Photochemistry of 2-Formylphenylnitrene: A Doorway to Heavy-Atom Tunneling of a Benzazirine to a Cyclic Ketenimine

Article abstract of DOI:10.1021/jacs.7b10495

The slippery potential energy surface of aryl nitrenes has revealed unexpected and fascinating reactions. To explore such a challenging surface, one powerful approach is to use a combination of a cryogenic matrix environment and a tunable narrowband radiation source. In this way, we discovered the heavy-atom tunneling reaction involving spontaneous ring expansion of a fused-ring benzazirine into a seven-membered ring cyclic ketenimine. The benzazirine was generated in situ by the photochemistry of protium and deuterated triplet 2-formylphenylnitrene isolated in an argon matrix. The ring-expansion reaction takes place at 10 K with a rate constant of ～7.4 × 10^-7 s^-1, despite an estimated activation barrier of 7.5 kcal mol^-1. Moreover, it shows only a marginal increase in the rate upon increase of the absolute temperature by a factor of 2. Computed rate constants with and without tunneling confirm that the reaction can only occur by a tunneling process from the ground state at cryogenic conditions. It was also found that the ring-expansion reaction rate is more than 1 order of magnitude faster when the sample is exposed to broadband IR radiation.

Full text of DOI:10.1021/jacs.7b10495

Subscriber access provided by READING UNIV
Article
Photochemistry of 2-Formyl-Phenylnitrene: A Doorway to
Heavy-Atom Tunneling of a Benzazirine to a Cyclic Ketenimine
Cláudio M. Nunes, Igor Reva, Sebastian Kozuch, Robert J. McMahon, and Rui Fausto
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.7b10495 • Publication Date (Web): 07 Nov 2017
Downloaded from http://pubs.acs.org on November 7, 2017
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 free 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 accessible to all
readers and 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.
Journal of the American Chemical Society 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 29
Journal of the American Chemical Society
INTRODUCTION
1
2
3
4
5
6
7
8
9
Aryl nitrenes (Ar−N) are highly reactive intermediates most commonly generated from aryl azides
(Ar−N3).1–3 To shed light on the aryl azides/nitrenes chemistry and on the nature of aryl nitrene species,
many studies have been carried out in last fifty years, in particular with the application of matrix
isolation, timeꢀresolved spectroscopy and quantumꢀmechanical computational approaches.1–7 This
knowledge has played a crucial role for the design of aryl azides that render aryl nitrenes with suitable
reactivity for uses in photoresists, in photoaffinity labeling of biomolecules, and in the design of
materials or organic molecules.8–11
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
It is now well established that the photolysis of aryl azides occurs with the release of molecular
nitrogen and the formation of openꢀshell singlet (OSS) aryl nitrene intermediates (Scheme 1). The OSS
aryl nitrenes are ~15 kcal mol−1 more energetic than their triplet ground state.6,12,13 Therefore, upon
formation of OSS aryl nitrenes, either relaxation by intersystem crossing (ISC) to triplet aryl nitrenes or
competitive rearrange to benzazirines takes place, or both, depending on the experimental conditions
and on the characteristics of the aryl substituents. With a few exceptions, the formation of benzazirines
is the rateꢀlimiting step of the process of aryl nitrene isomerization to cyclic ketenimines. The barrier of
ring expansion from benzazirines to cyclic ketenimines is quite small (~3 kcal mol−1 computed for
parent species),6,14 which makes very challenging and rare the capture and direct observation of
benzazirines.15–18
Scheme 1. General mechanism for the photochemistry of aryl azides.
2
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 29
The slippery potential energy surface of aryl nitrenes hides, however, other fascinating reactions.
Some of us have recently found the first indication of heavyꢀatom tunneling for the ring expansion of a
benzazirine to a cyclic ketenimine.15 It was observed that methylthioꢀsubstituted benzazirine, generated
in the photochemistry of 4ꢀmethylthioꢀphenylazide, spontaneously rearranges in cryogenic matrices to
the corresponding cyclic ketenimine. In contrast, under similar conditions, methoxyꢀsubstituted
benzazirine, generated in the photochemistry of 4ꢀmethoxyꢀphenylazide, was found to be stable.
Furthermore, we have also recently revealed the first direct evidence of a tunneling reaction in the aryl
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
3
nitrene chemistry.19 It was found that triplet 2ꢀformylꢀphenylnitrene 2ꢀh, generated from 2ꢀformylꢀ
phenylazide 1ꢀh, spontaneously rearranges into iminoꢀketene 3ꢀh in the dark at 10 K (Scheme 2). Under
3
the same conditions, the deuterium derivative 2ꢀd was observed to be stable, strongly supporting the
occurrence of a tunneling mechanism as opposed to a thermally activated process.
H
N3
H
N
H
N
dark, 10 K
O
h
O
O
Ng, 10 K
-N2
[1,4]-H shift
tunneling
1
-h
32
-h
3
-h
D
N3
D
dark, 10 K
no reaction
N
D
N
O
h
O
O
Ng, 10 K
-N2
1-d
32-d
3-d
3
Scheme 2. Tunneling reaction of triplet 2ꢀformylꢀphenylnitrene 2ꢀh to iminoꢀketene 3ꢀh directly
observed in (Ng = noble gas) cryogenic matrices.19
Quantum mechanical tunneling is an intriguing phenomenon that is becoming recognized as crucial to
understand the reactivity of some organic reactions and even some biological processes that involve
displacement of a light hydrogen atom.20–23 Because the quantum tunneling probability decreases
exponentially with the square root of the shifting mass, the observation of heavyꢀatom tunneling
(involving the displacement of nonꢀhydrogen atoms) is very rare.24 There are six reactions known where
experimental evidence was obtained for heavyꢀatom tunneling occurring under cryogenic conditions
3
ACS Paragon Plus Environment

Page 3 of 29
Journal of the American Chemical Society
from the ground vibrational state:25 the automerization of cyclobutadiene,26 the ring expansion reactions
methylcyclobutylfluorocarbene,27 noradamantylchlorocarbene,28
1
2
3
of
strained
cyclopropane
4
5
6
7
8
9
1Hꢀbicyclo[3.1.0]ꢀgexaꢀ3,5ꢀdienꢀ2ꢀone29 and methylthio substituted benzazirine,15 and the Cope
rearrangement of dimethylsemibullvalene30. Nonetheless, computational chemists have been
discovering and predicting that carbon atom tunneling would contribute to many more reactions.24,31
Herein, we report new results obtained in the quest for elusive reactions on potential energy surface of
aryl nitrenes. Using lowꢀtemperature matrix isolation coupled with IR spectroscopy and a tunable
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
narrowband radiation source, the photochemistry of protium 2ꢀh and deuterated 2ꢀd triplet 2ꢀformylꢀ
phenylnitrene was investigated and one interesting case of heavyꢀatom tunneling consisting in the ring
expansion reaction of the corresponding benzazirine to cyclic ketenimine was discovered. Computations
performed using small curvature tunneling approximation and canonical variational transition state
theory confirm that the reaction can only occur by a tunneling process from the ground state at
cryogenic conditions.
4
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 29
RESULTS AND DISCUSSION
Photochemistry of 2ꢀformylꢀphenylnitrene
1
2
3
4
5
6
7
8
Few minutes irradiation at λ = 308 nm of deuterated 2ꢀformylꢀphenylazide 1ꢀd, isolated in an argon
matrix at 10 K, gives mainly triplet deuterated 2ꢀformylꢀphenylnitrene 32ꢀd.32 As shown in Figure 1, the
9
3
experimental IR bands of the photoproduct are in excellent agreement with the IR spectrum of 2ꢀd
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
calculated at the B3LYP/6ꢀ311++G(d,p) level. The comprehensive assignment of the experimental IR
3
spectrum of triplet deuterated 2ꢀformylꢀphenylnitrene 2ꢀd is given in Table S1. The irradiation, under
similar conditions, of protium 2ꢀformylꢀphenylazide 1ꢀh produces mainly triplet 2ꢀformylꢀphenylnitrene
32ꢀh, which spontaneously rearranges by tunneling to iminoꢀketene 3ꢀh.19 In contrast to the behavior of
nitrene 32ꢀh, deuterated nitrene 32ꢀd is stable against tunneling in argon matrix at 10 K.
60
ꢀ
Simulated (b)
ꢀ
ꢀ
30
ꢀ
ꢀ
ꢀ
ov
ov
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
0
0.12
ꢀ
λ
= 308 nm (a)
ꢀ
ꢀ
ꢀ
0.06
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
0.00
1800
1600
1400
1200
1000
800
600
Wavenumber / cm−1
Figure 1. (a) Experimental difference IR spectrum obtained as the spectrum after UV irradiation at
λ = 308 nm (240 s, 3 mW) of deuterated 2ꢀformylꢀphenylazide 1ꢀd isolated in an argon matrix at 10 K
“minus” the spectrum of 1ꢀd before irradiation. The negative bands (truncated) are due to the consumed
1ꢀd (at this stage ∼60 %). The positive bands labeled with black circles (●) are assigned to triplet
3
3
deuterated 2ꢀformylꢀphenylnitrene 2ꢀd. (b) IR spectrum of 2ꢀd simulated at B3LYP/6ꢀ311++G(d,p)
3
level. The labels “ov” indicate IR bands of 2ꢀd overlapped with more intense bands due to the 1ꢀd
precursor.
3
Subsequently, the photochemistry of 2ꢀd was investigated by performing irradiations using longer
wavelengths, where the remained precursor 1ꢀd does not absorb. Starting with irradiations at λ = 600 nm
5
ACS Paragon Plus Environment

Page 5 of 29
Journal of the American Chemical Society
and gradually decreasing the wavelength, it was found that nitrene 32ꢀd starts to react at around λ = 530
nm.33 Figure 2 shows the results of irradiation at λ = 530 nm during a total time of 40 min, when the
1
2
3
4
5
6
3
3
consumption of 2ꢀd is ∼70 %. Concomitantly with consumption of 2ꢀd, at least three different
7
8
9
photoproducts labeled as A (2112 cm−1), B (1839 cm−1) and C (1695/1688 cm−1) were identified. The
amount of photoproduct A was found to reach its maximum at the current stage of irradiation (t = 40
min). The continuation of irradiation (λ = 530 nm) led to a decrease in the total amount of A and a
significant increase in the total amount of B (Figure S1). This kinetic trend strongly suggests that B is
not directly formed from the nitrene 32ꢀd.
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
0.07
λ
= 530 nm (b)
A
C
B
0.00
ꢀ
ꢀ
ꢀ
ov
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
-0.07
60
ꢀ
Simulated (a)
ꢀ
30
0
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
2200
2000
1800
1600
1400
1200
Wavenumber / cm−1
Figure 2. (a) IR spectrum of triplet deuterated 2ꢀformylꢀphenylnitrene 32ꢀd simulated at
B3LYP/6ꢀ311++G(d,p) level. (b) Experimental difference IR spectrum showing changes after
irradiation at λ = 530 nm (40 min, 90 mW), subsequent to irradiation at λ = 308 nm (see Figure 1). The
negative bands are due to the consumed nitrene 32ꢀd (at this stage ∼70 %). The positive bands labeled A,
B and C are the most characteristic bands of the identified photoproducts. The label “ov” indicates an IR
band of 32ꢀd overlapped with a more intense band due to the photoproduct A.
Indeed, after the irradiation at λ = 530 nm during a total time of 160 min (when almost all nitrene 32ꢀd
is consumed; Figure S1), it was found that subsequent irradiations at λ = 500 nm convert A into B
(Figure 3b).34 The experimental IR signature of the consumed A compares very well with the IR
6
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 29
spectrum of Nꢀdeuterated iminoꢀketene 3ꢀd calculated at B3LYP/6ꢀ311++G(d,p) level (Figure 3aꢀb).
The most intense band observed in the IR spectrum of A appears at 2112 cm−1, a frequency region
where ketenes usually have characteristic strong ν(C=C=O)as absorption (e.g. typically around
2100 cm−1).35 Actually, this band compares well with the ν(C=C=O)as band described at 2110 cm−1 for
1
2
3
4
5
6
7
8
9
3
the protium iminoꢀketene 3ꢀh analogue, which is formed by 2ꢀformylꢀphenylnitrene 2ꢀh tunneling.19
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
Other characteristic IR bands of A are observed at 1637, 1547, 1540 and 621/616 cm−1, in good
correspondence with the IR bands of 3ꢀd calculated at 1645 [ν(C=C)as], 1555 [ν(C=N)], 1541 [ν(C=C)s]
and 607 [τ(N–D)] cm−1. A more comprehensive assignment of the observed IR spectrum of product A,
identified as Nꢀdeuterated iminoꢀketene 3ꢀd, is given in Table S2.
40
20
320
160
ꢂ
Simulated (c)
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
0
0
0.17
0.025
λ
= 500 nm (b)
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
B
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
0.00
0.000
*
ꢁ
ꢁꢁ
*
*
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ A
-0.17
400
-0.025
50
ꢁ
ꢁ
Simulated (a)
ꢁ
200
0
25
0
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
2200
2000
1800
1600
1400
1200
1000
800
600
Wavenumber / cm−1
Figure 3. (a) IR spectrum of Nꢀdeuterated iminoꢀketene 3ꢀd simulated at B3LYP/6ꢀ311++G(d,p) level.
(b) Experimental difference IR spectrum showing changes after irradiation at λ = 500 nm (100 min,
3
60 mW), subsequent to irradiation at λ = 530 nm that consumed nitrene 2ꢀd (Figure S1). The negative
bands are due to the consumed A assigned to iminoꢀketene 3ꢀd (solid squares, ꢁ). The positive bands
are due to the formation of B assigned to benzoazetinone 4ꢀd (empty squares, ꢂ). Asterisks (*) indicate
3
the bands due to traces of nitrene 2ꢀd consumed. (c) IR spectrum of Nꢀdeuterated benzoazetinone 4ꢀd
simulated at B3LYP/6ꢀ311++G(d,p) level.
7
ACS Paragon Plus Environment

Page 7 of 29
Journal of the American Chemical Society
The experimental IR signature of the photoproduct B was found to compare very well with the IR
1
2
spectrum of Nꢀdeuterated benzoazetinone 4ꢀd calculated at B3LYP/6ꢀ311++G(d,p) level (Figure 3bꢀc).
The most intense band observed in the IR spectrum of B at 1836 cm−1 correlates well with the most
intense band predicted in the IR spectrum of 4ꢀd at 1852 [ν(C=O)] cm−1. This assignment is also
supported by the frequency proximity between the band at 1836 cm−1 and the strong carbonyl stretching
at 1843 cm−1 described for the Nꢀmethyl benzoazetinone analogue,36 which are typical of the ν(C=O)
frequencies where the carbonyl group is inserted in a 4ꢀmembered ring.37 Other characteristic IR bands
of B are observed at 2543/2533 and 1605 cm−1, fitting well with the IR bands of 4ꢀd calculated at 2535
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
[ν(N−D)] and 1608 [ring stretching] cm−1. A more complete assignment (total of 19 bands) of the
observed IR spectrum of B, identified as Nꢀdeuterated benzoazetinone 4ꢀd, is given in Table S3.
The photochemical transformation of an iminoꢀketene into a benzoazetinone is not unprecedented.
Dunkin et al.36 and Tomioka et al.38 reported that analogous Nꢀmethyl and Nꢀmethoxy iminoꢀketenes
undergo ring closure to the corresponding Nꢀsubstituted benzoazetinones in argon matrix, upon
irradiation with λ > 400 nm. The identification of those iminoꢀketenes and benzoazetinones was
performed mainly by their characteristic ν(C=C=O)as and ν(C=O) bands, respectively (Scheme S2). In
the current work, the transformation of iminoꢀketene 3ꢀd into benzoazetinone 4ꢀd was established much
more comprehensively, based on the experimental identification of the almost complete midꢀIR spectra
of these species.
Ring expansion of benzazirine into cyclic ketenimine
Interestingly, besides the photochemical conversion of 3ꢀd (A) into 4ꢀd (B), we found that C
spontaneously rearranges into a new product labeled as D (1890/1880 cm−1). To observe this
3
transformation in detail, a new experiment was performed where first the nitrene 2ꢀd was transformed
by irradiation at λ = 530 nm and then the sample was kept at 10 K in the spectrometer beam for 3 days.
After that time, the rearrangement of C into D was practically completed and the change in the
experimental IR spectrum were analyzed (Figure 4b). The IR spectrum of the consumed C is readily
8
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 29
assigned to deuterated formyl benzazirine 5ꢀd based on the good agreement between the experimental
and the calculated IR spectra (Figure 4a,b). The cyclization of nitrene 2ꢀd may proceed “towards” and
“away” from the formyl substituent giving two isomeric benzazirines; each having two possible
conformers regarding orientation of the formyl group (Table S4). Vibrational calculations on these four
structures revealed that only 5ꢀd (the most stable form) is compatible with the experimental IR spectrum
of C (Figure S2 and S3). The most characteristic IR bands of C are observed at ~2100, 1751, ~1700 and
1241 cm−1 in good agreement with the estimated IR bands of 5ꢀd at 2109 [ν(OC–D)], 1802 [ν(C=N)],
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
1710 [ν(C=O)] and 1221 [ν(ODC–C)] cm−1. A slight overestimation of the B3LYP calculated ν(C=N)
frequency of 5ꢀd (~50 cm−1) is consistent with the overestimation trend observed for other
benzazirines16 (see also data presented in Table S5). A more detailed assignment of the IR spectrum of
C, identified as benzazirine 5ꢀd, is given in Table 1.
80
ꢁ
Simulated (c)
ꢁ
40
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
0
0.03
Decay, 10 K (b)
ꢁ
ꢁ
D
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
H O
2
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
0.00
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
C
ꢂ
ꢂ
ꢂ
-0.03
80
ꢀ
Simulated (a)
40
0
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
1900
1700
1500
1300
1100
1
900
700
Wavenumber / cm−
Figure 4. (a) IR spectrum of deuterated formyl benzazirine 5ꢀd simulated at B3LYP/6ꢀ311++G(d,p)
level. (b) Experimental difference IR spectrum showing changes occurred after keeping the sample at
10 K in the spectrometer beam for ~3 days, subsequent to irradiation at λ = 530 nm that consumed
almost all triplet nitrene 32ꢀd. The negative bands are due to the consumed C assigned to benzazirine 5ꢀd
(solid triangles, ꢀ). The positive bands are due to the formation of D assigned to cyclic ketenimine 6ꢀd
(empty triangles, ꢁ). (c) IR spectrum of deuterated formyl cyclic ketenimine 6ꢀd simulated at
B3LYP/6ꢀ311++G(d,p) level.
9
ACS Paragon Plus Environment

Page 9 of 29
Journal of the American Chemical Society
Table 1. Experimental IR spectral data (argon matrix at 10 K), B3LYP/6ꢀ311++G(d,p) calculated
vibrational frequencies (ν, cm−1), absolute infrared intensities (Ath, km mol−1), and vibrational
assignment of deuterated formyl benzazirine 5ꢀd.a
1
2
3
4
5
6
Ar matrixb
Calculatedc
Approx. assignmentd
I
Ath
7
8
ν
ν
9
2108/2097/2093 ov/br/m
1751
1706/1695/1689/1685 ov/s/s/s
2109
1802
1710
1586
1484
1415
1351
1221
1183
1139
1108
1013
997
948
936
783
747
711
649
59.8
34.8
ν(C–D)
ν(C=N)
341.3 ν(C=O)
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
w
1595/1589
1488
ww/vw
m
3.7
26.3
3.9
7.0
57.4
5.5
ν(C=C)as
ν(C=C)s
δ1(C–H)
δ2(C–H)
ν(C2–C1)
δ3(C–H)
δ4(C–H)
ν(C3–C2–C1)as, δ(C–D)
δ1(6ꢀring)
δ(C–D)
γ1(C–H) + ν(C6–C5)
γ1(C–H) – ν(C6–C5)
γ2(C–H)
δ2(6ꢀring)
γ3(C–H)
1419
1355
vw
w
1241
s
1189
1143
1125
vw
vw
br
6.0
22.5
23.3
7.8
5.0
4.4
41.2
4.3
67.4
8.5
1018/1016
1004
br
vw
vw/ov
ꢀ
947
ꢀ
784/782
747
714/712
643/640
s/s
w/ov
s/s
m/w
τ1(6ꢀring)
aDeuterated formyl benzazirine 5ꢀd was generated by irradiation of deuterated 2ꢀformylꢀphenylnitrene 2ꢀd in an argon
matrix at 10 K. Only bands in the 2600−600 cm−1 region having calculated intensities above 3 km mol−1 are included.
bExperimental intensities are presented in qualitative terms: s = strong, m = medium, w = weak, vw = very weak, br = broad,
and ov = overlapped. cB3LYP/6ꢀ311++G(d,p) calculated scaled frequencies. dAssignments made by inspection of Chemcraft
animations. Abbreviations: ν = stretching, δ = bending, γ = rocking, τ = torsion, s = symmetric, as = antisymmetric, and
6ꢀring = sixꢀmembered ring. Signs “+” and “−” designate combinations of vibrations occurring in “syn”ꢀphase (“+”) and in
“anti”ꢀphase (“−”). eFor atom numbering see Scheme S1.
3
The IR spectrum of the produced D is assigned to deuterated formyl cyclic ketenimine 6ꢀd based on
the good match between the experimental and calculated IR spectra (Figure 4b,c). Cyclic ketenimines
are known to have a characteristic strong IR absorption near 1900 cm−1, due to ν(C=C=N)as
vibration,16,39–41 which correlates well with the most intense band observed in the IR spectrum of D at
1890/1880 cm−1. Other characteristic bands of D are observed at 2137/2117, 1698, 1576 and ~1525
cm−1 in good agreement with the estimated IR bands of 6ꢀd at 2109 [ν(OC–D)], 1711 [ν(C=O)], 1579
[ν(C=C)as] and 1519 [ν(C=C)s] cm−1. Four different cyclic ketenimine structures may, in principle, be
formed in the chemistry of nitrene 2ꢀd (Table S6). Based on vibrational calculations, only the 6ꢀd form
10
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 29
matches the experimental IR spectrum of D (Figure S4 and S5). The formation of this structure is in
accordance with the expectations because the ring expansion via C–C cleavage in benzazirine isomer
5ꢀd directly results in cyclic ketenimine isomer 6ꢀd.42 A detailed assignment of the IR spectrum of D,
identified as cyclic ketenimine 6ꢀd, is given in Table 2.
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
Table 2. Experimental IR spectral data (argon matrix at 10 K), B3LYP/6ꢀ311++G(d,p) calculated
vibrational frequencies (ν, cm−1), absolute infrared intensities (Ath, km mol−1), and vibrational
assignment of deuterated formyl cyclic ketenimine 6ꢀd.a
Ar matrixb
Calculatedc
Approx. assignmentd
I
Ath
ν
ν
2133/2117
1890/1880
1698
1576
1533/1525
1297
vw/ov
s/s
s/ov
m
vw/w
w
m
2128
1913
1711
1579
1519
1308
1288
1205
1154
1110
1037
1005
968
956
923
843
775
756
684
663
609
65.6
ν(C–D)
175.0 ν(C=C=N)as
236.7 ν(C=O)
50.5
66.5
3.9
46.3
6.7
38.0
26.2
4.9
3.7
9.4
5.8
5.5
ν(C=C)as
ν(C=C)s
δ1(C–H)
δ2(C–H)
δ3(C–H)
ν(C2–C1)
δ4(C–H)
ν(C7C6) – ν(C5C4)
δ(C–D)
ν(C7C6) + ν(C5C4)
γ1(C–H)
δ1(ring)
γ(C–D), δ2(ring)
γ2(C–H)
γ3(C–H)
γ4(C–H)
1286
1208
1164
w
m
w
ꢀ
1110/1004
ꢀ
ꢀ
ꢀ
965/962
950/949
916/914
ꢀ
vw/vw
w/w
vw/vw
ꢀ
3.1
776
m
24.2
68.1
67.7
25.9
3.9
759/756/755
686/681
658
m/m/s
m/m
w
δ(C=C=N)
τ1(ring)
602
vw
aDeuterated formyl cyclic ketenimine 6ꢀd was generated by ring expansion of deuterated formyl benzazirine 5ꢀd in an argon
matrix at 10 K. Only bands in the 2600−600 cm−1 region having calculated intensities above 3 km mol−1 are included.
bExperimental intensities are presented in qualitative terms: s = strong, m = medium, w = weak, vw = very weak, and
c
d
ov = overlapped. B3LYP/6ꢀ311++G(d,p) calculated scaled frequencies. Assignments made by inspection of Chemcraft
animations. Abbreviations: ν = stretching, δ = bending, γ = rocking, τ = torsion, s = symmetric, and as = antisymmetric.
Signs “+” and “−” designate combinations of vibrations occurring in “syn”ꢀphase (“+”) and in “anti”ꢀphase (“−”). eFor atom
numbering see Scheme S1.
A similar approach was applied to investigate the possibility of ring expansion in the protium
3
benzazirine 5ꢀh. Triplet 2ꢀformylꢀphenylnitrene 2ꢀh was first produced by irradiation of 2ꢀformylꢀ
phenylazide 1ꢀh and then immediately exposed to irradiation at λ = 530 nm. The photochemistry of 32ꢀh
11
ACS Paragon Plus Environment

Page 11 of 29
Journal of the American Chemical Society
is similar to that described for 32ꢀd; at least three different products labeled as Aʹ (2111 cm−1), Bʹ (1843
cm−1) and Cʹ (~1710 cm−1) were formed (Figure S6 and S7). However, it must be noted that tunneling
1
2
3
4
5
6
3
reaction of 2ꢀh to iminoꢀketene 3ꢀh (τ1/2 ~6 hours)19 occurs simultaneously with the photochemistry
7
8
3
triggered at λ = 530 nm. After 1 hour of irradiation at this wavelength only a few percent of 2ꢀh
9
(<10 %) remained. The nitrene 32ꢀh is difficult to convert in its totality through irradiation at 530 nm on
during this time scale because the formation of 3ꢀh (Aʹ) that also absorbs at λ = 530 nm and is
transformed into benzoazetinone 4ꢀh (Bʹ). Thus, after the irradiation at λ = 530 nm, the remaining
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
nitrene 2ꢀh was first allowed to disappear in the dark by tunneling and subsequently the sample was
kept at 10 K in the spectrometer beam for 3 days. At that time, the rearrangement of Cʹ into Dʹ was
completed and the IR difference spectrum corresponding to this transformation was obtained (Figure 5b).
80
Simulated (c)
ov
ꢁ
40
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
0
0.04
Decay, 10 K (b)
D'
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
H O
ꢁ
ꢁ
ꢁ
ꢁ
2
ꢁ
ꢁ
ꢁꢁ
ꢂ
0.00
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
C'
-0.04
80
ꢀ
Simulated (a)
H
N
O
40
0
5
-h
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ ꢀ
ꢀ
ꢀ
1900
1700
1500
1300
1100
900
700
Wavenumber / cm−1
Figure 5. (a) IR spectrum of formyl benzazirine 5ꢀh simulated at B3LYP/6ꢀ311++G(d,p) level. (b)
Experimental difference IR spectrum showing changes occurred after keeping the sample at 10 K in the
spectrometer beam for 3 days, subsequent to irradiation at λ = 530 nm and after the decay of the
3
remained triplet nitrene 2ꢀh. The negative bands are due to the consumed Cʹ assigned to benzazirine
5ꢀh (solid triangles, ꢀ). The positive bands are due to the formation Dʹ assigned to cyclic ketenimine
6ꢀh (empty triangles, ꢁ). (c) IR spectrum of formyl cyclic ketenimine 6ꢀh simulated at
B3LYP/6ꢀ311++G(d,p) level.
12
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 29
Based on the excellent agreement between the experimental and calculated IR spectra, the consumed
species Cʹ was assigned to formyl benzazirine 5ꢀh and the produced species Dʹ assigned to formyl
cyclic ketenimine 6ꢀh (Figure 5, Tables S7 and S8). Hence, in addition to the experiment described
above for the deuterated derivative, we also found that protium benzazirine 5ꢀh undergoes ring
expansion to cyclic ketenimine 6ꢀh.
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
Kinetics of the benzazirine ring expansion to cyclic ketenimine
To better understand the nature of the transformation of benzazirine into cyclic ketenimine, their
respective decrease and increase were followed by IR spectroscopy (using the bands at 1241 cm−1 for
5ꢀd and 1890 cm−1 for 6ꢀd for the deuterated derivative, and using the bands at 1242 cm−1 for 5ꢀh and
1883 cm−1 for 6ꢀh for the protium derivative) while keeping the sample under different given sets of
conditions. For “unfiltered IR light” conditions, the kinetics of rearrangement of benzazirine 5ꢀd into
cyclic ketenimine 6ꢀd was measured when the sample was kept at 10 K and permanently exposed to the
IR light source of the FTIR spectrometer. As shown in Figure 6, the data are nicely fitted by equations
of monoꢀexponential decay kinetics, with a rate constant of ~1.5 × 10−5 s−1 and a halfꢀlife of ~13 hours.
For “dark” conditions, new experiments were performed where the sample was kept at 10 K in dark
with only exception when monitoring the IR spectra, which was performed using a cutoff filter
transmitting only below 2200 cm−1 placed between the sample and the spectrometer IR light source. As
shown in Figure 7, under “dark” conditions (Kinetics 1) only ~25 % of 5ꢀd rearranged to 6ꢀd upon 5
days. After this period of 5 days, “unfiltered IR light” conditions were applied (Kinetics 2) to confirm
that a much faster rate of the rearrangement occurs under the influence of IR light. The “dark” kinetics
(Figure 7, Kinetics 1) shows some dispersive character, i.e. it deviates from the firstꢀorder kinetics and
exhibits a distribution of rate constants that decrease over time. This is commonly observed for reactions
in cryogenic matrices and is generally attributed to distributions of matrix sites and weak interactions
between substrate and the matrix environment.15,19,27–30 Anyway, for the sake of a rough estimate, the
data for “dark” kinetics were fitted by using classical equations of monoꢀexponential decay, resulting in
a rate constant of ~7.4 × 10−7 s−1 and a halfꢀlife of ~260 hours. This “dark” rate constant is more than an
order of magnitude smaller than the “unfiltered light” rate constant, unmistakably indicating a
significant difference in both kinetics.
13
ACS Paragon Plus Environment

Page 13 of 29
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
1.0
0.8
0.6
0.4
0.2
0.0
9
10
11
12
13
14
15
16
0
8
16
24
32
Time / hours
Figure 6. Kinetics of rearrangement of benzazirine 5ꢀd to cyclic ketenimine 6ꢀd in an argon matrix at 10
K, with the sample exposed to the IR light source of the FTIR spectrometer. Empty red () and blue
() circles represent the time evolution of the amounts of 5ꢀd (consumption, measured using the peak at
1241 cm−1) and 6ꢀd (production, measured using the peak at 1890 cm−1), respectively. Red and blue
dotted lines represent best fits obtained using firstꢀorder exponential kinetics equations. The rate
constants are k5ꢀd = 1.4 × 10−5 s−1 and k6ꢀd = 1.5 × 10−5 s−1. More details are given in the experimental
section.
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
Kinetics 1
Kinetics 2
1.0
0.8
0.6
0.4
0.2
0.0
5 days
10 days
240 300
Time / hours
0
60
120
180
360
Figure 7. Kinetics of rearrangement of benzazirine 5ꢀd (measured using the peak at 1241 cm−1) to
cyclic ketenimine 6ꢀd (measured using the peak at 1890 cm−1) in an argon matrix at 10 K. Kinetics 1
(initial 5 days, left): rearrangement of 5ꢀd (solid red circles (ꢀ), consumption) into 6ꢀd (solid blue
circles (ꢀ), production) with the sample kept in dark, except when the spectra were recorded, the
sample was then protected by an infrared longpass cutoff filter transmitting only below 2200 cm−1.
Kinetics 2 (after 5 days in dark, right): rearrangement of 5ꢀd (empty red circles (), consumption) into
6ꢀd (empty blue circles (), production) with the sample permanently exposed to the IR light source of
the FTIR spectrometer. Blue and red solid lines represent best fits obtained using firstꢀorder single
exponential kinetics equations, and adjustments were made using only the amounts of the reactant 5ꢀd
and the product 6ꢀd during the first 5 days of observation (sample in dark), corresponding to ~25% of
the conversion. The rate constants are k5ꢀd = 7.1 × 10−7 s−1 and k6ꢀd = 7.7 × 10−7 s−1. Blue and red dotted
lines (right) are presented to guide the eye only.
14
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 29
1
2
3
4
5
6
7
8
In other independent experiment, “filtered IR light” conditions were applied, whereas a cutoff filter
transmitting light only below 2200 cm−1 was placed between the IR light source of the FTIR
spectrometer and the sample. In that case, the sample was exposed to filtered light continuously (i.e.,
including periods between the registrations of spectra). The obtained kinetics for “filtered light” decay
was found to be similar to the “dark” decay kinetics.
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
Experiments were also carried out to measure the kinetics of rearrangement of protium benzazirine
5ꢀh into protium cyclic ketenimine 6ꢀh. As shown in Figure S8, under “dark” conditions (Kinetics 1) the
rearrangement after 5 days amounts to ~27 %. Afterwards, “unfiltered IR light” conditions were applied
(Kinetics 2) and a strong increase in the reaction rate was observed, similarly to the behavior found for
the deuterated derivative. Again, similarly to the deuterated compound, the kinetics of rearrangement
from 5ꢀh to 6ꢀh in the dark has some dispersive character. To obtain a rough estimate, the data for
“dark” kinetics were fitted by using equations of the firstꢀorder kinetics, resulting in a rate constant of
~8.8 × 10−7 s−1 and a halfꢀlife of ~219 hours. This rate constant of the rearrangement from 5 to 6 under
“dark” conditions of protium compound differs somewhat from the analogous rate found for the
deuterated analogue. Even within the measurement accuracy of peak intensities of the bands used and
also with the caveat that a singleꢀexponential fitting is, only approximately applicable to trace the
dispersive kinetics, these rates suggest a small decrease in the rate resulting from the substitution of
protium with deuterium in the carbonyl group of 5. Indeed, as it will show in the next section,
computations indicate that the secondary kinetic isotope effect is significant.
To verify if the rearrangement is not activated thermally, we carried out additional kinetic
measurements of the rearrangement of 5ꢀd to 6ꢀd at 20 K also under dark conditions (Figure S9). Again,
even though the kinetics has some dispersive character, the data were adjusted by using the firstꢀorder
kinetic equations and a rough rate constant was estimated as ~8.9 × 10−7 s−1 and a halfꢀlife of ~216
hours. Within the precision of estimations and considering the data from a semiꢀqualitative perspective,
the rate of the rearrangement hardly shows any increase upon increase of the absolute temperature by a
15
ACS Paragon Plus Environment

Page 15 of 29
Journal of the American Chemical Society
factor of two. The matrix softening induced by the temperature increase can be responsible for the
possible small acceleration of the rearrangement, as proposed for other cases of tunneling reactions in
matrices.15,27–30
1
2
3
4
5
6
CCSD(T) computations (see below) estimate an energy barrier of 7.5 kcal mol−1 for the rearrangement
of 5 to 6 (~1.4 kcal mol−1 more stable). This energy barrier is far too high to afford a thermal activation
at the low working temperatures up to 20 K under “dark” conditions. Moreover, attaining such a high
barrier seems not possible under the experimental “filtered light” condition, when the sample at 10 K is
exposed to light with wavenumbers below 2200 cm−1 (~6.3 kcal mol−1 or less). The data indicate
therefore that the rearrangement of benzazirine 5 to cyclic ketenimine 6 in low temperature matrix
occurs through heavyꢀatom tunneling. Because the reaction rate under “dark” conditions is similar to the
rate when the sample is exposed to “filtered IR light” conditions, the vibrationally assisted tunneling is
unlikely.
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
In the case of “unfiltered light” conditions, the observed significantly faster rate, compared to “dark”
or “filtered IR light” conditions, suggests the occurrence of IRꢀinduced photochemistry. Here the
sample is clearly exposed to light with energy well above the calculated barrier.43 The energy deposited
in vibrational modes of benzazirine 5 located above the reaction barrier, for instance into the CH
stretching modes or in any overtone/combination transition with frequency above ~2625 cm−1 (7.5 kcal
mol−1) may promote isomerization to cyclic ketenimine 6 (Figure S10). In the opposite direction, from 6
to 5, the estimated reaction barrier is higher, ca. ~8.9 kcal mol−1 (~3125 cm−1). This barrier is above the
energy of the CH stretching modes (~3000 cm−1) in cyclic ketenimine 6. Hence, the probability of
depositing energy in vibrational modes above the reaction barrier is higher in 5 than in 6, which could
explain why the broadband IR irradiation tends to promote IRꢀinduced chemistry toward 6.44
Tunneling Calculations
To support the experimental evidence of quantum tunneling effect in the process of formation of
ketenimine 6 from cyclic benzazirine 5, we computed the rate constants without and with tunneling
16
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 29
(using canonical variational transition state theory –CVT− and small curvature tunneling –SCT–,
1
2
3
4
5
6
7
8
9
respectively).45 The estimated CVT rate constant at 10 K is 1.8 × 10−177 s−1, which indicates that the 5ꢀh
to 6ꢀh reaction is impossible to occur. In contrast, the SCT rate constant for the 5ꢀh to 6ꢀh reaction at
10 K (from the ground state at this temperature) is 3.5 × 10−5 s−1 (a halfꢀlife time of ~6 hours), which is
comparable to experimental result (around 40 times faster). The difference is justifiable considering the
difference between the solid matrix experiment and the gas phase computation plus the error of the DFT
theoretical method [M06ꢀ2X gives an activation energy of 6.2 kcal mol−1, 1.3 kcal mol−1 lower than
CCSD(T)]. Thus, the results validate the computational method and at the same time corroborate the
observed tunneling effect of the experiments: going from 5 to 6 is a transformation only attainable by a
tunneling process from the ground state at cryogenic conditions (see Figure 8). At 20 K the computed
SCT rate constant is 3.6 × 10−5 s−1, almost the same as at 10 K, indicating that the first vibrational
excited quantum level of the reactant remains practically not thermally populated, and a thermally
activated tunneling process should be negligible. This indicates that a marginal increase of the
experimental rate at 20 K compared to the 10 K should be attributed to matrix softening.
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 8. Arrhenius graph for the benzazirine (5ꢀh and 5ꢀd) to cyclic ketenimine (6ꢀh and 6ꢀd) reaction
without (CVT) and with (SCT) tunneling correction. Computations indicate that below ~150 K
tunneling is the main component of the rate constant, and below 20 K the reaction is completely
independent of the temperature (tunneling exclusively from the ground state).
17
ACS Paragon Plus Environment

Page 17 of 29
Journal of the American Chemical Society
We then proceeded to compute the kinetic isotope effect (KIE) not only for the formyl hydrogen but
1
2
3
12
for all the other atoms as well (using the C/13C, 14N/15N, 16O/18O and H/D mass relations). The results
4
5
6
7
8
at 10 K are shown in Table 3. As can be seen, 5ꢀd indeed has a smaller rate of reaction compared to 5ꢀh,
as was observed experimentally. The experimental KIE value was approximately 1.20 (8.8 × 10−7
9
s−1/7.4× 10−7 s−1), while theory predicts a KIE of 1.88 (first row in Table 3); this difference can be
ascribed to the same reasons explained above, but still both results clearly indicate a significant
deceleration of the reaction. Notably, there is a significant negative ꢀZPE‡ value of −0.27 kJ mol−1 in
the H/D substitution at the formyl hydrogen which, opposite to the observed KIE, leads to an inverse
semiꢀclassical KIE and a potentially faster tunneling due to a lower barrier;46 nevertheless, the mass
effect of the deuterium substitution strongly overrides this effect. Other H/D substitutions also provide a
significant KIE (except for H4). In the case of heavyꢀatom tunneling, we can see that C3 has the largest
KIE, indicating that this is the “tunneling determining atom”, the one with the widest displacement. But
most of other atoms also have large KIE, denoting a holistic molecular rearrangement. Interestingly the
pivotal nitrogen, with a KIE of only 1.05, is comparatively static and is mostly a “spectator” in the
tunneling mechanism. In this sense, the heavy atom tunneling is predominantly carbon 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
12
14
Table 3. Computed kinetic isotope effect for the benzazirine 5 ring expansion for the C/13C, N/15N,
16O/18O and H/D isotopic substitutions at 10 K.
Atom KIE
Atom
KIE
C1
C2
C3
C4
C5
C6
C7
1.29
1.23
1.88
1.16
1.17
1.04
1.23
H1
H2
H3
H4
H5
N
1.88
1.80
2.16
1.05
1.53
1.05
1.17
O
18
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 18 of 29
CONCLUSIONS
1
2
3
3
3
The summary of the experimental results of the photochemistry of protium 2ꢀh and deuterated 2ꢀd
4
5
6
3
triplet 2ꢀformylꢀphenylnitrene is presented in Scheme 3. Irradiation of 2, generated in argon matrix at
7
8
9
10 K from 2ꢀformylꢀphenylazide 1, led to the formation of iminoꢀketene 3 and benzazirine 5. The
iminoꢀketene 3 was found to subsequently photorearrange to benzoazetinone 4. The benzazirine 5 was
found to undergo spontaneous ring expansion to cyclic ketenimine 6 in the dark at 10 K, despite an
estimated reaction barrier of 7.5 kcal mol−1. Moreover, the reaction rate was observed to barely show
any increase upon increase of the absolute temperature by a factor of two (from 10 to 20 K). These
experimental data clearly provide strong evidence that ring expansion of 5 to 6 occurs by heavyꢀatom
tunneling. Computed rate constants with and without tunneling (SCT and CVT approximation,
respectively) confirm that the ring expansion of 5 to 6 can only be attainable by a tunneling process
from the ground state at cryogenic conditions. In addition, a small secondary kinetic isotopic effect,
corresponding to the deceleration of the reaction upon substitution of protium with deuterium in the
formyl group of benzazirine, was predicted theoretically and measured experimentally. Also interesting,
it was observed that the ring expansion of 5 to 6 is stimulated by broadband IR radiation. The reaction
rate was measured to be more than one order of magnitude faster when the sample at 10 K is exposed to
the radiation from the IR light source of FTIR spectrometer.
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 3. Summary of the reactions experimentally observed on the potential energy surface of the aryl
nitrene 2.
19
ACS Paragon Plus Environment

Page 19 of 29
Journal of the American Chemical Society
Overall, the results unravel a remarkable feature on the potential energy surface of an aryl nitrene. We
discovered two isomeric species connected photochemically that both react by quantum tunneling: the
2ꢀformylꢀphenylnitrene by proton tunneling, as reported previously,19 and the corresponding benzazirine
by heavyꢀatom tunneling, as reported here. To the best of your knowledge, this is the first system where
interconverted isomers exhibit two different types of tunneling.
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
EXPERIMENTAL SECTION
Sample: Protium 2ꢀformylꢀphenylazide 1ꢀh and deuterated 2ꢀformylꢀphenylazide 1ꢀd were prepared
as described in our previous communication.19 The deuterium enrichment of sample 1ꢀd is ≥ 90 %.
Between the experiments, the samples of 1ꢀh or 1ꢀd were stored in a glass tube, under vacuum, in
refrigerator at +4°C, and in dark. Under such conditions the azides were stable for many weeks
Matrix Isolation IR Spectroscopy: A solid sample of 1ꢀh or 1ꢀd was placed in a glass tube that was
then connected to a needle valve (SS4 BMRG valve, NUPRO) attached to the vacuum chamber of a
heliumꢀcooled cryostat (APD Cryogenics closedꢀcycle refrigerator with DEꢀ202A expander). Prior to
deposition of the matrices, the samples were purified from the volatile impurities by pumping through
the cryostat at room temperature. The sample tube with 1ꢀh or 1ꢀd was subsequently cooled with an iceꢀ
water mixture (at 0°C) to reduce the saturated pressure of the compound and improve the metering
function of the needle valve. During the preparation of the matrices, the sample vapors were deposited,
together with a large excess of argon (N60, Air Liquide) onto a CsI window at 15 K, used as optical
substrate. After the deposition, all samples were cooled down to 10 K and kept at this temperature both
during irradiations and during the monitoring of spontaneous decays (unless stated otherwise). The
temperature was measured directly at the sample holder window by a silicon diode sensor connected to
a digital controller providing stabilization accuracy of 0.1 K.
The midꢀIR spectra (4000ꢀ400 cm−1 range) were recorded using a Thermo Nicolet 6700 Fourier
transform infrared spectrometer, equipped with a deuterated triglycine sulphate (DTGS) detector and a
20
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 20 of 29
KBr beam splitter, with 0.5 cm−1 resolution. To avoid interference from atmospheric H2O and CO2, a
stream of dry air was continuously purged through the optical path of the spectrometer. In some
experiments, to avoid exposing the sample matrix to the light source of the FTIR spectrometer with
wavenumbers higher than 2200 cm−1, the midꢀIR spectra were recorded only in the 2200ꢀ400 cm−1
range, with a standard Edmund Optics longꢀpass filter placed between the spectrometer sources and the
cryostat.
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
UVꢀVis Irradiation Experiments: The matrices were irradiated, through the outer KBr window of
the cryostat, using tunable narrowband (~0.2 cm−1 spectral width) light provided by a signal (visible
light) or a frequencyꢀdoubled signal (UV range) beam of the Spectra Physics QuantaꢀRay MOPOꢀSL
optical parametric oscillator pumped with a pulsed Nd:YAG laser (repetition rate = 10 Hz, duration = 10
ns). The pulse energy for visible and UVꢀlight irradiations was 9ꢀ10 mJ and ~1ꢀ3 mJ, respectively.
Irradiation with UVꢀlight at 308 nm was used to induce the initial photochemistry of the azide precursor
(1ꢀh and 1ꢀd) because it matches its first absorption and avoid, as much as possible, the absorption of
corresponding product triplet nitrene (32ꢀh and 32ꢀd) with a maximum at around 326 nm.
Kinetics: Monitoring of the decays of 5ꢀd into 6ꢀd and of 5ꢀh into 6ꢀh was carried out by successive
registrations of infrared spectra, as a function of time. For all of the three different sets of external light
(“dark”, “filtered IR light”, “unfiltered IR light”, described above) the infrared spectra were collected
using 64 scans. Under such conditions, the collection length of one spectrum equals to 261 seconds.
Already during these initial minutes, a part of the photoproduced benzazirine 5 undergoes decay into
cyclic ketenimine 6. In the kinetic analysis, the moment of registration of the first spectrum was
assumed to be the origin of decay time, the intensity of the benzazirine 5 bands present in this first
spectrum was assumed to be relative 100%, and the amount of cyclic ketenimine 6 present extracted
from the same first spectrum (formed between the irradiation and the registration of the first spectrum)
was assumed to be relative 0%. The amounts of particular species in the sample were followed by the
peak intensities of the most characteristic absorption bands, nonꢀoverlapping with absorptions of other
species. For this purpose, the following peaks were selected: 1241 cm−1 for 5ꢀd (Table 1), 1890 cm−1 for
21
ACS Paragon Plus Environment

Page 21 of 29
Journal of the American Chemical Society
6ꢀd (Table 2), 1242 cm−1 for 5ꢀh (Table S7), 1883 cm−1 for 6ꢀh (Table S8). For “unfiltered IR light”
kinetics, (Figure 6) the decay of 5ꢀd was monitored for 32 hours and, upon that time, ~20% of the
benzazirine remained in the sample. The “dark” decay kinetics of 5ꢀd (Figure 7, left) was monitored for
120 hours and, upon that time, ~75% of benzazirine 5ꢀd still remained in the sample. After that period,
the “unfiltered IR light” conditions were implemented, and the decay was followed during subsequent
65 hours (Figure 7, right). During that time, the amount of benzazirine 5ꢀd decreased from ~75% to just
~5%. In a similar experiment with the protium analogue 5ꢀh, ~73% of benzazirine 5ꢀh remained in the
sample after 120 hours of applying “dark” conditions. After subsequent 72 hours of observation using
the “unfiltered IR light conditions”, the amount of 5ꢀh decreased from ~73% to less than 5%. The
amounts of cyclic ketenimine 6 generated in the process of tunneling were then normalized in such a
way so that at the end of each monitoring period the sum of 5 and 6 would be 100%. These
normalizations assumed a direct quantitative transformation of 5 into 6, which was indeed confirmed
spectroscopically. The decays observed in present experiments (presented in Figures 6, 7, S8 and S9)
were fitted, for rough estimation of the rate constants, using the equations of a singleꢀexponential decay:
[R]t=[R]0 exp[−k t]. The rate constants k extracted from these fits, and the corresponding halfꢀlives, are
discussed along the text.
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
Theoretical Calculation: Calculations performed at B3LYP/6ꢀ311++G(d,p) level were carried using
Gaussian0947. With the aim of modeling the IR spectra, geometry optimizations at B3LYP/6ꢀ
311++G(d,p) level were followed by harmonic frequency calculations at the same level. To correct the
vibrational anharmonicity, basis set truncation, and the neglected part of electron correlation, the
calculated frequencies were scaled by 0.9848 (0.97 for the ν(N−D) mode). The scaled frequencies and
the calculated IR intensities were then convoluted with Lorentzian functions having an fwhm = 2 cm–1.
The integral area of the simulated band equals to the theoretical calculated IR intensity. Due to the
broadening, the peak intensities in the simulated spectra are reduced compared with the calculated
intensities (in km mol−1), and therefore, are shown in arbitrary units of “relative intensity”. Assignments
of vibrational modes were made by inspection of ChemCraft49 animations.
22
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 22 of 29
Tunneling Calculation: Single point energy values were computed at the CCSD(T)ꢀF12/ccꢀpVTZꢀ
F12//M06ꢀ2X/6ꢀ311+G(2d,p) level50–52 (coupled cluster computations were carried out with Orca53
version 4.0.1, DFT optimizations with Gaussian0947). All the computed reaction rates including
quantum tunneling were carried out using the multidimensional small curvature tunneling method
(SCT)54, with step size of 0.001 Bohr, quantized reactant state tunneling (QRST) for the reaction
coordinate mode, and at the M06ꢀ2X/6ꢀ311+G(2d,p) level. Using a multidimensional method to account
for “cornerꢀcutting” was critical, as the results using the unidimensional zeroꢀcurvature tunneling (ZCT)
are three orders of magnitude less for this reaction. All the rates were computed with Polyrate55, using
Gaussrate56 as the interface with Gaussian.
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
SUPPORTING INFORMATION
The Supporting Information is available free of charge on the ACS Publications website at DOI:
Additional experimental results, IR assignments, and computational data.
AUTHOR INFORMATION
Corresponding Author:
*cmnunes@qui.uc.pt
ORCID
Cláudio M. Nunes: 0000ꢀ0002ꢀ8511ꢀ1230
NOTES
The authors declare no competing financial interest.
ACKNOWLEDGMENT
This work was supported by the Portuguese “Fundação para a Ciência e a Tecnologia” (FCT). The
Coimbra Chemistry Centre is supported by the FCT through the project UID/QUI/0313/ 2013, cofunded
by COMPETE. C.M.N. and I.R. acknowledge the FCT for Postdoctoral Grant No.
SFRH/BPD/86021/2012 and Investigador FCT Grant, respectively.
23
ACS Paragon Plus Environment

Page 23 of 29
Journal of the American Chemical Society
REFERENCES
1
2
3
4
(1) Nitrenes and Nitrenium Ions; Falvey, D. E.; Gudmundsdóttir, A. D., Eds.; John Wiley & Sons,
5
6
2013.
7
8
9
(2) Gritsan, N.; Platz, M. S. Photochemistry of Azides: The Azide/Nitrene Interface. In Organic Azides:
Syntheses and Applications; Bräse, S.; Banert, K., Eds.; John Wiley & Sons, 2010; pp. 311–372.
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) Platz, M. S. Nitrenes. In Reactive Intermediate Chemistry; Moss, R. A.; Platz, M. S.; Jones, M. J.,
Eds.; John Wiley & Sons, 2004; pp. 501–559.
(4) Gritsan, N.; Borden, W. T; Platz, M. The Study of Nitrenes by Theoretical Methods. In
Computational Methods in Photochemistry; Kutateladze, A. G., Eds.; CRC Press, 2005; pp. 235–356.
(5) Wentrup, C. Chem. Rev. 2017, 117, 4562–4623.
(6) Gritsan, N. P.; Platz, M. S. Chem. Rev. 2006, 106, 3844–3867.
(7) Borden, W. T.; Gritsan, N. P.; Hadad, C. M.; Karney, W. L.; Kemnitz, C. R.; Platz, M. S. Acc.
Chem. Res. 2000, 33, 765–771.
(8) Smith, E.; Collins, I. Future Med. Chem. 2015, 7, 159–183.
(9) Tasdelen, M. A.; Yagci, Y. Angew. Chem. Int. Ed. 2013, 52, 5930–5938.
(10) Liu, L.; Yan, M. Acc. Chem. Res. 2010, 43, 1434–1443.
(11) Bräse, S.; Gil, C.; Knepper, K.; Zimmermann, V. Angew. Chem. Int. Ed. 2005, 44, 5188–5240.
(12) Lineberger, W. C.; Borden, W. T. Phys. Chem. Chem. Phys. 2011, 13, 11792–11813.
(13) Wijeratne, N. R.; Da Fonte, M. ; Ronemus, A.; Wyss, P. J.; Tahmassebi, D.; Wenthold, P. G.
J. Phys. Chem. A 2009, 113, 9467–9473.
(14) Karney, W.; Borden, W. J. Am. Chem. Soc. 1997, 119, 1378–1387.
(15) Inui, H.; Sawada, K.; Oishi, S.; Ushida, K.; McMahon, R. J. J. Am. Chem. Soc. 2013, 135, 10246–
10249.
(16) Grote, D.; Sander, W. J. Org. Chem. 2009, 74, 7370–7382.
(17) Pritchina, E. A.; Gritsan, N. P.; Bally, T. Phys. Chem. Chem. Phys. 2006, 8, 719–727.
(18) Morawietz, J.; Sander, W. J. Org. Chem. 1996, 61, 4351–4354.
24
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 24 of 29
1
2
3
(19) Nunes, C. M.; Knezz, S. N.; Reva, I.; Fausto, R.; McMahon, R. J. J. Am. Chem. Soc. 2016, 138,
4
5
15287–15290.
6
7
8
9
(20) Gerbig, D.; Schreiner, P. R. Angew. Chem. Int. Ed. 2017, 9445–9448.
(21) Ley, D.; Gerbig, D.; Schreiner, P. R. Org. Biomol. Chem. 2012, 10, 3781–3790.
(22) Layfield, J. P.; HammesꢀSchiffer, S. Chem. Rev. 2014, 114, 3466–3494.
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
(23) Lambert, N.; Chen, Y.ꢀN.; Cheng, Y.ꢀC.; Li, C.ꢀM.; Chen, G.ꢀY.; Nori, F. Nat. Phys. 2012, 9,
10ꢀ18.
(24) Borden, W. T. WIREs Comput. Mol. Sci. 2015, 6, 20–46.
(25) Two possible additional cases of heavyꢀatom tunneling are the ring closure reactions of triplet
1,3ꢀdiradicals, cyclopentaneꢀ1,3ꢀdiyl and cyclobutaneꢀ1,3ꢀdiyl. See refs. 25a and 25b, respectively.
Because both reactions require that the triplet 1,3ꢀdiradical undergoes ISC to form the singlet product,
the rationalization of the mechanism involved is more complex. (25a) Buchwalter, S. L.; Closs, G. L.
J. Am. Chem. Soc. 1979, 101, 4688–4694. (25b) Sponsler, M. B.; Jain, R.; Coms, F. D.; Dougherty, D.
A. J. Am. Chem. Soc. 1989, 111, 2240–2252.
(26) Orendt, A. M.; Arnold, B. R.; Radziszewski, J. G.; Facelli, J. C.; Malsch, K. D.; Strub, H.; Grant,
D. M.; Michl, J. J. Am. Chem. Soc. 1988, 110, 2648–2650.
(27) Zuev, P. S.; Sheridan, R. S.; Albu, T. V.; Truhlar, D. G.; Hrovat, D. A.; Borden, W. T. Science
2003, 299, 867–870.
(28) Moss, R. A.; Sauers, R. R.; Sheridan, R. S.; Tian, J.; Zuev, P. S. J. Am. Chem. Soc. 2004, 126,
10196–10197.
(29) Ertelt, M.; Hrovat, D. A.; Borden, W. T.; Sander, W. Chem. A Eur. J. 2014, 20, 4713–4720.
(30) Schleif, T.; MierezꢀPerez, J.; Henkel, S.; Ertelt, M.; Borden, W. T.; Sander, W. Angew. Chem. Int.
Ed. 2017, 56, 10746–10749.
(31) Nandi, A.; Gerbig, D.; Schreiner, P. R.; Borden, W. T.; Kozuch, S. J. Am. Chem. Soc. 2017, 139,
9097–9099.
(32) Only the antiꢀform of the deuterated 2ꢀformylꢀphenylazide 1ꢀd precursor, considering the
orientation of the –CDO moiety, was populated in the gas phase and deposited in the matrix.
3
Accordingly, only the antiꢀform of nitrene 2ꢀd was generated in the matrix upon photolysis of 1ꢀd
25
ACS Paragon Plus Environment

Page 25 of 29
Journal of the American Chemical Society
(see also ref. 19).
1
2
3
4
5
6
7
3
(33) In the experimental difference UVꢀVis spectrum obtained for triplet nitrene 2 (ref. 19), the
presence of a strong absorption above 350 nm due to imino ketene 3 precluded the identification of
3
weaker absorptions due to the nitrene in this region. However, the occurrence of photochemistry for 2
8
9
with onset at 530 nm is coherent with the estimation of very weak electronic transitions at 449 and 536
nm by TDDFT calculations (ref. 19). Indeed, it has been observed experimentally that triplet
arylnitrenes have a characteristic weak feature that extends to 500 nm or higher (ref. 6 and 33a). (33a)
Carra, C.; Nussbaum, R.; Bally, T. ChemPhysChem 2006, 7, 1268–1275.
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
(34) The conversion of A into B was also observed during irradiations at 530 nm, but the reaction is
much more efficient at 500 nm.
(35) Nunes, C. M.; Reva, I.; Pinho e Melo, T. M. V. D.; Fausto, R. J. Org. Chem. 2012, 77, 8723–8732.
(36) Dunkin, I. R.; Lynch, M. A.; Withnall, R.; Boulton, A. J.; Henderson, N. J. Chem. Soc. Chem.
Commun. 1989, 1777.
(37) Breda, S.; Lapinski, L.; Reva, I.; Fausto, R. J. Photochem. Photobiol. A Chem. 2004, 162, 139–151.
(38) Tomioka, H.; Ichikawa, N.; Komatsu, K. J. Am. Chem. Soc. 1993, 115, 8621–8626.
(39) Chapman, O. L.; Le Roux, J.ꢀP. J. Am. Chem. Soc. 1978, 100, 282.
(40) Dunkin, I. R.; Thomson, P. C. P. J. Chem. Soc. Chem. Commun. 1980, 499–501.
(41) Donnelly, T.; Dunkin, I. R.; Norwood, D. S. D.; Prentice, A.; Shields, C. J.; Thomson, P. C. P. J.
Chem. Soc. Perkin Trans. 1985, 307–310.
(42) Bally et al., have developed a protocol of irradiations allowing the selective photoconversion
between the three isomeric species in cryogenic matrices: 2,6ꢀdifluorophenylnitrene, the corresponding
benzazirine and cyclic ketenimine. The approach was based on their different UVꢀVis absorption
spectra (see ref. 33a): triplet arylnitrenes have a characteristic weak feature that extends to 500 nm or
more, benzazirines have a broad feature around 300 nm, and the corresponding cyclic ketenimines have
a broad band around 400 nm. Similar protocol was applied by other authors (See refs. 15, 16, 42a, 42b).
3
In the present case, the study of photochemical interconversions between triplet arylnitrene 2,
benzazirine 5, and cyclic ketenimine 6 might be hampered by an additional photochemical pathway
3
from nitrene 2 to iminoꢀketene 3 (which is also photoreactive) and the existence of two tunneling
reactions. (42a) Sander, W.; Winkler, M.; Cakir, B.; Grote, D.; Bettinger, H. F. J. Org. Chem. 2007, 72,
715–724; (42b) Grote, D.; Finke, C.; Neuhaus, P.; Sander, W. Eur. J. Org. Chem. 2012, 3229–3236.
(43) The IR light source (ETC EverGlo) of our FTIR spectrometer, running in normal operating mode,
26
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 26 of 29
produces radiation essentially only in the 7400 to 50 cm−1 region. This clearly rules out the possibility of
the rearrangement of 5 to 6 being promoted by an electronic transition, when the sample is kept under
“unfiltered IR light” conditions. Actually, if any radiation capable of inducing electronic excitation
reached the sample under “unfiltered IR light” conditions, one would expect the observation of 3 to 4
rearrangement. As described, the rearrangement of 3 to 4 is induced by irradiation at 500 nm, and at this
wavelength species 5 does not absorb (TDDFT calculation estimates the first electronic transition of 5 at
λ = 332 nm [f=0.0093]).
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
(44) Examples of isomerization reactions in matrices that consist of two components, tunneling and
IRꢀinduced conversion (which may compete), have been previously described. See refs. 30 and 44aꢀc.
(44a) Reva, I.; Nunes, C. M.; Biczysko, M.; Fausto, R. J. Phys. Chem. A 2015, 119, 2614–2627.
(44b) Lapinski, L.; Reva, I.; Rostkowska, H.; Fausto, R.; Nowak, M. J. J. Phys. Chem. B 2014, 118,
2831–2841. (44c) Reva, I.; Nowak, M. J.; Lapinski, L.; Fausto, R. J. Chem. Phys. 2012, 136, 64511.
(45) Some of us have been working on the predictions of heavyꢀatom tunneling using SCT
computations. For some of the most recent works see refs. 31 and 45aꢀb. (a) Kozuch, S. Phys. Chem.
Chem. Phys. 2014, 16, 7718–7727. (b) Kozuch, S.; Zhang, X.; Hrovat, D. A; Borden, W. T. J. Am.
Chem. Soc. 2013, 135, 17274–17277.
(46) Zhang, X.; Datta, A.; Hrovat, D. A.; Borden, W. T. J. Am. Chem. Soc. 2009, 131, 16002–16003.
(47) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.;
Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.;
Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.;
Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.;
Vreven, T.; Montgomery, Jr., J. A.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.;
Kudin, K. N.; Staroverov, V. N.; Keith, T.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.;
Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.;
Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin,
A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth,
G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, Ö.; Foresman, J. B.; Ortiz,
J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09, Revision D.01; Gaussian, Inc.: Wallingford, CT, 2013.
(48) Nunes, C. M.; Reva, I.; Fausto, R.; Begue, D.; Wentrup, C. Chem. Commun. 2015, 51, 14712–
14715.
(49) Zhurko, G. A. ChemCraft, Version 1.8. http://www.chemcraftprog.com (2016). Last accessed on
November 3, 2017.
27
ACS Paragon Plus Environment

Page 27 of 29
Journal of the American Chemical Society
(50) Tenꢀno, S.; Noga, J. WIREs Comput. Mol. Sci. 2012, 2, 114–125.
(51) Zhao, Y.; Truhlar, D. G. Theor. Chem. Acc. 2008, 120, 215–241.
(52) Peterson, K. A.; Adler, T. B.; Werner, H. J. J. Chem. Phys. 2008, 128, 84102.
(53) Neese, F. WIREs Comput. Mol. Sci. 2012, 2, 73–78.
1
2
3
4
5
6
7
8
9
(54) FernandezꢀRamos, A.; Ellingson, B. A.; Garrett, B. C.; Truhlar, D. G. In Rev. Comput. Chem.;
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
Lipkowitz, K. B., Cundari, T. R., Eds.; John Wiley & Sons, 2007; Vol. 3, pp 125–232.
(55) Zheng, J.; Bao, J. L.; MeanaꢀPañeda, R.; Zhang, S.; Lynch, B. J.; Corchado, J. C.; Chuang, Y.ꢀY.;
Fast, P. L.; Hu, W.ꢀP.; Liu, Y.ꢀP.; Lynch, G. C.; Nguyen, K. A.; Jackels, C. F.; Fernandez Ramos, A.;
Ellingson, B. A.; Melissas, V. S.; Villà, J.; Rossi, I.; Coitiño, E. L.; Pu, J.; Albu, T. V.; Ratkiewicz, A.;
Steckler, R.; Garrett, B. C.; Isaacson, A. D.; Truhlar, D. G. Polyrate–version 2016ꢀ2A, University of
Minnesota, Minneapolis, 2016.
(56) Zheng, J.; Zhang, S.; Corchado, J. C.; MeanaꢀPañeda; Chuang, Y.ꢀY.; Coitiño, E. L.; Ellingson, B.
A.; Truhlar, D. G.; Gaussrate 2016, Department of Chemistry and Supercomputing Institute, University
of Minnesota, Minneapolis, Minnesota 55455.
28
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 28 of 29
1
2
3
4
Graphical TOC image
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
29
ACS Paragon Plus Environment