Solvent-controlled excited state behavior: 2-(2'-pyridyl)indoles in alcohols

Article abstract of DOI:10.1021/ja952797d

Efficient fluorescence quenching caused by rapid internal conversion from the first excited singlet state is observed in alcohol solutions of 2-(2'-pyridyl)indoles, molecules which may simultaneously act as hydrogen bouding donors and acceptors. Electroni

Full text of DOI:10.1021/ja952797d

3508
J. Am. Chem. Soc. 1996, 118, 3508-3518
Solvent-Controlled Excited State Behavior:
2-(2′-Pyridyl)indoles in Alcohols
Jerzy Herbich,† Chi-Ying Hung,‡ Randolph P. Thummel,‡ and Jacek Waluk*,†
Contribution from the Institute of Physical Chemistry, Polish Academy of Sciences, Kasprzaka 44,
01-224 Warsaw, Poland, and Department of Chemistry, UniVersity of Houston,
Houston, Texas 77204-5641
ReceiVed August 14, 1995X
Abstract: Efficient fluorescence quenching caused by rapid internal conversion from the first excited singlet state is
observed in alcohol solutions of 2-(2′-pyridyl)indoles, molecules which may simultaneously act as hydrogen bonding
donors and acceptors. Electronic and infrared absorption studies show that upon adding alcohols to nonpolar solutions
of pyridylindoles, 1:1 complexes are formed in the ground state, with hydrogen bonding occurring to the indole NH
group. At higher alcohol concentrations 1:n (n g 2) solvates dominate. Investigations of the fluorescence intensity
as a function of temperature and deuterium substitution in the hydroxylic group of the alcohol, and the results obtained
for the N-methylated derivatives and 2-phenylindole show that the quenching may be described by a stepwise
mechanism. First, a 1:1 cyclic, doubly hydrogen-bonded complex between alcohol and pyridylindole is formed
after photoexcitation. This process is controlled by solvent reorientation. Excited state double proton transfer in
such a complex opens up a fast internal conversion channel. The driving force for phototautomerization arises as a
result of the excited state increase of the basicity of the pyridine nitrogen atom and the enhanced acidity of the
indole NH group. This effect was also predicted by calculations of excited state valence electron potentials, parameters
that can be used as acido-basic reactivity indices.
Introduction
formation of cyclic complexes, usually with alcohol, may lead
to excited state tautomerization involving the cooperative
movement of two or even more protons. The most investigated
system of this kind are alcohol complexes of 7-azaindole;22-53
similar behavior has been observed for 1-azacarbazole,54-58
Formation of intermolecular hydrogen bonding has been
shown to influence the photophysical behavior of many het-
eroaromatic systems.1-98 For instance, different energy shifts
of close-lying nπ* and ππ* states upon hydrogen bonding may
result in the reversal of state ordering or lead to large changes
in the strength of vibronic or spin-orbit coupling (“proximity
effects”).4 Another mechanism of altering the photophysical
properties operates when the hydrogen bonding strength is not
the same in the ground and excited states. This leads to different
shapes and positions of the minima in the potential energy
hypersurfaces of the two states, which, in turn, may activate
rapid internal conversion from S1.90-96 In many systems,
fluorescence quenching has been explained on the basis of
charge transfer interactions between the proton donor and
acceptor5-20 or by a hydrogen transfer21 mechanism.
(13) Martin, M. M.; Miyasaka, H.; Karen, A.; Mataga, N. J. Phys. Chem.
1985, 89, 182.
(14) Miyasaka, H.; Tabata, A.; Kamada, K.; Mataga, N. J. Am. Chem.
Soc. 1993, 115, 7335.
(15) Miyasaka, H.; Tabata, A.; Ojima, S.; Ikeda, N.; Mataga, N. J. Phys.
Chem. 1993, 97, 8222.
(16) Talavera, E. M.; Quintero, B.; Alvarez, J. M. J. Photochem.
Photobiol. A 1992, 66, 171.
(17) Kikuchi, K.; Watarai, H.; Koizumi, M. Bull. Chem. Soc. Jpn. 1973,
46, 749.
(18) Yamamoto, S.; Kikuchi, K.; Kokubun, H.; Bull. Chem. Soc. Jpn.
1976, 49, 2950; J. Photochem. 1977, 7, 177.
(19) Okada, T.; Karaki, I.; Mataga, N. J. Am. Chem. Soc. 1982, 104,
7191.
(20) Poizat, O.; Buntinx, G.; Valat, P.; Wintgens, W.; Bridoux, M. J.
Phys. Chem. 1993, 97, 5905.
(21) Rehm, D.; Weller, A. Isr. J. Chem. 1970, 8, 259.
(22) Taylor, C. A.; El-Bayoumi, M. A.; Kasha, M. Proc. Natl. Acad.
Sci. U.S.A. 1969, 63, 253.
(23) Ingham, K. C.; Abu-Elgheit, M.; El-Bayoumi, M. A. J. Am. Chem.
Soc. 1971, 93, 5023.
(24) Ingham, K. D.; El-Bayoumi, M. A. J. Am. Chem. Soc. 1974, 96,
1674.
A particular case occurs for molecules that can simultaneously
act as hydrogen bonding donors and acceptors. In such systems,
† Polish Academy of Sciences.
‡ University of Houston.
X Abstract published in AdVance ACS Abstracts, March 15, 1996.
(1) Kasha, M. J. Chem. Soc., Faraday Trans. 2 1986, 82, 2379; J. Acta
Phys. Pol. 1987, A71, 717.
(2) Arnaut, L. G.; Formosinho, S. J. J. Photochem. Photobiol. A 1993,
75, 1; Formosinho, S. J.; Arnaut, L. G. J. Photochem. Photobiol. A 1993,
75, 21.
(25) El-Bayoumi, M. A.; Avouris, P.; Ware, W. R. J. Chem. Phys. 1975,
62, 2499.
(3) Ormson, S. M.; Brown, R. G. Prog. Reaction Kinetics 1994, 19, 45.
(4) Lim, E. C. J. Phys. Chem. 1986, 90, 6770.
(5) Mataga, N.; Tsuno, S. Naturwissenschaften 1956, 10, 305. Bull.
Chem. Soc. Jpn. 1957, 30, 711.
(26) Avouris, P.; Yang, L. L.; El-Bayoumi, M. A. Photochem. Photobiol.
1976, 24, 211.
(27) Hetherington, W. M., III; Micheels, R. M.; Eisenthal, K. B. Chem.
Phys. Lett. 1979, 66, 230.
(6) Mataga, N. Bull. Chem. Soc. Jpn. 1958, 31, 481.
(7) Mataga, N.; Torihashi, Y.; Kaifu, Y. Z. Z. Phys. Chem. (Frankfurt/
Main) 1962, 34, 379.
(8) Mataga, N.; Tanaka, F.; Kato, M. Acta Phys. Pol. 1968, 34, 733.
(9) Martin, M. M.; Ware, W. R. J. Phys. Chem. 1978, 82, 2770.
(10) Ikeda, N.; Okada, T.; Mataga, N. Chem. Phys. Lett. 1980, 69, 251;
Bull. Chem. Soc. Jpn. 1981, 54, 1025.
(28) Waluk, J.; Bulska, H.; Grabowska, A.; Pakula, B.; Sepiol, J. J.
Luminescence 1981, 24/25, 519.
(29) Collins, S. J. Phys. Chem. 1983, 87, 3202.
(30) Fuke, K.; Yoshiuchi, H.; Kaya, K.; Achiba, Y.; Sato, K.; Kimura,
K. Chem. Phys. Lett. 1984, 108, 179.
(31) Bulska, H.; Grabowska, A.; Pakula, B.; Sepiol, J.; Waluk, J.; Wild,
U. P. J. Luminescence 1984, 29, 65.
(11) Martin, M. M.; Ikeda, N.; Okada, T.; Mataga, N. J. Phys. Chem.
1982, 86, 4148.
(12) Ikeda, N.; Miyasaka, H.; Okada, T.; Mataga, N. J. Am. Chem. Soc.
1983, 105, 5206.
(32) Herbich, J.; Sepiol, J.; Waluk, J. J. Mol. Struct 1984, 114, 329.
(33) Fuke, K.; Yoshiuchi, H.; Kaya, K. J. Phys. Chem. 1984, 88, 5840.
(34) Tokumura, K.; Watanabe, Y.; Itoh, M. Chem. Phys. Lett. 1984, 111,
379; J. Phys. Chem. 1986, 90, 2362.
0002-7863/96/1518-3508$12.00/0 © 1996 American Chemical Society

2-(2′-Pyridyl)Indoles in Alcohols
J. Am. Chem. Soc., Vol. 118, No. 14, 1996 3509
hydroxyquinolines,59-67 indoloquinoxalines,68,69 2-hydroxy-4,5-
benzotropone,70 2-(2′-pyridyl)benzimidazole,71-73 3-hydroxy-
flavone,74-82 and lumichromes.83-86
Although the phototautomerization of 7-azaindole alcohol and
water complexes has been studied for more than two decades,
basic controversies still remain. The structure of the ground
state species is not definitely solved. The nature of the rate-
determining step for phototautomerization is being discussed:
it may involve proton movement, solvent reorientation, or a
combination of both.
In this work, we address problems of the ground state
structure and the excited state deactivation mechanism in order
to understand the process of fluorescence quenching by alcohols
observed in 2-(2′-pyridyl)indoles.97,98 We have postulated that
this phenomenon must involve a cyclic, doubly hydrogen-
bonded complex with alcohol, which strongly suggests that the
shift of two protons, occurring in the excited state, is responsible
for the process. Still, several problems remained to be solved.
It was not clear whether the attainment of a structure predisposed
to tautomerization occurs prior to or after excitation. The origin
of the rate-determining step for the quenching has not been
elucidated.
The results presented in this work are based on a study of
the ground state equilibria followed by calculations and the
investigation of the quenching efficiency as a function of
temperature, solvent properties, and the isotope substitution. We
show that the stepwise mechanism of quenching involves: (i)
solvent rearrangement leading to the formation of a doubly
hydrogen-bonded alcohol complex and (ii) internal conversion
induced by the concerted motion of two protons.
(35) McMorrow, D.; Aartsma, T. J. Chem. Phys. Lett. 1986, 125, 581.
(36) Konijnenberg, J.; Huizer, A. H.; Varma, C. A. G. O. J. Chem. Soc.,
Faraday Trans. 2 1988, 84, 1163.
(37) Moog, R. S.; Bovino, S. C.; Simon, J. D. J. Phys. Chem. 1988, 92,
6545.
(38) Matsui, K.; Matsuzuka, T.; Fujita, H. J. Phys. Chem. 1989, 93, 4991.
(39) Kim, S. K.; Bernstein, E. R. J. Phys. Chem. 1990, 94, 3531.
(40) Jang, D.-J.; Brucker, G. A.; Kelley, D. F. J. Phys. Chem. 1986, 90,
6809.
Experiment and Calculations
The synthesis of 2-(2′-pyridyl)indole (Pyln-0), 3,3′-methylene-2-
(2′-pyridyl)indole (Pyln-1), 3,3′-dimethylene-2-(2′-pyridyl)indole (Pyln-
2), 3,3′-trimethylene-2-(2′-pyridyl)indole (Pyln-3), and 3,3′-tetramethylene-
2-(2′-pyridyl)indole (Pyln-4) has been described previously.99 The
N-methylated derivatives were obtained as follows:
(41) Share, P. E.; Sarisky, M. J.; Pereira, M. A.; Repinec, S. T.;
Hochstrasser, R. M. J. Lumin. 1991, 48/49, 204.
(42) Moog, R. S.; Maroncelli, M. J. Phys. Chem. 1991, 95, 10359.
(43) Ne´grerie, M.; Gai, F.; Bellefeuille, S. M.; Petrich, J. W. J. Phys.
Chem. 1991, 95, 8663.
(44) Gai, F.; Chen, Y.; Petrich, J. W. J. Am. Chem. Soc. 1992, 114, 8343.
(45) Chou, P.-T.; Martinez, M. L.; Cooper, W. C.; Collins, S. T.;
McMorrow, D. P.; Kasha, M. J. Phys. Chem. 1992, 96, 5203.
(46) Chapman, C. F.; Maroncelli, M. J. Phys. Chem. 1992, 96, 8430.
(47) Ne´grerie, M.; Gai, F.; Lambry, J.-C.; Martin, J.-L.; Petrich, J. W.
J. Phys. Chem. 1993, 97, 5046.
1-Methyl-2-(2′-pyridyl)indole (MePyln-0). A mixture of 2-acetyl-
pyridine (0.88 g, 7.27 mmol) and 1-methyl-1-phenylhydrazine (1.10
g, 9.02 mmol) in EtOH (5 mL) with one drop of concentrated H2SO4
was refluxed for 2 h. The mixture was cooled, and the crude product
was collected and purified by chromatography on SiO2, eluting with
CH2Cl2, to obtain 1.22 g (74%) of the corresponding N-methylphenyl-
hydrazone. This hydrazone (0.85 g, 3.78 mmol) was heated with
polyphosphoric acid (PPA, 1.15 g) at 115 °C for 2 h. The mixture
was cooled, made basic with 50% NaOH, and extracted with CH2Cl2.
Chromatography on SiO2, eluting with 10:1 CH2Cl2/EtOAc provided
0.29 g (37%) of 1-methyl-2-(2′-pyridyl)indole, mp 86-86.5 °C: 1H
(48) Rich, R. L.; Chen, Y.; Neven, D.; Ne´grerie, M.; Gai, F.; Petrich, J.
W. J. Phys. Chem. 1993, 97, 1781.
(49) Chen, Y.; Rich, R. L.; Gai, F.; Petrich, J. W. J. Phys. Chem. 1993,
97, 1770.
(50) Chen, Y.; Gai, F.; Petrich, J. W. J. Am. Chem. Soc. 1993, 115, 10158;
Chem. Phys. Lett. 1994, 222, 329.
(51) Gai, F.; Rich, R. L.; Petrich, J. W. J. Am. Chem. Soc. 1994, 116,
735.
(52) Chang, C.-P.; Wen-Chi, H.; Meng-Shin, K.; Chou, P.-T.; Clements,
H. J. Phys. Chem. 1994, 98, 8801.
(74) Woolfe, G. J.; Thistlethwaite, P. J. J. Am. Chem. Soc. 1981, 103,
6916.
(75) Itoh, M.; Tokumura, K.; Tanimoto, Y.; Okada, Y.; Takeuchi, H.;
Obi, K.; Tanaka, I. J. Am. Chem. Soc. 1982, 104, 4146.
(76) Itoh, M.; Fujiwara, Y. J. Phys. Chem. 1983, 87, 4558.
(77) Strandjord, A. J. G.; Courtney, S. H.; Friedrich, D. M.; Barbara, P.
F. J. Phys. Chem. 1983, 87, 1125.
(53) Ray, J. G.; Sengupta, P. K. Chem. Phys. Lett. 1994, 230, 75.
(54) Chang, C.; Shabestary, N.; El-Bayoumi, M. A. Chem. Phys. Lett.
1980, 75, 107.
(55) Waluk, J.; Grabowska, A.; Pakula, B.; J.; Sepiol, J. J. Phys. Chem.
1984, 88, 1160.
(78) McMorrow, D.; Kasha, M. J. Phys. Chem. 1984, 88, 2235.
(79) Strandjord, A. J. G.; Barbara, P. F. J. Phys. Chem. 1985, 89, 2355.
(80) Brucker, G. A.; Kelley, D. F. J. Phys. Chem. 1987, 91, 2856.
(81) Lo´pez-Arbeloa, T.; Lo´pez-Arbeloa, F.; Tapia, M. J.; Lo´pez-Arbeloa,
I.; J. Phys. Chem. 1993, 97, 4704.
(56) Waluk, J.; Komorowski, S. J.; Herbich, J. J. Phys. Chem. 1986, 90,
3868.
(57) Waluk, J.; Herbich, J.; Oelkrug, D.; Uhl, S. J. Phys. Chem. 1986,
90, 3866.
(82) Schwartz, B. J.; Peteanu, L. A.; Harris, C. B. J. Phys. Chem. 1992,
96, 3591.
(83) Song, P. S.; Sun, M.; Koziolowa, A.; Koziol, J. J. Am. Chem. Soc.
1974, 96, 4319.
(58) Fuke, K.; Tsukamoto, K.; Misaizu, F.; Kaya, K. J. Chem. Phys.
1991, 95, 4074.
(59) Mason, S. F.; Philip, J.; Smith, B. E. J. Chem. Soc. (A) 1968, 3051.
(60) Thistlethwaite, P. J. Chem. Phys. Lett. 1983, 96, 509.
(61) Itoh, M.; Adachi, T.; Tokumura, K. J. Am. Chem. Soc. 1983, 105,
4828.
(84) Fugate, R. D.; Song, P. S. Photochem. Photobiol. 1976, 24, 479.
(85) Koziolowa, A. Photochem. Photobiol. 1979, 29, 459.
(86) Choi, J. D.; Fugate, R. D.; Song, P. S. J. Am. Chem. Soc. 1980,
102, 5293.
(87) Shizuka, H.; Kameta, K.; Shinozaki, T. J. Am. Chem. Soc. 1985,
107, 3956.
(62) Itoh, M.; Adachi, T.; Tokumura, K. J. Am. Chem. Soc. 1984, 106,
850.
(88) Shizuka, H.; Serizawa, M. J. Phys. Chem. 1986, 90, 4573.
(89) Shizuka, H.; Serizawa, M.; Okazaki, K.; Shioya, S. J. Phys. Chem.
1986, 90, 4694.
(90) Grabowska, A.; Pakula, B.; Sepiol, J. NouV. J. Chim. 1979, 3, 287.
(91) Barbara, P. F.; Rentzepis, P. M.; Brus, L. E. J. Am. Chem. Soc.
1980, 102, 2786.
(92) Houston, A.; Scott, G.; Gupta, A. J. Chem. Phys. 1982, 76, 4978.
(93) Woessner, G.; Goeller, G.; Kollat, P. A.; Stezowski, J. J.; Hauser,
M.; Klein, U. K. A.; Kramer, H. E. A. J. Phys. Chem. 1984, 88, 5544.
(94) McMorrow, D.; Kasha, M. J. Phys. Chem. 1984, 105, 2235; J. Am.
Chem. Soc. 1983, 105, 5733.
(95) Flom, S. R.; Barbara, P. F. Chem. Phys. Lett. 1983, 94, 488; J.
Phys. Chem. 1985, 89, 4489.
(63) Konijnenberg, J.; Ekelmans, G. B.; Huizer, A. H.; Varma, C. A. G.
O. J. Chem. Soc., Faraday Trans. 2 1989, 85, 39.
(64) Jalink, C. J.; Van Ingen, W. M.; Huizer, A. H.; Varma, C. A. G. O.
J. Chem. Soc., Faraday Trans. 1991, 87, 1103.
(65) Kang, W. K.; Cho, J. S.; Lee, M.; Kim, D. H.; Ryoo, R.; Jung, K.
H.; Jang, D.-J. Bull. Korean Chem. Soc. 1992, 13, 140.
(66) Lavin, A.; Collins, S. Chem. Phys. Lett. 1993, 204, 96; Chem. Phys.
Lett. 1993, 207, 513; J. Phys. Chem. 1993, 97, 13615.
(67) Chou, P. T.; Martinez, S. S. Chem. Phys. Lett. 1995, 235, 463.
(68) Waluk, J.; Komorowski, S. J. J. Mol. Struct. 1986, 142, 159.
(69) Waluk, J.; Komorowski, S. J. Chem. Phys. Lett. 1987, 133, 368.
(70) Jang, D.-J.; Brucker, G. A.; Kelley, D. F. J. Phys. Chem. 1986, 90,
6808.
(96) Inoue, H.; Hida, H.; Naksashima, N.; Yoshihara, K. J. Phys. Chem.
1982, 86, 3184.
(71) Kondo, M. Bull. Chem. Soc. Jpn. 1978, 51, 3027.
(72) Brown, R. G.; Entwistle, N.; Hepworth, J. D.; Hodgson, K. W.;
May, B. J. Phys. Chem. 1982, 86, 2418.
(73) Rodriguez Prieto, M. F.; Mosquera, M.; Novo, M. J. Phys. Chem.
1990, 94, 8536.
(97) Herbich, J.; Rettig, W.; Thummel, R. P.; Waluk, J. Chem. Phys.
Lett. 1992, 195, 556.
(98) Herbich, J.; Waluk, J.; Thummel, R. P.; Hung, C. J. J. Photochem.
Photobiol. A 1994, 80, 157.
(99) Thummel, R. P.; Hegde, V. J. Org. Chem. 1989, 54, 1720.

3510 J. Am. Chem. Soc., Vol. 118, No. 14, 1996
Herbich et al.
a
Table 1. The maxima of Room Temperature Fluorescence (in Units of 103 cm-1
)
n-hexane
benzene
ethyl ether
BuCN
pyridine
DMSO
CH3CN
BuOH
PrOH
EtOH
MeOH
Pyln-0
MePyln-0
Pyln-1
Pyln-2
MePyln-2
Pyln-3
Pyln-4
2-PI
27.5
26.8
27.8
26.3
25.5
26.5
26.2
28.5
26.6
25.9
27.0
25.5
24.9
25.7
25.7
27.7
26.8
26.5
26.9
25.6
25.3
25.9
25.7
27.9
26.4
25.8
26.7
25.6
25.0
25.7
25.2
27.7
26.0
25.5
26.2
24.9
24.6
25.3
24.9
26.9
25.6
25.3
26.0
24.9
24.5
24.9
24.4
27.1
26.1
25.5
26.4
25.3
24.7
25.4
24.6
27.6
24.6
25.3
25.9
25.1
24.5
25.3
24.5
27.6
24.5
25.0
25.8
25.0
24.4
25.3
24.2
27.6
24.3
24.6
25.6
25.0
24.4
25.1
23.9
27.6
24.4
24.7
25.4
24.8
24.2
24.9
23.6
27.6
a The error is estimated to be (200 cm-1
.
NMR (300 MHz, CDCl3) δ 8.70 (d, 1H, J ) 4.4 Hz), 7.8-7.7
(overlapping m, 2H), 7.66 (d, 1H, J ) 7.8 Hz), 7.41 (d, 1H, J ) 8.2
Hz), 7.3-7.2 (overlapping M, 2H), 7.14 (t, 1H, J ) 7.7 Hz), 6.87 (s,
1H), 4.07 (s, 3H); IR (KBr) 3010, 2960, 1550, 1525, 1430, 1395, 1295,
1130 cm-1
.
1-Methyl-3,3′-dimethylene-2-(2′-pyridyl)indole (MePyln-2). Fol-
lowing the procedure described above, the reaction of 5,6,7,8-tetrahydro-
8-quinolone100 (0.62 g, 4.22 mmol) with 1-methyl-1-phenylhydrazine
(0.72 g, 5.9 mmol) afforded a crude hydrazone which was purified by
chromatography on Al2O3 eluting with CH2Cl2. The hydrazone was
then treated with PPA (10.5 g) in toluene (4 mL) at 115 °C for 2 h.
Workup provided a crude product which was purified by chromatog-
raphy on SiO2 eluting with CH2Cl2 to afford the pure indole (0.41 g,
41%), mp 99-100 °C: 1H NMR (300 MHz, CDCl3) 8.43 (d, 1H, J )
4.9 Hz), 7.59 (d, 1H, J ) 7.7 Hz), 7.51 (d, 1H, J ) 7.1 Hz), 7.38 (d,
1H, J ) 8.4 Hz), 7.27 (t, 1H, J ) 7.2 Hz), 7.13 (t, 1H, J ) 7.4 Hz),
7.03 (dd, 1H, J ) 4.9, 7.5 Hz), 4.32 (s, 3H), 3.01 (m, 4H); IR (KBr)
2600, 1528, 1426, 1204, 782, 735 cm-1. Anal. Calcd for C16H14N2-
0.2H2O: C, 80.80; H, 6.06; N, 11.78. Found: C, 80.85; H, 5.64; N,
11.89.
Commercially available 2-phenylindole (Aldrich, tech.), was purified
by thin layer chromatography and recrystallization from n-hexane and/
or methanol.
Fluorescence quantum yields of pyridylindoles were measured on a
Jasny spectrofluorimeter and corrected for the sensitivity of the
instrument. Quinine sulfate (φ ) 0.51) was used as a standard. Special
care was taken to ensure that the solvent did not contain impurities
that might quench the luminescence. Solvents: n-hexane, ethyl ether,
acetonitrile, methanol, ethanol, n-propanol, butanol, pyridine, benzene,
1,1,1,3,3,3-hexafluoropropanol-2, and dimethylsulfoxide (DMSO) were
of fluorescence or spectral grade (Merck). Butyronitrile (Merck, for
synthesis) was purified as previously described.97 The solutions used
for isotope substitution studies were prepared in a drybox under a
nitrogen atmosphere.
Absorption spectra were run on a Shimadzu UV 3100 spectropho-
tometer. Transient absorption spectra were recorded on a custom-built
nanosecond spectrophotometer with dye laser probing and computer
control.101 Infrared spectra were obtained on a NICOLET SX-170 FTIR
spectrometer.
Excited state energies were computed using the INDO/S method,
with 200 singly excited configurations taken into account in the CI
procedure. Ground state geometry optimizations were performed by
molecular mechanics (MMX force field, PCMODEL). In some cases,
the AM1 method was also used.
Figure 1. Fluorescence spectra of Pyln-2 and MePyln-2 at 293 K in
various solvents: n-hexane (1), ethyl ether (2), benzene (3), butyronitrile
(4), acetonitrile (5), pyridine (6), DMSO (7), butanol (8), n-propanol
(9), ethanol (10), and methanol (11).
local solvent polarity parameter, π*,102,103 fall on a straight line.
Pyln-2 reveals the same behavior only for non-hydrogen
bonding solvents, whereas in pyridine, DMSO, and even in ethyl
ether the fluorescence is shifted to the red more than could have
been expected on the basis of the solvatochromic curve. These
results point to an increase in the hydrogen bonding strength
between the indole NH group and the proton-accepting solvent
upon excitation to S1. We observe this effect for all the
investigated compounds that contain the NH group.
In alcohols, the situation is more complicated. Fluorescence
shifts may now reflect excited state changes in two types of
hydrogen bonds, occurring between the alcohol and either the
indole NH group or the pyridine nitrogen atom. In the former,
Results
Figure 1 presents room temperature fluorescence spectra of
Pyln-2 and MePyln-2 in various solvents. Fluorescence
maxima of all the compounds studied are given in Table 1.
Although not dramatic, solvent dependence of the fluorescence
band position clearly indicates specific intermolecular hydrogen-
bonding interactions. This is illustrated in Figure 2 for the pair
Pyln-2/MePyln-2. For the N-methylated derivative, the values
of the fluorescence maxima plotted against the values of the
(100) Thummel, R. P.; Lefoulon, F.; Cantu, D.; Mahadevan, R. J. Org.
Chem. 1984, 49, 2208.
(101) Jasny, J.; Sepiol, J.; Karpiuk, J.; Gilewski, J. ReV. Sci. Instr. 1994,
65, 1631.
(102) Kamlet, M. J.; Abboud, J. L. M.; Taft, R. M. Progr. Phys. Org.
Chem. 1981, 13, 485.
(103) Kamlet, M. J.; Abboud, J. L. M.; Abraham, M. H.; Taft, R. M. J.
Org. Chem. 1983, 48, 2877.

2-(2′-Pyridyl)Indoles in Alcohols
J. Am. Chem. Soc., Vol. 118, No. 14, 1996 3511
Chart 1. Formulae
Figure 2. Fluorescence maxima of Pyln-2 (triangles) and MePyln-2
(squares) at 293 K plotted against the values of the local polarizability
parameter, π*.102,103 (1) n-hexane, (2) diethyl ether, (3) benzene, (4)
butyronitrile, (5) acetonitrile, (6) pyridine, and (7) DMSO.
Table 2. Room Temperature Fluorescence Quantum yieldsa
methanol
ethanol
n-propanol
butanol
the alcohol molecule acts as proton acceptor, in the latter, as a
donor. The proton donating and accepting abilities of the
alcohols vary in opposite directions in the sequence: butanol,
n-propanol, ethanol, and methanol. Thus, it is not possible to
predict a priori the spectral shifts along the series of alcohols,
since they will depend on which of the two hydrogen bond types
is more affected by excitation.
Pyln-0
MePyln-0
Pyln-1
Pyln-2
MePyln-2
Pyln-3
Pyln-4
2PI
0.015
0.008
0.004
0.003
0.54
0.004
0.004
0.70
0.08
0.05
0.11
0.09
0.12
0.14
0.009
0.005
0.67
0.007
0.007
0.80
0.009
0.006
0.69
0.011
0.011
0.75
0.012
0.008
0.71
0.017
0.012
0.79
The experimentally observed energies corresponding to
fluorescence maxima in alcohols are generally lower than those
in pyridine and DMSO. The values of the local polarity
parameter, π*, are smaller for the alcohols (0.87, 1.00, 0.46,
0.51, 0.54, and 0.60 for pyridine, DMSO, butanol, n-propanol,
ethanol, and methanol, respectively); the values of the solvent
basicity parameter, â,102,103 are similar (0.64, 0.76, 0.88, 0.78,
0.77, and 0.62, in the same order). Therefore, the additional
red shift can be explained by the increased hydrogen bonding
strength between the pyridine nitrogen atom and the alcohol.
This is in agreement with the observation that the emission shift
in alcohols can be correlated with the parameter R, which
describes the hydrogen bonding acidity of the solvent103 (0.79,
0.78, 0.83, and 0.93 for butanol, n-propanol, ethanol, and
methanol).
We thus conclude that upon electronic excitation both the
indole NH group and the pyridine nitrogen atom increase their
hydrogen bonding ability: the former becomes more acidic, the
latter, more basic. A possibility of excited state proton transfer
between the two centers can be envisaged, provided that a
chemical “link” between the two parts of the chromophore could
be established. We have postulated the existence of a cyclic
1:1 complex with alcohol in order to explain the photophysical
properties of pyridylindoles.97,98 In particular, a strong fluo-
rescence quenching in alcohols is observed only for the
compounds containing both groups (Table 2). No quenching
occurs in the MePyln-2 as well as in 2-PI.
a Accuracy (20%.
Ground State Structure of Alcohol Complexes. In order
to accurately determine the stoichiometry of alcohol solvates,
it is crucial to work in the range where the concentration of
alcohol oligomers is negligible. To ensure that, we have
measured the IR spectra of methanol and butanol in CCl4. The
formation of oligomeric species is readily detected by the
appearance of broad bands around 3300-3500 cm-1. We found
that in CCl4 at 293 K, the concentration range where only
monomeric species of the alcohol are observed is below 2 ×
10-2 M for butanol and 5 × 10-2 M for methanol.
Adding alcohol to a solution of pyridylindole in a nonpolar
solvent results in spectral changes that testify to the formation
of ground state complexes. A new absorption band appears on
the low energy side of the electronic spectrum, and isosbestic
points are observed (Figure 3). This effect is quite strong for
Pyln-0, Pyln-1, and Pyln-2, weaker for Pyln-3 and Pyln-4 and
even weaker for 2-PI. These variations correlate with the
H-bonding ability of the N-H group determined from the NMR
chemical shift.99 If, instead of the alcohol, DMSO (which can
only act as a hydrogen bond acceptor) is added to the solution,
very similar spectral behavior is observed. No changes are
observed upon adding alcohols to solutions of the N-methylated
derivatives.
These results strongly suggest that the bonding with alcohol
in the ground state occurs preferentially to the NH proton. The
direct proof is provided by the IR measurements (Figure 4). In
CCl4, a band at 3473 cm-1 for Pyln-2 may be assigned to the
NH stretch of the unbound molecule. Adding alcohol leads to
a decrease in the intensity of this peak. Simultaneously, a broad
Fluorescence quenching was also observed in pyridine. The
mechanism of this process will be dealt with separately.104 Here,
we focus on the ground and excited state structure and dynamics
of alcohol complexes.
(104) Herbich, J.; Dobkowski, J.; Karpiuk, J.; Thummel, R. P.; Waluk,
J. Manuscript in preparation.
band appears at 3288 ( 5 cm-1
.

3512 J. Am. Chem. Soc., Vol. 118, No. 14, 1996
Herbich et al.
is the optical density measured at an alcohol concentration [A].
Plotting ln[(OD - OD0)/[(OD∞ - OD0)] vs ln[A] should yield
the slope equal to n, the number of alcohol molecules in a
complex. An example is shown in Figure 5 for Pyln-2 in
butanol. The slope obtained for various wavelengths was in
the range 1.04 ( 0.06. Thus, the 1:1 stoichiometry was
determined, in accordance with our earlier investigations.97
The above result is based on the assumption that the
absorption coefficients of both forms are known. While this is
true for the uncomplexed species, the value for the complex
must be approximated from the measurements performed in
large excess of the alcohol. We have therefore checked the
validity of the results by using different procedures. The
following equations may be written,105 relating the measured
quantities to the equilibrium constants of the complex formation:
Figure 3. Absorption changes observed upon adding butanol to a 10-5
M solution of Pyln-2 in n-hexane at 293 K. The concentrations of
butanol were 0, 8.8 × 10-4, 1.76 × 10-3, 3.51 × 10-3, 7.02 × 10-3
,
([P]0[A]0 + [PA]2) [P]0 + [A]0
1.05 × 10-2, 1.40 × 10-2, and 1.76 × 10-2 M. The arrow indicates
1
)
+
(3)
(4)
increasing alcohol concentration.
∆OD
∆ꢀ
K11∆ꢀ
[P]0K11[A](∆ꢀ1 + ∆ꢀ2K12[A])
1 + K11[A] + K11K12[A]2
∆OD )
[P]0K11[A](1 + 2K12[A])
1 + K11[A] + K11K12[A]2
[A]0 ) [A] +
(5)
Equation 3 assumes that only a 1:1 complex is formed, while
eqs 4 and 5 describe the formation of both 1:1 and 1:2 species,
with equilibrium constants K11 and K12 (note that the units of
both K11 and K12 are the same (M-1), since the latter is now
expressed as [PA2]/[PA][A]). ∆OD ) OD - OD0, [P]0 and
[A]0 are the initial concentrations of pyridylindole and alcohol,
respectively, [PA] is the concentration of a 1:1 complex, ∆ꢀ1
) ꢀ11 - ꢀP; ∆ꢀ2 ) ꢀ12 - ꢀP, where ꢀP is the extinction coefficient
of the uncomplexed molecule.
Using the nonlinear least squares method, we have fitted the
experimental data to eq 3 or simultaneously to eqs 4 and 5,
optimizing the values of K11, K12, ∆ꢀ11, and ∆ꢀ12.
Figure 4. Portion of the IR spectrum of Pyln-2 (c ) 1.7 × 10-3 M)
in CCl4; changes in the spectrum after adding methanol: 1.3 × 10-2
M (b), 3.5 × 10-2 M (c) and 5 × 10-2 M (d).
Yet another approach14,15 is to assume the formation of a 1:1
complex and to use the relation
The ratio of the intensity of the OH stretching band of the
alcohol (3644 cm-1) to the intensities of the other bands, not
involving the OH vibrations (e.g., 2938 cm-1), does not change
in mixed solutions of pyridylindoles and alcohols in CCl4 and
remains the same as observed in the absence of pyridylindoles.
This would mean that the hydroxylic proton of the alcohol, and
thus the pyridine nitrogen atom of pyridylindole, is not engaged
in the ground state hydrogen bonding. However, this result must
be treated with caution, since the alcohol concentration was
always larger than that of the pyridylindoles, so that the fraction
of complexed alcohol molecules did not exceed 10%. Still, this
observation agrees well with the finding that adding alcohol
did not lead to any spectral changes in the IR spectrum of
pyridylindoles in the region 950-1100 cm-1, i.e., the expected
location of a band corresponding to a vibration involving the
pyridine nitrogen atom (and thus sensitive to hydrogen bonding
to the pyridine nitrogen).
(1 - OD0/OD)/[A] ) -K11 + (ꢀ11/ꢀP)K11(OD0/OD) (6)
The intercept of the plot of (1 - OD0/OD)/[A] vs (OD0/OD)
should yield K11.
The analysis of the experimental data using these three
different procedures is shown in Figure 5 for Pyln-2 in butanol
at 293 K. All the methods yield very similar values of the
equilibrium constant K11 ) 40 ( 5 M-1
.
Figure 6 shows the temperature dependence of the complex-
ation, measured by the changes in the absorption. For Pyln-2,
we have determined the values of K11 in mixtures of n-hexane
with methanol and butanol in the temperature range 278-313
K at intervals of 5 K. From the van’t Hoff plot (Figure 6) we
obtain ∆H ) -8.6 ( 0.2 kcal/mol and -8.7 ( 0.2 kcal/mol,
∆S ) -22.3 ( 0.6 eu and -23.0 ( 0.6 eu for butanol and
methanol, respectively. From somewhat less detailed studies
of the other pyridylindoles, we estimate that the thermodynamic
parameters for Pyln-0 and Pyln-1 are similar to those of Pyln-
2, whereas alcohol complexes of Pyln-3 and Pyln-4 have
smaller values of the heats of formation (∆H ≈ -7 kcal/mol
for the former, -5 kcal/mol for the latter). It is observed that
the equilibrium constants are always larger in butanol than in
methanol. This confirms that the alcohol binds to the NH group,
since butanol is a stronger base but a weaker acid than methanol.
There are several ways to determine the stoichiometry of the
complex. For a reaction
P + nA ) PAn K1n ) [PAn]/([P][A]n)
(1)
the equilibrium constant may be expressed as
K1n ) (OD - OD0)/{(OD∞ - OD0)‚[A]n}
(2)
where OD0 and OD∞ denote the optical density measured when
only the uncomplexed or complexed forms are present, and OD
(105) Pipkin, J. D.; Stella, V. J. J. Am. Chem. Soc. 1982, 104, 6672.

2-(2′-Pyridyl)Indoles in Alcohols
J. Am. Chem. Soc., Vol. 118, No. 14, 1996 3513
noticed that the slope of ln[(OD - OD0)/[(OD∞ - OD0)] vs
ln[A] systematically deviated from the values close to 1 toward
larger values upon discarding the experimental points corre-
sponding to low alcohol concentrations. This can be interpreted
as a sign of formation of 1:2 (or even higher order) complexes.
Since we could not extend the alcohol concentration range to
larger values, due to the aggregation of alcohol molecules, we
used the approximate procedure for the evaluation of K12: ꢀ11
and ꢀ12 were assumed to be equal; the concentration of the
uncomplexed species, [P], was calculated as [P]0(OD∞ - OD)/
[(OD∞ - OD0). Then, [PA] ≈ K11[P][A] (true for [A]0 . [P]0:
in our case, the latter was always smaller than 2 × 10-5 M).
Finally, [PA2] ) [P0] - [PA], and the value of K12 could be
estimated. For Pyln-2 in hexane at 293 (K11 ) 40 M-1) with
admixtures of butanol in the range 0.01-0.02 M, the obtained
values of K12 varied between 10 and 50 M-1. More important
than the exact value is the finding that the 1:2 complexes should
begin to dominate the 1:1 species already at relatively small
alcohol concentrations. Let us denote by R(P), R(PA), and
R(PA2) the molar fractions of the uncomplexed species, 1:1 and
1:2 solvates, respectively. For Pyln-2 in hexane at 293, even
if the smallest value of K12 ) 10 M-1 is assumed, the fractions
of 1:1 and 1:2 species will be equal at [A] ) 0.1 M: R(P) )
0.11, R(PA) ) R(PA2) ) 0.44. For K12 ) 30 M-1, we get
R(P) ≈ 0.22, R(PA) ≈ 0.35, and R(PA2) ≈ 0.42 when [A] )
0.03 M. In the bulk alcohol, more than 99% of the complexes
should not be present as a 1:1 form. Naturally, with increasing
alcohol concentration the formation of higher order (1:n, n g
2) solvates with alcohol oligomers is also possible.
Figure 5. Determination of the stoichiometry of the alcohol solvates
of Pyln-2 using eq 2 (top left), eqs 3-5 (top right) and eq 6 (bottom).
The absorbance values at 27 000 cm-1 were used for calculations.
Similar results were obtained using the values at 26 800 and 27 500
cm-1
.
The results we obtain from the ground state studies may be
summarized as follows: upon adding alcohols to nonpolar
solutions of pyridylindoles, 1:1 complexes are formed, with the
bonding occurring to the NH group. No spectral evidence for
the participation of the pyridine nitrogen atom in the hydrogen
bonding was found. Thus, cyclic ground state 1:1 complexes
seem highly improbable. As the alcohol concentration increases,
1:n species (n g 2) begin to dominate. The second alcohol
molecule may become attached either to the pyridine nitrogen
atom or to the alcohol molecule that is already involved in the
complex. Several observations support the latter possibility:
(i) the electronic spectra in alcohols and DMSO (which can
only act as a hydrogen bond acceptor) are very similar; (ii) no
significant spectral shifts are detected upon passing from mixed
solutions of nonpolar solvents with alcohols (where mostly 1:1
species are present) to bulk alkohols; (iii) the absorption spectra
of the N-methylated derivatives in aprotic solvents are com-
pletely insensitive to the addition of alcohols; (iv) the IR spectra
clearly indicate formation of hydrogen bonding between alcohol
and the NH group but not between the alcohol and the pyridine
nitrogen. The somewhat surprising absence of the ground state
hydrogen bonding to the pyridine nitrogen atom may be related
to the pKa values (Table 5), which are significantly lower than
5.2, the pKa value of pyridine.106
Figure 6. Top, temperature dependence of absorption of Pyln-2 in
n-hexane with 10-2 M of butanol. The temperature was varied from
278 to 313 K in 5 K steps. Bottom, the van’t Hoff plots for methanol
(squares) and butanol (circles), obtained from absorption changes at
The most important implication of the ground state studies
is that the cyclic, doubly hydrogen-bonded 1:1 complex, crucial
for the understanding of the fluorescence quenching process,
must be formed after photoexcitation.
Temperature Dependence of Fluorescence Intensity. In
all pyridylindoles that reveal quenching by alcohols, temperature
decrease leads to the recovery of the emissive properties. At
low temperatures, both quantum yields (higher than 60%) and
lifetimes (2.3-2.8 ns) are very similar for all the compounds,
including the methylated derivatives, and practically solvent-
27 000 cm-1
.
Comparison of the results of the fitting procedures that assume
only the formation of 1:1 complexes with those based on the
method which optimized both K11 and K12 shows that the
spectral behavior in the low alcohol concentration range is
satisfactorily explained by the presence of only 1:1 solvates.
Such a result does not warrant, however, that the situation is
the same at higher alcohol concentrations. Indeed, we have
(106) Kasende, O.; Zeegers-Huyskens, Th. J. Phys. Chem. 1984, 88,
2132.

3514 J. Am. Chem. Soc., Vol. 118, No. 14, 1996
Herbich et al.
Table 3. Arrhenius Parameters for Pyridylindoles in Alcohols
In Pyln-0 and MePyln-0, emission intensity changes with
temperature obey a completely different pattern than that
observed in the bridged, nonmethylated derivatives. Lower
activation energies and preexponential factors are obtained. This
and other observations discussed below provide arguments that
the Arrhenius parameters extracted from the temperature
dependence of fluorescence of Pyln-0 do not describe the
species that undergoes efficient quenching, but another con-
former.
a
∆E (kcal/mol)
k′ (s-1
)
Pyln-2
Methanol
ethanol
ethanol-O-d
propanol
butanolb
ethanol
ethanol-O-d
propanol
methanol
ethanol
ethanol-O-d
propanol
methanol
ethanol
3.5 ( 0.5
3.9 ( 0.5
3.7 ( 0.5
4.6 ( 0.5
5.0 ( 0.5
4.3 ( 0.2
4.4 ( 0.2
5.0 ( 0.3
1.8 ( 0.4
1.7 ( 0.4
2.1 ( 0.5
1.7 ( 0.4
2.0 ( 0.5
2.2 ( 0.4
3.2 ( 0.4
3.5 ( 0.5
(2.6 ( 1.3) × 1013
(3.0 ( 1.5) × 1013
(1.0 ( 0.5) × 1013
(8.6 ( 4.3) × 1013
(3.6 ( 1.8) × 1013
(4.4 ( 1.0) × 1013
(3.0 ( 0.9) × 1013
(9.3 ( 4.0) × 1013
(3.0 ( 0.7) × 1011
(3.4 ( 1.7) × 1010
(3.0 ( 1.5) × 1010
(3.0 ( 1.5) × 1010
(7.4 ( 3.7) × 1011
(1.4 ( 0.7) × 1011
(4.4 ( 2.2) × 1011
(3.8 ( 1.9) × 1011
Pyln-3
Pyln-0
Photophysics of Pyln-0 and MePyln-0. Pyln-0 and its
methylated derivative may exist as syn and anti conformers,
whereas the syn conformation of all the other pyridylindoles is
imposed by bridging. Spectral evidence points to the existence
of two conformations of Pyln-0. The fluorescence decay is
biexponential in alcohols and in mixtures of alcohol with either
a nonpolar or a polar, aprotic solvent.97 The value of the shorter
component is practically the same as the decay time observed
for the bridged pyridylindoles. Therefore, it is assigned to the
syn conformation. The other component would thus correspond
to the anti conformation. Characteristically enough, its fluo-
rescence lifetime, 3.3 ns in both n-propanol and butanol, is
longer than that observed for Pyln-0 in aprotic solvents (1.3 ns
in n-hexane, 1.8 ns in butyronitrile97).
Me-Pyln-0
propanol
butanolb
a Obtained assuming the radiative rate of S1 depopulation, kr ) 2 ×
108 s-1 b From fluorescence lifetime measurements.
.
independent. The values of the radiative rate constant, kr,
calculated from quantum yields and lifetimes, are also similar
(all lie in the range 2-3 × 108 s-1) and only weakly dependent
on the solvent. For instance, the kr values calculated for Pyln-2
are (2.9 ( 0.8) × 108 s-1 (n-hexane, 293 K), (2.2 ( 0.7) × 108
s-1 (butyronitrile, 293 K), and (2.2 ( 0.7) × 108 s-1 (butanol,
77 K).
The quantum yields measured for Pyln-0 at 293 K in alcohols
are much larger than those of the bridged compounds (cf. Table
2). On the other hand, these yields are smaller than the yield
measured for Pyln-0 in aprotic solvents (0.48 in n-hexane, 0.47
in ethyl ether, 0.53 in butyronitrile, and 0.55 in benzene98). It
should also be noted that the emission maxima of Pyln-0 in
alcohols are shifted considerably to the red from those measured
in hydrogen-bonding solvents that can only act as proton
acceptors (e.g., pyridine and DMSO). This solvent dependence
is not observed for the other derivatives. These findings indicate
that the emission of Pyln-0 observed at room temperature in
alcohols is mostly due to the anti-conformer. Indeed, fluores-
cence decay studies performed in mixtures of n-hexane with
n-propanol or butanol show that the fraction of the anti-
conformer increases upon adding the alcohol. In pure alcohols,
more than 90% of the population may be assigned to the anti
form. It is noteworthy that the radiative constant of the
fluorescence of the anti-conformer is about one order of
magnitude smaller than that of the syn-conformer of all
pyridylindoles (4 × 107 s-1 as compared to 2-3 × 108 s-1).
This difference suggests a conformational change, such as
twisting. Transient singlet-singlet absorption spectra for Pyln-
0, with nanosecond resolution, reveal a band at around 580 nm,
which decays with the same rate (about 3.5 ns) as the longer-
lived fluorescence component. Such a band is not observed in
other pyridylindoles (cf. Figure 8). A more detailed analysis
of the photophysics of the anti conformation of Pyln-0 is being
carried out and will be presented elsewhere.104
Thus, the quenching process in alcohols is strongly temper-
ature-dependent. If one assumes that the nonradiative rate of
excited singlet state deactivation may be expressed as k0
+
nr
k′exp(-∆E/RT), a sum of a temperature independent and a
temperature dependent process, the Arrhenius-type expression
results
1/φ(T) ) 1 + (k0nr/kr) + (k′/kr)exp(-∆E/RT) )
a + b exp(-∆E/RT) (7)
Fitting the experimental data to this equation enabled the
determination of the Arrhenius parameters, k′ and ∆E. The
results are presented in Table 3. An example of the fit is shown
in Figure 7.
The activation energy of the temperature-dependent quench-
ing process in Pyln-2 is correlated with the activation energy
of viscous flow of the solvent (2.5, 3.5, 4.4, and 4.6 kcal/mol
for methanol, ethanol, n-propanol, and butanol, respectively).107
A very significant result is that the activation energy in Pyln-2
and Pyln-3 does not change upon replacing the hydroxylic
proton of the alcohol by deuterium. The isotope effect is
observed, however, in the decreased value of the preexponential
factor in deuterated ethanol. This finding will be the basis of
the model we use for the explanation of the quenching
mechanism.
The values of k0nr/kr obtained from the fit are of the order of
one. This is in agreement with the fact that the fluorescence
quantum yields observed when no quenching occurs are of the
order of 50%. Thus, the radiative and radiationless deactivation
processes occur with similar rates. We attribute the nonradia-
tive, temperature-insensitive channel to intersystem crossing,
based on the results of transient absorption measurements which
will be discussed below.
Table 2 shows that the quantum yields of Pyln-0 and
MePyln-0 in alcohols are quite similar. However, this similarity
is fortuitous. In contrast to the emission of Pyln-0, the
fluorescence decay of MePyln-0 measured in n-propanol and
butanol is purely monoexponential. The radiative rate constant,
determined from the values of fluorescence quantum yields and
lifetimes, is the same as that obtained for aprotic solvents and
similar to the values for other pyridylindoles (2 × 108 s-1).
Thus, there is no evidence for the existence of two conforma-
tions of MePyln-0 that would decay according to different
mechanisms.
Both the unbridged and methylated derivatives reveal com-
pletely different temperature dependence. In 2PI and MePyln-
2, fluorescence yields are very high already at room temperature.
(107) Calculated from the data given in Viswanath, D. S.; Natarajan, G.
Data Book on the Viscosity of Liquids; Hemisphere Publishing: New York,
1989; Landolt-Bo¨rnstein, Bd. II, 5 Teil, Springer Verlag: 1969; pp 209-
214.
The quenching of MePyln-0 fluorescence in alcohols is much
smaller than that of the syn form of Pyln-0, the species capable
of forming a double cyclic hydrogen bond. The fluorescence

2-(2′-Pyridyl)Indoles in Alcohols
J. Am. Chem. Soc., Vol. 118, No. 14, 1996 3515
Table 4. Experimental and Computed Ground and Excited States Parameters
E(S1)exp (103 cm-1
)
Θa
(deg)
E(S1)cal.
µ(S0)b
(D)
µ (S1)c
(D)
WN(S0)d
(eV)
∆WN(S1)e
(eV)
WNH(S0)f
(eV)
∆WNH(S1)e
nph
pi
(103 cm-1
)
(eV)
Pyln-0 (syn)
Pyln-0 (anti)
MePyln-0 (syn)
MePyln-0 (anti)
Pyln-1
Pyln-2
MePyln-2
Pyln-3
0
0
4
40
0
12
14
14
36
17
5
30.7
30.9
31.0
31.0
31.0
32.1
31.4
29.9
30.0
30.5
31.5
29.4
31.5
2.29(2.59)
3.10(4.07)
2.13(2.37)
2.90(4.16)
2.09(1.88)
2.46(2.27)
2.48(2.21)
2.71(2.71)
2.90(2.97)
(3.35)
1.23
4.57
0.50g
3.63g
2.47
2.55
2.51
2.07
1.59
2.69
2.04
-11.86
-11.81
-11.83g
-11.83g
-11.71
-11.74
-11.72
-11.77
-11.81
-11.69
1.05
1.15
-0.17g
0.18g
1.00
1.14
1.13
1.13
0.41
1.38
-5.57
-5.90
-0.33
-0.53
31.3
31.5
30.4
29.5
29.7
29.9
31.0
30.6
29.6
30.0
30.3
31.5
-5.68
-5.57
-0.32
-0.26
-5.42
-5.52
-5.26
-5.77
-0.25
-0.35
-0.23
-0.59
Pyln-4
Pyln-4j
2-PI
32.7
32.2
1.19(1.85)
a The dihedral angle between pyridine and indole rings. b Ground state dipole moment; first value obtained by MMX, second, by INDO/S.
c Calculated by INDO/S. d Effective valence electron potential of the pyridine type nitrogen atom obtained by INDO/S. e ∆W(S1) ) W(S1) - W(S0).
f WNH ) (1/2)(WN + WH) for the NH bond of the indole group. g Results for the second calculated excited state, see text. h Absorption maximum
in a nonpolar solvent (n-pentane). i Polar solvent (acetonitrile). j AM1 optimized ground state geometry.
Table 5. Ground State pKa Values
pKa
Pyln-0
Pyln-1
Pyln-2
Pyln-3
Pyln-4
3.8 ( 0.1
4.4 ( 0.1
4.4 ( 0.2
4.4 ( 0.2
4.3 ( 0.2
Figure 7. Plot of 1/φ vs 1/T for Pyln-3 in ethanol. Circles show the
experimental values, solid line, result of the fit to eq 7.
lifetime of MePyln-0 in n-propanol and butanol (about 0.6 and
0.7 ns at 293 K, respectively) is about 3-4 times lower than
that observed in aprotic solvents (2.5 ns in butyronitrile),
whereas for the syn form of Pyln-0 about 20-fold decrease is
observed (the lifetime is shorter than 100 ps, too short to be
accurately measured with the time resolution of our apparatus).
The extent of emission quenching of the methylated derivatives
is related to two factors. The first factor is the alcohol acidity:
methanol shows a much stronger effect than ethanol, n-propanol,
and butanol. In even more acidic alcohol, 1,1,1,3,3,3-hexafluo-
ropropanol-2, the fluorescence quantum yield drops below 0.001.
The second factor is the rigidity of the molecule: fluorescence
of the bridged, methylated derivative, MePyln-2 is not quenched
by alcohols.
Figure 8. Transient absorption of Pyln-2 in n-hexane (a) and butanol
(b). Top, absorption spectrum recorded 40 ns after excitation; bottom,
time profile of the absorption monitored at 435 nm.
bond and increased internal conversion. As shown below, a
large basicity increase upon excitation is determined experi-
mentally and predicted by calculations. Such a mechanism
previously has been observed for various azaaromatic com-
pounds.90 In the present case, comparison of the behavior of
MePyln-0 with the syn conformer of Pyln-0 shows the effect
of cooperativity, resulting in a much higher rate of quenching
in a molecule capable of forming two hydrogen bonds, both as
a donor and an acceptor.
Calculations of Ground and Excited State Acido-Basic
Properties. For a quantitative description of excitation-induced
changes of the hydrogen bonding abilities on both nitrogen
atoms, we have used the concept of effective valence electronic
Quenching of MePyln-0 fluorescence in alcohols may be
explained by the increased basicity of the pyridyl nitrogen in
the excited state, followed by the formation of the hydrogen

3516 J. Am. Chem. Soc., Vol. 118, No. 14, 1996
Herbich et al.
different values of Θ, the dihedral angle between the pyridine
and indole rings. Indeed, in Pyln-4 varying Θ from 36° (the
value obtained by MMX) to 17° (the result of AM1 opotimiza-
tion) results in a calculated pKa change of about one unit.
The calculations predict that the molecules with longer bridges
(Pyln-3 and Pyln-4) should have weaker acidities. Assuming
that acidities and hydrogen bonding donor properties are related,
this prediction is confirmed by our measurements of the alcohol
complex formation.
For pyridine, we obtain WN(S0) ) -11.66 eV. The calcula-
tions thus predict that pyridylindoles should be less basic than
pyridine and that the pKa difference should be about 1 unit.
This agrees nicely with experiment and may help explain the
lack of the ground state hydrogen bonding to the pyridine
nitrogen atom in pyridylindoles.
A huge increase in basicity upon excitationsabout 10 pKa
unitssis calculated for nearly all the compounds. Experimental
evaluation, based on the Fo¨rster cycle111 yields ∆pKa ≈ 7 for
Pyln-0 and Pyln-2.
Figure 9. The proposed mechanism of fluorescence quenching: Fl,
fluorescence; IC, internal conversion.
potentials.108-110 These parameters are defined as simple
functions of atomic charges and intermolecular distances. For
an atomic orbital µ on atom A
WA ) WA(q ) +
WA(q )
(8)
The results obtained for the WN(S1) in MePyln-0 are much
different from those computed for the other species. However,
this may be due to a computational artefact: In all compounds
except MePyln-0, the first excited singlet state is very well
represented by a single configuration corresponding to HOMO-
LUMO electron jump. The second and third calculated singlet
transitions have small oscillator strengths and different orbital
parentage. In MePyln-0, however, strong configurational
mixing occurs. The three lowest singlet states are computed
with similar, fairly large oscillator strengths, and with compa-
rable contributions of three different configurations. Thus, there
is no straightforward correlation between the S1 state of
MePyln-0 and the lowest excited state of other molecules. The
excited state parameters given in Table 4 for MePyln-0 actually
refer to the second calculated transition, because it resembles,
albeit very weakly, the S1 state in other pyridylindoles. Since
the absorption spectra of Pyln-0 and MePyln-0 are very similar,
we do not think that the puzzling results for the latter have real
physical meaning.
The calculations reproduce well the position of the first
absorption maximum and the relative shifts caused by bridging.
A small blue shift, observed upon passing from a nonpolar to
polar protic solvent indicates a decrease of the dipole moment
in the first excited state. This is also in agreement with
calculations. For 2PI, an increase of dipole moment upon
excitation, and thus a red shift in absorption in polar solvents
is predicted by calculations and experimentally confirmed (Table
4). Actually, for the anti isomer of Pyln-0 a red shift in polar
solvents should be expected, as the calculated dipole moment
in S1 is larger than in S0. Since the opposite is observed, it
seems that the syn conformer is the dominant form. MMX
calculations predict it to be more stable than the anti species
form by 7.1 kcal/mol; the AM1 method gives 3.1 kcal/mol.
For our present purpose, the most relevant result of the
calculations is the prediction of a large increase of basicity of
the pyridine nitrogen and of acidity of the indole proton upon
electronic excitation. Thus, a driving force for the excited state
proton transfer is provided.
∑
µ
µ
A
µ
B
B*A
WAµ (qA) ) -(a + bqA)3/2
WAµ (qB) ) γAB(qB-Zc)
(9)
(10)
qA and qB are the gross populations; the summation in (8) occurs
over all centers in the molecule except A; γAB is the repulsion
integral, Zc is the number of valence electrons in the neutral
atom B. The constants a and b are derived from the valence
state ionization potentials for the positive and negative ions.
It has been shown that the effective potentials WA2p are very
reliable in the prediction of ground108-110 and excited states109
acidities. The less negative values indicate higher basicity.
Relative pKa values of azaarenes are reproduced also in cases
where an approach based only on consideration of charge
densities on nitrogen atoms fails.
One can also formally define a bond potential,110 e.g., for
the NH group
W
NH ) (1/2)(WN + WH)
(11)
In this case, the increased acidity is accompanied by WNH
becoming more negative.
Using the INDO/S method, we have calculated the values of
the effective valence electron potentials for the pyridine nitrogen
and for the NH bond in the ground and the first excited singlet
state. The results are shown in Table 4. Ground state basicities
of the pyridine nitrogen and the acidities of the NH group are
predicted to be similar throughout the series. The difference
of 0.1 eV in WN(S0) corresponds to about 1 pKa unit. Thus, for
instance, pKa values for Pyln-1 and Pyln-0 should differ by
that amount, the former being more basic. We have experi-
mentally determined the pKa values for pyridylindoles using
spectrophotometric titration. The results are shown in Table
5. We obtain pKa ) 3.8 ( 0.1 for Pyln-0 and pKa ) 4.4 ( 0.1
for Pyln-1, in nice agreement with the computational results.
The pKa values of all bridged derivatives are practically the
same, while the calculations predict somewhat smaller values
for Pyln-3 and, in particular, for Pyln-4. However, the latter
two molecules may exist in several conformeric forms, with
Nature of the Quenching Process. Figure 8 presents the
transient absorption spectra of Pyln-2 taken in a nonpolar and
a protic solvent. A band centered around 430 nm is observed
in n-hexane, while there is practically no signal in butanol. The
decay of the transient absorption in n-hexane occurs in several
tens of nanoseconds, suggesting the assignment to a triplet-
(108) Spanget-Larsen, J. J. Chem. Soc., Perkin Trans. 2 1985, 417.
(109) Waluk, J.; Rettig, W.; Spanget-Larsen, J. J. Phys. Chem. 1988,
92, 6930. Herbich, J.; Grabowski, Z. R.; Wo´jtowicz, H.; Golankiewicz,
K. J. Phys. Chem. 1989, 93, 3439.
(111) Fo¨rster, T. Z. Elektrochem. Ber. Bunsenges. Physik. Chem. 1950,
54, 42.
(110) Spanget-Larsen, J. J. Phys. Org. Chem. 1995, 8, 496.

2-(2′-Pyridyl)Indoles in Alcohols
J. Am. Chem. Soc., Vol. 118, No. 14, 1996 3517
triplet transition (fluorescence lifetime is 1.7 ns97). Lack of a
signal in alcohol indicates that the triplet population efficiency
was greatly reduced and that the quenching process does not
involve enhancement of the intersystem crossing to the triplet
state. Thus, one can exclude the fluorescence quenching due
to solvent-induced shifts of the singlet and triplet levels with
respect to each other (in particular, the nπ* states with respect
to those of ππ* character).
Two other processes may be responsible for the decrease of
fluorescence upon hydrogen bonding. One is the excited state
double proton transfer. Such a process, leading to the formation
of a nonfluorescent tautomeric species, has been proposed to
explain the photophysical behavior of 2-(2′-pyridyl)benzimid-
azole, a molecule structurally similar to pyridylindoles.71,72
However, in our case the calculations performed for the tautomer
predict that the energy of the transition to the first excited singlet
state is about 8500 cm-1 lower than that of the normal form,
and the oscillator strength is about six times weaker. We should
thus be able to observe the tautomeric fluorescence in alcohols
in the region of 16 000-17 000 cm-1, unless it is very efficiently
quenched. Only an extremely weak emission, too weak to
measure its excitation spectrum is detected in this region. Also,
the absence of transient absorption in alcohol indicates that no
significant population of tautomeric species is formed.
There is no doubt that the quenching involves the double
hydrogen bond and some form of proton motion. We therefore
propose that the mechanism of radiationless depopulation
involves an efficient internal conversion process that is activated
by the movement of hydrogen-bonded protons. There are two
possible sequences for such an event: (i) efficient internal
conversion occurring during proton translocation, before reach-
ing the tautomeric conformation, and (ii) rapid deactivation of
the phototautomer, faster than the rate of its creation. While
both would lead to the same final resultsefficient primary
fluorescence quenching and lack of the tautomeric emissionsthey
are, in principle, distinguishable kinetically. In the former case,
the rate of the process would correspond to internal conversion
itself, while for the latter it is the slower step, the proton transfer,
which is rate-limiting. The values of the preexponential factor
obtained from the temperature dependence of fluorescence are
of the order of 1013 s-1 (Table 3), typical of internal conversion.
The analogous factors obtained for azaindole, where the decay
of the initially excited state is due to double proton transfer,
are 1-2 orders of magnitude smaller.42 Therefore, it seems
that the mechanism (i) is dominant, and the excited molecule
becomes deactivated sometimes “on the way” between the initial
structure and the phototautomer.
also occurs with alcohol molecules serving as bridges between
proton donor and acceptor centers. The interpretation of the
unusual isotope effect is based on a two-step model. According
to this scheme, the excited state proton transfer must be preceded
by the attainment of a favorable structure of the alcohol complex,
which is equivalent to a solvent reorientation process. The
reaction may be depicted as
k2
98 C
A {kk } B
1
-1
where A, B, and C correspond to the primarily excited complex,
the structure obtained after solvent reorientation, and the
phototautomer, respectively. The rate constants k1 and k-1 refer
to solvent rearrangement, and k2 describes the deactivation of
the “prepared” structure involving proton transfer. Thus, k2
should depend on deuterium substitution, whereas k1 and k-1
should not.
For k1 , k-1, i.e., for the case where A is the dominant
species, the decay rate of A may be written as
kobs ) k1k2/(k-1 + k2)
(12)
In the high temperature region, where the solvent reorientation
is fast, the rate limiting step would correspond to proton transfer,
while in the other extreme the reaction is controlled by the
solvent rearrangement. The activation energy of the latter
should be correlated with the activation energy of viscous flow
and not dependent on the isotope substitution.
Fitting of the experimentally observed decay rate of A to eq
12 requires, in general, the determination of six parameters (three
preexponential factors and three activation energies) and may
only be attempted after making several assumptions, not
necessarily well-justified. Instead of trying this procedure, we
note that the present results are similar to the ones obtained for
7-azaindole and 7-hydroxyquinoline. In excited pyridylindoles
the first step, solvent reorientation, should involve expulsion
of one or more alcohol molecules from the complex before a
1:1 structure can be achieved. A driving force for such a process
is provided by the excited state increase of basicity on the
pyridine nitrogen atom which, as our results suggest, is not
hydrogen bonded in the ground state. If both hydrogen bonding
centers of pyridylindoles were engaged in the hydrogen bonds
already in the ground state, each with a different alcohol
molecule, one would predict the increase in the strength of both
bonds upon excitation rather than breaking one of them.
The second step in our case actually involves two processes,
proton transfer and internal conversion, of which the first
provides the driving force for the other. While the rate of both
should not be strongly temperature-dependent, the population
of molecules able to undergo this process is solvent-controlled
and rapidly decreases upon increasing the viscosity or lowering
of temperature. Thus, the radiative properties of pyridylindoles
are fully recovered in alcohol glasses, which shows that the
ground state structure of the complexes is not prone to
tautomerization.
It thus seems that the mechanism of excited state proton
transfer in hydrogen-bonded solvates can now be generalized
to encompass various species. Attainment of a suitable
conformation is the prerequisite for a rapid phototautomerization.
This process is often associated with the alcohol rearrangement
and thus provides an example of a solvent-controlled reaction.
There are cases, however, when an appropriate structure of the
complex already exists in the ground state. In such a situation,
the intermolecular excited state proton transfer can be extremely
rapid, even faster than the intramolecular process. Excited state
proton transfer in methanol solution of 3-hydroxyflavone is
Discussion
The results of ground and excited state studies allow us to
propose the cycle of events occurring after excitation of
pyridylindoles in alcohols. In the bulk alcoholic solvent, 1:n
complexes are predominant, where n, the number of alcohol
molecules is greater than one. After excitation, the complexes
may rearrange to form a 1:1 cyclic structure, which undergoes
a double proton transfer reaction. During this process, a rapid
internal conversion channel brings the complex to the ground
state.
Crucial for the understanding of the excited state phenomena
was the observation that deuterium substitution in the hydroxylic
group of the alcohol does not lead to a change of activation
energy. The isotope effect is expressed only through a
difference in the preexponential factor. The same behavior has
been recently observed in alcohol solutions of 7-azaindole42 and
7-hydroxyquinoline,63,67 where a double phototautomerization

3518 J. Am. Chem. Soc., Vol. 118, No. 14, 1996
Herbich et al.
faster than 125 fs, whereas the rate of the intramolecular reaction
occurring in a nonpolar solvent was determined to be 240 (
50 fs.82 In 1:1 cyclic hydrogen bonded complexes of 7-azain-
dole with carboxylic acid, the rate of photoinduced double
proton transfer is so fast that only the tautomer emission is
observed.52 This observation is in contrast to the same reaction
occurring in alcohols, where the much slower dynamics involves
solvent reorganization.35,42
We are currently investigating systems related to pyridylin-
doles with more than two hydrogen bonding centers. Prelimi-
nary results for dipyridocarbazole indicate that the quenching
of fluorescence by alcohols is even more efficient than in
pyridylindoles. A related line of research involves quenching
of emission based on coupled photoinduced proton and electron
transfer reactions. The results, to be published shortly, depict
the quenching properties of azaaromatic hydrogen bond and
electron acceptors.104
Acknowledgment. Technical assistance from Mrs. A.
Zielinska, Mrs. G. Orzanowska, Mr. R. Rodowski, and Mr. J.
Karpiuk is deeply appreciated. R.P.T. thanks the Robert A.
Welch Foundation and the National Science Foundation (CHE-
9224686) for financial support.
JA952797D