Capture of an elusive nitrile ylide as an intermediate in isoxazole-oxazole photoisomerization

Article abstract of DOI:10.1021/jo4015672

The unimolecular photochemistry of 3,5-dimethylisoxazole (1) induced by a narrow-band tunable UV laser was studied using low-temperature matrix isolation coupled with infrared spectroscopy. Monomers of 1 were isolated in argon matrices at 15 K and characterized spectroscopically. Irradiation of matrix-isolated 1 at λ = 222 nm (near its absorption maximum) led to the corresponding 2H-azirine 3 and ketenimine 6 as primary photoproducts and also to nitrile ylide 4 and 2,5-dimethyloxazole (5). The photoproducts were identified (i) by comparison with infrared spectra of authentic matrix-isolated samples of 3 and 5 and (ii) using additional irradiations at longer wavelengths (where 1 does not react) which induce selective photoisomerizations of 4 and 6. In particular, irradiation with λ = 340 nm led to the unequivocal identification of the nitrile ylide anti-4, which was transformed into oxazole 5. The details of the 1,5-electrocyclization of the carbonyl nitrile ylide 4 and its structural nature (propargyl-like versus allene-like geometry) were also characterized using theoretical calculations. Thus, the elusive carbonyl nitrile ylide 4 was captured and characterized for the first time as an intermediate in the isoxazole-oxazole photoisomerization.

Full text of DOI:10.1021/jo4015672

Article
pubs.acs.org/joc
Capture of an Elusive Nitrile Ylide as an Intermediate in Isoxazole−
Oxazole Photoisomerization
Claudio M. Nunes,* Igor Reva,* and Rui Fausto
́
Department of Chemistry, University of Coimbra, P-3004-535 Coimbra, Portugal
S
* Supporting Information
ABSTRACT: The unimolecular photochemistry of 3,5‑dime-
thylisoxazole (1) induced by a narrow-band tunable UV laser
was studied using low-temperature matrix isolation coupled
with infrared spectroscopy. Monomers of 1 were isolated in
argon matrices at 15 K and characterized spectroscopically.
Irradiation of matrix-isolated 1 at λ = 222 nm (near its
absorption maximum) led to the corresponding 2H-azirine 3
and ketenimine 6 as primary photoproducts and also to nitrile
ylide 4 and 2,5-dimethyloxazole (5). The photoproducts were
identified (i) by comparison with infrared spectra of authentic
matrix-isolated samples of 3 and 5 and (ii) using additional
irradiations at longer wavelengths (where 1 does not react)
which induce selective photoisomerizations of 4 and 6. In
particular, irradiation with λ = 340 nm led to the unequivocal identification of the nitrile ylide anti-4, which was transformed into
oxazole 5. The details of the 1,5‑electrocyclization of the carbonyl nitrile ylide 4 and its structural nature (propargyl-like versus
allene-like geometry) were also characterized using theoretical calculations. Thus, the elusive carbonyl nitrile ylide 4 was captured
and characterized for the first time as an intermediate in the isoxazole−oxazole photoisomerization.
nitrene II and nitrile ylide IV reactive intermediates is not
required.24 At the MO-CI (STO-3G) ab initio level, the
photorearrangement of isoxazole I to 2H-azirine III was
theoretically found to occur in a concerted manner through the
lowest singlet excited state (S1). The excitation of 2H-azirine
III to the S2 excited singlet state induces isomerization to
oxazole V without any intermediate involved.24 However,
recently we have shown that vinyl nitrenes should play a central
role in isoxazole thermal reactivity and also that nitrile ylides
should be intermediates in the isomerization of 2H-azirines III
to oxazoles V.25 Nevertheless, concerning the mechanism of
isoxazole thermal or photochemical reactivity, until now the
postulated reactive intermediates species involved, i.e. the vinyl
nitrene II and the nitrile ylide IV, have escaped from direct
experimental identification and their existence still remains to
be proved.26
INTRODUCTION
■
Isoxazoles are five-membered heterocyclic compounds with
important applications in areas such as pharmaceuticals,
agrochemistry, molecular electronics, corrosion inhibitors,
liquid crystalline materials, and others.1−3 Isoxazoles are also
important building blocks in organic synthesis. Their aromatic
character gives enough stabilization to allow the manipulation
of their substituents, and on the other hand, their weak N−O
bond can be easily cleaved under different conditions to give
functionalized compounds.1−3
Under photochemical conditions isoxazoles rearrange to
oxazoles via 2H-azirine intermediates, which show a remarkable
wavelength-dependent photochemistry.4−6 This fascinating
photochemical transformation was first reported by Ullman
and Singh in 1966.4 Afterward, several other studies were
carried out and isoxazole−oxazole isomerization has been
established as the most common pathway in isoxazole
photochemistry.7−23 From a mechanistic point of view, the
first step of this photoisomerization reaction has been
suggested to occur through cleavage of the N−O bond of
isoxazole I with the generation of a diradical, the vinyl nitrene
type reactive intermediate II, which rapidly collapses to give the
2H-azirine III (Scheme 1). Then the 2H-azirine III can
undergo C−C bond cleavage to generate a zwitterionic species,
the nitrile ylide type reactive intermediate IV, which quickly
isomerizes to oxazole V.
In a previous paper,27 we investigated the photochemistry of
the parent isoxazole and found surprisingly no indication of
isomerization to oxazole. Instead, we observed the formation of
nitriles via the corresponding ketenimine, which was established
to be a key intermediate in that reaction. According to the
mechanistic model proposed, it is conceivable that the nitrile
formation will take place preferentially in the case of the
photochemistry of 3-unsubstituted isoxazoles.
According to an early theoretical study performed for the
Received: July 18, 2013
isoxazole−oxazole photoisomerization, the existence of vinyl
Published: September 27, 2013
© 2013 American Chemical Society
10657
dx.doi.org/10.1021/jo4015672 | J. Org. Chem. 2013, 78, 10657−10665

The Journal of Organic Chemistry
Article
Scheme 1. Photochemical Isomerization of Isoxazoles I to Oxazoles V via 2H-Azirines III
Herein, we report on the photochemistry of 3,5-dimethyl-
isoxazole (1, R1 = R3 = Me, R2 = H; see Scheme 1). In this
study we took advantage of a combination of narrow-band
tunable UV laser irradiation and low-temperature matrix
isolation coupled with infrared spectroscopy, used as
experimental techniques. As a result, for the first time, it was
possible to capture and identify the elusive nitrile ylide reactive
intermediate in the isoxazole−oxazole photoisomerization.
use it in the analysis of the experimental IR spectra obtained in
the photochemistry study.
Photochemistry of 3,5-Dimethylisoxazole at λ = 222
nm. The photochemistry of 3,5-dimethylisoxazole (1) isolated
in an argon matrix was induced using narrow-band UV laser
light tuned at λ = 222 nm. This wavelength is the lowest
available in our laser system, and also, it is close enough to the
absorption maximum of 1 (UV-vis (ACN) λmax ∼215 nm, ε ≈
5500 dm3 mol−1 cm−1). Under these conditions, two types of
irradiation experiments were carried out: using either shorter (a
few seconds) or longer (several minutes) irradiation times. The
changes in the samples were followed by infrared spectroscopy
after each irradiation.
Figure 2a presents the results of a total irradiation time of 1
min, in which about 2% of 1 was consumed and two
photoproducts A and B were generated. Figure 2b presents
the results of a total irradiation time of 80 min, in which about
46% of 1 was consumed and, in addition to A and B,
photoproduct C was observed in a considerable amount. The
progress of the photochemistry of 1 from 0 to 180 s is shown in
Figure S3 (Supporting Information), in the 2150−1630 cm−1
region, where the most characteristic bands of the photo-
products appear. The kinetics of the formation of photo-
products shows two relevant facts: (i) initial stage irradiations,
i.e. from 0 to 60 s (total time), indicate that photoproducts A
and B are formed simultaneously and that only they appear at
the very beginning of the photochemistry; (ii) prolonged
irradiations, i.e. from 60 to 180 s (total time), allow
identification of another photoproduct, labeled as C.
RESULTS AND DISCUSSION
■
IR Spectrum of Matrix-Isolated 3,5-Dimethylisoxa-
zole. The experimental FTIR spectrum of monomeric 3,5-
dimethylisoxazole (1) isolated in an argon matrix at 15 K and
the simulated theoretical IR spectrum are shown in Figure 1.
By comparison of the results of short and longer irradiation
times (parts a and b of Figure 2), it is evident that for longer
irradiation times the ratio between photoproducts A and B
decreases (and therefore A and B are different species). The
data also show the accumulation of photoproduct C for longer
irradiation times, which is manifested by an increase in the C:A
or C:B ratio in IR spectra. Finally, the observation of
photoproducts labeled as X after the formation of the
photoproducts A−C, and after prolonged irradiation, should
also be mentioned. These probably result from photo-
decomposition and will not be discussed here.
Identification of 3,5-Dimethylisoxazole Photoisome-
rization Products A−C. As shown before, by monitoring IR
spectra during a series of irradiations of 3,5-dimethylisoxazole
(1) (at λ = 222 nm), it was possible to determine the formation
of two primary photoproducts, A and B, and a secondary
product, C. In this section, we will present the identification of
these photoproducts with the aid of IR spectra of authentic
samples or with the aid of subsequent irradiation experiments
using longer wavelengths: i.e., under conditions where 1 is
unreactive but its photoproducts can be transformed.
Figure 1. (a) Experimental FTIR spectrum of 3,5-dimethylisoxazole
(1) isolated in an argon matrix at 15 K. (b) IR spectrum of 1 simulated
at the B3LYP/6-311++G(d,p) level of theory (for details, see the
Experimental Section).
The intense bands observed at ∼1616 and ∼1415 cm−1 in the
experimental IR spectrum of 1 exhibit a site-splitting effect.
Such an effect results from the fact that the rigid matrix
accommodates the guest molecules in different local environ-
ments that may present slightly different physicochemical
properties. The scaling factor of 0.980 used to correct the
calculated frequencies was obtained from a least-squares linear
fit of the experimental versus calculated frequencies in the
fingerprint region of the IR spectrum (see the Supporting
Information, Figure S1). The assignment of the experimental
IR spectrum of 1 isolated in an argon matrix, together with the
calculated IR spectrum at the B3LYP level, and the results of
normal coordinate analysis are presented in Table S1
(Supporting Information). The fact that the density functional
theory B3LYP method provides an accurate estimation of
vibrational frequencies of 1 gives us additional confidence to
Identification of Photoproduct A. For photoproduct A the
most distinctive band appears at 1713 cm−1, a characteristic
region for ν(CO) vibrations. With this information, 2-acetyl-
3-methyl-2H-azirine (3) arises as a possible candidate for
photoproduct A, also because it is known that the photo-
10658
dx.doi.org/10.1021/jo4015672 | J. Org. Chem. 2013, 78, 10657−10665

The Journal of Organic Chemistry
Article
Figure 2. Experimental difference IR spectra obtained as the spectrum after UV irradiation at λ = 222 nm of 1 in an argon matrix minus the spectrum
of the sample before irradiation. Total exposition of the samples to UV light: (a) total irradiation time of 1 min; (b) total irradiation time of 80 min.
The rectangle in the 2150−1630 cm−1 region (specified with a gray background) indicates where the most characteristic bands due to
photoisomerization products (labeled as A−C) appear (see text). Monomeric water bands (at 1624 and 1608 cm−1) were removed from spectra for
clarity.28
Figure 3. (a) Experimental IR spectrum of photoproducts of 1 in an argon matrix obtained after 1 min of UV irradiation at λ = 222 nm (see also
Figure 2a). (b) Experimental IR spectrum of authentic 2-acetyl-3-methyl-2H-azirine (3) isolated in an argon matrix in a separate experiment.25 The
cross symbol (×) designates the bands of photoproduct A assigned to 3. The plus sign symbol (+) indicates bands due to photoproduct A which also
have contribution due to photoproduct B.
chemistry of 1 in solution leads to 3, though with a low yield.16
In fact, as shown in Figure 3, where the experimental IR
spectrum of photoproducts of 1 in an argon matrix (λ = 222
nm; 1 min) is compared with the IR spectrum of an authentic 3
isolated in an argon matrix in a separate experiment,25 the
identification of photoproduct A as the 2H-azirine 3 is
unequivocal. The data on experimental and B3LYP calculated
IR spectra of 2-acetyl-3-methyl-2H-azirine (3) are given in
Table S3 (Supporting Information).
A comparison of the spectrum of 3 in Figure 3b and the
spectrum of photoproduct A in Figure 3a shows some
differences in the intensities of the IR bands, which arise
probably due to a different conformer ratio. Calculations at the
B3LYP/6-311++G(d,p) level indicate that the acetyl moiety in
2-acetyl-3-methyl-2H-azirine (3) adopts syn and anti orienta-
tions, with the syn more stable by only about 0.15 kJ mol−1 (see
Figure S4, Supporting Information). The syn-3:anti-3 pop-
ulation ratio in the gas-phase equilibrium at room temperature,
estimated using the Boltzmann distribution, is about 52:48.
This ratio is expected to be trapped in the matrix of 3 prepared
using an authentic sample evaporated at room temperature
(shown in Figure 3b). However, the ratio syn-3:anti-3 produced
by in situ irradiation of 1 isolated in an argon matrix (the
photoproduct A shown in Figure 3a) could be different due to
distinct experimental conditions.
Identification of Photoproduct B. The most characteristic
bands of photoproduct B appear at 2085 and 2064 cm−1,
spectral region where ketenimines (R2CCNR) generally
absorb strongly,25,29 and between 1690 and 1670 cm−1. These
absorptions suggest that 3-acetyl-N-methylketenimine (6) is a
10659
dx.doi.org/10.1021/jo4015672 | J. Org. Chem. 2013, 78, 10657−10665

The Journal of Organic Chemistry
Article
Figure 4. (a) Experimental difference IR spectrum showing changes in the spectrum of photoproduct B after 2 min of UV irradiation at λ = 260 nm
(after irradiations at 222 nm and at 340 nm). Growing bands are displayed upward. (b) IR spectrum simulated at the B3LYP/6-311++G(d,p) level
considering the consumption of syn-3-acetyl-N-methylketenimine (syn-6) and the production of anti-3-acetyl-N-methylketenimine (anti-6).
Figure 5. (a) IR spectrum of nitrile ylide anti-4 simulated at the B3LYP/6-311++G(d,p) level. (b) Experimental difference IR spectrum showing
changes after 3 min of UV irradiation at λ = 340 nm (subsequent to 80 min of irradiation at λ = 222 nm). Growing bands are displayed upward. The
violet crosses (×) designate the bands formed at the expense of C and assigned to 2,5-dimethyloxazole (5) (positive bands). The red crosses (×)
designate the bands of C assigned to anti-4 (negative bands). The plus signs (+) indicate bands due to photoproduct B, as assigned in Figure 4. (c)
Experimental IR spectrum of authentic 2,5-dimethyloxazole (5) isolated in an argon matrix in a separate experiment.25 The asterisks (*) indicate
bands of monomeric water.
possible candidate for photoproduct B. Indeed, 6, generated
from the reaction of 2,5-dimethylisoxazolium iodine with
triethylamine at −77 °C, was characterized by an IR absorption
band at 2075 cm−1,15 which agrees well with the mean
frequency of the two bands of B observed in this region.
The definitive assignment of photoproduct B to ketenimine
6 was established in this study on the basis of results of
additional irradiations performed on the photolyzed matrix of
1. After the isoxazole 1 was irradiated at λ = 222 nm in an argon
matrix for 80 min (Figure 2b), further irradiations were
performed using longer wavelengths to investigate only the
photochemistry of photoproducts of 1. Thus, after the
consumption of photoproduct C by irradiations in the 340
nm region (as presented below), it was observed that
subsequent irradiations at 260 nm selectively induce changes
in photoproduct B. As shown in Figure 4a, irradiations at 260
10660
dx.doi.org/10.1021/jo4015672 | J. Org. Chem. 2013, 78, 10657−10665

The Journal of Organic Chemistry
Article
nm result in redistribution of intensities within the bands of
photoproduct B, which is consistent with the possibility of
photoisomerization. At the B3LYP/6-311++G(d,p) level,
ketenimine 6 was found to adopt two conformations, syn-6
and anti-6, which are characterized by different relative
orientations of the acetyl moiety and approximately the same
energy (see Figure S5, Supporting Information). Indeed, as
shown in Figure 4, the theoretical difference IR spectrum
simulating the quantitative isomerization syn-6 → anti-6
demonstrates an excellent agreement with the experimental
difference IR spectrum.
Table 1. Experimental FTIR Spectrum of the Photoproduct
C and Selected Vibrational Frequencies and Infrared
Intensities of Nitrile Ylide anti-4 Calculated at the B3LYP/6-
a
311++G(d,p) Level
Ar matrix (15 K)
calculated
b
c
d
ν/cm−1
I
ν/cm−1
I
approximate assignment
νas(CNC)
2037
s
2036
1667
1454
1450
1440
1440
1401
1365
1359
1240
1228
1027
1021
1009
981
395.6
543.1
10.7
9.8
1654
1446
s
?
ν(CO)
δas(CH3)′nitrile
δas(CH3)″carbonyl
δas(CH3)′carbonyl
δas(CH3)″nitrile
νs(CNC) + δ(NCH)
δs(CH3)nitrile
For example, the decreasing bands at 2085 and 1674 cm−1
match well with frequencies of the νas(CCN) and
ν(CO) vibrations of syn-6, calculated at 2118 and 1685
cm−1, respectively. On the other hand, the increasing bands at
2064 and 1688/1681 cm−1 show a good correlation with the
νas(CCN) and ν(CO) frequencies of anti-6, calculated
at 2098 and 1691 cm−1, respectively. As presented in Table S4
(Supporting Information), these data clearly reveal the
photoisomerization of B induced by irradiation at λ = 260
nm, the unequivocal identification of photoproduct B as 3-
acetyl-N-methylketenimine (6), and the comprehensive assign-
ment of the IR bands of syn-6 and anti-6 conformers.
1437
1391
?
s
5.5
9.8
134.6
4.1
1355
1244
1233
m
s
31.8
166.1
20.4
1.5
δs(CH3)carbonyl
δ(NCH) − ν(CC)
?
δ(NCH) + ν(CC) − νs(CNC)
γ′(CH3)nitrile + γ′(CH3)carbonyl
γ″(CH3)nitrile
γ′(CH3)nitrile − γ′(CH3)carbonyl
γ″(CH3)carbonyl
1036/1031
1026
m
vw
s
100.5
15.7
124.5
21.3
8.8
984/982
879/877
777
w
874
ν(CC)nitrile
vw
w
781
ν(CC)carbonyl
Identification of Photoproduct C. As mentioned before, C
appears as a secondary photoproduct in the photochemistry of
3,5-dimethylisoxazole (1) at λ = 222 nm and then plausibly
results from photochemistry of A or B. Interestingly, after 80
min of irradiation of isoxazole 1 at λ = 222 nm (Figure 2b), it
was found that photoproduct C was consumed selectively by
subsequent irradiation at λ = 340 nm. As shown in Figure 5b,
several new low-intensity bands emerge in the IR spectrum
concomitant with the consumption of product C. A comparison
of these new bands with the experimental IR spectrum of
authentic 2,5-dimethyloxazole (5) isolated in an argon matrix
(see Figure 5c) clearly establishes that irradiation at λ = 340 nm
transforms product C into 5. Some traces of 5 were also
detected in the spectrum obtained after irradiation of isoxazole
1 (λ = 222 nm, 80 min). The experimental and calculated IR
spectra of 5 are given in Table S5 (Supporting Information;
note that all infrared intensities of 5 are intrinsically very low).
From a mechanistic point of view, the last step of the
isoxazole−oxazole photoisomerization might occur through a
reactive nitrile ylide intermediate, which should be formed by
C−C bond cleavage of the 2H-azirine (see Scheme 1). In fact,
as shown in Figure 5, the secondary photoproduct C, captured
in the photochemistry of matrix-isolated isoxazole 1, is
undoubtedly assigned to the nitrile ylide anti-4 on the basis
of an excellent correspondence between the calculation and the
experiment. For example, the most prominent infrared bands of
nitrile ylide appear at 2037 and 1654 cm−1 and can be readily
assigned to the νas(CNC) and ν(CO) of anti-4,
calculated to appear at 2036 and 1667 cm−1, respectively.
The absence of the syn-4 form was confirmed by comparison of
absorptions of photoproducts with calculated IR spectra of syn-
4 (see Figure S6, Supporting Information). A comprehensive
assignment of the experimental spectrum of C to the calculated
spectrum of nitrile ylide anti-4 is presented in Table 1.
741
717
38.9
γ(CH)
a
Photoproduct C (same as nitrile ylide 4) was generated during the
b
photochemistry of isoxazole 1 at 222 nm. Experimental intensities are
presented in qualitative terms. Abbreviations: s, strong; m, medium; w,
c
weak; vw, very weak. Calculated intensities (I) are given in km mol−1.
The calculated B3LYP/6-311++G(d,p) frequencies were scaled by
0.980. Legend: ν, stretching; δ, bending; γ, rocking; +, in syn phase;
−, in anti phase; s, symmetric; as, antisymmetric,
d
6 as primary photoproducts and also to nitrile ylide 4 and
oxazole 5. It is likely that the first step in the photochemistry of
1 involves N−O bond cleavage with the formation of the very
reactive vinyl nitrene 2.25,27 However, no indications of this
intermediate were obtained, even when experiments were
performed in nitrogen matrices at 12 K.30 Although the
existence of the vinyl nitrene intermediate 2 remains to be
proven, the simultaneous appearance of 3 and 6 during the first
seconds of irradiation of 1 suggests the participation of 2 as a
key intermediate. In this mechanism, the formation of 2 is
followed by its rapid isomerization by two competitive paths:
(i) decay to 2H-azirine 3 (Scheme 2, green) and (ii) a [1,2]-
methyl shift to form ketenimine 6 (Scheme 2, blue).
In order to investigate if two different electronic config-
urations of 2, the triplet ground state and the open shell singlet
excited state, are each independently responsible for the
formation of 3 or 6, the photochemistry of 1 was also studied
in xenon matrices. The heavy-atom xenon matrix might favor
the path involving the triplet manifold31,32 of vinyl nitrene 2.
Nevertheless, the results obtained in xenon matrices did not
show significant differences from those obtained in argon
matrices.
As already mentioned, upon irradiation experiments at λ =
222 nm, performed on matrix-isolated isoxazole 1, a series of
additional laser irradiations was carried out at longer wave-
lengths where 1 does not absorb, but its photoproducts do
react. These additional irradiations allowed inducing selective
transformations in some of photoproducts of 1 and acquire
important information about their structures. The irradiation at
λ = 260 nm (performed after irradiation at λ = 340 nm) led to
Mechanistic Discussion of the Photochemistry of
3,5‑Dimethylisoxazole. A summary of the experimental
results and the mechanistic proposal for the photochemistry
of 3,5‑dimethylisoxazole (1) are presented in Scheme 2.
Irradiation experiments of 1 at 222 nm, near its absorption
maximum, led to the formation of 2H-azirine 3 and ketenimine
10661
dx.doi.org/10.1021/jo4015672 | J. Org. Chem. 2013, 78, 10657−10665

The Journal of Organic Chemistry
Article
Scheme 2. Summary of the Experimental Observations and the Mechanism for Photochemistry of 3,5-Dimethylisoxazole 1
Isolated in a Low-Temperature Argon Matrix
photoisomerization of 3-acetyl-N-methylketenimine (6). Start-
ing from an approximately equivalent ratio of syn-6 and anti-6
conformers, trapped in the samples upon irradiation of
isoxazole 1 at λ = 222 nm, the irradiation at λ = 260 nm
induced photoisomerization of syn-6 to anti-6 by intramolecular
rotation of the acetyl moiety. This observation could be
interpreted as a wavelength-dependent photochemical equili-
brium; i.e. the photoequilibrium was strongly shifted toward
anti-6 for UV irradiations at 260 nm, whereas syn-6 and anti-6
are in similar amounts in the photoequilibrium after initial
irradiations at λ = 222 nm. UV-induced conformational
isomerization of the acetyl group has been reported for some
matrix-isolated molecules such as hydroxyacetophenone33 and
acetylketene,34 the latter being structurally very similar to 6. A
wavelength-dependent photochemical equilibrium was also
identified in some syn−anti photoisomerizations.35
However, the most important result was obtained when the
photoproducts of isoxazole 1 were irradiated at λ = 340 nm.
This led to the transformation of a secondary photoproduct
(initially designated as C) and then to its unequivocal
identification as being the elusive nitrile ylide anti-4. The
product that results from the photochemistry of anti-4 was
reliably identified as 2,5-dimethyloxazole (5). Traces of 5 were
also detected after irradiation 1 at λ = 222 nm over 80 min. It
could be expected that, for extended irradiations of 1, oxazole 5
would accumulate in the sample as the end product, according
to the mechanistic rationalization presented in Scheme 2.
Nature of Nitrile Ylide 4 as an Intermediate in
Isoxazole−Oxazole Photoisomerization. The nitrile ylides
are species with a CNC framework containing six electrons in π
and n orbitals.36 Typically, due to their high reactivity, nitrile
ylides are classified as elusive reactive intermediate species.37
Therefore, it not surprising that most of the experimental
information about the molecular structure of nitrile ylides has
been acquired by low-temperature matrix isolation and IR
spectroscopy.38 Nevertheless, to date, only a few acyclic nitrile
ylides have been captured and studied using this experimental
approach.29,39−44
isomerization was isolated here and characterized for the first
time. The conjugation of the carbonyl group with the nitrile
ylide allows for the possibility of an especially facile
intramolecular 1,5-electrocyclization, which makes it distinct
from all the nitrile ylides so far reported experimentally and
even more difficult to observe. The 1,5-electrocyclization
reaction to oxazole occurs via a planar or nearly planar
nonrotatory transition state involving the in-plane attack of the
lone pair of the oxygen carbonyl atom.45 However, only the
nitrile ylide syn-4 is perfectly aligned to provide the possibility
of cyclization and must be the precursor of oxazole 5. anti-4
lacks such a geometric alignment, and precisely this conformer
of the nitrile ylide 4 was captured in the matrix and identified in
the current work. By analogy with photoproduct 3-acetyl-N-
methylketenimine (6), it could be anticipated that UV
irradiation of the matrices induces conformational isomer-
ization also in nitrile ylide 4.46 Thus, according to our
interpretation, the observed transformation of anti-4 to oxazole
5 must have occurred via the syn-4 intermediate, which has a
very low energy barrier toward 1,5-electrocyclization and thus is
more difficult to capture. Indeed, the calculated potential
energy surface connecting anti-4, syn-4, and 5 corroborates our
hypothesis (see Figure 6).
The calculated potential energy barrier separating the nitrile
ylide syn-4 from the much more stable oxazole 5 is only 10 kJ
mol−1. Low barriers can be easily overcome at 15 K in
cryogenic matrices by the effect known as conformational
cooling.47,48 Indeed, if the considered molecules are formed in
matrices during UV-induced isomerization, they must have an
excess of vibrational energy and their thermal relaxation over
such low barriers should become even easier. That is why syn-4,
when formed, should instantly collapse to 5. On the other
hand, the anti-4 conformer is not aligned regarding isomer-
ization toward 5, and the pathway from anti-4 to 5 should go
across the syn-4 intermediate. Therefore, the effective barrier
between anti-4 and syn-4 is the barrier which stabilizes anti-4 in
the matrices. It is equal to 54.5 kJ mol−1, according to the
B3LYP/6-311++G(d,p) calculation (see Figure 6). Thus, when
anti-4 is produced, such a barrier is high enough to prevent
thermal relaxation and allow its capture in the matrix.
To the best of our knowledge, a carbonyl nitrile ylide reactive
intermediate postulated for the isoxazole−oxazole photo-
10662
dx.doi.org/10.1021/jo4015672 | J. Org. Chem. 2013, 78, 10657−10665

The Journal of Organic Chemistry
Article
Figure 6. Relaxed potential energy profiles for (right, red) the
intramolecular rotation of the acetyl group in nitrile ylide 4 around the
NCCO dihedral angle and (left, purple) the ring-contraction from
nitrile ylide syn-4 to 2,5-dimethyloxazole (5) as a function of the OC
distance. The energy of 5 was chosen as the relative zero. Both scans
were carried out at the B3LYP/6-311++G(d,p) level of theory.
41
40
■
▲
Figure 7. Experimentally observed (bottom, black; 7 ( ), 8 ( ),
29
◆
●
anti-9 ( ), and anti-4 ( ; present work)) and theoretically
calculated (top, red and blue) frequencies of the νas(CNC) vibration
in nitrile ylides. All theoretical frequencies were obtained for
geometries fully optimized at the B3LYP/6-311++G(d,p) level of
theory and scaled by 0.980. The equilibrium CN ylide (red) and NC
nitrile (blue) calculated bond lengths are shown as functions of the
Finally, it appears instructive to us to comment on the
structure of nitrile ylides. It is known that nitrile ylides can be
represented by two different resonance structures, propargyl-
like and allene-like (Scheme 3). Theoretical studies have shown
○
calculated νas(CNC) frequency. Note that syn-4 ( ) does not have an
experimental counterpart, as explained in the text. The phenyl nitrile
■
ylide 8 ( , dashed lines) appears in the experiment as a split band due
to the Fermi resonance.40 Abbreviations: Np, 1-naphthyl; Ac, acetyl.
Scheme 3. Propargyl-Like (Left) And Allene-Like (Right)
Resonance Structures of Nitrile Ylides
present work, ●, |CN| = 130.2 pm and |NC| = 119.6 pm) this
difference is more than 10 pm and the νas(CNC) frequency is
the highest. In other terms, a small difference between the two
“NC” bonds means that both of them tend to adopt the double-
bond character (allene-like structure), while a greater
separation in bond lengths indicates a trend toward the
propargyl-like structure.52 The fact that the nitrile ylide trapped
in this work stands apart from other nitrile ylides should be
explained due to the presence of the electron-withdrawing
substituent (the acetyl group) at the ylide side of the molecule,
which promotes a more propargyl-like structure.
that simple nitrile ylides (for example, R1 = Me, R2 = H, Cl, R3
= H, Cl; Scheme 3) have allene-like character.36,49 Indeed, the
detailed experimental work of the Pimentel group on the IR
spectra of nitrile ylides generated from photolysis of 3-phenyl-
2
2H-azirine (and their 13C, 15N, and H isotopically substituted
derivatives) in an N2 matrix at 12 K led to the proposal of an
allene-like rather than a propargyl-like structure.40 The IR band
of the antisymmetric CNC vibration of phenyl nitrile ylide (R1
= Ph, R2 = R3 = H; Scheme 3) was observed in that study as a
multiplet feature with main maxima at 1926 and 1903 cm−1.
However, in clear contrast, the nitrile ylide derived from the
photochemistry of isoxazole 1 in the present study exhibited an
absorption at a much higher frequency: i.e., at 2037 cm−1 (see
Figure 5). In both of the above cases, the experimentally
observed frequencies were in excellent accord with theoretical
calculations.
Figure 7 presents the experimental infrared data on matrix-
isolated nitrile ylides, available from different sources,50,51
juxtaposed with theoretical calculations carried out at the
B3LYP/6-311++G(d,p) theory level. As can be seen, the
νas(CNC) vibration may appear in a very broad frequency
range, and its position in the spectrum correlates with the
difference between the two “NC” bond lengths (CN ylide and
NC nitrile) and therefore with the structure of the nitrile ylide.
In the unsubstituted nitrile ylide 7 (▲, |CN| = 127.8 pm and
|NC| = 121.9 pm) and phenyl nitrile ylide 8 (■, |CN| = 128.0
pm and |NC| = 122.0 pm) this difference is the smallest (ca. 6
pm) and the νas(CNC) frequency is the lowest (ca. 1920
cm−1). In the acetyl-substituted nitrile ylide (anti-4 in the
CONCLUSION
■
Monomers of 3,5-dimethylisoxazole (1) were isolated in argon
matrices at 15 K and characterized spectroscopically. Photolysis
of 1 at λ = 222 nm yields two primary photoproducts, 2-acetyl-
3-methyl-2H-azirine (3) and 3-acetyl-N-methylketenimine (6).
Most probably, they are formed through a common acetyl vinyl
nitrene intermediate 2. Upon longer times UV irradiation at λ =
222 nm, two additional photoproducts are formed, being
subsequently identified as acetyl nitrile ylide 4 and 2,5-
dimethyloxazole (5).
The unequivocal spectroscopic identification of 3 and 5
generated from 1 was made possible by comparison with
infrared spectra of authentic matrix-isolated 3 and 5 obtained in
separate experiments. The identification of 4 and 6 was
achieved in additional photochemical experiments with UV
irradiation at longer wavelengths, where 1 does not react. In
particular, ketenimine 6 undergoes syn−anti isomerization on
irradiation with λ = 260 nm, while the nitrile ylide 4 transforms
into the oxazole 5 by irradiation at λ = 340 nm.
The kinetic data indicate that the elusive carbonyl nitrile
ylide reactive intermediate 4 is formed from 2H-azirine 3,
10663
dx.doi.org/10.1021/jo4015672 | J. Org. Chem. 2013, 78, 10657−10665

The Journal of Organic Chemistry
Article
during the isoxazole−oxazole photochemical conversion. The
nitrile ylide 4 can adopt two conformations. The syn-4
conformer has the proper geometry for 1,5-electrocyclization
to the oxazole 5 and could not be experimentally detected. In
contrast, the nitrile ylide anti-4 conformer is not properly
aligned regarding cyclization toward the oxazole 5, which
allowed its experimental capture in the cryogenic matrix and
spectroscopic characterization. Interestingly, the nitrile ylide
observed in this study, for the first time, stands apart from the
few other nitrile ylides reported experimentally, exhibiting a
trend toward a propargyl-like geometry instead of an allene-
type structure.
B3LYP/6-311++G(d,p) calculations. This material is available
free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Authors
*E-mail for C.M.N.: cmnunes@qui.uc.pt.
*E-mail for I.R.: reva@qui.uc.pt.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
These studies were partially funded by the Portuguese
■
“Fundaca
̧
o para a Cien
̂
cia e a Tecnologia” (FCT), FEDER,
̃
EXPERIMENTAL SECTION
■
and projects PTDC/QUI-QUI/111879/2009, PTDC/QUI-
QUI/118078/2010, FCOMP-01-0124-FEDER-021082, co-
funded by QREN-COMPETE-UE. C.M.N. acknowledges the
FCT for the Postdoc Grant No. SFRH/BPD/86021/2012.
Sample. A commercial sample of 3,5-dimethylisoxazole (1; 98%)
was used. The synthesis of 2-acetyl-3-methyl-2H-azirine (3) and 2,5-
dimethyloxazole (5), and experimental IR spectra of these species
isolated in argon matrices, are reported in ref 25.
Matrix-Isolation FTIR Spectroscopy. Prior to use, the sample
was degassed by the standard freeze−pump−thaw method. Afterward,
the sample vapor was premixed with high-purity argon (N60) in a ratio
of ∼1:1500, in a 3 L glass reservoir. The premixed sample was then
deposited effusively onto a CsI window, cooled to 15 K. A closed-cycle
helium refrigeration system was used. The temperature of the CsI
window was measured directly by a silicon diode sensor connected to a
digital controller, providing stabilization with accuracy of 0.1 K.
The IR spectra, in the 4000−400 cm−1 range, were obtained using a
Fourier transform infrared spectrometer, equipped with a deuterated
triglycine sulfate (DTGS) detector and a Ge/KBr beam splitter, with
0.5 cm−1 resolution. To avoid interference from atmospheric H2O and
CO2, the optical path of the spectrometer was continuously purged by
a stream of dry air.
UV-Laser Irradiation Experiments. The matrices were irradiated
through a quartz window of the cryostat, using a frequency-doubled
signal beam provided by an optical parametric oscillator (fwhm ∼0.2
cm−1) pumped with a pulsed Nd:YAG laser (repetition rate 10 Hz,
pulse energy ∼0.5 mJ, duration 10 ns).
Theoretical Calculations. All calculations were carried out using
the GAUSSIAN 09 program package.53 With the aim of modeling the
IR spectra, the geometry optimizations at the B3LYP/6-311++G(d,p)
level were followed by harmonic frequency calculations at the same
level of theory. To correct for the vibrational anharmonicity, basis set
truncation, and the neglected part of electron correlation, the
calculated frequencies were scaled by 0.980. The resulting frequencies,
together with the calculated intensities, were used to simulate the
spectra by convoluting each peak with a Lorentzian function with a full
width at half-maximum (fwhm) of 2 cm−1.54 The peak intensities of
the simulated spectra (in arbitrary units of “relative intensity”) are
several times less than the calculated intensities (in km mol−1).
REFERENCES
■
(1) Comprehensive Heterocyclic Chemistry; Lang, S. A., Lin, Y. I., Eds.;
Pergamon: Oxford, U.K., 1984; Vol. 6, Part 4B.
(2) Chemistry of Heterocyclic Compounds; Grunanger, P., Vita-Finzi,
̈
P., Eds.; Wiley: New York, 1991; Vol. 49, Part 1.
(3) Pinho e Melo, T. M. V. D. Curr. Org. Chem. 2005, 9, 925.
(4) Ullman, E. F.; Singh, B. J. Am. Chem. Soc. 1966, 88, 1844.
(5) Singh, B.; Ullman, E. F. J. Am. Chem. Soc. 1967, 89, 6911.
(6) Singh, B.; Zweig, A.; Gallivan, J. B. J. Am. Chem. Soc. 1972, 94,
1199.
(7) Kurtz, D. W.; Shechter, H. Chem. Commun. 1966, 6, 689.
(8) Good, R. H.; Jones, G. J. Chem. Soc. C 1971, 1196.
(9) Ferris, J. P.; Antonucci, F. R.; Trimmer, R. W. J. Am. Chem. Soc.
1973, 95, 919.
(10) Sato, T.; Yamamoto, K.; Fukui, K. Chem. Lett. 1973, 111.
(11) Sato, T.; Saito, K. J. Chem. Soc., Chem. Commun. 1974, 99, 781.
(12) Padwa, A.; Chen, E.; Ku, A. J. Am. Chem. Soc. 1975, 97, 6484.
(13) Murature, D. A.; Peres, J. D.; Bertorello, M. M.; Bertorello, H. E.
An. Asoc. Quim. Argent. 1976, 64, 337.
(14) Dietliker, K.; Gilgen, P.; Heimgartner, H.; Schmid, H. Helv.
Chim. Acta 1976, 59, 2074.
(15) Ferris, J. P.; Trimmer, R. W. J. Org. Chem. 1976, 41, 13.
(16) Sato, T.; Yamamoto, K.; Fukui, K.; Saito, K.; Hayakawa, K.;
Yoshiie, S. J. Chem. Soc., Perkin Trans. 1 1976, 783.
(17) Sauers, R. R.; Van Arnum, S. D. Tetrahedron Lett. 1987, 28,
5797.
(18) Sauers, R. R.; Hadel, L. M.; Scimone, A. A.; Stevenson, T. A. J.
Org. Chem. 1990, 55, 4011.
(19) D’Auria, M. Adv. Heterocycl. Chem. 2001, 79, 41.
(20) Fonseca, S. M.; Burrows, H. D.; Nunes, C. M.; Pinho e Melo, T.
M. V. D.; Rocha Gonsalves, A. M. d. Chem. Phys. Lett. 2005, 414, 98.
(21) Pavlik, J. W.; St. Martin, H.; Lambert, K. A.; Lowell, J. A.;
Tsefrikas, V. M.; Eddins, C. K.; Kebede, N. J. Heterocyl. Chem. 2005,
42, 273.
ASSOCIATED CONTENT
■
S
* Supporting Information
Tables and figures giving the experimental and calculated IR
spectra for 3,5-dimethylisoxazole (1), 2-acetyl-3-methyl-2H-
azirine (3), 3-acetyl-N-methylketenimine (6), and 2,5-dime-
thyloxazole (5), the definition of internal coordinates used in
the normal-mode analysis of 1, least-squares linear fit data of
the experimental versus calculated frequencies in the fingerprint
region of the IR spectrum of 1, experimental IR spectra
showing the progress from 0 to 180 s of the photochemistry of
1 at λ = 222 nm, calculated relaxed potential energy profiles for
the internal rotation of the C−C bonds of 3 and 6, a
comparison of calculated IR spectra of nitrile ylide, syn-4 and
anti-4, with experimental data, most significant resonance
structures and IR absorptions for five “non-classic” nitrile ylides,
and Cartesian coordinates and frequencies (1 and 3−9) from
(22) Fonseca, S. M.; Burrows, H. D.; Nunes, C. M.; Pinho e Melo, T.
M. V. D. Chem. Phys. Lett. 2009, 474, 84.
́
(23) Lopes, S.; Nunes, C. M.; Gomez-Zavaglia, A.; Pinho e Melo, T.
M. V. D.; Fausto, R. J. Phys. Chem. A 2011, 115, 1199.
(24) Tanaka, H.; Osamura, Y.; Matsushita, T.; Nishimoto, K. Bull.
Chem. Soc. Jpn. 1981, 54, 1293.
(25) 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.
(26) Some experimental evidence (not unequivocal) of vinyl nitrenes
II and nitrile ylides IV have been respectively reported in: (a) 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. (b) Ref 23.
(27) Nunes, C. M.; Reva, I.; Pinho e Melo, T. M. V. D.; Fausto, R. J.
Org. Chem. 2012, 77, 8723.
10664
dx.doi.org/10.1021/jo4015672 | J. Org. Chem. 2013, 78, 10657−10665

The Journal of Organic Chemistry
Article
(28) Michaut, X.; Vasserot, A.-M.; Abouaf-Marguin, L. Vib. Spectrosc.
2004, 34, 83.
(29) Inui, H.; Murata, S. J. Am. Chem. Soc. 2005, 127, 2628.
(30) It is known that nitrogen matrices may increase lifetimes of
metastable higher-energy species; see for example: (a) Lopes, S.;
of the XCHCHC(HCNCH) type, were generated under
conditions comparable to those in the present work (in cryogenic
noble gas matrices). These nitrile ylides were charaterized by the
infrared νas(CNC) bands at 1930 cm−1 (for X = CH, BW1), and at
1961/1947 cm−1 (for X = N, BW2). However, nitrile ylides BW1 and
BW2 reported in ref 43 could assume at least four different
conformations, and indeed more than one rotamer was observed for
BW2. The ring-contraction reaction has a low predicted barrier for
syn−syn conformers (only 3.5 kcal mol−1) and it must occur easily,
similarly as it was found in the present study for syn-4. Therefore, we
believe that the observed nitrile ylides must adopt one of the three
remaining conformations, with at least one of the dihedrals being anti.
According to our B3LYP/6-311++G(d,p) calculations, the different
conformers of BW1 and BW2 present a significative scattering in the
predicted νas(CNC) frequencies as well as variance in the two “NC”
bond lengths. Despite the fact that their precise conformational
assignment leaves doubts, BW1 and BW2 must belong to the group of
the allene-like nitrile ylides, according to their observed νas(CNC)
frequencies.
(51) Note also that Figure 7 does not include the data of five nitrile
ylides captured in low-temperature matrices in the sequence of ring-
opening reactions of fused-aromatic nitrenes.44 One of the reviewers of
this work pointed out, concerning the IR frequency νas(CNC)
vibration of nitrile ylides, that two more regions characteristic of fully
propargylic (up to 2300 cm−1) and carbenic (below 1900 cm−1)
structures should be explored. Indeed, some of the nitrile ylide species
listed in ref 44 are reported to have IR absorptions near 2200 cm−1.
Nevertheless, in these reported structures, the sp2 carbon atom of the
CNC nitrile ylide fragment is inserted in a six-membered ring, which
in some resonance structures can be fully aromatic (such as dipolar
species marked gray in Figure S7 (Supporting Information)). If this
type of resonance structure describes better the structure of the species
in question, they cannot be regarded as “classic” nitrile ylide species,
and then they cannot be directly comparable with those shown in
Figure 7.
Domanskaya, A. V; Fausto, R.; Rasanen, M.; Khriachtchev, L. J. Chem.
̈
̈
Phys. 2010, 133, 144507. (b) Nunes, C. M.; Lapinski, L.; Fausto, R.;
Reva, I. J. Chem. Phys. 2013, 138, 125101.
(31) Lundell, J.; Krajewska, M.; Rasanen, M. J. Phys. Chem. A 1998,
̈
̈
102, 6643.
(32) Kus,
Phys. 2012, 136, 144509.
̧
N.; Sharma, A.; Reva, I.; Lapinski, L.; Fausto, R. J. Chem.
(33) Lapinski, L.; Rostkowska, H.; Reva, I.; Fausto, R.; Nowak, M. J.
J. Phys. Chem. A 2010, 114, 5588.
(34) Kappe, C. O.; Wong, M. W.; Wentrup, C. J. Org. Chem. 1995,
60, 1686.
(35) Reva, I.; Nowak, M. J.; Lapinski, L.; Fausto, R. J. Phys. Chem. B
2012, 116, 5703.
(36) Escolano, C.; Duque, M. D.; Vaz
11, 741.
́
quez, S. Curr. Org. Chem. 2007,
(37) The first direct observation of a nitrile ylide, credited to Schimd
et al., was achieved in organic rigid matrices at −185 °C by means of
UV spectroscopy. See: Sieber, W.; Gilgen, P.; Chaloupka, S.; Hansen,
H.-J.; Schmid, H. Helv. Chim. Acta 1973, 56, 1679.
(38) Note that relevant information on nitrile ylide intermediates
have also been obtained by time-resolved spectroscopy. See ref 36 and
references cited therein.
(39) Wentrup, C.; Fischer, S.; Berstermann, H.-M.; Kuzaj, M.;
Luerssen, H.; Burger, K. Angew. Chem., Int. Ed. Engl. 1986, 25, 85.
̈
(40) Orton, E.; Collins, S. T.; Pimentel, G. C. J. Phys. Chem. 1986, 90,
6139.
(41) Maier, G.; Schmidt, C.; Reisenauer, H. P.; Endlein, E.; Becker,
D.; Eckwert, J.; Hess, B. A.; Schaad, L. J. Chem. Ber. 1993, 126, 2337.
(42) Naito, I.; Nakamura, K.; Kumagai, T.; Oku, A.; Hori, K.;
Matsuda, K.; Iwamura, H. J. Phys. Chem. A 1999, 103, 8187.
(43) Bednarek, P.; Wentrup, C. J. Am. Chem. Soc. 2003, 125, 9083.
(44) We are also aware of, at least, other five nitrile ylides captured in
low-temperature matrices. These five nitrile ylides, labeled here as
“non-classic”,51 were generated as intermediates in the sequence of
ring-opening reactions of fused-aromatic nitrenes. See: (a) Kvaskoff,
D.; Mitschke, U.; Addicott, C.; Finnerty, J.; Bednarek, P.; Wentrup, C.
(52) A larger separation in the two “NC” bond lengths (|CN| = 138.9
pm and |NC| = 113.1 pm) was observed in an exceptionally stable
nitrile ylide, which presents a propargylic structure, by X-ray
crystallography. See: Janulis, E. P.; Wilson, S. R.; Arduengo, A. J.
Tetrahedron Lett. 1984, 25, 405.
(53) 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, J. A.,Jr.;
Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin,
K. N.; Staroverov, V. N.; 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.;
Aust. J. Chem. 2009, 62, 275. (b) Addicott, C.; Luerssen, H.; Kuzaj, M.;
̈
Kvaskoff, D.; Wentrup, C. J. Phys. Org. Chem. 2011, 24, 999.
(c) Kvaskoff, D.; Vosswinkel, M.; Wentrup, C. J. Am. Chem. Soc. 2011,
133, 5413. (d) Wentrup, C.; Lan, N. M.; Lukosch, A.; Bednarek, P.;
Kvaskoff, D. Beilstein J. Org. Chem. 2013, 9, 743.
(45) Fabian, W.; Kappe, C.; Bakulev, V. J. Org. Chem. 2000, 65, 47.
(46) Two different conformers of a “non-classic” nitrile ylide were
observed as a result of a photoinduced ring-opening reaction of a cyclic
carbodiimide.44c This could suggest photoisomerization in the nitrile
ylide.
(47) Reva, I. D.; Lopes Jesus, A. J.; Rosado, M. T. S.; Fausto, R.;
Euseb
(48) Lopes Jesus, A. J.; Rosado, M. T. S.; Reva, I.; Fausto, R.;
Eusebio, M. E. S.; Redinha, J. S. J. Phys. Chem. A 2008, 112, 4669.
́
io, M. E. S.; Redinha, J. S. Phys. Chem. Chem. Phys. 2006, 8, 5339.
̈
Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.;
Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09,
Revision A.2; Gaussian, Inc., Wallingford, CT, 2009.
(54) Irikura, K. K. Program SYNSPEC; National Institute of
Standards and Technology, Gaithersburg, MD 20899, USA.
́
(49) Pliego, J. P., Jr. Chem. Phys. Lett. 2000, 142.
(50) Note that we opted not to add to Figure 7 the data for three
nitrile ylides reported in refs 39 and 43, due to the following reasons.
(i) In ref 39, a bis(trifluoromethyl)-tert-butyl nitrile ylide was
charaterized by the infrared band νas(CNC) at 2250 cm−1. We
optimized its geometry at the B3LYP/6-311++G(d,p) level, the same
as in the present study, and ran the vibrational calculations. The
calculated νas(CNC) frequency (ca. 2097 cm−1, scaled) shows a very
significant discrepancy from experiment. This discrepancy must result
from the environmental effects on the structure of bis(trifluorometh-
yl)-tert-butyl nitrile ylide generated in a neat KBr pellet at −196 °C.
Thus, this result is not directly comparable with IR data on monomeric
nitrile ylides isolated in inert (noble gas or nitrogen) matrices. (ii) In
the work of Bednarek and Wentrup (BW),43 conjugated nitrile ylides,
10665
dx.doi.org/10.1021/jo4015672 | J. Org. Chem. 2013, 78, 10657−10665