Oxidation of aryl-substituted cycloheptatrienes by photoinduced electron transfer

Article abstract of DOI:10.1039/a902662b

The photooxidation of the aryl-substituted cycloheptatrienes 7-(p-methoxyphenyl)cycloheptatriene (1a), 7-, 1- and 3-(p-dimethylaminophenyl)cycloheptatriene (1b, 1c and 1d) in deaerated acetonitrile solution at room temperature to the corresponding radical cations is accomplished by electron transfer to the electronically excited acceptors 9,10-dicyanoanthracene (A2), N-methylquinolinium perchlorate (A3), N-methylacridinium perchlorate (A4), 2,4,6-triphenylpyrylium tetrafluoroborate (A5) and 1,1′-dimethyl-4,4′-bipyridinium dichloride (A6). In the case of 1 back electron transfer is unimportant. The primarily formed radical cation deprotonates and the resultant cycloheptatrienyl radical undergoes a self reaction thus forming bitropyl. By contrast, 1b, 1c and, 1d decompose only very slightly, bitropyl is not formed and the acceptors are not consumed. Notably, the photooxidation of 1a in air-saturated acetonitrile solution containing HBF₄ (3.2 × 10^-2 M) and one of the acceptors results in the formation of tropylium salt. A tentative mechanism for this process is postulated.

Full text of DOI:10.1039/a902662b

View Article Online / Journal Homepage / Table of Contents for this issue
Oxidation of aryl-substituted cycloheptatrienes by photoinduced
electron transfer
Dirk Jacobi,a Werner Abraham,*a Uwe Pischel,a Lutz Grubert,a R. Stössera and
Wolfram Schnabelb
a Humboldt-University, Institute of Chemistry, D-10115 Berlin, Hessische Str. 1-2, Germany
b Hahn-Meitner-Institut GmbH, Department of Physical Chemistry, D-14109 Berlin,
Glienicker Str. 100, Germany
Received (in Cambridge) 1st April 1999, Accepted 18th June 1999
The photooxidation of the aryl-substituted cycloheptatrienes 7-(p-methoxyphenyl)cycloheptatriene (1a), 7-, 1- and
3-(p-dimethylaminophenyl)cycloheptatriene (1b, 1c and 1d) in deaerated acetonitrile solution at room temperature
to the corresponding radical cations is accomplished by electron transfer to the electronically excited acceptors
9,10-dicyanoanthracene (A2), N-methylquinolinium perchlorate (A3), N-methylacridinium perchlorate (A4),
2,4,6-triphenylpyrylium tetrafluoroborate (A5) and 1,1Ј-dimethyl-4,4Ј-bipyridinium dichloride (A6). In the case
of 1 back electron transfer is unimportant. The primarily formed radical cation deprotonates and the resultant
cycloheptatrienyl radical undergoes a self reaction thus forming bitropyl. By contrast, 1b, 1c and, 1d decompose
only very slightly, bitropyl is not formed and the acceptors are not consumed.
Notably, the photooxidation of 1a in air-saturated acetonitrile solution containing HBF4 (3.2 × 10Ϫ2 M) and one
of the acceptors results in the formation of tropylium salt. A tentative mechanism for this process is postulated.
ally excited acceptors in order to get insight into the mechanism
Introduction
of the stepwise oxidation of cycloheptatrienes to tropylium
salts. The experiments comprised various electron acceptors
and apart from 1a also 7-, 1- and 3-(p-dimethylaminophenyl)-
cyclohepta-1,3,5-triene (1b, 1c and 1d). The chemical structures
of these compounds are illustrated in Chart 1.
Recently, it has been shown that arylbitropyls can be trans-
formed into the related tropylium cations via the radical cations
of bitropyls generated by photoinduced electron transfer.1 In
the case of an aryl-substituted cycloheptatriene 1 cleavage of
one C–H bond in the seven-membered ring has to proceed in
order to enable a second oxidation step yielding the aryltro-
pylium ion. Commonly, deprotonation of radical cations is
observed.2
The generation of aryltropylium ions from arylcyclo-
heptatrienes is of special interest because various molecular
properties such as shape, color and electron acceptor strength
of the cycloheptatrienes differ strongly from those of the
tropylium cations. Accordingly, the properties of macrocycles3
illustrated in Scheme 1 and supramolecular systems such as
Since the singlet and triplet energies of the acceptors listed
in Chart 1 are lower than those of the arylcycloheptatrienes,
energy transfer from electronically excited A to 1, i.e. the
reaction A* ϩ 1→A ϩ 1*, can be excluded. Therefore, any
observable photoreaction between the reactants suggests the
occurrence of electron transfer. Actually, electron transfer is
quite feasible in these cases as can be inferred from the free
energy change ∆G0 which is estimated with the aid of the
et
Rehm–Weller equation (1).5 Here, the Coulomb term is
∆G0et = E0ox(1) Ϫ E0red(A) Ϫ ES or T(A)
(1)
neglected. E0ox(1) and E0red(A) denote the oxidation and the
reduction potential of 1 and A, respectively. ES or T(A) is the
singlet or triplet energy of electronically excited A. Relevant
hν, n X–
(A)
n X–
values of E0ox, E0 and ES or T are compiled in Table 1 and
red
from Table 2 it can be seen that in all cases electron transfer
from 1 to A* is highly exergonic and, therefore, its occurrence is
quite feasible.
CH2
CH2
OR
OR
n
n
The photo-induced electron transfer (PET) from 1 to elec-
tronically excited A results in the formation of the radical ions
ϩ and A Ϫ, as is depicted in Scheme 2. Back electron transfer
Scheme 1
1
᝽
᝽
(BET) from the reduced electron acceptor to the cyclo-
heptatriene radical cation is prevented if the latter undergoes
rotaxanes and catenanes containing cycloheptatriene building
blocks are expected to undergo significant changes in response
to the absorption of light.
ؒ
fast deprotonation thus forming radical 1 . Provided a suitable
ؒ
electron acceptor X is present radical 1 can be oxidized to the
Recently, the oxidation of 7-(p-methoxyphenyl)cyclohepta-
1,3,5-triene (1a) induced by electron transfer to photo-excited
triphenylpyrylium ion (A5) or to photo-excited p-benzoquinone
(A1), was studied4 and it turned out that the tropylium ion was
formed in the case of A5 provided that the UV-irradiation was
performed in aerated solution containing HBF4. This finding
has prompted us to study in some detail the donor behavior of
arylcycloheptatrienes during the electron transfer to electronic-
tropylium ion 3.
Regarding the experiments described below it was intended
to study the photo-reactivity of the acceptors listed in Chart 1
towards arylcycloheptatrienes with the aid of fluorescence
measurements. Moreover, we tried to detect intermediates by
applying the flash photolysis technique and to identify final
products by product analysis studies in conjunction with con-
tinuous irradiations.
J. Chem. Soc., Perkin Trans. 2, 1999, 1695–1702
1695

View Article Online
Table 1a Reduction potentials,a excitation energies and selected extinction coefficients of acceptors
A1
A2
A3
A4
A5
A6
E0red/V (vs. SCE)
E00/eV
Ϫ0.62
Ϫ0.89c
Ϫ0.85
Ϫ0.58
Ϫ0.50
Ϫ0.45
2.25 (T)b
2.86c
3.53d
2.77e
2.82 (S)f
3.08 (T)c
2.30 (T)c
ε347 nm/MϪ1 cmϪ1
2.37 × 103
3.28 × 104
a Own measurements (SCE) unless otherwise noted. b Ref. 8. c Ref. 29. d Ref. 24. e Ref. 30. f Ref. 31.
Table 1b Oxidation potentials and extinction coefficients of cycloheptatrienes
1a
1b
1c
1d
2b (3,3Ј)
cyclohepta-1,3,5-triene
1.43
Epa/Va (vs. SCE)
ε347 nm/MϪ1 cmϪ1
1.23
≈0
0.60
121.3
0.55
0.52
0.55
a Peak potential measured vs. the saturated calomel electrode (SCE) calibrated vs. the ferrocene/ferrocenium couple (0. 31 V).
Table 2 ∆Get (in eV) of the photo-induced electron transfer from 1 to excited acceptors and of the thermal electron transfer from cycloheptatrienyl
type radicals to acceptors in the ground state
A1
A2
A3
A4
A5
A6
1a
1b
Ϫ0.4 (T)
Ϫ1.0 (T)
Ϫ0.7 (S)
Ϫ1.4 (S)
Ϫ1.5 (S)
Ϫ2.1 (S)
Ϫ1.0 (S)
Ϫ1.6 (S)
Ϫ1.1 (S)
Ϫ0.6 (T)
Ϫ1.7 (S)
Ϫ1.2 (S)
0.14
Ϫ1.4 (T)
Ϫ2.0 (T)
ؒ
ؒ
1a
1b
0.26
0.11
0.53
0.38
0.49
0.34
0.22
0.07
0.09
Ϫ0.06
Ϫ0.01
Calculated according to eqn. (1) with the aid of the oxidation peak potentials listed in Table 1.
+•
R
NMe2
NMe2
R
R
PET
BET
+A–•
(A•)
+ A*([A+]*)
a: λmax270 nm
b: λmax310 nm
λ
max320 nm
λmax340 nm
+•
1
1
1
1d
1c
–H+
R
a) R = OMe
b) R = NMe2
R
R
CN
O
X–•
+
X
+
a: λmax435 nm
b: λmax567 nm
O
CN
3
A1
A2
1•
3
Ph
Scheme 2
Results
N
N
Ph
O
Ph
1. Fluorescence quenching
–
–
–
Me
Me
BF4
ClO4
ClO4
Experiments performed at room temperature with acetonitrile
solutions of the acceptors A2, A3, A4 and A5 showed that the
fluorescence of these compounds is effectively quenched by
both 1a and 1b. In the cases of A2 and A5 the dependence of
the fluorescence intensity on the quencher concentration was
measured. Stern–Volmer plots yielded straight lines and the fol-
lowing rate constants kq in MϪ1 sϪ1 are obtained with τS(A2) =
16 ns6 and τS(A5) = 4.2 ns7: 1 × 1010 (A2/1a), 2 × 1010 (A2/1b),
A3
A5
A4
Me
N
N Me
2 Cl-
A6
Chart 1
1696
J. Chem. Soc., Perkin Trans. 2, 1999, 1695–1702

View Article Online
Fig. 1 Transient absorption spectrum observed at the flash photolysis
of A2 (2.5 × 10Ϫ4 M) in deaerated acetonitrile solution containing 1a
(6 × 10Ϫ3 M). The spectrum was recorded 200 ns after the laser pulse.
Fig. 2 Transient absorption spectrum observed at the flash photolysis
of A5 (2 × 10Ϫ5 M) in deaerated acetonitrile solution containing 1a
(2.2 × 10Ϫ3 M). The spectrum was recorded 800 ns after the laser pulse.
3 × 1010 (A5/1a and A5/1b). These values are close to that calcu-
lated for encounter-controlled processes in acetonitrile at room
temperature (1.9 × 1010).8 From these results it can be inferred
that electron transfer occurs quite effectively.
yields and the triplet and singlet lifetimes are: ΦISC = 0.48 and
ΦF = 0.52, and τT = 10 µs11 and τS = 4.2 ns.7 Interestingly, both
1
3
excited states A5 and A5 can act as electron acceptors with
appropriate donors. At sufficiently low donor concentration
only triplet quenching occurs whereas both singlets and triplets
are quenched at substantially high donor concentration. The
2. Flash photolysis
The aim of these experiments was to detect radical ions or
radicals that were expected to be formed according to the
general mechanism depicted in Scheme 2 upon electron transfer
from 1a or 1b to electronically excited non-ionic or ionic
acceptors. For this purpose the non-ionic acceptor A2 and the
ionic acceptor A5 were selected as typical representatives of the
two acceptor classes and exposed to 20 ns flashes of 347 nm
light in argon-saturated acetonitrile solutions. The donor con-
centrations were chosen low enough to guarantee that the
photolyzing light was totally or almost totally absorbed by the
acceptor.
ؒ
absorption spectrum of the 2,4,6-triphenylpyranyl radical (A5 )
generated in this process possesses a broad absorption band
extending from about 440 to 600 nm with a peak around 540
nm.12 At the rather high concentration of 2.2 × 10Ϫ3 M 1a acts
as a quencher of both excited states and the transient absorp-
tion spectrum recorded 2 µs after the 20 ns flash (vide Fig. 2) is
ؒ
identical with the absorption spectrum of A5 reported in the
ؒ
literature. Apart from the band assigned to A5 the transient
absorption spectrum also exhibits a negative peak at about 410
nm indicating bleaching of the strong absorption band of A5
located at this wavelength.
At low concentration of 1a the characteristic optical absorp-
2.1 Experiments with A2. Fig. 1 shows the transient absorption
spectrum recorded with the system 1a/A2 at the end of the
flash. It has maxima at about 640 nm and 705 nm and resembles
fairly well the spectrum of the radical anion of A2 reported in
the literature.9 Obviously, electron transfer from 1a to A2*
3
tion of A5* between 600 and 800 nm11 was observed. In the
absence of 1a the triplet lifetime (measured at 750 nm) was 11.6
µs. It decreased with increasing concentration of 1a and on
the basis of the concentration dependence of the 1st order
halflife the rate constant kq (3A5* ϩ 1a) = 5 × 109
M
Ϫ1 sϪ1 was
occurs, but only the spectrum of the radical anion A2 Ϫ and not
᝽
that of the radical cation 1a ϩ is detectable.
obtained.
᝽
In another experiment performed with 1b at a rather low
concentration (1.9 × 10Ϫ5 M) the transient absorption spectrum
shown in Fig. 3 was recorded 1 µs after the flash. Apart from
the absorption band at λ < 600 nm this spectrum also contains
a rather strong band at λ > 600 nm with an intensity steadily
increasing up to the experimentally accessible limit (800 nm).
The formation of ionic species upon the reaction of A2* with
1a was substantiated by electrical conductivity measurements.
Simultaneously with the formation of the absorption due to
Ϫ
A2 a current was formed. Actually, a photocurrent was also
᝽
formed when A2 was irradiated in the absence of 1a with a 20 ns
flash of 347 nm light. In this case light absorption induces the
ejection of an electron from excited A2 molecules, i.e. some of
the excited A2 molecules undergo photoionization [eqn. (2)].
3
This part of the spectrum is assigned to the triplet state A5*.
Reportedly, the absorption spectrum of 3A5* possesses a
maximum at 850 nm.13 Fig. 3 also shows kinetic traces
recorded at 410, 540 and 790 nm. It can be seen that the
bleached absorption band at 410 nm recovers with a rate
equal to the decay rate of the triplet–triplet absorption at 790
nm. On the other hand it is noticeable that the absorption at
540 nm does not decrease down to zero, as does the absorp-
tion at 790 nm, but approaches a level definitely above zero.
A2* → A2 ϩ ϩ eϪ
(2)
᝽
Notably, the photocurrent was significantly increased when 1a
was present in the solution indicating that the reaction of A2*
with 1a results in the formation of ionic species. Both the decay
of the photocurrent and the decay of the intermediate radical
anion of A2 follow second order kinetics. The rate constant of
the decay of the radical anion of A2 was determined to be
ؒ
This indicates that A5 is formed concomitantly with the
decay of 3A5*.
3 × 108 MϪ1 sϪ1 (ε = 7700 MϪ1 cmϪ1 for 2 Ϫ 10).
᝽
2.3 Conclusions of the flash photolysis experiments. The
results described above yield evidence for the formation of the
Ϫ
2.2 Experiments with A5. Upon absorption of light the 2,4,6-
triphenylpyrylium cation (A5) is known to be transformed to
the long-lived excited triplet state 3A5* with a rather high yield.
Reportedly, the intersystem crossing and fluorescence quantum
radical anion A2 in solutions containing A2 and of the
᝽
ؒ
radical A5 in solutions containing A5. These intermediates are
expected to be formed upon electron transfer from 1a and 1b to
these acceptors. In both cases, it was, however, not possible
J. Chem. Soc., Perkin Trans. 2, 1999, 1695–1702
1697

View Article Online
Fig. 3 Transient absorption spectrum observed at the flash photolysis
of A5 (1.9 × 10Ϫ5 M) in deaerated acetonitrile solution containing 1b
(2 × 10Ϫ5 M). The spectrum was recorded 1 µs after the laser pulse.
Insets: kinetic traces depicting the decay of the optical density at 410
nm, 540 nm and 790 nm.
to detect the corresponding radical cations or intermediates
resulting from reactions of the radical cations.
3. Continuous irradiations
3.1 Photoexcitation in the absence of oxygen and acid. 3.1.2
HPLC Measurements. Degassed acetonitrile solutions contain-
ing 1a or 1b and A2, A4 or A5 were irradiated with UV light at
λinc = 365 or 405 nm, where A only is excited, and the progress
of the reaction was followed by high pressure liquid chrom-
atography (HPLC). In all experiments performed with 1a
bitropyls 2a were formed in high yield. Isomers of 2a were
identified with the help of authentic samples. By contrast,
bitropyl was not formed when solutions containing 1b and A
were irradiated. In this case, the HPLC analysis revealed that
the acceptors are photostable. 1b was slightly decomposed, but
the conversion was too low to detect reaction products.
Fig. 4 EPR spectra. (a) Spectrum recorded at Ϫ60 ЊC with the system
1b/A4 in deaerated solution, [1b] = 6 × 10Ϫ2 M, [A4] = 3 × 10Ϫ2 M.
Solvent: dioxane–hexafluoropropan-2-ol (75/25, v/v). (b) Spectrum of
the radical cation 1b ϩ simulated (program Compar 2) on the basis of
᝽
the hyperfine coupling constants given in the graph.
3.1.2 EPR Measurements. As pointed out above (Section 2.3)
the radical cations of 1a and 1b were not detectable in the flash
ϩ
photolysis experiments. The existence of 1b could be evi-
᝽
denced, however, when EPR measurements were performed
with the system 1b/A4 in another solvent (dioxane–hexafluoro-
propan-2-ol, 75/25, v/v) at Ϫ60 ЊC. As can be seen from Fig. 4
the EPR spectrum recorded upon irradiation of the solution
with the unfiltered light of a Hg/Xe-lamp compares fairly well
ϩ
with the simulated spectrum of 1b (obtained with the aid of
᝽
the program Compar 2). Apparently, the spectrum of 1b ϩ does
᝽
ؒ
not have the spectrum of the acridinyl radical A4 , which is
generated simultaneously with 1b ϩ, superimposed onto it,
᝽
14
ؒ
because A4 dimerizes rapidly.
3.2 Photoexcitation in the presence of oxygen and acid. Irradi-
ation of air-saturated acetonitrile solutions containing 1a and
A2, A3, A4 or A5 and in addition HBF4 (3.2 × 10Ϫ2 M) at
λinc = 365 or 405 nm resulted in the formation of tropylium ion
3a and H2O2 (detected according to the method of Fukuzumi et
al.15). Typical results are shown in Fig. 5 where optical absorp-
tion spectra of the system 1a/A4 recorded before and after
irradiation for given times are presented. Obviously, the bands
characteristic of the tropylium ion at 230, 275 and 430 nm grow
with increasing time of irradiation.
Fig. 5 Optical absorption spectra observed at the photooxidation of
1a (8 × 10Ϫ5 M) upon irradiation of A4 (3 × 10Ϫ5 M) in air-saturated
acetonitrile solution containing HBF4 (3.2 × 10Ϫ2 M).
lengths. In this respect it is notable that analogous observations
(according to UV-Vis and HPLC-studies) were made with the
other acceptors used in these experiments (A2, A3 and A5),
i.e. they were not consumed while 3a and H2O2 were formed.
Notably, 3a was not formed in air-equilibrated solutions in the
absence of protons.
On the other hand, the intensity of the bands characteristic
of A4 at 250 nm and 360 nm does not decrease simultaneously
with the increase in the absorption due to the tropylium ion
indicating that the acceptor is not consumed. Actually, the
bands at 250 nm and 360 nm become larger to some extent.
This increase is assigned to the contributions of the absorption
of the tropylium ion to the total absorption at these wave-
The quantum yield of the formation of 3a, Φ(3a), was meas-
ured in the case of the system 1a/A5 and it turned out (see Fig.
6, inset) that Φ(3a) increases with increasing concentration of
1a and eventually approaches a limiting value of about 0.3.
Moreover, it can be seen from Fig. 6 that the reciprocal quan-
tum yield depends linearly on the reciprocal concentration of
1a.
1698
J. Chem. Soc., Perkin Trans. 2, 1999, 1695–1702

View Article Online
absorption spectra of these species are relatively weak, i.e.
Discussion
1. Photooxidation in the absence of oxygen and acid
undetectable in the present work. In the case of systems of type
ϩ
1a/A evidence for the generation of radical cations 1 comes
᝽
from the product analysis studies according to which 2a is
The photoinduced electron transfer from 1 to electronically
excited acceptors A is expected to be favored in all cases accord-
formed in high yield. According to the mechanism presented
ϩ
ing to the free energy change ∆G0 calculated with the aid of
in Scheme 3 radical cations 1a deprotonate rapidly and the
᝽
et
cycloheptatrienyl radicals generated in this way combine thus
eqn. (1) (vide Table 2). Indeed, the occurrence of a fast reaction
between photo-excited acceptors A2 and A5 and the donors 1a
and 1b was evidenced by the results of the flash photolysis and
the fluorescence quenching experiments, and this fast reaction
is very likely to involve electron transfer. The fact that time
resolved absorption spectroscopy permitted the detection of
radical anions or radicals only and not that of the correspond-
ing radical cations is puzzling. Probably, the absorptivity of
the radical cations is low and, therefore, the transient optical
forming 2a. Actually, the fate of the radical cation 1 ϩ depends
᝽
on the competition between various reactions. Apart from
deprotonation it can undergo addition to the solvent, dimeriz-
ation and back electron transfer. Regarding deprotonation a
ϩ
prognosis on the basis of pKa-values of the radical cations 1
᝽
is not possible since those values are not accessible. On the
other hand, the deprotonation of 1a must be an important
process, provided the mechanism depicted in Scheme 3 holds,
since 2a is formed in a high yield. Analogously, it is concluded
that back electron transfer is unimportant in this case. Princi-
ϩ
᝽
ؒ
pally, the radical 1a could be converted to 3a provided the
acceptor (in the ground state) used for the photooxidation of 1a
would also be capable of undergoing reaction (3). However, it
Ϫ
ؒ
1a ϩ A → 3a ϩ A
(3)
᝽
can be excluded that the A compounds tested in this work can
play this role because of the endergonicity of reaction (3). As
can be seen from Table 2 the ∆G0 values are positive in all
et
ؒ
cases. Therefore, radicals 1a are very likely to dimerize.
Regarding the compounds 1b, 1c and 1d it is noteworthy that
neither dimer 2b nor the tropylium ion 3b is formed. Actually,
ؒ
the oxidation of radical 1 is feasible as far as the system 1b/
A5 is concerned because of the exergonicity (∆G0et = Ϫ0.01 eV)
ؒ
of the reaction 1b ϩ A5, but 3b is not formed. Therefore, it is
ϩ
ϩ
concluded that the radical cations 1b ϩ, 1c and 1d do not
᝽
᝽
᝽
deprotonate but undergo back electron transfer leading to the
bimolecular decay of the reduced acceptors observed with A2
and A5. Also in the case of the electron transfer from 1b to A5
in the triplet state the back electron transfer is fast. Therefore,
a delay between the decay of the triplet state of A5 and the
recovery of A5 from its radical was not observed (see Fig. 3).
A back electron transfer faster than the deprotonation of
the radical cation is feasible because of the diminished acid
strength of these radical cations (relative to 1a ϩ) and also
᝽
because of the smaller exergonicity of the back electron trans-
fer (Ϫ1.1 eV) with respect to 1a (Ϫ1.7 eV). Assuming that the
back electron transfer occurs in the Marcus inverted region it
should proceed faster than in the case of 1a.
Fig. 6 Dependence of the reciprocal quantum yield of 3a formation
vs. the reciprocal concentration of 1a. Inset: plot of the quantum yield
of 3a formation on the concentration of 1a.
2. Photooxidation in presence of oxygen and acid
Concerning the photolysis in acidic, air-equilibrated aceto-
1a
OMe
OMe
Ph
*
Ph
PET
BET
–
BF4
+
+
+
Ph
O
Ph
Ph
O
Ph
–
BF4
A5
1a
A5
–H+
1/2
OMe
OMe
MeO
1a
2a
Scheme 3
J. Chem. Soc., Perkin Trans. 2, 1999, 1695–1702
1699

View Article Online
ؒ
nitrile solution O2 is regarded to play the role of a one-electron
oxidant which enables the oxidation of the cycloheptatrienyl
radical. A mechanism illustrating how the acceptor A5 acts as a
photosensitizer for the oxidation of 1a by O2 is proposed in
Scheme 4. The essential feature of this mechanism is the elec-
assumed that in the presence of a protonic acid the radical A5
is oxidized to A5 by O2. Principally, the oxidation of 1a to 3a
can be initiated both by singlet- and triplet-excited A5 mole-
cules as is depicted in Scheme 5.
At relatively low concentration of 1a the singlet pathway is
unimportant and practically only A5* molecules are involved
3
in the formation of 3a. It is presumed that the rate of the reac-
1a
ؒ
H+
tion 1a ϩ O2 is much larger than the rate of deprotonation
A5 + 1a
ϩ
ϩ
ؒ
(1a → 1a ϩ H ), i.e. the rate of the latter reaction deter-
᝽
A5*
mines the formation rate of 3a. Then, eqn. (9) was derived on
the basis of the steady-state approximations (6)–(8).
hν
A5 + 1a
A5
ϩ
kET[1a][3A5*] = (3kBET ϩ 3kbr)[1a
]
(6)
(7)
᝽
O
2 + H+
HO2
kisc[1A5*] = (3kET[1a] ϩ 3τϪ1)[3A5*]
A5
O
2 + H+
3a + HO2
Φ(3a) =
HO2
H2O2 + O2
3τkisc kbr kET[1a]/(1 ϩ 3kET3τ[1a])(1kd ϩ kisc)(3kBET ϩ 3kbr) (8)
A transformation of eqn. (8) leads to eqn. (9).
1/Φ(3a) = (1kd ϩ kisc)(3kBET ϩ 3kbr)/kisc3kbr ϩ
3
3
Scheme 4
ؒ
tron transfer from 1a to O2 and the subsequent protonation of
O2 [reactions (4) and (5)]. At a sufficiently high proton concen-
Ϫ
(1kd ϩ kisc)(3kBET ϩ 3kbr)/kisc kbr3τkET[1a] (9)
ؒ
3
Ϫ
With a = (1kd ϩ kisc)(3kBET ϩ 3kbr)/kisc kbr, and m = (1kd ϩ kisc)-
ؒ
ؒ
3
1a ϩ O2 → 3a ϩ O2
(4)
(5)
(3kBET ϩ 3kbr)/kisc kbr3τkET, one obtains eqn. (10) and a/m = kET
3
Ϫ
ϩ
ؒ
ؒ
O2 ϩ H
HO2
1/Φ(3a) = a ϩ m/[1a]
(10)
tration the equilibrium (5) is shifted to the right side (in water
pK = 4.88;16 in acetonitrile pK = 1217) thus guaranteeing irre-
versibility of electron transfer according to reaction (4). In the
absence of acid the occurrence of reaction (5) is rather improb-
able because of its endothermicity, i.e. the oxidation potential
3τ and kET = (a/m)/3τ. A plot of 1/Φ(3a) vs. 1/[1a] yields a
straight line as can be seen from Fig. 6. With m and a, the slope
and the intercept with the ordinate of the straight line, respect-
ively, kET = 9 × 108 MϪ1 sϪ1 is calculated using τ = 5.4 µs (3τ:
3
ؒ
3
of 1a (Ϫ0.36 V) is more positive than the reduction potential
lifetime of A5* in air-equilibrated acetonitrile). The kET value
of O2 (Ϫ0.86 V).18 However, the occurrence of reaction (4)
becomes feasible if the reduction potential of O2 effective in the
presence of HBF4 is larger than Ϫ0.86 V. Actually, a shift of
the reduction potential of O2 to more positive values upon the
addition of acid has been proven by various authors, e.g. by
Sawyer et al.19 and by Fukuzumi.20
agrees satisfactorily with the kq value (5 × 109 MϪ1 sϪ1) obtained
from the triplet quenching experiments (see Results 2.2). The
extrapolated quantum yield at infinite quencher concentration
amounts to 0.25.
Conclusions
Regarding the mechanism depicted in Scheme 4 it should be
noted that the observation that the acceptor is not consumed
during the oxidation of 1a to 3a has been accounted for. It is
1. This paper demonstrates that the photooxidation of aryl-
substituted cycloheptatrienes to tropylium ions involves two
2a
kdim
1kbr
1
A5 1a
A5 + 1a
- H+
H+, O2
fast
2a
1kET + 1a
3a + HOO
kdim
1kBET
3kbr
3kET
kisc
1A5*
3
3A5*
A5 1a
A5 + 1a
1a
H+, O2
1kd
3τ-1
fast
hν
3kBET
3a + HOO
A5
Scheme 5
1700
J. Chem. Soc., Perkin Trans. 2, 1999, 1695–1702

View Article Online
electron transfer steps. The process is governed by the chemical
nature of the aryl substituent. The p-dimethylaminophenyl sub-
stituent on the seven-membered ring hinders the photoinduced
oxidation to the tropylium salt.
2. In the case of 1a deprotonation of the radical cation com-
petes successfully with back electron transfer to the reduced
electron acceptor thus making the forward electron transfer
irreversible. However, back electron transfer dominates if a
powerful donor substituent, namely the dimethylaminophenyl
group, is attached to the seven-membered ring and the form-
ation of the cycloheptatrienyl radical is prevented. This
obstacle can be circumvented by protonation of the amino
group.21
using a HEKA Elektronik potentiostat (model PG 285) with
the aid of an EG&G cell having a planar working electrode
(d = 1 mm). A saturated calomel reference electrode (located in
a lugging capillary with diaphragm) was employed. The reduc-
tion potential E0red(3a) = Ϫ0.35 V was calculated from the
measured peak potential Ep = Ϫ0.36 V according to eqn. (11).22
E0red(3a) = Ep Ϫ RT/{ln(k2[1a]RT/nFv) Ϫ 3.12}/3F (11)
Here, F is 96.485 C molϪ1, k2 = 2 × 106 MϪ1 sϪ1 23 is the rate
constant of the dimerization of 1a, n = 1, v = 1 V sϪ1 (scan rate)
and [1a] = 1 × 10Ϫ3 M. Analogously, E0red(3b) = Ϫ0.51 V was
ؒ
ؒ
obtained. The oxidation potentials of the radicals 1a and 1b
3. In the case of 1a it was shown for the first time that
tropylium cations 3a can be efficiently generated by photo-
induced electron transfer. The photooxidation is accomplished
with the support of protons and dioxygen. We have shown that
macrocycles such as calixarenes incorporating the cyclohepta-
triene moiety can be transformed into the tropylium derivative
with the aid of this method. Work to study the photochemically
induced alteration of the complexation properties of these
macrocycles is currently in progress.
were assumed to be equal to the reduction potentials of the
corresponding tropylium ions 3a and 3b, i.e. E0ox(1a ) =
ؒ
0
E0red(3a) and E0ox(1b ) = E red(3b).
ؒ
4. Materials
Acetonitrile (Riedel de Haen, HPLC grade) was used as
received. The acceptors A2 and A5 were commercial products.
A324 and A425 were synthesized as described in the literature.
Cyclohepta-1,3,5-triene (Fluka) was purified by rectification.
The cycloheptatrienes 1a,26 1b,26 1c,27 1d28 and the tropylium
perchlorates 3a and 3b26 were synthesized as described in the
literature and purified by recrystallization.
Experimental
1. Instrumentation
Absorption spectra were recorded on a Shimadzu UV-2101 PC
spectrophotometer and emission measurements were carried
out with a Perkin-Elmer spectrofluorimeter LS 50B. NMR
spectra were recorded on a Bruker instrument (DPX 300) and
EPR spectra on a ZWG instrument (EPR-300).
4.1 Synthesis of 3,3Ј-bis(4-methoxyphenyl)bi(cyclohepta-
1,3,5-trien-7-yl) (2a). A solution of CrCl2 prepared from 9 g
(36.2 mmol) CrCl3 in 20 ml HCl (5 m) and 10 g Zn dust (153
mmol) was filtered and transferred under argon into a three-
necked flask containing 250 ml of a suspension of 1 g 3a (3.4
mmol) in HCl (2 m). After stirring for one hour under argon, 40
ml tert-butyl methyl ether were added. The mixture was neutral-
ized with NaOH and cooled with ice. Subsequently, the organic
layer was separated and the aqueous phase extracted with tert-
butyl methyl ether. The combined organic extracts were dried
and the solvent was evaporated under reduced pressure. 620 mg
(94%) of the raw mixture of regioisomers of the substituted
1.2 Continuous irradiation. Continuous irradiations were
carried out with acetonitrile solutions in 1 cm quartz cuvettes
with a 500 W high pressure mercury lamp operated in con-
junction with a metal interference filter (Carl Zeiss Jena) or an
Oriel monochromator (model 77200). For some irradiations a
Rayonet photoreactor (lamp type: RPR 3500 or 4190) was
used. Prior to irradiation the solutions were degassed by argon
bubbling. Solutions containing HBF4 (3.2 × 10Ϫ2 M) were
prepared using aqueous HBF4 (50%). The absorbed irradi-
ation dose was determined with the aid of the ferric oxalate
actinometer.
1
bitropyl remained. As concluded from H NMR spectroscopy
at least four isomers were formed. This isomer mixture was used
to identify the products formed during the oxidation of 1a by
photoexcited A4 or A5.
Successive recrystallization from cyclohexane–hexane (1:1)
and from acetonitrile resulted in pale yellow plates of the pure
3,3Ј-isomer of 2a. Yield 80 mg (12%), mp: 191–194 ЊC (Found:
C, 84.90; H, 6.50. Calc. for C28H26O2: C, 85.25; H, 6.64%);
δH (CDCl3) 2.18 (2H, br m, 7-H); 3.85 (6H, s, MeO); 5.43, 5.50
(4H, m, 1-H, 6-H); 6.38 (2H, m, 5-H); 6.44 (2H, 2-H); 6.94
(4H, d, Ar); 7.00 (2H, d, 4-H); 7.47 (4H, d, Ar).
1.3 Flash photolysis. Laser flash photolysis of acetonitrile
solutions was performed in 1 cm quartz cuvettes with the aid
of a ruby laser (Korad Model K1 QS2, flash duration 20 ns)
operated in conjunction with a frequency doubler (λ = 347 nm).
The solutions were deaerated by argon bubbling.
2. Product analysis
HPLC analysis was performed using RP 18 phases and a diode
array detector (Shimadzu SPD-M10A). The mobile phase con-
sisted of a 80:20 mixture of acetonitrile and water that was
buffered with triethylamine–H3PO4 (pH = 2.7). Tropylium salt
yields were determined by analytical HPLC on the basis of a
calibration curve obtained with probes of given concentration.
The latter was obtained by measuring the optical absorption
at 435 nm (ε = 2.25 × 104 MϪ1 cmϪ1). The rate constant of
tropylium salt formation was calculated by the tangent method
using plots of the conversion vs. the time of irradiation. The
H2O2 concentration of the (neutralized) reaction solution was
determined via the oxidation of iodide to iodine using
Acknowledgements
The authors appreciate the partial financial support of
this work obtained from Fonds der Deutschen Chemischen
Industrie and gratefully acknowledge the technical assistance
of Mrs K. Buck.
References
1 D. Jacobi, W. Abraham, U. Pischel, L. Grubert and W. Schnabel,
J. Chem. Soc., Perkin Trans. 2, 1999, 1241.
2 T. Linker and M. Schmittel, Radikale und Radikalionen in der
Organischen Synthese, Wiley-VCH, Weinheim, 1998.
3 V. Wendel and W. Abraham, Tetrahedron Lett., 1997, 38, 1177.
4 D. Jacobi and W. Abraham, Tetrahedron Lett., 1996, 37, 7493.
5 D. Rehm and A. Weller, Israel J. Chem., 1970, 8, 259.
6 M. E. R. Marcondes, V. G. Toscano and R. G. Weiss, J. Am. Chem.
Soc., 1975, 97, 4485.
Ϫ
(C4H9)4NI.14 The concentration of I3 was determined by
measuring the optical density at 365 nm with the aid of a
calibration curve.
3. Determination of reduction and oxidation potentials
Rapid scan cyclic voltammetry (1 to 20 V sϪ1) was performed
7 P. Ramamurthy, F. Morlet-Savary and J. P. Fouassier, J. Chem. Soc.,
Faraday Trans., 1993, 89, 465.
J. Chem. Soc., Perkin Trans. 2, 1999, 1695–1702
1701

View Article Online
8 S. L. Murov, Handbook of Photochemistry, Marcel Dekker,
New York, 1973.
9 T. Shida, Phys. Sci. Data, 1988, 34, 246.
21 D. Jacobi, PhD Thesis, Humboldt-University, 1998.
22 M. L. Olmstead, R. G. Hamilton and R. S. Nicholson, Anal. Chem.,
1969, 41, 260.
10 I. R. Gould, D. Ege, J. E. Moser and S. Farid, J. Am. Chem. Soc.,
1990, 112, 4290.
11 M. A. Miranda and H. Garcia, Chem. Rev., 1994, 94, 1063.
12 F. Pragst, Electrochim. Acta, 1976, 21, 497.
13 P. Valat, S. Tripathi, V. Wintgens and J. Kossanyi, Nouv. J. Chem.,
1990, 14, 825.
23 B. Dreher, W. Abraham and F. Pragst, J. Prakt. Chem., 1983, 325,
104.
24 U. C. Yoon, P. S. Mariano, S. L. Quillen, R. Swanson, J. L.
Stavinoha and E. Bay, J. Am. Chem. Soc., 1983, 105, 1204.
25 S. Fukuzumi, S. Koumitsu, K. Hironaka and T. Tanaka, J. Am.
Chem. Soc., 1987, 109, 305.
14 N. W. Koper, S. A. Jonker, J. W. Verhoeven and C. van Dijk, Recl.
Trav. Chim. Pays-Bas, 1985, 104, 296.
15 S. Fukuzumi, M. Ishikawa and T. Tanaka, J. Chem. Soc., Perkin
Trans. 2, 1989, 1037.
26 C. Jutz and F. Voithenleitner, Chem. Ber., 1964, 97, 29.
27 J. J. Looker, J. Org. Chem., 1965, 30, 4180.
28 W. Abraham, K. Buck, Ch. Csongar, E. Henke and D. Kreysig,
J. Prakt. Chem., 1979, 321, 117.
16 D. Behar, G. Czapski, J. Rabani, L. M. Dorfman and H. A.
Schwarz, J. Phys. Chem., 1970, 74, 3209.
17 S. Fukuzumi, M. Chiba, M. Ishikawa, K. Ishikama and T. Tanaka,
J. Chem. Soc., Perkin Trans. 2, 1989, 1417.
18 D. T. Sawyer, T. S. Calderwood, K. Yamaguchi and C. T. Angelis,
Inorg. Chem., 1983, 22, 2577.
29 G. J. Kavarnos and N. J. Turro, Chem. Rev., 1986, 86, 401.
30 P. S. Marirano, J. L. Stavinoha and E. Bay, Tetrahedron, 1981, 37,
3385.
31 Y. Kuriyama, T. Arai, H. Skuragi and T. Tokumaru, Chem. Lett.,
1988, 1193.
19 D. T. Sawyer and J. S. Valentine, Acc. Chem. Res., 1981, 14, 393.
20 S. Fukuzumi, Adv. Electron Transfer Chem., 1992, 2, 118.
Paper 9/02662B
1702
J. Chem. Soc., Perkin Trans. 2, 1999, 1695–1702