Chemical
Kinetic
Study
of the
Reaction
of CH
2
OO
with
CH
3
O
2
Wen
Chao,
*
Charles
R. Markus,
Mitchio
Okumura,
Frank
A. F. Winiberg,
and Carl J. Percival
*
Cite
This:
J. Phys.
Chem.
Lett.
2024,
15,
3690−3697
Read
Online
ACCESS
Metrics
& More
Article
Recommendations
*
sı
Supporting
Information
ABSTRACT:
Criegee
intermediates
play an important
role in the oxidizing
capacity
of the
Earth’s troposphere.
Although
extensive
studies have been conducted
on Criegee
intermediates
in the past decade,
their kinetics
with radical species remain underexplored.
We investigated
the
kinetics
of the simplest
Criegee
intermediate,
CH
2
OO, with the methyl peroxy radical, CH
3
O
2
,
as a model system to explore
the reactivities
of Criegee
intermediates
with peroxy radicals.
Using a multipass
UV
−
Vis
spectrometer
coupled
to a pulsed-laser
photolysis
flow reactor,
CH
2
OO and CH
3
O
2
were generated
simultaneously
from the photolysis
of CH
2
I
2
/CH
3
I/O
2
/
N
2
mixtures
with CH
2
OO measured
directly
near 340 nm. We determined
a reaction
rate
coefficient
k
CH
d
2
OO+CH
d
3
O
d
2
= (1.7
±
0.5)
×
10
−
11
cm
3
s
−
1
at 294 K and 10 Torr, where the
influence
of iodine adducts
is reduced.
This rate coefficient
is faster than previous
theoretical
predictions,
highlighting
the challenges
in accurately
describing
the interaction
between
zwitterionic
and biradical
characteristics
of Criegee
intermediates.
C
arbonyl
oxides, also known as Criegee
intermediates,
play
critical
roles in the Earth’s
troposphere.
1
−
3
These
intermediates
arise from the ozonolysis
of unsaturated
organic
compounds,
such as isoprene,
and are believed
to enhance
the
atmosphere’s
oxidation
capacity.
They also contribute
to the
formation
of secondary
organic
aerosols
(SOA)
and ultrafine
particles,
especially
in biogenic-rich
atmospheres
such as that
of the Amazon
rainforest.
2
−
4
Laboratory
chamber
experiments
have shown that different
oligomers
are generated
as secondary
products
of ozonolysis
reactions
in the presence
of different
OH radical scavengers,
resulting
in the production
of different
peroxy radicals.
5,6
Moreover,
a recent study on the chemical
composition
of SOA from
β
-pinene
ozonolysis
revealed
that
the aerosol
composition
is predominantly
influenced
by
reactions
of stabilized
Criegee
intermediates
with the
corresponding
peroxy radicals.
7
In laboratory
studies,
small Criegee
intermediates
(with
three carbon
atoms or fewer)
are generated
through
the
photolysis
of diiodoalkanes.
8,9
Their spectroscopic
character-
istics have been extensively
studied,
providing
valuable
insights
into their reactivities.
10
As a result of the pronounced
zwitterionic
nature of the C
�
O
−
O
functional
group,
11
the
reactivities
of some Criegee
intermediates
depend
heavily
on
their structures.
For example,
the
anti
conformers
(where
a
hydrogen
atom aligns with the terminal
oxygen)
undergo
rapid
reactions
with hydrogen-bonding
molecules,
such as H
2
O,
NH
3
, and CH
3
OH, via a 1,2-insertion
mechanism
into the
hydrogen
bond.
12
Conversely,
the
syn
conformers
(where
other substitution
groups align with the terminal
oxygen)
react
slowly
with hydrogen-bonding
molecules
but can rapidly
decompose
to yield OH radicals.
13
The oxidation
of SO
2
and NO
2
via reactions
with Criegee
intermediates,
generating
SO
3
and Criegee-NO
2
adducts,
shows only weak structural
dependencies.
9,14
Isoprene-derived
Criegee
intermediates,
such as methyl vinyl ketone oxide
15
and
methacrolein
oxide,
16
display similar reactivity
in the reaction
with SO
2
to their smaller counterparts.
This suggests
that small
Criegee
intermediates
can be used as analogues
to understand
the behavior
of larger ones, greatly reducing
the computational
cost for high-level
ab initio
calculations.
While the reactivities
of small Criegee
intermediates
with
stable molecules
have been thoroughly
investigated,
10,17
there
is a notable
gap in our understanding
of their interactions
with
radical species.
Welz et al. determined
rate coefficients
of
≤
6
×
10
−
14
and 7
×
10
−
12
cm
3
s
−
1
for the reactions
of the simplest
Criegee
intermediate,
CH
2
OO, with NO and NO
2
, respec-
tively.
8
Further
studies on reactions
of
syn
-CH
3
CHOO,
9
anti
-
CH
3
CHOO,
9
and (CD
3
)
2
COO
18
with NO
2
have similar rate
coefficients
near 2
×
10
−
12
cm
3
s
−
1
. Chhantyal-Pun
et al.
19
measured
fast reaction
rate coefficients
for methyl
peroxy,
CH
3
O
2
, and acetyl peroxy radicals,
CH
3
C(O)O
2
, reporting
the
rate coefficients
k
(CH
2
OO + RO
2
, RO
2
= CH
3
O
2
or
CH
3
C(O)O
2
) = (2.4
±
1.2)
×
10
−
11
cm
3
s
−
1
with weak
temperature
(243
−
310
K) effects.
19
Criegee
intermediates
are unique,
exhibiting
both zwitter-
ionic and biradical
character
with respect to reaction
pathways.
From a zwitterionic
perspective,
the reactions
of Criegee
intermediates
with radicals
are expected
to be slow. However,
the biradical
perspective
predicts
a fast rate similar to that of a
Received:
January
16, 2024
Revised:
March 14, 2024
Accepted:
March 18, 2024
Published:
March 28,
2024
Letter
pubs.acs.org/JPCL
©
2024
@2024
California
Institute
of
Technology
Gov
sponsorship
acknowledged.
Published
by
American
Chemical
Society
3690
https://doi.org/10.1021/acs.jpclett.4c00159
J. Phys.
Chem.
Lett.
2024,
15,
3690
−
3697
Downloaded via CALIFORNIA INST OF TECHNOLOGY on April 15, 2024 at 22:57:43 (UTC).
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
typical
radical
−
radical
association
reaction.
20
Theoretical
calculations
indicate
that the CH
2
OO + CH
3
O
2
reaction
passes
through
a submerged
barrier,
yielding
a fast rate
coefficient
of 4
×
10
−
12
cm
3
s
−
1
at 290 K,
21
suggesting
pronounced
biradical
character
in contrast
to the widely
accepted
zwitterionic
perspective.
In this work, we investigated
the reaction
kinetics
of
CH
2
OO + CH
3
O
2
at 294 K and 10 Torr total pressure.
CH
2
OO and CH
3
O
2
were generated
through
the photolysis
of
diiodomethane,
CH
2
I
2
, and iodomethane,
CH
3
I, respectively,
in the presence
of oxygen.
We monitored
the near-UV
absorption
using a white cell multipass
UV
−
vis
spectrometer
coupled
to a pulsed-laser
photolysis
flow reactor
22
and
constructed
a chemical
kinetics
model to analyze
the measured
temporal
profiles.
The absorption
cross sections
of the species
that can be characterized
in our experimental
system
are
summarized
in Figure S1.
In the CH
2
I
2
/O
2
/N
2
photolysis
system
(CH
2
I
2
+
h
ν
→
CH
2
I + I, CH
2
I + O
2
→
CH
2
OO + I,
k
CH
d
2
I+O
d
2
= 1.5
×
10
−
12
cm
3
s
−
1
),
23
the CH
2
I radical exhibited
a lifetime
of less than 2
μ
s at 10 Torr of O
2
, resulting
in the prompt
formation
of
CH
2
OO. As anticipated,
the characteristic
absorptions
from
CH
2
OO near 340 nm and IO around
430 nm were clearly
visible. Additionally,
a negative
absorption
near 300 nm was
observed,
which was attributed
to the depletion
of CH
2
I
2
due
to photolysis
as illustrated
in Figure 1A.
Figure 1B presents
the spectra from the photolysis
of a N
2
/
O
2
/CH
2
I
2
/CH
3
I mixture
at 50 Torr total pressure.
The
CH
3
O
2
radicals
were generated
via the photolysis
of CH
3
I
followed
by the reaction
of the resulting
CH
3
radical with O
2
(CH
3
I +
h
ν
→
CH
3
+ I, CH
3
+ O
2
+ M
→
CH
3
O
2
+ M,
k
CH
d
3
+O
d
2
= 2.7
×
10
−
13
cm
3
s
−
1
at 50 Torr);
24
the CH
3
radical
lifetime
was roughly
12
μ
s at 10 Torr of O
2
, leading
to the
immediate
formation
of CH
3
O
2
. Based on the collisional
energy transfer
parameter
for modeling
the pressure
depend-
ence of the CH
3
+ O
2
reaction,
25
we estimate
that CH
3
O
2
reaches
thermal
equilibrium
within 0.5 ms even though
its
formation
is highly exothermic
(30 kcal mol
−
1
). The initial
signal intensity
for CH
2
OO remained
unchanged
upon the
introduction
of CH
3
I, suggesting
that the formations
of
CH
2
OO and CH
3
O
2
were both rapid and independent.
The
absorption
signals of IO were observed
at 3 ms, which can be
attributed
to the increased
iodine
atom concentration
following
CH
3
I photolysis.
However,
in the absence
of
CH
2
I
2
(Figure
1C), no IO was detected,
implying
that IO
primarily
stemmed
from the chemistry
involving
CH
2
I and
CH
2
OO.
CH
2
OO was monitored
near 340 nm, where both CH
3
O
2
and CH
3
I have negligible
absorption
cross sections.
Figure 2
shows the temporal
profiles
of CH
2
OO as a function
of [CH
3
I]
at total pressures
of 10 and 50 Torr. At 50 Torr (Figure
2B),
the decay rates increased
with [CH
3
I] but remained
the same
for [CH
3
I] > 8.5
×
10
14
cm
−
3
. Furthermore,
an increase
in
absorption
was observed
at longer times for higher [CH
3
I]
runs. At lower [CH
3
I], the long time signals were negative,
which was attributed
to the photolytic
depletion
of CH
2
I
2
.
However,
at higher [CH
3
I], positive
signals were recorded
at
longer times, with the intensity
being proportional
to [CH
3
I]
(Figure
S2). Additionally,
we noted that this shift became
more pronounced
at 90 Torr and 238 K (Figure
2C) and was
significantly
minimized
at 10 Torr and 294 K (Figure
2A).
We have considered
IO, ICH
2
OO, and CH
3
OOI as possible
candidates
that would explain
the observed
features
at long
delay times. First, we note that the absorption
of IO is far
smaller
than the expected
absorbance
of 0.02 near 427 nm,
σ
IO
(427 nm)/
σ
IO
(340 nm) = 1.44
×
10
−
17
cm
2
/2.84
×
10
−
19
=
50.7.
26
Second,
a residual
absorption
signal at 340 nm,
attributed
to the ICH
2
OO adduct,
has been reported
when
CH
2
OO was fully scavenged
by water vapor.
27
However,
we
did not observe
any distinct
absorption
features
from ICH
2
OO
within the detection
window.
Even under ideal conditions
for
the formation
of ICH
2
OO, where the yield is expected
to be
greater than 30% (20
−
90
Torr and 238 K),
28
the absorption
could be comprehensively
accounted
for by
Δ
CH
2
I
2
, CH
2
OO,
and IO (Figures
S4
−
S6),
indicating
that it is unlikely
to be the
carrier. Finally,
an absorption
signal was observed
when only a
CH
3
I/O
2
/N
2
mixture
was introduced
into the reactor at lower
temperatures
(Figure
S7, green trace).
The profile
was
indicative
of a secondary
product
from a strong absorption
feature with a peak near 290 nm (Figure
S8), which has been
previously
assigned
to CH
3
OOI.
29
Figure 1.
Representative
absorption
spectra
of the photochemical
systems
containing
(A) only CH
2
I
2
, (B) both CH
2
I
2
and CH
3
I, and
(C) only CH
3
I at selected
delay times (gate width = 117.5
μ
s and
1024 laser shot average)
at 50 Torr and 294 K. Each spectrum
is
composed
of two measurements
using a 600 grooves/mm
grating
with center wavelengths
of 280 and 413 nm (spectral
coverage
±
63
nm); thus, a small spectral
gap of 7 nm appears
in our spectra.
The
absorption
features
of CH
2
OO, IO, CH
2
I
2
, and CH
3
O
2
are indicated.
Note that the absorption
signals below 325 nm are different
at 0.1 ms
(red) and 3 ms (green)
in the presence
of CH
3
I.
The
Journal
of Physical
Chemistry
Letters
pubs.acs.org/JPCL
Letter
https://doi.org/10.1021/acs.jpclett.4c00159
J. Phys.
Chem.
Lett.
2024,
15,
3690
−
3697
3691
We observed
the absorption
feature
became
stronger
at
lower temperatures
(Figure
S3), in line with the formation
through
the CH
3
O
2
+ I
→
CH
3
OOI association
reaction.
Also,
the absorption
feature
at short wavelengths,
which
is
dominated
by CH
3
O
2
, remained
unchanged
at 10 Torr (Figure
S11) compared
to 50 Torr (Figure
1). Nevertheless,
we
observed
a small absorption
signal at 340 nm (<5
×
10
−
4
in
absorbance)
in the absence
of CH
2
I
2
for the photolysis
of the
CH
3
I/N
2
/O
2
mixture
(Figure
S15). The absorption
cross
section of CH
3
OOI near 290 nm
29
has been reported
to be 8.8
×
10
−
18
cm
2
, and we estimated
an absorption
cross section
near 340 nm
30
of 2
×
10
−
18
cm
2
based on the relative
profiles.
The maximum
[CH
3
OOI] in our experiments
at 10 Torr will
be less than 5
×
10
11
cm
−
3
(
L
eff
≈
450 cm). A background
subtraction
to eliminate
the minor influence
of CH
3
OOI
(<10% of the initial CH
2
OO signals,
Figure S15) was applied
to the CH
2
OO temporal
profiles.
As mentioned
previously,
these traces were analyzed
using a
chemical
kinetic model with the reactions
listed in Table 1.
These reactions
were classified
into those involving
CH
2
OO
(reactions
k
1
−
k
3
), iodine species (reactions
k
4
and
k
5
), CH
3
O
2
(reactions
k
6
−
k
11
), and CH
3
OOI (reactions
k
12
and
k
13
). We
also considered
the lumped
first-order
decay due to diffusion
or wall loss (
C
diff
). The CH
2
OO + I reaction
(
k
2
) and the
CH
2
OO self-reaction
(
k
3
) dominate
the observed
CH
2
OO
decay in the absence
of coreactants.
We found that the
modeled
[CH
2
OO] reproduced
the measured
temporal
profiles
at 10 Torr (Figure
S12). At [CH
2
I
2
] = 5.1
×
10
13
and 2.6
×
10
13
cm
−
3
, the obtained
C
diff
values were (796
±
11)
and (701
±
7) s
−
1
g
0.5
mol
−
0.5
, which correspond
to first-order
decay rates of (117
±
2) and (103
±
1) s
−
1
, respectively.
These
values agreed with the natural
decay of CH
2
OO in our flow
reactor (Figures
S13 and S14) and were adopted
for analyzing
the CH
2
OO + CH
3
O
2
reaction
rate coefficients.
These fits yielded
rate coefficients
k
2
= (2.30
±
0.12)
×
10
−
12
cm
3
s
−
1
and
k
3
= (7.81
±
0.32)
×
10
−
12
cm
3
s
−
1
, with
errors representing
one standard
deviation.
Previous
studies
have been conducted
to model the photolysis
of CH
2
I
2
/O
2
/N
2
mixtures,
monitoring
the signals
of CH
2
OO, IO, and the
iodine atom over 10
−
760
Torr.
31
−
34,38
−
40
For the CH
2
OO + I
reaction
(
k
2
), there is a prevailing
consensus
in the literature
indicating
a slow rate coefficient
of approximately
2
×
10
−
12
cm
3
s
−
1
below 10 Torr, which is in agreement
with our results.
We note that there is no clear agreement
in the literature
for
the rate coefficient
of CH
2
OO + I reaction
at at pressures
of 50
Torr and above;
31
−
34,38
−
40
therefore,
only the temporal
profiles
recorded
at 10 Torr were analyzed
to obtain the rate
coefficient
for CH
2
OO + CH
3
O
2
.
In previous
studies,
CH
3
OOI was assumed
to react rapidly
with I atoms (
k
13
= 1.5
×
10
−
10
cm
3
s
−
1
)
29
to regenerate
CH
3
O
2
and produce
I
2
in order to explain the fast formation
of
I
2
and the overestimation
of CH
3
O
2
concentration.
For the
CH
3
OO + I
→
CH
3
OOI reaction,
calculations
41
predict
a
strong pressure
dependence
with the falloff region from 10
−
4
to
1000 Torr, and Dillon et al.
29
reported
a rate coefficient
of
k
12
= 2
×
10
−
11
cm
3
s
−
1
at 70 Torr. By assuming
a linear pressure
dependence,
we estimated
k
12
= 2.9
×
10
−
12
cm
3
s
−
1
at 10
Torr. In the kinetic fitting, we varied
k
12
from 0 to 2
×
10
−
11
cm
3
s
−
1
to estimate
the influence
of CH
3
OOI formation,
which
effectively
decrease
the [CH
3
O
2
] and [I].
Figure 3 displays
a representative
fit of the temporal
profile
of CH
2
OO along with modeled
CH
3
OOI, IO, CH
3
O, CH
3
O
2
,
and I. The concentrations
of each species at a delay of 0.2 ms
were obtained
by performing
a least-squares
fit of the recorded
spectra (Figure
S16 as an example)
to the reported
absorption
cross section (Figure
S1). The effective
absorption
length (
L
eff
≈
450 cm) was characterized
by previous
measurements
of
NO
2
.
22
The fit was highly sensitive
to the accurate
determination
of
[CH
3
O
2
]
t
=0
, for which we utilized
two different
approaches.
First, we extrapolated
[CH
3
O
2
] to
t
= 0. However,
as shown in
Figure S16, there was considerable
spectral
overlap
of CH
3
I
and CH
3
O
2
, which could result in significant
errors. A second
approach
was to utilize the relative
depletion
of
Δ
CH
2
I
2
to
Δ
CH
3
I, assuming
that the photolysis
of CH
3
I yields 100%
CH
3
O
2
and using the known absorption
cross sections
of both
precursors
at 248 nm.
24
The [CH
3
O
2
] derived
from both
methods
was consistent
with an uncertainty
of 20% (one
standard
deviation)
in all except a few measurements
at lower
[CH
3
I] (Figure
S17).
The uncertainty
of the fitted CH
2
OO + CH
3
O
2
reaction
rate
coefficients
was estimated
using a Monte
Carlo fitting
procedure
by performing
multiple
fits with randomly
varied
fixed parameters
taking into account
the literature
uncertainty
(assuming
a normal
distribution)
in the rate coefficients
of
CH
2
OO + I (
k
2
), CH
2
OO + CH
2
OO (
k
3
), and CH
3
O
2
+
CH
3
O
2
(
k
7
) as well as the 20% error in [CH
3
O
2
] (Figures
S21
−
S23,
Table S1). The modeled
concentrations
of both
Figure 2.
Representative
temporal
profiles
near 340 nm at various
[CH
3
I] with total pressures
and temperatures
of (A) 10 Torr, 294 K,
(B) 50 Torr, 294 K, and (C) 90 Torr, 238 K. The O
2
pressures
are 9,
10, and 5 Torr for (A), (B), and (C), respectively.
Note that the
baseline
shift is less obvious
at 10 Torr and 294 K.
The
Journal
of Physical
Chemistry
Letters
pubs.acs.org/JPCL
Letter
https://doi.org/10.1021/acs.jpclett.4c00159
J. Phys.
Chem.
Lett.
2024,
15,
3690
−
3697
3692
CH
2
OO and CH
3
O
2
were found to be the most sensitive
to
these reactions
in our model.
This error analysis
procedure
produced
a distribution
of values for the target rate coefficient
(Figure
S21), which represents
the error propagation.
We
determined
the rate coefficient
as the mean value and the error
as one standard
deviation
of the distribution.
The mean rate coefficients
as a function
of [CH
3
I] are
summarized
in Figure 4. The modeled
reaction
rate coefficients
were constant,
within experimental
error, as a function
of
[CH
3
I] and [CH
2
I
2
], which were varied by an order of
magnitude
and a factor of 2, respectively.
A rate coefficient
of
k
CH
d
2
OO+CH
d
3
O
d
2
= (1.7
±
0.5)
×
10
−
11
cm
3
s
−
1
was derived
from
our data at 10 Torr and 294 K. The formation
of CH
3
OOI
increases
the determined
k
CH
d
2
OO+CH
d
3
O
d
2
by 6% by reducing
the
concentration
of CH
3
O
2
and I atoms, but the effect is relatively
small in comparison
to the overall uncertainty.
Table 1. Summary
of the Rate Coefficients
of the Chemical
Kinetics
Model for the Kinetic
Analysis
Comments
Reactions
a
Rate Coefficients
at 298 K and 10 Torr
b
Notation
CH
2
OO Chemistry
CH
2
I
2
+
h
ν
→
CH
2
I + I
Instantaneous
−
CH
2
I + O
2
→
CH
2
OO + I
1.13
×
10
−
12
c
k
1a
31
CH
2
I + O
2
→
H
2
CO + IO
1.5
×
10
−
13
k
1b
32
CH
2
I + O
2
→
Other Products
2.2
×
10
−
13
d
k
1c
31
CH
2
OO + I
→
H
2
CO + IO
(2.2
±
1.1)
×
10
−
12
k
2
33
CH
2
OO + CH
2
OO
→
2 H
2
CO + O
2
(7.4
±
0.6)
×
10
−
11
k
3
34
Iodine Chemistry
I + I
→
I
2
6.2
×
10
−
15
k
4
30
IO + IO
→
Products
9.9
×
10
−
11
k
5
35
CH
3
O
2
Chemistry
CH
3
I +
h
ν
→
CH
3
+ I
Instantaneous
−
CH
3
+ O
2
→
CH
3
O
2
9.2
×
10
−
14
k
6
24
CH
3
O
2
+ CH
3
O
2
→
2CH
3
O + O
2
(2.0
±
0.9)
×
10
−
13
k
7
36
CH
3
O + CH
3
O
→
H
2
CO + CH
3
OH
7.0
×
10
−
11
k
8
37
CH
3
O + O
2
→
H
2
CO + HO
2
1.9
×
10
−
15
k
9
24
CH
3
O + HO
2
→
Products
e
1.1
×
10
−
10
k
10
37
HO
2
+ HO
2
→
H
2
O
2
+ O
2
1.4
×
10
−
12
k
11
24
CH
3
OOI Chemistry
CH
3
O
2
+ I
→
CH
3
OOI
2.9
×
10
−
12
f
k
12
29
CH
3
OOI + I
→
CH
3
O
2
+ I
2
1.5
×
10
−
10
k
13
29
Target Reaction
CH
2
OO + CH
3
O
2
→
Adduct
(1.7
±
0.5)
×
10
−
11
k
CH
d
2
OO+CH
d
3
O
d
2
a
The first-order
diffusion
was simulated
using
C
diff
×
m
−
0.5
for radical species,
including
CH
2
OO, CH
3
O
2
, CH
3
O, HO
2
, IO, and the I atom, where
C
diff
denotes
the proportional
constant
and
m
denotes
the molar mass.
b
Unit: cm
3
s
−
1
. The reaction
rate coefficients
with uncertainty
were
considered
in the Monte Carlo error analysis
procedure.
c
The CH
2
I + O
2
→
products
reaction
rate coefficient
is
k
1
= 1.5
×
10
−
12
cm
3
s
−
1
. The
listed rate coefficients
are derived
based on the measured
yield of ref 28.
d
The reaction
rate coefficients
were derived
as
k
1c
=
k
1
−
k
1a
−
k
1b
to
match the yield of CH
2
OO.
e
Calculations
from ref 37 suggest
CH
2
O + H
2
O
2
and CH
3
OH + O
2
(
3
P) as products
on the singlet and triplet potential
energy surfaces,
respectively.
f
Assuming
a linear pressure
dependence
and derived
from 2
×
10
−
11
cm
3
s
−
1
at 70 Torr in ref 29.
Figure 3.
Representative
experimental
(blue) and modeled
(smooth
lines) temporal
profiles.
The residual
(black)
has been offset in the
figure to avoid overlap
with the measured
profiles.
The model profiles
of [CH
3
O
2
] (red) and [I] (orange)
were scaled for comparison.
The
magenta
line shows the product
formation
from the CH
2
OO +
CH
3
O
2
reaction;
P
O
d
2
= 9.5 Torr, [CH
2
I
2
] = 5.1
×
10
13
cm
−
3
, and
[CH
3
I] = 1.1
×
10
15
cm
−
3
, balanced
with N
2
to 10.0 Torr at 294 K.
The concentration
of CH
2
OO was derived
using
L
eff
= 450 cm and
σ
CH
d
2
OO
(340 nm) = 1.2
×
10
−
17
cm
2
. The profile was corrected
for the
negative
absorption
of
Δ
CH
2
I
2
, which
resulted
in a positive
concentration
before time zero.
Figure 4.
Summary
of the fitted CH
2
OO + CH
3
O
2
reaction
rate
coefficients,
k
CH
d
2
OO+CH
d
3
O
d
2
, for distinct
[CH
3
I] and [CH
2
I
2
] at 10 Torr
and 294 K. The CH
3
OO + I
→
CH
3
OOI rate coefficient
(
k
12
) was set
to 2.9
×
10
−
12
cm
3
s
−
1
. Error bars represent
one standard
deviation
of
the distribution
of
k
CH
d
2
OO+CH
d
3
O
d
2
, obtained
from the Monte Carlo error
analysis
procedure,
including
the uncertainty
of [CH
3
]
0
,
k
2
,
k
3
, and
k
7
.
The dashed gray line shows the average
of a total of 9 measurements.
The dotted light-gray
lines show the average
values using different
k
12
values in the model (Figures
S24 and S25).
The
Journal
of Physical
Chemistry
Letters
pubs.acs.org/JPCL
Letter
https://doi.org/10.1021/acs.jpclett.4c00159
J. Phys.
Chem.
Lett.
2024,
15,
3690
−
3697
3693
There is only one other study of the reaction
of CH
2
OO +
CH
3
O
2
in the literature
to date.
19
Acetone
photolysis
was
utilized
to generate
peroxy
radicals,
which resulted
in the
simultaneous
generation
of CH
3
C(O)O
2
and CH
3
O
2
, while
monochromatic
cavity ring down spectroscopy
was used for
the detection
of transient
species.
While highly sensitive,
the
single-wavelength
method
was unable
to independently
measure
CH
3
C(O)O
2
, CH
3
O
2
, and CH
2
OO. By assuming
the same rate coefficients
for reactions
of CH
2
OO + CH
3
O
2
and CH
2
OO + CH
3
C(O)O
2
, they reported
a rate coefficient
of
k
(CH
2
OO + RO
2
, RO
2
= CH
3
O
2
or CH
3
C(O)O
2
) = (2.4
±
1.2)
×
10
−
11
cm
3
s
−
1
.
19
This agrees with our measured
k
CH
d
2
OO+CH
d
3
O
d
2
values within
the large reported
experimental
uncertainty.
Nevertheless,
the results of both studies serve to
suggest
that the reactivities
of Criegee
intermediates
toward
peroxy radicals
are similar (i.e., fast and on the order of
∼
10
−
11
cm
3
s
−
1
). Clearly,
there is considerable
uncertainty,
and further
work on the reaction
between
CH
3
C(O)O
2
with CH
2
OO is
required.
Theoretical
calculations
suggest
that the terminal
oxygen
of
the peroxy radical approaches
the central carbon of the Criegee
intermediate
and donates
its unpaired
electron
to the central
carbon.
20,21
While this interaction
is repulsive
for the ground
state of CH
2
OO due to its zwitterionic
character,
mixing with a
triplet excited
state can lead to a strongly
bound potential.
20
The addition
reaction
can be compared
to that of radicals
and
a carbonyl
moiety,
such as formaldehyde.
The energy of the
lowest triplet excited
state of CH
2
OO is 1.4 eV (32.3 kcal
mol
−
1
),
20
2.7 times lower than the lowest excited state energy
(87.6 kcal mol
−
1
)
42
of formaldehyde.
Notably,
the reaction
rate
of HO
2
with formaldehyde
is quite slow (
∼
3
×
10
−
14
cm
3
s
−
1
),
43
implying
that the mixing between
the singlet ground
state and the triplet
excited
state may be weak for
formaldehyde,
likely because
of the significant
energy gap.
The CH
2
OO + CH
3
O
2
rate coefficient
of 4
×
10
−
12
cm
3
s
−
1
was calculated
at the CASPT2(19,
15)/aug-cc-pVTZ
level
21
and is 4 times slower than our measurements.
This discrepancy
points to the need for a more sophisticated
treatment
of the
mixing
of the excited
state at the transition
state structure
beyond
the perturbation
method
(e.g., the multireference
configuration
interaction
method)
in order to fully understand
the reaction
of Criegee
intermediates
with peroxy radicals.
Association
reactions
of Criegee
intermediates,
such as the
reaction
with organic
acids, results in the formation
of low-
vapor-pressure
reaction
products
and can contribute
to SOA
formation.
44
Such reactions
have been shown
to be of
particular
importance
in the Amazon
rainforest,
where Criegee
intermediates
are estimated
to achieve
concentrations
of as
high as 2000 cm
−
3
.
45
Field measurements
have revealed
significant
SOA formation,
yet current
atmospheric
models
struggle
to replicate
this observation.
4
Chamber
experiments
further
suggest
that oligomers
play an important
role in the
mechanism
of particle
formation.
7
By integrating
new reaction
pathways,
such as the fast reactions
involving
peroxy radicals
reported
in this study and water enhancement
effects
of
reactions
with hydrogen-bonding
molecules,
12
the gap between
empirical
models
and field measurements
might be bridged.
In this study, we have investigated
the CH
2
OO + CH
3
O
2
reaction
using a White cell multipass
UV
−
Vis
spectrometer
coupled
with a pulsed-laser
photolysis
flow reactor.
CH
2
OO
and CH
3
O
2
were generated
by photolyzing
CH
2
I
2
and CH
3
I,
respectively,
in the presence
of oxygen.
Photolysis
of the
CH
3
I/O
2
/N
2
mixture
produced
a prominent
absorption
signal
with a peak position
near 290 nm, which was attributed
to the
formation
of CH
3
OOI. The absorption
temporal
profiles
of
CH
2
OO were monitored
near 340 nm at low pressures,
which
mitigated
the impacts
of CH
3
O
2
+ I and CH
2
OO + I reactions,
and we were able to determine
the CH
2
OO + CH
3
O
2
reaction
rate coefficient
k
CH
d
2
OO+CH
d
3
O
d
2
= (1.7
±
0.5)
×
10
−
11
cm
3
s
−
1
at
10 Torr and 294 K.
■
EXPERIMENTAL
METHODS
The experimental
setup was described
in our previous
work.
22
In short, we utilized
a free-space
broadband
light source
(LDLS,
Energetiq
EQ-99)
that was collimated
using parabolic
mirrors
(Thorlab,
MPD149-F01,
RFL = 101.6 mm, 90
°
of the
OAP) and directed
through
a White cell. The photolysis
laser
passed through
the gap between
the object mirrors
and was
directed
colinearly
with the flow reactor
and fully overlapped
with the probe beam, achieving
an effective
absorption
path
length of
L
eff
≈
450 cm. After passing
through
the multipass
system,
the probe beam was focused
into a dual-exit
spectrograph.
We used two spectrographs
over the course of
this study, fitted with either a 600 grooves/mm
(Princeton
Instruments,
SpectraPro
HRS-300)
or 300 grooves/mm
grating
(Acton
Research
Corporation,
SpectraPro
300i), each
offering
distinct
spectral
resolution
and coverage.
A half-height
mirror in the spectrograph
splits the probe beam, directing
the
lower portion
to a photomultiplier
tube (Hamamatsu
R928)
and the upper portion
to an intensified
CCD camera
(Princeton
Instruments
PI-MAX4,
1024
×
256), allowing
for
simultaneous
spectral
and fixed-wavelength
temporal
light
collection.
A small stream
of nitrogen
flowed
through
a bubbler
containing
CH
2
I
2
(Sigma-Aldrich,
>99%) and copper shavings
as a stabilizer.
The bubbler
was held in a temperature-
controlled
bath at 292 K (Fisherbrand,
Isotemp
4100) and
copper shavings
as a stabilizer.
CH
3
I (Sigma-Aldrich,
>99.5%)
was purified
using freeze
−
pump
−
thaw
techniques
and stored
in a nitrogen-balanced
cylinder.
These sample
flows were
mixed with an oxygen
flow (Airgas,
Ultra High Purity)
and a
buffer flow of N
2
in a mixing volume
approximately
100 cm
long before
being directed
into a temperature-regulated
double-jacket
flow reactor
(240
−
298
K,
±
2 K). The reactor’s
pressure
was continuously
monitored
and controlled
using a
pressure
gauge (MKS 127AA-00100A,
0
−
100
Torr) and a
throttle
valve (MKS type 153). The gas mixture
underwent
photolysis
at 248 nm via an excimer
laser (Coherent
COMPex
205F, KrF), which was isolated
from the spectrograph
using a
long-pass
filter (Semrock
LP02-257RU-30x40).
Typical
concentrations
in this study were [CH
2
I
2
] = (2.6
−
5.1)
×
10
13
cm
−
3
and [CH
3
I] = (0
−
1.3)
×
10
15
cm
−
3
, with O
2
of 5
−
10 Torr balanced
with N
2
of 10
−
90 Torr and temperatures
set
at either 238 or 294 K.
Analysis
routines
were developed
for simulating
and fitting
the kinetic traces, which made use of the numerical
ordinary
differential
equation
solver from the SciPy library
46
and the
nonlinear
least-squares
fitting library LMFIT.
47
For the Monte
Carlo analysis,
specified
parameters
were varied as standard
normal
random
variables
with a specified
standard
deviation.
Each run was fit with 1024 different
random
inputs, and the
results were compiled
to provide
an estimated
error for the
floating
parameters.
The
Journal
of Physical
Chemistry
Letters
pubs.acs.org/JPCL
Letter
https://doi.org/10.1021/acs.jpclett.4c00159
J. Phys.
Chem.
Lett.
2024,
15,
3690
−
3697
3694
■
ASSOCIATED
CONTENT
*
sı
Supporting
Information
The Supporting
Information
is available
free of charge
at
https://pubs.acs.org/doi/10.1021/acs.jpclett.4c00159.
Supplementary
figures
for experimental
details
and
analyses
(PDF)
Transparent
Peer Review
report available
(PDF)
■
AUTHOR
INFORMATION
Corresponding
Authors
Wen
Chao
−
Division
of Chemistry
and Chemical
Engineering,
California
Institute
of Technology,
Pasadena,
California
91125,
United
States;
orcid.org/0000-0003-0602-1606;
Email: wchao@caltech.edu
Carl J. Percival
−
NASA
Jet Propulsion
Laboratory,
California
Institute
of Technology,
Pasadena,
California
91109-8099,
United
States;
orcid.org/0000-0003-2525-160X;
Email: carl.j.percival@jpl.nasa.gov
Authors
Charles
R. Markus
−
NASA
Jet Propulsion
Laboratory,
California
Institute
of Technology,
Pasadena,
California
91109-8099,
United
States;
orcid.org/0000-0003-2656-
0017
Mitchio
Okumura
−
Division
of Chemistry
and Chemical
Engineering,
California
Institute
of Technology,
Pasadena,
California
91125,
United
States;
orcid.org/0000-0001-
6874-1137
Frank
A. F. Winiberg
−
NASA
Jet Propulsion
Laboratory,
California
Institute
of Technology,
Pasadena,
California
91109-8099,
United
States;
orcid.org/0000-0003-2801-
5581
Complete
contact
information
is available
at:
https://pubs.acs.org/10.1021/acs.jpclett.4c00159
Author
Contributions
W.C. performed
the experiments
and data analysis.
C.R.M.
prepared
the Python
code for model
fitting and error
estimation.
All authors
have given approval
to the final version
of the manuscript.
Notes
The authors
declare
no competing
financial
interest.
■
ACKNOWLEDGMENTS
The experimental
research
herein was carried
out at the Jet
Propulsion
Laboratory,
California
Institute
of Technology,
under contract
with the National
Aeronautics
and Space
Administration
(NASA).
Financial
support
was provided
by
the NASA Upper Atmosphere
Research
and Tropospheric
Chemistry
Programs.
C.R.M.’s
research
was supported
by an
appointment
to the NASA Postdoctoral
Program
at the Jet
Propulsion
Laboratory,
administered
by Oak Ridge Associated
Universities
under contract
with NASA. The authors
thank Dr.
S. Sander for discussion.
■
REFERENCES
(1) Taatjes,
C. A.; Shallcross,
D. E.; Percival,
C. J. Research
Frontiers
in the Chemistry
of Criegee
Intermediates
and Tropospheric
Ozonolysis.
Phys.
Chem.
Chem.
Phys.
2014
,
16
(5), 1704
−
1718.
(2) Khan, M. A. H.; Percival,
C. J.; Caravan,
R. L.; Taatjes,
C. A.;
Shallcross,
D. E. Criegee
Intermediates
and Their Impacts
on the
Troposphere.
Environ.
Sci.: Processes
Impacts
2018
,
20
(3), 437
−
453.
(3) Chhantyal-Pun,
R.; Khan, M. A. H.; Taatjes,
C. A.; Percival,
C.
J.; Orr-Ewing,
A. J.; Shallcross,
D. E. Criegee
Intermediates:
Production,
Detection
and Reactivity.
Int. Rev.
Phys.
Chem.
2020
,
39
(3), 385
−
424.
(4) Zhao, B.; Shrivastava,
M.; Donahue,
N. M.; Gordon,
H.;
Schervish,
M.; Shilling,
J. E.; Zaveri, R. A.; Wang, J.; Andreae,
M. O.;
Zhao, C.; Gaudet,
B.; Liu, Y.; Fan, J.; Fast, J. D. High Concentration
of Ultrafine
Particles
in the Amazon
Free Troposphere
Produced
by
Organic
New Particle
Formation.
Proc.
Natl.
Acad.
Sci. U.S.A.
2020
,
117
(41), 25344
−
25351.
(5) Zhao, Y.; Wingen,
L. M.; Perraud,
V.; Greaves,
J.; Finlayson-
Pitts, B. J. Role of the Reaction
of Stabilized
Criegee
Intermediates
with Peroxy Radicals
in Particle
Formation
and Growth
in Air.
Phys.
Chem.
Chem.
Phys.
2015
,
17
(19), 12500
−
12514.
(6) Sakamoto,
Y.; Inomata,
S.; Hirokawa,
J. Oligomerization
Reaction
of the Criegee
Intermediate
Leads to Secondary
Organic
Aerosol
Formation
in Ethylene
Ozonolysis.
J. Phys.
Chem.
A
2013
,
117
(48), 12912
−
12921.
(7) Gong, Y.; Jiang, F.; Li, Y.; Leisner,
T.; Saathoff,
H. Impact
of
Temperature
on the Role of Criegee
Intermediates
and Peroxy
Radicals
in Dimer Formation
from
β
-Pinene
Ozonolysis.
Atmos.
Chem.
Phys.
2024
,
24
(1), 167
−
184.
(8) Welz, O.; Savee, J. D.; Osborn,
D. L.; Vasu, S. S.; Percival,
C. J.;
Shallcross,
D. E.; Taatjes,
C. A. Direct Kinetic
Measurements
of
Criegee
Intermediate
(CH
2
OO) Formed
by Reaction
of CH
2
I with
O
2
.
Science
2012
,
335
(6065),
204
−
207.
(9) Taatjes,
C. A.; Welz, O.; Eskola, A. J.; Savee, J. D.; Scheer,
A. M.;
Shallcross,
D. E.; Rotavera,
B.; Lee, E. P. F.; Dyke, J. M.; Mok, D. K.
W.; Osborn,
D. L.; Percival,
C. J. Direct Measurements
of Conformer-
Dependent
Reactivity
of the Criegee
Intermediate
CH
3
CHOO.
Science
2013
,
340
(6129),
177
−
180.
(10) Jr-Min Lin, J.; Chao, W. Structure-Dependent
Reactivity
of
Criegee
Intermediates
Studied
with Spectroscopic
Methods.
Chem.
Soc. Rev.
2017
,
46
(24), 7483
−
7497.
(11) Osborn,
D. L.; Taatjes,
C. A. The Physical
Chemistry
of
Criegee
Intermediates
in the Gas Phase.
Int. Rev. Phys.
Chem.
2015
,
34
(3), 309
−
360.
(12) Chao, W.; Yin, C.; Takahashi,
K.; Lin, J. J.-M. Hydrogen-
Bonding
Mediated
Reactions
of Criegee
Intermediates
in the Gas
Phase: Competition
between
Bimolecular
and Termolecular
Reac-
tions and the Catalytic
Role of Water.
J. Phys.
Chem.
A
2019
,
123
(39), 8336
−
8348.
(13) Lester, M. I.; Klippenstein,
S. J. Unimolecular
Decay of Criegee
Intermediates
to OH Radical
Products:
Prompt
and Thermal
Decay
Processes.
Acc. Chem.
Res.
2018
,
51
(4), 978
−
985.
(14) Caravan,
R. L.; Khan, M. A. H.; Rotavera,
B.; Papajak,
E.;
Antonov,
I. O.; Chen, M.-W.; Au, K.; Chao, W.; Osborn,
D. L.; Lin, J.
J.-M.; Percival,
C. J.; Shallcross,
D. E.; Taatjes,
C. A. Products
of
Criegee
Intermediate
Reactions
with NO
2
: Experimental
Measure-
ments and Tropospheric
Implications.
Faraday
Discuss.
2017
,
200
,
313
−
330.
(15) Caravan,
R. L.; Vansco,
M. F.; Au, K.; Khan, M. A. H.; Li, Y.-L.;
Winiberg,
F. A. F.; Zuraski,
K.; Lin, Y.-H.; Chao, W.; Trongsiriwat,
N.; Walsh, P. J.; Osborn,
D. L.; Percival,
C. J.; Lin, J. J.-M.; Shallcross,
D. E.; Sheps, L.; Klippenstein,
S. J.; Taatjes,
C. A.; Lester, M. I. Direct
Kinetic
Measurements
and Theoretical
Predictions
of an Isoprene-
Derived
Criegee
Intermediate.
Proc.
Natl.
Acad.
Sci. U.S.A.
2020
,
117
(18), 9733
−
9740.
(16) Lin, Y.-H.; Yin, C.; Takahashi,
K.; Lin, J. J.-M. Surprisingly
Long Lifetime
of Methacrolein
Oxide, an Isoprene
Derived
Criegee
Intermediate,
under Humid Conditions.
Commun.
Chem.
2021
,
4
(1),
12.
(17) Taatjes,
C. A. Criegee
Intermediates:
What Direct Production
and Detection
Can. Teach Us About Reactions
of Carbonyl
Oxides.
Annu.
Rev. Phys.
Chem.
2017
,
68
(1), 183
−
207.
(18) Chhantyal-Pun,
R.; Welz, O.; Savee, J. D.; Eskola, A. J.; Lee, E.
P. F.; Blacker,
L.; Hill, H. R.; Ashcroft,
M.; Khan, M. A. H.; Lloyd-
Jones, G. C.; Evans, L.; Rotavera,
B.; Huang,
H.; Osborn,
D. L.; Mok,
D. K. W.; Dyke, J. M.; Shallcross,
D. E.; Percival,
C. J.; Orr-Ewing,
A.
The
Journal
of Physical
Chemistry
Letters
pubs.acs.org/JPCL
Letter
https://doi.org/10.1021/acs.jpclett.4c00159
J. Phys.
Chem.
Lett.
2024,
15,
3690
−
3697
3695