of the above observations, the pseudo-first-
order rate coefficient, k′ALK, may be introduced as
1.5. Reagents. Chemicals were all commercially available
and used as received without further purification. α-Pinene
(>98%), 4M1P (98%) 2,2-dimethyl-propanal (96%), acetone
(>99.8%), and cyclohexane (99.5%) were purchased from
Sigma-Aldrich. 2-Methyl-1-propene (10 ppm in N2) was
purchased from Messer-Griesheim. tM2P (>98%), tM1P
(99%), and 2M2P (> 99%) were from Janssen Chimica.
Purified air was produced by an air purifier (AZ 2020,
Claind) and was characterized by a relative humidity of less
than 5% and an organic carbon content of max. 0.1 ppbv.
k′
= k × [ALK]
ALK 0
(III)
(IV)
(V)
ALK
Equation II becomes
−d ln[O ] /dt = k′
× [O ]
ALK 3 t
3 t
Integrating the rate law yields
ln([O ] /[O ] ) = k′ × t
3 0 3 t
ALK
According to the last relation, the slope of the linear fitting of
ln([O3]0/[O3]t) versus t gives the pseudo-first-order rate
coefficient k′ALK, whose value can be divided by the initial
alkene concentration to retrieve the ozonolysis rate coefficient
kALK through eq III.
Additional experiments were performed, in the case of
α-pinene, in reversed conditions, i.e., in excess of O3 and moni-
toring the decay rates of α-pinene. Though giving comple-
mentary results, this strategy was not pursued since the approach
was more time-consuming and generated higher uncertainties in
the α-pinene concentrations (and thus in the k′ values) as
compared with the general case described above.
1.4. Product Study in the Flow Reactor. Because of its
higher rate coefficient and the fact that the double bond is not
located on a terminal carbon, which facilitates product
identification with the available analytical techniques, tM2P
was chosen to validate the flow reactor from a mechanistic
point of view.
2. RESULTS AND DISCUSSION
2.1. Pentenes: Ozonolysis Rate Coefficients. The alkene
compounds investigated in the present study were chosen
in order to fill a wide range of rate coefficient values as a
certification of the versatility of the flow reactor. In addition,
literature data were available for these alkenes (at least one
study). Experimental conditions and results for both reactors
are presented in Table 1. Plots of ln([O3]t/[O3]0) versus time
obtained with the flow reactor are displayed in Figure 3 for
each of the four pentenes investigated at different initial con-
centrations. The uncertainties (1σ) quoted in the graphs of
Figure 3 represent the sum of the variation coefficients (V ≡
sx/x, where sx is the standard deviation of a set of samples xi and
̅
x is its mean) of the recorded series of values for [O3]0 and
̅
[O3]t as derived from the equation
The alkene ([tM2P]0 = (3.9−4.3) × 1012 molecules cm−3)
and ozone ([O3]0 = 5.1 × 1012 molecules cm−3) were
introduced into the flow reactor in the presence of carbon
monoxide ([CO]0 = 3 × 1016 molecules cm−3) as the OH
radical scavenger. A total flow of 900 mL min−1 was chosen in
order to obtain greater reaction times, required to ensure the
conversion of a sufficient amount of reagents at the sampling
time.
[O ]
3 0
⎧
⎩
⎫
⎬
⎭
⎛
⎞
⎟
⎠
⎨
σ ln⎜
[O ]
⎝
⎛
3 t
2
2
2
2
⎞
⎛
⎝
⎞
σ[O ]
σ[O ]
[O ]
σ[O ]
3 0
σ[O ]
3 0
3 t
3 t
⎜
⎜
⎟
⎟
⎜
⎜
⎟
⎟
=
+
+ 2 ×
×
[O ]
[O ]
[O ]
⎝
3 0 ⎠
3 t ⎠
3 0
3 t
(VI)
Since the sampling rate of the ozone analyzer (1300 mL min−1)
is higher than the total flow rate, an additional amount of dry
Nonweighted linear least-squares fits of the data yielded
high correlation coefficients and near-zero intercepts, and
6172
dx.doi.org/10.1021/jp211480x | J. Phys. Chem. A 2012, 116, 6169−6179

The Journal of Physical Chemistry A
Article
Table 1. Experimental Conditions and Results Obtained for the Ozonolysis Reactions in Excess of Alkenes at Ambient
Temperature and Pressure
a
b
c
d
SC: smog chamber; FR: flow reactor. Experiments in excess of alkene. Experiments in excess of ozone. Slope of the weighted linear least-squares
fit forced through zero for flow reactor and smog chamber data; the uncertainty was calculated at the 95% confidence level and only represents
statistical uncertainty.
Figure 3. Plots of ozone consumption versus time for (a) 4M1P, (b) 2M2P, (c) tM1P, and (d) tM2P in the flow reactor experiments.
the slopes led to the pseudo-first-order rate coefficients k′ALK
where quoted uncertainties correspond to the 1σ precision
of the fit.
Plotting k′ALK versus [ALK]0 for the flow reactor (open
marks) and the Teflon chamber (full marks) data (Figure 4)
gives the ozonolysis rate coefficients, kALK, for all the target
compounds, according to eq III. The uncertainties in Figure 4
represent one standard deviation (1σ) for both axes where the
,
errors on concentrations correspond to 1σ (repeatability) for
flow reactor data, and was calculated using the following
equation for smog chamber data:
2
2
⎛
⎜
⎞
⎟
2
⎛
⎞
σ
ρ
M
V
V
V
ALK
i
i
2
σ
=
N
+
σ
V
total
⎜
⎝
⎟
⎠
ALK
A
2
4
⎜
⎝
⎟
⎠
V
ALK
total
total
(VII)
6173
dx.doi.org/10.1021/jp211480x | J. Phys. Chem. A 2012, 116, 6169−6179

The Journal of Physical Chemistry A
Article
Figure 4. Plots of the pseudo-first-order rate coefficient k′ALK versus alkene concentration for (a) 4M1P, (b) 2M2P, (c) tM1P, and (d) tM2P. Open
squares: flow reactor experiments; filled circles: smog chamber experiments. Straight lines are linear regression fits of all the data, and curved lines are
the 95% confidence intervals.
with ρALK and MALK being the density and molar mass of the
alkene, respectively, NA is the Avogadro number, Vi the injected
volume of the compound, Vtotal is the total volume of the smog
chamber, σVi is the precision of the microsyringe, and σVtotal is
obtained by propagating uncertainties considering flows and
times for filling up the Teflon bag.
Fitting was made taking into account the weighting of the
standard deviation on k′ALK values. Table 1 shows the kALK
values obtained together with the 95% confidence interval (also
represented as dashed lines in Figure 4). This error was
combined with the one estimated for each compound
concentration given the repeatability of the measurements
(2M2P: 21%; 4M1P: 6%; tM1P: 7%; tM2P: 8%) to give the
overall uncertainty reported in Table 2 for each rate coefficient.
A very good agreement between the kinetic data obtained
from two very different systems (flow tube and smog chamber)
and for a very wide range of reaction times can be observed for
all compounds (Figure 5 and Table 2), as well as a good
comparison with literature data (Table 2).
values of (7.3 1.4) × 10−18 cm3 molecule−1 s−1 at 287 K36
and (7.9 0.3) × 10−18 cm3 molecule−1 s−1 at 292 K.37
While the compounds with substituted methyl groups at the
double bond (2M2P, tM2P) present larger rate coefficients, the
compounds with terminal double bond (4M1P, tM1P) are
clearly less reactive toward ozonolysis. Thus, the present data
support very well the increase of the reactivity of the ozone
electrophilic addition at the unsaturated carbon−carbon with the
degree of substitution, as stated in the literature.5,6,38,39
Furthermore, the presence of methyl groups at the vicinal
carbon of the double bond reduces the ozonolysis rate coefficient
as observed in the comparison between 2M2P and tM2P or
tM1P and 2,3,3-trimethyl-1-butene (see Table 4), probably due
to the contribution of steric effects.36 Finally, as also observed in
the literature, this reactivity is not strongly influenced by the size
of the substituent (comparison between 4M1P and tM1P).
2.2. α-Pinene: Ozonolysis Rate Coefficient. The ozono-
lysis of α-pinene is an important source of oxidized species
in the atmosphere, which contribute significantly to the atmos-
pheric formation of SOA. The O3 reaction rate coefficient has
been determined by many previous studies and has been re-
viewed by Atkinson et al. (2006).40 The ozonolysis of α-pinene
was carried out only in the flow reactor.
2M2P appears to be the most reactive compound toward
ozone among those investigated, which can be explained by the
alkyl substituents increasing the activity of the unsaturated
bond together with a minimal inhibiting steric effect due to the
lack of massive substituents. The rate coefficient derived from
Data were analyzed using the same procedure as described
previously for the methylated pentenes. Graphs of ln([O3]t/
[O3]0) versus time are presented in Figure 6a for different
initial concentrations of α-pinene in excess. Two experiments
were also performed in excess of ozone ([O3]0 = 1.4 and
2.6 ppmv). The corresponding slopes have been plotted
the flow tube and the smog chamber experiments (k2M2P
=
4.54 × 10−16 cm3 molecule−1 s−1) and the value from McGillen
et al. (2008) (k2M2P = 4.06 × 10−16 cm3 molecule−1 s−1) present
an excellent agreement. On the contrary, the 4M1P kinetics is
the slowest in the studied series (k4M1P = 8.23 × 10−18 cm3
molecule−1 s−1), confirming previous studies that determined
versus [α-pinene]0 or [O3]0 in Figure 6b. A value of kα‑pinene
=
(1.29 0.16) × 10−16 cm3 molecule−1 s−1 has been obtained
6174
dx.doi.org/10.1021/jp211480x | J. Phys. Chem. A 2012, 116, 6169−6179

The Journal of Physical Chemistry A
Article
Table 2. Summary of Results and Comparison with
Literature Data
T
kALK × 10−17
(K) (cm3 molecule−1 s−1
experimental
setup
a
alkene
4M1P
)
reference
296
1.06
SC; O3 excess,
GC/FID
Cox and Penkett
(1972)50
287
0.73 0.14
SC; VOC excess; Grosjean and
OA
Grosjean
(1995)36
292
297
297
297
297
295
297
297
297
290
297
297
297
297
297
0.79 0.03
SC; VOC excess; Leather et al.
OA
(2010)37
b
0.92
Leather et al.
(2010)37
c
1.28 0.09
0.82 0.05
FR; VOC excess; this work
OA
c
SC; VOC excess; this work
OA
c
0.823 0.050
40.6 4.9
FR and SC; VOC this work
excess; OA
2M2P
SC; VOC excess; McGillen et al.
OA
(2008)48
Figure 5. Comparison between the rate coefficients obtained from
flow reactor (kFR) and smog chamber (kSC) experiments plotted with
their overall uncertainty.
c
40.8 8.6
FR; VOC excess; this work
OA
c
46.2 9.7
SC; VOC excess; this work
OA
c
In the case of α-pinene, the rate coefficient obtained in the
present work is in very good agreement with the recommended
value.40 A recent kinetic study41 involving the use of a similar
flow reactor also reports a kα‑pinene value at the upper limit of the
IUPAC recommendation (1.1 0.1 × 10−16 cm3 molecule−1
s−1), consistent with our study.
2.3. tM2P: Product Study. Cartridges were sampled at
several reaction times in order to follow the formation of the
reaction products and the consumption of reagents. As
expected, the two main stable oxidation products were
identified by comparison with NIST mass spectra as acetone
and 2,2-dimethyl-propanal, and their temporal distribution
profiles are presented in Figure 7. A 30% consumption of tM2P
was observed at the maximum reaction time. For all reaction
times, there is a close-to-unity sum of the branching ratios
((95.3 5.2)% on average), with mean yields of (21.4 2.8)%
45.4 9.6
FR and SC; VOC this work
excess; OA
tM1P
0.97 0.22
SC; VOC excess; Leather et al.
OA
(2010)37
b
1.47
Leather et al.
(2010)37
c
1.65 0.15
FR; VOC excess; this work
OA
c
1.48 0.10
SC; VOC excess; this work
OA
c
1.48 0.11
FR and SC; VOC this work
excess; OA
tM2P
13.9 3.4
SC; VOC excess; Grosjean and
OA
Grosjean
(1996)39
c
297
297
297
13.1 1.1
FR; VOC excess; this work
OA
c
12.4 1.0
SC; VOC excess; this work
OA
c
12.5 1.0
FR and SC; VOC this work
excess; OA
and (73.9
7.3)% for acetone and 2,2-dimethyl-propanal,
respectively (Table 3). A previous FEP Teflon chamber study
at ambient temperature and atmospheric pressure42 reported
formation yields of (19 1)% and (84 4)% for acetone and
2,2-dimethyl-propanal, respectively, in excellent agreement with
our work.
α-pinene 298
9.0
IUPAC data
Atkinson et al.
(2006)40
c
297
12.9 1.6
FR; VOC excess; this work
OA
a
SC: smog chamber; FR: flow reactor; OA: ozone analyzer; GC/FID:
gas chromatography/flame ionization detection. Values calculated
using the proposed Arrhenius equations at 297 K. Overall uncertainty
taking into account the 95% confidence interval on the slope of the
weighted linear least-squares fit and the uncertainty on the
concentration of the compound in excess using the propagation of
uncertainty approach.
b
The experimental results are supportive of the general
Criegee mechanism for alkene ozonolysis, in agreement with
the literature,42−45 which consists in the concomitant formation
of primary carbonyl compounds (2,2-dimethyl-propanal and
acetone in the case of tM2P ozonolysis), and other stable
oxidized reaction products from the Criegee intermediates such
as hydroxy-acetone, methyl glyoxal, and formaldehyde. The
dual-energy-rich Criegee radicals can follow collision-stabiliza-
tion processes or unimolecular decomposition processes42,43,46
to become low energy chemical entities. When the biradical
involves a t-butyl substituent, intramolecular migration of a
hydrogen atom leading to the formation of a hydroperoxide is
not possible, thus implying the absence of hydroxy-carbonyl
and α-dicarbonyl products.6
c
for the slope of the weighted linear least-squares fit forced
through zero. Figure 6b also shows the IUPAC recommended
value of 9.0 × 10−17 cm3 molecule−1 s−1 from Atkinson et al.
(2006).40 A repeatability of 12% for α-pinene concentrations is
taken into account for the estimation of the overall uncertainty
reported in Table 2. Very good agreement could be observed
between the results obtained in excess of α-pinene and those
obtained in excess of ozone. The larger error bars in k′α‑pinene may
be related to the high concentrations of O3 used, which limits the
efficiency of the KI scrubbers used upstream of the sampled
cartridge, and thus the repeatability of the measurements.
In addition to the primary reaction products, the study
identified and quantified a nonoxidized product, 2-methyl-1-
propene, with an average contribution of (11.9 2.4)%, and
the profiles are reported in Table 3 and Figure 7 for different
reaction rates. A new reaction mechanism is proposed involving
6175
dx.doi.org/10.1021/jp211480x | J. Phys. Chem. A 2012, 116, 6169−6179

The Journal of Physical Chemistry A
Article
Figure 6. Kinetics of α-pinene ozonolysis. (a) Plots of ozone consumption versus time (excess of α-pinene). (b) Plot of the pseudo-first-order rate
coefficient k′ALK versus alkene concentration (excess of α-pinene, open circles) or versus ozone concentration (excess of ozone, filled circles) and
comparison with the IUPAC recommendation40 (dotted line). The straight line is the linear regression fit, and the gray area represents the 95%
confidence interval. The dashed lines represent the uncertainty limits given by the IUPAC (i.e., 5.7 and 14.3 × 10−17 cm3 molecule−1 s−1).
Figure 8. Suggested mechanism for the formation of 2-methyl-2-
propene via the Criegee intermediate.
presented in the literature regarding the SAR analysis for alkene
ozonolysis, involving quantum molecular orbital calculations47
or the topological SAR method.48 The second method was
chosen, as it seems to present a robust molecular approach for
Figure 7. Time profiles of the tM2P ozonolysis products in the gas
the estimation of the alkene ozonolysis kinetics, and the most
phase in presence of an OH scavenger. Lines in the figure are drawn to
accurate in the prediction of the rate coefficients.
guide the eye and do not represent a model fit.
The topological SAR methodology used is based on the
influence of the structure of the molecule over the alkene
reactivity in the ozonolysis process and is described in detail
elsewhere.48 The calculation of the ozonolysis rate coefficient is
based on the estimation of the inductive (I) and steric (S)
effects around the unsaturated bond for each molecule and is
characterized by the index x calculated as
electron rearrangement in the Criegee intermediate containing
the t-butyl substituent, with the formation of OH and
HCO radicals and a nonoxidized alkene: 2-methyl-1-propene
(Figure 8). Its formation yields present a lowering tendency
with reaction time, suggesting competitive reactions of ozone
(since OH radicals are scavenged by CO) with either the
formed alkene or tM2P, and they remain invariably below the
formation yields of acetone, which is the molecule formed
simultaneously in the suggested Criegee reaction mechanism.
The two expected stable products from the ozonolysis reaction
of 2-methyl-1-propene are acetone and formaldehyde.
x = (yS) + I
(VIII)
where y is an empirical constant and is equal to −4.04. On
the basis of the analysis of a series of C2−C10 alkenes, McGillen
et al. (2008)48 derived the following correlation between
the logarithm of the room temperature rate coefficient log
2.4. Structure−Activity Relationship Analysis. An SAR
analysis based on the data from the present study (Table 2) and
literature data (Table 4) was applied to check the consistency
of the obtained rate coefficients. Different approaches are
kALK and the SAR index x: log kALK = (1.28
(18.14 0.07).
0.05)x −
Table 3. tM2P Ozonolysis Reaction: Alkene Consumption and Product Formation Yields for Different Reaction Times
reaction time (s) Δ[tM2P] (1011 molecules cm‑3) acetone yield (%) 2,2-dimethyl-propanal yield (%) Σ branching ratios (%) 2-methyl-1-propene yield (%)
152
204
309
414
518
5.28
6.37
8.49
10.6
11.7
23.6
25.1
67.1
66.8
90.7
91.9
14.1
14.5
19.8
75.3
95.1
11.8
18.9
75.9
94.8
9.7
19.4
84.5
103.9
9.2
average yield (%)
Grosjean and Grosjean (1997)42
21.4 2.8
73.9 7.3
95.3 5.2
11.9 2.4
19
1
84
4
103
4
6176
dx.doi.org/10.1021/jp211480x | J. Phys. Chem. A 2012, 116, 6169−6179

The Journal of Physical Chemistry A
Article
Table 4. Ozonolysis Rate Coefficients Predicted by the Topological SAR Algorithm48 Taking into Account Inductive (I) and
Steric (S) Effects in the Estimation of the SAR Index (x) and Available in the Literature49 by Averaging Values at 297 3 K,
unless Stated Otherwise
kliterature
kliterature
kpredicted
kpredicted
( × 10‑18 cm3
( × 10‑18 cm3
( × 10‑18 cm3
( × 10‑18 cm3
alkene
molecule‑1 s‑1) molecule‑1 s‑1)
I
S
x
alkene
molecule‑1 s‑1) molecule‑1 s‑1)
I
S
x
ethene
1.65
10.7
12.5
10.4
13.9
9.51
0.72
13.8
25.5
9.34
17.3
6.32
11.7
0
1
2
1
2
1
2
0.000 0.000
0.000 1.000
0.196 1.208
0.033 0.867
0.229 1.076
0.066 0.735
0.262 0.943
3-hexene, (Z)-
3-hexene, (E)-
144
157
38.4
120
120
55.1
2
2
2
0.066 1.735
0.066 1.735
0.131 1.470
propene
propene, 2-methyl
1-butene
3-hexene, 2,5-
dimethyl, (E)-
3-hexene, 2,2-
dimethyl, (E)-
40.4
11.4
55.1
2
0.131 1.470
1-butene, 2-methyl
1-butene, 3-methyl
1-heptene
7.37
7.37
75.0
75.0
7.37
75.0
7.37
13.8
18.8
25.5
1
1
2
2
1
2
1
2
3
4
0.053 0.787
0.053 0.787
0.105 1.574
0.105 1.574
0.053 0.787
0.105 1.574
0.053 0.787
0.000 1.000
0.196 1.104
0.392 1.208
b
1-butene, 2,3-
dimethyl
10.0
b
1-octene
12.5
89.8
b
4-octene, (Z)-
4-octene, (E)-
1-decene
1-butene, 3,3-
dimethyl
3.90
4.27
7.90
1
2
0.098 0.602
0.294 0.811
c
131
b
1-butene, 2,3,3-
trimethyl
7.75
8.00
114
5-decene, (Z)-
1-tetradecene
1,3-butadiene
isoprene
1-butene, 2-ethyl
10.9
3.02
11.7
5.34
2
2
0.262 0.943
0.327 0.678
22.0
6.57
12.6
25.6
1-butene, 3-methyl-
2-(1-methylethyl)
2-butene, (Z)-
140
223
263
263
486
899
2
2
3
4
0.000 2.000
0.000 2.000
0.196 2.208
0.392 2.417
1,3-butadiene, 2,3-
dimethyl
2-butene, (E)-
2-butene, 2-methyl
486
1,3-pentadiene,
(Z)-
27.8
43.2
80.0
314
374
60.3
60.3
81.9
263
263
3
3
4
4
4
0.000 1.500
0.000 1.500
0.196 1.604
0.000 2.000
0.000 2.000
2-butene, 2,3-
dimethyl
1200
1,3-pentadiene,
(E)-
1-pentene
9.57
13.1
7.37
13.6
1
2
0.053 0.787
0.249 0.996
1-pentene, 2-
methyl
1,3-pentadiene, 2-
methyl
b
1-pentene, 3-
methyl
3.80
4.99
5.82
9.22
8.49
1
1
2
2
0.086 0.655
0.073 0.707
0.281 0.863
0.288 0.835
2,4-hexadiene, (Z,
E)-
a
1-pentene, 4-
methyl
8.70
2,4-hexadiene, (E,
E)-
1-pentene, 2,3-
dimethyl
5.12
1,4-pentadiene
14.5
13.2
13.8
18.8
2
3
0.000 1.000
0.196 1.104
1,4-pentadiene, 2-
methyl
a
1-pentene, 2,4,4-
trimethyl
14.8
1,5-hexadiene, 2-
methyl
20.7
14.2
3060
23.9
25.3
691
10.0
25.5
486
3
4
6
3
3
4
3
0.301 0.891
0.392 1.208
0.392 2.208
0.066 1.367
0.098 1.301
0.334 1.325
0.394 1.409
2-pentene, (Z)-
2-pentene, (E)-
168
237
178
178
329
2
2
3
0.033 1.867
0.033 1.867
0.229 2.076
1,5-Hexadiene, 2,5-
dimethyl
a
2-pentene, 2-
methyl
430
2,4-hexadiene, 2,5-
dimethyl
2-pentene, 3-
methyl, (Z)-
465
563
329
329
151
3
3
3
0.229 2.076
0.229 2.076
0.294 1.811
1,3-hexadiene, 5-
methyl
40.8
33.5
36.0
46.1
2-pentene, 3-
methyl, (E)-
1,3-hexadiene, 5,5-
dimethyl
a
2-pentene, 2,4,4-
trimethyl
132
1,6-octadiene, 3,7-
dimethyl
1-hexene
10.9
104
182
3.99
7.37
140
140
93.8
1
2
2
3
0.053 0.787
0.053 1.787
0.053 1.787
0.334 1.650
α-pinene
90.0
2-hexene, (Z)-
2-hexene, (E)-
2-hexene, 3,4-
diethyl, (Z)-
a
b
c
Averaged value between the literature data and the current study. Values at 288 K. Values at 290 K.
A series of alkenes and dialkenes ranging from C2 to C14 for
Within the obtained data set of the 59 considered
compounds, there are three statistical outliers presented in
Figure 9 (crosses) but not taken into account for the estimation
of the SAR correlation (open squares). It is worth noting that
all estimated rates of ozonolysis of the outliers (1,6-octadiene,
3,7-dimethyl;2-hexene, 3,4-diethyl, and 2,4-hexadiene, 2,5-di-
methyl) are measured only once,49 and various consistency
problems for these data are observed.48
Additionally, using the literature data listed in Table 4 and
our own values of rate coefficients, a new linear regression,
using a larger set of data, of log kALK as a function of x could be
obtained (Figure 9):
which room temperature ozonolysis rate coefficients are available
in the literature49 were selected (Table 4). The alkenes included
in the analysis were chosen on the grounds of similarity with the
molecules used for the validation of the flow reactor, presenting
an internal or a terminal double bond, various degrees of
branching, and a wide range of ozonolysis rate coefficient values,
spanning the range 10−15 to 10−18 cm3 molecule−1 s−1. Applying
the correlation from McGillen et al. (2008) to the selected
alkenes allowed calculating predicted O3 rate coefficients, which,
as reported in Table 4, present a general good agreement (within
a factor spanning from 0.46 to 2.98) to the measured rate
coefficients, including those determined in the present work.
log k
ALK
= (1.26 0.10)x − (18.10 0.14)
(IX)
6177
dx.doi.org/10.1021/jp211480x | J. Phys. Chem. A 2012, 116, 6169−6179

The Journal of Physical Chemistry A
Article
AUTHOR INFORMATION
■
Corresponding Author
*Phone: (33) 327 712 604; fax: (33) 327 712 914; e-mail:
veronique.riffault@mines-douai.fr.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Our laboratory participates in the Institut de Recherche en
ENvironnement Industriel (IRENI), which is financed by the
́
Communaute Urbaine de Dunkerque, the Nord-Pas de Calais
Regional Council, the French Ministry of Higher Education and
Research, the CNRS, and the European Regional Development
Fund (ERDF). R.I.O. acknowledges the EMD and the IRENI
for a 1-month Invited Professor Fellowship. M.D. is grateful for
a Ph.D. scholarship from the Nord-Pas de Calais Regional
Council and Armines.
Figure 9. Ozonolysis rate coefficient (log scale) as a function of the
SAR index x for the selected alkenes in Table 4 and the corresponding
95% confidence interval (dotted line). Filled circles: this work; open
squares: literature data; crosses: outliers of literature data.
where the errors on the slope and the intercept are calculated
for a 95% confidence interval, and a value of R2 ∼ 0.93 has been
obtained for the linear regression. This correlation (eq IX) is in
excellent agreement with the one quoted by McGillen et al.
(2008) who took into account a smaller database.
REFERENCES
■
(1) Siese, M.; Becker, K. H.; Brockmann, K. J.; Geiger, H.;
Hofzumahaus, A.; Holland, F.; Mihelcic, D.; Wirtz, K. Environ. Sci.
Technol. 2001, 35, 4660−4667.
(2) Atkinson, R.; Aschmann, S. M. Environ. Sci. Technol. 1993, 27,
1357−1363.
CONCLUSION
■
(3) Aschmann, S. M.; Arey, J.; Atkinson, R. Atmos. Environ. 2002, 36,
4347−4355.
The rate coefficients of the reactions of ozonolysis of four
methylated pentenes and α-pinene have been determined using
a new flow tube reactor and a Teflon smog chamber. The
results obtained have shown very good agreement with
literature data, indicating that we may rely on the reactor to
perform further kinetic studies on new systems.
(4) Grosjean, E.; Grosjean, D. Environ. Sci. Technol. 1996, 30, 2038−
2044.
(5) Grosjean, E.; Grosjean, D. Atmos. Environ. 1996, 30, 4107−4113.
(6) Grosjean, E.; Grosjean, D. Atmos. Environ. 1998, 32, 3393−3402.
(7) Hearn, J. D.; Smith, G. D. J. Phys. Chem. A 2004, 108, 10019−
10029.
The present measurements confirmed the strong increase of
the ozonolysis rate coefficients with alkyl substitution at the
double bond, probably due to the lowering of the ionization
potential of the olefin, while substitution with a bulky
substituent (such as t-butyl in tM2P) slightly reduces the
reactivity through steric effects. The topological SAR analysis
performed using the model of McGillen et al. (2008) provided
an excellent correlation with existing data and confirms the
consistency of the measured kinetic data.
(8) Stenby, C.; Poschl, U.; von Hessberg, P.; Bilde, M.; Nielsen, O. J.;
̈
Moortgat, G. K. Atmos. Chem. Phys. Discuss. 2006, 6, 10275−10297.
(9) Bonn, B.; Schuster, G.; Moortgat, G. K. J. Phys. Chem. A 2002,
106, 2869−2881.
(10) Jang, M.; Lee, S.; Kamens, R. M. Atmos. Environ. 2003, 37,
2125−2138.
(11) Czoschke, N. M.; Jang, M.; Kamens, R. M. Atmos. Environ. 2003,
37, 4287−4299.
(12) Morris, J. W.; Davidovits, P.; Jayne, J. T.; Jimenez, J. L.; Shi, Q.;
Kolb, C. E.; Worsnop, D. R.; Barney, W. S.; Cass, G. Geophys. Res. Lett.
2002, 29.
The product yields determined in the tM2P ozonolysis study
carried out in the flow reactor at short reaction times were in
excellent agreement with the literature data and also allowed
the identification of a new nonoxidized product (2-methyl-1-
propene), probably derived from the electronic rearrangement
of one Criegee intermediate. This strongly confirms the flow
reactor suitability to carry out further similar studies and its
potential to quantify reaction intermediates that are difficult to
identify by smog chamber reactors because of their high
reactivity, involving competitive formation and consumption.
In further studies, we will use readily available state-of-the-art
instrumentation to characterize SOA by coupling a scanning
mobility particle sizer (SMPS) to measure the number and size
distribution of particles formed and a high-resolution time-of-
flight aerosol mass spectrometer (HR-ToF-AMS) to determine
their chemical average composition. The gas phase will be
monitored both by online TD/GC/FID-MS and a proton-
transfer-reaction time-of-flight mass spectrometer (PTR-ToF-MS)
to try to identify more reactive intermediates.
(13) Ezell, M. J.; Johnson, S. N.; Yu, Y.; Perraud, V.; Bruns, E. A.;
Alexander, M. L.; Zelenyuk, A.; Dabdub, D.; Finlayson-Pitts, B. J.
Aerosol Sci. Technol. 2010, 44, 329−338.
(14) Lee, S.; Kamens, R. M. Atmos. Environ. 2005, 39, 6822−6832.
(15) Bilde, M.; Pandis, S. N. Environ. Sci. Technol. 2001, 35, 3344−
3349.
(16) Tolocka, M. P.; Heaton, K. J.; Dreyfus, M. A.; Wang, S.; Zordan,
C. A.; Saul, T. D.; Johnston, M. V. Environ. Sci. Technol. 2006, 40,
1843−1848.
(17) Schauer, J. J.; Kleeman, M. J.; Cass, G. R.; Simoneit, B. R. T.
Environ. Sci. Technol. 1999, 33, 1578−1587.
(18) Schauer, J. J.; Kleeman, M. J.; Cass, G. R.; Simoneit, B. R. T.
Environ. Sci. Technol. 2001, 35, 1716−1728.
(19) Hays, M. D.; Geron, C. D.; Linna, K. J.; Smith, N. D.; Schauer, J. J.
Environ. Sci. Technol. 2002, 36, 2281−2295.
(20) Pinho, P. G.; Pio, C. A.; Carter, W. P. L; Jenkin, M. E. J. Atmos.
Chem. 2006, 55, 55−79.
(21) Kroll, J. H.; Seinfeld, J. H. Environ. Sci. Technol. 2005, 39, 4159−
4165.
(22) Kroll, J. H.; Seinfeld, J. H. Atmos. Environ. 2008, 42, 3593−3624.
(23) Bennadji, H.; Glaude, P. A.; Coniglio, L.; Billaud, F. Fuel 2011,
90, 3237−3253.
6178
dx.doi.org/10.1021/jp211480x | J. Phys. Chem. A 2012, 116, 6169−6179

The Journal of Physical Chemistry A
Article
(24) Seeley, J. V.; Jayne, J. T.; Molina, M. J. Int. J. Chem. Kinet. 1993,
25, 571−594.
(25) Helmig, D. Atmos. Environ. 1997, 31, 3635−3651.
(26) Turpin, E.; Tomas, A.; Fittschen, C.; Devolder, P.; Galloo, J.-C.
Environ. Sci. Technol. 2006, 40, 5956−5961.
(27) Malkin, T. L.; Goddard, A.; Heard, D. E.; Seakins, P. W. Atmos.
Chem. Phys. 2010, 10, 1441−1459.
(28) Kroll, J. H.; Clarke, J. S.; Donahue, N. M.; Anderson, J. G.;
Demerjian, K. L. J. Phys. Chem. A 2001, 105, 1554−1560.
(29) Kroll, J. H.; Donahue, N. M.; Cee, V. J.; Demerjian, K. L.;
Anderson, J. G. J. Am. Chem. Soc. 2002, 124, 8518−8519.
(30) Tuazon, E. C.; Aschmann, S. M.; Arey, J.; Atkinson, R. Environ.
Sci. Technol. 1997, 31, 3004−3009.
(31) Hasson, A. S.; Chung, M. Y.; Kuwata, K. T.; Converse, A. D.;
Krohn, D.; Paulson, S. E. J. Phys. Chem. A 2003, 107, 6176−6182.
(32) Rickard, A. R.; Johnson, D.; McGill, C. D.; Marston, G. J. Phys.
Chem. A 1999, 103, 7656−7664.
(33) Presto, A. A.; Donahue, N. M. J. Phys. Chem. A 2004, 108,
9096−9104.
(34) Gutbrod, R.; Meyer, S.; Rahman, M. M.; Schindler, R. N. Int.
J. Chem. Kinet. 1997, 29, 717−723.
(35) Horie, O.; Moortgat, G. K. Chem. Phys. Lett. 1998, 288, 464−
472.
(36) Grosjean, E.; Grosjean, D. Int. J. Chem. Kinet. 1995, 27, 1045−
1054.
(37) Leather, K. E.; McGillen, M. R.; Percival, C. J. Phys. Chem.
Chem. Phys. 2010, 12, 2935−2943.
(38) Atkinson, R.; Carter, W. P. L Chem. Rev. 1984, 84, 437−470.
(39) Grosjean, E.; Grosjean, D. Int. J. Chem. Kinet. 1996, 28, 911−
918.
(40) Atkinson, R.; Baulch, D. L.; Cox, R. A.; Crowley, J. N.;
Hampson, R. F.; Haynes, R. G.; Jenkin, M. E.; Rossi, M. J.; Troe, J.
Atmos. Chem. Phys. 2006, 6, 3625−4055.
(41) Bernard, F.; Fedioun, I.; Peyroux, F.; Quilgars, A.; Daele, V.;
̈
Mellouki, A. J. Aerosol Sci. 2012, 43, 14−30.
(42) Grosjean, E.; Grosjean, D. Environ. Sci. Technol. 1997, 31,
2421−2427.
(43) Finlayson-Pitts, B. J.; Pitts, J. N. J. Chemistry and Physics of the
Upper and Lower Atmosphere; Wiley: New York, 1999.
(44) Neeb, P.; Moortgat, G. K. J. Phys. Chem. A 1999, 103, 9003−
9012.
(45) Seinfeld, J. H.; Pandis, S. N. Atmospheric Chemistry and Physics:
From Air Pollution to Climate Change; Wiley: New York, 1998; Vol. 2.
(46) Horie, O.; Moortgat, G. K. Acc. Chem. Res. 1998, 31, 387−396.
(47) King, M. D.; Canosa-Mas, C. E.; Wayne, R. P. Phys. Chem.
Chem. Phys. 1999, 1, 2231−2238.
(48) McGillen, M. R.; Carey, T. J.; Archibald, A. T.; Wenger, J. C.;
Shallcross, D. E.; Percival, C. J. Phys. Chem. Chem. Phys. 2008, 10,
1757−1768.
(49) NIST. Chemical Kinetics Database on the Web. In Standard
Reference Database 17, Version 7.0(Web Version), Release 1.6.0; 7.0 ed.,
2011.
(50) Cox, R. A.; Penkett, S. A. J. Chem. Soc., Faraday Trans. 1 1972,
68, 1735−1753.
6179
dx.doi.org/10.1021/jp211480x | J. Phys. Chem. A 2012, 116, 6169−6179