686
Chemistry Letters 2000
Concentration Dependent Photodimerization of Azobenzenes in Solution
Masanobu Kojima,* Tomoko Takagi, and Takashi Karatsu†
Department of Bioscience and Biotechnology, Faculty of Agriculture, Shinshu University, Asahi, Matsumoto, Nagano 390-8621
†Department of Materials Technology, Faculty of Engineering, Chiba University, Yayoi-cho, Inage-ku, Chiba 263-8522
(Received March 13, 2000; CL-000242)
Cis-trans photoisomerization of azobenzenes in higher
concentrations proceeds through a bimolecular reaction process,
probably involving the dimer biradical intermediates, in competi-
tion with a unimolecular isomerization process, which is the
predominant reaction in lower concentrations.
As one of the best known photochromic molecules used in
optical molecular switches, azobenzene (AB) has attracted
considerable interest. Many investigations into the mechanism of
cis-trans photoisomerization and its applications have been
reported since the 1950's.1 However, attempts to apply the
photoisomerization process in solid, polymer matrix, and
polymer chains have often encountered the problem of poor
formation of the cis isomer (c-AB), which has previously been
explained in terms of the smaller free volume in solid and
polymers compared to that in solution; namely, it has been
thought likely that the trans-to-cis photoisomerization is greatly
suppressed in the smaller spaces.2 Furthermore, it has been
reported that on irradiation in cyclohexane using 313-nm light
the cis/trans (c/t) isomer ratio of AB in the photostationary state
(PSS) reached ca. 80/20 in a concentration of 1 × 10-3 M (M =
mol/dm3) and was independent of the initial concentration of
AB.3 In our most recent study of the photoisomerization of AB
in zeolite cavities,4 we have noticed incidentally that irradiation
of 1 × 10-2 M AB in cyclohexane using 313-nm light gives the
isomers at a ratio of c/t = ca. 55/45 in PSS, which is a markedly
different ratio from that in the lower concentrations reported
previously.3 Even in the case of aromatic alkenes there have so
far been few examples of the concentration having an effect on
the photoisomerization.5 Therefore, we reinvestigated the
concentration dependence of the c/t ratio of AB in PSS. In this
paper, on the basis of the effect of initial concentration, excitation
wavelength, substituent, and solvent on the cis-trans photoiso-
merization of AB, we propose a novel photoisomerization mech-
anism in higher concentrations. This mechanism consists of a
bimolecular photoisomerization process favoring the formation
of the trans isomer (t-AB), which competes with a cis-trans uni-
molecular isomerization process dependent on the ratio between
the molar absorption coefficients (ε) of c- and t-AB (εc-AB and
mers), and a mixed solvent of hexane/ethyl acetate = 90/10 as
an eluent. In contrast, irradiation of 2.5 × 10-3 and 1 × 10-2
M
AB yielded a very low quantity of c-AB (ca. 60% and 50%,
respectively) in PSS as shown in Figure 1. The concentration
dependent c/t ratio in PSS was also observed in 1,1,2-trichloro-
1,2,2-trifluoroethane (CCl2FCClF2), but the ratio in benzene
(PhH) was almost constant at ca. 80/20 regardless of the initial
concentration, as can be seen in Figure 1 and Table 1. In order
to explain these findings, we propose here, as shown in Scheme
1, that the photoisomerization of AB in higher concentrations
proceeds through the excimers and exciplex of AB, followed by
the production of a dimer singlet biradical intermediate (1).
This finally collapses into t-AB, which is more stable than c-
AB. Fluorescence emission from excited singlet state AB and
from the excimers and exciplex was not observed under our
conditions probably because the lifetime in the singlet excite
state is too short. In fact, the fluorescence lifetime of t-AB has
been determined to be ca. 25 ps.7 However, it should be noted
that on excitation of AB in cyclohexane by 254-nm light from a
160-W low-pressure mercury lamp, the c/t ratio in PSS was
found to be almost independent of the initial concentration (c/t
= ca. 15/85 in 1 × 10-2 M and 10/90 in 2 × 10-4 M). This is
because a unimolecular photoisomerization process on using
254-nm light also favors the formation of t-AB over that of c-
AB probably because the ratio of εc-AB/εt-AB is 8300/2200 at the
wavelength.6 It should be noted that, as indicated above, the
initial concentration has no effect on the c/t ratio in PSS in PhH
on using 313-nm light. This is probably because the formation
of the excimers and exciplex was disturbed by π-interaction
between benzene molecules and the benzene rings of AB.
εt-AB, respectively) as well as on that between the quantum yields
of isomerization for t-AB to c-AB (Φt-c) and c-AB to t-AB (Φc-t)
according to the excitation wavelength used.
When an aerated cyclohexane solution of 2 × 10-4 – 1 × 10-3
M AB was irradiated through a Hoya U-340 glass filter (313-nm
band pass; εc-AB and εt-AB are 1160 and 21100 at 313 nm,
respectively) using a 400-W high-pressure mercury lamp at room
temperature, the c/t ratio in PSS was determined to be ca. 80/20
by analyzing the irradiated solutions at 40 °C using a JASCO
Finepack SIL column (inner diameter, 4.6 mm; length, 25 cm),
271-nm light as monitoring light (isosbestic point of the iso-
Using trans-4-methoxyazobenzene (t-MeOAB), we fur-
ther examined the effect of the initial concentration on photo-
isomerization. The cis (c-MeOAB) and trans isomers were
analyzed under conditions similar to those described above for
AB (monitoring light, 300 nm; eluent, hexane/ethyl acetate =
Copyright © 2000 The Chemical Society of Japan
Kojima, Masanobu
