Sydnone photochemistry: Direct observation of earl's bicyclic lactone valence isomers (oxadiazabicyclo[2.1.0] pentanones), formation of carbodiimides, reaction mechanism, and photochromism

Article abstract of DOI:10.1071/CH13536

The matrix photolyses of 3-phenyl-, 3-pyridyl, and 3,4-diphenylsydnones 16, 19, and 22 were investigated by matrixisolation infrared spectroscopy. The formation of the neutral, bicyclic lactone valence isomers postulated by Earl - the oxadiazabicyclo[2.1.0]pentanones exo-17 and exo-20 - was clearly observed in the first two cases and is also likely in the case of exo-23 (C=Oabsorptions in the IR at 1881, 1886, and 1874 cm ^-1, respectively). The efficient photodecomposition of sydnones to carbodiimides RN=C=NR0 (18, 21, and 24) and CO2 was established in all three cases. The formation of benzonitrile 27 and azacycloheptatetraene 29 in the matrix photolysis of diphenylsydnone 22 is indicative of diphenylnitrile imine PhCNNPh 26 as an intermediate (2340 cm ^-11). Neither bicyclic lactones nor carbodiimides have been observed previously in sydnone photochemistry. A general reaction mechanism for the formation of carbodiimides, nitrile imines, and photochromism is put forward. CSIRO 2014.

Full text of DOI:10.1071/CH13536

RESEARCH FRONT
CSIRO PUBLISHING
Aust. J. Chem. 2014, 67, 457–468
Full Paper
http://dx.doi.org/10.1071/CH13536
Sydnone Photochemistry: Direct Observation of Earl’s
Bicyclic Lactone Valence Isomers (Oxadiazabicyclo[2.1.0]
pentanones), Formation of Carbodiimides, Reaction
Mechanism, and Photochromism
A
A
,
A B
Rakesh N. Veedu, David Kvaskoff, and Curt Wentrup
A
School of Chemistry and Molecular Biosciences, The University of Queensland,
Brisbane, Qld 4072, Australia.
B
Corresponding author. Email: wentrup@uq.edu.au
The matrix photolyses of 3-phenyl-, 3-pyridyl, and 3,4-diphenylsydnones 16, 19, and 22 were investigated by matrix-
isolation infrared spectroscopy. The formation of the neutral, bicyclic lactone valence isomers postulated by Earl – the
oxadiazabicyclo[2.1.0]pentanones exo-17 and exo-20 – was clearly observed in the first two cases and is also likely in the
ꢀ
1
case of exo-23 (C¼O absorptions in the IR at 1881, 1886, and 1874 cm , respectively). The efficient photodecomposition
0
of sydnones to carbodiimides RN¼C¼NR (18, 21, and 24) and CO was established in all three cases. The formation of
2
benzonitrile 27 and azacycloheptatetraene 29 in the matrix photolysis of diphenylsydnone 22 is indicative of
ꢀ
1
diphenylnitrile imine PhCNNPh 26 as an intermediate (2340 cm ). Neither bicyclic lactones nor carbodiimides have
been observed previously in sydnone photochemistry. A general reaction mechanism for the formation of carbodiimides,
nitrile imines, and photochromism is put forward.
Manuscript received: 7 October 2013.
Manuscript accepted: 31 October 2013.
Published online: 27 November 2013.
Introduction
The sydnones were the first widely recognized class of
mesoionic compounds, although similar 1,3,4-triazolium-2-
imides and 1,3,4-thiadiazolium-2-thiolates had previously been
The sydnones (1,2,3-oxadiazol-3-ium-5-olates) were first pre-
pared by Earl and co-workers at the University of Sydney by
[
dehydration of N-nitroso-N-phenylglycine (Scheme 1). Earl
proposed the bicyclic lactone structure 2,1,5-oxadiazabicyclo
[2.1.0]pentan-3-one (1) for 3-phenylsydnone (Scheme 1). The
compound analyzed correctly for C H N O , had a cryoscopi-
1]
[4]
described as amphoions by Sch o¨ nberg and as mesomeric
5]
[
betaines by Jensen and Friediger.
The most important canonical structures of 2 are presented in
Eqn 1.
[1]
8
6 2 2
cally determined molecular weight of 166 (calculated 162), and
reverted to nitrosophenylglycine on treatment with KOH.
However, the chemical and spectroscopic properties of the
sydnones were not readily reconciled with the bicyclic lactone
structure 1, and Baker and Ollis proposed instead the mesoionic
ꢁ
ꢁ
N
N
N
N
ꢀ
N
N
N
N
C
O
ð1Þ
O
O
O
O
O
O
O
2
2a
2b
2c
(mesomeric, zwitterionic) structure 2, which is now generally
accepted (Scheme 2).
[
2,3]
The ketene-type ‘no-bond’ mesomeric structure 2c was origi-
[
nally dismissed by Baker and Ollis,
structure determinations revealed a five-membered ring struc-
2b]
but X-ray crystal
Ph
N
Ph
N
N
(AcO)2O
KOH
O
N
O
ture with a peculiarly long endocyclic C–O single bond
˚
O
1
O O
(
led to the proposal of the mesomeric structure 2c as one of
,1.41 A) and large exocyclic C–C–O angle (,1368), which
H
[6,7]
Scheme 1. Earl’s proposed bicyclic ‘sydnone’ structure 1.
several resonance structures of the sydnone ring (Eqn 1).
One can speculate about the possibility to break the long C–O
bond either thermally or photochemically, thus forming a ketene
valence isomer 3 (Eqn 2).
Ph
Ph
N
H
Ph
N
N
N
N
O
AcOH
ꢀ
O
N
O
O
O O
O
Rꢂ
R
Rꢂ
R
H
?
1
2
N
N
ꢀ
N
N
C
ð2Þ
O
O
O
O
Scheme 2. The mesoionic sydnone structure 2 proposed by Baker and
2
3
Ollis.
Journal compilation Ó CSIRO 2014
www.publish.csiro.au/journals/ajc

4
58
R. N. Veedu, D. Kvaskoff, and C. Wentrup
The mass spectra of sydnones show abundant losses of NO and
[
hν
8]
CO, and measurements of ionization potentials suggested
that the neutral sydnones exist in the ring-opened nitrosoketene
O
N
O
ꢀ
N
[
9,10]
O
C
H
H
form 3 before ionization,
such nitrosoketenes by other means have met with failure
see below). Gotthardt and Reiter first proposed the involvement
of nitrosoketene 3 as a reactive intermediate to explain aspects
but attempts to characterize
4
O
Ph
Ph
(
hν
S
ꢀ
O
S
N
N
C
O
[
11]
of the solution photochemistry of sydnones.
After the discovery that several sydnones display photochro-
O
O
5
[12]
?
mic behaviour in the solid state,
and sometimes also in
several theories were advanced to explain
this phenomenon, including colour centres similar to those
Δ or hν
[
13]
N
organic glasses,
N
ꢀ
O
C
O
O
O
[
14,15]
formed by alkali metals. Current thinking holds a mole-
[16–18]
cular process responsible for the photochromic behaviour.
6
The photochromic sydnones are colourless (or slightly yellow)
in the solid state with lmax up to ,330nm, but they rapidly
develop a blue colour on exposure to UV light. Solid 3-(3-pyridyl)
sydnone develops a blue colour (lmax ,640 nm) even on
exposure to daylight or fluorescent light. The colour is bleached
again by heating (or leaving the material at room temperature in
the dark) or by infrared irradiation. The N-nitrosaminoketene 3
has been considered a molecular candidate for the blue species.
Evidence for its formation has been adduced by Nespurek
Scheme 3. Valence isomerization in pyrrolopyridinylium olates 4, 3,1,2-
oxathiazolinylum olates 5, and possibly m u¨ nchnones 6.
Rꢂ
R
Rꢂ
R
R
N
Rꢂ
hν
N
N
?
ꢁ
N
ꢃ ꢁ
Rꢂ
C N N R
N
O
ꢀ
ꢃ
N
O
COO
O
O
2
1
7
8
Scheme 4. Trozzolo’s postulated zwitterion 7.
[16–18]
et al.
on the basis of the UV-vis and electron spectroscopy
for chemical analysis (ESCA) studies, and this was allegedly
supported by CNDO/S and STO-3G calculations. However, as
described below, our calculations at the TD-B3LYP/6–31þG**
level indicate that nitrosaminoketenes should not have absorp-
tions higher than 465 nm, in agreement with the fact that
nitrosamines are generally yellow, not blue (for example
N-nitroso-N-methylaniline and N-nitrosodiphenylamine are
yellow). Furthermore, in spite of the usually very strong
C¼C¼O stretching vibrations of ketenes, all attempts to detect
ketenes and any other intermediate by IR spectroscopy of the
Earl’s bicyclic lactone structure 1. The blue coloration was
claimed to be a thermal effect, taking place only on warming the
first-formed intermediate. Therefore, Trozzolo et al. dispute that
photochromism takes place at all. The second intermediate,
responsible for the blue colour, was assigned the azonium
carboxylate zwitterionic structure 7. If formed, 7 would be a
potential precursor of nitrile imines 8 (Scheme 4). However, it
has been pointed out that 7 should not possess long-wavelength
[18]
absorptions that could make it blue.
[
28]
George et al.
support the possibility of formation of
[14,16,17,19,20]
blue materials have failed.
Of course, it is possible
Trozzolo’s zwitterion in the photolyses of sydnones, but unfor-
tunately a clear conclusion was not reached in this work. We
have confirmed that a blue coloration forms rapidly on photoly-
sis at room temperature, but not on warming of cold, previously
irradiated sydnone matrices. The blue colours disappear when
leaving the material in the dark at room temperature, or more
rapidly on heating. Ketenes have not been detected in our low
temperature matrices by IR spectroscopy.
that the blue coloration of solids is a surface effect, perhaps
only a few nanometers deep, and that the quantum yield is very
low in those cases where photochromism is observable in
solution. Yet, an understanding of the mechanism of the blue
photochromism is lacking.
Ketene-type canonical structures have also been proposed
for six-memberd mesoionic rings (pyrimidinylium and oxazi-
nylium olates), and several of these were demonstrated to
Sydnones undergo thermal 1,3-dipolar cycloadditions across
the N(2)N(3)C(4) moiety, for example with acetylene dicarbox-
[
21–23]
undergo thermal valence isomerization to ketenes.
[
29]
Ketenes formed by photochemical ring opening of other five-
membered mesoionic ring compounds have only been observed
directly by matrix-isolation IR spectroscopy in a few cases, such
ylates.
regarded as an azomethine imine. The cycloadduct rapidly
extrudes CO with formation of pyrazoles. The photolysis of
This NNC moiety in sydnones can therefore be
2
[
24]
as pyrrolopyridinylium olates 4
and 3,1,2-oxathiazolinylum
sydnones in the presence of acetylene and alkene dipolarophiles
also gives pyrazoles in up to 90 % yields, but with a different
structure, as a result of initial photofragmentation of the syd-
[
25]
olates 5 (Scheme 3). In the case of the m u¨ nchnones 6, which
are the closest relatives of sydnones, chemical evidence both for
and against valence isomerization to ketenes (Scheme 3) has
been reported, and direct spectroscopic detection is lacking.
We have observed ketene bands in the IR spectra following
[
30–32]
nones to nitrile imines 8 (Scheme 5).
undergo 1,3-dipolar cycloaddition (‘click’ chemistry) not only
The nitrile imines
[
26]
with alkenes and alkynes, but with a variety of other dipolar-
[30,32]
[
23]
matrix photolysis of m u¨ nchnones in this laboratory,
but a
ophiles A¼B to afford compounds 10 (Scheme 5).
Thus,
1
4
14
conclusive identification of these ketenes awaits more detailed
investigation. However, several other types of five-membered
mesoionic compounds have been found to undergo thermal or
reaction with C-labelled CO yields C-labelled 2-phenyl-
2
[
33]
[30,32]
1,2,3-oxazol-5(2H)-one,
[
and addition to CS ,
2
30]
[32]
nitriles, and phenyl isocyanate occurs analogously. Intra-
molecular trapping to produce 1,2-diazepines has also been
[
27]
photochemical ring opening.
[
34]
Several other intermediates have been proposed as alterna-
tives to, or formed in parallel to, the ketenes 3. Trozzolo and co-
workers investigated the irradiation of 3-pyridylsydnone in a
KBr disc at 77 K by means of UV spectroscopy and observed
reported.
2,5-Disubstituted tetrazoles 9 are frequently used
as precursors of nitrile imines, and the same cycloaddition
products are obtained (Scheme 5). Indeed, the similarity with
tetrazole photochemistry is often used as evidence for the
formation of nitrile imines from sydnones.
[
19]
that a colourless species was formed first. This was assigned

Sydnone Photochemistry
459
R
Rꢂ
X
Y
X
ꢁ
Y
N
N
Y
N
ꢀ
O
ꢃ
N
N
COO
O
N
hν
ꢃCO2
Rꢂ
X
Y
O
O
AꢄB
N
ꢃ ꢁ
N
13
ꢃ ꢁ
ꢀ
2
Rꢂ
C
N
N
R
Y
C
N
N
X
N
B
N
O
ꢃ
R
O
Y
O
hν
ꢃN2
8
A
O
Rꢂ
X
C
1
0
N
N
ꢁN O
ꢁ
N
N
N
N
R
N
9
ꢃ
1
,3 -dipolar cycloaddition
O
X
1
4
Click chemistry
15
Scheme 5. Formation of nitrile imines 8 and cycloadducts 10 from
Scheme 8. Gotthardt’s proposed intermediates 13–15. X ¼ NR ; Y ¼ RS.
2
sydnones and tetrazoles.
1
sydnones just fluoresce (from the S p,p* state, calculated
energy 78.7 kcal mol for diphenylsydnone), but do not react.
However, blue photochromism of 3-phenylsydnones in frozen
1
R
Rꢂ
ꢀ1
N
?
ꢀ
N C N Rꢂ
N
O
R
O
[18]
solution at 77 K has been reported.
very efficient intersystem crossing (ISC) to the p,p* state
(
Eber et al. also reported
3
2
11
ꢀ
1
calculated energy 48.05 kcal mol for diphenylsydnone) and
concluded that it is this triplet state that undergoes photoreac-
Scheme 6. Are carbodiimides 11 formed in sydnone photochemistry?
[37]
[28,31a]
tion.
However, George et al.
concluded that the long-
Rꢂ
Rꢂ
R
Rꢂ
lived triplet states (lmax 380–400 nm; t ¼ 0.2–2 ms) are not
involved in the photochemical transformation to nitrile imines 8.
It is worth noting that nitrile imines can be either allenic or
R
hν
N
N
?
?
ꢃ
ꢁ
N
Rꢂ
C
N
N R
O
N
R
ꢀ
N
N
O
O
O
[41]
propargylic. The nitrile imines mentioned above are depicted
in the allenic form. Simple nitrile ylides such as PhCNNH and
HCNNPh are allenic, but C,N-diarylnitrile imines ArCNNAr are
2
1
12
8
Scheme 7. The usually assumed mechanism of sydnone photochemistry
and nitrile imine formation. Compounds 1 and 12 have never been observed.
[41]
propargylic. Some nitrile imines can exist in both allenic and
41]
[
propargylic states separated by small energy differences.
Thus, in summary, there are several sydnone problems, some
of which can be formulated as follows:
Thus the formation of nitrile imines 8 in sydnone photolysis
is not doubted. However, tetrazoles are known to afford both
nitrile imines and carbodiimides readily, and the nitrile imines
rearrange easily to carbodiimides. Why have carbodiimides 11
never been reported in sydnone photochemistry (Scheme 6)?
UVspectroscopic investigations using laser flash photo-
1. Is Earl’s bicyclic lactone 1 formed and observable?
2. What is the mechanism of formation of nitrile imines 8?
3. Why have carbodiimides 11 not been observed in sydnone
chemistry?
[
28,35–37]
lysis (LFP)
imines, which have the same UV spectra as those formed by
photolysis of 2,5-disubstituted tetrazoles 9 (typical l for C,
have corroborated the formation of nitrile
4
. What is the nature of the blue photochromic intermediate?
In this paper we will show that Earl’s bicyclic lactones 1 and
carbodiimides 11 are the principal products of the matrix
photolysis of sydnones.
max
[
28,35,38,39]
N-diarylnitrile imines 375–380 nm).
nitrile imines from sydnones is frequently postulated to take
place via formally anti-aromatic 1-substituted diazirenes 12
The formation of
Results
(
Scheme 7) formed by loss of CO from the bicyclic species
2
[
11,19,30,33,40]
1
.
In spite of numerous investigations, neither 1
nor diazirenes 12 have ever been observed in sydnone
photochemistry.
Other potential routes to the nitrile imines via the diazir-
enium carboxylate 13, the azoxyketene 14, and the four-
membered oxazete zwitterion 15 were postulated by Gotthardt
The photochemistry of three sydnones was investigated by
means of IR spectroscopy of the species isolated in noble
gas matrices (Ar or Xe) at ,12–80 K. The compounds are
3-phenylsydnone 16, 3-(3-pyridyl)sydnone 19 (in the following
called 3-pyridylsydnone), and 3,4-diphenylsydnone 22.
[
11]
3-Phenylsydnone 16
et al. on strong chemical grounds (Scheme 8).
In the laser photolysis of arylsydnones, the arylnitrile imines
were formed within the duration of the nanosecond laser
pulse, with the corollary that the postulated intermediate(s)
must be very short lived or not involved at all. The quantum
[28,37]
The crystalline 2-phenylsydnone features a very strong carbonyl
ꢀ1
ꢀ1
absorption at 1760 cm in a KBr disc and at 1758 cm when
isolated as a neat solid at 50 K (Fig. S1, Supplementary Mate-
rial). When the compound is isolated in an Ar matrix at 12 K, two
[
28]
ꢀ1
yields of nitrile imine formation are low (0.05 in benzene).
Additional transient intermediates with lmax up to 600 nm
and t ¼ 0.3–4 ms in benzene) were detected in the LFP of
peaks are observed at 1795 and 1762 cm (Fig. 1 and Figs S1–
S7, Supplementary Material). This is most probably a result of
matrix site effects, since annealing to 30 K causes changes in the
relative intensities of the two bands (Figs S2 and S3, Supple-
mentary Material). A similar phenomenon is observed in xenon
3
conclusion as to their identity was reached. The initial photo-
-arylsydnones carrying H or CH in position 4, but no firm
3
[
28]
excitation of sydnones is believed to lead to the S state, which
1
matrices, where the two carbonyl bands appear at 1787 and
ꢀ1
3
decays rapidly to the p,p* triplet state.
[37]
The quantum yield
1757 cm
(Figs S4 and S5, Supplementary Material).
of photoreaction in solution decreases dramatically with
decreasing temperature, and the reaction has essentially stopped
at 200 K, but the quantum yield of fluorescence and the
fluorescence lifetime go up. So, at lower temperatures, the
Furthermore, the relative intensities of the two bands depend on
the exact conditions of matrix deposition(compare, for example,
Fig. 1 and Fig. S7, Supplementary Material). The site respon-
ꢀ
1
sible for the high frequency band (1795 (Ar) or 1787 (Xe) cm
)

4
60
R. N. Veedu, D. Kvaskoff, and C. Wentrup
(a)
1795
Ph
N
C
NH
1762
18
(a)
CO2
7
19
1
476
1
447
761
49
1173 1017
1
506
9
1
227 1080
683
B
C
B
(b)
CO2
A
(b)
B
S
S
S
S
S
(
c)
d)
B
O
C
Ph
N
S
(
N
O
B
exo-17
C
(c)
500
(e)
2
2000
1500
ν [cmꢃ1
]
1000
500
4000
3000
2000
1500
1000
450
[
cmꢃ1
]
Fig. 2. (a) Calculated IR spectrum of N-phenylcarbodiimide, Ph-
N¼C¼NH, 18. (b) IR spectrum from photolysis of 3-phenylsydnone 16 in
Ar matrix at l ¼ 310–390 nm for 100 min. Negative bands: S ¼ sydnone 16.
Fig. 1. Photolysis of 3-phenylsydnone 16 at 310–390 nm (Ar matrix,
2 K). (a) IR spectrum of the deposited sample. (b) After 1 min irradiation.
c) After 3 min irradiation. (d) After 15 min irradiation. (e) After 100 min
1
ꢀ1
Positive bands: B ¼ bicyclic lactone exo-17, 1881 cm . C ¼ phenylcarbo-
(
ꢀ1
ꢀ1
diimide 18, 2163, 2128 cm . (c) Calculated IR spectrum of the bicyclic
lactone exo-17. All calculations at the B3LYP/6–31G* level with wave-
numbers scaled by 0.9613.
irradiation. Note that the most reactive site absorbing at 1795 cm dis-
ꢀ1
appears first. Products: B (1881 cm ) ¼ bicyclic lactone exo-17; C (2128
ꢀ1
and 2163 cm ) ¼ phenylcarbodiimide 18.
ꢀ
1
is the more reactive photochemically (see Fig. 1). Such site-
[
dependent photochemistry is known for other compounds,
and it is likely that the more reactive site at the higher frequency
corresponds to discrete, matrix-isolated molecules; the lower
frequency site may correspond to less well matrix-isolated
molecules or clusters.
calculated at 2144 cm and the combination n þ n , which
ꢀ1 ꢀ1
3 17
3
17
25]
falls within 9 cm at 2135 cm . The n and n modes involve
coupling between a ring deformation and the NCN stretch.
As expected, CO2 was formed in increasing amounts
[41]
throughout the photolysis (Figs 1 and 2). N-Phenylnitrile imine
2
1
8 (R ¼ Ph, R ¼ H) would be a logical precursor of the carbo-
[41]
Photolysis of 16 in an argon matrix at 310–390 nm resulted in
the eventual disappearance of both carbonyl bands and forma-
tion of new products (Fig. 1). Owing to the excellent agreement
with the calculated IR spectrum at the B3LYP/6–31G* level, the
diimide,
weak peak was observed at 2099 cm during the first 10 min of
but it was not observed in this photolysis. A very
ꢀ1
photolysis, but this does not match expectations for 8, which has
ꢀ1 [38b]
been observed previously in a PVC matrix at 2014 cm
and the best calculated, anharmonic absorption maximum is at
,
ꢀ1
major product with a strong absorption at 1881 cm is assigned
as the exo-bicyclic lactone 17 (Fig. 2). The high frequency is in
agreement with a carbonyl group in a strained, four-membered
ring. For example, benzopropiolactone (benz[b]oxete-2-one)
ꢀ1 [41]
2009 cm
. The isomeric C-phenylnitrile imine HNNCPh
was found to isomerize very rapidly to carbodiimide 18 on
matrix photolysis at 254 nm.
The reversibility of the formation of the bicyclic lactone 17
was probed by first preparing a matrix of the sydnone, irradiat-
ing it at 308 nm for 60 min to convert it completely into products
as in Fig. 2, and then irradiating the products at 222 nm for up to
2 h. This caused the 1881 cm band ascribed to lactone 17 to
decrease, and those of sydnone 16 (1795 cm ), carbodiimide
ꢀ1 ꢀ1
18 (2129, 2166 cm ), and CO (2343 cm ) to increase (Fig. 3).
2
As mentioned above, the 1795 cm band of the sydnone
corresponds to discrete sydnone molecules, and it is also the
one regenerated on photolysis at 222 nm.
When an Ar matrix of sydnone 16 was irradiated at 222 nm
from the beginning, only the carbodiimide 18 and CO were
[
41]
ꢀ
1
[42]
ꢀ1
absorbs at 1904 cm
(calculated 1901 cm
at the
B3LYP/6–31G* level), and we calculate the carbonyl absorp-
ꢀ1
tion in 4-methylpropiolactone at 1866 cm (B3LYP/6–31G*;
wavenumbers at this level are scaled by a factor 0.9613
throughout this paper). In addition, the matrix photolysis of 16
afforded another product with absorptions at 2163 and
ꢀ1
ꢀ1
ꢀ
1
2
128 cm (Figs 1 and 2). Both of these bands are readily
ꢀ
1
assigned to N-phenylcarbodiimide, PhN¼C¼NH 18, which
we have characterized recently as a product of matrix photolysis
the double
[
41]
[41]
of 5-phenyltetrazole.
As shown previously,
ꢀ1
band for 18 in the 2100 cm region arises from Fermi resonance
between the NCN symmetrical stretching vibration n22
2

Sydnone Photochemistry
461
3-Py-NꢄCꢄNH 21(C)
CO2
S
(
a)
b)
C
exo-20 (B)
A
(
CO2
B
1886
4000
3000
2000
1500
1000
440
B
2
135
[
cmꢃ1
]
C
(c)
Fig. 3. IR-difference spectrum showing the disappearance of the band of
ꢀ1
the bicyclic lactone 17 at 1881 cm (B; negative peak) on irradiation at
ꢀ
1
S
S
S
2
22 nm in Ar matix at 12 K. Products formed (positive peaks): S (1795 cm )
ꢀ1
active site of matrix-isolated sydnone 16; C (2128 and 2163 cm ) phenyl-
carbodiimide 18; and CO . 17 was originally formed on photolysis of 16 at
08 nm.
S
2
3
S
Ph
H
hν
308 nm
O
3000
2600
2200
1800
1400
1000
600
Ph
N
N
ꢀ
ν [cmꢃ1
]
N
N
O
O
O
222 nm
exo-17
16
Fig. 4. (a) Calculated IR spectrum of 3-pyridylcarbodiimide 21 (C).
(b) CalculatedIRspectrumofthebicycliclactoneexo-20 (B).(c)IR-difference
2
22 nm
spectrum from photolysis of 3-pyridylsydnone 19 in Ar matrix (12 K) at
ꢀ1
N C
N
H
3
08 nm for 20 min. Negative bands are due to 19 (S, 1799, 1780, 1765 cm );
ꢀ1
Ph
1
8
positive bands:
ꢀ1
B
(1886 cm ) ¼ bicyclic lactone exo-20,
C (2135,
ꢁ
CO
ꢀ1
2
2
148 cm ) ¼ 3-pyridylcarbodiimide, and CO
2
(2340 cm ). These calcula-
tions are at the B3LYP/6–31þG** level, wavenumbers scaled by 0.964.
Scheme 9. Photochemistry of 3-phenylsydnone 16.
ꢀ
1
at 1626 (669) and 1933 cm (780 km mol ). Likewise, there
ꢀ1
ꢀ
1
formed (Fig. S7, Supplementary Material). The 1881 cm band
was not observable under these conditions.
was no evidence for the formally anti-aromatic 1H-diazirene 12
calculated C¼N stretch 1718 cmꢀ (11 km mol )) or the
1
ꢀ1
(
diazirenium carboxylate 13 (1746 cm (747 km mol )).
In summary, photolysis of 3-phenylsydnone 16 gives rise to
the bicyclic lactone valence isomer exo-17 together with phe-
nylcarbodiimide PhN¼C¼NH 18. The bicyclic lactone exo-17
regenerates the sydnone on photolysis at 222 nm, but decompo-
ꢀ1
ꢀ1
3-Pyridylsydnone 19
sition to the carbodiimide 18 and CO also takes place under
2
The IR spectrum of solid 3-pyridylsydnone in a KBr disc
ꢀ1
these conditions. In other words,
exhibits a strong carbonyl band at 1742 cm with two shoulders
on the low-frequency side. These bands are resolved in an Ar
ꢀ1
1
6 þ hn (l ¼ 308 nm) - exo-17 þ 18 þ CO ; exo-17 þ hn
matrix, where they appear at 1799, 1780, and 1765 cm
(Fig. S8, Supplementary Material). Photolysis of this matrix at
08 nm caused the intensity of the high-frequency band at
2
(
l ¼ 222 nm) - 16 þ 18 (Scheme 9).
3
ꢀ1
There was no evidence in these experiments for the for-
mation of nitrosaminoketene 3, diazonium derivatives 7, 13,
1799 cm to decrease rapidly, with the other two bands dis-
appearing more slowly (Fig. S9, Supplementary Material). The
resulting spectrum features two important bands at 1886 and
0
and 15, or azoxy compound 14 (R(X) ¼ Ph, R (Y) ¼ H). s-E-
ꢀ
1
ꢀ1
Nitrosaminoketene 3 has a strong calculated C¼C¼O stretching
2135 cm , the latter with a shoulder at 2148 cm , which are
assigned to the bicyclic lactone exo-20 and the carbodiimide 21,
respectively (Scheme 2), based on the good agreement with the
calculated spectra (Fig. 4) and analogy with the photochemistry
of the phenyl derivative 16 detailed in the preceding section. As
described for phenylcarbodiimide 18 above, arylcarbodiimides
will often exhibit double bands attributable to coupled vibra-
ꢀ
1
ꢀ1
vibration at 2144 cm (759 km mol ); the s-Z-isomer is not a
calculated energy minimum, as it cyclizes to sydnone 16 without
a barrier (this is described further in the ketene and photochro-
mism section below). Azoxy-ketene 14 is predicted at 2151 or
ꢀ ꢀ1
1
2
160 cm (918 and 2160 km mol , respectively) for the s-E
and s-Z isomers. All observed absorptions in this region are
accounted for by the carbodiimide 19. The diazonium carbo-
xylate 7 has the strongest calculated absorption at 1192 (1943),
[
41]
tions engaging in Fermi resonance,
as has been observed in
but as this depends on coupling with ring
deformations, it can be influenced by substitution. Thus,
[43]
several cases,
ꢀ ꢀ1
1
1
793 (517), and 1921 cm (115 km mol ), and those for 15 are

4
62
R. N. Veedu, D. Kvaskoff, and C. Wentrup
hν
0.2
3-Py
H
CO2
3
08 nm 3-Py
N
O
1892
1874
N
ꢀ
N
O
N O
O
222 nm
2141
150
2117
2
exo-20
1
9
hν
hν ꢃCO2
2
22 nm
0
N
C
N
H
ꢁ CO2
2243 2237
3-Py
21
A
Scheme 10. Photochemistry of 3-pyridylsydnone 19.
ꢃ0.2
ꢃ0.4
anharmonic frequency calculations predict a strong fundamental
band for NCN stretching in 3-pyridylcarbodiimide 21 at
1
742
ꢀ
1
2
137 cm (n
tones at higher wavenumbers. The characteristic shape of the
) and weak combination bands and over-
N¼C¼N
[41]
1
787
ꢀ1
ꢀ1
band (2135 cm with a shoulder at 2148 cm ) is in good
ꢀ1 [44]
2400 2200 2000 1800 1600 1400 1200 1000 800 600
agreement with the literature values of 2136 and 2149 cm
.
[
cmꢃ1
]
The ratio of the photoproducts exo-20 and 21 depends on the
photolysis wavelength. A relatively larger proportion of the
bicyclic lactone exo-20 is obtained when irradiating at 310–
Fig. 5. IR-difference spectrum showing disappearance of 3,4-diphenyl-
sydnone 22 (negative peaks) on irradiation in Ar matrix at 222 nm for 2 h.
3
compelling evidence for the formation of ketenes or nitrile
90 nm (Fig. S9, Supplementary Material). There was no
ꢀ1
Products (positive peaks): 1874 cm bicyclic lactone exo-23 (tentative
ꢀ1
assignment); 1892 cm : 1-azacyclohepta-1,2,4,6-tetraene 29; 2117, 2141,
ꢀ1
1
2
imines (3 and 8, R ¼ 3-pyridyl, R ¼ H) in any of these photo-
ꢀ1
and 2150 (shoulder) cm : diphenylcarbodiimide 24; 2237 cm : benzoni-
ꢀ1
ꢀ
1
lyses. A very weak absorption appeared at 2088 cm during the
first few minutes of irradiation, and another very weak band
trile 27; 2243 cm : diphenylnitrile imine 26 (tentative assignment); CO
2
ꢀ
1
(2340 cm ).
ꢀ1
developed at 2176 cm , but neither is in good agreement with
ꢀ
1
Ph
O
the calculated maxima for the nitrosaminoketene 3 (2144 cm
1
)
Ph
Ph
Ph
N
O
ꢀ
hν
N
or the nitrile imine 8 (2055 cm ), and the intensities are far too
low to allow any positive assignment.
N
C
N
Ph ꢁ CO2
ꢀ
N
O
N
Ph
O
24
Further irradiation of a matrix containing exo-20 and 21 (e.g.
a matrix corresponding to Fig. 4) at 222 nm caused strong
exo-23
2
2
hν
ꢀ
1
)
reduction of the intensity of the lactone exo-20 (1886 cm
and increased absorptions ascribed to carbodiimide 21
Ph
N
hν
ꢁ ꢃ
ꢀ1
N
N
Ph
C
N
N
Ph
(
Thus, the photochemistry of 3-pyridylsydnone 19 can be sum-
marised as follows:
2135 cm ) and CO2 (Fig. S11, Supplementary Material).
Ph
ꢃN2
N
2
6
2
5
1
9 ¼ hn (l ¼ 308 nm) - exo-20 þ 21; exo-20 þ hn (222 nm)
2
CN ꢁ
N
-
(
[sydnone 19] - carbodiimide 21 þ CO or exo-20 þ hn
27
28
222 nm) - carbodiimide 21 þ CO (Scheme 10).
2
A blue coloration always develops when 3-pyridylsydnone is
irradiated at room temperature. This coloration was also
observed on the material deposited on warm (at or near room
temperature) surfaces of the matrix isolation equipment
N
29
(see Fig. S10, Supplementary Material), but never in the cold,
irradiated Ar matrices. In order to test Trozzolo’s assertion
[
19]
Scheme 11. Photochemistry of 3,4-diphenylsydnone 22.
that a blue colour develops on warming a colourless, irradiated
sample, 3-pyridylsydnone 19 was deposited neat (without a
matrix host) at 77 K and irradiated at 308 nm. This afforded the
bicyclic lactone exo-20 and the carbodiimide 21 in low yields as
30 min caused decomposition with formation of CO , but no
2
discrete photoproducts were discernible (Fig. S12, Supple-
mentary Material).
ꢀ1
observed by IR spectroscopy (1873 and 2131 cm , respec-
tively) but it was not possible to achieve complete conversion
into products under these conditions. The irradiated sample
remained colourless. The sample was then warmed slowly to
room temperature in the dark. At no stage did a blue coloration
appear, nor was there any evidence for the formation of
ketenes, or the zwitterion 7, while the absorptions at 1873 and
3
,4-Diphenylsydnone 22
Diphenylsydnone 22 shows a strong C¼O absorption at
ꢀ1
1748 cm in a KBr disc, and two bands at 1787 and 1742 cm
ꢀ1
in an Ar matrix (Fig. 5 and Fig. S13, Supplementary Material).
Photolysis at 310–390 nm was slow but resulted in the formation
ꢀ1
of CO (2340 cm ) as well as development of weak bands at
2
ꢀ
1
ꢀ1 ꢀ1
2
131 cm disappeared.
2117 and 2141 cm with a shoulder at 2150 cm (Fig. 5).
41]
[
An intense blue coloration was observed when a KBr disc of
-pyridylsydnone 19 was irradiated at 308 nm at room tem-
These bands are attributable to diphenylcarbodiimide 24.
Further irradiation at 254 or 222 nm caused a further increase in
these bands as well as the formation of very weak absorptions at
3
perature, but the IR spectrum did not show any evidence for the
presence of a ketene. Not surprisingly, the bicyclic lactone
exo-20 was not observed either. Continued irradiation for
ꢀ
1
[41]
2237 cm attributable to benzonitrile 27
ꢀ1
and at 1874 and
1892 cm (Fig. 5). The 1892 cm absorption corresponds to
ꢀ1

Sydnone Photochemistry
463
C
1454
A
1283
1078
1169
749
1069
A
1216
1595 1488
1341
N
2750
2400
2000
1800
1600
1400
1200
1000
800
600 500
[
cmꢃ1
]
ꢀ1
Fig. 6. IR-difference spectrum showing diphenylnitrile imine 26 (N, 2243 cm ; negative peaks), obtained by photolysis
of 2,5-diphenyltetrazole 25 at 254 nm in Ar matrix at 12 K, and its conversion to diphenylcarbodiimide 24 (C, 2117 and
ꢀ1
2
141 cm ; positive peaks) on subsequent photolysis at 338–485 nm. A very weak peak due to 1-azacyclohepta-1,2,4,6-
ꢀ1
ꢀ1
tetraene 29 (A, 1892 cm ) is also seen. A weak peak due to benzonitrile 27 (2237 cm ) is hidden under the strong nitrile
imine peak. Ordinate: absorbance in arbitrary units. See also Ref. [41].
ꢀ1
1
product of phenylnitrene 28.
-azacycloheptatetraene 29, which is the known ring expansion
[
mol at the B3LYP/6–31þG** level. This is a substantial
barrier, and it is no surprise, therefore, that this chemistry is not
observable by thermal means: Flash vacuum thermolysis (FVT)
of 16 at 7008C affords a complex mixture of products, from
45]
The formation of 1-azacycloheptatetraene 29 and benzoni-
trile 27 is very important. These compounds have been observed
as the products of cleavage of diphenylnitrile imine 26 by
Toubro and Holm, who photolyzed 2,5-diphenyltetrazole 25 in
which only benzene and CO were identified; diphenylsydnone
2
[
46]
yielded benzonitrile and CO as the major products at 4708C.
2
[38b]
ꢀ1
a PVC film.
photolysis of an Ar matrix of 26.
We have made the same observation on
[
The thermal barrier of 70 kcal mol corresponds to a wave-
length of 407 nm, which is close to the upper limit usually used
in sydnone photochemistry. Wavelengths of 254–366 nm are
41]
Therefore, although the
bands of 27 and 29 are very weak in the sydnone photolyses, they
provide strong evidence that diphenylnitrile imine 26 is actually
formed as an intermediate. A strong IR spectrum of diphenylni-
trile imine can be obtained by photolysis of an Ar matrix of
ꢀ
1
more usual (113–78 kcal mol ), and this will presumably
[37]
1
populate the excited p,p* singlet state.
The energy of the
[37]
1
ꢀ1
p,p* state was calculated as 78.7 kcal mol (364 nm).
1-Substituted diazirenes like 31 have often been postulated
[
41]
tetrazole 25 at 254 nm (Scheme 11 and Fig. 6).
This is not
achievable on analogous photolysis of diphenylsydnone 22. The
reason may be that the photolysis of the sydnone is much slower
than that of the tetrazole, so that the nitrile imine 26 may
rearrange to diphenylcarbodiimide 24 faster or at almost the
same rate as it is formed. Furthermore, carbodiimide 24 may
form by other routes from the bicyclic lactone exo-23
as intermediates in the solution photochemistry of sydnones,
[
28,37,47]
although they have never been observed.
The formation
of the diazirene 31 plus CO2 from sydnone 16 is nearly
thermoneutral at the B3LYP level (Scheme 12).
The formation of diazirenes has been claimed in matrix
[
48]
photolyses of a few 1,5-disubstituted tetrazoles
lated in several other cases under both thermal and photoche-
mical conditions,
and postu-
(
Scheme 11). These issues are discussed further below. The
ꢀ1
[39,41,44,49]
weak peak at 1874 cm (Fig. 5) does not belong to 29; it can be
assigned to the bicyclic lactone exo-23 owing to an excellent
agreement with the strongest calculated absorption for this
but more detailed photochemistry
to elucidate the relationship between diazirenes, imidoylni-
trenes, nitrile imines, and carbodiimides is required. There is
evidence for the involvement of diazirenes and/or imidoy-
ꢀ1
compound at 1876 cm . Finally, a very weak shoulder on the
ꢀ
1
benzonitrile peak at 2243 cm (Fig. 5) indicates that a small
lnitrenes in the thermal rearrangement of nitrile imines to
[50]
ꢀ
1 [41]
amount of diphenylnitrile imine 26 (2243 cm
be present. Diphenylnitrile imine 26 has the propargylic struc-
)
may in fact
carbodiimides.
1,5-disubstituted tetrazoles in solution leads to nitrile imines,
The photolysis of sydnones as well as
ture shown (Scheme 11), whereas monoarylnitrile imines have
[
allenic structures.
which have been observed by UV spectroscopy and trapped
[37,47]
41]
with dipolarophiles.
3
The calculated activation barrier for
1
ꢀ
1 - 32 is ,36 kcal mol (Scheme 12). It is emphasized that
Mechanisms of Carbodiimide and Nitrilimine Formation
the calculations in Scheme 12 pertain to the ground state energy
surface, whereas the reactions are actually photochemical, so for
a complete description calculations of the excited-state energy
surface will be required.
The observed photochemistry is rationalized in Scheme 12 using
3-phenylsydnone 16 as a model. The cyclization to the bicyclic
lactone exo-17 has a calculated activation barrier of 70 kcal

4
64
R. N. Veedu, D. Kvaskoff, and C. Wentrup
ꢃ
Ph
H
Ph
COO
O
ꢁ
N
N
N
hν
?
ꢁ
Ph
N
N
O
ꢃ
O
N
O
16
endo -17
33
0
33.0
36
hν
matrix
(TS 70.2)
TS 62.6
TS 36.2
TS 44.7
H
O
Ph
N
N
TS 36.0
TS 44.7
ꢁ
ꢃ
ꢁ
CO2
Ph
N
Ph
N
N
C
H
ꢁ CO2
O
N
3
1
32
ꢃ12.4
exo-17
9.3
3
5.1
(observed)
TS 5.9
H
3
4Z S1 ꢃ2.8
N
N
hν
34Z T0 ꢃ10.6
34E S1 ꢃ0.7
34E T0 ꢃ9.8
Ph
N
N
ꢃN2
N
N
Ph
3
4
3
4E S2 ꢃ0.3
37 (R ꢄ H)
←
TS (34Z S1 18) 13.7
←
TS (34E S1 18) 18.7
N
C
N
H
ꢁ CO
2
Ph
1
8
ꢃ60.8
observed)
(
Scheme 12. Reaction scheme for 3-phenylsydnone 16 with calculated energies of ground and transition states in
ꢀ1
kcal mol at the B3LYP/6–31þG** level. Very similar energies apply to the 3-pyridyl system 19, 20, 21 etc. The
2
energy for TS(31–34 S1) has a , S . value of 0.214 indicating minor spin contamination.
N
R
ꢁ
N
ꢃ
N
N
hν or Δ
ꢃN2
R
C
N
N
Ph
Ph
35
36
hν or Δ
R
N
Ph
N
N
R
R
hν or Δ
Ph
39
N
N
N
C
N
R
ꢃN2
Ph
N
N
N
Ph
40
37
38
Scheme 13. Formation of nitrile imines 36 and carbodiimides 40 from terazoles (R ¼ H, alkyl, aryl).
[
41]
The matrix photolysis of 2-substituted tetrazoles affords
photochemically – but not the other way round
(Schemes
[
41]
nitrile imines, e.g. 35 - 36 (Scheme 13). The corresponding
matrix photolysis of 1-substituted tetrazoles affords carbodii-
mides, e.g. 37 - 18 (Scheme 12) and 37 - 40 (Scheme 13) via
12 and 13). The photochemical isomerization of diphenylnitrile
imine 26 to diphenylcarbodiimide 24 (Fig. 6) is an example.
The endo isomers such as endo-17, endo-20, and endo-23
have not been observed in this work; the IR spectra reported
above are in good agreement with the calculated spectra of the
exo isomers (exo-17, exo-20, and exo-23) but do not agree with
those of the endo isomers. The calculated structure of endo-17 is
[
41]
the diazirene 38 and/or the imidoylnitrene 39 (Scheme 13).
The energy of the open-shell singlet nitrene 34 (S , Scheme 12)
1
[
51]
has been estimated using the Cramer–Ziegler correction.
summary, the formation of carbodiimides, e.g. 18, 21, 24, and
0, is preferred on energetic grounds. Moreover, nitrile imines
isomerize efficiently to carbodiimides, both thermally and
In
˚
4
unusual, having an extremely long N—O bond (2.14 A at the
B3LYP/6–31G* level), and the N-aryl substituent is nearly in

Sydnone Photochemistry
465
3
Py
UV-Vis spectrum
O
1
1
2000
0000
0.18
.16
0.14
0.12
O
N
3
Py
N
0
44
ꢁ
ꢃ
N
O
N
O
8
6
4
2
000
000
000
000
0
exo-20
7.9
endo-20
0
0
.10
.08
3
33.7
0.06
7
6.7
68.4
0
0
0
.04
.02
3
Py
TS
3Py
600
500
400
300
200
100
0
N
H
O
N
N
H
N
Excitation energy [nm]
O
C
O
C
Fig. 7. Calculated UV-vis spectrum of s-E-nitrosaminoketene 42 (TD-
B3LYP/6–31þG**).
O
41
42
24.3
photoreaction with decreasing temperature between room tem-
perature and ,160 K. In contrast, aryl- and diaryl-tetrazoles
readily give nitrile imines under these conditions. The diphe-
nylsydnones do decompose readily in solution at room temper-
3
Py
H
N
ꢀ
N
O
O
ature at 366 nm, but, importantly, the authors did not state
[31a]
1
9
0
that nitrile imines are formed.
3-Phenylsydnone did react
photochemically on irradiation at 320 nm in organic glasses at
7 K to produce a spectrum with a broad UV maximum
around 320 nm (sic!) extending into the visible, but the forma-
7
Scheme 14. Potential formation of nitrosaminoketene 42 (energies in
ꢀ1
kcal mol at the B3LYP/6–31G* level).
[31a]
tion of a nitrile imine was not unequivocally established.
It is
highly desirable to reinvestigate the solution photochemistry
of sydnones and determine their fate without added dipolaro-
philes. Also, the solution photochemistry of sydnones should be
investigated with modern time-resolved IR spectroscopy to
observe and differentiate nitrile imines, carbodiimides, and
indeed ketenes.
the plane of the diazirene ring (see Fig. S14, Supplementary
Material). In effect this species is better described as an inner
complex of a diazirenium carboxylate (Scheme 12). Breaking
[11]
this long (and weak) bond leads to Gotthardt’s
zwitterionic
diazirenium carboxylate 33, which is predicted to lose CO very
2
easily to form the nitrile imine 32 directly. The structure of
endo-17 is more normal at the QCISD(T) level with a still long
˚
N—O distance of 1.57 A and the N-aryl group being more
The Ketenes and the Blue Photochromism
decidedly endo oriented (see Fig. S15, Supplementary Material).
The formation of endo isomers such as endo-17 may be
sterically hindered in the rigid matrix environment (and in
frozen solution), as it necessitates a wide-angle motion of the
phenyl group.
The question of the nitrosaminoketenes 3 remains. We have
never observed them in this work, and the evidence for their
existence is weak (see Introduction). It would be possible in
principle to form ketenes such as 41 and 42 on energetic grounds
(Scheme 14) but the overall calculated activation barriers are
significantly higher than those for the formation of carbodii-
mides and nitrile imines (Scheme 12). Therefore, it may be
difficult to find reaction conditions that will permit the clearcut
observation of the ketenes. The cisoid nitrosaminoketene 41
cyclizes back to the sydnone without a barrier on optimization of
the geometry (Scheme 14).
An explanation of the differences between solution and
matrix photochemistry of sydnones is desirable. Nitrile imines
are observable and trappable in solution photochemistry, but in
matrix photochemistry they are at best observed as a minor
product in the case of diphenylnitrile imine 26 (Scheme 11;
Fig. 5). One explanation could be that nitrile imines rearrange to
carbodiimides faster or at the same rate as they are formed in the
sydnone photolyses. This is plausible, since we have found that
C-aryl nitrile imines are only observable during the first few
minutes of photolysis of tetrazoles at 254 nm, and they rearrange
Importantly, although the nitrosaminoketenes have been
[
16–18]
championed by Nespurek et al.
as carriers of the blue
colour formed on exposure of sydnones to UV light, our
calculations do not indicate that the ketenes could be blue; the
longest wavelength maximum for the ketene 42 derived from
3-pyridylsydnone is 405 nm (Fig. 7) in accord with the expecta-
tion that this compound should be yellow.
[
41]
very rapidly to carbodiimides at this wavelength.
The route endo-17 - 33 - 32 would provide a direct route
to nitrile imines in solution. As mentioned above, the formation
of the endo isomers may be hindered in the matrix and frozen
solution because of the constraint imposed by the solid matrix.
The route endo-17 - 31 - 34 - 18 is a low-energy route for
the formation of carbodiimides in the matrix; the formation of
nitrile imine 32 requires higher energy (Scheme 12).
In contrast, the calculated UV-vis spectrum of endo-17 at the
TD-B3LYP level has a maximum at 650 nm (Fig. 8), which
would make this compound blue. The corresponding maximum
for endo-20 is at 535 nm (see Supplementary Material for the
calculated UV-vis spectra of this as well as species 8, 14,
and 15). In effect, these long-wavelength absorptions can be
considered as charge-transfer transitions in the diazirenium
carboxylates. If the distance between the two charged centres
However, an explanation as to why carbodiimides are
readily observed in the matrices but never reported in solution
photochemistry is still needed. It is noted that diphenylsydnones
are not photoreactive at 366 nm in organic glasses (ether-
ꢀ
þ
(COO and NN ) is too long (as in 33, Scheme 12) charge
transfer will not occur. If it is too short, as when a covalent bond
is formed between N and O in the endo isomer, it can also not
[
31a]
isopentane-ethanol (EPA) and 2-MeTHF) at 77 K,
Eber et al.
and
found a rapidly decreasing quantum yield of
[37]

4
66
R. N. Veedu, D. Kvaskoff, and C. Wentrup
9000
8000
7000
6000
5000
4000
3000
2000
1000
0
0.20
0.18
0.16
0.14
0.12
0.10
0.08
0.06
0.04
0.02
0
1000
800
600
400
200
0
Excitation energy [nm]
Fig. 8. Calculated UV-vis spectrum of endo-17 (TD-B3LYP/6–31þG**).
occur (and for the same reason it can of course also not occur in
˚
Experimental
the exo isomer). It just happens that the ,2.1 A distances in
The sydnones were prepared according to literature proce-
[52]
dures. The apparatus for matrix isolation and photolysis was
as previously described.
endo-17 and endo-20 are just right. Naturally, this needs to be
confirmed at higher levels of theory, including multiconfigura-
tional ones. A further possibility is charge transfer between
neighbouring head-to-tail or sandwich oriented molecules in the
crystal or in solution, or charge-transfer transitions in the
mesoionic sydnone molecules themselves. One thing is clear
though: the photochromism is not a result of the supposed ketene
intermediates.
[53]
Samples were isolated in Ar or Xe
matrices on a KBr or CsI observation disc at 12–20 K, or neat at
2 or 77 K. Photolyses were carried out at 12 K using a low-
pressure Hg lamp (75 W, 254 nm; Gr a¨ ntzel, Karlsruhe), a high-
1
pressure Xe–Hg lamp (1000 W; Hanovia) equipped with a
ꢀ2
monochromator, or laser lamps operating at 222 (25 mW cm
ꢀ2
)
and 308 nm (50 mW cm ) (Ushio). Infrared spectra were
ꢀ
1
.
recorded at a resolution of 1 cm
Conclusion
Photolysis of 3-phenyl- and 3-(3-pyridyl)-sydnones 16 and 19 in
Ar matrices yields Earl’s bicyclic lactones (oxadiazabicyclo
Computational Methods
[2.1.0]pentanones) exo-17 and exo-20 as primary products and
carbodiimides 18 and 21 and CO as secondary products. Only
Calculations of structures, energies, and vibrational spectra in
the harmonic approximation were performed at the B3LYP/
2
the exo-lactones are observed. These reactions are very clean. In
the case of 3,4-diphenylsydnone 22 there is the further formation
of 1-azacyclohepta-1,2,4,6-tetraene 29 and benzonitrile 27 as
tell-tale products derived from diphenylnitrile imine 26, and a
very weak absorption at 2243 cm is in agreement with for-
mation of this nitrile imine as a transient.
6
–31G* or B3LYP/6–31þG** levels and in some cases the
[55]
[54]
QCISD/6–31þG** level as indicated using the Gaussian 03
or Gaussian 09
suites of programs. Electronic transitions
were calculated at the TD-B3LYP/6–31þG** level, and the
[57]
ꢀ1
[56]
UV-vis spectra were simulated using SWizard.
wavenumbers were scaled by a factor 0.9613.
Harmonic
The tem-
Carbodiimides have never been observed in the solution
photochemistry of sydnones, which instead leads to nitrile
imines, which have been observed directly by UV-vis spectro-
scopy and trapped in 1,3-dipolar cycloaddition reactions. Nitrile
imines rearrange very efficiently to carbodiimides on matrix
perature used was 298 K. All energies were corrected for
unscaled zero point vibrational energies (ZPVE). Transition
states were verified by intrinsic reaction coordinate calcula-
tions. Imaginary vibrational frequencies are listed in the Sup-
plementary Material.
[41]
photolysis. Thus the formation of carbodiimides in the matrix
photolyses may be a rapid two-step process via transient nitrile
imines. A reinvestigation of the solution photochemistry of
sydnones without added dipolarophiles is called for.
Supplementary Material
Additional matrix IR spectra, colour photographs of the blue
photochromic material, calculated UV-vis spectra, vibrational
frequencies, energies, and Cartesian coordinates for all calcu-
lated ground and transition state structures are available on the
Journal’s website.
The nitrosaminoketene valence isomers 3 of sydnones
Eqn 2), often postulated to be carriers of the blue photochro-
(
mism, are not predicted to have significant absorptions in the
visible spectrum (.400 nm) and cannot therefore be blue.
Also Trozzolo’s and Gotthardt’s zwitterions 7, 14, and 15
(
Schemes 6 and 8) are not predicted to have significant
visible absorptions beyond 400 nm and cannot be blue. In
contrast, the endo-lactones endo-17 and endo-20 are pre-
dicted to have long-wavelength maxima (535–650 nm) of a
charge-transfer type in the electronic spectra at the TD-
B3LYP level; these compounds thus emerge as possible
carriers of the photochromism.
Acknowledgements
This work was supported by the Australian Research Council, the Centre for
Computational Molecular Science at The University of Queensland, and the
National Computing Infrastructure facility supported by the Australian
Government (merit allocation scheme, grant g01). The authors thank Dr Lisa
George for some preliminary calculations.

Sydnone Photochemistry
467
References
[12] (a) J. M. Tien, I. M. Hunsberger, Chem. Ind. (London) 1955, 199.
b) J. M. Tien, I. M. Hunsberger, J. Am. Chem. Soc. 1955, 77, 6604.
doi:10.1021/JA01629A052
c) C. V. Greco, B. P. O’Reilly, J. Heterocycl. Chem. 1972, 9, 207.
(
[1] (a) J. C. Earl, A. W. Mackney, J. Chem. Soc. 1935, 899. doi:10.1039/
JR9350000899
(
(
(
b) J. C. Earl, Nature 1946, 158, 910. doi:10.1038/158910A0
c) R. A. Eade, J. C. Earl, J. Chem. Soc. 1946, 591. doi:10.1039/
doi:10.1002/JHET.5570090207
13] M. Sorm, S. Nespurek, Collect. Czech. Chem. Commun. 1975, 40,
[
JR9460000591
3
459. doi:10.1135/CCCC19753459
14] (a) H. S. Gutowsky, R. L. Rutledge, I. M. Hunsberger, J. Chem. Phys.
958, 29, 1183. doi:10.1063/1.1744675
b) T. Mill, A. van Roggen, F. C. Wackly, J. Chem. Phys. 1961, 35,
139. doi:10.1063/1.1701203
[2] (a) W. Baker, W. D. Ollis, Nature 1946, 158, 703. doi:10.1038/
[
158703A0
1
(
b) W. Baker, W. D. Ollis, V. D. Poole, J. Chem. Soc. 1949, 307.
(
1
doi:10.1039/JR9490000307
[3] J. Kenner, K. Mackay, Nature 1946, 158, 909. doi:10.1038/158909B0
[4] A. Sch o¨ nberg, J. Chem. Soc. 1938, 824. doi:10.1039/JR9380000824
[5] K. A. Jensen, A. Friediger, Kgl. Danske Videnskab. Selskab
[
[
[
[
[
15] F. I. Metz, W. C. Servoss, F. E. Welsh, J. Phys. Chem. 1962, 66, 2446.
doi:10.1021/J100818A031
16] S. Nespurek, S. B o¨ hm, J. Kuthan, J. Mol. Struct. THEOCHEM 1986,
1943, 20, 1.
1
36, 261. doi:10.1016/0166-1280(86)80141-5
17] S. Nespurek, J. Obrda, J. Lipinski, J. Photochem. Photobiol. Chem.
993, 75, 49. doi:10.1016/1010-6030(93)80159-7
[
6] (a) W. E. Thiessen, H. Hope, J. Am. Chem. Soc. 1967, 89, 5977.
doi:10.1021/JA00999A059
1
(
b) H. Hope, W. E. Thiessen, Acta Crystallogr. B 1969, 25, 1237.
18] S. Nespurek, J. Lukas, S. B o¨ hm, Z. Bastl, J. Photochem. Photobiol.
Chem. 1994, 84, 257. doi:10.1016/1010-6030(94)03870-8
19] A. M. Trozzolo, T. M. Leslie, A. S. Sarpotdar, R. D. Small,
G. J. Ferraudi, Pure Appl. Chem. 1979, 51, 261. doi:10.1351/
PAC197951020261
doi:10.1107/S0567740869003888
[7] (a) Other X-ray structure determinations of sydnones: J. T. King, P. N.
Preston, J. S. Suffolk, K. Turnbull, J. Chem. Soc., Perkin Trans. 2
1
979, 1751. doi:10.1039/P29790001751
b) S. Nespurek, J. Hasek, M. Sorm, K. Huml, J. Mol. Struct. 1982, 82,
5. doi:10.1016/0022-2860(82)85237-X
c) C. H. Ueng, P. L. Lee, Y. Wang, M. Y. Yeh, Acta Crystallogr. C
985, 41, 1776. doi:10.1107/S0108270185009416
d) J. Hasek, P. T. Beuerskens, J. Obrda, H. Schenk, K. Goubitz,
(
9
[
[
[
20] P. V. Bernhardt, D. Kvaskoff, R. N. Veedu, C. Wentrup, Aust. J. Chem.
2
012, 65, 376. doi:10.1071/CH12041
(
1
21] P. V. Bernhardt, C. Wentrup, Aust. J. Chem. 2012, 65, 371.
doi:10.1071/CH12040
22] (a) A. Fiksdahl, C. Pl u¨ g, C. Wentrup, J. Chem. Soc., Perkin Trans. 2
(
K. Huml, J. D. Schagen, Coll. Czech. Chem. Commun 1985, 50, 1764.
doi:10.1135/CCCC19851764
2
000, 1841. doi:10.1039/B003662P
b) H. G. Andersen, U. Mitschke, C. Wentrup, J. Chem. Soc., Perkin
Trans. 2 2001, 602. doi:10.1039/B007298M
c) H. Sheibani, P. V. Bernhardt, C. Wentrup, J. Org. Chem. 2005, 70,
859. doi:10.1021/JO050428M
d) H. Bornemann, C. Wentrup, J. Org. Chem. 2005, 70, 5862.
(
(
e) C. H. Ueng, Y. Wang, M. Y. Yeh, Acta Crystallogr. C 1987, 43,
122. doi:10.1107/S0108270187092801
f) S. Nespurek, J. Obrda, J. Lipinski, J. Photochem. Photobiol. A 1993,
5, 49. doi:10.1016/1010-6030(93)80159-7
g) A. Hergold-Brundic, A. Nagl, L. Spoljarik-Lukacic, Z. Krist.
New Cryst. Struct. 1997, 212, 407.
h) S. Thamotharan, V. Parthasarathi, R. Sanyal, B. Badami,
1
(
5
(
7
(
(
doi:10.1021/JO050429E
[
23] C. Pl u¨ g, B. Wallfisch, H. G. Andersen, P. V. Bernhardt, L. J. Baker,
G. R. Clark, M. Wong, C. Wentrup, J. Chem. Soc., Perkin Trans. 2
(
A. Linden, Acta Crystallogr. E 2003, 59, o894. doi:10.1107/
S1600536803011309
2
000, 2096. doi:10.1039/B003933K
24] X. Ye, J. Andraos, H. Bibas, M. W. Wong, C. Wentrup, J. Chem. Soc.,
[
[
[
(
i) D. A. Grossie, L. Sun, K. Turnbull, Acta Crystallogr. E 2007, 63,
o2042. doi:10.1107/S160053680701361X
j) D. A. Grossie, K. Turnbull, S. Felix-Balderrama, S. Raghavapuram,
Acta Crystallogr. E 2009, 65, o367. doi:10.1107/S1600536809002347
k) H.-K. Fun, J. H. Goh, B. Nithinchandra, Kalluraya, Acta Crystal
logr. E 2010, 66, o3252. doi:10.1107/S1600536810047422
l) H.-K. Fun, W.-S. Loh, B. Nithinchandra, Kalluraya, Acta Crystal
logr. E 2011, 67, o1489. doi:10.1107/S1600536811018484
m) H.-K. Fun, W.-S. Loh, B. Nithinchandra, Kalluraya, Acta Crys
tallogr. E 2011, 67, o1175. doi:10.1107/S1600536811013900
n) H.-K. Fun, T. S. Chia, B. Nithinchandra, S. Kalluraya, Shetty, Acta
Crystallogr. E 2012, 68, o2103. doi:10.1107/S1600536812026049
o) D. Grossie, L. Harrison, K. Turnbull, Acta Crystallogr. E 2013, 69,
o1196. doi:10.1107/S1600536813017765
p) M. A. Rahiman, G. N. Ravikumar, W.-S. Loh, I. A. Razak, Acta
Perkin Trans. 1 2000, 401. doi:10.1039/A906831G
25] N. Harrit, A. Holm, I. R. Dunkin, M. Poliakoff, J. J. Turner, J. Chem.
Soc., Perkin Trans. 2 1987, 1227. doi:10.1039/P29870001227
26] (a) Evidence for ketenes from m u¨ nchnones: H. O. Bayer, R. Huisgen,
R. Knorr, F. C. Schaefer, Chem. Ber. 1970, 103, 2581. doi:10.1002/
CBER.19701030830
(
(
(
(
b) E. Funke, R. Huisgen, Chem. Ber. 1971, 104, 3222. doi:10.1002/
CBER.19711041027
c) J. Lukac, J. H. Bieri, H. Heimgartnr, Helv. Chim. Acta 1977, 60,
657. doi:10.1002/HLCA.19770600521
d) J. Lukac, H. Heimgartner, Helv. Chim. Acta 1979, 62, 1236.
doi:10.1002/HLCA.19790620435
e) N. H. Toubro, B. Hansen, N. Harrit, A. Holm, K. T. Potts, Tetra
hedron 1979, 35, 229. doi:10.1016/S0040-4020(01)99486-4
f) R. G. Wilde, Tetrahedron Lett. 1988, 29, 2027. doi:10.1016/S0040-
039(00)87825-9
g) Evidence against ketenes from m u¨ nchnones: A. M. Knowles,
(
(
1
(
(
(
(
(
(
4
(
Crystallogr. E 2013, 69, o727. doi:10.1107/S1600536813008805
8] (a) J. H. Bowie, R. A. Eade, J. C. Earl, Aust. J. Chem. 1968, 21, 1665.
doi:10.1071/CH9681665
[
[
A. Lawson, G. V. Boyd, R. A. Newberry, Tetrahedron Lett. 1971, 485.
doi:10.1016/S0040-4039(01)96475-5
(
b) R. S. Goudie, P. N. Preston, M. H. Palmer, Org. Mass Spectrom.
969, 2, 953. doi:10.1002/OMS.1210021004
9] K. Undheim, M. A. F. El-Gendy, T. Hurum, Org. Mass Spectrom.
974, 9, 1242. doi:10.1002/OMS.1210091216
1
(
h) W. Friedrichsen, W.-D. Schr o¨ er, Liebigs Ann. Chem. 1980, 1836.
doi:10.1002/JLAC.198019801121
1
(
(
i) H. J. Gingrich, J. S. Baum, in Oxazoles, Heterocyclic Compounds
Ed. I. J. Turchi) 1986, Vol. 45, pp. 731–961 (Wiley: New York, NY).
[
10] Mass spectra of some other mesoionic compounds have been inter-
preted in terms of ring opening to ketene or isocyanate valence isomers
prior to ionization: K. T. Potts, R. Armbruster, E. Houghton, J. Kane,
Org. Mass Spectrom. 1973, 7, 203. doi:10.1002/OMS.1210070210
11] (a) H. Gotthardt, F. Reiter, A. Gieren, V. Lamm, Tetrahedron Lett.
[
27] (a) For ring-chain isomerization between pyrazolylium oxides and
hydrazonylketenes see: S. Ebner, B. Wallfisch, J. Andraos, I. Aitbaev,
M. Kiselewski, P. V. Bernhardt, G. Kollenz, C. Wentrup, Org. Biomol.
Chem. 2003, 1, 2550. doi:10.1039/B304070D
[
1
978, 19, 2331.
b) H. Gotthardt, F. Reiter, Chem. Ber. 1981, 114, 1737. doi:10.1016/
S0040-4039(01)91528-X
(b) Ring-chain isomerism between 1-isocyanatocarbonylpyrazole
and the pyrazolo[1,2-a][1,2,4]triazolylium olate: A. B o¨ ttcher,
(

4
68
R. N. Veedu, D. Kvaskoff, and C. Wentrup
T. Debaerdemaeker, J. G. Radziszewski, W. Friedrichsen, Chem. Ber.
988, 121, 895. doi:10.1002/CBER.19881210513
c) For ring-chain isomerizations involving isothiocyanates, see:
[43] F. Kurzer, K. Douraghi-Zadeh, Chem. Rev. 1967, 67, 107. doi:10.1021/
CR60246A001
[44] M. Pagacz-Kostrzewa, J. Krupa, A. Olbert-Majkut, M. Podruczna,
R. Bronisz, M. Wierzejewska, Tetrahedron 2011, 67, 8572.
doi:10.1016/J.TET.2011.08.055
1
(
R. M. Moriarty, R. Mukherjee, O. L. Chapman, D. R. Eckroth, Tetra
hedron Lett. 1971, 397. doi:10.1016/S0040-4039(01)96450-0
(
d) S. Chaloupka, H. Heimgartner, H. Schmid, H. Link,
[45] O. L. Chapman, J.-P. LeRoux, J. Am. Chem. Soc. 1978, 100, 282.
doi:10.1021/JA00469A049
P. Sch o¨ nholzer, K. Bernauer, Helv. Chim. Acta 1976, 59, 2566.
doi:10.1002/HLCA.19760590733
28] K. Bhattacharyya, D. Ramaiah, P. K. Das, M. V. George, J. Photo-
chem. 1987, 36, 63. doi:10.1016/0047-2670(87)87063-6
29] (a) R. Huisgen, H. Gotthardt, Chem. Ber. 1968, 101, 1059.
doi:10.1002/CBER.19681010343 and references therein.
[46] R. Harder, Etude de la formation et des r e´ arrangements des carb e` nes
et nitr e` nes, Th e` se de Doctorat 1978, Universit e´ de Lausanne,
Switzerland.
[
[
[47] W. Sieber, P. Gilgen, S. Chalupka, H.-J. Hansen, H. Schmid, Helv.
Chim. Acta 1973, 56, 1679. doi:10.1002/HLCA.19730560525
[48] (a) L. M. T. Frija, I. D. Reva, A. Gomez-Zavaglia, M. L. S. Cristiano,
R. Fausto, J. Phys. Chem. A 2007, 111, 2879. doi:10.1021/JP068556H
(b) A. Gomez-Zavaglia, I. D. Reva, L. Frija, M. L. S. Cristiano,
R. Fausto, J. Photochem. Photobiol. Chem. 2006, 180, 175.
doi:10.1016/J.JPHOTOCHEM.2005.10.012
(c) A. Gomez-Zavaglia, I. D. Reva, L. Frija, M. L. Cristiano, R. Fausto,
J. Phys. Chem. A 2005, 109, 7967. doi:10.1021/JP0517706
[49] M. Pagacz-Kostrzewa, I. D. Reva, R. Bronisz, B. M. Giuliano,
R. Fausto, M. Wierzejewska, J. Phys. Chem. A 2011, 115, 5693.
doi:10.1021/JP202925R
(
b) D. L. Browne, J. P. A. Harrity, Tetrahedron 2010, 66, 553.
doi:10.1016/J.TET.2009.10.085
[
[
30] (a) H. Gotthardt, F. Reiter, Tetrahedron Lett. 1971, 12, 2749.
doi:10.1016/S0040-4039(01)96970-9
(
(
b) H. Gotthardt, F. Reiter, Chem. Ber. 1979, 112, 1206.
c) H. Gotthardt, F. Reiter, Chem. Ber. 1979, 112, 1635. .
31] (a) C. S. Angadiyavar, M. V. George, J. Org. Chem. 1971, 36, 1589.
doi:10.1021/JO00811A004
(
b) K. Butkovi c´ , N. Basari c´ , K. Lovrekovi c´ , Z. Marini c´ , A. Vi sˇ njevac,
B. Koji c´ -Prodi c´ , M. Sindler-Kulyk, Tetrahedron Lett. 2004, 45, 9057.
doi:10.1016/J.TETLET.2004.10.036
[50] (a) S. Fischer, C. Wentrup, Chem. Commun. 1980, 502.
(b) M. Granier, A. Baceiredo, H. Grutzmacher, H. Pritzkow,
G. Bertrand, Angew. Chem. Int. Ed. 1990, 29, 659. doi:10.1002/
ANIE.199006591
[
[
[
[
[
[
[
32] K.-H. Pfoertner, J. Foricher, Helv. Chim. Acta 1980, 63, 653.
doi:10.1002/HLCA.19800630312
33] C. H. Krauch, J. Kuhls, H.-J. Piek, Tetrahedron Lett. 1966, 7, 4043.
doi:10.1016/S0040-4039(00)90284-3
[51] W. T. G. Johnson, M. B. Sullivan, C. J. Cramer, Int. J. Quantum Chem.
ˇ
ˇ
34] K. Butkovi c´ , Z. Marini c´ , K. Mol cˇ anov, B. Koji c´ -Prodi c´ , M. Sindler-
Kulyk, Tetrahedron 2010, 66, 9356. doi:10.1016/J.TET.2010.10.013
35] P. Leihkauf, V. Lohse, C. Csongar, G. Tomascewski, J. Wilda,
Z. Chem. 1985, 25, 266. doi:10.1002/ZFCH.19850250722
2001, 85, 492. doi:10.1002/QUA.1518
[52] (a) 3-Phenylsydnone: C. J. Thoman, D. J. Voaden, Synthesis 1965, 96.
(b) 3-(3-Pyridyl)sydnone: J. M. Tien, I. M. Hunsberger, J. Am. Chem.
Soc. 1955, 77, 6604. doi:10.1021/JA01629A052
36] K. Bhattacharyya, M. Muneer, R. K. Tikare, P. K. Das, M. V. George,
(c) J. M. Tien, I. M. Hunsberger, J. Am. Chem. Soc. 1961, 83, 178.
doi:10.1021/JA01462A035
(d) 3,4-Diphenylsydnone: Ref. [2b].
J. Indian Chem. Soc. 1986, 63, 43.
37] G. Eber, S. Schneider, F. D o¨ rr, Ber. Bunsenges. Phys. Chem 1980, 84,
281. doi:10.1002/BBPC.19800840318
[53] (a) C. Wentrup, D. Kvaskoff, Aust. J. Chem. 2013, 66, 286.
doi:10.1071/CH12502
38] (a) UV spectra of nitrilimines: H. Meier, W. Heinzelmann,
H. Heimgartner, Chimia 1980, 34, 504.
(b) D. Kvaskoff, P. Bednarek, L. George, S. Pankajakshan,
C. Wentrup, J. Org. Chem. 2005, 70, 7947. doi:10.1021/JO050898G
[54] M. J. Frisch, et al., Gaussian 03, Revision B.05 2003 (Gaussian, Inc.:
Pittsburgh, PA) (full citation in the Supplementary Material).
[55] M. J. Frisch et al., Gaussian 09, Revision A.2 2009 (Gaussian, Inc.:
Wallingford, CT) (full citation in the Supplementary Material).
[56] S. I. Gorelsky, A. B. P. Lever, J. Organomet. Chem. 2001, 635, 187.
doi:10.1016/S0022-328X(01)01079-8
(
b) N. H. Toubro, A. Holm, J. Am. Chem. Soc. 1980, 102, 2093.
doi:10.1021/JA00526A058
39] (a) Review on nitrile imines: G. Bertrand, C. Wentrup, Angew. Chem.
[
1994, 106, 549. doi:10.1002/ANGE.19941060505
(
b) G. Bertrand, C. Wentrup, Angew. Chem. Int. Ed. 1994, 33, 527.
doi:10.1002/ANIE.199405271
[
[
[
40] M. M a¨ rky, H. Meier, H. Heimgartner, H. Schmid, H.-J. Hansen, Helv.
Chim. Acta 1978, 61, 1477. doi:10.1002/HLCA.19780610432
41] D. B e´ gu e´ , G. Qiao, C. Wentrup, J. Am. Chem. Soc. 2012, 134, 5339.
doi:10.1021/JA2118442
[57] M. W. Wong, Chem. Phys. Lett. 1996, 256, 391. doi:10.1016/0009-
2614(96)00483-6
42] O. L. Chapman, C. L. McIntosh, J. Pacansky, G. V. Calder, G. Orr,
J. Am. Chem. Soc. 1973, 95, 4061. doi:10.1021/JA00793A046