Vinylnitrene formation from 3,5-diphenyl-isoxazole and 3-benzoyl-2-phenylazirine

Article abstract of DOI:10.1021/jo401819g

Photolysis of 1 in argon-saturated acetonitrile yields 2, whereas in oxygen-saturated acetonitrile small amounts of benzoic acid and benzamide are formed in addition to 2. Similarly, photolysis of 2 in argon-saturated acetonitrile results in 1 and a trace amount of 3, whereas in oxygen-saturated acetonitrile the major product is 1 in addition to the formation of small amounts of benzoic acid and benzamide. Laser flash photolysis of 1 results in an absorption due to triplet vinylnitrene 4 (broad absorption with λ_max at 360 nm, τ = 1.8 μs, acetonitrile) that is formed with a rate constant of 1.2 × 10⁷ s^-1 and decays with a rate constant of 5.6 × 10⁵ s^-1. Laser flash photolysis of 2 in argon-saturated acetonitrile likewise results in the formation of triplet vinylnitrene 4 but also ylide 5 (λ_max at 440 nm, τ = 13 μs). The rate constant for forming 4 in argon-saturated acetonitrile is 1.6 × 10⁷ s^-1. In oxygen-saturated acetonitrile, vinylnitrene 4 reacts to form the peroxide radical 6 (λ_max 360 nm, ～0.7 μs, acetonitrile) at a rate of 2 × 10⁹ M^-1s^-1. Density functional theory calculations were performed to aid in the characterization of vinylnitrene 4 and peroxide 6 and to support the proposed mechanism for the formation of these intermediates.

Full text of DOI:10.1021/jo401819g

Article
pubs.acs.org/joc
Vinylnitrene Formation from 3,5-Diphenyl-isoxazole and 3‑Benzoyl-
2-phenylazirine
Disnani W. Gamage, Qian Li, R. A. A. Upul Ranaweera, Sujan K. Sarkar, Geethika K. Weragoda,
Patrick L. Carr, and Anna D. Gudmundsdottir*
Department of Chemistry, University of Cincinnati, Cincinnati Ohio 45221-0172, United States
S
* Supporting Information
ABSTRACT: Photolysis of 1 in argon-saturated acetonitrile yields 2,
whereas in oxygen-saturated acetonitrile small amounts of benzoic acid
and benzamide are formed in addition to 2. Similarly, photolysis of 2 in
argon-saturated acetonitrile results in 1 and a trace amount of 3, whereas
in oxygen-saturated acetonitrile the major product is 1 in addition to the
formation of small amounts of benzoic acid and benzamide. Laser flash
photolysis of 1 results in an absorption due to triplet vinylnitrene 4 (broad absorption with λmax at 360 nm, τ = 1.8 μs,
acetonitrile) that is formed with a rate constant of 1.2 × 107 s−1 and decays with a rate constant of 5.6 × 105 s−1. Laser flash
photolysis of 2 in argon-saturated acetonitrile likewise results in the formation of triplet vinylnitrene 4 but also ylide 5 (λmax at
440 nm, τ = 13 μs). The rate constant for forming 4 in argon-saturated acetonitrile is 1.6 × 107 s−1. In oxygen-saturated
acetonitrile, vinylnitrene 4 reacts to form the peroxide radical 6 (λmax 360 nm, ∼0.7 μs, acetonitrile) at a rate of 2 × 109 M−1 s−1.
Density functional theory calculations were performed to aid in the characterization of vinylnitrene 4 and peroxide 6 and to
support the proposed mechanism for the formation of these intermediates.
Scheme 1
1. INTRODUCTION
Nitrenes are monovalent electron-deficient intermediates that
have either a triplet or singlet configuration.1−6 Singlet nitrenes
are generally highly reactive, short-lived intermediates that
primarily decay by unimolecular rearrangement. However,
some of the more stable singlet nitrenes are sufficiently long-
lived to undergo bimolecular reactions, and because they can
efficiently insert into neighboring molecules, they have been
used in various applications, such as surface modification,7
cross-polymerization8 and photoaffinity labeling.9 In compar-
ison, triplet nitrenes are generally less reactive and are thus
longer-lived intermediates because they require intersystem
crossing to form products. Due to their high spin properties,
triplet nitrenes have the potential to be building blocks for
high-spin assemblies.10 It is not only the spin state of nitrenes
that affects their reactivity. In addition, electron-donating
substituents, such as aryl- and alkyl groups, stabilize the nitrene
intermediates further, whereas electron-withdrawing substitu-
ents destabilize nitrenes. For example, triplet aryl- and
alkylnitrenes are long-lived intermediates with lifetimes on
the order of milliseconds because they do not react with the
solvent but rather decay by dimerization.11−14 The measured
rate at which triplet alkylnitrenes react with molecular oxygen is
∼5 × 104 M−1 s−1,11 and the rate for trapping triplet
phenylnitrenes with oxygen has been determined to be between
5 × 104 and 8 × 106 M−1 s−1,15−17 which is considerably slower
than diffusion, i.e., the rate at which most C-centered radicals
react with oxygen.
Although this triplet vinylnitrene has an electron-donating
vinyl substituent, it has a lifetime of only a few microseconds in
solution and is efficiently quenched with molecular oxygen at a
rate of 7 × 108 M−1 s−1.18 Spin density calculations show that
the triplet vinylnitrene has significant 1,3 biradical character
with substantial spin density on the β C-atom, which explains
why this vinylnitrene reacts efficiently with molecular oxygen at
rates that are comparable to those generally observed for C-
atom centered radicals. We were, however, not able to isolate
any photoproducts that result from trapping the triplet
vinylnitrene with oxygen, whereas Murata et al. have previously
reported that the vinylnitrene depicted in Scheme 2 can be
trapped with molecular oxygen to form stable compounds.19
Scheme 2
Recently, we reported the first direct detection of a triplet
vinylnitrene in solution by performing laser flash photolysis of
(3-methyl-2H-azirin-2-yl)phenylmethanone (Scheme 1).18
Received: August 29, 2013
Published: October 15, 2013
© 2013 American Chemical Society
11349
dx.doi.org/10.1021/jo401819g | J. Org. Chem. 2013, 78, 11349−11356

The Journal of Organic Chemistry
Article
In this study, we used laser flash photolysis to demonstrate
that both isoxazole 1 and azirine 2 can serve as precursors for
triplet vinylnitrene 4. As expected, vinylnitrene 4 is short-lived
and decays by intersystem crossing to reform 1 and 2. In
oxygen-saturated acetonitrile, triplet vinylnitrene 4 is inter-
cepted by oxygen to form peroxide radical 6, which can be
observed directly with laser flash photolysis. The rate for the
reaction of vinylnitrene 4 with oxygen was estimated to be 2.1
× 109 M−1 s−1. Thus, in this paper, we have further
demonstrated that triplet vinylnitrenes show 1,3-biradical
character, which is shown by their efficient reactivity with
molecular oxygen.
Scheme 5. Proposed Mechanism for the Reaction of 1 and 2
on Their Triplet Energy Surface
2. RESULTS
2.1. Product Studies. Irradiation of 1 in argon-saturated
CDCl3, through a Pyrex filter, resulted in the formation of 2 as
the only photoproduct (Scheme 3), whereas photolysis of 2
Scheme 6. Proposed Mechanism for Trapping Vinylnitrene 4
with Oxygen
Scheme 3. Photolysis of 1 in Argon- and Oxygen-Saturated
CDCl3
Scheme 7. Proposed Mechanism Showing the Singlet
Photoreactivity of 2
through a Pyrex filter in argon-saturated CDCl3 forms 1 and a
trace amount of 3 at 55% conversion (Scheme 4). The results
agree with those previously reported by Singh et al.20
In comparison, irradiation of 1 in oxygen-saturated CDCl3,
through a Pyrex filter, also resulted in formation of 2, benzoic
acid and benzamide in a ratio of 3:1:1 with 37% conversion
(Scheme 3). Similarly, photolysis of 2 in oxygen-saturated
CDCl3 results in formation of 1, benzoic acid, benzamide in the
ratio 8:1:1 as well as a trace amount of 3 with 54% conversion
(Scheme 4).
Thus, we theorize that upon irradiation, both 1 and 2 form
triplet excited states, which undergo C−N bond cleavage to
form triplet vinylnitrene 4 (Scheme 5). Triplet vinylnitrene 4
must decay by intersystem crossing and reforms both 1 and 2 in
argon-saturated CDCl3, whereas in oxygen-saturated CDCl3,
vinylnitrene 4 is intercepted by oxygen to form benzoic acid
and benzamide (Scheme 6).
Singh et al. proposed that 3 is formed by cleavage of the C−
C bond of the azirine moiety in 2 to form ylide 5 (Scheme 7).20
There are several examples in the literature demonstrating that
the photolysis of azirine derivatives yields ylides such as 521−25
and that these reactions proceed on the singlet surface of the
azirine.26 Thus, azirine 2 is a precursor to triplet vinylnitrene 4,
but 2 also undergoes singlet reactivity to form 3.
triplet surfaces of 1 and 2 using the B3LYP level of theory with
the 6-31G+(d) basis as implemented on Gaussian09.27−29
We compared the triplet excited states of 1 and 2. Time-
dependent density functional theory (TD-DFT) calculations
located the first excited singlet state (S1) of 1 at 103 kcal/mol
above the S0 of 1, whereas the first and second excited triplet
state (T1, T2) of 1 were located 65 and 76 kcal/mol above the
S0 of 1 (Figure 5), respectively.
We also optimized the T1 of 1 and found that it is located 63
kcal/mol above the S0. The energies for the T1 of 1 obtained
from TD-DFT and optimization calculations are in good
agreement, presumably because the geometries of T1 and S0 of
1 are fairly similar and therefore the vertical excitation and
equilibrium energies are comparable. In the optimized structure
of the T1 of 1, the O−C and the N−C bonds are 1.42 and 1.42
Å and are thus somewhat longer than the corresponding bonds
in the S0 of 1, which are 1.36 and 1.32 Å, respectively. In
addition, the CC bonds in the phenyl ring adjacent to the
oxygen atom in the T1 of 1 are no longer equivalent, as in the
S0 of 1. Thus, T1 is best described as a triplet excited state that
is delocalized throughout the isoxazole moiety and the phenyl
group adjacent to the oxygen atom. Spin density calculations
2.2. Calculations. To better compare the photoreactivities
of 1 and 2, we calculated stationary points on the singlet and
Scheme 4. Photolysis of 2 in Argon- and Oxygen-Saturated CDCl3
11350
dx.doi.org/10.1021/jo401819g | J. Org. Chem. 2013, 78, 11349−11356

The Journal of Organic Chemistry
Article
further support this notion, as they show that the unpaired spin
is mainly localized on the isoxazole moiety and the phenyl
group adjacent to the oxygen atom (Figure 1, Scheme 8).
Figure 2. Selected bond distances (Å) in the optimized structure of 2
and T1K of 2 and TA of 2.
Two minimal energy conformers were calculated for triplet
vinylnitrene 4, A and B, and A was slightly more stable, by ∼0.7
kcal/mol (Figure 3). The calculated spin density of 4
Figure 1. Selected bond distances (Å) in the optimized structure of 1
and the T1 of 1.
TD-DFT calculations of the optimized structure of 2 place
the first and second excited triplet state of the ketone (T1K and
T2K) of 2 at 68 and 73 kcal/mol, respectively, above its S0, or 82
and 87 kcal/mol with respect to the S0 of 1 (Figure 5).
Inspection of the molecular orbitals indicates that the T1K of 2
has an (n,π*) configuration, whereas the T2K of 2 has a (π,π*)
configuration.
The optimized structure of the T1K of 2 is located 65 kcal/
mol above the S0. The progression of the CO bond in the
T1K of 2 to 1.31 Å from 1.23 Å in the S0 of 2 indicates that this
triplet state has an (n,π*) configuration, which is in an
agreement with the TD-DFT calculations. The energy obtained
for the optimized structure of the T1K of 2 is only slightly less
than the energy obtained from the TD-DFT calculations,
although the geometries of T1K and S0 of 2 are different, thus
indicating that the potential energy surface of T1K of 2 must be
relatively flat. We have, however, previously shown that B3LYP
calculations underestimate the energy of triplet ketones with an
(n,π*) configuration.30 Spin density calculations further
support this because the spin density is primarily localized on
the oxygen atom in the CO moiety and the adjacent phenyl
ring (Scheme 8).
We also optimized the structure of the TA of 2 and found
that the energy level is located 63 kcal/mol above the S0 of 2. In
the TA of 2, the CO bond is 1.22 Å, which is similar to the
CO bond in the S0 of 2, 1.23 Å (Figure 2), whereas the C−C
bond in the azirine ring is shorter than in the S0 of 2.
Furthermore, the CN bond in the TA of 2 progressed to 1.42
Å compared to 1.26 Å in the S0 of 2, whereas the C−C bond in
the azirine ring is condensed to 1.37 Å in the TA of 2 compared
to 1.45 Å in the S0 of 2. The calculated spin density of the TA of
2 is shown in Scheme 8, which shows that the unpaired
electrons are mainly localized on the azirine ring and the
adjacent phenyl ring. Thus, the calculations show that the TA of
2 is localized on the azirine moiety and the phenyl group that is
conjugated with the azirine (Figure 2).
Figure 3. Relative energy of the minimal energy conformers A and B
of vinylnitrene 4 and selected bond lengths (Å).
demonstrates that the triplet vinylnitrene 4 has considerable
1,3-carbon iminyl biradical character, as the spin density is
mainly located on the N and the β-C atoms (Scheme 8). The
biradical character of vinylnitrene 4 can be further highlighted
by calculating the rotational barrier between conformer 4A and
4B, which is ∼7.5 kcal/mol and thus significantly lower than for
a simple CC bond (Figure 4).
The stationary points on the triplet surface of 1 and 2,
relative to S0 of 1, to form vinylnitrene 4 are shown in Figure 5.
The calculations show that vinylnitrene 4 can be formed from
both the T1K of 2 and the T1 of 1, but the transition state
barrier for forming vinylnitrene 4 is considerably larger for the
T1 of 1 than for the T1K of 2. However, it is possible that
vinylnitrene 4 is formed from the T2 of 1. Direct irradiation of
the S1 of 1 can result in intersystem crossing to the T2 of 1
rather than to the T1 of 1 because the S1 and T2 of 1 are closer
in energy.
The transition state energy for forming ylide 5 was calculated
to be 44 kcal/mol above the S0 of 2. IRC calculations correlate
this transition state with the S0 of 2 and ylide 5 (Figure 6).
Scheme 8. Calculated Spin Densities for the T1 of 1, T1K of 2, TA of 2 and 4
11351
dx.doi.org/10.1021/jo401819g | J. Org. Chem. 2013, 78, 11349−11356

The Journal of Organic Chemistry
Article
Figure 4. Calculated rotational barriers between vinylnitrene conformers 4A and 4B.
Figure 5. Calculated stationary points for the triplet surface of 1 and 2. The calculated energies are in kcal/mol and are relative to S0 of 1. The
energies of the S1 and T2, T1 (65) of 1, S1, T2K and T1K (82) of 2 were obtained from the TD-DFT calculations, whereas the energies of the
transition states (TS) and the S0 and T1 (63) of 1 and the S0, T1K (79), and TA of 2 were obtained by optimization.
Figure 6. Calculated stationary points for the singlet surface of 2. The calculated energies are in kcal/mol. The energy of the S1 of 2 was obtained by
TD-DFT calculations, whereas the energies of the transition states (TS) S0 of 2 and S0 of 2 and 3 were obtained by optimization.
We also explored the reactivity of vinylnitrene 4 with oxygen
to form radicals 6 and 7 (Figure 7). The calculated triplet
transition state for the addition of oxygen to 4 to form 6 is
located 6 kcal/mol above 4 and that of an oxygen molecule.
Radical 6 is 7 kcal/mol more stable than vinylnitrene 4 and
oxygen. In comparison, the transition state to form 7 is located
16 kcal/mol above 4, and that of an oxygen molecule and
radical 7 is 5 kcal/mol less stable than 4 and an oxygen
molecule. Thus, the calculations show that intercepting
vinylnitrene 4 with oxygen to form 6 is feasible at ambient
temperature and that the formation of 6 is strongly favored
over 7 (Figure 7).
2.3. Phosphorescence. The phosphorescence spectra of 1
and 2 were obtained on a phosphorometer in frozen ethanol
matrices at 77 K (Figure 8). Because the emission from 1 and 2
is not sufficiently resolved to locate the (0,0) transitions, we
used the onset of the emission at the shortest wavelength to
estimate the (0,0) transitions for the T1 of 1 and T1K of 2. The
estimated onset for phosphorescence for 1 is 471 nm, which
corresponds to the T1 of 1 being 61 kcal/mol above the S0 of 1
11352
dx.doi.org/10.1021/jo401819g | J. Org. Chem. 2013, 78, 11349−11356

The Journal of Organic Chemistry
Article
Figure 7. Energy diagram for forming triplet peroxide radicals 6 and 7. (Energies are in kcal/mol.)
Figure 8. Phosphorescence spectra of 1 and 2 in ethanol glasses at 77
K obtained with excitation at 340 nm.
(Table 1). In comparison, the estimated onset of phosphor-
escence of 2 is observed at ∼400 nm; thus, the T1K of 2 is ∼71
Table 1. Energies of the T1 of 1 and the T1K of 2
phosphorescence
Figure 9. Transient spectra obtained from laser flash photolysis of 1 in
(A) argon-saturated acetonitrile and (B) argon-saturated and (C)
oxygen-saturated methanol.
TD-
a
b
cmpd
wavelength (nm) energy (kcal/mol)
DFT
optimized
T1 of 1
471
400
61
71
65
68
63
65
T1K of 2
oxygen-saturated methanol and acetonitrile, and this resulted in
a broad transient spectrum with a λmax at 360 nm. This transient
spectrum does show absorption above 480 nm and therefore
does not correlate to either the T1 of 1 or vinylnitrene 4. Thus,
we theorized that in oxygen-saturated solution the transient
absorption results from the reaction of vinylnitrene 4 with
oxygen to form peroxyradical 6, because the observed spectrum
matches the calculated spectrum of 6 (Figures 9C and 10C).
Kinetic studies further support the assignments of the
transient spectra obtained from the laser flash photolysis of 1.
At 360 nm, the absorption is formed with a rate constant of 1.2
× 107 s−1 (τ = 83 ns) in argon-saturated acetonitrile, whereas
the decay of the transient absorption is best fitted as an
exponential function, which yields a decay rate constant of 5.5
× 105 s−1. Thus, vinylnitrene 4 is formed at a rate of 1.2 × 107
s−1 from its precursor and has a lifetime of 1.8 μs. It should be
highlighted that the transient absorption at 460 nm forms and
decays with the same rate constants as the absorption at 360
nm in argon-saturated solution. In oxygen-saturated acetonitrile
(Figure 11), the transient absorption is formed at a rate of 1.65
× 107 s−1 (τ = 61 ns) and it decays at a rate of 1.35 × 106 s−1 (τ
= 743 ns). Furthermore, the absorption intensity is higher at
360 nm in oxygen-saturated solution compared to argon-
saturated solution. Kinetic analysis at 460 nm shows that in
a
The energies were obtained from TD-DFT (B3LYP/6-31G+(d))
b
calculations. The energies were obtained from optimization (B3LYP/
6-31G+(d)) calculations.
kcal/mol above the S0 of 2. The calculated and measured
energies for the T1 of 1 and the T1K of 2 are in good agreement
(Table 1).
2.4. Laser Flash Photolysis. To directly identify vinyl-
nitrene 4 and its triplet precursors, we performed laser flash
photolysis of 1 and 2 (308 nm31). Laser flash photolysis of 1 in
argon-saturated acetonitrile or methanol resulted in a broad
absorption with a λmax of ∼360 nm that trails off to 550 nm
(Figure 9). The calculated spectrum of the T1 of 1 in
acetonitrile has major electronic transitions at 531 (f = 0.147),
512 (f = 0.104), 453 (f = 0.074), 392 (f = 0.057), and 333 (f =
0.31) nm (Figure 10), which agrees with the observed
spectrum. Likewise, the calculated spectra of 4A and 4B in
acetonitrile have major electronic transitions at 344 (f = 0.043),
347 (f = 0.024), 525 (f = 0.029), and 540 (f = 0.036) nm,
which also agree with the observed spectra, even though the
absorption is not as intense at longer wavelengths as the
calculations predict. To further aid in the assignment of this
transient absorption, we performed laser flash photolysis of 1 in
11353
dx.doi.org/10.1021/jo401819g | J. Org. Chem. 2013, 78, 11349−11356

The Journal of Organic Chemistry
Article
Figure 12. Transient spectra obtained from laser flash photolysis of 2
in (A) argon-saturated and (B) oxygen-saturated acetonitrile.
vinylnitrene 4 and ylide 5 based on the following. The observed
transient spectra was similar to the one obtained from the laser
flash photolysis of 1 in argon-saturated solution, thus
supporting the theory that transient absorption is due to
vinylnitrene 4. Furthermore, the calculated absorption spec-
trum of the TK of 2 does not match the observed spectra. In
oxygen-saturated acetonitrile, laser flash photolysis of 2 resulted
in a transient spectrum that is similar to the one obtained by
laser flash photolysis of 1 in oxygen-saturated acetonitrile, but
there is an additional absorption band at 440 nm that we assign
to ylide 5 in addition to the absorption due to formation of
peroxide radical 6. The major calculated transition for ylide 5 is
located at 394 nm (f = 1.04), which fits reasonably with the
band observed at 440 nm.
Figure 10. Calculated electronic transitions in acetonitrile for T1 of
(A) 1, (B) 4A and 4B, (C) 6A and 6B; TK of (D) 2 and (E) ylide 5.
As before, kinetic analysis further supported these absorption
assignments (Figure S1 in the Supporting Information [SI]).
The absorption of vinylnitrene 4 at 360 nm is formed with a
rate constant of 2.0 × 107 s−1 (τ = 50 ns), and it decays with a
rate constant of 6.0 × 105 s−1 (τ = 1.7 μs) in argon-saturated
acetonitrile. As expected, vinylnitrene 4 is formed with a
different rate constant from the excited states of 1 and 2,
whereas the lifetime of vinylnitrene 4 is the same, within
experimental error. The transient absorption at 440 nm can be
fitted as biexponential decay with rate constants of 6.5 × 105 s−1
(τ = 1.5 μs) and 7.68 × 105 s−1 (τ = 13 μs), with the shorter-
lived component corresponding to vinylnitrene 4 and the
longer-lived component to ylide 5. Ylide 5 has a lifetime similar
to those of other ylides that decay by intramolecular reaction,
but it is much shorter lived than those of ylides that decay by
bimolecular reactions.18,36 The transient absorption at 500 nm
forms and decays with the same rate constants as the
absorption at 360 nm.
Figure 11. Kinetic traces obtained at (A) 360 nm and (B) 460 nm for
the laser flash photolysis of 1 in argon-, air- and oxygen-saturated
acetonitrile.
In addition, kinetic analysis in oxygen-saturated acetonitrile
shows that the transient absorption at 360 nm is formed with a
rate constant of 1.63 × 107 s−1 (τ = 61 ns) and decays with a
rate constant of 1.40 × 106 s−1 (τ = 714 ns), which fits nicely
with the rate constants measured from the reactivity of 1.
Analysis of the transient kinetics at 500 nm shows that, in
oxygen-saturated solution, the absorption is fully quenched.
oxygen-saturated solution, the absorption is quenched and not
enhanced, as it is at 360 nm, thus further supporting that a
different transient is formed in oxygen-saturated solutions
compared to argon-saturated solutions. Therefore, we conclude
that triplet vinylnitrene 4 reacts with oxygen with a rate
constant of 1.65 × 107 s−1, and because the concentration of
oxygen in saturated acetonitrile is 0.009 M,32−34 the rate for
oxygen quenching of vinylnitrene 4 is 1.8 × 109 M−1s−1, which
is similar to what we have previously measured for another
vinylnitrene derivative.35
3. DISCUSSION
Laser flash photolysis has demonstrated that both 1 and 2 yield
vinylnitrene 4. Although the calculated transition state barrier
for forming vinylnitrene 4 from the T1 of 1 is considerably
higher than that for the T1K of 2, the rate of forming
The laser flash photolysis of 2 in argon-saturated acetonitrile
produced a transient spectrum showing a broad absorption with
the λmax at 360 nm (Figure 12). We assign this absorption to
11354
dx.doi.org/10.1021/jo401819g | J. Org. Chem. 2013, 78, 11349−11356

The Journal of Organic Chemistry
Article
vinylnitrene 4 from these two compounds is on the same order.
Thus, we theorize that the S1 of 1 does not intersystem cross to
the T1 of 1 but rather to a higher-lying triplet excited state that
is closer in energy to the S1 of 1 and then cleaves to form
vinylnitrene 4. Thus, both isoxazole and azirine are effective
precursors for forming triplet vinylnitrene intermediate 4.
Azirine 2 is, however, somewhat less selective than 1 because
excitation of the azirine chromophore in 2 leads to formation of
ylide 5, whereas excitation of the ketone chromophore in 2
yields the S1K of 2 that will intersystem cross to form the T1K of
2. The conjugation of the phenyl ring with the azirine moiety
lowers the triplet energy of the TA of 2 significantly, and its
triplet energy is comparable to the T1K of 2. However, the TD-
DFT calculations of the S0 of 2 show that the main absorption
above 300 nm is due to the ketone chromophore; thus, the
main reactivity observed is the formation of triplet vinylnitrene
4. Furthermore, 3-phenyl-2H-azirine (8) has been shown to
undergo singlet reactivity upon direct irradiation; thus, it is not
likely that the singlet excited state of the azirine moiety in 2
intersystem crosses to form the TA of 2 (Scheme 9).36
than those for 10 due to the involvement of the additional
phenyl chromophore. The rates of vinylnitrenes 4 and 10 when
reacting with oxygen are efficient, 7 × 108 and 2 × 109 M−1s−1,
respectively, because the oxygen adds to the β-carbon atom in
the vinylnitrenes and not the nitrogen atom, as in triplet alkyl-
and phenylnitrenes.
4. CONCLUSION
We have shown that azirine 2 and isoxazole 1 can both serve as
precursors to triplet vinylnitrene 4, which is a short-lived
intermediate. Spin density calculations show that triplet
vinylnitrene 4 has significant spin density on the β-carbon
atom, thus demonstrating that vinylnitrene 4 has significant 1,3-
biradical character. Therefore, it reacts efficiently with oxygen
to form peroxide radical 6.
EXPERIMENTAL SECTION
■
Calculation. All the geometries were optimized at the B3LYP level
of theory and with the 6-31G+(d) basis set as implemented in the
Gaussian03/09 programs.28,29,39 All transition states were confirmed to
have one imaginary vibrational frequency by analytical determination
of the second derivatives of the energy with respect to internal
coordinates. Intrinsic reaction coordinate (IRC) calculations were used
to verify that the optimized transition states corresponded to the
attributed reactant and product.40,41 The absorption spectra were
calculated using time-dependent density functional theory (TD-
DFT).42−45 The effect of solvation was calculated using the self-
consistent reaction field (SCRF) method with the integral equation
formalism polarization continuum model (IEFPCM) with acetonitrile
as the solvent.46−49 All calculations were performed at the Ohio
Supercomputer Center.
Scheme 9
Laser Flash Photolysis. Laser flash photolysis was performed with
an Excimer laser (308 nm, 17 ns).31 A stock solution of 1 or 2 in
acetonitrile or methanol was prepared with spectroscopic-grade
acetonitrile or methanol, so that the solutions had absorption between
0.6 and 0.4 at 308 nm. Typically, 3 mL of the stock solution was placed
in a 10 mm × 10 mm wide, 48 mm long quartz cuvette and it was
purged with argon or oxygen for 5 min. The rates were obtained by
fitting an average of three kinetic traces. Transient absorption spectra
were obtained by plotting average absorbance values collected from
decays at 10 or 20 nm intervals between 300 and 600 nm.
The transient spectrum of vinylnitrene 4 is somewhat similar
to the transient spectrum of vinylnitrene 10 (Scheme 9); they
both have broad absorptions with a λmax at 360 nm that trail out
to 500 nm.18 The absorption in vinylnitrenes 4 and 10 are due
to several electronic transitions between the lone pair on the
oxygen, the half-filled p-orbital on the N-atom and the π-
orbitals into the half-filled p-orbital on the N-atom, and the π*-
orbitals. Vinylnitrenes 4 and 10 also have similar lifetimes of a
few microseconds because they both decay by efficient
intersystem crossing to their singlet states. Presumably, the
intersystem crossing is efficient because the triplet vinylnitrenes
intersystem cross to their singlet vinylnitrenes, which have open
shell configurations.37,38
The calculated spin densities of vinylnitrenes 4 and 10 are
similar. The spin density is mainly located on the N-atom and
the β-C atom; thus, the increased conjugation of the
vinylnitrene chromophore to the additional phenyl group in
4 does not affect the spin density. The most significant
difference between triplet vinylnitrenes 4 and 10 is that
vinylnitrene 4 reacts with oxygen to form stable products,
benzoic acid and benzamide, whereas 10 did not yield any new
products in oxygen-saturated solutions (Scheme 9). It can be
theorized that in oxygen-saturated solution, the triplet ketone in
9 is efficiently quenched; thus, no new products were observed
from attempts to trap the vinylnitrene 10 with oxygen, or the
products were not stable enough to be isolated. More
importantly, it was possible to detect the peroxide radical 6
directly by performing laser flash photolysis of 1 and 2 in
oxygen-saturated acetonitrile, thereby experimentally verifying
that vinylnitrene 4 has significant 1,3-biradical character, as
oxygen adds to the β-C atom. In addition, the TD-DFT
calculations demonstrate that the intensities of the calculated
electronic transitions for vinylnitrene 4 are more significant
Preparation of 1. Dibenzoylmethane (2.24 g, 10 mmol) and
NH2OH·HCl (0.725 g, 10.5 mmol) in absolute ethanol (25 mL) were
refluxed for 3 h. The mixture was cooled and diluted with water until
3,5-diphenyl isoxazole precipitated. The precipitate was washed with
water and dried in a vacuum desiccator and recrystallized from diethyl
1
ether. The H NMR spectrum of 1 was identical to that previously
reported.50 1H NMR (400 MHz, CDCl3): δ 7.84−7.89 (m, 4H),
7.46−7.51 (m, 6H), 6.84 (s, 1H) ppm. GC−MS m/z (EI) 221 (M+,
30), 193, 165, 144, 116, 105 (100), 89, 77, 63, 51. IR (neat, νmax):
3114, 1489, 1457, 1046, 951, 914, 820, 763, 689 cm−1.
Preparation of 2. 3,5-Dibenzoyl isoxazole (500 mg, 2.2 mmol) in
argon-saturated CHCl3 was irradiated through a Pyrex filter until GC−
MS analysis of the reaction mixture showed 30% formation of 2 (30
mg, 0.14 mmol, 6% yields) and 70% of remaining starting material.
Azirine 2 was separated from the starting material using a silica column
1
eluted with 5% ethyl acetate in hexane. The H NMR spectrum of 2
matches the one reported in the literature.50 1H NMR (400 MHz,
CDCl3): δ 8.13−8.15 (d, J = 8 Hz, 2H), 7.87−7.89 (d, J = 8 Hz, 2H),
7.61−7.66 (m, 2H), 7.53−7.58 (m, 4H), 3.86 (s, 1H) ppm. GC−MS.
(EI) m/z: 221 (M+, 30), 193, 165, 144, 116, 105 (100), 89, 77 (66),
63, 51. IR (neat, νmax): 3062, 1775, 1672, 1597, 1579, 1449, 1351,
1230, 1009, 721, 688, 653 cm−1.
Photolysis of 1. A solution of 1 (75 mg, 0.33 mmol) in argon-
saturated CDCl3 was irradiated through a Pyrex filter for a day using a
medium-pressure mercury arc lamp. 1H NMR spectrum of the reaction
mixture showed the formation of 2 and remaining starting material in
the ratio of 1:2.
11355
dx.doi.org/10.1021/jo401819g | J. Org. Chem. 2013, 78, 11349−11356

The Journal of Organic Chemistry
Article
A solution of 1 (75 mg, 0.33 mmol) in oxygen-saturated CDCl3
were irradiated through a Pyrex filter for a day using a medium-
pressure mercury arc lamp. 1H NMR spectroscopy and GC−MS
analysis of the reaction mixture showed the formation of 2, benzoic
acid and benzamide in the ratio of 4:1:1 at 40% conversion.
Photolysis of 2. A solutions of 2 (30 mg, 0.13 mmol) in argon-
saturated CDCl3 was irradiated through a Pyrex filter for 24 h using a
medium-pressure mercury arc lamp. 1H NMR spectrum of the reaction
mixture showed the formation of 1 and remaining 2, along with trace
amounts of 3, at 45% conversion.
A solutions of 2 (30 mg, 0.13 mmol) in oxygen-saturated CDCl3
was irradiated through a Pyrex filter (>300 nm) for 24 h using a
medium-pressure mercury arc lamp. 1H NMR spectroscopy and GC−
MS analysis of the reaction mixture showed the formation of 1,
benzoic acid, and benzamide in the ratio of 8:1:1, respectively, along
with trace amount of 3 at 55% conversion.
(19) Inui, H.; Murata, S. J. Am. Chem. Soc. 2005, 127, 2628.
(20) Singh, B.; Zweig, A.; Gallivan, J. B. J. Am. Chem. Soc. 1972, 94,
1199.
(21) Barcus, R. L.; Hadel, L. M.; Johnston, L. J.; Platz, M. S.; Savino,
T. G.; Scaiano, J. C. J. Am. Chem. Soc. 1986, 108, 3928.
(22) Mueller, F.; Mattay, J.; Steenken, S. J. Org. Chem. 1993, 58,
4462.
(23) Orton, E.; Collins, S. T.; Pimentel, G. C. J. Phys. Chem. 1986, 90,
6139.
(24) Barcus, R. L.; Wright, B. B.; Platz, M. S.; Scaiano, J. C.
Tetrahedron Lett. 1983, 24, 3955.
(25) Padwa, A.; Rosenthal, R. J.; Dent, W.; Filho, P.; Turro, N. J.;
Hrovat, D. A.; Gould, I. R. J. Org. Chem. 1984, 49, 3174.
(26) Bornemann, C.; Klessinger, M. Chem. Phys. 2000, 259, 263.
(27) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B: Condens. Matter
1988, 37, 785.
(28) Frisch, M. J.; et al. Gaussian, 03; Gaussian: Wallingford CT,
2004.
(29) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
(30) Muthukrishnan, S.; Mandel, S. M.; Hackett, J. C.; Singh, P. N.
D.; Hadad, C. M.; Krause, J. A.; Gudmundsdottir, A. D. J. Org. Chem.
2007, 72, 2757.
(31) Muthukrishnan, S.; Sankaranarayanan, J.; Klima, R. F.; Pace, T.
C. S.; Bohne, C.; Gudmundsdottir, A. D. Org. Lett. 2009, 11, 2345.
(32) Clark, W. D. K.; Steel, C. J. Am. Chem. Soc. 1971, 93, 6347.
(33) Upul Ranaweera, R. A. A.; Sankaranarayanan, J.; Casey, L.; Ault,
B. S.; Gudmundsdottir, A. D. J. Org. Chem. 2011, 76, 8177.
(34) Muthukrishnan, S.; Sankaranarayanan, J.; Pace, T. C. S.;
Konosonoks, A.; DeMichiei, M. E.; Meese, M. J.; Bohne, C.;
Gudmundsdottir, A. D. J. Org. Chem. 2010, 75, 1393.
(35) Rajam, S.; Murthy, R. S.; Jadhav, A. V.; Li, Q.; Keller, C.; Carra,
C.; Pace, T. C. S.; Bohne, C.; Ault, B. S.; Gudmundsdottir, A. D. J. Org.
Chem. 2011, 76, 9934.
ASSOCIATED CONTENT
* Supporting Information
Cartesian coordinates and energies of 1−7 and NMR spectra of
1 and 2. This material is available free of charge via the Internet
at http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
*Anna.Gudmundsdottir@uc.edu
Notes
■
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
We thank the National Science Foundation (CHE-1057481)
and the Ohio Supercomputer Center for supporting this work.
■
(36) Albrecht, E.; Mattay, J.; Steenken, S. J. Am. Chem. Soc. 1997,
119, 11605.
(37) Nunes, C. M.; Reva, I.; Pinho e Melo, T. M. V. D.; Fausto, R.;
Solomek, T.; Bally, T. J. Am. Chem. Soc. 2011, 133, 18911.
REFERENCES
■
̌
(1) Platz, M. S. In Reactive Intermediate Chemistry; John Wiley &
Sons, Inc.: 2005, p 501.
(2) Gritsan, N. P.; Platz, M. S.; Borden, W. T. Mol. Supramol.
Photochem. 2005, 13, 235.
(3) Gritsan, N. P.; Platz, M. S. Chem. Rev. 2006, 106, 3844.
(4) Platz, M. S. Nitrenes; John Wiley & Sons, Inc., 2004.
(5) Bucher, G., Ed. Photochemical Reactivity of Azides; CRC Press:
Boca Raton, 2004.
(38) Wenthold, P. G. J. Org. Chem. 2012, 77, 208.
(39) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B: Condens. Matter
1988, 37, 785.
(40) Gonzalez, C.; Schlegel, H. B. J. Chem. Phys. 1989, 90, 2154.
(41) Gonzalez, C.; Schlegel, H. B. J. Phys. Chem. 1990, 94, 5523.
(42) Labanowksi, J. K., Andzelm, J. W., Eds. Density Functional
Methods in Chemistry; Springer-Verlag: New York, Weinheim, 1991.
(43) Bauernschmitt, R.; Ahlrichs, R. Chem. Phys. Lett. 1996, 256, 454.
(44) Foresman, J. B.; Head-Gordon, M.; Pople, J. A.; Frisch, M. J. J.
Phys. Chem. 1992, 96, 135.
(45) Stratmann, R. E.; Scuseria, G. E.; Frisch, M. J. J. Chem. Phys.
1998, 109, 8218.
(46) Cances, E.; Mennucci, B. J. Chem. Phys. 2001, 114, 4744.
(47) Cramer, C. J.; Truhlar, D. G. Chem. Rev. 1999, 99, 2161.
(48) Mennucci, B.; Cances, E.; Tomasi, J. J. Phys. Chem. B 1997, 101,
10506.
(6) Platz, M. S. Acc. Chem. Res. 1995, 28, 487.
(7) Mecomber, J. S.; Murthy, R. S.; Rajam, S.; Singh, P. N. D.;
Gudmundsdottir, A. D.; Limbach, P. A. Langmuir 2008, 24, 3645.
(8) Jadhav, A. V.; Gulgas, C. G.; Gudmundsdottir, A. D. Eur. Polym. J.
2007, 43, 2594.
(9) Platz, M. S. Photochem. Photobiol. 1997, 65, 193.
(10) Iwamura, H. Pure Appl. Chem. 1987, 59, 1595.
(11) Singh, P. N. D.; Mandel, S. M.; Sankaranarayanan, J.;
Muthukrishnan, S.; Chang, M.; Robinson, R. M.; Lahti, P. M.; Ault,
B. S.; Gudmundsdottir, A. D. J. Am. Chem. Soc. 2007, 129, 16263.
(12) Sankaranarayanan, J.; Rajam, S.; Hadad, C. M.;
Gudmundsdottir, A. D. J. Phys. Org. Chem. 2010, 23, 370.
(13) Sankaranarayanan, J.; Bort, L. N.; Mandel, S. M.; Chen, P.;
Krause, J. A.; Brooks, E. E.; Tsang, P.; Gudmundsdottir, A. D. Org.
Lett. 2008, 10, 937.
(49) Tomasi, J.; Mennucci, B.; Cammi, R. Chem Rev 2005, 105, 2999.
(50) Renfrow, W. B.; Witte, J. F.; Wolf, R. A.; Bohl, W. R. J. Org.
Chem. 1968, 33, 150.
(14) Muthukrishnan, S.; Ranaweera, R. A. A. U.; Gudmundsdottir, A.
D. In Nitrenes and Nitrenium Ions; Falvey, D. E., Gudmundsdottir, A.
D., Eds.; John Wiley & Sons: Hoboken, NJ, 2013; Vol. 6.
(15) Liang, T. Y.; Schuster, G. B. J. Am. Chem. Soc. 1987, 109, 7803.
(16) Pritchina, E. A.; Gritsan, N. P. J. Photochem. Photobiol., A 1988,
43, 165.
(17) Gritsan, N. P.; Pritchina, E. S. J. Inf. Rec. Mater. 1989, 17, 391.
(18) Rajam, S.; Murthy, R. S.; Jadhav, A. V.; Li, Q.; Keller, C.; Carra,
C.; Pace, T. C. S.; Bohne, C.; Ault, B. S.; Gudmundsdottir, A. D. J. Org.
Chem. 2011, 76, 9934.
11356
dx.doi.org/10.1021/jo401819g | J. Org. Chem. 2013, 78, 11349−11356