Photocyclization mechanism of halopyridinium salt tethered to arene: Flash photolysis observation of a pyridinium σ, cyclohexadienyl radicals, and a dihalide radical anion in aqueous solution

Article abstract of DOI:10.1021/ja970425u

The photocyclization reactions of 2-halopyridinium salt tethered to arene have been studied. The photocyclization of 2-halophenethylpyridinium salts produced a six-membered heterocyclic product, pyrido[2,1-a]-3H,4H-isoquinolinium salt. The 2-halopyridiniu

Full text of DOI:10.1021/ja970425u

J. Am. Chem. Soc. 1997, 119, 10677-10683
10677
Photocyclization Mechanism of Halopyridinium Salt Tethered
to Arene: Flash Photolysis Observation of a Pyridinium σ,
Cyclohexadienyl Radicals, and a Dihalide Radical Anion in
Aqueous Solution
Yong-Tae Park,*,† Nam Woong Song,‡ Chul-Gyun Hwang,†
Kwang-Wook Kim,† and Dongho Kim*,‡
Contribution from the Department of Chemistry, Kyungpook National UniVersity,
Taegu 702-701, Korea, and Spectroscopy Laboratory, Korea Research Institute of Standards and
Science, Taejeon 305-600, Korea
ReceiVed February 7, 1997X
Abstract: The photocyclization reactions of 2-halopyridinium salt tethered to arene have been studied. The
photocyclization of 2-halophenethylpyridinium salts produced a six-membered heterocyclic product, pyrido[2,1-a]-
3H,4H-isoquinolinium salt. The 2-halopyridinium salts tethered to phenyl group with methylene linkage (1 and 2a)
are more reactive than the 2-halopyridinium salts with an ethylene group (3 and 4a). As expected, the
2-bromopyridinium salt linked to the phenyl group with ethylene linkage (4a) is more reactive than the
2-chloropyridinium salt with the ethylene group (3), which is different from the 2-halopyridinium salt with the
methylene linkage. The transient intermediates such as pyridinium salt σ, 2,3-dihydrocyclohexadienyl radicals, and
a dibromide radical anion are detected with laser flash photolysis facility. The maximum absorption wavelengths of
the pyridinium salt σ, 2,3-dihydrocyclohexadienyl radicals, and dibromide radical anion are 250, 310, and 360 nm,
respectively. The lifetimes of the pyridinium salt σ, 2,3-dihydrocyclohexadienyl radicals, and dibromide radical
anion are 0.93 ms, 0.84 ms, and 25 µs, respectively. Thus, a photohomolytic radical mechanism is proposed for the
cyclization reaction. The excited singlet state is mainly involved in the photocyclization of 2-halopyridinium salts
tethered to phenyl group with ethylene linkage (3 and 4a), while both the singlet and triplet states are involved in
the photocyclization of the 2-halopyridinium salts with methylene linkage (1 and 2a).
Introduction
to heteroarene.6,7 Therefore, it is important to know the
mechanistic aspect of the photocyclization.
The photoreactions of haloarenes are diverse depending upon
reaction conditions. Generally three photochemical reactions
are known for the haloarene: in aromatic solvent, arylation
reaction is major, while in solvents able to donate hydrogen
atoms, reduction reaction is dominant; under certain conditions,
a nucleophilic substitution reaction can occur. The work in the
field of haloarene photochemistry has been reviewed extens-
ively.1-3 The photochemical reaction of haloarene is consider-
ably important, in that the reaction is not only valuable in
organic synthesis but also related to the degradation of the
environmental pollutents (halogenated arene). Particularly, the
photocyclization of haloarene tethered to arene or heteroarene
is simple and valuable for homo- and heterocyclic ring forma-
tion.4,5 Five- and six-membered heterocyclic ring systems can
be readily formed by the photocyclization of haloarene linked
Only a few mechanistic studies on the intermolecular pho-
toarylation and intramolecular photocyclization (or photoary-
lation) have been described,4a,8-10 although several studies on
the mechanistic aspect and transient species involved in the
photoreduction of haloarene have been reported.11,12 Perdersen
and Lohse8 reported that 1,4-dibromobenzene underwent pho-
toarylation in benzene to give 4-bromobiphenyl and p-terphenyl
Via transient cyclohexadienyl radical. The transient radical was
characterized as follows: it decayed in the second order process
with a half-life 0.7 ms. Grimshaw and de Silva reported a
photohomolysis of the carbon-halogen bond assisted by radical
complexation on the photocyclization of 1,3-diphenyl-5-(2′-
halogenophenyl)pyrazoles4a and 2-chlorobenzanilide.4b Grim-
shaw and Trecha-Grimshaw9 detected a transient intermediate,
a cyclohexdienyl radical, in the photocyclization of 1,3-diphenyl-
5-(2′-halogenophenyl)pyrazoles, supporting the mechanism of
homolytic carbon-halogen bond fission assisted by radical
complexation. Obviously, a direct observation of a transient
intermediate such as a cyclohexadienyl radical involved in the
photoreaction is strong evidence for supporting the proposed
mechanism.
† Kyungpook National University.
‡ Korea Research Institute of Standards and Science.
X Abstract published in AdVance ACS Abstracts, October 15, 1997.
(1) Bunce, N. J. In CRC Handbook of organic photochemistry and
photobiology; Horspool, W. H., Song, P.-S. Eds.; CRC: New York, 1995;
1181-1192.
(2) Kessar, S. V.; Mankotia, A. K. S. In CRC Handbook of organic
photochemistry and photobiology; Horspool, W. M., Song, P.-S. Eds.;
CRC: New York, 1995; pp 1218-1228.
(3) Grimshaw, J. J. Chem. Soc., Chem. ReV. 1981, 10, 181-203.
(4) (a) Grimshaw, J.; de Silva, P. A. J. Chem. Soc. 1979, 193-194. (b)
Grimshaw, J.; de Silva, P. A. J. Chem. Soc., Chem. Commun. 1980, 302-
303. (c) Grimshaw, J.; de Silva, P. A. Can. J. Chem. 1980, 58, 1880-
1888.
(6) (a) Forzard, A.; Brasher, C. K. J. Org. Chem. 1967, 32, 2966-2969.
(b) Portlock, D. E.; Kane, M. K.; Bristol, J. A.; Lyle, R. E. J. Org. Chem.
1973, 38, 2351-2357.
(7) (a) Park, Y.-T.; Joo, C.-H.; Choi, C.-D.; Park, K.-S. J. Heterocycl.
Chem. 1991, 28, 1083-1089. (b) Park, Y.-T.; Kim, K.-S.; Lee, I.-H.; Park,
Y.-C. Bull. Korean Chem. Soc. 1993, 14, 152-153. (c) Byun, Y.-S.; Jung,
C.-H.; Park, Y.-T. J. Heterocycl. Chem. 1995, 32, 1835-1837.
(8) Pedersen, C.; Lohse, C. Acta Chim. Scand. B 1979, 33, 649-652.
(9) Grimshaw, J.; Trocha-Grimshaw, J. J. Chem. Soc., Chem. Commun.
1984, 744-745.
(5) (a) Park, Y.-T.; Lee, I.-H.; Kim, Y.-H. J. Heterocycl. Chem. 1994,
31, 1625-1629. (b) Park, Y.-T.; Kim, Y.-H.; Hwang, C.-G. Bull. Korean
Chem. Soc. 1996, 17, 506-510.
S0002-7863(97)00425-3 CCC: $14.00 © 1997 American Chemical Society

10678 J. Am. Chem. Soc., Vol. 119, No. 44, 1997
Park et al.
Chart 1
Experimental Section
Material and General. Benzylbromide, phenethylbromide, pyri-
dine, 2-chloropyridine, 2-bromopyridine, and acetonitrile (Aldrich) were
used as received. Water distilled triply was used.
UV/visible spectra were recorded on a Shimadzu UV 265 spectro-
photometer. NMR spectra were recorded on a Varian Unity Plus
spectrometer operating at 300 MHz for proton. The 1H NMR spectra
are referenced with respect to water proton of solvent D2O. Infrared
spectra were recorded on an FT-IR (Mattson Galax Series 7020)
spectrophotometer on KBr crystal).
Synthesis of the Pyridinium Salts. N-Benzyl-2-chloropyridinium
and N-benzyl-2-bromopyridinium bromides are prepared as described
in ref 7a.
N-Benzyl-2-bromopyridinium perchlorate, 2b, was prepared by an
ion exchange reaction of 2a with perchloric acid: When excess
perchloric acid (60% perchloric acid 0.5 mL, 4.5 mmol) was added
slowly into ethanolic solution of 2a (500 mg, 1.5 mmol), a white crystal
appeared: the crystal was filtered and recrystallized from ethanol, yield
50%, mp 97-98 °C.
2-Chloro-N-phenethylpyridinium bromide (3) was prepared by
reaction of phenethyl bromide with 2-chloropyridine in acetonitrile,
yield 30%. Recrystallization (acetonitrile) gave the pyridinium salt,
3, as a white crystal (mp 179-180 °C).
2-Bromo-N-phenethylpyridinium bromide (4a) was prepared by
reaction of phenethyl bromide with 2-bromopyridine in acetonitrile,
yield 14%. Recrystallization (acetonitrile) gave the pyridinium salt 4a
as a white crystal (mp 150 °C).
2-Bromo-N-phenethylpyridinium perchlorate (4b) was prepared by
ion exchange reaction of 4a with perchloric acid: The method was the
same as 2b except the dissolved solvent methanol used. The pyridinium
salt 4b was recrystallized from methanol, yield 80%, mp 184-185 °C.
N-Phenethylpyridinium bromide (5) was prepared by reaction of
phenethyl bromide with pyridine in acetonitrile, yield 60%. Recrys-
tallization (acetonitrile) gave the pyridinium salt 5 as a white crystal
(mp 125-126 °C).
Recently, we observed the transient intermediates such as a
phenyl σ, 2,3-dihydrocyclohexadienyl radicals (a conjugated
radical), and a dihalide radical anion13 in the photolysis of the
aqueous N-(2-halobenzyl)pyridinium salts (1 and 2a in Chart
1).10 Thus, it was proposed that the photoreaction proceeds to
cyclized product pyrido[2,1-a]isoindolium salt Via a photo-
homolytic cleavage of the pyridinium ring-halogen bond,
followed by arylation of the phenyl σ radical to the neighboring
phenyl group and hydrogen atom ejection of the conjugated
radical.
2-Bromo-N-ethylpyridinium bromide (6) was prepared by reaction
of ethyl bromide with 2-bromopyridine, yield 50%. Recrystallization
(acetonitrile) gave the pyridium salt 6 as a white crystal (mp 194-195
°C).
An electron transfer between haloarene and arene or halide
ion followed by halide ion expelling, arylation of the resulting
phenyl radical with the neighboring phenyl ring cation, and
proton ejection may be assumed, since photochemically induced
electron transfer between haloarene and amine, diene, or
haloarene itself for the photoreduction is well documented.14-16
However, the electron transfer mechanism has not yet been
proposed for the photocyclization of haloarene linked to arene.
In other words, the photocyclization mechanism of haloarene
has not been clarified, albeit limited studies on the photo-
cyclization of haloarene tethered to arene support the photo-
homolytic mechanism.
In this paper, our objective has been to clarify the detailed
mechanism for the photocyclization of halopyridinium salt
tethered to arene. To this end, we have studied not only for
the reactivities of the steady-state photoreaction of 2-halopyri-
dinium salt tethered to phenyl ring (1, 2a, 3, and 4a, Chart 1)
but also for the detection and characterization of the transient
species involved in the laser flash photolysis of the halopyri-
dinium salts (3 and 4a).
Steady-State Photoreaction. The aqueous solution of the halo-
pyridinium salts 3, 4a, or 4b (2.0 × 10-4 M) in UV cuvette (path length
1 cm, 3 mL) was irradiated with a monochromatic light (278 nm) and
the UV/visible absorption change was measured for every 10 min,
Figure 1. The new peaks at about 313 and 260 nm refer to a cyclized
product, pyrido[2,1-a]isoquinolinium bromide (vide infra).
An aqueous solution of 2-chloro-N-phenethylpyridinium bromide (3,
500 mg, H2O 300 mL) was introduced into an immersion quartz
photovessel and then irradiated with a Hg-lamp for 7 h. The reaction
was followed by UV absorption change at about 313 nm. The reaction
mixture was dried by evaporation, and the residue was dissolved in 10
mL methanol. By adding 50 mL of acetone into the above methanolic
solution, the appearing flocculent material was filtered out. After the
solvent was dried, when 5 mL of ethanol was introduced, followed by
addition of 0.1 mL of perchloric acid, pyrido[2,1-a]-3H,4H-isoquino-
linium perchlorate (white crystal, 8) appeared, with a yield 30%. In
fact, the yield of the cyclized product 7 before the workup with
perchloric acid was 35% in HPLC. The retention times of the product
and reactant of reaction mixture were 4.8 and 7.1, respectively (Rainin
Instrument; C-18 column, 4.6 × 250 mm; mobile phase, methanol/
water (3/1); flow rate, 0.5 mL/min). Recrystallization (ethanol) gave
the pyridinium salt 8 as a white crystal (mp 142-143 °C).
The photochemical behavior of 4a or 4b is the same as that of 3:
The identical product is obtained in the photoreaction of 4a or 4b.
To measure the quantum yield of the photocyclization of the
pyridinium salts (1, 2a, 2b, 3, 4a, or 4b), the following procedure was
performed. The intensity of the monochromatic light was measured
by using standard ferrioxalate actinometry; the intensity at 278 nm was
1.02 × 10-6 ein/s. The aqueous solution of the pyridinium salt in UV
cuvette (1 cm path, 3 mL) under argon or oxygen was irradiated with
the above monochromatic light. The absorption change at 312 nm was
monitored. The molar absorptivities of the photocyclized products,
(10) Park, Y.-T.; Song, N. W.; Kim, Y.-H.; Hwang, C.-G.; Kim, S. K.;
Kim, D. J. Am. Chem. Soc. 1996, 118, 11399-11405.
(11) (a) Previtali, C. M.; Ebbesen, T. W. J. Photochem. 1984, 27, 9-15.
(b) Alfassi, Z. B.; Previtali, C. M. J. Photochem. 1985, 30, 127-132. (c)
Previtali, C. M.; Ebbensen, T. W. J. Photochem. 1985, 30, 259-267.
(12) (a) Smothers, W. K.; Schanze, K. S.; Saltiel, J. J. Am. Chem. Soc.
1979, 101, 1895-1896. (b) Bunce, N. J.; Bergsma, J. P.; Bergsma, M. D.;
De Graaf, W.; Kumar, Y.; Ravanal, L. J. Org. Chem. 1980, 45, 3708-
3713.
(13) The name of dihalide radical anion is used for halogen anion radical.
For example, dibromide radical anion is used for bromine anion radical.
(14) Chesta, C. A.; Cosa, J. J.; Previtali, C, M. J. Photochem. 1986, 32,
203-215.
(15) Chesta, C. A.; Cosa, J. J.; Previtali, C. M. J. Photochem. Photobiol.
A. Chem. 1988, 45, 9-15.
(16) Bunce, N. J. J. Chem. Soc., Perkin Trans. 1975, 1607-1610.

Halopyridinium Salt Tethered to Arene
J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10679
Scheme 1
Figure 1. The UV absorption change of an aqueous solution of 4a
(2.0 × 10-4 M) under argon Vs irradiation time (each interval 10 min).
Scheme 2
pyrido[2,1-a]isoindolium bromide and pyrido[2,1-a]isoquinolinium
perchlorate, are 1.0 × 104 and 1.3 × 104 M-1 cm-1, respectively, and
were used for the determination of the concentration change corre-
sponding to the absorption change.
For the Stern-Volmer plot, the following procedure was performed:
the aqueous solutions of the pyridinium salts (3 or 4, 2.0 × 10-4 M)
with and without methylviologen (0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0 × 10-5
M) under argon were irradiated with monochromatic light (278 nm,
450W Xe-lamp, Shimadzu monochromator). The absorption change
at 313 nm was monitored to measure the quantum yield of the cyclized
product.
Laser Flash Photolysis. A detailed description of the experimental
setup can be found elsewhere.10 The fourth harmonic (266 nm) output
from a Q-switched Nd:YAG laser (Spectron SL803G) was used as an
excitation source. The time duration of the excitation pulse was ca. 5
ns, and the pulse energy was typically 55 mJ. A cw Xe-arc lamp (Atago
Bussan Co. XC-150) was used as a probe light source for transient
absorption measurement. The spectral resolution was obtained by using
320 nm monochromator (Jobin-Yvon HR320) after the probe light
passed through the sample solution. A boxcar signal averager (Stanford
Research Sys. SR250) was used in recording the transient signals. The
temporal profile of the transient absorption signal was monitored by a
500 MHz digital storage oscilloscope (Hewlett Packard HP54503A).
Sample solutions were prepared by dissolving the reagents in triply
distilled water, and the concentration of the solution was adjusted to
be 0.8-1.5 in absorbance at 266 nm. The sample solution was
circulated from a Teflon bottle (2.5 L in volume) to the fluorometer
quartz cuvette of 10 mm in path length (flow type, Helma QS1.0) to
reduce the effect by the accumulation of the reaction product and
dilution of the reactant in the photolysis cell.
Table 1. The Quantum Yield of the Photocyclized Product of the
Aqueous Pyridinium Salts
Φ
Φ
reactant under Ar under O2 reactant under Ar under O2
1
2a
2b
0.22
0.20
0.21
0.07
0.08
0.09
3
4a
4b
0.03
0.06
0.06
0.02
0.05
0.04
1). The yield of the cyclized product 7 before the workup with
perchloric acid was 35% in HPLC. The identical product was
also obtained from the bromopyridinium salt 4a. For six-
membered heterocyclic ring formation, this reaction is a simple
and convenient method.
In order to study the effect of the counter ion of the
pyridinium salts, 2-bromo-N-phenethylpyridinium perchlorate
(2b) and 2-bromo-N-phenethylpyridinium perchlorate (4b) hav-
ing nonoxidizable counter anion were prepared and studied in
steady-state photoreaction and laser flash photolysis. The
identical photocyclized product (8) was obtained from the
steady-state photoreaction of 4b.
The quantum yields of the photocyclized product of the
pyridinium salts (1, 2a, 2b, 3, 4a, and 4b) were measured and
are shown in Table 1. The quantum yield of 1 or 2a, in which
the halopyridinium ring is connected to phenyl group by
methylene group, is 3-10 times as great as that of 3 or 4a, in
which the halopyridinium ring is linked to phenyl by ethylene
group, under argon. This result is understandable in consider-
ation of the conformation of the pyridinium salts: the pyridinium
salts with ethylene connector (3 and 4a) mainly have a staggered
conformation because of the different propensity of both the
Results and Discussion
Steady-State Photoreaction. When the aqueous solution of
2-bromo-N-phenethylpyridinium bromide (4a, 2.0 × 10-4 M,
3 mL) in UV-cuvette (path length 1 cm) was irradiated with
monochromatic light (278 nm, 450W Xe-lamp) under argon,
the UV absorption change Vs irradiation time (interval 10 min)
was measured and is shown in Figure 1. The absorption band
at about 278 nm corresponding to the absorption of the starting
pyridinium salt decreased meanwhile those at about 313 and
260 nm corresponding to a product increased. For chloro-
pyridinium salt 3, a parallel spectrum change was obtained (not
shown). The 313-nm absorbing product from 3 or 4a is readily
identified as a cyclized product because the absorbing wave-
length is similar to the photocyclized product of halopyridinium
salt 1 or 2a. The clear isosbestic points are seen at about 292
and 268 nm, indicating that only one product is formed.
When an aqueous solution of 2-chloro-N-phenethylpyridinium
bromide (3, 0.5 g, H2O 300 mL, quartz photovessel) was
irradiated with a high pressure mercury lamp for 7 h and worked
up with perchloric acid, a cyclized product, pyrido[2,1-a]3H,4H-
isoquinolinium salt (8), was obtained in 30% yield (Scheme

10680 J. Am. Chem. Soc., Vol. 119, No. 44, 1997
Park et al.
pyridinium and phenyl rings toward the solvent, water; the
pyridinium salts with methylene connector (1 and 2a) have only
a close conformation of both the pyridinium and phenyl rings
whatever the properties of the rings toward the water are. Thus,
it is indicative that the interacting pyridinium ring with phenyl
ring should be brought into a sufficiently close proximity for
cyclization to take place effectively.
Quantum yield of N-benzyl-2-chloropyridinium salt 1 is
higher than that of N-benzyl-2-bromopyridinium salt, 2a,
although the bond energy of phenyl-chlorine (96 kcal/mol) is
higher than that of phenyl-bromine (81 kcal/mol).17 This result
is in line with that of 2-halobenzanilide4b and 2-phenethyl-1-
halobenzene.5b However, for halopyridinium salts 3 and 4a, in
which the halopyridinium ring is connected with the ethylene
group, this is not the case; chloropyridinium salt 3 is less reactive
than bromopyridinium salt 4a. Probably, some assistance of
chlorine moiety of the excited pyridinium ring in the photocy-
clization of the pyridinium salt with methylene connector 1 may
exist for phenyl-chlorine bond cleavage and/or for bringing
into close proximity of the pyridinium and phenyl moieties. The
argument is not applicable to the pyridinium salt with longer-
chain connector, ethylene group, probably because of difficulty
of accessible conformation of close proximity of the pyridinium
and phenyl rings.
Figure 2. Transient absorption spectra obtained upon 266-nm excitation
of 3 in aqueous solution under argon (s) 3 µs after laser pulse and
(‚‚‚) after 300 µs.
The quantum yield of 1 or 2a is reduced to one-third in the
presence of oxygen, while that of 3 or 4a is not reduced
effectively. This means that at least two-thirds of the photo-
cyclization of 1 or 2a originate from the triplet state and the
photocyclization of 3 or 4a occur mainly from the singlet state,
albeit ineffective.
Figure 3. Transient absorption spectra obtained upon 266-nm excitation
of 4a and 5 (dashed line) in aqueous solution under argon (s) 3 µs
after laser pulse and (‚‚‚) after 300 µs.
The quantum yields of 2b and 4b having a nonoxidizable
counter anion under argon or oxygen are the same as 2a and
4a, respectively. These results imply that the counter anion
has no effect on the reaction and in turn the electron transfer to
excited pyridinium ring from anion can be excluded.
Stern-Volmer plot of photocyclization of 4a with triplet
quencher methylviologen has been performed. The kQτ value
is 3.8 × 103 M-1 which is one-tenth of 1 or 2a,10 indicating
that the triplet state is less significantly involved, compared with
that of 1 or 2a. This result is consistent with that of the forgoing
quantum yield measurement. If kQ is assumed as the diffusion
control rate (kQ ) 6.4 × 109 M-1 s-1 in water),18 the lifetime
of the triplet state of 4a is 0.6 µs (lower limit). For the
chloropyridinium salt, 3, the kQτ and lifetime are 5.9 × 103
M-1 and 0.9 µs (lower limit), respectively.
Laser Flash Photolysis. Transient absorption spectra were
obtained from the laser flash photolysis of 2-chloro-N-phen-
ethylpyridinium bromide (3) in argon-saturated water as shown
in Figure 2. There are three major transient absorption peaks
at about 250, 310, and 380 nm at 3 µs delay after laser flash.
At 300 µs delay after the laser flash, the absorption bands at
about 250 and 310 nm remain, and that around 380 nm
diminishes. A similar spectra were obtained for 2-bromo-N-
phenethylpyridinium salt 4a, as shown in Figure 3. At 3 µs
delay, three bands around 250, 310, and 350-420 nm appear.
At 300 µs delay, both peaks around 250 and 310 nm remain,
while the band at around 350-420 weakens significantly. Both
spectra from 3 and 4a are similar to those of N-benzyl-2-
halopyridinium salts (1 and 2a)10 except that the broad band in
the region of 350-420 nm of 3 and 4a is weaker and broader
than that of 1 and 2a.
In order to identify which bonds of the pyridinium salts are
raptured to give the transient species, the transient absorption
spectra of 2-bromo-N-phenethylpyridinium salt (4a) and N-
phenethylpyridinium salt (5, Chart 1) were obtained under argon
and are compared in Figure 3. For pyridinium salt 5 without
halogen substituent, no transient species was detected under our
experimental condition. These results imply that the transient
species of the 2-halopyridinium salts 3 or 4a originate from
the aryl-halogen bond cleavage, not from the benzyl-hydrogen
bond cleavage of the pyridinium salt (side chain of the
pyridinium salt).
Faria and Steenken19 reported that a benzyl radical and
solvated electron (measured at 650 nm, λmax 720 nm) were
generated by the laser photolysis of triphenylmethane and
diphenylmethane in several solvents. The reaction was de-
scribed in terms of a biphotonic process based on the quadratic
dependence of the benzyl-type radical yield on the laser light
intensity. The structure of the pyridinium system here is quite
similar to that of Faria and Steenken’s diphenylmethane.
Therefore, the possibility of the photoionization of the side chain
of the pyridinium salts 3 or 4a was examined by using DSO
(500 MHz). For the pyridinium salts 3, 4a, or 5, no sign of the
presence of hydrated electron has been detected in the region
of 400-750 nm. Consequently, based upon this and the
forgoing results, we suggest that photoionization of the benzyl-
hydrogen bond of (or side chain of) the pyridinium salts does
not occur.
o-Pyridinium σ Radical. The temporal profile of the
transient species at about 250 nm from the photolysis of 4a is
shown in Figure 4. The species decays in a first order process,
and its lifetime and rise times are 0.93 and 0.40 ms, respectively,
Table 2. The absorption wavelength is the same as that from
(17) Streitwieser, A.; Healthcook, C. H.; Kosower, E. M. Introduction
to Organic Chemistry, 4th ed.; Macmillan: New York, 1992; Appendix II.
(18) Calvert, J. G.; Pitts, J. N. Photochemistry; John Wiley and Sons:
New York, 1966; pp 625-627.
(19) (a) Faria, J. L.; Steenken, S. J. Am. Chem. Soc. 1990, 112, 1277-
1279. (b) Faria, J. L.; Steenken, S. J. Phys. Chem. 1993, 97, 1924-1930.

Halopyridinium Salt Tethered to Arene
J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10681
Table 2. The λmax and Lifetimes of the Transient Intermediates
Figure 5. Temporal profile of 310-nm transient from laser flash
photolysis of 4a under argon. Inset: same as above at µs time scale
under argon (a) and oxygen (b).
Figure 6. Temporal profile of 360-nm transient (s) from laser flash
photolysis of 4a (curve a) and 4b (curve b) under argon. Inset: Effect
of addition of ferrous ion on 360 nm transient formed from 4a under
Fe2+
argon (curve a
8 curve c).
Figure 4. Temporal profile of 250-nm transient from laser flash
photolysis of 2-bromo-N-phenethylpyridinium bromide (4a) under
argon. Inset: same as above at µs time scale.
dibromobenzene.8 Hence, the 310-nm transient is identified as
a conjugated radical, 2,3-dihydrocyclohexadienyl radical, Table
2.
1 or 2a, and the lifetime is a little longer than that from 1 or 2a
(0.7 ms). The 250-nm transient diminishes by the presence of
N-tert-butylphenylnitrone, radical scavenger. Thus, the 250-
nm transient is assigned unequivocally to o-pyridinium σ radical
formed from the photohomolytic cleavage of the aryl-bromine
bond. It is understandable that the lifetime of pyridinium σ
radical from 4a is longer than that from 1 or 2a, since the
photocyclization quantum yield of 4a is much lower than that
of 1 or 2a.
It is unusual to observe another short rise time (10 µs) for
the formation of the 250-nm transient (inset in Figure 4). One
possible explanation is that the short one is the rise time for
formation of o-pyridinium σ radical from the singlet state and
the long one is that from the triplet state. Another possible
explanation is that the short one is the rise time for formation
of o-pyridinium σ radical from the excited pyridinium salt and
the long one (400 µs) is that from the conjugated radical, 2,3-
dihydro-cyclohexadienyl radical (vide infra). We prefer the
latter explanation over the first one because the lower-limit
triplet lifetime is in the range of 0.6 µs.
2,3-Dihydrocyclohexadienyl Radical. The temporal profile
of a transient species at about 310 nm from 4a is shown in
Figure 5. The transient decays in a first order process, and its
lifetime and rise time are 0.84 and 0.33 ms, respectively. The
absorption wavelength is the same as that of a 2,3-dihydro-
cyclohexadienyl radical from 1 or 2a reported10 and that of a
cyclohexadienyl radical (λmax ) 310 nm) produced upon
γ-irradiation of benzene in methanol at 77 K.20 This lifetime
of 310 nm transient from 4a is similar to the corresponding
cyclohexadienyl radical from 1 or 210 and to that from, 1,4-
To confirm the assignment, the transient species from a half-
moiety model of the pyridinium salt 6, which cannot form 2,3-
dihydrocyclohexadienyl radical, were examined. Laser flash
photolysis of the pyridinium salt 6 in water under argon
produced 250- and 360-nm transients, but a 310-nm transient
at 300 µs delay after laser flash, supporting the above assignment
(vide infra).
There are also two different rise times for the 2,3-dihydro-
cyclohexadienyl radical: one is 10 µs (inset in Figure 5) and
the other is 327 µs under argon. The possible explanation is
that the short one is the rise time for the formation of the
conjugated radical from the excited pyridinium salt and the long
one is that from the pyridinium salt σ radical. The rise time of
the conjugated radical is not the same as the decay time of the
pyridinium σ radical. The reverse is also true. They are
understandable since either of the transients does not come from
only one source. The presence of oxygen diminished a little
amount of the transient (10%) at 300 µs delay, indicating the
fact that about 10% of the conjugated radical originates from
the triplet state. The 310-nm transient of the inset of Figure 5
is also reduced to only 11% in the presence of oxygen. In other
words, about 90% of the radical is generated from the singlet
state. This result is consistent with the lack of oxygen effect
on the quantum yield of the steady-state photoreaction.
Dihalide Radical Anion. The region of 350-420 nm from
4a in Figure 3 exhibits a complicated spectral feature. The
temporal profile of 360-nm transient shown in Figure 6 (curve
a) showed two component decays for the peaks: the short-lived
transient is 25 µs and the long-lived one is 780 µs. It has been
•-
reported that the absorption of dichloride radical anion (Cl2
)

10682 J. Am. Chem. Soc., Vol. 119, No. 44, 1997
Park et al.
Figure 7. Temporal profile on a transient at 400 nm formed from 4a
under argon.
Figure 8. A transient absorption spectra acquired from laser flash
photolysis of 2-bromo-N-ethylpyridinium bromide (6) under argon.
Solid line, at 3 µs delay and dotted line, at 300 µs delay after laser
flash.
was suppressed by addition of potassium hexacyanoferrate(II)
in the pulse radiolysis of aqueous sodium chloride.21,22 The
presence of ferrous ion diminishes the intensity of the short-
lived transient at 360 nm (Inset in Figure 6). Therefore, the
short-lived transient at 360 nm is dibromide radical anion, Table
2. The lifetime is somehow a little longer than that reported.10
The long-lived component (τ ) 780 µs) is not identified yet.
In order to clarify the role of counter anion, the property of
a transient at 360 nm was examined with 4b or 2b having a
nonoxidizable anion. The temporal profile of 360-nm transient
from the laser flash photolysis of 4b is shown in curve b in
Figure 6. Surprisingly, the transient absorption at 360 nm
(Br2•-) from 4b is seen, even though it is very weak compared
with that of 4a (curve a). This is probably formed from the
reaction of bromine radical produced in the cleavage of C-Br
bond with HBr produced as product. This explanation is
supported by the following observation: At the initial stage of
illumination of 2b (within 2 min) the transient at 360 nm was
hardly seen, while at the latter stage of illumination (within 10
min) the transient was obviously seen (not shown). These
observations support again the assignment of the short compo-
nent of the transient at 360 nm to Br•- in the reaction of 4a.
The 250-nm transient decays in a first order process, and the
lifetime and rise time are 1.4 ms and 364 µs, respectively, Table
2. The lifetime and rise time of 290-nm transient are the same
as those of 250-nm transient. Thus the 290-nm species belong
to a part of 250-nm transient. As expected, the 250-nm transient
is unambiguously recognized as pyridinium σ radical, because
of the above properties and diminishing effect of the transient
spectra by the addition of radical scavenger, N-tert-butyl-
phenylnitrone.
The broad band around 305 nm has a lifetime of 59 µs. The
transient diminishes in the presence of oxygen, while the
bleaching at 277 nm is less. The most interesting observation
is that the 305-nm transient disappeared at 300 µs delay, while
the bleaching at 277 nm weakened, indicating that the 305-nm
species is due to the excited state of 6. As expected, that 305-
nm transient from 6 is obviously different from the 310-nm
transient in the photolysis 3 or 4a. Previtali and Ebbesen11c
reported that the triplet-triplet absorption of chlorobenzene
appeared at 305 nm. The 305-nm transient from 6 is probably
the triplet-triplet absorption of the pyridinium salt 6. In other
words, the assignments of transients from 3 or 4a are supported
by experiments showing that the flash photolysis of 2-bromo-
N-ethylpyridinium salt 6 produced a phenyl σ radical and
bromine anion radical but not a 2,3-dihydrocyclohexadienyl
radical.
Mechanism. A mechanism of an electron transfer to the
pyridinium ring from the counter anion or phenyl ring is
excluded based on the following observations: (1) photocyclized
product is only observed in the steady-state photoreaction; (2)
no photoreduced product was observed; (3) quantum yield of
2b or 4b having nonoxidizable anion is the same as 2a and 4a,
respectively. Since the transients such as pyridinium salt σ,
2,3-dihydrocyclohexadienyl radicals, and dihalide radical anion
are detected in laser flash photolysis, a photohomolytic radical
mechanism for the photocyclization of 3 or 4a is proposed as
in the photoreaction of N-benzyl-2-halopyridinium salt,10 Scheme
2. The excited singlet state populated by the light absorption
of the pyridinium salt ground state undergoes either homolytic
cleavage of an aryl-hydrogen bond to give a pyridinium salt σ
radical or an intersystem crossing to give the triplet state. Both
the pyridinium salt σ radical and triplet state proceed to the
conjugated radical, 2,3-dihydrocyclohexadienyl radical, by the
cyclization reaction. The conjugated radical in turn proceeds
ultimately to the cyclized product, pyrido[2,1-a]-3,4-dihydroiso-
quinolinium salt, by ejecting hydrogen atom.
The temporal profile of the transient species at 400 nm was
examined with old and new batches of the pyridinium salt 4a
(Figure 7): this transient could be observed only in the old batch
which was illuminated for long period time, but not in the new
batch, which is illuminated in a very short period of time,
indicating that the transients are formed from product, pyrido-
[2,1-a]-3,4-dihydroisoquinolinium salt. It is possible that a
photochemical transient formation from photocyclized product
occurs probably because of the open structure of the product
for hydrogen abstraction of the benzylic hydrogen or of the rigid
structure of the product in which the energy cascade is not
efficient. The transient from the product will not be discussed
further here.
Transients from 2-Bromo-N-ethylpyridinium Salt (6). In
order to identify the transient species from 3 or 4a, the transient
absorption spectra of a structurally halopyridinium salt 6 were
measured and are shown in Figure 8. At 3 µs delay after laser
flash, the spectral features consist of three absorption peaks at
about 250, 300, and 360 nm. At 300 µs delay, 250- and
290-nm transients remain, and 305- and 360-nm transients
disappear, while the bleaching at 277 nm is less. The 360-nm
transient is unequivocally recognized as the dibromide radical
anion as expected because of its properties (lifetime 14 µs, λmax
360 nm, and diminished by the presence of ferrous ion).
Conclusion
(20) Shida, T.; Hanazaki, I. Bull. Chem. Soc. Jpn. 1970, 43, 646-651.
(21) Anbar, M.; Thomas, J. K. J. Phys. Chem. 1964, 668, 3829-3835.
(22) Scaiano, J. C.; Barra, M.; Calabrese, G.; Sinta, R. J. Chem. Soc.,
Chem. Commun. 1992, 1418-1419.
Five- and six-membered heterocyclic ring systems can be
readily formed by photocyclization of 2-halopyridinium salts

Halopyridinium Salt Tethered to Arene
J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10683
linked to phenyl rings with methylene (1 and 2a) and ethylene
groups (3 and 4a). Photocyclization of halopyridinium salts
with an ethylene connector is less effective than a halo-
pyridinium salt with a methylene group, probably because the
halopyridinium salts with an ethylene group take a staggered
conformation freely while the halopyridinium salts with a
methylene group take only an “eclipsed” conformation of their
pyridinium and phenyl rings. For the pyridinium salts with a
methylene group, the photocyclization reaction of chloropyri-
dinium salt predominates over a bromopyridinium salt with a
methylene group, whereas for the pyridinium salt with an
ethylene group, this is not the case. Probably, for the chloro-
pyridinium salt with a methylene group, some assistance of the
excited chlorine moiety can hold the neighboring phenyl ring
for cyclization to take place, but for a chloropyridinium salt
with an ethylene group, the assistance cannot overcome the
eclipsed conformation energy for the cyclization. Thereby, the
bromopyridinium salt with an ethylene group is more reactive
than a chloropyridinium salt with an ethylene group, as expected.
The transient intermediates such as pyridinium salt σ (λmax
250 nm), 2,3-dihydrocyclohexadienyl radicals (λmax 310 nm),
and dibromide radical anion (λmax 360 nm) were detected and
characterized by using the nanosecond laser flash photolysis
technique. The lifetime of pyridinium salt σ, 2,3-dihydro-
cyclohexadienyl radicals, and dibromide radical anion in argon-
saturated water are 0.93 ms, 0.84 ms, and 25 µs, respectively.
Thus a photohomolytic radical mechanism involving the inter-
mediates is proposed. For a halopyridinium salt with an
ethylene group, the photocyclization originates mainly from the
singlet state along the minor triplet-state path, albeit ineffective.
For a halopyridinium salt with a methylene group, both singlet
and triplet states are involved for the photocyclization.
A
photonic ionization of the side chain of 2-halopyridinium salt
linked to arene in water does not occur, although the photo-
ionization possibility of the side chain of the photocyclized
product may not be excluded.
Acknowledgment. This research was supported by the Basic
Science Research Institute Program, Ministry of Education, 1996
(Project No. BSRI-96-3402) (Y.-T.P.) and the Korea Science
and Engineering Foundation through the Center for Molecular
Science (D.K.).
Supporting Information Available: UV, IR, NMR, and
element analysis data for 2a, 3, 4a, 4b, 5, 6, and 8 (2 pages).
See any current masthead page for ordering and Internet access
instruction.
JA970425U