Excited state behavior of diarylethenes in the subnanosecond timescale: The role of an upper singlet

Article abstract of DOI:10.1021/ja961859c

Picosecond absorption measurements of the trans isomers of 9-styrylanthracene (9-StAn), n-styrylnaphthalene (n-StN, with n = 1 and 2), and n-styrylphenanthrene (n-StPh, with n = 1, 2, and 9) were carried out in a number of solvents at room temperature. In

Full text of DOI:10.1021/ja961859c

J. Am. Chem. Soc. 1996, 118, 10879-10887
10879
Excited State Behavior of Diarylethenes in the Subnanosecond
Timescale: The Role of an Upper Singlet
G. G. Aloisi,§ F. Elisei,§ L. Latterini,§,† U. Mazzucato,*,§ and M. A. J. Rodgers†
Contribution from the Dipartimento di Chimica, UniVersita` di Perugia, 06123 Perugia, Italy,
and the Center of Photochemical Sciences, Bowling Green State UniVersity,
Bowling Green, Ohio 43403
ReceiVed June 3, 1996X
Abstract: Picosecond absorption measurements of the trans isomers of 9-styrylanthracene (9-StAn), n-styrylnaph-
thalene (n-StN, with n ) 1 and 2), and n-styrylphenanthrene (n-StPh, with n ) 1, 2, and 9) were carried out in a
number of solvents at room temperature. In all cases, a short-lived transient (with τS ) 0.1 ÷ 0.8 ns) was produced
monophotonically within the laser pulse. This transient, which absorbs in a wide wavelength range (400-750 nm),
was assigned to the S2 state. At longer delay time, the S1 f Sn absorption (with lifetime τL) replaced that of S2, its
assignment being strongly supported by the comparison between τL and the fluorescence lifetime of these diarylethenes.
The S1 f Sn absorption of two distinct rotamers was detected for 2-StN in solvents of different polarity, while for
the other compounds only one longer-lived transient was detected. Complementary fluorimetric measurements by
phase-modulation technique confirmed the presence of a second decay component in the picosecond region with a
lifetime corresponding to the τS values measured by laser flash photolysis. The effect of temperature studied in both
picosecond and nanosecond regions suggests that the S2 f S1 internal conversion is not the only decay pathway of
S2. In particular, for 9-StAn, it is shown that both S1 and S2 are involved in the singlet-triplet intersystem crossing,
and, at least in the cases of 1-StN and 2-StPh, the S2 state (or both S1 and S2) is (are) involved in the trans f perp
decay process. An investigation of the effect of excitation wavelength on the fluorescence and trans f cis
photoisomerization quantum yields was found to support such hypothesis of the role played by upper excited states
in the relaxation processes of some of the styrylarenes investigated.
1. Introduction
ing on the nature of the larger arene and the position of the
styryl group. Recently, the importance of the nature of the
lowest excited singlet state (S1) of trans-1,2-diarylethenes
(DAEs) in determining the rates of the competitive reactive
(trans f cis) and physical relaxation processes was reported.5
The relative energies of the ethenic and arenic states and of the
La and Lb components of the latter are particularly important in
favoring the isomerization process or radiative decay. From
this point of view, it is interesting to obtain information on the
nature not only of S1 but also of the nearby located upper excited
states.
In the past, we dealt with cases where the quantum yields of
the reactive and radiative channels of relaxation were not easily
understandable only on the basis of the properties of the lowest
excited states of singlet and triplet multiplicity. The intervention
of upper excited states Sn or Tn, having deactivation channels
bypassing S1 and T1, respectively, offered an interesting
explanation of the observed behavior, even if such a hypothesis
could be considered to be rather speculative in the absence of
direct experimental evidence.
The cis-trans photoisomerization of stilbene-like molecules
where one (or both) phenyl groups are replaced by larger
polycyclic aryl groups has been studied extensively for almost
two decades.1-5 In our laboratory, particular attention has been
given to compounds bearing one naphthyl,6,7 phenanthryl,8 or
anthryl group.9,10 Quite different photochemical and photo-
physical behaviors were observed for these compounds depend-
* Author to whom correspondence should be addressed.
§ Universita` di Perugia.
† Bowling Green State University.
X Abstract published in AdVance ACS Abstracts, October 1, 1996.
(1) Saltiel, J.; D’Agostino, J.; Megarity, E. D.; Metts, L.; Neuberger, K.
R.; Wrighton M.; Zafiriou, O. C. Org. Photochem. 1973, 3, 1. Saltiel, J.;
Charlton, J. L. Rearrangement in the Ground and Excited States; de Mayo,
P., Ed.; Academic Press: New York, 1980; Vol. 3, p 25. Saltiel, J.; Sun,
Y-P. In Photochromism; Molecules and Systems; Du¨rr, H., Bouas-Laurent,
H., Eds., Elsevier: Amsterdam, 1990; Chapter 3.
(2) Wismontski-Knittel, T.; Fischer, G.; Fischer E. J. Chem. Soc., Perkin
Trans. II 1974, 1930. Castel, N.; Fischer, E.; Strauch, M.; Niemeyer, M.;
Lu¨ttke, W. J. Photochem. Photobiol., A: Chem. 1991, 57, 301 and references
therein.
A typical example has been reported with 1-StN,7 known to
isomerize in the singlet manifold near room temperature with a
high activation energy. The fluorescence (φF) and photoisomer-
ization (φtfc) quantum yields were found to be coupled and
complementary above ca. 300 K. When the temperature
decreased below 300-270 K (the value depends on the solvent),
φF remained practically constant down to the liquid nitrogen
temperature, with its value remaining below unity. In contrast,
φtfc decreased due to thermal and viscosity barriers and was
inhibited in rigid matrices. The constancy of φF in the low
temperature region seemed to indicate that the fluorescent state
(3) Mazzucato, U. Pure Appl. Chem. 1982, 54, 1705. Mazzucato, U. Gazz.
Chim. Ital. 1987, 117, 661. Mazzucato, U.; Aloisi, G. G.; Elisei, F. Proc.
Indian Acad. Sci. (Chem. Sci.) 1993, 105, 475.
(4) Arai, T.; Karatsu, T.; Misawa, H.; Kuriyama, Y.; Okamoto, H.;
Hiresaki, T.; Furuuchi, H.; Zeng, H.; Sakuragi, H.; Tokumaru, K. Pure Appl.
Chem. 1988, 60, 989. Tokumaru, K.; Arai, T. J. Photochem. Photobiol., A:
Chem. 1992, 65, 1. Arai, T.; Tokumaru, K. Chem. ReV. 1993, 93, 23.
(5) Bartocci, G.; Mazzucato, U.; Spalletti, A.; Orlandi, G.; Poggi, G. J.
Chem. Soc., Faraday Trans. 1992, 88, 3139.
(6) Bartocci, G.; Masetti, F.; Mazzucato, U.; Marconi, G. J. Chem. Soc.,
Faraday Trans. 2 1984, 80, 1093.
(7) Elisei, F.; Mazzucato, U.; Go¨rner H. J. Chem. Soc., Faraday Trans.
1 1989, 85, 1469.
(8) Bartocci, G.; Masetti, F.; Mazzucato, U.; Spalletti, A.; Baraldi, I.;
Momicchioli, F. J. Phys. Chem. 1987, 91, 4733.
(9) Bartocci, G.; Masetti, F.; Mazzucato, U.; Spalletti, A.; Orlandi, G.;
Poggi, G. J. Chem. Soc., Faraday Trans. 2 1988, 84, 385.
(10) Galiazzo, G.; Spalletti, A.; Elisei, F.; Gennari, G. Gazz. Chim. Ital.
1989, 119, 277.
S0002-7863(96)01859-8 CCC: $12.00 © 1996 American Chemical Society

10880 J. Am. Chem. Soc., Vol. 118, No. 44, 1996
Aloisi et al.
pulse).16,17 The effect of temperature on the decay rate constant was
determined employing a variable-temperature cell similar to that
designed by Wardman.18 A deconvolution program was used to obtain
singlet lifetimes from this setup.
Typical statistical errors in the lifetimes are (20% (for τ < 1 ns)
and (10% (for longer-lived singlets). To avoid photodegradation, the
solutions were circulated in a flow through cell.
was not the only reactive one. However, the S1 f T1
intersystem crossing (ISC) quantum yield remained very small
even at low temperature, and T1 could not account for all the
reactive quanta. A reactive upper state Sn or Tn has been
invoked to explain such behavior, but this explanation is not
without question since neither the fluorescence nor the T1
absorption increased in rigid matrices when the isomerization
was inhibited.
Corrected emission spectra were recorded by a Spex fluorimeter
(Fluorolog 112). The fluorescence quantum yields were measured
A somewhat clearer case is represented by 2-StPh,8,11-13 for
which a prevailing triplet mechanism has been hypothesized.
The fact that φF was independent of temperature indicated that
an activated twisting is not operative in S1. On the other hand,
T1 was not detectable by direct excitation in fluid solutions.
The T1 spectrum became clearly observable only in rigid
matrices at low temperatures, where isomerization was hindered.
This behavior was tentatively attributed to fast twisting in an
upper excited triplet state. However, the T1 yield at low
temperature should be known to have a complete picture of the
relaxation mechanism and to compare the ISC efficiency with
the quanta involved in the isomerization at room temperature.
A similar behavior was observed for 1- and 3-StPh but only in
a polar solvent.12
(estimated error (5%) as a function of the excitation wavelength (λexc
)
using anthracene in ethanol as reference (φF ) 0.27).19 For each
measurement, fresh deareated solutions with absorbance at λexc below
0.1 were used.
The fluorescence lifetimes, τF, were measured by a Spex Fluorolog-
τ2 system, which uses the phase-modulation technique (excitation
wavelength modulated in the 0.5-330 MHz range; time resolution ca.
10 ps). To avoid photodegradation of the samples, the intensity of the
excitation light was reduced by neutral filters, and, therefore, a broad
band of the emission light (selected by a cut-off filter) was collected
to increase the signal intensity. The frequency-domain intensity decays
(phase angle and modulation Vs frequency) were analyzed with the
Globals Unlimited (rev. 3) global analysis software.20
Trans f cis photoisomerization quantum yields (φtfc) were measured
(mean deviation of at least three independent experiments, ap-
proximately 7%) at various λexc using the lamp of the fluorimeter Spex
Fluorolog 112 (bandwidth ca. 2 nm). The conversion percentage,
measured spectrophotometrically, was held below 10% to avoid
competition from the back photoreaction.
The object of the present paper is the study of the transient
absorption and the fluorescence decay in the picosecond region
of some widely studied styrylarenes, bearing a naphthyl,
phenanthryl, or anthryl group, in order to have direct information
about the possible role of higher excited states in their
photochemistry and photophysics. The photokinetic behavior
of the observed transients and the effects of temperature and
excitation energy provided interesting evidence about the
implication of higher excited states in the reactive deactivation
of some of these compounds.
All measurements were carried out at 22 ( 2 °C unless otherwise
indicated; the solutions of the trans-styrylarenes (concentrations ca.
10-4 M) were deaerated by bubbling with oxygen-free argon.
3. Results
3.1. Transient Absorption. 9-Styrylanthracene. Upon
picosecond laser excitation of 9-StAn, a structured absorption
spectrum, with λmax of ca. 430, 580, and 730 nm, was recorded
just after the laser pulse in solvents of different polarity and
viscosity as CH, MeCN, and EtG (see Table 1 and Figure 1, as
an example). The change of absorbance (∆A) decreased with
time at wavelengths longer than 525 nm but rose between 420
and 525 nm, with an isosbestic point at 525 nm. The spectrum
taken after 3.9 ns (curve 7 of Figure 1) shows a main maximum
at 460 nm (which is reminiscent of the T1 f Tn spectrum of
9-StAn recorded in previous works)10,21,26 and a low signal at
λ > 525 nm. The absorption maximum at about 650 nm,
recorded only in a polar solvent, showed that the cation radical
9-StAn•+ 21 was produced within a few nanoseconds.
2. Experimental Section
The trans and cis isomers of 9-styrylanthracene (9-StAn), n-
styrylnaphthalene (n-StN, with n ) 1 and 2), and n-styrylphenanthrene
(n-StPh, with n ) 1, 2, and 9) were the same as used for previous
works.10,14,15 Anthracene (Baker Chem. Co.) and phenanthrene (Ph,
Aldrich 99.5+%) were used without further purification.
The solvents (Fluka, ACS grade) acetone (Me2CO), acetonitrile
(MeCN), bromopropane (BrP), cyclohexane (CH), 1,2-dichloroethane
(DCE), ethanol (EtOH), ethylene glycol (EtG), ethyl acetate (EtAc),
1-hexanol (HeOH), methanol (MeOH), n-hexane (n-H), and propanol
(PrOH) were used without further purification.
The decay kinetics recorded at λ > 525 nm were described
quite well by a biexponential fitting with τS (the short-lived
component) < 1 ns and τL (the longer-lived component) g 2.6
ns. The inset of Figure 1 shows the decay kinetics at 575 nm
and the growth of ∆A at 480 nm, which are clearly coupled. In
fact, the two kinetics were well-fitted with the same rate
constants (differences within the experimental error) and cor-
relation coefficients larger than 0.99. The value of τS was
affected by the solvent on going from 0.77 ns in MCH to 0.33
ns in the viscous EtG; the solvent effect on τL was much smaller.
Picosecond laser flash photolysis studies were carried out with the
third (λexc ) 355 nm, pulse width 30 ps and energy < 5 mJ) and fourth
(λexc ) 266 nm, pulse width 30 ps and energy < 2 mJ) harmonics of
a mode-locked Nd:YAG laser system (Quantel YG571). The basic
features of the pump-probe double-diode array spectrography and the
single-wavelength kinetic setup have been already described.16 Unless
otherwise indicated, an excitation wavelength of 355 nm was used.
The “zero” time was assumed to be the time just after the laser pulse
when the maximum of singlet-singlet absorbance was reached. Decay
kinetics were recorded up to 4 ns after the laser pulse.
Nanosecond flash photolysis experiments were performed with a
Q-switched Nd:YAG laser, which was tripled and quadrupled to give
single 7-ns pulses at 355 and 266 nm, respectively (energy < 10 mJ/
(16) Logunov, S. L.; Rodgers, M. A. J. J. Phys. Chem. 1992, 96, 2915.
Shand, M. A.; Rodgers, M. A. J.; Webber, S. E. Chem. Phys. Lett. 1991,
177, 11.
(17) Firey, P. A.; Ford, W. E.; Sounic, J. R.; Kenney, M. E.; Rodgers,
M. A. J. J. Am. Chem. Soc. 1988, 110, 7626.
(18) Wardman, P. J. J. Phys. E. 1972, 5, 17.
(11) Elisei, F.; Aloisi, G. G.; Mazzucato, U. J. Phys. Chem. 1990, 94,
5818.
(12) Aloisi, G. G.; Elisei, F.; Go¨rner, H. J. Phys. Chem. 1991, 95, 4225.
(13) Elisei, F.; Aloisi, G. G.; Go¨rner, H. Res. Chem. Interm. 1995, 21,
725.
(14) Aloisi, G. G.; Mazzucato, U.; Spalletti, A.; Galiazzo, G. Z. Phys.
Chem. (Munich) 1982, 133, 107; Galiazzo, G.; Spalletti, A.; Bartocci, G.;
Aloisi, G. G. Gazz. Chim. Ital. 1986, 116, 705.
(19) Eaton, D. F. In Handbook of Organic Photochemistry; Scaiano, J.
C., Ed.; CRC Press: Boca Raton, FL, 1989; Vol. I, p 235.
(20) Beechem, J. M.; Gratton, E.; Ameloot, M.; Kutson, J. R.; Brand,
L. Fluorescence Spectroscopy: Vol. 1 Principles and Techniques; Lakowicz,
J. R., Ed.; Plenum Press: New York, 1988; and references therein.
(21) Go¨rner, H.; Elisei, F.; Aloisi, G. G. J. Chem. Soc., Faraday Trans.
1992, 88, 29.
(15) Galiazzo, G.; Bortolus, P. J. Photochem. 1973/74, 2, 361.

Excited State BehaVior of Diarylethenes
J. Am. Chem. Soc., Vol. 118, No. 44, 1996 10881
Table 1. Singlet-Singlet Absorption Maxima (λmax) and Singlet
Lifetimes (τS and τL) of Some DAEs in Different Solvents at Room
Temperature Together with Their Fluorescence Lifetimes (τF)
compd solvent
λ
max (nm)
τS (ns) τL (ns)
τF (ns)
9-StAn CH
e430, 585, 720
0.77
3.9
4.7
2.9
1.7
0.83
3.3
3.6a
MeCN <450, 580, 660, 730 0.67
EtG
1-StN CH
MeCN
EtG
2-StN n-H
MeCN
EtG
1-StPh CH
MeCN
2-StPh CH
MeCN
EtG
9-StPh CH
4.2b
430, 580, 730
515, 590, 770
505, 580, >700
505, 585, >700
495
0.33
0.59
0.33
0.11
0.15
0.14
0.16
1.8c
0.9d
4.5, >10 5.2, 27.6e
2.3, >10 3.2, 17.8d
2.0, >10
490, >700
500, >700
530, 610, 730
0.43 10f
8.5f
0.40 32f
0.35 26f
0.32 27f
0.32
0.30
0.13
10g
500, 550, 610, 675 0.42
<430, 570, 620
550, 600
<430, 570, 620
500, 705
510, 690
510, 695
475, 575, 650
7.0h
34g
26.2h
5.1
5.3g
2.0h
MeCN
2.0
3.6
EtG
n-H
Ph
56f
57.5i
a In methylcyclohexane, from ref 5. b From ref 21. c From ref 7.
d From ref 22. e From ref 23. f From nanosecond flash photolysis (λexc
) 355 nm). g In n-hexane, from ref 24. h From ref 12. i From ref 25.
Figure 2. Time-resolved absorption spectra of 1-StPh in MeCN at
“zero” time (1) and at 0.27 (2), 0.57 (3), and 1.8 (4) ns, 2-StPh in
MeCN at “zero” time (1) and at 0.22 (2), 0.57 (3), 1.3 (4), and 1.8 (5)
ns, and 9-StPh in MeCN at “zero” time (1) and at 0.17 (2), 0.57 (3),
1.3 (4), and 3.3 (5) ns (λexc ) 355 nm, Ar-saturated solutions).
In the case of 2-StN, a main maximum (whose position was
practically unaffected by the solvent) was present at about 500
nm together with a broad absorption band at λ > 600 nm. At
least three transients were detected for this compound. In fact,
the initial absorption, detected over the whole 450-700 nm
range, decayed with a high rate constant (τS of ca. 0.15 ns in
n-H, MeCN, and EtG) and was replaced by two longer-lived
transients. The best fits of the decay kinetics recorded at 500
and 670 nm in MeCN were obtained by monoexponential plus
residual (at 500 nm) and biexponential (670 nm) equations. In
fact, τL values larger than 10 ns (rest absorption with ∆A
constant over 2 ns) and of a few nanoseconds were obtained
monitoring the absorption in the 500 nm region and at λ > 600
nm, respectively (see Table 1).
n-Styrylphenanthrene. As shown for the previous com-
pounds, a strong short-lived absorption was recorded in a wide
wavelength range (see Figure 2) for n-StPh. This absorption
decayed by a first-order kinetics with lifetime τS shorter than 1
ns (see Table 1).
The spectrum recorded in CH at “zero” time for 1-StPh (curve
1 of Figure 2a) was rather broad with three nonresolved maxima
at about 530, 610, and 730 nm. In MeCN, the absorption was
shifted at shorter wavelengths (λmax of ca. 500, 550, 610, and
675 nm); τS did not change with the solvent (∼0.43 ns). The
longer-lived transient shown in Figure 2a (curve 4) replaced
the short-lived component and decayed by a first-order kinetics
of 10 ns (in CH) and 8.5 ns (in MeCN). These two lifetimes
were measured by nanosecond laser flash photolysis experiments
with λexc of 355 nm.
Figure 1. Time-resolved absorption spectra of 9-StAn in MeCN at
“zero” time (1) and at 0.57 (2), 1.3 (3), 1.9 (4), 2.6 (5), 3.3 (6), and 3.9
(7) ns (λexc ) 355 nm, air-equilibrated solution). Inset: kinetics recorded
at 575 (O) and 480 (4) nm.
n-Styrylnaphthalene. Rather different spectral and kinetic
properties were found for the transients of the two trans isomers
1- and 2-StN, produced by 355 nm excitation. Preliminary
measurements were recently reported elsewhere.27
In particular, Ar-saturated solutions of 1-StN, when irradiated
with ca. 30 ps pulses, showed an instantaneously formed
absorption spectrum (Table 1). The spectrum had three broad
absorption bands with maxima near 515, 590, and 770 nm. All
the bands decayed almost completely over 4 ns with biexpo-
nential decays (Table 1). The short-lived component τS was
strongly affected by the solvent, being 0.59 ns in the nonpolar
CH and 0.11 ns in the polar and viscous EtG. On the other
hand, the longer-lived component τL was much longer in EtG
(3.3 ns) than in the other solvents.
(22) Aloisi, G. G.; Elisei, F.; Latterini, L. J. Chem. Soc., Faraday Trans.
1992, 88, 2139.
(23) Mazzucato, U.; Momicchioli, F. Chem. ReV. 1991, 91, 1679 and
references cited therein.
(24) Aloisi, G. G.; Elisei, F. J. Phys. Chem. 1990, 94, 5813.
(25) Murov, S. L.; Carmichael, I.; Hug, G. L. Handbook of Photochem-
istry; Dekker: New York, 1993.
(26) Wismontsky-Knittel, T.; Das, P. K. J. Phys. Chem. 1984, 88, 1168.
Bhattacharyya, K.; Chattopadhyay, S. K.; Baral-Tosh, S.; Das, P. K. J. Phys.
Chem. 1986, 90, 2646. Go¨rner, H. J. Photochem. Photobiol., A: Chem.
1987, 43, 263.
The behavior of 2-StPh was similar to that of isomer 1. The
τS values (Table 1) did not change significantly with the position
of the styryl group on the phenanthrene moiety. The main
differences are due to the shape of the absorption spectra (λmax
of ca. 570 and 620 nm for the isomer 2) recorded after ps-
pulses (Figure 2b) and to the lifetime of the longer-lived transient
which was more than 26 ns (obtained by nanosecond flash
photolysis).
(27) Aloisi, G. G.; Elisei, F.; Mazzucato, U.; Latterini, L.; Rodgers, M.
A. J. Am. Inst. Phys. Conf. Proc. 1996, 364, 139.

10882 J. Am. Chem. Soc., Vol. 118, No. 44, 1996
Aloisi et al.
Table 2. Solvent Effect on the Short-Lived (τS) and Longer-Lived
(τL) Components of the Singlet Decay of 9-StPh
spectrum and to longer wavelengths (350-370 nm), closer to
the 0-0 transition. In all cases, two decay components (τ1 and
τ2) were detected, the fast one having small amplitudes of 1-4%
or less with the exception of 1-StN (A1 in the range 4-14%)
(Table 3).
solvent
ꢀ
η (mN s m-2
)
τS (ns)
τL (ns)
n-H
CH
EtAc
BrP
DCE
HeOH
PrOH
Me2CO
EtOH
MeOH
MeCN
EtG
1.89
2.02
6.02
8.09
10.36
13.3
20.33
20.70
24.55
32.70
37.5
0.313
0.98
0.37
0.32
0.71
0.11
0.67
0.24
0.18
0.77
0.23
0.18
0.30
0.13
5.0
5.1
3.0
0.8
1.8
1.6
0.9
5.0
0.9
1.2
2.0
3.6
0.455
0.539
0.887
4.592
2.237
0.337
1.078
0.544
0.375
21
The frequency responses (phase and modulation) recorded
for 2-StPh and 9-StPh in MeCN are shown, as examples, in
Figures 4a and 5a, respectively. Treatments of the experimental
data with a global analysis method20 gave the results summarized
in Table 3 for mono- and biexponential fitting functions. The
analysis with a triexponential function (results not shown) gave
one pre-exponential factor that was practically zero and/or two
decay parameters that were very close to each other. ø2
parameters greater than those given by biexponential fittings
were obtained in all cases.
38.66
Generally, the monoexponential functions gave higher ø2
parameters (Table 3) and worse residual distributions (see
Figures 4b,c and 5b,c as examples) than the corresponding
biexponential functions. In a few cases (1-StN in MeCN and
9-StPh in CH), at λexc around 350 nm, the fitting parameter did
not improve by using a biexponential treatment thus indicating
the absence of the short-lived component at low excitation
energies. In the other cases, where two decay components were
clearly observable, the amplitude of the short component
generally increased with the excitation energy.
The longer-lived fluorescence component was found to be
quenched by oxygen at nearly diffusion-controlled rate. Obvi-
ously, the short component was unaffected by the presence of
oxygen.
Figure 3. Time-resolved absorption spectra of Ph in n-hexane at “zero”
time (λexc ) 266 nm, Ar-saturated solution). Inset: decay kinetics
recorded at 570 nm.
The longer component was in reasonable agreement with the
τF values (Table 1) previously measured by the single-photon
counting method by monoexponential deconvolution. In that
case, the time resolution (ca. 0.3 ns) and the different method
of analysis, which had to take into account the source profile,
did not allow the short component (with small amplitude) to
be detected. In fact, the monoexponential fitting led to
nonhomogeneous residual distribution in the short time range,
while the biexponential function did not reach convergence.
Picosecond-laser excitation of 9-StPh was performed in 12
solvents of different polarity and viscosity. In all cases, the
spectrum showed two maxima (Figure 2c) at about 510 and
700 nm (Table 1). As obtained for the other compounds, two
transients were produced in all the solvents. No clear trends
of the measured lifetimes with the properties of the medium
were found. The τS and τL values obtained by biexponential
fittings are collected in Tables 1 and 2. The shortest values of
τS were obtained in BrP and in the viscous EtG.
Phenanthrene. Picosecond excitation of Ph itself was
performed in n-H in order to compare the behavior of the
flexible styrylarenes described above with that of a polycyclic
hydrocarbon, which does not bear the styryl group, under the
same experimental conditions.
The results obtained are shown in Table 1 (last line) and
Figure 3. The singlet-singlet absorption spectrum, recorded
at “zero” time, shows three absorption bands (λmax ) 475, 575,
and 650 nm) whose positions and intensities remained practically
unchanged in the time-window accessible with our setup (4 ns).
The kinetic profile, independent of the wavelength of analysis,
is shown in the inset of Figure 3. For this compound, no
evidence was found of short-lived transients over the whole
wavelength range investigated (450-700 nm). A lifetime τL
of 56 ns was determined by nanosecond laser flash photolysis.
A comparison between τL and the fluorescence lifetime (τF )
57.5 ns)25 of Ph in a nonpolar solvent, indicates that the
absorption shown in Figure 3 is due to S1 f Sn transitions.
3.2. Fluorescence Lifetimes. To investigate the fluores-
cence lifetimes of DAEs in the subnanosecond time scale, the
emission decays were recorded with a time resolution of ca.10
ps (phase-shift method). Deaerated solutions of DAEs in CH
and MeCN were excited at two wavelengths corresponding to
the bathochromic maximum (300-330 nm) of the absorption
The values of the fluorescence lifetimes (τ1 and τ2) obtained
in this work can be compared with the lifetimes τS and τL
measured by singlet-singlet absorption (Tables 1 and 2). If
one takes into account the experimental error (larger in τ1 due
to its low pre-exponential factor), the lifetimes of the emission
and those of the decay of the singlet-singlet absorption appear
to be very close. The decay component in the subnanosecond
time scale was not detected by fluorescence measurements for
2-StN. However, the system is here more complicated showing
two-component decays with lifetimes longer than 3 ns (due to
the emission of two conformers23). It must be noted that the
large errors in τ1 and τ2, obtained by excitation of 2-StN at 320
nm in both solvents, seem to indicate that another component
was present, but the triexponential function could not improve
the fitting, probably due to the very small pre-exponential factor
of the third component.
Frequency-domain intensity decays of anthracene (in CH and
MeCN) and phenanthrene (in MeCN) were also recorded. In
these cases, the best fitting of the frequency responses was
obtained with a monoexponential equation with lifetimes not
affected by λexc (for anthracene, τF ) 5.1 and 4.7 ns in CH and
MeCN, respectively, and for phenanthrene, τF ) 54 ns in
MeCN) and close to the literature data.25
3.3. Effects of Excitation Wavelength and Laser Energy.
The emission measurements described in Section 3.2 showed
that the lifetimes τ1 and τ2 were independent of λexc, while the

Excited State BehaVior of Diarylethenes
J. Am. Chem. Soc., Vol. 118, No. 44, 1996 10883
Table 3. Mono- and Biexponential Analysis of DAE Fluorescence by the Phase Shift Methoda
c,d
compd
9-StAn
solvent
CH
λ
exc (nm)
τb (ns)
ø2 b
1.9
τ1c (ns)
τ2c (ns)
A1
ø2 c
330
370
330
370
302
360
302
355
320
350
320
350
320
360
320
360
320
360
325
360
300
350
325
370
3.5 ( 0.12
3.56 ( 0.04
3.88 ( 0.12
4.00 ( 0.09
1.67 ( 0.04
1.64 ( 0.01
0.68 ( 0.02
0.62 ( 0.01
15.0 ( 2.8
5.6 ( 0.9
0.68 ( 0.3
0.43 ( 0.20
0.69 ( 0.30
0.35 ( 0.20
0.62 ( 0.16
0.41 ( 0.15
0.32 ( 0.05
0.62 ( 1.2
4.8 ( 1.0
3.48 ( 0.03
3.59 ( 0.06
4.05 ( 0.07
4.10 ( 0.05
1.82 ( 0.05
1.75 ( 0.03
0.76 ( 0.02
0.62 ( 1.2
24.0 ( 3.0
22.3 ( 0.7
16.3 ( 2.0
18.1 ( 0.8
8.5 ( 0.05
8.5 ( 0.08
7.0 ( 0.3
e0.0001
0.0044
0.032
0.015
0.097
0.041
0.14
0.06
0.27
0.70
0.15
1.2
0.98
0.7
0.82
1.7
1.05
1.2
2.9
6.6
0.8
2.9
2.7
1.1
1.18
6.0
0.67
1.3
1.3
1.2
1.0
0.25
1.8
1.2
1.8
1.3
1.6
2.6
8.4
3.7
1.8
2.7
20.0
122
7.9
104
2.6
13.3
7.3
4.3
1.9
25.4
10
MeCN
CH
1-StN
2-StN
1-StPh
2-StPh
9-StPh
MeCN
CH
3.76 ( 0.05
3.1 ( 0.4
MeCN
CH
10.2 ( 1.9
5.7 ( 1.1
3.1 ( 0.08
0.30 ( 0.6
0.21 ( 0.09
0.83 ( 0.22
0.78 ( 0.15
0.30 ( 0.15
0.23 ( 0.18
0.49 ( 0.20
0.45 ( 0.14
0.40 ( 0.30
3.6 ( 3.4
0.58
8.45 ( 0.06
8.38 ( 0.43
6.9 ( 0.18
6.85 ( 0.14
33.7 ( 0.54
32.6 ( 1.2
29.2 ( 0.57
29.0 ( 0.9
4.08 ( 0.05
3.95 ( 0.03
1.77 ( 0.06
1.63 ( 0.03
0.001
0.016
0.02
MeCN
CH
7.0 ( 0.05
33.7 ( 0.4
32.9 ( 0.4
30.0 ( 0.2
29.2 ( 0.2
4.09 ( 0.03
4.3 ( 4.3
0.02
0.001
0.005
0.006
0.006
0.0006
0.47
MeCN
CH
7.3
2.2
1.8
5.7
4.2
MeCN
0.31 ( 0.09
0.33 ( 0.16
1.89 ( 0.03
1.69 ( 0.02
0.04
0.04
a Broad band emission. b Obtained from monoexponential analysis. c Obtained from biexponential analysis. d Amplitude of the component with
decay time τ1 determined by a nonlinear least-squares fitting.
Figure 4. Frequency response of 2-StPh in MeCN fitted by monoex-
Figure 5. Frequency response of 9-StPh in MeCN fitted by mono-
ponential (- -) and biexponential (s) functions (a) together with the
exponential (- -) and biexponential (s) functions (a) together with the
residual distribution for monoexponential (b) and biexponential (c)
residual distributions for monoexponential (b) and biexponential (c)
fittings (λexc ) 355 nm, broad band emission).
fittings (λexc ) 355 nm, broad band emission).
contribution of the short-lived component decreased with the
increase of λexc
compiled in Table 4, show that for 9-StAn and 9-StPh the φF
values were practically constant (within the experimental error),
while a marked effect was found for the other two compounds.
In fact, for 1-StN and 2-StPh, φF increased with λexc, the highest
values (φF ) 0.85 and 0.50 for 1-StN and 2-StPh, respectively)
being obtained using an excitation energy close to the 0-0
transition. The same investigation was not performed with
1-StPh and 2-StN, because the conformational equilibrium at
room temperature between two different rotamers23 could have
hidden such an effect.
.
Absorption picosecond laser experiments on 2-StPh in MeCN
were carried out to check the effect of λexc on the transients
produced. In fact, both the short- and longer-lived components
of the decay were detected upon excitation at 266 and 355 nm.
Similar τS values were obtained in the two experiments.
The presence of the short-lived component was also checked
at different laser intensities for some DAEs. The ∆A values
recorded at 550 nm (maximum of the singlet-singlet absorption)
obtained by irradiating an oxygen-free solution of 9-StPh in
n-H (λexc ) 355 nm) are plotted in Figure 6 versus the laser
dose. The linear trend, over a wide range of laser intensities,
which passes through zero, is typical of transients produced
monophotonically upon light absorption.
The effect of λexc on φtfc of 1-StN and 2-StPh was also
investigated (Table 4). The yield of 1-StN decreased markedly
when λexc increased from 320 (φtfc ) 0.16) to 365 nm (φtfc
)
0.07). A similar, but less evident, effect was found for 2-StPh,
for which φtfc values between 0.30 and 0.25 were measured in
the 280-370 nm range.
The effect of λexc was also investigated by steady-state
experiments. Fluorescence quantum yields (φF) of four DAEs
in n-H were measured in a wide range of λexc. The results,
In previous papers, it has been reported that, contrary to the
case of 2-StN, the fluorescence quantum yield of 1-StN did not

10884 J. Am. Chem. Soc., Vol. 118, No. 44, 1996
Aloisi et al.
Figure 6. Effect of laser dose on the absorbance change at 550 nm of
9-StPh in n-hexane at “zero” time (λexc ) 355 nm, Ar-saturated
solution).
Figure 8. Temperature effect on the relative change of absorbance
(∆A/∆Amax) at 600 nm (extrapolated at “zero” time) of 2-StPh in
n-hexane (O) and MeCN (4), on the longer lifetime τL (f) in n-hexane,
and on the relative φF (b) φtfc ([) values in methylcyclohexane (taken
from ref 8).
Table 4. Effect of the Excitation Wavelength (λexc, nm) on the
Fluorescence Quantum Yield (φF) and on the Trans f Cis
Photoisomerization Quantum Yield (φtfc) of DAEs in n-H
behaviors were recorded: (i) τS of 9-StAn was practically
constant over the temperature range investigated, (ii) for 1-StN
there was an increase of τS with the temperature (typical
behavior expected in the presence of an activated decay process),
and (iii) for the two StPh isomers, a decrease of the decay rate
constant was obtained on increasing the temperature between
zero and 25 °C, while above room temperature, τS remained
practically constant. This atypical behavior could not be
followed below 273 K owing to the limits of the picosecond
setup.
9-StAn
λexc φF
1-StN
2-StPh
9-StPh
φtfc λexc φF
λexc
φF
φtfc λexc
φF
-1
370 0.49 300 0.70
384 0.49 310 0.70
400 0.49 320 0.69 0.16 300 0.39
420 0.48 340 0.70 0.15 310 0.42
430 0.48 350 0.77 0.10 325
360 0.80 0.08 330 0.42
280 0.32 0.30 293 0.62
290 0.37
300 0.65
310 0.65
320 0.64
0.27 340 0.64
350 0.65
0.28 360 0.70
365 0.83 0.07 340
370 0.85
355 0.45 0.24 370 0.67
372 0.50 0.25
For 2-StPh, whose fluorescence lifetime is rather long (tenths
of nanoseconds), nanosecond flash photolysis experiments
allowed the relative population of the longer-lived transient
(measured as relative ∆AL at 600 nm) to be followed at different
temperatures (over a -75 ÷ 50 °C range). Figure 8 shows the
relative absorbance change (∆AL) obtained in the nonpolar n-H
and in the polar MeCN. The two, practically identical, trends
look like titration curves with a plateau below -25 °C
(maximum concentration of the transient), an inflection point
at about 0 °C, and another plateau above 25 °C (minimum
concentration of the transient). This behavior can be compared
with those of fluorescence and trans f cis photoisomerization
quantum yields of 2-StPh in methylcyclohexane8 shown in the
same figure. The effect of temperature on φF is reminiscent of
that on ∆AL, while φtfc increases with temperature over -60
÷ 25 °C and remains practically constant at higher values of
temperature.8 Figure 8 also shows the τL values obtained over
the same temperature range; since they are constant within the
experimental error, no activated decay processes are expected
for this longer-lived transient.
Figure 7. Temperature effect on the lifetime (τS) of the short-lived
component of the singlet-singlet absorptions of 1-StN (O), 9-StAn ([),
2-StPh (4), and 9-StPh (f) in Ar-saturated MeCN.
vary with the frequency of the exciting light.6,23,28 This was
taken as a demonstration of a shift of the ground state
conformational equilibrium toward the more stable rotamer of
1-StN. As a matter of fact, the λexc effect had been explored in
our laboratory in the 270-340 nm range. Therefore, our
measurements failed to show the increase of φF now observed
in the 340-370 nm range, a trend which is not consistent with
the presence of rotamers.
3.4. Temperature Effect. To check the presence of
activated decay processes in the singlet manifold, τS of four
DAEs (9-StAn, 1-StN, 2-StPh, and 9-StPh) was measured in
MeCN in the 0-70 °C temperature range; the results obtained
are summarized in Figure 7 where the plots of the decay rate
constants (τS-1) Vs temperature are shown. Three different
4. Discussion
4.1. Transient Assignment. All the DAEs investigated
displayed a biexponential decay of both the singlet-singlet
absorption transients produced by short-pulse laser excitation
(lifetimes τS and τL) and the fluorescence emission (lifetimes
τ1 and τ2).
The lifetimes, measured by the two techniques, coincide
within the experimental error indicating that they originate from
the same electronic states. Only in the case of 2-StN, the fast
component was not detected by fluorimetry, the two long-lived
components being assigned to the known s-trans and s-cis
rotamers of this molecule.6,23
While the similarity of the longer-lived components, τL and
τ2, to the known lifetimes measured by nanosecond fluorimetry
(28) Fischer, E. J. Phys. Chem. 1981, 85, 1770.

Excited State BehaVior of Diarylethenes
J. Am. Chem. Soc., Vol. 118, No. 44, 1996 10885
(τF in Table 1) clearly indicates that they refer to the S1 f Sn
and S1 f S0 transitions, respectively, the assignment of the
shorter ones, τS and τ1, is less straightforward and not discussed
in the literature for these stilbene-like molecules.
Other possible explanations for the short decay component
were considered.27 One could think, e.g., of τS reflecting the
rate of equilibration of the upper state toward S1. However,
this hypothesis seems to be ruled out by the largely different
ratios of the pre-exponential factors of the absorption (ca. 1)
and emission (generally, < 0.05) decays. In case of an
equilibration, one would expect to observe only one state and
then expect a ratio depending only on the equilibrium rate and
independent of the detection technique. In fact, the large
difference in the measured ratios should reflect the difference
in the extinction coefficients (absorption) and fluorescence yields
(emission) of the two states. A similar explanation could be
offered by the relaxation of the Franck-Condon (transoid) S1
state towards more stable twisted conformations. However, the
fact that the shape of the fluorescence spectra is independent
of λexc and temperature seems to rule out such hypothesis too.
Other possible assignments, such as transitions from the 1perp*
configuration, must be ruled out since the precursor twisting
from S1 (1trans* f 1perp*) would be too slow to be operative
during the laser pulse. Moreover, it would not be expected for
9-StAn, because it does not isomerize at room temperature. On
the other hand, the occurrence of some kind of solvent
rearrangement is expected to be faster than the lifetime of the
observed transient.
An important support for the assignment of the short-lived
component (τS) observed by absorption measurements to the
S2 state came from the decay component (with an amplitude,
A1, generally smaller than 4%) measured in emission with
practically the same lifetime (τ1). Another evidence for all the
DAEs investigated is the general increase of A1 on increasing
the excitation energy (see Table 3).
The reliability of the present assignment of the short-lived
component seems to receive further support by theoretical
calculations in progress.32 In fact, a perturbative treatment,
based on vibronic coupling, shows that the two lowest excited
states of 1-StN and other related molecules are mixed and that
they have noncommunicating decay pathways, characterized by
different rates. The fast component could remain hidden in the
emission spectrum being the minor one, as indicated by the small
amplitudes measured in the fluorescence decay.
The unexpected long lifetime assigned to S2 deserves some
comments. It has to be noted that the rate constant of the S2 f
S1 internal conversion (kIC) depends on several factors including
the energy gap, the vibronic coupling, the vibrational density
of the destination state and the Franck-Condon factor. When
the energy gap (∆E) is high (e.g., 20 000-30 000 cm-1), kIC
depends mainly on the energetic factor and increases when ∆E
decreases. However, when the two states are close (∆E < 3000
cm-1), kIC depends strongly on the vibronic coupling also. If
the nature of the two states is rather different, the adiabatic
coupling between S2 and S1 decreases. In the meantime, the
vibrational density of the destination state decreases by many
orders of magnitude. In the literature, there are several examples
of upper excited states, with lifetimes up to hundreds of
picoseconds, having rather slow IC when the two states are very
close.33 Quantum-mechanical (semiempirical) calculations on
our compounds have shown that the nature of S2 and S1 is rather
different (one state is more ethenic and the other more arylic,
depending on the compound5,6,32), and, consequently, the kIC
value predicted is in agreement with the long lifetime measured
for S2 in the present work.
The two decay components cannot be due (with the exception
of the fluorimetric measurements on 2-StN) to the presence of
s-trans-s-cis rotamers since for some compounds (9-StPh,
9-StAn) no rotamerism is expected and for others (1-StN,
1-StPh) the rotamer equilibrium is expected to be almost
completely shifted toward the more stable species.23 Moreover,
previous studies of the conformational equilibria of DAEs,
carried out in our and other laboratories, have shown that the
S1 lifetimes of the existing rotamers are generally in the
nanosecond time scale and relatively close one to each other.23
This does not exclude that conformations more or less deviating
from planarity (due to small angles of twisting about the ethenic
bridge or the 1-R bond between the ethenic carbons and the
aryl groups) may play a role in the relaxation processes of the
lowest excited singlet states, mainly affecting the internal
conversion (IC) rate (see below).
The fast decay component is more reasonably assigned to an
upper excited state which lives long enough to relax through
decay channels competitive with IC to S1.
Indications of a role of upper states in the photoisomerization
of trans-stilbene itself had been reported in the past by K. Fuke
et al. on the basis of the results obtained by using two-photon
fluorescence excitation and thermal lensing methods.29 Another
indication was previously reported by some of us for a similar
flexible molecule, trans-2-StAn.30 In addition to the dual slow
fluorescence, assigned to two equilibrated conformers, a third
emission component was detected for this anthryl derivative
above 200 K at λexc < 410 nm and assigned to an anti-Kasha
emission from a thermally populated S2 state.
Theoretical calculations6,8,9 have shown that these flexible
DAEs generally have the S2 state closely located to the lowest
fluorescent state S1, and, in fact, the fast transient was not
detected in the absence of the flexible styryl substituent (see
sections 3.1 and 3.2 for phenanthrene and anthracene).
Discussing first the absorption results, the fast transient
detected in the present work was assigned to a transition from
an upper excited state, probably implying the second excited
singlet (S2 f Sn), based on the following: (i) it was produced,
monophotonically, within the laser pulse; (ii) it decayed by a
first order decay kinetics; (iii) it was observed for all the DAEs
investigated at different λexc and in a number of solvents; (iV)
its lifetime is practically equal to the short-lived component of
the fluorescence decay; and (V) like S1, it interacted efficiently
with concentrated additives such as heavy-atom-containing
compounds (e.g., using BP as solvent) or electron donors or
acceptors, thus inducing the formation of triplet states or radical
ions, respectively.27,31
This assignment could appear not consistent with the small
changes observed for the singlet-singlet absorption spectra with
the delay time; anyway, it seems to us reasonable considering
the probable strong mixing between the two states (see below).
It is to be noted that, as already observed for n-StN,27 the
transient spectrum (a pure electronic property) depends linearly
on the oscillator strength of the transitions from the composite
states to Sn, while the dynamic properties depend exponentially
on the vibronic rearrangement of the two states during their
relaxations.
4.2. Decay Channels from S2 and S1. The competition of
(29) Fuke, K.; Sakamoto, S.; Ueda, M.; Itoh, M. Chem. Phys. Lett. 1980,
74, 546.
(30) Bartocci, G.; Mazzucato, U.; Spalletti, A.; Elisei, F. Spectrochimica
Acta 1990, 46A, 413.
(32) Marconi, G. Private communication.
(33) See, e.g.: Ferrante, C.; Kensy, U.; Dick, B. J. Phys. Chem. 1993,
97, 13457. Sahyun, M. R. V.; Sharma, D. K. Chem. Phys. Lett. 1992, 189,
571.
(31) Latterini, L. et al. Unpublished results.

10886 J. Am. Chem. Soc., Vol. 118, No. 44, 1996
Aloisi et al.
the relaxation processes from the S1 state of DAEs (fluorescence,
ISC, isomerization) has been described in previous papers from
our and other laboratories.1-9 In this discussion more attention
will be paid to the decay from the upper excited singlet state.
The dominant relaxation process seems to be very dependent
on the nature of the DAE.
Intersystem crossing to the triplet manifold, starting from S2,
was evidenced for the anthryl derivative, 9-StAn. This is shown
in Figure 1, where the biexponential decay kinetics of S1 and
S2 (precursors) correspond to the rise of the band at 460 nm,
assigned to the T1 f Tn absorption.10,21,26 The absence of
isomerization of 9-StAn, ascribed to high torsional barriers in
both the singlet and triplet manifolds,9 was here confirmed by
the constancy of the fluorescence quantum yield with λexc (Table
4) and of the lifetime τS with the temperature (Figure 7), which
indicates the absence of activated relaxation channels.
For the naphthyl derivatives, which have a substantial yield
of isomerization to cis but a negligible triplet yield,1-3,6,7 the
decay of S2 is generally faster than for 9-StAn (Tables 1 and
3), indicating a possible intervention of twisting already in the
upper state. Moreover, for these derivatives, which have been
reported to exist in two ground-state isomeric conformations,23
some interesting information was obtained from the longer-lived
transient. In the case of 1-StN, for which the equilibrium
between the two rotamers is markedly shifted toward the more
stable rotamer,6,23,28,34 only one slow decay component, inde-
pendent of the monitoring wavelength, was observed. This is
in agreement with the results of the fluorimetric measurements
which indicated only one rotamer in fluid solutions over a large
temperature range. On the other hand, two different, longer-
lived decay components, in addition to the fast one, were
observed in the transient absorption of 2-StN when monitoring
at 450 nm or at longer wavelengths. These transients were
assigned to the two rotational isomers, already described in
several papers.6,23,27,28,34,35 For 1-StN, the short lifetime τS
decreased with the increase of temperature, as reported for the
longer τF value,6,28,34 indicating activated twisting leading to
the cis isomer in both S2 and S1 states.
The transients of the three phenanthryl derivatives investigated
absorb in the same spectral region but display different behavior.
The solvent effect, particularly investigated for 9-StPh (Table
2), showed no clear trend with polarity and viscosity. The
shortest lifetimes for both transients were measured in BrP (as
expected, due to the favored spin-orbit coupling induced by the
heavy atom, which favors ISC in both excited singlet states),
in alcoholic solvents (probable implication of H-bonded struc-
tures which favor planar or quasi-planar conformations and their
radiationless deactivation by IC), and in EtG (possibly due to
fast IC from specific conformations favored in the viscous
solvent, see below).
It is not easy to explain the unusual temperature effect found
on 2-StPh and 9-StPh in MeCN (Figure 7), whose lifetime
τS increased with temperature from 0 to 25 °C remaining
practically constant at higher temperatures. It seems that at least
a second distinct decay process, probably IC, is influenced by
temperature in addition to the activated trans* f perp* rotation.
In fact, one cannot exclude that temperature modifies the
conformational distribution of these flexible molecules with a
subsequent change of the kIC value. Therefore, the increase in
the short lifetime with temperature can be considered the result
of two opposite effects, namely a slower IC process, dominant
for both compounds, and a faster twisting in S2, which is evident
only in the case of 2-StPh since the isomer 9 is known to
isomerize in the lowest S1 and T1 states.8,11,12 This tentative
explanation of the role of twisted geometries is speculative but
reasonable considering that such conformations can strongly
affect the vibronic coupling between S1 and S2 and then the
rate of internal conversion. A similar explanation could apply
to the effect of the viscous solvent cited above. It is to be noted
that these two phenanthrene derivatives are expected to isomer-
ize at a low rate from S1 due to the high torsional barrier
measured by fluorimetry.8 Moreover, isomer 2 has the lowest
state of Lb character,8 weakly coupled with the ethenic Bu
state, and is therefore expected to display an even lower
reactivity.
Figure 8, which refers to 2-StPh, offers further indications
on the behavior of this DAE. First of all, it shows that τL
remains constant indicating that no activated channels are
operative in S1, which can therefore be considered as a
nonisomerizable state. In the same temperature range, the trend
of the optical density of the longer-lived transient in n-H and
MeCN, practically the same in both solvents, can be compared
with the trends of the fluorescence and photoisomerization
quantum yields in methylcyclohexane. These trends of the
quantum yields and transient absorptions are in very close
agreement each other, considering the different solvent and
techniques involved. They show that from room temperature
upward the minimum population of S1 corresponds to the
minimum emission yield and maximum reactivity, while the
opposite situation is observed from -25 °C downward; the two
limiting situations are connected by an inflection of the various
parameters in the -25 to +25 °C range. Taken together, these
observations indicate that S2 is the state that is mainly
responsible for isomerization of 2-StPh, either by direct twisting
to the 1perp* configurations in the singlet manifold or through
ISC followed by twisting in the triplet manifold. Both decay
processes of S2 are probably operative. In fact, a contribution
of an upper triplet has been indicated by previous flash
photolytic measurements mentioned in the Introduction.11,12,24
However, considering that the T1 f Tn spectrum was readily
observable in fluid solution in the presence of perturbers (triplet
donor sensitizers or heavy-atom containing compounds), but was
not detectable by direct excitation, a contribution, at least partial,
For 1-StN and 2-StPh, an increase of φF with λexc was
observed, the maximum yield being obtained when the excitation
approached the 0-0 transition (Table 4). A parallel decrease
of φtfc with λexc was found for both compounds, particularly
evident for the naphthyl derivative. In the singlet manifold,
the diabatic isomerization mechanism accepted for these DAEs1-9
consists in the 1trans* f 1perp* twisting followed by IC to the
ground state 1perp and then partitioning (ca. 50:50) to the ground
state trans and cis isomers. On this basis, one can get
information on the nonradiative and nonreactive deactivations
by the difference [1 - (φF + 2×φtfc)]. For the compounds
listed in Table 4, this difference was never found to be higher
than 0.08, which indicates, taking into account the experimental
uncertainty, that the IC to the ground state is generally
negligible. The radiative quanta lost on exciting at an energy
level higher than that corresponding to the 0-0 transition are
then recovered at the upper level through the reactive pathway.
One can conclude that excitation to upper excited singlet states
of molecules whose twisting in S1 is highly activated can
increase the isomerization yields through a faster, less-activated,
twisting in S2.
(34) Haas, E.; Fischer, G.; Fischer, E. J. Phys. Chem. 1978, 82, 1638.
Fischer, E. J. Photochem. 1981, 17, 331.
(35) Alfimov, M. V.; Shenck, Yu. B.; Kovalenko, N. P. Chem. Phys.
Lett. 1976, 43, 154 and references therein. Matthews, A. C.; Sakurovs, R.;
Ghiggino, K. P. J. Photochem. 1982, 19, 235 and references therein.
Bartocci, G.; Mazzucato, U.; Masetti, F.; Aloisi, G. G. Chem. Phys. 1986,
101, 461 and references therein. Saltiel, J.; Sears, D. F., Jr.; Choi, J.-O.;
Sun, Y.-P.; Eaker, D. W. J. Phys.Chem. 1994, 98, 35.

Excited State BehaVior of Diarylethenes
J. Am. Chem. Soc., Vol. 118, No. 44, 1996 10887
of twisting in S2 itself can now be considered as a more
convincing explanation.
For 9-StPh in n-H, the negligible effect of λexc shown in Table
4 is in agreement with a trans* f perp* rotation in the T1 state,
as pointed out from previous works.8,11,12
This phenomenon appears to be more common than imagined
so far. It falls also into line with the results of the two-photon
fluorescence-excitation experiments on stilbene29 mentioned in
section 4.1, which showed the existence of a λexc-dependence
of φF assigned to the presence of a fast decay channel from an
upper singlet state, namely a barrierless twisting to the non-
fluorescent 1perp* state. Unlike the case of the twisting in the
S1 state, this process in the upper state could be fast enough to
compete with the vertical internal conversion to S1.
5. Conclusions
The results of the absorption and emission measurements in
the picosecond region allowed the decay kinetics of the upper
excited singlet state of the DAEs investigated to be spectro-
scopically detected. The role played by this upper state in the
reactive and radiationless relaxations was demonstrated for some
compounds, on the basis of temperature and λexc effects.
Depending on the nature of the polycyclic aryl group of the
DAE, a fast radiative relaxation of the upper singlet to the
ground state (2-StAn, previously reported30) or its fast nonra-
diative relaxation toward the triplet manifold (9-StAn) or toward
the isomerization channel (1-StN and 2-StPh) can compete with
the IC pathways to S1 and S0.
Acknowledgment. This work was supported by the “Con-
siglio Nazionale delle Ricerche” and by the “Ministero per
l’Universita` e la Ricerca Scientifica e Tecnologica”. One of
us (L.L.) thanks the Center for Photochemical Sciences of the
Bowling Green State University for a grant during her stay there
and Dr. S. Logunov for his assistance in the absorption
picosecond measurements.
JA961859C