Mechanism for photocleavage of N-(anthroyloxy)-9-fluorenylideneamines and dynamic behavior of anthroyloxyl radicals

Article abstract of DOI:10.1246/bcsj.75.2013

Photocleavage of the N-O bond of N-(9-anthroyloxy)-9-fluorenylideneamine and its 1-and 2-anthroyloxy derivatives takes place efficiently in acetonitrile in the excited singlet state attributed to the fluorenylidene moiety. This made it possible for the first time to directly observe anthroyloxyl radicals by the transient absorption method. The quantum efficiency for photocleavage decreases remarkably in benzene, in which the lowest excited singlet state is attributed to the anthroate moiety. All the three kinds of anthroyloxyl radicals are much less reactive in decarboxylation, addition to olefins, and hydrogen-atom abstraction than benzoyloxyl and 1- and 2-naphthoyloxyl radicals; 9-anthroyloxyl radicals supposedly undergo intramolecular addition-elimination in the ipso-position in equilibrium with α-lactonic spirodihydroanthryl radicals, as indicated by one-color and two-color laser photolyses.

Full text of DOI:10.1246/bcsj.75.2013

c. Jpn.
© 2002 The Chemical Society of Japan
Bull. Chem. Soc. Jpn., 75, 2013–2023 (2002) 2013
e Chemical
ciety of Ja-
n
e Chemical
ciety of Ja-
n
Mechanism for Photocleavage of N-(Anthroyloxy)-9-fluorenylidene-
amines and Dynamic Behavior of Anthroyloxyl Radicals
Dynamic Behavior of Anthroyloxyl Radicals
Y. Saitoh et al.
Yasuo Saitoh, Makoto Kaneko, Katsunori Segawa, Hiroki Itoh,† and Hirochika Sakuragi*
02
13
23
SJA8
09-2673
256
Department of Chemistry, University of Tsukuba, Tsukuba, Ibaraki 305-8571
†Department of Material and Biological Chemistry, Faculty of Science, Yamagata University, Yamagata 990-8560
(Received August 3, 2001)
Photocleavage of the N–O bond of N-(9-anthroyloxy)-9-fluorenylideneamine and its 1-and 2-anthroyloxy deriva-
tives takes place efficiently in acetonitrile in the excited singlet state attributed to the fluorenylidene moiety. This made it
possible for the first time to directly observe anthroyloxyl radicals by the transient absorption method. The quantum ef-
ficiency for photocleavage decreases remarkably in benzene, in which the lowest excited singlet state is attributed to the
anthroate moiety. All the three kinds of anthroyloxyl radicals are much less reactive in decarboxylation, addition to ole-
fins, and hydrogen-atom abstraction than benzoyloxyl and 1- and 2-naphthoyloxyl radicals; 9-anthroyloxyl radicals sup-
posedly undergo intramolecular addition–elimination in the ipso-position in equilibrium with α-lactonic spirodihydro-
anthryl radicals, as indicated by one-color and two-color laser photolyses.
01
02
.2
Diaroyl peroxides are useful precursors of aroyloxyl radi-
cals and have been used for studies on various kinds of ben-
zoyloxyl and naphthoyloxyl radicals.1–10 As previously studied
in photolyses of the peroxides, the decarboxylation of benzoyl-
oxyl radicals is accelerated by introduction of methyl groups
or chlorine atoms in the two ortho-positions resulting in twist-
ing of the carbonyloxyl group from the molecular plane on ac-
count of steric hindrance.3,4 On the contrary, the decarboxyla-
Scheme 1.
tion rate of 1-naphthoyloxyl radicals is reduced by their conju-
gate ring system, which is wider than in the case of benzoylox-
Results and Discussion
yl radicals.5
Anthroyloxyl (anthracenecarbonyloxyl) radicals are attrac-
tive species to study, since they have a more expanded conju-
gate ring system and the carbonyloxyl moiety suffers varying
extents of steric hindrance to become more or less twisted
from the molecular plane depending on its attached ring posi-
tion. Among some anthroate (anthracenecarboxylate) esters of
aromatic and aliphatic ketoximes employed as precursors in-
stead of intractable peroxides, only fluorenone oxime esters
have been found to produce the anthroyloxyl radicals efficient-
ly in polar solvents. According to the previously proposed
triplet mechanism,11–13 the fluorenone oxime esters should be
less reactive than acetophenone and benzophenone oxime es-
ters.
In order to get further insight into the electronic and steric
effects on the reactivity of aroyloxyl radicals, we have firstly
clarified the mechanism of photocleavage of the radical precur-
sors, N-(anthroyloxy)-9-fluorenylideneamines, from the view-
points of solvent and excitation wavelength effects.14 Then we
studied the behavior of the anthroyloxyl radicals in solution by
means of transient absorption spectroscopy.15 In this paper we
discuss the photocleavage mechanism of the oxime esters and
the structure-reactivity relationships of 1-, 2-, and 9-anthroyl-
oxyl radicals.
1. Photocleavage of the Oxime Esters.
N-(9-Anthroyl-
oxy)-9-fluorenylideneamine (9AFA, Scheme 1) exhibits ab-
sorption bands around 300 and 365 nm in acetonitrile and ben-
zene. In comparison with the spectra of N-methoxy-9-fluoren-
ylideneamine (MFA) and 9-methoxycarbonylanthracene
(9MA), these bands are attributed to π-π* transitions of the flu-
orenylidene and anthroate moieties, respectively, though the
300-nm band is slightly blue-shifted from MFA’s band, and the
365-nm band slightly red-shifted from 9MA’s and overlapped
with a weak n-π* band of the fluorenylidene moiety (Fig. 1).
Similar features of the absorption spectra were observed for N-
(1-anthroyloxy)- (1AFA) and N-(2-anthroyloxy)-9-fluoren-
ylideneamines (2AFA, Scheme 2). However, spectral profiles
of the longer wavelength bands depend on the structure of
AFAs: structured for 2AFA and 9AFA, but structureless for
1AFA. These features are common to 1-methoxycarbonyl-
(1MA), 2-methoxycarbonylanthracenes (2MA), and 9MA.
9AFA was photolyzed with 365-nm stationary light in a low
concentration (1 × 10−3 mol dm−3) in acetonitrile under argon
at ambient temperature to give 9-anthroic (9-anthracenecar-
boxylic) acid (9AA; 0.92 mol/mol 9AFA) and 9-fluorenone
azine (FAz; 0.42) as products (Scheme 3). Similar irradiation
of 9AFA at 300 nm gave almost the same results, as shown in

2014 Bull. Chem. Soc. Jpn., 75, No. 9 (2002)
Dynamic Behavior of Anthroyloxyl Radicals
detected in the products at all with either 300- or 365-nm light,
·
9-AnCO2 seems not to undergo decarboxylation at ambient
temperature.
1AFA and 2AFA were also photolyzed under similar condi-
tions to give the corresponding acids (1AA and 2AA, respec-
tively) and FAz in good yields, as shown in Table 1. These re-
sults suggest that the excitation of AFAs in either chromophore
results in the N–O bond cleavage to give the iminyl and the
corresponding anthroyloxyl radicals, and that the former
dimerize to give FAz and the latter undergo hydrogen-atom ab-
straction from the solvent to give the acids.
Stationary photolyses of AFAs were also performed in ben-
zene. AAs and FAz were found to be the major products, but
their yields were much lower than those in acetonitrile (Table
1). Several minor products were detected by HPLC, but their
structures have not been determined.
Fig. 1. Absorption spectra of 9AFA, 9MA, and MFA in ace-
tonitrile.
2. Direct Observation of the Radical Species. Pulsed la-
ser photolysis of 9AFA (5 × 10−4 mol dm−3) at 308 nm under
argon in acetonitrile at ambient temperature exhibited transient
absorption bands at 400–500 and 500–800 nm, as shown in
Fig. 2. Similar laser photolysis of 9AFA at 360 nm gave al-
most the same spectrum.14 The longer-wavelength band dis-
played essentially identical decay time constants of 8–10 µs at
580, 630, 680, and 720 nm, and can be reasonably assigned to
9-AnCO2· by comparison with the spectra of 2,6-dimethylben-
zoyloxyl and other aroyloxyl radicals.2–8 For this band, molec-
ular oxygen neither reduced spectral intensity nor affected
spectral and decay profiles. These facts not only support the
above assignment of the oxygen-centered radicals but also in-
dicate that excited triplet states are not involved in the radical
formation, i.e., N-O bond cleavage. The spectral assignment
was also supported by two-color laser photolysis of 9AFA.
Upon 308 plus 580 nm laser photolysis, the latter pulse in-
duced a rapid decrease in intensity of the initially rising band,
as monitored at 570, 630, 680, and 720 nm. This phenomenon
Scheme 2.
Table 1. Product Yields (mol/mol AFA) in Photolysis of
AFAs in Acetonitrile and Benzene
AFA
Light/nm
AA
FAz
In acetonitrile
1AFA
300
366
300
366
300
366
0.93
0.90
0.92
0.93
0.83
0.92
0.44
0.41
0.45
0.45
0.41
0.42
·
can be ascribed to photoinduced decarboxylation of 9-AnCO2 ,
2AFA
9AFA
as described in Subsection 4. Thus, the whole longer wave-
·
length band is assigned to a single species, 9-AnCO2 .
An identical transient absorption band was also observed in
the 500–800 nm region on laser photolysis of 1-(9-anthroyl-
oxy)-2-pyridone (9APy; 3.0 × 10−4 mol dm−3) at 308 nm in
acetonitrile (Scheme 5, Fig. 3). In Fig. 3 an absorption band
due to 2-pyridyloxyl radicals is overlapped around 390 nm.16,17
This spectral identity confirms the above assignment, since a
series of 1-benzoyloxy- and 1-(naphthoyloxy)-2-pyridones
have been shown to be photochemically dissociated into
benzoyloxyl17 and naphthoyloxyl radicals18 and 2-pyridyloxyl
radicals. Actually, in the case of 9APy with 300-nm light, the
formation of 9AA (0.51 mol/mol 9APy) and 2-pyridone (0.61)
In benzene
1AFA
2AFA
300
300
300
0.14
0.14
0.26
0.22
0.18
0.35
9AFA
Table 1. The formation of FAz indicates intermediacy of 9-flu-
orenylideneaminyl radicals,11 which must be generated togeth-
er with 9-anthroyloxyl radicals (9-AnCO2 ) via homolysis of
·
the N-O bond of 9AFA (Scheme 4). Since anthracene was not
Scheme 3.

Y. Saitoh et al.
Bull. Chem. Soc. Jpn., 75, No. 9 (2002) 2015
Fig. 3. Transient absorption spectra observed on 308-nm
pulsed laser excitation of 9APy in acetonitrile under argon.
Scheme 4.
Fig. 4. Transient absorption spectra observed on 308-nm
pulsed laser excitation of 1AFA in acetonitrile under ar-
gon.
Fig. 2. Transient absorption spectra observed on 308-nm
pulsed laser excitation of 9AFA in acetonitrile under ar-
gon.
Scheme 5.
was confirmed in the product analysis with HPLC.
Pulsed laser photolysis of 1AFA and 2AFA (5 × 10−4 mol
dm−3) at 308 and 360 nm under similar conditions exhibited
characteristic transient absorption bands, as shown in Figs. 4
and 5, respectively. These bands can be assigned to 1- and 2-
anthroyloxyl radicals, respectively, since their spectra and time
profiles were not affected by oxygen at all, and essentially
identical bands were observed on laser photolysis of 1-(1-
anthroyloxy)- (1APy) and 1-(2-anthroyloxy)-2-pyridones
(2APy), respectively, at 308 nm under similar conditions. The
Fig. 5. Transient absorption spectra observed on 308-nm
pulsed laser excitation of 2AFA in acetonitrile under ar-
gon.
lifetimes of the observed radicals are summarized in Table 2.
3. Mechanism for Photocleavage of the Oxime Esters.
The quantum yields for photocleavage of 9AFA (1–3 × 10−4

2016 Bull. Chem. Soc. Jpn., 75, No. 9 (2002)
Dynamic Behavior of Anthroyloxyl Radicals
Scheme 6.
Table 2. Lifetimes of Anthroyloxyl Radicals Generated
from 360-nm Laser Photolysis of AFAs and 308-nm
Laser Photolysis of APys in Acetonitrile at Ambient Tem-
perature
Radical
Precursor
Lifetime/µsa)
·
·
·
1-AnCO2
1AFA
1APy
2AFA
2APy
9AFA
9APy
9.9
10
14
14
6.5
6.3
·
2-AnCO2
9-AnCO2
a) Monitored at 580 and 720 nm for 1-AnCO2 , 720 nm
for 2-AnCO2 , and 580 and 630 nm for 9-AnCO2 .
·
·
Table 3. Quantum Yields for Photocleavage of AFAs in
Acetonitrile and Benzene
Fig. 6. Transient absorption spectra observed on 308-nm
pulsed laser excitation of MeOBDPM in acetonitrile under
argon.
Solvent
AFA
Light/nm
MeCN
0.09
0.10
0.12
0.11
0.10
0.09
PhH
0.004
0.003
0.003
0.002
0.035
0.014
gy of biacetyl (236 kJ mol−1 19
) is higher than those of anthroic
1AFA
302
365
302
365
302
365
acid (177)19 and 9-fluorenone (211),19 this observation sug-
gests that the photocleavage of 9AFA takes place from the ex-
cited singlet state but not from the excited triplet state. This is
consistent with the fact that the spectral intensity of 9-anthro-
yloxyl radicals was not affected by molecular oxygen.
2AFA
9AFA
The involvement of the excited singlet state in the photo-
cleavage of the present oxime esters is in contrast to the triplet
mechanism, which was previously proposed for photocleavage
of a series of oxime esters such as N-(p-methylbenzoyl-
oxy)diphenylmethanimine.11–13 In this study we have carried
out pulsed laser photolysis of N-(p-methoxybenzoyloxy)-
(MeOBDPM), N-(9-anthroyloxy)diphenylmethanimines (9A-
DPM), and N-(9-anthroyloxy)isopropylideneamine (9AIPA)
(Scheme 6). Upon laser photolysis at 308 nm, MeOBDPM
showed absorption bands in the wavelength regions of 300–
400 and 600–800 nm (Fig. 6). In Fig. 6, the shorter and longer
wavelength bands are long- (lifetime >300 µs) and short-lived
(5.5 µs), respectively, and the latter is ascribed to p-methoxy-
benzoyloxyl radicals after comparison with the spectrum from
bis(p-methoxybenzoyl) peroxide.2,6 The former can be, there-
fore, ascribed to the pair radicals, diphenylmethaniminyls. For
9ADPM and 9AIPA, a strong band due to the anthroate triplets
was observed at <450 nm; however, the intensities were not
enough in the 500–800 nm region to identify the spectrum.
The structure dependence of the photocleavage efficiency can
be interpreted in terms of the triplet energies of the imines;13 in
the methoxybenzoyloxy derivative the lowest triplet is located
in the diphenylmethanimino moiety, but in the anthroyloxy de-
rivatives it is in the anthroate moiety. The triplet energy of the
mol dm−3) were determined from consumption of the starting
material with 302- and 365-nm light in acetonitrile and ben-
zene. The results are summarized in Table 3 together with
those for 1AFA and 2AFA. In the above concentration range,
the quantum yields were not affected. In acetonitrile the quan-
tum yields are around 0.1 for all the three AFAs and are inde-
pendent of the irradiation wavelengths. In benzene, however,
the quantum yields are lower, particularly for 1AFA and
2AFA, and for 9AFA the quantum yield depends on the irradi-
ation wavelength; it is higher at 302 nm than at 365 nm. It is
noticeable that the quantum yield is dependent on the solvents
and the excitation wavelengths, suggesting that the energy lev-
els of the two chromophores change between the polar and
nonpolar solvents.
Pulsed laser photolysis of 9AFA (5 × 10−4 mol dm−3) at
444 nm in acetonitrile in the presence of biacetyl (1 × 10−2
mol dm−3) as a triplet sensitizer exhibited an absorption band
around 430 nm with a lifetime of 7.8 µs, but none of the ab-
sorption bands assignable to 9-anthroyloxyl radicals was de-
tected. The observed band is very similar to the T-T absorption
of 9MA (lifetime 8.5 µs), and can be assigned to the triplet
state of the anthroate moiety. Since the triplet excitation ener-

Y. Saitoh et al.
Bull. Chem. Soc. Jpn., 75, No. 9 (2002) 2017
Table 4. Features of Fluorescence Spectra of AFAs and MAs in Acetonitrile and Benzene; Fluorescence Maxima
(λemax/nm), Singlet Excitation Energies (ES/kJ mol−1), Quantum Yields (φf), and Lifetimes (τf/ns)
AFA
Solvent
λemax
ES
φf
τf
0.46
8.0
0.86
7.5
1.1
1AFA
MeCN
PhH
MeCN
PhH
MeCN
PhH
MeCN
PhH
MeCN
PhH
MeCN
PhH
MeCN
PhH
457
463
438
463
463
465
453
446
435
432
459
455
466
461
464a)
462a)
465a)
462a)
485a)
471a)
284b)
282b)
289b)
285b)
291b)
290b)
291
291
297
297
301
286
282
291
285
296
291
0.02
0.23
0.02
0.04
0.01
0.05
0.64
0.54
0.97
0.90
0.31
0.69
0.001
0.001
2AFA
9AFA
1MA
2MA
9MA
MFA
7.8
15.5
8.8
22.7
16.6
8.2
12.4
3.4
299
291b)
293
293
4.1
a) Excitation wavelength: 300 nm. b) From fluorescence spectra observed on 300 nm excitation.
Fig. 8. Absorption, fluorescence, and fluorescence excitation
spectra of 9AFA in benzene. The spectra were normalized
at a band maximum except for the fluorescence spectrum
measured on 300-nm excitation; this spectrum was depict-
ed relative to the fluorescence spectrum measured on 364-
nm excitation.
Fig. 7. Absorption, fluorescence, and fluorescence excitation
spectra of 9AFA in acetonitrile. The spectra were normal-
ized at a band maximum except for the fluorescence spec-
trum measured on 300-nm excitation; this spectrum was
depicted relative to the fluorescence spectrum measured on
364-nm excitation.
photoreactive MeOBDPM is assumed to be larger than the dis-
sociation energy of the N-O bond, but those of 9ADPM and
9AIPA are assumed to be smaller.13
excited singlet state was confirmed by quenching of the 9MA
fluorescence by MFA; linear Stern-Volmer plots of the fluores-
cence lifetimes gave quenching rate constants (kq) of 2.4 ×
1010 and 1.5 × 1010 mol−1 dm3 s−1 in acetonitrile and benzene,
respectively. These diffusion-controlled rate constants are
quite reasonable if one takes into account the fact that the sin-
glet energy of 9MA is larger than that of MFA (vide infra) and
that one finds a very small spectral overlap of the 9MA fluores-
cence and MFA absorption.
The singlet excitation energies (ES) of 9AFA, 9MA, and
MFA were estimated from the absorption and fluorescence
spectra, as listed in Table 4. As mentioned above, 9AFA and
9MA exhibit similar fluorescence spectra. Their ES values are
estimated to be 296 and 301 kJ mol−1, respectively, from their
absorption and fluorescence spectra, which were measured by
exciting at 364 nm in acetonitrile. These values can be as-
cribed to the anthroate moiety. Upon 300-nm excitation,
The fluorescence spectra of 9AFA (1 × 10−4 mol dm−3) are
structureless and very similar to those of 9MA (1 × 10−4 mol
dm−3) in acetonitrile (Fig. 7) and benzene (Fig. 8), respective-
ly, on excitation at 364 nm. The fluorescence quantum yield
(φf) and lifetime (τf) of 9AFA were, however, much lower and
shorter, respectively, than those of 9MA in acetonitrile and
benzene, as shown in Table 4. The results were not affected up
to a substrate concentration of 1 × 10−4 mol dm−3. These re-
sults suggest that the anthroate singlet is efficiently quenched
by the fluorenylidene moiety. 1AFA and 2AFA showed almost
the same fluorescence quantum yields in acetonitrile as did
9AFA. The features of the fluorescence spectra are summa-
rized in Table 4.
The strong interaction between the two chromophores in the

2018 Bull. Chem. Soc. Jpn., 75, No. 9 (2002)
Dynamic Behavior of Anthroyloxyl Radicals
9AFA exhibited a weak fluorescence spectrum more than 20
nm red-shifted from that observed on 364-nm excitation (Fig.
7), and this spectrum gives a slightly lower ES value of 291 kJ
mol−1. This value is identical with that for MFA, and can be
ascribed to the fluorenylidene moiety.
The fluorescence excitation spectrum of 9AFA monitored at
the emission peak of 462 nm in acetonitrile shows a 365-nm
band profile almost identical with the absorption band but
slightly shifted to the shorter wavelengths, as shown in Fig. 7.
However, no clear peak appeared around 300 nm, though the
absorption spectrum exhibits a band due to the fluorenylidene
moiety in this region. These observations indicate that the
300-nm excitation results in the excited singlet state located in
the fluorenylidene moiety through internal conversion from a
higher excited singlet state, and that the 364-nm excitation
leads to the initial formation of the excited singlet state of the
anthroate moiety, followed by energy transfer to the fluoren-
ylidene moiety, to result in the fluorenylidene excited singlet;
the initially formed anthroate singlet is responsible for the flu-
orescence emission. Thus, the lowest (S1) and the second ex-
cited singlet state (S2) of 9AFA are assigned to the fluoren-
ylidene and anthroate moieties, respectively.
In benzene, the absorption spectrum of 9AFA is slightly red-
shifted compared with that in acetonitrile, and the fluorescence
spectrum is not affected by the excitation wavelengths of 300
and 364 nm. These facts suggest a significantly lower ES value
(290 kJ mol−1) for the anthroate moiety compared with the
above ES in acetonitrile. The excitation spectrum of 9AFA ex-
hibited a weak but clear band around 300 nm (Fig. 8), indicat-
ing that the initial excitation of the fluorenylidene moiety re-
sults in a partial population of the excited singlet in the an-
throate moiety. Thus the lowest excited singlet state is attribut-
ed to the anthroate moiety. Since the excited singlet level of
MFA is estimated to be about 293 kJ mol−1, the S2 state of
9AFA is assigned to the fluorenylidene moiety, but the energy
difference between S1 and S2 may be quite small. The rever-
sion of the two singlet energy levels due to the anthroate and
fluorenylidene moieties in acetonitrile and benzene is consis-
tent with the observed solvent dependence of the quantum
yields for photocleavage of 9AFA, lower in benzene and high-
er in acetonitrile (Table 3). Inspection of absorption and fluo-
rescence spectra of 9MA and MFA indicates that the difference
in the singlet energy between 9MA and MFA is larger in aceto-
nitrile than in benzene, though the singlet level of 9MA is
higher than that of MFA in both solvents.
Schematic energy diagrams are depicted in Fig. 9. In ben-
zene the lowest excited singlet states of the fluorenylidene
(S2(F)) and the anthroate (S1(A)) moieties are similar in ener-
gy. However, in acetonitrile the lowest excited singlet state of
the fluorenylidene moiety (S1(F)) is clearly lower than that of
the anthroate moiety (S2(A)) and may be responsible for the
N–O cleavage. Thus, the irradiation of 9AFA at 300 nm (ca.
400 kJ mol−1) in acetonitrile results in the formation of a high-
er excited singlet state of the fluorenylidene moiety (probably
S3(F); Fig. 9) followed by its internal conversion to the lowest
excited singlet state, S1(F) (ES ca. 290 kJ mol−1), which under-
goes N–O bond cleavage to give the radicals. The weak fluo-
rescence results predominantly from S1(F). On the contrary,
the irradiation at 365 nm produces the S2(A) state (ca. 300 kJ
Fig. 9. Schematic energy diagrams of 9AFA in the polar and
the nonpolar solvent.
mol−1), which undergoes fluorescence emission or energy
transfer to the fluorenylidene moiety to give S1(F), leading to
the N–O bond cleavage. In benzene, the S2(F) state produced
by 300-nm irradiation followed by internal conversion under-
goes predominantly energy transfer to the anthroate moiety to
produce S1(A), which undergoes fluorescence emission or in-
tersystem crossing to give eventually the T1(A) state; thus the
N–O cleavage is less efficient.
1AFA and 2AFA were also photocleaved more efficiently in
acetonitrile than in benzene (Table 3). The ES values of 9MA,
2MA, and 1MA decrease in this order (301 → 297 → 291 kJ
mol−1 in acetonitrile), as shown in Table 4. The ES values for
the anthroate moieties in 9AFA, 2AFA, and 1AFA are lower
than those of MAs, respectively, and also decrease in this order
(296 → 291 → 286 kJ mol−1). For 1AFA and 2AFA as well as
9AFA, the ES values ascribable to the fluorenylidene moiety
are slightly lower (291, 289, and 284 kJ mol−1 for 9AFA,
2AFA, and 1AFA, respectively) than ES’s due to the anthroate
moiety, as estimated from the absorption spectra and the fluo-
rescence spectra measured with 300-nm excitation light. Thus,
the features of the energy levels of S2(A) and S1(F) in 1AFA
and 2AFA are similar to those of 9AFA. In benzene, 1AFA
and 2AFA exhibited almost identical fluorescence spectra in-
dependently of the excitation wavelengths at 300 and 364 nm

Y. Saitoh et al.
Bull. Chem. Soc. Jpn., 75, No. 9 (2002) 2019
Scheme 7.
wavelength band, the increment of the slow component de-
creased with increasing the olefin concentration (cf. Subsec-
tion 6). This fact can be interpreted in terms of the reaction of
·
9-AnCO2 . with cyclohexene to reduce the intermediate forma-
tion.
As for the structure of the intermediate, 9,10-dihydro-9-an-
thryl radicals derived from 9,10-dihydroanthracene have an ab-
sorption band around 350 nm,20 and the present intermediate
may exhibit a band maximum at the wavelengths shorter than
400 nm. Thus, we tentatively assign this intermediate to an α-
lactonic
oxiran-2-one-3-spiro-9ꢀ-(9ꢀ,10ꢀ-dihydroanthracen)-
10ꢀ-yl radical, which is produced through an intramolecular
addition of the oxygen-centered radical to the ipso-position
(Scheme 7). As will be described below, a twisted geometry of
·
9-AnCO2 . may promote this process. Intermediacy of similar
α-lactonic radicals was previously proposed in the thermolysis
of di(1-naphthoyl) peroxide.21
Analysis of the time profiles of the transient absorptions
suggests that the α-lactonic radicals are in addition-elimina-
tion equilibrium with the original 9-AnCO2 . radicals, which
Fig. 10. Time profiles of the transient absorption spectra
monitored at different wavelengths on one-color and two-
color laser photolysis of 9AFA at 308 (+580) nm in aceto-
nitrile under argon.
·
eventually abstract a hydrogen atom to give the anthroic acid.
This is an intramolecular case of the reversible addition of
aroyloxyl radicals to aromatic substrates.22 As mentioned
above, the diphenylmethaniminyl radicals are assumed to have
an absorption band in the 300–400 nm region. For 9AFA and
other AFAs, however, no absorption bands assignable to 9-flu-
orenylideneaminyl radicals were detected in the wavelengths
longer than 400 nm.23
and at 300 and 358 nm, respectively; the spectra are due to the
anthroate moiety. The ES values of the S1 states are estimated
to be 282 and 285 kJ mol−1 for 1AFA and 2AFA, respectively.
The ES values of the S2 states for the fluorenylidene moiety are
assumed to be nearly 291 kJ mol−1 or slightly lower. Thus, the
energy difference between S2 and S1 may be larger in 1AFA
and 2AFA than in 9AFA. These observations are consistent
with the aforementioned solvent dependence of the quantum
yields, and similar reaction mechanisms may be operative in
the three isomers.
The addition-elimination equilibration of 9-AnCO2· was ex-
amined by means of the two-color laser photolysis. This is
·
based on the fact that the decarboxylation of 9-AnCO2 is in-
duced by light irradiation. Actually, on laser excitation of
9AFA with 308 nm pulses in acetonitrile, the yield of an-
thracene as a decarboxylation product increased and that of an-
throic acid decreased with increasing the laser power, as shown
in Table 5. The two-color photolysis of 9AFA with 308 plus
580-nm laser pulses also produced anthracene at the expense
of anthroic acid (Table 5).
4. Dynamic Behavior of 9-Anthroyloxyl Radicals.
In
Fig. 2, the shorter wavelength band observed in the laser pho-
tolysis of 9AFA may be ascribed to a plural species, such as
9AFA triplets, 9-fluorenylideneaminyl radicals, and an inter-
·
mediate derived from 9-AnCO2 . Inspection of the time pro-
files of the spectrum at 430 and 470 nm (Fig. 10; 308-nm pho-
tolysis) indicates that there are at least two rise components, a
fast one (within the instrumental detection limit) and a slow
one. The fast component can be assigned to 9AFA triplets on
the basis of the results that the fast rise was almost completely
quenched under oxygen and that the triplet sensitization with
biacetyl exerted only a T-T absorption due to the anthroate
moiety at 430 nm. The slow component grows with a time
constant of ca. 10 µs with a concomitant decay of the longer
wavelength band with an isosbestic point around 500 nm (Fig.
2), and then decays with a time constant of ca. 100 µs. The de-
cay of the longer wavelength band is composed of two compo-
nents with time constants of 8–10 and 90–100 µs. In the pres-
ence of cyclohexene, which accelerated the decay of the longer
Pulsed laser excitation of 9-AnCO2· in the two-color photol-
ysis of 9AFA with a 308-nm laser pulse followed by a 580-nm
laser pulse with a 35-µs delay caused a rapid loss in absorb-
ance, for example, at 570 and 680 nm and a gradual loss at 430
and 470 nm, as shown in Fig. 10. The rapid loss in absorbance
in the longer wavelengths corresponds to the decrease of 9-
·
AnCO2 concentration by photoinduced decarboxylation, and
the concomitant gradual loss in the shorter wavelengths is ra-
·
tionalized by a shift of the equilibrium to the 9-AnCO2 side
caused by the decrease of 9-AnCO2· concentration.
The time profile of the slow rise component in the shorter
wavelengths was not affected much under oxygen atmosphere.
This observation means that the α-lactonic radical is almost

2020 Bull. Chem. Soc. Jpn., 75, No. 9 (2002)
Dynamic Behavior of Anthroyloxyl Radicals
Table 5. Yields of Anthracene (AH) and Anthroic Acids
(AA) in 308 nm One-Color and 308 + 580 nm Two-Color
Laser Photolyses of AFAs in Acetonitrile
Table 6. Rate Constants (k) and Activation Parameters (Ea
and A) for Disappearance of Anthroyloxyl Radicals in
Acetonitrile
Power/mJ
pulse−1
Yield/%
Radical
k/106 s−1 a)
Ea/kJ mol−1
log (A/s−1
)
Precursor
9AFA
Laser/nm
308
·
·
·
AH
38
25
10
9
16
26
4
AA
52
68
73
67
62
70
84
77
1-AnCO2
2-AnCO2
9-AnCO2
0.10
0.07
0.15
0.42
2.1
12
17
23
31
32
31
7.3
8.0
9.2
11.1
11.7
12.1
40
13
4
·b)
·b)
1-NpCO2
2-NpCO2
PhCO2
360
308 + 580
308
4
·c)
4 + 7
13
40
4.5
9APy
1AFA
2AFA
a) At ambient temperature. b) Ref. 18. c) Ref. 2.
308
308
40
11
·
tion and decarboxylation of 9-AnCO2 are 55.6 and 96.7 kJ
mol−1, respectively.25 In the transition state of the ipso-addi-
tion, the distance between the ipso-carbon and the oxygen rad-
ical center is 175.7 pm, while the Cipso-Cα bond length is 189.6
pm in the transition state of the decarboxylation. These results
are in keeping with the experimental results that 9-AnCO2· un-
dergoes ipso-addition but not decarboxylation (cf. Subsection
6).
The endothermicity in peroxyl-radical formation (addition
to molecular oxygen) was estimated for some related radicals
by UHF/PM3. Calculations predicted that the oxygen adduct
of the α-lactonic radical is 41.4 kJ mol−1 higher in energy than
the α-lactonic radical itself (plus a triplet oxygen molecule).
In the case of the oxygen-sensitive 9,10-dihydro-9-anthryl rad-
ical, its oxygen adduct is at a level 38.8 kJ mol−1 higher than
the parent radical. Similar calculations for a 9-phenyl-9-fluo-
renyl radical, which has been reported to be unreactive towards
molecular oxygen,24 predicted that the energy of its oxygen ad-
duct would be 53.0 kJ mol−1 higher than that of the parent rad-
ical. These results suggest that the endothermicity may be one
of the factors suppressing the radical reaction with molecular
oxygen; however, in the case of the α-lactonic radical the mag-
nitude of the endothermicity is similar to that of the oxygen-
sensitive 9,10-dihydro-9-anthryl radical. Consequently, the
lack of oxygen-sensitivity in the α-lactonic radical can be as-
cribed to the exothermal elimination reverting to the aroyloxyl
radical, which is unreactive towards molecular oxygen.
6. Reactivity of Anthroyloxyl Radicals. The activation
parameters for disappearance of 1-, 2-, and 9-anthroyloxyl rad-
icals were determined by monitoring their absorption spectra
at 720 (for 1AFA and 2AFA) or 680 nm (for 9AFA) on 308-nm
laser photolysis of AFAs in acetonitrile in the temperature
range of 6.5–46 °C. The results are summarized in Table 6. In
·
Fig. 11. Calculated energy surfaces for 9-AnCO2 . The fig-
ures refer to energies in kJ mol−1 relative to that of 9-
AnCO2 .
·
unreactive towards oxygen. This apparent loss of reactivity
with oxygen might be due to its equilibration with the anthro-
yloxyl radical, which has no reactivity with oxygen (cf. Sub-
section 5). Recently, some diaryl- and triarylmethyl radicals
were reported to be unreactive towards molecular oxygen.24
5. Theoretical Approach to the Behavior of 9-Anthroyl-
oxyl Radicals. Geometries of the anthroyloxyl radicals were
optimized by use of UHF/PM3 and UB3LYP/6-31G* calcula-
tions. For 9-AnCO2· and 2-AnCO2· a geometry perpendicular-
ly twisted at the Cipso-Cα bond and a planar geometry were op-
timized, respectively, by both methods. For 1-AnCO2· a nearly
60° twisted geometry was optimized by the UHF/PM3 meth-
od; however, the DFT method predicted that the energy mini-
ma were located at a twisted and a planar geometry but the
former was slightly higher in energy than the latter. The DFT-
·
this Table the data for naphthoyloxyl (NpCO2 ) and benzoyl-
·
oxyl radicals (PhCO2 ) are also listed for comparison. The ac-
tivation energies of 31–32 kJ mol−1 for benzoyloxyl and naph-
thoyloxyl radicals correspond to their decarboxylation; howev-
er, those for 1- and 2-anthroyloxyl radicals are rather small
(12–17 kJ mol−1) and are assigned to their hydrogen-atom ab-
straction by comparison with those for o-iodobenzoyloxyl (11
kJ mol−1; log(A/s−1) = 8.5)17 and o-methybenzoyloxyl radi-
cals (17 kJ mol−1; log(A/s−1) = 10.5)26 which disappear
through the intermolecular and intramolecular hydrogen-atom
abstraction, respectively. The intermediate Ea value for 9-an-
throyloxyl radicals does not seem to correspond to either of
these processes, but seems to be consistent with the ipso-addi-
·
optimized geometry of 9-AnCO2 has a Cipso-Cα bond signifi-
cantly longer (151.2 pm) than those for 1- and 2-AnCO2•
(147.3 and 146.9 pm, respectively). As for the relative energy,
·
the planar 1- and the twisted 9-AnCO2 are 7.5 and 7.1 kJ
mol−1 higher than the planar 2-AnCO2 , respectively.
·
The energy surfaces for the reactions of 9-AnCO2• and the
α-lactonic radical were estimated as shown in Fig. 11. The
UHF/PM3 calculations predict that the energies (heats of for-
mation) of the α-lactonic radical and the anthryl radical (plus a
CO2 molecule) are 40.2 and 31.8 kJ mol−1 higher than that of
·
9-AnCO2 , and that the energy barriers (Ea) to the ipso-addi-

Y. Saitoh et al.
Bull. Chem. Soc. Jpn., 75, No. 9 (2002) 2021
Table 7. Rate Constants (k2/105 dm3 mol−1 s−1) for Reactions of Aroyloxyl Radicals with
Substrates in Acetonitrile
·
Ar in ArCO2
Substrate
1-An
—
—
43
3.3
2.2
2-An
—
—
9.1
1.4
1.1
9-An
11
2.6
5.5
0.1
0.3
1-Npa)
15400
610
290
Phb)
13000
4600
520
370
70
1,3-Cyclohexadiene
1,3-Pentadiene
Cyclohexene
Cyclohexane
Styrene
3.5
—
a) Ref. 18. b) Ref. 26.
tion. The ∆G‡ values estimated from the above activation pa-
rameters are in the range of 43–45 kJ mol−1 at 25 °C for all
three kinds of AnCO2· radicals.
zoyloxyl and naphthoyloxyl radicals in hydrogen-atom ab-
straction and addition to olefins. This means that the reactivity
of the aroyloxyl radicals is decreased by expansion of the con-
jugate ring system even if the conjugation is inhibited by a
heavily twisted geometry.
Bimolecular rate constants, k2, for the reactions of 1- and 2-
·
·
AnCO2 as well as 9-AnCO2 with substrates were evaluated
from pseudo-first-order decay rates obtained by monitoring the
transient absorption at 720 or 680 nm in acetonitrile: kexptl = k0
+ k2[substrate], where k0 refers to first-order reactions by
which the aroyloxyl radical decays at zero substrate concentra-
tion. In Table 7 the k2 values are compared with those of
benzoyloxyl27 and 1-naphthoyloxyl radicals.18 None of the
·
The 9-AnCO2 radicals are supposed to undergo the in-
tramolecular ipso-addition and to be in addition-elimination
equilibrium with the formed oxiran-2-one-3-spiro-9ꢀ-(9ꢀ,10ꢀ-
dihydroanthracen)-10ꢀ-yl radical. This equilibrium is not af-
fected in the presence of molecular oxygen, but is shifted to the
9-AnCO2· side by its photoinduced decarboxylation.
·
three kinds of AnCO2 radicals undergo decarboxylation at
ambient temperature, and they are much less reactive in hydro-
gen-atom abstraction and addition to olefins. These results in-
dicate that the reactivity of the aroyloxyl radicals is decreased
by expansion of the conjugate ring system even if the conjuga-
tion is insufficient as seen in 9-AnCO2· with a heavily twisted
geometry.
Benzoyloxyl radicals have been reported to photochemical-
ly decarboxylate.6 As mentioned above, 9-anthroyloxyl radi-
cals were found to decarboxylate with 308 and 580 nm light.
1-Anthroyloxyl and 2-anthroyloxyl radicals also decarboxylat-
ed with 308 nm light; however, the efficiency seems to be low-
er than 9-anthroyloxyl radicals (Table 5).
Experimental
General. Proton and carbon 13 NMR spectra were recorded
on a Varian Gemini-200 (200 and 50 MHz, respectively) or a
JEOL FX-270 (270 and 67.5 MHz, respectively) spectrometer.
UV absorption and fluorescence spectra were measured on a JAS-
CO Ubest 55 or a Shimadzu 1600 spectrophotometer and a Hita-
chi F-4000 spectrofluorimeter, respectively. Elemental analyses
were performed at the Analysis Center of the University of Tsuku-
ba.
Materials. Solvents, acetonitrile and benzene (Dojin, spec-
trograde) were used as received. 9-Anthroic acid was commer-
cially available (Aldrich). 2-Anthroic acid was prepared accord-
ing to a reference procedure from anthraquinone-2-carboxylic ac-
id,28 and 1-anthroic acid was also prepared according to reference
procedures from benzanthrone via anthraquinone-1-carboxylic ac-
id.28,29 1-Methoxycarbonylanthracene (1MA) and 2-methoxycar-
bonylanthracene (2MA) were obtained by esterification of the cor-
responding acid in methanol in the presence of sulfuric acid. 9-
Methoxycarbonylanthracene (9MA) was prepared by dehydration
from 9-anthroic acid and methanol with dicyclohexylcarbodiimide
(DCC) in the presence of 4-(dimethylamino)pyridine (DMAP) in
dichloromethane. N-Methoxy-9-fluorenylideneamine (MFA) was
prepared from the reaction of 9-fluorenone and methoxyamine hy-
drochloride in alkaline aqueous ethanol.
Conclusion
The excitation of AFAs in either the fluorenylidene or an-
throate moiety results in the N–O bond cleavage to give the
iminyl and the anthroyloxyl radicals; the former dimerize to
give 9-fluorenone azine and the latter undergo hydrogen-atom
abstraction from the solvent acetonitrile to give the anthroic
acids. The photocleavage of 9AFA takes place from the excit-
ed singlet state but not from the excited triplet state. The quan-
tum yield for the bond cleavage is higher in acetonitrile than in
benzene independently of the excitation wavelengths. In ace-
tonitrile the lowest excited singlet state is attributed to the fluo-
renylidene moiety, and is responsible for the N–O cleavage. In
benzene the excited singlet state of the fluorenylidene moiety
is almost the same in energy as that of the anthroate moiety.
Similar mechanisms are operative in the photocleavage of
1AFA and 2AFA.
N-(9-Anthroyloxy)-9-fluorenylideneamine (9AFA).
N-(9-
Anthroyloxy)-9-fluorenylideneamine was prepared by dehydra-
tion from 9-anthroic acid (2.22 g, 10 mmol) and 9-fluorenone
oxime (2.34 g, 12 mmol) with DCC (3.1 g, 15 mmol) in the pres-
ence of DMAP (122 mg, 1 mmol) in dichloromethane (200 ml).
9AFA was purified by silica-gel column chromatography (ethyl
acetate/hexane (1:10)), followed by crystallization from ethanol
(1.5 g, 38%); mp 169 °C; UV (CH3CN) λmax 301 nm (ε 11300),
365 (9700), 385 (8800); UV (PhH) λmax 304 nm (ε 10700), 367
(9800), 387 (9000); 1H NMR (CDCl3) δ 6.84–6.92 (m, 1H), 7.24–
7.58 (m, 9H), 7.77 (d, 1H, J = 8 Hz), 8.02–8.09 (m, 3H), 8.24–
8.30 (m, 2H), 8.62 (s, 1H); 13C NMR (CDCl3) δ 120.2, 123.6,
Dynamic behavior of the anthroyloxyl radicals was directly
observed for the first time by means of the transient absorption
method. The anthroyloxyl radicals are much longer-lived than
the benzoyloxyl and naphthoyloxyl radicals. None of the three
·
kinds of AnCO2 radicals undergoes decarboxylation at ambi-
ent temperature, and each is much less reactive than the ben-

2022 Bull. Chem. Soc. Jpn., 75, No. 9 (2002)
Dynamic Behavior of Anthroyloxyl Radicals
125.2, 125.7, 128.5, 128.7, 129.2, 130.0, 130.37, 130.43, 131.0,
131.9, 132.7, 134.3, 141.3, 142.6, 159.4, 166.6; Found: C, 83.90;
H, 4.27; N, 3.50%; Calcd for C28H17NO2: C, 84.19; H, 4.29; N,
3.51%.
silica-gel column chromatography (ethyl acetate/hexane (1:5))
followed by crystallization from ethanol (1.4 g, 72%); mp 171–
172 °C; UV (CH3CN) λmax 301 nm (ε 4600), 366 (7700), 386
(6800); 1H NMR (CDCl3) δ 6.33 (t, 1H, J = 8 Hz), 6.91 (d, 1H, J
10 Hz), 7.45–7.71 (m, 6H), 8.08 (d, 2H, J = 8 Hz), 8.61–8.66 (m,
3H); 13C NMR (CDCl3) δ 105.0, 121.1, 122.7, 124.5, 125.3,
127.5, 128.1, 129.2, 130.1, 131.0, 134.0, 134.8, 139.0, 156.9,
164.8; Found: C, 76.06; H,4.27; N, 4.45%. Calcd for C20H13NO3:
C, 76.18; H, 4.16; N, 4.44%.
1-(1-Anthroyloxy)-2-pyridone (1APy) and 1-(2-Anthroyl-
oxy)-2-pyridone (2APy). 1-(1-Anthroyloxy)-2-pyridone and 1-
(2-anthroyloxy)-2-pyridone were similarly prepared by dehydra-
tion from the corresponding anthroic acid and 2-hydroxypyridine
N-oxide with DCC in the presence of DAMP in dichloromethane
(78 and 82%, respectively).
1APy: mp 186–187 °C; UV (CH3CN) λmax 387 nm (ε 6200);
1H NMR (CDCl3) δ 6.29 (t, 1H, J = 8 Hz), 6.85 (d, 1H, J = 10
Hz), 7.41–7.59 (m, 5H), 7.99–8.08 (m, 2H), 8.31 (d, 2H J = 8
Hz), 8.50 (s, 1H), 8.62 (d, 1H, J = 10 Hz), 9.53 (s, 1H); 13C NMR
(CDCl3) δ 162.8, 157.0, 138.9, 135.7, 135.3, 132.5, 132.3, 131.1,
130.9, 128.5, 127.8, 127.3, 127.0, 125.9, 125.8, 124.3, 122.9,
122.5, 121.2, 104.6; Found: C, 76.08; H,4.22; N, 4.29%. Calcd
for C20H13NO3: C, 76.18; H, 4.16; N, 4.44%.
N-(1-Anthroyloxy)-9-fluorenylideneamine (1AFA) and N-
(2-Anthroyloxy)-9-fluorenylideneamine (2AFA).
N-(1-An-
throyloxy)-9-fluorenylideneamine and N-(2-anthroyloxy)-9-fluo-
renylideneamine were similarly prepared by dehydration from the
corresponding anthroic acid and 9-fluorenone oxime with DCC in
the presence of DMAP in dichloromethane (83 and 76%, respec-
tively).
1AFA: mp 156–157 °C (EtOH); UV (CH3CN) λmax 301 nm
(ε 13900), 384 (7500); UV (PhH) λmax 304 nm (ε 13200), 387
1
(7500); H NMR (CDCl3) δ 7.16–7.64 (m, 9H), 8.00–8.11 (m,
3H), 8.21–8.34 (m, 3H), 8.51 (s, 1H), 9.52 (s, 1H); 13C NMR
(CDCl3) δ 120.2, 120.4, 123.6, 123.7, 125.3, 126.0, 126.3, 126.4,
127.3, 127.8, 128.4, 128.5, 128.6, 129.1, 130.1, 130.4, 131.7,
131.8, 132.6, 132.8, 134.5, 134.6, 141.2, 142.7, 158.9, 164.9;
Found: C, 84.10, H, 4.27, N 3.50%. Calcd for C28H17NO2: C,
84.19; H, 4.29; N, 3.51%.
2AFA: mp 184–185 °C (EtOH); UV (CH3CN) λmax 302 nm
(ε 26100), 383 (4600); UV (PhH) λmax 302 nm (ε 25700), 384
1
(5700); H NMR (CDCl3) δ 7.42–7.67 (m, 10H), 7.75–8.16 (m,
3H), 8.39 (d, 1H, J = 8 Hz), 8.47 (s, 1H), 8.62 (s, 1H), 8.93 (s,
1H); 13C NMR (CDCl3) δ 120.5, 120.9, 124.0, 124.2, 126.5,
126.8, 127.3, 128.6, 128.9, 129.0, 129.4, 129.5, 130.4, 130.6,
132.2, 132.6, 133.1, 133.5, 133.8, 141.6, 143.2, 159.5, 164.8;
Found: C, 84.03; H, 4.41; N, 3.44%. Calcd for C28H17NO2: C,
84.19; H, 4.29; N, 3.51%.
2APy: mp 213–214 °C; UV (CH3CN) λmax 387 nm (ε 3000);
1H NMR (DMSO-d6): δ 6.39 (t, 1H, J = 8 Hz), 6.70 (d, 1H, J =
10 Hz), 7.54–7.68 (m, 3H), 7.98 (d, 1H, J = 8 Hz), 8.13–8.31 (m,
4H), 8.72 (s, 1H), 8.99 (s, 1H), 9.09 (s, 1H); 13C NMR (DMSO-d6)
δ 162.9, 156.4, 140.6, 137.4, 133.9, 133.3, 132.4, 131.8, 129.6,
128.6, 128.3, 127.5, 126.6, 123.2, 122.3, 121.6, 105.2; Found: C,
75.98; H, 4.21; N, 4.31%. Calcd for C20H13NO3: C, 76.18; H,
4.16; N, 4.44%.
Photolyses of AFAs and APys. A solution of a radical pre-
cursor in acetonitrile or benzene was irradiated with 300-nm light
(Rayonet RPR-3000 lamps) or 365-nm light (Riko 400-W high-
pressure mercury lamp with a CuSO4 solution filter), and the prod-
ucts were analyzed qualitatively and quantitatively by HPLC.
Quantum Yield Measurements. The quantum yields were
determined by irradiating solutions of AFAs in acetonitrile and
benzene with 302- or 365-nm light from a Bunkokeiki SM-5 su-
per-monochro-irradiator equipped with a 300-W xenon lamp.
Consumption of AFAs was monitored by HPLC, and the amount
of light quanta was determined by potassium tris(oxalato)fer-
rate(ꢁ) actinometry.30
N-(9-Anthroyloxy)diphenylmethanimine (9ADPM) and N-
(9-Anthroyloxy)isopropylideneamine (9AIPA). N-(9-Anthro-
yloxy)diphenylmethanimine and N-(9-anthroyloxy)isopropylide-
neamine were similarly prepared by dehydration from 9-anthroic
acid and benzophenone and acetone oximes, respectively, with
DCC in the presence of DMAP in dichloromethane (51 and 46%,
respectively).
1
9ADPM: mp 151–152 °C (EtOH); H NMR (CDCl3) δ 8.45
(s, 1H), 8.05–7.90 (m, 4H), 7.72–7.68 (m, 2H), 7.49–7.26 (m,
12H); 13C NMR (CDCl3) δ 124.5, 124.8, 126.4, 127.6, 127.9,
128.2, 128.3, 128.6, 129.0, 129.2, 130.3, 130.5, 131.9, 134.0,
166.2, 166.5; Found: C, 83.51; H, 4.81; N, 3.46%. Calcd for
C28H19NO2: C, 83.77; H, 4.77; N, 3.49%.
9AIPA: mp 170 °C (EtOH); 1H NMR (CDCl3) δ 1.97 (s, 3H),
2.18 (s, 3H), 7.45–7.59 (m, 4H), 8.02 (d, 2H, J = 8 Hz); 8.15 (d,
2H, J = 8 Hz), 8.54 (s, 1H); 13C NMR (CDCl3) δ 16.8, 21.6,
124.5, 125.0, 125.6, 126.6, 128.0, 128.3, 129.1, 130.3, 164.8,
166.5; Found: C, 77.90; H, 5.52; N, 4.97%. Calcd for C18H15NO2:
C, 77.96; H, 5.45; N, 5.05%.
Fluorescence Measurements. Fluorescence quantum yields
were determined by using anthracene as a standard (φf = 0.27 in
ethanol).31 Fluorescence lifetimes were measured by a single pho-
ton counting technique (Horiba NAES 110 time-resolved fluoro-
photometer).
N-(p-Methoxybenzoyloxy)diphenylmethanimine (MeOB-
Laser-Flash Photolyses. Laser-flash photolyses were per-
formed by using an excimer laser (Lambda Physik LPX-100, Xe-
Cl, 308 nm, 10-ns fwhm, 70 mJ/pulse) and a pulsed xenon arc
(Ushio UXL-159, 150 W) as a monitoring light source. The de-
tails of the apparatus for laser-flash photolysis have been de-
scribed elsewhere.32 The temperature of each sample cell was
controlled with a thermostated cell holder and a circulating water
bath (Haake F3-K). An excimer laser-pumped dye laser (Lambda
Physik FL3002) was used to obtain 360- (10-ns fwhm, 7–8 mJ/
pulse) and 444-nm laser pulses (10-ns fwhm, 12–13 mJ/pulse)
with DMQ and Coumarin 120 dyes, respectively. In two-color la-
ser photolyses, 580-nm laser pulses were supplied from an exci-
mer laser (Lambda Physik Lextra, XeCl, 308 nm)-pumped dye la-
ser (Lambda Physik Scanmate; dye: Rhodamine 6G, 10-ns fwhm,
DPM). N-(p-Methoxybenzoyloxy)diphenylmethanimine
was
prepared from p-methoxybenzoyl chloride and benzophenone
oxime in pyridine. MeOBDPM was crystallized from ethanol; mp
1
153–154.5 °C; H NMR (CDCl3) δ 3.81 (s, 3H), 6.82–6.85 (m,
2H), 7.36–7.77 (m, 12H); 13C NMR (CDCl3) δ 55.4, 113.7, 121.0,
128.2, 128.3, 128.5, 128.8, 129.0, 129.5, 130.8, 131.4, 131.6,
132.9, 134.7, 163.4, 163.5, 165.0; Found: C, 76.06; H, 5.13; N,
4.19%. Calcd for C21H17NO3: C, 76.12; H, 5.17; N, 4.23%.
1-(9-Anthroyloxy)-2-pyridone (9APy). 1-(9-Anthroyloxy)-
2-pyridone was prepared by dehydration from 9-anthroic acid
(1.38 g, 6.22 mmol) and 2-hydroxypyridine N-oxide (0.83 g, 7.46
mmol) with DCC (2.55 g, 12.4 mmol) in the presence of DMAP
(76 mg, 0.62 mmol) in dichloromethane; 9APy was purified by

Y. Saitoh et al.
Bull. Chem. Soc. Jpn., 75, No. 9 (2002) 2023
10 mJ/pulse).
18 T. Najiwara, K. Segawa, and H. Sakuragi, Chem. Lett.,
2001, 1064.
19 S. L. Murov, I. Carmichael, and G. L. Hug, “Handbook of
Photochemistry,” 2nd ed, Revised and Expanded, Marcel Dekker
(1993), Table 1.
20 I. W. C. E. Arends, P. Mulder, K. B. Clark, and D. D. M.
Wayner, J. Phys. Chem., 99, 8182 (1995).
21 J. E. Leffler and R. G. Zepp, J. Am. Chem. Soc., 92, 3713
(1970).
22 M. J. Perkins, in “Free Radicals,” ed by J. K. Kochi, John
Wiley, New York (1973), Vol. 2, Chap. 16, p. 231.
23 G. Bucher, J. C. Scaiano, R. Sinta, G. Barclay, and J.
Cameron, J. Am. Chem. Soc., 117, 3848 (1995).
24 J. C. Scaiano, presented in 2000 International Chemical
Congress of Pacific Basin Societies, Honolulu, December 14-19
(2000); Book of Abstracts ORGN 0577.
MO Calculations. Semiempirical (UHF/PM3) calculations
were performed to determine the optimized geometries of an-
throyloxyl and other radicals and to elucidate the potential energy
curves for decarboxylation and ipso-addition of 9-anthroyloxyl
radicals using a program of WinMOPAC V2.0.33 In each transi-
tion state for the ipso-addition and decarboxylation, only one
imaginary frequency was calculated. The eigen vectors with the
imaginary frequency mainly involve the approach of the oxygen
radical center to the ipso-carbon and the elimination of the CO2
group from the ipso-carbon, respectively. Density functional the-
ory (DFT; UB3LYP/6-31G*) calculations were performed to de-
termine the optimized geometries of anthroyloxyl radicals using
the Gaussian 98 program.34 All calculations were carried out on a
Gateway Japan G6-400 personal computer.
25 The DFT method gives smaller values of the energy barri-
ers and differences; for example, 62.8 kJ mol−1 as the energy bar-
rier to decarboxylation of 9-AnCO2· and 20.1 kJ mol−1 as the en-
References
1
S. Yamauchi, N. Hirota, K. Sawabe, H. Misawa, S.
·
Takahara, H. Sakuragi, and K. Tokumaru, J. Am. Chem. Soc., 111,
ergy difference between 9-An· and 9-AnCO2 . Detailed results
4402 (1989).
from the DFT calculations will be reported elsewhere together
with those for other aroyloxyl radicals.
26 J. Wang, M. Tsuchiya, T. Tateno, H. Sakuragi, and K.
Tokumaru, Chem. Lett., 1992, 563; J. Wang, M. Tsuchiya, K.
Tokumaru, and H. Sakuragi, Bull. Chem. Soc. Jpn., 68, 1213
(1995).
27 J. Hashimoto, K. Segawa, H. Itoh, and H. Sakuragi, Chem.
Phys. Lett., 314, 261 (1999).
28 P. Arjunan and K. D. Berlin, Org. Prep. Prog. Intern., 13,
368 (1981).
2
H. Misawa, K. Sawabe, S. Takahara, H. Sakuragi, and K.
Tokumaru, Chem. Lett., 1988, 357.
J. Wang, M. Tsuchiya, H. Sakuragi, K. Tokumaru, and H.
Itoh, Tetrahedron Lett., 35, 6321 (1994).
J. Wang, H. Itoh, M. Tsuchiya, K. Tokumaru, and H.
Sakuragi, Tetrahedron, 51, 11967 (1995).
T. Tateno, H. Sakuragi, and K. Tokumaru, Chem. Lett.,
1992, 1883.
3
4
5
6
J. Chateauneuf, J. Lusztyk, and K. U. Ingold, J. Am. Chem.
Soc., 110, 2877 (1988).
29 A. G. Perkin, J. Chem. Soc., 117, 706 (1920).
30 S. L. Murov, I. Carmichael, and G. L. Hug, “Handbook of
Photochemistry,” 2nd ed, Revised and Expanded, Marcel Dekker
(1993), p. 299.
7
J. Chateauneuf, J. Lusztyk, and K. U. Ingold, J. Am. Chem.
Soc., 110, 2886 (1988).
8
H. G. Korth, J. Chateauneuf, J. Lusztyk, and K. U. Ingold,
J. Org. Chem., 56, 2405 (1991).
31 W. R. Dawson and M. W. Windsor, J. Phys. Chem., 72,
3251 (1968).
9
S. Oishi, H. Tsubaki, and H. Matsuzawa, Chem. Lett.,
1999, 805.
32 H. Furuuchi, T. Arai, Y. Kuriyama, H. Sakuragi, and K.
Tokumaru, Chem. Phys. Lett., 162, 211 (1989).
33 WinMOPAC V2.0, Fujitsu, Tokyo, 1998.
10 T. Karatsu, Y. Yoshida, K. Fukui, and A. Kitamura, Chem.
Lett., 1999, 877.
11 H. Sakuragi, S. Ishikawa, T. Nishimura, M. Yoshida, N.
Inamoto, and K. Tokumaru, Bull. Chem. Soc. Jpn., 49, 1949
(1976).
12 M. Yoshida, H. Sakuragi, T. Nishimura, S. Ishikawa, and
K. Tokumaru, Chem. Lett., 1975, 1125.
13 H. Sakuragi, M. Yoshida, H. Kinoshita, K. Utena, K.
Tokumaru, and M. Hoshino, Tetrahedron Lett., 1978, 1529.
14 Y. Saitoh, K. Segawa, H. Itoh, and H. Sakuragi, Tetrahe-
dron Lett., 41, 8353 (2000).
15 Y. Saitoh, K. Segawa, H. Itoh, and H. Sakuragi, Chem.
Lett., 2001, 82.
34 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A.
Montgomery, Jr., R. E. Stratmann, J. C. Burant, S. Dapprich, J. M.
Millam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J.
Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C.
Pomelli, C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y.
Ayala, Q. Cui, K. Morokuma, D. K. Malick, A. D. Rabuck, K.
Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz, B. B.
Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R.
Gomperts, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C.
Y. Peng, A. Nanayakkara, C. Gonzalez, M. Challacombe, P. M. W.
Gill, B. Johnson, W. Chen, M. W. Wong, J. L. Andres, C.
Gonzalez, M. Head-Gordon, E. S. Replogle, and J. A. Pople,
“Gaussian 98, Revision A.6,” Gaussian, Pittsburgh, PA (1998).
16 B. M. Aveline, I. E. Kochevar, and R. W. Redmond, J. Am.
Chem. Soc., 118, 10124 (1996).
17 J. Hashimoto, K. Segawa, H. Itoh, and H. Sakuragi, Chem.
Lett., 2000, 362.