Four-Carbon Criegee Intermediate from Isoprene Ozonolysis: Methyl Vinyl Ketone Oxide Synthesis, Infrared Spectrum, and OH Production

Article abstract of DOI:10.1021/jacs.8b06010

The reaction of ozone with isoprene, one of the most abundant volatile organic compounds in the atmosphere, produces three distinct carbonyl oxide species (RR′COO) known as Criegee intermediates: formaldehyde oxide (CH₂OO), methyl vinyl ketone oxide (MVK-OO), and methacrolein oxide (MACR-OO). The nature of the substituents (R,R′ = H, CH₃, CH=CH₂) and conformations of the Criegee intermediates control their subsequent chemistry in the atmosphere. In particular, unimolecular decay of MVK-OO is predicted to be the major source of hydroxyl radicals (OH) in isoprene ozonolysis. This study reports the initial laboratory synthesis and direct detection of MVK-OO through reaction of a photolytically generated, resonance-stabilized monoiodoalkene radical with O₂. MVK-OO is characterized utilizing infrared (IR) action spectroscopy, in which IR activation of MVK-OO with two quanta of CH stretch at ca. 6000 cm^-1 is coupled with ultraviolet detection of the resultant OH products. MVK-OO is identified by comparison of the experimentally observed IR spectral features with theoretically predicted IR absorption spectra. For syn-MVK-OO, the rate of appearance of OH products agrees with the unimolecular decay rate predicted using statistical theory with tunneling. This validates the hydrogen atom transfer mechanism and computed transition-state barrier (18.0 kcal mol^-1) leading to OH products. Theoretical calculations reveal an additional roaming pathway between the separating radical fragments, which results in other products. Master equation modeling yields a thermal unimolecular decay rate for syn-MVK-OO of 33 s^-1 (298 K, 1 atm). For anti-MVK-OO, theoretical exploration of several unimolecular decay pathways predicts that isomerization to dioxole is the most likely initial step to products.

Full text of DOI:10.1021/jacs.8b06010

Subscriber access provided by Kaohsiung Medical University
Article
Four Carbon Criegee Intermediate from Isoprene Ozonolysis: Methyl
Vinyl Ketone Oxide Synthesis, Infrared Spectrum, and OH Production
Victoria P. Barber, Shubhrangshu Pandit, Amy M Green, Nisalak
Trongsiriwat, Patrick J Walsh, Stephen J Klippenstein, and Marsha I. Lester
J. Am. Chem. Soc., Just Accepted Manuscript • Publication Date (Web): 03 Aug 2018
Downloaded from http://pubs.acs.org on August 3, 2018
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 57
Journal of the American Chemical Society
1
2
3
4
5
Four Carbon Criegee Intermediate from Isoprene Ozonolysis:
6
7
Methyl Vinyl Ketone Oxide Synthesis, Infrared Spectrum, and OH Production
8
9
Victoria P. Barber1, Shubhrangshu Pandit1, Amy M. Green1, Nisalak Trongsiriwat1, Patrick J. Walsh1,
Stephen J. Klippenstein2, and Marsha I. Lester1*
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1Department of Chemistry, University of Pennsylvania
Philadelphia, PA 19104ꢀ6323
2Chemical Sciences and Engineering Division, Argonne National Laboratory
Argonne, IL, 60439
Abstract
The reaction of ozone with isoprene, one of the most abundant volatile organic compounds in the
atmosphere, produces three distinct carbonyl oxide species (RR′COO) known as Criegee intermediates:
formaldehyde oxide (CH2OO), methyl vinyl ketone oxide (MVKꢀOO), and methacrolein oxide (MACRꢀ
OO). The nature of the substituents (R,R′ = H, CH3, CH=CH2) and conformations of the Criegee
intermediates control their subsequent chemistry in the atmosphere. In particular, unimolecular decay of
MVKꢀOO is predicted to be the major source of hydroxyl radicals (OH) in isoprene ozonolysis. This
study reports the initial laboratory synthesis and direct detection of MVKꢀOO through reaction of a
photolyticallyꢀgenerated, resonanceꢀstabilized monoiodoalkene radical with O2. MVKꢀOO is
characterized utilizing infrared action spectroscopy, in which infrared activation of MVKꢀOO with two
quanta of CH stretch at ca. 6000 cmꢀ1 is coupled with ultraviolet detection of the resultant OH products.
MVKꢀOO is identified by comparison of the experimentally observed IR spectral features with
theoretically predicted IR absorption spectra. For synꢀMVKꢀOO, the rate of appearance of OH products
agrees with the unimolecular decay rate predicted using statistical theory with tunneling. This validates
the hydrogen atom transfer mechanism and computed transition state barrier (18.0 kcal molꢀ1) leading to
OH products. Theoretical calculations reveal an additional roaming pathway between the separating
radical fragments, which results in other products. Master equation modeling yields a thermal
* Corresponding author email: milester@sas.upenn.edu
1
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 57
1
2
3
4
5
6
7
8
unimolecular decay rate for synꢀMVKꢀOO of 33 sꢀ1 (298 K, 1 atm). For antiꢀMVKꢀOO, theoretical
exploration of several unimolecular decay pathways predicts that isomerization to dioxole is the most
likely initial step to products.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
ACS Paragon Plus Environment

Page 3 of 57
Journal of the American Chemical Society
1
2
3
4
1. Introduction
5
6
7
8
Isoprene (2ꢀmethylꢀ1,3ꢀbutadiene) is the most abundant volatile organic compound (VOC) in the
Earth’s troposphere after methane, with global emissions estimated at ca. 500 Tg/year.1 In the
atmosphere, the predominant removal pathways for isoprene (and other alkenes) are by reaction with the
hydroxyl radical (OH), ozone (O3), and the nitrate radical (NO3; nighttime only).2ꢀ4 Globally, ozonolysis
removes ca. 10% of isoprene,5 and is an important pathway to OH radicals during daylight and nighttime
conditions.6 The OH radical is the most important oxidant in the lower atmosphere and is sometimes
referred to as “the atmosphere’s detergent”.7 The OH yield from isoprene ozonolysis, however, is not
well determined and may be pressure dependent.8 In the earlier works, experimentally measured OH
yields ranged from as low as ca. 20% to as high as ca. 70% depending on the method used for OH
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
detection.8ꢀ13 Recent measurements indicate an overall OH yield from isoprene ozonolysis of ca. 25%.5,
14ꢀ15
Scheme 1. Criegee intermediates and coꢀproducts formed in ozonolysis of isoprene.
Alkene ozonolysis proceeds via highly energized carbonyl oxide species, termed Criegee
intermediates, which represent a critical branching point controlling the products formed from this
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 57
1
2
3
4
5
6
7
8
9
important class of reactions. The alkene ozonolysis reaction (Scheme 1) begins by cycloaddition of O3
across a C=C double bond to form a primary ozonide (POZ).7 Isoprene has two distinct double bonds and
can therefore produce two different POZs. Due to the asymmetry of each double bond, subsequent
decomposition of each POZ can produce two Criegee intermediates and their corresponding carbonyl
coproducts.5, 9, 16ꢀ17 Ozonolysis at the C(1)=C(2) bond results in CH2OO or methyl vinyl ketone oxide
((CH2=CH)(CH3)COO, MVKꢀOO)18 with their corresponding coproducts, methyl vinyl ketone
((CH2=CH)(CH3)CO, MVK) or formaldehyde (CH2O), respectively. Ozonolysis at the C(3)=C(4) bond
results in either CH2OO or methacrolein oxide ((CH2=C(CH3))CHOO, MACRꢀOO) with their
corresponding coproducts, methacrolein ((CH2=C(CH3))CHO, MACR) or formaldehyde, respectively.
Combined experimental and theoretical evidence indicates an overall branching for isoprene ozonolysis of
ca. 58% CH2OO, ca. 23% MVKꢀOO and ca. 19% MACRꢀOO.5, 19 This work focuses on MVKꢀOO,
which is predicted to be the more abundant fourꢀcarbon unsaturated Criegee intermediate and the major
source of OH radicals from atmospheric isoprene ozonolysis.5 MVKꢀOO is expected to differ
substantially from previously investigated Criegee intermediates because of its extended π conjugation
across the vinyl and carbonyl oxide functional groups.20 MVKꢀOO is predicted to exhibit four
conformers as shown in Figure 1. Here, syn and anti refer to the orientation of the methyl group with
respect to the terminal oxygen of the carbonyl oxide group, while cis and trans refer to the orientation of
the C=C and C=O bonds with respect to one another. The relative stabilities of the four conformers of
MVKꢀOO differ by only ca. 3 kcal molꢀ1, yet significant barriers restrict interconversion between them, as
discussed below.16
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
4
ACS Paragon Plus Environment

Page 5 of 57
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. Optimized structures and zeroꢀpoint corrected energies (kcal molꢀ1) computed for the four
conformers of the methyl vinyl ketone oxide (MVKꢀOO) Criegee intermediate. Geometries are
optimized at the B2PLYPD3/ccꢀpVTZ level of theory. Energies are calculated at the ANL0ꢀB2F level
of theory (Table S2).
In general, the atmospheric fate of the RR′COO Criegee intermediates is highly dependent on the
conformation and nature of its substituents (R, R′).21 In the atmosphere, Criegee intermediates can
undergo prompt or thermal unimolecular decay to OH radicals and other products.7, 22ꢀ25 Alternatively,
collisionally stabilized Criegee intermediates can undergo bimolecular reaction with other atmospheric
species, including water vapor and SO2,26ꢀ32 which may result in aerosol formation and impact climate.33
For synꢀalkyl substituted Criegee intermediates, such as synꢀCH3CHOO, (CH3)2COO, and the syn
conformers of MVKꢀOO, unimolecular decay to OH is predicted to be the dominant atmospheric loss
process.24, 34ꢀ36 This reaction proceeds via a 1,4 Hꢀatom transfer mechanism, whereby a synꢀalkyl αꢀH
atom transfers to the terminal oxygen of the carbonyl oxide group via a 5ꢀmembered ring transition state
(TS) to produce a vinyl hydroperoxide (VHP) species, which subsequently decays to generate OH, as
shown in Scheme 2.23, 37ꢀ39 Criegee intermediates lacking a synꢀalkyl αꢀH atom, such as CH2OO, antiꢀ
CH3CHOO, anti conformers of MVKꢀOO (which have a more strongly bound vinylic αꢀH atom), and
5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 57
1
2
3
4
5
6
7
8
9
MACRꢀOO (which has no αꢀH atom) may decay unimolecularly via alternative pathways, but these are
not anticipated to result in significant OH production.16ꢀ17, 21, 40ꢀ41 In particular, Criegee intermediates with
a vinyl substituent in the syn position, such as antiꢀMVKꢀOO or synꢀMACRꢀOO, are predicted to undergo
an electrocyclic ringꢀclosing reaction via a low barrier to form a 5ꢀmembered ring structure (cycꢀ
CH2OOC(CH3)CH, dioxole).5, 16, 21 As a result, the syn conformers of MVKꢀOO are predicted to be the
major source of OH radicals from isoprene ozonolysis.5
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 2. Unimolecular decay of synꢀalkyl substituted Criegee intermediates to OH radical.
For Criegee intermediates with only an H atom substituent in the syn position, such as CH2OO or
antiꢀCH3CHOO, reactions with the water monomer and dimer are predicted to be the most important
removal pathways in the atmosphere.28, 30ꢀ31 However, for MVKꢀOO, reactions with water and water
dimer are theoretically predicted to be very slow for all four conformers.42ꢀ45 Indeed, experimentally,
there is no evidence for reaction between water vapor and MVKꢀOO during isoprene ozonolysis.5
In 2012, a new alternative synthetic route enabled direct experimental measurements of various
small, prototypical Criegee intermediates.46ꢀ47 In this method, a gemꢀdiiodo alkane precursor is
photolyzed to produce a monoiodo alkyl radical, which subsequently reacts with O2 to produce the
Criegee intermediate. This class of reactions has proven useful for the production of a variety of Criegee
intermediates including CH2OO,47ꢀ48 CH3CHOO,46, 49ꢀ50 (CH3)2COO,51ꢀ52 and CH3CH2CHOO.52 The fourꢀ
carbon MVKꢀOO and MACRꢀOO Criegee intermediates generated in isoprene ozonolysis, which are
arguably the most atmospherically relevant, have not been synthesized via this same method, because
their corresponding gemꢀdiiodo precursors are extremely reactive and cannot be prepared and purified.
As a result, there has been no direct detection of MVKꢀOO prior to this report.
6
ACS Paragon Plus Environment

Page 7 of 57
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Recently, this group has studied several small prototypical Criegee intermediates, including synꢀ
CH3CHOO,39 (CH3)2COO,38 and synꢀCH3CH2CHOO,53 using IR action spectroscopy with detection of
OH radical products under collisionꢀfree conditions. IR activation of Criegee intermediates with two
quanta of CH stretch (2νCH) accesses energies in the vicinity of the TS barrier to the 1,4 Hꢀatom transfer
reaction, and provides sufficient energy for the Criegee intermediates to surmount or tunnel through this
barrier enroute to OH products. This laboratory has also carried out rate measurements for unimolecular
decay of the Criegee intermediates using IR activation with timeꢀresolved detection of the resultant OH
radical products.36, 53 The experimentally determined OH appearance rates were found to be in excellent
agreement with complementary statistical RiceꢀRamspergerꢀKasselꢀMarcus (RRKM) calculations
utilizing high level theoretical evaluation of the TS barrier height and incorporating quantum mechanical
tunneling.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
We anticipate that the synꢀMVKꢀOO conformers will undergo a similar alkyl Hꢀatom transfer
mechanism to form OH radical products. As a result, we employ the sensitive and selective IR action
spectroscopy technique to study the unimolecular decay of MVKꢀOO to OH. Several groups have
predicted theoretical barriers for this process for the syn conformers of MVKꢀOO. For synꢀtransꢀMVKꢀ
OO, these predictions range from 17.5 to 18.8 kcal molꢀ1 (6120 to 6575 cmꢀ1),6, 16, 21, 44 depending on the
level of theory used, while for synꢀcisꢀMVKꢀOO, they range from 16.9 to 18.4 kcal molꢀ1 (5910 to 6435
cmꢀ1).6, 16, 44 [Note that barrier, well depth, and other energy terms utilized herein refer to zeroꢀpoint
(ZPE) corrected values.] IR activation of MVKꢀOO in the CH stretch overtone region at ca. 6000 cmꢀ1 is
therefore expected to provide sufficient energy to surmount or tunnel through the reaction barrier and lead
to OH products.
In this paper, we present the first direct observation of the MVKꢀOO Criegee intermediate. We
introduce a new synthetic method that is used to produce the MVKꢀOO Criegee intermediate. We report
the IR action spectrum of MVKꢀOO in the CH stretch overtone region with detection of OH radical
products, which provides an IR fingerprint for MVKꢀOO and demonstrates that IR activation of MVKꢀ
OO leads to OH and other products (OH products). Additionally, we present experimental measurements
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 57
1
2
3
4
5
6
7
8
of the energyꢀdependent unimolecular decay rates for MVKꢀOO to OH products. Finally, we report
theoretical calculations of electronic structure, vibrational frequencies, unimolecular decay mechanisms,
and microcanonical and thermal decay rates for MVKꢀOO.
9
2. Methods
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2.1. Synthesis
A novel synthetic route is used to generate the methyl vinyl ketone oxide (MVKꢀOO) Criegee
intermediate starting from a (Z/E)ꢀ1,3ꢀdiiodobutꢀ2ꢀene precursor as introduced in Scheme 3. The
synthesis and characterization of the precursor are described in detail in the Supporting Information (SI,
Sec. S1.1 and Figures S1ꢀS3). Photolysis of 1,3ꢀdiiodobutꢀ2ꢀene at 248 nm results in preferential
dissociation of the weaker allylic C(1)ꢀI bond, rather than the vinylic (sp2ꢀhybridized) C(3)ꢀI bond, due to
resonance stabilization of the resultant allylic monoiodoalkene radical product Int(1).54
Scheme 3. Alternate synthetic route to MVKꢀOO.
The unique aspect of the present approach is based on the two resonance structures of the allylic
monoiodoalkene radical Int(1). Although the radical center is initially formed at the C(1) carbon,
delocalization will lead to the preferred radical site on the C(3) carbon due to its higher degree of
substitution.54 Subsequent reaction with O2 is barrierless at the more stable carbon C(3) radical site as
explained below. Addition of O2 is expected to transiently form an energized iodoalkene peroxy radical
8
ACS Paragon Plus Environment

Page 9 of 57
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
Int(2), which can readily rotate about the C−C and C−O bonds before dissociation of I resulting in
formation of the four conformers of MVKꢀOO. The O2 reaction step is similar to that leading to other
Criegee intermediates from their corresponding geminal diiodo precursors.47ꢀ49, 52, 55
9
Our theoretical calculations (see SI Sec. S1.2) examine the minimum energy pathway (MEP) for
O2 addition to Int(1) at the C(3) vs. C(1) site. As shown in Figure S4, a barrierless MEP is found for O2
addition at C(3), while a saddle point is identified along the MEP for C(1), the latter causing a rateꢀlimiting
bottleneck for O2 addition. As a result, the barrierless O2 addition to Int(1) at the C(3) site is strongly
favored both energetically and kinetically over addition at C(1).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2.2. Experimental Methods
The production of Criegee intermediates in a pulsed jet expansion has been described
previously.48ꢀ49, 52 In the present study, the diiodoalkene precursor is heated (60 °C) in a pulsed valve
(Parker General Valve Series 9) with a Peltier thermoelectric heating module (Laird Technologies, PC4).
The temperature is monitored with a thermocouple (ColeꢀParmer, Type K digital thermometer). Heating
increases the vapor pressure of the precursor and substantially increases the MVKꢀOO signal. The
precursor is seeded in a 10% O2/Ar carrier gas (30 psig), and pulsed through a nozzle (1 mm orifice) into
a quartz capillary tube reactor (ca. 25 mm length, 1 mm ID). The precursor is photolyzed along the length
of the capillary using the cylindrically focused 248 nm output (25 mJ pulseꢀ1) of a KrF excimer laser,
which induces CꢀI bond dissociation. Subsequent reaction of the resonanceꢀstabilized monoiodoalkene
radical with O2 forms the MVKꢀOO Criegee intermediate. MVKꢀOO is initially detected by
photoionization using 10.5 eV (118 nm) vacuum ultraviolet (VUV) radiation on the m/z = 86 mass
channel of a timeꢀofꢀflight mass spectrometer (TOFꢀMS, RM Jordan) as shown in Figure S6 and
described in SI (Sec. S2; computed ionization energies for MVKꢀOO conformers are given in Table S1).
For IR pump – UV probe studies, MVKꢀOO is collisionally stabilized in the capillary and cooled
in a free jet expansion to a rotational temperature of ~10 K.38ꢀ39 Approximately 1 cm downstream in the
collision free region, the gas mixture is intersected by counterpropagating and spatially overlapped IR
pump and UV probe laser beams. The IR pump and UV probe lasers are focused to diameters of
9
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 57
1
2
3
4
5
6
7
8
approximately 2 and 5 mm, respectively, in the interaction region. Tunable IR radiation is used to excite
MVKꢀOO in the CH stretch overtone (2νCH) region, initiating unimolecular decay, and the UV laser then
probes the resultant OH X2Π3/2 (v=0) products via laser induced fluorescence (LIF).
9
The tunable IR radiation (~20 mJ/pulse, 6 ns FWHM) used for spectroscopic experiments is the
signal output of an optical parametric oscillator/amplifier (OPO/OPA, Laservision; 0.15 cmꢀ1 bandwidth),
pumped by an injectionꢀseeded Nd:YAG laser (Continuum Precision II 8000, 10 Hz). For rate
measurements, the OPO/OPA is pumped with an unseeded Nd:YAG laser, yielding tunable IR radiation
of ~30 mJ/pulse with a 0.9 cmꢀ1 bandwidth. The UV radiation (~4 mJ/pulse, 6 ns FWHM) is generated by
frequency doubling the output of an Nd:YAG (Continuum 7020; 532 nm, 20 Hz) pumped dye laser
(ND6000, Rhodamine 590 dye). The UV radiation is fixed on the OH A2Σ+ ꢀ X2Π3/2 (1,0) Q1(3.5)
transition. Fluorescence is collected on the OH A2Σ+ ꢀ X2Π3/2 (1,1) band with a gated photomultiplier tube
(Electron Tubes 9813 QB). The output signal is preamplified and displayed on a digital storage
oscilloscope (LeCroy WaveRunner 44Xi) interfaced with a computer for processing. An active
background subtraction scheme (IR onꢀIR off) is used to remove background OH, which arises from the
unimolecular decay of energized Criegee intermediates in the capillary and subsequent cooling in the free
jet expansion.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Two types of experiments are performed: (1) The IR pump laser is scanned over the 2νCH energy
region (5750ꢀ6260 cmꢀ1) at a fixed IRꢀUV time delay. Here, the time delay is set to 800 ns, which
provides the optimal signal. The experimental method for acquiring IR action spectra of Criegee
intermediates has been described previously.34ꢀ35, 38ꢀ39 (2) The IR pump wavelength is fixed on a
particular feature in the IR action spectrum, and the time delay between the IR pump and UV probe lasers
is stepped in 20 ns increments to measure the time dependent appearance rate of OH radical products
from IRꢀactivated MVKꢀOO, as described previously for other Criegee intermediates.34ꢀ36, 53
2.3. Theoretical Methods
Complementary electronic structure calculations are carried out to characterize the energies of
each MVKꢀOO conformer, barriers to conformational isomerization, as well as transition state (TS)
10
ACS Paragon Plus Environment

Page 11 of 57
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
barriers and products of various unimolecular reactions. The present ab initio calculations were mostly
performed with a slightly modified version of the ANL0ꢀF12 approach described previously.56 This
approach (which we label as ANL0ꢀB2F for ease of reference) is based on estimates of the complete basis
set (CBS) CCSD(T) limit, (via CCSD(T)ꢀF12/CBSꢀF12(TZꢀF12,QZꢀF12)//B2PLYPꢀD3/ccꢀpVTZ
calculations) together with a series of corrections as described in the SI (Sec. S3). It differs from the
ANL0ꢀF12 approach by the replacement of the CCSD(T)/ccꢀpVTZ harmonic and B3LYP anharmonic
rovibrational analyses with a B2PLYPꢀD3/ccꢀpVTZ analyses. We have also chosen to neglect the
diagonal Born Oppenheimer correction (DBOC) since it is often difficult to evaluate for transition states
and rarely significant.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Our prior studies of related Criegee intermediates dissociations indicate that for rovibrational
properties the B2PLYPꢀD3 approach is an effective alternative to CCSD(T) evaluations, at least for the
initial H transfer from a Criegee intermediate to a vinylhydroperoxide.36 Furthermore, for a set of 147
molecules the CCSD(T)/CBS/B2PLYPꢀD3 energies are lower than the CCSD(T)/CBS//CCSD(T)/ccꢀ
pVTZ ones, indicating that the B2PLYPꢀD3 geometries are more accurate that CCSD(T)/ccꢀpVTZ ones.
Meanwhile, for a set of 82 molecules the RMSD for B2PLYPꢀD3/ccꢀpVTZ and CCSD(T)/CBS ZPE
values is only 0.11 kcal molꢀ1. Thus, we expect the accuracy for the ANL0ꢀB2F approach to be fairly
similar to that for the ANL0ꢀF12 approach, i.e. 2σ error bars of ~0.2ꢀ0.3 kcal molꢀ1 for species with at
most modest multireference character56 (e.g., CCSDT(Q) corrections of 1.2 kcal molꢀ1 or less). It is not
clear how much the uncertainty is increased for species such as the Criegee intermediates of interest here
where the multireference character is quite significant (perhaps a factor of two larger is a reasonable
estimate; cf. Fig. 14 of Ref. 56). It is perhaps worth noting though that our prior calculations with this
methodology for various Criegee intermediate dissociations have yielded barrier heights that agree with
experiment to much better than even the smaller uncertainties.36, 53 The present ANL0ꢀB2F stationary
point energies are summarized in Table S2.
We are also interested in a number of aspects of the potential energy surface for which single
reference methods are not appropriate. In particular, pathways involving stationary points with radical
11
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 57
1
2
3
4
5
6
7
8
9
pairs or diradical configurations (see Tables S3ꢀS5) require a multireference description. To explore the
energetics of each of these pathways, we employ complete active space wave functions with both multiꢀ
reference second order perturbation theory (CASPT2) and multiꢀreference singles and doubles
configuration interaction (MRCISD+Q) including a fixed Davidson correction for the effect of higher
order excitations. The uncertainties in the energies from these multireference calculations are likely much
larger but still quite modest; i.e., 2σ error bars of 1 kcal molꢀ1 or less (depending on the configuration).
The electronic structure calculations described here were performed with Gaussian 0957 (for the B2PLYPꢀ
D3 calculations), MOLPRO58 (for the CASPT2, MRCISD+Q, and CCSD(T) evaluations), and CFOUR59
(for most of the CCSDT(Q) calculations).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The IR spectra of MVKꢀOO in the 2νCH energy region is computed at the B2PLYPꢀD3/ccꢀpVTZ
level of theory60ꢀ62 using vibrational secondꢀorder perturbation theory (VPT2) as implemented in Gaussian
09,57 which has been previously demonstrated to provide highly accurate anharmonic vibrational
frequencies for other Criegee intermediates.34ꢀ35, 63 The results of these calculations are given in Table S6.
For kinetic rate calculations, the rovibrational parameters (anharmonic fundamental vibrational
frequencies, anharmonicity matrix and potential energy surfaces for methyl and vinyl torsional motion) of
the reactants and transition states are also calculated at the same B2PLYPꢀD3/ccꢀpVTZ level of theory.
Calculated anharmonic fundamental vibrational frequencies are given in Table S7 for the reactants and
Table S8 for the TS.
Statistical RRKM theory is utilized to compute the energyꢀdependent unimolecular decay rate
k(E) of synꢀMVKꢀOO. The microcanonical rate constant k(E) is given by64
‡
ꢄ
ꢅꢆꢆ ꢇ ꢁꢂ − ꢂꢈꢃ
ꢁ ꢃ
ꢀ ꢂ =
ꢄꢅ‡ꢆꢆ
ℎꢉꢁꢂꢃ
where E0 is the TS barrier energy, G‡(EꢀE0) is the sum of vibrational states at the TS, N(E) is the density
of vibrational states in synꢀMVKꢀOO, and h is Planck’s constant. The effective symmetry numbers σeff
are 1 for the reactants with a plane of symmetry, and ½ for the transition states with no symmetry and 2
enantiomers. The RRKM calculations utilize the computed energies and rovibrational properties of synꢀ
12
ACS Paragon Plus Environment

Page 13 of 57
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
MVKꢀOO, the TS, and products. The sums and densities of states are evaluated using anharmonic
vibrational frequencies. The methyl and vinyl torsional modes of the synꢀMVKꢀOO reactants and the
vinyl torsional mode of the TS are treated as oneꢀdimensional hindered rotors, as described in SI (Sec.
S5). Tunneling is included in the RRKM calculations using either a oneꢀdimensional asymmetric Eckart
model as implemented in MESS65ꢀ66 or semiclassical transition state theory (SCTST)67ꢀ72 as implemented
in Multiwell.73ꢀ75 Additional RRKM calculations are carried out for unimolecular decay of antiꢀMVKꢀ
OO via ring closure to dioxole as described in SI (Sec. S5).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3. Results
The present study focuses on the IR action spectroscopy of MVKꢀOO, and the ensuing
unimolecular decay dynamics and OH radical production. The experimental IR action spectrum of MVKꢀ
OO with detection of OH products is obtained by scanning the IR pump laser in the CH stretch overtone
(2νCH) region at a fixed IR pumpꢀUV probe time delay. In separate experiments, the IRꢀUV time delay is
stepped to determine the rate of OH appearance following the vibrational activation of MVKꢀOO. In both
types of experiments, the OH (v=0, J=3.5) products are detected by laserꢀinduced fluorescence (LIF) on
the OH A2Σ+ꢀX2Π3/2 (1,0) Q1(3.5) transition to obtain the best signalꢀtoꢀbackground ratio; the OH signal
intensity peaks at J=2.5. Complementary theoretical calculations are carried out to predict the IR
absorption spectra of the MVKꢀOO conformers (Sec. 3.2) and energyꢀdependent unimolecular
dissociation rates k(E) from synꢀMVKꢀOO conformers to OH products (Sec. 3.3).
3.1. MVKꢀOO Conformers
The minimum energy structures and ZPE corrected energies of the four conformers of MVKꢀOO
are shown in Figure 1, and are in good agreement with previous theoretical predictions.6, 16 The four
MVKꢀOO conformers have similar energies with the highest energy antiꢀcis conformer predicted to lie
only 3.1 kcal molꢀ1 higher in energy than the lowest energy synꢀtrans conformer; all four conformers are
expected to be produced via the new synthetic route. The zeroꢀpoint corrected barriers to interconversion
between the conformers are also evaluated and shown in Figure S7. The barriers to interconversion
between the two syn conformers or the two anti conformers are predicted to be relatively low (<10 kcal
13
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 57
1
2
3
4
molꢀ1), while the barrier to interconversion between the syn and anti conformers is quite high (ca. 30 kcal
molꢀ1).
5
6
7
8
9
Scheme 4. Unimolecular decay pathway for IR activated synꢀtransꢀMVKꢀOOa to OH radical products
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
with zeroꢀpoint corrected energies (kcal molꢀ1).
a There is an analogous pathway for synꢀcisꢀMVKꢀOO.
The computed reaction coordinate for the unimolecular decay of synꢀMVKꢀOO conformers to
OH products is shown in Scheme 4 and Figure 2. The reaction proceeds via a 1,4 alkyl Hꢀatom transfer
mechanism through a 5ꢀmembered ring transition state (TS) to produce 2ꢀhydroperoxybutaꢀ1,3ꢀdiene
((CH2=CH)(CH2)COOH, HPBD, (1a)), followed by rapid homolysis of the OꢀO bond to produce OH and
oxybutadiene ((CH2=CH)(CH2)CO, OBD, (2a)) radicals. An expanded theoretical mapping of the HPBD
decay pathway is given in Sec. 4.1. The pathway from the lowest energy synꢀtransꢀMVKꢀOO conformer
(0 kcal molꢀ1) takes place via a TS barrier of 18.0 kcal molꢀ1, while the higher energy synꢀcisꢀMVKꢀOO
conformer (1.8 kcal molꢀ1) is predicted to have a TS barrier at 19.2 kcal molꢀ1. Due to the low barrier to
isomerization between the cis and trans conformers, unimolecular decay of the synꢀcis conformer will
mostly proceed via the lower energy TS with an effective barrier of 16.2 kcal molꢀ1. The incipient HPBD
(1a) product is computed to be 12.0 kcal molꢀ1 lower in energy than synꢀtransꢀMVKꢀOO. The resulting
OBD (2a) and OH radical products are predicted to be 7.8 kcal molꢀ1 higher in energy relative to synꢀ
transꢀMVKꢀOO.
14
ACS Paragon Plus Environment

Page 15 of 57
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Computed reaction coordinate for unimolecular decay of synꢀMVKꢀOO conformers to OH
products. The reaction proceeds via a 1,4 alkyl Hꢀatom transfer mechanism through a 5ꢀmembered
ringꢀlike transition state (TS) to produce 2ꢀhydroperoxyꢀbutaꢀ1,3ꢀdiene ((CH2=CH)(CH2)COOH,
HPBD, 1a), followed by rapid homolysis of the OꢀO bond to produce OH and oxybutadiene
((CH2=CH)(CH2)CO, OBD, 2a) radicals. Stationary points along the reaction for the lowest energy
synꢀtransꢀMVKꢀOO conformer are shown (cyan line). Also shown are stationary points for the less
stable synꢀcisꢀMVKꢀOO conformer (solid purple line) and its higher energy transition state (TS)
(dashed purple line). The energy scale denotes ZPE corrected energies for the stationary points. The
IR pump laser (red arrow) excites synꢀMVKꢀOO in the CH overtone region (2νCH). The OH products
are detected by UV probe laser (black arrow) induced fluorescence (LIF) on the OH AꢀX (1,0)
transition.
3.2. IR Action Spectroscopy of MVKꢀOO
3.2.1. Computed IR Absorption Spectrum
Anharmonic vibrational frequencies and IR intensities in the CH stretch overtone region
computed for the four conformers of MVKꢀOO are shown in Figure 3 and listed in Table S6. MVKꢀOO
exhibits six high frequency normal mode vibrations (2900ꢀ3200 cmꢀ1) attributed to CH stretches. Four of
these stretches involve inꢀplane motions (ν1ꢀν4) and two involve outꢀofꢀplane motions: one inꢀphase (ν5)
and one out of phase (ν21). Fundamental frequencies of these modes are given in Table S7. As a result,
21 IR transitions are expected for each conformer in the 2νCH region (5750ꢀ6300 cmꢀ1): Six CH stretch
overtones (e.g. 2ν1) and 15 CH stretch combinations comprised of one quantum each of two different CH
15
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 57
1
2
3
4
5
6
7
8
stretches (e.g. ν1+ν2), although some of these features are not predicted to have appreciable IR intensity.
It should be noted that the intensities shown in these stick spectra reflect only the intrinsic calculated
intensity of each IR transition and assume equal population of the conformers.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. Computed anharmonic IR stick spectra for the four conformers of MVKꢀOO in the CH
stretch overtone 2νCH region at the B2PLYPꢀD3/ccꢀpVTZ level of theory. (Top panel) synꢀcisꢀMVKꢀ
OO (purple) and synꢀtransꢀMVKꢀOO (cyan). (Bottom panel) antiꢀcisꢀMVKꢀOO (blue) and antiꢀtransꢀ
MVKꢀOO (orange). Strong transitions attributed to the vinyl outꢀofꢀphase CH stretching mode (2ν1)
and outꢀofꢀphase, outꢀofꢀplane, methyl CH stretch (2ν21) are labeled.
The highest energy transition predicted for each of the four MVKꢀOO conformers in the CH
stretch overtone region arises from two quanta of the vinyl outꢀofꢀphase CH stretching mode (2ν1). For
synꢀMVKꢀOO, the 2ν1 transitions of the trans and cis conformers lie within ca. 10 cmꢀ1 of one another
(6200.0 and 6210.4, respectively) and are the most intense transitions in this spectral region (with cis ca.
20% weaker than trans). In contrast, the 2ν1 transitions of the trans and cis conformers of antiꢀMVKꢀOO
are predicted to be separated by ca. 55 cmꢀ1, and are shifted to lower (trans at 6185.8 cmꢀ1) and higher (cis
at 6242.9 cmꢀ1) frequency of the synꢀMVKꢀOO transitions. The 2ν1 transition of the antiꢀtrans conformer
16
ACS Paragon Plus Environment

Page 17 of 57
Journal of the American Chemical Society
1
2
3
4
5
6
is 2ꢀfold stronger than that of antiꢀcis, and antiꢀtrans has an overlapping ν1+ ν2 transition. Similar trends
are predicted for the fundamental ν1 transitions (Table S7).
7
8
9
Other prominent transitions in the computed IR spectra of antiꢀMVKꢀOO arise from two quanta
excitation of the outꢀofꢀphase, outꢀofꢀplane, methyl CH stretch (2ν21) at 5871.3 and 5877.9 cmꢀ1 for the
antiꢀcis and antiꢀtrans conformers, respectively. By contrast, the 2ν21 transitions for the synꢀcis and synꢀ
trans conformers are predicted at significantly lower energy, 5822.3 and 5833.4 cmꢀ1, respectively, and
are much weaker in intensity. Most other strong IR transitions for the four conformers are clustered
between 5900 and 6150 cmꢀ1.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In the syn conformers, the methyl CH stretch overtone (2ν21) transitions are weaker in intensity
and shifted to lower energy due to the intramolecular interaction between methyl group and the terminal
oxygen of the carbonyl oxide group. In the anti conformers, the vinyl CH stretch overtone (2ν1)
transitions are perturbed by the interaction of the vinyl group with the terminal oxygen. These
interactions give rise to the distinctive 2ν21 and 2ν1 transitions (Figure 3) predicted for the syn and anti
conformers, respectively.
In general, the computed IR spectra for the two syn conformers are very similar to one another, as
are the computed IR spectra for the two anti conformers. Due to the high barrier (ca. 30 kcal molꢀ1)
separating the syn and anti conformers (see Figure S7), the syn and anti conformers are chemically
distinct species at the IR excitation energies explored in this work (5750ꢀ6300 cmꢀ1, 16.4ꢀ18.0 kcal molꢀ1).
However, the excitation energies are significantly higher than the predicted barriers separating the cis and
trans conformers (ca. 10 kcal molꢀ1), which may interconvert upon vibrational activation. Thus, in Figure
4, the computed IR transitions for the two syn conformers are combined as are those for the two anti
conformers, and then are convoluted with a Gaussian width of 3 cmꢀ1 (FWHM). This width corresponds
to the typical breadth of a simulated rotational band contour at 10 K as shown in Figure S8. In addition,
the broadened spectra for the syn and anti conformers are summed together in Figure 4 assuming equal
populations of the four conformers.
17
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 18 of 57
1
2
3
4
3.2.2. Experimental IR Action Spectrum
The experimental IR action spectrum of MVKꢀOO is obtained in the 5750ꢀ6260 cmꢀ1 region with
UV LIF detection of the resulting OH radical products at an IRꢀUV time delay of 800 ns, and is shown in
Figure 4. Peak positions and relative intensities are given in Table S9. The IRꢀUV time delay is selected
based on the experimentally determined OH appearance rates (Sec. 3.3.1). The most intense feature in the
spectrum appears at 6200.4 cmꢀ1, and is ascribed, based on its IR frequency and relative intensity, to the
overlapping 2ν1 transitions of synꢀcisꢀ and synꢀtransꢀMVKꢀOO, as indicated by cyan vertical shading in
Figure 4. Other relatively strong, sharp and isolated features in the experimental IR action spectrum agree
remarkably well with predicted transitions for antiꢀMVKꢀOO conformers and are denoted by orange
vertical shading in Figure 4. Specifically, the feature at 5869.6 cmꢀ1 agrees well in IR frequency and
relative intensity with the calculated overlapping 2ν21 transitions of the antiꢀtrans and antiꢀcis conformers.
In addition, the feature at 6174.2 cmꢀ1 agrees with the computed 2ν1 and overlapping ν1+ν2 transitions of
antiꢀtransꢀMVKꢀOO. A much smaller feature at 6234.2 cmꢀ1 is consistent with the predicted 2ν1
transition of antiꢀcisꢀMVKꢀOO.
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
18
ACS Paragon Plus Environment

Page 19 of 57
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. Experimental IR action spectrum of MVKꢀOO (black points) recorded in the CH stretch
overtone 2νCH region with UV laserꢀinduced fluorescence (LIF) detection of OH at an IRꢀUV time
delay of 800 ns. A smoothed fit (gray shaded area) that includes experimental uncertainty (±1.5 σ) is
superimposed on the experimental data. The computed IR absorption spectra for the four conformers
of MVKꢀOO (total, gray), the two syn conformers (syn, cyan) and the two anti conformers (anti,
orange) are shown for comparison. The computed IR spectra are a convolution of the stick spectra
(Figure 3) for the syn and anti conformers with a Gaussian profile (3 cmꢀ1 width), representing the
rotational band contour at ca. 10 K, and summed to give the total. The four conformers are assumed to
be equally populated. The vertical shaded areas represent segments of the experimental spectrum
attributed to the syn conformers (blue) and the anti conformers (orange) of MVKꢀOO in comparison
with theory. Asterisks indicate features at which experimental rate measurements are performed.
Additionally, the experimental IR action spectrum exhibits two relatively strong and quite broad
features centered at ca. 6040 and 5975 cmꢀ1. The feature at 6040 cmꢀ1 is consistent with a combination of
overlapping transitions predicted theoretically for syn and anti conformers, most notably the ν3+ν4 and 2ν3
transitions of antiꢀcisꢀMVKꢀOO, and 2ν2 and 2ν3 transitions of synꢀcisꢀMVKꢀOO. The feature at 5975
cmꢀ1 is broader and stronger than suggested from the sum of the calculated transitions for synꢀ and antiꢀ
MVKꢀOO. The predicted spectrum, however, considers only two quanta vibrational changes and secondꢀ
order couplings. This feature may arise from three or more quanta vibrational changes and/or higher
order couplings between vibrational modes of MVKꢀOO, which will be explored in future work. Several
19
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 20 of 57
1
2
3
4
5
6
7
8
weak features between 5750 and 5850 cmꢀ1 are also observed, where both syn and anti conformers are
predicted to have weak transitions. The overall shape of the experimentally observed IR action spectrum
agrees qualitatively with the total spectrum computed for MVKꢀOO.
9
It is important to note that IR action spectroscopy differs from IR absorption spectroscopy in that
the vibrationally excited molecule must generate OH products. For syn conformers, unimolecular decay
to OH products proceeds via the 1,4 alkyl Hꢀatom transfer pathway predicted for synꢀMVKꢀOO (Figure 2
and Scheme 4) and other synꢀalkyl substituted Criegee intermediates. However, the mere observation of
OH radical products following excitation of spectroscopic features attributed to anti conformers is
surprising. The anti conformers of MVKꢀOO lack a synꢀalkyl αꢀH atom available for facile transfer,
suggesting that OH is formed by a different mechanism. Possible pathways from antiꢀMVKꢀOO to OH
products are considered in the Discussion (Sec. 4.2). We also considered the possibility that other
isomers of MVKꢀOO, e.g. HPBD and dioxole, could contribute to the IR action spectrum. However,
these species have their own distinctive IR transitions (see Figure S9), are more stable (see Table S2) than
MVKꢀOO, and will not undergo unimolecular decay to OH at the experimental excitation energies.
3.3. Unimolecular Decay of MVKꢀOO to OH products
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3.3.1. Experimental Rate Measurements
The experimental rate of appearance of OH products is obtained from direct timeꢀdomain
measurements following excitation of several features in the IR action spectrum ascribed to MVKꢀOO by
comparison with theory. The OH products appear on a hundreds of nanosecond time scale, and fall off on
a greater than 1 ꢁs time scale. The temporal profiles of the OH products are fit to a dual exponential
function as in previous work.34ꢀ36, 53, 63 The exponential rise of OH products is attributed to the
unimolecular decay of MVKꢀOO, while the slower exponential fall arises from purely experimental
factors due to molecules moving out of the spatial region irradiated by the UV probe laser. A
representative temporal profile of the OH products (Figure S10) along with details of the fitting procedure
are given in the SI (Sec. S4).
20
ACS Paragon Plus Environment

Page 21 of 57
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
IR excitation of the strongest feature at 6200.4 cmꢀ1 results in an exponential OH appearance rate
of (2.3 ± 0.2) × 106 sꢀ1 (with ±1σ uncertainty), corresponding to a lifetime of 430 ± 30 ns. This feature is
attributed to the 2ν1 transition of synꢀMVKꢀOO. IR excitation at 6174.2 cmꢀ1 yields a faster exponential
rise rate of (3.2 ± 0.4) × 106 s−1, corresponding to a lifetime of 310 ± 40 ns. This feature is ascribed to the
2ν1 transition of antiꢀtransꢀMVKꢀOO based on theory, but lies on the shoulder of the stronger synꢀMVKꢀ
OO feature at 6200.4 cmꢀ1. (A reliable rate measurement was not feasible for a weaker feature attributed
to antiꢀMVKꢀOO at 5869.6 cmꢀ1.) Finally, IR excitation of the broad and unassigned feature at 5975.4
cmꢀ1 results in a slower OH appearance rate of (1.4 ± 0.3) × 106 sꢀ1, corresponding to a lifetime of 710 ±
150 ns. The experimental OH rise and fall rates are summarized in Table S10; the OH appearance rates
are also plotted in Figure 5.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3.3.2. RRKM Calculations
Statistical RRKM theory is utilized to compute the energyꢀdependent unimolecular decay rates
k(E) for the lowest energy synꢀtrans conformer of MVKꢀOO (see Sec. 3.1). Quantum mechanical
tunneling is included using an asymmetric Eckart model or semiꢀclassical transition state theory
(SCTST), which incorporate anharmonic vibrational frequencies, given in Table S7 for the reactants and
Table S8 for the TS. The computed RRKM rates are predicted to increase uniformly with excitation
energy as shown in Figure 5 and listed in Table S11. The computed unimolecular decay rate for synꢀ
transꢀMVKꢀOO at 6200 cmꢀ1 of 2.79 × 106 sꢀ1 (Eckart) or 3.08 × 106 sꢀ1 (SCTST) agrees well with the
experimentally measured rate for appearance of OH radical products following IR excitation of synꢀ
MVKꢀOO (2ν1) of (2.3 ± 0.2) × 106 sꢀ1. The 6200 cmꢀ1 excitation energy lies slightly below the computed
synꢀtrans TS barrier (6288.7 cmꢀ1) demonstrating the importance of tunneling in the generation of OH
products. The agreement between experiment and theory validates the 1,4 Hꢀatom transfer mechanism
and TS barrier for unimolecular decay of synꢀtransꢀMVKꢀOO to OH products. In addition, the agreement
demonstrates that the synꢀtrans conformer of MVKꢀOO is the principal spectral carrier near the peak of
21
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 22 of 57
1
2
3
4
5
the strong 2ν1 feature in the IR action spectrum at 6200.4 cmꢀ1 with minimal contribution from the 2ν1
transition of the synꢀcis conformer, theoretically predicted at 6210 cmꢀ1, as explained below.
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. Experimentally observed rates and corresponding lifetimes (semiꢀlog scales) for
unimolecular decay of IRꢀactivated MVKꢀOO to OH products (gray points) with ±1σ uncertainty from
repeated measurements (Table S10). Statistical RRKM rates for synꢀtransꢀMVKꢀOO (cyan) are shown
for comparison using asymmetric Eckart (solid cyan line) and SCTST (dashed cyan line) tunneling
models.
Due to the small barrier to isomerization between the cis and trans conformers of synꢀMVKꢀOO,
the two conformers are assumed to freely interconvert at energies near the TS barrier. As a result, the
higher energy synꢀcis conformer of MVKꢀOO (1.8 kcal molꢀ1) will mostly proceed via the lower energy
TS with an effective barrier of 16.2 kcal molꢀ1. Furthermore, 2ν1 excitation of the higher energy synꢀcis
conformer will access a higher vibrational state of synꢀMVKꢀOO. These factors lead to the prediction of
a ca. 4x faster RRKM rate for synꢀcisꢀMVKꢀOO [9.99 × 106 sꢀ1 (Eckart) or 1.06 × 108 sꢀ1 (SCTST) as
shown in Figure S11 (Table S11)] than computed for synꢀtransꢀMVKꢀOO at 6200 cmꢀ1. Indeed, the
predicted RRKM rates for synꢀcisꢀMVKꢀOO are much faster than any of the experimental rates (Table
S10). This suggests that the higher energy synꢀcis conformer makes minimal contribution at 6200 cmꢀ1
due to less abundance of this conformer or the 2ν1 transition of synꢀcisꢀMVKꢀOO is shifted to higher
wavenumber as predicted theoretically.
22
ACS Paragon Plus Environment

Page 23 of 57
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
The experimental OH appearance rate at 6174.2 cmꢀ1 is more rapid than that observed for the
higher energy feature at 6200.4 cmꢀ1 assigned to synꢀtransꢀMVKꢀOO. This is not consistent with the
general RRKM trend of slower rates at lower energies, e.g. as shown for synꢀtransꢀMVKꢀOO (Figure 5).
As discussed in Sec. 3.2.2, the position of the observed feature at 6174.2 cmꢀ1 is consistent with the 2ν1
transition predicted for antiꢀtransꢀMVKꢀOO. Nevertheless, the experimental rate at 6174.2 cmꢀ1, (3.2 ±
0.4) × 106 sꢀ1, does not differ sufficiently from the computed RRKM rate for synꢀtransꢀMVKꢀOO at this
energy [2.63 × 106 sꢀ1 (Eckart) or 2.92 × 106 sꢀ1 (SCTST)] to exclude the possibility of contribution to the
rate from partial overlap of the adjacent synꢀtransꢀMVKꢀOO feature at 6200.4 cmꢀ1. Despite the
similarity in rate, the position of this and other spectroscopic features in Figure 4 (orange shading) are
consistent with IR transitions predicted theoretically for antiꢀMVKꢀOO. In Sec. 4.2, we consider
plausible pathways from anti conformers of MVKꢀOO to OH products.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The computed unimolecular decay rate for synꢀtransꢀMVKꢀOO at 5975 cmꢀ1 of 1.66 × 106 sꢀ1
(Eckart) or 1.87 × 106 sꢀ1 (SCTST) is also in good agreement with the experimentally measured OH
appearance rate of (1.4 ± 0.3) × 106 sꢀ1 for IRꢀactivated MVKꢀOO at this energy. This suggests that there
may be a contribution from synꢀtransꢀMVKꢀOO to the broad feature in the IR action spectrum at 5975
cmꢀ1, although contributions from other conformers cannot be ruled out.
The energyꢀdependent unimolecular decay rates for MVKꢀOO to OH products may alter the
intensities observed in the IR action spectrum at a fixed IRꢀUV time delay. For synꢀtransꢀMVKꢀOO over
the energy range (6260 to 5750 cmꢀ1) of the IR action spectrum, the OH appearance time will increase
(280 ns to 1000 ns) toward lower energy. This will decrease the intensity observed at a fixed delay of 800
ns by a factor of 2 toward the lower energy extreme as shown in Figure S12, making it more difficult to
observe the weaker transitions predicted at lower energy. Nevertheless, we expect to be able to detect
vibrationally activated MVKꢀOO that decays to OH products on a microsecond or faster timescale (see SI
Sec. S4).
23
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 24 of 57
1
2
3
4
4. Discussion
5
6
7
8
Qualitative agreement between the overall shape of the experimental IR action spectrum and
calculated IR absorption spectrum (Figure 4) demonstrate that MVKꢀOO is the spectral carrier and
indicate the formation of OH radicals from MVKꢀOO. The OH production mechanism from synꢀMVKꢀ
OO is described in the Results (Sec. 3), illustrated in Scheme 4, and discussed further in Sec. 4.1.
Plausible mechanisms for OH production from antiꢀMVKꢀOO are presented in Sec. 4.2, but are less
certain and will require further investigation. The chemical names, chemical formulas, abbreviations, and
labels used in the schemes and text are summarized in Table S13.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
4.1 synꢀMVKꢀOO
It is well accepted that OH production from Criegee intermediates with synꢀalkyl substituents
proceeds via 1,4 alkyl Hꢀatom migration from the synꢀalkyl group to the terminal O atom.23, 37ꢀ39 Synꢀ
MVKꢀOO is predicted to follow this same mechanism (Figure 2 and Scheme 4). The highꢀlevel
calculation of the TS barrier in this work agrees within 1 kcal molꢀ1 of the previously predicted barriers
for this process (see Table 1). In addition, the computed RRKM rate with tunneling for unimolecular
decay of synꢀtransꢀMVKꢀOO is in quantitative agreement with the experimentally observed rate of OH
production near the peak of the strong feature at 6200.4 cmꢀ1. This confirms that the IR spectral feature
arises primarily from synꢀtransꢀMVKꢀOO and validates the TS barrier for the 1,4 alkyl Hꢀatom transfer
mechanism.
The unimolecular decay of synꢀtransꢀMVKꢀOO (430 ± 30 ns) is significantly slower than
previously observed for synꢀCH3CHOO (6.7 ± 0.2 ns at 5709 cmꢀ1), (CH3)2COO (18.5 ± 1.0 ns at 5971
cmꢀ1) and synꢀCH3CH2CHOO (39.2 ± 2.6 ns at 5961 cmꢀ1) upon 2νCH excitation.36, 53 The slower
unimolecular decay for synꢀtransꢀMVKꢀOO is primarily due to its higher TS barrier (18.0 kcal molꢀ1)
compared to synꢀCH3CHOO (17.05 kcal molꢀ1), (CH3)2COO (16.16 kcal molꢀ1) and synꢀCH3CH2CHOO
(16.46 kcal molꢀ1),36, 53 and larger density of states in the 2νCH region. For synꢀtransꢀMVKꢀOO and
24
ACS Paragon Plus Environment

Page 25 of 57
Journal of the American Chemical Society
1
2
3
4
previously studied synꢀalkyl substituted Criegee intermediates, Hꢀatom tunneling significantly enhances
the unimolecular decay rate and allows for OH production at energies much below the TS barrier.34ꢀ36, 53, 63
5
6
7
8
9
Table 1. Comparison of ZPE corrected stationary point energies (kcal molꢀ1).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
CCSD(T)//
B3LYPa
CBS-
CCSD(T)// Present
Species
Labels
MCG3c
0
QB3b
M06-2Xd
Work
Minima
syn-trans-MVK-OO
syn-cis-MVK-OO
anti-trans-MVK-OO
anti-cis-MVK-OO
trans-HPBD
cis-HPBD
dioxole
trans-dioxirane
cis-dioxirane
0
1.5
2.5
2.7
-14.9
-13.7
0
1.7
2.9
2.9
-14.7
-12.9
-25.6
-17.1
-18.6
-23.1
0
0
1.76
2.57
3.05
-13.83
-12.02
-24.28
-15.28
-16.77
-18.1
1b
1a
11
6a
6b
12
-17.0
-18.2
diradical
Transition States
syn-trans-MVK-OO to cis-HPBD
syn-cis-MVK-OO to trans-HPBD
anti-trans-MVK-OO to AHP
anti-cis-MVK-OO to dioxole
anti-cis-MVK-OO to cis-dioxirane
cis-dioxirane to syn-cis-MVK-OO
anti-trans-MVK-OO to trans-
dioxirane
18.8
19.9
24.4
18.0
19.1
21.7f
13.9
24.9
20.9
18.34
18.64e
17.5
17.96
19.19
24.57
15.06
25.38
21.35
23.6e
13.9e
24.9
21.0
21.0
20.0
22.1
19.7
19.8e
19.93
trans-dioxirane to syn-trans-MVK-OO
syn-trans to anti-trans MVK-OO
syn-cis to anti-cis MVK-OO
dioxole to diradical
diradical to β-dicarbonyl
22.2
26.0
23.5
-4.9
22.44
30.93
30.63
-2.5
-19.5
-18.7
a Ref. 6.
b Ref. 16.
c Ref. 44.
d Ref. 21. Note that this work does not distinguish between cis and trans conformers of MVK-OO; only
syn and anti conformers are identified.
e The energies of TS barriers in Ref. 21 are added to the ground state energies of the relevant reactant
reported in this work. We therefore add the TS barrier reported in Ref. 21 to the energy for syn-cis-
MVK-OO reported herein.
f We repeated the calculation for anti-trans-MVK-OO to AHP using the CBS-QB3 method and obtained a
barrier of 24.2 kcal mol-1, which is in much better accord with the present work.
25
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 26 of 57
1
2
3
4
5
6
Scheme 5. Unimolecular decay pathways for 2ꢀhydroperoxybutꢀ1,3ꢀdiene (HPBD) with zeroꢀpoint
corrected energies (kcal molꢀ1) relative to synꢀtransꢀMVKꢀOO (black/blue) and OH + oxybutadiene
(OBD) asymptote (red).
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Thus far, we have considered the rateꢀlimiting 1,4 Hꢀatom transfer pathway from synꢀMVKꢀOO
to HPBD (cisꢀHPBD, 1a), followed by a simple barrierless OꢀO bond dissociation to form OH + OBD
(cisꢀOBD, 2a) products (see Scheme 4). However, further theoretical investigation using multireference
methods reveals that dissociation of HPBD (1a) is not a simple, barrierless bond cleavage reaction, but
rather the potential energy surface for this process is fairly complex and may lead to additional products
as shown in Scheme 5. Note that in Scheme 5 and the discussion that follows in Sec. 4.1, we provide
energies relative to the separated OH + OBD (2b) products. This scheme is analogous to that recently
suggested by Kuwata et al. for thermal decay of synꢀCH3CHOO and (CH3)2COO.76 For simplicity, we
display the reaction pathway from the lowest energy conformer of HPBD (transꢀHPBD, 1b). [The cisꢀ
HPBD (1a) conformer is also accessible and will readily isomerize to the transꢀHPBD (1b).]
The unimolecular reaction pathway of HPBD (1b) begins with OꢀO bond elongation, passes over
a small submerged saddle point (SP) at ꢀ3.2 kcal molꢀ1, and leads to intermediate well RC(1) where OH +
OBD (2b) are weakly interacting in the longꢀrange region of the potential. Further elongation of the OꢀO
bond continues smoothly (no SP) to OH + OBD (2b) products. Within the shallow well RC(1) (ꢀ4.8 kcal
mol−1), the Hꢀside of the OH radical is in a strong hydrogen bonding interaction with the Oꢀsite of OBD.
26
ACS Paragon Plus Environment

Page 27 of 57
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
These features along the minimum energy dissociation coordinate of HPBD are consistent with previous
theoretical studies of the reaction coordinate for synꢀCH3CHOO to OH radical products.77ꢀ78
Alternatively, a roamingꢀinduced unimolecular decay process may occur that involves the
9
isomerization of the complex (RC(1)). In particular, the OH radical may roam in a relatively flat
potential region from one side of the resonance stabilized OBD (2b) moiety to the other side to form
another distinct weakly interacting hydrogen bonded complex (RC(2)), in which the Oꢀside of OH radical
is pointing toward the CH2 group of the OBD (2b) fragment. Notably, the roaming SP (ꢀ4.4 kcal molꢀ1)
for this OH reorientation process (RC(1) ↔ RC(2)) lies considerably below the asymptotic energy of OH
+ OBD (2b). The second complex (RC(2)) may also undergo a barrierless (no SP) dissociation to OH +
OBD (2b) products. An alternative roamingꢀinduced isomerization reaction may occur from RC(2) by
addition of the OH fragment to the CH2 group of the OBD fragment to yield 1ꢀhydroxybutꢀ3ꢀenꢀ2ꢀone
(HB, 3) in a deep well (ꢀ89.2 kcal molꢀ1). Under the collision free conditions of the present experiments,
HB (3) formed via roaming would likely dissociate via CC bond fission to different radical products (4 +
5) that lie lower in energy (ca. 8.3 kcal molꢀ1) than OO fission to OH + OBD (2b) products. Although not
explored here, the pathway from HB (3) to products (4 + 5) may include a sequence of submerged SP
(denoted ‘barrierless’ in Scheme 5). Under thermalized conditions, the roaming path may result
predominantly in collisional stabilization of HB (3) in its deep well. Such a roaming induced
isomerization pathway provides an explanation for the observation of hydroxy acetone in recent
experiments involving thermalized (CH3)2COO.76, 79 Thus, the roaming pathway from HPBD (1b) to HB
(3) could potentially reduce the OH yield from IRꢀactivated synꢀMVKꢀOO.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The MEP for the roaming mechanism from HPBD (1b) to HB (3) is shown in Figure 6 using
coordinates that reflect the distance and the angular orientation of OH relative to OBD (2b). The saddle
points (SP) along the roaming pathway lie at least 4 kcal molꢀ1 lower in energy than the OH + OBD (2b)
product asymptote. As a result, the roaming dynamics (i.e. the reorientation of OH between RC(1) and
RC(2)) and the roaming mediated isomerization from RC(2) to HB (3) become accessible at a lower
energy than OO bond fission. Thus, one might expect roaming to be a significant component of the
27
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 28 of 57
1
2
3
4
5
6
dynamics under atmospheric conditions, which could alter the overall OH yield and the products formed
from atmospheric isoprene ozonolysis.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. Minimum energy path for the roaming mechanism from transꢀ2ꢀhydroperoxybutaꢀ1,3ꢀdiene
(transꢀHPBD, (1b)) to hydroxybutenone (HB, (13)) is displayed as a function of two coordinates, CO
bond distance (RCO) and OCO angle (θOCO), as shown on the inset structure. The calculations are
performed using the CASPT2(6e,6o)/ccꢀpVTZ method/basis. The color gradient represents the
electronic energy (kcal molꢀ1) with zero taken to be the separated OH and transꢀoxybutadiene (transꢀ
OBD, (2b)) products. Points include roaming complexes RC(1) and RC(2) and saddle points (SP)
along the pathway from HPBD to HB.
4.2.
antiꢀMVKꢀOO
Some features observed in the experimental IR action spectrum in the CH stretch overtone (2νCH)
region are consistent with theoretically predicted IR transitions of antiꢀMVKꢀOO. In this section, we
explore several unimolecular reaction routes theoretically to ascertain their viability as a source of OH
from anti conformers of MVKꢀOO. In general, energies shown in schemes are relative to the lowest
energy synꢀtrans conformer of MVKꢀOO. Note that the antiꢀtrans and antiꢀcis conformers lie 2.6 and 3.1
kcal molꢀ1 higher in energy, respectively, resulting in lower effective barriers for unimolecular reactions of
anti conformers. An IR excitation energy of 5750ꢀ6260 cmꢀ1 (16.4ꢀ17.9 kcal molꢀ1) is utilized in the
present experiments, which is much greater than the isomerization barrier between the two antiꢀMVKꢀOO
28
ACS Paragon Plus Environment

Page 29 of 57
Journal of the American Chemical Society
1
2
3
4
5
conformers (ca. 8 kcal molꢀ1). As a result, the same reaction pathways should be relevant for both antiꢀcis
and antiꢀtrans conformers.
6
7
8
4.2.1. anti to syn conformational isomerization
9
Scheme 6. Two different isomerization pathways from antiꢀMVKꢀOO to synꢀMVKꢀOO with zeroꢀ
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
point corrected energies (kcal molꢀ1).
Two different anti to syn isomerization pathways for MVKꢀOO are shown in Scheme 6. The first
involves internal rotation about the C=O bond, as predicted previously.16, 21 We have explored this
isomerization pathway using a multireference description due to the diradical configurations of the
transition state. The lowest barrier to internal rotation is predicted to be 30.6 kcal molꢀ1 between the antiꢀ
cis and synꢀcis conformers. Our calculated barriers are ca. 5ꢀ9 kcal molꢀ1 higher than those predicted by
Kuwata et al.16 Most importantly, the predicted barriers are >10 kcal mol−1 greater than the excitation
energies used in this experiment, indicating that this isomerization process cannot be a feasible pathway
to OH radical products. We have also considered a second multistep anti to syn isomerization pathway in
our calculations: ring closure to a dioxirane (6a) intermediate, followed by ring opening to synꢀMVKꢀ
OO, as shown by the red arrows (antiꢀtransꢀMVKꢀOO to 6a to 6b to synꢀcisꢀMVKꢀOO) within the
dashed box in Scheme 6. The minimum energy route along this isomerization pathway involves
surmounting a relatively high barrier of 21.4 kcal molꢀ1 for the isomerization from 6b to synꢀcisꢀMVKꢀ
OO. At the excitation energies in this study, this rateꢀlimiting step is expected to be very slow (ca. 10ꢀ1
29
ACS Paragon Plus Environment