Mechanistic studies of direct and sensitized photolysis of methyl (4-nitrophenyl)diazoacetate in the presence of an electron-donating amine: Photochemical generation of the diazoalkane radical anion

Article abstract of DOI:10.1039/b202800j

The direct and perylene-sensitized photolysis of methyl (4-nitrophenyl)diazoacetate (1) in the presence of an electron-donating amine was investigated by the characterization of reaction products and by the direct observation of reactive species using a laser flash photolysis technique. In the direct photolysis of 1 in the presence of N,N,N′,N′-tetramethyl-p-phenylenediamine (TMPD), electron transfer from TMPD occurs and the photoproducts are obtained through an intermediate other than the carbene 3, which we suppose to be the carbene radical anion 3^-·. In the perylene-sensitized photolysis of 1 in the presence of an amine having a large electron-donating ability such as N,N-dimethylaniline and TMPD, the radical anion of 1 can be generated by perylene-mediated electron transfer from the amine to 1. The resulting 1^-· appears to undergo extrusion of dinitrogen to give 3^-· smoothly, the behavior of which is dependent on the amine employed as an additive. The product studies revealed that 3^-· reacts readily with O₂ to give the ketoester 5 or abstracts hydrogen atoms to give the ester 7, which is not inconsistent with the theoretical prediction that 3^-· has a σ radical character in its electronic structure.

Full text of DOI:10.1039/b202800j

View Article Online / Journal Homepage / Table of Contents for this issue
Mechanistic studies of direct and sensitized photolysis of methyl
(4-nitrophenyl)diazoacetate in the presence of an electron-donating
amine: photochemical generation of the diazoalkane radical anion
2
Tadashi Mizushima,a Go Monden,a Shigeru Murata,*a Kunihiko Ishiib and Hiro-o Hamaguchib
a Department of Basic Science, Graduate School of Arts and Sciences, The University of Tokyo,
Meguro-ku, Tokyo, 153-8902, Japan
b Department of Chemistry, School of Science, The University of Tokyo, Bunkyo-ku, Tokyo,
113-0033, Japan
Received (in Cambridge, UK) 19th March 2002, Accepted 26th April 2002
First published as an Advance Article on the web 30th May 2002
The direct and perylene-sensitized photolysis of methyl (4-nitrophenyl)diazoacetate (1) in the presence of an
electron-donating amine was investigated by the characterization of reaction products and by the direct observation
of reactive species using a laser flash photolysis technique. In the direct photolysis of 1 in the presence of N,N,NЈ,NЈ-
tetramethyl-p-phenylenediamine (TMPD), electron transfer from TMPD occurs and the photoproducts are obtained
through an intermediate other than the carbene 3, which we suppose to be the carbene radical anion 3Ϫ . In the
ؒ
perylene-sensitized photolysis of 1 in the presence of an amine having a large electron-donating ability such as
N,N-dimethylaniline and TMPD, the radical anion of 1 can be generated by perylene-mediated electron transfer from
Ϫ
the amine to 1. The resulting 1Ϫ appears to undergo extrusion of dinitrogen to give 3 smoothly, the behavior of
ؒ
ؒ
which is dependent on the amine employed as an additive. The product studies revealed that 3Ϫ reacts readily with O2
ؒ
to give the ketoester 5 or abstracts hydrogen atoms to give the ester 7, which is not inconsistent with the theoretical
prediction that 3Ϫ has a σ radical character in its electronic structure.
ؒ
reported so far to generate diazoalkane radical anions photo-
chemically. Tomioka and his co-workers found that the irradi-
Introduction
One-electron reduction is an attractive method for the decom-
ation of (4-nitrophenyl)diazomethane in diethylamine afforded
position of diazoalkanes, in which their radical anions are gen-
4-nitrotoluene predominantly, and they explained its formation
erated as the first-formed reactive intermediate. The behavior of
by a mechanism involving the radical anion of the diazo-
diazoalkane radical anions is the subject of great interest, par-
methane produced from an electron transfer from diethyl-
ticularly in that a loss of dinitrogen gives rise to the correspond-
amine to the diazomethane.11 To our knowledge, this is the sole
ing carbene radical anions, which are expected to have unique
example of the photoreaction in which the intervention of the
structural properties and chemical reactivities. The research on
one-electron reduction of diazoalkanes was initiated in 1963 by
Kauffmann and Hage, who reported that the reduction of
diazoalkane radical anion was proposed. It seems possible that
the photolysis of a diazoalkane having an electron-deficient
group in the presence of an electron-donating substance results
in an electron transfer from the substance to the diazoalkane,
diphenyldiazomethane (Ph2CN2) with sodium gave its radical
anion.1 Moreover, Webster proposed the formation of carbene
which gives rise to the diazoalkane radical anion. In this paper,
radical anion in the polarographic reduction of tetra-
we report our mechanistic studies of direct and sensitized
cyanodiazocyclopentadiene.2 The intervention of the parent
methylene radical anion, as well as cyclopentadienylidene
photolysis of methyl (4-nitrophenyl)diazoacetate (1) in the
presence of an electron-donating amine, in which we wish to
radical anion, was suggested in one-electron reduction of the
propose a reaction scheme involving the radical anion of 1
corresponding diazoalkane in the gas phase.3,4
formed from an electron transfer from an amine to 1.12
In the 1980s, the electrochemical reduction of diazoalkanes
was extensively investigated. McDonald and Hawley reported
that the electrochemical reduction of Ph2CN2 in DMF afforded
Results
benzophenone azine through the carbene radical anion,
Direct photolysis of the diazoacetate 1 in the presence of an
amine
Ph2CϪ , produced from electrogenerated Ph2CN2 by rapid
Ϫ
ؒ
ؒ
loss of dinitrogen.5 However, Bethell and his co-workers pro-
Ϫ
ؒ
posed that the azine was formed by the reaction of Ph2CN2
with Ph2CN2 or Ph2CN2 on the basis of the data obtained by
(1) Dependence of photoproducts on amines. The photo-
chemistry of the diazoacetate 1 was reported by Tomioka
and his co-workers, in which the irradiation of 1 in methanol
afforded methyl α-methoxy-(4-nitrophenyl)acetate (2).13 This
product was formed by the insertion of the photolytically
generated carbene 3 into the OH bond of methanol, which is
known as an excellent trapping agent for carbene.14 A mechan-
ism for the formation of 2 involving the Wolff rearrangement
of 3 into the ketene, followed by reaction with methanol, was
excluded, because the irradiation of 1 in methanol-d4 gave 2
deuterated not at the methyl group of the methoxycarbonyl
moiety (δ 3.75) but at that of the α-methoxy moiety (δ 3.47).
Ϫ
ؒ
using their skilful cyclic voltammetric experiments, which ruled
out the intervention of Ph2CϪ . Based on the results of electro-
6
ؒ
chemical reduction of various diazoalkanes,7–9 they established
that a unimolecular loss of dinitrogen from the diazoalkane
radical anion giving the corresponding carbene radical anion
occurred in the diazoalkanes having a carbonyl group at the
position adjacent to the diazo group, such as azibenzil and
diethyl diazomalonate.7,10
In contrast to this research on chemical and electrochemical
one-electron reduction of diazoalkanes, no attempt has been
1274
J. Chem. Soc., Perkin Trans. 2, 2002, 1274–1282
This journal is © The Royal Society of Chemistry 2002
DOI: 10.1039/b202800j

View Article Online
Table 1 Dependence of the photoproducts on additives in the photolysis of 1 in acetonitrile–methanol (9 : 1)a
Yield (%)b
Additives
2
4
5
7
Eox c/V vs. SCE
∆Gd/kcal molϪ1
None
TEA
ClDMA
DMA
MeODMA
TMPD
TMPDe
34
38
29
26
23
< 1
< 1
12
14
16
15
17
10
11
10
< 1
11
12
17
19
< 1
< 1
11
2
—
—
1.07
0.89
0.79
0.55
0.12
0.12
Ϫ4.3
Ϫ8.5
Ϫ11
Ϫ16
Ϫ26
Ϫ26
5
13
13
29
a [1] = 5 mM, [Additive] = 50 mM. b Yield based on the reacted material (30–40%). c Oxidation potential of the additive. d Free-energy change for
the electron transfer from the additive to 1*T, the excitation of which is assumed to be 58 kcal molϪ1 e Degassed by three freeze–thaw cycles.
.
TEA = triethylamine, ClDMA = 4-chloro-N,N-dimethylaniline, DMA = N,N-dimethylaniline, MeODMA = 4-methoxy-N,N-dimethylaniline,
TMPD = N,N,NЈ,NЈ-tetramethyl-p-phenylenediamine.
Since the reaction of carbene is very complicated in the absence
of a trapping agent, and it is reasonable to think that an elec-
tron transfer occurs favorably in a polar solvent, we selected
acetonitrile containing 10% (v/v) methanol as a solvent to
employ in the study of the photolysis of 1.
carbene 3 in its triplet state from TEA to give 7 in preference to
the reaction with O2 to give 5. However, we found that an
increase in the electron-donating ability of the amine, which is
evaluated by a lowering of its oxidation potential (Table 1),
caused an increase in the yields of 5 and 7 at the expense of 2.
Especially, it should be noted that the formation of 2 was
quenched completely in the photolysis of 1 in the presence
of N,N,NЈ,NЈ-tetramethyl-p-phenylenediamine (TMPD). This
result implies that a reactive intermediate other than the
carbene 3 would intervene in the photolysis of 1 in the presence
of TMPD. As shown in the table, when the photolysis was
carried out after the strict removal of O2, 5 was not detected but
the yield of 7 significantly increased instead. This result indi-
cates that the reactive intermediate generated in the presence of
TMPD could afford not 2 but both 5 and 7. The oxazole 4
could be produced when the formation of 2 was quenched,
suggesting that 4 is at least partly formed through a pathway in
which the carbene 3 does not participate.
When 1,4-diazabicyclo[2.2.2]octane (DABCO) was employed
as an additive, the azine 8 was obtained predominantly. It is
known that azines are produced by electrochemical reduc-
tion of diazoalkanes through diazoalkane radical anion as a
reactive intermediate.5–8 However, it is found that the azine 8
was formed when the photoreaction mixture obtained by the
irradiation of 1 in the presence of DABCO was concentrated
by evaporation of the solvent. Moreover, the photolysis in
an NMR sample tube confirmed that the irradiation of 1 in the
presence of DABCO in methanol-d4 gave not the azine 8
but the ether 2. Probably, the azine 8 was produced through
the deprotonation of 2 by DABCO, followed by the attack of
the resulting carbanion on the terminal nitrogen atom of the
starting material 1 remaining in the reaction mixture and
the elimination of methoxide ion (Scheme 1). The similar
conversion of diazo compounds into the corresponding
azines induced by carbanions was reported by Bethell and
McDowall.18
A solution of 1 in acetonitrile containing methanol was
purged with argon, and irradiated with a xenon arc lamp
(> 390 nm). Chromatographic separation of the reaction mix-
ture gave three products, the α-methoxyacetate 2,13 the oxazole
4,15 and the ketoester 5,16 which were identified by comparison
of their 1H NMR spectra with those of authentic samples. The
formation of the oxazole 4 is reasonably interpreted in terms of
the intramolecular cyclization of the nitrile ylide 6. Nitrile
ylides are accepted as reactive intermediates in the photolysis of
diazo compounds in acetonitrile, direct spectroscopic observ-
ation of which has been recently achieved by using a laser flash
photolysis technique.17 It seems that the ketoester 5 is produced
by the reaction of the carbene 3 in its triplet state with O2
remaining in the solution, because the formation of 5 was
suppressed in the photolysis of the solution degassed strictly by
freeze–thaw cycles.
When irradiation of 1 was carried out in the presence of an
amine (50 mM), a mixture composed of 2, 4, and 5 was
obtained, but the product distribution varied with the amine
employed as an additive. The results are shown in Table 1. In
the presence of triethylamine (TEA), methyl (4-nitrophenyl)-
acetate (7) was obtained instead of 5. This observation seems to
be explained in terms of the hydrogen abstraction of the
Scheme 1
J. Chem. Soc., Perkin Trans. 2, 2002, 1274–1282
1275

View Article Online
(2) Dependence of the product distribution on TMPD con-
centration. Table 1 shows that the formation of the ether 2 is
retarded completely in the photolysis of 1 in the presence of
50 mM of TMPD. We examined the dependence of the product
distribution on the TMPD concentration, [TMPD], which
would provide information about the reactive species quenched
by TMPD. As [TMPD] was increased from 0 to 5 mM, the yield
of 2 decreased simply from 34% to 1%, while the yield of 5
increased simply from 4% to 24%. The yield of the oxazole 4
depended in a complicated manner on [TMPD]; the yield of 4
increased from 12% to 19% upon addition of TMPD (0.5 mM),
and decreased gradually to 15% with increasing [TMPD] to
5 mM. Fig. 1 illustrates a plot of Φ(2)0/Φ(2) against [TMPD], in
Fig. 2 Transient absorption spectra of a solution of 1 (1.0 mM) in
acetonitrile recorded in the presence (solid line) and absence (dotted
line) of TMPD (10 mM) at the delay time of 300 ns after a laser pulse
(390 nm, 20 ns).
On the other hand, the transient absorption spectrum
recorded in the photolysis of 1 in the presence of TMPD
(10 mM) exhibited additional bands with maxima of 560 and
610 nm (a solid line in Fig. 2). The intensity of these bands
increased until 300 ns after the laser pulse, and decreased slowly
with a lifetime of > 5 µs. These bands were observed even in an
O2-saturated acetonitrile solution. We identified the transient
species having these bands as TMPD radical cation, TMPDϩ
,
ؒ
by comparison with an authentic spectrum reported in the
literature.19 Since the diazoacetate 1 is excited exclusively under
the conditions of laser flash photolysis, the direct observation
of TMPDϩ provides unambiguous evidence for the involve-
ؒ
Fig. 1 Dependence of the yield of 2 on the TMPD concentration in
the photolysis of 1 in acetonitrile–methanol (9 : 1). Solid and dotted
lines represent the curves fitted to the experimental data by using the
non-linear least-squares analysis according to eqn. (1) and eqn. (2),
respectively.
ment of a single electron transfer process in the photolysis of 1
in the presence of TMPD.
Sensitized photolysis of the diazoacetate 1 in the presence of an
amine
which Φ(2)0 and Φ(2) are the yields of 2 in the absence and
the presence of TMPD, respectively. As shown in the figure, the
plot gives a curve deviating upward from a straight line pre-
dicted by a Stern–Volmer type relationship. This observation
implies that a reaction scheme in which a single reactive species
leading to the formation of 2 is quenched by TMPD cannot be
applicable to this photoreaction. The analysis of these data is
shown in a subsequent section.
Even if an electron transfer occurs from the amine employed as
an additive to the photochemically excited diazoacetate 1, the
efficiency of the generation of 1Ϫ would be low because the
ؒ
extrusion of dinitrogen giving the carbene 3 competes with the
electron transfer process. However, when an electron is trans-
ferred from a sensitizer radical anion, which is generated by the
electron transfer from an amine to the excited sensitizer,
to 1, this problem could be overcome. Moreover, the charge
recombination between the resulting amine radical cation and
(3) Laser flash photolysis studies. In order to have inform-
ation about the mechanism of the photochemistry of 1 in the
presence of TMPD, we tried a direct observation of transient
intermediates using a nanosecond laser flash photolysis tech-
nique. Irradiation of 1 in acetonitrile (1.0 mM) degassed by
bubbling of N2 with the laser pulse (390 nm, 20 ns) gave an
intense absorption band with a maximum at 445 nm and a weak
broad absorption around 580 nm (a dotted line in Fig. 2). The
former absorption, the lifetime of which was estimated to be
> 5 µs, is assigned to the nitrile ylide 6 for the following three
reasons. First, there have been many precedents that the laser
flash photolysis of diazoalkanes in acetonitrile gives a transient
absorption around 400 nm having a long lifetime of the order
of µs, which is thought to be due to the absorption of the corre-
sponding nitrile ylide.17 Second, the absorption was observed
even in an O2-saturated acetonitrile solution and its lifetime was
unchanged, compared with that recorded in the degassed solu-
tion, excluding the possibility of species in the triplet state.
Finally, the absorption was not observed in a benzene solution.
The weak broad absorption around 580 nm, which had a
relatively short lifetime of less than 1 µs, was not observed in an
O2-saturated solution. We assigned this absorption to the T–T
absorption of the oxazole 4, which was confirmed by the laser
flash photolysis of an authentic sample of 4 in acetonitrile.
1Ϫ , which reduces the practical efficiency of the electron trans-
ؒ
fer, could be suppressed in the sensitized photolysis, compared
with the direct irradiation, because these radical ions are gener-
ated separately from each other. Thus, the electron transfer
from the photochemically generated sensitizer radical anion to
1 is thought to be a more reliable approach to the generation of
1Ϫ , compared with the electron transfer from an amine to the
ؒ
excited state of 1.
We selected perylene as a sensitizer, because perylene has
an absorption with a large absorptivity in the range of 400 to
450 nm, in which the diazoacetate 1 has a very weak absorption.
Moreover, perylene has a relatively large reduction potential of
Ϫ1.67 V (vs. SCE),20 which is favorable for an electron transfer
from an electron-donating amine. Unfortunately, because of
low solubility of perylene in acetonitrile–methanol, the selec-
tive excitation of perylene cannot be achieved. However, under
our irradiation conditions ([perylene] = 0.3 mM, [1] = 5 mM), it
is estimated that 83% of the incident light (at 440 nm) is
absorbed by the sensitizer.
(1) Dependence of photoproducts on amines. Irradiation
(> 440 nm) of a solution of 1 (5 mM) in acetonitrile containing
10% methanol in the presence of perylene (0.3 mM) afforded a
1276
J. Chem. Soc., Perkin Trans. 2, 2002, 1274–1282

View Article Online
Table 2 Dependence of the photoproducts on additives in the perylene-sensitized photolysis of 1 in acetonitrile–methanol (9 : 1)a
Yield (%)b
Additives
2
4
5
7
kq c/1010
M
Ϫ1 sϪ1
∆Gd/kcal molϪ1
None
TEA
ClDMA
DMA
MeODMA
TMPD
43
46
39
34
13
17e
5
19
18
20
9
11
4
9
5
27
37
< 1
< 1
< 1
2
< 1
< 1
—
—
0.23
1.2
1.3
1.6
2.0
Ϫ3.3
Ϫ7.5
Ϫ9.8
Ϫ15
Ϫ25
< 1
a [1] = 5 mM, [perylene] = 0.3 mM, [Additive] = 50 mM. b Yield based on the reacted material (30–40%). c Rate constant for quenching of perylene
fluorescence by the additive. d Free-energy change for the electron transfer from the additive to singlet excited perylene. e The isomer of 4 was
obtained, which was tentatively identified as 9. TEA = triethylamine, ClDMA = 4-chloro-N,N-dimethylaniline, DMA = N,N-dimethylaniline,
MeODMA = 4-methoxy-N,N-dimethylaniline, TMPD = N,N,NЈ,NЈ-tetramethyl-p-phenylenediamine.
mixture of the products, which were practically identical to
those obtained in the direct photolysis of 1. It was note-
worthy that the isomerization of the oxazole 4 occurred
under these irradiation conditions. The product was tentatively
identified as the isoxazole 9 by considering the ring current
effect of the 4-nitrophenyl moiety. The signal assigned to the
methyl group on the five-membered ring of 9 appeared in
the more shielding area (δ 2.17), compared with the corre-
sponding methyl signal of 4 (δ 2.44). A similar photochemical
isomerization of the related oxazoles was reported by Buu and
Edward.21
demonstrated that the rate constant for the quenching of peryl-
ene fluorescence by the amine increases with a decrease in
its oxidation potential. This result strongly suggests that the
quenching occurs by a mechanism involving electron transfer
from the amine to singlet excited perylene. The negative free-
energy changes for this electron transfer process calculated by
the Rehm–Weller equation,22 which are also collected in
Table 2, support this assumption.
(3) Laser flash photolysis studies. The transient absorption
spectrum was recorded by the laser excitation (434 nm, 40 ns)
of a deoxygenated solution of perylene (0.2 mM) in the
presence of DMA (50 mM), which is shown with a solid line in
Fig. 3. A transient species having an intense absorption band
Table 2 shows the dependence of the photoproduct distribu-
tion on the amine employed as an additive (50 mM). In contrast
to the irradiation in the absence of additives, the oxazole 4 was
obtained instead of 9, suggesting that an energy transfer from
excited perylene to 4 which causes its isomerization is sup-
pressed by quenching of excited perylene by additives. More-
over, in contrast to the direct irradiation, methyl (4-nitro-
phenyl)acetate (7) was hardly obtained. Similarly to the results
obtained in the direct irradiation, the product distributions
were very dependent on the additive. It was found that the
photoproducts originating from the carbene 3 were obtained
even in perylene-sensitized photolysis of 1 in the presence of an
amine having a relatively high oxidation potential, such as TEA
and N,N-dimethylaniline (DMA). However, it should be
emphasized again that as the electron-donating ability of the
additive was increased, the yield of the ether 2 decreased and
the ketoester 5 was predominantly obtained instead. Thus, the
addition of an amine having a large electron-donating ability,
such as TMPD, resulted in the complete suppression of the
formation of the ether 2, which is a unique product derived
from the carbene 3, indicating the intervention of a reactive
intermediate other than 3.
Fig.
3
Transient absorption spectra of a solution of perylene
(0.20 mM) in acetonitrile containing DMA (50 mM) recorded in the
absence (solid line) and presence (dotted line) of 1 (2.0 mM) at the delay
time of 400 ns after a laser pulse (434 nm, 40 ns).
with a maximum at 575 nm was detected, which can be charac-
terized as perylene radical anion, PeϪ , by comparison with
ؒ
its authentic spectrum.23 This observation clearly shows that
an electron transfer occurs from DMA to excited perylene,
supporting the assumption made on the basis of perylene
fluorescence quenching data. A weak broad absorption around
480 nm could be assigned to the overlap of absorptions of
triplet excited perylene23 and DMAϩ
.
24
ؒ
(2) Quenching of perylene fluorescence by amines. In order
to have a clue to disclosing the mechanism of the perylene-
sensitized photodecomposition of 1 in the presence of amines,
quenching of the perylene fluorescence by the amines employed
as additives was examined next. We obtained Stern–Volmer
constants, KSV = kqτ0, for the quenching of perylene fluores-
cence by the amines. The rate constants for the quenching of
singlet excited perylene by the amines, kq, which are evaluated
by employing the reported lifetime of perylene fluorescence in
acetonitrile (τ0 = 6 ns),20 are collected in Table 2. It is clearly
On the other hand, in the transient absorption spectrum
recorded by the laser excitation of the solution identical to that
described above, except for the presence of 1 (2 mM), no transi-
ent absorption band assigned to PeϪ could be detected, which
ؒ
is demonstrated with a dotted line in Fig. 3. Taking into
account the rate constants for the quenching of the perylene
fluorescence by DMA and 1 (vide infra) and their concen-
trations, it is reasonable to think that singlet excited perylene
formed by the laser excitation is predominantly quenched
by DMA. Therefore, this result strongly suggests that PeϪ
ؒ
J. Chem. Soc., Perkin Trans. 2, 2002, 1274–1282
1277

View Article Online
produced by the electron transfer from DMA to singlet excited
perylene is immediately quenched by 1.
probably due to the delocalization of two π electrons on the
cyclopentadiene π system to achieve the 6π aromaticity.
In order to gain information about the structure and reactiv-
ity of the diazoalkane radical anion 1Ϫ and the carbene radical
ؒ
Cyclic voltammetric measurement of the diazoacetate 1
anion 3Ϫ , we carried out DFT calculations on these reactive
ؒ
The cyclic voltammogram of 1 recorded in acetonitrile at a scan
species. The geometry was optimized by using the UB3LYP/
6-31G(d) method. Figs. 5 and 6 depict the optimized geometries
rate of 0.2 V sϪ1 is exhibited in Fig. 4. The electrochemical
Fig. 5 Structure of 1Ϫ calculated at the UB3LYP/6-31G(d) level of
ؒ
theory. Charge and spin densities (in parentheses) on the selected atoms
are given in the structure drawing.
Fig. 4 Cyclic voltammogram for the reduction of 1 (4.4 mM) recorded
in acetonitrile containing tetrabutylammonium perchlorate (0.1 M) at a
scan rate of 0.2 V sϪ1
.
reduction of 1 is irreversible, showing the reduction peak at
Ep,c = Ϫ1.26 V (vs. Ag/Agϩ). No corresponding peak for the
reoxidation was detected even at a scan rate of 5 V sϪ1. As
shown in Fig. 4, on the reoxidation sweep, two peaks were
observed at Ep,a = Ϫ0.43 and 0.01 V. Although the cyclic
voltammetric measurement was carried out on the products
obtained by the photolysis of 1 in acetonitrile, that is, the
oxazole 4, the ketoester 5, and the ester 7, their redox peaks
were not in agreement with the anodic peaks observed in the
measurement of 1. However, the cyclic voltammetric measure-
ment of the carbanion 10, which could be generated by the
addition of tetramethylammonium hydroxide in a solution of 7,
afforded an oxidation peak at 0.01 V. This observation implies
that one of the products formed in the electrochemical reduc-
tion of 1 is the carbanion 10. The formation of the correspond-
ing carbanion was reported in the electrochemical reduction
of various diazoalkanes, such as diphenyldiazomethane6 and
diethyl diazomalonate.7
Fig. 6 Structure of 3Ϫ calculated at the UB3LYP/6-31G(d) level of
ؒ
theory. Charge and spin densities (in parentheses) on the selected atoms
are given in the structure drawing.
Ϫ
of 1Ϫ and 3 , respectively, together with the charge and spin
ؒ
ؒ
densities on the selected atoms. It is noteworthy that the diazo
moiety of 1Ϫ is significantly bent with a C–N–N angle of
ؒ
132.7Њ, and that the bond angle of the divalent carbon atom of
3Ϫ is 135.7Њ which is nearly identical to that of the carbene in
ؒ
its triplet state 3T (134.9Њ) calculated by the use of the same
method.
Ϫ
It is found that the negative charge of both 1Ϫ and 3 exists
mainly on the nitro group and the methoxycarbonyl moiety,
implying that the negative charge is delocalized on the π conju-
gated system of the molecule. On the other hand, the spin of
ؒ
ؒ
1Ϫ and 3Ϫ is almost perfectly localized on the dinitrogen
moiety and on the divalent carbon, respectively. Especially, it
should be pointed out that the molecular orbital occupied by
ؒ
ؒ
Structure and reactivity of the diazoacetate radical anion 1؊ and
ؒ
the carbene radical anion 3؊ based on the theoretical studies
ؒ
ؒ
The electronic structure of carbene radical anions has been
of interest theoretically in recent years. Davidson and Hudak
carried out molecular orbital calculations using the UHF/STO-
an unpaired electron of 3Ϫ , which is localized on the divalent
carbon, has a σ character, indicating that the electronic struc-
ture of 3Ϫ is essentially described as σ1π2. Thus, 3Ϫ can be
predicted to be the species represented as the conjugate base of
1-(4-nitrophenyl)-2-methoxy-2-hydroxyvinyl radical 11. On the
ؒ
ؒ
Ϫ
25
ؒ
3G method on the parent methylene radical anion, CH2
.
Ϫ
ؒ
They showed that CH2 was similar in geometrical structure
to singlet methylene, and had the σ2π1 electronic structure.
Recently, Seburg and his co-workers drew the same conclusion
from density functional theory (DFT) calculations using the
UB3LYP/6-31ϩG(d) method on phenylcarbene radical anion,
in which the σ2π1 electronic structure was more stable than the
basis of its σ radical character, it is expected that 3Ϫ can react
ؒ
readily with O2 and abstract hydrogen atoms.
Furthermore, the enthalpy change for the extrusion of
Ϫ
dinitrogen from 1Ϫ to yield 3 , ∆∆Hf, was estimated from their
electronic energies calculated by using the UB3LYP/6-31ϩG(d)
method. On the basis of our calculations, ∆∆Hf for the trans-
ؒ
ؒ
σ1π2 structure by 11–12 kcal molϪ1 26
On the other hand, the
.
formation of 1Ϫ into 3Ϫ and dinitrogen is evaluated to be
ؒ
ؒ
semi-empirical calculation on cyclopentadienylidene radical
anion done by McDonald and his co-workers predicted that
this radical anion had the σ1π2 electronic structure in its ground
state.4 The stabilization of the σ1π2 electronic structure is
ϩ2.7 kcal molϪ1. Thus, the endothermicity for the extrusion of
dinitrogen of 1Ϫ is found to be considerably smaller, compared
ؒ
with that of the neutral diazoalkane 1 to yield the carbene 3T
1278
J. Chem. Soc., Perkin Trans. 2, 2002, 1274–1282

View Article Online
the corresponding ketone or abstracts hydrogen atoms to give
the methylene compound.14 Therefore, it is reasonable to think
that the ether 2 and the ketoester 5 obtained by the irradiation
of 1 in the absence of TMPD originate from 3S and 3T, respec-
tively. The formation of the nitrile ylide is reported in the
photolysis of diazoalkanes in acetonitrile,17 so that it is likely
that the oxazole 4 is also derived from 3. However, we propose
that a portion of 4 produced in the photolysis of 1 arises
directly from the electronically excited state of 1 on the basis of
the following two reasons. First, 4 was obtained even in the
photolysis of 1 in the presence of TMPD, in which the form-
ation of 2 was completely quenched. Secondly, the T–T absorp-
tion of 4 was observed immediately after the irradiation of 1
with the laser pulse.
and dinitrogen, which was calculated to be ϩ20.4 kcal molϪ1
by the use of the same method. On the basis of the small
endothermicity, it is expected that 1Ϫ is a thermally labile
ؒ
species, which is readily decomposed to give 3Ϫ . To validate
ؒ
this expectation, however, estimation of the activation energy
for the extrusion of dinitrogen is required.
Because it is known that the reaction of singlet carbene with
methanol is diffusion-controlled and the concentration of
TMPD employed as an additive is much smaller than that of
methanol, 3S is ruled out as a candidate for an electron
acceptor in the photolysis of 1 in the presence of TMPD.
Although it is difficult to estimate the free-energy change for the
electron transfer from TMPD to 3T, this should be an endo-
thermic process. As discussed in the following section, the
oxidation potential of the carbanion 10 (0.01 V vs. Ag/Agϩ) can
be employed to estimate the reduction potential of 3T, which
Discussion
Mechanism of the direct photolysis of the diazoacetate 1 in the
presence of an amine
As described in the previous section, the distribution of the
products obtained in the photolysis of 1 in acetonitrile–
methanol was dependent on the amine employed as an additive.
It should be noted again that an increase in the electron-
donating ability of the amine resulted in a decrease in the yield
of the ether 2, which is the unique product originating from the
carbene 3. Especially, 2 could not be afforded by the irradiation
of 1 in the presence of TMPD, which has the lowest oxidation
potential of the amines employed in this study. The mechanism
of the photodecomposition of 1 in the presence of an additive
having a large electron-donating ability is now discussed.
is equal to the oxidation potential of 3Ϫ , leading to a value of
ؒ
ϩ3 kcal molϪ1 for the free-energy change for the electron trans-
fer from TMPD to 3T. Thus, it cannot be expected that this
electron transfer process proceeds at a rate comparable to that
of the reaction of 3T with O2 or solvent. However, it seems
possible that an encounter of 3T with TMPD in a solution
leads to the formation of the complex, in which an electron is
partially transferred from TMPD to 3T.
It should be emphasized that the dependence of the yield of 2
on [TMPD] shown in Fig. 1 suggests that a reaction scheme
where a single reactive species leading to the formation of 2 is
simply quenched by TMPD cannot be acceptable, no matter
which is the reactive species, 1* or 3. We wish to propose the
following two schemes which explain the nonlinear relationship
between the yield of 2 and [TMPD]. The first is a reaction
scheme in which both 1* and 3 are quenched by TMPD, which
is illustrated in Scheme 2. In the scheme, for the sake of
The most plausible one is the mechanism involving an elec-
tron transfer process. As mentioned in the introductory part,
Tomioka’s group proposed that an electron transfer occurred
in the photolysis of (4-nitrophenyl)diazomethane in amines.11
Moreover, in the field of azide photochemistry, the partici-
pation of an electron transfer was proposed in the photolysis of
4-nitrophenyl azide in amines.27,28 It is reasonable to think that
the diazoacetate 1 having an electron-withdrawing methoxy-
carbonyl moiety has a larger electron-accepting ability than
(4-nitrophenyl)diazomethane. Thus, it is conceivable that an
electron transfer process participates in the photolysis of 1 in
the presence of an electron-donating amine. The direct observ-
ation of TMPDϩ in the photolysis of 1 in the presence of
ؒ
TMPD affords strong evidence in support of this assumption.
There are two candidates for an electron acceptor in the
photolysis of 1 in the presence of TMPD: the electronically
excited diazoacetate 1* and the carbene 3. Upon excitation of 1,
1* is initially produced in its singlet excited state (1*S), which
undergoes intersystem crossing to its triplet excited state (1*T).
It is likely that the lifetime of 1*S is extremely short, because
the nitro group, as well as the methoxycarbonyl moiety, is
expected to accelerate the intersystem crossing.29 Thus, 1*S can
be ruled out as a candidate for the electron acceptor. The free-
energy change for the electron transfer from TMPD to 1*T is
calculated by the Rehm–Weller equation to be Ϫ26 kcal molϪ1
,
assuming that the energy level of 1*T is estimated by that of
nitrobenzene (58 kcal molϪ1). Therefore, it appears that the
electron transfer from TMPD to 1*T is an energetically favor-
able process, which is expected to proceed with a diffusion-
Scheme 2
controlled rate constant (1.9 × 1010
M
Ϫ1 sϪ1 in acetonitrile).20
simplicity, a mixture of 1*S and 1*T is shown by 1*, and 3
The carbene 3 is another candidate for an electron acceptor
in the photolysis of 1 in the presence of TMPD. The extrusion
of dinitrogen from 1*S and 1*T gives the carbene in its singlet
state 3S and triplet state 3T, respectively. It is known that
phenylcarbenes are energetically more stable in their triplet
state than in their singlet state, but a rapid interconversion
occurs between these two spin states.14 Moreover, it is generally
accepted that singlet carbene reacts with methanol to give the
product derived from the insertion of the carbene into the OH
bond, while triplet carbene is readily trapped with O2 to yield
represents an interconverting mixture of 3S and 3T. Moreover,
Ϫ
Dϩ
3
shows the radical ion pair produced by the electron
ؒ
ؒ
transfer, in which D stands for TMPD, assuming that 1Ϫ
ؒ
extrudes rapidly dinitrogen to give 3Ϫ (vide infra). P and Q
ؒ
represent the products originating from the electronically
Ϫ
excited state of 1 and Dϩ
3
, respectively. The products derived
ؒ
ؒ
from 3 other than 2 are depicted by R. In the scheme, kd, kdec
,
and ke represent unimolecular rate constants for nonradiative
deactivation of 1*, decomposition of 1* to give 3, and charge
Ϫ
recombination of the radical ion pair Dϩ
3
to reproduce 3,
ؒ
ؒ
J. Chem. Soc., Perkin Trans. 2, 2002, 1274–1282
1279

View Article Online
while kp, kq, kr, and ks show pseudo-first-order rate constants
for the reactions giving P, Q, R, and 2, respectively. The second-
order rate constants for the electron transfer from TMPD to 1*
and 3 are depicted by k1 and k2, respectively. The ratio of the
yield of 2 in the absence of TMPD to that in the presence of
TMPD, Φ(2)0/Φ(2), is given by eqn. (1) by using a steady-state
approximation.
The rate constants for fluorescence quenching by amines, kq, are
close to the diffusion-controlled limit (1.9 × 1010 MϪ1 sϪ1 in
acetonitrile),20 except for triethylamine. Therefore, taking into
account the concentration of the amine (50 mM), it is reason-
able to think that the singlet excited perylene is quenched not by
1, but by the amine predominantly. The large exothermicity of
the electron transfer process evaluated by the Rehm–Weller
equation,22 which is shown in Table 2, implies that the quench-
ing of singlet excited perylene by the amines proceeds by the
electron transfer mechanism. Moreover, the direct observation
Φ(2)0/Φ(2) = (1 ϩ α[D])(1 ϩ β[D])/(1 ϩ γ[D])
(1)
of PeϪ by the irradiation of perylene in the presence of DMA
ؒ
In this equation, α, β, and γ are equal to k1/(kp ϩ kd ϩ kdec),
k2kq/(kr ϩ ks)(ke ϩ kq), and k1ke/kdec(ke ϩ kq), respectively. The
non-linear least-squares analysis revealed that the experimental
data shown in Fig. 1 were reproduced by eqn. (1) with α = 0.70,
β = 0.72, and γ = 0, as shown with a solid line in Fig. 1. As
mentioned above, the electron transfer from TMPD to 1*T is
with the laser pulse provides unambiguous evidence supporting
the electron transfer mechanism. Therefore, it can be concluded
that when the amine having an oxidation potential lower than
DMA is employed as an additive, singlet excited perylene is
quenched by the amine by the electron transfer mechanism to
give PeϪ
.
ؒ
expected to be diffusion-controlled (k1 = 1.9 × 1010 MϪ1 Ϫ1).
s
Next, a discussion on the electron transfer process from PeϪ
ؒ
Therefore, if this mechanism is correct, the lifetime of 1*T in
acetonitrile–methanol, which is equal to 1/(kp ϩ kd ϩ kdec), is
evaluated to be 37 ps. Unfortunately, the experimental inform-
ation on the lifetime of 1*T which would serve to validate this
mechanism is not available at the present stage.
The second scheme we propose to explain the nonlinear
relationship between Φ(2)0/Φ(2) and [TMPD] is illustrated in
Scheme 3, in which the complex formed from 3 and TMPD is
to 1 is given. The free-energy change for this process is evalu-
ated to be Ϫ9.5 kcal molϪ1, which implies that the electron
transfer is an energetically favorable process. Furthermore, laser
flash photolysis studies revealed that accumulation of PeϪ pro-
ؒ
duced by the electron transfer from DMA to singlet excited
perylene was not observed in the presence of 1. These data
strongly suggest that electron transfer from PeϪ to 1 proceeds
ؒ
in the perylene-sensitized photolysis of 1 in the presence of
DMA. Because the exothermicity of the electron transfer pro-
cess increases with a lowering of the oxidation potential of the
amine, we conclude that in the presence of an amine having an
oxidation potential lower than DMA, 1Ϫ can be generated by
ؒ
electron transfer from PeϪ to 1.
ؒ
Thus, evidence is obtained supporting the generation of 1Ϫ
ؒ
in the perylene-sensitized photolysis of 1 in the presence of an
electron-donating amine such as DMA and TMPD. However, it
should be pointed out that the product distribution obtained in
the presence of DMA is completely different from that in the
presence of TMPD (Table 2). The former is essentially identical
to the product distribution for the reaction involving the
carbene 3 as a reactive intermediate, while the formation of 2,
which is a unique product originating from 3, is completely
quenched in the presence of TMPD. These results suggest that
Ϫ
the reactivity of 1Ϫ generated by electron transfer from Pe to
ؒ
ؒ
1 is significantly dependent on the amine used as an electron
donor to produce PeϪ
.
ؒ
Scheme 3
These results can be explained by a mechanism involving the
generation of the carbene 3 by the charge recombination
further quenched by TMPD to yield the free carbene radical
between carbene radical anion 3Ϫ generated by the extrusion
ؒ
anion 3Ϫ . In this scheme, k1 and k2 represent the second-order
ؒ
of dinitrogen from 1Ϫ and the radical cation of the amine used
ؒ
rate constants for the charge transfer from TMPD to 3 to afford
the complex [Dδϩؒؒ3δϪ] and for the reaction of TMPD with the
complex, respectively. Moreover, kϪ1 represents the rate con-
stant for the dissociation of the complex to regenerate 3. The
steady-state treatment leads to eqn. (2) for Φ(2)0/Φ(2),
as an electron donor. The total scheme we propose for the
perylene-sensitized photolysis of 1 in the presence of an
electron-donating amine D is illustrated in Scheme 4. As men-
tioned above, it is plausible that 1Ϫ is generated by the quench-
ؒ
Ϫ
ing of Pe* by D to give the contact ion pair [Dϩ Pe ], followed
ؒ
ؒ
by its dissociation into a free radical ion PeϪ
and electron
ؒ
Φ(2)0/Φ(2) = 1 ϩ α[D]2/(1 ϩ β[D])
(2)
transfer from PeϪ to 1. In the scheme, kb1, kb2[Dϩ ]s, and
ؒ
ؒ
kb3[Dϩ ]s are the rate constants for the charge recombination
ؒ
where α and β are depicted by k1k2/kϪ1(kr ϩ ks) and k2/kϪ1
,
processes, in which [Dϩ ]s stands for the steady-state concen-
ؒ
respectively. By non-linear least squares analysis, it was found
that the experimental data were also fitted to eqn. (2) with
α = 1.70 and β = 0.23, which is demonstrated with a dotted line
in Fig. 1. At the present stage, we have no decisive evidence to
determine which mechanism is appropriate for the photolysis of
1 in the presence of TMPD. In order to establish the mechan-
ism, further kinetic data on the reaction of 1* and 3 are
required.
tration of the amine radical cation. Moreover, P and Q
represent the products originating from 3 and 3Ϫ , respectively.
ؒ
On the basis of theoretical calculations which imply that the
endothermicity of the process of the extrusion of dinitrogen
from 1Ϫ is very small, we propose that 1Ϫ decomposes
ؒ
ؒ
immediately to give 3Ϫ . Although 3 reacts smoothly with O2
Ϫ
ؒ
ؒ
to afford 5, the charge recombination between 3Ϫ and the
ؒ
radical cation of the amine Dϩ to produce the carbene 3
ؒ
should be considered. The rate constant for the charge
recombination, kb3, could be dependent on the amine owing to
the dependence of the free-energy change for the charge
recombination process on its oxidation potential, Eox. As sug-
gested by the theoretical calculations, the electronic structure of
Mechanism of the sensitized photolysis of the diazoacetate 1 in
the presence of an amine
As described in the previous section, the fluorescence of peryl-
ene is effectively quenched by amines employed as additives.
1280
J. Chem. Soc., Perkin Trans. 2, 2002, 1274–1282

View Article Online
well as the nitro group, of 1 is thought to be essential for the
Ϫ
stabilization of the radical anion species 1Ϫ and 3 . Work is in
progress to elucidate the general scheme for the photochemical
generation of diazoalkane radical anions.
ؒ
ؒ
Experimental
General methods
1H and 13C NMR spectra were recorded on a JEOL JNM-A500
spectrometer at 500 and 125 MHz, respectively. UV–vis spectra
were obtained with a JASCO V-560 spectrophotometer. Fluor-
escence spectra were recorded on a JASCO FP-777 spectro-
fluorometer. Gel permeation liquid chromatography (GPLC)
was carried out on a JAI LC-08 equipped with a HAIGEL-H
column. Preparative thin–layer chromatography (PTLC) was
carried out on a Merck kieselgel 60 PF254
.
Scheme 4
Materials
3Ϫ , in which the negative charge is delocalized on the
π-conjugated system of the molecule, is similar to that of
the corresponding carbanion 10. If we employ Eox of 10 (0.01 V
ؒ
Methyl (4-nitrophenyl)diazoacetate (1) was prepared according
to the procedure reported by Schank.30 Perylene was purchased
from Tokyo Kasei Kogyo Co., Ltd., and was recrystallized
from benzene before use. Triethylamine, N,N-dimethylaniline,
N,N-dimethyl-4-chloroaniline, and N,N-dimethyl-4-methoxy-
aniline were purified by distillation under reduced pressure.
N,N,NЈ,NЈ-Tetramethyl-p-phenylenediamine was obtained by
treatment of its commercially available hydrochloride with
an aqueous dilute sodium hydroxide solution, and purified
by recrystallization from ether. 1,4-Diazabicyclo[2.2.2]octane
(DABCO) was recrystallized from ether before use. Aceto-
nitrile was refluxed with, and distilled from CaH2 prior to use.
Methanol was purified by distillation.
vs. Ag/Agϩ) as that of 3Ϫ , the free-energy change for the charge
ؒ
Ϫ
recombination between Dϩ and 3 is evaluated to be Ϫ19 and
ؒ
ؒ
Ϫ3 kcal molϪ1 for D = DMA and TMPD, respectively. Because
the electron transfer from D to 3 is mediated by excited peryl-
Ϫ
ene, the radical ions Dϩ and 3 are not in contact with each
other. Therefore, it is expected that the rate constant for the
charge recombination increases with an increase in the exo-
thermicity of the process. Thus, it is reasonable to think that
when DMA is employed as an electron donor, the charge
ؒ
ؒ
recombination between DMAϩ and 3Ϫ proceeds rapidly to
yield the carbene 3, which undergoes spin interconversion and
gives a mixture of products originating from both 3S and 3T.
ؒ
ؒ
On the other hand, the rate of the charge recombination
Irradiation for product identification
Ϫ
between TMPDϩ and 3 is considerably reduced, so that the
ؒ
ؒ
In a typical run, a solution of the diazoacetate 1 (5 mM) and
amine (50 mM) in acetonitrile–methanol (9 : 1, 3 mL) was
placed in a Pyrex tube, and degassed by bubbling of argon for
5 min. The solution was irradiated with a 500-W xenon arc
lamp at room temperature through a cutoff glass filter (> 390
nm) for 10 min. After removal of the solvent under reduced
pressure, the residue was separated by GPLC with chloroform
eluent and PTLC to yield a mixture of the photoproducts: the
α-methoxyacetate 2,13 the oxazole 4,15 the ketoester 5,16 and
methyl (4-nitrophenyl)acetate (7). The products were identified
by comparison of the spectroscopic data with those of authen-
tic material. When DABCO was employed as an additive, the
azine 8 was obtained in 45% yield. Dimethyl α,αЈ-azinobis-
reaction products derived from 3Ϫ are obtained without the
ؒ
formation of the products arising from 3.
Conclusions
We report our attempts to produce a diazoalkane radical anion
by a photoinduced electron transfer from an electron-donating
amine to the diazoalkane 1. On the basis of the photoproducts
and the results obtained by the laser flash photolysis studies, we
conclude the following three points concerning the formation
Ϫ
and reactivity of 1Ϫ and 3 . First, in the direct photolysis of 1
in the presence of TMPD in acetonitrile–methanol, electron
transfer from TMPD occurs and the photoproducts are
obtained through an intermediate other than the carbene 3,
ؒ
ؒ
1
[(4-nitrophenyl)acetate] (8): yellow granules; mp 50–52 ЊC; H
NMR (CDCl3) δ 4.40 (6H, s), 7.96 (4H, d, J = 9.0 Hz), 8.32 (4H,
d, J = 9.0 Hz); 13C NMR (CDCl3) δ 52.8, 124.1, 129.1, 136.5,
149.9, 160.0, 164.0; IR (KBr) 1727, 1596, 1589 cmϪ1. Anal.
Found: C, 52.05; H, 3.48; N, 13.43. Calcd for C18H14N4O8: C,
52.18; H, 3.41; N, 13.53%. Perylene-sensitized photolysis was
carried out as follows; a degassed solution of 1 (5 mM) and
amine (50 mM) in acetonitrile–methanol (9 : 1, 3 mL) contain-
ing perylene (0.3 mM) was placed in a Pyrex tube, and irradi-
ated with a 500-W xenon arc lamp at room temperature
through a cutoff glass filter (> 440 nm) for 3 h. The irradiation
mixture was treated with a procedure identical to that described
for the direct photolysis. In the perylene-sensitized photolysis
of 1 in the absence of an amine, not the oxazole 4, but the
isomer of 4 was obtained, together with 2 and 5. Although the
purification of the isomer was not successful, it was tentatively
identified as the isoxazole 9. 5-Methoxy-3-methyl-4-(4-nitro-
phenyl)isoxazole (9); 1H NMR (CDCl3) δ 2.17 (3H, s), 3.82 (3H,
s), 7.75 (2H, d, J = 8.7 Hz), 8.26 (2H, d, J = 8.7 Hz).
which we suppose to be 3Ϫ . Second, in the perylene-sensitized
ؒ
photolysis of 1 in the presence of an amine having a large
electron-donating ability such as DMA and TMPD, 1Ϫ can be
ؒ
generated by perylene-mediated electron transfer from the
amine to 1. It seems that 1Ϫ undergoes the extrusion of
ؒ
dinitrogen smoothly to give 3Ϫ , the behavior of which is very
ؒ
dependent on the amine employed as an additive. When DMA
is used the rapid charge recombination between DMAϩ and
ؒ
3Ϫ causes the generation of the carbene 3, while the photo-
ؒ
products derived from 3Ϫ are obtained in the presence of
ؒ
TMPD. Finally, 3Ϫ reacts readily with O2 to give the ketoester 5
ؒ
or abstracts hydrogen atoms to give 7. These results imply that
3Ϫ has highly radical character in its reactivity, which is not
ؒ
inconsistent with the theoretical prediction that 3Ϫ can be
ؒ
represented as the conjugate base of the 2-hydroxyvinyl radical
11.
Thus, our results have revealed that diazoalkane radical
anion, as well as carbene radical anion, can be generated photo-
chemically in solutions. At the present stage the structural
requirement of diazoalkanes for the formation of their radical
anions is not known, although the methoxycarbonyl moiety, as
Irradiation for analytical experiments
A solution identical to that described in the irradiation for
J. Chem. Soc., Perkin Trans. 2, 2002, 1274–1282
1281

View Article Online
2 O. W. Webster, J. Am. Chem. Soc., 1966, 88, 4055.
product identification was prepared and irradiated. After
removal of the solvent, the residue was dissolved in CDCl3. The
consumption of 1 (30–40%) and the yield of the photoproducts
3 P. F. Zittel, G. B. Ellison, S. V. O’Neil, E. Herbst, W. C. Lineberger
and W. P. Reinhardt, J. Am. Chem. Soc., 1976, 98, 3731.
4 R. N. McDonald, A. K. Choudhury and D. W. Setser, J. Am. Chem.
Soc., 1980, 102, 6491.
1
were determined by the integration of H NMR in the crude
reaction mixture. Identifications of the products were estab-
lished by the agreement of their 1H NMR spectra with those of
authentic samples.
5 R. N. McDonald, F. M. Triebe, J. R. January, Y. J. Borhani and
M. D. Hawley, J. Am. Chem. Soc., 1980, 102, 7867.
6 D. Bethell and V. D. Parker, J. Chem. Soc., Perkin Trans. 2, 1982,
841.
7 D. Bethell, L. J. McDowall and V. D. Parker, J. Chem. Soc., Chem.
Commun., 1984, 308; D. Bethell and V. D. Parker, J. Am. Chem. Soc.,
1986, 108, 7194.
8 D. Bethell and J. J. McDowall, J. Chem. Soc., Perkin Trans. 2, 1984,
1531; D. Bethell and V. D. Parker, J. Am. Chem. Soc., 1986, 108,
895.
9 D. Bethell and V. D. Parker, Acc. Chem. Res., 1988, 21, 400.
10 D. A. V. Galen, M. P. Young, M. D. Hawley and R. N. McDonald,
J. Am. Chem. Soc., 1985, 107, 1465.
11 H. Tomioka, K. Tabayashi and Y. Izawa, J. Chem. Soc., Chem.
Commun., 1985, 906.
Fluorescence quenching studies
Solutions of perylene (1.5 × 10Ϫ6 M) in acetonitrile containing
various amounts of quenchers were placed in 10 mm quartz
cells. Fluorescence spectra were measured at room temperature
on excitation at 380 nm. When TMPD was used as a quencher,
the light of 440 nm was employed for the excitation. Relative
fluorescence intensities were determined by measuring the peak
heights for the maxima.
12 For a preliminary report, see: T. Mizushima, S. Ikeda, S. Murata,
K. Ishii and H. Hamaguchi, Chem. Lett., 2000, 1282.
13 H. Tomioka, K. Hirai, K. Tabayashi, S. Murata, Y. Izawa, S. Inagaki
and T. Okajima, J. Am. Chem. Soc., 1990, 112, 7692.
14 For reviews of reactivities of carbenes, see: W. Kirmse, Carbene
Chemistry, Academic Press, New York, 1971; R. A. Moss and
M. Jones, Jr., Eds., Carbenes, Wiley, New York, 1973, 1975, Vols. 1
and 2; C. Wentrup, Reactive Molecules, Wiley, New York, 1984,
Chapter 4.
15 T. Ibata, T. Yamashita, M. Kashiuchi, S. Nakano and H. Nakawa,
Bull. Chem. Soc. Jpn., 1984, 47, 2450.
16 A. J. Muller, K. Nishiyama, G. W. Griffin, K. Ishikawa and D. M.
Gibson, J. Org. Chem., 1982, 47, 2342.
Laser flash photolysis studies
The laser flash photolysis experiments were performed by using
a pulse Q-switch Nd : YAG laser (SOLAR LF114) and Ti :
sapphire laser (SOLAR CF131M) as the excitation source.
Transient absorption signals were monitored by using a 10-W
xenon flash lamp (HAMAMATSU E2608). A sample solution
was placed in a flask, purged with N2 before and during
measurements, and flowed into a quartz cuvette (5 × 10 mm) by
a magnetic gear pump. The transient absorption spectra were
recorded in the wavelength range of 400 to 700 nm.
17 P. B. Grasse, B.-E. Rrause, J. J. Zupanicic, K. J. Kaufmann and
G. B. Schuter, J. Am. Chem. Soc., 1983, 105, 6833; R. L. Barcus,
L. M. Hadel, L. S. Johnston, M. S. Platz, T. G. Savino and
H. C. S. Scaiano, J. Am. Chem. Soc., 1986, 108, 3928; Y. Wang,
T. Yuzawa, H. Hamaguchi and J. P. Toscano, J. Am. Chem. Soc.,
1999, 121, 2875.
Cyclic-voltammetric measurements
Cyclic voltammograms were measured in acetonitrile solution
degassed by bubbling of N2 (2–4 mM) with a platinum elec-
trode, tetrabutylammonium perchlorate (0.1 M) as a support-
ing electrolyte, and a silver/silver chloride reference electrode.
Although the electrochemical reduction of 1 is irreversible as
shown in Fig. 4, the reversible voltammograms were recorded
on the oxazole 4, the ketoester 5, and the ester 7, the half-wave
potentials, E1/2, of which were determined to be Ϫ1.10, Ϫ0.73,
and Ϫ1.13 V (vs. Ag/Agϩ), respectively.
18 D. Bethell and L. J. McDowall, J. Chem. Soc., Chem. Commun.,
1984, 1408.
19 N. Yamamoto, Y. Nakano and H. Tsubonura, Bull. Chem. Soc. Jpn.,
1966, 39, 2603; S. Nakamura, N. Kanamaru, S. Nohara,
H. Nakamura, Y. Saito, H. Tanaka, M. Sumitani, N. Nakashima
and K. Yoshihara, Bull. Chem. Soc. Jpn., 1984, 57, 145.
20 S. L. Murouv, I. Carmichael and G. L. Hug, Handbook of
Photochemistry, 2nd edn., Marcel Dekker, New York, 1993.
21 N. T. Buu and J. T. Edward, Can. J. Chem., 1972, 50, 3730.
22 D. Rehm and A. Weller, Isr. J. Chem., 1970, 8, 259.
23 H. Tagaya, T. Aruga, O. Ito and M. Matsuda, J. Am. Chem. Soc.,
1981, 103, 5484.
24 X. Zhang, S. Yeh, S. Hong, M. Freccero, A. Albini, D. E. Falvey and
P. S. Mariano, J. Am. Chem. Soc., 1994, 116, 4211.
25 R. B. Davidson and M. L. Hudak, J. Am. Chem. Soc., 1977, 99,
3918.
Calculations
The DFT calculations were carried out by the program package
Gaussian 98 on a DEC Alpha workstation computer. The
geometry optimizations were performed at the UB3LYP/
6-31G(d) level of theory. The electronic energies calculated by
using the UB3LYP/6-31ϩG(d) level of theory were as follows.
Ϫ
1Ϫ : Ϫ812.131206, 3 : Ϫ702.6027575, 1: Ϫ812.0562134, 3T:
Ϫ702.4994941, N2: Ϫ109.5241739 hartree. These values include
zero-point energies. The charge distribution was calculated by
Mulliken population analysis.
ؒ
ؒ
26 R. A. Seburg, B. T. Hill and R. R. Squires, J. Chem. Soc., Perkin
Trans. 2, 1999, 2249.
27 T.-Y. Liang and G. B. Schuster, J. Am. Chem. Soc., 1986, 108, 546;
T.-Y. Liang and G. B. Schuster, J. Am. Chem. Soc., 1987, 109,
7803.
28 S. Murata, Y. Mori, Y. Satoh, R. Yoshidome and H. Tomioka,
Chem. Lett., 1999, 597.
References
29 H. Görner and D. Schulte-Frohlinde, Chem. Phys. Lett., 1986, 124,
1 T. Kauffmann and S. Hage, Angew. Chem., Int. Ed. Engl., 1963, 2,
156.
321.
30 K. Schank, Liebigs Ann. Chem., 1969, 723, 205.
1282
J. Chem. Soc., Perkin Trans. 2, 2002, 1274–1282