S
1
SUPPORTING INFORMATION
Acetonyl
P
eroxy and
H
ydro
P
eroxy
S
elf
-
and
C
ross
-
R
eactions:
Temperature
-
D
ependent
K
inetic
P
arameters,
B
ranching
F
ractions, and
C
haperone
E
ffects
Kristen Zuraski,
a
Fred J. Grieman,
a,
b
Aileen O. Hui,
c
Julia Cowen,
a,
b
Frank A. F. Winiberg,
a
Carl
J. Percival,
a
Mitchio Okumura,
c
and Stanley P. Sander
a
a
NASA Jet Propulsion Laboratory, California Institute of Technology, 4800 Oak Grove Drive,
Pasadena, California 91109, United States
b
Seaver Chemistry Laboratory, Pomona College, Claremont, California 91711, United State
s
c
Arthur Amos Noyes Laboratory of Chemical Physics, California Institute of Technology,
Pasadena, California 91125, United States
Number of pages:
6
Number of Figures
: 2
Number of Tables
: 2
I. Response to
the
2022 Paper by Assali et al. ..................
...
.......................................S
1
II. Discussion of the
k
2
MCMC distributions ..............................................................
.
.S
5
References .....
...............................................................................................S
6
S
2
© 2023. All rights reserved.
I.
Response to
the
2022 Paper by Assali et al
.
1
The recent paper by Assali et al.
1
investigated the
acetonyl peroxy (CH
3
C(O)CH
2
O
2
) self
-
reaction at room
temperature using laser photolysis/continuous wave cavity ring down
spectroscopy. Similar to the experiments presented in our room
-
temperature paper
,
2
CH
3
C(O)CH
2
O
2
radicals
were
generated by the reaction of Cl atoms with acetone
(CH
3
C(O)CH
3
)
following
pulsed
laser photolysis of Cl
2
. Table S1 compar
es
the experimental
conditions, starting radical concentrations, species measured
during the experiments, and
reported values for the self
-
reaction rate coefficients
between the two studies.
In both
studies, the
CH
3
C(O)CH
2
O
2
reactant and HO
2
secondary product are monitored. Additionally,
the
secondary
products,
CH
3
C(O)O
2
and CH
3
O
2
,
were monitored by Assali et al. and the
secondary co
-
product
to HO
2
,
OH
,
were monitored
in our room
temperature study
.
2
Table S1:
Comparison of experiment
al conditions
and reported values
between Assali et al.
1
and
Zuraski et al.
2
Reference
Assali et al.
1
Zuraski et al.
2
Reactant Concentrations
(molecule cm
-
3
)
[Cl
2
]
0
(4
–
9)
×
10
15
(9
–
10)
×
10
15
[Cl]
0
(1.1
–
9)
×
10
13
(1
–
2)
×
10
14
[CH
3
C(O)CH
3
]
0
(0.5
–
7.2)
×
10
16
(1.7
–
2.8)
×
10
16
Species
Monitored
CH
3
C(O)CH
2
O
2
,
CH
3
C(O)O
2
, CH
3
O
2
, HO
2
CH
3
C(O)CH
2
O
2
, OH, HO
2
P(Torr), T (K)
100 (O
2
), RT
100 (N
2
), RT
Reported values
rate coefficient,
k
2
(
cm
3
s
-
1
)
(5.4
± 1.4)
× 10
-
12
(4.8 ± 0.
8
)
× 10
-
12
branching fraction,
k
2b
/
k
2
0.6 ± 0.1
0.33 ± 0.13
S
3
The reported rate constants,
k
2
, were in agreement between these two studies, h
owever,
the reported
branching fractions to the alkoxy pathway
(which leads to the OH and HO
2
secondary products),
k
2b
/
k
2
,
did not agree. The disagreement between
the reported
k
2b
/
k
2
values
was explained by Assali et al. to be the result of two additional reactions in the chemical
mechanism, R1
-
R2
.
The first, R1a
-
R1c
, describes the reaction between Cl atoms and the
CH
3
C(O)CH
2
O
2
radicals
. R1
has not been previously measured and the suggested product
pathways are based on the analogous reaction between Cl atoms and CH
3
C(O)O
2
.
In addition to
R1, Assali
et al. added the reaction between the Criegee intermediate and the acetone precursor,
R2, to their reaction mechanism. Without the inclusion of these reactions, they could not
reproduce the kinetic profiles for HO
2
observed in their work
.
T
herefore, they
includ
ed
R1 and
R2 to
their reaction mechanism
through necessity
to fit their observed HO
2
kinetic profiles
and
reported the rate coefficients for these reaction as (1.35 ± 0.8)
×
10
-
10
and (4.5 ± 2.0)
×
10
-
14
cm
3
s
-
1
, respectively.
CH
3
C(O)CH
2
O
2
+ Cl
→
CH
3
C(O)CH
2
O
+ ClO
(R1a)
→
CH
3
C(O)CHO
2
+ HCl
(R1b)
→
CH
3
C(O)CHO
+ HClO
(R1c)
CH
3
C(O)CH3 + CH
3
C(O)CHO
2
→
HO
2
+ products
(R2)
Assali et al. suggested these reactions to be non
-
negl
igible under our experimental
conditions.
As R1 and R2
were not included in the
previous
analysis for our room temperature
work
,
2
we
investigated the importance of
these reactions
on our work
.
As shown in Figure S1,
the addition of these reactions
drastically changes the simulated kinetic profiles. However,
adding these reactions makes fitting our observed data to not be feasible. The [CH
3
C(O)CH
2
O
2
]
S
4
would have to be incorrect
by a factor of two and the alkoxy pathway would require a branching
fracti
on of ~1.25 (a non
-
physical value) to match our observations.
Figure S1.
(black) Kinetic traces of (a) CH
3
C(O)CH
2
O
2
, (b) HO
2
, and (c) OH observed at T =
298 K and P = 100 Torr N2. (Red, solid line) Output of our kinetic model simulation based on
the chem
ical mechanism and our reported values for the rate coefficient and branching fraction
from our room temperature paper.
2
(Green, dashed line) Output of our kinetic model simulation
with the addition of reactions R1 and R2 using reported values the rate coe
fficient and branching
fraction from Assali et al.
1
In both studies, determination of
k
2b
/
k
2
results from analysis of secondary chemistry as
oppose to monitoring the direct products of the R2 reaction. Our analysis used a MCMC
approach which allows all of the uncertainties in our reaction mechanism to be sampled during
the fit. In the work by Ass
ali et al., the MCMC approach was not used in the analysis and, as a
result, large uncertainties in some of the reactions in their simulations were not considered.
Through necessity to
fit their observed
kinetic profiles
for HO
2
resulting from the secondar
y
chemistry of the CH
3
C(O)CH
2
O
2
self
-
reaction, they instead had to add R1 and R2 to their
reaction mechanism and use analogous reactions to estimate the rate coefficients of these
reactions
.
Without a more direct measurement of these rate coefficients and
because inclusion of
these reactions with the rate coefficients proposed by Assali et al.
resulted in non
-
feasible fits to
our data
, we have decided to not include them in our analysis
. Furth
er
more, the unfeasibility
of
Assali’s values for
k
2b
/
k
2
would onl
y become more unfeasible during the analysis of the lower
1.4 x 10
14
1.2
1
.8
.6
.4
.2
0
[CH
3
C(O)CH
2
O
2
], molecules cm
-3
5
4
3
2
1
0
Time, ms
10 x 10
12
8
6
4
2
0
[HO
2
], molecules cm
-3
14
12
10
8
6
4
2
0
Time, ms
250 x 10
9
200
150
100
50
0
-50
[OH], molecule cm
-3
2.0
1.5
1.0
0.5
0.0
Time, ms
(a)
(b)
(c)
S
5
temperature results, where the
k
2b
/
k
2
increases by 60% compared to the value at room
temperature.
II.
Discussion of the
k
2
MCMC distributions
At
T = 270
–
290 K
,
a bimodal distribution was observed in the MCMC outputs of the
k
2b
/
k
2
branching fraction. Values for the geometric means and uncertainties (resulting from twice
the full
-
width half
-
max of the Gaussian fits to the distributions) for all of the distributions
are
given in Table S2. The distributions with the lower
k
2b
/
k
2
values are
wide
r (reflecting
significantly higher uncertainty
).
As shown in Figure S
2
, the lower
k
2b
/
k
2
values are also
inconsistent with the trend
line
of the higher temperature values.
Table
S2
.
Results from the MC simulations for the acetonyl peroxy self
-
reaction.
2
s
uncertainties are given in the parenthesis.
T (K)
k
2
,
10
-
12
cm
3
molecule
-
1
s
-
1
Branching Fraction to
Alkoxy Channel (R2b)
Secondary Branching Fraction
to Alkoxy Channel (R2b)
330
(1)
2.85 (0.76)
0.19 (0.07)
320
(1)
3.19 (0.38)
0.23 (0.05)
310
(1)
3.97 (0.69)
0.29 (0.07)
298
(2)
*
4.8 (0.8)
0.33 (0.13)
290
(1)
5.05 (0.59)
0.48 (0.06)
0.35 (0.08)
280
(1)
5.45
(0.51)
0.51 (0.05)
0.26 (0.12)
270
(1)
5.88 (0.44)
0.53 (0.05)
0.13 (0.10)
*Values from Zuraski
et al.
17
S
6
Figure S
2
.
(a)
Arrhenius plot for the temperature dependence of
k
2
(black, filled circles) and (b)
the linear temperature dependence (red, filled circles) for the branching fraction,
k
2b
/
k
2
.
The
values and uncertainties for the secondary outputs (red, open triangles) are shown alongside the
linear temperature dependence.
In contrast to the room temperature results for this reaction,
2
the branching fractions for
R2a and R2c could not be disentangled during the analysis for the temperature results.
The work
by Berndt et al.
3
would be an example of work that, if extended t
o
higher and
lower
temperatures, could allow this work to be reanalyzed to determine the branching fraction of all
three pathways
.
Results of work studying one of the other pathways for this reaction could also
allow
for an additional constraint on the R2b
outputs, which would help to rule out
potentially
unfeasible values
from our MCMC outputs
.
REFERENCES
1.
Assali, M.; Fittschen, C. Self
-
reaction of acetonyl peroxy radicals and their reaction with
Cl atoms. J. Phys. Chem. A.
2022
,
126
, 4585
-
4597.
2.
Zuraski, K.; Hui, A. O.; Grieman, F. J.; Darby, E.; Møller, K. H.; Winiberg, F. A. F.;
Percival, C.J.; Smarte, M. D.; Okumura, M.; Kjaergaard, H. G.; Sander, S. P. Acetonyl
peroxy and hydro peroxy self
-
and cross
-
reactions: Kinetics, Mechanism, and Chapero
ne
Enhancement from the perspective of the hydroxyl radical product.
J. Phys. Chem. A
,
2020
,
124
, 8128
-
8143.
3.
Berndt, T.
;
Scholz,
W
.;
Mentler,
B.;
Fischer, L
.;
Herrmann,
H.;
Kulmala, M.
; Hansel, A.
Accretion Product Formation from Self
-
and Cross
-
Reactions
of RO
2
Radicals in the
Atmosphere.
Angew. Chem. Int. Ed.
2018,
57
.
-27.0
-26.5
-26.0
ln (
k
2
cm
3
molecule
-1
s
-1
)
3.8
3.6
3.4
3.2
3.0
1000K / T
0.6
0.5
0.4
0.3
0.2
0.1
0.0
Branching Fraction,
k
2b
/
k
2
330
320
310
300
290
280
270
Temperature, K
(a)
(b)