The photochemistry of trans-ortho-, -meta-, and -para-aminostilbenes

Article abstract of DOI:10.1021/ja992335q

The photochemical behavior of the three positional isomers of trans-aminostilbene is reported and compared to that for unsubstituted trans-stilbene. The absorption spectrum of the para isomer displays a single intense long-wavelength band; however the met

Full text of DOI:10.1021/ja992335q

J. Am. Chem. Soc. 1999, 121, 12045-12053
12045
The Photochemistry of trans-ortho-, -meta-, and -para-Aminostilbenes
Frederick D. Lewis,* Rajdeep S. Kalgutkar, and Jye-Shane Yang†
Contribution from the Department of Chemistry, Northwestern UniVersity, 2145 Sheridan Road,
EVanston, Illinois 60208
ReceiVed July 6, 1999. ReVised Manuscript ReceiVed October 15, 1999
Abstract: The photochemical behavior of the three positional isomers of trans-aminostilbene is reported and
compared to that for unsubstituted trans-stilbene. The absorption spectrum of the para isomer displays a single
intense long-wavelength band; however the meta and ortho isomers display two less intense bands as a
consequence of configuration interaction. All three isomers are fluorescent and display similar solvent-induced
shifts of their fluorescence maxima. The fluorescence rate constant of the para isomer is larger than that of the
ortho or meta isomer. The para isomer has a short singlet lifetime and high photoisomerization quantum yield,
similar to those for trans-stilbene. In contrast, the ortho and meta isomers have long singlet lifetimes and low
photoisomerization quantum yields. Isomerization of the para isomer is a singlet-state process, whereas
isomerization of both the ortho and meta isomers is a triplet-state process. The lower bound for the barrier for
singlet-state torsion of the ortho and meta isomers is 7 kcal/mol. The barrier height is found to be determined
by the effects of the amino substituent upon the relative energies of the fluorescent singlet and twisted singlet
states.
Introduction
trans-3-aminostilbenes display long singlet lifetimes and large
fluorescence quantum yields. The behavior of meta- and para-
substituted stilbenes is in most cases similar; however, a few
exceptions to this generalization have been reported.6,7 More-
over, meta-substituted benzenoid compounds are observed to
be more reactive than their para isomers toward photochemical
solvolysis reactions, a phenomenon known as the “meta effect”.8
Recent studies by Zimmerman using CASSCF calculations led
to the suggestion that ortho substituents should display enhanced
reactivity similar to the meta isomers, the “meta-ortho” effect.8
Whereas a small number of studies have compared the
photochemical behavior of meta- and para-substituted stilbenes,
ortho-substituted stilbenes have been largely ignored.
We report here the results of an investigation of the
photochemical behavior of the trans isomers of ortho- (2), meta-
(3), and para-aminostilbene (4). As previously observed for
several donor-acceptor disubstituted aminostilbenes, the meta
isomer is found to have a substantially longer singlet lifetime
than the para isomer.5 In addition, the behavior of the ortho
isomer is found to be similar to that of the meta isomer. The
origin of this pronounced effect of positional isomerization on
photochemical reactivity is elucidated.
The photochemical behavior of trans-stilbene (1) and its
derivatives has been the subject of intense scrutiny.1,2 Two decay
processes can account for the behavior of trans-stilbene in
solution; fluorescence and torsion about the central double bond
to yield a short-lived twisted intermediate which decays to yield
an approximately equal mixture of trans- and cis-stilbene.1 There
is a small thermal barrier for the torsional process but not for
fluorescence. As a consequence, the lifetime of the fluorescent
singlet state is very short, and isomerization is the predominant
decay process at room temperature. Intersystem crossing
competes with singlet state isomerization and fluorescence in
the case of some halogenated and nitro-substituted stilbenes.3,4
Other monosubstituted stilbenes display behavior similar to that
of trans-stilbene.1c
We recently reported that the photochemical behavior of
several disubstituted aminostilbenes is highly dependent upon
the position of the amine substituent.5 Substituted trans-4-
aminostilbenes display behavior similar to that of stilbene (short
singlet lifetime and large isomerization quantum yields), whereas
* To whom correspondence should be addressed. E-mail: lewis@
chem.nwu.edu.
† Present address: Department of Chemistry, National Central University,
Taiwan, R. O. C.
Results and Discussion
(1) (a) Saltiel, J.; Sun, Y.-P. Photochromism, Molecules and Systems;
Durr, H., Boaus-Laurent; H., Eds.; Elsevier: Amsterdam, 1990; p 64 and
references therein. (b) Waldeck, D. H. Chem. ReV. 1991, 91, 415. (c) Saltiel,
J.; Charlton, J. L. Rearrangements in Ground and Excited States; de Mayo,
P., ed.; Academic Press: New York, 1980; Vol. 3, p 25.
(2) Recent representative publications in the area of donor-acceptor
stilbenes are: (a) Abraham, E.; Oberle, J.; Jonusauskas, G.; Lapouyade,
R.; Rulliere, C. J. Photochem. Photobiol., A 1997, 105, 101. (b) Il’ichev,
Y. V.; Ku¨hnle, W.; Zachariasse, K. A. Chem. Phys. 1996, 211, 441. (c)
LeBreton, H.; Bennetau, B.; Letard, J. F.; Lapouyade, R.; Rettig, W. J.
Photochem. Photobiol., A 1996, 95, 7. (d) Go¨rner, H. Ber. Bunsen Phys.
Chem. 1998, 102, 726. (e) Rettig, W. Top. Curr. Chem. 1994, 169, 253.
(3) Saltiel, J.; Marinari, A.; Chang, D. W.-L.; Mitchener, J. C.; Megarity,
E. D. J. Am. Chem. Soc. 1979, 101, 2982.
Electronic Absorption Spectra. The absorption spectra of
aminostilbenes 2-4 in hexane solution are shown in Figure 1.
The spectrum of 4 resembles that of trans-stilbene (1), consisting
of a single intense long wavelength band.9 The absorption
maximum of 410,11 is intermediate in energy between that of 1
and 4-dimethylaminostilbene12 as expected on the basis of the
(6) Gu¨sten, H.; Klasinc, L. Tetrahedron Lett. 1968, 26, 3097.
(7) Lapouyade, R.; Kuhn, A.; Letard, J.-F.; Rettig, W. Chem. Phys. Lett.
1993, 208, 48.
(8) Zimmerman, H. E. J. Phys. Chem. A. 1998, 102, 5616 and references
therein.
(4) Go¨rner, H.; Schulte-Frohlinde, D. J. Phys. Chem. 1979, 83, 3107.
(5) Lewis, F. D.; Yang, J.-S. J. Am. Chem. Soc. 1997, 119, 3834.
(9) Gegiou, D.; Muzkat, K. A.; Fischer, E. J. Am. Chem. Soc. 1968, 90,
3907.
10.1021/ja992335q CCC: $18.00 © 1999 American Chemical Society
Published on Web 12/08/1999

12046 J. Am. Chem. Soc., Vol. 121, No. 51, 1999
Table 1. Observed and Calculated Spectroscopic Parameters
Lewis et al.
absorption max
molar absorptivity
(log ꢀmax
fl. max
(λmax/nm)
0,0
(λ/nm)
Stokes shift/
cm-1 b
oscillator
(λmax/nm)a
)
strength (f)c
hexane
acetonitrile
hexane
acetonitrile
hexane
acetonitrile
hexane
294, (307, 321)d
294, (308, 320)
286, 334
4.45d
4.47, (4.46, 4.26)
4.25, 4.11
347e
347
407
445
387
446
380
423
326e
325
373
398
356
384
354
378
2545
2663
2319
2993
2719
3906
2367
3127
0.75f
0.76
0.23
0.23
0.12
0.12
0.86
0.92
1
2
3
4
291, 346
4.25, 4.08
298, (311, 329)
300, (312, 332)
316, (332)
4.38, (4.28, 3.94)
4.41, (4.30, 3.90)
4.50, (4.47)
acetonitrile
(318), 336
(4.49), 4.51
a Shoulders in parentheses. b Calculated by Berlman’s method.17a c f ) 4.3 × 10-9 ∫(νj)dνj. d From ref 15, in heptane. e From ref 36, in
methylpentanes. f From ref 11a.
for ortho than for meta compounds.14 Molecular orbital calcula-
tions13 suggest that the origin of this behavior in disubstituted
benzene derivatives may be due to symmetry-induced config-
uration interaction in the ortho and meta cases but not in the
para case. In the case of para-disubstituted benzene derivatives
(C2V), transitions of unlike symmetry (A and B) do not show
configuration interaction as they belong to different irreducible
representations. In the case of the ortho- and meta-disubstituted
benzene derivatives (Cs), the reduction of symmetry results in
extensive configuration interaction between pure transitions and
a resultant splitting of the state energy levels that is slightly
larger for ortho vs meta compounds. The absorption spectra of
2 and 3 (Figure 1) are similar to those of disubstituted benzenes.
Hence, 2 and 3 may be thought of as substituted anilines that
possess lowered symmetry. It should be pointed out that the
splitting of absorption bands occurs only in the case of
mesomeric substituents and not in the case of alkyl substituents.
Sterically hindered stilbenes such as 2,4,6-trimethylstilbene and
2,2′,4,4′,6,6′-hexamethylstilbene display a single long wave-
length band that has reduced intensity and vibronic structure
and is blue-shifted compared to stilbene, as expected from
theoretical considerations of orbital energies.9 A similar loss of
intensity along with a blue shift of the long wavelength band
Figure 1. Molar absorptivity (solid lines) and emission (arbitrary
has been observed for 4-(dimethylamino)-2′,6′-dimethyl-
intensity, long dashed lines) spectra of 2-4 in hexane solution. Gaussian
deconvolutions of the longest wavlength transition in the absorption
stilbene.11
The oscillator strengths of the long wavelength absorption
band have been estimated for 2-4 in hexane and acetonitrile
spectra of 2 and 3 are shown as dotted lines.
donor strength of the substituent. The spectra of 2 and 3 are
more complex than that of 1 or 4. Two long wavelength maxima
are observed for 2, whereas a maximum with shoulders are
observed for 3.10 More complex spectra might arise from the
presence of two ground-state rotamers; however, both experi-
mental and computational results (vide infra) indicate that both
rotamers of 2 and 3 should have the same spectra. The
absorption spectra of 3-4 show only modest solvent induced
shifts in acetonitrile vs hexane solution (e4 nm), whereas a
slightly larger solvent shift (12 nm) is observed for 2 in
acetonitrile (Table 1).
The appearance of the absorption spectra of 2-4 are similar
to those of some disubstituted benzene derivatives possessing
two mesomeric substituents.13 Whereas para-disubstituted ben-
zene derivatives display only a single long wavelength band,
ortho- and meta-disubstituted derivatives generally possess two
long wavelength bands, with the band splitting being greater
solvents (Table 1). Integration of the molar absorptivity curve
is straightforward for 4, since the long wavelength band is well-
resolved from other, higher energy bands. In the case of 2 and
3, the lowest-energy transitions overlap partially with more
intense, higher-energy transitions and therefore their lowest-
energy transitions are approximated as Gaussian curves (Figure
1). The basic assumption in this approach is that each spectral
band represents an excitation to a single state.15 Integration of
the deconvoluted Gaussian bands for 2 and 3 provided the
integral area required to compute the oscillator strength.16 The
calculated oscillator strengths are given in Table 1. Values for
2 and 3 are smaller than those for 1 and 4 in both hexane and
acetonitrile solution.
Fluorescence Spectra. The fluorescence spectra of 2-4 in
hexane are also shown in Figure 1. They display varying degrees
of vibronic structure in nonpolar solvents. The fluorescence
spectra of 1 and 4 are similar in appearance and the fluorescence
maximum of 4 is intermediate between those of 1 and
(10) Jungmann, H.; Gu¨sten, H.; Schulte-Frohlinde, D. Chem. Ber. 1968,
101, 2690.
(11) (a) Beale, R. N.; Roe, E. M. F. J. Am. Chem. Soc. 1952, 74, 2302.
(b) Haddow, A.; Harris, R. J. C.; Kon, G. A. R.; Roe, E. M. F. Philos.
Trans. R. Soc. 1948, 241, 147.
(12) Le´tard, J.-F.; Lapouyade, R.; Rettig, W. J. Am. Chem. Soc. 1993,
115, 2441.
(13) Murrell, J. N. The Theory of the Electronic Spectra of Organic
Molecules; John Wiley And Sons: New York, 1963.
(14) (a) Grinter, R.; Heilbronner, E.; Godfrey, M.; Murrell, J. N.
Tetrahedron Lett. 1961, 21 771. (b) Grinter, R.; Heilbronner, E. HelV. Chem.
Acta 1962, 45, 2496.
(15) Jaffe, H. H.; Orchin, M. Theory and Applications of UltraViolet
Spectrocopy; John Wiley and Sons: New York, 1962.
(16) Klessinger, M.; Michl, J. Excited States And Photochemistry Of
Organic Molecules; VCH Publishers: New York, 1995.

Photochemistry of Isomers of trans-Aminostilbenes
J. Am. Chem. Soc., Vol. 121, No. 51, 1999 12047
Table 2. Ground and Excited-State Dipole Moments for 2-4
solvatochromic
slope/cm-1 a
µg/Db
µe/Dc,d
2
3
4
2208
9881
6641
1.5
1.5
2.1
6.0
11.9
10.2
a Calculated based on eq 1. b From ref 19. c Assuming solvent cavity,
a ) 5 Å. d 95% confidence interval is ( 1.0 D.
Figure 2. Lippert-Mataga plot for 2 (0), 3 (O), and 4 (4). The ∆f
parameter is calculated according to eq 1 for hexane, cyclohexane,
dibutyl ether, diethyl ether, tetrahydrofuran, and acetonitrile in order
of increasing polarity.
4-dimethylaminostilbene.12 The S0 f S1 transition energies for
3 and 4 in hexane are similar, indicating that both isomers show
similar stabilization of the relaxed Franck-Condon state (Figure
1) whereas the S0 f S1 transition for 2 in hexane is red-shifted
by about 4 kcal/mol relative to the S0 f S1 transition for 3 or
4. The red-shifted S0 f S1 transition for 2 may result from the
larger splitting of the absorption band due to configuration
interaction (vide infra).14b The Stokes shifts for 1-4 can be
calculated according to Berlman’s method as the difference
between the 0,0 transition and the center of gravity of the
fluorescence spectrum.17a The Stokes shift for 2-4 in hexane
is 2500 ( 200 cm-1 and slightly larger in acetonitrile (3450 (
450 cm-1) suggesting that there is no “extra” stabilization of
the emissive state via large amplitude nuclear motion (Table
1).
The spectra of 2-4 in medium polar-to-polar solvents are
broad and lack vibronic structure. Upon increasing the polarity
of the medium, the fluorescence spectra shift to longer
wavelengths (Table 1 and Figure 2). The solvent-polarity-
induced shift can be used to determine the dipole moment of
the excited state using the Lippert-Mataga equation (equation
1).18
Figure 3. Temperature dependence of the fluorescence spectra of 2
and 3 in MTHF. The spectra are at the following temperatures: s 80
K; - - - - 120 K; ‚‚‚ 200 K; -‚-‚-‚ 280 K. Baseline offset increases with
increasing temperature.
< 4 < 3. The smaller excited-state dipole moment of 2 may be
attributed to the large solvatochromic shift in its absorption
spectra, possibly due to specific solvation of its nonplanar
structure, thereby reducing the observed Stokes shift and
ultimately its excited-state dipole moment. The excited-state
dipole moments of 2-4 are smaller than that of 4-(dimethyl-
amino)-4′-cyanostilbene (21 D), consistent with the absence of
strong acceptor in 2-4.2b Although excited-state dipole moments
are not expected to be a sum of component bond moments as
they are in the ground state, the observed decrease in the ortho
position is expected based on earlier work by Sinha and Yates.20
The temperature dependence of the fluorescence spectra of
2 and 3 has been studied in 2-methyltetrahydrofuran (MTHF)
between 280 and 80 K. The fluorescence spectra of 2 and 3
show no change in shape as the temperature is lowered, although
there is a slight red shift (6 and 12 nm, respectively) upon
cooling from 280 to 200 K (Figure 3). On further cooling a
blue shift is observed, and below 120 K the fluorescence spectra
of 2 and 3 are blue-shifted relative to the room-temperature
spectra. The emission spectra of both 2 and 3 at 80 K resemble
their room-temperature spectra in nonpolar solvents (Figure 3).
This behavior is consistent with the temperature dependence
of the dielectric constant of MTHF which is reported to increase
with decreasing temperature followed by a sudden drop near
the glass transition temperature of the solvent (approximately
120 K).21 Due to the absence of any large spectral shifts on
lowering the temperature, it can be concluded that there are no
large amplitude excited-state geometry changes for 2 and 3.
The quantum yields for fluorescence, φf, of 2, 3, and 4 have
been measured in cyclohexane and acetonitrile solution and the
results are reported in Table 3. The φf values for 2-4 decrease
2µe(µe - µg)
νjabs - νjfluo
)
‚∆f,
hca3
ꢀ - 1
2ꢀ + 1
η2 - 1
2η + 1
where ∆f )
-
(1)
2
[
]
where νabs and νfluo are the absorption and fluorescence maxima,
µg and µe are the ground and excited-state dipole moments, a
is the solvent cavity radius in Å, ꢀ is the solvent dielectric, and
η is the solvent refractive index. The ground-state dipole
moments of 2-4 have previously been determined from their
measured dielectric constants and refractive indices (Table 2).19
The excited-state dipole moments of 2-4, calculated from eq
1 are given in Table 2, and their values increase in the order 2
(17) (a) Berlman, I. B. Handbook of Fluorescent Spectra of Aromatic
Molecules, 2nd ed.; Academic Press: New York, 1971; pp 27-28. (b)
Berlman, I. B. Handbook of Fluorescent Spectra of Aromatic Molecules,
2nd ed.; Academic Press: New York, 1971; pp 74-75.
(18) (a) Lippert, E. Z. Elektrochem. 1957, 61, 962. (b) Mataga, N.; Kaifu,
Y.; Koizumi, M. Bull. Chem. Soc. Jpn. 1956, 29, 465. (c) Liptay, W. Z. Z.
Naturforsch. 1965, 20a, 1441.
(20) Sinha, H.; Yates, K. J. Am. Chem. Soc. 1991, 113, 6062.
(21) Furutsuka, T.; Imura, T.; Kojima, T.; Kawabe, K. Technol. Rep.
Osaka UniV. 1974, 367.
(19) Everard, K. B.; Kumar, L.; Sutton, L. E. J. Chem. Soc. 1951, 2807,
2812.

12048 J. Am. Chem. Soc., Vol. 121, No. 51, 1999
Lewis et al.
Table 3. Quantum Yields of Fluorescence (φf), Photoisomerization (φi), Singlet Lifetimes (τs), and Rate Constants of Fluorescence (kf)
a
φf
φi
τs/ns
0.070e
kf (expt)/s-1
kf (calc)/s-1 b
kf0 (fit)/s-1 h
kf0 (rt)/s-1 i
1
2
hexane
0.04c
0.88
0.69
0.78
0.40
0.05
0.03
0.52d
0.04
0.12
0.09
0.23
0.49f
0.52
6.0 × 108 c
5.9 × 108 c
1.4 × 108 g
1.1 × 108
0.8 × 108 g
0.5 × 108
5.9 × 108 g
4.6 × 108
3.8 × 108 c
cyclohexane
acetonitrile
cyclohexane
acetonitrile
cyclohexane
acetonitrile
3.7
5.4
7.5
11.7
∼0.1
∼0.1
2.4 × 108
1.3 × 108 h,j
1.3 × 108 j
6.0 × 107 j
1.3 × 108
1.0 × 108
5.8 × 107 h,j
3
4
0.3 × 108
∼5.0 × 108
∼3.0 × 108
a Excitation at 313 nm. b Calculated by eq 2. c From ref 22. d In pentane, from ref 1b. e From ref 26. f From ref 6. g In hexane solution. h Calculated
by fitting temperature dependent τs to eq 4 without the activation term. i Calculated from room-temperature data: φfτs-1/η2; η ) 1.35141 at 25 °C
(ref 30). j In isopentane.
Table 4. Temperature Dependence of τs for 1-4 in Isopentane
τs/ns
on going from nonpolar to polar solvents. Values of φf for 4
are similar to those determined for trans-stilbene;1,22 however,
values of φf for 2 and 3 are substantially higher. Such high
quantum yields are unusual for nonrigid stilbenes, but have been
reported for 4,4′-dimethoxystilbene in hexane (φf ) 0.9123) and
4,4′-diaminostilbene in dioxane (φf ) 0.7624).
T (K)
1
2
3
4
77
80
1.65a
1.2
2.4
2.3
2.3
3.0
3.2
3.6
3.5
3.6
3.8
3.7
3.8
3.7
6.0
5.9
6.2
6.2
7.3
7.3
7.3
7.3
7.4
7.4
7.4
7.5
100
120
140
160
180
200
220
240
260
280
298
Singlet lifetimes, τs, have been measured for 2-4 in cyclo-
hexane and acetonitrile solutions and the results are given in
Table 3. The lifetime of 4 is similar to that of 1 (0.108 ns in
3:2 methylcyclohexane/isohexane,25 0.070 ns in hexane at 296.2
K26), and both molecules have similar fluorescence quantum
yields.1 Thus, the fluorescence rate constants (kf ) φf ‚ τs-1
)
for both compounds are similar. The singlet lifetimes of 2 and
3 are unusually long when compared to those for 4 and 1.
Single-exponential decay is observed for 2 and 3, even though
both are expected to exist as a mixture of rotamers. Either one
rotamer predominates, or both rotamers have the same decay
time. It is also possible that one rotamer is nonfluorescent;
however, in view of the large fluorescence quantum yields in
nonpolar solvent (Table 3) it would have to be the minor
rotamer. Increasing solvent polarity causes a decrease in kf for
2-4, resulting in a increase in τf and decrease in φf (Table 3).
The rate constant of fluorescence can be calculated using the
Strickler-Berg relationship (equation 2)27
0.070b
∼ 0.1
a In 3:2 methylcyclohexane/isopentane. From ref 25. b From ref 26.
increase has been observed for 1.25 The singlet lifetimes of 2
and 3 have been determined in isopentane from 77 K to ambient
temperature and the data obtained is presented in Table 4.
Single-exponential decay is observed at all temperatures. It is
of interest to note that unlike the behavior of trans-stilbene25
or 4, the lifetimes of 2 and 3 do not decrease between 160 and
300 K, and in fact a slight increase due to changes in solvent
refractive index is noted.22,28 Below 160 K, there is an sudden
decrease in the lifetime for 3, and thereafter the lifetime remains
constant on further cooling, whereas in the case of 2 there is a
continuous decrease in the fluorescent lifetime (Table 4).
Changes in the fluorescent lifetime upon cooling below 160 K
may be interpreted as arising from either loss of a nuclear
relaxation pathway12 or from loss of orientational polarization
of an excited-state dipole by slowing down solvent reorganiza-
tion.2b However, the data above 160 K for 2 and 3 clearly
indicates that no activated decay process such as twisting about
the double bond competes with fluorescence.
Electronic Structure. The electronic structure and spectra
of the aminostilbenes have been further investigated by means
of semiempirical INDO/S-SCF-CI (ZINDO) calculations using
the algorithm developed by Zerner and co-workers.29 Ground-
state molecular structures were obtained from AM1 calculations.
The calculated ground-state geometry of 3 is quasi-planar
(aniline-styrene torsion angle is 26° for 3 and 0° for 4), whereas
the ground state of 2 has a styryl-aniline dihedral angle of 42-
48°. The ZINDO-derived frontier molecular orbitals for 3 are
shown in Figure 4. Corresponding orbitals for 2 and 4 are
provided in the Supporting Information. The HOMO and LUMO
orbitals of 2-4 resemble those of 1, aside from the presence of
the nitrogen p orbital which has a larger orbital coefficient in
∞f(νj)dνj
∞f(νj)νj-3 dνj
∫
∞ꢀ(νj)dνj
0
kf ) 2.889 × 10-9 η2
(2)
∫
0
νj
∫
0
where f(νj) and ꢀ(νj) are the fluorescence intensity and molar
absorptivity obtained from the fluorescence and absorption
spectra, respectively, and η is the refractive index of the solvent.
The calculated values of kf agree quite well with those obtained
experimentally based on the fluorescence quantum yields and
singlet lifetime data (Table 3). The calculated kf’s for 2 and 3
are lower than those for 4, in accord with lower oscillator
strength for the S0 f S1 transition in 2 and 3. The calculated kf
value for 4 in hexane is similar to that of 1.1a The decreased
value of kf in polar solvents may result, at least in part from
the a shift in the ν˜-3 -1 term in the Strickler-Berg equation
av
to lower energy, although the Franck-Condon factors may also
be important.
The singlet lifetime of 4 in isopentane solution increases from
∼0.1 ns at room temperature to 1.2 ns at 77 K. A similar
(22) Saltiel J.; Waller, A. S.; Sears, D. F., Jr.; Garrett, C. Z. J. Phys.
Chem. 1993, 97, 2516.
(23) Zeglinski, D. M.; Waldeck, D. H. J. Phys. Chem. 1988, 92, 692.
(24) Smit, K. J.; Ghiggino, K. P. Chem. Phys. Lett. 1985, 122, 369.
(25) Sumitani, M.; Nakashima, N.; Yoshihara, K.; Nagakura, S. Chem.
Phys. Lett. 1977, 51, 183.
(26) Kim, S. K.; Courtney, S. H.; Fleming, G. R. Chem. Phys. Lett. 1989,
159, 543 and references therein.
(27) (a) Strickler, S. J.; Berg, R. A. J. Chem. Phys. 1962, 37, 814. (b)
Birks, J. B.; Dyson, D. J. Proc. R. Soc. London, Ser. A 1963, 275, 135.
(28) Ru¨ckert, I.; Demeter, A.; Morawski, O.; Ku¨hnle, W.; Tauer, E.;
Zachariasse, K. A. J. Phys. Chem. A 1999, 103, 1958.
(29) Zerner, M. C.; Loew, G. H.; Kirchner, R. F.; Mueller-Westerhoff,
U. T. J. Am. Chem. Soc. 1980, 102, 589.

Photochemistry of Isomers of trans-Aminostilbenes
J. Am. Chem. Soc., Vol. 121, No. 51, 1999 12049
the fluorescent singlet states are shown to be similar to those
of the ground state.
(c) As a consequence of more extensive configuration
interaction in 2 and 3 vs 4, the singlet oscillator strengths and
fluorescence rate constants are lower for 2 and 3 than for 4.
(d) On the basis of the observation of temperature-dependent
lifetimes for 4, but not for 2 or 3, a thermally activated decay
process is shown to compete with fluorescence in the case of 4
but not 2 or 3.
Photoisomerization. Quantum yields for photoisomerization
(φi) of 2-4 in cyclohexane solution are reported in Table 3. As
previously observed by Gu¨sten and Klasinc6 the value of φi for
4 in cyclohexane is similar to that of stilbene. Significantly
smaller quantum yields are observed for 2 and 3. In fact, the
values of φi for 2 and 3 are, to our knowledge, the lowest that
have been observed for a monosubstituted stilbene derivative.1c
The efficiency of photoisomerization increases substantially with
increasing solvent polarity for 2 and 3, but only slightly for 4.
Both singlet- and triplet-state mechanisms have been thor-
oughly characterized for the photoisomerization of trans-stilbene
and its derivatives.1,3,4 Stilbene serves as the prototype for the
singlet-state mechanism. A small barrier (Eact ) 3.5 kcal/mol,
A ) 1012.6) for torsion about the central double bond separates
the planar fluorescent 1S* state from the perpendicular 1P* state
which decays with similar probability to the trans and cis ground
states.1 Thermally activated barrier crossing competes effectively
with fluorescence at room temperature, but less so at lower
temperatures, resulting in a temperature-dependent singlet
lifetime (Table 4).25 The stilbene triplet can be populated by
triplet sensitization and undergoes barrierless twisting to the
perpendicular triplet.1 Intersystem crossing from 1S* is estimated
to occur with a rate constant kisc ) 3.9 × 107 s-1 but does not
compete with fluorescence (kf ) 6.0 × 108 s-1) and torsion on
the singlet state surface at room temperature (k ≈ 1010 s-1 at
ambient temperature).1a However, significant intersystem cross-
ing is observed for singlet stilbenes with halogen or nitro
substituents.3,4 Intersystem crossing rates for these molecules
are substantially larger than for stilbene. For example, in the
case of 3-bromostilbene kisc ) 4.0 × 109 s-1 resulting in
intersystem crossing being a major decay pathway for its excited
singlet state.4
Figure 4. Frontier molecular orbitals for 3 as optimized by ZINDO.
the HOMO than the LUMO. To the extent that the HOMO f
LUMO transition contributes to the lowest-energy transition,
the lowest excited singlet states should possess aniline f styrene
charge-transfer character, in accord with the observation of
solvatochromic shifts for their fluorescence spectra.
The calculated energies of the ground and first two excited
singlet states of 1-4 are reported in Table 5 along with the
calculated oscillator strengths and configuration interaction
descriptions of the excited states. In the case of 2 and 3 data is
reported for both of the ground-state rotamers. The energy
difference, as calculated by AM1, between the lower-energy
(see Figure 1 for structures) and higher-energy rotamers is only
0.11 and 0.15 kcal/mol, respectively. Their calculated S1 and
S2 energies are identical. The calculated ground-state energies
for 2-4 are similar despite the nonplanarity of 2. The lowest-
energy transition for 4, like that of 1, is an allowed, essentially
pure HOMO to LUMO π,π* transition. The calculated and
observed singlet energies for 4 (Tables 1 and 5) are in good
agreement: however, the calculated oscillator strengths are
higher than the values obtained from the integrated absorption
spectra (Table 1). The second singlet states for both 4 and 1
are predicted to have extensive configuration interaction and
low oscillator strengths. Both of the two lowest singlet states
calculated for 2 and 3 have extensive configuration interaction
and more equal oscillator strengths, as previously reported for
other ortho- and meta-disubstituted benzene derivatives with
mesomeric substituents.14,16 The contribution of the HOMO f
The singlet lifetime of 4 increases from 0.1 ns at room
temperature to 1.2 ns at 77 K, similar to the increase reported
for 1 upon cooling.25 The room-temperature quantum yields for
photoisomerization and fluorescence of 4 and 1 are also similar
(Table 3). Thus, it is reasonable to assume that 4 isomerizes
via a singlet-state mechanism. The barrier for singlet-state
torsion in 4 can be estimated from the singlet lifetime and
photoisomerization quantum yield using eq 3, assuming the
value of A0 ) 1012.6 reported for stilbene. The calculated singlet
barrier for 4 is 3.5 kcal/mol, similar to that reported for 1.
1
LUMO configuration to the description of S* is substantially
smaller for 2 and 3 (∼50%) than for 4 or 1 (> 95%). The
ZINDO calculations are consistent with the observation of two
low-energy transitions for 2 and 3 (Figure 1). The calculated
and observed S2 energies are in reasonable agreement; however,
1
the calculated S* energies and oscillator strengths are higher
than the observed values, particularly in the case of 2. Splitting
of the lowest energy π,π* transition of 4-(dimethylamino)-3′-
nitrostilbene has previously been reported Lapouyade et al.7
based on CNDO/S calculations. In this case, the calculated
oscillator strength for the lower-energy transition is significantly
lower than for the higher-energy transition and no maximum
or shoulder is apparent in the absorption spectrum.
Several important conclusions concerning the fluorescent
singlet states of 2-4 can be drawn from the spectral data and
electronic structure calculations.
(a) On the basis of the fluorescence solvatochromic shifts
and calculations, the singlet states of 2-4 are shown to possess
similar dipole moments and degrees of aniline f styrene charge-
transfer character.
(b) On the basis of the absence of large Stokes shifts or
strongly temperature-dependent fluorescence, the geometry of
2φiτ-1
Eact ) -RT ln
(3)
(
)
A0
In contrast to the temperature-dependent lifetimes of 4 and
1, the singlet lifetimes of 2 and 3 are independent of temperature
between 298 and 160 K (Table 4). This suggests that no
activated process for decay of the fluorescent 1S* state competes
with the nonactivated processes, fluorescence and intersystem
crossing, at room temperature or below. Thus, photoisomeriza-
tion of 2 and 3 is concluded to occur via intersystem crossing
followed by barrierless torsion on the triplet surface. Assuming
that the twisted triplet decays to yield a 1:1 ratio of trans and

12050 J. Am. Chem. Soc., Vol. 121, No. 51, 1999
Lewis et al.
Table 5. Results of the INDO/S-CIS-SCF (ZINDO) Calculations for the Lowest Singlet States of 1-4
a
S0
S1
S2
∆Hf
kcal/mol
∆E/nm
fc
descriptionb
34:35 (0.97)
∆E/nm
fc
descriptionb
1
2
60.89
301
1.265
286
0.015
31:37 (0.12),
32:35 (0.13),
32:38 (0.12),
33:35 (0.13),
34:36 (0.21),
34:37 (0.25)
36:38 (0.12),
37:38 (0.26),
37:39 (0.12),
37:40 (0.17)
36:38 (0.30),
37:38 (0.44),
37:40 (0.13)
34:39 (0.16),
37:39 (0.62)
59.70
(59.81)d
306
(306)d
0.493
(0.397)d
36:41 (0.10),
37:38 (0.56),
37:40 (0.16)
286
(285)d
0.322
(0.114)d
3
4
59.51
(59.66)d
59.11
308
(308)d
316
0.577
(0.588)d
1.328
36:38 (0.12),
37:38 (0.51),
37:40 (0.20)
37:38 (0.95)
296
(297)d
302
0.693
(0.656)d
0.065
a Geometry optimized using the AM1 Hamiltonian as implemented by the MOPAC package under Cache, Release 3.5 (Cache Scientific). b Orbital
numbers of the pure configurations are given and the weight of each configuration is given in parentheses. Only configurations with 10% or greater
contribution are included. The HOMO for 1 is orbital 34 and for 2-4 is orbital 37. c Oscillator strength. d Values in parentheses are for the conformer
obtained by rotating the molecules shown in Figure 1 by 180° about the aniline-styrene bond. See Supporting Information for a complete table.
cis isomers, quantum yields and rate constants for intersystem
crossing can be estimated from the measured isomerization
quantum yields and lifetimes (φisc ) 2φi, kisc ) φiscτs-1). The
sum of the quantum yields for fluorescence and intersystem
crossing for both 2 and 3 in hexane solution are within the
experimental error of 1.0, in accord with the absence of singlet-
state isomerization or nonradiative decay. The calculated values
of kisc for 2 and 3 are 2.2 and 2.4 × 107 s-1, respectively, similar
to the value for 1 (3.9 × 107 ( 2.3 × 107 s-1) estimated by
Saltiel.1 Thus, the occurrence of triplet isomerization for 2 and
3 is a consequence of their abnormally long singlet lifetimes,
rather than rapid intersystem crossing such as that observed for
halostilbenes.4
Increasing solvent polarity results in an increase in both the
singlet lifetime and isomerization quantum yield. Assuming that
isomerization remains a triplet-state process in acetonitrile, the
calculated values of kisc ) 4.4 and 3.9 × 107 s-1 for 2 and 3,
Figure 5. Temperature dependence of fluorescence lifetime for 2 (0)
respectively. These values are approximately twice as large as
those obtained in cyclohexane solution, plausibly reflecting
smaller singlet-triplet splitting in the more polar solvent.
A lower bound for the barrier to torsion on the singlet surface
can be estimated by modeling the temperature dependence of
the singlet lifetimes, τs, using eq 4, assuming that both the gas-
phase radiative rate constant, kf0, and kisc are independent of
temperature, the refractive index, η, is temperature-dependent,30
and 3 (O) in isopentane between 160 and 300 K. The fitted lines are
based on eq 4 and are as follows: kisc (2) ) 2.2 × 107 s-1 and kisc (3)
) 2.4 × 107 s-1: Eact ) s 6.0 kcal/mol; - - - 7.0 kcal/mol; ‚‚‚ 8.0
kcal/mol with A ) 1012 (kf0 ) 1.3 × 108 s-1 for 2 and 5.8 × 107 s-1
for 3).
residuals from the η2 fits are within experimental error.31 The
kf0 term so obtained was then used in eq 4 to estimate the
minimum activation barrier to singlet isomerization. Literature
and that the pre-exponential for singlet torsion A0 ) 1012 s-1
.
values of A0 for alkene isomerization lie in the range 1012
-
1013 s-1 1b,c,4
Assumption of a value at the lower end of this
.
1
τs )
(4)
range provides the lowest activation energy, Eact. The results
of kinetic modeling (see Supporting Information) of the singlet
lifetime vs temperature for several trial values of Eact is shown
in Figure 5 along with experimental lifetimes. Inspection of
Figure 5 provides values of Eact g 7 kcal/mol for both 2 and 3.
To our knowledge, these barriers are the largest observed for a
monosubstituted, unconstrained stilbene derivative.
According to the original Orlandi-Siebrand theoretical
treatment of stilbene isomerization, the small barrier for singlet-
state torsion is the result of an avoided crossing between the
lowest energy singlet excited 1B state (1S*) and a higher energy
doubly excited 1A state.32 Subsequent theoretical treatments also
-Eact
RT
kf0‚η2 + kisc + A exp
(
)
0
The values for kf0 for 2 and 3 obtained by fitting the temperature-
dependent fluorescence lifetimes to eq 4 but without the
activated nonradiative term are similar to those obtained from
the fluorescence quantum yield data and singlet lifetimes at
ambient temperature (φf ‚ τs-1/η2)(Table 3). Improved fits are
obtained by allowing the refractive index exponent to assume
nonintegral values (see Supporting Information) although the
(30) The temperature dependence η for isopentane (1.35691-0.00055*(t
- 15) where t is in °C) may be obtained from the following reference:
Timmermans, J. Physico-Chemical Constants of Pure Organic Compounds;
Elsevier: New York, 1950; Vol. 1.
(31) For a discussion of nonintegral refractive index exponents see:
Shibuya, T. Chem. Phys. Lett. 1983, 103, 46 and reference 22.

Photochemistry of Isomers of trans-Aminostilbenes
J. Am. Chem. Soc., Vol. 121, No. 51, 1999 12051
Scheme 1
indicate the presence of a small thermal barrier on the singlet-
state surface.33 The experimental evidence as summarized by
Saltiel indicates that the barrier height is approximately 3.5 kcal/
1
1
mol and that the planar S* and perpendicular P* states are
approximately isoenergetic.34 Internal conversion from P* to
1
the ground state is sufficiently rapid (kic g 1012 s-1) to render
1
1
conversion of S* f P* largely irreversible, despite the low
barrier for the reverse process.34
To a first approximation, the height of the barrier for singlet-
state torsion should be dependent upon the relative energies of
the planar 1S* and perpendicular 1P* states, lowering the energy
of 1P* relative to 1S* decreasing the barrier and increasing the
energy of 1P* relative to 1S* increasing the barrier.35 The relative
energies of 1S* can be estimated from the positions of their 0,0
transitions (Table 1). The 0,0 energies of 3 and 4 are ap-
proximately 7 kcal/mol lower than that of 136 and the energy
of 2 is an additional 4 kcal/mol lower (Table 1). Extended or
through-conjugation is presumably responsible for the stabiliza-
tion of 4 (Scheme 1a). Extended conjugation is not possible
for 3 and is of reduced importance for 2 due to the large styryl-
aniline dihedral angle. The 0,0 energies of both 2 and 3 are
lowered by splitting of the lowest transition due to configuration
interaction (Figure 1 and Table 1) which is largely responsible
for the lower singlet energies of 2 and 3.
Figure 6. Potential energy diagram for ethylene bond torsion in the
lowest excited state for 1-4 in nonpolar solvents. Energies relative to
1S* for stilbene.
parameters for the amino group (-1.3 and -0.16, respec-
tively).40 The difference in these values is the largest for any
substituent in a standard tabulation of Hammett parameters.40
In the absence of a stabilized charge-transfer configuration (eq
5), the energy of 1P* for 3 should be similar to that of 1. Since
1
the energy of S* in 3 is ∼7 kcal/mol more stable than for 1
this should result in endergonic torsion with a barrier g7 kcal/
mol (Figure 6), in accord with experimental observation.
In the case of 2 resonance stabilization of 1P* is expected to
be smaller than that for 4 due to nonbonded repulsion between
the amine and styryl substituents. AM1 minimization of the
twisted geometry expected for 1P* provides a dihedral angle of
31° between the amine and benzyl groups, somewhat smaller
than the value for S0 (45-48°). Nakata et al.41 have computed
the destabilization energy of benzylic cations upon rotation of
the phenyl-methylene bond at the RHF/6-31 G* level and
find that it obeys a cos2(θ) dependence. Thus, the resonance
stabilization energy of 1P* in 2 might be expected to be
approximately 3/4 as large as that in 4 (∼5 kcal/mol). However,
since the stabilization of 1S* is 11 kcal/mol, twisting should be
endergonic with a barrier >6 kcal mol (Figure 6), in accord
with experimental observations.
The nature of the twisted 1P* state is the subject of continuing
1
controversy. Michl and Bonacˇic´-Koutecky´ have described P*
for simple alkenes as a resonance hybrid of biradical and charge-
transfer configurations (eq 5).37,38
ψP* ) c1ψbiradical(A B ) + c2ψCT(A B ) + c3ψCT(A B ) (5)
•
•
+
-
- +
An electron-donating or electron-withdrawing para-substituent
would be expected to stabilize one of the two charge-transfer
configurations and thus lower the energy of 1P*. Evidently, the
para-amino substituent in 4 stabilizes 1P* (Scheme 1b) to
Concluding Remarks. In summary, the exceptionally long
singlet lifetimes for the ortho- and meta-aminostilbenes 2 and
3 are a consequence of large barriers for torsion (>7 kcal/mol)
on the singlet state surface and small fluorescence rate constants.
Therefore, 2 and 3 decay exclusively via fluorescence and
intersystem crossing to the triplet, which undergoes barrier-
less torsion about the ethylene bond.1,42 In nonpolar solvents,
fluorescence is the major decay pathway, and thus the singlet
lifetime is determined mainly by the fluorescence rate constant
which is larger for 2 than for 3. Increasing the solvent polarity
results in a decrease in the fluorescence rate constant and an
increase in singlet lifetime. The calculated rate constants for
intersystem crossing for 2 and 3 (∼2 × 107 s-1) are similar to
the estimated value for trans-stilbene.1
1
approximately the same extent that it stabilizes S* (∼7 kcal/
mol). This would result in a barrier for singlet state torsion
similar to that of 1 (Figure 6), in accord with experimental
observation.
In the case of 3, the meta-amino substituent cannot stabilize
1
the charge-transfer configuration of P* by resonance. Dewar
et al.39 have calculated that the heats of formation of the 4- and
3-aminobenzyl cations using the MINDO Hamiltonian and find
the 3-aminobenzyl cation to be 24 kcal/mol less stable! This
difference is also reflected in the Hammett σ+
and σmeta
para
(32) (a) Orlandi, G.; Siebrand, W. Chem. Phys. Lett. 1975, 30, 352. (b)
Orlandi, G.; Palmieri, P.; Poggi, G. J. Am. Chem. Soc. 1979, 101, 3492.
(33) (a) Hohlneicher, G.; Dick, B. J. Photochem. 1984, 27, 215. (b) Troe,
J.; Weitzel, K.-M. J. Chem. Phys. 1988, 88, 7030.
(34) Saltiel, J.; Waller, A. S.; Sears, D. F., Jr. J. Am. Chem. Soc. 1993,
115, 2453.
The large barriers for singlet state torsion in 2 and 3 both
1
1
result from greater stabilization of S* vs P*. However, the
(35) (a) Hammond, G. S.; J. Am. Chem. Soc. 1955, 77, 334. (b) Leffler,
J. E.; Grunwald, E. Rates and Equilibria of Organic Reactions; Wiley: New
York, 1963.
(36) Dyck, R. H.; McClure, D. S. J. Chem. Phys. 1962, 36, 2326.
(37) Bonacˇic´-Koutecky´, V.; Ko¨hler, J.; Michl J. Chem. Phys. Lett. 1984,
104, 440.
(40) Brown, H. C.; Okamoto, Y. J. Am. Chem. Soc. 1958, 80, 4979.
(41) Nakata, K.; Fujio, M.; Saeki, Y.; Mishima, M.; Tsuno, Y.;
Nishimoto, K. J. Phys. Org. Chem. 1996, 9, 573.
(42) (a) Saltiel, J.; Chang, D. W. L.; Megarity, E. D.; Rousseau, A. D.;
Shannon, P. T.; Thomas, B.; Uriarte, A. K. Pure Appl. Chem. 1975, 41,
559. (b) Ni, T.; Caldwell, R. A.; Melton, L. A. J. Am. Chem. Soc. 1989,
111, 457.
(38) Bonacˇic´-Koutecky´, V.; Michl J. J. Am. Chem. Soc. 1985, 107, 1765.
(39) Dewar, M. J. S.; Landman, D. J. Am. Chem. Soc. 1977, 99, 7439.

12052 J. Am. Chem. Soc., Vol. 121, No. 51, 1999
Lewis et al.
origin of these barriers is somewhat different for the two
isomers. In both cases 1S* is stabilized by splitting of the lowest
singlet state resulting from configuration interaction, the stabi-
lization being larger for 2 than for 3. In the case of 3 the meta-
amino substituent is unable to stabilize the charge-transfer
configuration of the twisted intermediate 1P*. In the case of 2,
resonance stabilization of the charge-transfer configuration of
1P* by the ortho-amino substituent is possible, but the magnitude
of the stabilization is diminished by nonbonded interactions.
Elucidation of the origin of the barrier for singlet-state torsion
in 2 and 3 provides an explanation for the “meta effect”
previously reported for several donor-acceptor disubstituted
stilbenes.5 We suggested that the long-lived fluorescent state
for these stilbene derivatives might have TICT (twisted in-
tramolecular charge-transfer) character with twisting about the
aniline-styrene single bond. The current results suggest that
there is no need to invoke TICT character for the fluorescent
1S* states of the aminostilbenes.
prolong the singlet lifetime of donor-acceptor-substituted
stilbenes as well as monosubstituted stilbenes.
The greatly diminished reactivity of 3 vs 4 toward singlet
state photoisomerization follows the normal ground state pattern
for meta vs para electron-donating substituents. This result
stands in contrast to the enhanced reactivity of meta- vs para-
substituted methoxylated benzyl acetates toward photochemical
solvolysis, an example of the so-called “meta effect” (or “meta-
ortho effect”) which has been investigated extensively by
Zimmerman and others.8 Clearly electronic transmission in the
excited state determines the effects of substituents on stilbene
photoisomerization as well as benzyl acetate solvolysis. How-
ever the consequences are not the same.
Whereas an amino group in the ortho or meta position may
be unique in its ability to increase the stilbene singlet lifetime,
it may be possible to generalize the model presented in Figure
6 to other unimolecular photochemical reactions. Whenever an
activated nonradiative decay pathway competes with nonacti-
vated fluorescence and intersystem crossing, it should be
possible to alter the singlet lifetime by changing the barrier for
the activated process. Substituents which stabilize the fluorescent
singlet state more than the transition state for the activated
pathway will increase barrier and hence the singlet lifetime,
whereas substituents which stabilize the transition state more
than the singlet state will lower the barrier and decrease the
singlet lifetime. In cases where the activated process leads to
product formation, the efficiency of product formation will also
be affected.
The amino substituent is apparently unique in its ability to
dramatically increase the barrier for singlet-state isomerization
of a meta- or ortho-monosubstituted stilbene. A search of the
literature reveals that this effect was nearly discovered 30 years
ago by Gu¨sten and co-workers.6 They reported the photoisomer-
ization quantum yields for seven para- and seven meta-
substituted stilbenes in cyclohexane, all of which had quantum
yields for trans f cis isomerization >0.3. The amino group
was included among the para-substituents but not the meta
compounds. They subsequently reported the quantum yields for
photocyclization of the cis isomers of both 3 and 4.10 The
quantum yields reported for photoisomerization of the para-
and meta-methoxystilbenes are 0.40 and 0.31, suggesting that
the barrier for photoisomerization of the meta isomer is not
unusually large.6 Investigations of ortho-substituted stilbenes
appear to be limited to methyl substituents. The small fluores-
cence quantum yield reported for 2,4,6-trimethylstilbene (φf )
0.003) suggests that it has a small barrier for singlet torsion.9
The ortho-methyl substituents might be expected to stabilize
Experimental Section
Materials. The trans-aminostilbenes 2-4 were prepared by the
standard Wittig route.45 Corresponding ortho-, meta-, para-nitrobenz-
aldehydes (Aldrich) were reacted with triphenylphosphonium chloride
(Aldrich) in a CH2Cl2-H2O dual phase system using tetrabutylammo-
nium iodide (Aldrich) as a phase-transfer catalyst (10 mol %). The
reaction mixture was stirred at room temperature overnight under a N2
atmosphere. After the reaction was complete, the CH2Cl2 layer was
separated and washed with brine several times. Purification was carried
out by column chromatography (SiO2/hexanes-EtOAc (80:20), 230-
400 mesh SiO2) to remove the cis isomer. If necessary, the trans isomer
was enriched by refluxing a cis-trans mixture in benzene using a
catalytic amount of I2 prior to chromatography. The trans isomers were
further purified by recrystallization from MeOH to yield a pale yellow
solid (2: mp ) 105-106 °C, lit. mp46 ) 102-105 °C; 3: mp ) 119-
1
1P* more than S*.
The photochemistry of 4,4′-disubstitued stilbenes has recently
been the subject of extensive investigations. A high barrier for
photoisomerization of singlet 4,4′-dimethoxystilbene23,43 and
high fluorescence quantum yields for 4,4′-diaminostilbenes24 and
for several donor-acceptor stilbenes, including 3′-amino-4-
methoxycarbonylstilbene5 have been reported. In these examples
it seems likely that the substituents lower the energy of
fluorescent 1S* state more than that of the 1P* state. Papper et
al.44 have studied a variety of donor and acceptor stilbene and
find low barriers for singlet isomerization with relatively weak
donors (methoxy, methyl, and halogen) but large barriers for
the stronger dimethylamino donor. They have correlated their
results with the difference in the Hammett constants for the two
substituents.44
120 °C, lit. mp47 ) 120-121 °C; 4: mp ) 153-154 °C, lit. mp10
)
151 °C). Typical overall yields of the trans isomer were 40%. Reduction
of the nitro group to the amino group was carried out using Zn/HCl-
AcOH as the reducing agent.48 Typical yields from the reduction were
85%. Purification of the corresponding trans-aminostilbene was carried
out by recrystallization from HPLC grade MeOH. All materials were
found to have greater than 98.5% trans isomer as estimated by GLC.
1H NMR and HRMS was done to establish that the identity of all
compounds. All solvents used for spectroscopy were either spectro-
photometric or HPLC grade (Fisher) and were used as received.
Hexanes was used instead of hexane as a solvent for all spectroscopic
measurements.
Methods. 1H NMR spectra were measured on a Varian Gemini 300
spectrometer. GLC analysis was performed using a Hewlett-Packard
HP 5890 instrument equipped with a HP1 poly(dimethylsiloxane)
capillary column. UV-vis spectra were measured on a Hewlett-Packard
8452A diode array spectrometer using a 1 cm path length quartz cell.
Fluorescence spectra were measured on a SPEX Fluoromax spectrom-
eter. Low-temperature spectra were measured either in a Suprasil quartz
Lapouyade et al.7 compared the photophysical behavior of
meta- vs para-nitro-4′-dimethylaminostilbene and found the
meta isomer to have a significantly lower φf than the para
isomer. Simple resonance arguments suggest that moving an
electron-withdrawing substituent from the para to the meta
1
1
position should stabilize P* (eq 5) more than S* and hence
lower the torsional barrier. Thus, the amino group, when present
at the meta or ortho position, may be unique in its ability to
(43) Saltiel, J.; Waller, A. S.; Sears, D. F.; Hoburg, E. A.; Zeglinski, D.
M.; Waldeck, D. H. J. Phys. Chem. 1994, 98, 10689.
(44) Papper, V.; Pines, D.; Likhtenshtein, G.; Pines, E. J. Photochem.
Photobiol., A 1997, 111, 87.
(45) Lee, B. H.; Marvel, C. S. J. Polym. Sci. Chem. Ed. 1982, 20, 393.
(46) Ziegler, C. B.; Heck, R. F. J. Org. Chem. 1978, 43, 2941.
(47) Boyer, J. H.; Alul, H. J. Am. Chem. Soc. 1959, 81, 2136.
(48) Taylor, T. W. J.; Hobson, P. M. J. Chem. Soc. 1936, 181.

Photochemistry of Isomers of trans-Aminostilbenes
J. Am. Chem. Soc., Vol. 121, No. 51, 1999 12053
EPR tube using a quartz liquid nitrogen coldfinger dewar at 77K or in
a Oxford Cryogenics DN1704 cryostat fitted with a Oxford Instruments
ITC4 temperature controller. Anthracene (2 and 4, Φf ) 0.2749) or
phenanthrene (3, Φf ) 0.1450) was used as a external standard for the
measurement of fluorescence quantum yields. Fluorescence quantum
yields were measured by comparing the integrated area under the
fluorescence curve for the trans-aminostilbenes and the standard at equal
absorbance at the same excitation wavelength. All fluorescence spectra
are uncorrected, and the estimated error for the fluorescence quantum
yields is (10%, although the quantum yields were corrected for the
refractive index of the solvent. Fluorescence decays were measured
either on a Photon Technologies International LS-1 single photon
counting apparatus with a gated hydrogen arc lamp using a scatter
solution to profile the lamp or on a Photon Technologies International
Timemaster stroboscopic detection instrument with a gated hydrogen
or nitrogen lamp using a scatter solution to profile the instrument
response function. Nonlinear least-squares fitting of the decay curves
were fitted using the Levenburg-Marquardt algorithm as described
by James et al.51 as implemented by the Photon Technologies
International Timemaster (version 1.2) software. Goodness of fit was
determined by judging the ø2 (<1.3 in all cases), the residuals, and the
Durbin-Watson parameter (>1.6 in all cases). Measurements of
quantum yields of photoisomerization for 2 and 3 were measured on
optically dense degassed solutions (∼10-3 M) using an excitation
wavelength of 313 nm, and the extent of photoisomerization (<5%)
was quantified using GLC. Quantum yields of photoisomerization for
4 (<10% conversion) were determined by changes in absorbance at 2
wavelengths before and after irradiation at 313 nm. For this purpose,
the cis isomer of 4 was synthesized using the above methodology and
its molar absorptivity determined. Excitation at 313 nm was achieved
using a 150 W medium-pressure mercury arc lamp filtered through an
alkaline potassium dichromate solution (2 and 3) or a 150 W xenon
lamp filtered through a Bausch and Lomb high-intensity monochromator
(4). All spectroscopic measurements were performed on solutions that
were purged with dry N2 for 20-25 min. In some cases, solutions were
degassed under vacuum (<10-4 Torr) through five freeze-pump-thaw
cycles and flame-sealed to establish the validity of the N2-purged
experiments. INDO/S-CIS-SCF (ZINDO) calculations ( nine occupied
and nine unoccupied frontier orbitals) were performed on a Macintosh
IIfx computer using the ZINDO Hamiltonian as implemented by Cache
Release 3.5.29,52 All structures used in the ZINDO calculations were
based on ground state, SCF/AM1, optimized geometries using the
MOPAC (version 94.10) suite of programs as implemented under Cache
Release 3.5.52 Gaussian fitting procedures were carried out using
GRAMS/386 (version 3) software on a IBM-compatible PC running
Windows 3.1.53
Acknowledgment. We are grateful to Professors Jack Saltiel
and Siegfried Schneider for helpful discussions. Funding for
this project was provided by NSF grant CHE-9734941.
Supporting Information Available: ZINDO optimized
frontier orbitals for 2 and 4, kinetic modeling of the low-
temperature singlet lifetime data for 2 and 3 in isopentane, and
ZINDO calculated energies for the rotamers of 2 and 3 (PDF).
This material is available free of charge via the Internet at
http://pubs.acs.org.
JA992335Q
(49) Dawson, W. R.; Windsor, M. W. J. Phys. Chem. 1968, 72, 3251.
(50) Birks, J. B. Photophysics of Aromatic Molecules; John Wiley And
Sons: New York, 1970.
(51) James, D. R.; Siemiarczuk, A.; Ware, W. R. ReV. Sci. Instrum. 1992,
63, 1710.
(52) Oxford Molecular Group, Inc., Campbell, CA 95008. Telephone:
(800) 876-9994. Web site: http://www.oxmol.com.
(53) Galactic Industries Corp., Salem, NH 03079. Telephone: (603) 898
7600. Web site: http://www.galactic.com/.