Substituent-dependent photoinduced intramolecular charge transfer in N-aryl-substituted trans-4-aminostilbenes

Article abstract of DOI:10.1021/ja047604d

The photochemical behavior of trans-4-(N-arylamino)stilbene (1, aryl = 4-substituted phenyl) in solvents more polar than THF is strongly dependent on the substituent in the N-aryl group. This is attributed to the formation of a twisted intramolecular char

Full text of DOI:10.1021/ja047604d

Published on Web 09/11/2004
Substituent-Dependent Photoinduced Intramolecular Charge
Transfer in N-Aryl-Substituted trans-4-Aminostilbenes
Jye-Shane Yang,* Kang-Ling Liau, Chin-Min Wang, and Chung-Yu Hwang
Contribution from the Department of Chemistry and Center for Nano Science and Technology,
National Central UniVersity/UST, Chung-Li, Taiwan 32054
Received April 26, 2004; E-mail: jsyang@cc.ncu.edu.tw
Abstract: The photochemical behavior of trans-4-(N-arylamino)stilbene (1, aryl ) 4-substituted phenyl) in
solvents more polar than THF is strongly dependent on the substituent in the N-aryl group. This is attributed
to the formation of a twisted intramolecular charge transfer (TICT) state for those with a methoxy (1OM),
methoxycarbonyl (1CO), or cyano (1CN) substituent but not for those with a methyl (1Me), hydrogen (1H),
chloro (1Cl), or trifluoromethyl (1CF) substituent. On the basis of the ring-bridged model compounds 3-6,
the TICT states for 1CN and 1CO result from the twisting of the anilino-benzonitrilo C-N bond, but for
1OM it is from the twisting of the stilbenyl-anilino C-N bond, both of which are distinct from the TICT
states previously proposed for N,N-dimethylaminostilbenes.
Introduction
of the dimethylamino (donor, D)-benzonitrilo (acceptor, A)
single bond that results in a nearly perpendicular D-A geometry
The electronic excited states of arylamines possess more or
less the character of intramolecular charge transfer (ICT) from
the amino nitrogen to the arene, and those having the ICT
configuration as the main component are often referred to as
ICT states. While many ICT-based arylamines have been
investigated as nonlinear optical materials,1 two-photon-absorb-
ing chromophores,2 electrooptical switches,3 chemical sensors,4
and fluorescence probes,5 characterization of their ICT states
(e.g., electronic nature and molecular geometry) is far from
straightforward, even for a simple molecule such as 4-(N,N-
dimethylamino)benzonitrile (DMABN). Since the first observa-
tion by Lippert et al.,6 the phenomenon of dual fluorescence
for DMABN has led to numerous theoretical and experimental
studies to account for the origin of the ICT fluorescence.6-13
Several distinct models have been proposed, including a twisting
(twisted intramolecular charge transfer, TICT),7-10 an in-plane
bending of the cyano group (rehybridization by intramolecular
charge transfer, RICT),11 and a pyramidalization (wagged
(7) (a) Rotkiewicz, K.; Grellmann, K. H.; Grabowski, Z. R. Chem. Phys. Lett.
1973, 19, 315-318. (b) Grabowski, Z. R.; Rotkiewicz, K.; Siemiarczuk,
A.; Cowley, D. J. Baumann, W. NouV. J. Chim. 1979, 3, 443-454. (c)
Rettig, W. Angew. Chem., Int. Ed. Engl. 1986, 25, 971-988. (d) Rettig,
W.; Maus, M. In Conformational Analysis of Molecules in Excited States;
Waluk, J., Ed.; Wiley-VCH: New York, 2000; Chapter 1, pp 1-55. (e)
Grabowski, Z. R.; Rotkiewicz, K. Chem. ReV. 2003, 103, 3899-4031.
(8) (a) Rettig, W.; Bliss, B.; Dirnberger, K. Chem. Phys. Lett. 1999, 305, 8-14.
(b) Rettig, W.; Zietz, B. Chem. Phys. Lett. 2000, 317, 187-196. (c) Rettig,
W.; Lutze, S. Chem. Phys. Lett. 2001, 341, 263-271. (d) Dobkowski, J.;
Wo´jcik, J.; Kozminski, W.; Kolos, R.; Waluk, J.; Michl, J. J. Am. Chem.
Soc. 2002, 124, 2406-2407. (e) Rotkiewicz, K.; Rettig, W.; Detzer, N.;
Rothe, A. Phys. Chem. Chem. Phys. 2003, 5, 998-1002. (f) Dobkowski,
J.; Michl, J.; Waluk, J. Phys. Chem. Chem. Phys. 2003, 5, 1027-1031. (g)
Kwok, W. M.; Ma, C.; George, M. W.; Grills, D. C.; Matousek, P.; Parker,
A. W.; Phillips, D.; Toner, W. T.; Towrie, M. Phys. Chem. Chem. Phys.
2003, 5, 1043-1050. (h) Saigusa, H.; Iwase, E.; Nishimura, M. J. Phys.
Chem. A 2003, 107, 3759-3763.
(1) (a) Marder, S. R.; Perry, J. W. Science 1994, 263, 1706-1707. (b) Verbiest,
T.; Burland, D. M.; Jurich, M. C.; Lee, V. Y.; Miller, R. D.; Volksen, W.
Science 1995, 268, 1604-1606. (c) Whitaker, C. M.; Patterson, E. V.; Kott,
K. L.; McMahon, R. J. J. Am. Chem. Soc. 1996, 118, 9966-9973. (d) Nandi,
P. K.; Mandal, K.; Kar, T. Chem. Phys. Lett. 2003, 381, 230-238.
(2) (a) Albota, M.; Beljonne, D.; Bre´das, J. L.; Ehrlich, J. E.; Fu, J.-Y.; Heikal,
A. A.; Hess, S. E.; Kogej, T.; Levin, M. D.; Marder, S. R.; McCord-
Maughon, D.; Perry, J. W.; Ro¨ckel, H.; Rumi, M.; Subramaniam, G.; Webb,
W. W.; Wu, X.-L.; Xu, C. Science 1998, 281, 1653-1656. (b) Kogej, T.;
Beljonne, D.; Meyers, F.; Perry, J. W.; Marder, S. R.; Bre´das, J. L. Chem.
Phys. Lett. 1998, 298, 1-6. (c) Reinhardt, B. A.; Brott, L. L.; Clarson, S.
J.; Dillard, A. G.; Bhatt, J. C.; Kannan, R.; Yuan, L.; He, G. S.; Prasad, P.
N. Chem. Mater. 1998, 10, 1863-1874. (d) Lee, W.-H.; Lee, H.; Kim,
J.-A.; Choi, J.-H.; Cho, M.; Jeon, S.-J.; Cho, B. R. J. Am. Chem. Soc. 2001,
123, 10658-10667. (e) Wang, X.; Zhou, Y.; Zhou, G.; Jiang, W.; Jiang,
M. Bull. Chem. Soc. Jpn. 2002, 75, 1847-1854.
(9) (a) Mennucci, B.; Toniolo A.; Tomasi, J. J. Am. Chem. Soc. 2000, 122,
10621-10630. (b) Parusel, A. B. J.; Rettig, W.; Sudholt, W. J. Phys. Chem.
A 2002, 106, 804-815. (c) Jo¨dicke, C. J.; Lu¨thi, H. P. J. Am. Chem. Soc.
2003, 125, 252-264. (d) Jo¨dicke, C. J.; Lu¨thi, H. P. J. Chem. Phys. 2003,
119, 12852-12865. (e) Rappoport, D.; Furche, F. J. Am. Chem. Soc. 2004,
126, 1277-1284.
(10) (a) Serrano-Andre´s, L.; Mercha´n, M.; Roos, B. O.; Lindh, R. J. Am. Chem.
Soc. 1995, 117, 3189-3204. (b) Jo¨dicke, C. J.; Lu¨thi, H. P. J. Chem. Phys.
2002, 117, 4146-4156. (c) Jo¨dicke, C. J.; Lu¨thi, H. P. J. Chem. Phys.
2002, 117, 4157-4167.
(11) (a) Sobolewski, A. L.; Domcke, W. Chem. Phys. Lett. 1996, 250, 428-
436. (b) Sobolewski, A. L.; Domcke, W. Chem. Phys. Lett. 1996, 259,
119-127.
(12) Schuddeboom, W.; Jonker, S. A.; Warman, J. M.; Leinhos, U.; Ku¨hnle,
W.; Zachariasse, K. A. J. Phys. Chem. 1992, 96, 10809-10819.
(13) (a) von der Haar, T.; Hebecker A.; Il’ichev, Y.; Jiang, Y.-B.; Ku¨hnle, W.;
Zachariasse, K. A. Recl. TraV. Chim. Pays-Bas 1995, 114, 430-442. (b)
Zachariasse, K. A.; Grobys, M.; von der Haar, T.; Hebecker A.; Il’ichev,
Y. V.; Jiang, Y.-B.; Morawski, O.; Ku¨hnle, W. J. Photochem. Photobiol.
A: Chem. 1996, 102, 59-70. (c) Zachariasse, K. A.; Grobys, M.; von der
Haar, T.; Hebecker A.; Il’ichev, Y. V.; Morawski, O.; Ru¨ckert, I.; Ku¨hnle,
W. J. Photochem. Photobiol. A: Chem. 1997, 105, 373-383. (d) Il’ichev,
Y. V.; Ku¨hnle, W.; Zachariasse, K. A. J. Phys. Chem. A 1998, 102, 5670-
5680. (e) Zachariasse, K. A. Chem. Phys. Lett. 2000, 320, 8-13. (f)
Demeter, A.; Zachariasse, K. A. Chem. Phys. Lett. 2003, 380, 699-703.
(g) Zachariasse, K. A.; Druzhinin, S. I.; Bosch, W.; Machinek, R. J. Am.
Chem. Soc. 2004, 126, 1705-1715.
(3) La Clair, J. J. Angew. Chem., Int. Ed. 1999, 38, 3045-3047.
(4) (a) Morozumi, T.; Anada, T.; Nakamura, H. J. Phys. Chem. B 2001, 105,
2923-2931. (b) Xiao, Y.; Qian, X. Tetrahedron Lett. 2003, 44, 2087-
2091. (c) Yang, J.-S.; Hwang, C.-Y.; Hsieh, C.-C.; Chiou, S.-Y. J. Org.
Chem. 2004, 69, 719-726. (d) Yang, J.-S.; Lin, Y.-D.; Lin, Y.-H.; Liao,
F.-L. J. Org. Chem. 2004, 69, 3517-3525.
(5) La Clair, J. J. Angew. Chem., Int. Ed. 1998, 37, 325-329.
(6) (a) Lippert, E.; Lu¨der, W.; Moll, F.; Na¨gele, W.; Boos, H.; Prigge, H.;
Seibold-Blankenstein, I. Angew. Chem. 1961, 73, 695-706. (b) Lippert,
E.; Lu¨der, W.; Boos, H. In AdVances in Molecular Spectroscopy; Mangini,
A., Ed.; Pergamon Press: Oxford, 1962; pp 443-457.
9
10.1021/ja047604d CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 12325-12335
12325

A R T I C L E S
Yang et al.
intramolecular charge transfer, WICT)12 or a planarization
(planar intramolecular charge transfer, PICT) of the amino
group.13 Among them, the TICT model has gained widespread
acceptance and appears to be the conclusion of the numerous
debates over the past decades.7-13
stilbenes (1) and the related model compounds 2-6, which have
provided a unique opportunity for gaining insights into the ICT
states of aminostilbenes. When compared with the N,N-dialkyl
derivatives (e.g., DS), the N-aryl-substituted 4-aminostilbenes
have inherently greater fluorescence quantum yields and longer
fluorescence lifetimes due to the prominent “amino conjugation
effect”.25,26 Thus, a breakdown of such an N-aryl conjugation
effect by twisting either a C-N or a C-C single bond should
result in a state with distinct fluorescence properties. Provided
that structural relaxations of 1 toward a TICT state are present
and become more efficient in more polar solvents, as proposed
for the case of DS,15 a dramatic change of the ICT fluorescence
of 1 on going from nonpolar to polar solvents should be
observed. Another advantage offered by N-aryl vs N-alkyl
systems is that the donor strength could be tuned to a large
extent by changing the N-aryl substituents, without significantly
changing the size of molecules. It has been shown that the donor
strength plays an important role in observing the TICT
fluorescence for DMABN and its analogues.10 More importantly,
our results turn out to have a direct linkage to the TICT
paradigm of DMABN.7-10 The seven N-aryl derivatives in 1
could be divided into three distinct categories in terms of their
solvent-dependent fluorescence properties. While the one con-
sisting of 1CN and 1CO displays a DMABN-like TICT
fluorescence, the other two categories (1OM vs 1Me, 1H, 1Cl,
and 1CF) lack dual fluorescence but offer an analogy with and
a difference from the first one in other excited-state properties,
respectively. In conjunction with the N-methyl (2) and the ring-
bridged model compounds (3-6), the substituent-dependent ICT
states for 1 will be elucidated and discussed.
Despite the well-documented photochemistry and photo-
physics of trans-stilbene and its derivatives,14 the nature of the
ICT state of aminostilbenes is still under active discussion.15-24
When compared with DMABN, the participation of the double-
bond torsion (trans-cis isomerization) in the excited-state
manifold and the lack of steady-state dual fluorescence for
aminostilbenes in both nonpolar and polar solvents often
complicate the data analysis. On the basis of theoretical
predictions and experimental comparisons with ring-bridged
model compounds,15-19 an emissive TICT state resulting from
the twisting of the anilino-styrenyl C-C single bond has been
proposed for N,N-dimethylaminostilbene (DS) and its cyano-
and nitro-substituted derivatives in polar solvents. However,
controversies were soon raised regarding the nature of the
emissive state as a planar or a twisted configuration and the
necessity of a TICT state in interpreting the photochemical
properties of aminostilbenes.20-24 For example, a spectral
evolution of transient fluorescence could be attributed to a
relaxation of the solvent cage instead of the conformational
relaxation of the excited molecule.22 In addition, a pronounced
ring-bridging effect on the fluorescence quantum yield could
result from a particular substituent effect on the reaction
hypersurface from the planar (1t*) to the double-bond twisted
(1p*) states (the two-state model) rather than the participation
of an additional TICT state (the three-state model).21 Evidently,
further studies are required to reach a satisfactory conclusion
on these issues for aminostilbenes.
We report herein the results of our systematic studies on the
excited-state properties of N-aryl-substituted trans-4-amino-
(14) (a) Saltiel, J.; Charlton, J. L. In Rearrangements in Ground and Excited
States; de Mayo, P., Ed.; Academic Press: New York, 1980; Vol. 3, pp
25-89. (b) Saltiel, J.; Sun, Y.-P. In Photochromism, Molecules and Systems;
Du¨rr, H., Bouas-Laurent, H., Eds.; Elsevier: Amsterdam, 1990; pp 64-
164. (c) Waldeck, D. H. Chem. ReV. 1991, 91, 415-436. (d) Go¨rner, H.;
Kuhn, H. J. AdV. Photochem. 1995, 19, 1-117.
(15) Le´tard, J.-F.; Lapouyade, R.; Rettig W. J. Am. Chem. Soc. 1993, 115, 2441-
2447.
(16) (a) Rettig, W.; Majenz, W. Chem. Phys. Lett. 1989, 154, 335-341. (b)
Lapouyade, R.; Czeschka, K.; Majenz, W.; Rettig, W.; Gilabert, E.; Rullie`re,
C. J. Phys. Chem. 1992, 96, 9643-9650. (c) Abraham, E.; Oberle´, J.;
Jonusauskas, G.; Lapouyade, R.; Rullie`re, C. J. Photochem. Photobiol. A:
Chem. 1997, 105, 101-107. (d) Papper, V.; Pines, D.; Likhtenshtein, G.;
Pines, E. J. Photochem. Photobiol. A: Chem. 1997, 111, 87-96. (e)
Abraham, E.; Oberle´, J.; Jonusauskas, G.; Lapouyade, R.; Rullie`re, C. Chem.
Phys. 1997, 214, 409-423. (f) Pines, D.; Pines, E.; Rettig, W. J. Phys.
Chem. A 2003, 107, 236-242.
Results
Synthesis and Molecular Structure. The synthesis of
aminostilbenes 1 is straightforward, based on palladium-
catalyzed amination reactions27 between trans-4-bromostilbene
and the corresponding commercially available 4-substituted
(17) (a) Amatatsu, Y. Theor. Chem. Acc. 2000, 103, 445-450. (b) Amatatsu,
Y. Chem. Phys. 2001, 274, 87-98.
(18) (a) Gruen, H.; Go¨rner, H. J. Phys. Chem. 1989, 93, 7144-7152. (b)
Lapouyade, R.; Kuhn, A.; Le´tard, J.-F.; Rettig, W. Chem. Phys. Lett. 1993,
208, 48-58.
(24) (a) Lewis, F. D.; Kalgutkar, R. S.; Yang, J.-S. J. Am. Chem. Soc. 1999,
121, 12045-12053. (b) Lewis, F. D.; Weigel, W. J. Phys. Chem. A 2000,
104, 8146-8153. (c) Lewis, F. D.; Weigel, W.; Zuo, X. J. Phys. Chem. A
2001, 105, 4691-4696.
(19) Le´tard, J.-F.; Lapouyade, R.; Rettig, W. Chem. Phys. Lett. 1994, 222, 209-
216.
(20) Seydack, M.; Bendig, J. J. Phys. Chem. A 2001, 105, 5731-5733.
(21) (a) Il’ichev, Y. V.; Ku¨hnle, W.; Zachariasse, K. A. Chem. Phys. 1996,
211, 441-453. (b) Kovalenko, S. A.; Schanz, R.; Senyushkina, T. A.;
Ernsting, N. P. Phys. Chem. Chem. Phys. 2002, 4, 703-707.
(22) Eilers-Ko¨nig, N.; Ku¨hne, T.; Schwarzer, D.; Vo¨hringer, P.; Schroeder, J.
Chem. Phys. Lett. 1996, 253, 69-76.
(25) Yang, J.-S.; Chiou, S.-Y.; Liau, K.-L. J. Am. Chem. Soc. 2002, 124, 2518-
2527.
(26) Yang, J.-S.; Wang, C.-M.; Hwang, C.-Y.; Liau, K.-L.; Chiou, S.-Y.
Photochem. Photobiol. Sci. 2003, 2, 1225-1231.
(27) (a) Wolfe, J. P.; Wagaw, S.; Marcoux, J.-F.; Buchwald, S. L. Acc. Chem.
Res. 1998, 31, 805-818. (b) Hartwig, J. F. Angew. Chem., Int. Ed. 1998,
37, 2046-2067.
(23) Rechthaler, K.; Ko¨hler, G. Chem. Phys. Lett. 1996, 250, 152-158.
9
12326 J. AM. CHEM. SOC. VOL. 126, NO. 39, 2004

Substituent-Dependent Photoinduced Charge Transfer
A R T I C L E S
Scheme 1
Scheme 2
Figure 1. X-ray crystal structures of (a) 1CN, (b) 2CN, and (c) 1CO
(double bond disordered).
Scheme 3
anilines (Scheme 1). Typical procedures have been previously
reported for the synthesis of 1H.25
Model compounds 2-6 possessing a methoxy (e.g., 2OM)
or a cyano (e.g., 2CN) substituent also have been prepared.
Compounds 2 were prepared by N-methylation28 of the corre-
sponding compounds 1. Except for the cases of 4, the synthetic
methodology for the bridged derivatives is similar to that for 1,
namely, via a Pd-catalyzed coupling of bromostilbenes and
anilines with the desired bridged structures in either substrates.
The 5-methoxyindoline (7) for the synthesis of 3OM was
prepared from the reduction of 5-methoxyindole by sodium
cyanoborohydride.29 The corresponding precursor for compound
3CN is 5-bromoindoline (8), and the amination reaction was
followed by a cyanization reaction30 that converts the bromo
group to a cyano group. The bromostilbene 9 for the formation
of 5OM and 5CN could, in principle, be prepared by a
nucleophilic addition of benzylmagnesium bromide to 5-bromo-
1-indanone, followed by acid-catalyzed dehydration (Scheme
2), a method analogous to the synthesis of 1-benzalindane.31
However, the separation of 9 from its isomer 10 is not feasible.
Whereas a pure sample of the final product 5OM is not yet
available,32 compound 5CN could be readily purified by column
chromatography. The synthesis of compound 6H has recently
been reported,26 and the same intermediate 11 was employed
for the formation of 6OM and 6CN.
The X-ray crystal structures of 1CO, 1CN, and 2CN have
been determined (Figure 1). The stilbene moiety is essentially
planar in all three cases. Whereas the N-phenyl and the terminal
styrenyl groups in 2CN are located on the same side (the syn
conformation) with respect to the long molecular axis, they are
on the opposite side (the anti conformation) in 1CN, and there
is a syn-anti disorder in the case of 1CO. These observations
are in accord with our previous conformational analysis for
aminostilbenes 1H, 2H, and 6H, where the syn and the anti
conformers were calculated to be of similar energies, with a
difference less than 0.3 kcal/mol.25,26 Our previous results also
suggested a shallow minimum for their ground-state potential
energy surfaces, and in solutions there should exist a large
distribution of conformers with varied ring-ring torsional and
amino wagging angles.26 This is more likely a common feature
for all aminostilbenes 1-6.
Electronic Spectra. All the aminostilbenes 1-6 in hexane
and acetonitrile display a single intense long-wavelength absorp-
tion band. The corresponding absorption maxima (λabs) are
reported in Table 1, and typical spectra are presented in Figure
2 for 1OM and 1CN. For comparison, the data of DS are also
included. In general, an electron-donating (ED) or electron-
withdrawing (EW) substituent at the para position of the
N-phenyl group shifts the absorption maximum to the red (e.g.,
1OM vs 1H) and to the blue (e.g., 1CN vs 1H), respectively.
In addition, the spectra in acetonitrile are bathochromic,
hypochromic, and broadened when compared with those in
hexane. However, the solvatochromic shift is rather small, which
As shown in Scheme 3, the syntheses of 4OM and 4CN
started with the N-arylation of indoline, followed by a formyl-
ation33 at the 5-position of indoline. A conventional Horner-
Wadsworth-Emmons reaction34 then led to the desired ami-
nostilbenes 4.
(28) Borch, R. F.; Hassid, A. I. J. Org. Chem. 1972, 37, 1673-1674.
(29) Gangjee, A.; Vasudevan, A.; Queener, S. F. J. Med. Chem. 1997, 40, 479-
485.
(30) Ikan, R.; Rapaport, E. Tetrahedron 1967, 23, 3823-3827.
(31) Plentl, A. A.; Bogert, M. T. J. Am. Chem. Soc. 1941, 63, 989-995.
9
J. AM. CHEM. SOC. VOL. 126, NO. 39, 2004 12327

A R T I C L E S
Yang et al.
Table 1. Maxima of UV Absorption (λabs) and Fluorescence (λf),
Fluorescence Band Half-Width (∆ν1/2), 0,0 Transition (λ0,0), and
Stokes Shifts (∆νst) of Aminostilbenes 1-6 and DS in Hexane
(Hex) and Acetonitrile (MeCN)a
wavelength band can be attributed to electronic transitions
delocalized throughout the whole molecule with a charge-
transfer character of mainly the HOMO (amino nitrogen) f
LUMO (stilbene) transition.25,26 On the basis of the results of
semiempirical INDO/S-SCF-CI (ZINDO) calculations35 for the
AM1-optimized36 structures of 1OM and 1CN, the introduction
of substituents at the N-phenyl group has only a small effect
on the contribution of the HOMO f LUMO configuration to
the description of the lowest excited singlet state (S1) (i.e., ∼85-
86% for 1OM and 1CN as compared with ∼89% for 1H).
The fluorescence spectra of all aminostilbenes 1-6 are
structured in hexane but become less structured in toluene and
completely structureless in more polar solvents. In addition,
unlike the absorption spectra, the fluorescence spectra shift
significantly to the red with increasing solvent polarity, indicat-
ing a strong ICT character for the fluorescent states (1t*) of
these aminostilbenes. The fluorescence maxima (λf), the half-
bandwidth (∆ν1/2), the 0,0 transitions (λ0,0), and the Stokes shift
(∆νst) of 1-6 in hexane and acetonitrile are reported in Table
1.
For the compound series 1, the size of the solvatochromic
shifts and the shape of the fluorescence spectra in polar solvents
strongly depend on the substituent. As shown in Figure 3, the
structureless fluorescence spectra are in a normal Gaussian shape
for 1H, 1Me, 1Cl, and 1CF, but they are significantly broadened
for 1OM and become dual fluorescent for 1CO and 1CN in
solvents more polar than THF. It should be noted that an
emission longer than 650 nm would be less accurate due to the
limitation of our instrument. When the short emission bands
for 1CO and 1CN were taken into account, the magnitude of
the shift on going from hexane to acetonitrile is in the order
1OM > 1Me > 1H > 1Cl > 1CF > 1CN ∼ 1CO, which is
roughly parallel with the relative electron-donating ability of
the substituents. It is interesting to note that the energies of the
fluorescence maxima of 1 correlate better with the Hammett
σ+ than with the σ constants (Figure 4).37 In hexane, a nice
linear relationship could be observed for 1, except for 1CO and
1CN. The lower-than-predicted fluorescence peak energies for
the latter two species could be attributed to the conjugation effect
of the π-substituents. While the linear correlation among 1Me,
1H, 1Cl, and 1CF is retained in dichloromethane and acetoni-
trile, the data points for 1OM become off the line, which is
more severe in acetonitrile than in dichloromethane. It should
also be noted that both the fluorescence excitation and emission
spectra of 1 in hexane and acetonitrile are essentially indepen-
dent of the emission and excitation wavelengths, respectively,
and the former closely follow the absorption spectra.38 Evidently,
the conformers in 1 should have similar electronic properties,
as previously observed for 1H,25 and the dual emitters in 1CN
and 1CO should result from the same FC state.
λabs
(nm)
λf
(nm)b
∆ν1/
λ0,0
(nm)c
∆νst
2
1
1
d
compd
solvent
(cm-
)
(cm-
)
1H
Hex
MeCN
Hex
MeCN
Hex
MeCN
Hex
MeCN
Hex
MeCN
Hex
MeCN
Hex
MeCN
Hex
MeCN
Hex
MeCN
Hex
MeCNf
Hex
MeCN
Hex
MeCN
Hex
MeCN
Hex
MeCN
Hex
346
351
349
354
344
351
340
347
349
356
348
358
341
352
350
356
335
342
374
378
357
369
367
373
366
379
345
359
345
351
347
354
347
351
381 (399)
442
385 (402)
457
382 (399)
437
376 (395)
421
2956
3547
2808
3866
2913
3582
2775
3428
2782
6218
2526
8687
2712
8086
2946
370
393
374
398
369
390
366
383
378
398
373
388
370
387
386
382
379
397
403
2655
5785
2679
6367
2892
5607
2816
5065
2946
8170
2694
4292
3079
4880
3941
10759
4662
10859
4011
1Me
1Cl
1CF
1OM
1CO
1CN
2OM
2CN
3OM
3CN
4OM
4CN
5CN
6OM
6CN
DSh
389 (409)
502
384 (403)
425 [530]
381 (399)
425 [517]
406 (425)
577
397 (412)
[539]e
440
9658g
2851
5052
3856
391 (412)
438
408 (430)
483
407 (420)
[567]e
392 (406)
[535]e
385 (405)
472
391 (405)
[530]e
2147
3564
2362
3429
2759
5931g
3141
6019g
3117
6312
3129
5804g
3425
3306
383
400
398
423
393
408
376
392
375
389
375
388
2436
4269
2738
6106
2752
8749
3475
9164
3669
7304
3243
9381
2433
5763
MeCN
Hex
MeCN
Hex
379
440
MeCN
a Fluorescence data are from corrected spectra. b The second vibronic
band is given in parentheses, and the long-wavelength emission band is
given in brackets. c The value of λ0,0 was obtained from the intersection of
normalized absorption and fluorescence spectra. d ∆νst ) νabs - νf. e The
short-wavelength emission could not be resolved. f Fluorescence too weak
to be reliably determined. g Values are estimated due to the incomplete
spectra. h Data from ref 15.
The dipole moment (µe) of the fluorescent state can be
estimated from the slope (mf) of the plot of the energies of the
(32) Compound 5OM is weakly fluorescent and very sensitive to the room light,
and more than one photoproduct has been detected by HPLC.
(33) Jutz, C. AdV. Org. Chem. 1976, 9, 225-342.
Figure 2. UV-vis absorption spectra of 1OM (curves a and b) and 1CN
(curves c and d) in hexane (curves a and c) and acetonitrile (curves b and
d).
(34) Wadsworth, W. S., Jr. Org. React. 1977, 25, 73-253.
(35) Zerner, M. C.; Leow, G. H.; Kirchner, R. F.; Mueller-Westerhoff, U. T. J.
Am. Chem. Soc. 1980, 102, 589-599.
(36) Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J. J. P. J. Am.
Chem. Soc. 1985, 107, 3902-3909.
indicates a small difference between the dipole moments of the
Franck-Condon (FC) excited state and the ground state.
According to our previous studies on 1H, 2H, and 6H, the long-
(37) Lowry, T. H.; Richardson, K. S. Mechanism and Theory in Organic
Chemistry; Harper & Row: New York, 1987; p 144.
(38) See Supporting Information for details.
9
12328 J. AM. CHEM. SOC. VOL. 126, NO. 39, 2004

Substituent-Dependent Photoinduced Charge Transfer
A R T I C L E S
Figure 4. Correlation diagram of the energies of the fluorescence maxima
against the Hammett σ+ constants for 1 in hexane, dichloromethane, and
acetonitrile. The short-wavelength emission maxima are adopted for 1CN
and 1CO.
Table 2. Ground- and Excited-State Dipole Moments for 1, 3CN,
4OM, and DS
1
b
compd
a (Å)a
mf (cm-
)
µ
g (D)c
µe (D)
1OM
1Me
1H
4.93
4.84
4.76
4.66
4.96
5.07
4.90
5.04
5.06
4.53
14351
0.98
0.94
0.84
1.70
4.37
2.88
4.25
4.21
0.18
2.41
13.6
11493
9973
9851
8346
7880 (17132)
8672 (17106)
7329
11.8
10.8
1Cl
9.5d
1CF
1COe
1CNe
3CN
4OM
DSf
9.0d
9.4d (16.4)
9.1d (16.4)
8.7d
10467
12335
11.7
11.8
a Onsager radius calculated by eq 3 with d ) 1.0 for 1OM, 1Me, 1H,
1CO, 1CN, 3CN, and 4OM, 1.1 for 1CF, and 1.2 g/cm3 for 1Cl.
b Calculated on the basis of eq 1. c Calculated by AM1. d Calculated with
-0.5µg. e The values for the long-wavelength emission state are given in
parentheses. f Data from ref 15.
could be derived from the Avogadro number (N), molecular
weight (M), and density (d), and ꢀ, ꢀ0, and n are the solvent
dielectric, vacuum permittivity, and the solvent refractive index,
respectively. The value of µg was calculated using the MOPAC-
AM1 algorithm. The calculated ground-state dipole moment is
1CF > 1CN > 1CO > 1Cl > 1OM > 1Me > 1H, and the
dipole moment is essentially oriented toward the N-stilbenyl
and the N-aryl group for the ED- and the EW-substituted species,
respectively. Assuming that the angle between the ground- and
the excited-state dipoles of the EW-substituted derivatives is
ca. 120°, the calculated ground-state dipole would have a
component vector of 0.5µg (i.e., cos(120°)µg) in a direction
opposite to that of the N f stilbene excited-state dipole.
Accordingly, a negative value of -0.5µg was adopted for eq 1
in calculating the µe for 1Cl and 1CF and the short-wavelength
bands of 1CO and 1CN. In contrast, the long-wavelength
emitting states of 1CO and 1CN are more likely to have a dipole
in the same direction as that for the ground-state dipoles (vide
infra). These results, along with the corresponding data for DS,
are summarized in Table 2.
Figure 3. Normalized fluorescence spectra of aminostilbenes 1 in (a)
hexane, (b) toluene, (c) THF, (d) dichloromethane, (e) acetone, and (f)
acetonitrile.
fluorescence maxima against the solvent parameter ∆f according
to eq 1:15,38,39
νf ) -[(1/4πꢀ0)(2/hca3)] [µe(µe - µg)] ∆f + constant (1)
where
The dependence of the fluorescence spectra of 1OM and 1CN
in hexane and acetonitrile on temperature has been investigated
(Figure 5). In hexane, the fluorescence intensity of 1OM is
significantly reduced upon heating from -40 to 50 °C, but it is
insensitive to the change in temperature for 1CN. The results
∆f ) (ꢀ - 1)/(2ꢀ + 1) - 0.5(n2 - 1)/(2n2 + 1) (2)
and
a ) (3M/4Nπd)1/3
(3)
(39) (a) Liptay, W. Z. Z. Naturforsch. 1965, 20a, 1441. (b) Lippert, E. Z.
Elecktrochem. 1957, 61, 962-975. (c) Mataga, N.; Kaifu, Y.; Koizumi,
M. Bull. Chem. Soc. Jpn. 1956, 29, 465-470.
where νf is the fluorescence maximum, µg is the ground-state
dipole moment, a is the solvent cavity (Onsager) radius, which
9
J. AM. CHEM. SOC. VOL. 126, NO. 39, 2004 12329

A R T I C L E S
Yang et al.
Figure 6. Normalized fluorescence spectra of aminostilbenes 3CN and
4OM in (a) hexane, (b) toluene, (c) THF, (d) dichloromethane, (e) acetone,
and (f) acetonitrile.
Figure 5. Temperature dependence of the fluorescence spectra of 1OM
in (a) hexane and (b) acetonitrile and of 1CN in (c) hexane and (d)
acetonitrile recorded at intervals of 10 °C between -40 and 50 °C. The
arrows indicate the direction of fluorescence response upon raising the
temperature.
are different in acetonitrile, where the fluorescence is enhanced
and blue-shifted for both 1OM and 1CN upon raising the
temperature. Previous studies on 1H have shown that its
fluorescence intensity undergoes a monotonic decrease with
increasing temperature in both hexane and acetonitrile.25
Evidently, the excited-state behavior of aminostilbenes 1 is
significantly affected by the N-aryl substituent.
When compared with 1OM and 1CN, the additional N-methyl
or methylene bridging group in 2-6 affects the electronic spectra
more or less, depending on the nature of the N-aryl group and
the position of the methylene group (Table 1). For instance,
the introduction of an N-methyl group shifts the absorption
maxima of 1CN to the blue (i.e., 2CN), but it has only a small
effect on that of 1OM (i.e., 2OM). In addition, the ring-bridged
species have longer wavelength λabs and λf than the correspond-
ing nonbridged compounds 1 and 2, but compound 6OM is an
exception. The most intriguing observation is probably the
narrower fluorescence shape for 3CN and 4OM in comparison
to the other analogues in polar solvents (Figure 6). In other
words, the broad and largely Stokes-shifted fluorescence spectra
for 1OM and 1CN are also observed for 2OM-6OM and
2CN-6CN in dichloromethane (Figure 7) and acetonitrile, but
the solvatofluorochromic shifts and the excited-state dipole
moments for 3CN and 4OM are more like those for 1CF and
1Me (Table 2), respectively.
Quantum Yields and Lifetimes. The fluorescence quantum
yields (Φf) for aminostilbenes 1-6 in a variety of solvents have
been determined, and some of these data are reported in Table
3. For the compound series 1, all derivatives are strongly
fluorescent in hexane and toluene, but the values of Φf decrease
in more polar solvents. The plots of Φf against the microscopic
solvent polarity parameter, ET(30),40 are shown in Figure 8.
Whereas the group consisting of 1H, 1Me, 1Cl, and 1CF
Figure 7. Normalized fluorescence spectra of (a) 2OM, 3OM, and 6OM
and (b) 2CN and 4CN-6CN in dichloromethane. For comparison, the
spectra of 6OM in acetonitrile (the segmented curve in (a)) are included.
follows an essentially linear correlation (Figure 8a), it is a
sigmodial curve for the group of 1OM, 1CO, and 1CN (Figure
8b). It is interesting to note that all the substituted derivatives,
regardless of ED or EW substituents at the N-phenyl group,
have a larger value of Φf than does 1H in nonpolar solvents.
Quantum yields for trans f cis photoisomerization (Φtc) for
the compound series 1 in THF and dichloromethane are reported
in Table 3. Assuming that the decay of the double bond twisted
perpendicular excited state (p*) yields a 1:1 ratio of trans and
cis isomers,14 the sum of the fluorescence and double bond
torsion quantum yields (Φf + 2Φtc) for 1H, 1Me, 1Cl, and 1CF
in both solvents is within the experimental error of 1.0, but it
is relatively lower than 1.0 for 1OM, 1CO, and 1CN in THF
and even lower in dichloromethane. Evidently, fluorescence and
photoisomerization could account for the excited decay of the
former cases, but other channels of nonradiative decay should
be taken into account for the latter three compounds in polar
solvents.
The fluorescence lifetimes (τf) of the aminostilbenes 1 in
different solvents are provided in Table 3. All decays can be
well fit by single-exponential functions, although more than one
conformer is expected for 1 in solutions, and there is dual
fluorescence for 1CN and 1CO in polar solvents. Nonetheless,
as demonstrated in dichloromethane, the values of τf for 1CO
(40) Marcus, Y. Chem. Soc. ReV. 1993, 409-416.
9
12330 J. AM. CHEM. SOC. VOL. 126, NO. 39, 2004

Substituent-Dependent Photoinduced Charge Transfer
A R T I C L E S
Table 3. Quantum Yields for Fluorescence (Φf) and
Photoisomerization (Φtc), Fluorescence Decay Times (τf), Rate
Constants for Fluorescence Decay (kf), and Nonradiative Decay
(knr) for 1-6 and DS in Solutions
τf
(ns)a
kf
knr
(108 s-
1
1
compd
solvent
Φf
Φtc
(108 s-
)
)
1H
Hex
THF
CH2Cl2
MeCN
Hex
0.51
0.42
0.38
0.34
0.53
0.55
0.49
0.41
0.58
0.53
0.48
0.43
0.68
0.49
0.43
0.25
0.60
0.36
0.24
0.007
0.72
0.38
0.03
0.004
0.75
0.46
0.11
0.015
0.65
0.005
0.68
0.02
0.81
<0.001
0.81
0.81
0.81
0.75
0.66
0.68
0.78
0.014
0.056
0.005
0.84
0.031
0.87
0.007
0.030
0.037
0.24b
0.24
0.34
0.89
0.87
0.70
0.96
0.96
1.31
1.15
1.75
1.00
0.96
0.75
0.96
1.07
0.84
0.67
0.67
1.20
2.44
2.40
0.32
1.22
2.19
0.94d
0.41
1.30
2.70
2.90d
0.84
2.54
0.56
2.52
0.98
3.00
5.7
4.8
5.4
3.5
5.1
4.2
4.3
2.3
5.8
5.5
6.4
4.5
6.4
5.8
6.4
3.7
5.0
1.5
1.0
5.5
6.7
8.9
6.9
4.9
3.4
4.4
3.4
4.2
4.9
6.9
5.9
3.0
6.1
8.5
11.2
3.3
2.6
1Me
1Cl
THF
0.28
0.28
CH2Cl2
MeCN
Hex
THF
0.27
0.29
CH2Cl2
MeCN
Hex
Figure 8. Plots of the fluorescence quantum yield (Φf) against the solvent
parameter ET(30) for the compound series 1. The solvents [ET(30)] are (from
left to right) hexane [31.0], toluene [33.9], THF [37.4], dichloromethane
[40.7], and acetonitrile [45.6] for both plots (a) and (b), and three more
solvents, 1,4-dioxane [36.0], chloroform [39.1], and DMF [43.8], are added
for plot (b).
1CF
1OM
1CO
1CN
THF
0.18
0.22
CH2Cl2
MeCN
Hex
0.10b
0.05
0.09
THF
CH2Cl2
MeCN
Hex
3.2
31.0
2.3
0.22
5.9
1.7
0.32
0.10
5.8
THF
0.11
0.04
2.8
CH2Cl2
MeCN
Hex
10.3
24.3
1.9
2.0
3.1
11.7
1.4
17.7
1.4
10.0
0.63
0.16c
0.16
0.14
THF
1.7
CH2Cl2
MeCN
Hex
MeCN
Hex
MeCN
Hex
MeCN
Hex
CH2Cl2
MeCN
Hex
CH2Cl2
MeCN
Hex
MeCN
Hex
MeCN
Hex
0.38
0.18
2.6
0.18
2.6
2OM
2CN
3OM
3CN
0.20
2.7
1.43
1.45
1.71
1.54
1.91
2.36
1.56
0.38
0.12
<0.10
1.20
0.87
1.50
0.11
0.1f
5.7
5.6
4.7
4.0
3.5
2.4
5.0
0.37
4.7
>0.5
6.3
0.36
5.8
0.64
3.0
1.3
1.3
1.1
2.5
1.8
1.8
1.4
0.02
0.25
Figure 9. Simplified scheme for the formation and decay of the fluorescent
ICT state of trans-4-(N-arylamino)stilbenes 1. The CSt-N and CAr-N bonds
denote the stilbenyl-anilino and the aryl-anilino C-N bonds, respectively.
4OM
4CN
5CN
6OM
6CN
DSe
conversion is unimportant in accounting for the decay of 1H,
1Me, 1Cl, and 1CF.
25.9
78.7
>99.5
8.3
11.1
0.87
90.3
97
96
The values of Φf, τf, kf, and knr for 2-6 in hexane and
acetonitrile are also reported in Table 3. In accord with the
observations based on fluorescence spectra, the large values of
Φf and kf in acetonitrile for 3CN and 4OM differentiate them
from the other cyano- and methoxy-substituted derivatives. For
comparison, the values of Φtc for 3CN and 4OM in dichlo-
romethane were also determined. It should be noted that the
restriction of the CdC double bond torsion does not prevent
6CN and 6OM from having a low Φf value in acetonitrile. In
addition, the fluorescence quantum yield is low for 5CN even
in hexane. Nonetheless, the solvent effect on its Φf is still
significant.
MeCN
Hex
MeCN
Hex
MeCN
0.1f
3.7
a The value of τf was determined with excitation and emission around
the spectral maxima, unless otherwise noted. b Containing 10% of THF by
reason of solubility. c Containing 12% of THF by reason of solubility. d An
averaged value is adopted. e Data from ref 15. f Data determined in diethyl
ether and ethanol are both 0.1 ns.
(0.63-1.11 ns) and 1CN (2.31-3.50 ns) are dependent on the
emission wavelength and are larger at longer wavelength
emissions, a phenomenon not observed for the other derivatives
Discussion
of 1. In addition, the fluorescence rate constants (kf ) Φf τf-1
)
A generalized scheme for the photochemical behavior of
aminostilbenes 1 is shown in Figure 9. The substituents play
an important role in determining the relative energy levels of
the excited states and thus the resulting decay pathways. The
seven derivatives could be divided into three different groups
(I-III), based on their excited-state behavior, which are
discussed in the following.
are comparable for all seven compounds of 1 in hexane, but
they are more than 1 order of magnitude smaller for 1OM, 1CO,
and 1CN than the others in acetonitrile. The overall nonradiative
deactivation (knr ) 1/τf - kf) was also calculated (Table 3).
The absence of specific correlation between the log knr and the
emission energy (figure not shown) indicates that internal
9
J. AM. CHEM. SOC. VOL. 126, NO. 39, 2004 12331

A R T I C L E S
Yang et al.
1
the p* state of 1H (eq 5). As a result, a further enhancement
of its contribution to 1p* by ED R substituents should be rather
limited; however, a slight weakening of its contribution due to
the presence of an EW substituent could significantly raise the
1
energy of the p* state.
Figure 10. Qualitative representation of the substituent effect on the barrier
1
1
of t* f p* double bond torsion in the singlet excited state.
Excited-State Behavior of 1H, 1Me, 1Cl, and 1CF (Group
I). The photochemical behavior of 1H has recently been
reported.25,26 Like the parent trans-4-aminostilbene,24 the decay
of excited 1H is mainly via fluorescence and the double bond
torsion reaction (i.e., Φf + 2Φtc ≈ 1.0). However, 1H displays
much higher fluorescence quantum yields and lower trans f
cis isomerization quantum yields as a consequence of a larger
torsional barrier (Ea) in the singlet excited state. An analysis of
the vibrational structures observed in hexane and the correlations
between fluorescence maxima and solvent polarity further
suggested that the molecule becomes more planar in the excited
1
state, and such a planar t* (ICT) state is responsible for the
Excited-State Behavior of 1CN and 1CO (Group II). The
phenomenon of dual fluorescence observed for 1CN and 1CO
in dichloromethane and more polar solvents differentiates them
from the other derivatives of 1. The location of the long-
wavelength emission band (λf ) 517 nm in acetonitrile) and
the corresponding value of µe (16.4 D) for 1CN are reminiscent
of the TICT fluorescence of DMABN (λf ) 485 nm in
acetonitrile and µe ) 17 D).6,13 Further, it has recently been
reported that the TICT state of ethyl 4-(N-phenylamino)-
benzoate, an analogue of 1CO, dominates the fluorescence in
polar solvents (e.g., λf ) 486 nm in acetonitrile and µe ) 17.7
D).41 Indeed, as is demonstrated by the bridged model compound
3CN, restriction of the rotation of the anilino-benzonitrilo C-N
single bond removes the long-wavelength emission band and
thus the dual fluorescence property. Moreover, 3CN is strongly
fluorescent in both polar and nonpolar solvents, in analogy to
the behavior of the group I species. On the other hand, the long-
wavelength emission band is retained and even enhanced for
4CN-6CN (Figure 7), indicating that the rotation of the other
single bonds plays a negligible role in accounting for the TICT
state. Evidently, the formation of a DMABN-like TICT state is
responsible for the low quantum yields of fluorescence for 1CN
in polar solvents. The weak emission property observed for the
TICT state is consistent with the forbidden nature of the
fluorescence. In addition, a broad TICT emission is in accord
with the picture of a broad distribution of conformers with varied
twisted angles around the C-N single bond. Since the photo-
isomerization quantum yield is also low for 1CN in dichlo-
romethane, the decay process of TICT f 1p* or 3p* should be
unimportant, and the nonradiative decay of the TICT state might
be mainly via internal conversion, as proposed for the other
TICT systems.20,42 The similarity in photochemical behavior
between 1CO and 1CN also agrees with the common feature
observed fluorescence in both nonpolar and polar solvents.26
Values of kf decrease with increasing solvent polarity, but the
corresponding value of knr is slightly larger in acetonitrile vs
hexane (Table 3). These changes account for the decrease in
Φf with increasing solvent polarity.
The photochemical behavior of 1Me, 1Cl, and 1CF is similar
to that of 1H: namely, other decay channels except for
fluorescence and the double bond torsion are unimportant. It is
interesting to note that both ED and EW substituents at the para
position of the N-phenyl group enhance the fluorescence
quantum yield (Table 3). Since the values of kf are comparable
for all four compounds in the same solvent, the origin of
fluorescence enhancement could be attributed to an increase in
the torsional barrier that decreases the rate of isomerization and
thus the values of knr and Φtc. According to the 0-0 transition
1
energies (Table 1), the t* state of 1H is stabilized by ED
substituents (e.g., 1Me) but destabilized by EW substituents
(e.g., 1CF). To result in a larger torsional barrier for both cases,
1
the p* state should be less stabilized in the former and more
destabilized in the latter cases, as depicted in Figure 10. In other
1
1
words, the substituent effect on the p* vs the t* state of 1H
should be different. This is indeed in accord with the model of
resonance structures that was previously employed for the
rationalization of the amino conjugation effect in N-phenyl- vs
1
N-alkyl-substituted 4-aminostilbenes.25 For the t* state, the
amino lone pair electrons could delocalize to either the stilbene
(resonance structure A) or the phenyl ring (resonance structure
B) (eq 4). In principle, the importance of structure A would be
enhanced by ED substituents but diminished by EW substituents,
and the opposite would be true for structure B. Since structure
A corresponds to the HOMO f LUMO transition, such a
substituent effect agrees with the red vs blue shifts of the
absorption spectra (Table 1) and the larger vs smaller values of
µe for 1Me and 1CF vs 1H, respectively. In contrast, the
corresponding resonance structure A should be predominant in
(41) Ma, L.-H.; Chen, Z.-B.; Jiang, Y.-B. Chem. Phys. Lett. 2003, 372, 104-
113.
9
12332 J. AM. CHEM. SOC. VOL. 126, NO. 39, 2004

Substituent-Dependent Photoinduced Charge Transfer
A R T I C L E S
of dual fluorescence in both methyl 4-(N,N-dimethylamino)-
benzoate (DMABME) and DMABN.8,10
Excited-State Behavior of 1OM (Group III). Despite the
absence of well-resolved dual fluorescence in the steady-state
spectra, the great similarity in the solvent- and temperature-
dependent fluorescence behavior between 1OM and the group
II compounds indicates the presence of a TICT state for 1OM
in polar solvents. For example, the values of Φf for 1OM also
exhibit a sigmodial relationship with ET(30). In addition, the
sum of Φf + 2Φtc is also much less than 1.0 in dichloromethane
but not in hexane. Furthermore, the fluorescence in acetonitrile
is also enhanced upon heating. Therefore, the observed fluo-
rescence could result from an overlap of the emission of the
1t* and the TICT state, for two reasons: (1) the fluorescence
band becomes much narrower when the TICT state formation
is blocked (i.e., 4OM, vide infra), and (2) the correlation of the
fluorescence peak energies and the Hammett σ+ constants
The presence of a TICT state in 1CN also accounts for the
temperature-dependent fluorescence spectra (Figure 5). For
stilbenes such as the group I molecules that conform to the two-
state mechanism (1t* and 1p*), the fluorescence intensity would
be either decreased or nearly unchanged upon increasing the
temperature, corresponding to the activated and unactivated
processes of the double bond torsion in the singlet and the triplet
excited state, respectively.14 When there is an equilibrium TICT
a 1t*, the activated process TICT f 1t* could compensate for
1
1
1
the depopulation of t* in the activated process t* f p* as
long as the latter has a larger barrier (i.e., Ea > Ea′, Figure 9).
In fact, the double bond torsion in 1CN is more likely via the
unactivated triplet mechanism, presumably due to a large Ea
value (vide infra). As a result, the model of fluorescence kinetics
previously derived for DMABN7,13 could also apply to the case
of 1CN. Assuming that the change in fluorescence peak intensity
is proportional to that in Φf, the values of both Φf(1t*) and Φf-
(TICT) are increased but the ratio Φf(TICT)/Φf(1t*) is decreased
upon increasing the temperature for 1CN in acetonitrile (Figure
5). This is indeed consistent with an equilibrium between the
two states in the temperature range of -40 to 50 °C (i.e., kd >
kf′). A quantitative treatment of the spectra in Figure 5d was
not performed due to the poorly resolved and incomplete
fluorescence spectra.
1
predicts a less red-shifted maximum for the t* fluorescence
on going from hexane to dichloromethane and acetonitrile
(Figure 4). Accordingly, the two states are both highly polar
and the µe value reported in Table 2 for 1OM should be
considered as an average of their dipole moments. It should
also be noted that, unlike the case of 1CN, the photoisomer-
ization of 1OM is more likely involved with the singlet-state
pathway, because the fluorescence decreases monotonically in
hexane with increasing temperature (Figure 5a).
A comparison of the fluorescence behavior of 1OM-4OM
and 6OM has offered a clue to the single bond that is responsible
for the TICT state formation in 1OM. A common feature
observed for 1OM-3OM and 6OM is the dramatic decrease
in Φf on going from hexane to acetonitrile (Table 3). In
particular, the confinement of the central CdC bond in the fused
ring does not prevent 6OM from having a low value of Φf in
acetonitrile. The solvent effect on Φf for 4OM is different from
those for the other methoxy derivatives but similar to that for
the group I aminostilbenes, as is the case of 3CN among 1CN-
6CN. Although a pure sample of 5OM is not available, which
precludes a direct examination of the role of the styrenyl-anilino
C-C bond in the TICT state formation of 1OM, the experience
gained from the corresponding studies of 1CN-6CN has
allowed us to conclude that the TICT state of 1OM is associated
with the twisting of the stilbenyl-anilino C-N bond. Indeed, if
the rotation of the styrenyl-anilino C-C bond played an
important role in the excited-state behavior of 1OM, it would
affect 4OM as well, but this is not the case. Such a unique
TICT state formation observed for 1OM could be attributed to
the strong electron-donating methoxy substituent that results in
a stronger amino donor than that in the other derivatives of 1.
Several pieces of evidence have suggested that the process
of the C-N bond twisting toward the DMABN-like TICT state
is negligible for 1CN and 1CO in the nonpolar solvent
hexane: (a) The spectra show vibrational structures and no dual
fluorescence. (b) The fluorescence quantum yields are high and
the sum of Φf + 2Φtc is close to unity. (c) The values of kf are
large and comparable to those of the group I compounds, in
contrast to the low kf values for the TICT fluorescence in polar
solvents. (d) No fluorescence enhancement was observed for
1CN in hexane upon heating from -40 to 50 °C. Instead, the
fluorescence is only slightly perturbed, as previously observed
for trans-N,N-diphenylaminostilbene (DPhAS),25 indicating a
triplet-state mechanism of photoisomerization. Since the cyano
1
group is a strong EW group, the energy of the p* state for
1CN would be largely raised according to the substituent effect
1
shown in Figure 10. On the other hand, the corresponding t*
state is not destabilized (Table 1), presumably due to the
conjugation effect of the cyano group that compensates for its
induction effect. As a result, the torsional barrier in the singlet
excited state might be too high to be overcome at room
temperature. This could account not only for the large value of
Φf in hexane but also for the triplet-state mechanism of
photoisomerization (Φisc ) 2Φtc, kisc ) Φiscτf -1). The non-
radiative decay rate constant in hexane could thus be attributed
N-Alkyl- vs N-Aryl-Substituted Aminostilbenes. The ICT
states of several N,N-dimethylamino-substituted stilbenes, in-
cluding the parent molecule DS,15 the donor-acceptor-type
4-(N,N-dimethylamino)-4′-cyanostilbene (DCS)16,17 and 4-(N,N-
dimethylamino)-4′-nitrostilbene (DNS),18 and the donor-donor-
type 4,4′-tetramethyldiaminostilbene (DDS),15,19 have been
discussed by several research groups in favor of the TICT state
formation in polar solvents, although their steady-state spectra
lack dual fluorescence. In addition, despite the different nature
of these aminostilbenes, all of their TICT states have been
suggested to result from the twisting of the anilino-styrenyl C-C
single bond and to be responsible for the observed fluorescence,
even when the value of Φf is high for some cases. The emissive
nature of the TICT state was then attributed to either a large
to the rate constant for intersystem crossing (i.e., knr ≈ kisc
≈
2.1 × 108 s-1), a value nearly the same as that for DPhAS (kisc
≈ 2.0 × 108 s-1).25 (e) The sigmodial curves in the plots of Φf
against ET(30) (Figure 8) are analogous to the acid-base
titration curves, which indicates that the “equivalent point” of
1
the population of molecules in the TICT vs t* state is not
achieved until the solvent is more polar than THF.
(42) (a) Maliakal, A.; Lem, G.; Turro, N. J.; Ravichandran, R.; Suhadolnik, J.
C.; DeBellis, A. D.; Wood, M. G.; Lau, J. J. Phys. Chem. A 2002, 106,
7680-7689. (b) Nad, S.; Kumbhakar, M.; Pal, H. J. Phys. Chem. A 2003,
107, 4808-4816.
9
J. AM. CHEM. SOC. VOL. 126, NO. 39, 2004 12333

A R T I C L E S
Yang et al.
vibronic mixing with the other allowed states or an incomplete
twisting of the single bond (i.e., twisted angle < 90°).
possess a DS-like TICT state. However, regarding the distinct
fluorescence properties observed for the group I vs groups II
and III aminostilbenes, it would be difficult to understand why
the consequence of the C-C bond twisting is so different from
that of the C-N bond twisting in 1.
Experimental Section
Methods. Electronic spectra were recorded at room temperature.
UV spectra were measured on a Jasco V-530 double-beam spectro-
photometer. Fluorescence spectra were recorded on a PTI QuantaMaster
C-60 spectrofluorometer. The optical density of all solutions was about
0.1 at the wavelength of excitation. It should be noted that, unlike the
uncorrected spectra previously reported in refs 25 and 26, the
fluorescence spectra reported herein have been corrected for the
response of the detector. The fluorescence spectra at other temperature
were measured in an Oxford OptistatDN cryostat with an ITC502
temperature controller. A N2-bubbled solution of anthracene (Φf ) 0.27
in hexane)45 was used as a standard for the fluorescence quantum yield
determinations of 1-6 in acetonitrile under N2-bubbled conditions with
solvent refractive index correction. An error of (10% is estimated for
the fluorescence quantum yields. Fluorescence decays were measured
at room temperature by means of a PTI Timemaster apparatus with a
gated hydrogen arc lamp using a scatter solution to profile the
instrument response function. The goodness of nonlinear least-squares
fit was judged by the reduced ø2 value (<1.2 in all cases), the
randomness of the residuals, and the autocorrelation function. Quantum
yields of photoisomerization were measured on optically dense degassed
solutions (∼10-3 M) at 313 nm using a 75-W Xe arc lamp and
monochromator. trans-Stilbene was used as a reference standard (Φtc
) 0.50 in hexane).46 The extent of photoisomerization (<10%) was
determined using HPLC analysis (Waters 600 controller and 996
photodiode array detector, Thermo APS-2 Hypersil, heptane and ethyl
acetate mixed solvent). The reproducibility error was <10% of the
average. MOPAC-AM1 and INDO/S-CIS-SCF (ZINDO) calculations
were performed on a personal computer using the algorithms supplied
with the package of Quantum CAChe Release 3.2, a product of Fujitsu
Ltd. The X-ray diffraction measurements were performed on Bruker
Smart-CCD diffractometers (λ ) 0.71073 Å) at room temperature.
Intensity data were collected in 1315 frames with increasing ω (0.3°
per frame) and corrected for Lp and absorption effects using the
SADABS program. The structures were solved by direct methods.
Structural parameters were refined on the basis of F2. All calculations
were performed by using SHELXTL programs. All non-hydrogen atoms
were refined anisotropically. Hydrogen atoms were assigned idealized
locations and given isotropic thermal parameters 1.2× the thermal
parameter of the carbon atoms (but 1.5× for the hydrogens in methyl
groups) to which they were attached.
However, the arguments for such a three-state (1t*, 1p*, and
TICT) model for these N,N-dimethylaminostilbenes have been
challenged.21-23 First, a large solvatofluorochromism does not
necessarily correspond to a fluorescent TICT state, because the
1t* state could be also highly polar, as shown by femtosecond
dynamics for the case of DCS.22,23 In addition, the implication
1
of t* f TICT based on the precursor-successor relationship
observed in time-resolved fluorescence spectra could also be
explained by the longitudinal dielectric relaxation of solvent
molecules.21,22 Further, the observation of lifetime maxima at
intermediate temperature could be associated with the phase
(e.g., melting or glass) transition of the solvent molecules instead
of the presence of a fluorescent TICT state.21 Moreover, the
pronounced ring-bridging effect on the fluorescence quantum
yield and the nonradiative decay rate for DCSB vs DCS and
for DSB vs DS have been interpreted as results of the restriction
of the C-C single bond rotation, in favor of the TICT model.
However, the same phenomenon is also present for trans-1,1′-
biindanylidene (SB) vs trans-stilbene, a system that conforms
1
to the two-state (1t* and p*) model. Thus, the ring-bridging
effect could be simply due to the methylene substituent effect
that modifies the reaction hypersurface for the double bond
torsion and/or induces new reactions.32 Indeed, it has been
shown that the torsional barrier for SB is intrinsically lower
than that for trans-stilbene.43
A comparison of the fluorescence behavior between 1 and
DS has led to a new challenge to the TICT model for DS.44 On
the basis of (1) the shape of the fluorescence spectra, (2) the
relative values of ∆ν1/2 and ∆νst, and (3) the response of Φf to
the solvent polarity, the fluorescence behavior of DS is more
like that of the group I rather than the groups II and III species
of 1. In particular, the excited-state dipole moment for DS is
nearly the same as that for 1Me and much lower than those for
the TICT state of 1OM, 1CN, and 1CO. The observations of
(a) a much lower Φf value for 5CN vs the other cyano
derivatives in hexane and (b) a similar ratio of fluorescence
quantum yields for Φf(5CN)/Φf(DSB) (28) and Φf(1CN)/ Φf-
(DS) (25) in hexane also support a special methylene bridging
effect that enhances the nonradiative decay. Despite such a
special substituent effect, 5CN still displays a large decrease
in Φf on going from hexane to acetonitrile due to the C-N
bond twisting. Therefore, it should be that either the TICT state
proposed for DS is unimportant or the group I aminostilbenes
Materials. Solvents for organic synthesis were reagent grade or
HPLC grade, but all were HPLC grade for spectra and quantum yield
measurements. All other compounds were purchased from commercial
sources and were used as received. The characterizations of compounds
1-6 are provided in the Supporting Information.
Concluding Remarks
The substituent-dependent excited-state behavior of the seven
N-aryl-substituted 4-aminostilbene derivatives 1 has been
elucidated and divided into three categories. The group I consists
of 1H, 1Me, 1Cl, and 1CF, and their excited decay conforms
to the two-state model. Aminostilbenes 1CN and 1CO belong
to group II, the members of which display dual fluorescence in
polar solvents. The short- and the long-wavelength emission
bands have been attributed to the emission from the planar and
(43) (a) Saltiel, J.; D’Agostino, J. T. J. Am. Chem. Soc. 1972, 94, 6445-6456.
(b) Rothenberger, G.; Negus, D. K.; Hochstrasser, R. M. J. Chem. Phys.
1983, 79, 5360-5367.
(44) In view of the inherent difference in electronic structures among donor-
only, donor-acceptor-type, and donor-donor-type aminostilbenes, any
comparison to be made between 1 and the N-alkyl derivatives should be
more appropriately restricted to the same type of aminostilbenes, viz., the
donor-only DS.
(45) Dawson, W. R.; Windsor, M. W. J. Phys. Chem. 1968, 72, 3251-3260.
(46) Malkin, S.; Fischer, E. J. Phys. Chem. 1964, 68, 1153-1163.
9
12334 J. AM. CHEM. SOC. VOL. 126, NO. 39, 2004

Substituent-Dependent Photoinduced Charge Transfer
A R T I C L E S
twisted ICT states, respectively. The group III compound is
1OM, whose photochemical properties closely resemble the
group II molecules and could also be described by the three-
effect is a sensitive probe for the “degree of conjugation”
between the D and A groups of aminostilbenes in the excited
states. By the same token, the similarity in solvent-dependent
fluorescence properties for DS and the group I species suggests
that there is no need to invoke a fluorescent TICT state for DS.15
Further dynamic and theoretical studies on these and the related
systems would complement the current results and provide more
insights into the ICT states of aminostilbenes.
1
state model with an overlapped emission from the t* and the
TICT states. The TICT state formation in 1OM, 1CN, and 1CO
is only favorable in solvents more polar than THF. The
corresponding studies on model compounds 2-6 have allowed
us to deduce the TICT structures for the groups II and III
compounds. Whereas the TICT state for 1CN is DMABN-like,
resulting from the twisting of the benzonitrilo-anilino C-N
single bond, it is the stilbenyl-anilino C-N bond that twists in
the case of 1OM. Despite the difference in the bond that twists
and in the direction of the ICT transition, the TICT states for
1OM and 1CN have the common features of broad fluorescence
spectra and low fluorescence yields. These features are appar-
ently different from the TICT states previously proposed for
N,N-dialkylaminostilbenes such as DCS and DS. It is important
to note that the basis of our TICT arguments relies on the close
linkage of the steady-state fluorescence behavior for 1CN and
the TICT paradigm DMABN. A similar spectroscopic correla-
tion between 1OM and 1CN in turn leads to the conclusion of
the presence of a TICT state for 1OM. In this case, the group
I molecules have functioned as the “TICT-free controlled
compounds”, complementary to the bridged model molecules
3-6. These important correlations would not be possible without
the sensitive response of the fluorescence properties of 1 to the
solvent polarity. In other words, the N-aryl amino conjugation
Acknowledgment. We thank the National Science Council
of Taiwan, UST, and the NCU-ITRI Joint Research Center for
financial support, Mr. Jyh-Wei Tu for the synthesis of com-
pounds 3OM and 3CN, and Miss F.-L. Liao and Professor
S.-L. Wang at the Instrumentation Center of National Tsing Hua
University for resolving the crystal structures of 1CN, 2CN,
and 1CO.
Supporting Information Available: Detailed characterization
data for aminostilbenes 1-6; crystal refinement data for 1CN,
1CO, and 2CN; solvatofluorochromic plots for 1, 3CN, and
4OM; and fluorescence and excitation spectra of 1 in hexane
and acetonitrile recorded by changing the excitation and
emission wavelengths (PDF). X-ray experimental details (CIF).
This material is available free of charge via the Internet at
http://pubs.acs.org.
JA047604D
9
J. AM. CHEM. SOC. VOL. 126, NO. 39, 2004 12335