Ring-opening and -closure reaction dynamics of a photochromic dithienylethene derivative

Article abstract of DOI:10.1021/jp012614b

The influence of a bulky, strongly coupled substituent (benzoyl-phenyl-ethenyl, BPE) on the dynamics of the photoinduced ring-closure and ring-opening reactions of 1,2-bis[[2-methyl-thien-3-yl]]perfluorocyclopentene (BMTFP) in solution was investigated. UV/vis absorption spectroscopy in combination with chemical actinometry was employed to determine the reaction quantum yields. The relaxation and reaction dynamics were studied by transient absorption spectroscopy, exciting the S₀-S₂ transition of the open isomer with pump pulses at 288 nm, while 657 and 410 nm pulses were used to excite the S₁ and S₂ states of the closed isomer, respectively. Transient absorption spectra in the range of 350-950 nm were recorded using temporally delayed, white-light continuum probe pulses. After structural relaxation in the S₁ excited state, the ring-closure reaction takes place with high quantum yield through a conical intersection. This conical intersection acts also as relaxation funnel for the S₁ excited state of the closed isomer, from which it is separated by an energy barrier. The branching in the conical intersection favors the closed isomer, so that the ring-opening reaction yield is small (10^-3-10^-4). This yield was found to vary by a factor of 6 for different isomers obtained by rotation around the ethylene bond of the substituent.

Full text of DOI:10.1021/jp012614b

1654
J. Phys. Chem. A 2002, 106, 1654-1660
Ring-Opening and -Closure Reaction Dynamics of a Photochromic Dithienylethene
Derivative
Jan Ern,† Arthur T. Bens,‡ Hans-Dieter Martin,‡ Karla Kuldova,§ H. Peter Trommsdorff,§ and
Carola Kryschi*,†
Institut fu¨r Physikalische Chemie I, Friedrich-Alexander-UniVersita¨t Erlangen-Nu¨rnberg,
D-91058 Erlangen, Germany, Institut fu¨r Organische Chemie und Makromolekulare Chemie I,
Heinrich-Heine-UniVersita¨t, D-40225 Du¨sseldorf, Germany, and Laboratoire de Spectrome´trie Physique,
UniVersite´ Joseph-Fourier Grenoble I, CNRS (UMR 5588), B.P. 87, F-38402 St. Martin d’He`res, France
ReceiVed: July 11, 2001; In Final Form: October 10, 2001
The influence of a bulky, strongly coupled substituent (benzoyl-phenyl-ethenyl, BPE) on the dynamics of
the photoinduced ring-closure and ring-opening reactions of 1,2-bis[[2-methyl-thien-3-yl]]perfluorocy-
clopentene (BMTFP) in solution was investigated. UV/vis absorption spectroscopy in combination with
chemical actinometry was employed to determine the reaction quantum yields. The relaxation and reaction
dynamics were studied by transient absorption spectroscopy, exciting the S0-S2 transition of the open isomer
with pump pulses at 288 nm, while 657 and 410 nm pulses were used to excite the S1 and S2 states of the
closed isomer, respectively. Transient absorption spectra in the range of 350-950 nm were recorded using
temporally delayed, white-light continuum probe pulses. After structural relaxation in the S1 excited state,
the ring-closure reaction takes place with high quantum yield through a conical intersection. This conical
intersection acts also as relaxation funnel for the S1 excited state of the closed isomer, from which it is
separated by an energy barrier. The branching in the conical intersection favors the closed isomer, so that the
ring-opening reaction yield is small (10-3-10-4). This yield was found to vary by a factor of 6 for different
isomers obtained by rotation around the ethylene bond of the substituent.
1. Introduction
thiophene rings, on the reaction dynamics and quantum
yields.13-20 Substituents with various structural and electronic
properties, more or less strongly coupled with the photochromic
dithienylethene unit, have been introduced, and systems suitable
for the use in technologically relevant photochromic materials
such as photoswitchable liquid crystals and polymers were
prepared. As a result of this work, a more detailed understanding
of the various interactions between substituents and the pho-
tochromic dithienylethene switch became a central issue. Very
recently, we have reported femtosecond transient absorption
studies of the ring-opening and ring-closure reaction dynamics
of diformyl dithienylethene derivative (1,2-bis[5-formyl-2-
methyl-thien-3-yl]perfluorocyclopentene) in solution14,15 as well
as the dianthryl analogue (1,2-bis[5-anthryl-2-methyl-thien-3-
yl]perfluorocyclopentene).18
Photochromic 1,2-bis[2-methyl-thien-3-yl]perfluorocyclopen-
tene (BMTFP) derivatives have in recent years received
considerable attention because of their high potential for a
number of photonic applications.1-7 In addition, comprehensive
spectroscopic investigations of the fundamental photochemical
processes have been made.8-19 Irie and co-workers and,
independently, Lehn et al. optimized compounds with respect
to reaction quantum yield, resistance to photofatigue, and
thermal stability.1-12 The photochemical, electrocyclic ring-
opening and ring-closure reactions occur in the conrotatory
mode. For the dithienylethene derivatives, the ground-state
isomers are thermally stable because of the low aromatic
stabilization energy of the thiophene groups which gives rise
to a large potential barrier between the open-isomer and the
closed-isomer state.12 The open isomer is colorless, absorbing
in the near UV, whereas the closed isomer has a strong
absorption band in the visible range. Previous time-resolved
studies of electrocyclic reactions in other dithienylethene
derivatives have shown that both ring-opening and ring-closure
processes occur on the picosecond time scale.8-11,13,14,15,18
Our previous research activities on photochromic BMTFP
derivatives were aimed at the characterization of the influence
of π conjugated substituents, attached in the 5-position of the
The formyl substituent (CHO) is a π-electron donor that
interacts very strongly with the dithienylethene switch in the
first excited singlet state (S1). The transients of photoinduced
absorption and bleaching have been analyzed in terms of a
model potential extracted from calculations of the S0, S1, and
S2 potential energy hyperlines.14 Within the first picosecond after
excitation to the S1 state, the closed isomer is observed to
undergo a fast structural relaxation along the S1 potential energy
surface to a precursor state of the ring-opening process. The
rather long lifetime of this precursor, τ2 ) 13 ps, is dominated
by radiationless deactivation to the ground state, S0, of the closed
isomer, the quantum yield of the ring-opening reaction (3.6%)
being rather low. In contrast, the decay of the S1 state of the
open isomer is dominated by the ring-closure reaction with a
quantum yield of ΦRC ) 0.86.15 Both the decay of the S1 state
of the open isomer and the generation of the closed isomer are
* Prof. Dr. Carola Kryschi, Institut fu¨r Physikalische Chemie I, Friedrich-
Alexander-Universita¨t, Egerlandstrasse 3, D-91058 Erlangen, Germany.
Tel: +49-9131-85-27307. Fax: +49-9131-977172. E-mail: kryschi@
pctc.chemie.uni-erlangen.de.
† Friedrich-Alexander-Universita¨t Erlangen-Nu¨rnberg.
‡ Heinrich-Heine-Universita¨t.
§ Universite´ Joseph-Fourier Grenoble I.
10.1021/jp012614b CCC: $22.00 © 2002 American Chemical Society
Published on Web 02/06/2002

Photochromic Dithienylethene Derivative Dynamics
J. Phys. Chem. A, Vol. 106, No. 9, 2002 1655
657 nm and is about 630 fs for pump pulses at 288 nm. The
time dispersion of 1.4 ps over the spectral range of the white-
light continuum probe pulses was determined at spectral
intervals of about 80 cm-1 from the rise of the bleaching or
absorption transients, and all transient spectra shown here have
been corrected for this dispersion.
bpe-BMTFP was dissolved in spectrograde dichloromethane
at 2.5 · 10-5 mol/L. The solution was pumped at a velocity of
5 mL/s through a 400 µm flow cell to ensure a complete renewal
of the sample from pulse to pulse.
2.2. UV/vis Absorption Spectroscopy and Chemical Ac-
tinometry. Commercially (Aldrich) available solvents were
used. Absorption spectra were recorded with a Perkin-Elmer
lambda 9 spectrometer. Quantum yields were determined from
the evolution of the absorption spectra after irradiating the
solutions for various intervals of time. The experimental error
is estimated as 10%. The intensities of the light sources were
calibrated with the furylfulgide actinometer aberchrome 540.
The open isomers were irradiated with the UV lines of an Ar+
laser (351-363.5 nm), the closed isomers with laser lines at
514 nm (Ar+ laser), 575 and 690 nm (rhodamine 6G-dye laser),
and 632 nm (He-Ne laser).19
2.3. Syntheses. General. All syntheses, carried out under
argon or nitrogen atmosphere, were monitored by TLC using
aluminum-backed silica gel (230-400 mesh) or by HPLC using
a Hewlett-Packard ChemStation 1040 and 1050 series II
instrument and were protected strictly from light. Octafluoro-
cyclopentene was obtained as a donation from Zeon Corporation
(Japan). Solvents and other reagents were used as purchased
without further purification, unless stated otherwise. Column
chromatography was performed on silica gel (230-400 mesh)
using nitrogen- or argon-flushed solvents. Melting points:
uncorrected and Reichardt THERMOVAR. NMR: Varian VXR
300 and Bruker DRX 500 at frequencies of 300 and 500 MHz
for 1H and 470 MHz for 19F, respectively. Reference compounds
were TMS (1H) and CFCl3 (19F). UV/visible: Perkin-Elmer
lambda 19 spectrophotometer. IR: Perkin-Elmer 1420 spec-
trometer. EI-MS: Varian MAT 311A and MAT 8000. 1,2-Bis-
(2-methyl-5-(2-(4-benzoyl-phenyl-1-yl)ethen-1-yl)-thien-3-yl)-
perfluorocyclopentene 6 was synthesized according to Scheme
1. The details of the synthesis are described in the Appendix.
Figure 1. Structure formula of the open (top) and closed (bottom)
isomer of bpe-BMTFP.
exponential with a time constant of ca. 2 ps, so that the reaction
rate constant for ring-closure is kRC ≈ 4.3 ‚ 1011 s-1 15
.
These
results are characteristic for a small π-donating substituent. In
contrast, the coupling in the S1 state between anthracene
substituents, attached by a single bond to the 5 position of the
thiophene rings, and the dithienylethene switch is significantly
smaller because of the nearly perpendicular alignment of the
anthracene substituents with respect to the dithienylethene unit.
As a result, the S1 state of the closed isomer is less stabilized,
resulting in a relatively smaller ring-closure reaction quantum
yield and a larger ring-opening reaction quantum yield when
compared to CHO substituent.18,19
Here, we present a similar complete study of benzoyl-phenyl-
ethinyl substituted 1,2-bis[2-methyl-thien-3-yl]perfluorocyclo-
pentene (bpe-BMTFP, see Figure 1). A preliminary time-
resolved study of the closed isomer of this compound had shown
relaxation processes on the 1-10 ps time scale.13 The bpe
substituent is rather bulky and has an extended π electron
system, which is expected to couple strongly with that of the
dithienylethene unit. The reaction quantum yields and the
excited-state relaxation dynamics of the closed and open isomer
of bpe-BMTFP were investigated. AM1 computations of the
optimized geometry and heat of formation of the different
equilibrium conformations and isomers of bpe-BMTFP support
the analyses of the spectroscopic data.
3. Quantum Yields of Ring-Opening and Ring-Closure
Reaction
2. Experimental Section
2.1. Transient Absorption Spectroscopy. Transient absorp-
tion spectra with femtosecond time resolution were measured
using the pump-probe technique. Pump and probe pulses are
generated by a femtosecond laser system described in detail
elsewhere.14 The closed isomer was excited to the S1 state using
657 nm pulses from the optical parametric amplifier (OPA) or
to the S2 state using frequency-doubled amplified oscillator
pulses at 410 nm. The open isomer was excited at 288 nm to
the S2 state with the second harmonics of the 576 nm OPA
output. The intensities of the transmitted probe beam with and
without pump pulse excitation, I(p) and I(0), respectively, were
measured by chopping the pump pulse beam at a frequency of
6 Hz. The ratio I(p)/I(0) equals the pump pulse-induced change
of transmittance ∆T ) T(p)/T(0) in the sample. For the small
changes recorded here, 1 - ∆T equals the sum of the difference
spectra of all species contributing to the signal, weighted by
their respective concentrations. Transmission spectra of the
sample were obtained by dispersing the probe pulses with a
polychromator containing a 400 lines/mm grating in combination
with a CCD detector system (LOT, InstaSpec IV). The overall
instrumental time resolution is 120 fs for excitation at 410 and
Figure 2 shows stationary absorption spectra of the closed
isomer (solid line) and of the open isomer (dashed line) of bpe-
BMTFP. The lowest energy absorption band of the open isomer
at 355 nm correlates with the S0-S1 of the bpe substituents,
whereas the S0-S2 absorption band, peaking at 245 nm, involves
essentially the lowest-energy ππ* state of the dithienylethene
switch. For the closed isomer, the absorption maxima at 620
and 410 nm are assigned to the S0-S1 and S0-S2 transitions of
the dithienylethene unit, respectively. The excitation energy (620
nm) of the S0-S1 transition of the closed isomer is fairly low
as compared with different substituted dithienylethene deriva-
tives17 and is a proof of the strong coupling between the
π-electron systems of substituent and dithienylethene unit. For
example, the closed isomer of the unsubstituted dithienylethene
compound has its lowest energy transition at 520 nm and that
of the dianthryl derivative is located at 540 nm, while the strong
coupling in the closed isomer of formyl substituted compound
is responsible for the large redshift of the absorption to 620
nm.15 The open isomer of bpe-BMFTP may exist as two
conformers.18 The so-called parallel (p) and antiparallel (a-p)

1656 J. Phys. Chem. A, Vol. 106, No. 9, 2002
Ern et al.
Figure 2. Absorption spectra of the closed isomer (thick solid line)
and the open isomer (thick dashed line) of bpe-BMTFP in a 1.1 · 10-5
mol/L CH2Cl2 solution.
SCHEME 1
Figure 3. Structure formula of the three cis-trans isomers of
bpe-BMTFP.
behavior, in the sense that the quantum yield depends on both
the wavelength of excitation and the doses of irradiation. In
addition, the spectrum of the closed isomer was observed to
evolve as a function of the doses of irradiation. For example,
the quantum yield at the beginning of irradiation decreases from
1.3 ‚ 10-3 to 0.41 ‚ 10-3 when the wavelength of the irradiation
is shifted from 514 to 690 nm. After reduction of the maximum
absorption of the closed isomer by irradiation at 514 nm to 10%
of its initial value, the resulting spectrum is redshifted by 12
nm. During such an irradiation, the measured quantum yield
decreases by a factor of more than 4. The evolution of a solution
under irradiation was found to be only dependent on the doses
of irradiation and not on the intensity, so that an interruption
of the irradiation of the sample for several hours had no
influence on the results. All these observations were made in
solutions of toluene and are independent of the concentration
in the range of 5 ‚ 10-4 to 5 ‚ 10-5 mol/L.19 These results clearly
indicate the presence of a mixture of different isomers with
different reaction quantum yields, which do not interconvert
on the time scale of hours. If the minimal presence of just two
isomers (i1 and i2) is assumed, all observations can be
quantitatively analyzed, and the following properties of these
isomers are obtained: The initial relative concentrations of the
two isomers i1 and i2 are 0.6 and 0.4, respectively, and the
quantum yield of both isomers is wavelength-dependent and
decreases with increasing wavelength. Between 514 and 690
nm, the change is from 1.6 ‚ 10-3 to 0.55 ‚ 10-3 for i1 and
from 2.5 ‚ 10-4 to 0.9 ‚ 10-5 for i2. Analogous wavelength
dependence of quantum yields had also been reported by Irie
and Mohri22 and was attributed to the effect of excess vibrational
energy. Alternatively, the presence of more isomers or inter-
converting conformers could also contribute to this behavior.
However, because of the increased uncertainty associated with
the larger number of parameters, no analysis invoking a greater
number of species (isomers or conformers) was made here.
The most likely isomers are the three different cis-trans
isomers of bpe (see Figure 3). To avoid confusion with the open
and closed isomers of the dithienylethene unit, these isomers
will be referred to in the following as bpe isomers. AM1
computations21 of the optimized geometries result in different
values of the heat of formation, ∆HF, for the different isomers:
∆HF(c:cis-cis) ) -483 kJ/mol, ∆HF(c:cis-trans) ) -497 kJ/mol with
conformers are obtained from the planar, closed isomer by con-
and disrotatory rotations of the thiophene rings and have the
local symmetry of Cs and C2, respectively.20 For steric reasons,
only the a-p conformer may undergo the ring-closure reaction.
To evaluate the relative populations of the two conformers, the
heat of formation, ∆HF, was calculated using the AM1 procedure
of Hyperchem21 with the following result: ∆HF(o:p) ) -587.14
(o:a-p)
kJ/mol and ∆HF
) 592.22 kJ/mol. Neglecting entropy
differences and corrections due to differences in solvation
energy, the Boltzmann populations of both conformers, Fp and
F
a-p, are determined as Fp ≈ 12% and Fa-p ≈ 88%. The value
of Fa-p is the upper limit for the overall quantum yield of the
ring-closure reaction, consistent with the experimental value of
ΦRC ) 0.76 (363 nm excitation). Taking the above value of
Fa-p ≈ 88%, the quantum yield for the a-p conformer is
(a-p)
ΦRC
) 0.86.
In contrast, the quantum yield for the ring-opening reaction
is rather small, and the measurements expose a more complex

Photochromic Dithienylethene Derivative Dynamics
J. Phys. Chem. A, Vol. 106, No. 9, 2002 1657
Figure 4. Temporal evolution of the transient absorption spectra of
the open isomer excited at 288 nm and recorded at different delay times
τD as indicated. The S0-S1 absorption spectrum of the closed isomer
is shown in the top panel.
(c:trans-trans)
a statistical weight of 2, and ∆HF
) -513 kJ/mol.
We conclude that not only these different isomers may be
formed in the course of the chemical synthesis with different
probabilities, but also they do differ in their photochemical
behavior, and we suggest that these isomers are responsible for
the observed dependence of the quantum yield on the irradiation
dose and wavelength.
4. Transient Absorption Spectra
To analyze the time evolution of the transient spectroscopic
measurements, a decomposition of the spectra into a limited
number of components was made in the following way: for
each of the two (i.e., open and closed) isomers, the spectra at
all delay times were reproduced as the sum of a minimal number
of the same spectral components with different amplitudes. The
initial unrestricted decomposition clearly exposed, in addition
to components that originate from the population of excited state
species, components associated with the stable spectra of the
two isomers. To reduce the number of free parameters in the
final fitting, the position and width of these components were
forced to reproduce the stationary spectra. The sum of compo-
nents with the same time evolution was assumed to be associated
with one excited species, generated subsequent to the optical
excitation. In this way, the time evolution of the population of
all species was extracted.
The quantum yield measurements, presented above, clearly
indicate the presence of a mixture of different bpe isomers.
While the small ring-opening reaction yield strongly differs for
these species, the high ring-closure yield is insensitive and,
within the experimental uncertainty, no differences between bpe
isomers were found. Not surprisingly, the relaxation dynamics,
discussed below, reflects this inhomogeneous composition of
the samples.
Figure 5. Decomposition of the transient absorption spectra in Figure
4 into the spectral components a1-a4 and the ground-state absorption
spectrum of the closed isomer (marked “stationary”). (a) Example of
this decomposition at three different delay times. (b) Time evolution
of the spectral components.
nents, as described above, is indicted in Figure 5a, whereas in
Figure 5b the temporal evolution of these species is shown.
The temporal evolution of the transient absorption spectra
between -0.8 and +54 ps clearly exposes the generation of a
long-lived species with a rise time constant of 6 ps. This
component is attributed to the S0-S1 absorption of the closed
isomer generated by the ring-closure reaction (see the ground-
state absorption spectrum in Figure 4). In addition, shorter-lived
transient absorption bands, peaking at ca. 730 (a1), 580 (a2),
and 525 nm (a3) appear within the first picosecond and decay
with time constants around 6 ps. The good agreement between
these rise and decay times clearly indicates that the ring-closure
reaction is the direct relaxation pathway of these excited species.
The transient at 470 nm (a4) decays within the first picosecond.
In addition, a rise time of ca. 0.4 ps is observed for the 525 nm
(a3) transient, whereas the neighboring 580 nm (a2) band has
a correspondingly short decay component in addition to the 6
ps lifetime. These subpicosecond evolutions are attributed to a
structural relaxation within the S1 excited state of the open
isomer. The observation that the excited-state transients do not
4.1. Ring-Closure Reaction and Open Isomer Transients.
Transient absorption spectra of the open isomer of bpe-BMTFP
in dichloromethane, recorded with pump pulses tuned to the
S0-S2 transition at 288 nm, are shown in Figure 4. The spectrum
at τD ) -0.8 ps provides the baseline with ∆T(λ) ≈ 1. The
stationary absorption spectrum of the closed isomer (top panel)
contains the S0-S1 absorption band between 570 and 750 nm
and the onset of the S0-S2 transition below 425 nm. For some
of these spectra, the decomposition into five different compo-

1658 J. Phys. Chem. A, Vol. 106, No. 9, 2002
Ern et al.
Figure 7. Decomposition of the transient absorption spectra in Figure
6b into spectral components a1, a2, and the ground-state absorption
spectrum of the closed isomer (marked “stationary”). (a) Example of
this decomposition at three different delay times. (b) Time evolution
of the spectral components.
as is expected, since the contribution of the ring-opening reaction
to the decay dynamics is negligible and only the decay of S1 to
S0 of the closed isomer is observed. The bleaching of the ground
state is clearly identified, as are the components due to excited-
state absorption. The decay of all transients can be fitted with
biexponential functions of delay time, τD, as follows: ∆T ) 1
+ A1·exp(-τD/τ1) + A2·exp(-τD/τ2). The short time constant τ1
lies between 0.3 and 0.5 ps, and values between 5 and 11 ps
are obtained for the long time constant, τ2. The accuracy of the
data does not warrant any fit to sums of more exponentials,
even though it is likely that more than just two bpe isomers are
present (see above). Nevertheless the inhomogeneous behavior
of the sample is obvious from the different values obtained for
τ2, and, more important, the observation of a short as well as a
long time constant for each of the spectral components is firmly
established.
Figure 6. Temporal evolution of the transient absorption spectra of
the closed isomer recorded at different delay times τD as indicated:
excited at (a) 657 and (b) 410 nm.
decay to zero is ascribed to the presence of the open isomer in
the parallel conformation, which cannot decay via the fast ring-
closure reaction, and the decay of which is beyond the time
scale of the measurement. As might be expected, the spectra of
the different bpe isomers are not separated at the resolution of
the experiments, and the relative amplitude of the residual
absorption is therefore an indication of the population of the
nonreacting p conformer, the value obtained here, Fp ≈ 15%,
is consistent with the value of given above in Section 3.
4.2. Closed Isomer Transients. The excited-state dynamics
of the closed isomer was studied using pump-pulse excitation
wavelengths at 657 and 410 nm, resonant with the S1 and S2
excited states, respectively. Typical transient spectra are shown
in Figure 6a and b. In fact, aside from the presence of stray
light of the pump pulse at 657 nm in Figure 6a, the transient
spectra are virtually the same. This observation is consistent
with the fact that no rise times are resolved in the spectra of
Figure 6b, demonstrating that the S2 to S1 relaxation is faster
than our instrumental resolution. The decomposition in spectral
components, i.e., the ground-state absorption spectrum of the
closed isomer and excited-state spectra, a1 and a2, is shown
for some of the transient spectra of Figure 6b in Figure 7a, and
the corresponding time evolution of these components is shown
in Figure 7b. Overall, the picture that emerges is quite simple,
5. Discussion and Conclusions
The results obtained here for the ring-closure reaction are
fully consistent with observations made for other dithienylethene
derivatives. Figure 8 summarizes the overall picture of the
relaxation and reaction dynamics. After excitation, the open
isomer relaxes within less than 1 ps on the S1 potential energy
surface to a state (OP, “precursor state”) with a longer lifetime
of about τP ) 6 ps. Calculations indicate that the S1 potential
energy surface is fairly flat along the reaction coordinate down
to about a distance of about 3 Å, after which the descent to the
region of the conical intersection becomes steep.14,15,18 In fact,
the excitation energy of the S1 state of the open isomer is an
important factor, this slope becoming smaller for substituents
with low excitation energies. It is reasonable to accept that the
lifetime of the precursor state is the smaller the higher the slope
along the reaction coordinate, and in line with this proposition,

Photochromic Dithienylethene Derivative Dynamics
J. Phys. Chem. A, Vol. 106, No. 9, 2002 1659
in 400 mL icy acetic acid (p.a.). A solution of 25 mL bromine
(0.49 mol) in 200 mL icy acetic acid was added at room
temperature to the stirred mixture over 9 h. This darkened
solution was stirred intensively in the absence of light for a
further 16 h at room temperature, and then it was carefully added
to a 3.5 L solution of saturated aqueous sodium bicarbonate.
The product was extracted with 1.5 L diethyl ether in several
small amounts, and the organic phase was washed with aqueous
sodium bicarbonate solution (5% w/v, 3 · 500 mL) and with
water (2 · 500 mL). Then, the product was dried over Na2SO4,
filtered, and evaporated in vacuo to give a yellow-brown solid,
which was recrystallized in n-hexane to yield 61.9 g light yellow
crystals. Mp: 57.5 °C (reference mp: 58 °C).4,5,19 Yield: 76%.
1H NMR (300 MHz, CDCl3): δ ) 2.49 ppm (s, 3H, -CH3);
7.59 ppm (s, 1H, H-4); 9.78 ppm (s, 1H, -CHO).
Figure 8. Reaction scheme. The states Ci and Oi denote the states of
the closed and open isomers, respectively, the index P referring to the
precursor state reached, and CI stands for conical intersection. For
simplicity, only the important and dominant rate processes are indicated.
3-Bromo-2-methyl-5-thiophenecarboxaldehyde-dimethyl-
acetal 2. 3-Bromo-2-methyl-5-thiophenecarboxaldehyde (61.7
g, 0.3 mol) 1, 50 mL trimethyl orthoformiate, and 500 mg
p-toluoenesulfonic acid were suspended in 100 mL anhydrous
methanol. This mixture was refluxed for 8 h under a nitrogen
atmosphere. Water (500 mL) was then added, and the product
was extracted with 0.75 L diethyl ether and washed with aqueous
sodium bicarbonate solution (5% w/v, 2 · 250 mL) and water
(3 · 250 mL). Removal of the solvent gave a yellow residue
which was triturated in dichloromethane and subsequently
purified by CC on silica gel (n-hexane/dichloromethane 3:1
(v/v) as eluent). The yielded pale yellow oil was dried in vacuo
at 0.1 mbar, yielding 91% (reference yield: 97% yield).4,5,19
1H NMR (300 MHz, CDCl3): δ ) 2.37 ppm (s, 3H, -CH3);
3.34 ppm (s, 6H, -(OCH3)2); 5.52 ppm (s, 1H, -CH(OCH3)2);
6.88 ppm (s, 1H, H-4).
we observe that for the bpe derivative studied here (S1 at 350
nm) the lifetime of 6 ps is intermediate between that observed
for the bianthryl derivative (S1 at 390 nm, τP ) 10 ps) and the
biformyl derivative (S1 at 300 nm, τP ) 2 ps). This picture
confirms the view, obtained in studies of other dithienylethene
derivatives, that the ring-closure reaction is essentially barrierless
and downhill. Noticeable barriers are expected to exist, however,
when the excitation energy of the substituent becomes even
lower.
For the closed isomer, the transient spectra reflect just the
relaxation of the S1 excited state, because the quantum yield of
the ring-opening reaction is negligibly small. The spectrally
homogeneous, but temporally at least biexponential decay
indicates that an energy barrier exists between the minimum of
the S1 excited state and the relaxation funnel, which we propose
to be the conical intersection. This proposition is consistent with
the observation that the quantum yield of the ring-opening
reaction increases both with temperature and excitation en-
ergy.17,19 The branching between the open and the closed
isomers at the conical intersection is then dependent on the
excess energy, increasing in favor of the open isomer with
increasing energy. The fact that this branching is very sensitive
to the geometry of the substituent is a new, important result
found here. The change by a factor of 6 of the ring-opening
quantum yield, observed for different isomers of the bpe
derivative, demonstrates this extreme sensitivity to the structure
of the substituent.
The very small overall quantum yield of the ring-opening
reaction is consistent with the stabilization of the S1 excited
state of the closed isomer as compared to the unsubstituted
dithienylethene and the dianthryl derivatives. This stabilization
arises from the delocalization of the π electron system across
central unit and the bpe substituents and leads to a shift of the
S0-S1 absorption to longer wavelength (620 nm).17 Analogous
observations have been made for different polyene- and caro-
tinoid-substituted dithienylethene derivatives.16,18 The ring-
closure reaction yield, on the other hand, benefits from the bpe
substituents since the population of the reacting over nonreacting
conformers is favored.
1,2-Bis-(5-formyl-2-methyl-thien-3-yl)perfluorocyclopen-
tene 4. 3-Bromo-2-methyl-5-thiophenecarboxaldehyde-dimethyl-
acetal (36.42 g, 0.145 mol) 2 was dissolved in 800 mL
anhydrous diethyl ether. This solution was cooled under an argon
atmosphere to -78 °C. n-Butyllithium (93.75 mL, 0.150 mol,
1.6 M in n-hexane solution) was added dropwise over 20 min.
The yellow-orange reaction mixture was stirred for further 10
min at -78 °C, before 14.84 g (0.070 mol) perfluorocyclopen-
tene under cooling were added dropwise to the gradually
decoloring solution, and the reaction mixture was stirred for 1
h at -78 °C and a further 3 h at room temperature. The reaction
was terminated by diluting with 500 mL diethyl ether and
washing with dilute hydrochloric acid (1% v/v, 2 · 500 mL),
500 mL saturated sodium bicarbonate solution, and 2 · 500 mL
water. The combined aqueous phases were re-extracted with
500 mL diethyl ether, dried over MgSO4, filtered and evaporated
in vacuo to yield 33.8 g of the raw dimethylacetal 3. An
analytical sample of the yellow oil was obtained through CC
with dichloromethane/n-hexane 2:1 (p.v.). 1H NMR (300 MHz,
CDCl3): δ ) 1.88 ppm (s, 6H, -CH3); 3.33 ppm (s, 12H,
-OCH3); 5.53 ppm (s, 2H, -CH(-OCH3)2); 7.00 ppm (s, 2H,
H-4).
Dimethylacetal 3 was not isolated. The protecting group was
removed by direct acid hydrolysis:
Acknowledgment. This research has been supported by the
Volkswagenstiftung (PHOTONIK, Az.: I/71 939). A.T.B. is
grateful for partial support from A. Ro¨hm and the Ro¨hm
Foundation. The gift of octafluorocyclopentene by the Zeon
Corporation (Japan) is acknowledged.
16.9 g of the raw dimethylacetal 3 and 500 mg p-toluene-
sulfonic acid were dissolved in 500 mL diethyl ether. Then 1.5
L water with 1 mL aqueous 43% hydrobromic acid were added
to the intensively stirred solution. The resulting yellow suspen-
sion was thereafter stirred over 72 h at room temperature. After
this period, the complete reaction solution was extracted with
diethyl ether (2 · 250 mL), washed with saturated sodium
bicarbonate solution (2 · 250 mL) and 750 mL water, dried over
Na2SO4, and then filtered. Removal of the solvent in vacuo,
Appendix
3-Bromo-2-methyl-5-thiophenecarboxaldehyde 1. 5-Meth-
yl-thiophene-carboxaldehyde (50.0 g, 0.4 mol) was dissolved

1660 J. Phys. Chem. A, Vol. 106, No. 9, 2002
Ern et al.
followed by trituration in dichloromethane and by purification
with CC on silica gel (n-hexane/dichloromethane 3:1 (p.v.) as
eluent) gave 6.7 g bisaldehyde 4 as a pure white powder. Mp:
182 °C. Yield: 45% (reference yield and mp: 52% and 182
°C, respectively.4,5,19 1H NMR (300 MHz, CDCl3): open
form: δ ) 2.03 ppm (s, 6H, -CH3); 7.75 ppm (s, 2H, H-4);
9.86 ppm (2H, s, -CHO). Closed form: δ ) 2.18 ppm (s, 6H,
-CH3); 7.74 ppm (s, 2H, H-4); 9.86 ppm (s, 2H, -CHO). -
19F NMR (200 MHz, CDCl3, ref. CFCl3): open form: δ )
-109.60 ppm (t, 4F, CF2 at C-3 und C-5); -131.12 ppm (quint.,
2F, CF2 at C-4). Closed form: δ ) -113.65 and -114.50 ppm
(m, 4F, CF2 at C-3 und C-5); -134.091 ppm (quint., 2F, CF2
at C-4). IR (solid, KBr): ν˜[cm-1] ) 1280 cm-1 (C-F-Val.);
1670 cm-1 (CHO-Val.); 2870 cm-1 (C-H-Val.(w)). MS (70
eV, EI) m/z (%) ) 424 (100) [M+], 409 (65) [M+-CH3], 404
(8) [M+-HF], 32 (19) [S+]. C17H10F6O2S2 (424.38) calcd.:
48.11% C, 2.38% H. Found: 47.80% C, 2.45% H.
1,2-Bis-(2-methyl-5-(2-(4-benzoyl-phenyl-1-yl)ethen-1-yl)-
thien-3-yl)perfluorocyclopentene 6. 4-Benzoyl-benzyl-triph-
enylphosphoniumbromid (3285 mg, 6.0 mmol) 5 and 850 mg
(2.0 mmol) bisaldehyde 4 were suspended under an argon
atmosphere in 75 mL 1,2-epoxybutane. The initially white
suspension was stirred for 5 days at room temperature in the
dark, becoming a darkened mixture. Removal of the solvent
gave a residue which was triturated in dichloromethane and
subsequently purified by CC on silica gel (n-hexane/dichlo-
romethane/ethyl acetate 3:2:1 (v/v/v) as eluent). The pale yellow-
green solid was dried in vacuo at 0.1 mbar, giving 1187 mg of
6. Yield: 76%. Mp: 63 °C. 1H NMR (300 MHz, CDCl3 (product
contains a mixture of cis/trans isomeres): δ ) 1.78-1.97 ppm
(s, 6H, -CH3); 6.56-7.29 ppm (m, 4H, olefinic protons); 6.98
ppm, 7.09 ppm (2H, each d, thiophene-H at C-4_); 7.40-7.62
ppm (m, 10H, phenyl protons); 7.78-7.83 ppm (m, 8H, benzoyl
protons). 19F NMR (200 MHz, CDCl3, Ref. CFCl3) open form:
δ ) -110.54 ppm (m, 4F, CF2 at C-3 and C-5); -132.24 ppm
(m, 2F, CF2 at C-4). Closed form: δ ) -112.97 and -115.06
ppm (m, 4F, CF2 at C-3 und C-5); -135.14 ppm (m, 2F, CF2
at C-4). IR (solid, KBr): ν˜[cm-1] ) 710, 750, 800 (phenyl);
990 (C-H-Val. (conjug.-CdC-)); 1285 cm-1 (C-F-Val.);
1605 cm-1 (-CdC-Val.(phenyl)); 1660 cm-1 (-CdO-Val.-
(aryl)). MS (70 eV, EI) m/z (%) ) 781 (10) [M+], 602 (6)
[M+-C13H9O], 260 (56) [M3+3e-], 180 (2) [C13H9O+], 105 (66)
[C7H5O+], 77 (30) [C6H5+], 44 (100), 32 (58) [S+]. C45H30F6O2S2
(780.85) calcd.: 69.22% C, 3.87% H. Found: 69.63% C, 4.08%
H.
References and Notes
(1) Irie, M. Jpn. J. Appl. Phys. 1989, 28, 215.
(2) Tsujioka, T.; Tatezono, F.; Harada, T.; Kuroki, K.; Irie, M. Jpn. J.
Appl. Phys. 1994, 33, 5788.
(3) Hamano, M.; Irie, M. Jpn. J. Appl. 1996, 35, 1764.
(4) Gilat, S. L.; Kawai, S. H.; Lehn, J.-M. J. Chem. Soc., Chem.
Commun. 1993, 1439.
(5) Gilat, S. L.; Kawai, S. H.; Lehn, J.-M. Chem. Eur. J. 1995, 1, 275-
284.
(6) Tsivgoulis, G. M.; Lehn, J.-M. Angew. Chem. 1995, 107, 1188.
(7) Irie, M. Chem. ReV. 2000, 100, 1685.
(8) Tamai, N.; Miyasaka, H. Chem. ReV. 2000, 100, 1875.
(9) Miyasaka, H.; Araki, S.; Tabata, A.; Nobuto, T.; Mataga, N.; Irie,
M. Chem. Phys. Lett. 1994, 230, 249.
(10) Tamai, N.; Saika, T.; Shimidzu, T.; Irie, M. J. Phys. Chem. 1996,
100, 4689.
(11) Miyasaka, H.; Nobuto, T.; Itaya, A.; Tamai, N.; Irie, M. Chem.
Phys. Lett. 1997, 269, 281.
(12) Kobatake, S.; Shibata, K.; Uchida, K.; Irie, M. J. Am. Chem. Soc.
2000, 122, 12135.
(13) Ern, J.; Bens, A. T.; Bock, A.; Martin, H.-D.; Kryschi, C. J.
Luminescence 1998, 76-77, 90.
(14) Ern, J.; Bens, A. T.; Martin, H.-D.; Mukamel, S.; Schmid, D.;
Tretiak, S.; Tsiper, E.; Kryschi, C. Chem. Phys. 1999, 246, 115.
(15) Ern, J.; Bens, A. T.; Martin, H.-D.; Mukamel, S.; Schmid, D.;
Tretiak, S.; Tsiper, E.; Kryschi, C. J. Luminescence 2000, 87-88, 742.
(16) Bens, A.; Frewert, D.; Kodatis, K.; Martin, H.-D.; Kryschi, C.;
Trommsdorff, H. P. Eur. J. Org. Chem. 1998, 2333, 3-2338.
(17) Kuldova´, K.; Tsyganenko, K.; Corval, A.; Trommsdorff, H. P.;
Bens, A. T.; Kryschi, C. Synth. Met. 2000, 115, 163.
(18) Ern, J.; Bens, A. T.; Martin, H.-D.; Mukamel, S.; Tretiak, S.;
Tsyganenko, K.; Kuldova, K.; Trommsdorff, H. P.; Kryschi, C. J. Phys.
Chem. 2001, A105, 1741.
(19) Tsyganenko, K. Ph.D. Thesis, Universite Joseph-Fourier Grenoble
I, 1998.
(20) Uchida, K.; Tsuchida, E.; Aoi, Y.; Nakamura, Sh.; Irie, M. Chem.
Lett. 1999, 63.
(21) Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J. J. P. J.
Am. Soc. 1985, 107, 3902.
(22) Irie, M.; Mohri, M. J. Org. Chem. 1988, 53, 803.